Datasets:

Hack90

/

europe_pmc_articles_part_2

like 1

Bidirectional links between HIV and intimate partner violence in pregnancy: implications for prevention of mother-to-child transmission

<sec id="st1"><title>Introduction</title><p>Prevention of mother-to-child transmission (PMTCT) has the potential to eliminate new HIV infections among infants. Yet in many parts of sub-Saharan Africa, PMTCT coverage remains low, leading to unacceptably high rates of morbidity among mothers and new infections among infants. Intimate partner violence (IPV) may be a structural driver of poor PMTCT uptake, but has received little attention in the literature to date.</p></sec><sec id="st2"><title>Methods</title><p>We conducted qualitative research in three Johannesburg antenatal clinics to understand the links between IPV and HIV-related health of pregnant women. We held focus group discussions with pregnant women (<italic>n</italic>=13) alongside qualitative interviews with health care providers (<italic>n</italic>=10), district health managers (<italic>n</italic>=10) and pregnant abused women (<italic>n</italic>=5). Data were analysed in Nvivo10 using a team-based approach to thematic coding.</p></sec><sec id="st3"><title>Findings</title><p>We found qualitative evidence of strong bidirectional links between IPV and HIV among pregnant women. HIV diagnosis during pregnancy, and subsequent partner disclosure, were noted as a common trigger of IPV. Disclosure leads to violence because it causes relationship conflict, usually related to perceived infidelity and the notion that women are &#x0201c;bringing&#x0201d; the disease into the relationship. IPV worsened HIV-related health through poor PMTCT adherence, since taking medication or accessing health services might unintentionally alert male partners of the women's HIV status. IPV also impacted on HIV-related health via mental health, as women described feeling depressed and anxious due to the violence. IPV led to secondary HIV risk as women experienced forced sex, often with little power to negotiate condom use. Pregnant women described staying silent about condom negotiation in order to stay physically safe during pregnancy.</p></sec><sec id="st4"><title>Conclusions</title><p>IPV is a crucial issue in the lives of pregnant women and has bidirectional links with HIV-related health. IPV may worsen access to PMTCT and secondary prevention behaviours, thereby posing a risk of secondary transmission. IPV should be urgently addressed in antenatal care settings to improve uptake of PMTCT and ensure that goals of maternal and child health are met in sub-Saharan African settings.</p></sec>

<contrib contrib-type="author"><name><surname>Hatcher</surname><given-names>Abigail M</given-names></name><xref ref-type="corresp" rid="cor1">&#x000a7;</xref><xref ref-type="aff" rid="AF0001">1</xref><xref ref-type="aff" rid="AF0002">2</xref></contrib><contrib contrib-type="author"><name><surname>Woollett</surname><given-names>Nataly</given-names></name><xref ref-type="aff" rid="AF0001">1</xref></contrib><contrib contrib-type="author"><name><surname>Pallitto</surname><given-names>Christina C</given-names></name><xref ref-type="aff" rid="AF0003">3</xref></contrib><contrib contrib-type="author"><name><surname>Mokoatle</surname><given-names>Keneuoe</given-names></name><xref ref-type="aff" rid="AF0001">1</xref></contrib><contrib contrib-type="author"><name><surname>St&#x000f6;ckl</surname><given-names>Heidi</given-names></name><xref ref-type="aff" rid="AF0004">4</xref></contrib><contrib contrib-type="author"><name><surname>MacPhail</surname><given-names>Catherine</given-names></name><xref ref-type="aff" rid="AF0001">1</xref><xref ref-type="aff" rid="AF0005">5</xref></contrib><contrib contrib-type="author"><name><surname>Delany-Moretlwe</surname><given-names>Sinead</given-names></name><xref ref-type="aff" rid="AF0001">1</xref></contrib><contrib contrib-type="author"><name><surname>Garc&#x000ed;a-Moreno</surname><given-names>Claudia</given-names></name><xref ref-type="aff" rid="AF0003">3</xref></contrib>

Journal of the International AIDS Society

<sec sec-type="intro" id="S0001"><title>Introduction</title><p>Prevention of mother-to-child transmission (PMTCT) has reduced new infant HIV infections from an estimated 32% in the absence of treatment [<xref rid="CIT0001" ref-type="bibr">1</xref>, <xref rid="CIT0002" ref-type="bibr">2</xref>], to as low as 1% [<xref rid="CIT0003" ref-type="bibr">3</xref>, <xref rid="CIT0004" ref-type="bibr">4</xref>]. However, major gaps in achieving PMTCT coverage remain. In 21 priority African countries, PMTCT coverage is estimated to be 65% [<xref rid="CIT0005" ref-type="bibr">5</xref>]. A recent meta-analysis in low- and middle-income settings suggests that while 75% of pregnant women adhere to antiretroviral therapy (ART) during pregnancy, only 53% maintain adequate adherence levels in the postpartum period [<xref rid="CIT0006" ref-type="bibr">6</xref>]. Ensuring PMTCT adherence is crucial, particularly as countries increasingly move towards &#x0201c;Option B+,&#x0201d; a policy that offers immediate, lifelong treatment for pregnant women living with HIV [<xref rid="CIT0007" ref-type="bibr">7</xref>].</p><p>Many structural drivers influence PMTCT uptake and adherence. The literature has noted that structural factors such as stigma [<xref rid="CIT0008" ref-type="bibr">8</xref>&#x02013;<xref rid="CIT0013" ref-type="bibr">13</xref>], poverty [<xref rid="CIT0011" ref-type="bibr">11</xref>] and transport costs [<xref rid="CIT0010" ref-type="bibr">10</xref>] inhibit women's ability to adhere to PMTCT. Another key structural factor shaping access and adherence to PMTCT may be intimate partner violence (IPV). Fear and experience of IPV influence pregnant women's decisions to take up HIV services [<xref rid="CIT0014" ref-type="bibr">14</xref>, <xref rid="CIT0015" ref-type="bibr">15</xref>], and anticipated violence is associated with declines in HIV testing among pregnant women [<xref rid="CIT0012" ref-type="bibr">12</xref>, <xref rid="CIT0016" ref-type="bibr">16</xref>&#x02013;<xref rid="CIT0022" ref-type="bibr">22</xref>]. A history of physical or sexual violence decreases the likelihood of HIV-positive women using ART when medically eligible [<xref rid="CIT0023" ref-type="bibr">23</xref>, <xref rid="CIT0024" ref-type="bibr">24</xref>], and those who experience abuse are more likely to miss clinic visits and delay linkage to care [<xref rid="CIT0025" ref-type="bibr">25</xref>].</p><p>Little research to date has explored the association between IPV and PMTCT. In one qualitative study in South Africa, IPV was described as a common barrier to ART adherence in pregnancy [<xref rid="CIT0011" ref-type="bibr">11</xref>]. Healthy intimate partner relationships improve PMTCT uptake: male involvement in antenatal care predicted better adherence to nevirapine in one South African study [<xref rid="CIT0026" ref-type="bibr">26</xref>]; male antenatal attendance halved the risk of MTCT in a Kenyan study, an association that persisted after controlling for maternal viral loads [<xref rid="CIT0027" ref-type="bibr">27</xref>].</p><p>Using qualitative research methodology, we explored IPV as a potential structural driver of HIV-related health among pregnant women. This research aimed to contribute to literature suggesting that structural drivers shape the health and well-being of those already living with HIV, and may pose barriers to uptake of proven prevention strategies.</p></sec><sec sec-type="methods" id="S0002"><title>Methods</title><p>We conducted qualitative research to explore the links between IPV and HIV-related health among pregnant women and service providers in Johannesburg, South Africa. This research was a portion of a larger formative study, intended to help our team design an intervention to address IPV in pregnancy. In this setting, an estimated 29% of pregnant women are HIV positive [<xref rid="CIT0028" ref-type="bibr">28</xref>] and between 25 and 35% experience physical or sexual IPV in the 12 months leading up to pregnancy [<xref rid="CIT0029" ref-type="bibr">29</xref>&#x02013;<xref rid="CIT0032" ref-type="bibr">32</xref>].</p><sec id="S0002-S20001"><title>Conceptual framework</title><p>To explore the relationship between IPV and HIV-related health of pregnant women, we used an adapted socio-ecological conceptual framework (<xref ref-type="fig" rid="F0001">Figure 1</xref>), which posits that broader structural factors and relationship characteristics influence a woman's HIV-related health [<xref rid="CIT0033" ref-type="bibr">33</xref>]. This type of approach has been embraced by social scientists, who note that broader social and societal factors shape how women are able to adhere to ART [<xref rid="CIT0034" ref-type="bibr">34</xref>] and the extent to which they experience IPV [<xref rid="CIT0035" ref-type="bibr">35</xref>]. A socio-ecological framework highlights that the structural context influences the conditions and health outcomes of both IPV and HIV.</p><fig id="F0001" position="float"><label>Figure 1</label><caption><p>Conceptual framework.</p></caption><graphic xlink:href="JIAS-17-19233-g001"/></fig></sec><sec id="S0002-S20002"><title>Data collection</title><p>We conducted an exploratory qualitative study using in-depth interviews (IDIs) and focus group discussions (FGDs) with a wide range of stakeholders with the potential to take part in, deliver, or scale-up an intervention for violence in pregnancy. Participants included pregnant women, pregnant women experiencing IPV, health workers, non-governmental organizations, community leaders and policy makers (<xref ref-type="table" rid="T0001">Table 1</xref>).</p><table-wrap id="T0001" position="float"><label>Table 1</label><caption><p>Data collection methods</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1">Participant group</th><th align="center" rowspan="1" colspan="1">Group size</th><th align="center" rowspan="1" colspan="1">Method</th><th align="center" rowspan="1" colspan="1">Sampling</th><th align="center" rowspan="1" colspan="1">Example participants</th><th align="center" rowspan="1" colspan="1">Topics</th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">Pregnant women at ANC</td><td align="center" rowspan="1" colspan="1"><italic>n</italic>=13 women in 4 FGDs</td><td align="left" rowspan="1" colspan="1">Focus group discussions</td><td align="left" rowspan="1" colspan="1">Convenience</td><td align="center" rowspan="1" colspan="1">&#x02013;</td><td align="left" rowspan="1" colspan="1">Social and structural drivers of IPV; types of IPV in pregnancy; patterns of help seeking and available community resources for violence and HIV; barriers to disclosing IPV; receptivity to an antenatal intervention</td></tr><tr><td align="left" rowspan="1" colspan="1">Pregnant abused women</td><td align="center" rowspan="1" colspan="1"><italic>n</italic>=5</td><td align="left" rowspan="1" colspan="1">Semi-structured interviews</td><td align="left" rowspan="1" colspan="1">Convenience</td><td align="center" rowspan="1" colspan="1">&#x02013;</td><td align="left" rowspan="1" colspan="1">Existing needs and concerns of abused women; patterns of help seeking and available community resources for violence; links between IPV and HIV; receptivity to an antenatal intervention</td></tr><tr><td align="left" rowspan="1" colspan="1">Policy makers</td><td align="center" rowspan="1" colspan="1"><italic>n</italic>=10</td><td align="left" rowspan="1" colspan="1">Semi-structured interviews</td><td align="left" rowspan="1" colspan="1">Purposive</td><td align="left" rowspan="1" colspan="1">Department of Health managers, academic experts</td><td align="left" rowspan="1" colspan="1">Types of IPV in pregnancy; current health sector response to IPV; potential integration with HIV activities, including PMTCT</td></tr><tr><td align="left" rowspan="1" colspan="1">Health providers</td><td align="center" rowspan="1" colspan="1"><italic>n</italic>=8</td><td align="left" rowspan="1" colspan="1">Semi-structured interviews</td><td align="left" rowspan="1" colspan="1">Purposive</td><td align="left" rowspan="1" colspan="1">Doctors, nurses, lay counsellors in antenatal clinics</td><td align="left" rowspan="1" colspan="1">Types of IPV in pregnancy; knowledge and practice responding to IPV; receptivity of health workers to antenatal intervention; existing capacity in clinic</td></tr><tr><td align="left" rowspan="1" colspan="1">Non-governmental organizations</td><td align="center" rowspan="1" colspan="1"> <italic>n</italic>=6</td><td align="left" rowspan="1" colspan="1">Semi-structured interviews</td><td align="left" rowspan="1" colspan="1">Purposive</td><td align="left" rowspan="1" colspan="1">Shelters, police, counselling services</td><td align="left" rowspan="1" colspan="1">Psycho-social, legal and other needs of abused women; referral options for women living with IPV</td></tr><tr><td align="left" rowspan="1" colspan="1">Community leaders</td><td align="center" rowspan="1" colspan="1"> <italic>n</italic>=4</td><td align="left" rowspan="1" colspan="1">Semi-structured interviews</td><td align="left" rowspan="1" colspan="1">Convenience</td><td align="left" rowspan="1" colspan="1">Pastors, neighbourhood representatives, traditional healer</td><td align="left" rowspan="1" colspan="1">Community factors that support or prevent women from seeking IPV assistance during pregnancy</td></tr></tbody></table></table-wrap><p><italic>Pregnant women seeking antenatal care</italic> from two antenatal clinics were recruited for four FGDs (a total of <italic>n</italic>=13 women participated). Women were given group information about the study while they waited in queue for a clinic appointment. All FGDs were conducted in a private room in the clinic, led by trained moderators who were the same sex as participants and fluent in multiple local languages (Sotho, Zulu, Tswana). Semi-structured discussion guides explored several topics (<xref ref-type="table" rid="T0001">Table 1</xref>). Discussions were audio-recorded after obtaining participants&#x02019; permission and signing an informed consent form. The discussion groups lasted for about 1 hour and 30 minutes, and women were reimbursed R50 (US $6). Because of the nature of focus groups, additional confidentiality measures were implemented: during the informed consent process, researchers explained that questions about women's individual experiences of violence or HIV would not be asked, but rather the discussion would address the issue as observed in the community.</p><p> <italic>Pregnant women who were experiencing IPV</italic> were identified during the FGDs. Trained researchers explained that those women who had personal experience of IPV and were interested in participating in IDIs could approach the research team outside of the information giving session and privately indicate their interest in taking part in an interview. The interviews (<italic>n</italic>=5) took place in a private room at the clinics while the pregnant women were still waiting to be seen by clinic staff. As shown in <xref ref-type="table" rid="T0001">Table 1</xref>, the topics explored through structured interview guides were more focused on IPV-related help seeking and the relationship between IPV and HIV. On average, these interviews lasted about 60 minutes.</p><p>In depth interviews with <italic>Other Key Stakeholders</italic> were led by the research team and covered similar topics. This group comprised policy makers (<italic>n</italic>=10), health workers (<italic>n</italic>=8), non-governmental organizations (<italic>n</italic>=6) and community-based organizations (<italic>n</italic>=4). Stakeholder interviews focused on service provision and asked questions about available resources for women experiencing IPV. Some anecdotes of cases were shared, but this was not the main rationale for these interviews.</p><p>Several steps were taken to ensure confidentiality and provide additional support for participants during the research. In keeping with ethical considerations of researching IPV in pregnancy, all researches were conducted based on the World Health Organization's guidance on ethical and safety considerations in researching violence against women [<xref rid="CIT0036" ref-type="bibr">36</xref>]. Study staff were trained to describe research as the &#x0201c;social barriers&#x0201d; to use health services in the community, so as to reduce any undue risk associated with participating in a violence-related study. All participants were offered an information sheet containing contact information of local resources (counselling, legal advice and health care). Given the high prevalence of IPV in South Africa, it was likely that participants in categories other than &#x0201c;pregnant and experiencing IPV&#x0201d; category had experienced or witnessed IPV. If any individual demonstrated a need for additional assistance, that individual was offered an opportunity to speak to someone about his or her experience of IPV, and given referrals to organizations that could assist him or her. However, no participants asked for this referral during the course of the formative research.</p><p>All participation in this formative research was sought on the basis of informed consent and good clinical practice guidelines. Ethics approval was obtained by the World Health Organization (WHO A65780) and University of the Witwatersrand (HREC M110832).</p></sec><sec id="S0002-S20003"><title>Data analysis</title><p>The interview and FGD data were transcribed verbatim in the language in which they were conducted and, as necessary, translated from the local language (Sotho, Zulu, Tswana) into English by professional translators. To ensure accurate translation, each transcript was reviewed by a researcher, and queries were resolved through discussions among the researchers via phone or email. All identifying information about the participant or clinic setting was removed and transcripts were saved by a file name with no personal details.</p><p>Data were managed in QSR Nvivo 10, a qualitative analysis software package, following a two-day qualitative management and analysis training of the research team. Members of the research team collaboratively built an analytical framework of broad codes by creating a &#x0201c;start list&#x0201d; of possible themes and building upon the research questions. Each broad code, or wide thematic basket of ideas [<xref rid="CIT0037" ref-type="bibr">37</xref>], was applied to each transcript by two researchers using NVivo. The research team then held a series of meetings to collectively develop &#x0201c;fine codes&#x0201d; using an inductive approach &#x02013; deriving meaning from the data itself rather than imposing pre-formed ideas [<xref rid="CIT0038" ref-type="bibr">38</xref>]. Fine codes were developed by printing out a full set of excerpts (from each database) related to each code and identifying sub-themes emerging from the data.</p></sec></sec><sec sec-type="results" id="S0003"><title>Results</title><p>We found qualitative evidence of strong bidirectional links between IPV and HIV among pregnant women. Here, we present a conceptual framework (<xref ref-type="fig" rid="F0002">Figure 2</xref>) for understanding the ways in which IPV is related to HIV-related health of pregnant women.</p><fig id="F0002" position="float"><label>Figure 2</label><caption><p>IPV and HIV-related health among pregnant women.</p></caption><graphic xlink:href="JIAS-17-19233-g002"/></fig><sec id="S0003-S20001"><title>Pathway 1: HIV diagnosis leads to IPV via partner disclosure</title><p>HIV diagnosis during pregnancy was noted to be a trigger of IPV. One pregnant woman described how severe violence began following disclosure of her HIV-positive status during pregnancy:<disp-quote><p>He started telling me things, hurting me emotionally, telling me that I'm a fool, and stupid, I'm an idiot. And then he strangled me, That's when it started &#x02026; Maybe it's pregnancy, I don't know. I told him that I am HIV positive, so I don't know if that's what made him to do all these things. &#x02013; Pregnant abused woman 1</p></disp-quote> </p><p>HIV may lead to violence because it causes relationship conflict during the disclosure process. Usually, the conflict is related to perceived infidelity and blaming women for &#x0201c;bringing&#x0201d; the disease into the relationship:<disp-quote><p>Yes, if you're HIV positive, you start blaming each other. Because maybe the husband will be saying the wife brought it. So sometime, there's a connection [between HIV and violence] because you end up blaming each other. &#x02013; Pregnant woman, FGD 3</p></disp-quote> </p><p>Because HIV testing is coupled with antenatal care, women often learn of their HIV status in a clinic during access to health care during pregnancy. Within this health care context, women bear the brunt of disclosure to partners, who tend to use women's status as a &#x0201c;proxy&#x0201d; for their own.</p><p>In addition to physical violence, pregnant women described experiencing emotional abuse and abandonment following disclosure of HIV to a partner:<disp-quote><p>I have a sister, she was pregnant, &#x02026; then she came to be tested. When she tested she found out she is positive, and when she told her boyfriend everything turned around. And there was violence at home. He started coming late and when she started asking for things for her and the baby, he started to react badly up until he ended up leaving her. &#x02013; Pregnant woman, FGD 2</p></disp-quote> </p><p>Within a context where women fear violence, blame and abandonment, it is perhaps not surprising that many pregnant women chose not to disclose their HIV status to partners. Several pregnant women spoke about the fear of partner disclosure when women live in violent relationships:<disp-quote><p>Women who are in this abusive relationship, they do get HIV and they are scared what their partner will say. &#x02013; Pregnant abused woman 1</p></disp-quote> </p><p>Health workers talked about how women in violent relationships would be hesitant to disclose their status to partners:<disp-quote><p>When you counsel them &#x02026; after they have tested positive and when you have to issue the treatment she'll be saying, &#x02018;I am not going to disclose. I mustn't take this, I must hide it&#x02019;. Then you find out is it a problem for her to disclose because there's some emotional abuse or physical abuse from the partner. &#x02013; Female Health Worker 4</p></disp-quote> </p><p>Thus, fear of partner disclosure may be an early warning sign that pregnant women are in violent or unsupportive relationships and require additional assistance during antenatal care.</p></sec><sec id="S0003-S20002"><title>Pathway 2: IPV worsens HIV-related health via non-adherence</title><p>For women in violent relationships, adherence to PMTCT services was challenging, since taking medication or accessing health services might unintentionally alert male partners of their HIV status.</p><p>Health workers noted that non-adherence also served as warning sign that HIV-positive patients were in a violent relationship:<disp-quote><p> It was the very same patient that you had told she was HIV positive that was scared to go and disclose to their partner. It is the very same patient that will default on their medication because their partner does not know that they are taking the medication. &#x02013; Policy Maker 9</p></disp-quote> </p><p>In the antenatal clinic, an HIV diagnosis in the context of living with violence may cause patients to default on clinic visits:<disp-quote><p>But in your normal facility it is a bit difficult to avoid losing patients. I think we do. Especially the mere fact that you say to a patient, &#x0201c;you are HIV positive.&#x0201d; And this is a patient who is facing domestic violence! Some will just disappear. &#x02013; Policy maker 3</p></disp-quote> </p><p>Thus, the fear of being identified by a male partner as being HIV positive may preclude women from returning to clinic services that are essential for their health. While no participants mentioned this directly, it is important to note that non-adherence to PMTCT regimens greatly increases the risk that pregnant or breastfeeding women will transmit HIV to the infant.</p></sec><sec id="S0003-S20003"><title>Pathway 3: IPV worsens HIV-related health via mental health</title><p>Declines in mental health were noted in women experiencing IPV in pregnancy. In response to persistent violent relationships, women described internalizing the abuse and assuming that they had done something wrong to deserve it:<disp-quote><p>He used to beat her while she was pregnant. She just accepted it, and sometime she'd blame herself. Saying maybe I'm the one who's wrong that's why he's beating me. &#x02013; Pregnant woman, FGD 1</p></disp-quote> </p><p>Although IPV is associated with common mental health disorders in pregnancy, few patients or providers recognized these as having clinical implications. Most health providers equated mental health to severe cases of psychopathology and said they rarely encountered mental health disorders. For example, one health worker only considered mental health in relation to bipolar depression and pharmacologically treated patients:<disp-quote><p>Mental health, yes, I remember we've had three that we already on treatment, and will tell you, I have a bipolar patient. &#x02013; Health worker 8</p></disp-quote> </p><p>We found that health workers often fail to notice the mental health dynamics of IPV in pregnancy, choosing instead to focus on physical health sequelae of pregnancy. For example, one health provider was asked about stress related to IPV, but responded only in terms of how stress impacts hypertension while ignoring the relevant impact on a woman's mental health:<disp-quote><p>Stress is one of the predisposing factors to the development of hypertension. So it is still there, this stress, but as a predisposing factor. Sometimes because of the pregnancy itself, you can develop hypertension of pregnancy. &#x02013; Health worker 6</p></disp-quote> </p><p>This tendency towards equating mental health with severe illness may be related to the lack of capacity within South Africa's public health system. As one policy maker explained, in overlooking mental health issues, current health systems make it unlikely that patients will receive the crucial support that they require:<disp-quote><p>No one has time for mental health because there are so many other crises in the health system that need to be addressed, that are much more manifested. So that means that things like depressive disorder or mental health disorders, they're not addressed &#x02013; including partner depression, mental health and abuse and all of that. And people are not really encouraged to go and get support that they require. &#x02013; Policy Maker 9</p></disp-quote> </p><p>The notion of overlooking mental health is illustrated in an interview with one pregnant, abused woman when she described severe physical violence as leading to a state of being &#x0201c;a little depressed&#x0201d;:<disp-quote><p>Interviewer (I): Are you enjoying your pregnancy so far?</p><p>Participant (P): Being honest, a little depressed but I'm enjoying it.</p><p>I: Ok, so the depression is from what, if I may ask?</p><p>P: From the father of the baby. We are having problems.</p><p>I: What did he do, if you don't mind telling me.</p><p>P: He strangled me and then he let his cousin beat me up. &#x02013; Pregnant abused woman 1</p></disp-quote> </p><p>Not everyone in our sample ignored the impact of mental health on a woman's health and wellbeing. For example, poor mental health had concomitant effects on physical health for one HIV-positive participant, who described &#x0201c;going low&#x0201d; emotionally because of violence, and thereafter feeling worse physically:<disp-quote><p>I'm HIV positive and I'm in this domestic violence. And if you are HIV positive and then you have a partner who is abusing you emotionally &#x02026; or physically hits you, people can't talk. Maybe you can go low, maybe you can go sick. &#x02013; Pregnant abused woman 3</p></disp-quote> </p></sec><sec id="S0003-S20004"><title>Pathway 4: IPV leads to secondary HIV risk via lack of relationship control</title><p>IPV led to secondary HIV risk when women were in relationships with forced sex or without power to negotiate condom use.<disp-quote><p>When we are in relationships where our partner is abusive, sometimes we can't even negotiate things like using the condom. Let's say, for instance, you know that your partner is the kind of person that has other girlfriends, but because he uses power over you, you can't negotiate those things. &#x02013; Pregnant woman, FGD 4</p></disp-quote> </p><p>Male partners used their control over the relationship to dictate the terms and timing of sexual activity. In one instance, a FGD revealed a story about a newly diagnosed HIV-positive woman whose partner insisted that she have sex without condoms:<disp-quote><p>There's a friend of mine that was tested alone and she had a lot of problems. The man said, I'm not HIV positive, so I'm not going to test. So the man forced her to sleep with him without a condom. And that man said &#x02018;No! Why? We've been sleeping without a condom, but because today you went to the clinic, you're telling me we've to use a condom?&#x02019; &#x02013; Pregnant woman, FGD 1</p></disp-quote> </p><p>Pregnant women described balancing risks to physical safety (absence of physical harm to themselves or foetus) with health risks (of onwards HIV transmission to partners). They described making compromises between protecting themselves and the foetus and protecting themselves and partners from sexually transmitted infections:<disp-quote><p>If you are not compromising at all and you start saying &#x0201c;let's use condom,&#x0201d; he'll start having questions. Some things are better left unsaid, just for the safety part of it. &#x02013; Pregnant woman, FGD 2</p></disp-quote> </p><p>Many preferred staying silent on condom negotiation, in order to stay physically safe during pregnancy.</p></sec></sec><sec sec-type="discussion" id="S0004"><title>Discussion</title><p>We found that IPV and HIV-related health were connected concerns in the lives of pregnant women in Johannesburg. IPV and HIV seemed to have distinct pathways linking them to one another within the context of pregnancy. The initial HIV disclosure could serve as a trigger for violence in pregnancy. IPV, in turn, worsened HIV-related health through key pathways of lack of adherence and poor mental health. Finally, the experience of IPV led to secondary transmission risk behaviours &#x02013; both in terms of vertical transmission due to PMTCT non-adherence or secondary transmission due to risky sex.</p><p>According to our participants, IPV shapes HIV-related health outcomes among pregnant women primarily because it leads to non-adherence. While the effect of IPV on adherence has been confirmed in small studies in the United States [<xref rid="CIT0039" ref-type="bibr">39</xref>&#x02013;<xref rid="CIT0043" ref-type="bibr">43</xref>], this association is yet to be explored among pregnant women. Pregnant and postpartum women are a crucial population within which to understand IPV and adherence, since non-adherence leads not only to morbidity and mortality of the woman but also to risk of onwards HIV transmission to her infant [<xref rid="CIT0003" ref-type="bibr">3</xref>, <xref rid="CIT0004" ref-type="bibr">4</xref>]. Antenatal care provides a crucial moment to enable adherence, since a pregnant woman accesses the health system routinely and this is when many are first diagnosed with HIV.</p><p>Poor adherence among pregnant women may relate to challenges around partner disclosure [<xref rid="CIT0044" ref-type="bibr">44</xref>]. In a recent systematic review of PMTCT, partner disclosure was associated with poor PMTCT uptake in a majority of both quantitative (6 of 9) and qualitative (17 of 24) studies [<xref rid="CIT0045" ref-type="bibr">45</xref>]. We found that partner disclosure following HIV diagnosis in pregnancy led to enacted or feared violence. This aligns with extant literature, which suggests that fear of new or continued IPV may lead women to avoid disclosure of their status to male partners [<xref rid="CIT0011" ref-type="bibr">11</xref>]. In one Nigerian study among HIV-positive pregnant women, the prevalence of IPV was 17% before HIV testing and increased to 63% after testing for HIV and disclosing status [<xref rid="CIT0046" ref-type="bibr">46</xref>]. A Zimbabwean study showed that the risk of IPV in pregnancy was greatest among those women testing positive for HIV in antenatal care [<xref rid="CIT0047" ref-type="bibr">47</xref>]. Non-disclosure among pregnant women is a health risk in its own right, since it poses a risk for sexual transmission of HIV if the male partner is still HIV negative [<xref rid="CIT0048" ref-type="bibr">48</xref>&#x02013;<xref rid="CIT0051" ref-type="bibr">51</xref>] and may have an impact on the implementation of PMTCT [<xref rid="CIT0052" ref-type="bibr">52</xref>].</p><p>A related but distinct pathway linking IPV to PMTCT uptake may be mental health. A growing body of literature shows that IPV leads to depression and anxiety among pregnant women [<xref rid="CIT0029" ref-type="bibr">29</xref>, <xref rid="CIT0053" ref-type="bibr">53</xref>, <xref rid="CIT0054" ref-type="bibr">54</xref>], yet this link has been largely unexplored in sub-Saharan Africa in HIV-positive populations. Our findings reflect those of a qualitative study in Zambia, in which IPV, mental health and HIV are closely related in the experience of women [<xref rid="CIT0055" ref-type="bibr">55</xref>]. Such interrelated &#x0201c;syndemic&#x0201d; issues [<xref rid="CIT0056" ref-type="bibr">56</xref>] should be explored in future sub-Saharan African studies.</p><p>Existing research shows poor mental health has significant impact on ART adherence [<xref rid="CIT0057" ref-type="bibr">57</xref>&#x02013;<xref rid="CIT0060" ref-type="bibr">60</xref>] and among pregnant women depressive symptoms are associated with HIV disease progression and mortality [<xref rid="CIT0061" ref-type="bibr">61</xref>]. It is possible that IPV is one condition exacerbating the relationship between mental health and HIV outcomes. Indeed, one new study suggests that the link between mental health and ART adherence may be partly driven by partner conflict [<xref rid="CIT0062" ref-type="bibr">62</xref>]. Despite high rates of common mental health disorders in antenatal care [<xref rid="CIT0063" ref-type="bibr">63</xref>], little screening or treatment exists in South Africa [<xref rid="CIT0064" ref-type="bibr">64</xref>]. Mental health will be crucial to address among HIV-positive pregnant women because of its strong relationship to IPV and its association with the uptake of PMTCT regimens [<xref rid="CIT0065" ref-type="bibr">65</xref>].</p><p>Finally, IPV may worsen secondary prevention behaviours in pregnancy. Non-adherence to PMTCT regimens greatly increases the risk that pregnant or breastfeeding women will transmit HIV to the infant [<xref rid="CIT0066" ref-type="bibr">66</xref>], potentially in drug-resistant form [<xref rid="CIT0067" ref-type="bibr">67</xref>]. High viral loads related to non-adherence also increase the likelihood of secondary transmission to partners, particularly in the context of unsafe sex. Our research reflects existing knowledge by suggesting that IPV inhibits women's ability to negotiate condoms [<xref rid="CIT0068" ref-type="bibr">68</xref>]. These findings explore such dynamics within the context of pregnancy, thereby suggesting a dual risk of mother-to-child infection and secondary transmission risk to a partner.</p><p>Our findings echo calls for addressing IPV in pregnancy [<xref rid="CIT0069" ref-type="bibr">69</xref>]. Scholars note that antenatal care provides an important &#x0201c;window of opportunity&#x0201d; for women who are regularly accessing the health system [<xref rid="CIT0070" ref-type="bibr">70</xref>]. Although universal screening is not recommended in settings with limited referral options and overstretched providers [<xref rid="CIT0071" ref-type="bibr">71</xref>], some type of IPV assessment, provider training and targeted response may be appropriate for South African antenatal care. Indeed, a comprehensive health response to IPV will likely require either screening or case-finding &#x02013; both methods that may be acceptable in South African clinics [<xref rid="CIT0072" ref-type="bibr">72</xref>, <xref rid="CIT0073" ref-type="bibr">73</xref>].</p><sec id="S0004-S20001"><title>Limitations</title><p>The findings of this formative research should be examined in light of several limitations. Firstly, this study is exploratory in nature, resulting in small sample sizes of each participant group. While analysis suggested that we began to reach saturation through FGDs with pregnant women, the IDIs with pregnant women experiencing IPV were not sufficient to reach thematic saturation [<xref rid="CIT0074" ref-type="bibr">74</xref>]. Secondly, the socio-ecological perspective was brought to the data analysis process after data collection. Ideally, this conceptual approach would have informed the entire data collection process, rather than simply guiding the final interpretation of findings. However, since this was a preliminary, exploratory study, it was designed to explore several intersecting issues and we applied the conceptual framework during data analysis. Finally, some of the findings may be applicable for any woman experiencing IPV, and do not necessarily highlight the specific context of pregnancy. Further research should explore the perinatal time-period in detail to determine whether the link between IPV and HIV is somehow distinct during this life stage.</p></sec></sec><sec sec-type="conclusions" id="S0005"><title>Conclusions</title><p>IPV in pregnancy leads to declines in the physical and mental health of pregnant women. Our findings underscore the negative effects of IPV as a health issue in its own right and as a barrier to PMTCT. The connection between IPV and HIV medication adherence among pregnant women has yet to be explored quantitatively in sub-Saharan Africa. In future studies, it would be ideal to find systematic methods for recruiting more robust numbers of pregnant women who experience IPV and who are living with HIV. In the parent study [<xref rid="CIT0075" ref-type="bibr">75</xref>], we anticipate that by training health providers to ask about IPV confidentially and skilfully, it may be increasingly possible to reach this crucial population.</p><p>Beyond its marked impact on physical and mental health of women, IPV in pregnancy may have important implications for Option B+, as current cost-effectiveness models assume that women are willing and able to achieve 100% adherence [<xref rid="CIT0076" ref-type="bibr">76</xref>]. If Option B+ is to be adopted more broadly, the effect of IPV on adherence and mental health should be carefully considered. Addressing the inter-related issues of violence and HIV will be crucial to ensure that goals of maternal and child health are met in the sub-Saharan African region.</p></sec>

Service, training, mentorship: first report of an innovative education-support program to revitalize primary care social service in Chiapas, Mexico

<sec id="st1"><title>Background</title><p>The Mexican mandatory year of social service following medical school, or pasant&#x000ed;a, is designed to provide a safety net for the underserved. However, social service physicians (pasantes) are typically unpracticed, unsupervised, and unsupported. Significant demotivation, absenteeism, and underperformance typically plague the social service year.</p></sec><sec id="st2"><title>Objective</title><p>Compa&#x000f1;eros en Salud (CES) aimed to create an education-support package to turn the pasant&#x000ed;a into a transformative learning experience.</p></sec><sec id="st3"><title>Design</title><p>CES recruited pasantes to complete their pasant&#x000ed;a in CES-supported Ministry of Health clinics in rural Chiapas. The program aims to: 1) train pasantes to more effectively deliver primary care, 2) expose pasantes to central concepts of global health and social medicine, and 3) foster career development of pasantes. Program components include supportive supervision, on-site mentorship, clinical information resources, monthly interactive seminars, and improved clinic function. We report quantitative and qualitative pasante survey data collected from February 2012 to August 2013 to discuss strengths and weaknesses of this program and its implications for the pasante workforce in Mexico.</p></sec><sec id="st4"><title>Results</title><p>Pasantes reported that their medical knowledge, and clinical and leadership skills all improved during the CES education-support program. Most pasantes felt the program had an overall positive effect on their career goals and plans, although their self-report of preparedness for the Mexican residency entrance exam (ENARM) decreased during the social service year. One hundred percent reported they were satisfied with the CES-supported pasant&#x000ed;a experience and wished to help the poor and underserved in their careers.</p></sec><sec id="st5"><title>Conclusions</title><p>Education-support programs similar to the CES program may encourage graduating medical students to complete their social service in underserved areas, improve the quality of care provided by pasantes, and address many of the known shortcomings of the pasant&#x000ed;a. Additional efforts should focus on developing a strategy to expand this education-support model so that more pasantes throughout Mexico can experience a transformative, career-building, social service year.</p></sec>

<contrib contrib-type="author"><name><surname>Van Wieren</surname><given-names>Andrew</given-names></name><xref ref-type="aff" rid="AF0001">1</xref></contrib><contrib contrib-type="author"><name><surname>Palazuelos</surname><given-names>Lindsay</given-names></name><xref ref-type="aff" rid="AF0002">2</xref></contrib><contrib contrib-type="author"><name><surname>Elliott</surname><given-names>Patrick F.</given-names></name><xref ref-type="aff" rid="AF0001">1</xref><xref ref-type="aff" rid="AF0002">2</xref></contrib><contrib contrib-type="author"><name><surname>Arrieta</surname><given-names>Jafet</given-names></name><xref ref-type="aff" rid="AF0003">3</xref></contrib><contrib contrib-type="author"><name><surname>Flores</surname><given-names>Hugo</given-names></name><xref ref-type="aff" rid="AF0001">1</xref><xref ref-type="aff" rid="AF0002">2</xref></contrib><contrib contrib-type="author"><name><surname>Palazuelos</surname><given-names>Daniel</given-names></name><xref ref-type="aff" rid="AF0001">1</xref><xref ref-type="aff" rid="AF0002">2</xref><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="corresp" rid="cor1">*</xref></contrib>

Global Health Action

<p>Ever since the President of Mexico and the Dean of the Universidad Nacional Aut&#x000f3;noma de M&#x000e9;xico established an agreement in 1937, graduating Mexican medical students have been required to complete a year of primary care social service (called pasant&#x000ed;a) before obtaining their full medical license (<xref rid="CIT0001" ref-type="bibr">1</xref>, <xref rid="CIT0002" ref-type="bibr">2</xref>). Distributing social service physicians (pasantes) throughout Mexico is intended to provide a safety net for the underserved. In reality, however, pasantes are typically unsupervised and unpracticed (meaning that they are both inexperienced and have not undergone enough mentored training to adequately practice without further oversight). In addition, they are often distracted by pending residency entrance exams, and regularly try to secure placements in more comfortable urban environments (<xref rid="CIT0003" ref-type="bibr">3</xref>, <xref rid="CIT0004" ref-type="bibr">4</xref>). When they are assigned rural sites, lack of support, absenteeism, and underperformance often define their experience <xref rid="CIT0004" ref-type="bibr">4</xref>&#x02013;<xref rid="CIT0006" ref-type="bibr">6</xref>).</p><p>Compa&#x000f1;eros en Salud (CES), the Mexican branch of Partners in Health (PIH), is a non-governmental organization (NGO) whose mission is to build a model of excellence in rural primary care in partnership with the Ministry of Health (MOH). The founders of CES have each worked in the Sierra Madre Mountains of the state of Chiapas &#x02013; one of the poorest areas in Mexico &#x02013; for 5&#x02013;10 years, and found that access to primary care was limited by physician shortages (<xref rid="CIT0005" ref-type="bibr">5</xref>). Although Mexico's health system reform in 2003&#x02013;2004 expanded public health insurance coverage to the poorest Mexicans, its full promise has yet to be realized, as well-trained and logistically supported health care professionals are often not available to provide high-quality care in the rural areas where many of the poorest Mexicans live <xref rid="CIT0007" ref-type="bibr">7</xref>&#x02013;<xref rid="CIT0009" ref-type="bibr">9</xref>) . In an effort to expand effective access to care in the Sierra Madre region, CES began to recruit pasantes to work in MOH clinics that previously lacked a physician.</p><p>CES recognized that working with social service physicians also provided an opportunity to address known shortcomings of the pasant&#x000ed;a through creating a transformative educational experience. The CES approach to transformative education was inspired by a call to action from the Lancet Commission for Health Professions Education asking educators to not just inform students by imparting knowledge and skills or form them through teaching professional values, but rather transform students through developing leadership attributes and enlightening them as agents of change (<xref rid="CIT0010" ref-type="bibr">10</xref>). In this transformative spirit, CES staff and US physician collaborators created an innovative education-support program that aims to: 1) train pasantes to effectively deliver primary care, 2) expose pasantes to central concepts of global health and social medicine, and 3) foster continuing medical education and career development of pasantes. This paper describes the components, evaluation, and insights of the first 2 years of the CES education-support program.</p><sec sec-type="methods" id="S0002"><title>Methods</title><sec id="S0002-S20001"><title>Pasante recruitment</title><p>CES staff (LP, JA, HF, and DP) worked with faculty and a student group at the Monterrey Institute of Technology Medical School (ITESM) &#x02013; and subsequently with several other medical schools &#x02013; to identify graduating medical students interested in completing their pasant&#x000ed;a in Chiapas. Initial connections between CES and the ITESM were established through several of the CES staff being alumni or having previous affiliations with ITESM. CES staff have subsequently connected with global health-oriented student groups from medical schools throughout Mexico at national and international conferences, which has facilitated recruitment of pasantes from medical schools in many different areas of Mexico, including Nuevo Le&#x000f3;n, Mexico City, Quer&#x000e9;taro, Durango, Veracruz, Tamaulipas, Chihuahua, and Puebla. During recruitment, CES sought candidates with the following characteristics: academic excellence, interest in global health and social medicine, perceived ability to adapt to an impoverished rural community, and expressed desire to provide primary care to marginalized patients. To ensure that income is not a barrier to participation, CES offers pasantes a matching stipend to that provided by the government for a total combined stipend of about US$380 monthly (<xref rid="CIT0004" ref-type="bibr">4</xref>).</p></sec><sec id="S0002-S20002"><title>Site selection</title><p>CES staff (LP, HF, and DP) identified communities in the Sierra Madre region of Chiapas with public clinics that lacked a full-time physician. Local and state MOH authorities granted CES permission to recruit pasantes to these clinics and provide a package of training and logistical support. CES staff then met with community leaders to introduce the organization, offer a pasante to the local clinic, and request collaboration (including willingness to help pasantes find room and board). Two communities chose to participate in February 2012, and an additional four joined in August 2012. CES aims to expand to a total of 10 clinics by the end of 2015.</p></sec><sec id="S0002-S20003"><title>Program development and components</title><p>The education-support program strives to transform the entire pasant&#x000ed;a into a learning experience. The program harnesses both clinic and classroom time with five fundamental components: 1) monthly multi-day supportive supervision visits by CES staff, 2) biannual multi-week mentorship by medical resident and attending physician volunteers, 3) access to clinical information resources such as locally tailored evidence-based treatment algorithms and UpToDate&#x000a9;, 4) a monthly 2&#x02013;3 day interactive seminar utilizing adult learning best practices, and 5) improved functioning of the clinics where pasantes work. The CES approach to both patient care and pasante mentorship utilizes a style called &#x02018;accompaniment&#x02019;, which emphasizes solidarity and collective problem solving.</p><sec><title>Supportive supervision</title><p>CES hires a select group of graduated pasantes and external applicants as supervisors (in a ratio of approximately one supervisor per three to four pasantes) to provide holistic support for the next generation of pasantes. CES uses a model of supportive supervision informed by the World Health Organization (WHO) concept: supervision is helping to make things work, rather than checking to see what is wrong (<xref rid="CIT0011" ref-type="bibr">11</xref>). Supervisors visit each site for about 3 days per month and review a portfolio of functions with the pasante, including troubleshooting difficult cases, clinic management, facilitating referrals, and negotiating community relations. The pasante and supervisor jointly identify priority areas for improvement and make plans for how to address them. The system encourages performance by recognizing and praising when standards are met, and using a supportive rather than primarily punitive approach when they are not.</p></sec><sec><title>Clinical mentorship</title><p>Throughout the year and with intensified focus when pasantes first arrive in February and August, volunteer residents from the USA and Mexico of different specialties (e.g. Family Medicine, Internal Medicine, Pediatrics, Obstetrics/Gynecology, Psychiatry, and Neurology) pair with pasantes to provide intensive bedside teaching. Clinical mentors do not directly provide care, but rather precept and reinforce core skills, talking through each patient's diagnosis, counseling, and treatment with the pasante.</p></sec><sec><title>Clinical information resources</title><p>In an effort to promote high-quality and standardized care among pasantes, CES developed a series of clinical algorithms to guide pasante decision-making for common primary care problems (e.g. acute cough, dysuria, hypertension, diabetes mellitus, etc.). CES algorithms respect official MOH norms, incorporate likelihood ratios for history and physical exam findings, and use principles from care delivery value chains to encourage pasantes to provide evidence-based and locally tailored care <xref rid="CIT0012" ref-type="bibr">12</xref>&#x02013;<xref rid="CIT0014" ref-type="bibr">14</xref>). CES also provides pasantes with free access to UpToDate&#x000a9;.</p></sec><sec><title>Interactive seminar</title><p>CES designed a monthly course utilizing adult learning best practices such as building on existing knowledge, using a &#x02018;teach-back&#x02019; method, and project and case-based exercises. ITESM accredited the seminar as a 12-month certificate course in Global Health and Social Medicine. The curriculum has undergone two rounds of formal evaluation in an effort to improve its contents and delivery. Because new pasantes arrive to work with CES every 6 months, fundamental components of the curriculum are delivered every 6 months, allowing pasantes in the second half of their year to take on a &#x02018;teach-back&#x02019; role, whereas other components of the curriculum are delivered once every 12 months.</p><p>The course content focuses on three main areas: 1) clinical medicine taught through problem-based learning cases, case presentations done in &#x02018;morning report&#x02019; style, and use of ultrasound and other diagnostic tests in the primary care setting; 2) global health and social medicine delivered via case-based health systems analysis, socioeconomic determinants of health, policy, and cultural competency and humility; and 3) quality care delivery (e.g. clinical leadership skills, team-based care, quality improvement projects, morbidity and mortality review, feedback sessions, and humanistic curriculum). CES staff, volunteer residents, guest speakers, graduated pasante supervisors and, in many cases, pasantes in the latter half of their pasant&#x000ed;a all teach different sessions of the course, with all sessions supervised by CES leadership and often using a team-teaching strategy. The 2&#x02013;3 day monthly seminar also provides pasantes an opportunity to build community and support each other throughout the challenging social service year. As part of the course, pasantes also receive complementary EXARMED&#x000a9; study materials for the Mexican residency entrance exam (ENARM).</p></sec><sec><title>Improved function</title><p>CES pasantes leave their assigned communities for approximately 7&#x02013;8 pre-planned consecutive days each month, 3&#x02013;4 days of which are spent between the 2&#x02013;3 day educational seminar at the CES office, 1 day of turning in paperwork at the jurisdiction offices, and 1&#x02013;2 days of travel time. The remaining 3&#x02013;4 days are assigned as vacation. This model of concentrating days away from the community purposely differs from the traditional model of pasantes having 1&#x02013;2 days off a week, because the long trips required to exit and then re-enter isolated rural communities tend to diminish actual time off and promote absenteeism (<xref rid="CIT0004" ref-type="bibr">4</xref>, <xref rid="CIT0005" ref-type="bibr">5</xref>). CES-supported clinics also receive supplemental supplies to ensure there are no stock-outs of essential medicines, diagnostic tests, and basic infection control supplies.</p></sec></sec><sec id="S0002-S20004"><title>Motivating and retaining pasantes</title><p>Adequately supporting pasantes during their social service year requires many complex and interdependent investments. In a systematic review of factors that help motivate and retain health workers in underserved areas, Willis-Shattuck found that seven critical strategies combat &#x02018;brain drain&#x02019; and promote &#x02018;brain gain&#x02019; (<xref rid="CIT0015" ref-type="bibr">15</xref>). <xref ref-type="table" rid="T0001">Table 1</xref> provides examples of how CES has learned to address these seven areas programmatically. Furthermore, because new pasantes coming from different educational backgrounds have varied levels of preparedness for both the clinical and psychosocial demands of the social service year &#x02013; similar to new interns starting residency programs &#x02013; CES strives to quickly identify pasantes who require additional support to catch up to their peers and then provides supplemental supportive supervision and clinical mentorship during the first month of the pasant&#x000ed;a.</p><table-wrap id="T0001" position="float"><label>Table 1</label><caption><p>Seven key motivating factors (<xref rid="CIT0015" ref-type="bibr">15</xref>) and corresponding CES strategies to attract and retain pasantes</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1">Motivating factor</th><th align="center" rowspan="1" colspan="1">CES strategies</th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">Financial rewards</td><td align="left" rowspan="1" colspan="1">-Match pasante stipend, increasing the total from approximately ~US$190 to ~US$380 monthly</td></tr><tr><td align="left" rowspan="1" colspan="1">Career development</td><td align="left" rowspan="1" colspan="1">-Provide opportunities for pasantes to work in management roles within CES after completing the pasant&#x000ed;a<break/>-Provide opportunities for pasantes to visit other Partners in Health sites</td></tr><tr><td align="left" rowspan="1" colspan="1">Continuing education</td><td align="left" rowspan="1" colspan="1">-Monthly CES course<break/>-Supervisor and resident accompaniment in clinic<break/>-Clinical information resources such as clinical algorithms and access to UpToDate&#x000a9;</td></tr><tr><td align="left" rowspan="1" colspan="1">Work environment</td><td align="left" rowspan="1" colspan="1">-Providing electronic medical record and computers for clinic notes and for automatic generation of government paperwork that previously took hours to complete by hand</td></tr><tr><td align="left" rowspan="1" colspan="1">Resource availability</td><td align="left" rowspan="1" colspan="1">-Working with Mexican government to prevent &#x02018;stock-outs&#x02019; of essential medications<break/>-Supplementing medication stocks<break/>-Helping pasantes refer complex patients</td></tr><tr><td align="left" rowspan="1" colspan="1">Workplace management</td><td align="left" rowspan="1" colspan="1">-Teaching pasantes how to work as part of and lead a clinical team<break/>-Regular visits by supervisors and leadership to help build teamwork among clinic staff<break/>-Advising pasantes on how to manage patient flow and limit wait times</td></tr><tr><td align="left" rowspan="1" colspan="1">Recognition/appreciation</td><td align="left" rowspan="1" colspan="1">-Providing a certificate through the ITESM for the Global Health and Social Medicine course<break/>-Supporting pasantes to integrate with the communities they serve</td></tr></tbody></table></table-wrap></sec><sec id="S0002-S20005"><title>Survey instrument</title><p>The authors collected data about the education-support program and its influence on pasantes using pre-, mid-, and post-intervention surveys that ranged in length from 18 to 23 questions. The surveys incorporated both Likert scale and open-ended questions, and explored the following areas: pasantes&#x02019; self-reported medical knowledge (overall medical knowledge, and specialized knowledge in Internal Medicine, Pediatrics, Obstetrics/Gynecology, and Emergency Medicine); clinical leadership skills and understanding of the Mexican health care system; preparedness for the ENARM; residency and career plans (planning on applying for residency, type of medicine planning to pursue, desire to work with the poor or underserved); best and worst experiences as a pasante; best aspects and recommended changes for the education-support program; and degree of satisfaction with the CES program. Content of the three surveys was similar to facilitate measuring changes throughout the year. The survey instruments were written by AVW, LP, and DP in English and then translated into Spanish by a bilingual native Spanish-speaking CES volunteer.</p></sec><sec id="S0002-S20006"><title>Study design, data collection, and analysis</title><p>The present study was conceived as a case study of pasantes completing their social service requirement as part of the CES program. The data reported were collected from February 2012 through August 2013. The data reflect completed pre-, mid-, and post-intervention questionnaires for six pasantes, representing the first two classes of pasantes to complete the CES education-support program. Data are not reported for two pasantes who did not complete the study, one who left the CES-affiliated pasant&#x000ed;a due to a desire to be closer to family, and one who had incomplete data. To put the current size of the CES program in context, 263 of 932 (28.2%) public primary care clinics in Chiapas are staffed exclusively by pasantes (<xref rid="CIT0004" ref-type="bibr">4</xref>). Surveys were distributed to pasantes during a designated time at the seminars and pasantes were given as much time as needed to complete them. Pasantes wrote a code that only they would understand on their three surveys in order to concurrently track data and preserve anonymity. AVW, LP, and DP performed data analysis, which included conducting descriptive statistics on quantitative data, and identifying predominant themes and supporting quotations from qualitative data. This study was reviewed by the Partners Healthcare Institutional Review Board (Brigham and Women&#x02032;s Hospital) in Boston, MA, and was given IRB exemption.</p></sec><sec id="S0002-S20007"><title>Evaluation of the program</title><p>CES staff evaluates the CES education-support program on both an informal continuous basis and also at designated formal intervals at the end of each pasante cycle (every 6 months in January and July). Pasantes evaluate the program through the aforementioned pre-, mid-, and post-intervention questionnaires and through informal feedback obtained by staff and volunteers throughout the social service year. CES staff members verbally evaluate the program during regularly scheduled monthly meetings with CES leadership. CES volunteers provide either verbal or written evaluations of the program at the conclusion of their volunteer period. CES then uses feedback from pasantes, staff members, and volunteers to implement new changes to both the educational and support aspects of the curriculum at the beginning of each new pasante cycle (every 6 months in February and August). CES implements changes to the education-support program using a Plan-Do-Study-Act model, in which feedback is reviewed by CES leadership to form plans for change, which are then implemented, subsequently studied in the next semester's review process, and then the change plans are either adopted, adapted, or abandoned.</p></sec></sec><sec sec-type="results" id="S0003"><title>Results</title><p>The six pasantes who completed the study ranged in age from 22 to 26 years, included three men and three women, had all completed required coursework at an accredited medical school in Mexico making them eligible to complete their social service requirement, and all had completed the pasant&#x000ed;a as part of the CES education-support program in Chiapas.</p><sec id="S0003-S20001"><title>Knowledge and skills</title><p>Overall, pasantes reported that their medical knowledge and clinical skills improved while completing the education-support program during their social service year in Chiapas. <xref ref-type="fig" rid="F0001">Figure 1</xref> demonstrates how pasante self-report of general clinical knowledge and preparedness to practice Internal Medicine, Pediatrics, Obstetrics/Gynecology, and Emergency Medicine changed throughout the course of the education-support program. When asked to reflect on changes in knowledge and skills during the pasant&#x000ed;a, one pasante remarked, &#x02018;I have gained greater medical knowledge because of the accompaniment I received from residents, the feedback I received from patients, and from being able to longitudinally observe the patients I treated in my community&#x02019;. Another observed, &#x02018;My abilities to create differential diagnoses and treat patients are constantly improving because of the feedback and accompaniment I have received from the CES staff, residents, and volunteers&#x02019;.</p><fig id="F0001" position="float"><label>Fig. 1</label><caption><p>Percent of pasantes reporting good or very good general clinical knowledge and preparedness to practice several medical specialties before, halfway through, and at the end of the CES education-support program.</p></caption><graphic xlink:href="GHA-7-25139-g001"/></fig><p>Pasantes reported that clinical leadership skills and understanding of the Mexican health care system also improved throughout the course of the education-support program. The percentage of pasantes who reported feeling well or very well prepared to lead a primary care clinic increased from 50% at the beginning of the program to 83% by the end of the program. Whereas 67% of pasantes reported understanding the organization or hierarchy of the Mexican health care system well or very well at the beginning of the program, 100% of pasantes felt that way by the end of the program.</p><p>Throughout the course of the education-support program, pasante report of perceived preparedness for the Mexican residency entrance exam (ENARM) decreased and, by the end of the year, most pasantes had decided not to take the ENARM in the year following their pasant&#x000ed;a. Although 100% of pasantes believed the pasant&#x000ed;a would have a positive or very positive effect on their performance on ENARM at the beginning of the program, by the end of the year 83% of pasantes had decided against taking the ENARM in the year following their pasant&#x000ed;a.</p><p>Notably, pasantes reported learning more than just rote clinical knowledge. Pasantes also recalled developing more nuanced clinical skills &#x02013; such as the ability to use a referral network for more complex cases &#x02013; and beginning to appreciate and address the social determinants of health. One pasante reflected, &#x02018;I found that I was actually able to accurately diagnose and treat the majority of my patients; for the more complicated cases in which I did not know what to do, I fortunately had the support of the entire CES team to first figure out the diagnosis and then implement the best treatment plan&#x02019;. Another explained, &#x02018;I am beginning to understand how to break the cycle of poverty and disease by treating diseases that were previously incurable in the Sierra, diseases that affected whole generations and society as a whole&#x02019;.</p></sec><sec id="S0003-S20002"><title>Career goals and plans</title><p>Pasante career goals and plans changed in several ways throughout the course of the education-support program. When asked if they were considering a career in primary care at the beginning of the year, 33% responded &#x02018;probably no&#x02019;, 33% &#x02018;undecided&#x02019;, and 33% &#x02018;probably yes&#x02019;. By the end of program, 67% of pasantes reported they were either definitely or probably not considering a career in primary care, but 33% reported they definitely were considering a career in primary care and no pasantes remained undecided. The percentage of pasantes who reported a desire to work primarily with poor and underserved populations during their career increased from 83 to 100% between the beginning and end of the program. By the end of the year, 83% of pasantes felt the social service year had a positive or very positive effect on their career goals and plans. One pasante detailed this transition as follows: &#x02018;The trajectory of my life has taken a significant turn after working with CES. I would like to work with different communities, especially the poor communities where CES works&#x02019;. Another pasante reflected, &#x02018;Before I arrived to work with CES, my life plan was set. Now I'm not sure what I want to do. The only thing I'm sure of is that, whatever happens, I want to continue collaborating with CES in the short- and long-term&#x02019;.</p></sec><sec id="S0003-S20003"><title>Satisfaction with pasant&#x000ed;a</title><p>When asked at the end of the year, 100% of pasantes reported they were satisfied with their experience as a pasante and that they were glad they had done their social service year in CES-supported government clinics in Chiapas. When asked why they were satisfied with the pasant&#x000ed;a, one pasante wrote, &#x02018;I grew personally and professionally&#x02019;, whereas another pasante explained &#x02018;[the year] provided me a totally new perspective regarding professional opportunities that I had never imagined&#x02019;. At the end of the year, 67% of pasantes reported that they had achieved their goals for the pasant&#x000ed;a. Among the pasantes who reported achieving their goals, cited reasons why included: &#x02018;I learned how to practice medicine and feel confident as a doctor&#x02019;, and &#x02018;I completed all of my goals and even others I formed during the year&#x02019;. One of the pasantes who denied having achieved his or her goals for the year explained: &#x02018;I gained many clinical skills and helped many patients. However, I did not study for ENARM as much as I had hoped&#x02019;.</p></sec><sec id="S0003-S20004"><title>Satisfaction with education-support program</title><p> <xref ref-type="table" rid="T0002">Table 2</xref> shows highlights of what pasantes felt were the best and worst aspects of both the overall pasant&#x000ed;a and the CES education-support program. Regarding the overall pasant&#x000ed;a, pasantes appeared to appreciate the following: gaining clinical knowledge and confidence, developing relationships with their patients over time, seeing their patients improve, having resident and supervisor accompaniment, attending the monthly global health seminar, working on quality improvement projects, and exploring a new area of Mexico. In critiquing the overall pasant&#x000ed;a, pasantes cited the following: running out of available treatment options, patients not following up as requested, social isolation in the communities, frustration and fatigue throughout the course of the year, limited collaboration with nurses in the clinics, and growing pains of CES as a new organization. When asked to describe the best aspects of the CES education-support program, pasantes praised: the program's global health and social medicine focus, its case-based approach, the program being new and unique in Mexico, CES always striving to improve the program, resident and supervisor accompaniment, the ITESM certificate awarded at the conclusion of the program, and the program being tailored to the local reality of Chiapas. Asked to identify the worst aspects of the CES education-support program, pasantes criticized: the seminar's classes being too long and dense, the curriculum being poorly defined, the teaching approach being too informal, too many different instructors teaching the courses, too much focus on quality improvement, and limited distribution of written resources.</p><table-wrap id="T0002" position="float"><label>Table 2</label><caption><p>Pasante report of best and worst aspects of the overall pasant&#x000ed;a, and best and worst aspects of the CES education-support program</p></caption><table frame="hsides" rules="groups"><tbody><tr><td align="left" rowspan="1" colspan="1">Best aspects of overall pasant&#x000ed;a:</td><td align="left" rowspan="1" colspan="1">Worst aspects of overall pasant&#x000ed;a:</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Gaining clinical knowledge and confidence<break/>&#x02003;Seeing patients improve<break/>&#x02003;Long-term relationships with patients<break/>&#x02003;Resident and supervisor accompaniment<break/>&#x02003;Traveling to new area of Mexico<break/>&#x02003;Monthly global health seminar<break/>&#x02003;Working on quality improvement projects</td><td align="left" rowspan="1" colspan="1">&#x02003;Running out of available treatment options<break/>&#x02003;When patients did not follow up as requested<break/>&#x02003;Being alone in the community<break/>&#x02003;Frustration and fatigue<break/>&#x02003;Growing pains of CES as a new organization<break/>&#x02003;Limited collaboration with nurses</td></tr><tr><td align="left" rowspan="1" colspan="1">Best aspects of education-support program:<break/>&#x02003;Global health and social medicine focus<break/>&#x02003;Case-based approach<break/>&#x02003;Program is new and unique in Mexico<break/>&#x02003;CES is always working to improve program<break/>&#x02003;Resident and supervisor accompaniment<break/>&#x02003;Certificate awarded at end of course<break/>&#x02003;Locally oriented to Chiapas</td><td align="left" rowspan="1" colspan="1">Worst aspects of education-support program:<break/>&#x02003;Classes are too long<break/>&#x02003;Classes are too dense<break/>&#x02003;Curriculum is not defined well enough<break/>&#x02003;Too informal<break/>&#x02003;Too many different people teaching classes<break/>&#x02003;Too much focus on quality improvement<break/>&#x02003;Limited distribution of written resources</td></tr><tr><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1"/></tr></tbody></table></table-wrap></sec></sec><sec sec-type="discussion" id="S0004"><title>Discussion</title><p>Building on changes introduced by the Mexican health system reform of 2003&#x02013;2004, the Mexican health care system has recently been celebrated for providing universal health coverage to all Mexicans (<xref rid="CIT0007" ref-type="bibr">7</xref>, <xref rid="CIT0008" ref-type="bibr">8</xref>). Despite this high level of enrollment, gaps in effective coverage persist (<xref rid="CIT0016" ref-type="bibr">16</xref>). Truly achieving equitable access to high-quality care within Mexico will require an ongoing commitment to ensuring primary care providers arrive at the most underserved regions of Mexico and are provided logistical and training support so that they can perform at a high level (<xref rid="CIT0017" ref-type="bibr">17</xref>).</p><p>In Mexico, despite the fact that pasantes are still technically considered students until they complete their social service year, as many as one-third of public primary care clinics are staffed exclusively by pasantes and nearly three-quarters of pasantes are assigned to rural health centers (<xref rid="CIT0004" ref-type="bibr">4</xref>). This represents a large work force addressing rural health needs, but there is a growing concern that, similar to interns beginning residency, Mexican pasantes would provide higher quality and safer care if they were supervised, supported, and continued to receive education during the social service year (<xref rid="CIT0004" ref-type="bibr">4</xref>, <xref rid="CIT0005" ref-type="bibr">5</xref>). Innovative approaches to supporting and training pasantes during their social service year are sorely needed to help turn Mexico's universal enrollment into true universal effective coverage.</p><p>Given the limited number of studies evaluating strategies to address human resources concerns and strengthen primary health care systems in low- and middle-income countries (<xref rid="CIT0018" ref-type="bibr">18</xref>, <xref rid="CIT0019" ref-type="bibr">19</xref>), this program and its evaluation offer important insights to those interested in strengthening the primary health care system in Mexico and wherever social service providers are dispatched. Our results suggest that the CES education-support program has several important strengths and insights.</p><p>The results of this study show that the CES education-support program is a transformative experience for participants. The Lancet Commission for Health Professions Education has called for efforts at medical education reform to be guided by two main principles: transformative (rather than only informative or formative) learning, and interdependent and interdisciplinary education (<xref rid="CIT0010" ref-type="bibr">10</xref>). Involvement in the CES program fundamentally changed the way the pasantes considered their future; all experienced an increased commitment to work with the poor and underserved in the future. In addition, they reported greater clinical leadership skills and a more nuanced understanding of the Mexican health care system. The Lancet Commission for Health Professions Education has highlighted several programs that model transformative education, such as the Public Health Foundation of India which uses a similar public-private partnership as the CES model to transform public health professionals during a 12-month program that couples field work with classroom learning (<xref rid="CIT0010" ref-type="bibr">10</xref>). CES aims to draw from both internal program evaluation and lessons learned from others in an effort to constantly improve the CES program and make it as transformative as possible for pasantes. Although the MOH did provide pasantes with some basic supervision and training before development of the CES program, the public-private partnership between the MOH and CES has resulted in the development of a program that now teaches pasantes leadership skills and how to potentially be agents of change within the Mexican health care system.</p><p>The findings also highlight the importance of classroom training that goes beyond technical proficiencies to integrate systems-level concepts of global health, social medicine, and quality improvement. The Global Health and Social Medicine course appears to instill a sense of mission within pasantes, thus invigorating the care they provide to underserved patients. Interest in global health is rising to unprecedented levels among medical trainees, and seems to represent a revitalization of the sense of mission and altruism that is core to the medical profession (<xref rid="CIT0020" ref-type="bibr">20</xref>, <xref rid="CIT0021" ref-type="bibr">21</xref>). Inviting residents and attending physicians from Mexico, the USA, and elsewhere to offer clinical &#x02018;accompaniment&#x02019; of pasantes represents a powerful way in which interest in global health can be harnessed to improve training for not only Mexican providers but also the visiting physicians themselves (<xref rid="CIT0010" ref-type="bibr">10</xref>, <xref rid="CIT0019" ref-type="bibr">19</xref>). Efforts to create effective and sustainable approaches to pasante training and mentoring will benefit from new partnerships that have not been previously considered, including partnerships between the public sector, private NGOs, and academic institutions.</p><p>Although critically evaluating this type of education-support program will inform future efforts to support and educate pasantes, we must also consider how public policy changes could influence pasante behaviors and promote effective universal access (<xref rid="CIT0015" ref-type="bibr">15</xref>, <xref rid="CIT0019" ref-type="bibr">19</xref>). Indeed, some limitations of the pasant&#x000ed;a model may be better addressed through public policy changes than educational programs alone. For example, entrance to residency programs in Mexico is determined largely by the ENARM test, typically taken after the social service year. Despite providing complementary study preparation materials for ENARM, pasantes&#x02019; self-reported preparedness for the ENARM exam decreased during the social service year and most pasantes had decided to either postpone or not take ENARM by the end of the pasant&#x000ed;a. We suspect this phenomenon is the result of clinical responsibilities precluding pasantes from having sufficient study time. Perhaps a more realistic and transformative approach is that used within the Chilean social service system. In Chile, rather than having to worry about how social service time will negatively affect performance on residency entrance exams, pasantes who complete their social service in rural undeserved areas are awarded additional points on the medical residency entrance exam (<xref rid="CIT0004" ref-type="bibr">4</xref>).</p><p>Another example of how public policy changes could improve the social service year is by securing support within the most recent Mexican health reform to expand education-support programs for pasantes within the public sector. Although the CES program has created a unique educational environment within six government-run clinics, the state of Chiapas alone has 263 primary care clinics that are staffed exclusively by pasantes (<xref rid="CIT0004" ref-type="bibr">4</xref>). The CES program relies upon private donations and is run by a NGO, and is thus likely not sustainable at a large scale. We believe that the question is not whether there is enough money to better train and support the providers who care for the most underserved Mexicans, but rather whether sufficient political will exists to align funds for this purpose. As the Lancet Commission for Health Professions Education points out, despite the labor-intensive nature of medicine, only 2% of global health expenditures are spent on training &#x02013; a percentage that is even smaller in low-income countries (<xref rid="CIT0010" ref-type="bibr">10</xref>). Furthermore, in his critical appraisal of the pasant&#x000ed;a in Mexico, Nigenda proposes that, with relatively modest financial investment, the Mexican government could feasibly hire licensed and fully trained physicians to work alongside pasantes, guiding and training them during their social service year (<xref rid="CIT0004" ref-type="bibr">4</xref>). In fact, some Mexican districts already provide pasantes with limited intermittent training and supervision. However, our data support the deployment of more comprehensive training and support packages that convert the social service year into a transformative experience.</p><sec id="S0004-S20001"><title>Study limitations</title><p>Our study results have several important limitations. First, because we did not incorporate a comparison group of pasantes not participating in the education-support program into the study, the results could simply show the effect of completing a social service year rather than that of the education-support program. Furthermore, because many of the survey questions were retrospective about experiences during the preceding months, the results are subject to recall bias. Finally, because our sample size of pasantes is small (<italic>N</italic>=6), our quantitative results should be interpreted as suggesting possible trends rather than establishing statistically significant patterns.</p></sec><sec id="S0004-S20002"><title>Opportunities for future work</title><p>Ultimately, if we are to engage in true health systems strengthening, we will need to develop a strategy to fully transition the CES education-support program to the public sector (<xref rid="CIT0022" ref-type="bibr">22</xref>). Large-scale expansion of this type of education-support program with public funding, however, should only be undertaken after more rigorous studies have demonstrated a clear benefit of the education-support program by directly comparing it to the social service year alone. Notably, recruitment of pasantes for the program and study was likely facilitated by affiliations with ITESM, PIH, and Harvard Medical School &#x02013; all well-known and respected private institutions &#x02013; and, thus, efforts to expand this type of program within the public sector would need to consider how those organizational affiliations could either be maintained at larger scale or explore other types of recruitment incentives. However, there is an expanding body of literature indicating that a large percentage of medical students throughout the world are interested in global health suggesting that, if the pasant&#x000ed;a in Mexico (and elsewhere in Latin America) could be appropriately re-branded as a global health effort, enthusiasm for the social service requirement might surge at the national and international level (<xref rid="CIT0020" ref-type="bibr">20</xref>). Finally, subsequent studies of pasante education and support should attempt to move from pasante perceptions of outcomes, such as clinical knowledge and skills, to more robust measures such as test scores, and peer or patient evaluation of pasantes.</p></sec></sec><sec sec-type="conclusions" id="S0005"><title>Conclusions</title><p>Guaranteeing universal access to high-quality health care for the poor in Mexico requires not only financial and organizational reform, but also achieving an equitable distribution of well-trained and operationally supported health care providers. In Mexico and much of Latin America, the poor depend largely upon health care services provided by pasantes who are often inadequately prepared to practice independently. Efforts that focus on training and mentoring during the social service year represent a unique opportunity to improve the care that the poor and vulnerable receive, and build a cadre of providers motivated to serve them. Based on our data, education-support programs that combine on-site accompaniment of pasantes, access to clinical information resources, and regular seminars appear to be an effective way to recruit pasantes to underserved areas, increase their clinical knowledge and leadership skills, and address many of the known shortcomings of the social-service year. Additional efforts could focus on fully transitioning this type of education-support program to the public sector and developing an expansion model to reach pasantes throughout Mexico and Latin America.</p></sec><sec id="S0006"><title>Disclosures</title><p>AVW received funding for his travel, living expenses while in Chiapas, and limited supplies through a Martin Solomon Primary Care Scholarship through Brigham and Women&#x02032;s Hospital Department of Internal Medicine. CES is funded primarily by donations received through Partners in Health.</p></sec>

The effect of immunonutrition (glutamine, alanine) on fracture healing

<sec id="st1"><title>Background</title><p>There have been various studies related to fracture healing. Glutamine is an amino acid with an important role in many cell and organ functions. This study aimed to make a clinical, radiological, and histopathological evaluation of the effects of glutamine on fracture healing.</p></sec><sec id="st2"><title>Methods</title><p>Twenty rabbits were randomly allocated into two groups of control and immunonutrition. A fracture of the fibula was made to the right hind leg. All rabbits received standard food and water. From post-operative first day for 30 days, the study group received an additional 2 ml/kg/day 20% <sc>L</sc>-alanine <sc>L</sc>-glutamine solution via a gastric catheter, and the control group received 2 ml/kg/day isotonic via gastric catheter. At the end of 30 days, the rabbits were sacrificed and the fractures were examined clinically, radiologically, and histopathologically in respect to the degree of union.</p></sec><sec id="st3"><title>Results</title><p>Radiological evaluation of the control group determined a mean score of 2.5 according to the orthopaedists and 2.65 according to the radiologists. In the clinical evaluation, the mean score was 1.875 for the control group and 2.0 for the study group. Histopathological evaluation determined a mean score of 8.5 for the control group and 9.0 for the study group.</p></sec><sec id="st4"><title>Conclusion</title><p>One month after orally administered glutamine&#x02013;alanine, positive effects were observed on fracture healing radiologically, clinically, and histopathologically, although no statistically significant difference was determined.</p></sec>

<contrib contrib-type="author"><name><surname>K&#x000fc;&#x000e7;&#x000fc;kalp</surname><given-names>Abdullah</given-names></name><xref ref-type="aff" rid="AF0001">1</xref><xref ref-type="corresp" rid="cor1">*</xref></contrib><contrib contrib-type="author"><name><surname>Durak</surname><given-names>Kemal</given-names></name><xref ref-type="aff" rid="AF0002">2</xref></contrib><contrib contrib-type="author"><name><surname>Bayyurt</surname><given-names>Sarp</given-names></name><xref ref-type="aff" rid="AF0003">3</xref></contrib><contrib contrib-type="author"><name><surname>S&#x000f6;nmez</surname><given-names>G&#x000fc;rsel</given-names></name><xref ref-type="aff" rid="AF0004">4</xref></contrib><contrib contrib-type="author"><name><surname>Bilgen</surname><given-names>Muhammed S.</given-names></name><xref ref-type="aff" rid="AF0002">2</xref></contrib>

Food &#x00026; Nutrition Research

<p>The classical definition of a fracture is the breaking of the entirety of a bone from internal or external forces. The bone is healed by reconstruction of osseous tissue without any scarring (<xref rid="CIT0001" ref-type="bibr">1</xref>). Fracture healing starts immediately after the injury and continues until the broken ends unite with the new bone tissue (<xref rid="CIT0002" ref-type="bibr">2</xref>).</p><p>Studies have been made on many factors to accelerate fracture healing. Glutamine is an amino acid with an important role in several chemical reactions of many cell and organ functions (<xref rid="CIT0003" ref-type="bibr">3</xref>). Among these are the energy source of cells such as fibroblasts, epithelial cells, enterocytes, lymphocytes, and macrophages. In addition, it is the most important precursor substance regulating protein synthesis in the nucleic acid biosynthesis of cells. By protecting tissues from free radical damage, and as a major antioxidant, glutamine shows positive effects by promoting glutathione synthesis (<xref rid="CIT0004" ref-type="bibr">4</xref>&#x02013;<xref rid="CIT0006" ref-type="bibr">6</xref>). Several studies in literature have shown positive healing effects of glutamine on wound healing and on patients requiring critical care (<xref rid="CIT0007" ref-type="bibr">7</xref>, <xref rid="CIT0008" ref-type="bibr">8</xref>). It is thought that this study can make a positive contribution to fracture healing by considering the accelerating effect of glutamine on wound healing.</p><sec sec-type="materials|methods" id="S0002"><title>Materials and methods</title><p>This study was approved by Uludag University Animal Experiments Local Ethics Committee and was carried out on 20 New Zealand white rabbits, aged 3 months, weighing between 2.5 and 3.0 kg, obtained from the Experimental Animal Breeding and Research Centre with the permission of the Deanery of Uludag University Medical Faculty.</p><sec id="S0002-S20001"><title> Study plan &#x02013; surgical technique</title><p>The rabbits were randomly allocated into two groups of 10 as the immunonutrition group and the control group. Anaesthesia of 5 mg/kg xylazine and 35 mg/kg ketamine was administered intramuscularly to both groups. The rabbits were then laid down with the right hind leg uppermost. After disinfection of the fibular area with 10% free alcohol containing 10 g povidone iodine, a 1&#x02013;2 cm skin incision was made approximately 1 cm to the posterior of the fibula. By subcutaneous dissection, continuing to slide the skin over the fibula, the fibula shaft was reached. An oblique fracture was created using tissue scissors on the fibula corpus (<xref ref-type="fig" rid="F0001">Figs. 1</xref> and <xref ref-type="fig" rid="F0002">2</xref>).</p><fig id="F0001" position="float"><label>Fig. 1</label><caption><p>Surgical fracture of the exposed rabbit fibula.</p></caption><graphic xlink:href="FNR-58-24998-g001"/></fig><fig id="F0002" position="float"><label>Fig. 2</label><caption><p>Short oblique surgical fracture.</p></caption><graphic xlink:href="FNR-58-24998-g002"/></fig><p>After washing the surgical site with saline, the skin was closed. Povidone iodine was applied with a dressing, then transparent film dressing spray, leaving the wound uncovered. From the post-operative first day for 30 days, the immunonutrition group were given standard rabbit food and water with the addition of 2 ml/kg/day 20% L-alanine and L-glutamine solution via a gastric catheter. The control group received 2 ml/kg/day isotonic via gastric catheter in addition to standard rabbit food and water. All the rabbits were sacrificed at the end of 30 days. By disarticulating the lower extremity where the fracture had been made, from the hip joint, the surrounding soft tissue was removed without any damage to the callus tissue. The fractures were examined clinically, radiologically, and histopathologically in respect to the degree of union.</p><p>Anterior&#x02013;posterior and lateral radiographs were taken of the lower extremities removed from the rabbits. These radiographs were evaluated by five specialist doctors from the orthopaedics and traumatology department and five separate specialist doctors from the radiology department using the Lane and Sandhu (<xref rid="CIT0009" ref-type="bibr">9</xref>) radiological scoring system to give a numerical value. In this system, the radiological findings are scored 0&#x02013;4, where 0=no healing, 1=callus formation, 2=the start of bone union, 3=the fracture line is starting to disappear, and 4=full union. The fibula union was subjectively evaluated by examination in two planes as described by Dimar et al. (<xref rid="CIT0010" ref-type="bibr">10</xref>). This scoring system grades the amount and quality of bone callus tissue through physical movement. A score of 0 indicates movement in both planes and therefore no union, a score of 1 indicates movement in one plane with moderate union and with a score of 4, there is no movement and full union.</p><p>After the clinical and radiological evaluations, tissue from the prepared samples from the fracture union site were fixed in 10% concentration formalin and stored for 3 days. The electrolytic method was used for decalcification and accordingly the tissues were kept for 2 weeks in a decalcification solution, and changed every 2 days (<xref rid="CIT0011" ref-type="bibr">11</xref>). The tissues were washed for 4 h in running water, then for 1 h in each of the following strengths of alcohol; 70, 80, 90, 96, and 100%, kept overnight in xylol, then for 2 h in liquid paraffin before being fixed in paraffin blocks. The lamina prepared from 5-micron thickness sections cut by rotary microtome were stained with haematoxylin and eosin and examined by light microscope. Callus tissue was scored histopathologically using the system recommended by Huo et al. (<xref rid="CIT0012" ref-type="bibr">12</xref>).</p><p>Statistical analysis was made by SPSS 13.0 packet program. Descriptive statistics were given as median, minimum, and maximum values. The Mann&#x02013;Whitney <italic>U-</italic>test was used for comparison of the two independent groups and the Wilcoxon test for comparison of dependent groups. Spearman correlation analysis was used in the examination of the relationships between variables. A significance level of <italic>p</italic>&#x0003c;0.05 was accepted.</p></sec></sec><sec sec-type="results" id="S0003"><title>Results</title><p>In the first week of the study, two rabbits in the immunonutrition group aspirated the dipeptiven solution given by gastric catheter. One of the rabbits died within minutes of aspiration and the other on the following day. In the control group, in the first week, one rabbit trapped its right lower extremity in the cage while it was being cleaned and sustained a femoral fracture. A pelvipedal plaster was applied but the rabbit died 3 days later. In the second week, one rabbit in the control group was found dead in its cage of unknown causes. The study continued with eight rabbits in the immunonutrition group and eight rabbits in the control group. Throughout the study no surgical site infection developed in any of the rabbits. Apart from the rabbit that sustained a femoral fracture, no functional problems were observed in any rabbit other than the created fibula fracture.</p><sec id="S0003-S20001"><title>Radiological findings</title><p>The radiological evaluation results determined the control group points from the orthopaedic and traumatology specialists to be mean 2.5 (minimum 0.4&#x02013;maximum 3.6), and from the radiology specialists, mean 2.65 (minimum 0.6&#x02013;maximum 3.8) (<xref ref-type="fig" rid="F0003">Figs. 3</xref> and <xref ref-type="fig" rid="F0004">4</xref>).</p><fig id="F0003" position="float"><label>Fig. 3</label><caption><p>Control group anterior&#x02013;posterior radiograph.</p></caption><graphic xlink:href="FNR-58-24998-g003"/></fig><fig id="F0004" position="float"><label>Fig. 4</label><caption><p>Immunonutrition group anterior&#x02013;posterior radiograph.</p></caption><graphic xlink:href="FNR-58-24998-g004"/></fig><p>The immunonutrition group points from the orthopaedic and traumatology specialists were determined to be mean 3.42 (minimum 3.0&#x02013;maximum 3.8), and from the radiology specialists, mean 3.35 (minimum 2.4&#x02013;maximum 4.0) (<italic>p</italic>=0.13 orthopaedists, <italic>p</italic>=0.279 radiologists, <xref ref-type="fig" rid="F0005">Fig. 5</xref>).</p><fig id="F0005" position="float"><label>Fig. 5</label><caption><p>Mean points distribution given by Orthopaedics and Radiology physicians.</p></caption><graphic xlink:href="FNR-58-24998-g005"/></fig></sec><sec id="S0003-S20002"><title>Clinical examination results</title><p>An orthopaedic doctor who was not involved in the experimental study made a subjective clinical examination of the relevant disarticulated extremities after removal from the soft tissue. In the clinical evaluation, the control group mean points were 1.875 (minimum 1.0&#x02013;maximum 2.0), and the immunonutrition group mean points were 2.0 (minimum 2.0&#x02013;maximum 2.0) (<italic>p</italic>=0.721, <xref ref-type="fig" rid="F0006">Fig. 6</xref>).</p><fig id="F0006" position="float"><label>Fig. 6</label><caption><p>Clinical points distribution of callus tissue.</p></caption><graphic xlink:href="FNR-58-24998-g006"/></fig></sec><sec id="S0003-S20003"><title>Histopathological results</title><p>Following the clinical and radiological examinations, the preparations were histopathologically evaluated according to the criteria described by Huo et al. (<xref rid="CIT0012" ref-type="bibr">12</xref>). In this evaluation, the control group mean points were 8.5 (minimum 8.0&#x02013;maximum 9.0), and the immunonutrition group mean points were 9.0 (minimum 9.0&#x02013;maximum 9.0) (Figs. <xref ref-type="fig" rid="F0007">7</xref>&#x02013;<xref ref-type="fig" rid="F0011">11</xref> ).</p><fig id="F0007" position="float"><label>Fig. 7</label><caption><p>Histopathological points distribution.</p></caption><graphic xlink:href="FNR-58-24998-g007"/></fig><fig id="F0008" position="float"><label>Fig. 8</label><caption><p>Control group stage 8 recovery: Image shows immature bone (IB) and little cartilage (C) tissue (haemotoxylin and eosin staining &#x000d7;100 magnification).</p></caption><graphic xlink:href="FNR-58-24998-g008"/></fig><fig id="F0009" position="float"><label>Fig. 9</label><caption><p>Control group stage 8 recovery: Image shows immature bone (IB) and little cartilage (C) tissue (haemotoxylin and eosin staining &#x000d7;200 magnification).</p></caption><graphic xlink:href="FNR-58-24998-g009"/></fig><fig id="F0010" position="float"><label>Fig. 10</label><caption><p>Immunonutrition group stage 9 recovery: section from fracture line (FL) showing complete immature bone tissue (IB) (haemotoxylin and eosin staining&#x000d7;40 magnification).</p></caption><graphic xlink:href="FNR-58-24998-g010"/></fig><fig id="F0011" position="float"><label>Fig. 11</label><caption><p>Immunonutrition group stage 9 recovery: complete immature bone tissue (IB) (haemotoxylin and eosin staining &#x000d7;100 magnification).</p></caption><graphic xlink:href="FNR-58-24998-g011"/></fig></sec></sec><sec id="S0004"><title>Discussion and conclusion</title><p>In recent years several scientific studies have focussed on glutamine as a non-essential amino acid as it has been determined to have significant physiological roles. Glutamine facilitates the process of rapid cell reproduction as it is a leading source of both energy and biosynthesis (<xref rid="CIT0013" ref-type="bibr">13</xref>, <xref rid="CIT0014" ref-type="bibr">14</xref>). Particularly in critical cases such as multiple injuries, sepsis, or burns, glutamine is used as an essential amino acid as it undertakes the precursor task for nucleic acids, nucleotides, amino sugars, and proteins (<xref rid="CIT0003" ref-type="bibr">3</xref>). Under catabolic conditions, glutamine takes on the aspects of an amino acid (<xref rid="CIT0003" ref-type="bibr">3</xref>, <xref rid="CIT0015" ref-type="bibr">15</xref>). When muscle tissue is exposed to excessive stress, the most important of the amino acids which provide a repository of some amino acids, is glutamine (<xref rid="CIT0016" ref-type="bibr">16</xref>). A fall in the level of glutamine is associated with poor clinical results, as starting with a decreased amount in the process of muscle destruction, morbidity and mortality rates may increase. Parenteral or enteral administration of glutamine to critical patients leads to decreased complications associated with sepsis and better functional results (<xref rid="CIT0017" ref-type="bibr">17</xref>, <xref rid="CIT0018" ref-type="bibr">18</xref>). Glutamine has the essential properties of an amino acid in respect to affecting lymphocyte proliferation in the most appropriate way by neutrophil and macrophage functions <xref rid="CIT0019" ref-type="bibr">19</xref>&#x02013;<xref rid="CIT0022" ref-type="bibr">22</xref>). By eliminating a lack of glutamine, it is possible to restore nitrogen balance and improve immunosuppression (<xref rid="CIT0017" ref-type="bibr">17</xref>).</p><p>To the best of our knowledge there are no studies on the effect of orally administered glutamine on fracture healing. In this study, taking various beneficial effects into consideration, it was thought that glutamine could make a positive contribution to fracture healing, and in this context, the similarity of wound healing to fracture healing was the guide.</p><p>As a single indeterminate structure, glutamine quickly becomes denatured in water. The dipeptide structure formed when glutamine is combined with alanine may remain stable in aqueous solutions. L-amino acids are more easily absorbed than d-amino acids (<xref rid="CIT0016" ref-type="bibr">16</xref>). Therefore, in this study an l-alanine l-glutamine solution was used to achieve absorption with active transport of the glutamine and to prevent rapid denaturing.</p><p>In an experimental study of 50 rabbits with fibula fracture and femoral condyle defects, and orally nourished with essential amino acids containing lactose, Fini et al. (<xref rid="CIT0023" ref-type="bibr">23</xref>) reported a positive effect on fracture healing. Although the current study investigated the effect on rabbit fibula fracture of glutamine as a different essential amino acid, both studies showed a similarity in respect to the chosen animal model and the choice of amino acid given. In the fracture model created in the rabbits, cartilage islands started to form on the sixth day of fracture healing, on the 12th day the centre was filled with cartilage tissue, and in the third and fourth week bone tissue replaced the cartilage showing complete union. Thus, in the current study, the experiment was terminated in the fourth week as a sufficient time for fracture healing in rabbits.</p><p>Although various techniques are used for the radiological evaluation of fracture healing, most of these have been shown to be inadequate as they are not objective and vary from person to person. In some studies, the bridging status between the fractured ends has been evaluated from direct radiographs (<xref rid="CIT0009" ref-type="bibr">9</xref>, <xref rid="CIT0024" ref-type="bibr">24</xref>). In the current study, the widely used Lane and Sandhu (<xref rid="CIT0009" ref-type="bibr">9</xref>) scoring system was used in the radiological evaluation by more than one orthopaedic, traumatology, and radiology specialist (<xref rid="CIT0025" ref-type="bibr">25</xref>&#x02013;<xref rid="CIT0028" ref-type="bibr">28</xref>). We are of the opinion that this provides a more objective evaluation.</p><p>In clinical application, bone union is monitored with both physical examinations and radiological tests. Similarly, in the current study, the presence of bone union was investigated by forced movements to the fractured fibula in two planes, and radiology provided an additional evaluation. Dimar et al. (<xref rid="CIT0010" ref-type="bibr">10</xref>) emphasised clinical examination in two planes for stability evaluation and the approach in the current study supports this.</p><p>Some histopathological evaluations related to fracture healing have given qualitative results rather than numerical values (<xref rid="CIT0029" ref-type="bibr">29</xref>, <xref rid="CIT0030" ref-type="bibr">30</xref>). Allen et al. (<xref rid="CIT0031" ref-type="bibr">31</xref>&#x02013;<xref rid="CIT0033" ref-type="bibr">33</xref>) reported a 5-stage histopathological evaluation. However, Huo et al. (<xref rid="CIT0012" ref-type="bibr">12</xref>) recommended a more comprehensive method, which was used in the current study. By examination of 10 stages in this system, more detailed and accurate results are obtained (<xref rid="CIT0028" ref-type="bibr">28</xref>, <xref rid="CIT0034" ref-type="bibr">34</xref>, <xref rid="CIT0035" ref-type="bibr">35</xref>).</p><p>Physiological and pathological events in living organisms occur as a result of substances which may have harmful effects. The most important of these are free oxygen radicals which are balanced by the body's antioxidant system (<xref rid="CIT0036" ref-type="bibr">36</xref>). The effects of antioxidant molecules on fracture healing have been examined with particular focus on some vitamins and essential amino acids (<xref rid="CIT0004" ref-type="bibr">4</xref>, <xref rid="CIT0031" ref-type="bibr">31</xref>, <xref rid="CIT0033" ref-type="bibr">33</xref>, <xref rid="CIT0037" ref-type="bibr">37</xref>&#x02013;<xref rid="CIT0039" ref-type="bibr">39</xref>). As part of the antioxidant system, glutamine being a precursor of the glutathione-peroxidase enzyme, also serves as an antioxidant. Studies by Durak et al. (<xref rid="CIT0040" ref-type="bibr">40</xref>) and Saris&#x000f6;zen et al. (<xref rid="CIT0031" ref-type="bibr">31</xref>) emphasised that the effect of antioxidants in reducing free oxygen radicals found in fracture haematoma may have a positive effect on fracture healing. Therefore, in the current study it was thought that as well as other effects, glutamine as an antioxidant could positively influence fracture healing by reducing free oxygen radicals found in the fracture area, but indicators related to this were not examined.</p><p>In a study on rabbits by Sinha and Goal (<xref rid="CIT0041" ref-type="bibr">41</xref>), where the amino acids lysine and arginine were administered orally, increased callus vascularisation and mineralisation was reported and the healing period was accelerated to 2 weeks. In an experimental study by Kdolsky et al. (<xref rid="CIT0042" ref-type="bibr">42</xref>) a femoral diaphysis defect was created in guinea pigs and intramedullary fixation was applied to the femur with a k-wire. The findings were evaluated histologically, mechanically, and radiologically and it was concluded that the animals which had received L-Arginine had better fracture healing and mechanical stability. Comparing the current study to that one, the amino acid used was glutamine, the fibula was used rather than the femur, and there was no mechanical evaluation. Therefore, although the idea of the effect of amino acids on fracture healing is similar, from a technical aspect there is no similarity to the current study.</p><p>In the histological evaluation of the current study, all the rabbits that had received glutamine were seen to be at stage 9 full recovery, whereas half of the control group were at stage 8 recovery with cartilage and immature bone together and the others were at stage 9 recovery. This finding justifies the view that glutamine has a positive effect on fracture healing. However, no statistically significant difference was determined between the two groups. Similarly, in the radiological evaluation, the immunonutrition group was determined as mean 3.42 points from a total of 5 by the orthopaedic and traumatology specialists and mean 3.35 points by the radiology specialists whereas the control group points were low at mean 2.5 and 2.65 respectively. Despite the difference between them, there was no statistically significant difference. This is thought to be because of the low number of animals used in this study. It is also thought that histological evaluation at an earlier stage of healing may yield significant results.</p><p>In conclusion, despite the beneficial effects of glutamine on the speed of wound healing and reducing morbidity and mortality rates in sepsis, and the positive effect observed on bone healing in the immunonutrition group clinically, radiologically, and histopathologically, no statistically significant difference was determined compared to the control group in this study. Additional light will be shed on this subject in literature by subsequent research considering different parameters.</p></sec>

Foot-and-mouth disease virus leader proteinase: Structural insights into the mechanism of intermolecular cleavage

Could not extract abstract

<contrib contrib-type="author" id="au0005"><name><surname>Steinberger</surname><given-names>Jutta</given-names></name><xref rid="aff0005" ref-type="aff">a</xref></contrib><contrib contrib-type="author" id="au0010"><name><surname>Grishkovskaya</surname><given-names>Irina</given-names></name><xref rid="aff0010" ref-type="aff">b</xref></contrib><contrib contrib-type="author" id="au0015"><name><surname>Cencic</surname><given-names>Regina</given-names></name><xref rid="aff0005" ref-type="aff">a</xref><xref rid="fn1" ref-type="fn">1</xref></contrib><contrib contrib-type="author" id="au0020"><name><surname>Juliano</surname><given-names>Luiz</given-names></name><xref rid="aff0015" ref-type="aff">c</xref></contrib><contrib contrib-type="author" id="au0025"><name><surname>Juliano</surname><given-names>Maria A.</given-names></name><xref rid="aff0015" ref-type="aff">c</xref></contrib><contrib contrib-type="author" id="au0030"><name><surname>Skern</surname><given-names>Tim</given-names></name><email>timothy.skern@meduniwien.ac.at</email><xref rid="aff0005" ref-type="aff">a</xref><xref rid="cor1" ref-type="corresp">&#x0204e;</xref></contrib><aff id="aff0005"><label>a</label>Max F. Perutz Laboratories, Medical University of Vienna, Dr. Bohr-Gasse 9/3, A-1030 Vienna, Austria</aff><aff id="aff0010"><label>b</label>Max F. Perutz Laboratories, University of Vienna, Department of Structural and Computational Biology, Campus Vienna Biocenter 5, A-1030 Vienna, Austria</aff><aff id="aff0015"><label>c</label>Department of Biophysics, Escola Paulista de Medicina, Universidade Federal de S&#x000e3;o Paulo, Rua Tr&#x000ea;s de Maio 100, 04044-20 S&#x000e3;o Paulo, Brazil</aff>

Virology

<sec id="s0005"><title>Introduction</title><p id="p0030">Virally encoded proteinases play essential roles not only in the processing of the viral proteins but also in cleavage of host cell proteins in order to manipulate cellular processes to the advantage of the virus. One of the first such reactions to be documented was the modification of cellular translation factors during picornaviral replication leading to the shut-off of protein synthesis from capped cellular mRNA (<xref rid="bib13" ref-type="bibr">Etchison et al., 1982</xref>, <xref rid="bib27" ref-type="bibr">Leibowitz and Penman, 1971</xref>). This reaction was subsequently shown to be performed by the 2A proteinase (2A^pro) in enteroviruses (<xref rid="bib24" ref-type="bibr">Kr&#x000e4;usslich et al., 1987</xref>), a chymotrypsin-like cysteine proteinase (<xref rid="bib35" ref-type="bibr">Petersen et al., 1999</xref>), whereas in aphthoviruses, the proteolysis is performed by the leader proteinase (L^pro, illustrated in <xref rid="f0005" ref-type="fig">Fig. 1</xref>) (<xref rid="bib11" ref-type="bibr">Devaney et al., 1988</xref>), a papain-like cysteine proteinase (<xref rid="bib20" ref-type="bibr">Guarn<bold>&#x000e9;</bold> et al., 1998</xref>). The targets of both proteinases are the two homologues of the host protein eukaryotic initiation factor (eIF) 4G (<xref rid="bib15" ref-type="bibr">Gingras et al., 1999</xref>). Cleavage of the eIF4G homologues prevents recruitment of capped mRNAs to the ribosome (<xref rid="bib26" ref-type="bibr">Lamphear et al., 1995</xref>) whereas viral RNA can still be translated under these conditions as it initiates via an internal ribosome entry segment (IRES) (<xref rid="bib29" ref-type="bibr">Martinez-Salas and Ryan, 2010</xref>). In addition, L^pro has been shown to be involved in impairing the host innate immune defence by influencing NF-&#x003ba;B activation and to have deubiquitinase activity (<xref rid="bib8" ref-type="bibr">de Los Santos et al., 2007</xref>, <xref rid="bib9" ref-type="bibr">de los Santos et al., 2009</xref>, <xref rid="bib40" ref-type="bibr">Skern and Steinberger, 2014</xref>).<fig id="f0005"><label>Fig. 1</label><caption><p>Schematic drawing of FMDV L^pro self-processing and eIF4G cleavage. (A) The FMDV RNA genome is shown as a black line, the single open reading frame as a box with the names of the mature proteins and the position of the IRES. L^pro, being expressed either as Lab^pro or Lb^pro, is indicated in red. (B) Synthesis of the polyprotein from the FMDV genome showing that Lb^pro can either be freed by an intramolecular or intermolecular reaction. sLb^pro (shown in orange) is generated by self-processing at the C-terminus of Lb^pro. (C) The effect of eIF4G cleavage by Lb^pro or sLb^pro. The cellular mRNA is shown as a black line with the cap structure as a filled circle. Lb^pro and sLb^pro are shown in red and orange, respectively. The 40S ribosomal subunit, the polyA-binding protein (PABP), eIF4G, eIF4E, eIF4A and eIF3 are shown in different shades of grey. Following cleavage of eIF4G by Lb^pro or sLb^pro, the capped mRNA is no longer connected to the 40S subunit and cannot be translated. In contrast, the viral RNA can bind to the C-terminal fragment of eIF4G and thus to the 40S subunit via eIF3.</p></caption><graphic xlink:href="gr1"/></fig></p><p id="p0035">Given these involvements in such different reactions as intramolecular and intermolecular self-processing, eIF4G cleavage and deubiquitination, it is not surprising that L^pro has unusual specificity determinants. These are well illustrated by the sequences of the three L^pro cleavage sites that have been determined directly by protein sequencing: KVQRKLK&#x0204e;GAGQSS for both intra- and intermolecular cleavage on the viral polyprotein between the C-terminus of L^pro and VP4 (<xref rid="bib43" ref-type="bibr">Strebel and Beck, 1986</xref>), PSFANLG&#x0204e;RTTLST on eIF4GI (<xref rid="bib23" ref-type="bibr">Kirchweger et al., 1994</xref>) and VPLLNVG&#x0204e;SRRSQP on eIF4GII (<xref rid="bib17" ref-type="bibr">Gradi et al., 2004</xref>). Studies on L^pro intramolecular self-processing and cleavage of peptide substrates have revealed that L^pro can cleave before or after basic residues provided that the other amino acid before or after the scissile bond is glycine (<xref rid="bib16" ref-type="bibr">Glaser et al., 2001</xref>, <xref rid="bib33" ref-type="bibr">Nogueira Santos et al., 2012</xref>, <xref rid="bib38" ref-type="bibr">Santos et al., 2009</xref>). However, a peptide that contained basic residues before and after the scissile bond was refractory to cleavage and was subsequently shown to be an inhibitor in the micromolar range (<xref rid="bib38" ref-type="bibr">Santos et al., 2009</xref>). This information was then used to develop a nanomolar epoxide inhibitor based on E64, termed E64-R-P-NH₂ (<xref rid="bib38" ref-type="bibr">Santos et al., 2009</xref>); the structure and inhibitor parameters are shown in <xref rid="f0010" ref-type="fig">Fig. 2</xref>, together with those of the other inhibitors used or referred to in this work. The slow formation of the tight enzyme&#x02013;inhibitor complex indicates that inhibition follows slow-binding kinetics (<xref rid="bib38" ref-type="bibr">Santos et al., 2009</xref>, <xref rid="bib49" ref-type="bibr">Zhou et al., 1998</xref>).<fig id="f0010"><label>Fig. 2</label><caption><p>Chemical structures of inhibitors referred to in this work. The structures and kinetic parameters of the inhibitor E64-R-P-NH₂ (<xref rid="bib33" ref-type="bibr">Nogueira Santos et al., 2012</xref>) crystallised with sLb^pro are shown together with those of the inhibitors NS-134 (<xref rid="bib42" ref-type="bibr">Stern et al., 2004</xref>) and CA074 (<xref rid="bib48" ref-type="bibr">Yamamoto et al., 1997</xref>) whose structures were determined in complex with cathepsin B. The correspondence of side-chains in the inhibitors to substrate side-chains is shown using the nomenclature of <xref rid="bib39" ref-type="bibr">Schechter and Berger, (19670</xref>.</p></caption><graphic xlink:href="gr2"/></fig></p><p id="p0040">The structural basis for this unusual specificity has not been elucidated, as the present structures determined by X-ray crystallography and NMR (<xref rid="bib5" ref-type="bibr">Cencic et al., 2007</xref>, <xref rid="bib20" ref-type="bibr">Guarn&#x000e9; et al., 1998</xref>, <xref rid="bib19" ref-type="bibr">Guarn&#x000e9; et al., 2000</xref>, <xref rid="bib41" ref-type="bibr">Steinberger et al., 2013</xref>) only provide information on the S binding region but not on the S&#x02032; binding region of L^pro. The nomenclature for sites (S) on the enzyme binding to residues of substrate (P) is that of <xref rid="bib39" ref-type="bibr">Schechter and Berger (1967)</xref>; prime site residues are those C-terminal to the scissile bond. Indeed, information on the nature of the S&#x02032; region from related papain-like proteinases is also sparse (<xref rid="bib46" ref-type="bibr">Turk et al., 2012</xref>), with structural information only being available for cathepsin B (<xref rid="bib42" ref-type="bibr">Stern et al., 2004</xref>, <xref rid="bib45" ref-type="bibr">Turk et al., 1995</xref>, <xref rid="bib48" ref-type="bibr">Yamamoto et al., 1997</xref>) determined with inhibitors similar to E64-R-P-NH₂ (<xref rid="f0010" ref-type="fig">Fig. 2</xref>). However, cathepsin B is also unusual in being an exopeptidase, with an occluding loop that prevents access beyond the S2&#x02032; site, that is the site on the enzyme interacting with the P2&#x02032; residue of the substrate (<xref rid="bib42" ref-type="bibr">Stern et al., 2004</xref>). Thus, any information on the S&#x02032; binding region of FMDV L^pro will shed light on the nature of this region in papain-like cysteine proteinases generally.</p><p id="p0045">Understanding of the mechanism of L^pro is complicated by the presence of different forms of the protein in the infected cell (<xref rid="bib37" ref-type="bibr">Sangar et al., 1987</xref>, <xref rid="bib36" ref-type="bibr">Sangar et al., 1988</xref>). Two isoforms, Lab^pro and Lb^pro (<xref rid="f0005" ref-type="fig">Fig. 1</xref>), arise from the presence of two in-frame AUG codons for the initiation of protein synthesis on the viral RNA (<xref rid="bib37" ref-type="bibr">Sangar et al., 1987</xref>). Consequently, the Lab^pro possesses an additional 28 amino acids at the N-terminus than Lb^pro. <xref rid="bib4" ref-type="bibr">Cao et al. (1995)</xref> demonstrated in cell culture that Lb^pro was essential whereas Lab^pro was not; nevertheless, there may still be as yet unknown roles for Lab^pro during infection in the host organism. In addition, a shortened form of Lb^pro (sLb^pro) lacking 6 or 7 amino acids at the C-terminus has long been known (<xref rid="bib36" ref-type="bibr">Sangar et al., 1988</xref>). The truncation arises through Lb^pro self-cleavage (<xref rid="bib36" ref-type="bibr">Sangar et al., 1988</xref>) and can be observed in vitro when Lb^pro expressed in rabbit reticulocyte lysates (RRLs) is incubated for longer time periods (e.g. 1&#x000a0;h)(<xref rid="bib30" ref-type="bibr">Mayer et al., 2008</xref>). A separate function for sLb^pro has not been identified; however, one report suggested that Lb^pro and sLb^pro may differ in their cleavage efficiencies in intermolecular cleavage of the polyprotein substrate (<xref rid="bib5" ref-type="bibr">Cencic et al., 2007</xref>).</p><p id="p0050">One clear difference between Lb^pro and sLb^pro is the ability of Lb^pro to form homodimers through interactions of the C-terminal extension (CTE) of one monomer and the substrate binding site of the neighbouring one and vice versa. sLb^pro cannot form homodimers in this way because it lacks the six most C-terminal residues. The Lb^pro homodimer has been observed by X-ray crystallography and NMR (<xref rid="bib5" ref-type="bibr">Cencic et al., 2007</xref>, <xref rid="bib20" ref-type="bibr">Guarn&#x000e9; et al., 1998</xref>, <xref rid="bib19" ref-type="bibr">Guarn&#x000e9; et al., 2000</xref>, <xref rid="bib41" ref-type="bibr">Steinberger et al., 2013</xref>) with the K_D being estimated from NMR analyses to be in the millimolar range (<xref rid="bib5" ref-type="bibr">Cencic et al., 2007</xref>). Therefore, formation of the homodimer at concentrations of Lb^pro achieved when it is synthesised in the infected cell seems unlikely unless there is a high local concentration. In contrast, both Lb^pro and sLb^pro use an exosite featuring residues Tyr183 to Leu188 as well as Cys 133 to recognise binding sites located on the eIF4G homologues, located in both cases 20 to 30 amino acids from the cleavage site (<xref rid="bib14" ref-type="bibr">Foeger et al., 2005</xref>). How this binding favours Lb^pro or sLb^pro cleavage of the eIF4G homologues is not known.</p><p id="p0055">To investigate further the properties of sLb^pro, we set out to determine the structure of sLb^pro complexed with the inhibitor E64-R-P-NH₂ and to define differences in the cleavage of intermolecular polyprotein substrates by sLb^pro and Lb^pro.</p></sec><sec id="s0010"><title>Materials and methods</title><sec id="s0015"><title>Materials</title><p id="p0060">The bacterial expression plasmid pET-11d sLb^pro (FMDV residues 29&#x02013;195) was created by site-directed PCR mutagenesis of pET-11d sLb^pro C51A, described earlier (<xref rid="bib20" ref-type="bibr">Guarn&#x000e9; et al., 1998</xref>, <xref rid="bib23" ref-type="bibr">Kirchweger et al., 1994</xref>), to restore the catalytic cysteine.</p><p id="p0065">The plasmids that were used as templates for in vitro transcription pCITE-1d Lb^pro (residues 29&#x02013;201 of Lb^pro), pCITE-1d sLb^pro (residues 29&#x02013;195 of Lb^pro) and pCITE-1d Lb^pro C51A VP4/VP2 (residues 29&#x02013;201 of Lb^pro, all 85 residues of VP4 and 78 residues of VP2) have been described (<xref rid="bib16" ref-type="bibr">Glaser et al., 2001</xref>). The constructs pCITE-1d Lb^pro C51A VP4/VP2 containing the mutations at position P1 and P1&#x02032; of the Lb^pro-VP4 cleavage site (VQRKLG&#x0204e;RAGQ, VQRKLK&#x0204e;RAGQ, VQRKLG&#x0204e;AAGQ) were created by site-directed PCR mutagenesis of pCITE-1d Lb^pro C51A VP4/VP2. The construct pCITE-1d Lb^pro C51A VP4/VP2 containing the eIF4GI sequence SFANLG&#x0204e;RTTL at the Lb^pro-VP4 cleavage site, termed pCITE-1d Lb^pro C51A VP4/VP2 SFANLG&#x0204e;RTTL (FMDV residues 29&#x02013;195 of Lb^pro, residues 669&#x02013;678 of eIF4GI, residue 5&#x02013;85 of VP4 and 78 residues of VP2), has been described (<xref rid="bib5" ref-type="bibr">Cencic et al., 2007</xref>). The construct pCITE-1d Lb^pro C51A VP4/VP2 containing residues 599&#x02013;678 of eIF4GI, termed pCITE-1d Lb^pro C51A eIF4GI_599-668 VP4/VP2 SFANLG&#x0204e;RTTL, (FMDV residues 29&#x02013;195 of Lb^pro, residues 599&#x02013;678 of eIF4GI, residue 5&#x02013;85 of VP4 and 78 residues of VP2) was created by PCR amplification of residues 599&#x02013;678 of eIF4GI using the plasmid pKS eIF4GI 400&#x02013;739 as template and cloning of this fragment into pCITE-1d Lb^pro C51A VP4/VP2 via the restriction sites <italic>Bpu10</italic>I and <italic>Sac</italic>I.</p><p id="p0070">The inhibitor E64-R-P-NH₂ was prepared as described (<xref rid="bib33" ref-type="bibr">Nogueira Santos et al., 2012</xref>).</p></sec><sec id="s0020"><title>Protein expression and purification</title><p id="p0075">Protein expression and purification were performed as described by <xref rid="bib41" ref-type="bibr">Steinberger et al. (2013)</xref> with the following modifications. Proteins were expressed from the construct pET-11d sLb^pro transformed into BL21(DE3)LysE bacteria. To avoid degradation of the active protease, all purification steps were carried at a maximum of 10&#x000a0;&#x000b0;C.</p></sec><sec id="s0025"><title>Preparation of the sLb^pro-E64-R-P-NH₂ complex</title><p id="p0080">Purified sLb^pro was incubated with a fivefold molar excess of E64-R-P-NH₂ over night at 4&#x000a0;&#x000b0;C to allow complex formation. Subsequently, the complex was dialysed against a buffer containing 50&#x000a0;mM NaCl, 10&#x000a0;mM Tris HCl pH 8, 1&#x000a0;mM TCEP, 5% glycerol to remove excess inhibitor. The concentration was adjusted to 18&#x000a0;mg/ml and centrifuged at 18,000<italic>g</italic> for 10&#x000a0;min at 4&#x000a0;&#x000b0;C to remove precipitated protein.</p></sec><sec id="s0030"><title>Crystallisation, data collection, structure determination and refinement</title><p id="p0085">Crystals of the sLb^pro-E64-R-P-NH₂ complex were initially obtained in the Wizard I and II screen crystallisation screen (Emerald Bio), using the sitting-drop vapour diffusion technique and a nanodrop-dispensing robot (Phoenix RE; Rigaku Europe, Kent, United Kingdom), and optimised to 0.1&#x000a0;M sodium acetate pH 4.8, 0.9&#x000a0;M NaH₂PO₄ and 1.2&#x000a0;M K₂HPO₄ using the hanging drop vapour diffusion technique at 22&#x000a0;&#x000b0;C and seeding technique. The seed stock was produced by a &#x0201c;seed-bead&#x0201d; kit from Hampton Research (<xref rid="bib28" ref-type="bibr">Luft and DeTitta, 1999</xref>). The crystals were flash-frozen in liquid nitrogen in a reservoir solution supplemented with 25% glycerol prior to data collection.</p><p id="p0090">Diffraction data sets were collected at the ESRF Synchrotron (Grenoble) at beamline ID14-1 at 100&#x000a0;K using a wavelength of 0.93&#x000a0;&#x000c5; to 1.6&#x000a0;&#x000c5; resolution, processed using the XDS package (<xref rid="bib22" ref-type="bibr">Kabsch<bold>,</bold> 2010</xref>), converted to mtz format using POINTLESS and scaled with SCALA (<xref rid="bib47" ref-type="bibr">Winn et al., 2011</xref>).</p><p id="p0095">The crystal structure was solved by difference Fourier techniques using the protein atomic coordinates of the inactive mutant of sLb^pro from the Protein Data Bank (accession code 1QMY). Model building and refinement steps were performed with REFMAC and COOT. The structure was refined using the programs REFMAC (<xref rid="bib31" ref-type="bibr">Murshudov et al., 1997</xref>) and Phenix Refine (<xref rid="bib1" ref-type="bibr">Adams et al., 2010</xref>) and model building was done with the program Coot (<xref rid="bib12" ref-type="bibr">Emsley and Cowtan, 2004</xref>). Data collection and refinement statistics are shown in <xref rid="t0005" ref-type="table">Table 1</xref>. Stereo-chemistry and structure quality were checked using the MolProbity web server (<xref rid="bib7" ref-type="bibr">Davis et al., 2007</xref>).<table-wrap position="float" id="t0005"><label>Table 1</label><caption><p>X-ray parameters and refinement statistics.</p></caption><table frame="hsides" rules="groups"><thead><tr><th colspan="2" valign="middle"><bold>Data collection</bold></th></tr></thead><tbody><tr><td valign="middle">Source</td><td valign="middle">ID14-1, ESRF</td></tr><tr><td valign="middle">Wavelength (&#x000c5;)</td><td valign="middle">0.93</td></tr><tr><td valign="middle">Resolution (&#x000c5;)</td><td valign="middle">45.35&#x02013;1.6 (1.69&#x02013;1.6)<xref rid="tbl1fna" ref-type="table-fn">a</xref></td></tr><tr><td valign="middle">Space group</td><td valign="middle">P2₁</td></tr><tr><td rowspan="2" valign="middle">Unit cell (&#x000c5;, &#x000b0;)</td><td valign="middle"><italic>a</italic>=45.81 <italic>b</italic>=110.68, <italic>c</italic>=56.77</td></tr><tr><td valign="middle"><italic>&#x003b1;</italic>=<italic>&#x003b3;</italic>=90, <italic>&#x003b2;</italic>=98.12</td></tr><tr><td valign="middle">Molecules / a.u.</td><td valign="middle">3</td></tr><tr><td valign="middle">Unique reflections</td><td valign="middle">72906 (10232)</td></tr><tr><td valign="middle">Completeness (%)</td><td valign="middle">99.0 (95.0)</td></tr><tr><td valign="middle"><italic>R</italic>_merge<xref rid="tbl1fnb" ref-type="table-fn">b</xref></td><td valign="middle">0.037 (0.174)</td></tr><tr><td valign="middle"><italic>R</italic>_meas<xref rid="tbl1fnc" ref-type="table-fn">c</xref></td><td valign="middle">0.041 (0.213)</td></tr><tr><td valign="middle">Multiplicity</td><td valign="middle">4.9 (2.9)</td></tr><tr><td valign="middle">I/sig(I)</td><td valign="middle">29.9 (5.6)</td></tr><tr><td valign="middle">B_Wilson (&#x000c5;²)</td><td valign="middle">22.5</td></tr><tr><td colspan="2" valign="middle"><bold>Refinement</bold></td></tr><tr><td valign="middle"><italic>R</italic>_cryst<xref rid="tbl1fnd" ref-type="table-fn">d</xref>/ R_free<xref rid="tbl1fne" ref-type="table-fn">e</xref></td><td valign="middle">16.9/20.1</td></tr><tr><td valign="middle">R.m.s.d. bonds (&#x000c5;)</td><td valign="middle">0.011</td></tr><tr><td valign="middle">R.m.s.d. angles (&#x000b0;)</td><td valign="middle">1.4</td></tr><tr><td valign="middle">Ramachandran plot (%)</td><td valign="middle"/></tr><tr><td valign="middle">favored/allowed/outliers</td><td valign="middle">96.9/3.1/0</td></tr></tbody></table><table-wrap-foot><fn id="tbl1fna"><label>a</label><p id="ntp0005">Values in parentheses are for the highest resolution shell.</p></fn></table-wrap-foot><table-wrap-foot><fn id="tbl1fnb"><label>b</label><p id="ntp0010"><inline-formula><mml:math id="M1" altimg="si0001.gif" overflow="scroll"><mml:msub><mml:mrow><mml:mi>R</mml:mi></mml:mrow><mml:mrow><mml:mi>m</mml:mi><mml:mi>e</mml:mi><mml:mi>r</mml:mi><mml:mi>g</mml:mi><mml:mi>e</mml:mi></mml:mrow></mml:msub><mml:mo>=</mml:mo><mml:mstyle displaystyle="true"><mml:mfrac><mml:mrow><mml:msub><mml:mo>&#x02211;</mml:mo><mml:mrow><mml:mi>h</mml:mi><mml:mi>k</mml:mi><mml:mi>l</mml:mi></mml:mrow></mml:msub><mml:mrow><mml:msubsup><mml:mo>&#x02211;</mml:mo><mml:mrow><mml:mi>i</mml:mi><mml:mo>=</mml:mo><mml:mn>1</mml:mn></mml:mrow><mml:mi>N</mml:mi></mml:msubsup><mml:mrow><mml:mo>|</mml:mo><mml:msub><mml:mrow><mml:mi>I</mml:mi></mml:mrow><mml:mrow><mml:mi>i</mml:mi><mml:mo stretchy="false">(</mml:mo><mml:mi>h</mml:mi><mml:mi>k</mml:mi><mml:mi>l</mml:mi><mml:mo stretchy="false">)</mml:mo></mml:mrow></mml:msub><mml:mo>&#x02212;</mml:mo><mml:msub><mml:mrow><mml:mover accent="true"><mml:mi>I</mml:mi><mml:mo>&#x000af;</mml:mo></mml:mover></mml:mrow><mml:mrow><mml:mo stretchy="false">(</mml:mo><mml:mi>h</mml:mi><mml:mi>k</mml:mi><mml:mi>l</mml:mi><mml:mo stretchy="false">)</mml:mo></mml:mrow></mml:msub><mml:mo>|</mml:mo></mml:mrow></mml:mrow></mml:mrow><mml:mrow><mml:msub><mml:mo>&#x02211;</mml:mo><mml:mrow><mml:mi>h</mml:mi><mml:mi>k</mml:mi><mml:mi>l</mml:mi></mml:mrow></mml:msub><mml:mrow><mml:msubsup><mml:mo>&#x02211;</mml:mo><mml:mrow><mml:mi>i</mml:mi><mml:mo>=</mml:mo><mml:mn>1</mml:mn></mml:mrow><mml:mi>N</mml:mi></mml:msubsup><mml:mrow><mml:msub><mml:mrow><mml:mi>I</mml:mi></mml:mrow><mml:mrow><mml:mi>i</mml:mi><mml:mo stretchy="false">(</mml:mo><mml:mi>h</mml:mi><mml:mi>k</mml:mi><mml:mi>l</mml:mi><mml:mo stretchy="false">)</mml:mo></mml:mrow></mml:msub></mml:mrow></mml:mrow></mml:mrow></mml:mfrac></mml:mstyle></mml:math></inline-formula></p></fn></table-wrap-foot><table-wrap-foot><fn id="tbl1fnc"><label>c</label><p id="ntp0015"><inline-formula><mml:math id="M2" altimg="si0002.gif" overflow="scroll"><mml:msub><mml:mrow><mml:mi>R</mml:mi></mml:mrow><mml:mrow><mml:mi>m</mml:mi><mml:mi>e</mml:mi><mml:mi>a</mml:mi><mml:mi>s</mml:mi></mml:mrow></mml:msub><mml:mo>=</mml:mo><mml:mstyle displaystyle="true"><mml:mfrac><mml:mrow><mml:msub><mml:mo>&#x02211;</mml:mo><mml:mrow><mml:mi>h</mml:mi><mml:mi>k</mml:mi><mml:mi>l</mml:mi></mml:mrow></mml:msub><mml:mrow><mml:msqrt><mml:mrow><mml:mi>N</mml:mi><mml:mo>/</mml:mo><mml:mrow><mml:mo stretchy="false">(</mml:mo><mml:mi>N</mml:mi><mml:mo>&#x02212;</mml:mo><mml:mn>1</mml:mn><mml:mo stretchy="false">)</mml:mo></mml:mrow></mml:mrow></mml:msqrt><mml:msubsup><mml:mo>&#x02211;</mml:mo><mml:mrow><mml:mi>i</mml:mi><mml:mo>=</mml:mo><mml:mn>1</mml:mn></mml:mrow><mml:mi>N</mml:mi></mml:msubsup><mml:mrow><mml:mo>|</mml:mo><mml:msub><mml:mrow><mml:mi>I</mml:mi></mml:mrow><mml:mrow><mml:mi>i</mml:mi><mml:mo stretchy="false">(</mml:mo><mml:mi>h</mml:mi><mml:mi>k</mml:mi><mml:mi>l</mml:mi><mml:mo stretchy="false">)</mml:mo></mml:mrow></mml:msub><mml:mo>&#x02212;</mml:mo><mml:msub><mml:mrow><mml:mover accent="true"><mml:mi>I</mml:mi><mml:mrow><mml:mo stretchy="true">&#x000af;</mml:mo></mml:mrow></mml:mover></mml:mrow><mml:mrow><mml:mo stretchy="false">(</mml:mo><mml:mi>h</mml:mi><mml:mi>k</mml:mi><mml:mi>l</mml:mi><mml:mo stretchy="false">)</mml:mo></mml:mrow></mml:msub><mml:mo>|</mml:mo></mml:mrow></mml:mrow></mml:mrow><mml:mrow><mml:msub><mml:mo>&#x02211;</mml:mo><mml:mrow><mml:mi>h</mml:mi><mml:mi>k</mml:mi><mml:mi>l</mml:mi></mml:mrow></mml:msub><mml:mrow><mml:msubsup><mml:mo>&#x02211;</mml:mo><mml:mrow><mml:mi>i</mml:mi><mml:mo>=</mml:mo><mml:mn>1</mml:mn></mml:mrow><mml:mi>N</mml:mi></mml:msubsup><mml:mrow><mml:msub><mml:mrow><mml:mi>I</mml:mi></mml:mrow><mml:mrow><mml:mi>i</mml:mi><mml:mo stretchy="false">(</mml:mo><mml:mi>h</mml:mi><mml:mi>k</mml:mi><mml:mi>l</mml:mi><mml:mo stretchy="false">)</mml:mo></mml:mrow></mml:msub></mml:mrow></mml:mrow></mml:mrow></mml:mfrac></mml:mstyle></mml:math></inline-formula> where <inline-formula><mml:math id="M3" altimg="si0003.gif" overflow="scroll"><mml:msub><mml:mrow><mml:mover accent="true"><mml:mi>I</mml:mi><mml:mo>&#x000af;</mml:mo></mml:mover></mml:mrow><mml:mrow><mml:mo stretchy="false">(</mml:mo><mml:mi>h</mml:mi><mml:mi>k</mml:mi><mml:mi>l</mml:mi><mml:mo stretchy="false">)</mml:mo></mml:mrow></mml:msub></mml:math></inline-formula> is the mean intensity of multiple <inline-formula><mml:math id="M4" altimg="si0004.gif" overflow="scroll"><mml:msub><mml:mrow><mml:mi>I</mml:mi></mml:mrow><mml:mrow><mml:mi>i</mml:mi><mml:mo stretchy="false">(</mml:mo><mml:mi>h</mml:mi><mml:mi>k</mml:mi><mml:mi>l</mml:mi><mml:mo stretchy="false">)</mml:mo></mml:mrow></mml:msub></mml:math></inline-formula> observations of the symmetry-related reflections, <italic>N</italic> is the redundancy</p></fn></table-wrap-foot><table-wrap-foot><fn id="tbl1fnd"><label>d</label><p id="ntp0020"><inline-formula><mml:math id="M5" altimg="si0005.gif" overflow="scroll"><mml:msub><mml:mrow><mml:mi>R</mml:mi></mml:mrow><mml:mrow><mml:mi>c</mml:mi><mml:mi>r</mml:mi><mml:mi>y</mml:mi><mml:mi>s</mml:mi><mml:mi>t</mml:mi></mml:mrow></mml:msub><mml:mo>=</mml:mo><mml:mstyle displaystyle="true"><mml:mfrac><mml:mrow><mml:mo>&#x02211;</mml:mo><mml:mrow><mml:mo stretchy="true">|</mml:mo><mml:mrow><mml:mrow><mml:mo stretchy="true">|</mml:mo><mml:mrow><mml:msub><mml:mrow><mml:mi>F</mml:mi></mml:mrow><mml:mrow><mml:mi>o</mml:mi><mml:mi>b</mml:mi><mml:mi>s</mml:mi></mml:mrow></mml:msub></mml:mrow><mml:mo stretchy="true">|</mml:mo></mml:mrow><mml:mo>&#x02212;</mml:mo><mml:mrow><mml:mo stretchy="true">|</mml:mo><mml:mrow><mml:msub><mml:mrow><mml:mi>F</mml:mi></mml:mrow><mml:mrow><mml:mi>c</mml:mi><mml:mi>a</mml:mi><mml:mi>l</mml:mi><mml:mi>c</mml:mi></mml:mrow></mml:msub></mml:mrow><mml:mo stretchy="true">|</mml:mo></mml:mrow></mml:mrow><mml:mo stretchy="true">|</mml:mo></mml:mrow></mml:mrow><mml:mrow><mml:mo>&#x02211;</mml:mo><mml:mrow><mml:mo stretchy="true">|</mml:mo><mml:mrow><mml:msub><mml:mrow><mml:mi>F</mml:mi></mml:mrow><mml:mrow><mml:mi>o</mml:mi><mml:mi>b</mml:mi><mml:mi>s</mml:mi></mml:mrow></mml:msub></mml:mrow><mml:mo stretchy="true">|</mml:mo></mml:mrow></mml:mrow></mml:mfrac></mml:mstyle></mml:math></inline-formula></p></fn></table-wrap-foot><table-wrap-foot><fn id="tbl1fne"><label>e</label><p id="ntp0025"><italic>R</italic>_free is the cross-validation <italic>R</italic>_factor computed for the test set of reflections (5%) which are omitted in the refinement process.</p></fn></table-wrap-foot></table-wrap></p></sec><sec id="s0035"><title>In vitro transcription and translation</title><p id="p0100">In vitro transcription reactions were performed as described (<xref rid="bib32" ref-type="bibr">Neubauer et al., 2013</xref>) with the following modifications. The plasmids were cleaved with <italic>BamH</italic>I for the expression of active proteinases (pCITE-1d Lb^pro and pCITE-1d sLb^pro) and <italic>Sal</italic>I for the substrates (pCITE-1d Lb^pro C51A VP4/VP2 and derivatives thereof).</p><p id="p0105">In vitro translation reactions were performed as described (<xref rid="bib5" ref-type="bibr">Cencic et al., 2007</xref>) with the following modification. To translate substrate proteins and proteinases, RNA was added to the reaction at concentrations of 14&#x000a0;ng/&#x000b5;l.</p></sec><sec id="s0040"><title>Electrophoresis and immunoblotting</title><p id="p0110">Electrophoresis and immunoblotting for protein analysis were performed as described (<xref rid="bib32" ref-type="bibr">Neubauer et al., 2013</xref>), except for the separation of translation products when SDS-PAGE gels containing 17.5% acrylamide were used (<xref rid="bib6" ref-type="bibr">Dasso and Jackson, 1989</xref>).</p></sec><sec id="s0045"><title>Structural comparisons</title><p id="p0115">Structural alignments and superimpositions were done using Coot (<xref rid="bib12" ref-type="bibr">Emsley and Cowtan, 2004</xref>, <xref rid="bib25" ref-type="bibr">Krissinel and Henrick, 2004</xref>). All drawings were created using PyMOL (<xref rid="bib10" ref-type="bibr">DeLano<bold>,</bold> 2002</xref>). The electrostatic potential of sLb^pro was calculated using the Adaptive Poisson-Boltzmann Solver package (<xref rid="bib2" ref-type="bibr">Baker et al., 2001</xref>) within PyMOL.</p></sec><sec id="s0050"><title>Accession numbers</title><p id="p0120">Coordinates for the structure determined here have been deposited in the protein data bank (pdb accession code 4QBB). The PDB identifiers of the structures used for comparisons were 1QOL for Lb^pro, 1GEC for glycyl endopeptidase-complex with benzyloxycarbonyl-leucine-valine-glycine-methylene, 3CH3 for SERA5 from plasmodium falciparum, 1SP4 for bovine cathepsin B-complex with NS-134, 1QDQ for bovine cathepsin B-complex with CA074.</p></sec></sec><sec id="s0055"><title>Results and discussion</title><p id="p0125">The crystal structure of the inhibitor E64-R-P-NH₂ bound to sLb^pro has been determined. The correspondence of the side-chains in the inhibitor to substrate side-chains is illustrated in <xref rid="f0010" ref-type="fig">Fig. 2</xref>; a portion of the electron density in the final model of the inhibitor bound to the active site is shown in <xref rid="f0015" ref-type="fig">Fig. 3</xref>. Three chains, termed A, B and C were found in the asymmetric unit of the crystal lattice. Electron density was visible for sLb^pro residues 29&#x02013;187 of chain A, 29 to 184 of chain B and 29&#x02013;185 of chain C. For the inhibitor, electron density was visible for all atoms except for those of the P3 amino-alkyl guanidinium group (referred to as Arg-m in the text and figures) and P1&#x02032; Arg. For P3 Arg-m, density was visible up to the <italic>C</italic>_{<italic>&#x003b2;</italic>} atom for chain A, for all atoms of chain B (due to favourable interactions with an Asp residue from a symmetry related molecule) and to atom <italic>N</italic>_{<italic>&#x003b5;</italic>} for chain C. For the P1&#x02032; arginine residues, density up to the C_&#x003b2; atom for chain A was visible whereas for chains B and C density was observed to the <italic>C</italic>_{<italic>&#x003b3;</italic>} atom. The remaining atoms of these side-chains including the guanidinium group were modelled in <xref rid="f0020" ref-type="fig">Fig. 4</xref>, <xref rid="f0025" ref-type="fig">Fig. 5</xref>, <xref rid="f0030" ref-type="fig">Fig. 6</xref>, <xref rid="f0035" ref-type="fig">Fig. 7</xref> after the side-chain trace of <italic>C</italic>_{<italic>&#x003b1;</italic>} to <italic>C</italic>_{<italic>&#x003b3;</italic>} in the most likely conformation. Density for the covalent bond between the active site cysteine and the inhibitor (atom C1) was very clear in all three chains. Superimposition of the structure of sLb^pro bound to E64-R-P-NH₂ with the unbound Lb^pro structure of sLb^pro C51A C133S (PDB ID 1QMY, chainB) (<xref rid="bib19" ref-type="bibr">Guarn&#x000e9; et al., 2000</xref>) gave an r.m.s.d. of 0.35&#x000a0;&#x000c5; over 156<italic>C</italic>_{<italic>&#x003b1;</italic>} atoms superimposed. Given that the best resolution of the inhibitor was found in chain B, all structural analysis is based on this chain.<fig id="f0015"><label>Fig. 3</label><caption><p>Stereo view of the arrangement of the inhibitor E64-R-P-NH₂ and the substrate binding site of sLb^pro. 2F₀&#x02013;F_c maps contoured at 1&#x000a0;&#x003c3; are shown as grey mesh for the inhibitor and the sLb^pro residues Asp49, Cys51, Glu96 and Glu147. The inhibitor is shown as green sticks. Residues of sLb^pro interfacing with the inhibitor are shown as grey sticks. Oxygen, nitrogen and sulphur atoms are coloured red, blue and yellow, respectively. Due to the lack of electron density, no structure is shown for the P1&#x02018; Arg residue of E64-R-P-NH₂ from the C&#x003b4; atom onwards.</p></caption><graphic xlink:href="gr3"/></fig><fig id="f0020"><label>Fig. 4</label><caption><p>Comparison of the binding of E64-R-P-NH₂ and P1-P3 of the CTE. (A) The inhibitor (green sticks) is shown in the substrate binding site of sLb^pro. Side-chains of the inhibitor are labelled. In <xref rid="f0020" ref-type="fig">Fig. 4</xref>, <xref rid="f0025" ref-type="fig">Fig. 5</xref>, <xref rid="f0030" ref-type="fig">Fig. 6</xref>, <xref rid="f0035" ref-type="fig">Fig. 7</xref>, the atoms of the P1&#x02032; Arg residue from C_&#x003b4; onwards are modelled based on the most favourable conformation. Residues of the active site (Cys51, His148, Asp163) as well as the three acidic residues discussed in the text are shown as sticks. (B) As in A, with the P1-P3 residues (in yellow and labelled) of the CTE superimposed for comparison.</p></caption><graphic xlink:href="gr4"/></fig><fig id="f0025"><label>Fig. 5</label><caption><p>Electrostatic interactions involved in sLb^pro interaction with E64-R-P-NH₂ and the P1-P3 residues of the CTE. The electrostatic potential of sLb^pro was calculated using the Adaptive Poisson&#x02013;Boltzmann Solver package (<xref rid="bib2" ref-type="bibr">Baker et al., 2001</xref>) within PyMOL (<xref rid="bib10" ref-type="bibr">DeLano<bold>,</bold> 2002</xref>). The surface is coloured according to the electrostatic potential ranging from &#x02212;5 (red) to +5 (blue) kT/e. (A) The inhibitor E64-R-P-NH₂ is shown as green sticks, (B) residues P1-P3 of the CTE as yellow sticks. The representations on the right are rotated 90&#x000b0; on the <italic>x</italic>-axis relative to those on the left.</p></caption><graphic xlink:href="gr5"/></fig><fig id="f0030"><label>Fig. 6</label><caption><p>Comparison of arrangement of negatively charged residues in the substrate binding sites of sLb^pro, glycyl endopeptidase and SERA5. (A) sLb^pro bound to E64-R-P-NH₂ (green sticks). (B) Substrate binding site of SERA5. (C) Glycyl endopeptidase bound to the inhibitor benzyloxycarbonyl-leucine-valine-glycine-methylene (yellow sticks). (D) Stereo view of the superimposition of the three structures in A&#x02013;C. Residues referred to in the text are shown as sticks.</p></caption><graphic xlink:href="gr6"/></fig><fig id="f0035"><label>Fig. 7</label><caption><p>Comparison of the S1&#x02032; and S2&#x02032; binding sites of FMDV sLb^pro and cathepsin B. (A and B) E64-R-P-NH₂ (green sticks) bound in the substrate binding site of Lb^pro with the inhibitors NS-134 (A, blue sticks) and CA074 (B, magenta sticks) superimposed. (C and D) NS-134 (C) and CA074 (D) bound to the substrate binding site of cathepsin B with E64-R-P-NH₂ superimposed. The colour coding is as in A and B. The structure of Trp221 for which there is no equivalent in Lb^pro is shown as sticks as are residues referred to in the text.</p></caption><graphic xlink:href="gr7"/></fig></p><p id="p0130">To determine the binding of E64-R-P-NH₂ to sLb^pro, we first compared its arrangement in the substrate binding site of sLb^pro to that of the last three residues of the CTE observed in the crystal structure of Lb^pro (<xref rid="bib20" ref-type="bibr">Guarn<bold>&#x000e9;</bold> et al., 1998</xref>). <xref rid="f0020" ref-type="fig">Fig. 4</xref> shows that the positions of the P3 Arg-m side-chain of the inhibitor and the P3 Lys side-chain of the CTE occupy similar positions in the two structures. The <italic>C</italic>_{<italic>&#x003b3;</italic>} atom of the P1&#x02032; Arg residue of the inhibitor lies between the side-chains of Asp49 (distance from <italic>C</italic>_{<italic>&#x003b3;</italic>} to carboxy group of Asp49 is 4.2&#x000a0;&#x000c5;) and Glu147 (distance from <italic>C</italic>_{<italic>&#x003b3;</italic>} to <italic>C</italic>_{<italic>&#x003b2;</italic>} of Glu147 is 4.5&#x000a0;&#x000c5;). Given the uncertainty in the position of the guanidinium group (as mentioned earlier, the remaining atoms were modelled as no density was observed), a closer localisation is not possible. Nevertheless, the superimposition in <xref rid="f0020" ref-type="fig">Fig. 4</xref>B shows that the P1 Lys of the CTE lies almost equidistant between Asp49, Glu96 and Glu147. The disorder of the P1&#x02032; Arg in the structure of the inhibitor presented here indicates that the side-chain is flexible; in contrast, in the previously published structure of Lb^pro C51A, good density was observed to the P1 Lys residue in the substrate binding site of Lb^pro (<xref rid="bib20" ref-type="bibr">Guarn<bold>&#x000e9;</bold> et al., 1998</xref>). Given that the polypeptide chain is fully extended in both the CTE and E64-R-P-NH₂ bound structures, this explains how a peptide containing Lys and Arg at P1 and P1&#x02032; can be refractory to cleavage (<xref rid="bib33" ref-type="bibr">Nogueira Santos et al., 2012</xref>). If the Lys at P1 points away from the globular domain, an Arg side-chain at P1&#x02032; would have to point towards it. Thus, on oligopeptide substrates at least, the enzyme can only accommodate a basic residue at one of the positions, presumably because it requires a glycine with its greater freedom of rotation at the other. However, the data do not answer the question why a peptide containing Lys and Arg at P1 and P1&#x02032; can inhibit Lb^pro (<xref rid="bib33" ref-type="bibr">Nogueira Santos et al., 2012</xref>). This implies that the inhibitor may bind in a mode that has not yet been observed that moves the scissile bond out of the active site. However, additional structural information will be required to elucidate the nature of the binding of this peptide.</p><p id="p0135">Overall, comparison of the binding of the E64-R-P-NH₂ and the CTE residues (<xref rid="f0025" ref-type="fig">Fig. 5</xref>) show that the P1/P1&#x02032; binding area is a deep cleft surrounded by the acidic residues Asp49, Glu96 and Glu147. We set out to determine whether other papain-like cysteine proteinases have been identified that have a similar arrangement of three acidic residues in the vicinity of the S1/S1&#x02032; binding sites. Berti and Storer (<xref rid="bib3" ref-type="bibr">Berti and Storer, 1995</xref>) compared the sequences of 48 representative papain-like cysteine proteinases. Only one, SERA5 (Serine repeat antigen 5, termed PfalI in (<xref rid="bib3" ref-type="bibr">Berti and Storer, 1995</xref>)) from <italic>P. falciparum</italic> showed acidic amino acids at the equivalent positions to those in sLb^pro; these are Asp594, Glu638 and Asp761 which are equivalent to Asp49, Glu96 and Glu147 of sLb^pro ((<xref rid="bib21" ref-type="bibr">Hodder et al., 2009</xref>); <xref rid="f0030" ref-type="fig">Fig. 6</xref>A and B). However, little is known about the biochemistry of this protein; indeed, proteolytic activity has not been shown. Furthermore, the putative active site residue is serine, not cysteine. In addition, the authors suggested that Asp594 (equivalent to Asp49) of SERA5 is too near to the substrate binding site to allow substrate to bind.</p><p id="p0140">A second enzyme, glycyl endopeptidase (ppiv in <xref rid="bib3" ref-type="bibr">Berti and Storer, (1995)</xref>), also possesses two acidic residues, Glu23 and Asp158, equivalent to Asp49 and Glu147. The third residue (Asn64, equivalent to Glu96 in sLb^pro) is however not acidic and is followed by Arg65. As can be seen in <xref rid="f0030" ref-type="fig">Fig. 6</xref>C, the presence of Glu23 and Arg65 preclude the entry of any substrates with amino acids larger than glycine at P1, thus conferring the specificity referred to in the name glycyl endopeptidase.</p><p id="p0145">It should be noted that only these three papain-like enzymes have an amino acid other than glycine at the position equivalent to Gly23 in papain (equivalent to Asp 49 in sLb^pro). Superimposition of the three structures (<xref rid="f0030" ref-type="fig">Fig. 6</xref>D) shows that Asp49 in sLb^pro is further away from the substrate binding site than Glu23 or Asp594 in glycyl endopeptidase (<xref rid="bib34" ref-type="bibr">O&#x02019;Hara et al., 1995</xref>) and SERA5 (<xref rid="bib21" ref-type="bibr">Hodder et al., 2009</xref>). This is due to the presence of only four residues in sLb^pro lying between the oxyanion hole defining residue (Asn46) and the active site Cys51. In all other papain-like cysteine proteinases, five residues are present between the oxyanion-hole residue Gln19 and the active site nucleophile Cys25. Interestingly, Glu23 of glycyl endopeptidase is closer to the substrate binding site than Asp594 in SERA5, suggesting that the substrate binding site of SERA5 may be more open than previously thought. In contrast, Glu96 does not superimpose well with Glu638, with the <italic>C</italic>_{<italic>&#x003b1;</italic>} lying 3.7&#x000a0;&#x000c5; apart (<xref rid="f0030" ref-type="fig">Fig. 6</xref>D). Finally, as an important control for the accuracy of the structural superimposition, we note that the <italic>C</italic>_{<italic>&#x003b1;</italic>} of the catalytic histidines (H148, H762 and H159) superimpose well (<xref rid="f0030" ref-type="fig">Fig. 6</xref>D).</p><sec id="s0060"><title>Comparisons with inhibitor complexes from cathepsin B</title><p id="p0150">Information on the structural details of the S1&#x02032; and S2&#x02032; binding sites in papain-like cysteine proteinases is limited, especially for the S2&#x02032; site (<xref rid="bib44" ref-type="bibr">Turk et al., 1998</xref>, <xref rid="bib46" ref-type="bibr">Turk et al., 2012</xref>). Indeed, structures of compounds with residues bound in the S2&#x02032; position are only available for cathepsin B complexed with the inhibitors CA030, CA074 and NS-134 (<xref rid="f0010" ref-type="fig">Fig. 2</xref>; (<xref rid="bib42" ref-type="bibr">Stern et al., 2004</xref>; <xref rid="bib45" ref-type="bibr">Turk et al., 1995</xref>; <xref rid="bib48" ref-type="bibr">Yamamoto et al., 1997</xref>)). To compare the binding of the inhibitors CA074 and NS-134 to cathepsin B (CA030 differs only in the length and chemical bond of the N-terminal aliphatic moiety (<xref rid="bib45" ref-type="bibr">Turk et al., 1995</xref>)) and that of E64-R-P-NH₂ to sLb^pro, the structures of cathepsin B complexes with CA074 and NS-134 were superimposed on that of sLb^pro using the SSM tool of Coot (<xref rid="bib25" ref-type="bibr">Krissinel and Henrick, 2004</xref>). The r.m.s.d. values were 2.44&#x000a0;&#x000c5; for sLb^pro superimposed on cathepsin B complexed to CA074 (1QDQ) and 2.31&#x000a0;&#x000c5; for sLb^pro superimposed on cathepsin B complexed to NS-134 (1SP4). <xref rid="f0035" ref-type="fig">Fig. 7</xref>A and B show the positions of the inhibitors NS-134 and CA-074 relative to E64-R-P-NH₂ on sLb^pro; <xref rid="f0035" ref-type="fig">Fig. 7</xref>C and D show the relationships on the structure of cathepsin B.</p><p id="p0155">The P1&#x02032; residues of CA074 and NS-134 are Ile and Leu, respectively, in contrast to the Arg found in E64-R-P-NH₂. Despite the difference in chemical composition, however, the side-chains of these residues superimpose well and occupy the same relative space. The specificity of cathepsin B for these large hydrophobic residues is derived from the presence of a hydrophobic pocket made up of residues Phe174, Val176, Phe180, Leu181, M196 and Trp221 (<xref rid="f0035" ref-type="fig">Figs. 7</xref>C and D). In contrast, in sLb^pro, there is no equivalent loop to those in cathepsin B bearing residues Phe174 to Leu181 and Met196 or Trp221. This region is thus open and, as mentioned before, the presence of Glu147 and Asp49 enable sLb^pro to accept well the arginine residue.</p><p id="p0160">At the P2&#x02032; position, all three inhibitors have a proline residue. It is clear that the positions of the proline residues from CA074 and NS-134 on the one hand and sLb^pro on the other are different. In sLb^pro, the proline P2&#x02032; residue of E64-R-P-NH₂ lies closer to Asp163, the third member of the catalytic triad. Two factors appear to be responsible for this. The first is the absence of a residue equivalent to Trp221 in cathepsin B (Trp177 in papain) that pushes the proline residue away from Asn219 (equivalent to Asp163 in sLb^pro). Second, albeit only in chain A, the sLb^pro residue Asp164 forms a hydrogen bond (2.7&#x000a0;&#x000c5;) to the terminal nitrogen on the proline residue whereas in cathepsin B, the terminal carboxyl group of the proline is co-ordinated by His110 and His111 in the occluding loop, a structure unique to cathepsin B that is responsible for its exopeptidase activity. It is clear from the structure that there is no binding pocket for the P2&#x02032; residue in sLb^pro.</p></sec><sec id="s0065"><title>Lb^pro and sLb^pro differ in cleavage efficiencies on intramolecular substrates</title><p id="p0165">The structural analysis illustrates how sLb^pro can bind to an inhibitor bearing Leu, Gly and Arg at the P2, P1 and P1&#x02032; sites, respectively, the very residues found at the eIF4GII cleavage site. Nevertheless, a peptide corresponding to the eIF4GI peptide (SFANLG&#x0204e;RTTL) was a poor substrate for both Lb^pro and sLb^pro ((<xref rid="bib38" ref-type="bibr">Santos et al., 2009</xref>); unpublished data). However, the eIF4GI cleavage site when introduced into the background of the polyprotein substrate was efficiently cleaved by Lb^pro but was still refractory to cleavage by sLb^pro (<xref rid="bib5" ref-type="bibr">Cencic et al., 2007</xref>). Examination of the state of the endogenous eIF4GI in RRLs used for the experiments showed that it was cleaved by both Lb^pro and sLb^pro (<xref rid="bib5" ref-type="bibr">Cencic et al., 2007</xref>).</p><p id="p0170">To understand these observations and illuminate differences in cleavage efficiencies between Lb^pro and sLb^pro, we decided to investigate further the cleavage of intermolecular polyprotein substrates using the system described by <xref rid="bib5" ref-type="bibr">Cencic et al. (2007)</xref>. Here, a fragment of the FMDV polyprotein encoding an inactive form of Lb^pro, VP4 and part of VP2 (termed Lb^pro C51A VP4/VP2) is labelled with ³⁵S methionine by translation in RRLs (<xref rid="bib5" ref-type="bibr">Cencic et al., 2007</xref>). Subsequently, cold methionine is added and an mRNA encoding an active, mature Lb^pro or sLb^pro is added. The enzyme was synthesised from an RNA molecule rather than adding purified recombinant proteinase for two reasons. First, preparations of purified active Lb^pro always contain some sLb^pro that arises from self-processing, even when all purification steps are done at 4&#x000a0;&#x000b0;C (<xref rid="bib23" ref-type="bibr">Kirchweger et al., 1994</xref>). Second, the translation of the RNA followed by Lb^pro cleavage of the eIF4G isoforms in the RRLs resembles more closely the in vivo situation during an FMDV infection. The labelled substrate and products are separated by SDS-PAGE and detected by fluorography. Although it is very difficult to vary either the enzyme or substrate concentrations in this assay, it still provides qualitative information on differences in the rates of reaction between different forms of L^pro.</p><p id="p0175">A typical experiment is shown in <xref rid="f0040" ref-type="fig">Fig. 8</xref>A, with Lb^pro cleaving the wild-type sequence between 15 and 30&#x000a0;min. The Lb^pro moiety has four methionine residues compared to only two in the VP4/VP2 part, providing a partial explanation for the lower intensity of the latter band (<xref rid="bib16" ref-type="bibr">Glaser et al., 2001</xref>). In addition, we have evidence that the VP4/VP2 part is degraded in the RRLs (data not shown), with degradation being enhanced when a residue other than the wild-type glycine is present at the N-terminus of VP4 (see <xref rid="f0040" ref-type="fig">Fig. 8</xref>B&#x02013;E). We first investigated the effect of substituting residues at the P1 and P1&#x02032; positions with residues (underlined in <xref rid="f0040" ref-type="fig">Fig. 8</xref>B&#x02013;D), several of which had been shown to be detrimental to peptide cleavage (<xref rid="bib19" ref-type="bibr">Guarn&#x000e9; et al., 2000</xref>, <xref rid="bib33" ref-type="bibr">Nogueira Santos et al., 2012</xref>, <xref rid="bib38" ref-type="bibr">Santos et al., 2009</xref>). However, none of the modifications affected the cleavage efficiency of Lb^pro (<xref rid="t0010" ref-type="table">Table 2</xref>). These results show that there are clear differences between the activity of Lb^pro on peptides and polyprotein substrates, indicating that the conformation of the substrate may be different in the background of the polyprotein. In addition, as previously observed, replacement of the P5-P4&#x02032; residues of the polyprotein cleavage sequence with those of the eIF4GI site (<xref rid="f0040" ref-type="fig">Fig. 8</xref>E) also did not affect the efficiency of Lb^pro cleavage.<fig id="f0040"><label>Fig. 8</label><caption><p>Effects of P1 or P1&#x02032; site mutations on the intermolecular cleavage efficiency of Lb^pro. Intermolecular processing of the precursor Lb^pro C51A VP4/VP2 (A) and variants thereof (B&#x02013;E) by Lb^pro. The cleavage sequence present in the background of the polyprotein is shown in grey boxes. Differences from the wild-type sequence of the precursor are underlined. RRLs were programmed with RNA (14&#x000a0;ng/&#x000b5;l) coding for the polyprotein Lb^pro C51A VP4/VP2. Translation was performed for 20&#x000a0;min at 30&#x000a0;&#x000b0;C in the presence of [³⁵S]-Met and terminated by the addition of unlabelled Met for 10&#x000a0;min. RNA (14&#x000a0;ng/&#x000b5;l) coding for Lb^pro was added and translation was continued at 30&#x000a0;&#x000b0;C. The reaction was terminated by placing the samples on ice and the addition of Laemmli sample buffer containing excess unlabelled Met and Cys. 10&#x000a0;&#x000b5;l aliquots were analysed by 17.5% SDS-PAGE gels, followed by fluorography. Uncleaved precursor Lb^pro C51A VP4/VP2 and cleavage products Lb^pro C51A and VP4/VP2 are indicated. The asterisk (*) indicates an aberrant cleavage product. Negative controls devoid of any RNA (-sub, -prot) or comprising only RNA encoding the precursor (+sub, -prot) are shown on the right of each gel. Protein standards are shown on the left. Each cleavage reaction was performed twice; a representative autoradiogram for each is shown.</p></caption><graphic xlink:href="gr8"/></fig><table-wrap position="float" id="t0010"><label>Table 2</label><caption><p>Summary of the mutational analysis of the intermolecular cleavage efficiency of Lb^pro and sLb^pro. Data are taken from <xref rid="f0040" ref-type="fig">Fig. 8</xref>, <xref rid="f0045" ref-type="fig">Fig. 9</xref> and <xref rid="bib16" ref-type="bibr">Glaser et al., 2001</xref> for cleavage of eIF4GI by Lb^pro. The experiments were performed twice.</p></caption><table frame="hsides" rules="groups"><thead><tr><th/><th colspan="5"><bold>50% cleavage (min)</bold><hr/></th><th/></tr><tr><th/><th colspan="5"><bold>pCITE Lb</bold>^{<bold>pro</bold>}<bold>C51A VP4/VP2</bold><hr/></th><th><bold>endogenous eIF4GI</bold><hr/></th></tr><tr><th/><th><bold>VQRKLK^&#x0204e;GAGQ</bold></th><th><bold>VQRKL<underline>G</underline>^&#x0204e;<underline>R</underline>AGQ</bold></th><th><bold>VQRKLK^&#x0204e;<underline>R</underline>AGQ</bold></th><th><bold>VQRKL<underline>G</underline>^&#x0204e;AAGQ</bold></th><th><bold><underline>SFAN</underline>L<underline>G</underline>^&#x0204e;<underline>RTTL</underline> (eIF4GI)</bold></th><th><bold><underline>SFAN</underline>L<underline>G</underline>^&#x0204e;<underline>RTTL</underline></bold></th></tr></thead><tbody><tr><td><bold>Lb</bold>^{<bold>pro</bold>}</td><td align="center"><bold>15&#x02013;30</bold></td><td align="center"><bold>0</bold>&#x02013;<bold>15</bold></td><td align="center"><bold>15</bold>&#x02013;<bold>30</bold></td><td align="center"><bold>0</bold>&#x02013;<bold>15</bold></td><td align="center"><bold>15</bold>&#x02013;<bold>30</bold></td><td align="center"><bold>0</bold>&#x02013;<bold>15</bold></td></tr><tr><td><bold>sLb</bold>^{<bold>pro</bold>}</td><td align="center"><bold>30</bold></td><td align="center"><bold>30</bold></td><td align="center"><bold>45</bold>&#x02013;<bold>90</bold></td><td align="center"><bold>90</bold></td><td align="center"><bold>No cleavage</bold></td><td align="center"><bold>0</bold>&#x02013;<bold>15</bold></td></tr></tbody></table></table-wrap></p><p id="p0180">We next investigated the ability of sLb^pro to cleave the same five substrates. In all cases, sLb^pro cleavage was delayed compared to the Lb^pro cleavage. 50% cleavage of the wild-type substrate and a substrate bearing P1 Gly and P1&#x02032; Arg occurred at 30&#x000a0;min (<xref rid="f0045" ref-type="fig">Fig. 9</xref>A and B) compared to 15&#x02013;30&#x000a0;min and 0&#x02013;15&#x000a0;min, respectively, with Lb^pro (<xref rid="t0005" ref-type="table">Table 1</xref>). 50% cleavage of the substrate bearing two basic amino acids at the scissile bond was observed between 45&#x02013;90&#x000a0;min whereas the substrate lacking basic amino acids at the cleavage site only showed 50% cleavage after 90&#x000a0;min. Finally, as was shown previously by <xref rid="bib5" ref-type="bibr">Cencic et al. (2007)</xref>, the substrate with the eIF4GI cleavage site was not cleaved at all, even after 120&#x000a0;min of incubation. Thus, the cleavage efficiencies of sLb^pro on modified polyprotein substrates are similar to those on analogously modified oligopeptides, in contrast to those of Lb^pro. As a control, we examined the state of the endogenous eIF4GI present in RRLs by performing a Western blot of the cleavage reaction using an anti-eIF4GI antiserum (<xref rid="f0045" ref-type="fig">Fig. 9</xref>F). Endogenous eIF4GI was cleaved within 15&#x000a0;min after the start of incubation, an efficiency comparable to that previously observed with Lb^pro and sLb^pro (<xref rid="bib16" ref-type="bibr">Glaser et al., 2001</xref>). This cleavage efficiency on endogenous eIF4GI was also observed in all other reactions in <xref rid="f0045" ref-type="fig">Fig. 9</xref> (data not shown).<fig id="f0045"><label>Fig. 9</label><caption><p>Effects of P1 or P1&#x02032; site mutations on the intermolecular cleavage efficiency of sLb^pro. The intermolecular processing of the precursor Lb^pro C51A VP4/VP2 (A) and variants thereof (B&#x02013;E) by sLb^pro. The cleavage sequence present in the background of the polyprotein is shown in grey boxes. Differences from the wild-type sequence of the precursor are underlined. The translation and analysis was done as shown in <xref rid="f0040" ref-type="fig">Fig. 8</xref>. (F) The intermolecular cleavage of endogenous eIF4GI present in the RRLs from panel E. 10&#x000a0;&#x000b5;l aliquots were analysed on a 6% Dasso &#x00026; Jackson SDS-PAGE (<xref rid="bib6" ref-type="bibr">Dasso and Jackson, 1989</xref>), followed by immunoblotting with an antiserum detecting the N-terminal part of eIF4GI. Uncleaved eIF4GI and cleavage products (cp_N) are indicated. Negative controls devoid of any RNA (&#x02212;sub, &#x02212;prot) or comprising only RNA encoding the precursor (+sub, &#x02212;prot) are shown. Protein standards are shown on the left. Each cleavage reaction was performed twice; a representative autoradiogram for each is shown.</p></caption><graphic xlink:href="gr9"/></fig></p><p id="p0185">How can we explain the differences observed above between Lb^pro and sLb^pro in their behaviour towards oligopeptide and polyprotein substrates (<xref rid="t0010" ref-type="table">Table 2</xref>)? Why do the introduced mutations only affect sLb^pro cleavage and why is sLb^pro not capable of recognising the eIF4GI cleavage site in the background of the polyprotein? The difference in the cleavage efficiencies between Lb^pro and sLb^pro on polyprotein protein substrates can be explained by the ability of Lb^pro to bind to the cleavage site on the substrate with its canonical substrate binding site and through its own CTE to the &#x0201c;substrate binding site&#x0201d; of the substrate as shown in <xref rid="f0050" ref-type="fig">Fig. 10</xref>A and B. In contrast, sLb^pro can only make one of these interactions, as it lacks an intact CTE (<xref rid="f0050" ref-type="fig">Fig. 10</xref>C and D).<fig id="f0050"><label>Fig. 10</label><caption><p>Model of the intermolecular cleavage of polyprotein substrates by Lb^pro and sLb^pro. (A and C) Cleavage of wild-type Lb^pro C51A VP4/VP2 by Lb^pro and sLb^pro, respectively. (B and D) Cleavage of Lb^pro C51A VP4/VP2 containing the eIF4GI cleavage sequence by Lb^pro and sLb^pro, respectively. E. Cleavage of Lb^pro C51A eIF4GI_599-668 VP4/VP2 SFANLG*RTTL by sLb^pro. Lb^pro is in light blue, sLb^pro in dark blue, VP4 in light green, VP2 in dark green and the eIF4GI_599-668 fragment in red.</p></caption><graphic xlink:href="gr10"/></fig></p><p id="p0190">sLb^pro can efficiently cleave the eIF4GI site on the native protein present in the RRL because this involves residues Cys133 and Asp184-Leu188 of the CTE but not the last six residues of the CTE (<xref rid="bib14" ref-type="bibr">Foeger et al., 2005</xref>). Accordingly, we introduced 80 amino acids from eIF4GI containing the L^pro binding and cleavage sites into the Lb^pro C51A VP4/VP2 substrate (<xref rid="f0050" ref-type="fig">Fig. 10</xref>E). This modified substrate (termed Lb^pro C51A eIF4GI_599-668 VP4/VP2 SFANLG&#x0204e;RTTL) was cleaved by sLb^pro between 30 and 90&#x000a0;min, indicating that the availability of the two binding sites had allowed cleavage to take place (<xref rid="f0055" ref-type="fig">Fig. 11</xref>A).<fig id="f0055"><label>Fig. 11</label><caption><p>eIF4GI_599-668 is essential for the cleavage of the sub-optimal eIF4GI cleavage site SFANLG*RTTL by sLb^pro. A-C, left panels. Intermolecular processing of Lb^pro C51A eIF4GI_599-668 VP4/VP2 by sLb^pro and sLb^pro C133S Q185R E186K and Lb^pro C51A VP4/VP2 by sLb^pro C133S Q185R E186K. The cleavage sequence present in the background of the polyprotein is shown in grey boxes. Translation reactions and analyses of products were as in <xref rid="f0040" ref-type="fig">Fig. 8</xref>. Uncleaved precursors Lb^pro C51A eIF4GI_599-668 VP4/VP2 and Lb^pro C51A VP4/VP2 as well as cleavage products Lb^pro C51A eIF4GI_599-668, Lb^pro C51A and VP4/VP2 are indicated. A&#x02013;C, right panels. The intermolecular cleavage of endogenous eIF4GI present in the RRL. Analysis of the state of eIF4GI was as shown in <xref rid="f0045" ref-type="fig">Fig. 9</xref>. Uncleaved eIF4GI and cleavage products (cp_N) are indicated. Negative controls lacking added RNA (&#x02212;sub, &#x02212;prot) or comprising only of RNA encoding the precursor (+sub, &#x02212;prot) are shown. Protein standards are shown on the left.</p></caption><graphic xlink:href="gr11"/></fig></p><p id="p0195">As a control, we examined the ability of the variant sLb^pro C133S Q185R E186K that had previously been shown to be unable to cleave endogenous eIF4GI because the variant cannot recognise its binding site on this factor (<xref rid="bib14" ref-type="bibr">Foeger et al., 2005</xref>). This variant could also not cleave the Lb^pro C51A eIF4GI_599-668 VP4/VP2, but maintains the ability to cleave the wild-type substrate Lb^pro C51A VP4/VP2 (<xref rid="f0055" ref-type="fig">Figs. 11</xref>B and C, left panels). As previously reported, sLb^pro C133S Q185R E186K was however not able to cleave endogenous eIF4GI, even after 120&#x000a0;min of incubation (<xref rid="f0055" ref-type="fig">Figs. 11</xref>B and C, right panels).</p></sec></sec><sec id="s0070"><title>Concluding remarks</title><p id="p0200">The structural data presented here reveal that sLb^pro uses three acidic residues to bind to basic residues at the P1 or P1&#x02032; positions of a substrate and that this represents a unique arrangement that is not found in cellular papain-like proteinases. Differences in the cleavage efficiency of Lb^pro and sLb^pro were observed on modified polyprotein substrates. The presence of sLb^pro in infected cells (<xref rid="bib36" ref-type="bibr">Sangar et al., 1988</xref>) suggests that differences in the properties of Lb^pro and sLb^pro will be relevant to the success of viral replication. Hence, the removal of six C-terminal residues 40&#x000a0;&#x000c5; from the active site may represent a unique mechanism to modify the properties of a proteolytic enzyme during viral replication.</p></sec>

Transmission of allergen-specific IgG and IgE from maternal blood into breast milk visualized with microarray technology

Could not extract abstract

<contrib contrib-type="author" id="au1"><name><surname>Hochwallner</surname><given-names>Heidrun</given-names></name><degrees>PhD</degrees><xref rid="aff1" ref-type="aff">a</xref></contrib><contrib contrib-type="author" id="au2"><name><surname>Alm</surname><given-names>Johan</given-names></name><degrees>MD, PhD</degrees><xref rid="aff2" ref-type="aff">b</xref><xref rid="aff3" ref-type="aff">c</xref></contrib><contrib contrib-type="author" id="au3"><name><surname>Lupinek</surname><given-names>Christian</given-names></name><degrees>MD</degrees><xref rid="aff1" ref-type="aff">a</xref></contrib><contrib contrib-type="author" id="au4"><name><surname>Johansson</surname><given-names>Catharina</given-names></name><degrees>PhD</degrees><xref rid="aff4" ref-type="aff">d</xref></contrib><contrib contrib-type="author" id="au5"><name><surname>Mie</surname><given-names>Axel</given-names></name><degrees>PhD</degrees><xref rid="aff2" ref-type="aff">b</xref></contrib><contrib contrib-type="author" id="au6"><name><surname>Scheynius</surname><given-names>Annika</given-names></name><degrees>MD, PhD</degrees><xref rid="aff4" ref-type="aff">d</xref></contrib><contrib contrib-type="author" id="au7"><name><surname>Valenta</surname><given-names>Rudolf</given-names></name><degrees>MD</degrees><email>rudolf.valenta@meduniwien.ac.at</email><xref rid="aff1" ref-type="aff">a</xref></contrib><aff id="aff1"><label>a</label>Division of Immunopathology, Department of Pathophysiology and Allergy Research, Medical University of Vienna, Austria</aff><aff id="aff2"><label>b</label>Department of Clinical Science and Education, S&#x000f6;dersjukhuset, Karolinska Institutet, Stockholm, Sweden</aff><aff id="aff3"><label>c</label>Sachs' Children and Youth Hospital, S&#x000f6;dersjukhuset, Stockholm, Sweden</aff><aff id="aff4"><label>d</label>Department of Medicine Solna, Translational Immunology Unit, Karolinska Institutet and University Hospital, Stockholm, Sweden</aff>

The Journal of Allergy and Clinical Immunology

<p content-type="salutation">To the Editor:</p><p id="p0010">Data from experimental animal models have previously shown that allergen-specific IgG antibodies are transmitted from the mother to the offspring via breast milk<xref rid="bib1" ref-type="bibr">1</xref>, <xref rid="bib2" ref-type="bibr">2</xref> and have provided evidence that the transmitted allergen-specific IgG antibodies protect specifically against allergic sensitization.<xref rid="bib2" ref-type="bibr">²</xref> In some studies, it has been demonstrated for humans that breast-feeding has a prophylactic effect against atopic disease but there are also reports arguing against this and the underlying mechanisms are not known.<xref rid="bib3" ref-type="bibr">³</xref> There is evidence that IgG antibodies against bacterial antigens (ie, pneumococcal antigens) are transferred from the blood of mothers into their breast milk.<xref rid="bib4" ref-type="bibr">⁴</xref> Furthermore, it was possible to detect IgA and IgG against different food antigens in human serum, saliva, colostrums, and milk samples.<xref rid="bib5" ref-type="bibr">⁵</xref> Another study found that total IgE levels in breast milk and blood were associated but allergen-specific IgE was not analyzed.<xref rid="bib6" ref-type="bibr">⁶</xref></p><p id="p0015">With the FP7-funded European Union research program Mechanisms of the Development of ALLergy (MeDALL; <ext-link ext-link-type="uri" xlink:href="http://medall-fp7.eu/" id="intref0010">http://medall-fp7.eu/</ext-link>), we have recently developed a microarray containing a large number of purified natural and recombinant respiratory, food, and insect allergens that allows highly sensitive measurement of allergen-specific IgE and IgG levels with minute amounts of blood.<xref rid="bib7" ref-type="bibr">⁷</xref> A&#x000a0;major advantage of the microarray technology is that it allows one to measure antibody reactivities toward a large panel of different allergens. Here, we investigated whether the MeDALL chip is suitable for (1) the measurement of allergen-specific IgG and IgE levels in human breast milk samples, (2) whether there is a transmission of allergen-specific antibodies from blood into breast milk, and (3) whether the reactivity profile of allergens recognized by antibodies in blood and milk is similar. For this purpose, we analyzed plasma and breast milk samples from sensitized (n&#x000a0;= 23) and nonallergic mothers (n&#x000a0;= 6) from the ALADDIN birth cohort.<xref rid="bib8" ref-type="bibr">⁸</xref> None of the mothers was on allergen-specific immunotherapy. Maternal blood samples were collected in the period around delivery (&#x02212;1 to +2 months), and the breast milk samples were obtained 2 months after delivery. The study was approved by the local Research Ethical Committee, and written informed consent was obtained from all families.</p><p id="p0020">The breast milk samples were centrifuged for 10 minutes at 2500<italic>g</italic> before use to remove the lipids. For comparison of IgG titers in plasma and breast milk, the plasma samples were diluted 1:50, 1:100, 1:200, and 1:400 before analysis. Microarrays were incubated with 30 &#x003bc;L of the plasma dilutions or undiluted breast milk samples and allergen-specific IgG and IgE antibodies were detected with fluorophore-conjugated anti-IgG and anti-IgE antibodies, respectively.<xref rid="bib7" ref-type="bibr">⁷</xref> The fluorescence intensities were measured with a biochip scanner. Results were expressed in ISAC standardized units (Thermofisher, Uppsala, Sweden). Correlation coefficients were calculated with SPSS.</p><p id="p0025">Detailed analysis of allergen-specific IgG and IgE levels in plasma and breast milk samples indicated that allergen-specific IgG antibodies are transmitted from the blood into breast milk in a highly specific manner and that breast milk IgG mirrored the profile of IgG reactivity in the blood (see <xref rid="appsec1" ref-type="sec">Fig E1</xref> in this article's <xref rid="appsec1" ref-type="sec">Online Repository</xref> at <ext-link ext-link-type="uri" xlink:href="http://www.jacionline.org" id="intref0015">www.jacionline.org</ext-link>). A&#x000a0;comparison of allergen-specific IgG levels measured in 4 plasma dilutions with that of undiluted breast milk samples (<xref rid="fig1" ref-type="fig">Fig 1</xref>) indicated that allergen-specific IgG levels in breast milk were approximately 200- to 400-fold lower than in plasma. Allergen-specific IgG reactivities in plasma and breast milk were significantly correlated; for the 1:200 dilution, Spearman correlation coefficient was 0.608 (<italic>P</italic>&#x000a0;&#x0003c; .001) and for the 1:400 dilution, Spearman correlation coefficient was 0.604 (<italic>P</italic>&#x000a0;&#x0003c; .001) (<xref rid="fig2" ref-type="fig">Fig 2</xref>). Detailed results are displayed for each allergen in the heat map (<xref rid="appsec1" ref-type="sec">Fig E1</xref>). For the vast majority of allergens, plasma- and milk-derived IgG antibody reactivities were correlated. However, in certain instances (eg, milk allergens recognized by donor 3), specific IgGs were high in plasma but did not appear in milk; in some other cases, allergen-specific IgG was detected only in milk but not in plasma (<xref rid="appsec1" ref-type="sec">Fig E1</xref>). Possible explanations for lack of allergen-specific IgG binding in milk are that certain antigens are present in milk and inhibit IgG binding&#x000a0;and/or low affinity/avidity of IgG may prevent binding despite high titers in blood. In fact, the presence of certain respiratory and food allergens in breast milk has been recently demonstrated.<xref rid="bib9" ref-type="bibr">9</xref>, <xref rid="bib10" ref-type="bibr">10</xref> However, milk-specific IgG may appear because of local IgG production without corresponding IgG in blood.<fig id="fig1"><label>Fig 1</label><caption><p>Comparison of allergen-specific IgG levels (ISAC standardized unit [ISU]) measured in different plasma dilutions of 4 mothers with allergen-specific IgG levels in their breast milk samples.</p></caption><graphic xlink:href="gr1"/></fig><fig id="fig2"><label>Fig 2</label><caption><p>Correlation of allergen-specific IgG levels (ISAC standardized unit [ISU]) in plasma samples (<italic>left</italic>: dilution 1:200; <italic>right</italic>: dilution 1:400) from 4 mothers (<italic>the donors are labeled in different colors</italic>) with allergen-specific IgG levels in their corresponding undiluted breast milk samples.</p></caption><graphic xlink:href="gr2"/></fig></p><p id="p0030">The presence of allergen-specific IgG in breast milk may be due to specific transmission of allergen-specific IgG from the blood of mothers into their breast milk or eventually because of local production. For transmission, active transport by the IgG receptor FcRn, which in fact is expressed in the human mammary gland,<xref rid="bib11" ref-type="bibr">¹¹</xref> and/or transudation from blood into the mammary glands, as was found for allergen-specific IgG in mucosal fluids, may be considered.<xref rid="bib12" ref-type="bibr">¹²</xref> Interestingly, for mothers with high levels of allergen-specific IgE in their plasma (donors 1, 9, 13, 15, 16, and 19), we also found that specific IgE antibodies were present in breast milk. In all but 1 case (donor 3, Phl p 5&#x02013;specific IgE), the presence of allergen-specific IgE was associated with levels of IgE in the blood. The allergen-specific IgE reactivity profiles in blood and breast milk are summarized in detail in <xref rid="appsec1" ref-type="sec">Figs E1 and E2</xref> in this article's <xref rid="appsec1" ref-type="sec">Online Repository</xref> at <ext-link ext-link-type="uri" xlink:href="http://www.jacionline.org" id="intref0020">www.jacionline.org</ext-link>.</p><p id="p0035">In none of the 6 nonallergic mothers (donors 24-29), who served as negative controls, allergen-specific IgE was detected in plasma or in breast milk, demonstrating the specificity of IgE test results (<xref rid="appsec1" ref-type="sec">Fig E2</xref>).</p><p id="p0040">Our results thus demonstrate that allergen-specific IgG and IgE antibodies with similar specificity are present in blood and in breast milk. Furthermore, our study shows that the MeDALL allergen chip is suitable for the measurement of allergen-specific IgG and IgE antibodies not only in blood but also in breast milk, which opens the possibility to study the transmission of allergen-specific IgG antibodies from mothers to offsprings on the development of allergic sensitization in birth cohorts.</p>

Gain-of-function mutations in signal transducer and activator of transcription 1 (<italic>STAT1</italic>): Chronic mucocutaneous candidiasis accompanied by enamel defects and delayed dental shedding

Could not extract abstract

<contrib contrib-type="author" id="au1"><name><surname>Frans</surname><given-names>Glynis</given-names></name><degrees>MPharm</degrees><xref rid="aff1" ref-type="aff">a</xref><xref rid="fn1" ref-type="fn">&#x02217;</xref></contrib><contrib contrib-type="author" id="au2"><name><surname>Moens</surname><given-names>Leen</given-names></name><degrees>PhD</degrees><xref rid="aff1" ref-type="aff">a</xref><xref rid="fn1" ref-type="fn">&#x02217;</xref></contrib><contrib contrib-type="author" id="au3"><name><surname>Schaballie</surname><given-names>Heidi</given-names></name><degrees>MD</degrees><xref rid="aff1" ref-type="aff">a</xref><xref rid="aff7" ref-type="aff">g</xref></contrib><contrib contrib-type="author" id="au4"><name><surname>Van Eyck</surname><given-names>Lien</given-names></name><degrees>MD, MSc</degrees><xref rid="aff2" ref-type="aff">b</xref><xref rid="aff7" ref-type="aff">g</xref></contrib><contrib contrib-type="author" id="au5"><name><surname>Borgers</surname><given-names>Heleen</given-names></name><degrees>PhD</degrees><xref rid="aff1" ref-type="aff">a</xref></contrib><contrib contrib-type="author" id="au6"><name><surname>Wuyts</surname><given-names>Margareta</given-names></name><degrees>BSc</degrees><xref rid="aff1" ref-type="aff">a</xref></contrib><contrib contrib-type="author" id="au7"><name><surname>Dillaerts</surname><given-names>Doreen</given-names></name><degrees>MSc</degrees><xref rid="aff1" ref-type="aff">a</xref></contrib><contrib contrib-type="author" id="au8"><name><surname>Vermeulen</surname><given-names>Edith</given-names></name><degrees>MPharm</degrees><xref rid="aff3" ref-type="aff">c</xref></contrib><contrib contrib-type="author" id="au9"><name><surname>Dooley</surname><given-names>James</given-names></name><degrees>PhD</degrees><xref rid="aff2" ref-type="aff">b</xref></contrib><contrib contrib-type="author" id="au10"><name><surname>Grimbacher</surname><given-names>Bodo</given-names></name><degrees>MD, PhD</degrees><xref rid="aff4" ref-type="aff">d</xref></contrib><contrib contrib-type="author" id="au11"><name><surname>Cant</surname><given-names>Andrew</given-names></name><degrees>MD, PhD</degrees><xref rid="aff5" ref-type="aff">e</xref></contrib><contrib contrib-type="author" id="au12"><name><surname>Declerck</surname><given-names>Dominique</given-names></name><degrees>PhD</degrees><xref rid="aff6" ref-type="aff">f</xref></contrib><contrib contrib-type="author" id="au13"><name><surname>Peumans</surname><given-names>Marleen</given-names></name><degrees>PhD</degrees><xref rid="aff6" ref-type="aff">f</xref></contrib><contrib contrib-type="author" id="au14"><name><surname>Renard</surname><given-names>Marleen</given-names></name><degrees>MD</degrees><xref rid="aff7" ref-type="aff">g</xref></contrib><contrib contrib-type="author" id="au15"><name><surname>De Boeck</surname><given-names>Kris</given-names></name><degrees>MD, PhD</degrees><xref rid="aff7" ref-type="aff">g</xref></contrib><contrib contrib-type="author" id="au16"><name><surname>Hoffman</surname><given-names>Ilse</given-names></name><degrees>MD, PhD</degrees><xref rid="aff7" ref-type="aff">g</xref></contrib><contrib contrib-type="author" id="au17"><name><surname>Fran&#x000e7;ois</surname><given-names>Inge</given-names></name><degrees>MD, PhD</degrees><xref rid="aff7" ref-type="aff">g</xref></contrib><contrib contrib-type="author" id="au18"><name><surname>Liston</surname><given-names>Adrian</given-names></name><degrees>PhD</degrees><xref rid="aff2" ref-type="aff">b</xref></contrib><contrib contrib-type="author" id="au19"><name><surname>Claessens</surname><given-names>Frank</given-names></name><degrees>MD, PhD</degrees><xref rid="aff8" ref-type="aff">h</xref></contrib><contrib contrib-type="author" id="au20"><name><surname>Bossuyt</surname><given-names>Xavier</given-names></name><degrees>MD, PhD</degrees><xref rid="aff1" ref-type="aff">a</xref></contrib><contrib contrib-type="author" id="au21"><name><surname>Meyts</surname><given-names>Isabelle</given-names></name><degrees>MD, PhD</degrees><email>Isabelle.Meyts@uzleuven.be</email><xref rid="aff1" ref-type="aff">a</xref><xref rid="aff7" ref-type="aff">g</xref></contrib><aff id="aff1"><label>a</label>Department of Microbiology and Immunology, Experimental Laboratory Immunology, Katholieke Universiteit Leuven, Leuven, Belgium</aff><aff id="aff2"><label>b</label>Laboratory Genetics of Autoimmunity, Vlaams Instituut Biotechnologie, Leuven, Belgium</aff><aff id="aff3"><label>c</label>Department of Microbiology and Immunology, Laboratory for Clinical Bacteriology and Mycology, Katholieke Universiteit Leuven, Leuven, Belgium</aff><aff id="aff4"><label>d</label>Centre for Chronic Immunodeficiency, University Hospital Freiburg, Freiburg, Germany</aff><aff id="aff5"><label>e</label>Primary Immunodeficiency Group, Institute of Cellular Medicine, Newcastle University and Pediatric Immunology Service, Great North Children's Hospital, Newcastle upon Tyne, United Kingdom</aff><aff id="aff6"><label>f</label>Department of Conservative Dentistry, School for Dentistry, Katholieke Universiteit Leuven, University Hospitals Leuven, Leuven, Belgium</aff><aff id="aff7"><label>g</label>Department of Pediatrics, University Hospitals Leuven, Leuven, Belgium</aff><aff id="aff8"><label>h</label>Department of Cellular and Molecular Medicine, Laboratory of Molecular Endocrinology, Katholieke Universiteit Leuven, Leuven, Belgium</aff>

The Journal of Allergy and Clinical Immunology

<p content-type="salutation">To the Editor:</p><p id="p0010">Heterozygous gain-of-function mutations in signal transducer and activator of transcription 1 <italic>(STAT1)</italic> have recently been identified as a cause of chronic mucocutaneous candidiasis (CMC). Uzel et&#x000a0;al<xref rid="bib1" ref-type="bibr">¹</xref> described &#x0201c;<italic>STAT1</italic> gain-of-function mutations in patients with <italic>FOXP3</italic> wild-type immune dysregulation-polyendocrinopathy-enteropathy-X-linked syndrome.&#x0201d; They briefly mentioned the presence of poor enamel in 1 patient and structural and functional gastrointestinal defects in another patient. Here we present a patient with CMC associated with dental anomalies, diaphragmatic hernia, and esophageal dysmotility in whom the phenotype led to a broad differential diagnosis ranging from severe combined immunodeficiency inspired by the neonatal onset of infections to humoral immune&#x000a0;deficiency, immune dysregulation&#x02013;polyendocrinopathy&#x02013;enteropathy&#x02013;X-linked syndrome, nuclear factor &#x003ba;B essential modulator <italic>(NEMO)</italic> deficiency, Shwachman-Diamond syndrome, autosomal dominant hyper-IgE syndrome, and, finally, CMC caused by gain-of-function mutation in <italic>STAT1</italic>. We stress the early onset of respiratory tract infections, as well as the dental and gastrointestinal defects, in this patient.</p><p id="p0015">The patient was born at term small for gestational age (&#x02212;2 SD) as the third son of unrelated parents. He presented with recurrent lower respiratory tract infections from birth, intractable diarrhea, failure to thrive, seborrheic dermatitis, and CMC. Small-bowel biopsy showed villous atrophy interpreted as celiac-like disease, yet diarrhea was unresponsive to a gluten-free diet. Primary dentition showed enamel defects. Recurrent lower respiratory tract infections with <italic>Haemophilus influenzae and Streptococcus pneumoniae</italic> led to the development of bronchiectasis. Partial IgG₂ deficiency (0.34 g/L; normal range, 0.72-3.4 g/L) was found at 3 years of age. Anti-tetanus antibody levels were protective (1.2&#x000a0;mg/L; protective level, &#x0003e;1 mg/L), suggesting an intact anti-protein antibody response. Intravenous immunoglobulin (IVIG) substitution was initiated. Diaphragmatic hernia, gastroesophageal reflux disease, and disturbed esophageal motility were demonstrated and led to a Nissen fundoplication.</p><p id="p0020">At 13 years of age, the patient presented at the immunology clinic with severe growth retardation (<xref rid="fig1" ref-type="fig">Fig 1</xref>, <italic>A</italic>), pubertal delay, bronchiectasis, atonic esophagus with multiple diverticula (<xref rid="fig1" ref-type="fig">Fig&#x000a0;1</xref>, <italic>B</italic> and <italic>C</italic>), atrophic duodenal mucosa (<xref rid="fig1" ref-type="fig">Fig 1</xref>, <italic>D</italic>) corresponding to villous blunting or atrophy, joint hyperlaxity, osteopenia with recurrent fractures, delayed dental development with retention of primary teeth necessitating dental extractions (<xref rid="fig1" ref-type="fig">Fig&#x000a0;1</xref>, <italic>E</italic> and <italic>F</italic>), severe erosive tooth wear suggestive of enamel hypoplasia (<xref rid="fig1" ref-type="fig">Fig 1</xref>, <italic>E</italic> and <italic>F</italic>), seborrheic dermatitis, aphthous stomatitis, and CMC (see <xref rid="tblE1" ref-type="table">Table E1</xref> in this article's Online Repository at <ext-link ext-link-type="uri" xlink:href="http://www.jacionline.org" id="intref0010">www.jacionline.org</ext-link>). The autoimmune regulator <italic>(AIRE)</italic> and caspase recruitment domain family, member 9 <italic>(CARD9)</italic>, genes were sequenced, but no mutations were found. After withdrawal of IVIG, antibody response to unconjugated pneumococcal vaccine was tested and showed protective titers for 9 of 14 serotypes tested (see <xref rid="tblE2" ref-type="table">Table E2</xref> in this article's Online Repository at <ext-link ext-link-type="uri" xlink:href="http://www.jacionline.org" id="intref0015">www.jacionline.org</ext-link>).<xref rid="bib2" ref-type="bibr">²</xref> Increased IgG levels (21 g/L; normal range, 5.76-12.65 g/L) and persistent partial IgG₂ deficiency were noted (0.80 g/L; normal range, 1.06-6.10 g/L). An extended autoantibody screening panel was performed, including thyroid-related, adrenal gland&#x02013;related, and anti&#x02013;IFN-&#x003b1; and anti&#x02013;IFN-&#x003c9; antibodies. Only anti&#x02013;salivary gland antibodies were demonstrated (for the entire panel, see the&#x000a0;<xref rid="appsec1" ref-type="sec">Methods</xref> section in this article's Online Repository at <ext-link ext-link-type="uri" xlink:href="http://www.jacionline.org" id="intref0020">www.jacionline.org</ext-link>). Immunophenotyping showed a low percentage of switched memory B cells (1.3%; normal range, 5% to 10%; see <xref rid="tblE3" ref-type="table">Table E3</xref> in this article's Online Repository at <ext-link ext-link-type="uri" xlink:href="http://www.jacionline.org" id="intref0025">www.jacionline.org</ext-link>). Because withdrawal of IVIG was associated with an increased incidence of pneumonia, treatment was optimized by restarting IVIG and initiating azithromycin (both for antibacterial prophylaxis and its anti-inflammatory actions) and fluconazole prophylaxis, as well as overnight tube feeding. Growth hormone therapy and puberty induction led to a correction of the growth deficit (<xref rid="fig1" ref-type="fig">Fig 1</xref>, <italic>A</italic>). Finally, at the age of 18 years, hypothyroidism was diagnosed almost simultaneously with the initial reports on <italic>STAT1</italic> coiled-coil domain gain-of-function mutations.<xref rid="bib3" ref-type="bibr">3</xref>, <xref rid="bib4" ref-type="bibr">4</xref> A&#x000a0;mutation in the DNA-binding domain of <italic>STAT1</italic> was detected (c.1154C&#x0003e;T, p.T385M; <xref rid="fig2" ref-type="fig">Fig 2</xref>, <italic>A</italic>; for information on the analysis, see the <xref rid="appsec2" ref-type="sec">Results</xref> section in this article's Online Repository at <ext-link ext-link-type="uri" xlink:href="http://www.jacionline.org" id="intref0030">www.jacionline.org</ext-link>). The mutation was not found in the parents or the 2 male siblings of the index patient. As described previously, T385M is a gain-of-function mutation.<xref rid="bib1" ref-type="bibr">1</xref>, <xref rid="bib5" ref-type="bibr">5</xref>, <xref rid="bib6" ref-type="bibr">6</xref>, <xref rid="bib7" ref-type="bibr">7</xref> Likewise, we showed increased STAT1 phosphorylation in response to IFN-&#x003b1; and IFN-&#x003b3; in the patient compared with control values (<xref rid="fig2" ref-type="fig">Fig 2</xref>, <italic>B</italic>). Also, the electrophoretic mobility shift assay (EMSA) showed increased gamma-activated sequence (GAS) binding activity on stimulation with IFN-&#x003b3; (<xref rid="fig2" ref-type="fig">Fig 2</xref>, <italic>C</italic>).<fig id="fig1"><label>Fig 1</label><caption><p>Clinical characteristics. <bold>A,</bold> Growth charts showing severe growth retardation that only picks up after growth hormone therapy and puberty induction with tube feeding overnight. <italic>Arrows</italic> indicate onset of tube feeding and growth hormone therapy, respectively. <bold>B,</bold> Computed tomography showing an atonic esophagus <italic>(arrow)</italic> with air containing paraesophageal diverticula. <bold>C,</bold> Computed tomography of the chest showing multiple saccular bronchiectases and bronchial wall thickening. <bold>D,</bold> Atrophy in the duodenum, as&#x000a0;seen on endoscopy. <bold>E</bold> and <bold>F,</bold> Severe erosive tooth wear, caries, and retained primary teeth.</p></caption><graphic xlink:href="gr1"/></fig><fig id="fig2"><label>Fig 2</label><caption><p>The mutant T385M <italic>STAT1</italic> allele is a gain-of-phosphorylation and gain-of-function mutation. <bold>A,</bold>&#x000a0;Direct sequence analysis of exon 14 of <italic>STAT1</italic> (forward sequence) in a control subject and the patient with a c.1153C&#x0003e;T resulting in p.T385M. <bold>B,</bold> Intracellular staining of phosphorylated tyrosine 701 STAT1 <italic>(STAT1p)</italic> in lymphocytes after stimulation with IFN-&#x003b3; (2000 IU/mL, <italic>left panel</italic>) or IFN-&#x003b1; (10⁵ IU/mL, <italic>right panel</italic>) for 15 minutes. STAT1 and STATp are shown in a control subject (red) and in the T385M patient <italic>(blue)</italic>. Unstimulated conditions are represented as <italic>dashed lines</italic>. Results shown are representative of 2 independent experiments. <italic>MFI</italic>, Mean fluorescence intensity. <bold>C,</bold> Evaluation of STAT1, STAT1 phosphorylation, and STAT1p GAS DNA-binding capacity. Fibroblasts derived from wild-type (WT)/WT control subjects (C1 and C2), p.T385M/WT (patient P1), and p.K388E/WT (patient P2) were stimulated with 100 U/mL IFN-&#x003b1; <italic>(&#x003b1;)</italic> or 100 U/mL IFN-&#x003b3; <italic>(&#x003a5;)</italic> or left unstimulated <italic>(&#x02212;)</italic> for 60 minutes. <italic>a</italic>, Western blotting was carried out for detection of STAT1 and STAT1p levels in nuclear extracts (5 &#x003bc;g per sample). Heterogeneous nuclear ribonucleoprotein I (hnRNP I) was used as a loading control reference. <italic>b</italic>, STAT1 GAS DNA-binding capacity was evaluated by using EMSA. One microgram of nuclear extract was preincubated with 20,000 cpm of GAS probe at room temperature before nondenaturing PAGE separating free from STAT-bound probe.</p></caption><graphic xlink:href="gr2"/></fig></p><p id="p0025">The mechanism that leads to gain of function in the T385M mutation is not entirely clear. Takezaki et&#x000a0;al<xref rid="bib5" ref-type="bibr">⁵</xref> suggested an impaired dephosphorylation of STAT1. However, it is also possible that there is impaired dissociation from the DNA or a problem with the reciprocal association of the DNA-binding domain with the coiled-coil domain. After diagnosis, extended immunophenotyping of PBMCs was performed and showed absence of T_H17 cells, as described by Liu et&#x000a0;al.<xref rid="bib4" ref-type="bibr">⁴</xref></p><p id="p0030">The dental anomalies in the patient were impressive. Both primary and permanent teeth showed rapid loss of tooth substance, with severe caries and erosive tooth wear reminiscent&#x000a0;of the dental anomalies encountered in patients with Shwachman-Diamond syndrome or Ora1/Stromal interaction molecule 1 deficiency. Moreover, deciduous teeth had to be extracted because of delayed shedding. The latter feature resembles autosomal dominant hyper-IgE syndrome.</p><p id="p0035">Several hypotheses to explain the dental anomalies were put forward. First, antibiotic and antimycotic therapy and acidic hypercaloric nutrition were blamed. Second, malabsorption of calcium and vitamin D was investigated. Third, in the context of recurrent aphthous stomatitis, a sicca syndrome was suspected. Although salivary flow was low, treatment with oral saliva analogues did not improve the dental condition. There were no biochemical or clinical signs of hypothyroidism until age 18 years in our patient, excluding this as a cause for the delayed shedding of deciduous teeth. Although a role for <italic>STAT1</italic> signaling has been demonstrated in amelogenesis and dentinogenesis in rats, further research is needed to investigate the potential causal relationship between <italic>STAT1</italic> gain-of-function mutation and abnormal dental development.<xref rid="bib8" ref-type="bibr">8</xref>, <xref rid="bib9" ref-type="bibr">9</xref></p><p id="p0040">Aside from the persistent villous blunting, diaphragmatic hernia and esophageal dysmotility are remarkable gastrointestinal features. On computed tomographic (CT) scanning, as well as endoscopy, the esophagus appeared wide open and atonic, with multiple diverticula present. Thus defects in the development of the upper gastrointestinal tract seem to be a noteworthy feature of this syndrome, as hypothesized by Uzel et&#x000a0;al.<xref rid="bib1" ref-type="bibr">¹</xref> Whether the gastrointestinal manifestations are all secondary to CMC or a primary manifestation of disturbed <italic>STAT1</italic> signaling is yet to be determined.<xref rid="bib1" ref-type="bibr">1</xref>, <xref rid="bib6" ref-type="bibr">6</xref></p><p id="p0045">In conclusion, we report extensive dental anomalies, as well as diaphragmatic hernia and esophageal dysmotility, in a patient&#x000a0;with early onset of lower respiratory tract infections in the context of a gain-of-function mutation in the DNA-binding domain&#x000a0;of&#x000a0;<italic>STAT1</italic>. These features add to the complexity of the phenotype observed in patients with a gain-of-function mutation in <italic>STAT1</italic>.</p>

Regulation of the calcium-sensing receptor expression by 1,25-dihydroxyvitamin D₃, interleukin-6, and tumor necrosis factor alpha in colon cancer cells^{<xref ref-type="fn" rid="d35e73">&#x02606;</xref>}

Could not extract abstract

<contrib contrib-type="author" id="aut0005"><name><surname>Fetahu</surname><given-names>Irfete S.</given-names></name><email>irfete.fetahu@meduniwien.ac.at</email><xref rid="aff0005" ref-type="aff">a</xref></contrib><contrib contrib-type="author" id="aut0010"><name><surname>Hummel</surname><given-names>Doris M.</given-names></name><email>doris.hummel@meduniwien.ac.at</email><xref rid="aff0005" ref-type="aff">a</xref></contrib><contrib contrib-type="author" id="aut0015"><name><surname>Manhardt</surname><given-names>Teresa</given-names></name><email>teresa.manhardt@meduniwien.ac.at</email><xref rid="aff0005" ref-type="aff">a</xref></contrib><contrib contrib-type="author" id="aut0020"><name><surname>Aggarwal</surname><given-names>Abhishek</given-names></name><email>abhishek.aggarwal@meduniwien.ac.at</email><xref rid="aff0005" ref-type="aff">a</xref></contrib><contrib contrib-type="author" id="aut0025"><name><surname>Baumgartner-Parzer</surname><given-names>Sabina</given-names></name><email>sabina.baumgartner-parzer@meduniwien.ac.at</email><xref rid="aff0010" ref-type="aff">b</xref></contrib><contrib contrib-type="author" id="aut0030"><name><surname>K&#x000e1;llay</surname><given-names>Enik&#x00151;</given-names></name><email>enikoe.kallay@meduniwien.ac.at</email><xref rid="aff0005" ref-type="aff">a</xref><xref rid="cor0005" ref-type="corresp">&#x0204e;</xref></contrib><aff id="aff0005"><label>a</label>Department of Pathophysiology and Allergy Research, Medical University of Vienna, W&#x000e4;hringer G&#x000fc;rtel 18-20, Vienna, Austria</aff><aff id="aff0010"><label>b</label>Department of Internal Medicine III, Medical University of Vienna, W&#x000e4;hringer G&#x000fc;rtel 18-20, Vienna, Austria</aff>

The Journal of Steroid Biochemistry and Molecular Biology

<sec id="sec0005"><label>1</label><title>Introduction</title><p id="par0020">Epidemiological studies demonstrate an inverse correlation between calcium and vitamin D intake and risk of tumor development <xref rid="bib0005" ref-type="bibr">[1]</xref>, <xref rid="bib0010" ref-type="bibr">[2]</xref>. The calcium-sensing receptor (CaSR) is a putative tumor suppressor gene in the colon, which partially mediates the anti-proliferative and pro-differentiating actions of calcium in colonocytes (for review, see <xref rid="bib0015" ref-type="bibr">[3]</xref>, <xref rid="bib0020" ref-type="bibr">[4]</xref>). However, in colon cancer anti-proliferative effects of Ca²⁺ are lost <xref rid="bib0025" ref-type="bibr">[5]</xref>, <xref rid="bib0030" ref-type="bibr">[6]</xref>, and this could be due to loss of CaSR expression during colorectal tumorigenesis <xref rid="bib0035" ref-type="bibr">[7]</xref>. Very little is known about the factors that regulate the expression of CaSR in the colon. The <italic>CaSR</italic> gene contains 6 coding exons and two 5&#x02032;-untranslated exons (exons 1A and 1B), which are under the control of promoter 1 and 2, respectively, yielding alternative transcripts but coding for the same protein <xref rid="bib0040" ref-type="bibr">[8]</xref>, <xref rid="bib0045" ref-type="bibr">[9]</xref>. Several studies performed in rat parathyroid, thyroid, and kidney have mapped binding sites of numerous transcription factors, including NF-&#x003ba;B, STAT, SP1, and vitamin D response elements in both CaSR promoters (<xref rid="fig0005" ref-type="fig">Fig. 1</xref>) <xref rid="bib0045" ref-type="bibr">[9]</xref>, <xref rid="bib0050" ref-type="bibr">[10]</xref>, <xref rid="bib0055" ref-type="bibr">[11]</xref>, <xref rid="bib0060" ref-type="bibr">[12]</xref>. Currently, there is limited knowledge regarding the role of 1,25D₃ and of the proinflammatory cytokines TNF&#x003b1; and IL-6 on CaSR expression in the colon. Therefore, in the present study, we studied the impact of 1,25D₃, TNF&#x003b1;, and IL-6 on transcriptional and translational regulation of CaSR in two colon cancer cell lines with different proliferation and differentiation properties, mimicking different tumor stages.<fig id="fig0005"><label>Fig. 1</label><caption><p>Schematic illustration of the CaSR promoter region including exon 1A and exon 1B. Position of binding sites for regulatory elements is shown (signal transducer and activator of transcription (STAT), vitamin D response elements (VDRE), nuclear factor kappa B (NF-&#x003ba;B), specificity protein 1 (SP1)), which are critical for 1,25D₃, TNF&#x003b1;, and IL-6 responsiveness, as well as the CAAT and TATA boxes. Transcription start sites (TSS) 1 and 2 according to <xref rid="bib0060" ref-type="bibr">[12]</xref> were taken as point of reference for positioning the indicated binding sites in the corresponding promoters.</p></caption><graphic xlink:href="gr1"/></fig></p></sec><sec id="sec0010"><label>2</label><title>Materials and methods</title><sec id="sec0015"><label>2.1</label><title>Cell culture</title><p id="par0025">Caco2/AQ cells are a subclone of the Caco-2 cell line <xref rid="bib0065" ref-type="bibr">[13]</xref>. These carry a truncated APC and a missense mutation of &#x003b2;-catenin, and are able to differentiate spontaneously in culture. In the current study we used highly differentiated, 2 weeks post-confluent Caco2/AQ cells. Coga1A is a cell line derived from a moderately differentiated (G2) colon tumor <xref rid="bib0070" ref-type="bibr">[14]</xref>. These cells are heterozygous for truncated APC, without any known &#x003b2;-catenin mutations <xref rid="bib0075" ref-type="bibr">[15]</xref>. Confluent Caco2/AQ and Coga1A cells were treated for 6, 12, 24, and 48&#x000a0;h either with 10&#x000a0;nM 1,25D₃, 50&#x000a0;ng/mL TNF&#x003b1; (Sigma Aldrich, USA), 100&#x000a0;ng/mL IL-6 (Immunotools, Germany), or the combination of these compounds. Vehicle treated cells were used as controls.</p></sec><sec id="sec0020"><label>2.2</label><title>RNA isolation, reverse transcription, and real time qRT-PCR</title><p id="par0030">RNA isolation and reverse transcription were performed as described previously <xref rid="bib0080" ref-type="bibr">[16]</xref>. Real time qRT-PCR analyses were performed in StepOne Plus system using POWER SYBR GREEN Mastermix following the manufacturer's recommendations (Life Technologies, USA). Data were normalized to the expression of the reference genes: &#x003b2;2M or RPLP0 <xref rid="bib0085" ref-type="bibr">[17]</xref>, <xref rid="bib0090" ref-type="bibr">[18]</xref>, and set relative to the calibrator (Clontech, USA) to calculate the &#x00394;&#x00394;<italic>C</italic>_T value. Primer sequences for CaSR were: 5&#x02032;-AGCCCAGATGCAAGCAGAAGG-3&#x02032; forward, 5&#x02032;-TCTGGTGCGTAGAATTCCTGTGG-3&#x02032; reverse.</p></sec><sec id="sec0025"><label>2.3</label><title>Immunofluorescent staining of colon cancer cells</title><p id="par0035">Cells were grown on sterile glass cover slips. After treatments cells were fixed with 3.7% paraformaldehyde in PBS, permeabilized with 0.2% Triton-X (Sigma Aldrich, USA) for 20&#x000a0;min, and blocked with 5% goat serum (Jackson ImmunoResearch, USA). Cells were incubated either with rabbit polyclonal anti-CaSR antibody (1:100, Anaspec, USA) or mouse monoclonal anti-CaSR antibody (1:200, Abcam, UK) for 1&#x000a0;h at room temperature. As negative control we used rabbit or mouse IgG, respectively (Abcam, UK and Life Technologies, USA). As secondary antibody we used Dylight labeled 549 goat-anti-rabbit or Alexa Fluor 647 goat-anti-mouse IgG (1:500, Vector Laboratories and Life Technologies, USA). Nuclei were stained with DAPI (Roche, Switzerland). Images were acquired using TissueFAXS 2.04 (TissueGnostics, Austria).</p></sec><sec id="sec0030"><label>2.4</label><title>Statistical analysis</title><p id="par0040">All statistical analyses were performed with SPSS version 18 and graphs were drawn with GraphPad Prism version 5. In case of non-normal distribution, data were log transformed to achieve normal distribution and then subjected to one way ANOVA, followed by Tukey's multiple comparisons posttest. <italic>p</italic>-values smaller than 0.05 were regarded as statistically significant.</p></sec></sec><sec id="sec0035"><label>3</label><title>Results</title><sec id="sec0040"><label>3.1</label><title>Impact of 1,25D₃ on CaSR expression</title><p id="par0045">To study the role of vitamin D response elements on transcriptional regulation of CaSR expression we treated Caco2/AQ and Coga1A cells with 1,25D₃ for 6, 12, 24, and 48&#x000a0;h. In differentiated Caco2/AQ cells treatment with 1,25D₃ caused 2.4-fold induction of CaSR expression after 6&#x000a0;h. The maximal effect of 1,25D₃ on CaSR transcriptional activation in these cells was observed at 24&#x000a0;h (7.6-fold; <xref rid="fig0010" ref-type="fig">Fig. 2</xref>, <xref rid="fig0015" ref-type="fig">Fig. 3</xref>). In the less differentiated cells Coga1A 1,25D₃-induced CaSR transcription was 2.9-fold after 12&#x000a0;h and 4.2-fold after 24&#x000a0;h compared with the control group (<xref rid="fig0010" ref-type="fig">Fig. 2</xref>B). 1,25D₃ increased CaSR translation as well. Immunofluorescence staining demonstrated upregulation of the CaSR protein in Caco2/AQ after 24&#x000a0;h and Coga1A after 48&#x000a0;h (<xref rid="fig0015" ref-type="fig">Fig. 3</xref>C and D).<fig id="fig0010"><label>Fig. 2</label><caption><p>Transcriptional regulation of CaSR by 1,25D₃ in colon cancer cell lines. Caco2/AQ and Coga1A cells were treated with 10&#x000a0;nM 1,25D3 for the indicated time points. Bars represent mean&#x000a0;&#x000b1;&#x000a0;SEM of 2-3 independent experiments.</p></caption><graphic xlink:href="gr2"/></fig><fig id="fig0015"><label>Fig. 3</label><caption><p>Effect of 1,25D₃, TNF&#x003b1;, and IL-6 on CaSR expression. (A and B) mRNA expression of Caco2/AQ and Coga1A cells assessed by real time qRT-PCR. Data were log transformed to achieve normal distribution, then subjected to one way ANOVA and corrected with Tukey's posttest for multiple comparisons. Bars represent mean&#x000a0;&#x000b1;&#x000a0;SEM of 2-3 independent experiments, asterisks above bars indicate statistically significant changes compared with control. *<italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.05, **<italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.01. (C and D) Immunofluorescence staining of the CaSR protein (red) and nuclear staining (blue). Scale bar was 50&#x000a0;&#x003bc;m.</p></caption><graphic xlink:href="gr3"/></fig></p></sec><sec id="sec0045"><label>3.2</label><title>Impact of TNF&#x003b1; and IL-6 on CaSR expression</title><p id="par0050">We treated Caco2/AQ and Coga1A cells with TNF&#x003b1; and IL-6 for 6, 12, 24, and 48&#x000a0;h. In Caco2/AQ treatment with the proinflammatory cytokine TNF&#x003b1; caused only modest upregulation of CaSR expression. Treatment with IL-6 was accompanied by a 3.5-fold induction after 6&#x000a0;h compared with control. Combined treatment with TNF&#x003b1; and IL-6 induced CaSR mRNA expression in Caco2/AQ 10.3-fold (<italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.05) after 24&#x000a0;h and 10.2-fold (<italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.05) after 48&#x000a0;h. However, the combination of all three compounds either had no effect or reduced CaSR expression (<xref rid="fig0015" ref-type="fig">Fig. 3</xref>A).</p><p id="par0055">In Coga1A cells, treatment with TNF&#x003b1; induced CaSR robustly, especially at 48&#x000a0;h (134-fold, <italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.01). Treatment with IL-6 caused only marginal increases in CaSR mRNA expression. Furthermore, we observed upregulation of CaSR expression in the groups treated with TNF&#x003b1;/IL-6 (68.5-fold) and TNF&#x003b1;/1,25D₃ (121.2-fold, <italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.05) at 48&#x000a0;h. Similar results were observed in the groups that were treated with TNF&#x003b1;/IL-6/1,25D₃ at 6 and 48&#x000a0;h (18.8-fold, <italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.05 and 47.7-fold, <italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.05; <xref rid="fig0015" ref-type="fig">Fig. 3</xref>B).</p><p id="par0060">To address the question whether alterations on CaSR mRNA expression were translated into protein, we performed immunofluorescence staining. <xref rid="fig0015" ref-type="fig">Fig. 3</xref>C and D demonstrates the upregulation of the CaSR protein upon treatments with the proinflammatory cytokines using the rabbit polyclonal anti-CaSR antibody. Protein expression data were confirmed using the mouse monoclonal anti-CaSR antibody (data not shown). Both antibodies gave the same results.</p></sec></sec><sec id="sec0050"><label>4</label><title>Discussion</title><p id="par0065">Recent studies have demonstrated that murine CaSR activates the NLPR3 inflammasome, which in turn induces maturation and release of the inflammatory cytokine interleukin 1&#x003b2;, amplifying the inflammatory signal <xref rid="bib0095" ref-type="bibr">[19]</xref>, <xref rid="bib0100" ref-type="bibr">[20]</xref>. Inversely, mice double knockout for CaSR^{&#x02212;/&#x02212;}/PTH^{&#x02212;/&#x02212;} had increased inflammatory response after administration of dextran sodium sulfate compared with control mice expressing the receptor <xref rid="bib0105" ref-type="bibr">[21]</xref>. This suggests an important role for the CaSR in inflammation. Therefore, it is essential to understand how the expression of the CaSR is modulated in the colon.</p><p id="par0070">It has been demonstrated previously that activation of VDREs by 1,25D₃ and translocation of NF-&#x003ba;B to the nucleus after the treatment with interleukin 1&#x003b2; led to induction of CaSR expression in rat parathyroid, thyroid, and kidneys <xref rid="bib0045" ref-type="bibr">[9]</xref>, <xref rid="bib0050" ref-type="bibr">[10]</xref>. Furthermore, IL-6 injection in rats caused induction of CaSR transcription <italic>via</italic> Stat1/3 response elements in promoter 1 and Sp1/3 sites in promoter 2 <xref rid="bib0055" ref-type="bibr">[11]</xref>, but not much is known about the regulation of CaSR expression in the colon.</p><p id="par0075">Our study is the first to show that in colonocytes inflammatory cytokines are able to upregulate CaSR expression, and that this effect is time- and cell line-specific. In the present study, we investigated the role of 1,25D₃, TNF&#x003b1;, and IL-6 on the transcriptional and translational activation of the CaSR in two cell lines representing a highly differentiated and a moderately differentiated colorectal tumor.</p><p id="par0080">1,25D₃ is known for its anti-proliferative, pro-differentiating effects (for review, see <xref rid="bib0110" ref-type="bibr">[22]</xref>), and its involvement in regulating epigenetic mechanisms <xref rid="bib0115" ref-type="bibr">[23]</xref>. Inducing expression of CaSR, a putative tumor suppressor in the colon, might be one of the tumor preventive mechanisms of 1,25D₃. In the differentiated Caco2/AQ cells 1,25D₃ had more pronounced impact in inducing the expression of CaSR than in the less differentiated Coga1A cells. In Caco2/AQ cells treatment with 1,25D₃ reduced the expression of several proliferation markers also. This was much less evident in the Coga1A cells (data not shown), although the level of the vitamin D receptor is similar <xref rid="bib0075" ref-type="bibr">[15]</xref>.</p><p id="par0085">In Caco2/AQ cells, both TNF&#x003b1; and IL-6 increased CaSR expression to a lesser extent than 1,25D₃. In combination, however, they caused a strong upregulation at 6&#x000a0;h, which was lost at 12&#x000a0;h; at 24&#x000a0;h the effect became additive and the CaSR level remained high also after 48&#x000a0;h. We hypothesized that two different mechanisms were responsible: first, direct upregulation of CaSR expression due to a transient activation of CaSR promoters by NF-&#x003ba;B upon treatment with TNF&#x003b1; and Stat1/3 and Sp1/3 elements by IL-6. This was followed by a second induction of transcription that seems to be indirect. Some (still unknown) factors induced by TNF&#x003b1; and IL-6 might be needed for this more stable induction of CaSR expression. Unexpectedly, 1,25D₃ counteracted this additive effect, suggesting the existence of intricate feedback systems.</p><p id="par0090">In Coga1A cells, the CaSR was more sensitive to the proinflammatory cytokine TNF&#x003b1;, which was the main driver of CaSR expression in these cells. The low effectiveness of IL-6 in upregulating CaSR expression could be due to lower levels of the IL-6 receptor complex (both the IL-6 binding &#x003b1; chain and the signal transducing unit gp130) in Coga1A cells compared with Caco2/AQ <xref rid="bib0120" ref-type="bibr">[24]</xref>. Interestingly, in these cells the CaSR protein levels remained enhanced in all combined treatments. The robust increase of CaSR expression by TNF&#x003b1; treatment in Coga1A cells could be regarded as a defense mechanism against inflammation. Such protective mechanism was shown in murine macrophages, where lipopolysaccharide-induced TNF&#x003b1; release upregulated CaSR expression leading to inhibition of TNF&#x003b1; synthesis, in a negative feedback manner <xref rid="bib0125" ref-type="bibr">[25]</xref>.</p><p id="par0095">In conclusion, our results demonstrate for the first time that in colon cancer cells not only 1,25D₃, but also the proinflammatory cytokines TNF&#x003b1; and IL-6 were able to induce the expression of CaSR. How this observation can be translated <italic>in vivo</italic> and used for the treatment of inflammation in the gut, still needs to be explored.</p></sec>

Estimating the burden of disease in chronic pain with and without neuropathic characteristics: Does the choice between the EQ-5D and SF-6D matter?

Could not extract abstract

<contrib contrib-type="author" id="au005"><name><surname>Torrance</surname><given-names>Nicola</given-names></name><email>n.torrance@dundee.ac.uk</email><xref rid="af005" ref-type="aff">a</xref><xref rid="cor1" ref-type="corresp">&#x0204e;</xref></contrib><contrib contrib-type="author" id="au010"><name><surname>Lawson</surname><given-names>Kenny D.</given-names></name><xref rid="af010" ref-type="aff">b</xref></contrib><contrib contrib-type="author" id="au015"><name><surname>Afolabi</surname><given-names>Ebenezer</given-names></name><xref rid="af015" ref-type="aff">c</xref></contrib><contrib contrib-type="author" id="au020"><name><surname>Bennett</surname><given-names>Michael I.</given-names></name><xref rid="af020" ref-type="aff">d</xref></contrib><contrib contrib-type="author" id="au025"><name><surname>Serpell</surname><given-names>Michael G.</given-names></name><xref rid="af025" ref-type="aff">e</xref></contrib><contrib contrib-type="author" id="au030"><name><surname>Dunn</surname><given-names>Kate M.</given-names></name><xref rid="af015" ref-type="aff">c</xref></contrib><contrib contrib-type="author" id="au035"><name><surname>Smith</surname><given-names>Blair H.</given-names></name><xref rid="af005" ref-type="aff">a</xref></contrib><aff id="af005"><label>a</label>Medical Research Institute, University of Dundee, Dundee, Scotland, UK</aff><aff id="af010"><label>b</label>Centre for Research Excellence in the Prevention of Chronic Conditions in Rural and Remote Populations, James Cook University, Townsville City, Australia</aff><aff id="af015"><label>c</label>Research Institute for Primary Care &#x00026; Health Sciences, Keele University, Staffordshire, England, UK</aff><aff id="af020"><label>d</label>Academic Unit of Palliative Care, University of Leeds, Leeds, England, UK</aff><aff id="af025"><label>e</label>School of Medicine, University of Glasgow, Glasgow, Scotland, UK</aff>

Pain

<sec id="s0005"><label>1</label><title>Introduction</title><p id="p0005">Chronic pain is common, affecting up to half of the adult population <xref rid="b0085" ref-type="bibr">[17]</xref>, <xref rid="b0235" ref-type="bibr">[47]</xref>. Approximately 20% of the adult European population has significant chronic pain, and 7% to 8% of the population has chronic pain with neuropathic features <xref rid="b0035" ref-type="bibr">[7]</xref>, <xref rid="b0050" ref-type="bibr">[10]</xref>, <xref rid="b0235" ref-type="bibr">[47]</xref>. Health-related quality of life (HRQoL) is significantly poorer in people with chronic pain than in those without <xref rid="b0210" ref-type="bibr">[42]</xref>, and poorer in people with neuropathic pain than in those with nonneuropathic pain <xref rid="b0010" ref-type="bibr">[2]</xref>, <xref rid="b0080" ref-type="bibr">[16]</xref>, <xref rid="b0125" ref-type="bibr">[25]</xref>, <xref rid="b0215" ref-type="bibr">[43]</xref>, <xref rid="b0240" ref-type="bibr">[48]</xref>.</p><p id="p0010">HRQoL measures can be categorized into disease-specific and generic measures. It is important that these measures are valid, appropriate to the disease, and particularly for clinical trials, sensitive to detect changes. Although a disease-specific measure for neuropathic pain has been developed <xref rid="b0195" ref-type="bibr">[39]</xref>, many studies use a generic HRQoL measure alongside clinical assessment or a validated neuropathic pain screening tool <xref rid="b0010" ref-type="bibr">[2]</xref>, <xref rid="b0215" ref-type="bibr">[43]</xref>. A single summary score of overall HRQoL is generated by weighting responses to mental and physical health states by their perceived importance, using patient or general population preferences. This single summary score can be used to compare the impact of different conditions and how they vary across populations and to quantify the effectiveness of interventions. Economists term such scores health utilities or health utility index scores. Health utilities are measured on a scale where 0 represents a health state equivalent to death and 1 represents full health, with the potential for values less than zero representing states &#x0201c;worse than death&#x0201d; (WTD). A concern in only using generic measures is whether they are sensitive enough to discriminate between patients in whom important but complex differences might be detected by more specific measures.</p><p id="p0015">The EQ-5D <xref rid="b0090" ref-type="bibr">[18]</xref> and Short Form (SF) 12/36 <xref rid="b0265" ref-type="bibr">[53]</xref>, <xref rid="b0270" ref-type="bibr">[54]</xref> questionnaires are widely used generic HRQoL measures. They are a common means of generating health state values using an algorithm to derive health utility scores (SF-6D from SF-12/36) <xref rid="b0025" ref-type="bibr">[5]</xref> and are used by economists to calculate quality adjusted life years (QALYs) <xref rid="b0175" ref-type="bibr">[35]</xref>, <xref rid="b0245" ref-type="bibr">[49]</xref>, <xref rid="b0285" ref-type="bibr">[57]</xref> and in economic evaluations to evaluate the cost effectiveness of health care interventions.</p><p id="p0020">The association between neuropathic pain conditions and health utilities has been described in a multicenter European cross-sectional survey <xref rid="b0165" ref-type="bibr">[33]</xref> and a systematic review <xref rid="b0080" ref-type="bibr">[16]</xref>, both finding a significant relationship between increasing pain severity and reduced HRQoL. The EQ-5D has been used to measure HRQoL of patients with specific neuropathic pain diagnoses <xref rid="b0080" ref-type="bibr">[16]</xref>, but there are few published studies that have used SF-6D in patients with neuropathic pain <xref rid="b0145" ref-type="bibr">[29]</xref>, despite its widespread use and acceptance. Studies have directly compared EQ-5D and SF-6D utilities in patients with chronic painful conditions, such as arthritis <xref rid="b0040" ref-type="bibr">[8]</xref>, <xref rid="b0070" ref-type="bibr">[14]</xref>, <xref rid="b0115" ref-type="bibr">[23]</xref>, <xref rid="b0130" ref-type="bibr">[26]</xref>, <xref rid="b0155" ref-type="bibr">[31]</xref>, <xref rid="b0185" ref-type="bibr">[37]</xref>, <xref rid="b0285" ref-type="bibr">[57]</xref>, low back pain <xref rid="b0040" ref-type="bibr">[8]</xref>, <xref rid="b0070" ref-type="bibr">[14]</xref>, <xref rid="b0185" ref-type="bibr">[37]</xref>, and nonspecific neck pain <xref rid="b0275" ref-type="bibr">[55]</xref>, with most finding moderate levels of agreement. These patients were generally recruited according to strict criteria and are unlikely to represent HRQoL associated with these conditions in general populations and primary care, where most chronic and neuropathic pain is treated and managed <xref rid="b0230" ref-type="bibr">[46]</xref>.</p><p id="p0025">This study collected data on HRQoL and chronic pain status as part of a large UK general population questionnaire survey <xref rid="b0230" ref-type="bibr">[46]</xref>. We compared EQ-5D and SF-6D health utility index scores in individuals with chronic pain, with and without neuropathic characteristics. We also explored the ability of these generic measures to discriminate between the health utilities of patients with different pain severities.</p></sec><sec id="s0010"><label>2</label><title>Methods</title><sec id="s0015"><label>2.1</label><title>Sample selection</title><p id="p0030">In the UK, around 96% of the population is registered with a general practitioner (family doctor, GP) <xref rid="b0180" ref-type="bibr">[36]</xref>; a GP practice population therefore approximates to a general population sample. This study surveyed 10,000 individuals in 5 UK locations, with 2 GP practices in each locality generating a random sample of 1000 registered adult patients. Each practice&#x02019;s electronic register was used to generate a random sample of patients over age 18&#x000a0;years. The sample list was then screened by the GPs in each practice, to exclude patients in whom inquiry might be insensitive or inappropriate (for example, in terminal illness or with severe learning difficulties). Details of the sample selection procedures have been reported previously <xref rid="b0230" ref-type="bibr">[46]</xref>.</p></sec><sec id="s0020"><label>2.2</label><title>Patient questionnaire</title><p id="p0035">Individuals in the study sample were mailed a self-complete questionnaire that contained demographic items (age, sex, smoking, marital and employment status, educational attainment, and home ownership as a proxy for social class <xref rid="b0220" ref-type="bibr">[44]</xref>), HRQoL measures including the SF-12 and EQ-5D questionnaires, chronic pain identification and severity questions, and the Self-Complete Leeds Assessment of Neuropathic Symptoms and Signs (S-LANSS) questionnaire, which is used to identify pain with neuropathic characteristics (NC) <xref rid="b0020" ref-type="bibr">[4]</xref>.</p><sec id="s0025"><label>2.2.1</label><title>Pain ascertainment and characteristics</title><p id="p0040">Chronic pain was identified by affirmative answers to 2 questions: (1) Are you currently troubled by pain or discomfort, either all the time or on and off? (2) Have you had this pain or discomfort for more than 3&#x000a0;months? <xref rid="b0120" ref-type="bibr">[24]</xref> Identical case identification questions have been used in previous population-based research on chronic pain <xref rid="b0065" ref-type="bibr">[13]</xref>, <xref rid="b0085" ref-type="bibr">[17]</xref>, <xref rid="b0235" ref-type="bibr">[47]</xref>.</p><p id="p0045">The S-LANSS questionnaire is a validated 7-item questionnaire including 5 questions about pain characteristics and 2 self-examination items; its responses are weighted to provide a maximum score of 24, with a score &#x02a7e;12 indicating pain with NC <xref rid="b0020" ref-type="bibr">[4]</xref>. Pain severity was measured using the average pain intensity numeric rating scale (NRS) of the Chronic Pain Grade (CPG). This is a 0 to 10 NRS anchored at 0 for no pain and 10 for pain as bad as can be in the past 3&#x000a0;months <xref rid="b0255" ref-type="bibr">[51]</xref>.</p></sec><sec id="s0030"><label>2.2.2</label><title>HRQoL and generating health utility scores</title><p id="p0050">All respondents were asked to complete the SF-12 and EQ-5D HRQoL questionnaires before completion of the chronic pain screening questions. The SF-12 is a validated 12-item self-administered tool for measuring health status derived from the SF-36 <xref rid="b0265" ref-type="bibr">[53]</xref>. The SF-12 has been used in large general population questionnaire studies of chronic pain <xref rid="b0010" ref-type="bibr">[2]</xref>, <xref rid="b0060" ref-type="bibr">[12]</xref>, <xref rid="b0145" ref-type="bibr">[29]</xref>, <xref rid="b0150" ref-type="bibr">[30]</xref> and in studies of specific neuropathic pain conditions, such as postherpetic neuralgia <xref rid="b0030" ref-type="bibr">[6]</xref>. SF-12 scores can be calculated in 8 health domains: physical functioning, role physical, bodily pain, general health, vitality, social functioning, role emotional, and mental health. To generate a single health utility score from the SF-12, we used the standard SF-6D algorithm <xref rid="b0045" ref-type="bibr">[9]</xref>. This algorithm involves preference weighting of 6 of the SF-12 question responses (3 physical health and 3 mental health) by the desirability for different health states. These preference weights were derived from a survey representative of the UK general population <xref rid="b0045" ref-type="bibr">[9]</xref>. Summing across weighted question responses generated a health utility score for each respondent <xref rid="b0045" ref-type="bibr">[9]</xref>. The SF-6D generates a score on a 0.29 to 1.00 scale, with 1.00 indicating full health. The SF-6D can define 18,000 and 7500 health states for the SF-6D (SF-36) and SF-6D (SF-12), respectively <xref rid="b0280" ref-type="bibr">[56]</xref>.</p><p id="p0055">The EQ-5D is a generic measure of health status and defines health in terms of 5 dimensions: mobility, self-care, usual activities (work, study, housework, family, or leisure), pain or discomfort, and anxiety or depression, and is well validated in population studies <xref rid="b0090" ref-type="bibr">[18]</xref>, <xref rid="b0135" ref-type="bibr">[27]</xref>. A preference-based set of weights (or algorithm) is used to calculate a single EQ-5D index-based utility score. The EQ-5D generates a score on a &#x02212;0.59 to 1.00 scale, with 1.00 indicating full health and 0 equal to death. The negative EQ-5D utility values theoretically correspond to health states valued as WTD. (Negative utility values are not available with SF-6D.) The EQ-5D can define 243 distinct states <xref rid="b0280" ref-type="bibr">[56]</xref>.</p></sec></sec><sec id="s0035"><label>2.3</label><title>Data analysis</title><p id="p0060">A complete case analysis was conducted. Descriptive statistics show the sociodemographic characteristics of the whole study sample and by defined pain group (no chronic pain; chronic pain without neuropathic characteristics (S-LANSS&#x000a0;&#x0003c;&#x000a0;12; chronic pain without NC); chronic pain with neuropathic characteristics (S-LANSS&#x000a0;&#x02a7e;&#x000a0;12; chronic pain with NC). Linear regression analysis was conducted to determine differences in health utilities among the 3 defined pain subpopulations (no chronic pain vs chronic pain without NC, no chronic pain vs chronic pain with NC, chronic pain without NC vs chronic pain with NC) with adjustment for all significant demographic variables.</p><p id="p0065">The data analysis was divided into 3 elements: an assessment of the correlation between EQ-5D and SF-6D dimensions, the range of observed health utility index scores, and the sensitivity of both instruments to detect differences between chronic pain types and degrees of pain severity.</p><sec id="s0040"><label>2.3.1</label><title>Correlation between EQ-5D and SF-6D dimensions</title><p id="p0070">EQ-5D dimensions are what the respondents reported, and the EQ-5D health utility indices are derived using the algorithm. Similarly, SF-6D dimensions were obtained from the SF-12 responses, and health utility indices were calculated using the SF-6D algorithm. The first section of the analysis is based on EQ-5D and SF-6D self-reported dimensions with no reference to single index scores. This assessment of the degree of agreement between dimensions of the 2 instruments used Spearman rank correlations across the whole sample and by defined pain group. The value of the correlation coefficient can be interpreted thus: 1 is perfect, 0.7 to 0.9 is strong, 0.4 to 0.69 is moderate, 0.1 to 0.39 is weak, and 0 is no correlation <xref rid="b0075" ref-type="bibr">[15]</xref>. We hypothesized that there would be correlations between the dimensions purporting to capture similar aspects in SF-12 and EQ-5D. These similar dimensions are physical functioning and mobility, physical functioning and usual activities, role limitations and usual activities, role limitations and anxiety/depression, social functioning and usual activities, pain and pain/discomfort, mental health and anxiety/depression, vitality and usual activities <xref rid="b0275" ref-type="bibr">[55]</xref>.</p></sec><sec id="s0045"><label>2.3.2</label><title>Range of the health utility index scores</title><p id="p0075">Further analysis then explored the health utility indices with basic descriptive statistics, including means, medians, and ranges. The level of agreement between EQ-5D and SF-6D index scores also was examined by calculating the intraclass correlation coefficient (ICC), using a 2-way mixed model based on absolute agreement, where the 2 measures are treated as a source of variability <xref rid="b0040" ref-type="bibr">[8]</xref>, <xref rid="b0275" ref-type="bibr">[55]</xref>. The values of the ICC can theoretically range from 0 to 1, with a higher value indicating that less variance is due to other factors such as differences between observations. For the 2 instruments, the level used to interpret the ICC was 0.00 to 0.10&#x000a0;=&#x000a0;virtually no agreement, 0.11 to 0.40&#x000a0;=&#x000a0;slight, 0.41 to 0.60&#x000a0;=&#x000a0;fair, 0.61 to 0.80&#x000a0;=&#x000a0;moderate, and 0.81 to 1.00&#x000a0;=&#x000a0;substantial agreement <xref rid="b0205" ref-type="bibr">[41]</xref>, <xref rid="b0280" ref-type="bibr">[56]</xref>. Floor and ceiling effects (proportion of respondents with the best and worst possible theoretical scores, respectively) were also explored for both the EQ-5D and SF-6D indices.</p></sec><sec id="s0050"><label>2.3.3</label><title>Discriminating health utility index scores among degrees of pain severity</title><p id="p0080">In comparing the EQ-5D and SF-6D, a key criterion is whether the instruments are sensitive enough to discriminate differences in reported patients&#x02019; pain severity considered to be clinically meaningful. Because this was a cross-sectional study, we were unable to measure changes in utility scores over time. However, within an exploratory analysis, we were able to estimate the difference in utility scores for respondents with different levels of pain severity. This provides an indication of the potential change in utility scores that each instrument may detect if there was an intervention that moved patients between pain severity groups. This analysis is exploratory and intended to inform whether further research using a longitudinal study is necessary.</p><p id="p0085">Patients were first divided into 3 categories of reported pain severity. We used clinically validated cut-points for pain severity to create categories of mild, 0 to 3; moderate, 4 to 6; and severe, 7 to 10 chronic pain on the average pain intensity NRS of the Chronic Pain Grade <xref rid="b0255" ref-type="bibr">[51]</xref>, <xref rid="b0295" ref-type="bibr">[59]</xref>. We then compared mean utility scores from the EQ-5D and SF-6D across these 3 pain intensity groups as a means to infer discriminant ability regarding HRQoL. We also tested whether the difference in utility scores among pain groups is also meaningful to patients in terms of their perception of HRQoL. This was done by comparing the differences in utility scores to the minimally important difference (MID) according to Walters and Brazier <xref rid="b0260" ref-type="bibr">[52]</xref>, who in comparing the results of 11 longitudinal studies (the majority of which were chronic pain&#x02013;related conditions, including back pain and arthritis) found the mean MID for EQ-5D to be 0.074 (range 0.011 to 0.140) and for SF-6D to be 0.041 (range 0.011 to 0.097) <xref rid="b0260" ref-type="bibr">[52]</xref>.</p><p id="p0090">Because previous research has found HRQoL in individuals with neuropathic pain to be worse than in those with nonneuropathic pain of the same severity <xref rid="b0215" ref-type="bibr">[43]</xref>, we hypothesized that those respondents in the chronic pain with NC group would be more likely (than chronic pain without NC) to report WTD scores in EQ-5D and therefore to demonstrate a floor effect equivalent to worst possible health (&#x02a7d; 0) <xref rid="b0275" ref-type="bibr">[55]</xref>.</p></sec></sec><sec id="s0055"><label>2.4</label><title>Ethics approval</title><p id="p0095">The study was approved by North of Scotland Research Ethics Committee, REC reference number 09/S0802/103.</p></sec></sec><sec id="s0060"><label>3</label><title>Results</title><sec id="s0065"><label>3.1</label><title>Characteristics of study sample</title><p id="p0100">In total, 10,000 postal questionnaires were mailed, with 347 returned as undelivered or unable to be completed due to illness or learning disability. Of the rest, 4541 completed questionnaires were returned, giving an overall corrected response rate of 47%. Further information on the respondents and the study sample has been reported previously <xref rid="b0230" ref-type="bibr">[46]</xref>.</p><p id="p0105">Of the 4451 returned questionnaires, 4408 individuals completed both of the 2 screening questions for chronic pain status. Any chronic pain was reported by 2202 (48.5%; 95% confidence interval 47.0% to 49.9%) and 2206 respondents reported no chronic pain. S-LANSS questionnaires were incomplete in 192 of those with any chronic pain, and these individuals were excluded from further categorization into the chronic pain with or without neuropathic pain groups for analysis. Therefore 1611 individuals were categorized as chronic pain without NC and 399 as chronic pain with NC (S-LANSS&#x000a0;&#x02a7e;&#x000a0;12).</p><p id="p0110">Completion rates were high for both HRQoL questionnaires, with the EQ-5D completed by 4349 (95.8%) and the SF-12 completed by 4176 (92.0%) of all respondents. The characteristics of the whole study sample and by pain group are shown in <xref rid="t0005" ref-type="table">Table 1</xref>. There were significant differences in all of the measured sociodemographic characteristics between respondents reporting chronic pain with NC and those reporting chronic pain without NC except for age (mean (SD) 56.0 (15.4) years vs 56.3 (15.3) years, <italic>P</italic>&#x000a0;=&#x000a0;.673). Individuals with chronic pain with NC were more likely to be women, no longer married, and living in council rented accommodation than individuals whose chronic pain that did not have NC. They were also more likely to be unable to work due to illness or disability, to have no educational qualifications, and to be smokers. Pairwise comparisons between pain groups found those reporting chronic pain with NC to have significantly lower EQ-5D and SF-6D mean utility scores than the no Chronic pain and chronic pain without NC groups (<italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.001). Linear regression found the significant differences in health utilities between the pain groups persisted (<italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.001) after adjustment for all significant sociodemographic variables (sex, marital status, employment, housing, general health, education, and smoking).<table-wrap position="float" id="t0005"><label>Table 1</label><caption><p>Sample characteristics of respondents by chronic pain group, n (%).</p></caption><table frame="hsides" rules="groups"><thead><tr><th/><th>Whole sample (n&#x000a0;=&#x000a0;4541)</th><th>No chronic pain (n&#x000a0;=&#x000a0;2206)</th><th>Chronic pain without NC (n&#x000a0;=&#x000a0;1611)</th><th>Chronic pain with NC (n&#x000a0;=&#x000a0;399)</th></tr></thead><tbody><tr><td colspan="5"><italic>Age, n (%)</italic></td></tr><tr><td>18 to 39 y</td><td>968 (21.5)</td><td>654 (28.5)</td><td>234 (14.7)</td><td>60 (15.3)</td></tr><tr><td>40 to 59 y</td><td>1789 (39.4)</td><td>928 (40.4)</td><td>622 (39.0)</td><td>167 (42.5)</td></tr><tr><td>60+ y</td><td>1738 (38.3)</td><td>692 (30.1)</td><td>740 (46.4)</td><td>166 (42.2)</td></tr><tr><td colspan="5">&#x02028;&#x02028;</td></tr><tr><td colspan="5"><italic>Gender</italic></td></tr><tr><td>Men</td><td>1928 (42.5)</td><td>1016 (44.3)</td><td>684 (42.5)</td><td>146 (36.8)</td></tr><tr><td>Women</td><td>2609 (57.5</td><td>1280 (55.7)</td><td>925 (57.5)</td><td>251 (63.2)</td></tr><tr><td colspan="5">&#x02028;&#x02028;</td></tr><tr><td colspan="5"><italic>Marital status</italic></td></tr><tr><td>Never married</td><td>634 (14.0)</td><td>382 (16.6)</td><td>166 (10.4)</td><td>55 (14.0)</td></tr><tr><td>Living as married</td><td>3190 (70.6)</td><td>1614 (70.3)</td><td>1194 (74.5)</td><td>245 (62.2)</td></tr><tr><td>No longer married</td><td>693 (15.3)</td><td>290 (12.6)</td><td>243 (15.2)</td><td>94 (23.9)</td></tr><tr><td colspan="5">&#x02028;&#x02028;</td></tr><tr><td colspan="5"><italic>Housing tenure</italic></td></tr><tr><td>Owned/mortgaged</td><td>3657 (81.2)</td><td>1914 (85.4)</td><td>1321 (82.6)</td><td>261 (66.4)</td></tr><tr><td>Council rent</td><td>527 (11.7)</td><td>190 (8.5)</td><td>188 (11.8)</td><td>92 (23.4)</td></tr><tr><td>Private rent/other</td><td>322 (7.1)</td><td>178 (7.8)</td><td>91 (5.7)</td><td>40 (10.2)</td></tr><tr><td colspan="5">&#x02028;&#x02028;</td></tr><tr><td colspan="5"><italic>Employment</italic></td></tr><tr><td>Employed</td><td>2486 (55.1)</td><td>1483 (64.6)</td><td>774 (48.3)</td><td>151 (38.4)</td></tr><tr><td>Retired</td><td>1456 (32.3)</td><td>562 (24.5)</td><td>638 (39.9)</td><td>137 (34.9)</td></tr><tr><td>Unable to work</td><td>199 (4.4)</td><td>28 (1.2)</td><td>70 (4.4)</td><td>77 (19.6)</td></tr><tr><td>Not employed/other</td><td>369 (8.2)</td><td>212 (9.2)</td><td>119 (7.4)</td><td>28 (7.1)</td></tr><tr><td colspan="5">&#x02028;&#x02028;</td></tr><tr><td colspan="5"><italic>Education</italic></td></tr><tr><td>No qualifications</td><td>862 (19.3)</td><td>323 (14.1)</td><td>345 (21.7)</td><td>115 (29.6)</td></tr><tr><td>Secondary school/equivalent</td><td>1785 (40.1)</td><td>933 (40.6)</td><td>599 (37.7)</td><td>165 (42.5)</td></tr><tr><td>Higher education</td><td>1808 (40.6)</td><td>1010 (44.0)</td><td>644 (40.6)</td><td>108 (27.8)</td></tr><tr><td colspan="5">&#x02028;&#x02028;</td></tr><tr><td colspan="5"><italic>Smoking</italic></td></tr><tr><td>Smoker</td><td>798 (17.6)</td><td>372 (16.2)</td><td>264 (16.4)</td><td>105 (26.4)</td></tr><tr><td>Ex-smoker</td><td>1396 (30.9)</td><td>632 (27.5)</td><td>580 (36.1)</td><td>109 (27.5)</td></tr><tr><td>Never smoked</td><td>2329 (51.5)</td><td>1288 (56.1)</td><td>761 (47.3)</td><td>183 (46.1)</td></tr><tr><td colspan="5">&#x02028;&#x02028;</td></tr><tr><td colspan="5"><italic>Perceived health status</italic></td></tr><tr><td>Excellent</td><td>571 (12.7)</td><td>475 (20.8)</td><td>88 (5.5)</td><td>7 (1.8)</td></tr><tr><td>Very good</td><td>1611 (35.7)</td><td>1022 (44.7)</td><td>470 (29.4)</td><td>65 (16.6)</td></tr><tr><td>Good</td><td>1436 (31.8)</td><td>633 (27.7)</td><td>626 (39.1)</td><td>122 (31.1)</td></tr><tr><td>Fair</td><td>649 (14.4)</td><td>149 (6.5)</td><td>316 (19.8)</td><td>105 (26.8)</td></tr><tr><td>Poor</td><td>246 (5.5)</td><td>9 (0.4)</td><td>100 (6.2)</td><td>93 (23.7)</td></tr><tr><td colspan="5">&#x02028;&#x02028;</td></tr><tr><td>SF-6D, mean (SD)</td><td>0.767 (0.15)</td><td>0.826 (0.12)</td><td>0.728 (0.14)</td><td>0.619 (0.15)</td></tr><tr><td>EQ-5D index score, mean (SD)</td><td>0.794 (0.27)</td><td>0.932 (0.13)</td><td>0.702 (0.25)</td><td>0.468 (0.36)</td></tr><tr><td>EQ-VAS, mean (SD)</td><td>77.82 (19.38)</td><td>85.31 (13.5)</td><td>73.35 (19.3)</td><td>59.67 (24.0)</td></tr></tbody></table><table-wrap-foot><fn id="sp0025"><p>Pairwise comparisons (independent samples <italic>t</italic> tests) for SF-6D and EQ-5D found <italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.001 for no chronic pain vs chronic pain without NC, no chronic pain vs chronic pain with NC, chronic pain without NC vs chronic pain with NC.</p></fn><fn id="sp0030"><p>NC&#x000a0;=&#x000a0;neuropathic characteristics.</p></fn></table-wrap-foot></table-wrap></p></sec><sec id="s0070"><label>3.2</label><title>Correlation between EQ-5D and SF-6D dimensions</title><p id="p0115"><xref rid="t0010" ref-type="table">Table 2</xref> presents the relationships between EQ-5D and the SF-6D dimensions as measured by Spearman correlation coefficients for the whole dataset, and then separately for the 2 chronic pain groups. The similar dimensions that were expected to have the highest correlations are underlined <xref rid="b0040" ref-type="bibr">[8]</xref>, <xref rid="b0275" ref-type="bibr">[55]</xref> with the 5 actual most correlated dimensions shown in boldface type. We identified 8 paired dimensions that were expected to be among the highest correlations. Of these, the highest correlations were: between mobility and usual activities (EQ-5D) with physical functioning (SF-6D), between usual activities (EQ-5D) and role limitations and pain (SF-6D), between pain/discomfort (EQ-5D) and pain (SF-6D), and between anxiety/depression (EQ-5D) mental health (SF-6D). These all had a correlation coefficient &#x0003e;0.60, with the strongest correlation (0.75) found between the 2 pain dimensions in the whole sample. These results provide evidence for agreement, but lower correlations were found between other similar dimensions, notably between anxiety/depression (EQ-5D) and mental health (SF-6D). The 2-way mixed ICC for the whole sample was 0.61, which suggests moderate agreement between the 2 measures, with the ICC 0.44 for the chronic pain with NC and 0.57 for chronic pain without NC groups.<table-wrap position="float" id="t0010"><label>Table 2</label><caption><p>The correlation between EQ-5D and SF-6D dimensions (Spearman rank correlation).</p></caption><table frame="hsides" rules="groups"><thead><tr><th rowspan="2">SF-6D</th><th colspan="5">EQ-5D<hr/></th></tr><tr><th>Mobility</th><th>Self care</th><th>Usual activities</th><th>Pain/discomfort</th><th>Anxiety/depression</th></tr></thead><tbody><tr><td colspan="6"><italic>Whole sample, n&#x000a0;=&#x000a0;4541</italic></td></tr><tr><td>Physical functioning</td><td align="char"><bold><underline>0.65</underline></bold></td><td align="char">0.44</td><td align="char"><bold><underline>0.64</underline></bold></td><td align="char">0.55</td><td align="char">0.30</td></tr><tr><td>Role limitations</td><td align="char">0.63</td><td align="char">0.45</td><td align="char"><bold><underline>0.71</underline></bold></td><td align="char">0.60</td><td align="char"><underline>0.36</underline></td></tr><tr><td>Social functioning</td><td align="char">0.46</td><td align="char">0.41</td><td align="char"><underline>0.56</underline></td><td align="char">0.44</td><td align="char">0.56</td></tr><tr><td>Pain</td><td align="char">0.61</td><td align="char">0.41</td><td align="char"><bold>0.65</bold></td><td align="char"><underline>0.75</underline></td><td align="char"><bold>0.35</bold></td></tr><tr><td>Mental health</td><td align="char">0.24</td><td align="char">0.26</td><td align="char">0.33</td><td align="char">0.29</td><td align="char"><underline><bold>0.64</bold></underline></td></tr><tr><td>Vitality</td><td align="char">0.44</td><td align="char">0.35</td><td align="char"><underline>0.50</underline></td><td align="char">0.42</td><td align="char">0.42</td></tr><tr><td colspan="6">&#x02028;&#x02028;</td></tr><tr><td colspan="6"><italic>Chronic pain without NC, n&#x000a0;=&#x000a0;1611</italic></td></tr><tr><td>Physical functioning</td><td align="char"><bold><underline>0.64</underline></bold></td><td align="char">0.40</td><td align="char"><bold><underline>0.65</underline></bold></td><td align="char">0.46</td><td align="char">0.21</td></tr><tr><td>Role limitations</td><td align="char">0.59</td><td align="char">0.42</td><td align="char"><bold><underline>0.69</underline></bold></td><td align="char">0.46</td><td align="char"><underline>0.26</underline></td></tr><tr><td>Social functioning</td><td align="char">0.39</td><td align="char">0.38</td><td align="char"><underline>0.50</underline></td><td align="char">0.34</td><td align="char">0.27</td></tr><tr><td>Pain</td><td align="char">0.56</td><td align="char">0.39</td><td align="char"><bold>0.61</bold></td><td align="char"><underline>0.55</underline></td><td align="char">0.27</td></tr><tr><td>Mental health</td><td align="char">0.15</td><td align="char">0.22</td><td align="char">0.26</td><td align="char">0.19</td><td align="char"><underline><bold>0.66</bold></underline></td></tr><tr><td>Vitality</td><td align="char">0.36</td><td align="char">0.33</td><td align="char"><underline>0.46</underline></td><td align="char">0.31</td><td align="char">0.40</td></tr><tr><td colspan="6">&#x02028;&#x02028;</td></tr><tr><td colspan="6"><italic>Chronic pain with NC, n&#x000a0;=&#x000a0;399</italic></td></tr><tr><td>Physical functioning</td><td align="char"><bold><underline>0.67</underline></bold></td><td align="char">0.54</td><td align="char"><bold><underline>0.60</underline></bold></td><td align="char">0.50</td><td align="char">0.32</td></tr><tr><td>Role limitations</td><td align="char">0.60</td><td align="char">0.57</td><td align="char"><bold><underline>0.71</underline></bold></td><td align="char">0.53</td><td align="char"><underline>0.41</underline></td></tr><tr><td>Social functioning</td><td align="char">0.44</td><td align="char">0.51</td><td align="char"><underline>0.51</underline></td><td align="char">0.43</td><td align="char">0.57</td></tr><tr><td>Pain</td><td align="char">0.53</td><td align="char">0.51</td><td align="char">0.57</td><td align="char"><bold><underline>0.66</underline></bold></td><td align="char">0.39</td></tr><tr><td>Mental health</td><td align="char">0.25</td><td align="char">0.39</td><td align="char">0.37</td><td align="char">0.30</td><td align="char"><underline><bold>0.73</bold></underline></td></tr><tr><td>Vitality</td><td align="char">0.41</td><td align="char">0.43</td><td align="char"><underline>0.50</underline></td><td align="char">0.36</td><td align="char">0.43</td></tr></tbody></table><table-wrap-foot><fn id="sp0040"><p>The 5 most correlated dimensions are indicated in boldface type.</p></fn><fn id="sp0045"><p>The underlined correlations were identified as purporting to capture similar aspects of quality of life.</p></fn><fn id="sp0050"><p>NC&#x000a0;=&#x000a0;neuropathic characteristics.</p></fn></table-wrap-foot></table-wrap></p><p id="p0120">Further analysis was conducted into the responses to the pain questions in both EQ-5D and SF-6D (SF-12), where 94.3% (366 of 388) and 71.2% (282 of 396) of those with chronic pain with NC reported moderate to extreme pain, respectively, compared with 80.7% (1278 of 1584) and 39.2% (629 of 1605) of those with chronic pain without NC.</p></sec><sec id="s0075"><label>3.3</label><title>Range of the health utility index scores</title><p id="p0125">The mean EQ-5D and the SF-6D scores are shown in <xref rid="t0015" ref-type="table">Table 3</xref>. For the whole sample, the EQ-5D mean score was greater than the SF-6D by 0.02; however, for the 2 pain groups the mean SF-6D scores were greater, by 0.03 in the chronic pain without NC group, and by 0.15 in the chronic pain with NC. Overall, median scores were higher than mean, indicating a skewed distribution for both indices, except for the SF-6D index for the chronic pain with NC group (mean 0.62, median 0.60). Floor effects were small in both of the measures, and we did not reach the absolute theoretical floor effect of EQ-5D (&#x02212;0.594), with the lowest score found to be &#x02212;0.371.<table-wrap position="float" id="t0015"><label>Table 3</label><caption><p>Distribution of EQ-5D and SF-6D indices.</p></caption><table frame="hsides" rules="groups"><thead><tr><th/><th>No. of items</th><th>Theoretical range</th><th>Observed range</th><th>Floor effect,<xref rid="tblfn1" ref-type="table-fn">&#x0204e;</xref> n (%)</th><th>Ceiling effect,<xref rid="tblfn2" ref-type="table-fn">&#x02020;</xref> n (%)</th><th>Mean (SD)</th><th>Median (IQR)</th><th>ICC of indices</th></tr></thead><tbody><tr><td colspan="9"><italic>Whole sample, n&#x000a0;=&#x000a0;4541</italic></td></tr><tr><td>EQ-5D, n&#x000a0;=&#x000a0;4349</td><td align="char">5</td><td>&#x02212;0.594 to 1</td><td>&#x02212;0.371 to 1.00</td><td>119 (2.7)</td><td>1850 (42.5)</td><td>0.79 (0.27)</td><td>0.85 (0.73 to 1.00)</td><td align="char" rowspan="2">0.61</td></tr><tr><td>SF-6D, n&#x000a0;=&#x000a0;4176</td><td align="char">6</td><td>0.29 to 1</td><td>0.345 to 1.00</td><td>9 (0.2)</td><td>174 (4.2)</td><td>0.77 (0.15)</td><td>0.80 (0.66 to 0.92)</td></tr><tr><td colspan="9">&#x02028;&#x02028;</td></tr><tr><td colspan="9"><italic>Chronic pain without NC, n&#x000a0;=&#x000a0;1611</italic></td></tr><tr><td>EQ-5D, n&#x000a0;=&#x000a0;1551</td><td align="char">5</td><td>&#x02212;0.594 to 1</td><td>&#x02212;0.235 to 1.00</td><td>54 (3.8)</td><td>239 (15.4)</td><td>0.70 (0.25)</td><td>0.76 (0.69 to 0.80)</td><td align="char" rowspan="2">0.57</td></tr><tr><td>SF-6D, n&#x000a0;=&#x000a0;1499</td><td align="char">6</td><td>0.29 to 1</td><td>0.345 to 1.00</td><td>1 (0.1)</td><td>18 (1.2)</td><td>0.73 (0.14)</td><td>0.72 (0.62 to 0.86)</td></tr><tr><td colspan="9">&#x02028;&#x02028;</td></tr><tr><td colspan="9"><italic>Chronic pain with NC, n&#x000a0;=&#x000a0;399</italic></td></tr><tr><td>EQ-5D, n&#x000a0;=&#x000a0;373</td><td align="char">5</td><td>&#x02212;0.594 to 1</td><td>&#x02212;0.371 to 1.00</td><td>64 (17.2)</td><td>14 (3.8)</td><td>0.47 (0.36)</td><td>0.62 (0.89 to 0.73)</td><td align="char" rowspan="2">0.44</td></tr><tr><td>SF-6D, n&#x000a0;=&#x000a0;358</td><td align="char">6</td><td>0.29 to 1</td><td>0.345 to 1.00</td><td>5 (1.4)</td><td>3 (0.8)</td><td>0.62 (0.15)</td><td>0.60 (0.52 to 0.71)</td></tr></tbody></table><table-wrap-foot><fn><p>IQR&#x000a0;=&#x000a0;interquartile range; ICC&#x000a0;=&#x000a0;intraclass correlation coefficient, NC&#x000a0;=&#x000a0;neuropathic characteristics.</p></fn></table-wrap-foot><table-wrap-foot><fn id="tblfn1"><label>&#x0204e;</label><p id="np005">Floor effect, less than 0 (worse than death) for EQ-5D and minimum value for SF-6D&#x000a0;=&#x000a0;0.345.</p></fn></table-wrap-foot><table-wrap-foot><fn id="tblfn2"><label>&#x02020;</label><p id="np010">Ceiling effect&#x000a0;=&#x000a0;1 for both instruments.</p></fn></table-wrap-foot></table-wrap></p><p id="p0130">We observed 119 respondents with a score &#x0003c;0 or WTD on the EQ-5D (2.7% of the whole sample). Only 1 of these individuals reported no chronic pain, 54 individuals reported chronic pain without NC (from total n&#x000a0;=&#x000a0;1551), representing 45.4% of all WTD scores and 3.4% of chronic pain without NC. The remaining 64 individuals with WTD utility scores reported chronic pain with NC, representing 53.8% of all WTD scores and 17.2% of the chronic pain with NC group (n&#x000a0;=&#x000a0;64 of 373 with complete data). The majority of respondents who were WTD in both pain groups reported severe average pain (&#x02a7e;7 of 10), 98.1% (n&#x000a0;=&#x000a0;53 of 54) of the chronic pain without NC and 83.6% (n&#x000a0;=&#x000a0;51 of 1) of chronic pain with NC (&#x003c7;² test, <italic>P</italic>&#x000a0;=&#x000a0;.03). The lowest score observed for SF-6D was 0.345 in 1.4% (n&#x000a0;=&#x000a0;5 of 358) of those with chronic pain with NC, none of the respondents reached the theoretical lowest threshold of 0.29.</p><p id="p0135">The EQ-5D scores were negatively skewed with a ceiling effect that was most apparent in the whole general population sample, where 42.5% (n&#x000a0;=&#x000a0;1850) attained a maximum score of 1 (<xref rid="t0015" ref-type="table">Table 3</xref>). The highest possible score on EQ-5D was also found in 15.4% of those with chronic pain without NC and in 3.8% of the chronic pain with NC group, and there were few scores between 0.88 and 1. This compares to a ceiling effect in the SF-6D found in 4.2%, 1.2%, and 0.8% in the whole sample, and the chronic pain without and with NC groups, respectively. The SF-6D is also negatively skewed in the whole sample, but then these scores approximate a normal distribution in the 2 pain groups because the SF-6D was more positively skewed and is limited at the floor (0.29). The distribution of EQ-5D scores was distributed across the range of scores in the chronic pain groups. It also appears that the SF-6D was more continuous, whereas the EQ-5D appeared more discrete with gaps between states, with an apparent bimodal distribution in both the chronic pain groups. The frequency distributions of the utility scores for the whole sample and the 2 pain groups are shown in <xref rid="f0005" ref-type="fig">Fig. 1</xref>.<fig id="f0005"><label>Fig. 1</label><caption><p>Frequency distributions by pain group.</p></caption><graphic xlink:href="gr1"/></fig></p></sec><sec id="s0080"><label>3.4</label><title>Discriminating health utility scores by level of pain severity</title><p id="p0140"><xref rid="t0020" ref-type="table">Table 4</xref> shows the mean EQ-5D and SF-6D utilities by mild, moderate, and severe average pain. Of the total 2010 respondents with chronic pain, 1972 completed the average pain NRS, 22.6% (n&#x000a0;=&#x000a0;445) reporting mild chronic pain, 41.7% (n&#x000a0;=&#x000a0;822) reporting moderate chronic pain, and 35.8% (n&#x000a0;=&#x000a0;705) reporting severe chronic pain. Overall, those with severe pain had the lowest utilities for both EQ-5D and SF-6D as expected; furthermore, the EQ-5D utilities were lower in chronic pain (with and without NC) compared with the SF-6D. In the chronic pain with NC group, mean EQ-5D health utilities for those reporting mild pain intensity was 0.72, for those with moderate pain the mean was 0.63, and for those with severe pain it was 0.33. Therefore, the mean between-group differences (for pain intensity) were 0.09 (moderate-mild) and 0.39 (severe-moderate), and above the published mean MID of 0.074 <xref rid="b0260" ref-type="bibr">[52]</xref>. For SF-6D utilities, in the chronic pain with NC group, the mean difference in health utilities between those with moderate and mild pain was 0.08, with the same mean difference between severe and moderate pain intensity, and also above the mean MID of 0.041 <xref rid="b0260" ref-type="bibr">[52]</xref>. Notably, in comparing the EQ-5D and SF-6D, utility scores as measured by the EQ-5D were considerably lower in those with any severe pain (mean 0.48 vs 0.63), and the EQ-5D showed approximately twice the range of between-group differences (0.55 vs 0.33) compared with the SF-6D scores (0.65 vs 0.58).<table-wrap position="float" id="t0020"><label>Table 4</label><caption><p>Health utilities by pain severity, mean (SD).</p></caption><table frame="hsides" rules="groups"><thead><tr><th/><th>Mild (n&#x000a0;=&#x000a0;445)</th><th>Moderate (n&#x000a0;=&#x000a0;822)</th><th>Severe (n&#x000a0;=&#x000a0;705)</th><th>Mean difference mild-moderate</th><th>Mean difference moderate-severe</th><th>Mean difference mild-severe</th><th><italic>P</italic> value<xref rid="tblfn3" ref-type="table-fn">&#x0204e;</xref></th></tr></thead><tbody><tr><td colspan="8"><italic>Any chronic pain</italic></td></tr><tr><td>EQ-5D</td><td>0.82 (0.17)</td><td>0.72 (0.20)</td><td>0.48 (0.35)</td><td align="char">0.10</td><td align="char">0.24</td><td align="char">0.34</td><td align="char">&#x0003c;.001</td></tr><tr><td>SF-6D</td><td>0.79 (0.12)</td><td>0.73 (0.13)</td><td>0.63 (0.15)</td><td align="char">0.06</td><td align="char">0.10</td><td align="char">0.16</td><td align="char">&#x0003c;.001</td></tr><tr><td colspan="8">&#x02028;&#x02028;</td></tr><tr><td colspan="8"><italic>Chronic pain without NC</italic></td></tr><tr><td>EQ-5D</td><td>0.83 (0.16)</td><td>0.74 (0.17)</td><td>0.55 (0.32)</td><td align="char">0.09</td><td align="char">0.19</td><td align="char">0.28</td><td align="char">&#x0003c;.001</td></tr><tr><td>SF-6D</td><td>0.79 (0.11)</td><td>0.74 (0.13)</td><td>0.65 (0.15)</td><td align="char">0.05</td><td align="char">0.09</td><td align="char">0.14</td><td align="char">&#x0003c;.001</td></tr><tr><td colspan="8">&#x02028;&#x02028;</td></tr><tr><td colspan="8"><italic>Chronic pain with NC</italic></td></tr><tr><td>EQ-5D</td><td>0.72 (0.24)</td><td>0.63 (0.27)</td><td>0.33 (0.36)</td><td align="char">0.09</td><td align="char">0.30</td><td align="char">0.39</td><td align="char">&#x0003c;.001</td></tr><tr><td>SF-6D</td><td>0.74 (0.13)</td><td>0.66 (0.14)</td><td>0.58 (0.14)</td><td align="char">0.08</td><td align="char">0.08</td><td align="char">0.16</td><td align="char">&#x0003c;.001</td></tr></tbody></table><table-wrap-foot><fn id="sp0070"><p>Mean Minimally Important Difference (MID) for EQ-5D 0.074 and mean MID for SF-6D 0.041 (Walters &#x00026; Brazier, 2005 <xref rid="b0260" ref-type="bibr">[52]</xref>).</p></fn><fn id="sp0075"><p>NC&#x000a0;=&#x000a0;neuropathic characteristics.</p></fn></table-wrap-foot><table-wrap-foot><fn id="tblfn3"><label>&#x0204e;</label><p id="np015">ANOVA for mild, moderate and severe pain by pain group.</p></fn></table-wrap-foot></table-wrap></p></sec></sec><sec id="s0085"><label>4</label><title>Discussion</title><p id="p0145">This study compared the health utility scores derived from 2 widely used generic HRQoL measures, the EQ-5D and SF-6D, in a large general UK population sample focusing on chronic pain with and without neuropathic characteristics. We looked at the agreement between instruments in measuring individual health dimensions, the scoring range of the health utility scores, and whether scores could detect at least an MID between clinically meaningful differences in pain. Both instruments were able to discriminate important differences in pain groups and in pain intensity. Among those with chronic neuropathic pain, 17% had HRQoL scores equivalent to WTD on the EQ-5D.</p><sec id="s0090"><label>4.1</label><title>Key findings and implications</title><sec id="s0095"><label>4.1.1</label><title>Comparison between individual dimensions</title><p id="p0150">Overall, we found only moderate agreement between individual dimensions. The most highly correlated dimensions were mental health and anxiety/depression and role limitations and usual activities. The pain and pain/discomfort dimensions were more highly correlated in the chronic pain with NC group (compared to without NC).</p></sec><sec id="s0100"><label>4.1.2</label><title>Health utility scores: range, floor, and ceiling effects</title><p id="p0155">We confirmed the considerable ceiling effect of EQ-5D observed in previous general population surveys <xref rid="b0025" ref-type="bibr">[5]</xref>, <xref rid="b0070" ref-type="bibr">[14]</xref>, <xref rid="b0190" ref-type="bibr">[38]</xref>, with 43% of our whole sample reporting full health, compared with only 4.2% who were classified in full health on the SF-6D. The EQ-5D appears insensitive at the top (healthy end) of the scale, and a gap exists between 0.88 and 1. The SF-6D does not seem to have a ceiling effect and may capture smaller health changes toward the top of the scale <xref rid="b0055" ref-type="bibr">[11]</xref>. However, this may be less relevant in chronic neuropathic pain, in which the proportion of respondents attaining a maximum utility score with EQ-5D was relatively small (only 3.8% reported full health with EQ-5D and 0.8% with SF-6D).</p><p id="p0160">In total, 17% of chronic pain with NC and 3% of chronic pain without NC respondents had a score of below 0 or WTD on EQ-5D. Almost all of those with a WTD score also reported severe pain (&#x02a7e;7 of 10). Other studies of rheumatoid arthritis (RA) report similar findings: in a study of patients with established RA <xref rid="b0005" ref-type="bibr">[1]</xref>, 17% had WTD scores at baseline (before biological therapy) and 7% at 12-month follow-up. Extreme pain scores were strongly associated with a state WTD in a study of early arthritis, in which 11% had a negative EQ-5D score <xref rid="b0105" ref-type="bibr">[21]</xref>, and in patients with RA <xref rid="b0110" ref-type="bibr">[22]</xref>, 9% of trial participants had states WTD. In these studies, extreme pain/discomfort was the key EQ-5D domain associated with a WTD state, plus moderate problems in &#x02a7e;3 other domains <xref rid="b0110" ref-type="bibr">[22]</xref>. Whitehurst et al. (2011) <xref rid="b0280" ref-type="bibr">[56]</xref> suggest that the EQ-5D may be better suited to capture the magnitude of severity for poorer health states. Notably, however, the 20% of patients who on the EQ-5D had a score &#x02a7d;0 did not actually reach the SF-6D floor of 0.29. This raises interesting issues regarding the true HRQoL state of such patients. For instance, if such states are really considered as WTD, as estimated by EQ-5D, it may be legitimately expected that the utility scores of SF-6D in these patients would have clustered at the SF-6D floor of 0.29. However, the mean score of patients with chronic pain with neuropathic characteristics was 0.34.</p></sec><sec id="s0105"><label>4.1.3</label><title>Mean scores between patient groups</title><p id="p0165">There were differences in the mean utility scores for the 2 instruments, with EQ-5D utility scores higher in the whole sample and the no chronic pain groups, whereas SF-6D scores were higher in both chronic pain groups. This is most striking in chronic pain with NC, in which the mean SF-6D scores were 0.15 higher. Other studies have found the average differences in means to be around 0.05 <xref rid="b0040" ref-type="bibr">[8]</xref>, <xref rid="b0190" ref-type="bibr">[38]</xref> although this varies, with higher mean differences (0.15) reported in studies of severe pain conditions such as severe knee osteoarthritis <xref rid="b0185" ref-type="bibr">[37]</xref>, <xref rid="b0285" ref-type="bibr">[57]</xref> and inflammatory RA <xref rid="b0115" ref-type="bibr">[23]</xref>. These results suggest that it may be unreliable, perhaps even invalid, to compare studies of severe pain-related conditions that have used different health utility measures.</p></sec><sec id="s0110"><label>4.1.4</label><title>Inferring the potential sensitivity to detect a change in utility</title><p id="p0170">In an exploratory exercise, we attempted to use these cross-sectional data to infer the potential sensitivity of the 2 instruments if there was an intervention that led to a clinically meaningful reduction in pain. Respondents were classified according to clinically validated cut-points for mild, moderate, and severe chronic pain <xref rid="b0295" ref-type="bibr">[59]</xref>. It has been suggested that a clinically important outcome would be to reduce a patient&#x02019;s level of pain down to no worse than mild <xref rid="b0170" ref-type="bibr">[34]</xref>. We estimated the difference in mean utility scores between different pain severity groups and compared this to the MID <xref rid="b0260" ref-type="bibr">[52]</xref>. We found the mean between-group differences were above the MID for both scores. However, the differences were higher using the EQ-5D. In particular, the difference in utility between moderate-severe pain was 5-fold, 0.39 using EQ-5D and 0.08 using SF-6D. If such findings were to be found in a randomized trial, in which a patient&#x02019;s pain severity reduced from severe to moderate, this would have substantial implications regarding the cost effectiveness (cost-utility analysis) of the intervention. This analysis points to the need for further research using trial and longitudinal data. This is discussed further in Section <xref rid="s0125" ref-type="sec">4.4</xref>. Other studies that compared the utility estimates of EQ-5D and SF-6D in randomized trials found the choice between the instruments to be very important regarding the cost-utility estimates produced <xref rid="b0040" ref-type="bibr">[8]</xref>, <xref rid="b0130" ref-type="bibr">[26]</xref>, <xref rid="b0200" ref-type="bibr">[40]</xref>.</p></sec></sec><sec id="s0115"><label>4.2</label><title>Study limitations</title><p id="p0175">This study comprises a large dataset derived from a random sample of adults generated from GP practices in the UK. The relatively low response rate is similar to previous surveys of pain prevalence <xref rid="b0050" ref-type="bibr">[10]</xref>, <xref rid="b0160" ref-type="bibr">[32]</xref> and is an increasingly common problem in epidemiological research <xref rid="b0100" ref-type="bibr">[20]</xref>, <xref rid="b0140" ref-type="bibr">[28]</xref>, <xref rid="b0160" ref-type="bibr">[32]</xref>, <xref rid="b0250" ref-type="bibr">[50]</xref>. However, the prevalence of chronic pain and chronic pain with NC was similar to that in other population studies with higher response rates <xref rid="b0235" ref-type="bibr">[47]</xref>, <xref rid="b0290" ref-type="bibr">[58]</xref>. In relation to possible response bias and practicality, completion rates were high for both the EQ-5D and SF-12, with slightly higher response rate in favor of the EQ-5D. Similar differences in completion rates have been reported elsewhere <xref rid="b0015" ref-type="bibr">[3]</xref>, <xref rid="b0275" ref-type="bibr">[55]</xref>.</p></sec><sec id="s0120"><label>4.3</label><title>Future research directions</title><p id="p0180">Researchers and clinicians should consider using generic health utility instruments in pain-related burden of disease studies. The rationale for generating health utility scores is that it provides a generic, preference-weighted index that enables the severity of different conditions to be estimated consistently. The ultimate intention is to assist health care planners to allocate resources on a consistent and transparent basis among different diseases and interventions.</p><p id="p0185">There are a number of utility instruments <xref rid="b0045" ref-type="bibr">[9]</xref>, <xref rid="b0090" ref-type="bibr">[18]</xref>, <xref rid="b0095" ref-type="bibr">[19]</xref> commonly used in clinical practice and research. The findings from this study, that the choice between EQ-5D and SF-6D results in major differences in the estimation of utility scores for severe chronic (and neuropathic) pain, questions the validity of comparing studies that have used different instruments. This discordance warrants further investigation in other pain populations.</p><p id="p0190">Our exploratory analysis, using cross-sectional data, compared the mean utility scores between the instruments for patients classified by different cut-points for pain severity. The difference between the mean utility scores of patients with severe and moderate pain was 5 times greater when estimated using EQ-5D compared with SF-6D. Therefore, the choice between using the EQ-5D and SF-6D not only may be important in estimating the absolute burden of disease, but also may have major implications regarding the economic evaluation of interventions that take a cost-utility approach. It is important to compare these instruments in longitudinal studies to assess their sensitivity to detect changes in pain severity.</p><p id="p0195">Finally, it is important to note that cost utility analysis (cost effectiveness analysis) uses measures of relative change, in which improvement in health utilities scores are valued equally, irrespective of the level of utilities postintervention <xref rid="b0175" ref-type="bibr">[35]</xref>. This potentially raises concerns if an intervention involving patients with WTD scores (as measured by EQ-5D) improves health utilities but the patients remain in a state WTD. The National Institute for Health and Clinical Excellence recommends the use of general population preferences when generating utility index scores <xref rid="b0175" ref-type="bibr">[35]</xref>, meaning that it is not actually the patients themselves who would prefer death but the general population who score these health states. As people with chronic pain come to accept and adjust over time <xref rid="b0225" ref-type="bibr">[45]</xref>, the meaning and interpretation of WTD seems to raise both ethical and practical considerations regarding measuring burden of disease and in the assessment of utility.</p><p id="p0200">Overall, the rationale for attempting to generate utility index scores to generate a consistent HRQoL outcome measure is important. However, there are a number of practical issues that this study has raised, in common with other research <xref rid="b0040" ref-type="bibr">[8]</xref>, <xref rid="b0130" ref-type="bibr">[26]</xref>, <xref rid="b0200" ref-type="bibr">[40]</xref>.</p></sec><sec id="s0125"><label>4.4</label><title>Conclusions</title><p id="p0205">The measurement of HRQOL is important in chronic pain research, and health utilities derived from generic instruments such as EQ-5D and SF-12/36 can discriminate among group differences for chronic pain with and without NC and greater pain severity. This study demonstrates the substantial lack of agreement between EQ-5D and SF-6D when estimating the burden of disease for severe chronic pain. Future research should include longitudinal and clinical studies to test the validity of utility scores to understand the true health state of patients and also to assess the sensitivity of scores to detect changes in HRQoL as individuals&#x02019; pain severity ratings change. The choice between the instruments has substantial implications regarding the estimation of HRQoL in chronic pain patients.</p></sec></sec><sec id="s0130"><title>Conflict of interest statement</title><p id="p0210">N.T., K.L., E.A., and K.D. have no conflicts of interest. B.S. has received occasional lecture and consultancy fees, on behalf of his institution, from companies involved in the manufacture of drugs used in treating neuropathic pain. M.S has received research support, consulting fees, or honoraria in the past 3&#x000a0;years from Astellas, Astra Zenica, Gr&#x000fc;nenthal, GW Pharmaceuticals, Lilly, NAPP, and Pfizer. M.B. has received consultancy fees and lecturer honoraria from Pfizer, Astellas, and Grunenthal in the last 3&#x000a0;years. Kate Dunn is supported by the <funding-source id="gp005"><institution-wrap><institution-id institution-id-type="doi">10.13039/100004440</institution-id><institution>Wellcome Trust</institution></institution-wrap></funding-source> [083572]. The authors assert no personal pecuniary or other conflict of interest in the writing of this article. No writing assistance was used in the production of this article.</p></sec>

Differential structural and resting state connectivity between insular subdivisions and other pain-related brain regions

Could not extract abstract

<contrib contrib-type="author" id="au005"><name><surname>Wiech</surname><given-names>K.</given-names></name><email>katja.wiech@ndcn.ox.ac.uk</email><xref rid="af005" ref-type="aff">a</xref><xref rid="af010" ref-type="aff">b</xref><xref rid="cor1" ref-type="corresp">&#x0204e;</xref></contrib><contrib contrib-type="author" id="au010"><name><surname>Jbabdi</surname><given-names>S.</given-names></name><xref rid="af005" ref-type="aff">a</xref></contrib><contrib contrib-type="author" id="au015"><name><surname>Lin</surname><given-names>C.S.</given-names></name><xref rid="af005" ref-type="aff">a</xref><xref rid="af010" ref-type="aff">b</xref></contrib><contrib contrib-type="author" id="au020"><name><surname>Andersson</surname><given-names>J.</given-names></name><xref rid="af005" ref-type="aff">a</xref></contrib><contrib contrib-type="author" id="au025"><name><surname>Tracey</surname><given-names>I.</given-names></name><xref rid="af005" ref-type="aff">a</xref><xref rid="af010" ref-type="aff">b</xref></contrib><aff id="af005"><label>a</label>Oxford Centre for Functional Magnetic Resonance Imaging of the Brain, University of Oxford, John Radcliffe Hospital, Headley Way, Oxford OX3 9DU, England, UK</aff><aff id="af010"><label>b</label>Nuffield Department of Clinical Neurosciences, Nuffield Division Anaesthetics, University of Oxford, John Radcliffe Hospital, Headley Way, Oxford OX3 9DU, England, UK</aff>

Pain

<sec id="s0005"><label>1</label><title>Introduction</title><p id="p0005">Of the various brain regions that have been implicated in the perception of pain, the insula has most consistently been reported across studies <xref rid="b0085" ref-type="bibr">[17]</xref>. Insular activation has been found during acute <xref rid="b0045" ref-type="bibr">[9]</xref>, <xref rid="b0130" ref-type="bibr">[26]</xref>, <xref rid="b0205" ref-type="bibr">[41]</xref>, <xref rid="b0310" ref-type="bibr">[62]</xref> and chronic pain <xref rid="b0235" ref-type="bibr">[47]</xref>. Although often referred to as the insula or the insular cortex, studies in primates indicate that this structure consists of an anterior, mid, and posterior part, which differ considerably in anatomy and cytoarchitecture <xref rid="b0025" ref-type="bibr">[5]</xref>. In accordance with these findings, functional neuroimaging studies on pain (and other sensory experiences) in humans suggest a functional segregation (see <xref rid="b0070" ref-type="bibr">[14]</xref> for review). Activity in the posterior (PI) and mid insula (MI) predominantly reflects sensory aspects of pain <xref rid="b0005" ref-type="bibr">[1]</xref>, <xref rid="b0175" ref-type="bibr">[35]</xref>, <xref rid="b0220" ref-type="bibr">[44]</xref>. The anterior insula (AI), in contrast, has been associated with the cognitive&#x02013;affective dimension of pain that is sensitive to contextual manipulations <xref rid="b0200" ref-type="bibr">[40]</xref>, <xref rid="b0310" ref-type="bibr">[62]</xref>.</p><p id="p0010">Recent findings from functional connectivity studies support the notion of a functional segregation. During noxious stimulation, PI exhibits increased functional connectivity with the primary somatosensory cortex <xref rid="b0180" ref-type="bibr">[36]</xref> that is pivotal for the processing of sensory-discriminative information. AI, in contrast, showed increased connectivity with the prefrontal and anterior cingulate cortex <xref rid="b0180" ref-type="bibr">[36]</xref> known to be involved in evaluative processing. Moreover, increased functional coupling between the left AI and the midcingulate cortex that has also been implicated in emotional salience monitoring <xref rid="b0240" ref-type="bibr">[48]</xref>, <xref rid="b0280" ref-type="bibr">[56]</xref> precedes the perception of ambiguous stimuli (ie, stimuli of an intensity calibrated at the pain detection threshold) as painful <xref rid="b0310" ref-type="bibr">[62]</xref>.</p><p id="p0015">Tracer studies in primates have shown direct white matter connections between the AI and prefrontal regions whereas the MI/PI is most strongly connected to somatosensory regions <xref rid="b0025" ref-type="bibr">[5]</xref>, <xref rid="b0145" ref-type="bibr">[29]</xref>, <xref rid="b0155" ref-type="bibr">[31]</xref>. So far, explorative diffusion magnetic resonance imaging (MRI) tractography studies in humans have confirmed the dissociation between AI and MI/PI connections to prefrontal cortex and postcentral gyrus <xref rid="b0060" ref-type="bibr">[12]</xref>, <xref rid="b0065" ref-type="bibr">[13]</xref>. However, a more comprehensive analysis on connections with the different brain regions that have been implicated in pain processing is still missing.</p><p id="p0020">Here, we compared both structural and resting state connectivity of the 3 insular subdivisions with pain-related brain regions in healthy volunteers. More specifically, we used probabilistic tracto-graphy <xref rid="b0040" ref-type="bibr">[8]</xref> to (1) identify pain-related brain regions that are predominantly connected to the 3 subdivisions and (2) identify the main insular connection target for each pain-related region. Furthermore, we (3) searched for the pain-related brain regions that showed the strongest resting state connectivity for each insular subdivision and (4) identified the part of the insula each pain-related region exhibited the strongest resting state connectivity with.</p></sec><sec id="s0010"><label>2</label><title>Methods</title><p id="p0025">The study focuses on 2 different aspects of the interaction of the 3 insular subdivisions with other pain-related brain regions: structural connectivity and resting state connectivity. Because the 2 aspects were investigated in different samples and required different analyses, they are described separately below.</p><sec id="s0015"><label>2.1</label><title>Structural connectivity (probabilistic tractography)</title><p id="p0030"><italic>Participants.</italic> Sixteen healthy right-handed volunteers participated in the study. Data from 1 subject suffered technical failure and were consequently excluded; therefore, data from 15 subjects were entered into the analysis (mean&#x000a0;&#x000b1;&#x000a0;SD age 20.6&#x000a0;&#x000b1;&#x000a0;3.6, 11 women). All participants gave written informed consent. The study was approved by the Oxford Research Ethics Committee.</p><p id="p0035"><italic>Diffusion imaging and preprocessing.</italic> The diffusion-weighted images were acquired using echo planar imaging (EPI) with a voxel size of 2&#x000a0;&#x000d7;&#x000a0;2&#x000a0;&#x000d7;&#x000a0;2&#x000a0;mm. The diffusion weighting was isotropically distributed along 60 directions using a b-value of 1500&#x000a0;s&#x000a0;mm^&#x02212;2. Two sets of diffusion-weighted data sets were acquired in total for subsequent averaging, to improve the signal-to-noise ratio. A dual-echo blip-reversed sequence was used to reduce EPI distortions <xref rid="b0020" ref-type="bibr">[4]</xref>. In brief, for each set, non-diffusion-weighted (b0) images with opposite phase-encode directions (blip up/down) were combined to estimate field inhomogeneities using the TOPUP tool (part of FSL5). Then the 2 averages were corrected for EPI distortions (using the calculated field), as well as eddy currents and motion, using the EDDY_CORRECT tool in FSL. The images obtained using this method had virtually no residual EPI-induced distortions, which generally are particularly pronounced in the frontal lobes.</p><p id="p0040"><italic>Probabilistic tractography.</italic> Probabilistic tractography was performed using the FMRIB&#x02019;s Diffusion Toolbox (FDT, <ext-link ext-link-type="uri" xlink:href="http://fsl.fmrib.ox.ac.uk/fsl/fdt" id="ir0005">http://fsl.fmrib.ox.ac.uk/fsl/fdt</ext-link>). We first used FLIRT <xref rid="b0110" ref-type="bibr">[22]</xref>, <xref rid="b0115" ref-type="bibr">[23]</xref> for linear image transformation, then FNIRT <xref rid="b0010" ref-type="bibr">[2]</xref>, <xref rid="b0015" ref-type="bibr">[3]</xref> for nonlinear image transformation between the diffusion and the individual structural space and between the individual structural and MNI152 standard space. This transformation was used throughout all the tractography analyses, and tractography results were directly resampled in standard space. We performed probabilistic tracking to calculate the number of streamline samples starting from any location of the brain passing through a combined region of interest mask, which consists of a seed mask and a target mask. Therefore, any found tract was constrained by the seed mask at one end and the target mask at the other. This way of sampling has been shown to increase robustness in identifying long connections and connections that traverse regions with multiple crossing fibers, such as the connections between the insular and cingulate cortex <xref rid="b0255" ref-type="bibr">[51]</xref>. Note that this approach differs from those starting from voxels within the seed mask only but uses a standard algorithm as described in <xref rid="b0035" ref-type="bibr">[7]</xref>. To quantify the uncertainty of tracking results, we drew 100 sample streamlines for each voxel in the brain. The output file was split into a seed-to-target map, which revealed how many streamline samples can successfully reach the target region from each voxel in the seed region. For each insular subdivision, 24 tractography analyses were performed for the 12 targets, separately in both hemispheres. Only ipsilateral connections (eg, between the right AI to the right VLPFC) were analyzed. To exclude indirect tracts (ie, tracts that pass through more than one pain-related region before reaching the target region), we used a stop mask at the target region and used the other targets as exclusion masks.</p><p id="p0045"><italic>Psychological assessment.</italic> All participants filled in the following questionnaires to assess pain-related psychological traits: (1) Center for Epidemiological Studies Depression Scale (CES-D) <xref rid="b0215" ref-type="bibr">[43]</xref>, (2) State and Trait Anxiety Inventory <xref rid="b0260" ref-type="bibr">[52]</xref>, (3) Pain Vigilance and Awareness Questionnaire (PVAQ) <xref rid="b0135" ref-type="bibr">[27]</xref>, and (4) the Pain Catastrophizing Scale <xref rid="b0270" ref-type="bibr">[54]</xref>. All questionnaires were analyzed according to their respective manual.</p></sec><sec id="s0020"><label>2.2</label><title>Resting state connectivity</title><p id="p0050">We used resting state data from a cohort of healthy subjects that participated in previous studies conducted in our laboratory <xref rid="b0250" ref-type="bibr">[50]</xref>. Note that this sample did not overlap with the sample in which the structural connectivity was investigated.</p><p id="p0055"><italic>Participants.</italic> Resting state data sets were acquired from 36 healthy adult subjects (age range, 20&#x02013;35&#x000a0;years; mean, 28.5&#x000a0;years; 15 women). The study was approved by the institutional review board and subjects provided informed consent.</p><p id="p0060"><italic>Data acquisition.</italic> Resting fMRI blood oxygenation level dependent (BOLD) data were acquired on a 3&#x000a0;T Siemens Trio MRI scanner, using a 12-channel head coil with a standard gradient echo echo-planar-imaging (EPI) acquisition, TR&#x000a0;=&#x000a0;2&#x000a0;s, TE&#x000a0;=&#x000a0;28&#x000a0;ms, flip angle&#x000a0;=&#x000a0;89&#x000b0;, resolution 3&#x000a0;&#x000d7;&#x000a0;3&#x000a0;&#x000d7;&#x000a0;3.5&#x000a0;mm, whole-head coverage except for the lowest parts of the cerebellum in some subjects. The resting fMRI scan lasted 6&#x000a0;min. Ambient light was minimized, and the subjects were instructed to lie with eyes open, think of nothing in particular, and not to fall asleep. In order to aid the registration of the functional data into a common standard brain coordinate system (MNI152), structural brain images were acquired by using a T1-weighted 3--D MPRAGE sequence with whole-head coverage, TR&#x000a0;=&#x000a0;2.04&#x000a0;s, TE&#x000a0;=&#x000a0;4.7&#x000a0;ms, flip angle&#x000a0;= 8&#x000b0; resolution 1&#x000a0;&#x000d7;&#x000a0;1&#x000a0;&#x000d7;&#x000a0;1&#x000a0;mm, total acquisition time 12&#x000a0;min.</p></sec><sec id="s0025"><label>2.3</label><title>Preparation of masks</title><p id="p0065">The same set of anatomically defined masks was used in the analysis of structural, and resting state connectivity. Thirty masks of 3 bilateral seed (ie, AI, MI, and PI) and 12 pain-related target regions of interest were defined according to published atlases and manual drawings based on individual T1-weighted images (<xref rid="f0005" ref-type="fig">Fig. 1</xref>). Pain-related target regions comprised the thalamus (THA), primary and secondary somatosensory cortices (SI and SII), the subgenual, dorsal, and rostral anterior cingulate cortex (sgACC, dACC, and rACC), posterior midcingulate cortex (pMCC), dorsolateral and ventrolateral prefrontal cortex (DLPFC and VLPFC), amygdala (AMYG), orbitofrontal cortex (OFC), and periaqueductal grey (PAG). The THA and the AMYG were selected from the Harvard-Oxford Subcortical Atlas (Harvard Center for Morphometric Analysis), thresholded at 25%. SI and SII were selected from the Juelich Histological Atlas (The Research Center Juelich), thresholded at 25%. The masks of the sgACC, dACC, rACC, pMCC, DLPFC, VLPFC, AMYG, OFC, PAG, and the 3 subdivisions of the insula were defined on each subject&#x02019;s T1-weighted image and hand drawn using fslview (in FSL, FMRIB Software Library, <ext-link ext-link-type="uri" xlink:href="http://www.fmrib.ox.ac.uk/fsl" id="ir0010">http://www.fmrib.ox.ac.uk/fsl</ext-link>) starting in one view, eg, coronal view, and double-checked using multiplanar views.<fig id="f0005"><label>Fig. 1</label><caption><p>Seed and target regions of interest for probabilistic tractography. Seeds comprise 3 insular subdivisions: anterior (AI), mid (MI), and posterior (PI). Targets comprise 12 pain-related targets: thalamus (THA), primary and secondary somatosensory cortex (SI and SII), dorsal and rostral anterior cingulate cortex (dACC and rACC), dorsolateral and ventrolateral prefrontal cortex (DLPFC and VLPFC), amygdala (AMYG), orbitofrontal cortex (OFC), periaqueductal grey (PAG), posterior midcingulate cortex (pMCC), and subgenual anterior cingulate cortex (sgACC). Note that same regions of interest were used in analyses on resting state functional connectivity.</p></caption><graphic xlink:href="gr1"/></fig></p><p id="p0070"><italic>Anterior, mid, and posterior insula.</italic> The insula was divided into anterior, mid, and posterior regions as defined by Brooks et al. <xref rid="b0045" ref-type="bibr">[9]</xref>. The mask for the AI includes tissue from the anterior gyrus brevis, the MI comprises the posterior gyrus brevis, and the PI was defined as the anterior gyrus longus.</p><p id="p0075"><italic>Dorsal and rostral anterior cingulate cortex.</italic> The mask for the dACC corresponds to the anterior midcingulate cortex (Brodmann area 24&#x02032; and 32&#x02032;), while the rACC was defined as the perigenual ACC (area 24 and 32) following the division proposed by Vogt et al. <xref rid="b0290" ref-type="bibr">[58]</xref>.</p><p id="p0080"><italic>Subgenual anterior cingulate cortex.</italic> The subgenual ACC is located underneath the genu of the corpus callosum and corresponds mainly to BA25 and caudal portions of BA32 and BA24. The mask was drawn in sagittal view, starting from the inferior border of rACC and moving posteriorly to the end of the genu.</p><p id="p0085"><italic>Dorsolateral and ventrolateral prefrontal cortex.</italic> The mask for the DLPFC comprises BA8, 9, 46, and 9/46 located in the superior and middle frontal gyrus. VLPFC was defined as Brodmann areas 44 (pars opercularis), 45 (pars triangularis), and the lateral part of area 47/12 of the inferior frontal gyrus.</p><p id="p0090"><italic>Orbitofrontal cortex.</italic> The OFC mask comprises Brodmann areas 10, 11, 12, 13, 14, and the orbital part of area 47/12. It extends from the horizontal ramus of the lateral fissure on the lateral surface to the orbital surface and onto the medial surface to include the gyrus ventral to cingulate sulcus and the subcallosal cingulate areas. The boundary on the medial surface is from the rostral sulcus to the horizontal ramus of the lateral fissure.</p><p id="p0095"><italic>Periaqueductal grey.</italic> The outline of PAG was first delineated in the sagittal view, and the mask was drawn in the axial view according to the Duvernoy&#x02019;s atlas of the human brain stem <xref rid="b0160" ref-type="bibr">[32]</xref>.</p><p id="p0100"><italic>Posterior midcingulate cortex.</italic> The pMCC is located between the corpus callosum inferiorly and the cingulate sulcus superiorly. The posterior border is the continuation of the central sulcus, while the anterior border is defined by the sulcus midway between its posterior border and the posterior border of the dorsal ACC. The mask was outlined in sagittal view and filled up in coronal view.</p></sec><sec id="s0030"><label>2.4</label><title>Data analysis</title><sec id="s0035"><label>2.4.1</label><title>Structural connectivity</title><p id="p0105"><italic>Definition of connection probability indices.</italic> In the following analyses, we refer to connection probabilities as the number of streamlines from probabilistic tractography that satisfy a given condition (eg, connect to the AI), divided by the total number of streamlines seeded that were not rejected by the exclusion criteria. This probability reflects our uncertainty on the fiber orientation measurements and the tractography process. In analyses 1 and 2, we identify the dominant connection targets of each insular subdivision (analysis 1) and the dominant insular connection target for each pain-related brain region (analysis 2), separately for both hemispheres. In these analyses, we report relative target probabilities that represent the probability that a tract reaches a particular target region, relative to all targets (in %).</p><p id="p0110"><italic>Analysis 1: Main pain-related connection targets for each insular subdivision.</italic> In order to identify the dominant connection targets for each insular subdivision, we performed 2-way ANOVAs with the factors HEMISPHERE (left, right) and TARGET (sgACC, rACC, dACC, pMCC, SI, SII, THA, PAG, OFC, AMYG, DLPFC, and VLPFC) on the relative target probability.</p><p id="p0115"><italic>Analysis 2: Main insular connection target for each pain-related brain region.</italic> In this analysis, we compared the structural connectivity with the 3 insular subdivisions, separately for each pain-related brain region using a 2-way ANOVA with the factors HEMISPHERE (left, right) and SUBDIVISION (AI, MI, PI).</p><p id="p0120"><italic>Analysis 3: Overall structural connectivity of the insular subdivisions with pain-related brain regions (across all target regions).</italic> For each insular subdivision in both hemispheres, we first averaged the seed-to-target probabilities for all 12 ipsilateral targets to calculate the mean seed-to-target probability for each subdivision. Subsequently, we performed a repeated-measures ANOVA with the factors HEMISPHERE (left, right) and SUBDIVISION (AI, MI, PI) on these indices.</p><p id="p0125"><italic>Analysis 4: Correlation with pain-relevant psychological traits.</italic> To test whether the structural connectivity between the insular subdivisions and target regions is related to pain-relevant psychological traits, Pearson correlations were calculated between the relative connectivity strength of each connection (averaged across both hemispheres) and the questionnaire scores (corrected for multiple comparisons).</p></sec><sec id="s0040"><label>2.4.2</label><title>Resting state connectivity</title><p id="p0130">Data preprocessing was carried out with FSL tools. The following prestatistics processing was applied for each subject: head motion correction by using MCFLIRT <xref rid="b0110" ref-type="bibr">[22]</xref>; nonbrain removal by using BET <xref rid="b0245" ref-type="bibr">[49]</xref>; spatial smoothing by using a Gaussian kernel of FWHM 5 mm; grand-mean intensity normalization of the entire 4-D data set by a single multiplicative factor; high-pass temporal filtering (subtraction of Gaussian-weighted least-squares straight-line fitting, with sigma&#x000a0;=&#x000a0;50.0&#x000a0;s). Registration of each subject&#x02019;s fMRI data to that subject&#x02019;s high-resolution structural image was carried out by using FLIRT <xref rid="b0110" ref-type="bibr">[22]</xref>. Registration from the high-resolution structurals to MNI152 standard space was achieved by using FLIRT affine registration and then further refined by using FNIRT nonlinear registration <xref rid="b0010" ref-type="bibr">[2]</xref>, <xref rid="b0015" ref-type="bibr">[3]</xref>.</p><p id="p0135">All subjects&#x02019; 4-D fMRI time series data were transformed into standard space at 2&#x000a0;&#x000d7;&#x000a0;2&#x000a0;&#x000d7;&#x000a0;2&#x000a0;mm resolution using the registration transformations derived as described above. For each subject, functional connectivity between insular subdivisions and other pain-related regions was calculated using the fsl_sbca tool (as described in <xref rid="b0165" ref-type="bibr">[33]</xref>). Briefly, the first Eigen-time-series of each of the pain-related regions were calculated; conditional correlations between the 3 insular subdivisions and the pain-related brain regions were then calculated. To this end, the Eigen-time-series of any 2 out of the 3 insular subregions were regressed out from the third subregion and the pain-related brain regions before calculating the correlation coefficient. The result is a set of partial correlation coefficients that determine interdependence between insula time series and other pain regions that are not shared across the insular subregions.</p><p id="p0140"><italic>Analysis 5: Identification of differences in resting state connectivity between targets.</italic> The conditional correlation coefficients we calculated were entered into a 2-way ANOVA with the factors HEMISPHERE (left, right) and TARGET (sgACC, rACC, dACC, pMCC, SI, SII, THA, PAG, OFC, AMYG, DLPFC, and VLPFC) separately for each of the 3 insular subdivisions.</p><p id="p0145"><italic>Analysis 6: Identification of differences in resting state connectivity for each pain-related brain regions with the 3 insular subdivisions.</italic> Conditional correlation coefficients characterizing the resting state connectivity were entered into a 2-way ANOVA with the factors HEMISPHERE (left, right) and SUBDIVISION (AI, MI, PI).</p><p id="p0150">Significant ANOVA results were followed up by post hoc <italic>t</italic> test which were adjusted for multiple comparisons using Bonferroni correction. Results reaching a <italic>P</italic> value of &#x0003c;.05 were considered statistically significant. To increase the readability of the results for analyses 2 and 5, test statistics and <italic>P</italic> values are provided in <xref rid="s0095" ref-type="sec">Tables S1 and S2</xref>.</p></sec></sec></sec><sec id="s0045"><label>3</label><title>Results</title><sec id="s0050"><label>3.1</label><title>Structural connectivity</title><p id="p0155"><italic>Analysis 1: Main connection targets of each insular subdivision.</italic> The analysis for the AI showed a main effect of TARGET (<italic>F</italic>(1.94,27.12)&#x000a0;=&#x000a0;178.89, <italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.001; <xref rid="f0010" ref-type="fig">Fig. 2</xref>), indicating that this subdivision was differentially connected with the various pain-related brain regions. Pairwise comparisons between targets revealed a significantly higher connection probability with the OFC (left: 40.2%, right: 47.9.1%) and VLPFC (left: 38.3.4%, right: 34.7%) compared to the other targets. The difference between OFC and VLPFC only reached statistical significance for the right hemisphere with stronger connectivity of the OFC (<italic>t</italic>(14)&#x000a0;=&#x000a0;5.02, <italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.001). Furthermore, the AI showed a significant interaction between TARGET and HEMISPHERE (<italic>F</italic>(1.95,27.31)&#x000a0;=&#x000a0;8.27, <italic>P</italic>&#x000a0;=&#x000a0;.002), suggesting that the differential connectivity with pain-related brain regions varied between hemispheres. Pairwise comparisons between hemispheres revealed that the AI is more strongly connected with the OFC in the right than the left hemisphere (<italic>t</italic>(14)&#x000a0;=&#x000a0;5.53, <italic>P</italic>&#x000a0;=&#x000a0;.001). All other comparisons did not reach significance (ie, <italic>P</italic>&#x000a0;&#x0003e;&#x000a0;.05).<fig id="f0010"><label>Fig. 2</label><caption><p>Differential structural connectivity of insular subdivisions with pain-related brain regions. (A) Anterior insula showed significantly higher connection probability with orbitofrontal cortex (OFC) and ventrolateral prefrontal cortex (VLPFC) than other pain-related regions. (B) Mid insula showed higher connection probability with primary (SI) and secondary (SII) somatosensory cortex, VLPFC, and pMCC compared to other pain-related regions. (C) Posterior insula was preferentially connected to SI, SII, and pMCC. dACC, dorsal anterior cingulate cortex; rACC, rostral anterior cingulate cortex; SI, primary somatosensory cortex; SII, secondary somatosensory cortex; THA, thalamus; PAG, periaqueductal grey; OFC, orbitofrontal cortex; AMYG, amygdala; DLPFC, dorsolateral prefrontal cortex; VLPFC, ventrolateral prefrontal cortex; pMCC, posterior midcingulate cortex; sgACC, subgenual anterior cingulate cortex.</p></caption><graphic xlink:href="gr2"/></fig></p><p id="p0160">The MI showed a main effect of TARGET (<italic>F</italic>(2.11,29.51)&#x000a0;=&#x000a0;55.81, <italic>P</italic>&#x000a0;&#x0003c;&#x000a0;<italic>P</italic>.001) (<xref rid="f0010" ref-type="fig">Fig. 2</xref>). The highest connection probability was found for SII (left: 37.7%, right: 41.8%), followed by the VLPFC (left: 13.5%, right: 18.0%), SI (left: 15.2%, right: 12.2%) and pMCC (left: 10.3%, right: 12.6%). The connection probability for SII was significantly higher than for any other target. The MI also exhibited a significant interaction between TARGET and HEMISPHERE (MI: (<italic>F</italic>(3.21,44.96)&#x000a0;=&#x000a0;4.12, <italic>P</italic>&#x000a0;=&#x000a0;.01). However, pairwise comparisons revealed no difference in connection probability between hemispheres when corrected for multiple comparisons.</p><p id="p0165">The PI showed a main effect of TARGET (<italic>F</italic>(1.78,24.95)&#x000a0;=&#x000a0;56.10, <italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.001) (<xref rid="f0010" ref-type="fig">Fig. 2</xref>). The highest connection probability was for SII (left: 48.1%, right: 50.3%), followed by SI (left: 18.6%, right: 17.3%) and the pMCC (left: 9.3%, right: 18.1%). The probability for SII was significantly higher than for any other target. In contrast to the AI and MI, we found no interaction between HEMISPHERE and TARGET (<italic>P</italic>&#x000a0;&#x0003e;&#x000a0;.05) for the posterior subdivision.</p><p id="p0170"><italic>Analysis 2: Main insular connection targets of each pain-related brain region.</italic> As revealed by a 2-way ANOVA, the structural connectivity of the rACC, AMYG, and DLPFC with the insula (across subdivisions) was more pronounced in the left than the right hemisphere (main effect HEMISPHERE). In contrast, the OFC showed a higher connectivity for the right than the left side. A main effect of SUBDIVISION, indicating differential structural connectivity with the 3 insular subdivisions was found for all pain-related brain regions, except the ACC targets (ie, dACC, rACC, and sgACC). A posterior&#x02013;anterior gradient in connectivity (ie, decrease in connectivity from PI to AI) was found for SI, SII, pMCC, and the PAG. An anterior&#x02013;posterior gradient (ie, a decrease in connectivity from AI to PI) was revealed for the VLPFC. The OFC and amygdala exhibited stronger connectivity with both AI and PI relative to MI. The DLPFC and thalamus showed a preference for MI relative to AI and PI. <xref rid="f0015" ref-type="fig">Figure 3</xref> displays the connectivity pattern of the target regions that showed the strongest differential connectivity in Analysis 1 (ie, SI, SII, OFC, and VLPFC). Details on significant interactions are provided in <xref rid="s0095" ref-type="sec">Table S1</xref>.<fig id="f0015"><label>Fig. 3</label><caption><p>Structural connectivity of main connection targets with insular subdivisions. Primary (A) and secondary (B) somatosensory cortices were more strongly connected to mid (MI) and posterior insula (PI) than anterior insula (AI). In contrast, orbitofrontal cortex (C) and ventrolateral prefrontal cortex (D) were preferentially connected with AI.</p></caption><graphic xlink:href="gr3"/></fig></p><p id="p0175"><italic>Analysis 3: Overall structural connectivity of the insular subdivisions with pain-related brain regions (across all target regions).</italic> The 2-way ANOVA revealed a significant main effect of SUBDIVISION on the mean seed-to-target probabilities (<italic>F</italic>(1.26,17.65)&#x000a0;=&#x000a0;65.06, <italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.001; <xref rid="f0020" ref-type="fig">Fig. 4</xref>). Pairwise comparisons between the subdivisions showed that the AI had a significantly higher connection probability than the MI (<italic>t</italic>(14)&#x000a0;=&#x000a0;8.29, <italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.001) and PI (<italic>t</italic>(14)&#x000a0;=&#x000a0;8.53, <italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.001). The connection probabilities for the MI and PI were not significantly different (<italic>P</italic>&#x000a0;&#x0003e;&#x000a0;.5). The effect of HEMISPHERE and the interaction between the 2 factors did not reach statistical significance (both <italic>P</italic>&#x000a0;&#x0003e;&#x000a0;.05).<fig id="f0020"><label>Fig. 4</label><caption><p>Overall structural connectivity of 3 insular subdivisions with ipsilateral pain-related brain regions. Structural connectivity of the anterior insula (AI) across all ipsilateral pain-related brain regions was significantly higher than the connectivity of the mid (MI) and posterior insula (PI).</p></caption><graphic xlink:href="gr4"/></fig></p><p id="p0180"><italic>Analysis 4: Correlation with pain-relevant psychological traits.</italic> The correlation analyses revealed that pain vigilance and awareness was positively related to the structural connectivity between the AI and amygdala (<italic>r</italic>&#x000a0;=&#x000a0;0.71, <italic>P</italic>&#x000a0;=&#x000a0;.04; <xref rid="f0025" ref-type="fig">Fig. 5</xref>A) and negatively related to the connectivity between AI and rACC (<italic>r</italic>&#x000a0;=&#x000a0;&#x02212;0.69; <italic>P</italic>&#x000a0;=&#x000a0;.05; <xref rid="f0025" ref-type="fig">Fig. 5</xref>B).<fig id="f0025"><label>Fig. 5</label><caption><p>Correlation between structural connectivity of anterior insula (AI) (averaged across both hemispheres) and pain vigilance and awareness. Pain vigilance and awareness as a relevant pain-related psychological trait is positively correlated with structural connectivity between AI and amygdala (A) and negatively correlated with connectivity between AI and rostral anterior cingulate cortex (B).</p></caption><graphic xlink:href="gr5"/></fig></p></sec><sec id="s0055"><label>3.2</label><title>Resting state connectivity</title><p id="p0185"><italic>Analysis 5: Main connection target of each insular subdivision.</italic> Resting state connectivity of the AI showed a significant main effect of REGION (<italic>F</italic>(6.70,234.59)&#x000a0;=&#x000a0;35.54, <italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.001; <xref rid="f0030" ref-type="fig">Fig. 6</xref>; <xref rid="s0095" ref-type="sec">Table S2</xref>). Post hoc tests revealed that across hemispheres, the AI was most strongly connected to the thalamus, followed by the VLPFC (<italic>t</italic>(35)&#x000a0;=&#x000a0;2.21, <italic>P</italic>&#x000a0;=&#x000a0;.034) and SII (<italic>t</italic>(35)&#x000a0;=&#x000a0;3.77; <italic>P</italic>&#x000a0;=&#x000a0;.001). We found no significant main effect for HEMISPHERE (<italic>F</italic>(1,35)&#x000a0;=&#x000a0;0.22, <italic>P</italic>&#x000a0;=&#x000a0;.644) or the interaction between REGION and HEMISPHERE (<italic>F</italic>(7.17,250.96)&#x000a0;=&#x000a0;1.15, <italic>P</italic>&#x000a0;=&#x000a0;.329).<fig id="f0030"><label>Fig. 6</label><caption><p>Resting state connectivity of each insular subdivision with pain-related brain regions. (A) Anterior insula showed strongest resting state connectivity with the thalamus, followed by VLPFC. (B) For the mid insula, strongest results were found for the thalamus, SII, and VLPFC. (C) The posterior insula was most strongly connected to SII and the thalamus. dACC, dorsal anterior cingulate cortex; rACC, rostral anterior cingulate cortex; SI, primary somatosensory cortex; SII, secondary somatosensory cortex; THA, thalamus; PAG, periaqueductal grey; OFC, orbitofrontal cortex; AMYG, amygdala; DLPFC, dorsolateral prefrontal cortex; VLPFC, ventrolateral prefrontal cortex; pMCC, posterior midcingulate cortex; sgACC, subgenual anterior cingulate cortex.</p></caption><graphic xlink:href="gr6"/></fig></p><p id="p0190">For MI, the resting state connectivity differed significantly between regions (main effect REGION (<italic>F</italic>(7.31,255.96)&#x000a0;=&#x000a0;97.31, <italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.001); <xref rid="f0030" ref-type="fig">Fig. 6</xref>), whereas the main effect of HEMISPHERE (<italic>F</italic>(1,35)&#x000a0;=&#x000a0;0.02; <italic>P</italic>&#x000a0;=&#x000a0;.885) and the interaction (<italic>F</italic>(6.93,242.69)&#x000a0;=&#x000a0;1.81, <italic>P</italic>&#x000a0;=&#x000a0;.088) were not significant. As revealed by post hoc tests across hemispheres, MI showed the strongest resting state connectivity with the thalamus and SII, which did not differ significantly (<italic>t</italic>(35)&#x000a0;=&#x000a0;0.83, <italic>P</italic>&#x000a0;=&#x000a0;.414). The connectivity with these 2 targets was significantly stronger than with the VLPFC (comparison with THA: <italic>t</italic>(35)&#x000a0;=&#x000a0;6.94, <italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.001; SII: <italic>t</italic>(35)&#x000a0;=&#x000a0;7.28, <italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.001).</p><p id="p0195">In contrast, the PI showed a main effect of REGION (<italic>F</italic>(7.31,255.53)&#x000a0;=&#x000a0;111.69; <italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.001; <xref rid="f0030" ref-type="fig">Fig. 6</xref>). Post hoc tests across hemispheres revealed that PI was most strongly connected with SII and the thalamus with significantly weaker connectivity for the latter (<italic>t</italic>(35)&#x000a0;=&#x000a0;3.11, <italic>P</italic>&#x000a0;=&#x000a0;.004). Furthermore we found a significant interaction between REGION and HEMISPHERE (<italic>F</italic>(7.06,247.15)&#x000a0;=&#x000a0;7.37, <italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.001). Post hoc tests showed a side difference only for the thalamus with a stronger resting state connectivity with PI on the right side (<italic>t</italic>(35)&#x000a0;=&#x000a0;6.14; <italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.001). The main effect of HEMISPHERE was not significant (<italic>F</italic>(1,35)&#x000a0;=&#x000a0;0.04; <italic>P</italic>&#x000a0;=&#x000a0;.850).</p><p id="p0200"><italic>Analysis 6: Main insular connection targets of each pain-related brain region.</italic> Of the 12 pain-related brain regions, the thalamus, PAG, and pMCC showed a significant main effect of HEMISPHERE with stronger resting state connectivity on the right for the thalamus and pMCC and on the left for the PAG (<xref rid="s0095" ref-type="sec">Table S2</xref>). A main effect of SUBDIVISION was found for all pain-related brain regions, except the thalamus, amygdala, pMCC, and sgACC, indicating that most pain-related regions show a different functional connectivity with the 3 insular divisions during rest. Post hoc tests revealed an anterior&#x02013;posterior gradient in connectivity in the dACC, VLPFC and DLPFC (<xref rid="f0035" ref-type="fig">Fig. 7</xref>). A posterior&#x02013;anterior gradient was found in SI and SII (<xref rid="f0035" ref-type="fig">Fig. 7</xref>). The rACC, PAG, and OFC showed stronger connectivity in AI and PI than in MI. A significant interaction between both factors was found in the thalamus, VLPFC, and sgACC. For the thalamus, the stronger connectivity with AI relative to PI was more pronounced in the left hemisphere while the connectivity for PI than MI was significantly higher on the right side. For the VLPFC, the difference between PI and MI was stronger on the left than the right side. Post hoc tests for the sgACC did not reveal any significant differences after correction for multiple comparisons.<fig id="f0035"><label>Fig. 7</label><caption><p>Resting state connectivity of insular subdivisions with pain-related brain regions. (A) Secondary somatosensory cortex (SII) showed stronger resting state connectivity with mid and posterior insula compared to anterior insula (AI). (B) In contrast, ventrolateral prefrontal cortex (VLPFC) exhibited stronger connectivity with AI than with posterior insula.</p></caption><graphic xlink:href="gr7"/></fig></p></sec></sec><sec id="s0060"><label>4</label><title>Discussion</title><p id="p0205">In this study, we investigated the differential structural and resting state connectivity of the AI, MI, and PI with other pain-related brain regions. The analyses revealed largely overlapping findings for the 2 types of connectivity. AI was predominantly connected to (pre-)frontal brain regions, namely the OFC (structural connectivity) and VLPFC (structural and resting state connectivity). PI showed the strongest connectivity with SI (structural connectivity) and SII (structural and resting state connectivity). MI displayed a hybrid pattern of connectivity with dominant connections to the VLPFC, SII (structural and resting state connectivity) and SI (structural connectivity).</p><sec id="s0065"><label>4.1</label><title>AI and (pre-)frontal cortex</title><p id="p0210">Our observation that AI is predominantly connected to the OFC and VLPFC confirms recent evidence on direct connections between the insula and both (pre-)frontal regions <xref rid="b0065" ref-type="bibr">[13]</xref>. Functionally, AI activation has been implicated in a number of tasks and contexts involving affective processing (see <xref rid="b0125" ref-type="bibr">[25]</xref> for review), including pain <xref rid="b0045" ref-type="bibr">[9]</xref>, <xref rid="b0130" ref-type="bibr">[26]</xref>, <xref rid="b0205" ref-type="bibr">[41]</xref>, <xref rid="b0310" ref-type="bibr">[62]</xref>. AI activation has been found when participants are awaiting an aversive outcome <xref rid="b0195" ref-type="bibr">[39]</xref>, <xref rid="b0205" ref-type="bibr">[41]</xref>, <xref rid="b0310" ref-type="bibr">[62]</xref> and the level of activation has been shown to scale with subsequent pain avoidance behavior <xref rid="b0275" ref-type="bibr">[55]</xref>. It has therefore been argued that activation in the AI represents a &#x02018;global emotional moment&#x02019; <xref rid="b0225" ref-type="bibr">[45]</xref> that reflects the net evaluation of the affective impact of an impending situation. In response to salient events, the AI can gate subsequent processing by activating a cognitive control network including the DLPFC and posterior parietal cortex <xref rid="b0265" ref-type="bibr">[53]</xref>. Our data indicate that the AI is anatomically well situated to also influence affective processing through its connections with key regions involved in evaluative processing (ie, OFC) and emotion regulation (ie, VLPFC).</p><p id="p0215">Like the AI, the OFC and VLPFC are both sensitive to emotional stimuli. However, the OFC primarily responds to the reward value of the stimulus (including negative value) rather than its sensory features. Importantly, OFC responses also encode the anticipation of future outcome <xref rid="b0120" ref-type="bibr">[24]</xref>, which makes it well suited for guiding subsequent decisions.</p><p id="p0220">The VLPFC plays a key role in modulating the impact of painful and other emotionally relevant stimuli on behavior and subjective emotional states <xref rid="b0150" ref-type="bibr">[30]</xref>. Activation in this brain region has been associated with voluntary attenuation of emotions <xref rid="b0170" ref-type="bibr">[34]</xref> and suppression of negative emotions <xref rid="b0190" ref-type="bibr">[38]</xref>. In the context of pain, VLPFC activation has been found during perceived control over pain <xref rid="b0230" ref-type="bibr">[46]</xref>, <xref rid="b0305" ref-type="bibr">[61]</xref>, placebo analgesia <xref rid="b0185" ref-type="bibr">[37]</xref>, and belief-related pain modulation <xref rid="b0300" ref-type="bibr">[60]</xref>. It has recently been argued that emotion regulation strategies might be implemented via the interplay between VLPFC and AI <xref rid="b0150" ref-type="bibr">[30]</xref>. In line with this notion, patients with bipolar disorder seem to lack the negative correlation between VLPFC and right AI that was found in healthy controls during performance of a cognitive interference task <xref rid="b0210" ref-type="bibr">[42]</xref>. Moreover, voluntary control of AI activation using real-time fMRI feedback leads to concomitant changes in VLPFC activity <xref rid="b0285" ref-type="bibr">[57]</xref> and increased functional connectivity between both regions has been found during noxious stimulation <xref rid="b0180" ref-type="bibr">[36]</xref>. On the basis of these findings, Menon and Uddin postulated that the core function of the AI is to detect relevant stimuli and engage executive brain regions (such as the VLPFC) to ensure an appropriate response <xref rid="b0140" ref-type="bibr">[28]</xref>. The strong anatomical connections between AI and OFC as well as VLPFC may therefore underlie the critical role of the anterior division in cognitive&#x02013;affective pain modulation.</p><p id="p0225">The link between the AI and psychological pain processing is further corroborated by our correlation analyses with PVAQ scores reflecting the disposition to attend to pain. This disposition is known to scale with pain-related anxiety <xref rid="b0135" ref-type="bibr">[27]</xref>, is related to a heightened perception of acute pain and is a relevant predictor for the development of chronic pain. High PVAQ scores were seen in those participants with stronger connectivity between AI and the amygdala (<xref rid="f0025" ref-type="fig">Fig. 5</xref>A). This result is in accordance with the recent observation of a positive correlation between AI-amygdala structural connectivity and trait anxiety in healthy volunteers <xref rid="b0030" ref-type="bibr">[6]</xref>.</p><p id="p0230">PVAQ scores were also negatively correlated with the degree of structural connectivity between AI and the rACC (<xref rid="f0025" ref-type="fig">Fig. 5</xref>B), which is - together with the PAG - a key component of the descending pain inhibitory pathway. Although speculative at this stage, a weak relative structural connectivity between AI and rACC might be related to only limited access to this pain-attenuating pathway even though the AI might have detected a painful event as highly salient. In accordance with this interpretation, Ploner et al. reported a negative correlation between PVAQ scores and the functional connectivity between AI and PAG <xref rid="b0205" ref-type="bibr">[41]</xref>.</p><p id="p0235">A rather surprising finding was the low probability for structural connections between AI and the dorsal anterior cingulate cortex (dACC). Several studies have reported increased functional connectivity between both regions in the context of pain <xref rid="b0310" ref-type="bibr">[62]</xref> and other emotional experiences <xref rid="b0240" ref-type="bibr">[48]</xref>. Furthermore, there is evidence for direct anatomical connections from a diffusion tensor imaging (DTI) study in humans <xref rid="b0100" ref-type="bibr">[20]</xref>, although this study did not differentiate between subdivisions of the insula and ACC. Besides an actual weak connectivity, our result might have been confounded by the problem of resolving crossing fibers <xref rid="b0295" ref-type="bibr">[59]</xref>. Connections between the insula and ACC must pass through several large white matter tracts (eg, cortico-spinal, superior longitudinal fascicle, and corpus callosum). We could not resolve such complex fiber architecture with the data used in this study. However, we only report relative connection probabilities that reflect connectivity relative to other targets. Recent advances in diffusion MRI acquisition and preprocessing, as part of the Human Connectome Project initiative <xref rid="b0095" ref-type="bibr">[19]</xref>, may enable more robust tracking of insula&#x02013;ACC connections (see <xref rid="b0255" ref-type="bibr">[51]</xref>).</p></sec><sec id="s0070"><label>4.2</label><title>PI and somatosensory cortices</title><p id="p0240">PI showed strongest connections with SII in both the structural and resting state connectivity and with SI in the structural connectivity analysis (<xref rid="f0010" ref-type="fig">Fig. 2</xref>, <xref rid="f0030" ref-type="fig">Fig. 6</xref>). Our findings thereby confirm observations from tracer studies in macaques <xref rid="b0145" ref-type="bibr">[29]</xref>, <xref rid="b0155" ref-type="bibr">[31]</xref>, previous DTI studies in humans <xref rid="b0060" ref-type="bibr">[12]</xref>, and resting state analyses <xref rid="b0055" ref-type="bibr">[11]</xref>, <xref rid="b0080" ref-type="bibr">[16]</xref>. Together with lateral thalamic nuclei and PI, both somatosensory regions constitute the lateral pain system that conveys sensory-discriminatory aspects of pain. Activation in the PI reflects the transition from nonpainful to painful sensations with increasing stimulus intensity <xref rid="b0175" ref-type="bibr">[35]</xref> and is specific for noxious relative to innocuous stimulation <xref rid="b0130" ref-type="bibr">[26]</xref>. The link between PI (and MI) and sensory-discriminative processing is further corroborated by the moderately strong structural connectivity with the pMCC where nociceptive neurons have been identified in vivo <xref rid="b0105" ref-type="bibr">[21]</xref> and ex vivo <xref rid="b0090" ref-type="bibr">[18]</xref>.</p><p id="p0245">Given the evidence from animal tracer studies for direct connections between the thalamus and PI <xref rid="b0075" ref-type="bibr">[15]</xref>, the low connection probability for the thalamus in our analysis on structural connectivity is surprising. Information about the physiological state of the body is projected from lamina 1 of the spinal cord to the ventromedial thalamic nucleus, which in turn projects to the MI/PI <xref rid="b0075" ref-type="bibr">[15]</xref>. It has, however, been pointed out that direct inferences from these animal data to structural connectivity in humans are difficult because the insula has undergone considerable expansion in humans <xref rid="b0070" ref-type="bibr">[14]</xref> and the anatomy varies significantly across species <xref rid="b0050" ref-type="bibr">[10]</xref>.</p></sec><sec id="s0075"><label>4.3</label><title>Hybrid connectivity pattern of the MI</title><p id="p0250">The structural connectivity of the MI has previously been reported in connection with the results for PI <xref rid="b0060" ref-type="bibr">[12]</xref>, <xref rid="b0065" ref-type="bibr">[13]</xref>, despite the cytoarchitectonic differences between both subdivisions <xref rid="b0145" ref-type="bibr">[29]</xref>, <xref rid="b0155" ref-type="bibr">[31]</xref>. However, our findings show minor but potentially relevant differences. In addition to the somatosensory cortices that are also key targets of PI, the MI is strongly connected to the VLPFC (<xref rid="f0010" ref-type="fig">Fig. 2</xref>, <xref rid="f0020" ref-type="fig">Fig. 4</xref>). It is therefore the only insular subdivision that has strong connections with sensory-discriminative (ie, SI and SII) as well as to a cognitive&#x02013;affective brain region (ie, the VLPFC). Widespread connections of MI have previously been reported in a whole-brain DTI study <xref rid="b0060" ref-type="bibr">[12]</xref> where MI showed connections with parietal and temporal regions, the inferior frontal gyrus, OFC and premotor cortex. Together, this finding of a hybrid connectivity pattern <xref rid="b0065" ref-type="bibr">[13]</xref> of MI is in line with the notion that this region integrates sensory and cognitive&#x02013;emotional information <xref rid="b0070" ref-type="bibr">[14]</xref>.</p></sec><sec id="s0080"><label>4.4</label><title>Conclusion</title><p id="p0255">Taken together, our data indicate that AI, MI, and PI are differentially connected with pain-related brain regions and that their connectivity is linked to pain-relevant behaviour. On the basis of these findings, it is tempting to speculate that these differences account for the differential functional role of the insular subdivisions in the perception of pain and beyond. Given that the insula is a multimodal region and that it is the most consistently activated brain region across all neuroimaging studies <xref rid="b0315" ref-type="bibr">[63]</xref>, our findings are equally relevant for other domains. Studies combining further behavioural indicators with functional and structural neuroimaging measures will aid in unraveling the functional significance of our findings. Finally, future studies should explore whether interindividual differences in behavioural and functional pain processing are indeed paralleled by differences in structural and resting state connectivity within distinct subsystems of the pain processing system. A more detailed understanding of the relationship between structure, function and behaviour promises insights into the resilience and susceptibility to conditions such as chronic pain.</p></sec></sec><sec id="s0085"><title>Conflict of interest</title><p id="p0260">The authors report no conflict of interest.</p></sec>

Multiple roles for Na_V1.9 in the activation of visceral afferents by noxious inflammatory, mechanical, and human disease&#x02013;derived stimuli

Could not extract abstract

<contrib contrib-type="author" id="au005"><name><surname>Hockley</surname><given-names>James R.F.</given-names></name><xref rid="af005" ref-type="aff">a</xref><xref rid="af010" ref-type="aff">b</xref></contrib><contrib contrib-type="author" id="au010"><name><surname>Boundouki</surname><given-names>George</given-names></name><xref rid="af005" ref-type="aff">a</xref><xref rid="af010" ref-type="aff">b</xref></contrib><contrib contrib-type="author" id="au015"><name><surname>Cibert-Goton</surname><given-names>Vincent</given-names></name><xref rid="af010" ref-type="aff">b</xref></contrib><contrib contrib-type="author" id="au020"><name><surname>McGuire</surname><given-names>Cian</given-names></name><xref rid="af010" ref-type="aff">b</xref></contrib><contrib contrib-type="author" id="au025"><name><surname>Yip</surname><given-names>Ping K.</given-names></name><xref rid="af015" ref-type="aff">c</xref></contrib><contrib contrib-type="author" id="au030"><name><surname>Chan</surname><given-names>Christopher</given-names></name><xref rid="af010" ref-type="aff">b</xref></contrib><contrib contrib-type="author" id="au035"><name><surname>Tranter</surname><given-names>Michael</given-names></name><xref rid="af010" ref-type="aff">b</xref></contrib><contrib contrib-type="author" id="au040"><name><surname>Wood</surname><given-names>John N.</given-names></name><xref rid="af020" ref-type="aff">d</xref></contrib><contrib contrib-type="author" id="au045"><name><surname>Nassar</surname><given-names>Mohammed A.</given-names></name><xref rid="af025" ref-type="aff">e</xref></contrib><contrib contrib-type="author" id="au050"><name><surname>Blackshaw</surname><given-names>L. Ashley</given-names></name><xref rid="af005" ref-type="aff">a</xref><xref rid="af010" ref-type="aff">b</xref></contrib><contrib contrib-type="author" id="au055"><name><surname>Aziz</surname><given-names>Qasim</given-names></name><xref rid="af005" ref-type="aff">a</xref></contrib><contrib contrib-type="author" id="au060"><name><surname>Michael</surname><given-names>Gregory J.</given-names></name><xref rid="af015" ref-type="aff">c</xref></contrib><contrib contrib-type="author" id="au065"><name><surname>Baker</surname><given-names>Mark D.</given-names></name><xref rid="af015" ref-type="aff">c</xref></contrib><contrib contrib-type="author" id="au070"><name><surname>Winchester</surname><given-names>Wendy J.</given-names></name><xref rid="af030" ref-type="aff">f</xref></contrib><contrib contrib-type="author" id="au075"><name><surname>Knowles</surname><given-names>Charles H.</given-names></name><xref rid="af010" ref-type="aff">b</xref></contrib><contrib contrib-type="author" id="au080"><name><surname>Bulmer</surname><given-names>David C.</given-names></name><email>d.bulmer@qmul.ac.uk</email><xref rid="af005" ref-type="aff">a</xref><xref rid="af010" ref-type="aff">b</xref><xref rid="cor1" ref-type="corresp">&#x0204e;</xref></contrib><aff id="af005"><label>a</label>Wingate Institute of Neurogastroenterology, Blizard Institute, Barts and the London School of Medicine and Dentistry, Queen Mary University of London, London E1 2AJ, UK</aff><aff id="af010"><label>b</label>National Centre for Bowel Research and Surgical Innovation, Blizard Institute, Barts and the London School of Medicine and Dentistry, Queen Mary University of London, London E1 2AT, UK</aff><aff id="af015"><label>c</label>Centre for Neuroscience and Trauma, Blizard Institute, Barts and the London School of Medicine and Dentistry, Queen Mary University of London, London E1 2AT, UK</aff><aff id="af020"><label>d</label>Molecular Nociception Group, Wolfson Institute for Biomedical Research, University College London, London WC1E 6BT, UK</aff><aff id="af025"><label>e</label>Department of Biomedical Science, The University of Sheffield, Sheffield S10 2TN, UK</aff><aff id="af030"><label>f</label>Neusentis (Pfizer Ltd), The Portway Building, Granta Science Park, Cambridge CB21 6GS, UK</aff>

Pain

<sec id="s0005"><label>1</label><title>Introduction</title><p id="p0005">Chronic dysregulation of visceral sensation can lead to abdominal pain, which is clinically challenging to treat. A hallmark of chronic visceral pain is heightened sensitivity to gut motility and distension of visceral organs, typically in response to past or ongoing inflammation <xref rid="b0130" ref-type="bibr">[26]</xref>. In some patient groups, therapeutic inhibition of mechanosensitivity in visceral afferents has proven effective in treating pain <xref rid="b0255" ref-type="bibr">[51]</xref>, <xref rid="b0260" ref-type="bibr">[52]</xref>. However, opportunities for improved intervention will likely come from targeting mechanisms of activation common to both inflammatory and mechanical stimulation of the viscera.</p><p id="p0010">The activity of voltage gated sodium channels (Na_V) underpins electrogenesis in excitable cells. The 9 genes (<italic>SCN1A&#x02013;SCN5A</italic> and <italic>SCN8A&#x02013;SCN11A</italic>) encoding the Na_V1.1 to Na_V1.9 &#x003b1;-subunits exhibit tissue specific expression and describe proteins with varying biophysical characteristics. Of those expressed in peripheral neurons, 3 channels (Na_V1.7, Na_V1.8, and Na_V1.9) are strongly associated with pain behaviours in both rodents and humans. Specifically, Na_V1.9 mediates a slowly inactivating persistent sodium current that is activated at voltages close to the resting membrane potential of sensory neurones. As a consequence of these biophysical properties, Na_V1.9 is proposed to contribute to the resting membrane potential and regulate nerve terminal excitability <xref rid="b0015" ref-type="bibr">[3]</xref>. More recently, we, as well as others, have shown that the Na_V1.9 sodium current may be enhanced by the algogenic mediator, ATP, and the inflammatory mediator PGE₂, acting through G protein&#x02013;coupled pathways, suggesting that Na_V1.9 may also play an important role in the activation of sensory nerves by inflammatory mediators <xref rid="b0010" ref-type="bibr">[2]</xref>, <xref rid="b0150" ref-type="bibr">[30]</xref>, <xref rid="b0220" ref-type="bibr">[44]</xref>. This hypothesis is supported for somatic pain by behavioural studies demonstrating reduced hypersensitivity to inflammatory stimuli in rodents where Na_V1.9 has been deleted or knocked down <xref rid="b0005" ref-type="bibr">[1]</xref>, <xref rid="b0140" ref-type="bibr">[28]</xref>, <xref rid="b0195" ref-type="bibr">[39]</xref>. However, evidence for a role of Na_V1.9 in visceral pain processing is controversial, with behavioural studies in Na_V1.9 knockout mice reporting either reduced or enhanced pain behaviours to noxious stimuli <xref rid="b0135" ref-type="bibr">[27]</xref>, <xref rid="b0155" ref-type="bibr">[31]</xref>, <xref rid="b0205" ref-type="bibr">[41]</xref>. Similarly the extent to which Na_V1.9 is expressed within visceral afferents is far from clear <xref rid="b0020" ref-type="bibr">[4]</xref>, <xref rid="b0120" ref-type="bibr">[24]</xref>. As a consequence, the contribution of Na_V1.9 to visceral afferent sensitivity remains to be defined.</p><p id="p0015">We hypothesized that the persistent sodium current mediated by Na_V1.9 plays a significant role in the polymodal sensitivity of visceral afferents to inflammatory mediators and the subsequent development of hypersensitivity to mechanical stimuli. We therefore investigated effects of Na_V1.9 gene deletion on the response of visceral afferents to inflammatory and mechanical stimuli, and Na_V1.9 expression in colon-projecting sensory neurons.</p></sec><sec id="s0010"><label>2</label><title>Materials and methods</title><p id="p0020">All experimental studies were performed in accordance with the UK Animal (Scientific Procedures) Act of 1986. Human tissue was collected and used with approval of the East London and the City HA Local Research Ethics Committee (NREC 10/H0703/71).</p><sec id="s0015"><label>2.1</label><title>In vitro mouse colonic splanchnic afferent preparations</title><p id="p0025">Na_V1.9^{&#x02212;/&#x02212;} mice were rederived from Na_V1.9^+/&#x02212; breeding pairs and originally generated by homologous recombination on a C57/BL6 background, as described previously <xref rid="b0180" ref-type="bibr">[36]</xref>. Adult Na_V1.9^+/+ or Na_V1.9^{&#x02212;/&#x02212;} mice of either sex were euthanized by rising concentration of CO₂, and the distal colon with associated lumbar splanchnic nerves removed. For whole-nerve experiments, tissues were cannulated, luminally perfused (100&#x000a0;&#x003bc;L/min), and serosally superfused (7&#x000a0;mL/min; 32&#x02013;34&#x000b0;C) with carbogenated Krebs buffer (in mM: 124 NaCl, 4.8 KCl, 1.3 NaH₂PO₄, 2.5 CaCl₂, 1.2 MgSO₄&#x000b7;7H₂O, 11.1 glucose, and 25 NaHCO₃) supplemented with nifedipine (10&#x000a0;&#x003bc;M) to block smooth muscle contraction, and indomethacin (3&#x000a0;&#x003bc;M) to block endogenous prostanoid production. Atropine (10&#x000a0;&#x003bc;M), the muscarinic acetylcholine receptor antagonist, was also added to the Krebs buffer to further prevent smooth muscle contractions. For single-fibre recordings, tissues were perfused with supplemented Krebs as described above; however, the colon was opened along the antimesenteric border and pinned with the flat mucosal side up. This preparation has been described in detail previously <xref rid="b0030" ref-type="bibr">[6]</xref>, <xref rid="b0040" ref-type="bibr">[8]</xref>, <xref rid="b0095" ref-type="bibr">[19]</xref>.</p></sec><sec id="s0020"><label>2.2</label><title>Electrophysiological recordings and characterization of colonic splanchnic afferent properties</title><p id="p0030">Borosilicate glass suction electrodes were used to record multiunit activity from whole lumbar splanchnic nerves (rostral to the inferior mesenteric ganglia) or single-unit activity from fibres teased from lumbar splanchnic nerves. Signals were amplified, band pass filtered (gain 5&#x000a0;K; 100&#x02013;1300&#x000a0;Hz; Neurolog, Digitimer Ltd, UK), and digitally filtered for 50&#x000a0;Hz noise (Humbug, Quest Scientific, Canada). Raw traces were digitized at 20&#x000a0;kHz (micro1401; Cambridge Electronic Design, UK), and action potential firing counts were determined using a threshold of twice the background noise (typically 100&#x000a0;&#x003bc;V). All signals were displayed on a PC using Spike 2 software. In flat sheet preparations, distinct receptive fields were identified and characterized according to previously published classifications by graded stimulus&#x02013;response to punctate von Frey hair (vFh) probing (0.07&#x000a0;g, 0.16&#x000a0;g, 0.4&#x000a0;g, 1&#x000a0;g, and 2&#x000a0;g; each applied 3 times for a period of 3&#x000a0;s), circumferential stretch (0&#x000a0;g, 5&#x000a0;g, and 10&#x000a0;g; each weight applied for 1&#x000a0;min, with an interval of 1&#x000a0;min between applications), and mucosal stroking with light vFh (0.16&#x000a0;g; applied 10 times) <xref rid="b0030" ref-type="bibr">[6]</xref>, <xref rid="b0095" ref-type="bibr">[19]</xref>, <xref rid="b0105" ref-type="bibr">[21]</xref>. A cantilever system was used to apply stretch via a thread attached with a purpose-made claw to the tissue adjacent to the receptive field. Addition of weights to the end of the cantilever system initiated colonic stretch. Four distinct classes of afferent fibre were identified on the basis of these responses: muscular (those responding to low-intensity circumferential stretch [&#x02a7d;5&#x000a0;g], but not fine mucosal stroking); mucosal (those responding to light vFh stroking), mesenteric (those responding to focal compression of the mesentery), and serosal (those responding to focal compression of the colon wall, but not low-intensity stroke or low-intensity circular stretch). In addition, conduction velocity was calculated for serosal and mesenteric units by dividing evoked action potentials latency elicited after electrical stimulation (0.5&#x000a0;Hz, 15&#x000a0;V, 1&#x000a0;ms) of the receptive field with concentric stimulating electrodes by the distance from the recording electrode to receptive field.</p></sec><sec id="s0025"><label>2.3</label><title>Retrograde labelling of colonic sensory neurones</title><p id="p0035">Fast Blue (FB; 2% in saline, Polysciences Gmbh, Germany) was injected into the wall of the distal colon of Na_V1.9^+/+ and Na_V1.9^{&#x02212;/&#x02212;} mice and male Sprague Dawley rats (150&#x02013;250&#x000a0;g). Briefly, animals were anaesthetized with isoflurane then an approximate 1.5&#x000a0;cm laparotomy performed. Five injections of 0.2&#x000a0;&#x003bc;L FB per site, at a rate of 0.4&#x000a0;&#x003bc;L/min via a microinfusion pump, were made into the wall of the distal colon using a fine-pulled glass needle. The muscle layer was sutured and the skin secured with Michel clips. Postoperative analgesia (buprenorphine 0.05&#x02013;0.1&#x000a0;mg/kg daily) and care (monitoring body weight and soft diet) was provided.</p><p id="p0040">After 3&#x000a0;days for mice and 7&#x000a0;days for rats, animals were euthanized with sodium pentobarbital (200&#x000a0;mg/kg i.p.) and transcardially perfused with saline (0.9%) followed by paraformaldehyde (4% in 0.1&#x000a0;M phosphate buffer; pH 7.4). Dorsal root ganglia (DRG; Na_V1.9^+/+ and Na_V1.9^{&#x02212;/&#x02212;} mice, T13&#x02013;L1; rats, L2) were removed and postfixed in 4% paraformaldehyde for 2&#x000a0;h and cryoprotected in 30% sucrose (w/v phosphate-buffered saline). Cryostat sections (10&#x000a0;&#x003bc;m) were collected sequentially across 10 slides per DRG.</p></sec><sec id="s0030"><label>2.4</label><title>Immunohistochemistry</title><p id="p0045">Sections were blocked in antibody diluent (10% horse serum and 0.2% (v/v) Triton X-100 in 0.1&#x000a0;M phosphate-buffered saline) for 1&#x000a0;h, followed by overnight incubation with primary antibodies (rabbit anti-Na_V1.9 antibody [1:1000; Alomone, Israel] and goat anti-CGRP antibody [1:2000; Abcam, UK]) and a 4&#x000a0;h incubation with fluorophore conjugated secondary antibodies (donkey anti-rabbit IgG-Alexafluor-488 and donkey anti-goat IgG-Alexafluor-568 [1:1000; Invitrogen, UK]) and/or isolectin B4 [IB4] from <italic>Griffonia simplicifolia</italic>&#x02013;Alexafluor-647 (2.5&#x000a0;&#x003bc;g/mL; Invitrogen, UK). No labelling was observed in control experiments where the primary antibody was excluded or in the presence of a competing blocking peptide.</p></sec><sec id="s0035"><label>2.5</label><title>In situ hybridization</title><p id="p0050">Oligonucleotide probes complementary to bases 968 to 1001 (probe 1) and 2641 to 2674 (probe 2) of the mouse <italic>SCN11A</italic> sequence, accession number NM 011887.3, were synthesized (Sigma, UK), end labelled with ³⁵S, and hybridized to DRG sections using previously described protocols and visualized by silver grain development in radiographic emulsion <xref rid="b0160" ref-type="bibr">[32]</xref>. No specific labelling was observed in DRG sections of Na_V1.9^{&#x02212;/&#x02212;} mice hybridized with <italic>SCN11A</italic> probes; reactions where an excess of unlabelled probe was used resulted in only background signal.</p></sec><sec id="s0040"><label>2.6</label><title>Imaging and quantitation</title><p id="p0055">Sections were imaged and the relative intensities of reaction products after immunostaining for Na_V1.9 and other markers and <italic>SCN11A</italic> mRNA in situ hybridization were determined for all DRG cells with visible nuclei (ImageJ 1.45S analysis software, NIH, USA). The mean background absorbance was subtracted to control for variability in illumination. Percentage relative intensities of background-subtracted cell absorbance were determined by comparison with least intensely (0%) and most intensely (100%) labelled profiles. For in situ hybridization, relative intensity of DRG cells was measured in Na_V1.9^{&#x02212;/&#x02212;} sections to set the threshold for positive labelling in wild-type mice. Cells with intensity values greater than the mean intensity of the 10 cells with highest background intensity values in Na_V1.9^{&#x02212;/&#x02212;} sections plus 2 times its SD were considered positively labelled. For Na_V1.9, CGRP and IB4-like immunostaining, intensity of staining was scored on a scale of 0 to 5 by 2 independent observers, with 0 representing negative and 5 strongly positive. The mean absorbance of these cells taken from ImageJ analysis correlated with staining scores (eg, Pearson <italic>r</italic>&#x000a0;=&#x000a0;0.87; <italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.0001; <italic>n</italic>&#x000a0;=&#x000a0;149) and a threshold for positive staining was determined as performed by Fang et al. <xref rid="b0065" ref-type="bibr">[13]</xref> (eg, mean absorbance for Na_V1.9&#x000a0;&#x02a7e;&#x000a0;32%).</p></sec><sec id="s0045"><label>2.7</label><title>Generation of human tissue supernatants</title><p id="p0060">Resected human colon was obtained after full written consent was obtained from patients undergoing elective surgery at Barts Health NHS Trust, London, after approval by the local research ethics committee (NREC 10/H0703/71). Patient details are outlined in <xref rid="t0005" ref-type="table">Table 1</xref>. Control supernatants were generated using macroscopically normal colon (&#x0003e;10&#x000a0;cm from tumour margin) obtained from patients (male/female ratio, 2/1; mean age, 61&#x000a0;years) undergoing colectomy as part of their normal surgical treatment for bowel cancer. Disease tissue supernatants were derived from chronically inflamed colon or intestine obtained from patients undergoing operations as part of their standard surgical treatment for Crohn disease (CD) (male/female ratio, 7/3; mean age, 30&#x000a0;years) or ulcerative colitis (UC) (male, 4; mean age, 22&#x000a0;years). All patients had active disease which was unresponsive to medical treatment. Tissue samples were incubated in fresh carbogenated Krebs buffer at 37&#x000b0;C for 25&#x000a0;min at a fixed volume of 2.5&#x000a0;mL/g of tissue. After incubation, the tissue was removed and the buffer centrifuged at 2000&#x000a0;&#x000d7;&#x000a0;<italic>g</italic> for 10&#x000a0;min. The remaining supernatant was formed into aliquots and stored at &#x02212;80&#x000b0;C until use.<table-wrap position="float" id="t0005"><label>Table 1</label><caption><p>Patient details of human tissue used in the present studies.<xref rid="tblfn1" ref-type="table-fn">a</xref></p></caption><table frame="hsides" rules="groups"><thead><tr><th>Patient no.</th><th>Disease</th><th>Operation</th><th>Age</th><th>Sex</th><th>Cytokine analysis</th><th>Whole nerve</th><th>Single fibre</th></tr></thead><tbody><tr><td align="char">1</td><td>Cancer</td><td>Laparoscopic anterior resection</td><td align="char">50</td><td>F</td><td>Y</td><td>Y</td><td/></tr><tr><td align="char">2</td><td>Cancer</td><td>Right hemicolectomy</td><td align="char">76</td><td>M</td><td>Y</td><td>Y</td><td/></tr><tr><td align="char">3</td><td>Cancer</td><td>Subtotal colectomy</td><td align="char">57</td><td>M</td><td>Y</td><td>Y</td><td/></tr><tr><td/><td/><td><italic>Mean age/M:F ratio</italic></td><td align="char"><italic>61</italic></td><td><italic>2:1</italic></td><td/><td/><td/></tr><tr><td align="char">5</td><td>CD</td><td>Laparotomy and proceed</td><td align="char">21</td><td>F</td><td>Y</td><td>Y</td><td>Y</td></tr><tr><td align="char">6</td><td>CD</td><td>Redo right hemicolectomy</td><td align="char">30</td><td>F</td><td>Y</td><td>Y</td><td>Y</td></tr><tr><td align="char">7</td><td>CD</td><td>Right hemicolectomy</td><td align="char">21</td><td>M</td><td>Y</td><td>Y</td><td/></tr><tr><td align="char">8</td><td>CD</td><td>Completion colectomy</td><td align="char">25</td><td>M</td><td/><td>Y</td><td/></tr><tr><td align="char">9</td><td>CD</td><td>Subtotal colectomy</td><td align="char">30</td><td>M</td><td/><td/><td>Y</td></tr><tr><td align="char">10</td><td>CD</td><td>Small bowel resection</td><td align="char">24</td><td>F</td><td/><td/><td>Y</td></tr><tr><td align="char">11</td><td>CD</td><td>Small bowel resection</td><td align="char">42</td><td>M</td><td/><td/><td>Y</td></tr><tr><td align="char">12</td><td>CD</td><td>Colectomy and end ileostomy</td><td align="char">28</td><td>M</td><td/><td/><td>Y</td></tr><tr><td align="char">13</td><td>CD</td><td>Extended right hemicolectomy</td><td align="char">16</td><td>M</td><td/><td/><td>Y</td></tr><tr><td align="char">14</td><td>CD</td><td>Subtotal colectomy</td><td align="char">64</td><td>M</td><td/><td/><td>Y</td></tr><tr><td/><td/><td><italic>Mean age/M:F ratio</italic></td><td align="char"><italic>30</italic></td><td><italic>2.3:1</italic></td><td/><td/><td/></tr><tr><td align="char">15</td><td>UC</td><td>Subtotal colectomy</td><td align="char">21</td><td>M</td><td>Y</td><td>Y</td><td/></tr><tr><td align="char">16</td><td>UC</td><td>Restorative proctocolectomy</td><td align="char">20</td><td>M</td><td>Y</td><td>Y</td><td/></tr><tr><td align="char">17</td><td>UC</td><td>Subtotal colectomy</td><td align="char">30</td><td>M</td><td>Y</td><td>Y</td><td/></tr><tr><td align="char">18</td><td>UC</td><td>Restorative proctocolectomy</td><td align="char">18</td><td>M</td><td>Y</td><td>Y</td><td/></tr><tr><td/><td/><td><italic>Mean age/M:F ratio</italic></td><td align="char"><italic>22</italic></td><td><italic>4:0</italic></td><td/><td/><td/></tr></tbody></table><table-wrap-foot><fn><p>CD, Crohn disease; UC, ulcerative colitis.</p></fn></table-wrap-foot><table-wrap-foot><fn id="tblfn1"><label>a</label><p id="np005">Supernatants generated from these tissues have been used for inflammatory cytokine quantification (cytokine analysis), 20&#x000a0;mL bath superfusion of whole-nerve recordings from the lumbar splanchnic nerve (whole nerve), or discrete ring application to characterized receptive fields in a flat sheet preparation and single-fibre recording from the lumbar splanchnic nerve (single fibre).</p></fn></table-wrap-foot></table-wrap></p></sec><sec id="s0050"><label>2.8</label><title>Inflammatory cytokine quantification</title><p id="p0065">Quantitative analysis of protein cytokine levels was performed on supernatant samples using established capture sandwich immunoassay and magnetic microsphere methodology <xref rid="b0170" ref-type="bibr">[34]</xref>. Samples were prepared as per manufacturer&#x02019;s instructions (Invitrogen, UK) and analysed with the Luminex MAGPIX detection system (Luminex, TX, USA) for IL-1&#x003b2;, IL-6, GM-CSF, TNF-&#x003b1;, and IL-8.</p></sec><sec id="s0055"><label>2.9</label><title>Electrophysiology protocols</title><p id="p0070">In multiunit experiments, drugs were applied after a stabilizing period of 30&#x000a0;min by bath superfusion of a 20&#x000a0;mL volume. Where repeat concentrations of drugs were given, a minimum 60&#x000a0;min interval was maintained. Increasing concentrations of ATP (0.1, 1.0, and 3.0&#x000a0;mM) or PGE₂ (3&#x000a0;&#x003bc;M) were applied in separate colonic preparations. Supernatants derived from CD, UC, and control tissues were applied by 20&#x000a0;mL bath superfusion in separate colonic preparations. Ramp distensions were performed by blocking luminal perfusion out-flow of the cannulated colon producing noxious pressures known to both evoke pain behaviours in vivo in mice and robustly activate all known afferent mechanoreceptors (0&#x02013;80&#x000a0;mm&#x000a0;Hg; <xref rid="b0105" ref-type="bibr">[21]</xref>, <xref rid="b0175" ref-type="bibr">[35]</xref>). In later studies, ramp distensions from 0 to 145&#x000a0;mm&#x000a0;Hg (&#x0223c;4&#x02013;5&#x000a0;min) were conducted to provide a supraphysiological pressure. Response to 0 to 80&#x000a0;mm Hg ramp distension was also investigated after pretreatment (20&#x000a0;min) with and in the presence of intraluminally perfused IS. In some studies, ramp distensions were performed once firing rates had returned to baseline activity after drug application. Sets of 6 rapid phasic ramp distensions (0&#x02013;80&#x000a0;mm&#x000a0;Hg, 60&#x000a0;s at 9&#x000a0;min intervals) were implemented in separate experiments. Single-fibre studies were only performed on muscular, mesenteric, or serosal units as a result of the sparse nature of mucosal splanchnic afferents. In all 3 types of units, mechanosensitivity was examined in response to graded vFh probing (0.07&#x000a0;g, 0.16&#x000a0;g, 0.4&#x000a0;g, 1&#x000a0;g, and 2&#x000a0;g) and circumferential stretch (0&#x000a0;g, 5&#x000a0;g, and 10&#x000a0;g). In mesenteric and serosal units, the effect of applying either an IS (consisting of 1&#x000a0;&#x003bc;M bradykinin, 1&#x000a0;mM ATP, 10&#x000a0;&#x003bc;M histamine, 10&#x000a0;&#x003bc;M PGE₂, and 10&#x000a0;&#x003bc;M 5HT; 2&#x000a0;min) or CD inflammatory supernatant (5&#x000a0;min) to receptive fields isolated by a metal ring was also examined on ongoing nerve discharge. Additionally, changes in the response to vFh probing with a 2&#x000a0;g hair were examined before and after application of IS or supernatant.</p></sec><sec id="s0060"><label>2.10</label><title>Data analysis</title><p id="p0075">In multiunit experiments, peak changes and time profiles of electrophysiological nerve activity were determined by subtracting baseline firing (5&#x000a0;min before drug application) from increases in nerve activity after drug/supernatant superfusion or ramp distension. Changes in nerve activity were compared between Na_V1.9^+/+ and Na_V1.9^{&#x02212;/&#x02212;} animals by Student&#x02019;s <italic>t</italic> test or 2-way ANOVA with Bonferroni post hoc test, as appropriate. In teased single-fibre flat sheet experiments, average spikes/s per stimulus were compared between genotype or before and after ring application of drugs/supernatants. All cytokine quantification data were analysed nonparametrically. If detectable levels of cytokines were present in control supernatants, Mann-Whitney <italic>U</italic> tests were performed, or else Wilcoxon signed rank tests comparing to a theoretical median of 0.0 were performed. Statistical significance was set at <italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.05. Data are presented as mean&#x000a0;&#x000b1;&#x000a0;SEM, <italic>N</italic>&#x000a0;=&#x000a0;number of animals, and <italic>n</italic>&#x000a0;=&#x000a0;number of observations.</p></sec><sec id="s0065"><label>2.11</label><title>Drugs</title><p id="p0080">Stock concentrations of ATP (300&#x000a0;mM; water), PGE₂ (1&#x000a0;mM; ethanol), bradykinin (10&#x000a0;mM; water), histamine (100&#x000a0;mM; water), 5HT (10&#x000a0;mM; water), atropine (10&#x000a0;mM; ethanol), indomethacin (3&#x000a0;mM; DMSO) and nifedipine (10&#x000a0;mM; DMSO) were all purchased from Sigma Aldrich (UK) and prepared as described. IS was prepared in advance and aliquots frozen until use. All compounds were diluted to working concentrations in buffer on the day of experimentation.</p></sec></sec><sec id="s0070"><label>3</label><title>Results</title><sec id="s0075"><label>3.1</label><title>Localization of Na_V1.9 in visceral afferent neurons</title><p id="p0085">We first examined the expression of Na_V1.9 transcripts in colon-projecting thoracolumbar DRG sections of mouse via in situ hybridization (ISH). Under polarized light, Na_V1.9-positive labelling was observed as clusters of silver grains visible over cells and was present in 69.0&#x000a0;&#x000b1;&#x000a0;3.0% of all neurons comparable with previous studies in rat (<xref rid="f0005" ref-type="fig">Fig. 1</xref>) <xref rid="b0055" ref-type="bibr">[11]</xref>, <xref rid="b0090" ref-type="bibr">[18]</xref>. Injections of FB retrograde tracer into the colon wall labelled 6.7&#x000a0;&#x000b1;&#x000a0;1.8% of thoracolumbar DRG neurons in Na_V1.9^+/+ mice (<xref rid="f0005" ref-type="fig">Fig. 1</xref>Aii), and of these, 50.5&#x000a0;&#x000b1;&#x000a0;3.3% expressed Na_V1.9 transcripts. A similar proportion of FB-positive colonic neurons were found in Na_V1.9^{&#x02212;/&#x02212;} mice (8.3&#x000a0;&#x000b1;&#x000a0;0.9% vs Na_V1.9^+/+ FB positive, <italic>P</italic>&#x000a0;=&#x000a0;.46), indicating no loss or lesion of these cells. Importantly, no specific Na_V1.9 expression was observed in sections from Na_V1.9^{&#x02212;/&#x02212;} mice (<xref rid="f0005" ref-type="fig">Fig. 1</xref>B) and those incubated with excess unlabelled probe (data not shown). Size&#x02013;frequency analysis of all neurons revealed that Na_V1.9 mRNA was expressed to a higher extent in small (&#x0003c;20&#x000a0;&#x003bc;m diameter, 83.4&#x000a0;&#x000b1;&#x000a0;2.9%) compared to medium to large (&#x0003e;20&#x000a0;&#x003bc;m diameter, 50.2&#x000a0;&#x000b1;&#x000a0;2.9%) neurons. In the FB-positive population, a similar distribution was seen with 71.1&#x000a0;&#x000b1;&#x000a0;10.9% small-diameter and 44.5&#x000a0;&#x000b1;&#x000a0;3.6% medium-large diameter cells expressing Na_V1.9 transcripts (<xref rid="f0005" ref-type="fig">Fig. 1</xref>D).<fig id="f0005"><label>Fig. 1</label><caption><p>Expression of Na_V1.9 mRNA in visceral colonic sensory neurons in mouse. Radiographic in situ hybridization of Na_V1.9 mRNA transcript expression (bright field with Giemsa counterstain, i; polarized light, iii) in thoracolumbar dorsal root ganglia (DRG) section from Na_V1.9^+/+ (A) and Na_V1.9^{&#x02212;/&#x02212;} (B) mice combined with retrograde labelling of sensory neurons from the colon with Fast Blue (FB) (FB only, ii; merge, iv; scale bar&#x000a0;=&#x000a0;50 &#x003bc;m). (C) Expanded top-right region of Na_V1.9^+/+ ganglia section (A) showing 4 retrogradely labelled sensory neurons (scale bar&#x000a0;=&#x000a0;30&#x000a0;&#x003bc;m). Yellow arrows indicate colon projecting neurons positive for Na_V1.9 expression. Yellow arrowheads indicate colon projecting neurons negative for Na_V1.9 expression. (D) Cross-sectional area histogram of Fast Blue&#x02013;positive neurons from Na_V1.9^+/+ mice superimposed with those which also exhibit Na_V1.9 mRNA transcript expression (<italic>N</italic>&#x000a0;=&#x000a0;4).</p></caption><graphic xlink:href="gr1"/></fig></p><p id="p0090">To further characterize Na_V1.9 protein expression and confirm our mRNA distribution data, we used a commercially available Na_V1.9 antibody in rat DRG sections. Additionally, this was performed in conjunction with CGRP and IB4 labelling, on the basis of the strong correlation between IB4 binding and Na_V1.9 expression reported previously (Fang 2006). The Na_V1.9 antibody produced nonspecific labelling in mouse DRG but proved highly specific in rat tissues <xref rid="b0265" ref-type="bibr">[53]</xref>. No specific staining was seen in control experiments after omission of the Na_V1.9 antibody or preincubation with blocking peptide (data not shown). Rat Na_V1.9 immunoreactivity was observed in a similar proportion of total neurons (64.5&#x000a0;&#x000b1;&#x000a0;2.4%, <xref rid="f0010" ref-type="fig">Fig. 2</xref>A) to mouse ISH, and colonic injection of FB labelled a similar proportion of neurones in rat DRGs (9.2&#x000a0;&#x000b1;&#x000a0;1.6%). In agreement with the mouse ISH data in FB-positive neurones, 51.9&#x000a0;&#x000b1;&#x000a0;5.8% of rat FB-positive cells stained positive for Na_V1.9 protein. As expected, a greater proportion of CGRP-expressing cells (80.6&#x000a0;&#x000b1;&#x000a0;8.3%) was observed in the colon-projecting population compared to all neurons (34.8&#x000a0;&#x000b1;&#x000a0;1.1%, <xref rid="f0010" ref-type="fig">Fig. 2</xref>B, C), indicative of an enriched peptidergic population <xref rid="b0210" ref-type="bibr">[42]</xref>. The extensive colocalization of Na_V1.9 with IB4 staining predicted by previous studies was also observed in FB-negative neurons (73.4&#x000a0;&#x000b1;&#x000a0;3.5% of Na_V1.9; <xref rid="t0010" ref-type="table">Table 2</xref>) <xref rid="b0070" ref-type="bibr">[14]</xref>. By contrast, Na_V1.9/IB4 colocalization was less extensive (34.7&#x000a0;&#x000b1;&#x000a0;7.3% of Na_V1.9) in FB-positive colonic populations, with an enrichment of colocalization with CGRP (from FB negative: 32.7&#x000a0;&#x000b1;&#x000a0;1.1% to FB positive: 89.6&#x000a0;&#x000b1;&#x000a0;7.8% of Na_V1.9, <xref rid="f0010" ref-type="fig">Fig. 2</xref>B, C). As such, within the colon-projecting population, Na_V1.9 appears to be almost exclusively expressed in CGRP-positive neurons.<fig id="f0010"><label>Fig. 2</label><caption><p>Expression of Na_V1.9 in visceral colonic sensory neurons in rat. (A) Multiple fluorescent immunohistochemistry of Na_V1.9 (i) and CGRP (iii) immunoreactivity and IB4 binding (iv; scale bar&#x000a0;=&#x000a0;50 &#x003bc;m) in dorsal root ganglia from rat combined with retrograde labelling of sensory neurons from the colon (ii, merge of Na_V1.9-IR and FB). Solid arrows indicate Fast Blue (FB)-positive neurons with Na_V1.9, CGRP, and IB4 (white) or Na_V1.9 and CGRP (orange). Open arrows indicate example FB-negative neurons with Na_V1.9, CGRP, and IB4 (white) or Na_V1.9 and IB4 (light blue). (B) Proportional Venn diagram of coimmunoreactivity for Na_V1.9 and CGRP and IB4 binding in neuronal populations retrogradely labelled from the colon in rat. Approximately 13.3% of FB-positive neurons were negative for Na_V1.9, CGRP, and IB4 binding (<italic>N</italic>&#x000a0;=&#x000a0;3). (C) Proportional Venn diagram of coimmunoreactivity for Na_V1.9 and CGRP and IB4 binding in noncolonic neuronal populations (FB negative) in rat. Approximately 21.6% of FB-negative neurons were negative for Na_V1.9, CGRP, and IB4 binding (<italic>N</italic>&#x000a0;=&#x000a0;3).</p></caption><graphic xlink:href="gr2"/></fig><table-wrap position="float" id="t0010"><label>Table 2</label><caption><p>Comparison of Na_V1.9 and CGRP immunohistochemical expression and IB4 binding in colonic and noncolonic sensory neurons.<xref rid="tblfn2" ref-type="table-fn">a</xref></p></caption><table frame="hsides" rules="groups"><thead><tr><th>Characteristic</th><th colspan="2">Colonic (FB positive; N&#x000a0;=&#x000a0;4, <italic>n</italic>&#x000a0;=&#x000a0;247)</th><th colspan="2">Noncolonic (FB negative; N&#x000a0;=&#x000a0;4, <italic>n</italic>&#x000a0;=&#x000a0;2556)</th></tr></thead><tbody><tr><td>Na_V1.9-IR</td><td colspan="2">51.9&#x000a0;&#x000b1;&#x000a0;5.8%</td><td colspan="2">65.9&#x000a0;&#x000b1;&#x000a0;3.0%</td></tr><tr><td>CGRP-IR</td><td colspan="2">80.6&#x000a0;&#x000b1;&#x000a0;8.3%</td><td colspan="2">30.0&#x000a0;&#x000b1;&#x000a0;0.4%</td></tr><tr><td>IB4 binding</td><td colspan="2">26.0&#x000a0;&#x000b1;&#x000a0;5.3%</td><td colspan="2">54.1&#x000a0;&#x000b1;&#x000a0;0.9%</td></tr><tr><td/><td>% of Na_V1.9</td><td>% of CGRP</td><td>% of Na_V1.9</td><td>% of CGRP</td></tr><tr><td>Na_V1.9 and CGRP-IR</td><td>89.6&#x000a0;&#x000b1;&#x000a0;7.8%</td><td>57.7&#x000a0;&#x000b1;&#x000a0;5.1%</td><td>32.7&#x000a0;&#x000b1;&#x000a0;1.1%</td><td>71.7&#x000a0;&#x000b1;&#x000a0;3.5%</td></tr><tr><td/><td>% of Na_V1.9</td><td>% of IB4</td><td>% of Na_V1.9</td><td>% of IB4</td></tr><tr><td>Na_V1.9-IR and IB4 binding</td><td>34.7&#x000a0;&#x000b1;&#x000a0;7.3%</td><td>68.5&#x000a0;&#x000b1;&#x000a0;7.7%</td><td>73.4&#x000a0;&#x000b1;&#x000a0;3.5%</td><td>89.2&#x000a0;&#x000b1;&#x000a0;2.1%</td></tr></tbody></table><table-wrap-foot><fn><p>FB, Fast Blue; CGRP, calcitonin gene&#x02013;regulated peptide; IB4, isolectin-B4; IR, immunoreactivity.</p></fn></table-wrap-foot><table-wrap-foot><fn id="tblfn2"><label>a</label><p id="np010">Percentage coexpression in terms of total stated marker.</p></fn></table-wrap-foot></table-wrap></p></sec><sec id="s0080"><label>3.2</label><title>Deletion of Na_V1.9 inhibits colonic afferent responses to ATP and PGE₂</title><p id="p0095">To investigate the role of persistent tetrodotoxin-resistant sodium current activation in responses of visceral afferents to ATP and PGE₂, we made extracellular recordings of the lumbar splanchnic nerve in Na_V1.9^+/+ and Na_V1.9^{&#x02212;/&#x02212;} mice. Multiunit recordings provided an unbiased observation of the involvement of Na_V1.9 in lumbar splanchnic nerve responses to superfused application of ATP and PGE₂. Baseline firing was reduced in Na_V1.9^{&#x02212;/&#x02212;} mice compared to control animals (0.7&#x000a0;&#x000b1;&#x000a0;0.1&#x000a0;spikes/s, <italic>N</italic>&#x000a0;=&#x000a0;41 vs 1.4&#x000a0;&#x000b1;&#x000a0;0.3&#x000a0;spikes/s, <italic>N</italic>&#x000a0;=&#x000a0;49; <italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.05). The sequential application of increasing ATP concentrations led to robust and concentration-dependent increases in afferent firing (<xref rid="f0015" ref-type="fig">Fig. 3</xref>) in control mice. PGE₂ (3&#x000a0;&#x003bc;m) also evoked comparable afferent excitation (<xref rid="f0020" ref-type="fig">Fig. 4</xref>). However, in Na_V1.9^{&#x02212;/&#x02212;} mice, responses to ATP and to PGE₂ were significantly diminished for the duration of the response. Even at the highest concentrations of ATP applied (3&#x000a0;mM), firing rates and nerve terminal sensitivity were reduced by 73% in Na_V1.9^{&#x02212;/&#x02212;} afferents.<fig id="f0015"><label>Fig. 3</label><caption><p>Effect of increasing concentrations of ATP on colonic splanchnic nerve activity in Na_V1.9^+/+ and Na_V1.9^{&#x02212;/&#x02212;} mice. Example rate histogram and neurogram response in Na_V1.9^+/+ (A) and Na_V1.9^{&#x02212;/&#x02212;} (B) mice to 0.1&#x000a0;mM, 1&#x000a0;mM, and 3&#x000a0;mM ATP. Below, expanded traces at baseline and after addition of 3&#x000a0;mM ATP and example action potential for each genotype. (C) Response profiles to addition of 1&#x000a0;mM ATP in Na_V1.9^+/+ (<italic>N</italic>&#x000a0;=&#x000a0;6) and Na_V1.9^{&#x02212;/&#x02212;} mice (<italic>N</italic>&#x000a0;=&#x000a0;6). (D) Peak increase in nerve activity after addition of 0.1&#x000a0;mM, 1&#x000a0;mM, and 3&#x000a0;mM ATP above baseline activity in Na_V1.9^+/+ (<italic>N</italic>&#x000a0;=&#x000a0;6) and Na_V1.9^{&#x02212;/&#x02212;} (<italic>N</italic>&#x000a0;=&#x000a0;6) mice. ^{&#x02217;&#x02217;}<italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.01, ^{&#x02217;&#x02217;&#x02217;}<italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.001.</p></caption><graphic xlink:href="gr3"/></fig><fig id="f0020"><label>Fig. 4</label><caption><p>Effect of 3&#x000a0;&#x003bc;M PGE₂ on colonic splanchnic nerve activity in Na_V1.9^+/+ and Na_V1.9^{&#x02212;/&#x02212;} mice. Example rate histogram and neurogram response in Na_V1.9^+/+ (A) and Na_V1.9^{&#x02212;/&#x02212;} (B) mice to 3&#x000a0;&#x003bc;M PGE₂. Below, expanded traces at baseline and after addition of 3&#x000a0;&#x003bc;M PGE₂ and example action potential for each genotype. (C) Response profiles to addition of 3&#x000a0;&#x003bc;M PGE₂ in Na_V1.9^+/+ (<italic>N</italic>&#x000a0;=&#x000a0;6) and Na_V1.9^{&#x02212;/&#x02212;} mice (<italic>N</italic>&#x000a0;=&#x000a0;6). (D) Peak increase in nerve activity after addition of 3&#x000a0;&#x003bc;M PGE₂ in Na_V1.9^+/+ (<italic>N</italic>&#x000a0;=&#x000a0;6) and Na_V1.9^{&#x02212;/&#x02212;} (<italic>N</italic>&#x000a0;=&#x000a0;6) mice. ^{&#x02217;&#x02217;}<italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.01, ^{&#x02217;&#x02217;&#x02217;}<italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.001.</p></caption><graphic xlink:href="gr4"/></fig></p></sec><sec id="s0085"><label>3.3</label><title>Effects of Na_V1.9 deletion on responses of colonic afferents to IS and subsequent mechanical hypersensitivity</title><p id="p0100">We next investigated a more generic role for Na_V1.9 in afferent responses to multiple inflammatory mediators. To achieve this, single-fibre recordings were made in colonic flat sheet preparations, and receptive fields in the mesentery and serosa were identified. A mixture of multiple inflammatory mediators commonly used in pain studies was applied <xref rid="b0075" ref-type="bibr">[15]</xref>, <xref rid="b0115" ref-type="bibr">[23]</xref>, <xref rid="b0245" ref-type="bibr">[49]</xref>. Discrete application of IS (bradykinin, ATP, histamine, serotonin, and PGE₂) for 2&#x000a0;min in a small chamber surrounding the receptive field led to robust activation in 100% of serosal and 80% of mesenteric units in Na_V1.9^+/+ mice (<xref rid="f0025" ref-type="fig">Fig. 5</xref>). In line with our previous whole-nerve recordings of responses to ATP and PGE₂, the increase in firing of single afferent fibres after IS was statistically reduced in Na_V1.9^{&#x02212;/&#x02212;} mice, and only 43% of serosal (<italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.05) and 56% of mesenteric (<italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.05) units responded (<xref rid="f0025" ref-type="fig">Fig. 5</xref>C).<fig id="f0025"><label>Fig. 5</label><caption><p>Chemical and mechanical stimulation of single-fibre mechanoreceptive fields in mouse colon. Example colonic nerve activity to ring application (2&#x000a0;min) of inflammatory soup (IS) over a receptive field located in the serosa of Na_V1.9^+/+ (A) and Na_V1.9^{&#x02212;/&#x02212;} (B) mice. Probes with von Frey hair (2&#x000a0;g) were performed prior to and after IS application. (C) Response to IS ring application in serosal and mesenteric receptive fields. (D) Afferent hypersensitivity after IS application was observed in Na_V1.9^+/+ mice in both serosal and mesenteric units. This was not present in Na_V1.9^{&#x02212;/&#x02212;} mice.</p></caption><graphic xlink:href="gr5"/></fig></p><p id="p0105">After responses to IS, mechanical hypersensitivity of colonic afferents was seen. We compared responses to 2&#x000a0;g vFh probing before and after IS application in control animals and observed a significant increase in responses after IS (<italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.05, <xref rid="f0025" ref-type="fig">Fig. 5</xref>D). This is consistent with previous studies using the application of bradykinin <xref rid="b0035" ref-type="bibr">[7]</xref>. IS-induced mechanical hypersensitivity to a 2&#x000a0;g probe was completely lost in Na_V1.9^{&#x02212;/&#x02212;} mice (serosal, <italic>P</italic>&#x000a0;=&#x000a0;.76; mesenteric, <italic>P</italic>&#x000a0;=&#x000a0;.63), indicating that Na_V1.9 may directly influence mechanical hypersensitivity of visceral afferents induced by IS.</p></sec><sec id="s0090"><label>3.4</label><title>Na_V1.9 deletion reduces colonic afferent responses to human tissue&#x02013;derived inflammatory supernatants</title><p id="p0110">We investigated whether ablation of Na_V1.9 would also alter responses to supernatants derived from inflamed human tissue. To do this, resected bowel tissue from patients undergoing surgery for inflammatory bowel disease (IBD; CD and UC) was incubated in Krebs solution to generate individual IBD supernatants. Separately, macroscopically normal tissue taken from bowel cancer resections (&#x0003e;10&#x000a0;cm away from the cancer itself, cancer margins, and lymph nodes) was used to generate control supernatants. These control supernatants enabled possible excitatory effects caused by surgical manipulation, removal of tissue, and temporal ischaemia to be discounted from afferent activation driven by inflammatory mediators released from native tissue. Importantly, no further steps were taken to concentrate the tissue supernatant in any way. To confirm the inflammatory status of the supernatants, cytokine quantification for IL-6, IL-8, IL-1&#x003b2;, GM-CSF, and TNF-&#x003b1; was performed. In control supernatants, there were detectable levels of IL-6 (7.5&#x000a0;&#x000b1;&#x000a0;3.3&#x000a0;pg/mL) and IL-8 (43.4&#x000a0;&#x000b1;&#x000a0;13.8&#x000a0;pg/mL), with levels of the remaining cytokines below detection limits (<xref rid="f0030" ref-type="fig">Fig. 6</xref>D). Cytokine levels were all significantly elevated in IBD supernatants (IL-1&#x003b2;: 35.4&#x000a0;&#x000b1;&#x000a0;22.7; IL-6: 598.6&#x000a0;&#x000b1;&#x000a0;276.3; GM-CSF: 37.5&#x000a0;&#x000b1;&#x000a0;14.9; TNF-&#x003b1;: 31.5&#x000a0;&#x000b1;&#x000a0;23.6; IL-8: 3419.3&#x000a0;&#x000b1;&#x000a0;2083.7&#x000a0;pg/mL; each cytokine vs control, <italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.05), with cytokine levels greatest in supernatants generated from CD tissue compared to those of UC [(IL-1&#x003b2;: 8.8&#x000a0;&#x000b1;&#x000a0;3.6; IL-6: 186.4&#x000a0;&#x000b1;&#x000a0;32.5; GM-CSF: 15.7&#x000a0;&#x000b1;&#x000a0;3.0; TNF-&#x003b1;: 6.2&#x000a0;&#x000b1;&#x000a0;2.1; IL-8: 1040.0&#x000a0;&#x000b1;&#x000a0;231.2&#x000a0;pg/mL) CD: (IL-1&#x003b2;: 75.7&#x000a0;&#x000b1;&#x000a0;46.3; IL-6: 1157.0&#x000a0;&#x000b1;&#x000a0;507.8; GM-CSF: 66.8&#x000a0;&#x000b1;&#x000a0;27.9; TNF-&#x003b1;: 64.0&#x000a0;&#x000b1;&#x000a0;54.4; IL-8: 6446.0&#x000a0;&#x000b1;&#x000a0;4715.0&#x000a0;pg/mL)]. It should be mentioned that the influence on cytokine production by immunosuppressant and anti-TNF-&#x003b1; antibody therapies was not accounted for during patient selection. Supernatants were applied to mouse whole-nerve lumbar splanchnic recordings in a cannulated tubular preparation, similar to that used for ATP and PGE₂ application. IBD supernatants elicited a robust increase in afferent firing compared to control supernatants in wild-type animals (3.5&#x000a0;&#x000b1;&#x000a0;0.5&#x000a0;spikes/s vs control supernatant 0.7&#x000a0;&#x000b1;&#x000a0;0.1&#x000a0;spikes/s, <italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.01, <xref rid="f0030" ref-type="fig">Fig. 6</xref>C). In Na_V1.9^{&#x02212;/&#x02212;} mice, the response to IBD supernatants was greatly attenuated and almost comparable in magnitude to control supernatant response in wild-type mice (1.2&#x000a0;&#x000b1;&#x000a0;0.3&#x000a0;spikes/s, vs Na_V1.9^+/+ mice, <italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.01). We then investigated the effect of application of CD supernatants to individual isolated receptive fields using single-fibre recordings in flat sheet preparations. Responses to application of CD supernatant were seen in 64% of serosal (1.37&#x000a0;&#x000b1;&#x000a0;0.36&#x000a0;spikes/20&#x000a0;s) and 50% of mesenteric afferents (1.13&#x000a0;&#x000b1;&#x000a0;0.22&#x000a0;spikes/20&#x000a0;s) (<xref rid="f0030" ref-type="fig">Fig. 6</xref>E). The proportions of units responding in Na_V1.9^{&#x02212;/&#x02212;} mice were comparable (<italic>P</italic>&#x000a0;=&#x000a0;1.00, <xref rid="f0030" ref-type="fig">Fig. 6</xref>F); however, the degree of afferent activation elicited by application of CD supernatant to receptive fields was significantly less (serosal: 0.31&#x000a0;&#x000b1;&#x000a0;0.11&#x000a0;spikes/20&#x000a0;s, <italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.05 vs Na_V1.9^+/+; mesenteric: 0.42&#x000a0;&#x000b1;&#x000a0;0.16&#x000a0;spikes/20&#x000a0;s, <italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.05 vs Na_V1.9^+/+). Unlike effects of experimental IS, no mechanical hypersensitivity was observed after CD supernatant in either Na_V1.9^+/+ or Na_V1.9^{&#x02212;/&#x02212;} mice to 2&#x000a0;g vFh probe (data not shown), regardless of whether afferents showed direct excitation to supernatant. However, it is clear that Na_V1.9 contributes to direct afferent excitation by both artificial and natural inflammatory milieu of differing composition.<fig id="f0030"><label>Fig. 6</label><caption><p>Effect of inflammatory bowel disease (IBD) supernatants on colonic nerve activity. (A) Example rate histograms of colonic nerve activity to control supernatant in Na_V1.9^+/+ mice and to Crohn disease (CD) tissue supernatant in Na_V1.9^+/+ and Na_V1.9^{&#x02212;/&#x02212;} mice. (B) Example rate histograms of colonic nerve activity to ulcerative colitis (UC) supernatant in both Na_V1.9^+/+ and Na_V1.9^{&#x02212;/&#x02212;} mice. (C) Peak change in activity after addition of control (<italic>N</italic>&#x000a0;=&#x000a0;3), IBD supernatants in Na_V1.9^+/+ (<italic>N</italic>&#x000a0;=&#x000a0;7), and Na_V1.9^{&#x02212;/&#x02212;} mice (<italic>N</italic>&#x000a0;=&#x000a0;6). In addition, CD and UC responses are plotted separately (right; <italic>N</italic>&#x000a0;=&#x000a0;3&#x02013;4 per genotype). ^&#x02217;<italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.05, ^{&#x02217;&#x02217;}<italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.01 vs control; ^#<italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.05, ^##<italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.01 Na_V1.9^+/+ vs Na_V1.9^{&#x02212;/&#x02212;}. (D) Cytokine (IL-1&#x003b2;, IL-6, GM-CSF, TNF-&#x003b1;, and IL-8) levels in IBD (<italic>N</italic>&#x000a0;=&#x000a0;7) and control (<italic>N</italic>&#x000a0;=&#x000a0;3) supernatants applied to afferent nerve recordings. ^&#x02217;<italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.05 vs control. (E) Responses to ring application of CD supernatants over serosal and mesenteric receptive fields in Na_V1.9^+/+ and Na_V1.9^{&#x02212;/&#x02212;} mice. (F) Number of responsive vs nonresponsive serosal and mesenteric units when CD supernatants were applied.</p></caption><graphic xlink:href="gr6"/></fig></p></sec><sec id="s0095"><label>3.5</label><title>Deletion of Na_V1.9 increases mechanosensory activation thresholds of afferent fibres and reduces maintenance of repeated responses</title><p id="p0115">Our demonstration of a role for Na_V1.9 in mechanical hypersensitivity in response to IS prompted us to also determine how Na_V1.9 influences mechanosensation. To do this, we used whole-nerve recordings of distension responses in a cannulated tubular preparation of the distal colon, plus single-fibre recording techniques in a flat sheet preparation. Phasic distension of the bowel to 80&#x000a0;mm&#x000a0;Hg intraluminal pressure led to a robust initial increase in afferent activity after adapting to a plateau phase for the remainder of the 1&#x000a0;min distension (<xref rid="f0035" ref-type="fig">Fig. 7</xref>). Second and third repeat distensions (at 9&#x000a0;min intervals) evoked smaller peak responses until a stabilized peak response was reached by the fourth to sixth distension. In Na_V1.9^{&#x02212;/&#x02212;} mice, peak responses to first distension were similar to control animals (<italic>P</italic>&#x000a0;=&#x000a0;.17, <xref rid="f0035" ref-type="fig">Fig. 7</xref>C). However, afferent responses to subsequent repeat distension in Na_V1.9^{&#x02212;/&#x02212;} mice showed greater tachyphylaxis with significant reductions observed by the third to sixth distensions (<italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.001, <xref rid="f0035" ref-type="fig">Fig. 7</xref>B, E).<fig id="f0035"><label>Fig. 7</label><caption><p>Example rate histogram of colonic splanchnic nerve response to repeat phasic distension (0&#x02013;80&#x000a0;mm&#x000a0;Hg; 60&#x000a0;s; 10&#x000a0;min intervals) in Na_V1.9^+/+ (A) and Na_V1.9^{&#x02212;/&#x02212;} mice (B). Average response profiles to the first (C) and sixth (D) phasic ramp distension in Na_V1.9^+/+ and Na_V1.9^{&#x02212;/&#x02212;} mice. (E) Peak change in firing rate during sequential phasic distensions (^&#x02217;<italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.05; ^{&#x02217;&#x02217;}<italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.01; ^{&#x02217;&#x02217;&#x02217;}<italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.001, unpaired <italic>t</italic> test).</p></caption><graphic xlink:href="gr7"/></fig></p><p id="p0120">These findings suggested a particular role for Na_V1.9 in responses to persistent stimuli. To investigate this role further, we used a slow ramp distension paradigm. Initial ramp distensions up to 80&#x000a0;mm&#x000a0;Hg intraluminal pressure evoked maintained whole-nerve responses with a linear correlate between afferent activity and pressure, reaching a maximum firing rate of 13.3&#x000a0;&#x000b1;&#x000a0;3.7&#x000a0;spikes/s (<italic>N</italic>&#x000a0;=&#x000a0;6) in control animals. Na_V1.9^{&#x02212;/&#x02212;} mice, by contrast, elicited minimal activity during comparable ramp distensions, with a small burst of firing observed at &#x0223c;30&#x000a0;mm&#x000a0;Hg (maximum 1.4&#x000a0;&#x000b1;&#x000a0;0.6&#x000a0;spikes/s, <italic>N</italic>&#x000a0;=&#x000a0;6, <italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.0001). We reasoned that rapid phasic distension in experiments described above develops a more substantial stimulus to mechanoreceptive afferents compared to a slow ramp fill, even up to equivalent final pressures (80&#x000a0;mm&#x000a0;Hg). Therefore, it should be possible to overcome the lack of response to slow ramp distension by increasing the stimulus to supraphysiologic distension pressures. The maximum distension pressure achievable in mouse colon before rupture was determined at 159.6&#x000a0;&#x000b1;&#x000a0;4.3&#x000a0;mm Hg (<italic>N</italic>&#x000a0;=&#x000a0;10). Ramp distension to 145&#x000a0;mm&#x000a0;Hg was thus used to provide a stronger stimulus. Ramp distensions beyond 80&#x000a0;mm&#x000a0;Hg continued to linearly increase afferent firing rates to a maximum of 34.6&#x000a0;&#x000b1;&#x000a0;7.3&#x000a0;spikes/s in Na_V1.9^+/+ mice (<xref rid="f0040" ref-type="fig">Fig. 8</xref>A). However, in Na_V1.9^{&#x02212;/&#x02212;} mice, afferent firing remained just above baseline (&#x0003c;2&#x000a0;spikes/s) up to &#x0223c;95&#x000a0;mm&#x000a0;Hg, above which an exponential increase in firing was observed reaching firing rates almost comparable to Na_V1.9^+/+ animals at 145&#x000a0;mm&#x000a0;Hg (25.4&#x000a0;&#x000b1;&#x000a0;7.8&#x000a0;spikes/s, <xref rid="f0040" ref-type="fig">Fig. 8</xref>B, C). Mechanical sensitization of responses to noxious ramp distensions (80&#x000a0;mm&#x000a0;Hg) was investigated by intraluminal perfusion of IS. In Na_V1.9^+/+ mice, perfusion of IS led to direct activation of basal afferent firing rates, which was not observed in Na_V1.9^{&#x02212;/&#x02212;} mice (3.8&#x000a0;&#x000b1;&#x000a0;0.6&#x000a0;spikes/s, <italic>N</italic>&#x000a0;=&#x000a0;5 vs 0.0&#x000a0;&#x000b1;&#x000a0;0.4&#x000a0;spikes/s <italic>N</italic>&#x000a0;=&#x000a0;6; <italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.001, unpaired <italic>t</italic> test). Responses to ramp distension during IS perfusion were potentiated in Na_V1.9^+/+ mice, an effect not observed in Na_V1.9^{&#x02212;/&#x02212;} mice up to 50&#x000a0;mm&#x000a0;Hg (<xref rid="f0040" ref-type="fig">Fig. 8</xref>D). At greater ramp distension pressures, sensitized responses were observed in Na_V1.9^{&#x02212;/&#x02212;} tissues compared to control experiments, which may reflect either the sensitization of afferent fibres negative for Na_V1.9 or recruitment of alternative sensitizing mechanisms within Na_V1.9-containing fibers.<fig id="f0040"><label>Fig. 8</label><caption><p>Example rate histogram and raw trace of colonic splanchnic nerve response to ramp distension (0&#x02013;145&#x000a0;mm&#x000a0;Hg) in Na_V1.9^+/+ (A) and Na_V1.9^{&#x02212;/&#x02212;} (B) mice with intraluminal pressure trace. (C) Firing rates to distension pressures of &#x0003c;95&#x000a0;mm&#x000a0;Hg were greatly attenuated in Na_V1.9^{&#x02212;/&#x02212;} animals compared to Na_V1.9^+/+ mice (<italic>N</italic>&#x000a0;=&#x000a0;6), with an increase in firing at greater pressures in Na_V1.9^{&#x02212;/&#x02212;} mice (<italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.0001, 2-way ANOVA with Bonferroni&#x02019;s post hoc test, ^&#x02217;<italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.05; ^{&#x02217;&#x02217;}<italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.01; ^{&#x02217;&#x02217;&#x02217;}<italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.001). (D) Firing rates to noxious 80&#x000a0;mm Hg ramp distension in Na_V1.9^+/+ and Na_V1.9^{&#x02212;/&#x02212;} mice were increased by intraluminal perfusion with inflammatory soup (IS) (<italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.0001, Na_V1.9^+/+ vs Na_V1.9^{&#x02212;/&#x02212;}, 2-way ANOVA with Bonferroni&#x02019;s post hoc test, ^&#x02217;<italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.05; ^{&#x02217;&#x02217;}<italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.01; ^{&#x02217;&#x02217;&#x02217;}<italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.001; <italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.0001, Na_V1.9^+/+ in presence of IS vs Na_V1.9^{&#x02212;/&#x02212;} in presence of IS, 2-way ANOVA with Bonferroni&#x02019;s post hoc test, ^#<italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.05, ^##<italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.01, ^###<italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.001).</p></caption><graphic xlink:href="gr8"/></fig></p><p id="p0125">The deficits in afferent mechanosensation that we observed could be due to increased adaptation above threshold and/or altered thresholds for activation in particular fibers. To identify which distension-sensitive afferent subtype was altered in Na_V1.9^{&#x02212;/&#x02212;} mice, we used established single-fibre recording techniques in a colonic flat sheet preparation. Although von Frey probes enable the highly reproducible and directed stimuli required to interrogate specific afferent subtypes, they are unlikely to represent an in vivo stimulus encountered by the intact bowel. There are 4 subtypes of lumbar splanchnic nerve afferents characterized by their location and responses to stroke, von Frey probe, and circumferential stretch: mucosal, muscular, mesenteric, and serosal afferents <xref rid="b0030" ref-type="bibr">[6]</xref>. Stimulus&#x02013;response curves to probe with increasing weight of vFhs of serosal and muscular units was unchanged between genotype (<italic>P</italic>&#x000a0;=&#x000a0;.94 and <italic>P</italic>&#x000a0;=&#x000a0;.24, respectively), with no significant differences observed in either mechanosensory thresholds or responses to circumferential stretch of the preparation (<xref rid="f0045" ref-type="fig">Fig. 9</xref>). We did observe a significant reduction in mesenteric afferent responses to vFh probing, specifically lower-intensity probing (0.16&#x02013;1&#x000a0;g vFh), with a corresponding increase in mechanosensory threshold at these intensities (<xref rid="f0045" ref-type="fig">Fig. 9</xref>B, E). No difference in conduction velocity was observed between mechanoreceptors in control and Na_V1.9^{&#x02212;/&#x02212;} mice (mean conduction velocity Na_V1.9^+/+ mice, 0.50&#x000a0;&#x000b1;&#x000a0;0.03&#x000a0;m/s, <italic>n</italic>&#x000a0;=&#x000a0;33, <italic>N</italic>&#x000a0;=&#x000a0;10 vs Na_V1.9^{&#x02212;/&#x02212;} mice, 0.52&#x000a0;&#x000b1;&#x000a0;0.05 m/s, <italic>n</italic>&#x000a0;=&#x000a0;52, <italic>N</italic>&#x000a0;=&#x000a0;16; <italic>P</italic>&#x000a0;=&#x000a0;.82), signifying that action potential propagation is unaffected by loss of Na_V1.9. Together, these data suggest that high-intensity mechanical stimulation of colonic afferents (eg, 2&#x000a0;g vFh probing, &#x0003e;95&#x000a0;mm&#x000a0;Hg ramp distension, and initial rapid phasic distension) is unaffected by Na_V1.9 gene deletion. However, knockout of Na_V1.9 appears to decrease the sensitivity of mesenteric afferents to low-intensity stimulation, which accounts for the significant deficits observed in whole-nerve recordings during ramp distension.<fig id="f0045"><label>Fig. 9</label><caption><p>Stimulus&#x02013;response curves to von Frey hair (vFh) probing for afferent fibre subtypes in Na_V1.9^+/+ and Na_V1.9^{&#x02212;/&#x02212;} mice. Stimulus&#x02013;response curves to vFh probing (0.07&#x02013;2&#x000a0;g) for serosal (A) and mesenteric (B) afferent fibres (2-way ANOVA with Bonferroni&#x02019;s post hoc test, ^&#x02217;<italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.05; ^{&#x02217;&#x02217;}<italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.01; ^{&#x02217;&#x02217;&#x02217;}<italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.001). Associated activation thresholds of vFh probing for serosal (C) and mesenteric (D) afferent fibres (Fisher&#x02019;s exact test at each probe weight, ^{&#x02217;&#x02217;}<italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.01; ^{&#x02217;&#x02217;&#x02217;}<italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.001).</p></caption><graphic xlink:href="gr9"/></fig></p></sec></sec><sec id="s0100"><label>4</label><title>Discussion</title><p id="p0130">Our data demonstrate that Na_V1.9 is required for direct excitatory responses of colonic afferents to noxious inflammatory mediators and is the subsequent development of mechanical hypersensitivity. We show that Na_V1.9 is involved in determining the mechanosensory thresholds of a subset of colonic afferents with receptive fields in the mesenteric attachments, and we provide evidence of a significant role for Na_V1.9 during persistent (repeat phasic distensions) or sustained (ramp distension) noxious distension of the colon. Finally, we show that Na_V1.9 is important in the activation of visceral nociceptors located in the colonic serosa and mesentery by mediators released during human inflammatory disease.</p><p id="p0135">Given the polymodal nature of visceral afferents responsive to noxious stimuli, there are likely points of convergence in the molecular signalling pathways underlying action potential generation at their endings. These may provide valuable targets by which future treatments could counteract mechanical hypersensitivity induced by on-going inflammation. Voltage-gated sodium channels may represent such a convergence point. Of the peripherally expressed sodium channels, valuable insight has been obtained regarding the function of Na_V1.7 and Na_V1.8 in visceral chronic pain, not least due to causal links to human pain syndromes <xref rid="b0085" ref-type="bibr">[17]</xref>, <xref rid="b0200" ref-type="bibr">[40]</xref>. However, functional evidence for a role of Na_V1.9 in the activation of visceral afferents has not been extensively explored. The data presented here suggest that Na_V1.9 represents an important regulator of specific aspects of sustained visceral afferent excitability, especially in responses to inflammatory mediators.</p><p id="p0140">It is known that the persistent sodium current defined by Na_V1.9 can be greatly enhanced by inflammatory mediators acting at GPCRs via G_q/11/PKC and G_i/o/PKA intracellular signalling pathways <xref rid="b0010" ref-type="bibr">[2]</xref>, <xref rid="b0050" ref-type="bibr">[10]</xref>, <xref rid="b0150" ref-type="bibr">[30]</xref>, <xref rid="b0220" ref-type="bibr">[44]</xref>. In some studies, this has required the application of multiple mediators <xref rid="b0150" ref-type="bibr">[30]</xref>, whereas in other studies, the application of single mediators alone is sufficient to increase Na_V1.9 currents <xref rid="b0010" ref-type="bibr">[2]</xref>, <xref rid="b0220" ref-type="bibr">[44]</xref>. Our study extends these observations by examining the contribution of Na_V1.9 to nociceptor activation at the level of the nerve ending. We observe significantly reduced colonic afferent firing in Na_V1.9^{&#x02212;/&#x02212;} mice to individually applied ATP and PGE₂, even at supramaximal concentrations, demonstrating that Na_V1.9 is required in the activation of nerve fibres by these mediators. These findings suggest that the application of single mediators is sufficient to enhance Na_V1.9 currents and trigger action potentials in visceral afferents. However, it is important to stress that other inflammatory mediators may also be present at the afferent terminal due to ongoing interactions with immune cells or gut microbiota. As a consequence, we cannot rule out the possibility that multiple mediators may be acting to enhance Na_V1.9 currents. Another consideration is that both ATP and PGE₂ can enhance currents generated by other sodium channels (eg, Na_V1.8 <xref rid="b0010" ref-type="bibr">[2]</xref>) or ion channels <xref rid="b0225" ref-type="bibr">[45]</xref>, <xref rid="b0250" ref-type="bibr">[50]</xref>; however, our data suggest that visceral afferent excitability to these mediators appears highly dependent on Na_V1.9. The magnitude of the effect to ATP was potentially surprising given the relatively small proportion of colon afferent somata responsive to P2X_1,3 agonist (30&#x02013;35% <xref rid="b0125" ref-type="bibr">[25]</xref>) and may reflect the engagement of alternative purinoceptors, including P2Y receptors.</p><p id="p0145">In addition to attenuating responses to single mediators, in Na_V1.9^{&#x02212;/&#x02212;} mice we also observed a substantial reduction in serosal and mesenteric afferent firing after the application of an experimental IS (consisting of bradykinin, histamine, PGE₂, ATP, and 5HT). This indicates that Na_V1.9 is required for visceral afferent activation to a range of inflammatory mediators. To explore this further, we generated supernatants by incubating resected tissue from patients with IBD. These disease derived IBD supernatant&#x02013;evoked increases in afferent activity in whole-nerve recordings from mouse colon, and in single-unit recordings from serosal and mesenteric afferents; both were significantly attenuated in Na_V1.9^{&#x02212;/&#x02212;} mice. Collectively, this suggests that during human inflammatory disease where multiple inflammatory stimuli are present, Na_V1.9 contributes significantly to visceral afferent activation.</p><p id="p0150">Deletion of Na_V1.9 also caused significant deficits in mechanosensation and in the development of mechanical hypersensitivity after application of inflammatory mediators. These deficits were investigated using different mechanical stimuli, including von Frey probing of isolated visceral nociceptor receptive fields, repeated phasic distension of the colon at noxious pressures, and ramp distension of the colon into the noxious pressure range. Such studies have proven invaluable in the development of our current understanding of how noxious mechanical stimulation of the gut is transduced by visceral nociceptors <xref rid="b0040" ref-type="bibr">[8]</xref>, <xref rid="b0080" ref-type="bibr">[16]</xref>, <xref rid="b0190" ref-type="bibr">[38]</xref>. We used vFh probes to elicit brief, intense, and graded mechanical stimuli. These responses were sensitized in both serosal and mesenteric afferents after application of IS, an effect that is absent in Na_V1.9^{&#x02212;/&#x02212;} mice. Furthermore, the response to noxious ramp distension was also sensitized by IS, which, at least during lower distension pressures, was dependent on Na_V1.9. This is consistent with the proposed role for Na_V1.9 in regulating afferent nerve terminal sensitization during inflammation <xref rid="b0155" ref-type="bibr">[31]</xref>.</p><p id="p0155">We observed reduced baseline mechanosensitivity in mesenteric but not serosal or muscular afferents, suggesting that Na_V1.9 may contribute to the regulation of excitability in this afferent subtype. Interestingly, mechanosensitivity is reduced in serosal, but not mesenteric, afferents in ASIC2^{&#x02212;/&#x02212;} mice <xref rid="b0190" ref-type="bibr">[38]</xref>, <xref rid="b0240" ref-type="bibr">[48]</xref>. It remains to be seen whether differences in the expression of mechanotransducers or other known regulators of neuronal excitability (such as T-type calcium channels or HCN2) present in serosal vs mesenteric afferents contributes to the importance of Na_V1.9 in regulating afferent firing remains <xref rid="b0060" ref-type="bibr">[12]</xref>, <xref rid="b0230" ref-type="bibr">[46]</xref>.</p><p id="p0160">Next we showed that the activation of visceral afferents by an initial rapid phasic distension is unaffected by Na_V1.9 deletion. This observation supports the data generated using high-intensity vFh (2&#x000a0;g) stimulation which produced comparable responses across all types of units regardless of genotype. It suggests that the reduction in afferent excitability caused the loss of Na_V1.9 may be overcome by greater mechanical stimulation, and presumably greater activation of stimulus-transducing channels. In contrast to the initial phasic distension, Na_V1.9 is required for the maintenance of afferent sensitivity to persistent mechanical stimuli. The desensitization in responses typically seen to repeat colorectal distension <xref rid="b0110" ref-type="bibr">[22]</xref>, <xref rid="b0235" ref-type="bibr">[47]</xref> did not stabilize in Na_V1.9^{&#x02212;/&#x02212;} mice but instead continued to decline with subsequent distensions. One explanation for this observation that is worthy of further investigation is that repeated mechanical stimuli may progressively desensitise activity in mechanosensitive channels, such as TRPV4, TRPA1, and ASIC3; however, a concurrent up-regulation of Na_V1.9 maintains a consistent level of afferent activation. This mechanism could be crucial in vivo, where pain is evoked by repeated contractions of the colon around a bolus or stricture <xref rid="b0045" ref-type="bibr">[9]</xref>. Consistent with this proposed role for Na_V1.9 in persistent or sustained stimuli, we observed substantial reductions in the afferent response to ramp colorectal distension in Na_V1.9^{&#x02212;/&#x02212;} tissue at all pressures, including those within the noxious range. When these pressures were extended into the supramaximal range (100&#x02013;150&#x000a0;mm&#x000a0;Hg), increases in afferent activity were observed in both genotypes. This is in keeping with our previous suggestion that increased stimulus strength may overcome the loss of excitability after deletion of Na_V1.9.</p><p id="p0165">Considerable efforts were taken to ensure that changes observed in this study were not confounded by the loss of Na_V1.9 within intrinsic primary afferents neurones (IPANs) of the enteric nervous system (ENS) <xref rid="b0185" ref-type="bibr">[37]</xref>, <xref rid="b0215" ref-type="bibr">[43]</xref>. Potential impact of the loss of Na_V1.9 in IPANs on local motor reflexes was controlled by inhibiting smooth muscle contractility through the addition of nifedipine and atropine to the perfusion buffer. To reduce the probability of recording activity from intestinofugal neurones within the ENS, our recordings were made distil to the primary site of intestinofugal afferent termination in the inferior mesenteric ganglia <xref rid="b0145" ref-type="bibr">[29]</xref>. We believe these steps, in conjunction with the magnitude of the effects observed and paucity of evidence for cross-talk between the ENS and extrinsic afferents, strongly suggest that the changes we observed are due to modulation of neuronal excitability by Na_V1.9 in visceral extrinsic afferents <xref rid="b0165" ref-type="bibr">[33]</xref>. Visceral afferents responsive to noxious stimuli are relatively unspecialized anatomically and therefore require mechanisms for regulating neuronal excitation induced by a range of innocuous and noxious mechanical and inflammatory stimuli <xref rid="b0240" ref-type="bibr">[48]</xref>. Indeed, visceral nociceptors, in particular serosal or mesenteric afferents, are prone to modulation by inflammatory mediators, and they may be recruited after being previously silent <xref rid="b0025" ref-type="bibr">[5]</xref>, <xref rid="b0035" ref-type="bibr">[7]</xref>, <xref rid="b0075" ref-type="bibr">[15]</xref>, <xref rid="b0080" ref-type="bibr">[16]</xref>. Our results suggest that Na_V1.9 represents a significant mechanism responsible for regulating visceral nociceptor sensitivity to sustained inflammatory and mechanical stimuli (<xref rid="f0050" ref-type="fig">Fig. 10</xref>). The recent identification of gain-of-function <italic>SCN11A</italic> mutations in humans resulting in both familial episodic pain and painful neuropathy provides important evidence indicating that Na_V1.9 is present in pain-sensing nerves in humans and that modulating Na_V1.9 function produces pain <xref rid="b0100" ref-type="bibr">[20]</xref>, <xref rid="b0270" ref-type="bibr">[54]</xref>.<fig id="f0050"><label>Fig. 10</label><caption><p>Schematic of Na_V1.9 interacting with membrane-bound receptors and ion channels in a visceral afferent terminal. When present at the visceral afferent terminal, Na_V1.9, as well as contributing to setting the resting membrane potential, acts to amplify generator potentials evoked by mechanosensitive channels and functions as a key transducer of inflammatory mediators and other sensitizing stimuli.</p></caption><graphic xlink:href="gr10"/></fig></p><p id="p0170">In summary, our data support an important physiological role for Na_V1.9 in the regulation of visceral afferent sensitivity during inflammation, and one that will likely have clinical relevance to human visceral pain. Further studies are required to fully understand how inflammatory pathways interact with Na_V1.9 within the afferent fibre. The role of distinct sodium channel activity in human afferent function has been indirectly addressed here by combining human ex vivo samples with the use of null mutant mice. This approach allows us to conclude that Na_V1.9 is an important target for the treatment of visceral pain.</p></sec><sec id="s0105"><title>Conflict of interest statement</title><p id="p0175">WW is an employee of Neusentis (Pfizer Ltd). The other authors report no conflict of interest.</p></sec>

Anti-hyperalgesic effects of a novel TRPM8 agonist in neuropathic rats: A comparison with topical menthol

Could not extract abstract

<contrib contrib-type="author" id="au005"><name><surname>Patel</surname><given-names>Ryan</given-names></name><email>ryan.patel.10@ucl.ac.uk</email><xref rid="af005" ref-type="aff">a</xref><xref rid="cor1" ref-type="corresp">&#x0204e;</xref></contrib><contrib contrib-type="author" id="au010"><name><surname>Gon&#x000e7;alves</surname><given-names>Leonor</given-names></name><xref rid="af005" ref-type="aff">a</xref></contrib><contrib contrib-type="author" id="au015"><name><surname>Leveridge</surname><given-names>Mathew</given-names></name><xref rid="af010" ref-type="aff">b</xref></contrib><contrib contrib-type="author" id="au020"><name><surname>Mack</surname><given-names>Stephen R.</given-names></name><xref rid="af010" ref-type="aff">b</xref></contrib><contrib contrib-type="author" id="au025"><name><surname>Hendrick</surname><given-names>Alan</given-names></name><xref rid="af010" ref-type="aff">b</xref></contrib><contrib contrib-type="author" id="au030"><name><surname>Brice</surname><given-names>Nicola L.</given-names></name><xref rid="af010" ref-type="aff">b</xref></contrib><contrib contrib-type="author" id="au035"><name><surname>Dickenson</surname><given-names>Anthony H.</given-names></name><xref rid="af005" ref-type="aff">a</xref></contrib><aff id="af005"><label>a</label>Department of Neuroscience, Physiology and Pharmacology, University College London, London, UK</aff><aff id="af010"><label>b</label>Takeda Cambridge Ltd., Cambridge, UK</aff>

Pain

<sec id="s0005"><label>1</label><title>Introduction</title><p id="p0005">Cold temperatures can be refreshing and relieving, but can also evoke sensations of aching, burning, and pricking. Disturbances in cold sensitivity are a feature of a range of neuropathic conditions including complex regional pain syndrome, chemotherapy-induced neuropathy, and trigeminal neuralgia <xref rid="b0010" ref-type="bibr">[2]</xref>, <xref rid="b0095" ref-type="bibr">[20]</xref>, <xref rid="b0215" ref-type="bibr">[44]</xref>. The cold mimetic compound menthol has been used to alleviate several painful conditions including musculoskeletal pain <xref rid="b0295" ref-type="bibr">[60]</xref>. A case study has suggested that high-concentration topical menthol reduces ongoing pain in a patient undergoing chemotherapy <xref rid="b0070" ref-type="bibr">[14]</xref>, but also cold and pinprick hyperalgesia after peripheral neuropathy in a small of group of patients <xref rid="b0230" ref-type="bibr">[47]</xref>, <xref rid="b0315" ref-type="bibr">[64]</xref>. High concentration menthol applied to uninjured areas has been reported to induce cold hyperalgesia, pinprick hyperalgesia, and in some instances secondary hyperalgesia <xref rid="b0040" ref-type="bibr">[8]</xref>, <xref rid="b0130" ref-type="bibr">[27]</xref>, <xref rid="b0320" ref-type="bibr">[65]</xref>.</p><p id="p0010">The cold and menthol sensitive channel TRPM8 has a multifaceted role in cold sensitivity. Selective fibre blocks suggest that A&#x003b4;-fibres are likely responsible for the cooling effects of menthol and cold temperatures, whereas C-fibres are implicated in the paradoxical burning sensation <xref rid="b0310" ref-type="bibr">[63]</xref>, <xref rid="b0320" ref-type="bibr">[65]</xref>. TRPM8 knockout mice display deficiencies of innocuous cold sensitivity and partly of noxious cold sensitivity <xref rid="b0030" ref-type="bibr">[6]</xref>, <xref rid="b0065" ref-type="bibr">[13]</xref>, <xref rid="b0090" ref-type="bibr">[19]</xref>, <xref rid="b0165" ref-type="bibr">[34]</xref>. The evidence for TRPA1 being a noxious cold sensor is less compelling <xref rid="b0025" ref-type="bibr">[5]</xref>, <xref rid="b0165" ref-type="bibr">[34]</xref>, <xref rid="b0180" ref-type="bibr">[37]</xref>, and may have a role only in cold hyperalgesia <xref rid="b0240" ref-type="bibr">[49]</xref>. Both menthol and cooling induced analgesia are dependent on TRPM8 <xref rid="b0200" ref-type="bibr">[41]</xref>, <xref rid="b0270" ref-type="bibr">[55]</xref>, although a TRPM8 negative population of cold sensitive primary afferents are also important. Although cooling is analgesic in both phases of the formalin test in wild-type mice, cooling-induced analgesia is absent in the first phase, but not in the second phase, after genetic ablation of TRPM8 <xref rid="b0090" ref-type="bibr">[19]</xref>. Through genetic or pharmacological approaches, TRPM8 has been implicated in the development of cold hypersensitivity in several pre-clinical models <xref rid="b0065" ref-type="bibr">[13]</xref>, <xref rid="b0080" ref-type="bibr">[17]</xref>, <xref rid="b0170" ref-type="bibr">[35]</xref>. Thus, both activating and attenuating the functional roles of TRPM8 channels could be helpful in reducing pain. There are similarities with the TRPV1 channel, a member of the same family of TRP channels, where the agonist, capsaicin, can produce both pain and pain relief and where antagonists are being developed as possible analgesics <xref rid="b0235" ref-type="bibr">[48]</xref>.</p><p id="p0015">Clinical trials involving topical menthol are currently being undertaken for chronic pain conditions including osteoarthritis (clinical trials identifier: <ext-link ext-link-type="ClinicalTrials.gov" xlink:href="NCT01565070" id="ir0005">NCT01565070</ext-link>), muscle pain (<ext-link ext-link-type="ClinicalTrials.gov" xlink:href="NCT01542827" id="ir0010">NCT01542827</ext-link>), and chemotherapy-induced neuropathy (<ext-link ext-link-type="ClinicalTrials.gov" xlink:href="NCT01855607" id="ir0015">NCT01855607</ext-link>). In this study, we have addressed whether the menthol-induced hyperalgesia model in humans is replicable in a rodent model by combining behavioural tests and recordings from WDR neurones in the deep dorsal horn, which integrate low-threshold and noxious mechanical and thermal stimuli. We have also examined whether menthol induces similar analgesic effects in a rodent model of neuropathy, and finally, the systemic effects of a novel TRPM8 agonist, M8-Ag, on cold sensitivity in sham-operated and SNL rats.</p></sec><sec id="s0010"><label>2</label><title>Methods</title><sec id="s0015"><label>2.1</label><title>Animals</title><p id="p0020">Male Sprague-Dawley rats (250&#x02013;300&#x000a0;g; Biological Services, UCL, UK) were used for behavioural and electrophysiological experiments. Animals were group housed on a 12&#x000a0;h:12&#x000a0;h light&#x02013;dark cycle; food and water were available ad libitum. All procedures described here were approved by the Home Office, adhered to the Animals (Scientific Procedures) Act 1986, and were designed to reduce numbers and undue suffering in accordance with IASP ethics guidelines <xref rid="b0350" ref-type="bibr">[71]</xref>.</p></sec><sec id="s0020"><label>2.2</label><title>Calcium imaging</title><p id="p0025">TRPM8 agonist activity was determined by measuring changes in intracellular calcium levels using a Ca²⁺-sensitive fluorescent dye. The changes in fluorescent signal were monitored by FLIPR (Molecular Devices, UK). HEK293 cells stably expressing human TRPM8 were seeded in cell culture medium in black, clear-bottom poly-<sc>d</sc>-lysine-coated 384-well plates 24&#x000a0;hours before the assay (BD Biosciences, Oxford, UK) and grown overnight at 37&#x000a0;&#x000b0;C, 5% CO₂. On the day of the assay, cell culture media were removed and cells were loaded with Calcium 4 Dye (Molecular Devices, UK) for 1&#x000a0;hour at 37&#x000a0;&#x000b0;C, 5% CO₂. M8-Ag (4-[5-(4-chlorophenyl)-4-phenyl-4H-1,2,4-triazol-3-yl]morpholine; synthesised in-house) was added to cells and monitored on the FLIPR for 80&#x000a0;seconds. TRPM8-mediated increases in intracellular Ca²⁺ concentration was readily detected upon activation with compound and normalised to a positive control (1&#x000a0;&#x003bc;mol/L icilin, EC₁₀₀). The EC₅₀ values were determined from an 8-point half-log concentration&#x02013;response curve, generated using the average of 2 wells for each data point.</p></sec><sec id="s0025"><label>2.3</label><title>Spinal nerve ligation surgery</title><p id="p0030">Spinal nerve ligation (SNL) surgery was performed as described elsewhere <xref rid="b0140" ref-type="bibr">[29]</xref>. Rats (125&#x02013;135&#x000a0;g) were maintained under 2% v/v isoflurane anaesthesia delivered in a 3:2 ratio of nitrous oxide and oxygen. Under aseptic conditions, a paraspinal incision was made and the left tail muscle excised. Part of the L5 transverse process was removed to expose the L5 and L6 spinal nerves, which were then isolated with a glass nerve hook (Ski-Ry Ltd, London, UK) and ligated with a nonabsorbable, 6-0 braided silk thread proximal to the formation of the sciatic nerve. The surrounding skin and muscle were closed with absorbable 3-0 sutures. All rats groomed and gained weight normally in the days that followed after surgery.</p></sec><sec id="s0030"><label>2.4</label><title>Behavioural testing</title><p id="p0035">Behavioural testing of SNL rats was performed 14 days after surgery. Rats were placed inside Perspex chambers on a wire mesh floor and allowed to acclimatise. Using the up&#x02013;down method <xref rid="b0055" ref-type="bibr">[11]</xref>, 50% withdrawal thresholds were determined with von Frey filaments (Touch-Test; North Coast Medical, Gilroy, CA) proving forces of 1.4&#x000a0;g, 2&#x000a0;g, 4&#x000a0;g, 6&#x000a0;g, 8&#x000a0;g, 10&#x000a0;g, and 15&#x000a0;g. Filaments were applied until they buckled for 5 to 6&#x000a0;seconds. All rats had mechanical hypersensitivity defined as at least a 50% difference in paw withdrawal threshold between ipsilateral and contralateral paws. Cold hypersensitivity was tested by applying a drop of acetone to the plantar surface of the paw using a modified 1-mL syringe. Acetone was applied 5 times, and rats were allowed time to recover between applications. Flinching, licking, shaking, or other directed behaviours were considered positive responses to cold or mechanical stimulation. For the M8-Ag study, only rats with significant cold hypersensitivity were used for analysis, which was defined as 2 or more withdrawals out of 5 on the nerve-injured side (4 of 18 rats were excluded). Rats were then dosed intraperitoneally with vehicle (85% normal saline solution, 10% cremophor (Sigma, UK) 5% dimethylsulfoxide (Sigma, UK), or 30 mg/kg M8-Ag in dose volumes of 2&#x000a0;mL/kg. Behavioural testing was repeated 30, 60, and 90&#x000a0;minutes after dosing. The experimenter was blinded to the treatment during behavioural testing. For the menthol study, 150&#x000a0;&#x003bc;L 1%, 10%, and 40% <sc>l</sc>-menthol (Alfa Aesar, Heysham, UK) (in absolute ethanol) were cumulatively applied to the entire plantar surface of the left hind paw. Behavioural testing was repeated 15&#x000a0;minutes after each application.</p></sec><sec id="s0035"><label>2.5</label><title>In vivo electrophysiology</title><p id="p0040">In vivo electrophysiology was conducted as previously described <xref rid="b0300" ref-type="bibr">[61]</xref>. Spinal nerve-ligated and sham-operated rats were used between days 15 and 18 post surgery. Animals were anaesthetised with 3.5% v/v isoflurane delivered in 3:2 ratio of nitrous oxide and oxygen. Once areflexic, a tracheotomy was performed and rats were subsequently maintained on 1.5% v/v isoflurane for the remainder of the experiment. Rats were secured in a stereotaxic frame, and a laminectomy was performed to expose L4 to L5 segments of the spinal cord. Extracellular recordings were made from deep dorsal horn wide dynamic range (WDR) spinal neurones (lamina V/VI) with receptive fields on the glabrous skin of the toes using parylene-coated tungsten electrodes (A-M Systems, Sequim, WA).</p><p id="p0045">Electrical stimulation of WDR neurones was delivered transcutaneously via needles inserted into the receptive field. A train of 16 electrical stimuli (2-millisecond pulses, 0.5&#x000a0;Hz) was applied at 3 times the threshold current for C-fibre activation. Responses evoked by A&#x003b2;- (0&#x02013;20&#x000a0;milliseconds), A&#x003b4;- (20&#x02013;90&#x000a0;milliseconds), and C-fibres (90&#x02013;350&#x000a0;milliseconds) were separated and quantified on the basis of latency. Neuronal responses occurring after the C-fibre latency band were classed as post-discharge. The input and the wind-up were calculated as Input&#x000a0;=&#x000a0;(action potentials evoked by first pulse)&#x000a0;&#x000d7;&#x000a0;total number of pulses (16), wind-up&#x000a0;=&#x000a0;(total action potentials after 16 train stimulus)&#x000a0;&#x02212;&#x000a0;Input. The receptive field was also stimulated using a range of natural stimuli (brush, von Frey filaments [2&#x000a0;g, 8&#x000a0;g, 15&#x000a0;g, 26&#x000a0;g, and 60&#x000a0;g and heat [35&#x000a0;&#x000b0;C, 42&#x000a0;&#x000b0;C, 45&#x000a0;&#x000b0;C, and 48&#x000a0;&#x000b0;C]) applied over a period of 10&#x000a0;seconds per stimulus and the evoked response quantified. The heat stimulus was applied with a constant water jet onto the centre of the receptive field. A 100-&#x003bc;L quantity of acetone and of ethyl chloride (Miller Medical Supplies, UK) were applied as an evaporative innocuous cooling and a noxious cooling stimulus, respectively. Both acetone and ethyl chloride have been previously demonstrated to reliably and reproducibly induce cooling to temperatures comparable to the boundaries for the perception of innocuous and noxious cold. Ethyl chloride is also able to evoke a withdrawal reflex <xref rid="b0185" ref-type="bibr">[38]</xref>. The neuronal response to room-temperature water was subtracted from acetone and ethyl chloride-evoked responses to control for concomitant mechanical stimulation during application. Receptive fields were determined before electrical stimulation using 8-g and 60-g von Frey filaments. An area was considered part of the receptive field if a response of greater than 50&#x000a0;Hz was obtained. A rest period of 30&#x000a0;seconds between applications was used to avoid sensitisation. Receptive field sizes are expressed as a percentage area of a standardised paw measured using ImageJ (National Institutes of Health, Bethesda, MD).</p><p id="p0050">Data were captured and analysed by a Cambridge Electronic Design 1401 interface coupled to a computer with Spike2 software (Cambridge, UK) with post-stimulus time histogram and rate functions. After 3 consecutive stable baseline responses to natural stimuli (&#x0003c;10% variation; data were averaged to give control values), sham-operated and SNL animals were injected subcutaneously into the contralateral flank with 30&#x000a0;mg/kg M8-Ag. Responses to electrical and natural stimuli were measured 30&#x000a0;minutes post dosing and then every 20&#x000a0;minutes for the next 80&#x000a0;minutes. <sc>l</sc>-menthol (1%, 10%, and 40%; Alfa Aesar, Heysham, UK) (in absolute ethanol) were cumulatively applied to the receptive field in naive and SNL rats (approximately 20&#x02013;30&#x000a0;&#x003bc;L). Neuronal responses were tested 15&#x000a0;minutes later. One neurone was characterised per rat.</p></sec><sec id="s0040"><label>2.6</label><title>Statistical analysis</title><p id="p0055">Statistical analyses were performed using SPSS v22 (IBM, Armonk, NY). For in vivo electrophysiology measures, statistical significance was tested using a 1-way or 2-way repeated-measures (RM) analysis of variance (ANOVA), followed by a Bonferroni post hoc test for paired comparisons. Sphericity was tested using Mauchly&#x02019;s test, and the Greenhouse-Geisser correction was applied if violated. Behavioural time courses and receptive field sizes were tested with the Friedman test, followed by a Wilcoxon post hoc and Bonferroni correction for paired comparisons. Area under the curve (AUC) values were calculated using the trapezoid rule and compared with a 1-way ANOVA followed by a Bonferroni post hoc test.</p></sec></sec><sec id="s0045"><label>3</label><title>Results</title><sec id="s0050"><label>3.1</label><title>Summary of behavioural and neuronal responses in naive, sham, and SNL rats</title><p id="p0060">Behavioural sensitivity to mechanical and cooling stimuli was examined 14&#x000a0;days post sham or SNL surgery and in age-matched naive rats. SNL rats displayed guarding behaviour of the injured ipsilateral hind paw, which was absent on the uninjured contralateral side and in sham-operated rats. SNL rats used in this study, but not sham or naive rats, displayed significantly reduced withdrawal thresholds to punctate mechanical stimulation (Wilcoxon test, <italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.05) (<xref rid="f0005" ref-type="fig">Fig. 1</xref>A) and increased withdrawals to acetone-induced innocuous cooling (Wilcoxon test, <italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.05) (<xref rid="f0005" ref-type="fig">Fig. 1</xref>B) compared to contralateral responses.<fig id="f0005"><label>Fig. 1</label><caption><p>Behavioural hypersensitivity in spinal nerve-ligated (SNL) rats 14 days after surgery. (A) Unilateral ligation of L5 and L6 spinal nerves reduces mechanical withdrawal thresholds and (B) increases responsiveness to acetone-induced innocuous cooling in SNL rats. Sham-operated rats exhibited behavioural sensitivity similar to that in naive rats. Data represent mean&#x000a0;&#x000b1;&#x000a0;SEM. ^{&#x02217;&#x02217;&#x02217;}<italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.001. Naive n&#x000a0;=&#x000a0;7, sham n&#x000a0;=&#x000a0;7, SNL n&#x000a0;=&#x000a0;25. PWT, paw withdrawal threshold.</p></caption><graphic xlink:href="gr1"/></fig></p><p id="p0065">In vivo electrophysiology was performed to examine the effects of either menthol or M8-Ag on nociceptive processing by wide dynamic range (WDR) spinal neurones in neuropathic and uninjured conditions. Neurones were characterised from depths relating to lamina V/VI in the deep dorsal horn (naive 862&#x000a0;&#x000b1;&#x000a0;69&#x000a0;&#x003bc;m, sham 704&#x000a0;&#x000b1;&#x000a0;66&#x000a0;&#x003bc;m, and SNL 691&#x000a0;&#x000b1;&#x000a0;34&#x000a0;&#x003bc;m) <xref rid="b0325" ref-type="bibr">[66]</xref>, and were confirmed as WDR on the basis of responses to noxious heat, dynamic brushing, and noxious punctate mechanical stimulation. No significant difference was observed in the electrical thresholds for activation of A- or C-fibres or for electrically evoked neuronal responses (1-way ANOVA, <italic>P</italic>&#x000a0;&#x0003e;&#x000a0;.05). Thermal and mechanical coding of WDR neurones were also similar among naive, sham, and SNL rats (2-way ANOVA, <italic>P</italic>&#x000a0;&#x0003e;&#x000a0;.05) (<xref rid="s0100" ref-type="sec">Table S1</xref>).</p></sec><sec id="s0055"><label>3.2</label><title>Menthol (10% and 40%) alleviates cold hypersensitivity in SNL rats</title><p id="p0070">Increasing concentrations of menthol were unilaterally and cumulatively applied to the left hind paw of SNL rats 14&#x000a0;days post surgery and weight/age matched naive rats. Sporadic flinching/shaking behaviours were infrequently observed after application of menthol (naive 2/7, SNL 1/7). In naive rats, no significant sensitisation (2/7 rats sensitised) to acetone-induced innocuous cooling was observed after application of low and high concentrations of menthol (Friedman test, <italic>P</italic>&#x000a0;&#x0003e;&#x000a0;.05) (<xref rid="f0010" ref-type="fig">Fig. 2</xref>A). In contrast, high concentrations of topical menthol reduced acetone-induced withdrawals in SNL rats (Friedman test, <italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.05, followed by Dunn&#x02019;s post hoc) (<xref rid="f0010" ref-type="fig">Fig. 2</xref>A). Electrophysiological recordings of dorsal horn WDR neurones were made to examine spinal processing of threshold and supra-threshold stimuli. In naive rats, menthol had no overall effect on neuronal responses to acetone-induced innocuous cooling or ethyl chloride-induced noxious cooling (1-way RM ANOVA, <italic>P</italic>&#x000a0;&#x0003e;&#x000a0;.05) (<xref rid="f0010" ref-type="fig">Fig. 2</xref>C). In a further 6 non-acetone-responsive neurones recorded from, only 1 developed cooling sensitivity (data not shown). Only a weak trend towards reduced acetone evoked responses by 40% menthol was observed in SNL rats, with no effect on noxious cold responses (1-way RM ANOVA, <italic>P</italic>&#x000a0;&#x0003e;&#x000a0;.05) (<xref rid="f0010" ref-type="fig">Fig. 2</xref>E). No ongoing neuronal activity was observed after menthol application in either SNL or naive rats.<fig id="f0010"><label>Fig. 2</label><caption><p>Behavioural and neuronal correlates of mechanical and cold sensitivity after menthol application in naive and spinal nerve-ligated (SNL) rats. (A) Menthol alleviates cold hypersensitivity in SNL rats, with no significant effect in naive rats (n&#x000a0;=&#x000a0;7). Corresponding neuronal responses to innocuous and noxious evaporative cooling in (C) naive rats (n&#x000a0;=&#x000a0;7) and (D) SNL rats (n&#x000a0;=&#x000a0;6). (B) Menthol does not significantly affect mechanical withdrawal thresholds in naive or SNL rats. Neuronal responses to mechanical stimulation in (D) naive and (F) SNL rats corroborate behavioural observations. Data represent mean&#x000a0;&#x000b1;&#x000a0;standard error of the mean. ^&#x02217;<italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.05. PWT, paw withdrawal threshold.</p></caption><graphic xlink:href="gr2"/></fig></p><p id="p0075">Menthol had minimal effects on mechanical sensitivity in naive rats (Friedman test, <italic>P</italic>&#x000a0;&#x0003e;&#x000a0;.05) (<xref rid="f0010" ref-type="fig">Fig. 2</xref>B), a feature that was mirrored by a lack of effect on mechanical coding of dorsal horn WDR neurones (2-way RM ANOVA, <italic>P</italic>&#x000a0;&#x0003e;&#x000a0;.05) (<xref rid="f0010" ref-type="fig">Fig. 2</xref>D). In SNL rats, menthol was unable to attenuate significantly mechanical hypersensitivity (Friedman test, <italic>P</italic>&#x000a0;&#x0003e;&#x000a0;.05) (<xref rid="f0010" ref-type="fig">Fig. 2</xref>B), although a subset (3/7) appeared to exhibit a small increase in withdrawal thresholds. Similarly, neuronal responses to mechanical stimulation were also unaltered (2-way RM ANOVA, <italic>P</italic>&#x000a0;&#x0003e;&#x000a0;.05) (<xref rid="f0010" ref-type="fig">Fig. 2</xref>F). Ethanol alone had no effect on neuronal responses (data not shown).</p></sec><sec id="s0060"><label>3.3</label><title>Menthol does not induce sensitisation of deep dorsal horn neurones in naive rats</title><p id="p0080">We also examined the effect of menthol on features of neuronal excitability. High-concentration menthol did not induce an expansion of receptive field sizes to innocuous (8&#x000a0;g) or noxious (60&#x000a0;g) punctate mechanical stimulation in naive rats (Friedman test, <italic>P</italic>&#x000a0;&#x0003e;&#x000a0;.05) (<xref rid="f0015" ref-type="fig">Fig. 3</xref>A) and were also unaltered in SNL rats (Friedman test, <italic>P</italic>&#x000a0;&#x0003e;&#x000a0;.05) (<xref rid="f0015" ref-type="fig">Fig. 3</xref>B). In addition, menthol had no influence on neuronal responses to heat stimulation (2-way RM ANOVA, <italic>P</italic>&#x000a0;&#x0003e;&#x000a0;.05) (<xref rid="f0015" ref-type="fig">Fig. 3</xref>C, D) or dynamic brushing (1-way RM ANOVA, <italic>P</italic>&#x000a0;&#x0003e;&#x000a0;.05) (<xref rid="f0015" ref-type="fig">Fig. 3</xref>E, F) in naive or SNL rats. Furthermore, menthol did not sensitise A&#x003b4;- and C-fibre-evoked responses or increase the wind-up and post-discharge of WDR neurones in naive rats (1-way RM ANOVA, <italic>P</italic>&#x000a0;&#x0003e;&#x000a0;.05) (<xref rid="f0015" ref-type="fig">Fig. 3</xref>G). After nerve injury, menthol was unable to reduce evoked activity attributed to primary afferent fibres or subsequent wind-up of spinal neurones (1-way RM ANOVA, <italic>P</italic>&#x000a0;&#x0003e;&#x000a0;.05) (<xref rid="f0015" ref-type="fig">Fig. 3</xref>H).<fig id="f0015"><label>Fig. 3</label><caption><p>Neuronal responses to natural and electrical stimulation of the receptive field after menthol application in naive and spinal nerve-ligated (SNL) rats. After topical menthol application, no change in receptive field size was observed (A, B) or in heat (C, D), dynamic brush (E, F), or electrically (G, H) evoked responses in naive rats (left panels) or SNL rats (right panels). Data represent mean&#x000a0;&#x000b1;&#x000a0;standard error of the mean. Naive n&#x000a0;=&#x000a0;7, SNL n&#x000a0;=&#x000a0;6. I, input; PD, post-discharge; RF, receptive field; WU, wind-up.</p></caption><graphic xlink:href="gr3"/></fig></p></sec><sec id="s0065"><label>3.4</label><title>M8-Ag evokes TRPM8-mediated Ca²⁺ currents in vitro</title><p id="p0085">We next examined the in vitro and in vivo effects of a novel TRPM8 agonist. The ability of M8-Ag (<xref rid="f0020" ref-type="fig">Fig. 4</xref>A) to evoke calcium currents was investigated in HEK293 cells stably expressing TRPM8 channels. M8-Ag activated TRPM8 channels in a dose-dependent manner with an EC₅₀ of 44.97&#x000a0;&#x000b1;&#x000a0;1.2&#x000a0;nmol/L (<xref rid="f0020" ref-type="fig">Fig. 4</xref>B). M8-Ag also activated TRPA1 but with an EC₅₀ of &#x0003e;4000&#x000a0;nmol/L (<xref rid="s0100" ref-type="sec">Fig. S1A</xref>). Pharmacokinetic analysis revealed after an intraperitoneal injection of 10 mg/kg M8-Ag in rats, the average peak blood concentration was 4.11&#x000a0;&#x003bc;g/mL (12.06&#x000a0;&#x003bc;mol/L) after 78&#x000a0;minutes with a half-life of 2.6&#x000a0;hours (<xref rid="s0100" ref-type="sec">Fig. S1B</xref>).<fig id="f0020"><label>Fig. 4</label><caption><p>(A) Chemical structure of M8-Ag (4-[5-(4-chlorophenyl)-4-phenyl-4H-1,2,4-triazol-3-yl]morpholine). (B) M8-Ag activates TRPM8 stably expressed in HEK293 cells in a dose-dependent manner. Data are expressed as mean and range of 2 wells.</p></caption><graphic xlink:href="gr4"/></fig></p></sec><sec id="s0070"><label>3.5</label><title>Neuronal responses to innocuous and noxious cooling attenuated by M8-Ag in SNL rats</title><p id="p0090">In vivo electrophysiology was performed to examine the effects of systemically dosed M8-Ag on cold sensitive neurones within an integrated system. After obtaining stable baseline neuronal recordings, rats were dosed subcutaneously with 30 mg/kg M8-Ag. In SNL rats, compared to baseline, M8-Ag inhibited neuronal responses to noxious cold stimulation with significantly reduced responses at 30, 70, and 110&#x000a0;minutes after dosing (1-way RM ANOVA, <italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.01, followed by Bonferroni post hoc test) (<xref rid="f0025" ref-type="fig">Fig. 5</xref>A). In addition, M8-Ag decreased overall neuronal responses to innocuous cooling with acetone (1-way RM ANOVA, <italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.05) (<xref rid="f0025" ref-type="fig">Fig. 5</xref>A). These effects of M8-Ag were absent in sham-operated rats (1-way RM ANOVA, <italic>P</italic>&#x000a0;&#x0003e;&#x000a0;.05) (<xref rid="f0025" ref-type="fig">Fig. 5</xref>B). Furthermore, M8-Ag did not alter neuronal responses to heat stimulation (2-way RM ANOVA, <italic>P</italic>&#x000a0;&#x0003e;&#x000a0;.05) (<xref rid="f0025" ref-type="fig">Fig. 5</xref>C, D), punctate mechanical stimulation (2-way RM ANOVA, <italic>P</italic>&#x000a0;&#x0003e;&#x000a0;.05) (<xref rid="f0025" ref-type="fig">Fig. 5</xref>E, F) or dynamic brushing of the receptive field (1-way RM ANOVA, <italic>P</italic>&#x000a0;&#x0003e;&#x000a0;.05) (<xref rid="f0025" ref-type="fig">Fig. 5</xref>G, H) in either sham or SNL rats. Neuronal events evoked by global stimulation of A&#x003b2;-, A&#x003b4;-, and C-fibres were also unaffected in SNL and sham-operated rats, indicating that activating TRMP8 does not increase overall afferent drive under normal or pathological conditions (1-way RM ANOVA, <italic>P</italic>&#x000a0;&#x0003e;&#x000a0;.05) (<xref rid="f0025" ref-type="fig">Fig. 5</xref>I, J). Vehicle alone in naive rats had no effect on mechanical, cold, heat, or electrically evoked responses (data not shown).<fig id="f0025"><label>Fig. 5</label><caption><p>M8-Ag inhibits deep dorsal horn lamina V/VI neuronal responses to innocuous and noxious cold stimulation in spinal nerve-ligated (SNL) rats. After 30&#x000a0;minutes postdosing and then at 20-minute intervals, recordings were made of neuronal responses to cold (A, B), heat (C, D), punctate mechanical (E, F), dynamic brush (G, H), and electrical stimuli (I, J). Left panels, SNL; right panels, sham. (+) represents statistically significant main effect (2-way repeated-measures analysis of variance); asterisks denote significant difference from baseline (Bonferroni post hoc). ^&#x02217;<italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.05, ^{&#x02217;&#x02217;}<italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.01, n&#x000a0;=&#x000a0;7. I, input; PD, post-discharge; WU, wind-up.</p></caption><graphic xlink:href="gr5"/></fig></p></sec><sec id="s0075"><label>3.6</label><title>M8-Ag reverses behavioural hypersensitivity to acetone-induced cooling in SNL rats</title><p id="p0095">The ability of M8-Ag to alleviate behavioural signs of cold hypersensitivity was examined in SNL rats 14 days after nerve ligation. Rats were dosed intraperitoneally with either 30&#x000a0;mg/kg M8-Ag or vehicle and tested up to 90&#x000a0;minutes after dosing. The majority of rats dosed with M8-Ag exhibited sporadic &#x02018;wet dog shakes&#x02019; up to 1&#x000a0;hour after dosing (7/9). These behaviours were absent in vehicle-treated rats. M8-Ag reduced the behavioural response to innocuous evaporative cooling of the injured paw compared to pre-drug responses (Friedman test <italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.01<italic>,</italic> followed by Wilcoxon post hoc and Bonferroni correction) (<xref rid="f0030" ref-type="fig">Fig. 6</xref>A), whereas the vehicle alone had no significant effect (Friedman test, <italic>P</italic>&#x000a0;&#x0003e;&#x000a0;.05<italic>)</italic> (<xref rid="f0030" ref-type="fig">Fig. 6</xref>A). Contralateral responses to acetone-induced cooling were minimal in comparison and were not affected by M8-Ag (Friedman test, <italic>P</italic>&#x000a0;&#x0003e;&#x000a0;.05) (<xref rid="f0030" ref-type="fig">Fig. 6</xref>B). Overall behavioural signs of cold hypersensitivity are reduced by M8-Ag but not by vehicle treatment (1-way ANOVA <italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.01, followed by Bonferroni post hoc) (<xref rid="f0030" ref-type="fig">Fig. 6</xref><italic>C</italic>). The behavioural observations correlate with electrophysiological recordings of neuronal responses to acetone. Comparing the effects of 30&#x000a0;mg/kg M8-Ag on total neuronal events in uninjured and SNL rats between 30 and 110&#x000a0;minutes after dosing reveals that cooling-evoked responses were significantly reduced in SNL rats (unpaired Student <italic>t</italic> test, <italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.05) (<xref rid="f0030" ref-type="fig">Fig. 6</xref>D)<fig id="f0030"><label>Fig. 6</label><caption><p>M8-Ag selectively reduces behavioural and neuronal responses to cooling after spinal nerve ligation. (A) M8-Ag, 30 mg/kg, reversed the behavioural response to acetone-induced evaporative cooling on the nerve-injured ipsilateral side compared to pre-drug values (n&#x000a0;=&#x000a0;7), whereas vehicle alone had no significant effect (n&#x000a0;=&#x000a0;7). (B) Contralateral responses to innocuous cooling were not affected by either treatment. (C) Area under the curve (AUC) analysis confirms a significant attenuation of cold hypersensitivity by M8-Ag compared to vehicle treatment. (D) A decrease in overall neuronal responses to acetone after dosing of M8-Ag was also observed in spinal nerve-ligated (SNL) rats (n&#x000a0;=&#x000a0;7) compared to sham-operated rats (n&#x000a0;=&#x000a0;7). (E) M8-Ag did not alter mechanical withdrawal thresholds of the injured ipsilateral paw, or (F) the uninjured contralateral side (n&#x000a0;=&#x000a0;9). (G) AUC analysis confirms that rats exhibited significant mechanical hypersensitivity, which was not affected by M8-Ag or vehicle treatment. (H) Neuronal responses to mechanical stimulation in SNL rats (n&#x000a0;=&#x000a0;7) and sham-operated rats (n&#x000a0;=&#x000a0;7) were also unaffected by 30 mg/kg M8-Ag. Data represent mean&#x000a0;&#x000b1;&#x000a0;standard error of the mean. ^&#x02217;<italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.05, ^{&#x02217;&#x02217;}<italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.01, ^{&#x02217;&#x02217;&#x02217;}<italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.001. APs, action potentials; a.u., arbitrary units; PWT, paw withdrawal threshold.</p></caption><graphic xlink:href="gr6"/></fig></p><p id="p0100">M8-Ag did not alter mechanically evoked withdrawals on either the ipsilateral (<xref rid="f0030" ref-type="fig">Fig. 6</xref>E) or contralateral (<xref rid="f0030" ref-type="fig">Fig. 6</xref>F) side compared to pre-drug values (Friedman test, <italic>P</italic>&#x000a0;&#x0003e;&#x000a0;.05<italic>).</italic> AUC analysis also confirms no overall effect of either drug or vehicle treatment on mechanical sensitivity (1-way ANOVA, <italic>P</italic>&#x000a0;&#x0003c;&#x000a0;.001<italic>,</italic> followed by Bonferroni post hoc test) (<xref rid="f0030" ref-type="fig">Fig. 6</xref>G). Correspondingly, M8-Ag did not affect mechanical coding of WDR neurones to below threshold and threshold stimuli in SNL or sham-operated rats (2-way ANOVA, <italic>P</italic>&#x000a0;&#x0003e;&#x000a0;.05) (<xref rid="f0030" ref-type="fig">Fig. 6</xref>H).</p></sec></sec><sec id="s0080"><label>4</label><title>Discussion</title><p id="p0105">Several studies support menthol-induced hyperalgesia as a rapidly inducing surrogate model of cold hypersensitivity in human subjects <xref rid="b0040" ref-type="bibr">[8]</xref>, <xref rid="b0130" ref-type="bibr">[27]</xref>, <xref rid="b0230" ref-type="bibr">[47]</xref>, <xref rid="b0320" ref-type="bibr">[65]</xref>. In addition, menthol also possesses analgesic and counterirritant properties <xref rid="b0250" ref-type="bibr">[51]</xref>, <xref rid="b0315" ref-type="bibr">[64]</xref>. However, some discrepancies exist between these human studies focused on TRPM8, but also in animal models, as menthol and icilin have been proposed to have complex pro- and anti-nociceptive effects in both uninjured and injured animals <xref rid="b0045" ref-type="bibr">[9]</xref>, <xref rid="b0050" ref-type="bibr">[10]</xref>, <xref rid="b0100" ref-type="bibr">[21]</xref>, <xref rid="b0105" ref-type="bibr">[22]</xref>, <xref rid="b0115" ref-type="bibr">[24]</xref>, <xref rid="b0155" ref-type="bibr">[32]</xref>, <xref rid="b0160" ref-type="bibr">[33]</xref>, <xref rid="b0200" ref-type="bibr">[41]</xref>, <xref rid="b0270" ref-type="bibr">[55]</xref>. By combining behavioural measures and electrophysiological recordings of lamina V/VI neurones, we have further examined the translational value of this model in uninjured and neuropathic rats (summarised in <xref rid="t0005" ref-type="table">Table 1</xref> and <xref rid="t0010" ref-type="table">Table 2</xref>).<table-wrap position="float" id="t0005"><label>Table 1</label><caption><p>Comparison of effects of menthol in naive rats and normal human subjects.</p></caption><table frame="hsides" rules="groups"><thead><tr><th/><th>Rats</th><th colspan="5">Normal human subjects<hr/></th></tr><tr><th/><th>Naive</th><th>N&#x000a0;=&#x000a0;12</th><th>n&#x000a0;=&#x000a0;39</th><th>n&#x000a0;=&#x000a0;10</th><th>n&#x000a0;=&#x000a0;10</th><th>n&#x000a0;=&#x000a0;10</th></tr><tr><th>Increase in:</th><th/><th>Binder et al. (2011)</th><th>Hatem et al. (2006)</th><th>Wasner et al. (2004)</th><th>Namer et al. (2008)</th><th>Olsen et al. (2014)</th></tr></thead><tbody><tr><td>Dynamic mechanical</td><td>&#x02212;</td><td>+/&#x02212;</td><td>&#x02212;</td><td>&#x02212;</td><td>+/&#x02212;</td><td/></tr><tr><td>Punctate mechanical</td><td>&#x02212;</td><td>+</td><td>&#x02212;</td><td>+</td><td/><td>+</td></tr><tr><td>Cold</td><td>+/&#x02212;</td><td>+</td><td>+</td><td>+</td><td>+</td><td>+</td></tr><tr><td>Heat</td><td>&#x02212;</td><td>+</td><td>&#x02212;</td><td>&#x02212;</td><td>&#x02212;</td><td>&#x02212;</td></tr><tr><td>Wind-up</td><td>&#x02212;</td><td>&#x02212;</td><td/><td>&#x02212;</td><td/><td/></tr><tr><td>Secondary hyperalgesia/receptive field</td><td>&#x02212;</td><td>+</td><td/><td>+/&#x02212;</td><td/><td/></tr></tbody></table><table-wrap-foot><fn><p>+, Increased response frequently observed, +/&#x02212;, occasionally observed; &#x02212;, never/rarely observed; blank, not tested.</p></fn></table-wrap-foot></table-wrap><table-wrap position="float" id="t0010"><label>Table 2</label><caption><p>Comparison of effects of menthol in neuropathic rats and neuropathic patients.</p></caption><table frame="hsides" rules="groups"><thead><tr><th/><th>Rats</th><th colspan="4">Neuropathic pain patients<hr/></th></tr><tr><th/><th>SNL</th><th>PNP/PHN/CPSP n&#x000a0;=&#x000a0;8</th><th>CI n&#x000a0;=&#x000a0;12</th><th>CIN n&#x000a0;=&#x000a0;1</th><th>PHN n&#x000a0;=&#x000a0;1</th></tr><tr><th>Decrease in:</th><th/><th>Wasner et al. (2008)</th><th>Namer et al. (2008)</th><th>Colvin et al. (2008)</th><th><xref rid="b9000" ref-type="bibr">[16]</xref></th></tr></thead><tbody><tr><td>Dynamic mechanical</td><td>&#x02212;</td><td>&#x02212;</td><td>&#x02212;</td><td/><td>&#x02212;</td></tr><tr><td>Punctate mechanical</td><td>+/&#x02212;</td><td>+/&#x02212;</td><td>+/&#x02212;</td><td>+</td><td>&#x02212;</td></tr><tr><td>Cold</td><td>+</td><td>+</td><td>+/&#x02212;</td><td/><td/></tr><tr><td>Heat</td><td>&#x02212;</td><td/><td>&#x02212; (increased)</td><td/><td/></tr><tr><td>Wind-up</td><td>&#x02212;</td><td/><td/><td>+</td><td/></tr><tr><td>Receptive field</td><td>&#x02212;</td><td/><td/><td/><td/></tr><tr><td>Ongoing pain</td><td/><td/><td/><td>&#x02212;</td><td>+</td></tr></tbody></table><table-wrap-foot><fn><p>+, Decreased response frequently observed; +/&#x02212;, occasionally observed; &#x02212;, never/rarely observed; blank, not tested; CI, cold injury; CIN, chemotherapy induced neuropathy; CPSP, central post-stroke pain; PHN, post-herpetic neuralgia; PNP, polyneuropathy; spinal nerve-ligated.</p></fn></table-wrap-foot></table-wrap></p><p id="p0110">We have previously demonstrated that antagonism of the TRPM8 receptor inhibits behavioural and spinal neuronal responses to innocuous and noxious cooling in SNL but not naive rats <xref rid="b0260" ref-type="bibr">[53]</xref>. Surprisingly, the TRPM8 agonist reported here displayed identical effects. This would imply that TRPM8 is not essential for all forms of cold transduction in the absence of injury in rats, and that the effects of modulators of TRPM8 are dependent on changes induced by the pathophysiological state. Previous studies have demonstrated that menthol sensitises peripheral afferents and trigeminal neurones to cold stimulation <xref rid="b0135" ref-type="bibr">[28]</xref>, <xref rid="b0275" ref-type="bibr">[56]</xref>, <xref rid="b0330" ref-type="bibr">[67]</xref>, and that TRPM8 antagonists can reverse menthol-induced activity in peripheral nerve endings <xref rid="b0210" ref-type="bibr">[43]</xref>. Menthol does not appear consistently to induce ongoing pain-like behaviours in rats when applied to the paw, mirrored by an absence of ongoing spinal neuronal activity. Menthol may, however, render cool temperatures more aversive in temperature preference assays <xref rid="b0155" ref-type="bibr">[32]</xref>.</p><p id="p0115">Menthol-induced hyperalgesia was designed as a surrogate human model of cold hyperalgesia analogous to the capsaicin model, and some parallels may exist between the two. Hyperexcitability of dorsal horn neurones is considered the neural basis of capsaicin-induced hyperalgesia, in which a prolonged stimulus results in ongoing A&#x003b4;- and C-fibre activity and subsequent wind-up, central sensitisation, and secondary hyperalgesia (reviewed by O&#x02019;Neill et al. <xref rid="b0235" ref-type="bibr">[48]</xref>). Menthol does not appear to induce comparable changes in neuronal excitability in the rat, which in all likelihood reflects low overall afferent drive from TRPM8+ fibres. WDR neurones did not exhibit potentiated wind-up post menthol application; a similar lack of effect on a perceptual correlate of wind-up was observed in human subjects <xref rid="b0040" ref-type="bibr">[8]</xref>, <xref rid="b0320" ref-type="bibr">[65]</xref>. All studies in humans have demonstrated an increase in the cold pain threshold (i.e., cold pain experienced at warmer temperatures) after topical menthol with varied effects on other modalities (<xref rid="t0005" ref-type="table">Table 1</xref>). Binder et al. <xref rid="b0040" ref-type="bibr">[8]</xref> reported incidences of secondary hyperalgesia and pin-prick hyperalgesia as well as heat pain, although over longer time periods than other comparable studies. We did not observe any alterations in evoked neuronal activity by 10% or 40% menthol up to 80&#x000a0;minutes after application (data not shown).</p><p id="p0120">The impact of topical menthol in neuropathic patients has been less extensively examined compared to that in normal subjects. Heterogeneity in the mechanisms of cold allodynia complicates direct comparisons with each other and with pre-clinical models. As a case in point, menthol appears to exacerbate wind-up-like pain in amputees <xref rid="b0305" ref-type="bibr">[62]</xref>, whereas menthol reversed wind-up in a case of chemotherapy-induced neuropathy <xref rid="b0070" ref-type="bibr">[14]</xref>. The effects of topical menthol were examined in a nerve injury model exhibiting cold hypersensitivity (<xref rid="t0010" ref-type="table">Table 2</xref>). The analgesic properties of menthol in SNL rats appeared restricted to alleviating cooling hypersensitivity, although with no effect on innocuous or noxious cold-evoked neuronal responses. During behavioural characterisations, menthol was applied to the entire plantar surface, whereas with the electrophysiological characterisations, application was restricted to the receptive field of the neurone. Given the paucity of cutaneous TRPM8+ afferents <xref rid="b0285" ref-type="bibr">[58]</xref>, the spatial summation of menthol-evoked activity may be critical in mediating analgesia. Systemic M8-Ag alleviated cold hypersensitivity after SNL and reduced WDR neuronal responses to both innocuous and noxious cooling. The apparent lack of effect of menthol on noxious cold evoked neuronal activity could be attributed to the maximum concentration levels achievable through dermal diffusion. In neuropathic patients, however, the effects of menthol on noxious cold temperatures is difficult to ascertain, as pain is reported at threshold temperatures, and further decreases in temperature cannot be tolerated. These data are consistent with a restoration of spinal gating of cold temperatures at threshold levels by menthol but not of noxious temperatures.</p><p id="p0125">The effects of TRPM8 ligands could result from different consequences of activation of TRPM8 at peripheral levels, and this may be more apparent with topical application. Activation of cutaneous TRPV1 by capsaicin can be sensitising but can also have a desensitising effect due to a combination of acute desensitisation, tachyphylaxis, and withdrawal of epidermal nerve fibres <xref rid="b0235" ref-type="bibr">[48]</xref>. The mechanism of desensitisation differs between chemical agonists of TRPM8 <xref rid="b0175" ref-type="bibr">[36]</xref>. Systemic M8-Ag reversal of cold hypersensitivity and hyperalgesia could result from a depolarisation block of TRPM8+ afferents through Ca²⁺-dependent second messenger pathways downregulating TRPM8 activity to prevent calcium excitotoxicity through prolonged activation of channels <xref rid="b0075" ref-type="bibr">[15]</xref>, <xref rid="b0265" ref-type="bibr">[54]</xref>.</p><p id="p0130">There could also be central consequences, either as a result of peripheral TRPM8 targeting or actions on the receptor elsewhere, and these could be more marked with systemic dosing. A disinhibition hypothesis has been proposed to underlie the paradoxical burning associated with small decreases in temperature after nerve damage <xref rid="b0245" ref-type="bibr">[50]</xref>. If sensitisation of A&#x003b4;- and C-fibres underlies the hyperalgesia associated with topical menthol, in an already-sensitised state, rather than exacerbating cold hyperalgesia, one possibility is that activation of TRPM8-expressing afferents with menthol or M8-Ag recruits inhibitory circuits within the dorsal horn, restoring a loss of spinal inhibition <xref rid="b0315" ref-type="bibr">[64]</xref>. The paradox of a noxious stimulus inducing pain and analgesia according to context is also demonstrable with noxious cold temperatures applied to areas sensitised by capsaicin. The perception of relief could in part be determined by competing aversive and appetitive states converging on descending inhibitory pathways <xref rid="b0225" ref-type="bibr">[46]</xref>. Functional magnetic resonance imaging analysis suggests that the supraspinal integration of ascending spinal activity and the engagement of descending inhibitory pathways through the periaqueductal grey shapes relief or pain associated with topical menthol or cooling <xref rid="b0195" ref-type="bibr">[40]</xref>, <xref rid="b0225" ref-type="bibr">[46]</xref>, <xref rid="b0280" ref-type="bibr">[57]</xref>. Although not directly related to the present study, patients with post-stroke pain often have abnormal cold sensitivity, and this is thought to be due to damage to a cool-signaling lateral thalamic pathway that causes a disinhibition of a medial thalamic pathway promoting pain, resulting in the observed burning, cold, ongoing pain and cold allodynia <xref rid="b0125" ref-type="bibr">[26]</xref>. It is possible that peripheral and then spinal mechanisms via TRPM8 could finally have an impact on this system.</p><p id="p0135">In terms of central mechanisms, TRPM8+ afferents predominantly innervate the superficial dorsal horn, although some afferents terminate in the deeper laminae <xref rid="b0085" ref-type="bibr">[18]</xref>, <xref rid="b0285" ref-type="bibr">[58]</xref>. Cold temperatures activate a distinct subpopulation of spinothalamic and spinoparabrachial lamina I neurones described as &#x02018;cool&#x02019; responsive or &#x02018;polymodal-nociceptive&#x02019; <xref rid="b0005" ref-type="bibr">[1]</xref>, <xref rid="b0035" ref-type="bibr">[7]</xref>. These responses converge onto lamina V/VI neurones, which exhibit graded intensity-dependent responses to decreasing temperatures <xref rid="b0150" ref-type="bibr">[31]</xref>. Under normal conditions, local glycinergic and GABAergic inhibitory interneurones tonically control inhibitory tone within the dorsal horn and excitability of projection neurones <xref rid="b0290" ref-type="bibr">[59]</xref>. A substantial cross-inhibition exists within the dorsal horn, whereby low- and high-threshold afferents can moderate central neuronal activity <xref rid="b0205" ref-type="bibr">[42]</xref>, <xref rid="b0220" ref-type="bibr">[45]</xref>. Group III metabotropic glutamate receptors are highly expressed in the superficial laminae of the dorsal horn, as are group II receptors, which are primarily associated with interneurones in lamina IIi, in proximity to small myelinated fibres, with a large overlap with GABAergic terminals <xref rid="b0020" ref-type="bibr">[4]</xref>, <xref rid="b0145" ref-type="bibr">[30]</xref>, <xref rid="b0190" ref-type="bibr">[39]</xref>. Both groups of receptors have inhibitory roles within the dorsal horn and are subject to neuroplastic changes after injury <xref rid="b0060" ref-type="bibr">[12]</xref>, <xref rid="b0120" ref-type="bibr">[25]</xref>, <xref rid="b0345" ref-type="bibr">[70]</xref>. This neuroplasticity is the likely determinant of the reversal of icilin-induced analgesia by group II and III antagonists LY341495 and UBP1112 in neuropathic mice <xref rid="b0270" ref-type="bibr">[55]</xref>. TRPM8+ afferents also synapse near GABAergic terminals in the superficial dorsal horn <xref rid="b0085" ref-type="bibr">[18]</xref>. Menthol evokes excitatory postsynaptic potentials in GABAergic interneurones, presumed to activate cold afferents, subsequently gating transmission in projection neurones <xref rid="b0340" ref-type="bibr">[69]</xref>. Activation of endogenous opioid signalling pathways by menthol has also been implicated in analgesia. Contradictory reports exist of the reversal of icilin and menthol analgesia by naloxone. The discord between these studies could be attributed to the systemic route of administration of menthol <xref rid="b0100" ref-type="bibr">[21]</xref>, <xref rid="b0200" ref-type="bibr">[41]</xref> as opposed to topically applied icilin <xref rid="b0270" ref-type="bibr">[55]</xref>, as a centrally acting menthol could reduce neuronal excitability through TRPM8-independent, partly opioid mechanisms <xref rid="b0255" ref-type="bibr">[52]</xref>. Several molecular targets of menthol-mediated analgesia have been proposed, including the cumulative inactivation of sodium channels <xref rid="b0110" ref-type="bibr">[23]</xref>, activation of GABA_A currents <xref rid="b0335" ref-type="bibr">[68]</xref>, and inhibition of 5-HT₃ receptors <xref rid="b0015" ref-type="bibr">[3]</xref>. The efficacy of menthol is lost in TRPM8 knockout mice, suggesting that these non-TRP mechanisms comprise minor components of anti-hyperalgesia by menthol <xref rid="b0200" ref-type="bibr">[41]</xref>.</p><sec id="s0085"><label>4.1</label><title>Conclusion</title><p id="p0140">In summary, the data presented here suggest that aspects of menthol-induced analgesia are comparable between rats and humans, whereas menthol does not appear to induce central sensitisation as appears to be the case in normal human subjects. Furthermore, M8-Ag attenuated cold behaviours and neuronal responses in neuropathic rats but not in the absence of injury. Thus, overall, an interplay between complex peripheral and central effects appears to underlie the bi-directional effects of TRPM8 ligands, and changes in these functions appear to be driven by damage to peripheral nerves.</p></sec></sec><sec id="s0090"><title>Conflict of interest statement</title><p id="p0145">The authors have no conflicts of interest to declare.</p></sec>

Deletion and deletion/insertion mutations in the juxtamembrane domain of the <italic>FLT3</italic> gene in adult acute myeloid leukemia

<p>In contrast to <italic>FLT3</italic> ITD mutations, in-frame deletions in the <italic>FLT3</italic> gene have rarely been described in adult acute leukemia. We report two cases of AML with uncommon in-frame mutations in the juxtamembrane domain of the <italic>FLT3</italic> gene: a 3-bp (c.1770_1774delCTACGinsGT; p.F590_V592delinsLF) deletion/insertion and a 12-bp (c.1780_1791delTTCAGAGAATAT; p.F594_Y597del) deletion. We verified by sequencing that the reading frame of the <italic>FLT3</italic> gene was preserved and by cDNA analysis that the mRNA of the mutant allele was expressed in both cases. Given the recent development of <italic>FLT3</italic> inhibitors, our findings may be of therapeutic value for AML patients harboring similar <italic>FLT3</italic> mutations.</p>

<contrib contrib-type="author"><name><surname>Deeb</surname><given-names>Kristin K.</given-names></name><xref rid="aff0005" ref-type="aff">a</xref></contrib><contrib contrib-type="author"><name><surname>Smonskey</surname><given-names>Matthew T.</given-names></name><xref rid="aff0005" ref-type="aff">a</xref></contrib><contrib contrib-type="author"><name><surname>DeFedericis</surname><given-names>HanChun</given-names></name><xref rid="aff0005" ref-type="aff">a</xref></contrib><contrib contrib-type="author"><name><surname>Deeb</surname><given-names>George</given-names></name><xref rid="aff0015" ref-type="aff">c</xref></contrib><contrib contrib-type="author"><name><surname>Sait</surname><given-names>Sheila N.J.</given-names></name><xref rid="aff0020" ref-type="aff">d</xref></contrib><contrib contrib-type="author"><name><surname>Wetzler</surname><given-names>Meir</given-names></name><xref rid="aff0010" ref-type="aff">b</xref></contrib><contrib contrib-type="author"><name><surname>Wang</surname><given-names>Eunice S.</given-names></name><xref rid="aff0010" ref-type="aff">b</xref></contrib><contrib contrib-type="author"><name><surname>Starostik</surname><given-names>Petr</given-names></name><email>petr.starostik@roswellpark.org</email><xref rid="aff0005" ref-type="aff">a</xref><xref rid="cor1" ref-type="corresp">&#x0204e;</xref></contrib>

Leukemia Research Reports

<sec sec-type="intro" id="s0005"><label>1</label><title>Introduction</title><p>Mutations in the <italic>FLT3</italic> gene have been described in about 25% of acute myeloid leukemia (AML). They are somewhat more common in acute promyelocytic leukemia (APL), and have been associated with an increased risk of relapse, decreased disease-free survival, decreased event-free survival, and decreased overall survival <xref rid="bib1" ref-type="bibr">[1]</xref>. These mutations result in constitutive activation of the FLT3 protein and are of two types: internal tandem duplication (ITD) mutations in exon 14 resulting from the duplication and tandem insertion of a portion of the juxtamembrane (JM) domain of the <italic>FLT3</italic> gene and missense mutations in exon 20 which alter the aspartic acid residue at position 835 (D835) within the kinase domain of the FLT3 protein. In the case of ITD mutations, the duplicated segment length ranges in size from 3 to several hundred base pairs and is always in-frame and therefore expected to produce a functional protein <xref rid="bib2" ref-type="bibr">[2]</xref>. Rare deletion and deletion/insertion mutations affecting the <italic>FLT3</italic> juxtamembrane region have been described in childhood acute lymphoblastic leukemia <xref rid="bib3 bib4" ref-type="bibr">[3,4]</xref>. Here, we report two cases of deletion and deletion/insertion mutations in the juxtamembrane domain of <italic>FLT3</italic> in adult AML. Proper identification of these mutations may have prognostic and therapeutic significance for AML patients.</p></sec><sec sec-type="methods" id="s0010"><label>2</label><title>Methods</title><sec id="s0015"><label>2.1</label><title>Patients</title><sec id="s0020"><label>2.1.1</label><title>Patient #1</title><p>A 47 year-old man presented with complaints of shortness of breath, fatigue, and weakness over several days. He had WBC of 42.3&#x000d7;10⁹/L and hemoglobin of 4.8&#x000a0;g/dL. Bone marrow morphology showed 95% cellularity with 83% blasts and the case was classified as AML M0 with myelodysplasia-related changes based on the detection of del(5q) by FISH, as the minimal differentiation of the leukemic blasts made the assessment of multilineage dysplasia rather difficult. Molecular diagnostic studies detected wild-type <italic>NPM1</italic> gene and atypically mutated <italic>FLT3</italic> gene. The patient underwent induction with cytarabine and idarubicin-based chemotherapy, but had evidence of primary refractory <italic>FLT3</italic> mutation-positive AML on bone marrow biopsy performed 14 days after initiation of therapy. He then received high-dose cytarabine and mitoxantrone re-induction therapy. Repeat bone marrow evaluation upon count recovery revealed remission with 2% blasts and no evidence of <italic>FLT3</italic> mutation. He subsequently underwent allogeneic stem cell transplantation from his sister and remained in remission for five years, after which the AML relapsed with a <italic>FLT3</italic> D835 mutation and del(q5). He died shortly afterwards of infectious complications following re-induction chemotherapy.</p></sec><sec id="s0025"><label>2.1.2</label><title>Patient #2</title><p>A 54 year-old man with a prior medical history of coronary artery disease, diabetes, hypercholesterolemia, and hyperlipidemia presented with new onset of widespread bruising and blood in stool. Physical exam demonstrated scattered ecchymoses. Blood work revealed WBC of 8.6&#x000d7;10⁹/L, hemoglobin of 9.8&#x000a0;g/L, and platelet count of 26&#x000d7;10⁹/L. Prothrombin time was slightly elevated at 15.6&#x000a0;s (INR 1.25) with normal activated partial thromboplastin and a reduced fibrinogen level of 163&#x000a0;mg/dL. Hematopathologic evaluation of blood and bone marrow confirmed the diagnosis of acute promyelocytic leukemia (APL) with 91% marrow blasts/abnormal promyelocytes. Cytogenetics revealed a reciprocal translocation between the long arms of chromosomes 15 and 17 in 19/20 cells, t(15;17)(q24;q21). Molecular studies demonstrated a high level of the <italic>PMLRARalpha</italic> t(15;17) fusion transcript (164% of control) by quantitative RT-PCR. An atypical <italic>FLT3</italic> mutation was also identified. The patient was initiated on differentiation therapy with oral retinoic acid (ATRA) 45&#x000a0;mg/m² and arsenic trioxide 0.15&#x000a0;mg/kg intravenously daily as previously described <xref rid="bib5" ref-type="bibr">[5]</xref>. Pseudotumor cerebri, scrotal ulcerations, and persistent headaches necessitated ATRA dose reduction. The patient was subsequently found to have CNS involvement by APL and received multiple intrathecal methotrexate injections. He was discharged home with count recovery two months after diagnosis and in complete remission.</p></sec></sec><sec id="s0030"><label>2.2</label><title><italic>FLT3</italic> ITD and D835 mutation fragment analysis</title><p>DNA was extracted from blood or bone marrow samples using the EZ1 DNA Blood Kit (Qiagen, Germantown, MD) on the BioRobot EZ1 system (Qiagen). The <italic>FLT3</italic> PCR-based fragment analysis assay was performed as previously described <xref rid="bib6" ref-type="bibr">[6]</xref>.</p></sec><sec id="s0035"><label>2.3</label><title><italic>FLT3</italic> juxtamembrane domain Sanger sequencing</title><p>Mutations detected in the juxtamembrane domain of the <italic>FLT3</italic> gene underwent Sanger sequencing in both forward and reverse directions with the Big Dye Terminator v 3.1 Cycle Sequencing Kit (Life Technologies, Carlsbad, CA). Results were analyzed in Sequencing Analysis v5.2 software (Life Technologies) and Lasergene SeqMan Pro v10.0 (DNAStar, Madison, WI), and aligned to the <italic>FLT3</italic> reference gene (NCBI RefSeq NM_004119.2).</p></sec><sec id="s0040"><label>2.4</label><title><italic>FLT3</italic> mRNA analysis</title><p>RNA was extracted from patient samples using the miRNeasy Mini Kit (Qiagen) and converted to cDNA with the RevertAid First Strand cDNA Synthesis Kit (Thermo Scientific, Pittsburgh, PA), which was subsequently amplified with primers FLT3F: 5&#x02032;-6-FAM-GCCAGCTACAGATGGTACAGG-3&#x02032; and FLT3R: 5&#x02032;-TTGCGTTCATCACTTTTCCA-3&#x02032;. PCR products were analyzed on the ABI 3130<italic>xl</italic> Genetic Analyzer instrument (Life Technologies).</p></sec></sec><sec id="s0045"><label>3</label><title>Results and discussion</title><p>Upon <italic>FLT3</italic> ITD fragment analysis during routine molecular diagnostics work-up at presentation, both patient samples showed an unusual peak in the electropherogram (<xref rid="f0005" ref-type="fig">Fig. 1</xref>A). Besides the wild-type allele of 330&#x000a0;bp, a shorter PCR product in the same reaction pointed to the presence of a mutated allele showing a deletion in the PCR-amplified juxtamembrane domain region. Fragments shorter by 3-bp (327-bp) in patient #1 with a mutant allele/wild-type <italic>FLT3</italic> ratio of 0.29, and by 12-bp (318-bp) in patient #2 with a ratio of 0.49 were detected. These fragments were further analyzed by Sanger sequencing to elucidate the nature of the deletions. Compared to the wild-type <italic>FLT3</italic> sequence, patient #1 had a 5-bp deletion (CTACG) mutation combined with a 2-bp (GT) insertion: c.1770_1774delCTACGinsGT mutation (<xref rid="f0005" ref-type="fig">Fig. 1</xref>B), giving an overall 3&#x000a0;bp deletion as detected by <italic>FLT3</italic> fragment analysis. The deletion resulted in a p.F590_V592delinsLF amino acid change in the juxtamembrane domain. Patient #2 had a c.1780_1791delTTCAGAGAATAT (12-bp deletion) mutation (<xref rid="f0005" ref-type="fig">Fig. 1</xref>B) resulting in p.F594_Y597del amino acid deletion in the juxtamembrane domain. Notably, these deletion and deletion/insertion mutations were in-frame and the reading frame of the <italic>FLT3</italic> gene was preserved in both cases. Subsequently, the samples were tested for mutant versus wild type allele expression using cDNA fragment analysis. Both wild-type and mutant alleles were expressed at ratios comparable to the results of the <italic>FLT3</italic> ITD assay (<xref rid="f0010" ref-type="fig">Fig. 2</xref>).</p><p><italic>FLT3</italic> deletion and deletion/insertion mutations were previously reported in cases of pediatric acute lymphoblastic leukemia <xref rid="bib3 bib4" ref-type="bibr">[3,4]</xref>, but seldom described in adult acute leukemia. While the biological significance of this type of <italic>FLT3</italic> mutations is unknown in human disease, a small 10-amino acid (Tyr589 to Tyr599) deletion in the juxtamembrane domain of <italic>FLT3</italic> has been previously shown to lead to constitutive activation of the FLT3 protein in transformed murine IL3-dependent myeloid progenitor 32D cell line <xref rid="bib7" ref-type="bibr">[7]</xref>. Similar deletion mutations are found in receptor tyrosine kinase <italic>KIT</italic> in gastrointestinal stromal tumors (GIST) <xref rid="bib8 bib9" ref-type="bibr">[8,9]</xref>. An in-frame deletion of 7-amino acids in the juxtamembrane domain of the <italic>KIT</italic> gene resulted in receptor autophosphorylation and malignant transformation of mast cells <xref rid="bib10" ref-type="bibr">[10]</xref>. These studies and our findings that both patients showed in-frame deletions with mRNA expressed (unfortunately, the samples did not yield enough material for a Western blot) suggest that deletion and deletion/insertion mutations in <italic>FLT3</italic> juxtamembrane domain may lead to receptor activation. Animal models would be a way to prove this hypothesis and show, if inhibition of <italic>FLT3</italic> can be therapeutically exploited in such cases. Whether the presence of these mutations in adult acute leukemia has prognostic significance warrants further investigation of a larger patient cohort.</p></sec><sec id="s0050"><title>Conflict of interest</title><p>The authors declare no conflicts of interest.</p></sec>

An unusual case of splenomegaly and increased lactate dehydrogenase heralding acute myeloid leukemia with eosinophilia and RUNX1&#x02013;MECOM fusion transcripts

<p>We report the first case of acute myeloid leukemia (AML) with RUNX1&#x02013;MECOM fusion transcripts, showing marked eosinophilia. A 63-year old man admitted in August 2013, had previously been observed in April 2013, because of persisting homogeneous splenomegaly and increased LDH, which were initially attributed to both minor &#x003b2;-thalassemia and previous acute myocardial infarction. However, based upon the retrospective analysis of clinical features combined with the documentation of both JAK2 V617F and c-KIT D816V mutations at AML diagnosis, an aggressive leukemic transformation with eosinophilia of a previously unrecognized myeloproliferative neoplasm, rather than the occurrence of <italic>de novo</italic> AML, may be hypothesized.</p>

<contrib contrib-type="author"><name><surname>Forghieri</surname><given-names>Fabio</given-names></name><email>fabio.forghieri@unimore.it</email><xref rid="cor1" ref-type="corresp">&#x0204e;</xref></contrib><contrib contrib-type="author"><name><surname>Bigliardi</surname><given-names>Sara</given-names></name></contrib><contrib contrib-type="author"><name><surname>Morselli</surname><given-names>Monica</given-names></name></contrib><contrib contrib-type="author"><name><surname>Potenza</surname><given-names>Leonardo</given-names></name></contrib><contrib contrib-type="author"><name><surname>Fantuzzi</surname><given-names>Valeria</given-names></name></contrib><contrib contrib-type="author"><name><surname>Faglioni</surname><given-names>Laura</given-names></name></contrib><contrib contrib-type="author"><name><surname>Nasillo</surname><given-names>Vincenzo</given-names></name></contrib><contrib contrib-type="author"><name><surname>Messerotti</surname><given-names>Andrea</given-names></name></contrib><contrib contrib-type="author"><name><surname>Paolini</surname><given-names>Ambra</given-names></name></contrib><contrib contrib-type="author"><name><surname>Luppi</surname><given-names>Mario</given-names></name></contrib>

Leukemia Research Reports

<p>The t(3;21)(q26;q22) is a rare cytogenetic abnormality, reported in approximately 1% of all myelodysplastic syndrome (MDS) or acute myeloid leukemia (AML) cases and mainly occurring in therapy-related myeloid neoplasms, in accelerated or blast phase of chronic myeloid leukemia or other myeloproliferative neoplasms (MPN), and, rarely, in <italic>de novo</italic> AML <xref rid="bib1 bib2 bib3 bib4 bib5 bib6" ref-type="bibr">[1&#x02013;6]</xref>. Previous treatment with hydroxycarbamide or other antimetabolites is strongly implicated as a contributory cause <xref rid="bib4" ref-type="bibr">[4]</xref>. The presence of t(3;21)(q26;q22), isolated or associated with other chromosomal abnormalities, is associated with a very poor prognosis <xref rid="bib5" ref-type="bibr">[5]</xref>. In this translocation, portions of the RUNX1 gene have been reported to be variably fused to MECOM (currently preferred designation for the genes located within the 3q26 region, namely RPL22, MDS1, EVI1), as a result of alternative intergenic splicing, with production of multiple fusion transcripts <xref rid="bib1 bib2 bib3" ref-type="bibr">[1&#x02013;3]</xref>. RUNX1&#x02013;MECOM (formerly AML1&#x02013;MDS1&#x02013;EVI1) fusion products directly contribute to leukemogenesis or leukemic transformation, can block myeloid differentiation and promote proliferation by exerting a dominant-negative effect over RUNX1-induced normal transcriptional activation, antagonize the growth-inhibitory effects of transforming growth factors, block JNK activity and therefore prevent stress-induced apoptosis, and enhance AP-1 activity <xref rid="bib7 bib8" ref-type="bibr">[7,8]</xref>. We report here the unusual case of an elderly patient observed because of persisting splenomegaly and increased lactate dehydrogenase (LDH), heralding AML with eosinophilia and RUNX1&#x02013;MECOM fusion transcripts.</p><p>In details, a 63-year old Caucasian man with previous history of minor &#x003b2;-thalassemia, arterial hypertension, hypothyroidism, fibromyalgia, benign prostatic hyperplasia and acute myocardial infarction in 2009 was admitted to our outpatient&#x005f3;s Hematology Unit in April 2013 because of splenomegaly and persistently increased LDH, mainly LDH2 isoform, since 2009. The blood examinations documented white blood cell (WBC) count 4.7&#x000d7;10⁹/L, with a differential count showing 47% neutrophils, 34% lymphocytes, 9% monocytes, 5% eosinophils, 5% basophils, hemoglobin (Hb) level 15.4&#x000a0;g/dl with MCV 76&#x000a0;fl, platelet (Plt) count 338&#x000d7;10⁹/L, serum LDH 1053&#x000a0;IU/L. The morphological examination of the peripheral blood (PB) smear did not document cytological abnormalities, while neutrophil alkaline phosphatase activity was elevated. Neither abnormal myeloid cells nor atypical lymphocytes were detected by flow cytometry on PB samples. Microbiological, virological and autoimmunity examinations were negative. Homogeneous splenomegaly (bipolar diameter 16.2&#x000a0;cm) without signs of portal hypertension was observed on ultrasonography. Neither bone marrow (BM) examination nor cytogenetic and molecular analyses were performed at this time-point, whereas the morphological examination of a BM aspirate performed in 2008 because of transient mild neutropenia was unrevealing. In August 2013 the patient was subsequently admitted because of fever, fatigue, abdominal pain and drenching night sweats. The laboratory investigations revealed WBC count 12.2&#x000d7;10⁹/L with a differential count showing 25% neutrophils, 11% lymphocytes, 2% monocytes, 43% eosinophils, 1% basophils and 18% blasts, Hb level 9.7&#x000a0;g/dl, Plt count 17&#x000d7;10⁹/L, LDH 548&#x000a0;IU/L. Unfortunately, serum tryptase level was not measured. Homogeneous splenomegaly was worsened (bipolar diameter 20&#x000a0;cm), as observed on abdominal ultrasonography. The morphological examination of BM aspirate (<xref rid="f0005" ref-type="fig">Fig. 1</xref>A) and trephine biopsy (<xref rid="f0005" ref-type="fig">Fig. 1</xref>B) showed hypercellularity (90%), with marked proliferation of eosinophil granulocytopoiesis (40%), multilineage myelodysplastic features, especially with small hypolobated megakaryocytes, and a blast cell count 20&#x02013;30%. Blast cells were minimally differentiated, with high nuclear/cytoplasmic ratio and basophilic cytoplasm. They were CD34+, CD33+/&#x02212;, CD13+, CD117+, HLA-DR+, CD38+, c-MPO&#x02212; by flow cytometry analysis performed on the BM aspirate. Moreover, mild BM fibrosis was documented on Gomori methenamine silver staining (<xref rid="f0005" ref-type="fig">Fig. 1</xref>C). The presence of dense mast cell collections or atypical spindle-shaped mast cells was morphologically excluded. Moreover, neither immunohistochemical examinations for CD68 and CD117 nor flow-cytometry analyses to investigate the expression of CD2, CD25 and CD117, performed on BM trephine biopsy and BM aspirate, respectively, documented atypical or aggregated mast cells (not shown). Unfortunately, immunostaining for tryptase was not performed. Based upon these features, AML with MDS-related changes and eosinophilia was thus diagnosed. Conventional G-banding showed 46,XY,t(3;21)(q26;q22) karyotype in all the 20 metaphase cells analyzed (<xref rid="f0005" ref-type="fig">Fig. 1</xref>D). Molecular examinations, namely reverse transcriptase (RT)-PCR (<xref rid="f0005" ref-type="fig">Fig. 1</xref>E) and subsequent sequencing analyses (<xref rid="f0005" ref-type="fig">Fig. 1</xref>F) performed on both PB and BM samples, also documented RUNX1&#x02013;MECOM, alternatively spliced multiple fusion transcripts <xref rid="bib1 bib2 bib3" ref-type="bibr">[1&#x02013;3]</xref>.</p><p>Further extensive FISH and molecular studies failed to detect either BCR&#x02013;ABL, PDGFRA, PDGFRB and FGFR1 rearrangements or NPM1 and FLT3 mutations, whereas c-KIT D816V mutation was found on PB and BM mononuclear cells. Moreover, allele-specific PCR demonstrated JAK2 V617F mutation on both PB and BM polymorphonuclear cells, but JAK2 V617F allele burden was not assessed at AML diagnosis or during the subsequent course of the disease. An aggressive clinical behavior was observed. The patient was refractory to remission induction chemotherapy (cytarabine 100&#x000a0;mg/m² on days 1&#x02013;7 and daunorubicin 45&#x000a0;mg/m² on days 1&#x02013;3) and subsequent salvage chemotherapeutic FLAG regimen (fludarabine 30&#x000a0;mg/m² and cytarabine 2&#x000a0;g/m² on days 1&#x02013;5, G-CSF, 5&#x000a0;mcg/kg/day from day &#x02212;1 to day +5). Thereafter, he received best supportive care and was lost of follow-up.</p><p>To the best of our knowledge, we have described here the first case of AML with RUNX1&#x02013;MECOM fusion transcripts and concurrent JAK2 V617F and c-KIT D816V mutations, showing marked peripheral and BM eosinophilia. AML with eosinophilia is usually found in core binding factor (CBF) leukemias, with cytogenetic abnormalities such as inv(16)(p13q22)/t(16;16)(p13;q22) resulting in CBFB&#x02013;MYH11 fusion transcript or t(8;21)(q22;q22) resulting in RUNX1&#x02013;RUNX1T1 fusion transcript. However, it is also rarely associated with other translocations such as t(16;21)(q24;q22), which produces the RUNX1&#x02013;CBFA2T3 fusion transcript <xref rid="bib9" ref-type="bibr">[9]</xref>. Of interest, c-KIT mutations are documented in approximately 25&#x02013;30% of cases of CBF leukemia, whereas they are infrequent in other AML subtypes <xref rid="bib10" ref-type="bibr">[10]</xref>. It has also recently been reported that patients with t(8;21)-positive AML showed one or two additional gene mutations in up to 50% or 15% of the cases, respectively <xref rid="bib11" ref-type="bibr">[11]</xref>. The most common mutated genes, in that series, were c-KIT (23/139 patients; 16.5%), NRAS (18/139 patients; 12.9%) and ASXL1 (16/139 patients; 11.5%), but also FLT3, CBL, KRAS, IDH2 and JAK2 were mutated in 2.9&#x02013;5% of the cases <xref rid="bib11" ref-type="bibr">[11]</xref>. Of note, our patient was previously observed, a few months before AML diagnosis, because of persisting homogeneous splenomegaly and increased LDH, which, in the absence of any other features suspected for hematologic malignancy, except for a mild increase of morphologically normal basophils, were initially attributed to both minor &#x003b2;-thalassemia and previous acute myocardial infarction. However, based upon the retrospective analysis of the clinical features combined with the documentation of both JAK2 V617F and c-KIT D816V mutations on PB and BM samples at AML diagnosis, an aggressive leukemic transformation with eosinophilia and t(3;21)(q26;q22) translocation of a previously unrecognized MPN, rather than the occurrence of a <italic>de novo</italic> AML, may be hypothesized in our case <xref rid="bib4 bib5" ref-type="bibr">[4,5]</xref>. Unfortunately, PB samples adequate for molecular examinations were not stored in April 2013, so we could not retrospectively investigate the presence of JAK2 and KIT mutations at first patient&#x005f3;s observation. In the absence of available samples for BM morphological examinations or cytogenetic/molecular analyses before AML diagnosis, we cannot have confirmation of a potential MPN diagnosis preceding AML. Furthermore, neither JAK2 V617F allele burden at different time-points nor JAK2 V617F mutation analysis on purified myeloid blasts were available, so that we cannot exclude that such mutation would have been limited to polymorphonuclear cells. Of note, the latter possibility would not be surprising, because it has previously been reported that leukemic blasts are frequently negative for the JAK2 V617F mutation in transformed JAK2 V617F-positive MPN <xref rid="bib12" ref-type="bibr">[12]</xref>. The unusual clinical picture of our patient may suggest the need to perform either BM morphological examinations or cytogenetic/molecular analyses, at least on PB samples, to rule out MPN in similar cases observed for persisting/worsening splenomegaly and increased LDH, even in the absence of significant morphological and immunophenotypic abnormalities on PB. Furthermore, we have reported the association of AML with eosinophilia with another chromosomal translocation involving RUNX1, located at 21q22 region, suggesting the need to further investigate molecular mechanisms underlying the presence of abnormal eosinophilia in hematologic malignancies.</p><sec id="s0010"><title>Conflict of interest</title><p>The authors indicated no potential conflicts of interest.</p></sec><sec id="s0015"><title>Authors&#x005f3; contributions</title><p>FF and SB analyzed clinical, cytogenetic and molecular data and wrote the manuscript; MM, LP, VF, LF, VN, AM, AP took care of the patient, collected and analyzed clinical data and critically revised the manuscript; ML critically revised the manuscript. All authors approved the final version of the manuscript.</p><p>FF and SB equally contributed to the work.</p></sec>

Pathogenicity study in sheep using reverse-genetics-based reassortant bluetongue viruses

Could not extract abstract

<contrib contrib-type="author" id="aut0005"><name><surname>Celma</surname><given-names>Cristina C.</given-names></name><xref rid="aff0005" ref-type="aff">a</xref></contrib><contrib contrib-type="author" id="aut0010"><name><surname>Bhattacharya</surname><given-names>Bishnupriya</given-names></name><xref rid="aff0005" ref-type="aff">a</xref></contrib><contrib contrib-type="author" id="aut0015"><name><surname>Eschbaumer</surname><given-names>Michael</given-names></name><email>michael.eschbaumer@ars.usda.gov</email><xref rid="aff0010" ref-type="aff">b</xref><xref rid="fn1" ref-type="fn">1</xref></contrib><contrib contrib-type="author" id="aut0020"><name><surname>Wernike</surname><given-names>Kerstin</given-names></name><xref rid="aff0010" ref-type="aff">b</xref></contrib><contrib contrib-type="author" id="aut0025"><name><surname>Beer</surname><given-names>Martin</given-names></name><xref rid="aff0010" ref-type="aff">b</xref></contrib><contrib contrib-type="author" id="aut0030"><name><surname>Roy</surname><given-names>Polly</given-names></name><email>polly.roy@lshtm.ac.uk</email><email>pollyroy.office@lshtm.co.uk</email><xref rid="aff0005" ref-type="aff">a</xref><xref rid="cor0005" ref-type="corresp">&#x0204e;</xref></contrib><aff id="aff0005"><label>a</label>Department of Pathogen Molecular Biology, Faculty of Infectious and Tropical Diseases, London School of Hygiene and Tropical Medicine, United Kingdom</aff><aff id="aff0010"><label>b</label>Institut f&#x000fc;r Virusdiagnostik, Friedrich-Loeffler-Institut, Greifswald-Insel Riems, Germany</aff>

Veterinary Microbiology

<sec id="sec0005"><label>1</label><title>Introduction</title><p id="par0020">Bluetongue (BT), an insect-transmitted, non-contagious viral disease of domestic and wild ruminants is caused by bluetongue virus (BTV). The disease is characterized by inflammation of the mucous membranes, congestion, swelling, hemorrhages and is often accompanied with high mortality in sheep (<xref rid="bib0070" ref-type="bibr">Erasmus, 1975</xref>, <xref rid="bib0095" ref-type="bibr">Maclachlan et al., 2009</xref>). Although cattle and goats usually carry the virus for a certain period of time without showing any apparent clinical signs of disease, they are capable of transmitting the virus to other ruminants via biting <italic>Culicoides</italic> midges. BTV is present in a broad band of countries extending approximately between 40&#x000b0;N and 35&#x000b0;S (<xref rid="bib0115" ref-type="bibr">Mellor et al., 2000</xref>, <xref rid="bib0130" ref-type="bibr">Purse et al., 2005</xref>). Until 15 years ago, Europe was essentially BT-free apart from Cyprus; however, since 1998 at least one of the 26 serotypes of BTV has been active on the continent every year, mainly in the Mediterranean basin (<xref rid="bib0100" ref-type="bibr">Maclachlan and Guthrie, 2010</xref>, <xref rid="bib0120" ref-type="bibr">Mellor et al., 2008</xref>, <xref rid="bib0130" ref-type="bibr">Purse et al., 2005</xref>). In 2006 a highly pathogenic BTV-8 strain emerged for the first time in Northern Europe, spreading very rapidly and affecting thousands of herds. The same serotype re-emerged in 2007 and 2008, causing devastating disease not only in sheep but also in cattle with high morbidity and mortality (<xref rid="bib0055" ref-type="bibr">Elbers et al., 2008a</xref>, <xref rid="bib0060" ref-type="bibr">Elbers et al., 2008b</xref>, <xref rid="bib0145" ref-type="bibr">Wilson and Mellor, 2009</xref>). Studies involving molecular epidemiology have also shown that the most severe disease in northern European sheep and cattle was caused by BTV-8 (<xref rid="bib0040" ref-type="bibr">Dal Pozzo et al., 2013</xref>, <xref rid="bib0105" ref-type="bibr">Martinelle et al., 2013</xref>, <xref rid="bib0130" ref-type="bibr">Purse et al., 2005</xref>). The phenotypic differences between BTV-8 compared with less virulent strains suggested that genetic background may be partly responsible. However, the mechanism of pathogenicity is still very poorly understood.</p><p id="par0025">BTV is a member of the <italic>Orbivirus</italic> genus within the <italic>Reoviridae</italic> family. Like other members of the family, BTV has a genome of 10 segmented double-stranded RNA (segments S1&#x02013;S10) that are enclosed within two capsids. While the inner core is made up of 5 highly conserved proteins (VP1, VP3, VP4, VP6 and VP7), the outer capsid consists of two variable proteins, VP2 (receptor-binding protein and serotype determinant) and VP5 (membrane penetration protein). In addition, BTV also encodes for 4 non-structural proteins (NS1&#x02013;NS4), of which NS3 encoded by S10 is more variable than NS1 and NS2. NS3 is shown to be involved in virus trafficking and release from the infected host (<xref rid="bib0010" ref-type="bibr">Beaton et al., 2002</xref>, <xref rid="bib0025" ref-type="bibr">Celma and Roy, 2009</xref>). Recently it has been shown that NS3 is also involved in the regulation of the induction of interferon type 1 (<xref rid="bib0035" ref-type="bibr">Chauveau et al., 2013</xref>), suggesting a role in the innate immune response.</p><p id="par0030">In this study, we designed reassortant viruses between BTV-8 and BTV-1 to establish the genetic basis of BTV pathogenicity. The rationale for designing reassortant viruses was based on the two most variable proteins of the outer capsid (VP2 and VP5) and the non-structural protein NS3, which is the most variable within BTV NS proteins. Reassortant viruses were generated using a reverse genetics (RG) system replacing these three RNA segments (S2, S6 and S10) of the low virulent strain, BTV-1, with that of highly virulent BTV-8, either singly or in combinations. The phenotypic characteristics of the disease caused by these reassortant viruses were analyzed by infection of sheep. Our results suggested that all three proteins together are involved in the disease outcome and that the molecular basis of BTV pathogenicity is highly complex.</p></sec><sec id="sec0010"><label>2</label><title>Methods</title><sec id="sec0015"><label>2.1</label><title>Cells and viruses</title><p id="par0035">BSR cells (BHK-21 subclone) were maintained in Dulbecco modified eagle medium (DMEM, Sigma&#x02013;Aldrich) supplemented with 5% (v/v) fetal bovine serum (FBS, Invitrogen). PT and SFT-R cells (ovine-derived kidney and thymus cells respectively, Collection of Cell Lines in Veterinary Medicine, Friedrich-Loeffler-Institut, Insel Riems, Germany) were maintained in minimum essential medium eagle (MEM, Sigma&#x02013;Aldrich) supplemented with 10% (v/v) FBS.</p><p id="par0040">BTV-1 (South African strain) and BTV-8 (Ardennes isolate) viral stocks were generated by infection of BSR cells and kept at 4&#x000a0;&#x000b0;C until use.</p></sec><sec id="sec0020"><label>2.2</label><title>Recovery of reassortant BTV-1/BTV-8 viruses</title><p id="par0045">Segments S2 (VP2), S6 (VP5) and S10 (NS3) (GenBank accession numbers: KJ872780&#x02013;KJ872782) of BTV-8 were obtained using a sequence-independent cloning system as previously described (<xref rid="bib0020" ref-type="bibr">Boyce et al., 2008</xref>, <xref rid="bib0090" ref-type="bibr">Maan et al., 2007</xref>, <xref rid="bib0110" ref-type="bibr">Matsuo et al., 2011</xref>). Briefly, dsRNAs from purified core particles were ligated to a self-annealing primer before RT-PCR amplification using a specific primer. Each cDNA amplified from segments S2, S6 and S10 of BTV-8 was cloned into pUC19 and fully sequenced (Source Bioscience) before insertion of the T7 promoter at the 5&#x02032; end and insertion of a unique restriction enzyme site that generates the correct end of the segment at the 3&#x02032; end.</p><p id="par0050">For synthesis of uncapped T7 transcripts for segments S2, S6 and S10, RiboMAX Large-Scale RNA Production System T7 (Promega) or TranscriptAid T7 High Yield Transcription Kit (Thermo Scientific) kits were used according to manufacturer's instructions. Reassortant viruses between BTV-1 and BTV-8 were recovered from confluent monolayers of BSR cells after transfection with a full set of BTV T7 transcripts as described before (<xref rid="bib0020" ref-type="bibr">Boyce et al., 2008</xref>). Individual plaques were picked, amplified and virus stocks were kept at 4&#x000a0;&#x000b0;C.</p></sec><sec id="sec0025"><label>2.3</label><title>Virus growth kinetics and characterization</title><p id="par0055">The genomic dsRNA profile was analyzed for each reassortant. Monolayers of BSR cells were infected with reassortant or parental viruses and upon complete cytopathic effect, infected cells were harvested. The genomic dsRNA was purified with Tri reagent (Sigma) using standard methods. The parental origin of segments S2, S6 and S10 was determined by differential mobility in non-denaturing PAGE or by sequencing.</p><p id="par0060">For virus growth study, monolayers of PT and SFT-R cells were synchronously infected at a multiplicity of infection (MOI) of 1 and samples were collected at 0, 12, 24 and 48&#x000a0;h (hours) post-infection (p.i.). Cells and supernatant were harvested, subjected to two freeze/thaw cycles and the total titer was determined by plaque assay in triplicate and expressed as plaque formation units per ml (PFU/ml) or by tissue culture infective dose 50 (TCID₅₀). The mean, standard deviation and the <italic>p</italic> values were also determined by Excel (Microsoft). Viral protein expression was determined by Western blot (WB) using specific antibodies against structural VP5, VP7 and non-structural NS3 proteins. As loading control, an antibody against &#x003b2;-actin (Sigma) was used. Each blotting experiment was repeated twice and the amount of protein expression was quantitated by ImageJ software.</p></sec><sec id="sec0030"><label>2.4</label><title>Pathogenicity studies</title><p id="par0065">Forty-six sheep (East Frisian breed) were segregated in groups of 4 animals and subcutaneously injected with parental BTV-1 (South African strain), BTV-8 (Ardennes isolate) or one of the reassortant viruses at &#x0223c;1&#x000a0;&#x000d7;&#x000a0;10⁷&#x000a0;PFU/animal. Control groups received only saline buffer. Whole blood samples were taken at regular intervals before and after infection. For three weeks, a clinical score was assigned daily to each sheep based on a range of clinical signs. These included body temperature, feed intake, attitude, and lameness, appearance of conjunctiva, scleral blood vessels and cornea, appearance of skin and mucous membranes of the nose and mouth, redness and hemorrhaging, nasal and ocular discharge, salivation, respiratory rate and sounds, as well as edema of the tongue, face, ears, periorbital and submandibular regions. Depending on severity, up to four points were awarded per category. All animal procedures were in compliance with European ethical regulations, i.e. directives 91/628, 92/65 and 86/609/EEG, regarding the protection of vertebrate animals used for experimental and/or scientific purposes.</p><p id="par0070">A mixed-design analysis of the variance model (implemented in R, <ext-link ext-link-type="uri" xlink:href="http://www.r-project.org/" id="intr0005">http://www.r-project.org/</ext-link>) was used to test for significant differences in clinical scores between groups. <italic>P</italic> values from post hoc pairwise comparisons were adjusted for multiplicity by the Bonferroni method.</p></sec><sec id="sec0035"><label>2.5</label><title>Viral load</title><p id="par0075">To determine virus replication in animal hosts, RNA was extracted from blood samples at different time points with the NucleoSpin 96 virus core kit (Macherey-Nagel) and analyzed by real-time RT-PCR as described previously (<xref rid="bib0140" ref-type="bibr">Toussaint et al., 2007</xref>) using a LightCycler 480II (Roche Diagnostics). A NS1 (S5) RNA standard was used for absolute quantification.</p></sec><sec id="sec0040"><label>2.6</label><title>Immunological tests</title><p id="par0080">The presence of antibodies against BTV VP7 was measured by ELISA. Serum samples were analyzed by a double-recognition BTV antibody ELISA (PrioCHECK BTV DR, Prionics GmbH). For neutralization tests, serum samples were tested against parental BTV-1 and BTV-8 viruses and the neutralization titer was determined as the highest dilution able to neutralize 100 TCID₅₀ of virus in 50% of replicate wells.</p></sec></sec><sec id="sec0045"><label>3</label><title>Results</title><sec id="sec0050"><label>3.1</label><title>Recovery and characterization of reassortant BTV-1 and BTV-8 viruses</title><p id="par0085">Initially BTV-8 (Ardennes isolate) (<xref rid="bib0085" ref-type="bibr">Le Gal et al., 2008</xref>) exact-copy segments S2, S6 and S10, encoding VP2, VP5 and NS3 respectively, were cloned using a sequence-independent method (<xref rid="bib0020" ref-type="bibr">Boyce et al., 2008</xref>, <xref rid="bib0090" ref-type="bibr">Maan et al., 2007</xref>). Subsequently, after ensuring the accuracy of the clones by sequencing, a T7 promoter and a specific restriction enzyme site were introduced by PCR and transcripts were generated using the fully digested plasmids as templates. For virus recovery, BSR cells were transfected with ssRNAs S2, S6 and S10 of BTV-8 in different combinations together with the remaining ssRNAs (either 7 or 8 segments) of BTV-1 (<xref rid="bib0020" ref-type="bibr">Boyce et al., 2008</xref>). Following incubation, reassortant viruses were recovered as described previously (<xref rid="bib0020" ref-type="bibr">Boyce et al., 2008</xref>) and assessment of the plaque morphology established that the recovered BTV1/8VP2.5 (with segments S2 and S6 of BTV-8) and BTV1/8VP2.5.NS3 (with segments S2, S6 and S10 of BTV-8) reassortant viruses had plaque formation similar to the parental viruses (data not shown).</p><p id="par0090">Purification of the genomic dsRNAs from cells infected with independent plaques of BTV1/8VP2.5 and BTV1/8VP2.5.NS3 and their subsequent analysis on a non-denaturing polyacrylamide gel confirmed that the recovered reassortant viruses had the expected dsRNA genome patterns that were equivalent to the parental origin of the segments (<xref rid="fig0005" ref-type="fig">Fig. 1</xref>A, left panel). In addition, RT-PCR and sequencing also validated the authenticity of the segments that could not be assessed by their mobility in non-denaturing polyacrylamide gels.<fig id="fig0005"><label>Fig. 1</label><caption><p>Characterization of reassortant BTV-1/BTV-8 viruses. (A) Left panel, genomic profile of reassortant viruses generated by RG. DsRNA was extracted from cells infected with either BTV-1, BTV-8, BTV1/8VP2.5.NS3 or BTV1/8VP2.5 and analyzed in a non-denaturing polyacrylamide gel. Arrows indicate reassortant segments S2, S6 or S10 of BTV-8 in each generated virus. Right panel, virus growth profile in the ovine kidney-derived PT cell line. Cells were infected with parental BTV-1 or BTV-8, or reassortant BTV1/8VP2.5.NS3 or BTV1/8VP2.5 and samples were harvested at the indicated times. Titers at each time point were determined in triplicate by plaque assay, expressed as virus growth in plaque formation units per ml (PFU/ml) and plotted in logarithmic scale. Error bars indicate standard deviation. Asterisk and hashtag indicates the significance (<italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.05) of the difference in titers of BTV1/8VP2.5.NS3 or BTV1/8VP2.5 at 12, 24 and 48&#x000a0;h p.i. to WT BTV-1 (*) or BTV-8 (#). (B) Characterization of BTV1/8VP2.5.NS3 in ovine thymus-derived SFT-R cells. Left panel, virus growth. Cells were infected with BTV1/8VP2.5.NS3 or parental BTV-1 or BTV-8 and processed as indicated in (A). The <italic>p</italic> values were designated the same way as in (A). Right panel, protein expression profile from cells infected with BTV1/8VP2.5.NS3 at the indicated time points. Viral proteins were detected by Western blot using specific antibodies against VP5, VP7 or NS3. As loading control, an antibody against cellular actin was used. Protein profile from cells infected with BTV-1 and BTV-8 was used as controls.</p></caption><graphic xlink:href="gr1"/></fig></p><p id="par0095">Since pathogenicity of BT disease is more pronounced in sheep, ovine cells were evaluated for growth kinetics and protein expression at different times p.i. For this purpose, first an ovine cell line (PT) derived from kidney was infected at MOI of 1 with either wild-type BTV-1 or BTV-8 or reassortant virus BTV-1/8VP2.5 or BTV-1/8VP2.5.NS3. Virus growth kinetics was monitored by plaque assay (<xref rid="fig0005" ref-type="fig">Fig. 1</xref>A, right panel) of samples taken at 0, 12, 24 and 48&#x000a0;h p.i. Although both reassortant viruses BTV1/8VP2.5 and BTV1/8VP2.5.NS3 consisted of a BTV-1 backbone, the growth characteristics of only the triple reassortant were different to that of BTV-1. Reassortant virus BTV1/8VP2.5 exhibited a similar profile to BTV-1, with only a slight difference at 12&#x000a0;h (<italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.05) but not at (<italic>p</italic>&#x000a0;&#x0003e;&#x000a0;0.05) either 24 or 48&#x000a0;h p.i. In contrast, the triple reassortant BTV-1/8VP2.5.NS3 virus titer at 12&#x000a0;h p.i. was significantly different from BTV-1 (<italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.05) but not from BTV-8 at the same time point (<italic>p</italic>&#x000a0;&#x0003e;&#x000a0;0.05). However, growth of this virus at later times (24 and 48&#x000a0;h p.i.) exhibited a somewhat different (<italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.05) profile in comparison to both parental viruses, BTV-1 and BTV-8, albeit closer to BTV-8.</p><p id="par0100">Since triple reassortant virus BTV1/8VP2.5.NS3 appeared to have a different growth profile to that of BTV-1, the growth patterns of these two viruses were further analyzed using an alternate ovine cell line, ovine thymus SFT-R cells. As VP2, VP5 and NS3 in the triple reassortant virus BTV1/8VP2.5.NS3 were taken from BTV-8, the latter was also included as a control. Analysis of the growth profiles of the three viruses in SFT-R cells (<xref rid="fig0005" ref-type="fig">Fig. 1</xref>B, left panel) exhibited that the virus titers of the reassortant virus, BTV-1 and BTV-8 at 12&#x000a0;h p.i. were similar and there was no statistically significant difference (<italic>p</italic>&#x000a0;&#x0003e;&#x000a0;0.05) between them. Although virus titers of reassortant and BTV-8 plateaued by 24&#x000a0;h p.i., the only virus that still continued to grow at the later time points was the BTV-1. Further, Western blot analysis of viral protein expression, using antibodies for NS3, VP5 and a major structural protein VP7 (<xref rid="fig0005" ref-type="fig">Fig. 1</xref>B, right panel), revealed that all three viruses expressed these proteins as early as 12&#x000a0;h p.i. Since the virus titers of the triple reassortant virus, BTV-1 and BTV-8 were not statistically significant from each other at 12&#x000a0;h p.i., the protein expression for all three viruses was measured at 24 and 48&#x000a0;h p.i. Further, densitometric analysis of protein expression at each time point indicated that the increase in expression of VP7 and VP5, but not NS3, was similar to BTV-8; while NS3 was more similar to BTV-1 (results not shown). Altogether, this suggested that exchanging S2, S6 and S10 segments of BTV-1 with that of BTV-8 resulted in a reassortant virus that was more similar to BTV-8 than BTV-1. In addition, the effect of the reassortment was more pronounced in the thymus-derived (SFT-R) than in the kidney (PT) cells.</p></sec><sec id="sec0055"><label>3.2</label><title>Pathogenicity in sheep infected with BTV reassortant viruses</title><p id="par0105">Since in tissue culture the triple reassortant virus BTV1/8VP2.5.NS3 behaved more similarly to BTV-8 and BTV-1/8VP2.5 to BTV-1, experiments were designed to analyze the pathogenicity of these reassortant viruses in animal hosts. For this purpose, four groups containing four sheep in each group were infected with BTV-1 or BTV-8 or BTV1/8VP2.5 or BTV1/8VP2.5.NS3. Two animals injected with saline buffer were used as the control group. The pathogenicity of each virus infection was analyzed by clinical observation of the animals in all groups. A clinical score was assigned daily to each sheep based on a range of clinical symptoms as described in Section <xref rid="sec0010" ref-type="sec">2</xref>. Depending on severity, up to four points were awarded per category. Animals infected with the reassortant BTV1/8VP2.5.NS3 presented a slightly early onset of the highest clinical score compared to BTV1/8VP2.5 or BTV-1 (<xref rid="fig0010" ref-type="fig">Fig. 2</xref>A). As the same pattern was also observed for BTV-8 with a maximum clinical score at 6 days p.i., this data suggested that the NS3 might influence virus pathogenicity and might play an important role in the onset of BT disease. Although the clinical scores in all four infection groups were different from the control group, the observed differences between the infection groups were not statistically significant.<fig id="fig0010"><label>Fig. 2</label><caption><p>Clinical manifestation and virus replication in sheep infected with reassortant BTV1/BTV-8 viruses. (A) Animals were inoculated with reassortants BTV1/8VP2.5.NS3, BTV1/8VP2.5 or parental BTV-1 or BTV-8. Infected animals were examined daily. Attitude, body temperature and other clinical signs were recorded and scored as outlined in the text. Means and standard deviation were calculated for each group. (B) Animals were inoculated with virus as above and blood samples for detection of NS1 gene by real-time RT-PCR were taken at the indicated time points. As control, a group of animals injected with saline buffer was included in the study.</p></caption><graphic xlink:href="gr2"/></fig></p><p id="par0110">Subsequently, viremia in these infected animals was estimated by analyzing the whole blood samples that were taken before virus inoculation and on days 1 to 7, 9, 11, 14, 17, 19, and 21 p.i. using real-time RT-PCR (<xref rid="bib0140" ref-type="bibr">Toussaint et al., 2007</xref>). Animals infected with the reassortant viruses showed an increase in the number of NS1 gene copies over the first week after infection (<xref rid="fig0010" ref-type="fig">Fig. 2</xref>B) similar to BTV-1 and BTV-8 viruses, indicating that these viruses can replicate as efficiently as parental viruses in an animal host. In comparison, samples from control sheep were free of BTV RNA. (<xref rid="fig0010" ref-type="fig">Fig. 2</xref>B, control). While productive infection measured by detecting the presence of antibodies against VP7, the group-specific antigen, by ELISA, also demonstrated that all animals in the infection groups had seroconverted by day 7 after inoculation, no BTV antibody response was detected in any sample from the negative control sheep (data not shown). In addition, the serum neutralization response tested in samples taken at 21 days p.i. confirmed that animals infected with the reassortant viruses BTV1/8VP2.5.NS3 and BTV1/8VP2.5 showed a strong neutralization response against BTV-8 but not BTV-1 (<xref rid="tbl0005" ref-type="table">Table 1</xref>). In comparison, animals infected with BTV-1 presented a response only against the same serotype. Overall these data indicated that the reassortant viruses are able to replicate in animals and they also generate a strong immune response similar to that of the parental strains. Since the sheep infected with the reassortant BTV1/8VP2.5.NS3 showed a clinical score that was comparable to that of BTV-8, it was hypothesized that NS3 might be playing an important role in the pathogenicity of BT disease.<table-wrap position="float" id="tbl0005"><label>Table 1</label><caption><p>Serum neutralization titer in animals infected with BTV1/8VP2.5.NS3, BTV1/8VP2.5 or parental BTV-1 or BTV-8.</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left">Infection group</th><th align="left">BTV-1 (dilution)<xref rid="tblfn0005" ref-type="table-fn">a</xref></th><th align="left">BTV-8 (dilution)<xref rid="tblfn0005" ref-type="table-fn">a</xref></th></tr></thead><tbody><tr><td rowspan="4" align="left" valign="middle">BTV-1 (4 animals)</td><td align="left">1/640</td><td align="left">Negative</td></tr><tr><td align="left">1/240</td><td align="left">Negative</td></tr><tr><td align="left">1/240</td><td align="left">Negative</td></tr><tr><td align="left">1/480</td><td align="left">Negative</td></tr><tr><td colspan="3" align="left">&#x02028;&#x02028;</td></tr><tr><td rowspan="4" align="left" valign="middle">BTV-8 (4 animals)</td><td align="left">Negative</td><td align="left">1/480</td></tr><tr><td align="left">Negative</td><td align="left">1/240</td></tr><tr><td align="left">Negative</td><td align="left">1/240</td></tr><tr><td align="left">Negative</td><td align="left">1/320</td></tr><tr><td colspan="3" align="left">&#x02028;&#x02028;</td></tr><tr><td rowspan="4" align="left" valign="middle">BTV1/8VP2.5 (4 animals)</td><td align="left">Negative</td><td align="left">1/240</td></tr><tr><td align="left">Negative</td><td align="left">1/240</td></tr><tr><td align="left">Negative</td><td align="left">1/120</td></tr><tr><td align="left">Negative</td><td align="left">1/120</td></tr><tr><td colspan="3" align="left">&#x02028;&#x02028;</td></tr><tr><td rowspan="4" align="left" valign="middle">BTV1/8VP2.5.NS3 (4 animals)</td><td align="left">Negative</td><td align="left">1/480</td></tr><tr><td align="left">Negative</td><td align="left">1/320</td></tr><tr><td align="left">Negative</td><td align="left">1/480</td></tr><tr><td align="left">Negative</td><td align="left">1/120</td></tr><tr><td colspan="3" align="left">&#x02028;&#x02028;</td></tr><tr><td rowspan="2" align="left" valign="middle">PBS (2 animals)</td><td align="left">Negative</td><td align="left">Negative</td></tr><tr><td align="left">Negative</td><td align="left">Negative</td></tr></tbody></table><table-wrap-foot><fn><p>Negative indicates no detectable neutralizing activity in a 1/10 dilution.</p></fn></table-wrap-foot><table-wrap-foot><fn id="tblfn0005"><label>a</label><p id="npar0005">Neutralizing activity in serum samples from sheep 21 days post-infection with reassortant or parental virus was determined as the highest dilution that neutralized 100 TCID₅₀ virus in 50% of replicate wells.</p></fn></table-wrap-foot></table-wrap></p></sec><sec id="sec0060"><label>3.3</label><title>Importance of NS3 in bluetongue disease in sheep</title><p id="par0115">To further determine the importance of NS3 in pathogenicity, we generated a reassortant virus in which S10 of BTV-1 was exchanged with S10 of BTV-8, BTV1/8NS3 by RG system (<xref rid="fig0015" ref-type="fig">Fig. 3</xref>A). In addition, two other reassortants BTV1/8VP2.NS3 and BTV1/8VP5.NS3 were generated to determine the contribution of each segment in the context of NS3. Further, a sequence comparison between the segment S10 of BTV-1 and BTV-8 showed that only 12 residues in the NS3 sequence are different between these two serotypes, of which 5 are located at the putative extracellular domain of NS3. To determine if these differences play any role in pathogenicity, we introduced the substitutions L₁₄₉I/S₁₅₃K/A₁₅₄T/I₁₅₆V/Q₁₅₈S in the BTV-1 NS3 sequence and generated a mutant virus BTV1/NS3_IKTVS that expressed an NS3 protein containing an extracellular domain similar to that of BTV-8. Each reassortant virus was grown in BSR cells and stocks were stored at 4&#x000a0;&#x000b0;C. Although all viruses yielded titers between 1.5&#x000a0;&#x000d7;&#x000a0;10⁷ and 4.6&#x000a0;&#x000d7;&#x000a0;10⁷&#x000a0;PFU/ml, BTV1/8VP5.NS3 yielded consistently lower titers (&#x0223c;1&#x000a0;&#x000d7;&#x000a0;10⁶) than the other reassortant viruses.<fig id="fig0015"><label>Fig. 3</label><caption><p>Characterization of reassortants BTV1/8VP2.NS3, BTV1/8VP5.NS3, BTV1/8NS3 or mutant BTV1/NS3_IKTVS viruses. (A) Genomic profile of reassortant viruses generated by reverse genetics. Genomic dsRNA was extracted from cells infected with either reassortant or parental viruses as indicated in each lane and analyzed in a non-denaturing polyacrylamide gel. (B) Virus growth in SFT-R cell line. Cells were infected with parental BTV-1 or BTV-8 or triple reassortant BTV1/8VP2.5.NS3 or dual reassortant BTV1/8VP2.NS3, BTV1/8VP5.NS3 and BTV1/8VP2.5 or single BTV1/8NS3 and BTV1/NS3_IKTVS viruses and samples were harvested at 48&#x000a0;h p.i. Total titers were determined by TCID₅₀ and expressed as percentage of BTV-8 titer considered as 100%.</p></caption><graphic xlink:href="gr3"/></fig></p><p id="par0120">Further, the growth of each virus was examined in ovine SFT-R cells. Interestingly, both BTV1/8NS3 and BTV1/NS3_IKTVS single reassortant viruses behaved like BTV-1, not BTV-8. The other two dual reassortant viruses, BTV1/8VP2.NS3 and BTV1/8VP5.NS3, behaved as an intermediate between two parental viruses as shown in <xref rid="fig0015" ref-type="fig">Fig. 3</xref>B, suggesting that NS3 might be acting in concert with each of the outer capsid proteins.</p><p id="par0125">For pathogenicity studies, animals were divided into six infection groups and a control group with four sheep in each group. Sheep in the infection groups were inoculated with BTV-1 or BTV-8 or one of four different reassortant BTV-1/BTV-8 viruses. Animals were clinically scored daily and blood samples were taken before and at days 1&#x02013;11, 14, 17, and 21 days after infection. All the generated viruses (reassortants and mutant) were able to replicate in sheep, as real-time RT-PCR demonstrated an increasing amount of S6 RNA in blood samples over time (data not shown). When the animals were clinically scored, similar to the parental BTV-1, a delay in pathogenicity was observed for the virus BTV1/8VP2.NS3 with a maximum score at 8 days p.i. (<xref rid="fig0015" ref-type="fig">Fig. 3</xref>B). The reassortants BTV1/8VP5.NS3, BTV1/8NS3, mutant BTV1/NS3I_KTVS and BTV-8 reached a maximum score at seven days p.i. (<xref rid="fig0020" ref-type="fig">Fig. 4</xref>A and B). The clinical scores in all infection groups were significantly different from the control group; however, the observed differences between the infection groups were not statistically significant.<fig id="fig0020"><label>Fig. 4</label><caption><p>Pathogenicity of reassortants BTV1/8VP2.NS3, BTV1/8VP5.NS3, BTV1/8NS3 or mutant BTV1/NS3_IKTVS viruses. (A, B) Clinical signs were recorded for sheep infected with BTV1/8VP2.NS3 (A), BTV1/8VP5.NS3 (A), BTV1/8NS3 (B) or mutant BTV1/NS3_IKTVS (B) viruses using the same scoring described in <xref rid="fig0010" ref-type="fig">Fig. 2</xref>. Parental BTV-1 and BTV-8 viruses were also included as control (same set of data was plotted in each graph).</p></caption><graphic xlink:href="gr4"/></fig></p></sec></sec><sec id="sec0065"><label>4</label><title>Discussion</title><p id="par0130">BTV serotypes, BTV-1 and BTV-8, have displayed a marked ability to spread and severe virulence has been reported during European outbreaks in domestic ruminants, sheep being among the most severely affected livestock species (<xref rid="bib0005" ref-type="bibr">Allepuz et al., 2010</xref>, <xref rid="bib0060" ref-type="bibr">Elbers et al., 2008b</xref>). Detailed investigations in sheep to elucidate the pathogenic mechanisms of BTV-1 under experimental conditions are scarce (<xref rid="bib0075" ref-type="bibr">Hamblin et al., 1998</xref>, <xref rid="bib0125" ref-type="bibr">Perez de Diego et al., 2011</xref>). In comparison, infections carried out with BTV-8 in sheep have been accompanied by a significant variation in results (<xref rid="bib0045" ref-type="bibr">Darpel et al., 2007</xref>, <xref rid="bib0065" ref-type="bibr">Elbers et al., 2008c</xref>, <xref rid="bib0155" ref-type="bibr">Worwa et al., 2010</xref>). Recently, two studies have also reported a comparative study involving pathogenicity of BTV-1 and -8 in sheep and cattle (<xref rid="bib0040" ref-type="bibr">Dal Pozzo et al., 2013</xref>, <xref rid="bib0135" ref-type="bibr">Sanchez-Cordon et al., 2013</xref>). This is the first study to investigate how the difference in the clinical manifestations between BTV-1 and BTV-8 is influenced by the genetic origin of the virus. Since there is some evidence in previous reports regarding the involvement of VP2 and VP5, the two most exposed viral capsid proteins as determinants of virus virulence (<xref rid="bib0050" ref-type="bibr">DeMaula et al., 2000</xref>, <xref rid="bib0080" ref-type="bibr">Huismans and Howell, 1973</xref>), we have explored whether variations in pathogenicity are likely to be associated with the two most variable BTV genomic segments, S2 and S6 that encode VP2 and VP5 respectively. In addition, as S10, the gene encoding for NS3 has also been found to be the most variable amongst the segments encoding for the non-structural proteins in BTV infected cells, this protein was also hypothesized to play a potentially important role in BTV pathogenesis. Interestingly, in BTV infected cells it has also been demonstrated that the interaction of VP2 and VP5 with NS3 has an important role in virus assembly and trafficking (<xref rid="bib0010" ref-type="bibr">Beaton et al., 2002</xref>, <xref rid="bib0015" ref-type="bibr">Bhattacharya and Roy, 2008</xref>, <xref rid="bib0025" ref-type="bibr">Celma and Roy, 2009</xref>, <xref rid="bib0030" ref-type="bibr">Celma and Roy, 2011</xref>, <xref rid="bib0150" ref-type="bibr">Wirblich et al., 2006</xref>).</p><p id="par0135">For this purpose, the BTV RG system was utilized to generate reassortant viruses that consisted of a backbone of BTV-1, with VP2, VP5 and NS3 of BTV-8 in five different combinations (BTV1/8VP2.5, BTV1/8VP2.5.NS3, BTV1/8VP2.NS3, BTV1/8VP5.NS3 and BTV1/8NS3). In each of the reassortants the equivalent genome segment of BTV-1 was replaced by BTV-8. This is the first time that reassortant viruses generated by RG technology have been tested in an animal host.</p><p id="par0140">In order to investigate the influence of tissue specificity in the growth pattern of the reassortant viruses and to compare it with the two parental strains of BTV-1 and BTV-8, sheep kidney (PT) and thymus (SFT-R) cell lines were infected with the respective viruses. Since our primary objective was to assess the importance of the VP2, VP5 and NS3 in BTV pathogenesis, growth curves and virus protein production of reassortant virus BTV1/8VP2.5.NS3 were assessed in the two different ovine cell types. Our results showed that both wild-type viruses as well as the reassortant viruses had different growth profiles in each cell line, indicating that there may be a tissue specificity related to a particular strain that may play a role in virus pathogenicity. In addition, our results also revealed that the growth pattern of BTV1/8VP2.5 and not BTV1/8VP2.5.NS3 was similar to that of BTV-1 in the ovine kidney cell line. Experiments undertaken with BTV1/8VP2.5.NS3, BTV-1 and BTV-8 in ovine thymus cell lines demonstrated a more pronounced similarity in growth profile between BTV1/8VP2.5.NS3 and BTV-8 but not BTV-1. This confirmed that substituting all 3 segments, S2, S6 and S10 of BTV-1, with that of BTV-8 resulted in a reassortant virus that was more similar in behavior to BTV-8 than BTV-1. Since the effect was more pronounced in the ovine thymus SFT-R cell line, this also validated that in cell culture systems the origin of the cultured cells does play a significant role in determining virus growth characteristics. Hence, on the basis of these studies it was postulated that the growth profile of a particular BTV strain was dependent on VP2, VP5 and NS3.</p><p id="par0145">Consequently, experiments designed to analyze the pathogenicity of BTV1/8VP2.5 and BTV1/8VP2.5.NS3 in sheep revealed that parental BTV-8 and reassortant BTV1/8VP2.5.NS3 triggered an earlier onset of clinical disease symptoms than parental BTV-1 and reassortant BTV1/8VP2.5. Interestingly, our results are in direct contrast to another recent investigation undertaken in sheep that have demonstrated that BTV-1 is more pathogenic than BTV-8. The difference in pathogenicity of BTV-1 and BTV-8 between our current study and that undertaken by <xref rid="bib0135" ref-type="bibr">Sanchez-Cordon et al. (2013)</xref> can be attributed to the fact that the BTV-1 strain used to recover our reassortant viruses was a tissue culture adapted strain obtained from South Africa. Subsequently, RT-PCR on blood samples procured from infected sheep and serum neutralization tests also confirmed that BTV1/8VP2.5 and BTV1/8VP2.5.NS3 were able to replicate in animals and that they also were able to generate a strong immune response similar to that of the parental strains. Since sheep infected with reassortant BTV1/8VP2.5.NS3 showed similar clinical scores to that of BTV-8, it led us to further hypothesize that of the 3 proteins, VP2, VP5 and NS3, the non-structural protein NS3 might be an important contributing factor for the development of BT disease. Hence, three more reassortant viruses were generated by RG system in which S10 of BTV-8 was either present singly (BTV1/8NS3) or in a combination with either BTV-8 VP2 (BTV1/8VP2.NS3) or VP5 (BTV1/8VP5.NS3). Sequence comparison of S10 belonging to BTV-1 and BTV-8 highlighted the presence of only 12 amino acid residues that were variable between the two serotypes. Since a cluster containing 5 of these variable residues was located in the putative extracellular domain of NS3, a mutant virus was generated (BTV1/NS3_IKTVS) that only contained amino acid substitutions (L₁₄₉I/S₁₅₃K/A₁₅₄T/I₁₅₆V/Q₁₅₈S) in this region of NS3. Consequently, this domain of BTV-1 NS3 became identical to the extracellular domain of BTV-8. When the reassortant and the mutant virus were inoculated in sheep, real-time RT-PCR of blood samples demonstrated that all the newly generated viruses and the two parental strains could replicate in sheep. Although pathogenicity analysis of the reassortant viruses suggested that BTV1/8VP2.NS3 behaved similarly to BTV-1, while the other two reassortant viruses (BTV1/8NS3 and BTV1/8VP5.NS3) and the mutant virus BTV1/NS3_IKTVS were similar to BTV-8, the overall clinical scores were not significantly different. Based on our result it was hypothesized that the two outer capsid proteins (VP2 and VP5) and the non-structural protein NS3 are all acting in concert with each other and are to a certain extent responsible for the pathogenicity of a particular strain. Interestingly, earlier reports have demonstrated that the interaction of NS3 with the two outer capsid proteins and cytoplasmic proteins have important consequences for both virus assembly and trafficking in infected cells (<xref rid="bib0010" ref-type="bibr">Beaton et al., 2002</xref>, <xref rid="bib0015" ref-type="bibr">Bhattacharya and Roy, 2008</xref>, <xref rid="bib0025" ref-type="bibr">Celma and Roy, 2009</xref>, <xref rid="bib0030" ref-type="bibr">Celma and Roy, 2011</xref>, <xref rid="bib0150" ref-type="bibr">Wirblich et al., 2006</xref>). Since mutant virus BTV1/NS3_IKTVS, consisting of an extracellular loop region identical to that of BTV-8 NS3, did not show a statistical significant difference in virus pathogenicity when compared to the two parental strains, it can be concluded that the loop region of NS3 or its interaction with various cellular partners does not have a significant contribution to BT pathogenesis. Although the role of the remaining seven highly conserved BTV genomic segments cannot be negated, further work on this aspect is beyond the remit of this paper. Clinical experiments undertaken with BTV-1 and BTV-8 showed a different profile in two different sets of experiments (<xref rid="fig0010" ref-type="fig">Fig. 2</xref>, <xref rid="fig0020" ref-type="fig">Fig. 4</xref>). Since these two animal experiments were conducted over one year apart, it is possible that in spite of every effort that was made to keep conditions consistent between experiments, some variations occurred. However, it is unlikely that the small variation between experiments invalidates the findings made within each of the experiments. The role of NS3 in regulating interferon response has recently been reported (<xref rid="bib0035" ref-type="bibr">Chauveau et al., 2013</xref>). It will be interesting to investigate interferon responses of the triple reassortant virus versus the two wild-type viruses in the animal. However, these will require a series of experiments and detailed analysis that are beyond the remit of this study.</p></sec><sec id="sec0070"><label>5</label><title>Conclusion</title><p id="par0150">The clinical disease and pathogenicity of BTV is a very complex process and cannot be explained by the presence or absence of only one genome segment or its protein product. Much more work needs to be undertaken to clearly understand the genetic basis of BT pathogenesis and to our knowledge, this is the first report in this particular research area.</p></sec>

The binding characteristics and orientation of a novel radioligand with distinct properties at 5-HT₃A and 5-HT₃AB receptors

Could not extract abstract

<contrib contrib-type="author" id="au1"><name><surname>Thompson</surname><given-names>Andrew J.</given-names></name><xref rid="aff2" ref-type="aff">b</xref></contrib><contrib contrib-type="author" id="au2"><name><surname>Verheij</surname><given-names>Mark H.P.</given-names></name><xref rid="aff1" ref-type="aff">a</xref></contrib><contrib contrib-type="author" id="au3"><name><surname>Verbeek</surname><given-names>Joost</given-names></name><xref rid="aff3" ref-type="aff">c</xref></contrib><contrib contrib-type="author" id="au4"><name><surname>Windhorst</surname><given-names>Albert D.</given-names></name><xref rid="aff3" ref-type="aff">c</xref></contrib><contrib contrib-type="author" id="au5"><name><surname>de Esch</surname><given-names>Iwan J.P.</given-names></name><xref rid="aff1" ref-type="aff">a</xref></contrib><contrib contrib-type="author" id="au6"><name><surname>Lummis</surname><given-names>Sarah C.R.</given-names></name><email>sl120@cam.ac.uk</email><xref rid="aff2" ref-type="aff">b</xref><xref rid="cor1" ref-type="corresp">&#x02217;</xref></contrib><aff id="aff1"><label>a</label>Amsterdam Institute for Molecules Medicines and Systems (AIMMS), Division of Medicinal Chemistry, Faculty of Sciences, VU University Amsterdam, Amsterdam, The Netherlands</aff><aff id="aff2"><label>b</label>Department of Biochemistry, University of Cambridge, Cambridge, UK</aff><aff id="aff3"><label>c</label>VU University Medical Center, Dept Radiology &#x00026; Nuclear Medicine, Amsterdam, The Netherlands</aff>

Neuropharmacology

<sec id="sec1"><label>1</label><title>Introduction</title><p id="p0035">5-HT₃ receptors are transmembrane ligand-gated ion-channels that are responsible for fast synaptic neurotransmission in the central and peripheral nervous systems. They are composed of five subunits, each of which contains an extracellular, a transmembrane and an intracellular domain (<xref rid="bib21" ref-type="bibr">Thompson et&#x000a0;al., 2008a</xref>; <xref rid="bib10" ref-type="bibr">Miller and Smart, 2012</xref>). <italic>In&#x000a0;vivo</italic> 5-HT₃ receptor activation can result in nausea and vomiting, and for over three decades competitive antagonists of these receptors have been used to alleviate these symptoms arising from cancer therapy and general anaesthetics. There is also a limited use of antagonists for treating irritable bowel syndrome and pre-clinical interest in the use of partial agonists for the same disorder (<xref rid="bib24" ref-type="bibr">Thompson and Lummis, 2007</xref>, <xref rid="bib32" ref-type="bibr">Walstab et&#x000a0;al., 2010</xref>, <xref rid="bib20" ref-type="bibr">Thompson, 2013</xref>).</p><p id="p0040">There are currently five 5-HT₃ receptor subunits (5-HT3A&#x02013;5-HT3E), with further complexity arising from splice variants and species differences (<xref rid="bib32" ref-type="bibr">Walstab et&#x000a0;al., 2010</xref>). 5-HT3A subunits can form homomeric receptors, but the subunits 5-HT3B&#x02013;5-HT3E must combine with 5-HT3A subunits to function. The functional properties of these receptor subtypes have been reported by several groups, but to date only the pharmacologies of 5-HT₃A and 5-HT₃AB receptors have been studied in detail (<xref rid="bib7" ref-type="bibr">Holbrook et&#x000a0;al., 2009</xref>, <xref rid="bib32" ref-type="bibr">Walstab et&#x000a0;al., 2010</xref>, <xref rid="bib28" ref-type="bibr">Thompson et&#x000a0;al., 2013</xref>, <xref rid="bib23" ref-type="bibr">Thompson and Lummis, 2013</xref>). Until recently only pore-blocking antagonists were known to have different properties at 5-HT₃A and 5-HT₃AB receptors, and these differences could be attributed to the varying pore-lining amino acids of the 5-HT3A and 5-HT3B subunits (<xref rid="bib23" ref-type="bibr">Thompson and Lummis, 2013</xref>). However, the utility of these compounds is limited as they tend to be of low affinity (&#x003bc;M range) and also target other receptor types. More recently there have been descriptions of two compounds with other sites of action that discriminate between 5-HT₃A and 5-HT₃AB receptor subtypes. One of these, topotecan, primarily an anticancer drug, was found to inhibit 5-HT₃A and potentiate 5-HT₃AB receptors, although this compound also has a relatively low (&#x003bc;M) potency (<xref rid="bib12" ref-type="bibr">Nakamura et&#x000a0;al., 2013</xref>). The second compound is VUF10166 (2-chloro-3-(4-methylpiperazin-1-yl)quinoxaline), which is highly potent, with an affinity at 5-HT₃A receptors (p<italic>K</italic>_i&#x000a0;&#x0223c;&#x000a0;10) that is &#x0223c;100-fold greater than at 5-HT₃AB receptors (<xref rid="bib27" ref-type="bibr">Thompson et&#x000a0;al., 2012</xref>). We previously showed that VUF10166 binds to the orthosteric binding site of both 5-HT₃A and 5-HT₃AB receptors (formed at the interface of two 5-HT3A subunits, A+A&#x02212;) and that a second, allosteric, binding site (A+B&#x02212;) in the 5-HT₃AB receptor was responsible for causing ligands at the A+A&#x02212; binding site to dissociate more rapidly.</p><p id="p0045">Here we perform a detailed characterisation of VUF10166 binding to 5-HT₃A and 5-HT₃AB receptors with a radiolabelled version of this compound and use mutagenesis to explore the residues that interact with VUF10166 at the A+A&#x02212; binding site.</p></sec><sec id="sec2"><label>2</label><title>Experimental procedures</title><sec id="sec2.1"><label>2.1</label><title>Synthesis of [³H]VUF10166</title><p id="p0050">60&#x000a0;&#x003bc;l [³H]methyl nosylate (0.7&#x000a0;GBq/ml, 19&#x000a0;mCi/ml) in hexane/ethyl acetate (10/2 v/v) was injected into a closed reaction screwcap reaction vessel and the solvent evaporated under argon at 60&#x000a0;&#x000b0;C. 2-chloro-3-(piperazin-1-yl)quinoxaline hydrochloride (7.2&#x000a0;mg, 0.025&#x000a0;mmol) in dry DMF (150&#x000a0;&#x003bc;l) and DIPEA (30.7&#x000a0;&#x003bc;L, 0.176&#x000a0;mmol) were added for 1&#x000a0;h at room temperature. The reaction was quenched with 500&#x000a0;&#x003bc;l semi-prep HPLC eluent and subjected to semi-preparative HPLC purification, using a Reprosphere C18-DE 5&#x000a0;&#x003bc;M, 50*8&#x000a0;mm column as stationary phase (Dr. Maisch, Ammerbuch-Entringen, Germany) and acetonitrile/water 75/25 (v/v) with 0.1% diisopropylethylamine as eluent at a flow of 3&#x000a0;ml&#x000a0;min^&#x02212;1, with UV monitoring at 254&#x000a0;nm (Jasco UV-1575, Jasco, de Meern, Netherlands). 30&#x000a0;s fractions were collected, 5&#x000a0;&#x003bc;l of each added to 5&#x000a0;ml scintillation fluid, and counted for 1&#x000a0;min in a beta well counter (Rackbeta 1219 LSC, LKB-Wallac, Netherlands). Fractions containing 2-chloro-3-(4-[³H]methylpiperazin-1-yl)quinoxaline were diluted with 45&#x000a0;ml sterile water and passed over a preconditioned Waters tC18 plus Sep-Pak, washed with 20&#x000a0;ml of water, and the product obtained by elution with 1.5&#x000a0;ml ethanol; 35&#x000a0;MBq (83% radiochemical yield) of [³H]VUF10166 was obtained. The specific activity of the product was 3.13&#x000a0;TBq/mmol (84.5&#x000a0;Ci/mmol) and the radiochemical purity was &#x0003e;98%, as determined by HPLC with a Platinum C18 100a, 5&#x000a0;&#x003bc;M 250*4.6&#x000a0;mm column (Grace Alltech, Breda, Netherlands) as stationary phase and acetonitrile/10&#x000a0;mM ammoniumdihydrogen phosphate buffer pH 2.5 50/50 (v/v) as eluent at a flow of 1&#x000a0;ml&#x000a0;min^&#x02212;1, with UV monitoring at 254&#x000a0;nm (Jasco UV-1575) and radioactivity monitoring (Lablogic &#x003b2;-RAM model 4, Metorix, Goedereede, Netherlands).</p></sec><sec id="sec2.2"><label>2.2</label><title>Site-directed mutagenesis</title><p id="p0055">Mutagenesis was performed using the QuikChange method (Agilent Technologies Inc., California, USA) on human 5-HT3A cDNA (accession number: P46098) cloned into pcDNA3.1 (Invitrogen, Paisley, UK). Cysteine residues were substituted for amino acids throughout each of the binding loops A&#x02013;E (<xref rid="fig1" ref-type="fig">Fig.&#x000a0;1</xref>). To facilitate comparisons with previous work, we use the numbering of the equivalent residues in the mouse 5-HT3A subunit; for human numbering 5 should be subtracted from each residue number.<fig id="fig1"><label>Fig.&#x000a0;1</label><caption><p>Radioligand binding at 5-HT₃A receptors. (<bold>a</bold>) Representative binding curves for 5-HT₃A receptors. <italic>Inset</italic> competition binding of unlabelled VUF10166 with [³H]granisetron. (<bold>b</bold>) Association of [³H]VUF10166 was fit with a mono-exponential function to yield <italic>k</italic>_obs. (<bold>c</bold>) Linear regression was used to fit <italic>k</italic>_obs against the radioligand concentration, yielding the <italic>k</italic>_on (slope) and <italic>k</italic>_off (intercept at <italic>y</italic>&#x000a0;=&#x000a0;0) values in <xref rid="tbl1" ref-type="table">Table&#x000a0;1</xref>. (<bold>d</bold>) Dissociation of [³H]VUF10166 was best fit with a single exponential (<italic>k</italic>_off&#x000a0;=&#x000a0;0.011&#x000a0;&#x000b1;&#x000a0;0.001&#x000a0;min^&#x02212;1, <italic>n</italic>&#x000a0;=&#x000a0;4). (<bold>e</bold>) For [³H]granisetron, association was also best fit with mono-exponential functions that were used to plot <italic>k</italic>_obs against the concentration to yield the <italic>k</italic>_on and <italic>k</italic>_off values in <xref rid="tbl1" ref-type="table">Table&#x000a0;1</xref>. (<bold>f</bold>) Dissociation of [³H]granisetron (<italic>k</italic>_off&#x000a0;=&#x000a0;0.011&#x000a0;&#x000b1;&#x000a0;0.001&#x000a0;min^&#x02212;1, <italic>n</italic>&#x000a0;=&#x000a0;5).</p></caption><graphic xlink:href="gr1"/></fig></p></sec><sec id="sec2.3"><label>2.3</label><title>Cell culture and transfection</title><p id="p0060">Human embryonic kidney (HEK) 293&#x000a0;cells were maintained as monolayer cultures grown on 90&#x000a0;mm tissue culture plates in DMEM:F12 (Dulbecco's Modified Eagle Medium/Nutrient Mix F12 (1:1)) with GlutaMAX&#x02122; I media (Gibco BRL, Paisley, U.K.) containing 10% foetal calf serum (HyClone, Thermo Scientific, Cramlington, UK), at 37&#x000a0;&#x000b0;C and 7% CO₂, with a humidified atmosphere. Cells were transfected using polyethyleneimine (PEI, Polysciences Inc., Eppelheim, Germany), and incubated for 2&#x02013;3 days before harvesting.</p></sec><sec id="sec2.4"><label>2.4</label><title>Radioligand binding</title><p id="p0065">Transfected HEK 293&#x000a0;cells were washed twice with phosphate buffered saline (PBS) at room temperature, scraped into 1&#x000a0;ml of ice-cold HEPES buffer (10&#x000a0;mM, pH 7.4), homogenised and frozen. After thawing, they were washed with HEPES buffer, resuspended, and 50&#x000a0;&#x003bc;g of cell suspension incubated in 0.5&#x000a0;ml HEPES buffer and the relevant concentration of radioligand at 0&#x000a0;&#x000b0;C. Non-specific binding was determined using 2&#x000a0;mM quipazine. Equilibrium reactions were incubated for at least 3&#x000a0;h for [³H] granisetron (63.5&#x000a0;Ci/mmol, PerkinElmer, Boston, Massachusetts, USA) and 48&#x000a0;h for [³H]VUF10166. Incubations were terminated by vacuum filtration onto GF/B filters pre-soaked in 0.3% polyethyleneimine, followed by three rapid washes with 3.5&#x000a0;ml ice cold buffer. Radioactivity was determined by scintillation in Ecoscint A (National Diagnostics, Atlanta, Georgia) using a Beckman LS6000SC (Fullerton, California, USA). Each method was performed on at least three independent cell samples on at least three separate days.</p><sec id="sec2.4.1"><label>2.4.1</label><title>Saturation binding</title><p id="p0070">To construct saturation binding curves a range of [³H]granisetron (0.25&#x02013;2&#x000a0;nM) or [³H]VUF10166 (0.04&#x02013;50&#x000a0;nM) concentrations were used according to the conditions described above. Final counts were monitored to ensure that binding never exceeded 10% of the added concentrations of radioligands.</p></sec><sec id="sec2.4.2"><label>2.4.2</label><title>Competition binding</title><p id="p0075">Affinities of unlabelled Cys-loop receptor ligands were determined by adding a range (2 pM&#x02013;2&#x000a0;mM) of concentrations to samples containing 0.2&#x000a0;nM [³H]VUF10166 or 0.7&#x000a0;nM [³H]granisetron for 5-HT₃A receptors, and 0.6&#x000a0;nM [³H]VUF10166 or 0.7&#x000a0;nM [³H]granisetron for 5-HT₃AB receptors.</p></sec><sec id="sec2.4.3"><label>2.4.3</label><title>Kinetic measurements</title><p id="p0080">To determine the association rate (<italic>k</italic>_on), the observed association rate (<italic>k</italic>_obs) was measured for a range of radioligand concentrations. The experiment was started (<italic>t</italic>&#x000a0;=&#x000a0;0) by the addition of radioligand to 500&#x000a0;&#x003bc;l cell suspension in HEPES buffer and harvested at varying time points to construct association curves.</p><p id="p0085">Dissociation was measured by allowing each radioligand to reach equilibrium according to the times described above and then adding a final concentration of 2&#x000a0;mM quipazine (&#x0223c;<italic>K</italic>_d&#x000a0;&#x000d7;&#x000a0;10⁶) to each tube for varying time periods.</p></sec></sec><sec id="sec2.5"><label>2.5</label><title>Data analysis</title><p id="p0090">All data were analysed using GraphPad Prism 4.03. Individual saturation binding experiments were fitted to Equ <xref rid="fd1" ref-type="disp-formula">(1)</xref>, and the values averaged to obtain mean&#x000a0;&#x000b1;&#x000a0;sem:<disp-formula id="fd1"><label>(1)</label><mml:math id="M1" altimg="si1.gif" overflow="scroll"><mml:mrow><mml:mi>y</mml:mi><mml:mo>=</mml:mo><mml:msub><mml:mi>B</mml:mi><mml:mrow><mml:mi>max</mml:mi></mml:mrow></mml:msub><mml:mo>&#x000d7;</mml:mo><mml:mo>[</mml:mo><mml:mi>L</mml:mi><mml:mo>]</mml:mo><mml:mfrac><mml:mrow><mml:msub><mml:mi>K</mml:mi><mml:mtext>d</mml:mtext></mml:msub></mml:mrow><mml:mrow><mml:msub><mml:mi>K</mml:mi><mml:mtext>d</mml:mtext></mml:msub><mml:mo>+</mml:mo><mml:mo>[</mml:mo><mml:mi>L</mml:mi><mml:mo>]</mml:mo></mml:mrow></mml:mfrac></mml:mrow></mml:math></disp-formula>where <italic>B</italic>_max is maximum binding at equilibrium, <italic>K</italic>_d is the equilibrium dissociation constant and [<italic>L</italic>] is the free concentration of radioligand. Individual competition binding experiments were analysed by iterative curve fitting using the following equation and the values averaged to obtain the mean&#x000a0;&#x000b1;&#x000a0;sem:<disp-formula id="fd2"><label>(2)</label><mml:math id="M2" altimg="si2.gif" overflow="scroll"><mml:mi>y</mml:mi><mml:mo>=</mml:mo><mml:msub><mml:mi>B</mml:mi><mml:mi>min</mml:mi></mml:msub><mml:mo>+</mml:mo><mml:mfrac><mml:mrow><mml:msub><mml:mi>B</mml:mi><mml:mi>max</mml:mi></mml:msub><mml:mo>-</mml:mo><mml:msub><mml:mi>B</mml:mi><mml:mi>min</mml:mi></mml:msub></mml:mrow><mml:mrow><mml:mn>1</mml:mn><mml:mo>+</mml:mo><mml:msup><mml:mn>10</mml:mn><mml:mrow><mml:mfenced open="[" close="]"><mml:mi>L</mml:mi></mml:mfenced><mml:mo>&#x000b7;</mml:mo><mml:mi>log</mml:mi><mml:mspace width="0.25em"/><mml:mi>I</mml:mi><mml:msub><mml:mi>C</mml:mi><mml:mn>50</mml:mn></mml:msub></mml:mrow></mml:msup></mml:mrow></mml:mfrac></mml:math></disp-formula>where <italic>B</italic>_min is the non-specific binding, <italic>B</italic>_max is the maximum specific binding, [<italic>L</italic>] is the concentration of competing ligand and <italic>IC</italic>₅₀ is the concentration of competing ligand that blocks half of the specific bound radioligand.</p><p id="p0095">A simple bimolecular binding scheme for receptor and ligand can be represented as:<disp-formula id="fd3"><label>(3)</label><mml:math id="M3" altimg="si3.gif" overflow="scroll"><mml:mrow><mml:mi>L</mml:mi><mml:mo>+</mml:mo><mml:mi>R</mml:mi><mml:munderover><mml:mo>&#x021cc;</mml:mo><mml:mrow><mml:msub><mml:mi>k</mml:mi><mml:mrow><mml:mtext>off</mml:mtext></mml:mrow></mml:msub></mml:mrow><mml:mrow><mml:msub><mml:mi>k</mml:mi><mml:mrow><mml:mtext>on</mml:mtext></mml:mrow></mml:msub></mml:mrow></mml:munderover><mml:mi>L</mml:mi><mml:mi>R</mml:mi></mml:mrow></mml:math></disp-formula>where <italic>L</italic> is the free ligand concentration, <italic>R</italic> receptor concentration, <italic>LR</italic> bound receptor concentration, and <italic>k</italic>_on and <italic>k</italic>_off microscopic association and dissociation rate constants. In a simple scheme such as this, the equilibrium dissociation constant (<italic>K</italic>_d) is equal to the ratio of dissociation to association rate constants, such that:<disp-formula id="fd4"><label>(4)</label><mml:math id="M4" altimg="si4.gif" overflow="scroll"><mml:mrow><mml:msub><mml:mi>K</mml:mi><mml:mtext>d</mml:mtext></mml:msub><mml:mo>=</mml:mo><mml:mfrac><mml:mrow><mml:msub><mml:mi>k</mml:mi><mml:mrow><mml:mtext>off</mml:mtext></mml:mrow></mml:msub></mml:mrow><mml:mrow><mml:msub><mml:mi>k</mml:mi><mml:mrow><mml:mtext>on</mml:mtext></mml:mrow></mml:msub></mml:mrow></mml:mfrac></mml:mrow></mml:math></disp-formula></p><p id="p0100">Dissociation data were fitted to either a single or double exponential decay to yield <italic>k</italic>_off. Association data were fitted to a single exponential association to calculate <italic>k</italic>_obs. If <italic>k</italic>_obs is plotted against the radioligand concentration, according to a simple model, the slope of this plot equals the association constant (<italic>k</italic>_on) and the <italic>y</italic>-intercept of this line (at <italic>x</italic>&#x000a0;=&#x000a0;0) is the dissociation constant (<italic>k</italic>_off). <italic>k</italic>_on can also be calculated as described by Hill (<xref rid="bib6" ref-type="bibr">Hill, 1909</xref>), where <italic>k</italic>_off is predetermined from radioligand dissociation rate experiments.<disp-formula id="fd5"><label>(5)</label><mml:math id="M5" altimg="si5.gif" overflow="scroll"><mml:mrow><mml:msub><mml:mi>k</mml:mi><mml:mrow><mml:mtext>on</mml:mtext></mml:mrow></mml:msub><mml:mo>=</mml:mo><mml:mfrac><mml:mrow><mml:msub><mml:mi>k</mml:mi><mml:mrow><mml:mtext>obs</mml:mtext></mml:mrow></mml:msub><mml:mo>&#x02212;</mml:mo><mml:msub><mml:mi>k</mml:mi><mml:mrow><mml:mtext>off</mml:mtext></mml:mrow></mml:msub></mml:mrow><mml:mrow><mml:mrow><mml:mo>[</mml:mo><mml:mi>L</mml:mi><mml:mo>]</mml:mo></mml:mrow></mml:mrow></mml:mfrac></mml:mrow></mml:math></disp-formula></p></sec><sec id="sec2.6"><label>2.6</label><title>Homology modelling</title><p id="p0105">The protein sequence of the human 5-HT3A subunit (accession number; P46098) was aligned with a tropisetron bound AChBP template (PDB ID; 2WNC) using FUGUE. Using Modeller 9.9, five homology models were generated using default parameters and the best model selected using Ramachandran plot analysis. For the ligand, the protonated form of VUF10166 was constructed in Chem3D Ultra 7.0 (CambridgeSoft, Cambridge, UK). The binding site was defined as being within 5&#x000a0;&#x000c5; of the &#x003b1;-carbon of W183, a residue critical in the binding of other 5-HT₃ competitive ligands. VUF10166 was docked into this site using the GOLD docking program (version 3.0, The Cambridge Crystallographic Data Centre, Cambridge, UK) with the GOLDScore function and default settings. Ten docking poses were generated for each of the five homology models and the poses visualised with PyMol v1.3.</p></sec></sec><sec id="sec3"><label>3</label><title>Results</title><sec id="sec3.1"><label>3.1</label><title>[³H]VUF10166 binding at 5-HT₃A receptors</title><p id="p0110">[³H]VUF10166 showed high affinity saturable binding at 5-HT₃A receptors with low levels (&#x0003c;5%) of non-specific binding. The <italic>K</italic>_d value was similar to the <italic>K</italic>_i value from competition of unlabelled VUF10166 with [³H]granisetron (<xref rid="fig1" ref-type="fig">Fig.&#x000a0;1</xref>a, <xref rid="tbl1" ref-type="table">Table&#x000a0;1</xref>). <italic>B</italic>_max values for [³H]VUF10166 (2229&#x000a0;&#x000b1;&#x000a0;158&#x000a0;fmol/mg, <italic>n</italic>&#x000a0;=&#x000a0;6) were comparable to those with [³H]granisetron (2263&#x000a0;&#x000b1;&#x000a0;101, <italic>n</italic>&#x000a0;=&#x000a0;6) on paired samples, suggesting that both ligands bind to the same receptor population.<table-wrap position="float" id="tbl1"><label>Table&#x000a0;1</label><caption><p>Binding parameters for VUF10166 and BRL43694.</p></caption><table frame="hsides" rules="groups"><thead><tr><th>Receptor</th><th><italic>k</italic>_on (M^&#x02212;1&#x000a0;min^&#x02212;1)</th><th><italic>k</italic>_off (min^&#x02212;1)</th><th><italic>K</italic>_d(nM) (k_on/k_off)</th><th><italic>K</italic>_d (nM) saturation</th><th><italic>K</italic>_i (nM) competition<xref rid="tbl1fna" ref-type="table-fn">a</xref></th></tr></thead><tbody><tr><td colspan="6"><italic>VUF10166</italic></td></tr><tr><td>5-HT₃A</td><td>6.25&#x000a0;&#x000d7;&#x000a0;10⁷</td><td>0.010</td><td>0.16</td><td>0.18&#x000a0;&#x000b1;&#x000a0;0.04 (11)</td><td>0.24&#x000a0;&#x000b1;&#x000a0;0.11 (12)</td></tr><tr><td>5-HT₃AB</td><td><xref rid="tbl1fnb" ref-type="table-fn">b</xref>6.15&#x000a0;&#x000d7;&#x000a0;10⁷<break/><xref rid="tbl1fnb" ref-type="table-fn">b</xref>7.23&#x000a0;&#x000d7;&#x000a0;10⁶</td><td>0.024<break/>0.162</td><td>0.38<break/>22.4</td><td>&#x02013;</td><td>36.7&#x000a0;&#x000b1;&#x000a0;12.4 (12)</td></tr><tr><td colspan="6"><italic>BRL43694</italic></td></tr><tr><td>5-HT₃A</td><td>5.90&#x000a0;&#x000d7;&#x000a0;10⁷</td><td>0.064</td><td>1.08</td><td>0.68&#x000a0;&#x000b1;&#x000a0;0.05 (12)</td><td>&#x02013;</td></tr><tr><td>5-HT₃AB</td><td>1.20&#x000a0;&#x000d7;&#x000a0;10⁸</td><td>0.074</td><td>0.62</td><td>0.74&#x000a0;&#x000b1;&#x000a0;0.10 (4)</td><td>&#x02013;</td></tr></tbody></table><table-wrap-foot><fn id="tbl1fna"><label>a</label><p id="ntpara0010">Competition binding was performed with [³H]BRL43694 and unlabelled VUF10166. <italic>k</italic>_on and <italic>k</italic>_off were calculated from plots of <italic>k</italic>_obs versus ligand concentration (<xref rid="fig1" ref-type="fig">Fig.&#x000a0;1</xref>, <xref rid="fig5" ref-type="fig">Fig.&#x000a0;5</xref>). &#x02013; not determined.</p></fn></table-wrap-foot><table-wrap-foot><fn id="tbl1fnb"><label>b</label><p id="ntpara0015">Not significantly different to 5-HT₃A (<italic>p</italic>&#x000a0;&#x0003e;&#x000a0;0.05, Student's <italic>t</italic>-test).</p></fn></table-wrap-foot></table-wrap></p></sec><sec id="sec3.2"><label>3.2</label><title>VUF10166 kinetic parameters at 5-HT₃A receptors</title><p id="p0115">Association curves for [³H]VUF10166 were best fit with a single exponential function (<xref rid="fig1" ref-type="fig">Fig.&#x000a0;1</xref>b), and the resultant rates (<italic>k</italic>_obs) plotted against ligand concentration to yield <italic>k</italic>_on and <italic>k</italic>_off (<xref rid="fig1" ref-type="fig">Fig.&#x000a0;1</xref>c, <xref rid="tbl1" ref-type="table">Table&#x000a0;1</xref>). The value for <italic>k</italic>_on was similar to values determined directly from <italic>k</italic>_obs values using Equ <xref rid="fd5" ref-type="disp-formula">(5)</xref> (8.24&#x000a0;&#x000d7;&#x000a0;10⁷&#x000a0;M&#x000a0;min^&#x02212;1). Dissociation of [³H]VUF10166 in the presence of excess cold quipazine was also monophasic (<xref rid="fig1" ref-type="fig">Fig.&#x000a0;1</xref>d), with <italic>k</italic>_off values that were similar to those determined from plots of <italic>k</italic>_obs against ligand concentration (<xref rid="tbl1" ref-type="table">Table&#x000a0;1</xref>). <italic>K</italic>_d values calculated from these kinetic measurements (Equ <xref rid="fd4" ref-type="disp-formula">(4)</xref>) were similar to those derived from the saturation and competition binding (<xref rid="tbl1" ref-type="table">Table&#x000a0;1</xref>). These results indicate [³H]VUF101666 binding can be best described by a simple bi-molecular binding scheme.</p></sec><sec id="sec3.3"><label>3.3</label><title>Specificity of binding</title><p id="p0120">A range of competitive and non-competitive ligands of 5-HT₃ and related Cys-loop receptors were tested for their ability to compete with [³H]VUF10166 binding (<xref rid="tbl2" ref-type="table">Table&#x000a0;2</xref>). All tested 5-HT₃ receptor competitive ligands (agonists and antagonists) displaced specific [³H]VUF10166 binding. Binding was unaffected by the non-competitive ligands bilobalide, ginkgolide and picrotoxin, or the majority of competitive ligands of other Cys-loop receptors. Exceptions were strychnine (glycine receptor antagonist) and nicotine (nACh receptor agonist); these were later shown to also compete with [³H]granisetron.<table-wrap position="float" id="tbl2"><label>Table&#x000a0;2</label><caption><p>Competition of Cys-loop receptor ligands with [³H]VUF10166.</p></caption><table frame="hsides" rules="groups"><thead><tr><th rowspan="2">Compound</th><th colspan="2">p<italic>IC</italic>₅₀<hr/></th></tr><tr><th>5-HT₃A</th><th>5-HT₃AB</th></tr></thead><tbody><tr><td>Allosetron</td><td>11.14&#x000a0;&#x000b1;&#x000a0;0.01 (4)</td><td>11.15&#x000a0;&#x000b1;&#x000a0;0.10 (4)</td></tr><tr><td>Quipazine</td><td>8.84&#x000a0;&#x000b1;&#x000a0;0.03 (4)</td><td>8.60&#x000a0;&#x000b1;&#x000a0;0.75 (5)</td></tr><tr><td>MDL72222</td><td>12.90&#x000a0;&#x000b1;&#x000a0;0.01 (3)</td><td>13.23&#x000a0;&#x000b1;&#x000a0;0.11(1)</td></tr><tr><td><italic>m</italic>CPBG</td><td>7.49&#x000a0;&#x000b1;&#x000a0;0.06 (4)</td><td>6.07&#x000a0;&#x000b1;&#x000a0;0.20 (5)</td></tr><tr><td>Granisetron</td><td>10.48&#x000a0;&#x000b1;&#x000a0;0.08 (4)</td><td>10.35&#x000a0;&#x000b1;&#x000a0;0.10 (3)</td></tr><tr><td><italic>d</italic>-Tubocurarine</td><td>5.41&#x000a0;&#x000b1;&#x000a0;0.06 (4)</td><td>5.44&#x000a0;&#x000b1;&#x000a0;0.30 (4)</td></tr><tr><td>5-HT</td><td>4.54&#x000a0;&#x000b1;&#x000a0;0.07 (3)</td><td>4.49&#x000a0;&#x000b1;&#x000a0;0.09 (3)</td></tr><tr><td>ACh</td><td>NB (4)</td><td>NB (3)</td></tr><tr><td>GABA</td><td>NB (4)</td><td>NB (3)</td></tr><tr><td>Glycine</td><td>NB (4)</td><td>NB (3)</td></tr><tr><td>Gabazine</td><td>NB (4)</td><td>NB (3)</td></tr><tr><td>Bicuculline</td><td>NB (5)</td><td>NB (3)</td></tr><tr><td>Strychnine</td><td>5.83&#x000a0;&#x000b1;&#x000a0;0.09 (4)</td><td>6.26&#x000a0;&#x000b1;&#x000a0;0.01 (2)</td></tr><tr><td>Picrotoxin</td><td>NB (3)</td><td>NB (3)</td></tr><tr><td>Bilobalide</td><td>NB (3)</td><td>NB (2)</td></tr><tr><td>Ginkgolide</td><td>NB (3)</td><td>NB (3)</td></tr><tr><td>Nicotine</td><td>6.81&#x000a0;&#x000b1;&#x000a0;0.23 (4)</td><td>6.76&#x000a0;&#x000b1;&#x000a0;0.09 (2)</td></tr></tbody></table></table-wrap></p><p id="p0125">Previously we showed that unlabelled VUF10166 does not compete with [³H]epibatidine at &#x003b1;7 nACh receptors (the closest pharmacologically related receptor) (<xref rid="bib27" ref-type="bibr">Thompson et&#x000a0;al., 2012</xref>). Here we performed saturation binding experiments on &#x003b1;7 nACh receptors using [³H]VUF10166 which revealed no specific saturable binding (data not shown).</p><p id="p0130">These results show that classical 5-HT₃ receptor competitive antagonists compete with [³H]VUF10166, showing it binds at the orthosteric site.</p></sec><sec id="sec3.4"><label>3.4</label><title>Granisetron binding at 5-HT₃A receptors</title><p id="p0135">To compare [³H]VUF10166 with a well-established 5-HT₃ receptor competitive ligand, experiments were also conducted using [³H]granisetron. As expected, [³H]granisetron showed high affinity binding at 5-HT₃A receptors (<xref rid="tbl1" ref-type="table">Table&#x000a0;1</xref>). Competition binding with a range of known 5-HT₃ receptor agonists and antagonists gave <italic>K</italic>_i values similar to those determined using competition with [³H]VUF10166 (<xref rid="tbl3" ref-type="table">Table&#x000a0;3</xref>) and to those published elsewhere (<xref rid="bib3" ref-type="bibr">Brady et&#x000a0;al., 2001</xref>). Similar to [³H]VUF10166, nicotine and strychnine competed with [³H]granisetron.<table-wrap position="float" id="tbl3"><label>Table&#x000a0;3</label><caption><p>Competition of Cys-loop receptor ligands with [³H]BRL43694.</p></caption><table frame="hsides" rules="groups"><thead><tr><th rowspan="2">Compound</th><th colspan="2">p<italic>IC</italic>₅₀<hr/></th></tr><tr><th>5-HT₃A</th><th>5-HT₃AB</th></tr></thead><tbody><tr><td>Quipazine</td><td align="char">8.60&#x000a0;&#x000b1;&#x000a0;0.02 (5)</td><td align="char">8.12&#x000a0;&#x000b1;&#x000a0;0.18 (5)</td></tr><tr><td>MDL72222</td><td align="char">8.05&#x000a0;&#x000b1;&#x000a0;0.09 (3)</td><td align="char">7.96&#x000a0;&#x000b1;&#x000a0;0.15 (3)</td></tr><tr><td><italic>m</italic>CPBG</td><td align="char">6.88&#x000a0;&#x000b1;&#x000a0;0.13 (7)</td><td align="char">6.64&#x000a0;&#x000b1;&#x000a0;0.12 (5)</td></tr><tr><td>Granisetron</td><td align="char">9.12&#x000a0;&#x000b1;&#x000a0;0.05 (7)</td><td align="char">9.14&#x000a0;&#x000b1;&#x000a0;0.09 (4)</td></tr><tr><td><italic>d</italic>-Tubocurarine</td><td align="char">4.61&#x000a0;&#x000b1;&#x000a0;0.17 (3)</td><td align="char">4.29&#x000a0;&#x000b1;&#x000a0;0.47 (3)</td></tr><tr><td>5-HT</td><td align="char">6.38&#x000a0;&#x000b1;&#x000a0;0.35 (6)</td><td align="char">5.64&#x000a0;&#x000b1;&#x000a0;0.45 (5)</td></tr><tr><td>Nicotine</td><td align="char">6.01&#x000a0;&#x000b1;&#x000a0;0.61 (3)</td><td align="char">6.56&#x000a0;&#x000b1;&#x000a0;0.12 (3)</td></tr><tr><td>Strychnine</td><td align="char">4.30&#x000a0;&#x000b1;&#x000a0;0.09 (3)</td><td align="char">4.85&#x000a0;&#x000b1;&#x000a0;0.19 (3)</td></tr></tbody></table></table-wrap></p><p id="p0140">[³H]granisetron association rates were best fit with a monophasic curve. <italic>k</italic>_obs increased with free ligand concentration and a straight line was fitted (<xref rid="fig1" ref-type="fig">Fig.&#x000a0;1</xref>e) to yield the <italic>k</italic>_on and <italic>k</italic>_off values in <xref rid="tbl1" ref-type="table">Table&#x000a0;1</xref>. <italic>K</italic>_d values calculated from these kinetic measurements (Equ <xref rid="fd4" ref-type="disp-formula">(4)</xref>) were in agreement with affinities calculated from our saturation binding studies (<xref rid="tbl1" ref-type="table">Table&#x000a0;1</xref>). Dissociation was also monophasic and the rate agreed well with that from our <italic>k</italic>_obs versus concentration plots described above (<xref rid="fig1" ref-type="fig">Fig.&#x000a0;1</xref>f, <xref rid="tbl1" ref-type="table">Table&#x000a0;1</xref>).</p><p id="p0145">These observations show that using a well-established radiolabelled 5-HT₃ receptor antagonist ([³H]granisetron) we are able to accurately reproduce the binding characteristics reported elsewhere and, similar to [³H]VUF10166, they are consistent with a simple bi-molecular binding scheme.</p></sec><sec id="sec3.5"><label>3.5</label><title>Homology modelling &#x00026; docking</title><p id="p0150">To gain insights into the residues that potentially interact with VUF10166 at the orthosteric site (A+A&#x02212; interface), five 5-HT₃A receptor homology models were generated and <italic>in silico</italic> docking of VUF10166 performed on each one (<xref rid="fig2" ref-type="fig">Fig.&#x000a0;2</xref>). A total of 50 docked poses were generated and for each of these the amino acids within 5&#x000a0;&#x000c5; of VUF10166 were identified (<xref rid="tbl4" ref-type="table">Table&#x000a0;4</xref>). 26% of residues were common to all models, comparable to a previous docking study with granisetron, where 31% of residues were common to all of the predicted binding orientations (<xref rid="bib26" ref-type="bibr">Thompson et&#x000a0;al., 2005</xref>). A selection of these residues were chosen for mutagenesis based upon the following criteria, 1) side chains accessible to the ligand, 2) residues known to interact with other 5-HT₃ ligands or, 3) residues present in a limited number of docked poses to provide support for specific orientations. Of the 39 amino acids identified, 23 were mutated to cysteine (<xref rid="fig3" ref-type="fig">Fig.&#x000a0;3</xref>); cysteine substitution of these residues was chosen as all of the Cys mutants have been previously shown to express on the cell-surface, and the residue positions have been similarly used for the study of our radioligand standard, [³H]granisetron (<xref rid="bib26" ref-type="bibr">Thompson et&#x000a0;al., 2005</xref>, <xref rid="bib25" ref-type="bibr">Thompson et&#x000a0;al., 2011</xref>).<fig id="fig2"><label>Fig.&#x000a0;2</label><caption><p>Predicted binding clusters for VUF10166 docked into five different homology models of the 5-HT₃ receptor A+A&#x02212; binding site. All 10 predicted ligand poses are shown for each model. The 5-HT₃ receptor residues within 5&#x000a0;&#x000c5; of VUF10166 in each of the docked poses are shown in <xref rid="tbl4" ref-type="table">Table&#x000a0;4</xref>.</p></caption><graphic xlink:href="gr2"/></fig><table-wrap position="float" id="tbl4"><label>Table&#x000a0;4</label><caption><p>Residues within 5&#x000a0;&#x000c5; of docked VUF10166 in 5 different homology models of the 5-HT₃A receptor binding site.</p></caption><graphic xlink:href="fx1"/></table-wrap><fig id="fig3"><label>Fig.&#x000a0;3</label><caption><p>An amino acid sequence alignment showing the positions of residues mutated in this study (white text, grey boxes). The six recognised binding loops are indicated by black lines above the text. Positions of &#x003b2;-sheets are shown by grey lines beneath the text. Numbering of residues and structural features are taken from the AChBP protein crystal structure (<xref rid="bib35" ref-type="bibr">Celie et&#x000a0;al., 2004</xref>). The proteins are the human 5-HT3A subunit (P46098) and <italic>Lymnaea stagnalis</italic> AChBP (58154).</p></caption><graphic xlink:href="gr3"/></fig></p></sec><sec id="sec3.6"><label>3.6</label><title>Effects of mutations</title><p id="p0155">The binding affinity of [³H]VUF10166 at each of the mutant receptors is shown in <xref rid="tbl5" ref-type="table">Table&#x000a0;5</xref>, and their locations in <xref rid="fig4" ref-type="fig">Fig.&#x000a0;4</xref>. Changing 3 of the 23 residues resulted in no significant change in affinity, suggesting these residues do not play a role in ligand binding (I71, K112, S114). For the remaining 20 mutants there were differences in the binding affinities when compared to wild type receptors, indicating that these residues may have a role in VUF10166 binding. For 9 of these residues [³H]VUF10166 had reduced affinities (R92, L126, N128, I139, R145, Q151, Y153, H185, F226) and for 11 no saturable binding (<italic>K</italic>_d&#x000a0;&#x0003e;&#x000a0;10&#x000a0;nM) was detected (R58, W90, E129, Y141, Y143, T179, T181, W183, H185, D189, Y234, E236). All these mutant receptors have been previously shown to express in oocytes (<xref rid="bib27" ref-type="bibr">Thompson et&#x000a0;al., 2012</xref>).<table-wrap position="float" id="tbl5"><label>Table&#x000a0;5</label><caption><p>Saturation binding of [³H]VUF10166 at 5-HT₃A receptor mutants.</p></caption><graphic xlink:href="fx2"/></table-wrap><fig id="fig4"><label>Fig.&#x000a0;4</label><caption><p>Binding site residues mutated in this study, colour coded according to the change in <italic>K</italic>_d. A large number of residues that abolish binding (affinity&#x000a0;&#x0003e;&#x000a0;10&#x000a0;nM) are clustered around loop B, with further significant changes in loop D (W90) and loop E (Y141 &#x00026; Y143). Not all changes are likely to result from ligand interactions, such as the effect of D189C which is consistent with it maintaining the hydrogen bond network in the tight loop at the C-terminal end of loop B. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)</p></caption><graphic xlink:href="gr4"/></fig></p><p id="p0160">These data show that [³H]VUF10166 binds to the orthosteric site and are consistent with our findings that [³H]VUF10166 competes with other 5-HT₃ receptor competitive ligands.</p></sec><sec id="sec3.7"><label>3.7</label><title>VUF10166 binding at 5-HT₃AB receptors</title><p id="p0165">VUF10166 was previously shown to discriminate between 5-HT₃ receptors subtypes (<xref rid="bib27" ref-type="bibr">Thompson et&#x000a0;al., 2012</xref>) and so binding properties of the new radioligand were also tested at 5-HT₃AB receptors. [³H]VUF10166 showed high affinity binding at 5-HT₃AB receptors, but unlike at 5-HT₃A receptors, it was complex and could not be fit with a single site model (<xref rid="fig5" ref-type="fig">Fig.&#x000a0;5</xref>a). Dissociation of [³H]VUF10166 at these receptors was best fit with a double exponential curve, which contained both a fast and a slow component; the latter was not significantly different (<italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.05) to the single rate measured at 5-HT₃A receptors (<xref rid="fig5" ref-type="fig">Fig.&#x000a0;5</xref>b, <xref rid="tbl1" ref-type="table">Table&#x000a0;1</xref>). Association curves were monophasic (<xref rid="fig5" ref-type="fig">Fig.&#x000a0;5</xref>c), but when <italic>k</italic>_obs was plotted against radioligand concentration, the data were also best approximated by a two site fit (<xref rid="fig5" ref-type="fig">Fig.&#x000a0;5</xref>d, <xref rid="tbl1" ref-type="table">Table&#x000a0;1</xref>). At concentrations of [³H]VUF10166&#x000a0;&#x0003c;&#x000a0;3&#x000a0;nM the <italic>k</italic>_off and <italic>k</italic>_on values were similar to 5-HT₃A receptors; below 3&#x000a0;nM, average <italic>k</italic>_on values determined from <italic>k</italic>_obs (Equ <xref rid="fd5" ref-type="disp-formula">(5)</xref>) were also similar to 5-HT₃A receptors (8.77&#x000a0;&#x000d7;&#x000a0;10⁷&#x000a0;M&#x000a0;min^&#x02212;1). At concentrations &#x0003e;3&#x000a0;nM, <italic>k</italic>_off and <italic>k</italic>_on had slower rates that yielded a <italic>K</italic>_d (22.4&#x000a0;nM; Equ <xref rid="fd4" ref-type="disp-formula">(4)</xref>) close to the value from competition binding (36.7&#x000a0;nM; <xref rid="tbl1" ref-type="table">Table&#x000a0;1</xref>). Competition binding with a range of ligands was performed using 0.6&#x000a0;nM [³H]VUF10166 and <italic>K</italic>_i values were similar to values at 5-HT₃A receptors (<xref rid="tbl2" ref-type="table">Table&#x000a0;2</xref>).<fig id="fig5"><label>Fig.&#x000a0;5</label><caption><p>Radioligand binding at 5-HT₃AB receptors. (<bold>a</bold>) Binding at 5-HT₃AB receptors could not be well fit with a standard one site model; deviation occurs at a radioligand concentration of &#x0223c;3&#x000a0;nM (arrow). <italic>Inset</italic> competition binding of unlabelled VUF10166 with [³H]granisetron. (<bold>b</bold>) Dissociation was best fit with a double exponential at 5-HT₃AB receptors (0.010&#x000a0;&#x000b1;&#x000a0;0.003&#x000a0;min^&#x02212;1 and 0.227&#x000a0;&#x000b1;&#x000a0;0.056&#x000a0;min^&#x02212;1, <italic>n</italic>&#x000a0;=&#x000a0;8). (<bold>c</bold>) Association was mono-exponential, but a plot of <italic>k</italic>_obs against radioligand concentration. (<bold>d</bold>) revealed two components, showing that it was rate-limited at higher concentrations. (<bold>e</bold>) The association of [³H]granisetron was best fit with a mono-exponential function, but unlike [³H]VUF10166, the fit of <italic>k</italic>_obs against the radioligand concentration was linear at across all concentrations, yielding the values for <italic>k</italic>_on and <italic>k</italic>_off in <xref rid="tbl1" ref-type="table">Table&#x000a0;1</xref>. (<bold>f</bold>) Consistent with this plot, dissociation of [³H]granisetron was also best described by a single exponential function (<italic>k</italic>_off&#x000a0;=&#x000a0;(0.012&#x000a0;&#x000b1;&#x000a0;0.002&#x000a0;min^&#x02212;1, <italic>n</italic>&#x000a0;=&#x000a0;5)) that was not significantly different to 5-HT₃A receptors (<italic>p</italic>&#x000a0;&#x0003e;&#x000a0;0.05, Student's <italic>t</italic>-test).</p></caption><graphic xlink:href="gr5"/></fig></p><p id="p0170">These results show that [³H]VUF10166 has different binding properties at 5-HT₃A and 5-HT₃AB receptors. In the latter effects are complex and some only become apparent at higher concentrations of [³H]VUF10166.</p></sec><sec id="sec3.8"><label>3.8</label><title>Granisetron binding at 5-HT₃AB receptors</title><p id="p0175">Unlike [³H]VUF10166, [³H]granisetron saturation binding at 5-HT₃AB receptors yielded <italic>K</italic>_d values that were the same as those at 5-HT₃A receptors, as reported elsewhere (<xref rid="tbl1" ref-type="table">Table&#x000a0;1</xref>) (<xref rid="bib3" ref-type="bibr">Brady et&#x000a0;al., 2001</xref>). Association (<xref rid="fig5" ref-type="fig">Fig.&#x000a0;5</xref>e), dissociation (<xref rid="fig5" ref-type="fig">Fig.&#x000a0;5</xref>f) and <italic>K</italic>_i values from competition binding (<xref rid="tbl3" ref-type="table">Table&#x000a0;3</xref>) were also the same as those at 5-HT₃A receptors.</p><p id="p0180">These results show that the binding properties of [³H]granisetron are the same at 5-HT₃A and 5-HT₃AB receptors unlike those of [³H]VUF10166.</p></sec></sec><sec id="sec4"><label>4</label><title>Discussion</title><p id="p0185">[³H]VUF10166 binds specifically and with high affinity to 5-HT₃A and 5-HT₃AB receptors, with evidence of a second, lower affinity, binding site in 5-HT₃AB receptors. The effects of this second site are apparent at concentrations of [³H]VUF10166&#x000a0;&#x0003e;&#x000a0;3&#x000a0;nM, and are consistent with previous work that identified an additional allosteric binding site for unlabelled VUF10166 at the A+B&#x02212; interface (<xref rid="bib27" ref-type="bibr">Thompson et&#x000a0;al., 2012</xref>). Docking of this competitive ligand into the orthosteric (A+A&#x02212;) binding site, combined with data from mutagenesis, suggest that VUF10166 is oriented with its quinoxaline rings close to W183 and its basic nitrogen extended towards loop E. Individual residues, many of which have been previously shown to be important in studies of other 5-HT₃ receptor ligands (including <italic>d</italic>-tubocurarine, granisetron, lerisetron, <italic>meta</italic>-chlorophenylbiguanide and tropisetron) are also important for VUF10166 binding (<xref rid="bib8" ref-type="bibr">Hope et&#x000a0;al., 1999</xref>, <xref rid="bib11" ref-type="bibr">Mochizuki et&#x000a0;al., 1999</xref>, <xref rid="bib29" ref-type="bibr">Venkataraman et&#x000a0;al., 2002a</xref>, <xref rid="bib15" ref-type="bibr">Price and Lummis, 2004</xref>). The residues are discussed in more detail below.</p><sec id="sec4.1"><label>4.1</label><title>The role of loop A residues</title><p id="p0190">VUF10166 binding was abolished by Cys substitution of E129, slightly modified by L126C (&#x0223c;4 fold change in <italic>K</italic>_d) and not altered by N128C. E129 was previously identified as an important 5-HT₃ receptor binding residue and may form a hydrogen bond with bound ligand, which is consistent with our data (<xref rid="bib14" ref-type="bibr">Price et&#x000a0;al., 2008</xref>). However, data from 5HTBP (a modified AChBP with high affinity binding for 5-HT₃ receptor ligands) suggest that E129 may hydrogen bond with the side chain of T179 (<xref rid="bib9" ref-type="bibr">Kesters et&#x000a0;al., 2013</xref>), and therefore might have primarily a structural role. L126 may also have a structural role but is less important as the effects of altering this residue were small, while N128 has been shown to play a role in gating but not binding (<xref rid="bib14" ref-type="bibr">Price et&#x000a0;al., 2008</xref>, <xref rid="bib9" ref-type="bibr">Kesters et&#x000a0;al., 2013</xref>).</p></sec><sec id="sec4.2"><label>4.2</label><title>The role of loop B residues</title><p id="p0195">Loop B has been previously identified as both a critical structural component of the binding pocket, and it contributes to ligand binding. W183 is especially important as a constituent of the &#x02018;aromatic box&#x02019; that exists in all Cys-loop receptor binding sites (<xref rid="bib1" ref-type="bibr">Beene et&#x000a0;al., 2002</xref>, <xref rid="bib22" ref-type="bibr">Thompson et&#x000a0;al., 2008b</xref>, <xref rid="bib5" ref-type="bibr">Duffy et&#x000a0;al., 2012</xref>). Other residues (T179, H185, D189) are known to stabilise the binding site structure via hydrogen bonds (<xref rid="bib22" ref-type="bibr">Thompson et&#x000a0;al., 2008b</xref>, <xref rid="bib9" ref-type="bibr">Kesters et&#x000a0;al., 2013</xref>). It is therefore not surprising that all our loop B mutations altered or abolished [³H]VUF10166 binding and we suggest that T181 and W183 interact with VUF10166 while T179, H185 and D189 have a structural role.</p></sec><sec id="sec4.3"><label>4.3</label><title>The role of loop C residues</title><p id="p0200">F226 and Y234 are also constituents of the aromatic box and mutations here alter or eliminate VUF10166 binding. F226A has no effect on granisetron binding affinity, indicating this residue is more important for VUF10166 binding (<xref rid="bib26" ref-type="bibr">Thompson et&#x000a0;al., 2005</xref>). In 5HTBP Y234 (Y193) interacts with 5-HT and also contributes to a conserved water network that stabilises the granisetron-bound structure (<xref rid="bib9" ref-type="bibr">Kesters et&#x000a0;al., 2013</xref>); a conserved water network is also seen at this location in many AChBP crystal structures and may be important in many Cys-loop receptors. E236C also abolished VUF10166 binding, consistent with studies where substitutions affect binding of both GR65630 and granisetron, as well as altering the maximal current and <italic>EC</italic>₅₀ of 5-HT responses (<xref rid="bib17" ref-type="bibr">Schreiter et&#x000a0;al., 2003</xref>, <xref rid="bib13" ref-type="bibr">Nyce et&#x000a0;al., 2010</xref>). However E236 mutations may adversely affect the correct assembly of the binding site rather than interfering with specific ligand interactions as <xref rid="bib13" ref-type="bibr">Nyce et&#x000a0;al. (2010)</xref> and <xref rid="bib17" ref-type="bibr">Schreiter et&#x000a0;al. (2003)</xref> showed that some E236 mutant receptors are trapped within the cell. As this hypothesis is supported by the lack of interactions in the 5HTBP structure, we consider it unlikely that E236 contributes to VUF10166 binding (<xref rid="bib9" ref-type="bibr">Kesters et&#x000a0;al., 2013</xref>).</p></sec><sec id="sec4.4"><label>4.4</label><title>The role of loop D residues</title><p id="p0205">W90 is another aromatic box residue that contributes to binding. In 5HTBP the equivalent residue (W53) is involved in van der Waals interactions with granisetron and W90 may have a similar role in binding VUF10166 (<xref rid="bib18" ref-type="bibr">Spier and Lummis, 2000</xref>, <xref rid="bib15" ref-type="bibr">Price and Lummis, 2004</xref>, <xref rid="bib26" ref-type="bibr">Thompson et&#x000a0;al., 2005</xref>, <xref rid="bib34" ref-type="bibr">Yan and White, 2005</xref>). Substitutions at W90 decrease the affinity of other potent 5-HT₃ receptor-specific ligands such as curare, lerisetron and 5-HT (<xref rid="bib33" ref-type="bibr">Yan et&#x000a0;al., 1999</xref>, <xref rid="bib29" ref-type="bibr">Venkataraman et&#x000a0;al., 2002a</xref>) R92 interacts with granisetron in 5HTBP (R55), and the effects of its substitution on the affinity of VUF10166, ondansetron, granisetron and MDL72222, suggest an interaction with all of these ligands (<xref rid="bib26" ref-type="bibr">Thompson et&#x000a0;al., 2005</xref>, <xref rid="bib34" ref-type="bibr">Yan and White, 2005</xref>).</p></sec><sec id="sec4.5"><label>4.5</label><title>The role of loop E residues</title><p id="p0210">All of the mutations in loop E (Y141, Y143, R145, Q151, Y153) caused significant changes to [³H]VUF10166 binding. In the 5HTBP crystal structure granisetron does not extend towards loop E, but instead lies horizontally between loops B and D, similar to the orientations of the closely related ligands tropisetron (2WNC) and cocaine (2PGZ) in AChBP. In contrast, in 5HTBP 5-HT hydrogen bonds with the backbone carbonyls of I104 (Y141 in 5-HT₃) and I116 (Y153), and has hydrophobic interactions with M114 (Q151), explaining why 5-HT activation is strongly affected by mutations at these locations, but effects on granisetron are less apparent (<xref rid="bib30" ref-type="bibr">Venkataraman et&#x000a0;al., 2002b</xref>, <xref rid="bib15" ref-type="bibr">Price and Lummis, 2004</xref>, <xref rid="bib25" ref-type="bibr">Thompson et&#x000a0;al., 2011</xref>, <xref rid="bib9" ref-type="bibr">Kesters et&#x000a0;al., 2013</xref>). Here the affinity of VUF10166 was decreased 10-fold by Y153C and abolished by Y143C, indicating that bound VUF10166 extends towards, and may interact with, loop E residues. As VUF10166 is also a low efficacy partial agonist at &#x003bc;M concentrations, and must therefore induce the same structural changes as 5-HT, it is likely it adopts an orientation that at least partially mimics that of 5-HT.</p></sec><sec id="sec4.6"><label>4.6</label><title>The orientation of VUF10166 in the ligand binding pocket</title><p id="p0215">Our results show that VUF10166 binding is affected by many of the residues previously identified as important for binding 5-HT₃ receptor antagonists, while mutation of R58, I71, K112 and S114, which are close to VUF10166 in models 1, 2 and 5, did not alter its affinity, suggesting that these models are less probable. Also in model 1 the predicted ligand orientations do not extend towards Loop E and yet residues here were important for VUF10166 binding. Similarly R145 is within 5&#x000a0;&#x000c5; of VUF10166 in model 2, but our mutagenesis data show that altering this residue has little effect on binding affinity. Model 3 seems unlikely as these poses are positioned closer to the complimentary face of the binding site, and do not significantly interact with key principal face residues such as T181, W183 and Y234. Models 4 and 5 have quite similar docked poses with only F226 distinguishing them; F226C mutant receptors had a 7-fold lower affinity than wild type receptors suggesting that this residue is close enough to interfere with VUF10166 binding, which would best fit with model 4.</p><p id="p0220">In previous work we presented a structure-activity study (SAR) of VUF10166 analogues (<xref rid="bib31" ref-type="bibr">Verheij et&#x000a0;al., 2012</xref>, <xref rid="bib28" ref-type="bibr">Thompson et&#x000a0;al., 2013</xref>) and the active analogues from these studies would fit well into model 4 in two distinct orientations (<xref rid="fig6" ref-type="fig">Fig.&#x000a0;6</xref>). These data showed substitutions of the chlorine atom in VUF10166 (<xref rid="fig6" ref-type="fig">Fig.&#x000a0;6</xref>a, region 2) are poorly tolerated, suggesting an important interaction at this location; in both poses in <xref rid="fig6" ref-type="fig">Fig.&#x000a0;6</xref> the chlorine atom is closely located to R92 and W90. In contrast, substitutions in regions 1 and 3 are fairly well tolerated, providing that they are not too large; neither of the poses in <xref rid="fig6" ref-type="fig">Fig.&#x000a0;6</xref> are sterically restricted around these regions of VUF10166. The poses also explain the importance of the charged <italic>N</italic>-methylpiperazine nitrogen atom, as there are possible cation&#x02013;&#x003c0; interactions with W183 and Y234 in one pose, with these residues contributing to &#x003c0;&#x02013;&#x003c0; stacking of the quinoxaline ring in the other.<fig id="fig6"><label>Fig.&#x000a0;6</label><caption><p>Chemical structure of VUF10166 and its binding mode. (<bold>a</bold>) Three regions of the ligand are identified and are described in the text. Its protonation site, which is also its tritiation site, is indicated. (<bold>b</bold>&#x02013;<bold>c</bold>) The volume occupied by the two main docked pose clusters in model 4. In (c) cation&#x02013;&#x003c0; interactions are possible with W183 (5.06&#x000a0;&#x000c5; away) and Y234 (4.46&#x000a0;&#x000c5;). VUF10166 is shown as a stick and wire mesh representation (white), with the residues mutated in this study colour coded similar to <xref rid="fig4" ref-type="fig">Fig.&#x000a0;4</xref>. (<bold>d</bold>&#x02013;<bold>e</bold>) Cartoons showing our interpretation of the binding to heteromeric receptors. Below 3&#x000a0;nM, VUF10166 binds to a single population of binding sites at the A+A&#x02212; interface of both 5-HT₃A and 5-HT₃AB receptors; consequently, at these concentrations both receptors share common values for <italic>k</italic>_on and <italic>k</italic>_off. At concentrations of VUF10166&#x000a0;&#x0003e;&#x000a0;3&#x000a0;nM, binding also occurs at a second A+B&#x02212; binding site and allosterically influences the adjacent A+A&#x02212; site; therefore, additional rates are apparent and saturation binding is confounded by rates associated with multiple binding sites and allosteric interactions. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)</p></caption><graphic xlink:href="gr6"/></fig></p><p id="p0225">We therefore suggest that the docked poses in model 4 are most consistent with the mutagenesis data described here and our previously published SAR. It is difficult to predict whether the <italic>N</italic>-methylpiperazine ring or the quinoxaline ring is positioned toward loop E, but the orientation in <xref rid="fig6" ref-type="fig">Fig.&#x000a0;6</xref>c is most reminiscent of varenicline co-crystallised into AChBP (PDBID&#x000a0;=&#x000a0;4AFG &#x00026; 4AFT) and 5-HT in 5HTBP (2YMD), both of which are agonists at 5-HT₃ receptors (<xref rid="bib2" ref-type="bibr">Billen et&#x000a0;al., 2012</xref>, <xref rid="bib16" ref-type="bibr">Rucktooa et&#x000a0;al., 2012</xref>). This similarity in orientation may explain why VUF10166 also displays partial agonist activity (<xref rid="bib27" ref-type="bibr">Thompson et&#x000a0;al., 2012</xref>). However, it should be stressed that we must exercise caution when making these predictions as the physiological relevance of these structures have not yet been fully ascertained, for example three ligand molecules have been observed in a single AChBP binding site, something we would not have predicted (<xref rid="bib4" ref-type="bibr">Brams et&#x000a0;al., 2011</xref>, <xref rid="bib19" ref-type="bibr">Stornaiuolo et&#x000a0;al., 2013</xref>).</p></sec></sec><sec id="sec5"><label>5</label><title>Conclusion</title><p id="p0230">Our results show that VUF10166 interacts with several of the core binding site residues found at the A+A&#x02212; interface and, combined with homology modelling and ligand docking, we propose it adopts an orientation similar to that of other 5-HT₃ receptor agonists in AChBP and 5HTBP crystal structures. At 5-HT₃ receptors our kinetic measurements are consistent with a single A+A&#x02212; binding site, but at 5-HT₃AB an additional fast component is seen. This is consistent with the lower affinity of VUF10166 for the 5-HT₃AB receptor and is likely to result from an allosteric effect that is evident when the concentration of VUF10166 exceeds 3&#x000a0;nM (as summarised in <xref rid="fig6" ref-type="fig">Fig.&#x000a0;6</xref>).</p></sec>

Real-time use of instantaneous wave&#x02013;free ratio: Results of the ADVISE <italic>in-practice</italic>: An international, multicenter evaluation of instantaneous wave&#x02013;free ratio in clinical practice

Could not extract abstract

<contrib contrib-type="author" id="au0005"><name><surname>Petraco</surname><given-names>Ricardo</given-names></name><degrees>MD</degrees><xref rid="af0005" ref-type="aff">a</xref></contrib><contrib contrib-type="author" id="au0010"><name><surname>Al-Lamee</surname><given-names>Rasha</given-names></name><degrees>MBBS</degrees><xref rid="af0005" ref-type="aff">a</xref></contrib><contrib contrib-type="author" id="au0015"><name><surname>Gotberg</surname><given-names>Matthias</given-names></name><degrees>MD, PhD</degrees><xref rid="af0010" ref-type="aff">b</xref></contrib><contrib contrib-type="author" id="au0020"><name><surname>Sharp</surname><given-names>Andrew</given-names></name><degrees>MBBS, PhD</degrees><xref rid="af0015" ref-type="aff">c</xref></contrib><contrib contrib-type="author" id="au0025"><name><surname>Hellig</surname><given-names>Farrel</given-names></name><degrees>MD</degrees><xref rid="af0020" ref-type="aff">d</xref></contrib><contrib contrib-type="author" id="au0030"><name><surname>Nijjer</surname><given-names>Sukhjinder S.</given-names></name><degrees>MBChB</degrees><xref rid="af0005" ref-type="aff">a</xref></contrib><contrib contrib-type="author" id="au0035"><name><surname>Echavarria-Pinto</surname><given-names>Mauro</given-names></name><degrees>MD</degrees><xref rid="af0025" ref-type="aff">e</xref></contrib><contrib contrib-type="author" id="au0040"><name><surname>van de Hoef</surname><given-names>Tim P.</given-names></name><degrees>MD</degrees><xref rid="af0030" ref-type="aff">f</xref></contrib><contrib contrib-type="author" id="au0045"><name><surname>Sen</surname><given-names>Sayan</given-names></name><degrees>MBBS PhD</degrees><xref rid="af0005" ref-type="aff">a</xref></contrib><contrib contrib-type="author" id="au0050"><name><surname>Tanaka</surname><given-names>Nobuhiro</given-names></name><degrees>MD PhD</degrees><xref rid="af0035" ref-type="aff">g</xref></contrib><contrib contrib-type="author" id="au0055"><name><surname>Van Belle</surname><given-names>Eric</given-names></name><degrees>MD, PhD</degrees><xref rid="af0040" ref-type="aff">h</xref></contrib><contrib contrib-type="author" id="au0060"><name><surname>Bojara</surname><given-names>Waldemar</given-names></name><degrees>MD</degrees><xref rid="af0045" ref-type="aff">i</xref></contrib><contrib contrib-type="author" id="au0065"><name><surname>Sakoda</surname><given-names>Kunihiro</given-names></name><degrees>MD</degrees><xref rid="af0050" ref-type="aff">j</xref></contrib><contrib contrib-type="author" id="au0070"><name><surname>Mates</surname><given-names>Martin</given-names></name><degrees>MD, PhD</degrees><xref rid="af0055" ref-type="aff">k</xref></contrib><contrib contrib-type="author" id="au0075"><name><surname>Indolfi</surname><given-names>Ciro</given-names></name><degrees>MD</degrees><xref rid="af0060" ref-type="aff">l</xref></contrib><contrib contrib-type="author" id="au0080"><name><surname>De Rosa</surname><given-names>Salvatore</given-names></name><degrees>MD, PhD</degrees><xref rid="af0060" ref-type="aff">l</xref></contrib><contrib contrib-type="author" id="au0085"><name><surname>Vrints</surname><given-names>Christian J.</given-names></name><degrees>MD, PhD</degrees><xref rid="af0065" ref-type="aff">m</xref></contrib><contrib contrib-type="author" id="au0090"><name><surname>Haine</surname><given-names>Steven</given-names></name><degrees>MD</degrees><xref rid="af0065" ref-type="aff">m</xref></contrib><contrib contrib-type="author" id="au0095"><name><surname>Yokoi</surname><given-names>Hiroyoshi</given-names></name><degrees>MD</degrees><xref rid="af0070" ref-type="aff">n</xref></contrib><contrib contrib-type="author" id="au0100"><name><surname>Ribichini</surname><given-names>Flavio L.</given-names></name><degrees>MD</degrees><xref rid="af0075" ref-type="aff">o</xref></contrib><contrib contrib-type="author" id="au0105"><name><surname>Meuwissen</surname><given-names>Martjin</given-names></name><degrees>MD, PhD</degrees><xref rid="af0080" ref-type="aff">p</xref></contrib><contrib contrib-type="author" id="au0110"><name><surname>Matsuo</surname><given-names>Hitoshi</given-names></name><degrees>MD, PhD</degrees><xref rid="af0085" ref-type="aff">q</xref></contrib><contrib contrib-type="author" id="au0115"><name><surname>Janssens</surname><given-names>Luc</given-names></name><degrees>MD</degrees><xref rid="af0090" ref-type="aff">r</xref></contrib><contrib contrib-type="author" id="au0120"><name><surname>Katsumi</surname><given-names>Ueno</given-names></name><degrees>MD</degrees><xref rid="af0085" ref-type="aff">q</xref></contrib><contrib contrib-type="author" id="au0125"><name><surname>Di Mario</surname><given-names>Carlo</given-names></name><degrees>MD, PhD</degrees><xref rid="af0095" ref-type="aff">s</xref></contrib><contrib contrib-type="author" id="au0130"><name><surname>Escaned</surname><given-names>Javier</given-names></name><degrees>MD, PhD</degrees><xref rid="af0025" ref-type="aff">e</xref></contrib><contrib contrib-type="author" id="au0135"><name><surname>Piek</surname><given-names>Jan</given-names></name><degrees>MD, PhD</degrees><xref rid="af0030" ref-type="aff">f</xref></contrib><contrib contrib-type="author" id="au0140"><name><surname>Davies</surname><given-names>Justin E.</given-names></name><degrees>MBBS, PhD</degrees><email>justin.davies@imperial.ac.uk</email><xref rid="af0005" ref-type="aff">a</xref><xref rid="cr0005" ref-type="corresp">&#x0204e;</xref></contrib><aff id="af0005"><label>a</label>International Centre for Circulatory Health, National Heart and Lung Institute, Imperial College London and Imperial College Healthcare NHS Trust, London, United Kingdom</aff><aff id="af0010"><label>b</label>Lund University, Lund, Sweden</aff><aff id="af0015"><label>c</label>Royal Devon and Exeter Hospital, Exeter, United Kingdom</aff><aff id="af0020"><label>d</label>Sunninghill &#x00026; Sunward Park Hospitals, Johannesburg, South Africa</aff><aff id="af0025"><label>e</label>Cardiovascular Institute, Hospital Clinico San Carlos, Madrid, Spain</aff><aff id="af0030"><label>f</label>Academic Medical Centre, Amsterdam, the Netherlands</aff><aff id="af0035"><label>g</label>Tokyo University Hospital, Tokyo, Japan</aff><aff id="af0040"><label>h</label>H&#x000f4;pital cardiologique CHRU Lille, Lille, France</aff><aff id="af0045"><label>i</label>Koblenz-Mayen Hospital, Koblenz, Germany</aff><aff id="af0050"><label>j</label>St Luke's international hospital, Tokyo, Japan</aff><aff id="af0055"><label>k</label>Na Homolce Hospital, Prague, Czech Republic</aff><aff id="af0060"><label>l</label>University Magna Graecia, Catanzaro, Italy</aff><aff id="af0065"><label>m</label>Antwerp University Hospital, Antwerp, Belgium</aff><aff id="af0070"><label>n</label>Kokura Memorial Hospital, Kitakyushu, Japan</aff><aff id="af0075"><label>o</label>Universit&#x000e0; di Verona, Verona, Italy</aff><aff id="af0080"><label>p</label>Amphia Hospital, Breda, the Netherlands</aff><aff id="af0085"><label>q</label>Gifu Heart Center, Gifu, Japan</aff><aff id="af0090"><label>r</label>Imelda Hospital Bonheiden, Antwerp, Belgium</aff><aff id="af0095"><label>s</label>Cardiovascular Biomedical Research Unit, Royal Brompton Hospital, London, United Kingdom</aff>

American Heart Journal

<p id="p0005">The instantaneous wave&#x02013;free ratio (iFR) is a recently proposed index of coronary disease severity, which uses a trans-lesional pressure ratio as a measure of functional stenosis severity. It can be calculated using conventional pressure guide wires and differs fundamentally from fractional flow reserve (FFR)<xref rid="bb0005" ref-type="bibr">¹</xref> because it does not require vasodilators such as adenosine for its calculation.</p><p id="p0010">The relationship between iFR and FFR has been extensively evaluated in more than 2,000 stenoses,<xref rid="bb0005" ref-type="bibr">[1]</xref>, <xref rid="bb0010" ref-type="bibr">[2]</xref>, <xref rid="bb0015" ref-type="bibr">[3]</xref>, <xref rid="bb0020" ref-type="bibr">[4]</xref> and their agreement in lesion classification ranges from 80% to 90%, depending on whether the comparison is made in clinical populations, with predominantly intermediate lesions, or in samples with more severe stenoses.<xref rid="bb0010" ref-type="bibr">²</xref> When compared with independent arbiters of stenosis severity, iFR and FFR have demonstrated equal diagnostic agreement against invasive flow<xref rid="bb0025" ref-type="bibr">[5]</xref>, <xref rid="bb0030" ref-type="bibr">[6]</xref> and myocardial perfusion imaging.<xref rid="bb0035" ref-type="bibr">⁷</xref> Also, offline studies consistently demonstrated 0.89 to 0.90 as the optimal iFR cutoff to match an FFR of 0.80.<xref rid="bb0010" ref-type="bibr">[2]</xref>, <xref rid="bb0015" ref-type="bibr">[3]</xref>, <xref rid="bb0040" ref-type="bibr">[8]</xref></p><p id="p0015">To date, however, in all iFR-FFR studies, pressure data and iFR calculation have been processed offline in a core laboratory, after procedural termination and appropriate data extraction. Although necessary for any new technology during its early development, this methodology does not reflect the future application of iFR in clinical practice, which will be performed by clinicians, using appropriate software installed in hemodynamic consoles. It is not known whether these practical aspects could affect the performance of iFR.</p><p id="p0020">In this study we explored important aspects of the relationship between iFR and FFR, when both indices are measured in real time by clinicians. First, using a <italic>clinical</italic> FFR cutoff of 0.80 as a reference, we evaluated whether the diagnostic performance and optimal cutoff of iFR are maintained, when compared with offline studies. Second, we extended the analysis to the original 0.75 <italic>isch</italic>em<italic>ic</italic> FFR cutoff and explored the performance of real-time iFR to match such classification of stenoses. Finally, we aimed to validate with real-time measurements the previously reported iFR-FFR hybrid strategy, using a predefined deferral iFR cutoff of &#x0003e;0.93 and a predefined treatment iFR cutoff of &#x0003c;0.86.<xref rid="bb0045" ref-type="bibr">⁹</xref></p><sec id="s0010"><title>Methods</title><sec id="s0030"><title>Study population</title><p id="p0025">Instantaneous wave&#x02013;free ratio and FFR were measured from hemodynamic consoles in 16 centers in Europe, Asia, and Africa. The study included 392 stenoses from 313 consecutive patients who, as part of clinical investigation, required functional intracoronary assessment with pressure guide wires.</p></sec><sec id="s0035"><title>Hemodynamic data collection and analysis</title><p id="p0030">Cardiac catheterization was performed according to standard practice. Unfractionated intravenous heparin (5000 IU) and 300 to 600 &#x003bc;g intracoronary nitrates were given at the start of the procedure. Acquisition of physiological data for FFR calculation was performed according to conventional practice<xref rid="bb0050" ref-type="bibr">¹⁰</xref> using commercially available FFR systems (S5i and Prestige pressure guide wire; Volcano Corporation, San Diego, California). The method to induce pharamacologic hyperemia varied according to conventional clinical practice at each center. Adenosine (or adenosine triphosphate) was used for induction of hyperemia: in 39% of the cases, it was administered via a central line, with doses ranging from 140 to 180 &#x003bc;g kg^&#x02212;1 min^&#x02212;1 (median 140 &#x003bc;g kg^&#x02212;1 min^&#x02212;1); in 61% of the cases, the intracoronary route was used with a median dose of 60 &#x003bc;g (interquartile range 36-240 &#x003bc;g). Each reported iFR value and its FFR counterpart were obtained directly from the hardware console. Instantaneous wave&#x02013;free ratio was calculated using software embedded onto the hemodynamic consoles, which uses proprietary iFR algorithms acting on electrocardiogram (ECG)-gated, time-aligned pressure traces, as previously described.<xref rid="bb0005" ref-type="bibr">¹</xref> Instantaneous wave&#x02013;free ratio was automatically calculated as a ratio of distal (Pd) to proximal (Pa) coronary pressures at the baseline iFR window. Fractional flow reserve was automatically calculated as the ratio of whole-cycle Pd/Pa during hyperemia. Anatomical severity of coronary stenoses was measured using quantitative coronary angiography.</p><p id="p0035">Prior to iFR and FFR measurements, hemodynamic consoles ensure that the pressure data are appropriately calibrated. These include steps required for both FFR and iFR calculation, such as ensuring that catheter (Pa) and wire (Pd) pressures are equal or <italic>normalized</italic> before wire insertion. In addition, a specific step is needed before iFR measurement, which is the adjustment of temporal delays between Pa and Pd signals. This essential process occurs automatically and simultaneously with pressure normalization (<xref rid="f0010" ref-type="fig">Figure&#x000a0;1</xref>). <xref rid="f0015" ref-type="fig">Figure&#x000a0;2</xref> demonstrates the importance of ECG identification and the reliability of iFR calculation in sinus rhythm and atrial fibrillation.<fig id="f0010"><label>Figure&#x000a0;1</label><caption><p>Pressure normalization, temporal alignment, and iFR calculation using the hemodynamic console.</p></caption><graphic xlink:href="gr1"/></fig><fig id="f0015"><label>Figure&#x000a0;2</label><caption><p>Importance of ECG detection for accurate iFR measurement.</p></caption><graphic xlink:href="gr2"/></fig></p></sec><sec id="s0040"><title>Statistical analysis</title><p id="p0040">The classification agreement (and sensitivity, specificity, negative predictive value, and positive predictive value) between iFR and FFR as well the area under the receiver operating characteristic curves (ROC_AUC) were calculated for 2 different FFR cutoffs. First, a 0.80 <italic>clinical</italic> FFR cutoff (as per FAME, FAME II trials, and current clinical guidelines) was used (FFR or iFR &#x02264;0.8 as a reference test to define significant stenoses). For this clinical FFR cutoff, we established a prespecified iFR cutoff of 0.90 and evaluated whether it would also represent the optimal cutoff by ROC analysis (defined as the best sum of sensitivity and specificity). Second, the diagnostic performance of iFR was evaluated for a 0.75 <italic>ischemic</italic> FFR cutoff, as per original validation studies and DEFER<xref rid="bb0055" ref-type="bibr">[11]</xref>, <xref rid="bb0060" ref-type="bibr">[12]</xref> (FFR or iFR &#x02264;0.75 as a reference test to define significant stenoses). Finally, the diagnostic performance of iFR was explored when allowing for the 0.75 to 0.80 FFR gray zone (zone of values which, according to DEFER and FAME/FAME II studies, is <italic>both</italic> safe to defer and treat stenoses). Comparisons between proportions (classification agreement between iFR and FFR) observed in this study with previous offline data sets were made with the &#x003c7;² test for homogeneity of proportions. Areas under the receiver operating characteristic curve were compared using the method described by DeLong et al<xref rid="bb0065" ref-type="bibr">¹³</xref> in STATA, version 11 (StataCorp, College Station, Texas). Data are expressed as mean &#x000b1; SD. A <italic>P</italic> value of &#x0003c;.05 was considered statistically significant.</p></sec><sec id="s0045"><title>Hybrid iFR-FFR analysis</title><p id="p0045">The real-time iFR and FFR values reported in this sample were evaluated in a hybrid decision-making strategy, according to the methodology previously described.<xref rid="bb0045" ref-type="bibr">⁹</xref> We aimed to prospectively apply the following decision-making arms into this sample: when iFR &#x0003e;0.93, stenoses would be deferred; when iFR &#x0003c;0.86, stenoses would be treated; when iFR would fall between 0.86 and 0.93, FFR would be calculated. For such strategy, we evaluated the overall agreement in stenoses classification with an FFR-only strategy and the proportion of stenoses spared from vasodilator administration.</p></sec><sec id="s0050"><title>Funding sources</title><p id="p0050">Dr Petraco (FS/11/46/28861) and Dr. J.E. Davies (FS/05/006) are British Heart Foundation fellows. Dr Sen (G1000357) and Dr Nijjer (G1100443) are Medical Research Council fellows. Dr Di Mario is a senior National Institute for Health Research investigator. No other extramural funding was used to support this work. The authors are solely responsible for the design and conduct of this study, all study analyses, the drafting and editing of the manuscript, and its final contents.</p></sec></sec><sec id="s0015"><title>Results</title><sec id="s0055"><title>Sample characteristics</title><p id="p0060">The 392 stenoses (from 313 patients) included in this study demonstrated a unimodal distribution of FFR and diameter stenosis values. Population demographic and angiographic data are presented in <xref rid="t0005" ref-type="table">Table</xref>. Fractional flow reserve and iFR were measured in all attempted cases. Mean FFR was 0.82 &#x000b1; 0.10, and mean iFR was 0.89 &#x000b1; 0.11. This sample was formed predominantly by physiologically intermediate lesions, with 71% of stenoses falling between FFR 0.70 and 0.90. The proportion of stenoses with FFR values lower than 0.80, 0.75, 0.60, and 0.50, respectively, was 39%, 18%, 4.6%, and 1.3%. Mean percentage diameter stenosis was 56% &#x000b1; 13%, with 84% of lesions falling within the 40% to 80% group. <xref rid="f0020" ref-type="fig">Figure&#x000a0;3</xref> presents a histogram of FFR and diameter stenosis values.<table-wrap position="float" id="t0005"><label>Table</label><caption><p>Demographic and angiographic data</p></caption><table frame="hsides" rules="groups"><tbody><tr><td align="left">No. of stenoses (patients)</td><td align="center">392 (313)</td></tr><tr><td align="left">Age (y), mean &#x000b1; SD</td><td align="center">67 &#x000b1; 11</td></tr><tr><td align="left">Male %</td><td align="center">79</td></tr><tr><td align="left">Comorbidities (%)</td><td/></tr><tr><td align="left">&#x000a0;Hypertension</td><td align="center">74</td></tr><tr><td align="left">&#x000a0;Hypercholesterolemia</td><td align="center">67</td></tr><tr><td align="left">&#x000a0;Smoking history</td><td align="center">51</td></tr><tr><td align="left">&#x000a0;Diabetes</td><td align="center">30</td></tr><tr><td align="left">&#x000a0;Ejection fraction, mean &#x000b1; SD</td><td align="center">58 &#x000b1; 12</td></tr><tr><td align="left">Clinical presentation (%)</td><td/></tr><tr><td align="left">&#x000a0;Stable angina</td><td align="center">73</td></tr><tr><td align="left">&#x000a0;Unstable angina (nonculprit vessel)</td><td align="center">27</td></tr><tr><td align="left">Coronary anatomy (%)</td><td/></tr><tr><td align="left">&#x000a0;Single-vessel CAD</td><td align="center">36</td></tr><tr><td align="left">&#x000a0;Multivessel CAD</td><td align="center">63</td></tr><tr><td align="left">&#x000a0;LAD</td><td align="center">66</td></tr><tr><td align="left">&#x000a0;LCx</td><td align="center">10</td></tr><tr><td align="left">&#x000a0;RCA</td><td align="center">14</td></tr><tr><td align="left">&#x000a0;Other</td><td align="center">10</td></tr><tr><td align="left">&#x000a0;Diameter stenosis (%), men &#x000b1; SD</td><td align="center">56 &#x000b1; 13</td></tr><tr><td align="left">Adenosine route (%)</td><td/></tr><tr><td align="left">&#x000a0;Intravenous</td><td align="center">39</td></tr><tr><td align="left">&#x000a0;Intracoronary</td><td align="center">61</td></tr></tbody></table><table-wrap-foot><fn><p>Abbreviations: <italic>CAD</italic>, Coronary artery disease; <italic>LAD</italic>, left anterior descending artery; <italic>LCx</italic>, left circumflex artery; <italic>RCA</italic>, right coronary artery.</p></fn></table-wrap-foot></table-wrap><fig id="f0020"><label>Figure&#x000a0;3</label><caption><p>Frequency distribution of FFR and percentage diameter stenosis values in the study.</p></caption><graphic xlink:href="gr3"/></fig></p></sec><sec id="s0060"><title>Classification agreement between iFR and FFR</title><sec id="s0065"><title>Using a clinical FFR cutoff of 0.80</title><p id="p0070">In this sample, the predefined iFR cutoff of 0.90 was also identified as the ROC-derived optimal cutoff, yielding a classification agreement with FFR of 80%. The ROC_AUC for iFR to was 0.87 (CI 0.84-0.91). The diagnostic agreement between iFR and FFR when measured in real time (80%) was not different from previously reported offline studies performed in clinical samples (RESOLVE study<xref rid="bb0040" ref-type="bibr">⁸</xref> 80%, N = 1,593; ADVISE Registry<xref rid="bb0010" ref-type="bibr">²</xref> study 80%, N = 339; and by Park et al<xref rid="bb0015" ref-type="bibr">³</xref> 82%, N = 238; <italic>P</italic> = .95 for all comparisons). The detailed diagnostic performance of iFR is summarized in <xref rid="f0025" ref-type="fig">Figure&#x000a0;4</xref>. For this clinical FFR cutoff, the diagnostic relationship between iFR and FFR remained close within the intermediate anatomical lesion range (iFR-FFR ROC_AUC within 40%-80% diameter stenosis = 0.86 [0.83-0.90]).<fig id="f0025"><label>Figure&#x000a0;4</label><caption><p>Diagnostic agreement between iFR and FFR.</p></caption><graphic xlink:href="gr4"/></fig></p></sec><sec id="s0070"><title>Using a ischemic FFR cutoff of 0.75</title><p id="p0075">For the ischemic FFR cutoff, classification match between iFR and FFR was 88% with an ROC_AUC of 0.90 (CI 0.86-0.94). To match an FFR of 0.75, the ROC-derived optimal iFR cutoff was 0.85 (<xref rid="f0025" ref-type="fig">Figure&#x000a0;4</xref>). For this ischemic FFR cutoff, the diagnostic relationship between iFR and FFR also remained unchanged within the intermediate anatomical lesion range (iFR-FFR ROC_AUC within 40%-80% diameter stenosis = 0.88 [0.85-0.91]).</p><p id="p0080">When the FFR values falling into the 0.75 to 0.80 gray zone were considered equally safe to be classified as <italic>normal or abnormal</italic>, the ROC_AUC increased to 0.93 (CI 0.90-0.96) and the classification agreement between iFR and FFR to 92%. Both ROC_AUC and classification agreement were significantly higher when accounting for the FFR gray zone, when compared with single FFR cutoffs (<italic>P</italic> &#x0003c; .001 for comparisons against FFR 0.75 and FFR 0.80).</p></sec></sec><sec id="s0075"><title>Hybrid iFR-FFR decision-making analysis</title><p id="p0090">The previously reported iFR cutoffs of a hybrid iFR-FFR approach<xref rid="bb0045" ref-type="bibr">⁹</xref> were also validated in this real-time iFR sample. Using a predefined deferral iFR cutoff of &#x0003e;0.93 and a predefined treatment iFR cutoff of &#x0003c;0.86 (and measuring FFR only in those lesions with iFR between 0.86 and 0.93) would generate a 94% overall agreement with an FFR classification of lesions, while sparing 61% of patients from adenosine, proportions that are not different from the previous offline report (Petraco et al,<xref rid="bb0045" ref-type="bibr">⁹</xref> N = 577; <italic>P</italic> = .92). <xref rid="f0030" ref-type="fig">Figure&#x000a0;5</xref> summarizes other possible iFR-FFR diagnostic strategies.<fig id="f0030"><label>Figure&#x000a0;5</label><caption><p>Decision-making strategies of revascularization, using iFR only (bottom panel) and a hybrid iFR-FFR approach (top panel). FFR gray zone (0.75-0.80) refers to a region within which is known to be safe to defer <italic>and</italic> treat stenoses with equivalent clinical outcomes.</p></caption><graphic xlink:href="gr5"/></fig></p></sec></sec><sec id="s0020"><title>Discussion</title><p id="p0095">In this study, we present the results of the first clinical application of real-time iFR measurements in patients undergoing invasive functional assessment of intermediate coronary stenoses. We found that (1) for a clinical FFR cutoff of 0.80, a predefined iFR value of 0.90 provides the optimal cutoff, with a classification match of 80%, similar to what has been reported in previous offline studies; (2) when the originally validated 0.75 ischemic FFR cutoff is used as a reference comparison, the agreement between iFR and FFR increases to 88% with the optimal iFR cutoff being 0.85; (3) when accounting for the known 0.75 to 0.80 FFR gray zone, the classification match between iFR and FFR increases to 93%; and (4) confirming previous reports,<xref rid="bb0040" ref-type="bibr">[8]</xref>, <xref rid="bb0045" ref-type="bibr">[9]</xref> a hybrid decision-making strategy with iFR and FFR could spare 61% of patients from vasodilator, while maintaining 94% overall agreement with FFR classification of lesions.</p><sec id="s0080"><title>Real-time iFR measurement in the catheterization laboratory is feasible</title><p id="p0105">One potential limitation applicable to all studies that so far evaluated the relationship between iFR and FFR is the fact that all hemodynamic analyses were performed offline. This inevitable stage in the development of a new technology creates the theoretical possibility of <italic>expertise</italic> bias where only operators specifically trained in the offline analysis of hemodynamic traces, such as in a core laboratory, perform the analysis. We have found that when iFR is implemented in catheter laboratories around Europe, Africa, and Asia and the measurements are performed by clinicians who routinely perform functional assessment using pressure guide wires, the close agreement between iFR and FFR is maintained. Our results, therefore, demonstrate the feasibility of iFR use by the clinical community and the applicability of earlier offline core laboratory reports to clinical practice. <xref rid="f0035" ref-type="fig">Figure&#x000a0;6</xref> shows 2 examples of lesion interrogation by iFR and FFR.<fig id="f0035"><label>Figure&#x000a0;6</label><caption><p>Screenshots of measurements of iFR and FFR. <bold>A</bold>, An example of interrogation in the left circumflex artery (horizontal arrow), in which both iFR and FFR were negative, above their respective cutoffs of 0.90 and 0.80; revascularization was deferred. <bold>B</bold>, An example in which both iFR and FFR revealed a functionally significant stenosis in the proximal segment of the left anterior descending artery (oblique arrow); percutaneous revascularization was performed. Because iFR is performed without the need for vasodilator administration, the time of lesion interrogation is typically reduced to around 5 to 10 seconds, from 60 to 120 seconds for FFR.</p></caption><graphic xlink:href="gr6"/></fig></p></sec><sec id="s0085"><title>Validation of the clinical iFR cutoff and match with FFR classification</title><p id="p0115">In this study, when a clinical FFR cutoff of 0.80 was used as a reference comparison, a prespecified iFR value of 0.90 was also found to be the optimal cutoff identified by ROC, yielding a classification agreement of 80%, with an ROC_AUC of 0.87. This comparison with an FFR of 0.80 is important because it reflects current clinical guidelines,<xref rid="bb0070" ref-type="bibr">¹⁴</xref> is in line with FAME<xref rid="bb0075" ref-type="bibr">¹⁵</xref> and FAME II studies,<xref rid="bb0080" ref-type="bibr">¹⁶</xref> and validates with real-time measurements the results of all previous iFR-FFR studies in clinical populations.<xref rid="bb0010" ref-type="bibr">[2]</xref>, <xref rid="bb0015" ref-type="bibr">[3]</xref>, <xref rid="bb0040" ref-type="bibr">[8]</xref></p></sec><sec id="s0090"><title>Performance of iFR to identify FFR-ischemic stenoses</title><p id="p0125">Despite its clinical use for decision making, it must be remembered that FFR 0.80 is not an ischemia-derived FFR optimal cutoff. Studies that evaluated FFR against measures of myocardial perfusion more consistently identified FFR 0.74 to 0.75 as the optimal FFR cutoff.<xref rid="bb0055" ref-type="bibr">¹¹</xref> The change from FFR 0.75 to 0.80 was subsequently implemented in large clinical trials<xref rid="bb0075" ref-type="bibr">[15]</xref>, <xref rid="bb0080" ref-type="bibr">[16]</xref> and was aimed to increase the negative predictive value of FFR-based decisions and avoid significant lesions being missed by a lower cutoff of FFR 0.75. Therefore, the original and extensively validated <italic>isch</italic>em<italic>ic</italic> 0.75 FFR cutoff^{12 13} is also essential for new indices to be validated against, as it has been shown to provide the best match to previous perfusion modalities.<xref rid="bb0055" ref-type="bibr">[11]</xref>, <xref rid="bb0070" ref-type="bibr">[14]</xref>, <xref rid="bb0075" ref-type="bibr">[15]</xref> Indeed, in ADVISE <italic>in-practice</italic> when such a comparison was made, the agreement between iFR and FFR increased to 88% when an iFR cutoff of 0.85 was used. In addition, the finding of an iFR cutoff of 0.85 being equivalent to an FFR of 0.75 is very supportive to the ischemic iFR cutoff found in the CLARIFY study (iFR 0.86), which used invasive flow as a reference discriminator.<xref rid="bb0025" ref-type="bibr">⁵</xref> Also, similarly to what has been observed with FFR, a higher iFR value of &#x0003e;0.90 increases the negative predictive value of iFR to 95% to exclude ischemic stenoses (FFR &#x02264; 0.75).</p><p id="p0130">Therefore, across studies reported so far, in more than 2,000 patients, there appears to be a consistent ischemic iFR cutoff of 0.85 to 0.86, which matches an FFR 0.75, and a clinical iFR cutoff of iFR 0.89 to 0.90, which provides the best agreement with the clinically used FFR of 0.80. Reassuringly, similar to FFR, our findings confirm the presence of these 2 distinct iFR cutoffs (ischemic and clinical) when measurements are performed in real time, by interventionalists in the catheter laboratory.</p></sec><sec id="s0095"><title>The FFR 0.75 to 0.80 gray zone: safety implications for the development of new indices</title><p id="p0140">Early studies that validated FFR as a new diagnostic method identified optimal cutoffs that varied slightly from 0.66, 0.72, 0.74, to 0.75.<xref rid="bb0055" ref-type="bibr">¹¹</xref> The validity of the 0.75 cutoff was subsequently confirmed in 2 landmark studies, which laid the foundations of FFR as a clinical tool. First, the multitesting study that compared FFR against 3 noninvasive functional methods<xref rid="bb0085" ref-type="bibr">¹⁷</xref> undoubtedly demonstrated that no ischemia was detected when FFR fell between 0.75 and 0.80 and only 1 case when between 0.74 and 0.83. This safety was later translated into clinical outcomes with the results of the DEFER study,<xref rid="bb0060" ref-type="bibr">¹²</xref> which documented a rate of major cardiac events of less than 0.6% per year for deferred stenoses with FFR values of &#x02265;0.75.<xref rid="bb0090" ref-type="bibr">¹⁸</xref> Recently, however, FAME and FAME II have also demonstrated the safety and prognostic importance of revascularizing lesions with FFR &#x02264;0.80.<xref rid="bb0075" ref-type="bibr">[15]</xref>, <xref rid="bb0080" ref-type="bibr">[16]</xref> Therefore, the 0.75 to 0.80 FFR zone is a region within which it is known to be equally safe to defer <italic>and</italic> treat stenoses, where ischemia is almost certainly absent but cannot be always excluded.</p><p id="p0145">This DEFER-FAME gray zone not only is clinically relevant (as it permits clinicians to defer <italic>or</italic> treat stenoses with the same confidence)<xref rid="bb0095" ref-type="bibr">¹⁹</xref> but also has implications to the development of new indices, which use the safety of FFR classification of lesions as a reference comparison. For instance, we have found that when this gray zone is taken into account, and either a DEFER or FAME approach is taken, iFR classification match with FFR increases to 92% with an ROC_AUC of 0.93. This suggests that most of iFR-FFR disagreements are unlikely to be prognostic, as they predominantly fall in the FFR gray zone. Future trials with clinical outcomes will need to confirm this finding prospectively.</p></sec><sec id="s0100"><title>Hybrid approach confirms adenosine-sparing potential of iFR application</title><p id="p0155">Until clinical outcome studies judge the merits of iFR as an independent diagnostic method, clinicians might prefer to maintain a higher magnitude (&#x0003e;90%) of classification agreement with FFR, given its established role to select lesions for revascularization. For this purpose, a hybrid decision-making strategy using both iFR and FFR can be applied. A hybrid analysis on this real-time iFR data set with prespecified cutoffs has reproduced the results reported previously.<xref rid="bb0045" ref-type="bibr">⁹</xref> Our findings confirm the potential feasibility of using iFR and FFR together in a hybrid decision-making approach, in which 61% of patients could be spared from vasodilator, while maintaining the safety of a 94% match with FFR classification of lesions (<xref rid="f0030" ref-type="fig">Figure&#x000a0;5</xref>).</p></sec><sec id="s0105"><title>Study samples of patients undergoing invasive physiological assessment in clinical practice: how do they differ from large randomized clinical trials?</title><p id="p0165">The present study also highlights an important feature of unselected clinical FFR cohorts: they are predominantly formed by intermediate FFR values. In the present sample, mean FFR was 0.82 &#x000b1; 0.10, with 71% of stenoses falling within the 0.70 and 0.90 FFR range and only 20% &#x02264;0.75 and 39% &#x02264;0.80. Importantly, physiologically severe lesions appear extremely rare in samples in which FFR is used to guide clinical decisions, with only 4.6% having an FFR &#x02264;0.6 and 1.3% &#x02264;0.5. These figures are very similar to what has been previously reported in large independent clinical cohorts from Europe<xref rid="bb0010" ref-type="bibr">²</xref> and Asia<xref rid="bb0015" ref-type="bibr">³</xref> and reflect routine clinical application of FFR as recommended per guidelines.<xref rid="bb0070" ref-type="bibr">¹⁴</xref> Importantly, they differ markedly from the distribution of FFR values of previous large FFR clinical trials. For instance, mean FFR of treated lesions in DEFER<xref rid="bb0060" ref-type="bibr">¹²</xref> was 0.57 &#x000b1; 0.16, while in FAME,<xref rid="bb0075" ref-type="bibr">¹⁵</xref> it was 0.60 &#x000b1; 0.14. Notably, in the medical arm of FAME II study,<xref rid="bb0080" ref-type="bibr">¹⁶</xref> 19% and 26% of the stenoses had FFR less than 0.50 and 0.60, respectively, which contrasts with the rarity of such severe lesions in clinical samples.</p><p id="p0170">These differences between unselected clinical cohorts and studies in which patients are actively recruited by investigators are very relevant. First, they demonstrate that interventionalists have an excellent overall clinical judgement as to which stenoses require functional interrogation, as most clinically selected <italic>anatomically</italic> intermediate stenoses are also <italic>physiologically</italic> intermediate (<xref rid="f0020" ref-type="fig">Figure&#x000a0;3</xref>). Second, it appears that clinical populations such as found in this study provide the most conservative scenario with respect to the magnitude of agreement between iFR and FFR, because they are predominantly formed by intermediate FFR values, when disagreements are naturally higher.<xref rid="bb0010" ref-type="bibr">[2]</xref>, <xref rid="bb0100" ref-type="bibr">[20]</xref> In populations with more severe lesions, such as in DEFER or FAME studies, classification agreement between iFR and a clinical FFR cutoff of 0.80 is expected to be higher, around 90%, as previously demonstrated.<xref rid="bb0005" ref-type="bibr">[1]</xref>, <xref rid="bb0105" ref-type="bibr">[21]</xref></p></sec><sec id="s0110"><title>Clinical implications of our findings</title><p id="p0180">Routine clinical application of a vasodilator-free index such as iFR could potentially expand the benefits of physiology-guided revascularization to many more patients with coronary disease, as the need for vasodilator is one of the reasons for the low adoption of FFR.<xref rid="bb0110" ref-type="bibr">²²</xref> In South Africa, for example, 15 ampoules of a specific adenosine preparation are required for one intravenous infusion, at considerable cost. Similar restrictions occur in other countries, such as Russia, Turkey, and in Latin America, where papaverine is still frequently used as a hyperemic agent with its incumbent risks of serious cardiac arrhythmias.<xref rid="bb0115" ref-type="bibr">²³</xref> The costs and procedural delays associated with vasodilator administration (particularly long intravenous infusions, the recommended route for a complete vessel assessment) are particularly relevant to patients with multivessel disease, who might need 3-vessel interrogation. This group of patients is perhaps the one that benefits the most from invasive evaluation of intermediate stenoses,<xref rid="bb0075" ref-type="bibr">[15]</xref>, <xref rid="bb0080" ref-type="bibr">[16]</xref> although it appears to be the most affected by the lack of its widespread adoption.</p></sec><sec id="s0115"><title>Future directions: the need for clinical trials</title><p id="p0190">An accumulation of evidence supports iFR as an index of stenosis severity. However, clinical outcome data are required to further advance the global adoption of iFR as a single measure in clinical decision making. The DEFINE-FLAIR trial (NCT 02053038) will compare strategies of revascularization guided by iFR and FFR in 2,500 patients requiring physiological interrogation in clinical practice. The primary outcome will be the rate of major adverse cardiac events (composite of death, myocardial infarction, and unplanned revascularization) at 1 year (<xref rid="f0040" ref-type="fig">Figure&#x000a0;7</xref>). Other studies, such as the SYNTAX II, Prospect II, and iFR-SWEDEHEART will further evaluate the merits of iFR in clinical decision making.<fig id="f0040"><label>Figure&#x000a0;7</label><caption><p>The FLAIR trial will evaluate the clinical merits of iFR guided revascularization.</p></caption><graphic xlink:href="gr7"/></fig></p></sec><sec id="s0120"><title>Limitations</title><p id="p0200">The present study only performed a direct comparison between the classification of stenoses by iFR and FFR. No alternative noninvasive method was applied in this analysis, which prevents a more in-depth interpretation of iFR-FFR disagreements.</p><p id="p0205">There was no active recruitment of study subjects. Patients included in this study represent the selection of clinicians based on a clinical indication for FFR measurement. Also, no formal measurement protocol for iFR or FFR was suggested, and technical aspects were left to the discretion of the interventionalist according to their routine clinical practice. Also, iFR and FFR measurements were not blinded from each other, as the clinical operator had access to both results.</p><p id="p0210">Although each individual FFR value might have been different if different doses or routes of adenosine were used, it is likely that on average, across the whole study population, this would not affect the overall relationship with iFR because this variability would likely occur in both ways (more or less hyperemia). Importantly, as a retrospective analysis of the routine clinical application of FFR, our study used no fixed protocol of adenosine administration, which means that our results entirely reflect the choice of clinicians and the way FFR is used everyday around the world. This can be seen as a strength of our study because the relationship between iFR and FFR, therefore, reflects the clinical relationship expected when iFR is used for clinical reasons. Also, the recently presented ADVISE II study,<xref rid="bb0120" ref-type="bibr">²⁴</xref> which used a rigid protocol of central IV adenosine administration, demonstrated a level of agreement between iFR and FFR equal to our study, which again confirms the validity of our findings.</p><p id="p0215">Finally, the present study cannot evaluate the impact of iFR use on clinical outcomes because we simply described its diagnostic agreement with FFR in clinical practice. This merit will be addressed by the DEFINE-FLAIR study (<xref rid="f0040" ref-type="fig">Figure&#x000a0;7</xref>) and other future trials.</p></sec></sec><sec id="s0025"><title>Conclusion</title><p id="p0220">Real-time iFR measurement in the cardiac catheterization laboratory is feasible and accessible to clinicians, delivering the same magnitude of agreement between iFR and FFR as described in core laboratory studies. Outcome studies are needed to evaluate the clinical value of iFR as an independent decision-making tool in such population.</p></sec><sec id="s0130"><title>Disclosures</title><p id="p0230">J.E. Davies holds patents pertaining to iFR technology, which is under licence to Volcano Corporation. J.E. Davies is a consultant for Volcano Corporation. The other authors have no conflicts of interest to declare.</p></sec>

Measuring Recovery Related Outcomes: A Psychometric Investigation of the Recovery Markers Inventory

Could not extract abstract

<contrib contrib-type="author"><name><surname>Lusczakoski</surname><given-names>Kate DeRoche</given-names></name><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author"><name><surname>Olmos-Gallo</surname><given-names>P. Antonio</given-names></name><address><email>Antonio.Olmos@du.edu</email></address><xref ref-type="aff" rid="Aff2"/></contrib><contrib contrib-type="author" corresp="yes"><name><surname>McKinney</surname><given-names>C. J.</given-names></name><address><email>publications@mhcd.org</email></address><xref ref-type="aff" rid="Aff3"/></contrib><contrib contrib-type="author"><name><surname>Starks</surname><given-names>Roy</given-names></name><xref ref-type="aff" rid="Aff3"/></contrib><contrib contrib-type="author"><name><surname>Huff</surname><given-names>Steve</given-names></name><xref ref-type="aff" rid="Aff3"/></contrib><aff id="Aff1"><label/>Accelrys, San Diego, CA USA </aff><aff id="Aff2"><label/>Morgridge College of Education, University of Denver, Denver, CO USA </aff><aff id="Aff3"><label/>Mental Health Center of Denver, Denver, CO USA </aff>

Community Mental Health Journal

<p>Spurred by increased interest of members of consumer advocacy groups and recent political influence, professionals in the field of community mental health have undertaken a systemic transformation in order to incorporate the ideas and concepts that facilitate consumers&#x02019; recovery (Olmos-Gallo et al. <xref ref-type="bibr" rid="CR26">2012</xref>). At the same time, mental health researchers and evaluators have focused on a change from the traditional outcomes (e.g., symptom frequency, recidivism, and hospitalization rates) to recovery-oriented outcomes (Davidson and Roe <xref ref-type="bibr" rid="CR6">2007</xref>; Farkas et al. <xref ref-type="bibr" rid="CR9">2005</xref>; Olmos-Gallo and DeRoche <xref ref-type="bibr" rid="CR25">2010</xref>). Typically, recovery outcomes include the assessment of change in consumers&#x02019; perspectives of recovery, commonly defined as changes in: hope, active growth, safety, symptom management, social support, and wellness among others (Lusczakoski et al. <xref ref-type="bibr" rid="CR17">2013</xref>; Onken et al. <xref ref-type="bibr" rid="CR27">2007</xref>). However, due to the deeply individualized nature of the recovery process (Anthony <xref ref-type="bibr" rid="CR1">2000</xref>), direct measurement of this process can be a challenge. In this article, we suggest that in addition to instruments that can be used to measure consumers&#x02019; assessment of their own recovery (Lusczakoski et al. <xref ref-type="bibr" rid="CR17">2013</xref>; O&#x02019;Hare et al. <xref ref-type="bibr" rid="CR24">2003</xref>; Ridgway <xref ref-type="bibr" rid="CR31">2003</xref>; Trauer and Tobias <xref ref-type="bibr" rid="CR35">2004</xref>), providers can use measures related to the consumer&#x02019;s <italic>recovery</italic>-<italic>related factors</italic> to assist in the assessment of change in recovery. This paper describes a measure to assess recovery related factors.</p><p>As implied by the term, recovery-related factors are actions and events that tend to be correlated with consumers&#x02019; recovery, even though the consumers may not necessarily associate them with their own personal journey. In that sense, recovery-related factors can be an <italic>indicator</italic> or <italic>marker of</italic> growth in recovery. For example, participation in mental health services may foster recovery, despite the fact that consumers can recover without the involvement/assistance of professional help (Anthony <xref ref-type="bibr" rid="CR1">2000</xref>). Similarly, a provider&#x02019;s assessment of the consumer&#x02019;s interest in employment or symptom management can be associated with the consumer&#x02019;s recovery, even though recovery is dependent upon the consumer taking control and responsibility for his/her own life (Jacobson and Curtis <xref ref-type="bibr" rid="CR13">2000</xref>). Based upon this rationale, the provider&#x02019;s assessment is defined as an evaluation of recovery-related factors and not as recovery, <italic>per se</italic>. In this article, we define recovery-related factors as including: employment, self-education/learning, participation in services, housing, symptom management, active growth, and substance use/abuse. Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref> presents the operational definitions of these recovery-related factors.<table-wrap id="Tab1"><label>Table&#x000a0;1</label><caption><p>Operational definitions of recovery-related factors</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left">Recovery-related factor</th><th align="left">Operational definition</th></tr></thead><tbody><tr><td align="left">Employment</td><td align="left">Actions toward looking for or maintaining employment, with markers of employment progressing from no interest in employment to searching for employment opportunities, supportive employment, extending up to full time independent employment</td></tr><tr><td align="left">Education/learning</td><td align="left">Actions undertaken to continuing education, with markers including the actions of looking up information on the internet, newspapers, and books, taking skills-oriented classes (e.g., cooking class), as well as vocational and formal education</td></tr><tr><td align="left">Participation in services</td><td align="left">Actions undertaken to self-direct a consumer&#x02019;s recovery, ranging from not engaged in services to promote recovery, engagement, and the ideal level of directing own recovery (i.e., not associated with the frequency of services received)</td></tr><tr><td align="left">Housing</td><td align="left">Actions taken toward obtaining stable housing, with markers ranging from homelessness to residential housing, group homes, up to independent living</td></tr><tr><td align="left">Symptom management</td><td align="left">Consumers&#x02019; ability to cope with their symptoms within their daily lives ranging from <bold>No Impact</bold> (i.e., no impact on ability to function) to <bold>Very High</bold> (i.e., very high impact on the ability to interact with other people, or engage in work, etc.)</td></tr><tr><td align="left">Active growth</td><td align="left">Actions taken to seek and engage in activities within and outside the mental health center</td></tr><tr><td align="left">Substance use/abuse</td><td align="left">Consumers&#x02019; level of use across different substances and their stage of change (Prochaska et al. <xref ref-type="bibr" rid="CR29">1992</xref>) regarding their use</td></tr></tbody></table></table-wrap></p><p>The Recovery Markers Inventory (RMI; developed by the authors) is a short survey completed by case managers/clinicians, intended to measure the recovery-related factors described in Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>. Initial development of the RMI included a review of the literature and use of data collected from consumer focus groups. This process was supplemented by feedback provided by therapists, consumers, and members of the Mental Health Center of Denver&#x02019;s (MHCD) Recovery Committee (see: Olmos-Gallo et al. <xref ref-type="bibr" rid="CR26">2012</xref> for background about the MHCD Recovery Committee). RMI items have been used to generate a summed score, implying the measure is unidimensional, but no dimensionality analyses have been conducted to support that use.</p><p>The purpose of this study was to determine the unidimensionality, reliability, construct, and convergent validities of the RMI in a sample of adults with severe and persistent mental illness, who received services at a community mental health center at a major metropolitan city in the United States. We estimated reliability and unidimensionality using Rasch analysis (Bond and Fox <xref ref-type="bibr" rid="CR5">2001</xref>). We further tested unidimensionality conducting a confirmatory factor analysis in three steps, as suggested by Bollen (<xref ref-type="bibr" rid="CR4">1989</xref>) and Tabachnick and Fidell (<xref ref-type="bibr" rid="CR34">2013</xref>): (1) we split the original sample into two random halves, (2) fit a unidimensional model to one of the samples (calibration) and (3) cross-validated results with the second sample (validation). Finally, we tested convergent validity by calculating the correlation between the RMI and two measures of general functioning, and one intended to measure recovery from the consumer&#x02019;s perspective.</p><sec id="Sec1" sec-type="materials|methods"><title>Methods</title><sec id="Sec2"><title>Participants and Procedure<xref ref-type="fn" rid="Fn1">1</xref></title><p>A total of 1,513 consumers, who had RMIs completed by the case manager/clinician who regularly works with the consumer, in the MHCD Management Information System (MIS) during the month of July 2009, were included in the study. The MIS prompts clinicians/case managers every quarter to complete the instrument for every consumer in his/her caseload, and this reminder will not be removed from the to-do list until the RMI is completed. The RMI scores were matched through use of a unique ID number to demographic information and three other outcomes instruments: the Colorado Client Assessment Record (CCAR; Ellis et al. <xref ref-type="bibr" rid="CR8">1984</xref>); the Global Assessment of Functioning (GAF; Greenberg and Rosenheck <xref ref-type="bibr" rid="CR11">2005</xref>; Jones et al. <xref ref-type="bibr" rid="CR14">1995</xref>); and the Consumer Recovery Measure (CRM; Lusczakoski et al. <xref ref-type="bibr" rid="CR17">2013</xref>). These three instruments were completed within 30&#x000a0;days of the RMI. The CCAR is completed by the clinician/case manager who regularly works with the consumer; the GAF is completed by the psychiatrist who oversees the consumer&#x02019;s treatment during the most recent visit, and the CRM is completed by the consumer him/herself. The participants&#x02019; demographic information is presented in Table&#x000a0;<xref rid="Tab2" ref-type="table">2</xref>.<table-wrap id="Tab2"><label>Table&#x000a0;2</label><caption><p>Participant demographics</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left">Variable</th><th align="left">Range</th><th align="left">Mean (SD)</th></tr></thead><tbody><tr><td align="left">Age</td><td align="left">18&#x02013;95&#x000a0;years</td><td align="left">44.07 (12.11)</td></tr><tr><td align="left">Time in treatment at MHCD</td><td align="left">0&#x02013;240&#x000a0;months</td><td align="left">68&#x000a0;months (72.5)</td></tr></tbody></table><table frame="hsides" rules="groups"><thead><tr><th align="left">Variable</th><th align="left">Frequency</th><th align="left">Percent</th></tr></thead><tbody><tr><td align="left">Gender</td><td align="left"/><td align="left"/></tr><tr><td align="left">&#x000a0;Female</td><td char="." align="char">832</td><td char="." align="char">55.0</td></tr><tr><td align="left">&#x000a0;Male</td><td char="." align="char">681</td><td char="." align="char">45.0</td></tr><tr><td align="left">Martial status</td><td align="left"/><td align="left"/></tr><tr><td align="left">&#x000a0;Never married</td><td char="." align="char">849</td><td char="." align="char">56.1</td></tr><tr><td align="left">&#x000a0;Divorced</td><td char="." align="char">430</td><td char="." align="char">28.4</td></tr><tr><td align="left">&#x000a0;Separated</td><td char="." align="char">94</td><td char="." align="char">6.2</td></tr><tr><td align="left">&#x000a0;Married</td><td char="." align="char">85</td><td char="." align="char">5.6</td></tr><tr><td align="left">&#x000a0;Widowed</td><td char="." align="char">47</td><td char="." align="char">3.1</td></tr><tr><td align="left">Ethnicity</td><td align="left"/><td align="left"/></tr><tr><td align="left">&#x000a0;White</td><td char="." align="char">818</td><td char="." align="char">54.1</td></tr><tr><td align="left">&#x000a0;African American</td><td char="." align="char">392</td><td char="." align="char">25.9</td></tr><tr><td align="left">&#x000a0;Hispanic</td><td char="." align="char">312</td><td char="." align="char">20.6</td></tr><tr><td align="left">&#x000a0;American Indian</td><td char="." align="char">74</td><td char="." align="char">4.9</td></tr><tr><td align="left">&#x000a0;Asian</td><td char="." align="char">11</td><td char="." align="char">0.7</td></tr><tr><td align="left">&#x000a0;Hawaiian</td><td char="." align="char">7</td><td char="." align="char">0.5</td></tr><tr><td align="left">Primary diagnosis category</td><td align="left"/><td align="left"/></tr><tr><td align="left">&#x000a0;Bipolar</td><td char="." align="char">442</td><td char="." align="char">29.2</td></tr><tr><td align="left">&#x000a0;Schizoaffective</td><td char="." align="char">330</td><td char="." align="char">21.8</td></tr><tr><td align="left">&#x000a0;Depression</td><td char="." align="char">321</td><td char="." align="char">21.2</td></tr><tr><td align="left">&#x000a0;Schizophrenia</td><td char="." align="char">260</td><td char="." align="char">17.2</td></tr><tr><td align="left">&#x000a0;Other</td><td char="." align="char">104</td><td char="." align="char">6.9</td></tr><tr><td align="left">&#x000a0;Post traumatic stress</td><td char="." align="char">56</td><td char="." align="char">3.7</td></tr></tbody></table></table-wrap></p></sec></sec><sec id="Sec3"><title>Instruments</title><sec id="Sec4"><title>Recovery Marker Inventory (RMI)</title><p>As described earlier, the RMI was designed to measure recovery-related factors defined as the provider&#x02019;s assessment in the areas of employment, self-education/learning, participation in services, housing, symptom management, active growth, and substance use/abuse. With the exception of substance use/abuse, every area is measured by a single question with a different number of options. The different number of response options is intended to include all the significant measureable changes in the trait of interest. Thus, employment has six options, self-education/learning has 12, participation in services, 6; housing, 11; symptom management, 5; and finally, active growth orientation, has 6 options. It is worth noting that housing has 11 responses because they are also used for State and Federal reporting; however, for scoring purposes, the responses are collapsed into three general categories: (1) unstable/transitional housing (i.e., street, friends/motel homelessness); (2) stable housing (i.e., assisted living, congregate apartments, or single room occupancy); and (3) independent living. The assessment for substance use/abuse requires two steps. First, the provider assesses the consumer&#x02019;s level of use for ten substances (Alcohol, Cannabis, Cocaine, Hallucinogens, Inhalants, Methamphetamines, Opiates, Over-the-counter, PCP, Sedatives/Hypnotics/Anxiolytics, and Stimulants). Second, if substance use is reported, the consumer&#x02019;s stage of change is measured for the specific substance (Prochaska et al. <xref ref-type="bibr" rid="CR29">1992</xref>). Given that approximately one-half of all consumers report substance use/abuse (National Alliance on Mental Illness <xref ref-type="bibr" rid="CR23">2012</xref>), in the present study, the highest level of consumer use among all substances was applied as a moderator. The RMI was completed by the consumer&#x02019;s regular case manager or clinician, through the on-line Medical Information System.</p></sec><sec id="Sec5"><title>Colorado Client Assessment Record (CCAR)-Recovery and Global Assessment of Functioning (GAF)</title><p>The CCAR is the outcomes instrument used by the Division of Behavioral Health in the State of Colorado to assess mental health (Ellis et al. <xref ref-type="bibr" rid="CR8">1984</xref>), and is administered on a scheduled annual basis to all MHCD consumers. A revised version of this instrument includes items which measure: social support, hope, activity involvement, empowerment, and interpersonal relationships. A 9 point rating response scale is used, and this instrument is termed, the CCAR <italic>recovery scale</italic> (Menefee, personal communication, April 18, 2008). For the current sample, the CCAR recovery scale showed acceptable internal consistency (&#x003b1;&#x000a0;=&#x000a0;.85).</p><p>The GAF scale (Greenberg and Rosenheck <xref ref-type="bibr" rid="CR11">2005</xref>; Jones et al. <xref ref-type="bibr" rid="CR14">1995</xref>) is used by psychiatrists at MHCD to rate consumers&#x02019; functioning on a regular basis (i.e., at least once a year). The GAF is administered as part of the assessment submitted to the Division of Behavioral Health. The scores range from 0 (i.e., lowest functioning) to 100 (i.e., highest functioning). The GAF scores are used to assess overall functioning for individuals who struggle with their mental health, and they are used to assess the concurrent validity of other recovery scales (Fisher et al. <xref ref-type="bibr" rid="CR10">2009</xref>). The Office of Behavioral Health at the State level provides training as part of their CCAR package, but no estimates of the reliability of the GAF are yet available. However, strong, statistically significant correlations have been observed against the CCAR&#x02019;s symptom severity, and overall level of functioning scales (Mahalik personal communication, February <xref ref-type="bibr" rid="CR19">2014</xref>). Although no intraclass reliability has been calculated, psychiatrists receive regular refreshers on scoring.</p></sec><sec id="Sec6"><title>Consumer Recovery Measure (CRM)</title><p>The Consumer Recovery Measure (CRM: Lusczakoski et al. <xref ref-type="bibr" rid="CR17">2013</xref>) is used to evaluate consumers&#x02019; perceptions of their own recovery, and it is completed by consumers every quarter-year at MHCD. The survey consists of 17, 4 point rating response items, from Strongly Agree to Strongly Disagree. It is used to rate consumers&#x02019; self-perceptions of hope, safety, symptom management, social network, and active growth. The instrument was developed with use of a Rasch rating scale model (Bond and Fox <xref ref-type="bibr" rid="CR5">2001</xref>) and has a person separation reliability of 0.83 and an item separation reliability of 0.96 (Lusczakoski et al. <xref ref-type="bibr" rid="CR17">2013</xref>).</p></sec><sec id="Sec7"><title>Data Analysis</title><p>To estimate the psychometric properties of the RMI, three analytical techniques were applied. These were: (a) reliability estimation and construct validity using Rasch analysis; (b) construct validity using confirmatory factor analysis (CFA). We randomly split the original sample into two halves, then fit a unidimensional model to one of the samples (calibration) and cross-validated the results with the second sample (validation); and (c) concurrent validity through Pearson correlations between the RMI and the CCAR Recovery Scale (Menefee <xref ref-type="bibr" rid="CR21">2008</xref>, personal communication), the GAF (Greenberg and Rosenheck <xref ref-type="bibr" rid="CR11">2005</xref>), and the CRM (Luzczakoski et. al. 2012).</p><p>First, a Rasch Partial Credit Model was estimated with the use of Winsteps 3.64 (Linacre, <xref ref-type="bibr" rid="CR16">2007</xref>) to determine the person and item reliability and dimensionality, as well as in fit and outfit estimates (Bond and Fox <xref ref-type="bibr" rid="CR5">2001</xref>).</p><p>Next, the CFA models were estimated with use of LISREL 8.80 (J&#x000f6;reskog and S&#x000f6;rbom <xref ref-type="bibr" rid="CR15">2006</xref>), based on the procedures suggested by Bollen (<xref ref-type="bibr" rid="CR4">1989</xref>) and Tabachnick and Fidell (<xref ref-type="bibr" rid="CR34">2013</xref>). CFA is a confirmatory technique used to test a theory about measure structure (Tabachnick and Fidell <xref ref-type="bibr" rid="CR34">2013</xref>); in the present case, a single factor defining recovery-related factors. Participants&#x02019; were randomly divided into two datasets: (a) a calibration sample (<italic>n</italic>&#x000a0;=&#x000a0;737) to estimate the structure components and (b) a cross-validation sample (<italic>n</italic>&#x000a0;=&#x000a0;776) to verify them (Bollen <xref ref-type="bibr" rid="CR4">1989</xref>; Tabachnick and Fidell <xref ref-type="bibr" rid="CR34">2013</xref>). To account for the ordinal nature of the items, both CFA models were estimated using polychoric correlation matrices with robust maximum likelihood estimation procedures. Multivariate normality was evaluated and found to hold across all items. Model fit was assessed using Hu and Bentler&#x02019;s (<xref ref-type="bibr" rid="CR12">1999</xref>) guidelines: non-normed fit index &#x02265;.95 (NNFI; Bentler and Bonett <xref ref-type="bibr" rid="CR3">1980</xref>); comparative fit index &#x02265;.96 (CFI; Bentler <xref ref-type="bibr" rid="CR2">1990</xref>); and root mean square error of approximation (RMSEA; Steiger <xref ref-type="bibr" rid="CR33">1989</xref>) &#x02264;.05 indicating good fit or &#x02264;.08 for reasonable fit (MacCallum et al. <xref ref-type="bibr" rid="CR18">1996</xref>). Finally, Pearson-product moment correlations were computed between the RMI and the CCAR recovery scale, GAF score, and the MHCD CRM with use of SPSS 13 (SPSS Inc <xref ref-type="bibr" rid="CR32">2005</xref>).</p></sec></sec><sec id="Sec8" sec-type="results"><title>Results</title><sec id="Sec9"><title>Rasch Analysis</title><p>The RMI showed acceptable reliability estimates, with a person separation reliability of 0.74 and an item separation reliability of 1.00. As displayed in Table&#x000a0;<xref rid="Tab3" ref-type="table">3</xref>, all six items in the RMI produced in fit and outfit values within an acceptable range (0.5&#x02013;1.5). In terms of difficulty to endorse, self-education and employment were the two most difficult items, followed by active growth, participation, symptoms, and housing. As evidence of the unidimensionality, the strength of the first contrast following a principal components analysis of residuals was less than the suggested cutoff value of 2.0 (Bond and Fox <xref ref-type="bibr" rid="CR5">2001</xref>; Linacre <xref ref-type="bibr" rid="CR16">2007</xref>). In terms of model misfit, the majority of consumers&#x02019; Rasch model misfit was due to unexpectedly low scores in the housing item. That is, these consumers improved in every other area, but were still homeless or living in transitional housing.<table-wrap id="Tab3"><label>Table&#x000a0;3</label><caption><p>Infit, outfit, and difficulty estimates for the RMI items</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left">Items</th><th align="left">Infit</th><th align="left">Outfit</th><th align="left">Item difficulty</th></tr></thead><tbody><tr><td align="left">Self-education</td><td char="." align="char">.89</td><td char="." align="char">1.05</td><td char="." align="char">1.13</td></tr><tr><td align="left">Employment</td><td char="." align="char">1.04</td><td char="." align="char">.98</td><td char="." align="char">.90</td></tr><tr><td align="left">Participation</td><td char="." align="char">1.01</td><td char="." align="char">1.02</td><td char="." align="char">&#x02212;.01</td></tr><tr><td align="left">Active growth orientation</td><td char="." align="char">.82</td><td char="." align="char">.83</td><td char="." align="char">&#x02212;.03</td></tr><tr><td align="left">Symptom management</td><td char="." align="char">.94</td><td char="." align="char">.95</td><td char="." align="char">&#x02212;.37</td></tr><tr><td align="left">Housing</td><td char="." align="char">1.26</td><td char="." align="char">1.48</td><td char="." align="char">&#x02212;1.62</td></tr><tr><td align="left">Average</td><td char="." align="char">.99</td><td char="." align="char">1.05</td><td char="." align="char">&#x02013;</td></tr></tbody></table></table-wrap></p></sec><sec id="Sec10"><title>Construct Validity</title><p>The specification of the calibration model included a single latent factor (i.e., recovery-related factor). Regarding substance use, the highest level among all ten substances served as a moderator variable (i.e., exogenous variable). The CFA model structure is presented in Fig.&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>. The consumer&#x02019;s highest level of substance use was negatively correlated to the recovery-related factor (<italic>r</italic>&#x000a0;=&#x000a0;&#x02212;25, <italic>p</italic>&#x000a0;&#x0003c;&#x000a0;.01), which suggests that, as a consumer&#x02019;s substance use increases, the recovery-related factors decrease. Given that the fit statistics were acceptable in the calibration model (NNFI&#x000a0;=&#x000a0;.96, CFI&#x000a0;=&#x000a0;.97, RMSEA&#x000a0;=&#x000a0;.06, CI&#x000a0;=&#x000a0;.04, .08), the model for the validation sample was specified according to the factor loadings and variance&#x02013;covariance matrices estimated in the calibration model. The fit of the validation model was also acceptable (NNFI&#x000a0;=&#x000a0;1.01, CFI&#x000a0;=&#x000a0;1.00, RMSEA&#x000a0;&#x0003c;&#x000a0;.01, CI&#x000a0;&#x0003c;&#x000a0;.01, .01). Furthermore, a comparison between the validation and the calibration models showed no significant deviations (S&#x02013;B &#x003c7;²&#x000a0;=&#x000a0;0.0 (<italic>df</italic>&#x000a0;=&#x000a0;28), <italic>p</italic>&#x000a0;&#x0003e;&#x000a0;.05). The findings from the construct validity analysis supported by Rasch analysis provided evidence that the six items in the RMI can be combined to create a total recovery-related latent factor score, which is moderated by a consumer&#x02019;s highest level of substance use.<fig id="Fig1"><label>Fig.&#x000a0;1</label><caption><p>CFA model with standardized factor load. <italic>Boxes</italic> represent observed variables, the <italic>large circle</italic> represents the latent variable and the <italic>small</italic>, <italic>dark circles</italic> represent measurement errors associated with observed variables</p></caption><graphic xlink:href="10597_2014_9728_Fig1_HTML" id="MO1"/></fig></p></sec><sec id="Sec11"><title>Concurrent Validity Evidence</title><p>The RMI exhibited significant correlations with other recovery measures. The strongest relationship was with the CCAR recovery scale (<italic>r</italic>&#x000a0;=&#x000a0;&#x02212;.46, <italic>p</italic>&#x000a0;&#x0003c;&#x000a0;.01, 95&#x000a0;% CI &#x02212;.39, &#x02212;.52). The RMI showed statistically significant correlations with the GAF (<italic>r</italic>&#x000a0;=&#x000a0;.23, <italic>p</italic>&#x000a0;&#x0003c;&#x000a0;.01, 95&#x000a0;% CI .15, .32) and the CRM (<italic>r</italic>&#x000a0;=&#x000a0;.19, <italic>p</italic>&#x000a0;&#x0003c;&#x000a0;.01, 95&#x000a0;% CI .12, .25). It is worth noting that the negative correlation with the CCAR is due to the fact that the CCAR is used to measure severity, not recovery (Fig.&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>).<fig id="Fig2"><label>Fig.&#x000a0;2</label><caption><p>Item/person map of the Recovery Markers Inventory. <italic>Note</italic> The distribution of consumers&#x02019; ability scores are shown on the <italic>left side</italic> of the line; difficulty scores are depicted on the <italic>right side</italic> of the ruler</p></caption><graphic xlink:href="10597_2014_9728_Fig2_HTML" id="MO2"/></fig></p></sec></sec><sec id="Sec12" sec-type="discussion"><title>Discussion</title><p>The purpose of the study was to investigate the psychometric properties of the RMI, an instrument used to measure the factors associated with recovery from mental illness. As explained in the introduction, the RMI is not intended to measure recovery, but actions and events associated with growth that can be associated with it.</p><p>Evidence was provided by this study to support the RMI as a valid and reliable measure of recovery-related factors. The results from this study demonstrated that, with a community-based sample of consumers, the RMI showed (acceptable internal reliability estimates, unidimensionality as a measure of recovery-related factors, and evidence for concurrent validity in the measurement of consumer recovery and general overall functioning. Since the RMI has been used only with consumers receiving services at community-based mental health centers, it is suggested that similar settings can effectively adopt it. It is also worth notice that several of the indicators included under the RMI have been used on a regular basis to report outcomes to the Federal and State governments. Thus, the RMI can play a dual role by: (a) creating an accountability system for external stakeholders and funders; and (b) helping clinical staff, consumers, and their family members to observe improvements over time.</p><p>The RMI represents one of several instruments in the multi-pronged measurement approach undertaken by the Mental Health Center of Denver (MHCD) to understand recovery from mental illness. As described elsewhere (Olmos-Gallo et al. <xref ref-type="bibr" rid="CR26">2012</xref>), we have to date developed several instruments to explore recovery from multiple perspectives, in a way that can be easily integrated into clinical practice. The combination of multiple instruments has allowed MHCD staff to develop an outcomes-centered approach to recovery, which has already helped to demonstrate significant improvements. For example, since MHCD started to track recovery, with use of the RMI and CRM (Lusczakoski et al. <xref ref-type="bibr" rid="CR17">2013</xref>), staff of the two succeed in Employment program has reported a significant increase (38&#x000a0;%) in employment status (Olmos-Gallo et al.), and over 25&#x000a0;% obtained employment within 8&#x000a0;months of enrollment in the program. Approximately 55&#x000a0;% of the consumers in the same program increased their interest in education/learning activities (Outcomes Quarterly, Fall <xref ref-type="bibr" rid="CR28">2010</xref>).</p><p>The MHCD multi-measure approach has been used to develop a holistic approach toward recovery that takes into account the fact that recovery happens over time, and that there is always the possibility of relapse (Olmos-Gallo and DeRoche <xref ref-type="bibr" rid="CR25">2010</xref>). We have developed models with use of hierarchical linear modeling (Raudenbush and Bryk <xref ref-type="bibr" rid="CR30">2002</xref>), which can accommodate the hierarchical nature of the longitudinal data collected,<xref ref-type="fn" rid="Fn2">2</xref> in order to explore the nature of change over time and the potential impact of program characteristics (e.g., modality, intensity) or background variables (e.g., gender, ethnicity, diagnostic). More recently, we have added the use of statistical control charts (Montgomery <xref ref-type="bibr" rid="CR22">2008</xref>) that allow multiple stakeholders (e.g., managers, clinicians, consumers) to monitor and provide feedback on consumers&#x02019; progress as they move forward in their recovery. Outcomes information is combined with other relevant information (e.g., diagnosis, demographic) in a dashboard that presents it in a user-friendly manner. The combination of valid and reliable recovery-outcome measures, development of predictive models, and statistical control charts has helped MHCD staff to monitor clinical data with an emphasis on quality improvement (McLean et al. <xref ref-type="bibr" rid="CR20">2010</xref>).</p><p>More importantly, being able to measure recovery from different perspectives allows the staff to ask questions that can help them in their practice. For example: Are medications just relieving symptoms, or helping people improve? Are there some interventions that are better suited to consumers with specific characteristics? What kinds of services help the most when it comes to improving the lives of those with a severe and persistent mental illness? What is the impact of hope in recovery? The answer to these and similar questions will not be possible in the absence of valid and reliable instruments to measure recovery-based outcomes. In this study, we have shown that the RMI holds promise as a reliable and valid measure of factors associated with recovery from mental illness that can be used to answer such questions. </p></sec>

False and true pre-treatment predictors of weight loss in obese patients starting a program for lifestyle change

Could not extract abstract

<contrib contrib-type="author"><name><surname>Cresci</surname><given-names>Barbara</given-names></name><xref ref-type="aff" rid="Aff1">1</xref></contrib><contrib contrib-type="author"><name><surname>Pala</surname><given-names>Laura</given-names></name><xref ref-type="aff" rid="Aff1">1</xref></contrib><contrib contrib-type="author"><name><surname>Poggiali</surname><given-names>Roberta</given-names></name><xref ref-type="aff" rid="Aff2">2</xref></contrib><contrib contrib-type="author"><name><surname>Guarnieri</surname><given-names>Cosetta</given-names></name><xref ref-type="aff" rid="Aff2">2</xref></contrib><contrib contrib-type="author"><name><surname>Mannucci</surname><given-names>Edoardo</given-names></name><xref ref-type="aff" rid="Aff3">3</xref></contrib><contrib contrib-type="author"><name><surname>Bigiarini</surname><given-names>Michela</given-names></name><xref ref-type="aff" rid="Aff4">4</xref></contrib><contrib contrib-type="author" corresp="yes"><name><surname>Rotella</surname><given-names>Carlo Maria</given-names></name><address><email>carlomaria.rotella@unifi.it</email></address><xref ref-type="aff" rid="Aff4">4</xref></contrib><aff id="Aff1"><label>1</label>Endocrinology Unit, Careggi University Hospital, Florence, Italy </aff><aff id="Aff2"><label>2</label>Dietetic Unit, Careggi University Hospital, Florence, Italy </aff><aff id="Aff3"><label>3</label>Diabetes Agency, Careggi University Hospital, Florence, Italy </aff><aff id="Aff4"><label>4</label>Department of Biomedical Experimental and Clinical Sciences and Obesity Agency, University of Florence, Viale Pieraccini 6, 50134 Florence, Italy </aff>

Eating and Weight Disorders

<sec id="Sec1" sec-type="introduction"><title>Introduction</title><p>Weight loss programs for obese patients, particularly those lifestyle change-oriented and without any pharmacological treatment, are known to be time consuming and usually show low percentages of success in terms of weight loss for a relevant number of different reasons [<xref ref-type="bibr" rid="CR1">1</xref>]. At this purpose, weight loss treatment effectiveness and cost-effectiveness may be improved, not only by the identification of factors affecting weight loss, but mainly by the identification of patients who are sufficiently motivated and thus more able to participate and gain benefit from the intervention.</p><p>Many parameters have been identified with a potential capacity of predicting success/failure in a weight loss program [<xref ref-type="bibr" rid="CR2">2</xref>&#x02013;<xref ref-type="bibr" rid="CR4">4</xref>]. Among these, the most suggestive could be considered: self-motivation [<xref ref-type="bibr" rid="CR5">5</xref>&#x02013;<xref ref-type="bibr" rid="CR7">7</xref>]; self-efficacy [<xref ref-type="bibr" rid="CR8">8</xref>]; active lifestyle [<xref ref-type="bibr" rid="CR9">9</xref>]; more relevant initial weight loss [<xref ref-type="bibr" rid="CR10">10</xref>]; social and/or family support [<xref ref-type="bibr" rid="CR11">11</xref>, <xref ref-type="bibr" rid="CR12">12</xref>].</p><p>Our group has already explored this area, reporting that therapeutic success is predicted by basal TRE-MORE test scores and by estimated muscle mass. In fact, we have found that a higher muscle mass, as estimated through bioimpedance analysis (BIA), is associated with a greater weight loss in obese patients [<xref ref-type="bibr" rid="CR13">13</xref>]. Moreover, we demonstrated that therapeutic success is predicted by TRE-MORE test scores. In particular, TRE-MORE total score is a predictor of failure, but not of attendance, whereas drop-out patients showed a lower score only in TREMORE-3 subscale which investigates lifestyle habits [<xref ref-type="bibr" rid="CR14">14</xref>, <xref ref-type="bibr" rid="CR15">15</xref>].</p><p>A substantial body of literature exists suggesting that weight loss can be achieved by varying the macronutrient distribution and composition of dietary factors. Champagne et al. [<xref ref-type="bibr" rid="CR16">16</xref>] have investigated the effect of dietary intake modifications on weight loss and maintenance during an intensive behavioral weight loss program, concluding that success was mainly associated with increased intake of proteins, fruits and vegetables, and low-fat dairy. Moreover, many studies have explored the possibility of using the variation in the intake of different macronutrients as predictors of weight loss [<xref ref-type="bibr" rid="CR17">17</xref>&#x02013;<xref ref-type="bibr" rid="CR19">19</xref>]. Very few studies, on the contrary, explore pre-treatment dietary habits impact on treatment success. Recently, Byrne et al. [<xref ref-type="bibr" rid="CR20">20</xref>] evaluated the effects of pre-treatment self-efficacy for diet and exercise, as well as changes in self-efficacy occurring during treatment, on weight loss success, demonstrating that treatment attendance and changes in exercise self-efficacy during treatment were stronger predictors of weight loss than changes in diet.</p><p>It is well known that obesity is frequently associated to many other morbidities. The majority of these complications are related to comorbid conditions that include coronary artery disease, hypertension, type 2 diabetes mellitus, respiratory disorders and dyslipidemia. Therefore, patients starting a weight loss program are usually screened for biohumoral parameters related to these complications, such as lipid profile, glycemia, insulinemia, blood pressure levels and liver function, the potentiality of which as predictors has not yet been demonstrated.</p><p>Aim of the present study is to identify easily available predictors that could be used as additional predictors of weight loss among data present in the medical records of obese/overweight patients attending an outpatient clinic.</p></sec><sec id="Sec2" sec-type="materials|methods"><title>Methods</title><sec id="Sec3"><title>Patients</title><p>The study was carried out in the Outpatient Clinic of the Obesity Agency of the University of Florence. Complete baseline data including blood samples results were collected for 268 patients out of the enrolled 331 seeking treatment for overweight/obesity in our outpatient clinic [<xref ref-type="bibr" rid="CR15">15</xref>].</p><p>Inclusion and exclusion criteria are shown in Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>.<table-wrap id="Tab1"><label>Table&#x000a0;1</label><caption><p>Inclusion and exclusion criteria of patients</p></caption><table frame="hsides" rules="groups"><tbody><tr><td align="left">Inclusion criteria (A)</td></tr><tr><td align="left">&#x000a0;Obesity or overweight (BMI&#x000a0;&#x02265;&#x000a0;27&#x000a0;kg/m²)</td></tr><tr><td align="left">&#x000a0;Age 18&#x02013;65</td></tr><tr><td align="left">&#x000a0;Residence within 40&#x000a0;km from the Clinic</td></tr><tr><td align="left">&#x000a0;Informed consent</td></tr><tr><td align="left">Exclusion criteria (B)</td></tr><tr><td align="left">&#x000a0;Patients living at distances more than 40&#x000a0;km</td></tr><tr><td align="left">&#x000a0;Uncontrolled endocrine disorders, such as hypo- or hyperthyroidism</td></tr><tr><td align="left">&#x000a0;Diabetes</td></tr><tr><td align="left">&#x000a0;Pregnancy</td></tr><tr><td align="left">&#x000a0;Illiteracy, or inadequate knowledge of the Italian language</td></tr><tr><td align="left">&#x000a0;Any condition interfering with the possibility of regular physical exercise (e.g., severe cardiac dysfunction, severe respiratory insufficiency, major neurologic disorders, etc.)</td></tr><tr><td align="left">&#x000a0;Diagnosis of major depression, bipolar disorder, obsessive&#x02013;compulsive disorder, schizophrenia, or mental retardation</td></tr><tr><td align="left">&#x000a0;Current treatment with antipsychotics, antiepileptics, tricyclic antidepressants, or lithium</td></tr><tr><td align="left">&#x000a0;Intention to move more than 40&#x000a0;km far from Florence in the following 12&#x000a0;months</td></tr></tbody></table></table-wrap> </p><p>Before the collection of data, during the first routine visit, the procedures of the study were fully explained; after that, the patients were asked to provide their written informed consent to the participation to the study. The study protocol had been previously approved by the Local Ethical Committee. The treatment protocol was carried out as previously described [<xref ref-type="bibr" rid="CR15">15</xref>].</p></sec><sec id="Sec4"><title>Measurements</title><p>At baseline, an Endocrinologist collected an accurate medical history, performed a complete physical examination and measured anthropometric parameters (BMI, waist circumference, blood pressure and heart rate). Body fat (BF) and lean body mass were estimated through BIA, performed by a tetrapolar single frequency (50&#x000a0;kHz) phase-sensitive impedance analyzer (QUANTUM/S; AkernSrl, Firenze, Italy). In the same day, the patients also met a dietician (R.P. and C.G.) who assessed usual food intake on the basis of a 30-day recall, using a dedicated software (Gedip by Solutions S.n.c, Italy), together with photo atlas for the determination of portions (Scotti-Bassani Institute, Italy). A goal of a 500&#x000a0;kcal/day reduction from usual food intake was agreed upon with the patient. To reach that objective, the patients were asked to self-monitor their food intake for at least a week before each monthly follow-up visit with the dietitian [<xref ref-type="bibr" rid="CR15">15</xref>]. Blood samples were drawn after 12-h overnight fast for glucose (Beckman instruments, Fullerton, CA, USA), TSH (Electrochemiluminescent, Modular Roche, Milan, Italy), transaminases, gamma-GT, insulin (electrochemiluminescence immunoassay, Roche Diagnostics, Mannheim, Germany), cholesterol, HDL-cholesterol, triglycerides (for lipid panel: Abbott Bichromatic Analyser (Abbott Diagnostics, South Pasadena, CA, USA). All laboratory determinations were performed in the central Laboratory of Careggi Hospital in Florence.</p><p>Final measures, including BMI and waist, were considered those collected after 6&#x000a0;months from the enrollment in the protocol.</p></sec><sec id="Sec5"><title>Definition of outcomes</title><p>Therapeutic success was defined as a weight loss at 6&#x000a0;months of at least 5&#x000a0;% from baseline to assess this outcome.</p></sec><sec id="Sec6"><title>Statistical analysis</title><p>Continuous variables were reported as mean&#x000a0;&#x000b1;&#x000a0;standard deviations (SD), whereas categorical variables were reported as median and limits of confidence. Between-group comparisons of continuous variables were performed using unpaired Student&#x02019;s<italic> t</italic> test or Mann&#x02013;Whitney <italic>U</italic> test, for variables with normal or non-normal distribution, respectively. Correlation analyses were performed with Spearman&#x02019;s method. Logistic regression analyses were applied for dichotomous outcomes (successes vs. failures).</p><p>All analyses were performed using SPSS for Windows 15.0 (Chicago Inc, USA).</p></sec></sec><sec id="Sec7" sec-type="results"><title>Results</title><p>268 patients, 74 men and 195 women (age 43.2&#x000a0;&#x000b1;&#x000a0;11.9&#x000a0;years, weight 105.6&#x000a0;&#x000b1;&#x000a0;22.6&#x000a0;kg, BMI 38.9&#x000a0;&#x000b1;&#x000a0;6.8&#x000a0;kg/m², waist 116.1&#x000a0;&#x000b1;&#x000a0;16.2&#x000a0;cm) were enrolled. Among these patients, only 26 (35.6&#x000a0;%) men and 44 (22.7&#x000a0;%) women completed the 6-month protocol. Among participants, 50.7&#x000a0;% (<italic>n</italic>&#x000a0;=&#x000a0;70) lost at least 5&#x000a0;% initial body weight after 6&#x000a0;months (and will be thereafter called SUCCESSES), while 49.3&#x000a0;% (<italic>n</italic>&#x000a0;=&#x000a0;68) failed (FAILURES).</p><p>Baseline nutritional parameters, assessed as previously described, were also analyzed (Table&#x000a0;<xref rid="Tab2" ref-type="table">2</xref>). In particular, total Kcal intake, lipid intake, carbohydrate intake, protein intake, together with baseline alcohol intake were not significantly different in successes when compared to failures. Furthermore, the spontaneous difference in total kcal intake, lipid intake, carbohydrate intake, protein intake, alcohol intake recorded after 1-week self-monitoring was not significantly different in successes when compared to failures.<table-wrap id="Tab2"><label>Table&#x000a0;2</label><caption><p>Baseline nutritional parameters: differences between successes and failures</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left">Baseline intake</th><th align="left">Successes</th><th align="left">Failures</th><th align="left"> <italic>p</italic> </th></tr></thead><tbody><tr><td align="left">Kcal</td><td align="left">2,265.1&#x000a0;&#x000b1;&#x000a0;887.4</td><td align="left">2,032.7&#x000a0;&#x000b1;&#x000a0;856.1</td><td char="." align="char">0.121</td></tr><tr><td align="left">Proteins (g)</td><td align="left">88.3&#x000a0;&#x000b1;&#x000a0;33.8</td><td align="left">83.7&#x000a0;&#x000b1;&#x000a0;31.5</td><td char="." align="char">0.513</td></tr><tr><td align="left">Lipids (g)</td><td align="left">82.1 (68.3&#x02013;106.1)</td><td align="left">81.0 (65.0&#x02013;102)</td><td char="." align="char">0.784</td></tr><tr><td align="left">Saturated lipids (g)</td><td align="left">25.0&#x000a0;&#x000b1;&#x000a0;10.4</td><td align="left">22.2&#x000a0;&#x000b1;&#x000a0;12.1</td><td char="." align="char">0.228</td></tr><tr><td align="left">Carbohydrates (g)</td><td align="left">293.8&#x000a0;&#x000b1;&#x000a0;136.6</td><td align="left">253.4&#x000a0;&#x000b1;&#x000a0;118.2</td><td char="." align="char">0.067</td></tr><tr><td align="left">Alcohol (g)</td><td align="left">5.0 (1.2&#x02013;23.5)</td><td align="left">4.4 (1.0&#x02013;19.9)</td><td char="." align="char">0.511</td></tr></tbody></table><table-wrap-foot><p>Data are shown as media&#x000a0;&#x000b1;&#x000a0;SD, whereas categorical variables were reported as median and limits of confidence</p></table-wrap-foot></table-wrap> </p><p>Differences between baseline data of successes (losing at least 5&#x000a0;% of initial body weight at the intention-to-treat analysis) and failures (completers losing &#x0003c;5&#x000a0;% of initial body weight and drop-outs) are summarized in Table&#x000a0;<xref rid="Tab3" ref-type="table">3</xref>. A significant difference was observed only for diastolic blood pressure (DBP) (78&#x000a0;&#x000b1;&#x000a0;6 vs. 75&#x000a0;&#x000b1;&#x000a0;9&#x000a0;mmHg; <italic>p</italic>&#x000a0;=&#x000a0;0.033); free fat mass (FFM) (62.8&#x000a0;&#x000b1;&#x000a0;15.1 vs. 57.8&#x000a0;&#x000b1;&#x000a0;9.3&#x000a0;%; <italic>p</italic>&#x000a0;=&#x000a0;0.021); muscle mass (MM) (45.2&#x000a0;&#x000b1;&#x000a0;12.6 vs. 40.6&#x000a0;&#x000b1;&#x000a0;7.6&#x000a0;%; <italic>p</italic>&#x000a0;=&#x000a0;0.011); total body water (TBW) (46.3&#x000a0;&#x000b1;&#x000a0;11.1 vs. 42.5&#x000a0;&#x000b1;&#x000a0;6.7&#x000a0;%; <italic>p</italic>&#x000a0;=&#x000a0;0.016); HDL (44.9&#x000a0;&#x000b1;&#x000a0;11.9 vs. 50.5&#x000a0;&#x000b1;&#x000a0;14.8&#x000a0;mg/dl; <italic>p</italic>&#x000a0;=&#x000a0;0.020); ALT (29.6&#x000a0;&#x000b1;&#x000a0;30.8 vs. 20&#x000a0;&#x000b1;&#x000a0;5&#x000a0;mg/dl; <italic>p</italic>&#x000a0;=&#x000a0;0.017); AST (40.4&#x000a0;&#x000b1;&#x000a0;51.2 vs. 25.5&#x000a0;&#x000b1;&#x000a0;12.3&#x000a0;mg/dl; <italic>p</italic>&#x000a0;=&#x000a0;0.027); &#x003b3;GT (38.6&#x000a0;&#x000b1;&#x000a0;52.3 vs. 22.2&#x000a0;&#x000b1;&#x000a0;11.4&#x000a0;mg/dl; <italic>p</italic>&#x000a0;=&#x000a0;0.019).<table-wrap id="Tab3"><label>Table&#x000a0;3</label><caption><p>Differences between baseline data of successes (losing at least 5&#x000a0;% of initial body weight at the intention-to-treat analysis) and failures (completers losing &#x0003c;5&#x000a0;% of initial body weight and drop-outs)</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left"/><th align="left">Successes</th><th align="left">Failures</th><th align="left"> <italic>p</italic> </th></tr></thead><tbody><tr><td align="left">DBP (mmHg)</td><td char="&#x000b1;" align="char">78&#x000a0;&#x000b1;&#x000a0;6</td><td char="&#x000b1;" align="char">75&#x000a0;&#x000b1;&#x000a0;9</td><td char="&#x000b1;" align="char">0.033</td></tr><tr><td align="left">FFM (%)</td><td char="&#x000b1;" align="char">62.8&#x000a0;&#x000b1;&#x000a0;15.1</td><td char="&#x000b1;" align="char">57.8&#x000a0;&#x000b1;&#x000a0;9.3</td><td char="&#x000b1;" align="char">0.021</td></tr><tr><td align="left">MM (%)</td><td char="&#x000b1;" align="char">45.2&#x000a0;&#x000b1;&#x000a0;12.6</td><td char="&#x000b1;" align="char">40.6&#x000a0;&#x000b1;&#x000a0;7.6</td><td char="&#x000b1;" align="char">0.011</td></tr><tr><td align="left">TBW (%)</td><td char="&#x000b1;" align="char">46.3&#x000a0;&#x000b1;&#x000a0;11.1</td><td char="&#x000b1;" align="char">42.5&#x000a0;&#x000b1;&#x000a0;6.7&#x000a0;%</td><td char="&#x000b1;" align="char">0.021</td></tr><tr><td align="left">HDL (mg/dl)</td><td char="&#x000b1;" align="char">44.9&#x000a0;&#x000b1;&#x000a0;11.9</td><td char="&#x000b1;" align="char">50.5&#x000a0;&#x000b1;&#x000a0;14.8</td><td char="&#x000b1;" align="char">0.020</td></tr><tr><td align="left">ALT (mg/dl)</td><td char="&#x000b1;" align="char">29.6&#x000a0;&#x000b1;&#x000a0;30.8</td><td char="&#x000b1;" align="char">20&#x000a0;&#x000b1;&#x000a0;5</td><td char="&#x000b1;" align="char">0.017</td></tr><tr><td align="left">AST (mg/dl)</td><td char="&#x000b1;" align="char">40.4&#x000a0;&#x000b1;&#x000a0;51.2</td><td char="&#x000b1;" align="char">25.5&#x000a0;&#x000b1;&#x000a0;12.3</td><td char="&#x000b1;" align="char">0.027</td></tr><tr><td align="left">&#x003b3;GT (mg/dl)</td><td char="&#x000b1;" align="char">38.6&#x000a0;&#x000b1;&#x000a0;52.3</td><td char="&#x000b1;" align="char">22.2&#x000a0;&#x000b1;&#x000a0;11.4</td><td char="&#x000b1;" align="char">0.019</td></tr></tbody></table><table-wrap-foot><p>Data are shown as media&#x000a0;&#x000b1;&#x000a0;SD</p></table-wrap-foot></table-wrap> </p><p>None of these differences was retained at multivariate analysis with only one exception. In fact, after dividing into quartiles the not-normally distributed variables, it resulted that successes have AST values above median (3rd and 4th quartiles; <italic>&#x003c7;</italic> ²&#x000a0;=&#x000a0;0.003), even if a trend was observed for ALT (3rd and 4th quartiles; <italic>&#x003c7;</italic> ²&#x000a0;=&#x000a0;0.068) and &#x003b3;GT (3rd and 4th quartiles; <italic>&#x003c7;</italic> ²&#x000a0;=&#x000a0;0.074). At multivariate analysis (logistic regression), after adjusting for age and waist, the OR for AST was 3.34 (1.42&#x02013;7.85; <italic>p</italic>&#x000a0;=&#x000a0;0.006).</p></sec><sec id="Sec8" sec-type="discussion"><title>Discussion</title><p>The present study tries to identify new possible predictors of outcome in a non-pharmacologic lifestyle change-centered weight loss program. For this purpose, we analyzed a series of baseline blood parameters, which are usually studied in obese patients to assess the presence of eventual comorbidities. In particular, successes resulted to have GOT/AST values above median.</p><p>It is well known that obesity is frequently associated with a cluster of risk factors including dysglycemia, hypertension and dyslipidemia, in a few words defined by the concept of metabolic syndrome (MS). As a consequence, obese patients are often affected by non-alcoholic fatty liver disease (NAFLD). Impaired hepatic fatty acid (FA) turnover together with insulin resistance are key players in NAFLD pathogenesis [<xref ref-type="bibr" rid="CR21">21</xref>]. Most individuals are asymptomatic and are usually discovered incidentally because of abnormal liver function tests. Elevated liver biochemistry is found in 50&#x000a0;% of patients with simple steatosis. Change in liver function tests is considered as surrogate marker of liver injury and non-alcholic fatty liver disease (NAFLD). Previous studies have demonstrated that circulating concentration of liver function tests like &#x003b3;-glutamyltransferase (&#x003b3;GT), alanine aminotransferase (ALT) and aspartate aminotransferase (AST) is increased in individuals with insulin resistance and the metabolic syndrome [<xref ref-type="bibr" rid="CR22">22</xref>]. In addition, these components of liver function tests have been shown to be positively associated with the risk of future type 2 diabetes [<xref ref-type="bibr" rid="CR23">23</xref>, <xref ref-type="bibr" rid="CR24">24</xref>]. In particular, available data indicate moderate associations of ALT and &#x003b3;GT with risk of type 2 diabetes events, and no evidence for an increased risk with AST [<xref ref-type="bibr" rid="CR25">25</xref>, <xref ref-type="bibr" rid="CR26">26</xref>].</p><p>Moreover, recent data [<xref ref-type="bibr" rid="CR27">27</xref>] speculate that abnormal levels of ALT and AST are associated with a deregulation of normal amino acid metabolism in the liver, including aromatic amino acid, and then special compounds such as glutamate are released into the general circulation. This hypothesis attempts to illustrate the critical role of the &#x0201c;liver metabolism&#x0201d; in the pathogenesis of the MS and postulates that before the liver becomes fatty, abnormal levels of liver enzymes might reflect high levels of hepatic transamination of amino acids in the organ. Anyway, elevated transaminase levels could represent an early at-risk situation of pre-NAFLD.</p><p>In our patients, baseline liver enzyme levels (AST in particular), but not baseline quantitative (total Kcal) and qualitative (lipids, carbohydrates, proteins, alcohol) dietary intake, were significantly different in successes versus failures and could therefore represent a predictor of success. Successes have AST values above median. A possible explanation could be that AST is raised in acute liver damage, but is also present in red blood cells, and cardiac and skeletal muscle and is therefore not specific to the liver. For instance, in athletes, the interpretation of serum aminotransferases concentrations should consider the release of AST from muscle and of ALT mainly from the liver [<xref ref-type="bibr" rid="CR28">28</xref>]. Therefore, we could speculate that the predictive value of AST but not of ALT or &#x003b3;GT in our population could be related to the significantly different amount of baseline muscle mass in the two populations. This could be consistent with our previous findings that successes have better baseline grade of fitness in terms of initial muscle mass when compared to failures [<xref ref-type="bibr" rid="CR15">15</xref>].</p><p>Both genetic and environmental factors have been proposed to be involved in the etiology of NAFLD. Thus, nutrition is reasonably considered to be a potential environmental factor affecting the risk for this disease. Although there is consistent evidence that overweight due to energy overconsumption increases the risk for and the prevalence of fatty liver, the role of diet composition, in terms of macro- or micronutrients, in the pathogenesis of the disease remains controversial [<xref ref-type="bibr" rid="CR29">29</xref>]. Dietary habits shown by patients before starting the lifestyle change program have not been demonstrated to be significantly different between those losing at least 5&#x000a0;% of initial body weight and failures. The same result was obtained comparing the basal and 1&#x000a0;week after quantitative and qualitative dietary intake, spontaneously modified by patients without the dietician intervention. Anyway, we should point out that both baseline total Kcal and carbohydrate intake, even if not reaching the statistical significance, show different values when comparing successes and failures. In particular, failures report lower total Kcal intake and lower carbohydrate intake before entering the program. It should be taken into consideration that data regarding qualitative and quantitative dietary intake were collected based on information self-reported by patients. The tools used by the operators, such as a dedicated software and a photo atlas representing different portion sizes, were aimed at reducing the bias, but the problem related to an altered perception from the patient&#x02019;s point of view could nevertheless be still present.</p><p>Moreover, despite worldwide guidelines and recommendations, research examining diet composition for the management of obesity remains controversial. Previous studies have found that low-fat diets promote short-term weight loss [<xref ref-type="bibr" rid="CR30">30</xref>]; however, some studies suggest that low-carbohydrate, high-protein, and high-fat diets may also result in substantial weight loss [<xref ref-type="bibr" rid="CR31">31</xref>]. We can therefore conclude that a substantial body of literature exists suggesting that weight loss can be achieved by varying the macronutrient distribution and composition of dietary factors, even if evidence for type of diet on long-term weight maintenance remains debatable [<xref ref-type="bibr" rid="CR16">16</xref>, <xref ref-type="bibr" rid="CR32">32</xref>]. On the other hand, few data are present regarding quality of eating patterns as predictors of weight loss. Hart et al. [<xref ref-type="bibr" rid="CR33">33</xref>] demonstrated that early changes in eating habits, but not baseline eating behaviors may promote greater BMI reductions in a cohort of adolescent patients. These findings are consistent with our data, even if referred to a completely different population.</p><p>Based on our data, we can at the moment conclude that neither usual dietary habits nor self-management operated by patients based on personal beliefs could represent a good predictor of success for this kind of programs. Finally, these results confirm once again that baseline grade of fitness (i.e., initial muscle mass) is a better predictor of success, rather than dietary habits, when starting a lifestyle modification program, as already demonstrated by our group [<xref ref-type="bibr" rid="CR15">15</xref>].</p><p>Moreover, we can conclude that AST could represent a usually available biomarker that could be used as a predictor of weight loss in obese patients starting a lifestyle change program. Nevertheless, this should be considered as a pilot study which includes a reasonable number of patients and therefore the efficacy of present data, which aim to identify a weight loss predictor of outcome, will need to be verified through specifically designed longer-term randomized clinical trials enrolling a larger number of patients.</p></sec>

Eating behavior and perception of body shape in Japanese university students

Could not extract abstract

<contrib contrib-type="author"><name><surname>Ohara</surname><given-names>Kumiko</given-names></name><xref ref-type="aff" rid="Aff1">1</xref><xref ref-type="aff" rid="Aff2">2</xref></contrib><contrib contrib-type="author"><name><surname>Kato</surname><given-names>Yoshiko</given-names></name><xref ref-type="aff" rid="Aff1">1</xref></contrib><contrib contrib-type="author"><name><surname>Mase</surname><given-names>Tomoki</given-names></name><xref ref-type="aff" rid="Aff1">1</xref><xref ref-type="aff" rid="Aff3">3</xref></contrib><contrib contrib-type="author"><name><surname>Kouda</surname><given-names>Katsuyasu</given-names></name><xref ref-type="aff" rid="Aff4">4</xref></contrib><contrib contrib-type="author"><name><surname>Miyawaki</surname><given-names>Chiemi</given-names></name><xref ref-type="aff" rid="Aff1">1</xref><xref ref-type="aff" rid="Aff5">5</xref></contrib><contrib contrib-type="author"><name><surname>Fujita</surname><given-names>Yuki</given-names></name><xref ref-type="aff" rid="Aff4">4</xref></contrib><contrib contrib-type="author"><name><surname>Okita</surname><given-names>Yoshimitsu</given-names></name><xref ref-type="aff" rid="Aff6">6</xref></contrib><contrib contrib-type="author" corresp="yes"><name><surname>Nakamura</surname><given-names>Harunobu</given-names></name><address><phone>+81-78-803-7740</phone><email>hal@kobe-u.ac.jp</email></address><xref ref-type="aff" rid="Aff1">1</xref></contrib><aff id="Aff1"><label>1</label>Graduate School of Human Development and Environment, Kobe University, 3-11 Tsurukabuto, Nada, Kobe, Hyogo 657-8501 Japan </aff><aff id="Aff2"><label>2</label>Research Fellow of Japan Society for the Promotion of Science, 5-3-1 Kojimachi, Chiyoda, Tokyo, 102-0083 Japan </aff><aff id="Aff3"><label>3</label>Department of Childhood Education, Kyoto Seibo College, 1 Fukakusa Taya-cho, Fushimi, Kyoto, 612-0878 Japan </aff><aff id="Aff4"><label>4</label>Department of Public Health, Kinki University Faculty of Medicine, 377-2 Oono-Higashi, Osaka-Sayama, Osaka, 589-8511 Japan </aff><aff id="Aff5"><label>5</label>Heian Jogakuin (St.Agnes&#x02019;) College, 5-81-1 Nampeidai, Takatsuki, Osaka, 569-1092 Japan </aff><aff id="Aff6"><label>6</label>Graduate School of Science and Technology, Shizuoka University, 3-5-1 Johoku, Naka, Hamamatsu, Shizuoka, 432-8011 Japan </aff>

Eating and Weight Disorders

<sec id="Sec1"><title>Introduction</title><p>It is well known that many girls and women around the world feel driven to remain or become thin. Studies of large populations found that the mean body mass index (BMI) of many young females in European countries and the united States is &#x0003e;25 (overweight) [<xref ref-type="bibr" rid="CR1">1</xref>], and their self-reported ideal body shape was smaller than their current body shape [<xref ref-type="bibr" rid="CR2">2</xref>]. Young Japanese females have an average BMI of about 20 [<xref ref-type="bibr" rid="CR3">3</xref>]. However, many of them attempt to achieve a thinner physique as their ideal body shape [<xref ref-type="bibr" rid="CR4">4</xref>].</p><p>In previous studies, eating behavior was found to be associated with body image, and with a drive for thinness. For instance, high score on the Eating Attitude Test-26 (EAT-26) was positively associated with a thin ideal physique or drive for thinness [<xref ref-type="bibr" rid="CR5">5</xref>, <xref ref-type="bibr" rid="CR6">6</xref>]. The EAT-26 was developed by Garner [<xref ref-type="bibr" rid="CR7">7</xref>] for the screening of patients with eating disorders, in which a score of 20 is considered the clinical cut-off.</p><p>A different measure, the Dutch Eating Behaviour Questionnaire (DEBQ), was developed by van Strien to assess eating behavior, including that of normal subjects, in terms of three eating patterns, such as restrained eating, external eating, and emotional eating [<xref ref-type="bibr" rid="CR8">8</xref>]. However, there are few evidences on the relationship between DEBQ and body shape. Restrained eating of DEBQ was positively associated with the difference between clinical BMI and self-reported BMI [<xref ref-type="bibr" rid="CR9">9</xref>], positively associated with perception of body weight, which means whether subjects thought they were overweight or underweight [<xref ref-type="bibr" rid="CR10">10</xref>], and negatively associated with ideal body image [<xref ref-type="bibr" rid="CR11">11</xref>]. The above reports have referred to only restrained eating of DEBQ. Thus, little is known about an association between DEBQ and body shape from the viewpoint of three eating pattern of DEBQ. In addition, recent studies have also found that many males are preoccupied with their own body shapes [<xref ref-type="bibr" rid="CR12">12</xref>&#x02013;<xref ref-type="bibr" rid="CR14">14</xref>], but there are few studies about males&#x02019; perception of body shape, or studies that compare the perception of body shape between males and females. Moreover, in males, the drive for muscularity which is a desire to enhance one&#x02019;s musculature has been reported [<xref ref-type="bibr" rid="CR15">15</xref>&#x02013;<xref ref-type="bibr" rid="CR18">18</xref>]. Therefore, it is hypothesized that there is a difference in a direction of ideal body shape between males and females or between normal weight and underweight. It is also hypothesized that there is a difference in a relation of DEBQ and body image between males and females or between normal weight and underweight.</p><p>In the present study, to address the shortage in the literature, we investigated the relationship between eating behavior as measured by the DEBQ and the perception of body shape between the normal weight and underweight in each gender.</p></sec><sec id="Sec2"><title>Methods</title><sec id="Sec3"><title>Participants</title><p>The survey was conducted using an anonymous, self-administered questionnaire during university classes in 2011&#x02013;2012. The participants were 618 university students who attended the classes in the liberal arts. Participants were provided no remuneration. The questionnaires were delivered to the all attendance (618 students), and all of them were collected after completion of questionnaire. Out of 618 questionnaires, 548 questionnaires showed valid responses. Thus, the response rate, which was calculated by dividing the valid responses by delivered questionnaires, was 88.7&#x000a0;% (<italic>n</italic>&#x000a0;=&#x000a0;548, 252 males and 296 females, 19.2&#x000a0;&#x000b1;&#x000a0;0.9&#x000a0;years). We classified the participants into three groups according to the World Health Organization (WHO) criteria: overweight (BMI&#x000a0;&#x02265;&#x000a0;25.0&#x000a0;kg/m²), normal weight (18.5&#x000a0;&#x02264;&#x000a0;BMI&#x000a0;&#x0003c;&#x000a0;25.0&#x000a0;kg/m²), and underweight (BMI&#x000a0;&#x0003c;&#x000a0;18.5&#x000a0;kg/m²). The participants included 94 underweight students (33 males and 61 females), 436 normal-weight students (210 males and 226 females), and 18 overweight students (9 males and 9 females). The overweight individuals were excluded from the present analysis because of small number comparing with the underweight or normal-weight students.</p><p>All participants gave informed consent to participate, and the study was approved by the Human Ethics Committee of the Graduate School of Human Development and Environment, Kobe University.</p></sec><sec id="Sec4"><title>Measures</title><p>The questionnaire identified the participants&#x02019; physical status, perception of body shape, and eating behaviors. The physical status questions included four items: age, height, weight, &#x0201c;ideal height,&#x0201d; and &#x0201c;ideal weight&#x0201d; as answered by the participant. Concerning heights and weights, we asked subjects to write down their self-reported heights and weights according to previous reports [<xref ref-type="bibr" rid="CR19">19</xref>&#x02013;<xref ref-type="bibr" rid="CR21">21</xref>]. Each participant&#x02019;s BMI (kg/m²) was calculated by dividing his or her weight in kilograms by the square of the height in meters. Each participant&#x02019;s &#x0201c;ideal BMI&#x0201d; was calculated in the same way as the BMI using the ideal weight and ideal height he or she provided. A discrepancy of height, weight, and BMI was calculated by &#x0201c;ideal value&#x0201d; minus &#x0201c;current value&#x0201d;.</p><p>The participants&#x02019; perception of body shape (BSh) was assessed by a 27-item interval scale using four gender-specific BMI-based silhouettes of which validity and reliability have been evaluated in the previous study [<xref ref-type="bibr" rid="CR22">22</xref>]. The participants were asked to select which single item of the 27-item interval scale most closely represented their actual and ideal body size. A discrepancy of BSh was also calculated by &#x0201c;ideal BSh&#x0201d; minus &#x0201c;current BSh&#x0201d;.</p><p>Eating behavior was assessed by the Japanese version of the DEBQ of which validity and reliability have been evaluated in the previous study [<xref ref-type="bibr" rid="CR23">23</xref>]. The DEBQ is a 33-item self-rated questionnaire and is divided into three subscales: restrained eating (10 items), emotional eating (13 items), and external eating (10 items). Restrained eating means paradoxically dietary restraint (food intake is initially reduced to lose or maintain body weight, but followed by increased consumption and binge eating). Emotional eating means eating in response to negative emotions. External eating means eating in response to the sight or smell of food [<xref ref-type="bibr" rid="CR24">24</xref>]. The participants were asked to rate each question from 1 for &#x02018;never&#x02019; to up to 5 for &#x02018;very often.&#x02019; Responses to each question were added together in each subscale, and then divided by the number of questions included in each subscale to produce a score between 1 and 5.</p></sec><sec id="Sec5"><title>Statistical analysis</title><p>Student&#x02019;s <italic>t</italic> test was used to evaluate the differences in physical status or DEBQ score between the normal-weight and underweight students in each gender. Pearson&#x02019;s correlation coefficients were calculated on BMI and each DEBQ scores. A multiple linear regression analysis was used to investigate the association between DEBQ scores and discrepancy of physical index, adjusting for height. A two-way repeated measures analysis of variance (ANOVA) was used to investigate the effects of &#x02018;current&#x02019; and &#x02018;ideal&#x02019; physical status, the effects of gender, and the interaction effects between &#x02018;current&#x02019; and &#x02018;ideal&#x02019; physical status and gender on the parameters of physical status. For a post hoc analysis, we used the Bonferroni test. The statistical level for significance was set at 0.05. All statistical analyses were performed by SPSS^&#x000ae; 19.0&#x000a0;J for Windows (IBM, Tokyo).</p></sec></sec><sec id="Sec6"><title>Results</title><p>After excluding the 18 overweight participants, the means and standard deviations of height, weight, and BMI for all participants were 171.6&#x000a0;&#x000b1;&#x000a0;5.7&#x000a0;cm, 61.1&#x000a0;&#x000b1;&#x000a0;8.5&#x000a0;kg, and 20.7&#x000a0;&#x000b1;&#x000a0;2.4&#x000a0;kg/m² in the males, and 157.7&#x000a0;&#x000b1;&#x000a0;5.5&#x000a0;cm, 50.3&#x000a0;&#x000b1;&#x000a0;6.5&#x000a0;kg, and 20.2&#x000a0;&#x000b1;&#x000a0;2.5&#x000a0;kg/m² in the females, respectively. According to the classification of BSh by BMI, 17.7&#x000a0;% (<italic>n</italic>&#x000a0;=&#x000a0;94) of the participants were underweight, while 82.3&#x000a0;% (<italic>n</italic>&#x000a0;=&#x000a0;436) were normal weight. In the males, 13.1&#x000a0;% (<italic>n</italic>&#x000a0;=&#x000a0;33) of the participants were underweight and 83.3&#x000a0;% (<italic>n</italic>&#x000a0;=&#x000a0;210) were normal weight. In the females, 20.6&#x000a0;% (<italic>n</italic>&#x000a0;=&#x000a0;61) of the participants were underweight and 76.4&#x000a0;% (<italic>n</italic>&#x000a0;=&#x000a0;226) were normal weight.</p><p>As shown in Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>, the weight, BMI, perception of BSh, ideal weight, and ideal BMI were all significantly lower in the underweight participants than in the normal-weight participants, in both the males and females (ideal weight in males, <italic>p</italic>&#x000a0;=&#x000a0;0.001; other index, <italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.001).<table-wrap id="Tab1"><label>Table&#x000a0;1</label><caption><p>Physical status of underweight and normal-weight subjects</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="2"/><th align="left" colspan="2">Male</th><th align="left" colspan="2">Female</th></tr><tr><th align="left">Underweight (<italic>n</italic>&#x000a0;=&#x000a0;33)</th><th align="left">Normal weight (<italic>n</italic>&#x000a0;=&#x000a0;210)</th><th align="left">Underweight (<italic>n</italic>&#x000a0;=&#x000a0;61)</th><th align="left">Normal weight (<italic>n</italic>&#x000a0;=&#x000a0;226)</th></tr></thead><tbody><tr><td align="left">Height (cm)</td><td char="&#x000b1;" align="char">171.8&#x000a0;&#x000b1;&#x000a0;5.7^b </td><td char="&#x000b1;" align="char">171.4&#x000a0;&#x000b1;&#x000a0;5.6^c </td><td char="&#x000b1;" align="char">158.5&#x000a0;&#x000b1;&#x000a0;5.6</td><td char="&#x000b1;" align="char">157.7&#x000a0;&#x000b1;&#x000a0;5.4</td></tr><tr><td align="left">Weight (kg)</td><td char="&#x000b1;" align="char">51.9&#x000a0;&#x000b1;&#x000a0;4.5^ab </td><td char="&#x000b1;" align="char">61.6&#x000a0;&#x000b1;&#x000a0;6.3^c </td><td char="&#x000b1;" align="char">44.2&#x000a0;&#x000b1;&#x000a0;3.7^a </td><td char="&#x000b1;" align="char">51.3&#x000a0;&#x000b1;&#x000a0;4.9</td></tr><tr><td align="left">BMI (kg/m²)</td><td char="&#x000b1;" align="char">17.6&#x000a0;&#x000b1;&#x000a0;0.9^a </td><td char="&#x000b1;" align="char">20.9&#x000a0;&#x000b1;&#x000a0;1.6</td><td char="&#x000b1;" align="char">17.6&#x000a0;&#x000b1;&#x000a0;0.9^a </td><td char="&#x000b1;" align="char">20.6&#x000a0;&#x000b1;&#x000a0;1.5</td></tr><tr><td align="left">Perception of BSh</td><td char="&#x000b1;" align="char">7.1&#x000a0;&#x000b1;&#x000a0;2.4^a </td><td char="&#x000b1;" align="char">12.4&#x000a0;&#x000b1;&#x000a0;3.8</td><td char="&#x000b1;" align="char">7.5&#x000a0;&#x000b1;&#x000a0;2.8^a </td><td char="&#x000b1;" align="char">12.3&#x000a0;&#x000b1;&#x000a0;3.4</td></tr><tr><td align="left">Ideal height(cm)</td><td char="&#x000b1;" align="char">176.1&#x000a0;&#x000b1;&#x000a0;5.2^b </td><td char="&#x000b1;" align="char">176.0&#x000a0;&#x000b1;&#x000a0;5.1^c </td><td char="&#x000b1;" align="char">160.1&#x000a0;&#x000b1;&#x000a0;6.1</td><td char="&#x000b1;" align="char">159.4&#x000a0;&#x000b1;&#x000a0;4.5</td></tr><tr><td align="left">Ideal weight (kg)</td><td char="&#x000b1;" align="char">60.0&#x000a0;&#x000b1;&#x000a0;6.1^ab </td><td char="&#x000b1;" align="char">64.9&#x000a0;&#x000b1;&#x000a0;7.4^c </td><td char="&#x000b1;" align="char">44.9&#x000a0;&#x000b1;&#x000a0;6.5^a </td><td char="&#x000b1;" align="char">47.3&#x000a0;&#x000b1;&#x000a0;4.0</td></tr><tr><td align="left">Ideal BMI (kg/m²)</td><td char="&#x000b1;" align="char">19.4&#x000a0;&#x000b1;&#x000a0;1.5^ab </td><td char="&#x000b1;" align="char">20.9&#x000a0;&#x000b1;&#x000a0;1.8^c </td><td char="&#x000b1;" align="char">17.5&#x000a0;&#x000b1;&#x000a0;1.3^a </td><td char="&#x000b1;" align="char">18.6&#x000a0;&#x000b1;&#x000a0;1.2</td></tr><tr><td align="left">Ideal BSh</td><td char="&#x000b1;" align="char">11.2&#x000a0;&#x000b1;&#x000a0;3.7^b </td><td char="&#x000b1;" align="char">12.1&#x000a0;&#x000b1;&#x000a0;3.0^c </td><td char="&#x000b1;" align="char">6.5&#x000a0;&#x000b1;&#x000a0;2.8^a </td><td char="&#x000b1;" align="char">7.5&#x000a0;&#x000b1;&#x000a0;2.2</td></tr><tr><td align="left">Discrepancy of height (cm)</td><td char="&#x000b1;" align="char">4.2&#x000a0;&#x000b1;&#x000a0;4.4^b </td><td char="&#x000b1;" align="char">4.6&#x000a0;&#x000b1;&#x000a0;4.7^c </td><td char="&#x000b1;" align="char">1.6&#x000a0;&#x000b1;&#x000a0;4.7</td><td char="&#x000b1;" align="char">1.8&#x000a0;&#x000b1;&#x000a0;3.5</td></tr><tr><td align="left">Discrepancy of weight (kg)</td><td char="&#x000b1;" align="char">8.1&#x000a0;&#x000b1;&#x000a0;5.7^ab </td><td char="&#x000b1;" align="char">3.3&#x000a0;&#x000b1;&#x000a0;5.9^c </td><td char="&#x000b1;" align="char">0.7&#x000a0;&#x000b1;&#x000a0;7.1^a </td><td char="&#x000b1;" align="char">&#x02212;3.9&#x000a0;&#x000b1;&#x000a0;3.4</td></tr><tr><td align="left">Discrepancy of BMI (kg/m²)</td><td char="&#x000b1;" align="char">1.8&#x000a0;&#x000b1;&#x000a0;1.4^ab </td><td char="&#x000b1;" align="char">&#x02212;0.03&#x000a0;&#x000b1;&#x000a0;3.8^c </td><td char="&#x000b1;" align="char">&#x02212;0.1&#x000a0;&#x000b1;&#x000a0;1.7^a </td><td char="&#x000b1;" align="char">&#x02212;2.0&#x000a0;&#x000b1;&#x000a0;1.3</td></tr><tr><td align="left">Discrepancy of BSh</td><td char="&#x000b1;" align="char">4.1&#x000a0;&#x000b1;&#x000a0;3.5^ab </td><td char="&#x000b1;" align="char">&#x02212;0.3&#x000a0;&#x000b1;&#x000a0;3.8^c </td><td char="&#x000b1;" align="char">&#x02212;1.0&#x000a0;&#x000b1;&#x000a0;3.9^a </td><td char="&#x000b1;" align="char">&#x02212;4.8&#x000a0;&#x000b1;&#x000a0;3.1</td></tr></tbody></table><table-wrap-foot><p>Discrepancy: ideal value minus current value. Values are means&#x000a0;&#x000b1;&#x000a0;standard deviations</p><p> <italic>BMI</italic> body mass index, <italic>BSh</italic> body shape</p><p> ^aSignificantly different from normal weight in each gender (Student&#x02019;s <italic>t</italic> test with Bonferroni correction)</p><p> ^bSignificantly different from female underweight (Student&#x02019;s <italic>t</italic> test with Bonferroni correction)</p><p> ^cSignificantly different from female normal weight (Student&#x02019;s <italic>t</italic> test with Bonferroni correction)</p></table-wrap-foot></table-wrap> </p><p>Among the females, the perception of ideal BSh was significantly lower in the underweight participants than in the normal-weight participants (<italic>p</italic>&#x000a0;=&#x000a0;0.004). In the group of female normal-weight participants, discrepancies of weight, BMI, and BSh were negative, and significantly lower than those in female underweight participants (all three indexes, <italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.001). In male underweight participants, discrepancies of weight, BMI, and BSh were positive, and significantly higher than those in male normal-weight participants (all three indexes, <italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.001). Among all of the underweight participants, the males&#x02019; values for height, weight, ideal height, ideal weight, ideal BMI, ideal BSh, discrepancy of height, discrepancy of weight, discrepancy of BMI, and discrepancy of BSh were significantly larger as compared to the females (<italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.05). Similarly, among all of the normal-weight participants, the males&#x02019; values for height, weight, ideal height, ideal weight, ideal BMI, ideal BSh, discrepancy of height, discrepancy of weight, discrepancy of BMI, and discrepancy of BSh were significantly larger as compared to the females (<italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.05).</p><p>The DEBQ scores for restrained, emotional, and external eating were significantly higher in the females than in the males (all three scores: <italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.001). In addition, as shown in Table&#x000a0;<xref rid="Tab2" ref-type="table">2</xref>, the restrained eating values were lower in the underweight participants than in the normal-weight participants in both males and females (males, <italic>p</italic>&#x000a0;=&#x000a0;0.003; females, <italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.001). Among all of the underweight participants, the females had significantly higher restrained eating and external eating scores as compared to the males (<italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.05). Among all of the normal-weight participants, the females had significantly higher restrained, emotional, and external eating scores as compared to the males (<italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.05). Cronbach&#x02019;s alpha was 0.91 for restrained eating, 0.95 for emotional eating, and 0.79 for external eating.<table-wrap id="Tab2"><label>Table&#x000a0;2</label><caption><p>DEBQ scores of underweight and normal-weight subjects</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="2"/><th align="left" colspan="2">Male</th><th align="left" colspan="2">Female</th></tr><tr><th align="left">Underweight (<italic>n</italic>&#x000a0;=&#x000a0;33)</th><th align="left">Normal weight (<italic>n</italic>&#x000a0;=&#x000a0;210)</th><th align="left">Underweight (<italic>n</italic>&#x000a0;=&#x000a0;61)</th><th align="left">Normal weight (<italic>n</italic>&#x000a0;=&#x000a0;226)</th></tr></thead><tbody><tr><td align="left">Restrained</td><td align="left">1.8&#x000a0;&#x000b1;&#x000a0;0.6^ab </td><td align="left">2.2&#x000a0;&#x000b1;&#x000a0;0.8^c </td><td align="left">2.6&#x000a0;&#x000b1;&#x000a0;0.9^a </td><td align="left">3.0&#x000a0;&#x000b1;&#x000a0;0.8</td></tr><tr><td align="left">Emotional</td><td align="left">1.9&#x000a0;&#x000b1;&#x000a0;0.9</td><td align="left">1.8&#x000a0;&#x000b1;&#x000a0;0.6^c </td><td align="left">2.2&#x000a0;&#x000b1;&#x000a0;0.9</td><td align="left">2.3&#x000a0;&#x000b1;&#x000a0;1.0</td></tr><tr><td align="left">External</td><td align="left">3.0&#x000a0;&#x000b1;&#x000a0;0.7^b </td><td align="left">3.1&#x000a0;&#x000b1;&#x000a0;0.7^c </td><td align="left">3.4&#x000a0;&#x000b1;&#x000a0;0.7</td><td align="left">3.4&#x000a0;&#x000b1;&#x000a0;0.7</td></tr></tbody></table><table-wrap-foot><p>Values are means&#x000a0;&#x000b1;&#x000a0;standard deviations</p><p> ^aSignificantly different from normal weight in each gender (Student&#x02019;s <italic>t</italic> test with Bonferroni correction)</p><p> ^bSignificantly different from female underweight (Student&#x02019;s <italic>t</italic> test with Bonferroni correction)</p><p> ^cSignificantly different from female normal weight (Student&#x02019;s <italic>t</italic> test with Bonferroni correction)</p></table-wrap-foot></table-wrap> </p><p>Discrepancies in height, weight, BMI, and BSh between the participants&#x02019; current data and their ideal values are shown in Figs.&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>, <xref rid="Fig2" ref-type="fig">2</xref>, <xref rid="Fig3" ref-type="fig">3</xref>, <xref rid="Fig4" ref-type="fig">4</xref>. The two-way repeated measures ANOVA showed that there was a significant interaction effect between gender and height discrepancy (<italic>F</italic>&#x000a0;=&#x000a0;61.7, <italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.001). Both gender and height discrepancy had a significant main effect on height. After a post hoc test, both height and ideal height in the males were significantly higher than in the females. The ideal heights of both the males and females were significantly higher than their current heights.<fig id="Fig1"><label>Fig.&#x000a0;1</label><caption><p>Comparison of current and ideal height. Male current height and ideal height are shown by <italic>open circles</italic>, and the female current height and ideal height are shown by <italic>closed circles</italic>. There was also a significant interaction effect between current-ideal and gender on height (<italic>F</italic>&#x000a0;=&#x000a0;61.7, <italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.001). There were also significant main effects of current-ideal and gender on height <italic>single asterisk</italic> significant difference (<italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.05) between males and females. <italic>Dagger</italic> significant difference (<italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.05) between current height and ideal height</p></caption><graphic xlink:href="40519_2014_130_Fig1_HTML" id="MO1"/></fig> <fig id="Fig2"><label>Fig.&#x000a0;2</label><caption><p>Comparison of current and ideal body weight. Male current body weight and ideal body weight are shown by <italic>open circles</italic>, and the female current body weight and ideal body weight are shown by <italic>closed circles</italic>. There was also a significant interaction effect between current-ideal and gender on body weight (<italic>F</italic>&#x000a0;=&#x000a0;214.7, <italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.001). There were also significant main effects of current-ideal and gender on body weight. <italic>Single asterisk</italic> significant difference (<italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.05) between males and females. <italic>Dagger</italic> significant difference (<italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.05) between current body weight and ideal body weight</p></caption><graphic xlink:href="40519_2014_130_Fig2_HTML" id="MO2"/></fig> <fig id="Fig3"><label>Fig.&#x000a0;3</label><caption><p>Comparison of current and ideal body mass index. Male current body mass index and ideal body mass index are shown by <italic>open circles</italic>, and the female current body mass index and ideal body mass index are shown by <italic>closed circles</italic>. There was also a significant interaction effect between current-ideal and gender on body mass index (<italic>F</italic>&#x000a0;=&#x000a0;158.5, <italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.001). There were also significant main effects of current-ideal and gender on body mass index. <italic>Single asterisk</italic> significant difference (<italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.05) between males and females. <italic>Dagger</italic> significant difference (<italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.05) between current body mass index and ideal body mass index</p></caption><graphic xlink:href="40519_2014_130_Fig3_HTML" id="MO3"/></fig> <fig id="Fig4"><label>Fig.&#x000a0;4</label><caption><p>Comparison of current and ideal body shape. Male current body shape and ideal body shape are shown by <italic>open circles</italic>, and the female current body shape and ideal body shape are shown by <italic>closed circles</italic>. There was also a significant interaction effect between current-ideal and gender on body shape (<italic>F</italic>&#x000a0;=&#x000a0;163.2, <italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.001). There were also significant main effects of current-ideal and gender on body shape. <italic>Single asterisk</italic> significant difference (<italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.05) between males and females. <italic>Dagger</italic> significant difference (<italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.05) between current body shape and ideal body shape</p></caption><graphic xlink:href="40519_2014_130_Fig4_HTML" id="MO4"/></fig> </p><p>The two-way repeated measures ANOVA also revealed a significant interaction effect between gender and weight discrepancy (<italic>F</italic>&#x000a0;=&#x000a0;214.7, <italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.001). Both gender and weight discrepancy had a significant main effect on weight. After a post hoc test, weight and ideal weight in males were significantly larger than in the females. The ideal weight values were significantly larger than the current weight values in the males, whereas in the females, the ideal weight values were significantly smaller than the current weight values.</p><p>A significant interaction effect between gender and BMI discrepancy (<italic>F</italic>&#x000a0;=&#x000a0;158.5, <italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.001) was also observed. Both gender and BMI discrepancy had a significant main effect on BMI. After a post hoc test, the current BMI and ideal BMI were significantly larger in the males than in the females. In the males, the ideal BMI values were significantly larger than the BMI values, whereas in the females the ideal BMI values were significantly smaller than the BMI values.</p><p>There was also a significant interaction effect between gender and BSh discrepancy (<italic>F</italic>&#x000a0;=&#x000a0;163.2, <italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.001). Both gender and BSh discrepancy had a significant main effect on BSh. After a post hoc test, the ideal BSh values reported by the males were significantly larger than those reported by the females. In the females, the ideal BSh was significantly smaller than body shape.</p><p>BMI was significantly positively correlated with the score of restrained and external eating in males (restrained eating: <italic>r</italic>&#x000a0;=&#x000a0;0.213, <italic>p</italic>&#x000a0;=&#x000a0;0.001; external eating: <italic>r</italic>&#x000a0;=&#x000a0;0.183, <italic>p</italic>&#x000a0;=&#x000a0;0.004). In females, BMI was significantly positively correlated with the score of restrained and emotional eating (restrained eating: <italic>r</italic>&#x000a0;=&#x000a0;0.217, <italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.001; emotional eating: <italic>r</italic>&#x000a0;=&#x000a0;0.148, <italic>p</italic>&#x000a0;=&#x000a0;0.012).</p><p>Table&#x000a0;<xref rid="Tab3" ref-type="table">3</xref> shows the results of our multiple linear regression analysis between the DEBQ scores and the discrepancies of weight, BMI and perception of BSh, adjusting for height. In the males, restrained eating was significantly negatively associated with the discrepancy of weight (<italic>&#x003b2;</italic>&#x000a0;=&#x000a0;&#x02212;&#x000a0;0.185, <italic>p</italic>&#x000a0;=&#x000a0;0.005), the discrepancy of BMI (<italic>&#x003b2;</italic>&#x000a0;=&#x000a0;&#x02212;&#x000a0;0.363, <italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.001), and the discrepancy of BSh (<italic>&#x003b2;</italic>&#x000a0;=&#x000a0;&#x02212;&#x000a0;0.455, <italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.001). External eating was negatively associated with the discrepancy of BMI (<italic>&#x003b2;</italic>&#x000a0;=&#x000a0;&#x02212;&#x000a0;0.124, <italic>p</italic>&#x000a0;=&#x000a0;0.049).<table-wrap id="Tab3"><label>Table&#x000a0;3</label><caption><p>DEBQ scores and discrepancy of physical index</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="2"/><th align="left" colspan="2">Discrepancy of weight</th><th align="left" colspan="2">Discrepancy of BMI</th><th align="left" colspan="2">Discrepancy of body shape</th></tr><tr><th align="left"> <italic>&#x003b2;</italic> </th><th align="left"> <italic>p</italic> value</th><th align="left"> <italic>&#x003b2;</italic> </th><th align="left"> <italic>p</italic> value</th><th align="left"> <italic>&#x003b2;</italic> </th><th align="left"> <italic>p</italic> value</th></tr></thead><tbody><tr><td align="left" colspan="7">Male</td></tr><tr><td align="left">&#x000a0;Restrained</td><td char="." align="char">&#x02212;0.185</td><td char="." align="char">0.005*</td><td char="." align="char">&#x02212;0.363</td><td char="." align="char">&#x0003c;0.001*</td><td char="." align="char">&#x02212;0.455</td><td char="." align="char">&#x0003c;0.001*</td></tr><tr><td align="left">&#x000a0;Emotional</td><td char="." align="char">0.049</td><td char="." align="char">0.470</td><td char="." align="char">0.092</td><td char="." align="char">0.155</td><td char="." align="char">0.071</td><td char="." align="char">0.255</td></tr><tr><td align="left">&#x000a0;External</td><td char="." align="char">&#x02212;0.079</td><td char="." align="char">0.233</td><td char="." align="char">&#x02212;0.124</td><td char="." align="char">0.049*</td><td char="." align="char">&#x02212;0.034</td><td char="." align="char">0.579</td></tr><tr><td align="left" colspan="7">Female</td></tr><tr><td align="left">&#x000a0;Restrained</td><td char="." align="char">&#x02212;0.159</td><td char="." align="char">0.007*</td><td char="." align="char">&#x02212;0.236</td><td char="." align="char">&#x0003c;0.001*</td><td char="." align="char">&#x02212;0.276</td><td char="." align="char">&#x0003c;0.001*</td></tr><tr><td align="left">&#x000a0;Emotional</td><td char="." align="char">&#x02212;0.107</td><td char="." align="char">0.114</td><td char="." align="char">&#x02212;0.145</td><td char="." align="char">0.023*</td><td char="." align="char">&#x02212;0.157</td><td char="." align="char">0.015*</td></tr><tr><td align="left">&#x000a0;External</td><td char="." align="char">&#x02212;0.025</td><td char="." align="char">0.709</td><td char="." align="char">0.034</td><td char="." align="char">0.602</td><td char="." align="char">&#x02212;0.062</td><td char="." align="char">0.337</td></tr></tbody></table><table-wrap-foot><p> <italic>&#x003b2;</italic> standard coefficient in multiple linear regression analysis, adjusting for height</p><p> <italic>BMI</italic> body mass index</p><p>* Significantly correlated with a discrepancy of physical index</p></table-wrap-foot></table-wrap> </p><p>Among the females, restrained eating was negatively associated with the discrepancy of weight (<italic>&#x003b2;</italic>&#x000a0;=&#x000a0;&#x02212;&#x000a0;0.159, <italic>p</italic>&#x000a0;=&#x000a0;0.007), the discrepancy of BMI (<italic>&#x003b2;</italic>&#x000a0;=&#x000a0;&#x02212;&#x000a0;0.236, <italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.001), and the discrepancy of BSh (<italic>&#x003b2;</italic>&#x000a0;=&#x000a0;&#x02212;&#x000a0;0.276, <italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.001). Emotional eating was negatively associated with the discrepancy of BMI (<italic>&#x003b2;</italic>&#x000a0;=&#x000a0;&#x02212;&#x000a0;0.145, <italic>p</italic>&#x000a0;=&#x000a0;0.023) and the discrepancy of BSh (<italic>&#x003b2;</italic>&#x000a0;=&#x000a0;&#x02212;&#x000a0;0.157, <italic>p</italic>&#x000a0;=&#x000a0;0.015).</p></sec><sec id="Sec7"><title>Discussion</title><p>We attempted to clarify the relationship between eating behavior and body image in Japanese university students. Our main findings show that the ideal weight and ideal BMI were higher than the current weight and BMI in males, but lower in females. On the other hand, the females&#x02019; ideal body shape was smaller than their perception of current body shape. Restrained eating, emotional eating, and external eating in the DEBQ were higher in the females than the males among the normal-weight participants, and restrained eating and external eating were higher in the females than the males among the underweight participants. In addition, in the results of multiple linear regression analysis adjusting for height, higher score of the restrained eating showed less discrepancy of weight, BMI, and body shape in both the males and females. Emotional eating was negatively associated with the discrepancy in BMI and body shape only among the females.</p><sec id="Sec8"><title>Gender differences in ideal body image</title><p>We asked the participants to state their ideal height in addition to their ideal body weight and ideal body shape. Few studies have included ideal height as a study point, and here we found that the questionnaire&#x02019;s item about ideal height helped to clarify the gender difference in ideal body shape. Namely, the ideal height, ideal weight, ideal BMI, and ideal body shape were all larger than the current values in the males, whereas in the females, the ideal body weight and ideal body shape were lower than the current values, although ideal height was higher than the current height.</p><p>In previous reports, females tended to aspire to a slender body shape [<xref ref-type="bibr" rid="CR25">25</xref>&#x02013;<xref ref-type="bibr" rid="CR27">27</xref>], which is consistent with our present findings. It is interesting that young Japanese females still tend to aspire to a lean body even though they are not obese as compared with females in Europe or the US [<xref ref-type="bibr" rid="CR3">3</xref>, <xref ref-type="bibr" rid="CR27">27</xref>]. The reason why non-obese Japanese females aspire to a lean body and the reason why underlying the gender difference in ideal body shape are not yet clear. Females might prefer a lean body shape irrespective of their height or current body shape. In males, McCreary and Sasse [<xref ref-type="bibr" rid="CR15">15</xref>] defined the drive for muscularity, which is a desire to enhance one&#x02019;s musculature, and other studies with similar results have been published [<xref ref-type="bibr" rid="CR16">16</xref>&#x02013;<xref ref-type="bibr" rid="CR18">18</xref>]. This may be one possible reason for the present results. Further investigations are needed to draw conclusions about the direction and gender differences in body shape.</p></sec><sec id="Sec9"><title>Eating behavior and anthropometric index and body image</title><p>Almost all of the previous studies that used the DEBQ reported that only restrained eating was used to examine the association of DEBQ with body image. In the present study, we examined not only restrained eating, but also all three eating indexes of the DEBQ to explore the gender difference. It has been reported that the scores of restrained eating were higher in females than in males [<xref ref-type="bibr" rid="CR28">28</xref>&#x02013;<xref ref-type="bibr" rid="CR30">30</xref>], which are consistent with the present study. Nguyen-Rodriguez et al. [<xref ref-type="bibr" rid="CR31">31</xref>] reported that scores of emotional eating were not different between boys and girls in junior high school, whereas the score of emotional eating in females was significantly higher than in males in the present study. This disparity indicates that age may have some relations with emotional eating. Indeed, previous studies in adults showed that the scores of emotional eating in females were significantly higher than those in males [<xref ref-type="bibr" rid="CR29">29</xref>, <xref ref-type="bibr" rid="CR30">30</xref>].</p><p>Moreover, the scores on all three eating indexes were higher in the females than the males among the normal-weight participants, and restrained eating and external eating were higher in the females than the males among the underweight participants. In light of these results, it appears that both normal-weight and underweight females may be more sensitive to eating than their male counterparts.</p><p>Our multiple linear regression analysis showed that in the females, restrained eating was negatively associated with the discrepancies in weight, BMI, and body shape. Emotional eating in females was also negatively associated with the discrepancies of BMI and body shape. These results indicate that restrained and emotional eating were associated with a drive for thinness in females, which is consistent with previous reports [<xref ref-type="bibr" rid="CR10">10</xref>, <xref ref-type="bibr" rid="CR32">32</xref>, <xref ref-type="bibr" rid="CR33">33</xref>]. Among the present study&#x02019;s male participants, restrained eating was negatively associated with the discrepancies in weight, BMI, and body shape, similar to the present results for the females. On the other hand, emotional eating was not significantly associated with any discrepancies in the males. External eating was significantly associated with the discrepancy of BMI.</p><p>The reason for the gender differences in the present results is unclear. Emotional eating is most often defined as eating in response to negative affect [<xref ref-type="bibr" rid="CR34">34</xref>]. Nguyen-Rodriguez et al. showed that emotional eating was positively associated with negative mood, such as perceived stress and worries, only in females [<xref ref-type="bibr" rid="CR31">31</xref>]. Thus, females may be more sensitive to emotional eating than males. This concept should be studied more precisely in the future.</p></sec><sec id="Sec10"><title>Limitations</title><p>The limitations of this study should be noted. First, the samples were collected from a limited area. Second, the present participants were all Japanese students. Regarding the cut-off points of underweight and overweight, WHO criteria are also used for Japanese people. However, in general, Japanese individuals are considerably thinner and shorter than age-matched North American and European individuals. Therefore, the number of overweight subjects was not enough large to analyze, and it may be difficult to generalize the present results.</p></sec><sec id="Sec11"><title>Conclusions</title><p>In the present study, the ideal weight and ideal BMI values were higher than the current weight and current BMI in the males, but lower in the females, whereas ideal body shape was smaller than the perception of current body shape in the females, but not significantly different in the males. In addition, restrained eating, emotional eating, and external eating as measured by the DEBQ were higher in the females than in the males. Among the normal-weight participants, all three of these eating indexes were higher in the females as compared to the males, and among the underweight participants, restrained eating and external eating were higher in the females than the males. Taken altogether, our results indicate that at least in Japanese university students, the gender differences regarding ideal body shape are related to eating behavior.</p></sec></sec>

Dissecting conformational contributions to glycosidase catalysis and inhibition

Could not extract abstract

<contrib contrib-type="author" id="aut0005"><name><surname>Speciale</surname><given-names>Gaetano</given-names></name><xref rid="aff0005" ref-type="aff">1</xref></contrib><contrib contrib-type="author" id="aut0010"><name><surname>Thompson</surname><given-names>Andrew J</given-names></name><xref rid="aff0010" ref-type="aff">2</xref></contrib><contrib contrib-type="author" id="aut0015"><name><surname>Davies</surname><given-names>Gideon J</given-names></name><xref rid="aff0010" ref-type="aff">2</xref></contrib><contrib contrib-type="author" id="aut0020"><name><surname>Williams</surname><given-names>Spencer J</given-names></name><email>sjwill@unimelb.edu.au</email><xref rid="aff0005" ref-type="aff">1</xref></contrib><aff id="aff0005"><label>1</label>School of Chemistry and Bio21 Molecular Science and Biotechnology Institute, University of Melbourne, Parkville, Victoria 3010, Australia</aff><aff id="aff0010"><label>2</label>Department of Chemistry, University of York, Heslington, York YO10 5DD, United Kingdom</aff>

Current Opinion in Structural Biology

<p id="par0070"> <boxed-text id="tb0005"><p id="par0025"><bold>Current Opinion in Structural Biology</bold> 2014, <bold>28</bold>:1&#x02013;13</p><p id="par0030">This review comes from a themed issue on <bold>Carbohydrate-protein interactions and glycosylation</bold></p><p id="par0035">Edited by <bold>Harry J Gilbert</bold> and <bold>Harry Brumer</bold></p><p id="par9220">For a complete overview see the <ext-link ext-link-type="uri" xlink:href="http://www.sciencedirect.com/science/journal/0959440X/28" id="intr9005"><underline>Issue</underline></ext-link> and the <ext-link ext-link-type="doi" xlink:href="10.1016/j.sbi.2014.09.001" id="intr9010"><underline>Editorial</underline></ext-link></p><p id="par9240">Available online 10th July 2014</p><p id="par0040"> <ext-link ext-link-type="doi" xlink:href="10.1016/j.sbi.2014.06.003" id="intr0005"><bold>http://dx.doi.org/10.1016/j.sbi.2014.06.003</bold></ext-link> </p><p id="par0045">0959-440X/&#x000a9; 2014 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (<ext-link ext-link-type="uri" xlink:href="http://creativecommons.org/licenses/by/3.0/" id="intr9015">http://creativecommons.org/licenses/by/3.0/</ext-link>).</p></boxed-text> </p><p id="par0075">Glycoside hydrolases catalyze the hydrolytic cleavage of the glycosidic bond. They are enzymes of enduring interest owing to the ubiquity of carbohydrates in nature and their importance in human health and disease, the food, detergent, oil &#x00026; gas and biotechnology industries. Glycoside hydrolases generally, but not quite exclusively, perform catalysis with a net retention or inversion of anomeric stereochemistry. The gross mechanisms of glycosidases were postulated by Koshland in 1953 [<xref rid="bib0005" ref-type="bibr">1^{&#x02022;&#x02022;}</xref>], and his prescient insights remain largely true to this day. The glycoside hydrolases are an immensely varied group of enzymes and are usefully classified on the basis of sequence according to the CAZy system (<ext-link ext-link-type="uri" xlink:href="http://www.cazy.org;/" id="intr0010">www.cazy.org;</ext-link> see also Cazypedia: <ext-link ext-link-type="uri" xlink:href="http://www.cazypedia.org/" id="intr0015">www.cazypedia.org</ext-link>), which reveals a growing and formidable diversity of proteins (133 families as of 2014) [<xref rid="bib0010" ref-type="bibr">2</xref>]. What continues to occupy the attention of mechanistic enzymologists is a complete description of the fine details of the overall reaction coordinate. The free energy profile of catalysis is a composite of terms including: bond-making and breaking; the establishment and disbandment of stereoelectronic effects; and conformational effects. Conformational interactions include substrate-based: vicinal (e.g. eclipsing, gauche, &#x00394;2), 1,3-diaxial, and 1,4-bridgehead; and enzyme-based: local and global conformational changes of the enzyme that occur on the time-scale of catalysis [<xref rid="bib0015" ref-type="bibr">3</xref>].</p><p id="par0080">Two major areas of inquiry are active in the area of conformation and glycoside hydrolases:<list list-type="simple" id="lis0010"><list-item id="lsti0025"><label>1.</label><p id="par0085">What are the conformational changes that occur during catalysis upon substrate binding, at the transition state(s), intermediates (if relevant), and product? Aside from the elemental interest in this question, there is the potential for utilizing this information to develop glycosidase inhibitors that take advantage of the considerable amounts of energy used to selectively bind the transition state (for a glycosidase with a catalytic rate enhancement of 10¹⁷, the calculated transition state affinity is 10^&#x02212;22&#x000a0;M [<xref rid="bib0020" ref-type="bibr">4</xref>]), with the enticing possibility that differences in transition state conformation may allow the development of glycosidase-selective inhibitors.</p></list-item><list-item id="lsti0030"><label>2.</label><p id="par0090">Once transition-state structural information is acquired and used to inspire inhibitor development, do the resulting inhibitors actually bind by utilizing the same interactions that are used to stabilize the transition state&#x000a0;&#x02014;&#x000a0;that is, are they genuine transition state mimics? The answers to this question speak to our abilities to realize this unique form of rational inhibitor design.</p></list-item></list></p><p id="par0095">In this review we cover recent developments in the understanding of conformational reaction coordinates and how such information is acquired; and what constitutes good transition state mimicry by inhibitors. This work extends two recent comprehensive reviews [<xref rid="bib0025" ref-type="bibr">5</xref>, <xref rid="bib0030" ref-type="bibr">6&#x02022;</xref>].</p><sec id="sec0005"><title>Contortions along the reaction coordinate</title><p id="par0100">Substantial evidence has accrued that retaining and inverting glycoside hydrolases perform catalysis through an oxocarbenium ion-like transition state with significant bond breakage to the departing group and limited bond formation to the attacking nucleophile (<xref rid="fig0005" ref-type="fig">Figure 1</xref>a) [<xref rid="bib0035" ref-type="bibr">7</xref>]. On the basis of the four idealized half-chair and boat conformations expected for the transition state (see <xref rid="tb0010" ref-type="boxed-text">Side Panel A</xref>), four &#x02018;classical&#x02019; conformational itineraries may be identified (<xref rid="fig0005" ref-type="fig">Figure 1</xref>b). In these simplified presentations, it is apparent that C1 scribes an arc along the conformational reaction coordinate as it undergoes an electrophilic migration from the leaving group to a nucleophile. However, other ring atoms also change positions, in particular O5 and C2. The subtle change in the position of O5 has little mechanistic consequence other than to allow development of the partial double bond. Interactions at C2 are usually (but not always, see: [<xref rid="bib0040" ref-type="bibr">8</xref>]) significant and for the &#x003b2;-glucosidase Abg from <italic>Agrobacterium</italic> sp. or for &#x003b1;-glucosidase of <italic>Saccharomyces cerevisiae</italic> [<xref rid="bib0045" ref-type="bibr">9</xref>] have been shown to contribute 18&#x02013;22&#x000a0;kJ&#x000a0;mol^&#x02212;1 to transition state stabilization [<xref rid="bib0050" ref-type="bibr">10</xref>], highlighting that the repositioning of C2 and its substituent and other electronic changes accompanying formation of the oxocarbenium ion-like transition state can provide substantial amounts of stabilization energy. The ground state conformations and those of intermediates and transition states need not sit squarely on the graticules of the major meridians and latitudes but may be located within the conformational space nearby (see <xref rid="tb0010" ref-type="boxed-text">Side Panel A</xref>).<fig id="fig0005"><label>Figure 1</label><caption><p><bold>(a)</bold> Mechanisms of classical (i) inverting and retaining glycosidases that utilize (ii) an enzymic nucleophile or (iii) substrate-assisted catalysis. <bold>(b)</bold> Classical conformational itineraries around planar, oxocarbenium ion-like transition states in (i,ii) half-chair (<italic>H</italic>) or (iii,iv) boat (<italic>B</italic>) conformations. <bold>(c)</bold> Strategies and reagents used to study key species along the reaction coordinate.</p></caption><graphic xlink:href="gr1"/></fig><boxed-text id="tb0010"><label>Side panel A</label><caption><title>Theoretical considerations of the transition state conformation</title></caption><p id="par0105">It is a stereoelectronic requirement that the development of a partial double bond between O5 and C1 results in flattening of the system C2&#x02014;C1&#x02014;O5&#x02014;C5, with the remaining pyranose C3 and C4 atoms having freedom to move [<xref rid="bib0055" ref-type="bibr">11</xref>]. The possible idealized transition state structures that satisfy these requirements are ³<italic>H</italic>₄, ⁴<italic>H</italic>₃, ^2,5<italic>B</italic> and <italic>B</italic>_2,5 (and the closely related but usually higher energy ⁴<italic>E</italic>, ³<italic>E</italic>, <italic>E</italic>₄ and <italic>E</italic>₃). As the oxocarbenium ion-like transition states of glycosidases are &#x02018;central&#x02019; transition states it is appropriate to invoke the principle of least nuclear motion, which states that elementary reactions that involve the least change in atomic position and electronic configuration will be favoured [<xref rid="bib0060" ref-type="bibr">12</xref>, <xref rid="bib0065" ref-type="bibr">13</xref>]. Accordingly, the conformational reaction coordinate will most likely involve ground state conformations (corresponding to Michaelis, intermediate and product complexes) that are close neighbors to the transition state conformations. The conformational relationships of pyranose rings [<xref rid="bib0070" ref-type="bibr">14</xref>] may be conveniently summarized using a Cremer-Pople sphere [<xref rid="bib0075" ref-type="bibr">15</xref>] or its equivalent Mercator and polar projections (<xref rid="fig0035" ref-type="fig">Figure I</xref>).<fig id="fig0035"><label>Figure I</label><graphic xlink:href="gr1b1"/></fig></p></boxed-text></p><p id="par9110">Powerful computing resources allow the calculation of the energy of every possible conformation of individual sugars providing a so-called free energy landscape (FEL). Each carbohydrate stereoisomer possesses a unique FEL, owing to the presence of various substituents, the resulting hydrogen-bonding interactions, local steric interactions, and the contribution of the anomeric effect. This computational approach was first applied to &#x003b2;-glucopyranose and 9 energetic minima were identified [<xref rid="bib0080" ref-type="bibr">16</xref>]. Aside from the global minimum of ⁴<italic>C</italic>₁, the remaining 8 local minima were approximate <italic>B</italic> and <italic>S</italic> conformations; however these differed from the canonical <italic>B</italic> and <italic>S</italic> conformations owing to the lack of rigidity of the ring, and the presence of attractive (hydrogen bonding) and repulsive (eclipsing and 1,3-diaxial) interactions. Several of these conformations were identified as pre-activated for catalysis, with pseudo-axial C1&#x02014;O1 bonds leading to a lengthening of the C1&#x02014;O1 bond and a shortening of the O5&#x02014;C1 bond owing to the developing anomeric effect, and partial charge at C1 (see <xref rid="tb0015" ref-type="boxed-text">Side Panel B</xref>). Enticingly, the pre-activated conformations are those that are most frequently observed in so-called Michaelis complexes (representing E.S complexes) studied by X-ray crystallography for GHs that process &#x003b2;-glucosides. FEL analysis has been extended to include &#x003b1;-glucopyranose, &#x003b2;-xylopyranose, &#x003b2;-mannopyranose, and &#x003b2;-<italic>N</italic>-acetylglucopyranosamine [<xref rid="bib0085" ref-type="bibr">17</xref>], and &#x003b1;-mannopyranose [<xref rid="bib0090" ref-type="bibr">18^{&#x02022;&#x02022;}</xref>], with the data supporting the conclusion that for these sugars the catalytically relevant conformations are frequently the energetically predisposed distorted structures. These observations are harmonious with the suggestion of Wolfenden that the fact that many enzymes achieve rates approaching the diffusion limit suggests that they have evolved to recognize species that are reasonably populous in solution [<xref rid="bib0095" ref-type="bibr">19</xref>].<boxed-text id="tb0015"><label>Side panel B</label><caption><title>On substrate distortion and Michaelis complexes</title></caption><p id="par0110">The concept of substrate distortion upon binding to an enzyme in the enzyme-substrate (Michaelis) complex has long been proposed since the earliest X-ray crystallographic data of glycosidases became available. Four major principles are invoked to justify the need for a substrate to distort from a ground-state and less reactive conformation to a distorted and typically higher energy conformation: (1) the principle of least nuclear motion, which favours distorted conformations that require less nuclear movement to achieve the transition state (see Side Panel A); (2) the geometric demand that nucleophilic attack on the anomeric centre needs to be &#x02018;in line&#x02019; with the departing nucleophile, which arises from the stereoelectronic requirement that the electrons derived from the nucleophile will populate the &#x003c3;* orbital of the glycosidic bond; (3) the stereoelectronic requirement for development of a partial double bond at the oxocarbenium ion-like transition state, which requires electron donation by a suitably located n-type lone pair on the endocyclic oxygen; and (4) the fact that general acid catalytic residues of glycosidases are located within the plane of the ring, either syn or anti to the C1&#x02014;O5 bond [<xref rid="bib0100" ref-type="bibr">20</xref>]. Although often invoked as a rationale for substrate distortion, Sinnott has persuasively argued that the &#x02018;antiperiplanar lone pair hypothesis&#x02019;, viz. that sp³ lone pairs of electrons on a heteroatom direct the departure of a leaving group from an adjacent tetrahedral carbon centre, requires implausible contortions of the pyranose ring [<xref rid="bib0105" ref-type="bibr">21</xref>].</p></boxed-text></p></sec><sec id="sec0010"><title>Defining the conformational coordinate: Is seeing believing?</title><p id="par0120">X-ray crystallography is a powerful technique that provides a detailed molecular description of the catalytic machinery of an enzyme. While unliganded (apo) structures provide some information that can be used to help understand mechanism, the acquisition of complexes, with substrates (or substrate analogues), sugar-shaped inhibitors, mechanism-based inhibitors, or products have the potential to reveal intricate details of the amino acid residues involved in catalysis and the conformations of enzyme-bound species (<xref rid="fig0005" ref-type="fig">Figure 1</xref>c). Three main strategies for the acquisition of Michaelis (E.S)-like complexes of retaining and inverting glycosidases are: firstly, co-crystallization of non-hydrolyzable substrate mimics with wildtype enzyme (<xref rid="fig0005" ref-type="fig">Figure 1</xref>c(i)) [<xref rid="bib0110" ref-type="bibr">22</xref>], or secondly, co-crystallization of substrate with catalytically inactive mutant enzymes, or thirdly, co-crystallization of substrate and wildtype enzyme at a pH at which it is inactive. Occasionally, Michaelis complexes have been obtained serendipitously at pH values under which the enzyme is active; the reason for lack of hydrolysis in these cases is unclear [<xref rid="bib0115" ref-type="bibr">23</xref>, <xref rid="bib0120" ref-type="bibr">24</xref>].</p><p id="par0125">For retaining glycoside hydrolases that proceed through a glycosyl enzyme intermediate, fairly effective methods have been developed to allow the trapping of kinetically competent intermediates [<xref rid="bib0125" ref-type="bibr">25</xref>]; the general principle is to rapidly access the glycosyl enzyme intermediate using a good leaving group, but to modify the sugar such that its turnover to product is slowed, allowing its accumulation and study. The initial work by Legler involved addition to glycals to generate 2-deoxyglycosyl enzymes [<xref rid="bib0130" ref-type="bibr">26</xref>] or use of aryl 2-deoxyglycosides [<xref rid="bib0135" ref-type="bibr">27</xref>], and was elegantly extended by Withers to 2-deoxy-2-fluoro-, 2-deoxy-2,2-difluoro- and 5-fluoro glycosyl fluorides (and closely related 2,4-dinitrophenyl, and other activated glycosides) [<xref rid="bib0125" ref-type="bibr">25</xref>] (<xref rid="fig0005" ref-type="fig">Figure 1</xref>c(ii)). For reasons that are not entirely clear, for &#x003b1;-glycosidases the use of C5-inverted 5-fluoro-glycosyl fluorides for the corresponding enzymes usually yield better trapping results than for the stereochemically-matched alternative. For retaining enzymes that proceed by anchimeric assistance from a 2-acylamido group, sulfur mimics of the proposed oxazoline (or oxazolinium ion) intermediate, most notably NAGthiazoline [<xref rid="bib0140" ref-type="bibr">28</xref>, <xref rid="bib0145" ref-type="bibr">29</xref>], have proved effective as inhibitors and informative as mechanistic probes for crystallographically studying the intermediate (<xref rid="fig0005" ref-type="fig">Figure 1</xref>c(iii)).</p><p id="par0130">It is important to recognize that all X-ray structures of protein&#x02013;ligand complexes are by their very nature not catalytically competent and thus care must be taken in how to interpret the important clues they provide in the proposal of a credible conformational itinerary. Kinetically trapped species recapitulate the major bond-forming and breaking events but the structural modifications made to allow kinetic trapping may perturb substrate interactions that are important for defining the conformational reaction coordinate. Occasionally, crystallization efforts at non-optimal pH have led to the acquisition of apparently bonafide Michaelis complexes; however even these constitute complexes with catalytically-incompetent enzymes and the interpretation of these structures must recognize that these do not lie on the reaction coordinate. Complexes with substrate and enzyme may be &#x02018;pre-Michaelis&#x02019; complexes that represent enzyme-bound species that precede the formation of the true Michaelis complex, or may be catalytically irrelevant species that are actively misleading. Product bound to enzyme may have relaxed from its first formed conformation as the lack of a sizeable anomeric substituent prevents the enzyme from utilizing +1 subsite interactions to stabilize its conformation. Nonetheless, with some exceptions, most pseudo-Michaelis, glycosyl enzyme intermediate or thiazoline intermediate, and product complexes are sufficiently akin to hypothetical bonafide species on the reaction coordinate to allow cautious but probably reasonable insights into mechanism.</p><p id="par0135">Caution must also be exercised when studying complexes with sugar-shaped heterocycles that function as competitive inhibitors (<xref rid="fig0005" ref-type="fig">Figure 1</xref>c(iii)). While superficially these compounds resemble aspects of the proposed transition state, there are intrinsic limitations of what can be mimicked in a chemically stable compound, including hybridization changes, partial charge development, and fractional bond orders. A study of the FELs of two inhibitors that display superficial transition state mimicry: isofagomine and mannoimidazole revealed dramatic differences (<xref rid="fig0010" ref-type="fig">Figure 2</xref>a) [<xref rid="bib0150" ref-type="bibr">30^{&#x02022;&#x02022;}</xref>]. Isofagomine is strongly biased toward a ⁴<italic>C</italic>₁ conformation, with potential transition state mimicking ⁴<italic>H</italic>₃ and <italic>B</italic>_2,5 conformations lying 12 and 8&#x000a0;kcal&#x000a0;mol^&#x02212;1 higher, respectively, and importantly with a significant barrier to attaining those conformations. On the other hand while mannoimidazole prefers ⁴<italic>H</italic>₃ and ³<italic>H</italic>₄ conformations (with a 1&#x000a0;kcal&#x000a0;mol^&#x02212;1 preference for the latter), the <italic>B</italic>_2,5 conformation is also energetically accessible. Overlaying the observed conformations of isofagomine-type and mannoimidazole-type inhibitors from X-ray structures with mannose-processing enzymes of various GH families reveals all isofagomine complexes adopt a ⁴<italic>C</italic>₁ conformation, whereas for mannoimidazole the <italic>B</italic>_2,5 conformation is observed on enzymes of families GH2, 26, 38, 92 and 113, implying a ¹<italic>S</italic>₅&#x02192;<italic>B</italic>_2,5^&#x02021;&#x02192;^O<italic>S</italic>₂ conformational itinerary. One interesting footnote is that non-ground state conformations of isofagomine-type inhibitors have been observed on family GH6 cellulases in either ^2,5<italic>B</italic>/²<italic>S</italic>_O or ²<italic>S</italic>_O conformations [<xref rid="bib0155" ref-type="bibr">31</xref>, <xref rid="bib0160" ref-type="bibr">32</xref>]. The energetic difficulties in attaining such conformations highlight their special significance when seen and in these cases they reflect the proposed ²<italic>S</italic>_O&#x02192;^2,5<italic>B</italic>^&#x02021;&#x02192;⁵<italic>S</italic>₁ itinerary.<fig id="fig0010"><label>Figure 2</label><caption><p>Computational studies, in concert with X-ray crystallography and inhibitor design and synthesis, assist in assigning conformational itineraries. <bold>(a)</bold> Assigning the conformational itinerary of <italic>Cellvibrio japonicas</italic> GH26 &#x003b2;-mannanase Man26C. (i) Free energy landscapes reveal mannoimidazole, unlike isofagomine, is able to attain the conformations relevant to glycosidase catalysis; (ii) X-ray structures of a Michaelis complex (1GVY), glycosyl enzyme intermediate (1GW1), and transition state mimicking &#x003b2;-mannosyl-1,4-mannoimidazole complex (4CD5); (iii) proposed conformational itinerary. <bold>(b)</bold> Assigning the conformational itinerary of <italic>Caulobacter</italic> strain K31 GH47 &#x003b1;-mannosidase. (i) Free energy landscapes highlight substrate preactivation off-enzyme, and reshaping of the available conformations on-enzyme; (ii) X-ray structures of Michaelis complex (4AYP), transition state mimicking mannoimidazole complex (4AYQ), and product mimicking noeuromycin complex (4AYR); (iii) proposed conformational itinerary.</p></caption><graphic xlink:href="gr2"/></fig></p><p id="par0140">While the FEL of an isolated carbohydrate is often biased toward those conformations pre-activated for catalysis, further substrate distortion presumably occurs upon binding to enzyme. Quantum mechanics/molecular mechanics calculations of &#x003b1;-mannopyranose revealed that a FEL determined within the constraints of a GH47 &#x003b1;-mannosidase is moulded by the enzyme to dramatically limit the conformations accessible by the substrate to a previously inaccessible region of the FEL for the substrate off-enzyme (<xref rid="fig0010" ref-type="fig">Figure 2</xref>b) [<xref rid="bib0090" ref-type="bibr">18^{&#x02022;&#x02022;}</xref>]. In support of the theoretical predictions, X-ray analysis of &#x02018;snapshot&#x02019; complexes of the enzyme with a substrate analogue, transition state, and product mimics supported a ³<italic>S</italic>₁&#x02192;³<italic>H</italic>₄^&#x02021;&#x02192;¹<italic>C</italic>₄ conformational itinerary predicted on the basis of the FEL remodelling.</p></sec><sec id="sec0015"><title>Sugars getting into shape: News dispatches from the families</title><p id="par0145"><xref rid="tbl0005" ref-type="table">Table 1</xref> summarizes well-defined conformational itineraries for a range of GH enzymes (see Supporting Information for a more detailed listing). We present a few highlights from the last two years that are not covered in detail elsewhere.<table-wrap position="float" id="tbl0005"><label>Table 1</label><caption><p>Conformational itineraries around various transition state conformations. Listed are families for which strong evidence in support of a conformational itinerary is available. For a more comprehensive listing see Supporting Information Table S1</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left">Transition state conformation</th><th align="center">GH families</th><th align="center">Configuration of substrate</th><th align="center">Enzymatic activities</th></tr></thead><tbody><tr><td align="left">³<italic>H</italic>₄^&#x02021; (⁴<italic>H</italic>₅^&#x02021; for sialidases)</td><td align="left">29, 33, 34, 47</td><td align="left"><sc>l</sc>-<italic>fuco</italic> (=<sc>l</sc>-<italic>galacto</italic>) sialic acid,</td><td align="left">&#x003b1;-Fucosidase</td></tr><tr><td/><td/><td align="left"><sc>d</sc>-<italic>manno</italic></td><td align="left">&#x003b1;-Mannosidase</td></tr><tr><td align="left">⁴<italic>H</italic>₃^&#x02021;</td><td align="left">1, 2, 3, 5, 7, 10, 12, 16, 20, 22, 26, 27, 30, 84</td><td align="left"><sc>d</sc>-<italic>gluco</italic>/<sc>d</sc>-<italic>manno</italic></td><td align="left">&#x003b2;-Glucosidase/cellulase/lichenases/b-mannosidase</td></tr><tr><td/><td/><td align="left"><sc>d</sc>-<italic>galacto</italic></td><td align="left">&#x003b2;-Galactosidase</td></tr><tr><td/><td/><td align="left"><sc>d</sc>-<italic>gluco</italic></td><td align="left">&#x003b1;-Glucosidase/&#x003b1;-glucanase</td></tr><tr><td/><td/><td align="left"><sc>d</sc>-<italic>xylo</italic></td><td align="left">&#x003b2;-Xylosidase/xylanase</td></tr><tr><td/><td/><td align="left"><sc>d</sc>-<italic>gluco/galacto</italic></td><td align="left">&#x003b2;-Hexosaminidase</td></tr><tr><td align="left">^2,5<italic>B</italic>^&#x02021;</td><td align="left">6, 8, 11</td><td align="left"><sc>d</sc>-<italic>gluco</italic></td><td align="left">&#x003b2;-Glucosidase, chitinase</td></tr><tr><td/><td/><td align="left"><sc>d</sc>-<italic>xylo</italic></td><td align="left">Xylanase</td></tr><tr><td align="left"><italic>B</italic>_2,5^&#x02021;</td><td align="left">2, 26, 38, 92, 113</td><td align="left"><sc>d</sc>-<italic>manno</italic></td><td align="left">&#x003b2;-Mannosidase, &#x003b2;-mannanase, &#x003b1;-mannosidase</td></tr><tr><td align="left">⁴<italic>E</italic>^&#x02021;</td><td align="left">117</td><td align="left">3,6-anhydro-<sc>l</sc>-<italic>galacto</italic></td><td align="left">&#x003b1;-1,3-<sc>l</sc>-Neoagarobiase</td></tr></tbody></table></table-wrap></p><p id="par0150">&#x003b2;-Hexosaminidases of family GH3 perform catalysis through a two step mechanism with the initial substitution step occurring by an enzymatic nucleophile to afford a glycosyl enzyme intermediate (<xref rid="fig0005" ref-type="fig">Figure 1</xref>a(ii)). Insight into the reaction coordinate has been obtained through trapping a glycosyl enzyme intermediate in a ⁴<italic>C</italic>₁ conformation using the mechanism based inhibitor 5F-GlcNAcF, suggesting a ¹<italic>S</italic>₃&#x02192;⁴<italic>H</italic>₃^&#x02021;&#x02192;⁴<italic>C</italic>₁ itinerary [<xref rid="bib0165" ref-type="bibr">33^&#x02022;</xref>]. Interestingly, a complex with an inactive mutant with substrate, and of wild-type with product, also revealed ⁴<italic>C</italic>₁ conformations. In this case there is strong evidence that the complex with substrate is not a bonafide Michaelis complex as a loop containing the putative histidine general acid/base undergoes a dramatic movement. In the product complex it appears that the sugar has relaxed to a more stable conformation.</p><p id="par0155">Family GH39 &#x003b1;-<sc>l</sc>-iduronidase (IDUA) is a retaining lysosomal enzyme that assists in the stepwise degradation of heparin sulfate and dermatan sulfate, and which is of interest for enzyme replacement therapy of the associated lysosomal storage disorder (LSD) mucopolysaccharidosis type I [<xref rid="bib0170" ref-type="bibr">34^&#x02022;</xref>]. Michaelis complexes with iduronate analogues in ²<italic>S</italic>_O conformations, and the trapping of a glycosyl enzyme on IDUA in a ⁵<italic>S</italic>₁/^2,5<italic>B</italic> conformation using 2-deoxy-2-fluoro-&#x003b1;-<sc>l</sc>-idopyranosyluronic acid fluoride, imply a ²<italic>S</italic>_O&#x02192;^2,5<italic>B</italic>^&#x02021;&#x02192;⁵<italic>S</italic>₁ conformational itinerary (<xref rid="fig0015" ref-type="fig">Figure 3</xref>a) [<xref rid="bib0175" ref-type="bibr">35^{&#x02022;&#x02022;}</xref>].<fig id="fig0015"><label>Figure 3</label><caption><p>New glycosidase conformational assignments. <bold>(a)</bold> A likely conformational itinerary for the GH39 human &#x003b1;-<sc>l</sc>-iduronidase based on structures of a Michaelis complex (4KGJ) and a glycosyl enzyme intermediate (4KH2). <bold>(b)</bold> A likely conformational itinerary for an &#x003b1;-<sc>l</sc>-neoagarobiase based on a Michaelis complex (4AK7). <bold>(c)</bold> A possible conformational itinerary for a GH99 endo-&#x003b1;-mannosidase based on a proposed mechanism that proceeds through a 1,2-anhydro sugar intermediate.</p></caption><graphic xlink:href="gr3"/></fig></p><p id="par0160">Family GH59 &#x003b2;-galactocerebrosidase (GALC) degrades glycosphingolipids and its deficiency leads to another LSD, Krabbe disease. X-ray &#x02018;snapshots&#x02019; of a Michaelis complex (using 4-nitrophenyl &#x003b2;-<sc>d</sc>-galactopyranoside), a 2-deoxy glycosyl enzyme (from galactal addition), and product (with galactose) showed the sugar ring in a ⁴<italic>C</italic>₁ conformation in all structures [<xref rid="bib0180" ref-type="bibr">36^&#x02022;</xref>]. This surprising result, reminiscent of that seen with the GH2 &#x003b2;-galactosidase LacZ [<xref rid="bib0185" ref-type="bibr">37</xref>], was interpreted as suggesting that no distortion of the ring occurs along the reaction coordinate. Interestingly, the acid/base residue in the substrate complex is incorrectly positioned suggesting that this is not a Michaelis complex; additionally, the product complex may have relaxed from its first-formed conformation.</p><p id="par0165">Family GH117 &#x003b1;-1,3-<sc>l</sc>-neoagarobiase has been described as a keystone enzyme owing to its role in agarose degradation, which provides the capability for the human gut microbiota to degrade seaweed diets [<xref rid="bib0190" ref-type="bibr">38^{&#x02022;&#x02022;}</xref>]. These interesting (probably inverting) enzymes act on 3,6-anhydro-&#x003b1;-<sc>l</sc>-galactosides (<xref rid="fig0015" ref-type="fig">Figure 3</xref>b). The 3,6-anhydro bridge imparts significant rigidity on the sugar to prefer ⁴<italic>C</italic>₁ and ^1,4<italic>B</italic> conformations. A Michaelis complex of an inactive mutant with neoagarobiose highlighted a histidine residue as a potential catalytic general acid and revealed a ^1,4<italic>B</italic> conformation, suggestive of a ^1,4<italic>B</italic>&#x02192;⁴<italic>E</italic>^&#x02021;&#x02192;⁴<italic>C</italic>₁ conformational itinerary.</p><p id="par0170">GH11 xylanases are an as yet unresolved case. Early structures of intermediate complexes trapped with 2-fluoro sugars were interpreted as ^2,5<italic>B</italic> conformations [<xref rid="bib0195" ref-type="bibr">39</xref>, <xref rid="bib0200" ref-type="bibr">40</xref>], suggesting a possible ²<italic>S</italic>_O&#x02192;^2,5<italic>B</italic>^&#x02021;&#x02192;⁵<italic>S</italic>₁ itinerary. Very recently a long sought apparent Michaelis complex of xylohexaose bound to the xylanase XynII from <italic>Trichoderma reesi</italic> revealed a slightly distorted ⁴<italic>C</italic>₁ conformation [<xref rid="bib0205" ref-type="bibr">41</xref>]. Confounding this issue, product complexes with GH11 enzymes reveal a range of different conformations that are inconsistent with the proposed itinerary.</p></sec><sec id="sec0020"><title>On the importance of being mannose</title><p id="par0175">The majority of common, naturally occurring sugars in their ground-state chair conformation either have an equatorial hydroxyl at C2 (galactosides, glucosides, xylosides, fucosides), or no substituent (sialosides; formally C3). Mannosides and rhamnosides, bearing axial 2-hydroxyls in the ground state conformations, provide exceptions that have interesting and significant consequences for reactivity that lies at the heart of what has been described the recalcitrant chemistry of mannose. For &#x003b1;-mannosides in a ⁴<italic>C</italic>₁ conformation, in addition to the existence of the stabilizing anomeric effect that dissuades substrate distortion, the presence of the strongly electron-withdrawing OH at C2 results in opposing dipoles at C1 and C2 that provide additional ground state stabilization. For &#x003b2;-mannosides in a ⁴<italic>C</italic>₁ conformation, the anomeric effect provides little stabilization. In addition, other destabilizing effects are operative. Collectively, these can be described as a &#x00394;2 effect, a term first coined by Reeves [<xref rid="bib0210" ref-type="bibr">42</xref>]. The &#x00394;2 effect describes the destabilizing effect of an oxygen on one carbon that bisects two oxygens substituted on an adjacent carbon, aligning dipoles and causing gauche&#x02013;gauche interactions between the vicinal oxygens. Nucleophilic substitutions at C1 of &#x003b1;-mannosides have to contend with a developing &#x00394;2 effect. Reflecting these complexities, nature has devised some remarkable strategies for enzymes to solve these problems.</p><p id="par0180">&#x003b1;-Mannosidases of families GH38, 47 and 92 are metal dependent, with crystallographic evidence for the divalent metal cation (Zn²⁺ or Ca²⁺) binding O2 and O3. This interesting observation may provide a means to overcome the high stability of the unreactive ⁴<italic>C</italic>₁ conformation of &#x003b1;-mannosides, and encourage contraction of the ground state O2&#x02014;C2&#x02014;C3&#x02014;O3 torsion angle within the ⁴<italic>C</italic>₁ conformation of 60&#x000b0; toward the 0&#x02013;15&#x000b0; angle expected at the <italic>B</italic>_2,5 transition state [<xref rid="bib0215" ref-type="bibr">43</xref>]. In addition the flexible coordination number and geometry of calcium may allow coordination and delivery of the nucleophilic water, thereby providing a way to overcome a developing &#x00394;2 effect [<xref rid="bib0215" ref-type="bibr">43</xref>]. A computational study of a GH38 &#x003b1;-mannosidase suggested that Zn²⁺ coordination may stabilize charge that develops on O2 at the oxocarbenium ion like transition state [<xref rid="bib0220" ref-type="bibr">44</xref>]. Remarkably, this calculation revealed that the charge on zinc varies reciprocally with the charge developing on the oxocarbenium ion-like TS.</p><p id="par0185">There is now compelling evidence that family GH2, 26 and 113 retaining &#x003b2;-mannosidases, and GH38 retaining and GH92 inverting &#x003b1;-mannosidases, utilize <italic>B</italic>_2,5 transition states with Michaelis complexes in a ¹<italic>S</italic>₅ (for &#x003b2;-) or ^O<italic>S</italic>₂ (for &#x003b1;-), and thus operate through ¹<italic>S</italic>₅&#x02194;<italic>B</italic>_2,5^&#x02021;&#x02194;^O<italic>S</italic>₂ conformational itineraries [<xref rid="bib0150" ref-type="bibr">30^{&#x02022;&#x02022;}</xref>]. The Michaelis complex conformation provides a pseudo axial arrangement of the anomeric leaving group and permits inline attack of the nucleophile; and importantly the ¹<italic>S</italic>₅ conformation relieves the &#x00394;2 effect. One question that logically arises from studies of transition state conformation is whether all enzymes within a family operate with the same conformational itinerary? Family GH26 contains enzymes that act on &#x003b2;-mannosides, &#x003b2;-glucosides and &#x003b2;-xylosides: lichenases (which hydrolyse the mixed linkage &#x003b2;-1,3-; &#x003b2;-1,4-glucan lichenan), &#x003b2;-mannanase and &#x003b2;-1,3-xylanases. Studies of Michaelis complexes and trapped glycosyl enzymes provides good evidence for an alternative ¹<italic>S</italic>₃&#x02192;⁴<italic>H</italic>₃^&#x02021;&#x02192;⁴<italic>C</italic>₁ itinerary for lichenases [<xref rid="bib0225" ref-type="bibr">45</xref>] and &#x003b2;-1,3-xylanases [<xref rid="bib0230" ref-type="bibr">46^&#x02022;</xref>]. The different conformations of the transition state of the <sc>d</sc>-<italic>gluco</italic>/<sc>d</sc>-<italic>xylo</italic> and <sc>d</sc>-<italic>manno</italic> configured substrates result in the substituents at C2 being pseudo-equatorial in both cases and lying at essentially the same place in space, explaining how the conserved catalytic machinery of different GH26 family members can tolerate differently configured sugars, with the specificity arising from a large difference in the positions of the C3 substituents [<xref rid="bib0225" ref-type="bibr">45</xref>], a relationship which is highlighted by the common inhibition of &#x003b2;-mannosidases and &#x003b2;-glucosidases by isofagomine lactam [<xref rid="bib0235" ref-type="bibr">47</xref>].</p><p id="par0190">Uncertainty surrounds the conformational itineraries of &#x003b1;-mannosidases of families GH76, 99 and 125. No complexes are available for GH76 that could provide any insight into a possible itinerary. For GH99, which contains retaining endo-acting &#x003b1;-mannosidases, the only complexes available are with isofagomine and deoxymannojirimycin-derived inhibitors, and these bind in ⁴<italic>C</italic>₁ conformations which match the ground state of the inhibitors, so it is not clear whether these complexes represent enzyme-induced or substrate-biased conformations. However, on the basis of an inability to identify a catalytic nucleophile in the complex with &#x003b1;-glucosyl-1,3-isofagomine, a neighboring group participation mechanism for GH99 was suggested that proceeded through a 1,2-anhydro sugar [<xref rid="bib0240" ref-type="bibr">48^&#x02022;</xref>]. This proposal implies the intermediate adopts a ⁴<italic>H</italic>₅ conformation, and least nuclear motion would predict a ⁴<italic>C</italic>₁&#x02192;⁴<italic>E</italic>^&#x02021;&#x02192;⁴<italic>H</italic>₅ itinerary (<xref rid="fig0015" ref-type="fig">Figure 3</xref>c). For the inverting GH125 &#x003b1;-mannosidases, a pseudo Michaelis complex is available which has the &#x02212;1 sugar in an undistorted ⁴<italic>C</italic>₁ conformation, which matches that observed with a complex with deoxymannojirimycin [<xref rid="bib0245" ref-type="bibr">49</xref>]. The lack of distortion for enzyme bound to the non-hydrolyzable substrate is surprising.</p></sec><sec id="sec0025"><title>Neuraminidases: of conformational itineraries and transition state mimicry by inhibitors</title><p id="par0195">Neuraminidases (sialidases) are glycosidases that cleave sialic acid residues, with the family GH34 viral surface, retaining neuraminidases being significant as the eponymous enzymes in the HXNY classification system of influenza viruses. Influenza virus neuraminidases play key roles in the infection of cells by the virus and the ability of progeny virions to detach from an infected cell and infect new cells. In two related studies, Withers and co-workers [<xref rid="bib0250" ref-type="bibr">50^{&#x02022;&#x02022;}</xref>] and Gao and co-workers [<xref rid="bib0255" ref-type="bibr">51^&#x02022;</xref>] designed a series of neuraminidase inhibitors that combine features of the deoxyfluorosugar inhibitors modified to incorporate structural features of the clinically-approved drugs zanamivir (Relenza) and oseltamivir (Tamiflu). X-ray structures of an elusive tyrosyl enzyme intermediate revealed a ²<italic>C</italic>₅ conformation (<xref rid="fig0020" ref-type="fig">Figure 4</xref>a). While a ⁴<italic>S</italic>₂&#x02192;⁴<italic>H</italic>₅^&#x02021;&#x02192;²<italic>C</italic>₅ (equivalent to a ³<italic>S</italic>₁&#x02192;³<italic>H</italic>₄^&#x02021;&#x02192;¹<italic>C</italic>₄ for a hexopyranose) is consistent with this data and was proposed for neuraminidases of GH33 [<xref rid="bib0260" ref-type="bibr">52</xref>], Bennet reported kinetic isotope effect analysis of the GH33 <italic>Micromonospora viridifaciens</italic> sialidase that implied a Michaelis complex in a ⁶<italic>S</italic>₂ (⁵<italic>S</italic>₁ for aldose) conformation [<xref rid="bib0265" ref-type="bibr">53</xref>], a conformation also seen in the Michaelis complex of a GH33 transialidase from <italic>Trypanosoma cruzi</italic>, and consistent with a ⁶<italic>S</italic>₂&#x02192;⁴<italic>H</italic>₅^&#x02021;&#x02192;²<italic>C</italic>₅ (equivalent to a ⁵<italic>S</italic>₁&#x02192;³<italic>H</italic>₄^&#x02021;&#x02192;¹<italic>C</italic>₄ for an aldose) conformational itinerary [<xref rid="bib0260" ref-type="bibr">52</xref>].<fig id="fig0020"><label>Figure 4</label><caption><p>Conformational itinerary of influenza GH34 neuraminidases and conformational transition state mimicry by inhibitors. <bold>(a)</bold> The conformational itinerary of influenza GH34 neuraminidases informed by an X-ray structure of a glycosyl enzyme intermediate (4H52). <bold>(b)</bold> Complexes of anti-influenza drugs with influenza neuraminidases reveals that oseltamivir (2HU4), unlike zanamivir (1NNC), provides good conformational mimicry of the proposed sialidase transition state.</p></caption><graphic xlink:href="gr4"/></fig></p><p id="par0200">Zanamivir and oseltamivir are potent competitive inhibitors of viral neuraminidases and bear some similarity to the proposed transition state of neuraminidase, yet it is not clear whether either compound achieves its potency through transition state mimicry. As espoused by Wolfenden [<xref rid="bib0270" ref-type="bibr">54</xref>] and Thompson [<xref rid="bib0275" ref-type="bibr">55</xref>] and elegantly summarized by Bartlett [<xref rid="bib0280" ref-type="bibr">56</xref>], critical analysis of transition-state mimicry can be achieved by comparing the effects of equivalent structural perturbations on the affinity of the true transition state (via effects on substrate <italic>k</italic>_cat/<italic>K</italic>_M) and on the affinity of the transition state analogue (via <italic>K</italic>_I values), plotted as a linear free energy relationship. Two approaches may be used to introduce perturbations: firstly, modifications to the inhibitor and the corresponding substrate and measurement of kinetic parameters with the wild-type enzyme, or secondly, mutation of enzymatic active site residues to afford mutant enzymes, which are studied with the same inhibitor and substrate. A limitation of the former method is the effort that needs to be expended on synthesis of derivatives, but which allow atomic level modifications to be made limited only by the imagination and synthetic chemistry. Limitations of the latter include the rather blunt tool of site-directed mutagenesis which is restricted to the genetically encoded natural amino acids, and the possibility that mutational perturbations may affect the fundamental reaction mechanism of the enzyme.</p><p id="par0205">Zanamivir was designed to improve the known sialidase inhibitor neuraminic acid glycal (Neu5Ac2en) by the rational inclusion of an enzyme-specific guanidine group targeting a negatively charged pocket near the active site [<xref rid="bib0285" ref-type="bibr">57</xref>]. Neu5Ac2en bears some similarity to the transition state by virtue of <italic>sp</italic>² hybridization at C2. It can be argued that a transition state mimicking inhibitor has the potential to provide potent inhibitors that should be resistant to mutations within the active site, as mutations that affect the ability of the inhibitor to bind should also affect the catalytic proficiency of the enzyme to similar degrees, resulting in loss of fitness for the virus. By making alterations in the structure of zanamivir at the 4-position and relating the effects of these changes upon inhibitor <italic>K</italic>_I values to the equivalent changes to the substrate and their effect upon <italic>k</italic>_cat/<italic>K</italic>_M or <italic>K</italic>_M Bennet and co-workers showed that zanamivir is not a transition state analogue and is better considered a ground state analogue [<xref rid="bib0290" ref-type="bibr">58^&#x02022;</xref>]. Notably, this is confluent with the observation that influenza strains resistant to zanamivir possess reduced binding avidity for this drug but still possess catalytic competence. With impressive foresight, this possibility was suggested in the earliest publication describing the invention of zanamivir [<xref rid="bib0285" ref-type="bibr">57</xref>]. The failure of zanamivir in this Bartlett analysis is perhaps not overly surprising. Zanamivir has a double bond between C2<private-char name="E00C" description="bond, double bond"><!-- PBM data was replaced with SVG by xgml2pxml: <glyph-data id="pc-E00C" format="PBM" resolution="300" x-size="12" y-size="9"> 000000000000 000000000000 000000000000 111111111111 000000000000 111111111111 000000000000 000000000000 000000000000 </glyph-data> --> <glyph-data id="pc-E00C" format="SVG" resolution="300" x-size="12" y-size="9"> <![CDATA[ <svg xmlns="http://www.w3.org/2000/svg" version="1.0" width="20.666667pt" height="16.000000pt" viewBox="0 0 20.666667 16.000000" preserveAspectRatio="xMidYMid meet"><metadata> Created by potrace 1.16, written by Peter Selinger 2001-2019 </metadata><g transform="translate(1.000000,15.000000) scale(0.019444,-0.019444)" fill="currentColor" stroke="none"><path d="M0 440 l0 -40 480 0 480 0 0 40 0 40 -480 0 -480 0 0 -40z M0 280 l0 -40 480 0 480 0 0 40 0 40 -480 0 -480 0 0 -40z"/></g></svg> ]]> </glyph-data></private-char>C3, and cannot adopt the ⁴<italic>H</italic>₅ conformation predicted for the transition state of GH33 sialidases; indeed an <italic>E</italic>₅ conformation is observed for zanamivir in complexes (<xref rid="fig0020" ref-type="fig">Figure 4</xref>b). On the other hand oseltamivir (Tamiflu) is a carbocycle with a double bond located at the appropriate position to mimic the partial C2&#x02014;O5 double bond at the transition state, and is observed to bind to sialidases in a ⁴<italic>H</italic>₅ conformation matching that of the transition state [<xref rid="bib0295" ref-type="bibr">59</xref>]. It will be interesting to see if Bartlett analysis applied to oseltamivir provides evidence of transition state mimicry. This situation is worth comparing with the powerful &#x003b1;-glucosidase inhibitor acarbose, which has a double bond C5<private-char name="E00C" description="bond, double bond"><glyph-ref glyph-data="pc-E00C" /></private-char>C6 (using pyranose numbering) and cannot adopt a planar conformation matching that expected for a glycosidase transition state; Bartlett analysis of acarbose with a GH14 cyclodextrin glycosyltransferase gives good correlation of log <italic>K</italic>_I with log <italic>k</italic>_cat/<italic>K</italic>_M, but also good correlation with log <italic>K</italic>_M suggesting both substrate and transition state mimicry [<xref rid="bib0300" ref-type="bibr">60</xref>].</p></sec><sec id="sec0030"><title>Conclusions</title><p id="par0210">A sophisticated view of glycoside hydrolase catalysis is now evident in which conformational changes occur that predispose substrates to react through oxocarbenium ion like transition states that are in accord with stereoelectronic and least nuclear motion principles. The challenges of these studies include the fact that X-ray crystal structures in complex with ligands by their very nature result in perturbation of the system for species nominally on or near the reaction coordinate, and for species off the reaction coordinate, great care needs to be taken to ensure that ground state conformational preferences do not bias interpretations. Kinetic isotope effect measurements and computational analysis can provide much needed help in assigning conformational itineraries. Compelling data is now available to assign conformational itineraries for a large number of GH families, yet as highlighted above, there are examples in which crystallographic data alone do not allow proposal of conformational itineraries. In these cases application of KIE analyses and theoretical approaches may help reveal a likely itinerary.</p><p id="par0215">There is a growing need for glycosidase inhibitors that exhibit selectivity against specific glycosidases, both to enable chemical biology approaches in glycobiology such as unravelling the roles of specific glycoside hydrolases in complex biochemical pathways [<xref rid="bib0305" ref-type="bibr">61</xref>], and in translational applications, for example, as folding chaperones for treatment of lysosomal storage disorders [<xref rid="bib0310" ref-type="bibr">62</xref>], and as enzyme inhibitors targeting aberrant glycosylation [<xref rid="bib0315" ref-type="bibr">63</xref>]. One of the long-term goals of conformational analysis of the glycosidase reaction coordinate is the hope that such information can inform the design of potent inhibitors, and in addition that these may be specific for particular conformational itineraries. While using such information in the design of inhibitors is not the primary focus of this review it is probably fair to say that as a general rule while the destination is now clear, the path to achieve this is not. Some success has been achieved with inhibitors that have particular conformational biases such as the selectivity of kifunensine for GH47 &#x003b1;-mannosidases. However, the crude attempts to achieve unusual conformations by structural means such as the introduction of bridges across the molecule are typically not tolerated by glycosidase active sites, although the 3,6-anhydrosugars processed by the GH117 &#x003b1;-<sc>l</sc>-neoagarobiose might constitute a logical target for applying such an approach. More generally, smarter, less intrusive ways are needed to control the conformation of inhibitors through the application of stereoelectronic principles and hybridization. A better understanding of the intrinsic conformational preferences of existing glycosidase inhibitors would greatly assist in directing these efforts.</p></sec><sec id="sec0035"><title>Conflicts of interest</title><p id="par0220">None declared.</p></sec><sec id="sec0040"><title>References and recommended reading</title><p id="par0225">Papers of particular interest, published within the period of review, have been highlighted as:<list list-type="simple" id="lis0015"><list-item id="lsti0035"><p id="par0230">&#x02022; of special interest</p></list-item><list-item id="lsti0040"><p id="par0235">&#x02022;&#x02022; of outstanding interest</p></list-item></list></p></sec>

Characterization of the contributions of Hp-MMP 9 to the serum acute phase protein response of lipopolysaccharide challenged calves

Could not extract abstract

<contrib contrib-type="author"><name><surname>Hinds</surname><given-names>Charles A</given-names></name><address><email>ahinds@uidaho.edu</email></address><xref ref-type="aff" rid="Aff1"/><xref ref-type="aff" rid="Aff4"/></contrib><contrib contrib-type="author"><name><surname>Niehaus</surname><given-names>Andrew J</given-names></name><address><email>Niehaus.25@osu.edu</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author"><name><surname>Premanandan</surname><given-names>Christopher</given-names></name><address><email>Premanandan.1@osu.edu</email></address><xref ref-type="aff" rid="Aff2"/></contrib><contrib contrib-type="author"><name><surname>Rajala-Schultz</surname><given-names>Paivi J</given-names></name><address><email>Rajala-Schultz.1@osu.edu</email></address><xref ref-type="aff" rid="Aff3"/></contrib><contrib contrib-type="author"><name><surname>Rings</surname><given-names>Donald M</given-names></name><address><email>Rings.1@osu.edu</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author" corresp="yes"><name><surname>Lakritz</surname><given-names>Jeffrey</given-names></name><address><email>Lakritz.1@osu.edu</email></address><xref ref-type="aff" rid="Aff1"/></contrib><aff id="Aff1"><label/>Department of Veterinary Clinical Sciences, College of Veterinary Medicine, The Ohio State University, 601 Vernon L Tharp Street, Columbus, Ohio 43210 USA </aff><aff id="Aff2"><label/>Department of Veterinary Biosciences, College of Veterinary Medicine, The Ohio State University, 1900 Coffey Road, Columbus, Ohio 43210 USA </aff><aff id="Aff3"><label/>Department of Veterinary Preventive Medicine, College of Veterinary Medicine, The Ohio State University, 1900 Coffey Road, Columbus, Ohio 43210 USA </aff><aff id="Aff4"><label/>Current address: University of Idaho, Caine Veterinary Teaching Center, 1020 East Homedale Road, Caldwell, ID 83607 USA </aff>

BMC Veterinary Research

<sec id="Sec1" sec-type="introduction"><title>Background</title><p>Bovine respiratory disease (BRD) and other acute inflammatory diseases have major impacts on livestock productivity. Recent reports indicate the incidence of BRD morbidity in large feedlots is approximately 5-11% and mortality attributed to BRD is approximately 0.6 &#x02013; 1.1% [<xref ref-type="bibr" rid="CR1">1</xref>]. However, 10% morbidity of received animals approximates 1 million head and 1% mortality is &#x0003e;10,000 head [<xref ref-type="bibr" rid="CR2">2</xref>,<xref ref-type="bibr" rid="CR3">3</xref>]. While there is some fluctuation in rates of disease, the incidence of BRD remains relatively constant despite application of thoughtful management decisions, development of newer pharmaceutical therapies and biological preventatives. Accurate and early diagnosis of diseases requiring administration of therapeutic agents would be beneficial to cattle industries.</p><p>Under experimental and field conditions, the use of APP to detect animals requiring treatment and animals developing lung lesions is useful. For example, haptoglobin (Hp) responses to inflammation in cattle have been evaluated in acute bronchopneumonia [<xref ref-type="bibr" rid="CR4">4</xref>&#x02013;<xref ref-type="bibr" rid="CR7">7</xref>], acute rumen acidosis [<xref ref-type="bibr" rid="CR8">8</xref>], coliform mastitis [<xref ref-type="bibr" rid="CR9">9</xref>,<xref ref-type="bibr" rid="CR10">10</xref>], hepatic lipidosis [<xref ref-type="bibr" rid="CR11">11</xref>], and transport stress [<xref ref-type="bibr" rid="CR12">12</xref>,<xref ref-type="bibr" rid="CR13">13</xref>]. Serum concentrations of Hp in acutely ill cattle increase (&#x0003e;100-fold) reaching maximum concentrations between 48 and 96&#x000a0;h; [<xref ref-type="bibr" rid="CR14">14</xref>&#x02013;<xref ref-type="bibr" rid="CR17">17</xref>] however, serum Hp appears to be a better indicator of clinical responses of calves with BRD to intervention than as a diagnostic for morbidity [<xref ref-type="bibr" rid="CR5">5</xref>,<xref ref-type="bibr" rid="CR18">18</xref>]. As the response of a particular APP demonstrate tremendous species and temporal differences, the gradual increase in serum concentrations of Hp over 24-48 hours, while dramatic, appear perhaps less sensitive than other APP in diagnosis of acute disease [<xref ref-type="bibr" rid="CR1">1</xref>,<xref ref-type="bibr" rid="CR14">14</xref>]. Other acute phase proteins, such as serum amyloid A (SAA) and alpha 1 acid glycoprotein (AGP) have also been studied in cattle undergoing LPS-challenge and experimental or naturally occurring disease [<xref ref-type="bibr" rid="CR1">1</xref>,<xref ref-type="bibr" rid="CR14">14</xref>,<xref ref-type="bibr" rid="CR19">19</xref>,<xref ref-type="bibr" rid="CR20">20</xref>]. Several studies indicate that LPS- challenge or experimental bacterial infection elicited earlier increases in serum SAA concentrations, suggesting SAA is more sensitive than Hp due to more rapid production and release [<xref ref-type="bibr" rid="CR1">1</xref>,<xref ref-type="bibr" rid="CR14">14</xref>,<xref ref-type="bibr" rid="CR19">19</xref>,<xref ref-type="bibr" rid="CR20">20</xref>]. Likewise, studies evaluating AGP after experimental LPS-challenge and <italic>M. hemolytica</italic> A1-challenge demonstrate more marked increases in serum concentrations of Hp and AGP after live bacterial challenge [<xref ref-type="bibr" rid="CR14">14</xref>]. In contrast, when serum Hp, SAA and AGP were evaluated under field conditions, serum concentrations of Hp were more useful in predicting the presence of respiratory disease and response to therapy, whereas SAA and AGP did not discriminate between animals which became sick and those that did not [<xref ref-type="bibr" rid="CR21">21</xref>].</p><p>In the previous work, we demonstrated that phorbol ester stimulation of isolated peripheral blood neutrophils are is associated with appearance of Hp-MMP 9 complexes in culture medium within 30&#x000a0;minutes [<xref ref-type="bibr" rid="CR22">22</xref>]. We also identified covalent, heteromeric complexes of Hp in complex with matrix metalloproteinase 9 (Hp-MMP 9), within the serum of cattle with clinically apparent acute onset of septic inflammation of the abdomen or thorax [<xref ref-type="bibr" rid="CR23">23</xref>]. In these cases, sepsis was associated with the presence of Hp-MMP 9 complexes when serum was analyzed by ELISA. In contrast to free serum Hp, whose main source is the liver during inflammation, serum Hp-MMP 9 complexes are only produced by neutrophils. As such, Hp-MMP 9 complexes, in serum, represent neutrophil degranulation [<xref ref-type="bibr" rid="CR22">22</xref>].</p><p>Intravenous LPS injection has been shown to produce physiologic and biochemical alterations in cattle including increases in APP (Hp [<xref ref-type="bibr" rid="CR14">14</xref>,<xref ref-type="bibr" rid="CR24">24</xref>], seromucoid [<xref ref-type="bibr" rid="CR14">14</xref>], ceruloplasmin [<xref ref-type="bibr" rid="CR14">14</xref>], &#x003b1;-1 proteinase inhibitor [<xref ref-type="bibr" rid="CR14">14</xref>], and SAA [<xref ref-type="bibr" rid="CR24">24</xref>]), decreased feed intake [<xref ref-type="bibr" rid="CR24">24</xref>&#x02013;<xref ref-type="bibr" rid="CR26">26</xref>], increased rectal temperature [<xref ref-type="bibr" rid="CR24">24</xref>,<xref ref-type="bibr" rid="CR25">25</xref>,<xref ref-type="bibr" rid="CR27">27</xref>], dyspnea [<xref ref-type="bibr" rid="CR24">24</xref>,<xref ref-type="bibr" rid="CR26">26</xref>,<xref ref-type="bibr" rid="CR27">27</xref>], increased cytokines (TNF&#x003b1; [<xref ref-type="bibr" rid="CR24">24</xref>&#x02013;<xref ref-type="bibr" rid="CR27">27</xref>], IL-1&#x003b2; [<xref ref-type="bibr" rid="CR27">27</xref>], IL-6 [<xref ref-type="bibr" rid="CR27">27</xref>], and IFN-&#x003b3; [<xref ref-type="bibr" rid="CR27">27</xref>]), increased cortisol [<xref ref-type="bibr" rid="CR25">25</xref>,<xref ref-type="bibr" rid="CR27">27</xref>]. We used this reproducible model to determine the time course of release of Hp-MMP 9 in comparison to other APP (Hp, SAA, AGP) after an acute LPS challenge.</p><p>As neutrophils play key roles in the early onset of bovine diseases, we sought to evaluate a biomarker specific to neutrophils for monitoring very early inflammation responses in cattle. The kinetics of most APP involves recognition of pathogens/pathogen products, mediator production and release, gene expression and protein synthesis and release of protein into the circulation [<xref ref-type="bibr" rid="CR28">28</xref>&#x02013;<xref ref-type="bibr" rid="CR32">32</xref>]. Release of neutrophil granule proteins is rapid after <italic>in vitro</italic> stimulation, occurring within 30&#x000a0;minutes of phorbol ester treatment [<xref ref-type="bibr" rid="CR22">22</xref>]. The objective of the present study was to characterize the time course of serum Hp-MMP 9 complex appearance in relation to changes in the hemogram, serum C concentrations and by comparing with other acute phase proteins in calves (Hp, SAA, AGP).</p><p>We propose that Hp-MMP 9 complexes, observed after phorbol ester stimulation of isolated peripheral blood neutrophils <italic>in vitro</italic> and found in acute phase sera, have specific functional significance differing from un-complexed forms of Hp or MMP 9 alone and as a consequence, serum concentrations of Hp-MMP 9 may serve as an independent indicator of clinically important events occurring during acute inflammation.</p></sec><sec id="Sec2" sec-type="materials|methods"><title>Methods</title><sec id="Sec3"><title>Experimental design</title><p>The following experimental protocol was approved by The Ohio State University, Laboratory Animal Care and Use Committee. The study group consisted of 9 healthy Jersey bull calves between 65-82 days of age (average weight &#x02013; 75&#x02009;&#x000b1;&#x02009;13&#x000a0;kg; range &#x02013; 58-100&#x000a0;kg), acclimated to grass hay and 0.7&#x000a0;kg mixed grain diet for 7&#x000a0;days prior to the start of the study. The calves were born at The Ohio State University, Waterman Dairy Farm and were transferred to our facility for use in these studies. Calves were housed in groups of 2 animals, with one group of 3 calves, in climate controlled stalls (average temperature 21.7&#x000b0;C). Physical exams were performed daily throughout the study and body weights were monitored weekly. Twelve hours prior to LPS challenge, calves were fitted with indwelling jugular venous catheters^a after clipping the hair and aseptic cleansing the skin using 1% iodine scrub followed by 70% isopropyl alcohol. A 30&#x000a0;cm extension line with infusion port was attached and the catheter was held in place with elastic tape. The catheter was flushed with heparinized saline. Blood was collected from the catheter for CBC and serum biochemical analysis. Serum was collected from clotted whole blood that was centrifuged at 2,000 &#x000d7; g for 20&#x000a0;minutes after clotting at room temperature. The serum was removed and stored at -80&#x000b0;C until analysis.</p><p>At time (T) =0&#x000a0;hours, a physical exam was performed and blood was collected for a CBC and for serum collection which was stored for later analysis. All rectal temperature measurements were made using a digital thermometer.^b Lipopolysaccharide (LPS),^c 2.5&#x000a0;&#x003bc;g/kg body weight, that was diluted in 10&#x000a0;ml of autologous serum and allowed to incubate at 37&#x000b0;C for 30&#x000a0;minutes. After incubation, the LPS solution was administered rapidly via the IV catheter. After administration, each catheter was flushed with 10&#x000a0;mL heparinized saline (10&#x000a0;IU/mL). Physical exams and blood collection were performed at T&#x02009;=&#x02009;-24, 0, 0.5, 1, 1.5, 2, 3, 4, 6, 8, 12, 16, 24, 36, 48, 72, 96&#x000a0;hr post-LPS infusion. After each blood collection, equal volume of normal saline (0.9% NaCl for injection) was administered to maintain blood volume. IV catheters were removed after the last collection time point.</p></sec><sec id="Sec4"><title>Serum and blood analyses</title><p>All serum analyses were performed using commercially available ELISA kits (Hp, SAA), single radial immunodiffusion (AGP, SRID) according to the manufacturers recommendations, whereas the Hp-MMP 9 ELISA is an in house laboratory assay. Serum C concentrations were determined by use of a solid phase, competitive chemiluminescent enzyme immunoassay and an automated analysis system as described^d [<xref ref-type="bibr" rid="CR33">33</xref>] by an accredited veterinary clinical pathology laboratory. Briefly, aliquots of each calf&#x02019;s serum were placed into individual test units for analysis. The calibration range of the assay is 28 &#x02013; 1,380 nmol/L and analytical sensitivity is 5 nmol/L. All samples from each calf were analyzed in a single run.</p><p>Serum total Hp was determined using a commercially available bovine Hp ELISA kit.^e The analysis was conducted according to the manufacturer&#x02019;s instructions. Specifically, all serum samples from all calves were diluted 1:2,000 in sample buffer prior to analysis. Serum concentrations were determined from the concentration vs. absorbance relationship of the standard haptoglobin concentrations (7.8-250 ng/mL). All calf serum sample concentrations were corrected for dilution (2,000 fold dilution). Analytical variation between samples on the same day and on multiple days is &#x0003c;8.8 and 12.9% respectively.</p><p>Bovine Hp-MMP 9 complexes were determined as described previously [<xref ref-type="bibr" rid="CR23">23</xref>]. All serum samples were diluted 1:5 with sample diluent (TBS +1% Bovine serum albumin +0.05% Tween 20. After blocking the wells, known concentrations of Hp-MMP 9 (serum, pre-characterized and shown to contain ~912.6&#x000a0;ng/mL Hp-MMP 9) and the LPS challenged calf serum samples were added to wells. If sample absorbance fell outside of the linear portion of the concentration-absorbance line, samples were further diluted to ensure linearity. Between plate variability of calibrators from 5 different plates were less than 3% (median =1.8%; range 0.98-2.7%). The average coefficient of correlation determined from linear regression of the absorbance versus concentration of calibrator was 0.91 (range 0.85 &#x02013; 0.95). The analytical sensitivity of the assay is 3.5&#x000a0;ng/mL.</p><p>Serum concentrations of SAA were determined using a commercial multi-species ELISA (&#x0201c;Phase&#x0201d; Serum Amyloid A assay^f) used according to the manufacturer&#x02019;s instructions. All serum samples were diluted 1:500 with sample diluent buffer, and 50 &#x003bc;L of sample or calibrator were added to each well containing the detection antibody and the absorbance determined on a plate reader.^g The intra- and inter assay coefficients of variation for the assay were &#x0003c;11% and the analytical sensitivity of the bovine assay is 0.3 &#x003bc;g/mL.</p><p>Serum alpha₁ acid-glycoprotein (AGP) concentrations were determined using a commercial single radial immuno-diffusion assay.^h After addition of 5 &#x003bc;L each, of AGP calibrators (1,000 &#x003bc;g/mL, 250 &#x003bc;g/mL, 125 &#x003bc;g/mL) and calf serum samples to individual wells on each plate, the plates were incubated for 48 hours in a humidified container at 37&#x000b0;C. After incubation, the plates were imaged on a light table and the diameter of the rings measured using a 10x scale loupe with metric reticule. The diameter of precipitin rings of calibrators was plotted against the concentration of AGP to obtain an equation of the line. Repeated assay of the calibrators on 27 unique plates, produced coefficient&#x02019;s of variation between 2.7-3.5% over the range of the calibrators (125-1000 &#x003bc;g/mL), the average coefficient of correlation was 0.997.</p></sec><sec id="Sec5"><title>Statistical analysis</title><p>Data from the LPS challenge study (temperature, pulse, respiration, WBC count (total WBC, neutrophil counts, band neutrophil counts, lymphocyte counts) differential count, serum concentrations of C, Hp, Hp-MMP 9 complex, SAA, and AGP concentrations) were tabulated by time point and evaluated graphically in a commercial spreadsheet.ⁱ After visual comparison of the data, the changes in physiologic variables (temperature, heart rate, respiration), number of peripheral white blood cells, differential counts (total WBC, PMN, band, lymphocyte) and concentrations of the analytes (C, Hp, Hp-MMP 9 complexes, SAA, AGP) over time were examined using PROC MIXED in SAS (v.9.3).^j To account for the correlated data structure of the repeated measures from individual calves over time, four covariance structures were tested (compound symmetry, first order autoregressive, heterogeneous first order autoregressive, and unstructured). Time was included in the model as the main variable of interest to evaluate how the different parameters changed in response to the LPS challenge. Baseline (T = 0 hr) was used as the reference level and significance was set at <italic>P</italic> &#x02264;0.05. All measurements at different time-points were compared with the baseline value at T = 0. First order autoregressive covariance structure fitted the data best for SAA, AGP and C as outcomes, compound symmetry covariance structure was used with other outcomes. The areas under the concentration-time (AUC), for each acute phase protein analyte (Hp, Hp-MMP 9, SAA and AGP) were calculated using standard formulae, from time =0 to time =96 hours. No extrapolation of the terminal portion of the curve was conducted. The ratio of AUC_{Hp-MMP 9} (ng*hr/mL), to each of the other analytes (Hp, SAA and AGP) were expressed as a percentage and compared using the Kruskal-Wallis, 1 way ANOVA on ranks with Dunn&#x02019;s multiple comparison test. Differences in ranks were considered significant when p &#x0003c; 0.05 for each comparison.</p></sec></sec><sec id="Sec6" sec-type="results"><title>Results</title><p>The most consistent clinical indicators of illness were tachypnea and dyspnea developing within 30&#x000a0;minutes after LPS infusion (p&#x02009;&#x0003c;&#x02009;0.001 in comparison to T&#x02009;=&#x02009;-24&#x000a0;hours; Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>). Respiratory rates remained significantly greater than baseline until 4&#x000a0;h post-LPS and were not significantly greater than baseline by 6&#x000a0;h post-LPS (p&#x02009;=&#x02009;0.1542; Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>). Changes in rectal temperature and heart rate in these calves were minimal and not significantly different from baseline (Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>). Significant changes in respiratory rate, occurred with a marked reduction in peripheral leukocytes and rapid increases in serum C (Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>). Baseline serum C concentrations were 27.6&#x02009;&#x000b1;&#x02009;7.8&#x000a0;nmol/L (range: 30 &#x02013; 49.7&#x000a0;nmol/L) at time =0&#x000a0;hours and peaked at 174&#x02009;&#x000b1;&#x02009;40&#x000a0;nmol/L (range: 102-254&#x000a0;nmol/L) by 3&#x000a0;hours post-LPS infusion (p&#x02009;&#x0003c;&#x02009;0.0001; Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>). Total white blood cells (WBC), lymphocytes and neutrophil counts declined from baseline, reaching a nadir at 4&#x000a0;hours (p&#x02009;&#x0003c;&#x02009;0.0001) post-LPS (Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>). Total WBC returned to levels that were not different from baseline by 24-36 hours post-LPS challenge (Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>). Peripheral neutropenia was associated with the appearance of &#x0201c;band&#x0201d; PMN at 8&#x000a0;hours (0.2&#x02009;&#x000b1;&#x02009;0.36 &#x000d7; 10 [<xref ref-type="bibr" rid="CR9">9</xref>]/L; range: 0 &#x02013; 1.2 &#x000d7; 10 [<xref ref-type="bibr" rid="CR9">9</xref>]/L; P&#x02009;&#x0003c;&#x02009;0.0007; data not shown) and remained significantly greater than baseline until 24&#x000a0;hours post LPS (0.66&#x02009;&#x000b1;&#x02009;1.6 &#x000d7; 10 [<xref ref-type="bibr" rid="CR9">9</xref>]/L; p&#x02009;&#x0003c;&#x02009;0.0001 compared to baseline; data not shown). Other peripheral leukocyte types (monocytes, eosinophils, basophils) did not change significantly throughout the study period.<fig id="Fig1"><label>Figure 1</label><caption><p> <bold>Mean&#x02009;&#x000b1;&#x02009;SD heart rate (beats/minute), respiratory rate (breaths/minute), rectal temperarture (&#x000b0;C) and serum C concentrations (nmol/L) observed after intravenous administration of</bold> <bold><italic>E. coli</italic></bold> <bold>LPS (O111:B4; 2.5 ug/kg solubilized in autologous serum).</bold> Statistically higher respiratory rates were observed by 0.5&#x000a0;hours post-LPS (p&#x02009;&#x0003c;&#x02009;0.001) and remained higher than baseline (-24&#x000a0;hour time point) until 6&#x000a0;hours post LPS infusion (p&#x02009;&#x0003c;&#x02009;0.0034). Serum C values were significantly greater than baseline by 1&#x000a0;hour post LPS (p&#x02009;&#x0003c;&#x02009;0.0001). There was no statistically significant change in the HR and RT for these calves.</p></caption><graphic xlink:href="12917_2014_261_Fig1_HTML" id="MO1"/></fig><fig id="Fig2"><label>Figure 2</label><caption><p> <bold>Mean&#x02009;&#x000b1;&#x02009;SD total white blood cell counts, lymphocyte count and neutrophil counts in peripheral blood of Jersey calves immediately prior to and up to 96&#x000a0;hours after an intravenous bolus dose of</bold> <bold><italic>E. coli</italic></bold> <bold>LPS (O111:B4; 2.5 ug/Kg solubilized in autologous serum).</bold> Plasma WBC were significantly lower than pre-LPS (-24&#x000a0;hr time point) from 0.5&#x000a0;hour &#x02013; 12&#x000a0;hour time points (p&#x02009;&#x0003c;&#x02009;00001). Plasma lymphocyte and neutrophil counts showed dramatic drop from the 0.5&#x000a0;hour &#x02013; 16&#x000a0;hour post LPS challenge time points (P&#x02009;&#x0003c;&#x02009;0.05).</p></caption><graphic xlink:href="12917_2014_261_Fig2_HTML" id="MO2"/></fig></p><p>Administration of a single intravenous dose of LPS resulted in increased serum Hp concentrations. Serum concentrations of total Hp were &#x0003c;10 &#x003bc;g/mL until 8 hours post-LPS infusion (Figure&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>A), and increased to significantly greater than baseline by 12 hours post-LPS infusion (127 &#x000b1; 100 &#x003bc;g/mL; range: 0-368 &#x003bc;g/mL; p = 0.0006) (Figure&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>A). Maximum serum concentrations of Hp were 400 &#x000b1; 73 &#x003bc;g/mL (range: 257-460 &#x003bc;g/mL) 36 hours post-LPS infusion in all 9 experimental animals (Figure&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>A). Serum concentrations of Hp remained significantly greater than baseline through the 96 hour time point (p = 0.0087) in 2 of the nine calves; the remaining 7/ 9 calves had serum haptoglobin &#x0003c;30 &#x003bc;g/mL at the 96 hour time point. The increase in serum Hp concentrations occurred after resolution of clinical signs and was associated with return of serum C to baseline concentrations, increasing peripheral leukocyte counts.<fig id="Fig3"><label>Figure 3</label><caption><p> <bold>Mean &#x000b1; SD of serum concentrations of Hp-MMP 9, total Hp, SA A and serum AGP in peripheral blood of Jersey calves after an intravenous bolus dose of</bold> <bold><italic>E. coli</italic></bold> <bold>LPS (O111:B4; 2.5 ug/Kg solubilized in autologous serum).</bold> Concentrations of serum Hp-MMP 9 <bold>(B)</bold> is several orders of magnitude lower than total haptoglobin (Hp) <bold>(A)</bold> and far lower than SAA <bold>(C)</bold> or AGP <bold>(D)</bold>. Serum concentrations of Hp-MMP 9 are in ng/mL. Serum concentrations of Hp, SAA and AGP were converted to ng/mL for comparison to Hp-MMP 9 and as shown, the serum concentrations are reported as logarithm (base 10). Significant increases in serum Hp were detected at 12 hours post-LPS challenge in comparison to pre-LPS time points (-24 hour time points; p &#x0003c; 0.0006) and remained significant through the 96 hour sampling time (p &#x0003c; 0.0087 compared to -24 hour time point). Serum Hp-MMP 9 complex concentrations appeared by 0.5 hrs and were significantly greater than baseline at 16 hour post-LPS challenge (p &#x0003c; 0.008 compared to -24 sample) and returned to baseline values after the 48 hour sampling time (72 hour; p = 0.1254 compared to -24 hour time point). Serum AGP increased to significantly greater than the -24 hour sampling point at 16 hours post-LPS challenge (p &#x0003c; 0.05) and remained significantly greater than baseline through the 96 hour post-LPS time point (p &#x0003c; 0.007). Serum amyloid A concentrations were significantly greater than baseline (-24 hour time point) by 8 hours post-LPS (p &#x0003c; 0.05) and remained greater than baseline through the 96 hour time point (p &#x0003c; 0.05, compared to -24 hour time point).</p></caption><graphic xlink:href="12917_2014_261_Fig3_HTML" id="MO3"/></fig></p><p>Baseline concentrations of AGP were 591 &#x000b1; 124 &#x003bc;g/mL and did not change significantly until 16 hours post-LPS challenge when serum AGP concentrations were 843 &#x000b1; 452 ug/mL (p = 0.0434) (Figure&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>D). Serum AGP concentrations reached peak levels 24 hours post-LPS (920 &#x000b1; 466 &#x003bc;g/mL; p = 0.0065). Changes in serum AGP concentration also occurred after clinical signs and serum C concentrations returned to baseline and peripheral blood total WBC and neutrophil counts returned to baseline post LPS (p &#x0003c; 0.0001). Serum concentrations of AGP remained &#x0003e;800 &#x003bc;g/mL throughout the remainder of the study period.</p><p>Serum concentrations of SAA responded to LPS challenge increasing from 29 &#x000b1; 35 &#x003bc;g/mL at baseline to 162 &#x000b1; 121 &#x003bc;g/mL; 12 h post-LPS challenge; p = 0.0049), and remained significantly greater than baseline at 96 hours post-LPS infusion (98.4 &#x000b1; 72 &#x003bc;g/mL; p = 0.03) (Figure&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>C). Like serum concentrations of Hp and AGP, changes in SAA were significantly greater than baseline by 8 hours and did not correspond to changes in clinical signs, serum C and peripheral WBC counts. Serum amyloid A concentrations peaked 12 hours post LPS, and remained above baseline concentrations through the end of the study.</p><p>Serum concentrations of Hp-MMP 9 complexes were detectable (&#x0003e;3.5 ng/mL) from 0.5 &#x02013; 12 hours post-LPS, although these values were not significantly different from baseline (p = 0.82) (Figure&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>B). Serum concentrations of Hp-MMP 9 continued to increase until 36 hrs post-challenge, reaching concentrations of 665 &#x000b1; 308 ng/mL (range 298-1311 ng/mL; p &#x0003c; 0.0001). Unlike the other APP markers (Hp, SAA, AGP), serum concentrations of Hp-MMP 9 decreased to concentrations that were not significantly different from baseline by 72 hours post-LPS infusion (p = 0.1254). Serum concentrations of Hp-MMP 9 were &#x0003c;55 ng/mL in 7/9 calves at 96 hours.</p><p>The average exposure (AUC), to LPS induced serum acute phase proteins demonstrated that exposure (concentration &#x000d7; time) to Hp-MMP 9 differed significantly from that of Hp (p&#x02009;&#x0003c;&#x02009;0.05) and AGP (p&#x02009;&#x0003c;&#x02009;0.001), but not SAA Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>). Area under the curve for Hp-MMP 9 (AUC_{Hp-MMP 9}) was 0.14&#x02009;&#x000b1;&#x02009;0.05% of that for Hp, 0.13&#x02009;&#x000b1;&#x02009;0.06% of that for SAA and 0.05&#x02009;&#x000b1;&#x02009;0.03% of that for AGP (P&#x02009;&#x0003c;&#x02009;0.01 for Hp-MMP 9 compared to Hp and AGP responses; Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>). The AUC_AGP was significantly greater than AUC_SAA (p&#x02009;&#x0003c;&#x02009;0.01).<fig id="Fig4"><label>Figure 4</label><caption><p> <bold>Total systemic exposure of animal to serum acute phase proteins. Serum concentrations of serum acute phase proteins over time were converted to ng/mL and the area under the curve was calculated using the trapezoidal rule.</bold> Data for each analyte are reported as AUC_{0-96 hr} and plotted on log-linear graph. Area under the curve of serum Hp-MMP 9 was significantly lower than all 3 other serum acute phase proteins (p&#x02009;&#x0003c;&#x02009;0.01) when evaluated using Kruskal-Wallis, 1 way ANOVA on ranks with Dunn&#x02019;s multiple comparison test. There was no significant difference between total serum Hp and SAA over the time course of analysis and both SAA and Hp had significantly lower exposure in comparison to the areas observed for AGP. Differences in the letters over bars indicate significant differences between AUC for each analyte.</p></caption><graphic xlink:href="12917_2014_261_Fig4_HTML" id="MO4"/></fig></p></sec><sec id="Sec7" sec-type="discussion"><title>Discussion</title><p>The purpose of this study was to evaluate the time course of a neutrophil biomarker (Hp-MMP 9 complexes) after LPS challenge in calves in comparison to responses of established serum APP markers in cattle. Using a single challenge stimulus producing consistent clinical, hematologic, physiologic and APP responses, the concentration versus time profile of this neutrophil biomarker may prove useful to define the activation of neutrophils early in the course of inflammation. While not fully replicating the major clinical and pathologic findings associated with natural BRD, the characteristics of the LPS stimulus should allow dissection of neutrophil from other APP responses occurring in inflammation [<xref ref-type="bibr" rid="CR27">27</xref>,<xref ref-type="bibr" rid="CR34">34</xref>,<xref ref-type="bibr" rid="CR35">35</xref>]. In clinical terms, we believe early and specific detection of neutrophil activation should provide an indication of early responses to infection. Further, this response is distinct from that of the liver [<xref ref-type="bibr" rid="CR36">36</xref>].</p><p>Bovine neutrophils have been shown to play key roles in host responses to infection [<xref ref-type="bibr" rid="CR34">34</xref>]. Experimental bacterial pneumonia models have demonstrated rapid recruitment of neutrophils to the lung is associated with tachypnea, dyspnea, hypercortisolemia, peripheral leukopenia and influx of neutrophils into the lungs [<xref ref-type="bibr" rid="CR27">27</xref>,<xref ref-type="bibr" rid="CR34">34</xref>,<xref ref-type="bibr" rid="CR35">35</xref>,<xref ref-type="bibr" rid="CR37">37</xref>&#x02013;<xref ref-type="bibr" rid="CR40">40</xref>]. Our previous studies demonstrated a unique form of neutrophil matrix-metalloproteinase 9, covalently linked to haptoglobin (Hp-MMP 9) are stored in and released by neutrophils; and these complexes are present within the serum of cattle with acute, polymicrobial sepsis [<xref ref-type="bibr" rid="CR22">22</xref>,<xref ref-type="bibr" rid="CR23">23</xref>]. Therefore, as a unique marker produced by the neutrophil, serum Hp-MMP 9 complexes should herald early changes in neutrophils associated with the host response to inflammation.</p><p>Both natural and experimental infection studies demonstrate the appearance of APP supporting their use in detection of inflammation and response to therapy [<xref ref-type="bibr" rid="CR1">1</xref>,<xref ref-type="bibr" rid="CR14">14</xref>,<xref ref-type="bibr" rid="CR18">18</xref>,<xref ref-type="bibr" rid="CR19">19</xref>,<xref ref-type="bibr" rid="CR21">21</xref>,<xref ref-type="bibr" rid="CR24">24</xref>,<xref ref-type="bibr" rid="CR41">41</xref>&#x02013;<xref ref-type="bibr" rid="CR47">47</xref>]. Although some APP have proven useful in the evaluation of illness in cattle, their contributions to the bovine serum acute phase response, represent contributions of hepatocytes or other tissues in response to proximate mediators produced by other cells [<xref ref-type="bibr" rid="CR1">1</xref>,<xref ref-type="bibr" rid="CR14">14</xref>,<xref ref-type="bibr" rid="CR44">44</xref>,<xref ref-type="bibr" rid="CR45">45</xref>]. Timing of APP responses vary by APP marker and do not necessarily represent events occurring in the sub-clinical phase of disease such as when changes in clinical signs, serum C and peripheral leukopenia occur. Upon activation, peripheral blood neutrophils are a source of many proteins, including matrix metalloproteinase 9, Hp and AGP. These neutrophil APP in serum are a part of the acute phase proteome; however, current assays do not identify sources of these proteins [<xref ref-type="bibr" rid="CR48">48</xref>&#x02013;<xref ref-type="bibr" rid="CR50">50</xref>].</p><p>As a component of neutrophil granules, Hp-MMP 9 should not be present within the circulation of healthy animals and prior studies demonstrated Hp-MMP 9 was not present in the serum of healthy cows [<xref ref-type="bibr" rid="CR23">23</xref>]. After an intravenous dose of LPS, serum concentrations of Hp-MMP 9 were above the limit of detection of our ELISA (&#x0003e;3.5&#x000a0;ng/mL) by 1&#x000a0;hour. Serum Hp-MMP 9 complexes are detected when changes in respiratory rate, serum C and WBC numbers occur. We have also demonstrated cattle undergoing experimental bacterial pneumonia and naturally occurring cases of acute poly-microbial sepsis were associated with increased serum concentrations of Hp-MMP 9 complexes [<xref ref-type="bibr" rid="CR23">23</xref>,<xref ref-type="bibr" rid="CR24">24</xref>]. Experimental bacterial infection is associated with lung lesions consistent with <italic>Mannheimia hemolytica</italic> infection in these calves and with rapid increases in serum concentrations of Hp and Hp-MMP 9 complexes [<xref ref-type="bibr" rid="CR51">51</xref>]. Similarly, transportation is associated with increased incidence of respiratory disease and Hp is proposed to be a marker of non-inflammatory stress [<xref ref-type="bibr" rid="CR52">52</xref>,<xref ref-type="bibr" rid="CR53">53</xref>]. Stress induced induction of serum Hp still involves release of production of cytokines and hepatic expression of Hp and other APP. [<xref ref-type="bibr" rid="CR53">53</xref>] A recent study demonstrated the appearance of Hp-MMP 9 in serum after transportation suggesting transportation stress is also associated with measurable responses of neutrophils [<xref ref-type="bibr" rid="CR52">52</xref>,<xref ref-type="bibr" rid="CR54">54</xref>]. These studies support the further evaluation of Hp-MMP 9 complex as a marker of early inflammation.</p><p>Both neutrophil Hp-MMP 9 and SAA showed rapid increases in serum concentrations in LPS challenged calves, consistent with earlier studies [<xref ref-type="bibr" rid="CR1">1</xref>,<xref ref-type="bibr" rid="CR19">19</xref>,<xref ref-type="bibr" rid="CR45">45</xref>,<xref ref-type="bibr" rid="CR55">55</xref>]. However, individual calf data for SAA varied somewhat, making them significantly different from baseline only by 8&#x000a0;hours post-LPS. However, serum Hp-MMP 9, and SAA concentrations observed after LPS challenge increased concurrently with onset of tachypnea, peak serum C concentrations and reduction of circulating leukocytes. Our results are consistent with these findings; however, we observed sustained SAA concentrations to at least 96&#x000a0;hours post-LPS challenge [<xref ref-type="bibr" rid="CR1">1</xref>,<xref ref-type="bibr" rid="CR19">19</xref>,<xref ref-type="bibr" rid="CR27">27</xref>,<xref ref-type="bibr" rid="CR43">43</xref>,<xref ref-type="bibr" rid="CR45">45</xref>,<xref ref-type="bibr" rid="CR55">55</xref>,<xref ref-type="bibr" rid="CR56">56</xref>]. Other studies conducting i.v. LPS challenge, characterized the induction of SAA out to 10&#x000a0;hours post-challenge [<xref ref-type="bibr" rid="CR1">1</xref>,<xref ref-type="bibr" rid="CR27">27</xref>,<xref ref-type="bibr" rid="CR45">45</xref>]. Prolonged elevations of SAA observed in our study suggest continued production after LPS stimulation. It is plausible then, that use of SAA in animals may be useful in diagnosis of acute inflammation. However, most of the SAA produced in association with LPS-induced inflammation also represents hepatic responses [<xref ref-type="bibr" rid="CR56">56</xref>].</p><p>In contrast, the clinical, hematologic and C responses occurring in our calves after intravenous LPS challenge preceded increased serum Hp and AGP concentrations by several hours. Serum Hp concentrations remained normal until 8 hours; however, 5/9 had no detectable Hp until 12 hours post LPS challenge. Similarly, serum AGP concentrations were observed to increase significantly at 16 hours post LPS. As previous studies demonstrate, serum Hp and AGP concentrations remain significantly greater than baseline until the end of the study period, well after the LPS induced changes in respiratory function, serum C and leukopenia had resolved [<xref ref-type="bibr" rid="CR18">18</xref>&#x02013;<xref ref-type="bibr" rid="CR20">20</xref>]. Like Hp, Hp-MMP 9, AGP is also produced by bovine neutrophils [<xref ref-type="bibr" rid="CR49">49</xref>]. It is plausible that serum AGP concentrations as detected with currently available methods may reflect contributions from the neutrophil. However, these methods do not specifically identify this form autonomously of others present.</p><p>Serum APP terminal half-life is been proposed to be useful in terms of effectiveness of empirical antimicrobial therapy for pneumonia in humans [<xref ref-type="bibr" rid="CR57">57</xref>]. The serum concentration versus time profile of Hp-MMP 9 (AUC) was smaller than that of other APP measured. Area&#x02019;s under the Hp-MMP 9, concentration-time curve (AUC_{HP-MMP 9}) in our calves were 0.14% of that measured for serum total Hp (AUC_HP), 0.13% of AUC_SAA and 0.05% of AUC_AGP. This suggests that the animal&#x02019;s exposure to Hp-MMP 9 is much lower than that for Hp, SAA and AGP. This seems plausible since neutrophils and their granule proteins are limited in comparison to hepatic APP induced by inflammatory mediators. Characterization of the elimination processes of each APP may help to define the half-life of these proteins after induction. This information may provide a means for evaluation of the rate of normalization associated with therapy.</p></sec><sec id="Sec8" sec-type="conclusion"><title>Conclusions</title><p>As Hp-MMP 9 is detectable early after a consistent inflammatory stimulus when animals develop clinical signs, we believe that like SAA, it may serve as a useful marker of early inflammation. The serum content of this neutrophil protein complex induced by inflammation is limited in extent and is masked by APP produced by the liver. Dissecting the contribution of the neutrophil to the APP responses produced by other sources may prove useful as an adjunct to the clinical examination after arrival at the feedlot, for pre-slaughter exams and other situations. The availability of a test providing objective data regarding the status of the neutrophil, so intimately involved in acute inflammation, may have potential in clinical algorithms and decision making used in diseased cattle.</p></sec><sec id="Sec9"><title>Endnotes</title><p>^aAngiocath, 16GA, 3.5 inch; Becton Dickinson, Sandy, Utah</p><p>^bGLA M700 Digital Thermometer; GLA Agricultural electronics, San Luis Obispo, CA 93401</p><p>^c<italic>Escherichia coli</italic> O111:B4; L2630 Sigma-Aldrich, St. Louis, MO.</p><p>^dIMMULITE 1000 Cortisol, Immunoassay system. DPS, Los Angeles, CA</p><p>^eLife diagnostics, Haptoglobin ELISA test kit; catalogue #2410-7. West Chester, PA 19380; <ext-link ext-link-type="uri" xlink:href="http://www.lifediagnostics.com/">www.lifediagnostics.com</ext-link></p><p>^fMultispecies Cat no. TP-802; Tri-Delta Development LTD, Kildare, Ireland</p><p>^gLabsystems, Multiskan MS; P97180; Vienna, VA, 22182</p><p>^hCardiotech services, P0101-1; Louisville KY 40205</p><p>ⁱMicrosoft Excel, Microsoft, Corp. Redmond, WA 98052-6399</p><p>^jSAS, Version 9.2, SAS Institute, Cary NC 27513-2414; <ext-link ext-link-type="uri" xlink:href="http://www.sas.com/">www.sas.com</ext-link></p></sec>

Violations of local stochastic independence exaggerate scalability in Mokken scaling analysis of the Chinese Mandarin SF-36

Could not extract abstract

<contrib contrib-type="author"><name><surname>Watson</surname><given-names>Roger</given-names></name><address><email>r.watson@hull.ac.uk</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author"><name><surname>Wang</surname><given-names>Wenru</given-names></name><address><email>nurww@nus.edu.sg</email></address><xref ref-type="aff" rid="Aff2"/></contrib><contrib contrib-type="author" corresp="yes"><name><surname>Thompson</surname><given-names>David R</given-names></name><address><email>david.thompson@acu.edu.au</email></address><xref ref-type="aff" rid="Aff3"/><xref ref-type="aff" rid="Aff4"/><xref ref-type="aff" rid="Aff5"/></contrib><aff id="Aff1"><label/>Faculty of Health and Social Care, University of Hull, Hull, HU6 7RX UK </aff><aff id="Aff2"><label/>Alice Lee Centre for Nursing Studies, National University of Singapore, Singapore, Singapore </aff><aff id="Aff3"><label/>Centre for the Heart and Mind, Australian Catholic University, Melbourne, VIC 3000 Australia </aff><aff id="Aff4"><label/>Department of Psychiatry, University of Melbourne, Melbourne, VIC 3050 Australia </aff><aff id="Aff5"><label/>Department of Epidemiology and Preventive Medicine, Monash University, Melbourne, VIC 3800 Australia </aff>

Health and Quality of Life Outcomes

<sec id="Sec1" sec-type="introduction"><title>Background</title><sec id="Sec2"><title>Local stochastic independence</title><p>Questionnaires are commonly used to measure quality of life and a prime example is the Short Form 36-item health survey (SF-36). Questionnaires are comprised of a series of questions, commonly referred to as items, which measure a latent trait (eg quality of life) and in larger questionnaires it is common for sets of items to be grouped under themes or subscales which purport to measure different aspects of the latent trait (eg physical health and mental health). The relationship between items may be merely conceptual but this is commonly supported mathematically using any one of a range of methods under the umbrella of multivariate statistics. These methods, essentially, study the way sets of items correlate or covary. However, some assumptions about the relationship between items should be met and one of these is local stochastic independence (LSI).</p><p>In the study of latent variables, LSI is a crucial property [<xref ref-type="bibr" rid="CR1">1</xref>], whether these are being studied using methods under the umbrella of classical test theory (CTT), for example factor analysis, or item response theory (IRT), for example Rasch analsyis. When LSI can be assumed, any observed relationships between items in a questionnaire&#x02014;covariance&#x02014;can be assumed to be a result of the latent trait being measured rather than some other property of the items such as overlap between items or formation of items chains among items. However, LSI is usually assumed rather than estimated and, in fact, for some it is a hard concept to grasp: both its nature and its necessity.</p><p>Confusion over the nature of LSI arises because items forming scales have to covary [<xref ref-type="bibr" rid="CR2">2</xref>] and this can imply some stochastic dependence; in other words, that variance in one item depends on the variance in another. However, the assumption of LSI implies that the item covariance is independent and, specifically, means that the observed covariance of items in a scale is a result only of the latent trait that they measure [<xref ref-type="bibr" rid="CR3">3</xref>]. Therefore, the endorsement of any specific item is independent of the responses to other items on the latent trait. For example, if the underlying latent trait is &#x02018;tendency to become depressed&#x02019;, the items &#x02018;I don&#x02019;t feel like getting out of bed in the morning&#x02019; and &#x02018;I feel like a worthless person&#x02019; may both be incorporated into a scale; however, the response to one item is not dependent on the other and, theoretically, both items could be responded to by anyone without endorsing the other. In reality, when they are included in a series of questions about depression, both are likely to be endorsed. In IRT, where an order in the way items are endorsed is assumed [<xref ref-type="bibr" rid="CR4">4</xref>], the ordering of the items is also considered to be stochastically independent. The order of responses to items is a measure of the latent trait but it is not necessary to respond to one item before or after another; if items are ordered then it is assumed to be a result of the latent trait [<xref ref-type="bibr" rid="CR1">1</xref>]. We will expand on this below under the description of Mokken scaling.</p><p>The above explains the nature of LSI. The importance of the concept arises from the fact that, without assuming LSI, it is not possible to assume anything about a putative latent trait, only about some other relationship&#x02014;not necessarily a result of the latent trait&#x02014;between a set of items. Item dependence arises, as explained by Bakazs and de Boeck [<xref ref-type="bibr" rid="CR5">5</xref>] from two main sources: item chains; and item overlap. When items are in a chain, the success of any item may depend on a previous item; when items overlap, they include very similar concepts.</p><sec id="Sec3"><title>Item chains</title><p>An item chain may arise, for example, in tests of calculation where it is essential to compute the answer to one question and use that value in another question and, conversely, impossible to answer that question without answering the previous one. In the measurement of physical health, for example, an item chain would arise if a set of items were logically connected in terms of incremental ability such as ability to climb a set of stairs. If the questions were provided in numbers of steps (ie &#x02018;Can you climb 5 steps?&#x02019;; &#x02018;Can you climb 10 steps?&#x02019;) it has to be the case that someone unable to climb 5 steps will definitely be unable to climb 10 steps and so on for any number of steps.</p></sec><sec id="Sec4"><title>Item overlap</title><p>Item overlap, for example, arises in the following questions about motivation: &#x02018;I don&#x02019;t feel like getting out of bed in the morning&#x02019; and &#x02018;I just want to stay in bed in the morning&#x02019;. In this example, it is almost inconceivable that a person would answer one positively and the other negatively. In questions about preference, a series of questions related to liking for sport such as: &#x02018;Do you like playing football?; Do you like watching football on TV?; and Do you like going to football matches? would show considerable overlap. Such items are referred to in CTT as &#x02018;bloated specifics&#x02019; [<xref ref-type="bibr" rid="CR6">6</xref>] and the analogous phenomenon in regression analysis is known as multicollinearity.</p></sec></sec><sec id="Sec5"><title>Estimating LSI</title><p>Traditionally, LSI between items has been estimated using marginal frequencies in 2 x 2 contingency tables. If the items are dependent then the value of Chi-square will be large and statistically significant; if not then the value of Chi-square will be small and not statistically significant [<xref ref-type="bibr" rid="CR5">5</xref>]. This works well for dichotomous items; but for polytomous items, the degrees of freedom can be very large and some cells in the contingency table will have low values (down to and including 0 which are lower than the Yate&#x02019;s correction can account for). For parametric IRT (Rasch models) some sophisticated methods have been reported [<xref ref-type="bibr" rid="CR5">5</xref>] but for non-parametric IRT&#x02014;Mokken scaling, which is the focus of this paper&#x02014;methods remain in development [<xref ref-type="bibr" rid="CR7">7</xref>] and are not yet available.</p></sec><sec id="Sec6"><title>Mokken scaling and LSI</title><p>Mokken scaling is a non-parametric form of IRT derived from Guttman scaling [<xref ref-type="bibr" rid="CR1">1</xref>]. It is non-parametric in the sense that no assumption is made about the shape of the relationship between the score on an item and the score on the latent trait&#x02014;the item response functions (IRF)&#x02014;other than that IRFs are monotone and non-intersecting [<xref ref-type="bibr" rid="CR1">1</xref>]. Monotony refers to a property of an IRF whereby it is continually increasing over the range of the latent trait to which it relates. Non-intersection is now more commonly referred to as invariant item ordering (IIO) [<xref ref-type="bibr" rid="CR8">8</xref>] and is a property of IRFs whereby the IRFs for the total scores on a set of items are non-intersecting and the item step response functions (ISRFs) for each of the steps between response categories in polytomous items are also non-intersecting. When items are dichotomous, the IRFs and ISRFs are equivalent, and when they are non-intersecting IIO (formerly referred to as double monotony) also holds. Items which violate monotony can be identified in the diagnostics generated by Mokken scaling software and estimating IIO will be considered below.</p><p>Mokken scaling is described as a stochastic version of Guttman scaling because it envisages a stochastic&#x02014;rather than a deterministic&#x02014;relationship between the score on an item and the score on the latent trait. Nevertheless, the strength of a Mokken scale is judged by the number of Guttman errors [<xref ref-type="bibr" rid="CR9">9</xref>], whereby the relative endorsement of pairs of items is not in the expected direction. Like parametric forms of IRT, Mokken scales&#x02014;which are assumed to be unidimensional [<xref ref-type="bibr" rid="CR1">1</xref>]&#x02014;select items that form hierarchies on the basis of item difficulty. In this sense, &#x02018;difficulty&#x02019; refers to the likelihood of an item being endorsed; where endorsement of an item is indicated by a higher score on that item, then the most difficult items will have lower mean item scores [<xref ref-type="bibr" rid="CR8">8</xref>]. For example, in a scale which was designed to measure tendency to become depressed, an item about general lack of motivation would most likely be endorsed more readily than an item indicating suicidal ideation. Normally, we would expect the latter item to score lower than the former, and in a perfect Mokken scale (or a perfect Guttman scale) that would always be the case. However, in some cases items will be scored counter to expectations and these will be Guttman errors. The fewer the Guttman errors, the stronger the scale [<xref ref-type="bibr" rid="CR3">3</xref>]. The extent of Guttman errors in a Mokken scale is measured using Loevinger&#x02019;s coefficient H [<xref ref-type="bibr" rid="CR3">3</xref>], a measure of scalability, which can be reported for items (Hi), items pairs (Hij) and the overall scale (Hs). For items to be retained in a Mokken scale they must have Hi higher than some predetermined lowerbound level (c) which is normally set at 0.30. Items may also be judged by the 95% confidence intervals (CIs) around Hi and the CIs should not include the lowerbound value [<xref ref-type="bibr" rid="CR10">10</xref>]. For item pairs, the 95% CIs should not include 0 [<xref ref-type="bibr" rid="CR10">10</xref>]. For scales, the values of Hs can be considered as follows [<xref ref-type="bibr" rid="CR1">1</xref>]:<list list-type="bullet"><list-item><p>Hs&#x02009;&#x0003e;&#x02009;0.3 indicates a weak scale</p></list-item><list-item><p>Hs&#x02009;&#x0003e;&#x02009;0.4 indicates a moderate scale</p></list-item><list-item><p>Hs&#x02009;&#x0003e;&#x02009;0.5 indicates a strong scale.</p></list-item></list></p><p>It should be noted that strong scales and, especially, values of Hs greatly exceeding 0.50 are very rare in Mokken scales and very high values of H should be treated with caution and may indicate violations of LSI [<xref ref-type="bibr" rid="CR11">11</xref>]. In IRT this will arise, as introduced above, if items in a scale form a chain where responding to any question is dependent on or impossible without responding to another question in the scale. An example from soccer, used in a previous paper [<xref ref-type="bibr" rid="CR12">12</xref>], can illustrate this. If we consider that there is a latent trait &#x02018;ability at soccer&#x02019; then, over a football season, the team that wins the league&#x02014;where all other teams are played at least once&#x02014;can be considered to have achieved the highest level of difficulty. A position in a league, therefore, is analogous to a scale with LSI: the winning team, for example, does not have to win every match and it does not have to win any particular matches, only to win most and gain the highest number of points. Contrast this with a soccer cup competition where each stage is a &#x02018;knock-out&#x02019; for each team: only the winning team progresses to the next stage. Despite the fact that the latent trait of &#x02018;ability at soccer&#x02019; may well contribute to a team&#x02019;s position in the competition, a team&#x02019;s position is absolutely dependent on winning the previous stage of the competition; relative position in such a competition is not, therefore, independent of performance at another level.</p><p>Invariant item ordering is estimated using a coefficient which is analogous to Loevinger&#x02019;s coefficient H called Htrans (H^T) which is a measure of how close IRFs are [<xref ref-type="bibr" rid="CR13">13</xref>]; the closer they are then the more likely intersection is and the less likely that IIO holds; the range of values of H^T is as follows [<xref ref-type="bibr" rid="CR13">13</xref>]:<list list-type="bullet"><list-item><p>H^T&#x02009;&#x0003e;&#x02009;0.3 indicates weak IIO</p></list-item><list-item><p>H^T&#x02009;&#x0003e;&#x02009;0.4 indicates moderate IIO</p></list-item><list-item><p>H^T&#x02009;&#x0003e;&#x02009;0.5 indicates strong IIO.</p></list-item></list></p><p>Finally, the reliability of a Mokken scale can be estimated by a reliability coefficient Rho [<xref ref-type="bibr" rid="CR14">14</xref>], values of which should exceed 0.70 and the probability of obtaining a Mokken scale can be estimated using a Bonferroni method that accounts for multiple iterations in the method [<xref ref-type="bibr" rid="CR3">3</xref>]; the default setting is normally p &#x0003c;0.05.</p><sec id="Sec7"><title>Exploratory versus confirmatory Mokken scaling analysis</title><p>Mokken scaling analysis can be applied in either an exploratory or a confirmatory mode where the same criteria are used in both modes; the only difference is what is entered into the analysis. In exploratory Mokken scaling analysis a large pool of variables about which nothing is assumed or known in relation to the existence of Mokken scales is entered into the analysis. In exploratory Mokken scaling analysis, known or assumed scales are entered into the analysis and tested against the minimum criteria for Mokken scales. The two approaches are entirely complementary and flexible in the sense that in exploring the structures of established scales, new insights into existing scales can be gained and new scales developed; there is no hierarchy of methods. In the present study, as explained below, exploratory Mokken scaling analysis was considered appropriate.</p></sec></sec><sec id="Sec8"><title>The SF-36</title><p>The SF-36 is a generic instrument consisting of 36 questions to measure functional health and well-being from the patient&#x02019;s perspective. It is a practical, reliable and valid measure of physical and mental health that can be completed in five to ten minutes. The SF-36 provides scores for eight health domains (Physical Functioning, Role-Physical, Bodily Pain, General Health, Vitality, Social Functioning), and two measures of Mental Health, and provides psychometrically-based physical component summary and mental component summary scores [<xref ref-type="bibr" rid="CR15">15</xref>]. All items are rated on a three to six-point Likert scale, except for seven items in the role-physical and role emotional sub-scales, which are answered in a &#x02018;yes/no&#x02019; format. The SF-36 is designed for adults 18&#x000a0;years of age or older and can be self-administered or interview-administered. Scores are calibrated so that 50 is the average score or norm. Because the SF-36 uses norm-based scoring, comparisons can be made among other generic health surveys (SF-12 and SF-8). The SF-36 is a robust, widely used measure of quality of life, and has been translated into different languages, including Chinese Mandarin [<xref ref-type="bibr" rid="CR16">16</xref>]. The Chinese Mandarin version of the SF-36 (CM: SF-36) has been demonstrated to have good validity and reliability [<xref ref-type="bibr" rid="CR16">16</xref>,<xref ref-type="bibr" rid="CR17">17</xref>], and has been increasingly used to measure the quality of life of Chinese speaking patients, including patients with coronary heart disease [<xref ref-type="bibr" rid="CR18">18</xref>,<xref ref-type="bibr" rid="CR19">19</xref>].</p></sec><sec id="Sec9"><title>Mokken scaling of the SF-36</title><p>There have been two previous studies of the SF-36 using Mokken scaling [<xref ref-type="bibr" rid="CR20">20</xref>,<xref ref-type="bibr" rid="CR21">21</xref>]. The first study [<xref ref-type="bibr" rid="CR20">20</xref>] was of a Dutch language version of the SF-36 and Mokken scaling was used due to its less stringent nature compared with parametric forms of IRT; as such it was considered suitable for QoL measurement. This study [<xref ref-type="bibr" rid="CR20">20</xref>] was mainly concerned with the concepts of unidimensionality and reliability of the sub-scales of the SF-36 and Mokken scaling was considered a means of establishing unidimensionality while Cronbach&#x02019;s alpha&#x02014;the limitations of which are considered in the paper [<xref ref-type="bibr" rid="CR20">20</xref>]&#x02014;was used to estimate reliability. SF-36 subscales were analysed individually in a confirmatory manner; all subscales appeared to be unidimensional (Hs range 0.46 (Vitality; 4 items) &#x02013; 0.84 (Bodily Pain; 2 items)) with acceptable to high Cronbach&#x02019;s alpha (0.77 (General Health; 5 items) &#x02013; 0.93 (Physical Functioning; 10 items)). Item H values ranged from 0.40 &#x02013; 0.84, monotonicity of items was assumed and items were checked for violations of double monotonicity and it was suggested that the removal of several items with high violations of double monotonicity would lead to no violations. It should be noted that these two studies were carried out before the concept of IIO for Mokken scales had been reported and the means to calculate IIO were not available for polytomous items.</p><p>The second study [<xref ref-type="bibr" rid="CR21">21</xref>] reported the Mokken scaling of the SF-36 in older people who participated in three cohorts of a longitudinal study of ageing. Mokken scaling was applied in a confirmatory manner to each subscale of the SF-36 and to each of the cohorts separately. Scale H was taken as a measure of unidimensionality and reliability was assessed using Cronbach&#x02019;s alpha. All subscales appeared to be unidimensional (Hs range 0.42 (General Health; 5 items) &#x02013; 0.83 (Physical Functioning; 10 items)) with acceptable to high Cronbach&#x02019;s alpha (0.71 (Social Functioning; 2 items) &#x02013; 0.92 (Physical Functioning; 10 items)). Item H values ranged from 0.38 &#x02013; 0.79 but no further checks of Mokken scaling parameters were done.</p><p>The above studies have several common features. Both used MSA in a confirmatory as opposed to an exploratory manner [<xref ref-type="bibr" rid="CR22">22</xref>], thereby assuming that the underlying dimensions of the SF-36 were robust. In fact, it may have been more appropriate in this instance to explore the structure of the SF-36 to establish how many scales were present and to investigate whether there was some other structure of subscales according to the criteria of MSA. It would also have been appropriate to consider if all of the items were suitable or present in sufficient numbers for MSA. Neither study inspected the mean item values, nor any additional MSA diagnostics, to establish if there was a sensible hierarchy of items in the subscales and neither study considered the possibility, despite the nature of some of the items in the SF-36 (to be considered below), of their being violations of LSI. Both studies used Cronbach&#x02019;s alpha, as opposed to the unbiased estimator of reliability Rho [<xref ref-type="bibr" rid="CR14">14</xref>] available with Mokken scaling packages. It is well known that, in addition to other limitations [<xref ref-type="bibr" rid="CR23">23</xref>], Cronbach&#x02019;s alpha is sensitive to the number of items in a scale [<xref ref-type="bibr" rid="CR24">24</xref>]; specifically, alpha is inflated as the number of items increases [<xref ref-type="bibr" rid="CR25">25</xref>] as demonstrated through Monte Carlo simulation [<xref ref-type="bibr" rid="CR24">24</xref>], and this phenomenon is apparent in the reliability data from both studies. Furthermore, on inspection of the items of the SF-36, we consider that it is not appropriate to subject them all to MSA. For example, the response formats of the first two general health questions are not congruent and clearly overlapping. Questions 4 and 5 relating to physical and mental health have two questions in common and a third very similar question. The questions on pain (7 and 8) are not suitable for MSA as there are only two of them; there is only one question related to social health (10) and the three statements under question 11 on general health are clearly overlapping. Only the items in questions 3 and 9 measuring Physical Functioning and Mental Health, respectively, provide a sufficient number of questions for Mokken scaling. We consider some of these aspects of the above studies to be problematic and that the application of MSA to the SF-36 without fully considering the nature of Mokken scaling and the possibility that some aspects may have been violated, to be an incomplete application of the method.</p><sec id="Sec10"><title>Likelihood of violations of LSI in the SF-36</title><p>Inspecting the items of the SF-36, especially those in the measuring the physical dimensions of quality of life, it is likely that violations of LSI will take place. This is especially the case for those aspects that ask about cumulative walking distances; it is logical that ability to walk any particular distance will be predicated on walking shorter distances but may not imply ability to walk a longer distance. It is likely that violations of LSI did not take place among the items related to Mental Health.</p></sec></sec><sec id="Sec11"><title>The present study</title><p>We suspect that the extraordinarily high values of Hs obtained in both the above studies of the SF-36 indicate that the apparent scalability is artificially high. One explanation for this is that there are violations of LSI in the subscales of the SF-36 which is exaggerating the scalability.</p><p>The present study uses exploratory MSA to analyse two dimensions of items from the Chinese Mandarin form of the SF-36 (CM: SF-36) together to determine if there are underlying dimensions according to the criteria of Mokken scaling and, subsequently, to study the nature of any scales obtained. Therefore, the research question guiding the study was: &#x02018;Do violations of local stochastic independence exaggerate scalability in Mokken scaling analysis in the subscales of the CH: SF-36?&#x02019;.</p></sec></sec><sec id="Sec12" sec-type="materials|methods"><title>Methods</title><sec id="Sec13"><title>Participants</title><p>This is a secondary analysis of data from a cross-sectional study conducted at two university teaching hospitals in the People&#x02019;s Republic of China. A convenience sample consisted of patients who had a clinical diagnosis of coronary heart disease, were older than 18&#x000a0;years of age, were able to comprehend Chinese and did not have a known history of psychiatric disorders or a severe co-morbidity.</p><p>Ethical approval was obtained from both teaching hospitals of Xi&#x02019;an Jiaotong University. A research assistant administered the CM: SF-36 and collected demographic data from patients who agreed to participate in the study. Of 248 patients invited to participate, 202 agreed and completed the questionnaire. The mean age of these participants was 62.8 (SD =11.6) years and two-thirds of the participants were male.</p></sec><sec id="Sec14"><title>Analysis</title><p>As discussed above, inspection of the items of the SF-36 suggest that some are not suitable for MSA but also that the existence of any scales within the SF-36 had not been investigated fully using MSA. Therefore, in the present analysis we chose to analyze only the 19 items in questions 3 and 9 related, respectively, to the Physical Functioning and Mental Health aspects of the SF-36. The strategy was to check the number of putative Mokken scales present in the data and, if scales were identified, to analyze their Mokken scaling properties separately. Using the software <italic>MSP5 for Windows</italic> [<xref ref-type="bibr" rid="CR26">26</xref>] and the method of Hempker <italic>et al.</italic> [<xref ref-type="bibr" rid="CR27">27</xref>] and Meijer and Baneke [<xref ref-type="bibr" rid="CR28">28</xref>], as applied by Nader <italic>et al.</italic> [<xref ref-type="bibr" rid="CR2">2</xref>] and Shenkin et al. [<xref ref-type="bibr" rid="CR29">29</xref>] the data were explored for multiple dimensions. Using incremental values of c, starting at a lowerbound c =0.05, these are increased in 0.05 increments. This is continued until an appropriate balance is found between the number of scales which are reliable (rho &#x0003e;0.7) and an absence of trivial scales with fewer than three items. <italic>MSP5 for Windows</italic> was used only to study the effect on increasing the lowerbound threshold as this software is very convenient for this analysis. The remaining analysis was carried out, as described below, using the public domain software R as this uniquely, permits the calculation of H^T and also the plotting of IRF pairs.</p><p>CM: SF-36 data were analysed using package &#x02018;mokken&#x02019; (<ext-link ext-link-type="uri" xlink:href="http://cran.r-project.org/web/packages/mokken/mokken.pdf">http://cran.r-project.org/web/packages/mokken/mokken.pdf</ext-link>) in <italic>R</italic> [<xref ref-type="bibr" rid="CR30">30</xref>] (public domain software available at <ext-link ext-link-type="uri" xlink:href="http://www.r-project.org/">http://www.r-project.org/</ext-link>) and package &#x02018;foreign&#x02019; (<ext-link ext-link-type="uri" xlink:href="http://cran.r-project.org/web/packages/foreign/foreign.pdf">http://cran.r-project.org/web/packages/foreign/foreign.pdf</ext-link>) was used to convert SPSS&#x000a9; data into <italic>R</italic> data. Mokken scaling analysis (MSA) was run using the automated item selection procedure (aisp) in R. The aisp selects items and allocates them to scales in an hierarchical and iterative manner starting with those item pairs which scale best (ie with the highest values of Hij) and then building scales until no further items&#x02014;ie those with Hi below the selected lowerbound threshold, for example, 0.30&#x02014;can be incorporated into the scales. Items with Hi below the lowerbound value are excluded from the Mokken scale. The subsequent scales are then checked for items violating montonicity (check.monotonicity in R), reliability (check.reliability in R) and IIO (check.iio in R) and confidence intervals of Hi and Hij were calculated by hand. Item pairs were plotted (using plot(check.iio(FileR))) and inspected visually for overlap and also for &#x02018;extreme&#x02019; items: those lying far from the remaining clusters of items which could also be, artificially, exaggerating IIO.</p></sec></sec><sec id="Sec15" sec-type="results"><title>Results</title><p>The procedure of using incremental lowerbound values of c supported this two scale structure. Between c =0.05 to c =0.30 only one reliable scale was apparent and at c =0.35 and c =0.40, two scales were apparent both with Hs &#x0003e;0.50. At c &#x0003e;0.40, a three scale structure was apparent but the third scale was trivial. The two scales at c =0.40 perfectly partitioned the items into those related to Physical Functioning and Mental Health. After separate analysis of these two sets of items the two Mokken scales formed from the data are shown in Tables&#x000a0;<xref rid="Tab1" ref-type="table">1</xref> and <xref rid="Tab2" ref-type="table">2</xref>. The scale in Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref> contains items exclusively from the Physical Functioning dimension of the CM: SF-36 and the scale in Table&#x000a0;<xref rid="Tab2" ref-type="table">2</xref> contains items exclusively from the Mental Health dimension of the CM: SF-36. None of the items in either scale violated monotonicity and the 95% CIs of all item pairs were acceptable. With one exception, the 95% Cis of the items were acceptable, therefore, no further items were excluded.<table-wrap id="Tab1"><label>Table 1</label><caption><p> <bold>Mokken scale &#x02018;Physical Functioning&#x02019; from the SF-36 (n&#x02009;=&#x02009;202)</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th> <bold>Item</bold> </th><th> <bold>Label</bold> </th><th> <bold>Mean</bold> </th><th> <bold>Hi (SE)</bold> </th></tr></thead><tbody><tr valign="top"><td>9</td><td>Walking 150 meters</td><td>2.75</td><td>0.78 (0.040)</td></tr><tr valign="top"><td>10</td><td>Bathing or dressing yourself</td><td>2.72</td><td>0.67 (0.056)</td></tr><tr valign="top"><td>5</td><td>Climbing one flight of stairs</td><td>2.66</td><td>0.78 (0.039)</td></tr><tr valign="top"><td>3</td><td>Lifting or carrying groceries</td><td>2.57</td><td>0.74 (0.037)</td></tr><tr valign="top"><td>8</td><td>Walking 800 meters</td><td>2.56</td><td>0.76 (0.037)</td></tr><tr valign="top"><td>6</td><td>Bending, kneeling, or stooping</td><td>2.56</td><td>0.70 (0.044)</td></tr><tr valign="top"><td>2</td><td>Moderate activities, such as moving a table, cleaning the floor</td><td>2.39</td><td>0.75 (0.036)</td></tr><tr valign="top"><td>7</td><td>Walking 1,600 meters</td><td>2.28</td><td>0.74 (0.034)</td></tr><tr valign="top"><td>4</td><td>Climbing several flights of stairs</td><td>2.20</td><td>0.71 (0.042)</td></tr><tr valign="top"><td>1</td><td>Vigorous activities, such as running, lifting heavy objects, participating in strenuous sports</td><td>1.54</td><td>0.59 (0.065)</td></tr></tbody></table><table-wrap-foot><p> <italic>Hi</italic>&#x02009;=&#x02009;item H; <italic>Hs</italic>&#x02009;=&#x02009;scale H =0.73(SE 0.031); Rho&#x02009;=&#x02009;0.93; H^T&#x02009;=&#x02009;0.70.</p></table-wrap-foot></table-wrap><table-wrap id="Tab2"><label>Table 2</label><caption><p> <bold>Mokken scale &#x02018;Role-Emotional&#x02019; from the SF-36 (n&#x02009;=&#x02009;202)</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th> <bold>Item</bold> </th><th> <bold>Label</bold> </th><th> <bold>Mean</bold> </th><th> <bold>Hi (SE)</bold> </th></tr></thead><tbody><tr valign="top"><td>3</td><td>Have you felt so down in the dumps that nothing could cheer you up?</td><td>4.79</td><td>0.44 (0.060)</td></tr><tr valign="top"><td>2</td><td>Have you been a very nervous person?</td><td>4.56</td><td>0.38 (0.054)*</td></tr><tr valign="top"><td>6</td><td>Have you felt downhearted and blue?</td><td>3.52</td><td>0.53 (0.045)</td></tr><tr valign="top"><td>4</td><td>Have you felt calm and peaceful?^&#x02020; </td><td>3.97</td><td>0.41 (0.059)</td></tr><tr valign="top"><td>8</td><td>Have you been a happy person?</td><td>3.66</td><td>0.44 (0.050)</td></tr></tbody></table><table-wrap-foot><p> <italic>Hi</italic>&#x02009;=&#x02009;item H; <italic>Hs</italic>&#x02009;=&#x02009;scale H =0.44 (SE 0.045); Rho&#x02009;=&#x02009;0.77; H^T&#x02009;=&#x02009;0.35; ^{<bold>&#x02020;</bold>}-reverse scored items; items with lowerbound 95% confidence interval &#x0003c;0.30; * 95% confidence interval included 0.30.</p></table-wrap-foot></table-wrap></p><sec id="Sec16"><title>Physical functioning</title><p>The Physical Functioning Mokken scale retained all 10 of the items related to that dimension in the CM: SF-36 and was a strong scale (Hs =0.73) with strong IIO (H^T&#x02009;=&#x02009;0.70). Inspection of item pair plots (Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>) showed that the IRF for item 1 in this scale was positioned far from the remaining items and could be contributing to the high IIO; removing item 1 and re-analysis of the scale properties reduced H^T to 0.42 but the scale remained strong at Hs &#x0003e;0.70. The hierarchy of items in the Physical Functioning scale runs from walking a moderate distance through longer distances and a range of activities of daily living and instrumental activities of daily living to vigorous activity. Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref> shows the item pair plots for the items in the Physical Functioning scale referring to walking 150 metres, 800 metres and 1,600 metres. Collectively, the three item pair plots show increasing difficulty with increasing distance; the IRF for 800 metres lies between those for 150 metres and 1,600 metres. Items excluded from the scale were those that did not meet the minimal criteria for Mokken scaling analysis outlined in the <xref rid="Sec1" ref-type="sec">Background</xref> section.<fig id="Fig1"><label>Figure 1</label><caption><p> <bold>Example of an CM: SF-36 item pair plot (for &#x02018;Vigorous activities&#x02019;) lying far from a selection of remaining item pair plots. a</bold> Item pair plots for &#x02018;Vigorous activities&#x02019; and &#x02018;Walking 150 metres&#x02019;. <bold>b</bold> Item pair plots for &#x02018;Vigorous activities&#x02019; and &#x02018;Bathing or dressing yourself&#x02019;. <bold>c</bold> Item pair plots for &#x02018;Vigorous activities&#x02019; and &#x02018;Climbing one flight of stairs&#x02019;. <bold>d</bold> Item pair plots for &#x02018;Vigorous activities&#x02019; and &#x02018;Lifting or carrying groceries&#x02019;.</p></caption><graphic xlink:href="12955_2014_149_Fig1_HTML" id="MO1"/></fig><fig id="Fig2"><label>Figure 2</label><caption><p> <bold>Example of a set of items that violate local stochastic independence. a</bold> Item pair plots for &#x02018;Walking 150 metres&#x02019; and &#x02018;Walking 800 metres&#x02019;. <bold>b</bold> Item pair plots for &#x02018;Walking 800 metres&#x02019; and &#x02018;Walking 1,600 metres&#x02019;. <bold>c</bold> Item pair plots for &#x02018;Walking 150 metres&#x02019; and &#x02018;Walking 1,600 metres&#x02019;.</p></caption><graphic xlink:href="12955_2014_149_Fig2_HTML" id="MO2"/></fig></p><sec id="Sec17"><title>Mental health</title><p>The Mental Health Mokken scale retained all five of the items related to that dimension in the SF-36 and was moderately strong (Hs =0.44) with weak IIO (H^T&#x02009;=&#x02009;0.35). However, it should be noted that, for one item the 95% confidence interval around Hi included the lowerbound 0.30 suggesting that this item could be removed from the scale. In the present study it was not removed as the confidence intervals are related to sample size and in a larger sample this item may well have 95% confidence intervals around Hi that do not include the lowerbound 0.30. It should be noted in the Mental Health scale that a low mean item score indicates endorsement and that item 4 (&#x02018;Have you felt calm and peaceful?&#x02019;) is reverse scored. Therefore, from the IRT perspective, the &#x02018;easiest&#x02019; item is &#x02018;Have you been a happy person?&#x02019;&#x02014;which gained the highest endorsement&#x02014;and the most &#x02018;difficult&#x02019; item is &#x02018;Have you felt so down in the dumps that nothing could cheer you up?&#x02019;.</p></sec></sec></sec><sec id="Sec18" sec-type="discussion"><title>Discussion</title><p>Recent work [<xref ref-type="bibr" rid="CR11">11</xref>] has shown that high scalability in Mokken scales is worth investigating. Therefore, we set out to study if violations of local stochastic independence exaggerate scalability in Mokken scaling analysis of the subscales of the CM: SF-36 using exploratory MSA. Our view is that previous work on the SF-36 using MSA was limited. Specifically, we chose to study one subscale of the CH:SF-36 where the items were likely to violate LSI (Physical Functioning) and one where this was less likely (Mental Health). As a preliminary step, we explored the dimensionality of the Physical Functioning and Mental Health items of the CM: SF-36 using MSA to see if the underlying structure of these two subscales was supported. We consider that an exploratory approach&#x02014;to see if there were subscales according to Mokken scaling criteria&#x02014;was advantageous in this instance. We reiterate that some of the methods applied in this study&#x02014;analysis of IIO, calculation of CIs and plotting of item pairs&#x02014;were unavailable to previous analysts [<xref ref-type="bibr" rid="CR20">20</xref>,<xref ref-type="bibr" rid="CR21">21</xref>]. In addition, both of the previous studies of the SF-36 using MSA [<xref ref-type="bibr" rid="CR20">20</xref>,<xref ref-type="bibr" rid="CR21">21</xref>] used the reliability coefficient Cronbach&#x02019;s alpha and not the reliability coefficient Rho which is available in Mokken scaling packages. There are well known limitations to Cronbach&#x02019;s alpha [<xref ref-type="bibr" rid="CR23">23</xref>], and Rho&#x02014;described as an unbiased estimator of reliability [<xref ref-type="bibr" rid="CR14">14</xref>]&#x02014;is considered an improvement on Cronbach&#x02019;s alpha and was used in this study.</p><p>The sample size in the present study was low for Mokken scaling and, until recently, little work had been done on sample size requirements. However, in simulated studies of sample size, Straat [<xref ref-type="bibr" rid="CR7">7</xref>] has shown that one of the parameters to which sample size is related&#x02014;and inversely proportional&#x02014;is Hi and in this study the values if Hi, especially for the items in the Physical Functioning subscale, were high and according to Straat&#x02019;s work, the sample size was probably adequate. As mentioned above, a larger sample size may have resulted in the inclusion of all the Mental Health items in the relevant Mokken scale.</p><p>Our study supports the underlying structure of the CM: SF-36 inasmuch as, according to MSA, two Mokken scales were derived in the present analysis of the 19 items related to Physical Functioning and Mental Health and each scale was derived exclusively of items related to their respective dimensions. The scale formed from the Physical Functioning subscale included ten items and had a very high Loevinger&#x02019;s coefficient H. The strength of the scale is unusually high and it is apparent, on inspecting the items in the Physical Functioning subscale, there is a strong possibility they are not stochastically independent. Specifically, items 7, 8 and 9 refer to walking for increasing distances and, likewise, items 4 and 5 refer to climbing increasing numbers of flights of stairs and items 1 and 2 refer to increasing extents of exercise. These are likely to violate LSI because it is logical that achievement at any level in these incremental measures of physical activity is predicated on achievement at the lower level and that achievement above the highest level is impossible. Therefore, it is highly likely that the phenomena of an item chain is present, which is a potential sources of violations of LSI [<xref ref-type="bibr" rid="CR5">5</xref>].</p><p>Only five items from the Mental Health subscale were retained in a Mokken scale which was moderately strong (taking the present sample size into account). It is likely that this is a true Mokken scale showing an item hierarchy that is determined by the latent trait and not due to an item chain or item overlap. Taking the fact that low scores on this subscale of the CM: SF-36 mean high endorsement, the items form a sensible hierarchy from being happy through to being impossible to cheer up and the wording of the items suggest LSI. These items are unlikely to have violated LSI because the responses are not dependent on each other and they probe different aspects of mental health. Missing items are simply a result of items failing to meet the criteria for a Mokken scale as outlined in the <xref rid="Sec12" ref-type="sec">Methods</xref> section. Clearly this leads to some construct underrepresentation; however, in the present study this may be the result of the low sample size.</p><p>Both scales were reliable as indicated by values of Rho exceeding 0.70. For the Physical Functioning scale, Rho was very high (0.93) and this was also observed for Cronbach&#x02019;s alpha in previous studies [<xref ref-type="bibr" rid="CR20">20</xref>,<xref ref-type="bibr" rid="CR21">21</xref>]. For the Mental Health scale, Rho was 0.77 indicating acceptable reliability. The very high levels of Rho (and Cronbach&#x02019;s alpha) for the Physical Functioning scale could indicate item redundancy [<xref ref-type="bibr" rid="CR31">31</xref>], from the scaling perspective, and this is very likely given the similar wording and overlapping concepts in the items. It should be noted that factor analysis of these items is likely to suffer from the same phenomenon, leading to an artificially highly loaded set of items on a putative factor and these items would be &#x02018;bloated specifics&#x02019; [<xref ref-type="bibr" rid="CR6">6</xref>]. In regression analysis these items would most likely demonstrate collinearity.</p></sec><sec id="Sec19" sec-type="conclusion"><title>Conclusion</title><p>Our study partly supports previous work on the SF-36 using MSA. However, we conclude that previous MSA of the SF-36 may have concluded wrongly that all the subscales were unidimensional, at least by the criteria for Mokken scales. In any case, since undimensionality in Mokken scales is related not only to simple covariance, but also to an hierarchical ordering of items in scales, MSA may not have been the most appropriate analytical procedure. This is especially the case for the items of Physical Functioning aspect of the SF-36, some of which are likely to violate LSI. We recommend in future applications of Mokken scaling, that the possibility of violations of LSI be considered either prior to the analysis and always where very high values of scalability are obtained. Nevertheless, we are aware that the properties of the CM: SF-36 may be unique and that the sample size in the present study was small.</p><p>The consequences of this study do not undermine the use of the Short Form health survey in any of its forms or translations. Indeed, further study of the English version of the SF-36&#x02014;and other language translations&#x02014;is warranted, especially where adequate sample sizes can be obtained. The outcome of the study does lead us to urge caution in the interpretation of putative scale properties in general&#x02014;not only the SF-36&#x02014;where fundamental assumptions that are crucial to the application of any psychometric method are likely to have been violated.</p></sec>

Lactose-free milk prolonged endurance capacity in lactose intolerant Asian males

Could not extract abstract

<contrib contrib-type="author"><name><surname>Sudsa-ard</surname><given-names>Kriyot</given-names></name><address><email>neo-one13@hotmail.com</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author" corresp="yes"><name><surname>Kijboonchoo</surname><given-names>Kallaya</given-names></name><address><email>kallaya.kij@mahidol.ac.th</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author"><name><surname>Chavasit</surname><given-names>Visith</given-names></name><address><email>visith.cha@mahidol.ac.th</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author"><name><surname>Chaunchaiyakul</surname><given-names>Rungchai</given-names></name><address><email>gmrungchai@gmail.com</email></address><xref ref-type="aff" rid="Aff2"/></contrib><contrib contrib-type="author"><name><surname>Nio</surname><given-names>Amanda Qing Xia</given-names></name><address><email>anio@cardiffmet.ac.uk</email></address><xref ref-type="aff" rid="Aff3"/></contrib><contrib contrib-type="author"><name><surname>Lee</surname><given-names>Jason Kai Wei</given-names></name><address><email>lkaiwei@dso.org.sg</email></address><xref ref-type="aff" rid="Aff4"/><xref ref-type="aff" rid="Aff5"/><xref ref-type="aff" rid="Aff6"/></contrib><aff id="Aff1"><label/>Institute of Nutrition, Mahidol University, Salaya, Phutthamonthon, Nakhon Pathom 73170 Thailand </aff><aff id="Aff2"><label/>College of Sports Science and Technology, Mahidol University, Salaya, Phutthamonthon, Nakhon Pathom 73170 Thailand </aff><aff id="Aff3"><label/>School of Sport, Cardiff Metropolitan University, Cardiff, UK </aff><aff id="Aff4"><label/>Defence Medical and Environmental Research Institute, DSO National Laboratories, Singapore, Singapore </aff><aff id="Aff5"><label/>Yong Loo Lin School of Medicine, National University of Singapore, Singapore, Singapore </aff><aff id="Aff6"><label/>Lee Kong Chian School of Medicine, Nanyang Technological University, Singapore, Singapore </aff>

Journal of the International Society of Sports Nutrition

<sec id="Sec1" sec-type="introduction"><title>Introduction</title><p>Nutrition is one aspect of an athlete&#x02019;s lifestyle that can be modified to enhance sporting performance. This encompasses individualized interventions adopted before, during, and after exercise. Appropriate nutrition after exercise can enhance recovery and augment performance in the subsequent exercise bout, which may in turn, encourage greater physiological adaptation to exercise training and result in improved performance during competition.</p><p>Milk has been proven to be an effective post-exercise drink for endurance activities [<xref ref-type="bibr" rid="CR1">1</xref>-<xref ref-type="bibr" rid="CR7">7</xref>]. Milk contains several ergogenic nutrients including carbohydrate, protein, fat, vitamins, and minerals. Following an exhaustive bout of exercise, athletes who drank milk could recover faster and exhibit better exercise performance compared with those who had commercial sports drinks or carbohydrate replacement drink [<xref ref-type="bibr" rid="CR5">5</xref>,<xref ref-type="bibr" rid="CR6">6</xref>]. Moreover, an electrolyte drink that was fortified with carbohydrate and protein could increase muscle glycogen by 128% more than a 6% carbohydrate drink [<xref ref-type="bibr" rid="CR8">8</xref>]. Beradi et al. [<xref ref-type="bibr" rid="CR9">9</xref>] explained that muscle glycogen resynthesis was greater following 6&#x000a0;h of recovery due to the addition of protein to the recovery drink.</p><p>Unfortunately, the benefits of protein and carbohydrate, especially from milk, in enhancing the recovery period of athletes are not applicable to lactose intolerance individuals, such as Asians. Symptoms of lactose intolerance can include nausea, vomiting, and diarrhea, with the severity of symptoms dependent on the level of lactose intolerance. Such an effect will likely impair exercise performance. A study performed in indigenous Asians in Singapore showed that all of the sampled 22 subjects aged 15&#x02013;42 y were lactose-intolerant [<xref ref-type="bibr" rid="CR10">10</xref>]. Asmawi et al. [<xref ref-type="bibr" rid="CR11">11</xref>] reported hypolactasia in 88% of Malaysian Malays, 91% of Malaysian Chinese, and 83% of Malaysian Indians. A study using the breath-hydrogen test after oral intake of 25&#x000a0;g lactose in Thai adults demonstrated that almost half of the cohort was lactose intolerant [<xref ref-type="bibr" rid="CR12">12</xref>].</p><p>Lactose-free milk (LFM) may be potentially ergogenic as a recovery beverage for lactose intolerant individuals. We hypothesized that a LFM drink could extend cycling time to exhaustion. To our knowledge, there is no study of LFM on prolonged exercise. We investigated the effects of LFM on endurance cycling capacity in healthy Thai males.</p></sec><sec id="Sec2" sec-type="materials|methods"><title>Methods</title><sec id="Sec3"><title>Study participants</title><p>Ten healthy males volunteered for this study (Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>). Experimental procedures were approved by Mahidol University Institutional Review Board, Thailand. Participants received a verbal explanation about the study before providing written informed consent.<table-wrap id="Tab1"><label>Table 1</label><caption><p> <bold>Physical and physiological characteristics of the participants (n&#x02009;=&#x02009;10)</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th/><th> <bold>Mean&#x02009;&#x000b1;&#x02009;SD</bold> </th><th> <bold>Range</bold> </th></tr></thead><tbody><tr valign="top"><td>Age (yr)</td><td>21.2&#x02009;&#x000b1;&#x02009;0.8</td><td>20&#x02013;22</td></tr><tr valign="top"><td>Height (cm)</td><td>174&#x02009;&#x000b1;&#x02009;3</td><td>168&#x02013;178</td></tr><tr valign="top"><td>Body weight (kg)</td><td>66.8&#x02009;&#x000b1;&#x02009;4.6</td><td>60.8&#x02013;73.8</td></tr><tr valign="top"><td>BMI (kg/m²)</td><td>22.1&#x02009;&#x000b1;&#x02009;1.5</td><td>20.3&#x02013;24.8</td></tr><tr valign="top"><td>Body fat (%)</td><td>12.6&#x02009;&#x000b1;&#x02009;4.3</td><td>9.7&#x02013;19.7</td></tr><tr valign="top"><td>VO₂max (ml/kg/min)</td><td>44&#x02009;&#x000b1;&#x02009;2</td><td>40&#x02013;47</td></tr><tr valign="top"><td>Peak power output (W)</td><td>288&#x02009;&#x000b1;&#x02009;41</td><td>280&#x02013;320</td></tr></tbody></table></table-wrap></p></sec><sec id="Sec4"><title>Laboratory protocol</title><p>Participants arrived at the laboratory in the morning following an overnight fast. Each participant completed four visits to the laboratory (ambient temperature: 25&#x02009;&#x000b1;&#x02009;1&#x000b0;C; relative humidity: 56%&#x02009;&#x000b1;&#x02009;2): a maximal incremental exercise test and three randomized experimental trials ingesting water (WT), a commercial sports drink (SPD), or lactose-free milk (LFM; manufactured by the Institute of Nutrition, Mahidol University, Thailand; Table&#x000a0;<xref rid="Tab2" ref-type="table">2</xref>).<table-wrap id="Tab2"><label>Table 2</label><caption><p> <bold>Nutritional content in 250&#x000a0;mL of each test drink</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th> <bold>Contents</bold> </th><th> <bold>Water</bold> </th><th> <bold>Sports drink</bold> </th><th> <bold>Lactose-free milk</bold> </th></tr></thead><tbody><tr valign="top"><td>Energy (kcal)</td><td>-</td><td>100</td><td>100</td></tr><tr valign="top"><td>Carbohydrate (g)</td><td>-</td><td>25.0</td><td>12.5</td></tr><tr valign="top"><td>Protein (g)</td><td>-</td><td>-</td><td>8</td></tr><tr valign="top"><td>Fat (g)</td><td>-</td><td>-</td><td>2</td></tr><tr valign="top"><td>Sodium (g)</td><td>-</td><td>0.092</td><td>0.439</td></tr><tr valign="top"><td>Potassium (g)</td><td>-</td><td>0.015</td><td>0.518</td></tr><tr valign="top"><td>Osmolality (mosmol/kg)</td><td>-</td><td>415</td><td>352</td></tr></tbody></table></table-wrap></p><p>Lactose-free milk was produced by adding 500&#x000a0;ppm &#x003b2;-galactosidase (lactase) enzyme (Ha-Lactase 5200 NLU/g, Chr-Hansen, Horsholm, Denmark) to an in-house low-fat pasteurized milk. The inoculated milk was incubated at 8&#x000b0;C for 24&#x000a0;h before pasteurization at 72&#x000b0;C for 15&#x000a0;s. The pasteurized milk was cooled to room temperature and fortified with sodium and potassium to increase its osmolality which was initially much lower than the sports drink.</p><p>The experimental trial consisted of an exercise-induced glycogen depletion session, a 2-h recovery period during which the test drink was ingested, and an endurance capacity test. Each trial was separated by 1&#x02013;2 weeks. Dietary intake in the three days prior to the first experimental trial was recorded and repeated for subsequent trials. The macronutrients ingested were calculated using an in-house dietary software programme (INMUCAL, Institute of Nutrition, Mahidol University, Thailand). Urine specific gravity (Refractometer model 300CL, Atago Inc, Japan) was recorded as a measure of hydration status before each experimental trial.</p></sec><sec id="Sec5"><title>Maximal incremental exercise test</title><p>Participants completed a standardized 2-min warm-up (cadence 60 revolutions [rev]/min, workload 0.5kp) on a cycle ergometer (Ergomedic 828 E, Sweden) followed by an incremental cycling test to volitional exhaustion at a cadence of 80 rev/min. Workload was increased by 1 kp every 2&#x000a0;min. Oxygen uptake was determined using a metabolic cart (Vmax Sensor Medics Metabolic, SensorMedics&#x000ae; Corporation, USA). Heart rate (Polar RS800CX POLAR&#x000ae;, Finland) was measured. The test was accepted if at least two of following criteria were met: i) respiratory exchange ratio (RER) greater than 1.1; ii) heart rate above 90% of age-predicted maximum heart rate; and iii) a VO₂ increase of less than 0.15&#x000a0;l/min from the previous workload [<xref ref-type="bibr" rid="CR13">13</xref>]. Maximum power output (Pmax) was defined as the power output attained during the final completed stage.</p></sec><sec id="Sec6"><title>Glycogen depletion session</title><p>The glycogen depletion exercise consisted of 2-min intervals at 60%&#x02013;90%Pmax interspersed with 50%Pmax recovery at 80 rev/min [<xref ref-type="bibr" rid="CR5">5</xref>]. Participants commenced the exercise with 2-min intervals at 90%Pmax and the workload was subsequently reduced to 80%Pmax when their cadence fell below 70 rev/min for more than 30&#x000a0;s. This criterion was repeated for further reductions to 70%Pmax and 60%Pmax, respectively. Participants continued cycling with the alternating 2&#x000a0;min periods at each work interval at 50%Pmax with 80 rev/min until they could no longer maintain the required cadence over a minute. Heart rate was monitored every minute. Height, weight, body fat, and body mass index (BMI) were measured using a body composition analyzer (Inbody 720, Biospace, Korea) before and immediately after the glycogen depletion session. Blood lactate was measured before and after the endurance capacity test using a lactate scout analyzer (SensLab, LSSY-170407-E, Germany). VO₂, VCO₂ and RER (Vmax Sensor Medics Metabolic, SensorMedics&#x000ae;, USA) were measured continuously during the endurance capacity test.</p></sec><sec id="Sec7"><title>Recovery period</title><p>After completing the glycogen depletion session, participants rested for 2&#x000a0;h in the laboratory. The volume of LFM provided was calculated such that participants received 1&#x000a0;g of CHO/kg BM [<xref ref-type="bibr" rid="CR5">5</xref>,<xref ref-type="bibr" rid="CR6">6</xref>] during the recovery period. For example, if the participant had a body mass of 66&#x000a0;kg, he would receive 1320&#x000a0;ml of LFM (12.5&#x000a0;g CHO/250&#x000a0;ml LFM). By selecting a commercially available sports drink with a caloric content similar to LFM, both volume and caloric content of test drinks ingested during the SPD and LFM trials were matched. An equal volume of water was provided during the WT trial. Each test drink was administered in three aliquots during the recovery period: 50% at 0&#x000a0;min, 25% at 30&#x000a0;min and the remaining 25% at 60&#x000a0;min. [<xref ref-type="bibr" rid="CR14">14</xref>,<xref ref-type="bibr" rid="CR15">15</xref>].</p></sec><sec id="Sec8"><title>Endurance capacity test</title><p>Following the recovery period, participants completed a standardized warm-up (2&#x000a0;min at 60 rev/min, workload of 0.5 kp) before embarking on the endurance capacity test at 80 rev/min at 70%VO₂max. The test was terminated when the participant&#x02019;s cadence fell below 70 rev/min for more than 30&#x000a0;s twice [<xref ref-type="bibr" rid="CR5">5</xref>]. VO₂, VCO₂, and RER were measured during the endurance capacity test. Heart rate was recorded continuously. Time to exhaustion was measured. Rating of perceived exertion (RPE) were measured every 5&#x000a0;min [<xref ref-type="bibr" rid="CR16">16</xref>]. Body weight and blood lactate were measured at the start and immediately at the end of the trial.</p></sec><sec id="Sec9"><title>Data analysis</title><p>Statistical analyses were performed using SPSS (IBM&#x000ae; Corp, SPSS&#x000ae;Statistics Version 21, USA). One-way repeated measures analysis of variance (ANOVA) was used to compare all variables (e.g. heart rate, oxygen uptake and cycle time to exhaustion) between trials (i.e. WT, SPD and LFM). Data were presented as mean&#x02009;&#x000b1;&#x02009;standard deviation (SD). Alpha was set <italic>a priori</italic> at 0.05.</p></sec></sec><sec id="Sec10" sec-type="results"><title>Results</title><p>Physical and physiological characteristics of participants are shown in Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>. Participants were similarly euhydrated (urine specific gravity; WT: 1.014&#x02009;&#x000b1;&#x02009;0.004, SPD: 1.014&#x02009;&#x000b1;&#x02009;0.004, LFM: 1.020&#x02009;&#x000b1;&#x02009;0.003; p&#x02009;=&#x02009;0.57) and had similar absolute (WT: 2019&#x02009;&#x000b1;&#x02009;573, SPD: 1722&#x02009;&#x000b1;&#x02009;487, LFM: 2113&#x02009;&#x000b1;&#x02009;389&#x000a0;kcal/day; p&#x02009;=&#x02009;0.95) and relative caloric intakes (WT: 30&#x02009;&#x000b1;&#x02009;3, SPD: 26&#x02009;&#x000b1;&#x02009;8, LFM: 31&#x02009;&#x000b1;&#x02009;6&#x000a0;kcal/kg/day; p&#x02009;=&#x02009;0.39) prior to undertaking the experimental trials.</p><p>Heart rate was similar across all three experimental trials (Table&#x000a0;<xref rid="Tab3" ref-type="table">3</xref>). Whilst oxygen consumption and RER were similar during all three glycogen depletion sessions, oxygen uptake was lower in LFM (p&#x02009;&#x0003c;&#x02009;0.05) during the endurance capacity test than in SPD and WT. In addition, RER was lower in WT than SPD during the endurance capacity test (p&#x02009;&#x0003c;&#x02009;0.05).<table-wrap id="Tab3"><label>Table 3</label><caption><p> <bold>Physiological responses during each trial</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th/><th> <bold>Water</bold> </th><th> <bold>Sports drink</bold> </th><th> <bold>Lactose-free milk</bold> </th></tr></thead><tbody><tr valign="top"><td> <bold>Glycogen depletion session</bold> </td><td/><td/><td/></tr><tr valign="top"><td>Heart rate (beats/min)</td><td>136&#x02009;&#x000b1;&#x02009;7</td><td>139&#x02009;&#x000b1;&#x02009;6</td><td>135&#x02009;&#x000b1;&#x02009;7</td></tr><tr valign="top"><td>VO₂ (ml/min/kg)</td><td>38&#x02009;&#x000b1;&#x02009;1</td><td>38&#x02009;&#x000b1;&#x02009;1</td><td>36&#x02009;&#x000b1;&#x02009;1</td></tr><tr valign="top"><td>VCO₂ (ml/min/kg)</td><td>42&#x02009;&#x000b1;&#x02009;2</td><td>43&#x02009;&#x000b1;&#x02009;2</td><td>41&#x02009;&#x000b1;&#x02009;2</td></tr><tr valign="top"><td>RER</td><td>1.11&#x02009;&#x000b1;&#x02009;0.02</td><td>1.11&#x02009;&#x000b1;&#x02009;0.01</td><td>1.12&#x02009;&#x000b1;&#x02009;0.02</td></tr><tr valign="top"><td> <bold>Endurance performance test</bold> </td><td/><td/><td/></tr><tr valign="top"><td>Heart rate (beats/min)</td><td>153&#x02009;&#x000b1;&#x02009;13</td><td>153&#x02009;&#x000b1;&#x02009;13</td><td>152&#x02009;&#x000b1;&#x02009;13</td></tr><tr valign="top"><td>VO₂ (ml/min/kg)</td><td>35&#x02009;&#x000b1;&#x02009;4</td><td>34&#x02009;&#x000b1;&#x02009;4</td><td>30&#x02009;&#x000b1;&#x02009;4*#</td></tr><tr valign="top"><td>VCO₂ (ml/min/kg)</td><td>33&#x02009;&#x000b1;&#x02009;5</td><td>34&#x02009;&#x000b1;&#x02009;4</td><td>29&#x02009;&#x000b1;&#x02009;4*#</td></tr><tr valign="top"><td>RER</td><td>0.95&#x02009;&#x000b1;&#x02009;0.05</td><td>0.99&#x02009;&#x000b1;&#x02009;0.05*</td><td>0.97&#x02009;&#x000b1;&#x02009;0.03</td></tr><tr valign="top"><td>RPE</td><td>16&#x02009;&#x000b1;&#x02009;4</td><td>16&#x02009;&#x000b1;&#x02009;4</td><td>14&#x02009;&#x000b1;&#x02009;5</td></tr></tbody></table><table-wrap-foot><p>*p&#x02009;&#x0003c;&#x02009;0.05 compared with water; # p&#x02009;&#x0003c;&#x02009;0.05 compared with sports drink.</p></table-wrap-foot></table-wrap></p><p>Time to exhaustion was greatest in LFM (p&#x02009;&#x0003c;&#x02009;0.05), followed by SPD and WT (Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>). The longer exercise duration in LFM also elicited the greatest body mass loss compared to the other two trials (LFM 1.0&#x02009;&#x000b1;&#x02009;0.3&#x000a0;kg; SPD 0.8&#x02009;&#x000b1;&#x02009;0.3&#x000a0;kg; WT 0.6&#x02009;&#x000b1;&#x02009;0.3&#x000a0;kg; p&#x02009;&#x0003c;&#x02009;0.05). Blood lactate was higher after the endurance capacity test with WT (2.4&#x02009;&#x000b1;&#x02009;1.9&#x000a0;mmol/dl) than SPD (1.0&#x02009;&#x000b1;&#x02009;0.7&#x000a0;mmol/dl; p&#x02009;&#x0003c;&#x02009;0.05), but neither was significantly different in comparison to LFM (1.7&#x02009;&#x000b1;&#x02009;1.4&#x000a0;mmol/dl; p&#x02009;&#x0003e;&#x02009;0.05).<fig id="Fig1"><label>Figure 1</label><caption><p> <bold>Endurance capacity between trials.</bold> *p&#x02009;&#x0003c;&#x02009;0.05 vs. water. #p&#x02009;&#x0003c;&#x02009;0.05 vs. sports drink.</p></caption><graphic xlink:href="12970_2014_49_Fig1_HTML" id="MO1"/></fig></p></sec><sec id="Sec11" sec-type="discussion"><title>Discussion</title><p>Our study is the first to demonstrate the ergogenic effects of ingesting lactose-free milk on endurance cycling capacity in young adults. The ingestion of lactose-free milk during recovery after glycogen depleting exercise almost doubled (93% increase) the subsequent cycle time to exhaustion in comparison with water, and further extended exercise duration by 34% compared to sports drink. Despite identical workloads during the endurance capacity test with all three drinks, the oxygen uptake and concomitant expired carbon dioxide during exercise was lower after the ingestion of lactose-free milk compared to water and sports drink. This may indicate that the ingestion of lactose-free milk during recovery enhances metabolic efficiency during the subsequent exercise bout.</p><p>Although muscle glycogen resynthesis was not measured in the present study, this may have been enhanced with the ingestion of lactose-free milk compared to the isocaloric sports drink [<xref ref-type="bibr" rid="CR9">9</xref>,<xref ref-type="bibr" rid="CR17">17</xref>]. Previous work has generally shown a greater muscle glycogen resynthesis with carbohydrate-protein drinks compared to carbohydrate-only drinks [<xref ref-type="bibr" rid="CR8">8</xref>,<xref ref-type="bibr" rid="CR18">18</xref>], although mixed results exist [<xref ref-type="bibr" rid="CR19">19</xref>]. The interpretation of these results, however, is limited by a higher caloric content in the carbohydrate-protein drinks than the carbohydrate-only drinks. Whilst it has been suggested that a greater carbohydrate intake may enhance glycogen resynthesis to match that after carbohydrate-protein intake [<xref ref-type="bibr" rid="CR20">20</xref>], other comparisons of isocaloric carbohydrate-only and carbohydrate-protein drinks nevertheless demonstrate a more efficient replenishment of muscle glycogen after carbohydrate-protein intake [<xref ref-type="bibr" rid="CR9">9</xref>,<xref ref-type="bibr" rid="CR17">17</xref>]. Moreover, drinks containing too much carbohydrate (8&#x02013;10%), similar to the sports drink in the present study, may also delay gastric emptying and fluid absorption [<xref ref-type="bibr" rid="CR15">15</xref>]. This could result in reduced muscle and liver glycogen stores. Taken together, it is thus possible that muscle glycogen resynthesis was enhanced with ingestion of lactose-free milk compared to sports drink (and water) in the present study. An improved glycogen repletion would enable a better metabolic efficiency during the cycle ride to exhaustion after ingestion of lactose-free milk and explain, in part, the lower oxygen uptake required to sustain exercise and the resultant extension in exercise duration [<xref ref-type="bibr" rid="CR8">8</xref>].</p><p>Differences in the type of carbohydrate in lactose-free milk compared to sports drink could also have contributed to the lower oxygen consumption observed during the cycle ride to exhaustion in the present study. Whilst SPD contained 6% glucose and 4% sucrose, LFM consisted of 2.5% glucose and 2.5% galactose. Exogenous glucose oxidation during exercise has a maximum of 1.0&#x02013;1.1&#x000a0;g/min, whereas galactose utilization is limited to ~0.4&#x000a0;g/min [<xref ref-type="bibr" rid="CR14">14</xref>]. This lower oxidation rate of galactose compared to glucose and sucrose [<xref ref-type="bibr" rid="CR21">21</xref>] is due to the conversion of glucose in the liver, before subsequent utilization by the skeletal muscles. The ingested galactose may also have been synthesized to form glycogen during the recovery period. It is thus possible that the galactose in lactose-free milk slowed the oxidation process, resulting in the lowered oxygen consumption and carbon dioxide production observed [<xref ref-type="bibr" rid="CR4">4</xref>].</p><p>A reduction in muscle damage (i.e. creatine phosphokinase levels [<xref ref-type="bibr" rid="CR18">18</xref>]) with lactose-free milk ingestion is another likely candidate to explain the physiological differences observed during the cycle ride to exhaustion in the present study. This may be achieved via a reduced rate of protein breakdown (i.e. improved whole body net protein balance [<xref ref-type="bibr" rid="CR22">22</xref>]) and greater myofibrillar muscle protein synthesis through p70S6K, downstream of mTOR [<xref ref-type="bibr" rid="CR23">23</xref>], following the ingestion of protein after exercise. Whether the performance benefits associated with lactose-free milk ingestion in the present study are also relevant to a more aerobically fit population (e.g. VO2peak 65 &#x000b1; 7 mL/min/kg [<xref ref-type="bibr" rid="CR24">24</xref>]) is unclear, and requires further investigation.</p><p>Whilst the underlying mechanism(s) for the lower oxygen consumption and carbon dioxide production with LFM cannot be elucidated from the present study, our results nonetheless clearly demonstrate the efficacy of LFM as a recovery drink to enhance subsequent exercise performance. The extended time to exhaustion after LFM ingestion observed in the present study is in agreement with previous work that have investigated chocolate milk as a recovery drink [<xref ref-type="bibr" rid="CR5">5</xref>,<xref ref-type="bibr" rid="CR6">6</xref>]. Those studies have reported that the ingestion of chocolate milk as a recovery drink results in at least comparable, if not better, subsequent exercise performance compared to other commercially available fluid- and carbohydrate-replacement drinks [<xref ref-type="bibr" rid="CR5">5</xref>,<xref ref-type="bibr" rid="CR6">6</xref>].</p><p>Further work is necessary to determine the mechanism(s) underlying the extended endurance capacity duration after lactose-free milk ingestion, compared to an isocaloric and isovolumic sports drink. In the present study, participants could not be blinded to the test drink and this may have psychologically affected them during the cycle ride to exhaustion. Whilst the efficacy of lactose-free milk as a recovery drink is clear, the product used the present study was manufactured in-house and fortified with sodium and potassium. This is a key limitation for the immediate applicability of our results to the lactose intolerant population, as the lactose-free milk investigated in the present study is not currently commercially available.</p></sec><sec id="Sec12" sec-type="conclusion"><title>Conclusion</title><p>This study demonstrates an increased cycle time to exhaustion after ingesting lactose-free milk as a recovery drink, as compared to water and sports drink. The extended exercise duration may be explained by a greater metabolic efficiency, as the amount of oxygen consumed and carbon dioxide produced were both reduced with lactose-free milk. Consequently, lactose-free milk may be appropriate as a recovery drink for the general population and can be used as a substitute for normal milk in lactose intolerant individuals.</p></sec>

Genome-wide profiling of mouse RNA secondary structures reveals key features of the mammalian transcriptome

Could not extract abstract

<contrib contrib-type="author"><name><surname>Incarnato</surname><given-names>Danny</given-names></name><address><email>danny.incarnato@hugef-torino.org</email></address><xref ref-type="aff" rid="Aff1"/><xref ref-type="aff" rid="Aff2"/></contrib><contrib contrib-type="author"><name><surname>Neri</surname><given-names>Francesco</given-names></name><address><email>francesco.neri@hugef-torino.org</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author"><name><surname>Anselmi</surname><given-names>Francesca</given-names></name><address><email>francesca.anselmi@hugef-torino.org</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author" corresp="yes"><name><surname>Oliviero</surname><given-names>Salvatore</given-names></name><address><email>salvatore.oliviero@hugef-torino.org</email></address><xref ref-type="aff" rid="Aff1"/><xref ref-type="aff" rid="Aff3"/></contrib><aff id="Aff1"><label/>Human Genetics Foundation (HuGeF), via Nizza 52, Torino, 10126 Italy </aff><aff id="Aff2"><label/>Dipartimento di Biotecnologie Chimica e Farmacia, Universit&#x000e0; di Siena, via Fiorentina, Siena, 1-53100 Italy </aff><aff id="Aff3"><label/>Dipartimento di Scienze della Vita e Biologia dei Sistemi, Universit&#x000e0; di Torino, Via Accademia Albertina, Torino, 13-10123 Italy </aff>

Genome Biology

<sec id="Sec1" sec-type="introduction"><title>Background</title><p>The development of high-throughput methods for the analysis of the epigenome and transcriptome have led to the discovery of thousands of previously unannotated transcripts [<xref ref-type="bibr" rid="CR1">1</xref>,<xref ref-type="bibr" rid="CR2">2</xref>], many of which lack the ability to encode proteins [<xref ref-type="bibr" rid="CR3">3</xref>-<xref ref-type="bibr" rid="CR6">6</xref>], as further proven by genome-wide ribosome profiling approaches [<xref ref-type="bibr" rid="CR7">7</xref>]. While mechanisms of action have been elucidated for a small fraction of these non-coding RNAs (ncRNAs), for most the ways by which they contribute to gene regulation still remain unclear. One of the most intriguing modes of action proposed for long ncRNAs (lncRNAs) is their potential to act as modular scaffolds for the assembly of large multi-protein complexes [<xref ref-type="bibr" rid="CR4">4</xref>,<xref ref-type="bibr" rid="CR8">8</xref>], although the mechanistic aspects of these interactions are largely unknown. As learned from small nuclear ribonucleic particle (snRNP) complexes [<xref ref-type="bibr" rid="CR9">9</xref>,<xref ref-type="bibr" rid="CR10">10</xref>], most ncRNAs are thought to exert their functions by folding into locally stable secondary structures that may provide anchoring sites for interacting proteins. For example, it has been shown in <italic>Drosophila melanogaster</italic> that both MLE and MSL2 proteins of the MSL complex act by binding to conserved structural domains of the <italic>roX1/2</italic> ncRNAs, which then mediate targeting to the X chromosome to regulate dosage compensation in fruitfly [<xref ref-type="bibr" rid="CR11">11</xref>,<xref ref-type="bibr" rid="CR12">12</xref>]. Furthermore, in differentiating mouse embryonic stem cells (ESCs), MLL1 protein has been shown to be required for the transcriptional activation of <italic>Hoxa6/7</italic> genes, and its recruitment to chromatin is mediated by interaction with a stem-loop structure located in the 3&#x02032; region of the <italic>Mistral</italic> lncRNA [<xref ref-type="bibr" rid="CR13">13</xref>].</p><p>The growing number of annotated transcripts has outpaced the efficient analysis of their structure; at present, structural information exists for only a very tiny minority of annotated RNAs. To address this need, over the past few years various enzymatic- and chemical-based approaches have been proposed for the discovery of secondary structures for thousands of RNAs at a time [<xref ref-type="bibr" rid="CR14">14</xref>-<xref ref-type="bibr" rid="CR18">18</xref>]; however, all these methods are based on the assumption that <italic>in vitro</italic> folding may be representative of native RNA structures <italic>in vivo</italic>. While for certain small RNAs the <italic>in vitro</italic> folding landscape recapitulates well the <italic>in vivo</italic> one [<xref ref-type="bibr" rid="CR19">19</xref>-<xref ref-type="bibr" rid="CR21">21</xref>], long RNAs often exhibit rugged folding landscapes that lead <italic>in vitro</italic> to the prevalence of kinetically trapped intermediates and misfolded structures [<xref ref-type="bibr" rid="CR22">22</xref>-<xref ref-type="bibr" rid="CR24">24</xref>]. For example, <italic>in vitro</italic> folding of the RNAse P ribozyme is a slow process that takes several minutes and requires escape from a kinetic trap [<xref ref-type="bibr" rid="CR23">23</xref>,<xref ref-type="bibr" rid="CR25">25</xref>]. Comparative analysis of <italic>in vivo</italic> and <italic>in vitro</italic> probing data on human telomerase RNA revealed that while the 3&#x02032;-terminal small nucleolar RNA (snoRNA)-like domain folds into comparable structures in the two conditions, the 5&#x02032; template domain exhibits very different foldings [<xref ref-type="bibr" rid="CR26">26</xref>].</p><p>Two main scenarios can explain the differences observed in RNA folding <italic>in vitro</italic> and <italic>in vivo</italic>. The first is based on the assumption that, in the cell, most nascent transcripts are likely to fold during transcription [<xref ref-type="bibr" rid="CR20">20</xref>,<xref ref-type="bibr" rid="CR27">27</xref>,<xref ref-type="bibr" rid="CR28">28</xref>]. In this perspective, the elongation rate of RNA polymerase, as well as the directionality of transcription, may influence the order and the speed of the folding events, thus preventing the formation of non-native, kinetically trapped intermediates [<xref ref-type="bibr" rid="CR29">29</xref>]. The second, which does not exclude the first, is that many specific as well as non-specific RNA binding proteins (RBPs) may act as RNA chaperones, thus directing and stabilizing RNA folding [<xref ref-type="bibr" rid="CR30">30</xref>-<xref ref-type="bibr" rid="CR33">33</xref>]. To overcome the issues introduced by the study of RNA folding <italic>in vitro</italic>, two recent reports analyzed the structures of <italic>Saccharomyces cerevisiae</italic> and <italic>Arabidopsis thaliana</italic> RNAs by treating the cells with dimethyl sulfate (DMS) [<xref ref-type="bibr" rid="CR34">34</xref>,<xref ref-type="bibr" rid="CR35">35</xref>].</p><p>We present here a new method, named chemical inference of RNA structures followed by massive parallel sequencing (CIRS-seq), that allows genome-wide investigation of native deproteinized RNA secondary structures by exploiting the capacity of DMS and N-cyclohexyl-N&#x02019;-(2-morpholinoethyl)carbodiimide metho-p-toluenesulfonate (CMCT) to specifically react with RNA unpaired bases. Our approach, applied to mouse ESCs, allowed us to obtain single-base resolution structural information for thousands of transcripts in their native deproteinized conformation, revealing the structural complexity of the mammalian transcriptome.</p></sec><sec id="Sec2" sec-type="results"><title>Results</title><sec id="Sec3"><title>CIRS-seq enables accurate transcriptome-wide inference of single-stranded RNA residues</title><p>The CIRS-seq method (Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>) is based on the use of DMS, which mainly methylates N1 of adenosine and N3 of cytosine [<xref ref-type="bibr" rid="CR36">36</xref>,<xref ref-type="bibr" rid="CR37">37</xref>], and CMCT, which primarily forms adducts with N1 and N3 of pseudouridine, N3 of uridine, and, to a lesser extent, N1 of guanosine and inosine [<xref ref-type="bibr" rid="CR38">38</xref>-<xref ref-type="bibr" rid="CR40">40</xref>] but only when these residues are in single-stranded conformation. Treatment of RNA with the two reagents enables the detection of unpaired nucleotide positions due to the modification-induced reverse transcription (RT) stop one nucleotide downstream of the modified residue. To carry out CIRS-seq, we first optimized treatments to achieve similar degrees of modification with the two reagents at different concentrations, as measured by reduction of the full-length reverse transcription product for a test RNA following reaction with either DMS or CMCT (Figure S1 in Additional file <xref rid="MOESM1" ref-type="media">1</xref>).<fig id="Fig1"><label>Figure 1</label><caption><p> <bold>Overview of the CIRS-seq method.</bold> Cells are harvested and lysed in isotonic buffer, then treated with Proteinase K to unmask protein-bound regions of RNAs. The whole cell population of RNAs in their native deproteinized conformation is probed with either DMS or CMCT to modify unpaired bases. A non-treated control is also produced to allow further mapping of natural RT stops. After modification, the RNAs from the three populations are reverse transcribed, and cDNA is adapter ligated for high-throughput sequencing. Mapping reads to the transcriptome provide information regarding how many RT stops occurred at each position of the analyzed transcripts. The non-treated (NT) signal at each position is then subtracted from the DMS and CMCT signals to obtain the raw reactivity profile at base resolution. After scaling each data point above the 90th percentile to the 90th percentile, reactivity at each position is divided by the 90th percentile (90% Winsorising) to obtain the normalized reactivity.</p></caption><graphic xlink:href="13059_2014_491_Fig1_HTML" id="MO1"/></fig></p><p>To perform transcriptome-wide probing of RNAs in their native deproteinized conformation, we lysed mouse ESCs in an isotonic buffer, and treated the lysate with Proteinase K to unmask regions of RNAs bound by proteins, without affecting the RNA structure (supplementary Materials and methods in Additional file <xref rid="MOESM1" ref-type="media">1</xref>). ESC lysates were then treated with DMS or CMCT, and total RNA was extracted following reaction quenching. Extracted RNA was subjected to random-primed RT. A non-treated control was also produced to determine naturally occurring RT stops. The generated cDNAs were adapter ligated and subjected to high-throughput sequencing using the Illumina platform, resulting in about 90&#x02009;&#x000d7;&#x02009;10⁶ deep-sequencing reads for each treatment, across two biological replicates.</p><p>Since proper analysis of RNA folding requires correct annotation of transcript sequences, reads were mapped to a recently published variant of the mm9 assembly that integrates single-nucleotide variants from the E14 ESC line [<xref ref-type="bibr" rid="CR41">41</xref>], and we obtained a similar distribution of read mappings across all samples (Figure S2A in Additional file <xref rid="MOESM1" ref-type="media">1</xref>). Estimation of transcript abundances using CIRS-seq data correlated well across treatments, and with canonical RNA-seq data (<italic>R</italic> &#x02265;0.9, Spearman correlation; Figure S2B in Additional file <xref rid="MOESM1" ref-type="media">1</xref>), showing that the CIRS-seq method enables unbiased probing of RNAs. At the current coverage, we obtained structural information for approximately 30,000 transcripts, belonging to approximately 13,000 genes (Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>a; Figure S2C in Additional file <xref rid="MOESM1" ref-type="media">1</xref>).<fig id="Fig2"><label>Figure 2</label><caption><p> <bold>Validation of CIRS-seq data. (a)</bold> Distribution of transcripts with at least one RT stop on average per base. <bold>(b)</bold> Scatter plot of normalized reactivities in the two biological replicates of CIRS-seq. Reactivities are averaged in 10-nucleotide windows, with an offset of 5 nucleotides (Pearson&#x02019;s correlation coefficient = 0.90. <bold>(c)</bold> Normalized reactivity profiles for the glutamic acid tRNA and overlay of reactivity data on the phylogenetically derived secondary structure. Yellow arrows indicate highly reactive positions (reactivity &#x0003e;0.7). Bases are color coded according to their reactivity. <bold>(c)</bold> Normalized reactivity profiles for the U5 snRNA and overlay of reactivity data on phylogenetically derived secondary structure. The structure of the U5 human homolog is also shown, with superimposed DMS/CMCT-reactive positions from [<xref ref-type="bibr" rid="CR55">55</xref>]. The colors correspond to different degrees of chemical modification (purple, strong; yellow, medium; green, weak). Yellow arrows indicate highly reactive positions (reactivity &#x0003e;0.7). Bases are color coded according to their reactivity.</p></caption><graphic xlink:href="13059_2014_491_Fig2_HTML" id="MO2"/></fig></p><p>As a quantitative measure of the probability of observing a RT stop specifically induced by our treatment, we calculated raw reactivity scores as the base 2 logarithm ratio of the normalized read counts for the DMS/CMCT treatment at a given position of a transcript, and the normalized read counts at the same position in the non-treated control. The final normalization (Supplementary methods in Additional file <xref rid="MOESM1" ref-type="media">1</xref>) yielded reactivity values ranging from 0 to 1, and positions with reactivities &#x0003e;0 and &#x0003c;0.3, 0.3 to 0.7, or &#x0003e;0.7 were designated as weakly, moderately, or highly reactive, respectively [<xref ref-type="bibr" rid="CR42">42</xref>]. Correlation analysis of reactivity values across the top 75th percentile of covered transcripts revealed the high reproducibility of CIRS-seq (<italic>R</italic> = 0.90, Pearson correlation; Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>b); therefore, we combined the two replicates for further analysis.</p><p>Collectively, we obtained structural data for 1,190,948 and 1,080,859 nucleotides in the DMS (weak, 13.6%; moderate, 49.9%; high, 36.5%) and CMCT (weak, 16.6%; moderate, 54.7%; high, 28.7%) treatments, respectively (Figure S2D in Additional file <xref rid="MOESM1" ref-type="media">1</xref>). To validate CIRS-seq, we overlaid reactivity data on the well characterized structures of tRNAs [<xref ref-type="bibr" rid="CR43">43</xref>,<xref ref-type="bibr" rid="CR44">44</xref>] (Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>c; Figure S3A,B in Additional file <xref rid="MOESM1" ref-type="media">1</xref>), and observed that all the highly reactive residues were almost completely confined to the tRNAs' D and anticodon arm loops, suggesting a high overall accuracy for our method.</p><p>Despite the respective strong preference of DMS and CMCT for A/C and G/U residues, we also observed non-canonical reactivities in both treatments. Our data are in agreement with previous reports showing DMS reactivity with G/U residues [<xref ref-type="bibr" rid="CR36">36</xref>,<xref ref-type="bibr" rid="CR45">45</xref>-<xref ref-type="bibr" rid="CR47">47</xref>] and CMCT reactivity with cytosines [<xref ref-type="bibr" rid="CR48">48</xref>-<xref ref-type="bibr" rid="CR50">50</xref>]. We observed a significant increase in the accuracy of the <italic>de novo</italic> prediction of structures when considering also these non-canonical reactivities as both the canonical and non-canonical reactivities lay within single-stranded regions (Figure S3C in Additional file <xref rid="MOESM1" ref-type="media">1</xref>). Moreover, overlaying reactivity data on the known structures of U5 and U1 small nuclear RNAs (snRNAs) and U3 snoRNA (Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>d; Figure S4A,B in Additional file <xref rid="MOESM1" ref-type="media">1</xref>) showed that the Proteinase K treatment enabled high-resolution determination of secondary structures at the level of protein-masked regions of RNAs without losing the proper folding. In fact, internal loop IL2/IL2&#x02032; of U5, box B/C of U3 and loop II of U1 are bound <italic>in vivo</italic> by, respectively, a 116&#x000a0;kDa protein (Snu114p yeast homolog) [<xref ref-type="bibr" rid="CR51">51</xref>], a 15.5&#x000a0;K protein [<xref ref-type="bibr" rid="CR52">52</xref>,<xref ref-type="bibr" rid="CR53">53</xref>], and the U1A protein [<xref ref-type="bibr" rid="CR10">10</xref>]; these regions showed very high reactivity to DMS/CMCT treatments and were almost completely resolved by CIRS-seq. Overall, for the set of analyzed structures (Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>), 80.6% of the highly reactive residues were located within single-stranded regions. Of the 19.4% of the highly reactive residues located within regions of the known structures annotated as double-stranded, 84.2% were positioned at the end of helices or adjacent to bulges/loops. These regions were previously shown to be subjected to structural flexibility, so chemical reagents can easily modify these terminal residues [<xref ref-type="bibr" rid="CR34">34</xref>,<xref ref-type="bibr" rid="CR54">54</xref>]. When accounting for these additional accessible positions, the overall true positive rate of our method rose to 96.3%.<table-wrap id="Tab1"><label>Table 1</label><caption><p> <bold>CIRS-seq efficiency on validated secondary structures</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th/><th/><th colspan="2"> <bold>Including helix termini</bold> </th><th colspan="2"> <bold>Excluding helix termini</bold> </th></tr><tr valign="top"><th> <bold>ENSEMBL ID</bold> </th><th> <bold>Symbol</bold> </th><th> <bold>TP (%)</bold> </th><th> <bold>FP (%)</bold> </th><th> <bold>TP (%)</bold> </th><th> <bold>FP (%)</bold> </th></tr></thead><tbody><tr valign="top"><td>ENSMUST00000082420</td><td>mt-Te</td><td>87.5</td><td>12.5</td><td>87.5</td><td>12.5</td></tr><tr valign="top"><td>ENSMUST00000082389</td><td>mt-Ti</td><td>75.0</td><td>25.0</td><td>100.0</td><td>0.0</td></tr><tr valign="top"><td>ENSMUST00000082399</td><td>mt-Tn</td><td>90.0</td><td>10.0</td><td>90.0</td><td>10.0</td></tr><tr valign="top"><td>ENSMUST00000083033</td><td>U1</td><td>83.4</td><td>16.6</td><td>94.7</td><td>5.3</td></tr><tr valign="top"><td>ENSMUST00000082496</td><td>U5</td><td>73.7</td><td>26.3</td><td>100.0</td><td>0.0</td></tr><tr valign="top"><td>ENSMUST00000082466</td><td>U3</td><td>75.0</td><td>25.0</td><td>96.9</td><td>1.1</td></tr><tr valign="top"><td colspan="2">Total</td><td>80.6</td><td>19.4</td><td>96.3</td><td>3.7</td></tr></tbody></table><table-wrap-foot><p>Percentages of true positive (TP) and false positive (FP) highly reactive positions for known secondary structures.</p></table-wrap-foot></table-wrap></p><p>Collectively, this analysis proves the high accuracy of CIRS-seq, and provides a nucleotide-resolution panorama of the mouse ESC RNA structurome.</p></sec><sec id="Sec4"><title>CIRS-seq data allow accurate secondary structure prediction</title><p>Next, we verified the ability of CIRS-seq to infer <italic>de novo</italic> secondary structures. Constraints derived from chemical probing data may significantly improve the accuracy of RNA secondary structure prediction tools [<xref ref-type="bibr" rid="CR42">42</xref>,<xref ref-type="bibr" rid="CR56">56</xref>]. We chose the U2 and low-abundance U12 snRNAs, and the valine and threonine tRNAs, whose structures were previously experimentally defined [<xref ref-type="bibr" rid="CR57">57</xref>,<xref ref-type="bibr" rid="CR58">58</xref>], or can be easily derived from phylogenetic analysis. We used the RNAStructure tool [<xref ref-type="bibr" rid="CR59">59</xref>] to devise secondary structures by imposing constraints for unpaired positions. This tool can accept chemical probing data in the form of SHAPE data files, allowing more comprehensive modeling of the structure according to the CIRS-seq-derived data compared with hard constraints-based methods. For both the unconstrained minimum free energy (MFE) and the CIRS-seq constrained secondary structures, we calculated the positive predictive value (PPV) as the fraction of base-pairs present in the predicted structure that are also present in the validated structure, and the sensitivity as the fraction of base-pairs present in the validated structure that are also in the predicted structure (Table&#x000a0;<xref rid="Tab2" ref-type="table">2</xref>). Notably, CIRS-seq-derived structures for all the four transcripts analyzed showed higher similarity to the known structures (Figure&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>a,b; Figure S5A,B in Additional file <xref rid="MOESM1" ref-type="media">1</xref>); on average, the CIRS-seq-guided folding outperformed the MFE unconstrained predictions in terms of both PPV and sensitivity (PPV 0.53 and sensitivity 0.57 for unconstrained MFE structures; PPV 0.95 and sensitivity 0.95 for CIRS-seq constrained structures). This analysis demonstrates that the use of CIRS-seq data improves the accuracy of RNA secondary structure prediction tools, and that low-abundance transcripts can be successfully probed by CIRS-seq.<table-wrap id="Tab2"><label>Table 2</label><caption><p> <bold>Statistics for CIRS-seq</bold> <bold><italic>de novo</italic></bold> <bold>inferred secondary structures</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th/><th/><th colspan="2"> <bold>Unconstrained (MFE)</bold> </th><th colspan="2"> <bold>CIRS-seq constrained</bold> </th></tr><tr valign="top"><th> <bold>ENSEMBL ID</bold> </th><th> <bold>Symbol</bold> </th><th> <bold>PPV</bold> </th><th> <bold>Sensitivity</bold> </th><th> <bold>PPV</bold> </th><th> <bold>Sensitivity</bold> </th></tr></thead><tbody><tr valign="top"><td>ENSMUST00000101806</td><td>U2</td><td>0.68</td><td>0.89</td><td>1.00</td><td>1.00</td></tr><tr valign="top"><td>ENSMUST00000083242</td><td>U12</td><td>0.80</td><td>0.84</td><td>1.00</td><td>0.95</td></tr><tr valign="top"><td>ENSMUST00000082389</td><td>mt-Tv</td><td>0.22</td><td>0.20</td><td>1.00</td><td>1.00</td></tr><tr valign="top"><td>ENSMUST00000083422</td><td>mt-Tt</td><td>0.41</td><td>0.35</td><td>0.81</td><td>0.85</td></tr><tr valign="top"><td colspan="2">Average</td><td>0.53</td><td>0.57</td><td>0.95</td><td>0.95</td></tr></tbody></table><table-wrap-foot><p>Positive predictive value (PPV) and sensitivity measures calculated for both the unconstrained minimum free energy (MFE) and CIRS-seq constrained structures.</p></table-wrap-foot></table-wrap><fig id="Fig3"><label>Figure 3</label><caption><p> <bold>CIRS-seq data allow correct inference of native deproteinized RNA secondary structures. (a)</bold> Normalized reactivity profiles for the U2 snRNA and overlay of reactivity data on the secondary structure inferred from chemical constraints. Bases are color coded according to their reactivity. The structure of the human ortholog with superimposed SHAPE-reactive positions from [<xref ref-type="bibr" rid="CR57">57</xref>], and the unconstrained MFE structure are also shown. <bold>(b)</bold> Normalized reactivity profiles for the low-abundance U12 snRNA and overlay of reactivity data on the secondary structure inferred from chemical constraints. Bases are color coded according to their reactivity. The structure of the U12 <italic>A. thaliana</italic> ortholog with superimposed DMS/SHAPE-reactive positions from [<xref ref-type="bibr" rid="CR58">58</xref>], and the unconstrained MFE structure are also shown.</p></caption><graphic xlink:href="13059_2014_491_Fig3_HTML" id="MO3"/></fig></p></sec><sec id="Sec5"><title>CIRS-seq reveals structural features of mammalian mRNAs and ncRNAs</title><p>Thanks to the high resolution enabled by CIRS-seq, we then investigated the structural features of mouse mRNAs, and looked for structural differences across transcript regions. We selected approximately 9,500 mRNAs, in which the DMS/CMCT treatment induced, on average, at least one RT stop per nucleotide (Supplementary methods in Additional file <xref rid="MOESM1" ref-type="media">1</xref>).</p><p>Meta-analysis of average reactivity across UTRs and coding regions revealed a strong reduction of reactivity scores in the 50 nucleotides of the 5&#x02032; UTR immediately preceding the Kozak sequence (average 0.165) compared with the first and last 100 nucleotides of the coding region (average 0.208, <italic>P</italic>-value 3.0e-374, Wilcoxon rank sum test; Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>a,b). Moreover, a significant reduction of reactivity was also observed in the first 50 nucleotides of the 3&#x02032; UTR immediately downstream of the stop codon (average 0.172, <italic>P</italic>-value 1.1e-243, Wilcoxon rank sum test). These results differ from what has been recently observed in <italic>A. thaliana</italic>, where the coding region is more structured than the UTRs [<xref ref-type="bibr" rid="CR34">34</xref>]. We also identified a significant increase of reactivity score at the level of the Kozak sequence (average 0.229) with a maximum of reactivity on the base immediately preceding the AUG (average 0.345), and on the stop codon beginning three nucleotides upstream (average 0.226), compared with the coding region (<italic>P</italic>-values 4.0e-24 and 6.5e-8, respectively, Wilcoxon rank sum test; Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>c), revealing a markedly reduced probability of base-pairing in these regions. The reduced base-pairing on the Kozak sequence and around the stop codon suggests that a more accessible context in these regions of protein-coding transcripts may facilitate both the entry and the detachment of ribosomes.<fig id="Fig4"><label>Figure 4</label><caption><p> <bold>Transcriptome-wide analysis of mRNAs reveals structural features of protein-coding and non-coding transcripts. (a)</bold> Meta-gene analysis across the last 50 nucleotides of the 5&#x02032; UTR, the first and last 100 nucleotides of the coding region, and the first 50 nucleotides of the 3&#x02032; UTR of approximately 9,500 mRNAs. <bold>(b)</bold> Average reactivity of the 5&#x02032; UTR, coding region, and 3&#x02032; UTR. <bold>(c)</bold> Average reactivity on the Kozak sequence (&#x02212;6/+1 nucleotides around AUG), coding region, and stop codon (+3 nucleotides upstream). <bold>(d)</bold> Average reactivity for the first, second, and third base of each coding sequence codon, and for the first, second, and third base of the 5&#x02032; UTR and 3&#x02032; UTR, respectively, in the first and last 99 nucleotides of the coding region, last 48 nucleotides of the 5&#x02032; UTR, and first 48 nucleotides of the 3&#x02032; UTR. <bold>(e)</bold> Box-plot of base-normalized average CIRS-seq reactivities for protein-coding and non-coding RNAs, calculated on all transcript positions with sequencing depth &#x0003e;50&#x02009;&#x000d7;&#x02009;.</p></caption><graphic xlink:href="13059_2014_491_Fig4_HTML" id="MO4"/></fig></p><p>We next analyzed the first and last 99 nucleotides of mRNA coding regions to determine if the previously reported periodic signal of three nucleotides [<xref ref-type="bibr" rid="CR14">14</xref>,<xref ref-type="bibr" rid="CR34">34</xref>] was conserved also in mouse. To this end, we observed that mouse protein-coding transcripts, similar to <italic>A. thaliana</italic> and <italic>S. cerevisiae</italic> mRNAs, exhibit a strong three-nucleotide periodicity across the coding region that was not observed within the UTRs (Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>d). The second and third nucleotides of each codon were highly structured and exhibited lower average reactivities (average 0.205 and 0.199, respectively), with the third nucleotide being the less reactive (<italic>P</italic>-value 1.7e-07, Wilcoxon rank sum test), while the first nucleotide was the less structured and significantly more reactive to DMS/CMCT treatment than the second and third (average 0.220, <italic>P</italic>-values 2.0e-12 and 8.2e-12, respectively, Wilcoxon rank sum test). Taken together these results suggest a deep involvement of RNA secondary structures in driving and regulating translation efficiency.</p><p>Analysis of the RNA structure is particularly relevant for ncRNAs as they are thought to exert their function by interacting with other molecules via their secondary structure. We then sought to determine whether an overall structural difference exists between protein coding RNAs and different classes of ncRNA transcripts. To avoid biases due to differential coverage, only transcript positions with sequencing depth greater than 50&#x000d7; were considered (Supplementary methods in Additional file <xref rid="MOESM1" ref-type="media">1</xref>). Analysis of normalized reactivity showed a significantly lower average reactivity of snoRNAs (average 0.282, <italic>P</italic>-value 1.1e-87, Wilcoxon rank sum test), snRNAs (average 0.295, <italic>P</italic>-value 2.1e-146, Wilcoxon rank sum test), tRNAs (average 0.251, <italic>P</italic>-value 2.5e-9, Wilcoxon rank sum test), and long intergenic non-coding RNA (lincRNAs; average 0.309, <italic>P</italic>-value 7.4e-48, Wilcoxon rank sum test) compared with mRNAs (average 0.366) (Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>e). Collectively, these data reveal a higher structuring of ncRNA transcripts compared with mRNAs.</p></sec><sec id="Sec6"><title>CIRS-seq identifies structural requirements of RNA binding proteins</title><p>RNA-protein interactions are strongly influenced by secondary structures. Determining the structural requirements for RBPs to bind to their cognate targets is required to understand their roles and mechanisms of action. To this end, we analyzed from a structural perspective the binding sites of the highly conserved RBP Lin28a. Lin28a is highly expressed in ESCs, and is one of the factors required for the reprogramming of human fibroblasts to induced pluripotent stem cells [<xref ref-type="bibr" rid="CR60">60</xref>]. To investigate the structural requirements of Lin28a binding, we analyzed a previously published CLIP-seq dataset of Lin28a in ESCs [<xref ref-type="bibr" rid="CR61">61</xref>]. We identified peaks of Lin28a enrichment across the mouse transcriptome, and calculated average reactivity on a window of 300 nucleotides surrounding summits of the peaks (Figure&#x000a0;<xref rid="Fig5" ref-type="fig">5</xref>a). While more distal regions around the Lin28a peaks showed a level of reactivity comparable to that of the coding sequence (average 0.21), in agreement with a preferential positioning of Lin28a binding sites within this region, we observed a significant and progressive increase in the accessibility proceeding toward the peak summits (maximum 0.34, &#x000b1;25 nucleotides average 0.27, <italic>P</italic>-value 6.2e-79, Wilcoxon rank sum test). Concordant with this observation, analysis of putative Lin28a binding sites revealed that the target A/G-rich motifs tends to assume a single-stranded conformation within the loop regions of hairpin-like structures (Figure&#x000a0;<xref rid="Fig5" ref-type="fig">5</xref>b). This result is in agreement with previous <italic>in silico</italic> predictions based on the analysis of Watson-Crick pair co-occurrence around the Lin28a consensus [<xref ref-type="bibr" rid="CR61">61</xref>].<fig id="Fig5"><label>Figure 5</label><caption><p> <bold>CIRS-seq reveals structural preferences of RNA binding proteins. (a)</bold> Average reactivity across 300 nucleotides surrounding summits of Lin28a peaks. <bold>(b)</bold> Representation of sample secondary structures for Lin28a binding sites. Bases are color coded according to their reactivity. The purine-rich motifs are highlighted in green.</p></caption><graphic xlink:href="13059_2014_491_Fig5_HTML" id="MO5"/></fig></p></sec></sec><sec id="Sec7" sec-type="discussion"><title>Discussion</title><p>In this work we have defined, for the first time, the complete RNA structurome of mouse ESCs. Our analysis revealed the structural features of mRNAs and ncRNAs, and identified the structural requirements for Lin28a RNA binding protein.</p><p>The introduction of the CIRS-seq method, which does not rely on a denaturation and re-folding approach, allowed us to massively probe RNAs in their natural context. By applying CIRS-seq to mouse ESC RNAs, we were able to probe protein-coding RNAs as well as ncRNAs in their native deproteinized conformation. Analysis of previously validated secondary structures showed that CIRS-seq is extremely precise, and RNA secondary structures inferred using CIRS-seq data to constrain folding algorithms exhibit higher accuracy than MFE structures predicted in the absence of chemical probing data. Moreover, the use of two compounds that modify distinct bases, together with the introduction of a deproteinization step, which enabled us to investigate protein-masked regions of transcripts without losing their correct folding, increased the resolution of our method.</p><p>The analysis of CIRS-seq data revealed a strong structural partitioning of protein-coding transcripts, revealing a higher degree of structuring of UTRs compared with coding regions. This was unexpected since it has been recently reported that in <italic>A. thaliana</italic> UTRs have a higher propensity to single-strandedness than coding regions [<xref ref-type="bibr" rid="CR34">34</xref>]<italic>.</italic> This difference may represent evolutionary structural diversity between metazoans and plant RNAs, as suggested by previous <italic>in vitro</italic> [<xref ref-type="bibr" rid="CR17">17</xref>] and <italic>in silico</italic> [<xref ref-type="bibr" rid="CR62">62</xref>] analyses, or it could be explained by reduced accessibility of transcript coding regions to DMS treatment, due to the ribosome occupancy, in the absence of a deproteinization step. However, the agreement of our data with a recent nuclease-based analysis conducted in human lymphoblastoid cells [<xref ref-type="bibr" rid="CR63">63</xref>] suggests that this structuring is conserved in mammalian mRNAs and may have a functional role.</p><p>The slightly higher reactivity observed for 3&#x02032; UTRs compared with 5&#x02032; UTRs in mouse mRNAs may be representative of the preference of microRNA recognition elements, which are highly enriched in 3&#x02032; UTRs [<xref ref-type="bibr" rid="CR64">64</xref>], to reside within more accessible contexts [<xref ref-type="bibr" rid="CR65">65</xref>,<xref ref-type="bibr" rid="CR66">66</xref>]. It must be also noticed that structural regulatory elements in the 3&#x02032; UTR are often short and dispersed in the UTR, which in many cases may be very long, thus leading to a lower overall structuring of this region compared with the 5&#x02032; UTR [<xref ref-type="bibr" rid="CR67">67</xref>].</p><p>Our analysis of Lin28a protein recognition elements demonstrated genome-wide that binding sites for this protein tend to preferentially assume a single-stranded conformation. We moreover observed that Lin28a motifs tend to reside within loop regions of hairpin-like structures.</p><p>Furthermore, the analysis of ncRNAs revealed a higher overall degree of structuring compared with protein-coding transcripts, and showed that lincRNAs exhibit structural features intermediate to those of mRNAs and structural ncRNAs. This is in agreement with the report that ncRNAs have higher melting temperatures than mRNAs, denoting higher structural stability [<xref ref-type="bibr" rid="CR18">18</xref>].</p><p>Collectively, our data demonstrate that CIRS-seq can be used to obtain genome-wide information on native deproteinized RNA structures. Moreover, CIRS-seq methodology represents an important tool for the study of the structural binding specificities of RBPs.</p></sec><sec id="Sec8" sec-type="conclusion"><title>Conclusions</title><p>We define for the first time the complete RNA structurome of mouse ESCs, by developing a high-throughput method for the analysis of RNA secondary structures in their native deproteinized conformation. This method achieved extremely high accuracy on validated secondary structures, and allowed the <italic>de novo</italic> prediction of RNA structures. Analysis of structural data for protein-coding RNAs revealed their strong structural partitioning between 5&#x02032; UTRs, coding sequences, and 3&#x02032; UTRs. Comparison with non-coding RNAs showed that ncRNAs are more structured than mRNAs, and that lincRNAs present an average structuring midway between protein coding and structural non-coding transcripts. We also reveal the structural requirements for binding of the RBP Lin28a, and demonstrate that our method can provide insight into the structural preferences of RBPs.</p></sec><sec id="Sec9" sec-type="materials|methods"><title>Materials and methods</title><sec id="Sec10"><title>Cell culture</title><p>Mouse E14 ESCs were grown on 0.1% gelatin-coated plates and maintained in DMEM (4.5&#x000a0;g/L D-glucose) supplemented with 15% heat-inactivated fetal bovine serum, 0.1&#x000a0;mM NEAA, 1&#x000a0;mM sodium pyruvate, 0.1&#x000a0;mM 2-mercaptoethanol, 25 U/ml penicillin, 25&#x000a0;&#x003bc;g/ml streptomycin and 1,500 U/ml LIF, as previously described [<xref ref-type="bibr" rid="CR68">68</xref>].</p></sec><sec id="Sec11"><title>Quantitative RT-PCR</title><p>Real-time quantitative PCR was performed using the SuperScript III Platinum One-Step Quantitative RT- PCR System (Invitrogen Carlsbad, CA, USA) as previously described [<xref ref-type="bibr" rid="CR69">69</xref>]. The primers for the Rpph1 test transcript are provided in Table S1 in Additional file <xref rid="MOESM1" ref-type="media">1</xref>.</p></sec><sec id="Sec12"><title>RNA-seq library preparation</title><p>For RNA-seq library preparation, approximately 1&#x000a0;&#x003bc;g of TRIzol (Invitrogen) isolated total RNA from ESCs was subjected to ribosomal RNA depletion using the Ribo-Zero Gold Kit (Epicentre Madison, Winsconsin, USA). rRNA-depleted RNA was used as the input for the RNA-seq library preparation using the TruSeq RNA Sample Prep Kit (Illumina) following the manufacturer&#x02019;s instructions.</p></sec><sec id="Sec13"><title>CIRS-seq</title><p>Cell lysis and chemical probing, library preparation, and sequencing are detailed in the supplementary Materials and methods in Additional file <xref rid="MOESM1" ref-type="media">1</xref>.</p></sec><sec id="Sec14"><title>RNA quality assessment</title><p>RNA sample quality was assessed with the Agilent Bioanalyzer 2100. All of the samples had an RNA integrity number ranging from 9.9 to 10.</p></sec><sec id="Sec15"><title>Data analysis</title><p>CIRS-seq data analysis, normalization and background subtraction, and transcript analysis are detailed in the supplementary Materials and methods in Additional file <xref rid="MOESM1" ref-type="media">1</xref>.</p></sec><sec id="Sec16"><title>Data access</title><p>CIRS-seq and RNA-seq data have been deposited in the Gene Expression Omnibus (GEO) under accession number GSE54106. Additional datasets and the source code for the analysis tool are available at [<xref ref-type="bibr" rid="CR70">70</xref>].</p></sec></sec><sec sec-type="supplementary-material"><title>Additional file</title><sec id="Sec17"><supplementary-material content-type="local-data" id="MOESM1"><media xlink:href="13059_2014_491_MOESM1_ESM.pdf"><label>Additional file 1:</label><caption><p> <bold>PDF file containing supplementary Materials and methods, Figures S1 to S5), Table S1, and supplementary references.</bold> </p></caption></media></supplementary-material></sec></sec>

Acute diarrhea in adults consulting a general practitioner in France during winter: incidence, clinical characteristics, management and risk factors

Could not extract abstract

<contrib contrib-type="author" corresp="yes" deceased="no" equal-contrib="no"><name><surname>Arena</surname><given-names>Christophe</given-names></name><address><email>christophe.arena@upmc.fr</email></address><xref ref-type="aff" rid="Aff1"/><xref ref-type="aff" rid="Aff2"/><xref ref-type="aff" rid="Aff3"/></contrib><contrib contrib-type="author" deceased="no" equal-contrib="no"><name><surname>Amoros</surname><given-names>Jean Pierre</given-names></name><address><email>jean-pierre.amoros@ch-ajaccio.fr</email></address><xref ref-type="aff" rid="Aff1"/><xref ref-type="aff" rid="Aff2"/><xref ref-type="aff" rid="Aff3"/></contrib><contrib contrib-type="author" deceased="no" equal-contrib="no"><name><surname>Vaillant</surname><given-names>V&#x000e9;ronique</given-names></name><address><email>v.vaillant@invs.sante.fr</email></address><xref ref-type="aff" rid="Aff4"/></contrib><contrib contrib-type="author" deceased="no" equal-contrib="no"><name><surname>Ambert-Balay</surname><given-names>Katia</given-names></name><address><email>katia.balay@chu-dijon.fr</email></address><xref ref-type="aff" rid="Aff5"/></contrib><contrib contrib-type="author" deceased="no" equal-contrib="no"><name><surname>Chikhi-Brachet</surname><given-names>Roxane</given-names></name><address><email>roxane.brachet@anrs.fr</email></address><xref ref-type="aff" rid="Aff6"/></contrib><contrib contrib-type="author" deceased="no" equal-contrib="no"><name><surname>Jourdan-Da Silva</surname><given-names>Nathalie</given-names></name><address><email>n.jourdan-dasilva@invs.sante.fr</email></address><xref ref-type="aff" rid="Aff4"/></contrib><contrib contrib-type="author" deceased="no" equal-contrib="no"><name><surname>Varesi</surname><given-names>Laurent</given-names></name><address><email>varesi@univ-corse.fr</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author" deceased="no" equal-contrib="no"><name><surname>Arrighi</surname><given-names>Jean</given-names></name><address><email>orscorse@orange.fr</email></address><xref ref-type="aff" rid="Aff3"/></contrib><contrib contrib-type="author" deceased="no" equal-contrib="no"><name><surname>Souty</surname><given-names>C&#x000e9;cile</given-names></name><address><email>cecile.souty@upmc.fr</email></address><xref ref-type="aff" rid="Aff2"/><xref ref-type="aff" rid="Aff7"/></contrib><contrib contrib-type="author" deceased="no" equal-contrib="no"><name><surname>Blanchon</surname><given-names>Thierry</given-names></name><address><email>thierry.blanchon@upmc.fr</email></address><xref ref-type="aff" rid="Aff2"/><xref ref-type="aff" rid="Aff7"/></contrib><contrib contrib-type="author" deceased="no" equal-contrib="no"><name><surname>Falchi</surname><given-names>Alessandra</given-names></name><address><email>falchi@univ-corse.fr</email></address><xref ref-type="aff" rid="Aff1"/><xref ref-type="aff" rid="Aff2"/></contrib><contrib contrib-type="author" deceased="no" equal-contrib="no"><name><surname>Hanslik</surname><given-names>Thomas</given-names></name><address><email>thomas.hanslik@apr.aphp.fr</email></address><xref ref-type="aff" rid="Aff2"/><xref ref-type="aff" rid="Aff8"/><xref ref-type="aff" rid="Aff9"/></contrib><aff id="Aff1"><label/>Laboratory of Virology, EA7310, University of Corsica, Corte, France </aff><aff id="Aff2"><label/>INSERM, UMR_S 1136, Institut Pierre Louis d&#x02019;Epid&#x000e9;miologie et de Sant&#x000e9; Publique, Paris, F-75013 France </aff><aff id="Aff3"><label/>Regional Observatory of Health of Corsica, Ajaccio, France </aff><aff id="Aff4"><label/>Department of Infectious Diseases, Institut de Veille Sanitaire (InVS) (French Institute for Public Health Surveillance), Saint-Maurice, France </aff><aff id="Aff5"><label/>National Reference Center for Enteric Viruses, Laboratory of Virology, CHU of Dijon, Dijon, France </aff><aff id="Aff6"><label/>French National Agency for Research on AIDS and Viral Hepatitis, Paris, France </aff><aff id="Aff7"><label/>Sorbonne Universit&#x000e9;s, UPMC Univ Paris 06, UMRS 1136, Institut Pierre Louis d&#x02019;Epid&#x000e9;miologie et de Sant&#x000e9; Publique, Paris, F-75013 France </aff><aff id="Aff8"><label/>Internal Medicine Department, Assistance Publique-H&#x000f4;pitaux de Paris, Ambroise Par&#x000e9; Hospital, Boulogne Billancourt, F-92100 France </aff><aff id="Aff9"><label/>UFR des Sciences de la Sant&#x000e9; Simone-Veil, Universit&#x000e9; de Versailles - Saint-Quentin-en-Yvelines, Versailles, F-78280 France </aff>

BMC Infectious Diseases

<sec id="Sec1"><title>Background</title><p>In industrialized countries, acute diarrhea (AD) is a major cause of morbidity and medical expenses, particularly in vulnerable populations, such as elderly patients who are more often hospitalized, stay in the hospital longer and die more often than younger individuals when AD occurs [<xref ref-type="bibr" rid="CR1">1</xref>]. Infectious AD can be caused by various microbiological pathogens such as bacteria, parasites or viruses. AD occurs year-round but exhibits a pronounced winter peak, related to an increase in AD of a viral origin, mainly due to noroviruses and group A rotavirus infections [<xref ref-type="bibr" rid="CR2">2</xref>]-[<xref ref-type="bibr" rid="CR4">4</xref>].</p><p>Few studies have described the epidemiology and management of viral AD in adults during the winter. During the winter of 1998&#x02013;1999 in France, human caliciviruses were shown to be the most frequently encountered viruses in 16- to 65-year-old patients consulting a general practitioner (GP) for AD, while group A rotavirus predominated in patients 65 years of age and older [<xref ref-type="bibr" rid="CR2">2</xref>]. During the 1995&#x02013;1996 winter, risk factors shown to be associated with AD in France included contact with a person with AD, living with a child &#x02264;2 years of age, and recent treatment with oral penicillin or cephalosporin [<xref ref-type="bibr" rid="CR5">5</xref>]. However, in this study, microbiological investigations were not required, and the results were presented for all age groups and not specifically for adults. In the Netherlands, hand hygiene and contact with a sick person were identified as risk factors for viral gastroenteritis related to caliciviruses and group A rotavirus infections, but approximately 90% of the included patients were &#x0003c;10 years of age [<xref ref-type="bibr" rid="CR6">6</xref>]. The management of viral AD in general practice was studied for rotavirus infections in children [<xref ref-type="bibr" rid="CR7">7</xref>], but to our knowledge, such data are not available for viral AD occurring in adults.</p><p>Thus, data describing the epidemiology and management of viral AD in adults seen in general practice are scant. The objective of this study was to identify the clinical characteristics, management and risk factors associated with the occurrence of viral AD in French adults consulting a GP.</p></sec><sec id="Sec2"><title>Methods</title><sec id="Sec3"><title>Study design</title><p>In France, continuous surveillance of AD is conducted by the French <italic>Sentinelles</italic> GPs network (<ext-link ext-link-type="uri" xlink:href="http://www.sentiweb.fr">www.sentiweb.fr</ext-link>) [<xref ref-type="bibr" rid="CR8">8</xref>],[<xref ref-type="bibr" rid="CR9">9</xref>]. <italic>Sentinelles</italic> GPs&#x02019; characteristics, such as regional distribution, proportion in rural practice, type of practice and types of main clinical skills, are comparable to those of all French GPs [<xref ref-type="bibr" rid="CR10">10</xref>].</p><p>The study was conducted over two consecutive winters from the 49^th week of 2010 (2010w49) to 2011w17 and then from 2011w49 to 2012w17.</p><p>The <italic>Sentinelles</italic> GPs reported (via the Internet) information regarding all adult individuals (&#x02265;18 years old) presenting with AD, which was defined as &#x0201c;at least 3 daily watery (or nearly so) stools, less than 14 days&#x0201d;. The age and sex of the patients were documented.</p><p>A sample of <italic>Sentinelles</italic> GPs participated in a complementary survey with the aim of investigating clinical characteristics, virology, and management of AD occurring in adults. They were asked to recruit one AD case per week. To ensure that the selection of patients remained random, the GP had to include the first patient seen in consultation and who met the inclusion criteria in that particular week. Patients with inflammatory bowel disease and patients with an obvious non-viral etiology of diarrhea (traveler&#x02019;s diarrhea, recent use of antibiotics, colchicine, non-steroidal anti-inflammatory drugs or laxatives, or recent administration of chemotherapy or radiotherapy) were excluded.</p><p><italic>Sentinelles</italic> GPs were also asked to include one age- and sex-matched patient per AD case for a nested case&#x02013;control study. The study&#x02019;s aim was to identify the risk factors associated with the occurrence of viral AD. This matched individual presented just after the AD case for a non-gastrointestinal disease and did not report any gastrointestinal symptoms during the month preceding the consultation.</p><p>The GPs completed and sent a case report form for all patients included in the complementary survey by postal mail. The case report included collected data on gender, age and potential risk factors. The studied risk factors were factors related to lifestyle (professional status, educational level, presence in the household of children &#x02264;2 years of age, contact with pets or farm animals, hand hygiene, suffering from a chronic disease), and exposure during the last 7 days (contact with persons with AD in and/or outside the household; having eaten an unusual meal; consumption of tap water, oysters, mussels or shellfish; having used public transport; and/or having gone to a swimming pool). Data on reported symptoms, medications, days of missed work, additional medical examinations, or required hospitalizations were also collected for each AD case.</p><p>Patients included in the complementary survey were asked to collect and send stool specimens by postal mail in triple packaging (according to the United Nations class 6.2 specifications). They were also asked to return a follow-up questionnaire the week after enrollment to indicate the duration of symptoms (AD patients) and to ascertain whether an AD had occurred or not (non-AD patients).</p></sec><sec id="Sec4"><title>Virological analysis</title><p>All stool specimens were tested for four enteric viral pathogens (astrovirus, group A rotavirus, human enteric adenovirus, and norovirus of genogroup I - NoVGI - and genogroup II - NoVGII) using the Seeplex&#x000ae; Diarrhea-V ACE assay (Seegene) according to the manufacturer&#x02019;s instructions. A recent study showed that the Seeplex&#x000ae; Diarrhea-V assay is a sensitive, specific, convenient and reliable method to simultaneously detect several viral pathogens found directly in stool specimens from patients with gastroenteritis [<xref ref-type="bibr" rid="CR11">11</xref>].</p></sec><sec id="Sec5"><title>Statistical analysis</title><p>The AD cases reported via the Internet by the <italic>Sentinelles</italic> GPs allowed the estimation of winter incidence rates for mainland France by age group (18 &#x02013; 39 years, 40 &#x02013; 59 years, 60 &#x02013; 79 years and &#x02265;80 years). The winter incidence rate was calculated as follows: the average number of cases notified by <italic>Sentinelles</italic> GPs (adjusted for participation and geographic distribution) was multiplied by the total number of private GPs practicing in France and then divided by the French population [<xref ref-type="bibr" rid="CR12">12</xref>],[<xref ref-type="bibr" rid="CR13">13</xref>]. Confidence intervals were estimated by assuming that the distribution of the number of reported cases followed a Poisson distribution.</p><p>The data collected during the complementary survey were entered twice to ensure consistency. Data analysis was performed using STATA (version 11.2, StataCorp LP, Texas, USA). Quantitative variables were described by using medians [interquartile range IQ] and means &#x02013; standard deviations and were compared by the Wilcoxon test. Qualitative variables were described by using proportions and compared using a chi-square or Fisher&#x02019;s exact test if the chi-square test were not applicable; the results were presented as odds ratio with 95% confidence intervals (OR [95% CI]).</p><p>For the nested case&#x02013;control study, a <italic>case</italic> was a patient with AD in which at least one enteric virus was identified; a <italic>control</italic> was a matched patient without AD in which no enteric virus was identified. Univariate analyses were conducted using the McNemar test. A conditional logistic regression model was used to study the independent effects of risk factors that were associated in the univariate analyses (p-value of &#x0003c;0.20). Variables for the model were chosen through automatic backwards selection using a significance level of 0.05. Assuming a control-to-case ratio of 1:1, an exposure rate of 15% among controls, a two-tailed level of significance of 5% and a power level of 80%, 87 cases were needed to detect a minimal odds ratio (OR) of 3.</p></sec><sec id="Sec6"><title>Ethics statement</title><p>Oral consent was obtained from the patients at the time of inclusion for their participation in the study and for the publication of the clinical and virological data.</p><p>The Hospital Ethics Committee (CHU Saint-Antoine, Paris, France) approved the study.</p></sec></sec><sec id="Sec7"><title>Results</title><sec id="Sec8"><title>Incidence rates in general practice</title><p>During the two winters studied, 370 GPs participated in the electronic surveillance, and 10,415 AD cases were reported. Figure <xref rid="Fig1" ref-type="fig">1</xref> shows the weekly incidence rates, and Table <xref rid="Tab1" ref-type="table">1</xref> presents the winter incidences and incidence rates by age groups. The median age of adult patients seen by the <italic>Sentinelles</italic> GPs over the two consecutive winters was 37 years (IQ = [27 &#x02013; 52]) and 36 years (IQ = [27 &#x02013; 51]), respectively; the proportion of men was 46.2% and 45.4% over the two winters, respectively.<fig id="Fig1"><label>Figure 1</label><caption><p> <bold>Weekly incidence rates of acute diarrhea in adults (&#x02265;18 years old) consulting a GP in France (estimated using the French</bold> <bold><italic>Sentinelles</italic></bold> <bold>GPs network).</bold> </p></caption><graphic xlink:href="12879_2014_Article_574_Fig1_HTML" id="d30e622"/></fig></p><table-wrap id="Tab1"><label>Table 1</label><caption><p> <bold>Incidence rates of acute diarrhea in France by age group per 100,000 cases estimated by the French</bold> <bold><italic>Sentinelles</italic></bold> <bold>GPs network during two consecutive winters</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr><th rowspan="2"/><th colspan="2">Winter 2010/2011</th><th colspan="2">Winter 2011/2012</th></tr><tr><th colspan="2">2010w49 &#x02013; 2011w17</th><th colspan="2">2011w49 &#x02013; 2012w17</th></tr><tr><th/><th>Incidence [IC 95%]</th><th>Incidence rate per 100,000 [IC 95%]</th><th>Incidence [IC 95%]</th><th>Incidence rate per 100,000 [IC 95%]</th></tr></thead><tbody><tr><td>18 years of age and older</td><td>1,691,959 [1,287,372 &#x02013; 2,096,954]</td><td>3,471 [2,641 &#x02013; 4,302]</td><td>1,472,351 [1,060,343 &#x02013; 1,885,771]</td><td>3,002 [2,162 &#x02013; 3,845]</td></tr><tr><td> <italic>18 &#x02013; 39 years</italic> </td><td>953,943 [784,323 &#x02013; 1,123,563]</td><td>5,388 [4,430 &#x02013; 6,346]</td><td>859,608 [678,755 &#x02013; 1,040,461]</td><td>4,920 [3,885 &#x02013; 5,956]</td></tr><tr><td> <italic>40 &#x02013; 59 years</italic> </td><td>470,312 [351,703 &#x02013; 588,921]</td><td>2,771 [2,072 &#x02013; 3,470]</td><td>390,424 [273,691 &#x02013; 5 07,157]</td><td>2,290[1,606 &#x02013; 2,975]</td></tr><tr><td> <italic>60 &#x02013; 79 years</italic> </td><td>194,142 [122,101 &#x02013; 266,248]</td><td>1,761 [1,107 &#x02013; 2,415]</td><td>162,982 [90,142 &#x02013; 235,822]</td><td>1,420 [785 &#x02013; 2,055]</td></tr><tr><td> <italic>&#x02265; 80 years</italic> </td><td>73,562 [29,245 &#x02013; 118,222]</td><td>2,200 [875 &#x02013; 3,536]</td><td>59,337 [17,755 &#x02013; 102,331]</td><td>1,668 [499 &#x02013; 2,876]</td></tr></tbody></table></table-wrap></sec><sec id="Sec9"><title>Clinical characteristics, management and virology</title><p>Among the 100 <italic>Sentinelles</italic> GPs who agreed to participate in the complementary survey, 65 enrolled 192 adult patients who were seen for AD. Their median age was 36 years (IQ = [28 &#x02013; 52]), and 111 (57.8%) were men. The reported clinical signs are presented in Table <xref rid="Tab2" ref-type="table">2</xref>.<table-wrap id="Tab2"><label>Table 2</label><caption><p> <bold>Reported clinical signs in adult patients consulting a GP for acute diarrhea (complementary survey)</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr><th/><th>Patients with acute diarrhea (N = 192) (%)</th></tr></thead><tbody><tr><td>Average time before consultation &#x000b1; sd (days)</td><td>1.6 &#x000b1; 1.8</td></tr><tr><td>Average duration of diarrhea &#x000b1; sd (days)</td><td>2.0 &#x000b1; 1.8</td></tr><tr><td>Average number of stools in the last 24 h &#x000b1; sd</td><td>5.7 &#x000b1; 2.8</td></tr><tr><td>Average max. number of stools per day &#x000b1; sd</td><td>6.0 &#x000b1; 2.9</td></tr><tr><td>Mucous diarrhea</td><td>29 (15.1%)</td></tr><tr><td>Bloody diarrhea</td><td>2 (1.0%)</td></tr><tr><td>Watery diarrhea</td><td>170 (88.5%)</td></tr><tr><td>Abdominal pain</td><td>175 (91.1%)</td></tr><tr><td>Nausea</td><td>160 (83.3%)</td></tr><tr><td>Vomiting</td><td>119 (62.0%)</td></tr><tr><td> <italic>Average duration</italic> &#x000b1; <italic>sd (days)</italic> </td><td> <italic>1.0 &#x000b1; 1.1</italic> </td></tr><tr><td>Fever</td><td>83 (43.2%)</td></tr><tr><td> <italic>Average body temperature</italic> &#x000b1; <italic>sd (&#x000b0;C)</italic> </td><td> <italic>38.4 &#x000b1; 0.5</italic> </td></tr><tr><td> <italic>Average duration</italic> &#x000b1; <italic>sd (days)</italic> </td><td> <italic>1.4 &#x000b1; 1.3</italic> </td></tr><tr><td>Dehydration</td><td>8 (4.2%)</td></tr><tr><td>Other symptoms</td><td>8 (4.2%)</td></tr></tbody></table></table-wrap></p><p>Overall, 183 (95.3%) patients received a drugs prescription, which were mostly intestinal antisecretory drugs (N = 98, 53.6%), antiemetics (N = 96, 52.4%), antispasmodics (N = 72, 39.3%), intestinal adsorbents (N = 65, 35.5%), analgesics/antipyretics (N = 54, 29.5%) and regulators of intestinal motility (N = 54, 29.5%). Among the 146 patients who reported having a job, 116 (79.5%) benefited from stopping work for a median duration of 3 days (IQ = [2 &#x02013; 3]), regardless of gender.</p><p>Stool samples from 145 patients with AD (75.5%) were returned. The median age of those patients was 37.5 years (IQ = [30 &#x02013; 54]), and 80 were men (55.2%). Stools tested positive for at least one of the four enteric viruses investigated in 94 cases (65%). The detailed results from the virological investigation are presented in Table <xref rid="Tab3" ref-type="table">3</xref>. Among the patients with AD, the reported clinical signs did not differ between adults with a virus in the stool sample and those with no virus found in the stool exam, neither in frequency nor in severity (Table <xref rid="Tab4" ref-type="table">4</xref>). Thus, the management of patients with AD who tested positive for a virus was not different from the management of patients who tested negative (data not shown). None of the cases required hospitalization.<table-wrap id="Tab3"><label>Table 3</label><caption><p> <bold>Results from the virological investigation of adult patients consulting a general practitioner for acute diarrhea in France from week 2010w49 to week 2011w17 and from week 2011w49 to week 2012w17 (complementary survey)</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr><th>Viruses detected</th><th>Patients with acute diarrhea (N = 145) (%)</th></tr></thead><tbody><tr><td>Norovirus GII</td><td>59 (40.7)</td></tr><tr><td>Norovirus GI</td><td>17 (11.7)</td></tr><tr><td>Astrovirus</td><td>5 (3.5)</td></tr><tr><td>Rotavirus</td><td>2 (1.4)</td></tr><tr><td>Adenovirus 40/41</td><td>0 (0.0)</td></tr><tr><td>Coinfections</td><td>11 (8.0)</td></tr><tr><td> <italic>Norovirus GI + GII</italic> </td><td> <italic>5 (3.5)</italic> </td></tr><tr><td> <italic>Norovirus GI + Rotavirus A</italic> </td><td> <italic>1 (0.7)</italic> </td></tr><tr><td> <italic>Norovirus GI + Astrovirus</italic> </td><td> <italic>1 (0.7)</italic> </td></tr><tr><td> <italic>Norovirus GII + Astrovirus</italic> </td><td> <italic>4 (2.8)</italic> </td></tr><tr><td>At least one virus detected</td><td>94 (64.8)</td></tr><tr><td>No virus detected</td><td>51 (35.2)</td></tr></tbody></table></table-wrap><table-wrap id="Tab4"><label>Table 4</label><caption><p> <bold>Reported clinical signs in adult patients consulting a GP for acute diarrhea in virus-positive and virus-negative stool samples (complementary survey)</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr><th/><th colspan="3">Patients with acute diarrhea</th></tr><tr><th/><th>At least one virus detected (N = 94) (%)</th><th>No virus detected (N = 51) (%)</th><th>p-value*</th></tr></thead><tbody><tr><td>Average age &#x000b1; sd (years)<bold>**</bold> </td><td>40.4 &#x000b1; 15.1</td><td>44.3 &#x000b1; 17.8</td><td>0.16</td></tr><tr><td>Men<bold>**</bold> </td><td>52 (55.9%)</td><td>28 (57.1%)</td><td>0.89</td></tr><tr><td>Average time before consultation &#x000b1; sd (days)</td><td>1.6 &#x000b1; 1.9</td><td>1.7 &#x000b1; 1.9</td><td>0.92</td></tr><tr><td>Average duration of diarrhea &#x000b1; sd (days)</td><td>1.6 &#x000b1; 1.5</td><td>2.1 &#x000b1; 1.8</td><td>0.23</td></tr><tr><td>Average number of stools in the last 24 h &#x000b1; sd</td><td>5.4 &#x000b1; 2.7</td><td>6.3 &#x000b1; 3.3</td><td>0.09</td></tr><tr><td>Average max. number of stools per day &#x000b1; sd</td><td>5.7 &#x000b1; 2.5</td><td>6.5 &#x000b1; 3.3</td><td>0.13</td></tr><tr><td>Mucous diarrhea</td><td>10 (11.0%)</td><td>8 (17.4%)</td><td>0.30</td></tr><tr><td>Bloody diarrhea</td><td>1 (1.1%)</td><td>1 (2.2%)</td><td>0.63</td></tr><tr><td>Watery diarrhea</td><td>83 (91.2%)</td><td>39 (84.8%)</td><td>0.26</td></tr><tr><td>Abdominal pain</td><td>85 (93.4%)</td><td>42 (91.3%)</td><td>0.66</td></tr><tr><td>Nausea</td><td>77 (82.8%)</td><td>36 (78.3%)</td><td>0.52</td></tr><tr><td>Vomiting</td><td>61 (66.3%)</td><td>24 (52.2%)</td><td>0.11</td></tr><tr><td> <italic>Average duration</italic> &#x000b1; <italic>sd (days)</italic> </td><td> <italic>0.8 &#x000b1; 0.9</italic> </td><td> <italic>1.3 &#x000b1; 1.4</italic> </td><td> <italic>0.12</italic> </td></tr><tr><td>Fever</td><td>42 (46.2%)</td><td>15 (33.3%)</td><td>0.16</td></tr><tr><td>&#x02003;<italic>Average body temperature</italic> &#x000b1; <italic>sd (&#x000b0;C)</italic> </td><td> <italic>38.3 &#x000b1; 0.4</italic> </td><td> <italic>38.5 &#x000b1; 0.7</italic> </td><td> <italic>0.42</italic> </td></tr><tr><td>&#x02003;<italic>Average duration</italic> &#x000b1; <italic>sd (days)</italic> </td><td> <italic>1.2 &#x000b1; 1.2</italic> </td><td> <italic>1.7 &#x000b1; 1.7</italic> </td><td> <italic>0.30</italic> </td></tr><tr><td>Dehydration</td><td>3 (3.3%)</td><td>1 (2.2%)</td><td>0.71</td></tr><tr><td>Other symptoms</td><td>5 (5.5%)</td><td>1 (2.2%)</td><td>0.39</td></tr></tbody></table><table-wrap-foot><p>*Logistic regression: adjustment for age and sex.</p><p>**Not adjusted for age and sex.</p></table-wrap-foot></table-wrap></p></sec><sec id="Sec10"><title>Risk factors for viral AD</title><p>The GPs enrolled 101 matched individuals for the nested case&#x02013;control study. Among them, 95 patients mailed back a stool specimen. Of the stools examined, 4 tested positive (4.2%) for one enteric virus (NoVGII) and were excluded from the case&#x02013;control study. Thus, 91 pairs (51 male and 40 female) were included in the analysis. The median age was 36 years (IQ = [28 &#x02013; 50]) for the cases and 37 years (IQ = [29 &#x02013; 53]) for the controls. Viral acute diarrheas were independently associated with having been in contact with a person who has suffered from an AD in the last 7 days, either within or outside the household, and having a job or student (Table <xref rid="Tab5" ref-type="table">5</xref>). The contact of cases with sick people outside the household had taken place either at work (59%) or other place (41%). The median duration between the contact with a sick person and the onset of the symptoms was 2 days (IQ = [1 &#x02013; 4]).<table-wrap id="Tab5"><label>Table 5</label><caption><p> <bold>Factors associated with viral acute diarrhea (cases) in 91 pairs of adult patients consulting a GP</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr><th/><th>Cases (N = 91) (%)</th><th>Controls (N = 91) (%)</th><th>OR uni [95% CI] (p-value)*</th><th>OR multi [95% CI] (p-value)*</th></tr></thead><tbody><tr><td>Professional status (employed or student/non employed or retired)</td><td>80 (87.9%)</td><td>67 (73.6%)</td><td>4.25 [1.43 &#x02013; 12.63] (0.01)</td><td>4.10 [1.27 &#x02013; 13.21] (0.02)</td></tr><tr><td>Educational level (high school and above/middle school)</td><td>80 (87.9%)</td><td>69 (75.8%)</td><td>2.83 [1.12 &#x02013; 7.19] (0.03)</td><td>2.37 [0.86 &#x02013; 6.57] (0.10)</td></tr><tr><td>Children &#x02264;2 years in household (yes/no)</td><td>20 (22.0%)</td><td>9 (9.9%)</td><td>2.57 [1.07 &#x02013; 6.16] (0.03)</td><td>1.87 [0.69 &#x02013; 5.09] (0.22)</td></tr><tr><td>Being in contact with pets or farm animals (yes/no)</td><td>45 (49.5%)</td><td>48 (52.8%)</td><td>0.80 [0.48 &#x02013; 1.58] (0.65)</td><td>n.i.</td></tr><tr><td>Washing hands before cooking (never-sometimes/often-always)</td><td>10 (11.6%)</td><td>13 (16.1%)</td><td>0.58 [0.23 &#x02013; 1.48] (0.26)</td><td>n.i.</td></tr><tr><td>Washing hands after using the toilet (never-sometimes/often-always)</td><td>7 (7.7%)</td><td>8 (8.8%)</td><td>0.86 [0.29 &#x02013; 2.55] (0.78)</td><td>n.i.</td></tr><tr><td>Washing hands after attending public places (never-sometimes/often-always)</td><td>45 (52.3%)</td><td>35 (40.7%)</td><td>1.50 [0.76 &#x02013; 2.95] (0.24)</td><td>n.i.</td></tr><tr><td>Suffering from a chronic disease (yes/no)</td><td>28 (30.8%)</td><td>29 (31.9%)</td><td>0.94 [0.48 &#x02013; 1.86] (0.86)</td><td>n.i.</td></tr><tr><td>Contact with persons with AD in the household (yes/no)</td><td>31 (34.1%)</td><td>10 (11.0%)</td><td>5.20 [2.00 &#x02013; 13.50] (0.01)</td><td>4.18 [1.54 &#x02013; 11.33] (&#x0003c;0.01)</td></tr><tr><td>Contact with persons with AD outside household (yes/no)</td><td>22 (24.2%)</td><td>9 (9.9%)</td><td>3.60 [1.34 &#x02013; 9.70] (0.01)</td><td>3.31 [1.03 &#x02013; 10.63] (0.04)</td></tr><tr><td>Having eaten an unusual meal (yes/no)</td><td>33 (36.3%)</td><td>28 (30.8%)</td><td>1.39 [0.68 &#x02013; 2.83] (0.37)</td><td>n.i.</td></tr><tr><td>Having consumed oysters, mussels, or shellfish (yes/no)</td><td>27 (29.7%)</td><td>29 (31.9%)</td><td>0.90 [0.48 &#x02013; 1.70] (0.75)</td><td>n.i.</td></tr><tr><td>Having consumed tap water (yes/no)</td><td>69 (75.8%)</td><td>73 (80.2)</td><td>0.76 [0.37 &#x02013; 1.58] (0.47)</td><td>n.i.</td></tr><tr><td>Having used public transportation (yes/no)</td><td>20 (22.0%)</td><td>14 (15.4%)</td><td>2.00 [0.75 &#x02013; 5.33] (0.17)</td><td>2.57 [0.71 &#x02013; 9.39] (0.15)</td></tr><tr><td>Going to a public swimming pool (yes/no)</td><td>4 (4.4%)</td><td>5 (5.5%)</td><td>0.80 [0.22 &#x02013; 2.98] (0.74)</td><td>n.i.</td></tr></tbody></table><table-wrap-foot><p>*Conditional logistic regression: matched for age and sex.</p><p>OR: odds-ratio; Uni: univariate; Multi: multivariate; CI: confidence interval; n.i: not included in the multivariate model.</p></table-wrap-foot></table-wrap></p></sec></sec><sec id="Sec11"><title>Discussion</title><p>This study presents the first analysis of the global burden of AD in adults who consulted a GP in France. Winter incidences, clinical characteristics, virological investigation, management and risk factors for viral AD were investigated.</p><sec id="Sec12"><title>Incidence rates in general practice</title><p>During the two studied winters, 3,471 and 3,002 cases per 100,000 French adults consulted a GP for an AD in winters 2010/2011 and 2011/2012, respectively. The data on AD incidences vary from country to country because of differences in case definition, surveillance systems, and/or the period of study. In France, a telephone survey estimated the incidence rate of acute gastroenteritis at 0.33 cases/person-year [<xref ref-type="bibr" rid="CR14">14</xref>]. In the Netherlands, a population-based study conducted in 1998/1999 estimated that the gastroenteritis incidence was 283 per 1,000 person-years [<xref ref-type="bibr" rid="CR15">15</xref>]. In both studies, the incidence rate peaked in children and then decreased in adults.</p></sec><sec id="Sec13"><title>Clinical characteristics, management and virology</title><p>In this study, more than 80.0% of patients reported abdominal pain, watery diarrhea, and/or nausea, while vomiting and fever were reported by 62.0% and 43% of patients, respectively. These results are in agreement with other French studies [<xref ref-type="bibr" rid="CR2">2</xref>],[<xref ref-type="bibr" rid="CR14">14</xref>].</p><p>Adults are less likely to consult a GP for gastroenteritis compared with children, as it remains a self-limiting disease [<xref ref-type="bibr" rid="CR14">14</xref>]. Patients with more severe symptoms are more prone to consulting a GP, which is illustrated by the fact that 80% of working adult cases had to stop working. Although no cases required hospitalization, the economic burden of AD related to outpatient visits could be significant, because the average annual incidence of AD in adults is 1 million cases (<ext-link ext-link-type="uri" xlink:href="http://www.sentiweb.fr">www.sentiweb.fr</ext-link>). In addition to the cost of outpatient visits, medical treatment and missed work days increase the heavy burden of viral AD cost in adults. Indeed, 95% of the patients in this study received a drug prescription. The management of AD is most likely amenable to a more appropriate drug prescription in France. For example, antiemetics are prescribed in a majority of cases, whereas their efficacy in this indication has never been validated, and their side effects may be serious [<xref ref-type="bibr" rid="CR16">16</xref>].</p><p>The feces samples were not screened to rule out bacterial and parasitic infections. However, we included patients in whom there was a very high suspicion of viral diarrhea (and a very low risk of bacterial or parasitic infection), as inclusions were done during winter and cases with an obvious non-viral etiology of diarrhea were excluded. During the winter, viral AD is predominant, but the reason is not clear. Hypotheses for these findings include that the clustering of people indoors during the winter months facilitates person-to-person transmission and the enhanced persistence of noroviruses at low temperatures [<xref ref-type="bibr" rid="CR17">17</xref>]. Noroviruses have been described as the leading cause of winter AD [<xref ref-type="bibr" rid="CR18">18</xref>], and the GII genogroup strains have been previously shown to predominate during winter, although the reason for this remains unclear [<xref ref-type="bibr" rid="CR19">19</xref>]. In this study, the proportion of adult patients with AD who were positive for at least one enteric virus (65%) was higher than in previous studies in general practice performed in France or Europe (15-39%) [<xref ref-type="bibr" rid="CR2">2</xref>]-[<xref ref-type="bibr" rid="CR4">4</xref>]. However, unlike these studies, the aim of our study was to generate a sample of patients who were positive for a virus; thus, patients with obvious non-viral diarrhea were excluded. In the patients included here, the clinical characteristics of AD, and thus its management, were not different for adults with or without an identified virus in the stool. It is possible that a study with more statistical power would have identified some clinical differences, such as more frequent occurrence of vomiting [<xref ref-type="bibr" rid="CR20">20</xref>].</p></sec><sec id="Sec14"><title>Risk factors for viral AD</title><p>Being previously in contact with an individual presenting with AD was identified as a risk factor for developing AD. Norovirus and rotavirus are among the most communicable pathogens responsible for AD. Experimentally, an inoculum as low as 500 (and even less) viable organisms is sufficient to establish an infection, and the virus is environmentally stable [<xref ref-type="bibr" rid="CR20">20</xref>]. Thus, enteric viruses have a high potential for person-to-person spread. The increased risk in people who have had a contact with a sick person in the household is consistent with this already well-known mode of transmission [<xref ref-type="bibr" rid="CR21">21</xref>]. In 1995 and 1996 in France, Lettrillard et al. [<xref ref-type="bibr" rid="CR5">5</xref>] showed that the risk of developing AD was 5 times higher in patients who had been in contact with a person suffering from AD in their household. However, the study included patients whose AD etiology was unknown (no stool sample). Studies in Germany [<xref ref-type="bibr" rid="CR4">4</xref>] and the Netherlands [<xref ref-type="bibr" rid="CR6">6</xref>] have confirmed this observation and estimated adjusted ORs ranging from 1.9 to 12.9 based on viral detection; however, the results of these studies were not stratified by age. In this study, patients who reported having a job and students were significantly more likely to suffer from viral AD than those who were unemployed or retired. Among the studies that have tried to identify the risk factors for acquiring an AD, none have investigated professional status. The result obtained in this study seems quite relevant and suggests that this population has increased contact with sick people, which is the main risk factor for infection. The acquisition of a viral AD may be associated with other factors that were not identified in this study. For example, De Wit et al. showed that norovirus AD risk was increased in people with poorer hand hygiene (OR = 1.3 [1.0 &#x02013; 1.7]) [<xref ref-type="bibr" rid="CR6">6</xref>], which was not observed here. It has also been shown that living with children &#x02264;2 years of age increases the risk of developing AD in the winter, regardless of the children&#x02019;s health status (AD or not) [<xref ref-type="bibr" rid="CR5">5</xref>]. The association between developing AD and living with children &#x02264;2 years that was identified in our univariate analysis did not persist after adjusting for other variables. No association between viral AD and tap water use, seafood consumption or an unusual meal was found.</p></sec></sec><sec id="Sec15"><title>Conclusions</title><p>During the winter, AD of viral origin is a frequent disease in adults with a significant burden in the population. Noroviruses are mainly responsible for the disease. Other than contact with a person suffering from AD, no other preventable risk factor was identified. Thus, at the present time, education related to hand hygiene remains the only way to reduce the burden of disease.</p></sec><sec id="Sec16"><title>Authors&#x02019; contributions</title><p>CA, JPA, VV, KAB, RCB, NJDS, LV, JA, TB, AF and TH co-conceived the study. CA and TH collected the epidemiological and microbiological data. AF, LV and KAB designed the microbiology experiments. AF performed the microbiology experiments and analyzed and interpreted the data. CA and CS analyzed and interpreted the statistical data. CA, JPA, VV, KAB, RCB, NJDS, LV, JA, CS, TB, AF and TH contributed to writing the paper and approved the final manuscript.</p></sec>

Factors associated with healthcare avoidance among transgender women in Argentina

Could not extract abstract

<contrib contrib-type="author" corresp="yes"><name><surname>Soc&#x000ed;as</surname><given-names>Mar&#x000ed;a Eugenia</given-names></name><address><email>eugenia.socias@huesped.org.ar</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author"><name><surname>Marshall</surname><given-names>Brandon DL</given-names></name><address><email>brandon_marshall@brown.edu</email></address><xref ref-type="aff" rid="Aff2"/></contrib><contrib contrib-type="author"><name><surname>Ar&#x000ed;stegui</surname><given-names>In&#x000e9;s</given-names></name><address><email>ines.aristegui@huesped.org.ar</email></address><xref ref-type="aff" rid="Aff1"/><xref ref-type="aff" rid="Aff3"/></contrib><contrib contrib-type="author"><name><surname>Romero</surname><given-names>Marcela</given-names></name><address><email>marcelaromero_40@yahoo.com.ar</email></address><xref ref-type="aff" rid="Aff4"/></contrib><contrib contrib-type="author"><name><surname>Cahn</surname><given-names>Pedro</given-names></name><address><email>pcahn@huesped.org.ar</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author"><name><surname>Kerr</surname><given-names>Thomas</given-names></name><address><email>drtk@cfenet.ubc.ca</email></address><xref ref-type="aff" rid="Aff5"/><xref ref-type="aff" rid="Aff6"/></contrib><contrib contrib-type="author"><name><surname>Sued</surname><given-names>Omar</given-names></name><address><email>omar.sued@huesped.org.ar</email></address><xref ref-type="aff" rid="Aff1"/></contrib><aff id="Aff1"><label/>Fundaci&#x000f3;n Hu&#x000e9;sped, Angel Peluffo 3932, Buenos Aires, C1202ABB Argentina </aff><aff id="Aff2"><label/>Department of Epidemiology, Brown University School of Public Health, 121 South Main Street, Box G-S-121-2, Providence, RI 02912 USA </aff><aff id="Aff3"><label/>Center for Psychology Research, School of Social Sciences, University of Palermo, Mario Bravo 1259, Buenos Aires, C1175ABW Argentina </aff><aff id="Aff4"><label/>Association of Transvestites, Transsexuals, and Transgenders of Argentina (A.T.T.T.A.), Callao 339 6th floor, Buenos Aires, C1022AAD Argentina </aff><aff id="Aff5"><label/>British Columbia Centre for Excellence in HIV/AIDS, St. Paul&#x02019;s Hospital, 608-1081 Burrard Street, Vancouver, BC V6Z 1Y6 Canada </aff><aff id="Aff6"><label/>Department of Medicine, University of British Columbia, St. Paul&#x02019;s Hospital, 608-1081 Burrard Street, Vancouver, BC V6Z 1Y6 Canada </aff>

International Journal for Equity in Health

<sec id="Sec1" sec-type="introduction"><title>Introduction</title><p>Transgender (TG) women, the term for individuals who were assigned male sex at birth, but assume a feminine gender expression or identity, often experience multiple forms of oppression for transgressing gender norms, such as stigma, discrimination, isolation and economic hardship [<xref ref-type="bibr" rid="CR1">1</xref>,<xref ref-type="bibr" rid="CR2">2</xref>]. This marginalization, that is consequence of family, social and institutional transphobia, contributes to an increased risk of mental health problems, substance use and sexually transmitted infections. Collectively, these problems reinforce and perpetuate a harmful cycle of marginalization and poor health outcomes within this population [<xref ref-type="bibr" rid="CR2">2</xref>-<xref ref-type="bibr" rid="CR5">5</xref>].</p><p>Further, the widespread discrimination that TG women face, expose them to different forms of violence, which ranges from harassment, bullying, verbal abuse, physical violence to sexual assault and hate crimes [<xref ref-type="bibr" rid="CR6">6</xref>-<xref ref-type="bibr" rid="CR9">9</xref>]. In addition, due to the marginalization that many TG women experience, they interact frequently with police. Arbitrary arrest and detentions, which are common among this population, have been reported as an excuse to exploit TG women for bribes or coerce them into providing sex in exchange for release from detention [<xref ref-type="bibr" rid="CR7">7</xref>,<xref ref-type="bibr" rid="CR9">9</xref>-<xref ref-type="bibr" rid="CR11">11</xref>].</p><p>TG women also face socio-structural inequalities within the health care system. Studies consistently show that TG individuals experience multiple challenges when attempting to access both routine and transition-related medical care, including denial of care, harassment, and lack of competent and sensitive providers with adequate knowledge of their specific needs [<xref ref-type="bibr" rid="CR1">1</xref>,<xref ref-type="bibr" rid="CR10">10</xref>,<xref ref-type="bibr" rid="CR12">12</xref>-<xref ref-type="bibr" rid="CR15">15</xref>]. As a result of these barriers, many TG women postpone or avoid urgent and medical care altogether. This, in turn, may help explain the poor health outcomes observed among this population, such as increased risk for HIV infection, substance use, and suicide attempts [<xref ref-type="bibr" rid="CR3">3</xref>,<xref ref-type="bibr" rid="CR10">10</xref>,<xref ref-type="bibr" rid="CR15">15</xref>-<xref ref-type="bibr" rid="CR18">18</xref>]. Indeed, despite the high burden of HIV infection among this population [<xref ref-type="bibr" rid="CR19">19</xref>], studies show low rates of HIV testing [<xref ref-type="bibr" rid="CR20">20</xref>], as well as poor outcomes at each step of the HIV continuum of care [<xref ref-type="bibr" rid="CR21">21</xref>-<xref ref-type="bibr" rid="CR23">23</xref>]. Furthermore, due to barriers to transition-related medical care, use of non-prescribed hormones or injection of industrial silicone in non-sterilized environments is widespread among TG communities, posing additional risks for their health [<xref ref-type="bibr" rid="CR24">24</xref>,<xref ref-type="bibr" rid="CR25">25</xref>].</p><p>Argentina has a universal but mixed health system that is composed by 3 sub-systems: (a) the public, (b) the social security which provides health coverage for formally employed workers and their families, and (c) the private sector. The Ministry of Health oversees all three subsectors. The network of public hospitals and health centers is open to anyone (including foreigners) and nominally free of charge. However, the public sector mostly provides care to low-income individuals without other forms of health insurance coverage, approximately 30% of the population [<xref ref-type="bibr" rid="CR26">26</xref>,<xref ref-type="bibr" rid="CR27">27</xref>]. Concerns persist regarding various social and structural factors that may constrain access to health services among TG individuals in this setting, thus significantly impacting their health. For example, life expectancy of TG women is approximately 35&#x000a0;years (compared to 79&#x000a0;years in other women) [<xref ref-type="bibr" rid="CR28">28</xref>]. Furthermore, HIV prevalence among TG women in Argentina is estimated to be 34.1%, compared to 0.4% in the general population [<xref ref-type="bibr" rid="CR29">29</xref>,<xref ref-type="bibr" rid="CR30">30</xref>]. Other infectious diseases, such as syphilis, tuberculosis and viral hepatitis, are also more frequent among TG women [<xref ref-type="bibr" rid="CR30">30</xref>,<xref ref-type="bibr" rid="CR31">31</xref>], contributing to higher morbidity and mortality in this population. In response to this, and other related public health and social challenges, Argentina passed a progressive &#x0201c;Gender Identity Law&#x0201d; in May 2012 [<xref ref-type="bibr" rid="CR32">32</xref>]. This law acknowledges the right to self-defined gender identity, allowing for changes to gender, image, or birth name on one&#x02019;s identity card (ID), and ensures the right to appropriate transgender health services.</p><p>Despite these ongoing challenges and novel policy developments, little is known about the specific barriers to healthcare among TG women in Argentina, including those factors that may lead some TG women to avoid healthcare altogether. Therefore, the aim of this study was to explore potential individual, socio-structural and environmental factors associated with healthcare avoidance among TG women in Argentina.</p></sec><sec id="Sec2" sec-type="materials|methods"><title>Methods</title><p>Data for the present analysis were derived from a nation-wide, semi-structured, cross-sectional survey, conducted by Fundaci&#x000f3;n Hu&#x000e9;sped and A.T.T.T.A. (Association of Transvestites, Transsexuals, and Transgenders of Argentina), between June 2013 and December 2013. The objective of this survey was to collect baseline data and validate an instrument designed to assess the impact that the Gender Identity Law has had on living conditions among transgender individuals in Argentina.</p><p>In order to maximize representativeness of the transgender population in Argentina, snowball sampling was combined with quota sampling. As there are no official estimates of the size of the TG population in Argentina, the quotas (by region, by age group, and by educational level) were calculated using statistics provided by the national registration office (Registro Nacional de las Personas, RENAPER); specifically, the number of new IDs issued since the implementation of the &#x0201c;Gender Identity Law&#x0201d; and other national reports of socio-demographic characteristics of transgender individuals in Argentina. The goal was to recruit 450 TG women and 50 TG men.</p><p>Recruitment was done by peer outreach efforts facilitated by A.T.T.T.A. Recruitment venues included sex work areas and community-based organizations, among other places known to be frequented by TG individuals. Self-identified transgender individuals were eligible to participate. After providing written informed consent, enrolled individuals completed an interviewer-administered questionnaire. The surveys were conducted in private rooms by transgender peers previously trained in interviewer-administered assessment methods. The study was approved by the institutional ethics committee of Fundaci&#x000f3;n Hu&#x000e9;sped. Participation was voluntary, and upon completing participants received a $100 ARS reimbursement (approximately $10 USD) for their time and effort. All data were de-identified to maximize confidentiality.</p><p>The interviewer-administered questionnaire captured information regarding socio-demographic characteristics, gender transition, HIV status, violence, interactions with police, healthcare access, housing, education, work, and experiences of stigma and discrimination in these settings.</p><p>For the current study, only TG women were included, and the primary outcome of interest was avoidance of healthcare due to transgender identity, defined as answering &#x0201c;Yes&#x0201d; to the following question: &#x0201c;Have you ever avoided going to a hospital or clinic because of your transgender identity?&#x0201d; With the selection of this dependent variable, we sought to measure and characterize avoidance of healthcare among TG women within Argentina&#x02019;s universal healthcare system, which allows for analyses that are free from the confounding effects of affordability of health services.</p><p>In accordance with the Risk Environment Framework [<xref ref-type="bibr" rid="CR33">33</xref>,<xref ref-type="bibr" rid="CR34">34</xref>], we have sought to identify a range of individual, social-structural and environmental correlates of healthcare avoidance. Specifically, the following factors were explored:<list list-type="alpha-lower"><list-item><p><italic>Individual level factors:</italic> age (less than versus greater than the median age), place of birth (Argentina versus other), having a job other than sex work (yes versus no), history of sex work involvement (yes vs. no), extended health insurance, defined as having either social security or private health coverage in addition to the universal public health coverage (yes versus no), high school education or higher (yes versus no), self-reported HIV infection status (yes, no, or unknown), and internalized stigma, defined as responding &#x0201c;yes&#x0201d; to any of the following questions: &#x0201c;have you ever felt any of the following emotions because of your transgender identity: ashamed, guilty, low self-esteem, feel that you should be punished&#x0201d;). Cronbach&#x02019;s alpha for this stigma measure was 0.61.</p></list-item><list-item><p><italic>Social-structural factors:</italic> we asked participants to report if they had ever been arrested (yes versus no), and/or ever experienced violence from the police (defined as answering &#x0201c;yes&#x0201d; to any of the following questions: &#x0201c;Have you ever been exposed to any of the following situations due to your transgender identity: a policeman threatened you, a policeman beat, kicked or physically hurt you, a policeman forced you to have sex against your will?&#x0201d;). Cronbach&#x02019;s alpha for this measure was 0.82. We also examined whether participants reported ever experiencing any perceived discrimination due to their transgender identity by either healthcare workers (including physicians, psychologists, social workers, and other administrative staff), or by patients (yes versus no). For the latter questions, and to provide contextualization, participants were provided with examples of specific situations of discrimination within the healthcare setting (e.g., denied health services, not called by their preferred name, mockery or threats).</p></list-item><list-item><p><italic>Environmental factors:</italic> residency (Buenos Aires metropolitan area, the biggest urban center in Argentina, vs. other), and current housing status (stable versus others).</p></list-item></list></p><p>Chi-square (for categorical variables) and Mann&#x02013;Whitney (for continuous non-normally distributed variables) tests were conducted to compare TG women who reported avoiding or not healthcare. Bi- and multivariable logistic regression analyses were performed to examine potential associations between each independent variable and avoidance of healthcare. Variables found to be associated with the outcome at <italic>p</italic>&#x02009;&#x0003c;&#x02009;0.10 in bivariable analyses were included in a multivariable logistic regression model. All associations were considered statistically significant at the two-tailed <italic>p</italic>-value&#x02009;&#x0003c;&#x02009;0.05 threshold. A complete case analysis approach was employed, where cases with missing observations were excluded from the multivariable analyses. Analyses were performed using Stata/SE version 11.1 (Stata Corp, College Station, Texas).</p></sec><sec id="Sec3" sec-type="results"><title>Results</title><p>Overall, 452 self-identified TG women completed the survey and were included in the study. Baseline characteristics of the participants are presented in Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>. The majority of them were born in Argentina (90.0%), and had a median age of 30&#x000a0;years (interquartile range: 25&#x02013;37). Three hundred and seventy eight participants (84.6%) reported a history of sex work (61.1% were currently engaged in sex work), and among those with previous HIV testing (n&#x02009;=&#x02009;380, 84.1%), 27.4% reported being HIV-infected. As we have previously reported, characteristics among HIV-positive and HIV-negative TG women were similar [<xref ref-type="bibr" rid="CR35">35</xref>]. The only significant difference was that compared to HIV-negative participants, TG women who self-reported HIV infection were more likely to be living in unstable housing conditions or had experienced police violence. Overall, 184 respondents (40.7%) reported that they had ever avoided seeking healthcare because of their transgender identity.<table-wrap id="Tab1"><label>Table 1</label><caption><p> <bold>Baseline characteristics of transgender women included in the study, by whether they reported avoiding healthcare (N&#x02009;=&#x02009;452)</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th valign="middle" rowspan="3"> <bold>Characteristic</bold> </th><th valign="middle"> <bold>Total (%)</bold> </th><th valign="middle" colspan="2"> <bold>Healthcare avoidance</bold> </th><th valign="middle" rowspan="3"> <bold><italic>p</italic></bold> <bold>- value</bold> </th></tr><tr valign="top"><th valign="middle" rowspan="2"> <bold>(</bold> <bold><italic>n</italic></bold>&#x02009; <bold>=&#x02009;452)*</bold> </th><th valign="middle"> <bold>Yes (</bold> <bold><italic>%</italic></bold> <bold>)</bold> </th><th valign="middle"> <bold>No (%)</bold> </th></tr><tr valign="top"><th valign="middle"> <bold>(</bold> <bold><italic>n</italic></bold>&#x02009; <bold>=&#x02009;184)*</bold> </th><th valign="middle"> <bold>(</bold> <bold><italic>n</italic></bold>&#x02009; <bold>=&#x02009;268)*</bold> </th></tr></thead><tbody><tr valign="top"><td valign="middle"> <italic>Individual level factors</italic> </td><td valign="middle" align="center"/><td valign="middle" align="center"/><td valign="middle" align="center"/><td valign="middle" align="center"/></tr><tr valign="top"><td valign="middle">&#x02003;&#x02009; <bold>Age (median, IQR)</bold> </td><td valign="middle" align="left">30 (25 &#x02013; 37)</td><td valign="middle" align="left">30 (25 &#x02013; 27)</td><td valign="middle" align="left">31 (24 &#x02013; 37)</td><td valign="middle" align="left">0.861</td></tr><tr valign="top"><td valign="middle">&#x02003;&#x02009; <bold>Foreign born</bold> </td><td valign="middle" align="left">45 (10.0)</td><td valign="middle" align="left">17 (9.2)</td><td valign="middle" align="left">28 (10.5)</td><td valign="middle" align="left">0.673</td></tr><tr valign="top"><td valign="middle">&#x02003;&#x02009; <bold>High school education or greater</bold> </td><td valign="middle" align="left">152 (33.8)</td><td valign="middle" align="left">54 (29.5)</td><td valign="middle" align="left">98 (36.7)</td><td valign="middle" align="left">0.113</td></tr><tr valign="top"><td valign="middle">&#x02003;&#x02009; <bold>Extended health insurance</bold> </td><td valign="middle" align="left">81 (18.5)</td><td valign="middle" align="left">25 (13.8)</td><td valign="middle" align="left">56 (21.7)</td><td valign="middle" align="left">0.036</td></tr><tr valign="top"><td valign="middle">&#x02003;&#x02009; <bold>Currently employed (other than sex work)</bold> </td><td valign="middle" align="left">108 (23.9)</td><td valign="middle" align="left">35 (19.0)</td><td valign="middle" align="left">73 (27.2)</td><td valign="middle" align="left">0.044</td></tr><tr valign="top"><td valign="middle">&#x02003;&#x02009; <bold>History of sex work involvement</bold> </td><td valign="middle" align="left">378 (84.6)</td><td valign="middle" align="left">163 (89.1)</td><td valign="middle" align="left">215 (81.4)</td><td valign="middle" align="left">0.028</td></tr><tr valign="top"><td valign="middle">&#x02003;&#x02009; <bold>Self-reported HIV infection</bold> </td><td valign="middle" align="left">104 (27.4)</td><td valign="middle" align="left">41 (26.8)</td><td valign="middle" align="left">63 (27.8)</td><td valign="middle" align="left">0.838</td></tr><tr valign="top"><td valign="middle">&#x02003;&#x02009; <bold>Any internalized stigma</bold> </td><td valign="middle" align="left">245 (54.2)</td><td valign="middle" align="left">118 (64.1)</td><td valign="middle" align="left">127 (47.4)</td><td valign="middle" align="left">&#x0003c;0.001</td></tr><tr valign="top"><td valign="middle"> <italic>Social-structural factors</italic> </td><td valign="middle" align="left"/><td valign="middle" align="left"/><td valign="middle" align="left"/><td valign="middle" align="left"/></tr><tr valign="top"><td valign="middle">&#x02003;&#x02009; <italic>Police-related experiences</italic> </td><td valign="middle" align="left"/><td valign="middle" align="left"/><td valign="middle" align="left"/><td valign="middle" align="left"/></tr><tr valign="top"><td valign="middle">&#x02003;&#x02009; <bold>Experienced police violence ever</bold> </td><td valign="middle" align="left">243 (53.8)</td><td valign="middle" align="left">128 (69.6)</td><td valign="middle" align="left">115 (42.9)</td><td valign="middle" align="left">&#x0003c;0.001</td></tr><tr valign="top"><td valign="middle">&#x02003;&#x02009; <bold>Ever arrested</bold> </td><td valign="middle" align="left">354 (79.0)</td><td valign="middle" align="left">157 (85.8)</td><td valign="middle" align="left">197 (74.3)</td><td valign="middle" align="left">0.003</td></tr><tr valign="top"><td valign="middle">&#x02003;&#x02009; <italic>Experiences of perceived discrimination in healthcare settings</italic> </td><td valign="middle" align="left"/><td valign="middle" align="left"/><td valign="middle" align="left"/><td valign="middle" align="left"/></tr><tr valign="top"><td valign="middle">&#x02003;&#x02009; <bold>By healthcare workers ever</bold> </td><td valign="middle" align="left">302 (66.8)</td><td valign="middle" align="left">155 (84.2)</td><td valign="middle" align="left">147 (54.9)</td><td valign="middle" align="left">&#x0003c;0.001</td></tr><tr valign="top"><td valign="middle">&#x02003;&#x02009; <bold>By other patients ever</bold> </td><td valign="middle" align="left">143 (32.1)</td><td valign="middle" align="left">84 (46.4)</td><td valign="middle" align="left">59 (22.4)</td><td valign="middle" align="left">&#x0003c;0.001</td></tr><tr valign="top"><td valign="middle"> <italic>Environmental factors</italic> </td><td valign="middle" align="left"/><td valign="middle" align="left"/><td valign="middle" align="left"/><td valign="middle" align="left"/></tr><tr valign="top"><td valign="middle">&#x02003;&#x02009; <bold>Current residency in Buenos Aires</bold> </td><td valign="middle" align="left">140 (31.0)</td><td valign="middle" align="left">69 (37.5)</td><td valign="middle" align="left">71 (26.5)</td><td valign="middle" align="left">0.013</td></tr><tr valign="top"><td valign="middle">&#x02003;&#x02009; <bold>Stable housing</bold> </td><td valign="middle" align="left">355 (78.6)</td><td valign="middle" align="left">143 (77.7)</td><td valign="middle" align="left">212 (79.1)</td><td valign="middle" align="left">0.724</td></tr></tbody></table><table-wrap-foot><p>*Totals may differ due to non-response on some questions.</p><p>IQR: interquartile range.</p></table-wrap-foot></table-wrap></p><p>Table&#x000a0;<xref rid="Tab2" ref-type="table">2</xref> shows the results of the unadjusted and adjusted analyses of potential correlates of healthcare avoidance among TG women in our sample. Factors significantly and positively associated with avoidance of healthcare in bivariable analysis included: currently living in the Buenos Aires metropolitan area, having a history of sex work, experiencing internalized stigma as a result of one&#x02019;s gender identity, having ever been arrested, having been exposed to police violence (e.g., threats, beating, or sexual abuse), and having experienced discrimination in healthcare settings, both by healthcare workers and by patients (all <italic>p</italic>&#x02009;&#x0003c;&#x02009;0.05). In contrast, TG women with extended health insurance or a job other than sex work were less likely to report avoidance of healthcare (both <italic>p</italic>&#x02009;&#x0003c;&#x02009;0.05).<table-wrap id="Tab2"><label>Table 2</label><caption><p> <bold>Bi- and multivariable logistic regression analyses of individual, environmental and socio-structural factors associated with healthcare avoidance among transgender women in Argentina</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th valign="middle" rowspan="2"> <bold>Characteristic</bold> </th><th valign="middle" colspan="2"> <bold>Odds ratio (OR)</bold> </th></tr><tr valign="top"><th valign="middle"> <bold>Unadjusted OR (95% </bold> <bold>CI)</bold> </th><th valign="middle"> <bold>Adjusted OR (95% </bold> <bold>CI)</bold> </th></tr></thead><tbody><tr valign="top"><td valign="middle"> <italic>Individual level factors</italic> </td><td align="center"/><td align="center"/></tr><tr valign="top"><td valign="middle">&#x02003;&#x02009; <bold>Age&#x02009;&#x02265;&#x02009;30&#x000a0;years old</bold> </td><td align="left">1.02 (0.70 &#x02013; 1.48)</td><td align="left">&#x02014;</td></tr><tr valign="top"><td valign="middle">&#x02003;&#x02009; <bold>Foreign born</bold> </td><td align="left">0.87 (0.46 &#x02013; 1.64)</td><td align="left">&#x02014;</td></tr><tr valign="top"><td valign="middle">&#x02003;&#x02009; <bold>High school education or greater</bold> </td><td align="left">0.72 (0.48 &#x02013; 1.08)</td><td align="left">&#x02014;</td></tr><tr valign="top"><td valign="middle">&#x02003;&#x02009; <bold>Extended health insurance</bold> </td><td align="left">0.58 (0.35 &#x02013; 0.97)*</td><td align="left">0.49 (0.26 &#x02013; 0.93)</td></tr><tr valign="top"><td valign="middle">&#x02003;&#x02009; <bold>Currently employed (other than sex work)</bold> </td><td align="left">0.63 (0.40 &#x02013; 0.99)*</td><td align="left">1.00 (0.56 &#x02013; 1.79)</td></tr><tr valign="top"><td valign="middle">&#x02003;&#x02009; <bold>History of sex work involvement</bold> </td><td align="left">1.86 (1.06 &#x02013; 3.25)*</td><td align="left">0.86 (0.41 &#x02013; 1.80)</td></tr><tr valign="top"><td valign="middle">&#x02003;&#x02009; <bold>Self-reported HIV status</bold> </td><td align="left">0.95 (0.60 &#x02013; 1.51)</td><td align="left">&#x02014;</td></tr><tr valign="top"><td valign="middle">&#x02003;&#x02009; <bold>Any internalized stigma</bold> </td><td align="left">1.98 (1.35 &#x02013; 2.92)*</td><td align="left">1.60 (1.02 &#x02013; 2.51)</td></tr><tr valign="top"><td valign="middle"> <italic>Social-structural factors</italic> </td><td align="left"/><td align="left"/></tr><tr valign="top"><td valign="middle">&#x02003;&#x02009; <italic>Police-related experiences</italic> </td><td align="left"/><td align="left"/></tr><tr valign="top"><td valign="middle">&#x02003;&#x02009; <bold>Experienced police violence ever</bold> </td><td align="left">3.04 (2.05 &#x02013; 4.52)*</td><td align="left">2.20 (1.26 &#x02013; 3.83)</td></tr><tr valign="top"><td valign="middle">&#x02003;&#x02009; <bold>Ever arrested</bold> </td><td align="left">2.08 (1.27 &#x02013; 3.43)*</td><td align="left">1.00 (0.48 &#x02013; 2.11)</td></tr><tr valign="top"><td valign="middle">&#x02003;&#x02009; <italic>Experiences of perceived discrimination in healthcare settings</italic> </td><td align="left"/><td align="left"/></tr><tr valign="top"><td valign="middle">&#x02003;&#x02009; <bold>By healthcare workers ever</bold> </td><td align="left">4.40 (2.77 &#x02013; 7.00)*</td><td align="left">3.36 (1.25 &#x02013; 5.70)</td></tr><tr valign="top"><td valign="middle">&#x02003;&#x02009; <bold>By other patients ever</bold> </td><td align="left">3.01 (1.99 &#x02013; 4.54)*</td><td align="left">2.57 (1.58 &#x02013; 4.17)</td></tr><tr valign="top"><td valign="middle"> <italic>Environmental factors</italic> </td><td align="left"/><td align="left"/></tr><tr valign="top"><td valign="middle">&#x02003;&#x02009; <bold>Current residency in Buenos Aires</bold> </td><td align="left">1.66 (1.11 &#x02013; 2.49)*</td><td align="left">2.32 (1.44 &#x02013; 3.76)</td></tr><tr valign="top"><td valign="middle">&#x02003;&#x02009; <bold>Stable housing</bold> </td><td align="left">0.92 (0.58 &#x02013; 1.45)</td><td align="left">&#x02014;</td></tr></tbody></table><table-wrap-foot><p>*Significant at <italic>p&#x02009;&#x0003c;</italic>&#x02009;0.10 and entered into the multivariable model.</p><p>CI: confidence interval.</p></table-wrap-foot></table-wrap></p><p>As indicated in Table&#x000a0;<xref rid="Tab2" ref-type="table">2</xref>, factors that remained independently and positively associated with avoiding seeking healthcare in the multivariable analysis were having been exposed to police violence (adjusted odd ratio [aOR]&#x02009;=&#x02009;2.20; 95% CI: 1.26 &#x02013; 3.83), internalized stigma (aOR&#x02009;=&#x02009;1.60, 95% CI: 1.02&#x02013;2.51), having experienced discrimination by healthcare workers (aOR&#x02009;=&#x02009;3.36: 95% CI: 1.25 &#x02013; 5.70) or patients (aOR&#x02009;=&#x02009;2.57; 95% CI: 1.58 &#x02013; 4.17), and currently living in the Buenos Aires metropolitan area (aOR&#x02009;=&#x02009;2.32; 95% CI: 1.44 &#x02013; 3.76). Having extended health insurance remained significantly and negatively associated with the outcome (aOR&#x02009;=&#x02009;0.49; 95% CI: 0.26 &#x02013; 0.93).</p></sec><sec id="Sec4" sec-type="discussion"><title>Discussion</title><p>In this cross-sectional study, we found that approximately 40% of TG women reported avoiding healthcare because of their transgender identity. Reports of avoiding healthcare were found to be independently associated with internalized stigma, having experienced discrimination by healthcare workers or patients, having been exposed to police violence, and currently living in the Buenos Aires metropolitan area. In contrast, TG women with extended health insurance were less likely to report avoiding healthcare. These findings raise concern about the role of various social-structural factors in shaping access to healthcare among TG women, as well as concerns about suboptimal access to care among a population that is contending with various risks, including ongoing high rates of HIV infection [<xref ref-type="bibr" rid="CR19">19</xref>].</p><p>Previous studies have documented constrained access to healthcare among transgender communities throughout the globe [<xref ref-type="bibr" rid="CR8">8</xref>,<xref ref-type="bibr" rid="CR13">13</xref>,<xref ref-type="bibr" rid="CR14">14</xref>,<xref ref-type="bibr" rid="CR17">17</xref>,<xref ref-type="bibr" rid="CR36">36</xref>-<xref ref-type="bibr" rid="CR39">39</xref>]. Challenges that transgender women face when trying to access healthcare include inappropriate or nonexistent care protocols for transgender clients, as well as untrained and often discriminatory attitudes among health providers and staff [<xref ref-type="bibr" rid="CR10">10</xref>,<xref ref-type="bibr" rid="CR12">12</xref>,<xref ref-type="bibr" rid="CR14">14</xref>,<xref ref-type="bibr" rid="CR40">40</xref>]. These challenges, in turn, may contribute to TG women&#x02019;s perceptions of poor quality of and disrespectful care, which further shape their healthcare seeking behaviors. Our findings are consistent with previous studies, as we found that experiencing internalized stigma and having suffered discrimination in healthcare settings were positively associated with avoiding healthcare in multivariable analysis. Choosing not to seek medical care could be a way in which TG women cope with low levels of availability of gender-competent care and avoid further experiences of discrimination. Accordingly, training healthcare providers to ensure greater knowledge of and sensitivity to transgender health issues is critical. Wherever possible, transgender individuals should be involved in these initiatives.</p><p>Interestingly, the results of this study show that discrimination within healthcare facilities may come not only from healthcare providers and staff, but also from patients, which could be a reflection of the pervasiveness of transphobia within the larger society. Supporting this finding, a nation-wide qualitative study conducted by the Ministry of Health among 218 LGTB individuals (57 transgender individuals) showed that experiences of stigma and discrimination among this group are widespread, not only within the healthcare setting, but also in educational, work, and other public settings [<xref ref-type="bibr" rid="CR7">7</xref>]. Altogether, these findings suggest that other interventions, besides educating healthcare providers, might be needed in order to achieve culturally competent transgender friendly health care services. Increasing employment opportunities for transgender individuals in healthcare settings could have the dual benefit of creating more job opportunities for this population, as well as helping to create a more welcoming environment for transgender individuals seeking medical care [<xref ref-type="bibr" rid="CR3">3</xref>].</p><p>We also found a strong and positive association between experiences of police violence and avoidance of healthcare. Specifically, TG women may choose not to seek healthcare as a way to avoid further negative interactions with police or other personnel (e.g., security guards) in these environments. This finding is consistent with a large and growing body of literature pointing to the important role that policing practices can play in shaping the health of marginalized populations at risk for or living with HIV disease. A previous study of female sex workers in Vancouver, Canada, showed that policing presence and violence often displaces sex workers into more remote locations that are far away from health programs [<xref ref-type="bibr" rid="CR41">41</xref>]. Other studies have shown that fear of police can foster reluctance among people who use drugs to access HIV prevention programs or carry sterile syringes [<xref ref-type="bibr" rid="CR42">42</xref>-<xref ref-type="bibr" rid="CR44">44</xref>]. To our knowledge, this is the first study to reveal a relationship between police violence and the avoidance of healthcare among TG women. This finding is particularly concerning given the frequent confrontations with police among TG women, as demonstrated by the high rates of TG women in our sample who have been arrested (79%) or experienced different forms of police violence (54%). Future research should seek to elucidate the individual-level interactions and structural mechanisms that may explain this association, to better understand how police violence mediates healthcare access in this population. Regardless, efforts should be made to: ensure more effective monitoring of and sanctions for police misconduct; encourage systems for reporting and redress among victims; and provide educational approaches for police that focus on the unique experiences and needs of TG women.</p><p>Similar to previous research in other settings [<xref ref-type="bibr" rid="CR36">36</xref>-<xref ref-type="bibr" rid="CR38">38</xref>], we also found that TG women with extended health insurance were less likely to avoid seeking healthcare. This finding suggests that TG women lacking extended health insurance (as a proxy for low socio-economic status) may have other competing interests (e.g., money, food sources, distance, time) that could make seeking healthcare a low priority, even within a universal health care system. Therefore, additional efforts are needed to reach TG women with lower socio-economic status, and ensure access to appropriate and high-quality healthcare services in all 3 sub-sectors.</p><p>Finally, TG women living in the Buenos Aires metropolitan area (the biggest urban center in Argentina), were more likely to report avoiding healthcare. While we cannot exclude unmeasured confounding factors, such as frequent concentration of persons of lower socioeconomic status, as well as large migrant and minority populations in cities, this association persisted in the multivariable analysis. A possible explanation is that TG women residing in smaller cities may be more likely to have family support and tighter bonds with their communities, which can facilitate their access to and navigation through the health system. Alternatively, it might also be possible that TG women in Buenos Aires are more likely to be exposed to or engaged in the large transgender sex work scene in this city, and are therefore also more likely to experience police violence than those TG women living in smaller cities or towns. Indeed, previous research has described the relationship between sex work and drug scenes and police presence and misconduct [<xref ref-type="bibr" rid="CR43">43</xref>,<xref ref-type="bibr" rid="CR45">45</xref>]. Further work is needed to examine specific reasons for healthcare avoidance behaviors among TG women in Buenos Aires, specifically.</p><p>Collectively, these findings add to the growing body of evidence highlighting the importance of social, structural and environmental factors as drivers of disease burden and healthcare access among marginalized populations, such as injection drug users, sex workers, and transgender individuals [<xref ref-type="bibr" rid="CR2">2</xref>,<xref ref-type="bibr" rid="CR16">16</xref>,<xref ref-type="bibr" rid="CR34">34</xref>,<xref ref-type="bibr" rid="CR46">46</xref>,<xref ref-type="bibr" rid="CR47">47</xref>]. Policy-level interventions that protect transgender rights have potential to improve access to health care among this population. The recently passed Argentinean &#x0201c;Gender Identity Law&#x0201d; [<xref ref-type="bibr" rid="CR32">32</xref>] establishes a framework based on equity and human rights, and acknowledges the right to self-defined gender identity. Unlike many settings [<xref ref-type="bibr" rid="CR48">48</xref>], this law allows for changes to gender, image, or birth name on one&#x02019;s identity card and birth certificate without a requirement of psychiatric evaluation or judge approval. Since the enactment of this law in May 2012, over 3,000 TG individuals have acquired new government issued identification [<xref ref-type="bibr" rid="CR49">49</xref>]. This law further supports the right to the full development of one&#x02019;s person in line with one&#x02019;s chosen gender identity, and ensures the right to appropriate health services. For many TG women, body modification procedures are an important part of one&#x02019;s gender identity affirmation, and therefore increasing access to transition-related medical care could potentially help to engage and retain TG women in care. However, the impact of this new law remains under-evaluated, and further research is required to examine its health and social benefits in TG populations in Argentina.</p><p>Our study has a number of limitations. First, as there are no official registries of TG women in Argentina, our sample was not randomly selected, and therefore we cannot assume that our results are generalizable to all TG women in Argentina. We tried to mitigate this potential source of bias by recruiting a large sample, and by using a sampling quota technique to ensure the recruitment of TG participants from different age groups, educational levels and regions. Nevertheless, the possibility of non-random sampling bias remains. Second, our analysis relied on self-reported data, which may be susceptible to recall and social desirability biases. Third, although we followed and adapted previous validated instruments to measure internalized stigma among marginalized populations in Argentina such as the HIV Stigma and discrimination Index [<xref ref-type="bibr" rid="CR50">50</xref>], and despite the questionnaire design was informed by previous focus groups [<xref ref-type="bibr" rid="CR51">51</xref>] and tested by the transgender interviewers themselves, the internal consistency of our constructed &#x0201c;internalized stigma variable&#x0201d; was only acceptable (&#x003b1;&#x02009;=&#x02009;0.61). Fourth, another limitation of our study is that participants may hold diverse understandings of the term &#x0201c;healthcare&#x0201d;, and because we did not assess specifically what type of medical care (e.g., curative or preventive, minor or major illnesses) the participants avoided, we were unable to further contextualize this health behavior. Fifth, as the study is cross-sectional, and we included both lifetime and current explanatory variables (and the outcome referred to a lifetime behavior), temporality and causal associations could not be determined. Sixth, responses to many of the key variables of interest were limited to simple forced-choice responses (e.g., yes/no), which have may have precluded more detailed and in-depth responses. Accordingly, the findings presented here should be explored in greater detail through in-depth qualitative and ethnographic research. Lastly, although we performed multivariable analysis to adjust for known relevant confounders, there may be unmeasured factors (e.g., substance abuse) that may confound the relationship between exposures and outcomes.</p></sec><sec id="Sec5" sec-type="conclusion"><title>Conclusions</title><p>In summary, we found that 40% of TG women in our sample reported avoiding healthcare due to their transgender identity. Despite the limitations acknowledged, our study provides important information regarding contextual factors shaping healthcare access among TG women in Argentina. Of particular concern is the finding that, aside from the stigma and discrimination experienced within the healthcare setting, police violence was also independently associated with healthcare avoidance. Given the high burden of disease, and the high rates of discrimination and violence experienced by TG women in Argentina, there is an urgent need to adapt and develop socio-structural interventions that are tailored to promote healthcare access among this vulnerable population.</p></sec>

Genome-wide analysis of <italic>Cyclophilin</italic> gene family in soybean (<italic>Glycine max</italic>)

Could not extract abstract

<contrib contrib-type="author"><name><surname>Mainali</surname><given-names>Hemanta Raj</given-names></name><address><email>hemanta.mainali@agr.gc.ca</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author"><name><surname>Chapman</surname><given-names>Patrick</given-names></name><address><email>patrick.chapman@agr.gc.ca</email></address><xref ref-type="aff" rid="Aff2"/></contrib><contrib contrib-type="author" corresp="yes"><name><surname>Dhaubhadel</surname><given-names>Sangeeta</given-names></name><address><email>sangeeta.dhaubhadel@agr.gc.ca</email></address><xref ref-type="aff" rid="Aff1"/><xref ref-type="aff" rid="Aff2"/></contrib><aff id="Aff1"><label/>Department of Biology, University of Western Ontario, London, ON Canada </aff><aff id="Aff2"><label/>Agriculture and Agri-Food Canada, 1391 Sandford Street, London, ON Canada </aff>

BMC Plant Biology

<sec id="Sec1" sec-type="introduction"><title>Background</title><p>Cyclophilins (CYPs) are ubiquitous proteins found in organisms ranging from archaea and bacteria to plants and animals [<xref ref-type="bibr" rid="CR1">1</xref>,<xref ref-type="bibr" rid="CR2">2</xref>]. As they were originally identified as receptors for the immunosuppressive drug cyclosporine A (CsA), CYPs are classified in the immunophilin family of proteins possessing peptidyl prolyl <italic>cis/trans</italic> isomerase activity. Multiple CYPs have been found in genomes of various prokaryotes, but only a few have been studied in detail. The <italic>Escherichia coli</italic> genome encodes two CYPs, a cytosolic form (EcCYP-18) and its periplasmic counterpart (EcCYP-20) [<xref ref-type="bibr" rid="CR3">3</xref>]. In the yeast <italic>Saccharomyces cerevisiae</italic> there are at least 8 different CYPs, <italic>Cpr1</italic> to <italic>Cpr8</italic> [<xref ref-type="bibr" rid="CR4">4</xref>]. These proteins are not essential for growth, but are crucial for survival after heat stress [<xref ref-type="bibr" rid="CR5">5</xref>]. The human genome encodes 16 unique CYPs, categorized into 7 major groups, namely human CYP A (hCYP-A), hCYP-B, hCYP-C, hCYP-D, hCYP-E, hCYP-40 and hCYP-NK [<xref ref-type="bibr" rid="CR6">6</xref>]. The hCYP-A binds to CsA, and forms a ternary complex with calcineurin. The CsA-hCYP-A binding to calcineurin inhibits the phosphatase activity of calcineurin that results in a cascade of activities leading to the inactivation of T-cells [<xref ref-type="bibr" rid="CR7">7</xref>].</p><p>Compared to human CYPs, very little is known about plant CYPs. The first plant <italic>CYPs</italic> were identified concurrently from tomato (<italic>Lycopersicon esculentum</italic>), maize (<italic>Zea mays</italic>), and oilseed rape (<italic>Brassica napus</italic>) [<xref ref-type="bibr" rid="CR8">8</xref>]. Recently, with the availability of whole genome sequencing, the identification and characterization of plant CYPs has progressed substantially. However, compared to other organisms, the total number of plant <italic>CYPs</italic> in databases is still small, which suggests that many plant <italic>CYPs</italic> remain to be identified [<xref ref-type="bibr" rid="CR9">9</xref>]. To date, <italic>Arabidopsis thaliana</italic> and rice (<italic>Oryza sativa</italic>) are the two plant species reported to have highest number of CYPs with 35 <italic>AtCYPs</italic> [<xref ref-type="bibr" rid="CR10">10</xref>,<xref ref-type="bibr" rid="CR11">11</xref>] and 28 <italic>OsCYPs</italic> [<xref ref-type="bibr" rid="CR10">10</xref>,<xref ref-type="bibr" rid="CR12">12</xref>], respectively. Among the identified <italic>AtCYPs</italic>, only 15 are characterized at the molecular level [<xref ref-type="bibr" rid="CR11">11</xref>,<xref ref-type="bibr" rid="CR13">13</xref>&#x02013;<xref ref-type="bibr" rid="CR21">21</xref>]. Their encoded proteins are found in the cytoplasm [<xref ref-type="bibr" rid="CR17">17</xref>,<xref ref-type="bibr" rid="CR19">19</xref>,<xref ref-type="bibr" rid="CR20">20</xref>], endoplasmic reticulum (ER) [<xref ref-type="bibr" rid="CR18">18</xref>,<xref ref-type="bibr" rid="CR21">21</xref>], chloroplast [<xref ref-type="bibr" rid="CR15">15</xref>,<xref ref-type="bibr" rid="CR16">16</xref>], and nucleus [<xref ref-type="bibr" rid="CR13">13</xref>]. An increase in the expression of <italic>ROC1,</italic> an <italic>AtCYP,</italic> in response to light is associated with phytochromes and cryptochromes [<xref ref-type="bibr" rid="CR19">19</xref>,<xref ref-type="bibr" rid="CR22">22</xref>]. <italic>roc1</italic> mutants display an early flowering phenotype [<xref ref-type="bibr" rid="CR22">22</xref>], while gain-of-function mutations in <italic>ROC1</italic> reduce stem elongation and increase shoot branching [<xref ref-type="bibr" rid="CR23">23</xref>]. In contrast, loss-of-function mutations in <italic>AtCYP40</italic> reduce the number of juvenile leaves, with no change in inflorescence morphology or flowering time, and <italic>Arabidopsis</italic> plants with a defective <italic>AtCYP20-3</italic> are found to be hypersensitive to oxidative stress conditions created by high light and high salt levels, and osmotic shock [<xref ref-type="bibr" rid="CR24">24</xref>]. In addition to the AtCYPs having roles in various developmental processes, AtCYP59, a multi-domain CYP with a RNA recognition motif (RRM), regulates transcription and pre-mRNA processing by binding to the C-terminal domain of RNA polymerase II [<xref ref-type="bibr" rid="CR13">13</xref>]. Collectively, these results show the roles of <italic>Arabidopsis</italic> CYPs in different cellular pathways, which necessitate further work to explore the function associated with each of the CYPs.</p><p>Compared to the <italic>Arabidopsis</italic> CYPs, little work has been done on the rice CYPs. Most of the studies on the latter show their roles in different types of stresses. <italic>OsCYP2</italic> has been reported to have a role in different abiotic stress responses [<xref ref-type="bibr" rid="CR25">25</xref>]. The expression of <italic>OsCYP2</italic> is up-regulated towards salt stress, and its over-expression in rice enhances tolerance towards the salt stress. Similarly, overexpression in <italic>Arabidopsis</italic> and tobacco of the thylakoid-localized <italic>OsCYP20-2</italic> increased tolerance towards osmotic stress, and to extremely high light conditions [<xref ref-type="bibr" rid="CR26">26</xref>]. The expression levels of several other <italic>OsCYPs</italic> were increased by abiotic stresses such as desiccation and salt stress [<xref ref-type="bibr" rid="CR10">10</xref>,<xref ref-type="bibr" rid="CR12">12</xref>], indicating a critical role of OsCYPs during stress conditions.</p><p>Soybean (<italic>Glycine max</italic> [L.] Merr) is a legume plant belonging to the <italic>Papilionoideae</italic> family and is a rich source of protein, oil and plant natural products such as isoflavonoids. The soybean genome contains 56,044 protein coding loci located on 20 different chromosomes. Soybean has undergone two whole genome duplication events approximately 59 and 13 million years ago, as a result of which 75% of the genes have multiple copies [<xref ref-type="bibr" rid="CR27">27</xref>]. Until now, not much was known about soybean CYPs except that a handful of <italic>CYP</italic> gene sequences had been deposited in the public databases. We present here a genome-wide identification of soybean <italic>CYP</italic>s, their phylogenetic analysis, chromosomal distribution, and structural and expressional analysis<italic>.</italic> Our results indicate that soybean contains 62 CYPs, the largest family of CYP known to date in any organism. Further, the study describes a genome-wide segmental and tandem duplication during expansion of the <italic>GmCYP</italic> gene family.</p></sec><sec id="Sec2" sec-type="results"><title>Results and discussion</title><sec id="Sec3"><title>The soybean genome contains 62 putative <italic>GmCYPs</italic></title><p>To identify all the members of the <italic>CYP</italic> gene family in soybean, a BLASTN search of the soybean genome database <italic>G. max</italic> Wm82.a2.v1 (<ext-link ext-link-type="uri" xlink:href="http://phytozome.jgi.doe.gov/pz/portal.html#!search?show=BLAST&#x00026;method=Org_Gmax">http://phytozome.jgi.doe.gov/pz/portal.html#!search?show=BLAST&#x00026;method=Org_Gmax</ext-link>) was performed using the nucleotide sequence of a previously identified soybean <italic>CYP</italic> cDNA (GenBank: AF456323) as a query. This search identified 11 unique <italic>CYP</italic> genes. Each of the 11 <italic>CYP</italic> genes was used separately as a query sequence in the BLAST search of soybean genome database. This process was repeated until no new <italic>CYP</italic> gene was found. A total of 62 soybean <italic>CYPs,</italic> located on 18 different chromosomes, were identified and named <italic>GmCYP1</italic> to <italic>GmCYP62</italic> (Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>). Of the 62 <italic>GmCYPs</italic>, 52 encoded a protein with a single cyclophilin-like domain (CLD) which is responsible for the <italic>cis/trans</italic> isomerization of the peptidyl prolyl peptide bond. The remaining 10 GmCYPs contained the CLD and additional domains. As shown in Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>, GmCYP8, GmCYP9, GmCYP16, and GmCYP17 each contained two tetratricopeptide repeats (TPRs) at the C-terminus. The TPR motif is degenerate in nature and consists of a 34 amino acid repeat unit typically arranged in tandem arrays [<xref ref-type="bibr" rid="CR28">28</xref>]. Such TPR motif containing proteins mediate protein-protein interactions and often help in the assembly of multi-protein complexes. AtCYP40 (AGI:At2g15790), the <italic>Arabidopsis</italic> ortholog of GmCYP8, GmCYP9, GmCYP16, and GmCYP17, contains 3 TPRs and is involved in microRNA-mediated gene regulation [<xref ref-type="bibr" rid="CR29">29</xref>]. Loss-of-function mutation of <italic>AtCYP40</italic> showed a precocious phase change with reduced number of juvenile leaves, but no alteration of flowering time [<xref ref-type="bibr" rid="CR20">20</xref>]. Moreover, the conserved amino acids of the TPR domain of AtCYP40 are required for the interaction between AtCYP40 and cytoplasmic Hsp90 proteins. This interaction is essential for the function of AtCYP40 <italic>in planta</italic> [<xref ref-type="bibr" rid="CR29">29</xref>] suggesting a critical role for the TPR domain in microRNA-mediated gene regulation. Here we speculate a possibly similar function for the TPR domain in GmCYP8, GmCYP9, GmCYP16 and/or GmCYP17.<table-wrap id="Tab1"><label>Table 1</label><caption><p> <bold>Soybean</bold> <bold><italic>cyclophilin</italic></bold> <bold>gene family</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr><th> <bold>Gene name</bold> </th><th> <bold>Predicted transcript size (bp)</bold> </th><th> <bold>Predicted protein size (AA)</bold> </th><th> <bold>Predicted subcellular location</bold> </th><th> <bold>Domain information</bold> </th></tr></thead><tbody><tr><td> <italic>GmCYP1</italic> </td><td>973</td><td>172</td><td>Cytosol</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP32</italic> </td><td>1595</td><td>337</td><td>Secretory</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP2</italic> </td><td>1224</td><td>172</td><td>Cytosol</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP33</italic> </td><td>1301</td><td>373</td><td>Secretory</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP3</italic> </td><td>854</td><td>172</td><td>Cytosol</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP34</italic> </td><td>1065</td><td>204</td><td>Secretory</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP4</italic> </td><td>775</td><td>172</td><td>Cytosol</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP35</italic> </td><td>2543</td><td>616</td><td>Cytosol</td><td>MD</td></tr><tr valign="top"><td> <italic>GmCYP5</italic> </td><td>354</td><td>117</td><td>Cytosol</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP36</italic> </td><td>2559</td><td>668</td><td>Nucleus</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP6</italic> </td><td>393</td><td>130</td><td>Cytosol</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP37</italic> </td><td>1982</td><td>493</td><td>Cytosol</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP7</italic> </td><td>1072</td><td>175</td><td>Cytosol</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP38</italic> </td><td>582</td><td>114</td><td>Cytosol</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP8</italic> </td><td>1611</td><td>360</td><td>Cytosol</td><td>MD</td></tr><tr valign="top"><td> <italic>GmCYP39</italic> </td><td>1233</td><td>232</td><td>Secretory</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP9</italic> </td><td>1241</td><td>360</td><td>Cytosol</td><td>MD</td></tr><tr valign="top"><td> <italic>GmCYP40</italic> </td><td>1264</td><td>236</td><td>Secretory</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP10</italic> </td><td>1380</td><td>253</td><td>Chloroplast</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP41</italic> </td><td>1238</td><td>236</td><td>Secretory</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP11</italic> </td><td>1062</td><td>175</td><td>Cytosol</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP42</italic> </td><td>1087</td><td>165</td><td>Cytosol</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP12</italic> </td><td>711</td><td>236</td><td>Chloroplast</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP43</italic> </td><td>3138</td><td>850</td><td>Nucleus</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP13</italic> </td><td>793</td><td>164</td><td>Cytosol</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP44</italic> </td><td>2766</td><td>167</td><td>Secretory</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP14</italic> </td><td>1253</td><td>260</td><td>Chloroplast</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP45</italic> </td><td>1085</td><td>226</td><td>Secretory</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP15</italic> </td><td>1200</td><td>221</td><td>Cytosol</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP46</italic> </td><td>2532</td><td>843</td><td>Nucleus</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP16</italic> </td><td>1745</td><td>361</td><td>Cytosol</td><td>MD</td></tr><tr valign="top"><td> <italic>GmCYP47</italic> </td><td>1836</td><td>387</td><td>Chloroplast</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP17</italic> </td><td>1770</td><td>361</td><td>Cytosol</td><td>MD</td></tr><tr valign="top"><td> <italic>GmCYP48</italic> </td><td>1751</td><td>439</td><td>Chloroplast</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP18</italic> </td><td>2576</td><td>597</td><td>Nucleus</td><td>MD</td></tr><tr valign="top"><td> <italic>GmCYP49</italic> </td><td>1324</td><td>225</td><td>Mitochondria</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP19</italic> </td><td>2292</td><td>597</td><td>Nucleus</td><td>MD</td></tr><tr valign="top"><td> <italic>GmCYP50</italic> </td><td>2922</td><td>227</td><td>Chloroplast</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP20</italic> </td><td>2554</td><td>616</td><td>Nucleus</td><td>MD</td></tr><tr valign="top"><td> <italic>GmCYP51</italic> </td><td>947</td><td>232</td><td>Mitochondria</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP21</italic> </td><td>967</td><td>194</td><td>Cytosol</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP52</italic> </td><td>1724</td><td>439</td><td>Chloroplast</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP22</italic> </td><td>947</td><td>183</td><td>Cytosol</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP53</italic> </td><td>1983</td><td>445</td><td>Chloroplast</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP23</italic> </td><td>1349</td><td>251</td><td>Chloroplast</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP54</italic> </td><td>3559</td><td>849</td><td>Nucleus</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP24</italic> </td><td>1118</td><td>204</td><td>Secretory</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP55</italic> </td><td>1175</td><td>286</td><td>Chloroplast</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP25</italic> </td><td>1061</td><td>235</td><td>Secretory</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP56</italic> </td><td>2537</td><td>633</td><td>Nucleus</td><td>MD</td></tr><tr valign="top"><td> <italic>GmCYP26</italic> </td><td>1459</td><td>238</td><td>Secretory</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP57</italic> </td><td>546</td><td>181</td><td>Cytosol</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP27</italic> </td><td>2693</td><td>659</td><td>Nucleus</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP58</italic> </td><td>2374</td><td>350</td><td>Chloroplast</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP28</italic> </td><td>1869</td><td>263</td><td>Chloroplast</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP59</italic> </td><td>2403</td><td>640</td><td>Nucleus</td><td>MD</td></tr><tr valign="top"><td> <italic>GmCYP29</italic> </td><td>1822</td><td>326</td><td>Secretory</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP60</italic> </td><td>1921</td><td>445</td><td>Chloroplast</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP30</italic> </td><td>1988</td><td>327</td><td>Secretory</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP61</italic> </td><td>1493</td><td>230</td><td>Mitochondria</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP31</italic> </td><td>1645</td><td>337</td><td>PM/Mitochondria#</td><td>SD</td></tr><tr valign="top"><td> <italic>GmCYP62</italic> </td><td>1489</td><td>292</td><td>Mitochondria</td><td>SD</td></tr></tbody></table><table-wrap-foot><p>#, prediction with low confidence.</p></table-wrap-foot></table-wrap><fig id="Fig1"><label>Figure 1</label><caption><p> <bold>Schematic representation of multi-domain GmCYPs.</bold> CLD, cyclophilin-like-domain; TPR, tetratricopeptide repeat; WD, tryptophan-aspartate repeat; RRM, RNA recognition motif; ZK, zink knuckle, U-box, U-box domain.</p></caption><graphic xlink:href="12870_2014_282_Fig1_HTML" id="MO1"/></fig></p><p>GmCYP20 and GmCYP35 contain three tryptophan-aspartate (WD) repeats at the N-terminus (Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>). WD repeat-containing proteins are involved in a wide variety of cellular functions, providing binding sites for two or more proteins, or fostering transient interactions with other proteins [<xref ref-type="bibr" rid="CR30">30</xref>,<xref ref-type="bibr" rid="CR31">31</xref>]. The <italic>Arabidopsis</italic> CYP, AtCYP71 (AGI:At3g44600), contains 2 WD repeats [<xref ref-type="bibr" rid="CR11">11</xref>] and functions in chromatin remodelling [<xref ref-type="bibr" rid="CR14">14</xref>]. The very high sequence identity of AtCYP71 with GmCYP20 (87%) and GmCYP35 (83%) suggests that these two GmCYPs may play similar roles in soybean.</p><p>The sequence analysis identified two soybean CYPs, GmCYP56 and GmCYP59, having an RNA recognition motif (RRM) and zinc knuckle (ZK) at the C-terminus along with CLD at the N-terminus end (Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>). RRM is a small RNA binding motif of 90 amino acids and is conserved in a wide variety of organisms [<xref ref-type="bibr" rid="CR32">32</xref>]. AtCYP59 (AGI:At1g53720), the <italic>Arabidopsis</italic> ortholog of GmCYP56 (80%) and GmCYP59 (66%) (Additional file <xref rid="MOESM1" ref-type="media">1</xref>) [<xref ref-type="bibr" rid="CR11">11</xref>], contains an RRM motif, and is a transcriptional regulator [<xref ref-type="bibr" rid="CR13">13</xref>] that interacts with the conserved sequence of unprocessed mRNA, leading to the inhibition of the PPIase activity <italic>in vitro</italic> [<xref ref-type="bibr" rid="CR33">33</xref>]. Based on the functional association of AtCYP59 in transcriptional regulation, we speculate that the multi-domain soybean CYPs GmCYP56 and GmCYP59 possibly play a role in regulation of transcription in soybean <italic>via</italic> their RRMs.</p><p>Lastly, GmCYP18 and GmCYP19 contain a U-box at the N-terminus end of the protein. The U-box domain is highly conserved in some ubiquitin ligases and predicted to be a part of the ubiquitination machinery. Mammalian CYC4 and <italic>Arabidopsis</italic> AtCYP65 are the U-box containing CYP where the CYP domain is predicted to have chaperone function [<xref ref-type="bibr" rid="CR11">11</xref>,<xref ref-type="bibr" rid="CR34">34</xref>].</p><p>Of the 62 <italic>GmCYPs</italic>, it was ascertained that 13 contain a chloroplast transit peptide, 13 contain a signal peptide, 5 contain a mitochondrial targeting peptide, 10 contain a nuclear localization signal, and the remaining 21 are cytosolic (Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>). Unlike <italic>Arabidopsis</italic> and rice CYPs [<xref ref-type="bibr" rid="CR10">10</xref>], none of the soybean CYPs are predicted to be localized to the ER or golgi or plasma membrane. Only one secretory GmCYP, GmCYP39, is predicted for localization in the mitochondrial inner membrane or plasma membrane. A search for ER retention signal did not locate KDEL or HDEL in any GmCYP. We also searched for <italic>CYP</italic> genes in the DFCI soybean gene index that contains 1,354,268 ESTs representing 73,178 TC sequences (<ext-link ext-link-type="uri" xlink:href="http://compbio.dfci.harvard.edu/tgi/">http://compbio.dfci.harvard.edu/tgi/</ext-link>). Screening this database confirmed that 15 of the 62 <italic>GmCYP</italic> genes we identified were represented with 99-100% identity, and 100% coverage, implying that at least 25% of the <italic>GmCYPs</italic> are transcribed in various soybean tissues during normal growth and development, or in response to stress. Additionally, 33 <italic>GmCYPs</italic> displayed greater than 95% sequence identity with TC sequences in the soybean EST database, but with less than 100% query coverage. The lower sequence identities could be due simply to cultivar-specific sequence differences between the two databases, with the whole genome sequence originating solely from the cultivar Williams82 [<xref ref-type="bibr" rid="CR27">27</xref>], and the DFCI soybean gene index comprising EST data from cDNA libraries of several different soybean cultivars, the number of transcribed <italic>GmCYPs</italic> in soybean can be expected to be more than 15. A list of all the soybean <italic>CYP</italic> gene family members and their detailed information is provided in Additional file <xref rid="MOESM1" ref-type="media">1</xref>.</p></sec><sec id="Sec4"><title>Chromosomal distribution and phylogenetic analysis of soybean <italic>CYP</italic> genes</title><p>To determine the genome organization and distribution of <italic>GmCYPs</italic> on different chromosomes in soybean, a chromosome map was constructed. The results showed that the 62 <italic>GmCYPs</italic> are located on 18 different chromosomes. As depicted in Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>, the gene density per chromosome is uneven. Chromosome 11 and 19 contain the most, and show a relatively dense occurrence of <italic>CYP</italic> genes (6 each), whereas only one <italic>CYP</italic> (<italic>GmCYP54</italic>) is present on chromosome 14. No <italic>CYPs</italic> were found on chromosome 8 or 16. Most <italic>GmCYPs</italic> were localized towards the chromosome ends, and only <italic>GmCYP52, GmCYP49</italic> and <italic>GmCYP54</italic> were found near centromeres (Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>), suggesting the possibility of inter-chromosomal rearrangements, after genome duplication, between different soybean chromosomes.<fig id="Fig2"><label>Figure 2</label><caption><p> <bold>Genomic distributions of</bold> <bold><italic>GmCYP</italic></bold> <bold>genes on soybean chromosomes</bold> <bold><italic>.</italic></bold> Chromosomal locations of <italic>GmCYPs</italic> are indicated based on the location of the genes, length of chromosomes and positions of centromeres. The chromosomes are drawn to scale and chromosome numbers are shown under each chromosome. The <italic>GmCYPs</italic> that are clustered together and speculated to have undergone segmental duplication are indicated by shaded boxes of the same color and connected to each other by a line. Centromeres are indicated by blue ovals.</p></caption><graphic xlink:href="12870_2014_282_Fig2_HTML" id="MO2"/></fig></p><p>To explore the evolutionary relationship among soybean CYPs, a phylogenetic analysis was performed using their predicted amino acid sequences (Figure&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>). As observed for many other genes in soybean, most of the predicted GmCYPs clustered together in pairs, reflecting the ancient genome duplication event [<xref ref-type="bibr" rid="CR27">27</xref>,<xref ref-type="bibr" rid="CR35">35</xref>]. Such events result in two copies of each gene which undergo shuffling and rearrangement, creating the potential for new diversity. There are four possible fates of duplicated genes [<xref ref-type="bibr" rid="CR36">36</xref>]. First, one copy of the gene may be deleted during the course of evolution, resulting in loss of functional redundancy. Second, both copies of the genes may be retained and share the ancestral function, but gradually develop partially different functions (sub-functionalization). Third, one copy of the gene may acquire new function(s) during the course of evolution (neo-functionalization). Finally, there may be an intermediate stage between sub- and neo-functionalization. Which of these outcomes occur depends on the role of the specific gene in plant growth and development. Only those genes that are associated with critical functions for normal plant growth and development are retained, while others may be lost. The large number of <italic>CYP</italic>s present in the soybean genome thus likely reflects a combination of duplication and the important role of <italic>GmCYPs</italic> in soybean during normal growth and development, as well as in response to environmental stimuli.<fig id="Fig3"><label>Figure 3</label><caption><p> <bold>Phylogenetic relationships of GmCYP proteins.</bold> A Neighbor-Joining tree was generated by MEGA5.1 software [<xref ref-type="bibr" rid="CR51">51</xref>] using putative amino acid sequence of 62 GmCYPs, and the tree was annotated using Interactive Tree of Life [<xref ref-type="bibr" rid="CR52">52</xref>]. The numbers next to the branch shows the 1000 bootstrap replicates expressed in percentage. The solid line represents the real branch length and dotted lines added later for better visualization. The multi-domain CYPs are underlined and the predicted subcellular locations of GmCYPs are shown by colors as indicated.</p></caption><graphic xlink:href="12870_2014_282_Fig3_HTML" id="MO3"/></fig></p><p>Of the 62 GmCYPs, 54 are clustered in pairs (27 pairs) in the phylogenetic tree. The remaining 8 GmCYPs branched-off from the terminal branch of another pair of GmCYPs. This analysis further revealed that the multi-domain GmCYPs cluster together.</p><p>We also attempted to correlate the clustering of GmCYPs in the phylogenetic tree with their predicted subcellular localization. Interestingly, the GmCYPs predicted to be targeted to the same subcellular compartment grouped together as a separate clade. For example, GmCYPs with chloroplast transit peptide (GmCYP10, GmCYP23, GmCYP14, GmCYP28 and GmCYP12) formed a distinct clade on the tree. Another 4 chloroplast-localizing GmCYPs (GmCYP48, GmCYP52, GmCYP53, and GmCYP60) also formed a distinct clade, but in a different location on the tree. Similarly, the GmCYPs with nuclear localization signal (GmCYP27, GmCYP36, and GmCYP54, GmCYP43, GmCYP46) also formed separate clades on the tree (Figure&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>). A similar pattern of gene clustering was observed for the GmCYPs predicted to localize in mitochondria or that were secretory.</p><p>By comparing the positions of <italic>GmCYPs</italic> on the chromosome map (Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>) and in the phylogenetic tree (Figure&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>), an interesting grouping pattern was observed. If the <italic>GmCYPs</italic> were localized together on a chromosome, their paralogs were also found together on a different chromosome. For example, <italic>GmCYP4, GmCYP39,</italic> and <italic>GmCYP46</italic> are clustered at the sub-telomere region of chromosome 4, and are most similar to <italic>GmCYP3, GmCYP38,</italic> and <italic>GmCYP43</italic>, respectively, which are clustered together in the sub-telomere region of chromosome 6 (Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>). Similarly, <italic>GmCYP36, GmCYP13,</italic> and <italic>GmCYP7</italic> (chromosome 3) paired with <italic>GmCYP27, GmCYP15,</italic> and <italic>GmCYP11,</italic> respectively, from chromosome 19, and <italic>GmCYP18</italic> and <italic>GmCYP32</italic> (chromosome 11) paired with <italic>GmCYP19</italic> and <italic>GmCYP31</italic>, respectively (chromosome 1), whereas <italic>GmCYP1</italic> and <italic>GmCYP41</italic> (chromosome 11) paired with <italic>GmCYP2</italic> and <italic>GmCYP40</italic> (chromosome 12). Moreover, <italic>GmCYP34</italic> and <italic>GmCYP26</italic>, from chromosome 11, paired up with <italic>GmCYP24</italic> and <italic>GmCYP25</italic>, respectively, from chromosome 18. These findings provide strong evidence for segmental duplication of chromosomal regions containing the <italic>GmCYPs</italic>, such as has been shown to play a vital role in the evolutionary generation of members of other gene families [<xref ref-type="bibr" rid="CR37">37</xref>,<xref ref-type="bibr" rid="CR38">38</xref>].</p></sec><sec id="Sec5"><title>Gene structures of <italic>GmCYPs</italic></title><p>Analysis of the exon-intron structure of the <italic>GmCYP</italic> genes showed several variations (Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>). Six <italic>GmCYP</italic> genes (<italic>GmCYP1- GmCYP4</italic>, <italic>GmCYP6</italic> and <italic>GmCYP7</italic>) contained no intron in their open reading frame (ORF). The number of introns varied from 1 to 13 in the ORFs of other <italic>GmCYP</italic> genes. The <italic>GmCYP5, GmCYP47, GmCYP50, GmCYP55</italic> and <italic>GmCYP58</italic> contained a single intron in their ORF while the largest numbers of introns were found in <italic>GmCYP56</italic>. The size of intron also varied considerably between different <italic>GmCYP</italic> gene family members with their size ranging from 39&#x000a0;bp (<italic>GmCYP5</italic>) to 9359&#x000a0;bp (<italic>GmCYP56</italic>) in the primary transcripts. Several other genes such as <italic>GmCYP22, GmCYP30, GmCYP34, GmCYP39-GmCYP42, GmCYP45, GmCYP49, GmCYP51- GmCYP53, GmCYP55- GmCYP57</italic> and <italic>GmCYP59</italic> contained introns larger than 4.0&#x000a0;kb in their ORFs. It has been suggested that the genome size may be correlated with intron size and that some elements of genome size evolution occurs within the gene [<xref ref-type="bibr" rid="CR39">39</xref>]. However, in <italic>Gossypium sps.,</italic> intron and genome size evolution are not coupled [<xref ref-type="bibr" rid="CR40">40</xref>]. In the regions of low recombination, longer introns are selectively advantageous as they improve recombination and possibly counterbalance the mutational bias towards deletion [<xref ref-type="bibr" rid="CR41">41</xref>]. A large-scale comparative analysis of intron positions among different kingdoms (animal, plant and fungus) identified a large number of positions that are likely to be ancestral [<xref ref-type="bibr" rid="CR42">42</xref>]. Analysis of intron sizes and positions in paralogs among <italic>GmCYP</italic> family did not show any specific pattern. However, in the majority of cases, the exon-intron numbers were similar in the genes that clustered together in the phylogenetic tree (Figure&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>), for example, <italic>GmCYP25</italic> and <italic>GmCYP26</italic> or <italic>GmCYP47</italic> and <italic>GmCYP58</italic> or <italic>GmCYP20</italic> and <italic>GmCYP35</italic>. The 5&#x02032; and 3&#x02032; untranslated regions (UTR) that border protein-coding sequences are important structural and regulatory elements of eukaryotic genes [<xref ref-type="bibr" rid="CR43">43</xref>] and also contain large numbers of introns [<xref ref-type="bibr" rid="CR44">44</xref>]. Out of 62 <italic>GmCYPs</italic>, 12 contained a single intron in the 5&#x02032;UTR region while remaining <italic>GmCYPs</italic> consisted of intronless 5&#x02032;UTR. The 3&#x02032;UTR of two <italic>GmCYPs, GmCYP16</italic> and <italic>GmCYP50,</italic> were interrupted by a single intron whereas <italic>GmCYP17</italic> and <italic>GmCYP44</italic> contained 2 and 5 introns, respectively. The number of exon and intron in each <italic>GmCYP</italic> gene is shown in Additional file <xref rid="MOESM2" ref-type="media">2</xref>.<fig id="Fig4"><label>Figure 4</label><caption><p> <bold>Schematic diagrams of the exon-intron structures, and splice variants of</bold> <bold><italic>GmCYPs</italic></bold> <bold>.</bold> Exon-intron structures of <italic>GmCYPs</italic> were compiled from Phyotozome database (<ext-link ext-link-type="uri" xlink:href="http://phytozome.jgi.doe.gov/pz/portal.html#!info?alias=Org_Gmax">http://phytozome.jgi.doe.gov/pz/portal.html#!info?alias=Org_Gmax</ext-link>). <italic>GmCYP</italic> with predicted alternate transcripts are shown below the corresponding genes. The green boxes, black boxes and lines indicate exons, UTRs and introns, respectively. Left to right direction of transcript shows &#x0201c;+&#x0201d; strand while right to left shows &#x0201c;-&#x0201d; strand, relative to the annotation of the genome sequence. Gene structure images are drawn to scale except for <italic>GmCYP50</italic>, <italic>GmCYP56</italic>, and <italic>GmCYP59</italic>, where diagrams are reduced to 0.5X, 0.35X, and 0.5X, respectively.</p></caption><graphic xlink:href="12870_2014_282_Fig4_HTML" id="MO4"/></fig></p></sec><sec id="Sec6"><title>Expression analysis of <italic>GmCYP</italic> genes</title><p>To determine expression patterns of <italic>GmCYP</italic> genes, we used publicly-available genome-wide transcript profiling data of soybean tissues as a resource (<ext-link ext-link-type="uri" xlink:href="http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE29163">http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE29163</ext-link>). The dataset contains RNAseq reads from soybean seeds across several stages of seed development (globular, heart, cotyledon, early-maturation, dry), and reproductive (floral buds) and vegetative (leaves, roots, stems, seedlings) tissues. As shown in Figure&#x000a0;<xref rid="Fig5" ref-type="fig">5</xref>, most of the <italic>GmCYP</italic> genes showed distinct tissue-specific expression pattern. Out of the 62 <italic>GmCYP</italic> genes, 26 were expressed in the vegetative tissues whereas 34 were expressed in floral buds and different stages of seed development. Two <italic>GmCYPs</italic>, <italic>GmCYP5</italic> and <italic>GmCYP6,</italic> contained no sequence read in any of the soybean tissues included in the study. In addition, there were no EST or TC sequences in the DFCI gene index database with a perfect match to <italic>GmCYP5</italic> and <italic>GmCYP6</italic> (Additional file <xref rid="MOESM1" ref-type="media">1</xref>). These evidences indicated that <italic>GmCYP5</italic> and <italic>GmCYP6</italic> are pseudogenes or expressed under special conditions or at specific developmental stages. The gene expression data revealed that the majority of <italic>GmCYPs</italic> (41%) were expressed in leaf tissue with the highest transcript accumulation level. Furthermore, it is interesting to note that the GmCYPs predicted to localize in the chloroplast were expressed in leaf tissues, suggesting their possible role in photosynthesis. Expression of several <italic>GmCYP</italic> genes in seed tissues during development indicates an important role of these genes in seed development.<fig id="Fig5"><label>Figure 5</label><caption><p> <bold>Expression analyses of soybean</bold> <bold><italic>CYP</italic></bold> <bold>genes.</bold> The transcriptome data of soybean across different tissues and developmental stages were obtained from the National Center for Biotechnology Information (<ext-link ext-link-type="uri" xlink:href="http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE29163">http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE29163</ext-link>) for heatmap generation. The color scale below the heat map indicates expression values, green indicating low transcript abundance and red indicating high levels of transcript abundance. Clustering of <italic>GmCYPs</italic> in <bold>(I)</bold> vegetative tissue or <bold>(II)</bold> reproductive or seed tissue is shown.</p></caption><graphic xlink:href="12870_2014_282_Fig5_HTML" id="MO5"/></fig></p></sec></sec><sec id="Sec7" sec-type="conclusion"><title>Conclusions</title><p>Taken together, we have performed a comprehensive sequence analysis of soybean <italic>CYP</italic> genes (<italic>GmCYPs</italic>), and provided detailed information on them. Specifically, our results show that the soybean genome contains 62 <italic>CYP</italic> genes, the largest <italic>CYP</italic> gene family identified in any organism to date. The presence of predicted motifs, subcellular localization and their sequence homology with other identified CYPs from other organisms provided insight into their putative function. Results of the present study indicate a genome-wide segmental and tandem duplication during expansion of the <italic>GmCYP</italic> gene family.</p></sec><sec id="Sec8" sec-type="materials|methods"><title>Methods</title><sec id="Sec9"><title>Database search for <italic>CYP</italic> genes in soybean</title><p>To identify all the <italic>CYPs</italic> present in the soybean genome, the nucleotide sequence of the <italic>GmCYP1</italic> (GenBank:AF456323) was used for a BLASTN [<xref ref-type="bibr" rid="CR45">45</xref>] query against the new soybean genome database (Wm82.a2.v1) (<ext-link ext-link-type="uri" xlink:href="http://phytozome.jgi.doe.gov/pz/portal.html">http://phytozome.jgi.doe.gov/pz/portal.html</ext-link>) [<xref ref-type="bibr" rid="CR46">46</xref>]. The newly identified sequences were subsequently used as queries to find other less similar <italic>GmCYP</italic>s. The chromosomal locations for all <italic>GmCYPs</italic> were obtained from the soybean genome database to draw the chromosomal map. The molecular weight for each GmCYP was calculated using ProtParam software [<xref ref-type="bibr" rid="CR47">47</xref>] (<ext-link ext-link-type="uri" xlink:href="http://web.expasy.org/protparam/">http://web.expasy.org/protparam/</ext-link>). TargetP1 [<xref ref-type="bibr" rid="CR48">48</xref>] (<ext-link ext-link-type="uri" xlink:href="http://www.cbs.dtu.dk/services/TargetP/">http://www.cbs.dtu.dk/services/TargetP/</ext-link>) and WoLF-PSORT [<xref ref-type="bibr" rid="CR49">49</xref>] (<ext-link ext-link-type="uri" xlink:href="http://wolfpsort.org/">http://wolfpsort.org/</ext-link>) were used to identify putative sub-cellular localization of the predicted protein sequences, and domain information was obtained from the soybean genome database [<xref ref-type="bibr" rid="CR27">27</xref>]. To identify the transcribed <italic>GmCYP</italic>s in soybean, the coding sequence of each <italic>GmCYP</italic> was used as a query to BLAST against the soybean gene index (<ext-link ext-link-type="uri" xlink:href="http://compbio.dfci.harvard.edu/tgi/">http://compbio.dfci.harvard.edu/tgi/</ext-link>). The Tentative Contig (TC) sequences in the soybean gene index database were aligned with the corresponding <italic>GmCYP</italic> sequences to identify the percentage identity and coverage. Similarly, to find the GmCYP orthologs in <italic>Arabidopsis</italic>, the amino acid sequences of GmCYPs were used as queries to BLAST against the <italic>Arabidopsis</italic> protein database (<ext-link ext-link-type="uri" xlink:href="http://www.arabidopsis.org/">http://www.arabidopsis.org/</ext-link>) [<xref ref-type="bibr" rid="CR50">50</xref>].</p></sec><sec id="Sec10"><title>Multiple sequence alignment and phylogenetic analyses</title><p>To investigate the phylogenetic relationships among GmCYP proteins, and their molecular evolution, a phylogenetic tree was generated. Multiple sequence alignment of the deduced amino acid sequences of all GmCYP proteins were aligned by Clustal X and the alignment was imported into MEGA5.1 to create a phylogenetic tree [<xref ref-type="bibr" rid="CR51">51</xref>]. Neighbour-Joining method was used with 1000 bootstrap replicates. The tree was exported into the Interactive Tree Of Life (<ext-link ext-link-type="uri" xlink:href="http://itol.embl.de">http://itol.embl.de</ext-link>) for annotation and manipulation [<xref ref-type="bibr" rid="CR52">52</xref>].</p></sec><sec id="Sec11"><title>Expression analysis of soybean <italic>CYP</italic> genes</title><p>To determine the expression patterns of <italic>CYP</italic> genes in soybean tissues, the publically available transcriptome data (<ext-link ext-link-type="uri" xlink:href="http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE29163">http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE29163</ext-link>) was used as a main source. The illumina sequencing of transcripts from ten different soybean tissues were downloaded from the NCBI database (<ext-link ext-link-type="uri" xlink:href="http://www.ncbi.nlm.nih.gov/">http://www.ncbi.nlm.nih.gov/</ext-link>) with accession numbers SRX062325-SRX062334. After normalization of the dataset, the value of each gene was centered by subtracting the mean normalized value for each gene and scaled by dividing the centered value by the standard deviation of the gene following Eisen et al. [<xref ref-type="bibr" rid="CR53">53</xref>]. The heatmap for <italic>GmCYP</italic> genes was generated in R using the heatmap.2 function from the gplots CRAN library (<ext-link ext-link-type="uri" xlink:href="http://cran.r-project.org/package=gplots">http://CRAN.R-project.org/package=gplots</ext-link>).</p></sec></sec><sec id="Sec12"><title>Availability of supporting data</title><p>Phylogenetic data (tree and data used to generate them) have been deposited in TreeBASE repository and is available under the URL <ext-link ext-link-type="uri" xlink:href="http://purl.org/phylo/treebase/phylows/study/TB2:S16455">http://purl.org/phylo/treebase/phylows/study/TB2:S16455</ext-link>.</p></sec>

Stroke, multimorbidity and polypharmacy in a nationally representative sample of 1,424,378 patients in Scotland: implications for treatment burden

Could not extract abstract

<contrib contrib-type="author"><name><surname>Gallacher</surname><given-names>Katie I</given-names></name><address><email>katie.gallacher@glasgow.ac.uk</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author"><name><surname>Batty</surname><given-names>G David</given-names></name><address><email>david.batty@ucl.ac.uk</email></address><xref ref-type="aff" rid="Aff2"/><xref ref-type="aff" rid="Aff3"/></contrib><contrib contrib-type="author"><name><surname>McLean</surname><given-names>Gary</given-names></name><address><email>gary.mclean@glasgow.ac.uk</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author"><name><surname>Mercer</surname><given-names>Stewart W</given-names></name><address><email>stewart.mercer@glasgow.ac.uk</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author"><name><surname>Guthrie</surname><given-names>Bruce</given-names></name><address><email>b.guthrie@dundee.ac.uk</email></address><xref ref-type="aff" rid="Aff4"/></contrib><contrib contrib-type="author"><name><surname>May</surname><given-names>Carl R</given-names></name><address><email>c.r.may@soton.ac.uk</email></address><xref ref-type="aff" rid="Aff5"/></contrib><contrib contrib-type="author"><name><surname>Langhorne</surname><given-names>Peter</given-names></name><address><email>peter.langhorne@glasgow.ac.uk</email></address><xref ref-type="aff" rid="Aff6"/></contrib><contrib contrib-type="author" corresp="yes"><name><surname>Mair</surname><given-names>Frances S</given-names></name><address><email>frances.mair@glasgow.ac.uk</email></address><xref ref-type="aff" rid="Aff1"/></contrib><aff id="Aff1"><label/>Institute of Health and Wellbeing, University of Glasgow, Glasgow, G12 9LX Scotland </aff><aff id="Aff2"><label/>Department of Epidemiology and Public Health, University College London, London, WC1E 6BT England </aff><aff id="Aff3"><label/>University of Edinburgh, Centre for Cognitive Ageing and Cognitive Epidemiology, Edinburgh, EH8 9JZ Scotland </aff><aff id="Aff4"><label/>University of Dundee, School of Medicine, Dundee, DD2 4BF Scotland </aff><aff id="Aff5"><label/>University of Southampton, Faculty of Health Sciences, Southampton, SO17 1BJ England </aff><aff id="Aff6"><label/>Institute of Cardiovascular and Medical Sciences, University of Glasgow, Glasgow, G4 0SF Scotland </aff>

BMC Medicine

<sec id="Sec1" sec-type="introduction"><title>Background</title><p>Multimorbidity, defined as the presence of two or more long-term conditions, is becoming a global challenge for policy-makers, clinicians, and patients [<xref ref-type="bibr" rid="CR1">1</xref>-<xref ref-type="bibr" rid="CR3">3</xref>]. Treatment advances and increasing sub-specialisation of health services have improved functional outcomes for those with long-term conditions, but such changes have resulted in an increasing burden of treatment demands on patients, particularly those with multimorbidity [<xref ref-type="bibr" rid="CR3">3</xref>,<xref ref-type="bibr" rid="CR4">4</xref>]. Treatment burden is defined as the workload of healthcare for patients and the impact of this on their wellbeing [<xref ref-type="bibr" rid="CR5">5</xref>]. It includes information gathering, attending multiple appointments, taking medications, enacting self-care, and, in countries that lack a health service that is free at the point of care, organising finances to pay for treatments [<xref ref-type="bibr" rid="CR5">5</xref>-<xref ref-type="bibr" rid="CR8">8</xref>]. There is a risk that patients become overburdened by their treatments, which can mean failure to adhere to management plans, thus resulting in ineffective treatment and wasted resources [<xref ref-type="bibr" rid="CR3">3</xref>,<xref ref-type="bibr" rid="CR9">9</xref>-<xref ref-type="bibr" rid="CR11">11</xref>].</p><p>One aspect of treatment burden described above is polypharmacy, which can contribute to other treatment burdens such as adverse drug events [<xref ref-type="bibr" rid="CR12">12</xref>,<xref ref-type="bibr" rid="CR13">13</xref>]. Polypharmacy is most commonly defined as the use of multiple (usually five or ten) prescribed medications [<xref ref-type="bibr" rid="CR14">14</xref>-<xref ref-type="bibr" rid="CR16">16</xref>]. Although there is no strong evidence to support the use of any particular threshold, the risk of drug-related problems seems to increase with each additional medication prescribed [<xref ref-type="bibr" rid="CR17">17</xref>,<xref ref-type="bibr" rid="CR18">18</xref>]. There is a known association between number of morbidities and polypharmacy [<xref ref-type="bibr" rid="CR19">19</xref>-<xref ref-type="bibr" rid="CR21">21</xref>], with a study using routine Scottish health records finding that of those with two clinical conditions, 20.8% were receiving four to nine medications, and 1.1% were receiving ten or more medications; for patients with six or more comorbidities, these values were 47.7% and 41.7%, respectively [<xref ref-type="bibr" rid="CR19">19</xref>]. A systematic literature review investigating the relationship between the number of chronic conditions and healthcare utilisation outcomes found that about 60% of elderly respondents with zero or one condition reported taking prescription medications. This percentage went up to more than 90% for those with two or three conditions, and approached 100% for those with more than five conditions [<xref ref-type="bibr" rid="CR20">20</xref>], supporting the premise that those with higher numbers of conditions to manage are more likely to experience higher levels of treatment burden [<xref ref-type="bibr" rid="CR3">3</xref>]. Other aspects of treatment burden such as healthcare utilisation have also been shown to be associated with multimorbidity [<xref ref-type="bibr" rid="CR20">20</xref>,<xref ref-type="bibr" rid="CR22">22</xref>].</p><p>Stroke is a condition that can have a considerable impact on an individual&#x02019;s life. A recent systematic review of the qualitative literature revealed that people who have had a stroke experience four main areas of treatment burden: making sense of stroke management and planning care, interacting with others, enacting management strategies, and reflecting on management [<xref ref-type="bibr" rid="CR23">23</xref>]. Poor communication between patients and professionals was a common experience, exacerbated by fragmentation of health services and poor communication between healthcare providers themselves, aspects of stroke care likely to be exacerbated by multimorbidity [<xref ref-type="bibr" rid="CR24">24</xref>-<xref ref-type="bibr" rid="CR26">26</xref>]. Surprisingly, there has been limited exploration of multimorbidity or polypharmacy in people with stroke, the field being characterised by small-scale studies and a small number of conditions under examination [<xref ref-type="bibr" rid="CR19">19</xref>,<xref ref-type="bibr" rid="CR27">27</xref>-<xref ref-type="bibr" rid="CR36">36</xref>]. Those studies that have examined stroke in relation to other long-term conditions have suggested that stroke is one of the diseases most significantly associated with polypharmacy [<xref ref-type="bibr" rid="CR19">19</xref>,<xref ref-type="bibr" rid="CR33">33</xref>], but there is a lack of large-scale studies examining a broad range of medications and comorbidities.</p><p>In the current study, using a large, nationally representative cross-sectional primary care dataset, we examined the prevalence of multimorbidity and polypharmacy in people with and without stroke.</p></sec><sec id="Sec2" sec-type="materials|methods"><title>Methods</title><sec id="Sec3"><title>Study design and participants</title><p>This was a cross-sectional study based on a nationally representative dataset managed by the Primary Care Clinical Informatics Unit at the University of Aberdeen in Scotland. This fully anonymised dataset contains clinical data on all people that were alive and permanently registered with 314 primary care practices in Scotland on 31 March 2007. Comprising approximately one-third of the Scottish adult population, this sample has been shown to be representative of this population [<xref ref-type="bibr" rid="CR37">37</xref>]. In the UK, registration with a medical practice is required for an individual to access National Health Service (NHS) healthcare in the community. It is estimated that over 98% of the population are registered with a medical practice [<xref ref-type="bibr" rid="CR38">38</xref>], which systematically records information on each patient in an electronic medical record, for the purposes of registration and subsequent everyday medical care. We examined data extracted from medical records and collated for a previous study of multimorbidity that had examined the presence of forty conditions [<xref ref-type="bibr" rid="CR1">1</xref>]. The NHS National Research Ethics Service approved the use of these data for research purposes. Patient consent was not deemed necessary due to full anonymisation of the data.</p></sec><sec id="Sec4"><title>Data collected and disease definition</title><p>The data examined consisted of the following variables: sex, age, socioeconomic deprivation (measured from patients&#x02019; postcodes using the Carstairs score [<xref ref-type="bibr" rid="CR39">39</xref>]), counts of regularly prescribed medications and the presence of 40 long-term conditions, including stroke.</p><p>There is no &#x02018;gold standard&#x02019; method for the measurement of multimorbidity, therefore the forty long-term conditions included had been chosen and defined based on a recent systematic review [<xref ref-type="bibr" rid="CR40">40</xref>] and expert consensus [<xref ref-type="bibr" rid="CR1">1</xref>]. Existing definitions for each long-term condition were used if possible, mainly those used in the Quality and Outcomes Framework (QOF) or by NHS Scotland [<xref ref-type="bibr" rid="CR1">1</xref>,<xref ref-type="bibr" rid="CR41">41</xref>,<xref ref-type="bibr" rid="CR42">42</xref>]. If no standard definition was available, or there was concern about under-recording, then conditions were defined by the clinical members of the research team. For example, depression was defined as the presence of a QOF Read Code in the past year or receipt of four or more prescriptions for antidepressant drugs (excluding low-dose tricyclics, which are usually used for chronic pain) in the past year [<xref ref-type="bibr" rid="CR1">1</xref>]. The definitions of all morbidities examined are given in supplementary material (see Additional file <xref rid="MOESM1" ref-type="media">1</xref>). Comorbidity was measured using a count of long-term conditions [<xref ref-type="bibr" rid="CR43">43</xref>], with morbidities being noted as either mental health or physical morbidities. The original analysis measured the presence of a combined group of stroke or transient ischaemic attack (TIA), but for the purposes of this analysis, the presence of stroke alone was defined using the QOF Business Rules code set [<xref ref-type="bibr" rid="CR41">41</xref>], and TIA was ignored.</p><p>As there are no standard definitions of regularly prescribed treatments or measure of polypharmacy, we utilised a count of current regular prescriptions, including tablets, inhalers, stoma care and topical therapies [<xref ref-type="bibr" rid="CR17">17</xref>,<xref ref-type="bibr" rid="CR18">18</xref>]. Regular (&#x02018;repeat&#x02019;) prescriptions are clearly distinguished in UK general practice electronic medical records from one-off (&#x02018;acute&#x02019;) prescriptions such as those for most antibiotics. For the purposes of this analysis, any regular prescription that was still active (that is, available for issue on request) on the date of extraction and that had been prescribed in the past 84&#x000a0;days was counted as current. This time frame was selected as this was the maximum length of a repeat prescription in Scotland at the time of data collection.</p></sec><sec id="Sec5"><title>Statistical analysis</title><p>Analyses were predicated on a comparison of the characteristics of people with stroke (cases) and those without stroke (controls). First, the numbers of morbidities and prescribed medications in stroke cases and controls were calculated, and proportions within each group computed. Second, logistic regression, which produces ORs, was used to summarise the relationship between stroke and the presence of comorbidities and prescribed medications. ORs were initially unadjusted &#x02013; for the purposes of comparison &#x02013; then adjusted for the key confounding factors of age, sex and socioeconomic deprivation. Age and deprivation were used as continuous variables. Deprivation was measured using the Carstairs score, which is widely used in health research. The Carstairs score is based on four census indicators: low social class, lack of car ownership, overcrowding and male unemployment. The scores have been described as a measure that reflects access to &#x02018;those goods and services, resources and amenities and of a physical environment which are customary in society&#x02019; [<xref ref-type="bibr" rid="CR39">39</xref>]. The scores therefore cannot be described as a measure of the extent of an individual&#x02019;s material wellbeing, but are rather a summary measure applied to populations contained within small geographic localities. Further adjustment for number of morbidities was made when polypharmacy was the characteristic of interest. Associations between numbers of morbidities and prescriptions were assessed using Spearman correlation coefficients. For the purposes of this analysis, a <italic>P</italic>&#x02009;&#x0003c;&#x02009;0.05 was deemed statistically significant. All analyses were carried out using IBM Statistical Package for the Social Sciences (SPSS) Statistics software (V21).</p></sec></sec><sec id="Sec6" sec-type="results"><title>Results</title><p>The analyses were based on 1,424,378 individuals (724,949 women) aged 18&#x000a0;years and over who were registered with a general practitioner. In total, 35,690 people (2.5%) had a diagnosis of stroke. As anticipated, the mean age of people in the stroke group (72.68&#x02009;&#x000b1;&#x02009;12.21) was higher than that of the controls 47.36&#x02009;&#x000b1;&#x02009;17.93). For the demographic characteristics for each group, see Additional file <xref rid="MOESM2" ref-type="media">2</xref>.</p><sec id="Sec7"><title>Co<italic>morbidities</italic></title><p>Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref> shows the number and percent of total morbidities, physical morbidities and mental health morbidities in the stroke and control groups, along with ORs for stroke in relation to these variables. Multimorbidity was common in stroke: of the study members with stroke, the percentage that had one or more additional morbidities present (94.2%) was almost twice that in the control group (48%) (OR adjusted for age, sex and deprivation 5.18; 95% CI 4.95 to 5.43). Disaggregating the data into type of morbidity revealed that physical morbidity was markedly more common in people with stroke (adjusted OR 4.50; 95% CI 4.31 to 4.68), and mental health morbidity was also more common but the relationship was less strong (adjusted OR 2.10; 95% CI 2.05 to 2.15). In terms of assessing whether these differences exist across different age groups, a sub-analysis for age groups 35&#x02013;44 years and 75+ years was performed (see Additional file <xref rid="MOESM3" ref-type="media">3</xref>). This indicated that differences were larger for the younger age group, and increased with the number of conditions (a similar picture was found for number of repeat prescriptions). However, the skewed distribution of stroke prevalence towards the oldest age groups make any assessment of differences by age problematic, owing to the small sample sizes in the youngest age groups.<table-wrap id="Tab1"><label>Table 1</label><caption><p> <bold>Stroke status and number of morbidities (N&#x02009;=&#x02009;1,424,378)</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th rowspan="2"/><th> <bold>Stroke N (%)</bold> </th><th> <bold>No stroke (%)</bold> </th><th rowspan="2"> <bold>Unadjusted OR (95% </bold> <bold>CI)</bold> ^{<bold>a</bold>} </th><th rowspan="2"> <bold>Age, gender and deprivation adjusted OR (95% </bold> <bold>CI)</bold> ^{<bold>a</bold>} </th></tr><tr valign="top"><th> <bold>35690 (100)</bold> </th><th> <bold>1388688 (100)</bold> </th></tr></thead><tbody><tr valign="top"><td>Total number of morbidities^b </td><td/><td/><td/><td/></tr><tr valign="top"><td>None</td><td>2053 (5.8)</td><td>721430 (52.0)</td><td>1</td><td>1</td></tr><tr valign="top"><td>One-three</td><td>17750 (49.7)</td><td>551295 (39.7)</td><td>11.31 (10.81 to 11.85)</td><td>4.35 (4.15 to 4.56)</td></tr><tr valign="top"><td>Four-six</td><td>12300 (34.5)</td><td>100500 (7.2)</td><td>43.01 (41.03 to 45.09)</td><td>8.59 (8.17 to 9.04)</td></tr><tr valign="top"><td>Seven or more</td><td>3587 (10.1)</td><td>15463 (1.1)</td><td>81.52 (77.04 to 86.26)</td><td>12.81 (12.05 to 13.61)</td></tr><tr valign="top"><td>Number of physical morbidities^b </td><td/><td/><td/><td/></tr><tr valign="top"><td>None</td><td>2769 (7.8)</td><td>800202 (57.6)</td><td>1</td><td>1</td></tr><tr valign="top"><td>One-three</td><td>20716 (58.0)</td><td>510846 (36.8)</td><td>11.72 (11.26 to 12.20)</td><td>4.03 (3.86 to 4.20)</td></tr><tr valign="top"><td>Four-six</td><td>10414 (29.2)</td><td>70709 (5.1)</td><td>42.56 (40.79 to 44.41)</td><td>7.32 (6.99 to 7.67)</td></tr><tr valign="top"><td>Seven or more</td><td>1791 (5.0)</td><td>6931 (0.5)</td><td>74.68 (70.05 to 79.61)</td><td>10.33 (9.64 to 11.05)</td></tr><tr valign="top"><td>Number of mental morbidities</td><td/><td/><td/><td/></tr><tr valign="top"><td>None</td><td>21961 (61.5)</td><td>1163095 (83.8)</td><td>1</td><td>1</td></tr><tr valign="top"><td>One-three</td><td>13533 (37.9)</td><td>223739 (16.1)</td><td>3.20 (3.13 to 3.27)</td><td>2.08 (2.04 to 2.13)</td></tr><tr valign="top"><td>Four or more</td><td>196 (0.5)</td><td>1854 (0.1)</td><td>5.60 (4.83 to 6.49)</td><td>3.56 (3.03 to 4.20)</td></tr></tbody></table><table-wrap-foot><p> ^aall p&#x02009;&#x0003c;&#x02009;0.001.</p><p> ^bexcluding stroke.</p></table-wrap-foot></table-wrap></p><p>The ten most frequent comorbidities present in people with a diagnosis of stroke were: hypertension (60.9%), coronary heart disease (29.5%), painful condition (21.9%), depression (20.7%), diabetes (18.8%), chronic kidney disease (14.3%), constipation (13.8%), atrial fibrillation (13.0%), thyroid disorders (11.9&#x000a0;%), and chronic obstructive pulmonary disease (11.9%). Prevalences of all morbidities are shown in supplementary material (see Additional files <xref rid="MOESM4" ref-type="media">4</xref> and <xref rid="MOESM5" ref-type="media">5</xref>).</p><p>Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref> displays the ORs (adjusted for age, sex and deprivation) for stroke in relation to the thrity one physical morbidities examined. The supplementary material (see Additional file <xref rid="MOESM4" ref-type="media">4</xref>) elaborates on this by showing both the unadjusted and adjusted ORs along with the crude prevalence of all physical morbidities in the stroke and control groups. In all, twenty eight of the thirty one physical morbidities examined were significantly more common in the stroke group, this was twenty seven after adjustment for potential confounding factors. For instance, epilepsy (adjusted OR 4.43; 95% CI 4.14 to 4.74), hypertension (adjusted OR 2.67; 95% CI 2.61 to 2.73), peripheral vascular disease (adjusted OR 2.47; 95% CI 2.37 to 2.58), AF (adjusted OR 2.44; 95% CI 2.36 to 2.53) and CHD (adjusted OR 2.06; 95% CI 2.01 to 2.11) were all more common in people experiencing a cerebrovascular disease event. By contrast, dyspepsia was markedly less common in the stroke group (adjusted OR 0.63; 95% CI 0.60 to 0.66). Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref> shows the ORs (adjusted for age, sex and deprivation) for stroke in relation to eight mental health morbidities. The unadjusted and adjusted ORs, along with the crude prevalence of all mental health morbidities in the stroke and stroke-free groups, are shown in supplementary material (see Additional file <xref rid="MOESM5" ref-type="media">5</xref>). In all, six of the eight mental health morbidities examined were significantly more common in the stroke group, and following adjustments, all eight mental health morbidities were significantly more common. These included drug and medication use problems (adjusted OR 2.34; 95% CI 2.25 to 2.43), depression (adjusted OR 2.09; 95% CI 2.03 to 2.15), alcohol problems (adjusted OR 2.05; 95% CI 1.96 to 2.15) and anxiety and stress (adjusted OR 1.61; 95% CI 1.55 to 1.66).<fig id="Fig1"><label>Figure 1</label><caption><p> <bold>Odds ratios (with 95%</bold> <bold>Cl) for physical morbidities in relation to stroke status (adjusted for age, sex, and deprivation).</bold> The stroke group comprised 35,690 people, and the stroke-free group comprised 1,388,688 people.</p></caption><graphic xlink:href="12916_2014_151_Fig1_HTML" id="MO1"/></fig><fig id="Fig2"><label>Figure 2</label><caption><p> <bold>Odds ratios (with 95% </bold> <bold>Cl) for mental health morbidities in relation to stroke status (adjusted for age, sex, and deprivation).</bold> The stroke group comprised 35,690 people, and the stroke-free group comprised 1,388,688 people.</p><p>Note: Drug related problems is any Read code which records psychoactive substance abuse which includes both drug misuse and prescription drug problems of multiple sources.</p></caption><graphic xlink:href="12916_2014_151_Fig2_HTML" id="MO2"/></fig></p></sec><sec id="Sec8"><title>Regular prescriptions</title><p>As anticipated, the number of regular prescriptions was significantly correlated with number of morbidities in the stroke (Spearman&#x02019;s &#x003c1;&#x02009;=&#x02009;0.58 <italic>P</italic>&#x02009;&#x0003c;&#x02009;0.001) and control (Spearman&#x02019;s &#x003c1;&#x02009;=&#x02009;0.75 <italic>P</italic>&#x02009;&#x0003c;&#x02009;0.001) groups. Table&#x000a0;<xref rid="Tab2" ref-type="table">2</xref> shows the number of repeat prescriptions in the stroke and control groups, and the ORs. Those with stroke were more likely than the controls to be on a repeat prescription (adjusted OR 4.53; 95% CI 4.33 to 4.74). In the stroke group, 12.6% had eleven or more repeat prescriptions compared with only 1.5% of the control group (OR adjusted for age, sex, deprivation and morbidity count 15.84; 95% CI 14.86 to 16.88).<table-wrap id="Tab2"><label>Table 2</label><caption><p> <bold>Stroke status and number of repeat medications (N&#x02009;=&#x02009;1,424,378)</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th rowspan="2"/><th> <bold>Stroke N (%)</bold> </th><th> <bold>No stroke N (%)</bold> </th><th rowspan="2"> <bold>Unadjusted OR (95% </bold> <bold>CI)</bold> ^{<bold>a</bold>} </th><th rowspan="2"> <bold>Age, gender and deprivation adjusted OR (95%</bold> <bold>CI)</bold> ^{<bold>a</bold>} </th><th rowspan="2"> <bold>Age, gender, deprivation and morbidity count adjusted OR (95%</bold> <bold>CI)</bold> ^{<bold>a</bold>} </th></tr><tr valign="top"><th> <bold>35690 (100)</bold> </th><th> <bold>1388688 (100)</bold> </th></tr></thead><tbody><tr valign="top"><td>Number of medications</td><td/><td/><td/><td/><td/></tr><tr valign="top"><td>None</td><td>2447 (6.9%)</td><td>863688 (62.2%)</td><td>1</td><td>1</td><td>1</td></tr><tr valign="top"><td>One-two</td><td>3038 (8.5%)</td><td>240721 (17.3%)</td><td>4.45 (4.22 to 4.70)</td><td>2.38 (2.26 to 2.52)</td><td>2.29 (2.17 to 2.42)</td></tr><tr valign="top"><td>Three-four</td><td>6566 (18.4%)</td><td>122518 (8.8%)</td><td>18.92 (18.05 to 19.82)</td><td>6.25 (5.95 to 6.57)</td><td>5.78 (5.49 to 6.08)</td></tr><tr valign="top"><td>Five-six</td><td>8185 (22.9%)</td><td>75512 (5.4%)</td><td>38.26 (36.55 to 40.05)</td><td>10.50 (9.99 to 11.03)</td><td>9.36 (8.89 to 9.86)</td></tr><tr valign="top"><td>Seven-eight</td><td>6721 (18.8%)</td><td>43344 (3.1%)</td><td>54.73 (52.20 to 57.38)</td><td>13.90 (13.20 to 14.63)</td><td>11.94 (11.29 to 12.62)</td></tr><tr valign="top"><td>Nine-ten</td><td>4219 (11.8%)</td><td>22536 (1.6%)</td><td>66.08 (62.76 to 69.57)</td><td>16.22 (15.34 to17.15)</td><td>13.44 (12.65 to 14.29)</td></tr><tr valign="top"><td>Eleven or more</td><td>4514(12.6%)</td><td>20369 (1.5%)</td><td>78.22 (74.32 to 82.32)</td><td>20.13 (19.05 to 21.27)</td><td>15.84 (14.86 to 16.88)</td></tr></tbody></table><table-wrap-foot><p> ^aall p&#x02009;&#x0003c;&#x02009;0.001.</p></table-wrap-foot></table-wrap></p></sec></sec><sec id="Sec9" sec-type="discussion"><title>Discussion</title><sec id="Sec10"><title>Summary of findings and implications</title><p>Analyses of a large, nationally representative sample of people in Scotland, a country with universal healthcare, showed that multimorbidity and polypharmacy were more common in people with a diagnosis of stroke. These findings are consistent with our knowledge that those with stroke are an elderly population with considerable cardiovascular disease risk [<xref ref-type="bibr" rid="CR44">44</xref>], for whom effective treatments are increasingly available to alleviate symptoms and address underlying causal factors [<xref ref-type="bibr" rid="CR45">45</xref>]. Diagnoses of most chronic conditions were more common in the stroke group, and this remained the case after adjustment for age, sex and deprivation. In our preliminary analyses (see Additional file <xref rid="MOESM2" ref-type="media">2</xref>), both age and deprivation were associated with stroke in the expected directions. This gives us confidence in the novel results presented herein.</p><p>Polypharmacy represents only one aspect of treatment burden, but is directly measurable, and may be a proxy measure of wider aspects of burden [<xref ref-type="bibr" rid="CR17">17</xref>,<xref ref-type="bibr" rid="CR18">18</xref>]. Multimorbidity is likely to increase treatment burden in several ways. First, as this study and others have shown, the number of medications increases with number of conditions [<xref ref-type="bibr" rid="CR20">20</xref>,<xref ref-type="bibr" rid="CR21">21</xref>]. Second, treatments may interact, leading to side effects [<xref ref-type="bibr" rid="CR5">5</xref>,<xref ref-type="bibr" rid="CR7">7</xref>,<xref ref-type="bibr" rid="CR46">46</xref>] and this has the potential to further increase the volume of work; for example, as new treatments are given to compensate for interactions [<xref ref-type="bibr" rid="CR47">47</xref>]. Third, multimorbidity is likely to increase healthcare contacts and affect the capacity of the individual to follow therapeutic regimens [<xref ref-type="bibr" rid="CR48">48</xref>]; for example, those with stroke and comorbid arthritis may find physiotherapy sessions more challenging [<xref ref-type="bibr" rid="CR49">49</xref>,<xref ref-type="bibr" rid="CR50">50</xref>]. Fourth, multimorbid patients who become overburdened, for example by complex medication regimens, may be less likely to adhere to therapies, leading to poor disease control and a further escalation of treatments by health professionals, further increasing treatment burden [<xref ref-type="bibr" rid="CR3">3</xref>,<xref ref-type="bibr" rid="CR9">9</xref>,<xref ref-type="bibr" rid="CR51">51</xref>]. While many pharmacological therapies may be beneficial for those with stroke, a key question is whether people with stroke have made informed decisions regarding whether or not to take so many medications, given their modest benefits. Although perceived treatment burden and capacity to cope with any given treatment burden will vary, we would recommend that patients with stroke are made aware of the relative benefits of their drugs, and are empowered to make their own decision whether to take them.</p><p>Acknowledging and addressing treatment burden in stroke, particularly for those with multimorbidity, may improve the patient experience, adherence to therapies, and health outcomes [<xref ref-type="bibr" rid="CR48">48</xref>]. Minimising unnecessary treatments, improving co-coordination of services and making care more patient-centred [<xref ref-type="bibr" rid="CR23">23</xref>] are likely to lessen treatment burden, but will necessitate changes from policy level down to the individual consultation [<xref ref-type="bibr" rid="CR3">3</xref>,<xref ref-type="bibr" rid="CR48">48</xref>,<xref ref-type="bibr" rid="CR52">52</xref>,<xref ref-type="bibr" rid="CR53">53</xref>]. Most stroke management guidelines fail to mention multimorbidity, or merely acknowledge the more common comorbidities briefly with a lack of practical advice for clinicians [<xref ref-type="bibr" rid="CR45">45</xref>,<xref ref-type="bibr" rid="CR54">54</xref>-<xref ref-type="bibr" rid="CR57">57</xref>]. We found only one stroke guideline that acknowledged the issue of polypharmacy, and again, detailed practical help was lacking [<xref ref-type="bibr" rid="CR56">56</xref>]. This issue has been gaining prominence [<xref ref-type="bibr" rid="CR58">58</xref>,<xref ref-type="bibr" rid="CR59">59</xref>]. Guidelines should be redesigned to take account of comorbidity and treatment burden; for example, by providing guidance on potential interactions from drug combinations commonly prescribed for those with stroke and multimorbidity and how to deal with the possible side effects or interactions that may arise [<xref ref-type="bibr" rid="CR47">47</xref>]. In the current study, 21.9% of people with stroke had a painful condition, 20.7% had depression and 13.0% had atrial fibrillation, increasing the risk of being prescribed non-steroidal anti-infammatory drugs (NSAIDs), anti-depressants, anti-platelet therapies and anti-coagulants concomitantly, which increases risk of adverse events, such as bleeding. Care pathways should be structured around the patient themselves, rather than the individual conditions, using a more generalist approach that considers issues such as multimorbidity as well as the individual&#x02019;s support network and financial resources [<xref ref-type="bibr" rid="CR9">9</xref>,<xref ref-type="bibr" rid="CR60">60</xref>,<xref ref-type="bibr" rid="CR61">61</xref>].</p></sec><sec id="Sec11"><title>Strengths and limitations</title><p>This analysis was undertaken using data from a large, nationally representative, primary care sample, and as far as we are aware, this is the first study on such a scale to examine multimorbidity and polypharmacy in stroke. This sample is representative of the Scottish population [<xref ref-type="bibr" rid="CR37">37</xref>]; however, it may not reflect experience in other countries and healthcare systems. The prevalence of stroke in this sample was similar to that shown in other studies [<xref ref-type="bibr" rid="CR44">44</xref>,<xref ref-type="bibr" rid="CR62">62</xref>], further validating the data; however, the data were collected for clinical rather than research purposes. No standard methods for measuring multimorbidity or polypharmacy exist, therefore a pragmatic approach was taken. We examined thirty nine long-term conditions, which is substantially more than in previous studies. The rationale for including the conditions examined and the rules for identifying the presence of each were described in detail by the team who previously collated the data [<xref ref-type="bibr" rid="CR1">1</xref>]. In addition, any medications bought over the counter or given from secondary care were not included. However, at the time of the analysis, prescriptions to people over sixty five years of age and to many people with chronic conditions were all free, with others being able to cap their out-of-pocket costs, thus suggesting a financial incentive to obtain medication via the primary care practice.</p><p>As this is a cross-sectional study, the data we have were taken from one particular point in time, and therefore no conclusions about temporality or causation can be made. The measure of comorbidity was unweighted, as the aim was to be descriptive rather than to assess outcomes. This was deemed to be the most appropriate method, and is similar to that used by others investigating the prevalence of multimorbidity [<xref ref-type="bibr" rid="CR1">1</xref>], but could be viewed as a limitation, especially as there may be a qualitative difference between the effects on perceived treatment burden of long-term conditions that produce regular symptoms (for example, heart failure) and those that are asymptomatic (for example, hypertension). We have no information about stroke severity, which is also a potential limitation. It should also be noted that due to the nature of the study, multiple analyses were carried out. Thus, the large numbers of cases and controls assessed in this study may have identified some associations that were statistically significant but not necessarily clinically significant; for example, for conditions such as cancer, glaucoma and asthma, which had ORs between 1.08 and 1.10 but were statistically significant with <italic>P</italic>&#x02009;&#x0003c;&#x02009;0.001.</p><p>Lastly, to explore treatment burden in stroke, this study examined multimorbidity and polypharmacy, however there are many more aspects of treatment burden still to be examined, such as clinic visits, continuity, coordination of care, and financial burden of therapies. The development of a patient-reported measure would enable a more detailed examination of treatment burden in stroke from the patient perspective.</p></sec></sec><sec id="Sec12" sec-type="conclusion"><title>Conclusion</title><p>In this study, we found that multimorbidity and polypharmacy were strikingly more common in those with stroke than those without. Polypharmacy can be thought of as a direct measure of one aspect of treatment burden, and we would suggest that people with stroke should be made aware of the relative benefits of their drugs so they can make informed decisions about therapeutic regimens. Both polypharmacy and multimorbidity are likely to be proxy markers for other aspects of treatment burden, as patients face the demands of managing multiple medications and conditions simultaneously. Clinical guidelines for stroke need to place greater emphasis on the management of multimorbidity, and further investigation of treatment burden in stroke is required to inform redesign of health services to improve patient outcomes.</p></sec>

Primary care treatment guidelines for skin infections in Europe: congruence with antimicrobial resistance found in commensal <italic>Staphylococcus aureus</italic> in the community

Could not extract abstract

<contrib contrib-type="author" corresp="yes"><name><surname>van Bijnen</surname><given-names>Evelien ME</given-names></name><address><email>e.vanbijnen@nivel.nl</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author"><name><surname>Paget</surname><given-names>W John</given-names></name><address><email>j.paget@nivel.nl</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author"><name><surname>den Heijer</surname><given-names>Casper DJ</given-names></name><address><email>casper.den.heijer@mumc.nl</email></address><xref ref-type="aff" rid="Aff2"/></contrib><contrib contrib-type="author"><name><surname>Stobberingh</surname><given-names>Ellen E</given-names></name><address><email>e.stobberingh@gmail.com</email></address><xref ref-type="aff" rid="Aff2"/></contrib><contrib contrib-type="author"><name><surname>Bruggeman</surname><given-names>Cathrien A</given-names></name><address><email>c.bruggeman@mumc.nl</email></address><xref ref-type="aff" rid="Aff2"/></contrib><contrib contrib-type="author"><name><surname>Schellevis</surname><given-names>Fran&#x000e7;ois G</given-names></name><address><email>f.schellevis@nivel.nl</email></address><xref ref-type="aff" rid="Aff1"/><xref ref-type="aff" rid="Aff3"/></contrib><contrib contrib-type="author"><collab>in collaboration with the APRES Study Team</collab></contrib><aff id="Aff1"><label/>Netherlands Institute for Health Services Research (NIVEL), Otterstraat 118-124, Utrecht, The Netherlands </aff><aff id="Aff2"><label/>Department of Medical Microbiology/School for Public Health and Primary Care (CAPHRI), Maastricht University Medical Centre, P. Debyelaan 25, Maastricht, The Netherlands </aff><aff id="Aff3"><label/>Department of General Practice and Elderly Care Medicine/EMGO Institute for Health and Care Research, VU University Medical Centre, De Boelelaan 1117, Amsterdam, The Netherlands </aff>

BMC Family Practice

<sec id="Sec1" sec-type="introduction"><title>Background</title><p>Antimicrobial resistance (AMR) has become an important public health threat across the globe during recent decades [<xref ref-type="bibr" rid="CR1">1</xref>&#x02013;<xref ref-type="bibr" rid="CR3">3</xref>]. The development of AMR is considered to be mainly driven by antibiotic use: exposure to antibiotics leads to the selection of resistant bacteria in the commensal microbiota [<xref ref-type="bibr" rid="CR4">4</xref>&#x02013;<xref ref-type="bibr" rid="CR6">6</xref>]. An important source of exposure is found in primary care as over 90% of all antibiotics for human medical use in Europe are prescribed in primary care [<xref ref-type="bibr" rid="CR5">5</xref>,<xref ref-type="bibr" rid="CR7">7</xref>]. Therefore, several studies have advocated cautious and appropriate prescribing of antibiotics to control the emergence of AMR [<xref ref-type="bibr" rid="CR1">1</xref>,<xref ref-type="bibr" rid="CR8">8</xref>]: empirical treatment with antibiotics should only take place if necessary and should ideally include appropriate agents which are effective against the most common pathogenic bacteria [<xref ref-type="bibr" rid="CR8">8</xref>].</p><p>An inappropriate antibiotic treatment will have several effects, in the first place for the patient: the effectiveness of the treatment will be limited. Secondly, unnecessary costs will occur for the health care system; and finally, the exposure to antibiotics could lead to a further increase of AMR [<xref ref-type="bibr" rid="CR1">1</xref>,<xref ref-type="bibr" rid="CR9">9</xref>,<xref ref-type="bibr" rid="CR10">10</xref>]. Several studies recommend the use of relevant AMR data when developing or revising primary care treatment guidelines for bacterial infections [<xref ref-type="bibr" rid="CR6">6</xref>,<xref ref-type="bibr" rid="CR8">8</xref>,<xref ref-type="bibr" rid="CR11">11</xref>]. However, since previous AMR studies have mainly obtained data from hospitalized populations with higher resistance levels [<xref ref-type="bibr" rid="CR6">6</xref>], primary care treatment guidelines might benefit by integrating AMR patterns from the community [<xref ref-type="bibr" rid="CR12">12</xref>,<xref ref-type="bibr" rid="CR13">13</xref>].</p><p><italic>S. aureus</italic> is a part of the commensal microbiota mainly manifesting as bacterial skin and soft tissue infections [<xref ref-type="bibr" rid="CR14">14</xref>,<xref ref-type="bibr" rid="CR15">15</xref>]. The incidence of these infections in primary care is relatively high, especially in children, hereby forming a considerable cause for antibiotic prescriptions [<xref ref-type="bibr" rid="CR14">14</xref>,<xref ref-type="bibr" rid="CR16">16</xref>,<xref ref-type="bibr" rid="CR17">17</xref>]. Traditionally, methicillin-resistant <italic>S. aureus</italic> (MRSA) was confined to hospitals and long-term-care facilities, but in the last decade MRSA infections have also appeared in healthy community-dwelling individuals [<xref ref-type="bibr" rid="CR18">18</xref>&#x02013;<xref ref-type="bibr" rid="CR21">21</xref>]. Several studies have established the importance of commensal microbiota as a natural reservoir of bacterial resistance, from which resistance can be acquired by pathogens [<xref ref-type="bibr" rid="CR22">22</xref>,<xref ref-type="bibr" rid="CR23">23</xref>]. By focusing on <italic>S. aureus</italic>, our study assessed the congruency of primary care treatment guidelines for skin infections with AMR data from the community, to optimize treatment effectiveness.</p></sec><sec id="Sec2" sec-type="materials|methods"><title>Methods</title><sec id="Sec3"><title>Study design</title><p>This study was part of the EC-funded APRES study, aimed at establishing the appropriateness of prescribing antibiotics in primary care in Europe, by collecting data on AMR in the community, antibiotic prescription behaviour and treatment guidelines in primary care. Nine countries across Europe participated in APRES, with varying patterns of antibiotic prescription rates [<xref ref-type="bibr" rid="CR5">5</xref>]: Austria, Belgium, Croatia, France, Hungary, the Netherlands, Spain, Sweden, and the United Kingdom. A detailed overview of the APRES study design and an analysis of the AMR results have been published elsewhere [<xref ref-type="bibr" rid="CR4">4</xref>,<xref ref-type="bibr" rid="CR24">24</xref>]. This paper relates the measured AMR patterns in the community to primary care treatment guidelines for skin infections and assesses their congruency.</p></sec><sec id="Sec4"><title>Study participants and AMR</title><p>In each of the nine countries, national GP networks selected 20 primary care practices representative of their total GP population. From each of these practices 200 nasal swabs from patients visiting the practice for non-infectious reasons were collected [<xref ref-type="bibr" rid="CR4">4</xref>,<xref ref-type="bibr" rid="CR24">24</xref>]. Previous studies [<xref ref-type="bibr" rid="CR25">25</xref>,<xref ref-type="bibr" rid="CR26">26</xref>] have shown carriage of <italic>S. aureus</italic> to be dynamic and occurring on multiple bodily sites. With the nares being a common site for <italic>S. aureus</italic> we assumed our sample to be representative of all carriage. In order to assess AMR levels in the commensal flora in the community (from which resistance can be acquired by pathogens), we excluded patients with known important risk factors for AMR: antibiotic use or hospitalisation in the past 3&#x000a0;months. Although <italic>S. aureus</italic> is not the sole pathogen causing skin infections, we selected it due to its impact on public health and relatively high nasal carriage rate [<xref ref-type="bibr" rid="CR25">25</xref>,<xref ref-type="bibr" rid="CR26">26</xref>]. After isolation of <italic>S. aureus</italic> in 8 national laboratories using standardised procedures, we determined in one central laboratory whether the isolates were resistant or susceptible to a range of commonly used antibiotics in primary care, using cut-off points from the Eucast guidelines [<xref ref-type="bibr" rid="CR4">4</xref>,<xref ref-type="bibr" rid="CR27">27</xref>].</p></sec><sec id="Sec5"><title>Treatment guidelines for skin infections</title><p>Coordinators of national GP networks in each participating country supplied the most commonly used and most recent primary care treatment guidelines for bacterial skin infections. With the exception of Croatia, all countries had issued national treatment guidelines for one or more bacterial skin infections. This resulted in a total of 13 national guidelines from 8 European countries (see Additional file <xref rid="MOESM1" ref-type="media">1</xref>: Table S1), from which we extracted the prescription recommendations. We focused the analysis on the antibiotic prescription recommendations for four common bacterial skin infections in primary care which are often caused by <italic>S. aureus</italic> [<xref ref-type="bibr" rid="CR13">13</xref>]: impetigo, cellulitis, folliculitis and furuncle. We have analysed the treatment recommendations for antibiotic therapy, distinguishing between first-choice recommendations and, if available, second-choice options. Since skin infections are common in children, we assessed the recommendations for children separately if this information was available.</p></sec><sec id="Sec6"><title>Data analysis</title><p>To assess the treatment guidelines issued on a national level, resistance levels for each antibiotic were aggregated to a national level by dividing the number of resistant <italic>S. aureus</italic> isolates per country by the total number of persons who carried a <italic>S. aureus</italic>. Separate rates were calculated for children (4-17 years old) and adults (18+), since treatment recommendations are often adapted for children. The recommended antibiotics in the treatment guideline were linked to the respective AMR levels in that country. Based on research regarding urinary tract infections, the antibiotic treatment recommendations were considered to be congruent if the resistance to the antibiotic did not exceed 20% [<xref ref-type="bibr" rid="CR28">28</xref>]. Carriership of <italic>S. aureus</italic> is linked to a higher risk of bacterial skin infection [<xref ref-type="bibr" rid="CR25">25</xref>,<xref ref-type="bibr" rid="CR29">29</xref>], however, evidence on the relationship between nasal <italic>S. aureus</italic> and pathogenic <italic>S. aureus</italic> isolated from skin infections is lacking. Therefore, in our current comparison we assumed that pathogenic <italic>S. aureus</italic> related to skin infections shows the same AMR patterns as nasal colonized <italic>S. aureus</italic>.</p><p>Not all antibiotics mentioned in the treatment guidelines were covered on a product level in the resistance testing of our study. When no AMR data for a specific antibiotic was available, we used two expert opinions (a medical microbiologist and pharmacist) to identify a similar antibiotic: e.g. since no clarithromycin resistance was tested we used data of azithromycin resistance (see Additional file <xref rid="MOESM2" ref-type="media">2</xref>: Table S2). Recent studies have shown that resistance to similar antibiotics can serve as a reliable indicator for the level of resistance to the original antibiotic [<xref ref-type="bibr" rid="CR30">30</xref>,<xref ref-type="bibr" rid="CR31">31</xref>].</p></sec></sec><sec id="Sec7" sec-type="results"><title>Results</title><p>Data were obtained from a total number of 32,206 swabs and in twenty-two percent (N = 6,956) <italic>S. aureus</italic> was present. After excluding patients for whom age was unknown 6037 (87,8%) adults (aged 18+) and 840 (12,2%) children (aged 4 to 17) were included in our study sample, of which 56% were female.</p><sec id="Sec8"><title>Prevalence of resistance</title><p>Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref> (adults) and Table&#x000a0;<xref rid="Tab2" ref-type="table">2</xref> (children) show the AMR levels of <italic>S. aureus</italic> for five antibiotics per country. The difference in resistance was high: on average, <italic>S. aureus</italic> showed almost no resistance to oxacillin (0.4%) while resistance to penicillin was high (73%). Resistance to topical antibiotics was low: averaging 0.4% for mupirocin and 2.8% for fusidic acid. The level of variation between countries was considerable, especially regarding AMR levels to azithromycin which ranged from 1.5% in Sweden to 16.9% in France. Sweden stood out with the lowest AMR levels to all but one antibiotic. For azithromycin and penicillin significant differences between adults and children were found in some countries, with the higher resistance levels found in children.<table-wrap id="Tab1"><label>Table 1</label><caption><p> <bold>Resistance rates of</bold> <bold><italic>S. aureus</italic></bold> <bold>isolates in nine European countries</bold> &#x02013; <bold>adults 18</bold>+</p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th/><th/><th/><th colspan="7"> <bold>Resistance rates</bold> <bold>(%)</bold> <bold>(95%</bold> <bold>confidence interval)</bold> </th></tr><tr valign="top"><th> <bold>Country</bold> </th><th> <bold>Swabs</bold> </th><th> <bold>Isolates of</bold> <bold><italic>S. aureus</italic></bold> </th><th> <bold>Azithromycin</bold> </th><th> <bold>Clindamycin</bold> </th><th> <bold>Erythromycin</bold> </th><th> <bold>Oxacillin</bold> </th><th> <bold>Penicillin</bold> </th><th> <bold>Fucidic acid</bold> </th><th> <bold>Mupirocin</bold> </th></tr></thead><tbody><tr valign="top"><td> <bold>Austria</bold> </td><td>3168</td><td>522 (16.5%)</td><td>12.8** (6.3-19.4)</td><td>11.1 (4.9 - 17.3)</td><td>12.6** (6.1-19.1)</td><td>1.5 (0-3.9)</td><td>64.4 (55-73.8)</td><td>1.0 (0-3.0)</td><td>0.2 (0-1.1)</td></tr><tr valign="top"><td> <bold>Belgium</bold> </td><td>2892</td><td>552 (19.1%)</td><td>16.3 (9.1-23.5)</td><td>14.4 (7.5-21.3)</td><td>16.3 (9.1-23.5)</td><td>2.2 (0-5.1)</td><td>72.5 (63.7-81.3)</td><td>3.4 (0-7.0)</td><td>0.4 (0-1.6)</td></tr><tr valign="top"><td> <bold>Croatia</bold> </td><td>3380</td><td>601 (17.8%)</td><td>5.8 (1.2-10.4)</td><td>5.5 (1-10)</td><td>5.8 (1.2-10.4)</td><td>2.3 (0-5.2)</td><td>75.4** (67-83.8)</td><td>0.2 (0-1.1)</td><td>0.7 (0-2.3)</td></tr><tr valign="top"><td> <bold>France</bold> </td><td>3536</td><td>777 (22.0%)</td><td>17.5 (10.1-25.0)</td><td>14.9* (7.9-21.9)</td><td>17.1 (9.7-24.5)</td><td>1.8 (0-4.4)</td><td>74.4 (65.8-83)</td><td>4.0 (0.2-7.8)</td><td>0.1 (0-0.7)</td></tr><tr valign="top"><td> <bold>Hungary</bold> </td><td>2883</td><td>359 (12.5%)</td><td>10.3** (4.3-16.3)</td><td>10.3** (4.3-16.3)</td><td>10.3** (4.3-16.3)</td><td>1.9 (0-4.6)</td><td>71.0** (62.1-79.9)</td><td>0.3 (0-1.4)</td><td>0.3 (0-1.4)</td></tr><tr valign="top"><td> <bold>NL</bold> </td><td>3491</td><td>947 (27.1%)</td><td>6.9 (1.9-11.9)</td><td>5.2 (0.8-9.6)</td><td>5.5 (1.0-10.0)</td><td>1.0 (0-3.0)</td><td>68.4 (59.3-77.5)</td><td>5.2 (0.8-9.6)</td><td>0</td></tr><tr valign="top"><td> <bold>Spain</bold> </td><td>3563</td><td>620 (17.5%)</td><td>11.5 (5.3-17.8)</td><td>9.5 (3.8-15.2)</td><td>11.0 (4.9-17.1)</td><td>1.3 (0-3.5)</td><td>86.0* (79.2-92.8)</td><td>1.1 (0-3.1)</td><td>1.6 (0-4.1)</td></tr><tr valign="top"><td> <bold>Sweden</bold> </td><td>2859</td><td>846 (29.6%)</td><td>1.3 (0-3.5)</td><td>11.8 (5.5-18.1)</td><td>13.0 (6.4-19.6)</td><td>0</td><td>64.3* (54.9-73.7)</td><td>1.9 (0-4.6)</td><td>0</td></tr><tr valign="top"><td> <bold>UK</bold> </td><td>3152</td><td>811 (25.7%)</td><td>8.6 (3.1-14.1)</td><td>7.5 (2.3-12.7)</td><td>8.9 (3.3-14.5)</td><td>1.6 (0-4.1)</td><td>73.4 (64.7-82.1)</td><td>7.8 (2.5-13.1)</td><td>0</td></tr></tbody></table><table-wrap-foot><p>**Significant difference with children under p&#x02009;&#x0003c;&#x02009;0.05.</p><p>*Significant difference with children under p&#x02009;&#x0003c;&#x02009;0.1.</p></table-wrap-foot></table-wrap><table-wrap id="Tab2"><label>Table 2</label><caption><p> <bold>Resistance rates of</bold> <bold><italic>S. aureus</italic></bold> <bold>isolates in nine European countries</bold> &#x02013; <bold>children</bold> &#x0003c;<bold>18</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th/><th/><th/><th colspan="7"> <bold>Resistance rates (%) (95% confidence intervals)</bold> </th></tr><tr valign="top"><th> <bold>Country</bold> </th><th> <bold>Swabs</bold> </th><th> <bold>Isolates of</bold> <bold><italic>S. aureus</italic></bold> </th><th> <bold>Azithromycin</bold> </th><th> <bold>Clindamycin</bold> </th><th> <bold>Erythromycin</bold> </th><th> <bold>Oxacillin</bold> </th><th> <bold>Penicillin</bold> </th><th> <bold>Fucidic acid</bold> </th><th> <bold>Mupirocin</bold> </th></tr></thead><tbody><tr valign="top"><td> <bold>Austria</bold> </td><td>111</td><td>23 (20.7%)</td><td>30.4** (21.4-39.4)</td><td>13.0 (6.4-19.6)</td><td>30.4** (21.4-39.4)</td><td>0</td><td>73.9 (65.3-82.5)</td><td>4.3 (0.3-8.3)</td><td>0</td></tr><tr valign="top"><td> <bold>Belgium</bold> </td><td>101</td><td>30 (22.9%)</td><td>16.7 (9.4-24)</td><td>16.6 (9.3-23.9)</td><td>16.6 (9.3-23.9)</td><td>0</td><td>63.3 (53.9-72.7)</td><td>0</td><td>0</td></tr><tr valign="top"><td> <bold>Croatia</bold> </td><td>562</td><td>152 (27.0%)</td><td>5.3 (0.9-9.7)</td><td>4.6 (0.5-8.7)</td><td>5.3 (0.9-9.7)</td><td>0.7 (0-2.3)</td><td>88.2** (81.9-94.5)</td><td>0</td><td>0</td></tr><tr valign="top"><td> <bold>France</bold> </td><td>309</td><td>94 (30.4%)</td><td>11.7 (5.4-18.0)</td><td>8.5 (3.0-14.0)</td><td>10.6 (4.6-16.6)</td><td>1.1 (0-3.1)</td><td>79.8 (71.9-87.7)</td><td>2.1 (0-4.9)</td><td>0</td></tr><tr valign="top"><td> <bold>Hungary</bold> </td><td>930</td><td>171 (18.4%)</td><td>16.4** (9.1-23.7)</td><td>16.4 (9.1-23.7)</td><td>16.4 (9.1-23.7)</td><td>0.6 (0-2.1)</td><td>86.0** (79.2-92.8)</td><td>0</td><td>0</td></tr><tr valign="top"><td> <bold>NL</bold> </td><td>323</td><td>119 (36.8%)</td><td>5.9 (1.3-10.5)</td><td>3.4 (0-7)</td><td>4.2 (0.3-8.1)</td><td>0</td><td>73.1 (64.4-81.8)</td><td>0.5 (0-1.9)</td><td>0</td></tr><tr valign="top"><td> <bold>Spain</bold> </td><td>427</td><td>146 (34.4%)</td><td>12.3 (5.9-18.7)</td><td>10.3 (4.3-16.3)</td><td>12.3 (5.9-18.7)</td><td>0.7 (0-2.3)</td><td>91.8* (86.4-97.2)</td><td>0</td><td>3.4 (0 -7.0)</td></tr><tr valign="top"><td> <bold>Sweden</bold> </td><td>345</td><td>104 (30.1%)</td><td>2.9 (0-6.2)</td><td>2.9 (0-6.2)</td><td>2.9 (0-6.2)</td><td>0</td><td>73.1* (64.4-81.8)</td><td>2.0 (0-4.7)</td><td>0</td></tr><tr valign="top"><td> <bold>UK</bold> </td><td/><td/><td colspan="7">No children in study sample due to ethical considerations</td></tr></tbody></table><table-wrap-foot><p>**Significant difference with adults under p&#x02009;&#x0003c;&#x02009;0.05.</p><p>*Significant difference with adults under p&#x02009;&#x0003c;&#x02009;0.1.</p></table-wrap-foot></table-wrap></p></sec><sec id="Sec9"><title>Recommendations in treatment guidelines</title><p>Some guidelines were not complete in their coverage of all four infections for both adults and children. Since folliculitis and furuncle are related infections, they were often discussed together in guidelines and the same recommendations were applied. Overall, the first-choice recommendations for skin infections were very consistent across Europe. Almost all recommended first-choice antibiotics were of the B-lactam class, mainly flucloxacillin. Austria and Sweden also recommended cephalosporins for impetigo, folliculitis and furuncle; while in the Netherlands macrolides were preferred for cellulitis in children. For the treatment of impetigo, all guidelines recommended to first start treatment with a topical antibiotic (most often fusidic acid). The second-choice recommendations consisted of a wider range of antibiotics. Most countries used the same antibiotic for adults and children, with an adjusted dosage for children.</p></sec><sec id="Sec10"><title>Congruency of first- and second-choice antibiotics with AMR patterns</title><p>One can assess the congruency of the recommendations with AMR patterns by determining whether the resistance to the antibiotics is &#x0003c;20% (Tables&#x000a0;<xref rid="Tab3" ref-type="table">3</xref> and <xref rid="Tab4" ref-type="table">4</xref>) [<xref ref-type="bibr" rid="CR28">28</xref>]. As previously mentioned low resistance to the topical agents was found in <italic>S. aureus</italic>, so the topical treatment recommendations were all congruent with AMR in the community.<table-wrap id="Tab3"><label>Table 3</label><caption><p> <bold>Congruency of treatment recommendations for skin infections in adults with national commensal</bold> <bold><italic>S. aureus</italic></bold> <bold>resistance rates</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th/><th> <bold>Topical AB</bold> </th><th> <bold>Resistance rate</bold>*</th><th> <bold>First choice systemic AB</bold> </th><th> <bold>Resistance rate</bold>*</th><th> <bold>Second choice systemic AB</bold> </th><th> <bold>Resistance rate</bold>*</th></tr></thead><tbody><tr valign="top"><td colspan="7">Impetigo</td></tr><tr valign="top"><td> <bold>Austria</bold> </td><td>Fusidic acid</td><td>1.0</td><td>Cephalosporin</td><td>No data</td><td>Amoxicillin&#x02009;+&#x02009;Clavulanic acid</td><td>1.5</td></tr><tr valign="top"><td> <bold>Belgium</bold> </td><td>Fusidic acid</td><td>3.4</td><td>Flucloxacillin</td><td>2.2</td><td>Clarithromycin</td><td>16.3</td></tr><tr valign="top"><td> <bold>France</bold> </td><td>Fusidic acid</td><td>4.0</td><td colspan="4">No specific advice</td></tr><tr valign="top"><td> <bold>Hungary</bold> </td><td colspan="6">No guideline</td></tr><tr valign="top"><td> <bold>Netherlands</bold> </td><td>Fusidic acid</td><td>5.2</td><td>Flucloxacillin</td><td>1.0</td><td>Azithromycin</td><td>6.9</td></tr><tr valign="top"><td> <bold>Spain</bold> </td><td>Mupirocin</td><td>1.6</td><td>Penicillin (IM)/Cloxacillin</td><td> <bold>86.0</bold> </td><td>Clindamycin</td><td>9.5</td></tr><tr valign="top"><td/><td/><td/><td/><td>1.3</td><td/><td/></tr><tr valign="top"><td> <bold>Sweden</bold> </td><td>Retapamulin</td><td>0</td><td>Flucloxacillin</td><td>0</td><td>Cefadroxil</td><td>No data</td></tr><tr valign="top"><td> <bold>UK</bold> </td><td>Fusidic acid</td><td>7.8</td><td>Flucloxacillin</td><td>1.6</td><td>Clarithromycin</td><td>8.9</td></tr><tr valign="top"><td colspan="7">Cellulitis</td></tr><tr valign="top"><td> <bold>Austria</bold> </td><td/><td/><td>Penicillin (parenteral)</td><td> <bold>64.4</bold> </td><td>Clindamycin</td><td>11.1</td></tr><tr valign="top"><td> <bold>Belgium</bold> </td><td/><td/><td>Flucloxacillin</td><td>2.2</td><td>Clindamycin</td><td>14.4</td></tr><tr valign="top"><td> <bold>France</bold> </td><td colspan="6">No guideline</td></tr><tr valign="top"><td> <bold>Hungary</bold> </td><td colspan="6">No guideline</td></tr><tr valign="top"><td> <bold>Netherlands</bold> </td><td/><td/><td>Flucloxacillin</td><td>1.0</td><td>Claritromycin</td><td>5.5</td></tr><tr valign="top"><td> <bold>Spain</bold> </td><td/><td/><td>Cloxacillin</td><td>1.3</td><td>Amoxicillin&#x02009;+&#x02009;Clavulanic acid</td><td>1.3</td></tr><tr valign="top"><td> <bold>Sweden</bold> </td><td colspan="6">No guideline</td></tr><tr valign="top"><td> <bold>UK</bold> </td><td/><td/><td>Flucloxacillin</td><td>1.6</td><td>Ery/Clarithromycin</td><td>8.9</td></tr><tr valign="top"><td colspan="7">Folliculitis and Furuncle</td></tr><tr valign="top"><td> <bold>Austria</bold> </td><td/><td/><td>Cephalosporin</td><td>No data</td><td>Amoxicillin + Clavulanic acid</td><td>1.5</td></tr><tr valign="top"><td> <bold>Belgium</bold> </td><td colspan="6">No guideline</td></tr><tr valign="top"><td> <bold>France</bold> </td><td colspan="6">No guideline</td></tr><tr valign="top"><td> <bold>Hungary</bold> </td><td colspan="6">No guideline</td></tr><tr valign="top"><td> <bold>Netherlands</bold> </td><td/><td/><td>Flucloxacillin</td><td>1.0</td><td colspan="2">No second choice</td></tr><tr valign="top"><td> <bold>Spain</bold> </td><td/><td/><td>Cloxacillin</td><td>1.3</td><td colspan="2">No second choice</td></tr><tr valign="top"><td> <bold>Sweden</bold> </td><td/><td/><td>Flucloxacillin</td><td>0</td><td>Cefadroxil</td><td>No data</td></tr><tr valign="top"><td> <bold>UK</bold> </td><td/><td/><td>Flucloxacillin</td><td>1.6</td><td>Ery/Clarithromycin</td><td>8.9</td></tr></tbody></table><table-wrap-foot><p>*A recommendation is congruent if the resistance rate in <italic>S. aureus</italic> to that antibiotic is &#x0003c;20%. Data in bold indicate a resistance rate of &#x0003e;20%.</p></table-wrap-foot></table-wrap><table-wrap id="Tab4"><label>Table 4</label><caption><p> <bold>Congruency of treatment recommendations for skin infections in children with national commensal</bold> <bold><italic>S. aureus</italic></bold> <bold>resistance rates</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th/><th> <bold>Topical AB</bold> </th><th> <bold>Resistance rate</bold>*</th><th> <bold>First choice systemic AB</bold> </th><th> <bold>Resistance rate</bold>*</th><th> <bold>Second choice systemic AB</bold> </th><th> <bold>Resistance rate</bold>*</th></tr></thead><tbody><tr valign="top"><td colspan="7">Impetigo</td></tr><tr valign="top"><td> <bold>Austria</bold> </td><td>Fusidic acid</td><td>4.3</td><td>Cephalosporin</td><td>No data</td><td>Amoxicillin&#x02009;+&#x02009;clavulanic acid</td><td>0</td></tr><tr valign="top"><td> <bold>Belgium</bold> </td><td>Fusidic acid</td><td>0</td><td>Flucloxacillin</td><td>0</td><td>Clarithromycin</td><td>16.6</td></tr><tr valign="top"><td> <bold>France</bold> </td><td colspan="6">No guideline</td></tr><tr valign="top"><td> <bold>Hungary</bold> </td><td colspan="6">No guideline</td></tr><tr valign="top"><td> <bold>Netherlands</bold> </td><td>Fusidic acid</td><td>0.5</td><td>Flucloxacillin</td><td>0</td><td>Azithromycin</td><td>5.9</td></tr><tr valign="top"><td> <bold>Spain</bold> </td><td colspan="6">No guideline</td></tr><tr valign="top"><td> <bold>Sweden</bold> </td><td>Retapamulin</td><td>0</td><td>Cefadroxil</td><td>No data</td><td>Flucloxacillin</td><td>0</td></tr><tr valign="top"><td> <bold>UK</bold> </td><td>Fusidic acid</td><td>No data</td><td>Flucloxacillin</td><td>No data</td><td>Clarithromycin</td><td>No data</td></tr><tr valign="top"><td colspan="7">Cellulitis</td></tr><tr valign="top"><td> <bold>Austria</bold> </td><td colspan="6">No guideline</td></tr><tr valign="top"><td> <bold>Belgium</bold> </td><td/><td/><td>Flucloxacillin</td><td>0</td><td colspan="2">No second choice</td></tr><tr valign="top"><td> <bold>France</bold> </td><td colspan="6">No guideline</td></tr><tr valign="top"><td> <bold>Hungary</bold> </td><td colspan="6">No guideline</td></tr><tr valign="top"><td> <bold>Netherlands</bold> </td><td/><td/><td>Clarithromycin</td><td>4.2</td><td>Azithromycin</td><td>5.9</td></tr><tr valign="top"><td> <bold>Spain</bold> </td><td colspan="6">No guideline</td></tr><tr valign="top"><td> <bold>Sweden</bold> </td><td colspan="6">No guideline</td></tr><tr valign="top"><td> <bold>UK</bold> </td><td/><td/><td>Flucloxacillin</td><td>No data</td><td>Ery/Clarithromycin</td><td>No data</td></tr><tr valign="top"><td colspan="7">Folliculitis and Furuncle</td></tr><tr valign="top"><td> <bold>Austria</bold> </td><td/><td/><td>Cephalosporin</td><td>No data</td><td>Amoxicillin + Clavulanic acid</td><td>0</td></tr><tr valign="top"><td> <bold>Belgium</bold> </td><td colspan="6">No guideline</td></tr><tr valign="top"><td> <bold>France</bold> </td><td colspan="6">No guideline</td></tr><tr valign="top"><td> <bold>Hungary</bold> </td><td colspan="6">No guideline</td></tr><tr valign="top"><td> <bold>Netherlands</bold> </td><td colspan="6">No guideline</td></tr><tr valign="top"><td> <bold>Spain</bold> </td><td colspan="6">No guideline</td></tr><tr valign="top"><td> <bold>Sweden</bold> </td><td colspan="6">No guideline</td></tr><tr valign="top"><td> <bold>UK</bold> </td><td/><td/><td>Flucloxacillin</td><td>No data</td><td>Ery/Clarithromycin</td><td>No data</td></tr></tbody></table><table-wrap-foot><p>*A recommendation is congruent if the resistance rate in <italic>S. aureus</italic> to that antibiotic is &#x0003c;20%.</p></table-wrap-foot></table-wrap></p><sec id="Sec11"><title>Adults</title><p>All first choice recommendations for oral treatment were congruent with the AMR patterns. In Austria (for cellulitis and erysipelas) and Spain (for impetigo) a parenteral treatment with penicillin was advised, which was not congruent with the high penicillin resistance rates found in <italic>S. aureus</italic>. The second-choice antibiotic treatment recommendations were also congruent, with measured resistance levels of &#x0003c;20%, although for Belgium some recommended antibiotics exceeded 15% resistance. We found that all recommendations in the Swedish guidelines concerned antibiotics with an AMR level of 0%.</p></sec><sec id="Sec12"><title>Children</title><p>Only oral therapy was advised for children, in most cases consisting of the same antibiotic that is used for adults (flucloxacillin) but with adjusted dosages. All recommended antibiotics showed a resistance level of &#x0003c;20% and were therefore congruent.</p></sec></sec></sec><sec id="Sec13" sec-type="discussion"><title>Discussion</title><p>This study assessed the congruency of primary care treatment guidelines for bacterial skin infections with nasal AMR levels of <italic>S. aureus</italic> in the community in nine European countries.</p><sec id="Sec14"><title>Congruency of recommendations</title><p>To assess the congruency of recommendations we used a threshold of 20%: antibiotics to which <italic>S. aureus</italic> has resistance rates of &#x0003c;20% are considered congruent [<xref ref-type="bibr" rid="CR28">28</xref>]. Our study showed that most of the first- and second-choice recommendations in the treatment guidelines were congruent with AMR patterns in nasal <italic>S. aureus</italic> in the community, except for two recommendations for penicillin. Azithromycin was appropriate in the Netherlands, but the relatively high resistance rates in other countries (up to 30%) warrant a cautious use of this antibiotic for skin infections.</p><p>Given the resistance levels to penicillin in nasal <italic>S. aureus</italic> in the community, our findings suggest that it should not be used as a first- or second-choice antibiotic for <italic>S. aureus</italic> infections in primary care. Most guidelines for skin infections that we assessed were already congruent with this finding as they did not recommend the use of penicillin, except for two first-choice recommendations for penicillin in Austria and Spain. The penicillin recommendation for Austria was also used for erysipelas, which is often caused by a streptococcus. The same was true for non-bullous impetigo, for which penicillin was recommended in Spain [<xref ref-type="bibr" rid="CR32">32</xref>]. Literature regarding AMR levels in streptococci indicates a high susceptibility for penicillin [<xref ref-type="bibr" rid="CR33">33</xref>], implying these recommendations might be congruent as well in Austria and Spain.</p></sec><sec id="Sec15"><title>Strengths and limitations</title><p>The strength of our study is the broad scope of the data from nine countries across Europe (North, South, East and West), with a high variation in antibiotic use [<xref ref-type="bibr" rid="CR5">5</xref>]. Our study is also unique as it assesses the congruency of treatment guidelines for <italic>S. aureus</italic> skin infections in primary care based on the prevalence of antibiotic resistance patterns. The treatment guidelines are issued nationally and have been supplied by experts who are aware of the most frequently used guidelines in their countries. Our study is complete by covering both recommendations for adults and for children.</p><p>A previous paper presented information on the dosage and duration of the treatment recommendations. Although the relationship between certain dosage-duration regimes and the development of resistance is not fully clear, it is noteworthy that the treatment guidelines from Sweden recommend higher dosages, while at the same time low AMR was observed [<xref ref-type="bibr" rid="CR34">34</xref>].</p><p>One limitation of our study is that although it tested resistance to a wide range of antibiotics, not all antibiotic recommendations in the guidelines were covered on a one-on-one basis and in some cases we had to use the prevalence of resistance to a similar antibiotic [<xref ref-type="bibr" rid="CR30">30</xref>,<xref ref-type="bibr" rid="CR31">31</xref>]. Also, since we excluded patients with current infections or risk factors for AMR (antibiotic use or hospitalisation in the past 3&#x000a0;months), the level of AMR might be an underestimation for the total population in the community. However, given that antibiotic resistance can linger for up to one year [<xref ref-type="bibr" rid="CR6">6</xref>], we assume our sample to be a good approximation of primary care patients.</p><p>Another limitation is the assumption we made that the AMR patterns found in nasal <italic>S. aureus</italic> are similar to those of pathogenic <italic>S. aureus</italic> found in SSTIs. To our knowledge, this relationship has not been conclusively studied, and future research might be able to fill this knowledge gap.</p><p>Our study is also limited in its choice of pathogen: we focussed on <italic>S. aureus</italic> due to its relatively high prevalence and impact on public health but skin infections can also be caused by a <italic>Streptococcus</italic> bacteria which may have other AMR patterns and would be relevant to also consider in treatment guidelines. However, since the main pathogen for these skin infections is <italic>S. aureus</italic> we emphasize its importance and recommend that resistance patterns of this pathogen are taken into account when updating or developing treatment guidelines for skin infections. Finally, although our study uses aggregated data, possible regional differences in AMR patterns of pathogens could also be integrated into empiric treatment guidelines.</p></sec><sec id="Sec16"><title>Implications for primary care</title><p>Most AMR studies present data from non-community settings (e.g. the hospital setting) [<xref ref-type="bibr" rid="CR34">34</xref>] and there is limited data on antibiotic resistance in the community. The prevalence of resistance of <italic>S. aureus</italic> we found in primary care is lower than the levels reported in hospitals [<xref ref-type="bibr" rid="CR4">4</xref>,<xref ref-type="bibr" rid="CR35">35</xref>] and we recommend that recent national AMR data from the community should be taken into account to create more effective and evidence-based treatment guidelines for primary care. In such initiatives other factors affecting evidence-based practice, such as the implementation process or adherence to guidelines, should also be incorporated [<xref ref-type="bibr" rid="CR36">36</xref>]. Evidence-based guidelines are, however, a first step to control the development of antibiotic resistance.</p></sec></sec><sec id="Sec17" sec-type="conclusion"><title>Conclusions</title><p>Our comparison of primary care treatment guidelines with AMR patterns of commensal <italic>S. aureus</italic> in the community showed that not all European countries have developed national guidelines for the treatment of common skin infections in primary care and emphasizes the need to develop treatment guidelines in these countries. The first- and second-choice recommendations in the available treatment guidelines proved to be congruent with the national AMR patterns found in nasal colonized <italic>S. aureus</italic>: almost all recommendations concerned antibiotics to which <italic>S. aureus</italic> had low resistance levels (&#x0003c;20%). Given the high resistance to penicillin that has been demonstrated for commensal <italic>S. aureus</italic>, we recommend that this antibiotic should not be used in primary care treatment of <italic>S. aureus</italic> related bacterial skin infections. Based on the variation in antimicrobial resistance levels between countries, age groups and health care settings, national data regarding antimicrobial resistance in the community should be taken into account when updating or developing primary care treatment guidelines.</p></sec>

Barriers to the implementation of mobile phone reminders in pediatric HIV care: a pre-trial analysis of the Cameroonian MORE CARE study

Could not extract abstract

<contrib contrib-type="author" corresp="yes"><name><surname>Bigna</surname><given-names>Jean Joel R</given-names></name><address><email>bignarimjj@yahoo.fr</email></address><xref ref-type="aff" rid="Aff1"/><xref ref-type="aff" rid="Aff2"/><xref ref-type="aff" rid="Aff3"/></contrib><contrib contrib-type="author"><name><surname>Noubiap</surname><given-names>Jean Jacques N</given-names></name><address><email>noubiapjj@yahoo.fr</email></address><xref ref-type="aff" rid="Aff4"/></contrib><contrib contrib-type="author"><name><surname>Plottel</surname><given-names>Claudia S</given-names></name><address><email>claudia.plottel@nyumc.org</email></address><xref ref-type="aff" rid="Aff5"/><xref ref-type="aff" rid="Aff6"/></contrib><contrib contrib-type="author"><name><surname>Kouanfack</surname><given-names>Charles</given-names></name><address><email>charleskouanfack@yahoo.fr</email></address><xref ref-type="aff" rid="Aff1"/><xref ref-type="aff" rid="Aff7"/></contrib><contrib contrib-type="author"><name><surname>Koulla-Shiro</surname><given-names>Sinata</given-names></name><address><email>koullasinata@yahoo.fr</email></address><xref ref-type="aff" rid="Aff1"/><xref ref-type="aff" rid="Aff8"/></contrib><aff id="Aff1"><label/>Faculty of Medicine and Biomedical Sciences, University of Yaound&#x000e9; 1, Yaound&#x000e9;, Cameroon </aff><aff id="Aff2"><label/>Faculty of Medicine, University of Montpellier 1, Montpellier, France </aff><aff id="Aff3"><label/>Preventing Mother to Child Transmission Unit, Goulfey District Hospital, Goulfey, Cameroon </aff><aff id="Aff4"><label/>Internal Medicine Unit, Ed&#x000e9;a Regional Hospital, Ed&#x000e9;a, Cameroon </aff><aff id="Aff5"><label/>Department of Medicine, New York University Langone Medical Center, New York, USA </aff><aff id="Aff6"><label/>Department of Medicine, Division of Translational Medicine, New York University School of Medicine, New York, USA </aff><aff id="Aff7"><label/>Accredited Treatment Centre, Yaound&#x000e9; Central Hospital, Yaound&#x000e9;, Cameroon </aff><aff id="Aff8"><label/>Infectious Disease Unit, Yaound&#x000e9; Central Hospital, Yaound&#x000e9;, Cameroon </aff>

BMC Health Services Research

<sec id="Sec1" sec-type="introduction"><title>Background</title><p>The Global Observatory for eHealth defines mobile health (mHealth) as medical and public health practice supported by mobile devices, such as mobile phones, patient monitoring devices, personal digital assistants, and other wireless devices [<xref ref-type="bibr" rid="CR1">1</xref>]. The United Nations Joint Program on HIV/AIDS (UNAIDS) encourages the use of mHealth in addressing human immunodeficiency virus (HIV) related illnesses and treatments in resource-limited settings [<xref ref-type="bibr" rid="CR2">2</xref>,<xref ref-type="bibr" rid="CR3">3</xref>]. According to the International Telecommunications Union, the number of mobile-broadband subscriptions has reached 2.3 billion worldwide in 2014, with 55% of subscriptions based in developing countries. Africa leads in mobile-broadband growth [<xref ref-type="bibr" rid="CR4">4</xref>]. MHealth is increasingly recognized as an effective adjunct to HIV control measures [<xref ref-type="bibr" rid="CR5">5</xref>,<xref ref-type="bibr" rid="CR6">6</xref>]. Cameroon, a sub-Saharan African country, has seen the proportion of mobile phone users increase rapidly [<xref ref-type="bibr" rid="CR4">4</xref>,<xref ref-type="bibr" rid="CR7">7</xref>] as elsewhere in Africa [<xref ref-type="bibr" rid="CR8">8</xref>], with 60% of the Cameroonian population owning a mobile phone in 2012 [<xref ref-type="bibr" rid="CR9">9</xref>]. The use of mobile phones has shown effectiveness in health related behavior change, in screening campaigns, and as a supportive tool in treatment, diagnosis, and data collection [<xref ref-type="bibr" rid="CR3">3</xref>,<xref ref-type="bibr" rid="CR6">6</xref>,<xref ref-type="bibr" rid="CR10">10</xref>&#x02013;<xref ref-type="bibr" rid="CR23">23</xref>]. Challenges to the implementation of mHealth include lack of mobile phone ownership and an inability to communicate in national official languages (NOL) both via mobile phone call (oral) and mobile phone text message (SMS) (written). In 2011, about two-third of the adult population in Cameroon were literate [<xref ref-type="bibr" rid="CR8">8</xref>]. An individual&#x02019;s unwillingness to use a mobile phone for health-related communications represents another barrier to acceptance of mHealth. In certain resource-limited settings adherence to mHealth has been shown to be high, despite potential obstacles [<xref ref-type="bibr" rid="CR6">6</xref>,<xref ref-type="bibr" rid="CR24">24</xref>&#x02013;<xref ref-type="bibr" rid="CR26">26</xref>].</p><p>Notwithstanding the proven benefits of the use of mobile phones in the fight against HIV, challenges to implementation remain. Before adopting mobile technologies into the clinical arena, it is imperative to know exactly what e-health architecture to employ within a health system [<xref ref-type="bibr" rid="CR18">18</xref>]. In Cameroon, the challenges that need to be addressed to effectively and efficiently implement SMS and mobile phone calls in pediatric HIV care remain undefined. In the recently reported MORE CARE (Mobile Reminders for Cameroonian children Requiring HIV treatment) study, we investigated the efficacy and efficiency of mobile phone appointment reminders on the attendance of HIV-exposed or infected children at their previously scheduled follow-up medical appointments [<xref ref-type="bibr" rid="CR27">27</xref>]. The aim of the present study was to evaluate three factors as potential obstacles to the implementation of using mobile phones (SMS and calls) as medical appointment reminder tools for HIV-exposed and HIV-infected children in Cameroon. More specifically, we sought to investigate the extent of mobile phone non-ownership, the inability to communicate in a NOL, and the refusal to receive mobile phone reminders as variables influencing the acceptance of mobile phones as part of mHealth.</p></sec><sec id="Sec2" sec-type="materials|methods"><title>Methods</title><sec id="Sec3"><title>Ethical consideration</title><p>The study was approved by the Faculty of Medicine and Biomedical Sciences, University of Yaound&#x000e9; 1. We obtained ethical clearance from the National Research Ethics Committee for Human Health of Cameroon (Ethical approval N&#x000b0; 2013/03/232/L/CNERSH/SP). All patients included in the study gave their verbal and written consent.</p></sec><sec id="Sec4"><title>Study design</title><p>This is a pre-trial cross-sectional study of adults included in the MORE CARE trial and addresses mobile phone ownership, an ability to communicate in NOL, and a willingness to receive health-related mobile phone appointment reminders. MORE CARE has been described in detail previously [<xref ref-type="bibr" rid="CR23">23</xref>,<xref ref-type="bibr" rid="CR27">27</xref>].</p></sec><sec id="Sec5"><title>Settings and participants</title><p>Potential participants in the MORE CARE study were enrolled in an urban, a semi-urban, and a rural setting at the Essos National Hospital Insurance Fund, the Kouss&#x000e9;ri Annex Regional Hospital, and the Goulfey District Hospital respectively [<xref ref-type="bibr" rid="CR23">23</xref>,<xref ref-type="bibr" rid="CR27">27</xref>]. Adults aged 18&#x000a0;years or older who accompanied a child to a medical appointment for HIV-related care were invited to enroll in the MORE CARE trial. Only one caregiver present at the time of recruitment was identified for each child. Although that adult caregiver was responsible for ensuring the child&#x02019;s attendance at their next scheduled medical appointment, we did not require that the adult accompanying the child at that follow-up visit be the same adult since MORE CARE was designed to examine each child&#x02019;s presence or absence at their follow-up visit for HIV care.</p></sec><sec id="Sec6"><title>Appointment reminders in the MORE CARE study</title><p>We briefly report here the MORE CARE mHealth appointment reminder methods, which are further detailed in the published study protocol [<xref ref-type="bibr" rid="CR27">27</xref>]. The text message sent to caregiver participants was in French or in English (national official languages in Cameroon), and was edited manually. It included the physician's name and the date, time, and location of the appointment. In order to maintain confidentiality and to protect caregiver and patient privacy, the message did not contain information concerning the health status and name of the child, nor the name of the accompanist. The message was sent and each voice phone call placed 48 to 72&#x000a0;hours before the previously scheduled appointment. The content of phone calls was the same as that of the SMS. The communication during phone call was in English or in French, based on the caregiver&#x02019;s stated preference.</p></sec><sec id="Sec7"><title>Variables</title><p>We collected the age and sex of caregivers and of the children they accompanied. Obstacles to mHealth appointment reminders investigated were:<list list-type="order"><list-item><p>The non-possession of mobile phone; assessed by asking a potential participant the question, &#x0201c;Do you own a mobile phone?&#x0201d; The answer could only be &#x0201c;Yes&#x0201d; or &#x0201c;No&#x0201d;.</p></list-item><list-item><p>The inability to understand and read English or French (SMS communication); assessed by the potential participant&#x02019;s inability to read and/or comprehend the informed consent form.</p></list-item><list-item><p>The inability to understand and speak English or French (voice phone call communication); assessed by the potential participant&#x02019;s replies to the questions, &#x0201c;What is your name?&#x0201d;, &#x0201c;How old are you?&#x0201d;, and &#x0201c;Where do you live?&#x0201d;, &#x0201c;How did you spend your last holidays?&#x0201d;, and the potential participant&#x02019;s ability to have a brief one-on-one discussion about certain aspects of HIV infection.</p></list-item><list-item><p>The refusal to receive a SMS or a phone call as follow-up medical appointment reminders; assessed by asking the potential participant directly if he or she wished to receive a reminder via their mobile phone, or not.</p></list-item></list></p></sec><sec id="Sec8"><title>Analysis plan</title><p>Data was coded, entered, and analyzed using the Statistical Package for Social Science (SPSS) version 20.0 for Windows (IBM Corp. Released 2011. IBM SPSS Statistics for Windows, Version 20.0. Armonk, NY: IBM Corp.). We described continuous variables using means with standard deviations, and binary variables using their frequencies and percentages. The chi-square test or its equivalents was used to compare binary variables. All statistical tests were performed using two-sided tests at a 0.05 level of significance. We compared the obstacles between study sites, and between communication by SMS and phone call. The Benjamini-Hochberg method was used to adjust the level of significance of comparison between obstacles in each site [<xref ref-type="bibr" rid="CR28">28</xref>]. WinPepi version 11.25 was used to specifically adjust <italic>p</italic> values [<xref ref-type="bibr" rid="CR29">29</xref>].</p></sec></sec><sec id="Sec9" sec-type="results"><title>Results</title><p>We enrolled 301 subjects: 119, 142, and 40 respectively in rural, semi-urban and urban areas. Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref> shows the general characteristics of the study population. The mean age of caregivers was 42.9&#x000a0;years (SD 13.4) and 46 caregivers (15.3%) were male. Most of them, 148 (49.2%) had completed a primary level of education.<table-wrap id="Tab1"><label>Table 1</label><caption><p> <bold>General characteristics and obstacles to the use of mobile phone reminders for mHealth in Cameroon</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th rowspan="2"/><th> <bold>Rural</bold> </th><th> <bold>Semi-urban</bold> </th><th> <bold>Urban</bold> </th><th> <bold>Total</bold> </th></tr><tr valign="top"><th> <bold>n = 119</bold> </th><th> <bold>n = 142</bold> </th><th> <bold>n = 40</bold> </th><th> <bold>N = 301</bold> </th></tr></thead><tbody><tr valign="top"><td> <bold><italic>General characteristics</italic></bold> </td><td/><td/><td/><td/></tr><tr valign="top"><td>Children</td><td/><td/><td/><td/></tr><tr valign="top"><td>- Mean age, years</td><td>3.6 (3.8)</td><td>2.6 (4.0)</td><td>3.0 (3.7)</td><td>3.1 (3.9)</td></tr><tr valign="top"><td>- Boys</td><td>46 (38.7)</td><td>70 (49.3)</td><td>27 (67.5)</td><td>143 (47.5)</td></tr><tr valign="top"><td>Caregivers</td><td/><td/><td/><td/></tr><tr valign="top"><td>- Mean age, years</td><td>42.2 (12.9)</td><td>43.1 (13.7)</td><td>43.8 (13.8)</td><td>42.9 (13.4)</td></tr><tr valign="top"><td>- Men</td><td>22 (18.5)</td><td>19 (13.4)</td><td>5 (12.5)</td><td>46 (15.3)</td></tr><tr valign="top"><td>Caregivers level of education</td><td/><td/><td/><td/></tr><tr valign="top"><td>- No formal</td><td>34 (28.6)</td><td>39 (27.5)</td><td>8 (20.0)</td><td>81 (26.9)</td></tr><tr valign="top"><td>- Primary</td><td>55 ( 46.2)</td><td>67 (47.2)</td><td>26 (65.0)</td><td>148 (49.2)</td></tr><tr valign="top"><td>- Secondary</td><td>16 (13.4)</td><td>23 (16.2)</td><td>4 (10.0)</td><td>43 (14.3)</td></tr><tr valign="top"><td>- University</td><td>14 (11.8)</td><td>13 (9.2)</td><td>2 (5.0)</td><td>29 (9.6)</td></tr><tr valign="top"><td>Time to scheduled appointment, days</td><td>32.1 (17.7)</td><td>27.7 (12.5)</td><td>27.8 (15.9)</td><td>29.5 (15.3)</td></tr><tr valign="top"><td> <bold><italic>Obstacles</italic></bold> </td><td/><td/><td/><td/></tr><tr valign="top"><td>At least one obstacle</td><td>47 (39.5)</td><td>9 (6.3)</td><td>3 (7.5)</td><td>59 (19.6)</td></tr><tr valign="top"><td>Without mobile phone</td><td>14 (11.8)</td><td>1 (0.7)</td><td>0</td><td>15 (5.0)</td></tr><tr valign="top"><td>Unable to communicate via text message in NOL</td><td>41 (34.5)</td><td>5 (3.5)</td><td>1 (2.5)</td><td>47 (15.6)</td></tr><tr valign="top"><td>Unable to communicate via voice phone call in NOL</td><td>27 (22.7)</td><td>4 (2.8)</td><td>0</td><td>31 (10.3)</td></tr><tr valign="top"><td>Declined to receive text message</td><td>3 (2.5)</td><td>5 (3.5)</td><td>3 (7.5)</td><td>11 (3.7)</td></tr><tr valign="top"><td>Declined to receive voice phone call</td><td>0</td><td>2 (1.4)</td><td>1 (2.5)</td><td>3 (1.0)</td></tr></tbody></table><table-wrap-foot><p>NOL: National official languages.</p><p>Data are n (%) or mean (standard deviation).</p></table-wrap-foot></table-wrap></p><p>This study revealed that 80.1% of the study population did not present any of the obstacles to receiving mobile phone reminders. Regarding each study site, the distribution of the absence of obstacles was: 60.5% in rural, 93.7% in semi &#x02013; urban, and 92.5% in urban settings. The greatest obstacle was the inability to read an SMS message (15.6%) followed by the inability to communicate orally (10.3%) in NOL. Very few caregivers refused to receive a SMS (3.7%) or a phone call (1.0%) to remind them of the child&#x02019;s upcoming medical appointment. The extent of non-possession of a mobile phone was also low (5.0%) (Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>).</p><p>The occurrence of at least one obstacle to mobile reminders was more frequent in rural than in semi-urban (<italic>p</italic> &#x0003c;0.001) and urban (<italic>p</italic> &#x0003c;0.001) areas. Caregivers without a mobile phone were more common in rural than in semi-urban (<italic>p</italic> &#x0003c;0.001) and urban (<italic>p</italic> = 0.03) areas. The inability to use a NOL for text messaging was more prevalent among caregivers living in a rural area as compared to caregivers living in semi-urban (<italic>p</italic> &#x0003c;0.001) and urban (<italic>p</italic> = 0.002) areas. There were no differences between geographic areas regarding the refusal to receive text messaging reminder and voice phone call reminders. Also, there was no difference between urban and semi-urban areas regarding the mHealth impediments we evaluated (Table&#x000a0;<xref rid="Tab2" ref-type="table">2</xref>).<table-wrap id="Tab2"><label>Table 2</label><caption><p> <bold>Comparison of impediments to mobile phone reminders for mHealth between sites (</bold> <bold><italic>p</italic></bold> <bold>values) in Cameroon</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th> <bold>Obstacles</bold> </th><th> <bold>Urban vs. semi-urban</bold> </th><th> <bold>Urban vs. rural</bold> </th><th> <bold>Semi-urban vs. rural</bold> </th></tr></thead><tbody><tr valign="top"><td>At least one obstacle</td><td>.73</td><td>&#x0003c; .001</td><td>&#x0003c; .001</td></tr><tr valign="top"><td>Without mobile phone</td><td>1.0</td><td>.03</td><td>&#x0003c; .001</td></tr><tr valign="top"><td>Unable to communicate via text message in NOL</td><td>1.0</td><td>&#x0003c; .001</td><td>&#x0003c; .001</td></tr><tr valign="top"><td>Unable to communicate via voice phone call in NOL</td><td>.58</td><td>.002</td><td>&#x0003c; .001</td></tr><tr valign="top"><td>Reject to receive text message</td><td>.57</td><td>.51</td><td>.73</td></tr><tr valign="top"><td>Reject to receive voice phone call</td><td>.53</td><td>.06</td><td>.75</td></tr></tbody></table><table-wrap-foot><p>NOL: National official languages.</p></table-wrap-foot></table-wrap></p><p>Non-ownership of a mobile phone was associated with geographic areas of residence (<italic>p</italic> &#x0003c;0.001), and with the inability to use a NOL for text messaging (<italic>p</italic> &#x0003c;0.001) and voice phone calling (<italic>p</italic> &#x0003c;0.001) (Table&#x000a0;<xref rid="Tab3" ref-type="table">3</xref>). There was no association between caregiver age, sex, level of education attained, or time until the scheduled appointment and the refusal to receive appointment reminder by text message or voice phone call (Table&#x000a0;<xref rid="Tab4" ref-type="table">4</xref>). Impediments to using SMS were not significantly different than those to using voice phone calls (Table&#x000a0;<xref rid="Tab5" ref-type="table">5</xref>).<table-wrap id="Tab3"><label>Table 3</label><caption><p> <bold>Comparison of adult caregivers of children requiring follow-up medical care for HIV with and without mobile phone</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th/><th> <bold>With phone (n = 286)</bold> </th><th> <bold>Without mobile phone (n = 15)</bold> </th><th> <bold><italic>p</italic></bold> <bold>value</bold> </th></tr></thead><tbody><tr valign="top"><td>Sites</td><td/><td/><td/></tr><tr valign="top"><td>- Rural</td><td>105 (36.7)</td><td>14 (93.3)</td><td rowspan="3">&#x0003c; .001</td></tr><tr valign="top"><td>- Semi-urban</td><td>141 (49.3)</td><td>1 (6.7)</td></tr><tr valign="top"><td>- Urban</td><td>40 (14.0)</td><td>0</td></tr><tr valign="top"><td>Caregivers mean age, years</td><td>42.6 (13.5)</td><td>46.9 (11.5)</td><td>.23</td></tr><tr valign="top"><td>Caregivers male</td><td>43 (15.0)</td><td>3 (20.0)</td><td>.71</td></tr><tr valign="top"><td>Caregivers level of education</td><td/><td/><td/></tr><tr valign="top"><td>- No formal/Primary</td><td>219 (76.6)</td><td>10 (66.7)</td><td rowspan="2">.36</td></tr><tr valign="top"><td>- Secondary/University</td><td>67 (23.4)</td><td>5 (33.3)</td></tr><tr valign="top"><td>Unable to communicate via text message in NOL</td><td>38 (13.3)</td><td>9 (60.0)</td><td>&#x0003c; .001</td></tr><tr valign="top"><td>Unable to communicate via voice phone call in NOL</td><td>24 (8.4)</td><td>7 (46.7)</td><td>&#x0003c; .001</td></tr><tr valign="top"><td>Declined to receive text message</td><td>9 (3.1)</td><td>2 (13.3)</td><td>.09</td></tr><tr valign="top"><td>Declined to receive voice phone call</td><td>3 (1.0)</td><td>0</td><td>1.0</td></tr></tbody></table><table-wrap-foot><p>NOL: National official languages.</p><p>Data are n (%) or mean (standard deviation).</p></table-wrap-foot></table-wrap><table-wrap id="Tab4"><label>Table 4</label><caption><p> <bold>Comparison between adult caregivers who rejected or adhered to SMS/voice phone call reminders</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th/><th colspan="3"> <bold>SMS reminder</bold> </th><th colspan="3"> <bold>Voice call reminder</bold> </th><th colspan="3"> <bold>One of both reminder</bold> </th></tr><tr valign="top"><th/><th> <bold>Adhere (n = 290)</bold> </th><th> <bold>Reject (n = 11)</bold> </th><th> <bold><italic>p</italic></bold> </th><th> <bold>Adhere (n = 298)</bold> </th><th> <bold>Reject (n = 3)</bold> </th><th> <bold><italic>p</italic></bold> </th><th> <bold>Adhere (n = 287)</bold> </th><th> <bold>Reject (n = 14)</bold> </th><th> <bold><italic>p</italic></bold> </th></tr></thead><tbody><tr valign="top"><td>Caregivers&#x02019; age, years</td><td>50.0 (11.9)</td><td>42.6 (13.4)</td><td>.07</td><td>42.9 (13.4)</td><td>38.3 (11.5)</td><td>.557</td><td>47.5 (12.4)</td><td>42.6 (13.4)</td><td>.18</td></tr><tr valign="top"><td>Male caregivers</td><td>42 (14.5)</td><td>4 (36.4)</td><td>.07</td><td>46 (15.4)</td><td>0</td><td>1.0</td><td>42 (14.6)</td><td>14 (100.0)</td><td>.24</td></tr><tr valign="top"><td>Caregivers&#x02019; level of education</td><td/><td/><td/><td/><td/><td/><td/><td/><td/></tr><tr valign="top"><td>- No formal/Primary</td><td>221 (76.2)</td><td>8 (72.7)</td><td rowspan="2">.73</td><td>226 (75.8)</td><td>3 (100.0)</td><td rowspan="2">1.0</td><td>218 (76.0)</td><td>11 (78.6)</td><td rowspan="2">1.0</td></tr><tr valign="top"><td>- Secondary/University</td><td>69 (23.8)</td><td>3 (27.3)</td><td>72 (24.2)</td><td>0</td><td>69 (24.0)</td><td>3 (21.4)</td></tr><tr valign="top"><td>Time to scheduled appointment</td><td>29.4 (15.4)</td><td>31.8 (14.5)</td><td>.60</td><td>29.5 (15.4)</td><td>28.3 (3.5)</td><td>.90</td><td>29.4 (15.4)</td><td>31.1 (12.9)</td><td>.69</td></tr></tbody></table><table-wrap-foot><p>Data are n (%) or mean (standard deviation).</p></table-wrap-foot></table-wrap><table-wrap id="Tab5"><label>Table 5</label><caption><p> <bold>Comparison of impediments to the use of text message and phone call as appointment reminders</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th rowspan="2"> <bold>Obstacles, n (%)</bold> </th><th> <bold>Text message</bold> </th><th> <bold>Voice phone call</bold> </th><th rowspan="2"> <bold><italic>p</italic></bold> <bold>value</bold> </th></tr><tr valign="top"><th> <bold>n = 301</bold> </th><th> <bold>n = 301</bold> </th></tr></thead><tbody><tr valign="top"><td>Refusal</td><td>11 (3.7)</td><td>3 (1.0)</td><td>.054</td></tr><tr valign="top"><td>Unable to communicate</td><td>47 (15.6)</td><td>31 (10.3)</td><td>.052</td></tr></tbody></table></table-wrap></p></sec><sec id="Sec10" sec-type="discussion"><title>Discussion</title><p>This study reveals that the use of mobile phones for medical follow-up mHealth appointment reminders in pediatric HIV could potentially apply to 80% of the overall population in Cameroon. Considering each study site separately, the potential penetration of such mHealth use would be different, as we captured 60.5% of caregivers in rural, 93.7% of caregivers in semi - urban and 92.5% of caregivers in urban areas. The greatest obstacle to mobile phone reminders was an adult caregiver&#x02019;s inability to read an SMS message, followed an inability to communicate orally in English or French, which are Cameroon&#x02019;s two national official languages. Very few subjects refused to receive a SMS or a phone call to remind them of the child&#x02019;s medical appointment. The rate of mobile phone non-possession was also low. All impediments to mobile reminders were more frequent in the rural setting, except for the refusal to receive SMS or phone call. SMS or phone call showed no difference in their difficulty of use.</p><p>Language illiteracy was the major barrier in our study, as in others [<xref ref-type="bibr" rid="CR24">24</xref>], and was more pronounced in rural areas in our study. In Uganda, in a region where 80% of people live in rural areas, most persons speak a local language different than NOL [<xref ref-type="bibr" rid="CR30">30</xref>]. Before SMS and phone calls can be widely implemented as medical appointment reminders, it will be necessary to assess the feasibility of a particular linguistic communication program specific to each health care setting. The literacy rate will need to be improved, especially in rural areas, to help achieve gains in globalization and automation of medical appointment reminders via SMS and phone calls. Health care providers could also adapt and use local languages and dialects to deliver messages about upcoming medical appointments. We suggest that medical assistants in each HIV care center be trained to communicate in the most widely spoken local language (or dialect) in their catchment area in order to reach and include patients and caregivers who communicate exclusively in their local language. The challenge lies in the great number of ethno-linguistic groups in Cameroon, recently assessed at 286 [<xref ref-type="bibr" rid="CR31">31</xref>]. It will be therefore necessary to target the most widely spoken local languages in each health district.</p><p>In our study, 95% of subjects owned a mobile phone. The rate of ownership is higher than that found in Durban, South Africa in 2010 where 81% of people living with HIV had a mobile phone [<xref ref-type="bibr" rid="CR26">26</xref>] and in Nigeria where 68% of a diabetes population owned a mobile phone [<xref ref-type="bibr" rid="CR32">32</xref>]. The finding is explained by the exponential increase in the number of subscriptions to mobile phone companies each year [<xref ref-type="bibr" rid="CR7">7</xref>,<xref ref-type="bibr" rid="CR8">8</xref>,<xref ref-type="bibr" rid="CR24">24</xref>]. Mobile phone ownership in our 2013 study is however greater than that found in Kenya in the same year (61.2%) [<xref ref-type="bibr" rid="CR33">33</xref>]. The difference may reflect the greater proportion of the population living below the national poverty line in Kenya (45.9% in 2005) as compared to Cameroon (39.9% in 2007), and by Cameroon&#x02019;s higher per capita Gross National Income as compared to Kenya&#x02019;s [<xref ref-type="bibr" rid="CR34">34</xref>,<xref ref-type="bibr" rid="CR35">35</xref>]. In our study, subjects in urban and semi-urban areas owned more mobile phones compared to Kenya&#x02019;s rural areas [<xref ref-type="bibr" rid="CR33">33</xref>], due to the fact that the socioeconomic level is lower in rural areas [<xref ref-type="bibr" rid="CR36">36</xref>]. The observed regional and socioeconomic heterogeneity of mobile phone ownership was also demonstrated in Kenya [<xref ref-type="bibr" rid="CR37">37</xref>] where factors like gender, educational level, literacy and income are also thought to have an influence on mobile phone ownership [<xref ref-type="bibr" rid="CR37">37</xref>]. The present study also reveals that people unable to communicate in NOL by text message or orally are most likely not to own a mobile phone. In contrast to another study from sub-Saharan Africa, we found that neither gender nor educational level was a factor in mobile phone ownership [<xref ref-type="bibr" rid="CR37">37</xref>]. This implies that increasing the NOL literacy rate could lead to an increase in the rate of mobile phone ownership. Although it might be of interest to assess the feasibility and acceptability of using a shared or borrowed mobile phone for appointment reminders and delivery of other healthcare related information for caregivers not owning a mobile phone, we believe that such a &#x0201c;solution&#x0201d; carries significant ethical concerns, especially in regards to privacy and confidentiality.</p><p>The acceptability of mobile phones for our mHealth reminder intervention is high (95%) in our study, as it is in other studies in sub-Saharan Africa [<xref ref-type="bibr" rid="CR24">24</xref>&#x02013;<xref ref-type="bibr" rid="CR26">26</xref>,<xref ref-type="bibr" rid="CR30">30</xref>,<xref ref-type="bibr" rid="CR38">38</xref>]. In our study, age, sex, and educational level of the caregivers did not influence acceptability. Also, the interval of time until the scheduled follow-up appointment did not influence acceptability. The adult caregivers were agreeable to receiving an appointment reminder by mobile phone independently of how close or far away the upcoming appointment was in time. The willingness of an adult caregiver to receive mobile phone reminders for a child&#x02019;s follow-up HIV care likely reflects an interest in the well-being and continued optimal care of the child and incorporates the perception that a reminder helps ensure better monitoring and delivery of required ongoing care. It will thus be informative and relevant in future studies to investigate the motivations of caregivers of children exposed to or infected with HIV who refuse to receive mobile phone reminders.</p><p>Our study found no significant differences in the rates of refusal between SMS and voice phone calls. Crankshaw et al. found a 99% acceptability for phone calls and 96% for SMS [<xref ref-type="bibr" rid="CR26">26</xref>]. Their result is very similar to ours; we achieved 99% acceptability for calls and 96.3% for SMS. This suggests that we could freely choose to use SMS or phone calls in sub-Saharan Africa in terms of communication difficulty.</p><p>Our study has some limitations. Barriers in addition to those we assessed may impede the use of mobile phones as a medical appointment reminder aid. Examples of obstacles that we did not examine include the timing of sending messages, the unavailability and fluctuations of the wireless network, the phone being powered off, low motivation of medical assistants, and privacy concerns about health [<xref ref-type="bibr" rid="CR19">19</xref>,<xref ref-type="bibr" rid="CR24">24</xref>,<xref ref-type="bibr" rid="CR26">26</xref>,<xref ref-type="bibr" rid="CR39">39</xref>].</p></sec><sec id="Sec11" sec-type="conclusion"><title>Conclusions</title><p>Despite the proven benefits of mobile health in HIV in sub-Saharan Africa [<xref ref-type="bibr" rid="CR3">3</xref>,<xref ref-type="bibr" rid="CR5">5</xref>], the use of mobile phone SMS or voice calls to caregivers providing scheduled appointment reminders for pediatric HIV care in Cameroon encountered a relatively high proportion of obstacles. The most important obstacle was the inability to use spoken or written NOL, followed by non-ownership of a mobile phone. These impediments were higher in a rural area than in urban or semi-urban areas. Rural areas, during the implementation of any health policy based on SMS and phone calls, would thus benefit from a targeted program addressing difficulties in communicating in verbal and written NOL and low mobile phone ownership rate. In terms of acceptability and efficiency, both SMS and phone calls were equivalent when used as mHealth reminder tools in our Cameroonian population of adult caregivers of children requiring HIV care.</p></sec>

Barriers and facilitators to using NHS Direct: a qualitative study of &#x02018;users&#x02019; and &#x02018;non-users&#x02019;

Could not extract abstract

<contrib contrib-type="author" corresp="yes"><name><surname>Cook</surname><given-names>Erica J</given-names></name><address><email>erica.cook@beds.ac.uk</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author"><name><surname>Randhawa</surname><given-names>Gurch</given-names></name><address><email>gurch.randhawa@beds.ac.uk</email></address><xref ref-type="aff" rid="Aff2"/></contrib><contrib contrib-type="author"><name><surname>Large</surname><given-names>Shirley</given-names></name><address><email>shirley.large2@nhs.net</email></address><xref ref-type="aff" rid="Aff3"/></contrib><contrib contrib-type="author"><name><surname>Guppy</surname><given-names>Andy</given-names></name><address><email>andy.guppy@beds.ac.uk</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author"><name><surname>Chater</surname><given-names>Angel M</given-names></name><address><email>a.chater@ucl.ac.uk</email></address><xref ref-type="aff" rid="Aff4"/></contrib><contrib contrib-type="author"><name><surname>Ali</surname><given-names>Nasreen</given-names></name><address><email>nasreen.ali@beds.ac.uk</email></address><xref ref-type="aff" rid="Aff2"/></contrib><aff id="Aff1"><label/>Department of Psychology, University of Bedfordshire, Park Square, Luton, UK </aff><aff id="Aff2"><label/>Institute for Health Research, University of Bedfordshire, Putteridge Bury, Hitchin Road, Luton, UK </aff><aff id="Aff3"><label/>NHS England, Horley, UK </aff><aff id="Aff4"><label/>UCL School of Pharmacy, BMA House, Tavistock Square, London, UK </aff>

BMC Health Services Research

<sec id="Sec1" sec-type="introduction"><title>Background</title><p>In 2012, NHS Direct was replaced with the new non-emergency &#x02018;111&#x02019; telephone-based healthcare service. It was first introduced in 2010, followed by a national rollout in 2013. The aim was to provide a more integrated non-emergency service to provide a gateway for all non-urgent healthcare needs [<xref ref-type="bibr" rid="CR1">1</xref>]. Although the introduction of &#x02018;111&#x02019; has marked the end of NHS Direct [<xref ref-type="bibr" rid="CR2">2</xref>], it has highlighted the increased role that telephone-based healthcare has within the NHS structure. Therefore, to understand patterns of NHS Direct uptake have provided an opportunity to learn valuable lessons about access and uptake of telephone healthcare based services. This knowledge can be applied to the &#x02018;111&#x02019; telephone-based healthcare service, as well as services internationally, as countries worldwide adopt similar models of remote healthcare delivery [<xref ref-type="bibr" rid="CR3">3</xref>-<xref ref-type="bibr" rid="CR6">6</xref>].</p><p>NHS Direct provided 24&#x000a0;hour/7&#x000a0;day a week nurse led telephone-based healthcare advice and information to the public in England and Wales [<xref ref-type="bibr" rid="CR7">7</xref>,<xref ref-type="bibr" rid="CR8">8</xref>] (see Additional file <xref rid="MOESM1" ref-type="media">1</xref>). This service, introduced in 1998, marked a strategic shift towards the self-care movement [<xref ref-type="bibr" rid="CR9">9</xref>] which encouraged the population to take an increased responsibility for their own health [<xref ref-type="bibr" rid="CR8">8</xref>,<xref ref-type="bibr" rid="CR10">10</xref>,<xref ref-type="bibr" rid="CR11">11</xref>]. Evidence suggests that self-care is linked to improved health outcomes, improved quality of life, increased empowerment and patient satisfaction [<xref ref-type="bibr" rid="CR11">11</xref>-<xref ref-type="bibr" rid="CR13">13</xref>] and has been viewed as beneficial in reducing hospital admissions [<xref ref-type="bibr" rid="CR14">14</xref>]. Consequently, self-care is now being viewed as an inextricable part of the individual care pathway, from maintaining a healthy lifestyle to caring for minor, acute and long-term health conditions [<xref ref-type="bibr" rid="CR15">15</xref>].</p><p>NHS Direct has been at the leading edge of remote healthcare systems, directing healthcare into the 21st Century through the application of new technology solutions in primary care [<xref ref-type="bibr" rid="CR16">16</xref>]. By 2011, NHS Direct received 8 million calls per year with reported high levels of satisfaction [<xref ref-type="bibr" rid="CR17">17</xref>]. Whilst evidence suggests that there is an increasing shift towards self-care [<xref ref-type="bibr" rid="CR11">11</xref>], with over 90% of people cited as being interested in taking more ownership of their health [<xref ref-type="bibr" rid="CR18">18</xref>], the pattern is not uniform across all sections of society. For example, self-care uptake (and NHS Direct usage) has previously been reported to be substantially lower in those who are older (85+) [<xref ref-type="bibr" rid="CR19">19</xref>], among the less affluent and deprived [<xref ref-type="bibr" rid="CR20">20</xref>,<xref ref-type="bibr" rid="CR21">21</xref>] and minority ethnic groups [<xref ref-type="bibr" rid="CR22">22</xref>].</p><p>Uptake of telephone-based healthcare services has been explained by the technical performance and functional reliability of technology [<xref ref-type="bibr" rid="CR23">23</xref>], concerns of personal privacy and security [<xref ref-type="bibr" rid="CR22">22</xref>,<xref ref-type="bibr" rid="CR24">24</xref>], money, perceived confidence to engage with health technology [<xref ref-type="bibr" rid="CR20">20</xref>,<xref ref-type="bibr" rid="CR25">25</xref>,<xref ref-type="bibr" rid="CR26">26</xref>] and severity of health symptom(s) [<xref ref-type="bibr" rid="CR25">25</xref>]. Perceived confidence to engage with health technology and severity of symptoms suggests that if an individual has low confidence to use health technology and has high perceived severity of illness, they are more likely to prefer face-to-face contact with a healthcare professional [<xref ref-type="bibr" rid="CR26">26</xref>] and less likely to see the benefits in self-care [<xref ref-type="bibr" rid="CR27">27</xref>]. Factors enabling self-care include awareness of the services, and service recommendation and signposting by healthcare professionals [<xref ref-type="bibr" rid="CR18">18</xref>].</p><p>There is a dearth of evidence exploring explanations for usage and non-usage of NHS Direct. As the provision of healthcare moves away from face-to-face contact between patient and practitioner there is a pressing need to understand the reasons for usage and non-usage of telephone-based healthcare services to ensure that all sections of society are able to maximise opportunities for self-care. To examine the usage of NHS Direct this research makes a small, but valuable contribution, to help understand the barriers and facilitators to usage of telephone-based healthcare services.</p></sec><sec id="Sec2" sec-type="materials|methods"><title>Methods</title><sec id="Sec3"><title>Sampling and recruitment</title><p>Nine focus groups were conducted between October 2011 and January 2012. A trained researcher (EC) facilitated the focus groups with the support of a research assistant, both of whom had no direct connection with NHS Direct that ensured the focus groups were unbiased. Focus group methodology was used to generate data as it involves a group interaction which can help participants to explore and clarify their views in ways that may be less accessible in a one to one interview [<xref ref-type="bibr" rid="CR28">28</xref>,<xref ref-type="bibr" rid="CR29">29</xref>]. This methodology has been a commonly applied approach in health services research to identify views and attitudes towards health services [<xref ref-type="bibr" rid="CR30">30</xref>-<xref ref-type="bibr" rid="CR32">32</xref>].</p><p>Ethical approval was granted by the University of Bedfordshire ethics committee in March 2010 and the NHS Ethics Committee in April 2010 (REF: 11/H0301/8). All participants who took part in this study provided their written informed consent. Participants&#x02019; anonymity and confidentiality was ensured throughout.</p><p>Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref> presents the demographic composition of all focus groups and shows that each focus group comprised of between five and twelve participants (a total of 81 participants: 62 females and 19 males). Participants&#x02019; ages ranged between 21 and 94&#x000a0;years with the majority White British. A purposive stratified sampling strategy [<xref ref-type="bibr" rid="CR33">33</xref>] was used to recruit in three geographical areas in England to ascertain diversity of opinion.<table-wrap id="Tab1"><label>Table 1</label><caption><p> <bold>Focus group composition and recruitment of NHS Direct users and non-users</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th colspan="2"> <bold>Focus group</bold> </th><th> <bold>Location</bold> </th><th> <bold>Age</bold> </th><th> <bold>Gender</bold> </th><th> <bold>Ethnicity</bold> </th><th> <bold>Profile and characteristics</bold> </th><th> <bold>Focus group description</bold> </th></tr></thead><tbody><tr valign="top"><td>1</td><td>Users (N&#x02009;=&#x02009;8)</td><td> <bold>Mid Bedfordshire</bold> </td><td>21&#x02013;46</td><td>Female (9)</td><td>White British (7) Mixed: Black Caribbean (1)</td><td>High geographical usage area &#x02013; mothers with children (&#x0003c;5)</td><td>Participants recruited from Sure Start centres in Mid-Bedfordshire. Sure Start centres are open to parents, carers and children providing early learning and full day care for pre-school children.</td></tr><tr valign="top"><td>2</td><td>Users (N&#x02009;=&#x02009;9)</td><td> <bold>Mid Bedfordshire</bold> </td><td>23&#x02013;54</td><td>Female (9)</td><td>White British (9)</td><td>High geographical usage area &#x02013; mothers with children (&#x0003c;5)</td><td>Participants recruited from a range of Sure Start centres in Mid-Bedfordshire. Sure Start centres are open to all parents, carers and children providing early learning and full day care for pre-school children.</td></tr><tr valign="top"><td>3</td><td>Non-users (N&#x02009;=&#x02009;10)</td><td> <bold>Mendip: Moor</bold> </td><td>67&#x02013;93</td><td>Male (6); Female (4)</td><td>White British (10)</td><td>Older residents with high levels of deprivation residing in isolated rural community</td><td>Focus groups were held as part of an existing community group which provides retired adults mainly older (65+) a range of social activities and events.</td></tr><tr valign="top"><td>4</td><td>Non-users (N&#x02009;=&#x02009;11)</td><td> <bold>Mendip: Mells</bold> </td><td>67&#x02013;94</td><td>Female (11)</td><td>White British (11)</td><td>Older residents with high levels of deprivation residing in isolated rural community</td><td>Focus groups were held as part of an existing community group which provides retired adults mainly older (65+) a range of social activities and events.</td></tr><tr valign="top"><td>5</td><td>Non-users (N&#x02009;=&#x02009;9)</td><td> <bold>Mendip: Creech</bold> </td><td>64&#x02013;92</td><td>Female (9)</td><td>White British (9)</td><td>Older residents living in larger isolated rural community.</td><td>Focus groups were held as part of an existing community group which provides retired adults mainly older (65+) a range of social activities and events.</td></tr><tr valign="top"><td>6</td><td>Non-users (N&#x02009;=&#x02009;11)</td><td> <bold>Mendip: Beckington &#x00026; Rode</bold> </td><td>50&#x02013;87</td><td>Male (3) Female (8)</td><td>White British (11)</td><td>Middle income families living in moderate suburban semis in a rural area.</td><td>Focus groups were held as part of an existing community group which provides retired adults mainly older (65+) a range of social activities and events.</td></tr><tr valign="top"><td>7</td><td>Non-users (N&#x02009;=&#x02009;7)</td><td> <bold>Manchester: Baguley</bold> </td><td>36&#x02013;73</td><td>Male (2) Female (5)</td><td>White British (7)</td><td>Deprived ward resided by families in low rise social housing with high levels of benefit need.</td><td>Participants recruited from a range of community organisations which provide residents with their social, recreational and sporting needs.</td></tr><tr valign="top"><td>8</td><td>Non-users (N&#x02009;=&#x02009;11)</td><td> <bold>Manchester: Gorton North</bold> </td><td>16&#x02013;84</td><td>Male (3) Female (8)</td><td>White British (11)</td><td>Deprived ward characterised by low income workers in urban terraces.</td><td>Participants were recruited from a range of community organisations which provide residents with their social, recreational and sporting needs.</td></tr><tr valign="top"><td>9</td><td>Non-users (N&#x02009;=&#x02009;6)</td><td> <bold>Manchester: Longsight</bold> </td><td>26&#x02013;49</td><td>Male (6)</td><td>White British (2) Pakistani (2) Black African (2)</td><td>Deprived ward characterised by low income workers in urban terraces and culturally diverse areas.</td><td>Participants were recruited from a drop in community centre which provides residents a range of activities focusing on improving health and wellbeing.</td></tr></tbody></table></table-wrap></p></sec><sec id="Sec4"><title>NHS Direct &#x02018;users&#x02019;</title><p>NHS Direct &#x02018;users&#x02019; were purposefully chosen as mothers with young children (&#x0003c;5&#x000a0;years). Research suggests that this population group accounted for over 20% of all calls made [<xref ref-type="bibr" rid="CR21">21</xref>,<xref ref-type="bibr" rid="CR24">24</xref>] and represents the highest &#x02018;users&#x02019; of NHS Direct [<xref ref-type="bibr" rid="CR21">21</xref>]. NHS Direct &#x02018;users&#x02019; were recruited through two Children&#x02019;s Activity Centres in Mid-Bedfordshire, as these sites were based in high geographical usage areas [<xref ref-type="bibr" rid="CR24">24</xref>]. Prospective participants were approached by the lead researcher (EC) and invited to take part. If they were interested they were then screened to ensure that they met the inclusion criteria.</p><p>The inclusion criteria outlined that prospective participants were a mother of a child (&#x0003c;5&#x000a0;years) and had used NHS Direct at least once in the previous year for either themselves or their child. A total of two focus groups were held within this sample group before saturation was achieved [<xref ref-type="bibr" rid="CR34">34</xref>]. Participants (N&#x02009;=&#x02009;17) were aged between 21 and 54 (M&#x02009;=&#x02009;32.59; SD&#x02009;=&#x02009;8.4), the majority classified themselves as White British (N&#x02009;=&#x02009;16), with one participant who identified herself as Mixed White and Black Caribbean (Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>).</p></sec><sec id="Sec5"><title>Non NHS Direct &#x02018;users&#x02019;</title><p>Two Local Authorities were chosen, one urban and one rural with mortality used as a proxy to identify need. This approach allowed for the identification of geographical differences of life expectancy between regions, districts, wards and output areas [<xref ref-type="bibr" rid="CR35">35</xref>,<xref ref-type="bibr" rid="CR36">36</xref>]. Though the usage of expected life expectancy birth data, Local Authorities were chosen as (1) lowest life expectancy urban local authority area defined as urban 1 (predominantly major urban), and (2) the lowest life expectancy rural local authority area defined as urban 6 (predominantly rural 50/80) [<xref ref-type="bibr" rid="CR37">37</xref>].</p><p>Manchester was the chosen urban local authority which has the lowest life expectancy from birth which currently stands at 72.5 (CI: 72.1&#x02013;72.8). Mendip, located in the South West of England was the rural local authority chosen with the lowest life expectancy from birth which stands at 77.5 (CI: 76.8&#x02013;78.2) [<xref ref-type="bibr" rid="CR38">38</xref>]. Both geographical areas suffer from higher than average levels of deprivation. Manchester is ranked the fourth most deprived local authority in England [<xref ref-type="bibr" rid="CR39">39</xref>], whilst Mendip is shown to have high levels of unemployment with pockets of deprivation throughout [<xref ref-type="bibr" rid="CR40">40</xref>].</p><p>A stratified &#x02018;stratum&#x02019; sampling approach was then used on the basis of low geographical usage at ward level which was carried out and mapped NHS Direct call data and compared this to the concentration of calls by population through the use of geographical information system software ArcGIS [<xref ref-type="bibr" rid="CR41">41</xref>] (Figures&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref> and <xref rid="Fig2" ref-type="fig">2</xref>). The lowest usage wards were then explored using population segmentation (Mosaic) which provided detailed information that defined the population subgroups by a mix of demographic, cultural, behavioural, psychosocial, geographic factors [<xref ref-type="bibr" rid="CR42">42</xref>] (Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>). In relation to demography they should meet the characteristics of &#x02018;non-users&#x02019; of NHS Direct depending on the ward chosen.<fig id="Fig1"><label>Figure 1</label><caption><p> <bold>Penetration of calls to Mendip at ward area.</bold> </p></caption><graphic xlink:href="12913_2014_487_Fig1_HTML" id="MO1"/></fig><fig id="Fig2"><label>Figure 2</label><caption><p> <bold>Penetration of calls to Manchester at ward area</bold>.</p></caption><graphic xlink:href="12913_2014_487_Fig2_HTML" id="MO2"/></fig></p></sec><sec id="Sec6"><title>Mendip focus groups</title><p>A total of four focus groups were carried out in Mendip in the wards Moor, Mells, Creech and Beckington &#x00026; Rode. The Mendip sample were screened for age and were required to be &#x02264;50&#x000a0;years as this population sub-group represent the lowest users of NHS Direct [<xref ref-type="bibr" rid="CR24">24</xref>,<xref ref-type="bibr" rid="CR43">43</xref>,<xref ref-type="bibr" rid="CR44">44</xref>]. This sample (N&#x02009;=&#x02009;41) were predominantly White British females with ages ranging from 50&#x02013;94 (M&#x02009;=&#x02009;79.93, SD&#x02009;=&#x02009;10.08) (Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>).</p></sec><sec id="Sec7"><title>Manchester focus groups</title><p>A further three focus groups were organised with &#x02018;non-users&#x02019; in Manchester in the wards Gorton North, Longsight and Baguely. The Manchester sample (N&#x02009;=&#x02009;24) were also screened for representation of the geographical area and aimed to capture cultural diversity (Longsight) deprivation (Gorton North, Longsight and Baguely) and variation in gender within the residing wards (Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>).</p></sec><sec id="Sec8"><title>Setting</title><p>All community centres and day centres in the ward areas were visited by the lead researcher (EC) in person. The centre manager was provided with a recruitment poster, a lay overview of the study, alongside a participant information sheet which detailed the study and outlined the inclusion/exclusion criteria. This information was then disseminated to prospective participants and those who were interested were asked to provide their name and availability to the centre manager.</p><p>The focus groups were held in community or day centres. In Mendip some of the focus groups were held as part of an existing group; for Manchester, the focus groups were not part of an existing group. However, all participants were familiar with each other which allowed them to feel at ease within a familiar setting. Before the focus group took place, all participants were screened by a researcher (EC) to ensure that they met the inclusion criteria i.e. they reside in the ward defined and have not used NHS Direct service for either themselves or another person. Each focus group session lasted approximately 90&#x000a0;minutes and was audiotape-recorded with participants&#x02019; permission.</p></sec><sec id="Sec9"><title>Focus group process</title><p>Participants were asked questions surrounding their awareness of NHS Direct, why they have or have not used the service, the advantages and disadvantages around using this service, structural and perceived barriers relating to positive and negative attitudes. Questions also centred on the usefulness of the service and ease of use, alongside their attitudes towards communicating with healthcare professionals via the telephone. At the end of each focus group, the researcher gave the participants an opportunity to comment on the data and on the key themes that had emerged from the discussion to check and confirm accuracy. No factual errors were found in the data and there were no requests for amendments or amplifications. Data collection stopped once saturation had been reached.</p></sec><sec id="Sec10"><title>Analysis</title><p>Focus groups were audiotape-recorded and transcribed verbatim using pseudonyms by a member of the research team (EC). The framework approach [<xref ref-type="bibr" rid="CR45">45</xref>] was then used to thematically analyse the data which provides a generative analytical procedure that uses distinct connected stages of coding allowing for cases to be compared [<xref ref-type="bibr" rid="CR46">46</xref>]. The analysis closely followed the five distinct stages of analysis; (1) familiarisation, whereby the researchers independently reviewed a sample of the transcripts; (2) identify a thematic framework, whereby two researchers (EC &#x00026; NA) independently identified and organised key themes after coding the first few transcripts and cross checked. All major themes achieved consensus. The next stage (3) indexing, whereby the lead researcher (EC) independently applied the themes to the text. This was followed by stage (4), charting, where the data was managed and summarised using Excel whereby the summaries for each code were transferred into cells with assigned page numbers to track the narrative. The final stage (5), mapping and interpretation, occurred through exploration of the relationships and patterns within the data [<xref ref-type="bibr" rid="CR47">47</xref>], and was completed after research team discussions.</p></sec></sec><sec id="Sec11" sec-type="results"><title>Results</title><p>Five themes emerged throughout the analysis of the transcripts which related to awareness of the service, costs to the individual, ease of use, time/speed, and acceptability of non-face-to-face healthcare. Similarities and differences between &#x02018;users&#x02019; and &#x02018;non-users&#x02019; are identified for these themes where relevant (Table&#x000a0;<xref rid="Tab2" ref-type="table">2</xref>).<table-wrap id="Tab2"><label>Table 2</label><caption><p> <bold>Overview of similarities and differences of barriers/facilitators across the sample groups towards using NHS Direct</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th> <bold>Theme</bold> </th><th> <bold>User groups Bedfordshire</bold> </th><th> <bold>Non-user groups Manchester</bold> </th><th> <bold>Non-user groups Mendip</bold> </th></tr></thead><tbody><tr valign="top"><td rowspan="3"> <bold>Awareness of service</bold> </td><td>&#x02022; Good awareness and understanding of service</td><td>&#x02022; Lack of awareness</td><td>&#x02022; Lack of awareness</td></tr><tr valign="top"><td rowspan="2">&#x02022; Most participants had used a wide range of services NHS Direct had e.g. online self-assessment tool</td><td>&#x02022; Most participants had not heard of NHS Direct or services they provide</td><td>&#x02022; Most participants had not heard of NHS Direct or services they provide</td></tr><tr valign="top"><td>&#x02022; Some misunderstandings of what NHS Direct is</td><td>&#x02022; Some misunderstandings of what NHS Direct is</td></tr><tr valign="top"><td rowspan="2"> <bold>Cost to the individual</bold> </td><td>&#x02022; Most participants were not aware of the cost from a mobile phone</td><td>&#x02022; Viewed as very expensive</td><td>&#x02022; Expense was not viewed as a barrier</td></tr><tr valign="top"><td>&#x02022; All participants had a landline phone</td><td>&#x02022; Many of the participants did not have a landline phone</td><td>&#x02022; All participants had a landline phone</td></tr><tr valign="top"><td rowspan="5"> <bold>Ease of use</bold> </td><td>&#x02022; All participants found the service easy to use</td><td>&#x02022; Some participants felt that this would be an easy to use service</td><td>&#x02022; Difficulties in hearing over the phone</td></tr><tr valign="top"><td rowspan="4">&#x02022; Viewed easier than using conventional out-of hours services</td><td>&#x02022; Concern of complicated phone service with lots of options</td><td>&#x02022; Dislike of answering lots of questions over phone</td></tr><tr valign="top"><td>&#x02022; Being passed from person to person</td><td>&#x02022; Difficulty of understanding foreign accents</td></tr><tr valign="top"><td rowspan="2">&#x02022; Language barriers e.g. non English speaking</td><td>&#x02022; Technical issues e.g. afraid of being cut off</td></tr><tr valign="top"><td>&#x02022; Memory would make it difficult to use</td></tr><tr valign="top"><td rowspan="3"> <bold>Time/Speed</bold> </td><td>&#x02022; Seen as instant advice and reassurance</td><td>&#x02022; Concerned about waiting a long time for a call back</td><td>&#x02022; Concerned about waiting a long time for a call back</td></tr><tr valign="top"><td>&#x02022; Was viewed as a key advantage to using the service</td><td rowspan="2">&#x02022; Was viewed as wasting time</td><td rowspan="2">&#x02022; Was viewed as wasting time</td></tr><tr valign="top"><td>&#x02022; Sometimes there was a long time to wait for a call back from a nurse</td></tr><tr valign="top"><td rowspan="4"> <bold>Acceptability of non-face-to-face healthcare</bold> </td><td>&#x02022; Positive attitudes towards not having face-to-face contact</td><td>&#x02022; Preference for face-to-face healthcare</td><td>&#x02022; Preference for face-to-face healthcare</td></tr><tr valign="top"><td>&#x02022; Provided reassurance</td><td>&#x02022; Would feel that they are unable to express themselves</td><td>&#x02022; Would feel that they are unable to express themselves</td></tr><tr valign="top"><td rowspan="2">&#x02022; Viewed service as personable and professional</td><td>&#x02022; Would not provide reassurance</td><td>&#x02022; Would not provide reassurance</td></tr><tr valign="top"><td>&#x02022; Was not viewed as personable</td><td>&#x02022; Was not viewed as personable</td></tr></tbody></table></table-wrap></p><sec id="Sec12"><title>Awareness of service</title><p>Overall, NHS Direct users had a good awareness and understanding of the service. They were aware of all individual services on offer including the core triage provision, health information and medicine advice services. Many participants were also aware of the internet based services, including the health encyclopedia and the Self-Assessment Tool software, which many had used to receive a call back relating to symptoms either for themselves or their children. There was a variety of ways in which the participants had heard about NHS Direct. Many &#x02018;users&#x02019; were directed to NHS Direct through their GP answer phone machine when they had phoned their surgery out of hours.<disp-quote><p><italic>&#x02018;When I first called it I had called my doctor and the doctors surgery didn&#x02019;t have an out of hours so they actually give you the NHS Direct number so that&#x02019;s how I knew the number&#x02019; (NHS Direct &#x02018;user&#x02019;, FG1)</italic></p></disp-quote></p><p>However others were made aware of NHS Direct through their midwives when they had children.<disp-quote><p><italic>&#x02018;I think it was from the midwife when I had just given birth, she came to the house to do a check and she gave me the number then&#x02019; (NHS Direct &#x02018;user&#x02019;, FG1)</italic></p></disp-quote></p><p>One participant saw the service advertised through yellow pages (a telephone directory), and also recalled seeing through local level advertising. In fact, a number of participants recalled a small credit card leaflet which had the telephone number on which participants could keep in their wallet.<disp-quote><p><italic>&#x02018;I think I knew through getting information through the post&#x02026;.it was a white card with blue writing&#x02019; (NHS Direct &#x02018;user&#x02019;, FG2)</italic></p></disp-quote></p><p>Conversely, in Manchester and Mendip there was a distinct lack of awareness was evident across all &#x02018;non-user&#x02019; focus groups. Many of the participants had never heard about NHS Direct or the services that they provide. There were also uncertainties and misunderstandings of what services NHS Direct offered. For example, a number of participants thought that NHS Direct was a walk in clinic or provided an out of hours GP service.<disp-quote><p><italic>&#x02018;I&#x02019;ve heard about it it&#x02019;s supposed to make life easier or that&#x02019;s all I have heard it supposed to do with phone calls or Internet and that&#x02019;s about it&#x02019; (NHS Direct &#x02018;non-user&#x02019;, FG7)</italic></p><p><italic>&#x02018;I think some people myself included are getting confused with people ringing NHS Direct with people who ring their out of hours duty officer&#x02019; (NHS Direct &#x02018;non-user&#x02019;, FG6)</italic></p></disp-quote></p></sec><sec id="Sec13"><title>Costs to the individual</title><p>NHS Direct operated from a &#x02018;0845&#x02019; number, which is a cost of a local rate from a landline. However, the cost is substantially higher from a mobile phone when not covered by an inclusive minutes plan [<xref ref-type="bibr" rid="CR48">48</xref>]. It is important to note that the researcher did not explain the cost to identify awareness of this, so anything relating to cost was brought up by the participants.</p><p>Amongst the NHS Direct &#x02018;users&#x02019; only one participant mentioned the cost of the phone call, whereby she spoke of her friend who was a single parent and could not access the service because of the expense incurred on the use of her mobile phone. Many of the &#x02018;users&#x02019;, use landlines to phone NHS Direct and were not aware of the cost implications to use a mobile phone. However, when they realised this all participants said that this would not affect future usage.<disp-quote><p><italic>&#x02018;She&#x02019;s a single parent and she&#x02019;s only got her mobile phone and she said the only issue she has because it&#x02019;s an 0845 number and on her mobile it costs a lot&#x02026;.because she only has her mobile its three, four, five pounds&#x02019; (NHS Direct &#x02018;user&#x02019;, FG1)</italic></p></disp-quote></p><p>However, &#x02018;non-users&#x02019; in the focus groups in Manchester were much more aware of the cost incurred when using NHS Direct, whereby this service was viewed as very expensive. Many of the participants did not have landline phones so had to rely on using mobile phones to access the service.<disp-quote><p><italic>&#x02018;The cost is a big issue especially if you don&#x02019;t have a landline and if you have to do on a mobile phone if you are on pay-as-you-go then contract it&#x02019;s dearer&#x02019; (NHS Direct &#x02018;non-user&#x02019;, FG7)</italic></p><p><italic>&#x02018;It is a paid number it puts people off that it isn&#x02019;t a free number we only get credit once a fortnight when we get paid on our phones it&#x02019;s true we can&#x02019;t phone up no one here has landline phones&#x02019; (NHS Direct &#x02018;non-user&#x02019; FG8)</italic></p><p><italic>&#x02018;The area that we live a lot of people who do have mobile phones that are pay-as-you-go, and it&#x02019;s an extortionate amount that it costs on the phone. By the time you have got through your credit could go halfway through or even run out&#x02019; (NHS Direct &#x02018;non-user&#x02019;, FG7)</italic></p></disp-quote></p><p>Participants felt that if something was seriously wrong they would just phone &#x02018;999&#x02019; (emergency phone line in the UK) as this was a free number. The &#x02018;non-users&#x02019; felt that if NHS Direct was free to access they would be more likely to use the service. Although, there were discussions of concern that surrounded how the money to cover the cost of the call would be subsidised and if this would subsequently lead to further cuts to local NHS health services.<disp-quote><p><italic>&#x02018;If you are really poorly and you have a mobile phone and you have no credit on there then you can&#x02019;t ring NHS Direct but you can ring 999 and get an ambulance to you for free&#x02019; (NHS Direct &#x02018;non-user&#x02019;, FG7)</italic></p></disp-quote></p><p>In contrast, the majority of the Mendip participants did not mention the cost of the telephone call throughout any of the focus group session. However, many did not use mobile phones and they all had access to a landline phone. At the end of the focus group the researcher explained that the calls are charged at a national rate and the cost may be substantially higher when using a mobile phone, but no participants advised that this would impact on their decision to use this service.<disp-quote><p><italic>&#x02018;Well not if it&#x02019;s an emergency you would just pay it&#x02019; (NHS Direct &#x02018;non-user&#x02019;, FG6)</italic></p></disp-quote></p></sec><sec id="Sec14"><title>Ease of use</title><p>All of the participants who had used NHS Direct found the service easy to use with many participants highlighting that it was easier to use than using conventional out of hour&#x02019;s services e.g. GP co-operatives, Accident and Emergency, pharmacies. The main benefit disclosed was that you would not have to leave the house.<disp-quote><p><italic>&#x02018;You don&#x02019;t have to go through the process of packing and putting everyone in your car. You don&#x02019;t have to leave all the children with such and such the ability to have to deal with the problem without having to up sticks also if you are on your own. If you feel rubbish you wouldn&#x02019;t get in the car and drive&#x02019; (NHS Direct &#x02018;user&#x02019;, FG2)</italic></p></disp-quote></p><p>However, for participants in the Manchester sample there was a mixed response. Whilst there were a few participants who felt that they would find NHS Direct easy to use, the majority felt that to use the telephone would involve many deterring issues. For example, there was a perception through prior experiences of use of telephone services that there were too many options which would make it more complicated to use.<disp-quote><p><italic>&#x02018;It&#x02019;s supposed to make life easier but I spoke to a friend of mine who has used it because she&#x02019;s a mum and she had to press that many options that she found it easier to get the doctors to come out than use NHS Direct (NHS Direct &#x02018;non-user&#x02019;, FG7)</italic></p></disp-quote></p><p>Another perceived barrier which would impact on the ease of use, was the belief of being passed from person to person, which was felt as frustrating and would increase anxiety, especially when the call relates to an individual&#x02019;s health. There were a number of issues about speaking to somebody on the telephone as opposed to face-to-face. For example, one non-user was dyslexic and stated that he finds it easier to speak to his GP face-to-face due to the difficulties to express himself.<disp-quote><p><italic>&#x02018;I&#x02019;m dyslexic so it is better to see a doctor if I am ill so we can understand each other&#x02019; (NHS Direct &#x02018;non-user&#x02019;, FG9)</italic></p></disp-quote></p><p>Another issue related to language barriers. For example, not speaking English fluently was felt to impact negatively upon ease of use and confidence of using the service. The researcher did explain that NHS Direct did operate a translation service &#x02018;language line&#x02019;. However, none of the participants were aware that this service existed.<disp-quote><p><italic>&#x02018;Some people might not be able to call NHS Direct because some people can&#x02019;t speak English or their English isn&#x02019;t very good especially if someone is living on their own and their English isn&#x02019;t good or there&#x02019;s been very little English obviously they won&#x02019;t feel confident&#x02019; (NHS Direct &#x02018;non-user&#x02019;, FG9)</italic></p></disp-quote></p><p>Particularly for the Mendip sample, there were a range of barriers that would impact on ease of use. The biggest concern related to hearing, where many of the participants relied on using their hearing aids that made it difficult to communicate over the telephone. They felt that this would prove difficult when they have to explain symptoms when they could not hear what was being asked of them.<disp-quote><p><italic>&#x02018;Relies on the person giving the call giving accurate description of their symptoms so they&#x02019;re trying to explain how they feel and your elderly you can&#x02019;t hear very well and you&#x02019;re stressed and you&#x02019;re on your own it&#x02019;s not an ideal situation&#x02019; (NHS Direct &#x02018;non-user&#x02019;, FG5)</italic></p></disp-quote></p><p>Hearing was also a concern in relation to whom they would speak to. Participants from Mendip highlighted that they found foreign accents difficult to understand on the phone and often had to ask them to repeat themselves which they felt would prove difficult.<disp-quote><p><italic>&#x02018;I know there have been instances where you have been confronted by an Asian voice which is incredibly difficult to understand what she was saying which can be a massive language barrier&#x02019; (NHS Direct &#x02018;non-user&#x02019;, FG4)</italic></p></disp-quote></p><p>Participants from Mendip also discussed technical issues. For example, one participant from Creech, stated that there are a lot of technical issues related to the use of the telephone such as being cut off.<disp-quote><p><italic>&#x02018;In my opinion there is a lot of technical issues with the phone for example the line went dead so what do you do in that situation&#x02019; (NHS Direct &#x02018;non-user&#x02019;, FG5)</italic></p></disp-quote></p><p>Other physiological barriers related to memory, which was also suggested to impact on the ease of use.<disp-quote><p><italic>&#x02018;People with memory problems wouldn&#x02019;t be able to think or remember what to do, where to get the number etc.&#x02019; (NHS Direct &#x02018;non-user&#x02019;, FG6)</italic></p></disp-quote></p></sec><sec id="Sec15"><title>Time/speed</title><p>For NHS Direct &#x02018;users&#x02019;, speed to obtain healthcare advice was the key advantage of the service, whereby the majority of participants viewed this service to provide &#x02018;instant advice and reassurance&#x02019;, and valued being able to speak to a trained nurse or healthcare professional quickly.<disp-quote><p><italic>&#x02018;They give you immediate feedback on what you need to do when you are in that situation&#x02019; (NHS Direct &#x02018;user&#x02019;, FG2)</italic></p></disp-quote></p><p>However, some NHS Direct &#x02018;users&#x02019; did not agree with this perspective, and had some negative experiences that related to the amount of time it took to be called back by a nurse, and the time of day that they were called back e.g. being called during the middle of the night. For some participants, to wait a long time was perceived as reassurance, as it reflected that they were considered to be a low priority in terms of concern for their health condition.<disp-quote><p><italic>&#x02018;Apart from sometimes NHS Direct have taken 8&#x000a0;hours to phone me back I could have had an appointment in that time&#x02019; (NHS Direct &#x02018;user&#x02019;, FG2)</italic></p></disp-quote></p><p>NHS Direct &#x02018;non-users&#x02019; from Manchester and Mendip felt that waiting was a core barrier to use the service, whereby there was a distinct preference for instant face-to-face healthcare. Many of the participants shared concerns about the wait to be called back and did not like the thought of to wait on the telephone for long periods. There was a perception that NHS Direct was seen as a side step of out-of-hours care so was seen as &#x02018;wasting time&#x02019;.<disp-quote><p><italic>&#x02018;I know a young carer she&#x02019;s 24 looking after her mum with dementia who has seizures and every time she has got through (to NHS Direct) she has said it has been quicker to find a doctor and the doctors come out quicker than that because when her mum is bad she can&#x02019;t be spending 10&#x000a0;min on the phone&#x02019; (NHS Direct &#x02018;non-user&#x02019;, FG8)</italic></p></disp-quote></p><p>However in contrast, two &#x02018;non-users&#x02019;, from Longsight, Manchester, felt that NHS Direct could save time to provide instant reassurance instead of going straight to an Accident and Emergency Department in a hospital.<disp-quote><p><italic>&#x02018;Accident and emergency is reduced (and you) save time&#x02019; (NHS Direct &#x02018;non-user&#x02019;, FG9)</italic></p><p><italic>&#x02018;NHS Direct is more instant if a person does have a problem&#x02019; (NHS Direct &#x02018;non-user&#x02019;, FG9)</italic></p></disp-quote></p></sec><sec id="Sec16"><title>Communication and non-face-to-face healthcare</title><p>NHS Direct &#x02018;users&#x02019; felt the service gave them reassurance and enabled them to make the decision whether to escalate their health concerns or not. They also felt it gave them the reassurance that they had sought advice from a trained healthcare professional. None of the NHS Direct &#x02018;users&#x02019; were concerned that it was not a face-to-face service. In fact, many &#x02018;users&#x02019; highlighted that they preferred the lack of face-to-face contact, and viewed the service as both personable and professional which provided them with the level of reassurance they needed.<disp-quote><p><italic>&#x02018;I think the relief that it gives you in order to have someone to speak to and that you have actually looked into it. It&#x02019;s now like you can now get on and follow the guidance but knowing that it is the trained nurse that phones you back is just useful&#x02019; (NHS Direct &#x02018;user&#x02019; FG1)</italic></p></disp-quote></p><p>Conversely, &#x02018;non-users&#x02019; from both Manchester and Mendip outlined an overarching preference for face-to-face healthcare. &#x02018;Non-users&#x02019; felt that face-to-face healthcare offered more reassurance than speaking to somebody on the telephone. They also felt that if it was face-to-face they would be able to express themselves better and would feel more at ease to ask questions.<disp-quote><p><italic>&#x02018;If you felt that you needed reassurance you just take your children or yourself to hospital at least that way they can see you face-to-face or get the paramedic out then they would make that decision if you need to go to hospital&#x02026;..to be honest face-to-face is really important because this is what reassures you and this has to be the best option&#x02019; (NHS Direct &#x02018;non-user&#x02019;, FG7)</italic></p><p><italic>&#x02018;You can&#x02019;t talk about that you have got a high pressure you can&#x02019;t do that over the phone&#x02026;often physical symptoms are important aren&#x02019;t they so I think it&#x02019;s very necessary to see a doctor face-to-face&#x02019; (NHS Direct &#x02018;non-user&#x02019;, FG4)</italic></p></disp-quote></p><p>There were strong positive attitudes towards face-to-face communication. It was felt important by &#x02018;non-users&#x02019; that an individual could express themselves through body language. It was also more personable when speaking to someone face-to-face. Participants agreed that personal face-to-face interaction with a healthcare professional is an integral aspect when seeking healthcare advice, which presented a barrier to using telephone-based health services such as NHS Direct.<disp-quote><p><italic>&#x02018;Seeing someone in person is friendlier like if you went to see someone and you talk to them you can see them and see them smiling at you and treated sympathetically but on the phone it&#x02019;s different you don&#x02019;t see&#x02026;.I just think it is more personal rather than the telephone&#x02019; (NHS Direct &#x02018;non-user&#x02019;, FG9)</italic></p></disp-quote></p></sec></sec><sec id="Sec17" sec-type="discussion"><title>Discussion</title><p>This study has explored the barriers and facilitators to use NHS Direct, a hitherto under researched area. This research has uncovered explanations for usage and non-usage of NHS Direct. The core themes which emerged from the focus group discussions were related to awareness, costs to the individual, time/speed of the service and the acceptability of non-face-to-face communication. This research highlights that participants&#x02019; views on self-care varies by age, ethnicity and socio-demographic factors [<xref ref-type="bibr" rid="CR16">16</xref>].</p><p>NHS Direct &#x02018;users&#x02019; held a good awareness of all services that NHS Direct provide. However, there was a distinct lack of awareness among the &#x02018;non-users&#x02019;. Whilst many individuals from both Mendip and Manchester had heard of NHS Direct through media and out-of-hours signposting, there was a clear misunderstanding, with many who believed that it was a walk in service that operates out-of-hours. This supports research that has suggested awareness of this service is low [<xref ref-type="bibr" rid="CR43">43</xref>,<xref ref-type="bibr" rid="CR49">49</xref>] which indicates that the impact of previous advertising campaigns has been largely unsuccessful in reaching all sections of the population. It is clear that awareness is a core mechanism which impacts on health service uptake [<xref ref-type="bibr" rid="CR50">50</xref>] and, therefore, these findings reinforce the importance to provide clear information through tailored promotional campaigns to ensure all sections of the population are informed.</p><p>NHS Direct &#x02018;users&#x02019; suggested that they did not view the cost to use NHS Direct was a barrier, with many not aware of the cost implications to use the service. Conversely, &#x02018;non-users&#x02019; from Manchester felt the service was extremely costly, especially as many relied on pay as you go mobile phones. This view was not reported by &#x02018;non-users&#x02019; from Mendip, which suggests that the cost of the service appears to be an access barrier for those in deprived communities who are unable to afford to use the service. As such, it appears that NHS Direct and other telephone-based services should be aware of the impact that cost may have on uptake by individuals from more deprived communities. Nonetheless, as the new non-emergency &#x02018;111&#x02019; NHS phone line is rolled out nationally as a free service it will become even more important to communicate to the public that service has no cost, so this should not be a barrier to access.</p><p>A particular advantage of NHS Direct for &#x02018;users&#x02019; was that the service was accessible and easy to use. However, the predominantly older Mendip sample felt that there would be issues that relate to hearing and memory that would impact on discussing healthcare information via the telephone. Older peoples&#x02019; access to modern technology has been extensively debated with research that suggests that, not only physiological changes associated with ageing such as decrements of sight, hearing, dexterity, motor functioning, co-orientation and cognitive processing can impact on newer models of healthcare [<xref ref-type="bibr" rid="CR51">51</xref>-<xref ref-type="bibr" rid="CR53">53</xref>], but also a wide range of psycho-social factors. For example, uptake has been strongly dependent on income, education, experiences, and attitudes [<xref ref-type="bibr" rid="CR54">54</xref>], with confidence that relates to ease of use, shown to influence significantly older people&#x02019;s adoption and use of new technology [<xref ref-type="bibr" rid="CR54">54</xref>-<xref ref-type="bibr" rid="CR56">56</xref>]. There is an assumption that there should be a &#x02018;universal&#x02019; take-up of technology [<xref ref-type="bibr" rid="CR57">57</xref>,<xref ref-type="bibr" rid="CR58">58</xref>]. Whilst this assumption is challenged [<xref ref-type="bibr" rid="CR59">59</xref>], access to technology driven healthcare can be increased through two main ways: (1) ensure that the service is easy to use, and (2) through the provision of tailored information to enhance awareness of such services within the UK&#x02019;s diverse population.</p><p>A key advantage for &#x02018;users&#x02019; was that NHS Direct was a quick way to access advice and health information. However, &#x02018;non-users&#x02019; discussed the preference for &#x02018;instant&#x02019; face-to-face reassurance with NHS Direct viewed as a diversion. On the other hand, &#x02018;non-users&#x02019; suggested a clear preference for more traditional face-to-face health services both in and out-of-hours. This appears to support previous literature that has identified that older people [<xref ref-type="bibr" rid="CR30">30</xref>,<xref ref-type="bibr" rid="CR60">60</xref>], ethnic minority groups [<xref ref-type="bibr" rid="CR32">32</xref>,<xref ref-type="bibr" rid="CR61">61</xref>] alongside those from socially deprived communities [<xref ref-type="bibr" rid="CR62">62</xref>], prefer and have more confidence with face-to-face healthcare communication. This could also relate to the fact that ethnic minority groups [<xref ref-type="bibr" rid="CR63">63</xref>], older people and those who are from more deprived backgrounds prefer doctor-centred healthcare [<xref ref-type="bibr" rid="CR64">64</xref>,<xref ref-type="bibr" rid="CR65">65</xref>] and prefer to take a passive role in their health.</p><p>Whilst this research provided a wide overview of the facilitators and barriers of a telephone-based healthcare service there were some limitations that are noteworthy. Firstly, the NHS Direct &#x02018;users&#x02019; focus groups only focused on one high &#x02018;user&#x02019; group i.e. females with young children. This was also reflected by an imbalance between the numbers of participants in the &#x02018;user&#x02019; versus &#x02018;non-user&#x02019; focus groups (17 v 54). This imbalance is an outcome of the breadth of issues uncovered in the &#x02018;non-user&#x02019; focus groups alongside the inclusion of &#x02018;user&#x02019; focus groups which captured a diversity of opinion through a wide range of geographical and socio-cultural factors. Nonetheless, focus groups from other &#x02018;user&#x02019; groups, such as younger adults aged 20&#x02013;34, may have provided further insight into the barriers and facilitators of such health services.</p><p>Secondly, whilst there was an attempt to capture ethnic diversity, this was only evident in one focus group. As such, future research should aim to examine the barriers and facilitators of such services accounting for a wider variation of ethnicity. In particular, studies should focus on other &#x02018;non-users&#x02019; (e.g. Eastern European, Chinese and Black African [<xref ref-type="bibr" rid="CR24">24</xref>]) to determine the range of cultural factors that impact on the engagement of telephone-based healthcare. Finally, some of the focus groups were existing groups, in particular the Mendip sample. There were clear challenges to recruit older participants, and whilst this may have created some bias, it showed to be a useful way to reach a &#x02018;hard to reach&#x02019; community sample.</p></sec><sec id="Sec18" sec-type="conclusion"><title>Conclusions</title><p>This research uncovered a wide range of factors which impact on the uptake of NHS Direct. Acceptability of non-face-to-face healthcare was a key driver to use NHS Direct. Whilst &#x02018;users&#x02019; found the service both convenient and easy to use, &#x02018;non-users&#x02019; emphasised a clear preference for face-to-face healthcare. This was supported by a lack of confidence in discussing healthcare over the telephone, particularly in older groups who had cognitive and sensory difficulties. Awareness and cost also impacted on usage, whereby &#x02018;users&#x02019; showed a higher level of knowledge and awareness of the service. Conversely, &#x02018;non-users&#x02019; had a low awareness and the cost of phoning a premium &#x02018;0845&#x02019; number was also viewed as a barrier, particularly from those in deprived communities who rely on a mobile phone.</p><p>It is apparent that although some barriers are the same for both groups of &#x02018;non-users&#x02019; in Mendip and Manchester, there are some differences. This suggests that a one size fits all approach cannot be adopted. Instead socio-demographic factors need to be taken into account to identify the barriers to enable the service to become more accessible to all communities. Therefore, if other similar services such as the new &#x02018;111&#x02019; service are to become a more widely used model of remote healthcare then it is essential that the barriers and facilitators to access telephone-based services are addressed. Increased access will subsequently improve the patient experience and the urgent care pathway. In turn this will reduce the need of unnecessary visits to already overstretched healthcare services. A recognition of the factors that do and do not make people access and use services such as NHS Direct, will help to mobilise patients towards the self-care model and support them in to take responsibility for their own care [<xref ref-type="bibr" rid="CR3">3</xref>,<xref ref-type="bibr" rid="CR11">11</xref>,<xref ref-type="bibr" rid="CR12">12</xref>].</p></sec>

Ovarian metastasis from thyroid carcinoma: a case report and literature review

Could not extract abstract

<contrib contrib-type="author" corresp="yes"><name><surname>Corrado</surname><given-names>Giacomo</given-names></name><address><email>giacomo.corrado@alice.it</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author"><name><surname>Pomati</surname><given-names>Giulia</given-names></name><address><email>giuliapomati@tiscali.it</email></address><xref ref-type="aff" rid="Aff2"/></contrib><contrib contrib-type="author"><name><surname>Russo</surname><given-names>Andrea</given-names></name><address><email>russo71@ifo.it</email></address><xref ref-type="aff" rid="Aff3"/></contrib><contrib contrib-type="author"><name><surname>Visca</surname><given-names>Paolo</given-names></name><address><email>visca@ifo.it</email></address><xref ref-type="aff" rid="Aff3"/></contrib><contrib contrib-type="author"><name><surname>Vincenzoni</surname><given-names>Cristina</given-names></name><address><email>vincenzoni@ifo.it</email></address><xref ref-type="aff" rid="Aff4"/></contrib><contrib contrib-type="author"><name><surname>Patrizi</surname><given-names>Lodovico</given-names></name><address><email>lodovicop@libero.it</email></address><xref ref-type="aff" rid="Aff2"/></contrib><contrib contrib-type="author"><name><surname>Vizza</surname><given-names>Enrico</given-names></name><address><email>e.vizza@ifo.it</email></address><xref ref-type="aff" rid="Aff4"/></contrib><aff id="Aff1"><label/>Department of Oncological Surgery, Gynecologic Oncology Unit, &#x0201c;Regina Elena&#x0201d; National Cancer Institute, Via Elio Chianesi 53, 00144 Rome, Italy </aff><aff id="Aff2"><label/>Surgery Department, Gynecology Section and Obstetrics, Tor Vergata University, Rome, Italy </aff><aff id="Aff3"><label/>Pathology Department, &#x0201c;Regina Elena&#x0201d; National Cancer Institute, Rome, Italy </aff><aff id="Aff4"><label/>Surgery Department, Gynecologic Oncology Unit, &#x0201c;Regina Elena&#x0201d; National Cancer Institute, Rome, Italy </aff>

Diagnostic Pathology

<sec id="Sec1" sec-type="introduction"><title>Background</title><p>Papillary thyroid carcinoma is the most common histotype of thyroid carcinoma and it is associated to a good prognosis and to a loco regional spread. The presence of distant metastasis is an important prognostic factor, although it is a rare event. Distant metastasis from papillary thyroid carcinoma often occurs decades after the primary tumor and the 70% of patients who die for papillary thyroid carcinoma are disease free after the primary treatment. Moreover, the 30&#x000a0;years mortality rates increase to 43% as a result of a distant recurrence [<xref ref-type="bibr" rid="CR1">1</xref>,<xref ref-type="bibr" rid="CR2">2</xref>]. The most common metastatic sites are lung [<xref ref-type="bibr" rid="CR3">3</xref>] and, following, bone. Instead, rare metastatic sites are brain, parotid, breast, liver, kidney, adrenal glands, ovaries, muscle and skin [<xref ref-type="bibr" rid="CR4">4</xref>].</p><p>Ovaries are the most common metastatic sites from both genital and extragenital primaries, mostly originating in the gastrointestinal tract, and ovarian metastasis represent about 5% to 30% of all ovarian tumors [<xref ref-type="bibr" rid="CR5">5</xref>,<xref ref-type="bibr" rid="CR6">6</xref>].</p><p>We report a rare case of ovarian metastasis from thyroid carcinoma after 9&#x000a0;years from the diagnosis.</p></sec><sec id="Sec2"><title>Case presentation</title><p>In December 2013 a 51&#x000a0;years old woman presented to our Gynecologic Oncology Unit, for the presence of a pelvic mass originating from the left ovary, occasionally detected in the ultrasound imaging during a routine check. She was 1 gravida, 1 para, with no previous gynecological pathology in her history. She referred that in 2004, following diagnosis of thyroid carcinoma, she underwent a total thyroidectomy in another hospital. The histological examination revealed a papillary thyroid carcinoma, follicular variant, involving the left thyroid lobe. She received a radio-iodine metabolic adjuvant treatment by administration of 120&#x000a0;mCi ¹³¹I.</p><p>No evidence of disease was detected during follow up until December 2013 when an ovarian mass was revealed by ultrasound imaging and at the magnetic resonance it measured 76&#x02009;&#x000d7;&#x02009;46&#x02009;&#x000d7;&#x02009;62&#x000a0;mm (Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>). Normal value resulted for Ca125 and HE4 (respectively 23.9 UI/ML and 100 pmol/L), while thyroglobulin was detectable (0.2&#x000a0;ng/ml). She underwent laparoscopic bilateral salpingo-oophorectomy. The histological exam showed a papillary thyroid carcinoma involving left ovary. The ovarian tumor measured 10&#x02009;&#x000d7;&#x02009;5&#x02009;&#x000d7;&#x02009;6&#x000a0;cm and weighed 700&#x000a0;g. The section surface was solid and brown. We have performed one sample per centimeter of maximum dimension. Microscopic examination showed that ovarian parenchyma was nearly entirely occupied by thyroid-type neoplasm (Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>A-B) characterized by round follicles, of any size, and papillae (Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>C). Many follicles were lined by cuboidal, epithelial cells with moderate amounts of cytoplasm and round to oval and ground glass nuclei that exhibited frequent mitotic figures (Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>D). These cells were also positive to TTF-1 and Thyroglobulin antibodies (Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>E-F). Moreover, Keratin-19 (CK-19) e HBME-1 were positive while Galectin-3 (GAL-3) was negative (Figure&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>A-B-C).<fig id="Fig1"><label>Figure 1</label><caption><p> <bold>Pelvic magnetic resonance. A)</bold> Axial fat suppressed T1-weighted image after intravenous gadolinium enhancement. <bold>B)</bold> Sagittal T2-weighted image showing left solid ovarian mass.</p></caption><graphic xlink:href="13000_2014_193_Fig1_HTML" id="MO1"/></fig><fig id="Fig2"><label>Figure 2</label><caption><p> <bold>Microscopic examination and immunohistochemical stains. A)</bold> The ovarian parenchyma is occupied by a thyroid type neoplasm. Note the follicles. (HE 40&#x000d7;). <bold>B)</bold> Massive extension of the neoplasm in the parenchyma. There is no evidence of benign struma ovary or others components of teratoma (HE 40&#x000d7;). <bold>C)</bold> Follicular and papillary components of neoplasm (HE 100&#x000d7;). <bold>D)</bold> Papillary component of neoplasm. Papillae are lined by cells with ground glass nuclei. There are also some mitoses (HE 400&#x000d7;). <bold>E)</bold> Positive reaction of thyroglobulin antibody (HE 400&#x000d7;). <bold>F)</bold> Nuclear positive reaction of TTF-1 antibody (HE 100&#x000d7;).</p></caption><graphic xlink:href="13000_2014_193_Fig2_HTML" id="MO2"/></fig><fig id="Fig3"><label>Figure 3</label><caption><p> <bold>Other immunohistochemical characteristics. A)</bold> Positive reaction for CK-19 (HE 100&#x000d7;). <bold>B)</bold> Positive reaction for HBME-1 (HE 100&#x000d7;). <bold>C)</bold> Negative reaction for GAL-3 (HE 100&#x000d7;).</p></caption><graphic xlink:href="13000_2014_193_Fig3_HTML" id="MO3"/></fig></p><p>This neoplasm was suggestive for a metastasis because there was no evidence of benign struma ovary and the others teratomatous component.</p><p>Unfortunately, in March 2014 the CT/PET detected a left pelvic lymph nodes recurrence (SUV 10.8) and a paramedian nodular mass in proximity of the uterus (SUV 10.6). A laparoscopic evaluation showed left pelvic peritoneal carcinomatosis and a large left pelvic adenopathy (Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>). A left pelvic lymphadenectomy and a left pelvic peritonectomy were performed. Definitive histological examination showed a metastasis from papillary thyroid carcinoma. Actually, the patient is undergoing biological treatment with multikinase inhibitors.<fig id="Fig4"><label>Figure 4</label><caption><p> <bold>Laparoscopic and CT/PET characterization of the pelvic recurrence. A)</bold> Laparoscopic image of left pelvic peritoneal carcinomatosis. <bold>B)</bold> The large left pelvic lymphadenopathy. <bold>C)</bold> CT/PET image of a nodular mass near the uterus (SUV 10.6). <bold>D)</bold> Left pelvic lymph nodes (SUV 10.8).</p></caption><graphic xlink:href="13000_2014_193_Fig4_HTML" id="MO4"/></fig></p></sec><sec id="Sec3" sec-type="conclusion"><title>Conclusions</title><p>Papillary thyroid carcinoma is associated with a good prognosis and with a low metastatic power. A distant metastasis from papillary thyroid carcinoma is a rare event, above all when the recurrence occurs in less common sites. For this reason, rare metastasis is often not considered during the clinical setting.</p><p>When an ovarian mass is found to contain cells with features of thyroid carcinoma, a differential diagnoses should have to be considered between thyroid cancer arising from a struma ovarii and ovarian metastasis originating from a primary thyroid carcinoma, since the prognosis and clinical management are different. Thyroid carcinoma originating from a struma ovarii, presenting a papillary histotype in 70% of all cases, is reported to occur much more commonly than an ovarian metastasis from the thyroid. As a matter of fact, struma ovarii are the 5% of ovarian teratomas, 5&#x02013;10% results in malignant teratomas and metastatic diseases doesn&#x02019;t reach the 23% of cases. However, when no teratomatous elements and no normal thyroid epithelial tissue are detected in the ovarian lesion, the diagnosis of metastasis with a thyroid origin is suggestive [<xref ref-type="bibr" rid="CR7">7</xref>]. In our patient, the ovarian parenchyma was completely occupied by thyroid-type neoplasm, there was no evidence of benign struma ovary or others teratomatous component and cells were positive to TTF-1 and Thyroglobulin antibodies.</p><p>A review of literature from 1929 to 2013 can confirm the rarity of the ovarian metastasis from thyroid carcinoma. As it is shown in Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>, only four case reports of ovarian metastasis from thyroid are described in a comprehensive manner [<xref ref-type="bibr" rid="CR8">8</xref>-<xref ref-type="bibr" rid="CR11">11</xref>]. The table shows that most of patients affected by thyroid carcinoma were between the fourth and fifth decades of life at the moment of the first diagnosis and underwent I¹³¹ therapy after primary surgery. Moreover, ovarian metastasis seems to appear more commonly unilaterally. It can be inferred from the description of these cases that well differentiated thyroid carcinomas can give metastasis many years after the primary tumor. As a matter of fact, in Brogioni S et al. [<xref ref-type="bibr" rid="CR10">10</xref>] case report, the ovarian metastasis occurred almost 5&#x000a0;years after the first pulmonary metastasis and 7&#x000a0;years after the first diagnosis. Also in Pirvu A et al. [<xref ref-type="bibr" rid="CR11">11</xref>] report, the pulmonary metastasis occurred shortly after the thyroidectomy, while ovarian metastasis 11&#x000a0;years after the first diagnosis of thyroid carcinoma. Furthermore, in the well differentiated thyroid cancer group, papillary histotype seems to give ovarian metastasis more frequently than follicular. In our case ovaries had been the first metastatic site, while in three of the mentioned reports [<xref ref-type="bibr" rid="CR8">8</xref>,<xref ref-type="bibr" rid="CR10">10</xref>,<xref ref-type="bibr" rid="CR11">11</xref>], the ovarian metastasis was associated with a metastatic spread, probably pointing to a biologically more aggressive disease and to a worse prognosis associated to the ovarian involvement. Further 10 cases [<xref ref-type="bibr" rid="CR12">12</xref>-<xref ref-type="bibr" rid="CR18">18</xref>] of ovarian spread from thyroid carcinoma are mentioned in literature but, unfortunately, no more details were provided. Besic et al. [<xref ref-type="bibr" rid="CR12">12</xref>] in his autoptic series reported one case of ovarian metastasis from anaplastic thyroid carcinoma, while Silvesberg et al. [<xref ref-type="bibr" rid="CR13">13</xref>], always in autoptic series, reported two cases of ovarian metastasis from anaplastic thyroid carcinoma and one from medullary thyroid carcinoma. Others two cases of ovarian metastasis from medullary thyroid carcinoma are only mentioned by Ibanez et al. [<xref ref-type="bibr" rid="CR14">14</xref>] and Gordon et al. [<xref ref-type="bibr" rid="CR15">15</xref>]. In another article [<xref ref-type="bibr" rid="CR17">17</xref>] the ovarian involvement was bilateral. However, in literature there are not enough studies to draw conclusions about prognosis and best clinical management of ovarian metastasis from thyroid cancer. ¹³¹I scan and serum thyroglobulin are widely employed during the follow up of thyroid cancer and in the assessment of the best therapy to use after surgery, while immunohistochemical stain for thyroglobulin and TTF-1 is often essential in pathologic diagnosis as it has been in our experience. Moreover, Keratin-19 (CK-19) e HBME-1 were positive while Galectin-3 (GAL-3) was negative. This was due to because GAL-3 is a useful marker for diagnosis of low grade thyroid carcinomas [<xref ref-type="bibr" rid="CR19">19</xref>] while in our case the carcinoma was an high grade.<table-wrap id="Tab1"><label>Table 1</label><caption><p> <bold>Literature review of ovarian metastasis from thyroid carcinoma</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th> <bold>Author</bold> </th><th> <bold>Year</bold> </th><th> <bold>N&#x000b0;</bold> </th><th> <bold>Age</bold> </th><th> <bold>Primary treatment</bold> </th><th> <bold>Histotype</bold> </th><th> ^{<bold>131</bold>} <bold>I therapy</bold> </th><th> <bold>DFS (months)</bold> </th><th> <bold>Site of metastasis</bold> </th><th> <bold>Surgery of metastasis</bold> </th><th> ^{<bold>131</bold>} <bold>I therapy</bold> </th><th> <bold>Status (months)</bold> </th></tr></thead><tbody><tr valign="top"><td>Young RH [<xref ref-type="bibr" rid="CR8">8</xref>]</td><td>1994</td><td>1</td><td>17</td><td>Partial thyroidectomy</td><td>Follicular</td><td>-</td><td>144</td><td>Brain, ovaries</td><td>Right cystectomy</td><td>-</td><td>DOD, 150</td></tr><tr valign="top"><td>Logani S [<xref ref-type="bibr" rid="CR9">9</xref>]</td><td>2001</td><td>1</td><td>34</td><td>Total thyroidectomy with lymphadenectomy</td><td>Papillary</td><td>Yes</td><td>132</td><td>Ovaries</td><td>Left oophorectomy</td><td>Yes</td><td>NED, 140</td></tr><tr valign="top"><td>Brogioni S [<xref ref-type="bibr" rid="CR10">10</xref>]</td><td>2007</td><td>1</td><td>38</td><td>Total thyroidectomy, with lymphadenectomy</td><td>Papillary</td><td>Yes</td><td>24</td><td>Thymus, lungs, ovaries, brain</td><td>Bilateral oophorectomy</td><td>Yes</td><td>DOD, 92</td></tr><tr valign="top"><td>Pirvu A [<xref ref-type="bibr" rid="CR11">11</xref>]</td><td>2013</td><td>1</td><td>26</td><td>Total thyroidectomy with lymphadenectomy</td><td>Papillary</td><td>Yes</td><td>-</td><td>Lungs, ovaries</td><td>Left ovariectomy</td><td>Yes</td><td>AWD, 158</td></tr><tr valign="top"><td>Our experience</td><td>2014</td><td>1</td><td>42</td><td>Total tyroidectomy</td><td>Papillary</td><td>Yes</td><td>108</td><td>Ovaries</td><td>Laparoscopic bilateral oophorectomy</td><td>No</td><td>AWD, 111</td></tr></tbody></table><table-wrap-foot><p>NED&#x02009;=&#x02009;no evidence of disease. DOD&#x02009;=&#x02009;death of disease. AWD&#x02009;=&#x02009;alive with disease.</p></table-wrap-foot></table-wrap></p><p>In conclusion, the ovarian involvement by a primary thyroid cancer is a rare event, but it should be considered, since it seems to be a negative prognostic factor worsening the oncological outcome. The histopathologic evaluation, including immunohistochemical stain and the investigation of patient&#x02019;s history are crucial steps in the diagnosis and clinical management of ovarian metastasis from thyroid cancer.</p></sec><sec id="Sec4"><title>Consent</title><p>Written informed consent was obtained by patient for publication of this report and any accompanying images. A copy of the written consent is available for review by the Editor-in-Chief of this journal.</p></sec>

Del Nido Cardioplegia can be safely administered in high-risk coronary artery bypass grafting surgery after acute myocardial infarction: a propensity matched comparison

Could not extract abstract

<contrib contrib-type="author" deceased="no" equal-contrib="no"><name><surname>Yerebakan</surname><given-names>Halit</given-names></name><address><email>drhalid@gmail.com</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author" deceased="no" equal-contrib="no"><name><surname>Sorabella</surname><given-names>Robert A</given-names></name><address><email>ras2112@cumc.columbia.edu</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author" deceased="no" equal-contrib="no"><name><surname>Najjar</surname><given-names>Marc</given-names></name><address><email>mn2594@cumc.columbia.edu</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author" deceased="no" equal-contrib="no"><name><surname>Castillero</surname><given-names>Estibaliz</given-names></name><address><email>ec2929@cumc.columbia.edu</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author" deceased="no" equal-contrib="no"><name><surname>Mongero</surname><given-names>Linda</given-names></name><address><email>mongero@nyp.org</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author" deceased="no" equal-contrib="no"><name><surname>Beck</surname><given-names>James</given-names></name><address><email>beckjam@nyp.org</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author" deceased="no" equal-contrib="no"><name><surname>Hossain</surname><given-names>Maliha</given-names></name><address><email>mh3328@cumc.columbia.edu</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author" deceased="no" equal-contrib="no"><name><surname>Takayama</surname><given-names>Hiroo</given-names></name><address><email>ht2225@cumc.columbia.edu</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author" deceased="no" equal-contrib="no"><name><surname>Williams</surname><given-names>Mathew R</given-names></name><address><email>mw2@cumc.columbia.edu</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author" deceased="no" equal-contrib="no"><name><surname>Naka</surname><given-names>Yoshifumi</given-names></name><address><email>yn33@cumc.columbia.edu</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author" deceased="no" equal-contrib="no"><name><surname>Argenziano</surname><given-names>Michael</given-names></name><address><email>ma66@cumc.columbia.edu</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author" deceased="no" equal-contrib="no"><name><surname>Bacha</surname><given-names>Emile</given-names></name><address><email>eb2709@cumc.columbia.edu</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author" deceased="no" equal-contrib="no"><name><surname>Smith</surname><given-names>Craig R</given-names></name><address><email>crs2@cumc.columbia.edu</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author" corresp="yes" deceased="no" equal-contrib="no"><name><surname>George</surname><given-names>Isaac</given-names></name><address><email>ig2006@cumc.columbia.edu</email></address><xref ref-type="aff" rid="Aff1"/></contrib><aff id="Aff1">Department of Surgery, Division of Cardiothoracic Surgery, College of Physicians and Surgeons of Columbia University, New York Presbyterian Hospital, MHB 7GN-435, 177 Fort Washington Ave, New York, 10032 NY USA </aff>

Journal of Cardiothoracic Surgery

<sec id="Sec1"><title>Background</title><p>Despite advances in surgical technique and patient selection, mortality after coronary artery bypass grafting surgery (CABG) for acute myocardial infarction (AMI) remains high at 4-10% [<xref ref-type="bibr" rid="CR1">1</xref>],[<xref ref-type="bibr" rid="CR2">2</xref>]. Myocardial protection in this setting is complicated by subsequent ischemia-reperfusion injury, oxidative stress and intracellular Ca²⁺ overload [<xref ref-type="bibr" rid="CR3">3</xref>], all of which may contribute to post-operative myocardial dsyfunction. Current cardioplegia options include whole blood cardioplegia (WB), Buckberg solution, and warm blood cardioplegia. However, conflicting evidence exists regarding the superiority of one solution compared to another. Moreover, solutions designed specifically to address metabolic changes after AMI or cardiogenic shock, such as Buckberg solution [<xref ref-type="bibr" rid="CR4">4</xref>], can be cumbersome to deliver and are not universally adopted.</p><p>Del Nido (DN) cardioplegia was formulated to act as single-dose administration in pediatric patients. Compared to the traditional 4:1 blood cardioplegia, DN is more dilute (1:4, blood:crystalloid), has lower Ca⁺², and contains lidocaine (140 mg/L) (Table <xref rid="Tab1" ref-type="table">1</xref>) (Compass-Baxter Healthcare Inc., Edison, NJ). In pediatric patients, DN cardioplegia has been shown to result in lower postoperative troponin release compared to WB cardioplegia [<xref ref-type="bibr" rid="CR5">5</xref>]. However, its use in adults has not been studied, and its unknown effects on the complex derangements after AMI have led surgeons to question its use in this sick patient population. We sought to evaluate the clinical outcomes of DN cardioplegia in CABG after AMI compared to standard 4:1 WB solution.<table-wrap id="Tab1"><label>Table 1</label><caption><p> <bold>Composition of Whole Blood (WB) and del Nido (DN) cardioplegia solutions</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr><th/><th>4:1 Blood: Cardiogplegia</th><th>Del Nido</th></tr></thead><tbody><tr><td>Na (mmol/L)</td><td>136-152</td><td>143-153</td></tr><tr><td>K (mmol/L)</td><td>24</td><td>24</td></tr><tr><td>Cl (mmol/L)</td><td>126</td><td>132</td></tr><tr><td>Mg (mmol/L)</td><td>2</td><td>6.2</td></tr><tr><td>Ca (mmol/L)</td><td/><td>0.4</td></tr><tr><td>Lidocaine (mg/L)</td><td>200 (before XC release)</td><td>140</td></tr><tr><td>Mannitol (g/L)</td><td>12.5</td><td>2.6</td></tr><tr><td>NaHCO3</td><td>50</td><td>26</td></tr></tbody></table></table-wrap></p></sec><sec id="Sec2"><title>Methods</title><sec id="Sec3"><title>Patient population</title><p>Between July 2010 and July 2012, a total of 828 consecutive patients underwent CABG with or without concomitant cardiac surgical procedures at our institution. Out of 828 patients, 88 consecutive patients with AMI undergoing isolated CABG surgery with cardioplegic arrest were identified. AMI was defined as topononin I &#x0003e; 1 ng/mL within one week prior to operation. Demographic and clinical outcome data were retrospectively collected from chart review, including age, sex, race, body mass index (BMI), and comorbid medical conditions, including diabetes, hypertension, chronic obstructive pulmonary disease, renal failure, smoking history, severe aortic wall calcification, peripheral arterial disease, cerebrovascular disease, atrial fibrillation, and prior AMI, percutaneous coronary intervention, or cardiac surgery. The preoperative clinical status, including Society of Thoracic Surgeons (STS) risk score for mortality, New York Heart Association (NYHA) symptom class, echocardiographic data, and requirement for intra-aortic balloon pump (IABP) before surgery, was documented. This study met all guidelines of the Institutional Review Board of Columbia University.</p></sec><sec id="Sec4"><title>Study design</title><p>In the current study, clinical outcomes before and after adoption of DN were compared using a propensity score matching analysis. The first group (n = 40) was composed of a 1-year cohort (July 2010-June 2011) of patients who underwent isolated CABG following AMI with exclusive use of WB cardioplegia for myocardial protection (WB group). The second group (n = 48) was a similar 1-year cohort (July 2011-July 2012) of patients who underwent isolated CABG following AMI with exclusive use of DN cardioplegia for myocardial protection (DN group). A propensity score matching, based on Greedy 5 to 1 digital matching algorithm [<xref ref-type="bibr" rid="CR6">6</xref>], was used to reduce major patient characteristic differences between groups. Propensity score matching (1:1) identified 40 matched pairs (in each group) for analysis.</p><p>The primary outcomes included low output syndrome (LCOS) and in-hospital mortality. The secondary outcomes were duration of cross-clamp (XC) and cardiopulmonary bypass (CPB), volume of cardioplegia used, red blood cell transfusion rate, and in-hospital complications. LCOS was defined as, if the patient required an IABP or mechanical circulatory support (MCS) or extra-corporeal membrane oxygenation (ECMO) in the operating room in order to be weaned from CPB or in the intensive care unit because of hemodynamic compromise. LCOS was also diagnosed if the patient required inotropic medication (at least two of either vasopressin, dobutamine, milrinone, or epinephrine) to maintain systolic blood pressure. Patients who required a renal dose of dopamine (&#x02264;3 &#x003bc;g/kg) or a single inotropic medication support were not considered to have LCOS.</p></sec><sec id="Sec5"><title>Operative details, cardioplegia, and complications</title><p>All surgeries were performed using a standard general anesthesia protocol, median sternotomy approach, employing cardiopulmonary bypass with mild systemic hypothermia (30 to 34&#x000b0;C). Intraoperative transesophageal echocardiography was routinely employed. Myocardial protection was achieved with either WB or DN cardioplegia as follows. In both groups, the heart was arrested with an induction dose (1 liter) of cold (4&#x000b0;C) cardioplegia using antegrade and/or retrograde delivery (see Table <xref rid="Tab1" ref-type="table">1</xref> for details). In addition, repeated doses of WB cardioplegia was given via saphenous vein grafts or through the retrograde cannula at 20 min intervals in WB patients only. A second dose (500 ml) of DN was only given if the XC exceeded 90 minutes.</p><p>The operative details that were collected included priority of surgery (elective, urgent, emergent), cardio-pulmonary bypass time, global ischemic time, amount and method of cardioplegia, and number of transfusions. Postoperative in-hospital complications included need for IABP, ECMO, or MCS, inotrope dependence on intensive care unit admission, ventricular or atrial arrhythmia, need for permanent pace maker, respiratory failure, renal failure, sepsis or endocarditis, sternal wound infection, gastrointestinal bleeding, stroke, MI, reoperation for bleeding, and death before discharge.</p></sec><sec id="Sec6"><title>Propensity score matching and statistical analysis</title><p>To permit an independent comparison, logistic regression and the Greedy 5 to1 Digit Match macro was used to generate the propensity scores employed for matching [<xref ref-type="bibr" rid="CR6">6</xref>]. The multivariable logistic regression was run to compute propensity scores for the two groups (WB and DN) based on the following covariates: age, gender, STS score, NYHA class, body mass index (BMI), diabetes, hypertension, hyperlipidemia, preoperative cardiogenic shock, congestive heart failure, preoperative left ventricular ejection fraction (LVEF), number of diseased vessels, renal failure, peripheral vascular disease, chronic obstructive pulmonary disease, and reoperative cardiac surgery. Each patient who had WB was matched to one patient who had DN cardioplegia with the closest propensity score.</p><p>Descriptive statistics were used to describe patient characteristics. Categorical data were represented as frequency distributions and percentages. Continuous variables were expressed as mean &#x000b1; standard deviation (SD). Univariate analysis of continuous variables was performed using student <italic>t</italic>-tests, whereas categorical variables were compared using chi-square test of homogeneity and independence in contingency tables. All p-values were two-sided. Data were analyzed using IBM SPSS Statistics 21.0 (IBM Corp., Armonk, NY, USA).</p></sec></sec><sec id="Sec7"><title>Results</title><sec id="Sec8"><title>Baseline demographics</title><p>The mean age of the patients was 67.7 &#x000b1; 12.6 years (range 28-93 years), 57.5% were male, and mean BMI was 28.2 &#x000b1; 5.7 kg/m². Mean maximum preoperative serum troponin level was 15.1 ng/ml (range 1-84.5 ng/ml). Mean STS risk score for mortality of all patients was 4.3% (range 0.4-31.1%). The baseline clinical characteristics of the 2 propensity-matched groups (WB vs. DN) were balanced in all measured characteristics and summarized in Table <xref rid="Tab2" ref-type="table">2</xref>.<table-wrap id="Tab2"><label>Table 2</label><caption><p> <bold>Baseline demographics</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr><th/><th>WB</th><th>DN</th><th> <bold><italic>p</italic></bold> </th></tr></thead><tbody><tr><td>n</td><td>40</td><td>40</td><td/></tr><tr><td>Age (years)</td><td>66.6 &#x000b1; 13.5</td><td>68.7 &#x000b1; 11.7</td><td>0.472</td></tr><tr><td>Gender (male) (%)</td><td>21 (45.7)</td><td>25 (54.3)</td><td>0.366</td></tr><tr><td>BMI (kg/m²)</td><td>28.2 &#x000b1; 5.9</td><td>28.1 &#x000b1; 5.7</td><td>0.926</td></tr><tr><td>STS risk score (%)</td><td>4.6 &#x000b1; 6.9</td><td>3.9 &#x000b1; 4.2</td><td>0.545</td></tr><tr><td>Troponin I level (mean) (ng/ml)</td><td>15.6 &#x000b1; 23</td><td>14.8 &#x000b1; 21</td><td>0.915</td></tr><tr><td>NYHA class (%)</td><td/><td/><td/></tr><tr><td>1</td><td>5 (12.5)</td><td>6 (15)</td><td rowspan="4">0.921</td></tr><tr><td>2</td><td>22 (55)</td><td>21 (52.5)</td></tr><tr><td>3</td><td>12 (30)</td><td>11 (27.5)</td></tr><tr><td>4</td><td>1 (2.5)</td><td>2 (5)</td></tr><tr><td>Diabetes (%)</td><td>18 (45)</td><td>22 (55)</td><td>0.371</td></tr><tr><td>Insulin-dependent (%)</td><td>8 (20)</td><td>7 (17.5)</td><td>0.775</td></tr><tr><td>Hypertension (%)</td><td>34 (85)</td><td>34 (85)</td><td>1</td></tr><tr><td>Congestive heart failure (%)</td><td>5 (12.5)</td><td>10 (25)</td><td>0.152</td></tr><tr><td>LVEF (%)</td><td>49.4 &#x000b1; 12</td><td>42.4 &#x000b1; 12</td><td>0.13</td></tr><tr><td>Mitral regurgitation &#x0003e; Grade 2 (%)</td><td>10 (25)</td><td>9 (22.5)</td><td>0.925</td></tr><tr><td>Tricuspid regurgitation &#x0003e; Grade 2 (%)</td><td>3 (7.5)</td><td>4 (10)</td><td>0.283</td></tr><tr><td>Triple-vessel disease (%)</td><td>22 (55)</td><td>30 (75)</td><td rowspan="2">0.061</td></tr><tr><td>Quadruple-vessel disease (%)</td><td>12 (22.5)</td><td>8 (22.5)</td></tr><tr><td>Cardiogenic shock (%)</td><td>6 (15)</td><td>5 (12.5)</td><td>0.745</td></tr><tr><td>Pre-operative IABP support (%)</td><td>5 (12.5)</td><td>7 (17.5)</td><td>0.531</td></tr><tr><td>Previous cardiac surgery (%)</td><td>3 (7.5)</td><td>1 (2.5)</td><td>0.305</td></tr><tr><td>Preoperative arrhythmia (%)</td><td>10 (25)</td><td>6 (15)</td><td>0.264</td></tr><tr><td>Preoperative Atrial Fibrillation (%)</td><td>9 (22.5)</td><td>3 (7.5)</td><td>0.060</td></tr><tr><td>Preoperative stroke (%)</td><td>7 (17.5)</td><td>6 (15)</td><td>0.762</td></tr><tr><td>Peripheral vascular disease (%)</td><td>8 (20)</td><td>8 (20)</td><td>1</td></tr><tr><td>Preoperative renal failure (%)</td><td>4 (10)</td><td>6 (15)</td><td>0.499</td></tr><tr><td>Preoperative Dialysis (%)</td><td>1 (2.5)</td><td>0 (0)</td><td>0.314</td></tr><tr><td>COPD (%)</td><td>3 (7.5)</td><td>5 (12.5)</td><td>0.456</td></tr><tr><td>Cirrhosis (%)</td><td>1 (2.5)</td><td>1 (2.5)</td><td>1</td></tr></tbody></table><table-wrap-foot><p>Values are means &#x000b1; SD, or counts (%). BMI-body mass index, STS-Society of Thoracic Surgeons, NYHA-New York Heart Association, LVEF-left ventricular ejection fraction, COPD-chronic obstructive pulmonary disorder.</p></table-wrap-foot></table-wrap></p></sec><sec id="Sec9"><title>Operative data</title><p>Use of DN cardioplegia was associated with significantly shorter CPB and XC times (both, p &#x0003c; 0.001; Figure <xref rid="Fig1" ref-type="fig">1</xref>A). Additionally, administration of DN cardioplegia resulted in significant differences in manner and amount of delivered cardioplegia (all, p &#x0003c; 0.03, Figure <xref rid="Fig1" ref-type="fig">1</xref>B). An initial cold antegrade cardioplegia was used in all patients. A single cardioplegia dose was given in 33 DN vs. 5 WB patients (p &#x0003c; 0.001), and retrograde cardioplegia was used in only 8 DN vs. 31 WB patients (p &#x0003c; 0.001). Mean packed red blood cell (PRBC) transfusion requirements on CPB tended to be lower with DN versus WB during CBP (1.4 &#x000b1; 1.5 vs. 0.9 &#x000b1; 1.3, WB vs. DN, p = 0.113). The operative profiles of the 2 propensity-matched groups (WB vs. DN) are summarized in Table <xref rid="Tab3" ref-type="table">3</xref>.<fig id="Fig1"><label>Figure 1</label><caption><p> <bold>Intraoperative variables. A)</bold> Cardiopulmonary bypass and aortic cross clamp times between groups, <bold>B)</bold> Total cardioplegia volume given during operation by treatment group. (Abbreviations: CPB=cardiopulmonary bypass, DN=del Nido cardioplegia group, WB=whole blood cardioplegia group, XC=aortic cross-clamp).</p></caption><graphic xlink:href="13019_2014_Article_141_Fig1_HTML" id="d30e1094"/></fig></p><table-wrap id="Tab3"><label>Table 3</label><caption><p> <bold>Operative profile (Matched Groups)</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr><th/><th>WB</th><th>DN</th><th> <bold><italic>p</italic></bold> </th></tr></thead><tbody><tr><td>n</td><td>40</td><td>40</td><td/></tr><tr><td> <bold><italic>Intraoperative variables</italic></bold> </td><td/><td/><td/></tr><tr><td>Number of grafts</td><td>3.2 &#x000b1; 0.7</td><td>3.3 &#x000b1; 0.6</td><td>0.708</td></tr><tr><td>Lowest body temperature on CPB (&#x000b0;C)</td><td>32.3 &#x000b1; 1</td><td>34.1 &#x000b1; 2</td><td>&#x0003c;0.001</td></tr><tr><td>Cardioplegia Delivery (%)</td><td/><td/><td/></tr><tr><td>Antegrade</td><td>9 (22.5)</td><td>32 (80)</td><td>&#x0003c;0.001</td></tr><tr><td>Antegrade + Retrograde</td><td>31 (77.5)</td><td>8 (20)</td><td>&#x0003c;0.001</td></tr><tr><td>Repeated Dose (Ante or Retro)</td><td>35 (87.5)</td><td>7 (17.5)</td><td>&#x0003c;0.001</td></tr><tr><td>Cardioplegia Volume</td><td/><td/><td/></tr><tr><td>Total initial dose (ml)</td><td>1062 &#x000b1; 198</td><td>954 &#x000b1; 235</td><td>0.029</td></tr><tr><td>Antegrade-initial (ml)</td><td>786 &#x000b1; 295</td><td>944 &#x000b1; 235</td><td>0.010</td></tr><tr><td>Retrograde-initial (ml)</td><td>266 &#x000b1; 221</td><td>11 &#x000b1; 47</td><td>&#x0003c;0.001</td></tr><tr><td>Number of additional doses</td><td>3.2 &#x000b1; 2</td><td>0.2 &#x000b1; 0.4</td><td>&#x0003c;0.001</td></tr><tr><td>Additional dose amount (ml)</td><td>926 &#x000b1; 591</td><td>42 &#x000b1; 97</td><td>&#x0003c;0.001</td></tr><tr><td>Reinstitute CPB (%)</td><td>1 (2.5)</td><td>0 (0)</td><td>0.314</td></tr><tr><td>PRBC transfusion during CPB</td><td>1.4 &#x000b1; 1.5</td><td>0.9 &#x000b1; 1.3</td><td>0.113</td></tr></tbody></table><table-wrap-foot><p>Values are means &#x000b1; SD, or counts (%). CPB-cardioplumonary bypass, PRBC-packed red blood cells.</p></table-wrap-foot></table-wrap></sec><sec id="Sec10"><title>Postoperative outcomes</title><p>Postoperative outcome details are highlighted in Figure <xref rid="Fig2" ref-type="fig">2</xref> and Table <xref rid="Tab4" ref-type="table">4</xref>. In-hospital mortality was statistically comparable between groups (p = 0.314). One patient (2.5%) in DN group died due to profound cardiogenic shock and there was no hospital mortality in WB group. Additionally, the prevalence of LCOS was identical in the 2 groups (both 15%, p = 0.99). Newly required postoperative IABP support was equivalent between groups (WB: 15% vs. DN: 10%, p = 0.499). Use of postoperative inotropes and dosage used were similar between groups. Similarly, postoperative complications such as, unplanned reoperation, readmission to ICU, sepsis, renal failure, atrial fibrillation, and stroke were not significantly different between the 2 groups. Mean total post-operative PRBC transfusion requirement was significantly lower in DN versus WB patients (WB vs. DN, 2.3 &#x000b1; 2.4 vs.1.3 &#x000b1; 1.5, p = 0.033). A trend towards lower duration of ventilation, reduced incidence of postoperative atrial fibrillation, reduced length of ICU stay, and lower length of hospital stay was present in DN versus WB patients.<fig id="Fig2"><label>Figure 2</label><caption><p> <bold>In-hospital Complications by type of cardioplegia technique.</bold> Observed in-hospital post-operative complication rates in patients underdoing coronary artery bypass grafting after acute myocardial infarction stratified by cardioplegia technique with number of patients (n) represented next to event rate of each variable for whole blood and Del Nido, respectively. No significant differences were found for any subgroup except unplanned re-operation. The p values of each variable are presented next to the number of events.</p></caption><graphic xlink:href="13019_2014_Article_141_Fig2_HTML" id="d30e1343"/></fig></p><table-wrap id="Tab4"><label>Table 4</label><caption><p> <bold>Post-operative outcomes</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr><th/><th>WB</th><th>dN</th><th> <bold><italic>p</italic></bold> </th></tr></thead><tbody><tr><td>n</td><td>40</td><td>40</td><td/></tr><tr><td>Postoperative inotropic support</td><td>26 (65%)</td><td>28 (70%)</td><td>0.152</td></tr><tr><td>Number of inotropes</td><td>1.9 &#x000b1; 1</td><td>2.0 &#x000b1; 0.9</td><td>0.449</td></tr><tr><td>Dobutamine dose (&#x003bc;/kg/min)</td><td>4.2 &#x000b1; 1.1</td><td>6.0 &#x000b1; 1.4</td><td>0.124</td></tr><tr><td>Norepinephrine dose (&#x003bc;/min)</td><td>4.0 &#x000b1; 2.6</td><td>3.6 &#x000b1; 2.7</td><td>0.592</td></tr><tr><td>Vasopressin dose (units/hr)</td><td>2.4 &#x000b1; 1.5</td><td>2.7 &#x000b1; 1.5</td><td>0.600</td></tr><tr><td>Milrinone dose (&#x003bc;/kg/min)</td><td>0.3 &#x000b1; 0.1</td><td>0.3 &#x000b1; 0.1</td><td>0.057</td></tr><tr><td>Ventilation duration (hours)</td><td>44.7 &#x000b1; 87</td><td>30.6 &#x000b1; 47</td><td>0.373</td></tr><tr><td>Total PRBC Transfusion (mean) (units)</td><td>2.3 &#x000b1; 2.4</td><td>1.3 &#x000b1; 1.5</td><td>0.033</td></tr><tr><td>ICU stay (days)</td><td>7.5 &#x000b1; 14</td><td>4.0 &#x000b1; 3.0</td><td>0.133</td></tr><tr><td>Postoperative hospital stay (days)</td><td>20.2 &#x000b1; 24</td><td>12.4 &#x000b1; 9.0</td><td>0.048</td></tr></tbody></table><table-wrap-foot><p>Values are means &#x000b1; SD, or counts (%). PRBC-packed red blood cells, ICU-intensive care unit.</p></table-wrap-foot></table-wrap><p>Independent predictors of LCOS in the matched population are depicted in Table <xref rid="Tab5" ref-type="table">5</xref>. LVEF &#x0003c; 30% carried the highest risk of LCOS (OR 7.5, CI: 1.5-37.4) followed by reoperative cardiac surgery (OR 4.7, CI: 1.2- 45.8) and then preoperative cardiogenic shock (OR 4.2, CI: 1.6-28.9), respectively. Importantly, cardioplegia technique was not found to be an independent predictor of LCOS (OR 0.9, CI: 0.3-2.7). The multivariable model c statistic was 0.34 and the Hosmer-Lemeshow goodness of fit statistic was 0.63. No multicollinearity was identified.<table-wrap id="Tab5"><label>Table 5</label><caption><p> <bold>Predictors of low cardiac output syndrome (LCOS)</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr><th rowspan="2"/><th rowspan="2">Odds Ratio</th><th rowspan="2"> <bold><italic>p</italic></bold> </th><th colspan="2">95% C.I. for Odds Ratio</th></tr><tr><th>Lower</th><th>Upper</th></tr></thead><tbody><tr><td>DN cardioplegia</td><td>0.879</td><td>0.821</td><td>0.289</td><td>2.675</td></tr><tr><td>Female gender</td><td>2.056</td><td>0.222</td><td>0.648</td><td>6.526</td></tr><tr><td>Preoperative arrhythmia</td><td>2.799</td><td>0.117</td><td>0.773</td><td>10.127</td></tr><tr><td>Low LVEF, &#x0003c;30%</td><td>7.526</td><td>0.014</td><td>1.515</td><td>37.395</td></tr><tr><td>Previous Cardiac Surgery</td><td>4.735</td><td>0.039</td><td>1.154</td><td>45.809</td></tr><tr><td>Shock</td><td>4.198</td><td>0.029</td><td>1.610</td><td>28.914</td></tr><tr><td>NYHA class, &#x0003e;3</td><td>1.187</td><td>0.801</td><td>0.314</td><td>4.492</td></tr></tbody></table><table-wrap-foot><p>DN-Del Nido, LVEF-left ventricular ejection fraction, NYHA-New York Heart Association.</p></table-wrap-foot></table-wrap></p></sec></sec><sec id="Sec11"><title>Discussion</title><p>Despite advances in anesthesia, surgical techniques, perioperative care, and myocardial protection, low cardiac output syndrome after CABG for AMI confers substantial morbidity and mortality in this population, as high as 4-7% 30-day mortality in some recent series [<xref ref-type="bibr" rid="CR2">2</xref>],[<xref ref-type="bibr" rid="CR7">7</xref>]. Numerous myocardial protective solutions have been used after AMI, most notably warm whole blood and Buckberg solutions formulated to replenish ATP stores and essential electrolytes necessary for myocyte contractility [<xref ref-type="bibr" rid="CR8">8</xref>]. DN solution has not been studied in the AMI setting, and results of its use in the daily practice of adult cardiac surgery have yet to be documented. In this manuscript, we present our experience with Del Nido cardioplegia, a hyperkalemic, low calcium cardioplegic solution with lidocaine and magnesium additives, in patients undergoing CABG for AMI, and compare our outcomes to a propensity-matched cohort of patients who received our standard WB solution in the current era. Our primary findings were that: 1) there was no in-hospital mortality difference in patients receiving DB vs. WB cardioplegia, 2) intraoperative myocardial protection, as evidenced by need for inotropes or circulatory support, was equivalent in DN and WB patients, 3) DN use was associated with shorter CPB and XC times. At our institution, DN has been used for all adult cardiac surgery since its introduction in 2011.</p><p>Two distinct advantages of DN over WB cardioplegia are apparent from our study data. First, operations utilizing DN were shorter, despite having similar preoperative risk factors and undergoing similar numbers of bypass grafting, which we attribute to the reduced time required for cardioplegia administration. With DN, the need for retrograde and SVG cardioplegia administration is reduced (although not eliminated depending on coronary anatomy), and the hassle of repeated cardioplegia dosing is eliminated. Repeated dosing can be cumbersome, can interrupt the flow of the operation, and can be difficult to time properly. In this setting, a single-dose agent is highly preferable. Moreover, the reduced ischemic time (cross-clamp) could prove to have beneficial effects on clinical outcomes in a larger sample size.</p><p>A lower total volume of cardioplegia may reduce hemodilution while on bypass, thus lowering the requirement for transfusion. The benefits of preventing hemodilution has been shown with prior cardioplegia solutions - standard WB or undiluted blood (microplegia) cardioplegia offer superior myocardial protection when compared with crystalloid cardioplegia [<xref ref-type="bibr" rid="CR9">9</xref>]. Moreover, microplegia has been shown to improve recovery by limiting myocardial edema [<xref ref-type="bibr" rid="CR10">10</xref>]. Although the optimal dilution of cardioplegic solutions is still unknown, the added hemoconcentration afforded with DN administration should limit edema further than WB or microplegia. In our data, total blood transfusion requirement was lower with DN versus WB, likely as result of lower total cardioplegia volume. Secondary benefits may potentially include lower duration of ventilator support and reduced hospital stay.</p><p>During AMI, a complex series of biochemical and metabolic changes in myocardial tissue occur due to deprivation of oxygen and nutrient supply, causing myocardial tissue to behave differently than other conditions. Consequently mitochondrial damage and energy depletion impair myocardial contractile function [<xref ref-type="bibr" rid="CR11">11</xref>]. Anaerobic glycolysis due to the absence of oxygen results in the accumulation of lactate and intracellular pH reduction (to &#x0003c;7.0). The latter activates the Na⁺-H⁺ ion exchanger, thus extruding protons from the cell in exchange for Na⁺ entry. The impaired function of (Na + K)-ATPase contributes to exacerbate the intracellular Na⁺ and Ca²⁺ overload, which in turn worsens intolerance to ischemia [<xref ref-type="bibr" rid="CR3">3</xref>]. Moreover, during hyperkalemic cardioplegic arrest (K⁺ 16-20 mmol/L), membrane potential depolarization occurs at a membrane potential of -35 to -64 mV, at which point a small percentage of Na⁺ and Ca⁺² channels may continue to be active. The net result is continued intracellular accumulation of Ca⁺² during arrest, which the cell counteracts through energy requiring active transport mechanisms, and ultimately manifesting as myocardial dysfunction upon reperfusion. It is theorized that the lidocaine content in DN serves to increase Na⁺ channel blockade and minimize the potential for Na⁺ window current [<xref ref-type="bibr" rid="CR5">5</xref>]. This, in addition to its Mg²⁺ content acting as Ca²⁺ antagonist, may represent an important mechanism of benefit of DN cardioplegia. An effective reduction in intracellular calcium as a result of this mechanism has been demonstrated in animal hearts arrested with DN solution - diastolic intracellular calcium levels were significantly lower in DN hearts compared to standard WB solution, without a negative contractile effect after recovery [<xref ref-type="bibr" rid="CR12">12</xref>]. Furthermore, a study involving both animals and pediatric patients [<xref ref-type="bibr" rid="CR5">5</xref>] comparing WB and DN cardiplegia also demonstrated superior calcium handling of rat cardiomyocytes exposed to DN and reduced troponin T release after surgery in pediatric patients. Maintenance of primary calcium handling mechanisms may the most important biologic mechanism of cytoprotection of DN, but requires further investigation.</p><p>The limitations inherent to a retrospective analysis are present in this study. Although all patients in our recent experience were included and matched for preoperative risk using a propensity analysis, the possibility of selection bias exists due to differing surgeons, practice referral patterns for CABG, and the evolving management of AMI using PCI with modern stent platforms and pharmacology. Using recent data and consecutive patients should minimize this bias, although cannot be excluded short of a randomized trial. In addition to cohort size being small, some covariates such as intra- and postoperative hemodynamic parameters, completeness of revascularization, cardiac biomarker dynamics in postoperative period, and surgeon's level of experience could not be included in the analysis. Nonetheless, our operative techniques and postoperative management remained fairly similar during the time of these 2 cohorts.</p></sec><sec id="Sec12"><title>Conclusion</title><p>Our study showed that myocardial protection with DN solution in isolated CABG surgery following AMI was associated with mortality, complications, and preservation comparable to WB cardioplegia. The reduced CPB and XC times with DN solution, as well as lower total volume of cardioplegia, may help shorten lengths of stay and improve clinical outcomes. In addition, the added benefit of maintaining continuity of the operation without the need for repeated cardioplegia doses cannot be overlooked. Our study is the first to show clinical outcomes and benefits of dN cardioplegia in the setting of coronary revascularization for AMI. Further randomized clinical trial study of the clinical benefit of DN with single-dose administration, effect on cytoprotective mechanisms of membrane stabilization, and dosing regimen in this sick population is warranted.</p></sec>

Video-confidence: a qualitative exploration of videoconferencing for psychiatric emergencies

Could not extract abstract

<contrib contrib-type="author" corresp="yes"><name><surname>Trondsen</surname><given-names>Marianne Vibeke</given-names></name><address><email>marianne.trondsen@telemed.no</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author"><name><surname>Bolle</surname><given-names>Stein Roald</given-names></name><address><email>stein.roald.bolle@telemed.no</email></address><xref ref-type="aff" rid="Aff1"/><xref ref-type="aff" rid="Aff2"/></contrib><contrib contrib-type="author"><name><surname>Stensland</surname><given-names>Geir &#x000d8;yvind</given-names></name><address><email>geir.oyvind.stensland@unn.no</email></address><xref ref-type="aff" rid="Aff3"/></contrib><contrib contrib-type="author"><name><surname>Tjora</surname><given-names>Aksel</given-names></name><address><email>aksel.tjora@svt.ntnu.no</email></address><xref ref-type="aff" rid="Aff1"/><xref ref-type="aff" rid="Aff4"/></contrib><aff id="Aff1"><label/>Norwegian Centre for Integrated Care and Telemedicine, University Hospital of North Norway, P.O. Box 35, N-9038 Troms&#x000f8;, Norway </aff><aff id="Aff2"><label/>Division of Emergency Medical Services, University Hospital of North Norway, P.O. Box 45, N-9038 Troms&#x000f8;, Norway </aff><aff id="Aff3"><label/>Department South, General Psychiatric Clinic, University Hospital of North Norway, P.O. Box 6124, N-9291 Troms&#x000f8;, Norway </aff><aff id="Aff4"><label/>Department of Sociology and Political Science, Norwegian University of Science and Technology, N-7491 Trondheim, Norway </aff>

BMC Health Services Research

<sec id="Sec1" sec-type="introduction"><title>Background</title><p>De-institutionalisation, in the form of providing patient services through outpatient clinics and day treatment rather than hospital stays, is a strong and growing trend in healthcare services. In psychiatric hospitals, this process began in the 1960s and benefitted most psychiatric patients, although many were left without proper care [<xref ref-type="bibr" rid="CR1">1</xref>]. The process of de-institutionalisation occurs today partly for economic reasons and partly because new treatment options are making it increasingly feasible. In addition, maintaining life in patients&#x02019; own communities is also believed to help them better manage their disease and their lives [<xref ref-type="bibr" rid="CR1">1</xref>].</p><p>The Ministry of Health and Care Services in Norway aims to improve collaboration between the different levels of healthcare services, as well as to reduce unnecessary and compulsory admissions in acute psychiatric wards, by re-organising and de-institutionalising healthcare services [<xref ref-type="bibr" rid="CR2">2</xref>]. However, such changes may be particularly challenging in rural areas, especially for psychiatric emergencies, where the availability of specialists is limited.</p><p>Tele-psychiatry, via real-time videoconferences (VC), is increasingly being used to provide advanced consultative services in areas that lack local access to psychiatrists [<xref ref-type="bibr" rid="CR3">3</xref>]. The use of VC can improve this access by connecting psychiatrists to health personnel and patients from a great distance, and a broad range of mental health issues has been successfully handled using this technique. Studies have found that tele-psychiatric services used in non-urgent situations have increased patients&#x02019; access to therapy, increased patients&#x02019; satisfaction, saved time and reduced patients&#x02019; need for travel [<xref ref-type="bibr" rid="CR3">3</xref>-<xref ref-type="bibr" rid="CR10">10</xref>]. Young patients have stated that the use of VC communication in therapy alleviated their previous anxieties about consulting a psychiatrist [<xref ref-type="bibr" rid="CR5">5</xref>]. In addition, via regular access to specialists through VC, health personnel experienced increased knowledge and improved confidence in assisting patients locally [<xref ref-type="bibr" rid="CR3">3</xref>].</p><p>Both in psychiatric and somatic medicine, medical emergencies are characterised by complexity, uncertainty, lack of information, distances between patients and health personnel and decisions that may be retrospectively judged by others as suboptimal. The use of VC in these situations may therefore be thought of as challenging and as causing an unwanted shift of focus away from the patient and towards technology. At the same time, there is a need to develop new methods for providing proper emergency care for mentally ill patients in rural areas. The use of VC in psychiatric emergencies has been found to be safe, effective and satisfactory for both the patients and emergency health staff [<xref ref-type="bibr" rid="CR11">11</xref>,<xref ref-type="bibr" rid="CR12">12</xref>]. There is, however, limited research on the use of VC for psychiatric emergency consultations [<xref ref-type="bibr" rid="CR11">11</xref>,<xref ref-type="bibr" rid="CR12">12</xref>], and there is a need for increased knowledge of emergency tele-psychiatry in operation.</p><p>The aim of this article is to explore the experiences of patients, psychiatrists and nurses regarding the use of VC for consultations in psychiatric emergencies and how the technology influences their confidence. The article is based on a qualitative study of the implementation of the first Norwegian tele-psychiatric emergency service, which was established in 2011 by the University Hospital of North Norway [<xref ref-type="bibr" rid="CR13">13</xref>].</p></sec><sec id="Sec2" sec-type="materials|methods"><title>Methods</title><sec id="Sec3"><title>Site of research</title><p>The Department South of the General Psychiatric Clinic at the University Hospital of North Norway consists of three regional psychiatric centres and an acute psychiatric hospital ward that is located in the county capital (Troms&#x000f8;). The department serves 95,000 people living in the surrounding small cities, towns and rural communities. The regional centres are located in towns 2.5 to 4&#x000a0;hours away from the acute psychiatric ward in Troms&#x000f8;, and each of the centres have a local psychiatric ward with 12 beds, an ambulatory psychiatric team and outpatient services. The regional centres are responsible for delivering psychiatric emergency services in their geographical areas. However, because of the geographical remoteness of the regional psychiatric centres, the department have not succeeded in recruiting enough psychiatrists to the regional centres to provide robust 24-hour emergency care. To overcome this challenge, in 2011, Department South implemented a new decentralised on-call system for psychiatric emergencies in collaboration with the Norwegian Centre for Integrated Care and Telemedicine [<xref ref-type="bibr" rid="CR13">13</xref>]. VC studios were installed at the three regional psychiatric centres and in the offices or homes of six psychiatrists. The psychiatrists agreed to be accessible 24&#x000a0;hours a day, in a rotation scheme, for the regional centres&#x02019; ambulatory psychiatric teams and psychiatric wards and to be available to take part in direct patient consultations by VC in collaboration with nurses at the centres. The nurses are specially trained in psychiatric health care and are experienced in handling psychiatric crises. They generally work autonomously and only contact the psychiatrists when they need advice or help in evaluating the patient. The purpose of the new VC on-call service was to ensure that all patients had equal access to specialist assessments independent of their location.</p></sec><sec id="Sec4"><title>Research design</title><p>Because of the limited research evidence regarding the use of VC for psychiatric emergencies, a qualitative explorative research design was used in this study. With a focus on the users&#x02019; experiences of VC, we conducted semi-structured interviews with patients, psychiatrists working in the on-call system, and nurses at the three regional psychiatric centres that had participated in at least one VC consultation in a psychiatric emergency. This article is part of a larger study [<xref ref-type="bibr" rid="CR13">13</xref>] that explores how this tele-psychiatric on-call system affords clinical and organisational changes within emergency psychiatry.</p></sec><sec id="Sec5"><title>Participants</title><p>Twenty-nine participants were recruited for interviews: 5 patients (1 male and 4 females, ages 18 to 51), 5 psychiatrists (5 males, ages 40 to 60) and 19 nurses at the regional centres (4 males and 15 females, ages 39 to 63). All the psychiatrists that participated in the on-call system were interviewed except for the psychiatrist heading the department and who also initiated the new service. Both patients and the nurses represented all the three regional psychiatric centres that participated in the VC consultations. Six of the nurses were also clinical managers for each of the three local psychiatric wards and the three ambulatory psychiatric teams. At the time of their VC consultation, all patients were in an emergency situation during which they were considered for hospital admission. With consideration for their states of health, they were invited to participate in the study when they were discharged from emergency care. All potential participants were invited to participate in the study; i.e., all psychiatrists involved in the on-call system, all nurses at the regional centres that had tried VC and all patients who took part in a video consultation.</p></sec><sec id="Sec6"><title>Data generation and analysis</title><p>The interviews were all conducted individually between July 2012 and June 2013, which was approximately 1 to 2&#x000a0;years after the VC system became operational. The psychiatrists and the nurses were interviewed at their places of work. Four patient interviews took place at the closest regional psychiatric centre to their home, while one patient was interviewed at her home. The interviews focused on the participants&#x02019; experience of using VC consultations to provide psychiatric emergency care. Interview guides were used to structure the dialogue, while at the same time allowing the participants to rethink and explore their experiences in detail through &#x02018;grand tour&#x02019; questions [<xref ref-type="bibr" rid="CR14">14</xref>]. Because patients and service providers were expected to have quite different positions towards the application of VC, three different interview guides were created, each of which was tailored to the patients, nurses or psychiatrists. All participants were asked about their backgrounds, expectations and experiences with the on-call system and the VC consultations. The psychiatrists and nurses were also questioned about clinical, professional and organisational issues, while the patients were asked to explore in retrospect their experience of taking part in a VC consultation in a vulnerable situation. Based on the different interview guides, the interviews with the patients lasted 40 to 50&#x000a0;minutes, while the interviews with the psychiatrists lasted between 65 and 80&#x000a0;minutes. Because of variation in their experiences of using VC, the interviews with the nurses varied in length between 34 and 75&#x000a0;minutes. The interviews were all digitally recorded and then transcribed.</p><p>Our analysis followed a stepwise-deductive-inductive (SDI) approach [<xref ref-type="bibr" rid="CR15">15</xref>] directed towards identifying issues and themes inductively and by &#x02018;emergence&#x02019; [<xref ref-type="bibr" rid="CR16">16</xref>]. The transcripts were first coded in detail with the support of the HyperRESEARCH analysis software, maintaining in very detail the content of interviews. A total of 241 empirically close codes were generated in this process. These codes were grouped into 9 categories and then into 5 main themes. One of these five themes was titled &#x02018;Confidence&#x02019; and forms the empirical-analytical basis for this article. Using the stepwise-deductive&#x02019; control questions in the SDI method, each following step was assessed. For example, coding names were questioned for each code (e.g., Does it represents in detail what is being said in the interview?). After categories were developed inductively, we asked whether the categories were strong enough to cover the content of the categorised codes. By using this circular strategy, the SDI approach secured a tight connection between the empirical data and our interpretation [<xref ref-type="bibr" rid="CR15">15</xref>].</p></sec><sec id="Sec7"><title>Ethical considerations</title><p>The study has been approved by the Regional Committee for Medical Research Ethics in Norway. All participants were recruited based on voluntary participation, all patients provided written informed consent and all health personnel provided oral consent. The data material was depersonalised and securely handled according to the ethical recommendations of the Regional Committee for Medical Research Ethics in Norway.</p></sec></sec><sec id="Sec8" sec-type="results"><title>Results</title><p>The regional psychiatric centres were found to largely operate autonomously. On-call psychiatrists were regularly contacted via telephone by the nurses, although not on a daily basis. Because the use of the telephone was a well-established practice between the regional centres and the psychiatrists, the VC system was mostly seen as a supplement to the existing telephone communication. VC was used in the more challenging situations, in which the nurses or psychiatrists were uncertain about the patient&#x02019;s assessment and/or further treatment. In most cases, telephone calls were used as dyadic communication; e.g., between one nurse at a regional psychiatric centre and the on-call psychiatrist. When using VC, however, the patient and often a second nurse were also included in the communication.</p><p>In the analysis of the participants&#x02019; experience with applying VC, various accounts relating to confidence emerged and were grouped into a core empirical theme. We found that having access to the VC system increased the experience of confidence in challenging psychiatric emergencies in four different ways: (1) by strengthening patient involvement during psychiatric specialist assessments, (2) by reducing uncertainty, (3) by sharing responsibility for decisions and (4) by functioning as a safety net even when VC was not used. We explore these accounts in more detail in the following subsections.</p><sec id="Sec9"><title>Patient involvement</title><p>VC communication between the regional centres and the on-call psychiatrist involved the patient directly, as opposed to the established use of telephones, and strengthened patient involvement. Although some of the patients said they had been sceptical about communicating through VC in advance, they experienced stronger confidence and a feeling of being taken more seriously afterwards the VC consultation. The patients emphasised that the direct contact with the psychiatrist made them calmer and ensured that the right assessment and decisions were being made.<disp-quote><p><italic>VC gave me the possibility to speak directly with the psychiatrist while the nurse was sitting next to me. I could actively take part in the assessment.</italic> (Patient 5)</p><p><italic>I may have become annoyed, or maybe offended, if the communication about my mental health had been based on the view of the nurse. It was nice to be able to present my own story [directly to the psychiatrist].</italic> (Patient 3)</p></disp-quote></p><p>This statement by Patient 3 emphasises the need or wish for expert assessment from a specialist and not &#x02018;based on the view of the nurse&#x02019;. For patients in this situation, experiencing that extra efforts are being made to ensure the highest quality of care is very meaningful. Patient 3 also said, &#x02018;<italic>VC was helpful and calming. I understood that they took me seriously when they had to use VC&#x02019;.</italic> From this patient&#x02019;s perspective, the use of VC meant that all available resources were being used to provide the best help possible.</p><p>The nurses and psychiatrists also found that VC enabled stronger patient involvement.<disp-quote><p><italic>A telephone call is quicker. In spite of that, I am more satisfied with my work after VC because the patient has received a better service.</italic> (Nurse 4)</p><p><italic>Through VC I can see and talk directly with the patient, and then I get a better overview of the situation when I feel it is necessary. A VC consultation provides more than a telephone call.</italic> (Psychiatrist 5)</p><p><italic>When the psychiatrist meets the patient [..] more people have observed and evaluated [the patient], we all know each other&#x02019;s thoughts, and of course this improves the confidence both for the patient and for us.</italic> (Nurse 5)</p></disp-quote></p><p>The patients got more involved in the communication between the nurse and the psychiatrist when VC was used, and this changed the social dynamics of the consultation. When the voice of the patient was no longer mediated through the nurse, the nurse was relieved of a mediator&#x02019;s responsibility, as pointed out above. The patient felt that he or she was taken more seriously by being allowed to take part in direct dialogue with the highest level of formal expertise.</p></sec><sec id="Sec10"><title>Reducing uncertainty</title><p>In challenging psychiatric emergencies, the nurses at the regional psychiatric centres were sometimes uncertain about how to respond; they would then ask the on-call psychiatrist to see the patient. In these situations, they would also want to discuss the patient&#x02019;s situation and have access to help to collaboratively decide on the best treatment options. In these situations, using VC reduced uncertainty and ambivalence. One of the nurses said that using VC provided an opportunity &#x02018;<italic>to assess the patient together with someone who also can see the patient, who can see the same things as I, or maybe something I haven&#x02019;t seen, which makes the assessment sufficient&#x02019;</italic> (Nurse 10). When the psychiatrist had seen the patient, the nurse also gained professional support for his or her preliminary assessment, which improved their confidence in their own skills:<disp-quote><p><italic>What I consider most important is that the psychiatrist can watch the patient and assess the patient based on what he sees. This is a support for me, if I have observed the same things.</italic> (Nurse 9)</p></disp-quote></p><p>In challenging situations or when in doubt, the psychiatrists also wanted to see and talk with the patient in order to make well-considered decisions. When VC was not used, a telephone call was made by the nurse to the psychiatrist to present their view and interpretations and to discuss next steps. However, aspects of the patient&#x02019;s condition that were difficult to convey verbally and that potentially would require direct observation were potentially lost or overlooked when the psychiatrists only received information through the nurse. Therefore, the quality of assessment and decision-making was strengthened through the VC contact between psychiatrist and patient. The psychiatrists said that having the patient visible on the screen in front of them was important.<disp-quote><p><italic>I feel more confident when I in fact have seen the patient face to face.</italic> (Psychiatrist 2)</p><p><italic>I am more certain with a decision I might not make otherwise without having seen the patient. [Seeing the patient through VC] is something different than receiving a story told by others.</italic> (Psychiatrist 1)</p><p><italic>At times, when I don&#x02019;t know the nurse or the patient, and if it is a serious situation, I may feel uncertain. If the story told by the nurse is not in lined with the story from the referring general practitioner, my gut feeling tells me something is not right. Then I may have a need to enter the situation myself.</italic> (Psychiatrist 5)</p><p><italic>I think we have to see the patients. If we cannot meet them face-to-face, we should see them on VC. This service would have been less certain without it, and we might need to take chances. I believe the risk would have been greater if we didn&#x02019;t meet the patients ourselves with VC</italic> (Psychiatrist 3)</p></disp-quote></p><p>The use of VC was therefore regarded as a method to ensure high-quality decision making, which increased the confidence for all involved that the best decision was being made for each patient. This included the opportunity of receiving a second opinion from a psychiatrist or having immediate access to the experience and competence of a psychiatrist in a challenging situation. Also, during VC sessions, several participants are able to present their views, thereby increasing both the understanding as a group comprised of the clinicians and the patient and the collective confidence in&#x02014;and commitment to&#x02014;the decisions being made.</p></sec><sec id="Sec11"><title>Sharing responsibility</title><p>The use of VC increased the confidence of both the nurses and psychiatrists because more than one person was able to observe and communicate with the patient. VC made it easier to share the responsibility for patient treatment, which was especially important in challenging situations, such as for suicidal patients. For the nurses at the regional centres, sharing the responsibility for decision-making with a psychiatrist was of great support.<disp-quote><p><italic>My assessments are supported in situations in which I would otherwise feel alone.</italic> (Nurse 2)</p><p><italic>Assessment of suicidal patients [..] is a responsibility I do not want to have.</italic> (Nurse 8)</p><p><italic>For me, this is all about shared responsibility. [..] VC gives me the opportunity to discuss my concerns, the patient may participate in the discussion, and then decisions can be made.</italic> (Nurse 8)</p></disp-quote></p><p>The psychiatrists emphasised that the nurses at the regional centres in general had made safe and sound patient assessments before contacting the psychiatrists. In some cases, however, the nurses needed a psychiatrist&#x02019;s confirmation of their assessment. One of the nurses said, <italic>&#x02018;It is also important to receive confirmations that you have done and said the right things&#x02019;.</italic> (Nurse 10)</p><p>Although the use of VC was usually initiated by the nurses at the regional centres, some of the psychiatrists reported that they asked for VC when, during a telephone call, they heard that the nurse felt insecure. As Psychiatrist 2 said, <italic>&#x02018;I suggest VC when the regional staffs are uncertain and I am unable to form a picture of the patient and the problem&#x02019;.</italic> The combination of not being able to see the patient with their own eyes and feeling that the nurse was uncertain about, for instance, the seriousness of a situation, was a motivation for initiating VC. This led to the professional support from the on-call psychiatrists being more specific and sensitive to each patient&#x02019;s case. The psychiatrists&#x02019; direct view of the patients provided a stronger experience of a shared responsibility. They became involved as an observer together with the nurse and were not just an external advisor.</p></sec><sec id="Sec12"><title>A safety net</title><p>Finally, the nurses and the psychiatrists emphasised the availability of VC as a safety net. Although they did not actually use VC very often, the on-call system implied that the psychiatrists could be accessed by the regional centres through VC 24&#x000a0;hours a day. With a specialist only a VC connection away, the healthcare providers felt less uncertain in seriously and challenging psychiatric emergency situations.<disp-quote><p><italic>I feel much more confident. [..] I come to work with a much greater degree of calmness than before.</italic> (Nurse 10)</p><p><italic>There is always someone there. [..] I am never alone.</italic> (Nurse 8)</p></disp-quote></p><p>VC as a safety net was reported as being particularly important during the evenings, nights and weekends, when there were fewer health personnel at the centres.<disp-quote><p><italic>I often work night shifts, and I know there is always a psychiatrist on call to help me in the event of an emergency. [..] Yes, I do feel it has eased my work.</italic> (Nurse 7)</p><p><italic>I feel more confident when I know we have the VC opportunity. The staffs know it, and the patients know it to some degree. I believe it has a positive contagious effect on us all.</italic> (Psychiatrist 5)</p></disp-quote></p><p>The improved confidence among the nurses may have actually led to more seriously ill patients being admitted to the regional centres rather than sending them to the acute hospital ward in Troms&#x000f8;. Based on VC as a 24-hour safety net, the nurses knew that they could consult a psychiatrist for further assessment at any time. If the patient&#x02019;s condition changed, the patient could be reassessed, and decisions could be quickly revised if required.<disp-quote><p><italic>The availability of VC is almost as important as the actual [use of] VC. It gives us all a degree of confidence. We can just use it if we need it.</italic> (Psychiatrist 5)</p></disp-quote></p><p>The safety net argument is of major importance in emergency care, as this work is characterised by having to immediately solve problems using any tools at hand. The opportunity of expanding the telephone communication between the psychiatrists and the regional centres by VC has value as an option itself.</p></sec></sec><sec id="Sec13" sec-type="discussion"><title>Discussion</title><p>In this discussion, we reflect on our findings in a larger context, both theoretically and in relation to practical implications. It is important to note that the use of VC analysed in this paper is within the context of medical emergencies. In both psychiatric and somatic medicine, decisions need to be made within a short timeframe, during which confidence in actors, their handling of tasks and their decision-making is of great importance to effectively address any acute situation. It has been demonstrated that communication technologies that are well-integrated into team collaboration increase quality by applying knowledge and experience across personnel and locations [<xref ref-type="bibr" rid="CR17">17</xref>]. In previous trials of VC within the context of simulated cardiac arrest, visual contact through VC improved the confidence of both the rescuers and the nurses [<xref ref-type="bibr" rid="CR18">18</xref>,<xref ref-type="bibr" rid="CR19">19</xref>]. In addition, hospital teams worked together more efficiently and with greater confidence when VC was used as a collaboration tool for complex medical emergencies and traumas [<xref ref-type="bibr" rid="CR20">20</xref>,<xref ref-type="bibr" rid="CR21">21</xref>]. Our study has shown that VC may increase certainty, improve the sharing of information and contribute positively to teamwork in challenging psychiatric situations. The presence of VC as a treatment tool therefore supports preparedness for medical emergencies.</p><p>At a more conceptual level, we have identified two main mechanisms for generating stronger confidence within the provision of emergency psychiatric care: (1) using VC for collaborative problem-solving across locations and professional levels that involve patients and (2) maintaining VC as a feasible resource for challenging situations. Both the psychiatrists and the nurses emphasised how the use of VC for collaborative problem-solving across locations and professional levels increased their confidence. By solving a problem through a team process, their uncertainty about the right assessment was reduced, and they shared the responsibility for decisions made in challenging situations.</p><p>In our study, to determine if the psychiatrist needed to be contacted, the nurses always performed an initial assessment for each patient. Often the psychiatrists confirmed these assessments; hence, collaborative problem-solving improved the nurses&#x02019; confidence in their own skills and competence. This finding is in line with previous studies comparing the effectiveness of VC to telephone calls, which found that the VC groups had increased confidence among participants and improved their ability to collaborate as teams [<xref ref-type="bibr" rid="CR18">18</xref>-<xref ref-type="bibr" rid="CR22">22</xref>]. Including the patient in the VC-mediated team process allowed him or her to take an active part in their assessment and felt included in the decision-making process. Patient involvement meant that the patient&#x02019;s voice could be heard without first being filtered through the nurse.</p><p>The second main mechanism that we identified was maintaining VC as a safety net for challenging situations. This safety argument is of major interest because it challenges the idea of frequent use as a success criterion for technological innovations. Having the opportunity to use VC in a difficult situation is a source of confidence even when it is not used. Accordingly, it has been demonstrated that having a specialist&#x02019;s e-mail address is important for chronically ill patients even if no e-mails are ever sent [<xref ref-type="bibr" rid="CR23">23</xref>]. For the nurses at the regional psychiatric centres, just knowing that they could access the on-call psychiatrist through telephone or VC whenever they need it made them calmer and feel less alone. This opportunity is also an important aspect of confidence for the psychiatrist; i.e., knowing he can see and talk with the patient directly if he consider that it is necessary. VC as a safety net improved confidence among the nurses. Having the VC system set-up and being trained in how to use it strengthened the regional centres&#x02019; ability to handle and admit challenging and more seriously ill patients. A very important finding is that this effect was not dependent on the actual use (or frequency of use) of the VC system; rather, it was only related to having the VC equipment installed and being confident in using it. In addition, all parties knew that the nurses could contact the psychiatrist again later for reassessment if the patient&#x02019;s situation worsened.</p><p>The introduction of VC is a minor technological innovation, as VC technology has been well-established for many years. However, our study confirms that in psychiatric emergencies, including VC into a network between regional psychiatric centres and specialists at a central hospital may trigger significant organisational development. The nurses experienced the support of specialists, both as confirmation of their own assessment and as a safety net for dealing with challenging situations. At the same time, the patients experienced that the VC connection represented a major effort being made for their sake: It was a reassurance of the health personnel taking the needs and worries of patients seriously. Our analysis supports the sociological notion of a paternalistic public health service [<xref ref-type="bibr" rid="CR24">24</xref>], in which patients trust doctors&#x02019; advice more than other health personnel&#x02019;s assessment. Also, to some degree, this explains the effect of direct contact with the psychiatrist: Being taken seriously is related to the progress of the actual consultation, as well as to the formal position of the clinician who manages the consultation.</p><p>The VC installation afforded another level of regional collaboration for handling psychiatric emergencies. By applying an <italic>affordance perspective</italic> [<xref ref-type="bibr" rid="CR25">25</xref>,<xref ref-type="bibr" rid="CR26">26</xref>], we emphasise how various actors perceive the application of VC differently and how these different perceptions contribute to organisational changes within the delivery of emergency psychiatric services over time. By having direct access to the psychiatrist, the patients felt that they were taken more seriously. The nurses experienced VC as providing shared responsibility and as a safety net, and the psychiatrists were able to provide better assessments as the patients were visible and given a more active role. While these experiences reflect affordances of VC in psychiatric emergencies, the four themes in our analysis are related to various aspects of confidence that are not related to purely technical qualities. Rather, the potential use of VC provides a greater transparency between the central and local healthcare providers, as well as between these providers and the patients. This may represent a &#x02018;good circle&#x02019; of developing greater trust between actors in the provision of emergency psychiatric care.</p><p>Our analysis also supports a more reserved belief in the collaboration/education argument for implementing tele-medical applications: Stronger professional confidence among local healthcare staff may in the long run give them a more independent position in relation to centrally placed specialists. However, due to the inclusion of patients&#x02019; voices in our analysis, we have identified that patients in very vulnerable situations have a need for direct access to the highest level of formal expertise. Although well-founded assessments were made by the nurses in this study, patients submit to a taken-for-granted warranty in a doctor&#x02019;s clinical gaze. While health service provision is becoming more interdisciplinary, specialised health advice today comes from many more sources than just medical doctors. However, physicians&#x02019; knowledge and practice hold a strong position in the general population [<xref ref-type="bibr" rid="CR24">24</xref>], and this needs to be taken into consideration when handling particularly vulnerable mentally ill patients.</p><sec id="Sec14"><title>Limitations</title><p>This study is limited by the fact that this new VC system had only been in use for 1 to 2&#x000a0;years before the interviews were conducted. The patients and most of the nurses had therefore participated in only one VC session, although the psychiatrists had used VC several times. The participants may have focused on other aspects of their experience if the system had been well- established or if they had more experience with VC in this setting.</p><p>The study is also limited by the few patients willing to participate. Even if the total number of interviews were sufficient for understanding how VC is experienced and used, more patients could have helped us to better understand the patient perspective. We cannot rule out the fact that the patients willing to participate were more positive to VC than those who did not reply positively to recruitment.</p><p>Using semi-structured interviews enabled us to explore the nuances of the experiences of both the providers and patients in VC consultations. While the accounts of participants in this study are not statistically generalisable, the analysis in this article can serve as a basis for a conceptual [<xref ref-type="bibr" rid="CR15">15</xref>] or analytic [<xref ref-type="bibr" rid="CR27">27</xref>] generalisation, in which the four processes of confidence are relevant to other uses of VC in medical emergencies. Therefore, the contribution of this article is first and foremost an empirical exploration focused on this conceptual understanding of confidence.</p></sec></sec><sec id="Sec15" sec-type="conclusion"><title>Conclusion</title><p>This study has demonstrated that an emergency psychiatric service delivered by VC may improve the confidence of psychiatrists, nurses and patients in challenging psychiatric emergencies. Even if VC is not used often, it serves as a safety net for challenging situations, which increases the confidence of both nurses and psychiatrists. This illustrates that frequent use is not necessarily a success criterion for the application of technological innovations to emergency care.</p><p>When VC is used as an alternative to telephone calls, a group of people can see and hear each other at the same time. This means that all participants&#x02014;patients, nurses and psychiatrists&#x02014;can take an active role in the assessment and decisions-making process through face-to-face communication. VC is a richer mean of communication through which participants, especially patients, experience increased trust in each other and improved confidence in the decisions that are made. We therefore conclude that VC can be an important tool for building confidence in psychiatric emergencies and that patients, nurses and psychiatrists experienced that VC enables the provision of high-quality services for patients who need emergency psychiatric care that is located close to their homes.</p></sec>

Significance of ossificated ungular cartilages regarding the performance of cold-blooded trotters

Could not extract abstract

<contrib contrib-type="author" corresp="yes"><name><surname>Hedenstr&#x000f6;m</surname><given-names>Ulf O</given-names></name><address><email>ulf.hedenstrom@wangen.se</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author"><name><surname>Wattle</surname><given-names>Ove S</given-names></name><address><email>ove.wattle@slu.se</email></address><xref ref-type="aff" rid="Aff2"/></contrib><aff id="Aff1"><label/>National Equine Education Centre, W&#x000e5;ngen AB, SE-83593 Alsen, Sweden </aff><aff id="Aff2"><label/>Division of Diagnostics and Large Animal, Department of Clinical Sciences, Swedish University of Agricultural Sciences, SE-75007 Uppsala, Sweden </aff>

Acta Veterinaria Scandinavica

<sec id="Sec1" sec-type="introduction"><title>Background</title><p>Ossification of the ungular cartilages (OUC) in the foot of horses has been studied in different breeds for more than 100&#x000a0;years and the high heritability of this condition is well known [<xref ref-type="bibr" rid="CR1">1</xref>-<xref ref-type="bibr" rid="CR5">5</xref>]. For many decades, a number of breed clubs in Sweden and Norway have excluded stallions with high grades of OUC from their breeding programs. However, this measure is frequently debated among Swedish and Norwegian breeders, horse owners, farriers and veterinarians. One breeding organization, Swedish Ardennes, has recently ceased evaluating OUC when classifying stallions before breeding, since Tullberg and Wattle [<xref ref-type="bibr" rid="CR6">6</xref>] showed that at least 80% of the horses had OUC despite 70&#x000a0;years of breeding to lower its incidence in the population. Furthermore, it has not yet been shown that OUC, unless fractured, is a cause of lameness in the Swedish Ardennes horses. For other breeds too, there are only vague indications that OUC, if not fractured [<xref ref-type="bibr" rid="CR7">7</xref>], can cause lameness. In a study of 21 Finnhorses, Ruohoniemi <italic>et al.</italic> [<xref ref-type="bibr" rid="CR8">8</xref>] suggested that the performance of four of the horses might have been affected by OUC. Dyson and Nagy [<xref ref-type="bibr" rid="CR9">9</xref>] reported significant association between extensively ossified cartilages of the foot and collateral ligament and distal phalanx injury. The authors proposed that a possible clinically significant ossification was found in 13% of 462 cases [<xref ref-type="bibr" rid="CR9">9</xref>]. Mair and Sherlock [<xref ref-type="bibr" rid="CR10">10</xref>] described an association between injuries to multiple structures of the foot in 15 horses with high grades of OUC and another study on cadaver feet from six Finnhorses by Ruohoniemi <italic>et al.</italic> [<xref ref-type="bibr" rid="CR11">11</xref>] proposed incomplete fusion lines as a potential relevant clinical finding and highlighted separate centers of ossification and medial OUC as possibly more clinically significant. Neither Dyson and Nagy [<xref ref-type="bibr" rid="CR9">9</xref>], Mair and Sherlock [<xref ref-type="bibr" rid="CR10">10</xref>] nor Ruohoniemi <italic>et al.</italic> [<xref ref-type="bibr" rid="CR11">11</xref>] showed a correlation between clinical signs and OUC alone.</p><p>The Swedish-Norwegian cold-blooded trotter and the Finnish cold-blooded trotter are similar breeds, occasionally racing together, both slightly heavier than the standard- bred trotter. Among professional trainers, horses&#x02019; locomotion problems at high speed is an important clinical finding anecdotally suggested to be caused by OUC, but there are no studies including high speed work or other performance data available. Nevertheless, modern equine orthopaedic literature ranks OUC as one of top 10 causes of lameness in cold-blooded trotters [<xref ref-type="bibr" rid="CR12">12</xref>]. Based on governmental animal welfare regulations [<xref ref-type="bibr" rid="CR13">13</xref>] Swedish, Norwegian and Finnish trotting associations consider OUC a pathological condition and the annual Scandinavian breeding evaluations exclude 1&#x02013;2 high-performance stallions every five years due to high grades of OUC. However, the breeding plans are continuously updated on the basis of new data.</p><p>Using a 0&#x02013;5 point scale, Ruohoniemi <italic>et al.</italic> [<xref ref-type="bibr" rid="CR5">5</xref>] showed that the incidence of any grade of OUC was 80% in Finnhorses while the incidence of OUC&#x02009;&#x02265;&#x02009;grade 3&#x02013;5 was 35%. In a study of 2-year-old Swedish-Norwegian cold-blooded trotters, 70% had some grade of OUC while 13% had OUC&#x02009;&#x02265;&#x02009;grade 3 [<xref ref-type="bibr" rid="CR4">4</xref>].</p><p>The cold-blooded trotter has a high performance potential over a long career, from 2 to 15&#x000a0;years of age, and the capacity to be frequently trained and raced at high speed (&#x0003e;10&#x000a0;m/s) on different, but often hard, surfaces and in differently cambered curves. Many of these trotters are trained frequently from 1.5&#x000a0;years of age. A slightly decreased performance in training and racing is often the only or most obvious sign for which trainers and owners seek professional help. The availability of reliable sports data [<xref ref-type="bibr" rid="CR14">14</xref>] and many high performing individuals makes the cold-blooded trotter a good homogeneous population for clinical retrospective research. Many of these horses follow racing with an alternative career and are often kept as pleasure horses for numerous years. This makes it possible to locate and re-examine a large number of horses years after they have finished their racing career.</p><p>The aim of this study was to compare various performance parameters in Swedish-Norwegian cold-blooded trotters without and with different grades of OUC.</p><p>Our hypothesis was that OUC somehow affects the performance of cold-blooded trotters.</p></sec><sec id="Sec2" sec-type="materials|methods"><title>Methods</title><p>This project was approved by the Uppsala Animal Ethics Committee, Dnr C 88/6.</p><p>Breeding index or Best Linear Unbiased Prediction (BLUP) [<xref ref-type="bibr" rid="CR15">15</xref>,<xref ref-type="bibr" rid="CR16">16</xref>] was used for comparison of performance potential with the total population of cold-blooded trotters and between horses examined with different grades of OUC. A BLUP value of 100 is an objective mean value regarding breeding and selected performance markers in a continuously updated five year period of all horses in a breed. Individual data, including BLUP, were collected from official Swedish and Norwegian sports databases during 2009 [<xref ref-type="bibr" rid="CR14">14</xref>], when more than 99% of included horses had completed their racing careers.</p><p>Front hooves from 649 Swedish-Norwegian cold-blooded trotters, born between 1968&#x02013;1999, were evaluated for OUC. One examined horse was excluded because of uncertain identification. For 197 mares and 211 males (geldings and stallions), all born in 1995, the radiographs were taken at age 2.5&#x02009;&#x000b1;&#x02009;0.25&#x000a0;years, in connection with the study reported by Holm <italic>et al.</italic> [<xref ref-type="bibr" rid="CR4">4</xref>]. Twelve stallions were radiographed in connection with Norwegian pre-breeding evaluations (mean age 5, median 4&#x000a0;years), while 96 mares and 133 geldings/stallions (mean age 5, median 6&#x000a0;years) were radiographed in connection with gatherings of cold-blooded trotters through open invitations to horse owners living in a 160,000&#x000a0;km² area in the central third of Sweden. All horses in this study were over 1-year old and 95% were 2&#x000a0;years or older at the time of OUC evaluation. Horses &#x0003c;2-years-old were included since training on hard surfaces normally starts as yearlings in this breed which may affect both development of distal phalanx, OUC and pathological conditions of the front feet. In addition, 147 of the horses were re-examined, mean 9 and median 8&#x000a0;years after the first occasion and are also included in a study of ossification development over time [<xref ref-type="bibr" rid="CR17">17</xref>]. For these horses, the grade of OUC at the last examination was the OUC grade used in this study.</p><p>Radiographs were taken on non-sedated horses by different veterinarians at several equine clinics. A dorso-palmar view and horizontal beam was used, with the horse standing on the floor or with the front hooves on 4&#x02013;8&#x000a0;cm thick wooden blocks. Depending on the facility where the radiographs were taken, the film focus distance varied between 90 and 125&#x000a0;cm and the exposure between 66&#x02013;75&#x000a0;kV and 3.2&#x02013;5.0 mAs using a high performance generator and x-ray tube or a portable x-ray unit. For 95% of the horses, conventional film and a cassette with intensifying screens were used, while digital radiography was used for the remaining 5%. All radiographs were blind-coded and evaluated twice by the same person (OW) using both the scale of Ruohoniemi <italic>et al.</italic> [<xref ref-type="bibr" rid="CR5">5</xref>] and a new scale (NS) in which the navicular bone and the palmar level of the distal interphalangeal joint were excluded as references (Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>). The cartilage with the highest grade of ossification, including both the left and right front feet, determined the total score for an individual horse.<fig id="Fig1"><label>Figure 1</label><caption><p> <bold>OUC grading systems.</bold> Grading of ossification according to the Ruohoniemi <italic>et al.</italic> [<xref ref-type="bibr" rid="CR5">5</xref>] scale and a new scale developed in this study. Left side shows moderate ossification with separate center of ossification and the dotted line is the navicular bone. A separate center of ossification is generally located at a level that results in a grading of 4 or 5. Grading according to Ruohoniemi <italic>et al.</italic> [<xref ref-type="bibr" rid="CR5">5</xref>]. Grade 0: No ossification, inclination sagittal. Grade 1: Ossification maximum to the distal - palmar level of interphalangeal joint space. Grade 2: To the level of proximal edge of the palmar distal interphalangeal joint space. Grade 3: To a level of the proximal edge of the navicular bone (dotted line). Grade 4: Extending above the navicular bone up to distal half of the middle phalanx. Grade 5: Ossification above distal half of the middle phalanx. New grading according to SLU (new): Grade 0: Ossification not extending proximal of the distal edge of the middle phalanx. Grade 1: Ossification extending between the distal edge of the middle phalanx to a level of proximal edge of the palmar distal interphalangeal joint space. Grade 2 Up to distal half of the middle phalanx. Grade 3: Ossification above distal half of the middle phalanx.</p></caption><graphic xlink:href="13028_2014_74_Fig1_HTML" id="MO1"/></fig></p><p>Separate centers of ossification (kernels) were registered. However, we had to exclude registration of kernels on two radiographs since radiological examination technique and artifacts, such as small amounts of mud on hooves, made it impossible to be sure if small separate centers of ossification was present or not. Incomplete fusion lines were present in a few horses but not possible to register nor evaluate by radiography in a consistent way and therefore not included in this study.</p><p>All official sports data originate from strict protocols kept by groups of officials in trotting associations who use cameras, stopwatches and measuring devices for their evaluations. Swedish data from before 1995 are not digitized and hence were retrieved from microfilmed protocols. Unofficial races and non-competitive (qualification) races were excluded because they were considered to be a source of unreliable data or submaximal performance.</p><p>The following data was extracted: number of starts, time record (i.e. best average time over 1&#x000a0;km regardless of distance), career earnings and earnings/race in Swedish/Norwegian Crowns and races completed without official records of slightly irregular trot or breaking into a gallop or pace. The gait data were used to evaluate possible subclinical manifestations of OUC only detectable at high-speed work. Earnings in two different currencies (Swedish and Norwegian Crowns) were comparable [<xref ref-type="bibr" rid="CR18">18</xref>] over the 35-year-period, about three generations [<xref ref-type="bibr" rid="CR19">19</xref>,<xref ref-type="bibr" rid="CR20">20</xref>], of breeding and racing covered in this study.</p><p>Body size score (height&#x02009;+&#x02009;chest circumference at withers) of 100 horses, over 3-years-old, from both Sweden and Norway, was either measured by us using stick and tape or taken from official pre-breeding examination data using same methods [<xref ref-type="bibr" rid="CR14">14</xref>].</p><sec id="Sec3"><title>Statistical methods</title><p>Data on diagnosis of OUC were of an ordinal nature and were obtained using the two scales (Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>). To account for the use of four measurements on each horse, i.e. the lateral and medial grade of ossification in each front hoof, generalized linear mixed (GLM) models [<xref ref-type="bibr" rid="CR21">21</xref>-<xref ref-type="bibr" rid="CR24">24</xref>] were used for the analyses. The response variable was modeled using ordinal logistic models with a multinomial distribution and a cumulative logit link function. The horse was used as a random factor. The SAS [<xref ref-type="bibr" rid="CR23">23</xref>] procedure Glimmix was used for these analyses. Evaluation of the number of races during the career completed without recorded irregularities in gait was made in the GENMOD procedure with a negative binominal distribution in log-transformed (interval 0&#x02013;1) form. The variable &#x0201c;total earnings in career&#x0201d; displayed a rather skewed distribution and to reduce the effect of possible outliers, this variable was analysed in log-transformed form. Significance level was set at <italic>P</italic>&#x02009;&#x0003c;&#x02009;0.05.</p></sec></sec><sec id="Sec4" sec-type="results"><title>Results</title><p>The mean and median BLUP values were both 105 among the 648 horses examined (Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>). The corresponding values were 104 and 105 among horses with and without OUC, respectively, and 107 and 100 among raced and unraced horses, respectively (Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>).<table-wrap id="Tab1"><label>Table 1</label><caption><p> <bold>Data for mares, stallions/geldings and all horses included in the study</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th> <bold>Number of horses included</bold> </th><th> <bold>293 mares</bold> </th><th> <bold>355 stallions/geldings</bold> </th><th> <bold>All 648 horses</bold> </th></tr></thead><tbody><tr valign="top"><td>Best linear unbiased prediction value (BLUP)</td><td>104</td><td>105</td><td>105</td></tr><tr valign="top"><td>Mean no. of races throughout the career</td><td>23</td><td>48 (*)</td><td>36</td></tr><tr valign="top"><td>Record (average time in minutes and s/km)¹ </td><td>1.33,2</td><td>1.30,6 (*)</td><td/></tr><tr valign="top"><td>Total earnings/career [SEK/NOK]¹ </td><td>58,200</td><td>228,500 (*)</td><td/></tr><tr valign="top"><td>Average earnings/start [SEK/NOK]</td><td>2,530</td><td>4,760 (*)</td><td>4160</td></tr><tr valign="top"><td>% of races fulfilled without official record of irregular gait, galopp or pace.¹ </td><td>54</td><td>59</td><td/></tr><tr valign="top"><td>% of races with official record of irregular gait¹ </td><td>1.5</td><td>1.5</td><td/></tr><tr valign="top"><td>Horses with separate centers of ossification</td><td>20</td><td>23</td><td>43</td></tr><tr valign="top"><td>Average body size score (cm) for the 100 horses measured</td><td>348</td><td>349</td><td/></tr></tbody></table><table-wrap-foot><p> ¹Data for the 186 mares and 267 stallions/gelding that had competed in official races. *Significant difference (<italic>P</italic>&#x02009;&#x0003c;&#x02009;0.05).</p></table-wrap-foot></table-wrap></p><p>267 stallions or geldings and 186 mares had raced and these 453 horses completed a total of 23,583 official races, giving a mean of 52 races (median 65) per horse that had raced. There was no significant difference between horses with and without OUC and between horses with different grades of OUC as regards whether they started at all, or number of starts among horses with a racing carrier. Of the 648 horses (2592 cartilages) separate centers of ossification were present in a 43 horses (88 cartilages). Presence of separate centers of ossification did not have significant effect on any of performance parameters.</p><p>Regardless of OUC, mares had raced significantly less than stallions and geldings (<italic>P</italic>&#x02009;&#x0003c;&#x02009;0.0001) and had significantly lower best average time over 1&#x000a0;km (<italic>P</italic>&#x02009;&#x0003c;&#x02009;0.0001), career earnings (<italic>P</italic>&#x02009;&#x0003c;&#x02009;0.0001) and earnings/race (Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>).</p><p>The incidence of OUC evaluated with the Ruohoniemi <italic>et al.</italic> [<xref ref-type="bibr" rid="CR5">5</xref>] scale and the NS is shown in Table&#x000a0;<xref rid="Tab2" ref-type="table">2</xref>. Since the scale of Ruohoniemi <italic>et al.</italic> [<xref ref-type="bibr" rid="CR5">5</xref>] was found to be less specific than the NS in its range 0&#x02013;2 [<xref ref-type="bibr" rid="CR13">13</xref>], the incidence of OUC not related to gender (<italic>P</italic>&#x02009;=&#x02009;0.09) is only presented for the NS (Table&#x000a0;<xref rid="Tab2" ref-type="table">2</xref>).<table-wrap id="Tab2"><label>Table 2</label><caption><p> <bold>Grading of ossification of ungular cartilages in relation to grading system and gender</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th> <bold>Grade RS 0&#x02013;5 NS 0&#x02013;3</bold> </th><th> <bold>0</bold> </th><th> <bold>1</bold> </th><th> <bold>2</bold> </th><th> <bold>3</bold> </th><th> <bold>4</bold> </th><th> <bold>5</bold> </th></tr></thead><tbody><tr valign="top"><td>All 648 horses (RS)</td><td>169 (26%)</td><td>207 (32%)</td><td>154 (24%)</td><td>49 (7.5%)</td><td>39 (6%)</td><td>30 (4.5%)</td></tr><tr valign="top"><td>All 648 horses (NS)</td><td>454 (70%)</td><td>97 (15%)</td><td>67 (10%)</td><td>30 (5%)</td><td>_</td><td>_</td></tr><tr valign="top"><td>Males 355 horses (NS)</td><td>263 (74%)</td><td>45 (13%)</td><td>32 (9%)</td><td>15 (4%)</td><td>_</td><td>_</td></tr><tr valign="top"><td>Mares 293 horses (NS)</td><td>191 (65%)</td><td>52 (18%)</td><td>35 (12%)</td><td>15 (5%)</td><td>_</td><td>_</td></tr></tbody></table><table-wrap-foot><p>Grade of OUC according to the Ruohoniemi <italic>et al.</italic> [<xref ref-type="bibr" rid="CR5">5</xref>] scale (RS) and the new scale (NS).</p></table-wrap-foot></table-wrap></p><p>Average number of starts was not affected by any grade of OUC (<italic>P</italic>&#x02009;=&#x02009;0.98). Furthermore, it did not matter whether OUC was present in the lateral, medial or both sides of a hoof or in one or both hooves. In addition, average race running time was not related to any grade of OUC (<italic>P</italic>&#x02009;=&#x02009;0.21) or to the medial (<italic>P</italic>&#x02009;=&#x02009;0.56) or lateral (<italic>P</italic>&#x02009;=&#x02009;0.64) presence of OUC.</p><p>Total earnings in career were significantly higher if the horse had lateral OUC on both the left (<italic>P</italic>&#x02009;&#x0003c;&#x02009;0.0001) and right front hoof (p&#x02009;=&#x02009;0.018), but not when the OUC was medial (<italic>P</italic>&#x02009;=&#x02009;0.94 and <italic>P</italic>&#x02009;=&#x02009;0.62, for left and right hoof respectively). There was also a significant positive correlation between amount of money won by horses and median lateral OUC (<italic>P</italic>&#x02009;=&#x02009;0.004), but not median medial OUC (<italic>P</italic>&#x02009;=&#x02009;0.38). However, when the amount of money won was log-transformed, the overall effect of OUC was no longer significant (<italic>P</italic>&#x02009;=&#x02009;0.77) and the relationship with median medial (<italic>P</italic>&#x02009;=&#x02009;0.63) and lateral (<italic>P</italic>&#x02009;=&#x02009;0.60) OUC formation was non-significant. This suggests that the significant results seen in the untransformed data were caused by a small number of lateral OUC outliers with very large winnings.</p><p>Percentage of races without official pace deviations was not significantly related to gradee of OUC (<italic>P</italic>&#x02009;=&#x02009;0.28). However, horses with OUC grade 1 and 2 according to NS had a tendency (<italic>P</italic>&#x02009;=&#x02009;0.08) for more frequent records of irregular trot than horses without OUC or OUC of grade 3.</p></sec><sec id="Sec5" sec-type="discussion"><title>Discussion</title><p>The average BLUP index of 105 indicates that the horses studied accurately represented the Scandinavian population during the study period. As expected, the geldings and stallions were significantly stronger in general performance, but no gender differences were found in OUC or body size. The highly significant gender differences concerning performance data are already often compensated for, by separate races or 20 meters reduced racing distance for mares, when racing against males in Sweden as well as in Norway.</p><p>The use of objective official sports data for retrospective analysis is well established [<xref ref-type="bibr" rid="CR25">25</xref>] but unfortunately objective data were not available for other breeds with a high incidence of OUC, which made comparison between breeds impossible. Between 1973 and 2009 the cold-blooded trotter has through selected breeding become both faster and more rhythmical when trotting at high speed. Andersson <italic>et al.</italic> [<xref ref-type="bibr" rid="CR26">26</xref>] have identified mutations in certain gene affecting locomotion in many breeds including the cold-blooded trotter. This new research in equine genetics open new possibilities to distinguish physiological from pathological conditions causing gait irregularities in high speed trotting horses. Official reporting of irregular trotting is a subjective method used for improving animal welfare. Since 2009 irregular trotting is no longer reported officially in Sweden but it is still reported in Norway.</p><p>Unless fractured, OUC has been questioned as an obvious cause of lameness in several studies [<xref ref-type="bibr" rid="CR27">27</xref>,<xref ref-type="bibr" rid="CR28">28</xref>], but it has been suggested by Ruohoniemi <italic>et al.</italic> [<xref ref-type="bibr" rid="CR8">8</xref>] that it may have an impact on gait at high speed. When grouped together, horses with OUC of grade 1 and 2, according to the NS, had a tendency for more frequent records of irregular trot than horses without OUC or OUC of grade 3 (<italic>P</italic>&#x02009;=&#x02009;0.08). The gait in cold-blooded trotters was at the time of this study often irregular by nature but, as mentioned above, it has improved over the years through selected breeding, Therefore the tendency for getting an official record when having OUC of grade 1 or 2 should not be exaggerated, irrespective of what may have caused a slightly irregular gait.</p><p>Our results do not support theories regarding possible significant ossification [<xref ref-type="bibr" rid="CR8">8</xref>-<xref ref-type="bibr" rid="CR11">11</xref>] since official records showed no significant correlation to presence of OUC. In fact, OUC did not in any general way decrease the race performance in the population examined, nor did it cause reported poor performance by gait irregularities at high speed trotting. Hence, our hypothesis that OUC somehow affects the performance of cold-blooded trotters had to be rejected.</p><p>Anecdotal evidence sometimes suggests that OUC can cause lameness if the horse is lunged or worked on a volte, especially when highly asymmetrical OUC is present [<xref ref-type="bibr" rid="CR11">11</xref>]. However, training and racing of cold-blooded trotters are often performed on banked roads or left-handed tracks with cambered curves designed for the higher speed of standard-bred racehorses. The forces created by this way of uneven loading probably far exceed those present when lunging and do not seem to have different effects on horses with or without OUC regardless of body size, separate centers of ossification and regardless which ungular cartilage, lateral or medial, was ossified. Dyson <italic>et al.</italic> [<xref ref-type="bibr" rid="CR29">29</xref>] reported a positive correlation between extensively ossified cartilage and distal interphalangeal collateral ligament desmopathy diagnosed by clinical findings and MRI. However, it is not clear from their study how they limited the effect of the magic angle [<xref ref-type="bibr" rid="CR30">30</xref>] on the collateral ligament. Furthermore, adaptation to workload and strain on hooves at extreme speed and for long duration are not taken into consideration in any study on riding sport horses or draught horses. Actually, pathological and physiological conditions can develop separately or simultaneously in the young, growing, athletic horse and this fact must be taken into consideration when evaluating clinical findings in the adult horse.</p><p>It has been suggested that both the size and sex of the horse and the limb conformation influence whether and when in life OUC occurs [<xref ref-type="bibr" rid="CR1">1</xref>,<xref ref-type="bibr" rid="CR31">31</xref>]. OUC is a moderate to highly inherited condition manifesting itself early in life [<xref ref-type="bibr" rid="CR4">4</xref>] and is probably triggered by still unknown individual or environmental factors. Today the cold-blooded trotter is trained and raced earlier in life and further studies on &#x02264;2,5&#x000a0;year old individuals, using both pathology and established non-invasive diagnostic equipment such as MRI may provide additional and more relevant information on whether environmental factors affect the development of OUC or not. Furthermore, modern objective gait analysis systems may provide useful information about possible effects of OUC in high speed trotting on straight stretches as well as in banked curves.</p><p>From a breeding perspective, both lameness and OUC will occur, but our results strongly question whether OUC alone causes clinical or subclinical lameness in any way that affects performance or animal welfare. Thus compulsory radiological pre-breeding examination of stallions or mares possible grade of OUC should be of a low priority.</p></sec><sec id="Sec6" sec-type="conclusion"><title>Conclusions</title><p>Results from this study can assist many equine professionals in evaluating and interpreting the clinical relevance of radiological findings on ossified hoof cartilage among heavy and high-performing horses. Ossification of ungular cartilages in front hooves of cold-blooded trotters is, regardless of grade or position, not likely to cause decreased performance.</p></sec>

Observational study to characterise 24-hour COPD symptoms and their relationship with patient-reported outcomes: results from the ASSESS study

Could not extract abstract

<contrib contrib-type="author" corresp="yes"><name><surname>Miravitlles</surname><given-names>Marc</given-names></name><address><email>mmiravitlles@vhebron.net</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author"><name><surname>Worth</surname><given-names>Heinrich</given-names></name><address><email>med1@klinikum-fuerth.de</email></address><xref ref-type="aff" rid="Aff2"/></contrib><contrib contrib-type="author"><name><surname>Soler Catalu&#x000f1;a</surname><given-names>Juan Jos&#x000e9;</given-names></name><address><email>jjsoler@telefonica.net</email></address><xref ref-type="aff" rid="Aff3"/></contrib><contrib contrib-type="author"><name><surname>Price</surname><given-names>David</given-names></name><address><email>david@respiratoryresearch.org</email></address><xref ref-type="aff" rid="Aff4"/></contrib><contrib contrib-type="author"><name><surname>De Benedetto</surname><given-names>Fernando</given-names></name><address><email>debened@unich.it</email></address><xref ref-type="aff" rid="Aff5"/></contrib><contrib contrib-type="author"><name><surname>Roche</surname><given-names>Nicolas</given-names></name><address><email>nicolas.roche@cch.aphp.fr</email></address><xref ref-type="aff" rid="Aff6"/></contrib><contrib contrib-type="author"><name><surname>Godtfredsen</surname><given-names>Nina Skavlan</given-names></name><address><email>Nina.Skavlan.Godtfredsen@regionh.dk</email></address><xref ref-type="aff" rid="Aff7"/></contrib><contrib contrib-type="author"><name><surname>van der Molen</surname><given-names>Thys</given-names></name><address><email>T.van.der.Molen@med.umcg.nl</email></address><xref ref-type="aff" rid="Aff8"/></contrib><contrib contrib-type="author"><name><surname>L&#x000f6;fdahl</surname><given-names>Claes-G&#x000f6;ran</given-names></name><address><email>Claes-Goran.Lofdahl@med.lu.se</email></address><xref ref-type="aff" rid="Aff9"/></contrib><contrib contrib-type="author"><name><surname>Padull&#x000e9;s</surname><given-names>Laura</given-names></name><address><email>laura.padulles@almirall.com</email></address><xref ref-type="aff" rid="Aff10"/></contrib><contrib contrib-type="author"><name><surname>Ribera</surname><given-names>Anna</given-names></name><address><email>anna.ribera@almirall.com</email></address><xref ref-type="aff" rid="Aff10"/></contrib><aff id="Aff1"><label/>Pneumology Department, Hospital Universitari Vall d&#x02019;Hebron, Ciber de Enfermedades Respiratorias (CIBERES), P. de la Vall d&#x02019;Hebron, 119&#x02013;129, Barcelona, Spain </aff><aff id="Aff2"><label/>Medical Department I, F&#x000fc;rth Hospital, F&#x000fc;rth, Germany </aff><aff id="Aff3"><label/>Servicio de Neumolog&#x000ed;a, Hospital Arnau de Vilanova, Valencia, Spain </aff><aff id="Aff4"><label/>Centre of Academic Primary Care, University of Aberdeen, Aberdeen, UK </aff><aff id="Aff5"><label/>Pneumology Unit, Ospedale Clinicizzato SS. Annunziata, Chieti, Italy </aff><aff id="Aff6"><label/>Cochin Hospital, Paris Descartes University, Paris, France </aff><aff id="Aff7"><label/>Department of Respiratory Medicine, Bispebjerg University Hospital, Copenhagen, Denmark </aff><aff id="Aff8"><label/>University of Groningen, University Medical Center Groningen, Groningen, The Netherlands </aff><aff id="Aff9"><label/>Department of Respiratory Medicine and Allergology, Lund University Hospital, Lund, Sweden </aff><aff id="Aff10"><label/>Medical Affairs, Almirall, Barcelona, Spain </aff>

Respiratory Research

<sec id="Sec1"><title>Background</title><p>Despite being preventable and treatable, chronic obstructive pulmonary disease (COPD) is associated with considerable morbidity and mortality [<xref ref-type="bibr" rid="CR1">1</xref>,<xref ref-type="bibr" rid="CR2">2</xref>] and its prevalence is expected to increase in the coming decades [<xref ref-type="bibr" rid="CR3">3</xref>]. The characteristic symptoms of COPD include breathlessness, cough and increased sputum production and, based on cohort studies, there is now extensive evidence that COPD symptoms have a considerable impact on patients&#x02019; daily activities, health status and quality of life [<xref ref-type="bibr" rid="CR4">4</xref>-<xref ref-type="bibr" rid="CR8">8</xref>]. Furthermore, while COPD is diagnosed clinically based on persistent airflow limitation, it is the impact of symptoms on patients&#x02019; daily lives that generally drives them to seek a diagnosis. The importance of considering COPD symptoms in the overall assessment of COPD, and in determining appropriate treatment approaches, is now recognised [<xref ref-type="bibr" rid="CR9">9</xref>]. Reducing symptoms, improving health status and increasing physical activity are major goals in the management of stable COPD [<xref ref-type="bibr" rid="CR9">9</xref>].</p><p>COPD symptoms have been reported to vary throughout the day [<xref ref-type="bibr" rid="CR10">10</xref>-<xref ref-type="bibr" rid="CR12">12</xref>]. In cohort studies, patients with COPD who were receiving ongoing treatment with their normal COPD medication reported that their symptoms were worst in the morning [<xref ref-type="bibr" rid="CR10">10</xref>,<xref ref-type="bibr" rid="CR12">12</xref>]. Morning symptoms impact on patients&#x02019; normal activities [<xref ref-type="bibr" rid="CR8">8</xref>,<xref ref-type="bibr" rid="CR10">10</xref>,<xref ref-type="bibr" rid="CR12">12</xref>,<xref ref-type="bibr" rid="CR13">13</xref>] and have been demonstrated to be associated with worse health status and a higher risk of COPD exacerbations [<xref ref-type="bibr" rid="CR8">8</xref>,<xref ref-type="bibr" rid="CR13">13</xref>]. In the working population, morning symptoms were also significantly associated with increased annual absenteeism [<xref ref-type="bibr" rid="CR13">13</xref>]. With regard to night-time symptoms, a recent real-world study also demonstrated that patients with night-time symptoms had significantly worse health status, more sleep disturbances and higher healthcare resource utilisation than patients without night-time symptoms [<xref ref-type="bibr" rid="CR7">7</xref>].</p><p>In a pan-European, observational study, patients&#x02019; perception of the variability of their breathlessness was associated with both the severity of breathlessness and frequent exacerbations [<xref ref-type="bibr" rid="CR10">10</xref>], while the pattern of COPD symptom variability has been shown to be influenced by disease severity [<xref ref-type="bibr" rid="CR12">12</xref>]. Previous studies have also shown an association between morning or night-time symptoms and reduced lung function [<xref ref-type="bibr" rid="CR7">7</xref>,<xref ref-type="bibr" rid="CR13">13</xref>,<xref ref-type="bibr" rid="CR14">14</xref>]. However, the association between symptoms in each part of the 24-hour day and the severity of airflow obstruction and the inter-relationship between 24-hour COPD symptoms has not previously been investigated in a single patient cohort.</p><p>In this observational study, we investigated the prevalence and severity of night-time, early morning and daytime symptoms in patients with stable COPD being treated in clinical practice and explored the relationship between symptoms in each part of the 24-hour day. Additionally, to better understand the relationship between 24-hour symptoms and other aspects of a patient&#x02019;s overall well-being, we assessed their association with the severity of airflow obstruction and other patient-reported outcomes, including self-perceived dyspnoea, health status, anxiety and depression levels, sleep quality and physical activity level.</p></sec><sec id="Sec2" sec-type="methods"><title>Methods</title><sec id="Sec3"><title>Study design</title><p>This was a multinational, non-interventional, observational study conducted in 85 clinical practice centres (pulmonologists outpatients and primary care) across Denmark, France, Germany, Italy, The Netherlands, Spain, Sweden and UK (see Additional file <xref rid="MOESM1" ref-type="media">1</xref> for a list of investigators). Patients who met the eligibility criteria were identified consecutively at each site, with each site having a maximum quota to minimise selection bias. The study consisted of a baseline visit (Day 1) and a follow-up telephone interview after 6&#x000a0;months. There were no interventions beyond routine clinical care delivered at the discretion of the physician.</p><p>A steering committee, comprising the study country co-ordinators, was involved in the design of the study. The protocol was approved by all necessary ethics committees, as required by law for each country, before study initiation (see Additional file <xref rid="MOESM2" ref-type="media">2</xref> for a list of approval authorities for each country). All patients provided written informed consent.</p></sec><sec id="Sec4"><title>Study population</title><p>Patients were aged &#x02265;40&#x000a0;years with mild to very severe COPD according to the Global Initiative for Chronic Obstructive Lung Disease (GOLD) spirometric classification [<xref ref-type="bibr" rid="CR15">15</xref>], (spirometry data from the year before baseline were considered valid). Patients were current or former smokers with a smoking history of &#x02265;10 pack-years and had no history of COPD exacerbation in the previous month.</p><p>Exclusion criteria were: any change in maintenance COPD treatment in the previous 3&#x000a0;months; a previous diagnosis of asthma, sleep apnoea syndrome or chronic respiratory disease other than COPD; and any acute or chronic condition that would limit the patient&#x02019;s ability to complete the questionnaires.</p></sec><sec id="Sec5"><title>Assessments</title><p>Night-time, early morning and daytime symptoms, severity of airflow obstruction, dyspnoea severity, health status, anxiety and depression levels, sleep quality and physical activity levels were assessed at baseline. COPD symptoms were assessed using a Night-time, Morning and Daytime Symptoms of COPD questionnaire developed by the study sponsor. To ensure accurate translation and a clear understanding of the questionnaire in each participating country, linguistic validation based on five patients per country was performed before use of the questionnaire in the study. The Night-time, Morning and Daytime Symptoms of COPD questionnaire is a 33-item questionnaire that asks the patient about the prevalence, frequency and severity of COPD symptoms during each part of the day during (i) the week before baseline and (ii) a typical week in the month before baseline (defined as a week the patient considers most usual for them in the previous month). In addition, the questionnaire also contained questions related to sleep disturbances, rescue medication, anxiety due to symptoms, limitation of activities due to symptoms and concentration levels. The questionnaire consists of three parts (one part for each period during the 24-hours) and includes 13 items for night-time symptoms, ten items for morning symptoms and ten items for daytime symptoms. Night-time corresponds to the time from when the patient goes to bed until they get out of bed to start the day; morning is the time from getting out of bed until approximately 11&#x000a0;am; and daytime is from approximately 11&#x000a0;am until the patient goes to bed. Patients were asked about the frequency of symptoms related to breathlessness, coughing, bringing up phlegm or mucus, chest tightness, chest congestion and wheezing during each period. Patients were also asked about the overall severity of their night-time, early morning and daytime symptoms during the last week; symptom severity was scored as 1 (no symptoms); 2 (mild); 3 (moderate); 4 (severe) or 5 (very severe).</p><p>Dyspnoea was assessed using the modified Medical Research Council (mMRC) scale [<xref ref-type="bibr" rid="CR16">16</xref>] with patients assessing their perceived breathlessness on a scale of 0 (breathlessness with strenuous exercise) to 4 (too breathless to leave the house or breathless when dressing or undressing). Health status was assessed using the COPD Assessment Test (CAT; total score range 0&#x02013;40, &#x0003c;10 indicates low impact, 10&#x02013;20 medium impact, 21&#x02013;30 high impact, &#x0003e;30 very high impact on health status) [<xref ref-type="bibr" rid="CR17">17</xref>]. Self-perceived anxiety and depression levels were assessed using the Hospital Anxiety and Depression Scale [<xref ref-type="bibr" rid="CR18">18</xref>-<xref ref-type="bibr" rid="CR21">21</xref>] (HADS; total score range 0&#x02013;21 where &#x02265;8 indicates a probable diagnosis [<xref ref-type="bibr" rid="CR22">22</xref>,<xref ref-type="bibr" rid="CR23">23</xref>]). Sleep quality was assessed using the COPD and Asthma Sleep Impact Scale (CASIS) [<xref ref-type="bibr" rid="CR24">24</xref>,<xref ref-type="bibr" rid="CR25">25</xref>]. Patients assessed the frequency of a range of sleep problems on a scale of 1 (never) to 5 (very often; several items are reverse-scored); individual item scores were summed to give a total raw score, which was linearly transformed to total scale score (range 1&#x02013;100); higher scores indicate greater sleep impairment. At baseline, patients were also assessed as being sedentary (does not perform any type of physical activity), moderately active (patient performs some type of exercise two or three times a week) and active (patient plays sports or exercises more than three times a week).</p></sec><sec id="Sec6"><title>Study outcomes</title><p>Primary endpoints in the study were the prevalence, severity, and inter-relationships for night-time, early morning and daytime symptoms at baseline (assessed based on the Night-time, Morning and Daytime Symptoms of COPD questionnaire). To further explore the relationship between 24-hour symptoms and other aspects of COPD that affect patients&#x02019; well-being, secondary endpoints included the relationship between night-time, early morning and daytime symptoms and severity of COPD, dyspnoea severity, health status, levels of anxiety and depression, sleep quality and physical activity levels.</p></sec><sec id="Sec7"><title>Statistical analyses</title><p>All analyses were performed using the full analysis set, which comprised all patients who fulfilled the eligibility criteria and who completed the Night-time, Morning and Daytime Symptoms of COPD questionnaire. The data were analysed using only the available data for each outcome. Descriptive data are reported as mean&#x02009;&#x000b1;&#x02009;standard deviation (SD) or percentages, as appropriate. The relationship between symptoms in each part of the day and baseline characteristics or patient-reported outcomes was assessed using univariate analysis. The relationship between symptoms in each part of the day, symptoms and airflow limitation, symptoms and comorbidities and symptoms and physical activity level was assessed using a chi-squared test. The relationship between night-time, early morning or daytime symptoms and dyspnoea, health status, anxiety and depression levels, and sleep quality was assessed using a Wilcoxon rank sum test. All statistical tests were two-sided and used a 5% significance level; there was no adjustment for multiplicity. All statistical analyses were performed using SAS (version 9.1.3 or later; SAS Institute Inc., Cary, NC, USA).</p><p>A sample size of 680 patients offered a maximum margin of error (minimum precision) of 4% for estimating the percentage of patients with night-time, early morning and daytime symptoms, considering maximum indetermination (p&#x02009;=&#x02009;50%) and a confidence level of 95%. Anticipating that approximately 5% of patients would have missing data or a major protocol violation, the final sample size was set at 720 patients.</p></sec></sec><sec id="Sec8" sec-type="results"><title>Results</title><sec id="Sec9"><title>Patients</title><p>Of 743 patients who enrolled in the study and had a baseline visit, 727 were eligible for inclusion in the full analysis set. Demographics and baseline characteristics are shown in Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>; 72.4% of patients had a diagnosis of moderate or severe COPD (based on severity of airflow limitation) and 58.9% had dyspnoea assessed on the mMRC scale as grade &#x02265;2. Overall, 50.9% of patients were receiving treatment with triple therapy (long-acting &#x003b2;₂-agonists [LABA], long-acting muscarinic antagonists [LAMA] plus inhaled corticosteroids), with or without a phosphodiesterase 4 (PDE4) inhibitor (2.1% and 48.8%, respectively). In addition to COPD, 79.4% of patients had a comorbid medical condition; 45.1% of patients had a diagnosis of hypertension and 33.7% had cardiovascular disease. Based on physical activity level, 30.0% of patients were assessed as being sedentary, 38.1% as moderately active and 31.4% as active at baseline.<table-wrap id="Tab1"><label>Table 1</label><caption><p> <bold>Demographics and baseline characteristics</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th rowspan="2"> <bold>Characteristic</bold> </th><th> <bold>Eligible patients</bold> </th></tr><tr valign="top"><th> <bold>(N&#x02009;=&#x02009;727)</bold> </th></tr></thead><tbody><tr valign="top"><td>Sex, n (%), male</td><td>478 (65.8)</td></tr><tr valign="top"><td>Age, mean (SD), years (n&#x02009;=&#x02009;725)</td><td>67.2 (8.8)</td></tr><tr valign="top"><td>BMI, mean (SD), kg/m² (n&#x02009;=&#x02009;720)</td><td>26.4 (5.2)</td></tr><tr valign="top"><td>Current smoker, n (%)</td><td>202 (27.8)</td></tr><tr valign="top"><td>Smoking history, mean (SD), pack-years (n&#x02009;=&#x02009;723)</td><td>43.1 (24.8)</td></tr><tr valign="top"><td>Post-bronchodilator FEV₁, mean (SD), L (n&#x02009;=&#x02009;696)</td><td>1.4 (0.6)</td></tr><tr valign="top"><td>% predicted FEV₁, mean (SD) (n&#x02009;=&#x02009;718)</td><td>52.8 (20.5)</td></tr><tr valign="top"><td>COPD severity, n (%)</td><td/></tr><tr valign="top"><td>&#x02003;&#x02003;&#x02003;&#x02003;GOLD group I (mild)</td><td>63 (8.7)</td></tr><tr valign="top"><td>&#x02003;&#x02003;&#x02003;&#x02003;GOLD group II (moderate)</td><td>265 (36.5)</td></tr><tr valign="top"><td>&#x02003;&#x02003;&#x02003;&#x02003;GOLD group III (severe)</td><td>261 (35.9)</td></tr><tr valign="top"><td>&#x02003;&#x02003;&#x02003;&#x02003;GOLD group IV (very severe)</td><td>73 (10.0)</td></tr><tr valign="top"><td>mMRC grade, mean (SD)</td><td>1.8 (1.0)</td></tr><tr valign="top"><td>mMRC dyspnoea grade, n (%)</td><td/></tr><tr valign="top"><td>&#x02003;&#x02003;&#x02003;&#x02003;0</td><td>53 (7.3)</td></tr><tr valign="top"><td>&#x02003;&#x02003;&#x02003;&#x02003;1</td><td>244 (33.6)</td></tr><tr valign="top"><td>&#x02003;&#x02003;&#x02003;&#x02003;2</td><td>244 (33.6)</td></tr><tr valign="top"><td>&#x02003;&#x02003;&#x02003;&#x02003;3</td><td>140 (19.3)</td></tr><tr valign="top"><td>&#x02003;&#x02003;&#x02003;&#x02003;4</td><td>44 (6.1)</td></tr><tr valign="top"><td>Patients with an exacerbation in previous year, n (%)</td><td>392 (53.9)</td></tr><tr valign="top"><td>Number of COPD exacerbations in previous year, mean (SD) (n&#x02009;=&#x02009;724)</td><td>1.2 (1.6)</td></tr><tr valign="top"><td>Current COPD medication, n (%)^a </td><td/></tr><tr valign="top"><td>&#x02003;&#x02003;&#x02003;&#x02003;LABAs&#x02009;+&#x02009;LAMAs&#x02009;+&#x02009;ICS</td><td>355 (48.8)</td></tr><tr valign="top"><td>&#x02003;&#x02003;&#x02003;&#x02003;LABAs&#x02009;+&#x02009;ICS</td><td>100 (13.8)</td></tr><tr valign="top"><td>&#x02003;&#x02003;&#x02003;&#x02003;LABAs&#x02009;+&#x02009;LAMAs</td><td>70 (9.6)</td></tr><tr valign="top"><td>&#x02003;&#x02003;&#x02003;&#x02003;LABAs alone</td><td>66 (9.1)</td></tr><tr valign="top"><td>&#x02003;&#x02003;&#x02003;&#x02003;LAMAs alone</td><td>50 (6.9)</td></tr><tr valign="top"><td>&#x02003;&#x02003;&#x02003;&#x02003;Short-acting bronchodilators^b </td><td>22 (3.0)</td></tr><tr valign="top"><td>&#x02003;&#x02003;&#x02003;&#x02003;LABAs&#x02009;+&#x02009;LAMAs&#x02009;+&#x02009;ICS&#x02009;+&#x02009;PDE4 inhibitor</td><td>15 (2.1)</td></tr><tr valign="top"><td>&#x02003;&#x02003;&#x02003;&#x02003;LAMAs&#x02009;+&#x02009;ICS</td><td>8 (1.1)</td></tr><tr valign="top"><td>&#x02003;&#x02003;&#x02003;&#x02003;Other^c </td><td>19 (2.6)</td></tr><tr valign="top"><td>&#x02003;&#x02003;&#x02003;&#x02003;No treatment</td><td>22 (3.0)</td></tr><tr valign="top"><td>Total CAT score, mean (SD) (n&#x02009;=&#x02009;721)</td><td>16.5 (8.1)</td></tr><tr valign="top"><td>CAT score category, n (%)</td><td/></tr><tr valign="top"><td>&#x02003;&#x02003;&#x02003;&#x02003;CAT score &#x02264;10, n (%)</td><td>187 (25.7)</td></tr><tr valign="top"><td>&#x02003;&#x02003;&#x02003;&#x02003;CAT score 11&#x02013;20, n (%)</td><td>305 (42.0)</td></tr><tr valign="top"><td>&#x02003;&#x02003;&#x02003;&#x02003;CAT score 21&#x02013;30, n (%)</td><td>187 (25.7)</td></tr><tr valign="top"><td>&#x02003;&#x02003;&#x02003;&#x02003;CAT score &#x0003e;30, n (%)</td><td>42 (5.8)</td></tr><tr valign="top"><td>HADS anxiety score, mean (SD) (n&#x02009;=&#x02009;710)</td><td>6.1 (4.2)</td></tr><tr valign="top"><td>HADS depression score, mean (SD) (n&#x02009;=&#x02009;714)</td><td>5.5 (4.1)</td></tr><tr valign="top"><td>CASIS score, mean (SD) (n&#x02009;=&#x02009;712)</td><td>44.1 (19.1)</td></tr></tbody></table><table-wrap-foot><p>n&#x02009;=&#x02009;patients with available data for each outcome; percentages are based on N&#x02009;=&#x02009;727 patients.</p><p> ^aUsed by &#x0003e;1% of patients.</p><p> ^bIncludes: SABA alone; SABA&#x02009;+&#x02009;SAMA; SAMA alone.</p><p> ^cIncludes: ICS alone; ICS&#x02009;+&#x02009;PDE4 inhibitor; LABA&#x02009;+&#x02009;ICS&#x02009;+&#x02009;PDE4 inhibitor; LAMA&#x02009;+&#x02009;LABA&#x02009;+&#x02009;PDE4 inhibitor.</p><p>BMI, body mass index; CASIS, COPD and Asthma Sleep Impact Scale; CAT, COPD Assessment Test; COPD, chronic obstructive pulmonary disease; FEV₁, forced expiratory volume in 1&#x000a0;second; GOLD, Global Initiative for Chronic Obstructive Lung Disease; HADS, Hospital Anxiety and Depression Scale; ICS, inhaled corticosteroid; LABA, long-acting &#x003b2;₂-agonist; LAMA, long-acting muscarinic antagonist; mMRC, modified Medical Research Council; PDE4, phosphodiesterase 4; SABA, short-acting &#x003b2;₂ agonist; SAMA, short-acting muscarinic antagonist; SD, standard deviation.</p></table-wrap-foot></table-wrap></p></sec><sec id="Sec10"><title>Prevalence and severity of COPD symptoms in each part of the 24-hour day</title><p>The prevalence of COPD symptoms in each part of the 24-hour day is shown in Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>. In each part of the 24-hour day, &#x0003e;60% of patients experienced at least one COPD symptom in the week before baseline (Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>). Early morning and daytime symptoms were most common, however 63.0% of patients experienced at least one night-time symptom in the week before baseline and more than half of the patients (52.0%) reported having night-time symptoms at least three times during a typical week.<fig id="Fig1"><label>Figure 1</label><caption><p> <bold>Prevalence and frequency of night-time, early morning and daytime COPD symptoms (N&#x02009;=&#x02009;727).</bold> ^aA typical week refers to a week that the patient considered most usual for them during the previous month. COPD, chronic obstructive pulmonary disease.</p></caption><graphic xlink:href="12931_2014_122_Fig1_HTML" id="MO1"/></fig></p><p>Patients&#x02019; assessment of the severity of their night-time, early morning and daytime symptoms is shown in Table&#x000a0;<xref rid="Tab2" ref-type="table">2</xref>. In symptomatic patients the overall severity of symptoms was comparable for the night-time, early morning and daytime periods (Table&#x000a0;<xref rid="Tab2" ref-type="table">2</xref>). In each part of the 24-hour day, most people assessed their symptoms during the previous week as mild or moderate (night-time 89.5%, early morning 87.9% and daytime 89.3%).<table-wrap id="Tab2"><label>Table 2</label><caption><p> <bold>Patients&#x02019; assessment of night-time, early morning and daytime symptom severity in the week before baseline</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th rowspan="3"> <bold>Symptom severity</bold> </th><th colspan="3"> <bold>No. of patients (%)</bold> </th></tr><tr valign="top"><th> <bold>Night-time</bold> </th><th> <bold>Early morning</bold> </th><th> <bold>Daytime</bold> </th></tr><tr valign="top"><th> <bold>(n&#x02009;=&#x02009;409</bold> ^{<bold>a</bold>} <bold>)</bold> </th><th> <bold>(n&#x02009;=&#x02009;571</bold> ^{<bold>a</bold>} <bold>)</bold> </th><th> <bold>(n&#x02009;=&#x02009;589</bold> ^{<bold>a</bold>} <bold>)</bold> </th></tr></thead><tbody><tr valign="top"><td>Mild</td><td>191 (46.7)</td><td>252 (44.1)</td><td>254 (43.1)</td></tr><tr valign="top"><td>Moderate</td><td>175 (42.8)</td><td>250 (43.8)</td><td>272 (46.2)</td></tr><tr valign="top"><td>Severe</td><td>39 (9.5)</td><td>61 (10.7)</td><td>59 (10.0)</td></tr><tr valign="top"><td>Very severe</td><td>4 (1.0)</td><td>8 (1.4)</td><td>4 (0.7)</td></tr></tbody></table><table-wrap-foot><p> ^aPatients who reported symptoms during the previous week and provided data for symptom severity.</p><p>COPD, chronic obstructive pulmonary disease.</p></table-wrap-foot></table-wrap></p></sec><sec id="Sec11"><title>Individual COPD symptoms</title><p>When individual symptoms were assessed, symptoms related to breathlessness were most common (71.4% of patients) followed by coughing (65.9%), bringing up phlegm or mucus (59.6%), wheezing (41.4%), chest tightness (32.9%) and chest congestion (23.4%). The frequency and pattern of each individual symptom varied throughout the 24-hour day (Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>). The proportion of patients reporting breathlessness increased from night-time through the morning and into the daytime, whereas coughing and bringing up phlegm or mucus were most common early in the morning. Coughing and bringing up phlegm or mucus were the most common symptoms reported during the night-time.<fig id="Fig2"><label>Figure 2</label><caption><p> <bold>Prevalence of individual COPD symptoms throughout the 24-hour day in the week before baseline (N&#x02009;=&#x02009;727).</bold> COPD, chronic obstructive pulmonary disease<bold>.</bold> </p></caption><graphic xlink:href="12931_2014_122_Fig2_HTML" id="MO2"/></fig></p></sec><sec id="Sec12"><title>Relationship between COPD symptoms in each part of the 24-hour day</title><p>In the week before baseline, 90.5% of patients experienced COPD symptoms during at least one part of the 24-hour day (Figure&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>). More than half of patients (56.7%) experienced symptoms throughout the whole 24-hour day; 10.6% of patients had symptoms in only one part (Figure&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>). Almost 60% of patients had both night-time and early morning symptoms (Table&#x000a0;<xref rid="Tab3" ref-type="table">3</xref>). Among patients with night-time symptoms, 94.3% also had early morning symptoms while 73.3% of those with early morning symptoms also had night-time symptoms. A similar pattern was observed for the combinations of night-time and daytime symptoms (Table&#x000a0;<xref rid="Tab3" ref-type="table">3</xref>).<fig id="Fig3"><label>Figure 3</label><caption><p> <bold>Prevalence of COPD symptoms during one, two or three parts of the 24-hour day in the week before baseline (N&#x02009;=&#x02009;727).</bold> COPD, chronic obstructive pulmonary disease.</p></caption><graphic xlink:href="12931_2014_122_Fig3_HTML" id="MO3"/></fig><table-wrap id="Tab3"><label>Table 3</label><caption><p> <bold>Proportion estimates of night-time, early morning and daytime COPD symptom combinations</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th> <bold>Symptom combinations</bold> </th><th> <bold>% (n/N)</bold> </th><th> <bold>[95% CI]</bold> </th></tr></thead><tbody><tr valign="top"><td> <italic>Night-time (NT) and early morning (EM) symptoms</italic> </td><td/><td/></tr><tr valign="top"><td>Overall patients with both NT and EM symptoms</td><td>59.8 (432/723)</td><td>56.1, 63.3</td></tr><tr valign="top"><td>Patients with &#x02265;1 NT symptom (n&#x02009;=&#x02009;458) who also had &#x02265;1 EM symptom</td><td>94.3 (432/458)</td><td>91.8, 96.1</td></tr><tr valign="top"><td>Patients with &#x02265;1 EM symptom (n&#x02009;=&#x02009;589) who also had &#x02265;1 NT symptom</td><td>73.3 (432/589)</td><td>69.6, 76.8</td></tr><tr valign="top"><td> <italic>Night-time (NT) and daytime (DT) symptoms</italic> </td><td/><td/></tr><tr valign="top"><td>Overall patients with both DT and NT symptoms</td><td>59.4 (429/722)</td><td>55.8, 62.9</td></tr><tr valign="top"><td>Patients with &#x02265;1 NT symptom (n&#x02009;=&#x02009;458) who also had &#x02265;1 DT symptom</td><td>93.7 (429/458)</td><td>91.1, 95.6</td></tr><tr valign="top"><td>Patients with &#x02265;1 DT symptom (n&#x02009;=&#x02009;598) who also had &#x02265;1 NT symptom</td><td>71.7 (429/598)</td><td>68.0, 75.2</td></tr><tr valign="top"><td> <italic>Early morning (EM) and daytime (DT) symptoms</italic> </td><td/><td/></tr><tr valign="top"><td>Overall patients with both EM and DT symptoms</td><td>75.0 (544/725)</td><td>71.7, 78.1</td></tr><tr valign="top"><td>Patients with &#x02265;1 EM symptom (n&#x02009;=&#x02009;591) who also had &#x02265;1 DT symptom</td><td>92.1 (544/591)</td><td>89.6, 94.0</td></tr><tr valign="top"><td>Patients with &#x02265;1 DT symptom (n&#x02009;=&#x02009;601) who also had &#x02265;1 EM symptom</td><td>90.5 (544/601)</td><td>87.9, 92.6</td></tr></tbody></table><table-wrap-foot><p>n&#x02009;=&#x02009;patients with available data for each combination.</p></table-wrap-foot></table-wrap></p><p>When the relationships between symptoms during each part of the 24-hour day were assessed, there was a significant association for each potential symptom combination (night-time and early morning symptoms; night-time and daytime symptoms; and early morning and daytime symptoms; all p&#x02009;&#x0003c;&#x02009;0.001). The relationships between night-time, early morning and daytime symptoms were maintained for all symptom combinations, irrespective of the severity of airflow limitation (mild to very severe all p&#x02009;&#x0003c;&#x02009;0.05).</p></sec><sec id="Sec13"><title>Relationship between COPD symptoms in each part of the 24-hour day and other aspects of COPD</title><p>The overall proportion of patients with any COPD symptom and the prevalence of night-time, early morning and daytime symptoms, according to COPD severity (based on airflow limitation) are shown in Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>A and B. Irrespective of the severity of airflow obstruction, &#x0003e;80% of patients in each severity category experienced COPD symptoms (Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>A). Overall, there was a significant relationship between COPD severity and symptoms during the early morning and the daytime (both p&#x02009;&#x0003c;&#x02009;0.05). However, the relationship between night-time symptoms and COPD severity did not reach statistical significance and the proportion of patients with night-time symptoms was similar across all severities (58.9&#x02013;65.9%; Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>B). Interestingly, there was also no significant relationship between COPD severity and the number of parts of the 24-hour day when patients experienced symptoms (p&#x02009;=&#x02009;0.125); 47.6% of patients with mild COPD reported symptoms during the whole 24-hour day during the week before baseline compared with 63.0% of patients with very severe COPD.<fig id="Fig4"><label>Figure 4</label><caption><p> <bold>Prevalence of any COPD symptoms (A) overall and (B) during each part of the 24-hour day, according to COPD severity.</bold> n&#x02009;=&#x02009;patients in each group based on available data. COPD, chronic obstructive pulmonary disease.</p></caption><graphic xlink:href="12931_2014_122_Fig4_HTML" id="MO4"/></fig></p><p>There was a significant relationship between night-time, early morning and daytime symptoms and the severity of self-perceived dyspnoea (all p&#x02009;&#x0003c;&#x02009;0.001; Table&#x000a0;<xref rid="Tab4" ref-type="table">4</xref>). Mean mMRC grades were significantly higher in patients with symptoms compared with patients without symptoms in each corresponding part of the 24-hour day (Table&#x000a0;<xref rid="Tab4" ref-type="table">4</xref>). There was also an association between the number of parts of the 24-hour day when patients experienced symptoms and dyspnoea severity (p&#x02009;&#x0003c;&#x02009;0.001). Most patients who had grade &#x02265;2 dyspnoea assessed on the mMRC dyspnoea scale had symptoms throughout the whole 24-hour day (63.8%); this compares with 46.5% of patients assessed as mMRC grade &#x0003c;2. There was no significant relationship between night-time, early morning or daytime symptoms and the presence of comorbidities in these patients.<table-wrap id="Tab4"><label>Table 4</label><caption><p> <bold>Patient-reported outcomes in patients with/without COPD symptoms during each part of the 24-hour day</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th/><th colspan="2"> <bold>Night-time symptoms</bold> </th><th/><th colspan="2"> <bold>Early morning symptoms</bold> </th><th/><th colspan="2"> <bold>Daytime symptoms</bold> </th><th/></tr><tr valign="top"><th> <bold>Patient-reported outcome</bold> </th><th> <bold>No symptoms</bold> </th><th> <bold>&#x02265;1 symptom</bold> </th><th> <bold>p-value</bold> </th><th> <bold>No symptoms</bold> </th><th> <bold>&#x02265;1 symptom</bold> </th><th> <bold>p-value</bold> </th><th> <bold>No symptoms</bold> </th><th> <bold>&#x02265;1 symptom</bold> </th><th> <bold>p-value</bold> </th></tr></thead><tbody><tr valign="top"><td>mMRC grade,</td><td>1.6</td><td>1.9</td><td rowspan="3">&#x0003c;0.001</td><td>1.4</td><td>1.9</td><td rowspan="3">&#x0003c;0.001</td><td>1.4</td><td>1.9</td><td rowspan="3">&#x0003c;0.001</td></tr><tr valign="top"><td rowspan="2">mean (95% CI)</td><td>(1.5, 1.7)</td><td>(1.8, 2.0)</td><td>(1.2, 1.5)</td><td>(1.8, 2.0)</td><td>(1.2, 1.6)</td><td>(1.8, 2.0)</td></tr><tr valign="top"><td>(n&#x02009;=&#x02009;265)</td><td>(n&#x02009;=&#x02009;457)</td><td>(n&#x02009;=&#x02009;134)</td><td>(n&#x02009;=&#x02009;591)</td><td>(n&#x02009;=&#x02009;124)</td><td>(n&#x02009;=&#x02009;600)</td></tr><tr valign="top"><td>CAT score,</td><td>11.6</td><td>19.3</td><td rowspan="3">&#x0003c;0.001</td><td>9.8</td><td>18.1</td><td rowspan="3">&#x0003c;0.001</td><td>10.0</td><td>17.9</td><td rowspan="3">&#x0003c;0.001</td></tr><tr valign="top"><td rowspan="2">mean (95% CI)</td><td>(10.8, 12.4)</td><td>(18.6, 20.0)</td><td>(8.8, 10.9)</td><td>(17.4, 18.7)</td><td>(8.9, 11.1)</td><td>(17.3, 18.5)</td></tr><tr valign="top"><td>(n&#x02009;=&#x02009;263)</td><td>(n&#x02009;=&#x02009;455)</td><td>(n&#x02009;=&#x02009;133)</td><td>(n&#x02009;=&#x02009;588)</td><td>(n&#x02009;=&#x02009;124)</td><td>(n&#x02009;=&#x02009;596)</td></tr><tr valign="top"><td>HADS anxiety score,</td><td>4.6</td><td>6.9</td><td rowspan="3">&#x0003c;0.001</td><td>4.1</td><td>6.5</td><td rowspan="3">&#x0003c;0.001</td><td>4.2</td><td>6.5</td><td rowspan="3">&#x0003c;0.001</td></tr><tr valign="top"><td rowspan="2">mean (95% CI)</td><td>(4.1, 5.1)</td><td>(6.5, 7.3)</td><td>(3.4, 4.7)</td><td>(6.2, 6.9)</td><td>(3.5, 4.8)</td><td>(6.1, 6.8)</td></tr><tr valign="top"><td>(n&#x02009;=&#x02009;262)</td><td>(n&#x02009;=&#x02009;445)</td><td>(n&#x02009;=&#x02009;133)</td><td>(n&#x02009;=&#x02009;577)</td><td>(n&#x02009;=&#x02009;123)</td><td>(n&#x02009;=&#x02009;586)</td></tr><tr valign="top"><td>HADS depression score,</td><td>4.2</td><td>6.2</td><td rowspan="3">&#x0003c;0.001</td><td>3.4</td><td>6.0</td><td rowspan="3">&#x0003c;0.001</td><td>3.7</td><td>5.9</td><td rowspan="3">&#x0003c;0.001</td></tr><tr valign="top"><td rowspan="2">mean (95% CI)</td><td>(3.8, 4.6)</td><td>(5.8, 6.6)</td><td>(2.9, 4.0)</td><td>(5.6, 6.3)</td><td>(3.1, 4.4)</td><td>(5.5, 6.2)</td></tr><tr valign="top"><td>(n&#x02009;=&#x02009;263)</td><td>(n&#x02009;=&#x02009;448)</td><td>(n&#x02009;=&#x02009;132)</td><td>(n&#x02009;=&#x02009;582)</td><td>(n&#x02009;=&#x02009;123)</td><td>(n&#x02009;=&#x02009;590)</td></tr><tr valign="top"><td>CASIS score,</td><td>33.6</td><td>50.2</td><td rowspan="3">&#x0003c;0.001</td><td>34.4</td><td>46.3</td><td rowspan="3">&#x0003c;0.001</td><td>34.2</td><td>46.2</td><td rowspan="3">&#x0003c;0.001</td></tr><tr valign="top"><td rowspan="2">mean (95% CI)</td><td>(31.9, 35.2)</td><td>(48.4, 52.0)</td><td>(31.7, 37.1)</td><td>(44.8, 47.9)</td><td>(31.4, 37.0)</td><td>(44.7, 47.8)</td></tr><tr valign="top"><td>(n&#x02009;=&#x02009;260)</td><td>(n&#x02009;=&#x02009;449)</td><td>(n&#x02009;=&#x02009;131)</td><td>(n&#x02009;=&#x02009;581)</td><td>(n&#x02009;=&#x02009;122)</td><td>(n&#x02009;=&#x02009;589)</td></tr></tbody></table><table-wrap-foot><p>P values determined using Wilcoxon rank-sum test versus no symptoms in each period.</p><p>n&#x02009;=&#x02009;patients with available data for each outcome.</p><p>CASIS, COPD and Asthma Sleep Impact Scale; CAT, COPD Assessment Test; CI, confidence interval; COPD, chronic obstructive pulmonary disease; HADS, Hospital Anxiety and Depression Scale; mMRC, modified Medical Research Council.</p></table-wrap-foot></table-wrap></p><p>In each part of the 24-hour day, including night-time, there was a significant relationship between symptoms and health status, anxiety and depression levels, and sleep quality (all p&#x02009;&#x0003c;&#x02009;0.001 versus no symptoms; Table&#x000a0;<xref rid="Tab4" ref-type="table">4</xref>). In each period, mean CAT scores were &#x0003e;7.5 points higher in patients with symptoms versus patients without symptoms (Table&#x000a0;<xref rid="Tab4" ref-type="table">4</xref>). Based on HADS score at baseline, 34.5% of patients had anxiety and 27.6% had depression. When assessed according to patients with and without symptoms in each part of the 24-hour day, mean HADS anxiety and depression scores were significantly higher in patients with symptoms versus those without symptoms (p&#x02009;&#x0003c;&#x02009;0.001 for all; Table&#x000a0;<xref rid="Tab4" ref-type="table">4</xref>). Sensitivity analyses performed in patients with no medical history of anxiety (n&#x02009;=&#x02009;534) or depression (n&#x02009;=&#x02009;538) also showed a significant association between symptoms and HADS anxiety and depression scores in each part of the 24-hour day (all p&#x02009;&#x0003c;&#x02009;0.001). Patients with symptoms also had significantly higher CASIS scores compared with those without symptoms (p&#x02009;&#x0003c;&#x02009;0.001 for all; Table&#x000a0;<xref rid="Tab4" ref-type="table">4</xref>), indicating greater sleep impairment. When patients who were receiving sleep medications or treatment for benign prostatic hyperplasia (n&#x02009;=&#x02009;109) were excluded from the analyses of CASIS scores, the relationship between night-time, early morning and daytime symptoms and sleep quality remained significant in each period (data not shown).</p><p>In each part of the 24-hour day, there was a significant relationship between symptoms and patients&#x02019; physical activity level at baseline (assessed as sedentary, moderately active or active; p&#x02009;&#x0003c;&#x02009;0.05 for each part of the 24-hour day). There was also a significant relationship for the number of parts of the 24-hour day when patients experienced symptoms and physical activity levels (p&#x02009;=&#x02009;0.006). A higher proportion of patients who were sedentary had symptoms throughout the whole 24-hour day compared with patients who were active (64.2% versus 50.4%, respectively).</p><p>Mean CAT, HADS anxiety and depression, and CASIS scores according to each 24-hour symptom combination are shown in Figure&#x000a0;<xref rid="Fig5" ref-type="fig">5</xref>. Patients with symptoms throughout the whole 24-hour day had the worst health status and sleep quality and the highest levels of anxiety and depression. CAT scores were higher in patients with symptoms during two or more parts of the 24-hour day than in patients with only night-time, early morning or daytime symptoms (Figure&#x000a0;<xref rid="Fig5" ref-type="fig">5</xref>A). With the exception of patients with early morning and daytime symptoms, patients who reported night-time symptoms, either alone or in combination, had the highest anxiety levels (Figure&#x000a0;<xref rid="Fig5" ref-type="fig">5</xref>B) and patients with any combination of early morning and night-time symptoms had the highest depression levels (Figure&#x000a0;<xref rid="Fig5" ref-type="fig">5</xref>C). A similar pattern was generally observed when HADS scores were analysed in patients with no medical history of anxiety or depression (Additional file <xref rid="MOESM3" ref-type="media">3</xref>). Patients with any night-time symptoms had worse sleep quality than patients without night-time symptoms (Figure&#x000a0;<xref rid="Fig5" ref-type="fig">5</xref>D).<fig id="Fig5"><label>Figure 5</label><caption><p> <bold>(A) Health status, (B) anxiety, (C) depression and (D) sleep quality according to each combination of 24-hour COPD symptoms.</bold> n&#x02009;=&#x02009;patients with available data for each outcome. CASIS, COPD and Asthma Sleep Impact Scale; CAT, COPD Assessment Test; COPD, chronic obstructive pulmonary disease; HADS, Hospital Anxiety and Depression Scale; SD, standard deviation.</p></caption><graphic xlink:href="12931_2014_122_Fig5_HTML" id="MO5"/></fig></p></sec></sec><sec id="Sec14" sec-type="discussion"><title>Discussion</title><p>In this observational study, more than half of patients reported experiencing COPD symptoms throughout the whole 24-hour day, despite receiving ongoing treatment for their COPD and almost 80% of patients had symptoms during at least two parts of the 24-hour day. While early morning and daytime symptoms were most frequent, night-time symptoms were also very common and almost two-thirds of patients experienced at least one night-time symptom during the week before baseline. Importantly, any symptoms in the early morning, daytime or night-time were associated with worse outcomes across a range of patient-reported measures including more severe dyspnoea, higher anxiety and depression levels and worse health status and sleep quality.</p><p>The observation that a large majority of patients experienced symptoms during at least two parts of the 24-hour day is consistent with results from a recent real-world study in almost 1500 patients. The study by Roche et al. showed that most patients had symptoms during the daytime and night-time and only 34% of patients experienced COPD symptoms in isolation during one part of the 24-hour day [<xref ref-type="bibr" rid="CR13">13</xref>]. However, in contrast to the results reported here, in the previous study daytime symptoms were by far the most prevalent (97% of patients) with just over one-third of patients reporting symptoms when getting up in the morning. This discrepancy may relate to the different definitions of morning symptoms used; the Roche et al. study defined morning symptoms as those present on waking and did not include those symptoms that persisted later in the morning [<xref ref-type="bibr" rid="CR13">13</xref>].</p><p>There is no objective definition of &#x02018;night-time symptoms&#x02019; in patients with COPD, and it has been suggested that night-time symptoms may be under-reported by physicians or may not be reported by patients [<xref ref-type="bibr" rid="CR26">26</xref>]. The results of our study are consistent with a previous study in 2807 patients, which demonstrated that approximately 70% of patients reported experiencing night-time symptoms [<xref ref-type="bibr" rid="CR7">7</xref>]. Together these data suggest a high prevalence of night-time symptoms in patients with COPD. Lung function exhibits circadian variation with reduced airflow during the night-time period [<xref ref-type="bibr" rid="CR27">27</xref>]. The amplitude of this circadian variation has been shown to be increased in patients with COPD [<xref ref-type="bibr" rid="CR28">28</xref>,<xref ref-type="bibr" rid="CR29">29</xref>] and it may contribute to night-time symptoms [<xref ref-type="bibr" rid="CR26">26</xref>,<xref ref-type="bibr" rid="CR28">28</xref>]. In a previous study, wheezing was the most troublesome symptom at night, followed by cough [<xref ref-type="bibr" rid="CR10">10</xref>]. In the present study cough and bringing up phlegm were the most prevalent night-time symptoms suggesting that, in addition to reduced airflow, other mechanisms may be involved in mediating night-time symptoms including mucus hypersecretion, reduced ciliary activity or increased cough sensitivity. Further investigation of these processes is required to better understand the pathophysiology underlying night-time COPD symptoms.</p><p>Early morning symptoms have been reported to be most problematic for patients with COPD and can significantly impact on daily activities [<xref ref-type="bibr" rid="CR10">10</xref>-<xref ref-type="bibr" rid="CR12">12</xref>] and working life [<xref ref-type="bibr" rid="CR13">13</xref>]. Furthermore, in a previous study, a quarter of patients with COPD reported that night-time symptoms were most troublesome and night-time was the second most problematic time for patients with severe COPD [<xref ref-type="bibr" rid="CR12">12</xref>]. However, despite patients frequently reporting night-time symptoms, the impact that symptoms at night has on daily activities, such as getting up for work, is often under-estimated by physicians [<xref ref-type="bibr" rid="CR7">7</xref>]. Previous studies have shown a significant association between night-time symptoms and the severity of airflow obstruction in patients with COPD [<xref ref-type="bibr" rid="CR7">7</xref>,<xref ref-type="bibr" rid="CR14">14</xref>]. Interestingly, our results show that whilst there was a significant relationship between early morning and daytime symptoms and the severity of airflow limitation, this association was not significant for night-time symptoms and the prevalence of night-time symptoms was comparable across all severities of airflow limitation. Furthermore, symptoms in each part of the 24-hour day were inter-related, an observation that was consistent irrespective of COPD severity. These data suggest that the presence of night-time symptoms is not merely a consequence of more severe airflow limitation. Other mechanisms, such as decreased mucociliary clearance, could be involved. However, this study did not differentiate between different phenotypes of patients with COPD and further studies are required to determine if night-time symptoms are associated with a specific phenotype.</p><p>In each part of the 24-hour day, symptoms were associated with worse dyspnoea, health status, higher anxiety and depression levels, and greater sleep impairment. These are all outcomes that can impact on patients&#x02019; daily living and overall well-being. The difference in CAT scores between patients with and without symptoms in each period exceeded the estimated minimal clinically important difference (2 points) recently proposed [<xref ref-type="bibr" rid="CR30">30</xref>], suggesting that symptoms in any part of the 24-hour day may be associated with a clinically meaningful worsening of health status. Moreover, anxiety and depression levels were also significantly higher in patients with symptoms compared with patients without symptoms. In general, anxiety levels tended to be highest in patients who had any combination of night-time symptoms and depression levels were highest in patients with any combination of night-time/early morning symptoms. Depression is a common comorbidity in patients with COPD [<xref ref-type="bibr" rid="CR2">2</xref>] and patients with severe COPD have a 2.5-fold higher risk of depression compared with matched controls [<xref ref-type="bibr" rid="CR31">31</xref>]. Comorbid depression is associated with an increased risk of exacerbation and mortality in patients with COPD [<xref ref-type="bibr" rid="CR32">32</xref>]. Since symptoms of depression tend to be worse in the morning we cannot rule out that higher levels of depression contribute to night-time and morning COPD symptoms. Of note, examining questions on COPD symptoms and the HADS questionnaire does not reveal common items, making confounding by wording unlikely. Finally, a similar pattern in the magnitude of HADS scores and symptom combinations was observed in patients with no medical history of anxiety or depression. Sleep was also significantly impaired in patients with symptoms in any part of the 24-hour day compared with patients without symptoms. As expected, the greatest impairment was observed in patients with night-time symptoms. Poor sleep quality or sleep disturbance in patients with COPD has been shown to be associated with worse health status, more exacerbations, increased healthcare resource utilisation and increased mortality [<xref ref-type="bibr" rid="CR7">7</xref>,<xref ref-type="bibr" rid="CR33">33</xref>]. In this study, we also observed a significant relationship between symptoms in any part of the 24-hour day and physical activity levels: patients who were sedentary had more symptoms in each period than patients who were even moderately active. This may be important as low physical activity levels are significantly associated with poor quality of life and increased incidence of depression in patients with COPD [<xref ref-type="bibr" rid="CR34">34</xref>] and have been shown to be a strong predictor of mortality in patients with COPD [<xref ref-type="bibr" rid="CR35">35</xref>,<xref ref-type="bibr" rid="CR36">36</xref>], and improving physical activity is an important goal in the treatment of COPD [<xref ref-type="bibr" rid="CR9">9</xref>].</p><p>Overall, our results support previous studies showing that symptoms during the morning and the night-time are independently associated with worse outcomes in patients with COPD [<xref ref-type="bibr" rid="CR7">7</xref>,<xref ref-type="bibr" rid="CR13">13</xref>]. COPD symptoms when getting up in the morning have been shown to be independently associated with worse health status and more exacerbations, and have a negative impact on daily activities [<xref ref-type="bibr" rid="CR13">13</xref>]. Similarly, patients with night-time symptoms had significantly worse breathlessness and health status and were more likely to have morning symptoms than patients without night-time symptoms, even when these analyses were controlled for confounding factors such as disease severity [<xref ref-type="bibr" rid="CR7">7</xref>]. Our results extend these studies by demonstrating that there is an inter-relationship between symptoms in each part of the 24-hour day and that symptoms in any part of the day are associated with worse patient-reported outcomes.</p><p>While these results demonstrate significant relationships between symptoms in each part of the 24-hour day and various aspects of patients&#x02019; well-being, the analyses do not take into account confounding factors such as disease severity or comorbid conditions, which may also impact on patient-reported outcomes. Furthermore, no causal relationship can be inferred from the analyses as this was an observational study. Further investigation of the specific relationship between symptoms in each part of the 24-hour day and each outcome is required to establish whether symptoms are independently associated with the outcome, irrespective of underlying disease. While this study enrolled patients with mild to very severe COPD, only patients being treated in clinical practice (both primary care and specialist centres) were assessed. As such, the relevance of these observations for the wider population of patients with COPD, including those with undiagnosed COPD, requires further consideration.</p></sec><sec id="Sec15" sec-type="conclusions"><title>Conclusions</title><p>The results of this study demonstrate that despite receiving treatment for COPD, more than half of patients continued to have symptoms throughout the whole 24-hour day, including during the night-time and early morning periods. The relationship between night-time, early morning and daytime symptoms was observed irrespective of the severity of airflow obstruction. Patients with symptoms during any part of the 24-hour day also had significantly worse outcomes across a range of measures that impact on daily living, including health status, anxiety and depression levels and sleep quality compared with patients without symptoms. This suggests that current approaches to managing COPD may not adequately control symptoms, which can impact on a patient&#x02019;s overall well-being. Newer therapies, including long-acting bronchodilators that are administered twice-daily or ultra-long-acting bronchodilators, may be useful in improving symptom control during the night-time, whereas those with a rapid onset of action may have advantages in controlling early morning symptoms. It is important for physicians to manage patients&#x02019; symptoms throughout the 24-hour day, even in those with mild airflow obstruction.</p></sec>

Improving the accuracy of expression data analysis in time course experiments using resampling

Could not extract abstract

<contrib contrib-type="author"><name><surname>Walter</surname><given-names>Wencke</given-names></name><address><email>wwalter@cnic.es</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author"><name><surname>Striberny</surname><given-names>Bernd</given-names></name><address><email>bernd.ketelsen@uit.no</email></address><xref ref-type="aff" rid="Aff2"/></contrib><contrib contrib-type="author"><name><surname>Gaquerel</surname><given-names>Emmanuel</given-names></name><address><email>egaquerel@ice.mpg.de</email></address><xref ref-type="aff" rid="Aff1"/><xref ref-type="aff" rid="Aff3"/></contrib><contrib contrib-type="author"><name><surname>Baldwin</surname><given-names>Ian T</given-names></name><address><email>baldwin@ice.mpg.de</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author" corresp="yes"><name><surname>Kim</surname><given-names>Sang-Gyu</given-names></name><address><email>skim@ice.mpg.de</email></address><xref ref-type="aff" rid="Aff1"/><xref ref-type="aff" rid="Aff4"/></contrib><contrib contrib-type="author" corresp="yes"><name><surname>Heiland</surname><given-names>Ines</given-names></name><address><email>ines.heiland@uit.no</email></address><xref ref-type="aff" rid="Aff2"/></contrib><aff id="Aff1"><label/>Department of Molecular Ecology, Max Planck Institute for Chemical Ecology, Hans-Kn&#x000f6;ll-Stra&#x000df;e 8, D-07745 Jena, Germany </aff><aff id="Aff2"><label/>Department of Arctic and Marine Biology, UiT The Arctic University of Norway, Naturfagbygget, Dramsvegen 201, 9037 Troms&#x000f8;, Norway </aff><aff id="Aff3"><label/>Center for Organismal Studies, University of Heidelberg, Im Neuenheimer Feld 360, 69120 Heidelberg, Germany </aff><aff id="Aff4"><label/>Center for Genome Engineering, Institute for Basic Science, Gwanak-ro 1, Gwanak-gu, Seoul, 151-747 South Korea </aff>

BMC Bioinformatics

<sec id="Sec1" sec-type="introduction"><title>Background</title><p>Even with decreasing costs for sequencing and microarray experiments, time series experiments are still expensive and require a large number of samples. Thus, most time series currently have a very limited number of biological replicates. This makes it difficult to identify genes that truly show time-dependent expression patterns (true positives) and genes that just seem to have similar patterns due to biological variance (false positives). The biological variance is likely to be relatively high, especially when samples are collected from higher eukaryotes, because animals and plants are usually sampled from different individuals to avoid perturbation artifacts during sampling. Thus, in most time course experiments, the samples at each time point are usually from different individuals, resulting in a high biological variance among samples. This is the main reason why sufficient numbers of replicates are necessary. Lee <italic>et al</italic>. proposed that three replicates are sufficient, but this number also depends on the type of experiment [<xref ref-type="bibr" rid="CR1">1</xref>-<xref ref-type="bibr" rid="CR4">4</xref>]. However, the importance of biological replicates is often neglected in time series experiments, especially when circadian rhythms in gene expression are examined using transcriptomics datasets.</p><p>Many organisms have an endogenous clock, known as a circadian clock, to coordinate daily activities. The output of the circadian clock has the period of approximately 24&#x000a0;h; for example, the body temperature and sleep-wake cycle in humans, leaf movement in <italic>Mimosa</italic>, and flower opening in night-blooming jasmine all show 24&#x000a0;h diurnal rhythms under both light/dark and approx. 24&#x000a0;h rhythms under constant conditions [<xref ref-type="bibr" rid="CR5">5</xref>-<xref ref-type="bibr" rid="CR7">7</xref>]. Although the molecular components of circadian clocks are not conserved between animals and plants, negative and positive feedback loops in transcriptional and post-translational levels are the core system of circadian clocks in both animals and plants [<xref ref-type="bibr" rid="CR8">8</xref>]. These multiple interlocked feedback loops confer stability and protection from stochastic perturbations on the complexity of the circadian system [<xref ref-type="bibr" rid="CR9">9</xref>,<xref ref-type="bibr" rid="CR10">10</xref>]. To understand this complex network on a transcriptional level, time series microarrays have been frequently used to examine the oscillation of genes on a genomic scale [<xref ref-type="bibr" rid="CR11">11</xref>,<xref ref-type="bibr" rid="CR12">12</xref>].</p><p>Diurnal rhythms in transcript accumulation can be described in mathematical terms, including period, phase, and amplitude [<xref ref-type="bibr" rid="CR10">10</xref>]. There are several different algorithms that can be used to calculate these parameters from real data; they can furthermore be applied to identify oscillating genes in microarray or RNA-sequencing data. From the algorithms available we selected ARSER [<xref ref-type="bibr" rid="CR13">13</xref>], HAYSTACK [<xref ref-type="bibr" rid="CR14">14</xref>] as well as the algorithms implemented in BIODARE (<ext-link ext-link-type="uri" xlink:href="http://www.biodare.ed.ac.uk/">http://www.biodare.ed.ac.uk/</ext-link>) [<xref ref-type="bibr" rid="CR15">15</xref>,<xref ref-type="bibr" rid="CR16">16</xref>]. ARSER was selected as it has been shown to outperform earlier available algorithms such as COSOPT and Fisher&#x02019;s G-Test [<xref ref-type="bibr" rid="CR13">13</xref>]. The BIODARE platform is not originally designed for the analysis of gene expression data as the maximal list length of datasets that can be submitted is limited to 2500. Thus, gene expression data has to be split into multiple datasets. Nevertheless BIODARE has the advantage of providing 6 additional different algorithms for the analysis.</p><p>Using these algorithms we show the influence of replicates and resampling on the accuracy of predictions of rhythmically expressed genes. Although we perform the analysis to identify oscillating genes in circadian expression datasets, the resampling method can be similarly used to improve the detection of other time dependent expression patterns as long as the samples are collected from different individuals at the specific time points.</p></sec><sec id="Sec2" sec-type="results"><title>Results and discussion</title><p>The determination of oscillating genes is a binary classification. There are only two possible outcomes: either a gene is rhythmically expressed or it is not. The accuracy of this classification can be estimated by a confusion matrix. There are four fundamental members of the matrix: true positives (expression profiles correctly classified as periodic), false negatives (expression profiles incorrectly classified as non-periodic), true negatives (expression profiles correctly classified as non-periodic), and false positives (expression profiles incorrectly classified as periodic). As the number of true negatives and false negatives can be directly calculated from the total number of oscillating and non-oscillating genes and the number of true- and false positive genes identified, we only analyzed true- and false-positives in our calculations. The total number of oscillating and non-oscillating genes was set to 8400 in our simulated datasets (see <xref rid="Sec4" ref-type="sec">Methods</xref> section for details).</p><p>To calculate the performance of ARSER, HAYSTACK and the algorithms implemented in BIODARE, we simulated different conditions and wave forms for oscillating transcripts. To do this we used three different simulation procedures.</p><p>To simulate entrained, synchronized oscillations all simulations were done with a fixed period ranging from 22 to 28&#x000a0;h (LD-dataset). In contrast, the differences in free running period between different individuals under constant conditions were simulated by generating a dataset that contained 36 time courses that differed in period according to published standard deviations for individual cells [<xref ref-type="bibr" rid="CR17">17</xref>] (LL-dataset). In addition we generated a time course based on a published ordinary equation model of the mammalian circadian model [<xref ref-type="bibr" rid="CR18">18</xref>] (ODE-dataset) (see <xref rid="Sec4" ref-type="sec">Methods</xref> section for details).</p><p>For each simulation procedure 36 time courses were initially calculated, corresponding to the common experimental time courses for gene expression analysis in the literature that resample 2-day time courses with 4&#x000a0;h sampling intervals and 3 replicates. From these initial time courses we generated the initial dataset (3 replicate time courses) by randomly selecting one time point from each simulated time course. These initial datasets were in addition averaged to generate a fourth, averaged time course. True and false positives were then calculated for ARSER, HAYSTACK and using BIODARE. From BIODARE we initially tested all implemented algorithms but found that FFT-NLLS was performing best, confirming the observations form Zielinski <italic>et al</italic>. [<xref ref-type="bibr" rid="CR15">15</xref>,<xref ref-type="bibr" rid="CR16">16</xref>]. We therefore only present the results from this BIODARE algorithm (Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>). In comparison ARSER detects the largest number of true positives but at the cost of a relatively high number of false positives.<fig id="Fig1"><label>Figure 1</label><caption><p> <bold>Identification of oscillating transcripts in each replicate and the average dataset.</bold> True positives <bold>(A-</bold> <bold>C)</bold> and false positives <bold>(D)</bold> for LD. LL and ODE-based time courses using ARSER, HAYSTACK or BIODARE FFT-NLLS The results are displayed for each individual replicate and the average dataset. The replicates were generated as described in the <xref rid="Sec4" ref-type="sec">Methods section</xref>.</p></caption><graphic xlink:href="12859_2014_352_Fig1_HTML" id="MO1"/></fig></p><p>In the averaged time courses the number of true positives detected by ARSER is in most cases slightly higher than in the individual replicates but this again comes at the cost of a higher number of false positives. HAYSTACK and BIODARE FFT-NLLS show similar performance but HAYSTACK has more problems to detect oscillating genes in ODE-based simulations. As detected false positive genes can be experimentally quite costly in follow up studies, we wanted to improve the accuracy of the prediction without increasing the number of replicates or time points required as this too would be experimentally costly if not infeasible.</p><p>We hypothesized that transcripts identified several times in resampled datasets contain more true positive and fewer false positive transcripts. To test this hypothesis, we generated 36 resampled datasets and identified oscillating genes by ARSER and HAYSTACK algorithms in each resampled dataset. Subsequently, we calculated the consensus of detected oscillating genes in these 36 resampled datasets. A consensus of 10 means that the genes were detected in at least 10 out of the 36 resampled datasets. The consensus graphs for the analysis performed with ARSER and HAYSTACK are shown in Figures&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref> and <xref rid="Fig3" ref-type="fig">3</xref>, respectively. We compared the number of true and false positives to the number of true and false positives found in the averaged dataset, as well as to the consensus between the initial datasets and the initial simulations. The initial simulations represent the ideal situation that samples could be retrieved from the same individual, this is, however, not possible for gene expression analysis in most cases. It nevertheless represents the maximal detectable number of true positive transcripts in a noisy dataset. As can be seen from Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>, up to a required consensus between 15 datasets, the resampled datasets show a larger number of true positives compared to average and the overlap between initial datasets. To acquire the same consensus the number of false positives is 8 of 8400. A similar number of false positives is found if a full overlap between the 3 initial datasets is required. The number of true positives for the latter is, however, much lower for all types of simulations. We can therefore conclude that the resampling of datasets increases the number of true positive oscillating transcripts detected in a dataset without increasing the number of false positives compared to the initial replicates. Except when very low consensus is required (less than 7 for resampled dataset and 2 for initial datasets), the number of false positives detected with ARSER is always higher for averaged datasets, and hence not well suited to reliably identify oscillating genes.<fig id="Fig2"><label>Figure 2</label><caption><p> <bold>Performance evaluation of the resampling method to identify oscillating transcripts.</bold> For LL-, LD- and ODE-based time courses 36 time courses were simulated (original simulations). From these original simulations 3 replicates (initial dataset) for each type of time course were generated by randomly selecting one time point from each of the original simulations to mimic experimental sampling procedures. To generate the averaged dataset, the expression values of the 3 replicates at each time point were averaged. The initial datasets were furthermore used to generate 36 resampled datasets by random sampling at each time point. All datasets were analyzed with ARSER and true positives <bold>(A-</bold> <bold>C)</bold> and false positives <bold>(D)</bold> were calculated requiring increasing consensus between the datasets (see <xref rid="Sec4" ref-type="sec">Methods</xref> for details). A consensus of 10 thereby means that a gene is found in at least 10 different resampled or originally simulated datasets. For the averaged time course the true and false positives calculated are displayed as a line for comparison.</p></caption><graphic xlink:href="12859_2014_352_Fig2_HTML" id="MO2"/></fig><fig id="Fig3"><label>Figure 3</label><caption><p> <bold>Analysis of resampled time courses with HAYSTACK.</bold> For LL-, LD- and ODE-based time courses 36 time courses were simulated (original simulations). From these original simulations 3 replicates (initial dataset) for each type of time course were generated by randomly selecting from each of the original simulations one time point to mimic experimental sampling procedures. To generate the averaged dataset, the expression values of the 3 replicates at each time point were averaged. The initial datasets were furthermore used to generate 36 resampled datasets by random sampling at each time point. All datasets were analyzed with HAYSTACK and true positives <bold>(A</bold>-<bold>C)</bold> and false positives <bold>(D)</bold> were calculated requiring increasing consensus between the datasets (see <xref rid="Sec4" ref-type="sec">Methods</xref> for details). A consensus of 10 thereby means that a gene is found at least in 10 different resampled or originally simulated datasets. For the averaged time course the true and false positives calculated are displayed as a line for comparison.</p></caption><graphic xlink:href="12859_2014_352_Fig3_HTML" id="MO3"/></fig></p><p>We next analyzed the influence of the number of resampled datasets on the detection of true and false positive oscillating transcripts. As can be seen in Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>, higher consensus is required for a higher number of resampled datasets but the consensus range in which no false positives are found and in which the number of true positives remains high, is larger when a larger number of randomized datasets are analyzed. For the analysis of real data we therefore chose to generate 70 resampled datasets. Unfortunately there are very few circadian datasets available with sufficient replicates and time points. We found one study with two replicates performed in two different mouse tissues (liver and muscle) [<xref ref-type="bibr" rid="CR19">19</xref>] and one other mouse study with 3 replicates [<xref ref-type="bibr" rid="CR20">20</xref>]. The overall number of oscillating genes found in the dataset from Miller <italic>et al</italic>. [<xref ref-type="bibr" rid="CR19">19</xref>] is similar to that reported in the original article. The overlap, however, was not analyzed in the original work and we only found 2 and 3 transcripts, respectively (Figure&#x000a0;<xref rid="Fig5" ref-type="fig">5</xref>A and B) in both replicates. Using our resampling approach we identified 74 and 96 genes when requiring consensus between at least 10 sets and 10 and 5 transcripts, respectively, if a consensus of 20 was required.<fig id="Fig4"><label>Figure 4</label><caption><p> <bold>Dependency of the analysis on the number of resampled dataset.</bold> 3 initial datasets were generated as described in Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref> and in the <xref rid="Sec4" ref-type="sec">Methods section</xref> and from these either 10, 36 or 70 resampled datasets were generated by random resampling and the number of true (TP) and false positive (FP) transcripts calculated. The detection of oscillating genes was performed with ARSER.</p></caption><graphic xlink:href="12859_2014_352_Fig4_HTML" id="MO4"/></fig><fig id="Fig5"><label>Figure 5</label><caption><p> <bold>Analysis of published expression data.</bold> We reanalyzed 2 published circadian datasets with 4&#x000a0;hour sampling intervals and 12 time points using ARSER. The dataset by Miller <italic>et al</italic>. [<xref ref-type="bibr" rid="CR19">19</xref>] contained only two replicates each for two mouse tissues (liver <bold>(A)</bold> and muscle <bold>(B)</bold>). The dataset from Na <italic>et al</italic>. [<xref ref-type="bibr" rid="CR20">20</xref>] <bold>(C)</bold> contained 3 replicates.</p></caption><graphic xlink:href="12859_2014_352_Fig5_HTML" id="MO5"/></fig></p><p>For the dataset from Na <italic>et al</italic>. [<xref ref-type="bibr" rid="CR19">19</xref>] 147 transcripts were found in all 3 initial replicates (Figure&#x000a0;<xref rid="Fig5" ref-type="fig">5</xref>C). In our resampled dataset 796 transcripts were identified as oscillating when we require a consensus of 10, 183 genes remain if we require a consensus of 20.</p><p>As the study by Na <italic>et al</italic>. resulted in a larger number of oscillating transcripts we used our simulated LL datasets to analyze how the number of replicates influences the number of true and false negatives and thus the accuracy of the detection of oscillating transcripts. To do so we initially simulated 72 datasets. Those were used to generate the different numbers of initial replicated datasets. The analysis showed that the number of oscillating transcripts detected for a full overlap between all replicates is decreasing with the number of replicates (Figure&#x000a0;<xref rid="Fig6" ref-type="fig">6</xref>A) with increasing consensus required. But starting from a required consensus of 4, false positives were no longer detected in the initial datasets (Figure&#x000a0;<xref rid="Fig6" ref-type="fig">6</xref>C), thus a consensus of 4 is sufficient to accurately detect oscillating transcripts for initial datasets. Taken this into account the amount of true positives transcripts is higher for higher numbers of replicates as would be expected. Looking at the resampled datasets we see that with increasing number of replicates lower consensus is required to avoid detection of false positives, emphasizing the importance of replicates for the detection of circadian regulated transcripts.<fig id="Fig6"><label>Figure 6</label><caption><p> <bold>Impact of the number of replicates on the accuracy.</bold> From 72 original simulations, either 2,3,4,5, or 6 replicates were generated by random sampling<bold> (A</bold> <bold> and </bold> <bold>C)</bold> and analyzed using the ARSER algorithm as described in the <xref rid="Sec4" ref-type="sec">Methods</xref> section. The initial datasets were then used to generate 36 resampled datasets <bold>(B</bold> <bold> and </bold> <bold>D)</bold> as described in Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref> and the <xref rid="Sec4" ref-type="sec">Methods</xref> section. The number of true positives <bold>(A and </bold> <bold>B)</bold> and false positives <bold>(C and </bold> <bold>D)</bold> for different consensus required is shown.</p></caption><graphic xlink:href="12859_2014_352_Fig6_HTML" id="MO6"/></fig></p></sec><sec id="Sec3" sec-type="conclusion"><title>Conclusions</title><p>In this analysis, we conclude that in comparison to single replicates and averaged datasets, our resampling method improves the detection of oscillating transcripts without increasing the number of false positives. The resampling method particularly outperformed the average method to reduce the number of false positive transcripts. Furthermore, the resampling method shows that biological replicates are important to accurately identify true oscillating transcripts using time series gene expression datasets, and that the average method may result in a large number of false positives. To reliably identify oscillating transcripts, resampled datasets should be generated from at least 3 experimental samples per time point.</p></sec><sec id="Sec4" sec-type="materials|methods"><title>Methods</title><sec id="Sec5"><title>Simulated time series</title><p>As there is no way to determine whether an algorithm can distinguish true oscillating transcripts (true positives) from non-oscillating transcripts (false positives) in a real gene expression dataset, we generated artificial time series to analyze the performance of different algorithms. The artificially generated time series contained the expression values of 8400 transcripts. To generate periodic patterns for synchronized datasets (LD dataset), we used the formula by Yang and Su [<xref ref-type="bibr" rid="CR13">13</xref>]. Thus the model is defined by:<disp-formula id="Equ1"><label>1</label><alternatives><tex-math id="M1">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document} $$ {x}_t=SNR\kern0.1em \cdotp \kern0.1em 2 cos\frac{2\pi }{\tau}\left(t-\varphi \right)+{\varepsilon}_t $$ \end{document}</tex-math><graphic xlink:href="12859_2014_352_Article_Equ1.gif" position="anchor"/></alternatives></disp-formula>where SNR&#x02009;=&#x02009;2 is the signal-to-noise ratio; &#x003c4; is the period in the range of 22 and 28&#x000a0;hours; &#x003d5; is phase (0-28&#x000a0;h with 0.1&#x000a0;h intervals); and &#x003b5;_t is the normally distributed noise term (mean =0 standard deviation =1).</p><p>Desynchronizing individuals under constant condition (for example constant light (LL)) were simulated using the above formula but with a fixed period that was randomly selected for each of the 36 initially simulated time courses. The periods were normally distributed with a mean of 25&#x000a0;hours and a standard deviation of 3&#x000a0;h according to published experimental data [<xref ref-type="bibr" rid="CR9">9</xref>].</p><p>For more realistic circadian simulations we used the ODE-model of the mammalian circadian oscillators by Leloup <italic>et al</italic>. [<xref ref-type="bibr" rid="CR18">18</xref>]. We first generated time courses for all variable model species and then generated phase shifted copies thereof. Phase shifts had 0.1&#x000a0;h intervals. From these time courses, datasets with 4&#x000a0;h sampling intervals were generated. Normally distributed white noise was added as for the cosine wave simulations. To simulate non-periodic time series, we used normally distributed white noise with the same mean and standard deviation as above.</p><p>The simulations described above were repeated 36 times for each type of data. If not described otherwise we generated 3 initial datasets from these time courses by randomly selecting once from each original simulation to generate a new 4&#x000a0;hour interval time course. This mimics the sampling procedure from different individuals in real experiments.</p><p>Python scripts used to generate time series and initial datasets are provided as Additional file <xref rid="MOESM1" ref-type="media">1</xref>.</p></sec><sec id="Sec6"><title>ARSER, HAYSTACK and BIODARE</title><p>Recently, Yang and Su developed the algorithm ARSER, which combines frequency domain and time domain analyses [<xref ref-type="bibr" rid="CR13">13</xref>]. The algorithm first removes any linear trend from time series data (data preprocessing), and then the period is determined by AR spectral analysis (period detection). Because the period can differ from 24&#x000a0;h depending on the experimental conditions, the algorithm takes a range from 20&#x000a0;h to 28&#x000a0;h into account. With each period, ARSER employs harmonic regression to determine the four cyclic parameters: period, amplitude, mean level, and phase (rhythm modeling). Finally, false discovery rate (FDR) <italic>q</italic>-values are calculated for multiple comparisons and the output was filtered and only those transcripts with a <italic>q</italic>-value greater than 0.05 were consider in the analysis.</p><p>To exclude the possibility that our results depend on the chosen algorithm, the analysis was repeated with the HAYSTACK algorithm [<xref ref-type="bibr" rid="CR21">21</xref>] and the FFT-NLLS algorithm implemented in BIODARE [<xref ref-type="bibr" rid="CR15">15</xref>,<xref ref-type="bibr" rid="CR16">16</xref>]. HAYSTACK was designed to find periodic patterns in any large-scale dataset representing at least three data points. The web version and 120 cycling patterns are available at <ext-link ext-link-type="uri" xlink:href="http://haystack.mocklerlab.org/">http://haystack.mocklerlab.org/</ext-link>. The HAYSTACK algorithm compares gene expression profiles with predefined cycling patterns. Different cutoffs are used to detect oscillating patterns in gene expression. The most important parameter is the correlation coefficient. The higher value means a higher correlation between the experimental data and the predefined models. A coefficient of +1 indicates perfect positive correlation. Other cutoff values are the fold change and <italic>p-</italic>value, and these values are used to achieve statistical significance. The HAYSTACK algorithm searches for at least six different patterns, including &#x0201c;asymmetric,&#x0201d; &#x0201c;rigid,&#x0201d; &#x0201c;spike,&#x0201d; &#x0201c;cosine,&#x0201d; &#x0201c;sine,&#x0201d; and &#x0201c;box-like&#x0201d; patterns. The models that most successfully identify rhythmically expressed genes are &#x0201c;cosine&#x0201d; and &#x0201c;spike.&#x0201d;</p><p>BIODARE and the implemented algorithms are described elsewhere [<xref ref-type="bibr" rid="CR15">15</xref>]. Shortly, FFT-NLLS (Fast Fourier Transform - Non-Linear Least Square) is a curve fitting method which models a sum of cosine functions and calculates confidence levels for period, phase and amplitude. The BIODARE FFT-NLLS algorithm detection was limited to period range from 20 to 28&#x000a0;h to match the period range of ARSER. Linear detrending was applied.</p></sec><sec id="Sec7"><title>Resampling</title><p>The artificially generated time series dataset consists of 12 time points at 4&#x000a0;h sampling intervals, representing 48&#x000a0;h of observation. To generate resampled datasets, expression values of each gene were randomly selected from (if not stated otherwise) three initial replicate time series, and the values were combined to generate the new resampled dataset. Each expression value has an equal probability of selection, and the time points are treated independently of one another. If not stated otherwise, the procedure was repeated 36 times, and we created 36 different resampled datasets (python script provided as Additional file <xref rid="MOESM1" ref-type="media">1</xref>). Each resampled dataset was analyzed by the ARSER algorithm with the stringency threshold (<italic>q</italic>-value) set to 0.05. HAYSTACK algorithm was used with the following parameter: p-value&#x02009;=&#x02009;0.05; fold change&#x02009;=&#x02009;2.0, correlation cutoff&#x02009;=&#x02009;0.8; and background cutoff&#x02009;=&#x02009;0.01. Using the oscillating transcripts detected the consensus between the 36 resampled datasets were calculated.</p></sec></sec>

Feasibility and effectiveness of drop-off spots to promote walking to school

Could not extract abstract

<contrib contrib-type="author" corresp="yes"><name><surname>Vanwolleghem</surname><given-names>Griet</given-names></name><address><email>Griet.Vanwolleghem@Ugent.be</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author"><name><surname>D&#x02019;Haese</surname><given-names>Sara</given-names></name><address><email>Sara.DHaese@Ugent.be</email></address><xref ref-type="aff" rid="Aff1"/><xref ref-type="aff" rid="Aff2"/></contrib><contrib contrib-type="author"><name><surname>Van Dyck</surname><given-names>Delfien</given-names></name><address><email>Delfien.VanDyck@Ugent.be</email></address><xref ref-type="aff" rid="Aff1"/><xref ref-type="aff" rid="Aff2"/></contrib><contrib contrib-type="author"><name><surname>De Bourdeaudhuij</surname><given-names>Ilse</given-names></name><address><email>Ilse.DeBourdeaudhuij@Ugent.be</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author"><name><surname>Cardon</surname><given-names>Greet</given-names></name><address><email>Greet.Cardon@Ugent.be</email></address><xref ref-type="aff" rid="Aff1"/></contrib><aff id="Aff1"><label/>Department of Movement and Sport Sciences, Faculty of Medicine and Health Sciences, Ghent University, Watersportlaan 2, 9000 Ghent, Belgium </aff><aff id="Aff2"><label/>Research Foundation Flanders (FWO), Egmontstraat 5, 1000 Brussels, Belgium </aff>

The International Journal of Behavioral Nutrition and Physical Activity

<sec id="Sec1" sec-type="introduction"><title>Background</title><p>Engaging in walking and cycling to school is an important source of daily physical activity in 6-to-12-year olds [<xref ref-type="bibr" rid="CR1">1</xref>-<xref ref-type="bibr" rid="CR3">3</xref>]. Despite the numerous health benefits of active transport, many primary schoolchildren do not walk or cycle to school [<xref ref-type="bibr" rid="CR4">4</xref>-<xref ref-type="bibr" rid="CR6">6</xref>]. In some European countries (e.g. Belgium, the Netherlands, Denmark), the proportion of children that commutes actively to school is higher compared to other (non-) European countries (e.g. US, Australia, Spain, &#x02026;) [<xref ref-type="bibr" rid="CR7">7</xref>]. However, in Flanders (northern part of Belgium), still 47% of 6-to-12-year olds are driven to school by car [<xref ref-type="bibr" rid="CR8">8</xref>]. Eleven to twelve year old Flemish children commute actively to school more frequently, but still 41% are driven to school by car [<xref ref-type="bibr" rid="CR9">9</xref>].</p><p>Parental safety concerns (road safety and perceived danger from strangers) have been identified as important barriers for children&#x02019;s active commuting to school [<xref ref-type="bibr" rid="CR10">10</xref>,<xref ref-type="bibr" rid="CR11">11</xref>]. Therefore, previous interventions promoting active transport to school mainly focused on safety issues [<xref ref-type="bibr" rid="CR12">12</xref>,<xref ref-type="bibr" rid="CR13">13</xref>]. The &#x0201c;Safe Routes to School&#x0201d; intervention [<xref ref-type="bibr" rid="CR14">14</xref>] in the US aimed to provide several safe routes to school (funding the construction of safe pathways to school, providing crossing guards at major intersections, &#x02026;) and to provide support for schools for traffic safety education and organization of events. Furthermore, interventions like &#x0201c;Walking School Bus&#x0201d; [<xref ref-type="bibr" rid="CR15">15</xref>-<xref ref-type="bibr" rid="CR17">17</xref>] and &#x0201c;Bicycle Train&#x0201d; [<xref ref-type="bibr" rid="CR18">18</xref>] provided adult supervision (by teachers, parents, &#x02026;) during the active trip to school to deal with parents&#x02019; safety concerns. These programs have a fixed route with designated &#x0201c;stops&#x0201d; and &#x0201c;pick up times&#x0201d; where children can join a supervised group to walk or cycle to school. In a systematic review of Chill&#x000f3;n and colleagues [<xref ref-type="bibr" rid="CR13">13</xref>], all interventions promoting active transport reported an increase in the percentage of active transport to school (ranging from 3% to 64%). Additionally, 6 of the 14 interventions reported a small effect size (Cohen&#x02019;s d between 0.2 and 0.5) on active transport outcomes. Furthermore, Walking school bus interventions were effective in increasing walking to school (25%) among primary schoolchildren in the US and were positively evaluated [<xref ref-type="bibr" rid="CR12">12</xref>,<xref ref-type="bibr" rid="CR19">19</xref>]. Moreover, important benefits of walking school buses included strong social benefits, safety benefits and time-savings [<xref ref-type="bibr" rid="CR16">16</xref>]. Boarnet and colleagues [<xref ref-type="bibr" rid="CR20">20</xref>] demonstrated that the &#x0201c;Safe Routes to School&#x0201d; program was effective when the project was along the child&#x02019;s usual route to school.</p><p>Besides parental safety concerns, the home-school distance has also been identified as an important predictor of children&#x02019;s active commuting to school [<xref ref-type="bibr" rid="CR6">6</xref>,<xref ref-type="bibr" rid="CR9">9</xref>,<xref ref-type="bibr" rid="CR11">11</xref>,<xref ref-type="bibr" rid="CR21">21</xref>]. The home-school distance is negatively associated with active commuting to school, but it must be acknowledged that the distances, found to be feasible for active commuting to school, differ between countries and between environments. A study conducted in Australia reported that 6-to-12-year olds are more likely to commute actively to school if the home-school distance is less than 800 meters [<xref ref-type="bibr" rid="CR22">22</xref>]. In older primary schoolchildren (11&#x02013;12 years) in Flanders, D&#x02019;Haese et al. [<xref ref-type="bibr" rid="CR9">9</xref>] reported criterion distances of 1.5&#x000a0;km for walking and 3&#x000a0;km for cycling to school. In the latter study, 53% of the passive commuters to school lived further from school than the feasible active commuting distance (3&#x000a0;km). When developing interventions promoting active commuting to school it is important to also include those children living further away from school.</p><p>As the home-school distance is not easily modifiable, previous interventions (Walking School Bus&#x0201d; [<xref ref-type="bibr" rid="CR15">15</xref>-<xref ref-type="bibr" rid="CR17">17</xref>], &#x0201c;Bicycle Train&#x0201d; [<xref ref-type="bibr" rid="CR18">18</xref>]) focused mainly on children living within a feasible walking or cycling distance from school [<xref ref-type="bibr" rid="CR13">13</xref>]. In some previous walking school bus programs, children living further away could be dropped off by their parents along the route to join the walking school bus [<xref ref-type="bibr" rid="CR12">12</xref>]. However, having to match the timing of the drop-off of the child with the timing of the walking school bus can be an important barrier for parents.</p><p>Some studies indicated that drop-off spots may offer a solution to increase the daily walking in primary schoolchildren who are usually driven to school by their parents [<xref ref-type="bibr" rid="CR9">9</xref>,<xref ref-type="bibr" rid="CR23">23</xref>]. A drop-off spot is a location within a feasible walking distance from the school where parents can drop off or pick up their child before or after school hours. From this spot, children can walk to or from school independently or under supervision of teachers, parents or other volunteers. Drop-off spots may be an alternative for children who cannot actively commute to school due to the large home-school distance [<xref ref-type="bibr" rid="CR9">9</xref>,<xref ref-type="bibr" rid="CR23">23</xref>]. Additionally, in a drop-off spot intervention, safety issues can be taken into account in order to have children walking safely to school.</p><p>To our knowledge, no studies previously evaluated drop-off spots for the promotion of walking to school among primary schoolchildren. Before implementing drop-off spots, it is important to identify possible barriers, opportunities and practical concerns towards a drop-off spot intervention. Therefore, the first aim of this pilot study was to investigate the parental opinions concerning the feasibility of drop-off spots to promote walking to school among primary schoolchildren. When developing the intervention in collaboration with the schools (specifically developed for each school), those parental opinions were taken into account. A second aim of this study was to examine the effectiveness of drop-off spots on children&#x02019;s step counts and walking trips to and from school. A third aim was to study how the implemented drop-off spots were perceived by parents, teachers and school principals.</p></sec><sec id="Sec2" sec-type="materials|methods"><title>Methods</title><sec id="Sec3"><title>Participants and procedure</title><p>In Spring 2013, a convenience sample (n&#x02009;=&#x02009;8) of primary schools in West-Flanders (northwestern part of Belgium) was contacted by phone until two primary schools agreed to participate (one located in a suburban area, 150&#x02013;500 residents/km² (pupils: n&#x02009;=&#x02009;85), one located in an urban area, &#x0003e;500 residents/km² (pupils: n&#x02009;=&#x02009;228)).</p><sec id="Sec4"><title>Prior to the intervention: development and feasibility</title><p>Before the implementation of drop-off spots, two meetings with teachers and principals were organized in each school. The school staff was closely involved in developing the intervention to ensure that the intervention was tailored to the needs of each of the schools. During the first meeting, the protocol of the study was explained and the possibilities and barriers towards drop-off spots were discussed. In a second meeting, practical issues regarding the implementation of a drop-off spot were discussed and a specific proposal of the location, distance and organization of the drop-off spot was defined to propose to the parents. Based on the topics discussed during the school meetings, a feasibility questionnaire was developed to obtain parental opinions towards the implementation of drop-off spots. This questionnaire also included a school-specific proposal of the location of the drop-off spot. The feasibility questionnaires were given to all parents two weeks before baseline measurements and were distributed and collected through the schools. Parents of 313 primary schoolchildren (6&#x02013;12 years) received a questionnaire. In total, 216 parents (69%) completed the feasibility questionnaire (suburban school (n&#x02009;=&#x02009;56), urban school (n&#x02009;=&#x02009;160)). Based on those parental opinions and needs from each school, the intervention (organizing a drop-off spot near each school) was developed.</p></sec><sec id="Sec5"><title>Intervention</title><p>A within-subject design was used to study children&#x02019;s step counts and number of walking trips to and from school in usual conditions (baseline) and during the implementation of a drop-off spot (intervention). In each school, one drop-off spot was implemented during one school week. In total, 141 children (6&#x02013;12 years) were eligible to participate in the intervention study. Children were eligible if they used passive transport to school at least once a week, indicated by the parents in an additional question in the feasibility questionnaire. Parental informed consent to wear a pedometer during baseline and intervention was obtained from 60 parents of the 141 eligible children (response rate 43%). Both measurement periods (baseline and intervention) lasted one school week (Monday until Friday). During both weeks, children wore a pedometer and parents filled out a diary (11&#x02013;12 year old children completed the diary independently [<xref ref-type="bibr" rid="CR24">24</xref>]). There was a period of three to four weeks between baseline (April 2013) and intervention (May 2013). The weather conditions were similar in both measurement periods.</p><p>A teacher was present before the children arrived at the drop-off spot, waited for the children to arrive and walked together from the drop-off spot to school at an appointed time. Parents were asked to drop off their child during a specified time period. Parents were notified that after the appointed time, teacher supervision was no longer present. Besides the children who arrived by car, also children who already walked or cycled to school could stop at the drop-off spot and could walk together with the other children to school under supervision of a teacher. Children with a bike had to walk their bike. The organization of the drop-off spot was flexible and based on what each school indicated as feasible. Both schools organized the drop-off spot somewhat differently. In the suburban school, a drop-off spot was organized only before school hours. Parents could drop off their children between 8:15 and 8:25&#x000a0;AM. The drop-off spot was located in a residential area and parents could drop off their children in a cul-de-sac. In the urban school, children could use the drop-off spot before and after school hours. Before school hours, parents could drop off their children between 8:00 and 8:20&#x000a0;AM. Because of a high number of children from the urban school participated in the study, two departure times were organized to walk to school (a first group of children departed at 8:10&#x000a0;AM, the second at 8:20&#x000a0;AM). The urban school decided to organize the drop-off spot also after school hours because of the traffic congestion after school hours in the street of the school. Parents could pick up their child at the drop-off spot around 4:10&#x000a0;PM. In Belgium, primary schools run until 12:00&#x000a0;PM on Wednesdays. Because of practical limitations, the urban school decided not to organize the drop-off spot after school hours on Wednesday. The drop-off spot was located at a square along an approach road and was separated from the road. Prior to the intervention, a flyer with information was given in both schools to the children to hand to the parents. The information included the exact location of the drop-off spot and specific time periods when parents could drop off their children, the fact that a teacher would be present at the drop-off spot and would walk with the children to school. Parents were also informed that when they arrived later at the drop-off spot, teacher supervision was no longer present.</p></sec><sec id="Sec6"><title>After the intervention: perception of the intervention (process evaluation)</title><p>Within one week after the intervention, a questionnaire was given to the school principals, teachers and parents to collect data on how they perceived the intervention. Parents of 119 children were eligible to fill out the process evaluation questionnaire, including parents of all children wearing pedometers and parents of children not wearing pedometers but using the drop-off spot at least once a week. In total, both school principals, nine teachers (response rate 36%) and 44 parents of the 119 eligible children (response rate 37%) filled out the process evaluation questionnaire after the intervention. The present study was approved by the Ghent University Ethics Committee (EC UZG 2013/228).</p></sec></sec><sec id="Sec7"><title>Measurements</title><sec id="Sec8"><title>Parental feasibility questionnaire</title><p>The first section of the questionnaire contained general questions about the child (age, sex) and the parents (educational level of parents) to obtain socio-demographic information. Educational level of the parents was used as a proxy measure of children&#x02019;s socio-economic status (SES). The educational level was asked for mothers and fathers separately and was based on four options: completed elementary school, completed secondary school, completed college or completed university. Children were identified as being of high SES when at least one parent reached a college or university level, or of low SES when both parents did not reach a college or university level. Secondly, parents were asked to report their child&#x02019;s mode of transport to school using a question matrix [<xref ref-type="bibr" rid="CR25">25</xref>]. In this matrix, parents filled out per season on how many days per week their child went to school using different transport modes ((1) walking, (2) cycling, (3) driven by car and (4) using public transport). In a third part, parental opinions towards the implementation of drop-off spots (feasibility) were assessed. This part consisted of 16 questions concerning general characteristics of drop-off spots (benefits, barriers, environment, use). The specific questions with the corresponding response options are outlined in Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>.<table-wrap id="Tab1"><label>Table 1</label><caption><p> <bold>Parental feasibility towards implementation of drop-off spots</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th/><th/><th colspan="2"> <bold>School</bold> </th></tr><tr valign="top"><th/><th> <bold>All (n&#x02009;=&#x02009;216)</bold> </th><th> <bold>Suburban (n&#x02009;=&#x02009;56)</bold> </th><th> <bold>Urban (n&#x02009;=&#x02009;160)</bold> </th></tr><tr valign="top"><th/><th> <bold>% agree</bold> </th><th> <bold>% agree</bold> </th><th> <bold>% agree</bold> </th></tr></thead><tbody><tr valign="top"><td> <bold><italic>Benefits</italic></bold> ^{<bold><italic>1</italic></bold>} </td><td/><td/><td/></tr><tr valign="top"><td>Implementing a drop-off spot will be beneficial to children&#x02019;s health</td><td>78.9</td><td>80.0</td><td>78.5</td></tr><tr valign="top"><td>Children will enjoy the intervention</td><td>57.5</td><td>71.4^a </td><td>52.5</td></tr><tr valign="top"><td>Children will have more social contact because of the intervention</td><td>89.7</td><td>92.7</td><td>88.6</td></tr><tr valign="top"><td>I will save time when a drop-off spot will be implemented</td><td>68.3</td><td>45.5^a </td><td>76.5</td></tr><tr valign="top"><td> <bold><italic>Barriers</italic></bold> ^{<bold><italic>2</italic></bold>} </td><td/><td/><td/></tr><tr valign="top"><td>Weather</td><td>58.3</td><td>64.3</td><td>56.3</td></tr><tr valign="top"><td>Lack of time</td><td>34.7</td><td>19.6</td><td>40.0</td></tr><tr valign="top"><td> <bold><italic>Environment</italic></bold> ^{<bold><italic>1</italic></bold>} </td><td/><td/><td/></tr><tr valign="top"><td>Only a kiss and ride space should be available when implementing a drop-off spot</td><td>75.4</td><td>83.6</td><td>72.4</td></tr><tr valign="top"><td>There should be green space in the surroundings of the drop-off spot</td><td>78.9</td><td>75.9</td><td>80.0</td></tr><tr valign="top"><td>There should be no busy road in the surrounding of the drop-off spot</td><td>91.9</td><td>96.4</td><td>90.4</td></tr><tr valign="top"><td>The drop-off spot should be separated from the road (not only on sidewalk just next to the road)</td><td>95.2</td><td>92.5</td><td>96.1</td></tr><tr valign="top"><td>It is necessary that children do not have to cross over along the route from the drop-off spot to school</td><td>87.0</td><td>90.7</td><td>85.7</td></tr><tr valign="top"><td>The location of the drop-off spot should be on the route to parent&#x02019;s work</td><td>90.8</td><td>83.0 ^a </td><td>93.5</td></tr><tr valign="top"><td> <bold><italic>Use</italic></bold> </td><td/><td/><td/></tr><tr valign="top"><td>Is adult supervision necessary when arriving at the drop-off spot?</td><td/><td/><td/></tr><tr valign="top"><td>&#x02003;&#x02009; <italic>Never</italic> </td><td>0.9</td><td>0.0</td><td>1.3</td></tr><tr valign="top"><td>&#x02003;&#x02009; <italic>Sometimes</italic> </td><td>13.7</td><td>14.3</td><td>13.5</td></tr><tr valign="top"><td>&#x02003;&#x02009; <italic>Often</italic> </td><td>9.0</td><td>8.9</td><td>9.0</td></tr><tr valign="top"><td>&#x02003;&#x02009; <italic>Always</italic> </td><td>76.4</td><td>76.8</td><td>76.3</td></tr><tr valign="top"><td>Is adult supervision necessary during the route to school?</td><td/><td/><td/></tr><tr valign="top"><td>&#x02003;&#x02009; <italic>No</italic> </td><td>4.2</td><td>8.9</td><td>2.6</td></tr><tr valign="top"><td>&#x02003;&#x02009; <italic>Yes, for all children</italic> </td><td>45.8</td><td>42.9</td><td>46.8</td></tr><tr valign="top"><td>&#x02003;&#x02009; <italic>Yes, but only for youngest children (6-9 years)</italic> </td><td>50.0</td><td>48.2</td><td>50.6</td></tr><tr valign="top"><td>When (time of the day) would you use a drop-off spot?</td><td/><td/><td/></tr><tr valign="top"><td>&#x02003;&#x02009; <italic>Never</italic> </td><td>6.2</td><td>9.3</td><td>5.1</td></tr><tr valign="top"><td>&#x02003; <italic>Only before school</italic> </td><td>15.2</td><td>13.0</td><td>16.0</td></tr><tr valign="top"><td>&#x02003; <italic>Only after school</italic> </td><td>5.7</td><td>5.6</td><td>5.8</td></tr><tr valign="top"><td>&#x02003; <italic>Before and after school</italic> </td><td>72.9</td><td>72.2</td><td>73.1</td></tr><tr valign="top"><td>When (time of the year) would you use a drop-off spot?</td><td/><td/><td/></tr><tr valign="top"><td>&#x02003; <italic>Never</italic> </td><td>35.0</td><td>31.5^a </td><td>36.2</td></tr><tr valign="top"><td>&#x02003; <italic>Entire school year</italic> </td><td>30.0</td><td>24.1^a </td><td>32.2</td></tr><tr valign="top"><td>&#x02003; <italic>Spring/Summer</italic> </td><td>29.1</td><td>42.6^a </td><td>24.2</td></tr><tr valign="top"><td>&#x02003; <italic>Autumn/Winter</italic> </td><td>5.9</td><td>1.9^a </td><td>7.4</td></tr></tbody></table><table-wrap-foot><p> ¹scored on a 5-point Likert scale ranging from totally disagree to totally agree (% of response options &#x02018;rather agree&#x02009;+&#x02009;totally agree&#x02019; shown in table).</p><p> ²scored yes/no (% of response option &#x02018;yes&#x02019; shown in table).</p><p> ^asignificantly different from urban school.</p></table-wrap-foot></table-wrap></p><p>In a last part of the questionnaire, parental opinions towards a school-specific proposal of the organisation of the drop-off spot were asked.</p></sec><sec id="Sec9"><title>Step counts and self-reported transport to school</title><p>Weekday step counts were objectively assessed during the baseline and intervention week using a pedometer (Omron Walking Style Pro). This pedometer has been validated to measure step counts in children [<xref ref-type="bibr" rid="CR26">26</xref>] and it provides an hourly summary of the steps taken. Children wore a pedometer during 5 consecutive school days (Monday until Friday), at baseline and during the intervention. Children were asked to wear the pedometer during waking hours and to remove the pedometer for aquatic activities (e.g. swimming, showering) and for activities that prohibit the pedometers (e.g. contact sports).</p><p>During the intervention week, the daily number of times using the drop-off spot had to be reported in a diary, adding the reason for possible non-use. At baseline and during the intervention, parents of the 6&#x02013;10 year old children were asked to report their child&#x02019;s daily transport mode to school and the activities for which the pedometer was removed in the diary. The 11-to-12-year old children completed the diary independently. For every minute of reported moderate-to-vigorous physical activity for which the pedometer was removed, 150 steps were added to the daily number of step counts [<xref ref-type="bibr" rid="CR27">27</xref>].</p><p>In total, 58 children had valid pedometer data at both baseline and intervention measurements (minimum 3 school days excluding Wednesdays) and were included in the analyses. Total step counts during the entire day and step counts before and after school hours (7:00 to 9:00&#x000a0;AM/4:00&#x000a0;PM to 5:00&#x000a0;PM) were calculated. On Fridays, step counts after school hours were calculated from 3:00&#x000a0;PM to 4:00&#x000a0;PM, as the primary schools end at 3:00&#x000a0;PM on Fridays. Walking trips to and from school were obtained through the diaries. A walking trip was identified when a child walked to or from school, also when combined with another transport mode. To calculate step counts before and after school hours and walking trips to and from school, only step counts before school hours and walking trips to school were included for children from the suburban school as the suburban school only organized a drop-off spot before school hours. Step counts before and after school hours and walking trips to and from school were included for children from the urban school, because the urban school organized a drop-off spot twice a day. Total step counts per day, step counts per day before and after school hours and weekly number of walking trips to and from school were used as main outcomes to study intervention effects.</p></sec><sec id="Sec10"><title>Questionnaire on perception of the intervention (process evaluation)</title><p>To obtain information on how the drop-off spots were perceived, school principals, teachers and parents filled out a questionnaire. This questionnaire consisted of specific questions for school principals, teachers and parents concerning the usefulness and benefits of the drop-off spot intervention, experienced opportunities and difficulties during the intervention and future possibilities for the intervention.</p><p>The specific questions of the questionnaire for school principals, teachers and parents with the corresponding response options are outlined in Table&#x000a0;<xref rid="Tab2" ref-type="table">2</xref>.<table-wrap id="Tab2"><label>Table 2</label><caption><p> <bold>Perception of the intervention by the school principals, teachers and parents (n&#x02009;=&#x02009;53)</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th/><th> <bold>School principals (n&#x02009;=&#x02009;2)</bold> </th><th> <bold>Teachers (n&#x02009;=&#x02009;7)</bold> </th><th> <bold>Parents (n&#x02009;=&#x02009;44)</bold> </th></tr><tr valign="top"><th/><th> <bold>n agree</bold> </th><th> <bold>n agree</bold> </th><th> <bold>n (% ) agree</bold> </th></tr></thead><tbody><tr valign="top"><td> <bold><italic>Usefulness of intervention</italic></bold> ^{<bold><italic>1</italic></bold>} </td><td/><td/><td/></tr><tr valign="top"><td>The intervention was well organized</td><td>---</td><td>7</td><td>35 (92.1)</td></tr><tr valign="top"><td>It is possible to use the intervention in the future</td><td>2</td><td>2</td><td>31 (81.6)</td></tr><tr valign="top"><td>The school has to pay more attention to safety when organizing a drop-off spot compared to the usual conditions</td><td>2</td><td>6</td><td>---</td></tr><tr valign="top"><td> <bold><italic>Benefits</italic></bold> ^{<bold><italic>1</italic></bold>} </td><td/><td/><td/></tr><tr valign="top"><td>The intervention gives children, who are usually dropped off by car, the opportunity to walk to school</td><td>1</td><td>6</td><td>---</td></tr><tr valign="top"><td>Children enjoyed the intervention</td><td>1</td><td>7</td><td>31 (86.1)</td></tr><tr valign="top"><td>Children could have more social contact with others because of this intervention</td><td>1</td><td>4</td><td>23 (65.7)</td></tr><tr valign="top"><td> <bold><italic>Difficulties during intervention</italic></bold> ^{<bold><italic>2</italic></bold>} </td><td/><td/><td/></tr><tr valign="top"><td>Busy traffic on the way to the drop-off spot</td><td>---</td><td>---</td><td>2 (6.8)</td></tr><tr valign="top"><td>The time when the drop-off spot was organized did not fit</td><td>---</td><td>---</td><td>9 (20.5)</td></tr><tr valign="top"><td>Resistance teachers</td><td>0</td><td>---</td><td>---</td></tr><tr valign="top"><td>Resistance parents</td><td>1</td><td>1</td><td>---</td></tr><tr valign="top"><td>Resistance children</td><td>0</td><td>0</td><td>---</td></tr><tr valign="top"><td>Organizational limitations (e.g. willingness volunteers)</td><td>0</td><td>2</td><td>---</td></tr><tr valign="top"><td>School environment (e.g. busy traffic in the surrounding of the school environment)</td><td>0</td><td>2</td><td>---</td></tr><tr valign="top"><td>The intervention requires an additional load for the teachers</td><td>---</td><td>2</td><td>---</td></tr><tr valign="top"><td> <bold><italic>Opportunities for intervention</italic></bold> </td><td/><td/><td/></tr><tr valign="top"><td>How often would you continue to use this intervention?</td><td/><td/><td/></tr><tr valign="top"><td>&#x02003;&#x02009; <italic>Never</italic> </td><td>1</td><td>3</td><td>3 (7.0)</td></tr><tr valign="top"><td>&#x02003;&#x02009; <italic>1</italic>&#x02013;<italic>2 times per week</italic> </td><td>0</td><td>3</td><td>11 (25.6)</td></tr><tr valign="top"><td>&#x02003;&#x02009; <italic>3</italic>&#x02013;<italic>4 times per week</italic> </td><td>0</td><td>0</td><td>9 (20.9)</td></tr><tr valign="top"><td>&#x02003;&#x02009; <italic>Every day</italic> </td><td>1</td><td>1</td><td>20 (46.5)</td></tr><tr valign="top"><td>When (time of the day) would you continue to use this intervention?</td><td/><td/><td/></tr><tr valign="top"><td>&#x02003;&#x02009; <italic>Never</italic> </td><td>0</td><td>0</td><td>2 (4.7)</td></tr><tr valign="top"><td>&#x02003;&#x02009; <italic>Only before school</italic> </td><td>1</td><td>4</td><td>19 (44.2)</td></tr><tr valign="top"><td>&#x02003;&#x02009; <italic>Only after school</italic> </td><td>0</td><td>0</td><td>8 (18.6)</td></tr><tr valign="top"><td>&#x02003;&#x02009; <italic>Before and after school</italic> </td><td>1</td><td>3</td><td>14 (32.6)</td></tr><tr valign="top"><td>When (time of the year) would you continue to use this intervention?</td><td/><td/><td/></tr><tr valign="top"><td>&#x02003;&#x02009; <italic>Never</italic> </td><td>0</td><td>0</td><td>2 (4.7)</td></tr><tr valign="top"><td>&#x02003;&#x02009; <italic>Only during theme-related periods at school</italic> </td><td>0</td><td>3</td><td>0</td></tr><tr valign="top"><td>&#x02003;&#x02009; <italic>Entire school year</italic> </td><td>1</td><td>3</td><td>20 (46.5)</td></tr><tr valign="top"><td>&#x02003;&#x02009; <italic>Spring/Summer</italic> </td><td>1</td><td>1</td><td>21 (48.8)</td></tr><tr valign="top"><td>&#x02003;&#x02009; <italic>Autumn/Winter</italic> </td><td>0</td><td>0</td><td>0</td></tr><tr valign="top"><td>For which target group can this intervention be used in the future?</td><td/><td/><td/></tr><tr valign="top"><td>&#x02003;&#x02009; <italic>Nobody</italic> </td><td>0</td><td>0</td><td>2 (4.7)</td></tr><tr valign="top"><td>&#x02003;&#x02009; <italic>Only for oldest children (10&#x02013;12 years)</italic> </td><td>0</td><td>0</td><td>3 (7.0)</td></tr><tr valign="top"><td>&#x02003;&#x02009; <italic>All ages</italic> </td><td>2</td><td>7</td><td>38 (88.3)</td></tr><tr valign="top"><td>Is adult supervision necessary during the route to school in the future?</td><td/><td/><td/></tr><tr valign="top"><td>&#x02003;&#x02009; <italic>No</italic> </td><td>---</td><td>---</td><td>1 (2.4)</td></tr><tr valign="top"><td>&#x02003;&#x02009; <italic>Yes, for all children</italic> </td><td>---</td><td>---</td><td>11 (26.2)</td></tr><tr valign="top"><td>&#x02003;&#x02009;Yes, only <italic>for the youngest children (6&#x02013;9 years)</italic> </td><td>---</td><td>---</td><td>30 (71.4)</td></tr></tbody></table><table-wrap-foot><p> ¹scored on a 5-point Likert scale ranging from totally disagree to totally agree (% of response options &#x02018;rather agree&#x02009;+&#x02009;totally agree&#x02019; shown in table).</p><p> ²scored yes/no (% of response option &#x02018;yes&#x02019; shown in table).</p></table-wrap-foot></table-wrap></p></sec></sec><sec id="Sec11"><title>Data analysis</title><p>SPSS for Windows version 21 (SPSS Inc., Chicago, IL, USA) was used to describe the characteristics of the different samples. Descriptive statistics were used to describe the parental opinions towards the implementation of drop-off spots (feasibility) and the perception of the intervention by the school principals, teachers and parents. Additionally, chi square tests were conducted to test associations between parental opinions and the school (suburban and urban).</p><p>To determine intervention effects on total step counts/day, step counts/day before and after school hours and number of walking trips to and from school/week, three-level (class-participant-condition (i.e. baseline/intervention)) linear regression models with random intercept and fixed slope were conducted using MLwiN version 2.29. As only two primary schools were included, the clustering of participants within schools was not included as a level [<xref ref-type="bibr" rid="CR28">28</xref>,<xref ref-type="bibr" rid="CR29">29</xref>]. All analyses were controlled for age (continuous), sex, SES and school. When examining total step counts/day and step counts/day before and after school hours, analyses were controlled for pedometer wear time. Wednesdays were excluded from the analyses since no valid data were obtained at baseline and/or during the intervention (holiday, half day of school). The significance level was set at p&#x02009;&#x0003c;&#x02009;0.05.</p></sec></sec><sec id="Sec12" sec-type="results"><title>Results</title><sec id="Sec13"><title>Description of the samples</title><p>The characteristics of the different samples are shown in Table&#x000a0;<xref rid="Tab3" ref-type="table">3</xref>. Of the 216 children whose parents filled out the parental questionnaire before the intervention, 48.1% (n&#x02009;=&#x02009;104) were boys. Twenty-five percent went to the suburban school (n&#x02009;=&#x02009;56), the other 74.1% (n&#x02009;=&#x02009;160) to the urban school. Mean age was 9.6&#x02009;&#x000b1;&#x02009;1.7&#x000a0;years. In total, 11.6% of the children (n&#x02009;=&#x02009;25) mostly cycled to school, 27.4% (n&#x02009;=&#x02009;59) mostly walked to school, 48.4% (n&#x02009;=&#x02009;104) were mostly dropped off by car and 12.6% (n&#x02009;=&#x02009;27) mostly used public transport as travel mode.<table-wrap id="Tab3"><label>Table 3</label><caption><p> <bold>Descriptive characteristics of the samples</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th/><th> <bold>Parental feasibility sample (n&#x02009;=&#x02009;216)</bold> </th><th> <bold>Intervention sample (n&#x02009;=&#x02009;58)</bold> </th></tr></thead><tbody><tr valign="top"><td>Age (years) (Mean&#x02009;&#x000b1;&#x02009;SD)</td><td>9.6&#x02009;&#x000b1;&#x02009;1.7</td><td>9.6&#x02009;&#x000b1;&#x02009;1.7</td></tr><tr valign="top"><td>Sex (n, %)</td><td/><td/></tr><tr valign="top"><td>&#x02003;&#x02009; <italic>Boys</italic> </td><td>104 (48.1)</td><td>22 (37.9)</td></tr><tr valign="top"><td>&#x02003; <italic>Girls</italic> </td><td>112 (51.9)</td><td>36 (62.1)</td></tr><tr valign="top"><td>School (n, %)</td><td/><td/></tr><tr valign="top"><td>&#x02003; <italic>Suburban</italic> </td><td>56 (25.9)</td><td>29 (50.0)</td></tr><tr valign="top"><td>&#x02003; <italic>Urban</italic> </td><td>160 (74.1)</td><td>29 (50.0)</td></tr><tr valign="top"><td>SES (n, %)</td><td/><td/></tr><tr valign="top"><td>&#x02003; <italic>Low</italic> </td><td>114 (53.0)</td><td>30 (51.7)</td></tr><tr valign="top"><td>&#x02003; <italic>High</italic> </td><td>101 (47.0)</td><td>28 (48.3)</td></tr><tr valign="top"><td>Transport mode to school (n, %)</td><td/><td/></tr><tr valign="top"><td>&#x02003; <italic>Walking</italic> </td><td>59 (27.4)</td><td>4 (7.0)</td></tr><tr valign="top"><td>&#x02003; <italic>Cycling</italic> </td><td>25 (11.6)</td><td>2 (3.5)</td></tr><tr valign="top"><td>&#x02003; <italic>Driven by car</italic> </td><td>104 (48.4)</td><td>40 (70.2)</td></tr><tr valign="top"><td>&#x02003; <italic>Public transport</italic> </td><td>27 (12.6)</td><td>11 (19.3)</td></tr></tbody></table></table-wrap></p><p>In total, 58 children had valid pedometer data at baseline and during the intervention week. This sample consisted of 22 boys (37.9%) and 36 girls (62.1%). In total, 51.7% (n&#x02009;=&#x02009;30) had a low SES. Mean age was 9.7&#x02009;&#x000b1;&#x02009;1.6&#x000a0;years. The demographic characteristics (age, sex, SES) of the subsample of children with valid pedometer data (n&#x02009;=&#x02009;58) were comparable with those of the sample of children who dropped out (n&#x02009;=&#x02009;83) (no consent for participation, no valid pedometer data), except that the proportion of children going to an urban school was higher for the drop out sample. In the suburban school, 56% (n&#x02009;=&#x02009;14) of the children used the drop-off spot every day before school hours during the intervention. In the urban school, 15.4% (n&#x02009;=&#x02009;4) used the drop-off spot every day only before school hours, 11.5% (n&#x02009;=&#x02009;3) only after schools hours and 7.7% (n&#x02009;=&#x02009;2) before and after school hours. Of all children, 26.5% (n&#x02009;=&#x02009;13) never used the drop-off spot (12.0% (n&#x02009;=&#x02009;3) in the suburban school; 38.5% (n&#x02009;=&#x02009;10) in the urban school).</p></sec><sec id="Sec14"><title>Parental feasibility</title><p>Parental opinions (n&#x02009;=&#x02009;216) concerning the feasibility of drop-off spots prior to the intervention are presented in Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>. Of all parents, 89.7% agreed that there would be more social contact between children and 68.3% agreed that they would save time when a drop-off spot would be organized. Indicated barriers for using a drop-off spot were lack of time (34.7%) and weather conditions (58.3%). Of all parents, 76.4% agreed that providing adult supervision at the drop-off spot is needed and 95.8% expressed the need for adult supervision during the route to school. Of those 95.8%, 50.0% agreed that adult supervision was only needed for the youngest children (6&#x02013;9 years). Regarding the environment of a drop-off spot, the majority of all parents (75.4%) agreed that only a kiss and ride zone should be provided instead of parking space. Additionally, 95.2% of all parents agreed that the drop-off spot should be separated from the road and not on the sidewalk just next to the road. About 90.8% of all parents agreed that the drop-off spot should be on the route to their work. Parental concerns regarding supervision and location of the drop-off spot were included in the development of the intervention.</p></sec><sec id="Sec15"><title>Intervention effects</title><p>Intervention effects on total step counts per day, step counts per day before and after school hours and weekly number of walking trips to and from school are described in Table&#x000a0;<xref rid="Tab4" ref-type="table">4</xref>. Positive significant intervention effects were found for step counts per day before and after school hours (+732 step counts/day; X²&#x02009;=&#x02009;12.2; p&#x02009;&#x0003c;&#x02009;0.001) and number of walking trips to and from school (+2 trips/week; X²&#x02009;=&#x02009;52.9; p&#x02009;&#x0003c;&#x02009;0.001). No significant intervention effect was found for total step counts per day (X²&#x02009;=&#x02009;2.0; p&#x02009;=&#x02009;0.16).<table-wrap id="Tab4"><label>Table 4</label><caption><p> <bold>Intervention effects on children&#x02019;s step counts and number of walking trips to and from</bold> ^{<bold>1</bold>} <bold>school (n&#x02009;=&#x02009;58)</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th/><th> <bold>Mean total step counts per day (SD)</bold> ^{<bold>a</bold>} </th><th> <bold>Mean step counts per day before and after</bold> ^{<bold>1</bold>} <bold>school hours (SD)</bold> ^{<bold>a</bold>} </th><th> <bold>Mean walking trips per week</bold> _{<bold>+</bold>} <bold>to and from</bold> ^{<bold>1</bold>} <bold>school (SD)</bold> ^{<bold>b</bold>} </th></tr></thead><tbody><tr valign="top"><td>Baseline</td><td>12168 (3269)</td><td>1711 (961)</td><td>1 (2)</td></tr><tr valign="top"><td>Intervention</td><td>11261 (3252)</td><td>2443 (1074)</td><td>3 (2)</td></tr><tr valign="top"><td>&#x003a7;² </td><td>2.0</td><td>12.2***</td><td>52.9***</td></tr></tbody></table><table-wrap-foot><p>***p&#x02009;&#x0003c;&#x02009;0.001; SD = standard deviation.</p><p> ^aanalyses were controlled for: sex, age, socio-economic status, school and pedometer wear time.</p><p> ^banalyses were controlled for: sex, age, socio-economic status and school.</p><p> ¹not for children from the suburban school (suburban school only organized drop-off spot before school hours).</p><p> ₊excluding Wednesday.</p><p>&#x003a7;²&#x02009;=&#x02009;chi square.</p></table-wrap-foot></table-wrap></p></sec><sec id="Sec16"><title>Perception of the intervention (process evaluation)</title><p>Descriptive information of the questionnaire on how school principals (n&#x02009;=&#x02009;2), teachers (n&#x02009;=&#x02009;9) and parents (n&#x02009;=&#x02009;44) perceived the intervention is shown in Table&#x000a0;<xref rid="Tab2" ref-type="table">2</xref>.</p><p>All teachers (n&#x02009;=&#x02009;9) and 35 parents (92.1%) agreed that the drop-off spot was well organized. Both school principals and the majority of the parents (n&#x02009;=&#x02009;31; 81.6%) agreed that drop-off spots could be used in the future, while only two teachers agreed. Concerning opportunities and future possibilities for the intervention, both school principals, seven teachers and 38 parents (88.7%) agreed to organize drop-off spots for all ages. Most parents (n&#x02009;=&#x02009;30; 71.4%) agreed that adult supervision during the route to school is only needed for the youngest children (6&#x02013;9 years). Additionally, 20 parents (46.5%) suggested organizing drop-off spots every day. Most teachers were not willing to organize drop-off spots (n&#x02009;=&#x02009;3) or suggested to organize drop-off spots only one or two times per week (n&#x02009;=&#x02009;3). One school principal, four teachers and 19 parents (44.2%) agreed to organize drop-off spots only before school hours. The other principal would organize drop-off spots before and after school hours. Three teachers preferred to organize drop-off spots only during theme-related periods at school (e.g. the week of active mobility), one school principal and three teachers agreed to organize drop-off spots during the entire school year. However, one school principal and 21 parents (48.8%) prefer this only in spring or summer.</p><p>During the intervention, two teachers reported organizational limitations (e.g. willingness of volunteers to supervise) and two teachers reported that the environment in the surroundings of the drop-off spot was a limitation to organize a drop-off spot (e.g. busy traffic and traffic congestions). Moreover, two teachers expressed that the intervention was an additional load for teachers. Only two parents reported busy traffic on the way to the drop-off spot (6.8%) and nine parents (20.5%) reported the time period(s) of the organized drop-off spot as an experienced difficulty.</p></sec></sec><sec id="Sec17" sec-type="discussion"><title>Discussion</title><p>The present study provided evidence that implementing drop-off spots is feasible, effective and is positively perceived by school principals and parents to promote children&#x02019;s walking to school. However, teachers expressed doubts regarding future implementation.</p><p>Prior to the intervention, both schools mainly indicated organizational issues (e.g. time, location,&#x02026;) regarding the implementation of drop-off spots, while parents were mainly concerned about safety issues. A requirement for parents to make use of the drop-off spot was the provision of adult supervision at the drop-off spot and during the walk to school. This was not surprising as previous studies investigating determinants of active commuting to school identified parental safety concerns (road safety and perceived danger from strangers) as main barriers for children&#x02019;s active commuting to school [<xref ref-type="bibr" rid="CR10">10</xref>,<xref ref-type="bibr" rid="CR11">11</xref>].</p><p>Overall, we found that implementing drop-off spots in the proximity of primary schools was feasible, but that attention is required to several factors to enhance parental and teacher involvement and to ensure safety. These factors were comparable with the feasibility issues in walking school bus programs (e.g. willingness of volunteers, supervision, social benefits, time-savings) [<xref ref-type="bibr" rid="CR12">12</xref>,<xref ref-type="bibr" rid="CR19">19</xref>]. It is important to develop the intervention in close consultation with the schools, but some aspects of the intervention can be generalized across schools. Based on our findings, some general recommendations could be made to organize drop-off spots in the future. First, providing adult supervision is necessary in young children, but to stimulate children&#x02019;s independent mobility [<xref ref-type="bibr" rid="CR2">2</xref>,<xref ref-type="bibr" rid="CR30">30</xref>], older primary schoolchildren (11&#x02013;12 years) can walk independently to school. Secondly, the drop-off spot should be situated nearby approach roads as the majority of the parents indicated that they would use the drop-off spot if it is located on the route to their work. Third, the majority of the parents agreed that only a kiss and ride zone should be provided to drop off the children. Consequently, a drop-off spot does not necessarily have to be organized at a location with parking space, however, a zone which allows &#x0201c;kiss and ride&#x0201d; should be selected. This makes it easier for parents to drop off their children. With attention to safety, it is recommended that the drop-off spots are separated from the road (and not located on the sidewalk just next to the road). Cul-de-sacs, squares and playgrounds can be suitable locations. At these locations, children can play before they walk to school, which is beneficial for their daily physical activity levels. The feasibility study provided school-specific information to organize the drop-offs. Because a different approach is required for every school, it is important to check school and parental opinions before implementing drop-off spots and to take those school-specific opinions into account when developing the intervention.</p><p>Small but positive significant intervention effects were found for parameters regarding walking to school (steps before and after school hours; number of walking trips to and from school), demonstrating that drop-off spots are effective to promote walking to school among primary schoolchildren. The positive significant effects demonstrated that children who cannot commute actively to school (e.g. due to large home-school distance), can commute actively to school if drop-off spots are implemented in the proximity of the school. An explanation for the small effects could be the fact that the drop-off spots were not far enough from the school to induce large effects. Previous walking school bus programs reported higher increases of children&#x02019;s walking [<xref ref-type="bibr" rid="CR12">12</xref>,<xref ref-type="bibr" rid="CR31">31</xref>]. However, in those programs larger distances were traveled (ranging from 0.5&#x000a0;km to 2.5&#x000a0;km) when children walked to school. Furthermore, D&#x02019;Haese et al. [<xref ref-type="bibr" rid="CR9">9</xref>] reported criterion distances for Flemish 11-to-12-year olds of 1.5&#x000a0;km for walking to school. Nevertheless, the drop-off spots in the present study were located at less than 800&#x000a0;m from the school in order to reach young children as well. Additionally, the location of the drop-off spot and the distance from school were chosen in cooperation with the schools and both schools did not find it feasible to increase the distance (&#x0003e;800&#x000a0;m from the school). So, increasing the distance from the drop-off spot to school may be desirable from a health promotion perspective, however, the feasibility of more distant drop-off spots remains to be demonstrated. Another explanation for the small effects could be that the days when children did not use the drop-off spots were also included in the analyses. Subsequently, the findings showed that children did not use the drop-off spot on a daily basis during the intervention week. Moreover, the intervention period lasted only one school week. Parents and children could not make a habit of their behavior. When the intervention could be implemented over a longer period and parents and children would use the drop-off spot more frequently (e.g. twice on a daily basis), effects may be larger. It is important to mention that the intervention effects were reported for a group of mainly low SES children (51.7%). It has been demonstrated that children with lower educated parents are at increased risk of negative health behaviors and outcomes [<xref ref-type="bibr" rid="CR32">32</xref>] and that low SES parents are less likely to be reached and to participate in health promotion programs [<xref ref-type="bibr" rid="CR33">33</xref>]. Consequently this intervention could be a promising strategy to promote walking in this at-risk group.</p><p>Additionally, we found that the intervention did not contribute to children&#x02019;s total daily step counts. Possibly, the distance from the drop-off spots to the school was not large enough to contribute significantly to children&#x02019;s daily step counts. Another explanation for this finding could be that children engaged in compensation behavior during the intervention. Children may have been less active during the rest of the day (e.g. less active playing during recess before the school starts) as they already walked before or after school hours due to the implementation of drop-off spots [<xref ref-type="bibr" rid="CR34">34</xref>]. Moreover, children might have engaged in less after-school sports activities during the intervention period compared to the baseline measurements. In June (during the intervention period), community-based sports sessions in Flanders (for ball sports, gymnastics, dance,&#x02026;) end (summer pause). Conversely, the baseline measurements (end April 2013- early May 2013) occurred before the sports seasons ended. This could partly explain the absence of an intervention effect on children&#x02019;s daily step counts.</p><p>After the intervention, parents again reported the need to provide adult supervision during the route to school for the younger children. In general, the intervention was positively perceived by both school principals and parents. Nevertheless, the teachers expressed doubts regarding future implementation. Also the low response rate of the teachers (9 of 25 teachers filled out the questionnaire on how the intervention was perceived) demonstrates that teachers possibly experienced the intervention as an additional workload. Therefore, a possible solution could be to involve other volunteers (e.g. parents, grandparents, &#x02026;) to organize and supervise the drop-off spots. This has been demonstrated to be a feasible strategy in previous walking school bus programs [<xref ref-type="bibr" rid="CR12">12</xref>,<xref ref-type="bibr" rid="CR16">16</xref>]. It is also of interest to note that for the organization of drop-off spots, less supervising adults are needed compared to walking school bus programs (e.g. multiple supervised routes to school), which can increase feasibility. Furthermore, our findings show that it is important to motivate teachers in order for them to be willing to include this task in their job responsibilities. However, when implementing more drop-off spots, extra volunteers and motivated teachers are needed.</p><p>Overall, the intervention aimed to increase walking to school focusing on those children living further away from school and who are usually driven to school by their parents. A major advantage of the intervention is its flexibility, as every school can implement drop-off spots that are specifically tailored to the school&#x02019;s needs. When developing the intervention, the needs of the different schools and the parental opinions were taken into account, in order to create a real-life and most appropriate intervention for every school. Nevertheless, small adaptations to the intervention regarding organization (e.g. volunteers, distance) are desirable, depending on the school and its environmental context. Additionally, the intervention is free of cost and requires no large efforts from the school and the parents. Furthermore, implementing drop-off spots can be useful as part of a larger intervention to promote active transport to school in primary schools: drop-off spots can be easily implemented and commuting actively to school can become a daily habit. By implementing drop-off spots into a larger intervention (e.g. Walking School Bus, Safe Routes to School), also children not living within a feasible distance from school are reached.</p><p>The present study has some important limitations. First, it was a pilot study aiming to examine the feasibility and effectiveness of drop-off spots before implementing it into a larger-scale study. Therefore, the study involved only two schools with a small sample size, which limits power and generalizability. Also selection bias (self-selection of schools and participants; e.g. participation of most motivated parents) and the specific environment around the included schools limit generalizability. Secondly, the intervention period was rather short and no long-term effects were studied. Further research should focus on the long-term feasibility and effects of this intervention in a wider variety of primary schools. Third, the self-selection of parents to use the intervention or not could have influenced the results. In the current study, parents were free to decide whether they used the intervention or not, in contrast to other interventions at school where children do not have the choice to participate (e.g. playground interventions). Fourth, other influences on travel behavior (e.g. home location, household composition) might have influenced the results. However, it is assumed that these influences were limited since a within-subject design was used to determine the intervention effects. This study has important strengths. To our knowledge, this is the first study that implemented drop-off spots to increase children&#x02019;s walking to school, specifically for those children who cannot commute actively to school because of a large home-school distance. Additionally, this is the first study that investigated both the feasibility and effectiveness of the implemented drop-off spots, and added information on how the intervention was perceived by the school and the parents. Other strengths of this study were the within-subject design, which induces high external validity, and the relatively high proportion of low SES children involved in the study. Furthermore, the use of the Omron Walking Style Pro allowed to assess steps during the entire school day and steps before and after school hours.</p></sec><sec id="Sec18" sec-type="conclusion"><title>Conclusions</title><p>The present pilot study showed that implementing drop-off spots might be a promising strategy to increase children&#x02019;s walking to and from school and might provide an alternative for primary schoolchildren who cannot commute actively to school because of a large home-school distance. Implementing drop-off spots does not require major efforts from the schools and schools can choose how and when they organize and implement drop-off spots. Because teachers were less convinced and expressed doubts regarding future implementation, motivating teachers and involving other volunteers in the intervention may be desirable. Implementing drop-off spots may be useful as part of a larger intervention to promote active transport to school in primary schools.</p></sec>

Comparison of trace element concentration in bone and intervertebral disc tissue by atomic absorption spectrometry techniques

Could not extract abstract

<contrib contrib-type="author" corresp="yes"><name><surname>Kubaszewski</surname><given-names>&#x00141;ukasz</given-names></name><address><email>pismiennictwo1@gmail.com</email></address><xref ref-type="aff" rid="Aff1"/><xref ref-type="aff" rid="Aff4"/></contrib><contrib contrib-type="author"><name><surname>Zio&#x00142;a-Frankowska</surname><given-names>Anetta</given-names></name><address><email>ziola.a@gmail.com</email></address><xref ref-type="aff" rid="Aff2"/></contrib><contrib contrib-type="author"><name><surname>Frankowski</surname><given-names>Marcin</given-names></name><address><email>marcin.frankowski@amu.edu.pl</email></address><xref ref-type="aff" rid="Aff2"/></contrib><contrib contrib-type="author"><name><surname>Rogala</surname><given-names>Piotr</given-names></name><address><email>gabinet.rogala@gmail.com</email></address><xref ref-type="aff" rid="Aff3"/></contrib><contrib contrib-type="author"><name><surname>Gasik</surname><given-names>Zuzanna</given-names></name><address><email>zuzanna.gasik@poczta.onet.pl</email></address><xref ref-type="aff" rid="Aff4"/></contrib><contrib contrib-type="author"><name><surname>Kaczmarczyk</surname><given-names>Jacek</given-names></name><address><email>dr.kaczmarczyk@gmail.com</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author"><name><surname>Nowakowski</surname><given-names>Andrzej</given-names></name><address><email>BeataDeckert@interia.pl</email></address><xref ref-type="aff" rid="Aff3"/></contrib><contrib contrib-type="author"><name><surname>Dabrowski</surname><given-names>Mikolaj</given-names></name><address><email>md@twt.net.pl</email></address><xref ref-type="aff" rid="Aff3"/></contrib><contrib contrib-type="author"><name><surname>Labedz</surname><given-names>Wojciech</given-names></name><address><email>wlabedz@poczta.onet.pl</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author"><name><surname>Mi&#x00119;kisiak</surname><given-names>Grzegorz</given-names></name><address><email>gmiekisiak@gmail.com</email></address><xref ref-type="aff" rid="Aff5"/></contrib><contrib contrib-type="author"><name><surname>Gasik</surname><given-names>Robert</given-names></name><address><email>robert.gasik@ir.ids.pl</email></address><xref ref-type="aff" rid="Aff4"/></contrib><aff id="Aff1"><label/>Department of Orthopedics and Traumatology, W. Dega University Hospital, University of Medical Science Poznan, 28 Czerwca 1956r St., Pozna&#x00144;, 61-545 Poland </aff><aff id="Aff2"><label/>Department of Water and Soil Analysis, Faculty of Chemistry, Adam Mickiewicz University in Pozna&#x00144;, Umultowska 89b, Pozna&#x00144;, 61-614 Poland </aff><aff id="Aff3"><label/>Department of Spine Surgery, Oncologic Orthopaedics and Traumatology, W. Dega University Hospital, University of Medical Science Poznan, 28 Czerwca 1956r St., Pozna&#x00144;, 61-545 Poland </aff><aff id="Aff4"><label/>Clinic and Polyclinic of Neuroorthopedic and Neurology, Institute of Rheumatology, Warsaw, Sparta&#x00144;ska 1, Warsaw, 02-637 Poland </aff><aff id="Aff5"><label/>Department of Neurosurgery, Specialist Medical Center, Polanica-Zdroj, Poland </aff>

Journal of Orthopaedic Surgery and Research

<sec id="Sec1" sec-type="introduction"><title>Background</title><p>Trace element (TE) analysis in biota has a twofold purpose of monitoring environmental exposure to the pollutants and, secondarily, of indirect analysis of metabolism-related issues. The human body can be divided into several compartments with respect to element turnover. The most unstable compartments are serum, urine, and cerebrospinal fluid. Analysis of these tissues is useful for short periods of exposure or organismal response to specific metabolic impulses. In contrast is bone tissue, which is regarded as a TE repository that reflects turnover in the whole organism [<xref ref-type="bibr" rid="CR1">1</xref>]. The biological half life of heavy metals in bones may be up to 30 years and reflect up to 90% of the whole body content [<xref ref-type="bibr" rid="CR2">2</xref>], possibly because of the solid structure of the bone tissue along with optimal blood perfusion.</p><p>A distinct compartment of the human body is the intervertebral disc. Transformation during maturation causes loss of vascularity within the first decade of life [<xref ref-type="bibr" rid="CR3">3</xref>] and is followed by metabolism transformation to withstand low oxygen concentration and a highly acidic environment because of waste product concentration. IVD morphology is characterized by a dominating extracellular matrix seeded with a low density of chondrocyte-like cell clusters responsible for production and control of matrix turnover. Domination of the extracellular matrix is similar in bone and IVD tissue. Dissimilarities are related to blood perfusion that reflects metabolism. Such separateness of the IVD tissue makes it of interest for TE concentration analysis.</p><p>In this study, we evaluate the trace elements concentration in intervertebral disc tissue and femoral bone in patient with degenerative changes. There is high disproportion in a number of studies of two examined tissues. Only few papers present the concentration and accumulation potential for selected TEs in IVD [<xref ref-type="bibr" rid="CR4">4</xref>,<xref ref-type="bibr" rid="CR5">5</xref>], no comparison analysis between IVD and bone tissue has been presented.</p><p>The advantages of the analysis were comparison of the tissues that substantially differ in metabolism, blood perfusion, and separateness form adjoining tissues and organism. Nevertheless, tissues are similar, for biomechanical function and morphology with dominating extracellular matrix. As the pilot study, we have decided to choose the example elements from three groups: essential, potentially essential, and toxic to recognize the distribution differences in both tissues. Such analysis not only sheds a new light on the metabolism especially of the intervertebral disc but also evaluates the accumulation potential of the IVD in respect to the bone tissue. Addition value of the study was performing the analysis with the same methodology by the same laboratory.</p><p>Among the available analytical techniques, the GF-AAS analytical technique is better for determination of elements in biological samples because of, e.g., only few spectral interferences, good limits of detection, small sample volume, and low analysis costs. Besides for determination of single structural elements, (e.g., Mg and Zn) flame atomic absorption spectrometry is better because of the low cost. Also the GF-AAS technique is the most commonly used technique for the analysis of mineral and trace elements in biological samples, e.g., Pb in bones samples [<xref ref-type="bibr" rid="CR6">6</xref>]; Cr, Cd, Mn, Ni, and Pb in whole blood, urine, saliva, and axillary hair [<xref ref-type="bibr" rid="CR7">7</xref>]; and Al in bones [<xref ref-type="bibr" rid="CR8">8</xref>].</p><p>Another analytical techniques used in analysis of metals in biological samples are, e.g., X-ray fluorescence (Pb and Sr in bones) [<xref ref-type="bibr" rid="CR9">9</xref>,<xref ref-type="bibr" rid="CR10">10</xref>]; Prompt gamma neutron activation (Cd in liver and kidney) neutron activation analysis (Al in bones) [<xref ref-type="bibr" rid="CR11">11</xref>], ICP-AES (Cu, Co, Cr, Y, Yb, and Bi in biological samples) [<xref ref-type="bibr" rid="CR12">12</xref>]; ICP-MS (trace elements in human urine) [<xref ref-type="bibr" rid="CR13">13</xref>]; and HPLC-ICP-AES (Cu, Cd, and Zn in human liver) [<xref ref-type="bibr" rid="CR14">14</xref>].</p><p>The aim of the study was evaluation the differences of the trace elements concentration in intervertebral disc tissue and femoral bone in patient with degenerative changes with graphite furnace atomic absorption spectrometry (GF-AAS) technique.</p></sec><sec id="Sec2" sec-type="materials|methods"><title>Methods</title><p>This analysis involved two groups: patients with degenerative disc disease (DDD) and patients with idiopathic osteoarthritis of the hip joint (OA).</p><sec id="Sec3"><title>DDD group</title><p>Intervertebral discs from 22 patients were obtained during a surgical procedure. Twelve specimens (6 patients) were from the cervical spine and 18 (16 patients) were taken from the lumbar spine. The indication for the operation was degenerative disc disease with neurological changes in clinical examination: local neck or back pain with radicular symptoms. In the cervical spine, the discectomy was performed with an anterior approach with the removal of the intervertebral disc tissue followed with interbody fusion. In the lumbar spine, the intervertebral disc was approached from posterior. After tissue removal in seven cases, the interbody fusion with transpedicular stabilization was performed to restore the stability of the motion segment. During the acquisition of the biological material, only the intervertebral disc was taken to the analysis without the parts of the vertebral end-plate removed in the preparation process for interbody fusion. The samples were frozen in &#x02212;20&#x000b0;C.</p></sec><sec id="Sec4"><title>OA group</title><p>This study included the material of 26 femoral bone fragments from 26 patients, acquired during total hip arthroplasty. During the surgical procedure, the proximal part of the femur was resected with a motor saw. After resection, the sample was cleaned from adjoining soft tissues, i.e., joint capsule or muscles and frozen in &#x02212;20&#x000b0;C. The indication for the procedure was idiopathic osteoarthrosis of the hip joint. A separate analysis was performed for two anatomic regions of each type of resected fragment: the femoral head and femoral neck.</p></sec><sec id="Sec5"><title>Patient data and sample analysis</title><p>All patients were interviewed using a questionnaire to collect data on demography, health status, and occupational heavy metal exposure. In the interview, no patients had knowledge of inadvertent exposure to heavy metal pollution.</p><p>In all samples, the levels of Pb, Ni, Mo, Cu, Mg, and Zn were evaluated. In OA samples, TEs were evaluated separately for the femoral neck and head.</p><p>The frozen intervertebral disc samples were freeze-dried using a Lyovac lyophilizer GT2e (Steris, Germany) for 24 h (drying pressure of <italic>p</italic> = 6.5&#x02009;&#x000d7;&#x02009;10^&#x02212;1 mbar, ambient temperature under vacuum&#x02014;approximately &#x02212;55&#x000b0;C). After drying of the IVD samples, the femoral head and femoral neck samples were weighed, and nitric acid (Suprapur, Merck, Germany) was added to obtain a dilution factor of 10 (range of sample weight: 0.111&#x02013;0.489 g dry weight [dw]). The prepared samples were allowed to stand overnight to slow mineralization. Samples then were mineralized in a microwave oven (Mars Xpress 5, USA) based on modified 3051 EPA method [<xref ref-type="bibr" rid="CR15">15</xref>]. TE concentrations were calculated for the dry weight of the disc.</p><p>The TE concentrations were determined in three replications using an atomic absorption spectrometer (AA 7000, Shimadzu, Japan) with graphite furnace atomization (GF-AAS). The percent relative standard deviation (%RSD) for the GF-AAS analytical technique did not exceed 5%. The concentrations of Mg and Zn were determined in three replications using the AA 7000 (Shimadzu, Japan) with flame atomization (F-AAS). The %RSD for the F-AAS analytical technique did not exceed 7%.</p><p>In the analysis, the following values for the traits were derived: mean, median, maximum and minimum, and standard deviation (SD).</p><p>The Shapiro-Wilk test was used to confirm the normal distribution of the analyzed parameters. In case of normal distribution, Student&#x02019;s <italic>t</italic>-test for independent samples was used, if sample values were non-normally distributed, nonparametric statistical test was used (Mann-Whitney test). To evaluate the difference between groups DDD and OA in respect to sex, the Chi² test (<italic>p</italic> =1,0) was applied. Spearman rank correlation test was used for age correlation with TE concentration. <italic>p</italic> values &#x0003c;0.05 were considered statistically significant.</p></sec><sec id="Sec6"><title>Ethics statement</title><p>Ethical considerations were in agreement with the Helsinki Declaration. In all cases, patients were informed about the aim of study and gave written consent for participation in the study and for data publication. The use of tissue in the investigations was approved by the Bioethics Committee of the Institute of Rheumatology, Warsaw, on May 31, 2012, and Bioethics Committee of University of Medical Sciences, Poznan, reference nos. 406/13 and 172/14.</p></sec></sec><sec id="Sec7" sec-type="results"><title>Results and discussion</title><p>The TE values for both groups are presented in Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>. Among patients with DDD, the average age was 47.6 years (range 28&#x02013;64; SD 8.8). Among patients with OA, the average age was 57.8 (range 30&#x02013;64; SD 7.2). Among DDD patients, males constituted the majority (54.5%) of the group; in OA, the majority also were males (61.5%).<table-wrap id="Tab1"><label>Table 1</label><caption><p> <bold>Characteristics of trace element concentrations in the analyzed groups</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th/><th colspan="6"> <bold>Element</bold> </th></tr><tr valign="top"><th/><th> <bold>Pb</bold> </th><th> <bold>Ni</bold> </th><th> <bold>Mo</bold> </th><th> <bold>Cu</bold> </th><th> <bold>Mg</bold> </th><th> <bold>Zn</bold> </th></tr><tr valign="top"><th/><th> <bold>&#x003bc;g kg</bold> ^{<bold>&#x02212;1</bold>} <bold>dw</bold> </th><th> <bold>&#x003bc;g kg</bold> ^{<bold>&#x02212;1</bold>} <bold>dw</bold> </th><th> <bold>&#x003bc;g kg</bold> ^{<bold>&#x02212;1</bold>} <bold>dw</bold> </th><th> <bold>mg kg</bold> ^{<bold>&#x02212;1</bold>} <bold>dw</bold> </th><th> <bold>mg kg</bold> ^{<bold>&#x02212;1</bold>} <bold>dw</bold> </th><th> <bold>mg kg</bold> ^{<bold>&#x02212;1</bold>} <bold>dw</bold> </th></tr></thead><tbody><tr valign="top"><td>DDD</td><td colspan="6"/></tr><tr valign="top"><td>Min&#x02013;max</td><td>61.47&#x02013;2,233</td><td>25.48&#x02013;444.2</td><td>20.02&#x02013;143.2</td><td>0.971&#x02013;6.091</td><td>182.6&#x02013;2,132</td><td>10.56&#x02013;184.5</td></tr><tr valign="top"><td>Mean</td><td>686.7</td><td>215.4</td><td>54.33</td><td>2.71</td><td>800.1</td><td>39.61</td></tr><tr valign="top"><td>Median</td><td>510.9</td><td>169.9</td><td>47.75</td><td>2.504</td><td>624.5</td><td>31.99</td></tr><tr valign="top"><td>SD</td><td>565.6</td><td>132.1</td><td>27.53</td><td>1.352</td><td>525.5</td><td>35.95</td></tr><tr valign="top"><td>OA&#x02014;femoral neck</td><td colspan="6"/></tr><tr valign="top"><td>Min&#x02013;max</td><td>1,266&#x02013;5,090</td><td>294.0&#x02013;3,602</td><td>360.0&#x02013;1,900</td><td>0.550&#x02013;2.660</td><td>1,163&#x02013;2,243</td><td>52.67&#x02013;108.2</td></tr><tr valign="top"><td>Mean</td><td>2,898*</td><td>1,511*</td><td>1,244*</td><td>1.210^&#x02021; </td><td>1,661*^&#x02020; </td><td>72.23</td></tr><tr valign="top"><td>Median</td><td>3,227</td><td>1,139</td><td>1,425</td><td>1.117</td><td>1,617</td><td>67.71</td></tr><tr valign="top"><td>SD</td><td>1,359</td><td>1,290</td><td>502.2</td><td>0.540</td><td>318.9</td><td>15.19</td></tr><tr valign="top"><td>OA&#x02014;femoral head</td><td colspan="6"/></tr><tr valign="top"><td>Min&#x02013;max</td><td>1,436&#x02013;6,279</td><td>194.2&#x02013;6,438</td><td>1,105&#x02013;1,663</td><td>0.550&#x02013;2.550</td><td>900.4&#x02013;2,529</td><td>53.51&#x02013;96.89</td></tr><tr valign="top"><td>Mean</td><td>3,403*</td><td>1,467</td><td>1,376*</td><td>1.630*^&#x02021; </td><td>1,458*^&#x02020; </td><td>74.40*</td></tr><tr valign="top"><td>Median</td><td>3,416</td><td>838</td><td>1,338</td><td>1.760</td><td>1,413</td><td>73.50</td></tr><tr valign="top"><td>SD</td><td>1,540</td><td>1,727</td><td>213.5</td><td>0.540</td><td>384.3</td><td>13.32</td></tr></tbody></table><table-wrap-foot><p>*Normal distribution, ^&#x02021;statistically significant (Mann-Whitney test), ^&#x02020;statistically significant (Student&#x02019;s <italic>t</italic>-test).</p></table-wrap-foot></table-wrap></p><p>There were no statistically significant differences between two groups in respect to sex. Although the age ranges were similar in both groups, there was no ground to support the null hypothesis (<italic>p</italic> &#x0003e;0.05).</p><p>In the OA group, we confirmed the statistically significant higher concentration of the Ni in femoral neck in males. In femoral head group concentration of Zn, Cu, Ni, and Pb was higher in males, but the differences were not statistically significant. In the DDD group, we have found positive significant correlation of Pb concentration with age. No differences between sexes have been observed in respect to TE concentrations.</p><p>Stage of osteoarthritis in OA samples was moderate or severe. The cartilage of femoral head showed partial or total damage in the area of 50%&#x02013;90% of the cartilage.</p><p>In the DDD group, TE concentrations were within detection limits in the vast majority of samples. Only in the case of Mo (five samples) and Ni (one sample) were concentrations below the level of detection (LOD). The percentages of undetectable TE were 16% and 3%, respectively.</p><p>In bone samples acquired from the femoral neck, Pb was undetected in 17 samples (68%), Ni in 19 (76%), Mo in 13 (52%), and Cu in 12 (48%). In bone samples acquired from the femoral head, Pb was undetected in 16 samples (64%), Ni in 12 (48%), Mo in 20 (80%), and Cu in 10 (40%). In the statistical analysis, those measurements were excluded.</p><p>The mean TE values were higher in bone (from twofold to almost 26-fold) in all cases except for Cu, which was up to 1.9 times higher (2.71 mg kg^&#x02212;1) in the disc compared to the femoral neck and head (1.21 and 1.63 mg kg^&#x02212;1, respectively). The minimal level of Cu detected in bone tissue was lower compared to IVD (0.55 vs 0.97 mg kg^&#x02212;1, respectively). In the DDD group, the maximum value of Cu significantly exceeded levels observed in both femoral neck and head. In one IVD sample, Cu levels reached 23.64 mg kg^&#x02212;1, which was excluded from the statistical analysis as an outlier. After exclusion of this measurement from the analysis, the maximum values for Cu in IVD and the femoral neck and head were 6.09, 2.56, and 2.55 mg kg^&#x02212;1, respectively. The SD for the disc tissue was more than double that of bone (1.35 vs 0.54 mg kg^&#x02212;1, respectively).</p><p>Our groups were derived from two studies, so sample selection targeted obtaining comparable groups in respect to age and sex. Generating groups that are matched for age can be a special challenge, particularly with a large age variance in the analyzed pathologies, as in this case. Furthermore, obviously, the disc and femoral bone cannot be collected from the same individuals. In our opinion, the selection of groups for this analysis was as optimal as possible given these limitations.</p><p>In some of our samples, the analyzed elements were below the LOD. We have not encountered such a situation in the literature evaluating TE in human tissue. In chemometric analysis, extensively used in spectrophotometric studies, samples for which the concentration is below the LOD are usually substituted with 0.5&#x000d7; LOD values [<xref ref-type="bibr" rid="CR16">16</xref>]. However, we believe that presenting the data without this kind of substitution does not cause distortion and better reflects the clinical reality.</p><p>In both groups, there were samples with Mo and Ni concentrations below LOD. Otherwise, Mo and Ni were not detected in some samples of bone and IVD; in the latter case, the percentages were significantly lower than in bone. These elements are considered essential (Mo) or possibly essential (Ni) for humans [<xref ref-type="bibr" rid="CR17">17</xref>].</p><p>Molybdenum is responsible for stabilizing oxidized nitrogen and occurs in organic molecules such as amino acids and carbohydrates in the form of MoO₄^2&#x02212; [<xref ref-type="bibr" rid="CR18">18</xref>]. It is involved in xanthine oxidase synthesis, which is directly related to Mo content and affects protein synthesis and metabolism of purines and fats [<xref ref-type="bibr" rid="CR17">17</xref>]. There are studies indicating the inverse relation of the Mo and Mg concentration that may be organ specific [<xref ref-type="bibr" rid="CR19">19</xref>]. The element occurs in all human tissues in a range from 0.001 to 0.4 mg kg^&#x02212;1, with the lowest values for the blood and the highest for the kidneys and liver [<xref ref-type="bibr" rid="CR17">17</xref>]. The average content in soft tissues (of the &#x0201c;reference man&#x0201d;) is &#x0003c;0.075 mg kg^&#x02212;1 and in skeleton &#x0003c;0.48 mg kg^&#x02212;1 [<xref ref-type="bibr" rid="CR20">20</xref>]. Our study shows higher values of Mo in the bone, ranging from 0.360 to 1.9 mg kg^&#x02212;1. In the IVD, Mo concentration was approximately 25.8 times lower compared to bone and was in the &#x0201c;reference man&#x0201d; ranges for soft tissue.</p><p>In mammals, Ni is incorporated into superoxide dismutase, which uses divalent Zn [<xref ref-type="bibr" rid="CR21">21</xref>]. Ni deficiency can induce some dysfunction in fat metabolism, but except for toxicity data, findings describing a potential metabolic role of the element are insufficient [<xref ref-type="bibr" rid="CR17">17</xref>]. Studies indicate Ni supplementation to be bound with improved bone strength in birds [<xref ref-type="bibr" rid="CR22">22</xref>]. On our study, Ni concentration in femoral neck was found to be related with male sex, where osteoporotic changes are observed later than in females. Brodziak-Dopierala et al. [<xref ref-type="bibr" rid="CR23">23</xref>] estimated the Ni concentration in the femoral head at a medium level of 4.82 mg kg^-1 (SD 10.74 mg kg^&#x02212;1) and the cartilage at 4.40 mg kg^&#x02212;1 (SD 7.38 mg kg^&#x02212;1). The average Ni content in human soft tissues is estimated at 0.088 mg kg^&#x02212;1 [<xref ref-type="bibr" rid="CR17">17</xref>]. Its concentrations in soft human organs vary greatly, with the highest mean values for the lung (0.173 mg kg^&#x02212;1) and the lowest for the pancreas (0.034 mg kg^&#x02212;1) [<xref ref-type="bibr" rid="CR17">17</xref>]. Our study showed a bone concentration of Ni to be lower than in Brodziak-Dopierala et al. (mean 1.46&#x02013;1.51 vs 4.82 mg kg^&#x02212;1, respectively). Ni concentration in the IVD was still higher than reference values for soft tissues presented in the literature [<xref ref-type="bibr" rid="CR23">23</xref>].</p><p>In the OA group, in up to 68% of the samples, the concentration of Pb was below the LOD. It is interesting that Pb levels below LOD were observed in bone tissue even as the mean value for the remainder of the samples was more than four times that of the disc tissue. Also only in the DDD group, we confirmed the significant correlation of the element with the age. Pb level is purely related to environmental pollution with no identified metabolic role. Additionally, it is considered highly toxic and is present in every human tissue in the range of &#x0003c;0.2 to 4.8 mg kg^&#x02212;1 [<xref ref-type="bibr" rid="CR17">17</xref>]. The source can be food, water, or air. As an element, it is classified as a poor metal and a member of the carbon group. Pb influences heme synthesis by inhibiting porphobilinogen synthesis and ferrochelatase. It can lead to anemia by preventing formation of porphobilinogen and iron incorporation [<xref ref-type="bibr" rid="CR24">24</xref>]. In neural tissue, Pb may be substituted as a calcium analog, interfering with ion channels during impulse conduction [<xref ref-type="bibr" rid="CR17">17</xref>]. The lead exposure increases the risk of hip fracture in both males and females [<xref ref-type="bibr" rid="CR25">25</xref>]. No data appear to be available regarding Pb function in the connective tissue that is characteristic for IVDs or Pb concentration in this tissue. Brodziak-Dopierala et al. estimated Pb concentration to be 3.75 mg kg^&#x02212;1 (SD 5.86 mg kg^&#x02212;1) [<xref ref-type="bibr" rid="CR23">23</xref>]. Loska et al. identified lower levels with a mean of 1.35 mg kg^&#x02212;1 and SD of 0.68 mg kg^&#x02212;1 [<xref ref-type="bibr" rid="CR26">26</xref>]. In our analysis, Pb bone levels were closer to one, in agreement with Brodziak-Dopierala et al. In humans, more than 90% of Pb accumulates in bones, and over 70% of that is found in the cortical bone [<xref ref-type="bibr" rid="CR27">27</xref>]. The biological half-life of Pb in bones is estimated at about 30 years [<xref ref-type="bibr" rid="CR2">2</xref>], and data related to connective tissue are lacking. None of the participants in our study reported specific exposure to environmental pollution; therefore, we have to assume a rather uniform exposure of our study groups. For this reason, we infer from our results that IVD is less accessible to environmental pollutants even as it presents a more stable concentration.</p><p>The next interesting finding comes from the Cu analysis. Cu is the only TE found at a higher concentration in IVD compared to bone. It was detected in some samples from the OA group while we observed its consistent presence in disc tissue. Cu is an essential element. In humans, it is part of ceruloplasmin, albumins, and amino acids and serves as a co-factor of lysyl oxidase, which is essential in cross-linking of collagen and elastin fibers [<xref ref-type="bibr" rid="CR17">17</xref>]. Cu also acts together with iron in oxidation&#x02013;reduction reactions and hemoglobin synthesis and is found in cytochrome c oxidase and superoxide dismutases [<xref ref-type="bibr" rid="CR28">28</xref>]. In addition, this element influences polypeptide formation [<xref ref-type="bibr" rid="CR29">29</xref>]. In serum level analysis, copper concentration, unlike calcium, does not show correlation with degenerative disc disease [<xref ref-type="bibr" rid="CR30">30</xref>]. Also it does not show correlation with osteopenia or osteoporosis in postmenopausal women [<xref ref-type="bibr" rid="CR31">31</xref>,<xref ref-type="bibr" rid="CR32">32</xref>].</p><p>The range of Cu in human tissues is between 0.7 and 7.8 mg kg^&#x02212;1, with the lowest concentration in muscles and the highest in liver. Mean Cu level in tissues of the &#x0201c;reference man&#x0201d; is considered to be about 1 mg kg^&#x02212;1 [<xref ref-type="bibr" rid="CR20">20</xref>], and the mean Cu concentration documented in the bone is from 0.62 [<xref ref-type="bibr" rid="CR9">9</xref>] to 0.8 mg kg^&#x02212;1 dw [<xref ref-type="bibr" rid="CR33">33</xref>]. Lanocha et al. found that Cu concentration in bone was not much different from that in cartilage, with an average concentration of 0.79 mg kg^&#x02212;1 dw, ranging between 0.20 and 1.78 mg kg^&#x02212;1 [<xref ref-type="bibr" rid="CR33">33</xref>]. In our study, the Cu bone concentration was higher than what has been reported previously, with mean concentrations for femoral head and neck of 1.21 and 1.63 mg kg^&#x02212;1, respectively. Interestingly, the range and mean value of Cu in the IVD tissue was almost twice that of bone (0.97&#x02013;6.09 vs 2.71 mg kg^&#x02212;1, respectively), even after one outlier was excluded (23.64 mg kg^&#x02212;1). Given that IVD is generally an avascular compartment, Cu compounds such as ceruloplasmin and albumin should be excluded from consideration. The question arises then of Cu&#x02019;s relationship to oxygen transport in a low oxygen concentration compartment. In this case, a high Cu concentration should be rather linked with collagen formation and healing. The results at any rate suggest a significant role for Cu and its compounds in the IVD tissue. As was the case with Pb, the IVD is a more stable compartment for Cu detection compared to bone.</p><p>Zinc and Mg showed no deviation from previous findings. The biological role of magnesium ions is quite extensive [<xref ref-type="bibr" rid="CR17">17</xref>]: they take part in nucleic acid chemistry with DNA and RNA synthesis; an array of enzymes requires their presence as the reaction co-factor; and they have a role in energetic nucleotide formation (ATP as the chelate with Mg ion). Magnesium also plays a role in the active transport of calcium and potassium ions across cell membranes, a process that is important to nerve impulse conduction, muscle contraction, and normal heart rhythm. Magnesium levels are well documented in a variety of tissues, including IVD and the similar temporomandibular joint disc [<xref ref-type="bibr" rid="CR34">34</xref>]. Tohno et al. [<xref ref-type="bibr" rid="CR4">4</xref>] have reported Mg at almost all levels of the spine, with an average Mg content of 1,196 mg kg^&#x02212;1, ranging from 600 to 2,200 mg kg^&#x02212;1, in agreement with our results of 758.17 mg kg^&#x02212;1 (range 182.6&#x02013;2,132 mg kg^&#x02212;1 dw). Differing slightly from our findings are values for the temporomandibular joint disc [<xref ref-type="bibr" rid="CR34">34</xref>] reported by Takano et al. of 524.74 vs 758.17 mg kg^&#x02212;1 dw, respectively. Similar values for Mg were also reported in the posterior longitudinal ligaments of the cervical spine (445 mg kg^&#x02212;1) [<xref ref-type="bibr" rid="CR35">35</xref>], which also were less than our results (161 and 494.8 mg kg^&#x02212;1 of SD, respectively). In soft tissue of the stomach, the concentration seems to be the lowest compared to tissues described in the literature, ranging from 30 to 300 mg kg^&#x02212;1 dw [<xref ref-type="bibr" rid="CR36">36</xref>]. Compared to IVD, the average concentration of Mg in bones may be more than two times higher at 1,792.9 mg kg^&#x02212;1 [<xref ref-type="bibr" rid="CR26">26</xref>]. Our study confirms the literature data for bone and IVD performed separately. Magnesium concentration was approximately two times higher in bone (1,661.21 and 1,458.49 mg kg^&#x02212;1 in femoral neck and head, respectively) compared to disc (758.17 mg kg^&#x02212;1).</p><p>The majority of Mg ions are intracellular (39%), and only 1% is stored extracellularly [<xref ref-type="bibr" rid="CR37">37</xref>], which might lead to an explanation of these tissue differences based on their cell density characteristics. The literature on their relative cellularity is unclear. Cellularity of the bone ranges from 0.5 to 10 kcells mm^&#x02212;3 [<xref ref-type="bibr" rid="CR38">38</xref>], but we have to recall that bone is divided into two major compartments, the cortical and trabecular bone. An additional variable is the age of the patient. For the femoral head and neck, values for marrow cellularity in the newborn are 100% but fall to 60% in the 10-year-old and to 25% in the adult [<xref ref-type="bibr" rid="CR39">39</xref>,<xref ref-type="bibr" rid="CR40">40</xref>]. The cell density of IVD used in analytical calculations is 9 kcells mm^&#x02212;3 for annulus fibrosus, 4 kcells mm^&#x02212;3 for nucleus pulposus, and 15 kcells mm^&#x02212;3 for the cartilage end plate [<xref ref-type="bibr" rid="CR41">41</xref>]. Other sources cite that in extreme cases, such as the cartilage, there may be ~10 cells mm^&#x02212;3 [<xref ref-type="bibr" rid="CR42">42</xref>]. Considering the above, it is possible to infer that cellularity of the bone is double that of the disc and that the twofold higher concentration of Mg in the analyzed samples might relate to cell concentration.</p><p>Zinc, as well as Mg, is an enzyme co-factor [<xref ref-type="bibr" rid="CR43">43</xref>] and involved in DNA and protein synthesis and cell division [<xref ref-type="bibr" rid="CR44">44</xref>], as well as intracellular regulation [<xref ref-type="bibr" rid="CR45">45</xref>]. Zinc is believed to have antioxidant properties with anti-aging effects and an influence on the healing process [<xref ref-type="bibr" rid="CR46">46</xref>]. It is also believed to affect immune response mechanisms [<xref ref-type="bibr" rid="CR47">47</xref>]. The most important role of Zn in the IVD is probably the formation the matrix metalloproteinases that are Zn-dependent endopeptidases. The metal ions act as co-factors, distinguishing these endopeptidases from the others and commonly occur both in invertebrates and plants. The role of the metalloproteinases is degradation of the extracellular matrix, and its presence or a synthesis imbalance is related to IVD degeneration [<xref ref-type="bibr" rid="CR48">48</xref>]. It is known that Zn deficiency co-occurring in pinealectomized chickens can lead to developmental spine deformity [<xref ref-type="bibr" rid="CR49">49</xref>]. But the pathology can not be attributed to the pathological changes in the IVD.</p><p>The close relationship of the Mg and Zn may favor a concept of simultaneous regeneration and degeneration processes taking place in the IVD. Zinc concentration in the IVD determined in our study was similar to levels encountered in posterior longitudinal ligaments by Kumai et al. (34.61 and 36 mg kg^&#x02212;1 dw, respectively) [<xref ref-type="bibr" rid="CR50">50</xref>]. The average concentration in cartilage and bone was twice as high at 88.3 mg kg^&#x02212;1 (range 54.3&#x02013;163.8 mg kg^&#x02212;1) dw and 84.58 mg kg^&#x02212;1 (SD 17.68 mg kg^&#x02212;1), respectively [<xref ref-type="bibr" rid="CR26">26</xref>,<xref ref-type="bibr" rid="CR33">33</xref>]. In our study, Zn concentration, similar to the findings above, was approximately twice as high in bone compared to disc. Because of a lack of reference data for Zn in non-degenerated IVD, it is not possible to state whether we should link the concentration of the element with the biological activity of matrix metalloproteinases. Based on the concentration observed in the tendons, we would be prone to expect higher Zn values in degenerated disc tissue.</p><p>The only study comparing metal concentrations in both bone and IVD was performed by Minami et al. [<xref ref-type="bibr" rid="CR5">5</xref>] evaluating platinum levels in cis-platinum-treated patients. In that work, they showed that IVD concentration may be up to 4.3 times higher compared to bone. They determined the bone level in the vertebral body, which is the direct transportation route to and from the IVD. The concentration ratio of specific TE in bone and IVD may be related not only to exposure but also to tissue affinity and metabolic profile, which could be a basis for further studies.</p></sec><sec id="Sec8" sec-type="conclusion"><title>Conclusions</title><p>Except for Cu, the TE concentrations were higher in bone compared to IVD.</p><p>This study showed a higher concentration of Cu in disc tissue compared to bone, which may be related to cross-linking in collagen formation and healing processes and should be a subject of further study.</p><p>In addition, IVD tissue seems to be a more stable compartment for evaluating TE concentration, especially environmentally related TEs. In the case of disc tissue, a higher ratio of IVD samples had a concentration of Pb, Mo, and Ni within the detection threshold compared to bone. It may be better to consider IVD, compared to bone, as the indicator tissue in biochemical studies.</p></sec>

Early activation of pro-fibrotic WNT5A in sepsis-induced acute lung injury

Could not extract abstract

<contrib contrib-type="author" corresp="yes"><name><surname>Villar</surname><given-names>Jes&#x000fa;s</given-names></name><address><email>jesus.villar54@gmail.com</email></address><xref ref-type="aff" rid="Aff1"/><xref ref-type="aff" rid="Aff2"/><xref ref-type="aff" rid="Aff3"/></contrib><contrib contrib-type="author"><name><surname>Cabrera-Ben&#x000ed;tez</surname><given-names>Nuria E</given-names></name><address><email>nuriaecb@gmail.com</email></address><xref ref-type="aff" rid="Aff1"/><xref ref-type="aff" rid="Aff2"/></contrib><contrib contrib-type="author"><name><surname>Ramos-Nuez</surname><given-names>Angela</given-names></name><address><email>shefloanse@hotmail.com</email></address><xref ref-type="aff" rid="Aff1"/><xref ref-type="aff" rid="Aff2"/></contrib><contrib contrib-type="author"><name><surname>Flores</surname><given-names>Carlos</given-names></name><address><email>cflores@ull.edu.es</email></address><xref ref-type="aff" rid="Aff1"/><xref ref-type="aff" rid="Aff4"/></contrib><contrib contrib-type="author"><name><surname>Garc&#x000ed;a-Hern&#x000e1;ndez</surname><given-names>Sonia</given-names></name><address><email>soniagaher76@gmail.com</email></address><xref ref-type="aff" rid="Aff5"/></contrib><contrib contrib-type="author"><name><surname>Valladares</surname><given-names>Francisco</given-names></name><address><email>fvallapa@gmail.com</email></address><xref ref-type="aff" rid="Aff1"/><xref ref-type="aff" rid="Aff5"/></contrib><contrib contrib-type="author"><name><surname>L&#x000f3;pez-Aguilar</surname><given-names>Josefina</given-names></name><address><email>JLopezA@tauli.cat</email></address><xref ref-type="aff" rid="Aff1"/><xref ref-type="aff" rid="Aff6"/></contrib><contrib contrib-type="author"><name><surname>Blanch</surname><given-names>Llu&#x000ed;s</given-names></name><address><email>lblanch@tauli.cat</email></address><xref ref-type="aff" rid="Aff1"/><xref ref-type="aff" rid="Aff6"/></contrib><contrib contrib-type="author"><name><surname>Slutsky</surname><given-names>Arthur S</given-names></name><address><email>lblanch@tauli.cat</email></address><xref ref-type="aff" rid="Aff3"/><xref ref-type="aff" rid="Aff7"/></contrib><aff id="Aff1"><label/>CIBER de Enfermedades Respiratorias, Instituto de Salud Carlos III, Madrid, Spain </aff><aff id="Aff2"><label/>Multidisciplinary Organ Dysfunction Evaluation Research Network, Research Unit, Hospital Universitario Dr. Negrin, Las Palmas de Gran Canaria, Spain </aff><aff id="Aff3"><label/>Keenan Research Center for Biomedical Science, Li Ka Shing Knowledge Institute, St. Michael&#x02019;s Hospital, Toronto, Canada </aff><aff id="Aff4"><label/>Research Unit, Hospital Universitario NS de Candelaria, Santa Cruz de Tenerife, Spain </aff><aff id="Aff5"><label/>Department of Anatomy, Pathology &#x00026; Histology, Medical School University of La Laguna and Hospital Universitario de Canarias, La Laguna, Tenerife Spain </aff><aff id="Aff6"><label/>Critical Care Center, Corporaci&#x000f3; Sanitaria Parc Taul&#x000ed;, Sabadell, Barcelona Spain </aff><aff id="Aff7"><label/>Interdepartmental Division of Critical Care Medicine, University of Toronto, Toronto, ON Canada </aff>

Critical Care

<sec id="Sec1" sec-type="introduction"><title>Introduction</title><p>Acute respiratory distress syndrome (ARDS) is a severe inflammatory process caused by pulmonary or systemic insults to the lung alveolar-capillary barrier [<xref ref-type="bibr" rid="CR1">1</xref>-<xref ref-type="bibr" rid="CR3">3</xref>]. Sepsis is the most common predisposing factor underlying ARDS and is characterized by systemic inflammation in response to circulating microbes or microbial toxins such as lipopolysaccharide (LPS), also termed endotoxin, a component of the cell wall of gram-negative bacteria. Sepsis and sepsis-induced ARDS are common syndromes associated with high morbidity and mortality [<xref ref-type="bibr" rid="CR1">1</xref>,<xref ref-type="bibr" rid="CR2">2</xref>,<xref ref-type="bibr" rid="CR4">4</xref>]. Effective repair of the alveolar epithelium requires proliferation and migration of type-II alveolar epithelial cells, and their differentiation into type-I alveolar cells [<xref ref-type="bibr" rid="CR1">1</xref>,<xref ref-type="bibr" rid="CR5">5</xref>]. In addition, lung fibroblast migration and proliferation occur early after lung injury and are necessary for ongoing lung healing [<xref ref-type="bibr" rid="CR6">6</xref>-<xref ref-type="bibr" rid="CR8">8</xref>]. Damage to the alveolar epithelium can lead to abnormal repair that culminates in a vigorous fibroblastic response, leading to uncontrolled extracellular matrix deposition and destruction of lung parenchymal architecture [<xref ref-type="bibr" rid="CR8">8</xref>,<xref ref-type="bibr" rid="CR9">9</xref>].</p><p>The role of &#x003b2;-catenin-mediated wingless integration (Wnt) signaling is proving to be central to mechanisms of lung healing and fibrosis [<xref ref-type="bibr" rid="CR10">10</xref>,<xref ref-type="bibr" rid="CR11">11</xref>]. Tissue repair involves re-epithelialization, in which injured cells are replaced by cells of the same type and normal parenchyma may be replaced by connective tissue leading to fibrosis [<xref ref-type="bibr" rid="CR11">11</xref>]. K&#x000f6;nigshoff <italic>et al</italic>. [<xref ref-type="bibr" rid="CR10">10</xref>] showed that WNT ligands induce lung epithelial cell proliferation, fibroblast activation, and collagen synthesis, and is upregulated in a bleomycin-induced lung injury model and also in humans with idiopathic pulmonary fibrosis. Wnt binding to cognate Frizzled receptors results in cytosolic accumulation of &#x003b2;-catenin, which then translocates to the nucleus and participates in gene transcription [<xref ref-type="bibr" rid="CR11">11</xref>-<xref ref-type="bibr" rid="CR13">13</xref>]. Wnt/&#x003b2;-catenin signaling stimulates tissue remodeling and wound closure, or tissue remodeling and destruction through matrix metallopeptidases (MMPs) and other gene products [<xref ref-type="bibr" rid="CR14">14</xref>]. This activation stimulates many of the pro-inflammatory cytokines participating in inflammation-mediated lung destruction and hyaline membrane formation [<xref ref-type="bibr" rid="CR12">12</xref>], and induces expression of growth-associated genes such as cyclin D1 and vascular endothelial growth factor (VEGF) [<xref ref-type="bibr" rid="CR15">15</xref>]. MMP7 (also known as matrilysin) is a target gene of the Wnt signaling pathway found on the surface of lung epithelial cells and is a key regulator of pulmonary fibrosis [<xref ref-type="bibr" rid="CR16">16</xref>].</p><p>In the present study, we examined the hypothesis that the Wnt/&#x003b2;-catenin pathway is activated in the lungs very early after sepsis and plays a role in initiating the lung repair process. To test this hypothesis we used a well-established LPS-induced cell injury model using human lung cells based on the first steps in the development of sepsis and sepsis-induced ARDS [<xref ref-type="bibr" rid="CR17">17</xref>-<xref ref-type="bibr" rid="CR21">21</xref>]. Then, we validated the main gene targets of this pathway in a clinically relevant murine model of sepsis-induced ARDS by cecal ligation and perforation (CLP), and in lung biopsies obtained from patients who died within the first 24 h of septic ARDS.</p></sec><sec id="Sec2" sec-type="materials|methods"><title>Materials and methods</title><sec id="Sec3"><title><italic>In vitro</italic> studies</title><p>We used healthy human bronchial epithelial BEAS-2B cells (ATCC, Manassas, VA, USA) and human lung MRC-5 fibroblasts. BEAS-2B cells were cultured as previously described [<xref ref-type="bibr" rid="CR17">17</xref>] in Dulbecco&#x02019;s modified Eagle&#x02019;s medium supplemented with 10% FBS and penicillin and streptomycin, at 37&#x000b0;C in a 5% CO₂, 95% humidified air incubator. MRC-5 cells were obtained from the Department of Microbiology (Hospital Universitario Dr Negr&#x000ed;n, Las Palmas, Spain) and cultured in RPMI-1640 medium with 10% FBS under the same experimental conditions. We chose human BEAS-2B and MRC-5 cells as representative lines for studying changes during acute lung injury because these cell lines have been validated previously in experimental models addressing the first steps of sepsis-induced ARDS [<xref ref-type="bibr" rid="CR18">18</xref>-<xref ref-type="bibr" rid="CR24">24</xref>]. For all experiments, BEAS-2B and MRC-5 cells were stimulated with 100 ng/mL of LPS obtained from <italic>Escherichia coli</italic> (Sigma-Aldrich, St Louis, MO, USA), a concentration used in previous studies for induction of inflammatory responses [<xref ref-type="bibr" rid="CR18">18</xref>,<xref ref-type="bibr" rid="CR21">21</xref>] and validated to study LPS-induced effects [<xref ref-type="bibr" rid="CR22">22</xref>].</p></sec><sec id="Sec4"><title>Inhibition of cell proliferation</title><p>BEAS-2B and MRC-5 cells were suspended in 5 &#x000d7; 10⁶ cells/flask and inoculated in 75 cm² flasks. After 24 h, cells were exposed or not, to LPS (100 ng/mL) for 18 h, and then examined and photographed (Olympus Camedia digital camera) under a phase-contrast microscope (Olympus CK-40 F-200, Tokyo, Japan). The effects of LPS on cell growth were assessed using the Sulforhodamine B colorimetric assay (SRB, Sigma-Aldrich) [<xref ref-type="bibr" rid="CR25">25</xref>] (see Additional file <xref rid="MOESM1" ref-type="media">1</xref> for details).</p></sec><sec id="Sec5"><title>Western blotting</title><p>Protein levels of WNT5A, total &#x003b2;-catenin, non-phospho (Ser33/37/Thr41) &#x003b2;-catenin, MMP7, cyclin D1, and VEGF were measured by western blotting. For total protein extracts, cells were homogenized in radioimmunoprecipitation assay (RIPA) protein extract buffer, as described previously [<xref ref-type="bibr" rid="CR26">26</xref>] (see Additional file <xref rid="MOESM1" ref-type="media">1</xref> for further details). Bands were detected by chemiluminescence (Amersham Reagents, GE Healthcare, Fairfield, CN, USA) and blots were measured by Scion Image software package (Scion Corp, Frederick, MD, USA).</p></sec><sec id="Sec6"><title><italic>In vivo</italic> experimental animal model</title><p>In an attempt to translate the <italic>in vitro</italic> observations into the disease state of interest (sepsis and ARDS), we performed histological and immunohistochemical examination of lungs from a clinically relevant experimental animal model of sepsis-induced lung injury. The experimental protocol was approved by the Animal Care Committee at the Hospital Universitario Dr Negrin, Las Palmas de Gran Canaria, Spain (CEEBA#003/10), in accordance with the European Commission Directive 2010/63/EU for animal experimentation. This study followed the guidelines, Animal Research: Reporting of in Vivo Experiments (ARRIVE), for reporting animal research [<xref ref-type="bibr" rid="CR27">27</xref>].</p><p>We studied eight healthy male Sprague-Dawley rats weighing 300 to 350 g. After anesthesia with intraperitoneal injection of xylazine and ketamine hydrochloride, animals were randomized to control (sham-sepsis) (n = 3) or sepsis (n = 5). Sepsis was induced by CLP. A detailed description of this experimental model is provided elsewhere [<xref ref-type="bibr" rid="CR28">28</xref>]. Sham-CLP underwent the same surgical procedures as CLP rats: the cecum was exposed (but not ligated or punctured) and returned to the abdominal cavity, and the abdominal wall was then sutured. Eighteen hours later, control animals and the first three surviving septic animals were anesthetized and sacrificed. A midline thoracotomy/laparotomy was performed and the heart and lungs were removed <italic>en bloc</italic>. The lungs were isolated from the heart, the trachea was cannulated, and the right lung was fixed by intratracheal instillation of 3 mL of 10% formalin and floated in 10% formalin for a week. Lungs were serially sliced from apex to base and embedded in paraffin, cut (3-&#x003bc;m thickness sections) and stained with hematoxylin and eosin for microscope observation. Two pathologists (FV, SGH) were blinded to the sample identity. Three random sections from each animal were examined with particular reference to alveolar and interstitial damage defined by the presence of pulmonary edema, inflammatory cell infiltration, vascular congestion, and fibrosis. Slides were viewed using a Nikon Optiphot 2 microscope and photographed in a Nikon Digital Sight DS-5 M camera (Tokyo, Japan) at &#x000d7; 200 magnification.</p><p>We also used the Sirius-red staining technique [<xref ref-type="bibr" rid="CR29">29</xref>] for assessment of collagen content, as described elsewhere [<xref ref-type="bibr" rid="CR30">30</xref>]. We defined fibrosis as the presence of collagen. With this technique, collagen fibers are stained bright red and nuclei/cytoplasm are bright yellow. Slides were viewed with an Olympus (Bx50) microscope and photographed with an Olympus digital camera at &#x000d7; 200 magnification.</p></sec><sec id="Sec7"><title>Human lung tissue from autopsies</title><p>For translating the <italic>in vitro</italic> and <italic>in vivo</italic> observations into the human disease state of interest (sepsis and ARDS), we performed histological and immunohistochemical examination of human lungs from patients who died very early in their course of severe sepsis. Two pathologists (FV, SGH) analyzed the lungs of 12 patients from the archives of autopsies performed between 2007 and 2012 at the Department of Pathology of the University of La Laguna Medical School, Tenerife, Spain. A waiver of ethics was granted by the Ethics Committee for Clinical Research at the Hospital Universitario de Canarias (Tenerife, Spain), as informed consent is systematically obtained from patients&#x02019; relatives for both clinical autopsy and potential use of tissue samples in teaching and research purposes. An anonymized summary with clinical relevant information of patients who have had an autopsy is stored in a specific database of the Department of Pathology for further review when necessary. Control lungs were selected from six autopsies in patients who died from diseases without any lung involvement. Septic lungs were selected from autopsies performed in six patients meeting standard criteria for severe sepsis [<xref ref-type="bibr" rid="CR4">4</xref>] and ARDS [<xref ref-type="bibr" rid="CR1">1</xref>-<xref ref-type="bibr" rid="CR3">3</xref>], who did not receive mechanical ventilation and died within the first 24 h of developing severe sepsis. Pathologists were asked to select the autopsies of interest following a strict chronological order, starting with those performed in 2012 and continuing yearly backwards, without any preference or selection bias. After identification of the patients from the postmortem examination, they checked with the institutional database to confirm the clinical diagnosis.</p><p>Paraffin blocks of lung tissue collected during autopsy were retrieved from the Department of Pathology archives. In the routine autopsies, three to four fragments of lung parenchyma are obtained. In normal lungs, one fragment of lung tissue was collected from each lobe. The tissue had been fixed in 10% buffered formalin, routinely processed and paraffin embedded. Sections of 3-&#x003bc;m thickness were stained with hematoxylin and eosin and the Sirius-red technique, and evaluated for acute lung injury and collagen content.</p></sec><sec id="Sec8"><title>Immunocytochemistry</title><p>Immunocytochemical stains were performed by applying a standard avidin-biotin complex technique (see Additional file <xref rid="MOESM1" ref-type="media">1</xref> for further details). To view slides, we used an Olympus BX50 microscope and an Olympus Camedia digital camera at &#x000d7; 400 magnification.</p></sec><sec id="Sec9"><title>Statistical analysis</title><p>For the statistical power analysis for sample size calculations in both categories of autopsies (diseases with no lung involvement, septic ARDS), we estimated that to detect at least a 2-fold increase in the immunostaining intensity of fibrotic markers (WNT5A, MMP7) in septic lungs compared to the basal intensity in patients without sepsis, we would require six patients in each group, with an alpha of 0.05 and a power greater than 0.80.</p><p>Data are expressed as mean &#x000b1; SD, and were analyzed using Graph Pad Prism software version 5.0. Data are from different experiments and samples within each group. Comparisons involving all experimental cell groups were performed with one-way analysis of variance. We used the Bonferroni correction for multiple comparisons. For western blot experiments, densitometry data of the non-phospho (Ser33/37/Thr41) &#x003b2;-catenin bands were normalized to &#x003b2;-catenin and &#x003b2;-actin (as loading controls), and densitometry of the active form (20 kDa) of MMP7 was normalized to the inactive form (30 kDa) and then normalized to &#x003b2;-actin. Data are from at least three independent experiments. A two-tailed <italic>P</italic>-value &#x0003c;0.05 was considered significant.</p></sec></sec><sec id="Sec10" sec-type="results"><title>Results</title><sec id="Sec11"><title>In vitro studies</title><p>LPS decreased the proliferation of MRC-5 and BEAS-2B cells. The maximum effect on cell viability in both cell types was observed at 18 h using 100 ng/mL LPS (data not shown).</p></sec><sec id="Sec12"><title>WNT5A and associated proteins</title><p>WNT5A protein levels were significantly increased in MRC-5 and BEAS-2B cells (<italic>P</italic> &#x0003c;0.001) after LPS exposure (Figures&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>A and <xref rid="Fig2" ref-type="fig">2</xref>A, respectively). LPS stimulation led to a significant increase in non-phospho (Ser33/37/Thr41) &#x003b2;-catenin (Figures&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>B, <xref rid="Fig2" ref-type="fig">2</xref>B). The active form of the MMP7 protein was increased in both MRC-5 and BEAS-2B cells stimulated with LPS. LPS treatment also caused increased upregulation of cyclin D1 and VEGF (Figures&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>A, <xref rid="Fig2" ref-type="fig">2</xref>A). Immunocytochemical staining detected non-phospho Ser33/37/Thr41 &#x003b2;-catenin at the nuclei of MRC-5 and BEAS-2B cells stimulated with LPS (Figure&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>).<fig id="Fig1"><label>Figure 1</label><caption><p> <bold>Activation of WNT5A/&#x003b2;-catenin pathway by</bold> <bold><italic>E</italic></bold> <bold>.</bold> <bold><italic>coli</italic></bold> <bold>lipopolysaccharide (LPS) in human BEAS-2B cells. (A)</bold> Changes in total WNT5A, cyclin D1, metallopeptidase (MMP)7 and vascular endothelial growth factor (VEGF) proteins (representative blots and mean densitometric values) following LPS stimulation for 18 h. Densitometry analysis of the active form (20 kDa) of MMP7 was normalized to the inactive form (30 kDa) and then normalized to &#x003b2;-actin. <bold>(B)</bold> Changes in non-phosphorylated (Ser33/37/Thr41) &#x003b2;-catenin protein bands were normalized to total &#x003b2;-catenin and &#x003b2;-actin. ***<italic>P</italic> &#x0003c;0.001 versus control-vehicle <bold>(C)</bold>.</p></caption><graphic xlink:href="13054_2014_568_Fig1_HTML" id="MO1"/></fig><fig id="Fig2"><label>Figure 2</label><caption><p> <bold>Representative western blots for WNT5A/&#x003b2;-catenin pathway stimulated by LPS in human MRC-5 cells. (A)</bold> Changes in total WNT5A, cyclin D1, metallopeptidase (MMP)7 and vascular endothelial growth factor (VEGF) proteins following exposure to 100 ng/mL lipopolysaccharide (LPS) stimulation for 18 h. Densitometry analysis of the active form (20 kDa) of MMP7 was normalized to the inactive form (30 kDa) and then normalized to &#x003b2;-actin. <bold>(B)</bold> Changes in non-phospho Ser33/37/Thr41 &#x003b2;-catenin after 18 h of LPS stimulation. Densitometry was performed on at least three different blots per condition and normalized to the respective loading control (&#x003b2;-actin). Protein expression is expressed as fold-change relative to the respective control vehicle <bold>(C)</bold>. ***<italic>P</italic> &#x0003c;0.001 versus control vehicle <bold>(C)</bold>.</p></caption><graphic xlink:href="13054_2014_568_Fig2_HTML" id="MO2"/></fig><fig id="Fig3"><label>Figure 3</label><caption><p> <bold>Non-phospho Ser33/37/Thr41 &#x003b2;-catenin immunolocalization on BEAS-2B and MRC-5 cells stimulated with lipopolysaccharide (LPS).</bold> Red-pink colour indicates positive staining (3-amino-9-ethylcarbazole) for non-phospho Ser33/37/Thr41 &#x003b2;-catenin protein. Non-phospho Ser33/37/Thr41 &#x003b2;-catenin staining was found in nuclei (large arrows) in cells stimulated with LPS but not in control-vehicle cells (C). The images (at &#x000d7; 200 magnification) are representative of experiments performed in triplicate. Scale bars = 20 &#x003bc;m.</p></caption><graphic xlink:href="13054_2014_568_Fig3_HTML" id="MO3"/></fig></p></sec><sec id="Sec13"><title>Animal model</title><p>CLP induced typical signs of disease including lethargy, ruffled fur, generalized weakness, reduced gross motor activity, and weight loss, accordingly with the literature [<xref ref-type="bibr" rid="CR28">28</xref>]. Three out of five septic animals survived 18 h after CLP, and these animals were studied further. Lungs from septic animals showed acute inflammatory infiltrates, perivascular edema and collagen deposition in the parenchyma (Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>, panel B). The Sirius-red staining for collagen was negative in control animals (Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>, panel D). Healthy control animals had a basal intensity of WNT5A and MMP7 whereas septic lungs showed strong immunohistochemistry intensity of WNT5A and MMP7 (Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>, panels F and H).<fig id="Fig4"><label>Figure 4</label><caption><p> <bold>Representative histological features of healthy and septic rat lungs and immunohistochemical staining for WNT5A and MMP7 activation in healthy and septic rat lungs. (A-D)</bold> Histological features: <bold>(A)</bold> normal lung (healthy) and <bold>(B)</bold> septic lung showing pulmonary infiltrates and perivascular edema; <bold>(C)</bold> normal lung (healthy) and <bold>(D)</bold> septic lung stained with Sirius-red assay for collagen content (&#x000d7; 200 magnification). <bold>(E-</bold> <bold>H)</bold> Red-pink color indicates positive staining for WNT5A <bold>(E,</bold> <bold>F)</bold> and metallopeptidase (MMP)7 <bold>(G,</bold> <bold>H)</bold> and blue/violet indicates nuclei counterstained with hematoxylin. WNT5A and MMP7 were observed in alveolar walls and septa (&#x000d7; 200 magnification).</p></caption><graphic xlink:href="13054_2014_568_Fig4_HTML" id="MO4"/></fig></p></sec><sec id="Sec14"><title>Human lungs from autopsies</title><p>Clinical diagnoses of six patients who died with septic ARDS and six control subjects who died from non-pulmonary causes within 24 h of disease onset are presented in Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>. No relevant findings were found in the lungs from patients who died without lung disease. Lungs from septic patients showed features of diffuse lung damage, manifested by acute inflammatory infiltrates and perivascular edema (Figure&#x000a0;<xref rid="Fig5" ref-type="fig">5</xref>, panel B). Lungs from septic patients showed high intensity of collagen-rich areas in the parenchyma, providing evidence of the presence of a fibrotic response in the early stages of sepsis-induced lung injury (Figure&#x000a0;<xref rid="Fig5" ref-type="fig">5</xref>, panel D). Lungs from patients without pulmonary disease had a basal intensity of WNT5A and MMP7 whereas lungs from septic patients showed a strong immunohistochemical intensity of WNT5A and MMP7 (Figure&#x000a0;<xref rid="Fig5" ref-type="fig">5</xref>, panels F and H)<table-wrap id="Tab1"><label>Table 1</label><caption><p> <bold>Clinical diagnosis of patients with non-pulmonary diseases and sepsis-induced acute lung injury</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th> <bold>Main diagnosis</bold> </th><th> <bold>Cause/mechanism of death</bold> </th></tr></thead><tbody><tr valign="top"><td> <bold>Non-pulmonary diseases</bold> </td><td/></tr><tr valign="top"><td>Acute myocardial infarction</td><td align="left">Ventricular fibrillation</td></tr><tr valign="top"><td>Acute myocardial infarction</td><td align="left">Left ventricular rupture</td></tr><tr valign="top"><td>End-stage colon cancer</td><td align="left">Myocardial infarction</td></tr><tr valign="top"><td>End-stage liver cirrhosis</td><td align="left">Liver failure</td></tr><tr valign="top"><td>Coronary artery disease</td><td align="left">Severe cardiac arrhythmias</td></tr><tr valign="top"><td>Epilepsy, Down syndrome</td><td align="left">Severe hypoxic encephalopathy</td></tr><tr valign="top"><td> <bold>Sepsis-induced lung injury</bold> </td><td/></tr><tr valign="top"><td>Acute peritonitis</td><td align="left">Septic shock</td></tr><tr valign="top"><td>AIDS</td><td align="left">Pneumonia</td></tr><tr valign="top"><td>Pneumonia</td><td align="left">Septic shock</td></tr><tr valign="top"><td>Abdominal sepsis</td><td align="left">Septic shock</td></tr><tr valign="top"><td>Abdominal sepsis</td><td align="left">Septic shock</td></tr><tr valign="top"><td>Pneumonia</td><td align="left">Septic shock</td></tr></tbody></table></table-wrap><fig id="Fig5"><label>Figure 5</label><caption><p> <bold>Representative histological features of healthy and septic human lungs and immunohistochemical staining for WNT5A and MMP7 activation in normal and septic human lungs. (A-D)</bold> Histological features: <bold>(A)</bold> normal lung (healthy) and <bold>(B)</bold> septic lung showing pulmonary infiltrates and perivascular edema; <bold>(C)</bold> normal lung and septic lung <bold>(D)</bold> stained with Sirius-red assay for collagen content (&#x000d7; 200 magnification). <bold>(E-</bold> <bold>H)</bold> WNT5A <bold>(E,</bold> <bold>F)</bold> and metallopeptidase (MMP)7 <bold>(G,</bold> <bold>H)</bold> are shown in red-pink color. Tissues were counterstained with hematoxylin. There was increased immunoreactivity for WNT5A and MMP7 in alveolar walls and septa (&#x000d7; 200 magnification).</p></caption><graphic xlink:href="13054_2014_568_Fig5_HTML" id="MO5"/></fig></p></sec></sec><sec id="Sec15" sec-type="discussion"><title>Discussion</title><p>We examined the translational impact of the WNT/&#x003b2;-catenin pathway in an LPS-induced human lung cell injury model and validated the main gene targets of this pathway in the lungs of septic experimental animals and in human lungs from autopsies. The major findings of our study are: (1) WNT5A is expressed very early by human airway epithelial cells and lung fibroblasts in response to LPS; (2) upregulation of WNT5A expression and non-phospho Ser33/37/Thr41 &#x003b2;-catenin are associated with upregulation of downstream target genes that are involved in profibrotic transformation of injured tissues, such as MMP7, cyclin D1 and VEGF; and (3) pulmonary fibrosis is induced very early during sepsis-induced ARDS, both experimentally and clinically. These findings suggest that WNT5A and &#x003b2;-catenin contribute very early to repair the damage to lung tissue and may play a role in restructuring lung architecture during sepsis-induced ARDS.</p><p>We selected BEAS-2B and MRC-5 cell lines as representative human airway epithelial cells and lung fibroblasts because these cells have been implicated in the pathogenesis of sepsis-induced ARDS [<xref ref-type="bibr" rid="CR18">18</xref>-<xref ref-type="bibr" rid="CR24">24</xref>] and subsequent fibrosis [<xref ref-type="bibr" rid="CR31">31</xref>]. These cell models provide a powerful translational <italic>in vitro</italic> approach for recapitulating human ARDS. LPS-treated human BEAS-2B cells are an accepted and validated <italic>in vitro</italic> cell injury model of the acute lung inflammatory response based on the first steps in the development of sepsis and sepsis-induced ARDS [<xref ref-type="bibr" rid="CR18">18</xref>]. Lung airway epithelial cells and fibroblasts generate various immune effectors such as cytokines, chemokines, and several peptides in response to inflammatory stimuli [<xref ref-type="bibr" rid="CR23">23</xref>,<xref ref-type="bibr" rid="CR32">32</xref>], which control lung inflammation, lung injury and lung repair [<xref ref-type="bibr" rid="CR9">9</xref>,<xref ref-type="bibr" rid="CR12">12</xref>,<xref ref-type="bibr" rid="CR31">31</xref>,<xref ref-type="bibr" rid="CR33">33</xref>]. We selected <italic>E. coli</italic> LPS because it has been used in most endotoxin-induced lung injury models [<xref ref-type="bibr" rid="CR21">21</xref>,<xref ref-type="bibr" rid="CR34">34</xref>] and LPS is a key pathogen recognition molecule for sepsis [<xref ref-type="bibr" rid="CR33">33</xref>,<xref ref-type="bibr" rid="CR34">34</xref>]. Because previous <italic>in vitro</italic> studies using LPS-stimulated airway epithelial cells and fibroblasts focused on activation of pro-inflammatory mediators and increased cytokine release [<xref ref-type="bibr" rid="CR20">20</xref>,<xref ref-type="bibr" rid="CR35">35</xref>,<xref ref-type="bibr" rid="CR36">36</xref>], we examined the modulation of WNT5A, &#x003b2;-catenin, MMP7, cyclin D1 and VEGF molecules that contribute to lung repair and fibrosis [<xref ref-type="bibr" rid="CR12">12</xref>,<xref ref-type="bibr" rid="CR16">16</xref>,<xref ref-type="bibr" rid="CR37">37</xref>].</p><p>We extended our <italic>in vitro</italic> findings by confirming that collagen synthesis and the main target gene products of this pathway (WNT5A, MMP7) increased in a clinically relevant model of sepsis-induced lung injury and in lungs from patients who died with severe sepsis and ARDS. We used CLP as a clinically relevant and well characterized animal model to explore the fibrotic transformation in the lungs during the first 24&#x000a0;h of sepsis. CLP induced a reproducible and consistent septic and sepsis-induced ARDS condition in accordance with previous studies [<xref ref-type="bibr" rid="CR17">17</xref>,<xref ref-type="bibr" rid="CR28">28</xref>]. Histopathological features of CLP-induced ARDS in animals included atelectasis, pulmonary edema, and acute inflammatory infiltrates. Lung tissue damage is observed in 90% of patients dying from sepsis [<xref ref-type="bibr" rid="CR38">38</xref>]. Moreover, lung cells can activate mechanisms for initiating tissue repair, a process which involves re-epithelialization; injured cells are replaced by cells of the same type, but in some cases, normal parenchyma is replaced by connective tissue leading to fibrosis [<xref ref-type="bibr" rid="CR11">11</xref>]. There is evidence of fibrotic changes in the earliest stages of ARDS [<xref ref-type="bibr" rid="CR26">26</xref>,<xref ref-type="bibr" rid="CR39">39</xref>,<xref ref-type="bibr" rid="CR40">40</xref>]. &#x003b2;-catenin signaling stimulates tissue remodeling, cell migration, and wound closure through MMPs, but if the process is uncontrolled, it can drive tissue destruction through MMPs and other mediators [<xref ref-type="bibr" rid="CR11">11</xref>]. Wnt ligands induce lung epithelial cell proliferation, fibroblast activation and collagen synthesis [<xref ref-type="bibr" rid="CR16">16</xref>]. Collagen and other matrix extracellular molecules are the main components of the extracellular matrix, and MMP7 is a key mediator of pulmonary fibrosis [<xref ref-type="bibr" rid="CR16">16</xref>].</p><p>Several <italic>Wnt</italic> genes are expressed in the developing and adult lung. Of these, <italic>Wnt5a</italic> and <italic>Wnt7b</italic> are expressed at high levels in the airway epithelium [<xref ref-type="bibr" rid="CR14">14</xref>]. We chose to examine the modulation of WNT5A because it has been implicated in several pulmonary disorders [<xref ref-type="bibr" rid="CR11">11</xref>] and has not been studied in the context of sepsis and LPS-induced ARDS. In our study, WNT5A was detected with moderate intensity in alveolar walls and septa in the lungs of CLP rats and in the lungs of humans who died with early septic ARDS. Blumenthal <italic>et al</italic>. [<xref ref-type="bibr" rid="CR41">41</xref>] reported that the expression of WNT5A required Toll-like receptor signaling and NF-&#x003ba;B activation. In a previous report by our group, and using the same epithelial cell injury model as in the present study, we showed that LPS modulated the NF-&#x003ba;B activation through the Toll-like receptor signaling [<xref ref-type="bibr" rid="CR22">22</xref>]. The fact that &#x003b2;-catenin is rapidly upregulated in our epithelial/fibroblast cell injury model suggests that the WNT/&#x003b2;-catenin pathway could be continuously stimulated during ARDS and it could be a mechanism for perpetuating lung injury or for initiating lung repair. Thus, the activation of Wnt signaling after sepsis-induced ARDS likely represents a regenerative signal of the damaged epithelium [<xref ref-type="bibr" rid="CR42">42</xref>]. Using expression microarrays, Vuga <italic>et al</italic>. [<xref ref-type="bibr" rid="CR43">43</xref>] showed that WNT5A was significantly increased in fibroblasts isolated from lung tissues of patients with lung fibrosis compared with fibroblasts from normal lung tissues. They also reported increased cell proliferation when normal lung fibroblasts were treated with WNT5A.</p><p>Our findings parallel those of Chilosi <italic>et al</italic>. [<xref ref-type="bibr" rid="CR44">44</xref>] who found aberrant WNT/&#x003b2;-catenin pathway activation in lungs from patients with idiopathic pulmonary fibrosis, suggesting that this pathway could be responsible for dysfunctional lung repair processes leading to severe and irreversible pulmonary remodeling. This is a relevant translational finding because the development of pulmonary fibrosis has been found to have a direct correlation with severity of lung injury and mortality in ARDS patients [<xref ref-type="bibr" rid="CR45">45</xref>]. The cell cycle regulatory molecule cyclin D1 gene is one of the target genes for the Wnt/&#x003b2;-catenin signaling pathway, and VEGF is required for maintenance of adult lung alveolar structures. Any tissue repair involves coordinated cellular infiltration together with extracellular matrix deposition and re-epithelialization. Proteolytic degradation of the extracellular matrix requires MMPs which are regulated by Wnt signaling. It is uncertain why ARDS resolution involves fibrosis in some patients but not in others. Using western blot analysis of Wnt target gene products cyclin D1 and MMP7, K&#x000f6;nigshoff <italic>et al</italic>. [<xref ref-type="bibr" rid="CR16">16</xref>] demonstrated increased functional Wnt/&#x003b2;-catenin signaling in pulmonary fibrosis compared with patients without pulmonary fibrosis. Zuo <italic>et al</italic>. [<xref ref-type="bibr" rid="CR46">46</xref>] analyzed samples from patients with pulmonary fibrosis using microarray technology and found that <italic>Mmp7</italic> was the most upregulated gene, a finding that was confirmed by immunohistochemistry. The increased expression of cyclin D1, VEGF, and MMP7 in our study supports the importance of Wnt signaling in perpetuating lung inflammation and provides insights into the early development of a pro-fibrotic response during sepsis-induced ARDS. A greater understanding of modulators of WNT expression and the effects of WNT proteins in similar models will be paramount for clarifying the role of this pathway in lung inflammation and repair.</p><p>Our study does have some limitations. First, although the animal model used in the present investigation was CLP, we have examined autopsies from patients with different types of septic ARDS. However, there are no data suggesting that there is anything specific about pulmonary versus non-pulmonary insults in terms of different pulmonary fibrotic responses during severe sepsis. In an acid aspiration lung injury model, we found a similar fibrotic transformation as in our septic model [<xref ref-type="bibr" rid="CR26">26</xref>]. A recent study [<xref ref-type="bibr" rid="CR40">40</xref>] has shown that pulmonary fibrosis represents an early pathologic response in patients with ARDS, independent of the pulmonary or extrapulmonary nature of its cause. Second, we did not explore the effects of inhibitors of the Wnt pathway to irrefutably demonstrate that activation of Wnt pathway in the lung by a septic insult is responsible for the upregulation of downstream target genes (such as MMP7, cyclin D1, VEGF) that are involved in the pro-fibrotic transformation of injured tissues. However, studies by other investigators on selective inhibition of the Wnt/&#x003b2;-catenin signaling pathway [<xref ref-type="bibr" rid="CR44">44</xref>,<xref ref-type="bibr" rid="CR47">47</xref>,<xref ref-type="bibr" rid="CR48">48</xref>] have indicated that the WNT/&#x003b2;-catenin pathway is a target for anti-inflammatory and anti-fibrotic actions.</p></sec><sec id="Sec16" sec-type="conclusion"><title>Conclusion</title><p>In summary, our findings suggest that the WNT/&#x003b2;-catenin pathway may contribute to ongoing lung inflammation and lead to a pro-fibrotic response in the early stages of ARDS. We observed increased expression of WNT5A, cyclin D1, VEGF, and MMP7, all of which are <italic>Wnt</italic> target gene products that play an important role in pulmonary fibrosis. Further studies are needed to fully address unresolved questions regarding the modulation of the Wnt signaling pathway for attenuating lung inflammation and enhancing lung resolution and repair as a preventive or therapeutic approach in the setting of sepsis-induced ARDS.</p></sec><sec id="Sec17"><title>Key messages</title><p><list list-type="bullet"><list-item><p>The role of Wnt signaling is proving to be central to mechanisms of lung healing and fibrosis</p></list-item><list-item><p>Wnt/&#x003b2;-catenin pathway is activated in the lungs very early after sepsis and plays a role in initiating the lung repair process</p></list-item><list-item><p>Modulation of the Wnt/&#x003b2;-catenin pathway might represent a potential target for treatment in patients with sepsis and ARDS-induced pulmonary fibrosis</p></list-item></list></p></sec>

A comparison of principal component regression and genomic REML for genomic prediction across populations

Could not extract abstract

<contrib contrib-type="author"><name><surname>Dadousis</surname><given-names>Christos</given-names></name><address><email>christos.dadousis@studenti.unipd.it</email></address><xref ref-type="aff" rid="Aff1"/><xref ref-type="aff" rid="Aff2"/></contrib><contrib contrib-type="author"><name><surname>Veerkamp</surname><given-names>Roel F</given-names></name><address><email>roel.veerkamp@wur.nl</email></address><xref ref-type="aff" rid="Aff1"/><xref ref-type="aff" rid="Aff3"/></contrib><contrib contrib-type="author"><name><surname>Heringstad</surname><given-names>Bj&#x000f8;rg</given-names></name><address><email>bjorhe@umb.no</email></address><xref ref-type="aff" rid="Aff2"/><xref ref-type="aff" rid="Aff4"/></contrib><contrib contrib-type="author"><name><surname>Pszczola</surname><given-names>Marcin</given-names></name><address><email>mbee@jay.up.poznan.pl</email></address><xref ref-type="aff" rid="Aff1"/><xref ref-type="aff" rid="Aff3"/><xref ref-type="aff" rid="Aff5"/></contrib><contrib contrib-type="author" corresp="yes"><name><surname>Calus</surname><given-names>Mario PL</given-names></name><address><email>mario.calus@wur.nl</email></address><xref ref-type="aff" rid="Aff3"/></contrib><aff id="Aff1"><label/>Animal Breeding and Genomics Centre, Wageningen University, PO Box 338, Wageningen, 6700, AH The Netherlands </aff><aff id="Aff2"><label/>Department of Animal and Aquacultural Sciences, Norwegian University of Life Sciences, PO Box 5003, &#x000c5;s, N-1432 Norway </aff><aff id="Aff3"><label/>Animal Breeding and Genomics Centre, Wageningen UR Livestock Research, PO Box 338, Wageningen, 6700, AH The Netherlands </aff><aff id="Aff4"><label/>GENO Breeding and A. I. Association, PO Box 5003, &#x000c5;s, N-1432 Norway </aff><aff id="Aff5"><label/>Department of Genetics and Animal Breeding, Poznan University of Life Sciences, Wolynska 33, Poznan, 60-637 Poland </aff>

Genetics, Selection, Evolution : GSE

<sec id="Sec1" sec-type="introduction"><title>Background</title><p>For many years, dairy cattle breeding programs have been very successful in identifying the best animals via progeny-testing schemes. Progeny-testing was first implemented in Denmark and was soon used all over the world [<xref ref-type="bibr" rid="CR1">1</xref>]. One drawback of the progeny-testing scheme in dairy cattle breeding is the long generation intervals, which limits the response to selection, despite the high accuracy of selection achieved.</p><p>In order to reduce the generation interval by trying to obtain more accurate estimated breeding values (EBV) before progeny information is available, the use of molecular markers in connection with phenotypes to predict genetic merit has been investigated for some time [<xref ref-type="bibr" rid="CR2">2</xref>]. Recent advances in molecular techniques have made large-scale applications of such techniques possible. In 2001, Meuwissen et al. [<xref ref-type="bibr" rid="CR3">3</xref>] showed by simulation that genome-wide dense markers can adequately be used to estimate breeding values with a considerably high accuracy. Prediction of these EBV based on marker information is known as genomic prediction, and the subsequent selection step is known as genomic selection (GS). In GS, DNA information is used to predict the genetic merit of young animals, in order to reduce generation intervals. In recent years, GS has been implemented in dairy cattle breeding programs [<xref ref-type="bibr" rid="CR4">4</xref>&#x02013;<xref ref-type="bibr" rid="CR8">8</xref>] and has been described as the most promising molecular application in livestock [<xref ref-type="bibr" rid="CR9">9</xref>].</p><p>In practise, genomic prediction involves two steps. First, the effect of each SNP (single nucleotide polymorphism) is estimated in a reference population that consists of animals with both known phenotypes and marker genotypes. In the second step, genomic breeding values (GEBV) of young animals are estimated using only their marker information, to rank the animals for selection.</p><p>Despite the fact that several methods have been presented to estimate SNP effects, there are still many important questions and problems to be addressed, including statistical issues. These statistical issues concern mainly multicollinearity in the SNP dataset, due to linkage disequilibrium (LD) among markers, which leads to unstable estimates in least-squares regression. Moreover, a major problem in the statistical models used to estimate SNP effects is that the number of variables that needs to be estimated (<italic>p</italic>) is much larger than the number of observations (<italic>n</italic>), thereby removing least squares from possible analysis methods. In the field of statistics, these problems are frequently overcome by using principal component analysis (PCA) and subsequent regression on the principle components (PC) (PCR; principal component regression) instead of on the original variables.</p><p>In general, PCA can be used to solve multicollinearity problems among predictor variables and to reduce the dimensional space. In genetic studies, PCA has been used mainly for population studies and has been a powerful tool to identify population structures and migration patterns, and to correct for stratification in association studies by capturing genetic variation [<xref ref-type="bibr" rid="CR10">10</xref>&#x02013;<xref ref-type="bibr" rid="CR15">15</xref>]. One of the first applications of PCA in population genetics was by Menozzi et al. [<xref ref-type="bibr" rid="CR16">16</xref>] to produce maps of human genetic variation across mainland regions.</p><p>Likewise, in animal breeding, PCA has recently been used to infer population clusters from different breeds [<xref ref-type="bibr" rid="CR17">17</xref>] and to represent genotypes in the prediction of GEBV [<xref ref-type="bibr" rid="CR18">18</xref>&#x02013;<xref ref-type="bibr" rid="CR21">21</xref>]. Daetwyler et al. [<xref ref-type="bibr" rid="CR22">22</xref>] used PCA to investigate the impact of population structure on the accuracy of GEBV in a multi-breed sheep population. Results of these studies, which used either simulations or real data, describe PCA as a promising method for animal breeding to produce accurate GEBV. In these studies, the main benefits of using PCA were a significant reduction in data quantity (&#x0003e;90%) and fast computation. However, to date, there is only a limited number of studies based on real data that compare PCR for genomic prediction with a more commonly used genomic prediction model such as GBLUP (best linear unbiased prediction, in which the pedigree additive relationship matrix is replaced with a marker-derived relationship matrix) [<xref ref-type="bibr" rid="CR23">23</xref>]. Since PCA is able to recover population structure, it may be expected that using this information is beneficial for genomic prediction applied to data with strong population structure. One such application is across-population genomic prediction, e.g. genomic prediction based on reference data that only includes data from other populations and not from the predicted population itself. Whether the ability of PCA to detect population structure is also beneficial in applications of across-population genomic prediction is currently unknown.</p><p>The main objective of this research was to investigate the potential of PCR for across-population genomic prediction, as applied to yield traits in Holstein cows from different countries. More precisely, the objectives were (i) to compare the predictive accuracy of PCR with a REML model that uses a genomic relationship matrix (GREML) and (ii) to investigate the effect of alternative methods of extracting and selecting PC on the accuracy of genomic predictions.</p></sec><sec id="Sec2" sec-type="materials|methods"><title>Methods</title><sec id="Sec3"><title>Data</title><p>We used 66 116 daily records up to 45&#x000a0;weeks in lactation for milk, fat and protein yields from 1609 first lactation Holstein heifers. Heifers originated from four countries, Ireland (IRL; Teagasc, Moorepark Dairy Production), United Kingdom (UK; Scottish Agricultural College), the Netherlands (NLD; Wageningen UR Livestock Research) and Sweden (SWE; Swedish University of Agricultural Science). The UK data included animals from two divergent selection lines, a line selected for high fat and protein yield and a control line that represents the UK national average for fat and protein yield [<xref ref-type="bibr" rid="CR24">24</xref>]. These two lines were therefore considered as two groups (UK_1 and UK_2). All phenotypes were pre-adjusted to account for the mean overall lactation curve, herd, diet group, milking frequency, year-month of milk test-day by management group, and experimental treatments. For a full description, see [<xref ref-type="bibr" rid="CR24">24</xref>,<xref ref-type="bibr" rid="CR25">25</xref>]. For each animal, a single pre-adjusted phenotype was obtained as the average daily milk, fat and protein yields for lactation weeks 3 to 15, derived from individually predicted lactation curves. Descriptive statistics of the pre-adjusted phenotypes are in Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>.<table-wrap id="Tab1"><label>Table 1</label><caption><p> <bold>Descriptive statistics of pre-adjusted average daily data of the milk yield traits</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th> <bold>Trait</bold> </th><th> <bold>Mean</bold> </th><th> <bold>SD</bold> </th><th> <bold>SE</bold> </th><th> <bold>Min</bold> </th><th> <bold>Max</bold> </th><th> <bold>n</bold> </th></tr></thead><tbody><tr valign="top"><td>Milk yield (kg)</td><td>23.84</td><td>4.44</td><td>0.111</td><td>0.99</td><td>38.98</td><td>1609</td></tr><tr valign="top"><td>Fat yield (kg)</td><td>0.93</td><td>0.18</td><td>0.004</td><td>0.12</td><td>1.79</td><td>1609</td></tr><tr valign="top"><td>Protein yield (kg)</td><td>0.72</td><td>0.13</td><td>0.003</td><td>0.04</td><td>1.34</td><td>1609</td></tr></tbody></table></table-wrap></p><p>All animals were genotyped within the RobustMilk project (<ext-link ext-link-type="uri" xlink:href="http://www.robustmilk.eu">www.robustmilk.eu</ext-link>) with the Illumina BovineSNP50 Beadchip (Illumina Inc., San Diego, CA) containing 54 001 SNPs. Quality control checks on the SNPs used the following criteria: (1) a GenCall score greater than 0.20 and a GenTrain score greater than 0.55 for individual genotypes; (2) a call rate greater than 95%; (3) a minor allele frequency greater than 0.01 in each country; and (4) no extreme deviation from Hardy Weinberg Equilibrium (<italic>&#x003c7;</italic>²&#x02009;&#x0003c;&#x02009;600). After editing, 37 069 SNPs remained across the 29 autosomes and the X-chromosome.</p></sec><sec id="Sec4"><title>Reference and test datasets</title><p>The across-country dataset was split into five subsets and, in each analysis, four subsets were used as the reference set and the other one for testing. The first three test datasets included animals from only one country (Ireland, the Netherlands or Sweden), while the last two each contained one of the UK selection lines, such that each animal had its genomic breeding value predicted once for each trait and model. The number of cows in each subset ranged from 181 to 618 (Table&#x000a0;<xref rid="Tab2" ref-type="table">2</xref>). Accuracies of predicted genomic breeding values were calculated as Pearson correlations between the predicted genomic breeding values and the adjusted phenotypes within each test dataset (i.e. within country, and within line for the UK animals).<table-wrap id="Tab2"><label>Table 2</label><caption><p> <bold>Number of cows with phenotype records and genotypes from Ireland, the Netherlands, Sweden and two divergent selection lines from UK</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th> <bold>Population</bold> </th><th> <bold>Number of animals</bold> </th></tr></thead><tbody><tr valign="top"><td>UK_1</td><td>206</td></tr><tr valign="top"><td>UK_2</td><td>210</td></tr><tr valign="top"><td>Sweden</td><td>181</td></tr><tr valign="top"><td>Ireland</td><td>394</td></tr><tr valign="top"><td>The Netherlands</td><td>618</td></tr><tr valign="top"><td> <bold>Total</bold> </td><td> <bold>1609</bold> </td></tr></tbody></table></table-wrap></p></sec><sec id="Sec5"><title>Principal component analysis</title><p>Assume a matrix <bold>X</bold> of order <italic>(n</italic>&#x02009; <bold><italic>&#x000d7;</italic></bold>&#x02009; <italic>p)</italic> where <italic>n</italic> individuals have been genotyped for <italic>p</italic> SNPs. The elements of this matrix may be 0, 1 or 2, representing the genotype of each individual for each SNP (0 and 2 for homozygotes and 1 for heterozygotes). The main idea of PCA is to reveal hidden structure in the data, to reduce the number of variables in the dataset, and to solve the multicollinearity problem (high correlation between columns in <bold>X</bold>). It extracts the most important information, in terms of variation, and re-expresses the original dataset in a simplified way. Thus, PCA aims at finding a small set <italic>k (k&#x02009;&#x0003c;&#x02009;p)</italic> of PC that explain as much of the variability in <bold>X</bold> as possible. This is achieved through an orthogonal transformation of the original dataset such that as much of the original variability as possible is included in the first few PC. So, PC are linear combinations of a set of random variables in <bold>X</bold>, i.e. the matrix <bold>T</bold> with PC is obtained by:<disp-formula id="Equa"><alternatives><tex-math id="M1">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document} $$ \mathbf{T}={\mathbf{e}}^{\mathbf{\prime}}\mathbf{X}, $$ \end{document}</tex-math><graphic xlink:href="12711_2014_60_Article_Equa.gif" position="anchor"/></alternatives></disp-formula>where <bold>e</bold> represents the eigenvectors derived from spectral decomposition of the covariance (or correlation) matrix of <bold>X</bold>. In genomic data, the covariance (correlation matrix) of the SNP genotypes (of order <italic>p &#x000d7; p</italic>) can be used or alternatively the similarity matrix of the individuals (<bold>G</bold> matrix, of order <italic>n &#x000d7; n</italic>). The first PC is then defined as the vector:<disp-formula id="Equb"><alternatives><tex-math id="M2">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document} $$ {\mathbf{T}}_1={\mathbf{e}}_1^{\mathbf{\prime}}\mathbf{X}={e}_{11}{\mathbf{X}}_1+{e}_{12}{\mathbf{X}}_2+\cdots +{e}_{1p}{\mathbf{X}}_p, $$ \end{document}</tex-math><graphic xlink:href="12711_2014_60_Article_Equb.gif" position="anchor"/></alternatives></disp-formula>which captures the maximum variance in <bold>X</bold>, with the constraint that <bold>e</bold>&#x02032;<bold>e</bold>&#x02009;=&#x02009;1. For all PC combinations, it holds that: <italic>cov</italic>(<bold>T</bold>_i,&#x02009;<bold>T</bold>_j&#x02009;=&#x02009;0) for all <italic>i &#x02260; j (i,j&#x02009;=&#x02009;1,2,&#x02026;,p).</italic></p><p>The basis of PCA is either the spectral decomposition of the covariance (correlation) matrix of <bold>X</bold> or the singular value decomposition (SVD) of <bold>X</bold>. The SVD represents a more general view of the eigenvalue decomposition for non-square matrices <bold>X</bold>. In general, PCA based on SVD and eigen decomposition are expected to yield similar results if <bold>X</bold> is square and symmetric [<xref ref-type="bibr" rid="CR26">26</xref>]. Moreover, SVD on an <italic>n</italic>&#x02009; <bold><italic>&#x000d7;</italic></bold>&#x02009; <italic>p</italic> matrix <bold>X</bold> is expected to yield the same results as on its <italic>p</italic>&#x02009; <bold><italic>&#x000d7;</italic></bold>&#x02009; <italic>p</italic> correlation matrix.</p></sec><sec id="Sec6"><title>Principal component regression and genomic prediction</title><p>The concept of PCR, i.e. the use of PC in regression has been around for quite some time in the field of statistics [<xref ref-type="bibr" rid="CR27">27</xref>,<xref ref-type="bibr" rid="CR28">28</xref>]. For application in genomic prediction, first consider the general model to predict breeding values based on marker genotypes:<disp-formula id="Equc"><alternatives><tex-math id="M3">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document} $$ \mathbf{y}=\mathbf{1}\mu +\mathbf{X}\mathbf{b}+\mathbf{e}, $$ \end{document}</tex-math><graphic xlink:href="12711_2014_60_Article_Equc.gif" position="anchor"/></alternatives></disp-formula>where values in <bold>e</bold> are iid <inline-formula id="IEq1"><alternatives><tex-math id="M4">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document} $$ \sim N\left(\mathbf{0},\mathbf{I}{\upsigma}_{\mathrm{e}}^2\right), $$ \end{document}</tex-math><inline-graphic xlink:href="12711_2014_60_Article_IEq1.gif"/></alternatives></inline-formula><bold>y</bold> is a vector of phenotypic records, <bold>1</bold> is a vector of ones, <italic>&#x003bc;</italic> is the overall mean, <bold>X</bold> is a matrix (centred and possibly scaled) containing SNP genotypes, <bold>b</bold> is a vector of additive effects of all SNPs, <bold>e</bold> is a vector of residual effects, <bold>I</bold> is the identity matrix and <inline-formula id="IEq2"><alternatives><tex-math id="M5">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document} $$ {\upsigma}_{\mathrm{e}}^2 $$ \end{document}</tex-math><inline-graphic xlink:href="12711_2014_60_Article_IEq2.gif"/></alternatives></inline-formula> is the residual variance. The initial step for a PCR model is to perform PCA on the genotype matrix <bold>X</bold><italic>(n&#x02009;&#x000d7;&#x02009;p).</italic> For this purpose, we used SVD via the function &#x0201c;prcomp&#x0201d; in R [<xref ref-type="bibr" rid="CR29">29</xref>], which works as follows. Consider the SVD of <bold>X</bold>, <bold>X</bold>&#x02009;=&#x02009;<bold>U&#x003a3;V</bold>&#x02032;, where <bold>U</bold> and <bold>V</bold> are the left and right singular vectors of <bold>X</bold>, <bold>V</bold>&#x02032; is the transpose of <bold>V</bold> and <bold>&#x003a3;</bold> is a diagonal matrix containing the singular values. The matrix <bold>T</bold> (<italic>n&#x02009;&#x000d7;&#x02009;k</italic>) of PC scores is then calculated as:<disp-formula id="Equd"><alternatives><tex-math id="M6">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document} $$ \mathbf{T}=\mathbf{X}\mathbf{V}=\mathbf{U}\Sigma {\mathbf{V}}^{\mathbf{\prime}}\mathbf{V}=\mathbf{U}\Sigma, $$ \end{document}</tex-math><graphic xlink:href="12711_2014_60_Article_Equd.gif" position="anchor"/></alternatives></disp-formula></p><p>where <italic>k&#x02009;&#x0003c;&#x02009;r</italic>, where <italic>r</italic> is the rank of <bold>X</bold>, and <bold>V</bold><italic>(p&#x02009;&#x000d7;&#x02009;k)</italic> is the loading matrix derived from the SVD of <bold>X</bold>, which defines weights to the original <bold>X</bold> variables in each PC.</p></sec><sec id="Sec7"><title>PCA based on the reference dataset</title><p>Principal component analysis was initially performed only on the SNP matrix of the reference dataset, where the <bold>T</bold> matrix of PC was calculated as <bold>T</bold>_r&#x02009; <bold>=&#x02009;X</bold>_r<bold>V</bold>, where r denotes the reference dataset. The <bold>V</bold> matrix that was extracted from the reference dataset was also used to construct the <bold>T</bold> matrix for the test dataset as <bold>T</bold>_t&#x02009; <bold>=&#x02009;X</bold>_t<bold>V</bold>, where <bold>X</bold>_t contains the genotypes of the test dataset.</p><p>Following from the above, the PCR model that was applied is:<disp-formula id="Eque"><alternatives><tex-math id="M7">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document} $$ \mathbf{y}=\mathbf{1}\mu +\mathbf{Tg}+\mathbf{e}, $$ \end{document}</tex-math><graphic xlink:href="12711_2014_60_Article_Eque.gif" position="anchor"/></alternatives></disp-formula>where <bold>T</bold> is the matrix of PC, and <bold>g</bold> is a vector with regression coefficients for each PC in <bold>T</bold>. In this case, the derived transformed SNP effects (i.e. the values in <bold>g)</bold> are treated as fixed effects in contrast to what is commonly used in genomic prediction models that perform simultaneous regression on SNP genotypes treating SNP effects as random [<xref ref-type="bibr" rid="CR30">30</xref>].</p></sec><sec id="Sec8"><title>PCA based on all animals</title><p>In a second approach, PCA was performed on the matrix with all SNP genotypes, where genotypes of all reference and test datasets were included. For application in the PCR model, the <bold>T</bold> matrix was split in parts relating to the reference and test datasets (<bold>T</bold>_{<bold>r</bold>} and <bold>T</bold>_{<bold>t</bold>}, respectively), using methods that are briefly described in Additional file <xref rid="MOESM1" ref-type="media">1</xref>: Figure S1. In this approach, hereinafter referred to as semi-supervised PCR (SSPCR), the genotypic information of the individuals to be predicted is partly included in the training dataset of the prediction model. This is because the axes of variation, i.e. the singular vectors and the singular values of the SVD were extracted using all genomic information available in the dataset. This concept of semi-supervised PCA was borrowed from computer science and face recognition analysis [<xref ref-type="bibr" rid="CR31">31</xref>,<xref ref-type="bibr" rid="CR32">32</xref>].</p></sec><sec id="Sec9"><title>Selection of PC for inclusion in the PCR models</title><p>Two methods were tested to select sets of PC to be used in the subsequent PCR models. In the first method, PC were ranked based on decreasing eigenvalues (variation in the explanatory variables, i.e. the genotypes), which will be referred to as PCR_eigen. In the second method, the PC were ranked based on their contribution to the sum of squares (ss) of the regression (variation in the response variable), referred to as PCR_ss. These contributions were obtained from a PCR model for which only phenotypes and genotypes of the animals of the reference dataset were included.</p></sec><sec id="Sec10"><title>Selection of the optimal model in PCR</title><p>Once the order in which the PC should be added to the model is established, the question is how many PC should be used in the subsequent PCR model used for genomic prediction. There is no general consensus on which strategy should be followed for this. Inspecting plots of eigenvalues (via the so-called &#x0201c;scree plots&#x0201d; that plot the PC ranked based on decreasing eigenvalues) or keeping the number of PC that capture a given percentage of the original variation are two among a variety of methods (see [<xref ref-type="bibr" rid="CR33">33</xref>] for a detailed review). In our analyses, a cross-validation (CV) approach within the reference dataset (as in [<xref ref-type="bibr" rid="CR33">33</xref>]) was chosen in order to obtain the &#x0201c;optimum&#x0201d; number of PC to include in the PCR, which will be further used in the section on prediction of the test dataset. For CV, the reference dataset was either split by country (and line in the case of UK), which will be referred to as stratified CV hereafter, or split randomly in a 5-fold CV. In both these CV approaches, all PC were added in the PCR model one by one and the minimum mean squared error (MSE) of the predictions within the reference dataset was used as the target function to be optimized, which is briefly described in Additional file <xref rid="MOESM2" ref-type="media">2</xref>: Figure S2. Both CV approaches were performed using the R package &#x0201c;plsdof&#x0201d; [<xref ref-type="bibr" rid="CR34">34</xref>], with the appropriate modification for the semi-supervised PCA.</p></sec><sec id="Sec11"><title>GREML model</title><p>For the GREML model, the following individual animal model with a genomic relationship matrix was fitted in ASReml-R [<xref ref-type="bibr" rid="CR35">35</xref>]:<disp-formula id="Equf"><alternatives><tex-math id="M8">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document} $$ \mathbf{y}=\mathbf{1}\mu +\mathbf{W}\mathbf{u}+\mathbf{e}, $$ \end{document}</tex-math><graphic xlink:href="12711_2014_60_Article_Equf.gif" position="anchor"/></alternatives></disp-formula>where <bold>u</bold> is a vector of additive genetic effects for any of the considered traits and <bold>W</bold> is the design matrix that links <bold>u</bold> to the phenotypic records in <bold>y</bold>. For additive genetic and residual effects, the following normal distributions were assumed: <inline-formula id="IEq3"><alternatives><tex-math id="M9">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document} $$ \mathbf{u}\sim N\left(\mathbf{0},\mathbf{G}{\upsigma}_{\mathrm{u}}^2\right) $$ \end{document}</tex-math><inline-graphic xlink:href="12711_2014_60_Article_IEq3.gif"/></alternatives></inline-formula> and <inline-formula id="IEq4"><alternatives><tex-math id="M10">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document} $$ \mathbf{e}\sim N\left(\mathbf{0},\mathbf{I}{\upsigma}_{\mathrm{e}}^2\right). $$ \end{document}</tex-math><inline-graphic xlink:href="12711_2014_60_Article_IEq4.gif"/></alternatives></inline-formula> Note that this is a genomic BLUP model but with estimation of variances <inline-formula id="IEq5"><alternatives><tex-math id="M11">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document} $$ {\upsigma}_{\mathrm{u}}^2 $$ \end{document}</tex-math><inline-graphic xlink:href="12711_2014_60_Article_IEq5.gif"/></alternatives></inline-formula> and <inline-formula id="IEq6"><alternatives><tex-math id="M12">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document} $$ {\upsigma}_{\mathrm{e}}^2 $$ \end{document}</tex-math><inline-graphic xlink:href="12711_2014_60_Article_IEq6.gif"/></alternatives></inline-formula> together with estimation of the breeding values using restricted maximum likelihood (REML). Therefore, this model is more appropriately referred to as genomic REML (GREML) [<xref ref-type="bibr" rid="CR36">36</xref>]. The genomic relationship matrix was calculated following VanRaden [<xref ref-type="bibr" rid="CR23">23</xref>] as:<disp-formula id="Equg"><alternatives><tex-math id="M13">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document} $$ {\mathbf{G}}_{\mathbf{VR}}=\frac{\mathbf{Z}{\mathbf{Z}}^{\mathbf{\hbox{'}}}}{2{\displaystyle \sum {p}_i\left(1\hbox{-} {p}_i\right)}}, $$ \end{document}</tex-math><graphic xlink:href="12711_2014_60_Article_Equg.gif" position="anchor"/></alternatives></disp-formula>where <italic>p</italic>_{<italic>i</italic>} is the frequency at SNP <italic>i</italic> for which the homozygous genotype is coded 2, calculated across all genotyped animals, and <bold>Z</bold> is derived from genotypes of animals in the reference dataset by subtracting 2<italic>p</italic>_{<italic>i</italic>} from a matrix <bold>X</bold> that specifies the marker genotypes for each individual as 0, 1 or 2. Following Yang et al. [<xref ref-type="bibr" rid="CR37">37</xref>], <bold>G</bold>_{<bold>VR</bold>} was regressed back towards <bold>A</bold> (the pedigree relationship matrix) to account for errors in the estimation of <bold>G</bold>_{<bold>VR</bold>}, resulting in the computation of the genomic relationship matrix <bold>G</bold> as:<disp-formula id="Equh"><alternatives><tex-math id="M14">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document} $$ \mathbf{G}=b\times {\mathbf{G}}_{\mathbf{VR}}+\left(1-b\right)\times \mathbf{A}, $$ \end{document}</tex-math><graphic xlink:href="12711_2014_60_Article_Equh.gif" position="anchor"/></alternatives></disp-formula>where <italic>b</italic> is estimated according to Yang et al. [<xref ref-type="bibr" rid="CR37">37</xref>]. The value of <inline-formula id="IEq7"><alternatives><tex-math id="M15">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document} $$ \widehat{b} $$ \end{document}</tex-math><inline-graphic xlink:href="12711_2014_60_Article_IEq7.gif"/></alternatives></inline-formula> ranged from 0.975 to 0.997 across different bins of pedigree-based relationships. So, although in theory this adjustment of <bold>G</bold>_{<bold>VR</bold>} improves the properties of the <bold>G</bold> matrix [<xref ref-type="bibr" rid="CR37">37</xref>], in our case the adjusted <bold>G</bold> matrix was very close to the original matrix and is therefore expected to yield very similar predictions.</p></sec></sec><sec id="Sec12" sec-type="results"><title>Results</title><sec id="Sec13"><title>Characterization of the data</title><p>PCA was performed on all SNP genotypes to investigate differences in genotypes between the Holstein populations included in this study. Based on the plot of the 1st against the 2nd PC (Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>), one of the selection lines of the UK population could be distinguished from the rest with the first PC. However, it should be noted that the 1st and 2nd PC captured only 1.5% and 1.4% of the total original variability of the SNP data, respectively (Table&#x000a0;<xref rid="Tab3" ref-type="table">3</xref>). Comparison of relationships based either on pedigree or genomic relationships also confirmed that the UK_1 population had the weakest average relationship with the other populations (Table&#x000a0;<xref rid="Tab4" ref-type="table">4</xref>). In nearly all cases, standard deviations of the genomic relationships were higher than those of pedigree-based relationships, which indicates that the use of SNP information explains more variation in relationships than pedigree information. Averages and standard deviations of relationships were always higher within populations than between populations. This confirms that relationships among the five populations were low and that genomic predictions in these data indeed were &#x0201c;across populations&#x0201d;, in the sense that the reference data always included data only from other populations and not from the predicted population itself.<fig id="Fig1"><label>Figure 1</label><caption><p> <bold>Scatterplot of the first two principal components (PC1 vs. PC2).</bold> Principal component analysis performed on the whole dataset, with data from Ireland (IRL), the Netherlands (NLD), Sweden (SWE) and two divergent selection lines from United Kingdom (UK_1 and UK_2).</p></caption><graphic xlink:href="12711_2014_60_Fig1_HTML" id="MO1"/></fig><table-wrap id="Tab3"><label>Table 3</label><caption><p> <bold>Cumulative proportion of the original variability captured by principal components (PC)</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th> <bold>Number of PC</bold> </th><th> <bold>Cumulative proportion (%)</bold> </th></tr></thead><tbody><tr valign="top"><td>1</td><td>1.5</td></tr><tr valign="top"><td>2</td><td>2.9</td></tr><tr valign="top"><td>37</td><td>20</td></tr><tr valign="top"><td>138</td><td>40</td></tr><tr valign="top"><td>326</td><td>60</td></tr><tr valign="top"><td>668</td><td>80</td></tr><tr valign="top"><td>967</td><td>90</td></tr></tbody></table><table-wrap-foot><p>Principal component analysis was performed on the whole dataset.</p></table-wrap-foot></table-wrap><table-wrap id="Tab4"><label>Table 4</label><caption><p> <bold>Average (Av) and SD of pedigree and genomic relationships within and between countries (and selection lines)</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th/><th/><th/><th> <bold>UK_1</bold> </th><th> <bold>UK_2</bold> </th><th> <bold>Sweden</bold> </th><th> <bold>Ireland</bold> </th><th> <bold>NL</bold> </th></tr></thead><tbody><tr valign="top"><td>Av</td><td>Within¹ </td><td>Pedigree</td><td>0.0612</td><td>0.0763</td><td>0.0534</td><td>0.0556</td><td>0.0792</td></tr><tr valign="top"><td/><td/><td>Genomic</td><td>0.0560</td><td>0.0226</td><td>0.0268</td><td>0.0182</td><td>0.0096</td></tr><tr valign="top"><td/><td>Between¹ </td><td>Pedigree</td><td>0.0295</td><td>0.0490</td><td>0.0385</td><td>0.0424</td><td>0.0499</td></tr><tr valign="top"><td/><td/><td>Genomic</td><td>&#x02212;0.0088</td><td>&#x02212;0.0026</td><td>&#x02212;0.0024</td><td>&#x02212;0.0038</td><td>&#x02212;0.0028</td></tr><tr valign="top"><td>SD</td><td>Within</td><td>Pedigree</td><td>0.0894</td><td>0.0819</td><td>0.0919</td><td>0.0767</td><td>0.0645</td></tr><tr valign="top"><td/><td/><td>Genomic</td><td>0.0958</td><td>0.0860</td><td>0.0953</td><td>0.0788</td><td>0.0654</td></tr><tr valign="top"><td/><td>Between</td><td>Pedigree</td><td>0.0220</td><td>0.0322</td><td>0.0284</td><td>0.0305</td><td>0.0332</td></tr><tr valign="top"><td/><td/><td>Genomic</td><td>0.0265</td><td>0.0331</td><td>0.0283</td><td>0.0311</td><td>0.0342</td></tr></tbody></table><table-wrap-foot><p> ¹Averages (Av) and SD are computed for relationships within each population (Within) and for relationships between each population and all other populations (Between).</p></table-wrap-foot></table-wrap></p></sec><sec id="Sec14"><title>GREML versus optimal PCR and SSPCR</title><p>Accuracies of genomic predictions obtained using GREML, PCR and SSPCR models, after determining the optimal number of PC based on the reference data, are in Table&#x000a0;<xref rid="Tab5" ref-type="table">5</xref>. On average, across test datasets and traits, GREML outperformed the PCR and SSPCR models. Across the test datasets, the performance of the PCR and SSPCR models was closest to that of GREML for milk with PCR_eigen and 5-fold CV (0.15 vs. 0.14), and for fat with SSPCR_ss and 5-fold CV (0.07 for both models). For protein, the maximum accuracy obtained with PCR and SSPCR was 0.02 achieved in four cases (PCR_eigen and 5-fold CV, PCR_ss and stratified CV, SSPCR_ss and 5-fold CV, and SSPCR_ss and stratified CV) versus 0.05 for GREML.<table-wrap id="Tab5"><label>Table 5</label><caption><p> <bold>Accuracies</bold> ^{<bold>1</bold>} <bold>obtained for the PCR</bold> ^{<bold>2</bold>} <bold>and GREML</bold> ^{<bold>3</bold>} <bold> models</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th> <bold>Test</bold> </th><th> <bold>Trait</bold> </th><th> <bold>GREML</bold> </th><th colspan="2"> <bold>PCR_eigen</bold> </th><th colspan="2"> <bold>PCR_ss</bold> </th><th colspan="2"> <bold>SSPCR_eigen</bold> </th><th colspan="2"> <bold>SSPCR_ss</bold> </th></tr><tr valign="top"><th/><th/><th/><th> <bold>5-fold</bold> </th><th> <bold>stratified</bold> </th><th> <bold>5-fold</bold> </th><th> <bold>stratified</bold> </th><th> <bold>5-fold</bold> </th><th> <bold>stratified</bold> </th><th> <bold>5-fold</bold> </th><th> <bold>stratified</bold> </th></tr></thead><tbody><tr valign="top"><td>UK_1</td><td>Milk</td><td> <italic>0.26*</italic> </td><td>0.21 (249)</td><td>0.16 (204)</td><td>0.17 (142)</td><td>NA⁴ (0)</td><td>0.10 (71)</td><td>NA (0)</td><td>0.13 (821)</td><td>0.22 (537)</td></tr><tr valign="top"><td/><td>Fat</td><td>0.07</td><td>&#x02212;0.08 (35)</td><td>&#x02212;0.01 (103)</td><td>0.02 (95)</td><td>0.11 (1)</td><td>&#x02212;0.03 (40)</td><td>NA (0)</td><td>0.07 (859)</td><td> <italic>0.12 (614)</italic> </td></tr><tr valign="top"><td/><td>Protein</td><td>&#x02212;0.01</td><td>&#x02212;0.05 (121)</td><td>&#x02212;0.09 (69)</td><td>&#x02212;0.02 (91)</td><td>NA (0)</td><td>&#x02212;0.13 (91)</td><td>NA (0)</td><td>0.03 (910)</td><td> <italic>0.05 (570)</italic> </td></tr><tr valign="top"><td>UK_2</td><td>Milk</td><td>0.10</td><td>0.12 (233)</td><td>0.12 (217)</td><td>0.05 (342)</td><td> <italic>0.14 (13)</italic> </td><td>0.12 (89)</td><td>0.13 (10)</td><td>0.04 (941)</td><td>0.03 (527)</td></tr><tr valign="top"><td/><td>Fat</td><td>&#x02212;0.02</td><td>&#x02212;0.13 (96)</td><td>&#x02212;0.12 (37)</td><td>&#x02212;0.11 (35)</td><td>&#x02212;0.09 (28)</td><td>&#x02212;0.08 (33)</td><td>&#x02212;0.08 (27)</td><td>0.03 (909)</td><td> <italic>0.05 (523)</italic> </td></tr><tr valign="top"><td/><td>Protein</td><td>0.01</td><td>&#x02212;0.03 (89)</td><td>&#x02212;0.06 (46)</td><td>&#x02212;0.11 (91)</td><td> <italic>0.06 (7)</italic> </td><td>0.04 (83)</td><td>0.02 (10)</td><td>&#x02212;0.06 (780)</td><td>&#x02212;0.01 (505)</td></tr><tr valign="top"><td>SWE</td><td>Milk</td><td> <italic>0.16</italic> </td><td>0.15 (181)</td><td>0.14 (187)</td><td>0.12 (184)</td><td>0.04 (19)</td><td>0.03 (64)</td><td>0.04 (15)</td><td>0.09 (895)</td><td>0.07 (529)</td></tr><tr valign="top"><td/><td>Fat</td><td>0.09</td><td>0.10 (162)</td><td>0.10 (146)</td><td>0.04 (101)</td><td>0.07 (713)</td><td>0.07 (57)</td><td>&#x02212;0.02 (1)</td><td> <italic>0.16 (859)</italic> </td><td>0.04 (680)</td></tr><tr valign="top"><td/><td>Protein</td><td>0.06</td><td>0.00 (92)</td><td>&#x02212;0.04 (48)</td><td>0.01 (82)</td><td>&#x02212;0.06 (29)</td><td>&#x02212;0.10 (17)</td><td>&#x02212;0.10 (16)</td><td> <italic>0.10 (1025)</italic> </td><td> <italic>0.10 (574)</italic> </td></tr><tr valign="top"><td>IRL</td><td>Milk</td><td>0.06</td><td>0.05 (277)</td><td>&#x02212;0.10 (48)</td><td>&#x02212;0.03 (207)</td><td>&#x02212;0.08 (37)</td><td>&#x02212;0.04 (35)</td><td>&#x02212;0.13 (8)</td><td> <italic>0.14 (717)</italic> </td><td>0.13 (389)</td></tr><tr valign="top"><td/><td>Fat</td><td> <italic>0.08</italic> </td><td>0.06 (121)</td><td>0.04 (51)</td><td>0.02 (100)</td><td> <italic>0.08 (52)</italic> </td><td>0.07 (37)</td><td>0.02 (14)</td><td>0.07 (712)</td><td>0.06 (315)</td></tr><tr valign="top"><td/><td>Protein</td><td>0.04</td><td>0.05 (127)</td><td>0.04 (145)</td><td>0.00 (35)</td><td>&#x02212;0.04 (173)</td><td>&#x02212;0.05 (37)</td><td>&#x02212;0.12 (12)</td><td> <italic>0.12 (759)</italic> </td><td>0.11 (421)</td></tr><tr valign="top"><td>NLD</td><td>Milk</td><td>0.16</td><td>0.18 (50)</td><td> <italic>0.19 (28)</italic> </td><td>0.09 (55)</td><td>0.10 (7)</td><td>0.18 (25)</td><td>0.13 (2)</td><td>0.11 (95)</td><td>0.11 (95)</td></tr><tr valign="top"><td/><td>Fat</td><td> <italic>0.15</italic> </td><td>0.11 (47)</td><td>0.10 (99)</td><td>0.05 (75)</td><td>0.06 (7)</td><td>0.07 (39)</td><td>0.07 (7)</td><td>0.10 (197)</td><td>0.11 (196)</td></tr><tr valign="top"><td/><td>Protein</td><td>0.13</td><td>0.14 (31)</td><td>0.13 (34)</td><td>0.07 (42)</td><td>0.13 (7)</td><td>0.12 (28)</td><td> <italic>0.16 (2)</italic> </td><td>0.11 (48)</td><td>0.11 (48)</td></tr><tr valign="top"><td>Average</td><td>Milk</td><td> <italic>0.15</italic> </td><td>0.14 (198)</td><td>0.10 (137)</td><td>0.08 (186)</td><td>0.05 (15)</td><td>0.08 (57)</td><td>0.04 (7)</td><td>0.09 (694)</td><td>0.02 (415)</td></tr><tr valign="top"><td/><td>Fat</td><td> <italic>0.07</italic> </td><td>0.01 (92)</td><td>0.02 (87)</td><td>0.00 (81)</td><td>0.05 (160)</td><td>0.02 (41)</td><td>0.00 (10)</td><td> <italic>0.07 (707)</italic> </td><td>0.04 (466)</td></tr><tr valign="top"><td/><td>Protein</td><td> <italic>0.05</italic> </td><td>0.02 (92)</td><td>0.00 (68)</td><td>&#x02212;0.01 (68)</td><td>0.02 (43)</td><td>&#x02212;0.02 (51)</td><td>&#x02212;0.01 (8)</td><td>0.02 (704)</td><td>0.02 (424)</td></tr><tr valign="top"><td>SD</td><td>Milk</td><td>0.08</td><td>0.06 (90)</td><td>0.12 (91)</td><td>0.08 (105)</td><td>0.10 (14)</td><td>0.08 (26)</td><td>0.12 (6)</td><td>0.04 (345)</td><td>0.07 (189)</td></tr><tr valign="top"><td/><td>Fat</td><td>0.06</td><td>0.11 (53)</td><td>0.09 (44)</td><td>0.07 (28)</td><td>0.08 (310)</td><td>0.07 (9)</td><td>0.06 (11)</td><td>0.05 (295)</td><td>0.04 (204)</td></tr><tr valign="top"><td/><td>Protein</td><td>0.05</td><td>0.08 (38)</td><td>0.09 (45)</td><td>0.07 (27)</td><td>0.09 (73)</td><td>0.10 (34)</td><td>0.13 (7)</td><td>0.08 (382)</td><td>0.05 (219)</td></tr></tbody></table><table-wrap-foot><p>For the PCR models, the PC included (numbers are presented in brackets) were selected based on cross-validation in the reference population data, either in a 5-fold random or stratified split to select the optimum PCR model in respect to minimum mean squared error. Analyses were performed for three traits and five test populations.</p><p> ¹Accuracies were calculated as Pearson correlations between the predicted genomic breeding values and the adjusted phenotypes; ²selection of PCs was based either on the eigenvalues (eigen) or the regression sum of squares (ss); two different methods of applying principal component analysis, either separately for reference and test parts (PCR) or on the whole dataset (SSPCR), were compared; ³a REML based model with a genomic relationship matrix; ⁴all animals received the same prediction; *in italics the highest accuracies for each population and trait.</p></table-wrap-foot></table-wrap></p></sec><sec id="Sec15"><title>Comparison between optimal PCR and SSPCR</title><p>Accuracies of genomic prediction differed between the PCR and SSPCR models and also based on the two CV approaches that were used to obtain the optimum PCR (or SSPCR) model (Table&#x000a0;<xref rid="Tab5" ref-type="table">5</xref>). Interestingly, in some cases the stratified CV resulted in a null model, i.e. a model where only the intercept was included. In such cases, all predicted individuals had the same GEBV which was the mean of the reference dataset. This occurred only when predicting UK_1 and was independent of the method used (PCR or SSPCR), the approach of sorting the PC (eigen vs. ss), and trait. A closer look on the number of PC used in the various PCR and SSPCR models and in the CV methods (Table&#x000a0;<xref rid="Tab5" ref-type="table">5</xref>) showed that, in general, a stratified approach reduced the number of PC but also resulted in lower accuracy, on average, compared to using 5-fold random CV. Moreover, for all except the UK_1 fat predictions, SSPCR_eigen used fewer PC than PCR_eigen. In contrast, quite a large number of PC was included in the SSPCR_ss models for all traits and both CV approaches.</p></sec><sec id="Sec16"><title>GREML versus &#x0201c;best case scenario&#x0201d; of PCR and SSPCR</title><p>In the present study, a CV approach was used to select the PCR (or SSPCR) model that was used for genomic prediction. An additional objective was to investigate the full potential of PCR (or SSPCR) and the ability of CV to achieve this. To investigate this, the pattern of the accuracies when adding PC one by one in the model, was evaluated to identify the model with the highest accuracy (&#x0201c;best case scenario&#x0201d;). It should be noted that this is not possible in practical genomic prediction applications, because it involves the use of phenotypic information of the test dataset. Those best case scenarios PCR (or SSPCR) models always outperformed GREML (Table&#x000a0;<xref rid="Tab6" ref-type="table">6</xref>) and the optimal PCR and SSPCR models based on cross-validation (Table&#x000a0;<xref rid="Tab5" ref-type="table">5</xref>).<table-wrap id="Tab6"><label>Table 6</label><caption><p> <bold>Highest accuracies</bold> ^{<bold>1</bold>} <bold>obtained for PCR models</bold> ^{<bold>2</bold>} <bold>versus those obtained with the GREML model</bold> ^{<bold>3</bold>} </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th> <bold>Test</bold> </th><th> <bold>Trait</bold> </th><th> <bold>GREML</bold> </th><th> <bold>PCR_eigen</bold> </th><th> <bold>PCR_ss</bold> </th><th> <bold>SSPCR_eigen</bold> </th><th> <bold>SSPCR_ss</bold> </th></tr></thead><tbody><tr valign="top"><td>UK_1</td><td>Milk</td><td>0.26</td><td>0.44 (14)</td><td>0.36 (2)</td><td>0.45 (6)</td><td>0.23 (458)</td></tr><tr valign="top"><td/><td>Fat</td><td>0.07</td><td>0.16 (776)</td><td>0.13 (3)</td><td>0.18 (1)</td><td>0.21 (219)</td></tr><tr valign="top"><td/><td>Protein</td><td>&#x02212;0.01</td><td>0.25 (14)</td><td>0.16 (1)</td><td>0.28 (7)</td><td>0.09 (1)</td></tr><tr valign="top"><td>UK_2</td><td>Milk</td><td>0.10</td><td>0.16 (3)</td><td>0.15 (11)</td><td>0.19 (144)</td><td>0.15 (46)</td></tr><tr valign="top"><td/><td>Fat</td><td>&#x02212;0.02</td><td>0.08 (1061)</td><td>0.07 (751)</td><td>0.12 (593)</td><td>0.09 (1370)</td></tr><tr valign="top"><td/><td>Protein</td><td>0.01</td><td>0.07 (1)</td><td>0.08 (2)</td><td>0.10 (151)</td><td>0.08 (233)</td></tr><tr valign="top"><td>SWE</td><td>Milk</td><td>0.16</td><td>0.18 (1112)</td><td>0.16 (1060)</td><td>0.21 (365)</td><td>0.17 (344)</td></tr><tr valign="top"><td/><td>Fat</td><td>0.09</td><td>0.22 (46)</td><td>0.10 (991)</td><td>0.16 (1425)</td><td>0.18 (871)</td></tr><tr valign="top"><td/><td>Protein</td><td>0.06</td><td>0.11 (265)</td><td>0.08 (790)</td><td>0.25 (1424)</td><td>0.20 (1419)</td></tr><tr valign="top"><td>IRL</td><td>Milk</td><td>0.06</td><td>0.15 (967)</td><td>0.12 (758)</td><td>0.14 (92)</td><td>0.19 (288)</td></tr><tr valign="top"><td/><td>Fat</td><td>0.08</td><td>0.12 (954)</td><td>0.14 (572)</td><td>0.16 (790)</td><td>0.12 (965)</td></tr><tr valign="top"><td/><td>Protein</td><td>0.04</td><td>0.12 (749)</td><td>0.09 (245)</td><td>0.16 (94)</td><td>0.13 (273)</td></tr><tr valign="top"><td>NLD</td><td>Milk</td><td>0.16</td><td>0.21 (20)</td><td>0.17 (4)</td><td>0.19 (24)</td><td>0.17 (16)</td></tr><tr valign="top"><td/><td>Fat</td><td>0.15</td><td>0.17 (794)</td><td>0.19 (400)</td><td>0.17 (585)</td><td>0.18 (78)</td></tr><tr valign="top"><td/><td>Protein</td><td>0.13</td><td>0.17 (7)</td><td>0.16 (8)</td><td>0.17 (5)</td><td>0.17 (4)</td></tr><tr valign="top"><td>Average</td><td>Milk</td><td>0.15</td><td>0.23 (423)</td><td>0.19 (367)</td><td>0.24 (126)</td><td>0.18 (230)</td></tr><tr valign="top"><td/><td>Fat</td><td>0.07</td><td>0.15 (726)</td><td>0.13 (543)</td><td>0.16 (679)</td><td>0.16 (701)</td></tr><tr valign="top"><td/><td>Protein</td><td>0.05</td><td>0.14 (207)</td><td>0.11 (209)</td><td>0.19 (336)</td><td>0.13 (386)</td></tr><tr valign="top"><td>SD</td><td>Milk</td><td>0.08</td><td>0.12 (565)</td><td>0.10 (506)</td><td>0.12 (144)</td><td>0.03 (192)</td></tr><tr valign="top"><td/><td>Fat</td><td>0.06</td><td>0.05 (398)</td><td>0.05 (373)</td><td>0.02 (511)</td><td>0.05 (540)</td></tr><tr valign="top"><td/><td>Protein</td><td>0.05</td><td>0.07 (323)</td><td>0.04 (341)</td><td>0.07 (611)</td><td>0.05 (591)</td></tr></tbody></table><table-wrap-foot><p>Analyses were performed for three traits and five test populations.</p><p> ¹Accuracies were calculated as Pearson correlations between predicted genomic breeding values and adjusted phenotypes; ²selection of PC was based either on the eigenvalues (eigen) or the regression sum of squares (ss); two different methods of applying principal component analysis, either separately for reference and test parts (PCR) or on the whole dataset (SSPCR), were compared; ³a REML based model with a genomic relationship matrix.</p></table-wrap-foot></table-wrap></p></sec><sec id="Sec17"><title>PCR in the &#x0201c;best case scenario&#x0201d;</title><p>For the common PCR case, where PCA was applied on the genotypic data from the reference dataset, the pattern of accuracies was evaluated for an increasing number of PC that were included in the model based on decreasing eigenvalues (Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>). Several interesting observations can be made from these results. The PCR_eigen method generally resulted in higher accuracies than the PCR_ss method (Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>; Table&#x000a0;<xref rid="Tab6" ref-type="table">6</xref>). Accuracies using PCR_ss were slightly higher in only three cases. The pattern of the accuracies, when an increasing number of PC was included in the models, differed between traits and populations. In many cases, using very few PC (usually less than 50) gave accuracies very close to the maximum obtained across the whole range of number of PC included (Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>).<fig id="Fig2"><label>Figure 2</label><caption><p> <bold>Pattern of accuracies for principal component regression models with increasing numbers of principal components (PC).</bold> PCA was performed in the reference dataset. Selection of PC was based on eigenvalues (left panel) or on the sum of square contributions (right panel). Traits analysed were average daily milk, fat and protein yields for test populations from Ireland (IRL), the Netherlands (NLD), Sweden (SWE) and two divergent selection lines from United Kingdom (UK_1 and UK_2).</p></caption><graphic xlink:href="12711_2014_60_Fig2_HTML" id="MO2"/></fig></p></sec><sec id="Sec18"><title>SSPCR in the &#x0201c;best case scenario&#x0201d;</title><p>For this model, the whole SNP dataset, i.e. both the reference and testing data, was included in the PCA, while only phenotypes of the reference subset were used to estimate the regression coefficients in PCR. In this approach, genomic information on the test dataset, such as LD, is incorporated in the weights on the SNPs applied in each eigenvector. In some cases, this SSPCR_eigen method resulted in a slight increase in accuracies and in a reduction in the number of PC needed to achieve the highest accuracies compared to PCR_eigen (Table&#x000a0;<xref rid="Tab6" ref-type="table">6</xref>, Figure&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>). Interestingly, for the UK_1 population, accuracies of 0.45, 0.18 and 0.28 for milk, fat and protein yields were obtained with only the first six, one and seven PC, respectively. On average, SSPCR resulted in slightly higher accuracies than PCR when the genotypes of the test dataset were excluded from the PCA step.<fig id="Fig3"><label>Figure 3</label><caption><p> <bold>Pattern of accuracies for semi-supervised principal component regression models with increasing numbers of principal components (PC).</bold> PCA was performed on the whole dataset. Selection of PC was based on eigenvalues (left panel) or on the sum of square contributions (right panel). Traits analysed were average daily milk, fat and protein yields for test populations from Ireland (IRL), the Netherlands (NLD), Sweden (SWE) and two divergent selection lines from United Kingdom (UK_1 and UK_2).</p></caption><graphic xlink:href="12711_2014_60_Fig3_HTML" id="MO3"/></fig></p></sec></sec><sec id="Sec19" sec-type="discussion"><title>Discussion</title><p>Principal component regression enables data reduction in the regression model and solves problems of dependencies among variables (multicollinearity). The main advantage of PCR derives from the ability of PCA to capture a large proportion of the original variability of the dataset (e.g. &#x0003e;90%) in a small set of uncorrelated PC. As a result, generally a limited number of PC can replace the original variables with little loss of information. Based on this, we tested whether PCA and its regression (PCR) can provide a useful alternative method for genomic prediction. Our results showed that, on average, PCR yielded lower accuracies than the more commonly used GREML model, although it has the potential to yield considerably higher prediction accuracies than the GREML model. It should be noted that this potential was realized in the &#x0201c;best case scenario&#x0201d; that used both genotypic and phenotypic information from the test dataset animals to derive the optimal number of PC included in the model, which is not possible in practice. Nevertheless, the results of this scenario can be regarded as an upper limit of the achievable prediction accuracy, provided that the most appropriate PC can be selected in a practical application. Optimization (i.e. selection of PC) using the reference data by two different CV approaches (stratified vs. 5-fold random), proved to be unable to capitalize on the full potential of PCR, but still achieved levels of prediction accuracy close to those obtained with GREML. Although prediction accuracies appeared to be quite low for all models, it should be noted that the reported accuracies are Pearson correlations between observed phenotypes and predicted GEBV. Transforming those values to the accuracy of GEBV, which is defined as the correlation between true and predicted GEBV, involves division by the square root of the heritability of the trait. Since, for instance, the heritabilities of the adjusted phenotypes for milk yield used in our study ranged from 0.13 to 0.59 (results not shown) across countries, accuracies of GEBV for milk yield are predicted to be 1.3 to 2.8 times higher than the reported correlations.</p><p>Genomic relationships between reference and test datasets have been shown to have an important effect on prediction accuracy [<xref ref-type="bibr" rid="CR38">38</xref>,<xref ref-type="bibr" rid="CR39">39</xref>]. The average squared relationship between reference and test datasets has been shown to be a better predictor of accuracy of genomic prediction than the average relationship between reference and test datasets [<xref ref-type="bibr" rid="CR38">38</xref>]. Since the variance of relationships is closely related to the average squared relationship, we compared the standard deviation of relationships and the average relationships between populations (Table&#x000a0;<xref rid="Tab4" ref-type="table">4</xref>). The on average low relationships between the populations in our data and, in particular, the lower variance of relationships between populations compared to within populations, predicted that accuracies of genomic predictions would be low, which was indeed the case. Although we focussed on commonly measured milk yield traits, our results can be extended to other traits such as feed intake, for which, pooling existing research herd data is the only option to enable genomic prediction [<xref ref-type="bibr" rid="CR40">40</xref>]. In fact, pooling of such data has become possible by using genotypes, because models based on genomic relationships can overcome issues caused by low connectedness based on pedigree [<xref ref-type="bibr" rid="CR41">41</xref>].</p><sec id="Sec20"><title>Optimization of PCR and SSPCR models through cross-validation</title><p>An important question is how to select the optimal set of PC for the PCR model, using information from the reference data, such that the accuracy achieved is at least similar to the accuracy achieved with GREML. We used CV on the reference data to optimize the order and number of included PC. As a first observation, the &#x0201c;best&#x0201d; PCR model, i.e. the model obtained in the &#x0201c;best case scenario&#x0201d;, was never proposed with the CV approach. In our analyses, optimization of the CV was based on minimising the MSE. However, since the accuracy of EBV is important for animal breeding and affects response to selection, an alternative scenario could be to select the &#x0201c;best&#x0201d; PCR model after CV in the reference data based on maximum accuracy instead of minimum MSE.</p><p>The main advantage of PCR was that it reduced the dimension of the data by at least 96%. Despite the generally high reduction in data dimension, the highest accuracies after CV were achieved for a wide range of numbers of PC included in the PCR model, from only one to more than 1000 (Tables&#x000a0;<xref rid="Tab5" ref-type="table">5</xref> and <xref rid="Tab6" ref-type="table">6</xref>). This is a wider range than that reported in previous studies based on simulated data, in which the highest accuracies were achieved when the number of PC ranged from 250 to 350 [<xref ref-type="bibr" rid="CR19">19</xref>,<xref ref-type="bibr" rid="CR20">20</xref>]. However, it should be noted that for PCR_eigen, which is the most commonly used approach in the literature, the number of PC in the model was between 28 and 249 after CV and between 1 and 1112 for the &#x0201c;best case scenario&#x0201d;. By adding PC one by one in the model, it was shown that most PC affected the accuracy of predictions either positively or negatively and thus the trajectory of the accuracies was not stable but fluctuated (Figures&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref> and <xref rid="Fig3" ref-type="fig">3</xref>). Moreover, in some cases the first few PC resulted in the highest accuracies. As a result, using empirical thresholds to select PC for inclusion in the model (e.g. PC that together explain more than 90% of the original variability in the SNP genotypes based on eigenvalues) simply does not result in the highest accuracies that can be achieved in PCR. Thus, the number and selection of PC for inclusion in the PCR model should be derived empirically for each dataset.</p><p>The semi-supervised PCR model also used genotypic information from the test dataset. Despite our expectation that this would improve accuracy of predictions, because the model would be forced to define PC that explain variation in the genotypes of the test dataset, this model on average performed less well than the PCR model (Table&#x000a0;<xref rid="Tab5" ref-type="table">5</xref>). However, for the best case scenario, when accuracies were evaluated across different numbers of PC based on the phenotypes of the test dataset, the semi-supervised approach did yield slightly higher accuracies than the PCR model (Table&#x000a0;<xref rid="Tab6" ref-type="table">6</xref>). This indicates that the semi-supervised PCR has the potential to perform at least equally well as PCR, although it appears that identifying the optimal set of PC for the semi-supervised PCR model is even more difficult than for the PCR model. In the context of genomic prediction, the semi-supervised PCR may be more relevant when large differences exist between the genotypes of the reference and the test datasets. The most obvious application is across-breed or -line genomic prediction, where one breed or line is used to predict another. In that case, the semi-supervised PCR model may be able to better target the variance of the predicted line or breed. It should be noted that including animals of the predicted breed in the reference data, e.g. [<xref ref-type="bibr" rid="CR42">42</xref>,<xref ref-type="bibr" rid="CR43">43</xref>], may yield a similar result, regardless of the model used.</p></sec><sec id="Sec21"><title>Investigating the importance of the principal components in regression</title><p>In other studies that used PCA for genomic prediction [<xref ref-type="bibr" rid="CR18">18</xref>&#x02013;<xref ref-type="bibr" rid="CR21">21</xref>], the PC used only accounted for the variability captured in the original matrix <bold>X</bold> (SNPs) and not for the proportion of explained phenotypic variance in the reference population (as used here with PCR_ss in the common and semi-supervised approaches). However, it has been shown in statistical literature that the first principal components (accounting for most variation in <bold>X</bold>) can totally fail as predictors in PCR (in terms of accounting for variation in the response variable) and that even components that explain little variance in <bold>X</bold> can be important for prediction [<xref ref-type="bibr" rid="CR28">28</xref>,<xref ref-type="bibr" rid="CR44">44</xref>&#x02013;<xref ref-type="bibr" rid="CR46">46</xref>]. For instance, using Hald&#x02019;s data, Hadi and Ling [<xref ref-type="bibr" rid="CR47">47</xref>] showed that while the first three (out of four) PC accounted for 99.96% of the variability in <bold>X</bold>, they contributed nothing (zero sum of squares) to the fit of the regression model; instead, the last PC alone contributed everything. Thus, these authors proposed that the selection of PC should be based not only on the variance decomposition of the co-variables but also on the contribution of each PC to the regression sum of squares. However, despite the expectation that PCR_ss would yield at least equally accurate estimates as PCR_eigen, we observed the opposite in our analyses. It should be noted that PCR_ss selected PC based on associations with phenotype in the reference data. In that regard, PCR_ss is very similar to partial least squares regression [<xref ref-type="bibr" rid="CR21">21</xref>,<xref ref-type="bibr" rid="CR48">48</xref>]. Thus, our results suggest that the PC that show the strongest associations with phenotypes in the reference data do not necessarily have the strongest associations with the phenotypes in the testing data.</p></sec><sec id="Sec22"><title>GREML versus &#x0201c;best case scenario&#x0201d; for PCR and SSPCR</title><p>By investigating the pattern of the accuracies of the PCR models (Figures&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref> and <xref rid="Fig3" ref-type="fig">3</xref>), we observed that some specific combinations of PC resulted in relatively high prediction accuracies, considering the limited size of the reference dataset, and also compared to GREML. In our analyses, the highest accuracies from the PCR models across the numbers of PC included outperformed those from the GREML models in all cases. The data reduction achieved by PCR solves the &#x0201c;small <italic>n</italic> large <italic>p</italic>&#x0201d; issue and thereby enables the use of fixed regression, as done in our study, rather than the random regressions commonly used in genomic prediction models. An important question is why PCR_eigen and SSPCR_eigen in the &#x0201c;best case scenario&#x0201d; achieved a higher accuracy than GREML, while it uses a linear transformation of the SNP data used in GREML. The most likely explanation is the fact that by using fixed regression, the model is able to put as much or as little emphasis on any PC, without shrinking the effects, following the associations with phenotypic data. Other genomic prediction models such as GREML assume equal contributions of each SNP to the total variability, and generally include shrinkage of effects of individual SNP effects by modelling them as random effects. Although the variance explained by SNPs in a shrinkage model can still depart substantially from the prior assumptions based on evidence from the data, especially if the reference population is large [<xref ref-type="bibr" rid="CR30">30</xref>], estimated effects will still be affected by those prior assumptions. In this respect, PCR can be regarded as a variable selection method, albeit at the level of PC rather than individual SNPs. This implies that the accuracies reported for the &#x0201c;best case scenarios&#x0201d; provide an upper limit for the accuracy that could be achieved with variable selection models applied to PC rather than to SNPs.</p></sec><sec id="Sec23"><title>Further improvement of PCR</title><p>One of the underlying assumptions of PCA is linearity, such that the feature space is a linear transformation of the original data. In order to overcome the problem of linearity, Sch&#x000f6;lkopf et al. [<xref ref-type="bibr" rid="CR49">49</xref>] considered nonlinear component analysis as a &#x0201c;kernel eigenvalue problem&#x0201d; and introduced the term &#x0201c;kernel PCA&#x0201d;. The use of kernels has already been introduced in genomic prediction models [<xref ref-type="bibr" rid="CR50">50</xref>]. In addition, since Bayesian models are often used in genomic prediction, the use of probabilistic PCA [<xref ref-type="bibr" rid="CR51">51</xref>], where maximum likelihood is used to extract PC, could also be proposed for future research in genomic prediction.</p><p>Concerning the selection of SNPs the target function in our study was to minimize the prediction MSE. As suggested above, an alternative could be to select PC in the CV procedure based on maximum prediction accuracy. In addition, more sophisticated techniques such as the combination of statistical methods like PCA, neural networks and genetic algorithms could be applied, as has already been tested in other fields [<xref ref-type="bibr" rid="CR52">52</xref>]. However, a balance between benefits (e.g. higher prediction accuracies) and costs (e.g. computation time) should be taken into account.</p></sec><sec id="Sec24"><title>PCA in genetic studies</title><p>In general, PCA and multivariate analysis techniques have proven to be useful tools to extract information from markers. In addition, as an exploratory method, analyses with PCA can be performed without strong assumptions on the data (e.g. Hardy-Weinberg equilibrium, LD) [<xref ref-type="bibr" rid="CR53">53</xref>]. A disadvantage of PCA is that it does not take the response variable into account. However, in our study, this did not affect accuracies negatively (comparing PCR_eigen and PCR_ss in the common and semi-supervised approach). Nevertheless, it remains necessary to be very careful when applying multivariate analysis to genomic data, especially when interpreting the results. Jombart et al. [<xref ref-type="bibr" rid="CR53">53</xref>] provided a nice overview of a multivariate analysis application with genetic data and examined the incorrect use of multivariate analysis in different genetic datasets, as well as fallacies when interpreting the results. For instance, one assumption of PCA is that PC with large eigenvalues represent structure in the data, while those with low eigenvalues capture noise. This might not always be true and PC with small eigenvalues could contain predictive information, or perhaps reflect genotypes at a single SNP. Thus, PC should not be excluded from the analysis on the basis of their small contribution to the total variance in SNP genotypes. Our results confirmed that the optimum number of PC to be included in the PCR model can vary considerably across datasets and traits.</p></sec></sec><sec id="Sec25" sec-type="conclusion"><title>Conclusions</title><p>Our results show that PCR results in genomic prediction accuracies that are generally slightly lower than those obtained with a GREML model. In general, selecting PC based on their eigenvalues resulted in higher accuracies than selecting PC on decreasing correlations to the response variable in the reference dataset. Inclusion of genotypic information of the test animals when extracting the PC, i.e. the semi-supervised approach, unexpectedly decreased the accuracy when PC were selected based on the reference dataset after cross-validation. The semi-supervised approach did, however, slightly increase the potential of the model, i.e. the highest accuracies that can be achieved, provided that it is possible to select the optimal set of included PC. While the pattern of prediction accuracies across included PC showed that PCR had a higher potential than GREML, the model that was selected by CV within the reference data could not capitalize on this potential. On average, 5-fold random CV for PCR outperformed stratified CV. However, to capitalize on the full potential of PCR in practical applications, it is still unclear what the best way to select PC to be included in the model is.</p></sec>

Along with the privilege of authorship come important responsibilities

Could not extract abstract

<contrib contrib-type="author" corresp="yes"><name><surname>Moher</surname><given-names>David</given-names></name><address><email>dmoher@ohri.ca</email></address><xref ref-type="aff" rid="Aff1"/><xref ref-type="aff" rid="Aff2"/></contrib><aff id="Aff1"><label/>Knowledge Synthesis Group, Clinical Epidemiology Program, Ottawa Hospital Research Institute, The Ottawa Hospital, General Campus, 501 Smyth Rd, Room L1288, Ottawa, ON K1H 8L6 Canada </aff><aff id="Aff2"><label/>Department of Epidemiology and Community Medicine, University of Ottawa, Roger Guindon Hall, Room 3105, 451 Smyth Road, Ottawa, Ontario K1H 8M5 Canada </aff>

BMC Medicine

<sec id="Sec1"><title>Commentary</title><p>&#x0201c;<italic>Complexity is the enemy of transparency</italic>&#x0201d; [<xref ref-type="bibr" rid="CR1">1</xref>]. Today, <italic>BMC Medicine</italic> publishes another paper on journalology (publication science) [<xref ref-type="bibr" rid="CR2">2</xref>]. Attributing authorship, and authorship order, is complex and often a &#x02018;black box&#x02019; for prospective authors. Professor Maru&#x00161;i&#x00107; and colleagues have tried to peel back the black box concerning the assigning of authorship for industry-sponsored clinical trials. Their methods are good and reported in sufficient detail to allow interested readers to replicate them [<xref ref-type="bibr" rid="CR3">3</xref>]. The research team have used an integrated knowledge translation approach to developing their proposed five-steps for transparently disclosing authorship. Participants from pharmaceutical companies that conduct clinical trials, academics, editors, and the Medical Publishing Insights and Practices initiative were involved in the entire process; this facilitates buy-in and support for the process and outcome. These same people are likely to become front-line ambassadors and early adopters for disseminating and implementing the five-step framework within their own working environments, and hopefully, more broadly.</p><p>What is positive about this research is that the proposed attribution process for authorship is brief, and not complex; it&#x02019;s only five-steps. It is meant to augment the guidance provided by the International Committee of Medical Journal Editors [<xref ref-type="bibr" rid="CR4">4</xref>]. To enhance uptake of the framework it will be important for the team, or others, to develop a bank of worked examples for each step in the five-step process. Using worked examples from specific trials will likely facilitate implementation. The authors have started the process with seven case examples included in their publication. A dedicated website for the framework whereby authorship examples can be submitted by pharmaceutical companies and others, vetted and added to a bank of examples, freely accessible to anybody, is worth considering. This will help prospective clinical trialists ensure a transparent process in deciding on authorship.</p><p>If this authorship initiative is to be successful it requires endorsement and, more importantly, implementation. As Maru&#x00161;i&#x00107; and colleagues note [<xref ref-type="bibr" rid="CR2">2</xref>], previous efforts, such as contributorship, have not been broadly implemented. What is less clear is how the proposed framework is going to be endorsed and implemented. Important initial steps should include strong and consistent language of endorsement across all of the pharmaceutical companies involved in the development of the five-step process. Support and endorsement from umbrella groups, such as the Pharmaceutical Research and Manufacturers of America [<xref ref-type="bibr" rid="CR5">5</xref>], and others such as CONSORT [<xref ref-type="bibr" rid="CR6">6</xref>], is also worth considering. While endorsement is a useful step, it is difficult to measure and likely not the most relevant outcome. More important will be to develop plans based on appropriately developed approaches [<xref ref-type="bibr" rid="CR7">7</xref>,<xref ref-type="bibr" rid="CR8">8</xref>] to implement the framework. This is likely to be most effective when pharmaceutical companies modify their authorship practices and polices when conducting any clinical trial. An effective policy would require all clinical trials to implement the five-step framework at their inception. Without strong implementation the framework is less likely to affect positive change. This has been observed when trying to implement reporting guidelines in biomedical journals [<xref ref-type="bibr" rid="CR9">9</xref>,<xref ref-type="bibr" rid="CR10">10</xref>]. Part of any implementation plan also needs to include an evaluation of the framework. It is important to collect data that will inform its usefulness. There is little merit in maintaining policies that are not supported by evidence.</p><p>Maru&#x00161;i&#x00107; et al. are silent on whether their framework can be used when developing authorship for submitting clinical trial protocols for publication consideration [<xref ref-type="bibr" rid="CR2">2</xref>]. Making clinical trial protocols accessible is important and at least one Biomed Central journal &#x02013; <italic>Trials</italic> &#x02013; regularly publishes them. Additionally, with the requirement of trial registration, this framework could also be used when completing the investigator information part of the registration.</p><p>Most of us are not born authors. It is an acquired skill that often starts during graduate school. This is where all journalology issues, including those pertaining to authorship issues (e.g., attributing authorship, authorship order, and ghost and guest authorship, author responsibilities) should be formally taught and discussed. Developing such skills early can translate into something useful throughout a researcher&#x02019;s (author&#x02019;s) career. It is unfortunate that almost all universities, and other centres of higher learning, appear to have abdicated their responsibilities regarding formally teaching journalology; the irony is not lost on me. These institutions are the very same places developing the next generation of biomedical researchers. Universities need to set aside appropriate resources to enable and promote such courses, and others, related to journalology [<xref ref-type="bibr" rid="CR11">11</xref>].</p><p>While authors have rights and privileges, they also have important responsibilities that require much greater attention. Given the opportunity of authorship, it is equally important to assert this responsibility. Authors must ensure that papers baring their name are &#x0201c;fit for purpose&#x0201d; [<xref ref-type="bibr" rid="CR12">12</xref>]. Here, authors need to ensure that every report baring their name is a completely reported and transparent account of what was done (methods) and found (results) to enable interested readers to replicate the methods and use the results. Collectively, authors have not performed appropriately with regards to reporting their clinical trials. This avoidable waste is troublesome for shareholders of publicly traded pharmaceutical companies and tax payers of publicly-funded clinical trials. It is not a good return of a fiscal investment when reports of trials are so inadequate that their results cannot be used. For example, Duff et al. [<xref ref-type="bibr" rid="CR13">13</xref>] examined 262 reports of trials from the most prominent oncology journals assessing them for 10 essential elements regarding the description of their interventions, such as the drug&#x02019;s name and route of administration. The authors reported that only 11% of the articles reported all 10 characteristics. Although we have seen improvements over time in reporting the unique characteristics of randomized trials &#x02013; sequence generation, allocation concealment, and implementation &#x02013; these items are adequately reported in less than half of the trial reports [<xref ref-type="bibr" rid="CR14">14</xref>]. In some clinical specialties, the situation is much worse [<xref ref-type="bibr" rid="CR15">15</xref>].</p><p>Disclosing authorship transparently is important for any manuscript being submitted to a biomedical journal for publication consideration. The responsibilities associated with authorship must be taken seriously. This might help increase value and reduce avoidable waste of biomedical research.</p></sec>

Myofibroblastic reaction is a common event in metastatic disease of breast carcinoma: a descriptive study

Could not extract abstract

<contrib contrib-type="author" corresp="yes"><name><surname>Catteau</surname><given-names>Xavier</given-names></name><address><email>xavier.catteau06@gmail.com</email></address><xref ref-type="aff" rid="Aff1"/><xref ref-type="aff" rid="Aff2"/><xref ref-type="aff" rid="Aff5"/></contrib><contrib contrib-type="author"><name><surname>Simon</surname><given-names>Philippe</given-names></name><address><email>Philippe.Simon@erasme.ulb.ac.be</email></address><xref ref-type="aff" rid="Aff2"/><xref ref-type="aff" rid="Aff3"/></contrib><contrib contrib-type="author"><name><surname>No&#x000eb;l</surname><given-names>Jean-Christophe</given-names></name><address><email>Jean-Christophe.Noel@erasme.ulb.ac.be</email></address><xref ref-type="aff" rid="Aff2"/><xref ref-type="aff" rid="Aff4"/></contrib><aff id="Aff1"><label/>Institute of Pathology and Genetics, Gosselies, Belgium </aff><aff id="Aff2"><label/>Faculty of Medicine, Universit&#x000e9; Libre de Bruxelles, Brussels, Belgium </aff><aff id="Aff3"><label/>Gynaecology Unit, Erasme University Hospital-Universit&#x000e9; libre de Bruxelles, Brussels, Belgium </aff><aff id="Aff4"><label/>Gynaecopathology Unit, Pathology Department, Erasme University Hospital-Universit&#x000e9; Libre de Bruxelles, Brussels, Belgium </aff><aff id="Aff5"><label/>Department of Pathology, Institute of Pathology and Genetics, 25, Avenue Georges Lema&#x000ee;tre, Gosselies, 6041 Belgium </aff>

Diagnostic Pathology

<sec id="Sec1" sec-type="introduction"><title>Background</title><p>The importance of the stromal microenvironment has been suggested to play a major role in breast carcinoma by promoting tumour growth, progression and invasion [<xref ref-type="bibr" rid="CR1">1</xref>-<xref ref-type="bibr" rid="CR4">4</xref>]. In particular according to these data we and others have clearly demonstrated that the loss of CD34 fibrocytes and acquisition of peritumoral myofibroblasts expressing smooth muscle actin (SMA) is a fundamental step both in ductal carcinoma in situ (DCIS) and invasive carcinoma of no special type (NST) [<xref ref-type="bibr" rid="CR5">5</xref>,<xref ref-type="bibr" rid="CR6">6</xref>]. If the acquisition of a myofibroblastic differentiation is an important data in peritumoral connective tissue remodeling [<xref ref-type="bibr" rid="CR4">4</xref>], the morphological characterization of stromal microenvironment and particularly of myofibroblastic peritumoral cells in metastatic location is less understood. In preliminaries studies, some authors have suggested that the acquisition of a myofibroblastic differentiation could play a role in metastatic colonic adenocarcinoma [<xref ref-type="bibr" rid="CR7">7</xref>] but however, until now, these data have not been clearly described in breast metastatic sites. Therefore, to clarify this issue, the aim of the present study is to assess by immunohistochemistry, the topographic distribution of CD 34 positive fibrocytes and SMA positive myofibroblasts both in axillary lymph node and liver metastases which are frequent in breast carcinoma and strongly associated with an increased risk of distant metastasis and poor overall survival [<xref ref-type="bibr" rid="CR8">8</xref>].</p></sec><sec id="Sec2" sec-type="materials|methods"><title>Methods</title><sec id="Sec3"><title>Study population</title><p>We used a computer database from the Pathology and Genetics Institute (IPG) to identify 77 consecutive patients diagnosed between January 2008 and December 2012 with lymph node (n&#x02009;=&#x02009;41) and liver metastasis (n&#x02009;=&#x02009;36). The study protocol was approved by the institutional ethics (Ethics Committee Erasme Hospital) and research review boards. The belgian number (number of agreation) of this committee is OM021. The reference for this study is P2012/191. Consent has been established by the local ethics committee and is in accordance with Belgian and International law. For each patient, the following parameters including age, TNM classification, tumour grade and tumour size were performed according to the 4^th edition of WHO classification and are summarized in the Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>.<table-wrap id="Tab1"><label>Table 1</label><caption><p> <bold>Clinicopathological data of the study population</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th/><th> <bold>Liver metastases N&#x02009;=&#x02009;36</bold> </th><th> <bold>Lymph node metastases N =41</bold> </th></tr></thead><tbody><tr valign="top"><td/><td>No.</td><td>No.</td></tr><tr valign="top"><td>Age</td><td/><td/></tr><tr valign="top"><td>Mean</td><td>59.6</td><td>59</td></tr><tr valign="top"><td>Range</td><td>34 - 86</td><td>37 - 86</td></tr><tr valign="top"><td>Primary tumour size</td><td/><td/></tr><tr valign="top"><td>T1 (0.1- 2&#x000a0;cm)</td><td>18</td><td>21</td></tr><tr valign="top"><td>T2 (&#x0003e;2- 5&#x000a0;cm)</td><td>14</td><td>17</td></tr><tr valign="top"><td>T3 (&#x0003e;5&#x000a0;cm)</td><td>4</td><td>3</td></tr><tr valign="top"><td>Primary tumour grade</td><td/><td/></tr><tr valign="top"><td>Grade 1</td><td>3</td><td>8</td></tr><tr valign="top"><td>Grade 2</td><td>23</td><td>22</td></tr><tr valign="top"><td>Grade 3</td><td>10</td><td>11</td></tr></tbody></table></table-wrap></p></sec><sec id="Sec4"><title>Immunohistochemistry</title><p>The specimens were fixed in histology-grade 4% buffered formalin. Series paraffin sections were stained with haematoxylin and eosin and immunohistochemical detection was performed according to the manufacturer&#x02019;s protocols (Table&#x000a0;<xref rid="Tab2" ref-type="table">2</xref>). We used a fully automated immunohistochemical system (Autostainer Link A48 Dako).<table-wrap id="Tab2"><label>Table 2</label><caption><p> <bold>Antibodies used in this study</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th> <bold>Antigen</bold> </th><th> <bold>Clone</bold> </th><th/><th> <bold>Dilution</bold> </th><th> <bold>Source</bold> </th><th> <bold>Catalog number</bold> </th></tr></thead><tbody><tr valign="top"><td>CD 34</td><td>QBEnd-10</td><td>Monoclonal Mouse</td><td>Ready-to-use</td><td>Dako</td><td>IR63261</td></tr><tr valign="top"><td>Vimentine</td><td>V9</td><td>Monoclonal Mouse</td><td>Ready-to-use</td><td>Dako</td><td>IR63061</td></tr><tr valign="top"><td>&#x003b1;-SMA</td><td>1A4</td><td>Monoclonal Mouse</td><td>Ready-to-use</td><td>Dako</td><td>IR00611</td></tr><tr valign="top"><td>CKAE1/AE3</td><td>AE1/AE3</td><td>Monoclonal Mouse</td><td>Ready-to-use</td><td>Dako</td><td>IR05361</td></tr></tbody></table></table-wrap></p></sec><sec id="Sec5"><title>Semi-quantitative Assessment of Immunohistochemistry</title><p>We compared the distribution of CD34 and SMA between stromal areas located within the metastasis with areas of normal liver and lymph node tissue. The immunoreactivity of CD34 and SMA was assessed semi-quantitatively in the free tissue and the tumour. The percentage of stromal cells expressing each antigen was graded as &#x0201c;0&#x0201d;, &#x0201c;+&#x0201d;, &#x0201c;++&#x0201d;, &#x0201c;+++&#x0201d;, &#x0201c;++++&#x0201d; when up to 5%, more than 5% and up to 25%, more than 25% and up to 50%, more than 50% and up to 75% or more than 75% of stromal cells, disclosed immunoreactivity, respectively. Percentages were assessed by two independent observers, assuming that a high-power microscopic field (objective &#x000d7;40, microscopic magnification: &#x000d7;400) harboured 100 stromal cells (range: 75&#x02013;150) as previously described [<xref ref-type="bibr" rid="CR9">9</xref>]. The relationship between the staining pattern of SMA and different clinical and histological features (age, tumour size and grade, TNM classification) was compared using a Chi-squared test. A p value &#x0003c;0.05 was considered statistically significant. All analyses were performed using Statistica&#x000ae;.</p></sec></sec><sec id="Sec6" sec-type="results"><title>Results</title><sec id="Sec7"><title>CD 34 and SMA expression in lymph node metastases</title><p>In the normal lymph node, CD34 expression was limited both on the capsule and pericapsular fibrocytes but also in vessels within the parenchyma. CD34 fibrocytes were totally absent in peritumoral stroma and the immunoreactivity was restricted to the vasculature. By contrast, myofibroblasts were present in peritumoral stroma in 95% of cases with in a majority of cases more than 50% of stromal cells positive but this expression was not statistically correlated with clinical or pathological features (p&#x02009;&#x0003e;&#x02009;0.05) (Tables&#x000a0;<xref rid="Tab3" ref-type="table">3</xref> and <xref rid="Tab4" ref-type="table">4</xref>). The peritumoral myofibroblasts surrounded intimately the malignant cells (Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>). In the capsule both in the normal and peritumoral areas, a strong immunoreactivity for SMA was also observed. Lastly, in normal area, the reticular dendritic cells, stromal cells with myo&#x000ef;d features and vascular walls showed as previously described a discrete to moderate reactivity for SMA [<xref ref-type="bibr" rid="CR10">10</xref>-<xref ref-type="bibr" rid="CR12">12</xref>].<table-wrap id="Tab3"><label>Table 3</label><caption><p> <bold>Stromal SMA expression in lymph node and liver metastatic sites</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th> <bold>SMA expression</bold> </th><th> <bold>0</bold> </th><th> <bold>+ or ++</bold> </th><th> <bold>+++ or ++++</bold> </th><th> <bold>Total</bold> </th></tr></thead><tbody><tr valign="top"><td>Lymph node</td><td>2 (5%)</td><td>5 (12%)</td><td>34 (83%)</td><td>41</td></tr><tr valign="top"><td>Liver</td><td>1 (3%)</td><td>3 (8%)</td><td>32 (89%)</td><td>36</td></tr></tbody></table></table-wrap><table-wrap id="Tab4"><label>Table 4</label><caption><p> <bold>Relation of SMA stromal expression and clinicopathological features</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th/><th colspan="2"> <bold>Lymph node metastases</bold> </th><th/><th colspan="2"> <bold>Liver metastases</bold> </th><th/></tr><tr valign="top"><th/><th> <bold>Strong expression</bold> </th><th> <bold>Weak expression</bold> </th><th> <bold>p</bold> </th><th> <bold>Strong expression</bold> </th><th> <bold>Weak expression</bold> </th><th> <bold>p</bold> </th></tr></thead><tbody><tr valign="top"><td>Age</td><td/><td/><td/><td/><td/><td/></tr><tr valign="top"><td>&#x02264; 40</td><td>3</td><td>0</td><td/><td>12</td><td>2</td><td/></tr><tr valign="top"><td>&#x0003e; 40</td><td>31</td><td>7</td><td>0.4</td><td>20</td><td>2</td><td>0.6</td></tr><tr valign="top"><td>Tumour grade</td><td/><td/><td/><td/><td/><td/></tr><tr valign="top"><td>G1</td><td>6</td><td>2</td><td/><td>2</td><td>1</td><td/></tr><tr valign="top"><td>G2</td><td>17</td><td>4</td><td/><td>21</td><td>2</td><td/></tr><tr valign="top"><td>G3</td><td>11</td><td>1</td><td>0.6</td><td>9</td><td>1</td><td>0.9</td></tr><tr valign="top"><td>Tumour size (mm)</td><td/><td/><td/><td/><td/><td/></tr><tr valign="top"><td>&#x02264; 10</td><td>4</td><td>2</td><td/><td>3</td><td>1</td><td/></tr><tr valign="top"><td>&#x0003e; 10 and &#x02264; 20</td><td>14</td><td>2</td><td/><td>12</td><td>2</td><td/></tr><tr valign="top"><td>&#x0003e; 20</td><td>16</td><td>3</td><td>0.4</td><td>17</td><td>1</td><td>0.5</td></tr></tbody></table></table-wrap><fig id="Fig1"><label>Figure 1</label><caption><p> <bold>Typical example of SMA positive myofibroblasts within lymph node metastasis of invasive mammary carcinoma of NST (x100).</bold> Note that the myofibroblasts surround intimately the cancer cells at high power view (inset; x400).</p></caption><graphic xlink:href="13000_2014_196_Fig1_HTML" id="MO1"/></fig></p></sec><sec id="Sec8"><title>CD 34 and SMA expression in liver metastases</title><p>In normal liver parenchyma, CD 34 expression was limited to vascular walls and focal immunoreactivity in portal tract. In the peritumoral stroma as in the lymph node, the immunoreactivity was restricted to the vasculature. Myofibroblasts are found intimately surrounding tumoral metastatic cells in peritumoral stroma in 97% of cases and like in lymph node was not statistically correlated with clinical and pathological features (Tables&#x000a0;<xref rid="Tab3" ref-type="table">3</xref> and <xref rid="Tab4" ref-type="table">4</xref>) (Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>). In normal liver area, as previously described, both hepatic stellate cells (Ito cells) in perisinusoidal spaces, portal tracts and vascular walls were positive for SMA [<xref ref-type="bibr" rid="CR13">13</xref>].<fig id="Fig2"><label>Figure 2</label><caption><p> <bold>Myofibroblastic reaction within peritumoral stroma in liver metastasis (x200).</bold> </p></caption><graphic xlink:href="13000_2014_196_Fig2_HTML" id="MO2"/></fig></p></sec></sec><sec id="Sec9" sec-type="discussion"><title>Discussion</title><p>In preliminary papers, we have previously described that the stromal reaction and in particular tumor-associated myofibroblasts which are prominent in this stroma is a common feature both in situ and invasive breast carcinomas. In the present study, for the first time, we have demonstrated that myofibroblastic reaction was also common both in lymph node and liver metastases. Indeed, in more than 90% of metastatic lymph node and liver metastatic cases, a peritumoral myofibroblastic reaction is present and the myofibroblasts generally surrounded intimately the tumoral cells. In addition, like in primary breast carcinoma, we have not observed CD 34 fibrocytes in the stroma and the immunoreactivity for this marker was restricted to the vascular walls, which possibly represent &#x0201c;neovessels&#x0201d;.</p><p>In primary breast carcinoma, we have previously demonstrated that one of the potential origin of this peritumoral myofibroblasts is the transformation of resident fibrocytes CD 34 positive into myofibroblats by the way of the TGF-&#x000df; 1/TGF-&#x000df; 1 receptor. However, the precursor of these myofibroblasts remains hypothetical in metastatic process.</p><p>In liver, we have could demonstrate like others, firstly that myofibroblasts were absent in normal parenchyma and secondly that SMA positivity was observed in vascular walls, portal tract stroma and hepatic perisinusoidal cells [<xref ref-type="bibr" rid="CR13">13</xref>]. Therefore, as suggested by several authors in liver fibrosis, these cells could be potential precursor for myofibroblasts, which constitute the major source of collagen deposits [<xref ref-type="bibr" rid="CR14">14</xref>,<xref ref-type="bibr" rid="CR15">15</xref>]. By analogy, in the lymph node, stromal cells from the capsule, stromal cells with myo&#x000ef;d features and endothelial cells are potential precursors of myofibroblasts. In addition, both in lymph node and liver metastases, generation of myofibroblast either by epithelial-mesenchymal transition (EMT) process from epithelial carcinomatous cells, resident mesenchymal stem cells, or from totipotential bone marrow cells is still debate [<xref ref-type="bibr" rid="CR16">16</xref>,<xref ref-type="bibr" rid="CR17">17</xref>].</p><p>If the myofibroblastic reaction seems a constant event both in breast carcinoma lymph node and liver metastases, until now, it is unclear whether this event is favourable to development of the metastatic process or by opposition is just a secondary passive reaction remains unsettled. In diverse primary carcinomatous processes including breast carcinoma, myofibroblasts promote tumour growth, invasion and angiogenesis through the paracrine effects of multiple factors including TGF-&#x000df; 1 and matrix-metalloproteinases [<xref ref-type="bibr" rid="CR18">18</xref>]. However, these factors are actually poorly characterized in metastatic process actually.</p></sec><sec id="Sec10" sec-type="conclusion"><title>Conclusions</title><p>In summary, the presence of activated myofibroblasts in lymph node and liver metastases of breast carcinoma highlights the importance of the microenvironment in supporting cancers. Understanding the relationship between myofibroblasts and metastases is not just of prognostic significance, it could provide a new therapeutic target for the treatment of advanced cancer.</p></sec>

Deciphering clonality in aneuploid breast tumors using SNP array and sequencing data

Could not extract abstract

<contrib contrib-type="author" corresp="yes"><name><surname>L&#x000f6;nnstedt</surname><given-names>Ingrid M</given-names></name><address><email>ingrid.lonnstedt@gmail.com</email></address><xref ref-type="aff" rid="Aff1"/><xref ref-type="aff" rid="Aff2"/><xref ref-type="aff" rid="Aff3"/></contrib><contrib contrib-type="author"><name><surname>Caramia</surname><given-names>Franco</given-names></name><address><email>Franco.Caramia@petermac.org</email></address><xref ref-type="aff" rid="Aff3"/></contrib><contrib contrib-type="author"><name><surname>Li</surname><given-names>Jason</given-names></name><address><email>Jason.Li@petermac.org</email></address><xref ref-type="aff" rid="Aff3"/></contrib><contrib contrib-type="author"><name><surname>Fumagalli</surname><given-names>Debora</given-names></name><address><email>fumagalli@bordet.be</email></address><xref ref-type="aff" rid="Aff4"/></contrib><contrib contrib-type="author"><name><surname>Salgado</surname><given-names>Roberto</given-names></name><address><email>roberto.salgado@bordet.be</email></address><xref ref-type="aff" rid="Aff4"/></contrib><contrib contrib-type="author"><name><surname>Rowan</surname><given-names>Andrew</given-names></name><address><email>Andrew.rowan@cancer.org.uk</email></address><xref ref-type="aff" rid="Aff5"/></contrib><contrib contrib-type="author"><name><surname>Salm</surname><given-names>Max</given-names></name><address><email>Max.Salm@cancer.org.uk</email></address><xref ref-type="aff" rid="Aff6"/></contrib><contrib contrib-type="author"><name><surname>Kanu</surname><given-names>Nnennaya</given-names></name><address><email>n.kanu@ucl.ac.uk</email></address><xref ref-type="aff" rid="Aff7"/></contrib><contrib contrib-type="author"><name><surname>Savas</surname><given-names>Peter</given-names></name><address><email>Peter.savas@petermac.org</email></address><xref ref-type="aff" rid="Aff3"/></contrib><contrib contrib-type="author"><name><surname>Horswell</surname><given-names>Stuart</given-names></name><address><email>stuart.horswell@cancer.org.uk</email></address><xref ref-type="aff" rid="Aff6"/></contrib><contrib contrib-type="author"><name><surname>Gade</surname><given-names>Stephan</given-names></name><address><email>Stephan.Gade@germanbreastgroup.de</email></address><xref ref-type="aff" rid="Aff8"/></contrib><contrib contrib-type="author"><name><surname>Loibl</surname><given-names>Sibylle</given-names></name><address><email>Sibylle.Loibl@germanbreastgroup.de</email></address><xref ref-type="aff" rid="Aff8"/></contrib><contrib contrib-type="author"><name><surname>Neven</surname><given-names>Patrick</given-names></name><address><email>patrick.neven@uzleuven.be</email></address><xref ref-type="aff" rid="Aff9"/></contrib><contrib contrib-type="author"><name><surname>Sotiriou</surname><given-names>Christos</given-names></name><address><email>Christos.sotiriou@bordet.be</email></address><xref ref-type="aff" rid="Aff4"/></contrib><contrib contrib-type="author"><name><surname>Swanton</surname><given-names>Charles</given-names></name><address><email>Charles.swanton@cancer.org.uk</email></address><xref ref-type="aff" rid="Aff5"/><xref ref-type="aff" rid="Aff10"/></contrib><contrib contrib-type="author" corresp="yes"><name><surname>Loi</surname><given-names>Sherene</given-names></name><address><email>sherene.loi@petermac.org</email></address><xref ref-type="aff" rid="Aff2"/><xref ref-type="aff" rid="Aff3"/></contrib><contrib contrib-type="author"><name><surname>Speed</surname><given-names>Terence P</given-names></name><address><email>terry@wehi.edu.au</email></address><xref ref-type="aff" rid="Aff1"/><xref ref-type="aff" rid="Aff11"/></contrib><aff id="Aff1"><label/>Bioinformatics Division, The Walter and Eliza Hall Institute of Medical Research, 1G Royal Parade, Parkville, VIC 3052 Australia </aff><aff id="Aff2"><label/>University of Melbourne, Melbourne, VIC 3010 Australia </aff><aff id="Aff3"><label/>Division of Research and Cancer Medicine, Peter MacCallum Cancer Centre, East Melbourne, VIC 3002 Australia </aff><aff id="Aff4"><label/>Breast Cancer Translational Research Laboratory, Institut Jules Bordet, Brussels, Belgium </aff><aff id="Aff5"><label/>Cancer Research UK, London Research Institute, Translational Cancer Therapeutics Laboratory, 44 Lincoln&#x02019;s Inn Fields, London, WC2A 3LY UK </aff><aff id="Aff6"><label/>Bioinformatics and BioStatistics, Cancer Research UK, Lincoln&#x02019;s Inn Fields, Holborn, London WC2A 3LY UK </aff><aff id="Aff7"><label/>Translational Cancer Therapeutics Laboratory, UCL Cancer Institute, Paul O&#x02019;Gorman Building, University College London, 72 Huntley Street, London, WC1E 6DD UK </aff><aff id="Aff8"><label/>German Breast Group (GBG), Neu Isenburg, Germany </aff><aff id="Aff9"><label/>Multidisciplinary Breast Centre and Gynaecological Oncology, KU Leuven, University of Leuven, Department of Oncology, B-3000 Leuven, Belgium </aff><aff id="Aff10"><label/>UCL Cancer Institute, Paul O&#x02019;Gorman Building, University College London, 72 Huntley Street, London, WC1E 6DD UK </aff><aff id="Aff11"><label/>Department of Mathematics and Statistics, University of Melbourne, Melbourne, VIC 3010 Australia </aff>

Genome Biology

<sec id="Sec1" sec-type="introduction"><title>Background</title><p>Genomes can vary between cells within a tumor. Mutations and copy number (CN) alterations which appear during tumor development result in genomic subclones emerging. Subclonality of tumors is referred to as intra-tumor heterogeneity (ITH), a topic which has attracted much attention over the last few years [<xref ref-type="bibr" rid="CR1">1</xref>&#x02013;<xref ref-type="bibr" rid="CR17">17</xref>]. The subclones within a tumor may display different features and respond differently to drugs. It has been speculated that heterogeneity-related endpoints - a tumor&#x02019;s clonal architecture, features of the subclones, or whether mutations are clonal (present equally in all tumor cells) or subclonal - might serve as biomarkers for drug resistance [<xref ref-type="bibr" rid="CR5">5</xref>,<xref ref-type="bibr" rid="CR18">18</xref>,<xref ref-type="bibr" rid="CR19">19</xref>].</p><p>Heterogeneity of cancers has been studied by comparing mutations and CN alterations between spatially separated [<xref ref-type="bibr" rid="CR3">3</xref>,<xref ref-type="bibr" rid="CR6">6</xref>,<xref ref-type="bibr" rid="CR7">7</xref>] or sequential [<xref ref-type="bibr" rid="CR10">10</xref>] samples from the same tumor, or between primary and secondary tumors [<xref ref-type="bibr" rid="CR11">11</xref>] from the same patient.</p><p>To meet clinical demand, recent studies have attempted to assess heterogeneity from single tumor samples based on whole genome sequencing (WGS) [<xref ref-type="bibr" rid="CR4">4</xref>,<xref ref-type="bibr" rid="CR8">8</xref>,<xref ref-type="bibr" rid="CR9">9</xref>,<xref ref-type="bibr" rid="CR12">12</xref>,<xref ref-type="bibr" rid="CR14">14</xref>&#x02013;<xref ref-type="bibr" rid="CR16">16</xref>] or the cheaper whole exome sequencing (WES) [<xref ref-type="bibr" rid="CR1">1</xref>,<xref ref-type="bibr" rid="CR2">2</xref>,<xref ref-type="bibr" rid="CR5">5</xref>,<xref ref-type="bibr" rid="CR13">13</xref>,<xref ref-type="bibr" rid="CR17">17</xref>], usually in combination with genome-wide data from SNP arrays. In general, the average CN across all cells in the tumor sample is estimated at numerous genomic positions from SNP arrays or sequencing data, and these values are joined up into genome segments of constant CN (from now on called segmented CN data). Next, the variant allele fraction (VAF) of each somatic mutation identified in the sequencing data is compared to the local CN estimate, in order to classify the mutation as clonal or subclonal. Some papers proceed to construct a phylogenetic tree which visualizes the subclonal evolution of the tumor [<xref ref-type="bibr" rid="CR8">8</xref>,<xref ref-type="bibr" rid="CR9">9</xref>,<xref ref-type="bibr" rid="CR14">14</xref>&#x02013;<xref ref-type="bibr" rid="CR17">17</xref>].</p><p>We have looked at 52 single samples from newly diagnosed <italic>HER2</italic>-positive breast cancer tumors in the RESPONSIFY project [<xref ref-type="bibr" rid="CR20">20</xref>] using Affymetrix SNP 6.0 arrays, WES and pathologist purity estimates. Our tumors all show heterogeneity, in that most are highly aneuploid throughout most of the genome in only a fraction of the cells. The scientific question driving this methodology project was whether identified mutations are clonal or subclonal. In particular, we hoped to assess clonality of specific CN alterations, such as those of <italic>HER2</italic>, by inferring the status of mutations present at their genomic location. It turns out, as we will demonstrate, that classification of mutations in samples with heterogeneity is not always possible with the data we had.</p><p>The focus of this paper is on the stages of analysis preceding automatic approaches which take input data and return an estimated clonality status of each mutation. Our principal aim is to highlight challenges in CN estimation infrequently acknowledged in the literature which influence mutation classification. We also propose solutions that may aid in the quantification of ITH in tumor samples that have high levels of CN alteration. Such a method will help in understanding how ITH is related to prognosis, that is, survival of patients diagnosed with breast cancers, as well as drug resistance, as it will be applicable to large datasets with annotated treatment and clinical outcome information.</p><p>We make extensive use of grid plots, which help visualize the clonal architecture of aneuploid tumor samples and provide visual feedback on the absence or presence of bias in segmented CN data. We also describe the key issues and challenges in CN estimation of subclonal samples, and show how local subclonal integer CN estimates are vital for correct classification of mutations.</p><p>Our demonstrations are restricted to a handful of the 52 RESPONSIFY <italic>HER2</italic>-positive breast cancer samples. Complete analyses of all samples with medical results, including potential biomarkers for resistance to trastuzumab-based therapy, will be published separately.</p><p>Our results are divided into three parts (A to C). In part A we present grid plots and demonstrate key issues in the estimation of CN of subclonal tumor samples in a simulated setting, to show that even with no noise or bias, subclonal chromosomal CNs can only be estimated in some genome segments, in samples with simple subclonal architectures, and even then relying on subjective assumptions. In part B, still in a simulated framework with no noise or bias, we show how the subclonal chromosomal CNs play a vital role in the classification of mutations as clonal or subclonal. In part C we briefly discuss our data. We suggest a probabilistic strategy to separate subclonality from noise in segmented CN data, and to assign a clonality status to a mutation. We also supply a two-dimensional grid rotation method to adjust for B allele fraction bias, which is common in our datasets.</p><p>We will refer to the number of chromosomal copies at a genome position in specified cells as their (true) integer CN. The average integer CN across cells from a tumor sample at a genome position will be called the (true) average CN. SNP array signals, which have been preprocessed, segmented and possibly normalized towards germline array data so that they are supposedly proportional to average CNs apart from noise deviations, will be called array CNs. By cell fraction we mean the percentage of sample cells (out of both normal and tumor cells of a sample) that make up a specified subclone.</p></sec><sec id="Sec2" sec-type="results"><title>Results</title><sec id="Sec3"><title>A: Grid patterns and integer CN estimation in simulated aneuploid tumors</title><p>In this section we present grid plots and demonstrate key issues in the estimation of CN of subclonal tumor samples in a simulated setting, to show that even with no noise or bias, subclonal chromosomal CNs can only be estimated in some genome segments, in samples with simple subclonal architectures, and even then relying on subjective assumptions. This step is important for subsequent classification of mutations as clonal or subclonal, since the mutation VAFs depend on local integer CNs in the tumor cells.</p><sec id="Sec4"><title>Clonal tumors and grid plots</title><p>A normal, diploid cell has one copy of each parental chromosome in its nucleus. We say its integer CNs are (1,1). Aneuploid tumor cells exhibit integer CNs other than (1,1), including segments with loss of heterozygosity (LOH), such as (0,1) or (0,2), or CN gains, such as (1,2), (1,3) or (2,2). Each genome segment of constant CN in an aneuploid tumor cell may be represented by a point in a grid plot, a figure which displays all the combinations of CNs that occur throughout the genome in that cell, in a minor (smaller) versus major (larger) homologue CN pattern (extending the idea of TAPS plots in [<xref ref-type="bibr" rid="CR21">21</xref>]). Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>a is a grid plot of simulated integer CNs in a cell in which each possible major and minor combination of 0, 1, 2, 3 and 4 copies occurs somewhere along the genome.<fig id="Fig1"><label>Figure 1</label><caption><p> <bold>Grid plot of a simulated clonal tumor sample. (a)</bold> Grid plot of the integer CNs of a simulated single aneuploid tumor cell where all the combinations of major and minor integer CNs from 0 to 4 occur in the genome. <bold>(b)</bold> Grid plot of the average CNs from a simulated tumor sample with purity 80% and clonal CNs as in (a). Note how this grid is a version of (a) shrunken towards the point (1,1).</p></caption><graphic xlink:href="13059_2014_470_Fig1_HTML" id="MO1"/></fig></p><p>Tumor samples consist of thousands of tumor cells plus an unknown fraction of normal diploid cells, which we call normal contamination. We simulate a sample with clonal tumor CNs (identical integer CNs across all the tumor cells) as in Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>a with fraction (purity) <italic>p</italic> of tumor cells. Each average CN <italic>e</italic> of a given homologue and genome segment will then have the form:<disp-formula id="Equ1"><label>1</label><alternatives><tex-math id="M1">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ e=\left(1-p\right)+pc,\;c=0,\;1,\;2, \dots,\;4, $$\end{document}</tex-math><graphic xlink:href="13059_2014_470_Article_Equ1.gif" position="anchor"/></alternatives></disp-formula></p><p>where <italic>c</italic> is the integer CN in the tumor cells (grid plot in Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>b). Compared with Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>a, each point in Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>b is shifted (shrunken) towards the point (1,1), since each average CN is the average of the integer CNs (1,1) of the normal cells and the tumor cell integer CNs.</p><p>Integer CNs and the purity of a tumor sample can only be unambiguously estimated from the unbiased, noise free average CNs via Equation&#x000a0;<xref rid="Equ1" ref-type="">1</xref> if 1) the sample is known to be clonal, and 2) there are at least two points in the grid plot for which the difference is known on an integer CN scale. For example, it may be known that two consecutive vertical grid points reflect a difference of one copy in the minor homologue.</p><p>With tumor samples, it is seldom known that a sample is clonal (1), so we broaden the CN estimation framework to that of (potentially) subclonal tumors.</p></sec><sec id="Sec5"><title>Subclonal tumors</title><p>For tumors with heterogeneity, CN estimation comes down to estimation of the cell fraction and integer CNs of each subclone. As we shall see, this is a very difficult task with the data we consider.</p><p>Grid patterns from tumor samples with heterogeneity are more complicated than those in Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>. We simulate a sample consisting of 20% germline cells (<italic>p</italic> =80% purity), <italic>&#x003b1;</italic> =30% cells forming an aneuploid subclone A with integer CNs as in Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>a and <italic>&#x003b2;</italic> =50% cells forming another subclone B with integer CNs as in Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>b. Simulated average CNs are segment-specific averages of the subclonal integer CNs across all the sample cells, so in a given genome segment they take values of the form:<fig id="Fig2"><label>Figure 2</label><caption><p> <bold>Grid plot of a simulated, subclonal tumor sample. (a)</bold> Genome integer CNs in the aneuploid subclone A. <bold>(b)</bold> Genome integer CNs in the less variable subclone B. <bold>(c)</bold> Grid plot from sample with 20% normal cell contamination, 30% cells from subclone A and 50% cells from subclone B. In grid plots each data point represents average CNs of a genome segment. Different colors represent genome segments with different behaviors in terms of their average CNs across the sample cells. In this grid plot each third of the genome results in a separate grid pattern (blue for subclone B integer CNs (1,1), black for (0,1) or red for (1,2)) their size determined by the fraction of subclone A cells. The three grids are positioned in a larger grid for which the size is determined by the larger fraction of subclone B cells (green).</p></caption><graphic xlink:href="13059_2014_470_Fig2_HTML" id="MO2"/></fig><disp-formula id="Equ2"><label>2</label><alternatives><tex-math id="M2">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ e=\left(1-p\right)+\alpha {c}_a+\beta {c}_b, $$\end{document}</tex-math><graphic xlink:href="13059_2014_470_Article_Equ2.gif" position="anchor"/></alternatives></disp-formula></p><p>where <italic>&#x003b1;</italic>&#x02009;+&#x02009;<italic>&#x003b2;</italic>&#x02009;=&#x02009;<italic>p</italic> and <italic>c</italic>_{<italic>a</italic>} and <italic>c</italic>_{<italic>b</italic>} are the integer CNs of the subclone A and B cells in that segment. The grid plot for the sample in Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>c consists of three small grids, each of which originates from the CNs in A combined with the CNs in one of the three B segments. The size, or rather the density, of the small grids is due to the small fraction of cells in A. The positions of the small grids follow that of a more sparse grid, determined by the larger fraction of cells in B. Alternatively, the grid plot could be seen as many three-point sparse grids (the green circles being one of them), positioned according to the denser pattern of subclone A.</p><p>For one subclone (say A), the cell fraction (<italic>&#x003b1;</italic>) can be estimated from the perfect, noise-free average CNs via Equation&#x000a0;<xref rid="Equ2" ref-type="">2</xref> if <italic>Condition 1:</italic> There are at least two points in the grid plot for which the difference is known (on an integer CN scale) and known to be due only to a change of integer CNs in subclone A (so that all other subclones have constant CNs throughout these two segments).</p><p>Given the cell fraction of subclone A, its integer CNs can be estimated from unbiased, noise-free average CNs via Equation&#x000a0;<xref rid="Equ2" ref-type="">2</xref> (or its extensions to more than two subclones) if <italic>Condition 2:</italic> The integer CNs and cell fractions of all subclones other than A of the sample are also known.</p><p>Condition 2 seems to be a catch 22 in that no subclonal integer CNs can be estimated without knowing the integer CNs of the other subclones, but there is an important exception. If the grid pattern suggested by condition 1 includes the point (1,1), the integer CNs of all other subclones are normal (1,1) in the genome segments of the points for which condition 1 is true. However, condition 1 is seldom truly known for any points (Figure&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>a). Therefore, integer CN estimation in subclonal tumor samples can only be done from noise-free average CNs if the sample has certain properties and under certain assumptions. In <xref rid="Sec24" ref-type="sec">Materials and methods</xref> we further demonstrate CN estimation challenges caused by selected subclonal structures through Figure&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>, and outline properties and assumptions under which subclonal CNs can be estimated.<fig id="Fig3"><label>Figure 3</label><caption><p> <bold>Grid plots of simulated tumor samples with different subclonal architectures. (a)</bold> Aneuploid tumor with further subclonality in a small part of the genome. <bold>(b)</bold> Tumor with three subclones, all with CN alterations. <bold>(c)</bold> Tumor with two subclones of equal size. <bold>(d)</bold> Tumor with two subclones of different sizes. Note that each data point in a grid plot represents the average CN across all sample cells. Different colors do not represent different subclones, but highlight specific parts of the genome which we discuss further in <xref rid="Sec24" ref-type="sec">Materials and methods</xref>.</p></caption><graphic xlink:href="13059_2014_470_Fig3_HTML" id="MO3"/></fig></p></sec><sec id="Sec6"><title>Purity versus cell fraction</title><p>CN alterations in tumor cells appear diluted in average CNs because of the germline (normal) cells in the sample (Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>), which are always present. If a sample is known to be clonal, the purity of the sample can be deduced from the density of an observed average CN grid pattern via Equation 1: the distance between consecutive grid points is equal to the purity. However, when we study tumor samples we usually do not know whether or not that sample is clonal. In this case, as acknowledged by Durinck <italic>et al</italic>. [<xref ref-type="bibr" rid="CR2">2</xref>], further, indistinguishable dilution occurs when CN alterations are present only in part, but not all, of the tumor cells (Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>). With or without heterogeneity, the density of a grid pattern in a grid plot holds information about the cell fraction which express CN alteration throughout some genome segment(s) in which other subclones have constant CNs: the distance between consecutive grid points is equal to that cell fraction (Equation&#x000a0;<xref rid="Equ2" ref-type="">2</xref>). Average CNs do not carry sufficient information to deduce sample purity, although it is sometimes suggested that they do [<xref ref-type="bibr" rid="CR1">1</xref>,<xref ref-type="bibr" rid="CR22">22</xref>].</p></sec><sec id="Sec7"><title>Scaling: where is (1,1)?</title><p>Summarizing the preceding discussion, cell fractions and integer CNs can be quantified from unbiased, noise-free average CNs for some subclones and for some genome segments, if the tumor sample has some fortunate properties and we rely on a set of assumptions. Array CNs are at best proportional to the average CNs in the sample hybridized. Even if they were noise and bias free, array CNs are insufficient for determination of an identified subclone&#x02019;s integer CNs, its cell fraction and the scaling factor without further information [<xref ref-type="bibr" rid="CR1">1</xref>,<xref ref-type="bibr" rid="CR22">22</xref>,<xref ref-type="bibr" rid="CR23">23</xref>]. The colored points in the simulated grid plot of Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref> appear in a regular grid pattern as marked by dotted lines, but it is unknown which lattice point corresponds to integer copies (1,1): (<italic>g</italic>₂, <italic>g</italic>₂), (<italic>g</italic>₃, <italic>g</italic>₃) or (<italic>g</italic>₄, <italic>g</italic>₄)? Each of the colored points must have at least 0 minor integer copies. Therefore, the grid pattern suggests that g₄ is at least 1, or that (1,1) falls no higher than (<italic>g</italic>₄, <italic>g</italic>₄). In <xref rid="Sec24" ref-type="sec">Materials and methods</xref> we explain how the three proposed (1,1) scenarios originate from different integer CNs, cell fractions and scaling factors but result in identical array CNs, or, equivalently, identical total (minor&#x02009;+&#x02009;major) array CNs and B allele fractions.<fig id="Fig4"><label>Figure 4</label><caption><p> <bold>Grid plot and expected VAF levels of simulated tumor sample. (a)</bold> Grid plot of simulated, noise- and bias-free array CNs. The scale of the array CNs is unknown. The four colored points suggest the grid pattern drawn for subclone A, but it is unknown whether (1,1) integer copies happen at (<italic>g</italic> ₂, <italic>g</italic> ₂), (<italic>g</italic> ₃, <italic>g</italic> ₃) or (<italic>g</italic> ₄, <italic>g</italic> ₄). <bold>(b-d)</bold> The three scenarios are illustrated., Each scenario has one colored column for each of the colored genome segments. Two of the segments have mutations on them (red crosses), with VAFs as shown on the y-axes (simulated without noise or bias). The different scalings suggest different integer CNs (<italic>c</italic> ₁, <italic>c</italic> ₂) of the segments (labels on x-axes), which give different potential expected VAF levels under certain assumptions (horizontal lines). The mutation on the blue segment does not fit any suggested VAF level in (d), suggesting <italic>g</italic> ₄&#x02009;&#x02260;&#x02009;1. Assuming that some VAF levels are more plausible than others (thick rather than thin horizontal lines) also rules out (b) in favor of (c): <italic>g</italic> ₃&#x02009;=&#x02009;1 (see <xref rid="Sec24" ref-type="sec">Materials and methods</xref>).</p></caption><graphic xlink:href="13059_2014_470_Fig4_HTML" id="MO4"/></fig></p><p>Pounds <italic>et al</italic>. [<xref ref-type="bibr" rid="CR23">23</xref>] suggest solving this issue by identification of genomic regions with normal CNs (RAP, reference alignment procedure), which may be possible for some samples but not for all. In the context of heterogeneity, VAFs can sometimes provide sufficient information, and these are used in the software Absolute [<xref ref-type="bibr" rid="CR1">1</xref>] together with database knowledge about chromosome arm level CN alterations in common cancer types. Carter <italic>et al</italic>. [<xref ref-type="bibr" rid="CR1">1</xref>] and Pounds <italic>et al</italic>. [<xref ref-type="bibr" rid="CR23">23</xref>] both stress that manual care with each sample is vital for correct CN estimation. We examine circumstances under which knowledge of overall sample ploidy, matched normal sample array CNs or VAFs can resolve the scaling issue below.</p><sec id="d30e1143"><title><italic>Ploidy can sometimes help</italic></title><p>A sample&#x02019;s overall ploidy is the sample&#x02019;s average (minor&#x02009;+&#x02009;major) integer CN across the genome and across all subclones. In <xref rid="Sec24" ref-type="sec">Materials and methods</xref> we explain how an independent overall ploidy estimate (for example, from a fluorescence-activated cell sorting (FACS) run) may help us resolve the true position of (1,1). Often, overall ploidy estimates are not given by FACS, but with samples having simple subclonal architecture we may compare subclone-specific ploidies estimated for each potential position of (1,1) (<xref rid="Sec24" ref-type="sec">Materials and methods</xref>) to the suggested subclone ploidies from FACS, and deduce the true position of (1,1). Figure&#x000a0;<xref rid="Fig5" ref-type="fig">5</xref> shows FACS ploidy profiles and segmented SNP array data grid plots for two samples. Sample 11 (Figure&#x000a0;<xref rid="Fig5" ref-type="fig">5</xref>a,b) has several subclones and integer CNs cannot be located to specific subclones. Sample 29 (Figure&#x000a0;<xref rid="Fig5" ref-type="fig">5</xref>c,d) has most of its CN alteration in one subclone and the FACS and grid plots combined give clues to the scaling of array CNs.<fig id="Fig5"><label>Figure 5</label><caption><p> <bold>FACS ploidy profiles.</bold> FACS ploidy profiles (upper panels, number of cells versus cell ploidy) and grid plots (lower panels) of three actual data samples. Peaks at 50&#x000a0;K in the FACS profile correspond to diploid cells. <bold>(a)</bold> Sample 11 has multiple subclones suggested by multiple peaks in the FACS profile. Some peaks comprise approximately the same cell fraction (peak height). <bold>(b)</bold> A grid pattern appears in the sample 11 grid plot, but since the FACS profile reveals several subclones of the same size, assumption 1 is not reasonable and integer CNs cannot be located to specific subclones: CN alteration which agrees with the lattice points could originate from any of the subclones of the corresponding size. <bold>(c)</bold> The sample 29 FACS profile suggests that the largest non-diploid subclone has ploidy 3 to 4. <bold>(d)</bold> The sample 29 grid plot suggests a subclone with ploidy 3.05 if <italic>g</italic> ₆&#x02009;=&#x02009;1, ploidy 5.05 if <italic>g</italic> ₅&#x02009;=&#x02009;1, ploidy 7.05 if <italic>g</italic> ₄&#x02009;=&#x02009;1, and so on. The FACS profile and grid plot thereby together suggest that <italic>g</italic> ₆&#x02009;=&#x02009;1 for sample 29. DNA Index (DI) is a measurement of ploidy.</p></caption><graphic xlink:href="13059_2014_470_Fig5_HTML" id="MO5"/></fig></p></sec><sec id="d30e1215"><title><italic>Paired normal SNP array normalization helps in theory</italic></title><p>If array CNs have been normalized towards matched normal tissue SNP array CNs, segments with minor&#x02009;+&#x02009;major array CNs equal to 1 (red lines in the example samples of Figure&#x000a0;<xref rid="Fig6" ref-type="fig">6</xref>) and allelic balance (black lines in Figure&#x000a0;<xref rid="Fig6" ref-type="fig">6</xref>) - that is, segments at the intersection of the two lines - should theoretically correspond to normal integer CNs (1,1). Several CN packages (SOMATICS [<xref ref-type="bibr" rid="CR24">24</xref>], PICNIC [<xref ref-type="bibr" rid="CR25">25</xref>], SiDCoN [<xref ref-type="bibr" rid="CR26">26</xref>], GAP [<xref ref-type="bibr" rid="CR27">27</xref>] and ASCAT [<xref ref-type="bibr" rid="CR22">22</xref>]) rely on normalized array CNs and assume the solution with the minimal possible CNs that fit their (grid) pattern. ASCAT notes that they go wrong if that assumption is not correct. A look ahead at our actual data grid plots (Figure&#x000a0;<xref rid="Fig6" ref-type="fig">6</xref>) suggests that this method will not work in general for our samples.<fig id="Fig6"><label>Figure 6</label><caption><p> <bold>Array CNs normalized towards matched normal tissue SNP array CNs for four actual data samples.</bold> Theoretically, points in a cluster around the intersection of allelic balance (black line) and total array CN equal to 1 (red line) should reflect segments with integer CNs (1,1). In practice, the cluster closest to the intersection may reflect (1,1) (samples 5 and 9) but not always (samples 11 and 45).</p></caption><graphic xlink:href="13059_2014_470_Fig6_HTML" id="MO6"/></fig></p></sec><sec id="d30e1258"><title><italic>VAFs can sometimes help</italic></title><p>For samples with a reasonable amount of mutations in &#x02018;informative&#x02019; locations, VAFs can help deduce the scaling of array CNs if we rely on a set of chosen assumptions. We outline such a framework in <xref rid="Sec24" ref-type="sec">Materials and methods</xref> through Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>.</p></sec></sec><sec id="Sec8"><title>Estimation of cell fraction, integer CNs and average CNs</title><p>The cell fraction of subclone A and its integer CNs in genome segments that coincide with the subclone&#x02019;s lattice points (type A segments) can be estimated under the fortunate circumstances described, given the properly scaled array CNs. Let (<italic>g</italic>_{<italic>normal</italic>}, <italic>g</italic>_{<italic>normal</italic>}) be the position of (1,1) and &#x00394; the distance between two consecutive grid lines. Then we can derive the cell fraction <italic>&#x003b1;</italic>&#x02009;=&#x02009;&#x00394;/<italic>g</italic>_{<italic>normal</italic>} of subclone A and its integer CNs <italic>c</italic>₁&#x02009;=&#x02009;(<italic>a</italic>₁&#x02009;&#x02212;&#x02009;<italic>g</italic>_{<italic>normal</italic>}&#x02009;&#x02212;&#x02009;&#x00394;)/&#x00394; and <italic>c</italic>₂&#x02009;=&#x02009;(<italic>a</italic>₂&#x02009;&#x02212;&#x02009;<italic>g</italic>_{<italic>normal</italic>}&#x02009;&#x02212;&#x02009;&#x00394;)/&#x00394;, where <inline-formula id="IEq1"><alternatives><tex-math id="M3">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ {\overset{\rightharpoonup }{a}}_j=\left({a}_1,{a}_2\right) $$\end{document}</tex-math><inline-graphic xlink:href="13059_2014_470_Article_IEq1.gif"/></alternatives></inline-formula> are minor and major array CNs of unknown scale. The average CNs can be derived as <italic>e</italic>₁&#x02009;=&#x02009;<italic>a</italic>₁/<italic>g</italic>_{<italic>normal</italic>} and <italic>e</italic>₂&#x02009;=&#x02009;<italic>a</italic>_{<italic>e</italic>}/<italic>g</italic>_{<italic>normal</italic>}.</p></sec></sec><sec id="Sec9"><title>B: Clonal or subclonal mutations</title><p>With cell fractions and integer CNs of one or more subclones resolved and with knowledge of the sample purity, we can assess whether a VAF suggests the corresponding mutation is clonal (present in all tumor cells) or subclonal (not present in all tumor cells) if we rely on previously outlined and further properties and assumptions (see <xref rid="Sec24" ref-type="sec">Materials and methods</xref>).</p><p>The simulated sample of Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref> provides an example. Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>c gives the integer CNs and cell fraction of the sample&#x02019;s main subclone A. The genome segment of the blue grid point (Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>a) has no CN alteration in any cells. The blue column in Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>c shows expected VAF levels of heterozygous mutations present only in the cells of the main subclone (thick continuous horizontal line), present only in all other tumor cells (bottom thick dashed line) or present in all cells (top thick dashed horizontal line) of such genome segments. The observed VAF (red cross) coincides with the latter, so the corresponding mutation is estimated to be clonal. The red genome segment (Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>a) has CNs (1,2) in the main subclone, and normal CNs in all other cells. The pink column of Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>c shows the expected VAF levels given these CNs (thick horizontal lines). The top two thick dashed horizontal lines reflect expected VAF levels of clonal heterozygous mutations present on all its homologue&#x02019;s copies. The mutation on this segment is hence estimated to be clonal too. If it had coincided with one of the lower horizontal lines, we would have estimated it to be subclonal.</p><p>The four colored columns of Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>c show different expected VAF levels of clonal mutations (top one or two thick, dashed horizontal lines, one or two depending on whether the minor and major integer CNs are equal or not), and different expected VAF levels of mutations present only in the main subclone A (thick, solid horizontal lines), resulting from different local integer CNs. We also note that other cell fractions of subclone A, together with other integer CNs (Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>b,d), would give other expected VAF levels. Two important conclusions follow. First, in order to enable classification of mutations as clonal or subclonal from VAFs with any precision, correct estimation of subclonal integer CNs and cell fractions is vital. (The procedure will still rely on simplifying assumptions, even for mutations on fortunate segments on grid plot lattice points of an identifiable subclone, and when there is no noise or bias in VAFs or segmented CN data.) Second, one subclone is associated with a whole set of expected VAF levels, dependent on the subclone&#x02019;s cell fraction and integer CNs, for example, the thick continuous horizontal lines in Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>b-d. This contrasts with what has sometimes been suggested [<xref ref-type="bibr" rid="CR8">8</xref>]. We return to this point in the Discussion.</p></sec><sec id="Sec10"><title>C: Data examples</title><p>In this section we illustrate what we learned in the previous sections through selected analyses of single tumor and matched normal samples from a set of 52 newly diagnosed <italic>HER2</italic>-positive breast cancer tumors. The patients were all part of a European Union funded project (RESPONSIFY) investigating biomarkers of resistance to trastuzumab plus chemotherapy, which is standard treatment for newly diagnosed breast cancers that have <italic>HER2</italic> amplification. Clinical follow-up data are available for each patient through a median time of 5&#x000a0;years, including relapse status. SNP arrays and WES were run as described in <xref rid="Sec24" ref-type="sec">Materials and methods</xref>. Tumor sample purities (fractions of tumor cells) were estimated by a pathologist. The median purity was 87% across tumors. For details about SNP array preprocessing and detection of mutations, see <xref rid="Sec24" ref-type="sec">Materials and methods</xref>.</p><sec id="Sec11"><title>Bias in BAF can cause skewness in SNP array data</title><p>Our observed array CN grid plots display skewness (Figure&#x000a0;<xref rid="Fig7" ref-type="fig">7</xref>), so that segments with the same minor CN appear in clusters on a sloping rather than a horizontal line, and segments which have the same major CN appear in clusters on a sloping rather than a vertical line. This phenomenon is particularly pronounced in those of our samples that do not have matched normal sample SNP array data. This is an artifact caused by a systematic bias in our SNP array BAFs which needs to be removed in order to make the CN estimates comparable to WES VAFs. We do this by grid rotation and describe the BAF bias (see <xref rid="Sec24" ref-type="sec">Materials and methods</xref>). Unless otherwise stated, we refer to array CNs as rotated array CNs.<fig id="Fig7"><label>Figure 7</label><caption><p> <bold>Sample 45 grid plots.</bold> For the description of notation see <xref rid="Sec24" ref-type="sec">Materials and methods</xref>. <bold>(a)</bold> Original array CNs with skewness. <bold>(b)</bold> Array CNs after pre-rotation in search of start values (see <xref rid="Sec24" ref-type="sec">Materials and methods</xref>). <bold>(c)</bold> Array CNs after completed grid rotation, which removes the skewness. Unless otherwise stated, we refer to array CNs as rotated array CNs and we drop the prime from <inline-formula id="IEq2"><alternatives><tex-math id="M4">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ {\overset{\rightharpoonup }{\mathrm{a}}}^{\hbox{'}} $$\end{document}</tex-math><inline-graphic xlink:href="13059_2014_470_Article_IEq2.gif"/></alternatives></inline-formula>.</p></caption><graphic xlink:href="13059_2014_470_Fig7_HTML" id="MO7"/></fig></p></sec><sec id="Sec12"><title><italic>A typical</italic> HER2+ <italic>grid plot</italic></title><p>Most of our <italic>HER2</italic>-positive breast cancer sample grid plots are similar to Figure&#x000a0;<xref rid="Fig8" ref-type="fig">8</xref>, which suggests that they have an aneuploid subclone A in a small fraction of cells (because the regular grid pattern is small). Many of the grid plots also have some short segments with minor array CNs below the most prominent grid pattern, as in Figure&#x000a0;<xref rid="Fig8" ref-type="fig">8</xref>, which suggests that a larger subclone than the aneuploid one has some LOH.<fig id="Fig8"><label>Figure 8</label><caption><p> <bold>Sample 5 grid plot.</bold> The small, regular grid suggests that a low fraction of the sample cells form an aneuploid subclone (A). The segments with minor array CNs far below the regular grid may be caused by LOH in a different, larger subclone. Segments are classified into types A (blue, regular grid pattern), B (pink, breaking regular grid pattern), C (green, lower array CNs than the most evident grid pattern) and D (red, high enough array CNs for regular grid patterns not to appear).</p></caption><graphic xlink:href="13059_2014_470_Fig8_HTML" id="MO8"/></fig></p></sec><sec id="Sec13"><title>Probabilistic model to separate subclonality from noise and a simple endpoint quantifying ITH</title><p>In Figure&#x000a0;<xref rid="Fig8" ref-type="fig">8</xref> we identify a regular grid pattern (type A segments, blue), possibly caused by CN variation in a subclone, say A, of cells. We also spot array CNs that do not follow the grid pattern, in between the regular lattice points (type B segments, pink). In general we see lattice points (type A) as the default location of grid plot points, and it is only if we observe significant evidence to the contrary that we set the type of a segment to B according to the following process.</p><p>The classification between type A and B segments is made through the two-dimensional distribution of grid points <inline-formula id="IEq3"><alternatives><tex-math id="M5">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ \left\{{\overset{\rightharpoonup }{a}}_i\right\} $$\end{document}</tex-math><inline-graphic xlink:href="13059_2014_470_Article_IEq3.gif"/></alternatives></inline-formula> relative to their closest lattice points <inline-formula id="IEq4"><alternatives><tex-math id="M6">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ \left\{{\overset{\rightharpoonup }{e}}_i\right\} $$\end{document}</tex-math><inline-graphic xlink:href="13059_2014_470_Article_IEq4.gif"/></alternatives></inline-formula>, in effect overlaying all the lattice points into <inline-formula id="IEq5"><alternatives><tex-math id="M7">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ \left\{{\overset{\rightharpoonup }{x}}_i={\overset{\rightharpoonup }{a}}_i-{\overset{\rightharpoonup }{e}}_i\right\} $$\end{document}</tex-math><inline-graphic xlink:href="13059_2014_470_Article_IEq5.gif"/></alternatives></inline-formula> (Figure&#x000a0;<xref rid="Fig9" ref-type="fig">9</xref>a). We fit a two-dimensional <italic>t</italic>-distribution [<xref ref-type="bibr" rid="CR28">28</xref>] centered at the origin to the <inline-formula id="IEq6"><alternatives><tex-math id="M8">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ \left\{{\overset{\rightharpoonup }{x}}_i\right\} $$\end{document}</tex-math><inline-graphic xlink:href="13059_2014_470_Article_IEq6.gif"/></alternatives></inline-formula>, with maximum robustness (degrees of freedom =2) in order to capture the variation of observations in the dense central cluster (which may truly have CN alteration in subclone A only) but not that of the many outliers (which may not originate from CN alteration in subclone A). The estimated covariance matrix <italic>Q</italic> is used to calculate a segment length-weighted squared Mahalanobis distance <inline-formula id="IEq7"><alternatives><tex-math id="M9">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ {M}_i^2={\overset{\rightharpoonup }{x}}_i^T{\left(Q/{w}_i\right)}^{-1}{\overset{\rightharpoonup }{x}}_i $$\end{document}</tex-math><inline-graphic xlink:href="13059_2014_470_Article_IEq7.gif"/></alternatives></inline-formula> for each segment <italic>i</italic>, which should follow an exponential distribution with scale parameter &#x000bd; for segments within the dense centre cluster. We choose a cutoff <italic>m</italic> where the linearity in the exponential <italic>qq</italic>-plot starts to fail (Figure&#x000a0;<xref rid="Fig9" ref-type="fig">9</xref>b; commonly conservatively chosen to <italic>m</italic> =3), and classify segments as type A if <inline-formula id="IEq8"><alternatives><tex-math id="M10">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ {M}_i^2\le m $$\end{document}</tex-math><inline-graphic xlink:href="13059_2014_470_Article_IEq8.gif"/></alternatives></inline-formula> and type B otherwise. The segment length weight <inline-formula id="IEq9"><alternatives><tex-math id="M11">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ {w}_i=1-{e}^{-{l}_i/500000} $$\end{document}</tex-math><inline-graphic xlink:href="13059_2014_470_Article_IEq9.gif"/></alternatives></inline-formula>, where <italic>l</italic>_{<italic>i</italic>} is the length of segment <italic>i</italic>, downweights <inline-formula id="IEq10"><alternatives><tex-math id="M12">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ {M}_i^2 $$\end{document}</tex-math><inline-graphic xlink:href="13059_2014_470_Article_IEq10.gif"/></alternatives></inline-formula> values of short (&#x0003c;1&#x000a0;Mb) segments, since we think their deviance from the origin may be due to noise in the array CNs of such small segments, rather than to a true pattern-breaking deviance in CNs.<fig id="Fig9"><label>Figure 9</label><caption><p> <bold>Noise or heterogeneity in sample 5. (a)</bold> Positions of array CNs relative to their closest grid plot lattice points, with coloring by statistical distance. Note that positions have been scaled to fit the lattice point and its closest segments within &#x000b1;0.5. <bold>(b)</bold> Exponential(&#x000bd;) <italic>qq</italic>-plot with the same coloring as (a). Points above the horizontal cutoff (<italic>m</italic>) indicate segments with array CNs significantly different from the lattice points (type B segments).</p></caption><graphic xlink:href="13059_2014_470_Fig9_HTML" id="MO9"/></fig></p><p>The fraction of the genome covered by type B segments out of that covered by type A and B segments is a simple measure of the amount of ITH in a sample. This endpoint estimates the fraction of the genome in which the sample has CN alteration in other subclones than the main subclone A (possibly in addition to CN alteration in A). It has proved useful for prediction of relapse in the RESPONSIFY samples. Details will be published separately.</p></sec><sec id="Sec14"><title>Scaling: resolving location of (1,1) with help of VAFs and purity</title><p>We estimate the scaling of a sample&#x02019;s array CNs by the scenario that best fits VAFs estimated from WES data for mutations in segments classified to be of type A with respect to the sample&#x02019;s most evident subclone A. By our assumptions, these segments have CN variation only in the subclone A cells. Out of the 52 RESPONSIFY samples the scaling was resolved in this manner for 48 samples.</p><p>In Figure&#x000a0;<xref rid="Fig10" ref-type="fig">10</xref> we display a typical example rather than a perfect one (as, for example, that of Figure&#x000a0;<xref rid="Fig11" ref-type="fig">11</xref> below). Each panel shows the five expected VAF levels (y-axis, horizontal lines with different colors) for each type A CN segment (x-axis, ordered by decreasing expected VAF if present only on non-A cells and by increasing minor&#x02009;+&#x02009;major array CN) for one potential position of (1,1) of the sample introduced in Figure&#x000a0;<xref rid="Fig8" ref-type="fig">8</xref> under assumptions 3 to 6 in <xref rid="Sec24" ref-type="sec">Materials and methods</xref>. The observed mutation VAFs of type A segments are shown as red crosses. Each panel also gives SS, the sum of squared distances to each VAF&#x02019;s nearest expected VAF level. The figure suggests that <italic>g</italic>₃&#x02009;=&#x02009;1 or <italic>g</italic>₆&#x02009;=&#x02009;1, since in these panels the observed VAFs are, on average, closer to their expected levels (they have lower SS than the other panels). Note that we do not expect all mutations in type A segments to follow our assumptions and fit one of the expected levels, but we assume that most mutations do, in order to resolve the scaling of the array CNs.<fig id="Fig10"><label>Figure 10</label><caption><p> <bold>Sample 5 VAFs compared to expected VAF levels.</bold> As in Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>b-d, each panel shows observed VAFs (red crosses) and expected VAF levels given a potential position of (1,1) in the sample grid plot. Expected VAF levels are all under assumption 6, with mutations on the minor allele of subclone A expected on green VAF levels, on the major allele of subclone A expected on blue, on the single copy of all tumor cells not in A expected on black (the purity of this sample was estimated to 84%), clonal mutations on minor allele on light green and clonal mutations on major allele on light blue horizontal lines (see assumptions 3 to 5 in <xref rid="Sec24" ref-type="sec">Materials and methods</xref>). Genome segments have been ordered by decreasing expected VAF if present only on non-A cells and by increasing (minor&#x02009;+&#x02009;major) array CN (x-axis). SS gives the sum of squared distances from each observed VAF to its closest expected VAF.</p></caption><graphic xlink:href="13059_2014_470_Fig10_HTML" id="MO10"/></fig><fig id="Fig11"><label>Figure 11</label><caption><p> <bold>Sample 16 VAFs fit expected VAFs along chromosome 17.</bold> Sample 16 VAFs (points) and expected VAF levels (horizontal lines; Figure&#x000a0;<xref rid="Fig10" ref-type="fig">10</xref>) along chromosome 17 given subclone A CN estimates and its sample cell fraction <italic>&#x003b1;</italic>. This sample has 1,232 detected mutations, which is many more than the median 156 of the 52 RESPONSIFY samples. We see an almost perfect fit of the observed VAFs to those expected. For color coding of VAFs, see text. This figure also shows how the mutation rate differs across the chromosome, a different type of heterogeneity studied by Lawrence <italic>et al</italic>. [<xref ref-type="bibr" rid="CR29">29</xref>]. The purity of this sample is unknown and is imputed to 90%.</p></caption><graphic xlink:href="13059_2014_470_Fig11_HTML" id="MO11"/></fig></p><p>To further differentiate between the two suggested scenarios we transform these panels&#x02019; y-levels to show subclone A integer CN estimates under these scenarios (Figure&#x000a0;<xref rid="Fig12" ref-type="fig">12</xref>), also showing the segments by their genome position. Now, the (dark) green and blue horizontal lines are the minor and major integer CN estimates of subclone A for segments that have CN alteration only in subclone A. The red crosses&#x02019; y-levels show the mutation multiplicities (calculated under the assumption that they sit on A: <inline-formula id="IEq11"><alternatives><tex-math id="M13">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ \frac{D}{\alpha }VAF $$\end{document}</tex-math><inline-graphic xlink:href="13059_2014_470_Article_IEq11.gif"/></alternatives></inline-formula> ), which equal an integer CN estimate (green or blue horizontal line) if the mutation VAF equals the corresponding expected VAF level. The black line shows the expected y-level of the multiplicity for mutations on the single copy of all tumor cells not in A, which is equal across all segments. The light green and light blue horizontal lines show the expected y-levels of clonal mutation multiplicities. We see that if <italic>g</italic>₃&#x02009;=&#x02009;1, subclone A has no single allele integer CN below 3, a rough calculation (see <xref rid="Sec24" ref-type="sec">Materials and methods</xref>) suggests the sample has overall ploidy above 4, and all mutations seem to sit only on the assumed diploid cells (black horizontal line), not in subclone A. If <italic>g</italic>₆&#x02009;=&#x02009;1, subclone A has single allele integer CNs from 0 and above, the rough overall ploidy estimate is just over 2 and most mutations seem to sit on the subclone A cells. The majority of our samples end with a similar choice. The <italic>g</italic>₆&#x02009;=&#x02009;1 scenario sounds more reasonable and therefore we choose to proceed with that. When in doubt we choose the conservative scenario with the smallest integer CNs and smallest size <italic>&#x003b1;</italic> of subclone A.<fig id="Fig12"><label>Figure 12</label><caption><p> <bold>Mutation multiplicities compared to subclonal CNs.</bold> Sample 5 VAF fits to array CN data of type A segments and their mutations for the two most plausible positions of (1,1): <italic>g</italic> ₃&#x02009;=&#x02009;1 or <italic>g</italic> ₆&#x02009;=&#x02009;1. Each scenario results in different minor (green) and major (blue) subclone A integer CN estimates, and multiplicities of the mutations if these sit on A cells only (red). Light and black horizontal lines show expected y-levels of clonal mutations and mutations present on only non-A cells, all according to assumptions 3 to 6. For details, see text.</p></caption><graphic xlink:href="13059_2014_470_Fig12_HTML" id="MO12"/></fig></p><p>See <xref rid="Sec24" ref-type="sec">Materials and methods</xref> for our suggested estimation of subclonal architecture, cell fractions and integer CNs.</p></sec><sec id="Sec15"><title>Clonal or subclonal mutations</title><p>Our classification of mutations as clonal or subclonal is based on the methods outlined for simulated data. To acknowledge the uncertainty of real VAFs we run a set of non-inferiority, inferiority and equality tests for each VAF based on its binomial two-sided 90% confidence interval (CI) from the sequencing number of variant versus reference reads. For details, see <xref rid="Sec24" ref-type="sec">Materials and methods</xref>.</p></sec><sec id="Sec16"><title>Clonal or subclonal CN alterations</title><p>The vertical continuous line in Figure&#x000a0;<xref rid="Fig11" ref-type="fig">11</xref> denotes the position of the <italic>HER2</italic> (<italic>ERBB2</italic>) gene. Our samples have been diagnosed as <italic>HER2</italic>-enriched, and they do have a high, type D, major array CN at this position. Unfortunately no samples have VAFs that match the major homologue, so it is not possible to assign the clonality status or subclonal origin of <italic>HER2</italic> enrichment. The number of detected mutations in the 52 RESPONSIFY samples varies from 1 to 1,232 (median 156.5), and only a handful of samples have enough mutations (say &#x0003e;900) to enable assessment of subclonal origin of CN alteration on a large scale.</p></sec></sec></sec><sec id="Sec17" sec-type="discussion"><title>Discussion</title><p>The aim of this paper is to highlight challenges in CN estimation that influence mutation classification but are infrequently acknowledged in the literature as well as propose solutions that may aid in the quantification of ITH in tumor samples that have high levels of CN alteration. We have demonstrated how even with no noise or bias, integer CNs of tumor samples with ITH can only be estimated from segmented CN data in samples with a simple clonal architecture, given further information from, for example, WES VAFs or FACS images, and under a series of assumptions. Even with such samples, integer CNs can only be deduced for some subclones and only across a subset of the genome.</p><p>Classification of mutations as clonal or subclonal further requires knowledge of the sample purity, which cannot be deduced from segmented CN data. The classification relies on comparing observed VAFs to expected VAF levels given purity, subclonal cell fractions and local CNs. Therefore, the assumptions made in the CN estimation procedure will have a large influence on how mutations are classified, and on how the results can be interpreted.</p><p>We have also suggested a simple ITH endpoint for tumor samples with a high level of CN alterations based on segmented CN data alone and which does not require knowledge of subclonal cell fractions or integer CNs.</p><sec id="Sec18"><title>Estimate average CNs from sequencing</title><p>We have used SNP arrays to derive segmented CN data for the RESPONSIFY samples. Alternatively, sequencing depths could be used [<xref ref-type="bibr" rid="CR12">12</xref>&#x02013;<xref ref-type="bibr" rid="CR17">17</xref>], which has the advantage that it works well on formalin-fixed paraffin-embedded tissue, whereas SNP arrays usually require frozen tissue, which is less practical to validate. The expected (true) average CN patterns of tumor samples with heterogeneity are the same whether average CNs are estimated from SNP arrays or sequencing data. The challenges presented hence apply either way: there is ambiguity between purity and heterogeneity, there are difficulties deducing subclonal structures and assigning a subclonal origin to a segment with CN alteration, and the scaling of array CNs or SNP position sequencing depths relative to average CNs is unknown. Both sequencing and SNP array data may suffer from bias which needs attention before estimation of average CNs, although the types of bias are different. There may be BAF bias in SNP array data and GC bias in sequencing depths. Standardizing tumor sample sequencing depths to matched normal sample sequencing depths comes with challenges that are different from those of standardizing tumor sample array CNs to matched normal sample array CNs. We generally seek more evidence that the results after different steps of analysis look plausible than is typically presented in a literature dominated by model-based inferences. We find that just as important as detailed model descriptions. As for CN determination, a study of the two-dimensional grid plots (applicable equally well to SNP array CNs and sequencing depths) of average CN estimates can help reveal bias and give clues to sample architecture.</p></sec><sec id="Sec19"><title>Whole genome sequencing versus whole exome sequencing</title><p>WGS identifies many more mutations than WES (which can only find mutations in gene exons), but is comparatively more expensive. More mutations help in assigning CN alterations to identified subclones, and resolving the scaling of segmented CN data in relation to average CNs. Therefore, WGS is generally a benefit for assessment of integer CNs, clonality of mutations or phylogenetic trees (see below) in subclonal tumor samples.</p></sec><sec id="Sec20"><title>Phylogenetic trees</title><p>Given a set of identified subclones in a sample, trees can be inferred by assigning mutations to subclones and checking whether mutations close in genomic location but assigned to different subclones tend to co-appear or never co-appear on the same fragment. Co-appearance indicates that one of the subclones is in turn a subclone of the other, and no co-appearance indicates that the subclones belong to independent branches of the tree. Given that WGS identifies more mutations than WES, WGS is again a benefit. Since the majority of our samples have too complicated subclonal structures for more than one or two subclones to be identified in detail, and relatively few mutations identified by WES, detailed phylogenetic trees are not generally within reach. The number of identified mutations in the 52 RESPONSIFY samples ranges from 1 to 1,232 (median 156.5; samples were selected so that they had at least one identified mutation). The WES average coverage of the samples ranges from 25 to 179, with median 108.</p></sec><sec id="Sec21"><title>Clustering of cancer cell fractions</title><p>It has been suggested [<xref ref-type="bibr" rid="CR8">8</xref>,<xref ref-type="bibr" rid="CR9">9</xref>,<xref ref-type="bibr" rid="CR12">12</xref>&#x02013;<xref ref-type="bibr" rid="CR17">17</xref>] that, with WGS, subclones can be identified via groups of mutations present in similar fractions of cancer cells. On this topic we would first like to stress that clustering of a sample&#x02019;s VAFs is something different from clustering of the sample&#x02019;s cancer cell fractions. The former may cluster because of aneuploidy in the sample, even if the sample has no heterogeneity: a sample with aneuploidy has several expected VAF levels (like the thick continuous horizontal lines in Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>c), so each subclone corresponds to several VAF clusters. Also, different subclones may have overlapping expected VAF levels.</p><p>To the best of our knowledge, Papaemmanuil <italic>et al</italic>. [<xref ref-type="bibr" rid="CR9">9</xref>] do not take local CNs into account when classifying mutations as clonal or subclonal. They assume that the mutations with the highest VAFs are clonal, and classify mutations as subclonal if their CIs do not overlap with those of the &#x02018;clonal&#x02019; mutations. As seen in Figure&#x000a0;<xref rid="Fig11" ref-type="fig">11</xref>, expected clonal VAF levels (light green and light blue horizontal lines) may be very close to expected subclonal VAF levels (black horizontal lines). Therefore, we do not generally recommend classification of mutations by comparing a sample&#x02019;s VAFs within themselves with no reference to local integer CNs.</p><p>Nik-Zainal <italic>et al</italic>. [<xref ref-type="bibr" rid="CR8">8</xref>] estimate the cancer cell fraction (ccf) of each mutation by what we call the &#x02018;multiplicity&#x02019; of the mutation given integer CN estimates in the most evident identified subclone. For RESPONSIFY sample 5, this is exactly the y-levels of mutations in Figure&#x000a0;<xref rid="Fig12" ref-type="fig">12</xref>. More precisely, Nik-Zainal <italic>et al</italic>. [<xref ref-type="bibr" rid="CR8">8</xref>] estimate ccfs as the minimum of the multiplicity and 1, and look for clusters among the mutations with ccf &#x0003c;1. We see no clusters among the y-levels of mutations below the dotted horizontal line of 1 in Figure&#x000a0;<xref rid="Fig12" ref-type="fig">12</xref>, but perhaps we have too few mutations of type A detected from the WES data. We note that such clusters would only reveal very small subclones with low integer CNs and daughter subclones of the most evident identified subclone. We also note that with this ccf estimator, mutations of such a small subclone will get different ccf estimates if they sit on segments with different integer CNs in the most evident identified subclone, so several clusters may arise from the same subclone. Nevertheless, this method may help screening for long subclonal CN alterations to be verified by phasing of SNPs and mutations on the same sequencing reads, which is what Nik-Zainal <italic>et al</italic>. [<xref ref-type="bibr" rid="CR8">8</xref>] do.</p><p>The PyClone algorithm [<xref ref-type="bibr" rid="CR14">14</xref>] clusters mutations on the basis of their VAFs corrected for local CN, termed the &#x02018;cellular prevalence.&#x02019; To do so, at each mutation the algorithm splits the cells in the sample into the &#x02018;normal population&#x02019;, the &#x02018;reference population&#x02019;, consisting of all cancer cells which do not contain the mutation, and the &#x02018;variant population&#x02019;, consisting of all cancer cells with the mutation. It makes a &#x02018;key assumption&#x02019; that all cells within their three populations have the same genotype. We have applied the algorithm to the six samples discussed and made available in this paper. It produces an estimate of the number of subclones in a sample, and assigns mutations to subclones. The results are, in part, consistent with, but also complementary to, ours, bearing in mind that we do not attempt to estimate the number of subclones in a sample. For example, sample 16 depicted in Figure&#x000a0;<xref rid="Fig11" ref-type="fig">11</xref> has 1,232 somatic mutations, and PyClone infers 6 clusters, assigning over 900 to one and over 250 to a second. In data not shown, we inferred that CN alterations in a main aneuploid subclone only (segments of type A) comprised 90% of the genome and held 839 of the mutations (no others could be assigned to a specific subclone), while we found 8% of the genome to be segments of type B. This is a fair degree of consistency between rather different approaches to the same problem. On the other hand, sample 5 had its 199 mutations put into just 3 clusters by PyClone, but as can be seen from Figures&#x000a0;<xref rid="Fig8" ref-type="fig">8</xref>, <xref rid="Fig10" ref-type="fig">10</xref> and <xref rid="Fig12" ref-type="fig">12</xref> it has a considerable amount of subclonality, and we see evidence of more than 3 subclones. Most of our samples are like sample 5 in being highly heterogeneous, and it seems likely that the differences between PyClone&#x02019;s results and ours stem from a failure of their &#x02018;key assumption&#x02019;, in that we have different CNs between different subclones. This point is highlighted in [<xref ref-type="bibr" rid="CR16">16</xref>], where it is noted that clonal inference using CN aberrations and B-allele frequencies need not be the same as that using somatic aberrations. Our approach and that of PyClone are different ways of integrating these two data types, while the integrative analysis of [<xref ref-type="bibr" rid="CR16">16</xref>] is perhaps better than both if one has WGS data. Their method is not available to us as we do not have such data.</p></sec><sec id="Sec22"><title>CN estimation and mutation classification in the literature</title><p>Durinck <italic>et al</italic>. [<xref ref-type="bibr" rid="CR2">2</xref>] identify CN neutral LOH regions within one tumor subclone and classify mutations as homozygous or heterozygous within the subclone. This aim is slightly different to ours, but the paper deserves a mention because it acknowledges that an identified CN pattern reflects CN alteration in a subclone rather than in all tumor cells.</p><p>The software Absolute [<xref ref-type="bibr" rid="CR1">1</xref>] deduces integer CNs in pooled minor and major array CN histograms. The BAF bias of the RESPONSIFY array CNs cannot be spotted with one-dimensional histograms rather than two-dimensional grid plots, and in <xref rid="Sec24" ref-type="sec">Materials and methods</xref> we demonstrate how Absolute therefore does not work with our samples. But given data without bias, Absolute estimates integer CNs under the assumptions that (i) only one pattern of equally interspaced peaks can occur, and (ii) the pattern reflects the clonal CNs of all tumor cells in the sample. With the theoretically expected CN patterns of Figures&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref> and <xref rid="Fig3" ref-type="fig">3</xref> as background we suggest this approach may be useful for samples with most CN alteration taking place in most of the tumor cells, and only small subclones (accounting for up to say 10% of the sample cells) expressing further CN alteration.</p><p>To deduce the scaling of the array CNs, Absolute suggests the scenario for which the majority of (all the sample&#x02019;s) VAFs fit presence on one copy of one homologue of the large subclone. We acknowledge that this is different from our suggested scenario with VAFs (from the genome segments with CN alteration in the pronounced subclone) fitting presence on all copies of one homologue.</p><p>Nik-Zainal <italic>et al</italic>. [<xref ref-type="bibr" rid="CR8">8</xref>] estimates integer CNs and sample purity with ASCAT [<xref ref-type="bibr" rid="CR22">22</xref>], and thereby assumes the minimal CNs fitting array CNs (ignoring the unknown scaling) as well as assumptions (i) and (ii) above, as Absolute does. Again we suggest this approach may be useful for samples with most CN alteration taking place in most of the tumor cells, and only small subclones (accounting for up to say 10% of the sample cells) with other CN alteration. Nik-Zainal <italic>et al</italic>. [<xref ref-type="bibr" rid="CR8">8</xref>] further refine the precise integer CN estimates with help of WGS depths at SNP positions. This may or may not eliminate any BAF bias in average CN estimates; a reader of the paper cannot deduce which. ASCAT fails with most of the RESPONSIFY samples, which are highly aneuploid and subclonal.</p><p>The methods also differ in their interpretation of mutations as clonal or subclonal. In simple terms we call a mutation clonal if its VAF (is larger than or) fits presence on all copies of one homologue in subclone A plus on one copy of one homologue in the rest of the tumor cells (a fraction of cells determined via the pathologist purity estimate). Mutations with significantly smaller VAFs we call subclonal. Absolute calls a mutation clonal if the VAF gives a high likelihood of its presence on at least one copy of a homologue of the (large) subclone. Mutations with a high likelihood of presence in less than one copy are called subclonal. Nik-Zainal <italic>et al</italic>. [<xref ref-type="bibr" rid="CR8">8</xref>] similarly call a mutation clonal if it seems present in at least one copy of one homologue of the (large) subclone, except in segments with further subclonality (type B segments) where they require more. The methods will clearly classify mutations differently. Our method of calling clonal mutations is conservative, and will only find a few such mutations per sample (sometimes none, in particular since ambiguous mutations are not classified). The other three methods [<xref ref-type="bibr" rid="CR1">1</xref>,<xref ref-type="bibr" rid="CR8">8</xref>,<xref ref-type="bibr" rid="CR9">9</xref>] are conservative with calling subclonal mutations and will only call those that are present in a small fraction of cells. To our knowledge there is no clear answer to which of these interpretations is more appropriate biologically.</p></sec></sec><sec id="Sec23" sec-type="conclusion"><title>Conclusions</title><p>We have demonstrated that even with no noise or bias, integer CNs of tumor samples with ITH can only be estimated from SNP array data in samples with a simple clonal architecture, given further information from, for example, WES VAFs or FACS ploidy profiles, and only under a series of assumptions. Even with such samples, integer CNs can only be deduced for some subclones and only across a subset of the genome.</p><p>Estimation of local subclonal CNs has implications for the classification of mutations as clonal or subclonal. The classification also requires knowledge of the sample purity, which cannot be deduced from segmented CN data. The literature on this topic is divergent in assumptions and data analysis methods, with interpretational differences as a result. The insights demonstrated in this study impact research in heterogeneity and tumor evolution, with our emphasis being not only on data analysis methodology but also on the goals, design and interpretation of such studies.</p><p>We would like to stress the importance of illustrative figures to reveal bias and verify model assumptions in ITH studies. We think such evidence of performance is just as important as descriptions of analysis models in papers. As for CN determination, a study of two-dimensional grid plots of average CN estimates can help reveal biases and give clues to sample architecture.</p></sec><sec id="Sec24" sec-type="materials|methods"><title>Materials and methods</title><p>This section provide further details and demonstrations of the points made in the main text.</p><sec id="Sec25"><title>CN estimation challenges caused by selected subclonal structures</title><p>We aim to outline a set of assumptions under which subclonal cell fractions and integer CNs can be estimated from average CNs for some tumor samples. Let us first demonstrate some selected subclonal architectures with help from Figure&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>.</p><p>The clonal, aneuploid tumor of Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref> would have average CNs as shown in blue in Figure&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>a. We simulate a small subclone emerging from the tumor, so that part of a segment which originally had integer CNs (1,2) now splits up into small segments with different amounts of CN alteration relative to the original, main subclone. Figure&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>a shows the resulting grid plot, in which the affected small segments have been colored red. We note that small subclones with additional CN variation to that of a main subclone will cause average CNs between (and sometimes even on top of) the main subclone lattice points.</p><p>Next, imagine a subclonal tumor with 90% purity, which has two subclones as in Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref> plus <italic>&#x003b3;</italic> =10% cells forming another subclone C with integer CNs from 0 to 4, varying independently of the other subclonal integer CNs. Figure&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>b shows simulated average CNs of such a tumor sample, where segments from each third of the genome has been colored differently. Presented with such a grid plot, the underlying subclonal architecture is not easily detected. Even if we were told the number of subclones (three), each average CN is a combination of three subclonal integer CNs, so integer CNs for individual subclones could not be estimated from average CNs alone. We note that the pattern of average CNs quickly gets out of hand as subclonality increases, and that average CNs between regular lattice points may not be caused only by small subclones that deviate from a main subclone (Figure&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>a), but also by small subclones with integer CNs independent of those in a main subclone.</p><p>Even with only two subclones many samples cannot be resolved from average CNs. Figure&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>c is a grid plot from a simulated tumor sample with two subclones of the same size, which have independently sampled integer CNs from 0 to 4. We note that even though one regular grid pattern can be identified in the grid plot, it is not necessarily caused by just one subclone.</p><p>A further difficulty is that in reality not all integer CN combinations will occur, and in particular not in combination with each integer CN in other subclones. Figure&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>d shows the grid plot of a simulated sample with two subclones. Two separate regions of the plot show equally spaced grid points (blue and green). The blue points reflect segments with different integer CNs in the smaller of the two subclones, and (1,1) copies in the larger one. The possible lattice points on which such grid points can fall have been circled. The green points reflect segments with different integer CNs in the smaller subclone and (0,1) in the larger one.</p></sec><sec id="Sec26"><title>Properties and assumptions under which subclonal CNs can be estimated</title><p>In this section we describe some sample properties and assumptions under which conditions&#x000a0;<xref rid="Equ1" ref-type="">1</xref> and <xref rid="Equ2" ref-type="">2</xref> hold so that cell fractions and integer CNs of subclones can be estimated. Imagine a tumor sample for which the following holds.</p><p><italic>Property 1</italic>: The grid plot has a regular vertical/horizontal grid made up by at least two points. This indicates that there is CN alteration in a subclone or in all tumor cells throughout some genome segments where no other subclones have CN alteration. It may also result from the combined effect of CN alteration in two or more subclones. In order to proceed, we must simply assume (Assumption 1 below) that is not the case.</p><p>For example, the blue grid points of Figure&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>d satisfy property 1. Under the following assumption, condition&#x000a0;<xref rid="Equ1" ref-type="">1</xref> holds so we can correctly identify a subclone (say A) in the tumor sample by its cell fraction.</p><p><italic>Assumption 1:</italic> The regular spacing between the grid points of property 1 is caused by consecutive integer CNs in subclone A.</p><p>We now consider</p><p><italic>Property 2:</italic> The point (1,1) is part of the grid pattern suggested by property 1, even if there are no actual points at (1,1).</p><p>and</p><p><italic>Assumption 2:</italic> All grid points that fall on a lattice point of subclone A (circled in Figure&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>d), have normal integer CNs in all other subclones than A.</p><p>This assumption means that no grid points at the lattice points are due to CN variation in other subclones, like the top red point in Figure&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>a, or points of a second subclone with identical size to A as in Figure&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>c.</p><p>If in addition to assumptions 1 and 2, we have property 2 holding, then condition&#x000a0;<xref rid="Equ2" ref-type="">2</xref> is satisfied, and we can estimate integer CNs of subclone A in the genome segments which fall at lattice points of subclone A. We will call these segments type A segments with respect to subclone A.</p><p>Further subclones may be identified using the same strategy. The point (1,1) will be part of the lattice points for each grid caused by CN alteration in one subclone when the integer CNs of the other subclones are normal. Therefore, (1,1) may be regarded as an observed grid point in search of points fulfilling property 1, even if there is no observed point there. With real data, the position of (1,1) will not be identified until a first subclone like A is found, so only subsequent subclone identifications can make use of it.</p><p>Even other subclones may be quantified under additional assumptions, as exemplified next.</p></sec><sec id="Sec27"><title>Example of further subclonal cell fraction and integer CN estimation</title><p>Imagine a tumor sample with an identified subclone A according to properties 1 and 2 and assumptions 1 and 2, and with</p><p><italic>Property 3:</italic> The grid plot has at least one point below the lattice points of subclone A. (This indicates another subclone, with a larger cell fraction than A.)</p><p>For an example, see the green grid points of Figure&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>d. Under the following assumption (which could be varied in different ways), condition&#x000a0;<xref rid="Equ1" ref-type="">1</xref> holds, so we can correctly identify a subclone (say C) in the tumor sample by its cell fraction.</p><p><italic>Assumption 3:</italic> The horizontal distance between (1,1) and the average minor average CN of points below the lattice points of subclone A corresponds to a difference of integer CNs in subclone C of one.</p><p>We call segments with grid points falling below the lattice points of subclone A type C segments. If we further assume</p><p><italic>Assumption 4:</italic> All type C segments have integer CNs (0,1) in subclone C.</p><p>then we could continue to deduce integer CNs in subclone A for those type C segments with grid points on a new set of lattice points, based on assumptions parallel to assumption 2 above.</p></sec><sec id="Sec28"><title>Identical array CNs can originate from different integer CNs</title><p>Given the unbiased, noise-free array CNs of Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>, it is unknown which of the lattice points (<italic>g</italic>₂, <italic>g</italic>₂), (<italic>g</italic>₃, <italic>g</italic>₃) or (<italic>g</italic>₄, <italic>g</italic>₄) corresponds to (1,1) integer copies. The scenarios 2, 3 and 4 involve different fractions <italic>&#x003b1;</italic> of cells displaying the colored grid point CN alterations, different sets of integer CNs, and different scaling factors between array CNs and average CNs. The following algebra shows how two consecutive scenarios (<xref rid="Equ2" ref-type="">2</xref> and <xref rid="Equ3" ref-type="">3</xref>) can result in identical total (that is, minor&#x02009;+&#x02009;major) array CNs (TCNs) and BAFs, and hence identical array CNs.</p><p>For scenario 3, let <italic>c</italic>₁₃ and <italic>c</italic>₂₃ denote the integer CNs in the aneuploid fraction <italic>&#x003b1;</italic>₃ of cells of an arbitrary genome segment. Scenario 3 implies <italic>&#x003b1;</italic>₃&#x02009;=&#x02009;(<italic>g</italic>₃&#x02009;&#x02212;&#x02009;<italic>g</italic>₂)/<italic>g</italic>₃ (Equation&#x000a0;<xref rid="Equ2" ref-type="">2</xref>) and a scale factor <italic>f</italic>₃&#x02009;=&#x02009;<italic>g</italic>₃ relating array CNs to average CNs. Hence the segment has:<disp-formula id="Equa"><alternatives><tex-math id="M14">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ TC{N}_3=\left\{{\alpha}_3\left({c}_{13}+{c}_{23}\right)+2\left(1-{\alpha}_3\right)\right\}{g}_3 $$\end{document}</tex-math><graphic xlink:href="13059_2014_470_Article_Equa.gif" position="anchor"/></alternatives></disp-formula><disp-formula id="Equb"><alternatives><tex-math id="M15">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ BA{F}_3^{upper}=\frac{\alpha_3{c}_{23}+\left(1-{\alpha}_3\right)}{\alpha_3\left({c}_{13}+{c}_{23}\right)+2\left(1-{\alpha}_3\right)}. $$\end{document}</tex-math><graphic xlink:href="13059_2014_470_Article_Equb.gif" position="anchor"/></alternatives></disp-formula></p><p>Next consider scenario 2, for which <italic>&#x003b1;</italic>₂&#x02009;=&#x02009;(<italic>g</italic>₂&#x02009;&#x02212;&#x02009;<italic>g</italic>₁)/<italic>g</italic>₂, <italic>f</italic>₂&#x02009;=&#x02009;<italic>g</italic>₂ and its integer CNs would be <italic>c</italic>₁₃&#x02009;+&#x02009;1 and <italic>c</italic>₂₃&#x02009;+&#x02009;1 for the same segment. We note that (<italic>g</italic>₂&#x02009;&#x02212;&#x02009;<italic>g</italic>₁)&#x02009;=&#x02009;(<italic>g</italic>₃&#x02009;&#x02212;&#x02009;<italic>g</italic>₂)&#x02009;=&#x02009;<italic>&#x003b1;</italic>₂<italic>g</italic>₂&#x02009;=&#x02009;<italic>&#x003b1;</italic>₃<italic>g</italic>₃. Consequently<disp-formula id="Equc"><alternatives><tex-math id="M16">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ \begin{array}{l}TC{N}_2=\left\{{\alpha}_2\left({c}_{13}+{c}_{23}+2\right)+2\left(1-{\alpha}_2\right)\right\}{g}_2\\ {}\kern4.2em =\left\{\frac{\alpha_3{g}_3}{g_2}\left({c}_{13}+{c}_{23}+2\right)+2\frac{g_2-{\alpha}_3{g}_3}{g_2}\right\}{g}_2\\ {}\kern4.2em ={\alpha}_3{g}_3\left({c}_{13}+{c}_{23}+2\right)+2\left({g}_2-{\alpha}_3{g}_3\right)={\alpha}_3{g}_3\left({c}_{13}+{c}_{23}\right)+2{g}_2\\ {}\kern4.2em ={\alpha}_3{g}_3\left({c}_{13}+{c}_{23}\right)+2\left({g}_3-{\alpha}_3{g}_3\right)\\ {}\kern4.2em =\left\{{\alpha}_3\left({c}_{13}+{c}_{23}\right)+2\left(1-{\alpha}_3\right)\right\}{g}_3=TC{N}_3\end{array} $$\end{document}</tex-math><graphic xlink:href="13059_2014_470_Article_Equc.gif" position="anchor"/></alternatives></disp-formula><disp-formula id="Equd"><alternatives><tex-math id="M17">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ \begin{array}{l} BA{F}_2^{upper}=\frac{\alpha_2\left({c}_{23}+1\right)+\left(1-{\alpha}_2\right)}{\alpha_2\left({c}_{13}+{c}_{23}+2\right)+2\left(1-{\alpha}_2\right)}=\frac{\frac{\alpha_3}{\alpha_2}\left\{{\alpha}_2\left({c}_{23}+1\right)+\left(1-{\alpha}_2\right)\right\}}{\frac{\alpha_3}{\alpha_2}\left\{{\alpha}_2\left({c}_{13}+{c}_{23}+2\right)+2\left(1-{\alpha}_2\right)\right\}}=\\ {}\frac{\alpha_3{c}_{23}+{\alpha}_3+\left(\frac{\alpha_3}{\alpha_2}-{\alpha}_3\right)}{\alpha_3\left({c}_{13}+{c}_{23}\right)+2{\alpha}_3+2\left(\frac{\alpha_3}{\alpha_2}-{\alpha}_3\right)}=\frac{\alpha_3{c}_{23}+\frac{\alpha_3}{\alpha_2}}{\alpha_3\left({c}_{13}+{c}_{23}\right)+2\frac{\alpha_3}{\alpha_2}}=\frac{\alpha_3{c}_{23}+\frac{g_2}{g_3}}{\alpha_3\left({c}_{13}+{c}_{23}\right)+2\frac{g_2}{g_3}}=\frac{\alpha_3{c}_{23}+\frac{g_3-{\alpha}_2{g}_2}{g_3}}{\alpha_3\left({c}_{13}+{c}_{23}\right)+2\frac{g_3-{\alpha}_2{g}_2}{g_3}}=\\ {}\frac{\alpha_3{c}_{23}+\left(1-{\alpha}_3\right)}{\alpha_3\left({c}_{13}+{c}_{23}\right)+2\left(1-{\alpha}_3\right)}= BA{F}_3^{upper}\end{array} $$\end{document}</tex-math><graphic xlink:href="13059_2014_470_Article_Equd.gif" position="anchor"/></alternatives></disp-formula></p></sec><sec id="Sec29"><title>Resolving array CN scaling by approximate ploidy calculation</title><p>Given the position of (1,1) integer copies in a grid plot of noise-free array CNs, subclonal cell fractions and integer CNs can be derived for some segments and subclones in fortunate samples under certain assumptions. Unfortunately, the scaling (the position of (1,1)) of array CNs is generally unknown. By calculating subclone specific ploidies for each potential position of (1,1), FACS plots can sometimes help us resolve the scaling issue (Figure&#x000a0;<xref rid="Fig5" ref-type="fig">5</xref>).</p><p>This is how we estimate the ploidy of a selected subclone A under properties 1 and 2 and assumptions 1 and 2. Given the potential scale factor <italic>g</italic>_{<italic>i</italic>} =1, subclonal integer CNs of A can be estimated for each segment <italic>j</italic> on lattice points of subclone A&#x02019;s grid plot by <inline-formula id="IEq12"><alternatives><tex-math id="M18">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ {\hat{c}}_{1j}=\left({a}_{1j}-{g}_{i-1}\right)/\left({g}_i-{g}_{i-1}\right) $$\end{document}</tex-math><inline-graphic xlink:href="13059_2014_470_Article_IEq12.gif"/></alternatives></inline-formula>, <inline-formula id="IEq13"><alternatives><tex-math id="M19">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ {\hat{c}}_{2j}=\left({a}_{2j}-{g}_{i-1}\right)/\left({g}_i-{g}_{i-1}\right) $$\end{document}</tex-math><inline-graphic xlink:href="13059_2014_470_Article_IEq13.gif"/></alternatives></inline-formula>, where <inline-formula id="IEq14"><alternatives><tex-math id="M20">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ {\overset{\rightharpoonup }{a}}_j=\left({a}_{1j},{a}_{2j}\right) $$\end{document}</tex-math><inline-graphic xlink:href="13059_2014_470_Article_IEq14.gif"/></alternatives></inline-formula> are minor and major array CNs.</p><p>If</p><p><italic>Property 4:</italic> The fraction of the genome which cannot be resolved for integer CNs in subclone A is negligible with respect to the subclone&#x02019;s average CN.</p><p>then we can estimate the ploidy of subclone A by summing up these estimated integer CNs to <inline-formula id="IEq15"><alternatives><tex-math id="M21">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ {\displaystyle \sum_j}\left[{l}_j\left({\hat{c}}_{1j}+{\hat{c}}_{2j}\right)\right] $$\end{document}</tex-math><inline-graphic xlink:href="13059_2014_470_Article_IEq15.gif"/></alternatives></inline-formula>, where <italic>l</italic>_{<italic>j</italic>} is the genomic length of segment <italic>j</italic>, and dividing the result by <inline-formula id="IEq16"><alternatives><tex-math id="M22">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ {\displaystyle \sum_j}{l}_j $$\end{document}</tex-math><inline-graphic xlink:href="13059_2014_470_Article_IEq16.gif"/></alternatives></inline-formula>.</p><p>If at least one of the following holds, approximate overall ploidy estimates (across all the tumor cells) can be calculated from the array CNs for each potential position of (1,1), and an independent overall ploidy estimate from, for example, FACS runs may help resolve the array CN scaling.</p><p><italic>Property 5:</italic> The fraction of the genome which cannot be resolved for integer CNs (via subclones) is negligible with respect to the sample&#x02019;s average CN.</p><p><italic>Assumption 5:</italic> The average CN across the part of the genome which can be assessed for integer CNs (via subclones) is similar to the average CN across the rest of the genome.</p><p>This is how, under either property 5 or assumption 5, we estimate overall ploidy in a sample with one evident subclone A. Given the potential scale factor <italic>g</italic>_{<italic>i</italic>} =1, the subclone A cell fraction is <italic>&#x003b1;</italic>&#x02009;=&#x02009;(<italic>g</italic>_{<italic>i</italic>}&#x02009;&#x02212;&#x02009;<italic>g</italic>_{<italic>i</italic> &#x02212;1})/<italic>g</italic>_{<italic>i</italic>}. Let:<disp-formula id="Eque"><alternatives><tex-math id="M23">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ overall\; ploidy=\left(\frac{\alpha }{p}\right){\displaystyle \sum_j}\left[{\pi}_i\left({\hat{c}}_{1j}+{\hat{c}}_{2j}\right)\right]+2\left(\frac{1-\alpha }{p}\right) $$\end{document}</tex-math><graphic xlink:href="13059_2014_470_Article_Eque.gif" position="anchor"/></alternatives></disp-formula></p><p>where <italic>&#x003c0;</italic>_{<italic>j</italic>} is the fraction of the genome associated with segment <italic>j</italic> and <italic>p</italic> is a pathologist&#x02019;s estimate of sample purity. The relative size of subclone A among the tumor cells, <italic>&#x003b1;</italic>/<italic>p</italic>, is also known as the subclone&#x02019;s ccf<italic>.</italic></p><p>In samples with one evident subclone A as well as evidence of a larger subclone, C, we may refine the overall ploidy estimate with the integer CN estimates mentioned earlier.</p></sec><sec id="Sec30"><title>VAFs can sometimes help deduce the scaling of array CNs</title><p>In this section we use the example in Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref> to explain the use of mutation VAFs to deduce the scaling of array CNs. This procedure again requires a set of subjectively chosen assumptions and only works under fortunate circumstances.</p><p>A grid plot of simulated, noise- and bias-free array CNs is shown in Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>a. The scale of the array CNs is unknown. The four colored points suggest the grid pattern drawn for a subclone A, but it is unknown whether (1,1) integer copies happen at (<italic>g</italic>₂, <italic>g</italic>₂), (<italic>g</italic>₃, <italic>g</italic>₃) or (<italic>g</italic>₄, <italic>g</italic>₄). The three scenarios areis illustrated in Figure b-d, which all have one colored column for each of the colored genome segments in Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>a. Equally between the panels, two of the segments have mutations on them (red crosses), with VAFs as shown on the y-axes (simulated without noise or bias).</p><p>Each different scaling suggests different integer CNs (<italic>c</italic>₁, <italic>c</italic>₂) for the segments (labels on x-axes). For example, if <italic>g</italic>₂&#x02009;=&#x02009;1 (Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>b), the blue segment must have integer CNs (2,2) in subclone A. Given a pair of integer CNs (<italic>c</italic>₁, <italic>c</italic>₂), expected VAF levels can be derived under certain assumptions.</p><p>If</p><p><italic>Property 6:</italic> A mutation sits on a segment that falls on a lattice point of a subclone A.</p><p>and we assume</p><p><italic>Assumption 6:</italic> Mutations sit on a number of the <italic>c</italic>₁&#x02009;+&#x02009;<italic>c</italic>₂ local chromosomal copies in subclone A cells only.</p><p>in addition to relying on properties and assumptions 1 to 2, then we would expect VAF levels only in {<italic>&#x003b1;c</italic>/<italic>D</italic>, <italic>c</italic> =1, 2, &#x02026;, <italic>c</italic>₁&#x02009;+&#x02009;<italic>c</italic>₂}, where <italic>&#x003b1;</italic> is the cell fraction of subclone A, and <italic>D</italic> is the total (minor&#x02009;+&#x02009;major) average CN at the mutation&#x02019;s genomic position, <italic>D</italic>&#x02009;=&#x02009;<italic>&#x003b1;</italic>(<italic>c</italic>₁&#x02009;+&#x02009;<italic>c</italic>₂) +2(1&#x02009;&#x02212;&#x02009;<italic>&#x003b1;</italic>). Under these circumstances, and if <italic>g</italic>₂&#x02009;=&#x02009;1 (Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>b), the mutation on the blue segment in the example would sit on 1, 2, 3 or 4 of the 2&#x02009;+&#x02009;2 chromosomal copies. The four corresponding expected VAF levels, simulated with sample purity 90%, have been drawn as continuous, horizontal lines. The other scenarios, <italic>g</italic>₃&#x02009;=&#x02009;1 (Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>c) and <italic>g</italic>₄&#x02009;=&#x02009;1 (Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>d), suggest other VAF levels (continuous, horizontal lines).</p><p>If the sample purity <italic>p</italic> is known (for example, from a pathologist&#x02019;s examination) and if, instead of assumption 6, we assume</p><p><italic>Assumption 7:</italic> Mutations sit on one or both of the chromosomal copies of all tumor cells other than subclone A.</p><p>then we would expect VAF levels only in {(<italic>p</italic>&#x02009;&#x02212;&#x02009;<italic>&#x003b1;</italic>)<italic>c</italic>/<italic>D</italic>, <italic>c</italic> =1, 2}. For mutations present in both subclone A cells and all other tumor cells, assume</p><p><italic>Assumption 8:</italic> Mutations sit on a number of the <italic>c</italic>₁&#x02009;+&#x02009;<italic>c</italic>₂ local chromosomal copies in subclone A cells, and on one or both of the chromosomal copies of all tumor cells other than subclone A.</p><p>We call such mutations clonal, and for these we expect VAF levels only in {(<italic>&#x003b1;c</italic>&#x02009;+&#x02009;(<italic>p</italic>&#x02009;&#x02212;&#x02009;<italic>&#x003b1;</italic>)<italic>d</italic>)/<italic>D</italic>, <italic>c</italic> =1, 2, &#x02026;, <italic>c</italic>₁&#x02009;+&#x02009;<italic>c</italic>₂, <italic>d</italic> =1, 2}. If <italic>g</italic>₂&#x02009;=&#x02009;1 (Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>b), assumptions 7 and 8 and a purity of 90% give the 10 expected VAF levels drawn as dashed horizontal lines for the blue segment.</p><p>Pretending that assumptions 6 to 8 cover all possible locations of mutations on segments that fall on the grid plot lattice points fulfilling conditions&#x000a0;<xref rid="Equ1" ref-type="">1</xref> and <xref rid="Equ2" ref-type="">2</xref>, scenario <italic>g</italic>₄&#x02009;=&#x02009;1 (Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>d) can be ruled out - one mutation VAF in inexplicable as it does not coincide with a horizontal line. If we make the assumption that</p><p><italic>Assumption 9:</italic> Mutations are heterozygous and present on all the copies of its homologue (thick continuous or dashed lines).</p><p>we can also rule out the scenario <italic>g</italic>₂&#x02009;=&#x02009;1 (Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>b) and fix the average CNs of Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>c for this sample. Some samples have mutations which can help resolve the array CN scaling, like this. Other samples may have too few mutations, even in this optimal world with no noise in VAFs or segmented CN data.</p></sec><sec id="Sec31"><title>SNP array preprocessing and segmentation</title><p>Genome-wide SNP analysis of tumor and matched normal samples was performed at AROS Applied Biotechnologies a/s (Aarhus, Denmark) on Affymetrix Genome-Wide Human SNP Arrays 6.0 (Affymetrix, Santa Clara, CA, USA) following the manufacturer&#x02019;s instructions, with the 52 tumor samples and the 29 available matched normal samples. The arrays were preprocessed with the ASCRMAv2 single-array method in the aroma.affymetrix R package [<xref ref-type="bibr" rid="CR30">30</xref>,<xref ref-type="bibr" rid="CR31">31</xref>], and further adjusted for SNP-specific allelic crosstalk with CalMaTe [<xref ref-type="bibr" rid="CR32">32</xref>]. Total (signal A plus signal B) tumor SNP array signals were normalized (divided by) towards total SNP array signals of matched normal samples where available, or otherwise position-specific median total SNP array signals across the normal samples, giving TCNs for all tumors. BAFs were obtained and processed using TumorBoost. Allele-specific CN segments were identified from TCNs and BAFs with the paired or non-paired PSCBS method [<xref ref-type="bibr" rid="CR33">33</xref>] for samples with or without a matched normal sample. After this step we have a minor and a major array CN for each segment, equal to the median TCN(1 - BAF_upper) and TCN(BAF_upper) across the SNPs in the segment. Two arrays failed this preprocessing.</p><p>The segmented array CNs were refined with HAPSEG [<xref ref-type="bibr" rid="CR34">34</xref>], which phases the SNP alleles by comparing the sample-specific SNP data to large databases of normal sample SNP datasets. We let HAPSEG join up the adjacent segments we supplied with similar CNs to a limited extent (seg.merge.thresh =1 or 10^-10 for different samples). HAPSEG significantly reduced CN bias in segments with allelic balance, which originally occurred because segment BAFs were estimated by the median distance between individual SNP BAF levels and 0.5, which is &#x0003e;0 even for segments with allelic balance. It also rescales the segment CNs so that they average to 1 for single homologues. The resulting homologue-specific segment CNs are referred to as array CNs throughout this paper.</p><p>All data analyses in this study were made with R [<xref ref-type="bibr" rid="CR35">35</xref>] unless otherwise stated.</p></sec><sec id="Sec32"><title>WES variant detection</title><p>DNA was extracted using the DNeasy Blood and Tissue Kit&#x000ae; (Qiagen, Venlo, Netherlands) following the manufacturer&#x02019;s instructions. DNA concentration was measured using the NanoDrop 1000 instrument (Thermo Scientific, Waltham, MA, USA). Whole exome sequencing was performed at DNAVision (Gosselies, Belgium). Genomic libraries from the tumor and matched normal samples were generated using the SureSelectXT Reagent Kit HSQ (Agilent Technologies, Santa Clara, CA, USA) following the manufacturer&#x02019;s instructions. Enrichment was performed using the SureSelectXT Human All Exon V4&#x02009;+&#x02009;UTRs kit (Agilent) following the manufacturer&#x02019;s instructions.</p><p>Exome read alignment, filtering, variant calling and annotation were performed as follows. Cutadapt 1.1 [<xref ref-type="bibr" rid="CR36">36</xref>] was used for quality-based adaptor trimming, sequence reads were aligned to the GRCh37/hg19 human reference genome using bwa-aln 0.7.7-r441 [<xref ref-type="bibr" rid="CR37">37</xref>] and duplicate reads marked using Picard tools [<xref ref-type="bibr" rid="CR38">38</xref>]. Aligned reads for each tumor-normal sample pair were combined into one alignment file in BAM format, followed by local indel realignment and base quality recalibration using the Genome Analysis Tool Kit (GATK) software [<xref ref-type="bibr" rid="CR39">39</xref>]. The MuTect 2.7-1-g42d771f [<xref ref-type="bibr" rid="CR40">40</xref>] program was used to identify somatic point mutations. Predictions not labeled as &#x00152;REJECT¹ were accepted as confident somatic mutation predictions and considered for subsequent downstream validation and analysis steps. Variant annotation was performed using the Oncotator web-based service [<xref ref-type="bibr" rid="CR41">41</xref>]. VAFs denote the number of reads with the detected variant as a fraction of all reads at the corresponding genomic position.</p></sec><sec id="Sec33"><title>Scaling bias in SNP array B allele fractions</title><p>Grid plot skewness (Figure&#x000a0;<xref rid="Fig7" ref-type="fig">7</xref>) is adjusted for as follows.</p><p>We assume that an unknown fraction <italic>&#x003b1;</italic> of the (germline-contaminated) tumor sample contributes the most visible, regular CN grid in the plot, and refer to this as our main subclone A. Note that subclone A may be all the tumor cells in the sample, in which case <italic>&#x003b1;</italic> is the sample purity, or it may be a true subclone of tumor cells. Within subclone A, each true single homologue average CN <italic>e</italic> should, theoretically, follow:<disp-formula id="Equ3"><label>3</label><alternatives><tex-math id="M24">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ e\in \left\{\left(1-\alpha \right)+\alpha c,\;c=0,\;1,\;2, \dots \right\}, $$\end{document}</tex-math><graphic xlink:href="13059_2014_470_Article_Equ3.gif" position="anchor"/></alternatives></disp-formula></p><p>where <italic>c</italic> refers to the integer CNs in subclone A. We model the observed minor and major array CNs <inline-formula id="IEq17"><alternatives><tex-math id="M25">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ \overset{\rightharpoonup }{a}=\left({a}_1,{a}_2\right) $$\end{document}</tex-math><inline-graphic xlink:href="13059_2014_470_Article_IEq17.gif"/></alternatives></inline-formula> of an arbitrary CN segment as if they have been subject to a plane rotation <inline-formula id="IEq18"><alternatives><tex-math id="M26">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ F=\left(\begin{array}{cc}\hfill {f}_{11}\hfill &#x00026; \hfill {f}_{12}\hfill \\ {}\hfill {f}_{21}\hfill &#x00026; \hfill {f}_{22}\hfill \end{array}\right) $$\end{document}</tex-math><inline-graphic xlink:href="13059_2014_470_Article_IEq18.gif"/></alternatives></inline-formula> of the average CNs <inline-formula id="IEq19"><alternatives><tex-math id="M27">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ \overset{\rightharpoonup }{e}=\left({e}_1,{e}_2\right) $$\end{document}</tex-math><inline-graphic xlink:href="13059_2014_470_Article_IEq19.gif"/></alternatives></inline-formula>, which are functions of the integer CNs <inline-formula id="IEq20"><alternatives><tex-math id="M28">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ \overset{\rightharpoonup }{c}=\left({c}_1,{c}_2\right) $$\end{document}</tex-math><inline-graphic xlink:href="13059_2014_470_Article_IEq20.gif"/></alternatives></inline-formula> in A: <inline-formula id="IEq21"><alternatives><tex-math id="M29">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ \overset{\rightharpoonup }{a}=F\overset{\rightharpoonup }{e} $$\end{document}</tex-math><inline-graphic xlink:href="13059_2014_470_Article_IEq21.gif"/></alternatives></inline-formula>, that is:<disp-formula id="Equf"><alternatives><tex-math id="M30">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ \left(\begin{array}{c}\hfill {a}_1\hfill \\ {}\hfill {a}_2\hfill \end{array}\right)=\left(\begin{array}{cc}\hfill {f}_{11}\hfill &#x00026; \hfill {f}_{12}\hfill \\ {}\hfill {f}_{21}\hfill &#x00026; \hfill {f}_{22}\hfill \end{array}\right)\left(\begin{array}{c}\hfill 1-\alpha +\alpha {c}_1\hfill \\ {}\hfill 1-\alpha +\alpha {c}_2\hfill \end{array}\right). $$\end{document}</tex-math><graphic xlink:href="13059_2014_470_Article_Equf.gif" position="anchor"/></alternatives></disp-formula></p><p>Our aim is to estimate the rotation matrix <italic>F</italic>, assumed to be common to all the CN segments in the sample, and hence to derive the skewness adjusted array CNs <inline-formula id="IEq22"><alternatives><tex-math id="M31">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ {\overset{\rightharpoonup }{a}}^{\hbox{'}}={\hat{F}}^{-1}\overset{\rightharpoonup }{a} $$\end{document}</tex-math><inline-graphic xlink:href="13059_2014_470_Article_IEq22.gif"/></alternatives></inline-formula>. The matrix <italic>F</italic> involves a scale factor dependent on the unknown fraction <italic>&#x003b1;</italic>. For the sake of grid rotation we use the maximum <italic>&#x003b1;</italic> that fits the array CNs. This is the scaling scenario which corresponds to the smallest possible position of (1,1). (The final scaling step for estimation of <italic>&#x003b1;</italic> and the integer CNs {(<italic>c</italic>₁, <italic>c</italic>₂)} is based on observed VAFs as described below.)</p><p>We find M-estimates of <italic>&#x003b1;</italic> and the matrix <italic>F</italic> numerically [<xref ref-type="bibr" rid="CR42">42</xref>,<xref ref-type="bibr" rid="CR43">43</xref>] by minimizing of the sum of the distances from each CN segment <inline-formula id="IEq23"><alternatives><tex-math id="M32">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ \overset{\rightharpoonup }{a} $$\end{document}</tex-math><inline-graphic xlink:href="13059_2014_470_Article_IEq23.gif"/></alternatives></inline-formula> to its closest skewed lattice point <inline-formula id="IEq24"><alternatives><tex-math id="M33">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ F\overset{\rightharpoonup }{e} $$\end{document}</tex-math><inline-graphic xlink:href="13059_2014_470_Article_IEq24.gif"/></alternatives></inline-formula> weighted by the robust Tukey function and the length of the genome segment corresponding to each <inline-formula id="IEq25"><alternatives><tex-math id="M34">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ \overset{\rightharpoonup }{a} $$\end{document}</tex-math><inline-graphic xlink:href="13059_2014_470_Article_IEq25.gif"/></alternatives></inline-formula>. An M-estimator [<xref ref-type="bibr" rid="CR44">44</xref>,<xref ref-type="bibr" rid="CR45">45</xref>] is a generalization of the maximum likelihood (ML-) estimator. It minimizes the summed values of a function <italic>&#x003c1;</italic>, <inline-formula id="IEq26"><alternatives><tex-math id="M35">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ \hat{F}= argmi{n}_F{\displaystyle \sum_{i=1}^n}\rho \left({r}_i\right) $$\end{document}</tex-math><inline-graphic xlink:href="13059_2014_470_Article_IEq26.gif"/></alternatives></inline-formula>, where <italic>&#x003c1;</italic> is similar to but not necessarily a likelihood function. We let <inline-formula id="IEq27"><alternatives><tex-math id="M36">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ {r}_i={w}_i\left|{\overset{\rightharpoonup }{a}}_i-F{\overset{\rightharpoonup }{e}}_i\right| $$\end{document}</tex-math><inline-graphic xlink:href="13059_2014_470_Article_IEq27.gif"/></alternatives></inline-formula> for each segment <italic>i</italic>, where <inline-formula id="IEq28"><alternatives><tex-math id="M37">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ F{\overset{\rightharpoonup }{e}}_i $$\end{document}</tex-math><inline-graphic xlink:href="13059_2014_470_Article_IEq28.gif"/></alternatives></inline-formula> is the closest lattice point to <inline-formula id="IEq29"><alternatives><tex-math id="M38">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ {\overset{\rightharpoonup }{a}}_i $$\end{document}</tex-math><inline-graphic xlink:href="13059_2014_470_Article_IEq29.gif"/></alternatives></inline-formula>, and <italic>w</italic>_{<italic>i</italic>} is a weight &#x02264;1 determined by the length <italic>l</italic>_{<italic>i</italic>} of segment <italic>i</italic> (typically <inline-formula id="IEq30"><alternatives><tex-math id="M39">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ {w}_i=1-{e}^{-{l}_i/500000} $$\end{document}</tex-math><inline-graphic xlink:href="13059_2014_470_Article_IEq30.gif"/></alternatives></inline-formula> ) to downweight short segments (typically &#x0003c;1&#x000a0;Mb), which might have less reliable array CNs. Tukey&#x02019;s <italic>&#x003c1;</italic> function truncates its input in a smooth fashion, so that observations far away have a limited influence on our estimate. In this way, we avoid segments that violate the grid (for example, because they belong to a different subclone), blurring our estimate of <italic>F</italic>. Figure&#x000a0;<xref rid="Fig7" ref-type="fig">7</xref>c shows a resulting grid plot after grid rotation.</p><p>Starting values <italic>&#x003b1;</italic>₀ and <italic>F</italic>₀ for the grid rotation are derived in two steps. Step 1 estimates a pre-start matrix <inline-formula id="IEq31"><alternatives><tex-math id="M40">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ {F}^{\prime }=\left(\begin{array}{cc}\hfill {f}_{11}^{\hbox{'}}\hfill &#x00026; \hfill {f}_{12}^{\hbox{'}}\hfill \\ {}\hfill {f}_{21}^{\hbox{'}}\hfill &#x00026; \hfill {f}_{22}^{\hbox{'}}\hfill \end{array}\right) $$\end{document}</tex-math><inline-graphic xlink:href="13059_2014_470_Article_IEq31.gif"/></alternatives></inline-formula>, with <italic>F</italic>&#x02032;&#x02009;=&#x02009;<italic>&#x003b1;F</italic> so that a pre-rotation <inline-formula id="IEq32"><alternatives><tex-math id="M41">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ {\left({F}^{\prime}\right)}^{-1}\overset{\rightharpoonup }{a} $$\end{document}</tex-math><inline-graphic xlink:href="13059_2014_470_Article_IEq32.gif"/></alternatives></inline-formula> gives a non-skewed grid plot with unit increments between vertical and horizontal lattice points (Figure&#x000a0;<xref rid="Fig7" ref-type="fig">7</xref>b):<disp-formula id="Equg"><alternatives><tex-math id="M42">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ \left(\begin{array}{c}\hfill {a}_1\hfill \\ {}\hfill {a}_2\hfill \end{array}\right)=\frac{1}{\alpha }{F}^{\prime}\left(\begin{array}{c}\hfill 1-\alpha +\alpha {c}_1\hfill \\ {}\hfill 1-\alpha +\alpha {c}_2\hfill \end{array}\right) $$\end{document}</tex-math><graphic xlink:href="13059_2014_470_Article_Equg.gif" position="anchor"/></alternatives></disp-formula><disp-formula id="Equh"><alternatives><tex-math id="M43">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ ={F}^{\prime}\left(\begin{array}{c}\hfill \frac{1-\alpha }{\alpha }+{c}_1\hfill \\ {}\hfill \frac{1-\alpha }{\alpha }+{c}_2\hfill \end{array}\right), $$\end{document}</tex-math><graphic xlink:href="13059_2014_470_Article_Equh.gif" position="anchor"/></alternatives></disp-formula><disp-formula id="Equi"><alternatives><tex-math id="M44">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ {\left({F}^{\prime}\right)}^{-1}\left(\begin{array}{c}\hfill {a}_1\hfill \\ {}\hfill {a}_2\hfill \end{array}\right)=\left(\begin{array}{c}\hfill \frac{1-\alpha }{\alpha }+{c}_1\hfill \\ {}\hfill \frac{1-\alpha }{\alpha }+{c}_2\hfill \end{array}\right). $$\end{document}</tex-math><graphic xlink:href="13059_2014_470_Article_Equi.gif" position="anchor"/></alternatives></disp-formula></p><p>Rough settings are collected by manually selecting informative clusters in the skewed grid plot (Figure&#x000a0;<xref rid="Fig7" ref-type="fig">7</xref>a):<list list-type="bullet"><list-item><p><italic>d</italic>_{<italic>y</italic>}&#x02009;=&#x02009;<italic>vertical component of the distance between two</italic> &#x02018;<italic>vertically</italic>&#x02019; <italic>consecutive clusters</italic></p></list-item><list-item><p><italic>d</italic>_{<italic>x</italic>}&#x02009;=&#x02009;<italic>horizontal component of the distance between two</italic> &#x02018;<italic>horizontally</italic>&#x02019; <italic>consecutive clusters</italic></p></list-item><list-item><p><italic>slope</italic>_{<italic>y</italic>}&#x02009;=&#x02009;slope of the line through two &#x02018;<italic>vertically</italic>&#x02019; <italic>consecutive clusters</italic></p></list-item><list-item><p><italic>slope</italic>_{<italic>x</italic>}&#x02009;=&#x02009;slope of the line through two &#x02018;<italic>horizontally</italic>&#x02019; <italic>consecutive clusters</italic></p></list-item></list></p><p>Let (<italic>c</italic>_{1<italic>j</italic>}, <italic>c</italic>_{2<italic>j</italic>}) and (<italic>c</italic>_{1<italic>j</italic> +1}, <italic>c</italic>_{2<italic>j</italic> +1}) be the (unknown) integer CNs of subclone A seen as two consecutive, &#x02018;vertical&#x02019; clusters in the skewed grid plot, such that <italic>c</italic>_{1<italic>j</italic>} +1&#x02009;=&#x02009;<italic>c</italic>_{1<italic>j</italic> +1} and <italic>c</italic>_{2<italic>j</italic>}&#x02009;=&#x02009;<italic>c</italic>_{2<italic>j</italic> +1}. Then<disp-formula id="Equj"><alternatives><tex-math id="M45">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ {d}_y=\left({f}_{11}^{\hbox{'}}\left[\frac{1-\alpha }{\alpha }+{c}_{1j+1}\right]+{f}_{12}^{\hbox{'}}\left[\frac{1-\alpha }{\alpha }+{c}_{2j+1}\right]\right)-\left({f}_{11}^{\hbox{'}}\left[\frac{1-\alpha }{\alpha }+{c}_{1j}\right]+{f}_{12}^{\hbox{'}}\left[\frac{1-\alpha }{\alpha }+{c}_{2j}\right]\right)={f}_{11}^{\hbox{'}} $$\end{document}</tex-math><graphic xlink:href="13059_2014_470_Article_Equj.gif" position="anchor"/></alternatives></disp-formula></p><p>and similarly <inline-formula id="IEq33"><alternatives><tex-math id="M46">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ {d}_x={f}_{22}^{\hbox{'}} $$\end{document}</tex-math><inline-graphic xlink:href="13059_2014_470_Article_IEq33.gif"/></alternatives></inline-formula>, <inline-formula id="IEq34"><alternatives><tex-math id="M47">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ slop{e}_y={f}_{11}^{\hbox{'}}/{f}_{21}^{\hbox{'}} $$\end{document}</tex-math><inline-graphic xlink:href="13059_2014_470_Article_IEq34.gif"/></alternatives></inline-formula> and <inline-formula id="IEq35"><alternatives><tex-math id="M48">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ slop{e}_x={f}_{12}^{\hbox{'}}/{f}_{22}^{\hbox{'}} $$\end{document}</tex-math><inline-graphic xlink:href="13059_2014_470_Article_IEq35.gif"/></alternatives></inline-formula>. We estimate the pre-start matrix <italic>F</italic>&#x02032; by:<disp-formula id="Equk"><alternatives><tex-math id="M49">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ \left\{\begin{array}{c}\hfill {f}_{11}^{\hbox{'}}={d}_y\hfill \\ {}\hfill {f}_{22}^{\hbox{'}}={d}_x\hfill \\ {}\hfill {f}_{21}^{\hbox{'}}={f}_{11}^{\hbox{'}}/ slop{e}_y\hfill \\ {}\hfill {f}_{12}^{\hbox{'}}={f}_{22}^{\hbox{'}} slop{e}_x.\hfill \end{array}\right. $$\end{document}</tex-math><graphic xlink:href="13059_2014_470_Article_Equk.gif" position="anchor"/></alternatives></disp-formula></p><p>In step 2 we estimate <italic>&#x003b1;</italic>₀, the maximum possible fraction <italic>&#x003b1;</italic> such that 0&#x02009;&#x0003c;&#x02009;<italic>&#x003b1;</italic> &#x02264;1. Starting from a selected lattice point with allelic balance a grid with step size one is imposed on the pre-rotated Figure&#x000a0;<xref rid="Fig7" ref-type="fig">7</xref>b, stretching as close to zero as possible with all gridlines positive. Let the array CN levels of the first two horizontal gridlines be <italic>&#x003b1;</italic>_{<italic>i</italic>} and <italic>&#x003b1;</italic>_{<italic>ii</italic>}. Using Equation&#x000a0;<xref rid="Equ3" ref-type="">3</xref> we derive <inline-formula id="IEq36"><alternatives><tex-math id="M50">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ {\alpha}_{max}^x=\left({a}_{ii}-{a}_i\right)/{a}_{ii} $$\end{document}</tex-math><inline-graphic xlink:href="13059_2014_470_Article_IEq36.gif"/></alternatives></inline-formula>, and similarly for <inline-formula id="IEq37"><alternatives><tex-math id="M51">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ {\alpha}_{max}^y $$\end{document}</tex-math><inline-graphic xlink:href="13059_2014_470_Article_IEq37.gif"/></alternatives></inline-formula> from vertical gridlines. We set start values for the numerical optimization to <inline-formula id="IEq38"><alternatives><tex-math id="M52">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ {\alpha}_0=\left({\alpha}_{max}^x+{\alpha}_{max}^y\right)/2 $$\end{document}</tex-math><inline-graphic xlink:href="13059_2014_470_Article_IEq38.gif"/></alternatives></inline-formula> and <italic>F</italic>₀&#x02009;=&#x02009;<italic>F</italic>&#x02032;/<italic>&#x003b1;</italic>₀.</p><p>What is the origin of the grid bias, the skewness? We further investigate the SNP array components from which array CNs are computed: TCNs and BAFs. Note that the BAFs here refer to the upper BAF of each segment, which is always between 0.5 and 1.</p><p>We plot the observed total array CNs <italic>a</italic>₁&#x02009;+&#x02009;<italic>a</italic>₂ towards the rotated, supposedly unbiased <inline-formula id="IEq39"><alternatives><tex-math id="M53">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ {a}_1^{\hbox{'}}+{a}_2^{\hbox{'}} $$\end{document}</tex-math><inline-graphic xlink:href="13059_2014_470_Article_IEq39.gif"/></alternatives></inline-formula> (Figure&#x000a0;<xref rid="Fig13" ref-type="fig">13</xref>a) as well as observed (upper) BAFs <italic>a</italic>₂/(<italic>a</italic>₁&#x02009;+&#x02009;<italic>a</italic>₂) towards the rotated (upper) BAFs <inline-formula id="IEq40"><alternatives><tex-math id="M54">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ {a}_2^{\hbox{'}}/\left({a}_1^{\hbox{'}}+{a}_2^{\hbox{'}}\right) $$\end{document}</tex-math><inline-graphic xlink:href="13059_2014_470_Article_IEq40.gif"/></alternatives></inline-formula> (Figure&#x000a0;<xref rid="Fig13" ref-type="fig">13</xref>b). Assuming the rotated CNs are truly proportional to the true average CNs, the graph suggests the original total CNs (since proportional to rotated CNs) are indeed also proportional to the true average CNs. The biased total array CNs and the rotated ones have different scale factors, but that cannot be the cause or adjustment for skewness. In Figure&#x000a0;<xref rid="Fig13" ref-type="fig">13</xref>a single homologue (minor and major) original and rotated CNs are not just proportional to each other, but seem subject to bias related to single CN magnitude. Indeed, Figure&#x000a0;<xref rid="Fig13" ref-type="fig">13</xref>b suggests BAFs carry scaling bias centered at 0.5. If had plotted the lower BAFs instead (which are between 0 and 0.5), Figure&#x000a0;<xref rid="Fig13" ref-type="fig">13</xref>b would have shown points on the bottom left extension of the dotted line. Either way the deviation between BAFs before and after rotation is small for BAFs close to 0.5 and larger further away. To investigate whether such a BAF bias may cause grid plot skewness, we simulate a set of array CNs as in Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>, and derive its true BAFs. We then created a biased dataset with total array CNs as in Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref> but with BAFs biased (Figure&#x000a0;<xref rid="Fig13" ref-type="fig">13</xref>c) according to the estimated linear model in the real dataset (Figure&#x000a0;<xref rid="Fig13" ref-type="fig">13</xref>b). Plotting both the true and the biased simulated array CNs in Figure&#x000a0;<xref rid="Fig13" ref-type="fig">13</xref>d reveals that a 0.5 centered scaling bias of BAFs may indeed cause skewness in grid plots.<fig id="Fig13"><label>Figure 13</label><caption><p> <bold>Origin of skewness. (a)</bold> Sample 45 before versus after rotation TCNs are exactly proportional, whereas minor or major CNs show a more complicated difference. <bold>(b)</bold> Sample 45 BAFs before rotation (upper) carry a 0.5 centered scaling bias compared to after rotation (upper), a bias which causes grid plot skewness. <bold>(c)</bold> Imposed BAF bias on simulated array data in (d). <bold>(d)</bold> Grid plots of true and BAF-induced biased array CNs in a simulated dataset.</p></caption><graphic xlink:href="13059_2014_470_Fig13_HTML" id="MO13"/></fig></p><p>The connection between grid plot skewness and bias in BAFs introduces a relationship between the expected association in Figure&#x000a0;<xref rid="Fig13" ref-type="fig">13</xref>, <italic>BAF</italic>_{<italic>observed</italic>}&#x02009;=&#x02009;<italic>k</italic>&#x02009;+&#x02009;<italic>l</italic>&#x02009;&#x000d7;&#x02009;<italic>BAF</italic>_{<italic>true</italic>}, and the rotation matrix <italic>F</italic> up to a scaling constant <italic>C</italic>:<disp-formula id="Equ4"><label>4</label><alternatives><tex-math id="M55">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ F=C\left(\begin{array}{cc}\hfill 1-k\hfill &#x00026; \hfill 1-k-l\hfill \\ {}\hfill k\hfill &#x00026; \hfill k+l\hfill \end{array}\right). $$\end{document}</tex-math><graphic xlink:href="13059_2014_470_Article_Equ4.gif" position="anchor"/></alternatives></disp-formula></p><p>For the example sample 45 we estimated <italic>BAF</italic>_{<italic>observed</italic>}&#x02009;=&#x02009;&#x02212;0.286&#x02009;+&#x02009;1.582<italic>BAF</italic>_{<italic>rotated</italic>}, so we should have <italic>F</italic> proportional to <inline-formula id="IEq41"><alternatives><tex-math id="M56">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ \left(\begin{array}{cc}\hfill 1.286\hfill &#x00026; \hfill -0.286\hfill \\ {}\hfill -0.296\hfill &#x00026; \hfill 1.296\hfill \end{array}\right) $$\end{document}</tex-math><inline-graphic xlink:href="13059_2014_470_Article_IEq41.gif"/></alternatives></inline-formula>, and from the grid rotation we indeed estimated:<disp-formula id="Equl"><alternatives><tex-math id="M57">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ F=\left(\begin{array}{cc}\hfill 0.447\hfill &#x00026; \hfill -0.103\hfill \\ {}\hfill -0.101\hfill &#x00026; \hfill 0.452\hfill \end{array}\right)=0.349\left(\begin{array}{cc}\hfill 1.280\hfill &#x00026; \hfill -0.294\hfill \\ {}\hfill -0.289\hfill &#x00026; \hfill 1.295\hfill \end{array}\right). $$\end{document}</tex-math><graphic xlink:href="13059_2014_470_Article_Equl.gif" position="anchor"/></alternatives></disp-formula></p><p>Equation&#x000a0;<xref rid="Equ4" ref-type="">4</xref> imposes the restriction <italic>f</italic>₁₁&#x02009;+&#x02009;<italic>f</italic>₂₁&#x02009;=&#x02009;<italic>f</italic>₁₂&#x02009;+&#x02009;<italic>f</italic>₂₂ on <italic>F</italic>, which can sometimes help the numerical optimization.</p><p>The BAF bias causes segments with very low minor CNs to get upper BAFs biased down to 1 in the array preprocessing steps. Such segments will appear as a horizontal bottom line in original grid plots, and a sloped bottom line after grid rotation. We believe that these segments should have had a constant minor array CN, and so we project the corresponding grid plot points vertically down to the observed bottom line of constant minor array CNs when the latter is evident.</p><p>Unless otherwise stated, we refer to array CNs as rotated array CNs after projection of bottom sloped line array CNs, and we drop the prime from <inline-formula id="IEq42"><alternatives><tex-math id="M58">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ {\overset{\rightharpoonup }{a}}^{\hbox{'}} $$\end{document}</tex-math><inline-graphic xlink:href="13059_2014_470_Article_IEq42.gif"/></alternatives></inline-formula>.</p></sec><sec id="Sec34"><title>Estimation of subclonal architecture, cell fractions and integer CNs in RESPONSIFY samples</title><p>The cell fraction and integer CNs are estimated for type A segments of cells in the most evident subclone A. According to our assumptions all other cells are diploid in these segments.</p><p>CN alteration in type B segments may be due to further CN alteration in daughter subclones of A or by CN alteration in subclones independent of A. For some segments which have many mutations we can deduce the subclonal origin via the mutation VAFs as in the following example. In Figure&#x000a0;<xref rid="Fig11" ref-type="fig">11</xref>, VAFs in type A segments (blue) match the expected VAF levels of subclone A ({<italic>&#x003b1;c</italic>/<italic>D</italic>, <italic>c</italic>&#x02009;=&#x02009;<italic>c</italic>₁, <italic>c</italic>₂} with notation as above but now referring to estimates from real data), which is reassuring for our analyses. VAFs in type B segments (pink) also match expected VAF levels of subclone A ({<italic>&#x003b1;c</italic>/<italic>D</italic>, <italic>c</italic>&#x02009;=&#x02009;<italic>c</italic>₁, <italic>c</italic>₂} with <italic>c</italic>₁, <italic>c</italic>₂ being fractional rather than integer CNs). This suggests the true CNs indeed meet the fractional <italic>c</italic>₁, <italic>c</italic>₂ in a fraction <italic>&#x003b1;</italic> of the sample cells, which in turn suggests subclone A has daughter subclones with further CN alteration in these segments.</p><p>For 25 of the 48 samples resolved for array scaling we identified an additional subclone C by the existence of a lower grid pattern in the grid plot (see <xref rid="Sec24" ref-type="sec">Materials and methods</xref>). In these cases we call the lower grid pattern segments type C segments, estimate the subclone C cell fraction approximately, assign integer CNs (0,1) to the subclone C type C segments, and estimate integer CNs of subclone A in the type C segments via Equation&#x000a0;<xref rid="Equ2" ref-type="">2</xref>.</p><p>All our samples lack regular grid patterns in high array CN segments (red in Figure&#x000a0;<xref rid="Fig8" ref-type="fig">8</xref>). We call these type D segments. Their CN alteration may take place in subclone A (but the grid is not regular because the proportionality between array CNs and average CNs breaks down with high SNP array intensities), in subclone C or in any other subclone. As for type B segments, some segments which have many mutations can be assigned to a subclone via the mutation VAFs. In the example of Figure&#x000a0;<xref rid="Fig11" ref-type="fig">11</xref>, the VAFs in type D segments (red) match the expected VAF level of the minor homologue of subclone A ({<italic>&#x003b1;c</italic>₁/<italic>D</italic>} where <italic>c</italic>₁ is a fractional CN. This suggests that any CN alteration in the minor homologue takes place in subclone A.</p></sec><sec id="Sec35"><title>Clonal or subclonal mutations in RESPONSIFY samples</title><p>Continued from the &#x02018;Clonal or subclonal mutations&#x02019; section in <xref rid="Sec2" ref-type="sec">Results</xref>. Segment types are exemplified in Figure&#x000a0;<xref rid="Fig8" ref-type="fig">8</xref>.</p><p>For samples with only one identified subclone A, we classify each type A mutation as clonal if the CI falls above (<italic>&#x003b1;c</italic>₁&#x02009;+&#x02009;(<italic>p</italic>&#x02009;&#x02212;&#x02009;<italic>&#x003b1;</italic>))/<italic>D</italic>&#x02009;&#x02212;&#x02009;<italic>&#x003b4;</italic>, with <italic>&#x003b4;</italic> =0.1 (non-inferiority test at significance level 5%), <italic>&#x003b1;</italic>, <italic>c</italic>₁ and <italic>c</italic>₂ estimates for subclone A and <italic>D</italic> the estimate of the local minor&#x02009;+&#x02009;major average CN. We classify a mutation as subclonal if its CI falls below (<italic>&#x003b1;c</italic>₁&#x02009;+&#x02009;(<italic>p</italic>&#x02009;&#x02212;&#x02009;<italic>&#x003b1;</italic>))/<italic>D</italic> (one-sided inferiority test at significance level 5%). Some mutations will be called ambiguous. For subclonal mutations we further test whether they sit on subclone A or not, with equality tests significant if the CI falls within <italic>&#x003b1;c</italic>₁/<italic>D</italic>&#x02009;&#x000b1;&#x02009;<italic>&#x003b4;</italic> or <italic>&#x003b1;c</italic>₂/<italic>D</italic>&#x02009;&#x000b1;&#x02009;<italic>&#x003b4;</italic>.</p><p>If, in addition to a most evident subclone A, a sample has another identified subclone C in a fraction <italic>&#x003b3;</italic> of the cells, we assess whether the subclonal type A mutations seem to sit on A only (CI &#x02208; <italic>&#x003b1;c</italic>₁/<italic>D</italic>&#x02009;&#x000b1;&#x02009;<italic>&#x003b4;</italic> or CI &#x02208; <italic>&#x003b1;c</italic>₂/<italic>D</italic>&#x02009;&#x000b1;&#x02009;<italic>&#x003b4;</italic>), on C only (CI &#x02208; <italic>&#x003b3;</italic>/<italic>D</italic>&#x02009;&#x000b1;&#x02009;<italic>&#x003b4;</italic>) or on both A and C (CI &#x02208; (<italic>&#x003b1;c</italic>₁&#x02009;+&#x02009;<italic>&#x003b3;</italic>)/<italic>D</italic>&#x02009;&#x000b1;&#x02009;<italic>&#x003b4;</italic> or CI &#x02208; (<italic>&#x003b1;c</italic>₂&#x02009;+&#x02009;<italic>&#x003b3;</italic>)/<italic>D</italic>&#x02009;&#x000b1;&#x02009;<italic>&#x003b4;</italic>). The extended procedure creates more ambiguous mutations, since we only allow non-ambiguous classifications.</p><p>Mutations of type B or D are classified like A mutations (so that equivalence to a specific subclonal expected VAF level suggests further daughter subclones with CN alteration in B segments) unless the sample has an identified subclone C. In case of the latter, segments of type B or D may have (I) integer CNs (<italic>c</italic>_{1<italic>a</italic>}, <italic>c</italic>_{2<italic>a</italic>}) in subclone A and (1, 1) in other tumor cells, or, they may have (II) integer CNs (<italic>c</italic>_{1<italic>c</italic>}, <italic>c</italic>_{2<italic>c</italic>}) in subclone C and (1, 1) in other tumor cells, or something else. We classify the type B and D mutations as outlined for type A mutations (based on I) in parallel with an analogous procedure based on II. Only mutations for which both classifications agree are finally assigned a class different from ambiguous.</p><p>Type C mutations are classified as type B or D ones, except in samples with an identified subclone C, for which we set the minor homologue integer CNs in subclone C to (0, 1) instead of (1, 1), and adjust the subclone A integer CNs according to Equation&#x000a0;<xref rid="Equ2" ref-type="">2</xref>.</p><p>Approximately 35% of the 52 RESPONSIFY samples have the majority of type B and D VAFs matching expected VAF levels for CN alteration in subclone A. Most samples have no or only a handful of mutations in type C segments.</p></sec><sec id="Sec36"><title>Grid plots reveal bias</title><p>This section illustrates the benefit of two-dimensional grid plots compared with one-dimensional histograms, in order to reveal bias in array CNs.</p><p>Each point in the original (Figure&#x000a0;<xref rid="Fig14" ref-type="fig">14</xref>a) and rotated (Figure&#x000a0;<xref rid="Fig14" ref-type="fig">14</xref>b) array CN grid plots shows the minor and major array CN of a genome segment. In the Absolute software, the minor and major array CNs are pooled and shown in (one-dimensional) histograms (Figure&#x000a0;<xref rid="Fig14" ref-type="fig">14</xref>c for original and Figure&#x000a0;<xref rid="Fig14" ref-type="fig">14</xref>d for rotated array CNs) with heights proportional to segment lengths. Hence, each segment is represented twice in the histograms - once with the minor array CN and once with the major. Segments with subclone A two-dimensional CN estimates (1,1), (2,2), (0, 2) and (1,2) have been colored equally in all four panels (black, red, blue, green). The cyan segments have non-integer CNs (between 1 and 2) with respect to subclone A. They may have further heterogeneity within subclone A or originating from a subclone independent of A. Absolute searches for equally interspaced peak centers in the histogram with a maximum likelihood algorithm, and each peak is assigned an integer CN estimate. Segments that fall significantly far from their closest peak centers are classified as subclonal, under the assumptions that (i) only one pattern of equally interspaced peaks can occur, and (ii) the pattern reflects the clonal CNs of all tumor cells in the sample. According to the CN coloring in Figure&#x000a0;<xref rid="Fig14" ref-type="fig">14</xref>, three colored histogram peaks are expected: CN =0 (blue), CN =1 (black, green) and CN =2 (red, blue, green). In the histogram of original array CNs (Figure&#x000a0;<xref rid="Fig14" ref-type="fig">14</xref>a) it is hardly possibly to identify the three CN levels, their centers are not equally interspaced, and non-integer CNs (cyan) are intermixed with the integer CNs. Since all our samples have skewness, Absolute did not assign integer CNs optimally.<fig id="Fig14"><label>Figure 14</label><caption><p> <bold>Sample 11 grid plots and Absolute histograms. (a)</bold> Original grid plot. <bold>(b)</bold> Grid plot after rotation. <bold>(c)</bold> Absolute histogram based on original array CNs. <bold>(d)</bold> Absolute histogram based on rotated array CNs. Note that all plots show the array CNs scaled to equal average CNs, so that the level one (<xref rid="Equ1" ref-type="">1</xref>) corresponds to normal, single haplotype CNs.</p></caption><graphic xlink:href="13059_2014_470_Fig14_HTML" id="MO14"/></fig></p></sec><sec id="Sec37"><title>Data and implementation</title><p>The preprocessed array CN data for the six samples discussed in this paper are available as Additional files <xref rid="MOESM1" ref-type="media">1</xref>, <xref rid="MOESM2" ref-type="media">2</xref>, <xref rid="MOESM3" ref-type="media">3</xref>, <xref rid="MOESM4" ref-type="media">4</xref>, <xref rid="MOESM5" ref-type="media">5</xref> and <xref rid="MOESM6" ref-type="media">6</xref>, while the Oncotator annotated Mutect variants for two of these samples are available in Additional files <xref rid="MOESM7" ref-type="media">7</xref> and <xref rid="MOESM8" ref-type="media">8</xref>. A CRAN package to identify grid patterns, perform our grid rotation algorithm and calculate the ITH endpoint will be available shortly with full documentation under the name &#x02018;Gridith&#x02019;. A &#x0201c;Gridith&#x0201d; beta version is available at <ext-link ext-link-type="uri" xlink:href="https://github.com/fcaramia/GRIDITH">https://github.com/fcaramia/GRIDITH</ext-link>.</p></sec></sec>

Tanshinone II-A sodium sulfonate (DS-201) enhances human BK_Ca channel activity by selectively targeting the pore-forming &#x003b1; subunit

<sec><title>Aim:</title><p>Tanshinone II-A sodium sulfonate (DS-201), a water-soluble derivative of Tanshinone II-A, has been found to induce vascular relaxation and activate BK_Ca channels. The aim of this study was to explore the mechanisms underlying the action of DS-201 on BK_Ca channels.</p></sec><sec><title>Methods:</title><p>Human BK_Ca channels containing &#x003b1; subunit alone or &#x003b1; plus &#x003b2;1 subunits were expressed in HEK293 cells. BK_Ca currents were recorded from the cells using patch-clamp technique. The expression and trafficking of BK_Ca subunits in HEK293 cells or vascular smooth muscle cells (VSMCs) were detected by Western blotting, flow cytometry and confocal microscopy.</p></sec><sec><title>Results:</title><p>DS-201 (40&#x02013;160 &#x003bc;mol/L) concentration-dependently increased the total open probability of BK_Ca channels in HEK293 cells, associated with enhancements of Ca²⁺ and voltage dependence as well as a delay in deactivation. Coexpression of &#x003b2;1 subunit did not affect the action of DS-201: the values of EC₅₀ for BK_Ca channels containing &#x003b1; subunit alone and &#x003b1; plus &#x003b2;1 subunit were 66.6&#x000b1;1.5 and 62.0&#x000b1;1.1 &#x003bc;mol/L, respectively. In both HEK293 cells and VSMCs, DS-201 (80 &#x003bc;mol/L) markedly increased the expression of &#x003b1; subunit without affecting &#x003b2;1 subunit. In HEK293 cells, DS-201 enriched the membranous level of &#x003b1; subunit, likely by accelerating the trafficking and suppressing the internalization of &#x003b1; subunit. In both HEK293 cells and VSMCs, DS-201 (&#x02265;320 &#x003bc;mol/L) induced significant cytotoxicity.</p></sec><sec><title>Conclusion:</title><p>DS-201 selectively targets the pore-forming &#x003b1; subunit of human BK_Ca channels, thus enhancing the channel activities and increasing the subunit expression and trafficking, whereas the &#x003b2;1 subunit does not contribute to the action of DS-201.</p></sec>

<contrib contrib-type="author"><name><surname>Tan</surname><given-names>Xiao-qiu</given-names></name><xref ref-type="aff" rid="aff1">1</xref><xref ref-type="aff" rid="aff2">2</xref><xref ref-type="author-notes" rid="note1">^&#x00023;</xref></contrib><contrib contrib-type="author"><name><surname>Cheng</surname><given-names>Xiu-li</given-names></name><xref ref-type="aff" rid="aff1">1</xref><xref ref-type="author-notes" rid="note1">^&#x00023;</xref></contrib><contrib contrib-type="author"><name><surname>Yang</surname><given-names>Yan</given-names></name><xref ref-type="aff" rid="aff2">2</xref></contrib><contrib contrib-type="author"><name><surname>Yan</surname><given-names>Li</given-names></name><xref ref-type="aff" rid="aff1">1</xref></contrib><contrib contrib-type="author"><name><surname>Gu</surname><given-names>Jing-li</given-names></name><xref ref-type="aff" rid="aff1">1</xref></contrib><contrib contrib-type="author"><name><surname>Li</surname><given-names>Hui</given-names></name><xref ref-type="aff" rid="aff1">1</xref></contrib><contrib contrib-type="author"><name><surname>Zeng</surname><given-names>Xiao-rong</given-names></name><xref ref-type="aff" rid="aff2">2</xref><xref ref-type="corresp" rid="caf2">*</xref></contrib><contrib contrib-type="author"><name><surname>Cao</surname><given-names>Ji-min</given-names></name><xref ref-type="aff" rid="aff1">1</xref><xref ref-type="corresp" rid="caf1">*</xref></contrib><aff id="aff1"><label>1</label><institution>Department of Physiology, Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences, School of Basic Medicine, Peking Union Medical College</institution>, Beijing 100005, <country>China</country></aff><aff id="aff2"><label>2</label><institution>Key Laboratory of Medical Electrophysiology, Ministry of Education, Collaborative Innovation Center for Prevention and Treatment of Cardiovascular Disease/Institute of Cardiovascular Research, Luzhou Medical College</institution>, Luzhou 646000, <country>China</country></aff>

Acta Pharmacologica Sinica

<sec sec-type="intro"><title>Introduction</title><p>It is well known that arterial tone is regulated by functional balance of the ion channels responsible for cellular depolarization and hyperpolarization^{<xref ref-type="bibr" rid="bib1">1</xref>}. The large conductance calcium-activated potassium channels (BK_Ca channels), also called Maxi-K or Slo, are broadly expressed in vascular smooth muscle cells (VSMCs) and play crucial roles in regulating vascular tone^{<xref ref-type="bibr" rid="bib2">2</xref>,<xref ref-type="bibr" rid="bib3">3</xref>}. Activation of BK_Ca channels by elevation of the intracellular calcium concentration due to membrane depolarization increases the K⁺ conductance of the membrane and drives VSMC membrane hyperpolarization, which in turn closes the L-type voltage-dependent Ca²⁺ channels (LVDCCs), decreases global [Ca²⁺]_i, and induces vascular relaxation^{<xref ref-type="bibr" rid="bib4">4</xref>,<xref ref-type="bibr" rid="bib5">5</xref>}. The BK_Ca channel consists of a functional Slo &#x003b1;-subunit and an affiliated &#x003b2; subunit^{<xref ref-type="bibr" rid="bib6">6</xref>}. There are 4 types of affiliated subunits (&#x003b2;1&#x02013;&#x003b2;4). The &#x003b2;1 subunit is the major affiliated subunit of VSMCs, and it enhances not only BK_Ca &#x003b1;-subunit expression in the membrane but also the voltage and Ca²⁺ sensitivity of BK_Ca channels^{<xref ref-type="bibr" rid="bib7">7</xref>}. Previous reports have revealed that &#x003b2;1 subunit knock-out rats easily developed hypertension^{<xref ref-type="bibr" rid="bib8">8</xref>}. Our previous study showed that, in the VSMCs of patients with essential hypertension, the whole-cell current, spontaneous transient outward potassium currents (STOCs) and the Ca²⁺ sensitivity of BK_Ca channels were reduced due to the down-regulation of the &#x003b2;1 subunit both at the mRNA and protein levels, whereas &#x003b1;-subunit expression was maintained^{<xref ref-type="bibr" rid="bib9">9</xref>}. Therefore, rescuing &#x003b2;1 subunit function to restore the activity of BK_Ca channels could be a therapeutic approach for diseases with &#x003b2;1 subunit malfunction, such as hypertension. Alternatively, enhancing the activity of the &#x003b1; subunit directly might also be a therapeutic strategy for these diseases.</p><p>Danshen (<italic>Salvia miltiorrhiza</italic>), a traditional Chinese medicinal herb, has been widely used in China and many other countries with minimal side effects in therapies for cardiovascular and cerebrovascular diseases^{<xref ref-type="bibr" rid="bib10">10</xref>,<xref ref-type="bibr" rid="bib11">11</xref>,<xref ref-type="bibr" rid="bib12">12</xref>}. Tanshinone II-A is a type of diterpene quinine and a major effective component of Danshen, and tanshinone II-A sodium sulfonate (DS-201) (<xref ref-type="fig" rid="fig1">Figure 1</xref>) is a water-soluble derivative of tanshinone II-A after sulfonation. DS-201 retains the pharmacological efficacy of tanshinone II-A and is convenient for injection because of its water-soluble character. We showed in a previous report that DS-201 could induce the relaxation of isolated vascular rings, activate the macro-currents in a whole-cell configuration and increase open probability in inside-out patches in porcine coronary arterial smooth muscle cells^{<xref ref-type="bibr" rid="bib13">13</xref>}, suggesting that BK_Ca channel activation is involved in DS-201-mediated vasorelaxation. We further demonstrated that DS-201 activated BK_Ca channels mainly by shifting the kinetic properties and the Ca²⁺ dependence of the channel in mouse cerebral arterial smooth muscle cells^{<xref ref-type="bibr" rid="bib14">14</xref>}, suggesting that DS-201 likely plays a role similar to that of the &#x003b2;1 subunit in modulating the BK_Ca channel. However, it remains unknown whether the &#x003b2;1 subunit is necessary to the action of DS-201 on the BK_Ca channel and whether DS-201 can potentiate the expression and trafficking of BK_Ca channels, just as &#x003b2;1 subunit does.</p><p>To address the above questions, we manipulated HEK293 cells to express the &#x003b1; subunit alone or with the &#x003b2;1 subunit of the BK_Ca channels. Using these genetically engineered cell lines and cultured VSMCs and a series of approaches, including patch clamp, Western blotting, flow cytometry (FCM) and confocal microscopy, we investigated the acute effects of DS-201 on BK_Ca channel kinetics and the relatively chronic effects on the expression and trafficking of BK_Ca channel subunits, with a particular focus on the target protein of DS-201 in the BK_Ca channel complex. This study could increase our understanding of the mechanisms of DS-201 in mediating vasorelaxation and could provide further perspectives for promoting the clinical use of Danshen for treating diseases with vascular dysfunction, such as hypertension.</p></sec><sec sec-type="materials|methods"><title>Materials and methods</title><sec><title>Materials</title><p>DS-201, with the chemical formula of C₁₉H₁₇NaO₆S and the structure shown in <xref ref-type="fig" rid="fig1">Figure 1</xref>, was purchased from the National Institutes for Food and Drug Control (NIFDC, Beijing, China) (purity &#x02265;98%). The drug was dissolved in de-ionized water to obtain a 4-mmol/L stock solution and was added to the bath solution or culture medium to achieve the desired concentrations. K-aspartate, HEPES, EGTA and iberiotoxin (IbTX) were obtained from Sigma (MO, USA).</p></sec><sec><title>Plasmid construction</title><p>The plasmid pcDNA3.1-hSlo, containing the human BK_Ca &#x003b1; subunit (hSlo), was a kind gift from Prof Philip K AHRING (NeuroSearch A/S, Denmark). Using the overlapping PCR protocol, we constructed the expression plasmid pcDNA3.1-Flag-hSlo-EGFP (abbreviated as &#x0201c;Flag-hSlo-GFP&#x0201d; below), using the plasmid pcDNA3.1-hSlo as a template. The enhanced green fluorescence protein (EGFP) was connected to the C-terminal of hSlo, and a Flag tag (DYKDDDDK) was inserted into the extracellular loop between the S1 and S2 segments (SNPIES<underline>DYKDDDDK</underline>CQNFYKDF). The expression plasmid pEF1-myc-h&#x003b2;1 (h&#x003b2;1), containing the human &#x003b2;1 subunit with the His tag, was previously constructed in our laboratory. All of the plasmids were identified by DNA sequencing.</p></sec><sec><title>Cell culture and transfection</title><p>HEK293 cells (from the National Platform of Experimental Cell Resources for Sci-Tech, Beijing, China) and Sprague-Dawley (SD) rat thoracic aortic vascular smooth muscle cell lines (VSMCs) (A7r5 cell lines, from the American Type Culture Collection, ATCC, VA, USA) were cultured at 37 &#x000b0;C in a 95% air/5% CO₂ humidified incubator with Dulbecco's modified Eagle's medium (DMEM) containing 10% (<italic>v/v</italic>) fetal bovine serum (Gibco, Invitrogen, New York, USA). HEK293 cells were used for transfection and VSMCs for investigating BK_Ca protein expression. Transfection was carried out at 70%&#x02013;80% cell confluence, and 4 &#x003bc;g of total plasmid DNA (hSlo or Flag-hSlo-GFP, with or without h&#x003b2;1) was added to 35-mm cell culture dishes for transfection using the lipofection technique (Lipofectamine 2000, Invitrogen, NY, USA).</p></sec><sec><title>Patch clamp</title><p>Patch clamp experiments were performed in essentially the same manner as described previously^{<xref ref-type="bibr" rid="bib13">13</xref>}. Channel currents were recorded with configurations of whole-cell, inside-out and outside-out patches using EPC-10 amplifier (HEKA, Lambrecht/Pfalz, Germany). The total open probability (<italic>N</italic>Po), current amplitude and density, and kinetic characteristics of the BK_Ca channels were analyzed with the pCLAMP software (version 10.0, Molecular Devices, CA, USA). In the whole-cell configuration, the bath (extracellular) solution consisted of (in mmol/L): NaCl 140, KCl 5, CaCl₂ 1.8, MgCl₂ 2, and HEPES 10 (pH 7.4 with NaOH). The pipette (intracellular) solution consisted of (in mmol/L): K-aspartate (K-Asp) 100, KCl 40, MgCl₂ 5, EGTA 1, and HEPES 10 (pH 7.2 with KOH), and the intracellular Ca²⁺ concentration [Ca²⁺]_i was adjusted to 0.1 &#x003bc;mol/L. After the series resistance was compensated for by 70% and reached &#x0003c;10 M&#x003a9; to minimize voltage errors, the whole-cell macroscopic currents were recorded with step pulses (from &#x02212;70 mV to +70 mV), followed by N/P leak subtraction. In the inside-out patch experiments, the pipette (extracellular) solution consisted of (in mmol/L): K-Asp 40, KCl 100, HEPES 10, and EGTA 2 (pH 7.2 with KOH); and the bath (intracellular) solution (in mmol/L): K-Asp 100, KCl 40, HEPES 10, and EGTA 1 (pH 7.4 with KOH). In the outside-out patch, the pipette (intracellular) solution consisted of (in mmol/L): K-Asp 100, KCl 40, HEPES 10, and EGTA 1 (pH 7.4 with KOH); and the bath (extracellular) solution (in mmol/L): K-Asp 40, KCl 100, HEPES 10, and EGTA 2 (pH 7.2 with KOH). Because DS-201 activates the BK_Ca channels mainly from the cytoplasmic side of the membrane, we conducted a special configuration: the inside-out macro-patches. The advantages of this configuration were that the electrode tip had a larger size with low resistance and therefore could cover thousands of BK_Ca channels when an inside-out patch was formed, if also considering that the channel proteins were overexpressed in HEK293 cells. Thus, nanoampere currents could be obtained as macro-currents. Additionally, it was very convenient to change solution components at the &#x0201c;intracellular&#x0201d; side. Here, we used these inside-out macro-patches to investigate the voltage dependence and kinetics of the BK_Ca channels. The pipette and bath solutions were the same as those used in the inside-out patch experiments. To create serial intracellular free Ca²⁺ concentrations ([Ca²⁺]_i) of 0, 0.01, 0.1, 0.5, 1, or 10 &#x003bc;mol/L, the CaCl₂ concentration of the bath solution was set to 0, 0.11, 0.55, 0.86, 0.92, or 1 mmol/L, respectively^{<xref ref-type="bibr" rid="bib14">14</xref>}. The membrane potential (<italic>V</italic>_m) was expressed as that of the intracellular side. If there was no other special instruction, the intracellular free Ca²⁺ concentration ([Ca²⁺]_i) was 0.1 &#x003bc;mol/L, and <italic>V</italic>_m was +40 mV in the inside-out patch. All of the electrophysiological experiments were conducted at room temperature (22&#x000b1;2 &#x000b0;C).</p></sec><sec><title>Western blotting</title><p>Western blotting was conducted as described in a previous report^{<xref ref-type="bibr" rid="bib15">15</xref>}. Briefly, cells were harvested and lysed with lysis buffer containing 50 mmol/L Tris-Cl (pH 7.4), 150 mmol/L NaCl, 1% Triton X-100, 1% sodium deoxycholate and a series of protease inhibitors. Samples containing approximately 50 &#x003bc;g of total protein were separated with SDS-PAGE and were transferred to PVDF membrane, followed by blocking and incubation with anti-Slo (1:500, Alomone, Jerusalem, Israel) or anti-&#x003b2;1 (1:500, Alomone, Jerusalem, Israel) primary antibodies at 4 &#x000b0;C overnight. The blots were then incubated with a horseradish peroxidase-conjugated secondary antibody (1:5000, MBL, Nagoya, Japan) at room temperature for 1 h and were developed using an ECL system (Engreen Biosystem Co, Ltd, Beijing, China). Images were obtained and quantified using Quantity One software (Bio-Rad, CA, USA).</p></sec><sec><title>Biotinylation and isolation of cell surface proteins</title><p>Membrane biotinylation of HEK293 cells expressing BK_Ca channels were completed according to the manufacturer's protocol provided in the Cell Surface Protein Isolation Kit (Pierce, IL, USA). Briefly, cells were washed with cold PBS, followed by incubation with 0.25 mg/mL Sulfo-NHS-SS-Biotin in PBS for 30 min on ice. After quenching the biotinylation reaction, the cells were collected, lysed and incubated with beads for the isolation of the labeled proteins. Finally, the labeled proteins, which combined with the beads, were eluted using SDS-PAGE sample buffer and were analyzed by Western blotting.</p></sec><sec><title>Co-immunoprecipitation (co-IP)</title><p>Co-IP was used to investigate the interaction between hSlo and h&#x003b2;1 in HEK293 cells coexpressing Flag-hSlo-GFP and h&#x003b2;1. The experiment was conducted using Protein A/G PLUS Agarose Immunoprecipitation Reagent (Santa Cruz, TX, USA). Briefly, (2&#x02013;5)&#x000d7;10⁷ cells treated with DS-201 were lysed with mild lysis buffer (50 mmol/L Tris-Cl, pH 7.4, 150 mmol/L NaCl, 1 mmol/L EDTA, 1% NP-40, 0.5% sodium deoxycholate) containing protease inhibitors. One microgram (mg) of primary antibody (anti-Flag for hSlo or anti-His for h&#x003b2;1) was added to the sample and was incubated at 4 &#x000b0;C for 2 h. Then, the complex was incubated overnight with protein A/G agarose beads at 4 &#x000b0;C. The mixture was washed three times with cold PBS and was finally resuspended with 2&#x000d7; SDS sample buffer, was heated at 95 &#x000b0;C for 5 min and was centrifuged to acquire the supernatant. The samples were analyzed by Western blotting.</p></sec><sec><title>Flow cytometry to measure subcellular localization of channel proteins</title><p>Flow cytometry (FCM) was used to investigate the subcellular localization of BK_Ca channel proteins associated with protein trafficking in HEK293 cells expressing Flag-hSlo-GFP. Because the Flag tag was in the extracellular S1&#x02013;S2 loop, allophycocyanin (APC)-conjugated anti-Flag antibody could detect the levels of membranous BK_Ca channels, while GFP represented the global (cytoplasmic and membranous) levels of BK_Ca channels. To conduct the experiment, cells were collected and washed with cold PBS 3 times. Then, the cells were incubated with APC-conjugated anti-Flag antibody (1:200, Abcam, Cambridge, UK) at 4 &#x000b0;C for 2 h and were washed with cold PBS 3 times. Fluorescence was detected using the Accuri^&#x000ae; C6 cytometer (BD, MD, USA). The mean florescence intensity (MFI) was calculated with the following equation:</p><p><disp-formula id="equ1"><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="aps201485eq1.jpg"/></disp-formula></p><p>where FI_sample is the florescence intensity of the sample, FI_negative is the florescence intensity of the cells without expressing Flag-hSlo-GFP but incubated with APC conjugated anti-Flag antibody, and MFI_Flag and MFI_GFP are the expression levels of the membranous BK_Ca and total BK_Ca protein levels, respectively. The ratio MFI_Flag/MFI_GFP, representing the portion of membranous BK_Ca channel protein among the global pool of BK_Ca proteins, was used as an indicator of BK_Ca channel trafficking.</p></sec><sec><title>Flow cytometry to measure channel protein internalization</title><p>Forty-eight hours after transfection, HEK293 cells transiently expressing Flag-hSlo-GFP were harvested and stained with the mouse anti-Flag antibody (1:1000, Sigma, MO, USA) for 2 h at 4 &#x000b0;C. Then, the cells were switched to 37 &#x000b0;C for different time for the BK_Ca channels to internalize and for the internalization dynamic curves to be established. After this procedure, any remaining surface-labeled channels were stained with DyLight^&#x000ae; 650-conjugated donkey anti-mouse secondary antibody for 1 h (1:250, Abcam, Cambridge, UK) at 4 &#x000b0;C, and the fluorescence was detected using the Accuri^&#x000ae; C6 cytometer (BD, MD, USA). Because a portion of surface labeled BK_Ca channels would be internalized into the cytoplasm and could not be labeled by the secondary antibody, a decrease in fluorescence intensity denoted the internalization of the BK_Ca channels.</p></sec><sec><title>Confocal microscopy</title><p>HEK293 cells expressing Flag-hSlo-GFP were fixed in 4% paraformaldehyde for 15 min and were blocked with 5% BSA for 1 h. Whether 0.1% Triton-X 100 was used depended on the location at which the protein was detected. In detail, for staining cell surface BK_Ca channels, membrane permeabilization with Triton-X 100 was not performed, while this procedure was performed for 15 min to stain the global BK_Ca channel proteins. After blocking, the cells were incubated with the primary mouse anti-Flag antibody (1:2000, Sigma, MO, USA) overnight at 4 &#x000b0;C. DyLight^&#x000ae; 550-conjugated donkey anti-mouse secondary antibody (1:200, Abcam, Cambridge, UK) was incubated for 1 h at room temperature. Immunofluorescence-labeled samples were examined using an Olympus confocal laser scanning microscope (Tokyo, Japan). The laser lines (excitation/emission wave) were 358 nm/461 nm, 488 nm/507 nm, and 562 nm/576 nm for DAPI, GFP and DyLight^&#x000ae; 550-conjugated antibody, respectively. Negative control experiments were performed by pre-incubation of the primary antibody with the respective antigenic peptide (1:1), and these experiments did not show positive staining under the same experimental conditions.</p></sec><sec><title>Cytotoxicity assay with the Cell Counting Kit-8</title><p>The Cell Counting Kit-8 (CCK-8) was used to evaluate the potential cytotoxic effects of DS-201 on HEK293 cells and VSMCs. Briefly, cell suspension of 100 &#x003bc;L in volume was dispensed into each well of a 96-well plate (5&#x000d7;10³ cells/well) and was pre-incubated for 24 h. The cells were then exposed to various concentrations of DS-201 (0, 10, 20, 40, 80, 160, 320, 640, and 1280 &#x003bc;mol/L) and then were incubated for 12 h at 37 &#x000b0;C in an incubator. The wells were then washed and refilled with fresh culture medium. The cells in each well were incubated with 10 &#x003bc;L of CCK-8 solution for 2 h at 37 &#x000b0;C. Formazan was quantified spectroscopically at 450 nm using a microplate reader (Synergy&#x02122; 4, Biotek, VT, USA).</p></sec><sec><title>Statistical analysis</title><p>The data are expressed as mean&#x000b1;SEM. Student's <italic>t</italic> test and ANOVA were used for the statistical analysis, according to the experiments. A <italic>P</italic> value of &#x0003c;0.05 was considered statistically significant.</p><p>The relationship between drug concentration and normalized <italic>N</italic>Po was fitted to the Hill equation: <italic>y=x^b/(c^b+x^b</italic>). Here, <italic>x</italic> is the concentration of DS-201 or calcium, <italic>c</italic> is the half maximal effective concentration (EC₅₀) of DS-201 or Ca²⁺, the latter (EC₅₀ of Ca²⁺) reflecting the apparent Ca²⁺ sensitivity, and <italic>b</italic> is the slope factor (Hill coefficient, n_H). The conductance (<italic>G</italic>) of the BK_Ca channel macro-currents in the <italic>inside-out macro-patches</italic> was calculated by the slope of the I&#x02013;V curve, which was fitted to the following polynomial equation: <italic>y=a0+a1&#x000d7;x¹+a2&#x000d7;x²+...+a9&#x000d7;x⁹</italic>. <italic>G/G</italic>_max represents the normalized conductance (<italic>G</italic>) of BK_Ca channels at a <italic>V</italic>_m to the maximal conductance (<italic>G</italic>_max). Then, the <italic>G/G</italic>_max&#x02013;<italic>V</italic>_m curve was fitted by the Boltzmann function: <italic>G=1&#x02013;1/{1+exp[(V&#x02013;V_1/2</italic><italic>)/K]}</italic>. Here, <italic>V_1/2</italic> is the half maximal activation voltage, and <italic>K</italic> is the slope of the curve.</p></sec></sec><sec sec-type="results"><title>Results</title><sec><title>The electrophysiological properties of BK_Ca channels expressed in HEK293 cells</title><p>We first verified the electrophysiological characteristics of the BK_Ca channels heterologously expressed in HEK293 cells. <xref ref-type="fig" rid="fig2">Figure 2A</xref> shows a sketch map of the BK_Ca channel &#x003b1; subunit (hSlo) tagged with Flag and GFP. The Flag tag was inserted into the extracellular S1&#x02013;S2 loop of the &#x003b1; subunit and thus could detect the surface expression of hSlo by FCM and confocal microscopy, whereas the GFP tag was connected to the C-terminal of the &#x003b1; subunit and was used to detect the total cellular expression of hSlo. <xref ref-type="fig" rid="fig2">Figures 2B&#x02013;2E</xref> show the main electrophysiological properties of the BK_Ca channels expressed in HEK293 cells, which were consistent with those obtained from the native cells^{<xref ref-type="bibr" rid="bib13">13</xref>,<xref ref-type="bibr" rid="bib14">14</xref>}. The BK_Ca currents could not be recorded in HEK293 cells without the transfection of the BK_Ca channels (data not shown). <xref ref-type="fig" rid="fig2">Figure 2B</xref> shows typical recordings of the single BK_Ca channel currents under the inside-out patch configuration with the symmetrical 140 mmol/L K⁺ (<italic>V</italic>_m=+40 mV and [Ca²⁺]_i=0.1 &#x003bc;mol/L). Coexpression of the h&#x003b2;1 subunit increased the open time duration. Flag and GFP tags did not affect the open time duration. The <xref rid="tbl1" ref-type="table">Table 1</xref> provides the statistical results, showing that the Flag and GFP tags did not affect the kinetics and conductance of the single BK_Ca channel, while coexpression of the h&#x003b2;1 subunit increased the mean open time (<italic>T_o</italic>) and the mean closed time (<italic>T_c</italic>) of BK_Ca channels, compared with that of the &#x003b1; subunit expression alone. <xref ref-type="fig" rid="fig2">Figure 2C</xref> shows the Ca²⁺-dependent property of the BK_Ca channel opening. The BK_Ca currents were blocked by IbTX (a selective BK_Ca channel blocker) (200 nmol/L) when applied to the bath solution of the outside-out patch (<italic>V</italic>_m=+40 mV and [Ca²⁺]_i=0.1 &#x003bc;mol/L) (<xref ref-type="fig" rid="fig2">Figure 2D</xref>) and whole-cell experiments (<xref ref-type="fig" rid="fig2">Figure 2E</xref>).</p><p>In addition, confocal microscopy showed that anti-Flag antibody detected only the surface population of the BK_Ca channels (<xref ref-type="fig" rid="fig2">Figure 2F</xref>, the third panel of upper row), while after membrane permeabilization, anti-Flag antibody denoted the total BK_Ca population, as indicated by complete co-localization of Flag with GFP (<xref ref-type="fig" rid="fig2">Figure 2F</xref>, the fourth panel of lower row). Thus, the Flag tag was competent in examining the surface expression of the BK_Ca channels when without permeabilization, and Flag-hSlo-GFP was a valid and useful tool for studying BK_Ca channel trafficking.</p></sec><sec><title>Effects of DS-201 on BK_Ca channel currents recorded under inside-out patch and whole-cell configurations in HEK293 cells</title><p>We reported that DS-201 activated the BK_Ca channels mainly from the intracellular side of the membrane in mouse cerebral arterial VSMCs^{<xref ref-type="bibr" rid="bib14">14</xref>}. Here, we further investigated the effects of DS-201 on the BK_Ca channels expressed in HEK293 cells under whole-cell and inside-out configurations. <xref ref-type="fig" rid="fig3">Figure 3A</xref> shows typical recordings of single BK_Ca channel currents under an inside-out configuration in HEK293 cells, with or without co-expression of the h&#x003b2;1 subunit (<italic>V</italic>_m=+40 mV and [Ca²⁺]_i=0.1 &#x003bc;mol/L). DS-201 activated the BK_Ca channels in a concentration (0&#x02013;160 &#x003bc;mol/L)-dependent manner. Furthermore, the effects of DS-201 on BK_Ca channels were reversible after washout in the present study (data not shown). The curve of normalized <italic>N</italic>Po to DS-201 concentrations was fitted with Hill's equation (<xref ref-type="fig" rid="fig3">Figure 3B</xref>). When the [Ca²⁺]_i level was set to 0.1 &#x003bc;mol/L, the half maximum activation concentration (EC₅₀) was 62.04&#x000b1;1.07 &#x003bc;mol/L (with h&#x003b2;1, <italic>n</italic>=7) or 66.64&#x000b1;1.54 &#x003bc;mol/L (without h&#x003b2;1, <italic>n</italic>=6) (<italic>P</italic>&#x0003e;0.05). These results suggested that the &#x003b2;1 subunit did not affect the concentration-dependent effects of DS-201 on BK_Ca channels. DS-201 (80 &#x003bc;mol/L) also showed an agonist effect on the BK_Ca macroscopic currents recorded under a whole-cell configuration (<xref ref-type="fig" rid="fig3">Figures 3C</xref> and <xref ref-type="fig" rid="fig3">3D</xref>). At serial membrane voltages (from +30 mV to +70 mV), DS-201 increased the current densities (<italic>P</italic>&#x0003c;0.05 or <italic>P</italic>&#x0003c;0.01 <italic>vs</italic> control, <xref ref-type="fig" rid="fig3">Figure 3D</xref>).</p></sec><sec><title>Effects of DS-201 on the calcium and voltage dependence and on the kinetics of the BK_Ca channels expressed in HEK293 cells</title><p>We confirmed previously that DS-201 modulated the BK_Ca channels of mouse cerebral arterial VSMCs by increasing the calcium and voltage dependence and by shifting the channel kinetics^{<xref ref-type="bibr" rid="bib14">14</xref>}. The present study further investigated whether DS-201 exerted similar acute effects on the BK_Cachannels expressed in HEK293 cells. <xref ref-type="fig" rid="fig4">Figures 4A</xref> and <xref ref-type="fig" rid="fig4">4B</xref> show the effects of 20 &#x003bc;mol/L DS-201 on the calcium dependence of BK_Ca channels in an inside-out configuration (<italic>V</italic>_m=+40 mV). The open probability was normalized to the maximal <italic>N</italic>Po for each curve (<xref ref-type="fig" rid="fig4">Figure 4A</xref>). Coexpression of the h&#x003b2;1 subunit increased the calcium dependence of BK_Ca channels by shifting the EC₅₀ from 4.67&#x000b1;2.54 &#x003bc;mol/L (<italic>n</italic>=5) to 0.57&#x000b1;0.16 &#x003bc;mol/L (<italic>n</italic>=5), a result consistent with the previous report^{<xref ref-type="bibr" rid="bib7">7</xref>}. We reported that DS-201 at 20 &#x003bc;mol/L and 40 &#x003bc;mol/L increased the calcium dependence of BK_Ca channels in cerebral arterial VSMCs^{<xref ref-type="bibr" rid="bib14">14</xref>}. Here, we further demonstrated in HEK293 cells that 20 &#x003bc;mol/L DS-201, a concentration that did not activate BK_Ca channels significantly, shifted the EC₅₀ of BK_Ca channels for calcium activation from 0.57&#x000b1;0.16 &#x003bc;mol/L (control, <italic>n</italic>=5) to 0.30&#x000b1;0.03 &#x003bc;mol/L (with h&#x003b2;1, <italic>n</italic>=5) and from 4.67&#x000b1;2.54 &#x003bc;mol/L (control, <italic>n</italic>=5) to 1.38&#x000b1;0.12 &#x003bc;mol/L (without h&#x003b2;1, <italic>n</italic>=5), respectively (<xref ref-type="fig" rid="fig4">Figures 4A</xref> and <xref ref-type="fig" rid="fig4">4B</xref>). These results suggested that DS-201 increased the apparent calcium sensitivity of BK_Ca channels independently of the &#x003b2;1 subunit.</p><p>In addition, we examined the effects of DS-201 on the voltage dependence and kinetics of the BK_Ca channels (hSlo, without the h&#x003b2;1 subunit) in the inside-out macro-patches ([Ca²⁺]_i=0.1 &#x003bc;mol/L). <xref ref-type="fig" rid="fig4">Figure 4D</xref> shows that DS-201 (80 and 160 &#x003bc;mol/L) increased the current amplitude, consistent with <xref ref-type="fig" rid="fig3">Figure 3C</xref>. In addition, DS-201 shifted the <italic>G/G</italic>_max&#x02013;<italic>V</italic>_m curve (fitted from <xref ref-type="fig" rid="fig4">Figure 4D</xref>) leftward (<xref ref-type="fig" rid="fig4">Figure 4E</xref>), suggesting that DS-201 could activate BK_Ca channels at relatively more negative membrane potentials.</p><p>It is known that the current transition from a holding potential to a test pulse represents the shift of the BK_Ca channel from closed state to open state, and the tail current traces represent the transition from an open state to closed state. We found that 80 &#x003bc;mol/L and 160 &#x003bc;mol/L DS-201 did not affect the transition from closed state to open state (<xref ref-type="fig" rid="fig4">Figure 4C</xref>). The time constants of activation were 0.35 ms (control), 0.26 ms (80 &#x003bc;mol/L DS-201), and 0.31 ms (160 &#x003bc;mol/L DS-201). However, the same DS-201 concentrations (80 and 160 &#x003bc;mol/L) significantly slowed the transition from open state to closed state (<xref ref-type="fig" rid="fig4">Figure 4C</xref>). The time constants of deactivation were 0.25 ms (control), 0.91 ms (80 &#x003bc;mol/L DS-201), and 10.72 ms (160 &#x003bc;mol/L DS-201). These data were consistent with those observation in mouse cerebral arterial VSMCs in our previous report^{<xref ref-type="bibr" rid="bib14">14</xref>}, suggesting that DS-201 could bind to the BK_Ca channel &#x003b1; subunit directly and could retain it in the open state, thus inhibiting the transition from open to closed state.</p></sec><sec><title>DS-201 enhanced the expressions of the &#x003b1; subunit, but not of the &#x003b2;1 subunit, of BK_Ca channels in cultured VSMCs and HEK293 cells</title><p>It is well recognized that the &#x003b2;1 subunit increases the expression of the &#x003b1; subunit of BK_Ca channels in VSMCs^{<xref ref-type="bibr" rid="bib16">16</xref>}. Therefore, in addition to observing the &#x0201c;acute&#x0201d; effects of DS-201 on channel electrophysiological properties, we further investigated whether prolonged use of DS-201 would affect the expression of BK_Ca channel subunits both in cultured VSMCs and in HEK293 cells expressing the &#x003b1; and &#x003b2;1 subunits. After incubation of cells with DS-201 for 12 h, Western blotting was performed. As shown in <xref ref-type="fig" rid="fig5">Figures 5A</xref> and <xref ref-type="fig" rid="fig5">5B</xref>, DS-201 at 80 &#x003bc;mol/L increased the total and membranous expression levels of hSlo in HEK293 cells. However, DS-201 at the same concentration did not affect &#x003b2;1 subunit expression. A statistical summary of the subunit expression with or without DS-201 treatment is shown in <xref ref-type="fig" rid="fig5">Figure 5C</xref>. In cultured VSMCs, DS-201 at 80 &#x003bc;mol/L also increased the total protein expression level of the &#x003b1; subunit but did not affect the expression of the &#x003b2;1 subunit (<xref ref-type="fig" rid="fig5">Figure 5D</xref>), a result consistent with that observed in HEK293 cells. Confocal images further proved that 80 &#x003bc;mol/L DS-201 promoted the expression of the &#x003b1; subunit (<italic>green</italic>) and did not affect the expression of the &#x003b2;1 subunit (<italic>red</italic>) in cultured VSMCs (<xref ref-type="fig" rid="fig5">Figure 5E</xref>).</p><p>Furthermore, we investigated using co-IP assay whether DS-201 affected the expression of hSlo by accelerating the interaction of hSlo with h&#x003b2;1. The anti-Flag (for hSlo) and anti-His (for h&#x003b2;1) antibodies were used as the baits for immunoprecipitation. We calculated the portion of h&#x003b2;1 co-immunoprecipitated by hSlo, and vice versa. DS-201 did not change the ratio of hSlo to h&#x003b2;1 when either the anti-Flag or the anti-His antibody was used (<xref ref-type="fig" rid="fig5">Figure 5F</xref>). These results suggested that DS-201 accelerated the expression of hSlo, but not by affecting the interaction between hSlo and h&#x003b2;1.</p></sec><sec><title>DS-201 accelerated the trafficking of the &#x003b1; subunit independently of the &#x003b2;1 subunit of BK_Ca channels in HEK293 cells</title><p>Many ion channels function only when their subunits are transferred from the Golgi apparatus to the plasma membrane. Thus, investigation of channel trafficking is necessary for identifying the action of a drug on ion channels. Whether DS-201 promotes the trafficking of the pore-forming &#x003b1; subunit of BK_Ca channels is unknown. To address this question, plasmid Flag-hSlo-GFP was transfected to HEK293 cells to detect the subcellular localization of the &#x003b1; subunit, with the help of FCM and confocal microscopy. Incubation of the transfected HEK293 cells with DS-201 (80 &#x003bc;mol/L) for 12 h increased the membranous level of the &#x003b1; subunit (<xref ref-type="fig" rid="fig6">Figure 6</xref>). DS-201 at 80 &#x003bc;mol/L increased the percentage of GFP-positive cells from 31.7%&#x000b1;1.1% to 41.1%&#x000b1;1.3% (<italic>n</italic>=8, <italic>P</italic>&#x0003c;0.05) (<xref ref-type="fig" rid="fig6">Figures 6A</xref> and <xref ref-type="fig" rid="fig6">6B</xref>), indicating again that DS-201 increased the global expression level of the &#x003b1; subunit. In addition, within the GFP-positive cells, DS-201 (80 &#x003bc;mol/L) also increased the percentage of Flag-positive cells from 61.8%&#x000b1;1.7% to 75.2%&#x000b1;1.0% (<xref ref-type="fig" rid="fig6">Figure 6B</xref>) and shifted the MFI_Flag/GFP (an indicator of the membranous &#x003b1; subunit relative to the total &#x003b1; subunit pool) from 0.13 to 0.18 (<xref ref-type="fig" rid="fig6">Figure 6C</xref>), suggesting that DS-201 accelerated the trafficking of the &#x003b1; subunit toward the cell membrane. The confocal images further provided intuitive evidence that DS-201 (80 &#x003bc;mol/L) increased the membranous level of the &#x003b1; subunit in HEK293 cells (<xref ref-type="fig" rid="fig6">Figure 6D</xref>).</p></sec><sec><title>DS-201 stabilized the membranous retention of BK_Ca channel proteins in HEK293 cells</title><p>Endocytosis is an important mechanism for downregulating the functional expression of channels in the cell membrane. Therefore, the increase in membranous BK_Ca channels shown above might partially have been caused by decreased internalization (endocytosis) when DS-201 was administered. If this relationship was true, the BK_Ca channels might remain in the plasma membrane for a longer time. Here, we used flow cytometry to confirm this possibility. Surface BK_Ca channels were initially labeled with the anti-Flag antibody at 4 &#x000b0;C, and then the cell samples were switched to 37 &#x000b0;C to allow the channel protein to internalize for different lengths of time. After these procedures, some channels were supposed to be internalized, but others were not. The channels that were not internalized were stained with DyLight^&#x000ae;650-conjugated donkey anti-mouse IgG antibody and were detected by flow cytometry (<xref ref-type="fig" rid="fig7">Figure 7A</xref>). <xref ref-type="fig" rid="fig7">Figure 7B</xref> shows the single exponential curves indicating the channel internalization dynamics over time. The curves revealed that DS-201 decreased the internalization speed, with a shift of the <italic>t</italic>_1/2 from 6.04 min (control, <italic>n</italic>=5) to 11.21 min (DS-201, <italic>n</italic>=5) (<italic>P</italic>&#x0003c;0.05). <xref ref-type="fig" rid="fig7">Figure 7C</xref> shows the fluorescence intensity (FI) 60 min after DS-201 was administered: DS-201 reduced the decay of FI compared with the control, suggesting that DS-201 stabilized the anchoring of the BK_Ca channel in the plasma membrane to some extent.</p></sec><sec><title>DS-201 induced cell death only at very high concentrations</title><p>The potential cytotoxicity of DS-201 on HEK293 cells and VSMCs was evaluated by CCK-8 assay. DS-201 exposure for 12 h at relatively lower concentrations (10, 20, 40, 80, and 160 &#x003bc;mol/L) did not significantly affect cell viability. However, DS-201 at 320 &#x003bc;mol/L or higher concentration induced cell death and decreased cell viability in a dose-dependent manner (<xref ref-type="fig" rid="fig8">Figure 8</xref>). Therefore, the concentrations of DS-201 used in most of the experiments in this study (160 &#x003bc;mol/L or lower) were safe to cells and did not exert significant cytotoxic effects on HEK293 cells or VSMCs.</p></sec></sec><sec sec-type="discussion"><title>Discussion</title><p>It is well known that the native BK_Ca channel consists of four &#x003b1; subunits and four &#x003b2; subunits (&#x003b2;1-&#x003b2;4). The &#x003b1; subunit, encoded by <italic>Slo</italic> or the <italic>KCNMA1</italic> gene, is a pore-forming portion of the channel. <italic>Slo</italic> is the only gene encoding the &#x003b1;-subunit, and <italic>KCNMB1</italic> was the gene encoding the &#x003b2;1 subunit in VSMCs^{<xref ref-type="bibr" rid="bib2">2</xref>,<xref ref-type="bibr" rid="bib4">4</xref>,<xref ref-type="bibr" rid="bib17">17</xref>}. The present study first revealed that DS-201 modulated BK_Ca channels in a novel manner, independent of the &#x003b2;1 subunit. DS-201 directly modulated the pore-forming &#x003b1; subunit by increasing the calcium and voltage dependency and shifting the channel kinetics. These might be considered the acute electrophysiological effects of DS-201 on BK_Ca channels. In addition, DS-201 promoted the expression, membranous retention and trafficking of the &#x003b1; subunit when DS-201 acted for a longer period of time (12 h).</p><p>The role of the &#x003b2; subunit in BK_Ca channels has been extensively investigated using co-expression of different &#x003b2; subunits with the &#x003b1; subunit, including &#x003b2;1 to &#x003b2;4^{<xref ref-type="bibr" rid="bib18">18</xref>,<xref ref-type="bibr" rid="bib19">19</xref>}. As a whole, the &#x003b2; subunit shifts the voltage-dependent characteristics of the channel activity to a direction of more negative membrane potential by increasing Ca²⁺ sensitivity. The second role of the &#x003b2; subunit is that it increases the membrane expression and trafficking of BK_Ca channels^{<xref ref-type="bibr" rid="bib16">16</xref>}. Recently, we reported for the first time that, in Han Chinese patients with primary hypertension, the whole-cell current and Ca²⁺ sensitivity of BK_Ca channels were reduced in VSMCs due to downregulation of the &#x003b2;1 subunit but not of the &#x003b1; subunit^{<xref ref-type="bibr" rid="bib9">9</xref>}. These results suggested that the &#x003b1; and &#x003b2; subunits of BK_Ca channels were rather independently regulated by different molecular mechanisms. The basic function of BK_Ca channels can be expressed by the &#x003b1; subunit alone, and the &#x003b2; subunit has only supplementary action on BK_Ca channel function. There have been reports showing that some chemical molecules modified BK_Ca channel function by interacting with the &#x003b2; subunit^{<xref ref-type="bibr" rid="bib20">20</xref>}, whereas other molecules regulated the channel function by direct interaction with the &#x003b1; subunit^{<xref ref-type="bibr" rid="bib21">21</xref>}.</p><p>Here, we investigated the effects of DS-201 on BK_Ca channels, including the electrophysiological properties, channel protein expression and trafficking, with the convenience of HEK293 cell line, which allows for heterologous expression of the &#x003b1; subunit with or without the &#x003b2;1 subunit. Cultured VSMCs were also used as native vascular cells in this study. We first found that DS-201 activated BK_Ca channels reversibly in a dose-dependent manner, and the &#x003b2;1 subunit was not involved in this effect. The EC₅₀ of DS-201 was 62.04 &#x003bc;mol/L with the &#x003b2;1 subunit and 66.64 &#x003bc;mol/L without it. These EC₅₀ values were similar to our previous observation (68.5 &#x003bc;mol/L) in mouse cerebral arterial VSMCs^{<xref ref-type="bibr" rid="bib14">14</xref>}. Based on these observations, we used 80 &#x003bc;mol/L DS-201 in most of the experiments. At this concentration, DS-201 increased the whole-cell BK_Ca currents significantly. We further observed previously that 20 and 40 &#x003bc;mol/L DS-201 increased the apparent calcium sensitivity of BK_Ca channels in mouse cerebral arterial VSMCs^{<xref ref-type="bibr" rid="bib14">14</xref>}. The present study also chose a low concentration (20 &#x003bc;mol/L) of DS-201, which did not activate the BK_Ca channels but could increase the apparent calcium sensitivity of the channel, to test the effects of DS-201 on the calcium dependence of this channel in HEK293 cells expressing either the &#x003b1; subunit alone or with the &#x003b2;1 subunit. The results showed that 20 &#x003bc;mol/L DS-201 enhanced the apparent calcium sensitivity of the channel in HEK293 cells, and this effect was independent of the &#x003b2;1 subunit. However, it is interesting that DS-201, at lower concentrations (20 and 40 &#x003bc;mol/L), did not shift the voltage dependence and open/closed kinetics of the channel. The reason for this finding could be that BK_Ca channels are more sensitive to calcium than to voltage, or other mechanisms were involved in the phenomenon. As such, we used higher DS-201 concentrations (80 and 160 &#x003bc;mol/L) to avoid potential disputes in explaining the mechanisms. The data revealed that DS-201 increased the voltage dependence and inhibited the transition of the channel from the open state to the closed state in HEK293 cells. These results were quite consistent with those observed in the native BK_Ca channels of mouse cerebral VSMCs^{<xref ref-type="bibr" rid="bib14">14</xref>}. We further showed that DS-201 increased the protein expression of the &#x003b1; subunit but did not affect the expression of the &#x003b2;1 subunit, either in HEK293 cells or in cultured VSMCs. DS-201 also did not affect the interaction of the &#x003b1; and &#x003b2;1 subunits, as shown by the co-IP assay. Taken together, these data reveal that the &#x003b2;1 subunit did not contribute to the effects of DS-201 on BK_Ca function, including those effects on channel open probability, calcium and voltage dependence, shifting of channel kinetics, and channel &#x003b1; subunit expression. These effects of DS-201 are unique and quite different from those of other substances, compared even with many other Chinese medicinal herbs with vasorelaxing effects, such as Puerarin. The latter could induce vasodilation, and the &#x003b2;1 subunit was involved in this effect^{<xref ref-type="bibr" rid="bib22">22</xref>}.</p><p>As mentioned above, channel proteins must be transported and inserted to the cell membrane to function. Therefore, controlling channel trafficking dynamics with drugs could serve as a new pharmacological approach in treating diseases. We reported that BK_Ca currents were decreased in hypertension, together with the downregulation of &#x003b2;1 subunit expression but not &#x003b1; subunit expression^{<xref ref-type="bibr" rid="bib9">9</xref>}. Functional defects of the &#x003b2;1 subunit could lead to a reduction in BK_Ca &#x003b1; subunit trafficking, and the &#x003b1; subunit would be retained in the cytoplasm and could not be targeted to the cell membrane efficiently.</p><p>The trafficking processes of proteins include forward trafficking (toward the plasma membrane), internalization (endocytosis) and recycling to the membrane. Any change in one or more of these processes will affect the expression levels of channels in the plasma membrane. The present study first found that DS-201 increased the gene expression of the BK_Ca &#x003b1; subunit without affecting the &#x003b2;1 subunit, and, furthermore, DS-201 enhanced the trafficking of the BK_Ca channel &#x003b1; subunit again independently of the &#x003b2;1 subunit. However, we could not accurately evaluate whether one or more steps were involved in the effects of DS-201, because the three steps were dynamic and connected each other and were difficult to distinguish. We found that DS-201 slowed the internalization and stabilized the &#x003b1; subunit's anchoring in the plasma membrane. It is also possible that the recycling process was involved in the actions of DS-201, and this possibility might have affected the interpretations of the results. However, a study by McEwen <italic>et al</italic>^{<xref ref-type="bibr" rid="bib23">23</xref>} using similar methods, indicated that the amounts of recycling from internalization of the Kv1.5 channel were small (less than 10% for recycling and approximately 30% for internalization), and the recycling kinetics were slower than those of internalization. The present study showed that DS-201 increased the membranous levels of BK_Ca channels (an indicator of the comprehensive trafficking processes) by approximately 38.5%, as indicated by the MFI_Flag/GFP ratio (<xref ref-type="fig" rid="fig6">Figure 6C</xref>). However, in an experiment that examined only the internalization and potential recycling processes (<xref ref-type="fig" rid="fig7">Figure 7</xref>), the difference in membranous BK_Ca levels before and after DS-201 treatment was less than 20% (<xref ref-type="fig" rid="fig7">Figure 7C</xref>). Therefore, we inferred that DS-201 could accelerate forward trafficking. Certainly, it was a limitation of this study that we did not quantitatively examine the forward trafficking. If taken together, the electrophysiological results and the expression and trafficking results of this study demonstrated that DS-201 potentiated the function of the BK_Ca channel and did not require the presence of the &#x003b2;1 subunit. Therefore, DS-201 had complementary effects in diseases with &#x003b2;1 subunit deficiency, such as hypertension. These extraordinary effects of DS-201 on BK_Ca channels favored more widespread use of Danshen in cardiovascular medicine.</p><p>The data presented here suggested an interaction between DS-201 and the hSlo subunit, and this action could lead to enhancement of BK_Ca channel activity. Considering the amino acid compositions of the &#x003b2;1 subunit, there were 118 amino acid residues in the extracellular loop, while only 18 residues and 13 residues were located in the N- and C-terminals, respectively. However, DS-201 modulated the BK_Ca channels mainly from the cytoplasmic side of the membrane, so there was little opportunity for DS-201 to bind to the cytoplasmic N- or C-terminal of the &#x003b2;1 subunit. We infer that DS-201 could bind to the &#x003b1;-subunit directly, but the exact binding site is unknown and requires further study. Because BK_Ca channel activity in VSMCs generated outward currents that drive the membrane potential in the negative direction, eventually counteracting vascular contraction, DS-201-induced activation of BK_Ca could be an underlying mechanism, or at least contributing to the Danshen-induced relaxation of VSMCs.</p></sec><sec><title>Author contribution</title><p>Xiao-qiu TAN, Xiu-li CHENG, Xiao-rong ZENG, and Ji-min CAO designed the research; Xiao-qiu TAN, Xiu-li CHENG, Yan YANG, Li YAN, Jing-li, GU, and Hui LI performed the experiments; Xiao-qiu TAN, Xiu-li CHENG, and Ji-min CAO analyzed the data; Xiao-qiu TAN, Xiu-li CHENG, and Ji-min CAO wrote the paper.</p></sec>

Interest of low-dose hydrocortisone therapy during brain-dead organ donor resuscitation: the CORTICOME study

Could not extract abstract

<contrib contrib-type="author"><name><surname>Pinsard</surname><given-names>Michel</given-names></name><address><email>Michel.Pinsard@chu-poitiers.fr</email></address><xref ref-type="aff" rid="Aff69"/></contrib><contrib contrib-type="author"><name><surname>Ragot</surname><given-names>St&#x000e9;phanie</given-names></name><address><email>s.ragot@chu-poitiers.fr</email></address><xref ref-type="aff" rid="Aff70"/></contrib><contrib contrib-type="author"><name><surname>Mertes</surname><given-names>Paul Michel</given-names></name><address><email>paul-michelmertes@chru-strasbourg.fr</email></address><xref ref-type="aff" rid="Aff71"/></contrib><contrib contrib-type="author"><name><surname>Bleichner</surname><given-names>Jean Paul</given-names></name><address><email>jean-paul.bleichner@chu-rennes.fr</email></address><xref ref-type="aff" rid="Aff72"/></contrib><contrib contrib-type="author"><name><surname>Zitouni</surname><given-names>Samira</given-names></name><address><email>zitouni-s@chu-caen.fr</email></address><xref ref-type="aff" rid="Aff73"/></contrib><contrib contrib-type="author"><name><surname>Cook</surname><given-names>Fabrice</given-names></name><address><email>fabrice.cook@hmn.aphp.fr</email></address><xref ref-type="aff" rid="Aff74"/></contrib><contrib contrib-type="author"><name><surname>Pierrot</surname><given-names>Marc</given-names></name><address><email>mapierrot@chu-angers.fr</email></address><xref ref-type="aff" rid="Aff75"/></contrib><contrib contrib-type="author"><name><surname>Dube</surname><given-names>Laurent</given-names></name><address><email>LaDube@chu-angers.fr</email></address><xref ref-type="aff" rid="Aff76"/></contrib><contrib contrib-type="author"><name><surname>Menguy</surname><given-names>Edgard</given-names></name><address><email>edgard.menguy@chu-rouen.fr</email></address><xref ref-type="aff" rid="Aff77"/></contrib><contrib contrib-type="author"><name><surname>Lef&#x000e8;vre</surname><given-names>Laurent Martin</given-names></name><address><email>LaDube@chu-angers.fr</email></address><xref ref-type="aff" rid="Aff78"/></contrib><contrib contrib-type="author"><name><surname>Escaravage</surname><given-names>Laurence</given-names></name><address><email>lescaravage@chu-clermontferrand.fr</email></address><xref ref-type="aff" rid="Aff79"/></contrib><contrib contrib-type="author"><name><surname>Dequin</surname><given-names>Pierre-Fran&#x000e7;ois</given-names></name><address><email>dequin@med.univ-tours.fr</email></address><xref ref-type="aff" rid="Aff80"/></contrib><contrib contrib-type="author"><name><surname>Vignon</surname><given-names>Philippe</given-names></name><address><email>philippe.vignon@unilim.fr</email></address><xref ref-type="aff" rid="Aff81"/></contrib><contrib contrib-type="author" corresp="yes"><name><surname>Pichon</surname><given-names>Nicolas</given-names></name><address><email>nicolas.pichon@chu-limoges.fr</email></address><xref ref-type="aff" rid="Aff82"/></contrib><aff id="Aff69"><label/>Intensive Care Unit, Inserm U 1082, University Hospital Mil&#x000e9;trie, Poitiers, 86000 France </aff><aff id="Aff70"><label/>Center of Clinical Investigation, Inserm 0802, Poitiers, 86000 France </aff><aff id="Aff71"><label/>Department of Anesthesiology, Inserm U 1116, University Hospital of Strasbourg, Nancy, 54000 France </aff><aff id="Aff72"><label/>Intensive Care Unit, University Hospital Pontchaillou, Rennes, 35000 France </aff><aff id="Aff73"><label/>Intensive Care Unit, University Hospital C&#x000f4;te de Nacre, Caen, 14000 France </aff><aff id="Aff74"><label/>Intensive Care Unit, University Hospital Henri Mondor, Cr&#x000e9;teil, 94010 France </aff><aff id="Aff75"><label/>Critical Care Department, University Hospital of Angers, Angers, 49100 France </aff><aff id="Aff76"><label/>Intensive Care Unit, University Hospital of Angers, Angers, 49100 France </aff><aff id="Aff77"><label/>Intensive Care Unit, University Hospital of Rouen, Rouen, 76000 France </aff><aff id="Aff78"><label/>Intensive Care Unit, Hospital Les Oudairies, La Roche-sur-Yon, 85925 France </aff><aff id="Aff79"><label/>Department of Anesthesiology, University Hospital of Clermont-Ferrand, Clermont-Ferrand, 63000 France </aff><aff id="Aff80"><label/>Critical Care Department, University Hospital Bretonneau, Tours, 37000 France </aff><aff id="Aff81"><label/>Intensive Care Unit, University Hospital of Limoges, Limoges, 87042 France </aff><aff id="Aff82"><label/>Center of Clinical Investigation, INSERM 1435, CHU Dupuytren, 2 Avenue Martin Luther King, Limoges, 87042 France </aff>

Critical Care

<sec id="Sec1" sec-type="intro"><title>Introduction</title><p>Currently in France, increasing transplantation indications cannot be met because of graft shortage. It has been proven that the amount of procured grafts can be increased by an optimized management of brain-dead patients [<xref ref-type="bibr" rid="CR1">1</xref>&#x02013;<xref ref-type="bibr" rid="CR5">5</xref>]. Almost 80% of brain-dead patients exhibit circulatory failure and it is commonly associated with heart, lung, kidney, liver and pancreas dysfunction, which compromises organ procurement in 10% to 25% of cases [<xref ref-type="bibr" rid="CR6">6</xref>]. Although circulatory failure is controlled in more than 60% of brain-dead patients, primary function recovery of the grafts, especially heart, liver and pancreas grafts can be altered by increased vascular filling and administered vasopressors [<xref ref-type="bibr" rid="CR7">7</xref>&#x02013;<xref ref-type="bibr" rid="CR9">9</xref>].</p><p>Thyroid hormones and cortisol deficit have already been identified as circulatory failure factors [<xref ref-type="bibr" rid="CR10">10</xref>&#x02013;<xref ref-type="bibr" rid="CR12">12</xref>]. Thus, current British guidelines recommend giving thyroid hormones and corticosteroids to brain-dead patients with circulatory failure [<xref ref-type="bibr" rid="CR13">13</xref>&#x02013;<xref ref-type="bibr" rid="CR15">15</xref>]. Several studies showed an increased amount of procured grafts and less primary dysfunction in transplanted heart grafts when triple therapy with thyroid hormone, corticosteroids and arginine vasopressin was used [<xref ref-type="bibr" rid="CR6">6</xref>, <xref ref-type="bibr" rid="CR16">16</xref>]. However, systematic triple therapy remains debated because these studies are retrospective, whereas donor characteristics have considerably evolved with time (mean age of included patients in these studies is significantly inferior to the mean age of present potential donors for example), because the groups are not comparable and because the respective contribution of each agent of the triple therapy administered remains controversial.</p><p>This prospective multicenter study aimed to demonstrate that systemic administration of low-dose steroids during resuscitation of brain-dead donors makes vasopressor weaning possible in 25% of patients and also decreases by more than 15% the quantity of vasopressors needed to control circulatory failure. It also aimed at studying the impact of steroid administration on primary function recovery of grafts and on the number of procured grafts compared to the amount of potential donors who died with brain death.</p></sec><sec id="Sec2" sec-type="materials|methods"><title>Material and methods</title><p>The study was approved by the French human subject protection committee (CPP OUEST III, number 061026), which waived the need for written informed consent from the family. In fact, French law entitles the conduct of randomized studies among clinically deceased patients without any informed consent from the family of the patient concerned. The study was conducted in accordance with the ethical standards of the Declaration of Helsinski as well as the Declaration of Istanbul and in compliance with French guidelines on organ procurement. The families of the donors were informed of the study. The French <italic>Registre National des Refus</italic> was consulted systematically (mandatory in France) to eliminate any opposition of the donor to participation in a clinical trial or organ donation. The authors decided to conduct a prospective cluster study involving 22 ICUs during 15&#x000a0;months to compare two different resuscitation strategies: systematic hydrocortisone supplementation (steroid group) or no supplementation (control group) in brain-dead patients who were potential organ donors. Among the involved centers, 11 gave standard-care, low-dose hydrocortisone to brain-dead patients before organ procurement and 11 did not use that therapy. All the patients were treated in the same way in a given center and all the centers introduced vasopressor support on the same hemodynamic criteria so minimizing the bias bound to the absence of randomization and to the potential center effect.</p><sec id="Sec3"><title>Patients</title><p>All brain-dead patients, over 18&#x000a0;years old, hospitalized in the involved centers and considered for organ procurement were eligible for the study. Brain death was determined according to the usual criteria of French law: no motor response to nociceptive stimuli, no more brain stem reflexes and no more spontaneous breathing, which was confirmed with an apnea test when there were no other confounding factors such as hypothermia, collapsus and impregnation of the central nervous system by depressant drugs. Patients were included once the family has been informed about the study protocol. Patients over 18&#x000a0;years old, who received corticosteroids or had adrenal insufficiency before brain death, were excluded. Patients registered in <italic>Registre National des Refus</italic> (in France, where refusals to donate organs are registered) were secondarily excluded from the study. Two groups of equal numbers and patient ages could not be obtained because of the illegality of randomization of brain-dead patients in France and because of the cluster distribution of recruited patients.</p></sec><sec id="Sec4"><title>Hemodynamic evaluation</title><p>All patients were monitored through an artery catheter and a central venous catheter in the superior vena cava. Patients received controlled ventilation with a tidal volume (VT) of 7 to 10&#x000a0;ml/kg, positive end-expiratory pressure (PEEP) of 0 to 5 cmH₂O, and an adjusted rate to maintain normocapnia. Hypotension was defined as mean blood pressure &#x0003c;65&#x000a0;mmHg. Mean blood pressure was set between 65 and 75&#x000a0;mmHg. The preload dependence was evaluated with respiratory variation of pulse blood pressure or with variation of subaortic doppler speeds or echocardiography-measured vena cava diameter; common threshold values previously defined in the protocol were used [<xref ref-type="bibr" rid="CR17">17</xref>]. Vascular filling was performed with 500- to 1,000-ml aliquots of crystalloids administered during 10 to 15&#x000a0;minutes. Norepinephrine therapy was started if persistent hypotension and no preload dependence criteria were noted. If the mean blood pressure was superior to 85&#x000a0;mmHg, norepinephrine dose was decreased until complete vasopressor weaning.</p><p>When diabetes insipidus (defined by diuresis &#x0003e;3&#x000a0;ml/kg/h and urine density &#x0003c;1003 gm/cm³) occurred, diuresis needed to be offset volume for volume and desmopressin was administered to maintain diuresis between 1 and 3&#x000a0;ml/kg/h and natremia &#x0003c;160&#x000a0;mmol/L.</p></sec><sec id="Sec5"><title>Study design</title><p>Administration of replacement dose of hydrocortisone had to be started at a maximum 6&#x000a0;h after the diagnosis of brain death. Adrenal stimulation by adrenocorticotrophic hormone (ACTH) (250&#x000a0;&#x003bc;g injection of tetracosactrin, Synacthen^R; Novartis Pharma SAS, Rueil Malmaison, France) was investigated. Adrenal insufficiency was defined by plasma cortisol level inferior to 18&#x000a0;&#x003bc;g/dl at time of injection (zero minutes, T0) and/or by a variation of plasma cortisol level following ACTH injection (T60 to T0) inferior to 9&#x000a0;&#x003bc;g/dl (so-called non-responding patients) [<xref ref-type="bibr" rid="CR18">18</xref>]. After ACTH injection, when plasma cortisol level was superior to 9&#x000a0;&#x003bc;g/dl, patients were classed as responding. Patient then received a 50-mg injection of hydrocortisone (Roussel-Uclaf, Romainville, France) followed by a continuous infusion of 10&#x000a0;mg/h until the aortic clamping was performed in the operating room during organ retrieval. Plasma cortisol assays were done before ACTH injection (T0) and 60&#x000a0;minutes after injection (T60) by electrogenerated chemiluminescence (Roche automat modular). In the control group (patients did not receive hydrocortisone), the physician in charge of the patient decided whether or not to perform the ACTH test.</p></sec><sec id="Sec6"><title>Appraisal criteria</title><p>The main appraisal criterion was the quantity of norepinephrine weaning possible during resuscitation of brain-dead donors, or the decrease of the quantity of vasopressors needed to control circulatory failure. It was quantified with the average dose per hour, the variation percentage and the duration of administration of norepinephrine from the time of inclusion into the study (when brain death was diagnosed) to the aortic clamping during organ procurement.</p><p>Secondary appraisal criteria were: number of recovered organs compared with number of brain-dead donors and with number of organs considered for procurement when grafts were proposed to the <italic>Agence de la biom&#x000e9;decine</italic> (French organization responsible for census of brain-dead patients and for national distribution of grafts). Another secondary criterion was the frequency of delayed graft function (DGF) for each graft. For each organ, DGF was determined from clinical and biological data usually considered by French transplant physicians (Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>). Pancreas grafts were counted among recovered organs but primary dysfunction analyses were not studied since these organs were recovered for islets of Langerhans transplantation and not for whole organ transplantation. Finally, cold ischemia duration was registered for each recovered and transplanted organ.<table-wrap id="Tab1"><label>Table 1</label><caption><p> <bold>Delayed graft function (DGF) criteria</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr><th>Organs</th><th>DGF criteria</th></tr></thead><tbody><tr><td>Kidney</td><td align="center">Hemodialysis during first week or creatinine &#x0003e;250&#x000a0;&#x003bc;mol/L at day&#x02009;+&#x02009;10</td></tr><tr><td>Liver</td><td align="center">ASAT &#x0003e;1,500 or ALAT &#x0003e;1,000 or Quick Time &#x0003c;30% at h&#x02009;+&#x02009;72</td></tr><tr><td>Heart</td><td align="center">Circulatory support or left ventricle dysfunction or right ventricle dysfunction with pulmonary hypertension</td></tr><tr><td>Lungs</td><td align="center">PaO2/FiO2&#x02009;&#x0003e;&#x02009;100&#x000a0;mmHg and &#x0003c;300&#x000a0;mmHg and/or pulmonary edema</td></tr></tbody></table><table-wrap-foot><p>PaO₂/FiO₂, arterial partial pressure of oxygen/inspired oxygen fraction.</p><p>ASAT: Aspartate Aminotransferase.</p><p>ALAT: Alanine Aminotransferase.</p></table-wrap-foot></table-wrap></p></sec><sec id="Sec7"><title>Statistical analysis</title><p>Statistical analyses were performed using the SAS 9.2 software package (SAS Inc, Cary, NC, USA) and Statview 5.0 software (SAS Institute, Berkeley, CA, USA). Continuous variables were expressed as mean&#x02009;&#x000b1;&#x02009;SD and qualitative variables were expressed as absolute numbers and percentages. Comparisons between the steroid group and control group were performed using the Student <italic>t</italic>-test, or Mann-Whitney <italic>U</italic>-test when appropriate for the quantitative variables, and the chi-square test for the qualitative variables. For paired donor-kidneys, comparison of DGF between the two groups was adjusted for cold ischemia duration, serum creatinine value, norepinephrine dose, and simplified acute physiological score II (SAPSII), using logistic regression analysis with random effects to account for impairment, performed using the PROC GLIMMIX command in SAS. Kaplan-Meier curves were plotted to describe the probability of norepinephrine weaning. Curves were compared between groups using the log-rank test. The corresponding hazard ratio (HR) was estimated using a univariate Cox model. All statistical tests were two-sided and were conducted using the 0.05 level of significance.</p></sec></sec><sec id="Sec8" sec-type="results"><title>Results</title><p>During the study, 631 brain-dead patients were hospitalized in the 22 participating centers. Hemorrhage within the brain or meninges represented 72% of the etiologies of brain insult responsible for brain death (Table&#x000a0;<xref rid="Tab2" ref-type="table">2</xref>). Organs were recovered from 304 patients (48%) and 259 donors (41%) were included in the study. Finally, 208 donors (33%) were analyzed: 128 brain-dead patients were included in the control group and 80 in the steroid group (Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>). The mean age of each group was similar but in the steroid group, the number of patients over 65&#x000a0;years old was significantly higher, with an initial severity score represented by the highest SAPSII (Table&#x000a0;<xref rid="Tab2" ref-type="table">2</xref>). The mean dose of hydrocortisone received in the steroid group was 210&#x02009;&#x000b1;&#x02009;35&#x000a0;mg.<table-wrap id="Tab2"><label>Table 2</label><caption><p> <bold>Global population - general characteristics</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr><th/><th>All (n&#x02009;=&#x02009;208)</th><th>Control group (n&#x02009;=&#x02009;128)</th><th>Steroid group (n&#x02009;=&#x02009;80)</th><th> <bold><italic>P</italic></bold>-value</th></tr></thead><tbody><tr><td> <bold>Characteristics</bold> </td><td align="center"/><td align="center"/><td align="center"/><td align="center"/></tr><tr><td>Age, years, mean (SD)</td><td align="center">51.1 (16.7)</td><td align="center">48.1 (16.1)</td><td align="center">56.1 (16.6)</td><td align="center">0.77</td></tr><tr><td>Age &#x0003e;65&#x000a0;years, n (%)</td><td align="center">44 (21.1)</td><td align="center">17 (13.3)</td><td align="center">25 (31.2)</td><td align="center">0.001</td></tr><tr><td>Sex ratio</td><td align="center">1.39</td><td align="center">1.42</td><td align="center">1.35</td><td align="center">0.87</td></tr><tr><td>Body mass index, mean (SD)</td><td align="center">25.3 (4.7)</td><td align="center">25.5 (4.7)</td><td align="center">25.1 (4.8)</td><td align="center">0.54</td></tr><tr><td>Simplified acute physiology score, h 24, mean (SD)</td><td align="center">52.8 (19.4)</td><td align="center">47.6 (19.8)</td><td align="center">58.9 (17.1)</td><td align="center">0.0001</td></tr><tr><td>Cortisol time 0&#x000a0;minutes, &#x003bc;g/dl, mean (SD)</td><td align="center">17.8 (14.2)</td><td align="center">20.2 (14.1)</td><td align="center">16.5 (14.2)</td><td align="center">0.16</td></tr><tr><td>Cortisol time 60 to 0&#x000a0;minutes, &#x003bc;g/dl, mean (SD)</td><td align="center">16.9 (16.8)</td><td align="center">16.2 (19.6)</td><td align="center">18.7 (19.1)</td><td align="center">0.48</td></tr><tr><td>Adrenal insufficiency, n (%)</td><td align="center">94/121 (77.6)</td><td align="center">30/41 (73)</td><td align="center">64/80 (80)</td><td align="center">0.39</td></tr><tr><td>Adrenocorticotrophic hormone responders, n (%)</td><td align="center">77/121 (63.6)</td><td align="center">26/41 (63.4)</td><td align="center">51/80 (63.7)</td><td align="center">0.97</td></tr><tr><td>Average length of support, h, mean (SD)</td><td align="center">21 (8.5)</td><td align="center">21.5 (8.15)</td><td align="center">19.4 (9.5)</td><td align="center">0.007</td></tr><tr><td> <bold>Brain death etiology</bold> </td><td align="center"/><td align="center"/><td align="center"/><td align="center"/></tr><tr><td>Traumatic brain injury, n (%)</td><td align="center">60 (28.8)</td><td align="center">35 (27.3)</td><td align="center">25 (31.2)</td><td align="center">0.54</td></tr><tr><td>Brain hemorrhage, n (%)</td><td align="center">83 (39.9)</td><td align="center">57 (44.5)</td><td align="center">26 (32.5)</td><td align="center">0.08</td></tr><tr><td>Subarachnoid hemorrhage, n (%)</td><td align="center">67 (32.2)</td><td align="center">48 (37.5)</td><td align="center">19 (23.7)</td><td align="center">0.03</td></tr><tr><td>Cerebral ischemic injury, n (%)</td><td align="center">21 (10.1)</td><td align="center">12 (9.4)</td><td align="center">9 (11.2)</td><td align="center">0.66</td></tr><tr><td>Anoxic encephalopathy, n (%)</td><td align="center">31 (14.9)</td><td align="center">19 (14.8)</td><td align="center">12 (15)</td><td align="center">0.97</td></tr><tr><td>Post neurosurgery, n (%)</td><td align="center">4 (1.9)</td><td align="center">4 (3.1)</td><td align="center">0</td><td align="center">0.11</td></tr></tbody></table></table-wrap><fig id="Fig1"><label>Figure 1</label><caption><p> <bold>Study flow chart.</bold> </p></caption><graphic xlink:href="13054_2014_2596_Fig1_HTML" id="d30e1069"/></fig></p><p>The ACTH stimulation test was performed in the 80 patients in the steroid group and in 41 patients in the control group; it revealed adrenal insufficiency in 94/121 brain-dead patients (78%). In the steroid group, the mean time before administration of hydrocortisone was 168&#x02009;&#x000b1;&#x02009;130&#x000a0;minutes after brain death diagnosis. The mean quantity of hypotension episodes was comparable in control group and steroid group (1.2&#x02009;&#x000b1;&#x02009;1.4 versus 1.0&#x02009;&#x000b1;&#x02009;1.6, <italic>P</italic>&#x02009;=&#x02009;0.18). The mean vascular filling volume per hour was similar in the two groups (179&#x02009;&#x000b1;&#x02009;106&#x000a0;ml/h versus 219&#x02009;&#x000b1;&#x02009;165&#x000a0;ml/h, <italic>P</italic>&#x02009;=&#x02009;0.88). Although there were more patients in the steroid group who received norepinephrine before brain death (80% versus 66%, <italic>P</italic>&#x02009;=&#x02009;0.03), the mean dose of vasopressor administered after brain death was significantly lower than in the control group (1.18&#x02009;&#x000b1;&#x02009;0.92&#x000a0;mg/h versus 1.49&#x02009;&#x000b1;&#x02009;1.29&#x000a0;mg/h, <italic>P</italic>&#x02009;=&#x02009;0.03), duration of vasopressor support use was shorter than in control group (874&#x000a0;minutes versus 1,160&#x000a0;minutes: <italic>P</italic> &#x0003c;0.0001) and norepinephrine weaning before aortic clamping was more frequent (33.8% versus 9.5%, <italic>P</italic> &#x0003c;0.0001) (Table&#x000a0;<xref rid="Tab3" ref-type="table">3</xref>). Using a survival approach, probability of norepinephrine weaning was significantly different between the two groups (<italic>P</italic> &#x0003c;0.0001) with a probability of weaning 4.67 times higher in the steroid group than in the control group (95% CI 2.30, 9.49) (Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>). For the sub-groups of patients responding or non-responding to the ACTH test, no significant differences in norepinephrine weaning were noted (HR 1.84, 95% CI 0.67, 5.05, <italic>P</italic>&#x02009;=&#x02009;0.23 and HR 6.74, 95% CI 0.82, 55.19, <italic>P</italic>&#x02009;=&#x02009;0.07, respectively).<table-wrap id="Tab3"><label>Table 3</label><caption><p> <bold>Hemodynamic results</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr><th/><th>Control group (n&#x02009;=&#x02009;128)</th><th>Steroid group (n&#x02009;=&#x02009;80)</th><th> <bold><italic>P</italic></bold>-value</th></tr></thead><tbody><tr><td>Patients with norepinephrine before inclusion, n (%)</td><td align="center">85 (66.4)</td><td align="center">64 (80)</td><td align="center">0.03</td></tr><tr><td>Patients with norepinephrine during support, n (%)</td><td align="center">103 (80.4)</td><td align="center">76 (95)</td><td align="center">0.12</td></tr><tr><td>Norepinephrine dose (mg/h), mean (SD)</td><td align="center">1.49 (1.29)</td><td align="center">1.18 (0.92)</td><td align="center">0.03</td></tr><tr><td>Patients weaned norepinephrine, n (%)</td><td align="center">10/103 (9.7)</td><td align="center">26/76 (34.2)</td><td align="center">&#x0003c;0.0001</td></tr><tr><td>Varying dose in unweaned patients, % of initial dosing</td><td align="center">+ 46</td><td align="center">-20.9</td><td align="center">0.0004</td></tr><tr><td>Duration with norepinephrine, minutes, median</td><td align="center">1160</td><td align="center">874</td><td align="center">&#x0003c;0.0001</td></tr></tbody></table><table-wrap-foot><p>Varying dose in unweaned patients&#x02009;=&#x02009;(norepinephrin end dosing &#x02013; norepinephrin initial dosing)/norepinephrin initial dosing.</p></table-wrap-foot></table-wrap><fig id="Fig2"><label>Figure 2</label><caption><p> <bold>Kaplan-Meier curves for the probability of continuation of norepinephrine.</bold> </p></caption><graphic xlink:href="13054_2014_2596_Fig2_HTML" id="d30e1230"/></fig></p><p>The number of recovered organs compared to the number of brain-dead patients was similar in the steroid group and control group (3.31&#x02009;&#x000b1;&#x02009;1.36 versus 3.51&#x02009;&#x000b1;&#x02009;1.39, <italic>P</italic>&#x02009;=&#x02009;0.23) (Table&#x000a0;<xref rid="Tab4" ref-type="table">4</xref>). However, if compared to the number of organs considered for procurement, the percentage of recovered grafts in the steroid group was higher than in control group, but it did not reach the significance threshold (92% versus 88%, <italic>P</italic>&#x02009;=&#x02009;0.07) (Table&#x000a0;<xref rid="Tab4" ref-type="table">4</xref>). There was no significant difference in cold ischemia duration between the two groups (Table&#x000a0;<xref rid="Tab5" ref-type="table">5</xref>).<table-wrap id="Tab4"><label>Table 4</label><caption><p> <bold>Organs recovered/organs recoverable (%)</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr><th>Organs</th><th>All patients (n&#x02009;=&#x02009;208)</th><th>Control group (n&#x02009;=&#x02009;128)</th><th>Steroid group (n&#x02009;=&#x02009;80)</th><th> <bold><italic>P</italic></bold>-value</th></tr></thead><tbody><tr><td>Kidney</td><td align="center">394/403 (97.7)</td><td align="center">243/248 (98)</td><td align="center">151/155 (97.4)</td><td align="center">0.65</td></tr><tr><td>Liver</td><td align="center">162/172 (94.2)</td><td align="center">99/105 (94.3)</td><td align="center">63/67 (94)</td><td align="center">0.61</td></tr><tr><td>Heart</td><td align="center">66/80 (82.5)</td><td align="center">47/56 (83.9)</td><td align="center">19/24 (79.1)</td><td align="center">0.74</td></tr><tr><td>Lung</td><td align="center">71/93 (73.9)</td><td align="center">44/62 (70.9)</td><td align="center">27/34 (79.4)</td><td align="center">0.36</td></tr><tr><td>Pancreas</td><td align="center">21/47 (44.6)</td><td align="center">16/39 (41)</td><td align="center">5/8 (62.5)</td><td align="center">0.43</td></tr><tr><td>Total</td><td align="center">714/798 (89.5)</td><td align="center">449/510 (88)</td><td align="center">265/288 (92)</td><td align="center">0.07</td></tr><tr><td>Organs/donors, n (SD)</td><td align="center">3.43 (1.37)</td><td align="center">3.51 (1.39)</td><td align="center">3.31 (1.36)</td><td align="center">0.23</td></tr></tbody></table><table-wrap-foot><p>Results presented as number/total number (%) unless stated otherwise.</p></table-wrap-foot></table-wrap><table-wrap id="Tab5"><label>Table 5</label><caption><p> <bold>Cold ischemia duration (hours) by organ and strategy, mean (SD)</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr><th>Organs</th><th>Control group</th><th>Steroid group</th><th> <bold><italic>P</italic></bold>-value</th></tr></thead><tbody><tr><td>Right kidney</td><td align="center">19.05 (6.84)</td><td align="center">19.07 (6.4)</td><td align="center">0.88</td></tr><tr><td>Left kidney</td><td align="center">16.08 (5.76)</td><td align="center">16.12 (4.23)</td><td align="center">0.51</td></tr><tr><td>Liver</td><td align="center">8.51 (2.8)</td><td align="center">8.76 (2.2)</td><td align="center">0.31</td></tr><tr><td>Lung</td><td align="center">5.85 (1.65)</td><td align="center">5.21 (1.48)</td><td align="center">0.25</td></tr><tr><td>Heart</td><td align="center">3.04 (1.1)</td><td align="center">3.39 (0.97)</td><td align="center">0.24</td></tr><tr><td>Pancreas</td><td align="center">10.94 (1.47)</td><td align="center">11.68 (1.15)</td><td align="center">0.55</td></tr></tbody></table></table-wrap></p><p>Among the 714 recovered grafts, 72 grafts were not transplanted (10.1%). Among the 642 transplanted and studied grafts, we observed a DGF in one case out of three (212/642). DGF of kidney graft was more frequent in the steroid group (39% versus 28%, <italic>P</italic>&#x02009;=&#x02009;0.03) (Table&#x000a0;<xref rid="Tab6" ref-type="table">6</xref>). However this difference did not persist after adjustment for the other variables: logistic regression with random effects for donor showed that the probability of kidney graft DGF increased with age of the donor (<italic>P</italic>&#x02009;=&#x02009;0.0007), and with serum creatinine value before procurement (<italic>P</italic>&#x02009;=&#x02009;0.006) and decreased with norepinephrine dose (<italic>P</italic>&#x02009;=&#x02009;0.03) but was not modified by the strategy (<italic>P</italic>&#x02009;=&#x02009;0.26). When SAPSII was considered in the model instead of age, the only predictors of kidney graft DGF were serum creatinine value (<italic>P</italic>&#x02009;=&#x02009;0.04) and cold ischemia duration (<italic>P</italic>&#x02009;=&#x02009;0.01).<table-wrap id="Tab6"><label>Table 6</label><caption><p> <bold>Delayed grafts function (DGF) by organ and strategy, number (%)</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr><th>Organs</th><th>Control group</th><th>Steroid group</th><th> <bold><italic>P</italic></bold>-value</th><th>Adjusted <bold><italic>P</italic></bold>-value*</th></tr></thead><tbody><tr><td>Kidney</td><td align="center">65/230 (28.3)</td><td align="center">55/141 (39)</td><td align="center">0.03</td><td align="center">0.04</td></tr><tr><td>Liver</td><td align="center">25/94 (26.6)</td><td align="center">22/59 (37.3)</td><td align="center">0.16</td><td align="center">0.11</td></tr><tr><td>Lung</td><td align="center">12/31 (38.7)</td><td align="center">5/25 (20)</td><td align="center">0.13</td><td align="center">0.21</td></tr><tr><td>Heart</td><td align="center">21/44 (47.7)</td><td align="center">7/18 (38.8)</td><td align="center">0.52</td><td align="center">0.45</td></tr><tr><td>Total</td><td align="center">123/399 (30.8)</td><td align="center">89/243 (36.6)</td><td align="center">0.14</td><td align="center"/></tr></tbody></table><table-wrap-foot><p>*Adjusted for cold ischemia duration.</p></table-wrap-foot></table-wrap></p><p>For the other grafts, DGF frequency was comparable in both groups (Table&#x000a0;<xref rid="Tab6" ref-type="table">6</xref>).</p></sec><sec id="Sec9" sec-type="discussion"><title>Discussion</title><p>This multicenter prospective controlled study demonstrates that systemic administration of low-dose steroids during brain-death resuscitation of potential brain-dead organ donors makes vasopressor weaning possible in more than a third of patients (33.8%) and also decreases by more than 20% the quantity of vasopressors needed to control circulatory failure, allowing for a significant reduction in the need for inotropic support. This effect is not related to adrenal insufficiency identified by ACTH stimulation, but steroid administration alone fails to increase the number of organ recovered for transplantation.</p><p>When an organ donor with brain-death is resuscitated, one of the main objectives is to stabilize the hemodynamic state in order to limit ischemia and inflammation as far as possible in the different organs. This goal is usually achieved by a combination of fluid expansion and inotrope administration. Norepinephrine is associated with a decreased rate of high-yield procurement and it seems clinically relevant to reduce doses or use of norepinephrine in brain-dead donors to increase the rate of organ procurement. Thyroid hormones and cortisol deficit have already been identified as circulatory failure factors in brain-dead donors [<xref ref-type="bibr" rid="CR10">10</xref>&#x02013;<xref ref-type="bibr" rid="CR12">12</xref>]. Our study confirms that prescription of a replacement dose of hydrocortisone during resuscitation of a potential brain-dead donor makes vasopressor weaning possible and also decreases vasopressor doses, which are necessary to maintain a stable hemodynamic state in unweaned patients. These results are in agreement with those of a single-center observational cohort of 30 patients with brain death who were administered 50&#x000a0;mg of hydrocortisone: in 58% of patients norepinephrine doses were reduced by 30% after three hours [<xref ref-type="bibr" rid="CR19">19</xref>]. Another study compared two groups of donors including during two consecutive periods and studied the impact of high doses of methylprednisolone (15&#x000a0;mg/kg) versus low doses of hydrocortisone (300 to 500&#x000a0;mg): frequency of vasopressor weaning was 39% in the first group and 47% in the second group, with no significant difference between both groups [<xref ref-type="bibr" rid="CR20">20</xref>]. Our results for hemodynamic stability and decrease in vasopressor use following steroid administration are similar to the results obtained in several studies, in which identical doses of hydrocortisone were administered in patients who were not brain-dead [<xref ref-type="bibr" rid="CR21">21</xref>, <xref ref-type="bibr" rid="CR22">22</xref>].</p><p>Although in our study, adrenal insufficiency frequency (77.6%) was similar to the results of both studies cited above [<xref ref-type="bibr" rid="CR19">19</xref>, <xref ref-type="bibr" rid="CR20">20</xref>], no significant difference was noted in the frequency of vasopressor weaning of brain-dead patients compared to plasma cortisol level at baseline or to initial response to ACTH stimulation. Our results did not confirm those of Nicolas-Robin <italic>et al</italic>. [<xref ref-type="bibr" rid="CR19">19</xref>], who suggested that steroid administration would be more beneficial in patients with documented adrenal insufficiency. However our results are consistent with other studies conducted in different settings, such as in septic shock, showing that the ACTH stimulation test had no predictive value for hemodynamic response of patients receiving corticosteroids [<xref ref-type="bibr" rid="CR23">23</xref>&#x02013;<xref ref-type="bibr" rid="CR26">26</xref>].</p><p>Several reports suggest that multimodal hormonal therapy (thyroid hormone, corticosteroids and arginine vasopressin) might have beneficial effects on recovery of organs, with an increased amount of procured grafts and less primary dysfunction in transplanted heart grafts [<xref ref-type="bibr" rid="CR6">6</xref>, <xref ref-type="bibr" rid="CR16">16</xref>]; however, the respective contribution of each hormonal therapy remains controversial. Thyroid hormone administration by itself has been considered either to be beneficial, neutral or to have no significant impact on organ procurement. In our study, no significant difference was noted in the number of recovered organs per donor in either group. The number of recovered organs compared to the number of organs considered for procurement was slightly higher in the steroid group than in the control group, but the difference was not significant. DGF frequency was similar in both groups regardless of the grafted organ. Kidney grafts were the only exception with primary dysfunction significantly more frequent in the steroid group. Despite the benefits of a replacement dose of hydrocortisone on vasopressor consumption, our study did not demonstrate any benefits of steroid administration for primary function recovery of transplanted grafts. Some studies have either shown the absence of clinical impact of a replacement dose of hydrocortisone on primary function recovery of kidney [<xref ref-type="bibr" rid="CR27">27</xref>], liver [<xref ref-type="bibr" rid="CR28">28</xref>], heart and lung grafts; for those grafts, only a decrease in systemic inflammatory markers and their expression in tissue, and a decrease of extravascular lung water accumulation were beneficial for primary graft recovery [<xref ref-type="bibr" rid="CR29">29</xref>, <xref ref-type="bibr" rid="CR30">30</xref>].</p><p>Our study has several limitations: lack of randomization of brain-dead patients, cluster distribution of recruited patients, lack of data about receivers&#x02019; history, and comorbidities of receivers and their consequences for DGF. This might have limited our capacity to identify some potential beneficial effects of steroid administration during brain-dead donor resuscitation and we expect that practitioners would not change their usual practice in brain-dead donor resuscitation.</p></sec><sec id="Sec10" sec-type="conclusions"><title>Conclusion</title><p>Early substitutive administration of glucocorticoids in a potential brain-dead organ donor with circulatory failure makes it possible to significantly reduce the cumulative dose and administration duration of vasopressors. Whatever the case, based on our results we cannot reach a conclusion as to whether or not the routine use of steroids administration in potential brain-dead organ donors should be supported, independently of documented relative adrenal insufficiency, and no benefits for primary function recovery of transplanted grafts were observed in the study. Existing controversy in the literature suggests that a multiple strategy is required to achieve measurable effects in the standard care of organ donors. Routine steroid administration is probably an important component of that strategy to improve recovery of organs following brain death, but should not be used alone and probably should be considered along with other hormonal factors, for which the respective contribution remains to be defined.</p></sec><sec id="Sec11"><title>Key messages</title><p><list list-type="bullet"><list-item><p>&#x02009;Steroids and norepinephrine are equally effective in achieving hemodynamic stability in various different groups of brain-dead organ donors.</p></list-item><list-item><p>&#x02009;Steroid administration alone fails to increase the number of organs recovered for transplantation.</p></list-item><list-item><p>&#x02009;Steroid administration did not demonstrate any benefits for primary function recovery of transplanted grafts.</p></list-item><list-item><p>&#x02009;The decision to use steroids or norepinephrine was not observed to affect meaningful outcomes for hemodynamically stable brain-dead organ donors.</p></list-item></list></p></sec>

Modeling the economic impact of linezolid versus vancomycin in confirmed nosocomial pneumonia caused by methicillin-resistant <italic>Staphylococcus aureus</italic>

<sec><title>Introduction</title><p>We compared the economic impacts of linezolid and vancomycin for the treatment of hospitalized patients with methicillin-resistant <italic>Staphylococcus aureus</italic> (MRSA)&#x02013;confirmed nosocomial pneumonia.</p></sec><sec><title>Methods</title><p>We used a 4-week decision tree model incorporating published data and expert opinion on clinical parameters, resource use and costs (in 2012 US dollars), such as efficacy, mortality, serious adverse events, treatment duration and length of hospital stay. The results presented are from a US payer perspective. The base case first-line treatment duration for patients with MRSA-confirmed nosocomial pneumonia was 10&#x000a0;days. Clinical treatment success (used for the cost-effectiveness ratio) and failure due to lack of efficacy, serious adverse events or mortality were possible clinical outcomes that could impact costs. Cost of treatment and incremental cost-effectiveness per successfully treated patient were calculated for linezolid versus vancomycin. Univariate (one-way) and probabilistic sensitivity analyses were conducted.</p></sec><sec><title>Results</title><p>The model allowed us to calculate the total base case inpatient costs as $46,168 (linezolid) and $46,992 (vancomycin). The incremental cost-effectiveness ratio favored linezolid (versus vancomycin), with lower costs ($824 less) and greater efficacy (+2.7% absolute difference in the proportion of patients successfully treated for MRSA nosocomial pneumonia). Approximately 80% of the total treatment costs were attributed to hospital stay (primarily in the intensive care unit). The results of our probabilistic sensitivity analysis indicated that linezolid is the cost-effective alternative under varying willingness to pay thresholds.</p></sec><sec><title>Conclusion</title><p>These model results show that linezolid has a favorable incremental cost-effectiveness ratio compared to vancomycin for MRSA-confirmed nosocomial pneumonia, largely attributable to the higher clinical trial response rate of patients treated with linezolid. The higher drug acquisition cost of linezolid was offset by lower treatment failure&#x02013;related costs and fewer days of hospitalization.</p></sec>

<contrib contrib-type="author" id="A1"><name><surname>Patel</surname><given-names>Dipen A</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>dpatel@pharmerit.com</email></contrib><contrib contrib-type="author" id="A2"><name><surname>Shorr</surname><given-names>Andrew F</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>andrew.shorr@gmail.com</email></contrib><contrib contrib-type="author" id="A3"><name><surname>Chastre</surname><given-names>Jean</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>jean.chastre@psl.aphp.fr</email></contrib><contrib contrib-type="author" id="A4"><name><surname>Niederman</surname><given-names>Michael</given-names></name><xref ref-type="aff" rid="I4">4</xref><email>mniederman@winthrop.org</email></contrib><contrib contrib-type="author" id="A5"><name><surname>Simor</surname><given-names>Andrew</given-names></name><xref ref-type="aff" rid="I5">5</xref><email>andrew.simor@sunnybrook.ca</email></contrib><contrib contrib-type="author" corresp="yes" id="A6"><name><surname>Stephens</surname><given-names>Jennifer M</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>jstephens@pharmerit.com</email></contrib><contrib contrib-type="author" id="A7"><name><surname>Charbonneau</surname><given-names>Claudie</given-names></name><xref ref-type="aff" rid="I6">6</xref><email>claudie.charbonneau@pfizer.com</email></contrib><contrib contrib-type="author" id="A8"><name><surname>Gao</surname><given-names>Xin</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>cgao@pharmerit.com</email></contrib><contrib contrib-type="author" id="A9"><name><surname>Nathwani</surname><given-names>Dilip</given-names></name><xref ref-type="aff" rid="I7">7</xref><email>dilip.nathwani@nhs.net</email></contrib>

Critical Care

<sec sec-type="intro"><title>Introduction</title><p>Nosocomial pneumonia (NP) has been reported to be the second most frequent hospital-acquired infection in the United States [<xref ref-type="bibr" rid="B1">1</xref>]. Methicillin-resistant <italic>Staphylococcus aureus</italic> (MRSA) is responsible for a large number of cases of health-care&#x02013;associated pneumonia, hospital-acquired pneumonia and ventilator-associated pneumonia [<xref ref-type="bibr" rid="B2">2</xref>,<xref ref-type="bibr" rid="B3">3</xref>]. A longitudinal study showed that the proportion of <italic>Staphylococcus aureus</italic> isolates that were methicillin-resistant (that is, MRSA) increased from 35.9% in 1992 to 64.4% in 2003 in ICUs in the United States [<xref ref-type="bibr" rid="B4">4</xref>]; however, more recent data from nine metropolitan areas suggest that the incidence rates have declined among patients with health-care&#x02013;associated, community-onset or hospital-onset infections [<xref ref-type="bibr" rid="B5">5</xref>].</p><p>Despite the variation in incidence, MRSA infections remain a significant public health problem. MRSA-associated NP results in considerable patient morbidity, mortality and use of health-care resources with significant length of hospital stay [<xref ref-type="bibr" rid="B6">6</xref>-<xref ref-type="bibr" rid="B8">8</xref>]. The mean duration of hospitalization and associated costs of MRSA infections have been reported to be significantly higher than those of methicillin-susceptible <italic>S. aureus</italic> (MSSA) infections [<xref ref-type="bibr" rid="B9">9</xref>,<xref ref-type="bibr" rid="B10">10</xref>]. The majority of this cost difference can be attributed to excess hospitalization rather than to charges for antibiotic use, radiologic procedures or laboratory services.</p><p>Vancomycin and linezolid are the commonly recommended agents in clinical guidelines for the treatment of MRSA-related pneumonia [<xref ref-type="bibr" rid="B11">11</xref>,<xref ref-type="bibr" rid="B12">12</xref>]. In addition to these two agents, telavancin is the only other agent approved for the treatment of MRSA NP in the United States and Europe. Two large, prospective, randomized, double-blind trials demonstrated that linezolid (600&#x000a0;mg every 12&#x000a0;hours) was statistically noninferior to fixed-dose vancomycin (1&#x000a0;g twice daily) for the treatment of NP [<xref ref-type="bibr" rid="B13">13</xref>,<xref ref-type="bibr" rid="B14">14</xref>]. In a retrospective combined subgroup analysis of these two trials, researchers found significantly higher survival and clinical cure rates with linezolid treatment compared with vancomycin treatment [<xref ref-type="bibr" rid="B14">14</xref>]. Using <italic>post hoc</italic> data from the same studies, investigators have reported similar findings in patients with MRSA ventilator-associated pneumonia [<xref ref-type="bibr" rid="B15">15</xref>].</p><p>In a recent prospective, randomized, double-blind, controlled, multicenter study, specifically focused on MRSA-confirmed NP, researchers found greater clinical efficacy (defined as resolution of signs and symptoms, improved or lack of progression in chest imaging and no additional antibacterial treatment required) with linezolid than with adjusted-dose vancomycin [<xref ref-type="bibr" rid="B16">16</xref>]. That study&#x02019;s sample size for the modified intent-to-treat population (MRSA-confirmed population) was 224 patients in each arm, with an end-of-study success rate of 57.6% for linezolid-treated patients and 46.6% for vancomycin-treated patients (95% confidence interval (CI) for differences from 0.5 to 21.6; <italic>P</italic>&#x02009;=&#x02009;0.042). Linezolid was noninferior and statistically superior to vancomycin in end-of-treatment clinical and in end-of-treatment and end-of-study microbiologic outcomes. All-cause 60-day mortality rates were similar (15.7% for linezolid and 17.0% for vancomycin), as were the serious adverse event (SAE) rates.</p><p>Despite its higher acquisition costs, the overall cost for treating MRSA NP with linezolid may be lower because it is associated with better clinical outcomes compared with vancomycin. Yet, few researchers have investigated the costs associated with MRSA NP [<xref ref-type="bibr" rid="B17">17</xref>] and, in particular, the economic outcomes associated with treatments for MRSA NP. In two cost-effectiveness analyses based on a retrospective decision-analytic modeling approach, investigators found linezolid to be less costly and more efficacious than vancomycin for patients with suspected MRSA NP [<xref ref-type="bibr" rid="B18">18</xref>,<xref ref-type="bibr" rid="B19">19</xref>]. However, these earlier modeling studies either did not address the use of these agents in the US context [<xref ref-type="bibr" rid="B18">18</xref>] or were focused only on first-line therapy [<xref ref-type="bibr" rid="B19">19</xref>].</p><p>Our purpose in this economic analysis was to investigate the economic impact of improved clinical outcomes with linezolid compared with vancomycin in the treatment of hospitalized patients with MRSA-confirmed NP in the United States using a decision tree with a payer perspective and flexibility for real-world clinical conditions.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>Model design</title><p>We conducted a cost-effectiveness analysis of intravenous (IV) linezolid compared with IV vancomycin for the treatment of MRSA NP in hospitalized adults, which was based on a decision tree modeling approach. The decision tree model was developed to capture first-line and second-line therapy. Because of the short-term window for the clinical management of an NP episode, the model time horizon was up to 4&#x000a0;weeks, which was validated by practicing physicians. This time horizon spans periods typical for ICU and general ward stays during first-line and second-line treatment [<xref ref-type="bibr" rid="B18">18</xref>,<xref ref-type="bibr" rid="B20">20</xref>,<xref ref-type="bibr" rid="B21">21</xref>]. A total payer perspective (assuming a <italic>per diem</italic> basis of payment) was considered in the base case analysis, which was comprehensive and comprised all inpatient and outpatient health-care costs (antibiotic and medical). Because this was an economic model in which we used only previously published data to create a hypothetical patient pathway, ethical approval and informed consent were neither applicable nor required.</p><p>The hypothetical model population was assumed to be similar to the population included in a recent phase IV, prospective, double-blind, controlled, multicenter, international clinical trial of IV linezolid (600&#x000a0;mg every 12&#x000a0;hours) or IV vancomycin (15&#x000a0;mg/kg every 12&#x000a0;hours, dose-adjusted based on trough levels and renal function) for the treatment of MRSA NP [<xref ref-type="bibr" rid="B16">16</xref>]. The full details of the characteristics and resource use for this MRSA-confirmed trial population have been reported previously [<xref ref-type="bibr" rid="B21">21</xref>,<xref ref-type="bibr" rid="B22">22</xref>] and include the following data: mean age 62&#x000a0;years, 69% white, 66% male, 75% mechanically ventilated, 87% had at least 1&#x000a0;day in the ICU and 63% were from the United States. The population used for analysis in the model was hospitalized adult patients with a confirmed MRSA NP diagnosis.</p><p>Patients with suspected and/or confirmed Gram-positive NP could initially be treated with empiric IV antibiotic therapy (for example, vancomycin or linezolid in combination with ceftazidime, imipenem or piperacillin/tazobactam) for up to 3&#x000a0;days while laboratory confirmation of NP pathogen occurred (Figure&#x000a0;<xref ref-type="fig" rid="F1">1</xref>). This empiric treatment pathway was not included in the base case analysis. Following confirmation of MRSA NP, the economic model analysis began and patients were placed on first-line treatment (vancomycin or linezolid) for 10&#x000a0;days (Figure&#x000a0;<xref ref-type="fig" rid="F1">1</xref>). We focused on the component of treatment after MRSA confirmation when calculating cost-effectiveness, given the recent clinical trial data available [<xref ref-type="bibr" rid="B16">16</xref>] and because this is an important time point in clinical decision-making for reevaluation of the antibiotic treatment and coverage.Possible treatment outcomes associated with first-line therapy were (1) treatment success (defined as resolution of signs and symptoms of NP, improvement or lack of progression in chest imaging and no additional antibacterial treatment required among survivors), (2) failure due to lack of efficacy among survivors, (3) drug discontinuation due to SAEs and (4) failure due to death (Figure&#x000a0;<xref ref-type="fig" rid="F1">1</xref>). A penalty, described in the section below, was assigned for patients whose treatment failed due to lack of efficacy or was discontinued due to SAEs.</p><fig id="F1" position="float"><label>Figure 1</label><caption><p><bold>Decision model tree.</bold> NP, Nosocomial pneumonia; MRSA, Methicillin-resistant <italic>Staphylococcus aureus</italic>; SAEs, Serious adverse events.</p></caption><graphic xlink:href="cc13996-1"/></fig><p>Patients whose first-line treatment succeeded would finish their 10-day treatment duration and exit the model. In cases of any failure of first-line treatment, patients were switched to second-line treatment (for example, patients whose first-line treatment with linezolid failed were switched to second-line vancomycin, and vice versa) after 7&#x000a0;days, with the second-line treatment lasting 10&#x000a0;days. The model did not include a third-line treatment, given the lack of published data.</p></sec><sec><title>Model inputs, outcomes and assumptions</title><p>In the base case scenario, the model was based primarily on recent MRSA NP clinical trial data [<xref ref-type="bibr" rid="B16">16</xref>] (Table&#x000a0;<xref ref-type="table" rid="T1">1</xref>). Linezolid and vancomycin were the main treatment comparators. In the base case analysis, we used 10-day treatment duration for the first- and second-line therapies. Data on length of hospital stay, inpatient and outpatient resource use and associated costs, and drug costs were obtained by analysis of the recent clinical trial and published literature (Tables&#x000a0;<xref ref-type="table" rid="T1">1</xref> and <xref ref-type="table" rid="T2">2</xref>) [<xref ref-type="bibr" rid="B16">16</xref>,<xref ref-type="bibr" rid="B21">21</xref>,<xref ref-type="bibr" rid="B23">23</xref>]. Key resources included in the model were days of antibiotic treatment, hospital general ward stay, hospital ICU stay, mechanical ventilator use, days on IV therapy, inpatient visits (to physician, attending and/or intensivist), inpatient laboratory work and physician office visits.</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p><bold>Model inputs on clinical and resource use data</bold>^{<bold>a</bold>}</p></caption><table frame="hsides" rules="groups" border="1"><colgroup><col align="left"/><col align="left"/><col align="left"/><col align="left"/><col align="left"/></colgroup><thead valign="top"><tr><th align="left">&#x000a0;</th><th align="left"><bold>Linezolid base case value (range</bold>^{<bold>b</bold>}<bold>)</bold></th><th align="left"><bold>Vancomycin base case value (range</bold>^{<bold>b</bold>}<bold>)</bold></th><th align="left"><bold>Distribution</bold>^{<bold>b</bold>}</th><th align="left"><bold>Source</bold></th></tr></thead><tbody valign="top"><tr><td colspan="5" align="left" valign="bottom">Efficacy and safety end points^c,%<hr/></td></tr><tr><td align="left" valign="bottom">&#x02003;&#x000a0;Efficacy (in survivors)<hr/></td><td align="left" valign="bottom">54.8 (49.8^d to 66.7)<hr/></td><td align="left" valign="bottom">44.9 (35.5 to 52.9)<hr/></td><td align="left" valign="bottom">&#x003b2;<hr/></td><td align="left" valign="bottom">[<xref ref-type="bibr" rid="B16">16</xref>,<xref ref-type="bibr" rid="B20">20</xref>]<hr/></td></tr><tr><td align="left" valign="bottom">&#x02003;&#x000a0;Mortality^e<hr/></td><td align="left" valign="bottom">27.2<hr/></td><td align="left" valign="bottom">27.2<hr/></td><td align="left" valign="bottom">&#x02013;<hr/></td><td align="left" valign="bottom">&#x000a0;<hr/></td></tr><tr><td align="left" valign="bottom">&#x02003;&#x000a0;SAEs leading to discontinuation^f<hr/></td><td align="left" valign="bottom">1.8 (0^d to 5.2)<hr/></td><td align="left" valign="bottom">3.1 (0^d to 6.5)<hr/></td><td align="left" valign="bottom">&#x003b2;<hr/></td><td align="left" valign="bottom">[<xref ref-type="bibr" rid="B14">14</xref>,<xref ref-type="bibr" rid="B16">16</xref>]<hr/></td></tr><tr><td align="left" valign="bottom">&#x02003;&#x000a0;Failure leading to discontinuation^g<hr/></td><td align="left" valign="bottom">16.2<hr/></td><td align="left" valign="bottom">24.8<hr/></td><td align="left" valign="bottom">&#x02013;<hr/></td><td align="left" valign="bottom">&#x000a0;<hr/></td></tr><tr><td colspan="5" align="left" valign="bottom">Resource use<hr/></td></tr><tr><td align="left" valign="bottom">&#x02003;&#x000a0;Total days in hospital^h<hr/></td><td align="left" valign="bottom">17.9 (13.9^d to 18.8)<hr/></td><td align="left" valign="bottom">18.6 (14.6^d to 20.1)<hr/></td><td align="left" valign="bottom">&#x003b3;<hr/></td><td align="left" valign="bottom">[<xref ref-type="bibr" rid="B20">20</xref>,<xref ref-type="bibr" rid="B21">21</xref>]<hr/></td></tr><tr><td align="left" valign="bottom">&#x02003;&#x000a0;Days of treatmentⁱ<hr/></td><td align="left" valign="bottom">10.0 (7 to 14)<hr/></td><td align="left" valign="bottom">10.0 (7 to 14)<hr/></td><td align="left" valign="bottom">Uniform<hr/></td><td align="left" valign="bottom">[<xref ref-type="bibr" rid="B16">16</xref>]<hr/></td></tr><tr><td align="left" valign="bottom">&#x02003;&#x000a0;Days in ICU^h<hr/></td><td align="left" valign="bottom">10.1 (6.1^d to 12.2)<hr/></td><td align="left" valign="bottom">10.6 (6.6^d to 16.2)<hr/></td><td align="left" valign="bottom">&#x003b3;<hr/></td><td align="left" valign="bottom">[<xref ref-type="bibr" rid="B20">20</xref>,<xref ref-type="bibr" rid="B21">21</xref>]<hr/></td></tr><tr><td align="left" valign="bottom">Days on mechanical ventilation^h<hr/></td><td align="left" valign="bottom">8.3 (4.3^d to 10.4)<hr/></td><td align="left" valign="bottom">8.1 (4.1^d to 14.3)<hr/></td><td align="left" valign="bottom">&#x003b3;<hr/></td><td align="left" valign="bottom">[<xref ref-type="bibr" rid="B20">20</xref>,<xref ref-type="bibr" rid="B21">21</xref>]<hr/></td></tr><tr><td align="left" valign="bottom">&#x02003;&#x000a0;Additional days in hospital due to SAE<hr/></td><td align="left" valign="bottom">1.7 (0 to 5)^d<hr/></td><td align="left" valign="bottom">1.7 (0 to 5)^d<hr/></td><td align="left" valign="bottom">&#x003b3;<hr/></td><td align="left" valign="bottom">[<xref ref-type="bibr" rid="B21">21</xref>]<hr/></td></tr><tr><td align="left" valign="bottom">&#x02003;&#x000a0;Additional days in hospital due to treatment failure<hr/></td><td align="left" valign="bottom">2 (0 to 5)^d<hr/></td><td align="left" valign="bottom">2 (0 to 5)^d<hr/></td><td align="left" valign="bottom">Uniform<hr/></td><td align="left" valign="bottom">Expert input<hr/></td></tr><tr><td align="left" valign="bottom">&#x02003;&#x000a0;Number of days until switch to second-line after treatment failure/SAE with first- line<hr/></td><td align="left" valign="bottom">7 (5 to 10)^d<hr/></td><td align="left" valign="bottom">7 (5 to 10)^d<hr/></td><td align="left" valign="bottom">Uniform<hr/></td><td align="left" valign="bottom">Expert input<hr/></td></tr><tr><td align="left" valign="bottom">&#x02003;&#x000a0;Days receiving IV antibiotic^h<hr/></td><td align="left" valign="bottom">10.0<hr/></td><td align="left" valign="bottom">10.0<hr/></td><td align="left" valign="bottom">&#x02013;<hr/></td><td align="left" valign="bottom">[<xref ref-type="bibr" rid="B16">16</xref>]<hr/></td></tr><tr><td align="left" valign="bottom">&#x02003;&#x000a0;Antibiotic IV doses/day^h<hr/></td><td align="left" valign="bottom">2.0<hr/></td><td align="left" valign="bottom">2.0<hr/></td><td align="left" valign="bottom">&#x02013;<hr/></td><td align="left" valign="bottom">Product label<hr/></td></tr><tr><td align="left" valign="bottom">&#x02003;&#x000a0;Physician/attending/intensivist visit (inpatient)/dayⁱ<hr/></td><td align="left" valign="bottom">1.0<hr/></td><td align="left" valign="bottom">1.0<hr/></td><td align="left" valign="bottom">&#x02013;<hr/></td><td align="left" valign="bottom">Expert input<hr/></td></tr><tr><td align="left">&#x02003;&#x000a0;Lab work/wk^i,j</td><td align="left">7.0</td><td align="left">8.0</td><td align="left">&#x02013;</td><td align="left">Expert input</td></tr></tbody></table><table-wrap-foot><p>^aIV, Intravenous; SAE, Serious adverse event. Clinical response rate for modified intent-to-treat population at end-of-study time point was used [<xref ref-type="bibr" rid="B16">16</xref>,<xref ref-type="bibr" rid="B21">21</xref>,<xref ref-type="bibr" rid="B23">23</xref>]. ^bRanges and distributions are provided for variables that were used in sensitivity analyses. ^cSame clinical data were used for second-line treatment. ^dThis was an arbitrary assumption that was validated with expert opinion. ^eWeighted average, because model assumes equal mortality due to lack of significant mortality difference (linezolid&#x02009;=&#x02009;63 of 224; vancomycin&#x02009;=&#x02009;59 of 224). ^fLinezolid&#x02009;=&#x02009;4 of 224; vancomycin&#x02009;=&#x02009;7 of 224. ^gBecause this is a decision tree model, this probability was derived as [1 &#x02013; (probability of efficacy&#x02009;+&#x02009;probability mortality&#x02009;+&#x02009;probability of SAEs leading to discontinuation)]. ^hData input for first line treatment only. ⁱData input for first and second line treatment. ^jDaily serum creatinine levels and complete blood count for both antibiotics and once-weekly serum vancomycin levels for vancomycin.</p></table-wrap-foot></table-wrap><table-wrap position="float" id="T2"><label>Table 2</label><caption><p><bold>Model input data on unit costs of medical care (in 2012 US dollars)</bold>^{<bold>a</bold>}</p></caption><table frame="hsides" rules="groups" border="1"><colgroup><col align="left"/><col align="left"/><col align="left"/></colgroup><thead valign="top"><tr><th align="left"><bold>Cost inputs</bold></th><th align="left"><bold>Cost base case value (range</bold>^{<bold>b</bold>}<bold>)</bold></th><th align="left"><bold>Source</bold></th></tr></thead><tbody valign="top"><tr><td align="left" valign="bottom">Inpatient cost per day (general ward)<hr/></td><td align="left" valign="bottom">$1,973.7 ($1,480.3 to $2,467.1)^c<hr/></td><td align="left" valign="bottom">[<xref ref-type="bibr" rid="B24">24</xref>]<hr/></td></tr><tr><td align="left" valign="bottom">Inpatient cost per day (ICU)<hr/></td><td align="left" valign="bottom">$3,415.6 ($2,561.7 to $4,269.5)^c<hr/></td><td align="left" valign="bottom">[<xref ref-type="bibr" rid="B24">24</xref>]<hr/></td></tr><tr><td align="left" valign="bottom">Mechanical ventilation per day<hr/></td><td align="left" valign="bottom">$225.2 ($168.9 to $281.5)^c<hr/></td><td align="left" valign="bottom">[<xref ref-type="bibr" rid="B25">25</xref>]<hr/></td></tr><tr><td align="left" valign="bottom">Physician/attending/intensivist visit<hr/></td><td align="left" valign="bottom">$175.0<hr/></td><td align="left" valign="bottom">CPT 99233 [<xref ref-type="bibr" rid="B26">26</xref>]<hr/></td></tr><tr><td align="left" valign="bottom">Specialist inpatient visit<hr/></td><td align="left" valign="bottom">$251.2<hr/></td><td align="left" valign="bottom">CPT 99253 [<xref ref-type="bibr" rid="B26">26</xref>]<hr/></td></tr><tr><td align="left" valign="bottom">Laboratory test (serum creatinine levels)<hr/></td><td align="left" valign="bottom">$65.9<hr/></td><td align="left" valign="bottom">CPT 80069^d[<xref ref-type="bibr" rid="B26">26</xref>]<hr/></td></tr><tr><td align="left" valign="bottom">Laboratory test (serum vancomycin levels)<hr/></td><td align="left" valign="bottom">$36.3<hr/></td><td align="left" valign="bottom">CPT 80202^e[<xref ref-type="bibr" rid="B26">26</xref>]<hr/></td></tr><tr><td align="left" valign="bottom">Laboratory test (complete blood count)<hr/></td><td align="left" valign="bottom">$34.3<hr/></td><td align="left" valign="bottom">CPT 85025^d[<xref ref-type="bibr" rid="B26">26</xref>]<hr/></td></tr><tr><td align="left" valign="bottom">Outpatient parenteral antibiotic therapy/day<hr/></td><td align="left" valign="bottom">$204.2<hr/></td><td align="left" valign="bottom">[<xref ref-type="bibr" rid="B27">27</xref>]<hr/></td></tr><tr><td align="left" valign="bottom">Injection costs for administration<hr/></td><td align="left" valign="bottom">$7.6<hr/></td><td align="left" valign="bottom">[<xref ref-type="bibr" rid="B28">28</xref>]<hr/></td></tr><tr><td align="left" valign="bottom">IV linezolid 600&#x000a0;mg<hr/></td><td align="left" valign="bottom">$114.6 ($86.0 to $143.3)^c<hr/></td><td align="left" valign="bottom">[<xref ref-type="bibr" rid="B29">29</xref>]<hr/></td></tr><tr><td align="left">IV vancomycin 1&#x000a0;g</td><td align="left">$5.8 ($4.4 to $7.3)^c</td><td align="left">[<xref ref-type="bibr" rid="B29">29</xref>]</td></tr></tbody></table><table-wrap-foot><p>^aCPT, Current procedural terminology; ID, Infectious disease; IV, Intravenous. All costs were adjusted to US dollars using medical care component of the US Consumer Price Index [<xref ref-type="bibr" rid="B24">24</xref>-<xref ref-type="bibr" rid="B29">29</xref>]. ^bRanges are provided for variables that were used in sensitivity analyses. &#x003b3;-distribution was used for these variables for probabilistic sensitivity analysis. ^cArbitrary &#x000b1;25% range was used. ^dBased on testing once every day (while in hospital) for both linezolid and vancomycin. ^eBased on testing once a week for only vancomycin.</p></table-wrap-foot></table-wrap><p>In cases where a discontinuation due to an SAE or treatment failure occurred, patients were assumed to stay 1.7 or 2.0 additional days in the hospital, respectively, during first-line treatment compared with patients whose treatment was successful [<xref ref-type="bibr" rid="B18">18</xref>]. This additional length of stay was determined on the basis of <italic>post hoc</italic> analysis of recent clinical trial data [<xref ref-type="bibr" rid="B16">16</xref>,<xref ref-type="bibr" rid="B21">21</xref>,<xref ref-type="bibr" rid="B22">22</xref>] wherein bivariate analysis was conducted to compare length of stay in patients with or without moderate or severe adverse events and in patients with first-line treatment success versus failure. These values were further validated with expert opinion of the authors who reported the pertinent studies.</p><p>This study is primarily a cost-effectiveness analysis and not a cost&#x02013;utility analysis, because the treatment effect of interest is drug efficacy (that is, proportion of patients successfully treated), instead of quality-adjusted life years (QALYs) or life-years (LYs). The latter two outcomes (QALYs and Lys) were not considered ideal for this analysis and hence are not reported, because the model uses a short-term duration and the trial data used for this model suggest equal mortality rates between linezolid and vancomycin [<xref ref-type="bibr" rid="B16">16</xref>]. As a result, there were negligible differences in QALYs and no differences in LYs between the treatment arms.</p><p>The key result outcomes of this analysis, which are reported in the Results section, are total costs and effectiveness proportion for the two treatments, total cost per successfully treated patient for each treatment (calculated as ratio of total costs and total effectiveness) and incremental cost-effectiveness ratio (ICER), calculated as the difference in costs between treatments divided by the difference in the proportion of successfully treated patients receiving linezolid versus vancomycin.</p><p>The following key assumptions were made in the model:</p><p>&#x025cf;&#x02003;Every patient received treatment as long as they were hospitalized, and all patients were on IV therapy during their hospital stay.</p><p>&#x025cf;&#x02003;In the absence of published data for second-line treatment, the clinical inputs for second-line treatment were the same as those used for first-line treatment [<xref ref-type="bibr" rid="B18">18</xref>].</p><p>&#x025cf;&#x02003;Because we used the 60-day mortality rates reported in the clinical trial [<xref ref-type="bibr" rid="B16">16</xref>], which represented total mortality and included deaths associated with first-line and second-line treatment, mortality occurred only at the end of first-line treatment to avoid overestimation attributable to double-counting. Because the first-line mortality rates did not statistically differ between linezolid and vancomycin in the clinical trial, these rates were considered the same in the model.</p><p>&#x025cf;&#x02003;There were no patient dropouts due to failure or SAEs after first-line treatment.</p><p>&#x025cf;&#x02003;Patients whose second-line treatment failed and those who had SAEs were deemed to have completed the duration of therapy because no third-line therapy was available.</p><p>&#x025cf;&#x02003;Although the mean ICU stay was 10&#x000a0;days, ICU stay was considered to be 7&#x000a0;days if treatment duration was 7&#x000a0;days and the patient&#x02019;s first-line treatment succeeded. Alternatively, if the treatment duration was 14&#x000a0;days, then the ICU stay would be 10&#x000a0;days and the remaining 4&#x000a0;days would be in the general ward.</p></sec><sec><title>Sensitivity analyses</title><p>Univariate (one-way) sensitivity analysis was conducted to assess the impact of model uncertainties and the robustness of our analysis. Key model parameters were varied individually within the predefined sensitivity ranges (Tables&#x000a0;<xref ref-type="table" rid="T2">2</xref> and <xref ref-type="table" rid="T3">3</xref>), and ICERs were recorded. A published source was used for ranges whenever possible. In the absence of strong published data, an arbitrary range was used (such as &#x000b1;4&#x000a0;days for length of stay or &#x000b1;25% for costs). The results are presented in the form of a tornado diagram, with the variables stacked in decreasing order of impact on the ICER.</p><table-wrap position="float" id="T3"><label>Table 3</label><caption><p><bold>Detailed cost results of the base case scenario</bold>^{<bold>a</bold>}</p></caption><table frame="hsides" rules="groups" border="1"><colgroup><col align="left"/><col align="left"/><col align="left"/></colgroup><thead valign="top"><tr><th align="left"><bold>Cost items per patient</bold></th><th align="left"><bold>Linezolid</bold></th><th align="left"><bold>Vancomycin</bold></th></tr></thead><tbody valign="top"><tr><td align="left" valign="bottom">Drug treatment<hr/></td><td align="left" valign="bottom">$2,189<hr/></td><td align="left" valign="bottom">$746<hr/></td></tr><tr><td align="left" valign="bottom">Drug administration<hr/></td><td align="left" valign="bottom">$172<hr/></td><td align="left" valign="bottom">$182<hr/></td></tr><tr><td align="left" valign="bottom"><italic>Inpatient drug cost</italic><hr/></td><td align="left" valign="bottom"><italic>$</italic><italic>2,361</italic><hr/></td><td align="left" valign="bottom"><italic>$</italic><italic>928</italic><hr/></td></tr><tr><td align="left" valign="bottom">ICU<hr/></td><td align="left" valign="bottom">$34,217<hr/></td><td align="left" valign="bottom">$34,728<hr/></td></tr><tr><td align="left" valign="bottom">General ward<hr/></td><td align="left" valign="bottom">$2,451<hr/></td><td align="left" valign="bottom">$3,524<hr/></td></tr><tr><td align="left" valign="bottom">Mechanical ventilation<hr/></td><td align="left" valign="bottom">$1,869<hr/></td><td align="left" valign="bottom">$1,824<hr/></td></tr><tr><td align="left" valign="bottom">Physician/attending visit<hr/></td><td align="left" valign="bottom">$1,970<hr/></td><td align="left" valign="bottom">$2,091<hr/></td></tr><tr><td align="left" valign="bottom">Lab work<hr/></td><td align="left" valign="bottom">$1,137<hr/></td><td align="left" valign="bottom">$1,245<hr/></td></tr><tr><td align="left" valign="bottom">SAE/failure costs<hr/></td><td align="left" valign="bottom">$2,162<hr/></td><td align="left" valign="bottom">$2,651<hr/></td></tr><tr><td align="left" valign="bottom"><italic>Inpatient medical cost</italic><hr/></td><td align="left" valign="bottom"><italic>$</italic><italic>43,807</italic><hr/></td><td align="left" valign="bottom"><italic>$</italic><italic>46,064</italic><hr/></td></tr><tr><td align="left"><bold>Total costs</bold></td><td align="left"><bold>$</bold><bold>46,168</bold></td><td align="left"><bold>$</bold><bold>46,992</bold></td></tr></tbody></table><table-wrap-foot><p>^aSAE, Serious adverse event.</p></table-wrap-foot></table-wrap><p>A probabilistic sensitivity analysis (PSA) was also performed, wherein all parameters were varied simultaneously within their range using 10,000&#x000a0;second-order Monte Carlo simulations. &#x003b3;-distribution was specified for resource use and cost variables, and &#x003b2;-distribution for probability variables.</p></sec></sec><sec sec-type="results"><title>Results</title><sec><title>Base case analysis</title><p>Under the model base case settings (with no empiric treatment, a 10-day treatment duration, and discontinuation or switch of therapy possible after 7&#x000a0;days), the total inpatient (medical plus drug) costs were $46,168 for linezolid and $46,992 for vancomycin (Table&#x000a0;<xref ref-type="table" rid="T3">3</xref>). Although the drug costs were $1,433 higher with linezolid compared with vancomycin, the medical costs associated with linezolid were $2,256 lower with linezolid than with vancomycin. Overall, treatment with linezolid was associated with lower total costs (by a mean of $824) and greater effectiveness (+2.7% absolute difference in proportion of successfully treated patients) compared with vancomycin. The expected proportions of successfully treated patients were 62.9% and 60.2% for linezolid and vancomycin, respectively. Factoring in these expected success rates, the total costs per successfully treated patient were predicted to be $73,420 (linezolid) and $78,073 (vancomycin), for a total cost savings of $4,653. Thus, the ICER (in this case, incremental cost per successfully treated patient) was in favor of linezolid compared with vancomycin (that is, linezolid dominated vancomycin), owing to linezolid&#x02019;s lower total costs and greater efficacy in successfully treating patients.</p><p>We calculated that, within the model, approximately 80% of the total treatment costs were attributable to hospital stay, primarily ICU costs, because each patient stayed at least 10&#x000a0;days (plus additional days if first-line therapy failed) in the ICU and the cost per day of ICU stay in the United States is very high. General ward costs were higher with vancomycin compared with linezolid, because, even though the length of stay in the hospital was comparable between treatments, there was a higher percentage of patients for whom vancomycin failed as first-line therapy and thus were transitioned to second-line treatment and had an associated longer general ward stay. Moreover, the higher percentage of vancomycin-treated patients requiring second-line therapy may have led to marginally higher costs for additional physician visits and laboratory work. However, drug therapy, physician visits, laboratory tests and SAEs and/or treatment failure each accounted for no more than 5% of the total costs (Table&#x000a0;<xref ref-type="table" rid="T3">3</xref>).</p></sec><sec><title>Sensitivity analysis</title><p>The results of one-way sensitivity analysis (as seen in the Tornado diagram in Figure&#x000a0;<xref ref-type="fig" rid="F2">2</xref>) demonstrated variables that had the greatest impact on the model results. The ICERs ranged from a low of about&#x02009;&#x02212;&#x02009;$240,000, when ICU stay with linezolid was at its lower value of 6.1 (suggesting a dominant scenario for linezolid), to a high of around $210,000, when the clinical efficacy of vancomycin was at its higher value of 52.9% and $160,000, when ICU stay with vancomycin was at a low of 6.6&#x000a0;days. (These ICERs can be considered greater than the acceptable willingness-to-pay (WTP) threshold, making vancomycin the cost-effective option.) There is no clearly defined WTP threshold for successful treatment of one patient, and hence different WTP values were tested in the PSA cost-effectiveness acceptability curve.A cost-effectiveness acceptability curve generated from a PSA is presented in Figure&#x000a0;<xref ref-type="fig" rid="F3">3</xref>. This plot displays the percentages for linezolid being more cost-effective compared to vancomycin at the different WTP thresholds. Linezolid had a 64.4% likelihood of being cost-effective at a WTP threshold of $0 and a 90.8% chance at a WTP threshold of $120,000.</p><fig id="F2" position="float"><label>Figure 2</label><caption><p><bold>One-way sensitivity analyses of key parameters (Tornado diagram).</bold> IV, Intravenous; LIN, Linezolid; MV, Mechanical ventilation; SAE, Serious adverse events; tx, Treatment; VAN, Vancomycin.</p></caption><graphic xlink:href="cc13996-2"/></fig><fig id="F3" position="float"><label>Figure 3</label><caption><p>Cost-effectiveness acceptability curve of linezolid versus vancomycin.</p></caption><graphic xlink:href="cc13996-3"/></fig></sec></sec><sec sec-type="discussion"><title>Discussion</title><p>This economic decision tree analysis is the first, to our knowledge, to mirror real-world clinical conditions by allowing for a switch of therapy if needed (that is, it models first- and second-line treatment) and allowed us to assess the impact of varying treatment parameters, including treatment duration. To our knowledge, no other published studies of NP have researchers accounted for these factors from a US health-care payer perspective. Our results with this model show that linezolid is a cost-effective alternative to vancomycin for the treatment of MRSA-confirmed NP, owing primarily to the higher clinical response rate of linezolid-treated patients compared with vancomycin-treated patients. The higher acquisition cost of linezolid was offset by lower costs of treatment failure and SAEs, as well as fewer days spent in the hospital, when we accounted for combined first-line and second-line therapies. Only direct medical costs were included in the model, with a distinction made between inpatient and outpatient costs. For NP, inpatient costs accounted for the largest proportion of overall costs.</p><p>Linezolid was a more cost-effective treatment option in the majority of one-way sensitivity analyses (vancomycin was cost-effective only under two scenarios: low ICU stay and high vancomycin efficacy rate) and under varying WTP thresholds in PSA. Length of ICU stay and clinical efficacy rate appeared to be the most sensitive variables in one-way analysis, with the greatest impact on the ICER. This was expected, especially with regard to the length of ICU stay, because ICU stay <italic>per diem</italic> is very expensive in the United States and the cost of ICU stay accounts for the largest proportion of total treatment cost.</p><p>Our results are consistent with those reported in two previous cost-effectiveness analyses in which investigators found therapy initiated with linezolid to be less costly and more efficacious than vancomycin for patients with suspected MRSA NP [<xref ref-type="bibr" rid="B18">18</xref>,<xref ref-type="bibr" rid="B19">19</xref>]. Mullins <italic>et al</italic>. applied a retrospective decision-analytic model to pooled efficacy data derived from two clinical trials and health plan hospital claims and determined hospital costs for US patients with suspected NP. When median daily hospital charges and mean treatment durations were factored in, total hospitalization charges were estimated to be $32,636 for linezolid treatment compared with $32,024 for vancomycin treatment. The ICER for linezolid per life saved was $3,600. However, they based their efficacy estimates on a small sample of patients with MRSA NP (<italic>N</italic>&#x02009;=&#x02009;160) and examined the cost-effectiveness of only first-line linezolid or vancomycin treatment.</p><p>In a German cost-effectiveness analysis [<xref ref-type="bibr" rid="B18">18</xref>], the researchers used a decision-analytic model based on previously published clinical data [<xref ref-type="bibr" rid="B14">14</xref>] and found higher clinical cure and survival rates with linezolid, but at a small incremental cost compared with vancomycin, resulting in acceptable ICERs of cost per death avoided and cost per patient cured [<xref ref-type="bibr" rid="B18">18</xref>]. From a clinical standpoint, they demonstrated that linezolid had better efficacy than vancomycin for the treatment of MRSA NP (on the basis of trial data specifically in MRSA-confirmed patients), with fewer patients requiring a switch to second-line therapy. The longer hospital stays associated with switching from vancomycin as first-line treatment to a second-line therapy required additional resource use, including physicians&#x02019; and other health-care professionals&#x02019; time that could have been spent treating other patients.</p><p>Our present economic analysis included patients who received optional empiric therapy (2&#x000a0;days) followed by first- and/or second-line treatment once MRSA was confirmed, with the empiric treatment costs not included in the presented scenarios. Costs, therefore, were not considered in patients who did not have MRSA infection. In clinical practice, initiation with empiric antibiotic treatment is started as soon as MRSA is suspected, and antibiotic treatment success and the related costs of empiric therapy are determined by how well MRSA is predicted and by the proportion of patients with MRSA in the treated population. Our present analysis therefore does not include the costs of initial empiric therapy and the harm that comes from (1) not covering MSSA by using only MRSA coverage, (2) choosing vancomycin and the possibility of renal toxicity developing in a patient without MRSA and (3) not starting empiric therapy with either drug and having a delay in starting appropriate therapy until after culture results have been confirmed. Although we did not address these clinical aspects in our model, they are relevant and important and should be explored in future studies.</p><p>Vancomycin and linezolid are the most commonly recommended and prescribed treatment options for MRSA NP [<xref ref-type="bibr" rid="B11">11</xref>,<xref ref-type="bibr" rid="B12">12</xref>]. Vancomycin has been the mainstay generic for decades; however, challenges with tissue penetration at the site of infection, therapeutic drug monitoring and increased risk for renal dysfunction in NP patients makes the use of this agent more difficult in critically ill patients. In the economic analysis presented here, we used the recent and only clinical trial data specifically designed to evaluate clinical success in the treatment of patients with MRSA NP [<xref ref-type="bibr" rid="B16">16</xref>]. To date, linezolid is the only agent to have proven better clinical success rates in NP than vancomycin in a MRSA-only population. Linezolid is sold as ZYVOX and is under patent in the United States until the end of 2014; thus, use of this agent may increase further with the introduction of generic versions. However, there is another oxazolidinone drug for nosocomial pneumonia currently under development, tedizolid, in ongoing phase III trials. A newer glycopeptide, telavancin, became available in late 2013 for gram positive NP, and while the phase III trials included MRSA patients, the studies were not specifically designed to examine clinical success in the MRSA-only population. Thus, the only MRSA-specific NP study to date is the Wunderink <italic>et al</italic>. study [<xref ref-type="bibr" rid="B16">16</xref>], on which our economic analysis is based.</p><p>Our study has limitations. In the model&#x02019;s base case scenario, we considered the conditions under which the Wunderink clinical trial was performed [<xref ref-type="bibr" rid="B16">16</xref>], which may differ in real-life US clinical practice. Further, because the Wunderink trial enrolled US patients, the results may not be applicable to scenarios outside the United States. The model included only first-line and second-line treatments, not potential later treatment options. However, this is consistent with other published models [<xref ref-type="bibr" rid="B18">18</xref>] and is justifiable because the majority of the resources used and outcomes witnessed were within the first two lines of therapy. In the model, we estimated direct costs only and did not include indirect costs related to lost productivity incurred as a result of the length of hospital stay, convalescence or early mortality.</p><p>We used 60-day mortality data, calculated as a weighted average of the 60-day mortality rates for the modified intent-to-treat population in the clinical trial [<xref ref-type="bibr" rid="B16">16</xref>], which were the best available &#x0201c;proxy&#x0201d; data for this 4-week model, given that the difference between 30 and 60&#x000a0;days was found to be small based on the survival curve derived from the study. In addition, mortality rates were not statistically different in the clinical trial; thus, a cost per LY saved calculation was less relevant, given that the trial was never designed to show a difference in mortality. In fact, patients could have received up to 2&#x000a0;days of vancomycin before being randomized to the study drugs; thus, patients doing poorly on vancomycin would have been less likely to be enrolled in the study, where the chance of being randomized to vancomycin was 50&#x02013;50. Instead of focusing on LYs, we used &#x0201c;proportion of successfully treated patients&#x0201d; instead of QALYs as the efficacy measure in this model, which could be considered a drawback, especially because there is no clearly defined ICER threshold per successfully treated patient. However, we think that successful treatment is a clinically important efficacy measure for NP, and hence it can be argued to be relevant for this model.</p></sec><sec sec-type="conclusions"><title>Conclusion</title><p>Our US health-care system economic model using recent MRSA-specific clinical trial data shows that treatment with linezolid is less costly and more efficacious than treatment with vancomycin for MRSA-confirmed NP. Cost savings with linezolid were derived largely from lower treatment failure rates, fewer days of hospitalization and lower incidence of renal failure. We found our findings to be consistent in sensitivity analyses. In future analyses, researchers should use other country costs and resource-use data to test result generalizability and could model the empiric treatment phase before MRSA confirmation.</p></sec><sec><title>Key messages</title><p>&#x025cf;&#x02003;Linezolid is likely to be a cost-effective alternative compared to vancomycin for the treatment of MRSA NP, primarily owing to the former&#x02019;s better clinical success rate.</p><p>&#x025cf;&#x02003;Higher drug costs for linezolid are offset by lower overall medical costs due to fewer treatment failures and fewer serious adverse events, such as renal failure, as well as fewer days spent in the hospital, when accounting for combined first-line and second-line therapies.</p><p>&#x025cf;&#x02003;For MRSA NP, inpatient costs accounted for the largest proportion of overall costs.</p></sec><sec><title>Abbreviations</title><p>AE: Adverse event; CPT: Current procedural terminology; ICER: Incremental cost-effectiveness ratio; ID: Infectious disease; IV: Intravenous; MRSA: Methicillin-resistant <italic>Staphylococcus aureus</italic>; NP: Nosocomial pneumonia; PSA: Probabilistic sensitivity analysis; SAE: Serious adverse event.</p></sec><sec><title>Competing interests</title><p>This study was sponsored by Pfizer. The following authors declare receiving lecture or advisory board or research grant support: ASi, AFS, DN, MN and JC (Pfizer); AFS, DN and JC (Astellas Pharma); DN (AstraZeneca); AFS, DN and MN (Bayer); JC (Nektar-Bayer); DN (Basilea Pharmaceutica); JC (B&#x02022;R&#x02022;A&#x02022;H&#x02022;M&#x02022;S); AFS, DN and MN (Cubist Pharmaceuticals); DN (Durata Therapeutics); AFS (Forest Laboratories); JC (Janssen-Cilag Pharma); MN (Merck); JC and MN (Sanofi-Aventis); ASi (Sunovion Pharmaceuticals Canada); AFS (Tetraphase Pharmaceuticals); AFS (Theravance Biopharma); and AFS and JC (Trius Therapeutics). CC is an employee of Pfizer. DAP, JMS and XG are employees of Pharmerit. Pharmerit received research funding from Pfizer to develop the model. No funding was provided to the authors for manuscript development. Editorial and medical writing support was provided by Ray Beck, Jr, PhD, of Engage Scientific Solutions and was funded by Pfizer. The authors declare that they have no competing interests other than those described here.</p></sec><sec><title>Authors&#x02019; contributions</title><p>DP was responsible for model conceptualization and design, development of the model inputs and assumptions, programming, analyses, review and interpretation of results and manuscript writing. AFS, ASi, MN and DN were responsible for model conceptualization, design and assumptions; review and interpretation of results; and critical revision of the manuscript. JS was responsible for model conceptualization and design, development of the model inputs and assumptions, programming, analyses, review and interpretation of results and manuscript writing. CC was responsible for model conceptualization and design, financial support, review and interpretation of results and development of the manuscript. XG was responsible for model conceptualization and design, development of the model inputs and assumptions, programming, analyses, review and interpretation of results and critical revision of the manuscript. All authors read and approved the final manuscript.</p></sec>

¹⁸&#x02009;F-Fluoride positron emission tomography/computed tomography for noninvasive <italic>in vivo</italic> quantification of pathophysiological bone metabolism in experimental murine arthritis

<sec><title>Introduction</title><p>Evaluation of disease severity in experimental models of rheumatoid arthritis is inevitably associated with assessment of structural bone damage. A noninvasive imaging technology allowing objective quantification of pathophysiological alterations of bone structure in rodents could substantially extend the methods used to date in preclinical arthritis research for staging of autoimmune disease severity or efficacy of therapeutical intervention. Sodium ¹⁸&#x02009;F-fluoride (¹⁸&#x02009;F-NaF) is a bone-seeking tracer well-suited for molecular imaging. Therefore, we systematically examined the use of ¹⁸&#x02009;F-NaF positron emission tomography/computed tomography (PET/CT) in mice with glucose-6-phosphate isomerase (G6PI)&#x02013;induced arthritis for quantification of pathological bone metabolism.</p></sec><sec><title>Methods</title><p>F-fluoride was injected into mice before disease onset and at various time points of progressing experimental arthritis. Radioisotope accumulation in joints in the fore- and hindpaws was analyzed by PET measurements. For validation of bone metabolism quantified by ¹⁸&#x02009;F-fluoride PET, bone surface parameters of high-resolution &#x003bc;CT measurements were used.</p></sec><sec><title>Results</title><p>Before clinical arthritis onset, no distinct accumulation of ¹⁸&#x02009;F-fluoride was detectable in the fore- and hindlimbs of mice immunized with G6PI. In the course of experimental autoimmune disease, ¹⁸&#x02009;F-fluoride bone uptake was increased at sites of enhanced bone metabolism caused by pathophysiological processes of autoimmune disease. Moreover, ¹⁸&#x02009;F-fluoride signaling at different stages of G6PI-induced arthritis was significantly correlated with the degree of bone destruction. CT enabled identification of exact localization of ¹⁸&#x02009;F-fluoride signaling in bone and soft tissue.</p></sec><sec><title>Conclusions</title><p>The results of this study suggest that small-animal PET/CT using ¹⁸&#x02009;F-fluoride as a tracer is a feasible method for quantitative assessment of pathophysiological bone metabolism in experimental arthritis. Furthermore, the possibility to perform repeated noninvasive measurements <italic>in vivo</italic> allows longitudinal study of therapeutical intervention monitoring.</p></sec>

<contrib contrib-type="author" corresp="yes" id="A1"><name><surname>Irmler</surname><given-names>Ingo M</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>ingo.irmler@med.uni-jena.de</email></contrib><contrib contrib-type="author" id="A2"><name><surname>Gebhardt</surname><given-names>Peter</given-names></name><xref ref-type="aff" rid="I2">2</xref><xref ref-type="aff" rid="I3">3</xref><email>peter.gebhardt@med.uni-jena.de</email></contrib><contrib contrib-type="author" id="A3"><name><surname>Hoffmann</surname><given-names>Bianca</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>bianca.hoffmann@hki-jena.de</email></contrib><contrib contrib-type="author" id="A4"><name><surname>Opfermann</surname><given-names>Thomas</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>thomas.opfermann@med.uni-jena.de</email></contrib><contrib contrib-type="author" id="A5"><name><surname>Figge</surname><given-names>Marc-Thilo</given-names></name><xref ref-type="aff" rid="I2">2</xref><xref ref-type="aff" rid="I4">4</xref><email>thilo.figge@hki-jena.de</email></contrib><contrib contrib-type="author" id="A6"><name><surname>Saluz</surname><given-names>Hans P</given-names></name><xref ref-type="aff" rid="I2">2</xref><xref ref-type="aff" rid="I4">4</xref><email>hanspeter.saluz@hki-jena.de</email></contrib><contrib contrib-type="author" id="A7"><name><surname>Kamradt</surname><given-names>Thomas</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>thomas.kamradt@med.uni-jena.de</email></contrib>

Arthritis Research &#x00026; Therapy

<sec sec-type="intro"><title>Introduction</title><p>Rheumatoid arthritis (RA) is one of the most common autoimmune diseases, affecting approximately 1% of the population in Europe and North America. Bone erosion can be detected as early as several weeks after the onset of the first clinical signs and symptoms of RA [<xref ref-type="bibr" rid="B1">1</xref>]. Pathogenesis in preclinical models is similar to that in clinical RA; it is characterized by inflammation and bone destruction. In the past few decades, histopathological evaluation of joint sections has mainly been used for assessment of inflammation and bone destruction in small-rodent models of arthritis. Recently, we showed the benefit of state-of-the-art imaging modalities for visualization and quantitative assessment of inflammation in glucose-6-phosphate isomerase (G6PI)&#x02013;induced arthritis using 2-deoxy-2-¹⁸&#x02009;F-fluoroglucose (¹⁸&#x02009;F-FDG) positron emission tomography/computed tomography (PET/CT) [<xref ref-type="bibr" rid="B2">2</xref>]. Also, it has been shown that cell proliferation can be detected in experimental arthritis with the PET proliferation tracer 3&#x02032;-deoxy-3&#x02032;-¹⁸&#x02009;F-fluorothymidine [<xref ref-type="bibr" rid="B3">3</xref>].</p><p>¹⁸&#x02009;F-fluoride not only can be used to label glucose and other molecules of physiologic relevance but also has favorable properties in the form of sodium ¹⁸&#x02009;F-fluoride (¹⁸&#x02009;F-NaF) as a radiotracer <italic>per se</italic> in noninvasive <italic>in vivo</italic> imaging to investigate musculoskeletal diseases [<xref ref-type="bibr" rid="B4">4</xref>]. The use of ¹⁸&#x02009;F-NaF as a bone imaging probe was established by Blau <italic>et al</italic>. in the early 1960s [<xref ref-type="bibr" rid="B5">5</xref>], but it was subsequently replaced by ^99mTc-labeled tracers due to their availability, lower costs and the lower energy of 140-keV photons, allowing the use of &#x003b3;-cameras. ¹⁸&#x02009;F-fluoride PET is an increasingly used molecular imaging modality, not only in human skeletal disorders but also in small-animal preclinical research [<xref ref-type="bibr" rid="B6">6</xref>-<xref ref-type="bibr" rid="B8">8</xref>]. This is facilitated by the distribution and use of three-dimensional PET scanners with high spatial resolution in clinical medicine and advantages over ^99mTc-labeled bone agents used for skeletal scintigraphy, such as higher diagnostic accuracy and reduced scanning time. Compared to &#x003b3;-cameras, molecular imaging using PET provides the advantages of higher spatial resolution, higher sensitivity and three-dimensional tomographic image reconstruction. Furthermore, the combination of PET with &#x003bc;CT enables attenuation correction of radiotracer signaling, allowing quantitative measurements using ¹⁸&#x02009;F-fluoride PET/CT.</p><p>Applied ¹⁸&#x02009;F-NaF, dissociated into Na⁺ and ¹⁸&#x02009;F^&#x02212;, is rapidly cleared from the blood and accumulates in the bone, where, on the hydroxyapatite surface, an OH^&#x02212; ion is replaced by an ¹⁸&#x02009;F^&#x02212; ion to form fluorapatite. The incorporation of ¹⁸&#x02009;F-fluoride in the bone is determined by vascular perfusion and bone surface accessible for ion exchange, indirectly reflecting bone formation and bone resorption [<xref ref-type="bibr" rid="B9">9</xref>]. This means that ¹⁸&#x02009;F-NaF can be used not only for the common measurement of bone mineral deposition but also for visualization of osteolytic increases of exposed bone surface, such as in the context of musculoskeletal autoimmune disease [<xref ref-type="bibr" rid="B10">10</xref>].</p><p>In clinical oncology, primary bone tumors and skeletal metastasis can reliably be detected by ¹⁸&#x02009;F-fluoride PET [<xref ref-type="bibr" rid="B4">4</xref>,<xref ref-type="bibr" rid="B11">11</xref>]. In mice, pathological osteoblastic activity can be detected even earlier by ¹⁸&#x02009;F-fluoride PET/CT imaging than by radiography and corresponds to histological evaluation of increased bone formation [<xref ref-type="bibr" rid="B7">7</xref>]. As with bone tumor pathogenesis, a pathologically increased bone metabolism is a central feature of chronic arthritis, resulting in functional disorders of the joints [<xref ref-type="bibr" rid="B12">12</xref>]. Therefore, in our present study, we examined the use of ¹⁸&#x02009;F-NaF small-animal PET/CT for the quantitative and noninvasive <italic>in vivo</italic> assessment of pathophysiological bone metabolism in acute and chronic stages of experimental G6PI-induced murine arthritis. ¹⁸&#x02009;F-NaF PET quantification of bone destruction, visible as lesions in cortical bone surface and dysregulated bone formation, was validated using high-resolution CT measurements of the paws.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>Glucose-6-phosphate isomerase&#x02013;induced arthritis</title><p>DBA/1 mice were bred at the animal facility of the Jena University Hospital (Jena, Germany). All animal studies were approved by the local commission for animal protection (Th&#x000fc;ringer Landesamt f&#x000fc;r Verbraucherschutz, Bad Langensalza, Germany; registered number 02-045/08). Arthritis was induced as described elsewhere [<xref ref-type="bibr" rid="B13">13</xref>]. In brief, DBA/1 mice were immunized subcutaneously with 400&#x000a0;&#x003bc;g of recombinant human G6PI in emulsified complete Freund&#x02019;s adjuvant (Sigma-Aldrich, Taufkirchen, Germany). Macroscopic evaluation of arthritis was performed according to severity of clinical manifestations in wrist and ankle joints, in metacarpophalangeal and metatarsophalangeal joints, and in digits and toes. Swelling and redness in wrist and ankle joints and in metacarpophalangeal and metatarsophalangeal joints, respectively, were graded from 0 to 3. A score of 0 indicates no macroscopically recognizable signs of arthritis, 1 indicates swelling and redness, 2 means strong swelling and redness and 3 indicates massive swelling and redness. Additionally, the number of digits and toes with inflamed joints were divided in half to avoid assessment imbalance because inflammation in the G6PI-induced arthritis model is located mainly in proximal joints of the paws [<xref ref-type="bibr" rid="B2">2</xref>]. To calculate the total clinical score per animal, results from all paws were summed. For PET/CT measurements, mice were anesthetized with 1.5% to 2% isoflurane (Deltaselect, Dreieich, Germany) vaporized in oxygen (1.5&#x000a0;L/min) to prevent animal movement and reduce imaging artifacts. Respiration of mice under anesthesia was monitored. ¹⁸&#x02009;F-fluoride (half-life&#x02009;=&#x02009;109&#x000a0;minutes; Eckert &#x00026; Ziegler, Bad Berka, Germany) with an activity of 10.4&#x02009;&#x000b1;&#x02009;0.8&#x000a0;MBq was injected intravenously into the lateral tail vein. Longitudinal arthritis imaging was performed at various time points of acute and chronic clinical arthritis (<italic>n</italic>&#x02009;=&#x02009;3 to 6 mice per time point). We obtained dynamic PET scans for kinetic analysis of tracer uptake in nonimmunized mice (<italic>n</italic>&#x02009;=&#x02009;3).</p></sec><sec><title>Positron emission tomography/computed tomography <italic>in vivo</italic> imaging</title><p><italic>In vivo</italic>¹⁸&#x02009;F-NaF imaging was performed using a Siemens Inveon small-animal multimodality PET/CT system (Preclinical Solutions, Siemens Healthcare Molecular Imaging, Knoxville, TN, USA), characterized by the combination of two independently operating PET and CT scanners. Radial, tangential and axial resolutions at the center of the field of view of the PET module are 1.5&#x000a0;mm for this imaging modality [<xref ref-type="bibr" rid="B14">14</xref>,<xref ref-type="bibr" rid="B15">15</xref>]. PET image acquisition was carried out with a coincidence timing window of 3.4&#x000a0;ns and an energy window of 350 to 650&#x000a0;keV. PET data acquisition was performed for 1,800&#x000a0;seconds, starting 50&#x000a0;minutes after tracer application. In kinetic analysis, PET data acquisition lasted 5,400&#x000a0;seconds and started concomitantly with radiotracer injection. Images were reconstructed into three-dimensional images using Fourier rebinning and three-dimensional ordered-subset expectation maximization algorithm. Attenuation of PET data was corrected based on the CT measurements. The CT module consists of a cone-beam X-ray &#x003bc;CT source (50-&#x003bc;m focal spot size) and a 3,072&#x02009;&#x000d7;&#x02009;2,048&#x02013;pixel X-ray detector. In our &#x003bc;CT imaging protocol, we used an axial&#x02013;transaxial resolution of 2,048&#x02009;&#x000d7;&#x02009;2,048&#x02013;pixel, 80 kVp at 500&#x000a0;&#x003bc;A, 360&#x000b0; rotation and 360 projections per bed position for paw measurements and 3,072&#x02009;&#x000d7;&#x02009;2,048&#x02013;pixel, 220&#x000b0; rotation and 120 projections per bed position for whole-animal attenuation scans, respectively. CT images were reconstructed using a Shepp-Logan filter and cone-beam&#x02013;filtered back projection. To reduce stress due to overly prolonged measurement times, only high-resolution &#x003bc;CT data from hindpaws were acquired for correlation analysis of bone surface assessment with ¹⁸&#x02009;F-fluoride PET and CT.</p></sec><sec><title>Assessment of pathophysiological bone metabolism with ¹⁸&#x02009;F-fluoride or &#x003bc;CT</title><p>Quantitative analysis of ¹⁸&#x02009;F-fluoride accumulation in the period from 50 to 80&#x000a0;minutes after ¹⁸&#x02009;F-fluoride injection or 0 to 90&#x000a0;minutes, respectively, was enabled by image fusion technology in Siemens Inveon Research Workplace 4.0 software. For dynamic PET analysis, scans were started 5&#x000a0;seconds before radiotracer injection and continued for 90&#x000a0;minutes. The 90-minute data set was divided into 45 time frames (6&#x02009;&#x000d7;&#x02009;10&#x000a0;seconds, 6&#x02009;&#x000d7;&#x02009;20&#x000a0;seconds, 7&#x02009;&#x000d7;&#x02009;60&#x000a0;seconds, 10&#x02009;&#x000d7;&#x02009;120&#x000a0;seconds, 10&#x02009;&#x000d7;&#x02009;180&#x000a0;seconds and 6&#x02009;&#x000d7;&#x02009;300&#x000a0;seconds) during histogramming to construct time&#x02013;activity curves. For static analysis, data from 75 to 80&#x000a0;minutes after ¹⁸&#x02009;F-fluoride injection were analyzed in a single time frame. Radioisotope activity in the venous blood pool or in fore- or hindpaws, reflecting bone incorporation of ¹⁸&#x02009;F-fluoride, was measured as standardized uptake value (SUV; g/ml) using PMOD 3.15 software (PMOD Technologies Ltd, Zurich, Switzerland) or Siemens Inveon Research Workplace 4.0 software. Guided by maximum intensity projection images, volumes of interest were drawn as spheres (ellipsoids) over anatomic structures of bones and joints. A threshold of 40% (max&#x02013;min) was used to separate the site of tracer accumulation from background tissue signaling. Visualization of skeletal elements of the hindpaws was also done using PMOD 3.15 software.</p><p>Analyses of metatarsal CT data were performed using Definiens Developer XD&#x02122; 2.0.3 build 2015 software (Definiens, Munich, Germany). For segmentation of metatarsal bones, the automatic threshold routine was used. Then, individual bones were discriminated by watershed segmentation. Oversegmented fragments were merged manually to achieve three-dimensional objects of each bone. After identification of metatarsals 1 to 5, a local threshold was applied on each metatarsal bone to refine bone margins and remove remaining nonbone pixels. The threshold was calculated as the difference between mean pixel gray value and standard deviation of pixel gray values. Quantitative measurements of bone surface and bone volume were calculated for each metatarsal bone.</p></sec><sec><title>Statistical analysis</title><p>Statistical differences between groups were evaluated using the nonparametric Mann-Whitney <italic>U</italic> test. Correlation analyses (Spearman test) were performed for PET data and clinical scoring or PET and CT data, respectively, whereas CT surface data of all metatarsals, per paw, were summed. Statistical significance was accepted at <italic>P</italic>&#x02009;&#x0003c;&#x02009;0.05 (*<italic>P</italic>&#x02009;&#x0003c;&#x02009;0.05, **<italic>P</italic>&#x02009;&#x0003c;&#x02009;0.01, ***<italic>P</italic>&#x02009;&#x0003c;&#x02009;0.001). In bar charts and text, data are given as arithmetic means and standard errors of the mean (SEM). All calculations were performed using the software package IBM SPSS Statistics version 20.0 (IBM, Ehningen, Germany).</p></sec></sec><sec sec-type="results"><title>Results</title><sec><title>Inflammation and bone destruction in glucose-6-phosphate isomerase&#x02013;induced arthritis</title><p>Immunization of animals with G6PI induced severe paw inflammation in DBA/1 mice (Figure&#x000a0; <xref ref-type="fig" rid="F1">1</xref>A). Experimental disease was characterized by redness and swelling in wrist and ankle joints, as well as in dorsal areas of the paws and in metatarsophalangeal, metacarpophalangeal and phalangeal joints. Macroscopic scoring revealed onset of polyarthritis around 9&#x000a0;days after disease induction, maximum of inflammation at day 14 (d14) and a subsequent decrease until d33 (Figure&#x000a0; <xref ref-type="fig" rid="F1">1</xref>B). Histopathology revealed bone erosion in skeletal elements of the inflamed paws (Figure&#x000a0; <xref ref-type="fig" rid="F1">1</xref>C).</p><fig id="F1" position="float"><label>Figure 1</label><caption><p><bold>Joint inflammation in glucose-6-phosphate isomerase&#x02013;induced arthritis. (A)</bold> Photographs showing inflammation in fore- and hindpaws of arthritic mice (two at right) and nonimmunized control animals (two at left). <bold>(B)</bold> Clinical course of inflammatory arthritis in DBA/1 mice (<italic>n</italic>&#x02009;=&#x02009;7). GPI, Glucose-6-phosphate isomerase induction. <bold>(C)</bold> Arthritis was associated with destruction of bone. Shown is a representative image of metatarsal bone revealing cortical bone lesions (arrows) and infiltration of immune cells to the sites of inflammation. B, Bone; BM, Bone marrow; J, Joint; CI, Cellular infiltrate; MT, Metatarsal. Original magnification&#x02009;=&#x02009;20&#x000d7; (left) and 200&#x000d7; (right).</p></caption><graphic xlink:href="ar4670-1"/></fig></sec><sec><title>Kinetics of ¹⁸&#x02009;F-fluoride bone uptake in mice</title><p>Radiotracer applied intravenously in the tail vein was immediately (0 to 10&#x000a0;seconds postinjection) transported to the heart via the blood flow (Figure&#x000a0; <xref ref-type="fig" rid="F2">2</xref>A). Subsequent whole-body distribution of ¹⁸&#x02009;F-fluoride resulted in thoracic and intestinal PET signaling (10 to 20&#x000a0;seconds postinjection). After 3&#x000a0;minutes, the pattern of ¹⁸&#x02009;F-fluoride PET signaling indicated onset of bone accumulation, and signaling from the kidneys and bladder revealed partial radiotracer excretion. Twenty minutes after radiotracer injection, signaling of nonexcreted ¹⁸&#x02009;F-fluoride was restricted mainly to bone and continued to increase until the end of measurement at 90&#x000a0;minutes.</p><fig id="F2" position="float"><label>Figure 2</label><caption><p><bold>Biodistribution and bone accumulation of </bold>^{<bold>18</bold>}&#x02009;<bold>F-fluoride in na&#x000ef;ve mice. (A)</bold> Positron emission tomography (PET) scans (maximum intensity projection) of a na&#x000ef;ve mouse after intravenous ¹⁸&#x02009;F-fluoride tail vein injection. Radiotracer was immediately distributed in the whole organism via the circulation and rapidly accumulated in the skeletal elements. <bold>(B)</bold> The 90-minute ¹⁸&#x02009;F-fluoride time&#x02013;activity curve reveals rapid ¹⁸&#x02009;F-fluoride clearance from the bloodstream (via the venae cavae) and subsequent ¹⁸&#x02009;F-fluoride accumulation in the bones of the fore- and hindpaws. SUV, Standardized uptake value.</p></caption><graphic xlink:href="ar4670-2"/></fig><p>Concomitantly, the quantitative time&#x02013;activity curve analysis of ¹⁸&#x02009;F-fluoride paw uptake revealed rapid radiotracer accumulation beginning immediately after injection and tracer signaling remaining at almost the same level from 40 to 90&#x000a0;minutes postinjection (Figure&#x000a0; <xref ref-type="fig" rid="F2">2</xref>B). At 45&#x000a0;minutes postinjection, 90% of ¹⁸&#x02009;F-fluoride uptake was achieved (SUV_mean&#x02009;=&#x02009;1.8), compared to bone radiotracer signaling after 90&#x000a0;minutes (SUV_mean&#x02009;=&#x02009;1.9). In contrast, blood pool tracer activity measured in the <italic>venae cavae</italic> region showed a rapid decline after an activity maximum (SUV_max&#x02009;=&#x02009;15.2), within minutes following injection, reflecting rapid clearance of ¹⁸&#x02009;F-fluoride from the blood.</p></sec><sec><title>Localization of exact sites of ¹⁸&#x02009;F-fluoride paw bone uptake in glucose-6-phosphate isomerase&#x02013;induced arthritis</title><p>Detailed three-dimensional anatomical &#x003bc;CT measurements allowed identification of metatarsal and phalangeal bones and joints, and PET imaging revealed radiotracer uptake in the paws during the course of experimental arthritis (Figure&#x000a0; <xref ref-type="fig" rid="F3">3</xref>A). Low ¹⁸&#x02009;F-fluoride accumulation before immunization switched to a distinct tracer uptake in the acute (d14) and chronic (d28) stages of disease and declined during late remitting arthritis (d50). Coregistration of PET and CT data enabled exact localization of ¹⁸&#x02009;F-fluoride accumulation in arthritic mice.Detailed, three-dimensional anatomical information obtained by high-resolution &#x003bc;CT enabled visualization of the bone surfaces of phalangeal, metatarsal, tarsal and tibial bones. Compared to nonarthritic control animals (Figure&#x000a0; <xref ref-type="fig" rid="F3">3</xref>B, left), pathologic bone metabolism in the mice with G6PI-induced arthritis induced a distinct increase in bone roughness (Figure&#x000a0; <xref ref-type="fig" rid="F3">3</xref>B, right).</p><fig id="F3" position="float"><label>Figure 3</label><caption><p><bold>Localization of peripheral </bold>^{<bold>18</bold>}&#x02009;<bold>F-fluoride accumulation in glucose-6-phosphate isomerase&#x02013;induced arthritis. (A)</bold> In naive mice, administration of ¹⁸&#x02009;F-fluoride resulted in low bone tracer accumulation in the subtalar joints. Under acute arthritic conditions (day 14 (d14)), there was distinct radiotracer signaling at sites of manifested disease, whereas increased ¹⁸&#x02009;F-fluoride uptake was detectable in chronic arthritis (d28). In late remitting arthritis (d50), ¹⁸&#x02009;F-fluoride uptake declined. Shown are representative images of murine hindpaws obtained with micro&#x02013;computed tomography (&#x003bc;CT; top), ¹⁸&#x02009;F-fluoride positron emission tomography (PET; middle) and coregistered PET and CT (bottom). <bold>(B)</bold> High-resolution &#x003bc;CT images of hindpaws obtained before arthritis onset (left) and with chronic arthritis (right) reveal the impact of inflammatory autoimmune disease on bone and joint integrity.</p></caption><graphic xlink:href="ar4670-3"/></fig></sec><sec><title>Quantification of pathophysiological bone metabolism using ¹⁸&#x02009;F-fluoride PET in progressing murine arthritis</title><p>To analyze the usability of the ¹⁸&#x02009;F-fluoride PET imaging modality for quantification of pathological bone metabolism in experimental arthritis models, we measured ¹⁸&#x02009;F-fluoride uptake in all paws of mice with G6PI-induced arthritis at five different time points of disease. Before onset of clinical arthritis (that is, in bone tissue without pathological changes (d8), there were only low levels of ¹⁸&#x02009;F-fluoride accumulation in wrist (Figure&#x000a0; <xref ref-type="fig" rid="F4">4</xref>A) and ankle (Figure&#x000a0; <xref ref-type="fig" rid="F4">4</xref>B) joints, measured as SUV. In acute clinical arthritis at d11, paw inflammation was associated with an increase in bone metabolism and induced a significant increase of ¹⁸&#x02009;F-fluoride uptake compared to d8 (<italic>P</italic> =&#x02009;0.019 (wrist) and <italic>P</italic>&#x02009;=&#x02009;0.002 (ankle)). Mean SUVs at d8 were 2.08&#x02009;&#x000b1;&#x02009;0.1 in carpal joints and 4.0&#x02009;&#x000b1;&#x02009;0.2 in tarsal joints, and at d11 mean SUVs were 3.8&#x02009;&#x000b1;&#x02009;0.6 in carpal joints and 5.8&#x02009;&#x000b1;&#x02009;0.5 in tarsal joints. Mean ¹⁸&#x02009;F-fluoride uptake was further significantly increased at d18 (4.4&#x02009;&#x000b1;&#x02009;0.5 (carpal) and 6.3&#x02009;&#x000b1;&#x02009;0.5 (tarsal)) and d25 (5.6&#x02009;&#x000b1;&#x02009;0.6 (carpal) and 6.5&#x02009;&#x000b1;&#x02009;0.6 (tarsal)), followed by a decrease at d39 (4.4&#x02009;&#x000b1;&#x02009;0.5 (carpal) and 5.0&#x02009;&#x000b1;&#x02009;0.4 (tarsal)). PET data revealed a 1.3- to 2.8-fold increase in ¹⁸&#x02009;F-fluoride bone uptake in the paws at progressive stages of inflammatory arthritis, which always differed significantly from ¹⁸&#x02009;F-fluoride accumulation before onset of clinical disease. This finding indicated a rise of pathological bone-degrading and bone-forming processes leading to joint dysfunction, characteristic for RA and its animal models.</p><fig id="F4" position="float"><label>Figure 4</label><caption><p><bold>Quantification of </bold>^{<bold>18</bold>}&#x02009;<bold>F-fluoride uptake in progressive arthritis.</bold> Pathological effects in arthritic bone metabolism differed from ¹⁸&#x02009;F-fluoride accumulation before clinical arthritis onset can be seen above an arbitrary baseline of 2 and 4 standardized uptake value (SUV) for fore and hindpaws, respectively. <bold>(A)</bold> Two days after onset of clinical arthritis at day 9, uptake of ¹⁸&#x02009;F-fluoride in forelimbs was significantly enhanced (day 11). In chronic arthritis at days 18 and 25, ¹⁸&#x02009;F-fluoride accumulation was further increased significantly compared to day 8, whereas in late chronic arthritis (day 39), tracer uptake was declining (<italic>n</italic>&#x02009;=&#x02009;6 to 12 paws per time point). <bold>(B)</bold> Hindlimb pathological arthritic bone metabolism, reflected by ¹⁸&#x02009;F-fluoride signaling in the tarsal joint and metatarsal bones, coincided with forelimb ¹⁸&#x02009;F-fluoride signaling (<italic>n</italic>&#x02009;=&#x02009;6 to 12 paws per time point). <bold>(C)</bold> High-level ¹⁸&#x02009;F-fluoride uptake in distal femoral bone was similar at various time points in the course of arthritis pathogenesis, except for a significant decrease at day 39. <bold>(D)</bold> Low-level ¹⁸&#x02009;F-fluoride uptake in medial cortical femoral bone was comparable at various time points of arthritis pathogenesis. *<italic>P</italic>&#x02009;&#x0003c;&#x02009;0.05, **<italic>P</italic>&#x02009;&#x0003c;&#x02009;0.01, ***<italic>P</italic>&#x02009;&#x0003c;&#x02009;0.001.</p></caption><graphic xlink:href="ar4670-4"/></fig><p>To examine the general effects of arthritis on ¹⁸&#x02009;F-fluoride uptake in elements of the skeleton, we quantified PET signaling in medial and distal femoral bones. SUVs revealed high tracer uptake (SUV&#x02009;&#x0003e;&#x02009;10) in trabecular distal areas of the femur (Figure&#x000a0; <xref ref-type="fig" rid="F4">4</xref>C). Except for late remitting arthritis (d39), SUVs remained at a similar level at different stages of G6PI-induced arthritis, indicating only a slight influence of pathological bone metabolism associated with arthritis on distal femoral tracer uptake. In contrast to trabecular distal areas, mean SUVs in median cortical areas of the femur ranged from 2.1 to 2.2 in all stages of arthritis (Figure&#x000a0; <xref ref-type="fig" rid="F4">4</xref>D), additionally negating general effects of arthritis on bone accumulation of ¹⁸&#x02009;F-fluoride at sites without clinical manifestations of disease. ¹⁸&#x02009;F-fluoride PET/CT measurements in healthy mice without any pathological conditions revealed a nonequal pattern of bone tracer uptake (Additional file <xref ref-type="supplementary-material" rid="S1">1</xref>: Figures S1A and S1B). There were hotspots of PET signaling in proximal joints of the pectoral girdle; in the pelvis, including the knees; and in spinal and cranial bones. Volume-rendering of the knee joints revealed a coincidence of increased ¹⁸&#x02009;F-fluoride accumulation and trabecular bone (Additional file <xref ref-type="supplementary-material" rid="S1">1</xref>: Figure S1C). Therefore, we assume that the increased surface in trabecular bones, meeting the demands of distinct mechanical stress or saving weight, is the cause for the observed heterogeneous pattern of ¹⁸&#x02009;F-fluoride uptake in skeletal bone.</p></sec><sec><title>Quantification of pathophysiological bone metabolism using &#x003bc;CT</title><p>To validate assessment of pathophysiological bone metabolism by ¹⁸&#x02009;F-fluoride PET imaging, bone parameters in &#x003bc;CT data were quantified. Bone surface showed no difference for metatarsals 1 to 5 between unimmunized mice and mice with acute inflammatory arthritis (Figure&#x000a0; <xref ref-type="fig" rid="F5">5</xref>A, d14). In chronic and remitting arthritis (Figure&#x000a0; <xref ref-type="fig" rid="F5">5</xref>A, d28, d35 and d50), bone surface was significantly increased due to bone lesions or dysfunctional bone formation. The course of bone surface assessment was consistent with the course of SUV data obtained by ¹⁸&#x02009;F-fluoride PET. The volume of metatarsal bones revealed a slight increase at d35, followed by a subsequent decline to previous levels in late experimental arthritis at d50 (Figure&#x000a0; <xref ref-type="fig" rid="F5">5</xref>B). Overall, bone volume was largely unaffected by arthritis pathogenesis.</p><fig id="F5" position="float"><label>Figure 5</label><caption><p><bold>Quantification bone parameters on high-resolution &#x003bc;CT images. (A)</bold> Quantification of the surface of all five metatarsal bones (MT1 to MT5) revealed no difference in bone structure between na&#x000ef;ve mice and those with acute arthritis (day 14 (d14)), when inflammation was at its maximum. In contrast, bone surface area was significantly increased in chronic arthritis (d28 and d35). In late chronic arthritis (d50), surface area declined to preinflammation levels. <bold>(B)</bold> Volumes of metatarsal bones were almost unaffected by disease pathogenesis, except for significant increases in MT1, MT4 and MT5 at d28 and significant decreases in MT2 and MT3 at d50 (<italic>n</italic>&#x02009;=&#x02009;6 to 12 paws per time point). *<italic>P</italic>&#x02009;&#x0003c;&#x02009;0.05, **<italic>P</italic>&#x02009;&#x0003c;&#x02009;0.01, ***<italic>P</italic>&#x02009;&#x0003c;&#x02009;0.001.</p></caption><graphic xlink:href="ar4670-5"/></fig></sec><sec><title>Regression analysis of clinical parameters and quantitative PET and CT assessment of bone parameters</title><p>To validate quantification of pathological bone metabolism by ¹⁸&#x02009;F-fluoride PET imaging in experimental arthritis, we performed correlation analysis of quantitative PET results with semiquantitative clinical scoring data and assessment of bone surface roughness destruction by &#x003bc;CT. Although inflammation and bone destruction are associated but functionally not necessarily constrained mechanisms of arthritis pathophysiology, macroscopic clinical scoring and SUV of ¹⁸&#x02009;F-fluoride were significantly correlated (Figure&#x000a0; <xref ref-type="fig" rid="F6">6</xref>A). More importantly, there was a significant correlation of quantitative assessment of pathological bone metabolism by ¹⁸&#x02009;F-fluoride SUV and assessment of bone surface by &#x003bc;CT (Figure&#x000a0; <xref ref-type="fig" rid="F6">6</xref>B), demonstrating the feasibility of these modalities for quantification of bone destruction in experimental arthritis. Therefore, small-animal ¹⁸&#x02009;F-fluoride PET/CT is a reliable imaging method for <italic>in vivo</italic> quantification of arthritis-induced bone erosion and bone malformation, correlating with assessment of pathophysiologic bone surface alterations using &#x003bc;CT.</p><fig id="F6" position="float"><label>Figure 6</label><caption><p><bold>Correlation analysis of arthritis assessment by </bold>^{<bold>18</bold>}&#x02009;<bold>F-fluoride and &#x003bc;CT imaging. (A)</bold> Correlation analysis of quantitative ¹⁸&#x02009;F-fluoride positron emission tomography/computed tomography (PET/CT) hindpaw measurements and semiquantitative macroscopic scoring of paws at different stages of glucose-6-phosphate isomerase (G6PI)&#x02013;induced arthritis demonstrated a significant correlation of both parameters. SUV, Standardized uptake value. <bold>(B)</bold> Quantitative assessment of ¹⁸&#x02009;F-fluoride paw uptake PET and the sum of surface of metatarsal bones 1 to 5 according to &#x003bc;CT data were significantly correlated (<italic>n</italic>&#x02009;=&#x02009;60 paws). *<italic>P</italic>&#x02009;&#x0003c;&#x02009;0.05, **<italic>P</italic>&#x02009;&#x0003c;&#x02009;0.01, ***<italic>P</italic>&#x02009;&#x0003c;&#x02009;0.001.</p></caption><graphic xlink:href="ar4670-6"/></fig></sec></sec><sec sec-type="discussion"><title>Discussion</title><p>The usefulness of the small-animal PET/CT imaging modality using ¹⁸&#x02009;F-NaF as a radiotracer is not restricted to the detection of primary bone tumors and skeletal metastasis in cancer research. ¹⁸&#x02009;F-fluoride PET/CT imaging is also a valid technique for use in the assessment of disease severity according to pathological bone turnover in the field of preclinical arthritis research. In this study, in which we employed the G6PI-induced arthritis model, we focused on the feasibility of using ¹⁸&#x02009;F-fluoride PET to visualize and quantify pathological bone metabolism in distal murine arthritic joints noninvasively and <italic>in vivo</italic>. In addition to joint inflammation, which can easily be quantified by ¹⁸&#x02009;F-FDG PET/CT according to activation and proliferation of resident cells and migrated cells of the immune system, bone damage is the second major parameter used for assessment of arthritis severity [<xref ref-type="bibr" rid="B2">2</xref>].</p><p>In contrast to ¹⁸&#x02009;F-FDG, which is trapped in cells at sites of inflammation due to pathologically increased glucose metabolism, ¹⁸&#x02009;F-fluoride represents for specific radiotracer accumulation in the bone. Erosive processes degrading bone and cartilage and dysfunctional bone repair mechanisms in RA and its animal models are associated with an increased bone surface. Therefore, the increased mineral-binding capacity results in site-specific ¹⁸&#x02009;F-fluoride uptake in arthritic joints, which can easily be used for visualization and, more importantly, provides a measurement method for the quantification of pathological bone metabolism in preclinical arthritis models. The results of our studies show increased uptake of ¹⁸&#x02009;F-fluoride in the paws of arthritic mice and reveal that the ¹⁸&#x02009;F-fluoride PET/CT quantification of pathologic bone metabolism in arthritis pathogenesis significantly correlated with our assessment of pathophysiologic bone surface alterations based on high-resolution &#x003bc;CT measurements.</p><p>Radioisotope imaging is, to date, one of the most applicable imaging modalities for functional and whole-body <italic>in vivo</italic> quantification of bone metabolism in mice. Compared to other radiopharmaceuticals used for bone imaging, such as ^99mTc, ¹⁸&#x02009;F-fluoride has some beneficial attributes. First, ¹⁸&#x02009;F-fluoride has a high affinity to bone, resulting in favorable skeletal kinetics. Within 60&#x000a0;minutes after intravenous injection, only 10% of the injected dose is still located in the bloodstream [<xref ref-type="bibr" rid="B10">10</xref>]. Thus the concurrence of rapid bone uptake and fast blood-pool clearance yields a favorable bone-to-background ratio. Additionally, ¹⁸&#x02009;F-fluoride does not accumulate in inflamed soft tissue and only minimally binds to serum proteins [<xref ref-type="bibr" rid="B16">16</xref>]. Furthermore, small-animal ¹⁸&#x02009;F-NaF PET measurements have an excellent reproducibility in animal models of bone disease [<xref ref-type="bibr" rid="B6">6</xref>,<xref ref-type="bibr" rid="B17">17</xref>].</p><p>One limiting factor in ¹⁸&#x02009;F-fluoride PET imaging might be vascularization of the tissue restricting tracer delivery. In contrast to epithelial tissue, the circulation in well-vascularized bone tissue is less affected by exogenous factors, allowing reproducible data acquisition. In experimental arthritis, the increased vascularization and blood flow in inflamed tissue may influence tracer delivery and, therefore, PET signaling at stages of acute inflammation. Nevertheless, at time points of maximal tracer uptake in our experiments, clinical arthritis was already remitting and macroscopically visible signs of inflammation were diminished. Another important aspect of ¹⁸&#x02009;F-fluoride PET imaging is bone structure. Skeletal elements consisting of cortical and trabecular bone result in strong baseline PET signaling. This is presumably caused by the manifold increase in bone surface and less by increased metabolism at these sites, as binding capacities of cortical and trabecular bone are only slightly different. However, both decrease significantly if bone matrix is demineralized [<xref ref-type="bibr" rid="B18">18</xref>]. Therefore, if bones with cancellous and noncancellous structures are located near regions of interest, data analyses require a high degree of anatomical accuracy, which can be achieved with &#x003bc;CT. In quantification of experimental arthritis severity, we found that the effect of high baseline signaling in tibial and radial bones was of negligible relevance for ¹⁸&#x02009;F-fluoride PET imaging, because signaling hotspots of arthritis pathogenesis were located in tarsal, metatarsal and carpal and metacarpal bones. Another source of a high and pathological ¹⁸&#x02009;F-fluoride baseline PET signaling independent of autoimmune disease pathogenesis might be the preoccurrence of osteoarthritis resulting in mechanical bone erosion. Whereas this aspect is not relevant for imaging in experimental arthritis models, it should be considered in RA bone imaging, where age is a risk factor.</p></sec><sec sec-type="conclusions"><title>Conclusion</title><p>In our present study, we demonstrate the capability of ¹⁸&#x02009;F-fluoride PET to monitor and quantify pathological bone conditions in the model of G6PI-induced arthritis. Furthermore, we validated this bone imaging technique successfully by using bone surface &#x003bc;CT data. Because ¹⁸&#x02009;F-fluoride PET is a noninvasive and nondestructive way to measure bone metabolism <italic>in vivo</italic>, this molecular imaging modality is not only useful for numerous applications in basic animal science, where pathophysiological bone metabolism is of interest, but also is a valuable tool for use in preclinical arthritis research, where pathological bone destruction and inflammation are the major parameter used for assessment of disease severity.</p></sec><sec><title>Abbreviations</title><p>CT: Computed tomography; ¹⁸&#x02009;F-FDG: ¹⁸&#x02009;F-labelled fluorodeoxyglucose; G6PI: Glucose-6-phosphate isomerase&#x02013;induced arthritis; MT: Metatarsal bone; PET: Positron emission tomography; RA: Rheumatoid arthritis; SUV: Standardized uptake value.</p></sec><sec><title>Competing interests</title><p>The authors declare that they have no competing interests.</p></sec><sec><title>Authors&#x02019; contributions</title><p>II was responsible for study conception and design, animal experiments and data collection, PET data analysis and interpretation and manuscript writing. PG was responsible for study design, animal experiments and data collection, PET data analysis and manuscript writing. TO was responsible for animal experiments and data collection and PET data analysis. BH and MF were responsible for &#x003bc;CT data analysis and interpretation. HS and TK were responsible for data interpretation and critical revision of manuscript. All authors read and approved the final version of the manuscript.</p></sec><sec sec-type="supplementary-material"><title>Supplementary Material</title><supplementary-material content-type="local-data" id="S1"><caption><title>Additional file 1: Figure S1</title><p>^{<bold>18</bold>}<bold>F-fluoride accumulation in healthy mice.</bold> In healthy mice, application of ¹⁸F-fluoride resulted in distinct bone tracer accumulation in the spine, skull, pelvis, pectoral girdle, elbow and knee joints. <bold>(A)</bold> Coregistration of PET and CT data revealing exact sites of ¹⁸F-fluoride accumulation, and PET imaging of ¹⁸F-fluoride signaling 90 minutes after radiotracer injection in a na&#x000ef;ve mouse. <bold>(B)</bold> Coronal and transverse views of the pelvis&#x02013;knee region showing ¹⁸F-fluoride accumulation in trabecular bone. <bold>(C)</bold> Visualization of increased PET signaling in the knee joint (red) by volume rendering and trabecular structure of bone (right).</p></caption><media xlink:href="ar4670-S1.tiff"><caption><p>Click here for file</p></caption></media></supplementary-material></sec>

Cancer/stroma interplay via cyclooxygenase-2 and indoleamine 2,3-dioxygenase promotes breast cancer progression

Could not extract abstract

<contrib contrib-type="author"><name><surname>Chen</surname><given-names>Jing-Yi</given-names></name><address><email>e3336685@yahoo.com.tw</email></address><xref ref-type="aff" rid="Aff1">1</xref></contrib><contrib contrib-type="author"><name><surname>Li</surname><given-names>Chien-Feng</given-names></name><address><email>angelo.p@yahoo.com.tw</email></address><xref ref-type="aff" rid="Aff2">2</xref></contrib><contrib contrib-type="author"><name><surname>Kuo</surname><given-names>Cheng-Chin</given-names></name><address><email>kuocc@nhri.org.tw</email></address><xref ref-type="aff" rid="Aff3">3</xref></contrib><contrib contrib-type="author"><name><surname>Tsai</surname><given-names>Kelvin K</given-names></name><address><email>tsaik@nhri.org.tw</email></address><xref ref-type="aff" rid="Aff1">1</xref></contrib><contrib contrib-type="author"><name><surname>Hou</surname><given-names>Ming-Feng</given-names></name><address><email>mifeho@kmu.edu.tw</email></address><xref ref-type="aff" rid="Aff4">4</xref><xref ref-type="aff" rid="Aff5">5</xref><xref ref-type="aff" rid="Aff6">6</xref></contrib><contrib contrib-type="author" corresp="yes"><name><surname>Hung</surname><given-names>Wen-Chun</given-names></name><address><email>hung1228@nhri.org.tw</email></address><xref ref-type="aff" rid="Aff1">1</xref><xref ref-type="aff" rid="Aff6">6</xref><xref ref-type="aff" rid="Aff7">7</xref></contrib><aff id="Aff1"><label>1</label><institution-wrap><institution-id institution-id-type="ISNI">0000000406229172</institution-id><institution-id institution-id-type="GRID">grid.59784.37</institution-id><institution>National Institute of Cancer Research, </institution><institution>National Health Research Institutes, </institution></institution-wrap>No. 367, Shengli Road, Tainan, 704 Taiwan </aff><aff id="Aff2"><label>2</label><institution-wrap><institution-id institution-id-type="ISNI">0000 0004 0572 9255</institution-id><institution-id institution-id-type="GRID">grid.413876.f</institution-id><institution>Department of Pathology, </institution><institution>Chi-Mei Foundation Medical Center, </institution></institution-wrap>Tainan, 710 Taiwan </aff><aff id="Aff3"><label>3</label><institution-wrap><institution-id institution-id-type="ISNI">0000000406229172</institution-id><institution-id institution-id-type="GRID">grid.59784.37</institution-id><institution>Institute of Cellular and System Medicine, </institution><institution>National Health Research Institutes, </institution></institution-wrap>Maoli, 350 Taiwan </aff><aff id="Aff4"><label>4</label><institution-wrap><institution-id institution-id-type="ISNI">0000 0000 9476 5696</institution-id><institution-id institution-id-type="GRID">grid.412019.f</institution-id><institution>Department of Surgery, </institution><institution>College of Medicine, Kaohsiung Medical University, </institution></institution-wrap>Kaohsiung, 807 Taiwan </aff><aff id="Aff5"><label>5</label><institution-wrap><institution-id institution-id-type="ISNI">0000 0004 0477 6869</institution-id><institution-id institution-id-type="GRID">grid.415007.7</institution-id><institution>Department of Surgery, </institution><institution>Kaohsiung Municipal Ta-Tung Hospital, </institution></institution-wrap>Kaohsiung, 807 Taiwan </aff><aff id="Aff6"><label>6</label><institution-wrap><institution-id institution-id-type="ISNI">0000 0004 0620 9374</institution-id><institution-id institution-id-type="GRID">grid.412027.2</institution-id><institution>Cancer Center, </institution><institution>Kaohsiung Medical University Hospital, </institution></institution-wrap>Kaohsiung, 807 Taiwan </aff><aff id="Aff7"><label>7</label><institution-wrap><institution-id institution-id-type="ISNI">0000 0000 9476 5696</institution-id><institution-id institution-id-type="GRID">grid.412019.f</institution-id><institution>Graduate Institute of Medicine, </institution><institution>College of Medicine, Kaohsiung Medical University, </institution></institution-wrap>Kaohsiung, 807 Taiwan </aff>

Breast Cancer Research : BCR

<sec id="Sec1"><title>Background</title><p>Chronic inflammation is strongly associated with the development of cancer [<xref ref-type="bibr" rid="CR1">1</xref>]-&#x02013;[<xref ref-type="bibr" rid="CR3">3</xref>]. One of the crucial mediators of inflammatory reaction is cyclooxygenase (COX). The COX family of enzymes comprises two members (COX-1 and COX-2) and is the main controller of eicosanoid biosynthesis. Studies of human breast tumor tissues demonstrate that upregulation of COX-2 has been detected in approximately 40% of human breast tumor tissues, as well as preinvasive ductal carcinoma <italic>in situ</italic> lesions [<xref ref-type="bibr" rid="CR4">4</xref>]. Elevated expression of COX-2 is associated with large tumor size, advanced histologic grade, axillary node metastasis, and unfavorable disease-free survival [<xref ref-type="bibr" rid="CR4">4</xref>],[<xref ref-type="bibr" rid="CR5">5</xref>]. In addition, COX-2 expression also links with increased tumor angiogenesis [<xref ref-type="bibr" rid="CR6">6</xref>]. Epidemiologic investigations suggest that use of nonsteroidal antiinflammatory drugs or selective COX-2 inhibitors reduces breast cancer risk [<xref ref-type="bibr" rid="CR7">7</xref>],[<xref ref-type="bibr" rid="CR8">8</xref>].</p><p>Results of animal study also support an oncogenic role of COX-2. Transgenic COX-2 overexpression induces mammary tumor formation in mice [<xref ref-type="bibr" rid="CR9">9</xref>]. This tumorigenic transformation is highly dependent on PGE₂ production and angiogenic switch. In addition, <italic>HER-2/Neu</italic> oncogene-induced mammary tumorigenesis and angiogenesis are dramatically attenuated in COX-2 knockout mice, suggesting a key role of COX-2 in breast cancer [<xref ref-type="bibr" rid="CR10">10</xref>]. Recent studies also show that COX-2 inhibitors exhibit antitumor and antiangiogenic activities <italic>in vivo</italic> and exhibit chemopreventive effects against mammary carcinogenesis induced by 7,12-dimethyl-benz(a)anthracene in rats [<xref ref-type="bibr" rid="CR11">11</xref>]. All of the results suggest that COX-2 is involved in multiple steps of breast carcinogenesis and is a potential target for cancer prevention and therapy.</p><p>Interplay between breast cancer cells and cancer-associated fibroblasts (CAFs), the most abundant and active stromal cells, is crucial for tumor growth, progression, angiogenesis, and therapeutic resistance [<xref ref-type="bibr" rid="CR12">12</xref>]. Cancer cells release a number of factors to enhance the production of cytokines, chemokines, and matrix metalloproteinases (MMPs) from CAFs, which in turn facilitate cancer cell proliferation, migration, and invasion. Previous study demonstrated that stromal fibroblasts present in invasive breast carcinomas can secrete large amounts of stromal cell-derived factor 1 (SDF-1) to enhance tumor growth and angiogenesis [<xref ref-type="bibr" rid="CR13">13</xref>]. Co-injection of breast cancer cells and fibroblasts also promotes the progression of ductal carcinoma <italic>in situ</italic> to invasive breast carcinoma by stimulating chemokine (C-X-C motif) ligand 14 (CXCL14) and chemokine (C-X-C motif) ligand 12 (CXCL12) production [<xref ref-type="bibr" rid="CR14">14</xref>]. However, most studies addressing the crosstalk between cancer and stromal cells focus on protein factors like cytokines and chemokines. Whether other small molecules such as lipids or metabolites participate in cancer-stromal cell interaction is largely unknown.</p><p>The tumor-promoting role of CAFs via upregulation of COX-2 in ductal carcinoma <italic>in situ</italic> of the breast was first demonstrated by Hu <italic>et al</italic>. [<xref ref-type="bibr" rid="CR15">15</xref>]. The authors showed that co-culture with fibroblasts increases COX-2 expression in breast cancer cells and subsequently induces MMP-9 and MMP-14 in these cells to promote invasion. They also elucidated the underlying mechanism by demonstrating that inhibition of nuclear factor kappa-light-chain-enhancer of activated B cells (NF-&#x003ba;B) and COX-2 activity reduces the invasion-promoting effect of fibroblasts. These data suggest that fibroblasts secrete some factors to activate NF-&#x003ba;B-mediated transcription of COX-2 in breast cancer cells to enhance tumor progression.</p><p>However, several issues remain elusive. First, does PGE₂ generated by COX-2-expressing cancer cells also affect gene expression or behavior of stromal fibroblasts? Second, do CAFs secrete small molecules (other than proteins or peptides) to regulate cancer cell invasion? Finally, can the importance of cancer-stroma interaction in cancer progression be validated in clinical samples?</p><p>In this study, we address these questions and try to clarify the underlying mechanism.</p></sec><sec id="Sec2"><title>Methods</title><sec id="Sec3"><title>Cell culture</title><p>Human breast cancer cell lines MCF-7 and MDA-MB-231 were purchased from the Bioresource Collection and Research Center (BCRC) and ATCC. Immortalized human breast fibroblasts, RMF-EG [<xref ref-type="bibr" rid="CR16">16</xref>], were kindly provided by Dr. Charlotte Kuperwasser (Tufts University, Boston, MA, USA). These cells were cultured in DMEM/F12 containing 10% fetal bovine serum (FBS). Other experimental materials and procedures are provided in Additional file <xref rid="MOESM1" ref-type="media">1</xref>.</p></sec><sec id="Sec4"><title>Establishment of inducible COX-2-expression MCF-7 cell line</title><p>To establish an inducible COX-2-expression cell line, MCF-7 cells (1&#x02009;&#x000d7;&#x02009;10⁶) were resuspended in buffer R containing 2&#x000a0;&#x003bc;g pCMV-Tet3G plasmid. Transfection was performed by using Neon microporation transfection system at room temperature with 1,250&#x000a0;V, 20 milliseconds, and two pulses. After 48&#x000a0;hours, the cells were selected with 1&#x000a0;mg/ml&#x000a0;G418 for 2&#x000a0;weeks.</p><p>For the delivery of the second plasmid, pCMV-Tet3G stably transfected cells (1&#x02009;&#x000d7;&#x02009;10⁶) were resuspended in buffer R containing 2&#x000a0;&#x003bc;g of pTRE-mCherry-COX-2 plasmid. Transfection was performed by using Neon microporation transfection system at room temperature with 1,250&#x000a0;V, 20 milliseconds, and two pulses. After 48&#x000a0;hours, the cells were subjected to selection with 100&#x000a0;&#x003bc;g/ml hygromycin B. The stable cell line harbors both pCMV-Tet3G and pTRE-mCherry-COX-2 plasmid was used for induction experiment. The cells were maintained at 37&#x000b0;C in a 5% CO₂-humidified atmosphere and were incubated with doxycycline to induce COX-2 expression before co-culture assay.</p></sec><sec id="Sec5"><title>Co-culture assay</title><p>In the co-culture system, 1&#x02009;&#x000d7;&#x02009;10⁵ RMF-EG cells were grown in the bottom of a six-well plate in DMEM/F12 with 10% FBS, and 1&#x02009;&#x000d7;&#x02009;10⁶ breast cancer cells were seeded on the 0.4-&#x003bc;m polyester membrane of a transwell insert in the same medium. MCF-7 cells were treated with or without doxycycline (1&#x000a0;&#x003bc;g/ml) for 72&#x000a0;hours. The conditioned medium, breast cancer cells, and RMF-EG cells were harvested for metabolomics and Western blotting analysis.</p></sec><sec id="Sec6"><title>Metabolite profiling</title><p>The proteins in the conditioned medium were removed by using 3-kDa ultracentrifugation filter devices. The metabolites in the filtered medium were extracted by using iced 50% methanol and were subsequently dried by a speedvac. Metabolite profiles were analyzed with the Metabolomics Core of National Health Research Institutes by using a high-resolution ultraperformance liquid chromatography (UPLC) coupled online to a triple-quadrupole time-of-flight mass spectrometry system, as described previously [<xref ref-type="bibr" rid="CR17">17</xref>]. Metabolite identity was predicted with Human Metabolome Database [<xref ref-type="bibr" rid="CR18">18</xref>].</p></sec><sec id="Sec7"><title>RNA extraction and quantitative reverse transcription-PCR analysis</title><p>Total RNA was isolated from cells by using an RNA extraction kit (Qiagen, Valencia, CA, USA) and 1&#x000a0;&#x003bc;g of RNA was reverse-transcribed to cDNA. Target mRNAs were quantified by using real-time PCR reactions with SYBR green fluorescein, and actin served as an internal control. cDNA synthesis was performed at 95&#x000b0;C for 3&#x000a0;minutes, and the conditions for PCR were 28&#x000a0;cycles of denaturation (95&#x000b0;C/1&#x000a0;minute), annealing (60&#x000b0;C/1&#x000a0;minute) extension (72&#x000b0;C/1&#x000a0;minute), and 1&#x000a0;cycle of final extension (72&#x000b0;C/10&#x000a0;minutes). The primers used are tryptophan 2,3-dioxygenase (TDO)-forward: 5&#x02032;-GGGAACTACCTGCATTTGGA-3&#x02032;; TDO-reverse: 5&#x02032;-GTGCATCCGAGAAACAACCT-3&#x02032;; IDO-forward: 5&#x02032;-GCGCTGTTGGAAATAGCTTC-3&#x02032;; IDO-reverse: 5&#x02032;-CAGGACGTCAAAGCACTGAA-3&#x02032;; E-cadherin-forward: 5&#x02032;-CCTGGGACTCCACCTACAGA-3&#x02032;; E-cadherin-reverse: 5&#x02032;-GGATGAACACAGCGTGAGAGA-3&#x02032;; actin-forward: 5&#x02032;-TGTTACCAACTGGGACGACA-3&#x02032;; actin-reverse: 5&#x02032;-GGGGTGTTGAAGGTCTCAAA-3&#x02032;.</p></sec><sec id="Sec8"><title>Immunoprecipitation and Western blot analysis</title><p>MCF-7or COX-2-overexpressing MCF-7 cells were treated with or without 100&#x000a0;&#x003bc;<italic>M</italic> kynurenine for 24&#x000a0;hours; the cells were harvested with an RIPA buffer (50&#x000a0;m<italic>M</italic> Tris&#x02013;HCl, pH&#x000a0;7.4, 150&#x000a0;m<italic>M</italic> NaCl, 1% NP-40, 0.1% SDS, 0.5% sodium deoxycholate, 2&#x000a0;m<italic>M</italic> EDTA, and 50&#x000a0;m<italic>M</italic> NaF), and cellular lysates were incubated with anti-AhR antibody overnight at 4&#x000b0;C with rotation. Immunocomplexes were pulled down by Protein-G agarose bead, washed with RIPA buffer 3 times, and eluted with a sample buffer in boiled water for 10&#x000a0;minutes. The eluted samples were subjected to SDS-PAGE separation, and proteins were transferred to nitrocellulose membranes. Finally, the blots were probed with anti-E-cadherin or anti-Skp2 antibody and developed with enhanced chemiluminescence reagent.</p></sec><sec id="Sec9"><title>Migration assay</title><p>Migration assays were conducted in transwells with 8-&#x003bc;m-pore filter inserts. Then 1&#x02009;&#x000d7;&#x02009;10⁴ MCF-7 or COX-2-overexpressing MCF-7 cells were seeded in the upper chamber. The lower chambers were filled with DMEM medium containing 1% FBS and 100&#x000a0;&#x003bc;<italic>M</italic> kynurenine. After 24&#x000a0;hours, the cells on the upper surface were removed by wiping with a cotton swab, and the cells that migrated to the lower surface were fixed. The cells were stained with 4&#x02032;,6-diamidino-2-phenylindole (DAPI), and the cell number in 15 randomly selected fields was counted under a microscope (100&#x000d7;). Experiments were performed independently at least 3 times.</p></sec><sec id="Sec10"><title>Protein ubiquitination assay</title><p>MCF-7 cells treated with or without kynurenine were incubated with the proteasome inhibitor MG132 or the lysosome inhibitor chloroquine. The cells were harvested with a lysis buffer (20&#x000a0;m<italic>M</italic> Tris&#x02013;HCl at pH&#x000a0;7.5, 150&#x000a0;m<italic>M</italic> sodium chloride, 1&#x000a0;m<italic>M</italic> calcium chloride, and 1% Triton X-100 and protease inhibitors), and cellular lysates were incubated with an E-cadherin antibody overnight at 4&#x000b0;C with rotation. Protein-G beads were added to the samples and incubated for another 1&#x000a0;hour at 4&#x000b0;C. Immunocomplexes were eluted and were subjected to SDS-PAGE separation, and proteins were transferred to nitrocellulose membranes. Finally, the blots were probed by using an anti-ubiquitin antibody to detect the ubiquitination status of E-cadherin.</p></sec><sec id="Sec11"><title>Immunofluorescent staining and confocal microscopy</title><p>MCF-7 cells were treated with or without 100&#x000a0;&#x003bc;<italic>M</italic> kynurenine for 6&#x000a0;hours and fixed with 3.7% formaldehyde for 15&#x000a0;minutes at room temperature. Cells were washed twice with PBS and permeabilized by 0.1% Triton X-100 solution for 10&#x000a0;minutes. After permeabilization, cells were incubated with 0.05% BSA solution to block nonspecific binding. Anti-AhR mouse monoclonal antibody (1:80) or anti-E-cadherin goat polyclonal antibody (1:250) was added and incubated at room temperature for 1&#x000a0;hour. After extensive washing, Alexas Fluro 594 anti-mouse IgG or Alexas Fluro 488 anti-goat IgG was added and incubated for another 1&#x000a0;hour. Cell nuclei were stained with DAPI solution. Finally, coverslips were washed twice with PBS and subsequently placed in mounting solution. The fluorescent image was observed under a confocal microscope.</p></sec><sec id="Sec12"><title><italic>In vivo</italic>orthotopic animal study</title><p>MCF-7 or MCF-7-COX2 (8&#x02009;&#x000d7;&#x02009;10⁶) cells were mixed with RMF-EG (6&#x02009;&#x000d7;&#x02009;10⁶) cells in Hanks balanced salt solution and Matrigel (BD Transduction Laboratories, San Jose, CA, USA). Cells were inoculated into the fourth mammary fat pads of 6-week-old female BALB/cAnN.Cg-Foxn1nu/CrlNarl mice. Before the inoculation of the cancer cell/fibroblast mixture, all mice were primed with 6&#x000a0;mg/kg of 17&#x003b2;-estradiol twice a week for 3&#x000a0;weeks.</p><p>After inoculation, 17&#x003b2;-estradiol was continuously given to mice throughout the experiments. Measurement of tumor growth was begun at 4&#x000a0;weeks after injection, and tumor volume was calculated by using the equation: tumor volume&#x02009;=&#x02009;(length&#x02009;&#x000d7;&#x02009;width²)/2. After 10&#x000a0;weeks, mice injected with COX-2-overexpressing MCF-7 and RMF-EG produced tumors with volumes approximately 200&#x000a0;mm³ and were randomly divided into four groups that received vehicle (DMSO), NS-398 (10&#x000a0;mg/kg), L-1-methy-tryptophan (10&#x000a0;mg/kg), or both inhibitors 5 times per week.</p><p>Two weeks later, animals were killed, and tumors were isolated from mice. The statistical difference between experimental groups was evaluated with repeated-measures two-way ANOVA analysis. The animal-use protocol was approved by the Institutional Animal Care and Use Committee of National Health Research Institutes.</p></sec><sec id="Sec13"><title>Patients and statistical analysis</title><p>Paraffin-embedded human breast tumor tissues were obtained from Chi-Mei Medical Center (Tainan, Taiwan) between 1998 and 2004. The slides were stained with anti-COX-2 or anti-IDO antibodies. The COX-2 and IDO stainings were interpreted by using the H-score, defined by the following equation: H-score&#x02009;=&#x02009;&#x003a3;Pi (i&#x02009;+&#x02009;1), as previously described [<xref ref-type="bibr" rid="CR19">19</xref>], where i is the intensity of the stained tumor cells (0 to 3+), and Pi is the percentage of stained tumor cells with various intensities. We classified tumors with cancer cells and stromal cells showing H-scores no less than the median of all scored cases as having high COX-2 and IDO expression, respectively.</p><p>The follow-up duration ranged from 5.4 to 143.6&#x000a0;months, with a mean of 87.1&#x000a0;months. Survival analyses for disease-specific and metastasis-free survival were performed by using Kaplan-Meier plots and compared by using the log-rank test. The correlation between COX-2 and IDO expression with clinicopathologic parameters was examined with &#x003c7;² test. <italic>P</italic> value&#x02009;&#x0003c;&#x02009;0.05 was considered statistically significant. This study was approved by the Research Ethics Committee of National Health Research Institutes. Written informed consent was obtained from all patients participating in this study.</p></sec></sec><sec id="Sec14"><title>Results</title><sec id="Sec15"><title>COX-2-overexpressing breast cancer cells upregulated IDO expression in co-cultured fibroblasts</title><p>We analyzed the metabolite profile of the supernatant of RMF-EG human breast fibroblasts co-cultured with MCF-76 or COX-2-overexpressing MCF-7 cells and found that several metabolites were increased in the supernatant of COX-2-overexpressing MCF-7/RMF-EG co-culture. A peak with the m/z ratio of 209 was increased about twofold (Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>A). By using Human Metabolome Database search, a candidate metabolite was predicted to be kynurenine. UPLC/MS/MS analysis demonstrated that fragmentation of kynurenine standard yielded three peaks with m/z ratio of 209, 192, and 146, which is consistent with the reported data (accession: K0009019, MassBank, [<xref ref-type="bibr" rid="CR20">20</xref>]) (Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>B). Significant increase of kynurenine was confirmed with an ELISA assay (Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>C).<fig id="Fig1"><label>Figure 1</label><caption><p><bold>Increase of IDO expression and kynurenine production in fibroblasts co-cultured with COX-2-overexpressing breast cancer cells. (A)</bold> Metabolite profiling of the supernatant of RMF-EG fibroblasts co-cultured with MCF-7 (RMF-M) and COX-2-overexpressing MCF-7 (RMF-COX/M) cells and identified a peak with m/z ratio of 209 was increased. <bold>(B)</bold> UPLC/MS/MS fragmentation profile of the 209 (m/z) peak and the standard (L-kynurenine). <bold>(C)</bold> Increase of kynurenine in RMF-COX/M cells determined with an ELISA assay. The results from three independent assays were expressed as mean&#x02009;&#x000b1;&#x02009;SEM. Statistical significance was evaluated with the Student <italic>t</italic> test. *<italic>P</italic>&#x02009;&#x0003c;&#x02009;0.05. <bold>(D)</bold> Upregulation of IDO but not TDO in RMF-COX/M cells was assayed with quantitative RT-PCR. The results from three independent assays were expressed as mean&#x02009;&#x000b1;&#x02009;SEM. Statistical significance was evaluated with Student <italic>t</italic> test (i). *<italic>P</italic>&#x02009;&#x0003c;&#x02009;0.05. Protein level was also determined with Western blotting (ii). <bold>(E)</bold> MCF-7 cells were treated without (C) or with doxycycline (DOX, 1&#x000a0;&#x003bc;g/ml) for 72&#x000a0;hours to induce COX-2 expression. Protein level of COX-2 and IDO was studied with Western blotting. A 3.6-fold increase of COX-2 was found, whereas the expression of IDO was not detectable. <bold>(F)</bold> Protein level of COX-2 and IDO in MCF-7 and MDA-MB-231 cells was compared. In addition, IDO expression of RMF-EG cells co-cultured with MCF-7 or MDA-MB-231 cells was investigated.</p></caption><graphic xlink:href="13058_2014_410_Fig1_HTML" id="d29e806"/></fig></p><p>The rate-limiting enzymes in the generation of kynurenine are indoleamine 2,3-dioxygenase (IDO) and tryptophan 2,3-dioxygenase (TDO). We found a 2.5-fold of increase of IDO mRNA in RMF-EG cells co-cultured with COX-2-overexpressing MCF-7 cells, whereas the expression of TDO was not changed (Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>Di). A similar increase of IDO protein level was also found (Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>Dii). IDO was very low or undetectable in MCF-7- and COX-2-overexpressing MCF-7 cells, indicating that the kynurenine in the co-cultured medium was produced mainly by RMF-EG cells (Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>E). Co-culture of the COX-2-overexpressing MDA-MB-231 cells also upregulated IDO expression in RMF-EG cells (Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>F). These data suggest that COX-2-overexpressing breast cancer cells stimulate IDO expression and increase kynurenine secretion in co-cultured fibroblasts.</p></sec><sec id="Sec16"><title>PGE₂transcriptionally elevated IDO expression in RMF-EG fibroblasts through the EP4/STAT3 signaling pathway</title><p>We found that PGE₂ increased IDO mRNA and protein levels in RMF-EG cells (Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>Ai and <xref rid="Fig2" ref-type="fig">2</xref>Aii). In addition, our data showed that only PGE₂-alcohol (an EP4 agonist) significantly upregulated IDO expression (Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>B). Knockdown of EP4 abolished PGE₂-induced increase of IDO in these cells (Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>Ci and <xref rid="Fig2" ref-type="fig">2</xref>Cii). By using different <italic>IDO</italic> deletion promoters, we demonstrated that PGE₂ stimulated <italic>IDO</italic> transcription via the &#x02212;1,140/-844 region from the transcription start site (see Additional file <xref rid="MOESM2" ref-type="media">2</xref>: Figure S1). This region contained two potential &#x003b3;-interferon-activated sites (GASs) that could be activated by different signal transducer and activator of transcription (STAT) proteins [<xref ref-type="bibr" rid="CR21">21</xref>],[<xref ref-type="bibr" rid="CR22">22</xref>]. Both STAT1 and STAT3 have been implicated in the regulation of IDO expression [<xref ref-type="bibr" rid="CR23">23</xref>],[<xref ref-type="bibr" rid="CR24">24</xref>].<fig id="Fig2"><label>Figure 2</label><caption><p><bold>PGE</bold>_{<bold>2</bold>}<bold>upregulated IDO expression in fibroblasts via the EP4/STAT3 pathway. (A)</bold> RMF-EG cells were treated with DMSO or PGE₂ (2&#x000a0;&#x003bc;<italic>M</italic>) in 1% FCS medium for 48&#x000a0;hours. IDO mRNA (i) and protein (ii) were determined by quantitative RT-PCR and Western blotting. *<italic>P</italic>&#x02009;&#x0003c;&#x02009;0.05. <bold>(B)</bold> RMF-EG cells were treated with 17-phenyl-trinor-PGE₂ (an EP1 and EP3 receptor agonist), butaprost (an EP2 agonist), or PGE₂-alcohol (an EP4 agonist) for 48&#x000a0;hours, and IDO expression was studied with quantitative RT-PCR. *<italic>P</italic>&#x02009;&#x0003c;&#x02009;0.05. <bold>(C)</bold> RMF-EG cells were pretreated with EP4 shRNA for 24&#x000a0;hours and then cultured with MCF-7 (RMF-M) and COX-2-overexpressing MCF-7 (RMF-COX/M) cells for another 48&#x000a0;hours. (i) The IDO mRNA level was determined with quantitative RT-PCR. (ii) Protein level of IDO and EP4 was studied with Western blotting. *<italic>P</italic>&#x02009;&#x0003c;&#x02009;0.05. <bold>(D)</bold> RMF-EG cells were pretreated with STAT3 siRNA for 24&#x000a0;hours and then cultured with PGE₂-alcohol for another 48&#x000a0;hours. Protein levels of STAT3 and IDO were determined. <bold>(E)</bold> RMF-EG cells were transfected with pcDNA or STAT3 expression vector for 48&#x000a0;hours. Expression of STAT3 and IDO was studied with Western blotting.</p></caption><graphic xlink:href="13058_2014_410_Fig2_HTML" id="d29e932"/></fig></p><p>We performed a ChIP assay and found that the binding of STAT3 to <italic>IDO</italic> promoter was increased in RMF-EG cells co-cultured with COX-2-overexpressing MCF-7 cells, whereas the binding of STAT1 was decreased (see Additional file <xref rid="MOESM3" ref-type="media">3</xref>: Figure S2). Knockdown of STAT3 abolished the increase of IDO induced by the EP4 agonist in RMF-EG cells (Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>D). Ectopic expression of STAT3 upregulated IDO (2.8-fold) in these cells (Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>E). Thus, COX-2-overexpressing breast cancer cells upregulate IDO expression in fibroblasts through the PGE₂/EP4/STAT3 pathway.</p></sec><sec id="Sec17"><title>IDO-expressing fibroblasts enhanced the migration of breast cancer cells through downregulation of E-cadherin</title><p>We next studied the effect of kynurenine on breast cancer cells. Kynurenine did not affect the proliferation of MCF-7 cells (Figure&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>A). However, kynurenine significantly increased the motility of MCF-7 and COX-2-overexpressing MCF-7 cells (Figure&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>B). The conditioned medium of RMF-EG fibroblasts preincubated with COX-2-overexpressing MCF-7 cells also increased the motility of MCF-7 cells (Figure&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>C). We used 1-methyl-L-tryptophan to inhibit IDO activity in RMF-EG fibroblasts induced by co-culture with COX2-overexpressing MCF-7 cells and found that the stimulatory effect on cell motility was blocked (Figure&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>C). These data suggested that kynurenine released by IDO-expressing fibroblasts enhanced the migration of breast cancer cells.<fig id="Fig3"><label>Figure 3</label><caption><p><bold>Kynurenine induced E-cadherin ubiquitination and degradation and increased migration of breast cancer cells. (A)</bold> MCF-7 cells were treated with different concentrations of kynurenine for 48&#x000a0;hours, and cellular proliferation was investigated with MTT assay. <bold>(B)</bold> MCF-7- or COX-2-overexpressing MCF-7 cells were treated with 100&#x000a0;&#x003bc;<italic>M</italic> kynurenine, and cell migration was studied with transwell assays. The results from three independent assays were expressed as mean&#x02009;&#x000b1;&#x02009;SEM. Statistical significance was evaluated with Student <italic>t</italic> test. *<italic>P</italic>&#x02009;&#x0003c;&#x02009;0.05. <bold>(C)</bold> RMF-EG cells were cultured in the absence or presence of IDO inhibitor 1-methyl-L-tryptophan (L-1-MT) in the lower well of the transwell unit. MCF-7- or COX-2-overexpressing MCF-7 cells were seeded in the upper well. After 24&#x000a0;hours, migrated cell number was determined. *<italic>P</italic>&#x02009;&#x0003c;&#x02009;0.05. <bold>(D)</bold> MCF-7 cells were incubated without (&#x02212;) or with (+) 100&#x000a0;&#x003bc;<italic>M</italic> kynurenine for different times. Protein (i) and mRNA (ii) levels of E-cadherin were studied. <bold>(E)</bold> MCF-7 cells were incubated with kynurenine (100&#x000a0;&#x003bc;<italic>M</italic>) and MG132 (proteasome inhibitor, 10&#x000a0;&#x003bc;<italic>M</italic>) or chloroquine (CQ, lysosome inhibitor, 25&#x000a0;&#x003bc;<italic>M</italic>) for 24&#x000a0;hours. Protein level of E-cadherin was studied with Western blotting and normalized to actin. <bold>(F)</bold> Ubiquitination of E-cadherin was studied with immunoprecipitation of E-cadherin by specific antibody, and the ubiquitination status was detected with anti-ubiquitin antibody.</p></caption><graphic xlink:href="13058_2014_410_Fig3_HTML" id="d29e1018"/></fig></p><p>We investigated the expression of epithelial-to-mesenchymal markers in kynurenine-treated breast cancer cells and found that E-cadherin was reduced in a time-dependent manner (Figure&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>Di). E-cadherin began to decrease around 8&#x000a0;hours after addition of kynurenine, and a 70% of reduction was found at 24&#x000a0;hours. However, its mRNA did not decrease substantially (Figure&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>Dii). We found that kynurenine induced degradation of E-cadherin protein via a proteasome-dependent pathway, which could be rescued by MG132 (proteasome inhibitor) but not chloroquine (lysosome inhibitor) (Figure&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>E). In addition, ubiquitination of E-cadherin protein was increased in kynurenine-treated MCF-7 cells (Figure&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>F). These data suggest that kynurenine induces ubiquitination and degradation of E-cadherin to promote breast cancer cell motility.</p></sec><sec id="Sec18"><title>Kynurenine increased the degradation of E-cadherin in an AhR- and Skp2-dependent manner</title><p>Kynurenine has been shown to be an endogenous tumor-promoting ligand of the human AhR [<xref ref-type="bibr" rid="CR25">25</xref>]. In addition, AhR is involved in the degradation of sex steroid receptors via a cullin 4B-dependent ubiquitination pathway [<xref ref-type="bibr" rid="CR26">26</xref>]. We tested whether kynurenine reduced protein stability of E-cadherin through activation of AhR and found that the binding between AhR and E-cadherin was increased in kynurenine-treated MCF-7 cells (Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>A). Interestingly, Skp2, an F-box protein of the SCF E3 ligase, was co-immunoprecipitated with AhR, and the interaction was also increased by kynurenine. We did not detect the cullin 4B protein in the complex (data not shown). This is not a cell line-specific effect, because the interaction between AhR and E-cadherin protein was also elevated in kynurenine-treated A549 cells (see Additional file <xref rid="MOESM4" ref-type="media">4</xref>: Figure S3). The 3&#x02032;-methylcholanthrene (3-MC), another AhR ligand, also induced co-localization of AhR and E-cadherin at the cell membrane (Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>B). Knockdown of Skp2 reversed kynurenine-induced reduction of E-cadherin protein without affecting AhR expression (Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>C). The AhR antagonist, 3&#x02032;4&#x02032;-dimethoxyflavone (3&#x02032;4&#x02032;-DMF), also inhibited the decrease of E-cadherin induced by kynurenine (Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>D). Additionally, kynurenine-increased migration of MCF-7 cells was blocked by 3&#x02032;4&#x02032;-DMF (Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>E). These data suggest that kynurenine induces the formation of E-cadherin/AhR/Skp2 complex and causes E-cadherin degradation.<fig id="Fig4"><label>Figure 4</label><caption><p><bold>Kynurenine induced E-cadherin degradation in an AhR- and Skp2-dependent pathway. (A)</bold> MCF-7 cells were incubated with kynurenine for 6&#x000a0;hours, and cellular proteins were harvested. AhR protein was immunoprecipitated by specific antibody, and the binding of E-cadherin and Skp2 was studied with Western blotting. <bold>(B)</bold> Cells were incubated with kynurenine or 3-methylcholanthrene (an AhR agonist) for 6&#x000a0;hours, and the co-localization of E-cadherin and AhR was studied with confocal microscopy. <bold>(C)</bold> Cells were treated with Skp2 shRNA for 24&#x000a0;hours and then incubated with kynurenine for another 48&#x000a0;hours. The protein levels of Skp2, E-cadherin, and AhR were determined with Western blotting and normalized to actin level. <bold>(D)</bold> Cells were incubated with kynurenine, 3&#x02032;4&#x02032;-dimethoxyflavone (3&#x02032;4&#x02032;-DMF, an AhR antagonist) or both drugs for 24&#x000a0;hours. E-cadherin protein level was investigated with Western blotting. <bold>(E)</bold> Cells were also collected and subjected to migration assay. *<italic>P</italic>&#x02009;&#x0003c;&#x02009;0.05.</p></caption><graphic xlink:href="13058_2014_410_Fig4_HTML" id="d29e1088"/></fig></p></sec><sec id="Sec19"><title>COX-2 expression in breast cancer and IDO expression in stromal fibroblasts predicted poor disease-specific and metastasis-free survival</title><p>The correlation between COX-2 expression in tumor tissues and IDO expression in CAFs was confirmed by two approaches. First, we isolated CAFs from two breast tumor tissues without or with COX-2 overexpression and found that IDO expression in CAFs was upregulated in COX-2-overexpressing tumor (see Additional file <xref rid="MOESM5" ref-type="media">5</xref>: Figure S4). Second, we used immunohistochemical analysis to detect COX-2 and IDO expression in a cohort of breast cancer tissues (Figure&#x000a0;<xref rid="Fig5" ref-type="fig">5</xref>A). COX-2 expression in tumors was positively correlated with a high IDO expression in CAFs (65 of 101, 64%; <italic>P</italic>&#x02009;&#x0003c;&#x02009;0.001) (Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>). COX-2 was highly expressed in stage III (19 of 24, 79%; <italic>P</italic>&#x02009;&#x0003c;&#x02009;0.05), N1-N2 (61 of 85, 71%; <italic>P</italic>&#x02009;&#x0003c;&#x02009;0.001), and T3-4 stage (20 of24, 83%; <italic>P</italic>&#x02009;&#x0003c;&#x02009;0.05) tumor specimens. IDO expression in CAFs was significantly expressed in stage III (20 of 24, 83%; <italic>P</italic>&#x02009;&#x0003c;&#x02009;0.05), N1-N2 (60 of 85, 71%; <italic>P</italic>&#x02009;&#x0003c;&#x02009;0.001), and T3-4 stage (21 of 24, 88%; <italic>P</italic>&#x02009;&#x0003c;&#x02009;0.001) tumor specimens. The disease-specific and metastasis-free survival declined significantly in patients with high COX-2 expression in breast tumors (<italic>P</italic>&#x02009;=&#x02009;0.0043 and <italic>P</italic>&#x02009;&#x0003c;&#x02009;0.0001, respectively) (Figure&#x000a0;<xref rid="Fig5" ref-type="fig">5</xref>B and Table&#x000a0;<xref rid="Tab2" ref-type="table">2</xref>). Similarly, the disease-specific and metastasis-free survival declined significantly in patients with high IDO expression in CAFs (<italic>P</italic>&#x02009;=&#x02009;0.0045 and <italic>P</italic>&#x02009;&#x0003c;&#x02009;0.0001) (Figure&#x000a0;<xref rid="Fig5" ref-type="fig">5</xref>C and Table&#x000a0;<xref rid="Tab2" ref-type="table">2</xref>). More important, high COX-2 in tumors and high IDO1 expression in CAFs predicted worse disease-free and metastasis-free survivals in breast cancer patients (<italic>P</italic>&#x02009;&#x0003c;&#x02009;0.0001, Figure&#x000a0;<xref rid="Fig5" ref-type="fig">5</xref>D).<fig id="Fig5"><label>Figure 5</label><caption><p><bold>Coexpression of cancer COX-2 and stromal IDO predicted worse patient survival. (A)</bold> Immunohistochemical staining showed COX-2 expression in breast tumor tissues and IDO expression in tumor stroma. <bold>(B)</bold> High COX-2 expression in tumor tissues was associated with reduced disease-specific and metastasis-free survival. <bold>(C)</bold> High IDO expression in tumor stroma also was associated with reduced disease-specific and metastasis-free survival. <bold>(D)</bold> Coexpression of tumor COX-2 and stroma IDO predicted worse patient survival.</p></caption><graphic xlink:href="13058_2014_410_Fig5_HTML" id="d29e1176"/></fig></p><table-wrap id="Tab1"><label>Table 1</label><caption><p> <bold>Correlation between Cox-2 and Ido-1 expression and various clinicopathologic factors</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr><th rowspan="2">Parameters</th><th rowspan="2">Category</th><th rowspan="2">No. of cases</th><th colspan="2">Cox-2 expression (tumor)</th><th rowspan="2"><bold><italic>P</italic></bold>value</th><th colspan="2">Ido-1 expression (CAF)</th><th rowspan="2"><bold><italic>P</italic></bold>value</th></tr><tr><th>Low exp.</th><th>High exp.</th><th>Low exp.</th><th>High exp.</th></tr></thead><tbody><tr><td>Age (years)</td><td>&#x0003c;60&#x000a0;years</td><td>141</td><td>73</td><td>68</td><td>0.444</td><td>69</td><td>72</td><td>0.647</td></tr><tr><td/><td>&#x02267;60&#x000a0;years</td><td>61</td><td>28</td><td>33</td><td/><td>32</td><td>29</td><td/></tr><tr><td>Primary tumor (T)</td><td>T1</td><td>82</td><td>45</td><td>37</td><td> <bold>0.002</bold> </td><td>44</td><td>38</td><td> <bold>&#x0003c;0.001</bold> </td></tr><tr><td/><td>T2</td><td>96</td><td>52</td><td>44</td><td/><td>54</td><td>42</td><td/></tr><tr><td/><td>T3-4</td><td>24</td><td>4</td><td>20</td><td/><td>3</td><td>21</td><td/></tr><tr><td>Nodal status (N)</td><td>N0</td><td>117</td><td>77</td><td>40</td><td> <bold>&#x0003c;0.001</bold> </td><td>76</td><td>41</td><td> <bold>&#x0003c;0.001</bold> </td></tr><tr><td/><td>N1-N2</td><td>85</td><td>24</td><td>61</td><td/><td>25</td><td>60</td><td/></tr><tr><td>Stage</td><td>I</td><td>63</td><td>37</td><td>26</td><td> <bold>0.006</bold> </td><td>37</td><td>26</td><td> <bold>0.002</bold> </td></tr><tr><td/><td>II</td><td>115</td><td>59</td><td>56</td><td/><td>60</td><td>55</td><td/></tr><tr><td/><td>III</td><td>24</td><td>5</td><td>19</td><td/><td>4</td><td>20</td><td/></tr><tr><td>Histologic grade</td><td>Grade I</td><td>18</td><td>13</td><td>5</td><td>0.122</td><td>15</td><td>3</td><td> <bold>0.003</bold> </td></tr><tr><td/><td>Grade II</td><td>141</td><td>69</td><td>72</td><td/><td>71</td><td>70</td><td/></tr><tr><td/><td>Grade III</td><td>43</td><td>19</td><td>24</td><td/><td>15</td><td>28</td><td/></tr><tr><td>CAF Ido-1 expression</td><td>Low Exp. (&#x0003c;medium)</td><td>101</td><td>65</td><td>35</td><td> <bold>&#x0003c;0.001</bold> </td><td/><td/><td/></tr><tr><td/><td>High Exp. (&#x02267;medium)</td><td>101</td><td>35</td><td>65</td><td/><td/><td/><td/></tr></tbody></table><table-wrap-foot><p>Bold figures, Statistically significant.</p></table-wrap-foot></table-wrap><table-wrap id="Tab2"><label>Table 2</label><caption><p> <bold>Univariate log-rank analysis for disease-specific survival and metastasis-free survival</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr><th rowspan="2">Parameters</th><th rowspan="2">Category</th><th rowspan="2">No. of case</th><th colspan="2">DSS</th><th colspan="2">MeFS</th></tr><tr><th>No. of events</th><th><bold><italic>P</italic></bold>value</th><th>No. of events</th><th><bold><italic>P</italic></bold>value</th></tr></thead><tbody><tr><td>Age (years)</td><td>&#x0003c;60&#x000a0;years</td><td>141</td><td>13</td><td>0.9804</td><td>41</td><td>0.7650</td></tr><tr><td/><td>&#x02267;60&#x000a0;years</td><td>61</td><td>4</td><td/><td>16</td><td/></tr><tr><td>Primary tumor (T)</td><td>T1</td><td>82</td><td>5</td><td> <bold>0.0283</bold> </td><td>9</td><td> <bold>&#x0003c;0.0001</bold> </td></tr><tr><td/><td>T2</td><td>96</td><td>10</td><td/><td>34</td><td/></tr><tr><td/><td>T3-4</td><td>24</td><td>2</td><td/><td>14</td><td/></tr><tr><td>Nodal status (N)</td><td>N0</td><td>117</td><td>7</td><td> <bold>0.0079</bold> </td><td>19</td><td> <bold>&#x0003c;0.0001</bold> </td></tr><tr><td/><td>N1-2</td><td>85</td><td>10</td><td/><td>38</td><td/></tr><tr><td>Stage</td><td>I</td><td>63</td><td>3</td><td> <bold>0.0001</bold> </td><td>5</td><td> <bold>&#x0003c;0.0001</bold> </td></tr><tr><td/><td>II</td><td>115</td><td>10</td><td/><td>38</td><td/></tr><tr><td/><td>III</td><td>24</td><td>4</td><td/><td>14</td><td/></tr><tr><td>Histologic grade</td><td>Grade I</td><td>18</td><td>0</td><td>0.2066</td><td>1</td><td> <bold>0.0269</bold> </td></tr><tr><td/><td>Grade II</td><td>141</td><td>15</td><td/><td>41</td><td/></tr><tr><td/><td>Grade III</td><td>43</td><td>2</td><td/><td>15</td><td/></tr><tr><td>Cox-2 expression</td><td>Low Exp (&#x0003c;medium)</td><td>101</td><td>4</td><td> <bold>0.0043</bold> </td><td>15</td><td> <bold>&#x0003c;0.0001</bold> </td></tr><tr><td>(Tumor)</td><td>High Exp (&#x02267;medium)</td><td>101</td><td>13</td><td/><td>42</td><td/></tr><tr><td>Ido-1 expression</td><td>Low Exp (&#x0003c;medium)</td><td>101</td><td>4</td><td> <bold>0.0045</bold> </td><td>9</td><td> <bold>&#x0003c;0.0001</bold> </td></tr><tr><td>(CAF)</td><td>High Exp (&#x02267;medium)</td><td>101</td><td>13</td><td/><td>48</td><td/></tr></tbody></table><table-wrap-foot><p>Bold figures, Statistically significant.</p></table-wrap-foot></table-wrap></sec><sec id="Sec20"><title>COX-2 and IDO inhibitors suppressed growth of COX-2-overexpressing breast tumors <italic>in vivo</italic></title><p>The effect of COX-2 and IDO inhibitors was evaluated in an orthotopic model. Inoculation of MCF-7/RMF-EG- or COX-2-overexpressing MCF-7/RMF-EG cell mixture induced tumors in nude mice primed with 17&#x003b2;-estradiol injection (Figure&#x000a0;<xref rid="Fig6" ref-type="fig">6</xref>A). Tumor growth was higher in the COX-2-overexpressing MCF-7/RMF-EG group, and a 2.4-fold of increase of tumor volume was detected at 10&#x000a0;weeks (<italic>P</italic>&#x02009;&#x0003c;&#x02009;0.01). The COX-2-overexpressing MCF-7/RMF-EG group was randomly divided into four subgroups (<italic>n</italic>&#x02009;=&#x02009;3). Intratumoral injection of vehicle (DMSO, control), 10&#x000a0;mg/kg of NS398, 10&#x000a0;mg/kg of 1-methyl-L-tryptophan, or both inhibitors was conducted, and treatment was continuous for another 2&#x000a0;weeks. As shown in Figure&#x000a0;<xref rid="Fig6" ref-type="fig">6</xref>B, tumor volume of the groups treated with NS398 or 1-methyl-L-tryptophan was smaller than that of the control group. Co-treatment of COX-2 and IDO inhibitor induced a more obvious reduction in tumor size, although it did not show an additive effect.<fig id="Fig6"><label>Figure 6</label><caption><p><bold>Inhibition of tumor growth by COX-2 and IDO inhibitors</bold> <bold><italic>in vivo</italic></bold> <bold>. (A)</bold> MCF-7 or MCF-7-COX2 (8&#x02009;&#x000d7;&#x02009;10⁶) cells mixed with RMF-EG (6&#x02009;&#x000d7;&#x02009;10⁶) cells were inoculated into the fourth mammary fat pads of 6-week-old female nude mice. Before inoculation of the cancer cell/fibroblast mixture, all mice were primed with 6&#x000a0;mg/kg of 17&#x003b2;-estradiol twice a week for 3&#x000a0;weeks. Measurement of tumor growth was begun at 4&#x000a0;weeks after injection, and tumor volume was continuously monitored. The difference between the groups was evaluated by repeated measures two-way ANOVA analysis. n.s., no significance. *<italic>P</italic>&#x02009;&#x0003c;&#x02009;0.01. <bold>(B)</bold> After 10&#x000a0;weeks, mice injected with MCF-COX-2 and RMF-EG cells were randomly divided into four groups that received vehicle (DMSO), NS-398 (10&#x000a0;mg/kg), L-1-methy-tryptophan (10&#x000a0;mg/kg), or both inhibitors 5 times per week. Two weeks later, animals were killed, and tumors were isolated from mice.</p></caption><graphic xlink:href="13058_2014_410_Fig6_HTML" id="d29e2078"/></fig></p></sec></sec><sec id="Sec21"><title>Discussion</title><p>Previous studies demonstrated that IDO overexpression increases the secretion of kynurenine to inhibit effect T cells to promote immune escape and tumor progression in various human cancers [<xref ref-type="bibr" rid="CR27">27</xref>]-&#x02013;[<xref ref-type="bibr" rid="CR29">29</xref>]. The expression of IDO in cancer stroma has not been clarified. In addition, the clinical significance of stromal IDO is unclear.</p><p>In this study, we provide evidence that COX-2-overexpressing breasts cancer cells may secrete PGE₂ to induce IDO expression and kynurenine production in stromal fibroblasts. In addition, we show that kynurenine in the coculture-conditioned medium is produced mainly by CAFs because IDO is not induced by COX-2 overexpression in MCF-7 cells. An important upstream regulator of IDO is interferon-&#x003b3;. Yoshida <italic>et al</italic>. [<xref ref-type="bibr" rid="CR30">30</xref>] first reported that the pulmonary IDO was induced in the mouse after intraperitoneal administration of bacterial endotoxin or during <italic>in vivo</italic> virus infection, and this induction was triggered by interferon-&#x003b3; [<xref ref-type="bibr" rid="CR30">30</xref>]. Because interferon-&#x003b3; exhibits antitumor activity on various cancers <italic>in vitro</italic> and <italic>in vivo</italic>, it is unlikely that COX-2-overexpressing cancer cells produce interferon-&#x003b3; to stimulate stromal IDO. For the first time, we show that cancer cell-produced PGE₂ transcriptionally upregulates IDO expression through the EP4/STAT3 signaling pathway. <italic>In vivo</italic> binding of STAT3 to <italic>IDO</italic> gene promoter is confirmed by ChIP assay. Additionally, knockdown of STAT3 totally abolishes EP4 agonist-induced IDO expression. These data suggest that <italic>IDO</italic> is a direct transcriptional target of STAT3.</p><p>An unresolved question is why PGE₂ stimulates IDO expression in stromal fibroblasts but not in breast cancer cells, because both cell types express EP4 receptor [<xref ref-type="bibr" rid="CR31">31</xref>] and data not shown]. We are aware that the binding of STAT1 to <italic>IDO</italic> promoter is reduced by PGE₂ (Additional file <xref rid="MOESM3" ref-type="media">3</xref>: Figure S2); therefore, it is possible that the expression level of STAT1 and STAT3 and the competition between these two STATs may determine the response of cells to PGE₂ stimulation.</p><p>The concept of oncometabolite was established by the studies that mutations of isocitrate dehydrogenase 1 (IDH1) and IDH2 generate a novel metabolite 2-hydroxyglutarate (2-HG) that exhibits oncogenic activity in acute myeloid leukemia and glioma [<xref ref-type="bibr" rid="CR32">32</xref>],[<xref ref-type="bibr" rid="CR33">33</xref>]. Subsequently, 2-HG was shown to be a competitive inhibitor of &#x003b1;-ketoglutarate-dependent dioxygenases and inhibits histone demethylases like Tet methylcytosine dioxygenase 2 (TET2) to change promoter methylation and gene transcription [<xref ref-type="bibr" rid="CR34">34</xref>],[<xref ref-type="bibr" rid="CR35">35</xref>]. Kynurenine represents another oncometabolite, which acts as an immunosuppressor to create a favorable microenvironment for tumor formation and metastasis [<xref ref-type="bibr" rid="CR36">36</xref>]. A recent study demonstrated that the tryptophan catabolism enzyme TDO is overexpressed in human brain tumors, and elevated secretion of kynurenine promotes cell migration via an AhR-dependent pathway [<xref ref-type="bibr" rid="CR25">25</xref>].</p><p>However, the underlying mechanism by which kynurenine increases cell motility is still unclear. After screening of the EMT markers, we found that E-cadherin is decreased in kynurenine-treated breast cancer cells, and AhR is involved in this process. AhR has been shown to integrate as a component of a novel Cul4B ubiquitin E3 ligase complex and participated in the degradation of sex steroid receptors [<xref ref-type="bibr" rid="CR26">26</xref>]. We demonstrated that kynurenine increases the interaction between AhR and E-cadherin, and the AhR/E-cadherin complex also contains Skp2, an F-box protein of SCF E3 ligase. The formation of the E-cadherin/AhR/Skp2 complex and ubiquitination of E-cadherin induced by kynurenine is also detectable in A549 cells, indicating a general mechanism of kynurenine-induced proteolysis of E-cadherin in different cancer cells. Our results provide a novel oncometabolite function of kynurenine to enhance cancer cell migration by degrading E-cadherin.</p><p>The clinical validation of tumor COX-2 and stromal IDO in this study is important to verify the cancer-stroma interplay in cancer progression. Many histopathologic studies investigated the expression of two specific genes in the epithelial components of tumor tissues to show their association and to demonstrate the vertical regulation of these two genes. The correlation and clinical significance of genes separately expressed in tumor and stroma have received little attention.</p><p>However, the gene signatures in CAFs may provide more information than originally thought. West <italic>et al</italic>. [<xref ref-type="bibr" rid="CR37">37</xref>] first classified two stromal gene signature from tumors with solitary fibrous tumor (SFT) and desmoids-type fibromatosis (DTF) features and showed that patients with the expression of DTF had a favorable clinical outcome. Their subsequent study by using public databases and immunohistochemical approaches suggested that DTF fibroblast signature is a common tumor stroma signature in different types of cancers [<xref ref-type="bibr" rid="CR38">38</xref>]. Mercier <italic>et al</italic>. [<xref ref-type="bibr" rid="CR39">39</xref>] identified a hyperproliferative gene signature in CAFs and found that breast cancer patients with this signature had a poor prognosis with tamoxifen monotherapy and a great reduction in recurrence-free survival [<xref ref-type="bibr" rid="CR39">39</xref>]. By using a mouse model of squamous skin carcinogenesis, Erez [<xref ref-type="bibr" rid="CR40">40</xref>] demonstrated that carcinoma cells could educate CAFs to express proinflammatory genes to promote macrophage recruitment, neovascularization, and tumor growth. Additionally, this gene signature was also evident in mammary and pancreatic tumors in mice and in human cancers. By using metabolomics, molecular, and pathological approaches, we revealed that induction of stromal IDO by COX-2-overexpressing breast cancer cells promotes tumor progression and predicts poor patient survival.</p><p>Results of our animal study also clearly demonstrate the anticancer effect of COX-2 and IDO inhibitor on COX-2-overexpressing breast cancer <italic>in vivo</italic>. A combination of IDO and COX-2 inhibitor exhibits a more obvious effect on the inhibition of tumor growth. However, we did not find an additive effect. This can be because (a) the number of animals in each group is small, and (b) inhibition of COX-2 in cancer cells will attenuate stromal IDO expression, which reduces the anticancer activity of IDO inhibitor. Additional experiments are needed to clarify this issue.</p></sec><sec id="Sec22"><title>Conclusion</title><p>By using a metabolomics approach, we identified potential oncometabolites involved in the crosstalk between COX-2-overexpressing breast cancer cells and fibroblasts. Molecular study elucidates the underlying mechanism by which this cancer/stroma interplay via COX-2 and IDO promotes tumor progression. In addition, pathological investigation validates the importance of cancer COX-2 and stromal IDO in the prediction of the patient&#x02019;s survival. Simultaneous targeting of COX-2 and IDO may be a new strategy for breast cancer treatment.</p></sec><sec id="Sec23"><title>Additional files</title></sec><sec sec-type="supplementary-material"><title>Electronic supplementary material</title><sec id="Sec24"><p> <supplementary-material content-type="local-data" id="MOESM1"><media xlink:href="13058_2014_410_MOESM1_ESM.doc"><caption><p>Additional file 1: Supplementary materials and methods.(DOC 50 KB)</p></caption></media></supplementary-material> <supplementary-material content-type="local-data" id="MOESM2"><media xlink:href="13058_2014_410_MOESM2_ESM.tiff"><caption><p>Additional file 2: Figure S1.: PGE₂ stimulated <italic>IDO</italic> promoter activity. Different <italic>IDO</italic> promoter constructs were transfected into MCF-7 cells and stimulated by PGE₂. Promoter assay indicated that PGE₂ activated <italic>IDO</italic> via the &#x02212;1140/-844 promoter region. (TIFF 118 KB)</p></caption></media></supplementary-material> <supplementary-material content-type="local-data" id="MOESM3"><media xlink:href="13058_2014_410_MOESM3_ESM.tiff"><caption><p>Additional file 3: Figure S2.: <italic>In vivo</italic> binding of STAT3 on <italic>IDO</italic> gene promoter in EMF-EG fibroblasts and its regulation by co-culture of COX-2-overexpressing MCF7 cells. ChIP assay demonstrated that the binding of STAT3 to <italic>IDO</italic> promoter was increased, whereas the binding of STAT1 was reduced in EMF-EG fibroblasts after co-culture with COX-2-overexpressing MCF-7 cells. (TIFF 93 KB)</p></caption></media></supplementary-material> <supplementary-material content-type="local-data" id="MOESM4"><media xlink:href="13058_2014_410_MOESM4_ESM.tiff"><caption><p>Additional file 4: Figure S3.: Kynurenine induced the formation of E-cadherin/AhR/Skp2 complex in A549 lung cancer cells. A549 cells were treated without (&#x02212;) or with (+) kynurenine, and the interaction between E-cadherin and AhR or Skp2 was studied by immunoprecipitation and Western blotting. (TIFF 162 KB)</p></caption></media></supplementary-material> <supplementary-material content-type="local-data" id="MOESM5"><media xlink:href="13058_2014_410_MOESM5_ESM.tiff"><caption><p>Additional file 5: Figure S4.: IDO expression in cancer-associated fibroblasts (CAFs) was increased in COX-2-overexpressing breast cancer. (TIFF 153 KB)</p></caption></media></supplementary-material> </p><p>Below are the links to the authors&#x02019; original submitted files for images.<supplementary-material content-type="local-data" id="MOESM6"><media xlink:href="13058_2014_410_MOESM6_ESM.gif"><caption><p>Authors&#x02019; original file for figure 1</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="MOESM7"><media xlink:href="13058_2014_410_MOESM7_ESM.gif"><caption><p>Authors&#x02019; original file for figure 2</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="MOESM8"><media xlink:href="13058_2014_410_MOESM8_ESM.gif"><caption><p>Authors&#x02019; original file for figure 3</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="MOESM9"><media xlink:href="13058_2014_410_MOESM9_ESM.gif"><caption><p>Authors&#x02019; original file for figure 4</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="MOESM10"><media xlink:href="13058_2014_410_MOESM10_ESM.gif"><caption><p>Authors&#x02019; original file for figure 5</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="MOESM11"><media xlink:href="13058_2014_410_MOESM11_ESM.gif"><caption><p>Authors&#x02019; original file for figure 6</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="MOESM12"><media xlink:href="13058_2014_410_MOESM12_ESM.tiff"><caption><p>Authors&#x02019; original file for figure 7</p></caption></media></supplementary-material></p></sec></sec>

Nestin positively regulates the Wnt/&#x003b2;-catenin pathway and the proliferation, survival and invasiveness of breast cancer stem cells

Could not extract abstract

<contrib contrib-type="author"><name><surname>Zhao</surname><given-names>Zuowei</given-names></name><address><email>dlzhaozw@163.com</email></address><xref ref-type="aff" rid="Aff1">1</xref></contrib><contrib contrib-type="author"><name><surname>Lu</surname><given-names>Ping</given-names></name><address><email>luping2999@aliyun.com</email></address><xref ref-type="aff" rid="Aff1">1</xref></contrib><contrib contrib-type="author"><name><surname>Zhang</surname><given-names>Hao</given-names></name><address><email>yincailove@126.com</email></address><xref ref-type="aff" rid="Aff1">1</xref></contrib><contrib contrib-type="author"><name><surname>Xu</surname><given-names>Huanming</given-names></name><address><email>10169067@qq.com</email></address><xref ref-type="aff" rid="Aff2">2</xref></contrib><contrib contrib-type="author"><name><surname>Gao</surname><given-names>Ningning</given-names></name><address><email>465095929@qq.com</email></address><xref ref-type="aff" rid="Aff3">3</xref></contrib><contrib contrib-type="author"><name><surname>Li</surname><given-names>Man</given-names></name><address><email>liman126126@163.com</email></address><xref ref-type="aff" rid="Aff1">1</xref></contrib><contrib contrib-type="author" corresp="yes"><name><surname>Liu</surname><given-names>Caigang</given-names></name><address><email>angel-s205@163.com</email></address><xref ref-type="aff" rid="Aff1">1</xref></contrib><aff id="Aff1"><label>1</label><institution-wrap><institution-id institution-id-type="GRID">grid.452828.1</institution-id><institution>Department of Breast Surgery, </institution><institution>the Second Affiliated Hospital of Dalian Medical University, </institution></institution-wrap>465 Zhongshan Road, Dalian, 116027 Liaoning Province China </aff><aff id="Aff2"><label>2</label><institution-wrap><institution-id institution-id-type="GRID">grid.412644.1</institution-id><institution>Department of Transfusion, </institution><institution>the Fourth Affiliated Hospital of China Medical University, </institution></institution-wrap>4 Chongshan East Road, Shenyang, 110032 Liaoning Province China </aff><aff id="Aff3"><label>3</label><institution-wrap><institution-id institution-id-type="GRID">grid.412636.4</institution-id><institution>Department of Ultrasonic Diagnosis, </institution><institution>the First Affiliated Hospital of China Medical University, </institution></institution-wrap>155 Nanjing South Street, Shenyang, 110001 Liaoning Province China </aff>

Breast Cancer Research : BCR

<sec id="Sec1"><title>Introduction</title><p>Breast cancer is the most common malignancy in women, and its incidence is increasing worldwide. The disease remains a huge threat to women&#x02019;s health. Patients with &#x02018;triple-negative&#x02019; breast cancer, referring to absence in the expression of estrogen receptor (ER), progesterone receptor (PR) and human epidermal growth factor receptor 2 (HER2), are insensitive to hormonal therapy or HER2-targeted agents [<xref ref-type="bibr" rid="CR1">1</xref>]. These patients are also at increased risk for relapse and metastasis of breast cancer, which is often incurable and a leading cause of female mortality [<xref ref-type="bibr" rid="CR2">2</xref>]&#x02013;[<xref ref-type="bibr" rid="CR4">4</xref>]. Unfortunately, there is currently no effective therapy to control the recurrence and metastasis of triple-negative breast cancer. Therefore, understanding the molecular mechanisms underlying the recurrence and metastasis of triple-negative breast cancer may reveal new therapeutic targets, and can significantly improve the management of patients with triple-negative breast cancer.</p><p>Nestin, an intermediate filament protein, was initially identified as a neural stem cell marker [<xref ref-type="bibr" rid="CR5">5</xref>]&#x02013;[<xref ref-type="bibr" rid="CR7">7</xref>]. Nestin expression has been detected in malignant tumor tissues, and has been implicated in the development and metastasis of malignant tumors, such as brain malignancies [<xref ref-type="bibr" rid="CR8">8</xref>],[<xref ref-type="bibr" rid="CR9">9</xref>], melanoma [<xref ref-type="bibr" rid="CR10">10</xref>], colorectal [<xref ref-type="bibr" rid="CR11">11</xref>], prostate [<xref ref-type="bibr" rid="CR12">12</xref>], and pancreatic cancers [<xref ref-type="bibr" rid="CR13">13</xref>]. Nestin overexpression has been reported in metastatic breast cancers, (especially in triple-negative breast cancers) [<xref ref-type="bibr" rid="CR7">7</xref>],[<xref ref-type="bibr" rid="CR14">14</xref>]&#x02013;[<xref ref-type="bibr" rid="CR17">17</xref>], and has been associated with poor prognosis in a Caucasian breast cancer cohort [<xref ref-type="bibr" rid="CR16">16</xref>]. However, this association has not been verified in Chinese patients with triple-negative breast cancer.</p><p>Cancer stem cells (CSC) play an important role in the development and metastasis of breast cancer [<xref ref-type="bibr" rid="CR18">18</xref>]&#x02013;[<xref ref-type="bibr" rid="CR20">20</xref>]. Breast CSC exhibit potent tumorigenicity, and implantation with a few ESA⁺CD44⁺CD24^&#x02212; lineage^&#x02212; breast CSC has been demonstrated to induce solid tumor formation in severe combined immunodeficiency (SCID) mice [<xref ref-type="bibr" rid="CR21">21</xref>]. Our previous study shows that Nestin and other stemness factors are expressed in breast cancer tissues, and that their expression is associated with poor survival of patients with breast cancer [<xref ref-type="bibr" rid="CR22">22</xref>]. These findings suggest that Nestin may regulate the development and metastasis of breast cancers. However, little is known about the molecular mechanisms by which Nestin regulates cell proliferation, survival, and invasiveness of breast CSC. Previous studies have shown that the Wnt/&#x003b2;-catenin pathway is crucial for the development and progression of breast cancer [<xref ref-type="bibr" rid="CR23">23</xref>],[<xref ref-type="bibr" rid="CR24">24</xref>]. High levels of Wnt receptor and co-receptor expression, as well as aberrant activation of &#x003b2;-catenin, have been detected in breast cancer tissues. Downregulation of the Wnt/&#x003b2;-catenin pathway can inhibit the epithelial-mesenchymal transition (EMT) and reduce spontaneous invasion of breast cancer cells [<xref ref-type="bibr" rid="CR25">25</xref>],[<xref ref-type="bibr" rid="CR26">26</xref>]. Therapeutic targeting of the porcupine (PORCN) protein suppresses Wnt/&#x003b2;-catenin activation, and significantly reduces the spontaneous development of mammary tumors in transgenic mice [<xref ref-type="bibr" rid="CR27">27</xref>]. In addition, aberrant activation of the Wnt/&#x003b2;-catenin pathway has been associated with the development of resistance to radiotherapy and chemotherapy in breast cancer [<xref ref-type="bibr" rid="CR28">28</xref>],[<xref ref-type="bibr" rid="CR29">29</xref>]. The Wnt/&#x003b2;-catenin pathway is also important for self-renewal and migration of breast CSC [<xref ref-type="bibr" rid="CR30">30</xref>],[<xref ref-type="bibr" rid="CR31">31</xref>]. &#x003b2;-catenin is expressed constitutively in many types of tumor cells. Increased Wnt/&#x003b2;-catenin activation enhances the tumorigenicity of breast CSC [<xref ref-type="bibr" rid="CR32">32</xref>]. In the absence of Wnt binding to its receptor, &#x003b2;-catenin forms a degradation complex that includes the Axin, adenomatous polyposis coli (APC), and glycogen synthase kinase-3 beta (GSK-3&#x003b2;) proteins. GSK-3&#x003b2;/ casein kinase 1 (CK1) phosphorylates &#x003b2;-catenin, targeting it for &#x003b2;-TrCP1-mediated degradation, which is positively regulated by peroxisome proliferator-activated receptor alpha (PPARa) and PPARb. Binding of Wnt to its receptor and co-receptors causes the activation of Dishevelled (Dsh) proteins by phosphorylation. Activated Dsh then recruits GSK-3&#x003b2;, releasing &#x003b2;-catenin and promoting the nuclear translocation of &#x003b2;-catenin. Subsequently, &#x003b2;-catenin binds to the Tcf and Lef transcription factors in the nucleus, leading to the transcription of downstream genes, including <italic>c-Myc, cyclin D,</italic> and <italic>MMP-7</italic>. Although Wnt/&#x003b2;-catenin activation can promote the expression of various stemness factors, including Nestin, little is known about whether the modulation of Nestin expression can affect Wnt/&#x003b2;-catenin activation in breast CSC.</p><p>In the present study, we first characterized Nestin expression in 150 tumor specimens from patients with triple-negative breast cancer, and analyzed the potential association between levels of Nestin expression and the survival of patients. We isolated CD44⁺CD24^&#x02212; CSC from 26 triple-negative breast cancer tissues to generate Nestin-overexpressing (Nestin⁺), Nestin-silencing (Nestin-si), and control (Nestin-c) CSC. We also isolated CD44⁺CD24^&#x02212;, cell surface Nestin^high, or Nestin^low CSC from 12 triple-negative breast cancer tissues. Subsequently, we characterized the ability of these CSC to form mammospheres <italic>in vitro</italic>, and of Nestin^high or Nestin^low to induce tumors in SCID mice. We examined the effects of Nestin silencing on the cell cycle, survival, and apoptosis of breast CSC, and explored potential mechanisms underlying the action of Nestin in the tumorigenicity of breast CSC. Our data suggest that Nestin may promote the proliferation, survival, and migration of breast CSC by enhancing Wnt/&#x003b2;-catenin activation.</p></sec><sec id="Sec2"><title>Methods</title><sec id="Sec3"><title>Breast tissue specimens</title><p>A total of 150 breast cancer samples were obtained from patients with ER-/PR-/HER2- (triple-negative) breast cancer at the Department of Breast Surgery of the First Affiliated Hospital of China Medical University between January 2001 and December 2006. In addition, 12 patients with triple-negative breast cancer were recruited and ESA⁺CD44⁺CD24^&#x02212; lineage^&#x02212; CSC in the resected breast cancer tissues were isolated. The CSC were stained with anti-Nestin and sorted for cell surface positive Nestin^high or negative Nestin^low CSC. Patients with triple-negative breast cancer were diagnosed by histological examination of tissue samples, and they underwent radical surgery in the Department of Breast Surgery. The inclusion criteria were: (a) curative surgical resection; (b) pathological examination of the resected tumor; (c) pathological examination of &#x0003e;15 lymph nodes after surgery; and (d) availability of a complete medical record. The demographic and clinical data for individual patients were obtained from medical records. Individuals with breast cancer were excluded if they did not fulfill the criteria for inclusion. Written informed consent was obtained from individual patients and the experimental protocol was approved by the Ethics Committee of China Medical University.</p></sec><sec id="Sec4"><title>Immunohistochemical staining</title><p>Individual breast cancer tissue samples were fixed in 10% neutralized formalin (pH&#x000a0;7.0) and paraffin-embedded. The tissue sections (4&#x000a0;mm) were dewaxed, rehydrated, and treated with 3% H2O2 in methanol, followed by incubation overnight with primary anti-Nestin antibody (Santa Cruz Biotechnology, Santa Cruz, CA, USA). Subsequently, the tissue sections were incubated with Multi-Link biotinylated swine anti-goat/mouse/rabbit immunoglobulin (Ig)G (Dako, Carpinteria, CA, USA). After washing the cells, the bound antibodies were detected using horseradish peroxidase (HRP)-conjugated avidin-biotin complex (1:1000 dilution, Vector Laboratories, Burlingame, CA, USA), and visualized using 3,3-diaminobenzidine (DAB), followed by counterstaining with Gill&#x02019;s hematoxylin.</p><p>The intensity of anti-Nestin staining was scored semi-quantitatively. Cells with yellow to brown cytoplasmic staining were considered to be Nestin⁺ cells. Nestin expression levels in individual tumor tissues were assigned a score based on the following criteria: 0 if &#x0003c;1% neoplastic cells are Nestin⁺; 1+ if neoplastic cells were 1 to 10% Nestin⁺; and 2+ if &#x02265;10% Nestin⁺ neoplastic cells. Individual sections with 1+ or 2+ anti-Nestin staining were considered positive.</p></sec><sec id="Sec5"><title>Preparation and characterization of breast CSCs</title><p>Thirty-eight freshly resected breast tumor specimens were obtained for the preparation of CSC, as described previously [<xref ref-type="bibr" rid="CR33">33</xref>]. Briefly, fresh breast tumor specimens were cut into small pieces and were digested with 1&#x000a0;mg/mL of collagenase type III (5&#x000a0;mL/g tissue, Worthington Biochemical, Lakewood, NJ, USA) in 5% fetal bovine serum (FBS) containing RPMI medium at 37&#x000b0;C for 2&#x000a0;h, and centrifuged. The tumor cells (106/tube) were sequentially stained with FITC-anti-CD2, APC-anti-CD3, PE-anti-CD10, FITC-anti-CD16, APC-anti-CD18, PE-anti-CD31, and FITC-anti-CD326 lineage markers, as well as 7-AAD (BD Biosciences Pharmingen, San Diego, CA, USA). Lineage⁺ and dead cells were first eliminated by flow cytometric sorting. Subsequently, the unstained lineage-negative cells were stained with PE-anti-CD24 and FITC-anti-CD44, and CD44⁺CD24^&#x02212; breast CSC were purified by flow sorting. Finally, the sorted CD44⁺CD24^&#x02212; breast CSC were stained with rabbit anti-Nestin (#N5413; Sigma-Aldrich, St. Louis, MO, USA) and then with APC-anti-rabbit IgG to allow sorting of the cell surface positive (Nestin^high) and negative (Nestin^low) breast CSC.</p></sec><sec id="Sec6"><title>Transfection</title><p>The purified CD44⁺CD24^&#x02212; breast CSC from 26 specimens were transfected with human Nestin-specific (sc-36032, a mixture of sc-36032A, sense 5&#x02032;-CGAGGUCUUUAGAAGAAGAtt-3&#x02032; and antisense 5&#x02032;-UCUUCUUCUAAAGACCUCGtt-3&#x02032;; sc-36032B, sense 5&#x02032;-GCCUUUAGAUCUCUAGAAAtt-3&#x02032; and antisense 5&#x02032;- UUUCUAGAGAUCUAAAGGCtt-3; sc-36032C, sense 5&#x02032;- GGCAAUGAAUCCUCUAGAAtt-3&#x02032; and antisense 5&#x02032;-UUCUAGAGGAUUCAUUGCCtt-3&#x02032;) or control small interfering RNA (siRNA) (sc-37007, Santa Cruz Biotechnology) using Lipofectamine 2000 Reagent (Invitrogen, Carlsbad, CA, USA), according to the manufacturers&#x02019; protocols. Nestin-silenced (Nestin-si) and control siRNA-transfected (Nestin-c) CSC were harvested at 24 or 48&#x000a0;h posttransfection, and the efficacy of Nestin silencing was determined by western blotting. To generate Nestin overexpressing CSC, CD44⁺CD24^&#x02212; cells were transfected with pIRES-Nestin-EGFP (constructed in our laboratory) or control pIRES-EGFP (Clontech Laboratory, Mountain View, CA, USA) for 48&#x000a0;h. EGFP⁺ Nestin⁺ cells were then isolated by flow sorting. A preparation of EGFP⁺Nestin⁺ CSC with a purity of &#x0003e;95% was used for the subsequent experiments.</p></sec><sec id="Sec7"><title>Mammosphere formation and <italic>in vivo</italic>xenograft assays</title><p>Nestin⁺, Nestin-si, Nestin-c, Nestin^high, Nestin^low, and unmanipulated control breast CSC (5,000 cells/mL) were cultured in triplicate in Complete MammoCult&#x02122; Medium in six-well ultralow attachment plates (Corning Life Sciences, Tewksbury, MA, USA) for five days in the presence of human leukemia inhibitory factor (LIF, 50&#x000a0;ng/mL), an enhancer of mammosphere formation <italic>in vitro</italic>. The number of mammospheres (&#x0003e;50&#x000a0;mm in a dimension) in individual wells were counted in a blinded manner. In addition, Nestin^high CSC in the presence of LGK974 (1 nM), Nestin^low CSC in the presence of SB216763 (5&#x000a0;&#x003bc;M, Selleckchem, Houston, TX, USA), and control CSC in the presence of a vehicle were tested for the formation of mammospheres <italic>in vitro</italic>.</p><p>Eight-week-old female C57BL/6 SCID mice were obtained from The Jackson Laboratory (Beijing, China) and housed in a specific pathogen-free facility on our campus. C57BL/6 SCID mice were implanted with different numbers (10³ to 10⁵/mouse) of the purified ESA⁺CD44⁺CD24^&#x02212;lineage^&#x02212; Nestin^high, Nestin^low, and control CSC in 50&#x000a0;ml phosphate-buffered saline (PBS) in their mammary fat pads. The development of solid tumors was monitored for up to 28&#x000a0;days post xenotransplantation. The levels of Nestin expression in the dissected tumors were determined by western blotting. The experimental protocol was approved by the Animal Care and Research Committee of China Medical University.</p></sec><sec id="Sec8"><title>Flow cytometric analysis of cell cycling and apoptosis in CSC</title><p>Following transfection for 96&#x000a0;h, the Nestin-si and Nestin-c breast CSC were fixed by treatment with 70% cold ethanol, followed by staining with propidium iodide (PI, 30&#x000a0;&#x003bc;g/mL). The percentages of CSC in each phase of the cell cycle were determined by flow cytometry on a FACS Calibur device (BD Biosciences, San Jose, CA, USA). The Nestin-si and Nestin-c breast CSC were stained with FITC-Annexin V (2.5&#x000a0;&#x003bc;g/mL) and PI (5&#x000a0;&#x003bc;g/mL), and the percentages of spontaneously apoptotic Nestin-si and Nestin-c breast CSC were determined by flow cytometry.</p></sec><sec id="Sec9"><title>Transwell invasion assay</title><p>Nestin-si and Nestin-c breast CSC (1&#x02009;&#x000d7;&#x02009;10⁵ cells/well) were cultured in triplicate at 37&#x000b0;C for 1.5&#x000a0;h in the upper chambers of transwell plates (Corning) that had been loaded with 60 to 80&#x000a0;&#x003bc;L of diluted Matrigel (BD Biosciences). The cells that had migrated into the lower chamber were fixed with 95% ethanol and stained with crystal violet in gluteraldehyde, followed by visualization. The number of migrated cells in 10 random fields (magnification 400&#x000d7;) of each chamber were counted in a blinded manner.</p></sec><sec id="Sec10"><title>Quantitative reverse transcription PCR</title><p>The levels of Nestin, Axin, GSK-3&#x003b2;, APC, and &#x003b2;-catenin mRNA transcripts relative to &#x003b2;-actin were determined by quantitative reverse transcription PCR (qRT-PCR). Briefly, total RNA was isolated from Nestin-si and Nestin-c CSC using Trizol reagent, and reverse-transcribed into cDNA using the RevertAid&#x02122; First Strand cDNA synthesis kit (Fermentas, Pittsburgh, PA, USA), according to the manufacturer&#x02019;s instructions. The real-time PCR was performed using the SYBR Green PCR Master Mix and specific primers (Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>) on an Applied Biosystems 7500 Fast Real-Time PCR System (Applied Biosystems, Carlsbad, CA, USA). The amplification was performed at 95&#x000b0;C for 5&#x000a0;min and then 40&#x000a0;cycles of 94&#x000b0;C for 15&#x000a0;s, 55&#x000b0;C for 20&#x000a0;s, 72&#x000b0;C for 20&#x000a0;s, followed by 72&#x000b0;C for 7&#x000a0;min. The levels of mRNA transcripts were analyzed by the 2-&#x00394;&#x00394;Ct method. The relative levels of each gene transcript relative to &#x003b2;-actin in Nestin-c CSC were designated as 1.<table-wrap id="Tab1"><label>Table 1</label><caption><p> <bold>The sequences of primers</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr><th>Genes</th><th>Sequences</th><th>Sizes (bp)</th></tr></thead><tbody><tr><td rowspan="2">Axin</td><td>F: 5&#x02032;-CTCCAgTAgACggTACAgCgAAg-3&#x02032;</td><td rowspan="2">90</td></tr><tr><td>R: 5&#x02032;-gCATAgCCggCATTgACATA-3&#x02032;</td></tr><tr><td rowspan="2">GSK-3&#x003b2;</td><td>F: 5&#x02032;-CTACCAAATgggCGAGCATGAG-3&#x02032;</td><td rowspan="2">112</td></tr><tr><td>R: 5&#x02032;-CTGCTTGAATCCGAGCATGAG-3&#x02032;</td></tr><tr><td rowspan="2">Nestin</td><td>F: 5&#x02032;-CTCCAAGAATGGAGGCTGTAGGAA-3&#x02032;</td><td rowspan="2">75</td></tr><tr><td>R: 5&#x02032;-CCTATGAGATGGAGCAGGCAAGA-3&#x02032;</td></tr><tr><td rowspan="2">APC</td><td>F: 5&#x02032;-CCTCTGAAACAGTGCTGAACTTG-3&#x02032;</td><td rowspan="2">158</td></tr><tr><td>R: 5&#x02032;-CACCTGGTACTTGGCCACTA-3&#x02032;</td></tr><tr><td rowspan="2">&#x003b2;-catenin</td><td>F: 5&#x02032;-GTACGTCCATGGGTGGGACA-3&#x02032;</td><td rowspan="2">80</td></tr><tr><td>R: 5&#x02032;-GGCTCCGGTACAACCTTCAACTA-3&#x02032;</td></tr><tr><td rowspan="2">&#x003b2;-Actin</td><td>F: 5&#x02032;-TGGCACCCAGCACAATGAA-3&#x02032;</td><td rowspan="2">186</td></tr><tr><td>R: 5-CTAAGTCATAGTCCGCCTAGAAAGCA-3&#x02032;</td></tr></tbody></table></table-wrap></p></sec><sec id="Sec11"><title>Western blot analysis</title><p>Nestin-si and Nestin-c breast CSC were collected 48&#x000a0;h posttransfection and lysed in lysis buffer, followed by centrifugation. After quantification of protein concentrations using a BCA assay (Santa Cruz Biotechnology), the individual cell lysates (30&#x000a0;mg/lane) were separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto polyvinylidene fluoride (PVDF) membranes. The membranes were blocked with 5% fat-free milk powder in TBST and incubated with rabbit anti-Nestin, mouse anti-GSK-3&#x003b2;, anti-phosphorylated GSK-3&#x003b2;, anti-Axin, anti-APC (1:500; Cell Signaling Technology, Beverly, MA, USA), mouse anti-matrix metalloproteinase-2 (MMP-2), anti-MMP9 (1:400), anti-vimentin, anti-vascular endothelial growth factor (VEGF) (1:800), mouse anti-a-smooth muscle actin (a-SMA) (1:400), goat anti-E-cadherin (1:500; Santa Cruz Biotechnology), rabbit anti-N-cadherin (1:500), and rabbit anti-&#x003b2;-actin (1:5000; Abcam, Cambridge, UK) overnight at 4&#x000b0;C, respectively. After being washed, the bound antibodies were detected by incubation with HRP-conjugated anti-rabbit, anti-mouse, or anti-goat IgG at room temperature for 1&#x000a0;h, and visualized using enhanced chemiluminescence (Santa Cruz Biotechnology). Purified mouse, rabbit, or goat IgG were used as the negative controls. The levels of target proteins relative to &#x003b2;-actin were determined using the ImmuNe software. Furthermore, nuclear and cytoplasmic proteins were extracted from Nestin-si and Nestin-c CSC and subjected to western blot analysis. The relative levels of Nestin expression in the purified Nestin^high, Nestin^low, and unsorted control CSC as well as the dissected tumors from SCID mice that had been implanted with Nestin^high, Nestin^low, or control CSC were also analyzed by western blotting.</p></sec><sec id="Sec12"><title>Statistical analysis</title><p>Data are presented as mean&#x02009;&#x000b1;&#x02009;standard deviation (SD). Difference between groups was analyzed using Student&#x02019;s <italic>t</italic> test or the chi-square test. The cumulative survival of patients with Nestin⁺ or Nestin^&#x02212; triple-negative breast cancer was estimated using the Kaplan-Meier method, and analyzed using the log-rank test. All the statistical analyses were performed using the SPSS statistics 16.0 software package (SPSS Inc, Chicago, IL, USA). A <italic>P</italic> value of &#x0003c;0.05 was considered statistically significant.</p></sec></sec><sec id="Sec13"><title>Results</title><sec id="Sec14"><title>Nestin expression is associated with poor survival in patients with triple-negative breast cancer</title><p>Nestin is an intermediate filament protein expressed by neural precursors, muscle, and stem cells [<xref ref-type="bibr" rid="CR34">34</xref>], and has been shown to regulate cell proliferation. Our previous study showed that Nestin is expressed in human breast cancer tissue, and that its expression was associated with lymph node metastasis in a cohort of breast cancer patients [<xref ref-type="bibr" rid="CR16">16</xref>]. To further understand the role of Nestin in the development and progression of triple-negative breast cancer, we characterized the levels of Nestin expression in 150 specimens from patients with triple-negative breast cancer by immunohistochemistry (Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>A-C). Nestin expression was detected in myoepithelial cells in all of the matched adjacent nontumor areas (Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>A) and in the cytoplasm of tumor cells in 41 specimens (27.33%, Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>B and C), consistent with our previous observations [<xref ref-type="bibr" rid="CR16">16</xref>]. Survival analysis indicated that the survival of patients with Nestin⁺ triple-negative breast cancer was significantly reduced when compared with Nestin^&#x02212; triple-negative breast cancer (<italic>P</italic>&#x02009;=&#x02009;0.01, Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>D). These results further support a notion that Nestin regulates breast cancer progression, and suggest that Nestin expression may serve as a prognostic marker in patients with triple-negative breast cancers.<fig id="Fig1"><label>Figure 1</label><caption><p><bold>Nestin expression is associated with poor survival in patients with triple-negative breast cancer.</bold> Nestin expression in 150 surgically resected breast tissue samples from patients with triple-negative breast cancer was examined by immunohistochemistry. Survival was estimated using the Kaplan-Meier method, and the association between survival and Nestin expression was analyzed using the log-rank test. Data are representative images (400&#x000d7; magnification) and the cumulative survival of 150 patients. <bold>(A)</bold> Positive anti-Nestin staining in myoepithelial cells in the surrounding nontumor areas; <bold>(B, C)</bold> negative and positive anti-Nestin staining in breast tumors (white arrow indicates positive cytoplasmic staining). <bold>(D)</bold> Cumulative survival of patients with triple-negative breast cancer.</p></caption><graphic xlink:href="13058_2014_408_Fig1_HTML" id="d29e1002"/></fig></p></sec><sec id="Sec15"><title>Nestin^highbreast CSC exhibit an enhanced tumorigenicity</title><p>Breast CSC express Nestin and are crucial for the development and progression of breast cancer [<xref ref-type="bibr" rid="CR35">35</xref>]. To understand how Nestin regulates the proliferation and migration of breast CSC, we analyzed the expression of Nestin in ESA⁺CD44⁺CD24^&#x02212;lineage^&#x02212; CSC. We found that the proportion of CD44⁺CD24^&#x02212; Nestin^high CSC (57.45&#x02009;&#x000b1;&#x02009;18.27%) were significantly higher than that of CD44⁺CD24^&#x02212; Nestin^low CSC (42.55&#x02009;&#x000b1;&#x02009;16.4) in the 12 specimens isolated from patients with triple-negative breast cancer (<italic>P</italic> &#x0003c;0.05, Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>A). Western blot analysis indicated that the relative levels of Nestin expression in Nestin^high CSC were significantly higher than that of control CSC, and significantly higher than that in Nestin^low CSC (Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>B). Furthermore, Nestin^high CSC effectively formed mammospheres, with the mean size and numbers of mammospheres significantly greater than that of the control CSC (Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>C). In contrast, the Nestin^low CSC failed to form typical mammospheres, with the mean size of mammospheres significantly smaller than that of control CSC. Next, we tested the effect of Nestin expression on the tumorigenicity of breast CSC <italic>in vivo</italic>. We found that implantation of 10³ Nestin^high CSC resulted in tumor formation in 4 out of 10 C57BL/6 SCID mice, and the rates of tumor formation by Nestin^high CSC increased with increasing numbers of CSC, while implantation with 104 control CSC or 105 Nestin^low CSC was required to induce tumors (2 out of 10 mice, Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>D). More importantly, the relative levels of Nestin expression in the formed solid tumors from the different types of CSC was similar to that of the corresponding type of injected CSC (Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>E). These data indicate that Nestin^high CSC had at least 100-fold greater tumorigenicity than Nestin^low CSC in our experimental system. Taken together, the rapid formation of mammospheres <italic>in vitro</italic> and the efficient induction of solid tumors <italic>in vivo</italic> clearly demonstrated that higher levels of Nestin expression enhanced the tumorigenicity of CSC, which may contribute to the progression and metastasis of triple-negative breast cancer.<fig id="Fig2"><label>Figure 2</label><caption><p><bold>Nestin</bold>^{<bold>high</bold>}<bold>breast CSC have potent tumorigenicity.</bold> The isolated breast CD44⁺CD24^&#x02212; CSC from 12 freshly resected breast cancer tissues were stained with anti-Nestin, and the Nestin⁺ and Nestin^&#x02212; breast CSC were purified as Nestin^high or Nestin^low CSC by flow cytometry, respectively. The levels of Nestin expression in Nestin^high, Nestin^low, and unstained control CSC were determined by western blotting, and these CSC were tested for their capacity to form mammospheres <italic>in vitro</italic> and to induce solid tumors <italic>in vivo</italic>. Finally, the levels of Nestin expression in the solid tumors induced by Nestin^high, Nestin^low, or control CSC were determined by western blotting. Data are representative charts and images, and expressed as the mean&#x02009;&#x000b1;&#x02009;SD of each group of samples (n&#x02009;=&#x02009;12 per group). <bold>(A)</bold> Flow cytometric analysis of CSC. <bold>(B)</bold> Western blot analysis of the levels of Nestin expression. <bold>(C)</bold> Mammosphere formation. <bold>(D)</bold> The rates of tumor formation in SCID mice (n&#x02009;=&#x02009;10 animals per group) from three separate experiments. <bold>(E)</bold> Western blot analysis of Nestin expression in the formed tumors (n&#x02009;=&#x02009;2 to 10 per group). CSC, cancer stem cell; SCID, severe combined immunodeficiency; SD, standard deviation.</p></caption><graphic xlink:href="13058_2014_408_Fig2_HTML" id="d29e1175"/></fig></p></sec><sec id="Sec16"><title>Nestin silencing induces cell cycle arrest and apoptosis in breast CSC</title><p>To further examine the effect of Nestin expression on CSC, CD44⁺CD24^&#x02212; CSC were purified and transfected with pNestin-EGFP, control siRNA, or Nestin-specific siRNA to generate Nestin⁺, Nestin-c, and Nestin-si CSC, respectively. Compared to control Nestin-c CSC, Nestin⁺ CSC expressed higher levels of Nestin while Nestin-si CSC expressed much lower levels of Nestin at 24&#x000a0;h posttransfection, confirming Nestin overexpression and silencing, respectively (Figure&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>A and B). Nestin-si CSC did not effectively form typical mammospheres, while Nestin⁺ CSC under the same conditions formed numerous mammospheres &#x0003e;50&#x000a0;mm in diameter (Figure&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>C), suggesting that increased levels of Nestin expression enhanced the proliferation of breast CSC <italic>in vitro</italic>. Flow cytometry analysis indicated that the percentages of Nestin-si CSC at G2/M phases and apoptotic Nestin-si CSC were significantly higher than that of Nestin-c controls (52.03% vs. 19.99% for cells at G2/M; 43.53&#x02009;&#x000b1;&#x02009;4.78% vs. 16.24&#x02009;&#x000b1;&#x02009;3.22% for apoptotic cells, <italic>P</italic> &#x0003c;0.01 for both, Figure&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>D and E). Thus, knockdown of Nestin expression induced CSC cycle arrest at G2/M and promoted spontaneous CSC apoptosis <italic>in vitro</italic>, further supporting the hypothesis Nestin promoted the proliferation and survival of breast CSC.<fig id="Fig3"><label>Figure 3</label><caption><p><bold>Nestin silencing inhibits the mammosphere formation, induces cell cycle arrest at G2/M and promotes apoptosis in breast CSC.</bold> The isolated CD44⁺CD24^&#x02212; breast CSC from 26 specimens were transfected with control siRNA, Nestin-specific siRNA, pIRES-Nestin-EGFP, or control pIRES-EGFP to generate Nestin-c, Nestin-si, Nestin⁺ or Nestin-EGFP cells. The Nestin-c, Nestin-si, and Nestin⁺ CSC were tested for the levels of Nestin expression by western blotting and for their ability to form mammospheres <italic>in vitro.</italic> Their cell cycling progression and spontaneous apoptosis were determined by flow cytometry. Data are representative charts and images, and expressed as the mean&#x02009;&#x000b1;&#x02009;SD of each group (n&#x02009;=&#x02009;26 per group). <bold>(A)</bold>. Flow cytometric analysis of CD44⁺CD24^&#x02212; CSC. <bold>(B)</bold> Western blot analysis of Nestin expression in breast CSC. Nestin-EGFP and Nestin-c control cells showed similar levels of Nestin expression (data not shown). <bold>(C)</bold> The mammosphere formation. <bold>(D)</bold> Cell cycle analysis of Nestin-si and Nestin-c breast CSC; <bold>(E)</bold> apoptotic Nestin-si and Nestin-c breast CSC. ^*<italic>P</italic> &#x0003c;0.05 vs. the control. CSC, cancer stem cell; SD, standard deviation; siRNA, small interfering RNA.</p></caption><graphic xlink:href="13058_2014_408_Fig3_HTML" id="d29e1269"/></fig></p></sec><sec id="Sec17"><title>Nestin silencing inhibits the migration of breast CSC by downregulating the expression of EMT-related genes</title><p>Cancer metastasis is associated with the poor survival of patients with triple-negative breast cancer. We therefore tested the effect of Nestin silencing on the invasiveness of breast CSC using a transwell migration assay. We found that the numbers of migrated Nestin-si CSC were significantly less than control Nestin-c CSC (<italic>P</italic> &#x0003c;0.05, Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>A), indicating that Nestin silencing inhibited the migration of breast CSC <italic>in vitro</italic>.<fig id="Fig4"><label>Figure 4</label><caption><p><bold>Nestin silencing inhibits breast CSC migration by modulating the expression of EMT-related genes (A).</bold> The migration of Nestin-si and Nestin-c breast CSC. Nestin-si and Nestin-c breast CSC were incubated in the upper chambers of transwell plates at 37&#x000a0;&#x000b0;C for 1.5&#x000a0;h. The migrated cells were stained with crystal violet and counted. Results are presented as mean&#x02009;&#x000b1;&#x02009;SD of the numbers of migrated cells per group of cells. <bold>(B)</bold> and <bold>(C)</bold>. The relative levels of EMT-related protein expression in CSC. Nestin-si and Nestin-c breast CSC were incubated for 48&#x000a0;h, and the relative levels of EMT-related proteins to &#x003b2;-actin were determined by western blot analysis. Data shown are representative images and expressed as the mean&#x02009;&#x000b1;&#x02009;SD of each target protein in each type of cells from six separate experiments. ^*<italic>P</italic> &#x0003c;0.05 vs. the Nestin-c. CSC, cancer stem cell; EMT, epithelial-mesenchymal transition; SD, standard deviation.</p></caption><graphic xlink:href="13058_2014_408_Fig4_HTML" id="d29e1305"/></fig></p><p>Breast CSC migration is commonly associated with spontaneous EMT process. To investigate the potential mechanisms underlying the poor migration of Nestin-si CSC, we analyzed the relative levels of EMT-related molecules and metastatic regulators in Nestin-c and Nestin-si CSC by western blotting. The relative levels of N-cadherin, vimentin, and a-SMA expression were significantly lower in Nestin-si CSC than that in Nestin-c CSC, whereas the levels of E-cadherin were significantly higher (<italic>P</italic> &#x0003c;0.05, Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>B). Furthermore, significantly reduced levels of MMP2, MMP9 and VEGF were detected in Nestin-si CSC, when compared to Nestin-c CSC (<italic>P</italic> &#x0003c;0.05, Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>C). Together, these data indicate that knockdown of Nestin inhibited spontaneous EMT in breast CSC, contributing to their reduced invasiveness.</p></sec><sec id="Sec18"><title>Nestin silencing inhibits the Wnt/&#x003b2;-catenin signaling in breast CSC</title><p>Breast CSC proliferation and migration are regulated by a range of signaling pathways, including the Wnt/&#x003b2;-catenin pathway. We therefore examined the impact of Nestin silencing on the Wnt/&#x003b2;-catenin activation in breast CSC by western blotting (Figure&#x000a0;<xref rid="Fig5" ref-type="fig">5</xref>A). The relative levels of Axin, GSK-3&#x003b2;, and APC in Nestin-si CSC were significantly higher than that in Nestin-c CSC (<italic>P</italic> &#x0003c;0.05, Figure&#x000a0;<xref rid="Fig5" ref-type="fig">5</xref>A). A similar pattern of the relative levels of mRNA transcripts of these genes was detected in Nestin-si and Nestin-c CSC (data not shown). The relative levels of nuclear and cytoplasmic &#x003b2;-catenin, as well as the ratio of nuclear to cytoplasmic &#x003b2;-catenin, were significantly lower in Nestin-si cells than in Nestin-c CSC (<italic>P</italic> &#x0003c;0.05, Figure&#x000a0;<xref rid="Fig5" ref-type="fig">5</xref>A).<fig id="Fig5"><label>Figure 5</label><caption><p><bold>Nestin silencing inhibits the Wnt/&#x003b2;-catenin signaling in breast CSC.</bold> Nestin-si and Nestin-c breast CSC were cultured for 48&#x000a0;h, and the relative levels of target proteins and phosphorylation were determined by western blotting. Data are presented as the mean&#x02009;&#x000b1;&#x02009;SD of the levels of target proteins vs. the &#x003b2;-actin control, or the phosphorylated vs. the total form, from six separate experiments. In addition, Nestin^high in the presence or absence of LGK974 (1 nM), Nestin^low in the presence or absence of SB216763 (5&#x000a0;&#x003bc;M), and control CSC were tested for their proliferation via mammosphere formation assays. Western blot analysis. <bold>(A-C)</bold>. Quantitative analysis of the relative levels of target proteins. <bold>(D)</bold> Quantitative analysis of the formed mammospheres. ^*<italic>P</italic> &#x0003c;0.05, vs. the Nestin-c or control, except for specifically indicated. CSC, cancer stem cell; SD, standard deviation.</p></caption><graphic xlink:href="13058_2014_408_Fig5_HTML" id="d29e1369"/></fig></p><p>The relative levels of c-Myc, cyclin D1, and MMP-7 were significantly lower in Nestin-si CSC than that in Nestin-c CSC, while the relative levels of PPARa and GSK-3&#x003b2; expression and GSK-3&#x003b2; phosphorylation were significantly higher in Nestin-si CSC than that in Nestin-c CSC (<italic>P</italic> &#x0003c;0.05, Figure&#x000a0;<xref rid="Fig5" ref-type="fig">5</xref>B, C). Finally, we employed the mammosphere assay to determine the importance of the Wnt/&#x003b2;-catenin signaling in the proliferation of CSC. We found that treatment with LGK974, an inhibitor of the PORCN-related Wnt/&#x003b2;-catenin signaling pathway, significantly inhibited the formation of Nestin^high-mediated mammospheres; however, treatment with SB216763, an inhibitor of GSK-3&#x003b2;, significantly enhanced the formation of Nestin^low-mediated mammospheres <italic>in vitro</italic> (<italic>P</italic> &#x0003c;0.05 for both, Figure&#x000a0;<xref rid="Fig5" ref-type="fig">5</xref>D). These data indicate that Nestin silencing inhibits Wnt/&#x003b2;-catenin activation, which is crucial for the proliferation of breast CSC.</p></sec></sec><sec id="Sec19"><title>Discussion</title><p>Previous studies have shown that Nestin expression is associated with the disease progression and poor prognosis in breast cancers, particularly for those with advanced lymph node metastasis [<xref ref-type="bibr" rid="CR16">16</xref>],[<xref ref-type="bibr" rid="CR36">36</xref>]. In this study, we examined Nestin expression in 150 specimens of triple-negative breast cancers, and found that 41 (27.33%) out of them were positive for anti-Nestin staining. Importantly, Nestin expression was significantly associated with poor survival in patients with triple-negative breast cancer. To the best of our knowledge, this is the first report of an association between Nestin expression and disease prognosis in Chinese patients with triple-negative breast cancer. These novel data extend previous findings in other populations, and support a role for Nestin in promoting disease progression. Therefore, Nestin may be a therapeutic target and prognostic biomarker for triple-negative breast cancer.</p><p>Nestin is known to be a neural stem cell marker, and is expressed in CSC [<xref ref-type="bibr" rid="CR37">37</xref>], which play important roles in the progression and metastasis of breast cancer [<xref ref-type="bibr" rid="CR16">16</xref>]. In this study, we identified that breast Nestin^high, but not Nestin^low CSC had potent tumorigenicity, as evidenced by rapid mammosphere formation <italic>in vitro</italic> and efficient induction of solid tumors <italic>in vivo</italic>, consistent with our previous observations [<xref ref-type="bibr" rid="CR38">38</xref>]. Similarly, Nestin-overexpressing Nestin⁺, but not Nestin-silenced Nestin-si CSC effectively formed mammospheres. Nestin silencing induced cell cycle arrest at G2/M phrase and promoted the spontaneous apoptosis of Nestin-si CSC. These findings extended previous observations where downregulation of Nestin expression inhibited the proliferation of melanoma cells <italic>in vitro</italic>[<xref ref-type="bibr" rid="CR39">39</xref>], and the growth of implanted pancreatic tumors <italic>in vivo</italic>[<xref ref-type="bibr" rid="CR40">40</xref>]. Our findings suggest that Nestin may enhance the proliferation and survival of breast CSC, contributing to the progression of triple-negative breast cancer.</p><p>Cancer metastasis is associated with poor prognosis of triple-negative breast cancer. In breast CSC, the EMT process is related to increased invasiveness and metastatic potential [<xref ref-type="bibr" rid="CR26">26</xref>]. We observed that Nestin silencing mitigated the invasiveness of Nestin-si CSC, which was associated with significantly reduced levels of N-cadherin, vimentin and a-SMA, but increased levels of E-cadherin in breast CSC. In addition, Nestin silencing significantly reduced the relative levels of MMP2, MMP9, and VEGF in breast CSC. Together, our results indicate that Nestin silencing inhibits spontaneous EMT in breast CSC. Previous studies have showed that increased levels of EMT and expression of stemness markers are associated with progression and metastasis of invasive breast cancer, including triple-negative breast cancer [<xref ref-type="bibr" rid="CR16">16</xref>],[<xref ref-type="bibr" rid="CR36">36</xref>]. Our findings therefore suggest that Nestin may be a potential therapeutic target for the prevention of breast cancer metastasis. We are interested in further investigating the molecular mechanisms by which Nestin regulates the EMT process in breast CSC.</p><p>The Wnt/&#x003b2;-catenin pathway regulates the proliferation and tumorigenicity of stem cells.</p><p>Evidently, induction of Wnt/&#x003b2;-catenin overexpression enhances the tumorigenicity of breast CSC [<xref ref-type="bibr" rid="CR32">32</xref>]. In this study, we found that knockdown of Nestin expression significantly increased the levels of Axin, GSK-3&#x003b2;, and APC, and the ratios of cytoplasmic to nuclear &#x003b2;-catenin in breast CSC. Knockdown of Nestin expression significantly reduced the relative levels of c-Myc, cyclin D1, and MMP-7, but increased the relative levels of inhibitory PPARa in breast CSC. Clearly, Nestin silencing inhibited the Wnt/&#x003b2;-catenin activation in breast CSC, leading to poor proliferation, invasiveness, and cell cycle arrest of CSC. Importantly, we also found that inhibition of PORCN to mitigate spontaneous Wnt secretion and the Wnt/&#x003b2;-catenin activation significantly reduced the proliferation of Nestin^high CSC, while inhibition of GSK-3 promoted the proliferation of Nestin^low CSC <italic>in vitro</italic>. These data clearly indicate that spontaneous Wnt/&#x003b2;-catenin activation was crucial for Nestin to promote CSC proliferation, and further supported the notion that GSK-3&#x003b2; is a negative regulator of the Wnt/&#x003b2;-catenin activation in CSC [<xref ref-type="bibr" rid="CR41">41</xref>]. Since Nestin has no intrinsic phosphatase or protease activity, it is unlikely that Nestin directly mediates the degradation of phosphorylated GSK-3&#x003b2;. Indeed, many pathways, such as the ERK and PI3K/AKT pathways, can phosphorylate GSK-3&#x003b2; and downregulate its activity [<xref ref-type="bibr" rid="CR42">42</xref>],[<xref ref-type="bibr" rid="CR43">43</xref>]. It is possible that Nestin regulates the Wnt/&#x003b2;-catenin pathway by indirectly affecting the activity of GSK-3&#x003b2; in breast CSC. We are interested in further investigating how Nestin regulates GSK-3&#x003b2; and Wnt/&#x003b2;-catenin activation in breast CSC.</p></sec><sec id="Sec20"><title>Conclusions</title><p>Our data indicated that Nestin expression in breast cancer tissues was associated with poor survival of Chinese patients with triple-negative breast cancer. Furthermore, Nestin^high, but not Nestin^low breast CSC, had potent tumorigenicity to form mammospheres <italic>in vitro</italic>, and to induce solid tumors <italic>in vivo</italic>. Knockdown of Nestin expression inhibited the proliferation and invasiveness, but enhanced the spontaneous apoptosis of breast CSC. Nestin silencing inhibited the spontaneous process of EMT and the Wnt/&#x003b2;-catenin activation in breast CSC. Our data therefore suggest that Nestin may be a potential therapeutic target for triple-negative breast cancer. Our findings may provide new insights into the roles of Nestin in regulating key cellular processes in breast CSC. Although this study examined a relatively small number of breast tumor samples, our findings extended previous observations to confirm the importance of Nestin expression in the prognosis of triple-negative breast cancer. While we did not define the precise molecular mechanisms by which Nestin promoted the Wnt/&#x003b2;-catenin activation in breast CSC, our data clearly indicated that spontaneous Wnt/&#x003b2;-catenin activation was crucial for Nestin to promote CSC proliferation, which was negatively regulated by GSK-3&#x003b2;. We are interested in further investigating how Nestin inhibits the growth and metastasis of breast cancers <italic>in vivo</italic>. Therefore, further studies are warranted to validate these findings in a larger population.</p></sec>

Use of ¹H-nuclear magnetic resonance to screen a set of biomarkers for monitoring metabolic disturbances in severe burn patients

Could not extract abstract

<contrib contrib-type="author" corresp="yes"><name><surname>Zhang</surname><given-names>Yong</given-names></name><address><email>zhangyong79@tmmu.edu.cn</email></address><xref ref-type="aff" rid="Aff10"/></contrib><contrib contrib-type="author" corresp="yes"><name><surname>Cai</surname><given-names>Bin</given-names></name><address><email>bin.cai@traumabank.org</email></address><xref ref-type="aff" rid="Aff11"/></contrib><contrib contrib-type="author" corresp="yes"><name><surname>Jiang</surname><given-names>Hua</given-names></name><address><email>cdjianghua@gmail.com</email></address><xref ref-type="aff" rid="Aff11"/></contrib><contrib contrib-type="author"><name><surname>Yan</surname><given-names>Hong</given-names></name><address><email>yrandy@tom.com</email></address><xref ref-type="aff" rid="Aff10"/></contrib><contrib contrib-type="author"><name><surname>Yang</surname><given-names>Hao</given-names></name><address><email>hao.yang@traumabank.org</email></address><xref ref-type="aff" rid="Aff11"/></contrib><contrib contrib-type="author"><name><surname>Peng</surname><given-names>Jin</given-names></name><address><email>pengjin@scu.edu.cn</email></address><xref ref-type="aff" rid="Aff11"/></contrib><contrib contrib-type="author"><name><surname>Wang</surname><given-names>Wenyuan</given-names></name><address><email>wenyuan.wang@aibai.org</email></address><xref ref-type="aff" rid="Aff11"/></contrib><contrib contrib-type="author"><name><surname>Ma</surname><given-names>Siyuan</given-names></name><address><email>tmmu_msy@126.com</email></address><xref ref-type="aff" rid="Aff10"/></contrib><contrib contrib-type="author"><name><surname>Wu</surname><given-names>Xiuwen</given-names></name><address><email>wxwhappy2004@163.com</email></address><xref ref-type="aff" rid="Aff10"/></contrib><contrib contrib-type="author" corresp="yes"><name><surname>Peng</surname><given-names>Xi</given-names></name><address><email>pxlrmm@tmmu.edu.cn</email></address><xref ref-type="aff" rid="Aff10"/></contrib><aff id="Aff10"><label/>State Key Laboratory of Trauma, Burns and Combined Injury, Institute of Burns of PLA, Southwest Hospital, Third Military Medical University, Chongqing, 400038 People&#x02019;s Republic of China </aff><aff id="Aff11"><label/>Department of Computational Mathematics and Biostatistics, Metabolomics and Multidisciplinary Laboratory for Trauma Research, Sichuan Provincial People&#x02019;s Hospital, Sichuan Academy of Medical Sciences, No. 585, Da Mian Hong He Bei Lu, Chengdu, Sichuan Province 610101 China </aff>

Critical Care

<sec id="Sec1" sec-type="intro"><title>Introduction</title><p>Burn is a common injury with an incidence of about 0.2% in the normal population. Every year, approximately 3 million people in China and 0.8 million in the United States suffer from burns, with 200,000 and 40,000 requiring hospitalization, respectively [<xref ref-type="bibr" rid="CR1">1</xref>, <xref ref-type="bibr" rid="CR2">2</xref>]. In addition, more than one-third of burn patients are children under 14 years of age [<xref ref-type="bibr" rid="CR1">1</xref>, <xref ref-type="bibr" rid="CR3">3</xref>]. Therefore, the treatment course for burns is not only a public healthcare issue, but also a relevant matter in the growth and future of children.</p><p>Mild burn is easy to treat, and the cure rate is 95% or greater worldwide. However, severe burn, which covers more than 50% of the total body surface area (TBSA), is very difficult to treat, and the mortality rate is usually more than 30%. Among the extremely severe burn patients for whom more than 80% of the TBSA is burned, the death rate can reach 70% or higher [<xref ref-type="bibr" rid="CR1">1</xref>]. Although much research has been done and numerous advances have been made through the hard work of a generation of burn surgeons and scientists, the mortality of severe burn patients has not changed in the past decade [<xref ref-type="bibr" rid="CR4">4</xref>&#x02013;<xref ref-type="bibr" rid="CR6">6</xref>]. Determining how to reduce the mortality and improve the care of severe burn patients is a core issue in burn research. After severe burn injury, along with massive damage to the skin and subcutaneous tissue, multiple organs are also damaged. Pathophysiological conditions are complicated and are highly related to metabolic regulation [<xref ref-type="bibr" rid="CR7">7</xref>&#x02013;<xref ref-type="bibr" rid="CR9">9</xref>]. Therefore, understanding the complicated changes in metabolic networks is essential for developing the next generation of prognosis prediction tools and new treatment methods. However, metabolic regulatory networks involve large numbers of molecules and pathways. Conventional laboratory testing only includes a few of metabolic parameters and cannot measure global changes in metabolic networks in real-time. A metabolomics test based on ¹H-nuclear magnetic resonance (NMR) provides a unique high-throughput solution to resolve this challenge. It can be used to detect most small metabolic molecules in a single-use test [<xref ref-type="bibr" rid="CR10">10</xref>&#x02013;<xref ref-type="bibr" rid="CR12">12</xref>]. By using advanced mathematical modeling, researchers can visualize the global changes in metabolic networks (metabolic profile or metabolome) and extract a set of biomarkers. These biomarkers offer a new approach to quantitative, real-time monitoring for severe burn patients and would give clinical practitioners new opportunities to make better informed decisions.</p><p>One of the major challenges in analyzing NMR data from plasma samples is the high-dimension disaster of metadata. The only solution to address this challenge is to use a pattern recognizing technique. Principal component analysis (PCA) and partial least square (PLS) are two common algorithms that can be used for dimension reduction in NMR data analysis. Compared to PCA, PLS considers correlations between variables. Hence, both PCA and PLS are used as conventional mathematical tools in NMR data analysis. In our previous studies, we successfully used PCA and PLS to fit data according to the severity of spinal cord injury [<xref ref-type="bibr" rid="CR13">13</xref>, <xref ref-type="bibr" rid="CR14">14</xref>]. We have reasonable confidence that these algorithms can be used to establish a metabolomic profile for severe burn patients, who suffer much greater metabolic disturbances. In addition, with the release of the Human Metabolome Database (HMDB), matching peaks to metabolites is now becoming much easier than before [<xref ref-type="bibr" rid="CR15">15</xref>]. In brief, after peaks are screened using PCA and PLS, we can submit these peaks to HMDB and identify related metabolites. In the present study, by using a high-resolution NMR technique, we aimed to establish a plasma metabolomics fingerprint spectrum of severe burn patients and to use it to identify a set of biomarkers that can be used for clinical monitoring and to better understand metabolism disturbances after burns. With this effort, we expect to lay a foundation for formulating reasonable improvements to future treatment protocols.</p></sec><sec id="Sec2" sec-type="materials|methods"><title>Materials and methods</title><sec id="Sec3"><title>Subjects</title><p>Subjects were 21 adult severe burn patients admitted to the Institute of Burn Research of the Southwest Hospital of The Third Military Medical University between May 2012 and December 2012. Patients were recruited if they met the inclusion criteria of being between 18 and 65 years of age and having a burn area covering more than 50% TBSA. The exclusion criteria were: (1) special burns including chemical and electrical burns; (2) severe complications such as heart disease, hepatic disease, renal disease, and hematopoietic disease before burn; (3) oncologic disease; (4) history of endocrine disease including diabetes and hyperthyroidism; (5) obesity (body mass index &#x0003e;25 kg/m²); (6) pregnancy or lactation; (7) psychiatric disorder or mental state leading to failure to cooperate, inability for self-control, or trouble communicating; and (8) participation in other clinical trials.</p><p>Written informed consent was obtained from all participants, and the Committee of Medical Ethics of the Southwest Hospital of The Third Military Medical University approved the study protocol (approval number: KY201118).</p></sec><sec id="Sec4"><title>Clinical course of severe burn patients</title><p>All patients were admitted 2 to 24 hours post burn. When a patient was admitted, we applied the standardized fluid resuscitation protocol according to the Chinese Medical Association burn treatment guidelines to treat burn shock immediately. Silver sulfadiazine was applied to the wound, and systemic antibiotics were used to prevent infection. Escharectomy and skin grafting were performed three days after burn and three to four times within one month to help cover the wound surface.</p></sec><sec id="Sec5"><title>Collection and preparation of blood samples</title><p>Healthy controls were kept off food and water before blood collection at 8 a.m. Two milliliters of blood was collected from the median cubital vein using a citrate vacuum tube. The samples were centrifuged at 3,000 rpm for 10 minutes immediately, after which 1 ml of supernatant plasma was extracted. The supernatant was stored at -80&#x000b0;C until analysis. For severe burn patients, fasting blood samples were collected at 8 a.m. on the first morning after admission (24 to 48 hours post burn) and then processed as described for the controls.</p><p>Plasma samples were defrosted at room temperature and centrifuged at 16,000 rpm for 10 minutes. Then 450 &#x003bc;l supernatant was extracted from each sample and fully mixed with 50 &#x003bc;l deuterium oxide (D₂O) for 120 seconds. After standing for 10 minutes, samples were analyzed using 600-MHz NMR spectrometry.</p></sec><sec id="Sec6"><title>NMR measurements and data analysis</title><sec id="Sec7"><title>NMR measurements</title><p>We employed NMR measurements according to a protocol that was established and reported previously [<xref ref-type="bibr" rid="CR13">13</xref>]. All one-dimensional spectra were acquired at 600.13 MHz using a Bruker Avance DRx 600 600-MHz spectrometer (Bruker BioSpin GmbH, Rheinstetten, Germany) equipped with a proton observation probe (Bruker BBI inverse-broadband probe). Spectra were recorded at a room temperature of 300 K. Standard one-dimensional pulse sequences and Carr-Purcell-Meiboom-Gill (CPMG) sequences were used. A spin-spin relaxation delay of 64 ms was used for all samples, and water suppression irradiation was applied during the relaxation delay (2 s). Typically, in the standard one-dimensional and CPMG experiments, the spectral width was 20 ppm and 256 transients were collected into 32 k data points. CPMG experiments filter broad resonances from proteins and lipids, permitting latent biomarkers of smaller molecular weight to be visualized.</p></sec><sec id="Sec8"><title>Data processing</title><p>Clinical data were described as mean&#x02009;&#x000b1;&#x02009;standard deviation (SD) or as median and interquartile range (IQR) in the case of a skewed distribution. Differences between groups were assessed with the Student&#x02019;s <italic>t</italic> test for data presented as means. Differences in counts or percentages were evaluated with the Fisher&#x02019;s exact probability test. Differences were considered significant if a two-tailed <italic>P</italic> value was &#x0003c;0.05.</p><p>All plasma ¹H-NMR spectra were phased and baseline corrected within mestReC (version 4.9.9.9, Mestrelab Research SL, Rheinstetten, Germany), and the chemical shifts were referenced to a creatinine peak at &#x00394;3.05. These data were introduced into a Matlab (R2012b, The MathWorks, Inc, Natick, MA, USA) data structure, where each row comprised the integral descriptors for an NMR spectrum. To reduce the interference of huge water peaks, all spectra were analyzed to non-normalized data after removal of the spectral region containing the suppressed water resonance.</p></sec><sec id="Sec9"><title>Pattern recognition</title><p>All multivariate statistics and pattern recognition were performed using the Eigen victor toolbox (ver6.2.1) with two techniques: PCA and PLS on the Matlab. Before analyzing, scaling was applied to minimize the variation of the ¹H-NMR peak to ensure that the large peak did not overshadow the contribution of the small one. PCA score plots were constructed to visualize the inherent clustering of the samples based on burning. The toolbox can export the Q² value, which indicates how well the model predicts new data. A large Q² (&#x0003e;0.5) indicates good predictive capability.</p><p>For further analysis, PLS-discriminant analysis (PLS-DA) was used in the data processing. PLS is used to find the fundamental relationship between two matrices (X and Y), that is, a latent variable approach to modeling the covariance structures in these two spaces. Here the X is a 200&#x02009;&#x000d7;&#x02009;24 matrix, in which each row represents the integral value of the NMR spectrum of each patient, and Y represents the patient's condition where 1 indicates burn and 0 indicates health. A PLS model will try to find the multidimensional direction in the X space that explains the maximum multidimensional variance direction in the Y space; that is, it will try to find the spectrum variables in X that can explain the result of burn or health in Y. PLS is particularly suited when the matrix of predictors has more variables than observations.</p><p>In order to avoid excessive classification, we further adopted cross-validation (CV) to evaluate the stability of the model. We addressed the validation by cutting a single observation from the original sample as the validation data and the remaining observations as the training data. Each observation in the sample is used once as the validation data in turn. The Q² value represents the percentage of the variation in the dataset predicted by the model according to CV, that is, the Q² value represents the discriminating ability of the PPM of a particular segment. The formula for Q² is as follows: <disp-formula id="Equa"><graphic xlink:href="13054_2013_2899_Equa_HTML.gif" position="anchor"/></disp-formula></p><p>Here PRESS is the predictive residual sum of squares, and SSY is the sum of squares of the Y matrix. These measures can be equivalently expressed as standard error of prediction (SDEP or SEP), or standard error of CV (SECV).</p><p>Here, we get R²&#x02009;=&#x02009;0.87, Q²&#x02009;=&#x02009;0.76, and SECV&#x02009;=&#x02009;0.201 with SD&#x02009;=&#x02009;0.225</p><p>SECV is closed to the SD of the X matrix. It can be interpreted as the SEVC is closed to the NMR spectrum noise, so the stability of our model is acceptable.</p><p>Supporting vector machines (SVMs) have been successfully applied to various scientific problems, particularly in high-dimensional data, and a SVM generally achieves classification performance superior to that of many older methods. We employed a kernel function from quadratic, polynomial kernel, Gaussian Radial Basis, and multilayer perceptron to classify PLS scores.</p></sec><sec id="Sec10"><title>Establishing metabolome and gene function analysis</title><p>The HMDB was used to identify key metabolites related to enzymes and upstream genes. In order to determine the common functions of these metabolites, we used the Gene Ontology terminology (GO) system to analyze the enrichment condition of above selected enzymes (and the corresponding genes). All GO analysis was conducted using the G: profiler website [<xref ref-type="bibr" rid="CR16">16</xref>]. According to the website, the core algorithm in the program is the widely applied hypergeometric distribution for significance of the estimation principle for functional genomics of enrichment and analysis.</p></sec><sec id="Sec11"><title>Computing platform and tools</title><p>All computation processes were conducted using a high-performance computing platform (HPC, CPU XEON E7-8870 2.4G 6.4GT/s30M10C *4, GPU TESLA K20 5GB GENERIC, 512GB DDR3 1333MHz R-ECC; Environment: Unbutu12.04) of the Metabolomics and Multidisciplinary Laboratory of Sichuan Academy of Medical Sciences with computing software Matlab 2012b.</p></sec></sec></sec><sec id="Sec12" sec-type="results"><title>Results</title><sec id="Sec13"><title>Patients</title><p>Of the 21 participants initially recruited, none withdrew consent. The patients&#x02019; average age was 43.2&#x02009;&#x000b1;&#x02009;10.7 years, and they were admitted within 24 hours after injury to the participating hospitals. The patients&#x02019; average percentage of TBSA was 77&#x02009;&#x000b1;&#x02009;12% (IQR 55 to 97%) and percentage of full thickness surface area (FTSA) was 45&#x02009;&#x000b1;&#x02009;24% (IQR 5 to 95%). All patients were immediately given antishock fluid resuscitation upon admission, and all interventions were in accordance with the burn treatment guidelines issued by the Chinese Burns Medical Association. Four patients with severe burns died of multiple organ failure and sepsis. The overall mortality rate during the study period was 19%.</p></sec><sec id="Sec14"><title>Clinical assessments</title><p>All of the variables followed a normal distribution. Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref> demonstrates that the two groups were comparable for basic demographic data. The subjects were similar in age and body weight. However, there were significant differences between the groups in the percentage of TBSA of burns, breathing rate, blood pressure (BP), pulse, and temperature.<table-wrap id="Tab1"><label>Table 1</label><caption><p> <bold>Comparison of clinical data between severe burn patients and controls/</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr><th>Variables</th><th>Control (n&#x02009;=&#x02009;3)</th><th>Case (n&#x02009;=&#x02009;21)</th><th> <bold><italic>P</italic></bold> value</th></tr></thead><tbody><tr><td>Age (yr)</td><td align="center">45.1&#x02009;&#x000b1;&#x02009;7.4</td><td align="center">43.2&#x02009;&#x000b1;&#x02009;10.7</td><td align="center"> <italic>&#x0003e;0.05</italic> </td></tr><tr><td>Sex, male (female)</td><td align="center">2 (1)</td><td align="center">16 (5)</td><td align="center"> <italic>&#x0003e;0.05</italic> </td></tr><tr><td>Weight (kg)</td><td align="center">61&#x02009;&#x000b1;&#x02009;4</td><td align="center">64&#x02009;&#x000b1;&#x02009;13</td><td align="center"> <italic>&#x0003e;0.05</italic> </td></tr><tr><td>TBSA burn (%)</td><td align="center">0</td><td align="center">77&#x02009;&#x000b1;&#x02009;12</td><td align="center"/></tr><tr><td>Second-degree burn (%)</td><td align="center">0</td><td align="center">37&#x02009;&#x000b1;&#x02009;23</td><td align="center"/></tr><tr><td>Third-degree burn (%)</td><td align="center">0</td><td align="center">45&#x02009;&#x000b1;&#x02009;24</td><td align="center"/></tr><tr><td>Breathing rate (times/min)</td><td align="center">18.3&#x02009;&#x000b1;&#x02009;2.2</td><td align="center">22.2&#x02009;&#x000b1;&#x02009;2.1</td><td align="center"> <italic>&#x0003c;0.05</italic> </td></tr><tr><td>BP (mmHg)</td><td align="center">108&#x02009;&#x000b1;&#x02009;14/75&#x02009;&#x000b1;&#x02009;9</td><td align="center">123&#x02009;&#x000b1;&#x02009;18/82&#x02009;&#x000b1;&#x02009;11</td><td align="center"> <italic>&#x0003c;0.01</italic> </td></tr><tr><td>Pulse (times/min)</td><td align="center">78&#x02009;&#x000b1;&#x02009;12</td><td align="center">113&#x02009;&#x000b1;&#x02009;18</td><td align="center"> <italic>&#x0003c;0.01</italic> </td></tr><tr><td>T (&#x000b0;C)</td><td align="center">36&#x02009;&#x000b1;&#x02009;0.3</td><td align="center">36.8&#x02009;&#x000b1;&#x02009;0.9</td><td align="center"> <italic>&#x0003c;0.05</italic> </td></tr></tbody></table><table-wrap-foot><p>Data are expressed as n (%) or mean&#x02009;&#x000b1;&#x02009;standard deviation. TBSA, total body surface area; BP, blood pressure; T, temperature.</p></table-wrap-foot></table-wrap></p></sec><sec id="Sec15"><title>Plasma metabolome after severe burn</title><p>Typical 600.13-MHz NMR spectra demonstrated resonances arising from metabolites including glucose, histidine, and creatine. The differences between spectra from the severe burn patients and those from the controls were obvious on visual inspection (Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>), which demonstrates that there were significant alterations in the plasma metabolite profiles.<fig id="Fig1"><label>Figure 1</label><caption><p> <bold>Comparison of</bold> <bold>nuclear magnetic resonance</bold> <bold>(NMR) spectra from healthy controls and burn patients.</bold> The blue line is the ¹H-NMR plasma spectrum of healthy controls, and the red line is the NMR plasma spectrum of burn patients.</p></caption><graphic xlink:href="13054_2013_2899_Fig1_HTML" id="d30e799"/></fig></p><p>The variable importance in the projection (VIP) represents the value of each predictor in fitting the PLS model for both predictors and responses, and we used the method that was developed by Chong and Jun [<xref ref-type="bibr" rid="CR17">17</xref>] to calculate VIP scores. The VIP indicator can describe correlations between the variable (X) and response (Y). We used the VIP to identify metabolites correlated with severe burns and named these metabolites as the 'Eigen-metabolome&#x02019; of severe burns. We used a VIP score &#x0003e;1.5 as a threshold to obtain the determinant metabolites [<xref ref-type="bibr" rid="CR18">18</xref>]. Then we used the HMDB to identify 12 metabolites that are catalyzed by 103 enzymes (Tables&#x000a0;<xref rid="Tab2" ref-type="table">2</xref> and <xref rid="Tab3" ref-type="table">3</xref> and Figures&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref> and <xref rid="Fig3" ref-type="fig">3</xref>). These 12 metabolites represent the major metabolic changes that occur after severe burn and can be used as the Eigen-metabolome.<table-wrap id="Tab2"><label>Table 2</label><caption><p> <bold>Summary of Eigen-metabolome: metabolites related to severe burn</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr><th>HMDB metabolite</th><th colspan="7">HMDB enzyme - gene symbol</th></tr></thead><tbody><tr><td>3-Methylhistidine</td><td>CNDP1</td><td>PRMT3</td><td/><td/><td/><td/><td/></tr><tr><td>1,3-Diaminopropane</td><td>AOC2</td><td>AMD1</td><td>AOC3</td><td>SMS</td><td>DHPS</td><td>ABP1</td><td>ODC1</td></tr><tr><td>2-Hydroxybutyric acid</td><td>DLD</td><td>LDHB</td><td>LDHAL6B</td><td>LDHC</td><td>LDHAL6A</td><td>TDH</td><td/></tr><tr><td rowspan="3">2-Methoxyestrone</td><td>UGT1A1</td><td>UGT2B11</td><td>UGT2A3</td><td>UGT2B10</td><td>UGT1A5</td><td>SHBG</td><td>UGT1A8</td></tr><tr><td>UGT2B15</td><td>UGT1A7</td><td>UGT2B7</td><td>UGT2B4</td><td>UGT2B28</td><td>UGT1A3</td><td>UGT1A6</td></tr><tr><td>UGT2B17</td><td>COMT</td><td>UGT1A4</td><td>UGT2A1</td><td>UGT1A9</td><td>UGT1A10</td><td/></tr><tr><td rowspan="2">Deoxycorticosterone</td><td>HSD3B1</td><td>P450-CYP21B</td><td>CYP11B2</td><td>NR3C2</td><td/><td/><td/></tr><tr><td>HSD3B2</td><td>CYP11B1</td><td>CYP21A2</td><td>NR3C1</td><td/><td/><td/></tr><tr><td>Alpha ketoisovaleric acid</td><td>-</td><td/><td/><td/><td/><td/><td/></tr><tr><td>Iodotyrosine</td><td>TPO</td><td/><td/><td/><td/><td/><td/></tr><tr><td rowspan="2">Biotin</td><td>PCCA</td><td>SLC5A6</td><td>PC</td><td>PCCB</td><td>ACACA</td><td>MCCC1</td><td/></tr><tr><td>MCCC2</td><td>DKFZp686B20267</td><td>HLCS</td><td>ACACB</td><td>BTD</td><td/><td/></tr><tr><td>7-Dehydrocholesterol</td><td>SC5DL</td><td>HMGCS2</td><td>SCP2</td><td>DHCR24</td><td>CYP11A1</td><td>DHCR7</td><td/></tr><tr><td rowspan="3">Aldosterone</td><td>MLPH</td><td>SGK1</td><td>NR3C2</td><td>NPPB</td><td>CYP11B1</td><td/><td/></tr><tr><td>FN1</td><td>CTGF</td><td>NR3C1</td><td>AKR1D1</td><td>ADM</td><td/><td/></tr><tr><td>CYP11B2</td><td>AGTR1</td><td>PTGER4</td><td>EGFR</td><td>PRKD1</td><td/><td/></tr><tr><td rowspan="2">Dihydrobiopterin</td><td>TYR</td><td>TH</td><td>NOS3</td><td>DHFR</td><td>PCBD1</td><td/><td/></tr><tr><td>TPH1</td><td>QDPR</td><td>SPR</td><td>NOS1</td><td/><td/><td/></tr><tr><td rowspan="2">Butyric acid</td><td>HDAC1</td><td>TNF</td><td>PPARG</td><td>ACSM5</td><td>HDAC4</td><td>ACSM2A</td><td/></tr><tr><td>HDAC5</td><td>HDAC2</td><td>ACSM4</td><td>ACSM1</td><td>SLC16A1</td><td>HDAC3</td><td/></tr><tr><td/><td>ACSM2B</td><td>ACSM6</td><td>CASP3</td><td>ACSM3</td><td>HDAC9</td><td/><td/></tr></tbody></table><table-wrap-foot><p>Twelve characteristic metabolites found in nuclear magnetic resonance (NMR) metabolic spectra. They are closely related to substance metabolism, skeletal muscle and fat catabolism, or viscus functional disorder after severe burn injury. HMDB, Human Metabolome Database.</p></table-wrap-foot></table-wrap><table-wrap id="Tab3"><label>Table 3</label><caption><p> <bold>Biological processes associated with the 12 characteristic metabolites</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr><th> <bold><italic>P</italic></bold> value</th><th>T</th><th>Q</th><th>Q&#x00026;T</th><th>Q&#x00026;T/Q</th><th>Q&#x00026;T/T</th><th>Term ID</th><th>Main function</th><th>Gene dosage</th><th>Functional description</th></tr></thead><tbody><tr><td>7.18E-05</td><td align="center">42</td><td align="center">89</td><td align="center">6</td><td align="center">0.067</td><td align="center">0.143</td><td>GO:0006476</td><td align="center">BP</td><td align="center">7</td><td>Protein deacetylation</td></tr><tr><td>1.69E-02</td><td align="center">1274</td><td align="center">87</td><td align="center">18</td><td align="center">0.207</td><td align="center">0.014</td><td>GO:0009611</td><td align="center">BP</td><td align="center">80</td><td>Response to wounding</td></tr><tr><td>1.67E-04</td><td align="center">30</td><td align="center">72</td><td align="center">5</td><td align="center">0.069</td><td align="center">0.167</td><td>GO:0042312</td><td align="center">BP</td><td align="center">31</td><td>Regulation of vasodilation</td></tr><tr><td>3.43E-03</td><td align="center">28</td><td align="center">60</td><td align="center">4</td><td align="center">0.067</td><td align="center">0.143</td><td>GO:0055078</td><td align="center">BP</td><td align="center">77</td><td>Sodium ion homeostasis</td></tr><tr><td>1.22E-02</td><td align="center">2</td><td align="center">45</td><td align="center">2</td><td align="center">0.044</td><td align="center">1</td><td>GO:2001295</td><td align="center">BP</td><td align="center">41</td><td>Malonyl-coenzyme A biosynthetic process</td></tr><tr><td>1.80E-02</td><td align="center">32</td><td align="center">79</td><td align="center">4</td><td align="center">0.051</td><td align="center">0.125</td><td>GO:0048662</td><td align="center">BP</td><td align="center">52</td><td>Negative regulation of smooth muscle cell proliferation</td></tr><tr><td>4.14E-02</td><td align="center">89</td><td align="center">72</td><td align="center">5</td><td align="center">0.069</td><td align="center">0.056</td><td>GO:0046209</td><td align="center">BP</td><td align="center">45</td><td>Nitric oxide metabolic process</td></tr><tr><td>4.93E-33</td><td align="center">944</td><td align="center">88</td><td align="center">45</td><td align="center">0.511</td><td align="center">0.048</td><td>GO:0019752</td><td align="center">BP</td><td align="center">2</td><td>Carboxylic acid metabolic process</td></tr><tr><td>3.23E-02</td><td align="center">41</td><td align="center">71</td><td align="center">4</td><td align="center">0.056</td><td align="center">0.098</td><td>GO:0050999</td><td align="center">BP</td><td align="center">59</td><td>Regulation of nitric oxide synthase activity</td></tr><tr><td>2.75E-02</td><td align="center">83</td><td align="center">71</td><td align="center">5</td><td align="center">0.07</td><td align="center">0.06</td><td>GO:0051341</td><td align="center">BP</td><td align="center">53</td><td>Regulation of oxidoreductase activity</td></tr><tr><td>3.31E-02</td><td align="center">72</td><td align="center">85</td><td align="center">5</td><td align="center">0.059</td><td align="center">0.069</td><td>GO:0006096</td><td align="center">BP</td><td align="center">51</td><td>Glycolysis</td></tr><tr><td>4.64E-15</td><td align="center">7548</td><td align="center">88</td><td align="center">73</td><td align="center">0.83</td><td align="center">0.01</td><td>GO:0044444</td><td align="center">CC</td><td align="center">33</td><td>Cytoplasmic part</td></tr><tr><td>1.14E-02</td><td align="center">3100</td><td align="center">89</td><td align="center">31</td><td align="center">0.348</td><td align="center">0.01</td><td>GO:0031974</td><td align="center">CC</td><td align="center">74</td><td>Membrane-enclosed lumen</td></tr><tr><td>5.29E-05</td><td align="center">40</td><td align="center">89</td><td align="center">6</td><td align="center">0.067</td><td align="center">0.15</td><td>GO:0000118</td><td align="center">CC</td><td align="center">4</td><td>Histone deacetylase complex</td></tr><tr><td>2.54E-29</td><td align="center">33</td><td align="center">31</td><td align="center">14</td><td align="center">0.452</td><td align="center">0.424</td><td>GO:0015020</td><td align="center">MF</td><td align="center">21</td><td>Glucuronosyltransferase activity</td></tr><tr><td>7.80E-04</td><td align="center">185</td><td align="center">72</td><td align="center">8</td><td align="center">0.111</td><td align="center">0.043</td><td>GO:0005506</td><td align="center">MF</td><td align="center">27</td><td>Iron ion binding</td></tr><tr><td>1.94E-02</td><td align="center">3</td><td align="center">33</td><td align="center">2</td><td align="center">0.061</td><td align="center">0.667</td><td>GO:0004769</td><td align="center">MF</td><td align="center">81</td><td>Steroid delta-isomerase activity</td></tr><tr><td>1.05E-14</td><td align="center">58</td><td align="center">79</td><td align="center">12</td><td align="center">0.152</td><td align="center">0.207</td><td>GO:0033293</td><td align="center">MF</td><td align="center">60</td><td>Monocarboxylic acid binding</td></tr><tr><td>7.06E-09</td><td align="center">11</td><td align="center">89</td><td align="center">6</td><td align="center">0.067</td><td align="center">0.545</td><td>GO:0031078</td><td align="center">MF</td><td align="center">62</td><td>Histone deacetylase activity (H3-K14-specific)</td></tr><tr><td>3.84E-02</td><td align="center">149</td><td align="center">72</td><td align="center">6</td><td align="center">0.083</td><td align="center">0.04</td><td>GO:0020037</td><td align="center">MF</td><td align="center">54</td><td>Heme binding</td></tr><tr><td>4.77E-04</td><td align="center">1195</td><td align="center">89</td><td align="center">20</td><td align="center">0.225</td><td align="center">0.017</td><td>GO:0019899</td><td align="center">MF</td><td align="center">9</td><td>Enzyme binding</td></tr><tr><td>3.76E-06</td><td align="center">614</td><td align="center">68</td><td align="center">15</td><td align="center">0.221</td><td align="center">0.024</td><td>GO:0042803</td><td align="center">MF</td><td align="center">38</td><td>Protein homodimerization activity</td></tr></tbody></table><table-wrap-foot><p>Biological processes associated with the 12 characteristic metabolites.</p></table-wrap-foot></table-wrap><fig id="Fig2"><label>Figure 2</label><caption><p> <bold>A high-resolution</bold> ^{<bold>1</bold>} <bold>H-</bold> <bold>nuclear magnetic resonance</bold> <bold>spectrum of plasma demonstrating spectral assignments.</bold> Only the major metabolites are labeled.</p></caption><graphic xlink:href="13054_2013_2899_Fig2_HTML" id="d30e2032"/></fig><fig id="Fig3"><label>Figure 3</label><caption><p> <bold>Automatic separation of the sample score plot by support vector machine (SVM).</bold> The kernel function of the SVM is quadratic <bold>(A)</bold>, polynomial kernel <bold>(B)</bold>, Gaussian Radial Basis <bold>(C)</bold>, and multilayer perceptron <bold>(D)</bold>. The black line is the separating line between the burn (+) and healthy (*) samples.</p></caption><graphic xlink:href="13054_2013_2899_Fig3_HTML" id="d30e2056"/></fig></p><p>We were interested to use the Eigen-metabolome to establish a quantitative burn evaluation model. We employed a PLS regression model to establish a linear prediction model.</p><p>Then we obtained the discriminant equations for the relationship between plasma metabolites and injury severity: <disp-formula id="Equb"><graphic xlink:href="13054_2013_2899_Equb_HTML.gif" position="anchor"/></disp-formula></p><p>where x represents the ppm value from the NMR spectra, and a_ij represents the loadings. Finally, we obtained an injury severity discriminant model based on a SVM. We found that SVM equations successfully distinguished severe burn patients and healthy control individuals.</p></sec></sec><sec id="Sec16" sec-type="discussion"><title>Discussion</title><p>In this study, we found that the metabolomics fingerprint of severe burn patients was altered significantly. Twelve small molecular metabolites (Table&#x000a0;<xref rid="Tab2" ref-type="table">2</xref>) make up an Eigen-metabolome that distinguishes severe burn patients from healthy controls. Hence, this Eigen-metabolome comprises a set of biomarkers that can be used to monitor the metabolism disturbances after severe burn injury. To the best of our knowledge, this is the first study on human metabolomics fingerprinting after severe burn. In addition, we identified several interesting findings regarding metabolic pathway regulatory changes in metabolomics fingerprinting.</p><p>One very interesting metabolite included in the severe burn Eigen-metabolome was &#x003b1;-ketoisovaleric acid, an intermediate metabolite of valine, that is regarded as a marker of mitochondrial damage [<xref ref-type="bibr" rid="CR19">19</xref>, <xref ref-type="bibr" rid="CR20">20</xref>]. Valine can be converted to &#x003b1;-ketoisovaleric acid through a deamination reaction by branched-chain amino acid aminotransferase, and it is carried from the cytosol to mitochondria through the mitochondrial membrane transporter, where it is further converted to succinyl-coenzyme A by acyl-coenzyme A dehydrogenase and participates in the tricarboxylic acid cycle. Ketoisovaleric acid accumulates with disorder of the mitochondrial membrane transport system [<xref ref-type="bibr" rid="CR21">21</xref>, <xref ref-type="bibr" rid="CR22">22</xref>]. In this study, we found that the plasma ketoisovaleric acid level significantly increased after burn injury, indicating cell membrane damage and mitochondrial transport dysfunction in the early stage of burn injury. Although our and others&#x02019; previous studies have demonstrated that mitochondrial dysfunction can be found one day post burn, those data were derived from animal tissues [<xref ref-type="bibr" rid="CR23">23</xref>&#x02013;<xref ref-type="bibr" rid="CR26">26</xref>]. To date, there is no clinically available tool to monitor mitochondrial function. The present study indicates that through ¹H-NMR metabolic fingerprinting, &#x003b1;-ketoisovaleric acid can be used clinically as a new biomarker of mitochondrial dysfunction.</p><p>Researchers have found that the 3-methylhistidine (3-MH) level reflects skeleton muscle degeneration after burn. A clinical study indicated that urine 3-MH is significantly increased in burned children [<xref ref-type="bibr" rid="CR27">27</xref>]. It is seen as an important marker of catabolic metabolism. The traditional way to examine skeleton muscle decomposition is to detect 3-MH by high performance liquid chromatography (HPLC) and/or mass spectrometry (MS), but these methods are expensive and time-consuming. Because of the complexity of plasma contents, previous studies all used urine samples. The results from urine testing are not as accurate as those from plasma testing, which limits the use of 3-MH as a clinical marker. Our study found that as a high-throughput method, the ¹H-NMR metabolome accurately and timely duplicated results that only could be examined in complicated laboratory studies previously. Because of its cost-effectiveness, ¹H-NMR metabolome fingerprinting could be used as a sensitive monitoring tool for skeleton catabolism after severe burn.</p><p>After severe burn, the stress process is followed and represented by the release of a large amount of stress-related hormones and cytokines. This stress process leads to mal-metabolism of carbohydrates via insulin resistance and hyperglycemia [<xref ref-type="bibr" rid="CR28">28</xref>]. For a very long time, researchers and clinical practitioners have considered hyperglycemia post severe burn as a type of stress-related phenomena and quite different from diabetes. They believe that this type of hyperglycemia is not related to ketonemia during the early stage of burn and that ketonemia only occurs when patients are suffering from sepsis [<xref ref-type="bibr" rid="CR29">29</xref>]. Our study explored a different scenario: &#x003b2;-hydroxybutyric acid was increased in the metabolome of patients in the early stage post burn. Considering that &#x003b2;-hydroxybutyric acid is the major component of ketones (75% of ketones), the increasing level of &#x003b2;-hydroxybutyric acid indicated that ketogenic metabolism is enhanced by fatty acid decomposition in liver.</p><p>Through the above analysis of key metabolites from the Eigen-metabolome, we conclude that mitochondrial function and carbohydrate, protein, and fatty acid metabolism are significantly changed during the early stage of severe burn. The core cause of these types of changes is the decomposition of skeleton muscle and fat tissue to provide substrates for gluconeogenesis and ketogenic metabolism. The outcome of this metabolic adjustment is to fulfill the energy needs of brain and myocardial cells under stress conditions. All of this metabolic information could be obtained from a ¹H-NMR spectrum, which indicates that ¹H-NMR-based metabolomics fingerprinting can be used as a sensitive monitoring tool for severe burn patients. This also offers a new approach to understanding the complicated metabolic changes after severe illnesses and injuries such as burns.</p><p>Upon analysis of the 12 metabolites of the metabolome from severe burn patients, we found that they are catalyzed by 103 enzymes that mainly participate in biological processes including protein acetylation, wound healing, and dilation of blood vessels. From the cellular perspective, these enzymes are closely related to the deacetylation of histones, which means remodeling of chromatin and affects the dynamics of chromatin folding during gene transcription [<xref ref-type="bibr" rid="CR30">30</xref>]. Our results showed that levels of the histone deaceylase (HDAC) components (HDAC1-HDAC5, and HDAC9) were elevated significantly and the affinity between histones and DNA was increased, eventually leading to gene transcription repression [<xref ref-type="bibr" rid="CR30">30</xref>, <xref ref-type="bibr" rid="CR31">31</xref>]. However, the histone acetyltransferases (HATs), which reduce histone and DNA affinity and promote transcription, were not obviously changed. These results indicate that protein transcription and synthesis were inhibited and anabolism was restrained during the early stage of burn injury [<xref ref-type="bibr" rid="CR32">32</xref>, <xref ref-type="bibr" rid="CR33">33</xref>].</p><p>The present study indicates that after burn injury, the alterations of metabolism networks and patterns can be detected by a metabolomics techniques based on ¹H-NMR. On one hand, we found that 12 metabolites make up a set of biomarkers that can be used to monitor the severity of burns. Although the clinical standards for evaluating the severity of burns are well established and it is not difficult to distinguish severe and moderate burns, differences in the complex network of metabolism are not so easily understood. Our work used the disturbances in the metabolic fingerprints of 1H-NMR spectra to provide a quantitative method to describe the metabolic network disturbances after severe sepsis. It provides a systems biological approach to understand the relationships between metabolic network disturbances and the occurrence of morbidities (sepsis, multiple organ dysfunction syndrome, and so on) after severe burn.</p><p>In addition to establishing the Eigen-metabolomeour results demonstrate that &#x003b1;-ketoisovaleric acid can be used as a novel biomarker of mitochondrial dysfunction in the clinical setting. However, further analysis of the spectra of metabolites can go deeper and wider along the route of small molecular metabolite-enzyme-functional genomics, which provides innovative ideas for exploring pathophysiologic conditions, enhancing research efforts, and improving future treatment protocols. Finally, our study provides a novel approach for a clinical monitoring system with high sensitivity and accuracy in the future.</p></sec><sec id="Sec17" sec-type="conclusions"><title>Conclusions</title><p>To summary, we demonstrate that ¹H-NMR spectra can be used to establish Eigen-metabolome of severe burn patients. A set of biomarkers such as &#x003b1;-ketoisovaleric acid, 3-methylhistidine, and &#x003b2;-hydroxybutyric acid can characterize metabolic disturbances after severe burn. Our work also provides a systems approach to biomedicine that enable future researchers to integrate information from clinical settings, metabolomics and mathematical modeling to develop a new diagnostic monitoring tool for severe burn patients.</p></sec><sec id="Sec18"><title>Key messages</title><p><list list-type="bullet"><list-item><p>NMR spectra of plasma samples showed significant differences between burn patients and healthy individuals.</p></list-item><list-item><p>Using metabolomics techniques, we identified an Eigen-metabolome that consists of 12 metabolites, which are regulated by 103 enzymes in the global metabolic network.</p></list-item><list-item><p>&#x003b1;-ketoisovaleric acid, 3-methylhistidine, and &#x003b2;-hydroxybutyric acid were the most important biomarkers that were significantly increased during the early stage of burn injury.</p></list-item><list-item><p>Our results also show that the histone deacetylases, which are protein transcription suppressors, were remarkably increased and indicated that protein transcription was inhibited and anabolism restrained during the early stage of burn injury.</p></list-item></list></p></sec>

Targeting IL-6 by both passive or active immunization strategies prevents bleomycin-induced skin fibrosis

Could not extract abstract

<contrib contrib-type="author" corresp="yes"><name><surname>Desallais</surname><given-names>Lucille</given-names></name><address><email>lucille.desallais@gmail.com</email></address><xref ref-type="aff" rid="Aff67"/></contrib><contrib contrib-type="author" corresp="yes"><name><surname>Avouac</surname><given-names>J&#x000e9;r&#x000f4;me</given-names></name><address><email>jerome.avouac@cch.aphp.fr</email></address><xref ref-type="aff" rid="Aff68"/><xref ref-type="aff" rid="Aff69"/></contrib><contrib contrib-type="author"><name><surname>Fr&#x000e9;chet</surname><given-names>Maxime</given-names></name><address><email>maxime.frechet@gmail.com</email></address><xref ref-type="aff" rid="Aff69"/></contrib><contrib contrib-type="author"><name><surname>Elhai</surname><given-names>Muriel</given-names></name><address><email>muriel.elhai@inserm.fr</email></address><xref ref-type="aff" rid="Aff69"/></contrib><contrib contrib-type="author"><name><surname>Ratsimandresy</surname><given-names>Rojo</given-names></name><address><email>rratsima@gmail.com</email></address><xref ref-type="aff" rid="Aff70"/></contrib><contrib contrib-type="author"><name><surname>Montes</surname><given-names>Matthieu</given-names></name><address><email>matthieu.montes@cnam.fr</email></address><xref ref-type="aff" rid="Aff67"/></contrib><contrib contrib-type="author"><name><surname>Mouhsine</surname><given-names>Hadley</given-names></name><address><email>hadleymouhsine@gmail.com</email></address><xref ref-type="aff" rid="Aff67"/></contrib><contrib contrib-type="author"><name><surname>Do</surname><given-names>Herv&#x000e9;</given-names></name><address><email>herve.do@peptinov.fr</email></address><xref ref-type="aff" rid="Aff70"/></contrib><contrib contrib-type="author" corresp="yes"><name><surname>Zagury</surname><given-names>Jean-Fran&#x000e7;ois</given-names></name><address><email>zagury@cnam.fr</email></address><xref ref-type="aff" rid="Aff67"/></contrib><contrib contrib-type="author" corresp="yes"><name><surname>Allanore</surname><given-names>Yannick</given-names></name><address><email>yannick.allanore@cch.aphp.fr</email></address><xref ref-type="aff" rid="Aff68"/><xref ref-type="aff" rid="Aff69"/></contrib><aff id="Aff67"><label/>Chaire de Bioinformatique, Laboratoire G&#x000e9;nomique, Bioinformatique et Applications, EA 4627, Conservatoire National des Arts et M&#x000e9;tiers, 292 Rue Saint-Martin, 75003 Paris, France </aff><aff id="Aff68"><label/>Universit&#x000e9; Paris Descartes, Sorbonne Paris Cit&#x000e9;, Service de Rhumatologie A, H&#x000f4;pital Cochin, Universit&#x000e9; Paris Descartes, 27 rue du Faubourg Saint-Jacques, 75014 Paris, France </aff><aff id="Aff69"><label/>INSERM U1016/CNRS UMR8104, Institut Cochin, 22 Rue Mechain, 75014 Paris, France </aff><aff id="Aff70"><label/>Peptinov, Cochin Sant&#x000e9; Business Accelerator (&#x0201c;P&#x000e9;pini&#x000e8;re Cochin Sant&#x000e9;&#x0201d;), Cochin Hospital, 29 rue du Faubourg Saint Jacques, 75014 Paris, France </aff>

Arthritis Research &#x00026; Therapy

<sec id="Sec1" sec-type="intro"><title>Introduction</title><p>Systemic sclerosis (SSc, scleroderma) is a connective tissue disease of unknown etiology that affects particularly the skin. Early stages of SSc are characterized by vascular changes and inflammatory infiltrates in the lesional skin [<xref ref-type="bibr" rid="CR1">1</xref>]. Later stages of SSc are characterized by an excessive accumulation of extracellular matrix components, including collagen, leading to increased skin thickness.</p><p>Several lines of evidence suggest a pathologic role of cytokine overproduction in the pathogenesis of SSc, particularly in fibroblast activation, collagen synthesis, and subsequent fibrosis. Interleukin-6 (IL-6) is a pleiotropic cytokine whose activities stimulate the proliferation and differentiation of B and T lymphocytes, enhance antibody production, activate T cells, stimulate hematopoietic precursors to differentiate, influence the proliferative capacity of non-lymphoid cells, and activate acute-phase protein response [<xref ref-type="bibr" rid="CR2">2</xref>]. Preliminary data suggest that IL-6 might contribute to human SSc: levels of IL-6 are increased in the serum and in the lesional skin of patients with SSc, spontaneous production of IL-6 by peripheral blood leukocytes from patients with SSc is elevated compared with healthy controls, and IL-6 levels correlate with skin thickness score [<xref ref-type="bibr" rid="CR3">3</xref>&#x02013;<xref ref-type="bibr" rid="CR12">12</xref>]. In addition, two preliminary reports have showed that passive immunization with anti-IL-6 receptor (IL-6R) monoclonal antibody may alleviate skin disease in two mouse models of inflammation-driven dermal fibrosis [<xref ref-type="bibr" rid="CR13">13</xref>, <xref ref-type="bibr" rid="CR14">14</xref>]. However, the anti-fibrotic properties of IL-6 inhibition have not yet been assessed in mouse models of SSc that reflect later and non-inflammatory stages of SSc. Moreover, molecular targeted inhibition of IL-6 signaling <italic>in vivo</italic> was restricted to passive immunization, which may present several drawbacks, including primary and secondary resistances, repeated injections, side effects, and prohibitive costs. As an alternative and innovative strategy, our group has developed peptide-based anti-cytokine active immunization, which consists in inducing autoantibodies through an immunization against peptides of cytokines linked to a carrier protein (for example, keyhole limpet hemocyanin, or KLH) [<xref ref-type="bibr" rid="CR15">15</xref>&#x02013;<xref ref-type="bibr" rid="CR17">17</xref>]. This promising strategy has not been used so far for IL-6 but has been successfully established for other cytokines, including tumor necrosis factor-alpha (TNF&#x003b1;) and IL-1&#x003b2; and IL-23 in different autoimmune diseases [<xref ref-type="bibr" rid="CR15">15</xref>&#x02013;<xref ref-type="bibr" rid="CR18">18</xref>]. Therefore, in this study, our aim was to compare the antifibrotic properties of both passive and active immunization against IL-6 in complementary mouse models of SSc.</p></sec><sec id="Sec2" sec-type="materials|methods"><title>Materials and methods</title><sec id="Sec3"><title>Human skin biopsies</title><p>Paraffin-embedded sections of lesional skin biopsies were obtained from 10 patients with SSc and five healthy age- and sex-matched healthy volunteers. The median age of patients with SSc (eight females and two males) was 55 years (range 39 to 65 years), and disease duration was 4.5 years (range 1 to 12 years). Five patients with SSc had a disease duration of less than 5 years; four had the diffuse cutaneous subset, and six had the limited. No patient was treated with immunosuppressive or other potentially disease-modifying drugs. The median age of controls (four females and one male) was 57 years (range 31 to 62 years). All of the study aspects were approved by the local ethics review committee (Comit&#x000e9; Consultatif de Protection des Personnes dans la Recherche Biom&#x000e9;dicale Paris Ile de France III), and written informed consent was obtained from all patients and controls [<xref ref-type="bibr" rid="CR9">9</xref>, <xref ref-type="bibr" rid="CR19">19</xref>, <xref ref-type="bibr" rid="CR20">20</xref>].</p></sec><sec id="Sec4"><title>Animals</title><p>Four- and six-week-old male and female DBA/2 strain (Janvier, Le Genest Saint Isle, France) and five-week-old male and female tight skin-1 (Tsk-1) strain (The Jackson Laboratory, Bar Harbor, ME, USA) were bred and maintained at the animal care facilities of Montrouge dental university (Montrouge, France). All experimental procedures were conducted in compliance with animal health regulations, and the local ethics committee approved all animal experiments (Comit&#x000e9; National de R&#x000e9;flexion Ethique sur l&#x02019;Exp&#x000e9;rimentation Animale-34).</p></sec><sec id="Sec5"><title>Induction of bleomycin-induced dermal fibrosis</title><p>Skin fibrosis was induced in 6-week-old DBA/2 mice by administering local injections of bleomycin for 3 weeks (d0 to d21), as previously described [<xref ref-type="bibr" rid="CR8">8</xref>, <xref ref-type="bibr" rid="CR9">9</xref>, <xref ref-type="bibr" rid="CR21">21</xref>]. Briefly, 100 &#x003bc;L of bleomycin dissolved in 0.9% NaCl at a concentration of 0.5 mg/mL was administered every other day by subcutaneous injection into defined areas of 1 cm² on the upper back for 3 weeks. Subcutaneous injections of 100 &#x003bc;L of 0.9% NaCl were used as negative control.</p></sec><sec id="Sec6"><title>Tight skin mouse model</title><p>The Tsk-1 phenotype is caused by a dominant mutation in the fibrillin-1 gene [<xref ref-type="bibr" rid="CR22">22</xref>]. Tsk-1 mice are characterized by accumulation of collagen fibers in the hypodermis resulting in progressive hypodermal thickening. In contrast to bleomycin-induced fibrosis, inflammatory infiltrates are absent and the aberrant activation of fibroblasts is not caused by the release of inflammatory mediators from leukocytes. Similar to SSc fibroblasts, fibroblasts from Tsk-1 are endogenously activated with increased release of collagen that persists for several passages <italic>in vitro</italic>.</p></sec><sec id="Sec7"><title>Anti-IL-6 receptor monoclonal antibody treatment</title><p>Rat anti-mouse IL-6 receptor monoclonal antibody (clone MR16-1) described previously was provided by Chugai Pharmaceutical (Tokyo, Japan) [<xref ref-type="bibr" rid="CR13">13</xref>, <xref ref-type="bibr" rid="CR23">23</xref>]. Purified rat IgG1 (isotype-matched control antibody, MP Biomedicals, Illkirch, France) was administered as a negative control. DBA/2 mice received, in parallel of bleomycin injections, an intraperitoneal (i.p.) injection of 2 mg at day 0 followed by one i.p. injection of 1 mg at days 7 and 14. Mice were sacrificed at day 21. Tsk-1 mice received a first i.p. injection of 2 mg at the age of 5 weeks followed by one i.p. injection of 1 mg once a week. Mice were sacrificed by cervical dislocation at the age of 10 weeks.</p></sec><sec id="Sec8"><title>Selection of the peptide derivative of mouse IL-6</title><p>At the beginning of the study, no structural data on mouse IL-6 were available. Peptide was designed by using the human IL-6/IL-6R&#x003b1;/gp130 structure (Protein Data Bank (PDB) code: 1P9M). We could assume that the chosen site was transposable to mIL-6 because of the high conservation between mIL-6 and hIL-6 amino acid sequences. This hypothesis was verified after the publication of the mIL-6 structure in 2012 (PDB code: 2L3Y). Peptide was chosen in a loop exposed to the protein surface, which corresponded to a region involved in the interaction between the cytokine IL-6 and its receptor IL-6R&#x003b1;.</p></sec><sec id="Sec9"><title>Peptides synthesis and cyclisation</title><p>Peptides were synthesized by PolyPeptide Laboratories (Strasbourg, France). Peptides were analyzed by high-pressure liquid chromatography and mass spectrometry. Only fractions with purity superior to 85% were conserved. mIS200 peptide was produced in a cyclized form by the formation of intramolecular disulphide bonds between cysteine residues.</p></sec><sec id="Sec10"><title>Coupling of the peptide to the carrier protein</title><p>To ensure a specific coupling of the mIS200 peptide to KLH, an additional tyrosine residue was added at its C-terminus extremity before proceeding to the coupling with bis-diazobenzidine (BDB). Briefly, the free amines of the peptide were protected with citraconic acid at pH 8.5 to 9. Then, the peptide, the KLH, and a solution of BDB were mixed together in a borate buffer for 2 hours at 4&#x000b0;C. The coupled peptide was then submitted to two dialyses against 5% acetic acid followed by four dialyses against phosphate-buffered saline (PBS). mIS200 sequence: _{Ac&#x000e9;tyl+}C₇₆MNNDDALAENNLKLPE₉₂CY.</p></sec><sec id="Sec11"><title>Immunizations with mIS200 peptide</title><p>DBA/2 mice were immunized four times by intramuscular injections with mIS200 peptide (100 &#x003bc;g/mouse) or KLH alone (200 &#x003bc;g/mouse). The primo-injection was completed with complete Freund adjuvant (Sigma-Aldrich, Saint-Quentin Fallavier, France) and the three boosters, with an interval of 15 days, with incomplete Freund adjuvant (Sigma-Aldrich). Immunizations were performed 31, 17, and 3 days before the first injection of bleomycin and 11 days after the first injection of bleomycin.</p></sec><sec id="Sec12"><title>Mouse anti-IL-6 antibody ELISAs</title><p>Serum samples were obtained from mice at sacrifice, and the IgG response against mouse IL-6 was measured by enzyme-linked immunosorbent assay (ELISA). Briefly, 50 ng of mouse IL-6 (R&#x00026;D Systems, Lille, France) was adsorbed on microtitration plates overnight. After a step of saturation, sera from immunized mice were serially diluted in PBS bovine serum albumin 1% and added in coated wells. A wash was followed by an incubation with polyclonal anti-mouse IgG as secondary antibodies coupled to horseradish peroxidase. Plates were revealed with tetramethylbenzidine, and the reaction was stopped with 1 M sulfuric acid before reading on a spectrophotometer (Multiskan Ex, Thermo Scientific) at 450 nm. ELISA titers were expressed as those serum dilutions that lead to half-maximal optical density 450 (OD₄₅₀) (titer₅₀).</p></sec><sec id="Sec13"><title>Evaluation of dermal and hypodermal thickness</title><p>Lesional skin areas were excised, fixed in 4% formalin, and embedded in paraffin. Sections (5 &#x003bc;m thick) were stained with hematoxylin and eosin. The dermal thickness was analyzed at 100-fold magnification by measuring the distance between the epidermal-dermal junction and the dermal-subcutaneous fat junction at four sites from lesional skin of each mouse. The hypodermal thickness in Tsk-1 mice was determined by measuring the thickness of the subcutaneous connective tissue beneath the panniculus carnosus at four different sites of the upper back in each mouse [<xref ref-type="bibr" rid="CR8">8</xref>]. Two independent examiners performed the evaluation.</p></sec><sec id="Sec14"><title>Collagen measurements</title><p>The collagen content in lesional skin samples was evaluated by the hydroxyproline assay [<xref ref-type="bibr" rid="CR24">24</xref>]. After digestion of punch biopsy specimens (3 mm diameter) in 6 M HCl for 3 hours at 120&#x000b0;C, the pH of the samples was adjusted to 7. Afterwards, samples were mixed with 0.06 M chloramine T and incubated for 20 minutes at room temperature. Then, 3.15 M perchloric acid and 20% p-dimethylaminobenzaldehyde were added, and samples were incubated for an additional 20 minutes at 60&#x000b0;C. The absorbance was determined at 557 nm with a spectrophotometer. For direct visualization of collagen fibers, trichrome staining was performed.</p></sec><sec id="Sec15"><title>Immunohistochemical analysis of &#x003b1;-SMA, CD3, CD22, IL-6, and IL-6-R</title><p>Myofibroblasts were identified by staining for &#x003b1;-smooth muscle actin (&#x003b1;-SMA), as previously described [<xref ref-type="bibr" rid="CR8">8</xref>, <xref ref-type="bibr" rid="CR9">9</xref>, <xref ref-type="bibr" rid="CR21">21</xref>]. Cells positive for &#x003b1;-SMA in mouse skin sections were detected by incubation with monoclonal anti-&#x003b1;-SMA antibody (clone 1A4; Sigma-Aldrich, Saint-Quentin Fallavier, France) diluted 1:1,000 for 3 hours at room temperature. Polyclonal rabbit anti-mouse antibodies labeled with horseradish peroxidase (Dako, Les Ulis, France) were used as secondary antibodies for 1 hour at room temperature. The number of myofibroblasts was determined at 200-fold magnification in six different sections from each mouse by two blinded examiners.</p><p>To quantify the numbers of infiltrating T and B cells, lesional skin sections were stained for CD3 and CD22, respectively. Skin sections were incubated with polyclonal rabbit anti-human antibodies for CD3 or CD22 (Abcam, Paris, France). Polyclonal horseradish goat anti-rabbit antibodies labeled with horseradish peroxidase (Dako) were used as secondary antibodies. T and B cells were counted in a blinded manner, by two examiners, in six different sections of lesional skin of each mouse at 400-fold magnification.</p><p>To detect human IL-6 in lesional skin tissue, skin sections were incubated with monoclonal mouse anti-human antibodies against IL-6 (Abcam). Polyclonal rabbit anti-mouse antibodies labeled with horseradish peroxidase (Dako) were used as secondary antibodies. To detect mouse IL-6R in lesional skin tissue, polyclonal goat anti-mouse antibodies against IL-6R (R&#x00026;D Systems) were used. The intensity of immunostaining was quantified with ImageJ software, as described in the following webpage [<xref ref-type="bibr" rid="CR25">25</xref>].</p></sec><sec id="Sec16"><title>IL-6 measurement in lesional skin samples of bleomycin treated mice</title><p>IL-6 levels were measured in the skin of mice injected with bleomycin or NaCl by multiplex bead array technology (BD Biosciences, Le Pont de Claix, France), as previously described [<xref ref-type="bibr" rid="CR8">8</xref>, <xref ref-type="bibr" rid="CR9">9</xref>].</p></sec><sec id="Sec17"><title>Serum levels of IL-6 and IL-6R</title><p>Serum concentrations of IL-6 (pg/mL) and IL-6R (pg/mL) were measured in a previously described cohort of 187 patients with SSc and 48 unrelated age/sex-matched subjects by quantitative sandwich ELISAs (R&#x00026;D Systems, Minneapolis, MN, USA) in accordance with the recommendations of the manufacturer [<xref ref-type="bibr" rid="CR26">26</xref>]. Serum levels of IL-6 were also measured in mice subjected to bleomycin injections and passive or active immunization as well as in their respective controls by using a U-cytech sandwich mouse IL-6 ELISA kit (U-cytech Biosciences, Utrecht, The Netherlands).</p></sec><sec id="Sec18"><title>Statistics</title><p>Results were expressed as dot blots with median_(Q1,Q3). Mann-Whitney <italic>U</italic> test for non-related samples was used for statistical analyses. <italic>P</italic> values of less than 0.05 were considered significant.</p></sec></sec><sec id="Sec19" sec-type="results"><title>Results</title><sec id="Sec20"><title>Serum levels and skin expression of IL-6 are increased in patients with early systemic sclerosis</title><p>We first evaluated IL-6 and IL-6R expression in the serum of 187 patients with SSc as compared with 48 healthy controls. Higher median IL-6 serum concentrations were measured in patients with SSc compared with controls, although not reaching significant threshold (5.6 <italic>versus</italic> 4.0 pg/mL, <italic>P</italic> = 0.09) (Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>A). In the subset of patients with early disease (disease duration less than 5 years), median IL-6 serum levels were significantly higher than in controls (5.8 versus 4.0 pg/mL, <italic>P</italic> = 0.006) (Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>A). No difference was observed between SSc and controls regarding IL-6R serum levels (data not shown).<fig id="Fig1"><label>Figure 1</label><caption><p> <bold>Interleukin-6 (IL-6) is overexpressed in the serum and the skin of patients with systemic sclerosis (SSc). (A)</bold> Patients with SSc and early disease (less than 5 years) had increased IL-6 serum levels compared with age- and sex-matched healthy controls (<italic>P</italic> = 0.006). IL-6 was detected <italic>ex vivo</italic> by immunohistochemistry in patients with SSc <bold>(C)</bold> compared with controls <bold>(B)</bold>. Positive staining for IL-6 was observed in the epidermis and in several cell types of the dermis. In addition, the intensity of immunostaining assessed by the ImageJ software was significantly increased in patients with SSc compared with controls (<italic>P</italic> = 0.0007) <bold>(D)</bold>, particularly in those with early disease (<italic>P</italic> = 0.008) <bold>(E)</bold>. Bars represent median_(Q1,Q3).</p></caption><graphic xlink:href="13075_2013_4244_Fig1_HTML" id="d30e663"/></fig></p><p>We next assessed skin expression of IL-6 by immunohistochemistry in patients with SSc and controls. Consistent with the results obtained in serum, we observed overexpression of IL-6 in the skin of patients with SSc, including those with early disease (disease duration of less than 5 years) (Figures&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>B-E). The expression of IL-6 was detectable in the 10 patients with SSc and in the five controls. Positive staining for IL-6 was observed in the epidermis and in several cell types in the dermis, including perivascular inflammatory cells, fibroblasts, and endothelial cells (Figures&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>B and <xref rid="Fig1" ref-type="fig">1</xref>C). Moreover, the amount of immunostaining was more abundant in patients with SSc compared with controls (<italic>P</italic> = 0.008), including those with early disease (<italic>P</italic> = 0.008) (Figures&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>D and <xref rid="Fig1" ref-type="fig">1</xref>E). Taken together, these data support overexpression of IL-6 in patients with early SSc.</p></sec><sec id="Sec21"><title>MR16-1 prevents bleomycin-induced dermal fibrosis</title><p>MR16-1 treatment prevented the induction of bleomycin-induced dermal fibrosis. Upon bleomycin injections, dermal thickness was reduced by 25% in mice treated with MR16-1 compared with mice treated with the isotype control (<italic>P</italic> = 0.02) (Figures&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>A and <xref rid="Fig2" ref-type="fig">2</xref>B). Consistent with decreased dermal thickening, reduced accumulation of collagen upon bleomycin challenge was observed on trichrome-stained skin sections of mice treated with MR16-1 (Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>C). In addition, the hydroxyproline content in lesional skin was decreased by 30% in mice treated with MR16-1 compared with those treated with the isotype control (<italic>P</italic> = 0.007) (Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>D). The number of myofibroblasts upon challenge with bleomycin was also significantly reduced by 45% in mice treated with MR16-1 (<italic>P</italic> = 0.005) (Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>E).<fig id="Fig2"><label>Figure 2</label><caption><p> <bold>Mice treated with MR16-1 are protected against bleomycin-induced dermal fibrosis. (A)</bold> Reduced dermal fibrosis in mice treated with MR16-1, injected with bleomycin. Representative hematoxylin and eosin-stained skin sections are shown. <bold>(B)</bold> Decreased dermal thickness in mice treated with MR16-1 (change by median_(Q1,Q3) 1.0_0.7,1.2 versus 1.1_1.0,1.5 fold, <italic>P</italic> = 0.02). <bold>(C)</bold> Reduced accumulation of collagen in mice treated with MR16-1 following bleomycin treatment. Collagen fiber visualization by trichrome-staining skin sections is shown. <bold>(D)</bold> Reduced hydroxyprolin content in mice treated with MR16-1 following bleomycin treatment (change by median_(Q1,Q3) 1.1_0.9,1.4 versus 1.6_1.2,2.0 fold, <italic>P</italic> = 0.007). <bold>(E)</bold> Lower myofibroblast counts in mice treated with MR16-1 following bleomycin treatment (change by median_(Q1,Q3) 0.9_0.8,1.5 versus 1.5_1.3,2.5 fold, <italic>P</italic> = 0.005). Control mice were injected with NaCl, and the value for these mice was defined as 1; the results from the other groups were normalized to this value. Bars represent median_(Q1,Q3); 36 mice were used for these experiments (12 per group). Ab, antibody.</p></caption><graphic xlink:href="13075_2013_4244_Fig2_HTML" id="d30e783"/></fig></p></sec><sec id="Sec22"><title>MR16-1 exerts no antifibrotic effect in Tsk-1 mice</title><p>We next investigated whether treatment with MR16-1 may also be effective in a non-inflammatory model of SSc. We observed that Tsk-1 mice treated with MR16-1 showed no reduction of hypodermal thickness (Figures&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>A and <xref rid="Fig3" ref-type="fig">3</xref>B) and collagen content (Figure&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>C) compared with those receiving the isotype control.<fig id="Fig3"><label>Figure 3</label><caption><p> <bold>Tight skin-1 (Tsk-1) mice are not protected from skin fibrosis by MR16-1 treatment. (A, B)</bold> Tsk-1 mice showed increased hypodermal thickness compared with the pa/pa control mice. However, no difference was observed between Tsk-1 mice receiving either MR16-1 or the isotype control. <bold>(C)</bold> Tsk-1 mice showed increased hydroxyproline concentration compared with the pa/pa control mice. However, no difference was observed between Tsk-1 mice receiving either MR16-1 or the isotype control. pa/pa mice (control mice for Tsk-1 mice) were defined as 1.0, and all other results are normalized to this value. Bars represent median_(Q1,Q3); 17 mice were used for these experiments (five pa/pa mice, four Tsk-1 mice treated with isotype control, and eight Tsk-1 mice treated with MR16-1). Ab, antibody; NS, not significant.</p></caption><graphic xlink:href="13075_2013_4244_Fig3_HTML" id="d30e815"/></fig></p></sec><sec id="Sec23"><title>Active immunization against IL-6 in the bleomycin mouse model</title><sec id="Sec24"><title>Immunization of mice with mIS200 peptide is well tolerated</title><p>Following the encouraging results obtained with MR16-1 in the prevention of bleomycin-induced dermal fibrosis, we aimed to evaluate the protective effect of an anti-IL-6 peptide-based active immunization in this model. Immunization of mice with mIS200 peptide at a dose of 100 &#x003bc;g per injection was well tolerated. mIS200-immunized mice appeared healthy with normal activity, behavior, and texture of the fur. The body weight also did not differ between mice treated with mIS200 and KLH-immunized control.</p></sec><sec id="Sec25"><title>Anti-IL-6 autoantibody production</title><p>DBA/2 mice were immunized four times with mIS200 peptide, and bleomycin-induced fibrosis was induced after the second booster injection. Anti-IL-6 antibody production was evaluated by ELISA at sacrifice. Mice immunized against mIS200 peptide produced a high level of autoantibodies against IL-6, whereas mice immunized against KLH alone did not exhibit anti-IL-6 antibodies (Figure&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>). All mice showed antibody response to the carrier protein KLH (data not shown).<fig id="Fig4"><label>Figure 4</label><caption><p> <bold>Anti-mIL-6 autoantibody production in the group of mice immunized against mIS200 peptide and in the groups of mice immunized against keyhole limpet hemocyanin (KLH) alone.</bold> No anti-IL-6 autoantibodies were detected in serum samples from mice immunized against KLH alone compared with the mice immunized against mIS200 peptide. Bars represent mean &#x000b1; standard deviation.</p></caption><graphic xlink:href="13075_2013_4244_Fig4_HTML" id="d30e843"/></fig></p></sec><sec id="Sec26"><title>Anti-IL-6 peptide-based active immunization alleviates bleomycin-induced skin fibrosis</title><p>Mice immunized against the mIS200 peptide exhibited a significant reduction of dermal thickness by 20% compared with the group immunized against KLH alone (<italic>P</italic> = 0.02) (Figures&#x000a0;<xref rid="Fig5" ref-type="fig">5</xref>A and <xref rid="Fig5" ref-type="fig">5</xref>B). Consistently with decreased dermal thickening, a significant reduction of the collagen content following bleomycin challenge was observed on trichrome-stained skin sections from mice immunized against mIS200 peptide (Figure&#x000a0;<xref rid="Fig5" ref-type="fig">5</xref>C). Furthermore, the hydroxyproline content in the lesional skin of mIS200-immunized mice was significantly decreased by 25% compared with the skin of KLH-immunized mice (<italic>P</italic> = 0.005) (Figure&#x000a0;<xref rid="Fig5" ref-type="fig">5</xref>D). In addition, the number of myofibroblasts was decreased by 41% in mIS200-immunized mice (<italic>P</italic> = 0.01) (Figure&#x000a0;<xref rid="Fig5" ref-type="fig">5</xref>E).<fig id="Fig5"><label>Figure 5</label><caption><p> <bold>Mice immunized against mIS200 peptide are protected against bleomycin-induced dermal fibrosis. (A)</bold> Reduced dermal fibrosis in mice immunized against mIS200 peptide injected with bleomycin. Representative hematoxylin and eosin-stained skin sections are shown. <bold>(B)</bold> Decreased dermal thickness in mice immunized against mIS200 peptide treated with bleomycin (change by median_(Q1,Q3) 1.2_1.1,1.4 versus 1.4_1.3,1.6 fold, <italic>P</italic> = 0.02). <bold>(C)</bold> Reduced accumulation of collagen in mice immunized against mIS200 peptide challenged with bleomycin. Collagen fiber visualization by trichrome-staining skin sections is shown. <bold>(D)</bold> Reduced hydroxyprolin content in mice immunized against mIS200 peptide following bleomycin treatment (change by median_(Q1,Q3) 1.0_0.9,1.2 versus 1.4_1.1,1.6 fold, <italic>P</italic> = 0.005). <bold>(E)</bold> Lower myofibroblast counts in mice immunized against mIS200 peptide following bleomycin treatment (change by median_(Q1,Q3) 1.7_1.2,2.2 versus 2.4_1.7,3.6 fold, <italic>P</italic> = 0.01). Control mice were injected with NaCl, and the value for these mice was defined as 1; the results from the other groups were normalized to this value. Bars represent median_(Q1,Q3); 36 mice were used for these experiments (nine in the group keyhole limpet hemocyanin (KLH) NaCl, 16 in the group mIS200 bleomycin, and 11 in the group KLH bleomycin).</p></caption><graphic xlink:href="13075_2013_4244_Fig5_HTML" id="d30e939"/></fig></p></sec></sec><sec id="Sec27"><title>Passive or active immunization against IL-6 reduce T-cell infiltration in lesional skin</title><p>Inflammatory infiltrates are characteristic features of early stages of SSc, which are mimicked in the bleomycin-induced fibrosis mouse model. Infiltrating leukocytes contain mostly T cells, with a perivascular distribution, and stimulate fibroblast activation and collagen synthesis via the release of pro-fibrotic factors [<xref ref-type="bibr" rid="CR1">1</xref>]. To evaluate whether passive or active immunization against IL-6 influences the outcome of bleomycin-induced fibrosis by regulating leukocyte infiltration, we quantified the number of leukocytes in lesional skin. Inflammatory infiltrates upon bleomycin treatment were significantly reduced in mice treated with MR16-1 or immunized against the mIS200 peptide compared with mice injected with isotype control or KLH (<italic>P</italic> = 0.03 for both comparisons) (Figures&#x000a0;<xref rid="Fig6" ref-type="fig">6</xref>A-D). Since T cells are the main component of inflammatory infiltrates, we quantified the number of T cells in fibrotic skin. Consistent with the reduced number of leukocytes, T-cell counts were significantly lower in mice treated with MR16-1 or immunized against the mIS200 peptide compared with their respective control groups (<italic>P</italic> = 0.02 for both comparisons) (Figures&#x000a0;<xref rid="Fig6" ref-type="fig">6</xref>E and <xref rid="Fig6" ref-type="fig">6</xref>F). In contrast to T cells, the number of B cells did not significantly differ upon bleomycin challenge between mice treated with MR16-1 or immunized against the mIS200 peptide and those injected with isotype control or KLH (data not shown).<fig id="Fig6"><label>Figure 6</label><caption><p> <bold>Interleukin-6</bold> <bold>(IL-6) regulates leukocyte and T-cell infiltration into lesional skin. (A, B)</bold> Reduced inflammation in mice treated with MR16-1 or immunized against the mIS200 peptide and challenged with bleomycin. Representative sections stained by hematoxylin/eosin at 400-fold magnification. <bold>(C, D)</bold> Decreased leukocyte counts in lesional skin of mice treated with MR16-1 or immunized against the mIS200 peptide. <bold>(E, F)</bold> Reduced number of CD3⁺ T cell by immunohistochemistry in lesional skin of mice treated with MR16-1 or immunized against the mIS200 peptide. Bars represent median_(Q1,Q3); in regard to leukocyte quantification, 36 mice were evaluated for passive immunization and 36 for active immunization. In regard to T-cell quantification, 15 randomly chosen mice (five per group) were evaluated for passive immunization as well as for active immunization. All results are normalized to mice injected with NaCl. Ab, antibody; KLH, keyhole limpet hemocyanin.</p></caption><graphic xlink:href="13075_2013_4244_Fig6_HTML" id="d30e991"/></fig></p></sec><sec id="Sec28"><title>Serum and skin levels of IL-6 are regulated by passive or active immunization against IL-6</title><p>To better apprehend the effects of passive and active immunization in the bleomycin-induced dermal fibrosis mouse model, IL-6 and IL-6R expressions were assessed in the lesional skin of mice by multiplex bead array technology and immunohistochemistry, respectively. In addition, serum levels of IL-6 were measured by ELISA. Mice treated with MR16-1 and subjected to bleomycin injections displayed significantly higher IL-6 serum levels than mice receiving the isotype control (Figure&#x000a0;<xref rid="Fig7" ref-type="fig">7</xref>A), as reported in previous publications [<xref ref-type="bibr" rid="CR27">27</xref>]. Consistent with this result, a trend for higher skin levels of IL-6 was observed upon MR-16 treatment (Figure&#x000a0;<xref rid="Fig7" ref-type="fig">7</xref>B). Upon bleomycin challenge, IL-6 levels were significantly reduced in the serum and the skin of mice immunized against the mIS200 peptide compared with mice injected with KLH (Figures&#x000a0;<xref rid="Fig7" ref-type="fig">7</xref>C and <xref rid="Fig7" ref-type="fig">7</xref>D).<fig id="Fig7"><label>Figure 7</label><caption><p> <bold>Passive and active immunizations against interleukin-6 (IL-6) regulate serum levels and skin expression of IL-6. (A, B)</bold> In mice challenged by bleomycin, MR16-1 treatment led to a significant increase of IL-6 serum levels (<italic>P</italic> &#x0003c;0.0001) and a trend for increased IL-6 skin expression (<italic>P</italic> = 0.08). <bold>(C, D)</bold> Immunization against the mIS200 peptide conducted to a significant reduction of IL-6 in the serum and in the skin of mice challenged with bleomycin. Bars represent median_(Q1,Q3); in total, 36 mice were evaluated for passive immunization and 36 for active immunization. All results are normalized to mice injected with NaCl.</p></caption><graphic xlink:href="13075_2013_4244_Fig7_HTML" id="d30e1036"/></fig></p><p>Dermal expression of IL-6R was not modified by passive or active immunization against IL-6 in mice challenged with bleomycin injections (data not shown).</p></sec></sec><sec id="Sec29" sec-type="discussion"><title>Discussion</title><p>We first confirmed that IL-6 is overexpressed in the serum and the skin of patients with SSc and particularly in those with early disease. Since upregulation of serum IL-6 has been shown to be associated in SSc with disease activity, severity, disability, worse outcome, and reduced survival, targeting IL-6 may be particularly relevant in patients with early disease. Our results in mice also confirm the relevance of targeting IL-6 in early SSc since IL-6 exhibits a critical role in the development of bleomycin-induced dermal fibrosis, which reflects early and inflammatory stages of SSc. Indeed, inhibition of IL-6 through an innovative anti-IL-6 active immunization strategy exerted similar antifibrotic properties as passive immunization with MR16-1 (monoclonal antibody against IL-6-R) in the bleomycin-induced dermal fibrosis mouse model, by reducing the infiltration of inflammatory cells, especially T cells, into lesional skin. Conversely, MR16-1 did not prevent the development of fibrosis in the Tsk-1 mouse model, suggesting that IL-6 has no direct effects on fibroblast activation and collagen synthesis in this model of late and non-inflammatory stages of SSc.</p><p>We first performed passive immunization with MR16-1 in the bleomycin mouse model (i) to confirm previously reported findings in our in-house bleomycin-challenged mice and (ii) to obtain a range of effect of passive immunization efficacy, allowing further comparison with active immunization. We showed that mice treated with MR16-1 were protected from the development of dermal fibrosis, with a significant reduction of dermal thickness, collagen content, and infiltrating myofibroblasts in the lesional skin. Our results were in line to those of Kitaba <italic>et al.</italic> [<xref ref-type="bibr" rid="CR13">13</xref>], despite several differences in the experiments performed. In particular, we used DBA/2 mice in our study, which show higher susceptibility to bleomycin-induced fibrosis than C57BL/6 mice [<xref ref-type="bibr" rid="CR28">28</xref>]. Both studies differed for the bleomycin injection protocol (bleomycin at a concentration of 0.5 mg/mL injected subcutaneously every other day for 3 weeks in our study compared with a concentration of 1 mg/mL injected subcutaneously daily for 4 weeks). We have used this protocol in several previous projects [<xref ref-type="bibr" rid="CR8">8</xref>, <xref ref-type="bibr" rid="CR9">9</xref>, <xref ref-type="bibr" rid="CR21">21</xref>]. The treatment scheme with MR16-1 was also different between both studies (a first i.p. injection of 2 mg at day 0 then i.p. injections of 1 mg at days 7 and 14 in our study compared with a first intravenous injection of 2 mg at day 0 then i.p. injections of 0.5 mg at days 7, 14, and 21) [<xref ref-type="bibr" rid="CR13">13</xref>]. Our data showed that MR16-1 exerts antifibrotic effects in the model of bleomycin-induced skin fibrosis by reducing the infiltration of T cells into lesional skin, which was not reported before. Reduced T-cell infiltration may lead to decreased resident fibroblast activation in lesional skin. This is consistent with the reduction of myofibroblast counts in MR16-1-treated mice, upon bleomycin challenge, observed in our study and previously reported by Kitaba <italic>et al.</italic> [<xref ref-type="bibr" rid="CR13">13</xref>].</p><p>We also observed a significant increase of serum IL-6 levels and a trend for increased skin IL-6 expression in mice treated with MR16-1. This is in line with previous observations of increased blood IL-6 levels after anti-IL-6R antibody injection [<xref ref-type="bibr" rid="CR27">27</xref>]. This elevation seems to be the result of IL-6 clearance inhibition due to IL-6R blockade, rather than induction of IL-6 synthesis to counterbalance IL-6 blockade, or release of free IL-6 from complexes [<xref ref-type="bibr" rid="CR27">27</xref>].</p><p>Since previous studies focused on the effects of IL-6 inhibition with MR16-1 in mouse models of bleomycin-induced dermal fibrosis, we next aimed to assess the effect of MR16-1 in a non-inflammatory mouse model of SSc. Thus, we investigated the effects of MR16-1 in the Tsk-1 mouse model, in which increased serum IL-6 levels have been reported [<xref ref-type="bibr" rid="CR7">7</xref>]. However, MR16-1-treated mice did not exhibit a reduction of skin fibrosis, suggesting that other pathways or cytokines may be more important in this specific mouse model. This result also supports that IL-6 exerts indirect profibrotic effects in early inflammatory stages of SSc rather than by direct effects on the collagen synthesis by fibroblasts. This result contrasts with the direct profibrotic effects of IL-6 observed recently on dermal fibroblasts [<xref ref-type="bibr" rid="CR29">29</xref>]. Since these results were obtained <italic>in vitro</italic> in human fibroblasts, they may differ from those obtained <italic>in vivo</italic> in the very specific Tsk-1 mouse model. The absence of antifibrotic effects of MR16-1 in the Tsk-1 mice is also sustained by the controversial role of immune cells, especially T and B cells, in the development of fibrosis in this mouse model. Although a role for CD4⁺ T and B cells has been suggested in the activation of collagen synthesis, bone marrow transplantation experiments have challenged the contribution of these immune cells [<xref ref-type="bibr" rid="CR7">7</xref>, <xref ref-type="bibr" rid="CR30">30</xref>]. Indeed, the transfer of enriched B or T cells increased autoantibody production but did not cause skin fibrosis, and transfer of T and B lymphocytes led only to mild fibrotic lesions compared with massive fibrosis in Tsk-1 mice [<xref ref-type="bibr" rid="CR31">31</xref>]. In line with this finding, Tsk-1 mice lacking mature T and B lymphocytes developed a fibrotic phenotype in the absence of a functional immune system [<xref ref-type="bibr" rid="CR32">32</xref>]. Taken together, our results support that passive immunotherapy with MR16-1 has interest only in inflammation-driven dermal fibrosis; thus, this strategy should preferentially be tested in early inflammatory stages of SSc.</p><p>Following our results obtained with MR16-1 in the bleomycin mouse model, we evaluated the antifibrotic effects of IL-6 inhibition with an innovative anti-cytokine strategy. Since the late &#x02019;90s, anti-cytokine biologics (monoclonal antibodies and soluble receptors) have successfully been used to treat chronic inflammatory diseases. However, these biotherapies present several drawbacks, including primary and secondary resistances, repeated injections, side effects, and prohibitive costs. To circumvent these drawbacks, new anti-cytokine strategies, such as the anti-cytokine active immunization strategy, are in development [<xref ref-type="bibr" rid="CR33">33</xref>]. This is notably highlighted by the use of anti-TNF&#x003b1; kinoid in a phase-2 trial in rheumatoid arthritis (ClinicalTrials.gov identifier NCT01040715) and Crohn&#x02019;s disease (ClinicalTrials.gov identifier NCT01291810). In the present study, mice immunized against an IL-6 peptide exhibited a significant reduction of dermal fibrosis, collagen content, and myofibroblasts in the bleomycin mouse model with no adverse event. Moreover, this prevention was similar to the one conferred by the MR16-1 monoclonal antibody (25% and 20% reduction of dermal thickness, respectively, and 25% and 30% reduction of hydroxyproline content, respectively). Our results suggest that active immunization against IL-6 displays antifibrotic properties by decreasing T-cell infiltration into lesional skin, similarly as passive immunization, and by reducing IL-6 levels, as demonstrated by decreased IL-6 levels in the serum and the skin of mice immunized against the mIS200 peptide and challenged with bleomycin. This work shows the feasibility and the efficacy of targeting IL-6 by a peptide-based active immunization in order to block endogenous IL-6 pathological effects. Of note, this strategy was not evaluated in the Tsk-1 mouse model regarding the negative results obtained with the passive immunization strategy.</p><p>Evaluation of passive and active immunization was not evaluated to reverse established fibrosis, which is a limitation of our study. Kitaba <italic>et al.</italic> [<xref ref-type="bibr" rid="CR13">13</xref>] have previously shown that passive immunization with MR16-1 improves established bleomycin-induced dermal fibrosis. Further studies are required to assess whether active immunization may be curative in the modified bleomycin model of established dermal fibrosis.</p></sec><sec id="Sec30" sec-type="conclusions"><title>Conclusions</title><p>Using complementary mouse models of SSc, we demonstrated that passive and active immunization targeting IL-6 had similar antifibrotic effects in the mouse model of bleomycin-induced dermal fibrosis, which are mediated by the reduction of T-cell infiltration into lesional skin and by the decreased skin IL-6 levels in the case of active immunization. Translation to human disease is now required, and targeting early and inflammatory stages of SSc sounds the most appropriate. The passive immunization strategy is under investigation in a phase-3 clinical trial assessing the efficacy of tocilizumab to improve skin involvement in patients with early diffuse SSc (ClinicalTrials.gov identifier NCT01532869). Targeted innovative therapies are the most important issue in SSc, which is a very severe condition free of an efficient drug. It is awaited eagerly by the whole medical community, together with the patients, in order to stop the progression of tissue damage by this devastating disease.</p><p>Our results also highlight the relevance of anti-IL-6 peptide-based active immunotherapy to treat autoimmune diseases. Further investigations are needed to translate this strategy into humans, but it constitutes a promising therapeutic approach.</p></sec>

Broadening horizons: holistic viewpoints from the Biology of Genomes

<p>A report on the Cold Spring Harbor Laboratory 27th annual meeting on the Biology of Genomes, held in Cold Spring Harbor, New York, USA, 6-10 May 2014.</p>

<contrib contrib-type="author" id="A1"><name><surname>Carr</surname><given-names>Ambrose</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>ajc2205@columbia.edu</email></contrib><contrib contrib-type="author" corresp="yes" id="A2"><name><surname>Pe&#x02019;er</surname><given-names>Dana</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>dpeer@biology.columbia.edu</email></contrib>

Genome Biology

<sec><title>Fine mapping in <italic>cis</italic> discovers substantial functional regulatory variation</title><p>&#x02018;Biology of Genomes&#x02019; is a broad annual meeting that showcases the year&#x02019;s progress in the many subfields of genomics. With so many talks, it is impossible to cover all of them. However, despite the variety of subject matter, several trends recurred throughout the conference.</p><p>One result that consistently resurfaced across multiple sessions is the ability of enlarged studies of tens of thousands of individuals to move beyond linkage to genes and finely map traits to individual regulatory polymorphisms. Jeffrey Barret (Wellcome Trust Sanger Institute, Hinxton, UK) was able to identify specific genes as contributors to disease by using data from 12,000 patients in the UK10K project. Hailiang Hailiang (MIT, USA) presented results from the inflammatory bowel disease (IBD) consortium that allowed the identification of specific variants that contribute to IBD, and Stephen Parker (NIH Bethesda, USA) presented the results of the ongoing FUSION study of type 2 diabetes, where a combination of RNA-seq and genotyping was used to identify muscle expression quantitative trait loci (eQTLs) associated with different physiological stages of type 2 diabetes.</p><p>With the enormous amount of data being generated, it is logical to ask: how much functional variation is found in coding versus noncoding regions of the genome? Alexander Gusev (Harvard University, USA) answered this question by annotating the genome using data from Encyclopedia of DNA Elements (ENCODE) and Roadmap Epigenomics projects. He found substantial enrichment of psychophrenia QTLs in DNase hypersensitive sites, whereas there were essentially no QTLs in intron and intergenic sites. Together, these studies suggest that regulatory variation is responsible for a large amount of functional variation, and we are now capable of finding many concrete examples in <italic>cis</italic>, where chromatin state influences the cell or organism phenotype.</p><p>Even with these massive studies, there was still no mention of <italic>trans</italic>-QTL discoveries, suggesting that even use of tens of thousands of individuals leaves studies under-powered to locate distal effectors. Luke Jostins (Wellcome Trust Centre for Human Genetics, Oxford, UK) supported this hypothesis directly by reporting an inability to detect <italic>trans</italic>-eQTLs despite using a model specifically designed to detect them.</p></sec><sec><title>Transcriptional and translational regulation</title><p>Jacob Degner (European Molecular Biology Lab, Heidelberg, Germany) finely examined one method by which regulatory variation can have a functional impact on gene expression by using cap analysis of gene expression (CAGE), a method that precisely identifies transcriptional start sites (TSSs), on 80 cell lines at three time-points. By integrating genotype information, Degher defined promoter-shape QTLs, identifying two main classes of variants: those that cause the precise location of transcription initiation to vary at a higher rate, and those that shift the location of transcriptional initiation to a new position. It is not yet clear how these QTLs influence expression, but it suggests at least two mechanisms by which regulatory variation might regulate the 5&#x02032; mRNA, which could have downstream effects on mRNA secondary structure, splicing and stability.</p><p>Beyond transcription, there are many downstream regulatory layers that are under genetic influence. Protein translation and protein expression are two such layers that received substantial attention. Alexis Battle (Stanford University, USA) and Yoav Gilad (University of Chicago, USA) presented two sides of a collaboration that examines this question by comparing the incidence of eQTLs, &#x02018;ribo-QTLs&#x02019; and &#x02018;protein-QTLs&#x02019;, the latter of which were measured by comparing genotype information with ribosome profiling and high-throughput &#x02018;stable isotope labeling with amino acids in cell culture&#x02019; (SILAC) mass spectrometry. The results were heartening, showing that at least half of the functional variation that evokes an eQTL also associates with a protein-QTL. This indicates that RNA-based analyses are capturing a substantial portion of the cell phenotype. However, significantly, it shows the promise and importance of protein-based technologies that are capable of generating more-representative images of cell phenotypes.</p></sec><sec><title>Mutations and deletions</title><p>Clustered regularly interspaced short palindromic repeats (CRISPRs) were one of the big stories of the past year, bringing major improvements in the ability of scientists inexpensively to achieve complete gene knock-out disruption. Neville Sanjana (Broad Institute of Harvard and MIT, USA) presented a genome-scale CRISPR screen, over every gene in the genome, developed in Feng Zhang&#x02019;s group (Broad Institute, USA). They screened melanoma cell lines treated with vemurafenib, a highly effective late-stage melanoma drug against which most patients eventually develop resistance. In their screen, they identified many genes known to contribute to vemurafenib resistance, as well as some new candidates, highlighting the potential of CRISPR screens in drug discovery.</p><p>Several other groups asked questions about what can be learned from examining natural mutations. Michael Stratton (The Wellcome Trust Sanger Institute, Hinxton, UK) catalogued all the different mutation types present in the <italic>TP53</italic> gene across numerous tissues and cell lines and compared the mutation types with the effects of known mutagens. He found that over 50% of the mutations could be attributed to known mutagens such as UV light or aflatoxin and was able to identify the percentage of mutations that each mutagen was responsible for. However, much of the mutation load could not be explained by his panel, suggesting that there are novel mechanisms of DNA mutation that have not yet been investigated.</p><p>Minyoung Wyman (Columbia University, USA) leveraged mutation rates in an innovative way by comparing the rate at which germ cells accrue two different sets of mutations: mutations resulting from mitosis that require cell division, and non-mitotic mutations. As germ cells do not divide between birth and puberty, the two mutation types have different rates, and the historical difference in these rates allowed Wyman to estimate the average generation time throughout human history: a slow increase from 25 to 29&#x000a0;years.</p></sec><sec><title>Computation and technology</title><p>Perhaps the most consistent message throughout the conference was that, as data increase in magnitude, their utility is limited by the sophistication of our computational approaches. There were several excellent presentations that highlighted this fact, perhaps none more strongly than that of Joseph Pickrell (New York Genome Center, USA), who showed that, by using factor analysis to integrate multiple correlated variables, much more informative eQTLs can be discovered and used to distinguish causative linkages between symptoms and comorbid ones caused by the same functional variant. Dana Pe&#x02019;er (Columbia University, USA) showed how integrating multiple data types improves statistical power enough to detect novel driver mutations that are missed by any data type alone.</p><p>Matthew Stephens (University of Chicago, USA) reexamined the foundations of a ubiquitous statistic: the false discovery rate (FDR). Stephens showed that the bimodal null model using common FDR algorithms has the effect of depressing the corrected <italic>P</italic>-values output by the procedure. Therefore, we are likely over-estimating our experiments&#x02019; false discovery rates. Finally, Oliver Stegle (European Bioinformatics Institute, Hinxton, UK) examined the effect of the cell cycle on gene expression in single-cell sequencing. Unlike bulk expression analyses, which average over a population of cells, single-cell sequencing is very sensitive to the cell cycle, and Stegle demonstrated how an algorithm that corrects for cell-cycle stage reveals hidden information on the stage of differentiation in embryonic stem cells. Together, these talks demonstrate the importance of modeling the sources of variance in biological experiments and the power of integrative analyses to uncover biologically meaningful phenomena obscured under any single assay.</p></sec><sec sec-type="conclusions"><title>Conclusions</title><p>Holistically, Biology of Genomes 2014 provided a broad view into many cutting-edge projects. Promising results were reported from many corners of genomes: for example, large studies leveraged their improved power to identify specific regulatory SNPs; functional variation that affects expression corresponded to matched variation in protein expression in more than half of the cases; some early results of genome-wide deletion screens successfully identified genes involved in drug resistance; and new computational approaches were able to integrate multiple assays profitably to uncover a whole greater than its parts.</p></sec><sec><title>Abbreviations</title><p>CAGE: Cap analysis of gene expression</p><p>CRISPR: Clustered regularly interspaced short palindromic repeat</p><p>eQTL: Expression quantitative trait locus</p><p>FDR: False discovery rate</p><p>IBD: Inflammatory bowel disease</p><p>SILAC: Stable isotope labeling with amino acids in cell culture</p><p>TSS: Transcriptional start site</p></sec><sec><title>Competing interests</title><p>The authors declare that they have no competing interests.</p></sec>

How to manage aspergillosis in non-neutropenic intensive care unit patients

Could not extract abstract

<contrib contrib-type="author" corresp="yes"><name><surname>Bassetti</surname><given-names>Matteo</given-names></name><address><email>mattba@tin.it</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author"><name><surname>Righi</surname><given-names>Elda</given-names></name><address><email>elda.righi@libero.it</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author"><name><surname>De Pascale</surname><given-names>Gennaro</given-names></name><address><email>m.antonelli@rm.unicatt.it</email></address><xref ref-type="aff" rid="Aff2"/></contrib><contrib contrib-type="author"><name><surname>De Gaudio</surname><given-names>Raffaele</given-names></name><address><email>rdegaudio@tin.it</email></address><xref ref-type="aff" rid="Aff3"/></contrib><contrib contrib-type="author"><name><surname>Giarratano</surname><given-names>Antonino</given-names></name><address><email>antonino.giarratano@unipa.it</email></address><xref ref-type="aff" rid="Aff4"/></contrib><contrib contrib-type="author"><name><surname>Mazzei</surname><given-names>Tereesita</given-names></name><address><email>teresita.mazzei@unifi.it</email></address><xref ref-type="aff" rid="Aff5"/></contrib><contrib contrib-type="author"><name><surname>Morace</surname><given-names>Giulia</given-names></name><address><email>giulia.morace@unimi.it</email></address><xref ref-type="aff" rid="Aff6"/></contrib><contrib contrib-type="author"><name><surname>Petrosillo</surname><given-names>Nicola</given-names></name><address><email>nicola.petrosillo@inmi.it</email></address><xref ref-type="aff" rid="Aff7"/></contrib><contrib contrib-type="author"><name><surname>Stefani</surname><given-names>Stefania</given-names></name><address><email>stefanis@unicat.it</email></address><xref ref-type="aff" rid="Aff8"/></contrib><contrib contrib-type="author"><name><surname>Antonelli</surname><given-names>Massimo</given-names></name><address><email>m.antonelli@rm.unicatt.it</email></address><xref ref-type="aff" rid="Aff2"/></contrib><aff id="Aff1"><label/>Infectious Diseases Division, Santa Maria Misericordia University Hospital, 33100 Udine, Italy </aff><aff id="Aff2"><label/>Istituto di Anestesiologia e Rianimazione, Universit&#x000e0; Cattolica-Policlinico Universitario A.Gemelli, 00100 Roma, Italy </aff><aff id="Aff3"><label/>Department of Health Sciences, Anesthesiology and Intensive Care Section, University of Florence, 50100 Firenze, Italy </aff><aff id="Aff4"><label/>Anesthesia, Analgesia and Intensive Care Division, P.Giaccone University Hospital, School of Medicine DiBiMef-University of Palermo, 90100 Palermo, Italy </aff><aff id="Aff5"><label/>Department of Health Sciences - Section of Clinical Pharmacology and Oncology, University of Florence, 50100 Firenze, Italy </aff><aff id="Aff6"><label/>Department of Health Sciences, Universit&#x000e0; degli Studi di Milano, 20100 Milano, Italy </aff><aff id="Aff7"><label/>Second Division, Lazzaro Spallanzani National Institute for Infectious Diseases, 00100 Roma, Italy </aff><aff id="Aff8"><label/>Department of Bio-medical Sciences, University of Catania, 95100 Catania, Italy </aff>

Critical Care

<sec id="Sec1"><title>Review</title><sec id="Sec2" sec-type="introduction"><title>Introduction</title><p>Invasive aspergillosis (IA) is an opportunistic infection that occurs mainly among patients with hematological malignancies, most notably during prolonged periods of neutropenia, but also in subjects with solid tumors, critical illness, and HIV/AIDS, and those undergoing allogeneic stem cell transplantation and solid-organ transplantation [<xref ref-type="bibr" rid="CR1">1</xref>,<xref ref-type="bibr" rid="CR2">2</xref>]. In recent years, however, IA has increasingly been recognized as an emerging disease of non-neutropenic patients and in patients admitted to the ICU, even in the absence of an apparent predisposing immunodeficiency [<xref ref-type="bibr" rid="CR3">3</xref>-<xref ref-type="bibr" rid="CR8">8</xref>]. Although not uncommon, the features of IA among immunocompetent patients differ greatly from those of IA in neutropenic patients. The epidemiology, clinical characteristics, outcomes, and prognosis are not well known in immunocompetent patients. In the ICU, the incidence of IA ranges from 0.3% to 5.8% [<xref ref-type="bibr" rid="CR4">4</xref>,<xref ref-type="bibr" rid="CR5">5</xref>] with an overall mortality rate exceeding 80% [<xref ref-type="bibr" rid="CR9">9</xref>].</p><p>Several recent case series and single-center cohort reports have documented the expansion of patient populations at risk for IA that are different from the traditionally recognized risk groups. They include patients with chronic obstructive pulmonary disease (COPD) and other chronic lung or connective tissue diseases requiring corticosteroid therapy, decompensated liver cirrhosis, and solid cancer with or without treatment [<xref ref-type="bibr" rid="CR10">10</xref>,<xref ref-type="bibr" rid="CR11">11</xref>].</p><p>The diagnosis of IA in non-neutropenic critically ill patients is difficult because signs and symptoms are non-specific, and the initiation of additional diagnostic examinations is often delayed because of a low clinical suspicion. A high level of suspicion is needed to obtain an early diagnosis and a timely therapeutic intervention. A better understanding of the population at risk and the spectrum of diseases caused by IA in non-neutropenic patients may help to improve the outcome of this potentially treatable disease.</p><p>In this review, we describe the epidemiology of and the risk factors for pulmonary IA in non-neutropenic patients, limitations and advances in the diagnostic process, and the different approaches in antifungal therapy, including the main pharmacological properties of different antifungal drugs.</p></sec><sec id="Sec3"><title>Epidemiology</title><p>Despite a documented increase in the incidence of IA in ICUs, different rates are reported among subsets of ICU patients. Indeed, a high prevalence (17%) of IA has been observed in a cohort of 67 patients with severe hospital-acquired pneumonia admitted to the ICU [<xref ref-type="bibr" rid="CR12">12</xref>]. Among 40 critically ill patients with confirmed H1N1 infection, 9 (23%) developed IA 3&#x000a0;days after ICU admission [<xref ref-type="bibr" rid="CR13">13</xref>].</p><p>Retrospective, autopsy-controlled studies showed interesting results. Roosen and colleagues [<xref ref-type="bibr" rid="CR14">14</xref>] studied causes of death in the ICU, revealing 15 cases of IA, 5 of which were undiagnosed before death, among 100 autopsies. In a retrospective study, 127 patients out of 1,850 admissions (6.9%) had microbiological or histopathological evidence of <italic>Aspergillus</italic> during their ICU stay [<xref ref-type="bibr" rid="CR5">5</xref>]. Postmortem examination was done in 47 out of 71 patients, and 27 (59%) were identified with IA.</p><p>In a study comparing neutropenic and non-neutropenic patients with an IA diagnosis during a 6-year period, Cornillet and colleagues [<xref ref-type="bibr" rid="CR6">6</xref>] found a mean number of 15 IA cases per year; of these, approximately half were in the ICU. In an Italian study conducted in two mixed ICUs during 2&#x000a0;years, the incidence of IA was 0.2%, much lower than in other reports from similar ICUs [<xref ref-type="bibr" rid="CR15">15</xref>].</p><p>Risk factors for IA in non-neutropenic patients admitted to the ICU include prolonged treatment with corticosteroids before admission, COPD, liver cirrhosis with prolonged ICU stay (&#x0003e;7&#x000a0;days), solid organ cancer, HIV infection, and lung transplantation [<xref ref-type="bibr" rid="CR16">16</xref>]. However, most of these factors are frequent among non-neutropenic critically ill patients. An intriguing hypothesis on the cause of immunosuppression in the apparently immunocompetent patient with multiple-organ dysfunction relates to the biphasic response to sepsis. Indeed, the initial hyperinflammatory phase is followed by relative immunoparalysis. This latter process is characterized by neutrophil deactivation, and it may put the patient at risk of developing opportunistic infections, such as IA [<xref ref-type="bibr" rid="CR17">17</xref>].</p></sec><sec id="Sec4"><title>Risk factors</title><p>One of the most important risk factors for IA in non-neutropenic patients is COPD [<xref ref-type="bibr" rid="CR7">7</xref>]. Patients with COPD are susceptible to <italic>Aspergillus</italic> colonization of the lower tract of the respiratory airway and under particular circumstances this may lead to invasive infection [<xref ref-type="bibr" rid="CR18">18</xref>]. COPD patients present alterations in lung structure, an impaired immunologic response, reduced mucociliary clearance and mucosal lesions. Moreover, they are prone to frequent hospitalization, broad-spectrum antibiotic treatment and invasive procedures. All these factors could explain the high incidence of aspergillosis in COPD [<xref ref-type="bibr" rid="CR7">7</xref>]. Of note, they are frequently treated with corticosteroids and both inhaled and systemic therapy have been described as another important risk factor for aspergillosis [<xref ref-type="bibr" rid="CR19">19</xref>,<xref ref-type="bibr" rid="CR20">20</xref>]. Steroids are able to accelerate the <italic>in vitro</italic> growth of <italic>Aspergillus</italic> spp. since both the innate and acquired immune responses are impaired [<xref ref-type="bibr" rid="CR21">21</xref>]. Vandewoude and colleagues [<xref ref-type="bibr" rid="CR22">22</xref>] defined a total daily dose &#x02265;20&#x000a0;mg prednisone or equivalent among criteria for defining cases of IA. Both compensated and decompensated cirrhosis have been described as risk factors for IA and impaired phagocytosis has been proposed as a possible explanation in these groups [<xref ref-type="bibr" rid="CR23">23</xref>,<xref ref-type="bibr" rid="CR24">24</xref>]. Diabetes has been observed as another risk factor [<xref ref-type="bibr" rid="CR22">22</xref>]. Impaired innate and acquired immunity caused by hyperglycemia may explain this observation [<xref ref-type="bibr" rid="CR25">25</xref>]. Several authors report alcoholism and malnutrition as other possible risk factors for IA [<xref ref-type="bibr" rid="CR22">22</xref>,<xref ref-type="bibr" rid="CR26">26</xref>].</p><p>Patients in the ICU are subjected to several therapies (for example, broad spectrum antibiotics, mechanical ventilation) and/or maneuvers (for example, insertion of central venous catheter), which may affect the immune system defenses. Even though some of these conditions have been described as possible contributors, additional factors may be required for the development of IA [<xref ref-type="bibr" rid="CR5">5</xref>,<xref ref-type="bibr" rid="CR16">16</xref>,<xref ref-type="bibr" rid="CR26">26</xref>].</p><p>Immunosuppression has been described as a late stage of the biphasic response to sepsis and multiple organ failure syndrome [<xref ref-type="bibr" rid="CR27">27</xref>]. Hartemink and colleagues [<xref ref-type="bibr" rid="CR17">17</xref>] first proposed the association between this condition and IA development. This could be one of the main reasons why aspergillosis is frequent among patients not considered immunocompromised by classic criteria.</p></sec><sec id="Sec5"><title>Clinical diagnosis and case definition</title><p>Clinical manifestations of IA (for example, fever, cough, purulent sputum) may be initially indistinguishable from those of bacterial bronchopneumonia [<xref ref-type="bibr" rid="CR28">28</xref>]. The recovery of the same <italic>Aspergillus</italic> species from several respiratory samples in the course of antibiotic-resistant pneumonia in patients with risk factors is clearly evocative of the diagnosis [<xref ref-type="bibr" rid="CR10">10</xref>]. Therefore, it has been proposed that the isolation of an <italic>Aspergillus</italic> species from the respiratory tract in critically ill patients with risk factors (COPD after corticosteroid exposure, severe underlying disease) and clinical features of pneumonia should indicate a probable IA case.</p><p>The presence of a persistent pulmonary infection despite broad-spectrum antibiotics or abnormal thoracic imaging by CT scanning together with one of the risk factors should trigger further diagnostic exploration through collection of respiratory secretions and/or laboratory markers. Invasive infections in patients with negative cultures might be supported by positive molecular and serological tests, such as <italic>Aspergillus</italic> PCR and galactomannan (GM) antigen, which requires at least two sequentially positive samples. Radiological findings can be non-specific in non-neutropenic patients, and of the typical imaging findings observed in neutropenic patients, the air crescent sign was seen in only a small proportion of cases, while the halo sign was very rarely observed. The halo sign and air crescent sign in thoracic CT scans have a high sensitivity (80%) and specificity (60 to 98%) for IA among neutropenic patients with pulmonary infection [<xref ref-type="bibr" rid="CR29">29</xref>]. In non-neutropenic patients, a lower sensitivity (5 to 24%) is reported in the literature and these signs are less frequently observed [<xref ref-type="bibr" rid="CR30">30</xref>,<xref ref-type="bibr" rid="CR31">31</xref>]. Bronchoscopy manifestations were also non-specific in non-neutropenic patients, with a lack of consistent endoscopic features [<xref ref-type="bibr" rid="CR31">31</xref>].</p><p>The diagnosis of IA is particularly problematic. According to the revised definitions for invasive fungal disease of the European Organization for Research and Treatment of Cancer/Invasive Fungal Infections Cooperative Group and the National Institute of Allergy and Infectious Diseases Mycoses Study Group (EORTC/MSG) Consensus Group, IA is categorized into proven, probable, and possible invasive fungal disease [<xref ref-type="bibr" rid="CR32">32</xref>]. These diagnostic criteria have proven to be useful in research and practice in severely immunocompromised patients. The lack of specific criteria for diagnosing IA in non-neutropenic patients, however, hampers the timely initiation of appropriate antifungal therapy and may, as such, compromise the odds of survival. Recently, Blot and colleagues [<xref ref-type="bibr" rid="CR33">33</xref>] externally validated a clinical diagnostic algorithm (Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>) that aims to discriminate colonization from probable IA in ICU patients with <italic>Aspergillus</italic>-positive endotracheal aspirate cultures.<table-wrap id="Tab1"><label>Table 1</label><caption><p> <bold>Clinical algorithm for the diagnosis of invasive aspergillosis in non-neutropenic patients</bold> </p></caption><table frame="hsides" rules="groups"><tbody><tr valign="top"><td colspan="2"> <bold>Proven invasive pulmonary aspergillosis</bold> </td></tr><tr valign="top"><td> <bold>&#x025cf;</bold> </td><td> <bold>Follow EORTC/MSG criteria</bold> </td></tr><tr valign="top"><td colspan="2"> <bold>Putative invasive pulmonary aspergillosis (all four criteria must be met)</bold> </td></tr><tr valign="top"><td> <bold>&#x025cf;</bold> </td><td> <bold>1. Aspergillus-positive lower respiratory tract specimen culture</bold> </td></tr><tr valign="top"><td> <bold>&#x025cf;</bold> </td><td> <bold>2. Compatible signs and symptoms (one of the following)</bold> </td></tr><tr valign="top"><td/><td>&#x02003;&#x02003; <bold>Fever refractory to at least 3&#x000a0;days of appropriate antibiotic therapy</bold> </td></tr><tr valign="top"><td/><td>&#x02003;&#x02003; <bold>Recrudescent fever after a period of defervescence of at least 48&#x000a0;hours</bold> <bold>while still on antibiotics and without other apparent cause</bold> </td></tr><tr valign="top"><td/><td>&#x02003;&#x02003; <bold>Pleuritic chest pain</bold> </td></tr><tr valign="top"><td/><td>&#x02003;&#x02003; <bold>Pleuritic rub</bold> </td></tr><tr valign="top"><td/><td>&#x02003;&#x02003; <bold>Dyspnea</bold> </td></tr><tr valign="top"><td/><td>&#x02003;&#x02003; <bold>Hemoptysis</bold> </td></tr><tr valign="top"><td/><td>&#x02003;&#x02003; <bold>Worsening respiratory insufficiency in spite of appropriate antibiotic therapy and ventilatory support</bold> </td></tr><tr valign="top"><td> <bold>&#x025cf;</bold> </td><td> <bold>3. Abnormal medical imaging by portable chest X-ray or CT scan of the lungs</bold> </td></tr><tr valign="top"><td> <bold>&#x025cf;</bold> </td><td> <bold>4. Either 4a or 4b</bold> </td></tr><tr valign="top"><td/><td>&#x02003;&#x02003; <bold>4a. Host risk factors (one of the following conditions)</bold> </td></tr><tr valign="top"><td/><td>&#x02003;&#x02003;&#x02002;&#x02003; <bold>Neutropenia preceding or at the time of ICU admission</bold> </td></tr><tr valign="top"><td/><td>&#x02003;&#x02003;&#x02002;&#x02003; <bold>Underlying hematological or oncological malignancy treated with cytotoxic agents</bold> </td></tr><tr valign="top"><td/><td>&#x02003;&#x02003;&#x02002;&#x02003; <bold>Glucocorticoid treatment (prednisone equivalent, 20&#x000a0;mg/day)</bold> </td></tr><tr valign="top"><td/><td>&#x02003;&#x02003;&#x02002;&#x02003; <bold>Congenital or acquired immunodeficiency</bold> </td></tr><tr valign="top"><td/><td>&#x02003;&#x02003;&#x02002;&#x02003; <bold>COPD, decompensated cirrhosis</bold> </td></tr><tr valign="top"><td/><td>&#x02003;&#x02003; <bold>4b. </bold> <bold>Semiquantitative </bold> <bold><italic>Aspergillus</italic></bold> <bold>-positive culture of BAL fluid without bacterial growth together with a positive cytological smear showing branching hyphae</bold> </td></tr><tr valign="top"><td colspan="2"> <bold><italic>Aspergillus</italic></bold> <bold>respiratory tract colonization</bold> </td></tr><tr valign="top"><td> <bold>&#x025cf;</bold> </td><td> <bold>When &#x0003e;1 criterion necessary for a diagnosis of putative IPA is not met, the case is classified as</bold> <bold><italic>Aspergillus</italic></bold> <bold>colonization</bold> </td></tr></tbody></table><table-wrap-foot><p>BAL, bronchoalveolar lavage; COPD, chronic obstructive pulmonary disease; CT, computed tomography; EORTC/MSG, European Organization for Research and Treatment of Cancer/Invasive Fungal Infections Cooperative Group and the National Institute of Allergy and Infectious Diseases Mycoses Study Group; IPA, invasive pulmonary aspergillosis.</p></table-wrap-foot></table-wrap></p></sec><sec id="Sec6"><title>Microbiological diagnosis</title><p>The microbiological diagnosis of aspergillosis can be achieved using conventional and molecular approaches, including antigen detection and PCR assays [<xref ref-type="bibr" rid="CR34">34</xref>,<xref ref-type="bibr" rid="CR35">35</xref>]. The direct examination of clinical specimens by microscopy is particularly relevant to observe the fungal parasitism; this morphology can allow a presumptive diagnosis of aspergillosis. Microscopy is generally performed using wet preparations (potassium hydroxide, calcofluor) and Wright or Giemsa stains. Other specialized stains, like periodic acid-Schiff or Gomori methenamine silver, are generally performed in the histology laboratory [<xref ref-type="bibr" rid="CR34">34</xref>].</p><p>Since <italic>Aspergillus</italic> microscopic fungal elements can be confused with those of <italic>Fusarium</italic> and <italic>Scedosporium</italic> species, conventional culture methods are essential for isolating and identifying the etiological agent. Identification is largely based on an accurate analysis of the macro- and microscopic features of the colonies: the size, color and shape of the colony, microscopic visualization of conidiophores and conidial heads, morphology size and color of the conidia are important features useful to identify the isolate at the species level [<xref ref-type="bibr" rid="CR34">34</xref>,<xref ref-type="bibr" rid="CR36">36</xref>]. More recently, DNA sequencing and the matrix-assisted laser desorption/ionization-time of flight mass spectrometry proteomic approach have proven to be useful tools to identify non-sporulating isolates or isolates with atypical morphology [<xref ref-type="bibr" rid="CR37">37</xref>,<xref ref-type="bibr" rid="CR38">38</xref>]. It should be remembered that a negative result by both microscopy and culture does not exclude an active infection. The availability of clinical <italic>Aspergillus</italic> spp. isolates allows <italic>in vitro</italic> antifungal susceptibility testing, which can be useful to detect the emergence of resistance, especially to triazoles [<xref ref-type="bibr" rid="CR39">39</xref>].</p><p>The detection of antibodies against <italic>Aspergillus</italic> is strongly dependent on the immune status of the patient and has been proven to be of little value in the diagnosis of IA [<xref ref-type="bibr" rid="CR35">35</xref>].</p><p>The Platelia <italic>Aspergillus</italic> enzyme immunoassay (Bio-Rad Laboratories, Redmond, WA, USA) reveals the presence of GM, a polysaccharide of the outer cell wall layer of <italic>Aspergillus</italic>, in patients with suspected aspergillosis [<xref ref-type="bibr" rid="CR34">34</xref>,<xref ref-type="bibr" rid="CR35">35</xref>]. Because GM is produced at the apical hyphae of actively growing <italic>Aspergillus</italic>, the performance of this immunoassay decreases when antifungal therapy is successful [<xref ref-type="bibr" rid="CR34">34</xref>]. GM can be detected in body fluids, but serum levels in non-neutropenic patients do not seem to be accurate because circulating neutrophils are able to clear the antigen. Meersseman and colleagues [<xref ref-type="bibr" rid="CR40">40</xref>] demonstrated a high sensitivity and specificity of GM in bronchoalveolar lavage (BAL) for the diagnosis of IA; the sensitivity of BAL GM was 88% compared with 40% for serum GM. GM detection in BAL is, therefore, a valuable tool for the diagnosis of IA also in non-neutropenic patients. Alternatively, we could test for 1,3-&#x003b2;-D-glucan, a cell-wall component of many fungi, in sera of patients with suspected aspergillosis.</p><p>Encouraging results have been obtained using PCR techniques (that is, real time, nested) to detect <italic>Aspergillus</italic> DNA in the sera of patients with proven and probable aspergillosis. Although these tests have the advantage of being non-invasive and EU approved real time PCR kits could overcome the problems related to the absence of a standardized methodology, molecular detection of nucleic acids is not yet considered sufficiently reliable for use in the diagnosis of IA [<xref ref-type="bibr" rid="CR32">32</xref>,<xref ref-type="bibr" rid="CR41">41</xref>]. Moreover, conflicting results have been described in cases of histologically proven invasive aspergillosis when the PCR method was performed on BAL [<xref ref-type="bibr" rid="CR42">42</xref>,<xref ref-type="bibr" rid="CR43">43</xref>].</p></sec><sec id="Sec7"><title>Therapeutic approaches</title><p>Prompt administration of appropriate antifungal therapies for IA are immensely important to limit its mortality rate, which ranges from 60% to 90% [<xref ref-type="bibr" rid="CR16">16</xref>]. Hence, even patients without classic risk factors (that is, COPD, steroids and immunosuppressive agent use, hepatic failure, ICU-related immunoparalysis) should start adequate antifungal therapy upon suspicion of IA before obtaining definitive proof of infection. Early treatment initiation according to first-line therapy, at the stage of possible infection, has been reported to be associated with improved outcome in a retrospective cohort of 289 IA cases characterized by different predictors of death [<xref ref-type="bibr" rid="CR44">44</xref>].</p><p>Additionally, with the exclusion of neutropenic and allogenic hematopoietic stem cell transplantation recipients, the usefulness of anti-fungal prophylaxis has not been established. In non-neutropenic critically ill patients admitted to the ICU, this preventive approach is thus not recommended [<xref ref-type="bibr" rid="CR45">45</xref>].</p><p>Unlike the setting of febrile neutropenic episodes, there is no consensus about the exact time frame to use before starting empirical therapy without any diagnostic support in other critically ill patients at risk of IA [<xref ref-type="bibr" rid="CR46">46</xref>]. In a 6-year French survey, non-neutropenic patients with IA were less likely to show symptoms; nevertheless, microbiological samples, antigenemia assays and thoracic CT findings had sensitivities similar to those of neutropenic patients [<xref ref-type="bibr" rid="CR6">6</xref>]. In non-neutropenic patients, therefore, a pre-emptive approach based on microbiological biomarkers (GM, <italic>Aspergillus</italic> PCR, 1,3-beta-glucan) may be useful and should be implemented for early detection and prompt treatment of invasive fungal infections in the ICU [<xref ref-type="bibr" rid="CR11">11</xref>,<xref ref-type="bibr" rid="CR47">47</xref>].</p><p>Three classes of antifungal agents are available for the treatment of IA: azoles (voriconazole, posaconazole, itraconazole), amphotericin B, and echinocandins (Table&#x000a0;<xref rid="Tab2" ref-type="table">2</xref>). Current guidelines recommend voriconazole as first-line treatment for IA, including severely critically ill patients, where intravenous administration is preferred [<xref ref-type="bibr" rid="CR48">48</xref>]. During the past 10&#x000a0;years, voriconazole use has been widely and progressively used. In a randomized controlled trial in 2002 involving 277 patients with IA mainly affected by hematologic diseases, voriconazole use compared with amphotericin B was associated with statistically significant higher successful outcomes, survival rates and fewer severe adverse events [<xref ref-type="bibr" rid="CR49">49</xref>]. Voriconazole was the main antifungal used for the treatment of IA during a large prospective surveillance study conducted in North America between 2004 and 2008 [<xref ref-type="bibr" rid="CR50">50</xref>]. In a retrospective study of 289 IA patients, the authors observed that, after October 2002 (when amphotericin B formulations where replaced by voriconazole as the first-line anti-<italic>Aspergillus</italic> treatment), the overall survival rate increased from 47.5% to 60.4% (<italic>P</italic>&#x02009;=&#x02009;0.01), without concomitant modifications regarding diagnostic strategy [<xref ref-type="bibr" rid="CR44">44</xref>]. Recently, Burghi and colleagues [<xref ref-type="bibr" rid="CR51">51</xref>] analyzed data from 67 patients admitted to ICU with acute respiratory failure due to infection with <italic>Aspergillus</italic> spp<italic>.</italic> Voriconazole therapy was independently associated with lower mortality, confirming its primary role in the management of IA. A large retrospective cohort study investigating risk factors and outcome of ICU patients with IA (excluding those with classic risk factors) showed that a 1-day delay in starting effective antifungal therapy was associated with a longer length of stay (by 1.28&#x000a0;days) and 4% higher total costs per day (<italic>P</italic>&#x02009;&#x0003c;&#x02009;0.001). Voriconazole was the most frequently prescribed antifungal and its use appeared to improve the abovementioned outcome measures [<xref ref-type="bibr" rid="CR52">52</xref>]. Data collected from a large multinational randomized controlled trial, involving mainly hematological and transplanted patients, confirmed better outcomes for patients treated with voriconazole compared with conventional amphotericin B, even though total treatment costs were similar [<xref ref-type="bibr" rid="CR53">53</xref>].<table-wrap id="Tab2"><label>Table 2</label><caption><p> <bold>Treatment of invasive aspergillosis in non-neutropenic ICU patients</bold> [<xref ref-type="bibr" rid="CR48">48</xref>]</p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th> <bold>Setting</bold> </th><th> <bold>First choice</bold> </th><th> <bold>Alternatives</bold> </th></tr></thead><tbody><tr valign="top"><td>Primary therapy</td><td>Voriconazole (6&#x000a0;mg/kg every 12&#x000a0;hours intravenously on day 1, then 4&#x000a0;mg/kg every 12&#x000a0;hours intravenously)</td><td>Liposomial amphotericin B (3&#x02013;5&#x000a0;mg/kg/day intravenously) or Echinocandins (usual dosage)</td></tr><tr valign="top"><td>Salvage therapy</td><td colspan="2">Combination of voriconazole plus amphotericin B/echinocandins</td></tr></tbody></table></table-wrap></p><p>Itraconazole is considered a second-line therapeutic agent for the treatment of IA, especially in severely ill patients. However, its oral use has been described in non-life-threatening infections where the patients had already been stabilized with a more potent agent [<xref ref-type="bibr" rid="CR54">54</xref>]. Posaconazole is a broad-spectrum triazole with anti-<italic>Aspergillus</italic> activity similar to that of voriconazole. In a retrospective case&#x02013;control study involving 193 patients with IA and other mycoses, its use was associated with a 42% survival rate [<xref ref-type="bibr" rid="CR48">48</xref>]. However, limited clinical experience with it and the absence of intravenous formulations strongly reduce its applicability in critically ill patients. Although rare, triazole resistance in <italic>Aspergillus</italic> spp. (that is, <italic>Aspergillus fumigatus</italic>) has been reported. In these cases, alternative antifungal treatment should be adopted [<xref ref-type="bibr" rid="CR55">55</xref>].</p><p>Before the introduction of voriconazole, amphotericin B was the main treatment for IA. The deoxycholate formulation was associated with severe nephrotoxicity, infusion-related adverse events (fever, chills, arthralgias), and poor outcomes. Three lipid formulations have been approved and are associated with fewer renal toxicity and drug-related side effects, although optimal dosages have not been defined for any of these compounds [<xref ref-type="bibr" rid="CR56">56</xref>]. In a population of 201 patients with confirmed IA, Cornely and colleagues [<xref ref-type="bibr" rid="CR57">57</xref>] demonstrated that patients who received a high dose of liposomal amphotericin B (10&#x000a0;mg/kg/day) did not experience higher cure rates compared with standard doses, although relevant nephrotoxicity was observed. In a retrospective cohort of 16 COPD patients with IA treated with a deoxycholate formulation, the mortality rate was 100%, mainly due to septic shock or multiorgan failure. This poor prognosis raised doubts about the need for higher doses or lipid formulations in specific subgroups of patients [<xref ref-type="bibr" rid="CR3">3</xref>].</p><p>All echinocandins have been shown to have <italic>in vitro</italic> and <italic>in vivo</italic> activity against <italic>Aspergillus</italic> spp<italic>.</italic> However, only caspofungin is approved for the treatment of IA in patients who are intolerant to first-line compounds [<xref ref-type="bibr" rid="CR48">48</xref>]. In two phase II studies involving leukemic and hematopoietic stem cell transplantation patients treated with caspofungin, 12-week survival exceeded 50% [<xref ref-type="bibr" rid="CR58">58</xref>]. Although still not approved, two other echinocandins (anidulafungin and micafungin) are used in clinical practice, especially with non-neutropenic patients. In breakthrough IA and refractory diseases, combination therapy (for example, echinocandin plus voriconazole or liposomal amphotericin B) may be considered.</p><p>Although limited by the use of historical controls, some studies suggest the benefits of voriconazole-caspofungin combinations [<xref ref-type="bibr" rid="CR59">59</xref>,<xref ref-type="bibr" rid="CR60">60</xref>]. Furthermore, in a subgroup of 40 solid organ transplant recipients, this combination, as first-line therapy, was associated with significantly reduced mortality compared with amphotericin B [<xref ref-type="bibr" rid="CR61">61</xref>]. Similarly, a caspofungin-amphotericin B combination has been used with a more than 50% favorable antifungal response [<xref ref-type="bibr" rid="CR62">62</xref>,<xref ref-type="bibr" rid="CR63">63</xref>]. On the other hand, no clinical data support triazole-amphotericin B combinations due to possible antagonistic interactions. A phase III clinical trial investigating the effectiveness of a voriconazole-anidulafungin combination did not provide conclusive results [<xref ref-type="bibr" rid="CR64">64</xref>]. All-cause mortality rates at week 6 for proven or probable IA cases was 19.3% in the voriconazole-anidulafungin group versus 27.5% in the voriconazole group. A recent meta-analysis on combination therapy for IA concluded that the available clinical evidence is not conclusive and of moderate strength [<xref ref-type="bibr" rid="CR65">65</xref>].</p><p>The optimal duration of IA treatment is not known. Early assessment of treatment response is essential to confirm effectiveness. The site of infection, immunosuppressive status, baseline clinical conditions and subsequent therapeutic interventions may all influence physicians&#x02019; decisions. Generally, antifungals are not interrupted until all clinical signs have disappeared and radiological abnormalities have stabilized.</p><p>Recommendations regarding management of IA in non-neutropenic patients principally derive from evidence from hematological population studies. Large observational cohort studies and interventional trials are needed in order to define the most appropriate therapeutic approaches in non-neutropenic critically ill ICU patients.</p><sec id="Sec8"><title>Pharmacological properties of voriconazole</title><p>One of the main pharmacokinetic parameters of voriconazole is its excellent oral bioavailability [<xref ref-type="bibr" rid="CR66">66</xref>-<xref ref-type="bibr" rid="CR68">68</xref>]. It possesses the highest bioavailability among triazoles (&#x0003e;85 to 90%), which results in rapidly high plasma concentrations. The absorption of voriconazole is not affected by gastric pH but is decreased by co-administration with food [<xref ref-type="bibr" rid="CR69">69</xref>]. Peak plasma concentrations close to steady state are rapidly achieved via an intravenous loading dose followed by a maintenance dose within the first 24&#x000a0;hours of administration, but only after 5 to 7&#x000a0;days following multiple oral administrations. Thus, the intravenous route seems to be preferable for initial administration of voriconazole in critically ill patients suffering from IA in order to achieve therapeutic voriconazole levels as early as possible.</p><p>An analysis of pharmacokinetic data from several voriconazole clinical trials showed that median voriconazole plasma concentrations in older patients (&#x0003e;65&#x000a0;years) were approximately 80% to 90% higher than those in younger patients after both intravenous and oral administration [<xref ref-type="bibr" rid="CR70">70</xref>]. The estimated voriconazole oral bioavailability was lower (60%) than previously observed, which might be explained by altered gastrointestinal function, which is frequent in onco-hematological patients [<xref ref-type="bibr" rid="CR70">70</xref>]. Voriconazole is mainly eliminated by the liver, while kidney elimination is negligible, and less than 5% of the active drug is found in urine.</p><p>Voriconazole achieves therapeutically effective concentrations in the epithelial lining fluid after standard doses [<xref ref-type="bibr" rid="CR71">71</xref>-<xref ref-type="bibr" rid="CR73">73</xref>]. A recent experience assessing trough voriconazole concentrations in plasma and pulmonary epithelial lining fluid of lung transplant recipients receiving oral voriconazole showed a very high mean&#x02009;&#x000b1;&#x02009;standard deviation epithelial lining fluid/plasma ratio [<xref ref-type="bibr" rid="CR74">74</xref>]. This may by predictive of its efficacy in the treatment of pulmonary aspergillosis. Additionally, voriconazole is extensively transported across the blood&#x02013;brain and blood-eye barriers [<xref ref-type="bibr" rid="CR73">73</xref>,<xref ref-type="bibr" rid="CR75">75</xref>,<xref ref-type="bibr" rid="CR76">76</xref>]. A recent reference laboratory experience of clinically achievable voriconazole concentrations within cerebrospinal fluid (CSF) showed that, among 173 samples, the median quantifiable CSF level was 2.47&#x000a0;mg/L [<xref ref-type="bibr" rid="CR77">77</xref>]. The effective levels in CSF may support the results of a recent retrospective analysis assessing the efficacy of voriconazole in the treatment of 192 fungal central nervous system infections that documented a success rate of 48% [<xref ref-type="bibr" rid="CR78">78</xref>].</p><p>Variability of voriconazole serum concentrations is mainly due to metabolism via the CYP2C19 P450 enzyme [<xref ref-type="bibr" rid="CR79">79</xref>-<xref ref-type="bibr" rid="CR81">81</xref>]. Standard dosing in adults is outlined in Table&#x000a0;<xref rid="Tab3" ref-type="table">3</xref>.<table-wrap id="Tab3"><label>Table 3</label><caption><p> <bold>Standard dosing in adults and children</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th colspan="2"> <bold>Adults and adolescents (&#x0003e;12&#x000a0;years and &#x0003e;50&#x000a0;kg)</bold> </th></tr></thead><tbody><tr valign="top"><td colspan="2">Loading dose, for the first 24&#x000a0;hours</td></tr><tr valign="top"><td>&#x025cf;</td><td>IV: 6&#x000a0;mg/kg every 12&#x000a0;hours</td></tr><tr valign="top"><td>&#x025cf;</td><td>Oral &#x0003e;40&#x000a0;kg: 400&#x000a0;mg every 12&#x000a0;hours</td></tr><tr valign="top"><td>&#x025cf;</td><td>Oral &#x0003c;40&#x000a0;kg: 200&#x000a0;mg every 12&#x000a0;hours</td></tr><tr valign="top"><td colspan="2">Maintenance dose</td></tr><tr valign="top"><td>&#x025cf;</td><td>IV: 4&#x000a0;mg/kg every 12&#x000a0;hours</td></tr><tr valign="top"><td>&#x025cf;</td><td>Oral &#x0003e;40&#x000a0;kg: 200&#x000a0;mg every 12&#x000a0;hours</td></tr><tr valign="top"><td>&#x025cf;</td><td>Oral &#x0003c;40&#x000a0;kg: 100&#x000a0;mg every 12&#x000a0;hours</td></tr></tbody></table><table-wrap-foot><p>IV, intravenous.</p></table-wrap-foot></table-wrap></p><p>If a response to voriconazole is inadequate, the maintenance oral dose may be increased to 300&#x000a0;mg every 12&#x000a0;hours for patients weighing over 40&#x000a0;kg and to 150&#x000a0;mg every 12&#x000a0;hours for those &#x0003c;40&#x000a0;kg. Dose adjustment is required in case of hepatic failure. According to the prescribing information summary, dose adjustments are required for patients with mild to moderate hepatic dysfunction (Child-Pugh class A and B). The standard loading dose should be provided to these patients, but maintenance doses should be reduced by 50%. Studies have not adequately evaluated the safety of voriconazole in severe liver disease (Child-Pugh class C) [<xref ref-type="bibr" rid="CR82">82</xref>]. Caution should be exercised when administering the intravenous formulation to critically ill patients with renal dysfunction due to the presence of the solubilizing excipient sulfobutylether-beta-cyclodextrin. Indeed, two recent clinical experiences assessing the safety of intravenous voriconazole in patients with compromised renal function showed that the route of administration and baseline renal function were not predictors of worsening renal dysfunction [<xref ref-type="bibr" rid="CR83">83</xref>,<xref ref-type="bibr" rid="CR84">84</xref>].</p><p>Although voriconazole has many drug interactions, their clinical management can be relatively simple (Table&#x000a0;<xref rid="Tab4" ref-type="table">4</xref>) [<xref ref-type="bibr" rid="CR85">85</xref>].<table-wrap id="Tab4"><label>Table 4</label><caption><p> <bold>The main drugs interacting with voriconazole</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th> <bold>Drug</bold> </th><th> <bold>Interaction with voriconazole and management strategy</bold> </th></tr></thead><tbody><tr valign="top"><td>Drugs contraindicated</td><td/></tr><tr valign="top"><td>&#x02003;&#x02009;Astemizole, cisapride, ergot alkaloids, quinidine, sirolimus, terfenadine</td><td>Their levels are increased by voriconazole, avoid co-administration. Switch to a drug with no or with predictable interactions (for example, cyclosporine)</td></tr><tr valign="top"><td>&#x02003;&#x02009;Carbamazepine, long-acting barbiturates, rifampicin</td><td>They decrease voriconazole levels, avoid co-administration. Switch to a drug with no interactions (for example, levetiracetam)</td></tr><tr valign="top"><td>&#x02003;&#x02009;Rifabutin</td><td>Co-administration decreases voriconazole levels and increases rifabutin levels (contraindicated according to FDA, not according to EMA, see below), avoid co-administration</td></tr><tr valign="top"><td>Drugs not contraindicated but if co-administered the dose of voriconazole must be modified (increased)</td><td/></tr><tr valign="top"><td>&#x02003;&#x02009;Phenytoin</td><td>Increase voriconazole oral maintenance dose from 200&#x000a0;mg to 400&#x000a0;mg every 12&#x000a0;hours (100&#x02013;200&#x000a0;mg every 12&#x000a0;hours if &#x0003c;40&#x000a0;kg) and intravenous maintenance dose to 5&#x000a0;mg/kg every 12&#x000a0;hours; monitor for phenytoin toxicity</td></tr><tr valign="top"><td>&#x02003;&#x02009;Efavirenz</td><td>Increase voriconazole oral maintenance dose from 200&#x000a0;mg to 400&#x000a0;mg every 12&#x000a0;hours (100&#x02013;200&#x000a0;mg every 12&#x000a0;hours if &#x0003c;40&#x000a0;kg) and reduce efavirenz dose by 50% to 300&#x000a0;mg/day</td></tr><tr valign="top"><td>&#x02003;&#x02009;Rifabutin (according to FDA contraindicated as rifampicin)</td><td>According to EMA, increase oral voriconazole maintenance dose from 200 to 350&#x000a0;mg every 12&#x000a0;hours (100&#x02013;200&#x000a0;mg every 12&#x000a0;hours if &#x0003c;40&#x000a0;kg) and intravenous maintenance dose to 5&#x000a0;mg/kg every 12&#x000a0;hours; monitor for rifabutin toxicity</td></tr><tr valign="top"><td>Other drugs (apart from ritonavir, their levels are increased by voriconazole)</td><td/></tr><tr valign="top"><td>&#x02003;&#x02009;Low dose ritonavir (100&#x000a0;mg every 12&#x000a0;hours)</td><td>Co-administration decreases levels of both voriconazole and ritonavir; better avoided</td></tr><tr valign="top"><td>&#x02003;&#x02009;Cyclosporine, omeprazole, tacrolimus and warfarin</td><td>Their blood levels are increased by voriconazole and their dose should be reduced (by half for cyclosporine and by two-thirds for tacrolimus). Monitor serum levels of cyclosporine and tacrolimus or INR for warfarin</td></tr><tr valign="top"><td>&#x02003;&#x02009;Other drugs such as benzodiazepines, opioid analgesics (for example, oxycodone or fentanyl), sulfonylureas, statins, vinca alkaloids, calcium channel blockers</td><td>Their levels are increased by voriconazole co-administration. Monitor closely for their side effects, discontinue if toxicity is suspected or consider decreasing dosage immediately when voriconazole is started</td></tr></tbody></table><table-wrap-foot><p>EMA, European Medicines Agency; FDA, Food and Drug Administration; INR, international normalized ratio.</p></table-wrap-foot></table-wrap></p><p>As far as voriconazole dosing in special populations is concerned, supratherapeutic concentrations (4&#x000a0;mg/kg actual body weight) have recently been documented as a risk [<xref ref-type="bibr" rid="CR86">86</xref>]. Therefore, dosing voriconazole based on an ideal body weight or adjusted body weight has been recommended for morbidly obese patients [<xref ref-type="bibr" rid="CR86">86</xref>,<xref ref-type="bibr" rid="CR87">87</xref>]. Conversely, clearance of voriconazole during continuous veno-venous hemofiltration (CVVH) was not clinically significant, so voriconazole dose adjustment in critically ill patients undergoing the standard method of CVVH is not required [<xref ref-type="bibr" rid="CR88">88</xref>].</p><p>Several recent papers have underlined the crucial role of adequate plasma levels for maintaining efficacy during treatment of invasive fungal infections in immunocompromised patients [<xref ref-type="bibr" rid="CR87">87</xref>,<xref ref-type="bibr" rid="CR89">89</xref>-<xref ref-type="bibr" rid="CR91">91</xref>]. A trough concentration of at least 1&#x000a0;mg/L was associated with an approximately 70% response rate in adult patients, and to date the recommended range is between 1 and 5.5&#x000a0;mg/L [<xref ref-type="bibr" rid="CR70">70</xref>]. Interestingly, a reference laboratory experience of clinically achievable voriconazole bloodstream concentrations in a large number of subjects (n&#x02009;=&#x02009;14,370) showed that 50.6% of samples were within the recommended trough range [<xref ref-type="bibr" rid="CR77">77</xref>].</p><p>Although we still await definitive evidence-based guidelines on therapeutic drug monitoring of voriconazole, some practical indications, listed in order of importance, are summarized in Table&#x000a0;<xref rid="Tab5" ref-type="table">5</xref>.<table-wrap id="Tab5"><label>Table 5</label><caption><p> <bold>Practical indications, listed in order of importance, when therapeutic drug monitoring of voriconazole might be useful</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr valign="top"><th colspan="2"> <bold>Clinical situation</bold> </th></tr></thead><tbody><tr valign="top"><td>&#x025cf;</td><td>Suspected treatment failure</td></tr><tr valign="top"><td>&#x025cf;</td><td>Suspected suboptimal dosing - for example, due to interaction with other drugs such as phenytoin, in children or in cerebral infections<break/>Change in the administration of the drug from intravenous to oral route^a </td></tr><tr valign="top"><td>&#x025cf;</td><td>Suspected suboptimal absorption</td></tr><tr valign="top"><td>&#x025cf;</td><td>Suspected non-compliance</td></tr><tr valign="top"><td>&#x025cf;</td><td>Suspected neurologic toxicity possibly related to overdosing</td></tr><tr valign="top"><td>&#x025cf;</td><td>Suspected other toxicity possibly related to overdosing</td></tr></tbody></table><table-wrap-foot><p> ^aAs long as the patient is critical, intravenous therapy is preferred in order to avoid problems with absorption.</p></table-wrap-foot></table-wrap></p></sec><sec id="Sec9"><title>Pharmacological properties of echinocandins</title><p>The echinocandins are semisynthetic lipopeptides that act as noncompetitive inhibitors of 1,3-beta-D-glucan synthase, an enzyme complex within the fungal cell wall [<xref ref-type="bibr" rid="CR92">92</xref>]. All the echinocandins exert <italic>in vitro</italic> and <italic>in vivo</italic> activity against <italic>Aspergillus</italic> spp. [<xref ref-type="bibr" rid="CR93">93</xref>].</p><p>From a pharmacokinetic standpoint, the echinocandins are all similar for some aspects but differ for others [<xref ref-type="bibr" rid="CR92">92</xref>]. All are highly bound to plasma protein, do not diffuse through the blood&#x02013;brain barrier and/or the blood-ocular barrier, have a low propensity for drug-drug pharmacokinetic interaction (especially anidulafungin), are not renally cleared and have elimination half-lives long enough to allow once-daily administration. Recent studies suggest that the influence of continuous renal replacement therapy on anidulafungin, caspofungin or micafungin elimination in critically ill patients appears to be negligible, and that no dosage adjustments are needed for the echinocandins in patients undergoing CVVH ) [<xref ref-type="bibr" rid="CR94">94</xref>-<xref ref-type="bibr" rid="CR98">98</xref>].</p><p>It has been shown that hypoalbuminemic post-surgical patients might experience caspofungin underexposure due to increased clearance as a result of decreased plasma protein binding [<xref ref-type="bibr" rid="CR99">99</xref>]. Likewise, a recent study in critically ill patients suggested that standard doses of anidulafungin resulted in lower exposure than in the general patient population, even if no correlation between anidulafungin exposure and plasma protein concentrations was established [<xref ref-type="bibr" rid="CR99">99</xref>]. Additionally, it has been shown that dose optimization of caspofungin in obese patients may improve clinical success rates [<xref ref-type="bibr" rid="CR100">100</xref>].</p><p>Although these issues are not expected to greatly affect echinocandin efficacy against <italic>Candida</italic> strains [<xref ref-type="bibr" rid="CR101">101</xref>], they might become more relevant in the presence of less susceptible pathogens.</p><p>Although caspofungin is approved for second-line management of proven or probable IA at the standard dose of 50&#x000a0;mg once daily, it is worth noting that currently ongoing pharmacokinetic studies in patients with IA with higher doses ranging between 70 and 200&#x000a0;mg once daily suggest linear pharmacokinetics with no unpredictable accumulation across the investigated dosage range and good safety [<xref ref-type="bibr" rid="CR102">102</xref>,<xref ref-type="bibr" rid="CR103">103</xref>].</p></sec><sec id="Sec10"><title>Pharmacological properties of liposomal amphotericin B</title><p>Amphotericin B is a polyene antibiotic that binds to the ergosterol present in the fungal membrane. Among the various lipidic formulations of amphotericin B, liposomal amphotericin B (LAmB) has the more favorable pharmacokinetic behavior in terms of achieving higher peak plasma levels, having lower intracellular penetration rates and lower clearance through the reticuloendothelial system [<xref ref-type="bibr" rid="CR104">104</xref>]. Interestingly, both LAmB and amphotericin B lipid complex (ABLC) were shown to achieve therapeutically effective concentrations in the epithelial lining fluid of critically ill patients [<xref ref-type="bibr" rid="CR105">105</xref>]. However, experimental animal models suggest that only LAmB may achieve adequate levels in the CSF [<xref ref-type="bibr" rid="CR106">106</xref>] and the eye [<xref ref-type="bibr" rid="CR107">107</xref>].</p><p>The pharmacokinetic-pharmacodynamic relationships of the two most widely used lipid formulations of amphotericin B (LAmB and ABLC) were shown to differ markedly in an <italic>in vitro</italic> lung model of IA, considering that the concentrations producing a 50% maximal effect were about four-fold lower for LAmB than for ABLC [<xref ref-type="bibr" rid="CR108">108</xref>].</p><p>As far as LAmB dosing is concerned, it has been shown that dosages up to 10&#x000a0;mg/kg/daily gave no benefit for treatment of IA in comparison with the standard dose of 3 to 5&#x000a0;mg/kg/daily [<xref ref-type="bibr" rid="CR57">57</xref>,<xref ref-type="bibr" rid="CR109">109</xref>]. However it is worth noting that alternative dosing schedules based on higher dosages at longer dosing intervals are currently under evaluation for both prophylactic [<xref ref-type="bibr" rid="CR110">110</xref>] and therapeutic [<xref ref-type="bibr" rid="CR111">111</xref>] purposes.</p><p>Although potentially nephrotoxic, LAmB does not need dosage adjustment in the presence of renal insufficiency and recent clinical experience suggests that the impact of LAmB on the renal function of critically ill patients with impaired renal function was minimal [<xref ref-type="bibr" rid="CR112">112</xref>,<xref ref-type="bibr" rid="CR113">113</xref>].</p></sec></sec><sec id="Sec11"><title>Outcome and prognostic factors</title><p>Only a few clinical studies have investigated the outcome of IA in critically ill patients. Different studies are difficult to compare due to the absence of specific clinical signs, different diagnostic criteria and different coexisting diseases recognized as risk factors [<xref ref-type="bibr" rid="CR16">16</xref>].</p><p>Mortality rates for patients with proven or probable IA in the ICU range from 59% to 95% and seem to be higher in non-neutropenic patients [<xref ref-type="bibr" rid="CR114">114</xref>]. A mortality rate of 60% was observed for immunocompromised patients compared to 89% in non-neutropenic patients (<italic>P</italic>&#x02009;=&#x02009;0.007) [<xref ref-type="bibr" rid="CR6">6</xref>]. In the latter group, fungal infection was proven to be the main cause of death for 8 patients (22.2%). Russo and colleagues [<xref ref-type="bibr" rid="CR115">115</xref>] observed similar results: 14.3% of patients died as a direct consequence of <italic>Aspergillus</italic> infection. The mortality rate in these patients could be greater than in neutropenic patients. Compared to neutropenic patients, non-neutropenic patients could have a less symptomatic fungal infection with a complicated diagnosis, leading to suboptimal management and delayed therapy [<xref ref-type="bibr" rid="CR6">6</xref>].</p><p>In a retrospective analysis of fungal infections in non-neutropenic patients, Garbino and colleagues [<xref ref-type="bibr" rid="CR116">116</xref>] showed a mortality rate of 57.1% for patients with IA. Trof and colleagues [<xref ref-type="bibr" rid="CR11">11</xref>] showed that IA diagnosis was established post-mortem in 38% of patients, 94% of whom did not receive antifungal treatment. These data could explain the results observed by Meersseman and colleagues [<xref ref-type="bibr" rid="CR16">16</xref>] in a restrospective cohort study on 127 ICU patients with IA; patients with proven or probable infection without hematologic malignancy presented a two-fold increase in mortality rate compared with mortality expected by Simplified Acute Physiology Score II score.</p><p>Prognostic factors have been examined in a variety of studies. Isolation of <italic>Aspergillus</italic> in critically ill patients is associated with high mortality, irrespective of invasion or colonization [<xref ref-type="bibr" rid="CR11">11</xref>]. Cornillet and colleagues [<xref ref-type="bibr" rid="CR6">6</xref>] identified three factors associated with a poor prognosis: disseminated infection (100% mortality rate), co-infection (78% mortality rate) and bacterial pneumonia (78.5% mortality rate). In conclusion, it is possible that the overall mortality rate from IA is significantly higher in non-neutropenic patients.</p></sec></sec><sec id="Sec12" sec-type="conclusion"><title>Conclusion</title><p>The management of IA in non-neutropenic, critically ill patients represents a challenge for clinicians. Features of IA in this cohort may contribute to a delay in diagnosis and, consequently, to commencement of adequate antifungal therapy. The complex underlying conditions and the non-specificity of symptoms in non-neutropenic patients may be confounding and lead to underdiagnosis and underestimates of the disease prevalence in this population. Furthermore, current guidelines are mainly designed for recognizing and managing IA in hematological patients with severe and prolonged neutropenia. Although recent advances in microbiological techniques (GM analysis, PCR, and so on) showed promising results in identifying IA also in non-conventional subsets of patients, such as the critically ill, a high level of suspicion of IA should be maintained especially when risk factors (for example, COPD, steroid use) are present. Voriconazole still represents the drug of choice for IA in non-neutropenic patients. Since mortality resulting from IA in non-neutropenic, critically ill patients appears to be higher than in immunocompromised patients and its management is problematic, studies on large cohorts and trials to better define the characteristics of IA are encouraged.</p></sec>

Psychometric comparison of three behavioural scales for the assessment of pain in critically ill patients unable to self-report

Could not extract abstract

<contrib contrib-type="author"><name><surname>Chanques</surname><given-names>Gerald</given-names></name><address><email>g-chanques@chu-montpellier.fr</email></address><xref ref-type="aff" rid="Aff1"/><xref ref-type="aff" rid="Aff2"/><xref ref-type="aff" rid="Aff3"/></contrib><contrib contrib-type="author"><name><surname>Pohlman</surname><given-names>Anne</given-names></name><address><email>apohlman@medicine.bsd.uchicago.edu</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author"><name><surname>Kress</surname><given-names>John P</given-names></name><address><email>jkress@medicine.bsd.uchicago.edu</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author"><name><surname>Molinari</surname><given-names>Nicolas</given-names></name><address><email>nicolas.molinari@inserm.fr</email></address><xref ref-type="aff" rid="Aff4"/></contrib><contrib contrib-type="author"><name><surname>de Jong</surname><given-names>Audrey</given-names></name><address><email>audreydejong@hotmail.fr</email></address><xref ref-type="aff" rid="Aff4"/></contrib><contrib contrib-type="author"><name><surname>Jaber</surname><given-names>Samir</given-names></name><address><email>s-jaber@chu-montpellier.fr</email></address><xref ref-type="aff" rid="Aff2"/><xref ref-type="aff" rid="Aff3"/></contrib><contrib contrib-type="author" corresp="yes"><name><surname>Hall</surname><given-names>Jesse B</given-names></name><address><email>jhall@medicine.bsd.uchicago.edu</email></address><xref ref-type="aff" rid="Aff1"/></contrib><aff id="Aff1"><label/>Department of Medicine, Section of Pulmonary and Critical Care, University of Chicago, 5841 S. Maryland Avenue MC 6076, Chicago, IL 60637 USA </aff><aff id="Aff2"><label/>Department of Anaesthesia and Critical Care Medicine, University of Montpellier Saint Eloi Hospital, 80, Avenue Augustin Fliche, 34295 Montpellier, France </aff><aff id="Aff3"><label/>Unit&#x000e9; U1046 de l&#x02019;Institut National de la Sant&#x000e9; et de la Recherche M&#x000e9;dicale (INSERM), Universit&#x000e9; de Montpellier 1, Universit&#x000e9; de Montpellier 2, 34295 Montpellier, France </aff><aff id="Aff4"><label/>Department of Statistics, University of Montpellier Hospitals, 371, Avenue du Doyen Gaston Giraud, 34295 Montpellier, France </aff>

Critical Care

<sec id="Sec1" sec-type="intro"><title>Introduction</title><p>Pain is a frequent event in Intensive Care Unit (ICU) patients, with an incidence of up to 50% in medical as well as surgical patients [<xref ref-type="bibr" rid="CR1">1</xref>&#x02013;<xref ref-type="bibr" rid="CR3">3</xref>]. Pain is associated with an acute stress response including changes in neurovegetative system activity [<xref ref-type="bibr" rid="CR4">4</xref>], neuroendocrine secretion [<xref ref-type="bibr" rid="CR5">5</xref>, <xref ref-type="bibr" rid="CR6">6</xref>] and psychological distress often manifested as agitation [<xref ref-type="bibr" rid="CR7">7</xref>]. Improved pain management is associated with better patient outcomes in the ICU [<xref ref-type="bibr" rid="CR1">1</xref>, <xref ref-type="bibr" rid="CR8">8</xref>&#x02013;<xref ref-type="bibr" rid="CR10">10</xref>]. However, pain remains currently underevaluated and undertreated [<xref ref-type="bibr" rid="CR3">3</xref>, <xref ref-type="bibr" rid="CR11">11</xref>&#x02013;<xref ref-type="bibr" rid="CR14">14</xref>]. This relates to pain management being challenging in the ICU setting, particularly in patients unable to readily communicate their pain intensity, such as sedated patients and patients with delirium [<xref ref-type="bibr" rid="CR15">15</xref>]. These patients share the common feature of a cognitive dysfunction marked by an impaired level of vigilance. Several behavioural pain scales have been developed in order to standardise the assessment of pain by healthcare providers in those non-communicative patients. The recent Clinical Practice Guidelines for the Management of Pain, Agitation, and Delirium in Adult Patients in the Intensive Care Unit [<xref ref-type="bibr" rid="CR16">16</xref>] stated that both the Behavioural Pain Scale (BPS) [<xref ref-type="bibr" rid="CR17">17</xref>] and the Critical Care Pain Observation Tool (CPOT) [<xref ref-type="bibr" rid="CR18">18</xref>] demonstrated sufficient validity and reliability. However, these scales have never been compared to each other. Thus, we conducted a study in a medical ICU aimed at comparing the psychometric properties of the BPS and CPOT, as well as the Non-verbal Pain Scale (NVPS) [<xref ref-type="bibr" rid="CR19">19</xref>, <xref ref-type="bibr" rid="CR20">20</xref>], which is the usual behavioural pain tool routinely used by nurses at the host institution. Because inter-rater agreement of a pain tool is paramount regarding the necessity to standardise the recognition and treatment of pain by multiple caregivers in complex non-communicative patients, our primary hypothesis was that one pain tool would be superior to others with regard to inter-rater agreement. Secondary endpoints were to evaluate validity, responsiveness and users&#x02019; preference of each tool.</p></sec><sec id="Sec2" sec-type="materials|methods"><title>Materials and methods</title><sec id="Sec3"><title>Ethics approval</title><p>The protocol was approved by the Institutional Review Board of University of Chicago Hospitals (IRB # 11-0691; Protocol Version: 7 November, 2011; Consent Version: 1 December, 2011). Written consent was obtained from the legally authorized representative or a proxy/surrogate decision-maker (patient&#x02019;s next of kin) who gave consent on the patient&#x02019;s behalf.</p></sec><sec id="Sec4"><title>Patient population</title><p>The study took place in the 16-bed medical ICU of the University of Chicago Hospitals, an academic tertiary care hospital, from January 2012 to June 2012 (six months). All consecutive patients &#x02265;18&#x000a0;yrs old were eligible for enrolment if they had a Richmond Agitation Sedation Scale (RASS) [<xref ref-type="bibr" rid="CR21">21</xref>, <xref ref-type="bibr" rid="CR22">22</xref>] above -4 and were unable to self-rate their pain intensity with the Visually Enlarged 0 to 10 Numeric Rating Scale (0 to 10&#x000a0;V-NRS). This scale is adapted to ICU patients and demonstrated to be the most feasible self-report pain scale in the ICU setting [<xref ref-type="bibr" rid="CR23">23</xref>]. Exclusion criteria were neurological disorder, decision to withdraw life-support or unstable condition preventing planned routine care procedures.</p></sec><sec id="Sec5"><title>Conduct of the study</title><p>Investigators screened patients daily for eligibility including RASS assessment, self-report pain ability by the patient and possibilities to plan any routine procedures of care with the bedside nurse. After having obtained consent from the surrogate decision-maker and having enrolled the patient into the study, investigators planned different procedures of care with the bedside nurse including: (1) a simple repositioning of the patient in the bed (moving the patient up or onto their side), (2) a complete turning of the patient onto both sides in order to wash their back and change the sheets, (3) a tracheal suctioning if possible (intubated patients), and (4) a mobilisation by physiotherapist/occupational therapist if possible.</p></sec><sec id="Sec6"><title>Data handling</title><sec id="Sec7"><title>Pain</title><p>Pain evaluation using the three different behavioural pain tools (BPS, CPOT, NVPS) was independently performed at the same time by two or three paired evaluators (one or two investigators, and the bedside nurse) in three conditions for each patient: (1) at rest, before any procedure; (2) during the care procedure; and (3) 10&#x000a0;minutes after the procedure. Every patient was assessed during a simple repositioning and a complete turning on both sides. Patients were evaluated during tracheal suctioning or mobilisation if possible. Turning and suctioning were chosen because they are the most common and/or painful procedures in the ICU setting [<xref ref-type="bibr" rid="CR24">24</xref>, <xref ref-type="bibr" rid="CR25">25</xref>]. Repositioning, turning and mobilisation were chosen so that different intensities of stimulation could be compared to each other.</p><p>For all these measurements, investigators and the bedside nurse were blinded to each other, each observer using a separate sheet (see Additional file <xref rid="MOESM1" ref-type="media">1</xref>). Scale order was determined by randomisation software and printed as a list of combinations before the beginning of the study. Order of occurrence of a given scale was tested to assure that no scale would have a preferred order of occurrence. The randomisation of scale order was considered as a gold standard to take into account any learning effect or, on the contrary, any fatigability during a study procedure incorporating several pain tools [<xref ref-type="bibr" rid="CR26">26</xref>]. The nurse manager and the investigator team informed the bedside nurses about the study purposes before the study began. Moreover, pain tools descriptors and instruction for use were explained to the bedside nurses by the investigator team before the first procedure for each patient. Published educational tools for BPS/BPS-NI [<xref ref-type="bibr" rid="CR27">27</xref>] and CPOT [<xref ref-type="bibr" rid="CR28">28</xref>], as well as the most recent revised version of the NVPS [<xref ref-type="bibr" rid="CR20">20</xref>], were used for this educational purpose in the determined randomised order. Content details of the three tools are given in the additional file (see Additional file <xref rid="MOESM1" ref-type="media">1</xref>). All observers had to rate every domain of the pain tools on a sheet where descriptors of the tools were written to avoid any learning issues (see Additional file <xref rid="MOESM1" ref-type="media">1</xref>). A simplified comparison of the three tools structure is shown in Table&#x000a0; <xref rid="Tab1" ref-type="table">1</xref>. Each of the three tools requires observing three different kinds of behavioural domain related to pain: patient&#x02019;s face, muscular movements and/or tonus, breathing and/or vocalisation. In addition, NVPS requires observing physiological signs (Table&#x000a0; <xref rid="Tab1" ref-type="table">1</xref>).<table-wrap id="Tab1"><label>Table 1</label><caption><p> <bold>Structure comparison of the three behavioural pain tools</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr><th>BPS</th><th>CPOT</th><th>NVPS</th></tr></thead><tbody><tr><td align="center">Number of observation domains</td><td align="center">Number of observation domains</td><td align="center">Number of observation domains</td></tr><tr><td align="center">3</td><td align="center">4</td><td align="center">6</td></tr><tr><td align="center">Number of descriptors per domain</td><td align="center">Number of descriptors per domain</td><td align="center">Number of descriptors per domain</td></tr><tr><td align="center">4 (rated 1 to 4)</td><td align="center">3 (rated 0 to 2)</td><td align="center">3 (rated 0 to 2)</td></tr><tr><td align="center">Total score 3 - 12</td><td align="center">Total score 0 - 8</td><td align="center">Total score 0 - 12</td></tr><tr><td align="center" colspan="3"> <bold>Facial domains</bold> </td></tr><tr><td align="center">Face</td><td align="center">Face</td><td align="center">Face</td></tr><tr><td align="center" colspan="3"> <bold>Breathing domains</bold> </td></tr><tr><td align="center">Mechanical ventilation or vocalisation</td><td align="center">Mechanical ventilation or vocalisation</td><td align="center">Respiration</td></tr><tr><td align="center" colspan="3"> <bold>Muscular domains</bold> </td></tr><tr><td align="center" rowspan="2">Upper limbs movements</td><td align="center">Body movements</td><td align="center">Activity</td></tr><tr><td align="center">Muscle tension</td><td align="center">Guarding</td></tr><tr><td align="center" colspan="3"> <bold>Physiological domains</bold> </td></tr><tr><td align="center" rowspan="2" colspan="2"/><td align="center">Physiological I (vital signs)</td></tr><tr><td align="center">Physiological II (skin and pupils)</td></tr></tbody></table><table-wrap-foot><p>BPS, Behavioral Pain Scale; CPOT, Critical-Care Pain Observation Tool; NVPS, Non-verbal Pain Scale.</p></table-wrap-foot></table-wrap></p><p>Throughout the manuscript, we use the word BPS that includes both BPS and its adaptation for non-intubated patients (BPS-NI), similarly to the CPOT that includes both types of descriptors, either for intubated or non-intubated patients.</p></sec></sec><sec id="Sec8"><title>Demographic and medical data</title><p>Age, gender, height and weight, co-morbidities, and reason for admission to the ICU were recorded. Acute Physiology and Chronic Health Evaluation (APACHE) II score and Sequential Organ Failure Assessment (SOFA) score [<xref ref-type="bibr" rid="CR29">29</xref>] were calculated within 24&#x000a0;hours after ICU admission and before enrolment, respectively. Body mass index (BMI) was calculated as the ratio (kg/m²) between weight (kg) and height squared (m²). Type and doses of sedatives and analgesic drugs were collected before any procedures. In addition to the RASS measurement by investigators, delirium was assessed upon enrolment by the Confusion Assessment Method for the ICU (CAM-ICU) [<xref ref-type="bibr" rid="CR30">30</xref>, <xref ref-type="bibr" rid="CR31">31</xref>]. Physiological parameters (heart and respiratory rates, systolic, diastolic and mean arterial blood pressure, pulse oximetry) were continuously measured through bedside monitoring and retrospectively recorded by investigators to fit with the NVPS description [<xref ref-type="bibr" rid="CR20">20</xref>].</p></sec><sec id="Sec9"><title>Statistical analysis</title><sec id="Sec10"><title>Measurement of psychometric properties</title><p>Psychometric properties related to the use of pain tools were assessed using the new terminology [<xref ref-type="bibr" rid="CR32">32</xref>] as recommended by recent Clinical Practice Guidelines for the Management of Pain, Agitation, and Delirium in Adult Patients in the Intensive Care Unit [<xref ref-type="bibr" rid="CR16">16</xref>]. <list list-type="simple"><list-item><label>1.1</label><p><italic>Inter-rater reliability</italic></p><p>Inter-rater reliability of the three tools (primary endpoint) was tested by the weighted kappa coefficient. A kappa coefficient above 0.80, 0.60 and 0.40 is considered as measuring respectively a &#x02018;near perfect&#x02019;, &#x02018;important&#x02019; and &#x02018;moderate&#x02019; agreement [<xref ref-type="bibr" rid="CR33">33</xref>]. Comparisons of kappa coefficients between scales were made using the z test [<xref ref-type="bibr" rid="CR34">34</xref>].</p><p>To deal with repeated measurements, a sensitivity analysis was performed taking into account first assessments only, as previously described [<xref ref-type="bibr" rid="CR22">22</xref>]. Moreover, the inter-rater agreement within an error of one mark was calculated as the ratio, expressed in percentage, between the number of scores obtained with each scale that differed by not more than one point between different observers, and the total number of scores. Comparisons between scales were made using chi-square test.</p></list-item><list-item><label>1.2</label><p><italic>Internal consistency</italic></p><p>Internal consistency was measured using the Cronbach-&#x003b1; method [<xref ref-type="bibr" rid="CR35">35</xref>]. A Cronbach-&#x003b1; value higher than 0.7 reflects a satisfactory internal consistency, that is a high inter-relation between each domain of the tool [<xref ref-type="bibr" rid="CR35">35</xref>]. Cronbach-&#x003b1; coefficients were compared between the three scales using the method by Feldt [<xref ref-type="bibr" rid="CR36">36</xref>].</p></list-item><list-item><label>1.3</label><p><italic>Discriminant validation</italic></p><p>Discriminant validation was determined by comparing total scores obtained during different situations and stimuli, that is at rest and during a procedure (suctioning, repositioning or turning) as well as during procedures with different durations and intensities, that is during a simple repositioning and during a complete turning. The Mann-Whitney-Wilcoxon test was used to test the difference between two different situations. We tested the responsiveness of the three tools as another way to measure change, that is the ability to detect change regarding different situations even if those changes are small. The magnitude of this property was assessed by the effect size [<xref ref-type="bibr" rid="CR37">37</xref>]. The effect size coefficient is considered small if it is less than 0.20, moderate if it is near 0.50, and large if it is more than 0.80 [<xref ref-type="bibr" rid="CR37">37</xref>]. The modified Jackknife method was used to test any significant difference in responsiveness between two scales [<xref ref-type="bibr" rid="CR38">38</xref>].</p></list-item><list-item><label>1.4</label><p><italic>Feasibility</italic></p><p>Feasibility was assessed by administering a standardised questionnaire once to the bedside nurses during their initial participation in the study interventions. The nurses were asked to rate their preference of each particular pain scale, as well as the degree of accuracy when used for routine practice or research purposes, and the ease of learning.</p></list-item></list></p></sec><sec id="Sec11"><title>Primary endpoint and power analysis</title><p>The primary endpoint was the inter-rater reliability because this psychometric property is paramount and, if deficient, precludes implementation of a pain tool and associated diagnostic and therapeutic pain strategies by the ICU team [<xref ref-type="bibr" rid="CR1">1</xref>, <xref ref-type="bibr" rid="CR4">4</xref>, <xref ref-type="bibr" rid="CR16">16</xref>]. The number of paired assessments (assessment by investigators&#x02009;+&#x02009;assessment by the ICU clinical staff) needed to show a weighted kappa difference of 0.1 from a given kappa of 0.80 (&#x000b1;0.10), with an &#x003b1; of 0.05 and a &#x003b2; of 0.20, was determined to be n&#x02009;=&#x02009;167 paired assessments. Considering that post-procedure assessments might not be different than pre-procedure assessment, only the pre- and per-procedure assessments were included, that is at least 85 paired assessments before and 85 paired assessments during the procedure, which is equal to 170 paired assessments. Because each patient could be assessed during two to three procedures by two to three observers, the number of patients necessary to enrol was n&#x02009;=&#x02009;30 to reach these 170 paired assessments.</p></sec><sec id="Sec12"><title>Presentation of data</title><p>Quantitative data are shown as medians and 25^th to 75^th percentiles. A <italic>P</italic> value of &#x02264;0.05 was considered statistically significant. Data were analysed using the SAS software version 9.1 (SAS Institute, Cary, NC, USA).</p></sec></sec></sec><sec id="Sec13" sec-type="results"><title>Results</title><p>During the study period, 258 paired observations of pain behaviour were done with each pain tool in 30 patients by 24 observers (20 registered nurses (RNs), 4 investigators) during 75 procedures: repositioning, n&#x02009;=&#x02009;30; turning onto both sides for bathing, massage and changing the sheets, n&#x02009;=&#x02009;30; suctioning, n&#x02009;=&#x02009;14; mobilisation for physical therapy, n&#x02009;=&#x02009;1. A consort flow chart of patient enrolment is shown in Figure&#x000a0; <xref rid="Fig1" ref-type="fig">1</xref>. Table&#x000a0; <xref rid="Tab2" ref-type="table">2</xref> summarises patients&#x02019; demographic and medical characteristics.<fig id="Fig1"><label>Figure 1</label><caption><p> <bold>Study flow chart.</bold> </p></caption><graphic xlink:href="13054_2014_2900_Fig1_HTML" id="d30e832"/></fig></p><table-wrap id="Tab2"><label>Table 2</label><caption><p> <bold>Demographic and medical characteristics of the 30 patients included for analysis</bold> </p></caption><table frame="hsides" rules="groups"><tbody><tr><td>Age (years)</td><td align="center">67 [57-74]</td></tr><tr><td>Sex (F/M)</td><td align="center">19/11</td></tr><tr><td>Body mass index (kg/m^-2)</td><td align="center">26 [22-30]</td></tr><tr><td>Chronic pain syndrome, n (%)</td><td align="center">11 (36%)</td></tr><tr><td>Reason for admission to the ICU</td><td/></tr><tr><td> <italic>Acute respiratory failure, n (%)</italic> </td><td align="center">17 (57%)</td></tr><tr><td> <italic>Severe sepsis/septic shock, n (%)</italic> </td><td align="center">8 (27%)</td></tr><tr><td> <italic>Miscellaneous*, n (%)</italic> </td><td align="center">6 (20%)</td></tr><tr><td>Time between admission to ICU and enrolment (days)</td><td align="center">4 [2-7]</td></tr><tr><td>APACHE II score within 24&#x000a0;h after admission to ICU</td><td align="center">23 [20-29]</td></tr><tr><td>SOFA score upon enrolment</td><td align="center">8 [7-11]</td></tr><tr><td>Mechanical ventilation upon enrolment, n (%)</td><td align="center">19 (63%)</td></tr><tr><td>Sedation upon enrolment</td><td align="center">13 (43%)</td></tr><tr><td> <italic>Propofol, n (%)</italic> </td><td align="center"> <italic>12 (40%)</italic> </td></tr><tr><td> <italic>Dose (&#x003bc;g.kg</italic> ^{<italic>-1</italic>} <italic>.min</italic> ^{<italic>-1</italic>} <italic>)</italic> </td><td align="center"> <italic>10</italic> [5-11]</td></tr><tr><td> <italic>Dexmedetomidine, n (%)</italic> </td><td align="center"> <italic>1 (3%)</italic> </td></tr><tr><td>Analgesia upon enrolment</td><td align="center">16 (53%)</td></tr><tr><td> <italic>Fentanyl, n (%)</italic> </td><td align="center"> <italic>15 (50%)</italic> </td></tr><tr><td> <italic>Dose (&#x003bc;g.kg</italic> ^{<italic>-1</italic>} <italic>.h</italic> ^{<italic>-1</italic>} <italic>)</italic> </td><td align="center"> <italic>0.9 [0.6-1.2]</italic> </td></tr><tr><td> <italic>Hydromorphone, n (%)</italic> </td><td align="center"> <italic>1 (3%)</italic> </td></tr><tr><td>RASS level</td><td align="center">-1 [-3; +1]</td></tr><tr><td> <italic>RASS level&#x02009;=&#x02009;0, n (%)</italic> </td><td align="center"> <italic>4 (13%)</italic> </td></tr><tr><td> <italic>RASS level &#x0003e;0, n (%)</italic> </td><td align="center"> <italic>6 (20%)</italic> </td></tr><tr><td> <italic>RASS level &#x0003c;0, n (%)</italic> </td><td align="center"> <italic>20 (67%)</italic> </td></tr><tr><td>CAM-ICU positive in non-sedated patients, n/N (%)</td><td align="center">17/17 (100%)</td></tr></tbody></table><table-wrap-foot><p>Continuous data are expressed in median [25^th to 75^th percentiles]. *Miscellaneous reasons for admission to the ICU: metabolic, acute hepatitis, altered mental status, mechanical ventilation weaning, agitation post procedure.</p><p>ICU, Intensive Care Unit; APACHE II score, Acute Physiology And Chronic Health Evaluation II score; SOFA, Sequential Organ Failure Assessment; RASS, Richmond Agitation Sedation Scale; CAM-ICU, Confusion Assessment Method for the Intensive Care Unit.</p></table-wrap-foot></table-wrap><sec id="Sec14"><title>Inter-rater reliability (primary endpoint)</title><p>Inter-rater reliability was evaluated by weighted kappa coefficients, which are summarised in Table&#x000a0; <xref rid="Tab3" ref-type="table">3</xref>. The reliability was nearly perfect for BPS and CPOT and important for NVPS. Weighted kappa coefficients were significantly greater for BPS (0.81&#x02009;&#x000b1;&#x02009;0.03) and CPOT (0.81&#x02009;&#x000b1;&#x02009;0.03) than for NVPS (0.71&#x02009;&#x000b1;&#x02009;0.04, <italic>P</italic>&#x0003c;0.05 compared to BPS and CPOT). Using only the first assessments for each patient, the weighted kappa coefficients for BPS, CPOT and NVPS were unchanged at 0.88, 0.80 and 0.67, respectively.<table-wrap id="Tab3"><label>Table 3</label><caption><p> <bold>Inter-observer reliability measured by weighted kappa coefficients for each of the three pain tools</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr><th>BPS</th><th>CPOT</th><th>NVPS</th></tr></thead><tbody><tr><td align="center">Total score</td><td align="center">Total score</td><td align="center">Total score</td></tr><tr><td align="center">0.81 (0.03)^a </td><td align="center">0.81 (0.03)^a </td><td align="center">0.71 (0.04)</td></tr><tr><td align="center" colspan="3">Facial domains</td></tr><tr><td align="center">Face</td><td align="center">Face</td><td align="center">Face</td></tr><tr><td align="center">0.75 (0.03)</td><td align="center">0.81 (0.03)^a,c </td><td align="center">0.70 (0.04)^d </td></tr><tr><td align="center" colspan="3">Breathing domains</td></tr><tr><td align="center">Ventilation/vocalisation</td><td align="center">Ventilation/vocalisation</td><td align="center">Respiration</td></tr><tr><td align="center">0.78 (0.04)^a </td><td align="center">0.71 (0.05)^a,c </td><td align="center">0.54 (0.07)^e </td></tr><tr><td align="center" colspan="3">Muscular domains</td></tr><tr><td align="center">Upper limbs</td><td align="center">Body movements</td><td align="center">Activity</td></tr><tr><td align="center">0.61 (0.06)</td><td align="center">0.42 (0.07)^b </td><td align="center">0.52 (0.06)</td></tr><tr><td align="center" rowspan="2"/><td align="center">Muscle tension</td><td align="center">Guarding</td></tr><tr><td align="center">0.43 (0.07)^b </td><td align="center">0.32 (0.07)^b </td></tr><tr><td align="center" colspan="3">Physiological domains</td></tr><tr><td align="center" rowspan="4" colspan="2"/><td align="center">Physiological I</td></tr><tr><td align="center">0.46 (0.08)</td></tr><tr><td align="center">Physiological II</td></tr><tr><td align="center">0.02 (0.03)^f </td></tr></tbody></table><table-wrap-foot><p>All data are expressed in weighted kappa coefficient (standard deviation). ^a <italic>P</italic>&#x0003c;0.05 compared to NVPS; ^b <italic>P</italic>&#x0003c;0.05 compared to BPS; ^c <italic>P</italic>&#x0003c;0.05 compared to CPOT muscular domains; ^d <italic>P</italic>&#x0003c;0.05 compared to NVPS non-facial domains; ^e <italic>P</italic>&#x0003c;0.05 compared to NVPS guarding; ^f <italic>P</italic>&#x0003c;0.05 compared to NVPS non-physiological II domains. BPS, Behavioral Pain Scale; CPOT, Critical-Care Pain Observation Tool; NVPS, Non-verbal Pain Scale.</p></table-wrap-foot></table-wrap></p><p>Table&#x000a0; <xref rid="Tab3" ref-type="table">3</xref> shows inter-rater reliability for each tool&#x02019;s domain. For the facial domain, the greater reliability was demonstrated for CPOT, which was significantly greater than NVPS. For the muscular domains, the greater reliability was demonstrated for BPS, which was significantly greater than the two muscular domains of the CPOT and one of the NVPS muscular domains (Table&#x000a0; <xref rid="Tab3" ref-type="table">3</xref>). The three domains of the BPS demonstrated similar reliability. For the CPOT, both facial and breathing domains demonstrated a significantly greater reliability than muscular domains. For the NVPS, the facial domain demonstrated a significantly greater reliability than other domains. Apart from the facial domain, the breathing domain of the NVPS demonstrated the greater reliability and the physiological domain II the lowest. A subgroup analysis was performed on patients according to their intubation status. In intubated and non-intubated patients, BPS and CPOT had the highest inter-rater reliability but the difference was only significant between BPS and NVPS in non-intubated patients (0.89&#x02009;&#x000b1;&#x02009;0.04 vs. 0.74&#x02009;&#x000b1;&#x02009;0.05, <italic>P</italic>&#x0003c;0.05). Inter-rater reliability was not significantly different in intubated compared to non-intubated patients for NVPS (0.71&#x02009;&#x000b1;&#x02009;0.04 vs. 0.74&#x02009;&#x000b1;&#x02009;0.05) and CPOT (0.80&#x02009;&#x000b1;&#x02009;0.03 vs. 0.82&#x02009;&#x000b1;&#x02009;0.05). BPS had a significantly greater inter-rater reliability in non-intubated than intubated patients (0.89&#x02009;&#x000b1;&#x02009;0.04 vs. 0.77&#x02009;&#x000b1;&#x02009;0.04, <italic>P</italic>&#x0003c;0.05). Finally, within an error of one point, inter-rater agreement was significantly (<italic>P</italic>&#x0003c;0.01) greater for BPS (81%) and CPOT (77%) than for NVPS (65%) for all the observations (before and during the procedures), as well as for observations made during the procedures only (BPS, 73%; CPOT, 77%; NVPS, 57%; <italic>P</italic>&#x0003c;0.05 between NVPS and the two other scales).</p></sec><sec id="Sec15"><title>Internal consistency</title><p>Measurement of Cronbach-&#x003b1; coefficients showed a satisfactory internal consistency for each of the three scales: 0.80 for BPS, 0.81 for CPOT and 0.76 for NVPS. Cronbach-&#x003b1; was significantly greater for BPS (<italic>P</italic>&#x0003c;0.01) and CPOT (<italic>P</italic>&#x0003c;0.001) compared to NVPS. The difference between BPS and CPOT was not significantly different (<italic>P</italic>&#x02009;=&#x02009;0.48).</p><p>There was no significant difference in Cronbach-&#x003b1; coefficients between intubated and non-intubated patients for BPS (0.81 for intubated patients and 0.83 for non-intubated patients, <italic>P</italic>&#x02009;=&#x02009;0.15) and CPOT (0.82 for intubated patients and 0.81 for non-intubated patients, <italic>P</italic>&#x02009;=&#x02009;0.99) contrary to NVPS (0.79 for intubated patients and 0.46 for non-intubated patients, <italic>P</italic> &#x0003c;0.001).</p></sec><sec id="Sec16"><title>Discriminant validation</title><p>Figure&#x000a0; <xref rid="Fig2" ref-type="fig">2</xref> shows the median scores of the three tools evaluated by all the observers according to different situations. There was a significant increase in each of the three scores from baseline to procedure (<italic>P</italic>&#x0003c;0.001) and a significant decrease 10&#x000a0;minutes after the procedure (<italic>P</italic>&#x0003c;0.001). The median scores were not significantly different between observations made at baseline and observations made after the procedure (BPS, <italic>P</italic>&#x02009;=&#x02009;0.41: CPOT, <italic>P</italic>&#x02009;=&#x02009;0.74; NVPS, <italic>P</italic>&#x02009;=&#x02009;0.89). Discriminant validation was also tested comparing median scores observed during two similar situations differing by the intensity and the length of the procedures, that is repositioning and turning onto both sides. There was also a significant difference between these two procedures for each of the three tools (<italic>P</italic>&#x0003c;0.001). Finally, turning and suctioning were the most painful procedures (Figure&#x000a0; <xref rid="Fig2" ref-type="fig">2</xref>). Difference of pain scores between these two procedures was not significant (BPS, <italic>P</italic>&#x02009;=&#x02009;0.90: CPOT, <italic>P</italic>&#x02009;=&#x02009;0.68; NVPS, <italic>P</italic>&#x02009;=&#x02009;0.40).<fig id="Fig2"><label>Figure 2</label><caption><p> <bold>Median scores observed by all the observers with each of the three tools, according to different situations.</bold> This figure shows the median scores of the three tools evaluated by all the observers according to different situations: before, during and after repositioning, turning and suctioning. The left figures show that there was a significant increase in each of the three scores from baseline to procedure and a significant decrease 10&#x000a0;minutes after the procedure. The right figures showed the scores measured during the different procedures. Among them, turning and suctioning were significantly the most painful.</p></caption><graphic xlink:href="13054_2014_2900_Fig2_HTML" id="d30e1464"/></fig></p><p>Responsiveness of the scales was tested by the effect size coefficient, which was large (&#x0003e;0.80) for each of the three scales when calculated between baseline and observations done during the procedures: BPS&#x02009;=&#x02009;1.99; CPOT&#x02009;=&#x02009;1.55; NVPS&#x02009;=&#x02009;1.46. BPS and CPOT demonstrated a significantly higher responsiveness than NVPS, as well as BPS compared to CPOT. The effect size coefficients also remained large when calculated between the repositioning and turning procedures (BPS&#x02009;=&#x02009;0.90; CPOT&#x02009;=&#x02009;0.86; NVPS&#x02009;=&#x02009;0.92), without any significant differences between the three scales.</p></sec><sec id="Sec17"><title>Feasibility</title><p>The 20 RNs who participated in the study and the nurse manager (one of the investigators) rated the three tools at a median of 7 to 8 (0&#x02009;=&#x02009;the worst, 10&#x02009;=&#x02009;the best) for accuracy, usefulness and ease of learning. The BPS was rated higher with regard to ease of learning than the CPOT (<italic>P</italic>&#x02009;=&#x02009;0.02), but the BPS was the same as the NVPS (<italic>P</italic>&#x02009;=&#x02009;0.07): BPS, 8 [<xref ref-type="bibr" rid="CR7">7</xref>&#x02013;<xref ref-type="bibr" rid="CR10">10</xref>]; CPOT 8 [<xref ref-type="bibr" rid="CR5">5</xref>&#x02013;<xref ref-type="bibr" rid="CR8">8</xref>], NVPS 8 [<xref ref-type="bibr" rid="CR6">6</xref>&#x02013;<xref ref-type="bibr" rid="CR8">8</xref>]. There was no significant difference (all <italic>P</italic> values &#x0003e;0.49) between the three tools either with regard to accuracy (BPS, 7 [<xref ref-type="bibr" rid="CR7">7</xref>, <xref ref-type="bibr" rid="CR8">8</xref>]; CPOT 8 [<xref ref-type="bibr" rid="CR5">5</xref>&#x02013;<xref ref-type="bibr" rid="CR8">8</xref>], NVPS 7 [<xref ref-type="bibr" rid="CR6">6</xref>&#x02013;<xref ref-type="bibr" rid="CR8">8</xref>]) or usefulness (BPS, 7 [<xref ref-type="bibr" rid="CR5">5</xref>&#x02013;<xref ref-type="bibr" rid="CR8">8</xref>]; CPOT 8 [<xref ref-type="bibr" rid="CR5">5</xref>&#x02013;<xref ref-type="bibr" rid="CR8">8</xref>], NVPS 7 [<xref ref-type="bibr" rid="CR6">6</xref>&#x02013;<xref ref-type="bibr" rid="CR8">8</xref>]). Observers&#x02019; preference for the three tools is shown in Figure&#x000a0; <xref rid="Fig3" ref-type="fig">3</xref>. There was no difference between preference of use either for research or routine practice. The NVPS was chosen as the preferred tool the most often (43%), followed by the BPS (33%) and the CPOT (24%), but the difference was not significant. Among the nine observers who chose the NVPS as the preferred tool, four explained their choice resulting from their being more familiar with the scale. Reasons for preferential choice are given in Table&#x000a0; <xref rid="Tab4" ref-type="table">4</xref>. Most of the arguments were given by some observers as positive (explaining their first choice) but also by other observers as negative (explaining their last choice).<fig id="Fig3"><label>Figure 3</label><caption><p> <bold>Preference about the use of the three tools, rated by the 20 nurses and the nurse manager.</bold> This figure shows that NVPS was the preferred tool, following by the BPS but the difference was not significant compared to the others (<italic>P</italic>&#x02009;=&#x02009;0.68 for research and for practice). BPS, Behavioral Pain Scale; NVPS, Non-verbal Pain Scale.</p></caption><graphic xlink:href="13054_2014_2900_Fig3_HTML" id="d30e1559"/></fig></p><table-wrap id="Tab4"><label>Table 4</label><caption><p> <bold>Reasons of preferred tool choice by the 20 nurses and the nurse manager</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr><th/><th>Reasons given for first choice</th><th>Reasons given for last choice</th></tr></thead><tbody><tr><td rowspan="6"> <bold>BPS</bold> </td><td>n&#x02009;=&#x02009;7</td><td>n&#x02009;=&#x02009;6</td></tr><tr><td>Main reasons:</td><td>Main reasons:</td></tr><tr><td>Simplicity, easiness, n&#x02009;=&#x02009;4</td><td>Simplicity, n&#x02009;=&#x02009;1</td></tr><tr><td>Descriptors clear or precise, n&#x02009;=&#x02009;2</td><td>Descriptors less well described, n&#x02009;=&#x02009;1</td></tr><tr><td rowspan="2">4 descriptors instead of 3, n&#x02009;=&#x02009;1</td><td>Less specific, n&#x02009;=&#x02009;1</td></tr><tr><td>Less information, n&#x02009;=&#x02009;3</td></tr><tr><td rowspan="7"> <bold>CPOT</bold> </td><td>n&#x02009;=&#x02009;5</td><td>n&#x02009;=&#x02009;8</td></tr><tr><td>Main reasons:</td><td>Main reasons:</td></tr><tr><td>Descriptors more detailed, n&#x02009;=&#x02009;2</td><td>Descriptors too complex, n&#x02009;=&#x02009;2</td></tr><tr><td>Descriptors better described, n&#x02009;=&#x02009;2</td><td>Descriptors less well detailed or confusing, n&#x02009;=&#x02009;3</td></tr><tr><td>Vocalisation domain compared to NVPS, n&#x02009;=&#x02009;1</td><td>No reason, n&#x02009;=&#x02009;3</td></tr><tr><td> <italic>Other reason:</italic> </td><td rowspan="2"/></tr><tr><td> <italic>Ventilator alarm notified, n&#x02009;=&#x02009;1</italic> </td></tr><tr><td rowspan="9"> <bold>NVPS</bold> </td><td>n&#x02009;=&#x02009;9</td><td>n&#x02009;=&#x02009;7</td></tr><tr><td>Main reasons:</td><td>Main reasons:</td></tr><tr><td>Familiar with, n&#x02009;=&#x02009;4</td><td>Some descriptor not understandable, n&#x02009;=&#x02009;1</td></tr><tr><td>More information, n&#x02009;=&#x02009;3</td><td>Descriptors less well detailed, n&#x02009;=&#x02009;2</td></tr><tr><td>Vital signs notified, n&#x02009;=&#x02009;2</td><td>Vital signs not valid in ICU patients, n&#x02009;=&#x02009;3</td></tr><tr><td/><td>No reason, n&#x02009;=&#x02009;1</td></tr><tr><td> <italic>Other reasons:</italic> </td><td/></tr><tr><td> <italic>Vital signs notified, n&#x02009;=&#x02009;1</italic> </td><td/></tr><tr><td> <italic>Change over time notified, n&#x02009;=&#x02009;1</italic> </td><td/></tr></tbody></table><table-wrap-foot><p>BPS, Behavioral Pain Scale; CPOT, Critical-Care Pain Observation Tool; NVPS, Non-verbal Pain Scale.</p></table-wrap-foot></table-wrap></sec></sec><sec id="Sec18" sec-type="discussion"><title>Discussion</title><p>The main findings of this study are that BPS, CPOT and NVPS have good psychometric properties but BPS and CPOT have significantly higher inter-rater reliability, internal consistency and responsiveness than NVPS. Discriminant validation was good for all three scales. There was no difference in regards to feasibility except for BPS, which is rated a little easier to remember than the other scales, with only three domains of observation rather than four and six for CPOT and NVPS. Scales&#x02019; preference was variable among users, with no scale demonstrating any consensus. In all, either BPS or CPOT appear to be superior tools and should be chosen in the ICU where no behavioural pain scale has been implemented yet, consistent with the recent Practice Guidelines [<xref ref-type="bibr" rid="CR16">16</xref>].</p><p>These data are consistent with a recent study aimed at comparing CPOT and NVPS in mostly intubated patients, which found a better inter-rater reliability for CPOT [<xref ref-type="bibr" rid="CR39">39</xref>]. Moreover, our study showed that BPS and CPOT can be used in both intubated and non-intubated patients whereas NVPS demonstrated a poor internal consistency in non-intubated patients. NVPS was neither constructed nor validated in non-intubated patients [<xref ref-type="bibr" rid="CR19">19</xref>, <xref ref-type="bibr" rid="CR20">20</xref>] in contradistinction to the BPS and CPOT that are both constructed to be used either in intubated or non-intubated patients [<xref ref-type="bibr" rid="CR17">17</xref>, <xref ref-type="bibr" rid="CR18">18</xref>, <xref ref-type="bibr" rid="CR27">27</xref>]. It could not have been possible to compare BPS and CPOT in an ICU team trained to use one of those tools. In our institution, nurses are trained to use the NVPS, which consequently allows for an accurate comparison between BPS and CPOT in a team familiar with using a behavioural pain tool. Moreover, nurses in our institution routinely use the NVPS to also assess pain in non-intubated patients unable to self-report. NVPS&#x02019; internal consistency was indeed low in non-intubated patients. However, inter-rater reliability was not significantly different for NVPS depending on whether the patients were intubated or not. The reliability of the BPS was significantly greater in non-intubated patients. BPS requires assessing ventilator waveforms and asynchrony, which could be difficult while observing patients&#x02019; face and body at the same time. Listening to ventilator alarms like for the CPOT could be a useful alternative. Recent American Practice Guidelines recommended further assessment in non-intubated patients with a modified BPS (that is BPS-NI) or the CPOT. These new data should strengthen the rationale for BPS and CPOT use in ICU non-intubated non-communicative patients.</p><p>Pain is one of the most stressful events experienced by patients during their ICU stay [<xref ref-type="bibr" rid="CR40">40</xref>, <xref ref-type="bibr" rid="CR41">41</xref>]. At rest, surgical and trauma patients report surgery/trauma site as the most painful area although medical patients most likely report pain localised in back and limbs [<xref ref-type="bibr" rid="CR2">2</xref>]. Being moved for nursing-care procedures is one of the most painful procedures experienced by the patient during the ICU stay whatever the type of admission (medical, surgical or trauma) [<xref ref-type="bibr" rid="CR3">3</xref>, <xref ref-type="bibr" rid="CR24">24</xref>, <xref ref-type="bibr" rid="CR25">25</xref>, <xref ref-type="bibr" rid="CR42">42</xref>, <xref ref-type="bibr" rid="CR43">43</xref>]. Contrary to pain while moving the patient for nursing procedures, pain during active mobilisation for early rehabilitation had never been investigated in the ICU-setting [<xref ref-type="bibr" rid="CR44">44</xref>] until the recent EUROPAIN&#x02122; study [<xref ref-type="bibr" rid="CR25">25</xref>]. In this large multicentre study assessing 13 different procedures of care in ICU patients, active mobilization was the less painful procedure (NRS&#x02009;=&#x02009;2 [0;5]) while positioning and turning were associated with a higher pain intensity (3 [0;5] and 3 [0.25;6], respectively) [<xref ref-type="bibr" rid="CR25">25</xref>]. One of the differences between active and passive mobilization (that is rehabilitation vs. repositioning and turning) is that movements and pressure on body parts can be controlled by the patients or not. This could explain the difference in pain intensity between these procedures. However, whether pain could be a barrier toward early rehabilitation in specific ICU patients, such as surgical patients, remains unknown [<xref ref-type="bibr" rid="CR45">45</xref>, <xref ref-type="bibr" rid="CR46">46</xref>]. In the present study, we were able to enrol only one patient while being mobilised by a physiotherapist/occupational therapist. This was because mobilisation requires the patient to participate and be able to follow instructions and our inclusion criteria specifically enrolled patients unable to self-report their pain intensity, a less common feature in patients able to participate in early mobility. The one patient enrolled for mobilisation in our trial was effectively with delirium and was not able to use the 0 to 10 NRS. However, early mobilisation could prevent delirium in the ICU and is therefore recommended in patients able to participate. Along with delirium, pain is one other neuropsychological event for which an accurate management is highly recommended in ICU patients. Improved pain management based on an accurate assessment of patient&#x02019;s pain intensity is associated with better patient outcomes in the ICU [<xref ref-type="bibr" rid="CR1">1</xref>, <xref ref-type="bibr" rid="CR8">8</xref>&#x02013;<xref ref-type="bibr" rid="CR10">10</xref>]. Sequential studies using the BPS performed in surgical and medical ICUs reported that a multidisciplinary (nurse and physician) protocol to diagnose and manage pain, agitation and delirium was associated with a reduced duration of mechanical ventilation [<xref ref-type="bibr" rid="CR1">1</xref>, <xref ref-type="bibr" rid="CR10">10</xref>], ICU-acquired infections [<xref ref-type="bibr" rid="CR1">1</xref>], length of stay in ICU and hospital as well as 30-day mortality [<xref ref-type="bibr" rid="CR10">10</xref>]. A large multicentre observational study in 1,144 mechanically ventilated patients, in whom BPS was the most frequently used tool, showed that pain assessment was associated with reduced duration of mechanical ventilation and length of stay in ICU [<xref ref-type="bibr" rid="CR9">9</xref>]. That could be explained in part by a reduced use of sedatives and a greater use of analgesics [<xref ref-type="bibr" rid="CR9">9</xref>]. Implementation of the CPOT was also associated with a reduction of sedatives and change in analgesics ordering [<xref ref-type="bibr" rid="CR28">28</xref>, <xref ref-type="bibr" rid="CR47">47</xref>], suggesting that standardising pain assessment in critically ill patients may allow for a better match between analgesics requirements and administration. Recently, a multidisciplinary quality-improvement study based on pain assessment using the 0 to 10&#x000a0;V-NRS and BPS/BPS-NI along with an analgesia protocol showed that decreased incidence in severe pain while turning ICU patients was associated with decreased adverse outcomes [<xref ref-type="bibr" rid="CR4">4</xref>]. Therefore, pain management is highly challenging in the ICU setting and determining the most valid and reliable tool is paramount before any implementation of an analgesia protocol to a multidisciplinary team [<xref ref-type="bibr" rid="CR16">16</xref>]. The team&#x02019;s preference regarding the choice of a pain tool should also be taken into account but a consensus might be difficult to reach. Indeed, no tool reached a consensus among users in our study. One-third of users who chose NVPS as the preferred tool mentioned observation of vital signs as the reason. Inversely, almost half of the users who ranged NVPS as the less preferred tool mentioned that observation of vital signs was not accurate in critically ill patients. Indeed, the physiological domains of NVPS demonstrated poor to just moderate inter-rater reliability despite objective measurement and recording of vital signs. Because pain can be associated either with an increase or decrease in physiological variables [<xref ref-type="bibr" rid="CR48">48</xref>], which can moreover be influenced by many factors such as disease or treatment, variation of vital signs should be studied further in critically ill patients in order to standardise them as a possible domain in observational pain tools. Another example highlighting difficulties in reaching a consensus among users is the subjective assessment of tool&#x02019;s complexity. One-quarter of users found the BPS too simple or with less information whereas another quarter found the CPOT too complex or with descriptors less well detailed or confusing. However, complexity of a subjective tool may impact on inter-rater reliability. Thus, the higher reliability shown for the muscular domain of BPS compared to CPOT and NVPS might be potentially explained by the fact that both CPOT and NVPS have two muscular domains while BPS has only one.</p><p>Finally, if using tools demonstrating the best psychometric properties such as BPS or CPOT might be recommended, it is unknown whether a small but significant difference in psychometric measurement is clinically relevant or not in regard to patients&#x02019; outcome. Also, clinical studies are still needed to determine which threshold is the most effective in regard to ICU outcome (duration of mechanical ventilation, stress response-related events) but also in regard to outcome after ICU discharge (chronic pain syndrome, post-traumatic stress disorder (PTSD)). Then, further studies are needed to determine how it would be the most effective to educate, train and assess healthcare givers when using subjective behavioural pain tools to increase their reliability in research and routine use. Results of this study showed that repeated education and training is paramount to assure important inter-rater reliability of a tool as previously showed with the use of sedation and delirium tools in the ICU setting [<xref ref-type="bibr" rid="CR49">49</xref>]. A different education strategy and/or tool training prior to the present study might have resulted in different findings. Whether some investigators who could have been more experienced about NVPS or BPS/BPS-NI use might have impacted on the results should be considered as a possible bias and a limit of the study. In order to minimize educational issues, descriptors and instructions for use were clearly indicated on the data collection sheet for the three tools (see Additional file <xref rid="MOESM1" ref-type="media">1</xref>). Also, this could explain that all three tools demonstrated good psychometric properties.</p></sec><sec id="Sec19" sec-type="conclusions"><title>Conclusions</title><p>BPS, CPOT and NVPS demonstrate good inter-rater reliability in both intubated and non-intubated ICU patients unable to self-report their pain intensity. BPS and CPOT have significantly higher inter-rater reliability, internal consistency and responsiveness than NVPS, which psychometric properties remain, however, acceptable in general but not for the physiological domains. Discriminative validation is important for all three scales. There is no difference in regard to feasibility except for BPS, which is rated a little easier to remember. However, no scale demonstrated any consensus among users. Either BPS or CPOT should be used in intubated and non-intubated patients unable to self-report, particularly when no behavioural pain scale is already available in an ICU setting.</p></sec><sec id="Sec20"><title>Key messages</title><p><list list-type="bullet"><list-item><p>BPS and CPOT have significantly higher inter-rater reliability and internal consistency than NVPS in intubated and non-intubated ICU patients unable to self-report their pain intensity.</p></list-item><list-item><p>BPS demonstrates significantly highest responsiveness.</p></list-item><list-item><p>Psychometric properties are acceptable for NVPS in general but not for the physiological domains.</p></list-item><list-item><p>No scale demonstrates a better feasibility among users.</p></list-item><list-item><p>Because of significantly better psychometric properties, either BPS or CPOT should be used in intubated and non-intubated ICU patients unable to self-report.</p></list-item></list></p></sec><sec sec-type="supplementary-material"><title>Electronic supplementary material</title><sec id="Sec21"><supplementary-material content-type="local-data" id="MOESM1"><media xlink:href="13054_2014_2900_MOESM1_ESM.pdf"><caption><p>Additional file 1: <bold>Data sheet for observers&#x02019; pain assessments.</bold> This additional file provides the sheet used by the observers during the study to independently assess pain with each of the three tools: BPS, CPOT and NVPS. Note that descriptors and instruction of use were written for each tool to avoid any learning issues. (PDF 177 KB)</p></caption></media></supplementary-material></sec></sec>

Experimental and clinical evidences for glucose control in intensive care: is infused glucose the key point for study interpretation?

Could not extract abstract

<contrib contrib-type="author" corresp="yes"><name><surname>Mazeraud</surname><given-names>Aur&#x000e9;lien</given-names></name><address><email>aurelien.mazeraud@pasteur.fr</email></address><xref ref-type="aff" rid="Aff31"/><xref ref-type="aff" rid="Aff32"/></contrib><contrib contrib-type="author" corresp="yes"><name><surname>Polito</surname><given-names>Andrea</given-names></name><address><email>andrea.polito@rpc.aphp.fr</email></address><xref ref-type="aff" rid="Aff31"/><xref ref-type="aff" rid="Aff33"/><xref ref-type="aff" rid="Aff34"/></contrib><contrib contrib-type="author" corresp="yes"><name><surname>Annane</surname><given-names>Djillali</given-names></name><address><email>djillali.annane@rpc.aphp.fr</email></address><xref ref-type="aff" rid="Aff33"/><xref ref-type="aff" rid="Aff34"/></contrib><aff id="Aff31"><label/>Institut Pasteur, Human Histopathology and Animal Models, Pr Chretien F, 28, rue du docteur Roux, 75015 Paris, France </aff><aff id="Aff32"><label/>Universit&#x000e9; Paris Descartes Sorbonne Paris Cit&#x000e9;, Institut Pasteur, rue du Dr Roux, 75015 Paris, France </aff><aff id="Aff33"><label/>Medical and Surgical Intensive Care Unit, CHU Raymond Poincar&#x000e9; (AP-HP), University of Versailles Saint Quentin en Yvelines, 104 boulevard Raymond Poincar&#x000e9;, 92380 Garches, France </aff><aff id="Aff34"><label/>Universit&#x000e9; de Versailles Saint-Quentin en Yvellines, 55 Avenue de Paris, 78000 Versailles, France </aff>

Critical Care

<sec id="Sec1" sec-type="intro"><title>Introduction</title><p>In an ICU, stress induces insulin resistance and overproduction of glucose, resulting in a syndrome called stress-induced hyperglycemia (SIH) [<xref ref-type="bibr" rid="CR1">1</xref>]. SIH is common during critical illness and is associated with high mortality [<xref ref-type="bibr" rid="CR1">1</xref>&#x02013;<xref ref-type="bibr" rid="CR3">3</xref>]. Its incidence is approximately 50% in septic shock [<xref ref-type="bibr" rid="CR1">1</xref>] and 13% in surgical patients [<xref ref-type="bibr" rid="CR4">4</xref>]. Up to the beginning of the 21st century, hyperglycemia was considered an adaptive mechanism to stress. In 2001, the landmark Leuven study by Van den Berghe and colleagues [<xref ref-type="bibr" rid="CR4">4</xref>] reported a 34% relative risk reduction of in-hospital mortality when blood glucose was maintained at between 80 and 110&#x000a0;mg/dL. Since then, glucose metabolism during critical illness has been the focus of an increasing number of experimental and clinical studies. A decade later, after seven additional major randomized control trials, physicians remain confused about how to manage SIH. The heterogeneity of studies includes differences in population, ICU setting, staff experience, feeding strategy, blood glucose monitoring, variability, definition of hypoglycemia, insulin protocol, infusion site, its continuation after ICU discharge, and finally differences in the choice of the relevant major outcome. In the two &#x02018;positive&#x02019; trials from Leuven, mean non-protein daily caloric intake was approximately 20&#x000a0;kcal/kg per day, essentially via glucose administration initially given intravenously: up to 200 to 300&#x000a0;g/day in the 2001 trial, with a median total daily insulin administration of 71 units (confidence interval of 48 to 100). By contrast, in NICE-SUGAR (Normoglycemia in Intensive Care Evaluation and Surviving Using Glucose Algorithm Regulation), which suggested increased mortality with intensive insulin therapy, caloric intake was 11.04&#x02009;&#x000b1;&#x02009;6.08&#x000a0;kcal/kg per day, with 19.5% given intravenously, and cumulative mean daily dose of insulin was 50.2&#x02009;&#x000b1;&#x02009;38.1 units per day [<xref ref-type="bibr" rid="CR5">5</xref>]. Thus, there were two markedly different therapeutic approaches - that is, intensive gluco- and insulin therapy (liberal glucose intake, or the Leuven approach) and intensive insulin therapy (IIT) (restrictive glucose intake, or the NICE-SUGAR approach) [<xref ref-type="bibr" rid="CR6">6</xref>]. The aim of this review is to discuss experimental evidence of organ injury and insulin sensitivity during SIH and expose differences in strategies for its control that include a liberal or a rather restrictive glucose intake.</p></sec><sec id="Sec2"><title>The risk associated with hyperglycemia</title><p>After several large clinical trials, there is still no consensus on what blood glucose level (BGL) is &#x02018;too much&#x02019;. Whereas the Leuven trial demonstrated deleterious effects from uncontrolled glucose levels, subsequent trials comparing strategies to control BGL reported conflicting results [<xref ref-type="bibr" rid="CR4">4</xref>, <xref ref-type="bibr" rid="CR5">5</xref>, <xref ref-type="bibr" rid="CR7">7</xref>&#x02013;<xref ref-type="bibr" rid="CR11">11</xref>].</p><p>SIH is undoubtedly associated with mortality in stroke [<xref ref-type="bibr" rid="CR12">12</xref>], brain injury [<xref ref-type="bibr" rid="CR2">2</xref>], and myocardial infarction [<xref ref-type="bibr" rid="CR13">13</xref>] patients; trauma, cardiothoracic surgery, thermally injured, and mixed ICU patients [<xref ref-type="bibr" rid="CR3">3</xref>]; and non-ICU hospitalized patients. The aim of this section is to summarize current knowledge about the mechanisms of hyperglycemia toxicity.</p><sec id="Sec3"><title>Pathogenesis of stress-induced hyperglycemia</title><p>Critical illness is characterized by an imbalance between insulin and endogenous or exogenous counter-regulatory hormones (glucagon and glucocorticoids). As a result, glucose production is increased and its storage is decreased secondary to downregulated glycogen synthesis and enhanced glycogenolysis [<xref ref-type="bibr" rid="CR1">1</xref>]. In animal models, adrenaline infusion induces hyperglycemia via stimulation of hepatic and renal neoglucogenesis [<xref ref-type="bibr" rid="CR14">14</xref>]. This pathway represents the major source of endogenous glucose during critical illness [<xref ref-type="bibr" rid="CR1">1</xref>].</p><p>SIH is also the consequence of insulin resistance [<xref ref-type="bibr" rid="CR1">1</xref>]. Experimental data on the mechanisms of sepsis-induced acute insulin resistance are scarce. Most of the knowledge about the mechanisms of acute insulin resistance comes from trauma/hemorrhage experimental models and studies in type 2 diabetes [<xref ref-type="bibr" rid="CR15">15</xref>]. Under healthy conditions, insulin binding to its receptor results in the phosphorylation of insulin receptor substrates, which transmit insulin metabolic and growth signals. One major effector of the metabolic pathway is glucose transporter family (GLUT) 4, which facilitates glucose transport across cell membranes in muscle and adipose tissue. During injury, inhibitor of kappa B kinase and Jun B pathways are activated, leading to expression of inflammatory markers such as TNF. First, Jun B negatively regulates insulin receptor substrate and then TNF downregulates GLUT 4 gene transcription [<xref ref-type="bibr" rid="CR16">16</xref>]. These mechanisms could account for insulin resistance during sepsis [<xref ref-type="bibr" rid="CR17">17</xref>].</p><p>Whether this mechanism of response to aggression is deleterious during critical illness is still a matter of debate [<xref ref-type="bibr" rid="CR18">18</xref>]. In fact, insulin resistance is variable from one tissue to another and appears to be moderate in the heart and diaphragm but is major in skeletal muscles and adipocytes [<xref ref-type="bibr" rid="CR19">19</xref>]. In fact, despite this insulin resistance, glucose utilization is enhanced in sepsis secondary to different GLUT overexpression [<xref ref-type="bibr" rid="CR20">20</xref>].</p><p>Glucose transport across cell membranes is the rate-limiting step of cellular glucose metabolism. Each GLUT is characterized mostly by organ specificity, insulin sensitivity, and the Michaelis constant (K_m). K_m is defined as the transporter&#x02019;s specific value of glycemia for which 50% of transport capacities are reached. These characteristics (Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>) could explain differences in organ sensitivity to insulin and modulation of glucose uptake with glycemia. This heterogeneity may be seen as a protective effect of vital organs while placing other organs in a &#x02018;hibernating state&#x02019;.<table-wrap id="Tab1"><label>Table 1</label><caption><p> <bold>Summary of different glucose transporter characteristics</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr><th>Transporter</th><th>Insulin sensitivity</th><th align="left">Location</th><th>Michaelis constant (K_m), mg/dL</th><th align="left">Particularity</th></tr></thead><tbody><tr><td>GLUT 4</td><td align="center">Yes</td><td>Muscle, adipocytes</td><td align="center">460</td><td>Accounts for insulin resistance. Glycemia-dependent transporter</td></tr><tr><td>GLUT 1</td><td align="center">No</td><td>Blood&#x02013;brain barrier, astrocytes, cardiomyocytes, liver, endothelium</td><td align="center">25</td><td>Rate-limiting step of glucose transport in brain. Upregulated up to 1.7-fold in sepsis</td></tr><tr><td>GLUT 2</td><td align="center">No</td><td>Liver, kidney, beta pancreatic cells</td><td align="center">300</td><td>Glycemia-dependent transport</td></tr><tr><td>GLUT 3</td><td align="center">No</td><td>Brain</td><td align="center">25</td><td>Second important transporter in brain</td></tr></tbody></table><table-wrap-foot><p>GLUT, glucose transporter.</p></table-wrap-foot></table-wrap></p><p>Hyperglycemia is responsible for reactive oxygen species (ROS) overproduction in diabetes via four major pathways: advanced glycated endproduct release, <italic>de novo</italic> indirect activation of C kinase protein, increased polyol, and hexosamine pathway flux redox homeostasis [<xref ref-type="bibr" rid="CR21">21</xref>]. On one hand, some ROS production can be essential for immune cell &#x02018;respiratory burst&#x02019; to kill pathogens or for endothelial cell functions. On the other hand, during hyperglycemia, excess ROS may worsen organ failure [<xref ref-type="bibr" rid="CR22">22</xref>].</p></sec><sec id="Sec4"><title>Hyperglycemia and the central nervous system</title><p>Brain cells are dependent on glucose to maintain their membrane ionic gradient. However, the detrimental cerebral effects of hyperglycemia have been observed in critical illness [<xref ref-type="bibr" rid="CR2">2</xref>, <xref ref-type="bibr" rid="CR12">12</xref>, <xref ref-type="bibr" rid="CR23">23</xref>].</p><p>Brain glucose metabolism has some particularities:</p><p><list list-type="bullet"><list-item><p>Glucose crosses blood&#x02013;brain barrier and cellular membranes via a high-affinity and insulin-insensitive process involving GLUT 1 and GLUT 3 [<xref ref-type="bibr" rid="CR24">24</xref>].</p></list-item><list-item><p>Neurons and astrocytes cooperate to metabolize carbohydrates, as suggested by lactate shuttles between these cells.</p></list-item><list-item><p>During hypoxemic or hypoperfusion stress, GLUT 1 and 3 are upregulated (up to 300% in trauma) with a subsequent increase in glucose uptake [<xref ref-type="bibr" rid="CR25">25</xref>].</p></list-item></list></p><p>Hyperglycemia has been shown to enhance the breakdown of the blood&#x02013;brain barrier via induction of matrix metalloproteinase [<xref ref-type="bibr" rid="CR26">26</xref>] and to induce apoptosis [<xref ref-type="bibr" rid="CR23">23</xref>], mostly via enhanced superoxide production. Indeed, in epidemiologic studies on stroke, hyperglycemia is associated with edema, infarct size, mortality in non-diabetic patients, and poor functional status at 1&#x000a0;year [<xref ref-type="bibr" rid="CR12">12</xref>]. Nevertheless, trials aiming at controlling BGL in stroke, subarachnoid hemorrhage [<xref ref-type="bibr" rid="CR27">27</xref>], brain injury [<xref ref-type="bibr" rid="CR28">28</xref>], and neuro-intensive care [<xref ref-type="bibr" rid="CR29">29</xref>] patients did not report improved outcome with tight glucose control.</p></sec><sec id="Sec5"><title>Hyperglycemia and the peripheral nervous system</title><p>Neuromyopathy is a frequent complication of critical illness, such as septic shock and acute respiratory distress syndrome, with an incidence of up to 50% of patients with these conditions. In the first Leuven study, the risk of developing a critical illness neuromyopathy (CINM) was lowered from 49% to 25% (<italic>P</italic>&#x02009;&#x0003c;&#x02009;0.0001) in the interventional group, facilitating weaning from mechanical ventilation. The mechanisms of hyperglycemic neuromyopathy are poorly understood and may involve activation of apoptotic and inflammation pathways in response to acute hyperglycemia in muscles [<xref ref-type="bibr" rid="CR30">30</xref>] or ROS overproduction as suggested in type 2 diabetic neuropathy or CINM pathogenesis [<xref ref-type="bibr" rid="CR21">21</xref>].</p></sec><sec id="Sec6"><title>Hyperglycemia and other organs</title><sec id="Sec7"><title>Liver</title><p>In resting conditions, GLUT 2 is the predominant transporter for glucose in hepatic parenchymal cells [<xref ref-type="bibr" rid="CR24">24</xref>, <xref ref-type="bibr" rid="CR31">31</xref>]. This low-affinity transporter modulates glucose transport proportionally to BGL. After lipopolysaccharide (LPS) stimulation, GLUT 2 decreases, whereas GLUT 1 increases [<xref ref-type="bibr" rid="CR31">31</xref>], resulting in an enhanced insulin- and glycemia-independent uptake of up to 2.4-fold [<xref ref-type="bibr" rid="CR31">31</xref>, <xref ref-type="bibr" rid="CR32">32</xref>]. Analyses of liver cells from the control group of the Leuven study revealed dramatic lesions to the mitochondria. These mitochondria abnormalities may result from excessive glucose uptake, with subsequent overproduction of ROS [<xref ref-type="bibr" rid="CR33">33</xref>], which could have been diminished with BGL normalization.</p></sec><sec id="Sec8"><title>Immune system</title><p>In the Leuven study, patients with normoglycemia had almost 50% fewer bloodstream infections (7.8% versus 4.2%, <italic>P</italic>&#x02009;=&#x02009;0.003) [<xref ref-type="bibr" rid="CR4">4</xref>]. Indeed, each step of the immune response to stress is altered with hyperglycemia. First, in diabetes, chronic high BGL induces overexpression of surface and circulating cell adhesion molecules [<xref ref-type="bibr" rid="CR34">34</xref>&#x02013;<xref ref-type="bibr" rid="CR36">36</xref>], whereas LPS challenge is less effective in upregulating cell adhesion molecules [<xref ref-type="bibr" rid="CR34">34</xref>]. This enhanced overall immune cell adhesion paradoxically results in a less effective chemotactism and transmigration capacity of immune cells [<xref ref-type="bibr" rid="CR37">37</xref>]. Second, worse polymorphonuclear killing capacities against pathogens, as assessed by concentrations of lysosomal enzyme or burst respiratory intensity, are observed when BGL is poorly controlled with a dose-effect relationship in diabetes [<xref ref-type="bibr" rid="CR38">38</xref>]. Finally, the production of chemokines and other pro-inflammatory factors is decreased under hyperglycemic conditions [<xref ref-type="bibr" rid="CR39">39</xref>].</p></sec><sec id="Sec9"><title>Kidney</title><p>In septic shock, GLUT 2 and 3 expressions are decreased in the tubular epithelial cells of the kidney, whereas GLUT 1 expression is increased. This may account for enhanced glycosuria and acute renal failure during septic shock [<xref ref-type="bibr" rid="CR40">40</xref>]. In the Leuven study, renal replacement therapy was twice less frequent in patients with normoglycemia, whereas insulin <italic>per se</italic> was associated with worse renal outcome [<xref ref-type="bibr" rid="CR41">41</xref>]. This kidney protection may result directly from lower BGL, since high BGL directly inhibits transcription of an anti-apoptotic gene in renal tubules [<xref ref-type="bibr" rid="CR42">42</xref>] or from improvement in lipid profile, ROS production, and endothelial protection [<xref ref-type="bibr" rid="CR43">43</xref>].</p></sec><sec id="Sec10"><title>Heart and endothelium</title><p>SIH has been shown to be an important prognostic factor in acute coronary syndromes [<xref ref-type="bibr" rid="CR13">13</xref>]. The heart has a remarkable ability to switch from free fatty acid oxidation to carbohydrate oxidation under hypoxemic conditions [<xref ref-type="bibr" rid="CR44">44</xref>]. During acute myocardial infarction, SIH activates T cells in the atherosclerotic plaque and increases tissue levels of inflammatory markers and nitric oxide and ROS production, resulting in endothelial dysfunction [<xref ref-type="bibr" rid="CR45">45</xref>]. Consequently, coronary blood flow and reserve during myocardial infarction are impaired [<xref ref-type="bibr" rid="CR46">46</xref>]. Furthermore, acute hyperglycemia increases infarct size and suppresses cardioprotective signal transduction via mitochondrial potassium ATP channel inhibition [<xref ref-type="bibr" rid="CR47">47</xref>].</p><p>In shock, although BGLs are high, glucose represents only 12% of substrate oxidation by cardiomyocytes [<xref ref-type="bibr" rid="CR48">48</xref>]. Therefore, one could argue that hyperglycemia without insulin infusion does not confer a metabolic benefit and presents rather deleterious consequences on an ischemic heart.</p></sec></sec></sec><sec id="Sec11"><title>Evidence for gluco-insulinotherapy</title><sec id="Sec12"><title>Clinical trial calendar</title><p>See Table&#x000a0;<xref rid="Tab2" ref-type="table">2</xref>.<table-wrap id="Tab2"><label>Table 2</label><caption><p> <bold>Trials&#x02019; calendar</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr><th>Year</th><th align="left">Trial</th></tr></thead><tbody><tr><td>2001</td><td>The first Leuven RCT (1,548 patients) reported a 34% relative risk reduction in hospital mortality with maintenance of BGL of between 80 and 110&#x000a0;mg/dL [<xref ref-type="bibr" rid="CR4">4</xref>].</td></tr><tr><td>2003</td><td>Krinsley [<xref ref-type="bibr" rid="CR49">49</xref>], in an observational study (1,826 patients), confirmed the survival benefit associated with protocolized IIT targeting a BGL of less than 140&#x000a0;mg/dL.</td></tr><tr><td>2006</td><td>The second Leuven RCT (1,200 patients) confirmed a 10% absolute reduction in hospital mortality for long-stay medical ICU patients with maintenance of BGL of between 80 and 110&#x000a0;mg/dL [<xref ref-type="bibr" rid="CR11">11</xref>].</td></tr><tr><td>2008</td><td>De la Rosa <italic>et al</italic>. [<xref ref-type="bibr" rid="CR9">9</xref>] RCT (504 patients) failed to show any survival benefit in mixed ICU patients targeting BGL between 80 and 110&#x000a0;mg/dL but with less effective control.</td></tr><tr><td/><td>The VISEP study, with the same glycemic goals (537 patients with septic shock), was terminated prematurely because of an unacceptably high incidence of hypoglycemia (17.0% versus 4.1%; <italic>P</italic>&#x02009;&#x0003c;&#x02009;0.001) and no evidence for survival benefit at 90&#x000a0;days (39.7% versus 35.4%; <italic>P</italic>&#x02009;=&#x02009;0.31) [<xref ref-type="bibr" rid="CR50">50</xref>].</td></tr><tr><td/><td>Arabi <italic>et al</italic>. [<xref ref-type="bibr" rid="CR8">8</xref>] RCT (523 patients) also failed to show survival benefit (adjusted hazard ratio 1.09, 95% confidence interval 0.70 to 1.72) and showed increased hypoglycemic rates (28.6% versus 3.1% of patients; <italic>P</italic>&#x02009;&#x0003c;&#x02009;0.0001).</td></tr><tr><td/><td>SPRINT (BGL goal of 72 to 110&#x000a0;mg/dL) is an observational study with historic control. Nutritional and insulin protocols provided less variable and tighter glucose control (standard deviation of blood glucose was 38% lower compared with the retrospective control) with subsequent improvement in organ failures and outcome for long-stay ICU patients: failure-free days were different (SPRINT&#x02009;=&#x02009;41.6%; Pre-SPRINT&#x02009;=&#x02009;36.5%; <italic>P</italic>&#x02009;&#x0003c;&#x02009;0.0001) [<xref ref-type="bibr" rid="CR51">51</xref>].</td></tr><tr><td>2009</td><td>The Glucontrol (1,101 patients) was stopped prematurely for unintended protocol violations. The IIT was associated with increased hypoglycemia (8.7% versus 2.7%; <italic>P</italic>&#x02009;=&#x02009;0.0001) and a non-significant trend to higher mortality (15.3% versus 17.2%) while BGL was not optimally controlled [<xref ref-type="bibr" rid="CR10">10</xref>].</td></tr><tr><td>2009</td><td>The NICE-SUGAR trial (6,104 mixed ICU patients) compared a strategy of BGL control of between 81 and 108&#x000a0;mg/dL versus a more liberal strategy (&#x0003c;180&#x000a0;mg/dL). This RCT found an increase in mortality with IIT (27.5 versus 24.9; <italic>P</italic>&#x02009;=&#x02009;0.02) and increased incidence of hypoglycemia (6.8% versus 0.5%; <italic>P</italic>&#x02009;&#x0003c;&#x02009;0.001) [<xref ref-type="bibr" rid="CR7">7</xref>].</td></tr><tr><td>2010</td><td>COITTSS (509 patients with septic shock) compared a strategy of BGL control of between 80 and 110&#x000a0;mg/dL versus maintenance of BGL of less than 150&#x000a0;mg/dL. This trial did not find any difference in in-hospital mortality between the two strategies (45.9% versus 42.9%; <italic>P</italic>&#x02009;=&#x02009;0.05) [<xref ref-type="bibr" rid="CR8">8</xref>].</td></tr></tbody></table><table-wrap-foot><p>BGL, blood glucose level; COITTSS, Corticosteroids and Intensive Insulin Therapy for Septic Shock; IIT, intensive insulin therapy; NICE-SUGAR, Normoglycemia in Intensive Care Evaluation and Surviving Using Glucose Algorithm Regulation; RCT, randomized controlled trial; SPRINT, Specialized Relative Insulin and Nutrition Tables; VISEP, Efficacy of Volume Substitution and Insulin Therapy in Severe Sepsis.</p></table-wrap-foot></table-wrap></p></sec><sec id="Sec13"><title>Analysis of critical differences between trials</title><p>In the Leuven studies, patients were given intravenous glucose at 8 to 12&#x000a0;g/hour, with mean intravenous glucose feeding of 120&#x000a0;g during the first 15&#x000a0;hours, to a goal of 200 to 260&#x000a0;g/day afterward. This study shows higher intravenous glucose and insulin administration than in any other study (Table&#x000a0;<xref rid="Tab3" ref-type="table">3</xref>), whereas epidemiologic studies have shown that both are correlated with an increased mortality [<xref ref-type="bibr" rid="CR3">3</xref>, <xref ref-type="bibr" rid="CR41">41</xref>, <xref ref-type="bibr" rid="CR52">52</xref>].<table-wrap id="Tab3"><label>Table 3</label><caption><p> <bold>Summary of characteristics from the different major trials about glucose-insulin treatment</bold> </p></caption><table frame="hsides" rules="groups"><thead><tr><th>Study name</th><th colspan="2">VDB 2001 [<xref ref-type="bibr" rid="CR4">4</xref>]</th><th colspan="2">VDB 2006 [<xref ref-type="bibr" rid="CR11">11</xref>]</th><th colspan="2">Glucontrol [<xref ref-type="bibr" rid="CR10">10</xref>]</th><th colspan="2">NICE-SUGAR [<xref ref-type="bibr" rid="CR5">5</xref>]</th><th colspan="2">COITTSS [<xref ref-type="bibr" rid="CR7">7</xref>]</th><th colspan="2">VISEP [<xref ref-type="bibr" rid="CR50">50</xref>]</th></tr><tr><th/><th>Experimental</th><th>Control</th><th>Experimental</th><th>Control</th><th>Experimental</th><th>Control</th><th>Experimental</th><th>Control</th><th>Experimental</th><th>Control</th><th>Experimental</th><th>Control</th></tr></thead><tbody><tr><td>Morning mean BGL, mg/dL</td><td align="center">103.6</td><td align="center">154.5</td><td align="center">110.0</td><td align="center">160.9</td><td align="center">110.9</td><td align="center">140.0</td><td align="center">118.0</td><td align="center">144.9</td><td align="center">147.3</td><td align="center">154.5</td><td align="center">112.7</td><td align="center">152.7</td></tr><tr><td>SD or confidence interval</td><td align="center">20.0</td><td align="center">32.7</td><td align="center">30.0</td><td align="center">28.0</td><td align="center">100-123.6</td><td align="center">121.8- 160</td><td align="center">25.1</td><td align="center">26.0</td><td align="center">30.9</td><td align="center">34.5</td><td align="center">1.8</td><td align="center">3.6</td></tr><tr><td>Death at 90&#x000a0;days, %</td><td align="center">5</td><td align="center">7</td><td align="center">35.9</td><td align="center">37.7</td><td align="center">23.3</td><td align="center">19.4</td><td align="center">27.5</td><td align="center">24.9</td><td align="center">45.9</td><td align="center">42.9</td><td align="center">39.7</td><td align="center">35.4</td></tr><tr><td>Caloric intake, kcal/day</td><td align="center" colspan="2">550-1,600</td><td align="center">1,202</td><td align="center">1,237</td><td align="center">760</td><td align="center">760</td><td align="center">891</td><td align="center">872</td><td align="center" colspan="2">1,350</td><td align="center">1,217</td><td align="center">1,253</td></tr><tr><td>Quantity of glucose administered per day, g</td><td align="center" colspan="2">120</td><td align="center">202</td><td align="center">198</td><td align="center">73.7</td><td align="center">71.8</td><td align="center">23.4</td><td align="center">24.4</td><td align="center" colspan="2">25</td><td align="center">144</td><td align="center">144</td></tr><tr><td>Daily insulin dose, insulin units</td><td align="center">71</td><td align="center">33</td><td align="center">59</td><td align="center">10</td><td align="center">31.2</td><td align="center">7.68</td><td align="center">50.2</td><td align="center">16.9</td><td align="center">71</td><td align="center">46</td><td align="center">43</td><td align="center">29</td></tr><tr><td>SD or confidence interval</td><td align="center">48-100</td><td align="center">17-56</td><td align="center">37-86</td><td align="center">0-38</td><td align="center">15.6- 55.2</td><td align="center">30.48</td><td align="center">38.1</td><td align="center">29</td><td align="center">45-96</td><td align="center">30-65</td><td align="center">23-64</td><td align="center">15-51</td></tr><tr><td>Hypoglycemia rate, %</td><td align="center">0.8</td><td align="center">5</td><td align="center">18.7</td><td align="center">3.1</td><td align="center">8.7</td><td align="center">2.7</td><td align="center">6.8</td><td align="center">0.5</td><td align="center">16.4</td><td align="center">7.8</td><td align="center">10.1</td><td align="center">4.1</td></tr></tbody></table><table-wrap-foot><p>BGL, blood glucose level; COITTSS, Corticosteroids and Intensive Insulin Therapy for Septic Shock; NICE-SUGAR, Normoglycemia in Intensive Care Evaluation and Surviving Using Glucose Algorithm Regulation; SD, standard deviation; VDB, Van den Berghe; VISEP, Efficacy of Volume Substitution and Insulin Therapy in Severe Sepsis.</p></table-wrap-foot></table-wrap></p><p>Other trials of glucose control in the ICU used lower glucose intake and insulin doses. In none of these studies did the experimental intervention achieve maintenance of normal BGL like in the Leuven study. NICE-SUGAR is the only study showing increased mortality with tight BGL control. In that study, total caloric intake was much lower than in the Leuven study. In the Specialized Relative Insulin and Nutrition Tables (SPRINT) study, a 35% lowering of hospital mortality for patients with a long stay in the ICU (<italic>P</italic>&#x02009;=&#x02009;0.02) was observed after implementation of tight glucose control when glucose was administered enterally to allow a caloric intake of 25&#x000a0;kcal/kg per day. Likewise, the cumulative insulin dose per day was close to that observed in the experimental group of the Leuven study (67.2 units in SPRINT versus 71 units). These findings are in line with the latest IIT meta-analysis by Marik and Preiser [<xref ref-type="bibr" rid="CR6">6</xref>], who suggested that intravenous calorie administration plays a pivotal role for improvement of outcome during IIT. In contrast, the last Leuven trial, EPaNIC (Early versus late Parenteral Nutrition in Intensive Care), showed that parenteral nutrition administration to achieve a caloric intake of 20 to 25&#x000a0;kcal/kg per day might be detrimental. This raises the question of the effect of an exclusive and important glucose infusion during IIT in critical illness.</p><p>It was then suggested that, in the Leuven trial, difference in observed mortality was secondary to a higher mortality in the control group due to an excessive glucose load. Nevertheless, control mortality in the Leuven study matched the mortality expected from estimation of the EuroSCORE (European System for Cardiac Operative Risk Evaluation). Secondly, a recent meta-analysis suggested that intravenous glucose intake was an independent predictive factor for good outcome in the Leuven studies [<xref ref-type="bibr" rid="CR6">6</xref>]. But whether blood glucose control or insulin administration mediated positive effects in this study was not studied.</p><p>In 2003, Van den Berghe and colleagues [<xref ref-type="bibr" rid="CR41">41</xref>] performed a <italic>post-hoc</italic> analysis of their first study. The authors showed that both total amount of infused insulin and glycemic control were associated with lower mortality (independently of age, delayed ICU admission, Acute Physiology and Chronic Health Evaluation II score, reason for ICU admission, history of malignancy or diabetes, and at-admission hyperglycemia). The strength of association between the mortality rate and the mean BGL seemed to be stronger than with the total daily infused insulin [<xref ref-type="bibr" rid="CR41">41</xref>]. Nevertheless, no statistical comparison was made between these factors in this study. Furthermore, the respective effects of these two entwined factors could be analyzed only in an interventional study comparing gluco-insulinotherapy versus tight glycemic control. Then a recent study by Arabi and colleagues [<xref ref-type="bibr" rid="CR53">53</xref>] with a 2&#x02009;&#x000d7;&#x02009;2 factorial design compared IIT and permissive underfeeding (60% to 70% of daily recommended caloric intake versus 90% to 100%). Their study showed no mortality differences between groups but was underpowered and non-blinded, and the therapeutic goals were not achieved [<xref ref-type="bibr" rid="CR53">53</xref>].</p><p>Finally, in 2011, the Leuven group performed the EPaNIC study that evaluated the timing of parenteral nutrition introduction. In that study, a strategy of early parenteral nutrition initiation was performed with administration of 400&#x000a0;kcal (100&#x000a0;g) at day 1 and 800&#x000a0;kcal at day 2 exclusively via intravenous glucose administration, and then a relay with mixed parenteral and enteral nutrition was performed to achieve calculated daily physiological caloric intake [<xref ref-type="bibr" rid="CR54">54</xref>]. The control group received minimal glucose administration and enteral nutrition was started at day 2 if oral intake was insufficient. Results showed an increased rate of complications in the parenteral nutrition group (infection and cholestasis), whereas the late initiation of parenteral nutrition resulted in a shorter duration of renal replacement therapy, mechanical ventilation, and stay in the ICU. In that study, the amount of administered glucose was three times lower than in the 2001 study, and insulin doses were also lower in both groups: 31 insulin units (interquartile range (IQR) 19 to 48) in the control versus 58 insulin units (IQR 40 to 85) in the experimental group. Furthermore, parenteral nutrition contains lipid at recommended doses that could present detrimental effects as fat oxidation is a high oxygen-consuming metabolic pathway. A <italic>post-hoc</italic> analysis of EPaNIC concerning the first 2&#x000a0;days in the ICU in that study before introduction of parenteral nutrition might be of interest to clinicians and help them determine whether high glucose administration during IIT is beneficial for patients. We will present clinical and experimental evidence that may support the use of a glucose-insulin administration strategy.</p></sec><sec id="Sec14"><title>Is gluco-insulinotherapy associated with a decreased incidence of hypoglycemia?</title><p>The clinical signs of hypoglycemia are commonly masked in sedated patients. Thus, in clinical trials, hypoglycemia was defined empirically by a BGL value of less than 40&#x000a0;mg/dL. Its incidence varied from 5.1% to 18.7% in patients with IIT and from 0.5% to 4.1% in control groups. Seizures and comas have occasionally been observed following severe hypoglycemic episodes without establishing a clear causal relationship [<xref ref-type="bibr" rid="CR55">55</xref>]. Neuronal death during or following hypoglycemia has also been found in both animal and human models, but hypoglycemia does not seem to affect neurocognitive development in children [<xref ref-type="bibr" rid="CR56">56</xref>] but may contribute to long-term cognitive impairment following critical illness in adults [<xref ref-type="bibr" rid="CR57">57</xref>]. The existence of a direct causal link between hypoglycemia and mortality remains controversial, and hypoglycemia could reflect only a more severe illness. Some epidemiologic studies have found that only early or spontaneous hypoglycemia was independently associated with death in critically ill patients [<xref ref-type="bibr" rid="CR58">58</xref>, <xref ref-type="bibr" rid="CR59">59</xref>]. Preventive interventions are thus warranted in such a situation. Whether an increased daily amount of carbohydrate administered would decrease the risk of hypoglycermia during tight BGL is unknown.</p><p>In a retrospective study by Arabi and colleagues [<xref ref-type="bibr" rid="CR60">60</xref>], glucose intake was not a risk or protective factor of hypoglycemia whereas insulin daily dosage was an evident risk factor (73.5&#x02009;&#x000b1;&#x02009;36.7 in the group presenting hypoglycemia versus 47.5&#x02009;&#x000b1;&#x02009;51.8; <italic>P</italic>&#x02009;&#x0003c;&#x02009;0.0001) [<xref ref-type="bibr" rid="CR60">60</xref>]. Actually, caloric intake lowering (gastroparesis, intravenous glucose, or enteral nutrition lowering) without insulin adjustment may be one of the most frequent risk factors for hypoglycemia [<xref ref-type="bibr" rid="CR54">54</xref>, <xref ref-type="bibr" rid="CR55">55</xref>, <xref ref-type="bibr" rid="CR60">60</xref>, <xref ref-type="bibr" rid="CR61">61</xref>], and no study evaluated the effect of gluco-insulinotherapy on hypoglycemia rate. The recent EPaNIC study showed a decreased rate of hypoglycemia during IIT when early parenteral nutrition was initiated (1.9% versus 3.5%, <italic>P</italic>&#x02009;=&#x02009;0.001), suggesting a possible protective role of gluco-insulinotherapy to be explored in an interventional study [<xref ref-type="bibr" rid="CR54">54</xref>].</p></sec><sec id="Sec15"><title>Effects of high insulin and glucose intake on organs</title><p>Gluco-insulinotherapy consists of a high amount of glucose infused and higher insulinemia. Effects of insulin on glycemia lowering are mediated mostly by an increase in cellular uptake of glucose through GLUT 4 translocation to the membrane. GLUT 4 is located mostly on adipocytes and skeletal muscle cells [<xref ref-type="bibr" rid="CR1">1</xref>]. Thus, mostly GLUT 4-expressing cells consume glucose administered intravenously during IIT.</p><p>During early sepsis, metabolic stress resulted in glycogenolysis and depleted energetic reserves as shown in skeletal muscle biopsies [<xref ref-type="bibr" rid="CR33">33</xref>, <xref ref-type="bibr" rid="CR62">62</xref>]. This ATP depletion was correlated with poor outcome, and in addition recovery from sepsis was preceded by normalization of the phosphocreatine/ATP ratio [<xref ref-type="bibr" rid="CR62">62</xref>]. Indeed, energy depletion during sepsis could be a risk factor for CINM, an ICU complication associated with higher mortality [<xref ref-type="bibr" rid="CR63">63</xref>]. In fact, skeletal muscle protein levels were higher in the IIT group consistently with anabolic effects of insulin [<xref ref-type="bibr" rid="CR33">33</xref>]. Insulin augments the number of GLUT 4 receptors in adipose and skeletal muscle cells and the glucose uptake by myocytes [<xref ref-type="bibr" rid="CR64">64</xref>]. This higher amount of energetic substrate may be a protective factor against energy depletion and sarcopenia and may lower the risk to develop CINM. These findings are in line with the observed lower rate of CINM and the late improvement in survival in the Leuven trials [<xref ref-type="bibr" rid="CR4">4</xref>, <xref ref-type="bibr" rid="CR11">11</xref>].</p><p>Experiments from the Leuven cohort showed that maintenance of normal BGL protected the liver and skeletal muscles. In the liver, glucose uptake through GLUT 2 is independent of BGLs. Indeed, untreated hyperglycemia was associated with severely damaged mitochondria with altered complex I and IV activities [<xref ref-type="bibr" rid="CR33">33</xref>]. Furthermore, insulin administration during injury partly suppresses gluconeogenesis [<xref ref-type="bibr" rid="CR65">65</xref>]. Gluconeogenesis is an active process occurring mostly in the liver that requires four molecules of ATP and two molecules of GTP. This energy requirement may enhance hypoxic injury in liver during stress and could be counteracted with insulin administration [<xref ref-type="bibr" rid="CR64">64</xref>&#x02013;<xref ref-type="bibr" rid="CR66">66</xref>].</p><p>Whether gluco-insulinotherapy rather than BGL control protects the liver from hypoxic injury has been studied in few human and experimental studies. One study conducted by a different group showed beneficial effects of insulin independently of BGLs on hepatocyte apoptosis, cytolysis, and expression of inflammatory markers [<xref ref-type="bibr" rid="CR65">65</xref>]. Further studies in the Leuven group did not reproduce these results and showed a lower blood level of transaminases in burn-injured rabbits with BGL control rather than insulin administration [<xref ref-type="bibr" rid="CR67">67</xref>]. Liver injury may be mediated by a mitochondriopathy in reaction to cellular hyperglycemia and enhanced glycolysis and is likely to mediate organ damage [<xref ref-type="bibr" rid="CR68">68</xref>]. As insulin sensitivity is not overcome during IIT in the liver, glucose uptake by hepatocytes is likely to be dependent on glycemia rather than insulinemia [<xref ref-type="bibr" rid="CR66">66</xref>].</p><p>In sepsis models, insulin has been shown to improve immune cell function independently of glycemia. It inhibits the apoptosis of activated macrophages, may modulate antigen presentation, and improves chemotaxis and phagocytic properties. Finally, insulin may modulate the balance between lymphocyte T helper type and lymphocyte T helper type 2 cells, favoring anti-inflammation and repair function [<xref ref-type="bibr" rid="CR69">69</xref>]. Such effects of insulin on the immune system may account for the reduced rate of bloodstream infection during IIT [<xref ref-type="bibr" rid="CR4">4</xref>].</p><p>During sepsis, the heart shows little or no insulin resistance [<xref ref-type="bibr" rid="CR70">70</xref>] and lowers its glucose consumption [<xref ref-type="bibr" rid="CR48">48</xref>, <xref ref-type="bibr" rid="CR70">70</xref>]. In porcine models, glucose and insulin infusion favored glucose and lactate utilization. This results in improvement in inotropic function without higher oxygen consumption observed in different studies [<xref ref-type="bibr" rid="CR67">67</xref>, <xref ref-type="bibr" rid="CR71">71</xref>]. In fact, during acute coronary syndrome, glucose insulin potassium therapy was associated with substantial survival and may prevent arrhythmias [<xref ref-type="bibr" rid="CR72">72</xref>]. However, the benefit of glucose insulin potassium infusion in patients with myocardial infarction remains controversial. In the ICU, IIT was not associated with reduced time or doses of inotropic support [<xref ref-type="bibr" rid="CR4">4</xref>] and was associated with a higher incidence of cardiovascular death in NICE-SUGAR [<xref ref-type="bibr" rid="CR5">5</xref>], but the Leuven study introduced IIT with an important amount of intravenous glucose administered to cardiovascular post-surgical patients. Such an inotropic effect may have improved organ perfusion and contributed to the lower renal failure rate and the better outcome in these patients.</p><p>In summary, gluco-insulinotherapy may present protective effects on muscle or improve immune or cardiac function. Contrary to Marik and Bellomo [<xref ref-type="bibr" rid="CR18">18</xref>] in their recent comment, we hypothesize that gluco-insulinotherapy may be a more beneficial rather than a restrictive strategy. This issue needs to be further studied.</p></sec></sec><sec id="Sec16" sec-type="conclusions"><title>Conclusions</title><p>The era of glucose control in the ICU started in 2001. Untreated SIH no doubt favors morbidity and mortality. Critical analyses of randomized controlled trials have suggested that glucose control is more likely to be associated with survival benefit when strict normal glucose levels are achieved and early high glucose intake is provided. An interventional study evaluating liberal and restrictive glucose intake during IIT is warranted to provide reliable evidence.</p></sec>

Chipping away at major depressive disorder

<p>An intriguing recent study examines the role of miR-1202, a glutamate receptor regulating microRNA, in regulating major depressive disorder.</p>

<contrib contrib-type="author" corresp="yes" id="A1"><name><surname>Rucker</surname><given-names>James JH</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>james.rucker@kcl.ac.uk</email></contrib><contrib contrib-type="author" id="A2"><name><surname>McGuffin</surname><given-names>Peter</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>james.rucker@kcl.ac.uk</email></contrib>

Genome Biology

<sec><title>The challenge of major depressive disorder</title><p>Major depressive disorder (MDD), a common, yet debilitating and economically costly, psychiatric disorder [[<xref ref-type="bibr" rid="B1">1</xref>]], has proven surprisingly refractory to molecular-genetic investigation and the development of biomarkers for diagnosis and treatment. However, although there was initial excitement over positive findings in studies of genetic linkage [[<xref ref-type="bibr" rid="B2">2</xref>]], genome-wide association of single-nucleotide polymorphisms (SNPs) [[<xref ref-type="bibr" rid="B3">3</xref>]] and copy-number variants [[<xref ref-type="bibr" rid="B4">4</xref>]], none of these areas of investigation has resulted in replicated findings, and a recent genome-wide mega-analysis of SNPs was entirely negative [[<xref ref-type="bibr" rid="B5">5</xref>]]. Similarly, no biomarker for MDD has been found to be clinically useful, despite the pressing need in psychiatry for objective tests for diagnosis and treatment response. The usual genetic model that is used to explain liability in complex disorders is one of polygenic heterogeneity [[<xref ref-type="bibr" rid="B6">6</xref>]], with many variants of small effect co-acting to produce a liability to psychiatric disorders that develops over the lifetime in interplay with the environment. Such a model was first postulated for common psychiatric disorders nearly a half-century ago [[<xref ref-type="bibr" rid="B7">7</xref>]] and has proved to be enduring. However, it remains a general theory that is &#x02018;agnostic&#x02019; about what the component polygenes might be.</p><p>In search of further specific elements of the genome that could help explain the heritability of common psychiatric disorders, the role of microRNAs (miRNAs) has recently been investigated. Within mammals, it is estimated that 60% of all protein-coding genes are regulated by miRNAs, contributing widely to the regulation of most cellular biochemical processes [[<xref ref-type="bibr" rid="B8">8</xref>]]. In addition, miRNAs are found in plasma and might be useful as biomarkers [[<xref ref-type="bibr" rid="B9">9</xref>]] - thus, they are particularly relevant candidates for study in psychiatric disorders. MDD tends to develop after the onset of puberty, with a peak in incidence during early adult life, and a particular preponderance for the female sex. This might suggest regulatory biological elements that segregate by sex and occur at specific points in time during development.</p><p>Recently, Lopez and colleagues have published a thought-provoking article in <italic>Nature Medicine</italic> that implicates a particular miRNA in the development of MDD [[<xref ref-type="bibr" rid="B10">10</xref>]]. The study provides evidence from a number of different experimental paradigms that miR-1202 exists in a dose-dependent relationship with expression of the gene <italic>GRM4</italic> (encoding metabotropic glutamate receptor 4) in the human prefrontal cortex and that the expression of miR-1202 is related to successful antidepressant treatment.</p></sec><sec><title>miR-1202 is differentially expressed in patients with depression</title><p>The authors initially investigated differences in miRNA expression in the prefrontal cortices of deceased individuals with diagnoses of MDD compared with deceased, psychiatrically healthy control samples. The levels of miR-1202 expression were significantly decreased in brains from depressed individuals when compared with those of controls. Psychiatric diagnosis was made postmortem, based on medical records. An <italic>in silico</italic> investigation of evolutionary conservation of miR-1202 across the genomes of 100 animal species revealed that miR-1202 is present only in humans and primates. To confirm this finding experimentally, Lopez and colleagues measured expression in the brains of six representative animal species: human, cynomolgus monkey (<italic>Macaca fasciculari</italic>), rhesus monkey (<italic>Macaca mulatta</italic>), rat (<italic>Rattus norvegicus</italic>), mouse (<italic>Mus musculus</italic>) and chicken (<italic>Gallus gallus</italic>) [[<xref ref-type="bibr" rid="B10">10</xref>]]. The authors showed that miR-1202 was not found in rat, mouse and chicken brain but was found in primate brains, with the highest levels in human brain, which also had higher levels than in 10 other investigated forms of human tissue. To predict the functional consequence of miR-1202, the authors used five different miRNA target-prediction databases to generate potential gene targets, cross-referencing these with genes expressed and upregulated in the prefrontal cortices of subjects with depression. Of five potential genes, miR-1202 was found to correlate negatively only with expression of <italic>GRM4</italic>, encoding subtype four of the metabotropic glutamate receptor. In a replication analysis, the authors measured the expression of miR-1202 and <italic>GRM4</italic> in an independent sample of human prefrontal cortices that also included depressed individuals who were taking antidepressants at the time of death. Replicating their original findings, the authors also noted that miR-1202 levels were no different between controls and depressed individuals taking antidepressants, suggesting that not only are the levels of miR-1202 inversely related to the level of <italic>GRM4</italic> expression, but also that antidepressants modulate the levels of miR-1202 to affect <italic>GRM4</italic> expression.</p></sec><sec><title>Chronic antidepressant administration upregulates miR-1202</title><p>To investigate this modulatory effect further, human embryonic kidney (HEK) cells were used to perform functional experiments investigating the interaction between miR-1202, <italic>GRM4</italic> expression and antidepressants. HEK cells were used because they particularly express <italic>GRM4</italic> without expressing miR-1202. Treatment of HEK cells with a miR-1202 mimic resulted in a decreased expression of <italic>GRM4,</italic> and co-treatment with the miR-1202 mimic together with an agent that interfered with the predicted binding sites of miR-1202 to the transcribed <italic>GRM4</italic> mRNA resulted in <italic>GRM4</italic> expression levels returning to baseline. Further investigation of the relationship between miR-1202 and <italic>GRM4</italic> with agonists and antagonists of <italic>GRM4</italic> in neural progenitor cells (NPCs) suggested that miR-1202 exists in a bidirectional relationship with <italic>GRM4</italic> expression. To investigate the effects of antidepressants on this relationship, the authors treated NPCs (which show a serotonergic profile) with the archetypal tricyclic antidepressant imipramine, the selective serotonin reuptake inhibitor (SSRI) antidepressant citalopram or a control possessing no active drug. Although there was no effect of acute (24 hours) treatment with either drug on miR-1202 levels or <italic>GRM4</italic> expression, chronic treatment (15 days) with either drug resulted in an up-regulation of miR-1202 - findings that were also confirmed using immunohistochemistry. Of note, this effect was not observed when treating cells with the drugs valproate or lithium, neither of which has a direct effect on the sodium-dependent serotonin transporter (SERT). Knockdown experiments showed that the increase in miR-1202 concentrations is dependent on SERT and the reuptake blockade elicited by conventional antidepressants. To rule out global miRNA dysregulation as an explanation of the observed effects, the authors additionally measured the expression levels of miRNAs known to be ubiquitously expressed, but found no differences in expression after chronic antidepressant treatment.</p><p>To confirm these findings <italic>in vivo</italic>, the authors measured blood levels of miR-1202 in treatment-na&#x000ef;ve patients with MDD and healthy controls. The levels of mir-1202 were found to be decreased in patients with depression. Patients were then treated with citalopram for eight weeks and classified as responders or non-responders on the basis of relative changes in Hamilton Depression (HAM-D) rating-scale scores. Although those who achieved remission from symptoms as specified by the HAM-D score showed increased miR-1202 levels after eight weeks of treatment with citalopram, there was no difference in expression in miR-1202 levels between non-responders and psychiatrically healthy controls without major depressive disorder. The change in depression severity, as defined by HAM-D scores, was negatively correlated with miR-1202 expression levels.</p></sec><sec><title>Concluding remarks</title><p>This paper presents a striking investigative narrative, providing evidence from a number of different angles that a specific miRNA is a biologically plausible biomarker for detection of, and treatment response in, MDD and is potentially of considerable interest to the relevant research and clinical communities. However, it is sensible to urge caution in such circumstances. It would indeed be a remarkable finding if such a clinically and biologically heterogeneous disorder as MDD was reducible, even in part, to a single biological entity, and one might have expected indications towards this previously. The glutamate system is the major excitatory neurotransmitter in the brain, and it would be unsurprising to find biological links to MDD within it, but certainly surprising to find that molecular regulation of a particular subtype of glutamate receptor was associated with something as conceptually distant as the HAM-D score, a clinically applied measure of subjectively experienced symptoms and observable clinical signs in MDD. Genetic investigation into MDD has failed to explain why significant heritability figures are obtained in twin studies, but a tagging SNP would be expected to achieve genome-wide levels of statistical significance in mega-analyses of MDD [[<xref ref-type="bibr" rid="B5">5</xref>]] if any particular genetic or epigenetic explanation contributed very significantly to the disorder - although there are likely also to be exceptions to this observation. Overall, these provocative and interesting results will certainly require independent replication - however, they present a potentially novel and intriguing facet of the complex genetics of MDD.</p></sec><sec><title>Competing interests</title><p>PM has received consultancy fees and honoraria for participating in expert panels for pharmaceutical companies, including GlaxoSmithKline and Pfizer. JR declares no competing interests.</p></sec>

Have plants evolved to self-immolate?

<p>By definition fire prone ecosystems have highly combustible plants, leading to the hypothesis, first formally stated by Mutch in 1970, that community flammability is the product of natural selection of flammable traits. However, proving the &#x0201c;Mutch hypothesis&#x0201d; has presented an enormous challenge for fire ecologists given the difficulty in establishing cause and effect between landscape fire and flammable plant traits. Individual plant traits (such as leaf moisture content, retention of dead branches and foliage, oil rich foliage) are known to affect the flammability of plants but there is no evidence these characters evolved specifically to self-immolate, although some of these traits may have been secondarily modified to increase the propensity to burn. Demonstrating individual benefits from self-immolation is extraordinarily difficult, given the intersection of the physical environmental factors that control landscape fire (fuel production, dryness and ignitions) with community flammability properties that emerge from numerous traits of multiple species (canopy cover and litter bed bulk density). It is more parsimonious to conclude plants have evolved mechanisms to tolerate, but not promote, landscape fire.</p>

<contrib contrib-type="author"><name><surname>Bowman</surname><given-names>David M. J. S.</given-names></name><uri xlink:type="simple" xlink:href="http://community.frontiersin.org/people/u/137024"/></contrib><contrib contrib-type="author"><name><surname>French</surname><given-names>Ben J.</given-names></name><uri xlink:type="simple" xlink:href="http://community.frontiersin.org/people/u/137568"/></contrib><contrib contrib-type="author"><name><surname>Prior</surname><given-names>Lynda D.</given-names></name><xref ref-type="author-notes" rid="fn001">^*</xref><uri xlink:type="simple" xlink:href="http://community.frontiersin.org/people/u/114865"/></contrib>

Frontiers in Plant Science

<sec sec-type="introduction" id="s1"><title>Introduction</title><p>The combination of carbon rich biomass, atmospheric oxygen, and ignitions makes landscape fire inevitable on Earth (Bowman et al., <xref rid="B14" ref-type="bibr">2009</xref>). However, the occurrence, spread, and energy released by landscape fires is controlled by the physical environment. The most prominent environmental factor is climate because it influences the production of biomass, fuel arrangement across landscapes and its dryness, as well as providing lightning ignitions (Bradstock et al., <xref rid="B18" ref-type="bibr">2012</xref>). The only life-forms that make fire are humans, and we, like our antecedents, are powerful agents in influencing the occurrence and spread of fires, given our capacities to modify fuels, provide ignitions and suppress fires (Bowman et al., <xref rid="B13" ref-type="bibr">2011</xref>; Archibald et al., <xref rid="B7" ref-type="bibr">2012</xref>). To what degree plant life has influenced the occurrence, extent and intensity of landscape fire remains controversial (Bradshaw et al., <xref rid="B16" ref-type="bibr">2011a</xref>,<xref rid="B17" ref-type="bibr">b</xref>; Keeley et al., <xref rid="B54" ref-type="bibr">2011b</xref>). Mutch (<xref rid="B74" ref-type="bibr">1970</xref>) hypothesized that &#x0201c;fire dependent plant communities burn more readily than non-fire dependent communities because natural selection has favored characteristics that make them more flammable&#x0201d; (Table <xref ref-type="table" rid="T1">1</xref>). The &#x0201c;Mutch hypothesis&#x0201d; has logical appeal and is intellectually consequential for fire ecology and pyrogeography because it provides these disciplines with an evolutionary platform. However, because landscape fires affect entire plant communities rather than being restricted to individuals with heritable flammable characteristics, it is difficult to avoid group selection arguments (Snyder, <xref rid="B99" ref-type="bibr">1984</xref>; Troumbis and Trabaud, <xref rid="B104" ref-type="bibr">1989</xref>; Bond and Midgley, <xref rid="B11" ref-type="bibr">1995</xref>; Scarff and Westoby, <xref rid="B90" ref-type="bibr">2006</xref>).</p><table-wrap id="T1" position="float"><label>Table 1</label><caption><p><bold>Summary of hypotheses regarding evolution of flammable traits in plants, and possible examples</bold>.</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1"><bold>Syndrome</bold></th><th align="left" rowspan="1" colspan="1"><bold>Ancestral state</bold></th><th align="left" rowspan="1" colspan="1"><bold>Evolved state</bold></th><th align="left" rowspan="1" colspan="1"><bold>Example</bold></th><th align="left" rowspan="1" colspan="1"><bold>References</bold></th></tr></thead><tbody><tr><td valign="top" align="left" rowspan="1" colspan="1">Mutch</td><td valign="top" align="left" rowspan="1" colspan="1">Recovery/tolerance of fire</td><td valign="top" align="left" rowspan="1" colspan="1">High flammability</td><td valign="top" align="left" rowspan="1" colspan="1"><italic>Eucalyptus</italic></td><td valign="top" align="left" rowspan="1" colspan="1">Crisp et al., <xref rid="B25" ref-type="bibr">2011</xref></td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">Mutch's converse</td><td valign="top" align="left" rowspan="1" colspan="1">High flammability</td><td valign="top" align="left" rowspan="1" colspan="1">Recovery/tolerance of fire</td><td valign="top" align="left" rowspan="1" colspan="1">Serotiny and thick bark in <italic>Pinus</italic> Thick bark, xylopodia in savanna plants Fire-cued flowering in orchids</td><td valign="top" align="left" rowspan="1" colspan="1">He et al., <xref rid="B44" ref-type="bibr">2012</xref> Simon et al., <xref rid="B98" ref-type="bibr">2009</xref> Bytebier et al., <xref rid="B19" ref-type="bibr">2011</xref></td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">Midgley's alternative</td><td valign="top" align="left" rowspan="1" colspan="1">High flammability</td><td valign="top" align="left" rowspan="1" colspan="1">Low flammability</td><td valign="top" align="left" rowspan="1" colspan="1">Branch shedding in <italic>Pinus</italic></td><td valign="top" align="left" rowspan="1" colspan="1">He et al., <xref rid="B44" ref-type="bibr">2012</xref></td></tr></tbody></table><table-wrap-foot><p>It is important to note that proving evolution of flammable traits, fire tolerance and post fire recovery demands extraordinarily rigorous studies that are yet to be achieved (Bradshaw et al., <xref rid="B16" ref-type="bibr">2011a</xref>,<xref rid="B17" ref-type="bibr">b</xref>; Keeley et al., <xref rid="B54" ref-type="bibr">2011b</xref>). We call this stricture &#x0201c;Bradshaw's null.&#x0201d;</p></table-wrap-foot></table-wrap><p>A number of theoretical models have attempted to reconcile the evolution of flammability with individualistic selection theory by proposing ways that self-immolation can increase individual fitness or advantage to their offspring (Bond and Midgley, <xref rid="B11" ref-type="bibr">1995</xref>; Kerr et al., <xref rid="B55" ref-type="bibr">1999</xref>; Gagnon et al., <xref rid="B38" ref-type="bibr">2010</xref>). For instance, Bond and Midgley (<xref rid="B11" ref-type="bibr">1995</xref>) developed a &#x0201c;kill thy neighbor&#x0201d; model, which demonstrated that a trait promoting canopy flammability amongst a population of closely spaced conspecific individuals could increase reproductive fitness on the condition it also conferred other evolutionary advantages. Recently, Midgley (<xref rid="B65" ref-type="bibr">2013</xref>) has withdrawn his support for this model because of unrealistic assumptions, such as the need for the seed shadow of the flammable individual to closely align with the fire footprint, and for its seedlings that inherit the flammable trait to be more competitive in post-fire environments. Likewise, Midgley (<xref rid="B65" ref-type="bibr">2013</xref>) argues that the &#x0201c;pyrogenicity as protection&#x0201d; hypothesis (Gagnon et al., <xref rid="B38" ref-type="bibr">2010</xref>), which posits that flammable crowns are protective of soil seed banks and subterranean bud banks, shares similar flaws to the Bond and Midgley (<xref rid="B11" ref-type="bibr">1995</xref>) model.</p><p>A feature of the discussion about the evolution of flammability is that flammability traits have been conflated with strategies that enable plants to recover following fire, such as resprouting from basal or aerial bud banks, and storing seeds in aerial or soil seed banks (Saura-Mas et al., <xref rid="B88" ref-type="bibr">2010</xref>; Clarke et al., <xref rid="B21" ref-type="bibr">2013</xref>). Such strategies manifestly increase the fitness of individual plants in fire prone landscapes. Traits that unambiguously assist post-fire recovery and regeneration can be used in ancestral trait reconstructions, illuminating evolutionary processes within clades. Examples include fire-cued flowering (Bytebier et al., <xref rid="B19" ref-type="bibr">2011</xref>), the epicormic strands that allow eucalypts to resprout after fire (Crisp et al., <xref rid="B25" ref-type="bibr">2011</xref>), and xylopodia and thick corky bark in South American savanna species (Simon et al., <xref rid="B98" ref-type="bibr">2009</xref>) (Table <xref ref-type="table" rid="T1">1</xref>). In contrast, traits that purportedly increase flammability are not so obviously related to the fitness of individuals. Some authors have rejected the notion that plants have evolved any traits to be flammable, indeed questioning the entire basis of the plant -fire evolutionary nexus (Bradshaw et al., <xref rid="B16" ref-type="bibr">2011a</xref>). This leads to the basic question that is the subject of this review: &#x0201c;what plant traits and community attributes are known to increase flammability and could have arisen from natural selection through an evolutionary fire-feedback loop?&#x0201d; For the purposes of this review we define flammability as the propensity of living or dead plant material to ignite and sustain combustion.</p></sec><sec><title>Flammability traits</title><sec><title>Biomass water content</title><p>Water in plant tissue is a heat sink, increasing the amount of energy required for fuels to ignite and sustain combustion. Therefore moisture content of living and dead fuels is the most fundamental constraint on biomass flammability (Gill and Moore, <xref rid="B41" ref-type="bibr">1996</xref>; Alessio et al., <xref rid="B2" ref-type="bibr">2008b</xref>; De Lillis et al., <xref rid="B29" ref-type="bibr">2009</xref>; Alexander and Cruz, <xref rid="B3" ref-type="bibr">2013</xref>; Murray et al., <xref rid="B73" ref-type="bibr">2013</xref>) (Table <xref ref-type="table" rid="T2">2</xref>). Leaf moisture content strongly affects flammability and is highly variable amongst life forms and biomes, exceeding 95% in succulents (Lamont and Lamont, <xref rid="B58" ref-type="bibr">2000</xref>) and being as low as 20% in some sclerophyllous species (De Lillis et al., <xref rid="B29" ref-type="bibr">2009</xref>). Although drought tolerating plants typically have more combustible living and dead foliage than mesic species, this correlation largely reflects the effect of the environment rather than inherent features that have evolved to increase flammability (Dickinson and Kirkpatrick, <xref rid="B31" ref-type="bibr">1985</xref>; Berry et al., <xref rid="B10" ref-type="bibr">2011</xref>; Hoffmann et al., <xref rid="B46" ref-type="bibr">2012</xref>; Davies and Nafus, <xref rid="B27" ref-type="bibr">2013</xref>; Seo and Choung, <xref rid="B96" ref-type="bibr">2014</xref>). This point is exemplified by otherwise non-flammable rain forest foliage and litter beds burning under extreme drought conditions (Cochrane and Laurance, <xref rid="B23" ref-type="bibr">2008</xref>) (Figure <xref ref-type="fig" rid="F1">1A</xref>).</p><table-wrap id="T2" position="float"><label>Table 2</label><caption><p><bold>Summary of the evidence for the effects and evolutionary origin of potential flammability</bold>.</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1"><bold>Property</bold></th><th align="left" rowspan="1" colspan="1"><bold>Evidence of effect on flammability</bold></th><th align="left" rowspan="1" colspan="1"><bold>References</bold></th><th align="left" rowspan="1" colspan="1"><bold>Evidence of evolution for flammability</bold></th><th align="left" rowspan="1" colspan="1"><bold>References</bold></th></tr></thead><tbody><tr><td valign="top" align="left" colspan="5" rowspan="1"><bold>LEAF</bold></td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">Leaf moisture content</td><td valign="top" align="left" rowspan="1" colspan="1">Strong</td><td valign="top" align="left" rowspan="1" colspan="1">Gill and Moore, <xref rid="B41" ref-type="bibr">1996</xref>; Dimitrakopoulos and Papaioannou, <xref rid="B32" ref-type="bibr">2001</xref>; Alessio et al., <xref rid="B1" ref-type="bibr">2008a</xref>,<xref rid="B2" ref-type="bibr">b</xref>; De Lillis et al., <xref rid="B29" ref-type="bibr">2009</xref>; Page et al., <xref rid="B78" ref-type="bibr">2012</xref>; Alexander and Cruz, <xref rid="B3" ref-type="bibr">2013</xref>; Murray et al., <xref rid="B73" ref-type="bibr">2013</xref></td><td valign="top" align="left" rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1"/></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">Organic chemistry</td><td valign="top" align="left" rowspan="1" colspan="1">Strong</td><td valign="top" align="left" rowspan="1" colspan="1">Dickinson and Kirkpatrick, <xref rid="B31" ref-type="bibr">1985</xref>; White, <xref rid="B109" ref-type="bibr">1994</xref>; Owens et al., <xref rid="B77" ref-type="bibr">1998</xref>; Kerr et al., <xref rid="B55" ref-type="bibr">1999</xref>; Schwilk and Kerr, <xref rid="B94" ref-type="bibr">2002</xref>; De Lillis et al., <xref rid="B29" ref-type="bibr">2009</xref>; Holmes, <xref rid="B47" ref-type="bibr">2009</xref>; Ormeno et al., <xref rid="B76" ref-type="bibr">2009</xref>; Page et al., <xref rid="B78" ref-type="bibr">2012</xref>; but see (Alessio et al., <xref rid="B1" ref-type="bibr">2008a</xref>,<xref rid="B2" ref-type="bibr">b</xref>)</td><td valign="top" align="left" rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1"/></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">Inorganic chemistry</td><td valign="top" align="left" rowspan="1" colspan="1">Moderate</td><td valign="top" align="left" rowspan="1" colspan="1">Dickinson and Kirkpatrick, <xref rid="B31" ref-type="bibr">1985</xref>; Scarff and Westoby, <xref rid="B90" ref-type="bibr">2006</xref>; Scarff et al., <xref rid="B89" ref-type="bibr">2012</xref></td><td valign="top" align="left" rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1"/></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">Leaf dimensions</td><td valign="top" align="left" rowspan="1" colspan="1">Moderate</td><td valign="top" align="left" rowspan="1" colspan="1">Direct effect (Gill and Moore, <xref rid="B41" ref-type="bibr">1996</xref>; Murray et al., <xref rid="B73" ref-type="bibr">2013</xref>) and indirect effect through litter bed structure (Scarff and Westoby, <xref rid="B90" ref-type="bibr">2006</xref>; Schwilk and Caprio, <xref rid="B93" ref-type="bibr">2011</xref>; De Magalh&#x000e3;es and Schwilka, <xref rid="B30" ref-type="bibr">2012</xref>; Engber and Varner III, <xref rid="B36" ref-type="bibr">2012</xref>)</td><td valign="top" align="left" rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1"/></tr><tr><td valign="top" align="left" colspan="5" rowspan="1"><bold>WHOLE PLANT</bold></td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">Phenology</td><td valign="top" align="left" rowspan="1" colspan="1">Strong</td><td valign="top" align="left" rowspan="1" colspan="1">Bajocco et al., <xref rid="B8" ref-type="bibr">2010</xref>; Ripley et al., <xref rid="B83" ref-type="bibr">2010</xref>; Wittich, <xref rid="B111" ref-type="bibr">2011</xref>; De Angelis et al., <xref rid="B28" ref-type="bibr">2012</xref></td><td valign="top" align="left" rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1"/></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">Leaf retention</td><td valign="top" align="left" rowspan="1" colspan="1">Moderate</td><td valign="top" align="left" rowspan="1" colspan="1">He et al., <xref rid="B45" ref-type="bibr">2011</xref>; Santana et al., <xref rid="B87" ref-type="bibr">2011</xref></td><td valign="top" align="left" rowspan="1" colspan="1">Equivocal</td><td valign="top" align="left" rowspan="1" colspan="1">He et al., <xref rid="B45" ref-type="bibr">2011</xref></td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">Decorticating bark</td><td valign="top" align="left" rowspan="1" colspan="1">Moderate</td><td valign="top" align="left" rowspan="1" colspan="1">Ganteaume et al., <xref rid="B39" ref-type="bibr">2009</xref>; Koo et al., <xref rid="B57" ref-type="bibr">2010</xref>; Ellis, <xref rid="B34" ref-type="bibr">2011</xref></td><td valign="top" align="left" rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1"/></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">Branch retention</td><td valign="top" align="left" rowspan="1" colspan="1">Strong</td><td valign="top" align="left" rowspan="1" colspan="1">Schwilk and Ackerly, <xref rid="B92" ref-type="bibr">2001</xref>; Schwilk, <xref rid="B91" ref-type="bibr">2003</xref>; Ne'eman et al., <xref rid="B75" ref-type="bibr">2004</xref>; Keeley, <xref rid="B51" ref-type="bibr">2012</xref>; Seo and Choung, <xref rid="B96" ref-type="bibr">2014</xref></td><td valign="top" align="left" rowspan="1" colspan="1">Equivocal</td><td valign="top" align="left" rowspan="1" colspan="1">He et al., <xref rid="B44" ref-type="bibr">2012</xref></td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">Plant architecture</td><td valign="top" align="left" rowspan="1" colspan="1">Moderate</td><td valign="top" align="left" rowspan="1" colspan="1">Archibald and Bond, <xref rid="B6" ref-type="bibr">2003</xref>; Schwilk, <xref rid="B91" ref-type="bibr">2003</xref>; Mitsopoulos and Dimitrakopoulos, <xref rid="B67" ref-type="bibr">2007</xref>; Hoffmann et al., <xref rid="B46" ref-type="bibr">2012</xref>; Ledig et al., <xref rid="B59" ref-type="bibr">2013</xref></td><td valign="top" align="left" rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1"/></tr><tr><td valign="top" align="left" colspan="5" rowspan="1"><bold>COMMUNITY</bold></td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">Fuel moisture</td><td valign="top" align="left" rowspan="1" colspan="1">Strong</td><td valign="top" align="left" rowspan="1" colspan="1">Dickinson and Kirkpatrick, <xref rid="B31" ref-type="bibr">1985</xref>; Bowman and Wilson, <xref rid="B15" ref-type="bibr">1988</xref>; Rollins et al., <xref rid="B84" ref-type="bibr">2002</xref>; Ray et al., <xref rid="B82" ref-type="bibr">2005</xref>; Jolly, <xref rid="B49" ref-type="bibr">2007</xref>; Hoffmann et al., <xref rid="B46" ref-type="bibr">2012</xref>; Alexander and Cruz, <xref rid="B3" ref-type="bibr">2013</xref>; Davies and Nafus, <xref rid="B27" ref-type="bibr">2013</xref></td><td valign="top" align="left" rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1"/></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">Fuel load</td><td valign="top" align="left" rowspan="1" colspan="1">Strong</td><td valign="top" align="left" rowspan="1" colspan="1">Rossiter et al., <xref rid="B7new" ref-type="bibr">2003</xref>; Brooks et al., <xref rid="B1new" ref-type="bibr">2004</xref>; Mitsopoulos and Dimitrakopoulos, <xref rid="B67" ref-type="bibr">2007</xref>; Ganteaume et al., <xref rid="B113" ref-type="bibr">2011</xref>; Hoffmann et al., <xref rid="B46" ref-type="bibr">2012</xref>; McCaw et al., <xref rid="B5new" ref-type="bibr">2012</xref>; Scott et al., <xref rid="B95" ref-type="bibr">2014</xref>; but see (Saura-Mas et al., <xref rid="B88" ref-type="bibr">2010</xref>)</td><td valign="top" align="left" rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1"/></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">Fuel arrangement</td><td valign="top" align="left" rowspan="1" colspan="1">Strong</td><td valign="top" align="left" rowspan="1" colspan="1">Bowman and Wilson, <xref rid="B15" ref-type="bibr">1988</xref>; D'Antonio and Vitousek, <xref rid="B26" ref-type="bibr">1992</xref>; Lippincott, <xref rid="B4new" ref-type="bibr">2000</xref>; Rollins et al., <xref rid="B84" ref-type="bibr">2002</xref>; Archibald and Bond, <xref rid="B6" ref-type="bibr">2003</xref>; Mitsopoulos and Dimitrakopoulos, <xref rid="B67" ref-type="bibr">2007</xref>; Davies et al., <xref rid="B3new" ref-type="bibr">2009</xref>; Ganteaume et al., <xref rid="B39" ref-type="bibr">2009</xref>, <xref rid="B113" ref-type="bibr">2011</xref>; Berry et al., <xref rid="B10" ref-type="bibr">2011</xref>; De Magalh&#x000e3;es and Schwilk, <xref rid="B30" ref-type="bibr">2012</xref>; Trauernicht et al., <xref rid="B8new" ref-type="bibr">2012</xref>; Van Altena et al., <xref rid="B106" ref-type="bibr">2012</xref>; Castagneri et al., <xref rid="B2new" ref-type="bibr">2013</xref>; Davies and Nafus, <xref rid="B27" ref-type="bibr">2013</xref></td><td valign="top" align="left" rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1"/></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">Canopy cover</td><td valign="top" align="left" rowspan="1" colspan="1">Strong</td><td valign="top" align="left" rowspan="1" colspan="1">Ray et al., <xref rid="B82" ref-type="bibr">2005</xref>; Peterson and Reich, <xref rid="B80" ref-type="bibr">2008</xref>; Warman and Moles, <xref rid="B9new" ref-type="bibr">2009</xref>; Hoffmann et al., <xref rid="B46" ref-type="bibr">2012</xref>; Little et al., <xref rid="B62" ref-type="bibr">2012</xref>; Murphy and Bowman, <xref rid="B72" ref-type="bibr">2012</xref>; Trauernicht et al., <xref rid="B8new" ref-type="bibr">2012</xref></td><td valign="top" align="left" rowspan="1" colspan="1">No</td><td rowspan="1" colspan="1"/></tr></tbody></table><table-wrap-foot><p><italic>Evidence for an effect on flammability is a necessary but not sufficient condition for demonstrating selection for flammability</italic>.</p></table-wrap-foot></table-wrap><fig id="F1" position="float"><label>Figure 1</label><caption><p><bold>Diverse plant traits that affect vegetation flammability. (A)</bold> Surface fire in Amazonian rainforest leaf litter and ground cover vegetation during a severe drought, when leaf moisture context of living and dead foliage was very low (Photo: Mark Cochrane); <bold>(B)</bold> Koala (<italic>Phascolarctos cinereus</italic>), an iconic specialist mammalian herbivore involved in a co-evolutionary relationship with eucalypt leaf secondary chemical defenses. These defenses also make foliage exceptionally flammable (Photo Kath Handasyde); <bold>(C)</bold> New Caledonian maquis vegetation, which is dominated by sclerophyll species with phylogenetic links to Australian flammable heathland, yet has a poor capacity to recover from fire (Photo David Bowman); <bold>(D)</bold> leaf retention of <italic>Richea pandanifolius</italic>, a fire sensitive Gondwana rainforest giant heath, demonstrates that this trait is not universally associated with increasing flammability (Photo David Bowman); <bold>(E)</bold> low bulk density annual grass layer in eucalypt savanna is exceptionally flammable (Photo Don Franklin); <bold>(F)</bold> post-flowering die-off of the giant bamboo <italic>Bambusa arnhemica</italic> in frequently burnt eucalypt savanna. The dead bamboo is much less flammable than the grass layer in surrounding savanna (photo Don Franklin); <bold>(G)</bold> decorticating bark on a SE Asian tropical rainforest tree <italic>Cratoxylum cochinchinense</italic> demonstrates that this trait is not necessarily related to spreading fires via fire brands (Photo David Tng); <bold>(H)</bold> abrupt rain forest boundary in north Queensland which limits the spread of savanna fires, as evidenced by the shrubs burnt in the preceding dry season (Photo David Bowman).</p></caption><graphic xlink:href="fpls-05-00590-g0001"/></fig></sec><sec><title>Organic chemistry</title><p>Foliar organic chemistry has a secondary effect on flammability after LMC (Alessio et al., <xref rid="B1" ref-type="bibr">2008a</xref>,<xref rid="B2" ref-type="bibr">b</xref>; De Lillis et al., <xref rid="B29" ref-type="bibr">2009</xref>; Page et al., <xref rid="B78" ref-type="bibr">2012</xref>) (Table <xref ref-type="table" rid="T2">2</xref>). For example, volatile organic compounds (VOCs such as terpenes and phenolics) can reduce ignition temperatures of living and dead leaves (Owens et al., <xref rid="B77" ref-type="bibr">1998</xref>; Ormeno et al., <xref rid="B76" ref-type="bibr">2009</xref>). However, VOCs also play an important role in herbivore defense (Owens et al., <xref rid="B77" ref-type="bibr">1998</xref>; Page et al., <xref rid="B78" ref-type="bibr">2012</xref>; Loreto et al., <xref rid="B63" ref-type="bibr">2014</xref>), confounding their attribution as flammability adaptations (Dickinson and Kirkpatrick, <xref rid="B31" ref-type="bibr">1985</xref>; Kerr et al., <xref rid="B55" ref-type="bibr">1999</xref>; Schwilk and Kerr, <xref rid="B94" ref-type="bibr">2002</xref>; Holmes, <xref rid="B47" ref-type="bibr">2009</xref>). For example, variation in leaf terpenes of eucalypts, a notoriously flammable group of plants, is known to serve a wide variety of functions including influencing insect and mammalian herbivory and attracting insect pollinators, and has knock-on effects on decomposition and nutrient cycling (Keszei et al., <xref rid="B56" ref-type="bibr">2008</xref>). Indeed, there is evidence of co-evolution between the diversification of plant secondary compounds and the intensity of special mammalian herbivores on eucalypt foliage (Moore et al., <xref rid="B69" ref-type="bibr">2005</xref>) (Figure <xref ref-type="fig" rid="F1">1B</xref>).</p></sec><sec><title>Inorganic chemistry</title><p>Leaves of flammable sclerophylls, which typically occur on infertile soils, have high foliar silica contents and low concentrations of other nutrients, especially phosphorus and nitrogen, compared to non-sclerophyll leaves (Turner, <xref rid="B105" ref-type="bibr">1994</xref>). However, sclerophyllous foliage is imperfectly correlated with fire adapted vegetation (Midgley, <xref rid="B65" ref-type="bibr">2013</xref>). The maquis shrublands of New Caledonia, for example, are dominated by sclerophyllous species, of which only about 19% persist through fires (McCoy et al., <xref rid="B64" ref-type="bibr">1999</xref>) (Figure <xref ref-type="fig" rid="F1">1C</xref>), despite close phylogenetic links to fire-tolerant Australian heathland species. In principle, high phosphate concentrations in foliage could inhibit combustion given that phosphate is commonly used in fire retardants, yet little support has been found for this hypothesis (Scarff and Westoby, <xref rid="B90" ref-type="bibr">2006</xref>; Scarff et al., <xref rid="B89" ref-type="bibr">2012</xref>).</p></sec><sec><title>Leaf dimensions</title><p>Leaf dimensions (size, thickness, and shape) influence the flammability of individual leaves. Thinner leaves, which have a high surface area to volume ratio and high specific leaf area, and larger leaves, appear to be more ignitable (Gill and Moore, <xref rid="B41" ref-type="bibr">1996</xref>; Saura-Mas et al., <xref rid="B88" ref-type="bibr">2010</xref>; Murray et al., <xref rid="B73" ref-type="bibr">2013</xref>). However, species with small leaves tend to have narrow, frequently branched twigs and dense wood, which burn more intensely (Westoby and Wright, <xref rid="B108" ref-type="bibr">2003</xref>; Pickup et al., <xref rid="B81" ref-type="bibr">2005</xref>), potentially counteracting the lower flammability of small individual leaves. While flammability of live individual leaves may influence the spread of crown fires, surface fires are more strongly influenced by the flammability of litter beds. Large, long leaves may produce more flammable litter fuels because of lower packing density, which influences oxygen availability (Scarff and Westoby, <xref rid="B90" ref-type="bibr">2006</xref>; Belcher et al., <xref rid="B9" ref-type="bibr">2010</xref>; De Magalh&#x000e3;es and Schwilk, <xref rid="B30" ref-type="bibr">2012</xref>). For instance, an American study has found a link between abundance in litter fuels of <italic>Pinus</italic> species, which have long needle-shaped leaves, and fire severity (Schwilk and Caprio, <xref rid="B93" ref-type="bibr">2011</xref>). Importantly, individual species have non-additive effects on litter flammability, which tends to be driven by the most flammable leaves in the litter (De Magalh&#x000e3;es and Schwilk, <xref rid="B30" ref-type="bibr">2012</xref>; Van Altena et al., <xref rid="B106" ref-type="bibr">2012</xref>).</p><sec><title>Dead leaf retention</title><p>When leaves die they are typically shed, although some plants retain dead leaves for extended periods; these dead leaves have low LMC relative to live foliage (Page et al., <xref rid="B78" ref-type="bibr">2012</xref>). It has been suggested that dead leaf retention is an adaptation to promote plant flammability (He et al., <xref rid="B45" ref-type="bibr">2011</xref>) and community flammability (Santana et al., <xref rid="B87" ref-type="bibr">2011</xref>). He et al. (<xref rid="B45" ref-type="bibr">2011</xref>) used dated phylogenies to show that dead leaf retention in the Australian genus <italic>Banksia</italic> arose after the appearance of serotiny, suggesting that dead leaf retention could have evolved to increase the probability of fire and ensure that seeds are liberated. However, retention of dead foliage is not restricted to plants that occur in flammable environments: an example is the fire sensitive endemic Tasmanian rainforest arborescent monocot <italic>Richea pandanifolia</italic> (Figure <xref ref-type="fig" rid="F1">1D</xref>), signaling that this trait is not universally related to flammability. Indeed, it has been suggested that the retention of dead foliage in tall grasses is an adaptation to reduce the intensity of mammalian herbivory, but which may have also increased landscape fire activity (Mingo and Oesterheld, <xref rid="B66" ref-type="bibr">2009</xref>; Antonelli et al., <xref rid="B5" ref-type="bibr">2011</xref>).</p></sec></sec><sec><title>Phenology</title><p>In seasonally dry environments, phenology influences flammability by causing seasonal patterns in production and senescence of both leaves (deciduous plants) and of whole plants (annuals) (Keeley and Bond, <xref rid="B52" ref-type="bibr">1999</xref>; Elliott et al., <xref rid="B33" ref-type="bibr">2009</xref>; Bajocco et al., <xref rid="B8" ref-type="bibr">2010</xref>; Ripley et al., <xref rid="B83" ref-type="bibr">2010</xref>; De Angelis et al., <xref rid="B28" ref-type="bibr">2012</xref>; Davies and Nafus, <xref rid="B27" ref-type="bibr">2013</xref>). Obvious examples are senescence of annual herbs and grasses, leading to increased community flammability in the non-growing season because of high fine fuel loads (Wittich, <xref rid="B111" ref-type="bibr">2011</xref>) (Figure <xref ref-type="fig" rid="F1">1E</xref>), as well as the dry season combustion of leaf litter in tropical dry forests (Mondal and Sukumar, <xref rid="B68" ref-type="bibr">2014</xref>). This seasonal surge in available fuel has not been attributed to evolution, although Keeley and Bond (<xref rid="B52" ref-type="bibr">1999</xref>) hypothesized that synchronized mass flowering and die-off of bamboos is an evolutionary strategy to generate a &#x0201c;synchronous fuel load that significantly increases the potential for wildfire disturbance.&#x0201d; However, there is little evidence that fire is a key feature in the evolution of bamboo life-history (Saha and Howe, <xref rid="B86" ref-type="bibr">2001</xref>). Franklin and Bowman (<xref rid="B37" ref-type="bibr">2003</xref>) found no support for this hypothesis from the north Australian giant bamboo, <italic>Bambusa arnhemica</italic>, which grows in an environment where fire is extremely frequent. The seedlings of this species did not require fire to establish, and dead adult biomass had low flammability (Franklin and Bowman, <xref rid="B37" ref-type="bibr">2003</xref>) (Figure <xref ref-type="fig" rid="F1">1F</xref>).</p></sec><sec><title>Decorticating bark</title><p>Lofted pieces of burning fuel (termed firebrands) can create spot fires ahead of a fire-front and are a key mechanism promoting fire spread (Koo et al., <xref rid="B57" ref-type="bibr">2010</xref>). Decorticating eucalypt bark has been hypothesized to evolve to spread fires (Jackson, <xref rid="B48" ref-type="bibr">1968</xref>; Mount, <xref rid="B71" ref-type="bibr">1979</xref>). However, the individual fitness benefits of this trait are not obvious (Ellis, <xref rid="B35" ref-type="bibr">1965</xref>). In any case decorticating bark also occurs in non-flammable environments (Figure <xref ref-type="fig" rid="F1">1G</xref>), and has been suggested as defending against epiphyte infestation (Carsten et al., <xref rid="B20" ref-type="bibr">2002</xref>; Wyse and Burns, <xref rid="B112" ref-type="bibr">2011</xref>).</p></sec><sec><title>Self-pruning and branch retention</title><p>Shedding of dead lower branches reduces continuity between surface fuels and the canopy. Conversely, retained dead branches create fuel ladders and allow fires to reach the crown of individual trees and their neighbors (Schwilk, <xref rid="B91" ref-type="bibr">2003</xref>; Keeley, <xref rid="B51" ref-type="bibr">2012</xref>; Seo and Choung, <xref rid="B96" ref-type="bibr">2014</xref>). Phylogenetic analysis shows that shedding of branches may have evolved in the genus <italic>Pinus</italic> to reduce crown fires (He et al., <xref rid="B44" ref-type="bibr">2012</xref>), in contrast to the ancestral condition of branch retention that promotes crown fires. The latter is often associated with serotiny (Gauthier et al., <xref rid="B40" ref-type="bibr">1996</xref>; Schwilk and Ackerly, <xref rid="B92" ref-type="bibr">2001</xref>; Ne'eman et al., <xref rid="B75" ref-type="bibr">2004</xref>), a derived trait that apparently offered an alternative strategy to deal with high fire activity during the Cretaceous (He et al., <xref rid="B44" ref-type="bibr">2012</xref>).</p></sec><sec><title>Plant architecture and canopy morphology</title><p>Plant architecture may also influence flammability. For instance, frequent fire on the New Jersey Pine Plains has selectively maintained a dwarf, crooked form of <italic>Pinus rigida</italic> which is more flammable than the surrounding tall forest (Ledig et al., <xref rid="B59" ref-type="bibr">2013</xref>). In some Mediterranean environments, plants with fire-dependent seeding strategy have open crowns with fine leaves that promote flammability (Saura-Mas et al., <xref rid="B88" ref-type="bibr">2010</xref>), although this crown morphology also occurs in environments where fire is not central to plant regeneration, such as South American shrublands with similar climates (Keeley et al., <xref rid="B53" ref-type="bibr">2011a</xref>). Shading by dense canopies of individual trees influences understory floristics and local microclimate (Peterson and Reich, <xref rid="B80" ref-type="bibr">2008</xref>; Cohn et al., <xref rid="B24" ref-type="bibr">2011</xref>), thereby affecting fire regime. For example, closed crowned trees can suppress grasses in savannas (Hoffmann et al., <xref rid="B46" ref-type="bibr">2012</xref>) (Figure <xref ref-type="fig" rid="F1">1H</xref>).</p></sec></sec><sec sec-type="discussion" id="s2"><title>Discussion</title><p>Our review has not been able to identify any individual plant traits attributes that exclusively influence flammability (Table <xref ref-type="table" rid="T1">1</xref>). Further, we show that plant traits that increase flammability may exist in plant communities that are rarely burnt, suggesting they have evolved independently of landscape fire. It is probable that some traits related to flammability, such as foliar chemistry, may be &#x0201c;exaptations&#x0201d; (Gould and Vrba, <xref rid="B43" ref-type="bibr">1982</xref>)&#x02014;traits with another function that incidentally increases flammability (Trabaud, <xref rid="B103" ref-type="bibr">1976</xref>; Snyder, <xref rid="B99" ref-type="bibr">1984</xref>; Bradshaw et al., <xref rid="B16" ref-type="bibr">2011a</xref>). Such micro-evolutionary processes are apparent in the selection of more flammable genotypes of <italic>Ulex parviflorus</italic> (Mediterranean gorse) (Pausas and Moreira, <xref rid="B79" ref-type="bibr">2012</xref>; Moreira et al., <xref rid="B70" ref-type="bibr">2014</xref>). The benefit of increased flammability for plants that require fire disturbance to regenerate is possibly greatest in environments where background fire frequency is low, for example in tall eucalypt forests compared to tropical eucalypt savannas (Bowman and Wilson, <xref rid="B15" ref-type="bibr">1988</xref>; Murphy and Bowman, <xref rid="B72" ref-type="bibr">2012</xref>). Increased flammability may also be of selective benefit for plants that recover following fire disturbance, thereby deflecting successional pathways from less flammable mature forests. For example, such a seral &#x0201c;niche construction&#x0201d; model has been proposed to explain the dynamics of eucalypt forests and rainforests in high rainfall areas of Australia (Jackson, <xref rid="B48" ref-type="bibr">1968</xref>; Bowman, <xref rid="B12" ref-type="bibr">2000</xref>). The eucalypt forests require fire to regenerate so that unless fire occurs within their life span the eucalypts are replaced by comparatively fire sensitive, continuously regenerating rainforest species (Tng et al., <xref rid="B102" ref-type="bibr">2012</xref>). Clarke et al. (<xref rid="B22" ref-type="bibr">2014</xref>) tested this hypothesis and found that foliage and litter from eucalypt forest was not more flammable that from rainforest. Further, eucalypt forests regenerating after severe fire did not have more flammable litter compared to areas affected by less severe fire or long unburnt, so there was no evidence that fire selected for higher litter flammability. Likewise, Lindenmayer et al. (<xref rid="B60" ref-type="bibr">2011</xref>) have suggested that stands of <italic>Eucalyptus regnans</italic> regenerating following disturbance are inherently more flammable than long unburnt stands, yet a recent analysis shows this effect was not evident in stands burnt within the last 7 years, and was most pronounced in stands burnt around 15 years ago (Taylor et al., <xref rid="B100" ref-type="bibr">2014</xref>), discounting the influence of short-lived herbaceous fire weeds that characterize the post-fire plant community (Jackson, <xref rid="B48" ref-type="bibr">1968</xref>).</p><p>It is important to acknowledge that traits that influence plant combustion are not exclusively associated with flammability. This complicates macro-evolutionary ancestral state reconstructions by demanding joint consideration of the evolution of fire tolerating traits and recovery mechanisms with flammable traits. Mutch (<xref rid="B74" ref-type="bibr">1970</xref>) suggested that fire promoting traits followed the development of fire tolerating and recovery mechanisms, but it is possible that inherently flammable plants drove the evolution of plant recovery mechanisms&#x02014;an evolutionary pathway known as &#x0201c;Mutch's converse&#x0201d; (Kerr et al., <xref rid="B55" ref-type="bibr">1999</xref>; Schwilk and Ackerly, <xref rid="B92" ref-type="bibr">2001</xref>; Schwilk and Kerr, <xref rid="B94" ref-type="bibr">2002</xref>). The analysis of serotiny in <italic>Banksia</italic>, and self-pruning, bark thickness and serotiny in <italic>Pinus</italic> (e.g., He et al., <xref rid="B45" ref-type="bibr">2011</xref>, <xref rid="B44" ref-type="bibr">2012</xref>) suggest the latter, but many more ancestral trait reconstructions are required before generalizations can be drawn about the most typical evolutionary pathways, and how these patterns vary biogeographically. A confounding factor in such reconstruction is that plants that evolve traits to tolerate or recover from fire may be under less selection pressure to reduce their flammability, leading to positive correlations between flammability and fire tolerance without evolutionary selection for high flammability. Importantly, Midgley (<xref rid="B65" ref-type="bibr">2013</xref>) points out that selection for non-flammable traits, such as branch shedding, avoids many of the problems with the Mutch hypothesis, given the manifest individual fitness benefits of avoiding self-immolation. More research needs to be directed to this hypothesis, which we call &#x0201c;Midgley's alternative.&#x0201d;</p><p>The focus on flammability traits of individuals in both theoretical models and ancestral trait reconstructions obscures the fact that wildfire propagates through vegetation made up of multiple species, so the most appropriate unit of analysis should be the plant community. Community flammability is controlled by the interplay of climate with vegetation canopy cover, fuel continuity and litter bed characteristics (Table <xref ref-type="table" rid="T1">1</xref>). This is well illustrated by boundaries between vegetation types with sharply contrasting flammability, such as savanna and tropical rainforests: forests which have closed canopies result in microclimates characterized by higher humidity, lower wind velocities, cooler temperatures, reduced evaporation and hence reduced fire risk compared to open-canopied savannas (Bowman and Wilson, <xref rid="B15" ref-type="bibr">1988</xref>; Ray et al., <xref rid="B82" ref-type="bibr">2005</xref>; Hoffmann et al., <xref rid="B46" ref-type="bibr">2012</xref>; Little et al., <xref rid="B62" ref-type="bibr">2012</xref>; Veldman et al., <xref rid="B107" ref-type="bibr">2013</xref>). Litter beds are an emergent property of the plant community because the mix of dead foliage with different sizes and shapes affects fuel bulk density, which in turn influences flammability (Scarff and Westoby, <xref rid="B90" ref-type="bibr">2006</xref>; Kane et al., <xref rid="B50" ref-type="bibr">2008</xref>; Schwilk and Caprio, <xref rid="B93" ref-type="bibr">2011</xref>; De Magalh&#x000e3;es and Schwilk, <xref rid="B30" ref-type="bibr">2012</xref>; Engber and Varner III, <xref rid="B36" ref-type="bibr">2012</xref>; Van Altena et al., <xref rid="B106" ref-type="bibr">2012</xref>; Murray et al., <xref rid="B73" ref-type="bibr">2013</xref>; McGlone et al., <xref rid="B64a" ref-type="bibr">2014</xref>) (Figure <xref ref-type="fig" rid="F1">1H</xref>). Large, thin leaves and leaves with complex shapes (such as compound leaves or leaves with lobed margins) result in well aerated litter beds that typically dry out quickly and readily combust during dry periods (Scarff and Westoby, <xref rid="B90" ref-type="bibr">2006</xref>; Kane et al., <xref rid="B50" ref-type="bibr">2008</xref>; Schwilk and Caprio, <xref rid="B93" ref-type="bibr">2011</xref>; De Magalh&#x000e3;es and Schwilk, <xref rid="B30" ref-type="bibr">2012</xref>; Engber and Varner III, <xref rid="B36" ref-type="bibr">2012</xref>). The most extreme examples of this effect are tall tropical grasses, which produce highly combustible fuel beds, in contrast to denser leaf litter fuels: the difference in these fuel types reinforces forest-savanna boundaries (Hoffmann et al., <xref rid="B46" ref-type="bibr">2012</xref>) (Figure <xref ref-type="fig" rid="F1">1H</xref>).</p><p>The stark differences in flammability of grasses and broadleaved fuels also invites consideration of the flammability traits amongst Poaceae lineages. Some grass genera have high flammability due to massive accumulation of fine, well-aerated fuels (e.g., <italic>Andropogon</italic>) (Setterfield et al., <xref rid="B97" ref-type="bibr">2010</xref>), &#x0201c;haying-off&#x0201d; after the growing season (e.g., annual <italic>Sorghum</italic>) (Elliott et al., <xref rid="B33" ref-type="bibr">2009</xref>), retention of dead foliage, or resin-rich leaves [e.g., <italic>Triodia</italic> (Allan and Southgate, <xref rid="B4" ref-type="bibr">2002</xref>)]. Indeed, globally, many C4 savanna grasslands are maintained by fire (Scott et al., <xref rid="B95" ref-type="bibr">2014</xref>). However, some other grasses are less flammable than surrounding vegetation, for example dense swards of Australian alpine <italic>Poa</italic> compared to surrounding heathlands (Williams et al., <xref rid="B110" ref-type="bibr">2006</xref>). While invasive grasses can drive a grass fire cycle (D'Antonio and Vitousek, <xref rid="B26" ref-type="bibr">1992</xref>; Setterfield et al., <xref rid="B97" ref-type="bibr">2010</xref>), it is important to note that in many situations this feedback loop is driven by high anthropogenic ignitions and an absence of co-evolved grazers. More investigation of the flammable traits of grasses, and their evolutionary pathways, including co-evolutionary relationships with grazers (e.g., Linder and Rudall, <xref rid="B61" ref-type="bibr">2005</xref>; Antonelli et al., <xref rid="B5" ref-type="bibr">2011</xref>; McGlone et al., <xref rid="B64a" ref-type="bibr">2014</xref>) are warranted.</p><p>Clarke et al. (<xref rid="B22" ref-type="bibr">2014</xref>) used a mosaic of flammable eucalypt forest and far less flammable rainforest as an evolutionary &#x0201c;model system&#x0201d; to show there were no differences in the flammability of foliage of congeners in these contrasting forest types. They also found no differences in the flammability of litter fuels dried to a standard moisture content. This led them to reject the Mutch hypothesis that individual plant flammability is under natural selection; rather, they concluded that community flammability differences were related to the contrasting microclimates under the open eucalypt and the dense rainforest canopies. It is important to note that low flammability rainforest can establish beneath canopies of mature eucalypt forests growing in moist environments, blunting the view that eucalypt canopy openness is a specific adaptation to increase flammability (Tng et al., <xref rid="B102" ref-type="bibr">2012</xref>).</p><p>Keeley et al. (<xref rid="B54" ref-type="bibr">2011b</xref>) argue that the most profitable route to disclosing the evolutionary relationships between plants and landscape fire is to understand the nexus between fire regimes and plant traits. However, we suspect fire regimes are too fluid to provide a sufficiently strong evolutionary pressure to select for highly flammable traits. Fire regimes respond rapidly to changing patterns of ignitions, intensity and type of herbivory, new species of invasive plants and longer term climate changes. For example, the loss of Pleistocene megafauna in both North America (Gill et al., <xref rid="B42" ref-type="bibr">2009</xref>) and Australia (Rule et al., <xref rid="B85" ref-type="bibr">2012</xref>) appeared to change fire regimes due to the proliferation of woody biomass, which fuelled more intense fires. Likewise, invasive species can abruptly change flammability by altering vertical or horizontal fuel continuity, and hence facilitate the spread of fires into canopies or amongst otherwise spatially isolated plants. This is well illustrated by the invasion of dry rainforests in Queensland by the woody shrub <italic>Lantana camara</italic>, which changes fire type from surface litter fires to shrub canopy fires that can kill rainforest trees, or invasive <italic>Bromis tectorum</italic>, which changes horizontal fuel continuity, causing loss of succulents such as giant saguaro cacti (<italic>Carnegiea gigantea</italic>) (Thomas and Goodson, <xref rid="B101" ref-type="bibr">1992</xref>). Such shifting patterns of fire activity filtering numerous plant traits from multiple species make it difficult to sustain the notion that numerous species in communities have all evolved to collectively self-immolate. It is more parsimonious to view fire activity as a powerful filter that sorts plants with pre-existing flammabilities and hones regeneration strategies.</p></sec><sec><title>Author contributions</title><p>David Bowman conceived the ideas for the manuscript, and Ben French carried out the initial literature review. All authors contributed to the writing.</p><sec><title>Conflict of interest statement</title><p>The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.</p></sec></sec>

Heterologous over-expression of <italic>ACC SYNTHASE8 (ACS8)</italic> in <italic>Populus tremula x P. alba</italic> clone 717-1B4 results in elevated levels of ethylene and induces stem dwarfism and reduced leaf size through separate genetic pathways

<p>Plant height is an important agronomic and horticultural trait that impacts plant productivity, durability and esthetic appeal. A number of the plant hormones such as gibberellic acid (GA), auxin and ethylene have been linked to control of plant architecture and size. Reduction in GA synthesis and auxin transport result in dwarfism while ethylene may have a permissive or repressive effect on tissue growth depending upon the age of plant tissues or the environmental conditions considered. We describe here an activation-tagged mutant of <italic>Populus tremula x P. alba</italic> clone 717-1B4 identified from 2000 independent transgenic lines due to its significantly reduced growth rate and smaller leaf size. Named <italic>dwarfy</italic>, the phenotype is due to increased expression of <italic>PtaACC SYNTHASE8</italic>, which codes for an enzyme in the first committed step in the biosynthesis of ethylene. Stems of <italic>dwarfy</italic> contain fiber and vessel elements that are reduced in length while leaves contain fewer cells. These morphological differences are linked to <italic>PtaACS8</italic> inducing different transcriptomic programs in the stem and leaf, with genes related to auxin diffusion and sensing being repressed in the stem and genes related to cell division found to be repressed in the leaves. Altogether, our study gives mechanistic insight into the genetics underpinning ethylene-induced dwarfism in a perennial model organism.</p>

<contrib contrib-type="author"><name><surname>Plett</surname><given-names>Jonathan M.</given-names></name><xref ref-type="aff" rid="aff1">¹</xref><xref ref-type="aff" rid="aff2">²</xref><uri xlink:type="simple" xlink:href="http://community.frontiersin.org/people/u/98537"/></contrib><contrib contrib-type="author"><name><surname>Williams</surname><given-names>Martin</given-names></name><xref ref-type="aff" rid="aff3">³</xref></contrib><contrib contrib-type="author"><name><surname>LeClair</surname><given-names>Gaetan</given-names></name><xref ref-type="aff" rid="aff3">³</xref></contrib><contrib contrib-type="author"><name><surname>Regan</surname><given-names>Sharon</given-names></name><xref ref-type="aff" rid="aff1">¹</xref><xref ref-type="author-notes" rid="fn003">^&#x02020;</xref><uri xlink:type="simple" xlink:href="http://community.frontiersin.org/people/u/190732"/></contrib><contrib contrib-type="author"><name><surname>Beardmore</surname><given-names>Tannis</given-names></name><xref ref-type="aff" rid="aff3">³</xref><xref ref-type="author-notes" rid="fn001">^*</xref><xref ref-type="author-notes" rid="fn003">^&#x02020;</xref></contrib>

Frontiers in Plant Science

<sec sec-type="introduction" id="s1"><title>Introduction</title><p>Reduced plant height, or dwarfism, is an important agronomic trait linked to higher yields (Huang et al., <xref rid="B37" ref-type="bibr">1996</xref>; Yang and Hwa, <xref rid="B90" ref-type="bibr">2008</xref>), easier harvesting (Adkins et al., <xref rid="B1" ref-type="bibr">2010</xref>) and reduced nutrient demand on soils (Sieling and Kage, <xref rid="B71" ref-type="bibr">2008</xref>). Leaf size, meanwhile, is linked to productivity, predation (Faeth, <xref rid="B22" ref-type="bibr">1991</xref>) and the water status of the plant (Scoffoni et al., <xref rid="B70" ref-type="bibr">2011</xref>). While both height and leaf size are complex traits, they appear to be genetically regulated by a similar panel of plant hormones (Valdovinos et al., <xref rid="B83" ref-type="bibr">1967</xref>; Ephritikhine et al., <xref rid="B21" ref-type="bibr">1999</xref>; Qi et al., <xref rid="B62" ref-type="bibr">2011</xref>; Luo et al., <xref rid="B52" ref-type="bibr">2013</xref>) and cytochrome P450s (Zhang et al., <xref rid="B93" ref-type="bibr">2014</xref>), as well as abiotic factors such as temperature (Yang et al., <xref rid="B91" ref-type="bibr">2014</xref>) and photoperiod (Li et al., <xref rid="B49" ref-type="bibr">2014</xref>). Reductions in organ size are a result of two different physiological phenomena: smaller cells and impeded cellular division (Beemster et al., <xref rid="B5" ref-type="bibr">2003</xref>). These two factors may work independently or synergistically to affect plant stature and organ size (Beemster et al., <xref rid="B4" ref-type="bibr">2005</xref>; Skirycz et al., <xref rid="B72" ref-type="bibr">2010</xref>). Newly produced plant tissues first exhibit growth due to rapid cellular division, a phase that is replaced in a distal-proximal manner by cellular expansion in progenitor cells (Donnelly et al., <xref rid="B17" ref-type="bibr">1999</xref>). Due to the integrated control between these two processes, genetic mutations to single genes can have a drastic impact on plant stature as a whole or at the level of a specific tissue. Altered expression of genes such as <italic>ARABIDOPSIS VACUOLAR H+-PYROPHOSPHATASE1</italic> (<italic>AVP1</italic>; Li et al., <xref rid="B48" ref-type="bibr">2005</xref>), <italic>CYTOKININ RESISTANT1</italic> (<italic>CNR1</italic>; Guo et al., <xref rid="B28" ref-type="bibr">2010</xref>), and <italic>ISOPENTENYL TRANSFERASE3</italic> (<italic>IPT3</italic>; Nobusawa et al., <xref rid="B58" ref-type="bibr">2013</xref>) impact tissue size due to a difference in the total number of cells produced, while <italic>EXPANSIN10</italic> (<italic>EXP10</italic>; Cho and Cosgrove, <xref rid="B12" ref-type="bibr">2000</xref>), <italic>ARGOS-LIKE</italic> (Hu et al., <xref rid="B35" ref-type="bibr">2006</xref>), and <italic>RETINOBLASTOMA-RELATED PROTEIN1</italic> (<italic>RBR1</italic>; Sabelli et al., <xref rid="B68" ref-type="bibr">2013</xref>) change the final size of plant tissues as a function of altered cell expansion.</p><p>The best studied genetic influences on dwarfism are genes and signaling pathways related to hormone production and sensitivity. Within these studies, ethylene, gibberellic acid (GA), auxin, and brassinosteroids (BR) have all been implicated with a role in cell division, cellular growth and overall plant architecture. Blocked BR synthesis (Nakaya et al., <xref rid="B56" ref-type="bibr">2002</xref>) and reduced GA biosynthesis (Tong et al., <xref rid="B79" ref-type="bibr">2007</xref>; Li et al., <xref rid="B47" ref-type="bibr">2011</xref>) or increased GA catabolism (Busov et al., <xref rid="B9" ref-type="bibr">2003</xref>; Schomburg et al., <xref rid="B69" ref-type="bibr">2003</xref>; Curtis et al., <xref rid="B13" ref-type="bibr">2005</xref>; Lee and Zeevaart, <xref rid="B45" ref-type="bibr">2005</xref>; Dijkstra et al., <xref rid="B15" ref-type="bibr">2008</xref>; Zawaski et al., <xref rid="B92" ref-type="bibr">2011</xref>) induce dwarfism in a wide range of model plant systems. Auxin transport, meanwhile, is a critical component of proper plant stem elongation. In rice, auxin transport inhibition has been correlated to slower stem elongation (Yamamoto et al., <xref rid="B89" ref-type="bibr">2007</xref>; Domingo et al., <xref rid="B16" ref-type="bibr">2009</xref>) while reduced basipetal auxin transport in maize and <italic>Arabidopsis thaliana</italic> results in stunted plant development (Lantican and Muir, <xref rid="B44" ref-type="bibr">1969</xref>; Geisler et al., <xref rid="B25" ref-type="bibr">2003</xref>, <xref rid="B24" ref-type="bibr">2005</xref>; Multani et al., <xref rid="B55" ref-type="bibr">2003</xref>; Geisler and Murphy, <xref rid="B26" ref-type="bibr">2006</xref>). Treatment of plant tissues with ethylene, a gaseous plant hormone, results in stunting (Vahala et al., <xref rid="B82" ref-type="bibr">2013</xref>), a phenotype that has been linked to the induced expression of certain <italic>ETHYLENE RESPONSE FACTORs</italic> (<italic>ERFs</italic>; Dubois et al., <xref rid="B18" ref-type="bibr">2013</xref>; Vahala et al., <xref rid="B82" ref-type="bibr">2013</xref>). There also appears to be extensive cross-talk between the different hormone pathways with components of the ethylene pathway controlling GA biosynthesis (Qi et al., <xref rid="B62" ref-type="bibr">2011</xref>) and the activity of DELLA proteins (Luo et al., <xref rid="B52" ref-type="bibr">2013</xref>). Ethylene can also regulate auxin diffusion and biosynthesis (Valdovinos et al., <xref rid="B83" ref-type="bibr">1967</xref>; Stepanova et al., <xref rid="B74" ref-type="bibr">2005</xref>; Ruzicka et al., <xref rid="B67" ref-type="bibr">2007</xref>; Swarup et al., <xref rid="B78" ref-type="bibr">2007</xref>).</p><p>Here we characterize an activation tagged mutant of <italic>Populus tremula x P. alba</italic> clone 717, named &#x0201c;<italic>dwarfy</italic>,&#x0201d; exhibiting severe dwarfism with both reduced stature and smaller leaves. We show that the gene responsible for this phenotype is annotated as the poplar 1-aminocyclopropane-1-carboxylate synthase (ACS) gene <italic>PtaACS8</italic>. Ethylene is synthesized in two enzymatic steps from the substrate S-adenosyl-methionine (SAM). The first step is the conversion of SAM into 1-aminocyclopropane-1-carboxylic acid (ACC) by the activity of ACSs followed by the conversion of ACC to ethylene catalyzed by ACC OXIDASEs (ACOs). Ethylene is then perceived by a family of membrane bound receptors that induce the transcription of ETHYLENE RESPONSE FACTORs (ERFs) which, in turn, controls transcription and, ultimately, plant development. We demonstrate that increased expression of <italic>PtaACS8</italic> in the <italic>dwarfy</italic> line results in significantly higher levels of ethylene in all aerial tissues of the plant. Morphologically, the increased expression of <italic>PtaACS8</italic> in the stem results in shorter vessels and fibers in secondary growth while endogenous over-expression of the <italic>PtaACS8</italic> gene in the leaves results in the production of fewer cells. The reduced growth of stem cells is accompanied by a repression of auxin transport and signaling genes while reduction in cell number in leaves is concurrent with a large reduction in the transcript abundance of a number of cell-cycle genes. Therefore, we conclude that increased expression of <italic>PtaACS8</italic> induces stem dwarfism and reduced leaf size through separate genetic pathways.</p></sec><sec sec-type="materials|methods" id="s2"><title>Materials and methods</title><sec><title>Plant material</title><p>All plants used in Figures <xref ref-type="fig" rid="F1">1</xref>, <bold>4</bold>, <bold>6</bold> were grown under greenhouse conditions at the Canadian Forest Service (CFS), Fredericton, New Brunswick, Canada, while plants used for data analysis in Figures <xref ref-type="fig" rid="F2">2</xref>, <xref ref-type="fig" rid="F3">3</xref>, <bold>5</bold> were grown under greenhouse conditions at Queen's University, Kingston, Ontario, Canada. In the former situation, plants were grown under natural daylight and temperature, while in the latter situation, photoperiod was maintained at 16 h per day and temperature at 25&#x000b0;C. The <italic>dwarfy</italic> mutant was generated as described by Harrison et al. (<xref rid="B30" ref-type="bibr">2007</xref>) in a <italic>P. tremula x P. alba</italic> clone 717-1B4 background and all comparisons of <italic>dwarfy</italic> were made with this hybrid (wildtype). The <italic>dwarfy</italic> mutant was initially identified based on the dwarf characteristics such as plant height and leaf size among others in the mutant.</p><fig id="F1" position="float"><label>Figure 1</label><caption><p><bold>Endogenous over-expression of <italic>PtaACS8</italic> results in a dwarfed growth phenotype in <italic>P. tremula x P. alba</italic> clone 717-1B4. (A)</bold> A representative image of the <italic>dwarfy</italic> mutant <italic>P. tremula x P. alba</italic> clone as compared to wildtype (717) after 4 months of growth. <bold>(B)</bold> Graphical mean height growth of <italic>dwarfy</italic> (gray data points) and wildtype (black data points) over two growth seasons. <bold>(C)</bold> Graphical representation of the genes found within a 30 kb window around the insertion of the activation tagging T-DNA in the <italic>dwarfy</italic> mutant. <bold>(D)</bold> Fold change in the genes found in the genomic vicinity of the T-DNA insertion in the <italic>dwarfy</italic> mutant as compared to wildtype in immature leaves (black bars), mature leaves (gray bars), stem (white bars). <bold>(E)</bold> Comparison of ethylene evolution in three different tissues of <italic>dwarfy</italic> (gray bars) and wildtype (black bars). All values &#x000b1;Standard Error (SE). In <bold>(D,E)</bold>, ^*significantly different from wildtype (<italic>p</italic> &#x0003c; 0.05).</p></caption><graphic xlink:href="fpls-05-00514-g0001"/></fig><fig id="F2" position="float"><label>Figure 2</label><caption><p><bold>Plant growth retardation in <italic>dwarfy</italic> mutant is significantly correlated to the expression level of <italic>PtaACS8</italic>. (A)</bold> A representative image of one independent line of the <italic>35S::PtaACS8</italic> mutant <italic>P. tremula x P. alba</italic> clone as compared to wildtype (717) after 2 months of growth. Scale bar = 8 cm. <bold>(B)</bold> Relative expression of <italic>PtaACS8</italic> in wildtype, <italic>dwarfy</italic> and 6 independent transgenic lines containing the <italic>35S::PtaACS8</italic> construct. <bold>(C)</bold> Internode length of wildtype, <italic>dwarfy</italic> and 6 independent transgenic lines containing the <italic>35S::PtaACS8</italic> construct. <bold>(D)</bold> Correlation between <italic>PtaACS8</italic> expression levels and internode length in wildtype, <italic>dwarfy</italic> and 6 independent transgenic lines containing the <italic>35S::PtaACS8</italic> construct. All values &#x000b1;Standard Error (SE). In <bold>(B,C)</bold>, ^*significantly different from wildtype (<italic>p</italic> &#x0003c; 0.05).</p></caption><graphic xlink:href="fpls-05-00514-g0002"/></fig><fig id="F3" position="float"><label>Figure 3</label><caption><p><bold>Elevated expression levels of <italic>PtaACS8</italic> result in significant changes in stem architecture and physical characteristics. (A)</bold> Transverse cross section of wild-type and <italic>dwarfy</italic> stems at the leaf 10&#x02013;11 internode as observed under brightfield and UV autofluorescence. Scale bar = 1 mm for first two images and 100 &#x003bc;m for the third panel <bold>(B)</bold> Vessel and fiber lengths of wild-type (black bars) and <italic>dwarfy</italic> (gray bars) stems between the leaf 10&#x02013;11 internode. <bold>(C)</bold> Transverse cross section of wildtype and <italic>dwarfy</italic> stems at the leaf 20&#x02013;21 internode as observed under brightfield and UV autofluorescence. Scale bar = 0.5 mm for first two images and 100 &#x003bc;m for the third panel. <bold>(D)</bold> Vessel and fiber lengths of wild-type (black bars) and <italic>dwarfy</italic> (gray bars) stems between the leaf 20&#x02013;21 internode. Vessel <bold>(E)</bold> and fiber <bold>(F)</bold> density in wild-type and <italic>dwarfy</italic> stems between the leaf 20&#x02013;21 internode. <bold>(G)</bold> Vessel diameter in wild-type and <italic>dwarfy</italic> stems between the leaf 20&#x02013;21 internode. Relative percentage of carbon <bold>(H)</bold>, nitrogen <bold>(I)</bold> and sulfur <bold>(J)</bold> in the stems of wildtype and <italic>dwarfy</italic>. All values &#x000b1;Standard Error. ^*Significantly different from wildtype (<italic>p</italic> &#x0003c; 0.05).</p></caption><graphic xlink:href="fpls-05-00514-g0003"/></fig></sec><sec><title>T-DNA insertion analysis</title><p>Southern analysis of the <italic>dwarfy</italic> poplar mutant line was done and confirmed the presence of one T-DNA insertion event (Harrison et al., <xref rid="B30" ref-type="bibr">2007</xref>). Genomic DNA (gDNA) was extracted from CFS greenhouse <italic>dwarfy</italic> mutant leaves using the cetyltrimethylammonium bromide (CTAB) method and gDNA was quantified using an Nanodrop1000 spectrophotometer and quality was checked on 0.8% (w/v) agarose Tris-acetate EDTA ethidium bromide gel. To identify the site of T-DNA insertion, the Genome Walker&#x02122; universal kit (Clontech, <ext-link ext-link-type="uri" xlink:href="http://www.clontech.com">http://www.clontech.com</ext-link>) was used according to the manufacturer's protocol to create 4 restriction digested gDNA libraries. Each library was analyzed by primary and nested PCR using T-DNA vector specific primers (VSP 1 and VSP 2) designed from the T-DNA sequence and adapter primers AP1 &#x00026; AP2 provided in the kit (Table <xref ref-type="supplementary-material" rid="SM1">S1</xref>). Primary PCRs were done on 1 &#x003bc;l of each library except using High Fidelity Platinum Taq (Invitrogen) for the PCR reaction mix. Primary PCR reactions from each library were diluted 50 times in H₂O and 1 &#x003bc;l of the each dilution was used for nested PCR analysis using the same PCR reaction components except primers VSP2 &#x00026; AP2 were used. Cycling parameters for both primary &#x00026; nested PCRs were the same ones stated in the kit except that the elongation time was increased to 5 min. Primary and nested PCRs were analyzed by gel electrophoresis and bands from the nested PCR reaction that were over 1 kb in size were subcloned in pCR4-topo vector using the TOPO TA cloning kit (Invitrogen, <ext-link ext-link-type="uri" xlink:href="http://www.invitrogen.com">http://www.invitrogen.com</ext-link>) and fully sequenced at the McGill University and Genome Quebec Innovation Centre (<ext-link ext-link-type="uri" xlink:href="http://gqinnovationcenter.com">http://gqinnovationcenter.com</ext-link>). Based on flanking sequence information, a flanking gDNA primer FSP1 was designed and PCR was done on <italic>dwarfy</italic> gDNA using FSP1 and VSP 2 primers to confirm the location of the T-DNA insertion site. This amplicon was TOPO TA cloned and fully sequenced. Localisation of insertion site was determined by BLASTn using flanking sequence as query against the <italic>P. trichocarpa</italic> genome (Tuskan et al., <xref rid="B81" ref-type="bibr">2006</xref>) (<italic>Populus trichocarpa</italic> v3.0, DOE-JGI, <ext-link ext-link-type="uri" xlink:href="http://www.phytozome.net/poplar">http://www.phytozome.net/poplar</ext-link>).</p></sec><sec><title>Affymetrix array analysis</title><p>For gene expression analysis, total RNA was extracted from 0.5 g of different tissues of wildtype and <italic>dwarfy</italic> taken from CFS greenhouse grown plants using a modification of Chang et al., <xref rid="B11a" ref-type="bibr">1993</xref> and the RNeasy kit (QIAGEN). Total RNA quality and quantity was determined by Nanodrop1000 and by electrophoresis. Triplicate samples of <italic>dwarfy</italic> and wild-type leaf and stem total RNA were sent to the Microarray Centre (University Health Network, Toronto (UHN)) for sample processing and analysis. Sample quality was verified with the Agilent 2100 Bioanalyser before analysis with the GeneChip&#x000ae; Poplar Genome Array. Data was also analyzed by the Microarray Center (UHN) using Gene Spring software.</p></sec><sec><title>Isolation of ACS8 candidate gene coding sequence</title><p>The <italic>ACS8</italic> cDNA was isolated from wild-type leaf tissues using the Smart RACE kit (Clontech, <ext-link ext-link-type="uri" xlink:href="http://www.clontech.com">http://www.clontech.com</ext-link>) according to manufacturer's protocol. One microgram of total RNA was used to produce the 5&#x02032;&#x00026;3&#x02032;RACE ready products and <italic>ACS8</italic> gene specific primers ACS8RACE.fwd and ACS8RACE.rev were used along with the Universal Primer (UP) provided in the kit (Table <xref ref-type="supplementary-material" rid="SM1">S1</xref>). 5&#x02032;&#x00026;3&#x02032; RACE products were subcloned in pCR4-topo vector using the TOPO TA cloning kit and sequenced. Gene specific primers; ACS8-ATG.fwd and ACS8-Stop.rev primers were designed and used to isolate the full <italic>ACS8</italic> CDS using the 3&#x02032;RACE ready product previously generated and the amplicon was cloned in pBluescript II (+) (Fermentas, <ext-link ext-link-type="uri" xlink:href="http://www.fermentas.com">http://www.fermentas.com</ext-link>) using <italic>Hin</italic>dIII-<italic>Xba</italic>I restriction sites. The resulting construct carrying the full <italic>ACS8</italic> CDS was fully sequenced.</p></sec><sec><title>Agrobacterium tumefaciens mediated tranformation of <italic>P. tremula x P. Alba</italic> clone 717-1B4</title><p>In order to generate a binary plant transformation vector, <italic>ACS8/</italic>pBluescript II (+) construct was digested with <italic>Eco</italic>RI and subcloned in pART7 (Gleave, <xref rid="B27" ref-type="bibr">1992</xref>) and verified by restriction digest for correct orientation between the CaMV promoter and ocs 3&#x02032; region. The <italic>ACS8</italic>/pART7 construct was subsequently digested with NotI and the whole cassette was ligated in the binary vector pART 27 (Gleave, <xref rid="B27" ref-type="bibr">1992</xref>) prior to <italic>Agrobacterium tumefaciens</italic> transformation in line 717-1-B4 (Harrison et al., <xref rid="B30" ref-type="bibr">2007</xref>). Out of 17 independent transgenic lines generated, 6 lines survived the transfer to greenhouse conditions and these lines were analyzed. Total RNA from leaf tissue was extracted from newly transformed lines and RT-qPCRs were done for gene expression analysis of lines generated using the procedures as stated below. Each line was analyzed in duplicate technical replicates. Internode lengths were measured after 3 months of growth.</p></sec><sec><title>Quantitative gene expression analysis</title><p>For in gene expression analysis in transgenic 35S::<italic>PtaACS8</italic> lines, total RNA was extracted from 100 mg of shoot apical tissue using the RNeasy kit (QIAGEN). Total RNA quality and quantity was determined by Nanodrop1000 and by electrophoresis. Two to four micrograms of total RNA was treated with Turbo DNaseI (Ambion, <ext-link ext-link-type="uri" xlink:href="http://www.ambion.com">http://www.ambion.com</ext-link>) and RT-qPCR was done with 50 ng of total RNA/reaction using the one step Quantitect SYBR Green RT-PCR kit (QIAGEN, <ext-link ext-link-type="uri" xlink:href="http://www.qiagen.com">http://www.qiagen.com</ext-link>). RT-qPCR cycling conditions were: 30 min at 50&#x000b0;C for reverse transcriptase reaction and 15 min at 95&#x000b0;C for enzyme inactivation followed by 40 cycles of 15 s at 94&#x000b0;C, 15 s denaturation at 55&#x000b0;C (annealing) and 30 s at 72&#x000b0;C elongation followed by fluorescence measurement. The relative expression of <italic>PtaACS8</italic> was compared to the <italic>UBQ10</italic> reference gene (Plett et al., <xref rid="B61" ref-type="bibr">2010</xref>).</p><p>The amplification efficiencies of each gene primer set were determined by <italic>E</italic> = 10^{[&#x02212;1/slope}] and were calculated using the slopes of n-fold serial dilution standard curves. Fold change ratios were determined using the comparative Ct method (&#x00394;&#x00394;Ct method) since amplification efficiencies were approximately equal in all target and reference genes measured in the study. Samples were analyzed in triplicates of each wildtype and the <italic>dwarfy</italic> mutant plants. Each total RNA sample was analyzed in duplicate. A No Reverse Transcriptase (NoRT) for each sample was included and a No Template Control (NTC) was included for each primer pair to make sure no contamination was present in the experiments. Amplicon specificity was confirmed by electrophoresis (single band at the right size), by melt curve analysis (single peak and Tm) and by sequencing.</p></sec><sec><title>Ethylene determination</title><p>Leaf and stem samples were removed from wild-type and <italic>dwarfy</italic> poplar plants between 10 AM and 12 PM, and incubated in 20 mL headspace vials for 4 h at ambient temperature. Fresh weight was recorded and time between vial seal and sample injection were noted to have an exact incubation time. Ethylene content within this headspace was determined by gas chromatography coupled to a flame ionization detector (Gas Chromatography- Flame Ionization Detector GC-FID, Agilent 7890A) with an injector (splitless mode) temperature of 240&#x000b0;C and oven temperature at 60&#x000b0;C (isothermal) using helium as carrier gas (3 mL/min). A30 m &#x000d7; 0.53 mm ID (30 &#x003bc;m average thickness) Carboxen 1006 PLOT column (Supelco, Sigma Aldrich, Ontario Canada) was used to separate ethylene from the mixture. The FID (heated to 240&#x000b0;C) hydrogen:air:makeup flows were 30:400:25 (mL/min). Two measurements of 0.1 mL gas aliquot was taken from a headspace vial using a gas tight syringe (Hamilton 1700 series) and immediately injected for each sample. At these conditions, the observed retention time of ethylene was 3.14 min. A calibration curve was generated to cover the range of 0.1&#x02013;20 mg ethylene. Ethylene standard gas mixture was made by drawing a volume of 99.5% ethylene (Praxair) and injecting it in a previously vacuum-purged sealed headspace vial (volume determined by water capacity), then breaking the vacuum with a syringe needle and filling the vial with ambient air to atmospheric pressure. The diluted ethylene was allowed to stand for 1 h to reach dispersal equilibrium. Increasing volumes were injected to cover the desired ethylene range and each injection was repeated in triplicate.</p></sec><sec><title>Physical characteristics analysis</title><p>Cuttings of wildtype and <italic>dwarfy</italic> were established by cutting 4&#x02013;5 cm shoot explants from stock plants. Cuttings were planted in Jiffy 42 mm peat plugs grown under greenhouse conditions June to August (natural lighting, watered twice daily) for 7 weeks. After 7 weeks, cuttings were transferred to greenhouse pots (15 cm diameter, 19 cm long). After a total of 9 weeks, 3 trees were randomly selected every month and height was measured and data analyzed using basic statistical tools in Excel (Microsoft Office). Leaf cell size, trichome density and cell density were performed as per Plett et al. (<xref rid="B61" ref-type="bibr">2010</xref>). Fiber and vessel isolation and measurements were performed as per Chaffey et al. (<xref rid="B11" ref-type="bibr">2002</xref>).</p></sec><sec><title>GA and ethylene biosynthetic inhibitor growth effect analysis</title><p>Plants used for GA effect on growth were grown under normal greenhouse conditions as mentioned above. A triplicate (for GA) or a duplicate (for AVG) of wild-type and <italic>dwarfy</italic> plants of similar heights, grown for 5 weeks from cuttings, were used for each treatment for the experiment. Plant height was measured before the experiment and measured prior to each new application of GA or AVG. A total of 3 applications of 10 &#x003bc;l of 3 mM GA/water or ETOH (for GA analysis) or of water or 100 &#x003bc;M AVG were added every fourth day to the shoot apex of each plant and the total length of the experiment was 12 days. Data was analyzed using the height difference between the first measurement (before first application) and before 3rd application (3rd measurement), since 2 of the dwarfy/GA treated shoot apex samples had dried up and were dead before the final measurement.</p></sec><sec><title>Percent carbon, nitrogen and sulfur analysis</title><p>Dried leaf, stem, and roots samples from both wildtype and <italic>dwarfy</italic> were ground with a bead mill, and kept under vacuum in order to keep moisture out of the samples prior to carbon (C), nitrogen (N) and sulfur (S) analysis (CNS) by the CFS analytical laboratory according to the method of Kalra and Maynard (<xref rid="B41" ref-type="bibr">1991</xref>). A triplicate of each clone for each tissue types was measured and data was analyzed using basic statistical tools in Excel (Microsoft Office). Results presented are the measure of C, N and S from healthy mature leaves and internode tissues harvested in June of the growing season.</p></sec></sec><sec sec-type="results" id="s3"><title>Results</title><sec><title>Endogenous over-expression of PtaACS8 induces dwarfism in populus</title><p>From a large population of activation-tagged <italic>P. tremula x P. alba</italic> clone 717-1B4 (2000 independent transgenic lines; Harrison et al., <xref rid="B30" ref-type="bibr">2007</xref>), we identified one line with a consistent reduction in growth rate over multiple growing seasons (Figures <xref ref-type="fig" rid="F1">1A,B</xref>). This mutant was named <italic>dwarfy</italic>. Using Southern blotting only one T-DNA insert in <italic>dwarfy</italic> and located this insert on chromosome 2 using a modified TAIL PCR was identified. Within a window of &#x000b1;20 Kb around the T-DNA, 3 genes annotated in Phytozome (Figure <xref ref-type="fig" rid="F1">1C</xref>) were found as follows: a gene of unknown function (Potri.002G11400; +14.4 Kb up-stream), <italic>PtaACC SYNTHASE8</italic> (<italic>PtaACS8;</italic> Potri.002G113900; 13.1 Kb down-stream) and <italic>PtaEARLY-RESPONSE TO DEHYDRATION 4</italic> (<italic>ERD4;</italic> Potri.002G113800; 16.9 Kb down-stream). A quantification of the expression of these genes in the <italic>dwarfy</italic> mutant line relative to wild-type <italic>P. tremula x P. alba</italic> clone 717-1B4 demonstrated that only <italic>PtaACS8</italic> exhibited increased gene expression in all aerial tissues of the plant (Figure <xref ref-type="fig" rid="F1">1D</xref>). As ACC synthases are involved in the first step in the biosynthesis of the plant hormone ethylene, ethylene production was measured in the same three compartments as used for gene expression analysis in wild-type and mutant plants (i.e., young and mature leaves and stem tissues). Compared to wildtype, the mutant line produced 14&#x000d7; higher levels of ethylene in younger leaves and 6&#x000d7; higher levels of ethylene in mature leaves and the stems (Figure <xref ref-type="fig" rid="F1">1E</xref>).</p><p>To verify that increased transcript abundance of <italic>PtaACS8</italic> was indeed responsible for the dwarfism phenotype of the mutant, the Potri.002G113900 gene was cloned and expressed ectopically in the <italic>P. tremula x P. alba</italic> clone 717-1B4 genetic background under the control of the 35S-cauliflower mosaic virus promoter. We were able to regenerate six independent transgenic lines from callus culture which, when grown alongside age-equivalent wildtype (i.e., propagated at the same time and treated in the same manner as the <italic>35S::PtaACS8</italic> lines), displayed a dwarf phenotype (Figure <xref ref-type="fig" rid="F2">2A</xref>). This reduction in growth and internode length was significant as compared to wildtype in all lines tested although the plants were consistently bigger than <italic>dwarfy</italic> (Figures <xref ref-type="fig" rid="F2">2A,B</xref>). The discrepancy in height difference is likely due to the fact that none of the <italic>35S::PtaACS8</italic> transgenic lines displayed the same level of <italic>PtaACS8</italic> transcript accumulation as <italic>dwarfy</italic> (Figure <xref ref-type="fig" rid="F2">2C</xref>). As there was a significant correlation between the transcript abundance of <italic>PtaACS8</italic> and the dwarf phenotype in the transgenic lines (Figure <xref ref-type="fig" rid="F2">2D</xref>; <italic>r</italic> = 0.91; <italic>p</italic> &#x0003c; 0.001), we conclude that increased transcript abundance of <italic>PtaACS8</italic> in the original <italic>dwarfy</italic> transgenic line is responsible for the reduction in plant stature.</p></sec><sec><title>Increased transcript abundance of PtaACS8 leads to altered stem characteristics</title><p>The <italic>dwarfy</italic> mutant line exhibited alterations to the morphology of all aerial parts of the plant. While the internode length of the <italic>dwarfy</italic> line was significantly reduced (Figure <xref ref-type="fig" rid="F2">2B</xref>), there were also significant alterations to the microscopic anatomy of the stem (Figure <xref ref-type="fig" rid="F3">3</xref>). Due to the great difference in height of the two plants being compared, we used a plastochron index to identify and compare the same internode between the mutant line and wildtype. We used different microscopy techniques to observe different stem properties: brightfield to gain a general over-view of the stem architecture, UV excitation to differentiate chlorophyll autofluorescence (red signal) from secondary cell wall fluorescence (blue-green signal; Figures <xref ref-type="fig" rid="F3">3A,C</xref>). In young stems (internode between leaves 10 and 11), there was a reduction in the amount of secondary xylem formed in <italic>dwarfy</italic> as compared to wildtype (Figure <xref ref-type="fig" rid="F3">3A</xref>) as well as a significant reduction in the length of xylem fibers and vessels (Figure <xref ref-type="fig" rid="F3">3B</xref>; <italic>p</italic> &#x0003c; 0.05). In older stem tissues (internode between leaves 20 and 21), the reduction in secondary xylem formation (Figure <xref ref-type="fig" rid="F3">3C</xref>) and fiber/vessel lengths were still observed (Figure <xref ref-type="fig" rid="F3">3D</xref>). Detailed analysis of wood formation in these older tissues also revealed a difference in cell density: <italic>dwarfy</italic> had a higher density of xylem vessels per square millimeter with a significantly smaller outer diameter as compared to wild-type stems (Figures <xref ref-type="fig" rid="F3">3A,C,E&#x02013;G</xref>) while there was no significant difference in the density of fibers. As alterations to the cell make-up of the stem and alteration in growth rate may influence nutrient deposition in the stem, we analyzed the percentage of carbon, nitrogen and sulfur in these mature internodes of both wildtype and <italic>dwarfy</italic>. No significant difference in percent accumulation of carbon and sulfur in the stems of <italic>dwarfy</italic> and wildtype were observed while the stems of the former accumulated a significantly higher concentration of nitrogen-containing compounds (Figures <xref ref-type="fig" rid="F3">3H&#x02013;J</xref>; <italic>p</italic> &#x0003c; 0.05).</p></sec><sec><title>Hormone- and nutrient-related genes display altered abundance in <italic>dwarfy</italic> stems</title><p>In order to understand the transcriptomic profile of <italic>dwarfy</italic> stems, we performed a whole genome oligo-array transcriptomic analysis of whole stem tissues. We found 223 genes differentially expressed (&#x02265;2-fold; <italic>p</italic> &#x0003c; 0.05) as compared to wild-type <italic>P. tremula x P. alba</italic> clone 717-1B4 stems of the same age (Table <xref ref-type="supplementary-material" rid="SM2">S2</xref>). Within these genes we found that there were a number of ethylene and auxin related genes and genes coding for proteins involved in nutrient transport and biosynthesis (Table <xref ref-type="table" rid="T1">1</xref>). Genes related to the ethylene pathway included <italic>PtaACS8</italic> (&#x0003e;230-fold increase) a number of ETHYLENE RESPONSE FACTOR (ERF) genes, two serine-threonine receptor kinases (<italic>PtaCTR3, PtaCTR4</italic>) and two ethylene receptor genes (<italic>PtaETR1, PtaETR5</italic>). The majority of genes associated with the auxin pathway, meanwhile, were repressed in the stems of <italic>dwarfy</italic> while a gene encoding an IAA-amido-synthetase glycosyl-hydrolase (GH) family protein displayed increased abundance. Nutrient transport and synthesis was also affected with two sugar transporters and an amino acid transporter being repressed while the transcript accumulation of a glutamine synthase was increased (Table <xref ref-type="table" rid="T1">1</xref>).</p><table-wrap id="T1" position="float"><label>Table 1</label><caption><p><bold>Genes found to have significantly different abundance in the stems of <italic>dwarfy</italic> as compared to wildtype</bold>.</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" valign="top" rowspan="1" colspan="1"><bold>Probe</bold></th><th align="left" valign="top" rowspan="1" colspan="1"><bold>RefSeq protein ID</bold></th><th align="center" valign="top" rowspan="1" colspan="1"><bold><italic>E</italic>-value</bold></th><th align="center" valign="top" rowspan="1" colspan="1"><bold>Fold change</bold></th><th align="left" valign="top" rowspan="1" colspan="1"><bold>Gene title</bold></th></tr></thead><tbody><tr><td align="left" valign="top" colspan="5" rowspan="1"><bold>ETHYLENE RELATED</bold></td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">PtpAffx.202003.1.S1_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002302380</td><td align="center" valign="top" rowspan="1" colspan="1">0.00E+00</td><td align="center" valign="top" rowspan="1" colspan="1">233.9</td><td align="left" valign="top" rowspan="1" colspan="1">1-Aminocyclopropane-1-carboxylate 8 (<italic>ACS8</italic>)</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">Ptp.6619.1.S1_s_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002315490</td><td align="center" valign="top" rowspan="1" colspan="1">8.00E&#x02013;144</td><td align="center" valign="top" rowspan="1" colspan="1">30.5</td><td align="left" valign="top" rowspan="1" colspan="1">AP2/ERF domain-containing transcription factor</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">PtpAffx.75787.1.A1_s_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002297877</td><td align="center" valign="top" rowspan="1" colspan="1">0.00E+00</td><td align="center" valign="top" rowspan="1" colspan="1">20.4</td><td align="left" valign="top" rowspan="1" colspan="1">AP2/ERF domain-containing transcription factor</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">PtpAffx.75787.1.A1_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002304640</td><td align="center" valign="top" rowspan="1" colspan="1">0.00E+00</td><td align="center" valign="top" rowspan="1" colspan="1">14.3</td><td align="left" valign="top" rowspan="1" colspan="1">AP2/ERF domain-containing transcription factor</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">PtpAffx.129036.1.S1_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002316302</td><td align="center" valign="top" rowspan="1" colspan="1">1.00E&#x02013;16</td><td align="center" valign="top" rowspan="1" colspan="1">12.3</td><td align="left" valign="top" rowspan="1" colspan="1">Ethylene-responsive protein</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">PtpAffx.219707.1.S1_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002326299</td><td align="center" valign="top" rowspan="1" colspan="1">0.00E+00</td><td align="center" valign="top" rowspan="1" colspan="1">8.1</td><td align="left" valign="top" rowspan="1" colspan="1">AP2/ERF domain-containing transcription factor</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">PtpAffx.4624.1.S1_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002328620</td><td align="center" valign="top" rowspan="1" colspan="1">0.00E+00</td><td align="center" valign="top" rowspan="1" colspan="1">5.8</td><td align="left" valign="top" rowspan="1" colspan="1">AP2/ERF domain-containing transcription factor</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">PtpAffx.572.3.S1_a_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002315958</td><td align="center" valign="top" rowspan="1" colspan="1">0.00E+00</td><td align="center" valign="top" rowspan="1" colspan="1">5.3</td><td align="left" valign="top" rowspan="1" colspan="1">AP2/ERF domain-containing transcription factor</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">Ptp.162.1.A1_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002302732</td><td align="center" valign="top" rowspan="1" colspan="1">0.00E+00</td><td align="center" valign="top" rowspan="1" colspan="1">3.6</td><td align="left" valign="top" rowspan="1" colspan="1">Ethylene receptor 1 (<italic>PtETR1</italic>)</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">PtpAffx.79014.1.S1_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002316514</td><td align="center" valign="top" rowspan="1" colspan="1">0.00E+00</td><td align="center" valign="top" rowspan="1" colspan="1">3.5</td><td align="left" valign="top" rowspan="1" colspan="1">Serine/threonine protein kinase (<italic>PtCTR4</italic>)</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">Ptp.866.1.S1_s_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002310408</td><td align="center" valign="top" rowspan="1" colspan="1">4.00E&#x02013;118</td><td align="center" valign="top" rowspan="1" colspan="1">3.3</td><td align="left" valign="top" rowspan="1" colspan="1">AP2/ERF domain-containing transcription factor</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">PtpAffx.208193.1.S1_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002311669</td><td align="center" valign="top" rowspan="1" colspan="1">0.00E+00</td><td align="center" valign="top" rowspan="1" colspan="1">2.8</td><td align="left" valign="top" rowspan="1" colspan="1">Ethylene receptor 5 (<italic>PtETR5</italic>)</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">PtpAffx.122897.1.A1_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002308565</td><td align="center" valign="top" rowspan="1" colspan="1">1.00E&#x02013;96</td><td align="center" valign="top" rowspan="1" colspan="1">2.7</td><td align="left" valign="top" rowspan="1" colspan="1"><italic>REVERSION-TO-ETHYLENE SENSITIVITY1</italic> (<italic>RTE1</italic>)</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">PtpAffx.13062.4.S1_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002308982</td><td align="center" valign="top" rowspan="1" colspan="1">0.00E+00</td><td align="center" valign="top" rowspan="1" colspan="1">2.5</td><td align="left" valign="top" rowspan="1" colspan="1">ein3-binding f-box protein 4</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">Ptp.2044.2.S1_a_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002311967</td><td align="center" valign="top" rowspan="1" colspan="1">0.00E+00</td><td align="center" valign="top" rowspan="1" colspan="1">2.3</td><td align="left" valign="top" rowspan="1" colspan="1">Serine/threonine protein kinase (<italic>PtCTR3</italic>)</td></tr><tr><td align="left" valign="top" colspan="5" rowspan="1"><bold>AUXIN RELATED</bold></td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">PtpAffx.144034.1.S1_s_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002310372</td><td align="center" valign="top" rowspan="1" colspan="1">1.00E&#x02013;06</td><td align="center" valign="top" rowspan="1" colspan="1">3.2</td><td align="left" valign="top" rowspan="1" colspan="1"><italic>AUXIN-REGULATED GENE INVOLVED IN ORGAN SIZE</italic> (<italic>ARGOS</italic>)</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">Ptp.6069.1.S1_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002320183</td><td align="center" valign="top" rowspan="1" colspan="1">0.00E+00</td><td align="center" valign="top" rowspan="1" colspan="1">2.9</td><td align="left" valign="top" rowspan="1" colspan="1">GH3 family protein</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">Ptp.8069.1.S1_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002306504</td><td align="center" valign="top" rowspan="1" colspan="1">0.00E+00</td><td align="center" valign="top" rowspan="1" colspan="1">&#x02013;2.1</td><td align="left" valign="top" rowspan="1" colspan="1"><italic>NAKED PINS IN YUC MUTANTS 2</italic> (<italic>NPY2</italic>)</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">PtpAffx.155898.1.S1_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002320550</td><td align="center" valign="top" rowspan="1" colspan="1">5.00E&#x02013;109</td><td align="center" valign="top" rowspan="1" colspan="1">&#x02013;2.3</td><td align="left" valign="top" rowspan="1" colspan="1">Dopamine beta-monooxygenase</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">PtpAffx.97214.1.A1_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002302687</td><td align="center" valign="top" rowspan="1" colspan="1">2.00E&#x02013;132</td><td align="center" valign="top" rowspan="1" colspan="1">&#x02013;2.3</td><td align="left" valign="top" rowspan="1" colspan="1">Auxin-induced protein 5NG4</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">PtpAffx.117529.1.S1_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002323866</td><td align="center" valign="top" rowspan="1" colspan="1">0.00E+00</td><td align="center" valign="top" rowspan="1" colspan="1">&#x02013;3.0</td><td align="left" valign="top" rowspan="1" colspan="1">MDR family ABC transporter family</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">PtpAffx.210100.1.S1_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002317029</td><td align="center" valign="top" rowspan="1" colspan="1">2.00E&#x02013;157</td><td align="center" valign="top" rowspan="1" colspan="1">&#x02013;3.1</td><td align="left" valign="top" rowspan="1" colspan="1">Auxin:hydrogen symporter</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">PtpAffx.7696.4.S1_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002312567</td><td align="center" valign="top" rowspan="1" colspan="1">1.00E&#x02013;75</td><td align="center" valign="top" rowspan="1" colspan="1">&#x02013;4.3</td><td align="left" valign="top" rowspan="1" colspan="1">Auxin-responsive protein IAA4</td></tr><tr><td align="left" valign="top" colspan="5" rowspan="1"><bold>NUTRIENT SYNTHESIS/TRANSPORT</bold></td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">PtpAffx.2311.1.S1_s_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002313246</td><td align="center" valign="top" rowspan="1" colspan="1">0.00E+00</td><td align="center" valign="top" rowspan="1" colspan="1">6.2</td><td align="left" valign="top" rowspan="1" colspan="1"><italic>GLUTAMINE-DEPENDENT ASPARAGINE SYNTHASE 1</italic> (<italic>ASN1</italic>)</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">PtpAffx.217242.1.S1_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002331420</td><td align="center" valign="top" rowspan="1" colspan="1">3.00E&#x02013;169</td><td align="center" valign="top" rowspan="1" colspan="1">&#x02013;2.3</td><td align="left" valign="top" rowspan="1" colspan="1">Sugar transporter</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">Ptp.5882.1.S1_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002301819</td><td align="center" valign="top" rowspan="1" colspan="1">3.00E&#x02013;43</td><td align="center" valign="top" rowspan="1" colspan="1">&#x02013;2.1</td><td align="left" valign="top" rowspan="1" colspan="1">RS21-C6 protein</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">PtpAffx.111624.1.S1_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002302894</td><td align="center" valign="top" rowspan="1" colspan="1">0.00E+00</td><td align="center" valign="top" rowspan="1" colspan="1">&#x02013;10.2</td><td align="left" valign="top" rowspan="1" colspan="1">Amino acid transporter</td></tr></tbody></table><table-wrap-foot><p><italic>(p &#x0003c; 0.05; &#x0003e;2-fold differential regulation). Note: In this table there are no column lines as there are in Table <xref ref-type="supplementary-material" rid="SM2">2</xref></italic>.</p></table-wrap-foot></table-wrap><p>GA has been linked to enhanced growth phenotypes through the induction of auxin biosynthesis and polar transportation (Bj&#x000f6;rklund et al., <xref rid="B6" ref-type="bibr">2007</xref>). Therefore, as our transcriptional analysis of the <italic>dwarfy</italic> mutant indicated that auxin transport and signaling was affected, we tested whether GA application to the growing apex of <italic>dwarfy</italic> would be able to rescue the growth phenotype of the mutant. We found that the growth rate of <italic>dwarfy</italic> was significantly increased by treatment with GA (Figures <xref ref-type="fig" rid="F4">4A,B</xref>). Therefore GA is able to rescue the <italic>dwarfy</italic> phenotype. We also treated <italic>dwarfy</italic> with the ethylene biosynthetic inhibitor AVG. This treatment resulted in an increase in internode length (Figures <xref ref-type="fig" rid="F4">4C&#x02013;E</xref>), demonstrating that blocking ethylene synthesis also rescues the <italic>dwarfy</italic> phenotype.</p><fig id="F4" position="float"><label>Figure 4</label><caption><p><bold>Application of GA and AVG to <italic>dwarfy</italic> apexes induces faster growth rate. (A)</bold> Representative image of GA influence on the growth rates of <italic>dwarfy</italic> and wildtype (+GA) as compared to ethanol control (+EtOH) and untreated control (Cont.). <bold>(B)</bold> Mean heights of wild-type and <italic>dwarfy</italic> saplings treated with GA (+GA) as compared to ethanol control (+EtOH) and untreated control (Cont.). <bold>(C)</bold> Representative image of water and <bold>(D)</bold> AVG influence on the growth rates of <italic>dwarfy</italic> and wild-type. Parentheses indicate growth of main stem for the treatment period. Scale bar = 3 cm. <bold>(E)</bold> Mean internode lengths wild-type and <italic>dwarfy</italic> saplings treated with AVG as compared to water control (Cont.). All values &#x000b1;SE. ^*Significantly different from wildtype (<italic>p</italic> &#x0003c; 0.05).</p></caption><graphic xlink:href="fpls-05-00514-g0004"/></fig></sec><sec><title>Increased transcript abundance of PtaACS8 leads to altered leaf characteristics</title><p>Mature leaves in the <italic>dwarfy</italic> mutant also showed altered size when compared to wild-type leaves (Figure <xref ref-type="fig" rid="F5">5A</xref>). The leaves of <italic>dwarfy</italic> were much smaller than those of wildtype (Figure <xref ref-type="fig" rid="F1">1A</xref>). Despite the alterations in leaf size, the epidermal cell size of <italic>dwarfy</italic> was not significantly altered compared to wildtype (Figure <xref ref-type="fig" rid="F5">5B</xref>). Trichome density was also not affected, but stomate density was significantly higher in the <italic>dwarfy</italic> mutant (Figures <xref ref-type="fig" rid="F5">5C,D</xref>). Only nitrogen content was significantly higher in <italic>dwarfy</italic> stems, as compared to wildtype (Figure <xref ref-type="fig" rid="F3">3I</xref>). Unlike stems, the total percentage of carbon in leaves was significantly reduced in <italic>dwarfy</italic> as compared to wildtype (Figure <xref ref-type="fig" rid="F5">5E</xref>), while nitrogen levels were not altered (Figure <xref ref-type="fig" rid="F5">5F</xref>). Sulfur levels in mutant leaves showed a tendency toward a lower accumulation compared to wildtype, but this difference was not found to be significant (Figure <xref ref-type="fig" rid="F5">5G</xref>; <italic>p</italic> &#x0003c; 0.05). The date at which leaves became chlorotic and dropped off the stem in <italic>dwarfy</italic> as compared to wild-type plants was assessed as increased ethylene levels have been correlated to early leaf senescence (Breeze et al., <xref rid="B7" ref-type="bibr">2011</xref>; Koyama et al., <xref rid="B43" ref-type="bibr">2013</xref>). When grown under natural conditions, chlorosis of 1-year-old <italic>dwarfy</italic> leaves happens earlier as compared to wild-type plants (Figures <xref ref-type="fig" rid="F6">6A&#x02013;D</xref>) and significant leaf drop occurred in <italic>dwarfy</italic> plants in the month of November while there was no significant leaf drop in the same period in wild-type trees (Figures <xref ref-type="fig" rid="F6">6E,F</xref>). It is interesting to speculate that the reduced C and S observed in the <italic>dwarfy</italic> leaves may be related to the shorter growing season for these leaves.</p><fig id="F5" position="float"><label>Figure 5</label><caption><p><bold>Elevated expression levels of <italic>PtaACS8</italic> result in significant changes in leaf architecture and physical characteristics. (A)</bold> A representative image of fully expanded leaves of wildtype, one independent line of the <italic>35S::PtaACS8</italic> mutant <italic>P. tremula x P. alba</italic> clone 717-1B4 and <italic>dwarfy</italic>, respectively, after 2 months of growth. Scale bar = 2 cm. Epidermal cell density <bold>(B)</bold>, trichome density <bold>(C)</bold> and stomate density <bold>(D)</bold> in fully expanded leaves of <italic>dwarfy</italic> (gray bars) as compared to wild-type leaves (black bars). Relative percentage of carbon <bold>(E)</bold>, nitrogen <bold>(F)</bold>, and sulfur <bold>(G)</bold> in mature leaves of wildtype (black bars) and <italic>dwarfy</italic> (gray bars). All values &#x000b1;SE. ^*Significantly different from wildtype (<italic>p</italic> &#x0003c; 0.05).</p></caption><graphic xlink:href="fpls-05-00514-g0005"/></fig><fig id="F6" position="float"><label>Figure 6</label><caption><p><bold><italic>dwarfy</italic> mutants exhibit pre-mature leaf senescence</bold>. Comparison of leaf senescence rates in wildtype and <italic>dwarfy</italic> mutant clones within their first year of growth in August <bold>(A)</bold>, September <bold>(B)</bold>, October <bold>(C)</bold>, and November <bold>(D)</bold>. Leaf numbers on wild-type trees <bold>(E)</bold> and <italic>dwarfy</italic> trees <bold>(F)</bold> over the same time period are presented. &#x000b1;SE; superscript letters indicate significant differences between treatments as determined by One-Way analysis of variance (ANOVA) followed by a Tukey HSD (honestly significant difference) multiple comparison test (<italic>p</italic> &#x0003c; 0.05).</p></caption><graphic xlink:href="fpls-05-00514-g0006"/></fig></sec><sec><title>Senescence- and cell cycle-related genes exhibit altered abundance in <italic>dwarfy</italic> leaves</title><p>We found that 183 genes were significantly regulated in fully expanded leaves of <italic>dwarfy</italic> as compared to wild-type <italic>P. tremula x P. alba</italic> clone 717-1B4. A large number of hormone-related genes with altered transcription were observed in the stems of <italic>dwarfy</italic>, while only two of these genes (<italic>PtaACS8</italic> and a <italic>GH3</italic> family protein) were significantly differentially regulated in mature <italic>dwarfy</italic> leaves (&#x02265;2-fold; <italic>p</italic> &#x0003c; 0.05; Table <xref ref-type="table" rid="T2">2</xref>; Table <xref ref-type="supplementary-material" rid="SM3">S3</xref>). A number of nutrient transporters displayed altered transcript abundance, although they were different from those identified in <italic>dwarfy</italic> stems (Table <xref ref-type="table" rid="T1">1</xref>). Three other classes of genes were differentially regulated in mature <italic>dwarfy</italic> leaves that were not observed in the stems: defense-, senescence- and cell cycle/expansion-related genes (Table <xref ref-type="table" rid="T2">2</xref>). The majority of the defense-related genes were associated with pathogen attack, including a chitinase, a lipase and a glyoxal oxidase. As the leaf tissues were healthy at the time of harvest and displayed no infection structures, the activation of these genes is likely constitutive in the <italic>dwarfy</italic> background. Three genes associated with leaf senescence were also up-regulated. One group that only showed reduced levels of transcript abundance was that of genes associated with cell cycle and cellular growth (Table <xref ref-type="table" rid="T2">2</xref>). Within this group of genes were a number of cyclins, calmodulin-like proteins and one expansin.</p><table-wrap id="T2" position="float"><label>Table 2</label><caption><p><bold>Genes found to have significantly different abundance in the leaves of <italic>dwarfy</italic> as compared to wildtype</bold>.</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" valign="top" rowspan="1" colspan="1"><bold>Probe</bold></th><th align="left" valign="top" rowspan="1" colspan="1"><bold>RefSeq protein ID</bold></th><th align="center" valign="top" rowspan="1" colspan="1"><bold><italic>E</italic>-value</bold></th><th align="center" valign="top" rowspan="1" colspan="1"><bold>Fold change</bold></th><th align="left" valign="top" rowspan="1" colspan="1"><bold>Gene title</bold></th></tr></thead><tbody><tr><td align="left" valign="top" colspan="5" rowspan="1"><bold>HORMONE</bold></td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">PtpAffx.202003.1.S1_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002302380</td><td align="center" valign="top" rowspan="1" colspan="1">0.00E+00</td><td align="center" valign="top" rowspan="1" colspan="1">506.5</td><td align="left" valign="top" rowspan="1" colspan="1">1-Aminocyclopropane-1-carboxylate 8 (<italic>ACS8</italic>)</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">PtpAffx.211163.1.S1_s_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002319260</td><td align="center" valign="top" rowspan="1" colspan="1">0.00E+00</td><td align="center" valign="top" rowspan="1" colspan="1">2.8</td><td align="left" valign="top" rowspan="1" colspan="1">GH3 family protein</td></tr><tr><td align="left" valign="top" colspan="5" rowspan="1"><bold>NUTRIENT TRANSPORT</bold></td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">PtpAffx.79594.1.S1_s_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002318842</td><td align="center" valign="top" rowspan="1" colspan="1">8.00E&#x02013;164</td><td align="center" valign="top" rowspan="1" colspan="1">3.2</td><td align="left" valign="top" rowspan="1" colspan="1">Sorbitol dehydrogenase-like protein</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">PtpAffx.15690.1.S1_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002311043</td><td align="center" valign="top" rowspan="1" colspan="1">0.00E+00</td><td align="center" valign="top" rowspan="1" colspan="1">3.0</td><td align="left" valign="top" rowspan="1" colspan="1">Proline transporter</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">Ptp.3435.2.S1_s_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002302223</td><td align="center" valign="top" rowspan="1" colspan="1">0.00E+00</td><td align="center" valign="top" rowspan="1" colspan="1">2.6</td><td align="left" valign="top" rowspan="1" colspan="1">Amino acid permease</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">Ptp.1552.1.S1_s_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002302727</td><td align="center" valign="top" rowspan="1" colspan="1">0.00E+00</td><td align="center" valign="top" rowspan="1" colspan="1">2.1</td><td align="left" valign="top" rowspan="1" colspan="1">SUS3 (sucrose synthase 3)</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">Ptp.8110.1.S1_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002313213</td><td align="center" valign="top" rowspan="1" colspan="1">0.00E+00</td><td align="center" valign="top" rowspan="1" colspan="1">&#x02013;3.5</td><td align="left" valign="top" rowspan="1" colspan="1">Oligopeptide transporter</td></tr><tr><td align="left" valign="top" colspan="5" rowspan="1"><bold>DEFENSE</bold></td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">PtpAffx.77318.1.S1_x_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002312918</td><td align="center" valign="top" rowspan="1" colspan="1">1.00E&#x02013;114</td><td align="center" valign="top" rowspan="1" colspan="1">10.4</td><td align="left" valign="top" rowspan="1" colspan="1">Chitinase</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">PtpAffx.50871.1.A1_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002302379</td><td align="center" valign="top" rowspan="1" colspan="1">0.00E+00</td><td align="center" valign="top" rowspan="1" colspan="1">5.8</td><td align="left" valign="top" rowspan="1" colspan="1">Lipase</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">Ptp.6139.1.S1_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002304920</td><td align="center" valign="top" rowspan="1" colspan="1">0.00E+00</td><td align="center" valign="top" rowspan="1" colspan="1">4.8</td><td align="left" valign="top" rowspan="1" colspan="1">Cytochrome P450</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">PtpAffx.136901.1.S1_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002306296</td><td align="center" valign="top" rowspan="1" colspan="1">3.00E&#x02013;165</td><td align="center" valign="top" rowspan="1" colspan="1">3.6</td><td align="left" valign="top" rowspan="1" colspan="1">GCL1-like</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">Ptp.2230.1.S1_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002302409</td><td align="center" valign="top" rowspan="1" colspan="1">5.00E&#x02013;133</td><td align="center" valign="top" rowspan="1" colspan="1">3.2</td><td align="left" valign="top" rowspan="1" colspan="1">Sigma factor B</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">PtpAffx.55005.1.A1_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002322929</td><td align="center" valign="top" rowspan="1" colspan="1">0.00E+00</td><td align="center" valign="top" rowspan="1" colspan="1">2.8</td><td align="left" valign="top" rowspan="1" colspan="1">Glyoxal oxidase-related</td></tr><tr><td align="left" valign="top" colspan="5" rowspan="1"><bold>SENESCENCE</bold></td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">PtpAffx.57533.1.S1_a_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002320492</td><td align="center" valign="top" rowspan="1" colspan="1">6.00E&#x02013;165</td><td align="center" valign="top" rowspan="1" colspan="1">37.0</td><td align="left" valign="top" rowspan="1" colspan="1">Triacylglycerol lipase 2 precursor</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">Ptp.2629.1.S1_s_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002307593</td><td align="center" valign="top" rowspan="1" colspan="1">2.00E&#x02013;40</td><td align="center" valign="top" rowspan="1" colspan="1">2.7</td><td align="left" valign="top" rowspan="1" colspan="1">Senescence-associated protein-related</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">PtpAffx.85571.1.S1_s_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002299638</td><td align="center" valign="top" rowspan="1" colspan="1">1.00E&#x02013;31</td><td align="center" valign="top" rowspan="1" colspan="1">2.2</td><td align="left" valign="top" rowspan="1" colspan="1">Senescence-associated protein-related</td></tr><tr><td align="left" valign="top" colspan="5" rowspan="1"><bold>CELL CYCLE/EXPANSION</bold></td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">PtpAffx.222953.1.S1_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002318886</td><td align="center" valign="top" rowspan="1" colspan="1">0.00E+00</td><td align="center" valign="top" rowspan="1" colspan="1">&#x02013;2.3</td><td align="left" valign="top" rowspan="1" colspan="1">Trehalose-6-phosphate synthase</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">PtpAffx.50897.2.S1_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002310432</td><td align="center" valign="top" rowspan="1" colspan="1">3.00E&#x02013;55</td><td align="center" valign="top" rowspan="1" colspan="1">&#x02013;2.3</td><td align="left" valign="top" rowspan="1" colspan="1">Calmodulin 24-like protein</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">PtpAffx.200879.1.S1_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002298451</td><td align="center" valign="top" rowspan="1" colspan="1">2.00E&#x02013;156</td><td align="center" valign="top" rowspan="1" colspan="1">&#x02013;2.4</td><td align="left" valign="top" rowspan="1" colspan="1">Cyclin</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">Ptp.5638.1.S1_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002307791</td><td align="center" valign="top" rowspan="1" colspan="1">0.00E+00</td><td align="center" valign="top" rowspan="1" colspan="1">&#x02013;2.5</td><td align="left" valign="top" rowspan="1" colspan="1">Cyclin B</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">Ptp.7389.1.S1_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002319120</td><td align="center" valign="top" rowspan="1" colspan="1">0.00E+00</td><td align="center" valign="top" rowspan="1" colspan="1">&#x02013;2.5</td><td align="left" valign="top" rowspan="1" colspan="1">CDC20.1</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">PtpAffx.162051.1.S1_a_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002322260</td><td align="center" valign="top" rowspan="1" colspan="1">3.00E&#x02013;35</td><td align="center" valign="top" rowspan="1" colspan="1">&#x02013;2.8</td><td align="left" valign="top" rowspan="1" colspan="1">CDC2-like protein kinases</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">Ptp.1602.1.S1_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002307822</td><td align="center" valign="top" rowspan="1" colspan="1">1.00E&#x02013;177</td><td align="center" valign="top" rowspan="1" colspan="1">&#x02013;2.9</td><td align="left" valign="top" rowspan="1" colspan="1">Cyclin-dependent kinase</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">PtpAffx.63679.1.A1_s_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002306649</td><td align="center" valign="top" rowspan="1" colspan="1">0.00E+00</td><td align="center" valign="top" rowspan="1" colspan="1">&#x02013;4.0</td><td align="left" valign="top" rowspan="1" colspan="1">Cyclin A</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">Ptp.2869.1.A1_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002299019</td><td align="center" valign="top" rowspan="1" colspan="1">0.00E+00</td><td align="center" valign="top" rowspan="1" colspan="1">&#x02013;4.2</td><td align="left" valign="top" rowspan="1" colspan="1">Patellin-4</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">PtpAffx.206669.1.S1_s_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002308551</td><td align="center" valign="top" rowspan="1" colspan="1">0.00E+00</td><td align="center" valign="top" rowspan="1" colspan="1">&#x02013;4.6</td><td align="left" valign="top" rowspan="1" colspan="1">Calmodulin binding protein</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">PtpAffx.17914.3.A1_at</td><td align="left" valign="top" rowspan="1" colspan="1">XP_002312101</td><td align="center" valign="top" rowspan="1" colspan="1">7.00E&#x02013;140</td><td align="center" valign="top" rowspan="1" colspan="1">&#x02013;4.8</td><td align="left" valign="top" rowspan="1" colspan="1">Expansin</td></tr></tbody></table><table-wrap-foot><p><italic>p &#x0003c; 0.05; &#x0003e;2-fold differential regulation</italic>.</p></table-wrap-foot></table-wrap></sec></sec><sec sec-type="discussion" id="s4"><title>Discussion</title><p>Due to ornamental value and to wide-ranging applications within agriculture, the genetic traits that control cell size and dwarfism in plants have been widely studied (Valdovinos et al., <xref rid="B83" ref-type="bibr">1967</xref>; Ephritikhine et al., <xref rid="B21" ref-type="bibr">1999</xref>; Busov et al., <xref rid="B9" ref-type="bibr">2003</xref>; Qi et al., <xref rid="B62" ref-type="bibr">2011</xref>; Luo et al., <xref rid="B52" ref-type="bibr">2013</xref>; Li et al., <xref rid="B49" ref-type="bibr">2014</xref>; Yang et al., <xref rid="B91" ref-type="bibr">2014</xref>; Zhang et al., <xref rid="B93" ref-type="bibr">2014</xref>). Through breeding practices and applications of growth regulators, dozens of different dwarf plant varieties have been produced over the past few decades (Parker, <xref rid="B60" ref-type="bibr">2012</xref>; Jiang et al., <xref rid="B39" ref-type="bibr">2013</xref>; Wang et al., <xref rid="B85" ref-type="bibr">2014a</xref>,<xref rid="B86" ref-type="bibr">b</xref>). Largely studied in annual models such as <italic>Arabidopsis, Zea</italic> and <italic>Oryza</italic>, control of plant stature has been linked most readily to plant hormones. For example, <italic>Arabidopsis</italic> mutants with increased ethylene production (e.g., <italic>eto;</italic> Woeste et al., <xref rid="B88" ref-type="bibr">1999</xref>) exhibit thickening of the hypocotyl while increased ethylene signaling (e.g., <italic>ctr1-1; ACS6^DDD</italic>; Liu and Zhang, <xref rid="B50" ref-type="bibr">2004</xref>) has been found to result in reduced stature and smaller leaf size. Due to advancements in insertional mutagenesis and other transgenic technologies, it is now becoming feasible to also screen perennial plants for the genes that control dwarfism (Busov et al., <xref rid="B9" ref-type="bibr">2003</xref>; Harrison et al., <xref rid="B30" ref-type="bibr">2007</xref>; Vahala et al., <xref rid="B82" ref-type="bibr">2013</xref>). Here we characterize a mutant line of <italic>P. tremula x P. alba</italic> clone 717-1B4 that exhibits higher transcript accumulation of <italic>PtaACS8</italic> and that produces a significantly higher level of ethylene in all aerial tissues as compared to wild-type trees. Increased transcript abundance of <italic>PtaACS8</italic> is correlated to reduced plant stature and smaller leaves while treatment of <italic>dwarfy</italic> shoots induces increases in internode length. The transcriptional cascade induced by altered levels of <italic>PtaACS8</italic> are very different in stem and leaf tissue with a transcriptional reduction in genes associated with auxin transport and signaling evident in stems and repressed cell cycle genes in the leaves. These results place <italic>PtaACS8</italic>, and likely ethylene, as regulators that control two major morphological traits associated with dwarfism and reduced tissue size.</p><p>Using transformation technologies such as activation tagging is a very useful approach to identifying and characterizing the role of genes in a physiologically relevant manner. Rather than ectopic over-expression of a gene, the inserted enhancer used in activation tagging only enhances expression in its native expression pattern. This mutagenesis approach has been used in a number of model plant systems including <italic>Arabidopsis</italic> (Weigel et al., <xref rid="B87" ref-type="bibr">2000</xref>), tomato (Mathews et al., <xref rid="B53" ref-type="bibr">2003</xref>), rice (Jeong et al., <xref rid="B38" ref-type="bibr">2006</xref>), and poplar (Harrison et al., <xref rid="B30" ref-type="bibr">2007</xref>). Using this approach Busov et al. (<xref rid="B9" ref-type="bibr">2003</xref>) were able to identify a poplar <italic>GA2-OXIDASE</italic> that resulted in a plant with a very similar phenotype to that described here for <italic>dwarfy;</italic> reduced stature and smaller leaves. Since their publication, dwarfism in a native dwarf plum tree cultivar has also been linked to a <italic>GA2-OXIDASE</italic> (El-Sharkawy et al., <xref rid="B20" ref-type="bibr">2012</xref>) demonstrating that findings from activation tagging studies can be extended to natural plant populations.</p><p>As opposed to a strictly GA-dependent phenotype, our results support the hypothesis that growth retardation in <italic>dwarfy</italic> is driven largely by ethylene, the endpoint of the biochemical pathway in which <italic>PtaACS8</italic> operates. This is based on the evidence that increased <italic>PtaACS8</italic> transcripts are correlated to significant increases in ethylene production in the stem (Figure <xref ref-type="fig" rid="F1">1E</xref>), whose signal is being relayed by the activation of several ERF genes (Table <xref ref-type="table" rid="T1">1</xref>). Our results also demonstrate that blocking of ethylene biosynthesis resuces the <italic>dwarfy</italic> phenotype (Figures <xref ref-type="fig" rid="F4">4C&#x02013;E</xref>). Further, the reduction in xylem fiber and vessel length described here-in has also previously been observed after ethylene treatment of poplar stems (Junghans et al., <xref rid="B40" ref-type="bibr">2004</xref>; Love et al., <xref rid="B51" ref-type="bibr">2009</xref>; Vahala et al., <xref rid="B82" ref-type="bibr">2013</xref>). While we cannot rule out the possibility that the reduced stature in <italic>dwarfy</italic> is a result of increased ACC accumulation, our results support the hypothesis that stunting of the <italic>dwarfy</italic> stem is controlled in an ethylene-dependent manner. Increased ethylene, however, is likely not the only causative factor in explaining the stature of <italic>dwarfy</italic>. Rather, ethylene appears to be influencing another pathway associated with plant stature: the auxin pathway. We found evidence for a repression of auxin-homeostasis and transport genes in the stem of <italic>dwarfy</italic> (Table <xref ref-type="table" rid="T1">1</xref>). Ethylene has long been tied to a negative effect on auxin diffusion (von Guttenberg and Steinmetz, <xref rid="B84" ref-type="bibr">1947</xref>; Morgan and Gausman, <xref rid="B54" ref-type="bibr">1966</xref>; Valdovinos et al., <xref rid="B83" ref-type="bibr">1967</xref>; Suttle, <xref rid="B77" ref-type="bibr">1988</xref>; Andersson-Gunneras et al., <xref rid="B2" ref-type="bibr">2003</xref>; Ruzicka et al., <xref rid="B67" ref-type="bibr">2007</xref>; Stepanova et al., <xref rid="B75" ref-type="bibr">2007</xref>; Swarup et al., <xref rid="B78" ref-type="bibr">2007</xref>). As inhibition of auxin diffusion has been correlated to a reduction in stem cell elongation of poplar (Junghans et al., <xref rid="B40" ref-type="bibr">2004</xref>), pea (Lantican and Muir, <xref rid="B44" ref-type="bibr">1969</xref>), tomato (Higashide et al., <xref rid="B33" ref-type="bibr">2014</xref>), tulip (Okubo and Uemoto, <xref rid="B59" ref-type="bibr">1985</xref>), <italic>Arabidopsis</italic> (Franklin et al., <xref rid="B22a" ref-type="bibr">2011</xref>; Chae et al., <xref rid="B10" ref-type="bibr">2012</xref>), gourds (Wang et al., <xref rid="B85" ref-type="bibr">2014a</xref>) amongst many other systems. Our results give a molecular framework by which ethylene affects <italic>dwarfy</italic> height where increased expression of <italic>PtaACS8</italic> results in greater production of ethylene which, upon perception in plant stem tissue, represses genes related to auxin diffusion and synthesis which would then curtail cell elongation in the stem. GA treatment of growth apexes can also rescue the <italic>dwarfy</italic> phenotype, although we cannot conclude from present data if GA generates this phenotype by acting downstream of the ethylene signal in the <italic>dwarfy</italic> mutant or in a separate pathway.</p><p>A different genetic pathway is likely responsible for the observed reduction in leaf size in <italic>dwarfy</italic>. While increases in <italic>PtaACS8</italic> transcripts and ethylene evolution in the stem coincided with stunted fiber and vessel growth, no change in leaf epidermal cell size is observed despite higher levels of <italic>PtaACS8</italic> transcripts and higher ethylene evolution in the leaves. This would indicate that the leaf is smaller due to the absolute number of cells making up the tissue rather than the size of cell generated. It is interesting in the leaves of <italic>dwarfy</italic> that we see no evidence of compensation by leaf epidermal cells to maintain a larger leaf area. &#x0201c;Compensation&#x0201d; occurs when upstream inhibition of cell division initiates a secondary signaling pathway that increases cell size to maintain proper tissue growth (Hemerly et al., <xref rid="B32" ref-type="bibr">1995</xref>; DeVeylder et al., <xref rid="B14" ref-type="bibr">2001</xref>; Tsukaya, <xref rid="B80" ref-type="bibr">2002</xref>; Horiguchi et al., <xref rid="B34" ref-type="bibr">2006</xref>). Ethylene treatment has been associated with both stimulation of cell division (Love et al., <xref rid="B51" ref-type="bibr">2009</xref>) and inhibition of cellular division (Edwards and Miller, <xref rid="B19" ref-type="bibr">1972</xref>; Lee and LaRue, <xref rid="B46" ref-type="bibr">1992</xref>; Heidstra et al., <xref rid="B31" ref-type="bibr">1997</xref>; Dubois et al., <xref rid="B18" ref-type="bibr">2013</xref>; Luo et al., <xref rid="B52" ref-type="bibr">2013</xref>). In <italic>Arabidopsis</italic>, ethylene has been associated with reduced petal and leaf size in mutants with constitutive ethylene signaling (Kieber et al., <xref rid="B42" ref-type="bibr">1993</xref>; Roman and Ecker, <xref rid="B65" ref-type="bibr">1995</xref>; Luo et al., <xref rid="B52" ref-type="bibr">2013</xref>) and under water limiting conditions due to the activity of ERF6 through its control of GA2-OXIDASE (Dubois et al., <xref rid="B18" ref-type="bibr">2013</xref>). In the transcriptomic analysis of <italic>dwarfy</italic> leaves we do not see evidence of either <italic>ERF</italic> or <italic>GA2-OXIDASE</italic> genes accumulating at altered abundances. Rather, within the group of genes regulated in <italic>dwarfy</italic> leaves, we observed the repression of a large class of cell cycle genes including <italic>CYCLIN-DEPENDENT KINASE1</italic> (<italic>CDK1</italic>), <italic>CYCLIN A</italic>, and <italic>CYCLIN B1</italic> (Table <xref ref-type="table" rid="T2">2</xref>). In eukaryotic cells, CYCLIN A initiates the cellular transition from G2 to prophase after which CYCLIN B1 enters the nucleus and, together with CDK1, induces mitosis by phosphorylation and activation of enzymes regulating chromatin condensation, nuclear membrane breakdown and mitosis-specific microtubule and microfilament re-orientation (Nigg, <xref rid="B57" ref-type="bibr">2001</xref>; Smits and Medema, <xref rid="B73" ref-type="bibr">2001</xref>; Gavet and Pines, <xref rid="B23" ref-type="bibr">2010</xref>; Suryadinata et al., <xref rid="B76" ref-type="bibr">2010</xref>; Rattani et al., <xref rid="B64" ref-type="bibr">2014</xref>). As this whole suite of proteins is necessary for cellular division, repression of their transcription in the leaves of <italic>dwarfy</italic>, as compared to wild-type leaves, is likely the key pathway by which leaf size is being affected. These results are reminiscent of earlier observations that ethylene in <italic>Pisum sativum</italic> stopped cell division prior to entry into prophase (Apelbaum and Burg, <xref rid="B3" ref-type="bibr">1972</xref>).</p><p>The leaf drop date of natural grown-year old dwarfy and wild-type plants was assessed as increased ethylene levels have been correlated to early leaf senescence (Breeze et al., <xref rid="B7" ref-type="bibr">2011</xref>; Koyama et al., <xref rid="B43" ref-type="bibr">2013</xref>). Leaf yellowing, considered to be the first visible senescent event (Quirino et al., <xref rid="B63" ref-type="bibr">2000</xref>) was present in only the dwarfy basal leaves in October (Figure <xref ref-type="fig" rid="F6">6C</xref>), while in the wild-type, senescence related-color changes were prevalent in November basal leaves (Figure <xref ref-type="fig" rid="F6">6D</xref>) by which time dwarfy basal leaves had dehisced. Buchanan-Wollaston et al. (<xref rid="B8" ref-type="bibr">2003</xref>) noted that plants exposed to exogenous ethylene do exhibit premature senescence with the older leaves yellowing first; similar to the results here-in. November dwarfy and wild-type plants both had green apical leaves suggesting the onset of senescence and leaf dehiscence in all but the apical leaves in the dwarfy mutant were altered.</p><p>Our results support the hypothesis that there are two different developmental programs regulating tissue size in the <italic>dwarfy</italic> mutant. In the stem, we observe alterations to ethylene response factors and an inhibition of auxin homeostasis genes suggesting that ethylene inhibits stem elongation as previously observed in model organisms such as <italic>Arabidopsis</italic> (Guzman and Ecker, <xref rid="B29" ref-type="bibr">1990</xref>), poplar (Junghans et al., <xref rid="B40" ref-type="bibr">2004</xref>; Love et al., <xref rid="B51" ref-type="bibr">2009</xref>; Vahala et al., <xref rid="B82" ref-type="bibr">2013</xref>), tobacco (Romano et al., <xref rid="B66" ref-type="bibr">1993</xref>), and tomato (Huang and Lin, <xref rid="B36" ref-type="bibr">2003</xref>) through its influence on the auxin pathway. In the leaves, we find that the leaves of <italic>dwarfy</italic> produce fewer cells and are, thereby, smaller. This phenotype is likely tied to the differential expression of the protein group responsible for the induction of mitosis. Altogether, our study of the <italic>dwarfy</italic> mutant poplar has given insight into the genetics underpinning ethylene-induced dwarfism.</p><sec><title>Conflict of interest statement</title><p>The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.</p></sec></sec>

Circadian rhythms, Wnt/beta-catenin pathway and PPAR alpha/gamma profiles in diseases with primary or secondary cardiac dysfunction

<p>Circadian clock mechanisms are far-from-equilibrium dissipative structures. Peroxisome proliferator-activated receptors (PPAR alpha, beta/delta, and gamma) play a key role in metabolic regulatory processes, particularly in heart muscle. Links between circadian rhythms (CRs) and PPARs have been established. Mammalian CRs involve at least two critical transcription factors, CLOCK and BMAL1 (Gekakis et al., <xref rid="B51" ref-type="bibr">1998</xref>; Hogenesch et al., <xref rid="B70" ref-type="bibr">1998</xref>). PPAR gamma plays a major role in both glucose and lipid metabolisms and presents circadian properties which coordinate the interplay between metabolism and CRs. PPAR gamma is a major component of the vascular clock. Vascular PPAR gamma is a peripheral regulator of cardiovascular rhythms controlling circadian variations in blood pressure and heart rate through BMAL1. We focused our review on diseases with abnormalities of CRs and with primary or secondary cardiac dysfunction. Moreover, these diseases presented changes in the Wnt/beta-catenin pathway and PPARs, according to two opposed profiles. Profile 1 was defined as follows: inactivation of the Wnt/beta-catenin pathway with increased expression of PPAR gamma. Profile 2 was defined as follows: activation of the Wnt/beta-catenin pathway with decreased expression of PPAR gamma. A typical profile 1 disease is arrhythmogenic right ventricular cardiomyopathy, a genetic cardiac disease which presents mutations of the desmosomal proteins and is mainly characterized by fatty acid accumulation in adult cardiomyocytes mainly in the right ventricle. The link between PPAR gamma dysfunction and desmosomal genetic mutations occurs via inactivation of the Wnt/beta-catenin pathway presenting oscillatory properties. A typical profile 2 disease is type 2 diabetes, with activation of the Wnt/beta-catenin pathway and decreased expression of PPAR gamma. CRs abnormalities are present in numerous pathologies such as cardiovascular diseases, sympathetic/parasympathetic dysfunction, hypertension, diabetes, neurodegenerative diseases, cancer which are often closely inter-related.</p>

<contrib contrib-type="author"><name><surname>Lecarpentier</surname><given-names>Yves</given-names></name><xref ref-type="aff" rid="aff1">¹</xref><xref ref-type="author-notes" rid="fn001">^*</xref><uri xlink:type="simple" xlink:href="http://community.frontiersin.org/people/u/139386"/></contrib><contrib contrib-type="author"><name><surname>Claes</surname><given-names>Victor</given-names></name><xref ref-type="aff" rid="aff2">²</xref></contrib><contrib contrib-type="author"><name><surname>Duthoit</surname><given-names>Guillaume</given-names></name><xref ref-type="aff" rid="aff3">³</xref></contrib><contrib contrib-type="author"><name><surname>H&#x000e9;bert</surname><given-names>Jean-Louis</given-names></name><xref ref-type="aff" rid="aff3">³</xref></contrib>

Frontiers in Physiology

<sec sec-type="intro" id="s1"><title>Introduction</title><p>CRs are biological temporal processes that display endogenous, entrainable free-running periods that last approximately 24 h. They are driven by molecular internal clocks which can be reset by environmental light-dark cycles. Circadian clocks are transcriptionally based molecular mechanisms which comprise feedback loops (Edery, <xref rid="B40" ref-type="bibr">2000</xref>). The molecular basis of CRs have first been clarified in Drosophila and Neurospora, then in cyanobacteria, plants, and mammals (Reppert and Weaver, <xref rid="B136" ref-type="bibr">2002</xref>). All living organisms adjust their physiology and behavior to the 24-h day-night cycle under the governance of circadian clocks. Circadian clocks may provide a selective advantage of anticipation, thus allowing organisms to respond efficiently to various stimuli at the appropriate time. In mammals, sleep-awake and feeding patterns, hormone secretion, heart rate, blood pressure, energy metabolism, and body temperature exhibit CRs. Their disruptions may have deleterious effects. People submitted to shift working, frequent transmeridian air flight, exposure to artificial light exhibit a particularly high incidence of metabolic syndrome and obesity. CR dysfunctions in blood pressure and heart rate, which are both partly regulated by PPAR gamma are involved in arrhythmias which may lead to sudden cardiac death, myocardial infarction or stroke, often occurring at the early morning during the surge in blood pressure. CRs are dissipative structures due to a negative feedback produced by a protein on the expression of its own gene (Goodwin, <xref rid="B57" ref-type="bibr">1965</xref>; Hardin et al., <xref rid="B64" ref-type="bibr">1990</xref>). They operate far-from- equilibrium and generate order spontaneously by exchanging energy with their external environment (Prigogine et al., <xref rid="B132" ref-type="bibr">1974</xref>; Goldbeter, <xref rid="B55" ref-type="bibr">2002</xref>; Lecarpentier et al., <xref rid="B91" ref-type="bibr">2010</xref>).</p></sec><sec><title>The regulatory sites of circadian rhythms</title><p>The master regulator site of CRs is the suprachiasmatic nucleus (SCN) inside the hypothalamus in which core clock genes are rhythmically expressed (Weaver, <xref rid="B171" ref-type="bibr">1998</xref>). In addition to this central clock, each organ has its own biological clock system, termed peripheral clock. The SCN and most peripheral tissues such as heart, blood vessels, skeletal muscles, kidneys, liver, and fat, govern numerous functions that are synchronized with the sleep-awake cycle (Zylka et al., <xref rid="B191" ref-type="bibr">1998</xref>). In the cardiovascular system, circadian clocks have been described within numerous mammalian cells, such as cardiomyocytes, vascular smooth muscle cells, endothelial cells, and fibroblasts (McNamara et al., <xref rid="B110" ref-type="bibr">2001</xref>; Nonaka et al., <xref rid="B121" ref-type="bibr">2001</xref>; Durgan et al., <xref rid="B37" ref-type="bibr">2005</xref>, <xref rid="B38" ref-type="bibr">2007</xref>; Takeda et al., <xref rid="B159" ref-type="bibr">2007</xref>). Peripheral clocks have their own regulatory mechanisms, which are specific to each peripheral organ by regulating the expression of clock-controlled genes (<italic>Ccg</italic>). CRs have been demonstrated in approximately 8&#x02013;10% of total genes expressed in mouse heart and liver, more than 90% of them depending on self-autonomous local diurnal oscillators (Storch et al., <xref rid="B153" ref-type="bibr">2002</xref>).</p></sec><sec><title>Genes and proteins</title><p>Important genes are involved in CRs including <italic>Clock</italic> (<italic>Circadian locomotor output cycles kaput</italic>), <italic>Bmal1</italic> (<italic>brain and muscle aryl-hydrocarbon receptor nuclear translocator-like 1</italic>), <italic>Cry1</italic> (<italic>cryptochrome 1</italic>), <italic>Cry2</italic> (<italic>cryptochrome 2</italic>), <italic>Per1</italic> (<italic>Period 1</italic>), <italic>Per2</italic> (<italic>Period 2</italic>), <italic>Per3</italic> (<italic>Period 3</italic>), and <italic>Ccg</italic>. They organize transcription/translation autoregulatory feedback loops comprising both activating and inhibiting pathways (Reppert and Weaver, <xref rid="B136" ref-type="bibr">2002</xref>; Schibler and Sassone-Corsi, <xref rid="B145" ref-type="bibr">2002</xref>). A complex network is formed by all these genes which interlock feedback and forward subtle loops whose complete time course is approximately 24 h. Clock genes Per1, Per2, Bmal1, and Cry1 display rhythmic expression in human hearts (Leibetseder et al., <xref rid="B93" ref-type="bibr">2009</xref>). At the start of the day, transcription of Clock and Bmal1 begins. The proteins CLOCK and BMAL1 are synthesized and then associate as dimers which bind to regulatory DNA sequences (E-box elements) of the promoters of target genes. CLOCK: BMAL1 dimer activates circadian gene transcription of Period genes (Per1, 2, and 3), Chryptochrome genes (Cry 1 and 2), Rev-erb, Ror (related orphan receptor), and Ccg, and drives their rhythmic expression (Reppert and Weaver, <xref rid="B135" ref-type="bibr">2001</xref>, <xref rid="B136" ref-type="bibr">2002</xref>; Young and Kay, <xref rid="B185" ref-type="bibr">2001</xref>; Canaple et al., <xref rid="B19" ref-type="bibr">2006</xref>; Chen and Yang, <xref rid="B23" ref-type="bibr">2014</xref>). Into the cytoplasm, the PER and CRY proteins dimerize and, after translocation to the nucleus, modulate the transcriptional activity of CLOCK: BMAL1 (Kume et al., <xref rid="B89" ref-type="bibr">1999</xref>). Concentrations of BMAL1 and PER proteins cycle in counterpoint. PER2 is a positive regulator of the Bmal1 loop. Protein CRY is a negative regulator of both Per and Cry loops. ROR alpha enhances Bmal1 transcription (Akashi and Takumi, <xref rid="B1" ref-type="bibr">2005</xref>), while the nuclear receptor Rev-erb inhibits it (Ueda et al., <xref rid="B161" ref-type="bibr">2002</xref>). Rev-erb alpha binds to ROR-responsive element (RORE) in the Bmal1 promoter and represses its transcriptional activity (Preitner et al., <xref rid="B131" ref-type="bibr">2002</xref>). Rev-erb alpha protein is a member of the nuclear receptor family of intracellular transcription factors. The gene Rev-erb alpha is a major regulatory component of the circadian clock (Yin et al., <xref rid="B179" ref-type="bibr">2006</xref>) and among various properties, is involved in the circadian expression of plasminogen activator inhibitor type 1 (Wang et al., <xref rid="B168" ref-type="bibr">2006</xref>).</p><sec><title>Circadian rhythms and mutations of genes</title><p>Mutations or deletions of clock genes in mice have shown the key role of circadian clocks to ensure the proper timing of metabolic and cardiovascular processes. There is an increased pathological remodeling and vascular injury together with an aberrant CR in <italic>Bmal1</italic>-knockout and <italic>Clock</italic> mutant mice (Anea et al., <xref rid="B5" ref-type="bibr">2009</xref>). Aortas from <italic>Bmal1</italic>-knockout and <italic>Clock</italic> mutant mice exhibit endothelial dysfunction. Akt (protein kinase B) and subsequent nitric oxide signaling is significantly attenuated in arteries from <italic>Bmal1</italic>-knockout mice. <italic>Bmal1</italic> is a key regulator of myogenesis which may represent a temporal regulatory mechanism to fine-tune myocyte differentiation (Chatterjee et al., <xref rid="B22" ref-type="bibr">2013</xref>). <italic>Bmal1</italic> regulates adipogenesis <italic>via the</italic> Wnt signaling pathway (Guo et al., <xref rid="B60" ref-type="bibr">2012</xref>). Disruption of <italic>Bmal1</italic> in mice led to increased adipogenesis, adipocyte hypertrophy, and obesity. Attenuation of <italic>Bmal1</italic> function resulted in down-regulation of genes in the canonical Wnt pathway known to suppress adipogenesis. Promoters of these genes, i.e., <italic>beta-catenin</italic>, Disheveled <italic>(Dsh)</italic>, T cell-enhancing binding <italic>(Tcf)</italic> display <italic>Bmal1</italic> occupancy, indicating direct circadian regulation by Bmal1. Among several abnormalities, deletion of the clock gene <italic>Bmal1</italic> in mice adipose tissue induces obesity (Paschos et al., <xref rid="B126" ref-type="bibr">2012</xref>). The cardiomyocyte-specific clock mutant <italic>(Ccm)</italic> is a mouse model wherein the cardiomyocyte circadian clock is selectively suppressed (Young et al., <xref rid="B182" ref-type="bibr">2001b</xref>,<xref rid="B183" ref-type="bibr">c</xref>; Durgan et al., <xref rid="B39" ref-type="bibr">2006</xref>). <italic>Ccm</italic> presents a temporal suspension of the cardiomyocyte circadian clock at the wake-to-sleep transition (Young, <xref rid="B180" ref-type="bibr">2009</xref>). Numerous mutations of genes will be discussed in the following paragraphs of this review.</p></sec><sec><title>Circadian rhythms and heart performance</title><p>Loss of synchronization between the internal clock and external stimuli can induce cardiovascular organ damage. Discrepancy in the phases between the central and peripheral clocks also seems to contribute to progression of cardiovascular disorders (Takeda and Maemura, <xref rid="B158" ref-type="bibr">2011</xref>). Peripheral clocks have their own roles specific to each peripheral organ by regulating the expression of <italic>Ccg</italic>, although the oscillation mechanisms of the peripheral clock are similar to that of the SCN (Takeda et al., <xref rid="B159" ref-type="bibr">2007</xref>). Both the physiological and pathological functions of cardiovascular organs are closely related to CRs. Heart rate, blood pressure and endothelial function show diurnal variations within a day. A profound pattern exists in the time of day at which the death may occur (Takeda and Maemura, <xref rid="B158" ref-type="bibr">2011</xref>). The onset of cardiovascular disorders such as acute coronary syndrome, atrial arrhythmias, and subarachnoid hemorrhage exhibits impairment of diurnal oscillations. Stroke and heart attacks most frequently happen in the morning when blood pressure surges.</p><p>Over the course of the day, the normal heart anticipates, responds and adapts to physiological alterations within its environment. Contractile performance, carbohydrate oxidation, fatty acid oxidation (FAO), mitochondrial function, oxygen consumption, and expression of all metabolic genes show diurnal variations. The circadian clock plays an important role in cardiac homeostasis through the anticipation of daily workload. In wild-type mice, the ejection fraction (EF) and the shortening fraction (FS) show circadian variation (Wu et al., <xref rid="B173" ref-type="bibr">2011</xref>). The diurnal variations in EF and FS are altered in mice with disruptions of circadian clock genes and are significantly diminished under an imposed light regimen. The circadian variation in blood pressure and heart rate is disrupted in <italic>Bmal1</italic>(-/-) and <italic>Clock</italic> (mut) mice in which core clock genes are deleted or mutated (Curtis et al., <xref rid="B29" ref-type="bibr">2007</xref>). <italic>Bmal1</italic> deletion abolishes the 24-h frequency in cardiovascular rhythms. However, a shorter ultradian rhythm remains. Sympathetic adrenal function is disrupted in these mice.</p></sec><sec><title>Peroxisome proliferator-activated receptors (PPARs)</title><p>PPARs (alpha, beta/delta, and gamma) are nuclear receptors belonging to the nuclear receptor superfamily. They function as transcription factors within the cell nuclei and regulate the expression of several target genes. PPARs play a pivotal role in various physiological and pathological processes, especially in energy metabolism, development, carcinogenesis, extracellular matrix remodeling, and CRs (Lockyer et al., <xref rid="B101" ref-type="bibr">2009</xref>). PPARs heterodimerize with the retinoid X receptor (RXR). PPARs are activated by their respective ligands, either endogenous fatty acids or pharmaceutical drugs which are potential therapeutic agents. Numerous natural and synthetic compounds, i.e., fatty acids, eicosanoids, arachidonic acid, hypolipidemic fibrates activating PPAR alpha, and anti diabetic thiazolidinediones (TZD) activating PPAR gamma, serve as activators of PPARs. PPARs are involved in numerous pathologies such as obesity, dyslipidemia, insulin resistance, type 2 diabetes, hypertension, cardiac hypertrophy (Berger and Moller, <xref rid="B12" ref-type="bibr">2002</xref>; Kelly, <xref rid="B81" ref-type="bibr">2003</xref>). PPAR beta/delta was not studied in this review.</p></sec><sec><title>PPARs and circadian rhythms</title><p>PPARs integrate the mammalian clock and energy metabolism (Chen and Yang, <xref rid="B23" ref-type="bibr">2014</xref>). PPARs have been shown to be rhythmically expressed in mammalian tissues (Yang et al., <xref rid="B177" ref-type="bibr">2006</xref>) and to directly interact with the core clock genes (Inoue et al., <xref rid="B74" ref-type="bibr">2005</xref>). PPAR beta/delta has not been studied in this review.</p><sec><title>PPAR alpha</title><p>PPAR alpha presents CRs in several organs, i.e. heart, muscles, liver, and kidney (Lemberger et al., <xref rid="B94" ref-type="bibr">1996</xref>; Yang et al., <xref rid="B177" ref-type="bibr">2006</xref>). PPAR alpha expression is stimulated by stress, glucocorticoid hormones, and insulin whose secretion follows CRs (Lemberger et al., <xref rid="B95" ref-type="bibr">1994</xref>). Importantly, PPAR alpha is a direct target of genes (<italic>Bmal1</italic> and <italic>Clock</italic>) through an E-box process (Oishi et al., <xref rid="B122" ref-type="bibr">2005</xref>). The circadian expression of PPAR alpha mRNA is abolished in the liver of homozygous <italic>Clock</italic> mutant mice and is regulated by the peripheral oscillators in a CLOCK-dependent mechanism. In rodent liver, there is a regulatory feedback loop involving BMAL1 and PPAR alpha in peripheral clocks. This regulation occurs via a direct binding of PPAR alpha on a PPAR alpha response element located in the <italic>Bmal1</italic> gene promoter. Moreover, BMAL1 is an upstream regulator of <italic>PPAR alpha</italic> gene expression (Gervois et al., <xref rid="B52" ref-type="bibr">1999</xref>; Canaple et al., <xref rid="B19" ref-type="bibr">2006</xref>). Several genes such as those encoding for sterol regulatory element binding protein, HMG-CoA reductase, fatty acid synthase, are involved in the lipid metabolism. They display circadian fluctuations, and their activities are diminished or suppressed in <italic>PPAR alpha</italic> knockout mice (Patel et al., <xref rid="B128" ref-type="bibr">2001</xref>; Gibbons et al., <xref rid="B53" ref-type="bibr">2002</xref>). PPAR alpha directly regulates the transcriptional activity of <italic>Bmal1</italic> and <italic>Rev-erb alpha</italic> through the PPRE located in the promoter site of their respective genes. <italic>Per2</italic> interacts with nuclear receptors including PPAR alpha and Rev-Erb alpha and serves as a co-regulator of nuclear receptor-mediated transcription. The PPAR alpha agonist fenofibrate increases transcription and resets circadian expression of <italic>Bmal1, Per2, and Rev-erb alpha</italic> in mouse liver (Canaple et al., <xref rid="B19" ref-type="bibr">2006</xref>) and cultured hepatocytes (Gervois et al., <xref rid="B52" ref-type="bibr">1999</xref>). Moreover, bezafibrate can phase advance the rhythmic expression of <italic>Bmal1, Per2, and Rev-erb alpha</italic> in several mouse peripheral tissues (Shirai et al., <xref rid="B149" ref-type="bibr">2007</xref>; Oishi et al., <xref rid="B123" ref-type="bibr">2008</xref>).</p></sec><sec><title>PPAR gamma</title><p>PPAR gamma exhibits variations in diurnal expression in mouse fat, liver, and blood vessels (Yang et al., <xref rid="B177" ref-type="bibr">2006</xref>; Wang et al., <xref rid="B169" ref-type="bibr">2008</xref>). Deletion of <italic>PPAR gamma</italic> in mouse suppresses or diminishes diurnal rhythms (Yang et al., <xref rid="B175" ref-type="bibr">2012</xref>). CRs have been analyzed in two strains of whole-body <italic>PPAR gamma null</italic> mouse models, i.e., <italic>Mox2-Cre</italic> mice (<italic>MoxCre/flox</italic>) or induced by tamoxifen (<italic>EsrCre/flox/TM</italic>). Diurnal variations in blood pressure and heart rate are blunted <italic>in MoxCre/flox</italic> mice. Impaired rhythmicity of the canonical clock genes is observed in adipose tissue and liver. This shows the important role of PPAR gamma in the coordinated control of circadian clocks, metabolism, and cardiac performance (Yang et al., <xref rid="B175" ref-type="bibr">2012</xref>). Moreover, insulin resistance is correlated with a non-dipper type&#x02014;i.e., with no blood pressure decrease during the circadian cycle- in essential hypertension. TZD are oral hypoglycemic agents act as insulin sensitizers and possess antihypertensive properties. TZD therapy with pioglitazone transforms the CR of blood pressure from a non-dipper to a dipper type (Anan et al., <xref rid="B2" ref-type="bibr">2007</xref>). PPAR gamma contributes to maintain the diurnal variations of both blood pressure and heart rate.</p><p>Rev-Erb alpha, an orphan nuclear receptor and a core clock component, is expressed after PPAR gamma activation with rosiglitazone in rat. Activated PPAR gamma induces <italic>Rev-Erb alpha</italic> promoter activity by binding to the response element <italic>Rev-DR2</italic>. Mutations of the 5&#x02032; or 3&#x02032; half-sites of the response element suppress PPAR gamma binding and transcriptional activation (Fontaine et al., <xref rid="B45" ref-type="bibr">2003</xref>). PGC-1 alpha, a transcriptional co-activator that regulates energy metabolism, is rhythmically expressed in the liver and skeletal muscle of mice. PGC-1 alpha stimulates the expression of clock genes, notably <italic>Bmal1</italic> and <italic>Rev-erb alpha</italic>, through co-activation of the ROR family of orphan nuclear receptors. Mice lacking PGC-1 alpha show abnormal CRs of activity, body temperature, and metabolic rate (Liu et al., <xref rid="B99" ref-type="bibr">2007a</xref>). <italic>Nocturnin</italic>, a circadian-regulated gene, promotes adipogenesis by stimulating PPAR gamma nuclear translocation (Kawai et al., <xref rid="B80" ref-type="bibr">2010</xref>). Nocturnin binds to PPAR gamma and stimulates its transcriptional activity whereas its deletion suppresses PPAR gamma oscillations (Green et al., <xref rid="B59" ref-type="bibr">2007</xref>). The hormone-dependent interaction of the nuclear receptor RXR alpha with CLOCK negatively regulates CLOCK: BMAL1-mediated transcriptional activation of clock gene expression in vascular cells. RXR alpha can phase shift Per2 mRNA rhythmicity, providing a molecular mechanism for hormonal control of clock gene expression (McNamara et al., <xref rid="B110" ref-type="bibr">2001</xref>).</p></sec></sec><sec><title>Canonical Wnt/beta-catenin pathway</title><p>Beta-catenin plays a key role during epithelial-mesenchymal transition (EMT), that characterizes normal embryonic development, tissue regeneration and cancer proliferation (Heuberger and Birchmeier, <xref rid="B68" ref-type="bibr">2010</xref>). Beta-catenin is a normal constituent of the zonula adherens, a major cell-to-cell adhesion complex in pavement-like tissues. During EMT, the loss of cadherins disrupts the zonula adherens, thus liberating beta-catenin into the cytoplasm. This molecule then migrates to the cell nucleus where it activates the Wnt/beta-catenin target genes. A hallmark of the canonical Wnt pathway activation is the elevation of cytoplasmic beta-catenin protein levels, the subsequent nuclear translocation and further activation of beta-catenin specific gene transcription (Ben-Ze'ev and Geiger, <xref rid="B11" ref-type="bibr">1998</xref>; Klymkowsky et al., <xref rid="B84" ref-type="bibr">1999</xref>; Zhurinsky et al., <xref rid="B190" ref-type="bibr">2000</xref>; Moon et al., <xref rid="B114" ref-type="bibr">2002</xref>; Maeda et al., <xref rid="B105" ref-type="bibr">2004</xref>; Sen-Chowdhry et al., <xref rid="B148" ref-type="bibr">2005</xref>; Garcia-Gras et al., <xref rid="B49" ref-type="bibr">2006</xref>), (Figure <xref ref-type="fig" rid="F1">1</xref>). In the absence of Wnt ligands, beta-catenin is recruited into a destruction complex that contains adenomatous polyposis coli (APC) and Axin, which facilitate the phosphorylation of beta-catenin by glycogen synthase kinase 3- beta (GSK3-beta). GSK3-beta phosphorylates the N-terminal domain of beta-catenin, thereby targeting it for ubiquitination and proteasomal degradation. In the presence of a Wnt ligand, the binding of Wnt to Frizzled (Fzd) leads to activation of the phosphoprotein Disheveled (Dsh). Dsh recruits Axin and the destruction complex to the plasma membrane, where Axin directly binds to the cytoplasmic tail of LRP5/6. Axin is degraded, which decreases beta-catenin degradation. The activation of Dsh also leads to the inhibition of GSK3-beta by phosphorylation, which further reduces the phosphorylation and degradation of beta-catenin. The beta-catenin degradation complex is inactivated with recruitment of axin to the plasma membrane, thus stabilizing the non-phosphorylated beta-catenin which translocates to the nucleus. Beta-catenin binds to T cell/lymphoid-enhancing binding (Tcf/Lef) transcription factors. The resulting complex becomes active by displacing Grouchos, leading to activation of numerous target genes.</p><fig id="F1" position="float"><label>Figure 1</label><caption><p><bold>The Wnt/beta-catenin pathway. (A)</bold> In the absence of Wnt, cytosolic beta-catenin is phosphorylated by GSK3 beta. APS and AXIN complex with GSK3 beta and beta-catenin to enhance the destruction process into the proteasome. Phosphorylated beta-catenin is recognized by the ubiquitin ligase beta -TrCP, ubiquinated and degraded. The Wnt pathway is in an &#x0201c;off state.&#x0201d; <bold>(B)</bold> In the presence of Wnt, Wnt binds both Frizzled and LRP5/6 receptors to initiate GRK5/6-mediated LRP phosphorylation and disheveled-mediated Frizzled internalization. Disheveled membrane translocation leads to dissociation of the AXIN/APC/GSK3 beta complex. Beta-catenin phosphorylation is inhibited and accumulates into the cytosol. Beta-catenin then translocates to the nucleus to bind Lef-Tcf co-transcription factors, which induces the Wnt-response gene transcription. Abbreviations: APC, adenomatous polyposis coli; Dsh, Disheveled; GSK3 beta, glycogen synthase kinase 3 beta; LRP5/6, low density lipoprotein receptor-related protein 5/6; Fzd, Frizzled.</p></caption><graphic xlink:href="fphys-05-00429-g0001"/></fig></sec><sec><title>Canonical Wnt/beta-catenin pathway and PPAR gamma</title><p>Numerous studies have shown the direct interaction between beta-catenin and PPAR gamma (Moldes et al., <xref rid="B113" ref-type="bibr">2003</xref>; Jansson et al., <xref rid="B78" ref-type="bibr">2005</xref>; Garcia-Gras et al., <xref rid="B49" ref-type="bibr">2006</xref>). PPAR gamma activation inhibits the beta-catenin activation of Tcf/Lef transcription factors (Lu and Carson, <xref rid="B103" ref-type="bibr">2010</xref>). The TZD PPAR gamma agonists troglitazone, rosiglitazone, and pioglitazone, and the non-TZD PPAR gamma activator GW1929 inhibit the beta-catenin-induced transcription in a PPAR gamma dependent manner. Activation of the Wnt-beta catenin pathway leads to osteogenesis, not adipogenesis and its inhibition leads to an increase in transcription of PPAR gamma. Osteogenic pathway is linked to the stimulation of Wnt signal leading to the final transcriptional activation of early osteogenic markers such as RUNX-2 and ALP, mediated by beta-catenin. Conversely, the adipogenic pathway involves inhibition of Wnt pathway leading to ubiquitination/degradation of beta-catenin which results in the transcription of PPAR gamma, a pivotal initiator of adipogenesis. The canonical Wnt/beta-catenin-PPAR gamma system determines the molecular switching of osteablastogenesis vs. adipogenesis (Takada et al., <xref rid="B157" ref-type="bibr">2009</xref>). PPAR gamma is a prime inducer of adipogenesis that inhibits osteoblastogenesis. Two different pathways switch the cell fate decision from adipocytes to osteoblasts by suppressing the transactivation function of PPAR gamma. TNF-alpha- and IL-1-induced TAK1/TAB1/NIK signaling cascade attenuate PPAR gamma-mediated adipogenesis by inhibiting the binding of PPAR gamma to the DNA response element. PPAR gamma suppresses Wnt/beta-catenin signaling during adipogenesis (Moldes et al., <xref rid="B113" ref-type="bibr">2003</xref>). Wnt/beta-catenin pathway operates to maintain the undifferentiated state of preadipocytes by inhibiting adipogenic gene expression. Importantly, there is a reciprocal relationship between beta-catenin expression and PPAR gamma activity.</p></sec><sec><title>Diseases associated with deactivation of the Wnt/beta-catenin pathway and increased expression of PPAR gamma</title><p>Numerous diseases present a common denominator: activation of the Wnt/beta-catenin pathway decreased and the expression of PPAR gamma increased. In most cases, expression of PPAR alpha decreases.</p><sec><title>Arrhythmogenic right ventricular cardiomyopathy (ARVC)</title><p>ARVC is a rare human disease characterized by the development of a fibro-fatty tissue in both ventricles, prominently involving the right ventricular (RV) myocardium (Marcus et al., <xref rid="B107" ref-type="bibr">1982</xref>; Fontaine et al., <xref rid="B46" ref-type="bibr">1999</xref>) (Figure <xref ref-type="fig" rid="F2">2</xref>). Cardiac dysfunction progressively develops, initially located at the RV and becoming biventricular in about 20% of cases (Richardson et al., <xref rid="B137" ref-type="bibr">1996</xref>; Hebert et al., <xref rid="B65" ref-type="bibr">2004</xref>). ARVC is most often an autosomic family-related disease. Genetic mutations have been identified in about 50% of cases, occurring among the five desmosomal proteins so far identified in the ventricular cardiomyocyte, i.e., desmoglein 2 (DSG2), desmocollin 2 (DSC2), plakophilin 2 (PKP2), plakoglobin (PG), and desmoplakin (DSP) (Basso et al., <xref rid="B8" ref-type="bibr">2009</xref>; Fressard et al., <xref rid="B47" ref-type="bibr">2010</xref>). PPAR abnormalities have been reported in ARVC with an increase in PPAR gamma and a decrease in PPAR alpha in RV (Djouadi et al., <xref rid="B32" ref-type="bibr">2009</xref>).</p><fig id="F2" position="float"><label>Figure 2</label><caption><p><bold>Arrhythmogenic right ventricular cardiomyopathy (ARVC) histology. (A)</bold> Typical morphology of right ventricular transmural free wall section in a terminal ARVC heart transplant specimen, showing extensive fibro-fatty replacement. A mid-mural residual muscular core (black asterisk) is well-identified. Fibrosis is prominently located at the subendocardium. Note the layer of normal subepicardial fat (Hematoxylin Eosin Saffron staining, original magnification &#x000d7; 10). <bold>(B&#x02013;D)</bold> are fresh tissue snap frozen fragments representative of regions referred to as muscular <bold>(B,C)</bold> and fatty myocardium, <bold>(D)</bold> respectively, stained with oil red O (original magnification &#x000d7; 50). The red staining indicates neutral lipid accumulation. <bold>(B)</bold> Note the normal discrete perinuclear staining of the cardiomyocytes (white arrow) within the well-preserved myocardial core. <bold>(C)</bold> In contrast, there is an abnormal major accumulation of fatty droplets (not visible under standard staining) dispersed within the cells of the mid mural muscular zone located above the residual muscular core and surrounded by fatty tissue. Note the remaining normal central position of the nucleus within the myocardial cells (red arrow). <bold>(D)</bold> Finally, there is a direct transdifferentiation of myocardial cells into adipocytes within the upper muscular zone bordering the normal subepicardial fat. Take notice of the major confluence of the fatty droplets as well as of the final aspect of total fatty transformation with migration of the nucleus beneath the cell membrane (black arrows).</p></caption><graphic xlink:href="fphys-05-00429-g0002"/></fig><p>Molecular mechanisms underlying ARVC are now better understood. The link between PPAR gamma dysfunction and desmosomal genetic mutations implicates the Wnt/beta-catenin pathway. Thus, the suppression of canonical Wnt/beta-catenin signaling by nuclear PG recapitulates the phenotype of ARVD by exhibiting fat accumulation in cardiomyocytes, enhanced myocyte apoptosis, ventricular dysfunction, and ventricular arrhythmias in transgenic mice (Garcia-Gras et al., <xref rid="B49" ref-type="bibr">2006</xref>). The desmosomal PG also known as gamma-catenin, has structural and functional similarities to beta-catenin, which is the effector for canonical Wnt signaling (Moon et al., <xref rid="B114" ref-type="bibr">2002</xref>). PG interacts and competes with beta-catenin at multiple cellular levels with a net negative effect on the canonical Wnt/beta-catenin signaling pathway through Tcf/Lef transcription factors (Ben-Ze'ev and Geiger, <xref rid="B11" ref-type="bibr">1998</xref>; Klymkowsky et al., <xref rid="B84" ref-type="bibr">1999</xref>; Zhurinsky et al., <xref rid="B190" ref-type="bibr">2000</xref>; Maeda et al., <xref rid="B105" ref-type="bibr">2004</xref>). Mutating the desmosomal protein DSP by impairing desmosome assembly set free gamma-catenin from the desmosomes. As a consequence, gamma catenin translocates to the nucleus and after competition with beta-catenin suppresses signaling through the canonical Wnt/beta-catenin-Tcf/Lef pathway. Suppression of DSP expression responsible for human ARVC, leads to nuclear localization of PG and to suppression of canonical Wnt/beta-catenin-Tcf/Lef1 signaling in cultured atrial myocytes and in mouse hearts (Sen-Chowdhry et al., <xref rid="B148" ref-type="bibr">2005</xref>). Tcf/Lef1 suppression induces a transcriptional switch from myogenesis to adipogenesis (Ross et al., <xref rid="B139" ref-type="bibr">2000</xref>). This leads to enhanced adipogenesis, fibrogenesis, and myocyte apoptosis, thus summarizing the phenotype of human ARVC (Corrado et al., <xref rid="B28" ref-type="bibr">1997</xref>).</p></sec><sec><title>Cardiac hypoxia</title><p>Hypoxia up-regulates expression of <italic>PPAR gamma angiopoietin-related</italic> gene in cardiomyocytes (Belanger et al., <xref rid="B9" ref-type="bibr">2002</xref>). Hypoxia-inducible factor 1 alpha (HIF1 alpha) inhibits PPAR alpha expression during hypoxia (Narravula and Colgan, <xref rid="B119" ref-type="bibr">2001</xref>). Hypoxia leads to activation of HIF1 alpha (Krishnan et al., <xref rid="B88" ref-type="bibr">2009</xref>) and several genes involved in the regulation of glucose transporters, glycolytic enzyme, and pyruvate deshydrogenase kinase (PDK1). HIF1 alpha overexpression <italic>in vitro</italic> leads to triacylglycerol accumulation, and reduced FAO due to inhibition of PPAR alpha. Cardiac hypoxia represents a pathological state where expression of PPAR alpha is reduced whereas that of PPAR gamma is increased. Hypoxia triggers a cascade of cellular metabolic responses including a decrease in mitochondrial oxidative flux (Huss et al., <xref rid="B72" ref-type="bibr">2001</xref>). Under hypoxic conditions, myocytes exhibit significant accumulation of intracellular neutral lipid consistent with reduced carnitine palmitoyltransferase-1 (CPT-1) activity and diminished FAO capacity. Hypoxia reduces PPAR alpha/RXR binding activity and had no effect on the nuclear level of PPAR alpha protein. Hypoxia reduced the nuclear and cellular RXR levels and deactivates PPAR alpha by reducing the availability of its obligate partner RXR. In rat models of systemic hypoxia (Razeghi et al., <xref rid="B134" ref-type="bibr">2001</xref>), cardiac hypoxia induces a decrease in heart muscle transcript levels of <italic>PPAR alpha</italic> and <italic>PPAR alpha-regulated</italic> genes (PDK4), muscle CPT-1, and malonyl-CoA decarboxylase. This explains the increased reliance of the heart for glucose during hypoxia.</p><p>PPAR gamma co-activator 1 alpha (PGC-1 alpha) is a major regulator of mitochondrial biogenesis and activity in the cardiac muscle. Hypoxia stimulates the expression of PGC-1 alpha in cardiac myocytes (Zhu et al., <xref rid="B189" ref-type="bibr">2010</xref>). PGC-1 alpha stimulates the expression of clock genes, particularly <italic>Bmal1</italic> and <italic>Rev</italic>-<italic>erb alpha</italic> (Liu et al., <xref rid="B99" ref-type="bibr">2007a</xref>). Mice lacking PGC-1alpha present abnormal diurnal rhythms of activity, body temperature, and metabolic rate. Overexpression of PGC-1 alpha inhibits clock gene expression in both heart and skeletal muscles and decreases the expression of PPAR alpha. PGC-1 alpha overexpression abolishes the diurnal variation of EF (Wu et al., <xref rid="B173" ref-type="bibr">2011</xref>) and plays an important role on cardiac function by regulating CRs of metabolic genes.</p><p>Ephrins belong to the family of receptor tyrosine kinases. Interestingly, Ephrin-Eph cell signaling is linked to the Wnt/beta catenin pathway (Clevers and Batlle, <xref rid="B26" ref-type="bibr">2006</xref>) and favorably influences cardiomyocyte viability which ultimately preserves cardiac function after myocardial infarction. Ephrin-Eph signaling could potentially be a new therapeutic target in the treatment of myocardial infarction (O'Neal et al., <xref rid="B124" ref-type="bibr">2013</xref>). In non re-perfused hearts of mice with a functional deletion of the CR gene <italic>mPer2</italic>, myocardial infarct size is reduced. A decrease in infarct size in <italic>mPer2-M</italic> mouse hearts following ischemia-reperfusion injury and ischemic preconditioning is observed and improves preservation of myocardial viability. In the <italic>mPer2</italic>-mutant mouse myocardium cardio-protection occurs via the mechanisms connecting cardiac events, mitochondrial function, and <italic>mPer2</italic> (Virag et al., <xref rid="B164" ref-type="bibr">2013</xref>).</p></sec><sec><title>Cardiac hypertrophy and cardiac overload</title><p>Development of cardiac hypertrophy and progression to heart failure induce a change in myocardial metabolism, characterized by a switch from fatty acid utilization to glycolysis, and lipid accumulation. PPAR gamma and HIF-1 alpha are key mediators of lipid anabolism and glycolysis, respectively. They are jointly up-regulated in hypertrophic cardiomyopathy and cooperate to mediate key changes in cardiac metabolism (Krishnan et al., <xref rid="B88" ref-type="bibr">2009</xref>). In response to pathological stress, HIF-1 alpha activates glycolytic genes, and PPAR gamma. This results in increased glycolytic flux, glucose-to-lipid conversion via the glycerol-3-phosphate pathway, and contractile dysfunction. Ventricular deletion of HIF1 alpha in mice prevents hypertrophy-induced PPAR gamma activation, the consequent metabolic re-programming, and contractile dysfunction. HIF-1 alpha and PPAR gamma protein expression is up-regulated in human and mouse cardiac hypertrophy. HIF 1 alpha directly activates PPAR gamma transcription. PPAR gamma is a key downstream effector of HIF-1 alpha-driven triacylglycerol accumulation in cardiomyocytes (Krishnan et al., <xref rid="B88" ref-type="bibr">2009</xref>).</p><p>In pathological hypertrophied heart, PPAR alpha expression and activity are diminished, leading to a reduction in the capacity for FAO and increased rate of glucose utilization (Barger et al., <xref rid="B7" ref-type="bibr">2000</xref>). Alpha 1-adrenergic agonist-induced hypertrophy of cardiomyocytes in culture results in a switch in energy substrate preference from fatty acids to glucose and in a significant decrease in palmitate oxidation rates together with a reduction in the expression of the gene encoding muscle carnitine palmitoyltransferase 1 (<italic>M-CPT1</italic>). Cardiac myocyte transfection has shown that <italic>M-CPT1</italic> promoter activity is repressed during cardiomyocyte hypertrophic growth, an effect involving a <italic>PPAR alpha</italic> response element. Hypertrophied myocytes exhibited reduced capacity for cellular lipid homeostasis, as evidenced by intracellular fat accumulation. Thus, during cardiomyocyte hypertrophic growth, PPAR alpha is deactivated at several levels, leading to diminished capacity for myocardial lipid metabolism. The functional consequences of this metabolic switch from lipid to glucose may serve to preserve ventricular function in the context of chronic pressure overload (Young et al., <xref rid="B181" ref-type="bibr">2001a</xref>). During cardiac pressure overload-induced cardiac hypertrophy, the diurnal variation of metabolic gene expression is completely suppressed and the cardiac performance is impaired (Young et al., <xref rid="B182" ref-type="bibr">2001b</xref>). The induction of clock output genes is attenuated in the pressure-overloaded hypertrophied heart, providing evidence for a diminished ability of the hypertrophied heart to anticipate and subsequently to adapt to physiological alterations during the day (Young et al., <xref rid="B183" ref-type="bibr">2001c</xref>).</p></sec><sec><title>Osteoporosis</title><p>The Wnt pathway induces differentiation of bone-forming cells (osteoblasts) and suppresses the development of bone-resorbing cells (osteoclasts). It is controlled by antagonists that interact either with Wnt proteins (Wnts) or with Wnt co-receptors. Wnts function as key regulators in osteogenic differentiation of mesenchymal stem cells and bone formation. Aberrant Wnt pathways are associated with many osteogenic diseases (Rawadi and Roman-Roman, <xref rid="B133" ref-type="bibr">2005</xref>; Canalis, <xref rid="B18" ref-type="bibr">2013</xref>). Both human genetics and animal studies have pointed out the role of the Wnt/LRP5 pathway as a major regulator of bone mass. In mice, down-regulation or neutralization of Wnt antagonists enhances bone formation. Mutations in <italic>LRP5</italic> cause primary osteoporosis by reducing Wnt signaling activity and result in decreased bone formation (Korvala et al., <xref rid="B86" ref-type="bibr">2012</xref>). Heterozygous PPAR gamma-deficient mice exhibit high bone mass by stimulating osteoblastogenesis from bone marrow progenitors. Inhibition of PPAR gamma increases osteoblastogenesis and bone mass in male C57BL/6 Mice (Duque et al., <xref rid="B36" ref-type="bibr">2013</xref>). PPAR gamma inhibits osteoblast differentiation (Wan et al., <xref rid="B167" ref-type="bibr">2007</xref>).</p><p>Cardiovascular disease and osteoporosis are common age-related conditions associated with significant morbidity and mortality. An increasing body of biological and epidemiological evidences provides support for a link between cardiovascular disease and osteoporosis that cannot be explained by age alone (Farhat and Cauley, <xref rid="B41" ref-type="bibr">2008</xref>). Several hypotheses have been proposed to explain the link between osteoporosis and cardiovascular disease including shared risk factors, common pathophysiological mechanisms and common genetic factors.</p></sec><sec><title>Alzheimer disease (AD)</title><p>AD is a progressive neurodegenerative disorder, neuropathologically characterized by amyloid-beta (Abeta) plaques, and hyperphosphorylated tau accumulation with hereditary missense mutations in the amyloid precursor protein or presenilin-1 and -2 (<italic>PSEN1</italic> and <italic>PSEN2</italic>) genes. Presenilins are involved in modulating beta-catenin stability; therefore familial AD-linked PSEN-mediated effects can reduce the Wnt pathway (Boonen et al., <xref rid="B13" ref-type="bibr">2009</xref>). Tau phosphorylation is mediated by GSK-3 beta, a key antagonist of the Wnt pathway. Sustained loss of function of Wnt/beta-catenin signaling underlies the onset and progression of AD (Inestrosa and Toledo, <xref rid="B73" ref-type="bibr">2008</xref>; De Ferrari et al., <xref rid="B30" ref-type="bibr">2014</xref>). Downregulation of Wnt signaling induced by Abeta is associated with AD progression. Persistent activation of Wnt signaling through Wnt ligands, or inhibition of negative regulators of Wnt signaling, such as Dickkopf-1 and GSK-3 beta are able to protect against Abeta toxicity and ameliorate cognitive performance in AD (Wan et al., <xref rid="B166" ref-type="bibr">2014</xref>). A relationship between amyloid-beta-peptide -induced neurotoxicity and a decrease in the cytoplasmatic levels of beta-catenin has been observed. Although PPAR gamma is elevated in the brain of AD individuals (Jiang et al., <xref rid="B79" ref-type="bibr">2008</xref>), activation of the Wnt signaling pathway may be proposed as a therapeutic target for the treatment of AD.</p><p>In old mice engineered to lack <italic>Bmal1</italic>, there is evidence of brain cell damage that looked similar to that seen in AD. BMAL1 in a complex with CLOCK regulates cerebral redox homeostasis and connects impaired clock gene function to neurodegeneration (Musiek et al., <xref rid="B116" ref-type="bibr">2013</xref>). Altered CR synchronization has been reported in the brain of AD patients (Cermakian et al., <xref rid="B20" ref-type="bibr">2011</xref>). CR disturbances affect as many as a quarter of AD patients. Alterations in the SCN and melatonin secretion are the major factors linked with CR abnormalities. Daytime agitation, night-time insomnia, and restlessness are among the common behavioral alterations observed in AD. Normally, in the interstitial fluid, Abeta has a diurnal fluctuation with low levels during sleep and peak levels during wake. Prolonged wake and/or orexin administration increase levels of the Abeta in the interstitial fluid of the brain in mice. Orexin antagonist reduces amyloid deposits in brain areas. There is a strong causal association between AD and cardiovascular disease. Several cardiovascular risk factors including hypertension and diabetes are also risk factors for dementia (Stampfer, <xref rid="B152" ref-type="bibr">2006</xref>).</p></sec><sec><title>Bipolar disorder and schizophrenia</title><p>The Wnt pathway and its key enzyme, GSK 3 beta, which antagonizes the canonical Wnt pathway, play an important role in regulating synaptic plasticity, cell survival, and CRs in the mature central nervous system. This pathway is implicated in the pathophysiology and treatment of bipolar disorder (Gould and Manji, <xref rid="B58" ref-type="bibr">2002</xref>; Valvezan and Klein, <xref rid="B162" ref-type="bibr">2012</xref>). GSK3- beta-inhibitor lithium chloride enhances activation of Wnt canonical signaling (Hedgepeth et al., <xref rid="B66" ref-type="bibr">1997</xref>; Sinha et al., <xref rid="B151" ref-type="bibr">2005</xref>; Galli et al., <xref rid="B48" ref-type="bibr">2013</xref>). Lithium activates downstream components of the Wnt signaling pathway <italic>in vivo</italic>, leading to an increase of the beta-catenin protein. GSK3-beta phosphorylates and stabilizes the orphan nuclear receptor Rev-erb alpha, a negative component of the circadian clock. Lithium treatment of cells leads to rapid proteasomal degradation of Rev-erb alpha and activation of clock gene <italic>Bmal1</italic> (Yin et al., <xref rid="B179" ref-type="bibr">2006</xref>). The origin of cyclicity in bipolar disorders has been shown by means of a computational approach, and this disease enters the class of dissipative structures (Goldbeter, <xref rid="B56" ref-type="bibr">2013</xref>). Valproate, an effective medication for the prevention and treatment of mood symptoms in bipolar disorder causes a decrease of PPAR gamma signaling (Lan et al., <xref rid="B90" ref-type="bibr">2008</xref>). Many cardiovascular complications are seen in bipolar disorder (Swartz and Fagiolini, <xref rid="B156" ref-type="bibr">2012</xref>).</p><p>An emerging role for Wnt and GSK-3 beta signaling pathways has been found in schizophrenia (Singh, <xref rid="B150" ref-type="bibr">2013</xref>). Sleep and circadian rhythm disruption are seen in schizophrenia (Wulff et al., <xref rid="B174" ref-type="bibr">2012</xref>). Schizophrenia increases risks of cardiovascular disease, particularly coronary heart disease, dyslipidemia, diabetes and hypertension (Hennekens et al., <xref rid="B67" ref-type="bibr">2005</xref>; Andreassen et al., <xref rid="B3" ref-type="bibr">2013</xref>).</p></sec></sec><sec><title>Diseases associated with activation of the Wnt/beta-catenin pathway and decreased expression of PPAR gamma</title><p>Numerous diseases present a common denominator: the Wnt/beta-catenin pathway is overexpressed and the PPAR gamma expression is decreased. This explains why type 2 diabetes is commonly associated with hypertension, sympathetic- parasympathetic abnormalities, and cancers and why CR disruptions are often observed among these pathologies. PPAR alpha expression is often increased in these diseases.</p><sec><title>Impaired sympathetic-parasympathetic system</title><p>PPAR gamma and sympathetic nerve activity (SNA) antagonistically regulate energy metabolism and cardiovascular function with the former promoting anabolism and vasorelaxation and the later favoring catabolism and vasoconstriction (Yang et al., <xref rid="B176" ref-type="bibr">2013</xref>). Systemic inactivation of PPAR gamma can be generated constitutively by using Mox2-Cre mice (<italic>MoxCre/flox</italic>) or inducibly by using the tamoxifen system (<italic>EsrCre/flox/TM</italic>). There is an increase in heart rate in both strains of null mice. PPAR gamma deletion causes the activation of SNA. Rosuvastatin increases vascular endothelial PPAR gamma expression and corrects blood pressure variability in obese dyslipemic mice (Desjardins et al., <xref rid="B31" ref-type="bibr">2008</xref>). Sympathetic adrenal function is disrupted in both <italic>Bmal1(-/-)</italic> and <italic>Clock</italic> (mut) mice (Curtis et al., <xref rid="B29" ref-type="bibr">2007</xref>). Although a shorter ultradian rhythm remains, <italic>Bmal1</italic> deletion abolishes the 24-h frequency in cardiovascular rhythms. In humans, heart rate variability has been shown to be driven by an intrinsic mechanism (Hu et al., <xref rid="B71" ref-type="bibr">2004</xref>; Ivanov et al., <xref rid="B76" ref-type="bibr">2007</xref>). CRs and sleep modulate the sympathetic-parasympathetic balance. Sleep deprivation induces a decrease in the global variability, and an imbalance of the autonomous nervous system (ANS) with an increase in sympathetic activity and a loss of parasympathetic predominance. Human individuals homozygous for the longer allele PER3(5/5) compared with PER3(4/4) subjects present an elevated sympathetic predominance and a reduction of parasympathetic activity (Viola et al., <xref rid="B163" ref-type="bibr">2008</xref>). In mice, selective deletion of the <italic>Bmal1</italic> activator PPAR gamma in the vasculature induces a diminution in heart rate circadian variations (Wang et al., <xref rid="B169" ref-type="bibr">2008</xref>). The CCM mouse model exhibits a decrease in heart rate. Conversely, this model does not present differences in systolic, diastolic, and mean blood pressures as compared with controls (Bray et al., <xref rid="B15" ref-type="bibr">2008</xref>).</p></sec><sec><title>Type 2 diabetes</title><p>PPAR alpha activity and its downstream targets are abnormally activated in the diabetic heart, leading to a marked increase in both fatty acid uptake and oxidation (Finck et al., <xref rid="B42" ref-type="bibr">2005</xref>). Chronic activation of the cardiac PPAR alpha pathway which occurs in the diabetic heart, contributes to myocardial lipid accumulation and diabetic cardiomyopathy (Finck et al., <xref rid="B44" ref-type="bibr">2002</xref>). Diabetes alters the circadian clock in the heart. The clock in the heart loses normal synchronization with its environment during diabetes. Diabetes and fasting activate the expression of cardiac FAO. Excessive fatty acid import and oxidation may be a cause of pathological cardiac remodeling in the diabetic heart (Finck and Kelly, <xref rid="B43" ref-type="bibr">2002</xref>). In type 2 diabetes, PPAR alpha is overexpressed and expression of PPAR gamma is deceased. Some TZD PPAR agonists are used to treat type 2 diabetes. The Wnt/beta-catenin signaling pathway is involved in diabetes mellitus (Ip et al., <xref rid="B75" ref-type="bibr">2012</xref>). Expression of PGC-1 alpha is down-regulated in muscles of type 2 diabetic subjects (Liang and Ward, <xref rid="B98" ref-type="bibr">2006</xref>). PGC-1 alpha activates the expression of insulin-sensitive GLUT4 in skeletal muscle and plays a role in preventing insulin resistance and type 2 diabetes mellitus.</p><p>The mammalian clock (<italic>Bmal1, Clock, Cry1, Cry2, Per1, Per2, and Per3</italic>) expresses CRs and the phases of these CRs are altered in the hearts from streptozotocin-induced diabetic rats (Young et al., <xref rid="B184" ref-type="bibr">2002</xref>). Two BMAL1 haplotypes are associated with type 2 diabetes and hypertension. This provides evidence of a causative role of <italic>Bmal1</italic> variants in pathological components of the metabolic syndrome (Woon et al., <xref rid="B172" ref-type="bibr">2007</xref>). Rhythmic control of insulin release is deregulated in humans with diabetes. Disruption of the clock components <italic>Clock</italic> and <italic>Bmal1</italic> leads to hypoinsulinemia and type 2 diabetes. Pancreatic islets express self-sustained circadian gene and protein oscillations of the transcription factors CLOCK and BMAL1. The phase of oscillation of the islet genes is delayed in circadian mutant mice, and both <italic>Clock</italic> and <italic>Bmal1</italic> mutants show impaired glucose tolerance, reduced insulin secretion and defects in size and proliferation of pancreatic islets (Marcheva et al., <xref rid="B106" ref-type="bibr">2010</xref>). In rodent models of type II diabetes, mean blood pressure is mildly elevated. The elevation in blood pressure is accompanied by changes in the circadian variation of blood pressure as demonstrated in type 2 diabetes (db/db) mice. The daytime fall in blood pressure in mice is significantly blunted in type 2 diabetes db/db mice (Rudic and Fulton, <xref rid="B140" ref-type="bibr">2009</xref>).</p></sec><sec><title>Hypertension</title><p>PPAR gamma in vascular muscle plays a role in the regulation of vascular tone and blood pressure. Thus, mutations in PPAR gamma induce severe hypertension and type 2 diabetes. Transgenic mice with mutations in PPAR gamma in smooth muscle present vascular dysfunction and severe systolic hypertension (Halabi et al., <xref rid="B63" ref-type="bibr">2008</xref>). PPAR gamma ligands lower blood pressure in both animals and humans. PPAR gamma agonist rosiglitazone improves vascular function and lowers blood pressure in hypertensive transgenic mice (Ryan et al., <xref rid="B141" ref-type="bibr">2004</xref>). In mice, after vascular PPAR gamma deletion, circadian variations of blood pressure and heart rate are dampened through a dysregulation of <italic>Bmal1</italic> (Wang et al., <xref rid="B169" ref-type="bibr">2008</xref>). In a null mouse model with specific disruption of PPAR gamma in endothelial cells, PPAR gamma appears to be an important regulator of blood pressure and heart rate mimicking type 2 diabetes, and mediates the antihypertensive effects of rosiglitazone (Nicol et al., <xref rid="B120" ref-type="bibr">2005</xref>). PPAR gamma regulates the renin-angiotensin system activity in the hypothalamic paraventricular nucleus and ameliorates peripheral manifestations of heart failure (Yu et al., <xref rid="B186" ref-type="bibr">2012</xref>). Activation of PPAR gamma down-regulates the renin-angiotensin system. PPAR gamma is expressed in key brain areas involved in cardiovascular and autonomic regulation. Activation of central PPAR gamma reduces sympathetic excitation and improves peripheral manifestations of heart failure by inhibiting brain renin-angiotensin system activity. PPAR gamma ligands lower blood pressure in both animals and humans, possibly via the PPAR gamma-mediated inhibition of the angiotensin II type 1 receptor expression which results in the suppression of the renin-angiotensin system (Sugawara et al., <xref rid="B154" ref-type="bibr">2010</xref>). Genetic variation in BMAL1 is associated with the development of hypertension in man. BMAL1 dysfunction is associated with susceptibility to hypertension and type 2 diabetes. In conditions of constant darkness, <italic>Cry1/Cry2</italic> deficient mice are hypertensive in the daytime (Rudic and Fulton, <xref rid="B140" ref-type="bibr">2009</xref>). Targeted deletion of <italic>Bmal1</italic> in mice <italic>(Bmal1-KO)</italic> abolishes the CR in blood pressure. Mice with targeted deletion of <italic>PPAR gamma</italic> in the endothelium (<italic>EC-PPAR gamma-KO</italic>) exhibit a striking phenotypic resemblance to endothelial cell (EC)-specific deletion of <italic>Bmal1 (EC-Bmal1-KO)</italic>. The loss of PPAR gamma in the aorta of both <italic>EC-PPAR gamma-KO</italic> mice leads to reduced expression of <italic>Bmal1, Cry1, Cry2</italic>, and <italic>Per2</italic>. The ability of PPAR gamma to modulate blood pressure arises in part from its ability to transactivate <italic>Bmal1</italic>.</p></sec><sec><title>Atherosclerosis</title><p>Wnt/beta-catenin signaling plays a key role in atherosclerosis (Wang et al., <xref rid="B170" ref-type="bibr">2002</xref>). Besides Wnt/beta-catenin, GSK3-beta acts as a beta-catenin independent signal, and plays a crucial role in the regulation of cell proliferation and vascular homeostasis. The progression of atherosclerosis is prevented by PPAR gamma ligands in both animals and humans (Sugawara et al., <xref rid="B154" ref-type="bibr">2010</xref>). Monocyte adhesion to vascular endothelium is one of the early processes in the development of atherosclerosis (Lee et al., <xref rid="B92" ref-type="bibr">2006</xref>). Activation of the canonical Wnt/beta-catenin pathway enhances monocyte adhesion to endothelial cells.</p></sec><sec><title>Cardiac-restricted overexpression of PPAR alpha (MHC-PPAR)</title><p>In mice with cardiac-restricted overexpression of PPAR alpha <italic>(MHC-PPAR</italic>), the expression of PPAR alpha target genes is increased whereas that of genes involved in glucose transport and utilization is repressed (Finck et al., <xref rid="B44" ref-type="bibr">2002</xref>). The metabolic phenotype <italic>of MHC-PPAR</italic> mice mimics that of the diabetic heart. <italic>MHC-PPAR</italic> hearts exhibits profiles of diabetic cardiomyopathy including ventricular hypertrophy, activation of gene markers of pathological hypertrophic growth, and systolic ventricular dysfunction. Transgenic mice overexpressing PPAR alpha in muscle (<italic>MCK-PPAR alpha</italic> mice) developed glucose intolerance. Skeletal muscle of <italic>MCK-PPAR alpha</italic> mice exhibits increased FAO rates and reduced insulin-stimulated glucose uptake. The effects on muscle glucose uptake imply transcriptional repression of the <italic>GLUT4</italic> gene.</p></sec><sec><title>Aging</title><p>Aging is associated with various heart diseases, and this may be attributable, in part, to the prolonged exposure of the heart to cardiovascular risk factors. However, aging is also associated with heart disorders such as diastolic dysfunction that are not necessarily linked to the risk factors for cardiovascular diseases. A mechanistic link between Wnt signaling and premature aging or aging-related phenotypes has been demonstrated (Naito et al., <xref rid="B117" ref-type="bibr">2010</xref>). Tissues and organs from klotho-deficient animals showevidence of increased Wnt signaling. Both <italic>in vitro</italic> and <italic>in vivo</italic>, continuous Wnt exposure triggers accelerated cellular senescence. Thus, klotho appears to be a Wnt antagonist (Brack et al., <xref rid="B14" ref-type="bibr">2007</xref>; Liu et al., <xref rid="B100" ref-type="bibr">2007b</xref>). Specific mutations in the human gene encoding lamin A cause premature aging. In mice and humans, these mutations affect adult stem cells by interfering with the Wnt signaling pathway (Meshorer and Gruenbaum, <xref rid="B111" ref-type="bibr">2008</xref>). Overexpression of <italic>Per</italic> in the fruit fly <italic>Drosophila melanogaster</italic> enhances long-term memory, while in <italic>Per</italic> null flies memory is impaired. This supports a link for circadian genes in the processes of learning and memory (Sakai et al., <xref rid="B142" ref-type="bibr">2004</xref>). In aged animals, the normal photonic stimulation of <italic>Per1</italic> expression is reduced. The free-running period of <italic>Per1</italic>&#x02013;<italic>luc</italic> rhythmicity is shortened in aged animals and the amplitude of <italic>Clock</italic> and <italic>Bmal1</italic> expression is decreased (Kolker et al., <xref rid="B85" ref-type="bibr">2003</xref>).</p></sec><sec><title>Neurodegenerative diseases</title><p>The common denominator overexpression of the Wnt/beta-catenin pathway and the consequent decrease in PPAR gamma expression play a central role in numerous neurodegenerative diseases (Clevers, <xref rid="B24" ref-type="bibr">2006a</xref>; MacDonald et al., <xref rid="B104" ref-type="bibr">2009</xref>; Yang, <xref rid="B178" ref-type="bibr">2012</xref>). PPAR gamma agonists could potentially inhibit neuro-inflammation and subsequently neurodegeneration. This may partially occur through the ability of PPAR: RXR heterodimers to antagonize <italic>NF_&#x003ba;B</italic> mediated gene transcription of several inflammatory mediators such as COX-2, iNOS, and various proinflammatory cytokines. It is not surprising that abnormalities of the cardiovascular system and CRs dysfunction are often associated with neurodegenerative pathologies. Sleep disturbances may predict manifestation of neurodegenerative diseases (Postuma and Montplaisir, <xref rid="B130" ref-type="bibr">2009</xref>).</p></sec><sec><title>Huntington disease (HD)</title><p>HD is a dominantly inherited cytosine-adenine-guanine (CAG) repeat disorder with expanded polyglutamine (polyQ) tracts in huntingtin, causing striatal and cortical degeneration (Walker, <xref rid="B165" ref-type="bibr">2007</xref>). Huntingtin interacts with beta-catenin, beta -TrCP, and axin. Normal huntingtin acts as a scaffold protein, promoting the beta-catenin degradation by facilitating the recognition of beta-catenin by beta -TrCP within the destruction complex (Godin et al., <xref rid="B54" ref-type="bibr">2010</xref>). The binding of beta-catenin to the destruction complex is altered in HD. The presence of an abnormal polyQ expansion in mutant huntingtin leads to a decreased binding to beta-catenin therefore impairing the binding of beta-catenin to the destruction complex and subsequently resulting in beta-catenin accumulation into the cytosol. Thus, beta-catenin levels are up-regulated in HD. Mutant huntingtin alters the stability and levels of beta-catenin. Reducing the canonical Wnt signaling pathway confers protection against mutant huntingtin toxicity in Drosophila (Dupont et al., <xref rid="B35" ref-type="bibr">2012</xref>). Knockdown of Wnt ligands improves the survival of HD flies. Overexpression of armadillo/beta-catenin destruction complex component (AXIN, APC2, or GSK3-beta) increases the lifespan of HD flies.</p><p>Early-onset of cardiovascular disease is the second leading cause of death in HD patients. Due to the ubiquitous expression of huntingtin, all cell types with high energetic levels can be impaired. Expression of mutant huntingtin induces cardiac dysfunction in the transgenic model of HD (line R6/2). R6/2 mice develop cardiac dysfunction with cardiac remodeling (e.g. hypertrophy, fibrosis, apoptosis, beta1 adrenergic receptor down-regulation) (Mihm et al., <xref rid="B112" ref-type="bibr">2007</xref>). R6/1 transgenic mice exhibit profound autonomic nervous system-cardiac dysfunction involving both sympathetic and parasympathetic systems, leading to cardiac arrhythmias, and sudden death (Kiriazis et al., <xref rid="B83" ref-type="bibr">2012</xref>). A baroreceptor reflex dysfunction has been described in the BACHD mouse model of HD (Schroeder et al., <xref rid="B147" ref-type="bibr">2011</xref>). Several studies report dysfunction of the autonomic nervous system in HD patients. This may contribute to the increased incidence of cardiovascular events in this patient population that often leads to death. There is a blunted response of the baroreceptor reflex as well as a significantly higher daytime blood pressure in BACHD mice compared to WT controls, which are both indications of autonomic dysfunction. In humans, autonomic dysfunction is present even in the middle stages of HD and affects both the sympathetic and parasympathetic systems (Andrich et al., <xref rid="B4" ref-type="bibr">2002</xref>). Sleep and wake regions of the brain including the brainstem, thalamus, hypothalamus, and cortex are also affected in HD (Kremer et al., <xref rid="B87" ref-type="bibr">1991</xref>). The SCN pacemaker is functional in HD mouse models, so a dysfunction of the circadian circuitry has been proposed to contribute to circadian abnormalities (Pallier and Morton, <xref rid="B125" ref-type="bibr">2009</xref>). Central and peripheral clock gene expression is altered (Maywood et al., <xref rid="B109" ref-type="bibr">2010</xref>). The sleep/wake cycle is disrupted in HD patients characterized by sleep fragmentation at night and delayed sleep phase (Aziz et al., <xref rid="B6" ref-type="bibr">2010</xref>).</p></sec><sec><title>Amyotrophic lateral sclerosis (ALS)</title><p>ALS is a neurodegenerative disease resulting in the progressive loss of upper and lower limb motoneurons and leading to gradual muscle weakening ultimately causing paralysis and death. The Wnt/beta-catenin pathway plays a role in the neurodegeneration of motor neurons in an <italic>in vitro</italic> model of ALS (Pinto et al., <xref rid="B129" ref-type="bibr">2013</xref>). In ALS, a potentially therapeutic pathway may be the activation by PPAR gamma agonists due to their ability to block the neuropathological damage caused by inflammation (Kiaei, <xref rid="B82" ref-type="bibr">2008</xref>). The neuroprotective effect of pioglitazone has been demonstrated in G93A SOD1 transgenic mouse model of ALS and shows a significant increase in their survival. In ALS, PPAR gamma controls natural protective mechanisms against lipid peroxidation (Benedusi et al., <xref rid="B10" ref-type="bibr">2012</xref>).</p><p><italic>Ataxin-2</italic> gene (<italic>ATX2</italic>) is linked to a number of neurodegenerative disorders in humans including ALS and Parkinson disease (PD). ATX2 protein inhibits the production of certain proteins and plays a crucial role in the control of the circadian sleep/wake cycle. <italic>ATX2</italic> regulates the expression of the circadian protein Per in Drosophila. By reducing expression of <italic>ATX2</italic> in Drosophila, the flies are active two and half hours longer. Patients suffering from a form of the neurodegenerative disease spinocerebellar ataxia caused by <italic>ATX2</italic> mutations also experience rapid eye movement sleep disruptions. <italic>ATX2</italic> is necessary for PER accumulation in circadian pacemaker neurons and thus determines period length of circadian behavior. ATX2 is required for the function of TWENTY-FOUR, an activator of PER translation. In humans with ALS, CR of cortisol is impaired (Patacchioli et al., <xref rid="B127" ref-type="bibr">2003</xref>). Both sympathetic and parasympathetic dysfunctions are observed in ALS (Druschky et al., <xref rid="B34" ref-type="bibr">1999</xref>). There are sleep-wake disturbances in patients with ALS (Lo Coco et al., <xref rid="B102" ref-type="bibr">2011</xref>). In human ALS, heart failure is a frequent common cause of death (Gdynia et al., <xref rid="B50" ref-type="bibr">2006</xref>).</p></sec><sec><title>Parkinson disease (PD)</title><p>In a mouse model of PD, a cross talk between inflammatory and Wnt/beta-catenin signaling pathways is involved (L'Episcopo et al., <xref rid="B97" ref-type="bibr">2012</xref>). The Wnt1 regulated Frizzled-1/beta-catenin signaling pathway controls the mesencephalic dopaminergic neuron-astrocyte crosstalk (L'Episcopo et al., <xref rid="B96" ref-type="bibr">2011</xref>). The PPAR gamma agonist pioglitazone modulates inflammation and induces neuroprotection in PD monkeys (Swanson et al., <xref rid="B155" ref-type="bibr">2011</xref>) and mice (Schintu et al., <xref rid="B146" ref-type="bibr">2009</xref>). Expanded glutamine repeats of the ATX2 protein have been identified in fronto-temporal lobar degeneration in PD (Ross et al., <xref rid="B138" ref-type="bibr">2011</xref>). Moreover, a peripheral molecular clock, as reflected in the dampened expression of the clock gene <italic>Bmal1</italic> in leukocytes is altered in PD patients (Cai et al., <xref rid="B17" ref-type="bibr">2010</xref>). There is a disappearance of CRs in a PD dog model (Hineno et al., <xref rid="B69" ref-type="bibr">1992</xref>). Sleep disturbances in PD may be related to CR dysfunction (Hack et al., <xref rid="B62" ref-type="bibr">2014</xref>). Sleep complaints are present in almost half of PD patients. PD patients exhibit increased sleep latency and reduced sleep efficiency. In PD, there is a sustained elevation of serum cortisol levels, reduced circulating melatonin levels, and altered <italic>Bmal1</italic> expression (Breen et al., <xref rid="B16" ref-type="bibr">2014</xref>). PD causes dysfunction of the diurnal autonomic cardiovascular regulation. This dysfunction is profound in patients with severe PD (Haapaniemi et al., <xref rid="B61" ref-type="bibr">2001</xref>).</p></sec><sec><title>Multiple sclerosis (MS)</title><p>Wnt signaling is involved in the MS pathogenesis (Yuan et al., <xref rid="B187" ref-type="bibr">2012</xref>). Mice with experimental autoimmune encephalomyelitis (EAE) have been widely used as a MS model with central nervous system demyelination, neuro-inflammation, and motor impairments. Wnt3a, a Wnt ligand for the canonical pathway, is significantly increased in the spinal cord dorsal horn (SCDH) of the EAE mice. Beta-catenin is also significantly up-regulated. Wnt signaling pathways are up-regulated in the SCDH of the EAE mice and aberrant activation of Wnt signaling contributes to the development of EAE-related chronic pain. PPAR gamma agonists modulate the development of experimental EAE (Drew et al., <xref rid="B33" ref-type="bibr">2008</xref>). Moreover, the risk of myocardial infarction, stroke, heart failure, and atrial fibrillation or flutter is increased in MS patients (Jadidi et al., <xref rid="B77" ref-type="bibr">2013</xref>).</p></sec><sec><title>Friedreich ataxia (FRDA)</title><p>FRDA is a debilitating, life-shortening, degenerative neuromuscular disorder, due to frataxin (FXN) deficiency. FRDA is characterized by neuronal degeneration and heart failure, which are due to loss of transcription of the <italic>FXN</italic> gene caused by a trinucleotide repeat expansion. FXN is a mitochondrial protein involved in iron&#x02013;sulfur-cluster biogenesis, serving to bind and transfer iron to key electron transport complexes and cytochrome C. Diabetes mellitus and serious heart dysfunction (hypertrophic cardiomyopathy) are associated in most cases. The PPAR gamma agonist Azelaoyl PAF increases FXN protein and mRNA expression in human neuroblastoma cells SKNBE and in primary fibroblasts from skin biopsies from FRDA patients. This offers new implications for the FRDA therapy (Marmolino et al., <xref rid="B108" ref-type="bibr">2009</xref>). It has been shown a coordinate dysregulation of the PPAR gamma co-activator PGC-1 alpha and transcription factor Srebp1 in cellular and animal models of FXN deficiency, and in cells from FRDA patients. A genetic modulation of the PPAR gamma pathway affects FXN levels <italic>in vitro</italic>, supporting PPAR gamma as a new therapeutic target in FRDA (Coppola et al., <xref rid="B27" ref-type="bibr">2009</xref>).</p></sec><sec><title>Colon cancer</title><p>Activation of beta-catenin-Tcf signaling has been observed in colon cancer (Morin et al., <xref rid="B115" ref-type="bibr">1997</xref>). Activation of the Wnt signaling pathway via mutation of the <italic>APC</italic> gene is a critical event in the development of colon cancer (Najdi et al., <xref rid="B118" ref-type="bibr">2011</xref>). Inherited mutations in <italic>APC</italic> lead to the development of non-invasive colonic adenomas (polyps). Wnt pathway activation is a driving force in the development of adenomas. Activation of the Wnt/beta-catenin signaling pathway decreases PPAR gamma activity in colon cancer cells (Jansson et al., <xref rid="B78" ref-type="bibr">2005</xref>) and a loss-of-function mutations in PPAR gamma is associated with human colon cancer (Sarraf et al., <xref rid="B143" ref-type="bibr">1999</xref>).</p><p>Colorectal cancer is linked to CR dysregulation (Savvidis and Koutsilieris, <xref rid="B144" ref-type="bibr">2012</xref>). Down-regulation of <italic>Per2</italic> increases beta-catenin protein levels and its target cyclin D, leading to cell proliferation in colon cancer cell lines and colonic polyp formation. <italic>Per2</italic> gene activation suppresses tumorigenesis in colon by down-regulation of beta-catenin. Increased beta-catenin affects the circadian clock and enhances PER2 protein degradation in colon cancer. Suppression of human beta-catenin expression inhibits cellular proliferation in intestinal adenomas. Disruption of the peripheral intestinal CRs may contribute to intestinal epithelial neoplastic transformation of human colorectal cancer. The circadian expression of dihydropyrimidine dehydrogenase, an enzyme that is implicated in the metabolism of the anticancer drug 5-fluorouracil, may be regulated by <italic>Per1</italic> in high-grade colon tumors. The ephrin-Eph cell pathway is linked to the Wnt/beta catenin pathway and is involved in colon cancer (Clevers, <xref rid="B24" ref-type="bibr">2006a</xref>,<xref rid="B25" ref-type="bibr">b</xref>).</p><p>Functional bowel disorders are associated with autonomic disturbance (Tougas, <xref rid="B160" ref-type="bibr">2000</xref>). People with type 2 diabetes have an increased risk of developing colorectal cancer. Diabetes is associated with a higher risk of colon cancer (Yuhara et al., <xref rid="B188" ref-type="bibr">2011</xref>). Heart disease increases at twice the risk of bowel cancer. Colon cancer and coronary artery disease are known to share similar risk factors (smoking, high-fat diet, obesity, diabetes, high blood pressure, and sedentary lifestyle) which increase the risk of colon cancer. This suggests that the two diseases may be connected (Chan et al., <xref rid="B21" ref-type="bibr">2007</xref>). People with coronary artery disease are more likely to develop colon cancer than those without.</p></sec></sec></sec><sec><title>Synthesis</title><p>Circadian rhythms (CRs) are particularly fascinating phenomena. They go very far back in evolution. The existence of a CR is the signature of instability. Beyond a point of bifurcation, an unstable thermodynamic system can evolve spontaneously into a periodic state. These periodic oscillations correspond to a phenomenon of self-organization in time and have been called &#x0201c;dissipative structures&#x0201d; (Prigogine et al., <xref rid="B132" ref-type="bibr">1974</xref>). Dissipative structures are far-from-equilibrium systems, such as cyclones, hurricanes, lasers, B&#x000e9;nard cells, Belousov&#x02013;Zhabotinsky reactions, Turing structures, circadian rhythms, and more generally most of the living organisms. CRs are based on the existence of negative feedback loops. Oscillatory behavior gradually has been integrated into the living world to become one of its major characteristics. In the cardiovascular system, circadian genes show properties of anticipation and this makes it possible to coordinate lipid and carbohydrate metabolism with the cardiovascular function, especially for blood pressure and heart rate. Dysfunction of CRs can be associated with serious clinical problems and may induce a negative impact on quality of life, sometimes with a poor prognosis. Abnormalities of circadian gene function may result in the occurrence of metabolic syndrome, obesity, and even more seriously, stroke, or myocardial infarction.</p><p>Two major systems interfere with circadian genes, namely the canonical Wnt pathway, and the PPAR system. In some cases of diseases presenting profiles 1 or 2, there is not always evidence of the exact influence of CRs on the Wnt-beta catenin-PPAR gamma pathway and cardiac function. PPAR gamma controls the circadian <italic>Clock-Bmal1</italic> genes in the vascular system. Importantly, there is an opposite between activation of the Wnt pathway and PPAR gamma. This is attested by their respective profiles in numerous diseases, either cardiovascular diseases or pathologies with cardiovascular complications. There is a subtle thermodynamic regulation of CRs that run far-from-equilibrium; moreover, there is the need to maintain the balance between the two systems canonical Wnt and PPAR gamma. Indeed, activation of the Wnt system with inactivation of PPAR gamma favors diabetes, hypertension, several cancers, and neurodegenerative diseases. The reverse is observed in ARVD, osteoporosis, Alzheimer disease, bipolar disorder, schizophrenia, and myocardial ischemia. The extreme complexity of the Wnt-PPAR systems and their numerous inter-related pathways partly explain their involvement in numerous diseases. We remain surprised by both the number and the importance of these diseases, causing considerable morbidity, and mortality and heavy social and economic costs.</p><p>The discovery and use of new agonist or antagonist pharmacological agents acting on PPARs, and more generally, directly or indirectly implied in the canonical Wnt system, are particularly important. This leads to numerous novel therapeutic approaches. PPAR gamma is a key regulator of lipid metabolism and its activation by some TZD is used for the treatment of type 2 diabetes and protects against atherosclerosis. However, some TZD have been reported to cause a higher rate of fractures in human patients. Pharmacological inhibition of PPAR gamma represents a potential therapeutic approach for age-related bone loss. Induction of the Wnt pathway or inhibition of Wnt antagonists may offer therapeutic opportunities in treating bone disorders, including osteoporosis. Antibodies targeting the Wnt inhibitor sclerostin lead to increased bone mineral density in post-menopausal women. Lithium, often used to treat bipolar disorder, blocks a Wnt antagonist, decreasing the patient's risk of fractures. Lithium exerts effects on components of the Wnt signaling pathway. The Wnt signaling pathway plays an important role in the treatment of bipolar disorder. The future development of selective GSK3- beta inhibitors may have considerable utility not only for the treatment of bipolar disorder but also for a variety of neurodegenerative disorders. Therapies targeting the Wnt pathway are not without risk, and may lead to over-activation of Wnt/catenin and its association with many tumors. However, it is conceivable that targeting Wnt inhibitors may predispose the individuals to tumorigenic phenotypes.</p><sec><title>Conflict of interest statement</title><p>The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.</p></sec></sec>

Animal Models to Investigate the Pathogenesis of Rheumatic Heart Disease

<p>Rheumatic fever (RF) and rheumatic heart disease (RHD) are sequelae of group A streptococcal (GAS) infection. Although an autoimmune process has long been considered to be responsible for the initiation of RF/RHD, it is only in the last few decades that the mechanisms involved in the pathogenesis of the inflammatory condition have been unraveled partly due to experimentation on animal models. RF/RHD is a uniquely human condition and modeling this disease in animals is challenging. Antibody and T cell responses to recombinant GAS M protein (rM) and the subsequent interactions with cardiac tissue have been predominantly investigated using a rat autoimmune valvulitis model. In Lewis rats immunized with rM, the development of hallmark histological features akin to RF/RHD, both in the myocardial and in valvular tissue have been reported, with the generation of heart tissue cross-reactive antibodies and T cells. Recently, a Lewis rat model of Sydenham&#x02019;s chorea and related neuropsychiatric disorders has also been described. Rodent models are very useful for assessing disease mechanisms due to the availability of reagents to precisely determine sequential events following infection with GAS or post-challenge with specific proteins and or carbohydrate preparations from GAS. However, studies of cardiac function are more problematic in such models. In this review, a historical overview of animal models previously used and those that are currently available will be discussed in terms of their usefulness in modeling different aspects of the disease process. Ultimately, cardiologists, microbiologists, immunologists, and physiologists may have to resort to diverse models to investigate different aspects of RF/RHD.</p>

<contrib contrib-type="author"><name><surname>Rush</surname><given-names>Catherine M.</given-names></name><xref ref-type="aff" rid="aff1">¹</xref></contrib><contrib contrib-type="author"><name><surname>Govan</surname><given-names>Brenda L.</given-names></name><xref ref-type="aff" rid="aff1">¹</xref></contrib><contrib contrib-type="author"><name><surname>Sikder</surname><given-names>Suchandan</given-names></name><xref ref-type="aff" rid="aff1">¹</xref><uri xlink:type="simple" xlink:href="http://frontiersin.org/people/u/181160"/></contrib><contrib contrib-type="author"><name><surname>Williams</surname><given-names>Natasha L.</given-names></name><xref ref-type="aff" rid="aff1">¹</xref><uri xlink:type="simple" xlink:href="http://frontiersin.org/people/u/172181"/></contrib><contrib contrib-type="author"><name><surname>Ketheesan</surname><given-names>Natkunam</given-names></name><xref ref-type="aff" rid="aff1">¹</xref><xref ref-type="corresp" rid="cor1">*</xref><uri xlink:type="simple" xlink:href="http://frontiersin.org/people/u/135289"/></contrib>

Frontiers in Pediatrics

<sec sec-type="intro" id="S1"><title>Introduction</title><p>Animal experimentation has been used for centuries to better understand human physiology and pathological processes, to improve diagnosis and to develop safe intervention, treatment, and prevention strategies. References to animal experimentation are found in the writings of the Greek philosopher&#x02013;physician, Aristotle (384&#x02013;322 BC) who was the first to carry out dissections. Later, Erasistratus (304&#x02013;250 BC) was probably the first to perform experiments on living animals. However, it was Galen (AD 130&#x02013;200) who justified animal experimentation as an arduous path to the truth, believing that assertions not based on experimentation do not lead to scientific progress (<xref rid="B1" ref-type="bibr">1</xref>). The twelfth century Arab physician, Ibn Zuhr (Avenzoar), introduced animal testing as an experimental technique for testing procedures before applying them to human patients (<xref rid="B2" ref-type="bibr">2</xref>). A valid animal model of a specific disease requires the process that initiates and perpetuates the pathological mechanisms to be identical or similar in the animal to that of the human condition that is being modeled.</p><p>Group A streptococci that trigger rheumatic fever (RF) and rheumatic heart disease (RHD) are uniquely human pathogens, with no other known natural host or environmental reservoir. Major shortcomings associated with research into RF/RHD, which may rely on studying human heart tissue from surgery or autopsy tissue from patients include heterogeneity and the ability to obtain sufficient numbers of high quality tissue specimens. Therefore, the availability of an animal model in which the pathogenic mechanisms responsible for RF/RHD could accurately be reproduced is crucial for furthering our understanding of the disease process, designing treatment and for testing efficacy, and safety of vaccine candidates against group A streptococci (GAS). Following the review of published literature on animal models for RF/RHD research in the last 85&#x02009;years, we outline briefly the models that have been used to investigate the pathogenesis of RF/RHD. We outline the recent work on rodents and in particular rats, that have been instrumental in modeling some of the cardinal immunopathological features observed in RF/RHD.</p></sec><sec id="S2"><title>Early Experience with Animal Models</title><p>Developments in molecular biology, genomics, transgenic, and cloning techniques have enabled researchers to study human pathology in animals in greater depth. By investigating and identifying homologous genes across species, researchers can translate experimental data from animals to humans. The overwhelming majority of animals used in biomedical research are rodents such as mice and rats. These are ideal animal models because they are small, easy to handle, reproduce rapidly, have a relatively short life span and are relatively inexpensive to maintain in a laboratory setting. One of the major impediments in the field of RF/RHD research has been the lack of a universally accepted animal model. Furthermore animals are not easily infected by GAS: even when GAS infection is initiated, it is usually not sustained for extended periods of time.</p><p>Early experimental work to produce a suitable animal model of RF/RHD was based on the hypotheses that RF/RHD was caused either by persistent sub-clinical infection by GAS or by direct injury to cardiac tissue by GAS toxins. Therefore, most studies involved the introduction of whole bacteria or crude streptococcal preparations into various animals including mice, rats, guinea pigs (Table <xref ref-type="table" rid="T1">1</xref>), rabbits, or non-human primates. While it was possible to observe myocardial necrosis, myocarditis, and endocarditis following immunization with GAS, none of the pathological lesions were considered representative of the hallmarks of RF/RHD, the development of Aschoff nodules and valvulitis. Animal models of RF/RHD developed prior to the 1970s paralleled the hypotheses for the causation of the disease process that were dominant at the time. When it became increasingly evident that the development of the inflammatory process observed in RF/RHD was immune-mediated, animal models were increasingly used to determine the role of host cross-reactive antibodies that developed post GAS infection. Although functional studies may be preferable in larger animals such as non-human primates, the prohibitive costs, and paucity of specific immunological reagents make such models less appropriate to determine the pathogenesis of RF/RHD.</p><table-wrap id="T1" position="float"><label>Table 1</label><caption><p><bold>Immunopathological changes in rodents investigated as models for rheumatic heart disease</bold>.</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1">Antigen (route of inoculation)</th><th align="left" rowspan="1" colspan="1">Histological changes</th><th align="left" rowspan="1" colspan="1">Antibody response</th><th align="left" rowspan="1" colspan="1">T cell response cytokine production</th><th align="left" rowspan="1" colspan="1">Cross-reactivity</th><th align="left" rowspan="1" colspan="1">References</th></tr></thead><tbody><tr><td align="left" colspan="6" style="background-color:DarkGray;" rowspan="1"><bold>RATS (<italic>Rattus norvegicus</italic>)</bold></td></tr><tr><td align="left" rowspan="1" colspan="1">Whole GAS (FP, SC)</td><td align="left" rowspan="1" colspan="1"><bold>Myocarditis, valvulitis</bold> lymphocyte, monocyte, M&#x000d8;, giant cell, Aschoff-like cell, fibroblast</td><td align="left" rowspan="1" colspan="1">Anti-myocardial IgG</td><td align="left" rowspan="1" colspan="1">NA</td><td align="left" rowspan="1" colspan="1">Valvular protein, myocardial protein</td><td align="left" rowspan="1" colspan="1">Cavelti (<xref rid="B3" ref-type="bibr">3</xref>), Huang et al (<xref rid="B4" ref-type="bibr">4</xref>), Xie et al (<xref rid="B5" ref-type="bibr">5</xref>)</td></tr><tr><td align="left" rowspan="1" colspan="1">Recombinant proteins or peptides of GAS (SC, IP, FP)</td><td align="left" rowspan="1" colspan="1"><bold>Myocarditis, valvulitis</bold> T cell, MNC, neutrophil, Anitschkow-like cell</td><td align="left" rowspan="1" colspan="1">Anti-GAS IgG</td><td align="left" rowspan="1" colspan="1">CD3⁺, CD4⁺, CD8⁺, CD68⁺, TCR-&#x003b1;&#x003b2;⁺</td><td align="left" rowspan="1" colspan="1">Cardiac myosin</td><td align="left" rowspan="1" colspan="1">Quinn et al (<xref rid="B6" ref-type="bibr">6</xref>), Lymbury et al (<xref rid="B7" ref-type="bibr">7</xref>), Gorton et al (<xref rid="B8" ref-type="bibr">8</xref>), Gorton et al (<xref rid="B9" ref-type="bibr">9</xref>), Kirvan et al (<xref rid="B10" ref-type="bibr">10</xref>)</td></tr><tr><td align="left" colspan="6" style="background-color:DarkGray;" rowspan="1"><bold>MICE (<italic>Mus musculus</italic>)</bold></td></tr><tr><td align="left" rowspan="1" colspan="1">Cell wall fragments of GAS (IP)</td><td align="left" rowspan="1" colspan="1"><bold>Myocarditis</bold> MNC, PMNC, giant cell, Anitschkow-like cell</td><td align="left" rowspan="1" colspan="1">Anti-GAS IgG</td><td align="left" rowspan="1" colspan="1">NA</td><td align="left" rowspan="1" colspan="1">NA</td><td align="left" rowspan="1" colspan="1">Ohanian et al (<xref rid="B11" ref-type="bibr">11</xref>)</td></tr><tr><td align="left" rowspan="1" colspan="1">Recombinant protein of GAS (IP)</td><td align="left" rowspan="1" colspan="1">NA</td><td align="left" rowspan="1" colspan="1">Collagen IV reactive IgG</td><td align="left" rowspan="1" colspan="1">NA</td><td align="left" rowspan="1" colspan="1">Basement membrane collagen</td><td align="left" rowspan="1" colspan="1">Dinkla et al (<xref rid="B12" ref-type="bibr">12</xref>)</td></tr><tr><td align="left" colspan="6" style="background-color:DarkGray;" rowspan="1"><bold>GUINEA PIG (<italic>Cavia porcellus</italic>)</bold></td></tr><tr><td align="left" rowspan="1" colspan="1">Whole GAS (IP, IV)</td><td align="left" rowspan="1" colspan="1"><bold>Myocarditis, valvulitis</bold> MNC</td><td align="left" rowspan="1" colspan="1">NA</td><td align="left" rowspan="1" colspan="1">NA</td><td align="left" rowspan="1" colspan="1">NA</td><td align="left" rowspan="1" colspan="1">Gross et al (<xref rid="B13" ref-type="bibr">13</xref>)</td></tr><tr><td align="left" rowspan="1" colspan="1">Cell wall fragments of GAS (FP)</td><td align="left" rowspan="1" colspan="1">NA</td><td align="left" rowspan="1" colspan="1">Anti-GAS IgG</td><td align="left" rowspan="1" colspan="1">NA</td><td align="left" rowspan="1" colspan="1">Cardiac sarcolemmal membrane</td><td align="left" rowspan="1" colspan="1">Yang et al (<xref rid="B14" ref-type="bibr">14</xref>)</td></tr></tbody></table><table-wrap-foot><p><italic>NA, not assessed; IV, intra-venous; SC, sub-cutaneous; IP, intra-peritoneal; FP, foot pad; MNC, mononuclear cell; PMNC, polymorphonuclear cell; M&#x000d8;, macrophage; Ig, immunoglobulin; CD, cluster of differentiation; IFN, interferon; TCR, T cell receptor</italic>.</p></table-wrap-foot></table-wrap></sec><sec id="S3"><title>Current Concepts of Pathogenesis of Rheumatic Heart Disease</title><p>The pathogenesis of RF/RHD involves three principal elements: an infection caused by a specific strain of GAS, a susceptible host, and an aberrant immune response against GAS antigens that cross-react with host tissue (<xref rid="B15" ref-type="bibr">15</xref>, <xref rid="B16" ref-type="bibr">16</xref>). The manifestations of RF/RHD are due to inflammatory changes that occur in cardiac tissue, joints, brain, blood vessels, and skin. Rheumatic carditis is the most serious consequence of the disease process while migratory polyarthritis and the neurologic manifestation Sydenham chorea (SC) may present in conjunction with carditis. Other clinical signs include the development of erythema marginatum and subcutaneous nodules (<xref rid="B16" ref-type="bibr">16</xref>). Host&#x02013;GAS interactions initiating the immune responses cause carditis with subsequent GAS infections potentiating the disease process resulting in a cascade of events that cause hemodynamic changes culminating in irreversible cardiac damage and decompensatory cardiac failure.</p><p>Monoclonal antibodies derived from RF/RHD patients have provided evidence for cross-reactive autoantibodies that target GAS epitopes such as group A carbohydrate, <italic>N</italic>-acetyl-beta-<sc>d</sc>-glucosamine (GlcNAc), and M protein with host proteins including heart valve endothelium, laminin, and laminar basement membrane and cardiac myosin [Reviewed in Ref. (<xref rid="B17" ref-type="bibr">17</xref>, <xref rid="B18" ref-type="bibr">18</xref>)]. GAS M proteins have been studies extensively and it is one of the important GAS proteins involved in inducing the autoimmune process. Over 220 different variants of the M protein have been described (<xref rid="B19" ref-type="bibr">19</xref>). While identification of M proteins is useful for epidemiological and vaccine studies not all GAS M types are implicated in the development of RF/RHD. Peripheral blood and heart-infiltrating T cells from patients with RF/RHD have revealed cross-reactivity of GAS M protein specific T cells with cardiac myosin. Experimental and clinical evidence suggests that antibodies interact with valvular endothelium and activate adhesion molecules with subsequent extravasation of T cells through the activated endothelium into the valve leading to the formation of granulomatous lesions. The inflammatory process that is thus triggered by antibodies to GAS infection activates adhesion molecules such as VCAM and ICAM, which play a role in the migration of leukocytes into the valvular tissue (<xref rid="B17" ref-type="bibr">17</xref>, <xref rid="B18" ref-type="bibr">18</xref>). It has been demonstrated in patients with RF/RHD that specific chemokine genes, such as CXCL3 in the myocardium and CCL1 and CXCL9 in valvular tissue are upregulated (<xref rid="B20" ref-type="bibr">20</xref>). These chemokines could potentially mediate T cell recruitment to areas of inflammation. The specific cytokines produced by peripheral and heart-infiltrating T cells in patients have been identified (<xref rid="B21" ref-type="bibr">21</xref>) suggesting that Th1 and Th17 cells are involved in the development of carditis. Such cytokine profiles have also been observed in other autoimmune inflammatory conditions (<xref rid="B22" ref-type="bibr">22</xref>). Additionally the low levels of IL-4 produced by T cells infiltrating the valvular tissue may potentially contribute to the progression of valvular pathology (<xref rid="B21" ref-type="bibr">21</xref>).</p><p>Repeated GAS infection also leads to neovascularization within heart valves resulting in increased mononuclear cell infiltration resulting in cardiac and valvular damage. This process may also involve autoantibodies generated via epitope spreading following immune recognition of other components of the cardiac tissue such as vimentin and collagen released during tissue destruction (<xref rid="B17" ref-type="bibr">17</xref>).</p><p>One of the major manifestations associated with RF/RHD is SC. It has been found that antibodies from SC patients target the GAS carbohydrate epitope GlcNAc and react with gangliosides and dopamine receptors found on the surface of neuronal cells in the brain. These antibodies have been found to activate calcium calmodulin-dependent protein kinase II in neuronal cells and recognize the intracellular tubulin (<xref rid="B23" ref-type="bibr">23</xref>).</p><p>Molecular mimicry between GAS antigens and host antigens has been shown to initiate and potentiates the development of the clinical manifestations observed in RF/RHD. GAS have a variety of antigens and superantigens that are able to stimulate robust B and T cell responses to autoantigens (<xref rid="B17" ref-type="bibr">17</xref>, <xref rid="B18" ref-type="bibr">18</xref>). Both human and animal studies have provided evidence that have been useful in formulating the potential mechanisms that lead to the post-infectious autoimmune sequelae. Animal models of RF/RHD and SC have been important in characterizing mimicry in carditis and SC as well as attempting to identify the pathogenic epitopes of the autoantigens and GAS antigens involved in the pathogenesis of RF/RHD. However, several facets of the pathological process at a cellular and molecular level that initiate inflammatory changes in the different organs and tissue in RF/RHD remain undefined. Fundamental concepts including which GAS strains and what combination of GAS antigens are involved in the breakdown of tolerance during the disease process are unclear. Studies to address some of these fundamental issues require the use of well characterized animal models that encompass the features observed in RF/RHD.</p></sec><sec id="S4"><title>Use of Rodent Models to Determine the Autoimmune Process</title><p>To further investigate the pathogenesis of RF/RHD and in particular to determine the contribution of autoreactive B and T cells in the RF/RHD pathogenesis, suitable animal models were investigated. Huber and Cunningham (<xref rid="B24" ref-type="bibr">24</xref>) produced myocarditis in MRL+/+ mice immunized with specific N-terminal peptides of GAS M5. These studies were one of the first to demonstrate that mimicry between a pathogen derived epitope and a host protein could break tolerance and trigger an autoimmune process in a susceptible host. The same group at the Oklahoma University Health Sciences Center subsequently developed a more robust animal model for RF/RHD (<xref rid="B6" ref-type="bibr">6</xref>, <xref rid="B25" ref-type="bibr">25</xref>). Female Lewis rats immunized with GAS M protein (Figure <xref ref-type="fig" rid="F1">1</xref>A) exhibit myocardial lesions similar to those observed in patients with RF/RHD. More importantly, valvular pathology was observed for the first time using this model (<xref rid="B5" ref-type="bibr">5</xref>&#x02013;<xref rid="B8" ref-type="bibr">8</xref>, <xref rid="B10" ref-type="bibr">10</xref>, <xref rid="B25" ref-type="bibr">25</xref>). This model is discussed in more detail below.</p><fig id="F1" position="float"><label>Figure 1</label><caption><p><bold>Immunization protocol for the induction of autoimmune valvulitis in Lewis rats and the immunological, histological, and functional changes following immunization with recombinant streptococcal M protein</bold>. <bold>(A)</bold> The induction of valvulitis in the rat autoimmune valvulitis (RAV) model of RF/RHD involves a primary immunization of female Lewis rats (under isoflurane anesthesia) with 500&#x02009;&#x003bc;g GAS rM5 protein (or PBS as a negative control) in complete Freund&#x02019;s adjuvant (CFA) administered subcutaneously (s.c.) in the hock on day 0. On days 1 and 3, rats are injected intraperitoneally (i.p.) with an additional adjuvant being either 0.3&#x02009;&#x003bc;g commercially purchased pertussis toxin [PTx; (<xref rid="B7" ref-type="bibr">7</xref>)] or 10¹⁰ CFU formalin-killed <italic>Bordetella pertussis</italic> (<xref rid="B6" ref-type="bibr">6</xref>, <xref rid="B8" ref-type="bibr">8</xref>, <xref rid="B9" ref-type="bibr">9</xref>) each in 200&#x02009;&#x003bc;l PBS. On day 7, rats receive a booster immunization with 500&#x02009;&#x003bc;g GAS rM5 protein (or PBS as a negative control) in incomplete Freund&#x02019;s adjuvant (IFA) administered s.c. in the flank under anesthesia. On day 21, the rats are euthansed by CO₂ asphyxiation to harvest blood and organs for histological examination of heart tissue and determination of GAS rM5-specific antibody levels and assessment of T cell function. <bold>(B)</bold> GAS rM5-specific IgG antibodies in rat (<italic>n</italic>&#x02009;=&#x02009;5) serum were detected by ELISA. The highest serum dilution that was positive for GAS rM5-specific IgG antibodies (cut-off value 3 SD higher than the mean for the known negative control serum) was recorded as the serum titer. Serum from rats immunized with GAS rM5 contained significantly higher GAS rM5-specific antibodies compared to control (<italic>P</italic>&#x02009;=&#x02009;0.007). <bold>(C)</bold> Proliferative response of GAS rM5-specific T cells from spleens of rats (<italic>n</italic>&#x02009;=&#x02009;5) was determined by ³H-thymidine incorporation assay and found to be significantly higher than in controls immunized with PBS (<italic>P</italic>&#x02009;=&#x02009;0.009). Bars depict the mean&#x02009;&#x000b1;&#x02009;SEM. **<italic>P</italic>&#x02009;&#x02264;&#x02009;0.01. Immunohistological changes <bold>(D,E)</bold> in representative valvular tissue and myocardium (inset) from <bold>(D)</bold> controls and <bold>(E)</bold> rM5-immunized animals. Immunohistochemical staining of infiltrating mononuclear cells demonstrates the presence of CD4+ T cells in rM5-immunized animals compared to controls. Scale bar indicates 50&#x02009;&#x003bc;M (DAB staining). <bold>(F)</bold> ECG complexes from a control and a rM5-immunized rat demonstrate significantly longer P-R interval in the rM5-immunized rats (Images courtesy of Dr Lisa Chilton, James Cook University). <bold>(G)</bold> Echocardiographs from control and an rM5-immunized rat demonstrate reduction in fractional shortening due to reduced LV contractility. Bars represent the width of the left ventricle (LV) chamber during contraction (*Images courtesy of Drs Lisa Chilton and Jane Day, James Cook University).</p></caption><graphic xlink:href="fped-02-00116-g001"/></fig><p>In recent years extensive work on murine models have been conducted to determine the efficacy of anti-GAS vaccine candidates in terms of their ability to protect against GAS infections (<xref rid="B18" ref-type="bibr">18</xref>, <xref rid="B26" ref-type="bibr">26</xref>, <xref rid="B27" ref-type="bibr">27</xref>). These murine models have been used mostly if not exclusively to determine vaccine efficacy. Although an HLA class II transgenic mouse model has been used to investigate both the protective immune responses induced by a vaccine candidate and determine whether it induces histological changes in host tissue (<xref rid="B26" ref-type="bibr">26</xref>), this model may not be very appropriate for the purpose of determining cross-reactivity in host tissue. A year following immunization with the candidate vaccine preparation the investigators did not observe any histological changes in various organs in the immunized mice. Although the vaccine preparation was considered to be safe, control transgenic mice developed additional complications. These complications may deter researchers from using such a model to investigate tissue cross-reactivity and vaccine safety.</p></sec><sec id="S5"><title>Rat Autoimmune Valvulitis Model of Rheumatic Heart Disease</title><p>The Lewis rat model of RF/RHD developed in the late 90s (<xref rid="B6" ref-type="bibr">6</xref>), exhibited myocardial lesions similar to those observed in RF/RHD following immunization with GAS M6 protein (Figure <xref ref-type="fig" rid="F1">1</xref>). More importantly, this Lewis rat model is the first in which valvular changes akin to human pathology has been demonstrated. Valvular lesions observed at the valve surface endothelium spread into the valve. Anitschkow cells and verruca-like lesions have also been observed. Due to the valvular involvement observed in immunized animals, this model has been referred to as the rat autoimmune valvulitis (RAV) model (<xref rid="B8" ref-type="bibr">8</xref>). T cells from recombinant M6-immunized (rM6) rats proliferated <italic>in vitro</italic> in the presence of cardiac myosin. In addition a T cell line produced from GAS rM6-immunized rats proliferated in the presence of cardiac myosin and GAS rM6 protein. When Galvin and colleagues (<xref rid="B25" ref-type="bibr">25</xref>) co-cultures myosin-sensitized lymphocytes isolated from the hearts of Lewis rats with peptides of GAS M5 protein, heart-infiltrating lymphocytes proliferated in response to peptides within the B-repeat region of the GAS M protein. Their work provided evidence that an immune response against cardiac myosin could potentially lead to valvular heart disease and the infiltration of the heart by GAS M protein-reactive T cells.</p><p>Using the same protocol to initiate valvulitis, Lymbury et al. (<xref rid="B7" ref-type="bibr">7</xref>) demonstrated that 80% of Lewis rats immunized with a pool of 15, 20-mer overlapping peptides spanning the conserved C-repeat region of the GAS M5 developed inflammatory lesions in both the myocardium and valvular tissue. These studies highlighted the role for GAS M protein-specific autoreactive T cells in the development of cardiac lesions. T cells from rats immunized with the conserved region peptides proliferated in response to the immunogen and to cardiac myosin. Further proof of the role of both humoral and cellular responses (Figures <xref ref-type="fig" rid="F1">1</xref>B&#x02013;E) in the pathogenesis of RF/RHD was demonstrated by Gorton et al (<xref rid="B8" ref-type="bibr">8</xref>). It was found that GAS rM5 protein elicited opsonic antibodies in Lewis rats, which recognized epitopes within the B- and C-repeat regions of M5. A single peptide from the GAS M5 B-repeat region induced lymphocytes that responded to both recombinant M5 and cardiac myosin. Additionally, it was found that rats immunized with GAS rM5 protein developed valvular lesions (Figures <xref ref-type="fig" rid="F1">1</xref>D,E), distinguished by infiltration of CD3+, CD4+, and CD68+ cells into valve tissue, consistent with human studies. This suggests that RF/RHD is mediated by inflammatory responses involving both CD4+ T cells and macrophages. Recent proof of concept work undertaken by this group on the RAV model has also demonstrated that repetitive immunization with GAS rM5 increases both B and T cell sensitization leading to increased inflammatory cell infiltration that could potentially lead to severe cardiac damage. This observation further demonstrates that the immunopathology in the RAV model reflects the human condition, where repetitive GAS infections lead to exacerbation of RF/RHD, which culminates in cardiac failure.</p><p>The Lewis rat model has also been used to immunize with formalin-killed and sonicated GAS (<xref rid="B5" ref-type="bibr">5</xref>). The investigators were able to demonstrate in rats killed 12&#x02009;weeks following immunization only 50% (4/8) developed myocarditis and valvulitis. In contrast, animals sacrificed 24&#x02009;weeks following GAS immunization demonstrated myocardial and valvular damage and developed rheumatic-like myocarditis with 62.5% (5/8) developing chronic valvulitis. Histological manifestations of the hearts in this group demonstrated &#x0201c;Aschoff-like&#x0201d; cells, verrucous vegetation, and chronic lesions including fibrosis and neovascularization, hallmark of chronic rheumatic valvulitis.</p><p>To identify the epitopes of M5 protein that produce valvulitis, and to prove that M protein-specific T cells may be important mediators of valvulitis, Kirvan and colleagues (<xref rid="B10" ref-type="bibr">10</xref>) used synthetic peptides spanning all three repeat regions of GAS M5 (A, B, and C-repeat regions) contained within the extracellular domain of the streptococcal M5 protein to immunize Lewis rats. Peptides NT4, NT5/6, and NT7 from the A repeat region induced valvulitis similar to the pepsin fragment of M5 protein. T cell lines from rats with valvulitis also recognized peptides NT5/6 and NT6. They also conducted passive transfer of a NT5/6-specific T cell line into na&#x000ef;ve rats, which produced valvulitis with characteristic CD4+ T cell infiltration and upregulation of VCAM demonstrating experimentally that M protein-specific T cells are important mediators of valvulitis.</p><p>To our knowledge the RAV model has not been widely used to investigate the safety of anti-GAS vaccine candidates by assessing their potential to initiate autoimmune pathology. However, prior to a recent human Phase 1 clinical trial for a GAS vaccine based on the J8 construct, the RAV model was used to test for safety of the vaccine preparation. None of the animals immunized with the vaccine preparation in the authors&#x02019; laboratory developed either carditis or valulitis. While the RAV model has been useful in characterizing key aspects involved in the autoimmune process in RF/RHD pathogenesis, its potential has not yet been fully realized. Extensive characterization of cardiac function in the RAV model (Figures <xref ref-type="fig" rid="F1">1</xref>F,G) may lead to greater acceptance of the RAV model among researchers working on different aspects of RF/RHD.</p></sec><sec id="S6"><title>Rat Model for Sydenham&#x02019;s Chorea</title><p>One of the manifestations of RF/RHD is the development of the SC, which is a neurological disorder characterized by rapid, uncoordinated jerking movements affecting the face, hands, and feet. Although it is reported to occur in a quarter of patients with RF/RHD in some regions, it may also be the presenting symptom of RF/RHD (<xref rid="B16" ref-type="bibr">16</xref>). A team of researchers from Tel Aviv University and the Oklahoma University Health Sciences Center developed a model for SC using male Lewis rats, which were exposed to GAS antigen (<xref rid="B28" ref-type="bibr">28</xref>, <xref rid="B29" ref-type="bibr">29</xref>). Following immunization with GAS, the rats exhibited neurological motor symptoms and compulsive behavior. These symptoms were alleviated by the D2 blocker haloperidol and the selective serotonin reuptake inhibitor paroxetine, medications that are used to treat motor symptoms and compulsions in GAS-related neuropsychiatric disorders. Recent studies published by these investigators revealed that antibodies purified from the sera of GAS-exposed rats and infused into the striatum of na&#x000ef;ve rats led to behavioral and motor alterations mimicking those seen in GAS-exposed rats (<xref rid="B29" ref-type="bibr">29</xref>). IgG from GAS-exposed rats reacted with dopamine and serotonin receptors <italic>in vitro</italic>, demonstrating the potential pathogenic role of autoantibodies produced following exposure to GAS.</p></sec><sec id="S7"><title>Summary</title><p>There are many fundamental immunological questions that need to be answered if effective RF/RHD control strategies are to be implemented worldwide. These questions include: (1) the identity and dominant hierarchy of streptococcal antigens in driving immune-mediated cardiac pathology; (2) the precise cell-mediated and antibody mechanisms involved in disease initiation and progression; and (3) whether disease can be prevented by regulating or switching off pathogenic immune responses. Targeting key processes involved in disease mechanisms, i.e., regulation of destructive immune responses directed against the host, may provide adjunct management strategies for RF/RHD. The RAV model is currently the only model available that is suitable for such investigations.</p></sec><sec id="S8"><title>Conflict of Interest Statement</title><p>The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.</p></sec>

IPM thresholds for <italic>Agriotes</italic> wireworm species in maize in Southern Europe

Could not extract abstract

<contrib contrib-type="author" corresp="yes"><name><surname>Furlan</surname><given-names>Lorenzo</given-names></name><address><phone>+39-049-8293879</phone><fax>+39-049-8293815</fax><email>lorenzo.furlan@venetoagricoltura.org</email></address><xref ref-type="aff" rid="Aff1"/></contrib><aff id="Aff1">Veneto Agricoltura, Agripolis, via dell&#x02019;Universit&#x000e0; 14, Legnaro, PD Italy </aff>

Journal of Pest Science

<sec id="Sec1" sec-type="introduction"><title>Introduction</title><p>EU Directive 2009/128/EC on the sustainable use of pesticides makes it compulsory to implement integrated pest management (IPM) for annual crops in Europe from January 2014. IPM strategies have not played a significant role in these crops to date, yet they have been widely used for crops such as orchards and vineyards. Therefore, accurate information about IPM strategies for annual crops is needed urgently, but this information must take into account that arable farming has few resources in terms of income, labour and technology. Since the use of soil insecticides is widespread, this paper intends to provide reliable IPM information to tackle wireworms, the main soil pest in Europe (Furlan <xref ref-type="bibr" rid="CR10">2005</xref>). It has proved difficult to implement IPM strategies for wireworms in Europe due to a shortage of reliable information on how to assess population levels and the relative thresholds (Furlan <xref ref-type="bibr" rid="CR10">2005</xref>). Wireworms are the larvae of click beetles (Coleoptera: Elateridae), but damage-causing genera and species vary with geographic location (Furlan et al. <xref ref-type="bibr" rid="CR14">2000</xref>, <xref ref-type="bibr" rid="CR16">2001b</xref>, <xref ref-type="bibr" rid="CR13">2007</xref>a; Rusek <xref ref-type="bibr" rid="CR32">1972</xref>; Staudacher et al. <xref ref-type="bibr" rid="CR34">2013</xref>). In Europe, most larvae in agricultural land belong to the <italic>Agriotes</italic> genus, but the specific species must be established if we are to predict the potential damage to crops. For example, high populations of <italic>Agriotes ustulatus</italic> do not damage maize late in the spring (late May&#x02013;June) because most of the larvae are in a non-feeding phase (Furlan <xref ref-type="bibr" rid="CR8">1998</xref>); in the same period, however, <italic>Agriotes sordidus</italic> or <italic>A. brevis</italic> can seriously reduce the stand of maize crops (Furlan <xref ref-type="bibr" rid="CR9">2004</xref>). The adults (click beetles) of these species can be divided into two main groups: (i) adults that do not overwinter and lay eggs a few days after swarming (<italic>A. ustulatus</italic> Sch&#x000e4;ller and <italic>A. litigiosus</italic> Rossi); and (ii) adults that overwinter, live for months, and lay eggs for a long period after adult hardening (<italic>A. sordidus</italic> Illiger, <italic>A. brevis</italic> Candeze, <italic>A. lineatus</italic> L., <italic>A. sputator</italic> L., <italic>A. obscurus</italic> L., <italic>A. rufipalpis</italic> Brull&#x000e8;, and <italic>A. proximus</italic> Schwarz) (Furlan <xref ref-type="bibr" rid="CR10">2005</xref>). The life cycle of the species in both groups is about 24&#x02013;36&#x000a0;months. In spring, the larvae of group (i) entering the bait traps come from eggs laid two years before, but group (ii) larvae come mainly from eggs laid the previous year. Unfortunately, the vast majority of literature on this matter does not report which species were involved (Hinkin <xref ref-type="bibr" rid="CR24">1976</xref>; Chabert and Blot <xref ref-type="bibr" rid="CR3">1992</xref>). Therefore, this present research assesses the effect of various <italic>Agriotes</italic> species on maize and looks at thresholds based on wireworms caught in bait traps in order to establish a range of IPM strategies. The ultimate aim of the research is to provide practical information so that European farmers can implement reliable, feasible and affordable IPM strategies to prevent wireworms damaging their maize.</p></sec><sec id="Sec2" sec-type="materials|methods"><title>Materials and methods</title><sec id="Sec3"><title>Field sites</title><p>Research was conducted in north-east Italy (area covered: 45.64N, 12.96E and 45.05N 11.88E) from 1993 to 2011 (19 consecutive years) on fields with the following characteristics: (1) soil at field water capacity, i.e. no more water can be stably retained; after winter, all of the fields studied, and particularly the bare ones, i.e. no crops consuming water, are usually very humid due to rainfall, negligible evaporation and transpiration. Sometimes strong winds dried up the soil, but only the top-most layer and not where the traps were placed. Therefore, the soil layer containing the traps was always at field water capacity; (2) bare soil (no plants growing), since traps perform reliably when they do not have to compete with plants whose roots produce carbon dioxide, which attracts larvae (Doane et al. <xref ref-type="bibr" rid="CR4">1975</xref>); (3) several previous crops had been sown, such as maize, soybean, winter cereals and meadow (e.g. alfalfa, festuca); meadow must be ploughed at least three months before the bait traps are placed in order to make sure that all ploughed-up meadow plants have died (it was observed that this takes about three months); the main reason for this procedure is that it allows the bait traps to attract wireworms without the competition of plants, as described above. Each year, monitoring was conducted in fields representing a balanced sample of agronomic conditions in north-east Italy. Part of the soils was classified with the USDA soil texture triangle based on analyses carried out in accordance with official USDA methods. Soil pH was basic for all the fields and ranged between 7.9 and 8.3.</p></sec><sec id="Sec4"><title>Agronomic practices</title><p>Conventional agronomic practices were applied to all of the fields studied (i.e. ploughing, harrowing, fertilization with 240&#x02013;300&#x000a0;N&#x000a0;kg, 70,000&#x02013;76,000 seeds/ha, interrow width 75&#x000a0;cm, pre-emergence plus post-emergence herbicide treatments causing very low weed densities, and planting date from late March to late April). The following commercial hybrids were used: ANITA, COSTANZA, ALICIA, SENEGAL (1993&#x02013;2001); TEVERE (2002&#x02013;2004); DKC6530 (2005&#x02013;2006); DKC 6530, MITIC, KERMESS, KLAXON (2007&#x02013;2008); DKC6666, NK FAMOSO, PR31A34, PR32G44 (2009&#x02013;2010); and DKC6677, PR32G44 and NK FAMOSO (2011).</p></sec><sec id="Sec5"><title>Estimation of wireworm population level</title><p>Bait traps made and used in accordance with Chabert and Blot (<xref ref-type="bibr" rid="CR3">1992</xref>) were deployed to estimate wireworm population densities from late February to mid April. These and similar traps were found to be efficient at capturing wireworms after research in UK conditions (Parker <xref ref-type="bibr" rid="CR28">1994</xref>, <xref ref-type="bibr" rid="CR29">1996</xref>). Each trap comprised a plastic pot 10&#x000a0;cm in diameter with holes in the bottom. The pots were filled with vermiculite, 30&#x000a0;ml of wheat seeds and 30&#x000a0;ml of maize seeds; they were then moistened before being placed into the soil 4&#x02013;5&#x000a0;cm below the soil surface, after which they were covered with an 18-cm diameter plastic lid placed 1&#x02013;2&#x000a0;cm above the pot rim. Traps were hand-sorted after 10&#x000a0;days when the average temperature 10&#x000a0;cm beneath the surface was above 8&#x000a0;&#x000b0;C (Furlan <xref ref-type="bibr" rid="CR8">1998</xref>, <xref ref-type="bibr" rid="CR9">2004</xref>) to ensure that the bait traps stayed in the soil for an equal period of wireworm activity. <italic>Agriotes</italic> larvae do not feed, or feed very little, at lower temperatures. Generally, the traps were removed from the fields 2 to 8&#x000a0;days before maize seeding. No considerable differences in wireworm feeding activity were observed between 8 and 13&#x000a0;&#x000b0;C, which is the usual temperature range in early spring in northern Italy (Furlan <xref ref-type="bibr" rid="CR8">1998</xref>, <xref ref-type="bibr" rid="CR9">2004</xref>). Previous investigations (unpublished data) found that only a negligible number of larvae escaped from the traps since it was noted that numbers tended to increase as days passed (Furlan personal observation). The final number of larvae was assessed under the aforementioned conditions, regardless of larvae behaviour on individual days. Population levels were calculated only on days when humidity was close to field water capacity. Dry top-soil forces larvae to burrow deep beneath the surface, away from the bait traps (Furlan <xref ref-type="bibr" rid="CR8">1998</xref>), and high humidity (flooding in extreme cases) prevents larvae activity since all the soil pores are full of water and contain no oxygen. Therefore, any days on which these conditions occurred were excluded from calculations, regardless of the soil temperature. This obviously resulted in traps sometimes being kept in the soil for longer than 10&#x000a0;days. In the UK, Parker (<xref ref-type="bibr" rid="CR28">1994</xref>) caught large numbers of <italic>Agriotes</italic> wireworms in average soil temperatures that ranged from 5 to 10&#x000a0;&#x000b0;C. In order to recover as many larvae as possible, and thus increase research precision, after 10 or more days, the traps were inspected manually and the contents put into Berlese funnels fitted with a 0.5-cm mesh screen at the bottom. The trap contents were allowed to dry for at least 20&#x000a0;days in a sheltered place without lamps, and the larvae that fell into the collecting vials were counted and identified. A personal key (unpublished), developed by rearing single larvae to adults, was used to identify them. Some of the distinguishing characteristics complied with Rudolph (<xref ref-type="bibr" rid="CR31">1974</xref>). Adults were determined with the key in Platia (<xref ref-type="bibr" rid="CR30">1994</xref>). The traps were deployed on a grid (20&#x000a0;m&#x000a0;&#x000d7;&#x000a0;10&#x000a0;m apart); a minimum of nine bait traps was placed per field and the sample area varied between 0.2 and 1&#x000a0;ha. The larger the area to be covered, the higher the number of traps placed. A total of 5,400 traps were placed during this 19-year study (18 traps/field on average). This research encompassed only fields monitored in spring (early March to late April).</p></sec><sec id="Sec6"><title>Estimation of wireworm damage to maize</title><p>In the maize fields monitored, wireworm damage to seeds and plants was assessed only once it was sure that insecticides had not been used, or that random untreated maize strips/plots, 3 or 4.5&#x000a0;m wide, had been sown alternately with treated strips/plots. When strips/plots were treated, the most effective insecticides available were used: 1993&#x02013;1994: Diphonate&#x000ae; (Fonofos 4.75&#x000a0;% a.i.) 10&#x000a0;kg/ha of granules applied in-furrow; Dotan&#x000ae; (Chlormephos 4.95&#x000a0;% a.i.) 7&#x000a0;kg/ha of granules applied in-furrow; 1995&#x02013;2005: Fipronil (Regent TS&#x000ae;) 0.6&#x000a0;mg a.i./seed; Imidacloprid (Gaucho&#x000ae;) 1.2&#x000a0;mg a.i./seed; Regent 2G&#x000ae; (Fipronil 2&#x000a0;% a.i.) 5&#x000a0;kg/ha of granules applied in-furrow; 2006&#x02013;2010: Force&#x000ae; ST (Tefluthrin 0.5&#x000a0;% a.i.) 15&#x000a0;kg/ha of granules applied in-furrow; Clothiadinin (Poncho&#x000ae;) 0.5&#x000a0;mg a.i./seed; 2011: Force&#x000ae; ST (Tefluthrin 0.5&#x000a0;% a.i.) 15&#x000a0;kg/ha of granules applied in-furrow; Clothiadinin (Poncho&#x000ae;) 0.5&#x000a0;mg a.i./seed; Imidacloprid (Gaucho&#x000ae;) 1.2&#x000a0;mg a.i./seed.</p><p>One litre of the fungicide Metalaxil&#x000a0;+&#x000a0;Fludioxonil (Celest&#x000ae;) per tonne of seed was used to treat all of the maize seeds planted. In order to study the correlation between wireworm densities (larvae/bait trap) and the damage to maize, at the 2&#x02013;3 and 6&#x02013;8 leaf stages, two sub-plots of 4&#x000a0;&#x000d7;&#x000a0;20&#x000a0;m rows of maize per portion of untreated field (0.1&#x02013;0.2&#x000a0;ha) or untreated strip were chosen at random and the plants observed. During plot trials, all plants (healthy and damaged) at the centre of each untreated plot were counted; the plots covered an area of 15&#x02013;18&#x000a0;m&#x000a0;&#x000d7;&#x000a0;1.5&#x000a0;m. The location and the number of the sub-plots were the same in both the untreated/treated strips and completely untreated field. In order to assess wireworm damage on emerged plants, plants with typical symptoms (e.g. wilting central leaves, broken central leaf due to holes in the collar, wilting of whole small plants) were sought and the soil around the plants was dug up to a depth 5&#x02013;6&#x000a0;cm; any larvae found near the collar were collected and identified. Wherever maize plants were missing from the rows, the soil was dug up in order to assess possible wireworm damage to seeds and/or emerging seedlings. Total plant damage was calculated as the sum of damage to emerged plants and seeds. In order to establish the effect of wireworm damage on yield, the same observations were made on the treated strips/plots where used. Finally, the strips and the plots were harvested and the maize grain weighed. Maize grain samples were collected and their humidity measured with a Pfeuffer-Granomat (the same machine was used for all samples each year). The four fields in which maize stands were irregular and damaged due to factors other than wireworm activity (e.g. bird damage, low emergence due to low soil moisture, flooding) were not considered. In order to isolate the &#x0201c;wireworm damage effect&#x0201d;, analysis excluded the two fields under considerable pressure from factors other than wireworms (e.g. other parasites, viruses). Fields in which the general conditions were good, but the soil insecticide had not worked completely and the stand of treated maize plots was not optimal, were not used to evaluate the effect on yield (this happened in two cases only). In the remaining fields where the insecticides had worked effectively, the final stand of the treated strips/plots was suitable for assessing whether yield had been reduced (&#x0003e;90&#x000a0;% of the sown seeds).</p></sec><sec id="Sec7"><title>Statistical methods</title><p>All analyses were performed by SAS 9.3 (SAS Institute Inc., Cary, NC). Linear regression analysis was used to determine the relationship between damage to maize (total plant damage, emerged plant damage and seed damage) and pre-seeding catches of wireworms in bait traps for each species. A paired <italic>t</italic> test was used to assess the effect of wireworm damage on grain yields in treated and non-treated plots. Where soil characteristics were available, a generalized linear model (Nelder and Wedderburn <xref ref-type="bibr" rid="CR27">1972</xref>) with a Poisson distribution was used to determine the factors affecting the percentage of total damage for each species. The model included the effect of the main agronomic characteristics (soil texture as a fixed effect, plus organic matter content and pH as covariates) and captures/plant damage data (as covariates too). The soil types (levels of the variable) were classified as follows: clay, loam, clay loam, silt clay loam, loam, sandy loam, and loamy sand. Analysis produced least squares means estimates of parameters and risk ratio. Risk ratio measures relative effect expressed by the outcome in two groups, i.e. the ratio between the prevalence in the exposed group (numerator level) vs the non-exposed group (denominator or reference value). The type of soil with the highest damage level caused by each species was chosen as reference value. Analysis was performed with PROC GENMOD.</p></sec></sec><sec id="Sec8" sec-type="results"><title>Results</title><sec id="Sec9"><title>Species composition and factors affecting the level of damage</title><p>Wireworms were found in the bait traps in 206 fields (70&#x000a0;%). The main species found were <italic>A. brevis, A. sordidus</italic> and <italic>A. ustulatus</italic>. All of these species are widespread in central and southern Europe (Furlan <xref ref-type="bibr" rid="CR7">1996</xref>, <xref ref-type="bibr" rid="CR9">2004</xref>; Furlan et al. <xref ref-type="bibr" rid="CR14">2000</xref>, <xref ref-type="bibr" rid="CR13">2007</xref>a; Kausnitzer <xref ref-type="bibr" rid="CR25">1994</xref>) including areas with significantly different conditions from those of this study, e.g. in Austria, <italic>A. brevis</italic> were found in zones with acid pH (Staudacher et al. <xref ref-type="bibr" rid="CR34">2013</xref>). The presence of other Elateridae species (mainly <italic>Synaptus filiformis</italic> Fabricius, <italic>Melanotus</italic> spp., <italic>Adrastus rachifer</italic> Geoffroy in Fourcroy) was negligible. Bait traps caught a single species in 81.1&#x000a0;% of the fields. The combinations of different species observed in the other cases are described in Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>. Only four fields (1.9&#x000a0;%) had a considerably mixed population (two or three species in a single bait trap). Table&#x000a0;<xref rid="Tab2" ref-type="table">2</xref> covers the fields in which at least one trap caught wireworms and gives the average, standard deviation and maximum value of all the single averages, standard deviations and maximum numbers estimated in each of the fields monitored. The variability between bait traps was high, and the ratio between average mean and average standard deviations was one. The generalized linear model found that the percentage of total damage variability was mainly explained by wireworm density (the average number of larvae/bait trap) for all three of the species studied (Table&#x000a0;<xref rid="Tab3" ref-type="table">3</xref>, <italic>P</italic>&#x000a0;&#x0003c;&#x000a0;0.001). Soil texture affected the risk: loam soils were prone to higher damage risk by <italic>A. sordidus</italic>, while the risk of damage by <italic>A.</italic> <italic>ustulatus</italic> was much lower in clay soils. PH variations in the range of soils studied (mean&#x000a0;=&#x000a0;8.01, SD&#x000a0;=&#x000a0;0.11) did not influence the risk of damage by any of the species, but organic matter content (mean&#x000a0;=&#x000a0;1.93, SD&#x000a0;=&#x000a0;0.49) may vary the risk of damage by <italic>A.</italic> <italic>ustulatus</italic> (Table&#x000a0;<xref rid="Tab3" ref-type="table">3</xref>, <italic>P</italic>&#x000a0;&#x0003c;&#x000a0;0.001).<table-wrap id="Tab1"><label>Table&#x000a0;1</label><caption><p>Wireworms found in the fields monitored with bait traps pre-seeding; fields are divided in accordance with the number of species found in each one</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="2">Species</th><th align="left" colspan="6">Fields</th></tr><tr><th align="left">Total</th><th align="left">One species</th><th align="left"> <italic>Agriotes brevis</italic>&#x000a0;+&#x000a0;<italic>Agriotes sordidus</italic> </th><th align="left"> <italic>Agriotes brevis</italic>&#x000a0;+&#x000a0;<italic>Agriotes ustulatus</italic> </th><th align="left"> <italic>Agriotes</italic> <italic>sordidus</italic>&#x000a0;+&#x000a0;<italic>Agriotes</italic> <italic>ustulatus</italic> </th><th align="left">All three species</th></tr></thead><tbody><tr><td align="left">Fields (no.)</td><td char="." align="char">206</td><td char="." align="char">167</td><td char="." align="char">9</td><td char="." align="char">8</td><td char="." align="char">12</td><td char="." align="char">10</td></tr><tr><td align="left"> <italic>A. brevis</italic> larvae</td><td char="." align="char">2,431</td><td char="." align="char">1,959</td><td char="." align="char">89</td><td char="." align="char">197</td><td char="." align="char">0</td><td char="." align="char">186</td></tr><tr><td align="left"> <italic>A. sordidus</italic> larvae</td><td char="." align="char">1,486</td><td char="." align="char">1,353</td><td char="." align="char">85</td><td char="." align="char">0</td><td char="." align="char">30</td><td char="." align="char">18</td></tr><tr><td align="left"> <italic>A. ustulatus</italic> larvae</td><td char="." align="char">4,217</td><td char="." align="char">3,765</td><td char="." align="char">0</td><td char="." align="char">160</td><td char="." align="char">280</td><td char="." align="char">12</td></tr></tbody></table></table-wrap> <table-wrap id="Tab2"><label>Table&#x000a0;2</label><caption><p>Variability between the number of wireworms in the single bait traps placed in fields monitored</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="2"/><th align="left" colspan="3"> <italic>Agriotes brevis</italic> </th><th align="left" colspan="3"> <italic>Agriotes sordidus</italic> </th><th align="left" colspan="3"> <italic>Agriotes ustulatus</italic> </th></tr><tr><th align="left">Mean</th><th align="left">SD</th><th align="left">Max</th><th align="left">Mean</th><th align="left">SD</th><th align="left">Max</th><th align="left">Mean</th><th align="left">SD</th><th align="left">Max</th></tr></thead><tbody><tr><td align="left">Mean</td><td char="." align="char">0.61</td><td char="." align="char">0.41</td><td align="left">1.50</td><td align="left">0.38</td><td align="left">0.49</td><td align="left">1.60</td><td char="." align="char">1.04</td><td char="." align="char">1.13</td><td align="left">4.11</td></tr><tr><td align="left">SD</td><td char="." align="char">3.27</td><td char="." align="char">1.71</td><td align="left">6.43</td><td align="left">0.65</td><td align="left">0.61</td><td align="left">2.05</td><td char="." align="char">3.25</td><td char="." align="char">2.88</td><td align="left">10.58</td></tr><tr><td align="left">Max</td><td char="." align="char">24.88</td><td char="." align="char">14.18</td><td align="left">53</td><td align="left">3.58</td><td align="left">2.99</td><td align="left">9</td><td char="." align="char">21.80</td><td char="." align="char">17.51</td><td align="left">60</td></tr></tbody></table><table-wrap-foot><p>Average, standard deviation (SD) and maximum value of the all averages, standard deviations and maximum numbers calculated per each of the fields monitored. Only fields with average higher than zero have been considered (<italic>Agriotes brevis</italic> 48 fields, <italic>Agriotes sordidus</italic> 103 fields, <italic>Agriotes ustulatus</italic> 55 fields)</p></table-wrap-foot></table-wrap> <table-wrap id="Tab3"><label>Table&#x000a0;3</label><caption><p>Least squares means (% of total damage on plants) and risk ratio for <italic>Agriotes ustulatus</italic>, <italic>Agriotes sordidus</italic> and <italic>Agriotes</italic> <italic>brevis</italic> in different soils and different pH levels, percentage of organic matter and number of larvae/trap calculated with a generalized linear model</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left">Variable</th><th align="left">Number of fields</th><th align="left">Least squares means % Damage (SE)</th><th align="left">RR (95&#x000a0;% CI)</th><th align="left">Chi-square</th><th align="left"> <italic>P</italic> </th></tr></thead><tbody><tr><td align="left" colspan="6">Agriotes ustulatus</td></tr><tr><td align="left">&#x000a0;Soil</td><td char="." align="char"/><td char="(" align="char"/><td align="left"/><td char="." align="char">25.96</td><td char="." align="char">&#x0003c;0.001</td></tr><tr><td align="left">&#x000a0;&#x000a0;Clay</td><td char="." align="char">7</td><td char="(" align="char">0.27 (0.12)</td><td char="(" align="char">0.22 (0.19&#x02013;0.25)</td><td char="." align="char">12.86</td><td char="." align="char">&#x0003c;0.001</td></tr><tr><td align="left">&#x000a0;&#x000a0;Loam^a </td><td char="." align="char">31</td><td char="(" align="char">1.24(0.23)</td><td char="(" align="char"/><td char="." align="char"/><td char="." align="char"/></tr><tr><td align="left">&#x000a0;&#x000a0;Clay loam</td><td char="." align="char">3</td><td char="(" align="char">0.08 (0.13)</td><td char="(" align="char">0.06 (0.002&#x02013;1.78)</td><td char="." align="char">2.62</td><td char="." align="char">0.105</td></tr><tr><td align="left">pH</td><td char="." align="char"/><td char="(" align="char"/><td char="(" align="char">&#x02013;</td><td char="." align="char">&#x02013;</td><td char="." align="char">&#x02013;</td></tr><tr><td align="left">Organic matter (%)</td><td char="." align="char"/><td char="(" align="char"/><td char="(" align="char">2.04 (1.23&#x02013;3.39)</td><td char="." align="char">7.53</td><td char="." align="char">&#x0003c;0.001</td></tr><tr><td align="left">No. larvae/trap</td><td char="." align="char"/><td char="(" align="char"/><td char="(" align="char">1.25 (1.21&#x02013;1.28)</td><td char="." align="char">455.42</td><td char="." align="char">&#x0003c;0.001</td></tr><tr><td align="left" colspan="6">Agriotes sordidus</td></tr><tr><td align="left">&#x000a0;Soil</td><td char="." align="char"/><td char="(" align="char"/><td char="(" align="char"/><td char="." align="char">67.50</td><td char="." align="char">&#x0003c;0.001</td></tr><tr><td align="left">&#x000a0;&#x000a0;Silty clay loam</td><td char="." align="char">2</td><td char="(" align="char">1.37 (1.08)</td><td char="(" align="char">0.35 (0.07&#x02013;1.67)</td><td char="." align="char">1.74</td><td char="." align="char">0.187</td></tr><tr><td align="left">&#x000a0;&#x000a0;Loam</td><td char="." align="char">9</td><td char="(" align="char">1.30 (0.44)</td><td char="(" align="char">0.33 (0.16&#x02013;0.67)</td><td char="." align="char">9.29</td><td char="." align="char">0.002</td></tr><tr><td align="left">&#x000a0;&#x000a0;Clay loam</td><td char="." align="char">15</td><td char="(" align="char">0.63 (0.19)</td><td char="(" align="char">0.16 (0.09&#x02013;0.29)</td><td char="." align="char">36.05</td><td char="." align="char">&#x0003c;0.001</td></tr><tr><td align="left">&#x000a0;&#x000a0;Silty clay loam</td><td char="." align="char">12</td><td char="(" align="char">2.18 (0.63)</td><td char="(" align="char">0.55 (28&#x02013;1.06)</td><td char="." align="char">3.16</td><td char="." align="char">0.076</td></tr><tr><td align="left">&#x000a0;&#x000a0;Sandy loam</td><td char="." align="char">9</td><td char="(" align="char">2.29 (0.52)</td><td char="(" align="char">0.58 (0.36&#x02013;0.93)</td><td char="." align="char">5.07</td><td char="." align="char">0.024</td></tr><tr><td align="left">&#x000a0;&#x000a0;Loamy sand^a </td><td char="." align="char">32</td><td char="(" align="char">3.96 (0.46)</td><td char="(" align="char"/><td char="." align="char"/><td char="." align="char"/></tr><tr><td align="left">pH</td><td char="." align="char"/><td char="(" align="char"/><td char="(" align="char">0.24 (0.03&#x02013;2.22)</td><td char="." align="char">1.50</td><td char="." align="char">0.221</td></tr><tr><td align="left">Organic matter (%)</td><td char="." align="char"/><td char="(" align="char"/><td char="(" align="char">0.59 (0.25&#x02013;1.37)</td><td char="." align="char">1.74</td><td char="." align="char">0.188</td></tr><tr><td align="left">No. larvae/trap</td><td char="." align="char"/><td char="(" align="char"/><td char="(" align="char">1.96 (1.74&#x02013;2.21)</td><td char="." align="char">106.19</td><td char="." align="char">&#x0003c;0.001</td></tr><tr><td align="left" colspan="6">Agriotes brevis</td></tr><tr><td align="left" colspan="6">&#x000a0;Soil</td></tr><tr><td align="left">&#x000a0;&#x000a0;Clay</td><td char="." align="char">11</td><td char="(" align="char">11.73 (1.19)</td><td char="(" align="char">0.55 (0.32&#x02013;0.94)</td><td char="." align="char">4.78</td><td char="." align="char">0.029</td></tr><tr><td align="left">&#x000a0;&#x000a0;Loam</td><td char="." align="char">4</td><td char="(" align="char">2.84 (1.13)</td><td char="(" align="char">0.13 (0.04&#x02013;0.40)</td><td char="." align="char">12.69</td><td char="." align="char">&#x0003c;0.001</td></tr><tr><td align="left">&#x000a0;&#x000a0;Clay loam</td><td char="." align="char">8</td><td char="(" align="char">8.45 (1.26)</td><td char="(" align="char">0.40 (0.23&#x02013;0.67)</td><td char="." align="char">11.96</td><td char="." align="char">&#x0003c;0.001</td></tr><tr><td align="left">&#x000a0;&#x000a0;Loamy sand^a </td><td char="." align="char">2</td><td char="(" align="char">21.36 (5.58)</td><td char="(" align="char"/><td char="." align="char"/><td char="." align="char"/></tr><tr><td align="left">pH</td><td char="." align="char"/><td char="(" align="char"/><td char="(" align="char">15.40 (0.29&#x02013;&#x0003e;20)</td><td char="." align="char">1.82</td><td char="." align="char">0.178</td></tr><tr><td align="left">Organic matter (%)</td><td char="." align="char"/><td char="(" align="char"/><td char="(" align="char">1.27 (0.56&#x02013;2.88)</td><td char="." align="char">0.32</td><td char="." align="char">0.569</td></tr><tr><td align="left">No. larvae/trap</td><td char="." align="char"/><td char="(" align="char"/><td char="(" align="char">1.07 (1.06&#x02013;1.09)</td><td char="." align="char">128.49</td><td char="." align="char">&#x0003c;0.001</td></tr></tbody></table><table-wrap-foot><p> <italic>RR</italic> risk ratio<italic>, SE</italic> standard error, <italic>CI</italic> confidence interval</p><p> ^aRepresents the reference level of comparison in the calculation of risk ratio</p></table-wrap-foot></table-wrap> </p></sec><sec id="Sec10"><title>The correlation between species caught by bait traps and symptoms observed on maize plants</title><p>Symptoms on maize plants varied per wireworm species. Wherever <italic>A. ustulatus</italic> was the prevalent species, no significant symptoms were found on emerged maize plants (e.g. wilting central leaves); see Table&#x000a0;<xref rid="Tab5" ref-type="table">5</xref> and Fig.&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>. Symptoms on emerged plants were always caused by <italic>A. brevis</italic> and/or <italic>A. sordidus</italic> (Table&#x000a0;<xref rid="Tab2" ref-type="table">2</xref>; Fig.&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>). Only one case of serious seed damage by <italic>A. brevis</italic> was observed; the maize had been sown late and the seeds germinated in May due to a prolonged dry spell. No significant seed damage by <italic>A. sordidus</italic> larvae was observed. <italic>A. ustulatus</italic> larvae (Table&#x000a0;<xref rid="Tab5" ref-type="table">5</xref>) significantly affected plant stand by damaging seeds, which could not germinate or emerge when the population was high. Few maize plants had the central leaf broken by <italic>A. ustulatus</italic> feeding below ground; in the fields where <italic>A. ustulatus</italic> was the prevalent species, less than 0.1&#x000a0;% of plants were damaged (3 out of a total of 3,100 seeds&#x000a0;+&#x000a0;plants found damaged). Broken central leaves were restricted to the 3&#x02013;4 leaf stage. <italic>A. brevis</italic> and <italic>A. sordidus</italic> proved able to cause all of the possible symptoms and to damage even developed maize plants (up to the 8&#x02013;10 leaf stage). Most of the damaged plants had one or more wilted central leaves due to larval feeding on the collar, which sometimes killed them.<fig id="Fig1"><label>Fig.&#x000a0;1</label><caption><p>The relationship between wireworm density (number of wireworms/bait trap) and total plant damage (plants/m&#x000b2;) for <italic>Agriotes ustulatus</italic>, <italic>A. brevis</italic> and <italic>A. sordidus</italic> (&#x000b1;95&#x000a0;% average confidence level). <italic>Larger</italic> (rhomb) <italic>dots</italic> represent combinations that resulted in a significant yield reduction</p></caption><graphic xlink:href="10340_2014_583_Fig1_HTML" id="MO1"/></fig> </p></sec><sec id="Sec11"><title>The correlation between species caught by bait traps and damage to maize</title><p>All or most of the larvae collected from damaged seeds, seedlings or plants belonged to the prevalent species captured by the bait traps (Table&#x000a0;<xref rid="Tab4" ref-type="table">4</xref>). A significant correlation (for all species) was discovered between the average number of wireworms caught in bait traps and the total damage to maize (damage to seeds, plus damage to emerged plants; Table&#x000a0;<xref rid="Tab5" ref-type="table">5</xref>, Fig.&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>). <italic>A. brevis</italic> was the most harmful species, as even wireworm densities just over one wireworm/trap caused considerable plant damage (one to two plants attacked/m&#x000b2;), i.e. enough to reduce yield (Fig.&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>; Table&#x000a0;<xref rid="Tab6" ref-type="table">6</xref>). The graph shows either a very low or a high population (only three fields had a very high population) and almost nothing in between. During this 19-year research, high <italic>A. brevis</italic> populations were found in maize fields after meadow had been ploughed, or after a soil had been continuously covered with vegetation (e.g. soybeans just after winter-wheat in the same growing season). After the first year of maize, the wireworm populations decreased dramatically; this means that high populations are possible, but uncommon, as they occurred only in a few meadows and fields where crops were continuously planted. Low populations, however, were common, as levels fell the very next spring, and usually remained low for several years after. Intermediate populations are therefore rare. To cause the same level of damage in maize fields, five times more <italic>A. ustulatus</italic> larvae are needed (Fig.&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>; Table&#x000a0;<xref rid="Tab6" ref-type="table">6</xref>). In Fig.&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>, the notable outlier in the <italic>A. ustulatus</italic> graph concerns a 2010 trial; the results may be explained by a cold spring and a very compact soil, which significantly slowed the emergence of maize seedlings, leaving them in the soil for a long time (about 20&#x000a0;days). These soil and climatic conditions did not cause significant damage in other fields with lower wireworm populations. <italic>A.</italic> <italic>ustulatus</italic> caused almost identical total damage and seed damage because it harmed very few emerged plants; on the contrary, very few maize seeds were damaged by <italic>A.</italic> <italic>brevis</italic> and <italic>A. sordidus</italic>. <italic>A. sordidus</italic> was the second most harmful with wireworm densities above two larvae/trap leading to reduced yield (Fig.&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>; Table&#x000a0;<xref rid="Tab6" ref-type="table">6</xref>). In Fig.&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref> (<italic>A. sordidus</italic>), the outlier fields, which experienced a significant decrease in yield, had sandy loam soils. Similar population levels did not cause serious damage in heavy soils. In most fields (0&#x02013;1 larva/trap), wireworm damage was negligible and did not cause any visible effects on maize crops, i.e. less than 5&#x000a0;% of plants were attacked and, in most cases, they partially or completely recovered. In some cases, damage of over 1 plant/m&#x000b2; led to significant yield reduction (Fig.&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>). Nevertheless, in others, even very severe plant damage (&#x0003e;3 plants/m&#x000b2;; &#x0003e;40&#x000a0;%) did not result in reduced yield. For example, in the same year (2011), severe plant damage (&#x0003e;3 plants/m&#x000b2;; &#x0003e;50&#x000a0;%) resulted in significant yield reduction at one site, but another trial produced no difference between untreated plots (8.74&#x000a0;t/ha) and Imidacloprid-treated plots (8.59&#x000a0;t/ha), despite the treated plots giving much higher stands than untreated ones in both trials. Plant damage below 1 plant/m&#x000b2; never resulted in significant yield reduction, and there were very limited differences (ranging between 0.01 and 0.3&#x000a0;t/ha) between treated and untreated strips or plots (see Furlan et al. <xref ref-type="bibr" rid="CR17">2002</xref>, <xref ref-type="bibr" rid="CR18">2007</xref>, <xref ref-type="bibr" rid="CR19">2009a</xref>, <xref ref-type="bibr" rid="CR20">b</xref>, <xref ref-type="bibr" rid="CR22">2011</xref>). The 2011 study confirmed the previous long-term observations (Table&#x000a0;<xref rid="Tab7" ref-type="table">7</xref>). A further field infested by <italic>A. brevis</italic> (damage &#x0003e;3 plants/m&#x000b2;; &#x0003e;50&#x000a0;%) experienced a significant yield reduction of 4.2&#x000a0;t/ha. The hybrid was PR32G44. Wherever wireworm densities of <italic>A. ustulatus</italic> were lower than five larvae/trap and <italic>A. sordidus</italic> were lower than two larvae/trap (Fig.&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>), stand reduction was lower than 0.5 plants/m&#x000b2; (in most cases, less than 5&#x000a0;% of total plants); no fields experienced reduced yield (i.e. there were no significant differences between treated and untreated strips/plot (Table&#x000a0;<xref rid="Tab6" ref-type="table">6</xref>; Fig.&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>).<table-wrap id="Tab4"><label>Table&#x000a0;4</label><caption><p>Wireworm species identified as damaging maize seeds and plants in fields monitored with bait traps expressed as a percentage of the total number of larvae collected from damaged plants</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left">Fields (no.)</th><th align="left">Species in bait traps</th><th align="left"> <italic>Agriotes ustulatus</italic> </th><th align="left"> <italic>Agriotes brevis</italic> </th><th align="left"> <italic>Agriotes sordidus</italic> </th><th align="left">Others</th><th align="left">Total number of larvae</th></tr></thead><tbody><tr><td align="left">30</td><td align="left"> <italic>Agriotes ustulatus</italic> </td><td char="." align="char">99.5</td><td char="." align="char">0.2</td><td char="." align="char">0.2</td><td align="left">0.1</td><td char="." align="char">1,015</td></tr><tr><td align="left">31</td><td align="left"> <italic>Agriotes brevis</italic> </td><td char="." align="char">0.1</td><td char="." align="char">99.6</td><td char="." align="char">0.2</td><td align="left">0.1</td><td char="." align="char">754</td></tr><tr><td align="left">88</td><td align="left"> <italic>Agriotes sordidus</italic> </td><td char="." align="char">0.1</td><td char="." align="char">0.2</td><td char="." align="char">99.7</td><td align="left">0.0</td><td char="." align="char">622</td></tr></tbody></table><table-wrap-foot><p>This table considers only fields where bait traps caught larvae belonging to one species</p></table-wrap-foot></table-wrap> <table-wrap id="Tab5"><label>Table&#x000a0;5</label><caption><p>Statistical outputs of the linear relationships between damage to maize and pre-seeding catches of wireworms (<italic>Agriotes</italic> <italic>brevis</italic>, <italic>Agriotes sordidus</italic>, <italic>Agriotes ustulatus</italic>) in bait traps</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="2">Fields (no.)</th><th align="left" rowspan="2">Species</th><th align="left" colspan="2">Total plant damage (plants/m&#x000b2;)</th><th align="left" colspan="2">Seed damage (n/m&#x000b2;)</th><th align="left" colspan="2">Emerged plant damage (plants/m&#x000b2;)</th></tr><tr><th align="left">R&#x000b2;</th><th align="left"> <italic>P</italic> </th><th align="left">R&#x000b2;</th><th align="left"> <italic>P</italic> </th><th align="left">R&#x000b2;</th><th align="left"> <italic>P</italic> </th></tr></thead><tbody><tr><td align="left">69</td><td align="left"> <italic>Agriotes brevis</italic> </td><td align="left">0.621</td><td align="left">&#x0003c;0.0001</td><td align="left">0.002</td><td align="left">0.709</td><td align="left">0.610</td><td char="." align="char">&#x0003c;0.0001</td></tr><tr><td align="left">135</td><td align="left"> <italic>Agriotes sordidus</italic> </td><td align="left">0.380</td><td align="left">&#x0003c;0.0001</td><td align="left">Not found</td><td align="left">Not found</td><td align="left">0.380</td><td char="." align="char">&#x0003c;0.0001</td></tr><tr><td align="left">93</td><td align="left"> <italic>Agriotes ustulatus</italic> </td><td align="left">0.467</td><td align="left">&#x0003c;0.0001</td><td align="left">0.469</td><td align="left">&#x0003c;0.0001</td><td align="left">0.011</td><td char="." align="char">0.326</td></tr></tbody></table><table-wrap-foot><p>&#x0201c;Total plant damage&#x0201d; is number of missing plants due to wireworm feeding on seeds (seed damage)&#x000a0;+&#x000a0;number of emerged plants damaged by wireworms (e.g. wilting of central leaves due to feeding on plant collars, broken central leaves)</p></table-wrap-foot></table-wrap> <table-wrap id="Tab6"><label>Table&#x000a0;6</label><caption><p>Percentage of fields where significant yield reductions occurred at different densities of the <italic>Agriotes</italic> wireworm species being studied (the average numbers of wireworms/trap were considered)</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left">Wireworm species</th><th align="left">Wireworm catches (larvae/trap)</th><th align="left">Fields sampled (no.)</th><th align="left">Fields with yield reduction (no.)</th><th align="left">Fields with yield reduction (%)</th></tr></thead><tbody><tr><td align="left" rowspan="6"> <italic>Agriotes ustulatus</italic> </td><td align="left">0</td><td char="." align="char">38</td><td align="left">0</td><td char="." align="char">0.0</td></tr><tr><td align="left">0.1&#x02013;1</td><td char="." align="char">25</td><td align="left">0</td><td char="." align="char">0.0</td></tr><tr><td align="left">1.01&#x02013;2</td><td char="." align="char">7</td><td align="left">0</td><td char="." align="char">0.0</td></tr><tr><td align="left">2.01&#x02013;5</td><td char="." align="char">9</td><td align="left">0</td><td char="." align="char">0.0</td></tr><tr><td align="left"> <bold>5.01&#x02013;10</bold> </td><td char="." align="char">9</td><td align="left">1</td><td char="." align="char">11.1</td></tr><tr><td align="left"> <bold>&#x0003e;10.01</bold> </td><td char="." align="char"> <bold>5</bold> </td><td align="left"> <bold>2</bold> </td><td char="." align="char"> <bold>40.0</bold> </td></tr><tr><td align="left" rowspan="5"> <italic>Agriotes brevis</italic> </td><td align="left">0</td><td char="." align="char">21</td><td align="left">0</td><td char="." align="char">0</td></tr><tr><td align="left">0.1&#x02013;1</td><td char="." align="char">32</td><td align="left">0</td><td char="." align="char">0.0</td></tr><tr><td align="left"> <bold>1.01&#x02013;2</bold> </td><td char="." align="char"> <bold>6</bold> </td><td align="left"> <bold>2</bold> </td><td char="." align="char"> <bold>33.3</bold> </td></tr><tr><td align="left"> <bold>2.01&#x02013;5</bold> </td><td char="." align="char"> <bold>7</bold> </td><td align="left"> <bold>4</bold> </td><td char="." align="char"> <bold>57.1</bold> </td></tr><tr><td align="left"> <bold>&#x0003e;5.01</bold> </td><td char="." align="char"> <bold>3</bold> </td><td align="left"> <bold>1</bold> </td><td char="." align="char"> <bold>33.3</bold> </td></tr><tr><td align="left" rowspan="4"> <italic>Agriotes sordidus</italic> </td><td align="left">0</td><td char="." align="char">32</td><td align="left">0</td><td char="." align="char">0.0</td></tr><tr><td align="left">0.1&#x02013;1</td><td char="." align="char">83</td><td align="left">0</td><td char="." align="char">0.0</td></tr><tr><td align="left">1.01&#x02013;2</td><td char="." align="char">10</td><td align="left">0</td><td char="." align="char">0.0</td></tr><tr><td align="left"> <bold>&#x0003e;2.01</bold> </td><td char="." align="char"> <bold>10</bold> </td><td align="left"> <bold>3</bold> </td><td char="." align="char"> <bold>30.0</bold> </td></tr></tbody></table><table-wrap-foot><p>Bold values indicate the population levels that resulted in yield reduction</p></table-wrap-foot></table-wrap> <table-wrap id="Tab7"><label>Table&#x000a0;7</label><caption><p>Maize grain yield (t/ha of grain at 14&#x000a0;% humidity) in a random subset of fields with &#x0003c;5&#x000a0;% (0.2 plants/m&#x000b2;) wireworm (<italic>A. sordidus</italic> Illiger) damage in untreated and treated plots with two different maize hybrids in 2011</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left">Treatments/hybrids</th><th align="left"> <italic>KORIMBOS</italic> </th><th align="left"> <italic>DKC6677</italic> </th></tr></thead><tbody><tr><td align="left">Untreated</td><td align="left">11.19</td><td align="left">13.40</td></tr><tr><td align="left"> <italic>Tefluthrin</italic> </td><td align="left">11.34</td><td align="left">N/A</td></tr><tr><td align="left"> <italic>Clothiadinin</italic> </td><td align="left">N/A</td><td align="left">13.49</td></tr><tr><td align="left">Df/t/P</td><td align="left">27/-0.550/0.587</td><td align="left">21/-0.330/0.744</td></tr></tbody></table><table-wrap-foot><p> <italic>N/A</italic> unavailable data, <italic>Df/t/P</italic> degrees of freedom, <italic>t</italic>-value, <italic>P</italic>-value</p></table-wrap-foot></table-wrap> </p></sec></sec><sec id="Sec12" sec-type="discussion"><title>Discussion</title><p>This long-term research found a significant correlation between the number of wireworms caught in bait traps before seeding and damage to maize plants caused by three of Europe&#x02019;s main wireworm species: <italic>A. brevis, A. sordidus</italic> and <italic>A. ustulatus</italic>. Over the last 19&#x000a0;years, whatever the hybrid, and regardless of agronomic and climatic conditions, no yield reduction was observed when <italic>A. brevis</italic> populations were lower than one larva/trap, <italic>A. sordidus</italic> populations were lower than two larvae/trap and <italic>A. ustulatus</italic> populations were lower than five larvae/trap. These should be considered reliable thresholds for each species. Populations were assessed via the deployment of at least nine bait traps in a sample soil grid (20&#x000a0;m&#x000a0;&#x000d7;&#x000a0;10&#x000a0;m). Although statistical analyses show that much of the variability in wireworm plant damage cannot be explained by the wireworm densities estimated by the bait traps, i.e. high wireworm density does not always mean high damage, this study did demonstrate that serious plant damage resulting in yield reduction may only occur when wireworm populations are above the thresholds established above, provided that precise conditions occur.</p><sec id="Sec13"><title>Conditions needed to use the thresholds</title><p>In order to use the thresholds established, the following conditions have to be satisfied: (i) no alternative food sources are available, soil is bare, and if meadow (e.g. alfalfa, festuca) has been cultivated previously, the field must have been ploughed at least three months before the bait traps are placed (no other previously grown crops have any particular requirements); (ii) average soil temperature 10&#x000a0;cm beneath the surface is above 8&#x000a0;&#x000b0;C for 10&#x000a0;days (including non-consecutive days); soil humidity is near to field water capacity, but days when soil humidity is over water capacity (soil pores filled with water, i.e. flooding) are not to be considered, regardless of soil temperature, since the wireworms are not active. These can be considered reliable, prudent economic thresholds for the implementation of IPM in maize in Italy and probably in the countries where the studied species are present in similar agronomic and climatic conditions. When trap catches are below the established thresholds, the probability of economic damage is negligible. However, although significant yield reduction is a risk when thresholds are exceeded, it may not always occur, as a combination of climatic and agronomic factors (e.g. hybrid, soil, rainfall, fertilization, irrigation) may compensate for stand reduction. In most cases, yield did not fall. Several factors may influence trap catches, including: (i) alternative food sources (Parker <xref ref-type="bibr" rid="CR29">1996</xref>); (ii) soil temperature (Furlan <xref ref-type="bibr" rid="CR8">1998</xref>, <xref ref-type="bibr" rid="CR9">2004</xref>; Chabert and Blot <xref ref-type="bibr" rid="CR3">1992</xref>); and (iii) soil moisture usually suitable for wireworm activity in spring in Italy and many other European countries. Thresholds, however, do need to be evaluated for different species and, for the species considered in this manuscript, under other conditions.</p></sec><sec id="Sec14"><title>Practical implementation of thresholds</title><p>Thresholds expressed as the number of wireworms per m&#x000b2;, or per trap, that do not specify the species caught (e.g. Hinkin <xref ref-type="bibr" rid="CR24">1976</xref>) do little to help IPM. Chabert and Blot (<xref ref-type="bibr" rid="CR3">1992</xref>) suggest one wireworm/trap as a threshold for early planted maize based on their observations in northern France. Their work, however, does not discriminate the larvae captured and provides no statistics. From a practical point of view, the prevalent <italic>A.</italic> species in fields intended for maize crops need to be identified if the correct IPM thresholds are to be established. This could be achieved by: (a) a quick binocular observation of representative larvae samples collected from fields (this needs trained people; currently a trained technician can identify about 40 larvae/h); (b) PCR-based identification (Ellis et al. <xref ref-type="bibr" rid="CR5">2009</xref>; Staudacher et al. <xref ref-type="bibr" rid="CR33">2010</xref>); and (c) indirectly evaluating: (i) information from click beetle monitoring with pheromone traps (Furlan et al. <xref ref-type="bibr" rid="CR15">2001a</xref>; Furlan and T&#x000f3;th <xref ref-type="bibr" rid="CR13">2007</xref>; T&#x000f3;th et al. <xref ref-type="bibr" rid="CR35">2003</xref>) since captured click beetles may be correlated with the presence in the soil of same-species larvae, at least for the three main species considered herein (Furlan et al. <xref ref-type="bibr" rid="CR16">2001b</xref>) while this is uncertain for other important European species, such as <italic>A. obscurus</italic> L., <italic>A. lineatus</italic> L. and <italic>A. sputator</italic> L. (Benefer et al. <xref ref-type="bibr" rid="CR1">2012</xref>; Blackshaw and Hicks <xref ref-type="bibr" rid="CR2">2013</xref>; Landl et al. <xref ref-type="bibr" rid="CR26">2010</xref>); and (ii) the characteristics of the field (Blackshaw and Hicks <xref ref-type="bibr" rid="CR2">2013</xref>; Furlan et al. <xref ref-type="bibr" rid="CR22">2011</xref>; Hermann et al. <xref ref-type="bibr" rid="CR23">2013</xref>; Staudacher et al. <xref ref-type="bibr" rid="CR34">2013</xref>). From a practical point of view, when a restricted area is monitored, the main <italic>Agriotes</italic> species can be easily determined because the number of the main species is limited, and a trained IPM technician can therefore identify the larvae of the few species present based on their few discriminating characteristics. Furthermore, when field information (e.g. rotation, click beetle captures) is collected and mapped properly, technicians will only need to determine a few larvae to garner reliable information about the species involved, as the species in a field tends to remain the same for at least about 4&#x02013;5&#x000a0;years if conditions remain unchanged (Furlan, personal observation). Further studies on the agronomic factors influencing crop response to wireworm damage (e.g. hybrids compensating for stand reduction) may improve the correlation between wireworm density and maize damage, as well as provide accurate (probably higher) thresholds for other groups of hybrids and for a range of conditions (e.g. irrigated or non-irrigated fields).</p></sec></sec><sec id="Sec15" sec-type="conclusion"><title>Conclusion</title><p>The information herein may be used immediately to implement IPM and to tackle soil pests attacking maize in many European regions. As a result, it may lead to a considerable reduction in the use of soil pesticides and in a fall in the environmental impact of agriculture without negative repercussions on farmers&#x02019; income. This can be achieved with the procedure described in Furlan <xref ref-type="bibr" rid="CR10">(2005)</xref>: (i) locate the areas with a serious risk of wireworm attacks by assessing field/environmental factors (Hermann et al. <xref ref-type="bibr" rid="CR23">2013</xref>; Furlan and Talon <xref ref-type="bibr" rid="CR12">1997</xref>; Furlan et al. <xref ref-type="bibr" rid="CR22">2011</xref>; Staudacher et al. <xref ref-type="bibr" rid="CR34">2013</xref>); (ii) in areas at risk of wireworm attacks, assess current <italic>Agriotes</italic> populations with the aforementioned procedure, i.e. use bait traps and assess the actual average larval population, in fields intended for maize sowing; (iii) if the average number of wireworms does not exceed the thresholds established, maize may be sown without any treatment; if the average number of wireworms does exceed at least one of the thresholds, farmers have the option of moving maize to a no-risk field, as well as of applying organic treatments (Furlan <xref ref-type="bibr" rid="CR11">2007</xref>; Furlan et al. <xref ref-type="bibr" rid="CR20">2009b</xref>, <xref ref-type="bibr" rid="CR21">2010</xref>), or chemical treatments (Furlan et al. <xref ref-type="bibr" rid="CR18">2007</xref>, <xref ref-type="bibr" rid="CR22">2011</xref> and Ferro and Furlan <xref ref-type="bibr" rid="CR6">2012</xref>). The aforementioned procedure may be considered the first reliable practical contribution towards implementing IPM of wireworms in Europe in accordance with EU Directive 2009/128/EC.</p></sec>

Insights into Genetic and Epigenetic Determinants with Impact on Vitamin D Signaling and Cancer Association Studies: The Case of Thyroid Cancer

<p>Vitamin D is a key regulator of calcium metabolism and has been implicated as a cancer preventive agent. However, clinical studies have revealed conflicting results on its cancer preventive properties, attributed in part to multiple metabolic and regulatory factors susceptible to affect individual responses to exogenous vitamin D. Vitamin D is obtained from dietary sources and sun exposure, which depends on numerous parameters such as skin type, latitude, and lifestyle factors. Focusing on thyroid cancer (TC), we document that genetic and epigenetic determinants can greatly impact individual response to vitamin D and may outweigh the classical clinical correlative studies that focus on sun exposure/dietary intake factors. In particular, genetic determinants innate to host intrinsic metabolic pathways such as highly polymorphic cytochromes P450s responsible for the metabolic activation of vitamin D are expressed in many organs, including the thyroid gland and can impact vitamin D interaction with its nuclear receptor (VDR) in thyroid tissue. Moreover, downstream regulatory pathways in vitamin D signaling as well as VDR are also subject to wide genetic variability among human populations as shown by genome-wide studies. These genetic variations in multiple components of vitamin D pathways are critical determinants for the revaluation of the potential preventive and anticancer properties of vitamin D in TC.</p>

<contrib contrib-type="author"><name><surname>Morand</surname><given-names>Gr&#x000e9;goire B.</given-names></name><xref ref-type="aff" rid="aff1">¹</xref><xref ref-type="aff" rid="aff2">²</xref><uri xlink:type="simple" xlink:href="http://frontiersin.org/people/u/184739"/></contrib><contrib contrib-type="author"><name><surname>da Silva</surname><given-names>Sabrina Daniela</given-names></name><xref ref-type="aff" rid="aff1">¹</xref><xref ref-type="aff" rid="aff2">²</xref><uri xlink:type="simple" xlink:href="http://frontiersin.org/people/u/55730"/></contrib><contrib contrib-type="author"><name><surname>Hier</surname><given-names>Michael P.</given-names></name><xref ref-type="aff" rid="aff1">¹</xref></contrib><contrib contrib-type="author"><name><surname>Alaoui-Jamali</surname><given-names>Moulay A.</given-names></name><xref ref-type="aff" rid="aff2">²</xref><xref ref-type="corresp" rid="cor1">*</xref><uri xlink:type="simple" xlink:href="http://frontiersin.org/people/u/33447"/></contrib>

Frontiers in Oncology

<sec sec-type="introduction" id="S1"><title>Introduction</title><p>Thyroid cancer (TC) is the most common endocrine cancer malignancy worldwide (<xref rid="B1" ref-type="bibr">1</xref>) with a rising incidence in particular among young patients and women (<xref rid="B2" ref-type="bibr">2</xref>&#x02013;<xref rid="B4" ref-type="bibr">4</xref>). Overdiagnosis of subclinical disease, previously proposed as a contributor for the rising incidence, cannot explain the full extent of the increase (<xref rid="B5" ref-type="bibr">5</xref>, <xref rid="B6" ref-type="bibr">6</xref>). Risk factors such as exposure to ionizing radiation (<xref rid="B7" ref-type="bibr">7</xref>&#x02013;<xref rid="B10" ref-type="bibr">10</xref>), chemical genotoxins (<xref rid="B11" ref-type="bibr">11</xref>&#x02013;<xref rid="B13" ref-type="bibr">13</xref>), and obesity (<xref rid="B14" ref-type="bibr">14</xref>&#x02013;<xref rid="B17" ref-type="bibr">17</xref>), as well as a lack of protective factors, such as vitamin D deficiency have been implicated in TC increased incidence (<xref rid="B18" ref-type="bibr">18</xref>&#x02013;<xref rid="B21" ref-type="bibr">21</xref>).</p><p>Vitamin D, an active ingredient of cod-liver oil, was first identified as a cure for rickets in the nineteenth century and has emerged as a principal regulator of calcium homeostasis (<xref rid="B22" ref-type="bibr">22</xref>). Cutaneous exposure to sun and dietary intake are the two main natural sources of vitamin D. Vitamin D activity depends on metabolic activation through hydroxylation of the 25 followed by the 1 position of the molecule by cytochromes P450s, which generate the biologically active metabolite 1,25(OH)₂D3. The action of vitamin D occurs mainly through its binding to the nuclear vitamin D receptor (VDR), which acts as a hormone-regulated transcription factor (<xref rid="B23" ref-type="bibr">23</xref>). Upon activation, the VDR forms a heterodimer with related retinoid-X receptors and binds to vitamin D response elements (VDREs) on chromatin regions resulting in the regulation of the expression of several target genes (<xref rid="B24" ref-type="bibr">24</xref>&#x02013;<xref rid="B26" ref-type="bibr">26</xref>). VDRE binding by the VDR provides the principle mechanism by which the receptor can activate gene transcription. However, the hormone-bound receptor can also repress gene transcription by a variety of mechanisms (<xref rid="B27" ref-type="bibr">27</xref>). Downstream targets of the receptor are involved in mineral metabolism, but VDR also regulates a variety of other metabolic pathways, many of which are components of immune response and cancer signaling (<xref rid="B28" ref-type="bibr">28</xref>, <xref rid="B29" ref-type="bibr">29</xref>).</p><p>Independent studies support that circulating levels of vitamin D are inversely correlated to several malignancies, including colorectal cancer (<xref rid="B30" ref-type="bibr">30</xref>, <xref rid="B31" ref-type="bibr">31</xref>), prostate cancer (<xref rid="B32" ref-type="bibr">32</xref>), breast cancer (<xref rid="B33" ref-type="bibr">33</xref>, <xref rid="B34" ref-type="bibr">34</xref>), and head and neck squamous cell carcinoma (<xref rid="B35" ref-type="bibr">35</xref>, <xref rid="B36" ref-type="bibr">36</xref>). As well, a more recent meta-analysis reported a correlation between vitamin D deficiency and poorer prognosis in several tumor types (<xref rid="B37" ref-type="bibr">37</xref>). In TC, several studies point toward a role for impaired 1,25(OH)₂D3-VDR signaling in the occurrence and progression of the disease (<xref rid="B38" ref-type="bibr">38</xref>). This review addresses new insights into genetic and epigenetic determinants of vitamin D response in relation to cancer risk focusing on TC. We provide a systematic review and analysis of experimental and clinical data and the impact of genome-wide analyses on individual susceptibility to TC.</p></sec><sec sec-type="materials|methods" id="S2"><title>Materials and Methods</title><sec id="S2-1"><title>Genomic database</title><p>The UCSC Cancer Genomics Browser (<xref rid="B39" ref-type="bibr">39</xref>), a set of web-based tools to display, was used to investigate and analyze cancer genomics data and its clinical information associated with VDR. The browser provides whole-genome to base-pair level views of several different types of genomic data, including next-generation sequencing platforms. Biological pathways, collections of genes, genomic or clinical information were used to sort, aggregate, and zoom into a group of samples. The current release (2013) displays an expanding set of data from various sources, including 201 datasets from 22 The Cancer Genome Atlas (TCGA) cancers as well as data from Cancer Cell Line Encyclopedia and Stand Up To Cancer (<xref rid="B39" ref-type="bibr">39</xref>).</p></sec><sec id="S2-2"><title>Database of somatic mutations</title><p>To collect data on TC related to VDR mutation, the web-software BioMart Central Portal and the Catalog of Somatic Mutations in Cancer (COSMIC) database (<xref rid="B40" ref-type="bibr">40</xref>) were used. BioMart offers a one-stop shop solution to access a wide array of biological databases, such as the major biomolecular sequence, pathway, and annotation databases such as Ensembl, Uniprot, Reactome, HGNC, Wormbase, and PRIDE (<xref rid="B41" ref-type="bibr">41</xref>). The Cancer BioMart web-interface with the following criteria was used: (1) Primary site&#x02009;=&#x02009;&#x0201c;thyroid&#x0201d;; (2) Mutation ID is not empty. The first criterion ensures that the mutation occurs in thyroid tissues, and the second criterion helps to exclude the samples without mutation in a specific gene. Thereby, we obtained the list of mutations in TC.</p><p>Catalog of Somatic Mutations in Cancer (<xref rid="B40" ref-type="bibr">40</xref>) stores and displays somatic mutation information and related details on human cancers. COSMIC was developed, and is currently maintained, at the Welcome Trust Sanger Institute. It is designed to gather, curate, and organize information on somatic mutations in cancer and to make it freely available on-line. It combines cancer mutation data, manually curate from the scientific literature, with the output from the Cancer Genome Project (CGP). Genes are selected for full literature curation using the Cancer Gene Census. The current release (v64) describes over 913,166 coding mutations of 24,394 genes from almost 847,698 tumor samples. All genes selected for the COSMIC database came from studies in the literature and are somatically mutated in human cancer (<xref rid="B42" ref-type="bibr">42</xref>). Based on this authority resource, a dataset of TC mutation was constructed.</p></sec><sec id="S2-3"><title>Data extraction</title><p>Information was carefully extracted from all eligible publications including clinical and experimental studies assessing any relation between vitamin D and non-medullary TC. A search for studies in the electronic databases Ovid Medline, Ovid Embase, Web of Science, AMED, and the Cochrane Library was run using an elaborated search strategy (Supplemental Material). In order not to miss any appropriate study, no time or language limits were applied for the search. Review articles were included only temporarily to provide a manual search tool.</p><p>The selection of studies involved an initial screening of the title and the abstract. In doubtful cases, the full text was obtained. Articles were entered in the data management software and the duplicates were eliminated (Endnote 6^&#x000ae;, Thomson Reuters Inc.). For clinical studies, detailed information about participants (number of patients, study location(s), and demographics variables), exposure (sun irradiation, dietary intake, and vitamin D serum level), comparison group, and outcome was assessed.</p><p>The search retrieved 471 references published until July 4th, 2013, 12 from the Cochrane Library, 176 from Ovid Medline, 188 from Ovid Embase and AMED, and 95 from Web of Sciences. Crosschecking the references of the reviews led to the inclusion of four supplementary articles (<xref rid="B43" ref-type="bibr">43</xref>&#x02013;<xref rid="B46" ref-type="bibr">46</xref>). No clinical trial was available. The flow chart of study selection is shown in Figure <xref ref-type="fig" rid="F1">1</xref>.</p><fig id="F1" position="float"><label>Figure 1</label><caption><p><bold>Flow chart of study selection for systematic review</bold>.</p></caption><graphic xlink:href="fonc-04-00309-g001"/></fig><p>Overall 30 articles were included, of which 17 were clinical studies (Table <xref ref-type="table" rid="T1">1</xref>) and 13 experimental studies (Table <xref ref-type="table" rid="T2">2</xref>). These studies were published in English language from 1987 to 2013. Of the 17 clinical studies, 8 (47.0%) showed protective effect of vitamin D (<xref rid="B44" ref-type="bibr">44</xref>, <xref rid="B45" ref-type="bibr">45</xref>, <xref rid="B47" ref-type="bibr">47</xref>&#x02013;<xref rid="B52" ref-type="bibr">52</xref>), 6 (35.3%) no significant relationship (<xref rid="B43" ref-type="bibr">43</xref>, <xref rid="B46" ref-type="bibr">46</xref>, <xref rid="B53" ref-type="bibr">53</xref>&#x02013;<xref rid="B57" ref-type="bibr">57</xref>), and 2 (11.7%) revealed an increased TC risk with high vitamin D intake (<xref rid="B58" ref-type="bibr">58</xref>, <xref rid="B59" ref-type="bibr">59</xref>). No comparison could be drawn from the remaining study (5.8%) (<xref rid="B60" ref-type="bibr">60</xref>). TC incidence was assessed in all of these studies, mortality in two (<xref rid="B45" ref-type="bibr">45</xref>, <xref rid="B47" ref-type="bibr">47</xref>); and one report assessed both (<xref rid="B45" ref-type="bibr">45</xref>). Except for three studies involving Arab populations (<xref rid="B51" ref-type="bibr">51</xref>, <xref rid="B56" ref-type="bibr">56</xref>, <xref rid="B60" ref-type="bibr">60</xref>), all studies included Europeans&#x02019; descendants and/or Hispanic whites.</p><table-wrap id="T1" position="float"><label>Table 1</label><caption><p><bold>Summary of clinical studies reporting an association between thyroid cancer and vitamin D</bold>.</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1">First author</th><th align="center" rowspan="1" colspan="1">Pub year</th><th align="left" rowspan="1" colspan="1">Country (state/province)<xref ref-type="table-fn" rid="tfn1">^a</xref></th><th align="left" rowspan="1" colspan="1">Cases/controls</th><th align="left" rowspan="1" colspan="1">Outcome</th><th align="left" rowspan="1" colspan="1">Exposure</th><th align="left" rowspan="1" colspan="1">Results<xref ref-type="table-fn" rid="tfn2">^b</xref></th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">Akslen (<xref rid="B44" ref-type="bibr">44</xref>)</td><td align="center" rowspan="1" colspan="1">1998</td><td align="left" rowspan="1" colspan="1">Norway</td><td align="left" rowspan="1" colspan="1">2627/NA</td><td align="left" rowspan="1" colspan="1">Incidence</td><td align="left" rowspan="1" colspan="1">Seasonal variation</td><td align="left" rowspan="1" colspan="1">pro</td></tr><tr><td align="left" rowspan="1" colspan="1">Boscoe (<xref rid="B45" ref-type="bibr">45</xref>)</td><td align="center" rowspan="1" colspan="1">2006</td><td align="left" rowspan="1" colspan="1">USA</td><td align="left" rowspan="1" colspan="1">&#x0003e;4,000/&#x0003e;4,000</td><td align="left" rowspan="1" colspan="1">Incidence and mortality</td><td align="left" rowspan="1" colspan="1">Latitude</td><td align="left" rowspan="1" colspan="1">pro<xref ref-type="table-fn" rid="tfn3">^c</xref></td></tr><tr><td align="left" rowspan="1" colspan="1">D&#x02019;avanzo (<xref rid="B53" ref-type="bibr">53</xref>)</td><td align="center" rowspan="1" colspan="1">1997</td><td align="left" rowspan="1" colspan="1">Italy</td><td align="left" rowspan="1" colspan="1">399/617</td><td align="left" rowspan="1" colspan="1">Incidence</td><td align="left" rowspan="1" colspan="1">Intake</td><td align="left" rowspan="1" colspan="1">NS</td></tr><tr><td align="left" rowspan="1" colspan="1">Glattre (<xref rid="B54" ref-type="bibr">54</xref>)</td><td align="center" rowspan="1" colspan="1">1993</td><td align="left" rowspan="1" colspan="1">Norway</td><td align="left" rowspan="1" colspan="1">92/460</td><td align="left" rowspan="1" colspan="1">Incidence</td><td align="left" rowspan="1" colspan="1">Intake</td><td align="left" rowspan="1" colspan="1">NS</td></tr><tr><td align="left" rowspan="1" colspan="1">Grant (<xref rid="B47" ref-type="bibr">47</xref>)</td><td align="center" rowspan="1" colspan="1">2006</td><td align="left" rowspan="1" colspan="1">Spain</td><td align="left" rowspan="1" colspan="1">NR</td><td align="left" rowspan="1" colspan="1">Mortality</td><td align="left" rowspan="1" colspan="1">Latitude</td><td align="left" rowspan="1" colspan="1">pro</td></tr><tr><td align="left" rowspan="1" colspan="1">Greenlee (<xref rid="B58" ref-type="bibr">58</xref>)</td><td align="center" rowspan="1" colspan="1">2004</td><td align="left" rowspan="1" colspan="1">USA (WA)</td><td align="left" rowspan="1" colspan="1">305/64,226</td><td align="left" rowspan="1" colspan="1">Incidence</td><td align="left" rowspan="1" colspan="1">Intake</td><td align="left" rowspan="1" colspan="1">con</td></tr><tr><td align="left" rowspan="1" colspan="1">Haghpanah (<xref rid="B56" ref-type="bibr">56</xref>)</td><td align="center" rowspan="1" colspan="1">2007</td><td align="left" rowspan="1" colspan="1">Iran</td><td align="left" rowspan="1" colspan="1">71/82</td><td align="left" rowspan="1" colspan="1">Incidence</td><td align="left" rowspan="1" colspan="1">VDR polymorphism</td><td align="left" rowspan="1" colspan="1">NS</td></tr><tr><td align="left" rowspan="1" colspan="1">Jonklass (<xref rid="B46" ref-type="bibr">46</xref>)</td><td align="center" rowspan="1" colspan="1">2013</td><td align="left" rowspan="1" colspan="1">USA (DC)</td><td align="left" rowspan="1" colspan="1">48/17</td><td align="left" rowspan="1" colspan="1">Incidence</td><td align="left" rowspan="1" colspan="1">Serum 25(OH)D</td><td align="left" rowspan="1" colspan="1">NS</td></tr><tr><td align="left" rowspan="1" colspan="1">Laney (<xref rid="B55" ref-type="bibr">55</xref>)</td><td align="center" rowspan="1" colspan="1">2010</td><td align="left" rowspan="1" colspan="1">USA (NE)</td><td align="left" rowspan="1" colspan="1">24/42</td><td align="left" rowspan="1" colspan="1">Incidence</td><td align="left" rowspan="1" colspan="1">Serum 25(OH)D</td><td align="left" rowspan="1" colspan="1">NS</td></tr><tr><td align="left" rowspan="1" colspan="1">Mack (<xref rid="B43" ref-type="bibr">43</xref>)</td><td align="center" rowspan="1" colspan="1">2002</td><td align="left" rowspan="1" colspan="1">USA (CA)</td><td align="left" rowspan="1" colspan="1">292/292</td><td align="left" rowspan="1" colspan="1">Incidence</td><td align="left" rowspan="1" colspan="1">Intake</td><td align="left" rowspan="1" colspan="1">NS</td></tr><tr><td align="left" rowspan="1" colspan="1">Penna-Martinez (<xref rid="B48" ref-type="bibr">48</xref>)</td><td align="center" rowspan="1" colspan="1">2009</td><td align="left" rowspan="1" colspan="1">Germany</td><td align="left" rowspan="1" colspan="1">147/57</td><td align="left" rowspan="1" colspan="1">Incidence</td><td align="left" rowspan="1" colspan="1">Serum 1,25(OH)₂ D VDR Polymorphism</td><td align="left" rowspan="1" colspan="1">pro</td></tr><tr><td align="left" rowspan="1" colspan="1">Penna-Martinez (<xref rid="B49" ref-type="bibr">49</xref>)</td><td align="center" rowspan="1" colspan="1">2012</td><td align="left" rowspan="1" colspan="1">Germany</td><td align="left" rowspan="1" colspan="1">253/302</td><td align="left" rowspan="1" colspan="1">Incidence</td><td align="left" rowspan="1" colspan="1">Serum 1,25(OH)₂ D</td><td align="left" rowspan="1" colspan="1">pro</td></tr><tr><td align="left" rowspan="1" colspan="1">Peterson (<xref rid="B60" ref-type="bibr">60</xref>)</td><td align="center" rowspan="1" colspan="1">2011</td><td align="left" rowspan="1" colspan="1">USA (MI)</td><td align="left" rowspan="1" colspan="1">30/70</td><td align="left" rowspan="1" colspan="1">NA</td><td align="left" rowspan="1" colspan="1">Sun exposure</td><td align="left" rowspan="1" colspan="1">NA</td></tr><tr><td align="left" rowspan="1" colspan="1">Ron (<xref rid="B59" ref-type="bibr">59</xref>)</td><td align="center" rowspan="1" colspan="1">1987</td><td align="left" rowspan="1" colspan="1">USA (CT)</td><td align="left" rowspan="1" colspan="1">159/285</td><td align="left" rowspan="1" colspan="1">Incidence</td><td align="left" rowspan="1" colspan="1">Intake</td><td align="left" rowspan="1" colspan="1">con</td></tr><tr><td align="left" rowspan="1" colspan="1">Roskies (<xref rid="B50" ref-type="bibr">50</xref>)</td><td align="center" rowspan="1" colspan="1">2012</td><td align="left" rowspan="1" colspan="1">Canada (QC)</td><td align="left" rowspan="1" colspan="1">12/88</td><td align="left" rowspan="1" colspan="1">Incidence</td><td align="left" rowspan="1" colspan="1">Serum 25(OH)D</td><td align="left" rowspan="1" colspan="1">pro</td></tr><tr><td align="left" rowspan="1" colspan="1">Sahin (<xref rid="B51" ref-type="bibr">51</xref>)</td><td align="center" rowspan="1" colspan="1">2013</td><td align="left" rowspan="1" colspan="1">Turkey</td><td align="left" rowspan="1" colspan="1">344/116</td><td align="left" rowspan="1" colspan="1">Incidence</td><td align="left" rowspan="1" colspan="1">Serum 25(OH)D</td><td align="left" rowspan="1" colspan="1">pro</td></tr><tr><td align="left" rowspan="1" colspan="1">Stepien (<xref rid="B52" ref-type="bibr">52</xref>)</td><td align="center" rowspan="1" colspan="1">2010</td><td align="left" rowspan="1" colspan="1">Poland</td><td align="left" rowspan="1" colspan="1">50/26</td><td align="left" rowspan="1" colspan="1">Incidence</td><td align="left" rowspan="1" colspan="1">Serum 1,25(OH)₂ D</td><td align="left" rowspan="1" colspan="1">pro</td></tr></tbody></table><table-wrap-foot><fn id="tfn1"><p>^aWA, Washington; DC, District of Columbia; NE, Nebraska; CA, California; MI, Michigan; CT, Connecticut; QC, Quebec;</p></fn><fn id="tfn2"><p>^bpro, protective effect of vitamin D (or surrogates); NS, not significant; con, vitamin D (or surrogates) increasing risk; NA, not applicable;</p></fn><fn id="tfn3"><p><italic>^cfor women only</italic>.</p></fn></table-wrap-foot></table-wrap><table-wrap id="T2" position="float"><label>Table 2</label><caption><p><bold>Experimental studies using cell lines or preclinical models to assess vitamin D effect on thyroid cancer</bold>.</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1">First author</th><th align="center" rowspan="1" colspan="1">Pub year</th><th align="left" rowspan="1" colspan="1">Samples<xref ref-type="table-fn" rid="tfn4">^a</xref></th><th align="left" rowspan="1" colspan="1">Main results</th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">Balla (<xref rid="B61" ref-type="bibr">61</xref>)</td><td align="center" rowspan="1" colspan="1">2011</td><td align="left" rowspan="1" colspan="1">6 PTC</td><td align="left" rowspan="1" colspan="1">Overexpression of CYP24A1 mRNA</td></tr><tr><td align="left" rowspan="1" colspan="1">Bennett (<xref rid="B62" ref-type="bibr">62</xref>)</td><td align="center" rowspan="1" colspan="1">2012</td><td align="left" rowspan="1" colspan="1">TPC1, C643</td><td align="left" rowspan="1" colspan="1">Antiproliferative effect of calcitriol</td></tr><tr><td align="left" rowspan="1" colspan="1">Clinckspoor (<xref rid="B63" ref-type="bibr">63</xref>)</td><td align="center" rowspan="1" colspan="1">2011</td><td align="left" rowspan="1" colspan="1">FTC133, C643, 8505c, Hth74</td><td align="left" rowspan="1" colspan="1">Antiproliferative effect of calcitriol and superagonistic analog CD578</td></tr><tr><td align="left" rowspan="1" colspan="1">Clinckspoor (<xref rid="B64" ref-type="bibr">64</xref>)</td><td align="center" rowspan="1" colspan="1">2012</td><td align="left" rowspan="1" colspan="1">64 thyroid cancers</td><td align="left" rowspan="1" colspan="1">VDR, CYP24A1, CYP27B1 overexpression</td></tr><tr><td align="left" rowspan="1" colspan="1">Dackiw (<xref rid="B65" ref-type="bibr">65</xref>)</td><td align="center" rowspan="1" colspan="1">2004</td><td align="left" rowspan="1" colspan="1">15 SCID mice/WRO</td><td align="left" rowspan="1" colspan="1">Growth inhibition of orthotopic tumor and p27^kip1 restoration after calcitriol treatment</td></tr><tr><td align="left" rowspan="1" colspan="1">Khadzkou (<xref rid="B66" ref-type="bibr">66</xref>)</td><td align="center" rowspan="1" colspan="1">2006</td><td align="left" rowspan="1" colspan="1">44 PTC</td><td align="left" rowspan="1" colspan="1">Overexpression of VDR and CYP27B1 (FFPE)</td></tr><tr><td align="left" rowspan="1" colspan="1">Liu (<xref rid="B67" ref-type="bibr">67</xref>)</td><td align="center" rowspan="1" colspan="1">2002</td><td align="left" rowspan="1" colspan="1">NPA, WRO</td><td align="left" rowspan="1" colspan="1">Antiproliferative effect of calcitriol and superagonistic analog EB1089, p27 restoration</td></tr><tr><td align="left" rowspan="1" colspan="1">Liu (<xref rid="B68" ref-type="bibr">68</xref>)</td><td align="center" rowspan="1" colspan="1">2005</td><td align="left" rowspan="1" colspan="1">WRO</td><td align="left" rowspan="1" colspan="1">Calcitriol and its analog EB1089 restore PTEN-dependent fibronectin expression</td></tr><tr><td align="left" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1">SCID mice/WRO</td><td align="left" rowspan="1" colspan="1">Growth inhibition in heterotopic model with calcitriol and EB1089</td></tr><tr><td align="left" rowspan="1" colspan="1">Liu (<xref rid="B69" ref-type="bibr">69</xref>)</td><td align="center" rowspan="1" colspan="1">2011</td><td align="left" rowspan="1" colspan="1">WRO, MRO</td><td align="left" rowspan="1" colspan="1">Calcitriol inhibits CEACAM1</td></tr><tr><td align="left" rowspan="1" colspan="1">Okano (<xref rid="B70" ref-type="bibr">70</xref>)</td><td align="center" rowspan="1" colspan="1">1999</td><td align="left" rowspan="1" colspan="1">Nude mice/NPA</td><td align="left" rowspan="1" colspan="1">Trend to growth inhibition in heterotopic model with calcitriol and less-calcemic analog</td></tr><tr><td align="left" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1">NPA</td><td align="left" rowspan="1" colspan="1">Dose-dependent inhibition of calcitriol and less-calcemic analog</td></tr><tr><td align="left" rowspan="1" colspan="1">Sharma (<xref rid="B71" ref-type="bibr">71</xref>)</td><td align="center" rowspan="1" colspan="1">2010</td><td align="left" rowspan="1" colspan="1">TPC1, C643, Hth7, Hth74, 8505c, SW1736</td><td align="left" rowspan="1" colspan="1">Response to calcitriol/DP006 depending on VDR polymorphism and 24-hydroxylase levels</td></tr><tr><td align="left" rowspan="1" colspan="1">Somjen (<xref rid="B72" ref-type="bibr">72</xref>)</td><td align="center" rowspan="1" colspan="1">2013</td><td align="left" rowspan="1" colspan="1">NPA, ARO, MRO</td><td align="left" rowspan="1" colspan="1">Overexpression of VDR and CYP27B1</td></tr><tr><td align="left" rowspan="1" colspan="1">Suzuki (<xref rid="B73" ref-type="bibr">73</xref>)</td><td align="center" rowspan="1" colspan="1">1999</td><td align="left" rowspan="1" colspan="1">TPC1-4, TAC1, TTA1</td><td align="left" rowspan="1" colspan="1">Dose-dependent growth inhibition of calcitriol and less-calcemic analog</td></tr></tbody></table><table-wrap-foot><fn id="tfn4"><p><italic>^aCell line-corresponding histologic subtype: TPC1-4-PTC, KTC-PTC, BCPAP-PTC, NPA-PTC, KAT5-PTC, FTC133-FTC, FRO-FTC, MRO-FTC, WRO-FTC, C643-ATC, Hth7-ATC, Hth74-ATC, 8505c-ATC, SW1736-ATC, TAC-1-ATC, TTA-1-ATC. PTC, papillary thyroid cancer; FTC, follicular thyroid cancer; ATC, anaplastic thyroid cancer; SCID, severe combined immunodeficient</italic>.</p></fn></table-wrap-foot></table-wrap></sec></sec><sec id="S3"><title>Results and Discussion</title><sec id="S3-4"><title>Determinants of vitamin D levels and impact in TC</title><p>Solar UVB irradiation is the primary source of vitamin D and can be estimated by latitude of the living area. In TC, large epidemiological studies support an inverse relation between TC incidence and latitude (<xref rid="B45" ref-type="bibr">45</xref>, <xref rid="B47" ref-type="bibr">47</xref>) (Table <xref ref-type="table" rid="T1">1</xref>). These studies performed a multivariate analysis to adjust for confounding factors. However, vitamin D levels were not measured. Consequently, it is unclear if the multivariate analysis resulted in accurate vitamin D estimates. Indeed, vitamin D deficiency is highly prevalent among latitudes that benefit from high solar irradiation such as Africa, the Middle East, and Southern Asia. This may be due to skin pigmentation, traditional clothing, and sun avoidance seen in southern heat-exposed populations (<xref rid="B60" ref-type="bibr">60</xref>, <xref rid="B74" ref-type="bibr">74</xref>). In contrast, fair-skinned northern populations usually seek sun exposure and may also benefit from high intake of vitamin D rich diet such as fatty fish and cod-liver oil (<xref rid="B74" ref-type="bibr">74</xref>). Further, a mutation in the cutaneous structural protein filaggrin, which occurs in up to 10% of Europeans was shown to lead to higher circulating vitamin D levels (<xref rid="B75" ref-type="bibr">75</xref>). Nonetheless, North American and European studies have shown seasonal variations of vitamin D levels due to insufficient sun irradiation during winter (<xref rid="B76" ref-type="bibr">76</xref>). In TC, one study from Norway reported higher proliferation values for tumors resected during winter compared to other seasons (<xref rid="B44" ref-type="bibr">44</xref>). These results comply with above-mentioned studies showing an inverse relation between TC incidence and latitude (<xref rid="B45" ref-type="bibr">45</xref>, <xref rid="B47" ref-type="bibr">47</xref>). For studies estimating vitamin D consumption and TC risk, however, no convincing associations have been shown (Table <xref ref-type="table" rid="T1">1</xref>) (<xref rid="B43" ref-type="bibr">43</xref>, <xref rid="B53" ref-type="bibr">53</xref>, <xref rid="B54" ref-type="bibr">54</xref>, <xref rid="B58" ref-type="bibr">58</xref>, <xref rid="B59" ref-type="bibr">59</xref>). This may be due to the general poor correlation between vitamin D deficiency and estimates of vitamin D consumption (<xref rid="B57" ref-type="bibr">57</xref>).</p><p>A more accurate way to assess vitamin D is biological monitoring. Association studies investigating the relationship between levels of serum vitamin D and TC risk mostly point toward a protective effect of vitamin D (<xref rid="B48" ref-type="bibr">48</xref>&#x02013;<xref rid="B52" ref-type="bibr">52</xref>, <xref rid="B55" ref-type="bibr">55</xref>, <xref rid="B77" ref-type="bibr">77</xref>) (Table <xref ref-type="table" rid="T1">1</xref>). Pooling the data among these studies is not possible due to different cut-off levels for different vitamin D derivatives and control groups used in each of these studies. This would greatly limit the validity of a meta-analysis. The lack of consensus in cut-off levels may reflect the fact that those are differently defined depending on targeted clinical endpoints (<xref rid="B78" ref-type="bibr">78</xref>, <xref rid="B79" ref-type="bibr">79</xref>). Classical vitamin D targets, i.e., those implicated in calcium and bone homeostasis, do not allow conclusions on optimal level of vitamin D having anticancer properties. While doses up to 4,000 IU of daily vitamin D supplementation have been considered safe, studies have reported hypercalcemia, nephrolithiasis, vascular, and soft tissue calcification with high doses of vitamin D and also U-shaped relationship between vitamin D levels above 75&#x02009;nmol/l and certain cancer subtypes (<xref rid="B80" ref-type="bibr">80</xref>, <xref rid="B81" ref-type="bibr">81</xref>). One additional issue of most of these association studies is that vitamin D levels were measured only once, which does not permit distinction between outcome and exposure. Indeed, some studies have reported low serum vitamin D as a result of malignancy (<xref rid="B82" ref-type="bibr">82</xref>).</p><p>Above-mentioned skin types, alimentary, and social habits yet do not fully explain vitamin D variability among populations (<xref rid="B83" ref-type="bibr">83</xref>). One major determinant of individual susceptibility to vitamin D is the activity of vitamin D metabolizing enzymes. Three major cytochrome P-450 (CYP) hydroxylases are responsible for vitamin D activation through 25- followed by 1&#x003b1;-hydroxylation of the molecule, and deactivation through 24-hydroxylation. Multiple enzymes have been reported as vitamin D 25-hydroxylases, a step occurring constitutively and primarily in the liver. Unlike 25-hydroxylation, 1&#x003b1;-hydroxylation of 25(OH)D₃ by the CYP27B1 is a tightly regulated and rate-limiting step. It is regulated by calcium, 1&#x003b1;,25(OH)₂D₃ itself, PTH, calcitonin, and phosphate levels. Recently, fibroblast growth factor 23 (FGF23) was identified as a novel antagonist of PTH and is thought to play an important role in vitamin D regulation pathway (<xref rid="B84" ref-type="bibr">84</xref>). Although CYP27B1 and CYP24A1 are primarily expressed in the kidney, recent studies showed that they are expressed in many other tissues, including the thyroid (<xref rid="B61" ref-type="bibr">61</xref>, <xref rid="B62" ref-type="bibr">62</xref>). In TC, there is evidence that polymorphisms leading to impaired CYP27B1 function and/or increased CYP24A1 activity are associated with increased TC risk (<xref rid="B49" ref-type="bibr">49</xref>). Transcriptional profiling studies show that both enzymes are overexpressed in early TC (<xref rid="B61" ref-type="bibr">61</xref>), but their expression tends to decrease along with tumor progression (<xref rid="B64" ref-type="bibr">64</xref>, <xref rid="B66" ref-type="bibr">66</xref>).</p></sec><sec id="S3-5"><title>Determinants of predicted response to vitamin D</title><p>The action of vitamin D mainly occurs through binding to the VDR (<xref rid="B23" ref-type="bibr">23</xref>), whose levels are subject to genetic variations. Using the UCSC genomic database, we analyzed 552 thyroid samples that underwent genomic profiling using RNA Seq. The expression of VDR was down regulated in benign thyroid samples and up regulated in most TC cases (Figure <xref ref-type="fig" rid="F2">2</xref>). These results are confirmed by a few <italic>in vitro</italic> studies using TC cell lines (<xref rid="B72" ref-type="bibr">72</xref>) and independent clinical samples (<xref rid="B64" ref-type="bibr">64</xref>, <xref rid="B66" ref-type="bibr">66</xref>). However, VDR levels alone may translate poorly with response to vitamin D stimulation if polymorphisms of VDR are not taken into account (<xref rid="B71" ref-type="bibr">71</xref>, <xref rid="B85" ref-type="bibr">85</xref>, <xref rid="B86" ref-type="bibr">86</xref>). The analysis of the genomic organization of the VDR <italic>locus</italic> at chromosome 12q13.1 revealed the large <italic>VDR</italic> gene (about 100&#x02009;Kb) with an extensive promoter region capable of generating multiple tissue-specific transcripts (<xref rid="B87" ref-type="bibr">87</xref>). In view of the observed genome-wide frequency of single nucleotide polymorphisms (<xref rid="B88" ref-type="bibr">88</xref>), one can predict &#x0003e;100 functional polymorphisms to be present in the VDR region alone, including the promoter region (Figure <xref ref-type="fig" rid="F3">3</xref>). Point mutations in the VDR gene have been identified in various regions, including the VDR DNA binding domain (DBD) and the ligand-binding domain (LBD) (<xref rid="B89" ref-type="bibr">89</xref>). Such mutations can disrupt ligand-binding affinity to the receptor (<xref rid="B90" ref-type="bibr">90</xref>), heterodimerization of VDR with RXR (<xref rid="B91" ref-type="bibr">91</xref>), or interactions of the VDR receptor with partners such as coactivators (<xref rid="B92" ref-type="bibr">92</xref>). Other mutations such as in the initiation codon can create a premature termination (<xref rid="B93" ref-type="bibr">93</xref>) or alternative translation start sites to result in alternative splicing and formation of truncated proteins (<xref rid="B94" ref-type="bibr">94</xref>, <xref rid="B95" ref-type="bibr">95</xref>). The analysis of the COSMIC database showed a high proportion of missense mutations that were re-identified (67.44%), while complex mutations were not detected (Table <xref ref-type="table" rid="T3">3</xref>). The distribution of the mutations observed in the VDR gene in TC is shown in Figure <xref ref-type="fig" rid="F4">4</xref>. Only two studies investigated the association between VDR polymorphisms and TC risk, one showed an increased TC risk for patients with particular VDR polymorphism (<xref rid="B48" ref-type="bibr">48</xref>), while another could not point out any significant difference (<xref rid="B56" ref-type="bibr">56</xref>).</p><fig id="F2" position="float"><label>Figure 2</label><caption><p><bold>Gene expression profile showing VDR signature for 552 thyroid cancer cases (RNA Seq)</bold>. Each row corresponds to sample from a single case. Columns from the left correspond to genomic heatmap according to chromosomal location. The last two columns represent VDR expression profile (represented by red for overexpression and green for downregulation) in normal (pink) versus cancer (red) tissues. VDR is mostly overexpressed in malignant samples but almost absent in benign tissues. Source: UC Santa Cruz &#x02013; Cancer Genomics Browser.</p></caption><graphic xlink:href="fonc-04-00309-g002"/></fig><fig id="F3" position="float"><label>Figure 3</label><caption><p><bold>Schematic diagram of VDR gene showing different restriction site on chromosome 12</bold>.</p></caption><graphic xlink:href="fonc-04-00309-g003"/></fig><table-wrap id="T3" position="float"><label>Table 3</label><caption><p><bold>Mutations identified in VDR</bold>.</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1">Position (AA)</th><th align="center" rowspan="1" colspan="1">Mutation&#x02009;&#x02009;(CDS)</th><th align="center" rowspan="1" colspan="1">Mutation (amino acid)</th><th align="center" rowspan="1" colspan="1">Mutation &#x02003;type</th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">8</td><td align="center" rowspan="1" colspan="1">c.23C&#x02009;&#x0003e;&#x02009;T</td><td align="center" rowspan="1" colspan="1">p.T8I</td><td align="center" rowspan="1" colspan="1">Substitution &#x02013; missense</td></tr><tr><td align="left" rowspan="1" colspan="1">33</td><td align="center" rowspan="1" colspan="1">c.98G&#x02009;&#x0003e;&#x02009;A</td><td align="center" rowspan="1" colspan="1">p.G33D</td><td align="center" rowspan="1" colspan="1">Substitution &#x02013; missense</td></tr><tr><td align="left" rowspan="1" colspan="1">52</td><td align="center" rowspan="1" colspan="1">c.156G&#x02009;&#x0003e;&#x02009;A</td><td align="center" rowspan="1" colspan="1">p.M52I</td><td align="center" rowspan="1" colspan="1">Substitution &#x02013; missense</td></tr><tr><td align="left" rowspan="1" colspan="1">74</td><td align="center" rowspan="1" colspan="1">c.221G&#x02009;&#x0003e;&#x02009;A</td><td align="center" rowspan="1" colspan="1">p.R74H</td><td align="center" rowspan="1" colspan="1">Substitution &#x02013; missense</td></tr><tr><td align="left" rowspan="1" colspan="1">78</td><td align="center" rowspan="1" colspan="1">c.233C&#x02009;&#x0003e;&#x02009;G</td><td align="center" rowspan="1" colspan="1">p.A78G</td><td align="center" rowspan="1" colspan="1">Substitution &#x02013; missense</td></tr><tr><td align="left" rowspan="1" colspan="1">130</td><td align="center" rowspan="1" colspan="1">c.389G&#x02009;&#x0003e;&#x02009;A</td><td align="center" rowspan="1" colspan="1">p.R130H</td><td align="center" rowspan="1" colspan="1">Substitution &#x02013; missense</td></tr><tr><td align="left" rowspan="1" colspan="1">146</td><td align="center" rowspan="1" colspan="1">c.438C&#x02009;&#x0003e;&#x02009;G</td><td align="center" rowspan="1" colspan="1">p.T146T</td><td align="center" rowspan="1" colspan="1">Substitution &#x02013; coding silent</td></tr><tr><td align="left" rowspan="1" colspan="1">149</td><td align="center" rowspan="1" colspan="1">c.445G&#x02009;&#x0003e;&#x02009;T</td><td align="center" rowspan="1" colspan="1">p.D149Y</td><td align="center" rowspan="1" colspan="1">Substitution &#x02013; missense</td></tr><tr><td align="left" rowspan="1" colspan="1">154</td><td align="center" rowspan="1" colspan="1">c.460C&#x02009;&#x0003e;&#x02009;T</td><td align="center" rowspan="1" colspan="1">p.R154W</td><td align="center" rowspan="1" colspan="1">Substitution &#x02013; missense</td></tr><tr><td align="left" rowspan="1" colspan="1">158</td><td align="center" rowspan="1" colspan="1">c.472C&#x02009;&#x0003e;&#x02009;T</td><td align="center" rowspan="1" colspan="1">p.R158C</td><td align="center" rowspan="1" colspan="1">Substitution &#x02013; missense</td></tr><tr><td align="left" rowspan="1" colspan="1">159</td><td align="center" rowspan="1" colspan="1">c.477G&#x02009;&#x0003e;&#x02009;C</td><td align="center" rowspan="1" colspan="1">p.V159V</td><td align="center" rowspan="1" colspan="1">Substitution &#x02013; coding silent</td></tr><tr><td align="left" rowspan="1" colspan="1">161</td><td align="center" rowspan="1" colspan="1">c.481G&#x02009;&#x0003e;&#x02009;A</td><td align="center" rowspan="1" colspan="1">p.D161N</td><td align="center" rowspan="1" colspan="1">Substitution &#x02013; missense</td></tr><tr><td align="left" rowspan="1" colspan="1">162</td><td align="center" rowspan="1" colspan="1">c.484G&#x02009;&#x0003e;&#x02009;T</td><td align="center" rowspan="1" colspan="1">p.G162C</td><td align="center" rowspan="1" colspan="1">Substitution &#x02013; missense</td></tr><tr><td align="left" rowspan="1" colspan="1">169</td><td align="center" rowspan="1" colspan="1">c.507G&#x02009;&#x0003e;&#x02009;A</td><td align="center" rowspan="1" colspan="1">p.R169R</td><td align="center" rowspan="1" colspan="1">Substitution &#x02013; coding silent</td></tr><tr><td align="left" rowspan="1" colspan="1">181</td><td align="center" rowspan="1" colspan="1">c.541G&#x02009;&#x0003e;&#x02009;T</td><td align="center" rowspan="1" colspan="1">p.D181Y</td><td align="center" rowspan="1" colspan="1">Substitution &#x02013; missense</td></tr><tr><td align="left" rowspan="1" colspan="1">191</td><td align="center" rowspan="1" colspan="1">c.573C&#x02009;&#x0003e;&#x02009;A</td><td align="center" rowspan="1" colspan="1">p.I191I</td><td align="center" rowspan="1" colspan="1">Substitution &#x02013; coding silent</td></tr><tr><td align="left" rowspan="1" colspan="1">199</td><td align="center" rowspan="1" colspan="1">c.597G&#x02009;&#x0003e;&#x02009;A</td><td align="center" rowspan="1" colspan="1">p.S199S</td><td align="center" rowspan="1" colspan="1">Substitution &#x02013; coding silent</td></tr><tr><td align="left" rowspan="1" colspan="1">208</td><td align="center" rowspan="1" colspan="1">c.623G&#x02009;&#x0003e;&#x02009;T</td><td align="center" rowspan="1" colspan="1">p.S208I</td><td align="center" rowspan="1" colspan="1">Substitution &#x02013; missense</td></tr><tr><td align="left" rowspan="1" colspan="1">236</td><td align="center" rowspan="1" colspan="1">c.708C&#x02009;&#x0003e;&#x02009;A</td><td align="center" rowspan="1" colspan="1">p.Y236<xref ref-type="table-fn" rid="tfn5">^a</xref></td><td align="center" rowspan="1" colspan="1">Substitution &#x02013; nonsense</td></tr><tr><td align="left" rowspan="1" colspan="1">253</td><td align="center" rowspan="1" colspan="1">c.757G&#x02009;&#x0003e;&#x02009;T</td><td align="center" rowspan="1" colspan="1">p.D253Y</td><td align="center" rowspan="1" colspan="1">Substitution &#x02013; missense</td></tr><tr><td align="left" rowspan="1" colspan="1">274</td><td align="center" rowspan="1" colspan="1">c.820C&#x02009;&#x0003e;&#x02009;T</td><td align="center" rowspan="1" colspan="1">p.R274C</td><td align="center" rowspan="1" colspan="1">Substitution &#x02013; missense</td></tr><tr><td align="left" rowspan="1" colspan="1">296</td><td align="center" rowspan="1" colspan="1">c.887G&#x02009;&#x0003e;&#x02009;A</td><td align="center" rowspan="1" colspan="1">p.R296H</td><td align="center" rowspan="1" colspan="1">Substitution &#x02013; missense</td></tr><tr><td align="left" rowspan="1" colspan="1">320</td><td align="center" rowspan="1" colspan="1">c.960G&#x02009;&#x0003e;&#x02009;A</td><td align="center" rowspan="1" colspan="1">p.L320L</td><td align="center" rowspan="1" colspan="1">Substitution &#x02013; coding silent</td></tr><tr><td align="left" rowspan="1" colspan="1">339</td><td align="center" rowspan="1" colspan="1">c.1015G&#x02009;&#x0003e;&#x02009;A</td><td align="center" rowspan="1" colspan="1">p.V339I</td><td align="center" rowspan="1" colspan="1">Substitution &#x02013; missense</td></tr><tr><td align="left" rowspan="1" colspan="1">350</td><td align="center" rowspan="1" colspan="1">c.1049C&#x02009;&#x0003e;&#x02009;T</td><td align="center" rowspan="1" colspan="1">p.A350V</td><td align="center" rowspan="1" colspan="1">Substitution &#x02013; missense</td></tr><tr><td align="left" rowspan="1" colspan="1">350</td><td align="center" rowspan="1" colspan="1">c.1050G&#x02009;&#x0003e;&#x02009;A</td><td align="center" rowspan="1" colspan="1">p.A350A</td><td align="center" rowspan="1" colspan="1">Substitution &#x02013; coding silent</td></tr><tr><td align="left" rowspan="1" colspan="1">352</td><td align="center" rowspan="1" colspan="1">c.1056T&#x02009;&#x0003e;&#x02009;C</td><td align="center" rowspan="1" colspan="1">p.I352I</td><td align="center" rowspan="1" colspan="1">Substitution &#x02013; coding silent</td></tr><tr><td align="left" rowspan="1" colspan="1">353</td><td align="center" rowspan="1" colspan="1">c.1058A&#x02009;&#x0003e;&#x02009;T</td><td align="center" rowspan="1" colspan="1">p.E353V</td><td align="center" rowspan="1" colspan="1">Substitution &#x02013; missense</td></tr><tr><td align="left" rowspan="1" colspan="1">358</td><td align="center" rowspan="1" colspan="1">c.1072C&#x02009;&#x0003e;&#x02009;T</td><td align="center" rowspan="1" colspan="1">p.R358C</td><td align="center" rowspan="1" colspan="1">Substitution &#x02013; missense</td></tr><tr><td align="left" rowspan="1" colspan="1">365</td><td align="center" rowspan="1" colspan="1">c.1094C&#x02009;&#x0003e;&#x02009;T</td><td align="center" rowspan="1" colspan="1">p.T365M</td><td align="center" rowspan="1" colspan="1">Substitution &#x02013; missense</td></tr><tr><td align="left" rowspan="1" colspan="1">368</td><td align="center" rowspan="1" colspan="1">c.1103G&#x02009;&#x0003e;&#x02009;A</td><td align="center" rowspan="1" colspan="1">p.R368H</td><td align="center" rowspan="1" colspan="1">Substitution &#x02013; missense</td></tr><tr><td align="left" rowspan="1" colspan="1">379</td><td align="center" rowspan="1" colspan="1">c.1135C&#x02009;&#x0003e;&#x02009;T</td><td align="center" rowspan="1" colspan="1">p.L379F</td><td align="center" rowspan="1" colspan="1">Substitution &#x02013; missense</td></tr><tr><td align="left" rowspan="1" colspan="1">399</td><td align="center" rowspan="1" colspan="1">c.1196A&#x02009;&#x0003e;&#x02009;T</td><td align="center" rowspan="1" colspan="1">p.K399M</td><td align="center" rowspan="1" colspan="1">Substitution &#x02013; missense</td></tr><tr><td align="left" rowspan="1" colspan="1">402</td><td align="center" rowspan="1" colspan="1">c.1205G&#x02009;&#x0003e;&#x02009;C</td><td align="center" rowspan="1" colspan="1">p.R402P</td><td align="center" rowspan="1" colspan="1">Substitution &#x02013; missense</td></tr><tr><td align="left" rowspan="1" colspan="1">418</td><td align="center" rowspan="1" colspan="1">c.1254G&#x02009;&#x0003e;&#x02009;T</td><td align="center" rowspan="1" colspan="1">p.V418V</td><td align="center" rowspan="1" colspan="1">Substitution &#x02013; coding silent</td></tr><tr><td align="left" rowspan="1" colspan="1">420</td><td align="center" rowspan="1" colspan="1">c.1258G&#x02009;&#x0003e;&#x02009;A</td><td align="center" rowspan="1" colspan="1">p.E420K</td><td align="center" rowspan="1" colspan="1">Substitution &#x02013; missense</td></tr></tbody></table><table-wrap-foot><fn id="tfn5"><p><italic>^aNonsense mutation resulting in stop codon</italic>.</p></fn></table-wrap-foot></table-wrap><fig id="F4" position="float"><label>Figure 4</label><caption><p><bold>Pie-chart showing the percentage of the mutation type in VDR in thyroid cancer according to COSMIC database</bold>.</p></caption><graphic xlink:href="fonc-04-00309-g004"/></fig></sec><sec id="S3-6"><title>Downstream impact of VDR activation</title><p>Upon activation by vitamin D, VDR binds as a heterodimer with retinoid-X receptors to specific VDREs (<xref rid="B84" ref-type="bibr">84</xref>). VDREs usually bear a consensus sequence known as DR3 element located in the promoter region of the target genes. In addition to this classic mechanism, recent chromatin-immunoprecipitation (ChIP-seq) studies allowed to gain genome-wide insights of the binding sites of VDR (<xref rid="B96" ref-type="bibr">96</xref>). These studies showed that the ligand-bound heterodimer can bind to ~2000&#x02013;8000 sites in the genome. Interestingly, the majority of the binding sites do not bear the classical DR3-type sequence (<xref rid="B84" ref-type="bibr">84</xref>). A significant enrichment was seen in regions associated with active chromatin and histone modifications thus supporting a broad genetic and epigenetic regulatory role of vitamin D. Further enrichment of VDR binding was also found in proximity of genes involved in autoimmune diseases (e.g., multiple sclerosis, type-I diabetes, and Crohn&#x02019;s disease) and colorectal or breast cancer (<xref rid="B97" ref-type="bibr">97</xref>). For TC, only data relying on classical <italic>in vitro</italic> experiments is available.</p><p>In agreement with experimental studies in other cancer types, exposure of a variety of TC cells to vitamin D <italic>in vitro</italic> leads to antiproliferative and pro-differentiation properties (<xref rid="B62" ref-type="bibr">62</xref>, <xref rid="B63" ref-type="bibr">63</xref>, <xref rid="B67" ref-type="bibr">67</xref>, <xref rid="B70" ref-type="bibr">70</xref>, <xref rid="B71" ref-type="bibr">71</xref>, <xref rid="B73" ref-type="bibr">73</xref>) (Table <xref ref-type="table" rid="T2">2</xref>). These results have been confirmed by <italic>in vivo</italic> studies (<xref rid="B65" ref-type="bibr">65</xref>, <xref rid="B68" ref-type="bibr">68</xref>). Most studies are testing vitamin D itself and synthetic vitamin D analogs, as patient&#x02019;s exposure to pharmacologically high doses of vitamin D can be limited by the side-effects, mainly hypercalcemia (<xref rid="B63" ref-type="bibr">63</xref>, <xref rid="B67" ref-type="bibr">67</xref>, <xref rid="B70" ref-type="bibr">70</xref>, <xref rid="B71" ref-type="bibr">71</xref>, <xref rid="B73" ref-type="bibr">73</xref>).</p><p>Mechanistically, vitamin D was shown to inhibit proliferation through c-mac mRNA inhibition, which is a well-known proto-oncogene (<xref rid="B70" ref-type="bibr">70</xref>). Further, it can induce a growth arrest effect in part through stimulating accumulation of the cyclin-dependent kinase inhibitor p27^kip1 in the nucleus (<xref rid="B67" ref-type="bibr">67</xref>). Treatment with vitamin D is thought to prevent p27^kip1 phosphorylation, which was shown to increase its ubiquitin-dependent proteasome degradation (<xref rid="B67" ref-type="bibr">67</xref>). Further, vitamin D was shown to enhance cell&#x02013;cell adhesion through PTEN-dependent fibronectin upregulation (<xref rid="B68" ref-type="bibr">68</xref>). Those results could be confirmed <italic>in vivo</italic>. Interestingly, the antiproliferative effect of vitamin D was abolished when knocking down fibronectin (<xref rid="B68" ref-type="bibr">68</xref>) and was shown to be independent of CEACAM1 expression, a tumor-suppressive adhesion molecule (<xref rid="B69" ref-type="bibr">69</xref>).</p></sec></sec><sec id="S4"><title>Conclusion and Perspectives</title><p>The pleiotropic roles of vitamin D in cancer have been recognized through seminal preclinical studies although the preventive and therapeutic potential of vitamin D or its analogs remain debated due in part to the complex mode of action of this vitamin. Recent progress in high-throughput technologies to interrogate human genomic and epigenomic events has provided additional levels of regulatory loops and individual genetic variations that can impact on individual susceptibility to vitamin D. This knowledge opens up new tools to address confounding factors that contribute to discrepant results seen in previous association studies, in particular in relation to cancer prevention. As well, this knowledge impels an exciting avenue in the discovery of novel vitamin D analogs with enhanced preventive or therapeutic efficiency and limited side-effects.</p></sec><sec id="S5"><title>Author Contributions</title><p>Gregoire B. Morand performed the literature search, the retrieval of the studies, the data extraction, and wrote the main part of the manuscript under Sabrina Daniela da Silva and Moulay A. Alaoui-Jamali&#x02019;s supervision. All the authors participated substantially to the final manuscript and approved the final version.</p></sec><sec id="S6"><title>Conflict of Interest Statement</title><p>The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.</p></sec><sec sec-type="supplementary-material" id="S7"><title>Supplementary Material</title><p>The Supplementary Material for this article can be found online at <uri xlink:type="simple" xlink:href="http://www.frontiersin.org/Journal/10.3389/fonc.2014.00309/abstract">http://www.frontiersin.org/Journal/10.3389/fonc.2014.00309/abstract</uri></p><supplementary-material content-type="local-data" id="SM1"><media xlink:href="Data_Sheet_1.DOCX"><caption><p>Click here for additional data file.</p></caption></media></supplementary-material></sec>

The obesity paradox in acute coronary syndrome: a meta-analysis

Could not extract abstract

<contrib contrib-type="author" corresp="yes"><name><surname>Niedziela</surname><given-names>Jacek</given-names></name><address><phone>+48 32 3733619</phone><email>jacek@niedziela.org</email></address><xref ref-type="aff" rid="Aff1">1</xref></contrib><contrib contrib-type="author"><name><surname>Hudzik</surname><given-names>Bartosz</given-names></name><xref ref-type="aff" rid="Aff1">1</xref></contrib><contrib contrib-type="author"><name><surname>Niedziela</surname><given-names>Natalia</given-names></name><xref ref-type="aff" rid="Aff2">2</xref></contrib><contrib contrib-type="author"><name><surname>G&#x00105;sior</surname><given-names>Mariusz</given-names></name><xref ref-type="aff" rid="Aff1">1</xref></contrib><contrib contrib-type="author"><name><surname>Gierlotka</surname><given-names>Marek</given-names></name><xref ref-type="aff" rid="Aff1">1</xref></contrib><contrib contrib-type="author"><name><surname>Wasilewski</surname><given-names>Jaros&#x00142;aw</given-names></name><xref ref-type="aff" rid="Aff1">1</xref></contrib><contrib contrib-type="author"><name><surname>Myrda</surname><given-names>Krzysztof</given-names></name><xref ref-type="aff" rid="Aff1">1</xref></contrib><contrib contrib-type="author"><name><surname>Lekston</surname><given-names>Andrzej</given-names></name><xref ref-type="aff" rid="Aff1">1</xref></contrib><contrib contrib-type="author"><name><surname>Polo&#x00144;ski</surname><given-names>Lech</given-names></name><xref ref-type="aff" rid="Aff1">1</xref></contrib><contrib contrib-type="author"><name><surname>Rozentryt</surname><given-names>Piotr</given-names></name><xref ref-type="aff" rid="Aff1">1</xref></contrib><aff id="Aff1"><label>1</label>Third Department of Cardiology, Silesian Center for Heart Diseases, Medical University of Silesia, M. Curie-Sk&#x00142;odowskiej 9, 41-800 Zabrze, Poland </aff><aff id="Aff2"><label>2</label>Department of Neurology, Medical University of Silesia, Zabrze, Poland </aff>

European Journal of Epidemiology

<sec id="Sec1"><title>Background</title><p>The concept of obesity (from the Latin word obdere&#x02014;to eat all over: ob&#x02014;over, above; edere&#x02014;to eat) for the first time was used in the Oxford Dictionary in 1611, as a synonym for words: corpulent, thick [<xref ref-type="bibr" rid="CR1">1</xref>]. The oldest trace of obesity is believed to be a female Willendorf statuette, dated about 22,000&#x02013;24,000&#x000a0;years B.C. [<xref ref-type="bibr" rid="CR2">2</xref>].</p><p>The attitude toward obesity has been changing over centuries and cultures. In ancient Greece (Hippocrates) and India (Sushruta), it was considered as a pathology [<xref ref-type="bibr" rid="CR3">3</xref>]. In the Europe and the Far East, in the Middle Ages and the Renaissance, obesity was attractive and desirable. A corpulent silhouette was identified with wealth. In the twentieth and twenty-first century, obesity again became unpopular and unfashionable. Being slim has been considered as optimal weight status both for aesthetic and health reasons.</p><p>There are many parameters describing body weight status. Years of observation revealed that body mass and height were in certain proportions. Epidemiological significance of the same body weight is completely different in tall and short person. The most popular formula describing weight in relation to height is the Quetelet index, also known as Body Mass Index (BMI) [<xref ref-type="bibr" rid="CR4">4</xref>]. BMI is expressed as the ratio of body weight in kilograms and the square of the height in meters. Based on epidemiological observations linking various aspects of health status with BMI, the World Health Organization (WHO) has established a normal BMI for European and North American populations in the range of 18.5&#x02013;24.9&#x000a0;kg/m² [<xref ref-type="bibr" rid="CR5">5</xref>]. A BMI range of 25&#x02013;29.9&#x000a0;kg/m² defines overweight and a BMI of 30&#x000a0;kg/m² and more is regarded as obesity. BMI below 18.5&#x000a0;kg/m² indicates underweight.</p><p>In some populations, the BMI cut-off values for a diagnosis of obesity are different. For example, in the Japanese, South Korean and Chinese populations obesity is recognized for BMIs above 25&#x000a0;kg/m² [<xref ref-type="bibr" rid="CR6">6</xref>], 27.5&#x000a0;kg/m² [<xref ref-type="bibr" rid="CR7">7</xref>] and 28&#x000a0;kg/m² [<xref ref-type="bibr" rid="CR8">8</xref>], respectively.</p><p>BMI can be calculated easily and quickly and thus it is widely used both in research and clinical areas. It is also applied for body weight classification by WHO. It should be noted that BMI is not the only and probably not the most accurate measure of the cardiovascular risk associated with body weight.</p><p>The obesity, described as higher BMI, is considered as the risk factor for mortality in the general population. The lowest mortality is observed for the BMI range of 20&#x02013;24.9&#x000a0;kg/m² (for non-smokers in the American and European populations) and it increases below and above this range [<xref ref-type="bibr" rid="CR5">5</xref>, <xref ref-type="bibr" rid="CR9">9</xref>]. During the last two decades, reports on the favorable prognosis in chronically ill patients with overweight or obesity have been published. This phenomenon commonly called the obesity paradox or reversed epidemiology was recognized in patients with chronic kidney disease [<xref ref-type="bibr" rid="CR10">10</xref>], chronic heart failure [<xref ref-type="bibr" rid="CR11">11</xref>] and chronic obstructive pulmonary disease [<xref ref-type="bibr" rid="CR12">12</xref>]. Recently, a similar paradox linking higher BMI with better prognosis was described in coronary artery disease [<xref ref-type="bibr" rid="CR13">13</xref>, <xref ref-type="bibr" rid="CR14">14</xref>]. Due to acute metabolic imbalance during AMI and increased catabolism following AMI [<xref ref-type="bibr" rid="CR15">15</xref>], the occurrence of obesity paradox after AMI could be different than in stable CAD.</p></sec><sec id="Sec2"><title>Objectives</title><p>Our aim was to analyze the relationship between BMI and total mortality in patients after acute coronary syndrome (ACS).</p></sec><sec id="Sec3"><title>Methods</title><sec id="Sec4"><title>Study design</title><p>The meta-analysis were performed according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement [<xref ref-type="bibr" rid="CR16">16</xref>].</p></sec><sec id="Sec5"><title>Data sources</title><p>PubMed, ScienceDirect and Cochrane Library databases were systematically searched for studies which reported total mortality rates in relation to BMI in patients with acute coronary syndrome. Multiple queries using following keywords were performed on August 27, 2014: (&#x02018;body mass index&#x02019; OR BMI OR &#x02018;body weight&#x02019; OR obesity OR overweight OR underweight) AND (&#x02018;acute coronary syndrome&#x02019; OR &#x02018;myocardial infarction&#x02019; OR &#x02018;unstable angina&#x02019;) AND (mortality OR death).</p></sec><sec id="Sec6"><title>Study eligibility criteria for qualitative and quantitative synthesis</title><p>Inclusion and exclusion criteria for qualitative and quantitative analyses were presented in Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>. Studies fulfilling the eligibility criteria were included into analysis.<table-wrap id="Tab1"><label>Table&#x000a0;1</label><caption><p>PICOS criteria for inclusion and exclusion of studies into qualitative and quantitative (meta-analysis) analyses</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left">Parameter</th><th align="left">Inclusion criteria</th><th align="left">Exclusion criteria</th></tr></thead><tbody><tr><td align="left" colspan="3">Qualitative synthesis criteria</td></tr><tr><td align="left">&#x000a0;Patients</td><td align="left"><p>Adults with acute coronary syndrome (STEMI and/or NSTEMI and/or UA), regardless of treatment (MT, fibrinolysis, PCI, CABG)</p><p>General population&#x02014;studies with subgroups (i.e. age or sex) were included only if there was possibility to compile subgroups into one cohort</p></td><td align="left"><p>only Korean or Japanese population</p><p>Population limited to a subgroup (i.e. age&#x000a0;&#x0003e;&#x000a0;65&#x000a0;years old or men only included)</p></td></tr><tr><td align="left">&#x000a0;Intervention</td><td align="left">Groups of BMI</td><td align="left">Studies without BMI groups</td></tr><tr><td align="left">&#x000a0;Comparator</td><td align="left">Normal BMI group</td><td align="left">&#x02013;</td></tr><tr><td align="left">&#x000a0;Outcomes</td><td align="left">All-cause (total) mortality</td><td align="left">&#x02013;</td></tr><tr><td align="left">&#x000a0;Study design</td><td align="left"><p>Randomized controlled trials</p><p>Non-randomized controlled trials</p><p>Retrospective, prospective, or concurrent cohort studies</p><p>Cross sectional studies</p></td><td align="left"><p>Case reports</p><p>Editorials &#x00026; opinion pieces</p></td></tr><tr><td align="left" colspan="3">Quantitative synthesis criteria^a </td></tr><tr><td align="left">&#x000a0;Patients</td><td align="left">&#x02013;</td><td align="left">&#x02013;</td></tr><tr><td align="left">&#x000a0;Intervention</td><td align="left">Low BMI, overweight, obesity, severe obesity (at least one of them)</td><td align="left">No BMI groups</td></tr><tr><td align="left">&#x000a0;Comparator</td><td align="left">Normal BMI group</td><td align="left">No possibility to extract normal BMI group</td></tr><tr><td align="left">&#x000a0;Outcomes</td><td align="left">All-cause (total) mortality expressed as mortlaity ratio, odds ratio or risk ratio</td><td align="left">Lack of mortality defined in BMI groups</td></tr><tr><td align="left">&#x000a0;Study design</td><td align="left">&#x02013;</td><td align="left">&#x02013;</td></tr></tbody></table><table-wrap-foot><p> ^aQuantitative synhesis criteria contain criteria for qualitative synthesis</p><p> <italic>PICOS</italic> patients, intervention, comparator, outcomes, study design; <italic>ACS</italic> acute coronary syndrome; <italic>BMI</italic> body mass index</p></table-wrap-foot></table-wrap> </p><p>Selection process was shown on Fig.&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref> and had been performed according to PRISMA statement [<xref ref-type="bibr" rid="CR16">16</xref>].<fig id="Fig1"><label>Fig.&#x000a0;1</label><caption><p>Flow diagram of the study (according to PRISMA statement)</p></caption><graphic xlink:href="10654_2014_9961_Fig1_HTML" id="MO1"/></fig> </p></sec><sec id="Sec7"><title>Study appraisal</title><p>Studies included in meta-analysis were appraised independently using Newcastle-Ottawa Quality Assessment Scale. Due to restricted inclusion/exclusion criteria, all of the studies had high (at least **) ratings in adequacy of selection and outcomes assessment. Comparability differed between studies, but meta-analysis was conducted on the basis of unadjusted mortality rates (see &#x0201c;<xref rid="Sec9" ref-type="sec">Methodology</xref>&#x0201d;). Agreement for the quality of the studies was over 90&#x000a0;%.</p></sec><sec id="Sec8"><title>Data extraction</title><p>Two reviewers (J.N. and B.H.) screened independently the titles and abstracts for relevance. Discrepancies between reviewers were discussed until consensus was reached. The articles of selected titles/abstracts were reviewed for inclusion. Using the above-mentioned selection criteria, these 2 reviewers determined independently the articles which were included and excluded. The data from the relevant articles were extracted using predefined extraction forms (Supplemental Appendix Table&#x000a0;1, available online). Any disagreements in data extraction were discussed until consensus was reached.</p></sec><sec id="Sec9"><title>Methodology</title><p>Due to differences in BMI groups between studies in our analysis (see the footnote of Table&#x000a0;<xref rid="Tab2" ref-type="table">2</xref>), patients were qualified to the closest BMI group. For the purpose of our meta-analysis subjects were divided into 5 groups: Low BMI, Normal BMI, Overweight, Obesity and Severe obesity. Due to heterogeneity of definitions of underweight used in different studies, in our Low BMI category we included subgroups of patients with BMI below 20&#x000a0;kg/m². Again, Normal BMI was defined as a BMI range from 18.5 to 25&#x000a0;kg/m², because in studies various BMI intervals were used i.e. 20&#x02013;25 or 18.5&#x02013;24.9&#x000a0;kg/m² (Table&#x000a0;<xref rid="Tab2" ref-type="table">2</xref>). Patients with BMI 25&#x02013;30 or 30&#x02013;35&#x000a0;kg/m² were categorized as Overweight and Obesity, respectively. Severe obesity category comprised patients with BMI&#x000a0;&#x02265;&#x000a0;35&#x000a0;kg/m². Patients with BMI 35&#x02013;39.9&#x000a0;kg/m² and patients with BMI 40&#x000a0;kg/m² or more were pooled as Severe obese (&#x02265;35).<table-wrap id="Tab2"><label>Table&#x000a0;2</label><caption><p>The summary of studies included into meta-analysis</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="2">Author</th><th align="left" rowspan="2">Year</th><th align="left" rowspan="2">Location</th><th align="left" rowspan="2">Enrolment period</th><th align="left" rowspan="2">ACS type</th><th align="left" rowspan="2">Number of patients</th><th align="left" rowspan="2">Treatment</th><th align="left" rowspan="2">Men &#x000a0;%</th><th align="left" rowspan="2">BMI category</th><th align="left" rowspan="2">Follow up (months)</th><th align="left" colspan="5">Prevalence (%)</th></tr><tr><th align="left">Low BMI</th><th align="left">Normal BMI</th><th align="left">Overweight</th><th align="left">Obesity</th><th align="left">Severe obesity</th></tr></thead><tbody><tr><td align="left">Hoit [<xref ref-type="bibr" rid="CR17">17</xref>]</td><td align="left">1987</td><td align="left">USA</td><td align="left">1979&#x02013;1983</td><td align="left">AMI</td><td char="." align="char">1,760</td><td align="left">M</td><td char="." align="char">75.4</td><td align="left">I</td><td align="left">IH&#x000a0;+&#x000a0;12</td><td char="." align="char">&#x02013;</td><td char="." align="char">37.4</td><td char="." align="char">50.2</td><td char="." align="char">12.4</td><td char="." align="char">&#x02013;</td></tr><tr><td align="left">Lopez-Jimenez [<xref ref-type="bibr" rid="CR18">18</xref>]</td><td align="left">2004</td><td align="left">USA</td><td align="left">1979&#x02013;1998</td><td align="left">AMI</td><td char="." align="char">2,263</td><td align="left">M P T</td><td char="." align="char">57.7</td><td align="left">G</td><td align="left">68.4</td><td char="." align="char">&#x02013;</td><td char="." align="char">36.0</td><td char="." align="char">40.0</td><td char="." align="char">24.0</td><td char="." align="char">&#x02013;</td></tr><tr><td align="left">Rana [<xref ref-type="bibr" rid="CR19">19</xref>]</td><td align="left">2004</td><td align="left">USA</td><td align="left">1989&#x02013;1994</td><td align="left">AMI</td><td char="." align="char">1,898</td><td align="left">NA</td><td char="." align="char">69.4</td><td align="left">A</td><td align="left">45.6</td><td char="." align="char">&#x02013;</td><td char="." align="char">32</td><td char="." align="char">44</td><td char="." align="char">17</td><td char="." align="char">7</td></tr><tr><td align="left">Eisenstein [<xref ref-type="bibr" rid="CR20">20</xref>]</td><td align="left">2005</td><td align="left">International</td><td align="left">1997&#x02013;1999</td><td align="left">ACS</td><td char="." align="char">15,071</td><td align="left">M P T C</td><td char="." align="char">72.7</td><td align="left">E</td><td align="left">12</td><td char="." align="char">&#x02013;</td><td char="." align="char">27.0</td><td char="." align="char">44.5</td><td char="." align="char">20.4</td><td char="." align="char">8.1</td></tr><tr><td align="left">Kragelund [<xref ref-type="bibr" rid="CR21">21</xref>]</td><td align="left">2005</td><td align="left">Denmark</td><td align="left">1990&#x02013;1992</td><td align="left">AMI</td><td char="." align="char">6,168</td><td align="left">M T</td><td char="." align="char">67.4</td><td align="left">M</td><td align="left">96</td><td char="." align="char">2.6</td><td char="." align="char">42.2</td><td char="." align="char">42.3</td><td char="." align="char">12.9</td><td char="." align="char">&#x02013;</td></tr><tr><td align="left">Diercks [<xref ref-type="bibr" rid="CR22">22</xref>]</td><td align="left">2006</td><td align="left">USA</td><td align="left">2001&#x02013;2003</td><td align="left">UA/NSTEMI</td><td char="." align="char">80,845</td><td align="left">M P C</td><td char="." align="char">60.4</td><td align="left">D</td><td align="left">IH</td><td char="." align="char">2.9</td><td char="." align="char">26.6</td><td char="." align="char">35.9</td><td char="." align="char">20.8</td><td char="." align="char">18.8</td></tr><tr><td align="left">Goldberg [<xref ref-type="bibr" rid="CR23">23</xref>]</td><td align="left">2006</td><td align="left">USA</td><td align="left">1997, 1999, 2001, 2003</td><td align="left">AMI</td><td char="." align="char">3,513</td><td align="left">P C</td><td char="." align="char">57.2</td><td align="left">F</td><td align="left">IH</td><td char="." align="char">7.0*</td><td char="." align="char">38.5</td><td char="." align="char">29.1</td><td char="." align="char">15.5</td><td char="." align="char">9.9</td></tr><tr><td align="left">Iakobishvili [<xref ref-type="bibr" rid="CR24">24</xref>]</td><td align="left">2006</td><td align="left">Israel</td><td align="left">2002&#x02013;2003</td><td align="left">STEMI</td><td char="." align="char">164</td><td align="left">P</td><td char="." align="char">75.6</td><td align="left">J</td><td align="left">1.0</td><td char="." align="char">&#x02013;</td><td char="." align="char">36.0</td><td char="." align="char">42.1</td><td char="." align="char">21.9</td><td char="." align="char">&#x02013;</td></tr><tr><td align="left">Nikolsky [<xref ref-type="bibr" rid="CR25">25</xref>]</td><td align="left">2006</td><td align="left">International</td><td align="left">1997&#x02013;1999</td><td align="left">AMI</td><td char="." align="char">2,035</td><td align="left">P</td><td char="." align="char">73.1</td><td align="left">G</td><td align="left">12</td><td char="." align="char">&#x02013;</td><td char="." align="char">27</td><td char="." align="char">45</td><td char="." align="char">28</td><td char="." align="char">&#x02013;</td></tr><tr><td align="left">Wells [<xref ref-type="bibr" rid="CR26">26</xref>]</td><td align="left">2006</td><td align="left">USA</td><td align="left">2003&#x02013;2004</td><td align="left">AMI</td><td char="." align="char">284</td><td align="left">M P T C</td><td char="." align="char">68.3</td><td align="left">L</td><td align="left">IH</td><td char="." align="char">6.0</td><td char="." align="char">22.2</td><td char="." align="char">34.2 1</td><td char="." align="char">22.9</td><td char="." align="char">14.8</td></tr><tr><td align="left">Buettner [<xref ref-type="bibr" rid="CR27">27</xref>]</td><td align="left">2007</td><td align="left">Germany</td><td align="left">1996&#x02013;1999</td><td align="left">UA/NSTEMI</td><td char="." align="char">1,676</td><td align="left">P</td><td char="." align="char">66.0</td><td align="left">A</td><td align="left">17</td><td char="." align="char">0.5*</td><td char="." align="char">32.9</td><td char="." align="char">49.2</td><td char="." align="char">14.6</td><td char="." align="char">17.4*</td></tr><tr><td align="left">Mehta [<xref ref-type="bibr" rid="CR28">28</xref>]</td><td align="left">2007</td><td align="left">International</td><td align="left">1990&#x02013;1997</td><td align="left">AMI</td><td char="." align="char">2,325</td><td align="left">P T</td><td char="." align="char">73.9</td><td align="left">G</td><td align="left">IH</td><td char="." align="char">&#x02013;</td><td char="." align="char">30.2</td><td char="." align="char">44.7</td><td char="." align="char">25.1</td><td char="." align="char">&#x02013;</td></tr><tr><td align="left">Lopez-Jimenez [<xref ref-type="bibr" rid="CR29">29</xref>]</td><td align="left">2008</td><td align="left">USA</td><td align="left">1996&#x02013;2001</td><td align="left">AMI</td><td char="." align="char">1,676</td><td align="left">M P C</td><td char="." align="char">55.9</td><td align="left">K</td><td align="left">29</td><td char="." align="char">3.6</td><td char="." align="char">22.8</td><td char="." align="char">37.6</td><td char="." align="char">30.2</td><td char="." align="char">5.8*</td></tr><tr><td align="left">Mehta [<xref ref-type="bibr" rid="CR30">30</xref>]</td><td align="left">2008</td><td align="left">Germany</td><td align="left">1994&#x02013;2002</td><td align="left">STEMI</td><td char="." align="char">7,630</td><td align="left">P T</td><td char="." align="char">70.7</td><td align="left">G</td><td align="left">IH</td><td char="." align="char">&#x02013;</td><td char="." align="char">29.8</td><td char="." align="char">49.3</td><td char="." align="char">20.8</td><td char="." align="char">&#x02013;</td></tr><tr><td align="left">Wienbergen [<xref ref-type="bibr" rid="CR31">31</xref>]</td><td align="left">2008</td><td align="left">Germany</td><td align="left">1998&#x02013;2002</td><td align="left">STEMI</td><td char="." align="char">10,534</td><td align="left">M P T C</td><td char="." align="char">70.2</td><td align="left">D</td><td align="left">IH&#x000a0;+&#x000a0;14</td><td char="." align="char">&#x02013;</td><td char="." align="char">32.3</td><td char="." align="char">43.5</td><td char="." align="char">20.2</td><td char="." align="char">&#x02013;</td></tr><tr><td align="left">Aronson [<xref ref-type="bibr" rid="CR32">32</xref>]</td><td align="left">2010</td><td align="left">Israel</td><td align="left">2001&#x02013;2007</td><td align="left">AMI</td><td char="." align="char">2,157</td><td align="left">M P</td><td char="." align="char">78.7</td><td align="left">B</td><td align="left">26</td><td char="." align="char">1.2</td><td char="." align="char">28.7</td><td char="." align="char">44.2</td><td char="." align="char">20.1</td><td char="." align="char">5.8</td></tr><tr><td align="left">Hadi [<xref ref-type="bibr" rid="CR33">33</xref>]</td><td align="left">2010</td><td align="left">Middle East</td><td align="left">2006&#x02013;2007</td><td align="left">ACS</td><td char="." align="char">7,843</td><td align="left">P T</td><td char="." align="char">75.8</td><td align="left">G</td><td align="left">IH</td><td char="." align="char">&#x02013;</td><td char="." align="char">32.8</td><td char="." align="char">40.4</td><td char="." align="char">26.7</td><td char="." align="char">&#x02013;</td></tr><tr><td align="left">Mahaffey [<xref ref-type="bibr" rid="CR34">34</xref>]</td><td align="left">2010</td><td align="left">International</td><td align="left">2001&#x02013;2003</td><td align="left">UA/NSTEMI</td><td char="." align="char">9,873</td><td align="left">M P C</td><td char="." align="char">66.2</td><td align="left">L</td><td align="left">1.0</td><td char="." align="char">2.4</td><td char="." align="char">23.8</td><td char="." align="char">41.5</td><td char="." align="char">21.7</td><td char="." align="char">10.1</td></tr><tr><td align="left">Shechter [<xref ref-type="bibr" rid="CR35">35</xref>]</td><td align="left">2010</td><td align="left">Israel</td><td align="left">2002, 2004, 2006</td><td align="left">ACS</td><td char="." align="char">5,751</td><td align="left">M P C</td><td char="." align="char">77.0</td><td align="left">E</td><td align="left">12</td><td char="." align="char">0.8</td><td char="." align="char">29.7</td><td char="." align="char">46.9</td><td char="." align="char">22.6</td><td char="." align="char">&#x02013;</td></tr><tr><td align="left">Das [<xref ref-type="bibr" rid="CR36">36</xref>]</td><td align="left">2011</td><td align="left">USA</td><td align="left">2007&#x02013;2009</td><td align="left">STEMI</td><td char="." align="char">49,329</td><td align="left">P T</td><td char="." align="char">70.5</td><td align="left">D</td><td align="left">IH</td><td char="." align="char">&#x02013;</td><td char="." align="char">23.5</td><td char="." align="char">38.7</td><td char="." align="char">22.4</td><td char="." align="char">13.8</td></tr><tr><td align="left">Timoteo [<xref ref-type="bibr" rid="CR37">37</xref>]</td><td align="left">2011</td><td align="left">Portugal</td><td align="left">2005&#x02013;2008</td><td align="left">STEMI</td><td char="." align="char">539</td><td align="left">P</td><td char="." align="char">77.0</td><td align="left">C</td><td align="left">12</td><td char="." align="char">&#x02013;</td><td char="." align="char">34.9</td><td char="." align="char">46.2</td><td char="." align="char">18.9</td><td char="." align="char">&#x02013;</td></tr><tr><td align="left">Bucholz [<xref ref-type="bibr" rid="CR38">38</xref>]</td><td align="left">2012</td><td align="left">USA</td><td align="left">2003&#x02013;2008</td><td align="left">AMI</td><td char="." align="char">6,359</td><td align="left">M P C</td><td char="." align="char">67.4</td><td align="left">A</td><td align="left">12</td><td char="." align="char">&#x02013;</td><td char="." align="char">22.8</td><td char="." align="char">36.4</td><td char="." align="char">24.1</td><td char="." align="char">16.7</td></tr><tr><td align="left">Camprubi [<xref ref-type="bibr" rid="CR39">39</xref>]</td><td align="left">2012</td><td align="left">Spain</td><td align="left">2009&#x02013;2010</td><td align="left">ACS</td><td char="." align="char">824</td><td align="left">P</td><td char="." align="char">73.5</td><td align="left">C</td><td align="left">IH</td><td char="." align="char">&#x02013;</td><td char="." align="char">27.6</td><td char="." align="char">50.6</td><td char="." align="char">21.8</td><td char="." align="char">&#x02013;</td></tr><tr><td align="left">Lazzeri [<xref ref-type="bibr" rid="CR40">40</xref>]</td><td align="left">2012</td><td align="left">Italy</td><td align="left">2004&#x02013;2010</td><td align="left">STEMI</td><td char="." align="char">1,268</td><td align="left">P</td><td char="." align="char">73.2</td><td align="left">O</td><td align="left">IH&#x000a0;+&#x000a0;12</td><td char="." align="char">2.9</td><td char="." align="char">31.8</td><td char="." align="char">51.7</td><td char="." align="char">13.6</td><td char="." align="char">&#x02013;</td></tr><tr><td align="left">Herrmann [<xref ref-type="bibr" rid="CR41">41</xref>]</td><td align="left">2014</td><td align="left">International</td><td align="left">2005&#x02013;2007</td><td align="left">STEMI</td><td char="." align="char">3,579</td><td align="left">M P C</td><td char="." align="char">76.6</td><td align="left">H</td><td align="left">36</td><td char="." align="char">&#x02013;</td><td char="." align="char">29.5</td><td char="." align="char">64.3</td><td char="." align="char">6.2</td><td char="." align="char">&#x02013;</td></tr><tr><td align="left">Witassek [<xref ref-type="bibr" rid="CR42">42</xref>]</td><td align="left">2014</td><td align="left">Switzerland</td><td align="left">2006&#x02013;2012</td><td align="left">STEMI</td><td char="." align="char">6,938</td><td align="left">P</td><td char="." align="char">77.1</td><td align="left">A</td><td align="left">IH</td><td char="." align="char">1.0</td><td char="." align="char">33.1</td><td char="." align="char">45.0</td><td char="." align="char">15.9</td><td char="." align="char">5.0</td></tr><tr><td align="left">26 Studies</td><td align="left"/><td align="left"/><td align="left">1979&#x02013;2012</td><td align="left"/><td char="." align="char">218,532</td><td align="left"/><td align="left"/><td align="left"/><td align="left"/><td align="left"/><td align="left"/><td align="left"/><td align="left"/><td align="left"/></tr></tbody></table><table-wrap-foot><p> <italic>ACS</italic> acute coronary syndrome, <italic>AMI</italic> acute myocardial infarction, <italic>UA</italic> unstable angina, <italic>NSTEMI</italic> non-ST-elevation myocardial infarction, <italic>STEMI</italic> ST-elevation myocardial infarction, <italic>NA</italic> not applicable/not available, <italic>IH</italic> in-hospital; <italic>USA</italic> United States of America, *&#x000a0;No mortality rates/survival analysis for this BMI subgroup (only prevalence available)</p><p>Treatment: <italic>M</italic> medical treatment, <italic>T</italic> thrombolysis, <italic>P</italic> percutaneous revascularization, <italic>C</italic> coronary artery bypass surgery (CABG)</p><p>Reported BMI categories (kg/m²): A&#x02014;Underweight: &#x0003c;18.5; Normal: 18.5&#x02013;24.9; Overweight: 25&#x02013;29.9; Obese: 30&#x02013;34.9; Severe obese: &#x02265;35; B&#x02014;Underweight: &#x0003c;18.5; Normal: 18.5&#x02013;21 AND 21&#x02013;23.5 (reference) AND 23.5&#x02013;25; Overweight: 25&#x02013;26.5 AND 26.5&#x02013;28 (overweight referent) AND 28&#x02013;30; Obese: 30&#x02013;35; Severe obese: &#x02265;35; C&#x02014;Normal: &#x0003c;25; Overweight: 25&#x02013;29.9; Obese: &#x0003e;30; D&#x02014;Underweight &#x0003c;18.5; Normal: 18.5&#x02013;24.9; Overweight: 25&#x02013;29.9; Obese (class I) 30&#x02013;34.9; Obese: (class II) 35&#x02013;39.9; Obese: (class III) &#x02265;40 (severe obesity&#x000a0;=&#x000a0;class II&#x000a0;+&#x000a0;III obesity); E&#x02014;Underweight: &#x0003c;18.5; Normal: 18.5&#x02013;24.9; Overweight: 25&#x02013;29.9; Obese: (class I, II, III) &#x02265;30; F&#x02014;Normal: &#x0003c;25; Overweight: 25&#x02013;29.9; Obese: 30&#x02013;34.9, Severe obese: &#x02265;35; G&#x02014;Normal: &#x0003c;25; Overweight: 25&#x02013;29.9; Obese: &#x02265;30; H&#x02014;Normal: &#x0003c;24.5; Overweight: 24.5&#x02013;27 AND 27.1&#x02013;30.1; Obese: &#x0003e;30.1; I&#x02014;Normal: &#x0003c;25; Overweight: 25&#x02013;34.9; Obese: &#x0003e;35; J&#x02014;Normal: &#x02264;25; Overweight: 25&#x02013;30; Obese: &#x0003e;30; K&#x02014;Underweight: &#x0003c;20; Normal: 20&#x02013;24.9; Overweight: 25&#x02013;29.9; Obese: 30&#x02013;39.9; Morbidly obese: &#x02265;40 (obesity&#x000a0;=&#x000a0;&#x02265;30); L&#x02014;Underweight: &#x0003c;20; Normal: 20&#x02013;25; Overweight: 25&#x02013;30; Obese: 30&#x02013;35; Severe obese: &#x02265;35; M&#x02014;Underweight: &#x0003c;18.5; Normal: 18.5&#x02013;24.9; Overweight: 25&#x02013;29.9; Obese: &#x0003e;30</p></table-wrap-foot></table-wrap> </p></sec><sec id="Sec10"><title>Statistical analyses</title><p>A random effects model with inverse variance weighting was used to calculate pooled relative risks (RR) and 95&#x000a0;% confidence interval (CI). Total mortality after ACS was analyzed. Unadjusted mortality rates (2&#x000a0;&#x000d7;&#x000a0;2 or risk ratios) in BMI groups were extracted from studies. Normal BMI group was chosen as the reference one. Heterogeneity between studies was assessed using Cochran Q test and I2 statistic, which denotes the percentage of total variation across studies as a result of heterogeneity rather than chance. All heterogeneity results from analyses of each group were compared with those of the Normal-BMI group. Heterogeneity was considered significant if the P value for the heterogeneity test was less than 0.05. Publication bias was tested by using the Begg and Mazumdar rank correlation test and the Egger&#x02019;s regression intercept test. In case of significant bias, Duval and Tweedie&#x02019;s trim and fill method was applied to correct the funnel plot asymmetry. The effect of individual studies was examined by exclusion sensitivity analysis. Each study was removed at a time to assess the degree to which the meta-analysis estimate depends on that particular study.</p></sec></sec><sec id="Sec11"><title>Results</title><sec id="Sec12"><title>Study characteristics</title><p>Out of the 49 pre-selected articles, 26 met inclusion criteria for meta-analysis [<xref ref-type="bibr" rid="CR17">17</xref>&#x02013;<xref ref-type="bibr" rid="CR42">42</xref>].</p><p>218,532 patients with ACS, enrolled in years 1979&#x02013;2012 were included in the study. Each study contained more men (range between 55.9 and 78.7&#x000a0;%) than women.</p><p>Excluded articles with criterion for exclusion were shown in the frame on Fig.&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>. To avoid bias due to the differences in diagnostic criteria of overweight and obesity, data from Japanese and South Korean populations were excluded from the analysis (4 studies).</p></sec><sec id="Sec13"><title>Main analysis</title><p>The relative risk ratio for total mortality in patients after ACS with Low BMI was RR 1.74 (CI 1.47&#x02013;2.05)&#x02014;Fig.&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>. The Begg and Mazumdar rank correlation test was not significant (<italic>p</italic>&#x000a0;=&#x000a0;0.47), but Egger&#x02019; s regression intercept test showed significant bias for publications (<italic>p</italic>&#x000a0;=&#x000a0;0.006). The Duval and Tweedie&#x02019;s Trim and Fill method was used to impute 5 missing studies and estimate RR as 1.47 (1.24&#x02013;1.74).<fig id="Fig2"><label>Fig.&#x000a0;2</label><caption><p>Meta-analysis: total mortality risk for Low BMI versus Normal BMI in patients with acute coronary syndrome</p></caption><graphic xlink:href="10654_2014_9961_Fig2_HTML" id="MO2"/></fig> <fig id="Fig3"><label>Fig.&#x000a0;3</label><caption><p>Meta-analysis: total mortality risk for Overweight versus Normal BMI in patients with acute coronary syndrome</p></caption><graphic xlink:href="10654_2014_9961_Fig3_HTML" id="MO3"/></fig> </p><p>Overweight patients had 30&#x000a0;% lower mortality risk after ACS in comparison to those with Normal BMI&#x02013;RR 0.70 (CI 0.64&#x02013;0.76)&#x02014;Fig.&#x000a0;<xref rid="Fig4" ref-type="fig">3</xref>.<fig id="Fig4"><label>Fig.&#x000a0;4</label><caption><p>Meta-analysis: total mortality risk for Obesity versus Normal BMI in patients with acute coronary syndrome</p></caption><graphic xlink:href="10654_2014_9961_Fig4_HTML" id="MO4"/></fig> </p><p>Obesity was related to 40&#x000a0;% lower risk of death after ACS in comparison with Normal-BMI subjects&#x02014;RR 0.60 (95&#x000a0;% CI 0.53&#x02013;0.68)&#x02014;Fig.&#x000a0;<xref rid="Fig3" ref-type="fig">4</xref>.</p><p>Severely obese patients had 30&#x000a0;% lower mortality risk after ACS in comparison to those with Normal BMI&#x02014;RR 0.70 (CI 0.58&#x02013;0.86)&#x02014;Fig.&#x000a0;<xref rid="Fig5" ref-type="fig">5</xref>.<fig id="Fig5"><label>Fig.&#x000a0;5</label><caption><p>Meta-analysis: total mortality risk for Severe Obesity versus Normal BMI in patients with acute coronary syndrome</p></caption><graphic xlink:href="10654_2014_9961_Fig5_HTML" id="MO5"/></fig> </p><p>Both tests used for publication bias assessment were not significant for Overweight, Obesity nor Severe obesity groups.</p><p>The relation between risk of mortality and BMI groups was U-shaped&#x02014;Fig.&#x000a0;<xref rid="Fig6" ref-type="fig">6</xref>.<fig id="Fig6"><label>Fig.&#x000a0;6</label><caption><p>Risk ratios (RR) assessed in meta-analysis in groups of BMI</p></caption><graphic xlink:href="10654_2014_9961_Fig6_HTML" id="MO6"/></fig> </p></sec></sec><sec id="Sec14"><title>Discussion</title><sec id="Sec15"><title>Age and sex</title><p>In 20 of 26 studies, overweight and/or obese patients were younger (1&#x02013;10&#x000a0;years). Madala et al. [<xref ref-type="bibr" rid="CR43">43</xref>] observed that the first NSTEMI occurred 12&#x000a0;years earlier in severely obese than in normal BMI patients, whilst only 3.5&#x000a0;years earlier in less endangered overweight group. The finding of younger age of obese patients admitted for ACS therapy could be one of possible explanation for the better survival after ACS in people with BMI&#x000a0;&#x02265;&#x000a0;25&#x000a0;kg/m². Peto et al. [<xref ref-type="bibr" rid="CR44">44</xref>] showed that in general population patients with BMIs above 25&#x000a0;kg/m² had an expected lifetime about 10&#x000a0;years shorter than people with normal BMI. Thus, the percentage of obese people in the population decreases with increasing age.</p><p>In patients aged 65&#x000a0;years or older, mortality was higher among obese patients in comparison with those with overweight (<italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.01) and normal weights (<italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.001). Obesity in this age group was an independent risk factor for in-hospital mortality [<xref ref-type="bibr" rid="CR17">17</xref>].</p><p>There are different reports on sex distribution across BMI groups. In some studies (Aronson, Eisenstein) more women, while in others [<xref ref-type="bibr" rid="CR18">18</xref>, <xref ref-type="bibr" rid="CR28">28</xref>, <xref ref-type="bibr" rid="CR30">30</xref>] more men were included in the obese groups. Rana et al. [<xref ref-type="bibr" rid="CR19">19</xref>] showed more women in normal-weight and class 1 and 2 obesity with nadir in the overweight ones (39, 33, 40 and 22&#x000a0;%, respectively, <italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.001). Similar differences were found for cardiogenic shock with occurrence 9.0; 4.1; 3.1; 2.9 and 5.4&#x000a0;% for underweight, normal weight, overweight, class 1 and class 2/3 obesity (<italic>p</italic>&#x000a0;=&#x000a0;0.006), respectively [<xref ref-type="bibr" rid="CR42">42</xref>].</p></sec><sec id="Sec16"><title>Comorbidities and complications</title><p>Patients with BMI&#x000a0;&#x02265;&#x000a0;25&#x000a0;kg/m² had higher cardiovascular risk. Diabetes mellitus (20 studies), hypertension (20 studies) or hyperlipidemia (10 studies) were more prevalent in obese than in normal-BMI group. Nevertheless, two studies showed lower GRACE risk score in obese patients [<xref ref-type="bibr" rid="CR35">35</xref>, <xref ref-type="bibr" rid="CR38">38</xref>].</p><p>Better survival in overweight or obese patients might be due to the relatively short follow-ups in the studies. During in-hospital stay or even in 5&#x000a0;years after MI, diabetes mellitus or hypertension had little chance to evoke complications and impact the mortality.</p><p>Although overweight or obese patients smoked rarely [<xref ref-type="bibr" rid="CR19">19</xref>&#x02013;<xref ref-type="bibr" rid="CR21">21</xref>, <xref ref-type="bibr" rid="CR28">28</xref>, <xref ref-type="bibr" rid="CR33">33</xref>, <xref ref-type="bibr" rid="CR35">35</xref>, <xref ref-type="bibr" rid="CR41">41</xref>], mortality risk among current smokers was higher in these groups and rose with increasing BMI&#x02013;hazard ratio (HR) for BMI&#x000a0;&#x0003e;&#x000a0;35&#x000a0;kg/m² was 4.51 (95&#x000a0;% CI, 1.42&#x02013;14.3) in comparison to HR 1.18 (95&#x000a0;% CI, 0.42&#x02013;2.58) for former smokers [<xref ref-type="bibr" rid="CR19">19</xref>]. Only 8&#x000a0;% of underweight patients smoked in the past in comparison to 15, 16 and 17&#x000a0;% found in normal-weight, overweight and obese subjects respectively (<italic>p</italic>&#x000a0;=&#x000a0;0.001) [<xref ref-type="bibr" rid="CR21">21</xref>].</p><p>Obese patients had higher concentrations of C-reactive protein [<xref ref-type="bibr" rid="CR27">27</xref>], lower troponin and NT-proBNP levels [<xref ref-type="bibr" rid="CR45">45</xref>]. The finding of lower natriuretic peptides levels in obese heart failure patients has been recognized recently was and could be explained by clearance function of adipose tissue on these peptides [<xref ref-type="bibr" rid="CR46">46</xref>].</p><p>Compared to normal-BMI group, in obese patients higher estimated glomerular filtration rates by both, MDRD or Cockroft-Gault formulas were observed [<xref ref-type="bibr" rid="CR25">25</xref>, <xref ref-type="bibr" rid="CR36">36</xref>, <xref ref-type="bibr" rid="CR47">47</xref>]. The choice of renal function estimation may be important because in patients with coronary artery disease and serum creatinine within normal range, CKD-EPI formula (Chronic Kidney Diseases Epidemiology Initiative) which was derived based on populations with vaster distribution of BMI, predicted long-term outcome more accurately, than MDRD equation [<xref ref-type="bibr" rid="CR48">48</xref>].</p><p>Patients with BMI&#x000a0;&#x0003c;&#x000a0;25 had higher risk of bleeding [<xref ref-type="bibr" rid="CR25">25</xref>, <xref ref-type="bibr" rid="CR34">34</xref>]. Nikolsky et al. [<xref ref-type="bibr" rid="CR25">25</xref>] postulated that the difference had been determined by gastro-intestinal bleeding (2.7 vs 0.4, <italic>p</italic>&#x000a0;=&#x000a0;0.02 for normal weight and obesity, respectively). Moreover, overweight and obese more often had anemia [<xref ref-type="bibr" rid="CR41">41</xref>] and indication for blood transfusion [<xref ref-type="bibr" rid="CR25">25</xref>]. Noteworthy, the local groin bleeds (hematoma in the arterial puncture site) occurred also more frequently in patients with normal body weight, compared with overweight and obese (11, 6.8 and 7.6&#x000a0;%, respectively, <italic>p</italic>&#x000a0;=&#x000a0;0.014) [<xref ref-type="bibr" rid="CR28">28</xref>]. This phenomenon could be explained by ability of fat tissue to compress punctured femoral artery and staunch bleeding.</p><p>Obese patients had less often history of stroke [<xref ref-type="bibr" rid="CR18">18</xref>, <xref ref-type="bibr" rid="CR21">21</xref>] and rarely in-hospital stroke [<xref ref-type="bibr" rid="CR39">39</xref>], but this also could be explained by the differences in age.</p><p>Kragelund et al. [<xref ref-type="bibr" rid="CR21">21</xref>] showed that prevalence of cancer was more likely in underweight women group: 12 vs 5&#x000a0;%, 3 and 4&#x000a0;% in normal-weight, overweight and obese groups respectively (<italic>p</italic>&#x000a0;=&#x000a0;0.001). The observation was confirmed by Anger&#x000e5;s et al. [<xref ref-type="bibr" rid="CR49">49</xref>] (from 8.7&#x000a0;% in underweight to 1.9&#x000a0;% in patients with BMI&#x000a0;&#x02265;&#x000a0;35&#x000a0;kg/m², <italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.001).</p></sec><sec id="Sec17"><title>Diagnosis and treatment</title><p>Angiotensin converting enzyme inhibitors (ACEI) were used more frequently in obese as compared to normal weight patients with ACS in 9 studies. Similarly beta-blockers (BB) or statins were given with higher probability to obese patients in 12 and 11 studies respectively. Better pharmacological treatment in obese patients might be caused by existence of other indications for these drugs such as hypertension (20 studies) among obese.</p><p>In four studies coronary angiography was reported more often in obese patients [<xref ref-type="bibr" rid="CR22">22</xref>, <xref ref-type="bibr" rid="CR23">23</xref>, <xref ref-type="bibr" rid="CR33">33</xref>, <xref ref-type="bibr" rid="CR34">34</xref>]. Additionally, six studies reported less frequent percutaneous coronary revascularization in underweight or normal-weight patients with ACS [<xref ref-type="bibr" rid="CR20">20</xref>, <xref ref-type="bibr" rid="CR22">22</xref>, <xref ref-type="bibr" rid="CR23">23</xref>, <xref ref-type="bibr" rid="CR31">31</xref>, <xref ref-type="bibr" rid="CR32">32</xref>, <xref ref-type="bibr" rid="CR34">34</xref>].</p><p>The door-to-balloon time was significantly longer in obese compared with normal weight patients. Moreover, they had more often final TIMI flow grade 0 compared to normal-weight individuals (2.0 vs. 0.4&#x000a0;%, respectively; <italic>p</italic>&#x000a0;=&#x000a0;0.04) [<xref ref-type="bibr" rid="CR28">28</xref>]. Initial TIMI flow grade 0 or 1 was also differs between in normal-weight and overweight patients (1.8 vs 0.7&#x000a0;%, respectively, <italic>p</italic>&#x000a0;=&#x000a0;0.04), as well as between overweight and obese subjects (0.7 vs 2.1&#x000a0;%, respectively, <italic>p</italic>&#x000a0;=&#x000a0;0.01) [<xref ref-type="bibr" rid="CR25">25</xref>].</p><p>Multi-vessel coronary artery disease was more common in patients with a normal body weight than in obese with BMI&#x000a0;&#x02265;&#x000a0;40&#x000a0;kg/m², according to studies of Das et al. (28.4 vs 22.4&#x000a0;%) and Diercks et al. (30.0 vs 24.6&#x000a0;%) [<xref ref-type="bibr" rid="CR22">22</xref>, <xref ref-type="bibr" rid="CR36">36</xref>]. Nikolsky et al. [<xref ref-type="bibr" rid="CR25">25</xref>] did not confirm the higher occurrence of multi-vessel coronary artery disease in normal-weight with STEMI and showed the same frequency of percutaneous (and surgical) revascularization in all BMI ranges.</p><p>Despite the lack of differences in the effect of angioplasty, patients with normal weight required a longer hospital stay: 7.1, 6.9, and 6.7&#x000a0;days for normal weight, overweight, and obese, respectively; <italic>p</italic>&#x000a0;=&#x000a0;0.014. Major adverse cardiovascular events (MACE) at 6&#x000a0;months was also observed more often in the normal BMI range in comparison with overweight and obese cases: 8.8, 6.6, and 5.0&#x000a0;% respectively; <italic>p</italic>&#x000a0;=&#x000a0;0.031 [<xref ref-type="bibr" rid="CR28">28</xref>]. Major adverse cardiovascular or cerebrovascular events (MACCE) was also more frequent in normal-weight patients, comparing to overweight and obese subjects: 14.7, 12.7, 10.0&#x000a0;%, respectively for in-hospital outcome (<italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.001) and 12.6, 9.3, 8.7&#x000a0;%, respectively (<italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.001) for long-term follow-up [<xref ref-type="bibr" rid="CR31">31</xref>].</p></sec><sec id="Sec18"><title>Central obesity and weight loss</title><p>Only four studies highlighted the prognostic role of central obesity. Zeller et al. divided patients with myocardial infarction (MI) into the tertiles of BMI and waist circumference (WC). The group of lower or middle tertile of BMI and upper tertile of WC had 1-year mortality risk above 20&#x000a0;% in women and more than 18&#x000a0;% in men, whilst in lower WC and upper BMI tertiles mortality was 7.6 and 7.7&#x000a0;%, respectively [<xref ref-type="bibr" rid="CR50">50</xref>]. This finding was confirmed by Kadakia et al. [<xref ref-type="bibr" rid="CR45">45</xref>]. It may indicate the special significance of central obesity. Unfortunately, most of the studies did not report parameters allowing more detailed description of obesity phenotype. Kragelund et al. [<xref ref-type="bibr" rid="CR21">21</xref>] confirmed abdominal obesity assessed by waist-to-hip ratio, to be independent predictor of all-cause mortality in men (adjusted RR 1.22 (1.07&#x02013;1.38), <italic>p</italic>&#x000a0;&#x0003c;&#x000a0;0.01), but not in women subgroup after ACS [adjusted RR 1.13 (0.95&#x02013;1.34, <italic>p</italic>&#x000a0;=&#x000a0;0.2)].</p><p>Guidelines of European Society of Cardiology (ESC) for the prevention of cardiovascular disease in clinical practice, highlights that obesity in the general population is associated with an increased incidence of cardiovascular disease and cardiovascular mortality. Therefore, the recommendation (class I, level of evidence A) exists for a weight reduction of overweight or obese individuals who have not undergone any cardiovascular event. Body weight reduction to the normal range (BMI 20&#x02013;24.9 a kg/m²) has a positive effect on blood pressure and plasma lipids, which is reflected in a lower incidence of cardiovascular disease [<xref ref-type="bibr" rid="CR51">51</xref>]. So far, no studies have confirmed the mortality reduction after MI in patients who reduced their body weight [<xref ref-type="bibr" rid="CR52">52</xref>]. On the contrary, weight loss of more than 5&#x000a0;% after MI in patients with depression (found in 27&#x000a0;% of patients) was related to 70&#x000a0;% higher risk of all-cause and cardiovascular mortality and those finding were not associated with depression nor social support [<xref ref-type="bibr" rid="CR29">29</xref>]. Weight loss of more than 5&#x000a0;% in a South Korean population of patients following acute MI was associated with a higher 1-year rate of MACEs. Patients who gained weight also have a greater 1-year mortality risk [<xref ref-type="bibr" rid="CR7">7</xref>]. On the other hand, intentional weight loss during cardiac rehabilitation in patients with CAD (not MI) was a marker for favourable long-term (6.4&#x000a0;years) outcomes, in both subgroups with initial BMI&#x000a0;&#x0003c;&#x000a0;25 or &#x02265;25&#x000a0;kg/m² [<xref ref-type="bibr" rid="CR53">53</xref>].</p></sec><sec id="Sec19"><title>Comparison to general population</title><p>The collected data showed that in a population of patients with ACS, an obesity paradox may occur. However, a meta-analysis of 97 studies about mortality in the general population, published in January 2013, indirectly calls into question the existence of the obesity paradox in patients with ACS and chronic diseases. In the general population, the risk of death (HR) in people who were overweight and in the 1st class of obesity (BMI 25&#x02013;35&#x000a0;kg/m²) was lower than in individuals with normal weights (BMI 18.5&#x02013;25&#x000a0;kg/m²). Only patients with BMIs 35&#x000a0;kg/m² and greater had a higher risk of death [<xref ref-type="bibr" rid="CR54">54</xref>]. To compare the results of the studies about BMI and mortality in chronic diseases with the work of Flegal et al. [<xref ref-type="bibr" rid="CR54">54</xref>], the obesity paradox exists also in the general population. In the ACS, chronic diseases and the general population the lowest mortality was observed among individuals with BMI values above the normal WHO range.</p><p>Although results of our study seem to be clear and quite obvious, outcomes should be interpreted with caution. Despite obese patients more often had diabetes mellitus and/or hypertension, they were younger and had less bleeding complications. Therefore, to compare the mortality of obese patients with people with normal BMIs, the age of the patients and associated diseases should be taken into account in long enough follow-up. In other cases, the relationship between BMI and mortality may be disturbed.</p><p>In unadjusted analyses performed on data assessed from the studies, better survival in overweight, obesity and severe obesity group was confirmed in 16 out of 26 studies, 19 of 26 and 5 of 10 studies, respectively. In Low BMI group 7 of 9 studies showed worse survival, comparing to Normal BMI group. After adjustment, both for multivariate analysis (BMI as continuous variable) or models adjusted for various covariables (BMI groups), significant relation between lower BMI and worse survival was found in 15 out of 25 studies.</p></sec></sec><sec id="Sec20"><title>Conclusion</title><p>The existence of obesity paradox in patients with ACS is supported by our meta-analysis.</p><sec id="Sec21"><title>Limitations</title><p>Our study has some limitations and weaknesses.</p><p>The analyzed articles varied in methodology. Groups of BMI were categorized using 11 different classification (see footnote of Table&#x000a0;<xref rid="Tab2" ref-type="table">2</xref>). Thus, in some studies BMI 19&#x000a0;kg/m² was classified as &#x02018;Low BMI&#x02019;, in other&#x02014;as &#x02018;Normal BMI&#x02019;. In some publications, underweight patients were excluded from the analyses, because of the &#x02018;extreme high risk of mortality&#x02019; [<xref ref-type="bibr" rid="CR38">38</xref>].</p><p>There were lacks of detailed data on race, age, treatment or complications in most of studies, thus those parameters were not shown in the analysis.</p><p>The reliability of the data on height and weight is also an important issue. Significant discrepancies between the values measured by physicians and those reported by patients have been shown [<xref ref-type="bibr" rid="CR54">54</xref>]. Nevertheless, in most ACS cases, weight and height measurements are not possible to conduct, due to life-threatening condition.</p></sec></sec><sec sec-type="supplementary-material"><title>Electronic supplementary material</title><sec id="Sec22"><p>Below is the link to the electronic supplementary material. <supplementary-material content-type="local-data" id="MOESM1"><media xlink:href="10654_2014_9961_MOESM1_ESM.docx"><caption><p>Supplementary material 1 (DOCX 12&#x000a0;kb)</p></caption></media></supplementary-material> </p></sec></sec>

Supporting Pregnant Aboriginal and Torres Strait Islander Women to Quit Smoking: Views of Antenatal Care Providers and Pregnant Indigenous Women

Could not extract abstract

<contrib contrib-type="author" corresp="yes"><name><surname>Passey</surname><given-names>Megan E.</given-names></name><address><email>megan.passey@ucrh.edu.au</email></address><xref ref-type="aff" rid="Aff1">1</xref></contrib><contrib contrib-type="author"><name><surname>Sanson-Fisher</surname><given-names>Rob W.</given-names></name><xref ref-type="aff" rid="Aff2">2</xref></contrib><contrib contrib-type="author"><name><surname>Stirling</surname><given-names>Janelle M.</given-names></name><xref ref-type="aff" rid="Aff1">1</xref></contrib><aff id="Aff1"><label>1</label>University of Sydney, Lismore, NSW Australia </aff><aff id="Aff2"><label>2</label>University of Newcastle, Newcastle, NSW Australia </aff>

Maternal and Child Health Journal

<sec id="Sec1"><title>Introduction</title><p>Tobacco smoking among pregnant Aboriginal and Torres Strait Islander women remains three times as common as among non-Indigenous Australian pregnant women, with approximately 50&#x000a0;% of women smoking during pregnancy [<xref ref-type="bibr" rid="CR1">1</xref>]. Addressing this disparity is a priority for reducing the gap in health outcomes between Indigenous and non-Indigenous Australians. Disparities in smoking rates between Indigenous and non-Indigenous pregnant women are also marked in the United States, Canada and New Zealand [<xref ref-type="bibr" rid="CR2">2</xref>&#x02013;<xref ref-type="bibr" rid="CR4">4</xref>]. While interventions to reduce antenatal smoking are known to be effective in non-Indigenous populations [<xref ref-type="bibr" rid="CR5">5</xref>], to date effective interventions for pregnant Indigenous women have not been identified [<xref ref-type="bibr" rid="CR6">6</xref>&#x02013;<xref ref-type="bibr" rid="CR8">8</xref>].</p><p>Previous reviews of interventions for smoking cessation in Indigenous peoples have concluded that approaches that specifically target Indigenous populations can be successful [<xref ref-type="bibr" rid="CR9">9</xref>, <xref ref-type="bibr" rid="CR10">10</xref>], and that interventions targeting individuals, such as counselling and nicotine replacement therapy (NRT), which are known to be effective in other populations, are likely to be effective for Indigenous people [<xref ref-type="bibr" rid="CR11">11</xref>]. However, these reviews did not include trials with pregnant Indigenous women. A review of smoking cessation interventions specifically for pregnant Indigenous women identified only two relevant trials, neither of which increased cessation, highlighting the need for further research to identify effective strategies [<xref ref-type="bibr" rid="CR8">8</xref>]. In addition to considering approaches found to work in other pregnant population groups, a useful starting point for developing interventions is an exploration of the views of pregnant Indigenous women, and the staff providing their antenatal care.</p></sec><sec id="Sec2"><title>Aims</title><p>To assess support for a range of potential smoking cessation program strategies among pregnant Indigenous women who currently smoke tobacco, pregnant ex-smokers, and their antenatal care providers.</p></sec><sec id="Sec3"><title>Methods</title><p>Cross-sectional surveys with antenatal care providers and pregnant Indigenous women were undertaken in the Northern Territory (NT) and New South Wales (NSW). The project was guided by a community reference group (CRG) to ensure cultural security. The CRG was composed of Aboriginal women from the community (some of whom were pregnant), Aboriginal Health Workers (AHWs) and Community Midwives. Ethical approval for the research was provided by the Human Research Ethics Committees of the University of Newcastle, the NT Department of Human Services and Menzies School of Health Research, Hunter New England Health Service and the Aboriginal Health &#x00026; Medical Research Council of NSW.</p><sec id="Sec4"><title>Recruitment</title><p>The detailed methodology for both surveys is described elsewhere [<xref ref-type="bibr" rid="CR12">12</xref>, <xref ref-type="bibr" rid="CR13">13</xref>]. A brief summary follows.</p><sec id="Sec5"><title>Staff Survey</title><p>Briefly, staff providing antenatal care in remote medical services in the NT and through the Aboriginal Maternal and Infant Health Service (AMIHS) in NSW were eligible and were identified by their relevant health departments and services. All staff worked in community based services. Between September 2008 and July 2009, eligible staff were sent invitation letters, information sheets and self-completion questionnaires. They were asked to complete the anonymous questionnaires and return them in pre-paid envelopes. Reminder letters with additional copies of the documents were sent twice&#x02014;3&#x000a0;weeks after the initial invitation and again 1&#x000a0;month later. Return of the questionnaire was considered to imply consent.</p></sec><sec id="Sec6"><title>Women&#x02019;s Survey</title><p>Women were recruited by the AMIHS teams from July to December 2009, and from the maternity outpatient clinic of a major hospital from July to September 2010 and April to June 2011. Women were eligible if pregnant and if they or their partner were Indigenous. They were excluded if aged less than 16; being treated for mental illness; or unable to provide informed consent. Consecutive eligible women were invited to participate by the midwife, AHW or a female Aboriginal research assistant, who explained the study and provided women with information sheets. Written consent was obtained. Recruiting staff offered assistance to complete the questionnaire if required. Staff were asked to invite all eligible women to participate and to complete a recruitment log to track participation rates.</p></sec></sec><sec id="Sec7"><title>Questionnaire Development and Contents</title><p>Draft questionnaires were critically reviewed by the CRG and colleagues experienced in Indigenous health research and smoking cessation, to assess content validity, reduce redundancy and refine the wording to ensure cultural appropriateness. Minor revisions were made prior to pilot-testing with 12 antenatal service providers, and 15 pregnant Indigenous women, in NSW and Western Australia. Further minor modifications were made in consultation with the CRG.</p><p>The final questionnaires had Flesch-Kincaid reading levels of grade 9 (staff) and grade 6 (women) and both took 15&#x02013;20&#x000a0;min to complete. The questionnaires for staff and women differed with regards to some content, but of relevance to this paper, both included a question on strategies. For staff, the wording was &#x0201c;Please indicate how useful you think each of the following would be in helping pregnant women quit smoking&#x0201d;. They were then presented with a list of 12 possible strategies, and asked to indicate if they considered them to be &#x02018;very helpful&#x02019;, &#x02018;somewhat helpful&#x02019;, &#x02018;maybe helpful&#x02019;, &#x02018;not helpful&#x02019; or &#x02018;harmful&#x02019;. The women were asked &#x0201c;How useful do you think each of the following would be in helping pregnant women to quit smoking&#x0201d;, with the same list of strategies and response options. Additionally, both the staff and women&#x02019;s questionnaires included a question on smoking status&#x02014;current daily smoker, current occasional smoker, ex-smoker or never smoked. The women&#x02019;s questionnaire also asked the usual number of cigarettes smoked each day; and their age, education, and parity.</p></sec><sec id="Sec8"><title>Statistical Analysis</title><p>Responses to the question on smoking status were categorised into current smokers (current daily or occasional smokers), ex-smokers or never smokers. Responses to the questions on the helpfulness of the strategies were dichotomised into &#x02018;very or somewhat helpful&#x02019; or &#x02018;other&#x02019;. For the women&#x02019;s survey, only responses from smokers and ex-smokers were included in the analysis.</p><p>Summary statistics of respondent characteristics were obtained. For the women, mean age and number of cigarettes smoked were calculated. Years of school, and parity were categorised, and the number and percentage in each category reported. For the staff, the number and percentage for each profession was calculated.</p><p>The proportion of each group (women who were current smokers, ex-smokers and service providers) who considered each strategy &#x02018;very or somewhat helpful&#x02019; was calculated and 95&#x000a0;% confidence intervals generated. We also assessed the proportions in each group indicating that each strategy &#x02018;maybe helpful&#x02019;, was &#x02018;not helpful&#x02019; or was &#x02018;harmful&#x02019;.</p></sec></sec><sec id="Sec9"><title>Results</title><sec id="Sec10"><title>Description of the Sample</title><p>In total, 264 women responded to the survey, of whom 121 were current smokers and 55 were ex-smokers and included in this analysis. The response rate could not be calculated as not all teams returned recruitment logs, but among the teams which did, the response rate was 88&#x000a0;%. The majority of smokers (85, 70&#x000a0;%) reported smoking every day with the remaining 36 (30&#x000a0;%) smoking occasionally. The smokers reported an average of 10 cigarettes per day. Other characteristics of the current smokers and ex-smokers are presented in Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>.<table-wrap id="Tab1"><label>Table&#x000a0;1</label><caption><p>Characteristics of women who were current smokers (n&#x000a0;=&#x000a0;121) and ex-smokers (n&#x000a0;=&#x000a0;55)</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left"/><th align="left">Current smokers n&#x000a0;=&#x000a0;121</th><th align="left">Ex-smokers n&#x000a0;=&#x000a0;55</th></tr></thead><tbody><tr><td align="left">Age: mean (standard deviation)</td><td char="(" align="char">24.9 (5.69)</td><td char="(" align="char">24.4 (6.02)</td></tr><tr><td align="left">Completed year 12 at school: n (%)</td><td char="(" align="char">14 (12)</td><td char="(" align="char">10 (18)</td></tr><tr><td align="left">Post-secondary education: n (%)</td><td char="(" align="char">40 (33)</td><td char="(" align="char">32 (58)</td></tr><tr><td align="left">Primiparous: n (%)</td><td char="(" align="char">29 (24)</td><td char="(" align="char">21 (38)</td></tr></tbody></table></table-wrap> </p><p>127 of 184 (69&#x000a0;%) eligible service providers responded, of whom 30 (24&#x000a0;%) were AHWs, 89 (70&#x000a0;%) were nurses or midwives, and eight (5&#x000a0;%) were doctors. Nineteen (15&#x000a0;%) reported being current smokers [10 AHWs (33&#x000a0;%) and nine midwives (10&#x000a0;%)].</p></sec><sec id="Sec11"><title>Perceived Helpfulness of Suggested Strategies</title><p>The numbers of participants indicating that they thought each strategy would be very or somewhat helpful for pregnant women in quitting smoking are shown in Table&#x000a0;<xref rid="Tab2" ref-type="table">2</xref> and are presented in order of the proportion of current smokers indicating they thought the strategy would be helpful. Overall, a greater proportion of service providers were likely to consider each of the strategies helpful than the current smokers, with the ex-smokers generally between the providers and the current smokers. Four of the six strategies rated most highly by smokers (family support, advice and support from the midwife, doctor or AHW) were also in the top five supported strategies for ex-smokers and the top four for providers. Interestingly, rewards were the most popular strategy among ex-smokers (83&#x000a0;%) and the 2nd most popular with current smokers (63&#x000a0;%) but equal 10th among providers (56&#x000a0;%). Community activities were less supported by ex-smokers (51&#x000a0;%) than by either current smokers (59&#x000a0;%) or providers (74&#x000a0;%). Access to Quitline was supported by less than 50&#x000a0;% of respondents in all three groups. For each strategy, respondents who did not consider it likely to be helpful were split fairly evenly between &#x02018;maybe helpful&#x02019; and &#x02018;not helpful&#x02019;. The only strategies considered harmful by more than one person in any group were free NRT which was considered harmful by eight providers, five current smokers and one ex-smoker; and rewards for quitting which were considered harmful by six providers, one current smoker and one ex-smoker (not shown in table).<table-wrap id="Tab2"><label>Table&#x000a0;2</label><caption><p>Proportion of respondents considering each strategy very or somewhat helpful among antenatal service providers, pregnant women who smoke and pregnant ex-smokers</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="4">Strategy^a,b </th><th align="left" colspan="6">Very or somewhat helpful</th></tr><tr><th align="left" colspan="6">Women</th></tr><tr><th align="left" colspan="2">Current smokers N&#x000a0;=&#x000a0;121</th><th align="left" colspan="2">Ex-smokers N&#x000a0;=&#x000a0;55</th><th align="left" colspan="2">Service providers N&#x000a0;=&#x000a0;127</th></tr><tr><th align="left">n</th><th align="left">% (95&#x000a0;% CI)</th><th align="left">n</th><th align="left">% (95&#x000a0;% CI)</th><th align="left">n</th><th align="left">% (95&#x000a0;% CI)</th></tr></thead><tbody><tr><td align="left">Support for the whole family to help others quit</td><td char="." align="char">74</td><td char="(" align="char">64 (54, 73)</td><td char="." align="char">40</td><td char="(" align="char">74 (60, 85)</td><td char="." align="char">116</td><td char="(" align="char">91 (85, 96)</td></tr><tr><td align="left">Rewards for women who stop smoking with vouchers to get things for the mother or baby</td><td char="." align="char">74</td><td char="(" align="char">63 (54, 72)</td><td char="." align="char">43</td><td char="(" align="char">83 (70, 92)</td><td char="." align="char">70</td><td char="(" align="char">56 (46, 64)</td></tr><tr><td align="left">Advice and support from the midwife</td><td char="." align="char">74</td><td char="(" align="char">62 (53, 71)</td><td char="." align="char">39</td><td char="(" align="char">74 (60, 85)</td><td char="." align="char">108</td><td char="(" align="char">86 (78, 91)</td></tr><tr><td align="left">Advice and support from the doctor</td><td char="." align="char">72</td><td char="(" align="char">61 (52, 70)</td><td char="." align="char">41</td><td char="(" align="char">76 (62, 87)</td><td char="." align="char">108</td><td char="(" align="char">85 (78, 91)</td></tr><tr><td align="left">Community activities about quitting</td><td char="." align="char">68</td><td char="(" align="char">59 (49, 68)</td><td char="." align="char">27</td><td char="(" align="char">51 (37, 65)</td><td char="." align="char">93</td><td char="(" align="char">74 (66, 82)</td></tr><tr><td align="left">Advice and support from the AHW</td><td char="." align="char">66</td><td char="(" align="char">56 (47, 66)</td><td char="." align="char">41</td><td char="(" align="char">76 (62, 87)</td><td char="." align="char">114</td><td char="(" align="char">90 (83, 94)</td></tr><tr><td align="left">Free nicotine replacement therapy</td><td char="." align="char">66</td><td char="(" align="char">56 (47, 66)</td><td char="." align="char">33</td><td char="(" align="char">62 (48, 75)</td><td char="." align="char">92</td><td char="(" align="char">74 (66, 82)</td></tr><tr><td align="left">Peer support groups</td><td char="." align="char">60</td><td char="(" align="char">53 (45, 62)</td><td char="." align="char">40</td><td char="(" align="char">74 (60, 85)</td><td char="." align="char">102</td><td char="(" align="char">81 (73, 87)</td></tr><tr><td align="left">Brochures: harms of smoking and advice on quitting</td><td char="." align="char">61</td><td char="(" align="char">52 (43, 61)</td><td char="." align="char">27</td><td char="(" align="char">50 (36, 64)</td><td char="." align="char">71</td><td char="(" align="char">56 (47, 65)</td></tr><tr><td align="left">Stress management programs</td><td char="." align="char">57</td><td char="(" align="char">49 (39, 58)</td><td char="." align="char">38</td><td char="(" align="char">70 (56, 82)</td><td char="." align="char">92</td><td char="(" align="char">73 (64, 81)</td></tr><tr><td align="left">Support person</td><td char="." align="char">55</td><td char="(" align="char">47 (38, 57)</td><td char="." align="char">37</td><td char="(" align="char">69 (54, 80)</td><td char="." align="char">79</td><td char="(" align="char">63 (54, 71)</td></tr><tr><td align="left">Access to a Quitline</td><td char="." align="char">54</td><td char="(" align="char">46 (37, 56)</td><td char="." align="char">26</td><td char="(" align="char">49 (35, 63)</td><td char="." align="char">61</td><td char="(" align="char">49 (40, 58)</td></tr></tbody></table><table-wrap-foot><p> ^aOrdered by proportion of current smokers perceiving strategies to be very or somewhat helpful</p><p> ^b0&#x02013;5 missing responses for each variable</p></table-wrap-foot></table-wrap> </p></sec></sec><sec id="Sec12"><title>Discussion</title><p>This paper is the first we are aware of that explores the degree to which pregnant Indigenous women and antenatal care providers consider particular strategies helpful for antenatal smoking cessation. In general, current smokers were least supportive of most strategies, and providers were most supportive. The majority of strategies were supported by over half the participants in each group. The reasons for the lower support among current smokers than among ex-smokers and providers on most strategies is not known, but may reflect their personal struggles with quitting and recognition of the difficulty of quitting or a general sense of hopelessness regarding the prospects of success. While these results reflect the opinions of respondents, not the actual efficacy of strategies, establishing acceptability is a useful starting point for developing intervention trials.</p><sec id="Sec13"><title>Rewards for Smoking Cessation</title><p>A similar proportion of current smokers and providers considered rewards likely to be helpful (63.3 and 55.6&#x000a0;% respectively) but a higher proportion of ex-smokers indicated they thought rewards would be helpful (83&#x000a0;%), with rewards the most popular strategy in this group, 2nd most popular among current smokers and equal 10th among providers.</p><p>The Cochrane review of antenatal smoking cessation interventions identified provision of incentives, or rewards, as the most effective intervention, with incentives reducing smoking by 24&#x000a0;% compared to 6&#x000a0;% for all interventions combined [<xref ref-type="bibr" rid="CR5">5</xref>]. Incentives are considered most effective for simple, time-limited behaviours such as completing immunisation, but may be less effective where the required behaviour change is complex [<xref ref-type="bibr" rid="CR14">14</xref>]. For maintaining complex behaviour change, financial incentives may be a useful addition to multi-faceted programs that address the complex individual, social and economic factors affecting behaviour [<xref ref-type="bibr" rid="CR14">14</xref>].</p><p>Incentives for antenatal smoking cessation are already used in some parts of the British NHS [<xref ref-type="bibr" rid="CR15">15</xref>, <xref ref-type="bibr" rid="CR16">16</xref>], yet their use for health behaviour change remains controversial [<xref ref-type="bibr" rid="CR15">15</xref>]. In a survey of pregnant Australian women, the majority did not support paying pregnant smokers to quit, but smokers were more likely to do so [<xref ref-type="bibr" rid="CR17">17</xref>]. A qualitative study with social service staff and clients found that clients were supportive of rewards for quitting while staff were less so, and expressed concerns about the feasibility of implementation [<xref ref-type="bibr" rid="CR18">18</xref>]. Our results add to this body of work identifying significantly greater support for rewards among ex-smokers than among providers. Given the apparent efficacy of incentives in antenatal smoking cessation, further research is required to explore the reasons for the low support among providers relative to their support for other strategies.</p></sec><sec id="Sec14"><title>Involving Family</title><p>The strategy rated highest by both current smokers and providers was &#x0201c;support for the whole family to help others quit&#x0201d;. Smokers who are supported by their partners are more likely to succeed, but a recent systematic review of interventions aimed at enhancing partner support to improve smoking cessation found little evidence for effective interventions [<xref ref-type="bibr" rid="CR19">19</xref>]. Family based interventions have been recommended for Indigenous Australians because of the importance of family in influencing smoking behaviour [<xref ref-type="bibr" rid="CR20">20</xref>, <xref ref-type="bibr" rid="CR21">21</xref>]. The endorsement by women and service providers in our study provides additional evidence for their acceptability and further support for their inclusion in future trials to assess their efficacy.</p></sec><sec id="Sec15"><title>Health Professionals</title><p>Advice and support from the range of health professionals were each rated reasonably highly by all groups. Good evidence exists for efficacy of advice from doctors and nurses [<xref ref-type="bibr" rid="CR22">22</xref>, <xref ref-type="bibr" rid="CR23">23</xref>], however midwives, including midwives caring for Indigenous women, have expressed reluctance to address smoking, concerned that they may damage their relationship with their clients [<xref ref-type="bibr" rid="CR12">12</xref>, <xref ref-type="bibr" rid="CR24">24</xref>]. Similar concerns have been expressed by AHWs, with the additional concern that AHW smoking may impede providing advice [<xref ref-type="bibr" rid="CR25">25</xref>]. However, over half the women in our study indicated that support from each of the health professionals was likely to be helpful, suggesting this approach is acceptable, perceived to be effective and may be a fruitful approach.</p></sec><sec id="Sec16"><title>Other Strategies</title><p>Community activities were rated fifth and sixth by current smokers and providers respectively but 10th by ex-smokers. The reasons for the lower support among ex-smokers are not known. Previous studies have emphasised the preference of Indigenous Australians for programs to be community-based [<xref ref-type="bibr" rid="CR21">21</xref>]. Although community interventions increase knowledge of risks, change attitudes to smoking and increase quit attempts, they have not been shown to reduce the prevalence of smoking [<xref ref-type="bibr" rid="CR26">26</xref>, <xref ref-type="bibr" rid="CR27">27</xref>].</p><p>Other activities considered helpful by at least half of each group included free NRT, support groups and brochures. NRT is efficacious in non-pregnant populations, but evidence for its effectiveness in pregnancy is inconclusive [<xref ref-type="bibr" rid="CR28">28</xref>]. Pregnant Indigenous women have previously been found to have relatively low levels of nicotine dependency [<xref ref-type="bibr" rid="CR29">29</xref>], which may contribute to the lower rating for NRT in our study. Current guidelines state that NRT should be considered if a pregnant woman is otherwise unable to quit [<xref ref-type="bibr" rid="CR30">30</xref>], and it would therefore be reasonable to include free NRT as a component of future cessation trials. In non-pregnant populations, group programs are more effective than self-help and other low intensity interventions, but the limited research in this area has not provided an adequate evidence base to determine whether they are more effective than intensive individual counselling, or whether they provide additional benefit as an adjunct to individual support [<xref ref-type="bibr" rid="CR31">31</xref>]. Although generally supported by respondents in each group, the logistic challenges of running groups, particularly in rural areas, would need to be overcome if they were to be included in future smoking cessation trials. Low intensity interventions, including providing verbal or written advice, demonstrated a small benefit in the Cochrane review on antenatal smoking cessation [<xref ref-type="bibr" rid="CR5">5</xref>]. While unlikely to have a large impact, culturally appropriate brochures and other resources may be a useful prop to use when discussing smoking cessation.</p><p>Interestingly, less than half the current smokers thought that stress management programs would be helpful. Research on smoking among Indigenous Australians has emphasised stress as an impediment to cessation [<xref ref-type="bibr" rid="CR21">21</xref>, <xref ref-type="bibr" rid="CR32">32</xref>]. Although stress contributes to pregnant women failing to quit, and stress management techniques are included in some cessation programs, the evidence on their benefit is inconclusive [<xref ref-type="bibr" rid="CR33">33</xref>].</p></sec><sec id="Sec17"><title>Limitations</title><p>A number of limitations need to be considered in interpreting the results from this study. The response rate was higher among the women than the service providers. The reasons for this difference are unknown, but it may be due to differences in recruitment, with providers recruited by letter, and women recruited through a personal approach. Secondly, the sample is fairly small, despite the reasonably good response rates. However, Indigenous women are a small proportion of the population, and engaging them in research can be challenging. One of the strengths of this study is that it includes women from across two different states, and they are representative of pregnant Indigenous women nationally with regard to age and parity [<xref ref-type="bibr" rid="CR34">34</xref>]. Thirdly, a delay between the providers&#x02019; and the women&#x02019;s surveys may have impacted on the results. However, we are unaware of any specific programs or initiatives which occurred between the two surveys that could be considered to impact on the findings. A fourth limitation is that the data are drawn from cross-sectional surveys, with no opportunity to explain the proposed strategies in more detail, nor to explore the reasons for support or opposition to the strategies. More importantly, the apparent support may not translate into implementation or uptake, nor into actual changes in smoking behaviour. Further intervention research is required to explore the feasibility of implementing these strategies in real world settings, their uptake by pregnant women and their actual impact on smoking rates and health outcomes.</p></sec></sec><sec id="Sec18"><title>Conclusions</title><p>Exploring the views of stakeholders involved in antenatal smoking cessation&#x02014;the providers and the pregnant women, has identified the strategies which are most acceptable, and thus the ones most likely to be implemented if introduced in routine care. These strategies, if known to be effective in other pregnant populations, should be included in interventions and tested in trials to assess their real world uptake and their impact on smoking behaviours and health outcomes. Given the apparent efficacy of rewards in other population groups, further research is required to assess their efficacy among pregnant Indigenous women and to identify reasons for their lower support by providers.</p></sec>

Interleukin-2-Inducible T-Cell Kinase (ITK) Deficiency - Clinical and Molecular Aspects

Could not extract abstract

<contrib contrib-type="author"><name><surname>Ghosh</surname><given-names>Sujal</given-names></name><xref ref-type="aff" rid="Aff1">1</xref><xref ref-type="aff" rid="Aff2">2</xref></contrib><contrib contrib-type="author"><name><surname>Bienemann</surname><given-names>Kirsten</given-names></name><xref ref-type="aff" rid="Aff1">1</xref></contrib><contrib contrib-type="author"><name><surname>Boztug</surname><given-names>Kaan</given-names></name><xref ref-type="aff" rid="Aff3">3</xref></contrib><contrib contrib-type="author" corresp="yes"><name><surname>Borkhardt</surname><given-names>Arndt</given-names></name><address><phone>+49-211 811-7680</phone><email>arndt.borkhardt@med.uni-duesseldorf.de</email></address><xref ref-type="aff" rid="Aff1">1</xref></contrib><aff id="Aff1"><label>1</label>Department of Pediatric Oncology, Hematology and Clinical Immunology, Medical Faculty, Center of Child and Adolescent Health, Heinrich-Heine-University, Moorenstra&#x000df;e 5, 40225 Duesseldorf, Germany </aff><aff id="Aff2"><label>2</label>Institute of Child Health, Molecular and Cellular Immunology Section, UCL Institute of Child Health, 30 Guilford Street, London, WC1N 1EH UK </aff><aff id="Aff3"><label>3</label>CeMM Research Center for Molecular Medicine of the Austrian Academy of Sciences, Department of Paediatrics and Adolescent Medicine, Medical University of Vienna, Vienna, 1090 Vienna Austria </aff>

Journal of Clinical Immunology

<sec id="Sec1" sec-type="introduction"><title>Background</title><p id="Par2">Epstein-Barr virus (EBV), being one of the most common viruses in humans, is best known to cause infectious mononucleosis. In patients with underlying immunodeficiency, EBV may lead to severe immune dysregulation manifesting as fatal mononucleosis, Hodgkin and Non-Hodgkin lymphoma, lymphoproliferative disease (LPD), lymphomatoid granulomatosis, hemophagocytic lymphohistiocytosis (HLH) and dysgammaglobulinemia [<xref ref-type="bibr" rid="CR1">1</xref>]. Patients, who harbor alterations in genes coding for proteins of the lymphocytic cytotoxic pathway, T cell signaling or T-B cell interaction, may present with EBV associated disease [<xref ref-type="bibr" rid="CR2">2</xref>]. Mutations in genes of the cytotoxic pathway (e.g. <italic>PRF1)</italic> can lead to hemophagocytic lymphohistiocytosis (HLH) after an infectious trigger, terming these diseases collectively as hemophagocytic syndromes [<xref ref-type="bibr" rid="CR3">3</xref>, <xref ref-type="bibr" rid="CR4">4</xref>]. Furthermore the quite heterogeneous group of combined immunodeficiencies (leaky or atypical SCID) with defects affecting antigen receptor recombination (e.g. hypomorphic mutations in <italic>RAG1/2</italic>) and some distinct genes of T cell signaling (e.g. in <italic>SH2D1A, XIAP</italic>) can lead to EBV associated disease due to immune dysregulation, notably lymphoproliferation [<xref ref-type="bibr" rid="CR5">5</xref>, <xref ref-type="bibr" rid="CR6">6</xref>]. However, EBV pathophysiology varies in each of these diseases.</p><p id="Par3">Several newly discovered primary immunodeficiencies with EBV lymphoproliferation (STK4, CD27, MAGT1, Coronin-1A deficiency) have been described in the last few years [<xref ref-type="bibr" rid="CR7">7</xref>&#x02013;<xref ref-type="bibr" rid="CR11">11</xref>]; our group and collaborators were able to reveal the pathogenicity of mutations in the Interleukin 2-inducible T-cell kinase (<italic>ITK</italic>) gene&#x000a0;in a cohort of eight patients all presenting with massive EBV B-cell lymphoproliferation [<xref ref-type="bibr" rid="CR12">12</xref>&#x02013;<xref ref-type="bibr" rid="CR15">15</xref>]. One further patient with recurrent infections and CD4 lymphopenia, but without lymphoproliferation was just recently discovered [<xref ref-type="bibr" rid="CR16">16</xref>]. Hence, we will briefly summarize the clinical and immunological findings in patients suffering from ITK deficiency and draw comparisons with the extensively investigated function of ITK in vitro and in the murine model.</p></sec><sec id="Sec2"><title>Clinical Presentation</title><p id="Par4">In 2009, we described two sisters of consanguineous Turkish descent with EBV lymphoproliferation [<xref ref-type="bibr" rid="CR12">12</xref>]. The older sister (at the age of 6&#x000a0;years) presented with severe candida stomatitis, pneumocystis jirovici pneumonia, pleural and pericardial effusion, hepatosplenomegaly, cytopenia and progressive hypogammaglobulinemia. A biopsy of an axillary lymph node revealed oligoclonal polymorphic B cell lymphoproliferation of type II latency. Despite repeated courses of Rituximab and partial resolution of clinical symptoms 1 &#x000bd;&#x02009;year later the girl developed Hodgkin&#x02019;s lymphoma. Successfully treated according to standard chemotherapy protocol, T cells unfortunately were continuously declining and the girl died from respiratory failure after acquiring pneumocystis pneumonia at the age of 10 &#x000bd;&#x02009;years. The younger sister developed pancytopenia, hepatosplenomegaly, abdominal lymphadenopathy, ascites and pleural effusions due to impaired liver function. EBV associated Hodgkin lymphoma was detected in an inguinal lymph node biopsy; as the clinical situation deteriorated remission could not be achieved and she received haploidentical peripheral blood stem cell transplantation (SCT); unfortunately she died of ischemic brain injury following airway obstruction and cardiac arrest in aplasia. We performed genome-wide linkage analysis with eight family members identifying a region at 5q31-34 that segregated with the disease. Subsequent analysis of candidate genes led to the causative homozygous R335W mutation in the SH2 domain of ITK. Till now we have gathered information on further six patients presenting primarily with massive EBV B-cell lymphoproliferation further progressing to full malignant Hodgkin lymphoma in some cases: three patients were within a family of Palestinian origin [<xref ref-type="bibr" rid="CR15">15</xref>]; and a total of three single individuals from Iran [<xref ref-type="bibr" rid="CR14">14</xref>], India and Morocco. Patients with lymphoproliferative disease / Hodgkin lymphoma manifested between 3 and 13&#x000a0;years of age with fever, lymphadenopathy, hepatosplenomegaly and EBV viremia; additional viral infections in some patients included CMV and severe varicella infections indicating a general T cell deficiency. Given severe immune dysregulation, three patients developed autoimmune phenomena (cytopenia, nephritis and thyroiditis); two patients developed HLH. Pulmonary involvement with large interstitial nodules was observed in the majority of patients, as it is known from other EBV-related pathologies; remarkably, extensive pulmonary infiltration of EBV-LPD leading to respiratory distress was the only major primary manifestation in one patient [<xref ref-type="bibr" rid="CR14">14</xref>]. A ninth ITK deficient patient has just recently been discovered. Interestingly the 18&#x000a0;year-old male Turkish patient of consanguineous background suffered from recurrent progressive pulmonary infections, but no lymphoproliferation so far. Please refer to Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref> for further details.<table-wrap id="Tab1"><label>Table 1</label><caption><p>Clinical and laboratory findings in <italic>ITK</italic> deficient patients. abbreviations: AIHA autoimmune hemolytic anemia, Cx chemotherapy, HL Hodgkin Lymphoma, ITP immune thrombocytopenia, LBCL large B-cell lymphoma, LG lymphomatoid granulomatosis, LPD lymphoproliferative disorder, n.d. not determined, n.q. not quantified. &#x02193; decreased &#x02198; lower margin &#x02191; increased</p></caption><table frame="hsides" rules="groups"><thead><tr><th>origin/mutation</th><th>Patient 1 Turkey c.1003C&#x02009;&#x0003e;&#x02009;T: p.R335W</th><th>Patient 2 Turkey c.1003C&#x02009;&#x0003e;&#x02009;T: p.R335W</th><th>Patient 3 Palestine c.1764C&#x02009;&#x0003e;&#x02009;G: p.Y588X</th><th>Patient 4 Palestine c.1764C&#x02009;&#x0003e;&#x02009;G: p.Y588X</th><th>Patient 5 Palestine c.1764C&#x02009;&#x0003e;&#x02009;G: p.Y588X</th><th>Patient 6 Morocco c.86G&#x02009;&#x0003e;&#x02009;A: p.R29H</th><th>Patient 7 India c.1497delT: p.S499SfsX4</th><th>Patient 8 Iran c.468delT: p.P156PfsX109</th><th>Patient 9 Turkey c.49C&#x02009;&#x0003e;&#x02009;T: p.Q17X</th></tr></thead><tbody><tr><td>sex</td><td>female</td><td>female</td><td>female</td><td>male</td><td>male</td><td>male</td><td>female</td><td>female</td><td>male</td></tr><tr><td>age at diagnosis</td><td>5</td><td>6</td><td>4</td><td>5</td><td>3</td><td>11</td><td>6</td><td>13</td><td>18</td></tr><tr><td>current status</td><td>died at age 10</td><td>died at age 7</td><td>died at age 6</td><td>remission after Cx, age 12</td><td>well after HSCT, age 8</td><td>died at age 26</td><td>died after HSCT at age 8</td><td>died at age 15</td><td/></tr><tr><td>fever</td><td>+</td><td>+</td><td>+</td><td>+</td><td>+</td><td>+</td><td>+</td><td>+</td><td>+</td></tr><tr><td>lymphadenopathy</td><td>+</td><td>+</td><td>+</td><td>+</td><td>+</td><td>+</td><td>+</td><td>+</td><td>none</td></tr><tr><td>hepatosplenomegaly</td><td>+</td><td>+</td><td>+</td><td>none</td><td>+</td><td>unknown</td><td>none</td><td>+</td><td>none</td></tr><tr><td>pulmonary involvement</td><td>+</td><td>none</td><td>none</td><td>+</td><td>+</td><td>+</td><td>+</td><td>+</td><td>infections</td></tr><tr><td>histology</td><td>B cell LPD Hodgkin</td><td>HL-like B cell LPD</td><td>HL</td><td>HL</td><td>HL</td><td>B cell LPD</td><td>B cell LPD, LBCL, LG</td><td>B cell LPD</td><td>none</td></tr><tr><td>autoimmunity</td><td>none</td><td>none</td><td>none</td><td>nephritis, thyroiditis</td><td>thyroiditis</td><td>AIHA / ITP</td><td>none</td><td>none</td><td>none</td></tr><tr><td>HLH</td><td>none</td><td>(+)</td><td>+ (at relapse)</td><td>none</td><td>none</td><td>none</td><td>none</td><td>none</td><td>none</td></tr><tr><td>CD4+ cells</td><td>&#x02193;</td><td>&#x02193;</td><td>normal</td><td>&#x02193;</td><td>normal</td><td>&#x02193;</td><td>&#x02193;</td><td/><td>&#x02193;</td></tr><tr><td>CD8+ cells</td><td>normal</td><td>&#x02193;</td><td>&#x02191;</td><td>&#x02193;</td><td>normal</td><td>normal</td><td>normal</td><td/><td>normal</td></tr><tr><td>NKT cells</td><td>n.d.</td><td>&#x02198;</td><td>n.d</td><td>&#x02198;</td><td>n.d.</td><td>&#x02198;</td><td>&#x02198;</td><td/><td>&#x02198;</td></tr><tr><td>serology</td><td>VCA-G +, VCA-M -, EA-G +, EBNA-G -</td><td>VCA-G +, VCA-M&#x02013;</td><td>VCA-G -, VCA-M -, EBNA +</td><td>VCA-G +, VCA-M -, EBNA-G -</td><td>n.d.</td><td>VCA-G +</td><td/><td/><td>negative</td></tr><tr><td>viral load at presentation</td><td>+ (n.q.)</td><td>10^3</td><td>10^5</td><td>10^3</td><td>10^5</td><td>+ (n.q.)</td><td>10^3</td><td/><td>10^3</td></tr><tr><td>peak viral load</td><td>10^7</td><td>10^4</td><td>unknown</td><td>unknown</td><td>unknown</td><td>10^6</td><td>10^4</td><td/><td>10^3</td></tr></tbody></table></table-wrap> </p></sec><sec id="Sec3"><title>Immune Phenotype, Viral Burden and Lymphoproliferation</title><p id="Par5">Common immunological features in ITK deficient patients are progressive hypogammaglobulinemia and beside global lymphopenia, a progressive loss of CD4+ T cells. Notably, we observe a declining proportion of naive CD45RA+CD4+ T cells. After gating on CD3+ T cells, NKT cells are determined as TCR Vbeta11 and TCR Valpha24 double-positive cells [<xref ref-type="bibr" rid="CR17">17</xref>]. As seen in other syndromes susceptible to EBV-LPD (e.g. patients with SAP, XIAP, CD27 and Coronin-1A deficiency) NKT cells are severely reduced in the peripheral blood of ITK deficient patients [<xref ref-type="bibr" rid="CR18">18</xref>]; hence a critical role has been postulated for NKT cells in the response to EBV infection [<xref ref-type="bibr" rid="CR19">19</xref>]. Patients with EBV-LPD show a rather high EBV viremia though the observed peak viral load in the eight reported patients was quite heterogeneous (10^4 - 10^8 copies/ &#x003bc;g DNA). As the serological phenotype at the time of manifestation differs among the patients, it is only hypothetical to predict the duration between primary infection and clinical immune dysregulation. We did not observe EBV-VCA-IgM in any ITK deficient person; though negative EBNA1-IgG in the first patient could suggest a rather short latency period. No EBV-VCA-IgG seronegative symptomatic EBV-LPD patient has been identified so far (as has been documented in SAP deficiency), leading to the assumption that EBV infection might be essential to cause immune dysregulation in particular LPD, lymphoma and autoimmunity in ITK deficiency. However, the ninth patient lacking EBV-LPD shows a seronegative EBV-VCA status, though revealing a low level of viremia (10^3 copies/ &#x003bc;g DNA). In any case, the early detection and watchful clinical monitoring of ITK deficient patients prior to their first encounter to EBV would be highly desirable.</p></sec><sec id="Sec4"><title>Treatment and Outcome</title><p id="Par6">Patients manifested between 3 and 13&#x000a0;years of age. Six out of eight described patients died between 1 and 15&#x000a0;years after primary manifestation of ITK deficiency, five of them within 2&#x000a0;years after the first symptoms appeared. Given the fact, that treatment of EBV-LPD has been best investigated in post-transplant lymphoproliferative disorder (PTLD) patients with a rather type III latency phenotype [<xref ref-type="bibr" rid="CR20">20</xref>], experience to treat EBV-LPD in primary immunodeficiencies is limited and case-based. Rituximab led to clinical improvement in some ITK deficient patients, while steroids did not show any substantial benefit. However, susceptibility to infections and viremia due to T cell depletion and hypogammaglobulinemia is hardly controlled by antivirals or antibiotics. IgG substitution might be helpful, but however, increasing autoimmunity (nephritis, cytopenia) and Hodgkin lymphoma with the need of chemotherapy and potentially irradiation increase the risk of a fatal outcome. Remarkably, two patients of Palestine origin are alive [<xref ref-type="bibr" rid="CR15">15</xref>]. While one is in remission after chemotherapy in Hodgkin lymphoma, the other one received a matched sibling bone marrow graft. One girl of Indian origin died after hematopoetic stem cell transplantation due to severe graft-versus-host disease, so at this moment it will remain unclear whether the outcome depends on the nature of disease, pathogenicity of individual mutations or other circumstances unless a bigger cohort is investigated. At least each patient should be evaluated for potential HSCT.</p></sec><sec id="Sec5"><title>Interleukin-2-Inducible T-Cell Kinase (ITK)</title><p id="Par7">TEC family kinases comprises five members in mammals namely Bruton agammaglobulinemia tyrosine kinase (BTK), TXK tyrosine kinase (TXK, also RLK), BMX non-receptor tyrosine kinase (BMX, also ETK), tec protein tyrosine kinase (TEC) and Interleukin-2-inducible T-cell kinase (ITK, also LYK) [<xref ref-type="bibr" rid="CR21">21</xref>]. All proteins function as non-receptor protein-tyrosine kinases in the development and signaling in the lymphoid lineage. X-linked (Bruton&#x02019;s) agammaglobulinemia, as one of the very first detected primary immunodeficiencies was described in 1952 [<xref ref-type="bibr" rid="CR22">22</xref>]; the discovery of the corresponding gene happened approximately 40&#x000a0;years later, leading to deep molecular insight into the role of <italic>BTK</italic> in B cell commitment [<xref ref-type="bibr" rid="CR23">23</xref>]; however studies on ITK deficiency in mice and in vitro were conducted long before the first patient was discovered in 2009. Notably, patients with peripheral T cell lymphoma had been found with translocations of <italic>ITK</italic> with the <italic>SYK</italic> gene and subsequent fusion proteins [<xref ref-type="bibr" rid="CR24">24</xref>].</p><p id="Par8">The <italic>ITK</italic> gene encompasses a region of 112&#x000a0;kb and 17 exons on chromosome 5q and encodes 620 amino acids of a 71&#x000a0;kDa protein. Like BTK, it consists of a pleckstrin homology (PH) domain at the N-terminus, a Tec homology (TH) domain; a Src homology 3 (SH3) domain; a Src homology 2 (SH2) domain and a catalytic kinase domain at the C-terminus [<xref ref-type="bibr" rid="CR25">25</xref>]. In the context of adaptive immune response in T and NKT cells antigen presenting cells activate the T cell receptor. Subsequently a series of phosphorylation recruits ITK (which is bound through its pleckstrin homology domain to phosphatidylinositol monophosphates) to the cell membrane, where it is phosphorylated by LCK enhancing autophosphorylation and full activation. ITK itself phosphorylates PLCG1, which subsequent leads to the cleavage of its substrates. Further downstream of this pathway, the endoplasmic reticulum releases calcium in the cytoplasm and the nuclear activator of activated T-cells (NFAT) is translocated into the nucleus for transcriptional activity for further lymphokine production, T cell proliferation and differentiation. Among the six observed pedigrees (nine individuals) two families harbored mutations in the kinase, other three pedigrees had mutations in the SH2 and PH domain. In one patient a deletion led to a truncated SH2 and deleted kinase domain. Interestingly, in <italic>BTK</italic>, homologous mutations have been reported to cause X-linked agammaglobulinemia and the observed mutations in our eight patients have mutations in corresponding <italic>BTK</italic> residues indicating similar structural alterations [<xref ref-type="bibr" rid="CR13">13</xref>].</p><p id="Par9">Our group did comprehensive studies in transformed Herpesvirus samiri (HVS) cell lines to characterize these mutations and analyze functional consequences. Most mutations did not dramatically change mRNA levels of the <italic>ITK</italic> gene, however taking protein instability into mind, immunoblot analyses of endogenous ITK protein showed several protein variants. Only one patient showed abundance of mRNA levels likely due to nonsense-mediated mRNA decay after a premature stop codon in exon 14 in the kinase region. Protein half-life was determined by pulse-chase studies and showed significant reduced values compared to wild-type ITK. Calcium response was assayed by flux studies in transformed HVS patient T lymphocytes after CD3 antibody dependent TCR stimulation showing clearly reduced or nearly absent cytosolic release of calcium ions in most patients. TCR-mediated calcium mobilization was restored in murine <italic>Itk</italic> &#x02212;/&#x02212; thymocytes after transduction of a wild type ITK construct [<xref ref-type="bibr" rid="CR13">13</xref>].</p></sec><sec id="Sec6"><title><italic>Itk</italic> -/- Murine Phenotype</title><p id="Par10">It is not elaborated how impaired ITK function changes T cell development. Most studies have been undertaken in the <italic>Itk</italic> -/- mouse model with a BALB/c or C57BL/6 background. While, there are even differences in T cell response among these two models it is ambitious to draw a genuine conclusion from mice to human.</p><p id="Par11">In <italic>Itk</italic> -/- mice thymocyte development gives rise to an increased population of innate single positive CD8+ (CD8SP) thymocytes [<xref ref-type="bibr" rid="CR26">26</xref>]. <italic>Itk</italic> -/- CD8SP thymocytes resemble antigen-experienced T cells showing a CD122&#x02009;+&#x02009;CD44hiCXCR3+ phenotype with high levels of the transcription factor Eomesodermin (Eomes) and IFNy production upon stimulation. Furthermore splenocytes revealed less CD4 and CD8 expression with a rather mature effector phenotype (CD44&#x02009;+&#x02009;CD62L+), also showing differential transcriptional signatures with increased levels of Eomes and Tbet as also observed in peripheral CD8 cells of ITK deficient patients [<xref ref-type="bibr" rid="CR27">27</xref>&#x02013;<xref ref-type="bibr" rid="CR31">31</xref>]. As observed in humans, NKT cell development and function is impaired and peripheral survival is reduced [<xref ref-type="bibr" rid="CR32">32</xref>].</p></sec><sec id="Sec7"><title>TH1 Skewing</title><p id="Par12">The current model (in mice and human) of naive CD4 cell differentiation encompasses further maturation to Th1, Th2, Th17, Treg and Tfh cells, not including further more subsets to follow. However, most in vitro and in vivo murine studies addressing the molecular defects of <italic>Itk</italic> &#x02212;/&#x02212; CD4 T cells did focus on the Th1/Th2 paradigm giving some evidence that ITK might be responsible for a proper Th2 response [<xref ref-type="bibr" rid="CR25">25</xref>]. Following TCR stimulation ITK deficient T cells show decreased proliferation and effector cytokine production. In addition to reduced intracellular calcium release, which is shown in human as in mice, an altered NFAT nuclear translocation gives evidence of the biochemical disturbance [<xref ref-type="bibr" rid="CR33">33</xref>&#x02013;<xref ref-type="bibr" rid="CR35">35</xref>]. <italic>Itk</italic> &#x02212;/&#x02212; mice have decreased CD4 T cell numbers in the thymus and periphery [<xref ref-type="bibr" rid="CR26">26</xref>]. Though, ITK seems to be dispensable for general TCR signaling (otherwise the clinical presentation would be more severe), several in-vivo studies of parasite infection in <italic>Itk</italic> &#x02212;/&#x02212; mice indicate a role for ITK in Th2 response&#x000a0;(see Table <xref rid="Tab2" ref-type="table">2</xref>).<table-wrap id="Tab2"><label>Table 2</label><caption><p>published in-vivo studies of <italic>Itk</italic> &#x02212;/&#x02212; mouse models challenged with parasites</p></caption><table frame="hsides" rules="groups"><thead><tr><th>study</th><th>Fowell et al., 1999 [<xref ref-type="bibr" rid="CR33">33</xref>]</th><th>Fowell et al., 1999 [<xref ref-type="bibr" rid="CR33">33</xref>]</th><th>Schaeffer et al., 1999 [<xref ref-type="bibr" rid="CR37">37</xref>]</th><th>Schaeffer et al., 2001 [<xref ref-type="bibr" rid="CR36">36</xref>]</th></tr></thead><tbody><tr><td> <italic>Itk</italic>&#x02212;/&#x02212; background</td><td>BALB/c</td><td>BALB/c</td><td>C57BL6/J</td><td>C57BL6/J</td></tr><tr><td>parasite</td><td> <italic>Leishmania major</italic> </td><td> <italic>Nippostrongylus brasiliensis</italic> </td><td> <italic>Toxoplasma gondii</italic> </td><td> <italic>Schistosoma mansoni</italic> </td></tr><tr><td>induced immune response in wt</td><td>Th1</td><td>Th2</td><td>Th1</td><td>Th2</td></tr><tr><td>effects in <italic>Itk&#x02212;/&#x02212;</italic> vs wt</td><td>control of infection, IFNy &#x02191;, IL-4 &#x02193;</td><td>unable to expel adult worms from the intestines, IL-4 &#x02193;</td><td>identical brain cyst count, near normal IFNy production upon ConA and STAg stimulation <bold>BUT</bold> decreased mean survival time</td><td>poor granulomatous responses: size of granuloma and draining lymph nodes &#x02193;&#x0115f;IL-4, IL-5 and IL-10 &#x02191;, IFNy &#x02193;</td></tr></tbody></table></table-wrap> </p><p id="Par13"> <italic>Itk</italic> &#x02212;/&#x02212; mice showed poor granulomatous responses when challenged with <italic>Schistosoma mansoni</italic> eggs, a Th2 response inducing helminth [<xref ref-type="bibr" rid="CR36">36</xref>]. Beside a marked reduction of size of granuloma and draining lymph nodes, IL-4, IL-5 and IL-10 production was decreased compared to <italic>S. mansoni</italic> infection in the WT. Instead, IFNy production was significantly increased, suggesting ITK deficiency skewed TCR response towards Th1 differentiation. Paradoxically, the same investigations in the <italic>Itk</italic>&#x02212;/&#x02212;<italic>Rlk</italic>&#x02212;/&#x02212; double knock out mouse (expected to be impaired more severely) did not show any weaker response compared to WT. Another agent to induce a powerful Th2 response is the challenge with the nematode <italic>Nippostrongylos brasiliens</italic> [<xref ref-type="bibr" rid="CR33">33</xref>]. 12&#x000a0;days after infection with N. brasiliensis wild type BALB/c mice had cleared the worm infection in the gut; in contrast BALB/c mice were unable to expel adult worms and showed a reduction of IL-4 producing cells. Remarkably, <italic>Itk</italic> &#x02212;/&#x02212; mice show low susceptibility to some intracellular protozoa [<xref ref-type="bibr" rid="CR33">33</xref>]. BALB/c mice fail to clear infection with <italic>Leishmania major</italic>, as the usual Th2 response is incapable to clear the parasites; however <italic>Itk</italic> &#x02212;/&#x02212; mice show an increase of Th1 dependent IFNy response, leading to the control of infection. The role of <italic>Toxoplasma gondii</italic>, which is usually considered to promote Th1 mediated-immunity, needs to be further elaborated, as <italic>Itk</italic> &#x02212;/&#x02212; mice do succumb to this infection [<xref ref-type="bibr" rid="CR37">37</xref>]. To our knowledge, reports in which <italic>Itk</italic> &#x02212;/&#x02212; mice were challenged with the murine herpesvirus 68 (MHV-68), the closest infectious model resembling human EBV infection, have not been published yet. This is somehow surprising given the similarities between MVH-68 and EBV, which were reviewed elsewhere [<xref ref-type="bibr" rid="CR38">38</xref>&#x02013;<xref ref-type="bibr" rid="CR40">40</xref>].</p><p id="Par14">A further approach to examine the Th1/Th2 differentiation was undertaken by several groups investigating the T cell dependent airway hyperresponsiveness in <italic>Itk</italic> &#x02212;/&#x02212; mice. Asthma is characterized by the infiltration of Th2 cells into the lungs; <italic>Itk</italic> &#x02212;/&#x02212; mice have reduced airway hyperresponsiveness to allergen challenge most likely due to an impaired Th2 response [<xref ref-type="bibr" rid="CR41">41</xref>]. In humans, only epidemiological studies have focused on SNPs in the <italic>ITK</italic> gene and the susceptibility to developing asthma without any further functional validation [<xref ref-type="bibr" rid="CR42">42</xref>].</p><p id="Par15">Comprehensive studies beyond the Th1/Th2 paradigm were published by Gomez-Rodriguez 2009 and 2014 [<xref ref-type="bibr" rid="CR43">43</xref>, <xref ref-type="bibr" rid="CR44">44</xref>]. The main focus was the differentiation of Th17 and Treg cells in <italic>Itk</italic> &#x02212;/&#x02212; mice. Under Th17 polarizing conditions sorted CD4 cells expressed significantly less IL17A and more FoxP3 RNA compared with wild type cells confirmed by intracellular staining. Moreover, under Treg culture conditions, a larger amount of cells of <italic>Itk</italic> &#x02212;/&#x02212; na&#x000ef;ve CD4 cells progressed to FoxP3 compared to wild type CD4 cells. Unfortunately, in ITK deficient patients the Th17 and Treg development and function have not been studied yet. The authors further describe a reduced TCR-induced phosphorylation of mammalian target of rapamycin (mTOR) targets accompanied by an altered downstream metabolic profile.</p><p id="Par16">A recent study investigated the protective effect of ITK deficiency on autoimmune phenomena. <italic>CTLA-4</italic> has a quite important function in maintaining T cell tolerance to self, thus <italic>Ctla4 &#x02212;/&#x02212;</italic> mice die of an autoimmune lymphoproliferative disorder driven by self-reactive T cell activation and tissue infiltration [<xref ref-type="bibr" rid="CR45">45</xref>]. In contrast, the <italic>Ctla4&#x02212;/&#x02212; Itk</italic>&#x02212;/&#x02212; double knockout mouse lacks autoimmune pathology and shows a significant reduced lethality.</p></sec><sec id="Sec8"><title>ITK Inhibition</title><p id="Par17">The application of pharmacological ITK inhibitors significantly increased the lifespan of the above mentioned double knockout mouse. The authors further investigated ITK function in the non-obese diabetic mice (NOD), the mouse model of autoimmune insulin-dependent diabetes mellitus, which usually leads to complete destruction of pancreatic beta islets. ITK inhibition by pharmacological means was able to prevent the disease. In contrast to EBV, several viral infections like HIV, Influenza A and Coxsackie virus seem to have a replication cycle depending on ITK [<xref ref-type="bibr" rid="CR46">46</xref>&#x02013;<xref ref-type="bibr" rid="CR48">48</xref>]. Inhibition of ITK leads to inhibition of the replication and alleviation of symptoms in the disease-model. These and many further studies of pharmacological inhibition, often developed from known BTK inhibitors, will portray ITK deficiency from a different perspective. Enthusiastic investigations undertaken to achieve a selective Th1 response is &#x0201c;sought by the medical community given their potential to inhibit a number of Th2-dominant autoimmune, inflammatory, and infectious diseases ranging from cancer immunosuppression and atopic dermatitis to inflammatory bowel disease and even HIV/AIDS&#x0201d; [<xref ref-type="bibr" rid="CR49">49</xref>, <xref ref-type="bibr" rid="CR50">50</xref>]. However to conclude human <italic>ITK</italic> physiology and primary ITK immunodeficiency from these inhibitors is quite ambiguous as this model reflects the acquired rather than the inherited innate ITK impairment.</p></sec><sec id="Sec9"><title>Mast Cell Physiology and NK Cell Mediated Cytotoxicity Depends on ITK Function</title><p id="Par18">Given the restricted expression of ITK in T-cells and mast cells, the functional properties of mast cells lacking ITK have been tested in mouse. ITK has been attributed to differentially modulate mast cell degranulation and cytokine production by regulating expression and activation of NFAT proteins. However, the interactions of an allergen-induced mast cell response, serum IgE and histamine release and consecutive altered expression of the FceR are the main features of mast cell dysfunction in the <italic>Itk</italic> &#x02212;/&#x02212; mouse [<xref ref-type="bibr" rid="CR51">51</xref>].</p><p id="Par19">To the best of our knowledge only one study addressed an altered NK function due to ITK impairment. In a human NK cell model, ITK function was inhibited by the application of siRNA; the authors show, that in activated human NK cells, ITK differentially regulates distinct NK-activating receptors. They suggest that ITK positively regulates FcR-initiated cytotoxicity, while in contrast, enhanced ITK expression would negatively regulate NKG2D-initiated granule-mediated killing [<xref ref-type="bibr" rid="CR52">52</xref>].</p></sec><sec id="Sec10"><title>Summary</title><p id="Par20">Over the last 5&#x000a0;years, a couple of patients suffering from inherited ITK deficiency have been reported and it is quite likely that many more have to be identified. After infection with EBV, their clinical symptoms &#x02013; usually accompanied by ultra-high EBV viral load in the peripheral blood &#x02013; develop, namely severe lymphoproliferation, and Hodgkin&#x02019;s lymphoma. The frequent pulmonary involvement seems to emerge as one clinical hallmark, which should alert clinical immunologists to <italic>ITK</italic>. To date, it remains unknown if there is a major pathology in EBV-negative patients as all reported children with LPD were diagnosed after their first encounter with EBV. Definitive treatment recommendations cannot be given yet, but the current knowledge suggests that early stem cell transplantation is life saving. B cell depletion, e.g. by anti-CD20 therapy seem to be only temporary beneficial for those patients.</p><p id="Par21">Both, the physiological role of <italic>ITK</italic> in mouse models and the functional consequences of the loss-of function mutations identified in human patients have been comprehensively studied. While the murine model has been challenged with various parasites to study Th1 and Th2 responses, an &#x0201c;EBV-mimicking&#x0201d; MHV-68 infectious model is warranted to reveal the pathogenesis of EBV infection in ITK deficiency. Furthermore, a more detailed developmental and functional characterization of T-cell subsets (Tregs and Th17 cells) and mast cells lacking ITK will be challenging issues in further studies.</p></sec>

The Activation of the Sox2 RR2 Pluripotency Transcriptional Reporter in Human Breast Cancer Cell Lines is Dynamic and Labels Cells with Higher Tumorigenic Potential

<p>The striking similarity displayed at the mechanistic level between tumorigenesis and the generation of induced pluripotent stem cells and the fact that genes and pathways relevant for embryonic development are reactivated during tumor progression highlights the link between pluripotency and cancer. Based on these observations, we tested whether it is possible to use a pluripotency-associated transcriptional reporter, whose activation is driven by the SRR2 enhancer from the Sox2 gene promoter (named S4+ reporter), to isolate cancer stem cells (CSCs) from breast cancer cell lines. The S4+ pluripotency transcriptional reporter allows the isolation of cells with enhanced tumorigenic potential and its activation was switched on and off in the cell lines studied, reflecting a plastic cellular process. Microarray analysis comparing the populations in which the reporter construct is active versus inactive showed that positive cells expressed higher mRNA levels of cytokines (IL-8, IL-6, TNF) and genes (such as ATF3, SNAI2, and KLF6) previously related with the CSC phenotype in breast cancer.</p>

<contrib contrib-type="author"><name><surname>Iglesias</surname><given-names>Juan Manuel</given-names></name><xref ref-type="aff" rid="aff1">¹</xref><xref ref-type="aff" rid="aff2">²</xref><uri xlink:type="simple" xlink:href="http://frontiersin.org/people/u/188095"/></contrib><contrib contrib-type="author"><name><surname>Leis</surname><given-names>Olatz</given-names></name><xref ref-type="aff" rid="aff1">¹</xref><xref ref-type="aff" rid="aff3">³</xref></contrib><contrib contrib-type="author"><name><surname>P&#x000e9;rez Ruiz</surname><given-names>Est&#x000ed;baliz</given-names></name><xref ref-type="aff" rid="aff1">¹</xref></contrib><contrib contrib-type="author"><name><surname>Gumuzio Barrie</surname><given-names>Juan</given-names></name><xref ref-type="aff" rid="aff1">¹</xref><xref ref-type="aff" rid="aff3">³</xref><uri xlink:type="simple" xlink:href="http://frontiersin.org/people/u/189736"/></contrib><contrib contrib-type="author"><name><surname>Garcia-Garcia</surname><given-names>Francisco</given-names></name><xref ref-type="aff" rid="aff4">⁴</xref><xref ref-type="aff" rid="aff5">⁵</xref><uri xlink:type="simple" xlink:href="http://frontiersin.org/people/u/175789"/></contrib><contrib contrib-type="author"><name><surname>Aduriz</surname><given-names>Ariane</given-names></name><xref ref-type="aff" rid="aff1">¹</xref></contrib><contrib contrib-type="author"><name><surname>Beloqui</surname><given-names>Izaskun</given-names></name><xref ref-type="aff" rid="aff1">¹</xref></contrib><contrib contrib-type="author"><name><surname>Hernandez-Garcia</surname><given-names>Susana</given-names></name><xref ref-type="aff" rid="aff6">⁶</xref></contrib><contrib contrib-type="author"><name><surname>Lopez-Mato</surname><given-names>Maria Paz</given-names></name><xref ref-type="aff" rid="aff1">¹</xref></contrib><contrib contrib-type="author"><name><surname>Dopazo</surname><given-names>Joaquin</given-names></name><xref ref-type="aff" rid="aff4">⁴</xref><xref ref-type="aff" rid="aff5">⁵</xref><xref ref-type="aff" rid="aff7">⁷</xref><uri xlink:type="simple" xlink:href="http://frontiersin.org/people/u/38670"/></contrib><contrib contrib-type="author"><name><surname>Pandiella</surname><given-names>Atanasio</given-names></name><xref ref-type="aff" rid="aff6">⁶</xref></contrib><contrib contrib-type="author"><name><surname>Menendez</surname><given-names>Javier A.</given-names></name><xref ref-type="aff" rid="aff8">⁸</xref><xref ref-type="aff" rid="aff9">⁹</xref><uri xlink:type="simple" xlink:href="http://frontiersin.org/people/u/175802"/></contrib><contrib contrib-type="author"><name><surname>Martin</surname><given-names>Angel Garcia</given-names></name><xref ref-type="aff" rid="aff1">¹</xref><xref ref-type="aff" rid="aff3">³</xref><xref ref-type="corresp" rid="cor1">*</xref><uri xlink:type="simple" xlink:href="http://frontiersin.org/people/u/140710"/></contrib>

Frontiers in Oncology

<sec sec-type="introduction" id="S1"><title>Introduction</title><p>Cancer stem cells (CSCs) play a central role in tumor progression and recurrence, but our knowledge of their biology and origin is still limited. The lack of good CSC markers in solid tumors could explain our limited understanding of its biology and hampers the development of more efficient chemotherapy treatments. In breast cancer, fluorescent substrates (like Aldefluor), DNA dyes (such as Hoechst 33342 or Rhodamine 123 for the isolation of the side population) or different combinations of surface markers (CD24, CD44, CD133, CD49f, CD29, CD90, CD14) can be used to isolate little overlapping cell populations displaying enhanced tumor-initiating potential. To better understand the origin and dynamics of breast CSCs and to be able to use this knowledge to develop novel therapeutic approaches, new isolation methods and/or more specific combinations of markers are needed.</p><p>Cancer and developmental biology scientists realized over a century ago that genes and pathways relevant to cancer overlap with fetal development as reflected in the reactivation of embryonic genes during tumor progression. Consequently, the question was raised of whether tumors could arise from transformation of tissue stem cells or &#x0201c;retro-differentiation&#x0201d; of more differentiated cells (<xref rid="B1" ref-type="bibr">1</xref>). Nearly 40&#x02009;years latter, these ideas and questions are still hot spots in cancer research. The &#x0201c;retro-differentiation&#x0201d; concept can be now translated as cellular plasticity, a process by which non-stem differentiated cells can spontaneously acquire stem cell-like characteristics (<xref rid="B2" ref-type="bibr">2</xref>). This phenomenon has important implications for cancer therapy and a big impact on our current view of the CSC hypothesis. The CSC model holds that tumors are organized in a cellular hierarchy in which CSCs are the only cells with unlimited proliferation potential and responsible for tumor growth and propagation. Originally, the CSC hypothesis was a linear model with the CSC on the top of the hierarchy and the more differentiated cells on the bottom, but the concept of cellular plasticity and experimental observations are challenging this model (<xref rid="B3" ref-type="bibr">3</xref>).</p><p>It is striking that the similarity observed at the mechanistic level between tumorigenesis and the generation of induced pluripotent stem (iPS) cells from fibroblasts as described by Takahashi and Yamanaka (<xref rid="B4" ref-type="bibr">4</xref>). The production of these iPS cells required the over-expression of four transcription factors, Oct4, Sox2, Klf4, and c-Myc, although Klf4 and c-Myc can be replaced by Lin28 and Nanog (<xref rid="B5" ref-type="bibr">5</xref>) and may even be dispensable. The efficiency of this reprograming process is extremely low and remains so far an <italic>in vitro</italic> phenomenon since there is no evidence that it can naturally occur <italic>in vivo</italic>. The mechanisms underlying the reprogramming process are not well understood yet; however, the three main transcription factors Oct4, Sox2, and Nanog, called master regulators of pluripotency, have proved responsible for maintaining the undifferentiated state (<xref rid="B6" ref-type="bibr">6</xref>, <xref rid="B7" ref-type="bibr">7</xref>). Recently, the processes of reprograming and tumorigenesis have been linked as the p53 tumor suppressor, one of the main regulators of oncogenic transformation, controls the induction of pluripotency (<xref rid="B8" ref-type="bibr">8</xref>&#x02013;<xref rid="B10" ref-type="bibr">10</xref>).</p><p>Both processes, reprograming and transformation, need the expression or activation of oncogenes, inactivation of tumor suppressor genes, overriding the senescence and apoptotic barriers and both processes also involve epigenetic changes and a metabolic switch toward a glycolytic metabolism (<xref rid="B11" ref-type="bibr">11</xref>, <xref rid="B12" ref-type="bibr">12</xref>). The work from Illmensee and Mintz (<xref rid="B13" ref-type="bibr">13</xref>) in the mid 70s strengthens the bonds between pluripotency and cancer. They demonstrated that teratocarcinoma cells are developmentally pluripotent since single teratocarcinoma cells injected into mouse blastocysts can differentiate into many developmentally unrelated tissues. In recent years, the work from Gill Smith&#x02019;s group has shown that breast CSCs are at least multipotent. Their work clearly shows that CSCs when placed in the right microenvironment can behave as phenotypically normal and can contribute to all cell types within the mammary gland epithelium (<xref rid="B14" ref-type="bibr">14</xref>, <xref rid="B15" ref-type="bibr">15</xref>). Furthermore, it has been shown that breast CSCs have the ability to differentiate not only in epithelial but also in the endothelial lineage (<xref rid="B16" ref-type="bibr">16</xref>). This ability of CSCs to differentiate into unrelated cell types is also supported by the fact that glioblastoma stem/progenitor cells can differentiate into endothelial cells contributing to the vascularization of the tumor and hence to tumor progression (<xref rid="B17" ref-type="bibr">17</xref>).</p><p>Sox2 is a good example of a gene involved in embryonic development whose expression is reactivated during tumor generation, as Sox2 is critical to maintain the pluripotent phenotype in embryonic stem cells (ESCs) (<xref rid="B18" ref-type="bibr">18</xref>) and its expression is reactivated during tumor progression (<xref rid="B19" ref-type="bibr">19</xref>&#x02013;<xref rid="B22" ref-type="bibr">22</xref>). Furthermore, Sox2 is part of the original Yamanaka cocktail of transcription factors necessary to reprogram somatic adult cells into iPS cells. These observations, together with the lack of reliable surface markers to isolate breast CSCs, drove us to test whether a pluripotency transcriptional GFP reporter based on the SRR2 enhancer from the Sox2 gene, developed to isolate IPS cells (<xref rid="B23" ref-type="bibr">23</xref>), can be used to isolate cells with cancer stem-like properties from breast cancer cell lines (<xref rid="B24" ref-type="bibr">24</xref>, <xref rid="B25" ref-type="bibr">25</xref>). Our results showed that the activation of this transcriptional GFP reporter in breast cancer cell lines is dynamic and identifies a subpopulation of cells with enhanced tumorigenic potential. Furthermore, when cultures depleted of GFP-positive cells were established and followed over time, some cells switched on the reporter and after a while GFP-negative and GFP-positive populations reached a steady state. Interestingly, the cells in which the reporter is active display higher mRNA levels of IL6, IL8, TNF, ATF3, KLF6, or SNAI2, genes previously related with the CSC-like phenotype and cellular plasticity in breast tumors.</p></sec><sec sec-type="materials|methods" id="S2"><title>Materials and Methods</title><sec id="S2-1"><title>Cell lines and culture conditions</title><p>MCF7 and MDA-MB-231 breast carcinoma cell lines were obtained directly from ATCC (Manasses, VA, USA) and were grown in DMEM (Gibco, Carlsbad, CA, USA) supplemented with 10% fetal bovine serum (Sigma, St. Louis, MO, USA) and 1% Penicillin/Streptomycin (Sigma, St. Louis, MO, USA). MDA-MB-436 cell line was a kind gift from T. Stein (University of Glasgow, UK, previously obtained from ATCC, Manassas, VA, USA) and was grown in DMEM (Gibco, Carlsbad, CA, USA) supplemented with 10% fetal bovine serum (Sigma, St. Louis, MO, USA), 20&#x02009;ng/ml Insulin (Sigma, St. Louis, MO, USA) and 1% penicillin/streptomycin (Sigma, St. Louis, MO, USA). All the cell lines were kept at 37&#x000b0;C in a 5% CO₂ incubator.</p></sec><sec id="S2-2"><title>Mouse xenograft assays</title><p>Female 6-week-old athymic nude mice (Balb/c Nu/Nu) were purchased from Charles River, and were housed in specifically designed pathogen-free isolation animal facility. All animal procedures were performed in accordance with institutional animal care and use guidelines and approved by the IRB. GFP^High and GFP^Low MCF7 cells were resuspended in 200&#x02009;&#x003bc;l of PBS with matrigel and subcutaneously inoculated in left and right caudal mammary fat pads. In all, 2.5&#x02009;&#x000d7;&#x02009;10⁶; 0.5&#x02009;&#x000d7;&#x02009;10⁶, and 0.25&#x02009;&#x000d7;&#x02009;10⁶ GFP^High MCF7cells were inoculated in the right mammary fat pad, with their respective GFP^Low MCF7 controls in the left mammary fat pad. Mice were weighed and the inoculation sites were inspected by palpation at weekly intervals. When tumors become detectable manually, the growth rates were determined by weekly measurement of two diameters of the tumor with a Vernier caliper. The tumor volume was estimated as the volume of an ellipse using the following formula: <italic>V</italic>&#x02009;=&#x02009;4/3&#x02009;&#x000d7;&#x02009;(<italic>a</italic>/2)&#x02009;&#x000d7;&#x02009;(<italic>b</italic>/2)2, where &#x0201c;<italic>a</italic>&#x0201d; and &#x0201c;<italic>b</italic>&#x0201d; correspond to the longest and shortest diameter, respectively. Animals were euthanized when their tumors were harvested.</p></sec><sec id="S2-3"><title>Flow cytometry and microscopy</title><p>Cells were harvested by trypsinization, trypsin was inactivated with regular medium, and DNAse I was added at a final concentration of 0.2&#x02009;mg/ml, cell suspensions were incubated at 37&#x000b0;C for another 10&#x02009;min and spun down, and finally cell pellets were resuspended in a suitable volume of sorting buffer (PBS w/o Ca and Mg, 1% BSA, 5&#x02009;mM EDTA). TO-PRO-3 (Molecular Probes, Life Technologies) was added as dead cell indicator and BD FACSAria or BD FACSCanto machines were used for sorting and analysis experiments following the gating strategy depicted on the Figure S2 in Supplementary Material. When tracking the changes in the percentage of GFP^High cells over time the FACSCanto was calibrated prior to the analysis using the Spherotech Rainbow beads (Spherotech Inc., Lake Forest, IL, USA) to ensure consistent signals over the course of the experiment and verify proper function of the machine.</p></sec><sec id="S2-4"><title>Microarray analysis</title><p>Total RNA from freshly sorted MCF7S4+ GFP^High and GFP^Low cells was prepared using TRIzol (Life Technologies) and whole genome gene expression analysis was performed using the HumanHT-12 v4 Expression BeadChip platform (Illumina) containing 47323 probes per chip. Data were standardized using background correction and quantile normalization (<xref rid="B26" ref-type="bibr">26</xref>). Differential gene expression was carried out using the limma (<xref rid="B27" ref-type="bibr">27</xref>) package from Biocondui (<uri xlink:type="simple" xlink:href="http://www.bioconductor.org/">http://www.bioconductor.org/</uri>). We performed a statistical test for each probe according to Benjamini and Hochberg (<xref rid="B28" ref-type="bibr">28</xref>) methodology. Gene set analysis was carried out for the Gene Ontology terms using FatiScan (<xref rid="B29" ref-type="bibr">29</xref>) in Babelomics (<xref rid="B30" ref-type="bibr">30</xref>) (<uri xlink:type="simple" xlink:href="http://babelomics.bioinfo.cipf.es/">http://babelomics.bioinfo.cipf.es/</uri>). This is a web-based program for the functional interpretation of large-scale experiments. The test aims to directly test the behavior of blocks of functionally related genes, instead of focusing on single genes. This tool detects significantly up- or downregulated blocks of functionally related genes in lists of genes ordered by differential expression. FatiScan returns adjusted <italic>p</italic>-values based on false discovery rate (FDR) method (<xref rid="B28" ref-type="bibr">28</xref>, <xref rid="B31" ref-type="bibr">31</xref>). Significant GO terms were represented by directed acyclic graphs from Blast2GO (<xref rid="B32" ref-type="bibr">32</xref>). GO annotation for the genes in the microarray where taken from Ensembl 56 release (<uri xlink:type="simple" xlink:href="http://www.ensembl.org">http://www.ensembl.org</uri>).</p></sec><sec id="S2-5"><title>Lentiviral gene transfer</title><p>Lentiviral particles encoding the pluripotency transcriptional reporter pL-SIN-EOS-S(4+) EGFP (<xref rid="B23" ref-type="bibr">23</xref>) were produced in-house at the Viral Vectors Core Unit. Cell lines were plated the day before the infection in six-well plates at a cell density of 0.25&#x02009;&#x000d7;&#x02009;10⁶ cells per well and exposed to the lentiviral particles at a MOI 2.5 in serum free medium for 6&#x02009;h, cells were washed twice with serum free medium and kept in regular medium thereafter.</p></sec></sec><sec id="S3"><title>Results</title><sec id="S3-6"><title>The S4+ transcriptional reporter is active in breast cancer cell lines</title><p>The pL-SIN-EOS-S(4+) EGFP pluripotency transcriptional reporter (from now on S4+ reporter) was described by Hotta et al. (<xref rid="B23" ref-type="bibr">23</xref>) as a tool to isolate human iPS cells. The backbone of this reporter is based on the EOS lentiviral system and the synthetic promoter controlling the expression of the EGFP reporter is made of a minimal promoter sequence derived from the LTR promoter from an early transposon (ETn) and four tandem repeats of the SRR2 enhancer sequence from the Sox2 gene (Figure S1 in Supplementary Material). To test whether this pluripotency reporter is active in breast cancer cell lines, cell lines MCF-7S4+ (representing the most common luminal breast cancer type), MDA-MB-231S4+ (as example of mesenchymal-like breast carcinoma), and MDA-MB-436S4+ (representing BRCA1 deficient breast cancer) were generated from parental cell lines through lentiviral gene transfer of the pL-SIN-EOS-S(4+) EGFP transcriptional reporter. The activation of the transcriptional reporter was analyzed by fluorescence microscopy to detect GFP expression. The three S4+ derivative cell lines expressed different levels of GFP in individual cells as shown in Figure <xref ref-type="fig" rid="F1">1</xref>. To quantify the number of cells expressing GFP and its expression levels, FACS analysis was performed. As shown in Figure <xref ref-type="fig" rid="F1">1</xref>, most of the GFP-positive cells expressed low levels of GFP with just a few cells expressing high levels of GFP in the three cell lines.</p><fig id="F1" position="float"><label>Figure 1</label><caption><p><bold>S4+ pluripotency transcriptional reporter is active in breast cancer cell lines</bold>. <bold>(A)</bold> On the left, fluorescence images of MDA-MB-231, MDA-MB-436, and MCF7 cell lines infected with the S4+ reporter to detect GFP expression. On the right are shown the fluorescence image (in green) merge with the bright-field image. <bold>(B)</bold> FACS plots of the wild-type cell lines MDA-MB-231, MDA-MB-436, and MCF7 and the S4+ derivatives infected with the S4+ reporter. On the bigger plot, GFP fluorescence is displayed on the <italic>X</italic>-axis and the fluorescence collected through the 695/40 filter on the <italic>Y</italic>-axis. On the inset, GFP fluorescence is displayed on the <italic>X</italic>-axis and the forward scattering on the <italic>Y</italic>-axis.</p></caption><graphic xlink:href="fonc-04-00308-g001"/></fig></sec><sec id="S3-7"><title>Cells in which S4+ reporter is active are more tumorigenic</title><p>One of the first questions we made after we found out that the S4+ reporter is active in a small population of cells was if there is any difference in tumorigenic potential between the GFP+ and GFP&#x02212; cells. Before we could address this question, we performed a calibration experiment to find out the minimum GFP levels detected on the FACS machine that can be detected by the naked eye on the microscope to help us decide which populations to select for the assay. MCF7S4+ cells were used to establish regions of fluorescence intensity (termed P2-P11) so cells falling in gates P10, P11, and P3 were GFP fluorescent when examined under the fluorescent microscope, as shown in Figure S2 in Supplementary Material, thus we decided to use cells in gate P3 as GFP^High for further studies. We decided to select the gate P4 as GFP^Low and not one of the gates on its right because in the latter ones there is a potential mixture of cells in which the reporter is inactive and cells lacking any viral integration. As control, expression of Sox2 was checked through RT-PCR, showing increased expression of Sox2 in GFP^High cells compared to GFP^Low cells, as expected.</p><p>On the basis of these results, we tested if there is any difference in tumorigenic potential between GFP^High and GFP^Low cells in the MCF7S4+ cell line. GFP^High and GFP^Low populations were then FACS sorted, injected subcutaneously in each flank of nude female mice, and tumor growth was monitored for 8&#x02009;weeks. As shown in Figure <xref ref-type="fig" rid="F2">2</xref>, tumors coming from GFP^High cells grew out first and faster than tumors initiated by GFP^Low cells, and this difference is more evident when higher numbers of cells are injected. We used MCF7S4+ cells as model for tumorigenesis in xenograft experiments instead of a mesenchymal-like model of breast cancer (such as MDA-MB-231S4+) because mesenchymal-like breast carcinoma cells are very invasive and spread rapidly when xenografted to immunocompromised mice, making this model unfeasible to compare direct tumorigenicity.</p><fig id="F2" position="float"><label>Figure 2</label><caption><p><bold>Cells in which the S4+ reporter is active show higher tumorigenic potential in NOD/SCID mice</bold>. The outline of the experiment is shown on the left <bold>(A)</bold> and the outcome on the right <bold>(B)</bold>. <bold>(A)</bold> MCF7 cells were infected with the lentiviral reporter vector, 7&#x02009;days later GFP^High and GFP^Low populations were sorted, GFP expression verified by fluorescent microscopy and SOX2 mRNA differential expression assessed by qPCR. <bold>(B)</bold> The GFP^High and GFP^Low cells were culture for 2&#x02009;days and subcutaneously injected into the left (GFP^Low) or right (GFP^High) fat pads of 6-week-old female nude mice and tumor growth was monitored weekly. In these experiments, three animals per condition were used and the standard deviation is plotted for each time point.</p></caption><graphic xlink:href="fonc-04-00308-g002"/></fig></sec><sec id="S3-8"><title>The S4+ transcriptional reporter is dynamic in breast cancer cell lines</title><p>Hotta et al. have shown that the S4+ reporter is dynamic; it is off in non-pluripotent cells, such as fibroblasts, turns on in iPS cells, and turns off again when the iPS cells are induced to differentiate into any lineage. To test whether the reporter is also dynamic in breast cancer cell lines, GFP^High and GFP^Low populations where sorted, placed in culture and changes in fluorescence where monitored by FACS analysis at each passage. When GFP^Low cells are placed in culture, the heterogeneity of the parental cell line is restored after just few days in culture (Figure <xref ref-type="fig" rid="F3">3</xref>), the same is true for the GFP^High population. The S4+ transcriptional reporter is also dynamic in MDA-MB-231 and MDA-MB-436 breast carcinoma cell lines, as shown in Figure S3 in Supplementary Material.</p><fig id="F3" position="float"><label>Figure 3</label><caption><p><bold>The S4+ reporter is dynamic in MCF7 cells</bold>. <bold>(A)</bold> FACS plots of the parental MCF7S4+ cell line, the GFP^High and GFP^Low populations just after sorting, and after 3, 12, and 33&#x02009;days in culture. <bold>(B)</bold> Typical micrographs of cell cultures at indicated time points. The GFP^High and GFP^Low populations were cultured on its own and changes in fluorescence were monitored by FACS at the indicated time points. Population doublings after sorting are also indicated for each time point.</p></caption><graphic xlink:href="fonc-04-00308-g003"/></fig><p>To confirm that the transcriptional reporters are dynamic and the restoration of the heterogeneity observed in the original cell line is not due to contamination during the sorting process, a clonogenic assay was set up using the MCF7 S4+ cell line. To carry on this assay, individual GFP^High or GFP^Low cells were FACS sorted into each well of a 96-well plate, each well was checked for the presence of an individual cell at the microscope, and after 3&#x02009;weeks, the colonies were scored for the presence of mixed-colonies with GFP⁺ or GFP^&#x02212; cells by fluorescence microscopy (experimental outline depicted in Figure S4 in Supplementary Material). This assay shows that the frequency of firing is much lower than the frequency of extinction of GFP as it would be expected if GFP labels CSCs and the dynamic activity of the transcriptional reporter reflects cellular plasticity (Table <xref ref-type="table" rid="T1">1</xref>).</p><table-wrap id="T1" position="float"><label>Table 1</label><caption><p><bold>Frequency of reporter activation and inactivation through single cell plating</bold>.</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1"/><th align="left" rowspan="1" colspan="1">Number of positions</th><th align="left" rowspan="1" colspan="1">GFP^low</th><th align="left" rowspan="1" colspan="1">GFP^high</th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">Day 1</td><td align="left" rowspan="1" colspan="1">Analyzed</td><td align="left" rowspan="1" colspan="1">288</td><td align="left" rowspan="1" colspan="1">288</td></tr><tr><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1">Containing 1 cell</td><td align="left" rowspan="1" colspan="1">212</td><td align="left" rowspan="1" colspan="1">256</td></tr><tr><td align="left" rowspan="1" colspan="1">After 3&#x02009;weeks</td><td align="left" rowspan="1" colspan="1">With colonies</td><td align="left" rowspan="1" colspan="1">98 (46.2%)</td><td align="left" rowspan="1" colspan="1">144 (56.2%)</td></tr><tr><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1">With colonies made of GFP+ and GFP&#x02212;cells</td><td align="left" rowspan="1" colspan="1">9 (9.1%)</td><td align="left" rowspan="1" colspan="1">76 (52.7%)</td></tr></tbody></table></table-wrap><p>In these series of experiments, we observed that when MCF7S4+ GFP^Low cells are placed in culture they switch on the S4+ reporter, and after a few passages, the culture reached a steady state in which the percentage of GFP^High cells stays around 0.1&#x02013;0.3%.</p></sec><sec id="S3-9"><title>Identification of genes differentially expressed among GFP^High and GFP^Low populations</title><p>GFP^High and GFP^Low populations from the MCF7S4+ cell line were isolated by FACS and total RNA was prepared to perform microarray analysis on the Illumina HumanHT-12_V4 BeadChip platform. Results were normalized and analyzed using Bioconductor and Babelomics, showing that 42 genes were found differentially expressed between the two populations with an adj. <italic>p</italic>-value&#x02009;&#x0003c;&#x02009;0.1, with 40 of those genes showing higher expression in the GFP^High population (Figure <xref ref-type="fig" rid="F4">4</xref>). Among the genes upregulated in GFP^High cells are cytokines (IL-6, IL-8, or TNF) and transcription factors (KLF6, ATF3, SNAI2) that have been previously related with cancer stemness, cellular plasticity, or both. GO analysis showed enrichment in genes related to anti-apoptosis (GO:0006916) and positive regulation of nitric oxide biosynthetic process (GO:0045429) among others (Figure <xref ref-type="fig" rid="F4">4</xref>). Interestingly, increased nitric oxide synthase expression in estrogen receptor-negative breast cancer patients predicts poor survival (<xref rid="B33" ref-type="bibr">33</xref>) and the enrichment in anti-apoptotic genes can contribute the intrinsic chemoresistance characteristic of BCSCs (<xref rid="B34" ref-type="bibr">34</xref>). Further experimentation will be needed to validate the links between these processes, specially inflammation, and CSC induction.</p><fig id="F4" position="float"><label>Figure 4</label><caption><p><bold>Microarray profiling of GFP^High versus GFP^Low MCF7 cells</bold>. <bold>(A)</bold> GO terms enriched in the GFP^High population identified by FatiScan enrichment analysis. <bold>(B)</bold> Genes (columns) differentially expressed between GFP^High and GFP^Low cells (rows). Red color denotes high expression, blue low expression.</p></caption><graphic xlink:href="fonc-04-00308-g004"/></fig></sec></sec><sec sec-type="discussion" id="S4"><title>Discussion</title><p>Different combinations of surface markers have been described to isolate CSCs, but it is striking that little overlap has been found between CSC markers reported in different tumor types (<xref rid="B35" ref-type="bibr">35</xref>). Prominin (CD133) is a good marker for brain and colon CSCs, but has never been successfully used for isolating breast CSCs. Even within breast tumors, the accepted combination of surface markers CD44/CD24 shows differences among different subtypes, being the CD44+/CD24&#x02212; phenotype common in the basal subtype, specially in BRCA1 hereditary tumors, but surprisingly scarce in HER2-positive tumors (<xref rid="B36" ref-type="bibr">36</xref>). In this work, we utilize transcription programs unique in stem cells as a new method to report the activity of CSCs.</p><p>The work from Illmensee and Mintz (<xref rid="B13" ref-type="bibr">13</xref>) linked teratocarcinoma cells to pluripotency. Regulatory networks orchestrated by key transcription factors like SOX2, OCT4, and NANOG have been proposed to play an important role maintaining ESC identity (<xref rid="B6" ref-type="bibr">6</xref>, <xref rid="B7" ref-type="bibr">7</xref>). Interestingly, mRNA profiling studies suggest that ESC and CSCs share common transcriptional programs (<xref rid="B37" ref-type="bibr">37</xref>). Furthermore, transcriptional reporters containing regulatory regions derived from those genes had previously been successfully used to isolate CSCs (<xref rid="B38" ref-type="bibr">38</xref>&#x02013;<xref rid="B40" ref-type="bibr">40</xref>). These observations and the striking similarity observed at the mechanistic level between tumorigenesis and the generation of iPS cells prompted us to test whether a pluripotency transcriptional reporter developed to isolate human iPS cells could be also used in the isolation of CSCs. The pluripotency transcriptional reporter selected to test our hypothesis was the pL-SIN-EOS-S(4+) EGFP (in short, S4+ reporter) (<xref rid="B23" ref-type="bibr">23</xref>). The synthetic promoter driving the expression of GFP is made of four tandem repeats of the SRR2 enhancer from the Sox2 gene plus the LTR from the mouse ETn. These transposons are only active during early mouse embryogenesis in ESCs and embryonic carcinoma (EC) cells. This configuration using a minimal promoter only active in ESCs and ECs and an enhancer sequence derived from the regulatory region of one of the key transcription factor for the maintenance of the ESC identity may provide a more specific way of isolating CSCs and may allow the isolation of CSCs from different tumor sources. Previous work from our laboratory (<xref rid="B19" ref-type="bibr">19</xref>) and others (<xref rid="B41" ref-type="bibr">41</xref>, <xref rid="B42" ref-type="bibr">42</xref>) had shown that Sox2 gene is activated in the early phases of breast tumor development and necessary for tumorigenicity of MCF7 cells; therefore, a reporter based on Sox2 promoter elements seemed appropriate.</p><p>After breast cancer cell lines are infected with the S4+ lentiviral transcriptional reporter, only a small fraction of cells switched on the expression of the GFP repoter, when inspected under the fluorescence microscope (GFP^High). Interestingly, the GFP^High cells showed enhanced tumorigenicity when injected into immunocompromised female mice.</p><p>The S4+ pluripotency transcriptional reporter is active in iPS and ES cells, but it is turned off when iPS cells are induced to differentiate (<xref rid="B23" ref-type="bibr">23</xref>). Here, we show for the first time that this reporter is also dynamic in breast cancer cell lines, as cell cultures depleted of GFP^High cells show spontaneous conversion of GFP^Low cells (these are the cells GFP-negative at the microscope, but with background GFP expression &#x02013; measured through flow cytometry, demonstrating effective viral integration but not transcriptional activation of the reporter) into GFP^High and after a few passages the culture reached a steady state, similar to the parental culture. This phenomenon is reminiscent of cellular plasticity as described by Chaffer et al. (<xref rid="B2" ref-type="bibr">2</xref>), as the spontaneous conversion to a stem-like state of non-stem cells. The reverse is also true, when S4+ cell lines were depleted of GFP^Low cells, some GFP^High cells switched off the expression of the reporter becoming GFP^Low, this might be equivalent to a differentiation process. In the breast cancer cell lines tested, the percentage of cells in which the reporter is active, ranges between 0.4 and 8% after transduction, this percentage fell below 1% when GFP^Low cultures were established and allowed to reach its steady state. These discrepancies in the percentage of fluorescent cells may be due to differences in the transduction efficiency and non-specific activation of the reporter due to positional effects after the lentiviral integration in the genome. Working with steady-state cultures derived from GFP^Low cells reduces the unspecific activation of the reporter due to positional effects. We used cell lines representing the main subtypes of breast cancer to prevent cell line bias: MCF7 as luminal ER-dependent breast cancer, MDA-MB-231 as mesenchymal-like basal breast carcinoma, and MDA-MB-436 as a model of hereditary BRCA1-deficient breast cancer. Similar results were obtained from all cell lines. Expression of the reporter gene did not alter phenotypic features of the cell lines used, such as ER expression in MCF7 cells (data not shown). Moreover, we recently published (<xref rid="B43" ref-type="bibr">43</xref>) a link between E2/ERa signaling in breast cancer and pluripotency-like reporgramming, pointing to a mechanism where SOX2 can promote non-genomic E2 signaling that leads to nuclear phospho-Ser118-ERa, which exacerbates genomic ER signaling in response to E2. Since E2 stimulation has been recently shown to enhance breast tumor-initiating cell survival (through downregulation of miR-140), which targets SOX2, this suggests a bidirectional cross-talk interaction to regulate breast cancer activity.</p><p>In order to understand the mechanisms governing the interconversion of reporter positive and negative cells, transcriptional profiling was carried out. Comparison of GFP^High versus GFP^Low populations showed many genes previously related with CSC homeostasis upregulated in the GFP^High population, where the reporter is active. The cytokines IL-6 and IL-8 are among the upregulated genes in GFP^High cells. IL-6 secretion has been reported to modulate the inducible formation of breast CSCs and their dynamic equilibrium with non-stem cancer cells (<xref rid="B44" ref-type="bibr">44</xref>) and recombinant IL-8 increased mammosphere formation and the ALDEFLUOR-positive population in breast cancer cell lines (<xref rid="B25" ref-type="bibr">25</xref>). The transcription factor ATF3 acts as an oncogene in mouse mammary gland (<xref rid="B45" ref-type="bibr">45</xref>) and enhances TGF&#x003b2; signaling and CSC features in breast cancer cell lines (<xref rid="B46" ref-type="bibr">46</xref>). We found also genes related with epithelial-to-mesenchymal transition (EMT) upregulated in the GFP^High population, such as SLUG or NEDD9. Recent studies suggest a link between EMT and acquisition of stem cell properties (<xref rid="B47" ref-type="bibr">47</xref>, <xref rid="B48" ref-type="bibr">48</xref>) where Slug co-operates with a Sox family member (Sox9) in the reprogramming of differentiated luminal epithelial cells to a stem-like state in the mouse mammary gland and co-expression of both transcription factors in breast cancer is associated with patient survival (<xref rid="B49" ref-type="bibr">49</xref>). NEDD9 acts as a positive regulator of EMT in breast cancer cell lines (<xref rid="B50" ref-type="bibr">50</xref>), and it is also involved in mammary gland tumorigenesis (<xref rid="B51" ref-type="bibr">51</xref>, <xref rid="B52" ref-type="bibr">52</xref>).</p><p>These data are compatible with a model of inducible formation of CSCs and their dynamic equilibrium with non-stem cancer cells. Further experimentation is needed to fully understand the molecular determinants controlling this process, which may have significant impact in our understanding of tumor generation and progression, and therefore opening new possibilities for therapeutic intervention. In this work, we demonstrate the use of a pluripotency related promoter as a tool to track CSC phenotype acquisition in breast cancer, suitable for novel drug discovery targeting the CSC compartment.</p></sec><sec id="S5"><title>Conflict of Interest Statement</title><p>Juan Manuel Iglesias is an employee of the private commercial company Synpromics Ltd. Angel Garcia Martin and Olatz Leis own stock of the private commercial company StemTek Therapeutics. Juan Gumuzio Barrie is an employee of the private commercial company StemTek Therapeutics. The other co-authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.</p></sec><sec sec-type="supplementary-material" id="S6"><title>Supplementary Material</title><p>The Supplementary Material for this article can be found online at <uri xlink:type="simple" xlink:href="http://www.frontiersin.org/Journal/10.3389/fonc.2014.00308/abstract">http://www.frontiersin.org/Journal/10.3389/fonc.2014.00308/abstract</uri></p><supplementary-material content-type="local-data" id="SM1"><media xlink:href="Image_1.TIF"><caption><p>Click here for additional data file.</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="SM2"><media xlink:href="Image_2.TIF"><caption><p>Click here for additional data file.</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="SM3"><media xlink:href="Image_3.TIF"><caption><p>Click here for additional data file.</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="SM4"><media xlink:href="Image_4.TIF"><caption><p>Click here for additional data file.</p></caption></media></supplementary-material></sec>

Knowledge, Perception and Utilization of Postnatal Care of Mothers in Gondar Zuria District, Ethiopia: A Cross-Sectional Study

Could not extract abstract

<contrib contrib-type="author"><name><surname>Tesfahun</surname><given-names>Fikirte</given-names></name><address><email>seblefikir@yahoo.com</email></address><xref ref-type="aff" rid="Aff1">1</xref></contrib><contrib contrib-type="author"><name><surname>Worku</surname><given-names>Walelegn</given-names></name><address><email>walelegnw@yahoo.com</email></address><xref ref-type="aff" rid="Aff2">2</xref></contrib><contrib contrib-type="author"><name><surname>Mazengiya</surname><given-names>Fekadu</given-names></name><address><email>mazengiafek23@gmail.com</email></address><xref ref-type="aff" rid="Aff3">3</xref></contrib><contrib contrib-type="author" corresp="yes"><name><surname>Kifle</surname><given-names>Manay</given-names></name><address><email>manay2000@gmail.com</email></address><xref ref-type="aff" rid="Aff2">2</xref></contrib><aff id="Aff1"><label>1</label>Yekokeb Berhan Program for Highly Vulnerable Children, Mahibere Hiwot Reproductive Health and Social Development Organazation, P. O. Box. No. 631, Gondar, Ethiopia </aff><aff id="Aff2"><label>2</label>Department of Environmental and Occupational Health and Safety, College of Medicine and Health Science, University of Gondar, P. O. Box. No. 196, Gondar, Ethiopia </aff><aff id="Aff3"><label>3</label>Department of Midwifery, College of Medicine and Health Science, University of Gondar, P. O. Box. No. 196, Gondar, Ethiopia </aff>

Maternal and Child Health Journal

<sec id="Sec1"><title>Background</title><p> The health of mothers is mostly regarded as an indicator the health of the society. Globally, more than half a million women die each year from complications of pregnancy and childbirth [<xref ref-type="bibr" rid="CR1">1</xref>&#x02013;<xref ref-type="bibr" rid="CR3">3</xref>]. A large proportion of maternal and neonatal deaths occur during the first 48&#x000a0;h after delivery. Thus, postnatal care (PNC) is important for both the mother and the child to treat complications arising from the delivery, as well as to provide the mother with important information [<xref ref-type="bibr" rid="CR4">4</xref>]. Every year, four million infants die within their first month of life, representing nearly 40&#x000a0;% of all deaths of children under age 5&#x000a0;year old. Almost all newborn deaths are in developing countries, with the highest number in South Asia and the highest rates in sub-Saharan Africa [<xref ref-type="bibr" rid="CR5">5</xref>]. PNC coverage is low in Ethiopia; only 5&#x000a0;% of mothers received PNC within the critical first 2&#x000a0;days after delivery [<xref ref-type="bibr" rid="CR4">4</xref>]. Nationwide, 34.3&#x000a0;% of mothers in Ethiopia receive PNC within the first 6&#x000a0;weeks after delivery. In the Amhara region (in which Gondar is found), the number of women attending PNC jumps to 44.5&#x000a0;% [<xref ref-type="bibr" rid="CR6">6</xref>]. It is an incontrovertible fact that PNC services help to safeguard women from complications following delivery and provide important opportunities to assess the infant&#x02019;s development. Moreover, PNC services help to offer newborn care (e.g. Counseling on breast feeding and Preventing Mother -to- Child Transmission) and other services like immunization and family planning which are crucial for both the mother and the infant [<xref ref-type="bibr" rid="CR7">7</xref>&#x02013;<xref ref-type="bibr" rid="CR9">9</xref>].</p><p>In the course of a lifetime, an individual encounters the greatest risk of mortality during birth and the first 28&#x000a0;days of life (the neonatal period). The risk of maternal mortality and morbidity is also high at birth and in the immediate post-natal period. Each year, nearly 4 million newborns die during the neonatal period throughout the world. Three quarters of these deaths take place within 1&#x000a0;week of birth, 1&#x02013;2 million die during the first day following birth, and most of these deaths occur at home [<xref ref-type="bibr" rid="CR5">5</xref>]. The developing world has the highest prevalence of maternal and infant morbidity and mortality. A total of 99&#x000a0;% of all maternal deaths occur in developing countries, where 85&#x000a0;% of the world&#x02019;s population lives [<xref ref-type="bibr" rid="CR10">10</xref>]. This is especially true in sub-Saharan African countries. Over 13,000 mothers, newborns, and children die every day in sub-Saharan Africa [<xref ref-type="bibr" rid="CR11">11</xref>]. Ethiopia, one of the countries in sub-Saharan Africa, is among the six countries that contribute about 50&#x000a0;% of the maternal deaths worldwide [<xref ref-type="bibr" rid="CR12">12</xref>]. In 2005 in Ethiopia, the maternal mortality rate was found to be 673/100,000. This rate has declined to 470 in 2008 as reported in 2010/11 [<xref ref-type="bibr" rid="CR4">4</xref>, <xref ref-type="bibr" rid="CR13">13</xref>, <xref ref-type="bibr" rid="CR14">14</xref>].</p><p>The poor quality of post-natal care in Ethiopia is a result of weak health infrastructure, poorly trained health professionals, and inadequate supplies of drugs and equipment. This, coupled with patients&#x02019; poverty, results in low utilization of services. According to the 2005 Ethiopia Demographic and Health Survey, an estimated 95&#x000a0;% of mothers nationwide did not receive PNC in the critical first 2 days after delivery [<xref ref-type="bibr" rid="CR4">4</xref>]. Research indicates that maternal health care service utilization is affected by several factors including awareness of the health services, accessibility, socio-cultural beliefs, practices, individual attitudes, and health-care-seeking behaviors. However, the determinants of utilization of maternal health care services are not the same across different cultures and socioeconomic statuses within a society. There is little information about mothers&#x02019; knowledge, perception, and utilization of PNC and factors that influence the use of PNC in the Gondar Zuria District of Ethiopia. This study aimed to elucidate the various factors influencing the use of these services within the district.</p></sec><sec id="Sec2"><title>Methods</title><sec id="Sec3"><title>Study Design and Study Setting</title><p>The study was cross-sectional and community-based, using both quantitative and qualitative methods of data collection and analysis. The study was conducted from April to August 2011 in the Gondar Zuria District, which is among twenty-four districts in the North Gondar Administrative Zone in the Amhara Regional State, is located 700&#x000a0;km from Addis Ababa, Ethiopia. At the time of the study, the population of the district was 204,698, of which 101,009 were female. There are 3 urban and 35 rural wards, five health centers, and two heath posts working to maintain the health status of people in the district.</p></sec><sec id="Sec4"><title>Sampling Procedure</title><p>The study population consisted of mothers from 15 to 49&#x000a0;years who gave birth in the last year in the selected wards and were residents of the district for at least 6&#x000a0;months. Multistage sampling technique was used to select study participants. The district was stratified into urban and rural wards, and then one urban and 12 rural wards were chosen by simple random sampling technique. In the study area, 12,282 women were estimated to be eligible (women who gave birth within the last year). The total sample size 836 was distributed proportionally to each strata: 131 urban and 705 rural households sampled were selected by systematic random sampling technique. If there was more than one mother within the same household, a lottery method was used to select the mother to be included.</p><p>Purposive sampling was used to select participants for focus group discussion (FGDs). Three FGDs which comprise a total of 6&#x02013;8 individuals were conducted with mothers, health extension workers (HEWs), and community health workers (CHWs). A total of 16 mothers, three HEWs, and three CHWs participated in the FGDs.</p></sec><sec id="Sec5"><title>Data Collection Instrument</title><p>The questionnaire was developed through review of related Ethiopian and international literatures. The questionnaire was prepared in English then translated into Amharic which is the local language of the area and back to English in order to ensure its consistency. The questionnaires consisted of information on socio-demographic characteristics, knowledge and perception of mothers towards PNC, and utilization of PNC services. Pre-testing of the questionnaire was done in other, unselected wards of the district and modifications were made based on the outcome of the pre-test. The data were collected by interview using a structured questionnaire. For the qualitative data, guiding questions for the FDGs were developed in English and converted to Amharic and then checked for validity. Camera and tape recorders were used in the discussion to record every discussion on the topics. Data were collected by 13 diploma nursing students after a 2&#x000a0;days training. Three supervisors and the principal investigators closely followed the day to day data collection process and ensured completeness and consistency of the collected questionnaires. Each questionnaire was checked for completeness and consistency by supervisors and principal investigators and incompletely filled questionnaires were discarded.</p></sec><sec id="Sec6"><title>Data Processing and Analysis</title><p>The data were entered into EPI info version 3.5.1 and exported to SPSS version 16.0 for analysis. Descriptive statistics such as frequencies and percents were computed to describe the study population in relation to relevant variables. Bivariate and multivariate logistic regression analyses were employed. Those factors that were significant at the 20&#x000a0;% level in the Bivariate logistic regression analysis were considered for the multivariate logistic regression analysis. Odds ratios with 95&#x000a0;% Confidence Interval (CI) and <italic>p</italic> value of &#x0003c;0.05 were computed to assess the presence and degree of statistical association between dependent and independent variables. The qualitative data responses were coded, categorized, and then organized by content with thematic analysis.</p></sec><sec id="Sec7"><title>Ethical Considerations</title><p>Ethical clearance for the study was obtained from the Institutional Review Board of University of Gondar. Official letters were written from Institute of Public Health to North Gondar Health Office, to the District Health Office, and to each ward to get permission. Written and verbal consent was obtained from each participant after explaining the purpose and nature of the research. Participation in the study was on a voluntary basis and participants were informed their right to quit/refuse their participation at any stage of the study if they do not want to participate. Moreover, confidentiality of the information was assured by using an anonymous questionnaire.</p></sec></sec><sec id="Sec8"><title>Result of the Study</title><sec id="Sec9"><title>Socio-Demographic Characteristics of the Respondents</title><p>From the total 836 mothers 820 (98.09&#x000a0;%) completely filed and returned the questionnaire. The majority of mothers surveyed (84.88&#x000a0;%) were from rural areas, with the remainder (15.12&#x000a0;%) living in cities and towns. The mean age of study participants was 28.58&#x000a0;years, with a standard deviation of &#x000b1;7.71&#x000a0;years. A large proportion of participants (47.32&#x000a0;%) travel a distance of one to 2&#x000a0;h on foot and 19.88&#x000a0;% require more than 2&#x000a0;h, whereas 32.80&#x000a0;% of participants travel &#x0003c;1&#x000a0;h to nearby health centers (Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>).<table-wrap id="Tab1"><label>Table&#x000a0;1</label><caption><p>Socio-demographic and economic characteristics of respondents, North Gondar, Ethiopia, 2011 (n&#x000a0;=&#x000a0;820)</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left">Characteristic</th><th align="left">Frequency</th><th align="left">Percent (%)</th></tr></thead><tbody><tr><td align="left" colspan="3">Age (years)</td></tr><tr><td align="left">&#x000a0;15&#x02013;19</td><td char="." align="char">109</td><td char="." align="char">13.29</td></tr><tr><td align="left">&#x000a0;20&#x02013;24</td><td char="." align="char">178</td><td char="." align="char">21.71</td></tr><tr><td align="left">&#x000a0;25&#x02013;29</td><td char="." align="char">173</td><td char="." align="char">21.10</td></tr><tr><td align="left">&#x000a0;30&#x02013;34</td><td char="." align="char">181</td><td char="." align="char">22.07</td></tr><tr><td align="left">&#x000a0;&#x0003e;35</td><td char="." align="char">179</td><td char="." align="char">21.83</td></tr><tr><td align="left" colspan="3">Marital status</td></tr><tr><td align="left">&#x000a0;Single</td><td char="." align="char">34</td><td char="." align="char">4.15</td></tr><tr><td align="left">&#x000a0;Married</td><td char="." align="char">706</td><td char="." align="char">86.10</td></tr><tr><td align="left">&#x000a0;Separated due to work</td><td char="." align="char">27</td><td char="." align="char">3.29</td></tr><tr><td align="left">&#x000a0;Divorced</td><td char="." align="char">33</td><td char="." align="char">4.02</td></tr><tr><td align="left">&#x000a0;Widowed</td><td char="." align="char">20</td><td char="." align="char">2.44</td></tr><tr><td align="left" colspan="3">Educational level</td></tr><tr><td align="left">&#x000a0;Illiterate</td><td char="." align="char">677</td><td char="." align="char">82.56</td></tr><tr><td align="left">&#x000a0;Primary</td><td char="." align="char">101</td><td char="." align="char">12.32</td></tr><tr><td align="left">&#x000a0;Secondary</td><td char="." align="char">33</td><td char="." align="char">4.02</td></tr><tr><td align="left">&#x000a0;Diploma and above</td><td char="." align="char">9</td><td char="." align="char">1.10</td></tr><tr><td align="left" colspan="3">Husbands&#x02019; educational level of (n&#x000a0;=&#x000a0;753)</td></tr><tr><td align="left">&#x000a0;Illiterate</td><td char="." align="char">608</td><td char="." align="char">80.74</td></tr><tr><td align="left">&#x000a0;Primary</td><td char="." align="char">87</td><td char="." align="char">11.55</td></tr><tr><td align="left">&#x000a0;Secondary</td><td char="." align="char">37</td><td char="." align="char">4.91</td></tr><tr><td align="left">&#x000a0;Diploma and above</td><td char="." align="char">31</td><td char="." align="char">4.12</td></tr><tr><td align="left" colspan="3">Mothers&#x02019; occupation</td></tr><tr><td align="left">&#x000a0;Housewife</td><td char="." align="char">633</td><td char="." align="char">77.20</td></tr><tr><td align="left">&#x000a0;Farmer</td><td char="." align="char">81</td><td char="." align="char">9.88</td></tr><tr><td align="left">&#x000a0;Private employ</td><td char="." align="char">61</td><td char="." align="char">7.44</td></tr><tr><td align="left">&#x000a0;Government employ</td><td char="." align="char">10</td><td char="." align="char">1.22</td></tr><tr><td align="left">&#x000a0;Daily labor worker</td><td char="." align="char">33</td><td char="." align="char">4.02</td></tr><tr><td align="left">&#x000a0;Others</td><td char="." align="char">2</td><td char="." align="char">0.24</td></tr><tr><td align="left" colspan="3">Husbands&#x02019; occupation (n&#x000a0;=&#x000a0;753)</td></tr><tr><td align="left">&#x000a0;Farmer</td><td char="." align="char">620</td><td char="." align="char">82.34</td></tr><tr><td align="left">&#x000a0;Private employ</td><td char="." align="char">68</td><td char="." align="char">9.03</td></tr><tr><td align="left">&#x000a0;Government employ</td><td char="." align="char">38</td><td char="." align="char">5.05</td></tr><tr><td align="left">&#x000a0;Daily labor worker</td><td char="." align="char">25</td><td char="." align="char">3.32</td></tr><tr><td align="left">&#x000a0;Others</td><td char="." align="char">2</td><td char="." align="char">0.27</td></tr><tr><td align="left" colspan="3">Residence of participants</td></tr><tr><td align="left">&#x000a0;Urban</td><td char="." align="char">124</td><td char="." align="char">15.12</td></tr><tr><td align="left">&#x000a0;Rural</td><td char="." align="char">696</td><td char="." align="char">84.88</td></tr><tr><td align="left" colspan="3">Distance from health institution (h)</td></tr><tr><td align="left">&#x000a0;&#x0003c;1</td><td char="." align="char">269</td><td char="." align="char">32.80</td></tr><tr><td align="left">&#x000a0;1&#x02013;2</td><td char="." align="char">388</td><td char="." align="char">47.32</td></tr><tr><td align="left">&#x000a0;&#x0003e;2</td><td char="." align="char">163</td><td char="." align="char">19.88</td></tr></tbody></table></table-wrap> </p></sec><sec id="Sec10"><title>Awareness of Post Natal Care Service</title><p>Six hundred ninety-two (84.39&#x000a0;%) of mothers were aware that they should receive PNC services after delivery. Those women who were aware of the need for PNC cited the following reasons for attending a clinic in the post-natal period: 97.69&#x000a0;% of women mentioned the need to receive vaccinations; 42.49&#x000a0;% to be counseled on family planning; 37.57&#x000a0;% to prevent and treat delivery related problems; 22.98&#x000a0;% to receive nutritional advice; 7.08&#x000a0;% to discuss breastfeeding; 1.16&#x000a0;% to receive advice on danger signs of pregnancy.</p><p>Among mothers who were aware that they should receive PNC services after delivery 85.84&#x000a0;% were given information about PNC from HEWs, 17.77&#x000a0;% from nurses, 8.24&#x000a0;% from family, and 1.16&#x000a0;% from doctors.</p><p>Maternal age, marital status, place of residence, previous visit by community health agents/HEWs, and having follow up for antenatal care had association with awareness of mothers about PNC service in both bivariate and multivariate logistic regression analysis (<italic>p</italic> value &#x0003c;0.05) (Table&#x000a0;<xref rid="Tab2" ref-type="table">2</xref>).<table-wrap id="Tab2"><label>Table&#x000a0;2</label><caption><p>Association of factors with awareness of mothers about postnatal care service, North Gondar, Ethiopia, 2011 (n&#x000a0;=&#x000a0;820)</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="2">Variable</th><th align="left" colspan="2">Awareness of PNC</th><th align="left" rowspan="2">Crude OR (95&#x000a0;% CI)</th><th align="left" rowspan="2">Adjusted OR (95&#x000a0;% CI)</th></tr><tr><th align="left">Yes</th><th align="left">No</th></tr></thead><tbody><tr><td align="left" colspan="5">Age of participant (years)</td></tr><tr><td align="left">&#x000a0;15&#x02013;19</td><td char="." align="char">77</td><td char="." align="char">32</td><td char="." align="char">1.00</td><td align="left"/></tr><tr><td align="left">&#x000a0;20&#x02013;24</td><td char="." align="char">151</td><td char="." align="char">27</td><td align="left">2.32 (1.30&#x02013;4.16)</td><td align="left">2.56 (1.15&#x02013;5.70)</td></tr><tr><td align="left">&#x000a0;25&#x02013;29</td><td char="." align="char">150</td><td char="." align="char">23</td><td align="left">2.71 (1.48&#x02013;4.95)</td><td align="left">3.16 (1.29&#x02013;7.71)*</td></tr><tr><td align="left">&#x000a0;30&#x02013;34</td><td char="." align="char">159</td><td char="." align="char">22</td><td align="left">3.00 (1.64&#x02013;5.51)</td><td align="left">3.47 (1.31&#x02013;9.18)*</td></tr><tr><td align="left">&#x000a0;&#x0003e;35</td><td char="." align="char">155</td><td char="." align="char">24</td><td align="left">2.69 (1.48&#x02013;4.87)*</td><td align="left">2.93 (1.17&#x02013;7.33)*</td></tr><tr><td align="left" colspan="5">Marital status</td></tr><tr><td align="left">&#x000a0;Single</td><td char="." align="char">17</td><td char="." align="char">17</td><td char="." align="char">1.00</td><td align="left"/></tr><tr><td align="left">&#x000a0;Married</td><td char="." align="char">616</td><td char="." align="char">90</td><td align="left">6.84 (3.37&#x02013;13.89)**</td><td align="left">3.14 (1.17&#x02013;8.41)*</td></tr><tr><td align="left">&#x000a0;Separated due to work</td><td char="." align="char">20</td><td char="." align="char">7</td><td align="left">2.86 (0.96&#x02013;8.52)</td><td align="left">1.60 (0.37&#x02013;6.81)</td></tr><tr><td align="left">&#x000a0;Divorced</td><td char="." align="char">25</td><td char="." align="char">8</td><td align="left">3.13 (1.10&#x02013;8.86)*</td><td align="left">3.03 (0.77&#x02013;11.98)</td></tr><tr><td align="left">&#x000a0;Widowed</td><td char="." align="char">14</td><td char="." align="char">6</td><td align="left">2.33 (0.73&#x02013;7.51)</td><td align="left">2.41 (0.52&#x02013;11.13)</td></tr><tr><td align="left" colspan="5">Place of residence</td></tr><tr><td align="left">&#x000a0;Rural</td><td char="." align="char">573</td><td char="." align="char">123</td><td char="." align="char">1.00</td><td align="left"/></tr><tr><td align="left">&#x000a0;Urban</td><td char="." align="char">119</td><td char="." align="char">5</td><td align="left">5.11 (2.05&#x02013;12.77)**</td><td align="left">6.58 (2.09&#x02013;20.75)*</td></tr><tr><td align="left" colspan="5">Distance from health institution (h)</td></tr><tr><td align="left">&#x000a0;&#x0003c;1</td><td char="." align="char">273</td><td char="." align="char">32</td><td align="left">2.18 (1.29&#x02013;3.66)</td><td align="left">0.55 (0.27&#x02013;1.12)</td></tr><tr><td align="left">&#x000a0;1&#x02013;2</td><td char="." align="char">329</td><td char="." align="char">59</td><td align="left">1.64 (1.03&#x02013;2.59)*</td><td align="left">0.73 (0.40&#x02013;1.34)</td></tr><tr><td align="left">&#x000a0;&#x0003e;2</td><td char="." align="char">126</td><td char="." align="char">37</td><td char="." align="char">1.00</td><td align="left"/></tr><tr><td align="left" colspan="5">Number of children born alive</td></tr><tr><td align="left">&#x000a0;None</td><td char="." align="char">41</td><td char="." align="char">18</td><td char="." align="char">1.00</td><td align="left"/></tr><tr><td align="left">&#x000a0;1&#x02013;5</td><td char="." align="char">372</td><td char="." align="char">65</td><td align="left">2.51 (1.36&#x02013;4.64)**</td><td align="left">0.40 (0.16&#x02013;1.01)</td></tr><tr><td align="left">&#x000a0;&#x0003e;5</td><td char="." align="char">279</td><td char="." align="char">45</td><td align="left">2.72 (1.44&#x02013;5.15)**</td><td align="left">0.29 (0.09&#x02013; 0.85)</td></tr><tr><td align="left" colspan="5">ANC visit</td></tr><tr><td align="left">&#x000a0;Not used</td><td char="." align="char">48</td><td char="." align="char">47</td><td char="." align="char">1.00</td><td align="left"/></tr><tr><td align="left">&#x000a0;Once</td><td char="." align="char">143</td><td char="." align="char">39</td><td align="left">5.12 (3.07&#x02013;8.55)**</td><td align="left">2.82 (1.54&#x02013;5.17)**</td></tr><tr><td align="left">&#x000a0;Twice</td><td char="." align="char">321</td><td char="." align="char">12</td><td align="left">37.34 (18.82&#x02013;74.08)**</td><td align="left">26.50 (12.36&#x02013;56.82)**</td></tr><tr><td align="left">&#x000a0;Three and more</td><td char="." align="char">180</td><td char="." align="char">10</td><td align="left">25.03 (12.03&#x02013;52.42)**</td><td align="left">15.40 (6.76&#x02013;35.08)**</td></tr><tr><td align="left" colspan="5">Last 1&#x000a0;year visit health facility</td></tr><tr><td align="left">&#x000a0;Yes</td><td char="." align="char">613</td><td char="." align="char">80</td><td align="left">4.66 (3.04&#x02013;8.55)**</td><td align="left">1.44 (0.82&#x02013;&#x02013;2.63)</td></tr><tr><td align="left">&#x000a0;No</td><td char="." align="char">79</td><td char="." align="char">48</td><td char="." align="char">1.00</td><td align="left"/></tr><tr><td align="left" colspan="5">Visit by HEW</td></tr><tr><td align="left">&#x000a0;Yes</td><td char="." align="char">18</td><td char="." align="char">77</td><td align="left">4.08 (2.76&#x02013;6.03)**</td><td align="left">2.01 (1.19&#x02013;3.39)**</td></tr><tr><td align="left">&#x000a0;No</td><td char="." align="char">505</td><td char="." align="char">51</td><td char="." align="char">1.00</td><td align="left"/></tr><tr><td align="left" colspan="5">History of institutional delivery</td></tr><tr><td align="left">&#x000a0;Yes</td><td char="." align="char">75</td><td char="." align="char">6</td><td align="left">2.47 (1.05&#x02013;5.81)</td><td align="left">1.02 (0.39&#x02013;2.68)</td></tr><tr><td align="left">&#x000a0;No</td><td char="." align="char">617</td><td char="." align="char">122</td><td char="." align="char">1.00</td><td align="left"/></tr></tbody></table><table-wrap-foot><p>1.00&#x000a0;=&#x000a0;Referent category, *&#x000a0;statistically associated <italic>p</italic> value &#x0003c;0.05, **&#x000a0;statistically associated <italic>p</italic> value &#x0003c;0.001</p></table-wrap-foot></table-wrap> </p></sec><sec id="Sec11"><title>Mothers&#x02019; Perception Towards Postnatal Care</title><p>The majority of mothers (74.27&#x000a0;%) stated that PNC is necessary for women and their children. Table&#x000a0;<xref rid="Tab3" ref-type="table">3</xref> shows the factors that had an association with the positive perception to PNC.<table-wrap id="Tab3"><label>Table&#x000a0;3</label><caption><p>Association of factors with perception of mothers towards postnatal care service, North Gondar, Ethiopia, 2011 (n&#x000a0;=&#x000a0;820)</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="2">Variable</th><th align="left" colspan="2">Perceive positively</th><th align="left">Crude OR (95&#x000a0;% CI)</th><th align="left">Adjusted OR (95&#x000a0;% CI)</th></tr><tr><th align="left">Yes</th><th align="left">No</th><th align="left"/><th align="left"/></tr></thead><tbody><tr><td align="left" colspan="5">Age of participant(years)</td></tr><tr><td align="left">&#x000a0;15&#x02013;19</td><td char="." align="char">77</td><td char="." align="char">22</td><td char="." align="char">1.00</td><td align="left"/></tr><tr><td align="left">&#x000a0;20&#x02013;24</td><td char="." align="char">151</td><td char="." align="char">27</td><td align="left">2.54 (1.42&#x02013;4.55)*</td><td align="left">1.87 (0.98&#x02013;3.55)</td></tr><tr><td align="left">&#x000a0;25&#x02013;29</td><td char="." align="char">150</td><td char="." align="char">23</td><td align="left">1.19 (0.71&#x02013;2.04)</td><td align="left">0.69 (0.38&#x02013;1.27)</td></tr><tr><td align="left">&#x000a0;30&#x02013;34</td><td char="." align="char">159</td><td char="." align="char">22</td><td align="left">1.39 (0.82&#x02013;2.38)</td><td align="left">0.97 (0.54&#x02013;1.76)</td></tr><tr><td align="left">&#x000a0;&#x0003e;35</td><td char="." align="char">155</td><td char="." align="char">24</td><td align="left">0.79 (0.48&#x02013;1.33)</td><td align="left">0.59 (.33&#x02013;1.05)</td></tr><tr><td align="left" colspan="5">Place of residence</td></tr><tr><td align="left">&#x000a0;Rural</td><td char="." align="char">487</td><td char="." align="char">209</td><td char="." align="char">1.00</td><td align="left"/></tr><tr><td align="left">&#x000a0;Urban</td><td char="." align="char">122</td><td char="." align="char">2</td><td align="left">26.18 (6.41&#x02013;11.86)**</td><td align="left">11.22 (2.53&#x02013;49.80)**</td></tr><tr><td align="left" colspan="5">Distance from health Institution</td></tr><tr><td align="left">&#x000a0;&#x0003c;1&#x000a0;h</td><td char="." align="char">240</td><td char="." align="char">29</td><td align="left">2.60 (1.54&#x02013;4.41)*</td><td align="left">0.93 (0.50&#x02013;1.71)</td></tr><tr><td align="left">&#x000a0;1&#x02013;2&#x000a0;h</td><td char="." align="char">245</td><td char="." align="char">143</td><td align="left">0.54 (0.36&#x02013;0.82)</td><td align="left">0.34 (0.22&#x02013;0.58)</td></tr><tr><td align="left">&#x000a0;&#x0003e;2&#x000a0;h</td><td char="." align="char">124</td><td char="." align="char">39</td><td char="." align="char">1.00</td><td align="left"/></tr><tr><td align="left" colspan="5">ANC visit</td></tr><tr><td align="left">&#x000a0;Not used</td><td char="." align="char">59</td><td char="." align="char">56</td><td char="." align="char">1.00</td><td align="left"/></tr><tr><td align="left">&#x000a0;Once</td><td char="." align="char">156</td><td char="." align="char">26</td><td align="left">5.70 (3.28&#x02013;9.90)**</td><td align="left">3.04 (1.61&#x02013;5.74)*</td></tr><tr><td align="left">&#x000a0;Twice</td><td char="." align="char">249</td><td char="." align="char">84</td><td align="left">2.81 (1.81&#x02013;4.38)**</td><td align="left">1.34 (0.74&#x02013;2.41)</td></tr><tr><td align="left">&#x000a0;There and more</td><td char="." align="char">145</td><td char="." align="char">45</td><td align="left">3.06 (1.86&#x02013;5.02)**</td><td align="left">1.26 (0.67&#x02013;2.40)</td></tr><tr><td align="left" colspan="5">Visit by community health agent/HEW</td></tr><tr><td align="left">&#x000a0;Yes</td><td char="." align="char">451</td><td char="." align="char">105</td><td align="left">2.88 (2.08&#x02013;3.99)**</td><td align="left">1.95 (1.33&#x02013;2.86)**</td></tr><tr><td align="left">&#x000a0;No</td><td char="." align="char">158</td><td char="." align="char">106</td><td char="." align="char">1.00</td><td align="left"/></tr><tr><td align="left" colspan="5">Awareness about PNC</td></tr><tr><td align="left">&#x000a0;Yes</td><td char="." align="char">546</td><td char="." align="char">146</td><td align="left">3.47 (2.61&#x02013;5.71)**</td><td align="left">3.55 (2.11&#x02013;5.97)**</td></tr><tr><td align="left">&#x000a0;No</td><td char="." align="char">63</td><td char="." align="char">65</td><td char="." align="char">1.00</td><td align="left"/></tr></tbody></table><table-wrap-foot><p>1.00&#x000a0;=&#x000a0;Referent category, *&#x000a0;statistically associated <italic>p</italic> value &#x0003c;0.05, **&#x000a0;statistically associated <italic>p</italic> value &#x0003c;0.001</p></table-wrap-foot></table-wrap> </p></sec><sec id="Sec12"><title>Utilization of Post Natal Care</title><p>Among mothers who give birth in the last 1&#x000a0;year, 548 (66.83&#x000a0;%) of them attended postnatal services. Above half 496 (60.49&#x000a0;%) of the mothers attended for immunization of their babies, 175 (21.34&#x000a0;%) for family planning, 129 (15.73&#x000a0;%) for counseling on PNC, 29 (3.54&#x000a0;%) for counseling on breastfeeding, and 15 (1.83&#x000a0;%) for physical examination (Fig.&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>).<fig id="Fig1"><label>Fig.&#x000a0;1</label><caption><p>Utilization of postnatal care by mothers who give birth in the last 1&#x000a0;year, North Gondar, Ethiopia, 2011</p></caption><graphic xlink:href="10995_2014_1474_Fig1_HTML" id="MO1"/></fig> </p><p>Two hundred seventy-two (33.17&#x000a0;%) mothers who didn&#x02019;t use PNC service provide different reasons for not attending PNC services. As depicted in (Fig.&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>) majority of the mothers do not attend because of lack of time and the long distance required to travel in order to receive services.<fig id="Fig2"><label>Fig.&#x000a0;2</label><caption><p>Mothers&#x02019; reasons for not attending postnatal care service, North Gondar, Ethiopia, 2011</p></caption><graphic xlink:href="10995_2014_1474_Fig2_HTML" id="MO2"/></fig> </p><p>Among mothers who utilize the service, 67.70&#x000a0;% utilized once, 27.92&#x000a0;% twice and 4.38&#x000a0;% three and or more times within 6&#x000a0;weeks after delivery. Half of (52.19&#x000a0;%) mothers utilize services from HEWs and community health agents in outreach service, 46.90&#x000a0;% of mothers from health institutions, and the other 0.91&#x000a0;% of mothers receive PNC from trained birth attendants.</p><p>In a multivariate logistic regression analysis, the following factors were associated with the utilization of PNC: place of residence, distance from health institution, history of ANC, contact with community health agents within the last year, awareness about the need for PNC, history of institutional delivery, and the ability to make decisions regarding healthcare (Table&#x000a0;<xref rid="Tab4" ref-type="table">4</xref>).<table-wrap id="Tab4"><label>Table&#x000a0;4</label><caption><p>Association of factors with post natal care utilization, North Gondar, Ethiopia, 2011 (n&#x000a0;=&#x000a0;820)</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="2">Variable</th><th align="left" colspan="2">PNC utilization</th><th align="left" rowspan="2">COR (95&#x000a0;% CI)</th><th align="left" rowspan="2">AOR (95&#x000a0;% CI)</th></tr><tr><th align="left">Yes</th><th align="left">No</th></tr></thead><tbody><tr><td align="left" colspan="5">Age of participant</td></tr><tr><td align="left">&#x000a0;15&#x02013;19</td><td char="." align="char">70</td><td char="." align="char">39</td><td char="." align="char">1.00</td><td align="left"/></tr><tr><td align="left">&#x000a0;20&#x02013;24</td><td char="." align="char">106</td><td char="." align="char">72</td><td align="left">0.82 (0.50&#x02013;1.34)</td><td align="left">0.50 (0.26&#x02013;0.99)</td></tr><tr><td align="left">&#x000a0;25&#x02013;29</td><td char="." align="char">116</td><td char="." align="char">57</td><td align="left">1.13 (0.69&#x02013;1.87)</td><td align="left">0.55 (0.23&#x02013;1.12)</td></tr><tr><td align="left">&#x000a0;30&#x02013;34</td><td char="." align="char">116</td><td char="." align="char">65</td><td align="left">0.99 (0.61&#x02013;1.63)</td><td align="left">0.51 (0.24&#x02013;1.09)</td></tr><tr><td align="left">&#x000a0;&#x0003e;35</td><td char="." align="char">140</td><td char="." align="char">39</td><td align="left">2.00 (1.18&#x02013;3.39)*</td><td align="left">1.03 (0.48&#x02013;2.21)</td></tr><tr><td align="left" colspan="5">Marital status</td></tr><tr><td align="left">&#x000a0;Single</td><td char="." align="char">17</td><td char="." align="char">17</td><td char="." align="char">1.00</td><td align="left"/></tr><tr><td align="left">&#x000a0;Married</td><td char="." align="char">492</td><td char="." align="char">214</td><td align="left">2.30 (1.15&#x02013;4.59)*</td><td align="left">1.50 (0.62&#x02013;3.63)</td></tr><tr><td align="left">&#x000a0;Separated due to work</td><td char="." align="char">13</td><td char="." align="char">14</td><td align="left">0.93 (0.34&#x02013;2.55)</td><td align="left">0.79 (0.24&#x02013;2.58)</td></tr><tr><td align="left">&#x000a0;Divorced</td><td char="." align="char">15</td><td char="." align="char">18</td><td align="left">0.83 (0.32&#x02013;2.18)</td><td align="left">1.10 (0.35&#x02013;3.45)</td></tr><tr><td align="left">&#x000a0;Windowed</td><td char="." align="char">11</td><td char="." align="char">9</td><td align="left">1.22 (0.40&#x02013;3.70)</td><td align="left">1.22 (0.33&#x02013;4.48)</td></tr><tr><td align="left" colspan="5">Place of residence</td></tr><tr><td align="left">&#x000a0;Rural</td><td char="." align="char">449</td><td char="." align="char">247</td><td char="." align="char">1.00</td><td align="left"/></tr><tr><td align="left">&#x000a0;Urban</td><td char="." align="char">99</td><td char="." align="char">25</td><td align="left">2.18 (1.37&#x02013;3.47)**</td><td align="left">2.68 (1.45&#x02013;4.98)**</td></tr><tr><td align="left" colspan="5">Distance from health institution (h)</td></tr><tr><td align="left">&#x000a0;&#x0003c;1</td><td char="." align="char">183</td><td char="." align="char">32</td><td align="left">2.10 (1.41&#x02013;3.14)**</td><td align="left">0.95 (0.56&#x02013;1.62)</td></tr><tr><td align="left">&#x000a0;1&#x02013;2</td><td char="." align="char">283</td><td char="." align="char">105</td><td align="left">2.66 (1.82&#x02013;3.89)**</td><td align="left">2.21 (1.39&#x02013;3.51)**</td></tr><tr><td align="left">&#x000a0;&#x0003e;2</td><td char="." align="char">82</td><td char="." align="char">81</td><td char="." align="char">1.00</td><td align="left"/></tr><tr><td align="left" colspan="5">Number of children born alive</td></tr><tr><td align="left">&#x000a0;None</td><td char="." align="char">30</td><td char="." align="char">29</td><td char="." align="char">1.00</td><td align="left"/></tr><tr><td align="left">&#x000a0;1&#x02013;5</td><td char="." align="char">294</td><td char="." align="char">143</td><td align="left">1.99 (1.15&#x02013;3.44)*</td><td align="left">2.11 (0.97&#x02013;4.59)</td></tr><tr><td align="left">&#x000a0;&#x0003e;5</td><td char="." align="char">224</td><td char="." align="char">100</td><td align="left">2.17 (1.23&#x02013;3.80)*</td><td align="left">1.83 (0.77&#x02013;4.35)</td></tr><tr><td align="left" colspan="5">Know about maternal health service</td></tr><tr><td align="left">&#x000a0;Yes</td><td char="." align="char">279</td><td char="." align="char">97</td><td align="left">1.87 (1.39&#x02013;2.52)**</td><td align="left">1.38 (0.97&#x02013;1.96)</td></tr><tr><td align="left">&#x000a0;No</td><td char="." align="char">269</td><td char="." align="char">175</td><td char="." align="char">1.00</td><td align="left"/></tr><tr><td align="left" colspan="5">ANC visit</td></tr><tr><td align="left">&#x000a0;Not used</td><td char="." align="char">41</td><td char="." align="char">74</td><td char="." align="char">1.00</td><td align="left"/></tr><tr><td align="left">&#x000a0;Once</td><td char="." align="char">106</td><td char="." align="char">76</td><td align="left">2.52 (1.55&#x02013;4.08)**</td><td align="left">1.36 (0.76&#x02013;2.45)</td></tr><tr><td align="left">&#x000a0;Twice</td><td char="." align="char">253</td><td char="." align="char">80</td><td align="left">5.71 (3.61&#x02013;9.01)**</td><td align="left">2.36 (1.31&#x02013;4.23)*</td></tr><tr><td align="left">&#x000a0;There and more</td><td char="." align="char">148</td><td char="." align="char">42</td><td align="left">6.36 (3.81&#x02013;10.62)**</td><td align="left">2.60 (1.40&#x02013;5.06)*</td></tr><tr><td align="left" colspan="5">Know about PNC service</td></tr><tr><td align="left">&#x000a0;Yes</td><td char="." align="char">497</td><td char="." align="char">195</td><td align="left">3.85 (2.60&#x02013;5.69)**</td><td align="left">1.72 (1.03&#x02013;2.86)**</td></tr><tr><td align="left">&#x000a0;No</td><td char="." align="char">51</td><td char="." align="char">77</td><td char="." align="char">1.00</td><td align="left"/></tr><tr><td align="left" colspan="5">Visit by community health agent/HEW</td></tr><tr><td align="left">&#x000a0;Yes</td><td char="." align="char">410</td><td char="." align="char">146</td><td align="left">2.56 (1.89&#x02013;3.48)**</td><td align="left">2.66 (1.39&#x02013;5.06)**</td></tr><tr><td align="left">&#x000a0;No</td><td char="." align="char">138</td><td char="." align="char">126</td><td char="." align="char">1.00</td><td align="left"/></tr><tr><td align="left" colspan="5">History of institutional delivery</td></tr><tr><td align="left">&#x000a0;Yes</td><td char="." align="char">77</td><td char="." align="char">4</td><td align="left">10.95 (3.97&#x02013;30.26)**</td><td align="left">8.09 (2.78&#x02013;23.53)**</td></tr><tr><td align="left">&#x000a0;No</td><td char="." align="char">471</td><td char="." align="char">268</td><td align="left"/><td align="left"/></tr><tr><td align="left" colspan="5">Perception towards PNC</td></tr><tr><td align="left">&#x000a0;Positive</td><td char="." align="char">420</td><td char="." align="char">189</td><td align="left">1.44 (1.04&#x02013;1.99)*</td><td align="left">1.10 (0.72&#x02013;1.68)</td></tr><tr><td align="left">&#x000a0;Negative</td><td char="." align="char">128</td><td char="." align="char">83</td><td align="left"/><td align="left"/></tr><tr><td align="left" colspan="5">Ability to make decisions for utilization</td></tr><tr><td align="left">&#x000a0;Yes</td><td char="." align="char">196</td><td char="." align="char">152</td><td align="left">2.28 (1.69&#x02013;3.06)**</td><td align="left">1.86 (1.31&#x02013;2.65)**</td></tr><tr><td align="left">&#x000a0;No</td><td char="." align="char">352</td><td char="." align="char">120</td><td char="." align="char">1.00</td><td align="left"/></tr></tbody></table><table-wrap-foot><p>1.00&#x000a0;=&#x000a0;Referent category, *&#x000a0;statistically associated <italic>p</italic> value &#x0003c;0.05, **&#x000a0;statistically associated <italic>p</italic> value &#x0003c;0.001</p></table-wrap-foot></table-wrap> </p></sec><sec id="Sec13"><title>Results of the Focus Group Discussion</title><p>Focus group discussions were conducted in the Chira Mantebro, Des Denzez, and Degoma wards in order to triangulate the quantitative data. The qualitative data responses of the FDGs were grouped into three themes: awareness about PNC, perception towards the care, and factors associated with the utilization of PNC.</p></sec><sec id="Sec14"><title>Awareness of Post Natal Care Service</title><p>The majority of mothers demonstrated an awareness of immunization, family planning, and counseling on nutrition. Mothers reported that HEWs and community health agents informed them of the existence of PNC services. However, those who knew about the services did not have adequate information on when post-natal clinics are offered, or for whom. Most mothers assumed that the services were only given for children and vaccination 45&#x000a0;days after birth.<disp-quote><p>&#x02018;&#x02026;Community health workers inform us about the vaccination and child nutrition but we didn&#x02019;t practice because there is no full service here, the health facility is too far from here, and we do not have financial power&#x02026;&#x02019; (A 32&#x000a0;year-old mother, focused group discussion, Chira Mantebro ward).</p></disp-quote> </p><p>Similarly health extension works reported the existence of awareness on PNC among mothers. A HEA said: <italic>&#x02018;&#x02026;we usually told mothers about PNC services including immunization, counseling about breastfeeding, nutrition, and family planning however most mothers do have better awareness about vaccination&#x02026;</italic>&#x02019;<disp-quote><p>&#x02018;&#x02026; We usually inform the mothers on PNC services, and they do have an awareness about the service but most of them need to be motivated every time to use the service&#x02026;&#x02019; <italic>(</italic>A community health worker, focused group discussion).</p></disp-quote> </p></sec><sec id="Sec15"><title>Mothers&#x02019; Perceptions Toward PNC</title><p>Most participants in the discussions had a positive perception toward PNC and they encourage others to use PNC.<disp-quote><p> <italic>&#x02018;&#x02026;</italic> Health extension workers gave a drop to my child 15&#x000a0;days after I deliver at home and it helps to my child. I went to the health post before 40&#x000a0;days for vaccination&#x02026; it prevents my child form different diseases hence I want to have this service for my child to be healthy&#x02019; (A 25&#x000a0;year-old mother, focused group discussion, Des Denzaz Ward).</p></disp-quote> </p><p>In harmony with the mother, a HEA stated <italic>&#x02018;&#x02026;most mothers in this ward posses positive feelings towards PNC services yet remoteness of the area prevent some mothers not to have the service&#x02026;&#x02019;&#x02019;.</italic> <disp-quote><p> <italic>&#x02018;&#x02026; Some mothers need to use PNC and they are happy to go to the health post for the service&#x02026;&#x02019;</italic> (A community health worker)</p></disp-quote> </p></sec><sec id="Sec16"><title>Factors Associated with the Utilization of Postnatal Care</title><p>Distance from health organization was a major problem, especially in remote, rural wards, some participants complain that they needed to walk on foot for up to 2&#x000a0;h to reach the nearest health center.<disp-quote><p>&#x02018;&#x02026; A community health worker educate us about vaccination and how to care for our children but we fail to do because health post is too far and our husbands mostly did not allow us&#x02026;&#x02019; (A 24&#x000a0;year-old mother, focused group discussion, Degoma Ward).</p></disp-quote> </p><p>Congruent evidence was also gathered from a health extension workers: &#x02018;&#x02026; <italic>there are village very far from here with no health extension workers so for such areas voluntary members of a society gave training and some services to mothers, but when we go for vaccination every month it is difficult to say they are getting the care &#x02026;&#x02019;</italic> </p><p>Most women complained limited availability of health services (equipment and drugs): mainly in remote areas vaccines are less available. <italic>&#x02018;&#x02026; When my child gets sick there were no drug at the health post and full service even if they informed us about the care&#x02026;&#x02019;</italic> (A 20&#x000a0;year-old mother, focused group discussion, Des Degoma Ward). &#x02018;&#x02026; <italic>Only they vaccinate our children in our ward, there was no satisfactory service here and there was no drug &#x02026;&#x02019; (</italic>A 26&#x000a0;year-old mother, focused group discussion, Des Denzaz Ward).</p><p>Another constraint mentioned by mothers was absent or frequent travel of HEAs out of the ward. <italic>&#x02018;&#x02026; I gave birth before three months but when I went there after a first month for vaccination the health worker were not available&#x02026;&#x02019;</italic> (A 32&#x000a0;year-old mother, focused group discussion, Chira Mantebro Ward).</p><p>Most mothers are responsible for house work such as baby care, preparation of food, and farming in rainy season which will delay their receiving of PNC services.</p><p> <italic>&#x02018;&#x02026; Community health worker told us about vaccination and counseled about nutrition but we have a work load to practice it&#x02026;&#x02019;</italic> (A 22&#x000a0;year-old mother, focused group discussion, Des Denzaz Ward).</p><p>Some participants did not feel the need for PNC services unless their children and they were sick after delivery. &#x02018;&#x02026;<italic>Since I did not get sick I did not go to the health post and never used family planning&#x02026;&#x02019;</italic> (A 27&#x000a0;year-old mother, focused group discussion, Degoma Ward).</p></sec></sec><sec id="Sec17"><title>Discussion</title><p>This community based study assessed knowledge, perception, and utilization of postpartum health care among women who give birth in the last year in Gondar Zuria District. The result showed that among the 820 postpartum women, 548 (66.83&#x000a0;%) obtained PNC during the 6&#x000a0;weeks following delivery. This is high compared with studies done in four other regions of Ethiopia (Amhara, Oromiya, Southern Nations, Nationalities and People&#x02019;s, and Tigray) (10&#x000a0;%) [<xref ref-type="bibr" rid="CR15">15</xref>], Sidama zone (37.2&#x000a0;%) [<xref ref-type="bibr" rid="CR16">16</xref>], Uganda (58&#x000a0;%) [<xref ref-type="bibr" rid="CR7">7</xref>], Nepal (34&#x000a0;%) [<xref ref-type="bibr" rid="CR17">17</xref>], and Palestine (36.6&#x000a0;%) [<xref ref-type="bibr" rid="CR18">18</xref>]. The difference may be attributed to time, place, and social context variation between the present study and previous studies.</p><p>In Ethiopia PNC coverage was 34.3&#x000a0;% and, in the case of Amhara region 44.5&#x000a0;%, which is lower than from the present finding [<xref ref-type="bibr" rid="CR6">6</xref>]. The variation could be due to socio-demographic characteristics, methodology difference, implementation of the service, and accessibility of health organizations.</p><p>The government of Ethiopia endorsed a strategy which is aimed at strengthening the health system to provide quality care, particularly skilled attendance at birth and emergency obstetrical care through a functional referral system that includes zonal hospitals and four health centers that refer to it as well as health posts [<xref ref-type="bibr" rid="CR19">19</xref>]. Through the Health Extension Program, the government plans to extend primary healthcare to the rural poor through deployment of about 30,000 government-salaried HEAs, two per kebele. Kebele Councils with Woreda Councils recruit young, locally resident women who have completed grade 10 and speak the local language to become Health Extension Workers (HEWs). Two HEWs are posted at a health post for a population of approximately 5,000; they are to spend 75&#x000a0;% of their time in outreach activities, teaching by example through three approaches: model families (40&#x02013;60 families who are early adopters of desirable health practices), community organizations (e.g., Idir, Ekub, Mahber), and health post and outreach service delivery [<xref ref-type="bibr" rid="CR20">20</xref>].</p><p>The majority (84.39&#x000a0;%) of mothers were aware that they were supposed to receive PNC services after delivery which is consistent with the studies done in Uganda (70.3&#x000a0;%) [<xref ref-type="bibr" rid="CR7">7</xref>] and in Ethiopia [<xref ref-type="bibr" rid="CR21">21</xref>]. Findings from the FGDs also showed that mothers had awareness and usually utilize the service 40&#x000a0;days after delivery. This high rate of awareness could be due to the range of government and non-government programs involved in the distribution of health information to the women such as the Integrated Family Health Program (IFPH) trained CHWs/agents for motivation of mothers to use maternal health services.</p><p>Postpartum maternal health care service awareness was associated with a previous visit by community health agents. Women who had antenatal follow up were 1.72 times more likely to be aware than women who did not have antenatal follow up. Similar studies in India show that mothers who attend antenatal care (64.4&#x000a0;%) were more aware than women who did not have (33.3&#x000a0;%) antenatal care [<xref ref-type="bibr" rid="CR22">22</xref>]. It&#x02019;s also similar to studies conducted in Indonesia, Nepal and Uganda [<xref ref-type="bibr" rid="CR7">7</xref>, <xref ref-type="bibr" rid="CR17">17</xref>, <xref ref-type="bibr" rid="CR23">23</xref>]. The possible explanation might be that mothers receive health education and counseling during community and antenatal visits.</p><p>The majority of mothers (74.27&#x000a0;%) perceived that PNC is helpful to mothers, and children&#x02019;s health, but only 66.83&#x000a0;% of mothers utilize PNC services. Those mothers who had awareness about PNC, who were previously visited by community health agents, and who had antenatal follow-up have most likely considered the care as important for maintaining their health and for their child&#x02019;s health.</p><p>The most frequently cited reasons mentioned by the FGD participants for not utilizing PNC were believing that the treatments were not important unless mothers feel sick, negative experiences of women with the care, and considering the service accessible only for the child. Similar studies in Sweden show that women who had negative experiences were avoid PNC [<xref ref-type="bibr" rid="CR24">24</xref>] other studies in Indonesia and the West Bank show that women in those areas have similar reasons to our study population for neglecting to utilize PNC [<xref ref-type="bibr" rid="CR18">18</xref>, <xref ref-type="bibr" rid="CR23">23</xref>].</p><p>Place of residence, distance from a health organization, antenatal follow up, previous visit by community agents, and the ability to make a decision had effect on utilization of PNC service in this study. The participants in FGDs also raised these issues and stated that they affected the utilization of the service. Other studies on postpartum mothers revealed matching factors that have consequences for PNC utilization [<xref ref-type="bibr" rid="CR17">17</xref>, <xref ref-type="bibr" rid="CR18">18</xref>, <xref ref-type="bibr" rid="CR22">22</xref>, <xref ref-type="bibr" rid="CR25">25</xref>&#x02013;<xref ref-type="bibr" rid="CR27">27</xref>].</p><p>In Bangladesh, urban mothers were more likely to receive PNC than rural mothers from both health professionals and non-health professionals [<xref ref-type="bibr" rid="CR26">26</xref>]. Similarly, in this study, urban mothers were 2.68 times more likely to receive PNC than rural mothers (AOR 2. 68; 95&#x000a0;% CI 1.45&#x02013;4.98). This might be due to the fact that urban women may have information from different sources on PNC or because of the availability of a good number of health institutions in urban areas.</p><p>Distance from the health institution remains a major problem as shown in previous literature; utilization of health services is strongly associated with access to health services [<xref ref-type="bibr" rid="CR21">21</xref>, <xref ref-type="bibr" rid="CR23">23</xref>, <xref ref-type="bibr" rid="CR27">27</xref>&#x02013;<xref ref-type="bibr" rid="CR29">29</xref>]. In this study, one to 2&#x000a0;h travel to the nearby health centers resulted in women being 2.21 times more likely utilize the service than those who travel above 2&#x000a0;h. Comparable evidence was also gathered from FDG participants. This may be inferred to the cost of travel in terms of time, money, or energy. Mothers (71.91&#x000a0;%) who had antenatal care and mothers (73.74&#x000a0;%) who were previously visited by community health agents were more effective in utilizing PNC than those who had never been in any contact with the health system. This is in line with studies in Syria, Indonesia, Nepal, and Uganda [<xref ref-type="bibr" rid="CR9">9</xref>, <xref ref-type="bibr" rid="CR18">18</xref>, <xref ref-type="bibr" rid="CR23">23</xref>, <xref ref-type="bibr" rid="CR30">30</xref>]. This could be due to the awareness of the mothers on possible postnatal complications as a result of previous contact with healthcare workers.</p><p>The ability to make a decision had a significant association with utilization of postpartum care. Mothers who can make a decision were more likely to use the service than those who cannot make a decision (AOR 1. 86; 95&#x000a0;% CI 1.31&#x02013;2.65). Studies indicate autonomy of decision making as a factor in the utilization of maternal health services [<xref ref-type="bibr" rid="CR9">9</xref>, <xref ref-type="bibr" rid="CR29">29</xref>, <xref ref-type="bibr" rid="CR31">31</xref>]. The possible reason for mothers not to make a decision might be the community belief about the hierarchy of authority in the household and economic dependency of mothers on husbands. A contradictory result with the current research was identified in a study in Ethiopia [<xref ref-type="bibr" rid="CR32">32</xref>]. This may be attributed to the prevalence of paternalism in the study area.</p><p>Many findings in the literature indicate a significant association of PNC utilization with the occupation of women, education of women, and the husbands&#x02019; occupation [<xref ref-type="bibr" rid="CR9">9</xref>, <xref ref-type="bibr" rid="CR17">17</xref>, <xref ref-type="bibr" rid="CR18">18</xref>, <xref ref-type="bibr" rid="CR33">33</xref>]. On the other hand, this study revealed no significant association with these variables. The majority of the respondents were illiterate, housewives and farmers which may be contributing to the lack of association in the study area.</p></sec><sec id="Sec18"><title>Conclusion</title><p>The majority of the mothers had an awareness of PNC services but they did not know when they should seek those services. From the result it can be concluded that mothers&#x02019; awareness about PNC service is more focused on the immunization component than others. Most mothers have a positive perception toward PNC services; however, mothers in a rural area possess a negative perception. PNC utilization was high but utilization of the most crucial elements was very low and large segments of mothers utilize only immunization services. Place of residence, distance from health institution, antenatal follow-up, previous visit by community health agents, and the ability to make decisions were significant factors that influence utilization of postnatal services.</p></sec>

Breastfeeding Practices in Relation to Country of Origin Among Women Living in Denmark: A Population-Based Study

Could not extract abstract

<contrib contrib-type="author"><name><surname>Busck-Rasmussen</surname><given-names>Marianne</given-names></name><address><email>marianne_busck@yahoo.dk</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author" corresp="yes"><name><surname>Villadsen</surname><given-names>Sarah Fredsted</given-names></name><address><phone>+45-35-327962</phone><fax>+45-61-333455</fax><email>sfv@sund.ku.dk</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author"><name><surname>Norsker</surname><given-names>Filippa Nyboe</given-names></name><address><email>fino@sund.ku.dk</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author"><name><surname>Mortensen</surname><given-names>Laust</given-names></name><address><email>lamo@sund.ku.dk</email></address><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author"><name><surname>Andersen</surname><given-names>Anne-Marie Nybo</given-names></name><address><email>amny@sund.ku.dk</email></address><xref ref-type="aff" rid="Aff1"/></contrib><aff id="Aff1">Section of Social Medicine, Department of Public Health, University of Copenhagen, Oster Farimagsgade 5, Box 2099, 1014 Copenhagen K, Denmark </aff>

Maternal and Child Health Journal

<sec id="Sec1"><title>Introduction</title><p>The beneficial effect of breastfeeding on infant health is well described, and WHO recommends at least 6&#x000a0;months of exclusive breastfeeding for children globally [<xref ref-type="bibr" rid="CR1">1</xref>]. One of the most important short term effects of breastfeeding is the protection against infectious diseases, and breastfed infants have been demonstrated to have lower prevalence of inflammatory bowel diseases, childhood cancers, and type-1 diabetes [<xref ref-type="bibr" rid="CR2">2</xref>]. Breastfed children have been shown to have improved cognitive performance [<xref ref-type="bibr" rid="CR3">3</xref>&#x02013;<xref ref-type="bibr" rid="CR5">5</xref>], while the effects on overweight and obesity are uncertain [<xref ref-type="bibr" rid="CR4">4</xref>, <xref ref-type="bibr" rid="CR6">6</xref>].</p><p>Women&#x02019;s decision and ability to breastfeed are affected by their social context. In many western countries are young maternal age and poor socio-economic position risk factors of no initiation and short duration of breastfeeding [<xref ref-type="bibr" rid="CR7">7</xref>&#x02013;<xref ref-type="bibr" rid="CR10">10</xref>]. However, the influence of socio-demographic parameters on breastfeeding might differ between societies. Studies from countries in Western Europe and the US have shown that breastfeeding frequencies are higher among migrant women than in the native populations [<xref ref-type="bibr" rid="CR11">11</xref>&#x02013;<xref ref-type="bibr" rid="CR14">14</xref>]. It seems that with acculturation, measured by years of residency in the new country breastfeeding prevalences decline and, likewise, first generation descendants of migrants breastfeed less than the women who migrated [<xref ref-type="bibr" rid="CR11">11</xref>, <xref ref-type="bibr" rid="CR14">14</xref>]. A Swedish study reported that there were no ethnic differences in breastfeeding at 6&#x000a0;months, while more immigrant mothers continued breastfeeding until at least 12&#x000a0;months [<xref ref-type="bibr" rid="CR9">9</xref>].</p><p>In Denmark, breastfeeding initiation and continuation of full breastfeeding in 4&#x000a0;months is high, 95 and 60&#x000a0;%, respectively [<xref ref-type="bibr" rid="CR15">15</xref>]. The national recommendations are full breastfeeding for 6&#x000a0;months, but the guidelines acknowledge that some children will be ready for complementary feeding after 4&#x000a0;months. Further, partial breastfeeding is recommended to continue until at least 12&#x000a0;months [<xref ref-type="bibr" rid="CR16">16</xref>]. At the international level WHO and UNICEF have encouraged breastfeeding with the <italic>Baby Friendly Hospital Initiative</italic> [<xref ref-type="bibr" rid="CR17">17</xref>] that has set guidelines for breastfeeding support at health facilities, including: not giving formulas (unless medically indicated) and to help the mothers initiate breastfeeding within one-half hour of birth [<xref ref-type="bibr" rid="CR18">18</xref>]. In Denmark, additional support for breastfeeding is offered to all newborns and their mothers at home visits where health visitors (specially trained nurses) after delivery assist the family in ensuring the well-being of the child. Unfortunately, no official statistics are available, but the general understanding among leaders of this system and the authors&#x02019; experience are that the health visitors are widely accepted and used by the large majority (more than 95&#x000a0;%) of the families, including families with migrant background.</p><p>Breastfeeding practices among migrants have scarcely been studied in Denmark. A single study has described the breastfeeding prevalence as lower in migrant women than in women of Danish origin [<xref ref-type="bibr" rid="CR15">15</xref>]. However, in this study country of origin was not taken into account. It has been shown that reproductive and child health are diverging among the different migrant groups in Denmark [<xref ref-type="bibr" rid="CR19">19</xref>, <xref ref-type="bibr" rid="CR20">20</xref>], and to understand the mechanisms behind the potential increased vulnerability of the migrant groups, it is important not to treat migrants as one entity. Therefore, in order to guide future preventive strategies, we find it justified that a better understanding of patterns in breastfeeding prevalence and the mechanisms behind will be obtained using a country of origin approach. The objective of this study was to assess the risk of suboptimal breastfeeding among groups of migrant women as compared to women of Danish origin, and to describe how measures of socioeconomic position, acculturation and breastfeeding support affect these associations.</p></sec><sec id="Sec2"><title>Materials and Methods</title><sec id="Sec3"><title>The Danish Health Visitor&#x02019;s Child Health Database</title><p>The Danish Health Visitor&#x02019;s Child Health Database was established by the municipalities in the county of Copenhagen in 2002 with the aims of documentation and monitoring of infant health and the work of municipality-based health visitors. All health visitors in 18 municipalities provided data to the database during the period 2002&#x02013;2009: ten municipalities participated during the total period, the remaining during one or more of the years. The database encompasses information on 47,869 infants (corresponding to more than 90&#x000a0;% of the children born in the municipalities) that had at least one health visit during the period of coverage.</p><p>The data on breastfeeding stem from a standardized systematic registration file for every infant, kept by the health visitors who recorded details on the child, the family and service provided by the health visitors during the first year of life [<xref ref-type="bibr" rid="CR15">15</xref>]. From the same data source we used health visitor registered information about maternal parity (categorized as 1, 2, and 2+), and a dichotomous indication of whether the newborn was put to the mothers&#x02019; breast within 2&#x000a0;h of birth (yes/no) and whether the newborn was given formula during the stay at hospital after birth (yes/no). The latter two variables were registered by the health visitor at first visit after birth as indicators of breastfeeding support, as these are two of the Baby Friendly Hospital Initiative ten steps to successful breastfeeding [<xref ref-type="bibr" rid="CR18">18</xref>]. The health visitors also registered duration of full breastfeeding, defined as age until which the infant was exclusively given breast milk, supplemented only by water or at maximum one meal of formula per week. This information was used to establish the main outcome of interest in the study: full breastfeeding until the age of at least 4&#x000a0;months (yes/no).</p></sec><sec id="Sec4"><title>Socio-Demographic Data from Statistics Denmark</title><p>From the Population Registry in Statistics Denmark we obtained information on maternal and paternal unique personal identification number (PIN) that, beside carrying information about the year of birth for the mother, served as key variable for all linkages. For women born outside Denmark, country of origin and year of immigration were also obtained. Maternal age was categorized in four strata (&#x0003c;25, 25&#x02013;29, 30&#x02013;34, and 35+&#x000a0;years). From year of immigration, maternal year of birth, and infant year of birth we calculated maternal age at migration to Denmark (&#x0003c;10, 10&#x02013;18, &#x0003e;18&#x000a0;years) and number of years lived in Denmark before delivery (&#x0003c;10&#x000a0;years, 10+ years). From the Integrated Database for Labour Market Research at Statistics Denmark we obtained information about maternal and paternal educational level (recoded from ISCED to &#x0003c;10, 10&#x02013;12 and &#x0003e;12&#x000a0;years of formal education), maternal and paternal attachment to labour market, dichotomized into being in work (having paid work more than 50&#x000a0;% of the year, for mothers the calendar year before birth, for fathers in the calendar year of birth) and out of work, and, finally, household income (sum of maternal and paternal disposable income) calculated into percentiles of the study population born in the same year and categorized in five strata (0&#x02013;5, 6&#x02013;24, 25&#x02013;49, 50&#x02013;74, and 75&#x02013;100&#x000a0;%) of which the two first groups represent the quartile with the lowest income.</p></sec><sec id="Sec5"><title>Country of Origin</title><p>We used the available data in Statistics Denmark&#x02019;s regarding ethnicity to determine maternal country of origin: The country of origin is <italic>Danish</italic> if (at least) one parent is born in Denmark <italic>and</italic> has Danish citizenship. <italic>Migrants</italic> are born outside Denmark and none of the parents of a migrant is born in Denmark <italic>and</italic> has Danish citizenship. For migrants, the country of origin is the same as the country of birth. <italic>Descendants</italic> are born in Denmark and none of the parents of a descendant is born in Denmark and has Danish citizenship. A child born of two descendants with non-Danish citizenship is also classified as a descendant. For descendants, the country of origin is defined as country of citizenship. If maternal and paternal country of birth and citizenship are different and not Danish, country of origin is defined based on maternal country of birth and citizenship. As follows from this description, Danish citizenship is not obtained automatically by virtue of birth in the country if the parents are non-Danish citizens and double citizenship is not allowed. Citizenship can be obtained after application in a relatively complicated procedure.</p></sec><sec id="Sec6"><title>Study Population</title><p>The study population was grouped according to maternal country of origin and only groups that exceeded 250 individuals in the database were included.</p><p>Of the 47,869 infants in the Danish Health Visitor&#x02019;s Child Health Database, 986 (2.1&#x000a0;%) were excluded due to missing information on PIN-number, 1,044 (2.2&#x000a0;%) had no information about country of origin in the Population Registry, and 3,410 (7.1&#x000a0;%) had a country of birth did not fulfill the inclusion criteria. Thus, 42,420 infants were included.</p></sec><sec id="Sec7"><title>Statistical Analyses</title><p>Distributions of parental socio-demographic variables and indicators of breastfeeding support according to maternal counties of origin were cross tabulated and the proportion of infants, who were fully breastfed, was calculated according to maternal country of origin and stratified on parental characteristics.</p><p>The crude odds ratios with 95&#x000a0;% confidence intervals for suboptimal breastfeeding according to country of origin were estimated using logistic regression with women of Danish origin as the reference group. Analyses exploring the relation between strata of socio-demographic characteristics and maternal country of origin on the risk of suboptimal breastfeeding were estimated in logistic regression models using the most prevalent value of the characteristic among mothers of Danish origin as the reference.</p><p>The risk of suboptimal breastfeeding according to maternal country of origin was furthermore estimated in a multivariable logistic regression model adjusted for maternal age and parity and all socio-demographic variables.</p></sec><sec id="Sec8"><title>Approvals</title><p>The linkage of data from the Danish Health Visitor&#x02019;s Child Health Database and data from Statistics Denmark was approved by the steering committee for the Danish Health Visitor&#x02019;s Child Health Database and the Danish Data Protection Agency.</p></sec></sec><sec id="Sec9"><title>Results</title><p>Of the 42,420 infants, 87&#x000a0;% had mothers of Danish origin. Among the mothers of non-Danish origin, Turkish origin was the most frequent (4.9&#x000a0;%), followed by mothers of Pakistani origin (2.3&#x000a0;%). Other maternal countries of origin were Former Yugoslavia (including Serbia, Montenegro, Croatia, Bosnia and Hercegovina, Macedonia, Kosovo and Slovenia), Iraq, Morocco, Lebanon (including Palestine), and Afghanistan. The majority of mothers of non-Danish origin were migrants themselves, though, 20&#x02013;30&#x000a0;% of mothers with Turkish, Pakistani, Former Yugoslavian and Moroccan origin were descendants, i.e. born in Denmark of women, who migrated to Denmark.</p><p>In general, the infants of Danish and other Nordic origin were more likely to be first or second born, their mothers were older than mothers of non-Nordic origin, their parents had a longer education and were more likely to be active participants at the labour market (Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>). The household income were unequally distributed, with more than 75&#x000a0;% of the infants of Iraqi, Lebanese/Palestinian, and Afghan origin born in families in the lowest income quartile. Information about full breastfeeding until the age of 4&#x000a0;months was missing on a substantial proportion of the infants. However, missing information on breastfeeding did not appear to depend on maternal country of origin or social and demographic characteristics (data not shown).<table-wrap id="Tab1"><label>Table&#x000a0;1</label><caption><p>Selected characteristics of the study population (distributions in percentages). All children seen by health visitor in 18 Danish municipalities, 2002&#x02013;2009</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="2"/><th align="left" colspan="9">Maternal country of origin</th></tr><tr><th align="left">Denmark</th><th align="left">Turkey</th><th align="left">Pakistan</th><th align="left">Former Yugoslavia</th><th align="left">Iraq</th><th align="left">Morocco</th><th align="left">Lebanon/Palestine</th><th align="left">Afghanistan</th><th align="left">Other Nordic</th></tr></thead><tbody><tr><td align="left">N&#x000a0;=&#x000a0;42,420</td><td align="left">36,899</td><td align="left">2,091</td><td align="left">992</td><td align="left">729</td><td align="left">442</td><td align="left">292</td><td align="left">262</td><td align="left">266</td><td align="left">447</td></tr><tr><td align="left">Descendants of migrants (%)</td><td char="." align="char"/><td char="." align="char">24.1</td><td char="." align="char">29.2</td><td char="." align="char">23.0</td><td char="." align="char">0.5</td><td char="." align="char">19.9</td><td char="." align="char">5.7</td><td char="." align="char">0.0</td><td align="left">13.4</td></tr><tr><td align="left" colspan="10"> <italic>Maternal age at birth</italic> </td></tr><tr><td align="left">&#x0003c;25&#x000a0;years</td><td char="." align="char">9.9</td><td char="." align="char">26.7</td><td char="." align="char">18.5</td><td char="." align="char">23.7</td><td char="." align="char">26.7</td><td char="." align="char">14.7</td><td char="." align="char">37</td><td char="." align="char">25.2</td><td char="." align="char">9.2</td></tr><tr><td align="left">25&#x02013;29&#x000a0;years</td><td char="." align="char">27.9</td><td char="." align="char">40.3</td><td char="." align="char">44.6</td><td char="." align="char">41.8</td><td char="." align="char">31.7</td><td char="." align="char">33.2</td><td char="." align="char">29.4</td><td char="." align="char">35.3</td><td char="." align="char">22.1</td></tr><tr><td align="left">30&#x02013;34&#x000a0;years</td><td char="." align="char">38.8</td><td char="." align="char">22.2</td><td char="." align="char">26.6</td><td char="." align="char">24.3</td><td char="." align="char">24</td><td char="." align="char">28.4</td><td char="." align="char">21.4</td><td char="." align="char">24.4</td><td char="." align="char">41.4</td></tr><tr><td align="left">35+ years</td><td char="." align="char">23.5</td><td char="." align="char">10.2</td><td char="." align="char">10.4</td><td char="." align="char">10.2</td><td char="." align="char">17.7</td><td char="." align="char">23.6</td><td char="." align="char">12.2</td><td char="." align="char">15.0</td><td char="." align="char">27.3</td></tr><tr><td align="left" colspan="10"> <italic>Maternal parity</italic> </td></tr><tr><td align="left">1</td><td char="." align="char">48.8</td><td char="." align="char">36.8</td><td char="." align="char">32.5</td><td char="." align="char">42.5</td><td char="." align="char">37.0</td><td char="." align="char">32.9</td><td char="." align="char">40.0</td><td char="." align="char">38.7</td><td char="." align="char">48.7</td></tr><tr><td align="left">2</td><td char="." align="char">36.4</td><td char="." align="char">30.9</td><td char="." align="char">29.9</td><td char="." align="char">36.7</td><td char="." align="char">29.1</td><td char="." align="char">28.0</td><td char="." align="char">21.4</td><td char="." align="char">29.7</td><td char="." align="char">34.0</td></tr><tr><td align="left">3+</td><td char="." align="char">14.9</td><td char="." align="char">32.3</td><td char="." align="char">37.6</td><td char="." align="char">20.8</td><td char="." align="char">34.0</td><td char="." align="char">39.4</td><td char="." align="char">39.1</td><td char="." align="char">31.5</td><td char="." align="char">17.7</td></tr><tr><td align="left">Missing</td><td char="." align="char">15.4</td><td char="." align="char">14.8</td><td char="." align="char">13.7</td><td char="." align="char">12.2</td><td char="." align="char">19.2</td><td char="." align="char">13.0</td><td char="." align="char">16.0</td><td char="." align="char">16.5</td><td char="." align="char">16.3</td></tr><tr><td align="left" colspan="10"> <italic>Maternal educational level</italic> </td></tr><tr><td align="left">&#x0003c;10&#x000a0;years</td><td char="." align="char">14.9</td><td char="." align="char">58.5</td><td char="." align="char">38.3</td><td char="." align="char">34.6</td><td char="." align="char">47.6</td><td char="." align="char">49.2</td><td char="." align="char">62.9</td><td char="." align="char">51.1</td><td char="." align="char">7.5</td></tr><tr><td align="left">10&#x02013;12&#x000a0;years</td><td char="." align="char">39.4</td><td char="." align="char">31.8</td><td char="." align="char">44.8</td><td char="." align="char">48.0</td><td char="." align="char">37.6</td><td char="." align="char">39.1</td><td char="." align="char">30.8</td><td char="." align="char">35.2</td><td char="." align="char">34.3</td></tr><tr><td align="left">&#x0003e;12&#x000a0;years</td><td char="." align="char">45.7</td><td char="." align="char">9.7</td><td char="." align="char">16.9</td><td char="." align="char">17.4</td><td char="." align="char">14.8</td><td char="." align="char">11.7</td><td char="." align="char">6.3</td><td char="." align="char">13.7</td><td char="." align="char">58.1</td></tr><tr><td align="left">Missing</td><td char="." align="char">0.5</td><td char="." align="char">8.6</td><td char="." align="char">14.7</td><td char="." align="char">10.3</td><td char="." align="char">11.1</td><td char="." align="char">12.3</td><td char="." align="char">9.5</td><td char="." align="char">17.7</td><td char="." align="char">17.2</td></tr><tr><td align="left" colspan="10"> <italic>Paternal educational level</italic> </td></tr><tr><td align="left">&#x0003c;10&#x000a0;years</td><td char="." align="char">15.7</td><td char="." align="char">57.0</td><td char="." align="char">38.2</td><td char="." align="char">33.9</td><td char="." align="char">35.5</td><td char="." align="char">41.4</td><td char="." align="char">46.1</td><td char="." align="char">46.1</td><td char="." align="char">7.0</td></tr><tr><td align="left">10&#x02013;12&#x000a0;years</td><td char="." align="char">45.0</td><td char="." align="char">32.1</td><td char="." align="char">40.3</td><td char="." align="char">51.4</td><td char="." align="char">37.7</td><td char="." align="char">41.4</td><td char="." align="char">32.4</td><td char="." align="char">32.4</td><td char="." align="char">43.2</td></tr><tr><td align="left">&#x0003e;12&#x000a0;years</td><td char="." align="char">39.3</td><td char="." align="char">10.9</td><td char="." align="char">21.5</td><td char="." align="char">14.8</td><td char="." align="char">26.8</td><td char="." align="char">17.1</td><td char="." align="char">21.6</td><td char="." align="char">21.6</td><td char="." align="char">49.8</td></tr><tr><td align="left">Missing</td><td char="." align="char">1.1</td><td char="." align="char">9.1</td><td char="." align="char">8.9</td><td char="." align="char">10.9</td><td char="." align="char">9.4</td><td char="." align="char">12.8</td><td char="." align="char">14.5</td><td char="." align="char">14.5</td><td char="." align="char">21.8</td></tr><tr><td align="left" colspan="10"> <italic>Parental attachment to labour market</italic> </td></tr><tr><td align="left">Mother in work</td><td char="." align="char">87.6</td><td char="." align="char">57.1</td><td char="." align="char">39.8</td><td char="." align="char">66.1</td><td char="." align="char">18.0</td><td char="." align="char">44.3</td><td char="." align="char">27.7</td><td char="." align="char">23.5</td><td char="." align="char">82.1</td></tr><tr><td align="left">Father in work</td><td char="." align="char">93.9</td><td char="." align="char">75.9</td><td char="." align="char">83.2</td><td char="." align="char">83.4</td><td char="." align="char">48.3</td><td char="." align="char">77.8</td><td char="." align="char">55.3</td><td char="." align="char">59.6</td><td char="." align="char">90.1</td></tr><tr><td align="left" colspan="10"> <italic>Household income percentile</italic> </td></tr><tr><td align="left">&#x0003c;6&#x000a0;%</td><td char="." align="char">3.2</td><td char="." align="char">11.7</td><td char="." align="char">17.4</td><td char="." align="char">7.8</td><td char="." align="char">19.9</td><td char="." align="char">8.0</td><td char="." align="char">16.5</td><td char="." align="char">17.5</td><td char="." align="char">4.2</td></tr><tr><td align="left">6&#x02013;24&#x000a0;%</td><td char="." align="char">14.9</td><td char="." align="char">51.6</td><td char="." align="char">41.3</td><td char="." align="char">36.1</td><td char="." align="char">63.4</td><td char="." align="char">55.2</td><td char="." align="char">59.3</td><td char="." align="char">63.0</td><td char="." align="char">17.3</td></tr><tr><td align="left">25&#x02013;49&#x000a0;%</td><td char="." align="char">25.4</td><td char="." align="char">27.4</td><td char="." align="char">25.3</td><td char="." align="char">34.0</td><td char="." align="char">10.4</td><td char="." align="char">23.4</td><td char="." align="char">20.3</td><td char="." align="char">12.4</td><td char="." align="char">19.1</td></tr><tr><td align="left">50&#x02013;74&#x000a0;%</td><td char="." align="char">28.0</td><td char="." align="char">7.7</td><td char="." align="char">12.3</td><td char="." align="char">15.0</td><td char="." align="char">4.0</td><td char="." align="char">10.1</td><td char="." align="char">3.0</td><td char="." align="char">6.0</td><td char="." align="char">25.1</td></tr><tr><td align="left">75&#x02013;100&#x000a0;%</td><td char="." align="char">28.6</td><td char="." align="char">1.8</td><td char="." align="char">4.0</td><td char="." align="char">7.1</td><td char="." align="char">2.4</td><td char="." align="char">3.1</td><td char="." align="char">0.9</td><td char="." align="char">1.2</td><td char="." align="char">34.3</td></tr><tr><td align="left">Missing</td><td char="." align="char">2.2</td><td char="." align="char">4.2</td><td char="." align="char">4.7</td><td char="." align="char">5.1</td><td char="." align="char">4.3</td><td char="." align="char">2.1</td><td char="." align="char">9.9</td><td char="." align="char">5.6</td><td char="." align="char">2.9</td></tr><tr><td align="left" colspan="10"> <italic>Missing information on</italic> </td></tr><tr><td align="left">Breastfeeding at 4&#x000a0;month (%)</td><td char="." align="char">32.7</td><td char="." align="char">32.1</td><td char="." align="char">43.3</td><td char="." align="char">28.0</td><td char="." align="char">38.5</td><td char="." align="char">33.9</td><td char="." align="char">33.6</td><td char="." align="char">39.5</td><td char="." align="char">34.7</td></tr></tbody></table></table-wrap> </p><p>A higher proportion of the infants with mothers of other Nordic origin were fully breastfed for at least 4&#x000a0;months while infants with non-Nordic origin were significantly less likely to be fully breastfed as compared to infants of Danish origin (Table&#x000a0;<xref rid="Tab2" ref-type="table">2</xref>). Mothers of Turkish and Pakistani origin displayed the lowest breastfeeding prevalence. For Danish and other Nordic infants, the likelihood of being fully breastfed increased with maternal age at birth and parity, and was higher among women with a high socio-economic position. The pattern was less clear for infants of all other migrant groups. Only the Turkish minority had a clear age-related pattern that was similar to the Danish majority and first-borns were less likely to be fully breastfed in all minorities. However, the trend with parity was weaker and only significant for the Lebanese/Palestinian minority. The strong parental educational trend for full breastfeeding found for offspring of Danish women was weaker for non-Nordic minorities. In few cases reversed and no pattern was seen for household income, where the two largest minorities, the Turkish and the Pakistani, showed opposite trends. In the same two minorities, descendants of migrants were less likely to fully breastfeeding than migrants and for migrants there was a tendency of less breastfeeding the longer the migrant had lived in Denmark before the delivery and the younger the mother had been when she immigrated to Denmark.<table-wrap id="Tab2"><label>Table&#x000a0;2</label><caption><p>Proportion of children who were fully breastfed for 4&#x000a0;months or more according to parental characteristics and maternal country of origin and according to acculturation indicators and maternal country of origin</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="2"/><th align="left" colspan="9">Maternal country of origin</th></tr><tr><th align="left">Denmark</th><th align="left">Turkey</th><th align="left">Pakistan</th><th align="left">Former Yugoslavia</th><th align="left">Iraq</th><th align="left">Morocco</th><th align="left">Lebanon/Palestine</th><th align="left">Afghanistan</th><th align="left">Other Nordic</th></tr></thead><tbody><tr><td align="left">N&#x000a0;=&#x000a0;28,441</td><td char="." align="char">24,842</td><td char="." align="char">1,420</td><td char="." align="char">562</td><td char="." align="char">525</td><td char="." align="char">272</td><td char="." align="char">193</td><td char="." align="char">174</td><td char="." align="char">161</td><td char="." align="char">292</td></tr><tr><td align="left">Overall</td><td char="." align="char">61.8</td><td char="." align="char">42.7</td><td char="." align="char">45.6</td><td char="." align="char">50.1</td><td char="." align="char">50.0</td><td char="." align="char">49.2</td><td char="." align="char">45.8</td><td char="." align="char">57.8</td><td char="." align="char">71.6</td></tr><tr><td align="left" colspan="10"> <italic>Maternal age at birth</italic> </td></tr><tr><td align="left">&#x0003c;25&#x000a0;years</td><td char="." align="char">36.5</td><td char="." align="char">33.7</td><td char="." align="char">44.8</td><td char="." align="char">47.0</td><td char="." align="char">47.1</td><td char="." align="char">48.2</td><td char="." align="char">38.2</td><td char="." align="char">52.2</td><td char="." align="char">50.0</td></tr><tr><td align="left">25&#x02013;29&#x000a0;years</td><td char="." align="char">56.9</td><td char="." align="char">43.6</td><td char="." align="char">45.8</td><td char="." align="char">48.9</td><td char="." align="char">52.7</td><td char="." align="char">44.1</td><td char="." align="char">46.0</td><td char="." align="char">55.6</td><td char="." align="char">69.0</td></tr><tr><td align="left">30&#x02013;34&#x000a0;years</td><td char="." align="char">66.9</td><td char="." align="char">50.2</td><td char="." align="char">44.8</td><td char="." align="char">57,0</td><td char="." align="char">47.7</td><td char="." align="char">54.6</td><td char="." align="char">54.3</td><td char="." align="char">62.5</td><td char="." align="char">72.8</td></tr><tr><td align="left">35+ years</td><td char="." align="char">69.8</td><td char="." align="char">47.1</td><td char="." align="char">48.1</td><td char="." align="char">44,0</td><td char="." align="char">52.2</td><td char="." align="char">51.2</td><td char="." align="char">47.6</td><td char="." align="char">66.7</td><td char="." align="char">78.3</td></tr><tr><td align="left" colspan="10"> <italic>Maternal parity</italic> </td></tr><tr><td align="left">1</td><td char="." align="char">57.3</td><td char="." align="char">40.6</td><td char="." align="char">46.3</td><td char="." align="char">45.6</td><td char="." align="char">48.9</td><td char="." align="char">42.0</td><td char="." align="char">33.3</td><td char="." align="char">49.2</td><td char="." align="char">64.7</td></tr><tr><td align="left">2</td><td char="." align="char">64.8</td><td char="." align="char">43.0</td><td char="." align="char">44.7</td><td char="." align="char">50.9</td><td char="." align="char">46.3</td><td char="." align="char">45.3</td><td char="." align="char">40.0</td><td char="." align="char">60.0</td><td char="." align="char">75.6</td></tr><tr><td align="left">3+</td><td char="." align="char">67.5</td><td char="." align="char">45.2</td><td char="." align="char">48.7</td><td char="." align="char">59.8</td><td char="." align="char">54.9</td><td char="." align="char">54.2</td><td char="." align="char">59.7</td><td char="." align="char">66.7</td><td char="." align="char">70.2</td></tr><tr><td align="left" colspan="10"> <italic>Maternal educational level</italic> </td></tr><tr><td align="left">&#x0003c;10&#x000a0;years</td><td char="." align="char">37.0</td><td char="." align="char">41.1</td><td char="." align="char">39.2</td><td char="." align="char">44.2</td><td char="." align="char">46.9</td><td char="." align="char">44.1</td><td char="." align="char">39.6</td><td char="." align="char">59.1</td><td char="." align="char">47.4</td></tr><tr><td align="left">10&#x02013;12&#x000a0;years</td><td char="." align="char">56.2</td><td char="." align="char">44.6</td><td char="." align="char">46.1</td><td char="." align="char">50.0</td><td char="." align="char">51.7</td><td char="." align="char">50.0</td><td char="." align="char">56.6</td><td char="." align="char">60.4</td><td char="." align="char">73.6</td></tr><tr><td align="left">&#x0003e;12&#x000a0;years</td><td char="." align="char">74.1</td><td char="." align="char">52.8</td><td char="." align="char">47,0</td><td char="." align="char">57.7</td><td char="." align="char">55.3</td><td char="." align="char">65.2</td><td char="." align="char">45.5</td><td char="." align="char">45.5</td><td char="." align="char">74.8</td></tr><tr><td align="left" colspan="10"> <italic>Paternal educational level</italic> </td></tr><tr><td align="left">&#x0003c;10&#x000a0;years</td><td char="." align="char">44.5</td><td char="." align="char">39.3</td><td char="." align="char">45.6</td><td char="." align="char">48.2</td><td char="." align="char">41.1</td><td char="." align="char">51.6</td><td char="." align="char">42.4</td><td char="." align="char">58.5</td><td char="." align="char">47.2</td></tr><tr><td align="left">10&#x02013;12&#x000a0;years</td><td char="." align="char">58.2</td><td char="." align="char">46.5</td><td char="." align="char">45.8</td><td char="." align="char">52.4</td><td char="." align="char">59.6</td><td char="." align="char">54.1</td><td char="." align="char">51.9</td><td char="." align="char">60.8</td><td char="." align="char">71.6</td></tr><tr><td align="left">&#x0003e;12&#x000a0;years</td><td char="." align="char">73.1</td><td char="." align="char">53.2</td><td char="." align="char">50.9</td><td char="." align="char">57.5</td><td char="." align="char">59.3</td><td char="." align="char">40.5</td><td char="." align="char">53.6</td><td char="." align="char">67.5</td><td char="." align="char">78.9</td></tr><tr><td align="left" colspan="10"> <italic>Parental attachment to labour market</italic> </td></tr><tr><td align="left">Mother in work</td><td char="." align="char">63.2</td><td char="." align="char">43.3</td><td char="." align="char">43.0</td><td char="." align="char">49.9</td><td char="." align="char">52.4</td><td char="." align="char">48.2</td><td char="." align="char">44.9</td><td char="." align="char">55.0</td><td char="." align="char">74.2</td></tr><tr><td align="left">Mother out of work</td><td char="." align="char">51.5</td><td char="." align="char">41.8</td><td char="." align="char">47.0</td><td char="." align="char">50.6</td><td char="." align="char">49.8</td><td char="." align="char">51.0</td><td char="." align="char">45.5</td><td char="." align="char">58.6</td><td char="." align="char">58.0</td></tr><tr><td align="left">Father in work</td><td char="." align="char">63.0</td><td char="." align="char">44.0</td><td char="." align="char">45.1</td><td char="." align="char">49.5</td><td char="." align="char">49.2</td><td char="." align="char">49.0</td><td char="." align="char">44.7</td><td char="." align="char">62.2</td><td char="." align="char">73.1</td></tr><tr><td align="left">Father not in work</td><td char="." align="char">49.0</td><td char="." align="char">40.3</td><td char="." align="char">41.1</td><td char="." align="char">54.0</td><td char="." align="char">53.0</td><td char="." align="char">51.1</td><td char="." align="char">48.1</td><td char="." align="char">54.8</td><td char="." align="char">55.6</td></tr><tr><td align="left" colspan="10"> <italic>Household income percentile</italic> </td></tr><tr><td align="left">&#x0003c;6&#x000a0;%</td><td char="." align="char">50.6</td><td char="." align="char">37.1</td><td char="." align="char">55.0</td><td char="." align="char">41.7</td><td char="." align="char">51.9</td><td char="." align="char">47.4</td><td char="." align="char">36.7</td><td char="." align="char">55.2</td><td char="." align="char">54.6</td></tr><tr><td align="left">6&#x02013;24&#x000a0;%</td><td char="." align="char">47.2</td><td char="." align="char">41.7</td><td char="." align="char">45.5</td><td char="." align="char">44.3</td><td char="." align="char">50.9</td><td char="." align="char">49.5</td><td char="." align="char">50.6</td><td char="." align="char">56.8</td><td char="." align="char">64.4</td></tr><tr><td align="left">25&#x02013;49&#x000a0;%</td><td char="." align="char">54.6</td><td char="." align="char">44.5</td><td char="." align="char">44.1</td><td char="." align="char">53.0</td><td char="." align="char">41.4</td><td char="." align="char">48.7</td><td char="." align="char">50.0</td><td char="." align="char">79.0</td><td char="." align="char">69.0</td></tr><tr><td align="left">50&#x02013;74&#x000a0;%</td><td char="." align="char">64.9</td><td char="." align="char">51.0</td><td char="." align="char">40.6</td><td char="." align="char">54.7</td><td char="." align="char">72.7</td><td char="." align="char">50.0</td><td char="." align="char">0.0</td><td char="." align="char">57.1</td><td char="." align="char">68.5</td></tr><tr><td align="left">75&#x02013;100&#x000a0;%</td><td char="." align="char">74.0</td><td char="." align="char">68.0</td><td char="." align="char">45.5</td><td char="." align="char">63.6</td><td char="." align="char">66.7</td><td char="." align="char">60.0</td><td char="." align="char">50.0</td><td char="." align="char">50.0</td><td char="." align="char">81.8</td></tr><tr><td align="left" colspan="10"> <italic>Maternal acculturation indicators</italic> </td></tr><tr><td align="left">Migrated to Denmark</td><td align="left"/><td char="." align="char">46.0</td><td char="." align="char">48.1</td><td char="." align="char">49.5</td><td char="." align="char">50.0</td><td char="." align="char">50.0</td><td char="." align="char">45.4</td><td char="." align="char">57.8</td><td char="." align="char">73.4</td></tr><tr><td align="left">Descendent of migrants</td><td align="left"/><td char="." align="char">31.6</td><td char="." align="char">39.7</td><td char="." align="char">52.1</td><td char="." align="char">50.0</td><td char="." align="char">46.3</td><td char="." align="char">36.4</td><td char="." align="char">n.e.</td><td char="." align="char">60.0</td></tr><tr><td align="left" colspan="10"> <italic>Number of years lived in Denmark before delivery</italic> </td></tr><tr><td align="left">&#x0003c;10&#x000a0;years</td><td align="left"/><td char="." align="char">47.7</td><td char="." align="char">51.9</td><td char="." align="char">49.5</td><td char="." align="char">53.1</td><td char="." align="char">58.1</td><td char="." align="char">41.5</td><td char="." align="char">60.5</td><td char="." align="char">71.0</td></tr><tr><td align="left">10+ years</td><td align="left"/><td char="." align="char">44.9</td><td char="." align="char">42.3</td><td char="." align="char">49.5</td><td char="." align="char">38.7</td><td char="." align="char">44.4</td><td char="." align="char">46.7</td><td char="." align="char">50.0</td><td char="." align="char">77.3</td></tr><tr><td align="left" colspan="10"> <italic>Age at migration to Denmark</italic> </td></tr><tr><td align="left">&#x0003c;10&#x000a0;years</td><td align="left"/><td char="." align="char">43.4</td><td char="." align="char">39.5</td><td char="." align="char">45.7</td><td char="." align="char">14.3</td><td char="." align="char">46.7</td><td char="." align="char">39.3</td><td char="." align="char">55.6</td><td char="." align="char">70.6</td></tr><tr><td align="left">10&#x02013;18&#x000a0;years</td><td align="left"/><td char="." align="char">44.7</td><td char="." align="char">55.3</td><td char="." align="char">48.5</td><td char="." align="char">48.2</td><td char="." align="char">48.0</td><td char="." align="char">54.4</td><td char="." align="char">45.5</td><td char="." align="char">64.7</td></tr><tr><td align="left">&#x0003e;18&#x000a0;years</td><td align="left"/><td char="." align="char">49.9</td><td char="." align="char">48.7</td><td char="." align="char">51.5</td><td char="." align="char">53.5</td><td char="." align="char">52.8</td><td char="." align="char">38.5</td><td char="." align="char">65.0</td><td char="." align="char">74.1</td></tr></tbody></table></table-wrap> </p><p>The risk of suboptimal breastfeeding according to maternal country of origin and parental socio-demographic characteristics can be found in Table&#x000a0;<xref rid="Tab3" ref-type="table">3</xref>. Except for the Afghan minority, all non-Nordic minorities had statistically significant elevated risks compared to the Danish majority. The pattern of descendants having higher risk than migrants was found in the Turkish, Pakistani, Lebanese and Moroccan minority, but the difference did only reach statistically significance for the Turkish minority. Infants of mothers &#x0003c; 25&#x000a0;years had essentially the same significantly elevated risk of suboptimal breastfeeding, irrespective of country of origin. While the risk estimates for all strata of maternal age differed significantly from the reference group in all minority groups, the confidence intervals for the age-related risk estimates were overlapping within the minority groups. The same phenomenon was seen for maternal parity and all of the socio-economic variables.<table-wrap id="Tab3"><label>Table&#x000a0;3</label><caption><p>Crude odds ratios (95&#x000a0;% CI) for suboptimal breastfeeding^a according to socio-demographic characteristics and maternal country of origin</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="2"/><th align="left" colspan="10">Maternal country of origin</th></tr><tr><th align="left">Denmark</th><th align="left">Turkey</th><th align="left">Pakistan</th><th align="left" colspan="2">Former Yugoslavia</th><th align="left">Iraq</th><th align="left">Morocco</th><th align="left">Lebanon/Palestine</th><th align="left">Afghanistan</th><th align="left">Other Nordic</th></tr></thead><tbody><tr><td align="left" colspan="11"> <italic>Migrant status</italic> </td></tr><tr><td align="left">Overall</td><td align="left">1 (ref)</td><td align="left">2.17 (1.95&#x02013;2.42)</td><td align="left">1.93 (1.63&#x02013;2.29)</td><td align="left" colspan="2">1.61 (1.36&#x02013;1.92)</td><td align="left">1.62 (1.27&#x02013;2.05)</td><td align="left">1.67 (1.26&#x02013;2.22)</td><td align="left">1.99 (1.48&#x02013;2.69)</td><td align="left">1.18 (0.86&#x02013;1.62)</td><td align="left">0.64 (0.50&#x02013;0.83)</td></tr><tr><td align="left">Migrant</td><td align="left"/><td align="left">1.9 (1.68&#x02013;2.14)</td><td align="left">1.75 (1.43&#x02013;2.13)</td><td align="left" colspan="2">1.65 (1.36&#x02013;2.01)</td><td align="left">1.62 (1.27&#x02013;2.06)</td><td align="left">1.62 (1.18&#x02013;2.23)</td><td align="left">1.95 (1.43&#x02013;2.65)</td><td align="left">1.18 (0.86&#x02013;1.62)</td><td align="left">0.59 (0.44&#x02013;0.78)</td></tr><tr><td align="left">Descendants of migrants</td><td align="left"/><td align="left"><p>3.5</p><p>(2.77&#x02013;4.42)</p></td><td align="left"><p>2.46</p><p>(1.81&#x02013;3.35)</p></td><td align="left" colspan="2"><p>1.49</p><p>(1.03&#x02013;2.14)</p></td><td align="left"><p>1.62</p><p>(0.10&#x02013;25.9)</p></td><td align="left"><p>1.87</p><p>(1.01&#x02013;3.46)</p></td><td align="left"><p>2.83</p><p>(0.89&#x02013;9.67)</p></td><td align="left">n.e.</td><td align="left"><p>1.08</p><p>(0.57&#x02013;2.03)</p></td></tr><tr><td align="left" colspan="11"> <italic>Maternal age at birth</italic> </td></tr><tr><td align="left">&#x0003c; 25&#x000a0;years</td><td align="left"><p>3.51</p><p>(3.20&#x02013;3.85)</p></td><td align="left"><p>3.97</p><p>(3.20&#x02013;4.92)</p></td><td align="left"><p>2.49</p><p>(1.69&#x02013;3.67)</p></td><td align="left" colspan="2"><p>2.28</p><p>(1.58&#x02013;3.29)</p></td><td align="left"><p>2.27</p><p>(1.41&#x02013;3.67)</p></td><td align="left"><p>2.17</p><p>(1.02&#x02013;4.63)</p></td><td align="left"><p>3.25</p><p>(1.99&#x02013;5.32)</p></td><td align="left"><p>1.85</p><p>(1.04&#x02013;3.30)</p></td><td align="left"><p>2.02</p><p>(0.93&#x02013;4.36)</p></td></tr><tr><td align="left">25&#x02013;29&#x000a0;years</td><td align="left"><p>1.53</p><p>(1.43&#x02013;1.63)</p></td><td align="left"><p>2.61</p><p>(2.20&#x02013;3.10)</p></td><td align="left"><p>2.39</p><p>(1.87&#x02013;3.06)</p></td><td align="left" colspan="2"><p>2.12</p><p>(1.62&#x02013;2.75)</p></td><td align="left"><p>1.81</p><p>(1.20&#x02013;2.73)</p></td><td align="left"><p>2.55</p><p>(1.58&#x02013;4.13)</p></td><td align="left"><p>2.37</p><p>(1.36&#x02013;4.14)</p></td><td align="left"><p>1.61</p><p>(0.94&#x02013;2.77)</p></td><td align="left"><p>0.91</p><p>(0.52&#x02013;1.59)</p></td></tr><tr><td align="left">30&#x02013;34&#x000a0;years</td><td align="left">1 (ref)</td><td align="left"><p>2</p><p>(1.60&#x02013;2.52)</p></td><td align="left"><p>2.49</p><p>(1.79&#x02013;3.47)</p></td><td align="left" colspan="2"><p>1.52</p><p>(1.08&#x02013;2.14)</p></td><td align="left"><p>2.21</p><p>(1.36&#x02013;3.61)</p></td><td align="left"><p>1.68</p><p>(0.99&#x02013;2.86)</p></td><td align="left"><p>1.7</p><p>(0.87&#x02013;3.31)</p></td><td align="left"><p>1.21</p><p>(0.64&#x02013;2.30)</p></td><td align="left"><p>0.75</p><p>(0.51&#x02013;1.12)</p></td></tr><tr><td align="left">35&#x000a0;+&#x000a0;years</td><td align="left"><p>0.87</p><p>(0.82&#x02013;0.94)</p></td><td align="left"><p>2.67</p><p>(1.65&#x02013;3.12)</p></td><td align="left"><p>2.18</p><p>(1.26&#x02013;3.76)</p></td><td align="left" colspan="2"><p>2.57</p><p>(1.45&#x02013;4.49)</p></td><td align="left"><p>1.85</p><p>(1.04&#x02013;3.30)</p></td><td align="left"><p>1.93</p><p>(1.06&#x02013;3.51)</p></td><td align="left"><p>2.22</p><p>(0.94&#x02013;5.23)</p></td><td align="left"><p>1.01</p><p>(0.41&#x02013;2.50)</p></td><td align="left"><p>0.56</p><p>(0.33&#x02013;0.94)</p></td></tr><tr><td align="left" colspan="11"> <italic>Maternal parity</italic> </td></tr><tr><td align="left">1</td><td align="left"><p>1</p><p>(ref)</p></td><td align="left"><p>1.97</p><p>(1.62&#x02013;2.39)</p></td><td align="left"><p>1.56</p><p>(1.41&#x02013;2.12)</p></td><td align="left" colspan="2"><p>1.6</p><p>(1.21&#x02013;2.19)</p></td><td align="left"><p>1.41</p><p>(0.92&#x02013;2.14)</p></td><td align="left"><p>1.86</p><p>(1.06&#x02013;3.26)</p></td><td align="left"><p>2.69</p><p>(1.57&#x02013;4.60)</p></td><td align="left"><p>1.39</p><p>(0.83&#x02013;2.32)</p></td><td align="left"><p>0.73</p><p>(0.50&#x02013;1.07)</p></td></tr><tr><td align="left">2</td><td align="left"><p>0.73</p><p>(0.69&#x02013;0.77)</p></td><td align="left"><p>1.78</p><p>(1.46&#x02013;2.71)</p></td><td align="left"><p>1.66</p><p>(1.20&#x02013;2.29</p></td><td align="left" colspan="2"><p>1.3</p><p>(0.96&#x02013;1.75)</p></td><td align="left"><p>1.56</p><p>(0.96&#x02013;2.53)</p></td><td align="left"><p>1.62</p><p>(0.94&#x02013;2.79)</p></td><td align="left"><p>2.02</p><p>(0.97&#x02013;4.18)</p></td><td align="left"><p>0.9</p><p>(0.48&#x02013;1.69)</p></td><td align="left"><p>0.44</p><p>(0.27&#x02013;0.71)</p></td></tr><tr><td align="left">3+</td><td align="left"><p>0.65</p><p>(0.60&#x02013;0.70)</p></td><td align="left"><p>1.63</p><p>(1.33&#x02013;1.99)</p></td><td align="left"><p>1.42</p><p>(1.06&#x02013;1.90)</p></td><td align="left" colspan="2"><p>0.9</p><p>(0.60&#x02013;1.36)</p></td><td align="left"><p>1.1</p><p>(0.71&#x02013;1.71)</p></td><td align="left"><p>1.14</p><p>(0.71&#x02013;1.81)</p></td><td align="left"><p>0.91</p><p>(0.54&#x02013;1.55)</p></td><td align="left"><p>0.67</p><p>(0.35&#x02013;1.28)</p></td><td align="left"><p>0.57</p><p>(0.31&#x02013;1.07)</p></td></tr><tr><td align="left" colspan="11"> <italic>Maternal educational level</italic> </td></tr><tr><td align="left">&#x0003c;10&#x000a0;years</td><td align="left"><p>4.86</p><p>(4.49&#x02013;5.27)</p></td><td align="left"><p>4.1</p><p>(5.53&#x02013;4.76)</p></td><td align="left"><p>4.43</p><p>(3.28&#x02013;5.98)</p></td><td align="left" colspan="2"><p>3.6</p><p>(2.64&#x02013;4.91)</p></td><td align="left"><p>3.24</p><p>(2.23&#x02013;4.69)</p></td><td align="left"><p>3.63</p><p>(2.36&#x02013;5.60)</p></td><td align="left"><p>4.36</p><p>(2.89&#x02013;6.58)</p></td><td align="left"><p>1.98</p><p>(1.21&#x02013;3.24)</p></td><td align="left"><p>3.16</p><p>(1.28&#x02013;7.79)</p></td></tr><tr><td align="left">10&#x02013;12&#x000a0;years</td><td align="left"><p>2.23</p><p>(2.10&#x02013;2.36)</p></td><td align="left"><p>3.55</p><p>(2.91&#x02013;4.34)</p></td><td align="left"><p>3.35</p><p>(2.55&#x02013;4.39)</p></td><td align="left" colspan="2"><p>2.86</p><p>(2.20&#x02013;3.71)</p></td><td align="left"><p>2.78</p><p>(1.77&#x02013;4.04)</p></td><td align="left"><p>2.86</p><p>(1.75&#x02013;4.67)</p></td><td align="left"><p>2.19</p><p>(1.27&#x02013;3.78)</p></td><td align="left"><p>1.87</p><p>(1.05&#x02013;3.34)</p></td><td align="left"><p>1.03</p><p>(0.64&#x02013;1.66)</p></td></tr><tr><td align="left">&#x0003e;12&#x000a0;years</td><td align="left"><p>1</p><p>(ref)</p></td><td align="left"><p>2.56</p><p>(1.80&#x02013;3.63)</p></td><td align="left"><p>3.22</p><p>(2.09&#x02013;4.97)</p></td><td align="left" colspan="2"><p>2.1</p><p>(1.34&#x02013;3.29)</p></td><td align="left"><p>2.31</p><p>(1.22&#x02013;4.39)</p></td><td align="left"><p>1.52</p><p>(0.65&#x02013;3.60)</p></td><td align="left"><p>3.43</p><p>(1.05&#x02013;11.24)</p></td><td align="left"><p>3.43</p><p>(1.48&#x02013;7.94)</p></td><td align="left"><p>0.96</p><p>(0.65&#x02013;1.41)</p></td></tr><tr><td align="left" colspan="11"> <italic>Paternal educational level</italic> </td></tr><tr><td align="left">&#x0003c;10&#x000a0;years</td><td align="left"><p>3.4</p><p>(3.14&#x02013;3.68)</p></td><td align="left"><p>4.21</p><p>(3.60&#x02013;4.94)</p></td><td align="left"><p>3.26</p><p>(2.44&#x02013;4.34)</p></td><td align="left" colspan="2"><p>2.93</p><p>(2.09&#x02013;4.12)</p></td><td align="left"><p>3.9</p><p>(2.56&#x02013;5.95)</p></td><td align="left"><p>2.56</p><p>(1.56&#x02013;4.19)</p></td><td align="left"><p>3.7</p><p>(2.21&#x02013;6.22)</p></td><td align="left"><p>1.93</p><p>(1.04&#x02013;3.60)</p></td><td align="left"><p>3.04</p><p>(1.58&#x02013;5.86)</p></td></tr><tr><td align="left">10&#x02013;12&#x000a0;years</td><td align="left"><p>1.96</p><p>(1.85&#x02013;2.08)</p></td><td align="left"><p>3.14</p><p>(2.56&#x02013;3.84)</p></td><td align="left"><p>3.22</p><p>(2.42&#x02013;4.30)</p></td><td align="left" colspan="2"><p>2.47</p><p>(1.92&#x02013;3.19)</p></td><td align="left"><p>1.85</p><p>(1.22&#x02013;2.80)</p></td><td align="left"><p>2.31</p><p>(1.39&#x02013;3.83)</p></td><td align="left"><p>2.53</p><p>(1.48&#x02013;4.32)</p></td><td align="left"><p>1.76</p><p>(1.00&#x02013;3.09)</p></td><td align="left"><p>1.08</p><p>(0.71&#x02013;1.65)</p></td></tr><tr><td align="left">&#x0003e;12&#x000a0;years</td><td align="left"><p>1</p><p>(ref)</p></td><td align="left"><p>2.39</p><p>(1.71&#x02013;3.35)</p></td><td align="left"><p>2.62</p><p>(1.79&#x02013;3.85)</p></td><td align="left" colspan="2"><p>2.01</p><p>(1.26&#x02013;3.20)</p></td><td align="left"><p>1.87</p><p>(1.09&#x02013;3.23)</p></td><td align="left"><p>3.99</p><p>(2.07&#x02013;7.71)</p></td><td align="left"><p>2.36</p><p>(1.12&#x02013;4.97)</p></td><td align="left"><p>1.31</p><p>(0.68&#x02013;2.55)</p></td><td align="left"><p>0.73</p><p>(0.48&#x02013;1.12)</p></td></tr><tr><td align="left" colspan="11"> <italic>Parental attachment to labour market</italic> </td></tr><tr><td align="left">Mother in work</td><td align="left"><p>1</p><p>(ref)</p></td><td align="left"><p>2.25</p><p>(1.95&#x02013;2.60)</p></td><td align="left"><p>2.27</p><p>(1.75&#x02013;2.96)</p></td><td align="left" colspan="2"><p>1.73</p><p>(1.39&#x02013;2.14)</p></td><td align="left"><p>1.56</p><p>(0.85&#x02013;2.86)</p></td><td align="left"><p>1.84</p><p>(1.20&#x02013;2.82)</p></td><td align="left"><p>2.1</p><p>(1.20&#x02013;3.70)</p></td><td align="left"><p>1.32</p><p>(0.71&#x02013;2.47)</p></td><td align="left"><p>0.6</p><p>(0.45&#x02013;0.80)</p></td></tr><tr><td align="left">Mother out of work</td><td align="left"><p>1.61</p><p>(1.15&#x02013;1.74)</p></td><td align="left"><p>2.39</p><p>(2.03&#x02013;2.81)</p></td><td align="left"><p>1.93</p><p>(1.55&#x02013;2.41)</p></td><td align="left" colspan="2"><p>1.68</p><p>(1.24&#x02013;2.26)</p></td><td align="left"><p>1.73</p><p>(1.33&#x02013;2.25)</p></td><td align="left"><p>1.65</p><p>(1.18&#x02013;2.43)</p></td><td align="left"><p>2.06</p><p>(1.44&#x02013;2.95)</p></td><td align="left"><p>1.14</p><p>(0.79&#x02013;1.65)</p></td><td align="left"><p>1.24</p><p>(0.71&#x02013;2.18)</p></td></tr><tr><td align="left">Father in work</td><td align="left"><p>1</p><p>(ref)</p></td><td align="left"><p>2.17</p><p>(1.91&#x02013;2.45)</p></td><td align="left"><p>2.07</p><p>(1.72&#x02013;2.50)</p></td><td align="left" colspan="2"><p>1.73</p><p>(1.43&#x02013;2.10)</p></td><td align="left"><p>1.75</p><p>(1.24&#x02013;2.49)</p></td><td align="left"><p>1.77</p><p>(1.28&#x02013;2.46)</p></td><td align="left"><p>2.1</p><p>(1.37&#x02013;3.23)</p></td><td align="left"><p>1.03</p><p>(0.67&#x02013;1.58)</p></td><td align="left"><p>0.62</p><p>(0.48&#x02013;0.82)</p></td></tr><tr><td align="left">Father out of work</td><td align="left"><p>1.79</p><p>(1.61&#x02013;1.99)</p></td><td align="left"><p>2.52</p><p>(2.01&#x02013;3.15)</p></td><td align="left"><p>1.63</p><p>(1.07&#x02013;2.46)</p></td><td align="left" colspan="2"><p>1.45</p><p>(0.92&#x02013;2.28)</p></td><td align="left"><p>1.51</p><p>(1.08&#x02013;2.12)</p></td><td align="left"><p>1.63</p><p>(0.91&#x02013;2.92)</p></td><td align="left"><p>1.84</p><p>(1.17&#x02013;2.87)</p></td><td align="left"><p>1.4</p><p>(0.85&#x02013;2.31)</p></td><td align="left"><p>1.36</p><p>(0.54&#x02013;3.45)</p></td></tr><tr><td align="left" colspan="11"> <italic>Household income percentile</italic> </td></tr><tr><td align="left">&#x0003c;6&#x000a0;%</td><td align="left"><p>2.78</p><p>(2.38&#x02013;3.25)</p></td><td align="left"><p>4.82</p><p>(3.48&#x02013;6.68)</p></td><td align="left"><p>2.33</p><p>(1.49&#x02013;3.63)</p></td><td align="left" colspan="2"><p>3.98</p><p>(2.05&#x02013;7.74)</p></td><td align="left"><p>2.63</p><p>(1.52&#x02013;4.55)</p></td><td align="left"><p>3.16</p><p>(1.28&#x02013;7.79)</p></td><td align="left"><p>4.91</p><p>(2.33&#x02013;10.34)</p></td><td align="left"><p>2.31</p><p>(1.11&#x02013;4.81)</p></td><td align="left"><p>2.37</p><p>(0.72&#x02013;7.78)</p></td></tr><tr><td align="left">6&#x02013;24&#x000a0;%</td><td align="left"><p>3.18</p><p>(2.92&#x02013;3.46)</p></td><td align="left"><p>3.98</p><p>(3.40&#x02013;4.66)</p></td><td align="left"><p>3.4</p><p>(2.58&#x02013;4.78)</p></td><td align="left" colspan="2"><p>3.58</p><p>(2.64&#x02013;4.85)</p></td><td align="left"><p>2.74</p><p>(2.01&#x02013;3.74)</p></td><td align="left"><p>2.9</p><p>(1.97&#x02013;4.26)</p></td><td align="left"><p>2.78</p><p>(1.83&#x02013;4.23)</p></td><td align="left"><p>2.16</p><p>(1.44&#x02013;3.25)</p></td><td align="left"><p>1.57</p><p>(0.85&#x02013;2.90)</p></td></tr><tr><td align="left">25&#x02013;49&#x000a0;%</td><td align="left"><p>2.37</p><p>(2.20&#x02013;2.55)</p></td><td align="left"><p>3.54</p><p>(2.86&#x02013;4.39)</p></td><td align="left"><p>3.6</p><p>(2.61&#x02013;4.97)</p></td><td align="left" colspan="2"><p>2.52</p><p>(1.88&#x02013;3.39)</p></td><td align="left"><p>4.03</p><p>(1.92&#x02013;8.45)</p></td><td align="left"><p>2.99</p><p>(1.59&#x02013;5.62)</p></td><td align="left"><p>2.84</p><p>(1.48&#x02013;5.48)</p></td><td align="left"><p>0.76</p><p>(0.25&#x02013;2.29)</p></td><td align="left"><p>1.28</p><p>(0.73&#x02013;2.24)</p></td></tr><tr><td align="left">50&#x02013;74&#x000a0;%</td><td align="left"><p>1.54</p><p>(1.43&#x02013;1.65)</p></td><td align="left"><p>2.73</p><p>(1.85&#x02013;4.05)</p></td><td align="left"><p>4.16</p><p>(2.52&#x02013;6.86)</p></td><td align="left" colspan="2"><p>2.36</p><p>(1.49&#x02013;3.73)</p></td><td align="left"><p>1.07</p><p>(0.28&#x02013;4.02)</p></td><td align="left"><p>2.84</p><p>(1.23&#x02013;6.57)</p></td><td align="left">N.e.</td><td align="left"><p>2.13</p><p>(0.48&#x02013;9.54)</p></td><td align="left"><p>1.31</p><p>(0.80&#x02013;2.15)</p></td></tr><tr><td align="left">75&#x02013;100&#x000a0;%</td><td align="left"><p>1</p><p>(ref)</p></td><td align="left"><p>1.34</p><p>(0.58&#x02013;3.11)</p></td><td align="left"><p>3.41</p><p>(1.47&#x02013;7.91)</p></td><td align="left" colspan="2"><p>1.63</p><p>(0.80&#x02013;3.31)</p></td><td align="left"><p>1.42</p><p>(0.26&#x02013;7.77)</p></td><td align="left"><p>1.9</p><p>(0.32&#x02013;11.35)</p></td><td align="left"><p>2.84</p><p>(0.18&#x02013;45.48)</p></td><td align="left"><p>2.84</p><p>(0.18&#x02013;45.43)</p></td><td align="left"><p>0.63</p><p>(0.38&#x02013;1.06)</p></td></tr></tbody></table><table-wrap-foot><p> ^aSuboptimal breastfeeding: non-adherence to the Danish Health and Medicines Authority recommendation of exclusively breastfeeding at least to the age of 4&#x000a0;months</p></table-wrap-foot></table-wrap> </p><p>The risk of suboptimal breastfeeding was estimated on the subsample with full information on all co-variables. The estimated crude risk can be seen in Table&#x000a0;<xref rid="Tab4" ref-type="table">4</xref> together with the estimated odds ratios adjusted for maternal age and parity, maternal and paternal education, maternal and paternal attachment to labour market and household income. The crude estimates were comparable to those estimated in the full sample (Table&#x000a0;<xref rid="Tab3" ref-type="table">3</xref>), but adjustment for the co-variables attenuated the estimated risks considerably, leaving all estimates, except the estimate for the Pakistani minority, close to unity. Infants with Pakistani origin had an adjusted odds ratio of 1.31 for being suboptimally breastfed and infants of Afghan origin has a decreased risk [OR 0.56 (95&#x000a0;% CI 0.36&#x02013;0.87)] as compared to infant of Danish mothers.<table-wrap id="Tab4"><label>Table&#x000a0;4</label><caption><p>Odds Ratios for suboptimal breastfeeding^a according to maternal country of origin adjusted for socio-demographic characteristics^b </p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="2"/><th align="left" colspan="9">Maternal country of origin</th></tr><tr><th align="left">Denmark</th><th align="left">Turkey</th><th align="left">Pakistan</th><th align="left">Former Yugoslavia</th><th align="left">Iraq</th><th align="left">Morocco</th><th align="left">Lebanon/Palestine</th><th align="left">Afghanistan</th><th align="left">Other Nordic</th></tr></thead><tbody><tr><td align="left">N&#x000a0;=&#x000a0;23,596</td><td char="." align="char">21,103</td><td char="." align="char">1,028</td><td char="." align="char">369</td><td char="." align="char">375</td><td char="." align="char">183</td><td char="." align="char">129</td><td char="." align="char">112</td><td char="." align="char">95</td><td char="." align="char">202</td></tr><tr><td align="left" colspan="10"> <italic>Full information on covariats</italic> </td></tr><tr><td align="left">Crude OR</td><td align="left"><p>1</p><p>(ref)</p></td><td align="left"><p>2.09</p><p>(1.84&#x02013;2.37)</p></td><td align="left"><p>1.95</p><p>(1.58&#x02013;2.39)</p></td><td align="left"><p>1.52</p><p>(1.24&#x02013;1.86)</p></td><td align="left"><p>1.38</p><p>(1.03&#x02013;1.85)</p></td><td align="left"><p>1.76</p><p>(1.24&#x02013;2.49)</p></td><td align="left"><p>1.75</p><p>(1.21&#x02013;2.53)</p></td><td align="left"><p>0.95</p><p>(0.63&#x02013;1.44)</p></td><td align="left"><p>0.7</p><p>(0.52&#x02013;0.95)</p></td></tr><tr><td align="left">Adjusted OR</td><td align="left"><p>1</p><p>(ref)</p></td><td align="left"><p>1.06</p><p>(0.92&#x02013;1.22)</p></td><td align="left"><p>1.31</p><p>(1.06&#x02013;1.63)</p></td><td align="left"><p>0.98</p><p>(0.79&#x02013;1.21)</p></td><td align="left"><p>0.79</p><p>(0.58&#x02013;1.08)</p></td><td align="left"><p>1.1</p><p>(0.77&#x02013;1.57)</p></td><td align="left"><p>0.88</p><p>(0.60&#x02013;1.30)</p></td><td align="left"><p>0.56</p><p>(0.36&#x02013;0.87)</p></td><td align="left"><p>0.81</p><p>(0.59&#x02013;1.12)</p></td></tr></tbody></table><table-wrap-foot><p> ^aSuboptimal breastfeeding: non-adherence to the Danish Health and Medicines Authority recommendation of exclusively breastfeeding at least to the age of 4&#x000a0;months</p><p> ^bAdjusted for maternal age, parity, maternal and paternal education, maternal and paternal attachment to labour market, and household income</p></table-wrap-foot></table-wrap> </p><p>The proportion of infants, who were exposed to the two indicators of breastfeeding support, can be seen in Fig.&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>. From the figure follows that the proportions of infants that were put to the breast within 2&#x000a0;h of birth, and not being artificially fed (formula fed) at the hospital did, not differ according to maternal country of origin.<fig id="Fig1"><label>Fig.&#x000a0;1</label><caption><p>Distribution of indicators of breastfeeding support according to maternal country of orgin</p></caption><graphic xlink:href="10995_2014_1486_Fig1_HTML" id="MO1"/></fig> </p></sec><sec id="Sec10"><title>Discussion</title><p>Suboptimal breastfeeding was more frequent among women of non-Nordic origin than among women with Danish origin. Among descendants of Turkish and Pakistani immigrants, the breastfeeding frequency further decreased, thus the acculturation process did not favor breastfeeding. For all minority groups, but the Pakistani, adjustment for socio-demographic variables abolished the increased risk of suboptimal breastfeeding. There were no differences between migrants and non-migrants in the breastfeeding support provided at hospital level.</p><p>Previous international studies are in general concluding that breastfeeding rates are higher among the migrant than the native populations [<xref ref-type="bibr" rid="CR11">11</xref>&#x02013;<xref ref-type="bibr" rid="CR14">14</xref>]. It seems that this does not apply to the Danish context. This could be explained with the very high level of breastfeeding initiation in general in Denmark (95&#x000a0;%) relative to US (57&#x000a0;%) [<xref ref-type="bibr" rid="CR11">11</xref>], UK (70&#x000a0;%) [<xref ref-type="bibr" rid="CR12">12</xref>], and the Netherlands (90&#x000a0;%) [<xref ref-type="bibr" rid="CR14">14</xref>]. Comparison of breastfeeding prevalence across countries is difficult because of differences in measures of duration and exclusivity of breastfeeding. However, in Denmark, around 46&#x000a0;% of women with Pakistani origin were fully breastfeeding at 4&#x000a0;months, while in a British context 36&#x000a0;% of women with Pakistani origin reported to be breastfeeding without introduction of any solid foods at 4&#x000a0;months [<xref ref-type="bibr" rid="CR12">12</xref>]. Thus, the levels of breastfeeding in migrants in Denmark might not be different from the levels in other settings, and the gap to the women with Danish origin could be due to the high level in women of Danish origin.</p><p>Studies from the US have shown that the longer the migrant women have been in the country prior to birth, the lower breastfeeding prevalence, and further that descendants were breastfeeding less than immigrants [<xref ref-type="bibr" rid="CR11">11</xref>]. The same pattern was seen for second generation Mediterranean migrants in the Netherlands, when compared to the first generation [<xref ref-type="bibr" rid="CR14">14</xref>]. However, in Sweden, partial breastfeeding among migrants was the same among women, who had stayed less or more than 5&#x000a0;years [<xref ref-type="bibr" rid="CR9">9</xref>], thus no negative effect of acculturation was indicated. In Denmark the negative effect of acculturation was seen in the crude analysis, however not in direction towards the native population, but towards increased disparity.</p><p>An anthropological study of perceptions of breastfeeding among women of Turkish origin in Denmark have found that the women had intentions to breastfeed, had initiated breastfeeding and had high knowledge about breastfeeding, but experienced early cessation [<xref ref-type="bibr" rid="CR21">21</xref>]. The study elucidated that these women, in their process of acculturation were struggling to feel respected as modern women. To these women a modern life included to take part in the public spaces, but traditional norms made them consider breastfeeding as a practice only acceptable in domestic spaces. As a consequence, some women shortened breastfeeding duration. We speculate that this example of everyday life experiences of being between two cultures could be one example of many acculturation processes, and that these might explain the lower frequencies of optimal breastfeeding in women of non-Western origin in Denmark.</p><p>For the Danish-born women the breastfeeding frequency was higher among women in work than among women out of work, which we see as a result of the supportive Danish legislation for maternity leave. Among the migrants this tendency was less clear, except for the other Nordic immigrants, and this finding might reflect the contemporary weaker attachment to the Danish labour market among the migrant women.</p><p>Adjustment for maternal age, parity, and paternal education, employment status, and income abolished the increased risk of suboptimal breastfeeding in all groups, except the Pakistani. In a strict sense, these factors do not act as confounders, but may be mediators of the association demonstrated. In other words, the crude odds ratios do represent the true, not confounded, increased risk of suboptimal breastfeeding that offspring of non-Nordic mothers face. The results of the adjusted analyses indicate, however, that this increased risk is a result of the socio-economic circumstances women of non-Nordic origin in Denmark currently face. In the Nordic countries, suboptimal health care for migrants has been shown [<xref ref-type="bibr" rid="CR22">22</xref>, <xref ref-type="bibr" rid="CR23">23</xref>] and migrants have reported discrimination in the Danish health care system [<xref ref-type="bibr" rid="CR24">24</xref>]. However, in this study we found no indication of discrimination as there were no migrant specific differences in the breastfeeding support provided at hospital level.</p><sec id="Sec11"><title>Study Limitations</title><p>The data available for this study was a unique combination of health visitor data and population covering registries. The health visitor data included information from more than 90&#x000a0;% of children born in the included 18 municipalities in the study period and thus most likely less subject to selection bias than other (often survey) studies of breastfeeding.</p><p>There was high occurrence of missing data on full breastfeeding, but it was reassuring regarding risk of biased estimates that no particular skewness was found the distribution of the co-variables between women with and without information on breastfeeding.</p></sec></sec><sec id="Sec12"><title>Conclusions and Perspectives</title><p>Suboptimal breastfeeding was more frequent among women of non-Nordic origin than among women with Danish origin. Women, who were descendants of Turkish and Pakistani immigrants, had an even lower breastfeeding frequency than first generation migrants from these countries, indicating that the acculturation process did not favor breastfeeding. This finding is of public health relevance to the health professionals, who needs to be aware of vulnerability of these mothers. The negative effect of acculturation makes it relevant to further study the reasons for breastfeeding cessation among the Turkish and Pakistani women and a qualitative approached combined with richer and more detailed information on breastfeeding practices from survey data seems appropriate. Based on such studies strategies for how best to support these women, and the health of their infants/children, could be developed.</p></sec>

Combined Immunodeficiency Evolving into Predominant CD4+ Lymphopenia Caused by Somatic Chimerism in JAK3

Could not extract abstract

<contrib contrib-type="author"><name><surname>Ban</surname><given-names>Sol A.</given-names></name><xref ref-type="aff" rid="Aff1">1</xref></contrib><contrib contrib-type="author"><name><surname>Salzer</surname><given-names>Elisabeth</given-names></name><xref ref-type="aff" rid="Aff1">1</xref></contrib><contrib contrib-type="author"><name><surname>Eibl</surname><given-names>Martha M.</given-names></name><xref ref-type="aff" rid="Aff2">2</xref></contrib><contrib contrib-type="author"><name><surname>Linder</surname><given-names>Angela</given-names></name><xref ref-type="aff" rid="Aff2">2</xref></contrib><contrib contrib-type="author"><name><surname>Geier</surname><given-names>Christoph B.</given-names></name><xref ref-type="aff" rid="Aff2">2</xref></contrib><contrib contrib-type="author"><name><surname>Santos-Valente</surname><given-names>Elisangela</given-names></name><xref ref-type="aff" rid="Aff1">1</xref></contrib><contrib contrib-type="author"><name><surname>Garncarz</surname><given-names>Wojciech</given-names></name><xref ref-type="aff" rid="Aff1">1</xref></contrib><contrib contrib-type="author"><name><surname>Lion</surname><given-names>Thomas</given-names></name><xref ref-type="aff" rid="Aff3">3</xref><xref ref-type="aff" rid="Aff4">4</xref><xref ref-type="aff" rid="Aff5">5</xref></contrib><contrib contrib-type="author"><name><surname>Ott</surname><given-names>Raphael</given-names></name><xref ref-type="aff" rid="Aff1">1</xref></contrib><contrib contrib-type="author"><name><surname>Seelbach</surname><given-names>Christoph</given-names></name><xref ref-type="aff" rid="Aff6">6</xref></contrib><contrib contrib-type="author" corresp="yes"><name><surname>Boztug</surname><given-names>Kaan</given-names></name><address><phone>+43-1/40160-70 069</phone><email>kboztug@cemm.oeaw.ac.at</email></address><xref ref-type="aff" rid="Aff1">1</xref><xref ref-type="aff" rid="Aff3">3</xref></contrib><contrib contrib-type="author" corresp="yes"><name><surname>Wolf</surname><given-names>Hermann M.</given-names></name><address><phone>+43 1/403-14-50</phone><email>hermann.wolf@itk.at</email></address><xref ref-type="aff" rid="Aff2">2</xref></contrib><aff id="Aff1"><label>1</label>CeMM Research Center for Molecular Medicine of the Austrian Academy of Sciences, Lazarettgasse 14, AKH BT 25.3, A-1090 Vienna, Austria </aff><aff id="Aff2"><label>2</label>Immunology Outpatient Clinic, Schwarzspanierstra&#x000df;e 15/1, A-1090 Vienna, Austria </aff><aff id="Aff3"><label>3</label>Department of Pediatrics and Adolescent Medicine, Medical University of Vienna, Vienna, Austria </aff><aff id="Aff4"><label>4</label>Children&#x02019;s Cancer Research Institute, Vienna, Austria </aff><aff id="Aff5"><label>5</label>Labdia Labordiagnostik GmbH, Vienna, Austria </aff><aff id="Aff6"><label>6</label>Kardinal Schwarzenberg&#x02018;sches Krankenhaus, Schwarzach im Pongau, Austria </aff><aff id="Aff7"><label>7</label>Research Center for Molecular Medicine of the Austrian Academy of Sciences, Lazarettgasse 14, AKH BT 25.3, A-1090 Vienna, Austria </aff>

Journal of Clinical Immunology

<sec id="Sec1" sec-type="introduction"><title>Introduction</title><p id="Par5">Idiopathic CD4 lymphopenias (ICLs) constitute an enigmatic and heterogeneous group of disorders which are collectively characterized by prolonged low CD4+ T lymphocyte count of less than 300 cells/&#x003bc;l or less than 20 % of lymphocytes on more than one determination in the absence of known causes such as HIV infection, malignant disease or medication [<xref ref-type="bibr" rid="CR1">1</xref>&#x02013;<xref ref-type="bibr" rid="CR3">3</xref>]. The clinical phenotype of ICL is variable and ranges from asymptomatic laboratory abnormality of CD4+ T-cell count to increased susceptibility to infections, opportunistic infections and autoimmune diseases [<xref ref-type="bibr" rid="CR3">3</xref>, <xref ref-type="bibr" rid="CR4">4</xref>]. Current knowledge on the molecular pathogenesis of ICL is limited.</p><p id="Par6">Genetic studies of patients with combined immunodeficiency (CID) with predominant CD4 cell deficiency have revealed mutations in genes encoding regulatory factors of the expression of MHC class II molecules such as <italic>CIITA</italic>, <italic>RFXANK</italic>, <italic>RFX5</italic> or <italic>RFXAP</italic> [<xref ref-type="bibr" rid="CR5">5</xref>&#x02013;<xref ref-type="bibr" rid="CR9">9</xref>]. The associated disease is termed MHC class II deficiency, characterized by low numbers of CD4+ T-cells while numbers of CD8+ T-cells are normal or elevated [<xref ref-type="bibr" rid="CR10">10</xref>]. Furthermore, mutations in <italic>P56LCK</italic>, a tyrosine kinase in the downstream of the TCR activation pathway, were described to cause CID with CD4 deficiency [<xref ref-type="bibr" rid="CR11">11</xref>]. Recently, a mutation in <italic>MAGT1</italic>, a gene encoding a Mg2+ transporter, was reported to cause a disorder associated with CD4 deficiency denominated as XMEN &#x02013; X-linked immunodeficiency with magnesium defect and EBV infection and neoplasia [<xref ref-type="bibr" rid="CR12">12</xref>]. Other studies have reported CID with CD4 lymphopenia in patients bearing hypomorphic mutations in genes which are typically associated with severe combined immunodeficiency (SCID) phenotype, such as <italic>RAG1</italic>, which is known to cause SCID when mutated in amorphic manner [<xref ref-type="bibr" rid="CR13">13</xref>].</p><p id="Par7">In this study, we investigated a consanguineous family with two affected siblings suffering from CID that evolved into predominant CD4 lymphopenia in order to define hitherto unknown genetic etiologies underlying this condition.</p></sec><sec id="Sec2" sec-type="materials|methods"><title>Methods</title><sec id="Sec3"><title>Patients</title><p id="Par8">The protocol for this study was approved by the Ethics Committee at the Medical University of Vienna, Austria. Blood samples from index patients and their family members from a Turkish family were obtained with informed consent in agreement with the Declaration of Helsinki.</p></sec><sec id="Sec4"><title>DNA Isolation</title><p id="Par9">For isolation of genomic DNA from whole blood, a commercially available kit (Wizard&#x000ae; Genomic DNA Purification Kit, Promega Corporation) was employed according to the manufacturer&#x02019;s instruction.</p><p id="Par10">For isolation of DNA from FACS-sorted leukocyte subsets, a commercially available Qiagen DNA Micro Kit was used according to the manufacturer&#x02019;s instruction. Subsequently, DNA was amplified by Whole Genome Amplification using the commercially available Qiagen Repli-g kit.</p></sec><sec id="Sec5"><title>Capillary Sequencing</title><p id="Par11">Capillary sequencing of genomic DNA from both patients was performed with primers designed for the variant in the <italic>JAK3</italic> gene with PrimerZ (<ext-link ext-link-type="uri" xlink:href="http://www.primerz.org/">www.primerz.org</ext-link>) and purchased from Eurofins/MWG Operon (Ebersberg, Germany). The sequences of the primers are AAGTGCTCTGACTTGCCACA (forward) and CACCTTTCTGACCCCTTCAC (reverse). Expand High Fidelity PCR System (Roche, Basel, Switzerland) was applied for PCR amplification and Big Dye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems, Darmstadt, Germany) for capillary sequencing. Sequences were acquired using an ABI 3130xl Sequencer (Applied Biosystems) and analyzed using 3130xl Genetic Analyzer (Applied Biosystems) and Sequencher DNA Software 4.10.1 (Gene Codes Corporation, Ann Arbor, MI, USA).</p></sec><sec id="Sec6"><title>Homozygosity Mapping</title><p id="Par12">Homozygous intervals were determined as previously described [<xref ref-type="bibr" rid="CR14">14</xref>] using Affymetrix&#x000ae; Genome-Wide Human SNP Array 6.0 technology. The outcome data was analyzed using Affymetrix&#x000ae; Genotyping Console^&#x02122; software version 4.0.1.8.6. Homozygous intervals were mapped using Homozygosity Mapper (<ext-link ext-link-type="uri" xlink:href="http://www.homozygositymapper.org/">www.homozygositymapper.org/</ext-link>).</p></sec><sec id="Sec7"><title>Exome Sequencing and Data Analysis</title><p id="Par13">Exome sequencing was performed for patient 2. Illumina TruSeq DNA Sample Preparation Guide and the Illumina TruSeq Exome Enrichment Guide version 3 were used. Genomic DNA (1&#x000a0;&#x003bc;g) was sheared to fragments of 200&#x02013;300&#x000a0;bp. Blunt ending, adenylation and adapter-ligation allowing the fragments to hybridize onto the flow cell were carried out. Exonic DNA fragments were enriched and clusters were generated using the Illumina cBot Cluster Generation System following the TruSeq PE Cluster Kit v3 Reagent Preparation Guide. The DNA fragment clusters ran in a multiplexed pool with five other samples distributed on two lanes of the flow cell.</p><p id="Par14">Data analysis was performed as previously described [<xref ref-type="bibr" rid="CR14">14</xref>]. Reads were aligned using Burrows-Wheeler Aligner (BWA) to the human genome 19. Insertion/deletion realignment was performed as well as Genome Analysis Toolkit (GATK version 1.5)-based quality score recalibration. For single nucleotide variants (SNVs) and Deletion/Insertion variants (DIVs) calling, Unified Genotyper and GATK Variant quality score recalibration were performed. SNVs and DIVs lists were uploaded to SeattleSeq Annotation database with dbSNPbuild135. Variants present in 1000Genomes and dbSNP were excluded and the lists were filtered for nonsense, missense and splice-site variants present within the overlapping homozygous intervals of both patient. At last, SNVs were filtered according to a validation prediction score.</p></sec><sec id="Sec8"><title>Cell Sorting for Analysis of Somatic Chimerism</title><p id="Par15">Peripheral blood mononuclear cells (PBMCs) of both patients were isolated by density gradient centrifugation using Ficoll-Hypaque (GE Healthcare, Uppsala, Sweden) and stained with the following antibodies: CD3-FITC, CD4-APC (BD, Biosciences, Schwechat, Austria), CD8-PECy7 (Beckmann Coulter, Krefeld, Germany), CD19-PerCPCy5.5 (eBioscience, Vienna, Austria) and CD56-V450 (BD, Biosciences, Austria). Subsequently, the stained cells were sorted into different subgroups of leukocytes CD3+CD4+CD8-T-cells, CD3+CD4-CD8+&#x02009;T-cells and CD3-CD19+&#x02009;B-cells using a MoFlo Astrios Cell Sorter from Beckmann Coulter.</p></sec><sec id="Sec9"><title>Chimerism Analysis for Maternal Cells</title><p id="Par16">The screening for maternal cells in the peripheral blood of the patient investigated was performed by quantitative chimerism testing, as described previously [<xref ref-type="bibr" rid="CR15">15</xref>, <xref ref-type="bibr" rid="CR16">16</xref>]. Five informative microsatellite/short tandem repeat (STR) markers including D3S3045, D4S2366, D12S1064, D16S539, D17S1290, and SE-33 were analyzed by PCR analysis and capillary electrophoresis with fluorescence-based detection [<xref ref-type="bibr" rid="CR17">17</xref>].</p></sec><sec id="Sec10"><title>T-cell CDR3 V&#x003b2; Spectratyping</title><p id="Par17">The examination of TCR V&#x003b2; repertoire was performed by spectratyping analysis as described by Pannetier et al. [<xref ref-type="bibr" rid="CR18">18</xref>] with the modification of the sequence for primers listed in the table below.<table-wrap id="Taba"><table frame="hsides" rules="groups"><tbody><tr><td> <bold>Target of the primer</bold> </td><td> <bold>Sequence</bold> </td></tr><tr><td colspan="2"> <bold>Primers for variable regions</bold> </td></tr><tr><td>BV02</td><td>ACATACGAGCAAGGCGTCGA</td></tr><tr><td>BV04</td><td>CATCAGCCGCCCAAACCTAA</td></tr><tr><td>BV07</td><td>CAAGTCGCTTCTCACCTGAATGC</td></tr><tr><td>BV17</td><td>TGTGACATCGGCCCAAAAGAA</td></tr><tr><td>BV21</td><td>GGAGTAGACTCCACTCTAAG</td></tr><tr><td>BV24</td><td>CCCAGTTTGGAAAGCCAGTGACCC</td></tr><tr><td colspan="2"> <bold>Primer for the constant region</bold> </td></tr><tr><td>C&#x000df;B1 (used for BV05, BV06BC, BV20)</td><td>CGGGCTGCTCCTTGAGGGGCTGCG</td></tr><tr><td colspan="2"> <bold>Fluorescently labeled primer for the constant region</bold> </td></tr><tr><td>FAM-marked primer for the constant region</td><td>ACACAGCGACCTCGGGTGGG</td></tr></tbody></table></table-wrap> </p><p id="Par18">For the fragment analysis, sequences were acquired using an ABI 3130xl Sequencer (ABI Applied Biosystems) and analyzed using GeneMapper software version 3.7.</p></sec><sec id="Sec11"><title>Generation of an EBV-Transformed Cell Line</title><p id="Par19">PBMCs were isolated from heparinized peripheral blood by density gradient centrifugation (Lymphoprep, Axis-Shield PoC AS, Oslo, Norway) and transformed with Epstein-Barr virus (EBV) using the supernatant from the B 95&#x02013;8 marmoset cell line (ATCC, Rockville, MD) according to a standard protocol [<xref ref-type="bibr" rid="CR19">19</xref>]. Growing cells were expanded in complete RPMI 1640 medium supplemented with 10 % heat-inactivated fetal calf serum (FCS, PAA), 2&#x000a0;mM L-glutamine, 100&#x000a0;IU/ml penicillin, and 100&#x000a0;&#x003bc;g/ml streptomycin (Gibco, Paisley, Scottland) at 37&#x000a0;&#x000b0;C in the presence of 5 % CO2.</p></sec><sec id="Sec12"><title>Analysis of JAK3 and STAT5 Protein Expression</title><p id="Par20">EBV-transformed B-cells were lysed for 30&#x000a0;min in ice-cold lysis buffer [20&#x000a0;mM Tris&#x02013;HCl (pH&#x000a0;7.5), 150&#x000a0;mM NaCl, 2&#x000a0;mM EDTA, 1 % Nonidet P-40, protease inhibitor cocktail (Complete; Roche)] and insoluble material was removed by centrifugation (16,000 x g, 10&#x000a0;min, 4 &#x000b0;C). 20 micrograms of protein was resolved by 8 % SDS-polyacrylamide gel electrophoresis (SDS-PAGE), electrotransferred onto polyvinylidene difluoride membrane (Immobilon-P; Millipore), and immunoblotted with anti-JAK3 antibody (C-21; Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA), anti-Stat5 antibody (89; BD Biosciences) and anti-GAPDH antibody (Santa Cruz). Detection was performed using the SuperSignal West Pico ECL detection system (Thermo Scientific, Waltham, MA, USA).</p></sec><sec id="Sec13"><title>Analysis of JAK3 Signaling Function in B-cell Lines and CD4+ T-cells</title><p id="Par21">Measurement of STAT3 and STAT5 activation was performed by flow cytometry: Phosphorylation-state analysis was performed on EBV-transformed B-cells or CD4+ T-cells contained within the PBMC fraction using BD PhosFlow technology, according to the manufacturer&#x02019;s instructions (BD Biosciences, Mississauga, ON). After 2&#x000a0;h of culture in 1 % FCS, cells were stimulated with recombinant human IL-2 at 100, 1,000 and 10 000 U/ml (Bio-Rad AbD Serotec, Germany), IL-4 plus IL-21 at 100&#x000a0;ng/ml (Gibco/Life Technologies, Vienna, Austria) or IL-6 at 100&#x000a0;ng/ml (Bio-Rad AbD Serotec, Vienna, Austria) for 15&#x000a0;min at 37&#x000a0;&#x000b0;C, fixed with BD Cytofix Fixation buffer and permeabilized in ice-cold BD Perm buffer III (BD Biosciences). Cells were stained with Alexa Fluor 647-labeled anti-STAT3-pY705 or anti-STAT5-pY694 (both from BD Biosciences). CD4+ T-cells contained within the PBMC fraction were identified by gating in dual-colour flow cytometry. Fluorescence was measured by flow cytometry using a FACS Calibur (Becton Dickinson, Heidelberg, Germany) and data analyzed with the CellQuest software (Becton Dickinson).</p></sec><sec id="Sec14"><title>Crystal Structure</title><p id="Par22">Crystal structures of the kinase domain of JAK3 tyrosine kinase were created using the ICM-Browser Software (Molsoft LLC, San Diego, US).</p></sec></sec><sec id="Sec15" sec-type="results"><title>Results</title><sec id="Sec16"><title>Patient Characteristics</title><p id="Par23">Patient 1 (currently 15&#x000a0;years of age) is the first child of healthy consanguineous parents of Turkish origin. She was admitted to hospital for the first time at the age of 20&#x000a0;months because of recurrent episodes of infection of the upper and lower airways with pulmonary infiltrates at varying locations, subsequently complicated by development of atelectases. Parenteral antibiotic therapy was initiated and led to improvement of clinical symptoms and discharge from hospital. No further severe infectious episodes occurred thereafter.</p><p id="Par24">Because of the recurrent infectious episodes primary immunodeficiency was suspected and an immunological work-up was initiated which revealed CID with CD4 lymphopenia on multiple occasions, IgG2-IgG4-subclass deficiency, and selective antibody deficiency against bacterial polysaccharide antigens (Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref> and online supplementary Table&#x000a0;<xref rid="MOESM2" ref-type="media">2</xref>). Intravenous Immunoglobulin (IVIG) substitution therapy was initiated at the age of five years and continued for over a year. Re-evaluation of antibody production after cessation of IVIG therapy confirmed selective polysaccharide antibody deficiency and showed a decreased IgG antibody response against other antigens e.g. diphtheria toxoid vaccination (Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>). At the age of nine, subcutaneous immunoglobulin (SCIG) home therapy was started which is still being carried out.<table-wrap id="Tab1"><label>Table 1</label><caption><p>Immunological phenotype of both patients at representative time points</p></caption><table frame="hsides" rules="groups"><tbody><tr><td colspan="9">A. Serum levels of immunoglobulins, antibacterial antibodies and antibodies against vaccination antigens</td></tr><tr><td/><td colspan="4">Patient 1 (II - 1)</td><td colspan="4">Patient 2 (II - 2)</td></tr><tr><td/><td colspan="2">22&#x000a0;m</td><td colspan="2">9&#x000a0;years</td><td colspan="2">3&#x000a0;years</td><td colspan="2">8&#x000a0;years</td></tr><tr><td>&#x02003;IgG (mg/dl)</td><td>585</td><td char="(" align="char">(570&#x02013;1322)</td><td>743</td><td char="(" align="char">(790&#x02013;1700)</td><td>801</td><td char="(" align="char">(696&#x02013;1518)</td><td>943</td><td char="(" align="char">(790&#x02013;1700)</td></tr><tr><td>&#x02003;IgA (mg/dl)</td><td>114</td><td char="(" align="char">(23&#x02013;97)</td><td>144</td><td char="(" align="char">(76&#x02013;450)</td><td>48</td><td char="(" align="char">(46&#x02013;177)</td><td>43</td><td char="(" align="char">(76&#x02013;450)</td></tr><tr><td>&#x02003;IgM (mg/dl)</td><td>307</td><td char="(" align="char">(76&#x02013;187)</td><td>247</td><td char="(" align="char">(90&#x02013;350)</td><td>160</td><td char="(" align="char">(97&#x02013;228)</td><td>147</td><td char="(" align="char">(90&#x02013;350)</td></tr><tr><td>&#x02003;IgG1 (mg/dl)</td><td>497</td><td char="(" align="char">(457&#x02013;734)</td><td>554</td><td char="(" align="char">(500&#x02013;880)</td><td>616</td><td char="(" align="char">(400&#x02013;983)</td><td>732</td><td char="(" align="char">(500&#x02013;880)</td></tr><tr><td>&#x02003;IgG2 (mg/dl)</td><td>&#x0003c;23</td><td char="(" align="char">(56&#x02013;200)</td><td>21</td><td char="(" align="char">(150&#x02013;600)</td><td>60</td><td char="(" align="char">(70&#x02013;400)</td><td>108</td><td char="(" align="char">(150&#x02013;600)</td></tr><tr><td>&#x02003;IgG3 (mg/dl)</td><td>&#x0003c;6</td><td char="(" align="char">(20&#x02013;81)</td><td>22</td><td char="(" align="char">(20&#x02013;100)</td><td>24</td><td char="(" align="char">(20&#x02013;81)</td><td>36</td><td char="(" align="char">(20&#x02013;100)</td></tr><tr><td>&#x02003;IgG4 (mg/dl)</td><td>&#x0003c;8</td><td char="(" align="char">(0&#x02013;40)</td><td>&#x0003c;8</td><td char="(" align="char">(8&#x02013;120)</td><td>&#x0003c;6</td><td char="(" align="char">(0&#x02013;40)</td><td>&#x0003c;7</td><td char="(" align="char">(8&#x02013;120)</td></tr><tr><td>&#x02003;Tetanus-IgG (IU/ml)</td><td>0,45*³)</td><td>(&#x0003e;&#x02009;=&#x02009;0,40)</td><td>1,57*⁴)</td><td>(&#x0003e;&#x02009;=&#x02009;0,40)</td><td>0,6*³)</td><td>(&#x0003e;&#x02009;=&#x02009;0,40)</td><td>1,02*⁵)</td><td>(&#x0003e;&#x02009;=&#x02009;0,40)</td></tr><tr><td>&#x02003;Diphteria-IgG (IU/ml)</td><td>0,15*³)</td><td>(&#x0003e;&#x02009;=&#x02009;0,40)</td><td>0,17*⁴)</td><td>(&#x0003e;&#x02009;=&#x02009;0,40)</td><td>0,08*³)</td><td>(&#x0003e;&#x02009;=&#x02009;0,40)</td><td>0,09*⁵)</td><td>(&#x0003e;&#x02009;=&#x02009;0,40)</td></tr><tr><td>&#x02003;Pn23-IgG (reciprocal titer)</td><td>&#x0003c;20</td><td>(&#x0003e;&#x02009;=&#x02009;200)</td><td>81*¹)</td><td>(&#x0003e;&#x02009;=&#x02009;200)</td><td>&#x0003c;20</td><td>(&#x0003e;&#x02009;=&#x02009;200)</td><td>249*²)</td><td>(&#x0003e;&#x02009;=&#x02009;200)</td></tr><tr><td>&#x02003;Pn23-IgM (reciprocal titer)</td><td>198</td><td>(&#x0003e;&#x02009;=&#x02009;100)</td><td>393*¹)</td><td>(&#x0003e;&#x02009;=&#x02009;100)</td><td>414</td><td>(&#x0003e;&#x02009;=&#x02009;100)</td><td>999*²)</td><td>(&#x0003e;&#x02009;=&#x02009;100)</td></tr><tr><td>&#x02003;Hib-IgG (&#x003bc;g/ml)</td><td>5,41*³)</td><td>(&#x0003e;&#x02009;=&#x02009;1)</td><td>0,17</td><td>(&#x0003e;&#x02009;=&#x02009;1)</td><td>2,57*³)</td><td>(&#x0003e;&#x02009;=&#x02009;1)</td><td>0,47</td><td>(&#x0003e;&#x02009;=&#x02009;1)</td></tr><tr><td colspan="9">B. Lymphocyte subpopulations</td></tr><tr><td/><td colspan="4">Patient 1 (II - 1)</td><td colspan="4">Patient 2 (II - 2)</td></tr><tr><td/><td colspan="2">22&#x000a0;m</td><td colspan="2">10&#x000a0;year</td><td colspan="2">3&#x000a0;years</td><td colspan="2">8&#x000a0;years</td></tr><tr><td>&#x02003;CD4 (%Ly)</td><td>7</td><td char="(" align="char">(31&#x02013;66)</td><td>22</td><td char="(" align="char">(31&#x02013;66)</td><td>11</td><td char="(" align="char">(31&#x02013;66)</td><td>22</td><td char="(" align="char">(31&#x02013;66)</td></tr><tr><td>&#x02003;CD4 (abs.Nr/&#x003bc;l)</td><td>430</td><td char="(" align="char">(386&#x02013;2022)</td><td>326</td><td char="(" align="char">(386&#x02013;2022)</td><td>417</td><td char="(" align="char">(386&#x02013;2022)</td><td>554</td><td char="(" align="char">(386&#x02013;2022)</td></tr><tr><td>&#x02003;CD4&#x02009;+&#x02009;CD45RA&#x02009;+&#x02009;(%Ly)</td><td>2</td><td char="(" align="char">(11&#x02013;38)</td><td>5</td><td char="(" align="char">(11&#x02013;38)</td><td>4</td><td char="(" align="char">(11&#x02013;38)</td><td>7</td><td char="(" align="char">(11&#x02013;38)</td></tr><tr><td>&#x02003;CD4&#x02009;+&#x02009;CD45RA&#x02009;+&#x02009;(abs.Nr/&#x003bc;l)</td><td>95</td><td char="(" align="char">(170&#x02013;1097)</td><td>74</td><td char="(" align="char">(170&#x02013;1097)</td><td>152</td><td char="(" align="char">(170&#x02013;1097)</td><td>176</td><td char="(" align="char">(170&#x02013;1097)</td></tr><tr><td>&#x02003;CD8 (%Ly)</td><td>35</td><td char="(" align="char">(7&#x02013;41)</td><td>39</td><td char="(" align="char">(21&#x02013;43)</td><td>42</td><td char="(" align="char">(7&#x02013;41)</td><td>49</td><td char="(" align="char">(21&#x02013;43)</td></tr><tr><td>&#x02003;CD8 (abs.Nr/&#x003bc;l)</td><td>2150</td><td char="(" align="char">(107&#x02013;1175)</td><td>578</td><td char="(" align="char">(297&#x02013;1011)</td><td>1593</td><td char="(" align="char">(107&#x02013;1175)</td><td>1233</td><td char="(" align="char">(297&#x02013;1011)</td></tr><tr><td>&#x02003;CD8&#x02009;+&#x02009;CD62L&#x02009;+&#x02009;CD45RA&#x02009;+&#x02009;(% of CD8+)</td><td>n.a.</td><td/><td>10,7</td><td char="(" align="char">(25&#x02013;61)</td><td>n.a.</td><td/><td>7,8</td><td char="(" align="char">(25&#x02013;61)</td></tr><tr><td>&#x02003;CD19 (%Ly)</td><td>15</td><td char="(" align="char">(7&#x02013;23)</td><td>7</td><td char="(" align="char">(7&#x02013;23)</td><td>18</td><td char="(" align="char">(7&#x02013;23)</td><td>13</td><td char="(" align="char">(7&#x02013;23)</td></tr><tr><td>&#x02003;CD19 (abs.Nr/&#x003bc;l)</td><td>921</td><td char="(" align="char">(71&#x02013;549)</td><td>104</td><td char="(" align="char">(71&#x02013;549)</td><td>683</td><td char="(" align="char">(71&#x02013;549)</td><td>327</td><td char="(" align="char">(71&#x02013;549)</td></tr><tr><td>&#x02003;CD56 (%Ly)</td><td>25</td><td char="(" align="char">(6&#x02013;29)</td><td>29</td><td char="(" align="char">(6&#x02013;29)</td><td>34</td><td char="(" align="char">(6&#x02013;29)</td><td>25</td><td char="(" align="char">(6&#x02013;29)</td></tr><tr><td>&#x02003;CD56 (abs.Nr/&#x003bc;l)</td><td>1536</td><td char="(" align="char">(98&#x02013;680)</td><td>430</td><td char="(" align="char">(98&#x02013;680)</td><td>1289</td><td char="(" align="char">(98&#x02013;680)</td><td>629</td><td char="(" align="char">(98&#x02013;680)</td></tr><tr><td>&#x02003;CD3 (%Ly)</td><td>10</td><td char="(" align="char">(53&#x02013;85)</td><td>54</td><td char="(" align="char">(53&#x02013;85)</td><td>30</td><td char="(" align="char">(53&#x02013;85)</td><td>62</td><td char="(" align="char">(53&#x02013;85)</td></tr><tr><td>&#x02003;CD3 (abs.Nr/&#x003bc;l)</td><td>614</td><td char="(" align="char">(694&#x02013;2976)</td><td>800</td><td char="(" align="char">(694&#x02013;2976)</td><td>1138</td><td char="(" align="char">(694&#x02013;2976)</td><td>1560</td><td char="(" align="char">(694&#x02013;2976)</td></tr><tr><td>&#x02003;HLA-DR (%Ly)</td><td>82</td><td char="(" align="char">(4&#x02013;18)</td><td>48</td><td char="(" align="char">(10&#x02013;36)</td><td>53</td><td char="(" align="char">(10&#x02013;36)</td><td>74</td><td char="(" align="char">(10&#x02013;36)</td></tr><tr><td>&#x02003;HLA-DR (abs.Nr/&#x003bc;l)</td><td>5036</td><td char="(" align="char">(75&#x02013;505)</td><td>711</td><td char="(" align="char">(200&#x02013;800)</td><td>2010</td><td char="(" align="char">(200&#x02013;800)</td><td>1862</td><td char="(" align="char">(200&#x02013;800)</td></tr><tr><td>&#x02003;CD3&#x02009;+&#x02009;HLA-DR (%Ly)</td><td>7</td><td char="(" align="char">(1&#x02013;8)</td><td>23</td><td char="(" align="char">(2&#x02013;12)</td><td>13</td><td char="(" align="char">(2&#x02013;12)</td><td>44</td><td char="(" align="char">(2&#x02013;12)</td></tr><tr><td>&#x02003;CD3&#x02009;+&#x02009;HLA-DR (abs.Nr/&#x003bc;l)</td><td>430</td><td char="(" align="char">(19&#x02013;219)</td><td>341</td><td char="(" align="char">(20&#x02013;250)</td><td>493</td><td char="(" align="char">(20&#x02013;250)</td><td>1107</td><td char="(" align="char">(20&#x02013;250)</td></tr><tr><td colspan="9">C. Lymphoproliferative response to mitogenic stimulation (3H-thymidine incorporation)</td></tr><tr><td/><td colspan="4">Patient 1 (II - 1)</td><td colspan="4">Patient 2 (II - 2)</td></tr><tr><td/><td colspan="2">22&#x000a0;m</td><td colspan="2">10&#x000a0;year</td><td colspan="2">3&#x000a0;years</td><td colspan="2">8&#x000a0;years</td></tr><tr><td>&#x02003;PHA 1.6&#x003bc;g (dpm)</td><td>10135</td><td>(&#x0003e;&#x02009;=&#x02009;30000)</td><td>101821</td><td>(&#x0003e;&#x02009;=&#x02009;20000)</td><td>25678</td><td>(&#x0003e;&#x02009;=&#x02009;20000)</td><td>72068</td><td>(&#x0003e;&#x02009;=&#x02009;20000)</td></tr><tr><td>&#x02003;CON A 1.2&#x003bc;g (dpm)</td><td>1415</td><td>(&#x0003e;&#x02009;=&#x02009;5200)</td><td>69767</td><td>(&#x0003e;&#x02009;=&#x02009;5000)</td><td>42126</td><td>(&#x0003e;&#x02009;=&#x02009;5000)</td><td>63988</td><td>(&#x0003e;&#x02009;=&#x02009;5000)</td></tr><tr><td>&#x02003;PWM 1:100 (dpm)</td><td>7450</td><td>(&#x0003e;&#x02009;=&#x02009;40000)</td><td>45225</td><td>(&#x0003e;&#x02009;=&#x02009;20000)</td><td>4599</td><td>(&#x0003e;&#x02009;=&#x02009;20000)</td><td>47075</td><td>(&#x0003e;&#x02009;=&#x02009;20000)</td></tr><tr><td>&#x02003;Medium (dpm)</td><td>183</td><td>(&#x0003c;=600)</td><td>60</td><td>(&#x0003c;=400)</td><td>75</td><td>(&#x0003c;=400)</td><td>109</td><td>(&#x0003c;=400)</td></tr><tr><td colspan="9">D. B-cell subpopulations</td></tr><tr><td colspan="5"/><td colspan="2">Patient 1 (II - 1)</td><td colspan="2">Patient 2 (II - 2)</td></tr><tr><td colspan="5"/><td colspan="2">15&#x000a0;years</td><td colspan="2">11&#x000a0;year</td></tr><tr><td colspan="5">&#x02003;CD38+ CD24+ transitional B-cells (%CD19)</td><td>3,8</td><td>(3,9&#x02013;7,8)</td><td>10,7</td><td>(3,9&#x02013;7,8)</td></tr><tr><td colspan="5">&#x02003;CD27-IgD+ na&#x000ef;ve B-cells (%CD19)</td><td>76,3</td><td>(75,2&#x02013;86,7)</td><td>70,1</td><td>(75,2&#x02013;86,7)</td></tr><tr><td colspan="5">&#x02003;CD27+ IgD+ non-switched memory B-cells (%CD19)</td><td>7,6</td><td>(4,6&#x02013;10,2)</td><td>9,3</td><td>(4,6&#x02013;10,2)</td></tr><tr><td colspan="5">&#x02003;CD27+ IgD- switched memory B-cells (%CD19)</td><td>10,9</td><td>(3,3&#x02013;9,6)</td><td>13,3</td><td>(3,3&#x02013;9,6)</td></tr><tr><td colspan="5">&#x02003;CD27-IgD- memory B-cells (%CD19)</td><td>5,2</td><td>(2,3&#x02013;5,5)</td><td>7,2</td><td>(2,3&#x02013;5,5)</td></tr><tr><td colspan="5">&#x02003;CD24-CD38+ plasmablasts (%CD19)</td><td>2,2</td><td>(0,3&#x02013;1,7)</td><td>1,3</td><td>(0,3&#x02013;1,7)</td></tr></tbody></table><table-wrap-foot><p>Pn23, 23-valent pneumococcal polysaccharide vaccine; <italic>HiB</italic> Haemophilus influenzae Type B</p><p>%Ly, percentage of lymphocytes; abs.Nr/&#x003bc;l, absolute number/&#x003bc;l blood; n.a. = data not available</p><p> <italic>PHA</italic> phytohaemagglutinin; <italic>CON</italic> A Concanavalin A; <italic>PWM</italic> Pokeweed-Mitogen; dpm, disintegrations per minute</p><p>%CD19, percentage of CD19+ cells</p><p>Normal ranges are indicated in brackets, next to patient values</p><p>*¹) measured following three vaccinations with Pn23 at the age of 5, 7 and 8&#x000a0;years</p><p>*²) measured following two vaccinations with Pn23 at the age of 3,5 and 7&#x000a0;years</p><p>*³) measured following four vaccinations</p><p>*⁴) measured 3&#x000a0;months after tetanus-diphtheria booster vaccination</p><p>*⁵) measured following a total of six vaccinations against tetanus and diphtheria</p></table-wrap-foot></table-wrap> </p><p id="Par25">The siblings of patient 1 were tested for immunodeficiency due to positive family history. Only one sister (patient 2) showed CD4 lymphopenia on initial determination that improved during follow-up and IgG2-IgG4-subclass deficiency, while generation of selective polysaccharide antibodies was intact (Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>, online supplementary Table&#x000a0;<xref rid="MOESM2" ref-type="media">2</xref>). She received antibiotic prophylaxis until the age of six years. Immunological findings in the other two siblings were within normal range with the exception of a slight hypogammaglobulinemia but</p><p id="Par26">intact antibody production against all antigens tested. Immunological findings in the parents were unremarkable (data not shown).</p></sec><sec id="Sec17"><title>Immunological Phenotype</title><sec id="Sec18"><title>T-Cell Deficiency</title><p id="Par27">The immunological phenotype of both patients is depicted in Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>. A marked reduction in the number of CD4+CD45RA+na&#x000ef;ve T helper cells was observed in both patients, more pronounced in patient 1. Numbers of CD4+ memory T-cells were within the normal range while numbers of na&#x000ef;ve CD8+ cells were also significantly reduced(CD8+CD62L+CD45RA+cells shown in Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>). In contrast, numbers of na&#x000ef;ve B lymphocytes were comparable to an age-matched control. CD4+ and CD3+ lymphocytes were below the normal range in both patients, while numbers of NK cells as identified by CD56 staining were increased initially but returned to normal during follow-up (Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref> and online supplementary Fig&#x000a0;<xref rid="MOESM3" ref-type="media">1</xref>). In patient 1, this predominant CD56+ lymphocyte cell type was further characterized and found to be positive for several markers including CD8, HLA-DR-dim, CD2, CD7, CD11b and negative for CD4 and CD3 (data not shown). Patient 1 initially showed a T-B+NK+SCID phenotype at the age of 22&#x000a0;months with low numbers of both CD4+ and CD8+ T-cells but with a normal number of total CD8 cells, most of which constituted NK-cells. During follow-up, this phenotype evolved into predominant CD4 lymphopenia. A comparable development from a combined CD4+ and CD8+ T-cell deficiency to predominant CD4 lymphopenia was also observed in patient 2 (Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref> and supplementary Figure&#x000a0;<xref rid="MOESM3" ref-type="media">1</xref>). The reduction in na&#x000ef;ve CD4 cells is still detectable although it became less pronounced (Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref> and online supplementary Fig&#x000a0;<xref rid="MOESM3" ref-type="media">1</xref>).</p><p id="Par28">T-cell activation in response to mitogenic stimuli was decreased in both patients during younger age and normalized during follow up (Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>). Further investigation of the capacity of T-cells to respond to activation was performed in patient 1 and revealed a substantial impairment of TCR/CD3-dependent lymphoproliferative responses following stimulation with recall antigen (tetanus toxoid) despite four previous vaccinations; anti-CD3- and staphylococcal superantigen-induced T-cell proliferation was decreased as well (online supplementary table&#x000a0;<xref rid="MOESM1" ref-type="media">1</xref>). Repeated booster vaccinations normalized the T-cell response to recall antigen, while mitogen-induced IL-2 and IFN-gamma release was still impaired (online supplementary table&#x000a0;<xref rid="MOESM1" ref-type="media">1</xref>).</p></sec><sec id="Sec19"><title>Evaluation of Antibody Production</title><p id="Par29">In both patients, IgG2 and IgG4 subclass deficiency was present (Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>). The IgG-response following four tetanus- and <italic>haemophilus influenzae type B</italic> vaccinations was within the normal range, and neutralizing antibodies against all three polio vaccine strains were detectable in patient 1 (serum titer of 1:1,280 against strains I to III). However, IgG antibody production against other vaccination antigens was impaired, diphtheria IgG antibodies remained low in both patients despite four vaccinations in the first two years of life and were only borderline detectable at the age of 10&#x000a0;years despite repeated booster vaccinations (Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>). In patient 1, EBV capsid-IgG antibodies were positive and EBV&#x02013;IgM antibodies negative, indicating a history of primary EBV infection; EBNA-IgG-antibodies remained negative up to the age of nine years (time of last follow-up control without SCIG substitution therapy). She had also experienced previous CMV infection as CMV-IgG antibodies were positive and CMV&#x02013;IgM antibodies negative.</p><p id="Par30">Patient 1 displayed a selective IgG antibody deficiency against bacterial polysaccharide antigens while in patient 2 this B-cell function was within the lower normal range (Online supplementary table&#x000a0;<xref rid="MOESM2" ref-type="media">2</xref>). At the age of 2 to 3&#x000a0;years, pneumococcal IgG antibodies were undetectable in the serum of both patients who had never been vaccinated against pneumococcal polysaccharide before. Patient 1 failed to mount a significant IgG antibody response following each of three pneumococcal vaccinations at the age of 5, 7 and 8&#x000a0;years, while the IgM response was within the normal range. In contrast, IgG antibody production was within the lower normal range after two pneumococcal vaccinations in patient 2, and she developed a significant IgM response (online supplementary table&#x000a0;<xref rid="MOESM2" ref-type="media">2</xref>).</p></sec></sec><sec id="Sec20"><title>Molecular Genetic Analysis of the Underlying Defect</title><sec id="Sec21"><title>Analysis of Known Molecular Defects of T-Cell Deficiency</title><p id="Par31">Several molecular defects known to lead to primary T-cell deficiency were examined in patient 1 and found to be normal, adenosine deaminase and purine nucleoside phosphorylase deficiency (examined by measuring the respective enzyme in peripheral blood erythrocytes), a defect in chromosome breakage repair (examined following treatment of cultured cells with Diepoxybutan), <italic>NBS1</italic>-gene deletion, and 22q11 microdeletion were excluded. Levels of alpha-1-fetoprotein in serum as a potential indication for ataxia telangiectasia as well as flow cytometric analysis of IL-7R&#x003b1;, IL-2R&#x003b3;, CD3 und CD45 expression in peripheral blood leukocytes showed normal results.</p></sec><sec id="Sec22"><title>Identification of a Mutation in <italic>JAK3</italic></title><p id="Par32">Given the family history including consanguinity, an autosomal recessive Mendelian trait was suspected. Thus, we employed a combination of homozygosity mapping and exome sequencing to detect the underlying genetic cause of the disease.</p><p id="Par33">Homozygosity mapping of the affected siblings revealed seven and thirteen homozygous intervals, respectively. Three intervals on chromosome six, eight and nineteen were homozygous in both patients (Fig&#x000a0;<xref rid="Fig1" ref-type="fig">1a</xref>). As the causative genetic defect was assumed to be located in an overlapping homozygous interval, exome sequencing data were filtered for the described homozygous intervals (Fig&#x000a0;<xref rid="Fig1" ref-type="fig">1b</xref>). Among those was a homozygous variant in the <italic>JAK3</italic> gene (c.T3196C, p.Cys1066Arg) encoding a tyrosine kinase bound to the common gamma chain of various interleukin receptors [<xref ref-type="bibr" rid="CR20">20</xref>]. JAK3 deficiency is commonly known to cause SCID [<xref ref-type="bibr" rid="CR21">21</xref>]. Using Sanger sequencing, this variant could be validated and it showed perfect segregation with the disease (Fig&#x000a0;<xref rid="Fig1" ref-type="fig">1c</xref>). Multiple sequence alignment revealed that this position is highly conserved throughout evolution (Fig&#x000a0;<xref rid="Fig1" ref-type="fig">1d</xref>) with a potentially critical function within the kinase domain of JAK3 (Fig&#x000a0;<xref rid="Fig1" ref-type="fig">1e</xref>).<fig id="Fig1"><label>Fig. 1</label><caption><p> <bold>a</bold> Homozygosity mapping results showing homozygous intervals highlighted in red color <bold>b</bold> Filtering strategy for exome sequencing data <bold>c</bold> Pedigree of the index family with sequence of the <italic>JAK3</italic> mutation site highlighted in gray background <bold>d</bold> Multiple sequence alignment with the <italic>JAK3</italic> mutation site highlighted in red background <bold>e</bold> Crystal structure of the JAK3 kinase domain. The described mutation site is marked with a black arrow</p></caption><graphic xlink:href="10875_2014_88_Fig1_HTML" id="MO1"/></fig> </p></sec><sec id="Sec23"><title>V&#x003b2; Spectratyping Indicates Restricted T-Cell Receptor (TCR) Repertoire</title><p id="Par34">TCR V&#x003b2; spectratyping is a well-established method to assess TCR diversity and restriction. Several T-cell deficiencies are known to use a restricted TCR V&#x003b2; repertoire [<xref ref-type="bibr" rid="CR22">22</xref>, <xref ref-type="bibr" rid="CR23">23</xref>]. We employed TCR V&#x003b2; spectratyping to assess the clonality of the TCR repertoire and observed a restricted TCR repertoire in both index patients (Fig&#x000a0;<xref rid="Fig2" ref-type="fig">2a</xref> and online supplementary Fig&#x000a0;<xref rid="MOESM4" ref-type="media">2</xref>).<fig id="Fig2"><label>Fig. 2</label><caption><p> <bold>a</bold> Subfamilies V&#x003b2;20-22 of normal donor and patient 1 are shown as representatives of V&#x003b2; TCR Spectratyping data. <bold>b</bold> Sequencing of the <italic>JAK3</italic> in genomic DNA derived from sorted lymphocyte subsets of both index patients. The mutation site is highlighted with a gray background</p></caption><graphic xlink:href="10875_2014_88_Fig2_HTML" id="MO2"/></fig> </p></sec><sec id="Sec24"><title>Detection of Somatic Chimerism</title><p id="Par35">Somatic chimerism has been observed in several primary immunodeficiencies such as Wiskott Aldrich Syndrome [<xref ref-type="bibr" rid="CR24">24</xref>&#x02013;<xref ref-type="bibr" rid="CR27">27</xref>], ADA-deficiency [<xref ref-type="bibr" rid="CR28">28</xref>, <xref ref-type="bibr" rid="CR29">29</xref>] or X-linked SCID [<xref ref-type="bibr" rid="CR30">30</xref>, <xref ref-type="bibr" rid="CR31">31</xref>] and is possibly associated with an amelioration of the clinical phenotype of disease [<xref ref-type="bibr" rid="CR32">32</xref>]. Thus, we speculated whether somatic chimerism may have contributed to the relatively mild clinical phenotype despite the homozygous <italic>JAK3</italic> mutation in a critical protein domain in both index patients. Hence, CD4+ T-cells, CD8+ T-cells and CD19+ B-cells were FACS-sorted, followed by Sanger sequencing for an amplicon harboring the identified <italic>JAK3</italic> mutation. CD8+ T-cells of both index patients showed somatic chimerism whereas revertant CD4+ T-cells were present only in the second patient (Fig&#x000a0;<xref rid="Fig2" ref-type="fig">2b</xref>). Whether this finding contributes to the milder clinical phenotype in this patient is unclear at the moment, as further studies are required to formally prove this assumption. To exclude maternal chimerism, DNA derived from total leukocyte preparations from peripheral blood was assessed and revealed no indication of maternal cells above the detection limit of 1 % (data not shown).</p></sec></sec><sec id="Sec25"><title>Biological Relevance of the Observed <italic>JAK3</italic> Mutation for Cell Signaling Events</title><sec id="Sec26"><title>Analysis of JAK3 and STAT5 Protein Expression in B-cell Lines</title><p id="Par36">To further study possible biological consequences of the observed <italic>JAK3</italic> mutation on a cellular level, we generated EBV-transformed B-cell lines from the patients homozygous for the mutation, their heterozygous father and an unrelated healthy control. Sanger sequencing confirmed that the patients&#x02019; B-cell lines indeed showed a homozygous <italic>JAK3</italic> mutation, while heterozygous expression of this mutation was detected in the father&#x02019;s B-cell line (online supplementary Fig&#x000a0;<xref rid="MOESM6" ref-type="media">3</xref>). We then examined protein lysates from EBV-transformed B-cells to assess whether the mutation led to destabilization and consequently reduced protein levels of the altered JAK3. Western blot analysis showed presence of JAK3 protein in both patients at reduced levels compared to her father, the normal control or a &#x003b3;c-deficient XSCID patient (Fig&#x000a0;<xref rid="Fig3" ref-type="fig">3a</xref>). STAT5 protein expression was comparable in all cell lines tested (data not shown).<fig id="Fig3"><label>Fig. 3</label><caption><p> <bold>a</bold> Analysis of JAK3 protein expression in B-cell lines of a healthy control (1), father of the patients (2), patient 1 (3), patient 2 (4) and a &#x003b3;c-deficient SCID patient (5). Examination of GAPDH protein expression served as a loading control. <bold>b</bold> Analysis of JAK3 signaling function in B-cell lines of a healthy control, patient 1 (II-1), her father (I-1) and a &#x003b3;c-deficient SCID patient after stimulation with IL-4 and IL-21. Numbers in the top left and right corners of each histogram in panel B indicate the Mean Fluorescence Intensity (MFI) values of unstimulated and cytokine-stimulated cells, respectively. <bold>c</bold> Analysis of JAK3 signaling function in CD4+ peripheral blood T-cells of a healthy control, patient 1 (II-1) and patient 2 (II-2) after stimulation with IL-2. Histogram overlays represent intracellular levels of phosphorylated STAT5 in CD4+ T-cells without stimulation or after stimulation with IL-2 or IL-6. Numbers in the left top corner and middle part of each histogram indicate percentages of cells with a positive staining for pSTAT5 following stimulation with IL-2 or IL-6, respectively, while numbers in right corner constitute percentages of pSTAT5-positive unstimulated control cells</p></caption><graphic xlink:href="10875_2014_88_Fig3_HTML" id="MO3"/></fig> </p></sec><sec id="Sec27"><title>Analysis of JAK3 Signaling Function in B-cell Lines and CD4+ T-Cells</title><p id="Par37">IL-2 activates STAT5, IL-4 activates STAT5 and STAT6 [<xref ref-type="bibr" rid="CR33">33</xref>], while IL-21 activates STAT1, STAT3 and STAT5 in a JAK3-dependent manner [<xref ref-type="bibr" rid="CR34">34</xref>]. We exposed EBV-transformed B-cells from patient 1 and her father to IL-4 and IL-21 and assessed activation of STAT3 and STAT5 by flow cytometry with phospho-STAT (p-STAT) specific antibodies. Results show that STAT3 phosphorylation was clearly detectable but diminished to a different extent in patient 1 and her father as compared to the healthy control (Fig&#x000a0;<xref rid="Fig3" ref-type="fig">3b</xref>). Furthermore, STAT5 phosphorylation was severely decreased in patient 1 while being moderately reduced in her father. In a second run, B-cells of patient 1 and 2 were stimulated with IL-2 in order to quantify phosphorylation of STAT5 (online supplementary Fig&#x000a0;<xref rid="MOESM6" ref-type="media">4</xref>). Compared to the healthy control, it was remarkably diminished but still detectable. Both experiments showed that neither STAT3 nor STAT5 phosphorylation were activated in a &#x003b3;c-deficient XSCID B-cell line. We next investigated JAK3-signaling in CD4+ T-cells from the two patients and a healthy control. PBMCs were stimulated with IL-2 before activation of STAT5 was assessed by dual colour flow cytometry and gating on CD4+ lymphocytes. We observed that the presence of revertant CD4+ T-cells in patient 2 had no effect on IL-2-induced, JAK3-dependent STAT5-phosphorylation which was diminished to a comparable extent in both patients (Fig.&#x000a0;<xref rid="Fig3" ref-type="fig">3c</xref>). Since IL-6 is known to induce both STAT3 phosphorylation [<xref ref-type="bibr" rid="CR35">35</xref>] and, to a lesser extent, STAT5 phosphorylation [<xref ref-type="bibr" rid="CR36">36</xref>], in a JAK3-independent manner, we investigated intactness of these pathways in PBMCs from the patients under study. IL-6-induced STAT5-phosphorylation (shown in Fig.&#x000a0;<xref rid="Fig3" ref-type="fig">3c</xref>) as well as STAT3-phosphorylation (29,8 % in healthy control and 29,8 % in patient 2, data not illustrated) were both normal. This confirms the presence of a JAK3-specific signaling defect in the patients&#x02019; CD4+ T-cells.</p></sec></sec></sec><sec id="Sec28" sec-type="discussion"><title>Discussion</title><p id="Par38">JAK3 is an intracellular protein tyrosine kinase which is predominantly expressed in hematopoietic cells and belongs to the Janus kinase family. It binds to the common gamma chain (&#x003b3;c) of receptors of interleukin(IL)-2, IL-4, IL-7, IL-9, IL-15 and IL-21 and plays an essential role in cytokine receptor signaling pathway [<xref ref-type="bibr" rid="CR37">37</xref>]. Upon ligand binding of the IL receptors, JAK3 is activated by autophosphorylation, enabling binding and phosphorylation of different STAT proteins including STAT1, 3, 5 and 6 [<xref ref-type="bibr" rid="CR38">38</xref>, <xref ref-type="bibr" rid="CR39">39</xref>]. Consequently, STATs can build homodimers (and heterodimers [<xref ref-type="bibr" rid="CR40">40</xref>]) and translocate to the cell nucleus in order to regulate the expression of several genes involved in development, proliferation and function of lymphoid cells [<xref ref-type="bibr" rid="CR41">41</xref>].</p><p id="Par39">Similar to &#x003b3;c chain-deficient X-linked SCID, deficiency of JAK3 is known to cause autosomal recessive SCID with a reduced number of T-cells and natural killer cells as well as dysfunctional B-cells in normal cell count (T^&#x02212;B⁺NK^&#x02212; SCID) and hypogammaglobulinemia [<xref ref-type="bibr" rid="CR42">42</xref>]. As in other SCID subtypes, affected children typically suffer from recurrent or persistent infections, often with opportunistic pathogens, intractable diarrhea, thrush and failure to thrive. Without prompt and adequate therapy by means of stem cell transplantation, they have a significant mortality during the first two years of life [<xref ref-type="bibr" rid="CR41">41</xref>, <xref ref-type="bibr" rid="CR43">43</xref>, <xref ref-type="bibr" rid="CR44">44</xref>]. Recently, hypomorphic mutations in <italic>JAK3</italic> have been reported to widen the spectrum of clinical and immunological phenotypes of JAK3 deficiency to include T-cell lymphopenia with maternal T-cell engraftment and defective antibody responses [<xref ref-type="bibr" rid="CR45">45</xref>].</p><p id="Par40">Due to the severe phenotype associated with deficiency of JAK3, it was surprising to find a mutation in <italic>JAK3</italic> in the index family with a relatively mild phenotype in terms of CID evolving into predominant CD4 deficiency. In retrospect, patient 1 showed a T-B+NK+SCID at the age of 22&#x000a0;months with a low number of CD3 lymphocytes, low numbers of CD4+ and CD8+ T-cells but with a normal number of total CD8 cells, most of which constituted NK-cells. During follow-up, this phenotype evolved into predominant CD4 lymphopenia. A comparable development from (S)CID to predominant CD4 lymphopenia was also observed in patient 2 (Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref> and supplementary Fig.&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>).</p><p id="Par41">The described novel homozygous mutation in <italic>JAK3</italic> (c.T3196C, p.Cys1066Arg) affects a highly conserved amino acid site which is located in the kinase domain of JAK3 [<xref ref-type="bibr" rid="CR46">46</xref>]. Similar to other JAK3 deficient patients [<xref ref-type="bibr" rid="CR22">22</xref>, <xref ref-type="bibr" rid="CR23">23</xref>], we observed an oligoclonal restriction of the TCR V&#x003b2; repertoire in both index patients (Fig&#x000a0;<xref rid="Fig2" ref-type="fig">2a</xref> and online supplementary Fig&#x000a0;<xref rid="MOESM4" ref-type="media">2</xref>).</p><p id="Par42">To investigate the molecular etiology of the relatively mild clinical disease phenotype in both index patients, we hypothesized that (i) the novel missense mutation may be of hypomorphic nature or (ii) a molecular chimerism event may have occurred in these patients.</p><p id="Par43">In other primary immunodeficiency disorders, such molecular chimerism has been linked to atypical and relatively milder disease courses, including ADA-deficient SCID [<xref ref-type="bibr" rid="CR28">28</xref>, <xref ref-type="bibr" rid="CR29">29</xref>], X-linked ectodermal dysplasia and immunodeficiency [<xref ref-type="bibr" rid="CR47">47</xref>], &#x003b3;c deficiency [<xref ref-type="bibr" rid="CR30">30</xref>, <xref ref-type="bibr" rid="CR31">31</xref>] or Wiskott-Aldrich syndrome [<xref ref-type="bibr" rid="CR24">24</xref>&#x02013;<xref ref-type="bibr" rid="CR27">27</xref>]. Our results from Sanger sequencing of DNA extracted from FACS-sorted leukocyte subsets indeed revealed somatic chimerism in CD8+ T-cells in both patients and additional chimerism in CD4+ T-cells in patient 2 (Fig&#x000a0;<xref rid="Fig2" ref-type="fig">2b</xref>). In view of this constellation, it is interesting to note that patient 1 with no somatic reversion in CD4+ T-cells presented in childhood with recurrent infections of the respiratory tract and received IVIG substitution therapy whereas patient 2 with somatic chimerism in CD4+ T-cells showed no clinical sign of immunodeficiency and merely laboratory investigations showed CD4+ lymphocytopenia and IgG subclass deficiency. The presence of somatic reversion in both patients of the same family despite somatic reversion being a rare event indicates a selective proliferative advantage of lymphocytes with a reverted over non-reverted <italic>JAK3</italic> genotype due to restored development and differentiation. Although it is tempting to assume that the somatic chimerism may have attenuated the severity of the disease phenotype in the described individuals, formal proof of this assumption would clearly require further studies. Preliminary results show that the presence of revertant cells had no effect on the level of residual JAK3 signaling activity when CD4+ T-cells from both patients were compared. Furthermore, the present data do not allow for exact quantification of the percentage of revertant cells contained within the T-cell subsets.</p><p id="Par44">Interestingly, analyses of phosphorylation of STAT3 and STAT5 in an EBV-transformed B-cell line from patient 1 and from the heterozygous father after stimulation with IL-4 and IL-21 showed detectable levels of phosphorylated STATs that were diminished to a different extent. Although reduced JAK3 protein expression might have contributed to this finding, it suggests that the specific <italic>JAK3</italic> mutation in this family may be hypomorphic, thus allowing for residual &#x003b3;c- and JAK3-mediated signaling function, which is a likely explanation for the relatively mild clinical phenotype observed in our JAK3-deficient patients. The phenomenon of null or hypomorphic mutations in the same gene causing different phenotypes of immunodeficiencies has already been described previously. For instance, null mutations of <italic>RAG1</italic> are known to cause SCID while hypomorphic mutations have been associated with Omenn syndrome [<xref ref-type="bibr" rid="CR48">48</xref>] or idiopathic CD4 lymphopenia [<xref ref-type="bibr" rid="CR13">13</xref>].</p></sec><sec id="Sec29" sec-type="conclusion"><title>Conclusions</title><p id="Par45">Taken together, we here describe for the first time JAK3 deficiency due to a hypomorphic <italic>JAK3</italic> mutation and with somatic chimerism, causing a phenotype of T-cell deficiency evolving into predominant CD4 lymphopenia. It is conceivable that other patients with primary CD4 lymphopenia may, in a similar fashion, bear hypomorphic mutations and/or somatic chimerism in other genes which are usually associated with SCID phenotypes. However, it appears likely that other subgroups of patients with CD4 lymphopenia are caused by novel nosological entities involved in T-cell homeostasis. Due to improvement of genomic technologies, the number of gene defects causing an incomplete impairment of T-cell development is increasing, leading to a broader understanding of normal and pathologic immune system development [<xref ref-type="bibr" rid="CR49">49</xref>]. The availability of state-of-the art genomic technologies such as so-called &#x0201c;next generation sequencing&#x0201d; approaches will be instrumental in defining and classifying the range of genomic variations underlying this group of immunodeficiencies.</p></sec><sec sec-type="supplementary-material"><title>Electronic supplementary material</title><sec id="Sec30"><p>Below is the link to the electronic supplementary material.<supplementary-material content-type="local-data" id="MOESM1"><media xlink:href="10875_2014_88_MOESM1_ESM.pdf"><label>ESM 1</label><caption><p>(PDF 123 kb)</p></caption></media></supplementary-material> <supplementary-material content-type="local-data" id="MOESM2"><media xlink:href="10875_2014_88_MOESM2_ESM.pdf"><label>ESM 2</label><caption><p>(PDF 118 kb)</p></caption></media></supplementary-material> <supplementary-material content-type="local-data" id="MOESM3"><media xlink:href="10875_2014_88_MOESM3_ESM.pdf"><label>ESM 3</label><caption><p>(PDF 243&#x000a0;kb)</p></caption></media></supplementary-material> <supplementary-material content-type="local-data" id="MOESM4"><media xlink:href="10875_2014_88_MOESM4_ESM.pdf"><label>ESM 4</label><caption><p>(PDF 582&#x000a0;kb)</p></caption></media></supplementary-material> <supplementary-material content-type="local-data" id="MOESM5"><media xlink:href="10875_2014_88_MOESM5_ESM.pdf"><label>ESM 5</label><caption><p>(PDF 253&#x000a0;kb)</p></caption></media></supplementary-material> <supplementary-material content-type="local-data" id="MOESM6"><media xlink:href="10875_2014_88_MOESM6_ESM.pdf"><label>ESM 6</label><caption><p>(PDF 299&#x000a0;kb)</p></caption></media></supplementary-material> </p></sec></sec>

A Community Engagement Process Identifies Environmental Priorities to Prevent Early Childhood Obesity: The Children&#x02019;s Healthy Living (CHL) Program for Remote Underserved Populations in the US Affiliated Pacific Islands, Hawaii and Alaska

Could not extract abstract

<contrib contrib-type="author" corresp="yes"><name><surname>Fialkowski</surname><given-names>Marie Kainoa</given-names></name><address><phone>+808-956-8337</phone><email>mariekf@hawaii.edu</email></address><xref ref-type="aff" rid="Aff1">1</xref></contrib><contrib contrib-type="author"><name><surname>DeBaryshe</surname><given-names>Barbara</given-names></name><address><email>debarysh@hawaii.edu</email></address><xref ref-type="aff" rid="Aff1">1</xref></contrib><contrib contrib-type="author"><name><surname>Bersamin</surname><given-names>Andrea</given-names></name><address><email>abersamin@alaska.edu</email></address><xref ref-type="aff" rid="Aff2">2</xref></contrib><contrib contrib-type="author"><name><surname>Nigg</surname><given-names>Claudio</given-names></name><address><email>cnigg@hawaii.edu</email></address><xref ref-type="aff" rid="Aff1">1</xref></contrib><contrib contrib-type="author"><name><surname>Leon Guerrero</surname><given-names>Rachael</given-names></name><address><email>rachaeltlg@uguam.uog.edu</email></address><xref ref-type="aff" rid="Aff3">3</xref></contrib><contrib contrib-type="author"><name><surname>Rojas</surname><given-names>Gena</given-names></name><address><email>grojas@uguam.uog.edu</email></address><xref ref-type="aff" rid="Aff3">3</xref></contrib><contrib contrib-type="author"><name><surname>Areta</surname><given-names>Aufa&#x02019;i Apulu Ropeti</given-names></name><address><email>aareta@yahoo.com</email></address><xref ref-type="aff" rid="Aff4">4</xref></contrib><contrib contrib-type="author"><name><surname>Vargo</surname><given-names>Agnes</given-names></name><address><email>amsvargo@yahoo.com</email></address><xref ref-type="aff" rid="Aff4">4</xref></contrib><contrib contrib-type="author"><name><surname>Belyeu-Camacho</surname><given-names>Tayna</given-names></name><address><email>tayna.belyeu-camacho@marianas.edu</email></address><xref ref-type="aff" rid="Aff5">5</xref></contrib><contrib contrib-type="author"><name><surname>Castro</surname><given-names>Rose</given-names></name><address><email>rose.castro@marianas.edu</email></address><xref ref-type="aff" rid="Aff5">5</xref></contrib><contrib contrib-type="author"><name><surname>Luick</surname><given-names>Bret</given-names></name><address><email>bluick@alaska.edu</email></address><xref ref-type="aff" rid="Aff2">2</xref></contrib><contrib contrib-type="author"><name><surname>Novotny</surname><given-names>Rachel</given-names></name><address><email>novotny@hawaii.edu</email></address><xref ref-type="aff" rid="Aff1">1</xref></contrib><contrib contrib-type="author"><collab>the CHL Team</collab></contrib><aff id="Aff1"><label>1</label>University of Hawai&#x02018;i at Manoa, Honolulu, HI 96822 USA </aff><aff id="Aff2"><label>2</label>University of Alaska at Fairbanks, Fairbanks, AK USA </aff><aff id="Aff3"><label>3</label>University of Guam, Mangilao, GU USA </aff><aff id="Aff4"><label>4</label>American Samoa Community College, Mapusaga, AS USA </aff><aff id="Aff5"><label>5</label>Northern Marianas College, Saipan, MP USA </aff>

Maternal and Child Health Journal

<sec id="Sec1" sec-type="introduction"><title>Introduction</title><p>Childhood obesity prevalence and its associated health complications have become a major national and global public health issue. Obese and overweight children are at risk for serious chronic illnesses [<xref ref-type="bibr" rid="CR1">1</xref>&#x02013;<xref ref-type="bibr" rid="CR16">16</xref>]. Striking disparity is found in US childhood obesity prevalence; indigenous groups, including US Affiliated Pacific Islanders, Native Hawaiians and Alaska Natives, are disproportionately affected [<xref ref-type="bibr" rid="CR17">17</xref>&#x02013;<xref ref-type="bibr" rid="CR23">23</xref>]. For example, a state of emergency has been declared in the US Affiliated Pacific Islands (USAPI) due to the high prevalence of chronic health conditions in both adults and children in these island communities [<xref ref-type="bibr" rid="CR24">24</xref>].</p><p>Individual level obesity prevention efforts promote short-term behavior change but may not have a significant or sustainable impact on childhood obesity [<xref ref-type="bibr" rid="CR25">25</xref>] amidst an obesogenic environment [<xref ref-type="bibr" rid="CR26">26</xref>]. Since the environment, i.e., the cumulative living conditions surrounding a child, is associated with childhood obesity [<xref ref-type="bibr" rid="CR27">27</xref>&#x02013;<xref ref-type="bibr" rid="CR30">30</xref>], interventions for young children that use sustainable, multi-strategy, and multi-setting approaches are needed [<xref ref-type="bibr" rid="CR31">31</xref>]. An ecological approach, which targets the individual, social and built environments, and policies [<xref ref-type="bibr" rid="CR32">32</xref>, <xref ref-type="bibr" rid="CR33">33</xref>], expands the potential of prevention efforts to address critical upstream determinants of obesity-related behaviors, to influence larger populations and to have a long-term, sustainable impact [<xref ref-type="bibr" rid="CR34">34</xref>]. Although evidence from ecological approaches is promising [<xref ref-type="bibr" rid="CR35">35</xref>, <xref ref-type="bibr" rid="CR36">36</xref>], larger scale, adequately powered studies are needed.</p><p>A process for understanding which ecological approaches are most appropriate and have the highest probability of success over the long term is also needed. Young children are especially sensitive to environmental changes given their rapid growth and captive state as they are less able to exert personal choice within their family, school, and community environments [<xref ref-type="bibr" rid="CR37">37</xref>]. Focusing on the environment as a mode for intervening to prevent early childhood obesity requires partnering with people that have first-hand knowledge of that environment (i.e., the people who live/experience that setting) to ensure applicability. In remote, underserved minority populations, such as within the USAPI, Hawaii and Alaska (will be referred to as USAPI/HI/AK throughout the remainder of the paper), a community-based approach to environmental interventions may be especially appropriate [<xref ref-type="bibr" rid="CR38">38</xref>]. Community-based approaches allow for cultural context to be applied, trusting relationships to be forged and contributes to leveling the playing field between community members and researchers [<xref ref-type="bibr" rid="CR39">39</xref>&#x02013;<xref ref-type="bibr" rid="CR42">42</xref>]. Community-based approaches focus on establishing relationships with community &#x0201c;experts&#x0201d; to build the community&#x02019;s capacity to promote the desired outcome.</p><p>Community-based processes are also highly compatible with an assets-based philosophy, such as positive deviance, which is an approach to identify locally available, sustainable, and effective strategies suitable for a community [<xref ref-type="bibr" rid="CR43">43</xref>, <xref ref-type="bibr" rid="CR44">44</xref>]. Positive deviance is based on the observation that in communities a few at-risk individuals follow uncommon, beneficial practices that result in better health outcomes than their neighbors who share similar risks [<xref ref-type="bibr" rid="CR43">43</xref>]. Drawing on the local knowledge from these few individuals (i.e., positive deviants) to develop interventions has the potential to increase affordability, acceptability and sustainability of community-based action, since local culture is already well-integrated into the behaviors/practices that resulted in a positive outcome [<xref ref-type="bibr" rid="CR43">43</xref>, <xref ref-type="bibr" rid="CR45">45</xref>].</p><p>Community-based approaches may be particularly critical for indigenous populations for whom mainstream models of childhood obesity prevention may have limited application. Through community partnerships cultural attunement is ensured leading to the strengthening of study design and implementation to more effectively address complex problems [<xref ref-type="bibr" rid="CR41">41</xref>, <xref ref-type="bibr" rid="CR42">42</xref>, <xref ref-type="bibr" rid="CR46">46</xref>]. Community-based processes have been demonstrated to be a critical step to developing sustainable and successful young child obesity prevention programs for islands in the South Pacific [<xref ref-type="bibr" rid="CR47">47</xref>, <xref ref-type="bibr" rid="CR48">48</xref>].</p><p>This paper describes the community engagement process (CEP) used by the Children&#x02019;s Healthy Living (CHL) Program for remote underserved minority populations in the USAPI/HI/AK. CHL used the CEP to seek alignment and collaboration with community partners throughout Alaska, American Samoa, the Commonwealth of the Northern Mariana Islands (CNMI), Guam, and Hawai&#x02018;i to meaningfully address childhood obesity. This paper also highlights the overall priorities for environmental intervention strategies identified by communities for community-based, environmentally focused childhood obesity prevention in the USAPI/HI/AK Region and the lessons learned from the CEP. The information presented here can guide future children&#x02019;s obesity prevention programs and policies and serve as a model for other island regions with remote, underserved native populations at high risk for obesity.</p></sec><sec id="Sec2"><title>Description of the Region and Program</title><sec id="Sec3"><title>The US Affiliated Pacific Island (USAPI), Hawaii and Alaska Region (USAPI/HI/AK)</title><p>The USAPI/HI/AK region is vast and isolated, covering more area in the Pacific Ocean (one million square miles) than the contiguous US does on land. The ocean is viewed as part of the natural resource of the region, as land is on the American continent. The remote vastness of the region promotes a multitude of small, diverse, widely dispersed cultures (including subsistence cultures) living in unique environments with delicate ecosystems. Nonetheless, the USAPI/HI/AK is characterized by a number of shared strengths that can be leveraged to promote healthy living. Indigenous groups in the USAPI/HI/AK maintain a strong traditional culture that includes valuing the group in favor of the individual, respecting elders and the family unit, and prizing healthy subsistence foods (e.g., taro, fish). These attributes, coupled with the unifying US land grant college infrastructure throughout the region, present a unique opportunity to join together to create a larger voice to address childhood obesity and improve child health in the USAPI/HI/AK.</p></sec><sec id="Sec4"><title>The CHL Program</title><p>Collaborators from land grant colleges and universities [<xref ref-type="bibr" rid="CR49">49</xref>, <xref ref-type="bibr" rid="CR50">50</xref>] in Alaska, American Samoa, CNMI, Guam, the Federated States of Micronesia (FSM), Hawai&#x02018;i, the Republic of the Marshall Islands (RMI), and the Republic of Palau (RP) formed the CHL Program. The land grant system [<xref ref-type="bibr" rid="CR49">49</xref>, <xref ref-type="bibr" rid="CR50">50</xref>], one of the few unifying institutions across the USAPI/HI/AK, provided a suitable infrastructure for regional collaboration. The CHL Program was developed in response to the United States Department of Agriculture (USDA) Agriculture and Food Research Initiative to develop a multi-state/jurisdiction, multi-institutional, and multi-disciplinary team to integrate knowledge about child nutrition, physical activity, and social and environmental influences on childhood obesity in order to develop and implement a large-scale, multifaceted, community-based, and environmentally-focused intervention for preventing early childhood obesity (ages 2&#x02013;8&#x000a0;years).</p><p>The overall goal of the CHL program is to strengthen the children&#x02019;s environments to better promote active play and intake of healthy food in order to prevent early childhood obesity in the USAPI/HI/AK, which are located in the northern Pacific Ocean (except American Samoa). For CHL, the environment is broadly construed and includes the social, cultural, physical/built [<xref ref-type="bibr" rid="CR51">51</xref>], political, and economic contexts of children&#x02019;s lives. CHL seeks to develop, implement, and evaluate a community-based environmental intervention to address six target health behaviors proposed by the investigators and subsequently required by the funder (USDA): (a) increase the consumption of fruits and vegetables (b) increase water intake, (c) decrease intake of sugar-sweetened beverages, (d) increase physical activity, (e) increase the duration of sleep, and (f) decrease screen time (e.g., TV and recreational screen use).</p></sec></sec><sec id="Sec5"><title>Methods: The CHL Community Engagement Process</title><p>The overall goal of the CHL CEP was to foster partnerships with CHL communities to jointly develop a community-based, multi-level, sustainable environmental intervention to prevent childhood obesity. The CHL CEP was informed by the analysis grid for elements linked to obesity (ANGELO) action model, a community and ecologically based framework used to develop environmental interventions to reduce childhood obesity in three island nations and a community in Australia, located in the South Pacific region [<xref ref-type="bibr" rid="CR31">31</xref>, <xref ref-type="bibr" rid="CR37">37</xref>, <xref ref-type="bibr" rid="CR52">52</xref>, <xref ref-type="bibr" rid="CR53">53</xref>]. The ANGELO action model includes both a conceptual framework for analyzing obesigenic environments and a process model for engaging community stakeholders. In the ANGELO conceptual framework, environments are cross-categorized by size and type [<xref ref-type="bibr" rid="CR37">37</xref>]. In the ANGELO process model, community members and researchers use the conceptual framework to analyze the assets and liabilities of a community&#x02019;s environment, prioritize areas amendable to productive change, and develop an action plan [<xref ref-type="bibr" rid="CR37">37</xref>]. Therefore, the CHL CEP was a multi-step process guided by a CHL specific conceptual model that engaged key stakeholders through a local advisory committee, key informant interviews, community meetings (CM) and community feedback meetings (CFM) (see Fig.&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref> for a description of the purpose, membership, and process of each group and Fig.&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref> for the CHL conceptual framework for community engagement). In this paper we focus on the CM and CFM, the goals of which were to (a) identify each CHL community&#x02019;s assets and needs relating to healthy eating and active living, and (b) prioritize environmental intervention strategies relating to healthy eating and active living in order to inform intervention development.<fig id="Fig1"><label>Fig.&#x000a0;1</label><caption><p>Children&#x02019;s Healthy Living (CHL) Program community engagement process (CEP)</p></caption><graphic xlink:href="10995_2013_1353_Fig1_HTML" id="MO1"/></fig> <fig id="Fig2"><label>Fig.&#x000a0;2</label><caption><p>Children&#x02019;s Healthy Living (CHL) Program conceptual framework for community engagement.<italic> F&#x00026;V</italic> fruit &#x00026; vegetable,<italic> SSB</italic> sugar-sweetened beverage,<italic> PA</italic> physical activity. The<italic> double solid line boxes</italic> represent the CHL primary objective of promoting a healthy child through a healthy weight. The<italic> solid line boxes</italic> relate to the six CHL target health behavioral objectives required by the funding agency. The<italic> dash line boxes</italic> relate to factors that influence the attainment of the CHL target health behavioral objectives: identifying resource types, availability and ease of access; possible intervention strategies prioritized by importance and feasibility; existing challenges to healthy behavior and the potential malleability of these obstacles. The<italic> dotted line boxes</italic> relate to the environmental domains that the factors that influence the attainment of the CHL target health behavioral objectives (see schematic: resources/availability, strategies/importance, and challenges/chanegability) operate in.</p></caption><graphic xlink:href="10995_2013_1353_Fig2_HTML" id="MO2"/></fig> </p><p>Institutional Review Board approval from the University of Alaska Fairbanks, University of Guam, and University of Hawai&#x02018;i at Manoa were attained prior to the initiation of the CEP. American Samoa Community College and the Northern Mariana College ceded their Institutional Review to the University of Hawai&#x02018;i at Manoa.</p><p>In addition, approvals for working with PreSchool and Head Start (a US federally funded program that educates preschool-age children and their families) teachers and parents were received in coordination with the program directors and/or boards (e.g., Tribal), when appropriate, in all jurisdictions. Other local level approvals included approvals from the chiefs (matai) and ministers (faifeau) of pertinent American Samoan villages and the participating village mayors in Guam.</p><sec id="Sec6"><title>Selection of Target Communities within Each Jurisdiction</title><p>Four communities were selected in each of the five jurisdictions in order to form two matched pairs per jurisdiction for a total of 20 communities. Later, each community with a matched-pair was randomized to intervention or delayed intervention. Community selection was based on the following eligibility criteria, identified using the 2000 US Census tract data [<xref ref-type="bibr" rid="CR54">54</xref>]: population size of <underline>&#x0003e;</underline>1,000, <underline>&#x0003e;</underline>25&#x000a0;% of the population of indigenous/native descent (15&#x000a0;% in Alaska due to no census tract with population of 1,000 having more than 25&#x000a0;%), and <underline>&#x0003e;</underline>10&#x000a0;% of the population under age 10&#x000a0;years (based on combining census tract data groups of &#x0003c;5&#x000a0;years of age and 5&#x02013;9&#x000a0;years of age, in order to have sufficient population for CHL targeting of 2&#x02013;8&#x000a0;year olds). Additional criteria included adequate settings for sampling children (e.g., schools), that children live and go to school in the same community, minimal risk of contamination between matched-pair communities, reasonable accessibility for the CHL team, community cohesiveness, and sufficient settings for intervention (e.g., community centers, parks, churches, and stores).</p></sec><sec id="Sec7"><title>Recruitment Methods</title><p>A non-probability, convenience sampling scheme was used to recruit participants for the CM and CFM from each of the selected CHL communities (see Fig.&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>). Participants that either resided or worked in the target communities were invited to attend. Each jurisdiction used their Local Advisory Committees (LAC) and key informants to assist with developing lists of potential participants who would be good representatives of one of three constituent groups&#x02014;community leaders, teachers, and parents. Additional potential participants were identified using contact and member listings from consortia and organizations sharing similar objectives (e.g., Non-Communicable Disease Consortium) or serving similar populations (e.g., early childhood education centers) as CHL and CHL staff contacts within the communities.</p><p>The recruitment process in each jurisdiction followed specific cultural protocols. In Hawai&#x02018;i, Guam, and Alaska, CM and CFM recruitment flyers were distributed over professional networks or paper copies were posted at various locations in the community. Letters of invitation were also hand delivered and emailed. In CNMI, the CHL team recruited participants via email and telephone. In American Samoa, the recruitment method used an involved and protracted cultural protocol under the direction of a High Orator Chief (Author AARA) who met one on one with each potential participant based on a chain of contact system.</p></sec><sec id="Sec8"><title>Community Meetings (CM) and Community Feedback Meetings (CFM)</title><p>The goals of the CMs and CFMs were to identify assets and needs around healthy eating and active living among children ages 2&#x02013;8&#x000a0;years and to identify priorities for the intervention in CHL communities. Specifically, facilitated group discussions: (a) identified factors that promote or hinder healthy living, (b) identified community resources that could be leveraged to promote healthy living, and (c) prioritized potential environmental intervention strategies.</p><p>Questions for the CM facilitated group discussions were developed in accordance with the CHL conceptual framework (see Fig.&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>) and the positive deviance approach (e.g., focusing on the strengths) [<xref ref-type="bibr" rid="CR43">43</xref>]. Separate questions were written for each of the three constituent groups&#x02014;parents, teachers, and community leaders&#x02014;to elicit constituent-specific ideas (see Table <xref rid="Tab5" ref-type="table">5</xref> in &#x0201c;<xref rid="Sec14" ref-type="sec">Appendix</xref>&#x0201d;). For example, teachers were asked whether their school allowed sugar sweetened beverages, while community leaders were asked about local policies related to sugar-sweetened beverages. In instances when participants identified with more than one constituent group (e.g., parent and teacher), they were asked to select and participate in the group they most strongly identify with. Questions were pre-tested to ensure cultural appropriateness and clarity. The CFM were held after the CM and the facilitated group discussions focused on the assets, needs and resources identified in the CM.</p><p>The CMs and CFMs were guided by trained facilitators who were instructed to remain neutral to the discussion [<xref ref-type="bibr" rid="CR55">55</xref>]. Ground rules were agreed upon prior to the start of every meeting to ensure a safe and open venue for communication [<xref ref-type="bibr" rid="CR55">55</xref>]. CM and CFM discussions were recorded on flip chart paper. The written record served as the group memory and was used to facilitate CM and CFM discussion [<xref ref-type="bibr" rid="CR55">55</xref>]. Constituent group discussions were also recorded using digital recorders to provide further detail during analysis. Meetings were conducted in English in all jurisdictions except in American Samoa where Samoan was used.</p></sec><sec id="Sec9"><title>Qualitative Analysis&#x02014;Community Meetings (CM)</title><p>The group memory served as the primary tool for qualitative analysis for the CM. Participants were asked to prioritize key points elicited during the CM based on importance and changeability. Prioritization was determined by a facilitated group agreement process in which before moving on to the next discussion item participants were asked to confirm if they could live with and agree to the prioritized items [<xref ref-type="bibr" rid="CR55">55</xref>]. In Guam and in one of the constituent group (parents) in Hawai&#x02018;i, prioritization of the CM group memory was achieved by thematic content analysis through a clustering/coding process [<xref ref-type="bibr" rid="CR56">56</xref>] using transcribed group memory responses to yield a list of the three to five main priorities made in response to each question. The priorities identified in the group memories were then aggregated across the parent, teacher, and leader groups in each community to form community-specific priorities. Because questions for each group were similar but not identical, questions and responses were clustered by topic or content area. Data from the four communities per jurisdiction were then aggregated to identify jurisdiction priorities. Priorities from each jurisdiction were then compared to identify CHL-wide themes.</p></sec><sec id="Sec10"><title>Qualitative Analysis&#x02014;Community Feedback Meetings (CFM)</title><p>For the CFM, analysis was begun at each jurisdiction. For the purposes of intervention development, only the communities randomized to intervention were included in the analysis. The CFM in the delayed intervention communities focused on the CHL delayed intervention proposal and timeline. Jurisdiction- and community-specific priorities for environmental intervention strategies to promote each target behavior were shared. A facilitated discussion regarding the proposed environmental intervention strategies was held [<xref ref-type="bibr" rid="CR55">55</xref>] and then participants ranked in each CHL target health behavior their top two (based on importance and changeability) proposed environmental intervention strategies. Another facilitated discussion followed to gather further feedback on the ranking process. After the meeting, voting results were tabulated for each CHL target health behavior for each state/jurisdiction. Jurisdiction voting results were then combined to identify the most endorsed CHL wide priorities for environmental intervention strategies.</p></sec></sec><sec id="Sec11" sec-type="results"><title>Results</title><p>The CHL CEP was implemented over a 14&#x000a0;month period (April 2011&#x02013;June 2012). In this time period, each CHL jurisdiction met with their LAC at least twice. CM were conducted between November 2011 and February 2012 after the initial LAC meetings and multiple community key informant meetings. CFM were held between May and June 2012. Across the 5 jurisdictions 912 individuals representing a range of stakeholders participated in the CHL CEP (See Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>). Parents and teachers were especially well represented in the CHL CEP. In many instances, participants who attended the CM also attended the CFM.<table-wrap id="Tab1"><label>Table&#x000a0;1</label><caption><p>Community representatives for Local Advisory Committees (LAC), key informants (KI), community meetings (CM), and community feedback meetings (CFM) across all Children&#x02019;s Healthy Living (CHL) Program jurisdictions</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left">Community representatives</th><th align="left">LAC</th><th align="left">KI</th><th align="left">CM</th><th align="left">CFM</th><th align="left">TOTAL</th></tr></thead><tbody><tr><td align="left" colspan="6">Education</td></tr><tr><td align="left">&#x000a0;Head Start*</td><td char="." align="char">4</td><td char="." align="char">22</td><td char="." align="char">78</td><td char="." align="char">6</td><td char="." align="char">110</td></tr><tr><td align="left">&#x000a0;Preschool</td><td char="." align="char">1</td><td char="." align="char">15</td><td char="." align="char">35</td><td char="." align="char">7</td><td char="." align="char">58</td></tr><tr><td align="left">&#x000a0;Department of Education</td><td char="." align="char">4</td><td char="." align="char">14</td><td char="." align="char">18</td><td char="." align="char">5</td><td char="." align="char">41</td></tr><tr><td align="left">&#x000a0;Other^&#x02021; </td><td char="." align="char">13</td><td char="." align="char">35</td><td char="." align="char">41</td><td char="." align="char">26</td><td char="." align="char">115</td></tr><tr><td align="left">Health Services</td><td char="." align="char">11</td><td char="." align="char">32</td><td char="." align="char">23</td><td char="." align="char">21</td><td char="." align="char">87</td></tr><tr><td align="left">Social Services</td><td char="." align="char">0</td><td char="." align="char">10</td><td char="." align="char">7</td><td char="." align="char">3</td><td char="." align="char">20</td></tr><tr><td align="left">Government^&#x000a7; </td><td char="." align="char">15</td><td char="." align="char">27</td><td char="." align="char">32</td><td char="." align="char">32</td><td char="." align="char">106</td></tr><tr><td align="left">Food Supply</td><td char="." align="char">2</td><td char="." align="char">19</td><td char="." align="char">5</td><td char="." align="char">1</td><td char="." align="char">27</td></tr><tr><td align="left">Wellness^&#x02020; </td><td char="." align="char">1</td><td char="." align="char">9</td><td char="." align="char">9</td><td char="." align="char">3</td><td char="." align="char">22</td></tr><tr><td align="left">Other**</td><td char="." align="char">8</td><td char="." align="char">39</td><td char="." align="char">34</td><td char="." align="char">58</td><td char="." align="char">139</td></tr><tr><td align="left">Parents</td><td char="." align="char">3</td><td char="." align="char">28</td><td char="." align="char">127</td><td char="." align="char">29</td><td char="." align="char">187</td></tr><tr><td align="left">Total</td><td char="." align="char">62</td><td char="." align="char">250</td><td char="." align="char">409</td><td char="." align="char">191</td><td char="." align="char">912</td></tr></tbody></table><table-wrap-foot><p>*&#x000a0;US federally funded program that educates preschool-age children and their families</p><p> ^&#x02021;College, childcare centers/daycares, elementary schools, unspecified education type</p><p> ^&#x000a7;Supplemental Program for Women, infants and children (WIC), parks and recreation, chiefs, mayors, cooperative extension service, affairs office, Department of Health</p><p> ^&#x02020;Sports groups, gyms, health advocates</p><p>**&#x000a0;Church, businesses, associations, unspecified community representatives</p></table-wrap-foot></table-wrap> </p><p>Community meeting priorities for environmental intervention strategies that were identified in all CMs held in a CHL state/jurisdiction are presented in Table&#x000a0;<xref rid="Tab2" ref-type="table">2</xref>. The four communities in American Samoa shared the highest number of priorities while the four communities in CNMI shared the least. The priorities for environmental intervention strategies that were most commonly suggested across CHL are identified in Table&#x000a0;<xref rid="Tab3" ref-type="table">3</xref>. Access to healthy, locally-grown food was a priority common across all five jurisdictions. Influencing policies (both school and governmental) to incur healthier behaviors were also identified as important and changeable in four out of five jurisdictions. Limiting screen time was a priority only in American Samoa and Hawai&#x02018;i. All participants received a brief summary of community-specific meeting findings in formal letters. CM participants were in support of CHL working in partnership with their communities to develop the CHL Program.<table-wrap id="Tab2"><label>Table&#x000a0;2</label><caption><p>Children&#x02019;s Healthy Living (CHL) Program priorities for environmental intervention strategies identified in all community meetings held in each corresponding jurisdiction</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" colspan="5">CHL Jurisdiction</th></tr><tr><th align="left">Alaska</th><th align="left">American Samoa</th><th align="left">CNMI</th><th align="left">Guam</th><th align="left">Hawaii</th></tr></thead><tbody><tr><td align="left" colspan="5">Priorities for environmental intervention strategies</td></tr><tr><td align="left">&#x000a0;1. Value system emphasizing self-reliance (e.g., subsistence lifestyle) combined with a sense of community an asset for healthy living</td><td align="left">1. American Samoa Community College Community and Natural Resources should lead the dissemination of healthy eating and physical activity information for the community</td><td align="left">1. Nutrition Assistance Program (NAP) should mimic Supplemental Program for Women, Infants, and Children (WIC) to restrict food purchases of unhealthy food and drinks</td><td align="left">1. Activities should be focus on the broad-spectrum of the community involving adults that influence young children (e.g., parents, teachers, caregivers)</td><td align="left">1. Strategies locally and culturally based (e.g., incorporate concepts like makahiki, ahupua&#x02018;a, ohana, hula)</td></tr><tr><td align="left">&#x000a0;2. Family education on all aspects of healthy living</td><td align="left">2. Family plantations are important to increasing fruit and vegetable intake</td><td align="left"/><td align="left">2. Improve physical activity infrastructure development, maintenance and access</td><td align="left">2. Older siblings/children as healthy role models</td></tr><tr><td align="left" rowspan="2">&#x000a0;3. Increase awareness and access to the diversity of resources for healthy living</td><td align="left">3. Adequate water resources (e.g., water coolers) should be readily available so children can only be given water to drink</td><td align="left"/><td align="left"/><td align="left">3. Give families specific activities to replace screen time</td></tr><tr><td align="left">4. Importance of role models demonstrating healthy living</td><td align="left"/><td align="left"/><td align="left"/></tr></tbody></table><table-wrap-foot><p>CNMI&#x000a0;=&#x000a0;Commonwealth of the Northern Mariana Islands, Makahiki&#x000a0;=&#x000a0;traditional Hawaiian festival, Ahupua&#x02018;a&#x000a0;=&#x000a0;traditional Hawaiian land division usually extending from the uplands to the sea, Ohana&#x000a0;=&#x000a0;family in Hawaiian, Hula&#x000a0;=&#x000a0;traditional Hawaiian dance</p></table-wrap-foot></table-wrap> <table-wrap id="Tab3"><label>Table&#x000a0;3</label><caption><p>The most commonly suggested priorities for environmental intervention strategies identified in all Children&#x02019;s Healthy Living (CHL) program community meetings</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left">Overall CHL priorities for environmental intervention strategies</th><th align="left">Alaska</th><th align="left">American Samoa</th><th align="left">CNMI</th><th align="left">Guam</th><th align="left">Hawaii</th></tr></thead><tbody><tr><td align="left">1. Educate parents, siblings, grandparents, children, communities on healthy living</td><td align="left">X</td><td align="left">X</td><td align="left">X</td><td align="left">X</td><td align="left">X</td></tr><tr><td align="left">2. Better and more free community activities and resources to promote healthy living</td><td align="left">X</td><td align="left"/><td align="left">X</td><td align="left">X</td><td align="left">X</td></tr><tr><td align="left">3. Importance of family, teachers, leaders, other respected figures as role models setting a healthy living example</td><td align="left"/><td align="left">X</td><td align="left">X</td><td align="left">X</td><td align="left">X</td></tr><tr><td align="left">4. Improve drinking water access/facilities</td><td align="left"/><td align="left">X</td><td align="left">X</td><td align="left">X</td><td align="left">X</td></tr><tr><td align="left">5. Community resources maintained and accessible during all times making physical activity easier</td><td align="left">X</td><td align="left"/><td align="left">X</td><td align="left">X</td><td align="left">X</td></tr><tr><td align="left">6. School policies need to be changed to make school lunches healthier, encourage water intake, increase physical activity, and reduce sugar sweetened beverage</td><td align="left"/><td align="left"/><td align="left">X</td><td align="left">X</td><td align="left">X</td></tr><tr><td align="left">7. Limit screen time</td><td align="left"/><td align="left">X</td><td align="left"/><td align="left"/><td align="left">X</td></tr><tr><td align="left">8. Change government policies to promote healthy lifestyle, regulate use of government assistance</td><td align="left"/><td align="left">X</td><td align="left">X</td><td align="left">X</td><td align="left">X</td></tr><tr><td align="left">9. Healthy locally-grown food, easily accessible and affordable</td><td align="left">X</td><td align="left">X</td><td align="left">X</td><td align="left">X</td><td align="left">X</td></tr></tbody></table><table-wrap-foot><p> <italic>CNMI</italic> Commonwealth of the Northern Mariana Islands</p></table-wrap-foot></table-wrap> </p><p>Environmental intervention priorities focused on infrastructure, access, role modeling and education were the most endorsed (e.g., received the most votes) across the CHL region for the CHL target health behaviors (See Table&#x000a0;<xref rid="Tab4" ref-type="table">4</xref>). Feedback from community members during the CFM-facilitated discussion stressed that to ensure CHL program success and sustainability, communities need to take ownership of the CHL initiative (Guam and American Samoa) and that CHL needs to incorporate cultural practices (CNMI, Hawai&#x02018;i and American Samoa) and be a catalyst for enhancing local resources/programs already directed at tackling the childhood obesity issue (Alaska and Guam).<table-wrap id="Tab4"><label>Table&#x000a0;4</label><caption><p>Children&#x02019;s Healthy Living (CHL) program priorities for environmental intervention development endorsed by community members at community feedback meetings held across the CHL region to affect CHL target food and activity behaviors</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" colspan="6">Children&#x02019;s Healthy Living target food and activity behaviors</th></tr><tr><th align="left">Increase sleep</th><th align="left">Increase PA</th><th align="left">Decrease screen time</th><th align="left">Increase F/V</th><th align="left">Increase water</th><th align="left">Decrease SSB</th></tr></thead><tbody><tr><td align="left" colspan="6">Priorities for environmental intervention development</td></tr><tr><td align="left">&#x000a0;1. Healthy lifestyle education</td><td align="left">1. Provide affordable/free community resources and programs</td><td align="left">1. Provide/promote alternative, community, and sports activities</td><td align="left">1. Teach family and children about healthy living to promote F/V intake and role modeling</td><td align="left">1. Allow only water at events (e.g. church, school, birthdays, sports activities)</td><td align="left">1. Teach family and children about beverage options and benefits of water</td></tr><tr><td align="left">&#x000a0;2. Regular sleep times</td><td align="left">2. Organized activities and gear/equipment lending program for children and families</td><td align="left">2. Educate parents</td><td align="left">2. Promote home/community gardening through school and community gardening education</td><td align="left">2. Healthy lifestyle education to teach family and children about healthy beverage options and benefits of water</td><td align="left">2. Promote healthy nutrition</td></tr><tr><td align="left">&#x000a0;3. Physical activity schedules</td><td align="left">3. Animal control</td><td align="left">3. Build better infrastructure for alternative activities</td><td align="left">3. More F/V in school meals</td><td align="left">3. Increase access to clean water in school and public places</td><td align="left">3. Healthy lifestyle education</td></tr><tr><td align="left"/><td align="left">4. Build and maintain indoor/outdoor infrastructure</td><td align="left">4. Parents monitor children&#x02019;s screen time</td><td align="left"/><td align="left"/><td align="left">4. Limit access and consumption of through government assistance programs</td></tr></tbody></table><table-wrap-foot><p> <italic>PA</italic> physical activity, <italic>F</italic>/<italic>V</italic> fruit/vegetable, <italic>SSB</italic> sugar-sweetened beverages</p></table-wrap-foot></table-wrap> </p></sec><sec id="Sec12" sec-type="discussion"><title>Discussion</title><p>The multi-step CEP successfully prioritized environmental intervention strategies for intervention program development in participating communities from the USAPI/HI/AK. These priorities focused on policy development, enhancing access to locally grown fruit and vegetables, engaging identified role models (e.g., parents, grandparents, older siblings), increasing access for safe physical activity venues and to clean water, and the provision of education to young children ages 2&#x02013;8&#x000a0;years and to other influential adults to support healthy eating and physical activity. The priorities span the multiple levels of influence on a child&#x02019;s health, ranging from the individual level (e.g., dietary intake) to the governmental level (e.g., policy) suggesting that following an ecological, community-based approach such as the ANGELO model is a viable approach [<xref ref-type="bibr" rid="CR52">52</xref>]. Since recommendations have been made in the literature [<xref ref-type="bibr" rid="CR57">57</xref>] to encourage the publishing of formative research on program development, and because the USAPI/HI/AK populations are underrepresented in the literature, a major objective of this paper was to provide to the wider scientific audience the CHL CEP used in the underserved and minority populations of the USAPI/HI/AK.</p><p>Interestingly, many of the environmental intervention strategies are similar to other previously successful childhood obesity prevention approaches [<xref ref-type="bibr" rid="CR47">47</xref>, <xref ref-type="bibr" rid="CR58">58</xref>&#x02013;<xref ref-type="bibr" rid="CR65">65</xref>]. Group agreement and participant voting was the primary method used for the prioritization to ensure that identified priorities were community driven [<xref ref-type="bibr" rid="CR55">55</xref>], while aligning with program behavioral objectives. No information on evidence-based priorities was provided to participants prior to the CM or the CFM. The alignment between the evidence-base and the community&#x02019;s perspective suggests that the community is an appropriate resource to determine how CHL can positively affect the environment to promote healthy eating and physical activity, a finding well received by community members at the end of the CEP.</p><p>The ability to identify community priorities for environmental intervention strategies may have been an outcome of bridging the gap between constituencies. Community leaders, parents, and other members of the community invested in child health were invited to the table to share openly and honestly. For example, at the onset of all CM and CFM, CHL staff and participants agreed upon ground rules to create a safe space for discussion [<xref ref-type="bibr" rid="CR55">55</xref>]. Impartial facilitators promoted full participation ensuring that all had an opportunity to share while the use of facilitated small group discussions among like constituencies allowed for a less intimidating environment. We found that the communities were grateful for the opportunity to discuss the threat of childhood obesity in their community and were eager to provide input. Overall, participants assumed community responsibility and recognized the importance of all levels of the community working together to address the problem.</p><p>The CEP allowed CHL to identify priorities for environmental intervention strategies for the USAPI/HI/AK so that the CHL team could build a sustainable intervention operation framework for implementation. However, since no community is alike, especially spanning the Pacific Rim, applying the findings into developing each community&#x02019;s intervention program is the next step. A positive aspect of the CHL CEP was that it allowed for the unique priorities, assets, and resources of each participating community to be identified. Correspondingly, the process assisted in the identification of potential community champions, especially evident by community partners who were involved in multiple stages of the CEP, who are significant players in community-based intervention success [<xref ref-type="bibr" rid="CR52">52</xref>]. The CEP also ensured that a jurisdiction&#x02019;s unique issues could also be identified. For example, in American Samoa and the CNMI, the cultural preference for oral versus written communication of messages was expressed multiple times. Oral interaction was identified as important to ensure that the interpretation of results (e.g., priorities for environmental intervention strategies) is culturally-based. These findings will influence the dissemination strategies for intervention messages in American Samoa and CNMI. None-the-less, there was also motivation to be recognized as a USAPI/HI/AK Regional group, recognizing the potential additional power of a shared regional vision in affecting policy and other change.</p><p>One limitation to our community-based processes related to balancing the intentions of a community-based and participatory process with our requirements and obligations as a federal grantee. The CHL team recognizes that a truly community-based participatory research (CBPR) process [<xref ref-type="bibr" rid="CR66">66</xref>] was not used. Though the intent was to use CBPR, many of the parameters and structure of the program were required to be set prior to receipt of the grant. As a result, community members did not have complete leeway to set project goals, objectives, and outcomes. For example, the six outcomes and target behaviors to prevent childhood obesity were set during the initial grant application process (although they built on prior work with the USAPI and Hawaii communities). Also, meeting the scientific expectation of a level of standardization of methodological approach across the culturally diverse USAPI/HI/AK where access to resources (e.g., high speed internet) is quite varied is challenging. Grant writing, where structural domains are set in place, does not lend itself to truly being CBPR [<xref ref-type="bibr" rid="CR67">67</xref>] as it lacks the flexibility inherent to the CBPR process [<xref ref-type="bibr" rid="CR65">65</xref>]. However, the CHL Program is an outcome of many years of collaborative, community-based work among partners in the USAPI and Hawaii [<xref ref-type="bibr" rid="CR20">20</xref>, <xref ref-type="bibr" rid="CR68">68</xref>&#x02013;<xref ref-type="bibr" rid="CR72">72</xref>]. The pre-planning for CHL included a grant application regional planning meeting (May 2011), which involved a variety of stakeholders and professionals in the field from the USAPI/HI/AK, who all agreed upon the strongest leadership profile to ensure successful competition for the grant. A grant of this scale continues to pose the management challenge of balancing between research structures and being community-based. With many existing partners, bringing new partners into the management structure is challenging, so compromise and adherence to common protocols is needed to yield group results. In most cases, the USAPI/HI/AK&#x02019;s cultural value of the group and the power of contributing to something larger than one&#x02019;s own community facilitates CHL&#x02019;s progress.</p><p>One important outcome of the CHL CEP was the lessons learned. As mentioned earlier, the USAPI is especially resource-limited reinforcing the need to collaborate with not only land-grant institutions and governmental agencies (e.g., Department of Health) but also with community-based organizations and agencies that have a vested interest in health and who are ultimately going to sustain health in the local communities. Establishing community liaisons (e.g., community champions) became essential to ensure appropriate linkages with the agencies and organizations that span the USAPI/HI/AK were formed. Forming these linkages required an intensive investment of time to ensure that appropriate cultural protocols were followed, especially for US and Non-US Affiliated Pacific Islander populations who prefer oral and group processes [<xref ref-type="bibr" rid="CR73">73</xref>]. Though challenging in research protocols, being adaptable and flexible is also essential. For example, CHL staff needed to quickly adapt to situations in community meeting events when not everyone that is invited attends or when attending individuals were not originally invited. Jurisdictions also had to be willing to reschedule activities that conflicted with the holiday season or other cultural events. Another important aspect is that what works in one jurisdiction does not always work in another (e.g., comment cards were not considered appropriate in American Samoa). Rather than developing rigid protocols for implementation, guideline templates are put forth so each jurisdiction can identify and discuss with the coordinating work group how to make adaptations, if needed.</p><p>As demonstrated in the OPIC Study in the South Pacific [<xref ref-type="bibr" rid="CR47">47</xref>], engaging the community in the intervention development process significantly impacts CHL intervention effectiveness [<xref ref-type="bibr" rid="CR37">37</xref>, <xref ref-type="bibr" rid="CR52">52</xref>]. The CEP ensured that the communities not only provided the initial input but also prioritized and verified that community interpretations resonated and were culturally appropriate. For example, the CEP provided specific language, examples, and the culturally contextualized perspective. Repeated engagement allowed for community validation, which is important in collectivist cultures of the USAPI, Hawaii and the Native populations of Alaska, and will be influential during intervention implementation.</p></sec><sec id="Sec13" sec-type="conclusion"><title>Conclusion</title><p>The CHL CEP was a viable community-based process covering a vast region with a variety of cultures. It allowed for flexibility while integrating commonalities. The CHL CEP identified community-based priorities for environmental intervention strategies that would inform CHL intervention program development and implementation in Alaska, American Samoa, CNMI, Guam and Hawai&#x02018;i. These priorities focused on promoting healthy eating and physical activity policy, training and supporting role models, enhancing access to fruits, vegetables, water and safe play and providing education/training. The CHL CEP is being adapted for use in the FSM, RMI, and RP. The approach taken by CHL to develop a community-based environmentally-focused child obesity prevention intervention may also be useful for other regions of the Pacific or in other underserved, minority island populations.</p></sec>

A Multicenter Retrospective Study of Frameless Robotic Radiosurgery for Intracranial Arteriovenous Malformation

<p><bold>Introduction:</bold> CT-guided, frameless radiosurgery is an alternative treatment to traditional catheter-angiography targeted, frame-based methods for intracranial arteriovenous malformations (AVMs). Despite the widespread use of frameless radiosurgery for treating intracranial tumors, its use for treating AVM is not-well described.</p><p><bold>Methods:</bold> Patients who completed a course of single fraction radiosurgery at The University of North Carolina or Georgetown University between 4/1/2005&#x02013;4/1/2011 with single fraction radiosurgery and received at least one follow-up imaging study were included. All patients received pre-treatment planning with CTA&#x02009;&#x000b1;&#x02009;MRA and were treated on the CyberKnife (Accuray) radiosurgery system. Patients were evaluated for changes in clinical symptoms and radiographic changes evaluated with MRI/MRA and catheter-angiography.</p><p><bold>Results:</bold> Twenty-six patients, 15 male and 11 female, were included in the present study at a median age of 41&#x02009;years old. The Spetzler-Martin grades of the AVMs included seven Grade I, 12 Grade II, six Grade III, and one Grade IV with 14 (54%) of the patients having a pre-treatment hemorrhage. Median AVM nidal volume was 1.62&#x02009;cm³ (0.57&#x02013;8.26&#x02009;cm³) and was treated with a median dose of 1900 cGy to the 80% isodose line. At median follow-up of 25&#x02009;months, 15 patients had a complete closure of their AVM, 6 patients had a partial closure, and 5 patients were stable. Time since treatment was a significant predictor of response, with patients experience complete closure having on average 11&#x02009;months more follow-up than patients with partial or no closure (<italic>p</italic>&#x02009;=&#x02009;0.03). One patient experienced a post-treatment hemorrhage at 22&#x02009;months.</p><p><bold>Conclusion:</bold> Frameless radiosurgery can be targeted with non-invasive MRI/MRA and CTA imaging. Despite the difficulty of treating AVM without catheter angiography, early results with frameless, CT-guided radiosurgery suggest that it can achieve similar results to frame-based methods at these time points.</p>

<contrib contrib-type="author"><name><surname>Oermann</surname><given-names>Eric K.</given-names></name><xref ref-type="aff" rid="aff1">¹</xref><xref ref-type="corresp" rid="cor1">*</xref><uri xlink:type="simple" xlink:href="http://frontiersin.org/people/u/40933"/></contrib><contrib contrib-type="author"><name><surname>Murthy</surname><given-names>Nikhil</given-names></name><xref ref-type="aff" rid="aff2">²</xref><uri xlink:type="simple" xlink:href="http://frontiersin.org/people/u/182860"/></contrib><contrib contrib-type="author"><name><surname>Chen</surname><given-names>Viola</given-names></name><xref ref-type="aff" rid="aff3">³</xref></contrib><contrib contrib-type="author"><name><surname>Baimeedi</surname><given-names>Advaith</given-names></name><xref ref-type="aff" rid="aff1">¹</xref><uri xlink:type="simple" xlink:href="http://frontiersin.org/people/u/190248"/></contrib><contrib contrib-type="author"><name><surname>Sasaki-Adams</surname><given-names>Deanna</given-names></name><xref ref-type="aff" rid="aff4">⁴</xref></contrib><contrib contrib-type="author"><name><surname>McGrail</surname><given-names>Kevin</given-names></name><xref ref-type="aff" rid="aff2">²</xref></contrib><contrib contrib-type="author"><name><surname>Collins</surname><given-names>Sean P.</given-names></name><xref ref-type="aff" rid="aff3">³</xref><uri xlink:type="simple" xlink:href="http://frontiersin.org/people/u/25988"/></contrib><contrib contrib-type="author"><name><surname>Ewend</surname><given-names>Matthew G.</given-names></name><xref ref-type="aff" rid="aff4">⁴</xref></contrib><contrib contrib-type="author"><name><surname>Collins</surname><given-names>Brian T.</given-names></name><xref ref-type="aff" rid="aff3">³</xref><xref ref-type="corresp" rid="cor1">*</xref><uri xlink:type="simple" xlink:href="http://frontiersin.org/people/u/26488"/></contrib>

Frontiers in Oncology

<sec sec-type="introduction" id="S1"><title>Introduction</title><p>Intracranial arteriovenous malformations (AVMs) present one of the greatest clinical challenges for neurosurgeons, radiation oncologists, and neurointerventionalists. Classically, the treatment of these lesions involved careful patient selection followed by large, open surgical procedures, or more recently endovascular obliteration, radiosurgery, or a combination of these methods (<xref rid="B1" ref-type="bibr">1</xref>&#x02013;<xref rid="B3" ref-type="bibr">3</xref>). This trend of utilizing increasingly less invasive options, endovascular, and radiosurgical, has lead to the advent of frameless radiosurgical devices that do not require the traditional head frame for stereotaxic guidance (<xref rid="B4" ref-type="bibr">4</xref>, <xref rid="B5" ref-type="bibr">5</xref>). Despite the widespread adoption of these devices for treating both intracranial and extracranial pathologies, to the author&#x02019;s knowledge to date there has been only two reports on the results of frameless radiosurgical devices for the treatment of intracranial AVM (<xref rid="B5" ref-type="bibr">5</xref>, <xref rid="B6" ref-type="bibr">6</xref>). We report the retrospective results of two institutions with using CyberKnife frameless stereotactic radiosurgery (SRS) for the treatment of intracranial AVM.</p></sec><sec sec-type="methods" id="S2"><title>Methods</title><sec id="S2-1"><title>Patient selection and treatment</title><p>We performed a retrospective review of patients with intracranial AVMs treated with CyberKnife SRS from December 1st, 2005 to February 1st, 2011 at Georgetown University Hospital and the University of North Carolina at Chapel Hill. Patients who had undergone single fraction SRS for intracranial AVM with or without endovascular embolization and had received at least one follow-up imaging study were included. All patients were treated by an interdisciplinary team of radiation oncologists and neurosurgeons. High resolution CTA images with or without MRA were obtained from all patients for pre-treatment planning. A planning target volume (PTV) and critical structures were manually delineated by the treating neurosurgeon with the PTV encompassing the contour of the AVM with a 1&#x02009;mm margin (Figure <xref ref-type="fig" rid="F1">1</xref>). All treatment planning was performed on pre-treatment CTA imaging, and when available, using fused MRA/CTA imaging. The treating isodose and prescription dose were determined by the treating radiation oncologist in consultation with the treating neurosurgeon, and took into account the AVM nidus, overall volume, proximity to critical structures, and previous treatment history. Treatment plans were generated using an inverse planning method by the CyberKnife treatment software (Multiplan, Accuray).</p><fig id="F1" position="float"><label>Figure 1</label><caption><p><bold>Figure demonstrating treatment planning for a representative case</bold>. Planned treatment volume (red), 90% isodose (blue), and 50% isodose (yellow) can be seen in three planes on pre-treatment planning CTA.</p></caption><graphic xlink:href="fonc-04-00298-g001"/></fig></sec><sec id="S2-2"><title>Outcomes assessment</title><p>Patients were tracked as part of routine clinical follow-up by the interdisciplinary team. MRA scans with or without catheter-angiography confirmation were obtained at pre-defined annual intervals unless acute changes in neurological status warranted immediate imaging. Neurological symptoms were clinically assessed and recorded by the treating neurosurgeons. Complete closure was defined as total resolution of the AVM nidus and draining veins on imaging, with partial closure being defined as a decrease in size of the nidus with the persistence of large draining veins.</p></sec><sec id="S2-3"><title>Statistical analysis</title><p>All statistical analyses were performed utilizing SPSS Statistics v19 (IBM). Statistical analysis was performed in order to identify pre-treatment and treatment variables that correlated with AVM closure. The Kruskal&#x02013;Wallis test, a non-parametric equivalent to ANOVA, was utilized for comparison of continuous variables grouped by AVM closure outcomes. For analysis of volume and dose, Pearson Chi-square testing was employed. Alpha was set to 0.05 to yield a 95% confidence interval (CI) for all statistical tests. Averages were all reported as the median value and interquartile range (IQR), which is a more robust measure of dispersion than simple range.</p></sec></sec><sec id="S3"><title>Results</title><sec id="S3-4"><title>Patient and treatment characteristics</title><p>Twenty-six patients were identified as having undergone treatment for intracranial AVM and met all criteria for inclusion in the current study (Table <xref ref-type="table" rid="T1">1</xref>). Fifteen (58%) of the patients were male and 11 (42%) were female. The median age at time of treatment with radiosurgery was 41&#x02009;years (IQR, 26&#x02013;55&#x02009;years). The AVMs had a range of Spetzler&#x02013;Martin grades with 7 Grade I, 12 Grade II, 6 Grade III, and 1 Grade IV. Ten (38%) of the patients were either current smokers or had a history of smoking, and seven (23%) of the patients were hypertensive. Fourteen (54%) of the patients had a pre-treatment hemorrhage, and of the hypertensive patients, six out of seven (86%) experienced pre-treatment hemorrhage (<italic>p</italic>&#x02009;=&#x02009;0.027). The median AVM nidus volume was 1.62&#x02009;cm³ (IQR, 0.57&#x02013;8.26&#x02009;cm³). Pre-treatment embolization was performed in 11 patients (42%), with 9 patients being treated with Onyx and the others with <italic>n</italic>-butyl cyanoacrylate (NBCA). The median isodose was 80% (76&#x02013;83%), which was treated with a median prescription dose of 1900 cGy (IQR, 1800-2175 cGy). Seventeen (65%) of the patients had SRS as monotherapy, while nine underwent a combination of SRS and embolization or, in one case, surgical resection.</p><table-wrap id="T1" position="float"><label>Table 1</label><caption><p><bold>Summary of AVM patient characteristics</bold>.</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1">Variable</th><th align="center" rowspan="1" colspan="1">Value</th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">Subjects, <italic>N</italic></td><td align="center" rowspan="1" colspan="1">26</td></tr><tr><td align="left" rowspan="1" colspan="1">Median age, years (IQR)</td><td align="center" rowspan="1" colspan="1">41 (26&#x02013;55)</td></tr><tr><td align="left" rowspan="1" colspan="1">Gender</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Male, <italic>n</italic> (%)</td><td align="center" rowspan="1" colspan="1">15 (58)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Female, <italic>n</italic> (%)</td><td align="center" rowspan="1" colspan="1">11 (42)</td></tr><tr><td align="left" rowspan="1" colspan="1">Smoking status</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Current/prior history, <italic>n</italic> (%)</td><td align="center" rowspan="1" colspan="1">10 (38)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Never, <italic>n</italic> (%)</td><td align="center" rowspan="1" colspan="1">12 (46)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Unknown, <italic>n</italic> (%)</td><td align="center" rowspan="1" colspan="1">4 (15)</td></tr><tr><td align="left" rowspan="1" colspan="1">Pre-treatment neurological symptoms</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Headache, <italic>n</italic> (%)</td><td align="center" rowspan="1" colspan="1">12 (46)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Seizures, controlled/uncontrolled, <italic>n</italic>/<italic>n</italic> (%/%)</td><td align="center" rowspan="1" colspan="1">2/2 (8/8)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Motor deficits, <italic>n</italic> (%)</td><td align="center" rowspan="1" colspan="1">9 (35)</td></tr><tr><td align="left" rowspan="1" colspan="1">Pre-treatment hemorrhage, <italic>n</italic> (%)</td><td align="center" rowspan="1" colspan="1">14 (54)</td></tr><tr><td align="left" rowspan="1" colspan="1">Hypertension (%)</td><td align="center" rowspan="1" colspan="1">7 (27)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Pre-treatment hemorrhage, <italic>n</italic> (%)</td><td align="center" rowspan="1" colspan="1">6 (86)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;No pre-treatment hemorrhage, <italic>n</italic> (%)</td><td align="center" rowspan="1" colspan="1">1 (14)</td></tr><tr><td align="left" rowspan="1" colspan="1">Spetzler martin grade</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;I, <italic>n</italic> (%)</td><td align="center" rowspan="1" colspan="1">7</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;II, <italic>n</italic> (%)</td><td align="center" rowspan="1" colspan="1">12</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;III, <italic>n</italic> (%)</td><td align="center" rowspan="1" colspan="1">6</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;IV, <italic>n</italic> (%)</td><td align="center" rowspan="1" colspan="1">1</td></tr><tr><td align="left" rowspan="1" colspan="1">Median Nidus volume, cm³ (IQR)</td><td align="center" rowspan="1" colspan="1">1.62 (0.57&#x02013;8.26)</td></tr><tr><td align="left" rowspan="1" colspan="1">Intervention</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;SRS only, <italic>n</italic> (%)</td><td align="center" rowspan="1" colspan="1">17 (65)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;SRS&#x02009;+&#x02009;embolization or surgery, <italic>n</italic> (%)</td><td align="center" rowspan="1" colspan="1">9 (35)</td></tr><tr><td align="left" rowspan="1" colspan="1">Isodose, median % (IQR)</td><td align="center" rowspan="1" colspan="1">80 (76&#x02013;83%)</td></tr><tr><td align="left" rowspan="1" colspan="1">Dose, median cGy (IQR)</td><td align="center" rowspan="1" colspan="1">1900 (1800&#x02013;2175)</td></tr></tbody></table></table-wrap></sec><sec id="S3-5"><title>AVM Closure rates</title><p>At median follow-up for the cohort of 25&#x02009;months (IQR, 19-36&#x02009;months), 15 patients had a complete closure of their AVM, 6 patients had a partial closure, and 5 patients were stable (Figure <xref ref-type="fig" rid="F2">2</xref>). Time since treatment was a significant predictor of response, fully closed AVM had, on average, 11&#x02009;months more follow-up time than those with partial or no closure (<italic>p</italic>&#x02009;=&#x02009;0.03) (Table <xref ref-type="table" rid="T2">2</xref>). Nidal volume and dose did not correlate with AVM closure rate (<italic>p</italic>&#x02009;=&#x02009;0.63, 0.12). Spetzler&#x02013;Martin Grade did not correlate with AVM closure as well (<italic>p</italic>&#x02009;=&#x02009;0.26).</p><fig id="F2" position="float"><label>Figure 2</label><caption><p><bold>(A)</bold> Pre-radiosurgery, and <bold>(B)</bold> post-radiosurgery CT angios and angiograms for a representative case with follow-up images taken at 2&#x02009;years post-radiosurgery demonstrating complete nidal obliteration with no residual draining vein.</p></caption><graphic xlink:href="fonc-04-00298-g002"/></fig><table-wrap id="T2" position="float"><label>Table 2</label><caption><p><bold>Summary of AVM treatment characteristics and patient outcomes</bold>.</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1">Variable</th><th align="center" colspan="3" rowspan="1">Endpoint<hr/></th><th align="center" rowspan="1" colspan="1">(<italic>p</italic>-Value)</th></tr><tr><th align="left" rowspan="1" colspan="1"/><th align="left" rowspan="1" colspan="1">Complete closure (<italic>n</italic>&#x02009;=&#x02009;15)</th><th align="center" rowspan="1" colspan="1">Partial closure (<italic>n</italic>&#x02009;=&#x02009;6)</th><th align="center" rowspan="1" colspan="1">Stable (<italic>n</italic>&#x02009;=&#x02009;5)</th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">Median follow-up, months (IQR)</td><td align="center" rowspan="1" colspan="1">31 (24&#x02013;39)</td><td align="center" rowspan="1" colspan="1">17 (11&#x02013;22)</td><td align="center" rowspan="1" colspan="1">25 (18&#x02013;29)</td><td align="center" rowspan="1" colspan="1">0.03</td></tr><tr><td align="left" rowspan="1" colspan="1">Spetzler martin grade</td><td align="center" rowspan="1" colspan="1">&#x02013;</td><td align="center" rowspan="1" colspan="1">&#x02013;</td><td align="center" rowspan="1" colspan="1">&#x02013;</td><td align="center" rowspan="1" colspan="1">0.26</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;I, <italic>n</italic> (%)</td><td align="center" rowspan="1" colspan="1">5 (33)</td><td align="center" rowspan="1" colspan="1">3 (50)</td><td align="center" rowspan="1" colspan="1">0 (0)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;II, <italic>n</italic> (%)</td><td align="center" rowspan="1" colspan="1">8 (53)</td><td align="center" rowspan="1" colspan="1">3 (50)</td><td align="center" rowspan="1" colspan="1">2 (40)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;III, <italic>n</italic> (%)</td><td align="center" rowspan="1" colspan="1">3 (20)</td><td align="center" rowspan="1" colspan="1">0 (0)</td><td align="center" rowspan="1" colspan="1">3 (60)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;IV, <italic>n</italic> (%)</td><td align="center" rowspan="1" colspan="1">1 (6)</td><td align="center" rowspan="1" colspan="1">0 (0)</td><td align="center" rowspan="1" colspan="1">0 (0)</td></tr><tr><td align="left" rowspan="1" colspan="1">Median Nidus volume, cm³ (IQR)</td><td align="center" rowspan="1" colspan="1">1.15 (0.54&#x02013;4.66)</td><td align="center" rowspan="1" colspan="1">3.42 (0.71&#x02013;10.24)</td><td align="center" rowspan="1" colspan="1">4.42 (1.96&#x02013;9.07)</td><td align="center" rowspan="1" colspan="1">0.63</td></tr><tr><td align="left" rowspan="1" colspan="1">Intervention</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.64</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;SRS only, <italic>n</italic> (%)</td><td align="center" rowspan="1" colspan="1">8 (53)</td><td align="center" rowspan="1" colspan="1">5 (83)</td><td align="center" rowspan="1" colspan="1">4 (80)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;SRS&#x02009;+&#x02009;embolization or surgery, <italic>n</italic> (%)</td><td align="center" rowspan="1" colspan="1">7 (47)</td><td align="center" rowspan="1" colspan="1">1 (17)</td><td align="center" rowspan="1" colspan="1">1 (20)</td></tr><tr><td align="left" rowspan="1" colspan="1">Isodose, median % (IQR)</td><td align="center" rowspan="1" colspan="1">80 (76&#x02013;85%)</td><td align="center" rowspan="1" colspan="1">77 (72&#x02013;80%)</td><td align="center" rowspan="1" colspan="1">81 (80&#x02013;81%)</td><td align="center" rowspan="1" colspan="1">0.24</td></tr><tr><td align="left" rowspan="1" colspan="1">Dose, median Gy (IQR)</td><td align="center" rowspan="1" colspan="1">1800 (1750&#x02013;1950)</td><td align="center" rowspan="1" colspan="1">2050 (2000&#x02013;2175)</td><td align="center" rowspan="1" colspan="1">2000 (2000&#x02013;2200)</td><td align="center" rowspan="1" colspan="1">0.12</td></tr></tbody></table></table-wrap></sec><sec id="S3-6"><title>Neurological deficits and toxicity</title><p>One patient experienced a post-treatment hemorrhage at 22&#x02009;months requiring emergent surgical decompression (Table <xref ref-type="table" rid="T3">3</xref>). No other significant post-treatment adverse events were reported. The most common pre-treatment neurological symptom was headaches (46%), which improved in most cases after treatment with only four patients (15%) reporting them at the end of the study. Pre-treatment, controlled, and uncontrolled seizures were symptoms in 16% of the patients. By conclusion of the study, 12% of the patients had controlled seizures on oral medications, and no patients had uncontrolled seizures.</p><table-wrap id="T3" position="float"><label>Table 3</label><caption><p><bold>Summary of post-treatment adverse events and symptoms</bold>.</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1">Variable</th><th align="center" rowspan="1" colspan="1">Value</th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">Post-treatment hemorrhage, <italic>n</italic> (%)</td><td align="center" rowspan="1" colspan="1">1 (4)</td></tr><tr><td align="left" rowspan="1" colspan="1">Post-treatment neurological symptoms</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Headache, <italic>n</italic> (%)</td><td align="center" rowspan="1" colspan="1">4 (15)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Seizures, controlled/uncontrolled, <italic>n</italic>/<italic>n</italic> (%/%)</td><td align="center" rowspan="1" colspan="1">3/0 (12/0)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Motor deficits, <italic>n</italic> (%)</td><td align="center" rowspan="1" colspan="1">3 (12)</td></tr></tbody></table></table-wrap></sec></sec><sec sec-type="discussion" id="S4"><title>Discussion</title><p>Our results show that frameless SRS is a safe and effective technique for the treatment of intracranial AVM. A large recent study by Ding et al. reported an obliteration rate of 30% at 10&#x02009;years, and the present study with an obliteration rate of 58% at 3&#x02009;years compares favorably to these results (<xref rid="B7" ref-type="bibr">7</xref>). The higher rate of closure in the present study is may be due to generally smaller nidal volumes, and generally lower grades, yet roughly equivalent marginal doses (<xref rid="B7" ref-type="bibr">7</xref>). It is worth noting that Spetzler&#x02013;Martin grading incorporates size into its calculation of grade (as well as draining veins and eloquence of cortex), and therefore is unsurprisingly correlated with obliteration rates. The results of this article with regards to obliteration rates and dependent factors are consistent with observations made in similar studies of radiosurgical outcomes for treatment of intracranial AVM with Gamma Knife (<xref rid="B7" ref-type="bibr">7</xref>&#x02013;<xref rid="B10" ref-type="bibr">10</xref>). While other studies have shown a consistent and expected dependence of AVM closure on dose, volume, grade, and follow-up time, the present study only demonstrated a dependence on follow-up time (<xref rid="B8" ref-type="bibr">8</xref>). This lack of dependence upon dose and volume may be attributable to a small sample size and the variance within these factors, and therefore are negative results due to lack of statistical power rather than truly negative results.</p><p>Our minimally invasive approach of obtaining CTA with or without MRA for planning purposes prior to frameless SRS does come with a notable drawback when treating AVM after embolization with Onyx. Onyx, an ethylene vinyl alcohol polymer which is solvated in dimethyl-sulfoxide (DMSO), is radio-opaque and can cause artifact on CTA, which can make it difficult to properly visualize the AVM nidus for treatment planning and follow-up.</p><p>The distribution of post-treatment neurological complications in the present group compared similarly to reported series within the Gamma Knife literature as well, with a significant improvement occurring for major neurological symptoms including seizures and motor function compared to pre-operative symptoms (<xref rid="B9" ref-type="bibr">9</xref>, <xref rid="B11" ref-type="bibr">11</xref>). For pre-treatment headaches, there was 66% rate of total resolution, identical to the results of Steiner et al. (<xref rid="B11" ref-type="bibr">11</xref>).</p><p>Approximately, half of the patients in the present series experienced pre-treatment hemorrhage. Pre-treatment hemorrhage can vary greatly between studies in the literature, with some cohorts consisting almost entirely of patients with hemorrhage, and others entirely without (<xref rid="B7" ref-type="bibr">7</xref>, <xref rid="B11" ref-type="bibr">11</xref>). Recent studies have shown that post-radiosurgery hemorrhage can increase the time until AVM closure, and previous work by Flickinger demonstrated that pre-treatment hemorrhage can have a lasting impact on the resolution of neurologic sequelae, although this last finding has been disputed (<xref rid="B7" ref-type="bibr">7</xref>, <xref rid="B12" ref-type="bibr">12</xref>, <xref rid="B13" ref-type="bibr">13</xref>).</p></sec><sec id="S5"><title>Conclusion</title><p>This small pilot series demonstrates that frameless SRS is a safe and effective measure for treating intracranial AVM in utilizing the traditional single fraction approach. Due to advanced imaging and motion tracking technologies, it can achieve equivalent results to traditional frame-based methods without the need for pins and a stereotaxic frame. With further research, we may be able to maximize the benefits of this novel technology for the treatment of intracranial AVM.</p></sec><sec id="S6"><title>Conflict of Interest Statement</title><p>Brian T. Collins and Sean P. Collins have received honoraria from Accuray Inc. for previous work as consultants. The other co-authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.</p></sec>

Perinatal Distress in Women in Low- and Middle-Income Countries: Allostatic Load as a Framework to Examine the Effect of Perinatal Distress on Preterm Birth and Infant Health

Could not extract abstract

<contrib contrib-type="author" corresp="yes"><name><surname>Premji</surname><given-names>Shahirose</given-names></name><address><phone>+01-403-220-4081</phone><email>premjis@ucalgary.ca</email></address><xref ref-type="aff" rid="Aff1">1</xref><xref ref-type="aff" rid="Aff2">2</xref><xref ref-type="aff" rid="Aff3">3</xref></contrib><aff id="Aff1"><label>1</label>Faculty of Nursing, University of Calgary, 2500 University Drive NW, Calgary, AB T2N 1N4 Canada </aff><aff id="Aff2"><label>2</label>Department of Community Health Sciences, Faculty of Medicine, University of Calgary, TRW Building, 3rd Floor, 3280 Hospital Drive NW, Calgary, AB T2N 4Z6 Canada </aff><aff id="Aff3"><label>3</label>Alberta Children&#x02019;s Hospital Research Institute, Heritage Medical Research Building, 3300 Hospital Drive NW, Calgary, AB T2N 1N4 Canada </aff>

Maternal and Child Health Journal

<sec id="Sec1" sec-type="introduction"><title>Introduction</title><p>Both perinatal distress and preterm birth are world-wide problems that are especially burdensome in low- and middle- income countries (LMIC). Maternal prenatal and postnatal distress (i.e., stress, anxiety, or depression at any time in pregnancy and during the first year following birth of the infant), collectively referred to as &#x0201c;perinatal distress,&#x0201d; may be significantly higher in LMIC than high income countries [<xref ref-type="bibr" rid="CR1">1</xref>]. The prevalence of perinatal mental disorders reported for LMIC is comparable to certain high-risk groups of women living in high-income countries [<xref ref-type="bibr" rid="CR1">1</xref>&#x02013;<xref ref-type="bibr" rid="CR3">3</xref>]. In LMIC the determinants of women&#x02019;s and children&#x02019;s health are complex. Moreover, inequities in determinants of health and the social, cultural, and political contexts of women in LMIC negatively influence women&#x02019;s mental health. Consequently, differential vulnerability may exist not only to risk factors of perinatal distress, but also to predictors of pregnancy outcome [<xref ref-type="bibr" rid="CR4">4</xref>].</p><p>Stress, anxiety, or depression during pregnancy may contribute to preterm birth [<xref ref-type="bibr" rid="CR5">5</xref>, <xref ref-type="bibr" rid="CR6">6</xref>]. Every year, 15 million babies are born prematurely, and 1.1 million will die due to prematurity-related health issues globally [<xref ref-type="bibr" rid="CR7">7</xref>, <xref ref-type="bibr" rid="CR8">8</xref>]. Twelve of the 15 countries which contribute more than 60&#x000a0;% to the global burden of preterm birth are low or low-middle income countries [<xref ref-type="bibr" rid="CR9">9</xref>]. Preterm birth is one of the major contributors to infant mortality and morbidity [<xref ref-type="bibr" rid="CR7">7</xref>, <xref ref-type="bibr" rid="CR10">10</xref>], accounting for 80&#x000a0;% of the world&#x02019;s 1.1 million deaths [<xref ref-type="bibr" rid="CR9">9</xref>]. Africa and South Asia, with the exception of Pakistan, have made some progress in improving neonatal survival; however, death resulting from preterm birth is now the second leading cause of newborn deaths [<xref ref-type="bibr" rid="CR7">7</xref>, <xref ref-type="bibr" rid="CR9">9</xref>]. Up to 50&#x000a0;% of pediatric neurodevelopment problems (e.g., cerebral palsy, lower intelligence quotient) are estimated to be the result of preterm birth [<xref ref-type="bibr" rid="CR11">11</xref>&#x02013;<xref ref-type="bibr" rid="CR14">14</xref>].</p><p>Perinatal distress may also adversely influence infant survival, behavior, and development through poor quality of maternal&#x02013;infant interactions [<xref ref-type="bibr" rid="CR15">15</xref>&#x02013;<xref ref-type="bibr" rid="CR21">21</xref>]. A Taiwanese population-based study, that linked birth and death certificate registry, found the adjusted risk of mortality among preschool children up to age 5&#x000a0;years was 1.47 fold (95&#x000a0;% Confidence Interval, CI 1.16&#x02013;1.87) when mothers experienced depression in the first year following birth [<xref ref-type="bibr" rid="CR22">22</xref>]. Infants born in LMIC are already exposed to poverty, poor health, and poor nutrition, which reduces their developmental potential [<xref ref-type="bibr" rid="CR23">23</xref>]. Beyond these issues, infants of depressed mothers are less likely to be breastfed, have incomplete immunizations, have poorer weight gain, and are more likely to experience illnesses, such as diarrhea, which in turn, may increase the number of hospital admissions and contribute to higher mortality in children under 5&#x000a0;years of age [<xref ref-type="bibr" rid="CR24">24</xref>&#x02013;<xref ref-type="bibr" rid="CR31">31</xref>].</p><p>Clinicians typically rely on self-report questionnaires to assess perinatal distress. While very useful, self-report is prone to bias or error [<xref ref-type="bibr" rid="CR32">32</xref>]. An alternative is to use biomarkers that may offer a more objective and quantifiable indicator of the level of perinatal distress [<xref ref-type="bibr" rid="CR33">33</xref>]. The conceptual framework of allostatic load [<xref ref-type="bibr" rid="CR20">20</xref>] links perinatal distress and its physiological responses to multisystem dysregulation, which promotes a cascade of events ultimately impacting pregnancy outcome (i.e., preterm birth) and infant health (i.e., survival and development) [<xref ref-type="bibr" rid="CR16">16</xref>, <xref ref-type="bibr" rid="CR18">18</xref>&#x02013;<xref ref-type="bibr" rid="CR20">20</xref>, <xref ref-type="bibr" rid="CR34">34</xref>]. In this context, biomarkers that detect physiological compromise may be useful predictors of perinatal distress and its negative consequences. Specifically, perinatal distress may activate aspects of the hypothalamic&#x02013;pituitary&#x02013;adrenal (HPA) axis, sympathetic, immune and cardiovascular systems, and promote behavior changes (e.g., smoking, drinking) in the effort to restore allostasis [<xref ref-type="bibr" rid="CR13">13</xref>, <xref ref-type="bibr" rid="CR20">20</xref>]. Over time, given &#x0201c;wear and tear&#x0201d; on the brain and body, biological responses may be compromised, or fail outright. Allostasis refers to the continual changes in set points (i.e., lower or higher ranges) of physiologic systems to maintain constancy [<xref ref-type="bibr" rid="CR20">20</xref>, <xref ref-type="bibr" rid="CR34">34</xref>] with repeated and ongoing (i.e., chronic) exposure to determinants of perinatal distress over the course of pregnancy [<xref ref-type="bibr" rid="CR16">16</xref>, <xref ref-type="bibr" rid="CR35">35</xref>, <xref ref-type="bibr" rid="CR36">36</xref>]. The resulting dysregulation of interrelated systems may, over time, reach a &#x0201c;tipping-point&#x0201d; [<xref ref-type="bibr" rid="CR16">16</xref>] referred to as allostatic load or overload, that ultimately results in pathophysiological effects. In the case of perinatal distress, effects can include preterm birth [<xref ref-type="bibr" rid="CR20">20</xref>, <xref ref-type="bibr" rid="CR34">34</xref>] and altered maternal and infant behaviors that adversely influence infant survival and development [<xref ref-type="bibr" rid="CR15">15</xref>, <xref ref-type="bibr" rid="CR16">16</xref>, <xref ref-type="bibr" rid="CR18">18</xref>&#x02013;<xref ref-type="bibr" rid="CR21">21</xref>, <xref ref-type="bibr" rid="CR37">37</xref>].</p><p>A critical narrative and interpretive review [<xref ref-type="bibr" rid="CR38">38</xref>] was undertaken to: (a) determine the etiologic contribution of perinatal distress on preterm birth in pregnant women in LMIC; and (b) develop a conceptual framework that would explicate the potential casual links of perinatal distress to preterm birth and infant health (i.e., infant survival, and mother&#x02013;infant interaction). The goal of the review was to inform future research in LMIC by providing a conceptual framework to examine psychosocial and environmental factors as both risk factors and targets of intervention to prevent preterm birth (i.e., improve maternal health outcomes) and improve infant survival and development.</p></sec><sec id="Sec2" sec-type="materials|methods"><title>Methods</title><sec id="Sec3"><title>Search and Selection Strategy</title><p>We searched peer-reviewed electronic databases including MEDLINE (1946&#x02013;January 2013), Embase (1974&#x02013;January 2013), Global Health (1910&#x02013;January 2013), and Cumulative Index to Nursing and Allied Health Literature (CINHAL) (1990&#x02013;January 2013). Grey literature (e.g., unpublished theses, organizational websites), reference lists, and an existing network of experts in the area (including research team members from Pakistan, Kenya, and Tanzania) were also used in identifying relevant publications. A conventional review technique using the search strategy and selection strategy detailed in Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref> proved to be limiting given the dearth of literature in LMIC (see Figs.&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>, <xref rid="Fig2" ref-type="fig">2</xref>). In contrast, a search of the existing literature using all key words, combined terms, and exclusion criteria (i.e., etiology and conceptual framework) without limiting the country of origin generated 6,908 records.<table-wrap id="Tab1"><label>Table&#x000a0;1</label><caption><p>Key words, combined terms, and selection criteria</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left"/><th align="left">Key words</th><th align="left">Combined terms</th><th align="left">Selection criteria</th></tr></thead><tbody><tr><td align="left">Etiologic contribution of perinatal distress on preterm birth</td><td align="left">Stress; anxiety; depression; stress, maternal; stress, psychological; chronic stress; postpartum depression; perinatal distress; or perinatal depression</td><td align="left">Pregnancy; women; pregnant, women; perinatal outcomes; infant health; infant survival; mother&#x02013;infant interaction; maternal health; or preterm birth</td><td align="left">Pregnant or postpartum women; recruitment in low- and middle-income countries; assessed psychosocial health/factors; examined any determinant of health that would impact maternal psychosocial well-being and maternal or infant health; any study design; human; English</td></tr><tr><td align="left">Conceptual framework</td><td align="left">Allostasis or allostatic load</td><td align="left">Pregnancy; preterm birth; or brain</td><td align="left">Pregnant or postpartum women; recruitment in low- and middle-income countries; maternal health; infant health; any study design; human; English</td></tr></tbody></table></table-wrap> <fig id="Fig1"><label>Fig.&#x000a0;1</label><caption><p>Overview of trial flow through the search and selection process: Contribution of perinatal distress on preterm birth</p></caption><graphic xlink:href="10995_2014_1479_Fig1_HTML" id="MO2"/></fig> <fig id="Fig2"><label>Fig.&#x000a0;2</label><caption><p>Overview of trial flow through the search and selection process: Allostasis and allostatic load</p></caption><graphic xlink:href="10995_2014_1479_Fig2_HTML" id="MO3"/></fig> </p><p>Applying a precise review question or narrowing the search by assembling certain levels of evidence is restrictive when the intent of the literature review is also to generate a theory [<xref ref-type="bibr" rid="CR38">38</xref>]. Consequently, we used a critical narrative and interpretive synthesis approach [<xref ref-type="bibr" rid="CR38">38</xref>], based in dialectic process including both inductive and deductive reasoning, to guide our sampling of the extant literature, regardless of study type and location of study, while maintaining a focus on the aims of the review. As a starting point we used our earlier review [<xref ref-type="bibr" rid="CR39">39</xref>] on the relationship between prenatal stress, depression, cortisol and preterm birth, and the literature reviewed here. We then purposefully sampled the existing literature to elaborate on the phenomena of interest and our analysis of the literature. The approach we used to develop the conceptual framework was iterative and the emphasis of the review changed and was informed by our emerging understanding and analysis of the literature (i.e., recursive and reflexive). We continued to sample the literature until there was saturation, that is, similar ideas emerged repeatedly [<xref ref-type="bibr" rid="CR38">38</xref>]. A total of 73 articles identified through this iterative process complemented the eight articles identified in the initial search (see Figs.&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>, <xref rid="Fig2" ref-type="fig">2</xref>).</p></sec><sec id="Sec4"><title>Quality Assessment and Data Extraction</title><p>All types of studies were valued for their contribution, as they provided new ways of understanding our emergent conceptual framework and causal links between perinatal distress and preterm birth. Criteria for assessment included: (1) whether the study design was appropriate given the aim and objectives of the study; (2) appraisal of study reporting (e.g., data collection process described, appropriate method of analysis, enough data shared to support interpretation and conclusions); or (3) judgment about whether the study clarified what is known and what is not known, and informed the interpretation of concepts or the review in general [<xref ref-type="bibr" rid="CR38">38</xref>]. No papers were removed because of poor methodological standards.</p></sec><sec id="Sec5"><title>Consultation Exercise</title><p>Towards the end of the review, a group of stakeholders (researchers, clinicians, academics, and policy decision-makers) from Pakistan, Kenya, Tanzania, and Canada were brought together to add additional insights and refine the conceptual framework. Terminology, such as perinatal distress, was clarified and a common understanding was developed of concepts. Essential elements of the framework were identified and revisions were made to better illustrate relationships between components. Through an iterative and consensus building process with feedback received from peer-reviewers of this manuscript, we present the final conceptual framework&#x000a0;(see Fig. <xref rid="Fig3" ref-type="fig">3</xref>).<fig id="Fig3"><label>Fig.&#x000a0;3</label><caption><p>Perinatal distress and pathways to pregnancy outcome: Allostatic load as a conceptual framework</p></caption><graphic xlink:href="10995_2014_1479_Fig3_HTML" id="MO1"/></fig> </p></sec><sec id="Sec6"><title>Findings</title><p>Perinatal mental health of women living in LMIC, particularly mental health during pregnancy, received little attention until 2002. LMIC were represented in only 8 and 15&#x000a0;% of the pregnant- and post-partum related studies, respectively compared to 90&#x000a0;% of high income countries [<xref ref-type="bibr" rid="CR1">1</xref>, <xref ref-type="bibr" rid="CR40">40</xref>]. A systematic review [<xref ref-type="bibr" rid="CR1">1</xref>] and a report of the World Health Organization-United Nations Population Fund [<xref ref-type="bibr" rid="CR40">40</xref>] concluded that available literature from LMIC (41 studies) suggests a wide range in prevalence rates of perinatal mental disorders as a consequence of place of recruitment (e.g., tertiary hospital, provincial or district health services, and community facilities), and methodology (e.g., time of data collection, and screening instruments). An average prevalence of 15.9&#x000a0;% (95&#x000a0;% CI 15.0&#x02013;16.8&#x000a0;%) during pregnancy and 19.8&#x000a0;% (95&#x000a0;% CI 19.2&#x02013;20.6&#x000a0;%) postpartum were reported [<xref ref-type="bibr" rid="CR1">1</xref>], with depression and anxiety disorders being the most frequent diagnoses in both periods [<xref ref-type="bibr" rid="CR1">1</xref>, <xref ref-type="bibr" rid="CR40">40</xref>]. The conceptual framework to examine risk factors for perinatal distress varied among the 31 studies and therefore data could not be pooled [<xref ref-type="bibr" rid="CR1">1</xref>]. Findings related to risk and protective factors of perinatal distress were mixed and the majority did not address all the domains of the social determinants of health used to synthesize the literature, namely, socioeconomic factors, quality of relationship with intimate partner, family and social relationships, reproductive and general health, history of mental health problems, and infant characteristics [<xref ref-type="bibr" rid="CR1">1</xref>]. Social factors, particularly those beyond the women&#x02019;s control, seem to influence perinatal mental health of women in LMIC [<xref ref-type="bibr" rid="CR1">1</xref>]. Moreover, prevention of preterm birth has received little attention in these countries. Although we were able to identify nine studies focusing on preterm birth in LMIC, none examined the relationship between perinatal distress and preterm birth. Five studies [<xref ref-type="bibr" rid="CR7">7</xref>&#x02013;<xref ref-type="bibr" rid="CR10">10</xref>] reviewed epidemiology including global trends, causes, and interventions thus informed this review. However, four studies were excluded as they focused on the use of antenatal steroids in LMIC (two studies), or long-term neurodevelopmental outcomes of preterm infants born in LMIC (two studies). Thus, in LMIC there are missed opportunities to address mental health needs of women along the perinatal continuum and contribute to scientific knowledge and evidence-informed practices and policies to reduce preterm birth and improve infant health outcomes.</p></sec><sec id="Sec7"><title>Perinatal Distress Predicts Preterm Birth</title><p>The term &#x0201c;perinatal distress&#x0201d; encompasses a spectrum of psychological conditions in response to experiences of episodic and chronic stress associated with adverse socio-economic, cultural, and environmental phenomena [<xref ref-type="bibr" rid="CR41">41</xref>]. The North American literature suggests that <italic>pregnancy</italic>-<italic>related anxiety</italic>, which relates to the women&#x02019;s fears about the infant&#x02019;s health, delivery, her own health and survival during the birthing experience, and the impending responsibility of providing for the child [<xref ref-type="bibr" rid="CR42">42</xref>], is a stronger determinant of preterm birth than general anxiety [<xref ref-type="bibr" rid="CR5">5</xref>, <xref ref-type="bibr" rid="CR6">6</xref>, <xref ref-type="bibr" rid="CR42">42</xref>&#x02013;<xref ref-type="bibr" rid="CR44">44</xref>]. Though many North American and European studies have shown an association between <italic>general anxiety</italic> and preterm birth (e.g., [<xref ref-type="bibr" rid="CR45">45</xref>, <xref ref-type="bibr" rid="CR46">46</xref>]), the findings have been mixed (e.g., [<xref ref-type="bibr" rid="CR47">47</xref>, <xref ref-type="bibr" rid="CR48">48</xref>]). In one study, changes in anxiety level over time rather than the anxiety level at one time point predicted preterm birth [<xref ref-type="bibr" rid="CR49">49</xref>]. North American and European studies examining the relationship between <italic>depression</italic> and preterm birth have also shown inconsistent findings, with a minority of the studies finding a statistically significant association between depression and preterm birth (e.g., [<xref ref-type="bibr" rid="CR44">44</xref>, <xref ref-type="bibr" rid="CR46">46</xref>, <xref ref-type="bibr" rid="CR50">50</xref>]).</p><p>Many distinguishable forms of stress can be grouped into chronic stressors and episodic (i.e., acute) stressors. <italic>Chronic stress</italic> differs from acute stress, in that the threat or demand is long-lived, and often without resolution [<xref ref-type="bibr" rid="CR13">13</xref>]. The chronic stress of homelessness or household strain has been associated with preterm birth [<xref ref-type="bibr" rid="CR41">41</xref>]. A study of 739 low-income African-American pregnant women in the United States found that inadequacy of time and money for non-essentials (e.g., time to look nice, time with friends and family) were mediating factors for preterm birth, whereas multidimensional stress (money worries, family problems, and neighborhood crime) and locus of control were independent predictors of preterm birth [<xref ref-type="bibr" rid="CR51">51</xref>]. Neighborhood-level stressors, such as poverty, crime, and racial composition have also shown an independent impact on preterm birth [<xref ref-type="bibr" rid="CR41">41</xref>]. <italic>Episodic stressors</italic> include catastrophic events, such as natural disasters (e.g., hurricane, earthquake, and drought), and manmade calamities (e.g., political strife, and war), have shown varied impacts on pregnancy outcomes from no detected effect (e.g., [<xref ref-type="bibr" rid="CR52">52</xref>]), to lower [<xref ref-type="bibr" rid="CR53">53</xref>] and higher [<xref ref-type="bibr" rid="CR54">54</xref>] rates of preterm birth. The inconsistent findings may be explained by differences in levels of support, medical care, and changes in behavior following the event [<xref ref-type="bibr" rid="CR41">41</xref>].</p><p>Based on the current literature, a multidimensional approach for examining perinatal distress is evident. None of the studies located examined all of the above dimensions of perinatal distress in relation to preterm birth in the same sample. Whether perinatal distress predicts preterm birth in LMIC remains to be established, as none of the studies considered women in LMIC despite nine of the 11 countries with the highest rate of preterm birth being LMIC [<xref ref-type="bibr" rid="CR9">9</xref>]. In our pilot study [<xref ref-type="bibr" rid="CR55">55</xref>] the odds of preterm birth were 1.44 times higher in the depressed Pakistani women than in the non-depressed Pakistani women. The social, cultural, and environmental context of LMIC provide the potential for an in-depth investigation of the multidimensional nature of perinatal distress, which could not be achieved in high-income countries, as all dimensions of perinatal distress co-exist in one setting. Furthermore, the void of empirical literature stemming from LMIC on perinatal distress makes it imperative to examine the etiologic contribution of perinatal distress on preterm birth in LMIC.</p></sec><sec id="Sec8"><title>Explaining Causal Links of Perinatal Distress to Preterm Birth</title><p>In an attempt to adapt or maintain stability (i.e., allostasis), the body responds to perinatal distress (i.e., stress, anxiety, or depression) by producing multisystem physiologic responses through the production of hormonal and neurotransmitter mediators [<xref ref-type="bibr" rid="CR20">20</xref>, <xref ref-type="bibr" rid="CR56">56</xref>, <xref ref-type="bibr" rid="CR57">57</xref>]. In addition to being protective or adaptive, these mediators can have damaging effects [<xref ref-type="bibr" rid="CR58">58</xref>]. Over time, repeated fluctuations and elevated levels of physiologic activity can lead to inefficiency in allostasis where accumulation and overexposure to these mediators (i.e., allostatic load) may results in organ system failure [<xref ref-type="bibr" rid="CR56">56</xref>, <xref ref-type="bibr" rid="CR58">58</xref>]. According to the conceptual framework of allostatic load, composite measures of biomarkers (i.e., hormonal and neurotransmitter mediators) versus individual biomarkers may be a stronger predictor of negative consequences of perinatal distress [<xref ref-type="bibr" rid="CR56">56</xref>, <xref ref-type="bibr" rid="CR59">59</xref>]. The original set of ten parameters of allostatic load continues to expand [<xref ref-type="bibr" rid="CR60">60</xref>]. Empirically supported allostatic load biomarkers implicated in the pathophysiological process linking perinatal distress to preterm birth include:</p><sec id="Sec9"><title>Cortisol</title><p>The brain coordinates the interconnected set of neuroendocrine and behavioral responses to perinatal distress [<xref ref-type="bibr" rid="CR58">58</xref>, <xref ref-type="bibr" rid="CR61">61</xref>]. Cortisol, regulated via the HPA axis, is a primary hormone reported to be elevated in response to stress induced by physical, cognitive and psychosocial challenges [<xref ref-type="bibr" rid="CR58">58</xref>, <xref ref-type="bibr" rid="CR61">61</xref>]. Cortisol is also proposed to be a primary mediator contributing to allostatic load [<xref ref-type="bibr" rid="CR59">59</xref>, <xref ref-type="bibr" rid="CR61">61</xref>]. Although chronically high levels of cortisol have been the focus in the interplay between stress and allostatic load, low cortisol has also been implicated in adverse health outcomes [<xref ref-type="bibr" rid="CR61">61</xref>]. Consequently, response and recovery promoting optimal functioning of pathophysiologic processes following stress is important when considering allostatis [<xref ref-type="bibr" rid="CR61">61</xref>]. Thus, low values and high values may be predictive of preterm birth. Cortisol, measured in blood, has been reported in the majority of studies to have a positive association with preterm birth [<xref ref-type="bibr" rid="CR62">62</xref>].</p></sec><sec id="Sec10"><title>Corticotropin-Releasing Hormone</title><p>Pathologic levels of cortisol can increase the production of placental corticotropin-releasing hormone (CRH) in a dose response relationship [<xref ref-type="bibr" rid="CR63">63</xref>]. Placental CRH levels beyond a certain threshold can have a paradoxical effect of preparing for labor and initiating contractions [<xref ref-type="bibr" rid="CR63">63</xref>]. In the pregnant state, the diurnal variations in hormones, such as cortisol, are to a certain extent diminished [<xref ref-type="bibr" rid="CR34">34</xref>]. In an attempt to compensate for the dysregulation of cortisol, systemic responses of the metabolic, inflammatory, and cardiovascular systems may also experience dysregulation [<xref ref-type="bibr" rid="CR36">36</xref>, <xref ref-type="bibr" rid="CR64">64</xref>].</p></sec><sec id="Sec11"><title>Triglyceride, Total Cholesterol, Low-Density Lipoprotein, and High-Density Lipoprotein</title><p>Total cholesterol, and high-density lipoprotein (HDL<bold>)</bold>, represent the primary effects in response to dysregulation of cortisol [<xref ref-type="bibr" rid="CR59">59</xref>]. Hypercholesterolemia (a secondary mediator) may result in response to high levels of cortisol which mobilizes lipids from adipose tissues [<xref ref-type="bibr" rid="CR65">65</xref>]. Although high levels of cholesterol decreases uterine contractility [<xref ref-type="bibr" rid="CR65">65</xref>], in combination with the natural lipid profile of pregnancy [<xref ref-type="bibr" rid="CR66">66</xref>], an allostatic load effect may alter the vulnerability of the uterine smooth muscle thereby changing its propensity to remain quiescent during pregnancy. During pregnancy, the lipid profile of women changes (i.e., increase in triglyceride, total cholesterol, and low-density lipoprotein) in response to hormonal changes occurring with increasing gestational age [<xref ref-type="bibr" rid="CR66">66</xref>]. Alternation in lipid metabolism, specifically delayed clearance of triglycerides, has been implicated in pregnancy complications (e.g., hypertension and development of preeclampsia) that may lead to medically indicated preterm birth [<xref ref-type="bibr" rid="CR66">66</xref>], as well as adverse pregnancy and infant outcomes [<xref ref-type="bibr" rid="CR67">67</xref>].</p></sec><sec id="Sec12"><title>White Blood Cell Count, C-Reactive Protein, and Cytokines</title><p>Primary effects, such as changes in inflammatory biomarkers in response to primary neuromediators (i.e., cortisol) of stress, have been implicated in the pathway to preterm birth. A systematic review examining the association between inflammatory cytokines and risk of spontaneous preterm birth in asymptomatic women concluded that the maternal&#x02013;fetal interface, rather than systemic inflammation, plays a major role [<xref ref-type="bibr" rid="CR68">68</xref>]. Pregnancy-related anxiety has been associated with preterm birth [<xref ref-type="bibr" rid="CR6">6</xref>, <xref ref-type="bibr" rid="CR69">69</xref>], but among these two studies, only one found that inflammatory markers mediated this influence [<xref ref-type="bibr" rid="CR69">69</xref>]. Various scales were used to measure pregnancy-related anxiety and samples were drawn from high-income countries with low rates of preterm birth.</p><p>Immunosuppression of cellular and humoral immune activity resulting from dysregulation of neuroendocrine mediators, is either site specific (e.g., bacterial vaginosis) [<xref ref-type="bibr" rid="CR70">70</xref>] or systemic, and may increase risk of infections which may be monitored by examining changes in white blood cells counts. In a meta-analysis, bacterial vaginosis was identified as a strong risk factor for preterm birth, with individual studies repeatedly and consistently demonstrating an association [<xref ref-type="bibr" rid="CR70">70</xref>]. A connection has also been demonstrated between prenatal stress and C-reactive protein (CRP) [<xref ref-type="bibr" rid="CR69">69</xref>]. Increased inflammatory cytokines produced both in response to stress (primary mediators) and in response to the infection stimulates production of CRP and triggers prostaglandin production which is a mediator of labor [<xref ref-type="bibr" rid="CR71">71</xref>]. Typically increased cortisol levels serve as a negative feedback loop to decrease production of cytokines and hormones [<xref ref-type="bibr" rid="CR71">71</xref>]; however, the dysregulation of neuromediators most likely impairs this negative feedback loop.</p></sec><sec id="Sec13"><title>Immunoglobulin G</title><p>Immunoglobulin G, an antibody that crosses the placenta, is critical in protecting the infant from infection in the neonatal period. Lower transplacental ratios of immunoglobulin G have been reported in preterm infants [<xref ref-type="bibr" rid="CR72">72</xref>]. High levels of immunoglobulin G, in response to dysregulation of cortisol, is proposed to saturate binding sites, thereby limiting the placenta&#x02019;s efficiency in transfer of immunoglobulin G [<xref ref-type="bibr" rid="CR73">73</xref>]. Since the infant&#x02019;s humoral response is inefficient, the impaired transfer of immunoglobulin G may further compromise the infant&#x02019;s ability to fight infection in early life [<xref ref-type="bibr" rid="CR73">73</xref>] and increase risk of mortality.</p></sec><sec id="Sec14"><title>Blood Pressure and Heart Rate</title><p>Increased blood pressure and heart rate represent a disease state or disorders resulting from allostatic load, as a consequence of secondary outcomes and primary mediator of stress [<xref ref-type="bibr" rid="CR59">59</xref>]. Cardiovascular reactivity is normally reduced in pregnancy [<xref ref-type="bibr" rid="CR74">74</xref>]. However, increased levels of cortisol may increase maternal cardiovascular reactivity (e.g., increase blood pressure and heart rate&#x02014;secondary mediators) [<xref ref-type="bibr" rid="CR34">34</xref>] by altering maternal, placental or fetal hemodynamics [<xref ref-type="bibr" rid="CR75">75</xref>]. A relationship has been demonstrated between high diastolic blood pressure responses to stress during pregnancy and decreased gestational age at birth [<xref ref-type="bibr" rid="CR75">75</xref>&#x02013;<xref ref-type="bibr" rid="CR78">78</xref>]. A dose&#x02013;response pattern has been observed between the rise in blood pressure and spontaneous preterm birth [<xref ref-type="bibr" rid="CR79">79</xref>].</p><p>There is empirical support (approximately 60 studies) for an association between increased allostatic load and negative health consequences of stress (e.g., cardiovascular disease) [<xref ref-type="bibr" rid="CR80">80</xref>]. Notably, none of the documented studies (e.g., [<xref ref-type="bibr" rid="CR6">6</xref>, <xref ref-type="bibr" rid="CR55">55</xref>, <xref ref-type="bibr" rid="CR81">81</xref>&#x02013;<xref ref-type="bibr" rid="CR85">85</xref>]) examining the relationship between perinatal distress, biomarkers of stress, and preterm birth have made use of allostatic load in their conceptual framework. Moreover, the scales used to measure perinatal distress, biomarkers of stress examined, time periods of measurements and findings have varied between studies (see Table&#x000a0;<xref rid="Tab2" ref-type="table">2</xref>). Individual mediators of stress examined in these studies included cytokines (interleukin-10, interleukin-6 and tumor necrosis factor-alpha), CRP [<xref ref-type="bibr" rid="CR69">69</xref>], cortisol [<xref ref-type="bibr" rid="CR6">6</xref>], and CRH [<xref ref-type="bibr" rid="CR6">6</xref>]. Interrelated physiological (i.e., biochemical) response patterns [<xref ref-type="bibr" rid="CR86">86</xref>, <xref ref-type="bibr" rid="CR87">87</xref>] and composite measures involving several biochemical measures offer a more objective and quantifiable indicator of the level of perinatal distress in pregnant women in LMIC who are in difficult cultures, than self-report psychological measures of perinatal distress [<xref ref-type="bibr" rid="CR56">56</xref>, <xref ref-type="bibr" rid="CR59">59</xref>]. The risk of preterm birth will be higher when there is an inadequate response to prenatal distress (i.e., high perinatal distress and low allostatic load) or prolonged response to a previous stress (i.e., low perinatal distress and high allostatic load) [<xref ref-type="bibr" rid="CR20">20</xref>, <xref ref-type="bibr" rid="CR57">57</xref>]. Identifying high risk pregnant women in LMIC and understanding the pathophysiological process of poor pregnancy and health outcomes will guide the development and evaluation of therapeutic interventions to avert preterm birth.<table-wrap id="Tab2"><label>Table&#x000a0;2</label><caption><p>Summary of studies examining the relationship between prenatal stress, biomarkers of stress, and preterm birth</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="2">Study and country (region)</th><th align="left" rowspan="2">Design</th><th align="left" rowspan="2">Participants (n)</th><th align="left" colspan="3">Measures</th><th align="left" rowspan="2">Results</th></tr><tr><th align="left">Scales</th><th align="left">Specimen</th><th align="left">Time, gestation (weeks)</th></tr></thead><tbody><tr><td align="left"><p>Hobel et al. [<xref ref-type="bibr" rid="CR81">81</xref>]</p><p>USA (Los Angeles)</p></td><td align="left">Prospective case&#x02013;control study^a </td><td align="left"><p>Subsample of 524</p><p>Cases: 18 (spontaneous onset of preterm labor)</p><p>Control: 18 (delivered at term)</p><p>Inclusion/exclusion criteria: not specified</p></td><td align="left">PSS-8; STAIT-10</td><td align="left">Plasma CRH, ACTH, cortisol</td><td align="left"><p>18&#x02013;20</p><p>28&#x02013;30</p><p>35&#x02013;</p></td><td align="left">Higher plasma CRH levels and ACTH levels were reported at all three time periods and elevated cortisol levels at 18&#x02013;20&#x000a0;weeks&#x02019; gestation and 28&#x02013;30&#x000a0;weeks&#x02019; gestation in women who delivered preterm when compared to those who delivered at term. Stress levels did not differ between 18&#x02013;20&#x000a0;weeks&#x02019; gestation and 28&#x02013;30&#x000a0;weeks&#x02019; gestation. Variance in CRH at 28&#x02013;30&#x000a0;weeks&#x02019; gestation was explained by maternal stress level at 18&#x02013;20&#x000a0;weeks&#x02019; gestation and maternal age.</td></tr><tr><td align="left"><p>Erickson et al. [<xref ref-type="bibr" rid="CR82">82</xref>]</p><p>Denmark (Odense)</p></td><td align="left">Prospective case&#x02013;control cohort design^a </td><td align="left"><p>Subsample of 2,927</p><p>Cases: 84 (delivered preterm [idiopathic etiology] without complications)</p><p>Control: 224 (delivered at term and matched, at time of enrollment, to within 10&#x000a0;days of due date of cases)</p><p>Inclusion criteria: age &#x0003e;18&#x000a0;years, ability to understand Danish. Exclusion criteria: insufficient responses to the questionnaires, placental previa (diagnosed after 30 full gestational weeks), history of severe fetal congenital malformations in previous pregnancy, uterine cervix insufficiency treated with cervical circlage</p></td><td align="left"><p>Three questionnaires: (1) just before inclusion (past medical history); (2) 30&#x000a0;weeks&#x02019; gestation (social and demographic information); (3) birth; (urogenital and obstetric problems)</p><p>If delivered preterm, completed second and third questionnaire at same time</p></td><td align="left"><p>Plasma CRH, CRH-binding protein, cortisol</p><p>Venous blood sample taken during labor (delivered preterm), and 37&#x02013;43&#x000a0;weeks&#x02019; gestation (delivered at term)</p></td><td align="left"><p>7&#x02013;23</p><p>27&#x02013;37</p></td><td align="left"><p>7&#x02013;23&#x000a0;weeks: CRH and CRH-binding protein levels were higher in women who delivered preterm when compared to women who delivered at term.</p><p>27&#x02013;37&#x000a0;weeks&#x02019; gestation: CRH and cortisol levels were higher but CRH-binding protein levels were lower in women who delivered preterm when compared to women who delivered at term. Previous preterm delivery and engagement in some risk-taking behaviors were associated with preterm birth</p></td></tr><tr><td align="left"><p>Ruiz et al. [<xref ref-type="bibr" rid="CR83">83</xref>]</p><p>USA (central Texas)</p></td><td align="left">Prospective, longitudinal, observational study</td><td align="left"><p>Cases: 78</p><p>Inclusion: English speaking, &#x0003c;28&#x000a0;weeks&#x02019; gestational age, 18&#x02013;40&#x000a0;years of age, singleton pregnancy. Exclusion criteria: Rh isoimmunization, cervical cerclage, use of tocolytic agents during current pregnancy, diabetes mellitus requiring insulin, thyroid disorders, chronic renal or heart disease, misses more than 1&#x000a0;monthly prenatal check for data collection</p></td><td align="left">PSS-10 (23&#x02013;26, and 31&#x02013;35&#x000a0;weeks&#x02019; gestation)</td><td align="left">Blood cortisol (all time points); vaginal swabs for fetal fibronectin, chlamydia, and bacterial vaginosis screen (23&#x02013;26 and 27&#x02013;30&#x000a0;weeks&#x02019; gestation)</td><td align="left"><p>15&#x02013;19</p><p>20&#x02013;22</p><p>23&#x02013;26</p><p>27&#x02013;30</p><p>31&#x02013;35</p></td><td align="left">Cortisol was a poor predictor of both preterm labor and preterm birth; however an increase in cortisol level was noted in women with genitourinary infection. Change is PSS score, that is decrease in perceived stress during the 2nd trimester, was significantly associated with increase in length of gestation</td></tr><tr><td align="left"><p>Mancuso et al. [<xref ref-type="bibr" rid="CR84">84</xref>]</p><p>USA (Los Angeles)</p></td><td align="left">Case&#x02013;control study nested in a prospective cohort^a </td><td align="left"><p>Subsample of 688</p><p>Cases: 282</p><p>Inclusion criteria: singleton intrauterine pregnancy, gave birth to liveborn infant, received prenatal care in prenatal clinics and private practices in Los Angeles, California.</p><p>Exclusion criteria: age &#x0003c;18&#x000a0;years, stillborn births, multiple gestation births, lack of birth outcome data, and incomplete psychosocial data</p></td><td align="left">PSA</td><td align="left">Plasma CRH</td><td align="left"><p>18&#x02013;20</p><p>28&#x02013;30</p></td><td align="left">Women with high CRH levels and high maternal prenatal anxiety at 28&#x02013;30&#x000a0;weeks gestation delivered earlier. CRH levels were significantly higher at both times points in women delivered preterm than women who delivered at term. Mediation effect of CRH</td></tr><tr><td align="left"><p>Kramer et al. [<xref ref-type="bibr" rid="CR6">6</xref>]</p><p>Canada (Montreal)</p></td><td align="left">Prospective cohort and nested case&#x02013;control design</td><td align="left"><p>Subsample of a larger study</p><p>Cases: 207</p><p>Control: 444</p><p>Inclusion criteria: age &#x02265;18&#x000a0;years, singleton gestation, and able to speak English or French. Exclusion criteria: severe chronic illness with ongoing treatment (note: other than hypertension, asthma, or diabetes), placenta previa, diagnosis of incompetent cervix in previous pregnancy, impending delivery, or fetus with congenital anomaly</p></td><td align="left">DHS (lacked basic or essential needs subscale), MSS (chronic stress), AAS (conjugal violence), 5-item scale (injury, job related stress), MIS (intention of pregnancy), ASSIS (perceived social support), PLES (acute stressors), PSS, Dunkel-Schetter 4-item scale (pregnancy related anxiety), RSES, LOT (optimism and pessimism), CES-D, single item (woman&#x02019;s perception of her risk of birth complications), 8-item scale (commitment to pregnancy)</td><td align="left">Hair cortisol, histo-pathologic examination of vaginal swabs, placenta, and cord</td><td align="left">24&#x02013;26</td><td align="left">Only pregnancy related anxiety was consistently and independently associated with spontaneous preterm birth and a dose&#x02013;response was reported across quartiles. Hair cortisol was positively associated with gestational age but not CRH. Maternal plasma CRH, hair cortisol, placental histopathology (i.e., features of infection/inflammation, infarction, or maternal vasculopathy) were not associated with stress, anxiety, or distress measures</td></tr><tr><td align="left"><p>Pearce et al. [<xref ref-type="bibr" rid="CR85">85</xref>]</p><p>Denmark (Odense)</p></td><td align="left">Case&#x02013;control study nested in a prospective cohort</td><td align="left"><p>Subsample of 2,927</p><p>Cases: 60 [delivering preterm (&#x0003c;37&#x000a0;weeks) without a cause, as determined from clinical findings or laboratory investigations during pregnancy or at delivery]</p><p>Control: 123 (delivering at term)</p><p>Inclusion criteria: age &#x0003e;18&#x000a0;years, ability to understand Danish. Exclusion criteria: insufficient responses to the questionnaires, placental previa (diagnosed after 30 full gestational weeks), history of severe fetal congenital malformations in previous pregnancy, uterine cervix insufficiency treated with cervical circlage</p></td><td align="left">Questionnaire (stressful life events, risk-taking behavior indicated by lack of seat-belt usage)</td><td align="left">Serum measures of cortisol, MIF, CRP, CRH, interleukin-1 &#x000df;, interleukin-6, interleukin-10, tumor necrosis factor-alpha</td><td align="left">&#x0003c;24</td><td align="left">Individual biomarkers: MIF (strongest association), interleukin-10, CRP and tumor necrosis factor-alpha predicted preterm birth at various cutoff levels (e.g., 75th, 85th, and 90th percentile). Logistic regression models: MIF, CRP, risk-taking behavior, and low education consistently predicted preterm birth at various cutoffs; however, the 75th percentile cutoff was the best predictive model. MIF may be a psychobiological mediator</td></tr><tr><td align="left"><p>Shaikh et al. [<xref ref-type="bibr" rid="CR55">55</xref>]</p><p>Pakistan (Kirachi)</p></td><td align="left">Prospective cohort study design</td><td align="left"><p>Cases: 132 (125 with complete data)</p><p>Inclusion criteria: age 18&#x02013;40&#x000a0;years, 28&#x02013;30&#x000a0;weeks&#x02019; gestation. Exclusion criteria: diabetes mellitus, thyroid disorder, chronic renal or heart disease, or uterine and cervical abnormality, or on antidepressants or other psychotropic drugs, and did not deliver in setting where the study was based</p></td><td align="left">A&#x02013;Z Stress Scale, CES-D</td><td align="left">Serum cortisol</td><td align="left">28</td><td align="left">A significant positive relationship was reported between maternal depression and stress. No relationship was noted between cortisol value and stress scale or depression scale. Preterm birth was associated with higher parity, past delivery of a male infant, and higher levels of paternal education</td></tr></tbody></table><table-wrap-foot><p>Adopted from Shaikh et al. [<xref ref-type="bibr" rid="CR39">39</xref>]</p><p> <italic>AAS</italic> Abuse Assessment Screen (adapted), <italic>ACTH</italic> adrenocorticotropic hormone, <italic>ASSIS</italic> Arizona Social Support Interview Schedule, <italic>CES</italic>-<italic>D</italic> Centre for Epidemiology Studies Depression Scale, <italic>CRH</italic> corticotropin-releasing hormone, <italic>CRP</italic> C-reactive protein, <italic>DHS</italic> Daily Hassles Scale, <italic>LOT</italic> Life Orientation Test, <italic>MIF</italic> macrophage migration inhibitory factor, <italic>MIS</italic> Miller Intendedness Scale, <italic>MSS</italic> Marital Strain Scale of Pearlin and Schooler, <italic>PLES</italic> Prenatal Life Events Scale, <italic>PSA</italic> Pregnancy-Specific Anxiety Scale, <italic>PSS</italic>-<italic>8</italic> Perceived Stress Scale 8-item version, <italic>PSS</italic>-<italic>10</italic> Perceived Stress Scale 10-item version, <italic>RSES</italic> Rosenberg Self-Esteem Scale, <italic>STAIT</italic> Spielberger&#x02019;s State Anxiety Inventory 10-item version</p><p> ^aNot labelled</p></table-wrap-foot></table-wrap> </p></sec></sec><sec id="Sec15"><title>Explaining Causal Links of Perinatal Distress to Infant Health</title><p>Allostatic load or overload exerts its influence on biological indices or mediators of the HPA axis and sympathetic&#x02013;adrenal&#x02013;medullary systems involving a complex interplay between the mother and fetus [<xref ref-type="bibr" rid="CR13">13</xref>, <xref ref-type="bibr" rid="CR59">59</xref>]. The dysregulation of cortisol influences the permeability of the placenta to cortisol, thereby altering the placental and fetal environment [<xref ref-type="bibr" rid="CR64">64</xref>] and potentially increasing permeability of other mediators which typically do not cross the placenta (e.g., epinephrine). The health of the fetus and newborn &#x0201c;mirror&#x0201d; the health of the mother whereby the fetus or newborn mimics the biochemical profile of the mother. Allostatic load in the fetal brain may also alter behavioral systems which involve attachment/approach and avoidance behaviors that are integral to survival [<xref ref-type="bibr" rid="CR16">16</xref>, <xref ref-type="bibr" rid="CR18">18</xref>, <xref ref-type="bibr" rid="CR58">58</xref>]. Allostatic load may also alter the function (e.g., affective, cognitive, and social) and structure of the brain, and pathological levels may impact developmental outcomes [<xref ref-type="bibr" rid="CR19">19</xref>]. Perinatal distress may directly (e.g., alter structure and function of brain) or indirectly (i.e., through mother&#x02013;infant interaction) influence infant health and well-being.</p><p>In addition to activating the HPA axis, and sympathetic, immune and cardiovascular systems, psychosocial health during pregnancy has been linked to negative maternal health behavior (e.g., consuming non-nutritive substances like soil, consumption of alcohol, and cigarette smoking) [<xref ref-type="bibr" rid="CR88">88</xref>]. Maternal prenatal distress and postnatal distress may result in the same disorders simply manifested along the perinatal continuum [<xref ref-type="bibr" rid="CR89">89</xref>]. Altered parenting patterns (i.e., lack of responsivity to infants&#x02019; needs [<xref ref-type="bibr" rid="CR90">90</xref>, <xref ref-type="bibr" rid="CR91">91</xref>], inability to coordinate age-appropriate activities [<xref ref-type="bibr" rid="CR92">92</xref>], and harsh parenting style [<xref ref-type="bibr" rid="CR93">93</xref>] ) observed in mothers with PPD may contribute to infant stress, with cumulative stress influencing vulnerability to death, disease, or poor developmental outcomes through the effects of infant allostatic load [<xref ref-type="bibr" rid="CR15">15</xref>&#x02013;<xref ref-type="bibr" rid="CR21">21</xref>]. Although in LMIC there is limited evidence examining the contribution of prenatal distress to infant health outcomes (e.g., [<xref ref-type="bibr" rid="CR94">94</xref>]), there is extensive scientific evidence linking PPD and infant health [<xref ref-type="bibr" rid="CR95">95</xref>] that may be explained by the conceptual framework of allostatic load.</p><p>Our pilot data suggests that the odds of depression are 2.7 times greater (95&#x000a0;% CI 1.16&#x02013;6.17, <italic>p</italic>&#x000a0;=&#x000a0;0.015) in Pakistani mothers of preterm infants than Pakistani mothers of full-term infants [<xref ref-type="bibr" rid="CR96">96</xref>]. Thus, for infants born in LMIC, the interactive effects of biological vulnerability associated with being born premature, social vulnerability inherent in women&#x02019;s responses to their environment during the postpartum period and inequities in determinants of health (i.e., poverty, poor nutrition) places them at triple jeopardy to experience poorer health outcomes. In LMIC, premature infants&#x02019; chance of survival, well-being and lifetime developmental and behavioral success may depend on reducing or managing risk factors associated with perinatal distress. For example, implementing early interventions to reduce the risk of stress, anxiety or depression during pregnancy or improve maternal behavior (i.e., increase responsiveness to infant) in the months following the birth of the infant may be warranted.</p></sec></sec><sec id="Sec16" sec-type="discussion"><title>Discussion</title><p>The conceptual framework of allostatic load relates preterm birth to the social, environmental, and biological antecedent of perinatal distress, thereby enabling researchers to examine the interrelationships between various determinants of health. It provides an integrated model that is essential to examine the nature of risk (i.e., cumulative risk) across many systems at the same time and the temporal effects of the risk(s). The use of the conceptual framework of allostatic load to examine the etiologic contributions of perinatal distress on pregnancy and infant outcomes will necessitate longitudinal study designs with multiple time points (e.g., first trimester, early and late second trimester, and third trimester), and multiple measures of data collection (i.e., all dimensions of perinatal distress).</p><p>Although for our purpose we have focused on the negative pregnancy outcome of preterm birth, the conceptual model can be used to investigate pathways for positive pregnancy outcomes. A positive health focus may facilitate population level interventions directed at promoting mental health during pregnancy or &#x0201c;salutogenesis&#x0201d; within the context of their social, cultural, and political environment [<xref ref-type="bibr" rid="CR97">97</xref>]. In LMIC, focusing on what makes women resilient in the face of toxic stress (i.e., pervasive, uncontrollable stress)&#x02014;that is, improving their sense of coherence or &#x0201c;way of being in the world&#x0201d; [<xref ref-type="bibr" rid="CR98">98</xref>]&#x02014;may reduce the burden of health care service delivery. Furthermore, this type of research will facilitate identification of culture-sensitive strategies [<xref ref-type="bibr" rid="CR98">98</xref>] to promote the mental health of women along the perinatal continuum. However, it will be important to debate and discuss social and cultural norms and policies that undermine, both at an individual level and society level, women&#x02019;s mental health during pregnancy and postpartum and access to mental health services.</p><p>Building research capacity will be essential to addressing the under-representation of pregnancy and post-partum related studies in LMIC. Facilitating partnerships between researchers in high-income countries and LMIC to identify and resolve unique challenges related to ethical conduct of research will be important to generating new knowledge in LMIC. Key among these challenges is the communication and understanding of informed consent [<xref ref-type="bibr" rid="CR99">99</xref>]. Since women in LMIC are underprivileged (e.g., poor, with limited access to health care), they may be particularly vulnerable to coercion. Moreover, in keeping with the World Medical Association Declaration of Helsinki, the research should &#x0201c;be responsive to the health needs and priorities of this population or community&#x0201d; [<xref ref-type="bibr" rid="CR100">100</xref>]. Since mental health care services may be non-existent or limited and predominantly hospital based [<xref ref-type="bibr" rid="CR101">101</xref>] consideration should be given to developing or strengthening local mental health care referral services that will continue to serve the women after completion of the study. Strategies (e.g., referrals) will need to be developed to minimize risk and prevent harm to women participating in the study over the course of their pregnancy and following birth of their baby.</p><p>Aside from these ethical issues, studies involving blood sampling for allostatic load parameters need to critically consider the available laboratory infrastructure. Study procedures including procurement of laboratory samples, storage, transportation and processing may create technical and logistical difficulties. Establishing standard procedures, training and supervision of local researchers to develop research capacity, and assisting with knowledge transfer may mitigate logistical issues and ensure adherence to study protocols [<xref ref-type="bibr" rid="CR102">102</xref>, <xref ref-type="bibr" rid="CR103">103</xref>]. Furthermore, quality assurance measures may need to be established to ensure quality data [<xref ref-type="bibr" rid="CR104">104</xref>]. Recruitment and retention of subjects may present significant challenges [<xref ref-type="bibr" rid="CR105">105</xref>] as infrastructure, including communication to arrange follow-up visits and clinical facilities for care may be lacking [<xref ref-type="bibr" rid="CR102">102</xref>].</p></sec><sec id="Sec17" sec-type="conclusion"><title>Conclusion</title><p>Pregnant women in LMIC have been a neglected population in studies on perinatal distress and pregnancy and infant outcomes. Given inequities in determinants of health and the social, cultural, and political contexts of childbearing women in LMIC, these women may experience differential vulnerability to risk factors for perinatal distress and poor pregnancy outcomes. Prospective studies with multiple biological and psychosocial measures of stress, depression or depressive symptoms and its antecedents (e.g., childhood stress, major life events, etc.), state and trait anxiety, and pregnancy-related anxiety may add new knowledge and enhance our understanding about the etiologic contributions of psychosocial processes to preterm birth. A theoretical framework of allostatic load will enable researchers to concurrently examine social, environmental, and genetic antecedents of stress-related vulnerability and physiological (e.g., immune system, placenta) and behavioral responses that influence not only pregnancy outcomes of women in LMIC but also the life trajectories of health and wellness of the fetuses/infants (i.e., mortality and morbidity over time) [<xref ref-type="bibr" rid="CR64">64</xref>]. Interrelated physiological (i.e., biochemical) response patterns [<xref ref-type="bibr" rid="CR86">86</xref>, <xref ref-type="bibr" rid="CR87">87</xref>] and composite measures involving several biochemical measures offer a more objective and quantifiable indicator of the level of perinatal distress in pregnant women in LMIC. We propose that researcher maintain a positive health focus by identifying protective factors or processes that contribute to resilience in the face of toxic stress. When planning research studies using an integrative approach with both biological and psychosocial measures in LMIC, of critical importance is the adherence to principles of ethical conduct of research, engaging local researchers and other stakeholders to anticipate operational challenges to conducting research, and ensuring that the research is responsive to the needs of women during the perinatal period.</p></sec>

Maintenance of familiarity and social bonding via communal latrine use in a solitary primate (<italic>Lepilemur leucopus</italic>)

Could not extract abstract

<contrib contrib-type="author" corresp="yes"><name><surname>Dr&#x000f6;scher</surname><given-names>Iris</given-names></name><address><phone>++49 551 39 7345</phone><email>iris.droescher@gmail.com</email></address><xref ref-type="aff" rid="Aff1">1</xref></contrib><contrib contrib-type="author"><name><surname>Kappeler</surname><given-names>Peter M.</given-names></name><xref ref-type="aff" rid="Aff1">1</xref><xref ref-type="aff" rid="Aff2">2</xref></contrib><aff id="Aff1"><label>1</label>Behavioral Ecology &#x00026; Sociobiology Unit, German Primate Center, Kellnerweg 4, 37077 G&#x000f6;ttingen, Germany </aff><aff id="Aff2"><label>2</label>Department of Sociobiology/Anthropology, Johann-Friedrich-Blumenbach Institute of Zoology &#x00026; Anthropology, University of G&#x000f6;ttingen, Kellnerweg 6, 37077 G&#x000f6;ttingen, Germany </aff>

Behavioral Ecology and Sociobiology

<sec id="Sec1" sec-type="introduction"><title>Introduction</title><p id="Par2">Chemical signals can transmit a variety of information in vertebrates, such as species identity (Caspers et al. <xref ref-type="bibr" rid="CR23">2009</xref>), sexual identity (Ferkin and Johnston <xref ref-type="bibr" rid="CR45">1995</xref>), reproductive state (Ziegler <xref ref-type="bibr" rid="CR170">2013</xref>), and individual identity (Linklater et al. <xref ref-type="bibr" rid="CR89">2013</xref>). Many chemical signals derive from various excretory products, such as feces, urine, and gland secretions (Eisenberg and Kleiman <xref ref-type="bibr" rid="CR40">1972</xref>), and scent marking is defined as the application of these products to features in the environment (Macdonald <xref ref-type="bibr" rid="CR90">1980</xref>). The repeated use of specific locations for defecation/urination can result in an accumulation of feces and other excretory products at so-called latrine sites, and this behavior can be considered a special form of scent marking in cases where it serves a communicatory function (Wronski et al. <xref ref-type="bibr" rid="CR167">2013</xref>). Latrines have been described for several ungulates (e.g., <italic>Ourebia</italic>: Brashares and Arcese <xref ref-type="bibr" rid="CR18">1999</xref>; <italic>Tragelaphus</italic>: Apio et al. <xref ref-type="bibr" rid="CR2">2006</xref>; <italic>Mazama</italic>: Black-Decima and Santana <xref ref-type="bibr" rid="CR12">2011</xref>; <italic>Gazella</italic>: Wronski et al. <xref ref-type="bibr" rid="CR167">2013</xref>), carnivores (e.g., <italic>Suricata</italic>: Jordan et al. <xref ref-type="bibr" rid="CR76">2007</xref>; <italic>Vulpes</italic>: Darden et al. <xref ref-type="bibr" rid="CR30">2008</xref>; <italic>Meles</italic>: Kilshaw et al. <xref ref-type="bibr" rid="CR80">2009</xref>; <italic>Hyaena</italic>: Hulsman et al. <xref ref-type="bibr" rid="CR72">2010</xref>), primates (e.g., <italic>Lepilemur</italic>: Charles-Dominique and Hladik <xref ref-type="bibr" rid="CR24">1971</xref>; <italic>Cheirogaleus</italic>: Schilling <xref ref-type="bibr" rid="CR141">1980a</xref>; <italic>Hapalemur</italic>: Irwin et al. <xref ref-type="bibr" rid="CR74">2004</xref>), and a few other mammalian taxa (e.g., <italic>Arvicola</italic>: Woodroffe and Lawton <xref ref-type="bibr" rid="CR163">1990</xref>; <italic>Oryctolagus</italic>: Sneddon <xref ref-type="bibr" rid="CR149">1991</xref>). Feces are either deposited alone (e.g., <italic>Bassariscus astutus</italic>: Barja and List <xref ref-type="bibr" rid="CR6">2006</xref>; <italic>Ourebia ourebi</italic>: Brashares and Arcese <xref ref-type="bibr" rid="CR18">1999</xref>) or together with urine and/or secretions of specialized glands at latrine sites (e.g., <italic>Meles meles</italic>: Roper et al. <xref ref-type="bibr" rid="CR133">1986</xref>; <italic>Mazama gouazoubira</italic>: Black-Decima and Santana <xref ref-type="bibr" rid="CR12">2011</xref>). In several species (e.g., <italic>Vulpes velox</italic>: Darden et al. <xref ref-type="bibr" rid="CR30">2008</xref>; <italic>Hyaena</italic> spp.: Gorman and Mills <xref ref-type="bibr" rid="CR54">1984</xref>; <italic>Meles meles</italic>: Stewart et al. <xref ref-type="bibr" rid="CR151">2002</xref>), urination is the most common mark used in this context, and feces per se may not be the most important information component of a latrine (Darden et al. <xref ref-type="bibr" rid="CR30">2008</xref>). Similarly, for arboreal species, one could reasonably expect that any potential communicatory function may be rather related to olfactory signals obtainable from arboreally deposited urine than from terrestrial accumulation of feces, which may rather be a byproduct of localized urine marking.</p><p id="Par3">Among primates, the lemurs of Madagascar (Lemuriformes) represent a radiation whose members rely heavily on chemical signals for their social communication (Mertl <xref ref-type="bibr" rid="CR95">1976</xref>; Schilling <xref ref-type="bibr" rid="CR140">1979</xref>, <xref ref-type="bibr" rid="CR142">1980b</xref>; Perret <xref ref-type="bibr" rid="CR112">1992</xref>; Kappeler <xref ref-type="bibr" rid="CR78">1998</xref>; Heymann <xref ref-type="bibr" rid="CR66">2006b</xref>; Charpentier et al. <xref ref-type="bibr" rid="CR25">2008</xref>, <xref ref-type="bibr" rid="CR26">2010</xref>; Boulet et al. <xref ref-type="bibr" rid="CR14">2009</xref>, <xref ref-type="bibr" rid="CR15">2010</xref>; Crawford et al. <xref ref-type="bibr" rid="CR28">2009</xref>; Morelli et al. <xref ref-type="bibr" rid="CR101">2013</xref>), irrespective of their social organization (Kappeler and van Schaik <xref ref-type="bibr" rid="CR79">2002</xref>). The more than 20 species of sportive lemurs (genus <italic>Lepilemur</italic>) are all medium-sized nocturnal folivores. Like many other nocturnal lemurs, they exhibit urine marking (Schilling <xref ref-type="bibr" rid="CR140">1979</xref>, <xref ref-type="bibr" rid="CR142">1980b</xref>; Epple <xref ref-type="bibr" rid="CR41">1986</xref>). In addition, <italic>Lepilemur</italic> males possess anogenital scent glands, while females have no scent glands (Petter et al. <xref ref-type="bibr" rid="CR115">1977</xref>; Schilling <xref ref-type="bibr" rid="CR140">1979</xref>). Sportive lemurs are strictly arboreal, and patterns of defecation/urination produce terrestrial accumulations of feces (Charles-Dominique and Hladik <xref ref-type="bibr" rid="CR24">1971</xref>; Russel <xref ref-type="bibr" rid="CR138">1977</xref>; Irwin et al. <xref ref-type="bibr" rid="CR74">2004</xref>). Some species live in dispersed pairs, which are characterized by spatial overlap between one adult male and one adult female, but low cohesion between pair partners (Sch&#x000fc;lke and Kappeler <xref ref-type="bibr" rid="CR143">2003</xref>; Zinner et al. <xref ref-type="bibr" rid="CR171">2003</xref>; M&#x000e9;ndez-C&#x000e1;rdenas and Zimmermann <xref ref-type="bibr" rid="CR94">2009</xref>; Hilgartner et al. <xref ref-type="bibr" rid="CR68">2012</xref>; Dr&#x000f6;scher and Kappeler <xref ref-type="bibr" rid="CR36">2013</xref>). Pair partners living in dispersed pairs may never share sleeping sites or allogroom each other, and they may even show signs of active spatial avoidance (Dr&#x000f6;scher and Kappeler <xref ref-type="bibr" rid="CR36">2013</xref>). In addition, sportive lemurs are highly territorial, as indicated by minimal home range overlap between individuals of neighboring social units (Zinner et al. <xref ref-type="bibr" rid="CR171">2003</xref>; Rasoloharijaona et al. <xref ref-type="bibr" rid="CR124">2006</xref>; M&#x000e9;ndez-C&#x000e1;rdenas and Zimmermann <xref ref-type="bibr" rid="CR94">2009</xref>; Dr&#x000f6;scher and Kappeler <xref ref-type="bibr" rid="CR36">2013</xref>). This combination of traits makes sportive lemurs an interesting taxon to study various potential functions of latrines.</p><p id="Par4">Irwin et al. (<xref ref-type="bibr" rid="CR74">2004</xref>) reviewed latrine behavior in primates and discussed several hypotheses for the function of latrine use. In particular, they suggested that latrine use in lemurs is mainly linked to the defense of resources, such as specific food patches, mates or sleeping sites. While male sportive lemurs exhibit mate guarding and defend their territories against neighboring males (Hilgartner et al. <xref ref-type="bibr" rid="CR68">2012</xref>), they do not defend food resources for their pair mates, and competition for food is low within as well as between social units (Dr&#x000f6;scher and Kappeler <xref ref-type="bibr" rid="CR37">2014</xref>). However, systematic tests of this potential function of latrines have not been conducted yet.</p><p id="Par5">While latrines may be merely a by-product of a bimodal defecation rhythm that results in the concentration of defecations being deposited under repeatedly used sleeping sites (Julliot <xref ref-type="bibr" rid="CR77">1996</xref>; Gonz&#x000e1;lez-Zamora et al. <xref ref-type="bibr" rid="CR53">2012</xref>), the use of localized defecation sites can also be explained by several additional, non-exclusive functional hypotheses. Many hypotheses that are commonly formulated for the function of scent marking (e.g., Ralls <xref ref-type="bibr" rid="CR119">1971</xref>; Kappeler <xref ref-type="bibr" rid="CR78">1998</xref>; Brady and Armitage <xref ref-type="bibr" rid="CR16">1999</xref>; Lazaro-Perea et al. <xref ref-type="bibr" rid="CR85">1999</xref>; Rostain et al. <xref ref-type="bibr" rid="CR135">2004</xref>; Heymann <xref ref-type="bibr" rid="CR65">2006a</xref>; Lewis <xref ref-type="bibr" rid="CR88">2006</xref>) are also applicable to the function of latrine use, as latrine behavior is a special form of olfactory communication.</p><p id="Par6">In the following, we present hypotheses that are applicable to the social system of our study species (see below) and provide key references for each one of them. First, latrines may be used to demarcate territories, since many mammals are known to use urine, feces or other scent marks to delineate home range boundaries (Mertl-Milhollen <xref ref-type="bibr" rid="CR96">1979</xref>; Brashares and Arcese <xref ref-type="bibr" rid="CR18">1999</xref>; Stewart et al. <xref ref-type="bibr" rid="CR150">2001</xref>; &#x0201c;territory demarcation hypothesis&#x0201d;). Second, latrines may be used to communicate reproductive state, since male mammals seem to be able to detect chemical cues in female urine and/or feces related to reproductive state (Balestrieri et al. <xref ref-type="bibr" rid="CR5">2011</xref>; Archunan and Rajagopala <xref ref-type="bibr" rid="CR3">2013</xref>; &#x0201c;reproductive signaling hypothesis&#x0201d;). Third, latrines may serve to advertise the willingness to defend important resources such as food (Kruuk <xref ref-type="bibr" rid="CR83">1992</xref>; Miller et al. <xref ref-type="bibr" rid="CR97">2003</xref>; Remonti et al. <xref ref-type="bibr" rid="CR128">2011</xref>) or resting sites (Goszczynski <xref ref-type="bibr" rid="CR58">1990</xref>; Branch <xref ref-type="bibr" rid="CR17">1993</xref>; Brady and Armitage <xref ref-type="bibr" rid="CR16">1999</xref>; &#x0201c;resource defense hypothesis&#x0201d;). Fourth, latrines may function as information exchange centers for individuals that rarely associate or interact directly to facilitate the exchange of olfactory individual-specific information within social units to maintain social bonds (Kingdon <xref ref-type="bibr" rid="CR81">1982</xref>; Greene and Drea <xref ref-type="bibr" rid="CR60">2014</xref>; &#x0201c;social bonding hypothesis&#x0201d;). Finally, latrines may play a role in mate defense by advertising the commitment of resident males to defend resident females (Roper et al. <xref ref-type="bibr" rid="CR133">1986</xref>; Jordan et al. <xref ref-type="bibr" rid="CR76">2007</xref>; &#x0201c;mate defense hypothesis&#x0201d;).</p><p id="Par7">By detailing latrine density and distribution, seasonality and behavioral contexts of latrine use as well as age and sex of users, we aimed to test predictions of the above hypotheses. Specifically, (a) if latrines were used to demarcate territories, we expected that they would be located at territorial boundaries or in zones of home range overlap between neighboring social units rather than in core home range areas. (b) If latrines were used to communicate reproductive state, we predicted that frequency of latrine use would increase during the pronounced annual mating season. (c) If latrines were used to contribute to resource defense, we anticipated that latrines would be located in proximity to regular sleeping trees, that feeding effort would be higher within than outside the latrine area, and/or that animals would mark specific food trees by defecation/urination. (d) If latrines were used as information exchange centers for intra-group communication in a species in which individuals of a given social unit visit latrines independently, we expected all individuals of a social unit to visit the same latrines to facilitate information transfer. In addition, we predicted that latrines would be visited exclusively by individuals of a social unit, but not by individuals of neighboring units. (e) If latrines play a role in mate defense, we expected that the frequency of male latrine use would increase with perceived intruder pressure. In addition, we expected that males would place glandular scent marks preferentially in latrines. Finally, (f) since aggression in <italic>L. leucopus</italic> is directed towards roaming individuals rather than neighbors (Dr&#x000f6;scher and Kappeler <xref ref-type="bibr" rid="CR36">2013</xref>), we expected individuals to react more strongly to experimentally introduced feces of strange individuals than to those of familiar ones (Ydenberg et al. <xref ref-type="bibr" rid="CR168">1988</xref>; M&#x000fc;ller and Manser <xref ref-type="bibr" rid="CR102">2007</xref>).</p></sec><sec id="Sec2" sec-type="materials|methods"><title>Methods</title><sec id="FPar1"><title>Study site and animal capture</title><p id="Par8">We studied a population of white-footed sportive lemurs (<italic>Lepilemur leucopus</italic>) at Berenty (S 25.00&#x000b0;, E 46.30&#x000b0;), an approximately 200&#x000a0;km² private ecotourism reserve in southern Madagascar. We observed animals in a spiny forest fragment of about 5&#x000a0;ha (HAH Reserve Foresti&#x000e8;re parcel 1), which is connected to gallery forest on one side via a transitional forest and a further 40&#x000a0;ha spiny forest fragment on the other side (Norscia and Palagi <xref ref-type="bibr" rid="CR105">2008</xref>). To ensure continuing focal observations of single individuals, we equipped animals with radio-tracking transmitters. We used a blowpipe and 1&#x000a0;ml air pressured narcotic syringe projectiles (Telinject, Germany) to anesthetize animals with 0.4&#x000a0;ml Ketanest (100&#x000a0;mg/ml) in the mornings in their daytime sleeping sites. We fitted the animals with radio-collars (TW-3 button-cell tags, Biotrack, UK) while anesthetized. We kept the animals in an animal transport box until they were fully recovered and released them again at their capture site in the evening. We fitted 16 adult (eight males and eight females) and four subadult individuals (three males and one female) with radio-collars. We differentiated adult individuals from subadults by the degree of tooth wear and body mass. We did not radio-collar animals when radio-collars exceeded 4&#x000a0;% of their body mass. We removed all radio-collars after the end of the study. The research followed standard protocols for animal handling, capture, and radio-tracking and was approved by the Commission Tripartite CAFF of the Ministry for Water and Forests (Madagascar).</p></sec><sec id="FPar2"><title>Behavioral observations</title><p id="Par9">We collected behavioral and locational data between October 2011 and October 2012 for a total of 1530&#x000a0;hours on 20 radio-collared individuals. For the present study, we only considered focal individuals that were adult and belonged to social units in which both pair mates were radio-collared (<italic>N</italic>&#x02009;=&#x02009;14 individuals, observation time in sight&#x02009;=&#x02009;1097&#x000a0;hours). Five out of seven social units consisted of pairs; whereas in the remaining cases, an adult male was associated with two adult females each (social unit 1 and 3). However, these females had exclusive ranges since they were regularly seen within the range of the associated adult male, but never within the range of the other adult female. No behavioral observations could be conducted on these females because they were not equipped with radio-collars. For a detailed description on the identification of the social units within the study population see Dr&#x000f6;scher and Kappeler (<xref ref-type="bibr" rid="CR36">2013</xref>).</p><p id="Par10">We divided the study period into four biologically relevant seasons: birth and offspring care with lactation (early wet season from November to January), offspring care without lactation (late wet season from February to April), mating and early gestation (early dry season from May to July) and late gestation (late dry season from August to October). Each individual was watched for two full nights during each season, once by the first author and once by a Malagasy research assistant, using a TR-4 receiver and a RA-14K antenna (Telonics, USA; <xref ref-type="sec" rid="Sec6">Appendix A</xref>) to locate animals. However, we included data only for 7 observation nights for male m9 since he joined female f2 only after he displaced the previous resident male. Similarly, we include data only for 4 observation nights for male m10 since he only joined female f1B at the beginning of the mating season.</p><p id="Par11">The trees of the spiny forest have small and exposed canopies (Grubb <xref ref-type="bibr" rid="CR61">2003</xref>), permitting nocturnal observation of the subjects clearly and continuously (Hladik and Charles-Dominique <xref ref-type="bibr" rid="CR69">1974</xref>). We started continuous focal animal observations (Altmann <xref ref-type="bibr" rid="CR1">1974</xref>) when an animal left its sleeping site at dusk until it returned to its daytime sleeping site at dawn. Usually, when the first author watched an adult male, the Malagasy research assistant watched the corresponding adult female during the same night simultaneously and vice versa. An overview of the focal animal observations is given in <xref ref-type="sec" rid="Sec6">Appendix A</xref>. We tagged spatial locations of animals during continuous focal observations with biodegradable tape while recording the beginning and end of each behavior (i.e., resting, travelling, grooming, feeding, displaying, social interactions). We determined the exact position of the tagged trees with reference to a 10&#x02009;&#x000d7;&#x02009;10m study grid system. In addition, we recorded all occurrences of defecation, urination, scent marking (i.e., rubbing of the anogenital region on a substrate) and olfactory inspection (i.e., sniffing and licking of substrate) of the focal animals along with their spatial location. We distinguished between single-use and multiple-use defecation sites by investigating the degree of ground coverage by feces (a few scattered droppings that could have been produced by a single defecation event vs. concentrated accumulation of feces indicative of multiple use). In addition, ID recorded the same data every time she could observe an un-collared animal defecating/urinating. Each morning after a full-night follow, we located the sleeping trees of all radio-collared animals.</p></sec><sec id="FPar3"><title>Experimental translocation of feces</title><p id="Par12">To establish whether animals discriminate between feces of their own, neighboring and strange social units, we conducted latrine translocation experiments in June 2013 with males and females of 5 social units. We gathered feces from latrines from known neighboring social units (i.e., &#x0201c;neighbor treatment&#x0201d;) and from latrines we located in a neighboring forest parcel, to ensure that the feces originated from social units that were not familiar to the focal animals (i.e., &#x0201c;stranger treatment&#x0201d;). Similarly, we gathered feces from latrines of the focal social unit (i.e., &#x0201c;control treatment&#x0201d;). For the experiments, we spread the gathered feces on plastic sheets of approximately 1&#x000a0;m² (i.e., &#x0201c;experimental latrine&#x0201d;). We handled the feces using disposable plastic gloves. To ensure that the focal animals would encounter the experimental latrines, we determined through preliminary observations which latrine tree each of the focal animals would visit first after leaving the day-time resting tree. For the experiments, we introduced the feces in proximity to the identified latrine tree before sunset. For each experiment, we used an approximately equal amount of feces. We started to record behavioral responses (i.e., loud calling, displaying, glandular scent marking, and sniffing) from the moment the focal individual entered the experimental latrine tree and continued behavioral observations for 30&#x000a0;min. In addition, we recorded the amount of time the animal spent in the latrine tree. We randomized the order in which we presented the three experimental treatments to the focal individuals. We only conducted one experimental treatment on one social unit during a single night. We removed the plastic sheets with the experimental feces immediately after each experiment.</p></sec><sec id="FPar4"><title>Data analyses</title><p id="Par13">To determine whether animals discriminate between feces of their own, neighboring and strange social units, we used Friedman&#x02019;s ANOVA to test for differences between experimental treatments. We used rates of loud calling, sniffing, displaying, and glandular scent marking as measures of response intensity in males, but only rates of loud calling and sniffing in females. A new bout started when an individual interrupted the behavior for more than 5&#x000a0;s. In addition, we used the amount of time the animals spent in the experimental latrine tree as a response variable in both sexes. We based all calculations on the time the animals were in sight.</p><p id="Par14">To establish the number and to investigate the distribution of latrines within the territories of the 7 social units, we calculated the size of individual annual home ranges with the Animal Movement extension of ArcView and plotted all recorded defecation/urination events. Since kernel densities do not require serial independence of observations, we did not correct for spatial autocorrelation (De Solla et al. <xref ref-type="bibr" rid="CR32">1999</xref>). However, we based our home range estimates on a constant time interval (i.e., 5&#x000a0;min) that is biologically meaningful, since it allows individuals to traverse their home range at maximum speed (Rooney et al. <xref ref-type="bibr" rid="CR132">1998</xref>). We calculated home range size from 95&#x000a0;% fixed kernel home range utilization distributions (Worton <xref ref-type="bibr" rid="CR164">1989</xref>) using ad hoc smoothing (Silverman <xref ref-type="bibr" rid="CR147">1986</xref>). To establish whether defecation/urination occurred anywhere in an animal&#x02019;s home range (i.e., random distribution of events) or were restricted to certain areas (i.e., clumped distribution of events), we used the nearest neighbor analysis as implemented in the Animal Movement extension for ArcView (Hooge and Eichenlaub <xref ref-type="bibr" rid="CR71">1997</xref>). While <italic>R</italic> values of 1 indicate a random distribution, values of &#x0003c;1 and &#x0003e;1 indicate a tendency towards a clumped or a uniform distribution, respectively. Before running the analyses, we applied a small amount of random noise to the spatial location points of observed defecation/urination events to break ties between repeated observations at the same localities using the function &#x0201c;jitter&#x0201d; of the R software (R Core Team <xref ref-type="bibr" rid="CR127">2012</xref>).</p><p id="Par15">After ascertaining the spatial distribution of defecation/urination events via nearest neighbor analysis as being clumped, we established the number of latrines per territory by visual inspection of the spatial features in ArcView. Specifically, we considered a latrine as a cluster of defecation/urination events that were at a distance of up to 6&#x000a0;m of each other. We choose 6&#x000a0;m as a distance criterion because this was the minimum distance at which a cluster of defecation/urination events would not disintegrate in a larger number of smaller, non-continuous latrines in close proximity to each other. When testing the various functional hypotheses of latrine use, we only considered defecation/urination events that were clearly associated with latrine visitations by removing all random defecation/urination events (i.e., single-use defecation sites that were not in proximity to a latrine; <italic>N</italic>&#x02009;=&#x02009;32 or 5&#x000a0;% of all defecation/urination events recorded).</p><p id="Par16">To test the territory demarcation hypothesis, we established the number of defecation/urination events within the core vs. the boundary area as well as in the zones of home range overlap. We delineated core areas using a time maximizing function derived from kernel analyses (Vander Wal and Rodgers <xref ref-type="bibr" rid="CR157">2012</xref>).</p><p id="Par17">To test the resource defense hypothesis with regard to defense of food, we investigated whether animals spent less time feeding within than outside the latrine area. We defined food patches as single feeding trees in which animals were observed feeding. Each food patch that was located within 6&#x000a0;m of a latrine tree was assigned as being part of the general latrine area. We calculated the relative proportion of feeding time within and outside the latrine area for each focal individual. In addition, we calculated the relative proportion of the number of food patches located within and without the latrine area. We calculated an index of feeding effort that allows accounting for the fact that the latrine area is smaller than the remaining home range area, and hence, innately can only contain a smaller number of potential food patches. We divided the proportion of foraging time within the latrine area by the relative proportion of the number of food patches located within the latrine area to calculate an index of feeding effort inside the latrine area. Likewise, we divided the proportion of foraging time outside the latrine area by the relative proportion of the number of food patches located outside the latrine area to calculate an index of feeding effort outside the latrine area. We compared feeding effort within and outside the latrine area using Wilcoxon signed-ranks test for each focal individual.</p><p id="Par18">To test the resource defense hypothesis with regard to defense of sleeping sites, we investigated spatial dependence between defecation/urination sites and regular sleeping sites (i.e., sleeping trees that were used more than once by the focal animals). We conducted the analyses using the R package &#x0201c;spatsat&#x0201d; (Baddeley and Turner <xref ref-type="bibr" rid="CR4">2005</xref>). We defined the union home range of all study individuals as the sampling window. We used the L-cross function to describe the dependence in bivariate point patterns using the independence approach (Dixon <xref ref-type="bibr" rid="CR35">2002</xref>). We used the inhomogeneous L-cross function to adjust for spatially varying intensity. For formal hypothesis testing, we computed simulation envelopes by pointwise Monte Carlo test. We used 99 simulations of CSR (complete spatial randomness) to compute envelopes. The theory of the Monte Carlo test requires the distance (<italic>r</italic>) to be fixed in advance for hypothesis testing (Baddeley and Turner <xref ref-type="bibr" rid="CR4">2005</xref>). We used a value of 6&#x000a0;m as a critical distance. Spatial dependence between points of two types occurs when events of each type are either closer (clustering) or farther away (inhibition) than expected under the assumption that the two processes are independent. Likewise, to test the mate defense hypothesis we investigated spatial dependence between defecation/urination sites and male glandular scent marking sites.</p><p id="Par19">To test the reproductive signaling hypothesis, we used linear mixed models (LMM) to estimate the effect of season on latrine use frequency (model 1). Since season may have a different effect on latrine use frequency in the two sexes, we included season, sex, and their interaction in the model. We included individual identity nested within social unit as a random effect to control for pseudo-replication. In addition, to test the mate defense hypothesis, we used LMM to estimate the effect of intruder pressure on latrine use frequency in males (model 2). We considered observation nights in which focal males engaged in display behavior (i.e., branch bashing displays accompanied by loud calling) and/or placed glandular scent marks as nights with perceived intruder pressure. For each full-night observation, we calculated the frequency of latrine use by dividing the number of latrine visits by the amount of time the focal animal was in sight. We included individual identity as a random effect to control for repeated observations. We controlled for the effect of the number of latrines within an individual&#x02019;s home range as well as for the effect of the type of social organization the individual lived in (i.e., pairs vs. one-male, two-female units). We transformed response variables using the function &#x0201c;boxcox&#x0201d; of the package &#x0201c;MASS&#x0201d; (Venables and Ripley <xref ref-type="bibr" rid="CR158">2002</xref>) and <italic>z</italic>-transformed the covariate (i.e., number of latrines; Schielzeth <xref ref-type="bibr" rid="CR139">2010</xref>).</p><p id="Par20">We checked the distribution of the model residuals, plotted residuals against predicted values, conducted the Lev&#x000e8;ne&#x02019;s test and correlated absolute residuals with fitted values to check model validity. We visually inspected qq-plots and plots of residuals vs. fitted values. None of the diagnostics indicated deviations from the assumptions of normality and homogeneity of residuals (Quinn and Keough <xref ref-type="bibr" rid="CR117">2002</xref>; Field et al. <xref ref-type="bibr" rid="CR47">2012</xref>). We calculated Variance Inflation Factors (VIFs) using the R function &#x0201c;vif&#x0201d; of the package &#x0201c;car&#x0201d; (Fox and Weisberg <xref ref-type="bibr" rid="CR48">2011</xref>) running a standard linear model with the random effect excluded from the predictors. VIFs indicated collinearity not to be an issue (largest VIF for model 1&#x02009;=&#x02009;2.03 and for model 2&#x02009;=&#x02009;1.35, respectively; Field et al. <xref ref-type="bibr" rid="CR47">2012</xref>). For influence diagnostics (Cook&#x02019;s distance, dfbetas), we used the R package &#x0201c;influence.ME&#x0201d; for mixed effect models (Nieuwenhuis et al. <xref ref-type="bibr" rid="CR104">2012</xref>). The largest Cook&#x02019;s distance was only 0.14 for model 1. However, Cook&#x02019;s distances indicated some problems with model stability for model 2 (largest Cook&#x02019;s distance&#x02009;=&#x02009;1.55). Similarly, unstandardized DFBeta values reached 1.15 for model 2, whereas values did not indicate any problems for model 1 (largest DFBeta&#x02009;=&#x02009;0.68; Quinn and Keough <xref ref-type="bibr" rid="CR117">2002</xref>; Field et al. <xref ref-type="bibr" rid="CR47">2012</xref>). Running the second model without the influential case (male 4) did not lead to a different overall result, and hence, we report the results obtained for the complete dataset. To test whether season or intruder pressure, respectively, had an overall effect on latrine use frequency we compared the full model to a model in which only these predictors were removed (i.e., season and its interaction with sex or perceived intruder pressure, respectively), using a likelihood ratio test. We fitted the models in <italic>R</italic> using the function &#x0201c;lmer&#x0201d; in the package &#x0201c;lme4&#x0201d; (Bates et al. <xref ref-type="bibr" rid="CR10">2012</xref>) using Maximum Likelihood rather than Restricted Maximum Likelihood to achieve more reliable <italic>P</italic> values (Bolker et al. <xref ref-type="bibr" rid="CR13">2008</xref>). We derived <italic>P</italic> values for the individual effects based on Satterthwaite approximation for denominator degrees of freedom by using the function &#x0201c;summary&#x0201d; of the R package &#x0201c;lmerTest&#x0201d; (Kuznetsova et al. <xref ref-type="bibr" rid="CR84">2014</xref>). We considered <italic>P</italic>&#x02009;&#x02264;&#x02009;0.05 as statistically significant.</p></sec></sec><sec id="Sec3" sec-type="results"><title>Results</title><sec id="FPar5"><title>General latrine behavior</title><p id="Par21">Animals remained on average 5.8&#x02009;&#x000b1;&#x02009;9.4&#x000a0;min (mean&#x02009;&#x000b1;&#x02009;SD; <italic>N</italic>&#x02009;=&#x02009;678) in trees in which they defecated/urinated. Similarly, they spent in total only 6&#x000a0;% of the total observation time they were in sight in trees in which they defecated/urinated. They lifted their tail to defecate and urinate while clinging to tree trunks. While the feces dropped to the ground, the urine dripped down the main trunk of the tree and left visible stains even once the urine was dried. While <italic>Lepilemur</italic> feces were not very odorous, at least to the human nose, urine was characterized by a distinct species-specific odor. We could observe the focal animals on two occasions to lick and on 26 occasions to sniff the bark of a tree. On 15 of these occasions this behavior occurred in the general latrine area and on six occasions in an identified latrine tree. Outside the observation period, we could observe a male to sniff a wet urine stain that was deposited 8&#x000a0;min earlier by a female in the latrine. In addition, we could observe the animals on four occasions to lower themselves to less than 1&#x000a0;m above the ground in a latrine tree to inspect the ground.</p></sec><sec id="FPar6"><title>Experimental translocation of feces</title><p id="Par22">The time spent in the experimental latrine ranged between 11&#x000a0;% and 80&#x000a0;% (mean&#x02009;&#x000b1;&#x02009;SD&#x000a0;= 29&#x02009;&#x000b1;&#x02009;23) of the observation time in females and between 11 and 39&#x000a0;% (20&#x02009;&#x000b1;&#x02009;7) in males. Rates of loud calling ranged between 0 and 2 bouts/h in females (0.14&#x02009;&#x000b1;&#x02009;0.55) and males (0.27&#x02009;&#x000b1;&#x02009;70). While we could not observe females to engage in sniffing, rates of sniffing ranged between 0 and 8 bouts/h in males (1.21&#x02009;&#x000b1;&#x02009;2.49). We could not observe males to engage in display behavior during the experiment, but rates of scent marking ranged between 0 and 2 bouts/h (0.54&#x02009;&#x000b1;&#x02009;0.92). Response intensity did not differ significantly among the three experimental treatments in either males or females. More precisely, the amount of time spent in the latrine tree (females, <italic>&#x003c7;</italic> ²&#x02009;=&#x02009;1.3, <italic>df</italic>&#x02009;=&#x02009;2, <italic>P</italic>&#x02009;=&#x02009;0.522; males, <italic>&#x003c7;</italic> ²&#x02009;=&#x02009;5.7, <italic>df</italic>&#x02009;=&#x02009;2, <italic>P</italic>&#x02009;=&#x02009;0.058), rates of loud calling (females, <italic>&#x003c7;</italic> ²&#x02009;=&#x02009;0.3, <italic>df</italic>&#x02009;=&#x02009;2, <italic>P</italic>&#x02009;=&#x02009;0.861; males, <italic>&#x003c7;</italic> ²&#x02009;=&#x02009;0.3, <italic>df</italic>&#x02009;=&#x02009;2, <italic>P</italic>&#x02009;=&#x02009;0.861), sniffing (females, <italic>&#x003c7;</italic> ²&#x02009;=&#x02009;0.0, <italic>df</italic>&#x02009;=&#x02009;2, <italic>P</italic>&#x02009;=&#x02009;1; males, <italic>&#x003c7;</italic> ²&#x02009;=&#x02009;1.2, <italic>df</italic>&#x02009;=&#x02009;2, <italic>P</italic>&#x02009;=&#x02009;0.549), displaying (males, <italic>&#x003c7;</italic> ²&#x02009;=&#x02009;0.0, <italic>df</italic>&#x02009;=&#x02009;2, <italic>P</italic>&#x02009;=&#x02009;1), and scent marking (males, <italic>&#x003c7;</italic> ²&#x02009;=&#x02009;1.2, <italic>df</italic>&#x02009;=&#x02009;2, <italic>P</italic>&#x02009;=&#x02009;0.549) did not differ significantly among treatments.</p></sec><sec id="FPar7"><title>Spatial distribution of defecation/urination events</title><p id="Par23">Union home range size (95&#x000a0;% Kernel estimates) for the seven social units ranged between 0.28 and 0.47&#x000a0;ha (mean&#x02009;&#x000b1;&#x02009;SD&#x000a0;= 0.38&#x02009;&#x000b1;&#x02009;0.07&#x000a0;ha, <italic>N</italic>&#x02009;=&#x02009;7). Nearest neighbor analyses of the locations of defecation/urination events computed <italic>R</italic> values ranging between 0.15 and 0.48 for the union home ranges. Within all seven union home ranges the spatial distribution of the defecation/urination events differed significantly from a random spatial distribution (<italic>P</italic>&#x02009;&#x0003c;&#x02009;0.001, <italic>N</italic>&#x02009;=&#x02009;7), with a tendency towards clumping as opposed to towards an even distribution (Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>). We identified 3 to 4 latrines in each union home range (Fig.&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>).<table-wrap id="Tab1"><label>Table 1</label><caption><p>Spatial distribution of observed defecation/urination events within the union home ranges of seven social units of <italic>Lepilemur leucopus</italic> based on nearest neighbor analysis</p></caption><table frame="hsides" rules="groups"><thead><tr><th>Social unit</th><th># defecation events</th><th> <italic>Z</italic> value</th><th> <italic>R</italic> value</th><th> <italic>P</italic> value</th></tr></thead><tbody><tr><td>1</td><td>100</td><td>&#x02212;13.59</td><td>0.22</td><td>&#x0003c;0.001</td></tr><tr><td>2</td><td>135</td><td>&#x02212;16.73</td><td>0.17</td><td>&#x0003c;0.001</td></tr><tr><td>3</td><td>112</td><td>&#x02212;12.75</td><td>0.23</td><td>&#x0003c;0.001</td></tr><tr><td>4</td><td>72</td><td>&#x02212;8.09</td><td>0.48</td><td>&#x0003c;0.001</td></tr><tr><td>5</td><td>86</td><td>&#x02212;11.34</td><td>0.35</td><td>&#x0003c;0.001</td></tr><tr><td>6</td><td>115</td><td>&#x02212;17.20</td><td>0.15</td><td>&#x0003c;0.001</td></tr><tr><td>7</td><td>90</td><td>&#x02212;13.91</td><td>0.22</td><td>&#x0003c;0.001</td></tr></tbody></table></table-wrap> <fig id="Fig1"><label>Fig. 1</label><caption><p>Ninety-five percent kernel annual home ranges for individual adult males (<italic>m</italic>) and females (<italic>f</italic>) of <italic>Lepilemur leucopus</italic> at Berenty between October 2011 and October 2012 as well as the spatial arrangement of the latrines within the home ranges. <italic>Dots</italic> represent individual latrines trees, whereas the <italic>shaded areas</italic> represent a contagious buffer of 3&#x000a0;m around individual latrine trees to distinguish discrete latrines. Home ranges of pair partners overlap (Sex, <italic>m</italic>&#x02009;=&#x02009;male, <italic>f</italic>&#x02009;=&#x02009;female)</p></caption><graphic xlink:href="265_2014_1810_Fig1_HTML" id="MO1"/></fig> </p></sec><sec id="FPar8"><title>Territory demarcation hypothesis</title><p id="Par24">We recorded a total of 678 defecation/urination events. Using the time maximization function, core areas of individual ranges were delineated by 65&#x000a0;% isopleths. Union core areas (65&#x000a0;% Kernel estimates) represented 26&#x02009;&#x000b1;&#x02009;6&#x000a0;% (range&#x000a0;= 20&#x02013;37&#x000a0;%, <italic>N</italic>&#x02009;=&#x02009;7) of the union home ranges (95&#x000a0;% Kernel estimates) of the social units. However, the majority of defecation/urination events (mean&#x02009;&#x000b1;&#x02009;SD&#x000a0;= 82&#x02009;&#x000b1;&#x02009;7&#x000a0;%; range&#x000a0;= 72&#x02013;94&#x000a0;%, <italic>N</italic>&#x02009;=&#x02009;7) were located within the small union core areas of the social units, so that the density of defecation/urination events was significantly higher in the core area (mean&#x02009;&#x000b1;&#x02009;SD&#x000a0;= 875&#x02009;&#x000b1;&#x02009;391 events/ha) compared to the remaining home range area (72&#x02009;&#x000b1;&#x02009;54 events/ha; Wilcoxon signed-rank test, <italic>V</italic>&#x02009;=&#x02009;28, <italic>P</italic>&#x02009;=&#x02009;0.016, <italic>N</italic>&#x02009;=&#x02009;7). The overlap zones comprised only 1.35&#x000a0;% of the union of all individual home ranges. None of the defecation/urination events were located within overlap zones of neighboring territories.</p></sec><sec id="FPar9"><title>Resource defense hypothesis</title><p id="Par25">The relative proportion of foraging time within the latrine area ranged between 22&#x000a0;% and 43&#x000a0;% (mean&#x02009;&#x000b1;&#x02009;SD&#x000a0;= 31&#x02009;&#x000b1;&#x02009;7&#x000a0;%, <italic>N</italic>&#x02009;=&#x02009;14). The relative proportion of the number of patches located within the latrine area ranged between 23&#x000a0;% and 46&#x000a0;% (34&#x02009;&#x000b1;&#x02009;7&#x000a0;%). The index of feeding effort inside the latrine area ranged between 0.7 and 1.1 (0.9&#x02009;&#x000b1;&#x02009;0.1) and between 0.8 and 1.1 (0.9&#x02009;&#x000b1;&#x02009;0.1) for the feeding effort outside the latrine area. Feeding effort within the latrine area did not differ significantly from the feeding effort outside the latrine area (Wilcoxon signed-ranks test, <italic>V</italic>&#x02009;=&#x02009;56, <italic>N</italic>&#x02009;=&#x02009;14, <italic>P</italic>&#x02009;=&#x02009;0.851). The animals spent only between 2&#x000a0;% and 14&#x000a0;% (mean&#x02009;&#x000b1;&#x02009;SD&#x000a0;= 7&#x02009;&#x000b1;&#x02009;4&#x000a0;%, <italic>N</italic>&#x02009;=&#x02009;14) of the total feeding time eating in identified latrine trees. While we could record a total number of 1,584 food patches throughout the study, animals were only seen to defecate/urinate in 79 of them. In addition, animals were observed to forage in only 41&#x000a0;%&#x02009;&#x000b1;&#x02009;11&#x000a0;% (range&#x000a0;= 24&#x000a0;% - 55&#x000a0;%, <italic>N</italic>&#x02009;=&#x02009;14) of the identified latrines trees.</p><p id="Par26">The number of repeatedly used sleeping trees ranged between 5 and 10 (mean&#x02009;&#x000b1;&#x02009;SD&#x000a0;= 7&#x02009;&#x000b1;&#x02009;2) for the 7 social units. None of the latrine trees served as a sleeping tree. The computed empirical homogenous L-cross function fell within the simulation envelop at the critical distance of 6&#x000a0;m, indicating spatial independence between defecation/urination and sleeping sites (Fig.&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>).<fig id="Fig2"><label>Fig. 2</label><caption><p>Estimated inhomogeneous L-cross function and envelopes for the bivariate point pattern consisting of defecation/urination sites and sleeping trees. The <italic>solid line</italic> indicates the empirical L-cross function, the <italic>dotted line</italic> indicates the theoretical value for complete spatial randomness (CSR), and the <italic>gray band</italic> indicates the envelope from 99 simulations and <italic>r</italic> is the distance argument</p></caption><graphic xlink:href="265_2014_1810_Fig2_HTML" id="MO2"/></fig> </p></sec><sec id="FPar10"><title>Social bonding hypothesis</title><p id="Par27">Regarding the social units consisting of one adult male and two adult females (unit 1 and 3), all latrines located within the common range of the focal male and focal female were shared by both adult individuals. All latrines within the home ranges of social units consisting of one male and one female were shared by both pair partners, with the exception of social unit 2 where only 2 of 3 latrines were shared. We only once saw a focal individual (m6) to visit a neighbor&#x02019;s latrine (unit 7). In addition, we recorded 47 defecation/urination events by un-collared individuals. 46 of these defecation/urination events were associated with an identified latrine. In 41 of these cases, it was the offspring, which ranged within the parental territory. In 6 cases, it was the second adult un-collared female of unit 1 and 3, respectively. In total, we could observe co-use by un-collared individuals in 18 out of 25 identified latrines.</p></sec><sec id="FPar11"><title>Reproductive signaling hypothesis</title><p id="Par28">Latrine use frequency (number of latrine visitations/h) equaled 0.58&#x02009;&#x000b1;&#x02009;0.25 (mean&#x02009;&#x000b1;&#x02009;SD; <italic>N</italic>&#x02009;=&#x02009;25) during the early wet, 0.48&#x02009;&#x000b1;&#x02009;0.21 (<italic>N</italic>&#x02009;=&#x02009;26) during the late wet, 0.48&#x02009;&#x000b1;&#x02009;0.19 (<italic>N</italic>&#x02009;=&#x02009;28) during the early dry and 0.55&#x02009;&#x000b1;&#x02009;0.19 (<italic>N</italic>&#x02009;=&#x02009;28) during the late dry season. The result of the LMM to estimate the effect of season on latrine use frequency (model 1) indicated that the full model containing the effects of season and its interaction with sex was not significantly better in explaining the data than the null model (likelihood ratio test, <italic>&#x003c7;</italic> ²&#x02009;=&#x02009;8.639, <italic>df</italic>&#x02009;=&#x02009;7, <italic>P</italic>&#x02009;=&#x02009;0.279).</p></sec><sec id="FPar12"><title>Mate defense hypothesis</title><p id="Par29">During 25 observations nights, we observed focal males to place anogenital scent marks and during 21 nights they engaged in branch bashing and vocal displays. One or both of these behaviors were recorded during 37 out of 51 observation nights on adult males. The result of the LMM to estimate the effect of perceived intruder pressure (as indicated by display and scent marking behavior) on latrine use frequency in males (model 2) showed that the full model was significantly better in explaining the data than the null model (likelihood ratio test, <italic>&#x003c7;</italic> ²&#x02009;=&#x02009;6.3327, <italic>df</italic>&#x02009;=&#x02009;1, <italic>P</italic>&#x02009;=&#x02009;0.012). Latrine use frequency was significantly increased in males during nights of perceived intruder pressure (mean frequency of latrine visitation&#x02009;&#x000b1;&#x02009;SD: nights with intruder pressure&#x02009;=&#x02009;0.60&#x02009;&#x000b1;&#x02009;0.27 latrine visitations/h, nights without intruder pressure&#x02009;=&#x02009;0.46&#x02009;&#x000b1;&#x02009;0.18; P&#x02009;=&#x02009;0.011; Table&#x000a0;<xref rid="Tab2" ref-type="table">2</xref>). In total, we recorded 50 scent marking events by the 7 focal males. 32 of these scent marks were placed in an identified latrine tree. At the critical distance of 6&#x000a0;m, the computed empirical inhomogeneous L-cross function fell above the simulation envelop, indicating spatial dependence (attraction) between latrines and scent marking locations (Fig.&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>). <table-wrap id="Tab2"><label>Table 2</label><caption><p>Effects of perceived intruder pressure, number of latrines, and social organization on latrine use frequency in male <italic>Lepilemur leucopus</italic> (LMM)</p></caption><table frame="hsides" rules="groups"><thead><tr><th>Fixed Factor</th><th>&#x003b2;</th><th>SE</th><th> <italic>df</italic> </th><th> <italic>t</italic> </th><th> <italic>P</italic> </th></tr></thead><tbody><tr><td>Intercept</td><td>1.019</td><td>0.002</td><td>9.55</td><td>473.012</td><td>NA</td></tr><tr><td>Intruder pressure perceived (yes)</td><td>&#x02212;0.005</td><td>0.002</td><td>44.44</td><td>&#x02212;2.658</td><td>0.011</td></tr><tr><td>Number of latrines</td><td>0.003</td><td>0.002</td><td>6.69</td><td>1.474</td><td>0.186</td></tr><tr><td>Social organization (1 &#x02642; and 2 &#x02640;)</td><td>&#x02212;0.007</td><td>0.004</td><td>7.49</td><td>&#x02212;1.672</td><td>0.136</td></tr></tbody></table></table-wrap> <fig id="Fig3"><label>Fig. 3</label><caption><p>Estimated inhomogeneous L-cross function and envelopes for the bivariate point pattern consisting of defecation/urination and scent marking sites. The <italic>solid line</italic> indicates the empirical L-cross function, the <italic>dotted line</italic> indicates the theoretical value for complete spatial randomness (CSR), the <italic>gray band</italic> indicates the envelope from 99 simulations, and <italic>r</italic> is the distance argument</p></caption><graphic xlink:href="265_2014_1810_Fig3_HTML" id="MO3"/></fig> </p></sec></sec><sec id="Sec4" sec-type="discussion"><title>Discussion</title><p id="Par30">Our study revealed that defecation/urination events were highly clustered in space, resulting in 3&#x02013;4 latrines with terrestrial accumulations of feces in each territory. The study animals spent only a notably short time in trees they visited for defecation/urination, and therefore, the formation of latrines is not a mere by-product of animals remaining for a considerable time in a few preferred resting trees (Charles-Dominique and Hladik <xref ref-type="bibr" rid="CR24">1971</xref>; Schilling <xref ref-type="bibr" rid="CR140">1979</xref>). The number and locations of latrines were stable throughout the study period. We tested whether terrestrial accumulations of feces in an arboreal species can be considered to have an olfactory signaling function. We found no support for this notion and conclude that urine, which is more accessible to the animals for olfactory investigation, is the more important latrine component in this species. Additionally, we found empirical support for the hypotheses that latrines function in social bonding and mate defense, but a potential function in territory demarcation, resource defense, and signaling of reproductive state could not be shown. Below, we discuss these findings in relation to the social system of <italic>L. leucopus</italic> and in light of available data for other latrine-using mammals.</p><sec id="FPar13"><title>Experimental translocation of feces</title><p id="Par31">Most species that exhibit latrine use are terrestrial, and feces are, therefore, assumed to be salient sources of olfactory signals. However, <italic>L. leucopus</italic> did not react differently to experimentally introduced feces from neighboring or strange social units, compared to feces from familiar animals. In contrast, river otters (<italic>Lontra canadensis</italic>) investigate foreign scat more than local one when added to latrines (Oldham and Black <xref ref-type="bibr" rid="CR107">2009</xref>). Brown brocket deer (<italic>Mazama gouazoubira</italic>) investigate introduced dung from unknown individuals of the same sex significantly more than their own dung, and males counter-mark introduced dung with a greater frequency than females (Black-Decima and Santana <xref ref-type="bibr" rid="CR12">2011</xref>). Badgers (<italic>Meles meles</italic>) respond more intensely towards foreign feces, and the response is greatest during the breeding season (Palphramand and White <xref ref-type="bibr" rid="CR110">2007</xref>). Among primates, only <italic>Cheirogaleus</italic> spp<italic>.</italic> produce arboreal latrines by smearing feces on branches during repeated walking defecation, resulting in a fecal accumulation adhering to the branch (Petter <xref ref-type="bibr" rid="CR114">1962</xref>). In arboreal species, such as <italic>L. leucopus</italic>, terrestrial latrines may serve as an optical signal (Irwin et al. <xref ref-type="bibr" rid="CR74">2004</xref>). Moreover, urination above ground facilitates dispersal of the odor by wind and increases the evaporating surface as the urine drips downward (Sillero-Zubiri and Macdonald <xref ref-type="bibr" rid="CR146">1998</xref>). Because urine marking is an ancestral behavior in strepsirrhine primates (Delbarco-Trillo et al. <xref ref-type="bibr" rid="CR33">2001</xref>), more experimental studies of urine communication in solitary and nocturnal species are called for.</p></sec><sec id="FPar14"><title>Social bonding</title><p id="Par32">Scent marks may function as self-advertisement and simply signal an individual&#x02019;s presence and identity to mates, family members, neighbors, and/or intruders (Eisenberg and Kleiman <xref ref-type="bibr" rid="CR40">1972</xref>; Peters and Mech <xref ref-type="bibr" rid="CR113">1975</xref>; Wolff et al. <xref ref-type="bibr" rid="CR162">2002</xref>), and latrines may serve as information exchange centers of individual-specific information (Darden et al. <xref ref-type="bibr" rid="CR30">2008</xref>; Black-Decima and Santana <xref ref-type="bibr" rid="CR12">2011</xref>). Latrines are maintained by all individuals of a social unit in <italic>L. leucopus</italic>. In contrast, in European badgers (<italic>Meles meles</italic>), a species in which latrines function mainly in territorial defense and demarcation, sexually immature juveniles rarely defecate/urinate at latrines (Brown et al. <xref ref-type="bibr" rid="CR21">2009</xref>). Latrines have been suggested to help maintaining social bonds in some ungulates such as steenbok (<italic>Raphicerus campestris</italic>), oribi (<italic>Ourebia ourebi</italic>), and dikdik (<italic>Madoqua kirkii</italic>; Kingdon <xref ref-type="bibr" rid="CR81">1982</xref>; Apio et al. <xref ref-type="bibr" rid="CR2">2006</xref>). Behaviors that facilitate familiarity, and hence, intra-group recognition may be especially important in solitary foragers with minimal direct social contact between individuals (Dr&#x000f6;scher and Kappeler <xref ref-type="bibr" rid="CR36">2013</xref>). In contrast, mated pairs of swift foxes (<italic>Vulpes velox</italic>) exhibit high levels of den sharing that allows the exchange of information within the pair and to maintain the pair bond. Thus, latrines are not considered important for intra-pair communication and maintenance of social cohesion in <italic>V. velox</italic> (Darden et al. <xref ref-type="bibr" rid="CR30">2008</xref>). Latrine locations within the core areas of <italic>L. leucopus</italic> also support the idea that they function in social bonding since this form of placement should be particularly suited for information exchange between group members (Wronski et al. <xref ref-type="bibr" rid="CR167">2013</xref>).</p><p id="Par33">In Coquerel&#x02019;s sifakas (<italic>Propithecus coquereli</italic>), the quality of the pair bond of breeding pairs is reflected in their olfactory signals by chemical convergence, possibly due to similar volatile production by shared microbial communities obtained through the exchange of odorant-producing microbes for example via overmarking (Greene and Drea <xref ref-type="bibr" rid="CR60">2014</xref>). Similarly, anal gland secretions that coat or saturate badger feces seem to have a group-specific chemical composition (Davies et al. <xref ref-type="bibr" rid="CR31">1988</xref>). Analogously, convergence in vocal signals facilitates group and pair cohesion in some primate and avian species (Geissmann and Orgeldinger <xref ref-type="bibr" rid="CR51">2000</xref>; Tyack <xref ref-type="bibr" rid="CR156">2008</xref>; Sewall <xref ref-type="bibr" rid="CR145">2009</xref>; Candiotti et al. <xref ref-type="bibr" rid="CR22">2012</xref>). Sportive lemurs not only exchange chemical but also acoustic signals. While pairs of the Milne Edwards&#x02019; sportive lemur (<italic>L. edwardsi</italic>) coordinate loud calls in duets, perhaps to strengthen pair bonds (M&#x000e9;ndez-C&#x000e1;rdenas and Zimmermann <xref ref-type="bibr" rid="CR94">2009</xref>), neither red-tailed sportive lemurs (<italic>L. ruficaudatus</italic>; Fichtel and Hilgartner <xref ref-type="bibr" rid="CR46">2013</xref>) nor <italic>L. leucopus</italic> exchange vocalizations in coordinated duets. In addition, males and females of <italic>L. leucopus</italic> produce sex-specific loud calls and thus are not available for vocal convergence. It, therefore, remains to be determined what exactly social bonding entails in different species and which aspects of it can be communicated in different modalities.</p></sec><sec id="FPar15"><title>Mate defense</title><p id="Par34">Latrines may play a role in mate defense by advertising the commitment of resident males to defend co-resident females (Roper et al. <xref ref-type="bibr" rid="CR133">1986</xref>; Jordan et al. <xref ref-type="bibr" rid="CR76">2007</xref>). We found that male latrine use frequency increased during nights of perceived intruder pressure. Likewise, latrine use frequency increases in meerkats (<italic>Suricata suricatta</italic>) when prospecting males are present (Jordan et al. <xref ref-type="bibr" rid="CR76">2007</xref>). In European badgers (<italic>Meles meles</italic>), males visit boundary latrines more often than females (Roper et al. <xref ref-type="bibr" rid="CR134">1993</xref>; Stewart et al. <xref ref-type="bibr" rid="CR150">2001</xref>), presumably to signal their commitment to guarding females of their own social group (Roper et al. <xref ref-type="bibr" rid="CR133">1986</xref>). Similarly, male brown brocket deer defecate/urinate more often after detecting dung from unknown individuals near one of their latrines. By re-marking their latrine, residents are thought to affirm their dominant or resident status (Black-Decima and Santana <xref ref-type="bibr" rid="CR12">2011</xref>).</p><p id="Par35">We do not have systematic data on the behavior of intruders. However, outside the focal observation period, we could observe a resident and a roaming male to repeatedly visit the same latrine tree to defecate, urinate and place glandular scent marks. Male scent marking is linked to intra-sexual competition in several species (e.g., <italic>Microtu</italic> sp.: Jannett <xref ref-type="bibr" rid="CR75">1986</xref>; <italic>Myocastor coypus</italic>: Gosling and Wright <xref ref-type="bibr" rid="CR57">1994</xref>; <italic>Lemur catta</italic>: Kappeler <xref ref-type="bibr" rid="CR78">1998</xref>), and by strategically placing anogenital scent marks in latrines, which are composite olfactory signals of all members of a group, males of <italic>L. leucopus</italic> may also signal their competitive ability and willingness to defend their social unit to intruders (Rich and Hurst <xref ref-type="bibr" rid="CR129">1998</xref>).</p></sec><sec id="FPar16"><title>Signaling of reproductive state</title><p id="Par36">Males are often able to detect chemical cues in female urine and/or feces related to reproductive state (Rasmussen et al. <xref ref-type="bibr" rid="CR122">1982</xref>; Ghosal et al. <xref ref-type="bibr" rid="CR52">2012</xref>; Archunan and Rajagopala <xref ref-type="bibr" rid="CR3">2013</xref>). Contrary to our predictions, frequency of latrine use in <italic>L. leucopus</italic> did not increase during the mating season. In contrast, genets (<italic>Genetta genetta</italic>) exhibit increased scat deposition at latrine sites during the mating period (Barrientos <xref ref-type="bibr" rid="CR8">2006</xref>). Similarly, latrine visitation peaks during the mating season in <italic>M. meles</italic> (Pigozzi <xref ref-type="bibr" rid="CR116">1989</xref>; Roper et al. <xref ref-type="bibr" rid="CR134">1993</xref>). While females may scent mark to advertise their reproductive state to attract males (Converse et al. <xref ref-type="bibr" rid="CR27">1995</xref>; Heymann <xref ref-type="bibr" rid="CR64">1998</xref>; Kappeler <xref ref-type="bibr" rid="CR78">1998</xref>), males may mask female scent to hide their oestrous condition from competing males or to advertise their presence to other males (Trumler <xref ref-type="bibr" rid="CR155">1958</xref>; Klingel <xref ref-type="bibr" rid="CR82">1974</xref>; Rich and Hurst <xref ref-type="bibr" rid="CR129">1998</xref>; Lewis <xref ref-type="bibr" rid="CR87">2005</xref>; Jordan et al. <xref ref-type="bibr" rid="CR76">2007</xref>). Although we cannot exclude the possibility that reproductive status may be communicated at latrine sites in <italic>L. leucopus</italic>, the function of latrine use does not appear to be specifically related to male attraction or to over-marking signals of estrous females, since neither females nor males increased latrine use frequency during the mating season. However, estrus in sportive lemurs is seasonal and short (Randrianambinina et al. <xref ref-type="bibr" rid="CR121">2007</xref>; Hilgartner et al. <xref ref-type="bibr" rid="CR67">2008</xref>) and any effect may have been concealed by our method of data collection, because we did not follow pairs when females were apparently in estrus.</p></sec><sec id="FPar17"><title>Territory demarcation</title><p id="Par37">Urine and feces are common, readily available materials and many mammals use them to demarcate their territories or home ranges (e.g., <italic>Meles meles</italic>; Pigozzi <xref ref-type="bibr" rid="CR116">1989</xref>; <italic>Panthera tigris</italic>: Smith et al. <xref ref-type="bibr" rid="CR148">1989</xref>; <italic>Ourebia ourebi</italic>: Brashares and Arcese <xref ref-type="bibr" rid="CR18">1999</xref>). We found that the majority of defecation/urination events were localized within the core areas of the territories, even though <italic>L. leucopus</italic> is highly territorial (Dr&#x000f6;scher and Kappeler <xref ref-type="bibr" rid="CR36">2013</xref>). However, where latrines cannot be economically maintained because territory borders are too long, they should be placed in the centre of the territory (Jordan et al. <xref ref-type="bibr" rid="CR76">2007</xref>). For example, brown hyenas (<italic>Hyaena brunnea</italic>) exhibit boundary marking when they live in small territories but display center marking if they inhabit large territories (Mills and Gorman <xref ref-type="bibr" rid="CR98">1987</xref>). Since territory size in <italic>L. leucopus</italic> is only 0.3&#x000a0;ha and individuals can easily traverse their territories in no more than 5&#x000a0;min, it is unlikely that territory size in this species would preclude a border marking strategy. In <italic>M. meles</italic>, latrine use is primarily concentrated along territory boundaries and these are shared by members of the same and neighboring groups (Kilshaw et al. <xref ref-type="bibr" rid="CR80">2009</xref>) and are visited mainly by males (Roper et al. <xref ref-type="bibr" rid="CR134">1993</xref>). Besides boundary latrines, badgers also use hinterland latrines, which are visited by both sexes (Roper et al. <xref ref-type="bibr" rid="CR134">1993</xref>). In <italic>L. leucopus</italic>, all latrines were visited by both pair partners. Furthermore, we could observe only once a focal animal to visit a neighboring latrine, indicating that latrines in <italic>L. leucopus</italic> are not used for inter-group information transfer to monitor occupancy of surrounding territories (Jordan et al. <xref ref-type="bibr" rid="CR76">2007</xref>). Instead of latrines, sportive lemurs seem to use vocalizations to signal occupancy and to regulate spacing within and between social units (Rasoloharijaona et al. <xref ref-type="bibr" rid="CR124">2006</xref>; Fichtel and Hilgartner <xref ref-type="bibr" rid="CR46">2013</xref>).</p></sec><sec id="FPar18"><title>Resource defense</title><p id="Par38">Resources such as resting sites (Goszczynski <xref ref-type="bibr" rid="CR58">1990</xref>; Branch <xref ref-type="bibr" rid="CR17">1993</xref>; Brady and Armitage <xref ref-type="bibr" rid="CR16">1999</xref>) and food trees may be marked to identify ownership and to deter conspecifics (Kruuk <xref ref-type="bibr" rid="CR83">1992</xref>; Miller et al. <xref ref-type="bibr" rid="CR97">2003</xref>). Contrary to our prediction, spatial locations of latrine trees and sleeping trees were spatially independent from each other, notwithstanding the fact that sportive lemurs only use a few selected sleeping sites and appropriate sleeping sites are limited, potentially leading to competition within or between social units (Rasoloharijaona et al. <xref ref-type="bibr" rid="CR123">2003</xref>, <xref ref-type="bibr" rid="CR125">2008</xref>). Establishing ownership of sleeping sites, therefore, may be beneficial to individuals by ensuring protection from predators or adverse climatic conditions (Franklin et al. <xref ref-type="bibr" rid="CR50">2007</xref>). For example, weasel sportive lemurs (<italic>L. mustelinus</italic>) gouge trees after leaving sleeping sites and before moving around, suggesting that they use non-nutritive tree gouging to display ownership of sleeping sites (Rasoloharijaona et al. <xref ref-type="bibr" rid="CR126">2010</xref>). Tree gouging behavior is absent in <italic>L. leucopus</italic>, and if latrines were to function instead for sleeping site defense, one would expect latrine trees to be in proximity to sleeping trees. Conversely, scent marks can potentially be exploited by predators to localize prey (Cushing <xref ref-type="bibr" rid="CR29">1984</xref>; Viitala et al. <xref ref-type="bibr" rid="CR159">1995</xref>), and an intentional placement of latrine trees in proximity to sleeping trees would seem to be disadvantageous in terms of predator attraction. In addition, animals may mark food trees as a means of asserting ownership of food resources.</p><p id="Par39">Communal use of latrines in <italic>L. leucopus</italic> rejects the idea that they are used to signal resource use among members of a social unit. In contrast, otters (<italic>Lutra lutra</italic>) deposit spraints (i.e., token feces) to signal the use of feeding areas exploited by each individual (Kruuk <xref ref-type="bibr" rid="CR83">1992</xref>). Alternatively, members of a social unit of <italic>L. leucopus</italic> may use latrines to signal to other social units their willingness to defend their food resources. However, <italic>L. leucopus</italic> did not preferentially defecate/urinate in food trees since animals were observed to defecate/urinate in only 5&#x000a0;% of all identified food patches and to feed in less than 50&#x000a0;% of the identified latrine trees. In addition, the fact that individual feeding effort was equally distributed within and outside the latrine area indicates that latrines are not used to mark important feeding areas. These results are in line with the observation that <italic>L. leucopus</italic> exhibits low dietary selectivity, relies on the most common food species, and rarely engages in conflict over food neither within nor between social units (Dr&#x000f6;scher and Kappeler <xref ref-type="bibr" rid="CR37">2014</xref>).</p></sec></sec><sec id="Sec5" sec-type="conclusion"><title>Conclusions</title><p id="Par40">Latrines are found in solitary, pair-, and group-living mammals (Table&#x000a0;<xref rid="Tab3" ref-type="table">3</xref>). Latrine use appears to be common among species that are nocturnal, exhibit a dispersed social system, and are territorial. Since many species do not just defecate, but often also urinate and deposit glandular secrets at latrine sites, these signals may function to convey more than one message. Especially in arboreal species with terrestrial accumulations of feces, urine may be of greater importance for chemical signaling than feces. Despite comparative data being sparse, a general pattern emerges that latrines are used in intra-specific olfactory communication in many cases. Although not restricted to nocturnal species, latrine use may facilitate communication in species with limited habitat visibility. Furthermore, latrines can be considered to be economical in species with low inter-individual cohesion, since individuals can benefit from predictable areas for information exchange. Notwithstanding the fact of being more common among territorial species, latrine use does not appear to necessarily function in territory demarcation. Clearly, more experimental studies are required to investigate the relative importance and functions of different modes of olfactory signaling at latrine sites.<table-wrap id="Tab3"><label>Table 3</label><caption><p>Overview of mammalian latrine users and species-specific attributes such as habitat use (<italic>T</italic>&#x02009;=&#x02009;terrestrial, <italic>A</italic>&#x02009;=&#x02009;arboreal, <italic>AQ</italic>&#x02009;=&#x02009;aquatic), period of activity (<italic>D</italic>&#x02009;=&#x02009;diurnal, <italic>N</italic>&#x02009;=&#x02009;nocturnal, <italic>C</italic>&#x02009;=&#x02009;crepuscular), social organization (<italic>S</italic>&#x02009;=&#x02009;solitary, <italic>P</italic>&#x02009;=&#x02009;pair, <italic>G</italic>&#x02009;=&#x02009;group), and cohesiveness during foraging (<italic>G</italic>&#x02009;=&#x02009;gregarious, <italic>D</italic>&#x02009;=&#x02009;dispersed) as well as suggested function of latrine use (<italic>1</italic>&#x02009;=&#x02009;territory demarcation, <italic>2</italic>&#x02009;=&#x02009;resource defense, <italic>3</italic>&#x02009;=&#x02009;centers of information exchange, <italic>4</italic>&#x02009;=&#x02009;reproductive signaling, <italic>5</italic>&#x02009;=&#x02009;mate defense/intrasexual competition, <italic>6</italic>&#x02009;=&#x02009;signaling of social status)</p></caption><table frame="hsides" rules="groups"><thead><tr><th>Order</th><th>Species</th><th>Common name</th><th>Habitat</th><th>Activity</th><th>Social organization</th><th>Cohesion</th><th>Territoriality</th><th>Function</th><th>Reference</th></tr></thead><tbody><tr><td>Artiodactyla</td><td> <italic>Alcelaphus buselaphus</italic> </td><td>Hartebeest</td><td>T</td><td>D</td><td>G</td><td>G</td><td>Yes</td><td/><td>Gosling (<xref ref-type="bibr" rid="CR55">1974</xref>)</td></tr><tr><td/><td> <italic>Cervus eldi</italic> </td><td>Eld&#x02019;s deer</td><td>T</td><td>N/C</td><td>G</td><td>G</td><td>No</td><td/><td>Wemmer and Montali (<xref ref-type="bibr" rid="CR161">1988</xref>)</td></tr><tr><td/><td> <italic>Damaliscus korrigum</italic> </td><td>Topi</td><td>T</td><td>N/D</td><td>G</td><td>G</td><td>Yes</td><td>1</td><td>Gosling (<xref ref-type="bibr" rid="CR56">1987</xref>)</td></tr><tr><td/><td> <italic>Gazella dorcas</italic> </td><td>Dorcas gazelle</td><td>T</td><td>N/D/C</td><td>P/G</td><td>G</td><td>Yes</td><td/><td>Essghaier and Johnson (<xref ref-type="bibr" rid="CR43">1981</xref>)</td></tr><tr><td/><td> <italic>Gazella gazella</italic> </td><td>Mountain gazelle</td><td>T</td><td>D</td><td>G</td><td>G</td><td>Yes</td><td/><td>Wronski and Plath (<xref ref-type="bibr" rid="CR165">2010</xref>)</td></tr><tr><td/><td> <italic>Gazella granti</italic> </td><td>Grant&#x02019;s gazelle</td><td>T</td><td>N/D</td><td>G</td><td>G</td><td>Yes</td><td>1</td><td>Estes (<xref ref-type="bibr" rid="CR44">1991</xref>)</td></tr><tr><td/><td> <italic>Gazella thomsoni</italic> </td><td>Thomson&#x02019;s gazelle</td><td>T</td><td>N/D</td><td>G</td><td>G</td><td>Yes</td><td/><td>Walther (<xref ref-type="bibr" rid="CR160">1978</xref>)</td></tr><tr><td/><td> <italic>Hydropotes inermis</italic> </td><td>Water deer</td><td>T</td><td>C</td><td>S</td><td>D</td><td>Yes</td><td/><td>Sun et al. (<xref ref-type="bibr" rid="CR152">1994</xref>)</td></tr><tr><td/><td> <italic>Lama guanicoe</italic> </td><td>Guanaco</td><td>T</td><td>D</td><td>G</td><td>G</td><td>Yes</td><td/><td>Henriquez (<xref ref-type="bibr" rid="CR63">2004</xref>)</td></tr><tr><td/><td> <italic>Madoqua guentheri</italic> </td><td>Guenther&#x02019;s dik-dik</td><td>T</td><td>N/D</td><td>P</td><td>G</td><td>Yes</td><td>1</td><td>Ono et al. (<xref ref-type="bibr" rid="CR108">1988</xref>)</td></tr><tr><td/><td> <italic>Madoqua kirkii</italic> </td><td>Kirk&#x02019;s dik-dik</td><td>T</td><td>N/D</td><td>P</td><td>G</td><td>Yes</td><td>3</td><td>Hendrichs and Hendrichs (<xref ref-type="bibr" rid="CR62">1971</xref>)</td></tr><tr><td/><td> <italic>Mazama americana</italic> </td><td>Red brocket deer</td><td>T</td><td>N/D</td><td>S/P</td><td>D</td><td>Yes</td><td/><td>Rivero et al. (<xref ref-type="bibr" rid="CR130">2004</xref>)</td></tr><tr><td/><td> <italic>Mazama gouazoubira</italic> </td><td>Brown brocket deer</td><td>T</td><td>N</td><td>S</td><td>D</td><td>Yes</td><td>3,5</td><td>Black-Decima and Santana (<xref ref-type="bibr" rid="CR12">2011</xref>)</td></tr><tr><td/><td> <italic>Moschus chrysogaster</italic> </td><td>Alpine musk deer</td><td>T</td><td>N</td><td>G</td><td>D</td><td>Yes</td><td/><td>Qureshi et al. (<xref ref-type="bibr" rid="CR118">2004</xref>)</td></tr><tr><td/><td> <italic>Moschus moschiferus</italic> </td><td>Siberian musk deer</td><td>T</td><td>N</td><td>G</td><td>D</td><td>Yes</td><td/><td>Green (<xref ref-type="bibr" rid="CR59">1987</xref>)</td></tr><tr><td/><td> <italic>Muntiacus muntjak</italic> </td><td>Indian muntjac</td><td>T</td><td>N/D</td><td>S</td><td>D</td><td>Yes</td><td>1</td><td>Dubost (<xref ref-type="bibr" rid="CR39">1971</xref>)</td></tr><tr><td/><td> <italic>Muntiacus reevesi</italic> </td><td>Chinese muntjac</td><td>T</td><td>N/D</td><td>S</td><td>D</td><td>Yes</td><td>1</td><td>Dubost (<xref ref-type="bibr" rid="CR38">1970</xref>)</td></tr><tr><td/><td> <italic>Oreotragus oreotragus</italic> </td><td>Klipspringer</td><td>T</td><td>D</td><td>P</td><td>G</td><td>Yes</td><td>1</td><td>Roberts and Lowen (<xref ref-type="bibr" rid="CR131">1997</xref>)</td></tr><tr><td/><td> <italic>Ourebia ourebi</italic> </td><td>Oribi</td><td>T</td><td>D</td><td>S/P/G</td><td>G</td><td>Yes</td><td>1,3</td><td>Brashares and Arcese (<xref ref-type="bibr" rid="CR18">1999</xref>)</td></tr><tr><td/><td> <italic>Pudu puda</italic> </td><td>Southern pudu</td><td>T</td><td>N/D</td><td>S</td><td>D</td><td>Yes</td><td/><td>MacNamara and Eldridge (<xref ref-type="bibr" rid="CR91">1987</xref>)</td></tr><tr><td/><td> <italic>Raphicerus campestris</italic> </td><td>Steinbuck</td><td>T</td><td>D</td><td>P</td><td>D</td><td>Yes</td><td>3</td><td>Kingdon (<xref ref-type="bibr" rid="CR81">1982</xref>)</td></tr><tr><td/><td> <italic>Tragelaphus scriptus</italic> </td><td>Bushbuck</td><td>T</td><td>N/C</td><td>G</td><td>D</td><td>Yes</td><td>3,4</td><td>Wronski et al. (<xref ref-type="bibr" rid="CR166">2006</xref>)</td></tr><tr><td/><td> <italic>Vicugna pacos</italic> </td><td>Alpaca</td><td>T</td><td>D</td><td>G</td><td>G</td><td>Yes</td><td/><td>McGregor and Brown (<xref ref-type="bibr" rid="CR92">2010</xref>)</td></tr><tr><td>Perissodactyla</td><td> <italic>Ceratotherium simum</italic> </td><td>White rhinoceros</td><td>T</td><td>N/D</td><td>S/G</td><td>G</td><td>Yes</td><td/><td>Owen-Smith (<xref ref-type="bibr" rid="CR109">1975</xref>)</td></tr><tr><td/><td> <italic>Diceros bicornis</italic> </td><td>Black rhinoceros</td><td>T</td><td>N/D</td><td>S</td><td>D</td><td>Yes</td><td/><td>Linklater et al. (<xref ref-type="bibr" rid="CR89">2013</xref>)</td></tr><tr><td/><td> <italic>Rhinocerus unicornis</italic> </td><td>Indian rhinoceros</td><td>T</td><td>N/D</td><td>S</td><td>D</td><td>Yes</td><td/><td>Dinerstein and Wemmer (<xref ref-type="bibr" rid="CR34">1988</xref>)</td></tr><tr><td/><td> <italic>Tapirus terrestris</italic> </td><td>South American tapir</td><td>T</td><td>N/C</td><td>S</td><td>D</td><td>Yes</td><td/><td>Fragoso et al. (<xref ref-type="bibr" rid="CR49">2003</xref>)</td></tr><tr><td>Carnivora</td><td> <italic>Bassariscus astutus</italic> </td><td>Ring-tailed cat</td><td>T</td><td>N/C</td><td>S</td><td>D</td><td>Yes</td><td/><td>Barja and List (<xref ref-type="bibr" rid="CR6">2006</xref>)</td></tr><tr><td/><td> <italic>Canis aureus</italic> </td><td>Golden jackal</td><td>T</td><td>N/D</td><td>G</td><td>D</td><td>Yes</td><td/><td>Macdonald (<xref ref-type="bibr" rid="CR90">1980</xref>)</td></tr><tr><td/><td> <italic>Canis latrans</italic> </td><td>Coyote</td><td>T</td><td>N/D</td><td>S/P/G</td><td>D,G</td><td>Yes</td><td/><td>Ralls and Smith (<xref ref-type="bibr" rid="CR120">2004</xref>)</td></tr><tr><td/><td> <italic>Canis simensis</italic> </td><td>Ethopian wolf</td><td>T</td><td>D</td><td>G</td><td>D</td><td>Yes</td><td/><td>Sillero-Zubiri and Macdonald (<xref ref-type="bibr" rid="CR146">1998</xref>)</td></tr><tr><td/><td> <italic>Civettictis civetta</italic> </td><td>African civet</td><td>T</td><td>N</td><td>S</td><td>D</td><td>Yes</td><td/><td>Bearder and Randall (<xref ref-type="bibr" rid="CR11">1978</xref>)</td></tr><tr><td/><td> <italic>Crocuta crocuta</italic> </td><td>Spotted hyena</td><td>T</td><td>N</td><td>G</td><td>G</td><td>Yes</td><td/><td>Gorman and Mills (<xref ref-type="bibr" rid="CR54">1984</xref>)</td></tr><tr><td/><td> <italic>Genetta genetta</italic> </td><td>Common genet</td><td>T, A</td><td>N</td><td>S/P</td><td>D</td><td>Yes</td><td>4,5</td><td>Barrientos (<xref ref-type="bibr" rid="CR8">2006</xref>)</td></tr><tr><td/><td> <italic>Hyaena brunnea</italic> </td><td>Brown hyena</td><td>T</td><td>N</td><td>G</td><td>G</td><td>Yes</td><td>1</td><td>Mills et al. (<xref ref-type="bibr" rid="CR99">1980</xref>)</td></tr><tr><td/><td> <italic>Hyaena hyaena</italic> </td><td>Striped hyena</td><td>T</td><td>N</td><td>G</td><td>D</td><td>Yes</td><td/><td>Macdonald (<xref ref-type="bibr" rid="CR90">1980</xref>)</td></tr><tr><td/><td> <italic>Lontra canadensis</italic> </td><td>River otter</td><td>T, AQ</td><td>N/C</td><td>G</td><td>G</td><td>Yes</td><td>6</td><td>Rostain et al. (<xref ref-type="bibr" rid="CR135">2004</xref>)</td></tr><tr><td/><td> <italic>Martes martes</italic> </td><td>Pine marten</td><td>T,A</td><td>N</td><td>S</td><td>D</td><td>Yes</td><td/><td>Barja et al. (<xref ref-type="bibr" rid="CR7">2011</xref>)</td></tr><tr><td/><td> <italic>Meles meles</italic> </td><td>European badger</td><td>T</td><td>N/C</td><td>G</td><td>D</td><td>Yes</td><td>1,2,4,5</td><td>Roper et al. (<xref ref-type="bibr" rid="CR134">1993</xref>), Balestrieri et al. (<xref ref-type="bibr" rid="CR5">2011</xref>)</td></tr><tr><td/><td> <italic>Nyctereutes procyonoides</italic> </td><td>Raccoon dog</td><td>T</td><td>N</td><td>P</td><td>D</td><td>No</td><td>3</td><td>Ikeda (<xref ref-type="bibr" rid="CR73">1984</xref>)</td></tr><tr><td/><td> <italic>Procyon lotor</italic> </td><td>Northern raccoon</td><td>T</td><td>N</td><td>G</td><td>D</td><td>Variable</td><td/><td>Brown and Macdonald (<xref ref-type="bibr" rid="CR20">1985</xref>)</td></tr><tr><td/><td> <italic>Proteles cristatus</italic> </td><td>Aardwolf</td><td>T</td><td>N</td><td>P</td><td>D</td><td>Yes</td><td/><td>Nel and Bothma (<xref ref-type="bibr" rid="CR103">2002</xref>)</td></tr><tr><td/><td> <italic>Pteronura brasiliensis</italic> </td><td>Giant otters</td><td>T, AQ</td><td>D</td><td>G</td><td>G</td><td>Yes</td><td/><td>Leuchtenberger and Mour&#x000e3;o (<xref ref-type="bibr" rid="CR86">2009</xref>)</td></tr><tr><td/><td> <italic>Suricata suricatta</italic> </td><td>Meerkats</td><td>T</td><td>D</td><td>G</td><td>G</td><td>Yes</td><td>1,5</td><td>Jordan et al. (<xref ref-type="bibr" rid="CR76">2007</xref>)</td></tr><tr><td/><td> <italic>Urocyon cinereoargenteus</italic> </td><td>Gray fox</td><td>T</td><td>N/C</td><td>P</td><td>D</td><td>Yes</td><td/><td>Trapp (<xref ref-type="bibr" rid="CR154">1978</xref>)</td></tr><tr><td/><td> <italic>Vulpes macrotis</italic> </td><td>Kit fox</td><td>T</td><td>N</td><td>P</td><td>D</td><td>Yes</td><td/><td>Ralls and Smith (<xref ref-type="bibr" rid="CR120">2004</xref>)</td></tr><tr><td/><td> <italic>Vulpes velox</italic> </td><td>Swift fox</td><td>T</td><td>N</td><td>P</td><td>D</td><td>Yes</td><td>1</td><td>Darden et al. (<xref ref-type="bibr" rid="CR30">2008</xref>)</td></tr><tr><td>Dasyuromorphia</td><td> <italic>Dasyurus geoffroii</italic> </td><td>Western quoll</td><td>T</td><td>N/C</td><td>S</td><td>D</td><td>Yes</td><td/><td>Serena and Soderquist (<xref ref-type="bibr" rid="CR144">1989</xref>)</td></tr><tr><td/><td> <italic>Dasyurus hallucatus</italic> </td><td>Northern quoll</td><td>T</td><td>N</td><td>S</td><td>D</td><td>No</td><td/><td>Oakwood (<xref ref-type="bibr" rid="CR106">2002</xref>)</td></tr><tr><td/><td> <italic>Dasyurus maculatus</italic> </td><td>Tiger quoll</td><td>T</td><td>N</td><td>S</td><td>D</td><td>Yes</td><td/><td>Ruibal et al. (<xref ref-type="bibr" rid="CR136">2010</xref>)</td></tr><tr><td/><td> <italic>Myrmecobius fasciatus</italic> </td><td>Numbat</td><td>T</td><td>D</td><td>S</td><td>D</td><td>Yes</td><td>1</td><td>Hogan et al. (<xref ref-type="bibr" rid="CR70">2013</xref>)</td></tr><tr><td/><td> <italic>Sarcophilus harrisii</italic> </td><td>Tasmanian devil</td><td>T</td><td>N</td><td>S</td><td>D</td><td>No</td><td/><td>Pemberton (<xref ref-type="bibr" rid="CR111">1990</xref>)</td></tr><tr><td>Diprotodontia</td><td> <italic>Petropseudes dahli</italic> </td><td>Rock-haunting possum</td><td>T</td><td>N</td><td>P</td><td>G</td><td>Yes</td><td/><td>Runcie (<xref ref-type="bibr" rid="CR137">2004</xref>)</td></tr><tr><td>Hyracoidea</td><td> <italic>Dendrohyrax arboreus</italic> </td><td>Southern tree hyrax</td><td>A</td><td>N/D</td><td>S/P</td><td>D</td><td>Yes</td><td/><td>Milner and Harris (<xref ref-type="bibr" rid="CR100">1999</xref>)</td></tr><tr><td/><td> <italic>Dendrohyrax validus</italic> </td><td>Eastern tree hyrax</td><td>A</td><td>N</td><td>?</td><td>D</td><td>Yes</td><td/><td>Topp-J&#x000f8;rgensen et al. (<xref ref-type="bibr" rid="CR153">2008</xref>)</td></tr><tr><td/><td> <italic>Heterohyrax brucei</italic> </td><td>Yellow-spotted rock hyrax</td><td>T</td><td>D</td><td>G</td><td>G</td><td>Yes</td><td/><td>Barry and Shoshani (<xref ref-type="bibr" rid="CR9">2000</xref>)</td></tr><tr><td/><td> <italic>Procavia capensis</italic> </td><td>Rock hyrax</td><td>T</td><td>D</td><td>G</td><td>G</td><td>Yes</td><td/><td>Meadows et al. (<xref ref-type="bibr" rid="CR93">2010</xref>)</td></tr><tr><td>Lagomorpha</td><td> <italic>Oryctolagus cuniculus</italic> </td><td>European rabbit</td><td>T</td><td>N</td><td>G</td><td>G</td><td>Yes</td><td/><td>Sneddon (<xref ref-type="bibr" rid="CR149">1991</xref>)</td></tr><tr><td>Primates</td><td> <italic>Alouatta caraya</italic> </td><td>Black howler monkey</td><td>A</td><td>D</td><td>G</td><td>G</td><td>Yes</td><td/><td>Bravo and Zunino (<xref ref-type="bibr" rid="CR19">2000</xref>)</td></tr><tr><td/><td> <italic>Alouatta seniculus</italic> </td><td>Red howler monkey</td><td>A</td><td>D</td><td>G</td><td>G</td><td>Yes</td><td/><td>Julliot (<xref ref-type="bibr" rid="CR77">1996</xref>)</td></tr><tr><td/><td> <italic>Ateles geoffroyi</italic> </td><td>Geoffroy&#x02019;s spider monkey</td><td>A</td><td>D</td><td>G</td><td>G</td><td>Yes</td><td/><td>Gonz&#x000e1;lez-Zamora et al. (<xref ref-type="bibr" rid="CR53">2012</xref>)</td></tr><tr><td/><td> <italic>Cheirogaleus major</italic> </td><td>Greater dwarf lemur</td><td>A</td><td>N</td><td>P</td><td>D</td><td>Yes</td><td/><td>Petter (<xref ref-type="bibr" rid="CR114">1962</xref>)</td></tr><tr><td/><td> <italic>Cheirogaleus medius</italic> </td><td>Fat-tailed dwarf lemur</td><td>A</td><td>N</td><td>P</td><td>D</td><td>Yes</td><td/><td>Petter (<xref ref-type="bibr" rid="CR114">1962</xref>)</td></tr><tr><td/><td> <italic>Hapalemur griseus</italic> </td><td>Lesser bamboo lemur</td><td>A</td><td>D</td><td>G</td><td>G</td><td>Yes</td><td>2,4,5</td><td>Irwin et al. (<xref ref-type="bibr" rid="CR74">2004</xref>)</td></tr><tr><td/><td> <italic>Hapalemur meridionalis</italic> </td><td>Southern lesser bamboo lemur</td><td>A</td><td>D</td><td>G</td><td>G</td><td>Yes</td><td>1,2</td><td>Eppley and Donati (<xref ref-type="bibr" rid="CR42">2010</xref>)</td></tr><tr><td/><td> <italic>Lagothrix lagotricha</italic> </td><td>Woolly monkey</td><td>A</td><td>D</td><td>G</td><td>G</td><td>Yes</td><td/><td>Yumoto et al. (<xref ref-type="bibr" rid="CR169">1999</xref>)</td></tr><tr><td/><td> <italic>Lepilemur leucopus</italic> </td><td>White-footed sportive lemur</td><td>A</td><td>N</td><td>P</td><td>D</td><td>Yes</td><td>3,5</td><td>This study</td></tr><tr><td/><td> <italic>Lepilemur wrightae</italic> </td><td>Wright&#x02019;s sportive lemur</td><td>A</td><td>N</td><td>P</td><td>D</td><td>Yes</td><td>2,4,5</td><td>Irwin et al. (<xref ref-type="bibr" rid="CR74">2004</xref>)</td></tr><tr><td>Rodentia</td><td> <italic>Arvicola terrestris</italic> </td><td>Water vole</td><td>T, AQ</td><td>N</td><td>S</td><td>D</td><td>Yes</td><td>4</td><td>Woodroffe and Lawton (<xref ref-type="bibr" rid="CR163">1990</xref>)</td></tr></tbody></table></table-wrap> </p></sec><sec sec-type="supplementary-material"><title>Electronic supplementary material</title><sec id="Sec6"><p>Below is the link to the electronic supplementary material.<supplementary-material content-type="local-data" id="MOESM1"><media xlink:href="265_2014_1810_MOESM1_ESM.docx"><label>ESM 1</label><caption><p>(DOCX 21 kb)</p></caption></media></supplementary-material> </p></sec></sec>

Corrosion current density prediction in reinforced concrete by imperialist competitive algorithm

Could not extract abstract

<contrib contrib-type="author" corresp="yes"><name><surname>Sadowski</surname><given-names>Lukasz</given-names></name><address><email>lukasz.sadowski@pwr.edu.pl</email></address><xref ref-type="aff" rid="Aff1">1</xref></contrib><contrib contrib-type="author"><name><surname>Nikoo</surname><given-names>Mehdi</given-names></name><address><email>sazeh84@yahoo.com</email></address><xref ref-type="aff" rid="Aff2">2</xref></contrib><aff id="Aff1"><label>1</label>Faculty of Civil Engineering, Wroclaw University of Technology, Wybrzeze Wyspianskiego 27, 50-370 Wroc&#x00142;aw, Poland </aff><aff id="Aff2"><label>2</label>SAMA Technical and Vocational Training College, Islamic Azad University, Ahvaz Branch, Ahvaz, Iran </aff>

Neural Computing &#x00026; Applications

<sec id="Sec1"><title>Introduction</title><p>Reinforced concrete (RC) is one of the most commonly used construction materials in civil engineering industry, and reinforcement corrosion is a major problem. For this reason, inspection techniques are used to evaluate steel corrosion in concrete in order to protect RC structures [<xref ref-type="bibr" rid="CR1">1</xref>&#x02013;<xref ref-type="bibr" rid="CR9">9</xref>].</p><p>Tuutti developed the model for predicting the service life of reinforcing steel [<xref ref-type="bibr" rid="CR10">10</xref>]. According to this model, the corrosion process has two distinct stages: corrosion initiation and corrosion propagation (Fig.&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>a). Once this process has started, the time until damage occurs will mostly depend on relative humidity, the availability of oxygen, and temperature. When corrosion is initiated, active corrosion results in a volumetric expansion of the rust around the reinforcing bars against the surrounding concrete [<xref ref-type="bibr" rid="CR10">10</xref>]. In the corrosion initiation stage, the required chemical reactions take place in concrete cover to initiate corrosion process. These chemical reactions may be carbonation or chloride ion attack.<fig id="Fig1"><label>Fig.&#x000a0;1</label><caption><p>Models for corrosion of steel reinforcement in concrete: <bold>a</bold> model for predicting the service life of reinforcing steel [<xref ref-type="bibr" rid="CR10">10</xref>], <bold>b</bold> performance-based service life model [<xref ref-type="bibr" rid="CR11">11</xref>]</p></caption><graphic xlink:href="521_2014_1645_Fig1_HTML" id="MO1"/></fig> </p><p>Liu and Weyers [<xref ref-type="bibr" rid="CR11">11</xref>] developed a performance-based service life model presented in Fig.&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>b, which uses structural engineering performance criteria: serviceability and strength limit states. In service life model, the corrosion process represents three distinct phenomena: depassivation, propagation, and final state. Depassivation is the loss of oxide layer over the rebar due to the high alkalinity of concrete. Depassivation takes an initiation period <italic>T</italic> _i (time from completion of the new built structures to the time of corrosion initiation in the structure). The second life cycle is the propagation period <italic>T</italic> _s from the initiation of corrosion to corrosion-induced unserviceability of the structure. The third life cycle is the time period <italic>T</italic> _f from loss of serviceability to ultimate failure [<xref ref-type="bibr" rid="CR12">12</xref>&#x02013;<xref ref-type="bibr" rid="CR15">15</xref>]. </p><p>It is proper to note that once the corrosion is initiated, concrete resistivity plays an important role in deciding reinforcement corrosion [<xref ref-type="bibr" rid="CR16">16</xref>&#x02013;<xref ref-type="bibr" rid="CR18">18</xref>]. The concrete resistivity measurement provides additional information to assist in assessing corrosion process. As mentioned in [<xref ref-type="bibr" rid="CR12">12</xref>], corrosion propagation is the phase in which the accelerated corrosion leads to rust staining, cracking of the concrete cover, and the deterioration mechanism in RC structures generally occurring in the form of longitudinal cracking, spalling, and delamination of concrete cover. Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref> shows the deterioration mechanism of RC element as a result of corrosion process.<fig id="Fig2"><label>Fig.&#x000a0;2</label><caption><p>Deterioration mechanism of RC element as a result of corrosion process: <bold>a</bold> element prior to corrosion initiation, <bold>b</bold> expansive corrosion initiation, <bold>c</bold> crack propagation, <bold>d</bold> concrete cracking</p></caption><graphic xlink:href="521_2014_1645_Fig2_HTML" id="MO2"/></fig> </p><p>One of the most useful methods of providing a direct evaluation of the corrosion rate by corrosion current density measurement <italic>i</italic> _corr is using the linear polarization resistance (LPR). In this method, a specific voltage shift (typically 10&#x000a0;mV) is applied to an electrode in solution. As mentioned in [<xref ref-type="bibr" rid="CR16">16</xref>], the instantaneous corrosion current density <italic>i</italic> _corr is obtained by dividing a Stern&#x02013;Geary [<xref ref-type="bibr" rid="CR20">20</xref>] constant <italic>B</italic> by the polarization resistance <italic>R</italic> _{<italic>p</italic>} value:<disp-formula id="Equ1"><label>1</label><alternatives><tex-math id="M1">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$i_{\text{corr}} = B/R_{\text{p}}$$\end{document}</tex-math><mml:math id="M2" display="block"><mml:mrow><mml:msub><mml:mi>i</mml:mi><mml:mtext>corr</mml:mtext></mml:msub><mml:mo>=</mml:mo><mml:mi>B</mml:mi><mml:mo stretchy="false">/</mml:mo><mml:msub><mml:mi>R</mml:mi><mml:mtext>p</mml:mtext></mml:msub></mml:mrow></mml:math></alternatives></disp-formula>By measuring <italic>i</italic> _corr, a corrosion rate can be derived. Typical corrosion rates from LPR measurements are presented in Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>.<table-wrap id="Tab1"><label>Table&#x000a0;1</label><caption><p>Typical corrosion rates from LPR measurements [<xref ref-type="bibr" rid="CR19">19</xref>]</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left">Corrosion classification</th><th align="left">Corrosion current density <italic>i</italic> _corr (&#x003bc;A/cm²)</th><th align="left">Corrosion penetration rate^a (&#x003bc;m/year)</th></tr></thead><tbody><tr><td align="left">Passive/very low</td><td align="left">Up to 0.2</td><td align="left">Up to 2</td></tr><tr><td align="left">Low/mod</td><td align="left">0.2 to 0.5</td><td align="left">2 to 6</td></tr><tr><td align="left">Mod/high</td><td align="left">0.5 to 1.0</td><td align="left">6 to 12</td></tr><tr><td align="left">Very high</td><td char="." align="char">&#x0003e;1.0</td><td char="." align="char">&#x0003e;12</td></tr></tbody></table><table-wrap-foot><p> ^aLoss of reinforcement section from Faraday&#x02019;s Law, assuming Fe&#x000a0;&#x02192;&#x000a0;Fe</p></table-wrap-foot></table-wrap> </p><p>LPR has its disadvantages because it can only be effectively performed in relatively clean aqueous electrolytic environments and in terms of destroying concrete lining to measure the amount of steel corrosion [<xref ref-type="bibr" rid="CR21">21</xref>]. LPR will not work in gases or liquid emulsions where fouling of the electrodes will prevent measurements [<xref ref-type="bibr" rid="CR22">22</xref>, <xref ref-type="bibr" rid="CR23">23</xref>]. It has been presented that corrosion current density is one of the most important input parameter in the corrosion-induced damage models [<xref ref-type="bibr" rid="CR24">24</xref>]. Researchers developed corrosion current density prediction models based on the electrochemical principles of steel reinforcement corrosion [<xref ref-type="bibr" rid="CR25">25</xref>], experimental testing [<xref ref-type="bibr" rid="CR26">26</xref>], and statistical analysis [<xref ref-type="bibr" rid="CR11">11</xref>].</p><p>It seems sensible to employ more parameters in one model to obtain a more accurate answer concerning the corrosion current density. Artificial neural networks (ANN) are massively parallel distributed processors that have a natural propensity for storing experiential knowledge and making it available for use [<xref ref-type="bibr" rid="CR27">27</xref>, <xref ref-type="bibr" rid="CR28">28</xref>]. Nowadays, the ANN are well known and over the last few years have emerged as a powerful device that could be used in many engineering applications such for the prediction of corrosion of steel reinforcement in concrete [<xref ref-type="bibr" rid="CR29">29</xref>&#x02013;<xref ref-type="bibr" rid="CR35">35</xref>]. ANN has been used, for example, as the prediction of chloride permeability of concretes with obtaining an empirical model having high capability of estimation of permeability for both their experimental data and the ones obtained from the studies of other researchers [<xref ref-type="bibr" rid="CR36">36</xref>]. It has been also concluded that ANN model has a theoretical value in the prediction of the corrosion current rate of steel in concrete using corrosion current density without the need for a connection to the steel reinforcement [<xref ref-type="bibr" rid="CR23">23</xref>].</p></sec><sec id="Sec2"><title>Research significance</title><p>In previous research, backpropagation has been used for optimizing ANN. Although backpropagation has unquestionably been a major factor for the success of past ANN applications, it is plagued with inconsistent and unpredictable performances [<xref ref-type="bibr" rid="CR37">37</xref>, <xref ref-type="bibr" rid="CR38">38</xref>]. Today, the new techniques exploiting ANN model, based on optimization algorithms, are being widely used in engineering fields [<xref ref-type="bibr" rid="CR39">39</xref>]. The most popular seems to be imperialist competitive algorithms (ICA) and genetic algorithms (GA). The total number of papers related to corrosion of reinforced concrete published in Science Direct database increased from 42 in 2010 to 78 in 2013 (Fig.&#x000a0;<xref rid="Fig3" ref-type="fig">3</xref>). The number of papers related to corrosion by using neural networks is on the same level since 2010. There were few attempts to use GA for the corrosion modeling. It is hard to find in the existed literature the application of ICA for corrosion current density prediction in steel reinforced concrete. It is proper to note that the applications of ICA for solving various engineering problems increased from 4 applications in 2010 into 39 in 2013.<fig id="Fig3"><label>Fig.&#x000a0;3</label><caption><p>Number of papers related to corrosion of reinforced concrete published in Science Direct database since 2010</p></caption><graphic xlink:href="521_2014_1645_Fig3_HTML" id="MO4"/></fig> </p><p>ICA is a randomized population algorithm inspired from of the human political&#x02013;social evolution [<xref ref-type="bibr" rid="CR40">40</xref>&#x02013;<xref ref-type="bibr" rid="CR45">45</xref>]. A number of colonial countries along with their colonies try to find a general optimal point in solving optimization problem. Different methods have been introduced to solve optimization problems. Some find the cost function optimum point iteratively, based on the gradient. In spite of the high rate of these methods, there is still the problem of falling into the local optimum trap [<xref ref-type="bibr" rid="CR46">46</xref>]. The main objective of this research is using ICA to optimize the weights of ANN as a new optimization algorithm in determining steel corrosion in concrete. The advantages of this method [<xref ref-type="bibr" rid="CR46">46</xref>] have been listed below:<list list-type="bullet"><list-item><p>The innovation of the ICA basic idea as the first optimization algorithm based on socio-political process;</p></list-item><list-item><p>Ability to aligning and even higher optimization in comparing various optimization algorithms facing various optimization problems;</p></list-item><list-item><p>Finding the optimal solution speed.</p></list-item></list> </p><p>In the last few years, there were few attempts to apply ICA for the engineering problems modeling like for the prediction of soil compaction in soil bin facility [<xref ref-type="bibr" rid="CR47">47</xref>], for the prediction oil flow rate of the reservoir [<xref ref-type="bibr" rid="CR48">48</xref>], or for optimum cost design of cantilever retaining walls [<xref ref-type="bibr" rid="CR49">49</xref>].</p><p>GA is inspired by nature. Nature evolution or Darwin&#x02019;s theory is the basis of its formation in which the bests have the right to survival. In this method, chromosomes with high competence have a higher chance to repeat in the selected population in the replication process. This takes place by the selection process. Various methods have been proposed, and the wheel method is the most famous one. Also, the elitist selection is used to determine how many of the most graceful persons were transferred to the next generation, unchangeably [<xref ref-type="bibr" rid="CR50">50</xref>, <xref ref-type="bibr" rid="CR51">51</xref>]. In the last few years, there were few successful attempts to apply GA for modeling the concrete structures such as finding optimum reliability-based inspection plans for the service life of the hypothetical bridge deck [<xref ref-type="bibr" rid="CR52">52</xref>], finding optimal placements of control devices and sensors in seismically excited civil structures [<xref ref-type="bibr" rid="CR53">53</xref>] or active control of high-rise buildings [<xref ref-type="bibr" rid="CR54">54</xref>].</p><p>Considering the above in this paper, steel corrosion is determined and predicted in concrete using ICA&#x02013;ANN model. Thus, the different aspects of the network will be checked with 2, 3, and 4 inputs; temperature, AC resistivity over the steel bar, AC resistivity remote from the steel bar, and the DC resistivity over the steel bar are as input parameters in the ANN, and corrosion current density is considered as an output parameter. The model ICA&#x02013;ANN accuracy is evaluated compared with a GA, in three phases of training, testing, and prediction.</p></sec><sec id="Sec3"><title>Imperialist competitive algorithm (ICA)</title><p>ICA is a method in the field of evolutionary computing that seeks to find the optimal solutions in various optimization issues, and it offers an algorithm for solving mathematical optimization problems. ICA forms an initial set of possible responses with a particular process improving the initial responses (countries) gradually and providing the appropriate response of optimization problem (the ideal country). The foundations of this algorithm are consisted of assimilation policies, imperialistic competition (IC), and revolution [<xref ref-type="bibr" rid="CR46">46</xref>, <xref ref-type="bibr" rid="CR55">55</xref>]. ICA begins with random initial population, and some of the best elements of the population are selected as imperialists. The remaining population is considered as a colony. Depending on the colonial power, the colonies are absorbed by imperialists with a special process. The total power of the empire depends on both constituent parts, i.e., the imperialist country (as the central core) and its colonies [<xref ref-type="bibr" rid="CR46">46</xref>]. This dependence has been modeled as the sum of the power of the imperialist country, plus a percentage of its average colonial power. With the formation of the early empires, IC between them will begin. Each empire unable to compete, succeed, and increase its power (or at least prevents reduction of its influence) will be removed in the IC. So the survival of an empire depends on its ability to absorb the colonial empires of the rivals and bringing them under control. As a result, during the IC, gradually, the power of larger empires will increase, and weaker empires will be removed. In Fig.&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>a, the flowchart of ICA process has been presented [<xref ref-type="bibr" rid="CR46">46</xref>].<fig id="Fig4"><label>Fig.&#x000a0;4</label><caption><p>Flowcharts of: <bold>a</bold> ICA process [<xref ref-type="bibr" rid="CR46">46</xref>], <bold>b</bold> GA process [<xref ref-type="bibr" rid="CR60">60</xref>]</p></caption><graphic xlink:href="521_2014_1645_Fig4_HTML" id="MO5"/></fig> </p><p>Policy of assimilation was performed with the aim of analyzing the culture and social structure of colonies in culture of central government. Colonial countries, to increase their influence, began to build developmental infrastructures. In line with this policy, the colony will move as x units to the line connecting the colony to imperialist and will be drawn to a new position. X is a random number with uniform distribution. The distance between the imperialist and the colonized is shown by &#x0201c;<italic>d,</italic>&#x0201d; and the value of parameter <italic>d</italic> is shown by [<xref ref-type="bibr" rid="CR46">46</xref>] in Eq. (<xref rid="Equ2" ref-type="">2</xref>):<disp-formula id="Equ2"><label>2</label><alternatives><tex-math id="M3">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$x\sim \, U\left( {0,\beta *d} \right)$$\end{document}</tex-math><mml:math id="M4" display="block"><mml:mrow><mml:mi>x</mml:mi><mml:mo>&#x0223c;</mml:mo><mml:mspace width="0.166667em"/><mml:mi>U</mml:mi><mml:mfenced close=")" open="(" separators=""><mml:mrow><mml:mn>0</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="italic">&#x003b2;</mml:mi><mml:mrow/><mml:mo>&#x02217;</mml:mo><mml:mi>d</mml:mi></mml:mrow></mml:mfenced></mml:mrow></mml:math></alternatives></disp-formula>where <italic>&#x003b2;</italic> is a number greater than one and close to two. A good choice can be <italic>&#x003b2;</italic>&#x000a0;=&#x000a0;2. <italic>&#x003b2;</italic>&#x000a0;&#x02265;&#x000a0;1 coefficient causes the colony country to near the imperialist country from different directions during moving to it. In addition to this move, a small angular deviation is added to the path with a uniform distribution [<xref ref-type="bibr" rid="CR46">46</xref>]. Every empire unable to increase its strength and lose its competitiveness power will be removed in the IC. This process will be done gradually. This means that over time, weakened empires will lose their colonies and more powerful empire, will conquer them and increase their power&#x02014;the empire being removed is the weakest one. Thus, when repeating the algorithm, one or more of the weakest colonies of the weakest empire will be taken, and to take possession of the considered colonies, a competition among all empires will be created. The colonies, not necessarily will be seized by the most powerful empire, but the stronger empires, are more likely to acquire them.</p></sec><sec id="Sec4"><title>Genetic algorithm (GA)</title><p>In this method, chromosomes with high competence have a higher chance to repeat in the selected population in the replication process. This takes place by the selection process [<xref ref-type="bibr" rid="CR50">50</xref>]. After completion of the selection process, the operator is applied on the selected direction for reproduction. In the transplant process, with a constant transplant rate, a random number is generated for each chromosome. If the generated random number is less than the transplant rate, this chromosome is selected to intercourse the next chromosome with the above conditions. In this method, uniform transplantation has been used among different types of transplantation. Then, the mutation operator is applied [<xref ref-type="bibr" rid="CR50">50</xref>]. The aim of this work is to create more dispersion in the range of design space. In the mutation process, a random number is generated, with a constant mutation rate, for all bits of the chromosomes. If the generated random number is smaller than the mutation rate, the value of that bit changes, i.e., the value of zero becomes one and vice versa. The basic operators of natural genetics are reproduction, crossover, and mutation. The basic operators of natural genetics are reproduction, crossover, and mutation [<xref ref-type="bibr" rid="CR56">56</xref>, <xref ref-type="bibr" rid="CR57">57</xref>]. The GA can be expressed as in Fig.&#x000a0;<xref rid="Fig4" ref-type="fig">4</xref>b. GA ends when certain criteria such as certain number of generation or the average standard deviation performance of individuals are satisfied [<xref ref-type="bibr" rid="CR50">50</xref>, <xref ref-type="bibr" rid="CR58">58</xref>, <xref ref-type="bibr" rid="CR59">59</xref>].</p></sec><sec id="Sec5"><title>Experimental details and database</title><p>The data used for development of the models were obtained from the past experimental researchers [<xref ref-type="bibr" rid="CR23">23</xref>, <xref ref-type="bibr" rid="CR61">61</xref>]. Exemplary steel corrosion data have been presented in Table&#x000a0;<xref rid="Tab2" ref-type="table">2</xref>. Of 68 data patterns, 80&#x000a0;% of samples (54 patterns) were used for training and 20&#x000a0;% of the selected samples (14 patterns) were used to test the network.<table-wrap id="Tab2"><label>Table&#x000a0;2</label><caption><p>Exemplary steel corrosion data</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left">No.</th><th align="left"> <italic>T</italic> (&#x000b0;C)</th><th align="left"> <italic>&#x003c1;</italic> _AC,bar (k&#x003a9;cm)</th><th align="left"> <italic>&#x003c1;</italic> _AC,coc (k&#x003a9;cm)</th><th align="left"> <italic>&#x003c1;</italic> _DC (k&#x003a9;cm)</th><th align="left"> <italic>i</italic> _corr (&#x003bc;A/cm²)</th></tr></thead><tbody><tr><td align="left" colspan="6"> <italic>Training</italic> </td></tr><tr><td align="left">1</td><td char="." align="char">21.00</td><td char="." align="char">19.31</td><td char="." align="char">22.27</td><td char="." align="char">21.81</td><td char="." align="char">0.422</td></tr><tr><td align="left">2</td><td char="." align="char">20.80</td><td char="." align="char">19.33</td><td char="." align="char">22.28</td><td char="." align="char">21.83</td><td char="." align="char">0.423</td></tr><tr><td align="left">3</td><td char="." align="char">20.50</td><td char="." align="char">19.34</td><td char="." align="char">22.30</td><td char="." align="char">21.85</td><td char="." align="char">0.421</td></tr><tr><td align="left">4</td><td char="." align="char">20.10</td><td char="." align="char">19.35</td><td char="." align="char">22.31</td><td char="." align="char">21.91</td><td char="." align="char">0.439</td></tr><tr><td align="left">5</td><td char="." align="char">19.80</td><td char="." align="char">19.36</td><td char="." align="char">22.32</td><td char="." align="char">21.92</td><td char="." align="char">0.439</td></tr><tr><td align="left">6</td><td char="." align="char">19.50</td><td char="." align="char">19.36</td><td char="." align="char">22.33</td><td char="." align="char">21.94</td><td char="." align="char">0.456</td></tr><tr><td align="left">7</td><td char="." align="char">19.20</td><td char="." align="char">19.37</td><td char="." align="char">22.36</td><td char="." align="char">21.96</td><td char="." align="char">0.466</td></tr><tr><td align="left">8</td><td char="." align="char">19.00</td><td char="." align="char">19.38</td><td char="." align="char">22.38</td><td char="." align="char">21.98</td><td char="." align="char">0.476</td></tr><tr><td align="left">9</td><td char="." align="char">20.90</td><td char="." align="char">19.24</td><td char="." align="char">22.09</td><td char="." align="char">21.62</td><td char="." align="char">0.373</td></tr><tr><td align="left">.</td><td align="left">.</td><td align="left">.</td><td align="left">.</td><td align="left">.</td><td align="left">.</td></tr><tr><td align="left">.</td><td align="left">.</td><td align="left">.</td><td align="left">.</td><td align="left">.</td><td align="left">.</td></tr><tr><td align="left">.</td><td align="left">.</td><td align="left">.</td><td align="left">.</td><td align="left">.</td><td align="left">.</td></tr><tr><td align="left">54</td><td char="." align="char">20.50</td><td char="." align="char">19.36</td><td char="." align="char">22.31</td><td char="." align="char">21.84</td><td char="." align="char">0.429</td></tr><tr><td align="left" colspan="6"> <italic>Testing</italic> </td></tr><tr><td align="left">1</td><td char="." align="char">20.10</td><td char="." align="char">19.34</td><td char="." align="char">22.33</td><td char="." align="char">21.91</td><td char="." align="char">0.434</td></tr><tr><td align="left">.</td><td align="left">.</td><td align="left">.</td><td align="left">.</td><td align="left">.</td><td align="left">.</td></tr><tr><td align="left">.</td><td align="left">.</td><td align="left">.</td><td align="left">.</td><td align="left">.</td><td align="left">.</td></tr><tr><td align="left">.</td><td align="left">.</td><td align="left">.</td><td align="left">.</td><td align="left">.</td><td align="left">.</td></tr><tr><td align="left">14</td><td char="." align="char">19.10</td><td char="." align="char">19.30</td><td char="." align="char">22.22</td><td char="." align="char">21.77</td><td char="." align="char">0.421</td></tr></tbody></table></table-wrap> </p><p>Specimens sized 400&#x000a0;mm&#x000a0;&#x000d7;&#x000a0;300&#x000a0;mm&#x000a0;&#x000d7;&#x000a0;100&#x000a0;mm were available, with each specimen containing a single, short steel bar 30&#x000a0;mm in diameter, made from steel class A-III grade 34GS. The slabs were made from concrete class C 20/25 and from Portland cement CEM I 42.5R and aggregate of maximum size&#x02014;5&#x000a0;mm. Since the relative humidity of concrete has a significant influence on the concrete resistivity, the specimens were stored in a laboratory under constant relative concrete humidity conditions of 65&#x000a0;&#x000b1;&#x000a0;1&#x000a0;% up to the time of the tests. A view of the concrete resistivity measurement system, laboratory stand, and resistivity measurement locations on concrete specimen is presented in Fig.&#x000a0;<xref rid="Fig5" ref-type="fig">5</xref>. The study was carried out based on a database of concrete slabs of two different conditions with high and moderate corrosion rates, respectively, with laboratory-induced corrosion as presented previously [<xref ref-type="bibr" rid="CR61">61</xref>].<fig id="Fig5"><label>Fig.&#x000a0;5</label><caption><p>View of: <bold>a</bold> concrete resistivity measurement system, <bold>b</bold> laboratory stand, <bold>c</bold> resistivity measurement locations on concrete specimen [<xref ref-type="bibr" rid="CR61">61</xref>]</p></caption><graphic xlink:href="521_2014_1645_Fig5_HTML" id="MO7"/></fig> </p><p>To include all the input parameters in a numerical range as the inputs of ANN to provide more accurate and suitable results, all data, according to formula (<xref rid="Equ3" ref-type="">3</xref>), will be normalized:<disp-formula id="Equ3"><label>3</label><alternatives><tex-math id="M5">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$xN = \left( {x - {\text{Min}}X} \right)/ \, \left( {{\text{Max}}X - {\text{Min}}X} \right) \times 2- 1$$\end{document}</tex-math><mml:math id="M6" display="block"><mml:mrow><mml:mi>x</mml:mi><mml:mi>N</mml:mi><mml:mo>=</mml:mo><mml:mfenced close=")" open="(" separators=""><mml:mrow><mml:mi>x</mml:mi><mml:mo>-</mml:mo><mml:mtext>Min</mml:mtext><mml:mi>X</mml:mi></mml:mrow></mml:mfenced><mml:mo stretchy="false">/</mml:mo><mml:mspace width="0.166667em"/><mml:mfenced close=")" open="(" separators=""><mml:mrow><mml:mtext>Max</mml:mtext><mml:mi>X</mml:mi><mml:mo>-</mml:mo><mml:mtext>Min</mml:mtext><mml:mi>X</mml:mi></mml:mrow></mml:mfenced><mml:mo>&#x000d7;</mml:mo><mml:mn>2</mml:mn><mml:mo>-</mml:mo><mml:mn>1</mml:mn></mml:mrow></mml:math></alternatives></disp-formula>where <italic>xN</italic> are the normalized input data, xs are the input data, Min<italic>X</italic> is the minimum of all data, and Max<italic>X</italic> is the maximum of all the data.</p><p>Moreover, in the output parameter, the formula (<xref rid="Equ4" ref-type="">4</xref>) will be used in normalization:<disp-formula id="Equ4"><label>4</label><alternatives><tex-math id="M7">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$yN = \left( {y - {\text{Min}}Y} \right)/\left( {{\text{Max}}Y - {\text{Min}}Y} \right) \times 2- 1$$\end{document}</tex-math><mml:math id="M8" display="block"><mml:mrow><mml:mi>y</mml:mi><mml:mi>N</mml:mi><mml:mo>=</mml:mo><mml:mfenced close=")" open="(" separators=""><mml:mrow><mml:mi>y</mml:mi><mml:mo>-</mml:mo><mml:mtext>Min</mml:mtext><mml:mi>Y</mml:mi></mml:mrow></mml:mfenced><mml:mo stretchy="false">/</mml:mo><mml:mfenced close=")" open="(" separators=""><mml:mrow><mml:mtext>Max</mml:mtext><mml:mi>Y</mml:mi><mml:mo>-</mml:mo><mml:mtext>Min</mml:mtext><mml:mi>Y</mml:mi></mml:mrow></mml:mfenced><mml:mo>&#x000d7;</mml:mo><mml:mn>2</mml:mn><mml:mo>-</mml:mo><mml:mn>1</mml:mn></mml:mrow></mml:math></alternatives></disp-formula>where, <italic>yN</italic> are the normalized output data, <italic>y</italic>s are output data; Min<italic>Y</italic> are minimum of all data, Max<italic>Y</italic> are maximum of all data. Therefore, all normalized data should be located within numerical distance [&#x02212;1, +1].</p><p>ANN used in this research is called &#x0201c;Feed Forward&#x0201d; with the input parameters of temperature, AC resistivity over the steel bar, AC resistivity remote from the steel bar and the DC resistivity over the steel bar, and the output parameter corrosion current density; the network is shown in Fig.&#x000a0;<xref rid="Fig6" ref-type="fig">6</xref>.<fig id="Fig6"><label>Fig.&#x000a0;6</label><caption><p>ANN with a hidden layer 4-5-4-1</p></caption><graphic xlink:href="521_2014_1645_Fig6_HTML" id="MO10"/></fig> </p></sec><sec id="Sec6"><title>Results</title><p>The hidden layer nodes numbers were determined through using [<xref ref-type="bibr" rid="CR62">62</xref>] the empirical formula (<xref rid="Equ5" ref-type="">5</xref>).<disp-formula id="Equ5"><label>5</label><alternatives><tex-math id="M9">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$N_{H} \le 2N_{I} + 1$$\end{document}</tex-math><mml:math id="M10" display="block"><mml:mrow><mml:msub><mml:mi>N</mml:mi><mml:mi>H</mml:mi></mml:msub><mml:mo>&#x02264;</mml:mo><mml:mn>2</mml:mn><mml:msub><mml:mi>N</mml:mi><mml:mi>I</mml:mi></mml:msub><mml:mo>+</mml:mo><mml:mn>1</mml:mn></mml:mrow></mml:math></alternatives></disp-formula>where <italic>N</italic> _{<italic>H</italic>} is the maximum number of nodes in the hidden layers, and <italic>N</italic> _{<italic>I</italic>} is the number of inputs. Considering that the number of effective inputs obtained is equal to 4, the maximum number of nodes in the hidden layer will be equal to 9. The ICA is used to determine the ANN model weights.</p><sec id="Sec7"><title>Research conduction</title><p>Table&#x000a0;<xref rid="Tab3" ref-type="table">3</xref> summarizes each pattern optima structure as well as the ICA features. Also, in Table&#x000a0;<xref rid="Tab4" ref-type="table">4</xref>, analytical results obtained from the training and testing patterns with optimized structure in Table&#x000a0;<xref rid="Tab3" ref-type="table">3</xref> have been shown. In Table&#x000a0;<xref rid="Tab5" ref-type="table">5</xref>, three statistical parameters: mean absolute error (MAE), root mean squared deviations (RMSD), and root mean squared error (RMSE) have been applied.<table-wrap id="Tab3"><label>Table&#x000a0;3</label><caption><p>Optimized structure of ICA&#x02013;ANN model</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="2">No.</th><th align="left" rowspan="2">Models&#x02019; name</th><th align="left" colspan="5">ANN features</th><th align="left" colspan="3">Utilized initialization parameters in ICA</th></tr><tr><th align="left">Number of input</th><th align="left">Number of output</th><th align="left">Number of hidden layer</th><th align="left">Number of nodes in hidden layer</th><th align="left">Transfer function</th><th align="left">Number of country</th><th align="left">Number of imperialist</th><th align="left">Number of decade</th></tr></thead><tbody><tr><td align="left">1</td><td align="left">50GEN_2IN</td><td char="." align="char">2</td><td char="." align="char">1</td><td char="." align="char">2</td><td align="left">3-2</td><td align="left">tansig</td><td char="." align="char">300</td><td char="." align="char">20</td><td char="." align="char">50</td></tr><tr><td align="left">2</td><td align="left">50GEN_3IN</td><td char="." align="char">3</td><td char="." align="char">1</td><td char="." align="char">2</td><td align="left">4-3</td><td align="left">tribas</td><td char="." align="char">200</td><td char="." align="char">25</td><td char="." align="char">50</td></tr><tr><td align="left">3</td><td align="left">50GEN_4IN</td><td char="." align="char">4</td><td char="." align="char">1</td><td char="." align="char">1</td><td char="." align="char">9</td><td align="left">satlins</td><td char="." align="char">400</td><td char="." align="char">40</td><td char="." align="char">50</td></tr><tr><td align="left">4</td><td align="left">100GEN_2IN</td><td char="." align="char">2</td><td char="." align="char">1</td><td char="." align="char">1</td><td char="." align="char">5</td><td align="left">satlins</td><td char="." align="char">300</td><td char="." align="char">30</td><td char="." align="char">100</td></tr><tr><td align="left">5</td><td align="left">100GEN_3IN</td><td char="." align="char">3</td><td char="." align="char">1</td><td char="." align="char">1</td><td char="." align="char">7</td><td align="left">tansig</td><td char="." align="char">400</td><td char="." align="char">40</td><td char="." align="char">100</td></tr><tr><td align="left">6</td><td align="left">100GEN_4IN</td><td char="." align="char">4</td><td char="." align="char">1</td><td char="." align="char">2</td><td align="left">6-3</td><td align="left">tansig</td><td char="." align="char">500</td><td char="." align="char">50</td><td char="." align="char">100</td></tr><tr><td align="left">7</td><td align="left">150GEN_2IN</td><td char="." align="char">2</td><td char="." align="char">1</td><td char="." align="char">1</td><td char="." align="char">5</td><td align="left">poslin</td><td char="." align="char">400</td><td char="." align="char">20</td><td char="." align="char">150</td></tr><tr><td align="left">8</td><td align="left">150GEN_3IN</td><td char="." align="char">3</td><td char="." align="char">1</td><td char="." align="char">1</td><td char="." align="char">7</td><td align="left">hardlims</td><td char="." align="char">250</td><td char="." align="char">25</td><td char="." align="char">150</td></tr><tr><td align="left">9</td><td align="left">150GEN_4IN</td><td char="." align="char">4</td><td char="." align="char">1</td><td char="." align="char">1</td><td char="." align="char">9</td><td align="left">purelin</td><td char="." align="char">450</td><td char="." align="char">45</td><td char="." align="char">150</td></tr><tr><td align="left">10</td><td align="left">200GEN_2IN</td><td char="." align="char">2</td><td char="." align="char">1</td><td char="." align="char">1</td><td char="." align="char">4</td><td align="left">tribas</td><td char="." align="char">600</td><td char="." align="char">60</td><td char="." align="char">200</td></tr><tr><td align="left">11</td><td align="left">200GEN_3IN</td><td char="." align="char">3</td><td char="." align="char">1</td><td char="." align="char">1</td><td char="." align="char">7</td><td align="left">satlins</td><td char="." align="char">250</td><td char="." align="char">25</td><td char="." align="char">200</td></tr><tr><td align="left">12</td><td align="left">200GEN_4IN</td><td char="." align="char">4</td><td char="." align="char">1</td><td char="." align="char">3</td><td align="left">3-3-3</td><td align="left">poslin</td><td char="." align="char">500</td><td char="." align="char">50</td><td char="." align="char">200</td></tr><tr><td align="left">13</td><td align="left">250GEN_2IN</td><td char="." align="char">2</td><td char="." align="char">1</td><td char="." align="char">1</td><td char="." align="char">5</td><td align="left">logsig</td><td char="." align="char">500</td><td char="." align="char">50</td><td char="." align="char">250</td></tr><tr><td align="left">14</td><td align="left">250GEN_3IN</td><td char="." align="char">3</td><td char="." align="char">1</td><td char="." align="char">2</td><td align="left">4-3</td><td align="left">radbas</td><td char="." align="char">250</td><td char="." align="char">25</td><td char="." align="char">250</td></tr><tr><td align="left">15</td><td align="left">250GEN_4IN</td><td char="." align="char">5</td><td char="." align="char">1</td><td char="." align="char">2</td><td align="left">5-4</td><td align="left">tansig</td><td char="." align="char">400</td><td char="." align="char">50</td><td char="." align="char">250</td></tr></tbody></table></table-wrap> <table-wrap id="Tab4"><label>Table&#x000a0;4</label><caption><p>Results from of testing and training of ICA&#x02013;ANN model</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="2">No</th><th align="left" rowspan="2">Model</th><th align="left" colspan="2">Best fitting line in testing phase</th><th align="left" colspan="2">Best fitting line in training phase</th><th align="left" colspan="3">Results</th></tr><tr><th align="left">Equation</th><th align="left"> <italic>R</italic> ² </th><th align="left">Equation</th><th align="left"> <italic>R</italic> ² </th><th align="left">MSE test</th><th align="left">MSE train</th><th align="left">Best Cost</th></tr></thead><tbody><tr><td align="left">1</td><td align="left">50GEN_2IN</td><td align="left"> <italic>y</italic>&#x000a0;=&#x000a0;0.4714<italic>x</italic>&#x000a0;&#x02212;&#x000a0;0.1684</td><td char="." align="char">0.5487</td><td align="left"> <italic>y</italic>&#x000a0;=&#x000a0;0.7753<italic>x</italic>&#x000a0;&#x02212;&#x000a0;0.0179</td><td char="." align="char">0.8043</td><td char="." align="char">0.1892</td><td char="." align="char">0.0482</td><td char="." align="char">0.0482</td></tr><tr><td align="left">2</td><td align="left">50GEN_3IN</td><td align="left"> <italic>y</italic>&#x000a0;=&#x000a0;0.1945<italic>x</italic>&#x000a0;+&#x000a0;0.1017</td><td char="." align="char">0.3751</td><td align="left"> <italic>y</italic>&#x000a0;=&#x000a0;0.3443<italic>x</italic>&#x000a0;+&#x000a0;0.174</td><td char="." align="char">0.5533</td><td char="." align="char">0.2709</td><td char="." align="char">0.2006</td><td char="." align="char">0.2006</td></tr><tr><td align="left">3</td><td align="left">50GEN_4IN</td><td align="left"> <italic>y</italic>&#x000a0;=&#x000a0;0.5445<italic>x</italic>&#x000a0;&#x02212;&#x000a0;0.2061</td><td char="." align="char">0.4666</td><td align="left"> <italic>y</italic>&#x000a0;=&#x000a0;0.9307<italic>x</italic>&#x000a0;&#x02212;&#x000a0;0.0289</td><td char="." align="char">0.7533</td><td char="." align="char">0.2332</td><td char="." align="char">0.0704</td><td char="." align="char">0.0704</td></tr><tr><td align="left">4</td><td align="left">100GEN_2IN</td><td align="left">0.4734<italic>x</italic>&#x000a0;&#x02212;&#x000a0;0.1947</td><td char="." align="char">0.5391</td><td align="left"> <italic>y</italic>&#x000a0;=&#x000a0;0.4734<italic>x</italic>&#x000a0;&#x02212;&#x000a0;0.1947</td><td char="." align="char">0.5391</td><td char="." align="char">0.2004</td><td char="." align="char">0.0452</td><td char="." align="char">0.0452</td></tr><tr><td align="left">5</td><td align="left">100GEN_3IN</td><td align="left"> <italic>y</italic>&#x000a0;=&#x000a0;0.4282<italic>x</italic>&#x000a0;&#x02212;&#x000a0;0.1636</td><td char="." align="char">0.4646</td><td align="left"> <italic>y</italic>&#x000a0;=&#x000a0;0.7518<italic>x</italic>&#x000a0;&#x02212;&#x000a0;0.0277</td><td char="." align="char">0.7866</td><td char="." align="char">0.2147</td><td char="." align="char">0.0523</td><td char="." align="char">0.0523</td></tr><tr><td align="left">6</td><td align="left">100GEN_4IN</td><td align="left"> <italic>y</italic>&#x000a0;=&#x000a0;0.4578<italic>x</italic>&#x000a0;&#x02212;&#x000a0;0.1857</td><td char="." align="char">0.5922</td><td align="left"> <italic>y</italic>&#x000a0;=&#x000a0;0.8676<italic>x</italic>&#x000a0;&#x02212;&#x000a0;0.0221</td><td char="." align="char">0.8752</td><td char="." align="char">0.1861</td><td char="." align="char">0.0303</td><td char="." align="char">0.0303</td></tr><tr><td align="left">7</td><td align="left">150GEN_2IN</td><td align="left"> <italic>y</italic>&#x000a0;=&#x000a0;0.1746<italic>x</italic>&#x000a0;+&#x000a0;0.0918</td><td char="." align="char">0.3745</td><td align="left"> <italic>y</italic>&#x000a0;=&#x000a0;0.3095<italic>x</italic>&#x000a0;+&#x000a0;0.163</td><td char="." align="char">0.5768</td><td char="." align="char">0.2760</td><td char="." align="char">0.2027</td><td char="." align="char">0.2027</td></tr><tr><td align="left">8</td><td align="left">150GEN_3IN</td><td align="left"> <italic>y</italic>&#x000a0;=&#x000a0;0.7651<italic>x</italic>&#x000a0;+&#x000a0;0.0221</td><td char="." align="char">0.21</td><td align="left"> <italic>y</italic>&#x000a0;=&#x000a0;1.4463<italic>x</italic>&#x000a0;+&#x000a0;0.2862</td><td char="." align="char">0.5108</td><td char="." align="char">0.8106</td><td char="." align="char">0.5836</td><td char="." align="char">0.5836</td></tr><tr><td align="left">9</td><td align="left">150GEN_4IN</td><td align="left"> <italic>y</italic>&#x000a0;=&#x000a0;0.4931<italic>x</italic>&#x000a0;&#x02212;&#x000a0;0.1985</td><td char="." align="char">0.5176</td><td align="left"> <italic>y</italic>&#x000a0;=&#x000a0;0.8187<italic>x</italic>&#x000a0;&#x02212;&#x000a0;0.0323</td><td char="." align="char">0.8231</td><td char="." align="char">0.2072</td><td char="." align="char">0.0430</td><td char="." align="char">0.0430</td></tr><tr><td align="left">10</td><td align="left">200GEN_2IN</td><td align="left"> <italic>y</italic>&#x000a0;=&#x000a0;0.1761<italic>x</italic>&#x000a0;+&#x000a0;0.0998</td><td char="." align="char">0.3862</td><td align="left"> <italic>y</italic>&#x000a0;=&#x000a0;0.2953<italic>x</italic>&#x000a0;+&#x000a0;0.1641</td><td char="." align="char">0.6067</td><td char="." align="char">0.2764</td><td char="." align="char">0.2059</td><td char="." align="char">0.2059</td></tr><tr><td align="left">11</td><td align="left">200GEN_3IN</td><td align="left"> <italic>y</italic>&#x000a0;=&#x000a0;0.4827<italic>x</italic>&#x000a0;&#x02212;&#x000a0;0.1489</td><td char="." align="char">0.5297</td><td align="left"> <italic>y</italic>&#x000a0;=&#x000a0;0.7894<italic>x</italic>&#x000a0;&#x02212;&#x000a0;0.0063</td><td char="." align="char">0.8273</td><td char="." align="char">0.1882</td><td char="." align="char">0.0430</td><td char="." align="char">0.0430</td></tr><tr><td align="left">12</td><td align="left">200GEN_4IN</td><td align="left"> <italic>y</italic>&#x000a0;=&#x000a0;0.1397<italic>x</italic>&#x000a0;+&#x000a0;0.0778</td><td char="." align="char">0.3724</td><td align="left"> <italic>y</italic>&#x000a0;=&#x000a0;0.2625<italic>x</italic>&#x000a0;+&#x000a0;0.1457</td><td char="." align="char">0.5857</td><td char="." align="char">0.2878</td><td char="." align="char">0.2084</td><td char="." align="char">0.2084</td></tr><tr><td align="left">13</td><td align="left">250GEN_2IN</td><td align="left"> <italic>y</italic>&#x000a0;=&#x000a0;0.0752<italic>x</italic>&#x000a0;+&#x000a0;0.091</td><td char="." align="char">0.4603</td><td align="left"> <italic>y</italic>&#x000a0;=&#x000a0;0.1283<italic>x</italic>&#x000a0;+&#x000a0;0.123</td><td char="." align="char">0.762</td><td char="." align="char">0.3230</td><td char="." align="char">0.2487</td><td char="." align="char">0.2487</td></tr><tr><td align="left">14</td><td align="left">250GEN_3IN</td><td align="left"> <italic>y</italic>&#x000a0;=&#x000a0;0.2227<italic>x</italic>&#x000a0;+&#x000a0;0.1323</td><td char="." align="char">0.3683</td><td align="left"> <italic>y</italic>&#x000a0;=&#x000a0;0.3235<italic>x</italic>&#x000a0;+&#x000a0;0.1799</td><td char="." align="char">0.5516</td><td char="." align="char">0.2711</td><td char="." align="char">0.2097</td><td char="." align="char">0.2097</td></tr><tr><td align="left">15</td><td align="left">250GEN_4IN</td><td align="left"> <italic>y</italic>&#x000a0;=&#x000a0;0.7337<italic>x</italic>&#x000a0;&#x02212;&#x000a0;0.0618</td><td char="." align="char">0.8019</td><td align="left"> <italic>y</italic>&#x000a0;=&#x000a0;0.8877<italic>x</italic>&#x000a0;&#x02212;&#x000a0;0.0042</td><td char="." align="char">0.9045</td><td char="." align="char">0.1572</td><td char="." align="char">0.0234</td><td char="." align="char">0.0234</td></tr></tbody></table></table-wrap> <table-wrap id="Tab5"><label>Table&#x000a0;5</label><caption><p>Statistical results of optimized ICA&#x02013;ANN</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="2">No.</th><th align="left" rowspan="2">Model</th><th align="left" colspan="2">MAE</th><th align="left" colspan="2">RMSE</th><th align="left" colspan="2">RMSD</th></tr><tr><th align="left">Train</th><th align="left">Test</th><th align="left">Train</th><th align="left">Test</th><th align="left">Train</th><th align="left">Test</th></tr></thead><tbody><tr><td align="left">1</td><td align="left">50GEN_2IN</td><td char="." align="char">0.155</td><td char="." align="char">0.275</td><td char="." align="char">0.048</td><td char="." align="char">0.189</td><td char="." align="char">0.359</td><td char="." align="char">0.460</td></tr><tr><td align="left">2</td><td align="left">50GEN_3IN</td><td char="." align="char">0.309</td><td char="." align="char">0.394</td><td char="." align="char">0.201</td><td char="." align="char">0.271</td><td char="." align="char">0.475</td><td char="." align="char">0.559</td></tr><tr><td align="left">3</td><td align="left">50GEN_4IN</td><td char="." align="char">0.206</td><td char="." align="char">0.271</td><td char="." align="char">0.070</td><td char="." align="char">0.233</td><td char="." align="char">0.417</td><td char="." align="char">0.420</td></tr><tr><td align="left">4</td><td align="left">100GEN_2IN</td><td char="." align="char">0.410</td><td char="." align="char">0.490</td><td char="." align="char">0.264</td><td char="." align="char">0.360</td><td char="." align="char">0.586</td><td char="." align="char">0.635</td></tr><tr><td align="left">5</td><td align="left">100GEN_3IN</td><td char="." align="char">0.173</td><td char="." align="char">0.291</td><td char="." align="char">0.052</td><td char="." align="char">0.215</td><td char="." align="char">0.383</td><td char="." align="char">0.467</td></tr><tr><td align="left">6</td><td align="left">100GEN_4IN</td><td char="." align="char">0.410</td><td char="." align="char">0.490</td><td char="." align="char">0.264</td><td char="." align="char">0.360</td><td char="." align="char">0.586</td><td char="." align="char">0.635</td></tr><tr><td align="left">7</td><td align="left">150GEN_2IN</td><td char="." align="char">0.319</td><td char="." align="char">0.403</td><td char="." align="char">0.203</td><td char="." align="char">0.276</td><td char="." align="char">0.490</td><td char="." align="char">0.568</td></tr><tr><td align="left">8</td><td align="left">150GEN_3IN</td><td char="." align="char">0.660</td><td char="." align="char">0.726</td><td char="." align="char">0.584</td><td char="." align="char">0.811</td><td char="." align="char">0.765</td><td char="." align="char">0.772</td></tr><tr><td align="left">9</td><td align="left">150GEN_4IN</td><td char="." align="char">0.147</td><td char="." align="char">0.263</td><td char="." align="char">0.043</td><td char="." align="char">0.207</td><td char="." align="char">0.353</td><td char="." align="char">0.429</td></tr><tr><td align="left">10</td><td align="left">200GEN_2IN</td><td char="." align="char">0.331</td><td char="." align="char">0.410</td><td char="." align="char">0.206</td><td char="." align="char">0.276</td><td char="." align="char">0.513</td><td char="." align="char">0.584</td></tr><tr><td align="left">11</td><td align="left">200GEN_3IN</td><td char="." align="char">0.152</td><td char="." align="char">0.262</td><td char="." align="char">0.043</td><td char="." align="char">0.188</td><td char="." align="char">0.358</td><td char="." align="char">0.431</td></tr><tr><td align="left">12</td><td align="left">200GEN_4IN</td><td char="." align="char">0.335</td><td char="." align="char">0.421</td><td char="." align="char">0.208</td><td char="." align="char">0.288</td><td char="." align="char">0.512</td><td char="." align="char">0.588</td></tr><tr><td align="left">13</td><td align="left">250GEN_2IN</td><td char="." align="char">0.403</td><td char="." align="char">0.482</td><td char="." align="char">0.249</td><td char="." align="char">0.323</td><td char="." align="char">0.585</td><td char="." align="char">0.658</td></tr><tr><td align="left">14</td><td align="left">250GEN_3IN</td><td char="." align="char">0.320</td><td char="." align="char">0.388</td><td char="." align="char">0.210</td><td char="." align="char">0.271</td><td char="." align="char">0.486</td><td char="." align="char">0.558</td></tr><tr><td align="left">15</td><td align="left">250GEN_4IN</td><td char="." align="char">0.098</td><td char="." align="char">0.214</td><td char="." align="char">0.023</td><td char="." align="char">0.157</td><td char="." align="char">0.282</td><td char="." align="char">0.372</td></tr></tbody></table></table-wrap> </p><p>For models performance and determining the best model, the MSE train and MSE test criteria created by the models are compared with each other. According to the results shown, ANN model weights optimized by ICA with structure&#x000a0;4-5-4-1 and properties of 400 countries, 50 empires, and 250 iterations have been optimized, indicating the best results in the considered models.</p><p>As shown in Table&#x000a0;<xref rid="Tab4" ref-type="table">4</xref> in the model No. 15, the coefficient <italic>R</italic> ² for steel corrosion parameter in concrete in training and test phases is equal to 0.9045 and 0.8019, and also the slope of the straight line for this parameter is equal to 0.7337, 0.8877, which demonstrated the suitable accuracy of the model for modeling. Moreover, according to Table&#x000a0;<xref rid="Tab5" ref-type="table">5</xref>, MAE and RMSE and RMSD coefficients in both phases of training and testing of ANN with 4-5-4-1 structure and properties of 400 countries, 50 empires, and 250 iterations are lower than all models, which indicate less error of this network than other models. So ANN model under the title of 250GEN_4IN possesses higher accuracy than its peers.</p><p>The results of model 250GEN_4IN have been shown in Figs.&#x000a0;<xref rid="Fig7" ref-type="fig">7</xref>, <xref rid="Fig8" ref-type="fig">8</xref>, <xref rid="Fig9" ref-type="fig">9</xref>, and <xref rid="Fig10" ref-type="fig">10</xref> in comparison with observed data. The &#x0201c;mean cost&#x0201d; and the &#x0201c;minimum cost&#x0201d; curves are shown as the best models in Fig.&#x000a0;<xref rid="Fig11" ref-type="fig">11</xref>. According to Fig.&#x000a0;<xref rid="Fig10" ref-type="fig">10</xref>, &#x0201c;minimum cost&#x0201d; and &#x0201c;mean cost&#x0201d; coefficients are 0.0339 and 0.0234, respectively.<fig id="Fig7"><label>Fig.&#x000a0;7</label><caption><p>Determining corrosion current density in RC by ICA&#x02013;ANN model in training phase</p></caption><graphic xlink:href="521_2014_1645_Fig7_HTML" id="MO12"/></fig> <fig id="Fig8"><label>Fig.&#x000a0;8</label><caption><p>Comparing corrosion current density in RC by ICA&#x02013;ANN model in training phase based on the observed data</p></caption><graphic xlink:href="521_2014_1645_Fig8_HTML" id="MO13"/></fig> <fig id="Fig9"><label>Fig.&#x000a0;9</label><caption><p>Determining corrosion current density in RC by ICA&#x02013;ANN model in the test phase</p></caption><graphic xlink:href="521_2014_1645_Fig9_HTML" id="MO14"/></fig> <fig id="Fig10"><label>Fig.&#x000a0;10</label><caption><p>Comparing corrosion current density in RC by ICA&#x02013;ANN model in the test phase based on observed data</p></caption><graphic xlink:href="521_2014_1645_Fig10_HTML" id="MO15"/></fig> <fig id="Fig11"><label>Fig.&#x000a0;11</label><caption><p>Graph cost for 250 iterations in the model FF-ICA-250GEN_4IN as the best model</p></caption><graphic xlink:href="521_2014_1645_Fig11_HTML" id="MO16"/></fig> </p></sec><sec id="Sec8"><title>Validating the model</title><p>To evaluate the accuracy of ICA&#x02013;ANN model, it will be compared with GA. The properties of the models have been listed in Table&#x000a0;<xref rid="Tab6" ref-type="table">6</xref>. Besides, the ICA&#x02013;ANN model with structure&#x000a0;4-5-4-1 and excitation function of Tansig have been exploited for two algorithms.<table-wrap id="Tab6"><label>Table&#x000a0;6</label><caption><p>Introducing ICA and GA parameters</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left">GA</th><th align="left"/><th align="left">ICA</th><th align="left"/></tr></thead><tbody><tr><td align="left">Population</td><td align="left">150</td><td align="left">Countries</td><td align="left">400</td></tr><tr><td align="left">Mutation rate</td><td char="." align="char">15</td><td align="left">Revolution rate</td><td align="left">0.3</td></tr><tr><td align="left" rowspan="2">Crossover rate</td><td align="left" rowspan="2">50</td><td align="left">Empires</td><td align="left">50</td></tr><tr><td align="left">Uniting threshold</td><td align="left">0.02</td></tr><tr><td align="left">Generation</td><td char="." align="char">250</td><td align="left">Generation</td><td align="left">250</td></tr></tbody></table></table-wrap> </p><p>According to a survey conducted investigating training and testing data in the two models, <italic>R</italic> ² coefficients and straight line slope have been shown in Table&#x000a0;<xref rid="Tab6" ref-type="table">6</xref>.</p><p>Table&#x000a0;<xref rid="Tab7" ref-type="table">7</xref> shows the results of steel corrosion in concrete by the two. Table&#x000a0;<xref rid="Tab7" ref-type="table">7</xref> discussed the results of the survey.<table-wrap id="Tab7"><label>Table&#x000a0;7</label><caption><p>Results of various algorithms in both training and test phases</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="2">No.</th><th align="left" rowspan="2">Model</th><th align="left" colspan="2">Best fitting line in testing phase</th><th align="left" colspan="2">Best fitting line in training phase</th></tr><tr><th align="left">Equation</th><th align="left"> <italic>R</italic> ² </th><th align="left">Equation</th><th align="left"> <italic>R</italic> ² </th></tr></thead><tbody><tr><td align="left">1</td><td align="left">ICA&#x02013;ANN</td><td align="left"> <italic>y</italic>&#x000a0;=&#x000a0;0.7337<italic>x</italic>&#x000a0;&#x02212;&#x000a0;0.0618</td><td char="." align="char">0.8019</td><td align="left"> <italic>y</italic>&#x000a0;=&#x000a0;0.8877<italic>x</italic>&#x000a0;&#x02212;&#x000a0;0.0042</td><td char="." align="char">0.9045</td></tr><tr><td align="left">2</td><td align="left">GA&#x02013;ANN</td><td align="left"> <italic>y</italic>&#x000a0;=&#x000a0;0.6239<italic>x</italic>&#x000a0;&#x02212;&#x000a0;0.1381</td><td char="." align="char">0.6618</td><td align="left"> <italic>y</italic>&#x000a0;=&#x000a0;0.8098<italic>x</italic>&#x000a0;&#x02212;&#x000a0;0.0181</td><td char="." align="char">0.8505</td></tr></tbody></table></table-wrap> </p><p>Figures&#x000a0;<xref rid="Fig12" ref-type="fig">12</xref> and <xref rid="Fig13" ref-type="fig">13</xref> show the comparison results of steel corrosion in computational and observational concrete by the two algorithms in the training phase. The cost graph is also shown by two algorithms in Fig.&#x000a0;<xref rid="Fig13" ref-type="fig">13</xref>.<fig id="Fig12"><label>Fig.&#x000a0;12</label><caption><p>Comparing steel corrosion values in observational and computational concrete using algorithms ICA, GA during the training phase</p></caption><graphic xlink:href="521_2014_1645_Fig12_HTML" id="MO17"/></fig> <fig id="Fig13"><label>Fig.&#x000a0;13</label><caption><p>Comparing steel corrosion values in observational and computational concrete using algorithms ICA, GA in the test phase</p></caption><graphic xlink:href="521_2014_1645_Fig13_HTML" id="MO18"/></fig> </p><p>Due to the fitted lines equations presented on the computational and observational values in each model and determination coefficient related to each in Table&#x000a0;<xref rid="Tab8" ref-type="table">8</xref>, as well as the obtained results presented statistically in Table&#x000a0;<xref rid="Tab7" ref-type="table">7</xref>, it is clear that the ANN optimized by competitive ICA specifies steel corrosion much more carefully compared with GA. Given the determining of steel corrosion by two models in Figs.&#x000a0;<xref rid="Fig11" ref-type="fig">11</xref>, <xref rid="Fig12" ref-type="fig">12</xref> and <xref rid="Fig13" ref-type="fig">13</xref>, it is obvious that the ANN optimized by ICA is more accurate and flexible than the GA.<table-wrap id="Tab8"><label>Table&#x000a0;8</label><caption><p>Statistical results of ANN models optimized by ICA and GA in training and testing phases</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="2">No.</th><th align="left" rowspan="2">Model</th><th align="left" colspan="2">MAE</th><th align="left" colspan="2">RMSE</th><th align="left" colspan="2">RMSD</th></tr><tr><th align="left">Train</th><th align="left">Test</th><th align="left">Train</th><th align="left">Test</th><th align="left">Train</th><th align="left">Test</th></tr></thead><tbody><tr><td align="left">1</td><td align="left">ICA&#x02013;ANN</td><td char="." align="char">0.098</td><td char="." align="char">0.214</td><td char="." align="char">0.023</td><td char="." align="char">0.157</td><td char="." align="char">0.282</td><td char="." align="char">0.372</td></tr><tr><td align="left">2</td><td align="left">GA&#x02013;ANN</td><td char="." align="char">0.130</td><td char="." align="char">0.247</td><td char="." align="char">0.037</td><td char="." align="char">0.192</td><td char="." align="char">0.326</td><td char="." align="char">0.392</td></tr></tbody></table></table-wrap> </p><p>The cost graph has also been presented for the two models in Fig.&#x000a0;<xref rid="Fig14" ref-type="fig">14</xref>.<fig id="Fig14"><label>Fig.&#x000a0;14</label><caption><p>Cost graph of 250 iterations in three models ICA&#x02013;ANN, GA&#x02013;ANN</p></caption><graphic xlink:href="521_2014_1645_Fig14_HTML" id="MO19"/></fig> </p><p>To assess the optimum performance of ICA&#x02013;ANN model in predicting the steel corrosion, its results are compared with GA&#x02013;ANN results. Thus, 5 samples were used to predict, in two models. The main advantage of the models is that training and testing phases have not been used in any of them. Steel corrosion in concrete predicted by the two models has been listed in Table&#x000a0;<xref rid="Tab9" ref-type="table">9</xref>.<table-wrap id="Tab9"><label>Table&#x000a0;9</label><caption><p>Steel corrosion in concrete predicted by ICA&#x02013;ANN and GA&#x02013;ANN models</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="2">Row</th><th align="left" rowspan="2">Model</th><th align="left" colspan="2">Results obtained by the graphs in training phase</th><th align="left" colspan="3">Statistical parameters being evaluated</th></tr><tr><th align="left"> <italic>R</italic> ² </th><th align="left">Equation</th><th align="left">RMSD</th><th align="left">RMSE</th><th align="left">MAE</th></tr></thead><tbody><tr><td align="left">1</td><td align="left">ICA&#x02013;ANN</td><td char="." align="char">0.8087</td><td align="left"> <italic>y</italic>&#x000a0;=&#x000a0;0.6836<italic>x</italic>&#x000a0;&#x02212;&#x000a0;0.0735</td><td char="." align="char">0.510</td><td char="." align="char">0.110</td><td char="." align="char">0.287</td></tr><tr><td align="left">2</td><td align="left">GA&#x02013;ANN</td><td char="." align="char">0.7028</td><td align="left"> <italic>y</italic>&#x000a0;=&#x000a0;0.5985<italic>x</italic>&#x000a0;&#x02212;&#x000a0;0.1396</td><td char="." align="char">0.517</td><td char="." align="char">0.174</td><td char="." align="char">0.328</td></tr></tbody></table></table-wrap> </p><p>Figures&#x000a0;<xref rid="Fig15" ref-type="fig">15</xref> and <xref rid="Fig16" ref-type="fig">16</xref> provide comparing prediction of steel corrosion values in concrete for in terms of obtained results from calculations and actual data by the two models.<fig id="Fig15"><label>Fig.&#x000a0;15</label><caption><p>Predicting steel corrosion in concrete using various algorithms</p></caption><graphic xlink:href="521_2014_1645_Fig15_HTML" id="MO20"/></fig> <fig id="Fig16"><label>Fig.&#x000a0;16</label><caption><p>Comparing steel corrosion values in concrete for observational data and computational models in the prediction phase</p></caption><graphic xlink:href="521_2014_1645_Fig16_HTML" id="MO21"/></fig> </p><p>According to the presented fitted lines equations on steel corrosion rate in concrete in each model in Fig.&#x000a0;<xref rid="Fig15" ref-type="fig">15</xref> and the associated determination and statistical coefficients in Table&#x000a0;<xref rid="Tab8" ref-type="table">8</xref>, it can be concluded that the ICA&#x02013;ANN evolutionary ANN model possesses more accuracy in prediction compared with GA&#x02013;ANN model. Moreover, due to predicting steel corrosion in concrete by the two models in Fig.&#x000a0;<xref rid="Fig14" ref-type="fig">14</xref>, we can conclude that ICA&#x02013;ANN model is more accurate and flexible than GA&#x02013;ANN model.</p></sec></sec><sec id="Sec9"><title>Conclusions</title><p>A new proposition of corrosion current density prediction of reinforced concrete specimens subjected to corrosion test is presented in this numerical study. The proposed formulations are derived from the most popular algorithms used in ANN, namely ICA and GA. For this, available experimental data presented in the existing literature were gathered and used for prediction. Based on the analysis of the results, the following conclusions can be drawn:<list list-type="bullet"><list-item><p>The experimental results of the proposed method on the ANN architecture showed that the ICA is quite successful in finding the functions&#x02019; optimum point. The practical problems solved with this algorithm also demonstrated that the proposed optimization strategy can be successfully applied along with other techniques such as GA and particles swarm to help solving practical problems. Comparing the results obtained by the proposed algorithm with the optimization conventional methods indicated the relative superiority of this algorithm.</p></list-item><list-item><p>The ability to optimize the ICA can be used as a powerful tool to optimize ANN weights.</p></list-item><list-item><p>Comparing the results of training and testing ANN different models optimized using ICA, indicated that the ANN with 4-5-4-1 structure, the transfer function of Tansigmoid and the associated properties to Colonial Competitive Algorithm with 400 countries and 50 primary emperors and 250 iterations, is able to accurately predict corrosion current density in RC.</p></list-item><list-item><p>In the best ANN optimized by ICA in predicting steel corrosion rate in concrete in the training and test phase, the coefficient is 0.8019 and 0.9045, respectively, and also the line slope of this parameter is equal to 0.7337 and 0.8877, indicating the higher accuracy of model to the peers, also in this model, the statistical factor values of MAE, RMSE and RMSD are less than all models, which presented lower error in this model.</p></list-item><list-item><p>Results showed that the ANN optimized with ICA possesses higher accuracy and flexibility in predicting steel corrosion in concrete than GA.</p></list-item></list> </p></sec><sec id="Sec10"><title>Recommendations</title><p>The purpose of this paper is to predict corrosion current density in concrete using ANN combined with ICA used to optimize weights of ANN. The ICA&#x02013;ANN model has a theoretical value and potential practical significance in the prediction of the corrosion current rate of steel in concrete using corrosion current density without the need for a connection to the steel reinforcement, and it may help the engineers in practice to make a comparison for the corrosion performance of steel in concrete in RC elements. However, to increase the generalization capability, the database should be further enhanced by increasing the number of training samples, a wider range of concrete humidity, corrosion current density, or the concrete&#x02019;s resistivity.</p></sec>

Colony size, but not density, affects survival and mating success of alternative male reproductive tactics in a polyphenic mite, <italic>Rhizoglyphus echinopus</italic>

Could not extract abstract

<contrib contrib-type="author" corresp="yes"><name><surname>Radwan</surname><given-names>Jacek</given-names></name><address><phone>+48 661641380</phone><email>jradwan@amu.edu.pl</email></address><xref ref-type="aff" rid="Aff1">1</xref><xref ref-type="aff" rid="Aff2">2</xref></contrib><contrib contrib-type="author"><name><surname>&#x00141;ukasiewicz</surname><given-names>Aleksandra</given-names></name><xref ref-type="aff" rid="Aff1">1</xref></contrib><contrib contrib-type="author"><name><surname>Twardawa</surname><given-names>Mateusz</given-names></name><xref ref-type="aff" rid="Aff1">1</xref></contrib><aff id="Aff1"><label>1</label>Evolutionary Biology Group, Faculty of Biology, Adam Mickiewicz University, Pozna&#x00144;, Poland </aff><aff id="Aff2"><label>2</label>Institute of Environmental Sciences, Jagiellonian University, Gronostajowa 7, 30-387, Krak&#x000f3;w, Poland </aff>

Behavioral Ecology and Sociobiology

<sec id="Sec1" sec-type="introduction"><title>Introduction</title><p>Alternative reproductive tactics (ARTs) are among the most spectacular examples of discontinuous phenotypic variation within species (reviewed in: Gross <xref ref-type="bibr" rid="CR9">1996</xref>; Brockmann <xref ref-type="bibr" rid="CR1">2001</xref>; Shuster and Wade <xref ref-type="bibr" rid="CR32">2003</xref>; Oliveira et al. <xref ref-type="bibr" rid="CR22">2008</xref>). A common type of ART is found in species that have two distinct male phenotypes: aggressive, often territorial males and sneaker males that avoid direct competition for mates. These alternative behavioural phenotypes often coincide with morphological differences; for example, aggressive males may be armoured with horns (e.g. Emlen <xref ref-type="bibr" rid="CR5">1997a</xref>), forceps (e.g. Tomkins and Brown <xref ref-type="bibr" rid="CR36">2004</xref>), or sharpened appendages (e.g. Woodring <xref ref-type="bibr" rid="CR39">1969</xref>) used in combat, whereas sneaker males are often small in size, which facilitates sneaky behaviours (e.g. Gross <xref ref-type="bibr" rid="CR8">1991</xref>; Shuster and Wade <xref ref-type="bibr" rid="CR31">1991</xref>; Emlen <xref ref-type="bibr" rid="CR5">1997a</xref>). Alternative mating tactics were initially interpreted in terms of the Hawk and Dove evolutionarily stable strategy (ESS) model (Maynard Smith and Price <xref ref-type="bibr" rid="CR19">1973</xref>), which led to the prediction that, under certain conditions, genotypes coding for alternative ways of dealing with conflict over resources may co-exist in stable proportions, with equal fitness of alternatives enforced by negative-frequency dependence. However, later models incorporated the observation that individuals typically differ in traits that affect their success in fights (e.g. body size). Given such asymmetry, a conditional strategy in which a male undertakes risky fights only if he is stronger than his opponent (called the &#x0201c;assessor&#x0201d; strategy) is evolutionarily stable (Maynard Smith <xref ref-type="bibr" rid="CR18">1982</xref>). Not surprisingly, most cases of alternative mating tactics fit such a conditional strategy model, with less endowed individuals assuming best-of-a-bad-job tactics (Eberhard <xref ref-type="bibr" rid="CR4">1982</xref>; Gross <xref ref-type="bibr" rid="CR9">1996</xref>). Given the large environmental component to body size variation, expression of alternative tactics may also be influenced by conditions such as food availability during development (e.g. Radwan <xref ref-type="bibr" rid="CR24">1995</xref>; Emlen <xref ref-type="bibr" rid="CR6">1997b</xref>).</p><p>In addition to indirectly causing changes in an individual&#x02019;s condition, the environment can also influence the success of ARTs by directly affecting the costs and benefits of different tactics. For example, in the acarid mite <italic>Sancassania berlesei</italic>, the success of the aggressive male tactic was shown to be modulated by population size: the aggressive tactic was associated with a high mortality cost, but this cost was counterbalanced by the benefit of better access to females when the number of potential rivals was small (Radwan <xref ref-type="bibr" rid="CR23">1993</xref>). Population density has also been implicated as a determining factor in the success of ARTs in species including earwigs (Tomkins and Brown <xref ref-type="bibr" rid="CR36">2004</xref>), dung beetles (Moczek <xref ref-type="bibr" rid="CR20">2003</xref>; Buzatto et al. <xref ref-type="bibr" rid="CR2">2012</xref>), and European bitterlings (Reichard et al. <xref ref-type="bibr" rid="CR29">2004</xref>). Similarly, the density of ovipositing females significantly affected the respective success of perching vs. hovering tactics of males in the threadtail damselfly, <italic>Protoneura amatoria</italic> (Larison <xref ref-type="bibr" rid="CR15">2009</xref>). Furthermore, the relative success of ARTs among acarids has also been shown to be modulated by the structural complexity of their environment (Lukasik et al. <xref ref-type="bibr" rid="CR16">2006</xref>; Tomkins et al. <xref ref-type="bibr" rid="CR35">2011</xref>). Here, we investigate whether population size and/or density affects success of ARTs in an acarid mite, <italic>Rhizoglyphus echinopus</italic>.</p><p>Acarid mites are a convenient model with which to study ARTs because aggressive fighters can be readily distinguished from benign scramblers by the presence of a thickened, sharply terminated pair of legs that are used to kill rivals by puncturing their cuticle (Woodring <xref ref-type="bibr" rid="CR39">1969</xref>; Radwan <xref ref-type="bibr" rid="CR23">1993</xref>; Radwan et al. <xref ref-type="bibr" rid="CR26">2000</xref>). Although, in all male-dimorphic acarid species that have been investigated thus far, fighter males tend to develop from larger nymphs (Radwan <xref ref-type="bibr" rid="CR24">1995</xref>, <xref ref-type="bibr" rid="CR25">2001</xref>; Radwan et al. <xref ref-type="bibr" rid="CR28">2002</xref>; Smallegange <xref ref-type="bibr" rid="CR33">2011</xref>; Tomkins et al. <xref ref-type="bibr" rid="CR35">2011</xref>), acarid species differ in how their social environment affects morph determination. In <italic>S. berlesei</italic> and <italic>R. echinopus</italic>, for example, the fighter morph is only expressed in small colonies, while airborne substances (pheromones) emanating from dense colonies can completely suppress its expression (Timms et al. <xref ref-type="bibr" rid="CR34">1980</xref>; Radwan <xref ref-type="bibr" rid="CR25">2001</xref>; Radwan et al. <xref ref-type="bibr" rid="CR28">2002</xref>); in contrast, expression of the fighter morph in <italic>Rhizoglyphus robini</italic> does not depend on colony density during an individual&#x02019;s nymphal stage (Radwan <xref ref-type="bibr" rid="CR24">1995</xref>).</p><p>In <italic>S. berlesei</italic>, sensitivity to chemical cues that are correlated with population size has been demonstrated to be adaptive: in large, dense colonies, fighters suffered high combat-related mortality, whereas scramblers were involved in fights less often, survived better, and achieved higher reproductive success (Radwan <xref ref-type="bibr" rid="CR23">1993</xref>). The system of pheromonal regulation of morph expression is therefore akin to polyphenisms in which morphology is adaptively modulated by environmental cues, such as protective morphologies induced by cues indicating the presence of predators (reviewed by Roff <xref ref-type="bibr" rid="CR30">1996</xref>). In contrast to <italic>S. berlesei</italic>, no significant relationship between population size and relative mortality and reproductive success of male morphs was found in <italic>R. robini</italic> (Radwan and Klimas <xref ref-type="bibr" rid="CR27">2001</xref>). Taken together, these results suggest that there is a relationship between pheromone sensitivity and the degree to which the fitness of different ARTs changes in response to population parameters, but data from a larger number of species are necessary to determine the generality of this relationship. Here, we study the effect of population size and density on the survival and mating success of alternative male morphs in <italic>R. echinopus</italic>, which, despite being a congener of <italic>R. robini</italic>, is characterised by pheromonal suppression of the fighter morph (Radwan <xref ref-type="bibr" rid="CR25">2001</xref>). Based on this mode of morph determination, we predict that, in <italic>R. echinopus</italic>, males using the fighting tactic will have higher reproductive success than those using the scrambler tactic in small and/or low-density colonies, but the situation will be reversed in favour of scrambler males in large and/or dense colonies.</p><p>As a result of very high fecundity and short generation time, populations of acarid mites are usually very large (hundreds to thousands of individuals) and dense, but great variations in population size may occur when the mites colonise new patches of habitat by phoresis. Hypopi are often associated phoretically with dung beetles, and mites were reported to stay on the body of a dead beetle, and to develop and feed on it for some time after its death (Kranz <xref ref-type="bibr" rid="CR14">1978</xref>; Houck and Oconnor <xref ref-type="bibr" rid="CR11">1991</xref>; Diaz et al. <xref ref-type="bibr" rid="CR3">2000</xref>). The number of hypopi found on beetles varies from a few to hundreds (Houck and Oconnor <xref ref-type="bibr" rid="CR11">1991</xref>). In this study, we created experimental conditions similar to those occurring when hypopi develop and a number of adult males of approximately the same age compete for females. We first verified the colony size at which fighter tactic expression is suppressed, and then we manipulated colony size and density to understand how this influenced the survival and reproductive success of fighters.</p></sec><sec id="Sec2" sec-type="materials|methods"><title>Methods</title><sec id="Sec3"><title>Mites</title><p> <italic>R. echinopus</italic> is a broadly distributed acarid mite feeding on bulbs and tubers of various plants and on stored food products (Diaz et al. <xref ref-type="bibr" rid="CR3">2000</xref>). Thus, their natural habitat is often patchily distributed, but new patches of habitat are colonised by phoresy, mostly on scarab beetles, but also curculionids and other insects occurring on the same host plants (reviewed by Diaz et al. <xref ref-type="bibr" rid="CR3">2000</xref>). The generation time is about two weeks, and includes the following pre-adult stages: egg, larva, protonymph, optional migratory deuteronymph (hypopus), and tritonymph. Females lay hundreds of eggs during lifespan, such that population growth may be high under optimal conditions, leading to depletion of resources, which induces formation of hypopi (Diaz et al. <xref ref-type="bibr" rid="CR3">2000</xref>). Here, we follow the classification scheme of genus <italic>Rhizoglyphus</italic> adopted by Manson (<xref ref-type="bibr" rid="CR17">1972</xref>), that is, the study species has long <italic>Sci</italic> setae and an oval penis base. These features differentiate <italic>R. echinopus</italic> from <italic>R. robini</italic>, which occurs in similar habitats (Diaz et al. <xref ref-type="bibr" rid="CR3">2000</xref>).</p><p>Mites used in this experiment came from a large stock population originally collected from tulip bulbs and subsequently cultured in the Department of Entomology, SGGW, Warsaw, Poland. The mites were maintained in a controlled temperature chamber at 24&#x02009;&#x000b1;&#x02009;1&#x000a0;&#x000b0;C; to maintain the desired level of humidity in the environment (&#x0003e;90&#x000a0;%), the mites were kept in desiccators filled with KOH solution (153&#x000a0;g/l H₂O). During the experiments, the mites were reared in tubes, whose bases were made of plaster of Paris mixed with powdered charcoal in order to provide a dark background that facilitates behavioural observations. Tubes were secured with non-absorbent cotton wool and provided with food (powdered yeast) ad libitum. Vials from all treatments were kept in a common large desiccator.</p></sec><sec id="Sec4"><title>The choice of colony parameters</title><p>In order to determine colony parameters which should be used in experiments investigating relative survival and mating success of male morphs, we first determined the proportions of morphs at colonies of different size. To that aim, we placed 4, 6, 8, 10, 12, or 18 protonymphs in tubes (0.8&#x000a0;cm in diameter) and repeated this process 10 times using a total of 60 tubes. Individuals in different tubes were thus exposed to different amounts of pheromones during the sensitive period that determines morph development, which occurs at the early tritonymphal stage (Radwan <xref ref-type="bibr" rid="CR25">2001</xref>). Then, during the quiescent stage that precedes the emergence of adults, we isolated tritonymphs into individual cells to prevent fight-related mortality and counted the number of each morph that emerged in each treatment. Consistent with earlier findings (Radwan <xref ref-type="bibr" rid="CR25">2001</xref>), the number of nymphs developing into fighters decreased with group size (Fig.&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>). For colony sizes between 6 and 12 individuals, both morphs emerged at a similar frequency, so we used a colony size of 8 mites/tube as a benchmark at which we expected the survival and reproductive success of both morphs to be approximately equal. We then created colonies with sizes or densities that were smaller or larger than 8 mites/tube in order to test the prediction outlined in the <xref rid="Sec1" ref-type="sec">Introduction</xref> that fighters have better survival and higher mating success (relative to scramblers) in small and/or low-density colonies but the situation is reversed in large/dense colonies. We did not replace dead individuals to keep the density/size constant throughout the experiment, as we preferred to create conditions similar to those when new colonies are started by hypopi colonising a new habitat patch, because this is when variation in colony sizes is most likely to occur. Nonetheless, differences in size/density between treatments were large enough to persist throughout experiments.<fig id="Fig1"><label>Fig. 1</label><caption><p>The effect of colony size on the proportions of fighter and scrambler males that emerged in each colony</p></caption><graphic xlink:href="265_2014_1787_Fig1_HTML" id="MO1"/></fig> </p></sec><sec id="Sec5"><title>The effect of population size and density on survival and mating success of male morphs</title><p>Mites for the experiment were obtained from groups of eight protonymphs as described above and were used for the experiments up to 1&#x000a0;day after the final moult. In order to test the effect of population size, we established colonies of three different sizes, but each with a density of about 16 mites/cm², equal sex ratio, and equal proportions of male morphs. Small colonies (<italic>N</italic>&#x02009;=&#x02009;40) consisted of 4 individuals in 0.55-cm-diameter tubes, intermediate colonies (<italic>N</italic>&#x02009;=&#x02009;20) of 8 individuals in 0.8-cm-diameter tubes, and large colonies (<italic>N</italic>&#x02009;=&#x02009;10) of 32 individuals in 1.6-cm-diameter tubes.</p><p>In order to test the effect of population density, colonies of eight individuals (four females, two fighters, two non-fighters) were established in cells that measured 0.55&#x000a0;cm (<italic>N</italic>&#x02009;=&#x02009;18), 0.8&#x000a0;cm (<italic>N</italic>&#x02009;=&#x02009;20) and 1.6&#x000a0;cm (<italic>N</italic>&#x02009;=&#x02009;20) in diameter, which resulted in three density treatments: 34, 16, and 4 mites/cm², respectively.</p><p>In both the population size manipulation experiment and density manipulation experiment, we recorded daily survival and the number of copulations and fights by males three times a day between 9&#x000a0;am and 6&#x000a0;pm at 4&#x000a0;h intervals. Such an interval prevented scoring the same behaviour more than once, as both mating and fights do not last longer than a couple of hours (personal observations). The observations were carried out until new generation emerged, i.e. for 13&#x000a0;days in size manipulation experiment and 15&#x000a0;days in density manipulation experiment.</p></sec><sec id="Sec6"><title>Data analysis</title><p>Our aim was to compare fitness components of fighters relative to that of scramblers between colonies of different size or density. Therefore for each colony, we calculated two measures of relative success of male morphs: relative survival and relative mating success. The relative survival of fighters compared to scramblers was calculated as the number of days survived summed for all fighters in a given colony, divided by the summed number of days survived for scramblers. Given equal numbers of males of both morphs at the beginning of the experiment, higher relative survival of fighters implies that during the time of the experiment, i.e. before the new generation matured, an average fighter survived better and thus had more opportunities to inseminate females. Similarly, the relative mating success of fighters was calculated as (number of copulations by fighters in a colony&#x02009;+&#x02009;1)/(number of copulations by scramblers in a colony&#x02009;+&#x02009;1); the value of 1 was added to the number of mating attempts because, in some cases in small colonies, no copulations by scramblers were recorded. The relative survival and mating success was compared between population size treatments and between density treatments using a non-parametric Kruskal-Wallis test. To check if fights between fighters and between fighters and scramblers occurred with equal probability, we compared observed proportions to the expectation given by the number of fighters (f) and scramblers (s) in a colony, i.e. <inline-formula id="IEq1"><alternatives><tex-math id="M1">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ \left(\begin{array}{c}\hfill f\hfill \\ {}\hfill 2\hfill \end{array}\right)/\left[\left(\begin{array}{c}\hfill f\hfill \\ {}\hfill 2\hfill \end{array}\right)+ fs\right] $$\end{document}</tex-math><mml:math id="M2"><mml:mfenced close=")" open="("><mml:mtable columnalign="center"><mml:mtr columnalign="center"><mml:mtd columnalign="center"><mml:mi>f</mml:mi></mml:mtd></mml:mtr><mml:mtr columnalign="center"><mml:mtd columnalign="center"><mml:mn>2</mml:mn></mml:mtd></mml:mtr></mml:mtable></mml:mfenced><mml:mo stretchy="true">/</mml:mo><mml:mfenced close="]" open="["><mml:mrow><mml:mfenced close=")" open="("><mml:mtable columnalign="center"><mml:mtr columnalign="center"><mml:mtd columnalign="center"><mml:mi>f</mml:mi></mml:mtd></mml:mtr><mml:mtr columnalign="center"><mml:mtd columnalign="center"><mml:mn>2</mml:mn></mml:mtd></mml:mtr></mml:mtable></mml:mfenced><mml:mo>+</mml:mo><mml:mi mathvariant="italic">fs</mml:mi></mml:mrow></mml:mfenced></mml:math><inline-graphic xlink:href="265_2014_1787_Article_IEq1.gif"/></alternatives></inline-formula> (Radwan <xref ref-type="bibr" rid="CR23">1993</xref>).</p></sec></sec><sec id="Sec7" sec-type="results"><title>Results</title><p>Male mortality was much higher than female mortality in both experiments (Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>; Table&#x000a0;<xref rid="MOESM1" ref-type="media">S1</xref>), a result that was most likely due to fighting between males. Indeed, when there was only one fighter in a colony, fighter mortality was very low and similar to that of females (Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>; Table&#x000a0;<xref rid="MOESM1" ref-type="media">S1</xref>, small colonies). Fighter survival decreased with colony size (Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>, <italic>H</italic>&#x02009;=&#x02009;48.5, <italic>N</italic>&#x02009;=&#x02009;79, <italic>P</italic>&#x02009;&#x0003c;&#x02009;001), but that of scramblers did not (<italic>H</italic>&#x02009;=&#x02009;1.1, <italic>N</italic>&#x02009;=&#x02009;70, <italic>P</italic>&#x02009;=&#x02009;0.571). Interestingly, survival of fighters, but not scramblers increased with colony density (Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>; fighters, <italic>H</italic>&#x02009;=&#x02009;14.9, <italic>N</italic>&#x02009;=&#x02009;58, <italic>P</italic>&#x02009;=&#x02009;0.001; scramblers, H&#x02009;=&#x02009;2.2, B&#x02009;=&#x02009;58, <italic>P</italic>&#x02009;=&#x02009;0.324).<table-wrap id="Tab1"><label>Table 1</label><caption><p>Number of days survived by males (based on colony averages) in colonies differing in size (small, 4 mites; medium, 8 mites; large, 32 mites) or density (8 mites at densities of 34, 16, and 4 mites/cm²)</p></caption><table frame="hsides" rules="groups"><thead><tr><th rowspan="2" colspan="2">Colony type (<italic>N</italic>)</th><th colspan="3">Mean (SE)</th><th colspan="3">Median (range)</th></tr><tr><th>Females</th><th>Fighters</th><th>Scramblers</th><th>Females</th><th>Fighters</th><th>Scramblers</th></tr></thead><tbody><tr><td rowspan="3">Size</td><td>Small (40)</td><td align="left">12.7 (0.14)</td><td align="left">12.8 (0.25)</td><td align="left">9.5 (0.64)</td><td align="left">13 (8&#x02013;13)</td><td align="left">13 (8&#x02013;13)</td><td align="left">13 (1&#x02013;13)</td></tr><tr><td>Medium (20)</td><td align="left">12.8 (0.20)</td><td align="left">8.9 (0.35)</td><td align="left">9.7 (0.91)</td><td align="left">13 (11&#x02013;13)</td><td align="left">7.3 (7&#x02013;13)</td><td align="left">8.5 (6&#x02013;13)</td></tr><tr><td>Large (10)</td><td align="left">19.9 (0.28)</td><td align="left">8.1 (0.50)</td><td align="left">9.9 (1.29)</td><td align="left">13 (12.3&#x02013;13)</td><td align="left">7.8 (5.8&#x02013;11.2)</td><td align="left">10 (5.8&#x02013;13)</td></tr><tr><td rowspan="3">Density</td><td>Low (20)</td><td align="left">14.7 (0.21)</td><td align="left">8.72 (9.69)</td><td align="left">11.2 (0.70)</td><td align="left">15 (11.5&#x02013;15)</td><td align="left">8 (4&#x02013;15)</td><td align="left">10 (6.6&#x02013;15)</td></tr><tr><td>Medium (20)</td><td align="left">14.5 (0.21)</td><td align="left">10.9 (0.69)</td><td align="left">11.4 (0.70)</td><td align="left">15 (12&#x02013;15)</td><td align="left">9.3 (8&#x02013;15)</td><td align="left">11.3 (8&#x02013;15)</td></tr><tr><td>High (18)</td><td align="left">14.2 (0.22)</td><td align="left">12.5 (0.72)</td><td align="left">12.7 (0.74)</td><td align="left">15 (11.5&#x02013;15)</td><td align="left">8 (4&#x02013;15)</td><td align="left">10 (6.5&#x02013;15)</td></tr></tbody></table><table-wrap-foot><p>For size manipulation, maximum number of days was 13, for density manipulation 15</p></table-wrap-foot></table-wrap> </p><p>In 17 small colonies, a fighter killed its rival and monopolised females (Table&#x000a0;<xref rid="MOESM1" ref-type="media">S1</xref>). In intermediate-sized colonies, monopolisation occurred in only one colony, and no cases of monopolisation were recorded in large colonies. In the population density treatment (equivalent in size to intermediate colonies), monopolisation by a fighter was recorded in four low-density colonies. In one high-density colony, both fighters died by the end of the experiment.</p><p>Fights were observed both between pairs of fighters and between fighters and scramblers (Table&#x000a0;<xref rid="Tab2" ref-type="table">2</xref>). The total number of fights observed was low and most of them occurred during the first day of observations, so we only analysed data from that day. We compared frequencies of fights between pairs of fighters and between fighters and scramblers for large colonies, and for medium-size colonies pooled for all densities (as the frequencies did not differ significantly between densities, Yates <italic>&#x003c7;</italic> ²&#x02009;=&#x02009;0.079, df&#x02009;=&#x02009;2, <italic>P</italic>&#x02009;=&#x02009;0.961). In both cases, fights between fighters occurred disproportionately more often (large colonies, <italic>&#x003c7;</italic> ²&#x02009;=&#x02009;7.125, df&#x02009;=&#x02009;1, <italic>P</italic>&#x02009;=&#x02009;0.007; medium-sized colonies, <italic>&#x003c7;</italic> ²&#x02009;=&#x02009;15.244, df&#x02009;=&#x02009;1, <italic>P</italic>&#x02009;&#x0003c;&#x02009;0.001) than expected (Table&#x000a0;<xref rid="Tab2" ref-type="table">2</xref>). Likewise, when we pooled data from all colonies that contained eight individuals at medium density across both treatments, we also found that conflicts between fighters occurred significantly more often than expected (<italic>&#x003c7;</italic> ²&#x02009;=&#x02009;6.612, <italic>P</italic>&#x02009;=&#x02009;0.010).<table-wrap id="Tab2"><label>Table 2</label><caption><p>Number of fights that occurred between pairs of fighters (f-f) and pairs of fighters and scramblers (f-s) observed in colonies differing in size and density; the proportion of fights expected between fighters given the number of fighters (f) and scramblers (s) in a colony was <inline-formula id="IEq2"><alternatives><tex-math id="M3">\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ \left(\begin{array}{c}\hfill f\hfill \\ {}\hfill 2\hfill \end{array}\right)/\left[\left(\begin{array}{c}\hfill f\hfill \\ {}\hfill 2\hfill \end{array}\right)+ fs\right] $$\end{document}</tex-math><mml:math id="M4"><mml:mfenced close=")" open="("><mml:mtable columnalign="center"><mml:mtr columnalign="center"><mml:mtd columnalign="center"><mml:mi>f</mml:mi></mml:mtd></mml:mtr><mml:mtr columnalign="center"><mml:mtd columnalign="center"><mml:mn>2</mml:mn></mml:mtd></mml:mtr></mml:mtable></mml:mfenced><mml:mo stretchy="true">/</mml:mo><mml:mfenced close="]" open="["><mml:mrow><mml:mfenced close=")" open="("><mml:mtable columnalign="center"><mml:mtr columnalign="center"><mml:mtd columnalign="center"><mml:mi>f</mml:mi></mml:mtd></mml:mtr><mml:mtr columnalign="center"><mml:mtd columnalign="center"><mml:mn>2</mml:mn></mml:mtd></mml:mtr></mml:mtable></mml:mfenced><mml:mo>+</mml:mo><mml:mi mathvariant="italic">fs</mml:mi></mml:mrow></mml:mfenced></mml:math><inline-graphic xlink:href="265_2014_1787_Article_IEq2.gif"/></alternatives></inline-formula>, i.e., 0.45 in large colonies (32 individuals) and 0.375 in medium-sized colonies (8 individuals); for small colonies, containing one fighter only, f-f fights do not apply (n.a.)</p></caption><table frame="hsides" rules="groups"><thead><tr><th colspan="4">Colony size</th><th colspan="4">Colony density</th></tr></thead><tbody><tr><td/><td/><td>f-f</td><td>f-s</td><td/><td/><td>f-f</td><td>f-s</td></tr><tr><td rowspan="2">Small</td><td>Day 1</td><td>n.a.</td><td>10</td><td rowspan="2">Low</td><td>Day 1</td><td>6</td><td>4</td></tr><tr><td>Total</td><td>n.a.</td><td>11</td><td>Total</td><td>7</td><td>9</td></tr><tr><td rowspan="2">Intermediate</td><td>Day 1</td><td>10</td><td>5</td><td rowspan="2">Medium</td><td>Day 1</td><td>8</td><td>4</td></tr><tr><td>Total</td><td>11</td><td>10</td><td>Total</td><td>16</td><td>17</td></tr><tr><td rowspan="2">Large</td><td>Day 1</td><td>10</td><td>2</td><td rowspan="2">High</td><td>Day 1</td><td>4</td><td>1</td></tr><tr><td>Total</td><td>19</td><td>7</td><td>Total</td><td>12</td><td>9</td></tr></tbody></table></table-wrap> </p><p>Colony size had a significant effect both on the relative survival of fighters (Kruskal-Wallis <italic>H</italic>&#x02009;=&#x02009;15.5, <italic>N</italic>&#x02009;=&#x02009;70, <italic>P</italic>&#x02009;&#x0003c;&#x02009;0.001) and on their relative mating success (Kruskal-Wallis <italic>H</italic>&#x02009;=&#x02009;10.2, <italic>N</italic>&#x02009;=&#x02009;70, <italic>P</italic>&#x02009;=&#x02009;0.006). Relative fighter survival and mating success were highest in the smallest colonies (Fig.&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>).<fig id="Fig2"><label>Fig. 2</label><caption><p>The effect of colony size on fighters&#x02019; <bold>a</bold> survival and <bold>b</bold> mating success relative to scramblers. <italic>Error bars</italic> denote&#x02009;&#x000b1;&#x02009;SE</p></caption><graphic xlink:href="265_2014_1787_Fig2_HTML" id="MO2"/></fig> </p><p>The relative survival of fighters tended to increase with colony density (Fig.&#x000a0;<xref rid="Fig3" ref-type="fig">3a</xref>), but this effect was not statistically significant (Kruskal-Wallis <italic>H</italic>&#x02009;=&#x02009;5.0, <italic>N</italic>&#x02009;=&#x02009;58, <italic>P</italic>&#x02009;=&#x02009;0.082). Density did not significantly affect the relative mating success of fighters (Kruskal-Wallis <italic>H</italic>&#x02009;=&#x02009;1.5, <italic>N</italic>&#x02009;=&#x02009;58, <italic>P</italic>&#x02009;=&#x02009;0.485) (Fig.&#x000a0;<xref rid="Fig3" ref-type="fig">3b</xref>).<fig id="Fig3"><label>Fig. 3</label><caption><p>The effect of colony density on fighters&#x02019; <bold>a</bold> survival and <bold>b</bold> mating success relative to scramblers. <italic>Error bars</italic> denote&#x02009;&#x000b1;&#x02009;SE</p></caption><graphic xlink:href="265_2014_1787_Fig3_HTML" id="MO3"/></fig> </p></sec><sec id="Sec8" sec-type="discussion"><title>Discussion</title><p>Population parameters, such as size and density, are thought to play an important role in shaping the relative success of various ARTs (Radwan <xref ref-type="bibr" rid="CR23">1993</xref>; Knell <xref ref-type="bibr" rid="CR12">2009</xref>). For example, studies using correlational data on population density and horn allometry across several populations of the dung beetle (<italic>O. taurus</italic>) showed that the frequency of horned major males decreased at high population densities as a result of a change in the allometric scaling of horn size with body size (Moczek et al. <xref ref-type="bibr" rid="CR21">2002</xref>; Moczek <xref ref-type="bibr" rid="CR20">2003</xref>). This result is consistent with the hypothesis that sneaker tactics are adaptive and enhance reproductive success in higher-density populations. In contrast, in their study of earwigs, Tomkins and Brown (<xref ref-type="bibr" rid="CR36">2004</xref>) proposed a different explanation: in high-density populations, males with superior fighting ability will be favoured because of the increased likelihood of encounters with other males. In both dung beetles and earwigs, though, the effect of density on the relative success of different ARTs has not yet been demonstrated. In an experimental study of European bitterlings, Reichard et al. (<xref ref-type="bibr" rid="CR29">2004</xref>) manipulated the density of males and demonstrated that, at low densities, territorial males have a considerable fitness advantage over sneakers; however, at high densities, both alternative mating tactics had similar levels of success. Here, we have demonstrated that, in <italic>R. echinopus</italic>, colony size, rather than density, affects the relative success of ARTs.</p><p>Given that in <italic>R. echinopus</italic> fighter morph expression is suppressed as population size increases (Radwan <xref ref-type="bibr" rid="CR25">2001</xref>; Fig.&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>), our findings are consistent with the idea of adaptive developmental plasticity in ART expression: when the relative success of different ARTs depends on an environmental factor, and it is possible to detect cues that predict the environment in which certain ARTs will be expressed, developmental processes should evolve to achieve adaptive plasticity in ART expression. The fighter tactic was effective only when the population size was small because only under such conditions was it possible to monopolise females by killing rival males. In larger colonies, fighter tactics were associated with higher mortality and lower mating success. The higher mortality was most likely a result of the propensity of fighters to engage in combat with other males, which often resulted in death. Similar findings have been previously reported for another acarid, <italic>S. berlesei</italic> (Radwan <xref ref-type="bibr" rid="CR23">1993</xref>). Likewise, in dung beetles, anticipated increases in population density, which is believed to shape the relative success of horned and hornless males, has been recently demonstrated to affect the male phenotype via parental effects (Buzatto et al. <xref ref-type="bibr" rid="CR2">2012</xref>).</p><p>The period during which acarid mites are sensitive to cues regarding population size is also consistent with the hypothesis of adaptive plasticity. This period is the early tritonymphal stage, which follows the migratory, facultative deuteronymph stage. Radwan (<xref ref-type="bibr" rid="CR23">1993</xref>) argued that conditions favouring fighter males occur mostly when a small group of deuteronymphs, which are phoretic on insects, start a new colony. Due to their high reproductive output, acarid colonies grow rapidly, and conditions favouring fighters are probably rarely present past the early migration stage. The high dynamics of populations of acarids also suggest that resources are depleted quickly, and as a result, migration events could occur often enough to exert selective pressure favouring the expression of the fighter phenotype.</p><p>Radwan (<xref ref-type="bibr" rid="CR23">1993</xref>) argued that population size, not density, determines the relative success of fighter males in <italic>S. berlesei</italic>, as fighters succeed only if the colony size is small enough that they can kill all their rivals and subsequently monopolise females. However, the experimental design of his study confounded size and density. Here, we have separated the effects of colony size and density and have confirmed that colony size, but not density, affects the relative fitness of male morphs in <italic>R. echinopus</italic>. We had initially predicted that fighters would survive better than scramblers under low-density conditions but worse in high-density conditions, but, in fact, the relative survival of fighters tended to increase, rather than decrease, with colony density (Fig.&#x000a0;<xref rid="Fig3" ref-type="fig">3b</xref>). Likewise, fighter mortality was the lowest at the highest density. Further work utilising more intensive scoring of behaviours should reveal whether this effect could be due to decreased fighter aggression. Such behavioural plasticity may be adaptive, preventing fighters from performing costly behaviours under unfavourable conditions: high density would normally be correlated with large population size where fighter tactic is disfavoured, as shown in this study.</p><p>Under demographic conditions in which their fitnesses are approximately equal, both fighter and scrambler morphs are expected to occur. Indeed, we observed approximately the same degree of survival and mating success of each morph in groups of eight mites maintained in 0.8-cm-diameter cells (a condition that resulted in both morphs occurring in similar proportions; Fig.&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>). When tactics co-occur, their relative fitness may be related to body size; for example, adopting aggressive tactics could be more beneficial to larger and stronger males (Maynard Smith <xref ref-type="bibr" rid="CR18">1982</xref>; Gross <xref ref-type="bibr" rid="CR9">1996</xref>). Consistent with this idea is the observation that fighter males in <italic>R. echinopus</italic> emerge from heavier nymphs (Tomkins et al. <xref ref-type="bibr" rid="CR35">2011</xref>); the same is true in <italic>S. berlesei</italic> (Radwan et al. <xref ref-type="bibr" rid="CR28">2002</xref>). It thus appears that adaptive morph determination in these species may involve a complex interplay between responses to external demographic cues and internal cues of body condition. In this study, we demonstrated that the former mechanism does contribute to the maintenance of alternative male morphs in <italic>R. echinopus</italic> by causing the reversal of fitness ranking of the alternatives with increasing population size. However, the adaptive significance of the latter mechanism remains to be investigated.</p><p>We used equal proportion of males in our experiments, therefore we cannot exclude that relative success of male morphs is additionally affected by frequency dependence; for example, in large populations fighters would probably not suffer costs of fights if they occurred in low frequencies, such that encounters with other fighters would be rare. However, the benefits of monopolisation would not occur either, and in fact in large stock populations, fighter males do not occur (personal observations). Furthermore, in a congener species, <italic>R. robini</italic>, mating success of male morphs did not depend on their proportions in a colony (Radwan and Klimas <xref ref-type="bibr" rid="CR27">2001</xref>).</p><p>Previous work on two other species of acarids has shown that in <italic>S. berlesei</italic>, a species in which male morph expression is guided by pheromonal cues linked to population size, the relative success of fighters depends on population size (Radwan <xref ref-type="bibr" rid="CR23">1993</xref>), whereas in <italic>R. robini</italic>, neither morph expression nor fitness is related to population size. This suggests that the reaction norms that guide morph expression may evolve in response to a change in selective pressures on morph expression under varying demographic conditions and thus vary among species in an adaptive fashion. Our results lend support to this hypothesis by showing a pattern similar to that found in <italic>S. berlesei</italic>, a congener of <italic>R. robini</italic>: sensitivity to pheromones is associated with the reversal of the relative success of fighters and scramblers that occurs as population size increases.</p><p>At population densities at which both morphs are expressed, the degree of fighter morph expression has been shown to respond quickly to experimental selection in both <italic>S. berlesei</italic> (Unrug et al. <xref ref-type="bibr" rid="CR38">2004</xref>) and <italic>R. echinopus</italic> (Tomkins et al. <xref ref-type="bibr" rid="CR35">2011</xref>). Such changes could be due to a change in mite sensitivity to pheromones and/or the change in the reaction norm that translates body size into the probability of fighter expression. Irrespective of the mechanism, the proportion of fighters expressed at a given population density was observed to evolve rapidly. Similarly, populations of <italic>S. berlesei</italic> collected from different locations and held in common-garden conditions have been shown to differ in the proportion of fighters emerging at various colony sizes (Tomkins et al. <xref ref-type="bibr" rid="CR37">2004</xref>), implying genetic variation for this reaction norm. Furthermore, a study of experimental evolution in <italic>R. echinopus</italic> showed that when a change in environmental complexity caused a shift in the relative costs and benefits of expressing the fighter phenotype, the reaction norm evolved to decrease the probability of expressing the fighter phenotype in an unfavourable habitat (Tomkins et al. <xref ref-type="bibr" rid="CR35">2011</xref>). A genetic basis for variation in the threshold of responses to cues inducing phenotypically plastic alternative phenotypes has previously been documented in contexts other than that of sexual selection, such as in cases of predator-induced defences (Hazel and West <xref ref-type="bibr" rid="CR10">1982</xref>) or migratory polymorphism (Fairbairn and Yadlowski <xref ref-type="bibr" rid="CR7">1997</xref>; Knulle <xref ref-type="bibr" rid="CR13">2003</xref>). The existence of standing genetic variation for sensitivity to pheromone cues should facilitate the evolutionary adjustment of morph expression patterns to changes in benefits of ARTs under different colony sizes.</p><p>Given that <italic>R. echinopus</italic> is considered to be closely related to <italic>R. robini</italic>, and the two species live in similar habitats (Manson <xref ref-type="bibr" rid="CR17">1972</xref>; Diaz et al. <xref ref-type="bibr" rid="CR3">2000</xref>), it remains unclear why the success of alternative male morphs in <italic>R. robini</italic> does not depend on population size (Radwan and Klimas <xref ref-type="bibr" rid="CR27">2001</xref>). Despite morphological similarity to <italic>R. echinipus</italic>, <italic>R. robini</italic> is generally smaller and has slightly shorter and stouter legs (personal observations). Radwan and Klimas (<xref ref-type="bibr" rid="CR27">2001</xref>) proposed that the lower mobility associated with such features may make scrambler males of <italic>R. robini</italic> less capable of escaping attacks from fighters, thus removing the survival advantage in large colonies that would be experienced by the more agile <italic>S. berlesei</italic> or <italic>R. echinopus</italic>. However, this speculation has not yet been empirically tested.</p></sec><sec sec-type="supplementary-material"><title>Electronic supplementary material</title><sec id="Sec9"><p>Below is the link to the electronic supplementary material.<supplementary-material content-type="local-data" id="MOESM1"><media xlink:href="265_2014_1787_MOESM1_ESM.doc"><label>ESM 1</label><caption><p>(DOC 75&#x000a0;kb)</p></caption></media></supplementary-material> </p></sec></sec>

Queen and young larval pheromones impact nursing and reproductive physiology of honey bee (<italic>Apis mellifera</italic>) workers

Could not extract abstract

<contrib contrib-type="author" corresp="yes"><name><surname>Traynor</surname><given-names>Kirsten S.</given-names></name><address><phone>240-439-3053</phone><email>ktraynor@asu.edu</email></address><xref ref-type="aff" rid="Aff1">1</xref></contrib><contrib contrib-type="author"><name><surname>Le Conte</surname><given-names>Yves</given-names></name><xref ref-type="aff" rid="Aff2">2</xref></contrib><contrib contrib-type="author"><name><surname>Page</surname><given-names>Robert E.</given-names><suffix>Jr</suffix></name><xref ref-type="aff" rid="Aff1">1</xref></contrib><aff id="Aff1"><label>1</label>School of Life Sciences, Arizona State University, Tempe, AZ USA </aff><aff id="Aff2"><label>2</label>INRA, UR 406, Abeilles et Environnement, Site Agroparc, 84914 Avignon, France </aff>

Behavioral Ecology and Sociobiology

<p id="Par2">Within social insects, the chemical communication system has proven to be highly diversified and richly complex, enhanced by synergistic interactions and context-dependent messaging (Slessor et al. <xref ref-type="bibr" rid="CR83">2005</xref>). For example, at least 50 substances derived from queens, workers, and immatures are expressed within the colonies of honey bees (<italic>Apis mellifera</italic>) (reviewed in Pankiw <xref ref-type="bibr" rid="CR62">2004</xref>). A number of chemicals act as releasers of behavior (releaser pheromones), causing rapid but short-lived responses, such as the attraction/orienting behavior in response to pheromone emission from the dorsal Nasanov gland (Free <xref ref-type="bibr" rid="CR16">1987</xref>; Pickett et al. <xref ref-type="bibr" rid="CR69">1980</xref>). Other chemicals act as primers (primer pheromones) and slowly influence behavior through long-term physiological effects, thereby influencing broad aspects of colony organization, caste structure, and the division of labor (Le Conte and Hefetz <xref ref-type="bibr" rid="CR39">2008</xref>; Wilson and Bossert <xref ref-type="bibr" rid="CR93">1963</xref>; Winston and Slessor <xref ref-type="bibr" rid="CR94">1998</xref>). Several multifunctional pheromones have both releaser and primer effects, such as queen mandibular pheromone (QMP) and brood ester pheromones (BEPs) produced by larvae. There is increasing evidence that these multifunctional pheromones may have profound effects in shaping honey bee colony dynamics (reviewed in Alaux et al. <xref ref-type="bibr" rid="CR1">2010</xref>; Winston and Slessor <xref ref-type="bibr" rid="CR94">1998</xref>).</p><p id="Par3">One of the primary effects elicited by honey bee pheromones is the organization of care received by immature bees. Larvae are confined to a cell, cannot feed themselves, and must signal their needs to adult nurses. By emitting pheromones, the larvae influence the behavior and physiology of their nurses, stimulating them to provide appropriate nutritional resources (Arnold et al. <xref ref-type="bibr" rid="CR7">1994</xref>; Le Conte et al. <xref ref-type="bibr" rid="CR42">2001</xref>; Maisonnasse et al. <xref ref-type="bibr" rid="CR50">2010</xref>; Mohammedi et al. <xref ref-type="bibr" rid="CR52">1996</xref>; Mohammedi et al. <xref ref-type="bibr" rid="CR53">1998</xref>; Pankiw et al. <xref ref-type="bibr" rid="CR65">1998</xref>; Sagili and Pankiw <xref ref-type="bibr" rid="CR76">2009</xref>). Pheromone composition changes as larvae age, with young larvae emitting the volatile pheromone e-beta ocimene (e&#x003b2;) and old larvae predominantly emitting a blend of ethyl and methyl fatty acid esters known collectively as BEPs. Nurse bees tightly regulate larval growth by adjusting the larval feeding regime according to larval age (Leimar et al. <xref ref-type="bibr" rid="CR43">2012</xref>; Linksvayer et al. <xref ref-type="bibr" rid="CR47">2011</xref>; Wang et al. <xref ref-type="bibr" rid="CR89">2014</xref>), indicating that nurse bees may use larval pheromones to regulate larval diet and prime their physiology for brood care (Le Conte et al. <xref ref-type="bibr" rid="CR40">1994</xref>, <xref ref-type="bibr" rid="CR41">1995</xref>; Mohammedi et al. <xref ref-type="bibr" rid="CR52">1996</xref>).</p><p id="Par4">Both queens and brood emit primer pheromones that strongly impact cooperative brood care, a redundancy in control mechanisms that appears to be a common feature of pheromone-based signaling systems in eusocial insects (Hoover et al. <xref ref-type="bibr" rid="CR22">2003</xref>). Queens produce QMP, known to release a worker retinue response and impact worker behavior through induced changes to their endocrine and reproductive physiology (De Groot and Voogd <xref ref-type="bibr" rid="CR13">1954</xref>; Jay <xref ref-type="bibr" rid="CR28">1970</xref>, <xref ref-type="bibr" rid="CR29">1972</xref>; Jay and Jay <xref ref-type="bibr" rid="CR30">1976</xref>; Kaatz et al. <xref ref-type="bibr" rid="CR32">1992</xref>). Both QMP and BEP of older larvae suppress ovary activation and stimulate hypopharyngeal gland (HPG) development of facultatively sterile workers, priming them to forego reproduction and activate both HPG and mandibular glands for brood care (Hoover et al. <xref ref-type="bibr" rid="CR22">2003</xref>; Mohammedi et al. <xref ref-type="bibr" rid="CR52">1996</xref>, <xref ref-type="bibr" rid="CR53">1998</xref>; Peters et al. <xref ref-type="bibr" rid="CR67">2010</xref>). The paired HPGs of nurse-aged bees produce the protein-rich food fed to developing larvae (Snodgrass <xref ref-type="bibr" rid="CR84">1925</xref>). To activate their HPG, bees normally must consume protein and have contact with larvae for 3&#x000a0;days (Huang et al. <xref ref-type="bibr" rid="CR26">1989</xref>; Huang and Otis <xref ref-type="bibr" rid="CR25">1989</xref>). Young adult bees receive proportionally more brood food from nurse-aged bees than older bees (Crailsheim <xref ref-type="bibr" rid="CR10">1991</xref>, <xref ref-type="bibr" rid="CR11">1992</xref>). This protein-rich diet can trigger activation of the HPGs in young workers while poor worker nutrition negatively impacts HPG development (Peters et al. <xref ref-type="bibr" rid="CR67">2010</xref>). A restricted diet also suppresses ovary activation, because bees do not have the protein resources to develop oocytes (Hoover et al. <xref ref-type="bibr" rid="CR23">2006</xref>; Lin and Winston <xref ref-type="bibr" rid="CR45">1998</xref>). Recent research has shown that simultaneous exposure to QMP and BEP even in the absence of a protein resource can increase protein production in HPGs (Peters et al. <xref ref-type="bibr" rid="CR67">2010</xref>) suggesting that under the queenright conditions of a hive environment (i.e., presence of queen and brood pheromones), workers can catabolize bodily proteins for larval food production. BEP stimulates increased pollen foraging, which is directly canalized into rearing more brood (Sagili and Pankiw <xref ref-type="bibr" rid="CR76">2009</xref>; Sagili et al. <xref ref-type="bibr" rid="CR77">2011</xref>).</p><p id="Par5">The reproductive ground plan hypothesis proposes that reproductive physiology provided building blocks on which natural selection acted to establish a foraging division of labor. As originally proposed, bees with more ovarioles and higher titers of the egg yolk precursor vitellogenin bias their foraging toward pollen collection used for larval rearing, repurposing reproductive traits to establish a division of labor (Amdam et al. <xref ref-type="bibr" rid="CR4">2004</xref>, <xref ref-type="bibr" rid="CR5">2006</xref>; Page <xref ref-type="bibr" rid="CR55">2013</xref>; Page and Amdam <xref ref-type="bibr" rid="CR56">2007</xref>). Similarly, queen and larval cues have been modified by natural selection into effective pheromone signals that help coordinate brood care and impact the same fundamental building blocks of reproductive physiology. Queens influence worker behavior via QMP, stimulating retinue behavior (Keeling et al. <xref ref-type="bibr" rid="CR34">2003</xref>) and upregulating worker genes tied to nursing (Whitfield et al. <xref ref-type="bibr" rid="CR91">2003</xref>). Larvae, similarly dependent on care from the workers, influence worker behavior via brood pheromones, increasing protein foraging required for brood food production (Pankiw et al. <xref ref-type="bibr" rid="CR65">1998</xref>; Traynor <xref ref-type="bibr" rid="CR86">2014</xref>). Thus, both queen and larval pheromones suppress ovary development and enhance nurse physiology, suggesting that nursing and reproductive physiology are intimately linked as proposed by the reproductive ground plan.</p><p id="Par6">The effects of BEP on honey bee physiology have been well-investigated, but less is known about the priming effects of the volatile young larval pheromone e&#x003b2;. How e&#x003b2; interacts with QMP remains unknown. Pheromones are often context specific and may require the natural conditions of the hive to trigger physiological responses; however, studying the effects of pheromones on the physiology of workers in the context of the hive creates unique obstacles due to trophallactic transmission of pheromone signals among nestmates (Korst and Velthuis <xref ref-type="bibr" rid="CR37">1982</xref>; Leoncini et al. <xref ref-type="bibr" rid="CR44">2004</xref>), the impact of feeding larvae on worker physiology (Amdam et al. <xref ref-type="bibr" rid="CR6">2009</xref>), and the impact of the external environment on developmental maturation and resource foraging (Dreller et al. <xref ref-type="bibr" rid="CR15">1999</xref>). We thus resolved to study the effects of e&#x003b2; on the physiology of nurse-aged bees in the laboratory while mimicking the conditions of a natural hive in a controlled cage setting.</p><p id="Par7">In order to test the interactive effects of e&#x003b2; and QMP on nursing and reproductive physiology of nurse-aged bees (10&#x000a0;days) in a tightly controlled environment, we ran three preliminary experiments to eliminate potential confounding factors of synthetic QMP, diet, and e&#x003b2; dose on HPG development and ovary activation. We first addressed an earlier controversy (Willis et al. <xref ref-type="bibr" rid="CR92">1990</xref>; Winston and Slessor <xref ref-type="bibr" rid="CR94">1998</xref>) on the ability of synthetic QMP to significantly suppress ovary activation by comparing the effects of live mated queens, virgin queens, and synthetic QMP on ovary activation (experiment 1). Virgin queens do not emit the full suite of pheromones of a mated queen, lacking emission of e&#x003b2; and significantly differing in quantities of other pheromone components compared to mated queens (Gilley et al. <xref ref-type="bibr" rid="CR17">2006</xref>; Richard et al. <xref ref-type="bibr" rid="CR71">2007</xref>). Bees can only activate their HPGs and ovaries with sufficient access to protein-rich food, but an excess of protein increases mortality (Altaye et al. <xref ref-type="bibr" rid="CR2">2010</xref>; Pirk et al. <xref ref-type="bibr" rid="CR70">2010</xref>). In the hive, newly emerged bees are fed royal jelly by nurses. We hypothesized that royal jelly incorporated into the diet at 10&#x000a0;% could stimulate HPG development in young bees without increasing mortality, substituting for access to nurse bees (experiment 2). Next, we investigated the effects of high versus low e&#x003b2; dose on HPG development and ovary activation (experiment 3), since brood pheromones have often produced dose-dependent results (Mohammedi et al. <xref ref-type="bibr" rid="CR53">1998</xref>; Sagili et al. <xref ref-type="bibr" rid="CR77">2011</xref>). Finally, we tested the effects of e&#x003b2; and QMP in combination on nurse-aged bees, to see if the queen and young larval brood pheromones had interactive effects on HPG development and ovary activation&#x000a0;(experiment 4), key components of nursing, and reproductive physiology. Our hypotheses were that (1) QMP would significantly suppress ovary activation compared to controls; (2) 10&#x000a0;% royal jelly would be sufficient to develop HPG without activating ovaries or increasing mortality; (3) the high dose of e&#x003b2; would significantly stimulate HPG development for nursing and suppress ovary activation; and (4) both e&#x003b2; and QMP would synergistically suppress ovary activation and enhance HPG development, thus stimulating the development of the nurse bee phenotype primed for caring of her sisters instead of reproduction.</p><sec id="Sec1" sec-type="materials|methods"><title>Materials and methods</title><sec id="Sec2"><title>Bees</title><p id="Par8">For each experiment, combs of capped honey bee mature pupae were removed from five wild-type colonies of <italic>A. mellifera ligustica</italic> headed by commercial queens purchased from northern California and placed in an incubator at 34&#x000a0;&#x000b0;C in cages. The following morning, newly emerged bees less than 18&#x000a0;h old were collected, and six replicates were established. Within a single replicate, bees were randomly selected from only two of the five colonies, caged, and the cage was randomly assigned to a treatment group, so that each replicate was composed of two randomly selected genetic families (genotypes). Thus, replicate encompasses genetic variance between and among colonies. One hundred newly emerged bees were paint marked on the thorax according to treatment and placed in an acrylic cage similar in design to the pain cage (Pain <xref ref-type="bibr" rid="CR61">1966</xref>) with the addition of a divider that split the cage in half. This divider was either made of single mesh to provide access to nurse bees or was solid. The cages ensured that the pheromones and diet were distributed among all members via trophallaxis and removed additional pheromone exposure from other colony sources. The cages were maintained at 30&#x02009;&#x000b1;&#x02009;3&#x000a0;&#x000b0;C and 35&#x02009;&#x000b1;&#x02009;4&#x000a0;% humidity in individual, disposable incubators assembled from wax-coated cardboard. Cages for each treatment group were kept together&#x000a0;in a vented fume hood&#x000a0;with a radiant heat source. The bees were fed <italic>ad libitum</italic> with water, queen or royal jelly candy, and pollen paste, replaced every 1&#x02013;2&#x000a0;days as necessary. Queen candy was made from 80&#x000a0;% powdered sugar and 20&#x000a0;% honey produced by our apiaries in Arizona. Royal jelly candy was made from 10&#x000a0;% royal jelly, 10&#x000a0;% honey, and 80&#x000a0;% powdered sugar on a <italic>w</italic>/<italic>w</italic> basis. Pollen paste was made from frozen pollen pellets (Crockett Honey, Tempe, AZ) ground and mixed with distilled water until it had the consistency of dough.</p></sec><sec id="Sec3"><title>Data collection</title><p id="Par9">Bee mortality was recorded daily. After 10&#x000a0;days, the cages of bees were frozen, and for each cage, six to ten bees were randomly selected, dissected, and evaluated for HPG development, total number of ovarioles comprising each ovary and ovary activation. Ovarioles were counted because ovariole number is positively correlated with behavior and ovary activation (Amdam et al. <xref ref-type="bibr" rid="CR4">2004</xref>, <xref ref-type="bibr" rid="CR5">2006</xref>).</p></sec><sec id="Sec4"><title>Dissections</title><p id="Par10">Bees typically transition out of the brood nest and into other in-hive tasks at 10&#x02013;12&#x000a0;days of age (R&#x000f6;sch <xref ref-type="bibr" rid="CR74">1930</xref>; Seeley <xref ref-type="bibr" rid="CR80">1982</xref>, <xref ref-type="bibr" rid="CR81">1995</xref>; Seeley and Kolmes <xref ref-type="bibr" rid="CR82">1991</xref>). Worker HPG reach peak development at 6&#x000a0;days, then typically diminish in size by 15&#x000a0;days of age and atrophy as bees transition to foraging (Deseyn and Billen <xref ref-type="bibr" rid="CR14">2005</xref>). As we were interested in the impacts of e&#x003b2; on nurse bee physiology, dissections were conducted on bees at 10&#x000a0;days of age.</p><p id="Par11">Both HPGs were dissected from the head capsule and placed into a drop of saline (0.25&#x000a0;mol/l NaCl) on a microscope slide. A representative section was examined at 100&#x000d7;. The activity of HPGs is positively correlated with size (Knecht and Kaatz <xref ref-type="bibr" rid="CR35">1990</xref>). Numerous globular acini attach to the long, slender main channel of the HPG, and these acini increase in diameter until 6&#x000a0;days of age, when they begin to shrink (Deseyn and Billen <xref ref-type="bibr" rid="CR14">2005</xref>). The gland continues to diminish, so that by 15&#x000a0;days of age, when bees typically transition to foraging, their size corresponds to the still undeveloped gland of newly emerged bees (Deseyn and Billen <xref ref-type="bibr" rid="CR14">2005</xref>). HPG development was thus rated using an established scale (Hess <xref ref-type="bibr" rid="CR21">1942</xref>), which uses the shape and density of the acini as the main criterion for classification and ranks them from by stage of development: (1) atrophied; (2) slightly swollen with noticeable spacing between acini; (3) swollen with small spacing between acini, capable of producing brood food; and (4) fully developed and tightly clustered, channel obscured by acini. Glands were additionally assigned to one of three classes according to lobe morphology (Wegener et al. <xref ref-type="bibr" rid="CR90">2009</xref>) as models predict that e&#x003b2; can accelerate behavioral maturation (Maisonnasse et al. <xref ref-type="bibr" rid="CR50">2010</xref>) and we wanted to determine if forager HPG morphology was present&#x000a0;in our nurse-aged bees. Class 1, typical of young broodless workers, consists of glands with small acini showing an uneven surface. Class 2, representative of active nurse bees, is composed of medium-sized to large acini with a smooth surface and numerous secretory vesicles, giving them a yellowish color. Class 3 glands, representative of older foragers, consist of large, but slightly pale and translucent lobes. Class 3 was not found among our samples.</p><p id="Par12">Both ovaries were removed from the bees and placed in a drop of saline. The number of ovary filaments (ovarioles) was counted using a 100&#x000d7; dissecting microscope (Zeiss, Jena, Germany). The stage of ovary activation was classified using an established scale (Pernal and Currie <xref ref-type="bibr" rid="CR66">2000</xref>); similar to the&#x000a0;4-point scale of Hess (<xref ref-type="bibr" rid="CR21">1942</xref>) except absence of activation is scored as a 0 instead of 1: 0, no follicle development; 1, slight enlargement; 2, presence of distinct cells leading to swellings and constrictions; 3, egg volume exceeding that of the nutritive follicle; 4, presence of fully formed eggs. For both HPG development and ovary activation, the most developed score of the pair of organs was used for statistical analyses, as occasionally, there were disparities within a bee.</p></sec><sec id="Sec5"><title>Treatments</title><sec id="Sec6"><title>Experiment 1: queen comparison</title><p id="Par13">To determine if synthetic QMP was as effective as a live queen in suppressing ovary activation, we compared cages subjected to one of five treatments: (1) mated queen; (2) virgin queen; (3) virgin queen subjected to two successive CO₂ treatments, which results in oviposition within a few days despite the lack of mating flight (Mackensen <xref ref-type="bibr" rid="CR48">1947</xref>); (4) one slow release strip of synthetic QMP (PseudoQueen, formerly known as BeeBoost, Contech Industries, Victoria, British Columbia) attached near the top of the cage using a plastic zip tie to simulate a queen; or (5) control which received no queen or synthetic QMP. The live queens in the first three treatment groups were unconfined and free to interact with the workers as in a natural colony. No comb was included in the cages to prevent egg laying and rearing of larvae.</p></sec><sec id="Sec7"><title>Experiment 2: royal jelly compared to nurse bee environment</title><p id="Par14">To determine if direct access to royal jelly was sufficient to activate HPG development or if newly emerged bees required contact with nurse bees, cages either received 10&#x000a0;% royal jelly (RJ) candy incorporated into queen candy or had access to 100 nurse bees (N) through a single mesh screen. Nurse bees were collected from a comb of open larvae in wild-type colonies, where they were actively engaged in nursing behavior. Each cage also received a synthetic QMP strip as in experiment 1 to mimic in-hive conditions and replicate conditions of future experiments. Since QMP suppresses ovary activation and RJ incorporated into the diet has previously been linked with ovary activation (Altaye et al. <xref ref-type="bibr" rid="CR2">2010</xref>; Lin and Winston <xref ref-type="bibr" rid="CR45">1998</xref>; Pirk et al. <xref ref-type="bibr" rid="CR70">2010</xref>), we included an additional treatment without QMP as a baseline comparison for ovary activation (OA).</p></sec><sec id="Sec8"><title>Experiment 3: high versus low e-beta ocimene dose</title><p id="Par15">Live larvae suppress OA in attending worker bees via larval pheromones, though the effectiveness of pheromones is often dose dependent (Maisonnasse et al. <xref ref-type="bibr" rid="CR49">2009</xref>; Mohammedi et al. <xref ref-type="bibr" rid="CR53">1998</xref>). To confirm that e&#x003b2; can suppress OA, we subjected each cage to one of three treatments: (1) low e&#x003b2; dose of one larval equivalent (Leq)/bee; (2) high e&#x003b2; dose of 10 Leq/bee; (3) carrier control. Due to the high volatility of e&#x003b2; and in order to avoid pheromone saturation in the cages, the molecule was mixed with 1-ml paraffin oil, and a similar droplet was used as the control (Maisonnasse et al. <xref ref-type="bibr" rid="CR49">2009</xref>). Treatments were supplied in a mesh screened glass Petri dish below the screened floor of the cage; so, bees could not contact the chemicals directly (Maisonnasse et al. <xref ref-type="bibr" rid="CR50">2010</xref>). Treatments were replaced daily. Each cage received RJ candy as their carbohydrate source.</p></sec><sec id="Sec9"><title>Experiment 4: e&#x003b2; and QMP synergy</title><p id="Par16">Pheromones are often context specific and interact with other pheromone components. To determine if e&#x003b2; and QMP have interactive effects, each cage was subjected to one of four treatments: (1) e&#x003b2;&#x02212;/QMP&#x02212;, (2) e&#x003b2;&#x02212;/QMP+, (3) e&#x003b2;+/QMP&#x02212;; and (4) e&#x003b2;+/QMP+. The e&#x003b2; was supplied at 10 Leq/bee in 1-ml paraffin oil as in experiment 3. The QMP was supplied in a slow release strip of synthetic QMP (PseudoQueen, Contech Industries), as in experiments 1 and 2. Each cage received RJ candy as their carbohydrate source.</p></sec></sec><sec id="Sec10"><title>Statistics</title><p id="Par17">Daily mortality was compared using repeated measures MANOVA with replicate and treatment as factors. Total ovarioles, OA, and HPG development were compared using two-way ANOVA with replicate and treatment as factors. Bivariate correlations for total ovarioles, OA, and HPG development were calculated using nonparametric Spearman&#x02019;s rank correlations. All calculations were performed using JMP Pro 10.0.0 (SAS, Cary, NC).</p></sec></sec><sec id="Sec11" sec-type="results"><title>Results</title><sec id="Sec12"><title>Experiment 1: queen comparison</title><p id="Par18">We compared the effects of synthetic QMP and live queens on mortality and ovarian status in caged worker bees. The mated queen in replicate 2 died on day 6 of the experiment, and the cage was excluded from analysis. Mortality was significantly affected by treatment (Fig.&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>; <italic>F</italic> _4,20&#x02009;=&#x02009;0.854, <italic>p</italic>&#x02009;=&#x02009;0.012) and age (<italic>F</italic> _8,13&#x02009;=&#x02009;3.616, <italic>p</italic>&#x02009;=&#x02009;0.003). Control and QMP cages had significantly higher mortality than the treatments that received a live queen (<italic>t</italic>&#x02009;&#x0003e;&#x02009;2.50, <italic>p</italic>&#x02009;&#x0003c;&#x02009;0.013), though mean mortality never exceeded 1 bee/day for any of the treatment groups.<fig id="Fig1"><label>Fig. 1</label><caption><p> <italic>Experiment 1</italic> queen comparison cumulative mortality (+S.E.) per cage over 10&#x000a0;days. Control (<italic>blue</italic>), QMP (<italic>red</italic>)&#x02009;=&#x02009;synthetic strip of queen mandibular pheromone; VirginQueen (<italic>green</italic>)&#x02009;=&#x02009;virgin queen; Virgin CO2Queen (<italic>purple</italic>)&#x02009;=&#x02009;virgin queen exposed to two treatments of CO₂; MatedQueen (<italic>tan</italic>)&#x02009;=&#x02009;mated queen</p></caption><graphic xlink:href="265_2014_1811_Fig1_HTML" id="MO1"/></fig> </p><p id="Par19">OA differed significantly by both treatment (Fig.&#x000a0;<xref rid="Fig2" ref-type="fig">2a</xref>; <italic>F</italic> _{4, 149}&#x02009;=&#x02009;15.506, <italic>p</italic>&#x02009;&#x0003c;&#x02009;0.001) and replicate (Fig.&#x000a0;<xref rid="Fig7" ref-type="supplementary-material">S1</xref>; <italic>F</italic> _5,149&#x02009;=&#x02009;4.476, <italic>p</italic>&#x02009;=&#x02009;0.002), and there was a significant interaction of these two factors (Fig.&#x000a0;<xref rid="Fig2" ref-type="fig">2b</xref>; replicate 2 excluded; <italic>F</italic> _16,125&#x02009;=&#x02009;2.55, <italic>p</italic>&#x02009;=&#x02009;0.002). Bees with more ovarioles often activate their ovaries more readily, as seen in this experiment (Fig.&#x000a0;<xref rid="Fig2" ref-type="fig">2c</xref>; <italic>F</italic> _1,172&#x02009;=&#x02009;5.92, <italic>p</italic>&#x02009;=&#x02009;0.016). Bees with large ovaries (eight or more ovarioles) had significantly more activated ovaries than bees with small ovaries (<italic>t</italic>&#x02009;=&#x02009;2.43, <italic>p</italic>&#x02009;=&#x02009;0.016). Ovary size (large vs small) significantly influenced the effect of treatment on OA (Fig.&#x000a0;<xref rid="Fig2" ref-type="fig">2d</xref>; <italic>F</italic> _4,164&#x02009;=&#x02009;2.67, <italic>p</italic>&#x02009;=&#x02009;0.034). Regardless of ovary size, bees in the control group had significantly greater OA compared to the four other queen treatments (<italic>t</italic>&#x02009;&#x0003e;&#x02009;3.00, <italic>p</italic>&#x02009;&#x0003c;&#x02009;0.003). In bees with small ovaries, QMP suppressed OA as well as a live queen; however, in bees with large ovaries, QMP was not as effective as a live queen (<italic>t</italic>&#x02009;&#x0003e;&#x02009;2.40 1.71, <italic>p</italic>&#x02009;&#x0003c;&#x02009;0.02). Total ovarioles and OA are significantly positively correlated in bees reared in the control (Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>; <italic>&#x003c1;</italic>&#x02009;=&#x02009;0.430, <italic>n</italic>&#x02009;=&#x02009;30, <italic>p</italic>&#x02009;=&#x02009;0.018) environment, but not in any of the treatments with a live queen or with synthetic QMP (Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>; <italic>&#x003c1;</italic>&#x02009;=&#x02009;0.309, <italic>n</italic>&#x02009;=&#x02009;30, <italic>p</italic>&#x02009;=&#x02009;0.097).<fig id="Fig2"><label>Fig. 2</label><caption><p> <italic>Experiment 1</italic> queen comparison ovary activation. <bold>a</bold> Mean (+S.E.) ovary activation by treatment; <bold>b</bold> mean (+S.E.) ovary activation by replicate; <bold>c</bold> mean (+S.E.) total ovarioles by ovary state; <bold>d</bold> mean (+S.E.) ovary activation by ovary size and treatment. QMP&#x02009;=&#x02009;synthetic queen mandibular pheromone; CO2&#x02009;=&#x02009;virgin queen treated 2&#x000d7; with CO₂; large ovary&#x02009;=&#x02009;(8 or more ovarioles); small ovary (&#x0003c;8 ovarioles). <italic>N</italic>&#x02009;=&#x02009;180 bees, 36 per treatment, 30 per replicate, 10 bees per cage. <italic>Different letters</italic> indicate significant differences using LSD student <italic>t</italic> tests</p></caption><graphic xlink:href="265_2014_1811_Fig2_HTML" id="MO2"/></fig> <table-wrap id="Tab1"><label>Table 1</label><caption><p>Significant correlations by experiment</p></caption><table frame="hsides" rules="groups"><thead><tr><th>Experiment</th><th>Treatments</th><th>Ovary activation and total ovarioles</th><th>Ovary activation and HPG development</th><th>Total ovarioles and HPG development</th></tr></thead><tbody><tr><td rowspan="5">1</td><td>C</td><td>+</td><td>n/a</td><td>n/a</td></tr><tr><td>QMP</td><td>NS</td><td>n/a</td><td>n/a</td></tr><tr><td>CO2</td><td>NS</td><td>n/a</td><td>n/a</td></tr><tr><td>VQ</td><td>NS</td><td>n/a</td><td>n/a</td></tr><tr><td>MQ</td><td>NS</td><td>n/a</td><td>n/a</td></tr><tr><td rowspan="2">2</td><td>RJ</td><td>NS</td><td>+</td><td>NS</td></tr><tr><td>N</td><td>+</td><td>NS</td><td>NS</td></tr><tr><td rowspan="3">3</td><td>C</td><td>NS</td><td>NS</td><td>NS</td></tr><tr><td>Low e&#x003b2;</td><td>+</td><td>NS</td><td>NS</td></tr><tr><td>High e&#x003b2;</td><td>+</td><td>+</td><td>NS</td></tr><tr><td rowspan="4">4</td><td>e&#x003b2;&#x02212;/QMP&#x02212;</td><td>NS</td><td>NS</td><td>NS</td></tr><tr><td>e&#x003b2;&#x02212;/QMP+</td><td>+</td><td>NS</td><td>NS</td></tr><tr><td>e&#x003b2;+/QMP&#x02212;</td><td>+</td><td>+</td><td>NS</td></tr><tr><td>e&#x003b2;+/QMP+</td><td>NS</td><td>NS</td><td>NS</td></tr></tbody></table><table-wrap-foot><p>Significant correlations between total ovarioles, ovary activation, and HPG development are given for each of the four experiments that are indicated. Significant correlations are indicated by + or &#x02212;, depending on relationship. Untested correlations because the HPG were not dissected are indicated by n/a.</p></table-wrap-foot></table-wrap> </p></sec><sec id="Sec13"><title>Experiment 2: royal jelly compared to nurse bee environment</title><p id="Par20">In a colony, young bees are fed protein-rich RJ from nurse bees (Crailsheim <xref ref-type="bibr" rid="CR10">1991</xref>, <xref ref-type="bibr" rid="CR11">1992</xref>), which may impact survivorship and promote development of both the ovaries and HPGs. We investigated the effects of access to nurse bees versus direct access to RJ. Because RJ can stimulate OA and QMP suppresses OA, we included a third treatment group without nurse bees, RJ, or QMP as a baseline comparison for OA. Mortality remained below 1 bee per day, and there was no significant difference in mortality between the RJ and nurse bee (N) treatment groups (<italic>F</italic> _1,5&#x02009;=&#x02009;0.154, <italic>p</italic>&#x02009;=&#x02009;0.421).</p><p id="Par21">OA did not differ between RJ and N treatment groups (<italic>t</italic>&#x02009;=&#x02009;1.16, <italic>p</italic>&#x02009;=&#x02009;0.247); there was a significant difference by replicate (<italic>F</italic> _5,114&#x02009;=&#x02009;2.41, <italic>p</italic>&#x02009;=&#x02009;0.041), but only replicates 1 and 4 were significantly different (<italic>t</italic>&#x02009;=&#x02009;2.96, <italic>p</italic>&#x02009;=&#x02009;0.004). The OA treatment group differed significantly from both RJ and N (Fig.&#x000a0;<xref rid="Fig3" ref-type="fig">3a</xref>; RJ <italic>t</italic>&#x02009;=&#x02009;5.56, <italic>p</italic>&#x02009;&#x0003c;&#x02009;0.001; N <italic>t</italic>&#x02009;=&#x02009;4.51, <italic>p</italic>&#x02009;&#x0003c;&#x02009;0.001). Ovary size significantly influenced OA in bees exposed to nurses (<italic>t</italic>&#x02009;=&#x02009;2.80, <italic>p</italic>&#x02009;=&#x02009;0.007) but had no effect in the two other treatment groups (Fig.&#x000a0;<xref rid="Fig3" ref-type="fig">3b</xref>).<fig id="Fig3"><label>Fig. 3</label><caption><p> <italic>Experiment 2</italic> access to royal jelly versus nurse bees. <bold>a</bold> Mean (+S.E.) ovary activation by treatment; <bold>b</bold> mean (+S.E.) ovary activation by treatment and ovary size; <bold>c</bold> mean HPG development by treatment. <italic>N</italic>&#x02009;=&#x02009;bees exposed to candy, pollen, QMP, and 100 nurses; RJ&#x02009;=&#x02009;bees exposed to pollen, QMP and 10&#x000a0;% royal jelly incorporated into the queen candy; OA&#x02009;=&#x02009;an ovary activation treatment group without QMP as a baseline comparison. <italic>N</italic>&#x02009;=&#x02009;180 bees, 60 per treatment, 30 per replicate, 10 bees per cage. <italic>Different letters</italic> indicate significant differences using LSD student <italic>t</italic> tests</p></caption><graphic xlink:href="265_2014_1811_Fig3_HTML" id="MO3"/></fig> </p><p id="Par22">Nurse-aged bees typically have well-developed HPGs, needed to produce the protein-rich food they feed to larvae. RJ significantly increased HPG development compared to N (Fig.&#x000a0;<xref rid="Fig3" ref-type="fig">3c</xref>; <italic>t</italic>&#x02009;=&#x02009;3.69, <italic>p</italic>&#x02009;&#x0003c;&#x02009;0.001). Replicate had a significant impact on HPG development (<italic>F</italic> _5,108&#x02009;=&#x02009;6.24, <italic>p</italic>&#x02009;&#x0003c;&#x02009;0.001). HPG development and OA were significantly correlated for bees reared with RJ (<italic>&#x003c1;</italic>&#x02009;=&#x02009;0.259, <italic>n</italic>&#x02009;=&#x02009;60, <italic>p</italic>&#x02009;=&#x02009;0.046) but were not significant in bees with access to nurse bees (<italic>&#x003c1;</italic>&#x02009;=&#x02009;0.068, <italic>n</italic>&#x02009;=&#x02009;60, <italic>p</italic>&#x02009;=&#x02009;0.604). Total ovarioles and OA were positively correlated in the bees with access to nurse bees (<italic>&#x003c1;</italic>&#x02009;=&#x02009;0.348, <italic>n</italic>&#x02009;=&#x02009;60, <italic>p</italic>&#x02009;=&#x02009;0.006).</p></sec><sec id="Sec14"><title>Experiment 3: high versus low e-beta ocimene dose</title><p id="Par23">Mortality did not differ significantly by treatment (<italic>F</italic> _2,10&#x02009;=&#x02009;0.725, <italic>p</italic>&#x02009;=&#x02009;0.066) or age (<italic>F</italic> _8,3&#x02009;=&#x02009;11.078, <italic>p</italic>&#x02009;=&#x02009;0.134) but varied significantly by replicate (<italic>F</italic> _5,10&#x02009;=&#x02009;1.950, <italic>p</italic>&#x02009;=&#x02009;0.032).</p><p id="Par24">Treatment significantly impacted OA (Fig&#x000a0;<xref rid="Fig4" ref-type="fig">4a</xref>; <italic>F</italic> _2,162&#x02009;=&#x02009;20.73, <italic>p</italic>&#x02009;&#x0003c;&#x02009;0.001). Bees that received the high e&#x003b2; dose of 10 Leq/bee had significantly fewer developing oocytes than bees in the control group or receiving the low dose of 1 Leq/bee (<italic>t</italic>&#x02009;&#x0003e;&#x02009;4.40, <italic>p</italic>&#x02009;&#x0003c;&#x02009;0.001). As above, bees with more ovarioles had significantly more activated ovaries than bees with fewer ovarioles (Fig.&#x000a0;<xref rid="Fig4" ref-type="fig">4b</xref>; <italic>t</italic>&#x02009;=&#x02009;3.62, <italic>p</italic>&#x02009;&#x0003c;&#x02009;0.001). Bees with activated ovaries had significantly more ovarioles in both the control (Fig.&#x000a0;<xref rid="Fig4" ref-type="fig">4c</xref>; <italic>t</italic>&#x02009;=&#x02009;2.38, <italic>p</italic>&#x02009;=&#x02009;0.035) and low e&#x003b2; (<italic>t</italic>&#x02009;=&#x02009;3.59, <italic>p</italic>&#x02009;&#x0003c;&#x02009;0.001) treatment groups, but not in the high e&#x003b2; group (<italic>t</italic>&#x02009;=&#x02009;0.20, <italic>p</italic>&#x02009;=&#x02009;0.844). In bees exposed to the high dose, 50&#x000a0;% of bees had at least one ovary with OA at or above stage 1 (slight ovariole swelling) and 3&#x000a0;% at or above stage 2 compared to 97 and 18&#x000a0;%, respectively, in the low dose and 92 and 7&#x000a0;% in the control.<fig id="Fig4"><label>Fig. 4</label><caption><p> <italic>Experiment 3</italic> e&#x003b2; dose. <bold>a</bold> Mean (+S.E.) ovary activation by treatment; <bold>b</bold> mean (+S.E.) ovarioles by ovary state; <bold>c</bold> mean (+S.E.) ovarioles by ovary state and treatment; <bold>d</bold> mean (+S.E.) HPG by treatment and ovary size. C&#x02009;=&#x02009;control; Lo&#x02009;=&#x02009;low e&#x003b2; (1 Leq/bee); Hi&#x02009;=&#x02009;high e&#x003b2; (10 Leq/bee); activated&#x02009;=&#x02009;one ovary at stage 2 or more; large&#x02009;=&#x02009;8 or more ovarioles; small&#x02009;=&#x02009;&#x0003c;8 ovarioles; <italic>N</italic>&#x02009;=&#x02009;180 bees, 60 per treatment, 30 per replicate, 10 bees per cage. <italic>Different letters</italic> indicate significant differences using LSD student <italic>t</italic> tests</p></caption><graphic xlink:href="265_2014_1811_Fig4_HTML" id="MO4"/></fig> </p><p id="Par25">There was no significant effect of treatment on HPG development, indicating that e&#x003b2; did not increase HPG development compared to controls (<italic>F</italic> _2,162&#x02009;=&#x02009;1.06, <italic>p</italic>&#x02009;=&#x02009;0.348); HPG development differed across replicates (<italic>F</italic> _5,162&#x02009;=&#x02009;4.34, <italic>p</italic>&#x02009;=&#x02009;0.001). In the groups treated with e&#x003b2;, bees with large ovaries had significantly more developed HPG than bees with small ovaries (<italic>t</italic>&#x02009;=&#x02009;2.59, <italic>p</italic>&#x02009;=&#x02009;0.01), but there was no difference in the control group (<italic>t</italic>&#x02009;=&#x02009;0.56, <italic>p</italic>&#x02009;=&#x02009;0.57), see Fig.&#x000a0;<xref rid="Fig4" ref-type="fig">4d</xref>. In bees with large ovaries treated with high e&#x003b2;, 60.6&#x000a0;% had developed HPG glands capable of nursing (stage 3 or 4) compared to 40.7&#x000a0;% of the bees with small ovaries. A similar trend was seen with the low e&#x003b2; dose, where 56.4&#x000a0;% of bees with large ovaries had well-developed HPG, compared to 28.5&#x000a0;% with small ovaries (see Fig.&#x000a0;<xref rid="Fig5" ref-type="fig">5</xref>). Thus, total ovarioles were significantly correlated with OA for bees reared in the low e&#x003b2; (<italic>&#x003c1;</italic>&#x02009;=&#x02009;0.432, <italic>n</italic>&#x02009;=&#x02009;60, <italic>p</italic>&#x02009;&#x0003c;&#x02009;0.001) and high e&#x003b2; environment (<italic>&#x003c1;</italic>&#x02009;=&#x02009;0.341, <italic>n</italic>&#x02009;=&#x02009;60, <italic>p</italic>&#x02009;=&#x02009;0.008), but not in the control group (total ovarioles <italic>&#x003c1;</italic>&#x02009;=&#x02009;0.212, <italic>n</italic>&#x02009;=&#x02009;60, <italic>p</italic>&#x02009;=&#x02009;0.104). Pairwise correlations show that OA and HPG development are positively correlated in the high e&#x003b2; environment (<italic>r</italic>&#x02009;=&#x02009;0.270, <italic>n</italic>&#x02009;=&#x02009;60, <italic>p</italic>&#x02009;=&#x02009;0.037), but the correlation is not significant when converted to nonparametric ranks (<italic>&#x003c1;</italic>&#x02009;=&#x02009;0.243, <italic>n</italic>&#x02009;=&#x02009;60, <italic>p</italic>&#x02009;=&#x02009;0.062).<fig id="Fig5"><label>Fig. 5</label><caption><p> <italic>Experiment 3</italic> e&#x003b2; dose. HPG development by ovary size for bees treated with low e&#x003b2; (<italic>left</italic>) and high e&#x003b2; (<italic>right</italic>). HPG stages 1 and 2 (<italic>light grays</italic>) are incapable of nursing, while stages 3 and 4 (<italic>dark greys</italic>) are capable of nursing. Large (<italic>top row</italic>)&#x02009;=&#x02009;8 or more ovarioles; small (<italic>bottom row</italic>)&#x02009;=&#x02009;&#x0003c;8 ovarioles</p></caption><graphic xlink:href="265_2014_1811_Fig5_HTML" id="MO5"/></fig> </p></sec><sec id="Sec15"><title>Experiment 4: e&#x003b2; and QMP synergy</title><p id="Par26">Having established that both QMP and the high dose of e&#x003b2; significantly suppresses OA compared to controls, we tested the interactive effects of e&#x003b2; and QMP. Mortality did not differ significantly by treatment (<italic>F</italic> _3,15&#x02009;=&#x02009;0.182, <italic>p</italic>&#x02009;=&#x02009;0.460) but differed significantly by replicate (Fig.&#x000a0;<xref rid="Fig9" ref-type="supplementary-material">S3</xref>; <italic>F</italic> _5,15&#x02009;=&#x02009;1.107, <italic>p</italic>&#x02009;&#x0003c;&#x02009;0.032) and age (<italic>F</italic> _7,9&#x02009;=&#x02009;3.585, <italic>p</italic>&#x02009;&#x0003c;&#x02009;0.019).</p><p id="Par27">Total ovariole number per bee did not differ by treatment (<italic>F</italic> _3,216&#x02009;=&#x02009;0.30, <italic>p</italic>&#x02009;=&#x02009;0.822) but varied significantly by replicate (<italic>F</italic> _5,216&#x02009;=&#x02009;2.53, <italic>p</italic>&#x02009;=&#x02009;0.030). Replicate 3 had significantly more ovarioles than replicates 1, 4, 5, and 6. Treatment significantly impacted OA (Fig.&#x000a0;<xref rid="Fig6" ref-type="fig">6a</xref>; <italic>F</italic> _3,216&#x02009;=&#x02009;17.73, <italic>p</italic>&#x02009;&#x0003c;&#x02009;0.001). Bees reared with e&#x003b2; had significantly less-developed ovaries than bees reared without e&#x003b2; (control <italic>t</italic> ₂₁₆&#x02009;&#x0003e;&#x02009;5.39, <italic>p</italic>&#x02009;&#x0003c;&#x02009;0.001; QMP <italic>t</italic> ₂₁₆&#x02009;&#x0003e;&#x02009;3.08, <italic>p</italic>&#x02009;&#x0003c;&#x02009;0.003). The bees reared with QMP and no e&#x003b2; had significantly less-developed ovaries than control bees reared without either pheromone (<italic>t</italic> ₂₁₆&#x02009;=&#x02009;2.31, <italic>p</italic>&#x02009;=&#x02009;0.022). However, bees reared with only QMP had significantly more developed ovaries than bees exposed to e&#x003b2; (e&#x003b2; alone <italic>t</italic> ₂₁₆&#x02009;=&#x02009;4.24, <italic>p</italic>&#x02009;&#x0003c;&#x02009;0.001; e&#x003b2; and QMP <italic>t</italic> ₂₁₆&#x02009;=&#x02009;3.08, <italic>p</italic>&#x02009;=&#x02009;0.002), indicating that e&#x003b2; is more effective at suppressing OA than QMP. In the control group, 82&#x000a0;% of bees had at least stage 1 OA in one ovary, compared to 73&#x000a0;% of the bees exposed only to QMP, 40&#x000a0;% of the bees exposed to only e&#x003b2; and 50&#x000a0;% of the bees exposed to both e&#x003b2; and QMP. OA also differed by replicate (<italic>F</italic> _5,216&#x02009;=&#x02009;17.023, <italic>p</italic>&#x02009;&#x0003c;&#x02009;0.001), seemingly a consequence of differences in total ovarioles as replicates 2 and 3 had the most total ovarioles combined with the most activated ovaries. There was a significant interaction of treatment and replicate (Fig.&#x000a0;<xref rid="Fig6" ref-type="fig">6b</xref>; <italic>F</italic> _15,216&#x02009;=&#x02009;1.74, <italic>p</italic>&#x02009;=&#x02009;0.046). Once again, bees with large ovaries (eight or more ovarioles) had significantly more active ovaries than bees with small ovaries (<italic>t</italic> ₂₂₈&#x02009;=&#x02009;2.72, <italic>p</italic>&#x02009;=&#x02009;0.007), and ovary size was a significant factor of OA in bees treated with only one of the two pheromones (Fig.&#x000a0;<xref rid="Fig6" ref-type="fig">6c</xref>), but not in bees treated with both or in the control group. There were no bees with stage 2 activation in at least one ovary in either e&#x003b2; group. In the QMP-treated group, only bees with significantly more ovarioles were able to activate their ovaries at stage 2 or above (Fig.&#x000a0;<xref rid="Fig6" ref-type="fig">6d</xref>; <italic>t</italic> ₁₁₆&#x02009;=&#x02009;2.13, <italic>p</italic>&#x02009;=&#x02009;0.035), while ovariole number did not influence OA in the control group (<italic>t</italic> ₁₁₆&#x02009;=&#x02009;0.46, <italic>p</italic>&#x02009;=&#x02009;0.649).<fig id="Fig6"><label>Fig. 6</label><caption><p> <italic>Experiment 4</italic> e&#x003b2; and QMP interaction. <bold>a</bold> Mean (+S.E.) ovary activation by treatment; <bold>b</bold> mean (+S.E.) ovary activation by treatment and replicate; <bold>c</bold> mean (+S.E.) ovary activation by ovary size; <bold>d</bold> mean (+S.E.) ovarioles by ovary state and treatment; <bold>e</bold> mean (+S.E.) HPG by treatment and ovary state. e&#x003b2;&#x02009;=&#x02009;e-beta; QMP&#x02009;=&#x02009;synthetic queen mandibular pheromone, large&#x02009;=&#x02009;8 or more ovarioles; small&#x02009;=&#x02009;&#x0003c;8 ovarioles; activated&#x02009;=&#x02009;one ovary at stage 2 or more; for 6e) activated &#x0003e;0&#x02009;=&#x02009;one ovary at stage 1 or more; <italic>N</italic>&#x02009;=&#x02009;240 bees, 60 per treatment, 30 per replicate, 10 bees per cage. <italic>Different letters or connecting bars</italic> indicate significant differences</p></caption><graphic xlink:href="265_2014_1811_Fig6_HTML" id="MO6"/></fig> </p><p id="Par28">HPG stage development did not differ significantly by treatment (<italic>F</italic> _3,216&#x02009;=&#x02009;1.06, <italic>p</italic>&#x02009;=&#x02009;0.365) or replicate (<italic>F</italic> _5,216&#x02009;=&#x02009;1.41, <italic>p</italic>&#x02009;=&#x02009;0.223). Bees with OA above stage 1 had significantly more developed HPG than bees with inactive ovaries across all treatments (<italic>F</italic> _1,232&#x02009;=&#x02009;4.279, <italic>p</italic>&#x02009;=&#x02009;0.024), and this effect was significant within the e&#x003b2;+/QMP&#x02212; treatment group (Fig.&#x000a0;<xref rid="Fig6" ref-type="fig">6e</xref>; <italic>t</italic>&#x02009;=&#x02009;2.291, <italic>p</italic>&#x02009;=&#x02009;0.027), where 50&#x000a0;% of bees with activated ovaries had HPG capable of nursing (stage 3 or 4) compared to 19.5&#x000a0;% of bees with inactive ovaries. Thus, HPG development and OA were significantly correlated for bees reared in the e&#x003b2;+/QMP&#x02212; (<italic>&#x003c1;</italic>&#x02009;=&#x02009;0.295, <italic>n</italic>&#x02009;=&#x02009;60, <italic>p</italic>&#x02009;=&#x02009;0.022), but not in any of the other groups. Total ovarioles and OA also correlated significantly in the e&#x003b2;+/QMP&#x02212; (<italic>&#x003c1;</italic>&#x02009;=&#x02009;0.323, <italic>n</italic>&#x02009;=&#x02009;60, <italic>p</italic>&#x02009;=&#x02009;0.012) and the e&#x003b2;&#x02212;/QMP+ (<italic>&#x003c1;</italic>&#x02009;=&#x02009;0.443, <italic>n</italic>&#x02009;=&#x02009;60, <italic>p</italic>&#x02009;&#x0003c;&#x02009;0.001) environments, but not in the other treatment groups.</p></sec></sec><sec id="Sec16" sec-type="discussion"><title>Discussion</title><p id="Par29">Our results demonstrate how social insect pheromone communication is defined by complexity, context, and dose (Alaux et al. <xref ref-type="bibr" rid="CR1">2010</xref>; Slessor et al. <xref ref-type="bibr" rid="CR83">2005</xref>). Throughout our experiments, QMP significantly suppressed OA in worker bees compared to controls (Figs.&#x000a0;<xref rid="Fig2" ref-type="fig">2a</xref> and <xref rid="Fig6" ref-type="fig">6a</xref>), as did the e&#x003b2; pheromone of young larvae (Figs.&#x000a0;<xref rid="Fig4" ref-type="fig">4a</xref> and <xref rid="Fig6" ref-type="fig">6a</xref>). Our results also show that e&#x003b2; had significant effects on the reproductive and nursing physiology of worker bees, so that bees with more ovarioles had increased OA (Figs.&#x000a0;<xref rid="Fig4" ref-type="fig">4b&#x02013;c</xref> and <xref rid="Fig6" ref-type="fig">6d</xref>) and increased HPG development (Figs.&#x000a0;<xref rid="Fig4" ref-type="fig">4d</xref>, <xref rid="Fig5" ref-type="fig">5</xref>, and <xref rid="Fig6" ref-type="fig">6e</xref>). This trend of HPG development and OA in bees with more ovarioles started to appear in the low e&#x003b2;-treated bees (Fig.&#x000a0;<xref rid="Fig4" ref-type="fig">4d</xref>) and was significant and consistent in bees treated with high e&#x003b2; (Figs.&#x000a0;<xref rid="Fig4" ref-type="fig">4d</xref> and <xref rid="Fig6" ref-type="fig">6e</xref>). The correlation disappears in the presence of QMP (Fig.&#x000a0;<xref rid="Fig6" ref-type="fig">6e</xref> and Table <xref rid="Tab1" ref-type="table">1</xref>), suggesting that QMP and e&#x003b2; interact to suppress OA in bees with more ovarioles.</p><sec id="FPar1"><title>Replicate effects</title><p id="Par30">Within a single replicate, bees were randomly selected from two of the five donor colonies to minimize within cage variance. We did not prescreen colonies for ovariole number or colony wide OA, both of which vary genetically and influence behavior (reviewed in Page <xref ref-type="bibr" rid="CR55">2013</xref>). Replicate thus encompasses both individual cage and genetic differences. Replicate frequently proved a significant factor in the experiments, suggesting that genotype may influence individual response thresholds to pheromones, as has been demonstrated in other experiments (Amdam et al. <xref ref-type="bibr" rid="CR6">2009</xref>; Pankiw and Page <xref ref-type="bibr" rid="CR63">1999</xref>, <xref ref-type="bibr" rid="CR64">2001</xref>). While there were frequently significant differences between replicates, the replicates typically followed the same trend and only interacted with the treatment group when indicated (Figs.&#x000a0;<xref rid="Fig2" ref-type="fig">2b</xref> and <xref rid="Fig5" ref-type="fig">5b</xref>).</p></sec><sec id="FPar2"><title>Mortality</title><p id="Par31">Although daily mortality remained low (&#x0003c;1 bee/day across all experiments), the presence of a live queen significantly reduced mortality compared to synthetic QMP or control groups (Fig.&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref>). This suggests that live queens enhance survival compared to synthetic QMP, perhaps by reducing overall stress, reducing reproductive competition among workers, and adding to group cohesion by their presence.</p></sec><sec id="FPar3"><title>Ovary activation</title><p id="Par32">Egg laying in insects involves two distinct processes, the production of the egg yolk proteins from the egg yolk precursor vitellogenin (Vg) and the incorporation of these proteins into eggs, followed by the physical oviposition of developed eggs. QMP and e&#x003b2; appear to act on different components of the reproductive physiology in honey bee workers, with the former suppressing OA in bees with fewer ovarioles (Fig.&#x000a0;<xref rid="Fig1" ref-type="fig">1d</xref>), while the latter suppresses OA across all bees at the higher dose of ten larval equivalents per bees (Figs.&#x000a0;<xref rid="Fig4" ref-type="fig">4c</xref> and <xref rid="Fig6" ref-type="fig">6c</xref>).</p><p id="Par33">When queens are present in a colony, there are very low incidences of worker egg laying, though some level of OA is always present (Page and Erickson <xref ref-type="bibr" rid="CR57">1988</xref>). In queenless colonies, some workers become the dominant egg layers and act as false queens (Sakagami <xref ref-type="bibr" rid="CR78">1958</xref>) that attract a queen retinue and suppress physical egg laying in other workers by emitting a queen-like mandibular pheromone (Crewe and Velthuis <xref ref-type="bibr" rid="CR12">1980</xref>). When these false queens are removed, the other workers immediately begin laying eggs (Page and Robinson <xref ref-type="bibr" rid="CR58">1994</xref>; Robinson et al. <xref ref-type="bibr" rid="CR72">1990</xref>), illustrating that queen pheromones suppress egg laying but do not suppress OA (Jay and Nelson <xref ref-type="bibr" rid="CR31">1973</xref>) as well as larval pheromones (this experiment). Workers with activated ovaries are often found in queenright colonies that lack brood (Jay <xref ref-type="bibr" rid="CR29">1972</xref>) or when the brood nest is diminished just prior to swarming (Kropacova and Haslbachova <xref ref-type="bibr" rid="CR38">1970</xref>).</p><p id="Par34">Our queen comparison experiment showed that synthetic QMP significantly suppresses OA compared to controls, though live queens are more effective than QMP in suppressing OA in bees with more ovarioles (Fig.&#x000a0;<xref rid="Fig2" ref-type="fig">2d</xref>). Bees had continual access to QMP, frequently clustering over the synthetic strip. Throughout our experiments, bees with more ovarioles were most likely to activate their ovaries (Figs.&#x000a0;<xref rid="Fig2" ref-type="fig">2c</xref>, <xref rid="Fig3" ref-type="fig">3b</xref>, <xref rid="Fig4" ref-type="fig">4b, c</xref>, and <xref rid="Fig6" ref-type="fig">6c, d</xref>), as has been shown previously (Amdam et al. <xref ref-type="bibr" rid="CR5">2006</xref>; Graham et al. <xref ref-type="bibr" rid="CR18">2011</xref>; Linksvayer et al. <xref ref-type="bibr" rid="CR46">2009</xref>; Page and Amdam <xref ref-type="bibr" rid="CR56">2007</xref>; Page et al. <xref ref-type="bibr" rid="CR59">2006</xref>, <xref ref-type="bibr" rid="CR60">2012</xref>; Tsuruda et al. <xref ref-type="bibr" rid="CR87">2008</xref>; Wang et al. <xref ref-type="bibr" rid="CR88">2010</xref>). Ovariole number is a recognized marker of reproductive potential in honey bees (Makert et al. <xref ref-type="bibr" rid="CR51">2006</xref>; Tanaka and Hartfelder <xref ref-type="bibr" rid="CR85">2004</xref>) demonstrating that workers with the most ovarioles and thus greatest reproductive potential are most likely to escape ovary suppression.</p><p id="Par35">The inability of QMP to suppress OA as strongly as a live queen suggests that more factors are involved in reproductive suppression. Only live queens, who emit multiple pheromones (QMP, Dufour&#x02019;s gland, and tergal pheromones), can fully suppress OA in workers, though both live queens and QMP disassociated total ovarioles from OA (Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>) (Hoover et al. <xref ref-type="bibr" rid="CR22">2003</xref>; Katzav-Gozansky <xref ref-type="bibr" rid="CR33">2006</xref>; Slessor et al. <xref ref-type="bibr" rid="CR83">2005</xref>; Willis et al. <xref ref-type="bibr" rid="CR92">1990</xref>). This difference between QMP and live queens has been postulated to be a sign of a queen &#x0201c;control&#x0201d; and a continuing evolutionary arms race over male reproduction (Katzav-Gozansky <xref ref-type="bibr" rid="CR33">2006</xref>). Alternatively, the multicomponent pheromone could represent an honest signal of queen fecundity linked to reproductive state that encourages worker &#x0201c;cooperation&#x0201d; and informs the colony when the queen starts to fail (Kocher and Grozinger <xref ref-type="bibr" rid="CR36">2011</xref>).</p><p id="Par36">QMP suppresses juvenile hormone (JH) biosynthesis (Robinson et al. <xref ref-type="bibr" rid="CR73">1992</xref>). In honey bees, JH and Vg are normally coregulated in a double-repressor relationship (Amdam and Omholt <xref ref-type="bibr" rid="CR3">2003</xref>; Ihle et al. <xref ref-type="bibr" rid="CR27">2010</xref>); high circulating titers of JH suppress production of Vg and conversely high titers of Vg suppress JH. Since QMP suppresses JH production, these low JH titers in turn augment Vg titers, stimulating production of the egg yolk precursor required for OA.</p><p id="Par37">In the absence of QMP, the e&#x003b2; high dose of 10 Leq/bee significantly suppressed OA (Fig.&#x000a0;<xref rid="Fig4" ref-type="fig">4a</xref>) as seen in previous experiments (Maisonnasse et al. <xref ref-type="bibr" rid="CR49">2009</xref>), paralleling the effects of live larvae, which inhibit worker OA (Jay <xref ref-type="bibr" rid="CR29">1972</xref>; Jay and Jay <xref ref-type="bibr" rid="CR30">1976</xref>). A queenless hive can survive by rearing a replacement queen from larvae present in the colony (Hatch et al. <xref ref-type="bibr" rid="CR19">1999</xref>). However, workers made queenless refrain from rearing an emergency queen for 24&#x000a0;h in the presence of eggs and young larvae but start rearing queens immediately when only older larvae (3rd&#x02013;5th larval instar) are available (Pettis et al. <xref ref-type="bibr" rid="CR68">1997</xref>), indicating that the eggs and/or young larvae provide a fecundity signal that fades after 24&#x000a0;h in the absence of a queen. The low dose of 1 Leq/bee of e&#x003b2; had no effect on ovary suppression.</p><p id="Par38">Our e&#x003b2; and QMP synergy experiment (experiment 4) demonstrates that e&#x003b2; is more effective than synthetic QMP at suppressing OA, and there is no apparent interactive effect on OA between the two pheromones (Fig.&#x000a0;<xref rid="Fig6" ref-type="fig">6a</xref>), at least not at 10&#x000a0;days of age. Our results confirm that both young (current results) and old larval brood pheromones are very effective in suppressing OA and worker reproduction (Arnold et al. <xref ref-type="bibr" rid="CR7">1994</xref>; Maisonnasse et al. <xref ref-type="bibr" rid="CR49">2009</xref>; Mohammedi et al. <xref ref-type="bibr" rid="CR53">1998</xref>). Just as live queens and QMP resulted in a disassociation between total ovarioles and OA (Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>, experiment 1), suggesting suppression of OA regardless of the underlying reproductive physiology, a similar disassociation occurred in our e&#x003b2; and QMP synergy experiment in bees exposed to both brood and queen pheromones (Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref>, experiment 4).</p><p id="Par39">Throughout all of our experiments, we saw low levels of OA at 10&#x000a0;days of age, with mean OA never exceeding stage 1, classified as slight swelling at the top of the ovariole. Control bees consistently had 80&#x000a0;% or more bees with stage 1 OA and 8&#x02013;15&#x000a0;% of bees with vitellogenic ovaries. Bees typically transition out of the brood nest and into other in-hive tasks at 10&#x02013;12&#x000a0;days of age (R&#x000f6;sch <xref ref-type="bibr" rid="CR74">1930</xref>; Seeley <xref ref-type="bibr" rid="CR80">1982</xref>, <xref ref-type="bibr" rid="CR81">1995</xref>; Seeley and Kolmes <xref ref-type="bibr" rid="CR82">1991</xref>). As we were interested in the impacts of e&#x003b2; on nurse bee physiology, we limited the duration of our cage trials to 10&#x000a0;days. Thus, the possibility remains that synergy between QMP and e&#x003b2; on suppression of worker reproduction could occur in more prolonged experiments, with e&#x003b2; suppressing OA and QMP stopping egg laying, although no significant differences or trends were evident between e&#x003b2;+/QMP&#x02212; and e&#x003b2;+/QMP+ at 10&#x000a0;days.</p></sec><sec id="FPar4"><title>HPG development</title><p id="Par40">Incorporating RJ into the diet at 10&#x000a0;% was more effective than access to nurse bees in stimulating HPG development, resulting in almost twice as many bees with well-developed HPGs, classified as stage 3 or 4. Adequate nutrition is essential for both HPG development and OA (Haydak <xref ref-type="bibr" rid="CR20">1970</xref>; Hoover et al. <xref ref-type="bibr" rid="CR23">2006</xref>; Hrassnigg and Crailsheim <xref ref-type="bibr" rid="CR24">1998</xref>).</p><p id="Par41">Bees that experienced the high e&#x003b2; environment developed their HPG significantly more when they had large ovaries compared to small ovaries (Fig.&#x000a0;<xref rid="Fig4" ref-type="fig">4d</xref>). This suggests that worker bees may be more strongly influenced to activate their HPG for larval feeding, if they are predisposed to nursing by possessing more ovariole filaments. Additionally, they may be more prone to activate their ovaries if they have no larvae to receive the brood food, thus repurposing the Vg from their HPG (Amdam and Omholt <xref ref-type="bibr" rid="CR3">2003</xref>; Seehuus et al. <xref ref-type="bibr" rid="CR79">2007</xref>) into their ovaries to produce eggs. Early OA in bees with more ovarioles is correlated with higher titers of Vg that subsequently drop. It is hypothesized that ovariole number and the dynamics of Vg expression influence the onset of foraging and foraging behavior (Ihle et al. <xref ref-type="bibr" rid="CR27">2010</xref>; Nelson et al. <xref ref-type="bibr" rid="CR54">2007</xref>; Page <xref ref-type="bibr" rid="CR55">2013</xref>) in <italic>A. mellifera</italic>, except for subspecies <italic>Apis mellifera capensis</italic> (Roth et al. <xref ref-type="bibr" rid="CR75">2014</xref>), where bees with more ovarioles do not forage precociously. However, &#x0201c;the reproductive control system in <italic>A. m. capensis</italic> is unique when compared with other honeybee subspecies,&#x0201d; (Zheng et al. <xref ref-type="bibr" rid="CR95">2010</xref>) and thus should not be used to dismiss the coupling of reproductive and nursing physiology in all other <italic>A. mellifera</italic>. In our experiments, e&#x003b2; appears to have greater effects on bees with more ovarioles, priming them for both larval care and protein-rich pollen foraging, behavior that supports the nutritional development of the young larvae emitting the pheromone.</p></sec></sec><sec id="Sec17" sec-type="conclusion"><title>Conclusion</title><p id="Par42">Our current results reinforce the reproductive ground plan hypothesis that postulates that ancestral reproductive physiology was coopted and used to regulate foraging behavior (Amdam et al. <xref ref-type="bibr" rid="CR4">2004</xref>, <xref ref-type="bibr" rid="CR5">2006</xref>; Page and Amdam <xref ref-type="bibr" rid="CR56">2007</xref>; Page et al. <xref ref-type="bibr" rid="CR59">2006</xref>). Early nutritional differences in larval development lead to variation in worker ovariole number (Leimar et al. <xref ref-type="bibr" rid="CR43">2012</xref>; Wang et al. <xref ref-type="bibr" rid="CR89">2014</xref>) and thus may contribute to differential response thresholds to e&#x003b2; priming. In field trials, e&#x003b2; both releases and primes bees toward pollen collection (Traynor <xref ref-type="bibr" rid="CR86">2014</xref>), a pollen-foraging bias predicted by the reproductive ground plan hypothesis (Page <xref ref-type="bibr" rid="CR55">2013</xref>; Page and Amdam <xref ref-type="bibr" rid="CR56">2007</xref>; Page et al. <xref ref-type="bibr" rid="CR59">2006</xref>). Our results thus suggest that e&#x003b2; impacts worker physiology tied to maternal traits differentially in predisposed bees that possess more ovariole filaments at both life stages of worker development: During early adult life, e&#x003b2; improves nursing physiology by stimulating HPG development. After the transition to foraging, e&#x003b2; biases bees toward pollen collection to provide protein for the developing brood nest.</p><p id="Par43">Young adult bees actively tending the brood nest typically have the most developed HPG in a colony. The queen spends the majority of her time in the brood nest laying eggs in the vicinity of these nurse bees; thus, the nurse bees have the greatest opportunity for interaction with the queen. When the queen is absent, QMP is not present, and when her reproductive potential starts to fail, there is a reduction of brood and thus a diminishing e&#x003b2; signal. At this point, the nurse bees may detect the changes and reroute Vg from their HPG to their own ovaries for activation and an opportunity for reproduction (Bier <xref ref-type="bibr" rid="CR8">1954</xref>, <xref ref-type="bibr" rid="CR9">1958</xref>), as seen throughout our experiments in the control bees raised without e&#x003b2; or QMP.</p><p id="Par44">Our experimental results illustrate that pheromones in social insects provide complex signals that must be interpreted in context-dependent circumstances and are strongly impacted by individual worker physiology. Honey bee chemical communication has dynamic properties and functions as a property of a complex system (Pankiw <xref ref-type="bibr" rid="CR62">2004</xref>). QMP and e&#x003b2; play important roles in honey bee society as both primer and releaser pheromones that change putative response thresholds to different stimuli by altering reproductive physiology and interacting with innate response thresholds of different genotypes. The young larval pheromone e&#x003b2; suppresses OA across all bees and activates HPG predominantly in bees that have become tuned to nursing because of their heightened number of ovarioles. Larval e&#x003b2; primes these more responsive workers to enhance larval provisioning by increasing HPG development to produce more brood food and by activating their ovaries, tuning those workers to bias later foraging toward pollen collection. Additional field trials that examine the role of e&#x003b2; on honey bee physiology in the context of the hive are needed to complement our current results, as well as experiments that probe the interactions between young and old larval pheromones in concert with QMP.</p></sec><sec sec-type="supplementary-material"><title>Electronic supplementary material</title><sec id="Sec18"><p>Below is the link to the electronic supplementary material.<supplementary-material id="Fig7" content-type="local-data"><media id="MO7" xlink:href="265_2014_1811_Fig7_ESM.jpg"><label>Supplemental Fig. S1</label><caption><p>Experiment 2: access to royal jelly vs. nurse bee mean (+S.E.) ovary activation by replicated. N&#x02009;=&#x02009;180 bees, 60 per treatment, 30 per replicate, 10 bees per cage. Different letters indicate significant differences using LSD student t tests. (JPEG 60 kb)</p></caption></media></supplementary-material> <supplementary-material id="Fig8" content-type="local-data"><media id="MO8" xlink:href="265_2014_1811_Fig8_ESM.jpg"><label>Supplemental Fig. S2</label><caption><p>Experiment 3: e&#x003b2; dose cumulative mortality (+S.E.) per cage over 10&#x000a0;days. N&#x02009;=&#x02009;180 bees, 60 per treatment, 30 per replicate, 10 bees per cage. (JPEG 136 kb)</p></caption></media></supplementary-material> <supplementary-material id="Fig9" content-type="local-data"><media id="MO9" xlink:href="265_2014_1811_Fig9_ESM.jpg"><label>Supplemental Fig. S3</label><caption><p>Experiment 4: e&#x003b2; &#x00026; QMP cumulative mortality (+S.E.) per cage over 10&#x000a0;days. N&#x02009;=&#x02009;240 bees, 60 per treatment, 30 per replicate, 10 bees per cage. (JPEG 137 kb)</p></caption></media></supplementary-material> </p></sec></sec>

Cell-surface translational dynamics of nicotinic acetylcholine receptors

<p>Synapse efficacy heavily relies on the number of neurotransmitter receptors available at a given time. In addition to the equilibrium between the biosynthetic production, exocytic delivery and recycling of receptors on the one hand, and the endocytic internalization on the other, lateral diffusion and clustering of receptors at the cell membrane play key roles in determining the amount of active receptors at the synapse. Mobile receptors traffic between reservoir compartments and the synapse by thermally driven Brownian motion, and become immobilized at the peri-synaptic region or the synapse by: (a) clustering mediated by homotropic inter-molecular receptor&#x02013;receptor associations; (b) heterotropic associations with non-receptor scaffolding proteins or the subjacent cytoskeletal meshwork, leading to diffusional &#x0201c;trapping,&#x0201d; and (c) protein-lipid interactions, particularly with the neutral lipid cholesterol. This review assesses the contribution of some of these mechanisms to the supramolecular organization and dynamics of the paradigm neurotransmitter receptor of muscle and neuronal cells -the nicotinic acetylcholine receptor (nAChR). Currently available information stemming from various complementary biophysical techniques commonly used to interrogate the dynamics of cell-surface components is critically discussed. The translational mobility of nAChRs at the cell surface differs between muscle and neuronal receptors in terms of diffusion coefficients and residence intervals at the synapse, which cover an ample range of time regimes. A peculiar feature of brain &#x003b1;7 nAChR is its ability to spend much of its time confined peri-synaptically, vicinal to glutamatergic (excitatory) and GABAergic (inhibitory) synapses. An important function of the &#x003b1;7 nAChR may thus be visiting the territories of other neurotransmitter receptors, differentially regulating the dynamic equilibrium between excitation and inhibition, depending on its residence time in each domain.</p>

<contrib contrib-type="author"><name><surname>Barrantes</surname><given-names>Francisco J.</given-names></name><xref ref-type="author-notes" rid="fn002">^*</xref><uri xlink:type="simple" xlink:href="http://community.frontiersin.org/people/u/14093"/></contrib>

Frontiers in Synaptic Neuroscience

<sec sec-type="intro"><title>INTRODUCTION</title><p>The nAChR abbreviation&#x003b1;BTX, &#x003b1;-bungarotoxin; CDx, methyl-&#x003b2;-cyclodextrin; FCS, fluorescence correlation spectroscopy; FRAP, fluorescence recovery after photobleaching; MSD, mean square displacement; nAChR, nicotinic acetylcholine receptor; SPT, single particle tracking; TIRF, total internal reflection fluorescence.is the prototype of the family of Cys-loop receptors (<xref rid="B101" ref-type="bibr">Nys et al., 2013</xref>). This family belongs in turn to the superfamily of ligand-gated ion channels (LGICs), a collection of three evolutionarily unrelated families which include, in addition to the aforementioned Cys-loop receptors, the ionotropic glutamate receptors and ATP-gated channels. The Cys-loop family of pentameric proteins is composed of neurotransmitter receptors with associated anion-selective channels [the &#x003b3;-amino butyric acid type A (GABA_A), &#x003b3;-amino butyric acid type C (GABA_C), and the glycine receptor] and cation-selective members such as the 5-HT3 (serotonin) receptor and the nAChR (<xref rid="B101" ref-type="bibr">Nys et al., 2013</xref>). nAChRs are composed of five polypeptide subunits organized pseudo-symmetrically around a central pore. Each subunit contains an extracellular domain, four hydrophobic transmembrane segments arranged in the form of three concentric rings around the pore (<xref rid="B15" ref-type="bibr">Barrantes, 2003</xref>) and a short extracellular carboxy-terminal domain (<xref rid="B74" ref-type="bibr">Karlin, 2002</xref>).</p><p>In the peripheral nervous system, at the neuromuscular junction in adult myotubes, the receptor macromolecule is highly concentrated in a relatively small area of the cell, juxtaposed and restricted to the endplate, packed at the very high density of 10,000&#x02013;20,000 particles &#x003bc;m^-2. Receptor density drops abruptly in the rest of the plasma membrane (&#x0003c;100 particles &#x003bc;m^-2 <xref rid="B12" ref-type="bibr">Barrantes, 1979</xref>; <xref rid="B123" ref-type="bibr">Sanes and Lichtman, 2001</xref>). The functional efficacy of the neuromuscular junction, as well as other synapses, heavily depends on its strength. This in turn is directly related to the number of receptors present at the synapse, which depends on the equilibrium between two sets of factors: (i) lateral diffusion into and out of the synaptic region from non-synaptic (&#x0201c;extrasynaptic&#x0201d;) areas, and (ii) the trafficking and turnover of receptors at the cell surface, determined by the rate and extent of biosynthesis and exocytic delivery to the plasmalemma, plus the contribution of receptor recycling back to the surface, on the one hand, and removal of synaptic receptors by internalization (endocytosis) or 2-dimensional diffusion driving them away from the synaptic region, on the other. The latter phenomenon is uncommon in the peripheral synapse. The density of nAChRs at the synapse is also a consequence of the dynamic equilibrium between all these factors (<xref rid="B1" ref-type="bibr">Akaaboune et al., 1999</xref>; <xref rid="B26" ref-type="bibr">Bruneau and Akaaboune, 2006</xref>). Diffusion into the endplate region is also rare except for accidental or man-tailored conditions such as in denervation hypersensitivity, in which migration of extrasynaptic receptors to the motor endplate occurs in a transient fashion. Several pathological conditions of the neuromuscular junction are associated with an insufficient number of receptor molecules, myasthenia gravis probably being the most prominent example.</p><p>In central nervous system (CNS) synapses, the rapid lateral exchange of receptors at the synapse with those in non-synaptic areas is thought to underlie the plastic behavior of excitatory glutamatergic synapses (i.e., those operating through AMPA and NMDA receptors; <xref rid="B34" ref-type="bibr">Choquet and Triller, 2003</xref>, <xref rid="B35" ref-type="bibr">2013</xref>; <xref rid="B65" ref-type="bibr">Holcman and Triller, 2006</xref>; <xref rid="B143" ref-type="bibr">Triller and Choquet, 2008</xref>). Indeed, this dynamic trafficking, and the resulting effective residence time of excitatory synaptic receptors in the active region, directly affects synaptic efficacy and plasticity, that is, long-term potentiation (LTP), long-term depression (LTD) and other biologically important phenomena which lie at the roots of key cognitive functions. GABAergic and glycinergic receptors at inhibitory synapses are dynamically regulated in a similar fashion. As reviewed in this paper, the 2-D translational mobility of nAChRs may impinge on these important processes. &#x003b1;7 nAChRs reside for distinct periods in the neighborhood of glutamatergic and GABAergic synapses, and due to their high Ca²⁺ permeability, are able to differentially regulate the excitatory/inhibitory balance, LTP, and, indirectly, may influence important cognitive functions like learning and memory.</p><p>Several neurological and neuropsychiatric disorders have been claimed to be associated with dysfunction of receptors and ion channels, whose alterations are encompassed under the term &#x0201c;synaptopathies.&#x0201d; Diseases like depression, anxiety disorders, various forms of dementia, epilepsy, Parkinson&#x02019;s disease, autism spectrum disorder, migraine, fragile X syndrome, and schizophrenia are among these disorders, which cover a wide spectrum of pathological synaptic phenotypes, ranging from alterations in the number, size or morphology of dendritic spines, disposition of spines along the dendritic arborizations, etc. The related alterations in these synaptopathies (either hypo- or hyper-function of the synapse) are assumed to depend in turn on the underlying dysfunction of the receptors and channels, the so-called channelopathies (<xref rid="B75" ref-type="bibr">Kass, 2005</xref>), which should now be extended to encompass scaffolding and other non-receptor proteins, e.g., those misfolded and aggregated at the synapse, like in Alzheimer&#x02019;s, Huntington&#x02019;s or Parkinson&#x02019;s diseases (for a recent review see, e.g., <xref rid="B114" ref-type="bibr">Remmers et al., 2014</xref>).</p></sec><sec><title>ASSESSING THE MOTION OF PROTEINS IN MEMBRANES</title><p>The motion of proteins in membranes depends on a multiplicity of factors: the physicochemical properties of the host lipid bilayer, homotropic intermolecular associations of the protein in question (which may or may not be associated with aggregation or clustering), heterotropic association with other proteins (e.g., scaffolding, cytoskeletal, or motor proteins), or lipids, etc. Physicochemical properties of the lipid bilayer (e.g., viscosity) vary from cell to cell and between different membrane compartments in the same cell, but not to the extent that they become a determining factor in protein motion. By far the most important element that influences diffusion in the 2-D plane of the membrane is the degree of association with partner molecules (crowding and clustering), scaffolding proteins or cytoskeletal barriers (corrals), or tethering to the cytoskeleton (<xref rid="B82" ref-type="bibr">Kusumi et al., 1993</xref>, <xref rid="B81" ref-type="bibr">2005</xref>; <xref rid="B140" ref-type="bibr">Suzuki et al., 2005</xref>; <xref rid="B32" ref-type="bibr">Chen et al., 2006</xref>; <xref rid="B70" ref-type="bibr">Jacobson et al., 2007</xref>) or lipid platforms (<xref rid="B146" ref-type="bibr">Varma and Mayor, 1998</xref>, and see review in <xref rid="B112" ref-type="bibr">Rao and Mayor, 2004</xref>, <xref rid="B113" ref-type="bibr">2014</xref>).</p><p>Assessing the motion of proteins and lipid in membranes has essentially relied on three complementary techniques: FRAP, FCS, and SPT. For a comprehensive review of the introduction and evolution of these techniques (see, e.g., <xref rid="B41" ref-type="bibr">Day and Kenworthy, 2009</xref>). Briefly, FRAP consists of bleaching an area of the membrane containing the fluorescently labeled proteins or lipids in questions with a rapid and relatively intense pulse of light, and then following the time-dependent recovery of the fluorescence signal with a much lower illumination power. The replenishment of the fluorescence signal arises from the diffusion into the photobleached area of fluorescence molecules originally located outside this area. The fluorescence recovery curves are typically characterized by two parameters, a diffusion coefficient (<italic>D</italic>) and a mobile fraction (<italic>Mf</italic>). FCS is also an ensemble method enabling one to study the dynamics (diffusion coefficient), concentrations and molecular interactions (molecular aggregation, binding-unbinding, co-diffusion of two molecular entities, etc.) with high temporal and spatial resolution by following the passage of fluorescently labeled molecules through very small volumes of the cell and analyzing the statistics of fluorescence intensity fluctuations as a function of time (see review in <xref rid="B78" ref-type="bibr">Kim and Schwille, 2003</xref>). Recently, the combined application of FCS and superresolution optical microscopy (see section below) has enabled the observation of some of the above phenomena down to the nanometer scale (see, recent review in <xref rid="B48" ref-type="bibr">Eggeling et al., 2013</xref>).</p><p>Single particle tracking can interrogate the motion of membrane proteins in the native membrane milieu of a living cell by following multiple trajectories of a sufficiently large number of single (e.g., fluorescently labeled) molecules and extracting the apparent average diffusion coefficient from the MSD of the molecules. Some shortcomings of these techniques have been pointed out, such as the invasive nature of FRAP, the essentially &#x0201c;local&#x0201d; interrogation of FCS, and the need to observe isolated particles for relatively long periods of time of SPT (<xref rid="B45" ref-type="bibr">Digman and Gratton, 2009</xref>). The limited spatial and/or temporal resolution of these techniques is still subject to criticism, since they provide a &#x0201c;global&#x0201d; or &#x0201c;macroscopic&#x0201d; diffusion coefficient which reflects the overall mobility over areas of several square microns (<xref rid="B121" ref-type="bibr">Sahl et al., 2014</xref>). In spite of these criticisms, SPT (<xref rid="B133" ref-type="bibr">Simson et al., 1995</xref>, <xref rid="B134" ref-type="bibr">1998</xref>; <xref rid="B124" ref-type="bibr">Saxton and Jacobson, 1997</xref>; <xref rid="B44" ref-type="bibr">Dietrich et al., 2002</xref>; <xref rid="B32" ref-type="bibr">Chen et al., 2006</xref>) still remains the most common approach for analyzing molecular diffusion in membranes, followed by the FRAP technique (see, e.g., <xref rid="B73" ref-type="bibr">Kapitza et al., 1985</xref>; <xref rid="B84" ref-type="bibr">Ladha et al., 1994</xref>; <xref rid="B100" ref-type="bibr">Niv et al., 2002</xref>; <xref rid="B109" ref-type="bibr">Pucadyil et al., 2007</xref>). New analytical tools have appeared in recent years to extend the applicability of SPT analysis to more &#x0201c;real life&#x0201d; (e.g., crowding, anomalous diffusion), complicated membrane environments. One such approach is based on Bayesian and Akaike information criteria in information theory for classifying molecular trajectories (<xref rid="B96" ref-type="bibr">Monnier et al., 2012</xref>; <xref rid="B144" ref-type="bibr">T&#x000fc;rkcan et al., 2012</xref>; <xref rid="B145" ref-type="bibr">T&#x000fc;rkcan and Masson, 2013</xref>; <xref rid="B92" ref-type="bibr">Masson et al., 2014</xref>). The Bayesian method has also been combined with superresolution microscopy techniques such as STED to improve the determination of still positions in sub-diffraction images of GPI-anchored membrane proteins (<xref rid="B89" ref-type="bibr">Manzo et al., 2014</xref>). The reader is referred to a recent paper (<xref rid="B33" ref-type="bibr">Chenouard et al., 2014</xref>), resulting from a competition in which 14 available SPT analytical methods were compared on the same, complex data set &#x02013; an interesting experiment with no winners, but useful conclusions on applicability.</p></sec><sec><title>LATERAL MOBILITY OF DEVELOPING AND ADULT MUSCLE-TYPE nAChR: FRAP STUDIES</title><p>The pioneer study of <xref rid="B8" ref-type="bibr">Axelrod et al. (1976)</xref> using the FRAP technique demonstrated that in developing muscle cells the highly clustered nAChRs present in large (20&#x02013;60 &#x003bc;m) patches are practically immobile, with an effective lateral diffusion coefficient (<italic>D</italic>) of &#x0003c;10^-12 cm² s^-1 (&#x0003c;10^-4 &#x003bc;m² s^-1). The translational mobility of diffusely distributed nAChRs in other regions of the same plasma membrane is only slightly faster (<italic>D</italic> &#x0223c;0.5 &#x000d7; 10^-2 &#x003bc;m² s^-1; see <bold>Table <xref ref-type="table" rid="T1">1</xref></bold>).</p><table-wrap id="T1" position="float"><label>Table 1</label><caption><p>Diffusion coefficients of muscle-type nAChR measured by the FRAP technique.</p></caption><table frame="hsides" rules="groups" cellspacing="5" cellpadding="5"><thead><tr><th valign="top" align="left" rowspan="1" colspan="1">Condition</th><td valign="top" align="left" rowspan="1" colspan="1"><italic>Mf</italic></td><td valign="top" align="left" rowspan="1" colspan="1"><italic>D</italic> (&#x003bc;m² s^-1)</td><th valign="top" align="left" rowspan="1" colspan="1">Reference</th></tr></thead><tbody><tr><td valign="top" align="left" rowspan="1" colspan="1"><bold>Muscle cells</bold></td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">Developing myotubes, synaptic</td><td valign="top" align="left" rowspan="1" colspan="1"/><td valign="top" align="left" rowspan="1" colspan="1">&#x0003c;10^-4</td><td valign="top" align="left" rowspan="1" colspan="1"><xref rid="B8" ref-type="bibr">Axelrod et al. (1976)</xref></td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">Extrasynaptic</td><td valign="top" align="left" rowspan="1" colspan="1"/><td valign="top" align="left" rowspan="1" colspan="1">0.5 &#x000d7; 10^-2</td><td valign="top" align="left" rowspan="1" colspan="1"><xref rid="B8" ref-type="bibr">Axelrod et al. (1976)</xref></td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">Adult rat muscle fibers culture</td><td valign="top" align="left" rowspan="1" colspan="1"/><td valign="top" align="left" rowspan="1" colspan="1">0.25 &#x000d7; 10^-2</td><td valign="top" align="left" rowspan="1" colspan="1"><xref rid="B137" ref-type="bibr">Stya and Axelrod (1983)</xref></td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1"><bold>CHO-K1/A5 cells labeled with Alexa^{<bold>488</bold>}-</bold>&#x003b1;<italic><bold>-BTX</bold></italic></td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">Control cells</td><td valign="top" align="left" rowspan="1" colspan="1">0.56 &#x000b1; 0.09</td><td valign="top" align="left" rowspan="1" colspan="1">0.46 &#x000b1; 0.09 &#x000d7; 10^-2</td><td valign="top" align="left" rowspan="1" colspan="1"><xref rid="B10" ref-type="bibr">Baier et al. (2010)</xref></td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">10 mM CDx-treated</td><td valign="top" align="left" rowspan="1" colspan="1">*0.19 &#x000b1; 0.12</td><td valign="top" align="left" rowspan="1" colspan="1">*0.27 &#x000b1; 0.08 &#x000d7; 10^-2</td><td valign="top" align="left" rowspan="1" colspan="1"><xref rid="B10" ref-type="bibr">Baier et al. (2010)</xref></td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">20 mM Latrunculin A</td><td valign="top" align="left" rowspan="1" colspan="1">0.44 &#x000b1; 0.04</td><td valign="top" align="left" rowspan="1" colspan="1">0.67 &#x000b1; 0.18 &#x000d7; 10^-2</td><td valign="top" align="left" rowspan="1" colspan="1"><xref rid="B10" ref-type="bibr">Baier et al. (2010)</xref></td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">10 mM CDx + 20 mM Latrunculin A</td><td valign="top" align="left" rowspan="1" colspan="1">0.28 &#x000b1; 0.10</td><td valign="top" align="left" rowspan="1" colspan="1">0.49 &#x000b1; 0.21 &#x000d7; 10^-2</td><td valign="top" align="left" rowspan="1" colspan="1"><xref rid="B10" ref-type="bibr">Baier et al. (2010)</xref></td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">3.5 mM CDx-cholesterol (6:1)</td><td valign="top" align="left" rowspan="1" colspan="1">0.62 &#x000b1; 0.08</td><td valign="top" align="left" rowspan="1" colspan="1">0.63 &#x000b1; 0.26 &#x000d7; 10^-2</td><td valign="top" align="left" rowspan="1" colspan="1"><xref rid="B10" ref-type="bibr">Baier et al. (2010)</xref></td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">10 mM CDx-cholesterol (6:1)</td><td valign="top" align="left" rowspan="1" colspan="1">0.54 &#x000b1; 0.10</td><td valign="top" align="left" rowspan="1" colspan="1">*1.17 &#x000b1; 0.44 &#x000d7; 10^-2</td><td valign="top" align="left" rowspan="1" colspan="1"><xref rid="B10" ref-type="bibr">Baier et al. (2010)</xref></td></tr></tbody></table><table-wrap-foot><attrib><italic>CDx, methyl-&#x003b2;-cyclodextrin. Asterisk denotes statistically significant differences, <italic>p</italic> &#x0003c; 0.001.</italic></attrib><attrib><italic>*From <xref rid="B10" ref-type="bibr">Baier et al. (2010)</xref>.</italic></attrib></table-wrap-foot></table-wrap><p>The relative immobility of synaptic nAChRs at the neuromuscular junction is probably due to a multiplicity of factors. The muscle endplate and the electromotor synapse of electric fish are compact &#x0201c;islands&#x0201d; with a huge absolute number of receptor macromolecules densely packed at an extraordinarily high density. It is thus not surprising that receptors hardly diffuse in the plane of the membrane&#x02026; In order to dissect the contribution of intrinsic (e.g., receptor&#x02013;receptor interactions, clearly apparent, e.g., in early electron micrographs of the <italic>Torpedo</italic> electroplax postsynaptic membrane <xref rid="B63" ref-type="bibr">Heuser and Salpeter, 1979</xref>) and extrinsic (e.g., corralling by the submembrane cytoskeletal meshwork) protein clustering factors it is useful to resort to simpler model systems. Heterologous constitutive expression of receptors in cells is a compromise system offering the possibility to conduct a variety of studies under physiological conditions. The clonal cell line CHO-K1/A5 (<xref rid="B117" ref-type="bibr">Roccamo et al., 1999</xref>) robustly expresses adult muscle-type nAChR at densities lower than those of the endplate in an adult muscle cell or the motor plate in the electric fish synapses. Recycling of nAChRs is too slow to contribute to the cell-surface pool within the experimentally observed period (<xref rid="B80" ref-type="bibr">Kumari et al., 2008</xref>). Furthermore, since one has the possibility to increase the complexity of the model system one building block at a time, the lack of non-receptor scaffolding proteins like rapsyn or the clustering factor agrin make the CHO-K1/A5 a useful mammalian expression system to explore &#x0201c;intrinsic&#x0201d; factors involved in clustering and 2-D diffusion of the nAChR protein and to interrogate in a systematic manner for possible involvement of additional components.</p><p>Initial attempts to measure the 2-D mobility of the nAChR at the plasma membrane of CHO-K1/A5 cells and its dependence on membrane cholesterol levels were undertaken using the FRAP technique in the confocal mode (as in <xref rid="B153" ref-type="bibr">Zaal et al., 1999</xref>; <xref rid="B98" ref-type="bibr">Nehls et al., 2000</xref>). A defined 2-D region was selected from the confocal section of the cell membrane, thus restricting the analysis to a few thousand fluorescent-tagged nAChRs. The region was photobleached by transiently increasing the laser power of the confocal microscope, and the diffusive exchange of bleached proteins with nearby unbleached molecules was then followed using low-intensity laser excitation. Recovery into the bleached region can be described by two parameters, an apparent lateral diffusion coefficient, <italic>D</italic>, and a <italic>Mf</italic> (<xref rid="B47" ref-type="bibr">Edidin, 1994</xref>; <xref rid="B32" ref-type="bibr">Chen et al., 2006</xref>; <xref rid="B59" ref-type="bibr">Guo et al., 2008</xref>). <italic>D</italic> provides a measure of the kinetics of translational mobility, whereas <italic>Mf</italic> reports on the proportion of fluorescent molecules that are able to diffuse back into the bleached area over the time course of the assay (<xref rid="B77" ref-type="bibr">Kenworthy et al., 2004</xref>). Using the FRAP technique on &#x003b1;BTX-labeled nAChRs in CHO-K1/A5 cells, we estimated <italic>D</italic> to be 0.46 &#x000d7; 10^-2 &#x003bc;m² s^-1 (<xref rid="B10" ref-type="bibr">Baier et al., 2010</xref>; <bold>Table <xref ref-type="table" rid="T1">1</xref></bold>), indicating that the lateral diffusion coefficient of the muscle-type nAChR at the cell surface of these cells is quite similar to that of the mobile nAChR fraction in developing rat myotubes (0.5 &#x000d7; 10^-2 &#x003bc;m² s^-1; <xref rid="B8" ref-type="bibr">Axelrod et al., 1976</xref>) and that of diffusely distributed nAChR in adult rat muscle fibers in cell culture (0.25 x10^-2 &#x003bc;m² s^-1; <xref rid="B137" ref-type="bibr">Stya and Axelrod, 1983</xref>, <xref rid="B138" ref-type="bibr">1984</xref>; <bold>Table <xref ref-type="table" rid="T1">1</xref></bold>).</p></sec><sec><title>TRANSLATIONAL MOBILITY OF MUSCLE-TYPE nAChR MEASURED BY SPT ANALYSIS</title><p>Fluorescent-labeled (AlexaFluor⁴⁸⁸&#x003b1;-BTX) nAChR particles imaged with TIRF are diffraction-limited (<xref rid="B23" ref-type="bibr">Borroni et al., 2007</xref>; <xref rid="B76" ref-type="bibr">Kellner et al., 2007</xref>); yet time-series of up to a few thousand frames are amenable to SPT analysis and useful information can be extracted about their translational dynamics. The density of these puncta is high, yet there is enough contrast and their separation suffices to track the trajectories with a good signal-to-noise ratio. Using the SPT strategy of Danuser and co-workers (<xref rid="B71" ref-type="bibr">Jaqaman et al., 2008</xref>) all particles contained within multiple frames from selected sub-regions of CHO-K1/A5 cells were detected in time-series for total durations of &#x0223c;25&#x02013;40 s. <bold>Figure <xref ref-type="fig" rid="F1">1</xref></bold> shows the trajectories followed by nAChR particles at the surface of CHO-K1/A5 cells labeled with a monovalent ligand (AlexaFluor⁴⁸⁸&#x003b1;-BTX) or a multivalent ligand (anti-nAChR mAb210 monoclonal antibody followed by AlexaFluor⁴⁸⁸-conjugated IgG secondary antibody) at 4&#x000b0;C. The differences between the two sets of experimental conditions are already apparent from visual inspection of the traces. The motional data derived from the analysis (average dwell-time of the particles, length of their trajectories, average velocity, etc.) are listed in <bold>Table <xref ref-type="table" rid="T2">2</xref></bold> (c.f. <xref rid="B6" ref-type="bibr">Almarza et al., 2014</xref>). No particles fell within the region established for immobile particles (&#x0201c;stationary&#x0201d; regime). The microscopic apparent diffusion coefficient <italic>D</italic>_2-4 (<xref rid="B82" ref-type="bibr">Kusumi et al., 1993</xref>) of the receptor labeled with the monovalent ligand &#x003b1;-BTX, shifted from a wide distribution spanning from &#x0223c;6.7 &#x000d7; 10^-4 - 1 &#x003bc;m² s^-1 (&#x0223c;6.7 &#x000d7; 10^-12 - 1 &#x000d7; 10^-8 cm² s^-1) to a much narrower distribution with an upper limit close to 5.0 &#x000d7; 10^-4 &#x003bc;m² s^-1 upon cholesterol depletion (see <bold>Table <xref ref-type="table" rid="T2">2</xref></bold>; c.f. <xref rid="B6" ref-type="bibr">Almarza et al., 2014</xref>). As for antibody-labeled samples, the proportion of slow-moving particles was significantly higher, with a net displacement of particle motion toward the immobile confined regime. <italic>D</italic>_2-4 values as low as &#x0223c;3.3 &#x000d7; 10^-5 &#x003bc;m² s^-1 (lower limit) to &#x0223c;6.7 &#x000d7; 10^-2 &#x003bc;m² s^-1 (upper limit) were observed. Control samples labeled with mAb210 already exhibited a substantial proportion (19.4%) of immobilized particles. This proportion dramatically increased upon cholesterol depletion of the cells, especially during the initial 10 min (83.3%). Interestingly, this short exposure to CDx appears to suffice to alter the mobility properties of monoliganded and mAb-crosslinked nAChRs. The percentage of stationary particles fell to 57.1 and 26.7% after 20 and 40 min treatment with CDx, respectively (<xref rid="B6" ref-type="bibr">Almarza et al., 2014</xref>). The MSD of nAChR particles is listed in <bold>Table <xref ref-type="table" rid="T3">3</xref></bold>. Recently <xref rid="B132" ref-type="bibr">Simonson et al. (2010)</xref> reported a 2-D diffusion coefficient of 0.1 &#x003bc;m^-2 s^-1 for &#x003b1;7-5HT3 chimeric nAChRs heterologously expressed in HEK cells.</p><fig id="F1" position="float"><label>FIGURE 1</label><caption><p><bold>Multiple trajectories of nAChR particles labeled with BTX and mAb210, respectively.</bold> Sequence of 15 successive frames (out of a total of 1024) corresponding to control BTX- (left column) and mAb (right column)-labeled samples superimposed on the raw TIRF initial frames. Particles were initially localized using a fixed-width Gaussian fitting. Detected particles were subsequently analyzed for their trajectories with the software Localizer (<xref rid="B42" ref-type="bibr">Dedecker et al., 2012</xref>) run in an Igor-Pro environment. Typical total number of trajectories was in the order of 800 (4%) and 700 (ca. 10%) out of a total of 15,000 and 8,000 for BTX and mAb-labeled samples, respectively. Scale bar = 3 &#x003bc;m. From <xref rid="B6" ref-type="bibr">Almarza et al. (2014)</xref>.</p></caption><graphic xlink:href="fnsyn-06-00025-g001"/></fig><table-wrap id="T2" position="float"><label>Table 2</label><caption><p>Mobility parameters of nAChR particles in samples labeled with Alexa^{<bold>488</bold>}&#x003b1;-BTX or with a primary anti-nAChR monoclonal antibody (mAb210) followed by staining with Alexa^{<bold>488</bold>}-labeled secondary antibody, with or without treatment with 15 mM methyl-&#x003b2;-cyclodextrin (CDx).</p></caption><table frame="hsides" rules="groups" cellspacing="5" cellpadding="5"><thead><tr><th valign="top" align="left" rowspan="1" colspan="1">Experiment</th><th valign="top" align="left" rowspan="1" colspan="1">Average lifetime (s)</th><th valign="top" align="left" rowspan="1" colspan="1">Average displacement (&#x003bc;m)</th><th valign="top" align="left" rowspan="1" colspan="1">Average velocity (&#x003bc;m/ms)</th><th valign="top" align="left" rowspan="1" colspan="1">Total no. of particles (in all frames)</th><th valign="top" align="left" rowspan="1" colspan="1">Total no. compound tracks analyzed</th></tr></thead><tbody><tr><td valign="top" align="left" rowspan="1" colspan="1">BTX control</td><td valign="top" align="left" rowspan="1" colspan="1">4.06 &#x000b1; 0.78</td><td valign="top" align="left" rowspan="1" colspan="1">4.05 &#x000b1; 0.27</td><td valign="top" align="left" rowspan="1" colspan="1">0.0011 &#x000b1; 0.0002</td><td valign="top" align="left" rowspan="1" colspan="1">4535</td><td valign="top" align="left" rowspan="1" colspan="1">121</td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">BTX CDx (10 min)</td><td valign="top" align="left" rowspan="1" colspan="1">4.11 &#x000b1; 0.61</td><td valign="top" align="left" rowspan="1" colspan="1">4.54 &#x000b1; 0.36</td><td valign="top" align="left" rowspan="1" colspan="1">0.0010 &#x000b1; 0.0002</td><td valign="top" align="left" rowspan="1" colspan="1">3759</td><td valign="top" align="left" rowspan="1" colspan="1">101</td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">BTX CDx (15 min)</td><td valign="top" align="left" rowspan="1" colspan="1">5.30 &#x000b1; 0.80</td><td valign="top" align="left" rowspan="1" colspan="1">4.42 &#x000b1; 0.06</td><td valign="top" align="left" rowspan="1" colspan="1">0.0009 &#x000b1; 0.0001</td><td valign="top" align="left" rowspan="1" colspan="1">4574</td><td valign="top" align="left" rowspan="1" colspan="1">128</td></tr><tr><td valign="top" align="left" colspan="6" rowspan="1"><hr/></td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">mAb control</td><td valign="top" align="left" rowspan="1" colspan="1">10.47 &#x000b1; 0.31^a</td><td valign="top" align="left" rowspan="1" colspan="1">4.36 &#x000b1; 0.02^a</td><td valign="top" align="left" rowspan="1" colspan="1">0.0004 &#x000b1; 0.0001</td><td valign="top" align="left" rowspan="1" colspan="1">7772</td><td valign="top" align="left" rowspan="1" colspan="1">69</td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">mAb CDx (10 min)</td><td valign="top" align="left" rowspan="1" colspan="1">11.06 &#x000b1; 3.11^b</td><td valign="top" align="left" rowspan="1" colspan="1">2.13 &#x000b1; 0.25^b*</td><td valign="top" align="left" rowspan="1" colspan="1">0.0002 &#x000b1; 0.0001</td><td valign="top" align="left" rowspan="1" colspan="1">5987</td><td valign="top" align="left" rowspan="1" colspan="1">53</td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">mAb CDx (20 min)</td><td valign="top" align="left" rowspan="1" colspan="1">13.41 &#x000b1; 1.44^c</td><td valign="top" align="left" rowspan="1" colspan="1">4.76 &#x000b1; 0.72^c</td><td valign="top" align="left" rowspan="1" colspan="1">0.0005 &#x000b1; 0.0001</td><td valign="top" align="left" rowspan="1" colspan="1">3755</td><td valign="top" align="left" rowspan="1" colspan="1">41</td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">mAb CDx (40 min)</td><td valign="top" align="left" rowspan="1" colspan="1">19.96 &#x000b1; 0.68^d*</td><td valign="top" align="left" rowspan="1" colspan="1">5.70 &#x000b1; 1.18^d</td><td valign="top" align="left" rowspan="1" colspan="1">0.0003 &#x000b1; 0.0001</td><td valign="top" align="left" rowspan="1" colspan="1">4409</td><td valign="top" align="left" rowspan="1" colspan="1">29</td></tr></tbody></table><table-wrap-foot><attrib><italic>Asterisks denote statistically significant differences (ANOVA test, <italic>p</italic> &#x0003c; 0.05).</italic></attrib><attrib><italic>Average lifetime: d exhibited statistically significant difference with a, b, and c.</italic></attrib><attrib><italic>Average displacement: b exhibited statistically significant difference with a, c, and d.</italic></attrib><attrib><italic>*From <xref rid="B6" ref-type="bibr">Almarza et al. (2014)</xref>.</italic></attrib></table-wrap-foot></table-wrap><table-wrap id="T3" position="float"><label>Table 3</label><caption><p>Mean square displacement (MSD) of nAChR particles in samples labeled with Alexa^{<bold>488</bold>}&#x003b1;-BTX or with anti-nAChR monoclonal antibody (mAb210) followed by staining with Alexa^{<bold>488</bold>}-labeled secondary antibody, with or without treatment with 15 mM methyl-&#x003b2;-cyclodextrin (CDx).</p></caption><table frame="hsides" rules="groups" cellspacing="5" cellpadding="5"><thead><tr><th valign="top" align="left" rowspan="1" colspan="1">Experimental condition</th><th valign="top" align="left" rowspan="1" colspan="1">Total no. of particles (in all frames)</th><th valign="top" align="left" rowspan="1" colspan="1">No. of frames analyzed to determine trajectory</th><th valign="top" align="left" rowspan="1" colspan="1">Total no. compound tracks analyzed</th><th valign="top" align="left" rowspan="1" colspan="1">Mean square displacement (&#x003bc;m²)</th></tr></thead><tbody><tr><td valign="top" align="left" rowspan="1" colspan="1">BTX Control</td><td valign="top" align="left" rowspan="1" colspan="1">42,625</td><td valign="top" align="left" rowspan="1" colspan="1">15</td><td valign="top" align="left" rowspan="1" colspan="1">1702</td><td valign="top" align="left" rowspan="1" colspan="1">0.0589 &#x000b1; 0.0016</td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1"/><td valign="top" align="left" rowspan="1" colspan="1"/><td valign="top" align="left" rowspan="1" colspan="1">30</td><td valign="top" align="left" rowspan="1" colspan="1">1530</td><td valign="top" align="left" rowspan="1" colspan="1">0.0936 &#x000b1; 0.0028</td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1"/><td valign="top" align="left" rowspan="1" colspan="1"/><td valign="top" align="left" rowspan="1" colspan="1">60</td><td valign="top" align="left" rowspan="1" colspan="1">1462</td><td valign="top" align="left" rowspan="1" colspan="1">0.1550 &#x000b1; 0.0056</td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">BTX CDx</td><td valign="top" align="left" rowspan="1" colspan="1">18,728</td><td valign="top" align="left" rowspan="1" colspan="1">15</td><td valign="top" align="left" rowspan="1" colspan="1">1572</td><td valign="top" align="left" rowspan="1" colspan="1">0.0874 &#x000b1; 0.0079</td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1"/><td valign="top" align="left" rowspan="1" colspan="1"/><td valign="top" align="left" rowspan="1" colspan="1">30</td><td valign="top" align="left" rowspan="1" colspan="1">1120</td><td valign="top" align="left" rowspan="1" colspan="1">0.1597 &#x000b1; 0.0177</td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1"/><td valign="top" align="left" rowspan="1" colspan="1"/><td valign="top" align="left" rowspan="1" colspan="1">60</td><td valign="top" align="left" rowspan="1" colspan="1">862</td><td valign="top" align="left" rowspan="1" colspan="1">0.2255 &#x000b1; 0.0206</td></tr><tr><td valign="top" align="left" colspan="5" rowspan="1"><hr/></td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">mAb Control</td><td valign="top" align="left" rowspan="1" colspan="1">15,297</td><td valign="top" align="left" rowspan="1" colspan="1">15</td><td valign="top" align="left" rowspan="1" colspan="1">1476</td><td valign="top" align="left" rowspan="1" colspan="1">0.0436 &#x000b1; 0.0032</td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1"/><td valign="top" align="left" rowspan="1" colspan="1"/><td valign="top" align="left" rowspan="1" colspan="1">30</td><td valign="top" align="left" rowspan="1" colspan="1">1126</td><td valign="top" align="left" rowspan="1" colspan="1">0.0671 &#x000b1; 0.0049</td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1"/><td valign="top" align="left" rowspan="1" colspan="1"/><td valign="top" align="left" rowspan="1" colspan="1">60</td><td valign="top" align="left" rowspan="1" colspan="1">861</td><td valign="top" align="left" rowspan="1" colspan="1">0.0974 &#x000b1; 0.0077</td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">mAb CDx</td><td valign="top" align="left" rowspan="1" colspan="1">104,227</td><td valign="top" align="left" rowspan="1" colspan="1">15</td><td valign="top" align="left" rowspan="1" colspan="1">956</td><td valign="top" align="left" rowspan="1" colspan="1">0.0242 &#x000b1; 0.0019</td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1"/><td valign="top" align="left" rowspan="1" colspan="1"/><td valign="top" align="left" rowspan="1" colspan="1">30</td><td valign="top" align="left" rowspan="1" colspan="1">620</td><td valign="top" align="left" rowspan="1" colspan="1">0.0388 &#x000b1; 0.0031</td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1"/><td valign="top" align="left" rowspan="1" colspan="1"/><td valign="top" align="left" rowspan="1" colspan="1">60</td><td valign="top" align="left" rowspan="1" colspan="1">388</td><td valign="top" align="left" rowspan="1" colspan="1">0.0668 &#x000b1; 0.0105</td></tr></tbody></table><table-wrap-foot><attrib><italic>Analyzed with the software Localizer (<xref rid="B42" ref-type="bibr">Dedecker et al., 2012</xref>).</italic></attrib><attrib><italic>*From <xref rid="B6" ref-type="bibr">Almarza et al. (2014)</xref>.</italic></attrib></table-wrap-foot></table-wrap></sec><sec><title>LIPID PLATFORMS AND CHOLESTEROL, THE OBLIGATORY PARTNERS OF THE nAChR</title><p>Cholesterol is an abundant component in the postsynaptic membrane (<xref rid="B14" ref-type="bibr">Barrantes, 1989</xref>) and it has been demonstrated that this lipid is essential for the nAChR, affecting its distribution and several of its functional properties (<xref rid="B17" ref-type="bibr">Barrantes, 2010</xref>, <xref rid="B18" ref-type="bibr">2012</xref>). The lateral heterogeneity of lipids in the postsynaptic membranes of <italic>Torpedo</italic> electrocyte was an early biophysical finding: protein-associated lipids were shown to be immobilized with respect to bulk membrane lipid (<xref rid="B91" ref-type="bibr">Marsh and Barrantes, 1978</xref>), and subsequent work has shown that cholesterol-like molecules form part of this protein-immobilized pool (<xref rid="B16" ref-type="bibr">Barrantes, 2007</xref>). The functional implications of this finding became apparent when it was demonstrated that cholesterol is an essential component for maintaining nAChR agonist-dependent state transitions in the postsynaptic membrane (<xref rid="B36" ref-type="bibr">Criado et al., 1982</xref>). It has been proposed that there are two cholesterol populations in nAChR-rich membranes from <italic>Torpedo:</italic> an easily extractable fraction that influences the bulk fluidity of the membrane and a tightly bound receptor-associated fraction (<xref rid="B85" ref-type="bibr">Leibel et al., 1987</xref>). The lipid raft hypothesis proposes that specific lipid species self-associate to form microdomains or platforms that can intervene in protein partition, signaling and other functional events that occur in cell membranes (<xref rid="B131" ref-type="bibr">Simons and van Meer, 1988</xref>; <xref rid="B130" ref-type="bibr">Simons and Ikonen, 1997</xref>). A fraction of nAChRs occurs in raft domains in mammalian cells, as demonstrated <italic>in vitro</italic> and <italic>in vivo</italic> (<xref rid="B27" ref-type="bibr">Bruses et al., 2001</xref>; <xref rid="B90" ref-type="bibr">Marchand et al., 2002</xref>; <xref rid="B30" ref-type="bibr">Campagna and Fallon, 2006</xref>; <xref rid="B136" ref-type="bibr">Stetzkowski-Marden et al., 2006</xref>; <xref rid="B152" ref-type="bibr">Willmann et al., 2006</xref>; <xref rid="B154" ref-type="bibr">Zhu et al., 2006</xref>), although the purified nAChR protein <italic>per se</italic> exhibits no preference for raft domains <italic>in vitro</italic> (<xref rid="B20" ref-type="bibr">Bermudez et al., 2010</xref>). It has also been shown that cholesterol plays a key role in the trafficking of the nAChR along the early exocytic (<xref rid="B106" ref-type="bibr">Pediconi et al., 2004</xref>) and endocytic (<xref rid="B23" ref-type="bibr">Borroni et al., 2007</xref>; <xref rid="B80" ref-type="bibr">Kumari et al., 2008</xref>; <xref rid="B24" ref-type="bibr">Borroni and Barrantes, 2011</xref>) pathways and also affects nAChR distribution in the plasma membrane (<xref rid="B23" ref-type="bibr">Borroni et al., 2007</xref>; <xref rid="B76" ref-type="bibr">Kellner et al., 2007</xref>; <xref rid="B6" ref-type="bibr">Almarza et al., 2014</xref>).</p><p>Congruent with this series of observations on the multiple roles of cholesterol on nAChR structure and function, cholesterol depletion by CDx treatment produces the accelerated internalization of roughly half of the cell-surface nAChRs in the CHO-K1/A5 cell line (<xref rid="B23" ref-type="bibr">Borroni et al., 2007</xref>), an effect exactly opposite to that observed with most other membrane proteins (see, e.g., <xref rid="B77" ref-type="bibr">Kenworthy et al., 2004</xref>; <xref rid="B41" ref-type="bibr">Day and Kenworthy, 2009</xref>).</p><p>Cholesterol has multiple functional impacts on nAChRs. Thus, lowering cholesterol was found to affect nAChR channel properties, producing gain-of-function, as measured by mean open time distribution in single-channel patch-clamp recordings, whereas cholesterol enrichment had the opposite effect (<xref rid="B23" ref-type="bibr">Borroni et al., 2007</xref>). CDx-mediated depletion of cholesterol produces a reduction in the fraction of mobile nAChRs from 55 to 20% (<xref rid="B10" ref-type="bibr">Baier et al., 2010</xref>). Concomitantly, fluorescence recovery of the toxin-labeled receptor observed in FRAP experiments was clearly slower, yielding an apparent diffusion coefficient (2.1 &#x000b1; 0.7 &#x000d7; 10^-11 cm² s^-1) lower than that in control cells (4.4 &#x000b1; 0.4 &#x000d7; 10^-11 cm² s^-1; <bold>Table <xref ref-type="table" rid="T1">1</xref></bold>). Cholesterol enrichment had the opposite effect. This effect is commonly observed with a wide variety of membrane-embedded proteins (see review by <xref rid="B41" ref-type="bibr">Day and Kenworthy, 2009</xref>).</p><p>A series of recent publications emphasizes the importance of membrane cholesterol in the biogenesis and stability of nAChR clusters <italic>in vivo</italic> and <italic>in vitro</italic>. In muscle cells, cholesterol was found to influence the formation of micron-sized nAChR clusters induced by agrin (<xref rid="B30" ref-type="bibr">Campagna and Fallon, 2006</xref>). Signaling via the agrin/MuSK complex and interaction between the receptor and rapsyn appears to involve lipid platforms (<xref rid="B154" ref-type="bibr">Zhu et al., 2006</xref>). Using Laurdan two-photon fluorescence microscopy (<xref rid="B136" ref-type="bibr">Stetzkowski-Marden et al., 2006</xref>) it was concluded that nAChR clusters reside in ordered membrane domains. Another study (<xref rid="B152" ref-type="bibr">Willmann et al., 2006</xref>) proposed that these cholesterol-rich lipid microdomains and Src-family kinases both contribute to stabilizing nAChRs and the postsynaptic apparatus. As mentioned above, in our experimental clonal cell line, CHO-K1/A5, there are no nAChR-clustering proteins such as rapsyn and tyrosine kinases, and therefore homophilic protein&#x02013;protein interaction, heterophilic protein-lipid interactions, and links with the actin cytoskeleton are more likely candidates for maintaining the nAChR nanocluster assemblies.</p></sec><sec><title>CLUSTERING OF MUSCLE-TYPE nAChR IN CHO-K1/A5 CELLS</title><p>The spatial distribution of nAChRs and other neurotransmitter receptors has been the subject of intense research over the last decades, and it is interesting to gain perspective by looking at what we knew on the subject 35-odd years ago (<xref rid="B12" ref-type="bibr">Barrantes, 1979</xref>). Today we can learn about the supramolecular organization and dynamics of receptors, in living cells, with sub-diffraction resolution, as analyzed in the following section on CNS receptors.</p><p>TIRF movies of BTX- or antibody-labeled muscle-type nAChR particles recorded from live CHO-K1/A5 were recently analyzed using Ripley&#x02019;s K-function (<xref rid="B115" ref-type="bibr">Ripley, 1977</xref>, <xref rid="B116" ref-type="bibr">1979</xref>) and local point-pattern analysis based on the K-function (<xref rid="B105" ref-type="bibr">Owen et al., 2010</xref>; <xref rid="B150" ref-type="bibr">Williamson et al., 2011</xref>; <xref rid="B118" ref-type="bibr">Rossy et al., 2013</xref>), as shown in <bold>Figure <xref ref-type="fig" rid="F2">2</xref></bold>. These methods allow one to examine the spatial organization of the particles by comparing their bi-dimensional point distribution with patterns of complete spatial randomness. Clusters were defined by a maximum nearest neighbor inter-particle radial separation of 200 nm. This dimension is at the limit of the lateral resolution of the TIRF microscopy but is validated -i.e., physically meaningful- by the dimensions of the nAChR nanoclusters resolved by STED microscopy (<xref rid="B76" ref-type="bibr">Kellner et al., 2007</xref>). The maps (<xref rid="B105" ref-type="bibr">Owen et al., 2010</xref>) graphically categorize areas of particle isodensity; discrete &#x0201c;hot spots&#x0201d; showing the highest degree of particle aggregation can be clearly identified in the two-dimensional projections stemming from the entire series of frames. <bold>Figure <xref ref-type="fig" rid="F3">3A</xref></bold> shows the time-dependent evolution of the quantitative cluster maps of BTX-labeled nAChR particles in control and 10 and 15 min after CDx treatment of the cells, respectively. The right column in <bold>Figure <xref ref-type="fig" rid="F3">3A</xref></bold> depicts particles sorted according to their relative brightness. The extremely bright clusters identified by this criterion match the clustered regions sorted by positional recognition in the left column of <bold>Figure <xref ref-type="fig" rid="F3">3A</xref></bold>. The quantitative local point-pattern analysis based on Ripley&#x02019;s K-function (<xref rid="B105" ref-type="bibr">Owen et al., 2010</xref>; <xref rid="B150" ref-type="bibr">Williamson et al., 2011</xref>; <xref rid="B118" ref-type="bibr">Rossy et al., 2013</xref>) clearly indicated that nAChR particles were not randomly distributed but organized in clusters, which differed in size, brightness and density between BTX and antibody-treated samples (<bold>Figure <xref ref-type="fig" rid="F3">3B</xref></bold>). Their arrangement also changed as a function of time of exposure to CDx, reaching a maximum at 10 min treatment for BTX- and 20 min for mAb-labeled samples (<bold>Figure <xref ref-type="fig" rid="F3">3</xref></bold>), in accordance with the kinetics observed in SPT data (<bold>Table <xref ref-type="table" rid="T2">2</xref></bold>).</p><fig id="F2" position="float"><label>FIGURE 2</label><caption><p><bold>From raw TIRF images to the graphical rendering of cluster distribution. (A)</bold> TIRF image of CHO-K1/A5 cells stained with Alexa488-&#x003b1;-BTX. The first frame of a movie comprising 1024 frames is shown. <bold>(B)</bold> The output of the QuickPALM reconstruction procedure (<xref rid="B62" ref-type="bibr">Henriques et al., 2010</xref>) rendered the totality of particles thresholded above a certain brightness level in the entire movie. The area outlined in red corresponds to a 7.5 &#x003bc;m &#x000d7; 7.5 &#x003bc;m region manually selected for further analysis. <bold>(C)</bold> Cluster map resulting from local-point pattern analysis (<xref rid="B118" ref-type="bibr">Rossy et al., 2013</xref>) of the area outlined in red in <bold>(B)</bold>. Visual identification of &#x0201c;hot spots&#x0201d; of clustered particles (black dots) in the entire series of frames. <bold>(D)</bold> Graphical cluster map based on Ripley&#x02019;s K-function (<xref rid="B105" ref-type="bibr">Owen et al., 2010</xref>), pseudo-colored according to relative fluorescence intensity in each individually detected particle. From <xref rid="B6" ref-type="bibr">Almarza et al. (2014)</xref>.</p></caption><graphic xlink:href="fnsyn-06-00025-g002"/></fig><fig id="F3" position="float"><label>FIGURE 3</label><caption><p><bold>Time-dependent evolution of the cluster maps. (A)</bold> Alexa488-&#x003b1;-BTX labeled nAChR particles imaged with TIRF microscopy in CHO-K1/A5 cells. The left column shows the interpolated cluster maps resulting from local-point pattern analysis of 4 &#x003bc;m &#x000d7; 4 &#x003bc;m regions in control and CDx-treated cells at the indicated intervals (10, 15 min). The maps, based on Ripley&#x02019;s K-function (<xref rid="B105" ref-type="bibr">Owen et al., 2010</xref>) provide a graphical representation of the degree of aggregation of particles (black dots) per unit area in the entire series of frames. The right column corresponds to the map of clustered BTX-stained particles, pseudo-colored according to relative brightness of the detected particles. <bold>(B)</bold> Time-dependent evolution of the cluster maps of mAb-crosslinked nAChR particles. The left column corresponds to the interpolated cluster map based on Ripley&#x02019;s K-function applied to CHO-K1/A5 cells labeled with primary anti-nAChR monoclonal antibody (mAb210) followed by staining with Alexa⁴⁸⁸-labeled secondary antibody. The right column shows the map of clustered nAChR particles pseudo-colored according to brightness. Scale bar: 0.2 &#x003bc;m. From <xref rid="B6" ref-type="bibr">Almarza et al. (2014)</xref>.</p></caption><graphic xlink:href="fnsyn-06-00025-g003"/></fig><p>An additional ensemble analytical tool, the pair correlation function <italic>G(r)</italic> (<xref rid="B58" ref-type="bibr">Greenfield et al., 2009</xref>) was applied to the experimental data to measure the properties of individual clusters, averaged over many clusters. Values of <italic>G(r)</italic> &#x0003e; 1 indicate non-random distribution, which can be assumed to be particle clustering (<xref rid="B107" ref-type="bibr">Perry et al., 2006</xref>; <xref rid="B79" ref-type="bibr">Kiskowski et al., 2009</xref>; <xref rid="B125" ref-type="bibr">Sengupta et al., 2011</xref>). Particles exhibited a high degree of clustering at very short&#x02013;length scales in the control sample as compared to particles in cells treated with CDx, which extended its non-random, clustered pattern up to a radius &#x0003e;1 &#x003bc;m. The statistics of nAChRs distribution between disperse and clustered particles are shown in <bold>Table <xref ref-type="table" rid="T4">4</xref></bold>.</p><table-wrap id="T4" position="float"><label>Table 4</label><caption><p>Distribution of free and clustered nAChR particles in CHO/K1-A5 cells (see also <bold>Figure <xref ref-type="fig" rid="F3">3</xref></bold>).</p></caption><table frame="hsides" rules="groups" cellspacing="5" cellpadding="5"><thead><tr><th valign="top" align="left" rowspan="1" colspan="1">Experiment</th><th valign="top" align="left" rowspan="1" colspan="1">Total number of particles</th><th valign="top" align="left" rowspan="1" colspan="1">Particles in clusters</th><td valign="top" align="left" rowspan="1" colspan="1"/></tr></thead><tbody><tr><td valign="top" align="left" rowspan="1" colspan="1">BTX control</td><td valign="top" align="left" rowspan="1" colspan="1">938.1 &#x000b1; 214^a</td><td valign="top" align="left" rowspan="1" colspan="1">895.8 &#x000b1; 209 (95.1%)^a</td><td valign="top" align="left" rowspan="1" colspan="1"/></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">BTX CDx (10 min)</td><td valign="top" align="left" rowspan="1" colspan="1">514.4 &#x000b1; 192^b*</td><td valign="top" align="left" rowspan="1" colspan="1">470.6 &#x000b1; 182 (91.4%)^b*</td><td valign="top" align="left" rowspan="1" colspan="1"/></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">BTX CDx (15 min)</td><td valign="top" align="left" rowspan="1" colspan="1">931.2 &#x000b1; 262^c</td><td valign="top" align="left" rowspan="1" colspan="1">886.6 &#x000b1; 269 (95.2%)^c</td><td valign="top" align="left" rowspan="1" colspan="1"/></tr><tr><td valign="top" align="left" colspan="3" rowspan="1"><hr/></td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">mAb control</td><td valign="top" align="left" rowspan="1" colspan="1">736.7 &#x000b1; 474^d*</td><td valign="top" align="left" rowspan="1" colspan="1">680.7 &#x000b1; 455 (92.3%)^d*</td><td valign="top" align="left" rowspan="1" colspan="1"/></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">mAb CDx (10 min)</td><td valign="top" align="left" rowspan="1" colspan="1">8930.8 &#x000b1; 3200^e*</td><td valign="top" align="left" rowspan="1" colspan="1">8859.9 &#x000b1; 3183 (99.2%)^e*</td><td valign="top" align="left" rowspan="1" colspan="1"/></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">mAb CDx (20 min)</td><td valign="top" align="left" rowspan="1" colspan="1">5521.3 &#x000b1; 2776^f</td><td valign="top" align="left" rowspan="1" colspan="1">5487.3 &#x000b1; 2761 (99.34%)^f</td><td valign="top" align="left" rowspan="1" colspan="1"/></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">mAb CDx (40 min)</td><td valign="top" align="left" rowspan="1" colspan="1">1151.1 &#x000b1; 995^g</td><td valign="top" align="left" rowspan="1" colspan="1">1134.3 &#x000b1; 990 (98.54%)^g</td><td valign="top" align="left" rowspan="1" colspan="1"/></tr></tbody></table><table-wrap-foot><attrib><italic>Total number of particles: b exhibited statistically significant difference with a and c; d exhibited statistically significant difference with e, f, and g.</italic></attrib><attrib><italic>Particles in clusters: b exhibited statistically significant difference with a and c; d exhibited statistically significant difference with e, f, and g.</italic></attrib><attrib><italic>Brightness: a exhibited statistically significant difference with b and c; a exhibited statistically significant difference with b and c.</italic></attrib><attrib><italic>*From <xref rid="B6" ref-type="bibr">Almarza et al. (2014)</xref>.</italic></attrib></table-wrap-foot></table-wrap><p>One important piece of information stemming from this and other studies analyzed in this review is the relationship between the motional regimes that receptor molecules undergo and their supramolecular organization, as well as the effect of non-receptor scaffolding proteins on the latter. It has long since been known that rapsyn (formerly called 43K protein) affects nAChR distribution at the cell surface (<xref rid="B19" ref-type="bibr">Barrantes et al., 1980</xref>; <xref rid="B29" ref-type="bibr">Burden et al., 1983</xref>; <xref rid="B111" ref-type="bibr">Ramarao and Cohen, 1998</xref>). This effect varies along development, as illustrated in a recent study on the effect of rapsyn on nAChR mobility followed along myoblast development in culture (<xref rid="B108" ref-type="bibr">Piguet et al., 2011</xref>). The myristoylated N-terminus of rapsyn molecules anchors nAChRs to the plasma membrane in a 1:1 stoichiometry, playing a major role during myoblast differentiation and neuromuscular junction development. In myoblasts the majority of the receptors were found to be immobile, with 20% of the receptors exhibiting restricted diffusion in small domains of about 50 nm (<xref rid="B108" ref-type="bibr">Piguet et al., 2011</xref>). Before differentiation, only 2% of the nAChRs showed Brownian diffusion, 24% diffused in confined regions, and 74% were immobile. Upon differentiation into multinucleated myoblasts, a strong diminution of the immobile fraction was observed, in conjunction with an increase in the proportion of confined diffusing receptors from 20 to 34%, and Brownian-diffusing receptors from 2 to 10%. In a myoblast cell line devoid of rapsyn, the fraction of mobile nAChRs was higher, and was accompanied by a threefold decrease in the immobile population in comparison to rapsyn-expressing cells. About 50% of the mobile receptors were confined to domains of about 120 nm. Irrespective of the presence of the nAChR-anchoring protein rapsyn, nAChR was confined to domains : when rapsyn was present, the size of the domains diminished (<xref rid="B108" ref-type="bibr">Piguet et al., 2011</xref>). This study is in agreement with our study using direct imaging of nAChR nanoclusters using superresolution microscopy in cells devoid of rapsyn (<xref rid="B76" ref-type="bibr">Kellner et al., 2007</xref>).</p></sec><sec><title>EFFECT OF CHOLESTEROL ON nAChR TRANSLATIONAL MOBILITY</title><p>Several FRAP studies have shown that cholesterol depletion affects the mobility of various proteins at the plasma membrane although the nature, extent and sign of the changes remain a contentious subject. In FRAP experiments performed on cells treated with Mevinolin, a statin that inhibits cholesterol biosynthesis, we found that nAChR mobility was affected in a manner similar to that reported here using CDx mediated acute cholesterol depletion (<xref rid="B10" ref-type="bibr">Baier et al., 2010</xref>; <bold>Table <xref ref-type="table" rid="T1">1</xref></bold>). FCS in the confocal microscopy modality corroborated the results of FRAP microscopy. Whereas values of <italic>D</italic> of 5.3 &#x000b1; 0.4 &#x000d7; 10^-2 &#x003bc;m² s^-1 were observed in control cells, D was reduced to 3.7 &#x000b1; 0.3 &#x000d7; 10^-2 upon cholesterol depletion (<xref rid="B10" ref-type="bibr">Baier et al., 2010</xref>). On the basis of these observations, we can conclude that plasma membrane fluidity is not the main factor determining nAChR mobility.</p><p>Some authors reported that the mobility of raft- and non-raft resident proteins decreases when cholesterol is removed from the plasma membrane (<xref rid="B77" ref-type="bibr">Kenworthy et al., 2004</xref>; <xref rid="B102" ref-type="bibr">O&#x02019;Connell and Tamkun, 2005</xref>). Restricted diffusion of membrane proteins upon cholesterol depletion is believed to result from the formation of solid-like clusters in the membrane (<xref rid="B147" ref-type="bibr">Vrljic et al., 2005</xref>; <xref rid="B99" ref-type="bibr">Nishimura et al., 2006</xref>). <xref rid="B139" ref-type="bibr">Sun et al. (2007)</xref> postulate that cholesterol affects the mechanical properties of plasma membrane through the underlying cytoskeleton. Using single-molecule tracking methods, another group (<xref rid="B104" ref-type="bibr">Orr et al., 2005</xref>) found that cholesterol depletion produces confinement of the epidermal growth factor receptor and human epidermal growth factor receptor 2 mobility, whereas cholesterol enrichment extended the boundaries of the mobility-restricted areas. In contrast, other authors observed an increase in the lateral mobility of the raft-resident proteins CD44 and wild-type GFP-H-Ras after cholesterol depletion (<xref rid="B103" ref-type="bibr">Oliferenko et al., 1999</xref>; <xref rid="B100" ref-type="bibr">Niv et al., 2002</xref>). Removal of cholesterol, particularly with CDx, not only alters membrane viscosity but can also hinder membrane protein diffusion (<xref rid="B127" ref-type="bibr">Shvartsman et al., 2006</xref>).</p><p>There is evidence of interactions between lipids, lipid domains and the cytoskeleton (<xref rid="B93" ref-type="bibr">Maxfield, 2002</xref>; <xref rid="B86" ref-type="bibr">Lenne et al., 2006</xref>; <xref rid="B67" ref-type="bibr">Honigmann et al., 2014</xref>). According to <xref rid="B83" ref-type="bibr">Kwik et al. (2003)</xref> cholesterol depletion produces general effects on the architecture and function of the membrane, making the sub-membrane cytoskeleton and in particular the cortical actin network more stable. Such a reorganization of the actin meshwork would be associated with reduced receptor mobility. Using FCS and STED it was recently shown that membrane-bound actin networks influence lipid phase separation; a model combining the coupling of membrane composition, membrane curvature, and the actin pinning sites was postulated from this study (<xref rid="B67" ref-type="bibr">Honigmann et al., 2014</xref>). More recently, confocal FRAP distinguished two protein populations of membrane proteins, including some classical &#x0201c;synaptic&#x0201d; proteins in PC12 cells, having diffusion coefficients <italic>D</italic> of 0.22 and 0.01 &#x003bc;m² s^-1, respectively (<xref rid="B122" ref-type="bibr">Saka et al., 2014</xref>). When FCS in the superresolution mode (STED-FCS) was applied, the spatio-temporal resolution afforded the determination of <italic>D</italic> on fast diffusing molecules (slowly diffusing or immobile molecules do not traverse the observation spot and do not cause intensity fluctuations, thus precluding their detection). <italic>D</italic> was found to be 0.1&#x02013;0.6 &#x003bc;m² s^-1 for the highly mobile protein fraction, which varied inversely proportional to molecular density. Interestingly, cholesterol level was found to be the most important factor in determining protein mobility <italic>and</italic> stabilizing protein assemblies (clusters; <xref rid="B122" ref-type="bibr">Saka et al., 2014</xref>).</p></sec><sec><title>DIFFUSIONAL MODULATION AND CONFINEMENT OF nAChR ASSEMBLIES BY CYTOSKELETAL COMPONENTS AND SCAFFOLDING PROTEINS</title><p>Cytoskeletal interactions have been shown to modulate the diffusion and confinement of several membrane proteins (<xref rid="B142" ref-type="bibr">Triller and Choquet, 2003</xref>; <xref rid="B81" ref-type="bibr">Kusumi et al., 2005</xref>; <xref rid="B140" ref-type="bibr">Suzuki et al., 2005</xref>). In the case of the muscle-type nAChR, cholesterol depletion affected the long-range relationship of nAChR nano-clusters of &#x0223c;55 nm diameter, changing from a random to a non-random distribution (within a radius of 0.5&#x02013;1.5 &#x003bc;m) upon depletion (<xref rid="B76" ref-type="bibr">Kellner et al., 2007</xref>). Interactions of these nano-clusters with the cytoskeleton were invoked as a possible explanation for these changes since nAChR mobility at the plasma membrane appears to be sensitive to the integrity of the cytoskeleton (<xref rid="B137" ref-type="bibr">Stya and Axelrod, 1983</xref>; <xref rid="B22" ref-type="bibr">Bloch et al., 1989</xref>; <xref rid="B110" ref-type="bibr">Pumplin and Strong, 1989</xref>; <xref rid="B39" ref-type="bibr">Dai et al., 2000</xref>). Furthermore, interaction between nAChR molecules and the cytoskeleton is of physiological and developmental importance: it is a requisite step in the formation and stability of the neuromuscular junction (<xref rid="B64" ref-type="bibr">Hoch, 1999</xref>). In subsequent work from our laboratory the effects of cytoskeleton disruption on nAChR dynamics (<xref rid="B10" ref-type="bibr">Baier et al., 2010</xref>) were experimentally explored. Even though cholesterol depletion-induced loss of nAChR mobility was partially restored in cells incubated with Latrunculin A (<bold>Table <xref ref-type="table" rid="T1">1</xref></bold>; <xref rid="B10" ref-type="bibr">Baier et al., 2010</xref>), the percentage of mobile nAChRs in these cells did not reach control levels. From this we concluded that although the cortical actin meshwork is likely involved in receptor mobility at the cell surface in cholesterol-depleted cells, it is not necessarily the only factor influencing nAChR translational diffusion. Other cortical cytoskeletal proteins and/or actin-binding proteins may be involved, and direct interactions of cholesterol with the nAChR may also be implicated. Furthermore, inhibition of actin polymerization by cytochalasin <italic>D</italic>, which binds to the barbed end of the actin filament and blocks monomer addition, resulted in inhibition of nAChR internalization (<xref rid="B80" ref-type="bibr">Kumari et al., 2008</xref>). However, direct effects of cholesterol on the nAChR cannot be discarded when considering the profound influence of this lipid on the macromolecule&#x02019;s cell surface mobility.</p><p>The nAChR <italic>Mf</italic> may correspond to the nAChR oligomeric forms observed in negatively stained electron micrographs (<xref rid="B13" ref-type="bibr">Barrantes, 1982</xref>), which are exchangeable with relatively less mobile nAChR aggregates in larger nano-clusters (<xref rid="B76" ref-type="bibr">Kellner et al., 2007</xref>). We have hypothesized that lowering cholesterol levels would affect mostly the more rapidly diffusing nAChRs due to enhanced nAChR&#x02013;nAChR interactions, which would decrease their residence time at the cell surface (<xref rid="B16" ref-type="bibr">Barrantes, 2007</xref>) and result in their internalization (<xref rid="B23" ref-type="bibr">Borroni et al., 2007</xref>). The effect of homophilic interactions in membrane proteins was exemplified in an experimental work and model for syntaxin molecules&#x02019; self-organization at the plasma membrane (<xref rid="B129" ref-type="bibr">Sieber et al., 2007</xref>), in which weak homophilic protein&#x02013;protein interactions were responsible for syntaxin clustering, syntaxin molecules in these clusters dynamically exchanging with freely diffusing molecules.</p><p>In brain, glycine receptors are stabilized by microtubules in extrasynaptic regions, and by gephyrin and actin filaments in synaptic regions (<xref rid="B31" ref-type="bibr">Charrier et al., 2006</xref>); AMPA receptors become stable upon interaction with the protein GRIP1, which binds in turn to microfilaments (<xref rid="B5" ref-type="bibr">Allison et al., 1998</xref>). Disruption of the cytoskeleton or the microtubule networks with Latrunculin A or nocodazole, respectively, affected the mobility of the neuronal &#x003b1;7 nAChR but not its ability to form clusters, as we have observed in muscle-type nAChRs using superresolution microscopy (<xref rid="B149" ref-type="bibr">Wenz et al., 2010</xref>). The exact mechanisms of nAChR immobilization in CNS synapses and in particular the role of the cytoskeleton or other diffusional traps merit further investigation.</p><p>Which other factors may contribute to nAChR mobility, trafficking and clustering? Various post-translational modifications are known to occur in nAChRs: the macromolecule is the target of disulphide bond formation, glycosylation, phosphorylation, palmitoylation, and other modifications which might affect nAChR dynamics. Palmitoylation of assembling &#x003b1;7 subunits in the endoplasmic reticulum has been shown to play a role in the formation of functional &#x003b1;BTX sites (<xref rid="B46" ref-type="bibr">Drisdel et al., 2004</xref>; <xref rid="B2" ref-type="bibr">Alexander et al., 2010</xref>). A linear relationship has been found between average nAChR half-life and the percentage of nAChR with phosphorylated &#x003b2; subunit in cultured muscle cells. Phosphorylation occurs specifically at tyrosine residue 390 of the &#x003b2; subunit, and is induced by agrin. This unexpected role of agrin in downregulating AChR turnover most likely stabilizes nAChRs at developing synapses and contributes to the extended half-life of the receptors at adult NMJs (<xref rid="B120" ref-type="bibr">Rudell and Ferns, 2013</xref>). Phosphorylation-induced global conformational changes have been recently proposed to be a universal phenomenon among LGICs, and also to play a role in pathophysiological phenomena such as nicotine addiction in the specific case of the nAChR (<xref rid="B141" ref-type="bibr">Talwar and Lynch, 2014</xref>).</p></sec><sec><title>ANTIBODY-MEDIATED nAChR CROSSLINKING RESTRICTS RECEPTOR MOBILITY</title><p>Muscle nAChR is the target auto-antigen in the autoimmune disease myasthenia gravis. Neuromuscular dysfunction in this disease is caused primarily by the crosslinking of autoantibodies to the endplate nAChR, although other antigens such as muscle-specific tyrosine kinase and low-density lipoprotein receptor-related protein 4 are currently recognized as molecular targets in muscle (<xref rid="B128" ref-type="bibr">Sieb, 2014</xref>). Antibody binding results in impaired receptor function, diminished neuromuscular transmission and clinical symptoms: weakness and rapid-onset fatigue. Antibody binding also triggers the endocytic internalization of nAChRs in C2C12 muscle cells and in CHO-K1/A5 cells (<xref rid="B80" ref-type="bibr">Kumari et al., 2008</xref>). Thus, the effect of antibodies on muscle nAChRs is not only of biological but also of medical importance.</p><p>In agreement with nAChR crosslinking studies in rat myotubes in primary culture (<xref rid="B7" ref-type="bibr">Axelrod, 1980</xref>), antibody-induced crosslinking results in a marked diminution of receptor mobility in adult-type nAChR expressed in CHO-K1/A5 cells. Employing the SPT technique, instead of the long particle walks observed with a monovalent ligand such as &#x003b1;-BTX, the motion of antibody-crosslinked nAChR particles was restricted to much shorter trajectories confined within relatively small areas (<bold>Table <xref ref-type="table" rid="T2">2</xref></bold>; cf. <xref rid="B6" ref-type="bibr">Almarza et al., 2014</xref>).</p></sec><sec><title>NEURONAL-TYPE nAChR MOBILITY</title><p>The dynamics of neuronal nAChRs have also been studied with biophysical techniques. One preparation that has proved suitable for this type of studies is the mouse submandibular ganglion (<xref rid="B94" ref-type="bibr">McCann et al., 2008</xref>). In the synapses between pre- and post-ganglionic neurons in this ganglion, the density of synaptic receptors is normally maintained by the combination of exchange of receptors with non-synaptic regions, a diffusional phenomenon occurring in the time course of minutes, and the turnover of cell surface receptors, taking place in the course of hours. To measure the kinetics of &#x003b1;7 nAChR, <xref rid="B94" ref-type="bibr">McCann et al. (2008)</xref> resorted to various techniques. First, using fluorescent &#x003b1;-BTX they identified postsynaptic and non-synaptic populations of nAChRs. Postsynaptic nAChRs remained stable for days; non-synaptic nAChRs were more dynamic, being replaced in the course of days. Secondly, using the FRAP technique the authors studied nAChR lateral diffusion in the ganglionic neurons, measuring a t_1/2 of recovery of 47 &#x000b1; 7 min and 11 &#x000b1; 4 min for synaptic and non-synaptic &#x003b1;7 nAChR clusters, respectively. Thirdly, to measure the turnover rate of nAChRs <italic>in vivo</italic>, <xref rid="B94" ref-type="bibr">McCann et al. (2008)</xref> resorted to a fluorescence and pulse-chase technique (<xref rid="B1" ref-type="bibr">Akaaboune et al., 1999</xref>) which enabled them to follow the fate of the nAChRs in the living animal for several days. The rate of loss of cell-surface neuronal &#x003b1;7 nAChRs (350 &#x000b1; 47 min) was found to be 60-fold faster than that of muscle-type nAChRs at the neuromuscular junction (<xref rid="B1" ref-type="bibr">Akaaboune et al., 1999</xref>; <xref rid="B26" ref-type="bibr">Bruneau and Akaaboune, 2006</xref>). If living ganglion cell axons were severed, synaptic receptors showed enhanced lateral mobility and insertion of new receptors dramatically decreased, leading to near-complete loss of synaptic receptors and to acute synaptic depression. Disappearance of postsynaptic spines and presynaptic terminals ensued (<xref rid="B94" ref-type="bibr">McCann et al., 2008</xref>). The authors concluded that rapid changes in synaptic efficacy precede long-lasting structural changes in synaptic connectivity. FRAP continues to be applied to the study of neuronal nAChRs. In a recent study, FRAP revealed that the agonist nicotine, acting on &#x003b1;7 nAChRs in hippocampal postsynaptic neurons, induces the stabilization and accumulation of GluA1-type AMPA receptors (<xref rid="B60" ref-type="bibr">Halff et al., 2014</xref>).</p><p>In the CNS, the two most abundant forms of nAChR are the heteropentameric oligomer formed by &#x003b1;4 and &#x003b2;2 subunits and the homopentameric receptor formed exclusively by &#x003b1;7 subunits (<xref rid="B57" ref-type="bibr">Gotti et al., 2009</xref>). The &#x003b1;7 nAChR is found in the neuronal soma and also pre-, post-and peri-synaptically. Presynaptic &#x003b1;7 nAChRs modulate the release of various neurotransmitters, and postsynaptic &#x003b1;7 nAChRs are involved in the generation of postsynaptic currents (<xref rid="B37" ref-type="bibr">Cuevas and Berg, 1998</xref>). Postsynaptic &#x003b1;7 nAChRs can be associated with dendritic spines, in a peri-synaptic annulus (<xref rid="B51" ref-type="bibr">Fabian-Fine et al., 2001</xref>). Peri-synaptic &#x003b1;7 nAChRs are found in the vicinity of GABAergic and glutamatergic synapses (see below and, e.g., <xref rid="B28" ref-type="bibr">Buerli et al., 2010</xref>). The &#x003b1;7 nAChR exhibits unique functional properties that distinguish it from other nicotinic receptors: (a) fast desensitizing kinetics, (b) unusually high Ca²⁺ permeability, and (c) high affinity for binding &#x003b1;BTX (<xref rid="B4" ref-type="bibr">Alkondon et al., 1997</xref>; <xref rid="B3" ref-type="bibr">Alkondon and Albuquerque, 2004</xref>). The &#x003b1;7 nAChR is highly expressed in the hippocampus and in GABAergic interneurons in particular. The hippocampus is one of the brain regions mostly affected in Alzheimer disease, where it regulates inhibition of hippocampal networks: activation of &#x003b1;7 nAChR blocks the induction of short-term potentiation as well as LTP. It is involved in cognition and has been associated with pathological states other than Alzheimer disease, such as schizophrenia and Parkinson disease (<xref rid="B11" ref-type="bibr">Banerjee et al., 2000</xref>).</p><p>In the Introduction I described the mechanisms involved in the maintenance of physiological numbers of receptors at the synapse. In the case of CNS synapses, 2-D diffusion plays an additional role in this equilibrium since receptors need to abandon the postsynaptic region, diffusing away before undergoing endocytosis, a process which appears to occur exclusively in extrasynaptic areas (<xref rid="B21" ref-type="bibr">Blanpied et al., 2002</xref>).</p><p>In another study of &#x003b1;7 nAChR mobility in cultured hippocampal neurons, SPT was carried out on a small fraction of receptors labeled with quantum dot-coupled &#x003b1;-BTX (<xref rid="B28" ref-type="bibr">Buerli et al., 2010</xref>). It should be mentioned that in hippocampal neuronal cultures the GABAergic interneurons are not expected to receive cholinergic innervation, since they are deprived of inputs from distal anatomical brain regions such as the septum. In spite of the absence of synaptic input, &#x003b1;7 nAChRs clusters are present on the neuronal surface. Less than 20% of the receptors were found in clusters, categorized as &#x0201c;synaptic,&#x0201d; as opposed to those labeled with the presynaptic marker synapsin 1, which were assigned to dendritic, postsynaptic, nicotinic sites. The majority (78%) of the receptors were found in the form of aggregates in extrasynaptic areas and were either classified as &#x0201c;axonal&#x0201d; (20%, highly mobile, <italic>D</italic> &#x0003e; 0.1 &#x003bc;m² s^-1, Brownian motion with mostly linear trajectories) or peri-synaptic, i.e., in the vicinity of, but not colocalized with, excitatory glutamatergic (identified by mCherry-Homer 1c staining) and inhibitory GABAergic (labeled with EGFP-gephyrin) postsynaptic densities. The &#x003b1;7 nAChRs in perisynaptic locations differed in their mobility, too, with lowest receptor mobility (&#x0003e;66% of the peri-GABAergic with <italic>D &#x0223c;</italic>0.018 &#x000b1; 0.03 &#x003bc;m² s^-1 and &#x0003e;70% of the peri-glutamatergic with <italic>D &#x0223c;</italic>0.028 &#x000b1; 0.04 &#x003bc;m² s^-1), reflecting local confinement domains, these differences suggesting in turn that the tethering mechanisms holding these nicotinic receptors in the vicinity of excitatory and inhibitory synapses differed as well (<xref rid="B28" ref-type="bibr">Buerli et al., 2010</xref>). What are the possible physiological implications of these findings? Stimulation of &#x003b1;7 nAChRs in hippocampal interneurons modulates GABAergic inhibitory postsynaptic potentials, depressing them in some cases (<xref rid="B148" ref-type="bibr">Wanaverbecq et al., 2007</xref>) or exciting them in other instances (<xref rid="B72" ref-type="bibr">Ji and Dani, 2000</xref>). In the latter case, the ACh-induced excitation of the bicuculline-sensitive GABAergic interneurons could in turn excite or inhibit pyramidal neurons in the CA1 region. Methyllycaconitine-sensitive &#x003b1;7 nAChRs also appear to affect glutamatergic synapses, modulating back-propagating dendritic action potentials and, hence, LTP (<xref rid="B119" ref-type="bibr">Rosza et al., 2008</xref>). Activation of (seven AChRs influences postsynaptic NMDA receptors, relieving the Mg2+ block and thus enhancing the probability of LTP induction (<xref rid="B40" ref-type="bibr">Dani and Bertrand, 2007</xref>). From this type of evidence, the conclusion was reached that their peri-synaptic localization and their high Ca²⁺ permeability endows &#x003b1;7 nAChRs with the ability to regulate both excitatory and inhibitory CNS synapses independently of their endogenous transmitter (<xref rid="B28" ref-type="bibr">Buerli et al., 2010</xref>).</p><p>Chick ciliary ganglion neurons in culture express homomeric &#x003b1;7 and heteromeric &#x003b1;3 nAChR at their surface. nAChR lateral mobility was measured using biotinylated &#x003b1;-BTX and biotinylated monoclonal antibody against &#x003b1;3 nAChRs, respectively, followed by streptavidin-coated quantum dots with an emission wavelength of 605 nm (<xref rid="B53" ref-type="bibr">Fernandes et al., 2010</xref>). In the case of &#x003b1;3 nAChRs, only 34% were mobile. The resulting diffusion coefficient, <italic>D</italic>, was reported to be 0.070 &#x003bc;m² s^-1 and 0.188 &#x003bc;m² s^-1 in synaptic (roughly 50%) and extrasynaptic regions, respectively. In the case of &#x003b1;7 nAChRs the <italic>Mf</italic> was much higher (61%) and the measured <italic>D</italic> was 0.067 and 0.188 &#x003bc;m² s^-1 for synaptic and extrasynaptic locations, respectively (<xref rid="B53" ref-type="bibr">Fernandes et al., 2010</xref>). The dwell time at the synaptic region was about 0.5 ms for the two types of neuronal nAChRs. Analysis of the MSD indicated that synaptic receptors exhibited constrained motion, and extrasynaptic receptors displayed Brownian motion. That is, when either type of receptors is able to diffuse freely, they do so at similar rates, but when their motion is restricted, their constraints differ. In adult ciliary ganglia <italic>in vivo</italic> &#x003b1;7 nAChRs are localized in the peri-synaptic region; in cultured neurons, wide-field microscopy immunocytochemistry showed puncta in close proximity to synaptophysin labeling (<xref rid="B53" ref-type="bibr">Fernandes et al., 2010</xref>).</p></sec><sec><title>CHOLESTEROL AND SCAFFOLDING PROTEINS DIFFERENTIALLY AFFECT NEURONAL &#x003b1;3 AND &#x003b1;7 nAChR MOBILITIES</title><p>Ciliary ganglion neurons were the first test preparation where &#x003b1;7 nAChRs were reported to occur in lipid &#x0201c;rafts&#x0201d; in somatic spines (<xref rid="B27" ref-type="bibr">Bruses et al., 2001</xref>). In their quantum dot SPT study of chick ciliary ganglion neurons, Berg and coworkers (<xref rid="B53" ref-type="bibr">Fernandes et al., 2010</xref>) found that &#x003b1;7 and &#x003b1;3 nAChRs had similar mobilities, but differed in the nature of their synaptic restraints. Furthermore, cholesterol depletion by treatment with cholesterol oxidase increased the mobility of extrasynaptic &#x003b1;3 nAChRs from 0.188 to 0.208 &#x003bc;m² s^-1 without affecting the proportion of immobile &#x003b1;7 nAChRs.</p><p>In contrast, cholesterol depletion affected both synaptic and extrasynaptic &#x003b1;7 nAChRs, and the proportion of receptors visiting synaptic territory increased. Cholesterol depletion also raised the proportion of mobile &#x003b1;3 nAChRs from 34 to 54%, without affecting that of &#x003b1;7 nAChRs. Disruption of PDZ-containing scaffolds or of actin filaments in chick ciliary ganglion neurons increased the mobility of &#x003b1;7 nAChRs but not that of &#x003b1;3, as expected from the wealth of evidence on the role of the actin and PDZ-scaffolds in maintaining synapse, and in particular dendritic spine, architecture (<xref rid="B68" ref-type="bibr">Hotulainen and Hoogenraad, 2010</xref>). It has been previously reported that in one cell, a single species of protein can have one subset undergoing Brownian diffusion and other subsets undergoing confined or anomalous diffusion (<xref rid="B52" ref-type="bibr">Feder et al., 1996</xref>). Muscle-type nAChR mobility also displays a strong dependence on cytoskeletal integrity (<xref rid="B22" ref-type="bibr">Bloch et al., 1989</xref>; <xref rid="B110" ref-type="bibr">Pumplin and Strong, 1989</xref>; <xref rid="B39" ref-type="bibr">Dai et al., 2000</xref>) in developing myotubes and in the adult neuromuscular junction.</p></sec><sec><title>SIMILARITIES AND DIFFERENCES BETWEEN MUSCLE-TYPE AND NEURONAL-TYPE nAChRs: MOTIONAL DYNAMICS AND CLUSTERING ABILITY</title><p>Using FRAP and FCS, two ensemble methods suitable for interrogating membrane protein mobility, we found that the mobility of the adult murine muscle-type nAChR heterologously expressed in the clonal cell line CHO-K1/A5 (<xref rid="B117" ref-type="bibr">Roccamo et al., 1999</xref>) is dependent on cytoskeletal integrity (<xref rid="B10" ref-type="bibr">Baier et al., 2010</xref>). In these cells the nAChR does not form clusters several microns in length as in adult myotubes, but aggregates in the form of very small, nanometer-sized nanoclusters (<xref rid="B76" ref-type="bibr">Kellner et al., 2007</xref>), probably because CHO/K1-A5 cells lack rapsyn and other scaffolding or receptor-anchoring proteins like agrin. An equivalent assembly in the CNS cholinergic synapses has not been experimentally demonstrated to date. Our recent SPT study (<xref rid="B6" ref-type="bibr">Almarza et al., 2014</xref>) reinforces the conclusion of Berg and coworkers (<xref rid="B53" ref-type="bibr">Fernandes et al., 2010</xref>) on the receptor-subtype specificity of the motional regime adopted by different nAChRs. The muscle-type nAChR in CHO-K1/A5 cells is inherently mobile, and only a modest proportion (20%) is immobilized by antibody crosslinking. A dramatic (83%) but transient increase in the percentage of immobile receptors is observed upon cholesterol depletion of the cells, especially during the initial 10 min. The percentage of stationary particles fell thereafter to 57% (20 min) and 27% (40 min) when cells having antibody-crosslinked receptors were additionally depleted of cholesterol. Thus, antibody crosslinking and cholesterol depletion exhibited a synergistic, time-dependent effect (<xref rid="B6" ref-type="bibr">Almarza et al., 2014</xref>).</p><p>The stability of the adult muscle-type nAChR nanoclusters at the cell surface is modulated by the size of their supramolecular organization &#x02013; nAChR nanoclusters increase in size upon cholesterol depletion (<xref rid="B76" ref-type="bibr">Kellner et al., 2007</xref>) &#x02013; and hence by the number of receptor units in the nanocluster. Nanoclusters are subsequently internalized (<xref rid="B23" ref-type="bibr">Borroni et al., 2007</xref>), further reducing the density of nAChRs at the cell surface. Similarly, in ciliary ganglion neurons cholesterol depletion also reduces the number of &#x003b1;3 nAChRs but not that of &#x003b1;7 nAChRs at the cell surface (<xref rid="B53" ref-type="bibr">Fernandes et al., 2010</xref>). Cholesterol, synergistically coupled to other factors determining the size of the nAChR nanoclusters, could thus exert homeostatic control over receptor levels and the dynamics of the nAChR supramolecular assemblies at the cholinergic synapse.</p><p><italic>Caenorhabditis elegans</italic> provides an interesting model system to explore the interplay between neurotransmitter receptors and scaffolding proteins, and to exploit the genetic manipulation of this singular animal to gain insight into the mechanisms involved. At <italic>C. elegans</italic> NMJs, it has been possible to show <italic>in vivo</italic> that extracellular scaffolding proteins are required to cluster the levamisol-sensitive nAChRs (L-nAChRs) in the nematode. The ectodomain of the transmembrane protein LEV-10 and a second extracellular protein, LEV-9, are involved in this process (<xref rid="B55" ref-type="bibr">Gendrel et al., 2009</xref>). LEV-9 is a multidomain factor containing complement control protein modules. LEV-9 is secreted by the muscle cells and needs to be proteolytically cleaved at its C terminus to exert its function (<xref rid="B25" ref-type="bibr">Brise&#x000f1;o-Roa and Bessereau, 2014</xref>).</p></sec><sec><title>MOTION OF OTHER BRAIN NEUROTRANSMITTER RECEPTORS IN CROSSTALK WITH nAChRs</title><p>In brain, most neurotransmitter receptors not anchored to diffusional traps or scaffolding domains appear to freely diffuse on the plane of the membrane at rates between 0.1 and 0.5 &#x003bc;m² s^-1 (<xref rid="B95" ref-type="bibr">Meier et al., 2001</xref>; <xref rid="B34" ref-type="bibr">Choquet and Triller, 2003</xref>; <xref rid="B38" ref-type="bibr">Dahan et al., 2003</xref>; <xref rid="B142" ref-type="bibr">Triller and Choquet, 2003</xref>; <xref rid="B31" ref-type="bibr">Charrier et al., 2006</xref>; <xref rid="B65" ref-type="bibr">Holcman and Triller, 2006</xref>; <xref rid="B49" ref-type="bibr">Ehlers et al., 2007</xref>). &#x003b1;7 nAChR (<xref rid="B28" ref-type="bibr">Buerli et al., 2010</xref>) and glycine receptors (<xref rid="B50" ref-type="bibr">Ehrensperger et al., 2007</xref>) display similar motional behavior: both exhibit high mobility in extrasynaptic areas and confined, low motion in peri-synaptic and synaptic domains. Confinement is inversely correlated to mobility (<xref rid="B95" ref-type="bibr">Meier et al., 2001</xref>; <xref rid="B49" ref-type="bibr">Ehlers et al., 2007</xref>; <xref rid="B28" ref-type="bibr">Buerli et al., 2010</xref>). In a recent study of inhibitory glycinergic receptors and their scaffolding anchorage protein at the postsynaptic density, gephyrin, PALM time-resolved superresolution imaging showed that gephyrin clusters are comprised of several sub-clusters, and that these undergo dynamic changes in the time-course of minutes (<xref rid="B135" ref-type="bibr">Specht et al., 2013</xref>). According to these authors, the morphological changes may correspond to the splitting and merging of gephyrin clusters in the postsynaptic density, whose size determines the number of receptors it can accommodate. Furthermore, the number of the two key inhibitory neurotransmitters &#x02013; glycine and GABA_A- increased with the number of gephyrin clusters at spinal cord synapses. This is another reflection of gephyrin&#x02019;s ubiquity in inhibitory synapses: gephyrin is involved in the clustering of both glycine receptors and a major subset of GABA_A receptors; both compete for the same sites on the gephyrin molecule. Palmitoylation of Cys212 and Cys284 in gephyrin has recently been reported to be critical for the association of this protein with the postsynaptic membrane and also essential for its clustering (trimers, hexamers, and nonamers; <xref rid="B43" ref-type="bibr">Dejanovic et al., 2014</xref>). Lack of palmitoylation leads to mislocalization of gephyrin in non-synaptic regions. Conversely, increased palmitoylation is associated with gain-of-function, i.e., augmented inhibitory GABAergic transmission.</p><p>Interestingly, although the lifetime of the &#x003b1;7 nAChR in glutamatergic and GABAergic synapses was similar, the diffusion coefficient was faster in the inhibitory GABAergic peri-synaptic region, and a larger fraction of &#x003b1;7 nAChRs was found close to glutamatergic synapses. This latter may be related to the observation from the same authors that PICK1, a protein that regulates the trafficking of AMPA receptors (<xref rid="B61" ref-type="bibr">Hanley, 2008</xref>), also interacts with &#x003b1;7 nAChR, inhibiting its clustering (<xref rid="B9" ref-type="bibr">Baer et al., 2007</xref>). The peri-synaptic localization of &#x003b1;7 nAChRs on excitatory glutamatergic synapses may bear particular relevance during early postnatal development, when AMPA receptors are still absent from postsynaptic sites. During this period, and in particular during the first postnatal week (when their density is highest, even higher than in the adult), &#x003b1;7 nAChRs may be the effectors of LTP, either directly or in conjunction with NMDA receptors, and be able to depolarize dendritic spines, thus relieving voltage-dependent Mg²⁺ block mediated by NMDA receptors, and leading to synaptic plasticity (<xref rid="B51" ref-type="bibr">Fabian-Fine et al., 2001</xref>).</p></sec><sec><title>CURRENT APPROACHES AND FUTURE PROSPECTS: ASSESSING NEUROTRANSMITTER RECEPTOR MOBILITY WITH NANOSCOPIC HIGH-TEMPORAL RESOLUTION</title><p>Fluorescence microscopy has recently experienced a series of interesting developments that have opened new avenues to study the fine structure and dynamics of synaptic constituents with unprecedented resolution (see recent reviews in <xref rid="B48" ref-type="bibr">Eggeling et al., 2013</xref>; <xref rid="B151" ref-type="bibr">Willig and Barrantes, 2014</xref>). Localization-based superresolution imaging techniques based on stochastic activation of photo-convertible (switchable) fluorescent probes have in theory inherently low temporal resolution because of the need to collect a considerable number of photons per molecule to accurately localize a given particle. The limitation is particularly apparent when investigating live-cell dynamics. However, the difficulty can be circumvented when studying highly dense collections of particles, and specific analytic techniques have been used to address this subject, as is the case with DAOSTORM &#x02013;originally developed to study crowded stars in astronomy &#x02013; (<xref rid="B66" ref-type="bibr">Holden et al., 2011</xref>) or CSSTORM (compressed sensing STORM; <xref rid="B155" ref-type="bibr">Zhu et al., 2012</xref>). In Neurobiology, these techniques are already proving fruitful for studying the mobility of synaptic molecules. A nanoscopic stochastic technique coined &#x0201c;universal point-accumulation-for-imaging-in-nanoscale-topography&#x0201d; (uPAINT; <xref rid="B56" ref-type="bibr">Giannone et al., 2010</xref>) can be used to render single-molecule diffusion maps at very high particle densities. Anti-R2 subunit of the glutamate receptor followed by ATTO^647N antibody labeling enabled these authors to map a high number of trajectories of AMPA receptors in a single dendritic spine. Another superresolution SPT technique, the combination of single-molecule stochastic nanoscopy (PALM, fPALM) with SPT (&#x0201c;sptPALM&#x0201d;), successfully used to follow the trajectories of membrane proteins (<xref rid="B88" ref-type="bibr">Manley et al., 2008</xref>) and optimized for measuring changes in dendritic spine morphology (<xref rid="B54" ref-type="bibr">Frost et al., 2012</xref>), was applied to Eos2-labeled glutamate subunit-1 of the AMPA receptor in hippocampal dendrites. AMPA-R accumulation was shown to arise from interactions of the receptor with the membrane rather than from clustering (<xref rid="B69" ref-type="bibr">Hoze et al., 2012</xref>). A recent comparison between uPAINT and sptPALM imaging of endogenous and overexpressed AMPA receptors, respectively, showed a remarkable agreement between the two techniques in reporting the number of receptors, confirmed with STED and electron microscopy. AMPA receptor nanodomains were also shown to change in shape in a highly dynamic fashion, often colocalizing with the scaffolding protein PSD95 (<xref rid="B97" ref-type="bibr">Nair et al., 2013</xref>). In the sptPALM study a massive amount of trajectories was analyzed, showing that AMPA receptors in hippocampal synapses are concentrated in nanoclusters of &#x0223c;70 nm containing about 20 receptor molecules each (<xref rid="B97" ref-type="bibr">Nair et al., 2013</xref>). Some clusters partially co-localize with the scaffolding protein PSD-95. As expected, AMPA receptors are retained transiently in the postsynaptic region and exhibit constrained mobility, whereas they are free to diffuse in extrasynaptic areas.</p><p>Another recent work using sptPALM was used to study the adenosine triphosphate (ATP)-gated P2X7 receptors, members of the purinergic receptor family, labeled with Dendra2 (<xref rid="B126" ref-type="bibr">Shrivastava et al., 2013</xref>). P2X7 receptors hardly diffuse in the synaptic region, and two populations of receptors were found in extra-synaptic regions: a rapidly diffusing population and one stabilized within nanoclusters of &#x0223c;100 nm diameter. Another important synaptic membrane protein has recently been studied with sptPALM: Calcium/calmodulin-dependent protein kinase II (CaMKII), an enzyme involved in synaptic plasticity and, indirectly, underlying memory formation. sptPALM was applied to rat hippocampal neurons to track single molecules of CaMKII&#x003b1; (<xref rid="B87" ref-type="bibr">Lu et al., 2014</xref>). CaMKII&#x003b1; exhibits at least three kinetic subpopulations, the major one regulated by actin dynamics, and enzyme mobility in spines was consistently slower than in dendrites, indicating the presence of physical obstacles or binding partners. Interestingly, NMDA-R stimulation triggered CaMKII activation and prompted the immobilization and presumably the binding of CaMKII in dendritic spines (<xref rid="B87" ref-type="bibr">Lu et al., 2014</xref>).</p></sec><sec sec-type="conclusions"><title>CONCLUSION</title><p>Keeping synaptic strength at an adequate level is a functional requisite of both peripheral and central nervous system synapses, and it is the combination of the &#x000b1; mechanisms outlined above that concertedly operate to maintain the functionally adequate density of neurotransmitter receptors. The mechanisms utilized by cells to achieve this equilibrium are complex, and vary between peripheral and CNS. A common feature is the transient immobilization of receptors in nanoscale compartments of the synapse as opposed to extrasynaptic regions, commonly achieved by clustering or by interaction with scaffolding non-receptor proteins. Our ability to interrogate the dynamics of receptors is currently limited to brief glimpses of the molecules&#x02019; entire lifetime, from synthesis to degradation, but nonetheless these snapshots provide useful hints about the organization and the functionally relevant spatiotemporal behavior of these important molecules in the synapse.</p></sec><sec><title>Conflict of Interest Statement</title><p>The author declares that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.</p></sec>

Heart Rate Variability Biofeedback Intervention for Reduction of Psychological Stress During the Early Postpartum Period

Could not extract abstract

<contrib contrib-type="author"><name><surname>Kudo</surname><given-names>Naoko</given-names></name><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author"><name><surname>Shinohara</surname><given-names>Hitomi</given-names></name><xref ref-type="aff" rid="Aff1"/></contrib><contrib contrib-type="author" corresp="yes"><name><surname>Kodama</surname><given-names>Hideya</given-names></name><address><phone>81-(0)18-884-6513</phone><email>kodamah@hs.akita-u.ac.jp</email></address><xref ref-type="aff" rid="Aff1"/></contrib><aff id="Aff1">Department of Maternity Child Nursing, School of Health Science, Akita Graduate School of Medicine and Faculty of Medicine, 1-1-1 Hondo, Akita-shi, 010-8543 Japan </aff>

Applied Psychophysiology and Biofeedback

<sec id="Sec1"><title>Introduction</title><p>Immediately after delivery, mothers are required to adapt to a new lifestyle that focuses on childcare. Many mothers find the process of meeting the demands of their new lifestyle a joyful experience. However, some mothers have trouble getting used to the new routines and responsibilities and experience high stress levels. The early postpartum period is a critical time during which women have an increased risk for depression (Cox et al. <xref ref-type="bibr" rid="CR3">1993</xref>; Ross and Dennis <xref ref-type="bibr" rid="CR27">2009</xref>; Klainin and Arthur <xref ref-type="bibr" rid="CR13">2009</xref>). Therefore, effective interventions that help these women transition through this stressful period should be available.</p><p>The autonomic nervous system plays an important role in human stress reactions. During usual stress reactions, the introduction of a stressor activates the sympathetic nervous system; the system returns to its former state when the stress fades. When subjects are exposed to chronic stress beyond the range where physiological functions are reversible, their everyday autonomic balance shifts toward a sympathetic-predominant state as a result of parasympathetic withdrawal. However, this persistent attenuation of parasympathetic activity may deteriorate the regulatory capability of physiological functions for external stressors (Porges <xref ref-type="bibr" rid="CR23">1995</xref>; McEwen <xref ref-type="bibr" rid="CR21">2004</xref>; Thayer and Sternberg <xref ref-type="bibr" rid="CR34">2006</xref>). In late pregnancy, the balance of the autonomic nervous system of the resting period is shifted toward a sympathetic-predominant state with parasympathetic withdrawal, probably due to adaptive responses against hemodynamic changes and aortocaval compression caused by the enlarged uterus (Kuo et al. <xref ref-type="bibr" rid="CR14">2000</xref>; Walther et al. <xref ref-type="bibr" rid="CR36">2005</xref>; Matsuo et al. <xref ref-type="bibr" rid="CR18">2007</xref>). After delivery, this specific condition rapidly returns to a non-pregnant state, and the recovery process includes parasympathetic activation. If this recovery process does not proceed normally, that is, if sympathetic-predominant autonomic balance is not smoothly recovered, postpartum women became more vulnerable to external stressors and may develop physical and/or psychiatric disorders.</p><p>Heart rate variability (HRV) biofeedback is a training method to control one&#x02019;s breathing to the resonate frequency of about five to six breaths per minute, at which the amplitude of HRV is maximized; this may strengthen the baroreflex, thus improving autonomic functioning (Lehrer et al. <xref ref-type="bibr" rid="CR17">2003</xref>; Vaschillo et al. <xref ref-type="bibr" rid="CR35">2006</xref>). HRV biofeedback has been shown to contribute to the treatment of a variety of diseases with autonomic dysfunctions, including stress-related psychiatric disorders (Karavidas et al. <xref ref-type="bibr" rid="CR12">2007</xref>; Reiner <xref ref-type="bibr" rid="CR25">2008</xref>; Siepmann et al. <xref ref-type="bibr" rid="CR29">2008</xref>; Zucker et al. <xref ref-type="bibr" rid="CR40">2009</xref>; Weber et al. <xref ref-type="bibr" rid="CR37">2010</xref>; Tan et al. <xref ref-type="bibr" rid="CR32">2011</xref>; Beckham et al. <xref ref-type="bibr" rid="CR1">2013</xref>) or stress-related chronic pain (Hassett et al. <xref ref-type="bibr" rid="CR9">2007</xref>; Hallman et al. <xref ref-type="bibr" rid="CR8">2011</xref>). Furthermore, HRV biofeedback may be available as a stress management method for healthy subjects under relatively stressful conditions (Henriques et al. <xref ref-type="bibr" rid="CR10">2011</xref>; Ratanasiripong et al. <xref ref-type="bibr" rid="CR24">2012</xref>; Whited et al. <xref ref-type="bibr" rid="CR39">2014</xref>). Theoretically, HRV biofeedback is beneficial in most mothers whose autonomic balance tends to shift toward a sympathetic-predominant state. Some portable devices for HRV biofeedback are marketed worldwide (Ebben et al. <xref ref-type="bibr" rid="CR6">2009</xref>), and HRV biofeedback is a feasible intervention during the early postpartum period. However, it remains questionable whether HRV biofeedback results in favorable modifications in autonomic functioning of healthy subjects (Lehrer and Eddie <xref ref-type="bibr" rid="CR16">2013</xref>), and effectiveness of HRV biofeedback in healthy postpartum women should be carefully verified before recommending it to mothers as a health-promoting measure after childbirth.</p><p>The objective of the present study was to examine the effectiveness of HRV biofeedback intervention for reduction of psychological stress in women in the early postpartum period. We investigated whether implementation of HRV biofeedback for 4&#x000a0;weeks immediately after delivery could contribute to reduction of the Edinburgh Postnatal Depression Scale (EPDS), a standardized self-reported questionnaire to identify women who have postpartum depression (Cox et al. <xref ref-type="bibr" rid="CR2">1987</xref>). The EPDS has been shown to be able to detect perinatal anxiety disorders as well (Matthey <xref ref-type="bibr" rid="CR19">2008</xref>; Matthey et al. <xref ref-type="bibr" rid="CR20">2013</xref>). Additionally, resting HRV measures in each woman were evaluated as indicators of a fundamental autonomic neural state, and impacts of HRV biofeedback on the measures were assessed. Our hypothesis was that implementation of HRV biofeedback immediately after delivery would result in lower scores on the EPDS and increased HRV measures at 1&#x000a0;month postpartum and that there would be close correlations between EPDS and HRV measures.</p></sec><sec id="Sec2"><title>Methods</title><sec id="Sec3"><title>Study Subjects</title><p>The study protocol was approved by the Ethics Committee of Akita University Graduate School of Medicine and the Faculty of Medicine. Subjects were recruited from mothers who gave birth at Akita University Hospital between October 1, 2011 and September 30, 2013; recruitment took place 4&#x000a0;days postpartum. Only healthy mothers who had experienced vaginal deliveries of a single infant, without any medical complications, were included. Mothers who habitually drank alcohol or smoked were excluded. Written informed consent was obtained from mothers who agreed to participate in the study.</p><p>On postpartum day 4, subjects completed a questionnaire detailing demographic data, including age, gestational age, parity, height, and employment status. As a part of a routine health checkup, body weight, blood pressure, heart rate, and body temperature were measured. Around 4&#x000a0;days after birth, mothers often experience a transient mental disorder called maternity blues. The Stein scale for maternity blues (Stein <xref ref-type="bibr" rid="CR30">1980</xref>) was used to determine whether subjects suffered from this condition.</p></sec><sec id="Sec4"><title>Heart Rate Variability Biofeedback</title><p>All subjects received a brief explanation about HRV biofeedback on postpartum day 4. If subjects agreed to use HRV biofeedback at home, detailed directions regarding how to implement HRV biofeedback using a portable device (StressEraser, Helicor, Inc., New York, NY, USA) were provided. This device records blood vessel pulse waves in the index finger in real time and displays HRV as a waveform on the screen. When users synchronize the rhythm of their breathing with this waveform, they create a resonance between breathing-induced HRV and HRV due to Mayer waves from arterial pressure. When a resonance is completely established, their HRV becomes maximized, and parasympathetic tone is enhanced. The degree of consistency between the HRV waveform on the screen and breathing rhythm is shown on the screen in real time above each individual waveform as a point display ranging from 1 to 3, with 3 points representing the best synchronization.</p><p>Subjects who agreed to implement HRV biofeedback learned to use the device while they were in the hospital and took the device home about 6&#x000a0;days after delivery. According to instructions for the device, subjects were recommended to undergo HRV biofeedback daily with a score of 30 points or more per session and with a sufficient number of sessions a day to achieve a total score of 100 points or more. They were also asked to record their performance daily on a provided chart. Subjects took part in a telephone interview around 2&#x000a0;weeks after discharge to check their compliance with HRV biofeedback. After 4&#x000a0;weeks, subjects visited our hospital for a routine 1-month postnatal check-up. Subjects who did not agree to use biofeedback served as the control group.</p></sec><sec id="Sec5"><title>Heart Rate Variability Analysis</title><p>The resting HRV of all subjects was recorded on day 4 and 1&#x000a0;month postpartum by photoplethysmography (Heart Rhythm Scanner, Biocom Technologies, Poulsbo, WA, USA). Data were collected between 10:00 am and noon, after subjects had confirmed that they had not eaten, drank, or smoked during the previous 2&#x000a0;h. Subjects were instructed to rest in the supine position for 5&#x02013;10&#x000a0;min in a quiet room and breathe slowly. Next, the heart rate scanner optical ear clip sensor was attached to the pinna of the ear. Pulse intervals were recorded for 5&#x000a0;min, during which participants were requested to remain in the supine position. Data were immediately uploaded to a personal computer and HRV measures were calculated. The HRV measures of interest included the standard deviation of the normal heartbeat interval (SDNN), the high-frequency (HF) power in the 0.15&#x02013;0.4&#x000a0;Hz waveband, and the low-frequency (LF) power in the 0.04&#x02013;0.15&#x000a0;Hz wave band (Task Force <xref ref-type="bibr" rid="CR33">1996</xref>).</p></sec><sec id="Sec6"><title>Edinburgh Postnatal Depression Scale</title><p>On day 4 postpartum and at the 1-month postpartum check-up at our hospital, mental state was assessed in all subjects using the EPDS. The EPDS is a 10-item self-rating questionnaire developed to detect probable depression in the first 8&#x000a0;weeks after childbirth (Cox et al. <xref ref-type="bibr" rid="CR2">1987</xref>) and appears to detect perinatal anxiety disorders as well (Matthey <xref ref-type="bibr" rid="CR19">2008</xref>; Matthey et al. <xref ref-type="bibr" rid="CR20">2013</xref>). Each item is scored on a scale of 0&#x02013;3, and the total score ranges from 0 to 30. A score &#x02265;10 points indicates a high risk for postpartum depression.</p><p>Each woman completed the EPDS by herself on day 4 postpartum, but the EPDS at the 1-month postpartum was evaluated during a face-to face-interview with a clinical psychologist who had no direct connection to this study; all interviews took place in a private room. Because our hospital has a rule requiring that all mothers be asked to undergo an interview with a clinical psychologist 1&#x000a0;month after giving birth, the subjects of this study were unaware that the EPDS was being used as the study scale when they were interviewed. After interview, we obtained informed consent from each subject to use this score at the 1-month postpartum for outcome measures in this study.</p></sec><sec id="Sec7"><title>Statistics</title><p>Statistical analyses were performed using the Statistical Package for the Biosciences (Nankodo, Tokyo, Japan) or IBM SPSS Statistics (version 21.0 Static Base and Advanced Statistics, IBM Japan, Tokyo, Japan). Because the distributions for HF power and LF power (the frequency domain analysis values for HRV) approached a normal distribution, logarithmic conversion was performed before analysis. Intergroup comparisons and correlations were analyzed by parametric or nonparametric methods depending on whether or not data were normally distributed. Two-way factorial analysis of variance was used to compare the repeated HRV measures between mothers who underwent biofeedback with those who did not. Group differences of HRV measures at 1&#x000a0;month postpartum were examined by analysis of covariance, adjusted for maternal age, parity, systolic blood pressure, and body mass index. Data were expressed as mean&#x000a0;&#x000b1;&#x000a0;SD, with <italic>P</italic>&#x000a0;&#x0003c;&#x000a0;0.05 regarded as statistically significant.</p></sec></sec><sec id="Sec8"><title>Results</title><p>Fifty-five mothers were approved to participate in this study. Among them, 25 mothers who agreed to implement HRV biofeedback were grouped as the biofeedback group, and 30 mothers who did not want to use HRV biofeedback were grouped as the control group. Table&#x000a0;<xref rid="Tab1" ref-type="table">1</xref> presents comparisons of demographic factors, physical findings, and HRV measures on postpartum day 4 between groups. There were significant differences between groups in terms of parity, gestational age, and systolic blood pressure. The proportion of primiparous mothers was significantly higher in the biofeedback group. Maternity blues was diagnosed in 16 mothers (29.1&#x000a0;%), and the proportion of affected mothers was comparable between groups. There were no significant differences in HRV measures or EPDS between groups on postpartum day 4.<table-wrap id="Tab1"><label>Table&#x000a0;1</label><caption><p>Comparisons of demographic factors, physical findings, heart rate variability measures, and Edinburgh postnatal depression scale on postnatal day 4 between the biofeedback and the control groups</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left"/><th align="left">Biofeedback group<break/>n&#x000a0;=&#x000a0;25</th><th align="left">Control group<break/>n&#x000a0;=&#x000a0;30</th><th align="left"> <italic>P</italic> ^a </th></tr></thead><tbody><tr><td align="left" colspan="4"> <italic>Demographic factors</italic> </td></tr><tr><td align="left">Age (years)</td><td align="left">30.5&#x000a0;&#x000b1;&#x000a0;5.7</td><td align="left">33.4&#x000a0;&#x000b1;&#x000a0;6.6</td><td align="left">0.086</td></tr><tr><td align="left">Primiparous (number, %)</td><td align="left">22 (88.9)</td><td align="left">19 (65.3)</td><td align="left">0.032</td></tr><tr><td align="left">Gestational age (weeks)</td><td align="left">39.1&#x000a0;&#x000b1;&#x000a0;1.0</td><td align="left">38.5&#x000a0;&#x000b1;&#x000a0;0.8</td><td align="left">0.030</td></tr><tr><td align="left" colspan="4"> <italic>Physical findings</italic> </td></tr><tr><td align="left">Body mass index (kg/m²)</td><td align="left">23.5&#x000a0;&#x000b1;&#x000a0;3.1</td><td align="left">24.1&#x000a0;&#x000b1;&#x000a0;3.0</td><td align="left">0.412</td></tr><tr><td align="left">Systolic blood pressure (mmHg)</td><td align="left">109&#x000a0;&#x000b1;&#x000a0;10</td><td align="left">116&#x000a0;&#x000b1;&#x000a0;12</td><td align="left">0.023</td></tr><tr><td align="left">Maternity blues</td><td align="left">9 (38.9)</td><td align="left">7 (23.1)</td><td align="left">0.303</td></tr><tr><td align="left" colspan="4"> <italic>Heart rate variability measures</italic> </td></tr><tr><td align="left">Heart rate (beats/min)</td><td align="left">74.3&#x000a0;&#x000b1;&#x000a0;7.5</td><td align="left">77.1&#x000a0;&#x000b1;&#x000a0;7.6</td><td align="left">0.186</td></tr><tr><td align="left">SDNN (ms)</td><td align="left">38.7&#x000a0;&#x000b1;&#x000a0;11.5</td><td align="left">40.0&#x000a0;&#x000b1;&#x000a0;16.8</td><td align="left">0.743</td></tr><tr><td align="left">HF power (log, ms²)</td><td align="left">4.46&#x000a0;&#x000b1;&#x000a0;0.94</td><td align="left">4.69&#x000a0;&#x000b1;&#x000a0;0.84</td><td align="left">0.358</td></tr><tr><td align="left">LF power (log, ms²)</td><td align="left">4.59&#x000a0;&#x000b1;&#x000a0;0.98</td><td align="left">4.70&#x000a0;&#x000b1;&#x000a0;1.00</td><td align="left">0.684</td></tr><tr><td align="left" colspan="4"> <italic>Edinburgh postnatal depression scale</italic> </td></tr><tr><td align="left">Total score</td><td align="left">4.60&#x000a0;&#x000b1;&#x000a0;1.99</td><td align="left">4.20&#x000a0;&#x000b1;&#x000a0;2.08</td><td align="left">0.317^b </td></tr></tbody></table><table-wrap-foot><p>Values are mean&#x000a0;&#x000b1;&#x000a0;SD (range) or numbers (%)</p><p> <italic>SDNN</italic> standard deviation of normal-to-normal beat intervals, <italic>LF</italic> low frequency, <italic>HF</italic> high frequency</p><p> ^aGroup differences were examined by student <italic>t</italic> test or chi-square test</p><p> ^bGroup differences were examined by Wilcoxon test</p></table-wrap-foot></table-wrap> </p><p>According to daily charts from mothers, all mothers in the biofeedback group implemented at least one session of HRV biofeedback every day, and 20 mothers (80&#x000a0;%) achieved a total of 100 points in all sessions. Four mothers reported that they could not achieve 100 points on about 2&#x02013;5&#x000a0;days because they fell asleep while implementing HRV biofeedback before a score reached that point. One multiparous mother implemented HRV biofeedback and achieved less than 100 points on most days because the laborious care required for her older child.</p><p>Table&#x000a0;<xref rid="Tab2" ref-type="table">2</xref> presents comparisons of HRV measures or EPDS between groups from 4&#x000a0;days to 1&#x000a0;month postpartum. All measures exhibited significant time-dependent changes from 4&#x000a0;days to 1&#x000a0;month postpartum, including a decrease in heart rate, increases in other HRV measures, and a decrease in EPDS. Significant interactive differences that (group&#x000a0;&#x000d7;&#x000a0;time) change between groups were found for heart rate, SDNN, HF power, and EPDS, indicating that these measures changed in different ways between two groups. Figure&#x000a0;<xref rid="Fig1" ref-type="fig">1</xref> presents changes of mean values of each HRV measure or EPDS from 4&#x000a0;days to 4&#x000a0;weeks postpartum in the two groups. The magnitude of changes in heart rate, SDNN, and HF power appeared to be greater in the biofeedback group, as compared to those in the control group. The time-course decrease in EPDS was observed in the biofeedback group only.<table-wrap id="Tab2"><label>Table&#x000a0;2</label><caption><p>Comparisons of heart rate variability measures or Edinburgh postnatal depression scale from 4&#x000a0;days to 1&#x000a0;month postpartum between the biofeedback and control groups</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="2"/><th align="left" rowspan="2">4&#x000a0;days<break/>Mean&#x000a0;&#x000b1;&#x000a0;SD</th><th align="left" rowspan="2">1&#x000a0;month<break/>Mean&#x000a0;&#x000b1;&#x000a0;SD</th><th align="left" colspan="2">Time^a </th><th align="left" colspan="2">Time&#x000a0;&#x000d7;&#x000a0;Group^a </th><th align="left" colspan="2">Group^a </th></tr><tr><th align="left">F_(1,53) </th><th align="left"> <italic>P</italic> </th><th align="left">F_(1,53) </th><th align="left"> <italic>P</italic> </th><th align="left">F_(1,53) </th><th align="left"> <italic>P</italic> </th></tr></thead><tbody><tr><td align="left" colspan="9"> <italic>Heart rate (beats/min)</italic> </td></tr><tr><td align="left">Biofeedback</td><td char="&#x000b1;" align="char">74.3&#x000a0;&#x000b1;&#x000a0;7.5</td><td char="&#x000b1;" align="char">63.1&#x000a0;&#x000b1;&#x000a0;5.8</td><td char="." align="char" rowspan="2">78.34</td><td char="." align="char" rowspan="2">&#x0003c;0.001</td><td char="." align="char" rowspan="2">8.30</td><td char="." align="char" rowspan="2">0.006</td><td char="." align="char" rowspan="2">11.67</td><td char="." align="char" rowspan="2">0.198</td></tr><tr><td align="left">Control</td><td char="&#x000b1;" align="char">77.1&#x000a0;&#x000b1;&#x000a0;7.6</td><td char="&#x000b1;" align="char">71.4&#x000a0;&#x000b1;&#x000a0;6.5</td></tr><tr><td align="left" colspan="9"> <italic>SDNN (ms)</italic> </td></tr><tr><td align="left">Biofeedback</td><td char="&#x000b1;" align="char">38.7&#x000a0;&#x000b1;&#x000a0;11.5</td><td char="&#x000b1;" align="char">57.2&#x000a0;&#x000b1;&#x000a0;12.3</td><td char="." align="char" rowspan="2">21.65</td><td char="." align="char" rowspan="2">&#x0003c;0.001</td><td char="." align="char" rowspan="2">12.41</td><td char="." align="char" rowspan="2">0.001</td><td char="." align="char" rowspan="2">5.26</td><td char="." align="char" rowspan="2">0.090</td></tr><tr><td align="left">Control</td><td char="&#x000b1;" align="char">40.0&#x000a0;&#x000b1;&#x000a0;16.8</td><td char="&#x000b1;" align="char">42.5&#x000a0;&#x000b1;&#x000a0;12.8</td></tr><tr><td align="left" colspan="9"> <italic>HF power (log, ms</italic> ^{<italic>2</italic>} <italic>)</italic> </td></tr><tr><td align="left">Biofeedback</td><td char="&#x000b1;" align="char">4.46&#x000a0;&#x000b1;&#x000a0;0.94</td><td char="&#x000b1;" align="char">5.45&#x000a0;&#x000b1;&#x000a0;0.84</td><td char="." align="char" rowspan="2">19.17</td><td char="." align="char" rowspan="2">&#x0003c;0.001</td><td char="." align="char" rowspan="2">4.89</td><td char="." align="char" rowspan="2">0.031</td><td char="." align="char" rowspan="2">0.46</td><td char="." align="char" rowspan="2">0.501</td></tr><tr><td align="left">Control</td><td char="&#x000b1;" align="char">4.69&#x000a0;&#x000b1;&#x000a0;0.84</td><td char="&#x000b1;" align="char">5.01&#x000a0;&#x000b1;&#x000a0;0.64</td></tr><tr><td align="left" colspan="9"> <italic>LF power (log, ms</italic> ^{<italic>2</italic>} <italic>)</italic> </td></tr><tr><td align="left">Biofeedback</td><td char="&#x000b1;" align="char">4.59&#x000a0;&#x000b1;&#x000a0;0.98</td><td char="&#x000b1;" align="char">5.32&#x000a0;&#x000b1;&#x000a0;0.80</td><td char="." align="char" rowspan="2">11.00</td><td char="." align="char" rowspan="2">0.002</td><td char="." align="char" rowspan="2">3.46</td><td char="." align="char" rowspan="2">0.068</td><td char="." align="char" rowspan="2">0.51</td><td char="." align="char" rowspan="2">0.477</td></tr><tr><td align="left">Control</td><td char="&#x000b1;" align="char">4.70&#x000a0;&#x000b1;&#x000a0;1.00</td><td char="&#x000b1;" align="char">4.91&#x000a0;&#x000b1;&#x000a0;0.95</td></tr><tr><td align="left" colspan="9"> <italic>Edinburgh Postnatal Depression Scale</italic> </td></tr><tr><td align="left">Biofeedback</td><td char="&#x000b1;" align="char">4.20&#x000a0;&#x000b1;&#x000a0;2.08</td><td char="&#x000b1;" align="char">2.56&#x000a0;&#x000b1;&#x000a0;2.26</td><td char="." align="char" rowspan="2">4.60</td><td char="." align="char" rowspan="2">0.037</td><td char="." align="char" rowspan="2">19.43</td><td char="." align="char" rowspan="2">&#x0003c;0.001</td><td char="." align="char" rowspan="2">7.68</td><td char="." align="char" rowspan="2">0.008</td></tr><tr><td align="left">Control</td><td char="&#x000b1;" align="char">4.60&#x000a0;&#x000b1;&#x000a0;1.99</td><td char="&#x000b1;" align="char">5.17&#x000a0;&#x000b1;&#x000a0;2.45</td></tr></tbody></table><table-wrap-foot><p>Values are mean&#x000a0;&#x000b1;&#x000a0;SD</p><p> <italic>SDNN</italic> standard deviation of normal-to-normal beat intervals, <italic>LF</italic> low frequency, <italic>HF</italic> high frequency</p><p> ^aValues are F(<italic>P</italic>) of two-way (time&#x000a0;&#x000d7;&#x000a0;group) repeated measures ANOVA</p></table-wrap-foot></table-wrap> <fig id="Fig1"><label>Fig.&#x000a0;1</label><caption><p>Mean&#x000a0;&#x000b1;&#x000a0;SD of each HRV measure or Edinburgh Postnatal Depression Scale from 4&#x000a0;days to 1&#x000a0;month postpartum in the biofeedback (<italic>solid lines</italic>) and control (<italic>dash lines</italic>) groups. Significant interactive differences (group&#x000a0;&#x000d7;&#x000a0;time) were found for heart rate, SDNN, and HF power (see Table&#x000a0;<xref rid="Tab2" ref-type="table">2</xref>). <italic>SDNN</italic> standard deviation of normal-to-normal beat intervals, <italic>LF</italic> low frequency, <italic>HF</italic> high frequency, <italic>EPDS</italic> Edinburgh Postnatal Depression Scale</p></caption><graphic xlink:href="10484_2014_9259_Fig1_HTML" id="MO1"/></fig> </p><p>Figure&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref> presents distributions of EPDS in women at one month postpartum. Distributions of all women (n&#x000a0;=&#x000a0;55) are presented in Fig.&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>a, and only two women scored &#x02265;10, indicating a high risk for postnatal depression. Distributions of women in each group are separately presented in Fig.&#x000a0;<xref rid="Fig2" ref-type="fig">2</xref>b. In the biofeedback group, 22 of 25 mothers (88.0&#x000a0;%) presented the EPDS of below 5, and, in the control group, 25 of 30 mothers (83.3&#x000a0;%) presented the EPDS of above 4.<fig id="Fig2"><label>Fig.&#x000a0;2</label><caption><p>Distribution of Edinburgh Postnatal Depression Scale in women at 1&#x000a0;month postpartum. Distributions of all women (n&#x000a0;=&#x000a0;55) are presented in (<bold>a</bold>), and distributions of women in each group (<italic>black bars</italic> for biofeedback, n&#x000a0;=&#x000a0;25 and <italic>gray bars</italic> for control, n&#x000a0;=&#x000a0;30) are presented in (<bold>b</bold>)</p></caption><graphic xlink:href="10484_2014_9259_Fig2_HTML" id="MO2"/></fig> </p><p>Table&#x000a0;<xref rid="Tab3" ref-type="table">3</xref> presents comparisons of HRV measures and EPDS (each item and total score) at 1&#x000a0;month postpartum between groups. In terms of HRV measures after adjusting for maternal age, parity, systolic blood pressure, and body mass index, there were significant decreases in heart rate and increases in SDNN in the biofeedback group compared with the control group. There were significant differences between groups in total EPDS score (<italic>P</italic>&#x000a0;&#x0003c;&#x000a0;0.001, Wilcoxon test). Among the EPDS items, significant differences between groups were found in three items related to anxiety (items 3&#x02013;5), one item related to difficulty sleeping (item 7) and one item related to sad and miserable feelings (item 8). Only one woman in the biofeedback group and one woman in the control group scored &#x02265;10, indicating that they were at high risk for postnatal depression.<table-wrap id="Tab3"><label>Table&#x000a0;3</label><caption><p>Comparisons of heart rate variability measures and Edinburgh Postnatal Depression Scale (each item and total score) at 1&#x000a0;month postpartum between the biofeedback and the control groups</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left"/><th align="left">Biofeedback group<break/>n&#x000a0;=&#x000a0;25</th><th align="left">Control group<break/>n&#x000a0;=&#x000a0;30</th><th align="left"> <italic>P</italic> </th></tr></thead><tbody><tr><td align="left" colspan="4"> <italic>Heart rate variability measures</italic> ^{<italic>a</italic>} </td></tr><tr><td align="left">Heart rate (beats/min)</td><td char="&#x000b1;" align="char">63.3&#x000a0;&#x000b1;&#x000a0;5.4</td><td char="&#x000b1;" align="char">71.2&#x000a0;&#x000b1;&#x000a0;6.1</td><td char="." align="char">&#x0003c;0.001</td></tr><tr><td align="left">SDNN (ms)</td><td char="&#x000b1;" align="char">54.7&#x000a0;&#x000b1;&#x000a0;10.7</td><td char="&#x000b1;" align="char">44.6&#x000a0;&#x000b1;&#x000a0;10.1</td><td char="." align="char">0.002</td></tr><tr><td align="left">HF power (log, ms²)</td><td char="&#x000b1;" align="char">5.38&#x000a0;&#x000b1;&#x000a0;0.78</td><td char="&#x000b1;" align="char">5.07&#x000a0;&#x000b1;&#x000a0;0.48</td><td char="." align="char">0.110</td></tr><tr><td align="left">LF power (log, ms²)</td><td char="&#x000b1;" align="char">5.26&#x000a0;&#x000b1;&#x000a0;0.78</td><td char="&#x000b1;" align="char">4.96&#x000a0;&#x000b1;&#x000a0;0.76</td><td char="." align="char">0.196</td></tr><tr><td align="left" colspan="4"> <italic>Edinburgh Postnatal Depression Scale</italic> ^{<italic>b</italic>} </td></tr><tr><td align="left">1. I have been able to laugh and see the funny side of things</td><td char="&#x000b1;" align="char">0.04&#x000a0;&#x000b1;&#x000a0;0.20</td><td char="&#x000b1;" align="char">0.00&#x000a0;&#x000b1;&#x000a0;0.00</td><td char="." align="char">0.290</td></tr><tr><td align="left">2. I have looked forward to things with enjoyment</td><td char="&#x000b1;" align="char">0.04&#x000a0;&#x000b1;&#x000a0;0.20</td><td char="&#x000b1;" align="char">0.10&#x000a0;&#x000b1;&#x000a0;0.31</td><td char="." align="char">0.409</td></tr><tr><td align="left">3. I have blamed myself unnecessarily when things went wrong</td><td char="&#x000b1;" align="char">0.60&#x000a0;&#x000b1;&#x000a0;0.50</td><td char="&#x000b1;" align="char">1.23&#x000a0;&#x000b1;&#x000a0;0.63</td><td char="." align="char">&#x0003c;0.001</td></tr><tr><td align="left">4. I have been anxious or worried for no good reason</td><td char="&#x000b1;" align="char">0.60&#x000a0;&#x000b1;&#x000a0;0.65</td><td char="&#x000b1;" align="char">1.00&#x000a0;&#x000b1;&#x000a0;0.74</td><td char="." align="char">0.040</td></tr><tr><td align="left">5. I have felt scared or panicky for no good reason</td><td char="&#x000b1;" align="char">0.08&#x000a0;&#x000b1;&#x000a0;0.11</td><td char="&#x000b1;" align="char">0.60&#x000a0;&#x000b1;&#x000a0;0.68</td><td char="." align="char">0.001</td></tr><tr><td align="left">6. Things have been getting on top of me</td><td char="&#x000b1;" align="char">0.96&#x000a0;&#x000b1;&#x000a0;0.54</td><td char="&#x000b1;" align="char">1.27&#x000a0;&#x000b1;&#x000a0;0.58</td><td char="." align="char">0.050</td></tr><tr><td align="left">7. I have been so unhappy that I have had difficulty sleeping</td><td char="&#x000b1;" align="char">0.08&#x000a0;&#x000b1;&#x000a0;0.28</td><td char="&#x000b1;" align="char">0.37&#x000a0;&#x000b1;&#x000a0;0.49</td><td char="." align="char">0.014</td></tr><tr><td align="left">8. I have felt sad or miserable</td><td char="&#x000b1;" align="char">008&#x000a0;&#x000b1;&#x000a0;0.28</td><td char="&#x000b1;" align="char">0.37&#x000a0;&#x000b1;&#x000a0;0.49</td><td char="." align="char">0.014</td></tr><tr><td align="left">9. I have been so unhappy that I have been crying</td><td char="&#x000b1;" align="char">0.08&#x000a0;&#x000b1;&#x000a0;0.28</td><td char="&#x000b1;" align="char">0.13&#x000a0;&#x000b1;&#x000a0;0.35</td><td char="." align="char">0.542</td></tr><tr><td align="left">10. The thought of harming myself has occurred to me</td><td char="&#x000b1;" align="char">0.00&#x000a0;&#x000b1;&#x000a0;0.00</td><td char="&#x000b1;" align="char">0.10&#x000a0;&#x000b1;&#x000a0;0.31</td><td char="." align="char">0.112</td></tr><tr><td align="left">Total score</td><td char="&#x000b1;" align="char">2.56&#x000a0;&#x000b1;&#x000a0;2.26</td><td char="&#x000b1;" align="char">5.17&#x000a0;&#x000b1;&#x000a0;2.45</td><td char="." align="char">&#x0003c;0.001</td></tr></tbody></table><table-wrap-foot><p> <italic>SDNN</italic> standard deviation of normal-to-normal beat intervals, <italic>LF</italic> low frequency, <italic>HF</italic> high frequency</p><p> ^aGroup differences were examined by analysis of covariance, adjusted for maternal age, parity, systolic blood pressure, and body mass index</p><p> ^bGroup differences were examined by Wilcoxon test</p></table-wrap-foot></table-wrap> </p><p>Table&#x000a0;<xref rid="Tab4" ref-type="table">4</xref> presents the correlation coefficients of total EPDS with each HRV measure at 1&#x000a0;month postpartum in all mothers (n&#x000a0;=&#x000a0;55). The EPDS score exhibited a significant positive correlation with heart rate, and significant negative correlations with SDNN and HF power (Spearman&#x02019;s rank correlation test).<table-wrap id="Tab4"><label>Table&#x000a0;4</label><caption><p>Correlation coefficients of total Edinburgh Postnatal Depression Scale with each heart rate variable measure in all mothers at 1&#x000a0;month postpartum</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left"/><th align="left">R^a </th><th align="left"> <italic>P</italic> </th></tr></thead><tbody><tr><td align="left">Heart rate (beats/min)</td><td char="." align="char">0.476</td><td char="." align="char">&#x0003c;0.001</td></tr><tr><td align="left">SDNN (ms)</td><td char="." align="char">&#x02212;0.277</td><td char="." align="char">0.04</td></tr><tr><td align="left">HF power (log, ms²)</td><td char="." align="char">&#x02212;0.357</td><td char="." align="char">0.008</td></tr><tr><td align="left">LF power (log, ms²)</td><td char="." align="char">&#x02212;0.211</td><td char="." align="char">0.122</td></tr></tbody></table><table-wrap-foot><p> <italic>SDNN</italic> standard deviation of normal-to-normal beat intervals, <italic>LF</italic> low frequency, <italic>HF</italic> high frequency</p><p> ^aSpearman&#x02019;s rank correlation coefficient</p></table-wrap-foot></table-wrap> </p></sec><sec id="Sec9"><title>Discussion</title><p>EPDS scores at 1&#x000a0;month postpartum were significantly lower in the biofeedback group than in the control group, suggesting that use of HRV biofeedback after delivery contributed to reduction of psychological stress in postpartum women. Comparisons of each item in the EPDS between groups showed that HRV biofeedback contributed to alleviation of items related to anxiety (Matthey <xref ref-type="bibr" rid="CR19">2008</xref>) and difficulty sleeping. Anxiety is probably among the most common negative emotions for postpartum women. In a community sample of 8,323 pregnant women, approximately 15&#x000a0;% of women reported elevated anxiety in the antenatal period, and rates were comparable in the postnatal period (Heron et al. <xref ref-type="bibr" rid="CR11">2004</xref>). In fact, anxiety disorders are more common than depressive disorders in the perinatal period (Matthey et al. <xref ref-type="bibr" rid="CR20">2013</xref>). Therefore, reduction of anxiety symptoms with HRV biofeedback, which was reported by other studies (Reiner <xref ref-type="bibr" rid="CR25">2008</xref>; Henriques et al. <xref ref-type="bibr" rid="CR10">2011</xref>; Ratanasiripong et al. <xref ref-type="bibr" rid="CR24">2012</xref>), may be particularly beneficial for postpartum women. HRV biofeedback was reported to shorten sleep latency (Ebben et al. <xref ref-type="bibr" rid="CR6">2009</xref>), prolong deeper sleep stages (Sakakibara et al. <xref ref-type="bibr" rid="CR28">2013</xref>), and ameliorate insomnia (McLay and Spira <xref ref-type="bibr" rid="CR22">2009</xref>). Due to childcare responsibilities, postpartum women sleep less during the early weeks following delivery than during pregnancy and other periods of the reproductive age (Lee et al. <xref ref-type="bibr" rid="CR15">2000</xref>). These impaired sleep patterns are strongly correlated with depressive symptoms in postpartum women (D&#x000f8;rheim et al. <xref ref-type="bibr" rid="CR5">2009</xref>). Therefore, it is likely that sleep-promoting effects of HRV biofeedback also contribute to reduction of psychological stress in some postpartum women.</p><p>From 4&#x000a0;days to 1&#x000a0;month postpartum, there were significant reductions in heart rate and elevations in SDNN, HF power, and LF power. HF power is an established index of cardiac vagal tone, reflecting respiratory sinus arrhythmia. Although LF power was previously thought to reflect cardiac sympathetic outflow, several researchers believe that the HRV power spectrum, including the LF component, is mainly determined by the parasympathetic system (Grassi and Esler <xref ref-type="bibr" rid="CR7">1999</xref>; Reyes del Paso et al. <xref ref-type="bibr" rid="CR26">2013</xref>). Therefore, we regarded that increases in both LF and HF power would reflect increased parasympathetic activity after delivery. The magnitude of changes in heart rate, SDNN, and HF power were larger in the biofeedback group, and thus, HRV biofeedback may exaggerate parasympathetic activation during the early postpartum period. However, this difference may also be attributable to the fact that the demographics of subjects were biased. The biofeedback group included a larger proportion of primiparous mothers with relatively younger ages. The activation of parasympathetic tone after delivery may be more evident among younger, primiparous mothers. Therefore, we should attach more importance to results showing significant group differences in heart rate and SDNN at 1&#x000a0;month postpartum after controlling for the influence of covariates, including maternal age and parity.</p><p>There were significant positive correlations between EPDS and heart rate, and negative correlations between EPDS and both SDNN and HF powers. In general, low HRV is thought to indicate decreased parasympathetic activity. Therefore, a significant increase in resting SDNN, the index of overall HRV, may indicate that use of HRV biofeedback resulted in increased parasympathetic tone in the resting state. However, the meaning attached to increases in SDNN without increases in HF and/or LF power in the present study should be carefully considered because SDNN simply increases when heart rate decreases, as was observed in the present study. Previously, several studies analyzed HRV measures as an outcome of HRV biofeedback, and significant increases in SDNN or LF power during HRV biofeedback were constantly reported (Lehrer et al. <xref ref-type="bibr" rid="CR17">2003</xref>; Karavidas et al. <xref ref-type="bibr" rid="CR12">2007</xref>; Hassett et al. <xref ref-type="bibr" rid="CR9">2007</xref>). However, although several studies investigated whether there were carry-over effects of HRV biofeedback on resting HRV, conclusions were inconsistent. Some studies demonstrated positive impacts of HRV biofeedback on resting HRV measures, such as SDNN (Zucker et al. <xref ref-type="bibr" rid="CR40">2009</xref>; Del Pozo et al. <xref ref-type="bibr" rid="CR4">2004</xref>) or LF power (Hallman et al. <xref ref-type="bibr" rid="CR8">2011</xref>), whereas other studies reported that the influences were rare or nonexistent (Lehrer et al. <xref ref-type="bibr" rid="CR17">2003</xref>; Karavidas et al. <xref ref-type="bibr" rid="CR12">2007</xref>; Siepmann et al. <xref ref-type="bibr" rid="CR29">2008</xref>; Swanson et al. <xref ref-type="bibr" rid="CR31">2009</xref>; Henriques et al. <xref ref-type="bibr" rid="CR10">2011</xref>). Therefore, it may be true that positive effects of HRV biofeedback cannot be clearly explained by changes in daily autonomic functioning. Wheat and Larkin (<xref ref-type="bibr" rid="CR38">2010</xref>) stated that, because clinical and physiological outcome do not improve concurrently, the mechanism by which HRV biofeedback results in salutary effects is still unclear.</p><p>Compliance with HRV biofeedback was high among women in this study, with as many as 20 mothers (80&#x000a0;%) achieving a total of 100 points or more in all sessions. Clinical outcomes of the biofeedback group were favorable, although it was probable that these effects resulted, in part, from the intervention effects. In our experience, HRV biofeedback serves as a useful communication tool between medical staff and mothers, as we noted that investigators and some mothers using HRV biofeedback achieved a closer relationship throughout this study. In another study, dizziness occurred in 15&#x000a0;% of 24 patients with anxiety disorders who used HRV biofeedback, a side effect that may have resulted from hyperventilation (Reiner <xref ref-type="bibr" rid="CR25">2008</xref>). No mothers in this study complained about this symptom. Therefore, HRV biofeedback is a feasible, effective, and safe intervention for most postpartum women. Thus, if staff members recommend HRV biofeedback with enthusiasm, a considerable number of mothers may be willing to use this treatment. However, it remains questionable whether HRV biofeedback is really advantageous to healthy users. Lehrer and Eddie (<xref ref-type="bibr" rid="CR16">2013</xref>) stated that HRV biofeedback enhanced the negative feedback loop, including the baroreflex, but this might weaken reflexes dependent on oscillations at other frequencies. This raises the concern that frequent, long-term use of HRV biofeedback may weaken adaptability of the physical control system to external stressors. Therefore, for postpartum women, it may be preferable to implement HRV biofeedback for a relatively short period daily (about 20&#x000a0;min, as recommended by Lehrer and Eddie <xref ref-type="bibr" rid="CR16">2013</xref>), and to limit the period of HRV biofeedback to the first month after delivery, when stress is most likely to occur.</p><p>Several limitations of the present study warrant discussion. First, this was not a random study and demographics were biased, although use of appropriate statistical analyses was able to control for potential covariates to some extent. Second, the resting HRV was recorded by photoplethysmography, not by electrocardiography, and thus, accuracy of HRV measures was less than ideal. Third, subjective influences may have occurred in the biofeedback group when they answered questions from the EPDS interviewer, and the decrease in EPDS in the biofeedback group may be largely due to intervention effects. The absence of an active control made it difficult to validate genuine effects of HRV biofeedback. Forth, evaluation of individual stress levels relying on single EPDS may be incorrect in some women, and introduction of multilateral evaluation (e.g., simultaneous estimation of another stress scale or using biochemical markers) may have provide more accurate information on stress levels in postpartum women. Fifth, our results do not indicate that HRV biofeedback contributes to a reduction in the risk of postpartum depression. Our study subjects included only two women who scored &#x02265;10 on the EPDS, indicating a high risk for postnatal depression. There is no evidence that a difference in scores that fall within the normal range reflects a difference in the actual risk for postpartum depression.</p><p>In conclusion, results in this study partially supported our hypothesis that implementation of HRV biofeedback immediately after delivery resulted in lower EPDS scores and increased HRV measures at 1&#x000a0;month postpartum. The mothers who used HRV biofeedback were relatively free from anxiety and complained less of difficulties sleeping; however, the lack of a random study design and an active control group means that these findings should be interpreted with caution. HRV biofeedback intervention was found to reduce heart rate and increase SDNN in the resting period, but increases in SDNN without increases in HF or LF powers provide inconclusive evidence of parasympathetic activations. However, due to its clinical effectiveness and feasibility, HRV biofeedback appears to be recommendable for many postpartum women after childbirth, especially when they are worried about upcoming changes in routines and the responsibilities of childcare.</p></sec>

Exploring nicotinamide cofactor promiscuity in NAD(P)H-dependent flavin containing monooxygenases (FMOs) using natural variation within the phosphate binding loop. Structure and activity of FMOs from <italic>Cellvibrio</italic> sp. BR and <italic>Pseudomonas stutzeri</italic> NF13

Could not extract abstract

<contrib contrib-type="author" id="aut0005"><name><surname>Jensen</surname><given-names>Chantel N.</given-names></name><xref rid="aff0005" ref-type="aff">a</xref></contrib><contrib contrib-type="author" id="aut0010"><name><surname>Ali</surname><given-names>Sohail T.</given-names></name><xref rid="aff0010" ref-type="aff">b</xref></contrib><contrib contrib-type="author" id="aut0015"><name><surname>Allen</surname><given-names>Michael J.</given-names></name><xref rid="aff0010" ref-type="aff">b</xref></contrib><contrib contrib-type="author" id="aut0020"><name><surname>Grogan</surname><given-names>Gideon</given-names></name><email>gideon.grogan@york.ac.uk</email><xref rid="aff0005" ref-type="aff">a</xref><xref rid="cor0005" ref-type="corresp">&#x0204e;</xref></contrib><aff id="aff0005"><label>a</label>York Structural Biology Laboratory, Department of Chemistry, University of York, Heslington, York YO10 5DD, UK</aff><aff id="aff0010"><label>b</label>Plymouth Marine Laboratory, Prospect Place, Plymouth PL1 3DH, UK</aff>

Journal of Molecular Catalysis. B, Enzymatic

<sec id="sec0005"><label>1</label><title>Introduction</title><p id="par0025">Flavin-containing monooxygenases (FMOs) <xref rid="bib0005" ref-type="bibr">[1]</xref> catalyse the oxygenation of heteroatoms, such as nitrogen and sulfur, in various organic substrates, and have been studied both for their role in metabolism in higher eukaryotes, including humans <xref rid="bib0010" ref-type="bibr">[2]</xref>, <xref rid="bib0015" ref-type="bibr">[3]</xref>, and also for their contributions to microbial metabolism, in which they are able to catalyse the oxidation of amines <xref rid="bib0020" ref-type="bibr">[4]</xref> and amino acids such as ornithine <xref rid="bib0025" ref-type="bibr">[5]</xref>. In the case of microbial enzymes, the identification of interesting FMO activity has led to the exploitation of those enzymes for biotechnological applications <xref rid="bib0030" ref-type="bibr">[6]</xref>, <xref rid="bib0035" ref-type="bibr">[7]</xref>, in which the enzymatic characteristics of high turnover rates and chemo-, regio- and enantioselectivity, are very valuable. FMOs typically employ the phosphorylated nicotinamide cofactor NADPH to reduce a molecule of FAD within the enzyme, which then reacts with molecular oxygen to form a (hydro)peroxy flavin species that is the catalytic oxidant in reaction <xref rid="bib0005" ref-type="bibr">[1]</xref>. A well-studied example of a bacterial FMO is that from <italic>Methylophaga aminisulfidivorans</italic> (Uniprot Q83XK4, mFMO), an enzyme of monomer molecular weight 46&#x000a0;kDa that was identified on the basis of its ability to form the pigment indigo through oxidative transformation of indole <xref rid="bib0040" ref-type="bibr">[8]</xref>. mFMO has also been shown to catalyse the asymmetric sulfoxidation of a series of prochiral thioethers, when employed as part of a fusion enzyme with phosphite dehydrogenase for the recycling of the nicotinamide cofactor <xref rid="bib0045" ref-type="bibr">[9]</xref>. Studies of mFMO have shown a dependence for the phosphorylated cofactor NADPH, but the lower cost of the non-phosphorylated analogue, NADH, has meant that recent studies of FMOs have been directed towards enzymes that might employ that cofactor for flavin reduction.</p><p id="par0030">We recently reported the cloning, expression and structural characterisation of another FMO, named SMFMO, from the marine bacterium <italic>Stenotrophomonas maltophilia</italic> <xref rid="bib0050" ref-type="bibr">[10]</xref>. This target was interesting as it displayed the ability to use either NADPH or NADH as the cofactor for reduction of the flavin. SMFMO was hence able to use NADH, along with a formate dehydrogenase/sodium formate based recycling system, to catalyse the asymmetric oxidation of thioethers, and also the Baeyer&#x02013;Villiger oxidation of a strained, fused cyclobutanone substrate <xref rid="bib0050" ref-type="bibr">[10]</xref>. Related activities have recently been described by the group of Fraaije, who have characterised other FMOs, notably FMO-D, from <italic>Rhodococcus jostii</italic> RHA1, related to SMFMO, and which are similarly able to employ NADPH or NADH <xref rid="bib0055" ref-type="bibr">[11]</xref>. The structure of SMFMO was determined <xref rid="bib0050" ref-type="bibr">[10]</xref>, and analysis of the nicotinamide cofactor binding loop revealed differences between NADPH-dependent mFMO <xref rid="bib0060" ref-type="bibr">[12]</xref>, <xref rid="bib0065" ref-type="bibr">[13]</xref>, <xref rid="bib0070" ref-type="bibr">[14]</xref> and SMFMO that might be significant in the recognition of the NADPH 2&#x02032; ribose phosphate that distinguishes NADPH and NADH <xref rid="bib0050" ref-type="bibr">[10]</xref>. In this region in mFMO, Arg234 and Thr235 project towards the phosphate and interact directly with the phosphate oxygen atoms, whereas in SMFMO Gln193 and His194 are found in equivalent positions. A double mutant of SMFMO that was designed to mimic the phosphate binding loop of mFMO, changed the preference of the enzyme for NADPH to NADH from a ratio of 1.5:1 to 1:3.5 <xref rid="bib0075" ref-type="bibr">[15]</xref>. These mutations were not successful in removing activity with NADH, however. Multiple studies on the wider group of NAD(P)H-dependent flavoprotein monooxygenases (FPMOs) have shown that, whilst positively charged basic residues are often involved in the specific recognition of negatively-charged phosphate <xref rid="bib0080" ref-type="bibr">[16]</xref>, <xref rid="bib0085" ref-type="bibr">[17]</xref>, <xref rid="bib0090" ref-type="bibr">[18]</xref>, NADH-dependent activity can be engineered through the mutation of the cofactor binding loop to include a negatively charged carboxylate side chain enzymes that excludes phosphate, presumably through charge repulsion <xref rid="bib0095" ref-type="bibr">[19]</xref>. Engineering a glutamate residue into the cofactor-binding loop of SMFMO, in an attempt to generate a more NADH-specific variant, resulted in a mutant that was not produced in the soluble fraction of the <italic>Escherichia coli</italic> strain used for gene expression, however <xref rid="bib0075" ref-type="bibr">[15]</xref>. In this report, we describe the cloning, expression, and characterisation of two homologs of SMFMO, CFMO from <italic>Cellvibrio</italic> sp. BR (Uniprot code I3IEE4) and PSFMO from <italic>Pseudomonas stutzeri</italic> NF13 (M2V3J0). These homologs display natural variation in the cofactor-binding loop, Thr&#x02013;Ser in CFMO and Gln&#x02013;Glu in PSFMO, which suggested there may be altered cofactor preference compared to either mFMO or SMFMO. The enzyme activity with NADH and NADPH and a range of prochiral sulfides is assessed, and the structures of the enzymes, which reveal the context of the substituted amino acids within the putative cofactor binding loop, are presented.</p></sec><sec id="sec0010"><label>2</label><title>Experimental</title><sec id="sec0015"><label>2.1</label><title>Chemicals</title><p id="par0035">Chemicals, including media and buffer components, sulfide substrates and cofactors were purchased from Sigma-Aldrich (Poole U.K.).</p></sec><sec id="sec0020"><label>2.2</label><title>Gene synthesis, cloning, expression and protein purification</title><p id="par0040">The genes encoding CFMO and PSFMO were synthesised by GeneArt (Invitrogen), with sequences optimised for expression in <italic>E. coli</italic> using the GeneArt server program. Genes were then amplified by PCR from the commercial genes using the following primers: For CFMO: Forward: CCAGGGACCAGCAATGGATACACCGGTTATGG; Reverse: GAGGAGAAGGCGCGTTAGGCGCTATCCAGATACTG; For PSFMO: Forward: CCAGGGACCAGCAATGCCTCCGATTCTGG; Reverse: GAGGAGAAGGCGCGTTACGGACGACGGCTCGG. PCRs were analysed on agarose gels, and bands of the expected size were isolated using a PCR Cleanup kit^&#x000ae; (Qiagen). Target genes were then sub-cloned into the pET-YSBL-LIC-3C vector following a previously published procedure <xref rid="bib0100" ref-type="bibr">[20]</xref>. The recombinant plasmids were then used to transform cells of <italic>E. coli</italic> XL1-Blue (Novagen), which, after transformation and overnight growth on LB agar containing 30&#x000a0;&#x003bc;g&#x000a0;mL^&#x02212;1 kanamycin as antibiotic marker, were subjected to miniprep procedures that resulted in plasmids suitable for DNA sequencing. Once the sequence of the genes had been confirmed, gene expression was conducted by transforming cells of <italic>E. coli</italic> BL21 (DE3) with the recombinant plasmids. 5&#x000a0;mL of LB medium containing 30&#x000a0;&#x003bc;g&#x000a0;mL^&#x02212;1 kanamycin was inoculated with a single colony of the relevant strain. This starter culture was grown at 37&#x000a0;&#x000b0;C overnight with shaking at 180 r.p.m. Each 5&#x000a0;mL culture was then used to incolulate 500&#x000a0;mL LB broth containing 30&#x000a0;&#x003bc;g&#x000a0;mL^&#x02212;1 kanamycin in a 2&#x000a0;L Erlenmeyer flask. These larger cultures were grown with shaking at 37&#x000a0;&#x000b0;C until the optical density, as determined by measurement at 600&#x000a0;nm, had reached 0.8. The cultures were then induced through the addition of 1&#x000a0;mM isopropyl &#x003b2;-D-1-thiogalactopyranoside (IPTG), and growth continued at 18&#x000a0;&#x000b0;C overnight. Cells were then harvested by centrifugation for 15&#x000a0;min at 4225&#x000a0;g using a Sorvall GS3 rotor in a Sorvall RC5B Plus centrifuge. Following centrifugation, the resultant cell pellets were resuspended in 25&#x000a0;mL 50&#x000a0;mM Tris/HCl buffer pH 7.5, containing 300&#x000a0;mM sodium chloride (&#x02018;buffer&#x02019;) per L of cell growth. These suspensions were then subjected to cell disruption using an ultrasonicator for 3&#x000a0;&#x000d7;&#x000a0;30&#x000a0;s periods at 4&#x000a0;&#x000b0;C with intervals of 1&#x000a0;min. The soluble fraction after sonication was recovered by centrifuging the suspension for 30&#x000a0;min at 26,892&#x000a0;g in a Sorvall SS34 rotor. Supernatants were then filtered using a 2&#x000a0;&#x003bc;m Amicon filter, and then subjected t nickel affinity chromatography using a 5&#x000a0;mL His-Trap&#x02122; Chelating HP column. After loading the filtered protein solution, the column was washed with five column volumes of buffer containing imidazole (30&#x000a0;mM). The FMOs were then eluted from the column using a 30&#x02013;500&#x000a0;mM imidazole gradient over twenty column volumes. Column fractions containing FMOs were identified using SDS-PAGE and combined. Pooled fractions were concentrated, typically, to a volume of 4&#x000a0;mL using a Centricon^&#x000ae; filter membrane (10&#x000a0;kDa cut-off) and 2&#x000a0;mL of this solution then loaded onto an S75 Superdex&#x02122; 16/60 size exclusion column that had been pre-equilibrated with buffer. FMOs were eluted with buffer at a flow rate of 1&#x000a0;mL&#x000a0;min^&#x02212;1. Fractions were analysed by SDS-PAGE and those that contained pure FMOs pooled and stored at 4&#x000a0;&#x000b0;C for crystallisation, enzyme assays or biotransformations. For the purposes of crystallisation, the histidine tags of CFMO or PSFMO were cleaved using 3&#x000a0;C protease and using a procedure described previously <xref rid="bib0100" ref-type="bibr">[20]</xref>. Typical CFMO and PSFMO preparations yielded 20&#x000a0;mg and 7.5&#x000a0;mg pure protein per litre of cells, respectively.</p></sec><sec id="sec0025"><label>2.3</label><title>Enzyme assays</title><p id="par0045">Steady-state kinetic constants for the NADH and NADPH-dependent reduction of FAD by the FMOs were determined using the method employed previously <xref rid="bib0050" ref-type="bibr">[10]</xref>, <xref rid="bib0105" ref-type="bibr">[21]</xref>. In a 1&#x000a0;mL quartz cuvette containing Tris&#x02013;HCl buffer pH 7.5 (50&#x000a0;&#x003bc;mol) the decrease in absorbance at 340&#x000a0;nm was monitored for concentrations of NAD(P)H (10&#x02013;100&#x000a0;&#x003bc;M) after the addition of enzyme (CFMO or PSFMO, 3.9&#x000a0;nmol). All data points represented the average of three separate runs. Kinetic constants (<italic>K</italic>_{<italic>M</italic>} and <italic>k</italic>_cat) were calculated using a value for <italic>&#x0025b;</italic> of 6220&#x000a0;mol&#x000a0;dm^&#x02212;3&#x000a0;cm^&#x02212;1 using Grafit&#x02122; (Erithacus Sofware).</p></sec><sec id="sec0030"><label>2.4</label><title>Biotransformations</title><p id="par0050">Biotransformations using isolated enzymes with cofactor recycling were performed using the method previously described for SMFMO <xref rid="bib0050" ref-type="bibr">[10]</xref>. For NADH-dependent biotransformations: To a 10&#x000a0;mL round bottomed flask containing Tris&#x02013;HCl buffer pH 7.5 (5&#x000a0;mL) were added substrate(s) (<bold>1</bold>&#x02013;<bold>6</bold> in 100&#x000a0;&#x003bc;L ethanol to a final concentration of 5&#x000a0;mM), NADH (5&#x000a0;mg, a final concentration of 0.7&#x000a0;mM), formate dehydrogenase (5&#x000a0;mg), sodium formate (6.8&#x000a0;mg, 0.1&#x000a0;mmol) and CFMO or PSFMO (1&#x000a0;mL of a 5&#x000a0;mg&#x000a0;mL^&#x02212;1 solution, 0.13&#x000a0;&#x003bc;mol). The reactions were then stirred for 24&#x000a0;h at room temperature. Aliquots (500&#x000a0;&#x003bc;L) were taken at intervals and extracted with ethyl acetate (500&#x000a0;&#x003bc;L). The organic layer was transferred to a GC vial and analysed by GC as described previously <xref rid="bib0050" ref-type="bibr">[10]</xref>. For NADPH-dependent biotransformations: To a 10&#x000a0;mL round bottomed flask containing Tris&#x02013;HCl buffer pH 7.5 (5&#x000a0;mL) were added substrate(s) (<bold>1</bold>&#x02013;<bold>6</bold> in 100&#x000a0;&#x003bc;L ethanol to a final concentration of 5&#x000a0;mM), NADPH (5.7&#x000a0;mg, a final concentration of 0.7&#x000a0;mM), glucose-6-phosphate-dehydrogenase (0.14&#x000a0;mg), glucose-6-phosphate (5.2&#x000a0;mg, 0.02&#x000a0;mmol) and CFMO or PSFMO (1&#x000a0;mL of a 5&#x000a0;mg&#x000a0;mL^&#x02212;1 solution, 0.13&#x000a0;&#x003bc;mol). The reactions were then stirred for 24&#x000a0;h at room temperature and organic extracts of 500&#x000a0;&#x003bc;L aliquots analysed as previously <xref rid="bib0050" ref-type="bibr">[10]</xref>. Chiral analysis of sulfoxide products was carried out using BGB 173 and BGB-175 columns (30&#x000a0;m&#x000a0;&#x000d7;&#x000a0;0.25&#x000a0;mm&#x000a0;&#x000d7;&#x000a0;0.25&#x000a0;&#x003bc;m; each from BGB-Analytik) according to procedures described previously <xref rid="bib0050" ref-type="bibr">[10]</xref>.</p></sec><sec id="sec0035"><label>2.5</label><title>Protein crystallisation</title><p id="par0055">Pure PSFMO and CFMO was subjected to crystallisation trials using a range of commercially available screens in 96-well plates employing 300&#x000a0;nL drops at a range of protein concentrations (3, 10 and 20&#x000a0;mg&#x000a0;mL^&#x02212;1). The best crystals for PSFMO were obtained using the Clear Strategy Screen (CSS) <xref rid="bib0110" ref-type="bibr">[22]</xref> conditions containing 35% (w/v) tacsimate pH 7.0 and His₆-tag cleaved protein at a concentration of 20&#x000a0;mg&#x000a0;mL^&#x02212;1. The best crystals for CFMO were obtained using the CSS conditions containing 1.5&#x000a0;M ammonium sulfate and non-cleaved protein at 20&#x000a0;mg&#x000a0;mL^&#x02212;1. Larger crystals for diffraction analysis using optimised conditions were prepared using the hanging-drop vapour diffusion method in 24-well plate Linbro dishes and using crystallisation drops of 2&#x000a0;&#x003bc;L, comprised of 1&#x000a0;&#x003bc;L of reservoir solution and 1&#x000a0;&#x003bc;L of protein solution at 20&#x000a0;mg&#x000a0;mL^&#x02212;1. For PSFMO, the best crystals were again obtained in crystal drops containing 35% (w/v) tacsimate at pH 7.0 with no further additions. For CFMO, the best crystals were obtained in crystal drops containing 1.5&#x000a0;M ammonium sulfate and 1% propan-2-ol (v/v) at pH 7.0. Crystals were flash-cooled in a cryogenic solution containing the mother liquor with 10% (v/v) glycerol, and tested for diffraction using a Rigaku Micromax-007HF generator fitted with Osmic multilayer optics and a MARRESEARCH MAR345 imaging plate detector. Crystals that diffracted to a resolution of greater than 3&#x000a0;&#x0212b; were retained for full dataset collection at the synchrotron.</p></sec><sec id="sec0040"><label>2.6</label><title>Data collection, structure solution, model building and refinement of CFMO and PSFMO</title><p id="par0060">Datasets described herein were collected at the Diamond Light Source, Didcot, Oxfordshire, U.K. Data for CFMO and PSFMO were each collected on beamline I24. Data were processed and integrated using XDS <xref rid="bib0115" ref-type="bibr">[23]</xref> and scaled using SCALA <xref rid="bib0120" ref-type="bibr">[24]</xref> included in the Xia2 processing system <xref rid="bib0125" ref-type="bibr">[25]</xref>. Data collection statistics are given in <xref rid="tbl0005" ref-type="table">Table 1</xref>. The crystals of CFMO were in space group <italic>C</italic>2. The structure of CFMO was solved using MOLREP <xref rid="bib0130" ref-type="bibr">[26]</xref>, using a monomer model of SMFMO (PDB code 4a9w) <xref rid="bib0050" ref-type="bibr">[10]</xref>. The solution contained two molecules in the asymmetric unit, representing one dimer, and the solvent content was 55%. The crystals of PSFMO were in space group <italic>P</italic>3₂21. The structure of PSFMO was again solved using SMFMO as a model, but in this case, the solution contained only one molecule in the asymmetric unit. The solvent content in this case was 57%. The structures of CFMO and PSFMO were built and refined using iterative cycles using Coot <xref rid="bib0135" ref-type="bibr">[27]</xref> and REFMAC <xref rid="bib0140" ref-type="bibr">[28]</xref>, employing local NCS restraints in refinement. The final structures exhibited <italic>R</italic>_cryst and <italic>R</italic>_free values of 24.6 and 28.7 (CFMO) and 16.2 and 20.4% (PSFMO), respectively. Each structure was validated prior to deposition using PROCHECK <xref rid="bib0145" ref-type="bibr">[29]</xref>. Refinement statistics for all structures are presented in <xref rid="tbl0005" ref-type="table">Table 1</xref>. The Ramachandran plot for CFMO showed 93.4% of residues to be situated in the most favoured regions, 6.2% in additional allowed and 0.5% residues in outlier regions. For PSFMO, the corresponding values were 92.8%, 4.6% and 2.6%, respectively. The coordinates and structure factors for CFMO and PSFMO have been deposited in the Protein Data Bank with the accession codes 4usq and 4usr, respectively.<table-wrap position="float" id="tbl0005"><label>Table 1</label><caption><p>Data collection and resolution statistics for CFMO and PSFMO. Values for the highest resolution shells are given in parentheses.</p></caption><table frame="hsides" rules="groups"><thead><tr><th/><th align="left">CFMO</th><th align="left">PSFMO</th></tr></thead><tbody><tr><td align="center">Beamline</td><td align="center">Diamond i24</td><td align="center">Diamond i24</td></tr><tr><td align="center">Wavelength (&#x000c5;)</td><td align="center">0.96862</td><td align="center">0.96862</td></tr><tr><td align="center">Resolution (&#x000c5;)</td><td align="center">22.44&#x02013;2.39 (2.45&#x02013;2.39)</td><td align="center">35.94&#x02013;1.83 (1.88&#x02013;1.83)</td></tr><tr><td align="center">Space group</td><td align="center"><italic>C</italic>2</td><td align="center"><italic>P</italic>3₂21</td></tr><tr><td rowspan="2" align="center">Unit cell</td><td align="center"><italic>a</italic>&#x000a0;=&#x000a0;115.41; <italic>b</italic>&#x000a0;=&#x000a0;95.09; <italic>c</italic>&#x000a0;=&#x000a0;92.37</td><td align="center"><italic>a</italic>&#x000a0;=&#x000a0;<italic>b</italic>&#x000a0;=&#x000a0;63.56; <italic>c</italic>&#x000a0;=&#x000a0;189.82</td></tr><tr><td align="center"><italic>&#x003b1;</italic>&#x000a0;<italic>=</italic>&#x000a0;<italic>&#x003b2;</italic>&#x000a0;=&#x000a0;90.0; <italic>&#x003b3;</italic>&#x000a0;<italic>=</italic>&#x000a0;126.3</td><td align="center"><italic>&#x003b1;</italic>&#x000a0;<italic>=</italic>&#x000a0;<italic>&#x003b2;</italic>&#x000a0;<italic>=</italic>&#x000a0;90.0; <italic>&#x003b3;</italic>&#x000a0;=&#x000a0;120.0</td></tr><tr><td align="center">No. of molecules in the asymmetric unit</td><td align="center">2</td><td align="center">1</td></tr><tr><td align="center">Unique reflections</td><td align="center">31344 (2329)</td><td align="center">41893 (3176)</td></tr><tr><td align="center">Completeness (%)</td><td align="center">98.1 (98.4)</td><td align="center">100 (99.9)</td></tr><tr><td align="center"><italic>R</italic>_merge (%)</td><td align="center">0.09 (0.50)</td><td align="center">0.112 (0.62)</td></tr><tr><td align="center"><italic>R</italic>_{p. i. m.}</td><td align="center">0.09 (0.50)</td><td align="center">0.054 (0.30)</td></tr><tr><td align="center">Multiplicity</td><td align="center">3.2 (2.9)</td><td align="center">9.8 (9.9)</td></tr><tr><td align="center">&#x02329;<italic>I/&#x003c3;</italic>(<italic>I</italic>)&#x0232a;</td><td align="center">9.0 (2.1)</td><td align="center">14.9 (3.8)</td></tr><tr><td align="center">CC_1/2</td><td align="center">0.99 (0.76)</td><td align="center">1.00 (0.90)</td></tr><tr><td align="center">Overall <italic>B</italic> factor from Wilson plot (&#x000c5;²)</td><td align="center">25</td><td align="center">13</td></tr><tr><td align="center"><italic>R</italic>_cryst/<italic>R</italic>_free (%)</td><td align="center">24.6/28.7</td><td align="center">16.2/20.4</td></tr><tr><td align="center">r.m.s.d. 1&#x02013;2 bonds (&#x000c5;)</td><td align="center">0.014</td><td align="center">0.02</td></tr><tr><td align="center">r.m.s.d. 1&#x02013;3 bonds (&#x000b0;)</td><td align="center">1.66</td><td align="center">2.26</td></tr><tr><td align="center">Avge main chain B (&#x000c5;²)</td><td align="center">38</td><td align="center">19</td></tr><tr><td align="center">Avge side chain B (&#x000c5;²)</td><td align="center">40</td><td align="center">22</td></tr><tr><td align="center">Avge water B (&#x000c5;²)</td><td align="center">31</td><td align="center">24</td></tr></tbody></table></table-wrap></p></sec></sec><sec id="sec0045"><label>3</label><title>Results and discussion</title><sec id="sec0050"><label>3.1</label><title>Target selection</title><p id="par0065">Following the failure to incorporate a glutamate residue within the cofactor binding loop of SMFMO that might engender discriminatory binding of NADH over NADP in that enzyme, it was decided to select targets from the genomic databanks representative of natural variation within that loop, particularly corresponding to the Arg&#x02013;Thr and Gln&#x02013;His couples of mFMO of SMFMO, respectively. It was hoped that these variants might be sufficiently different within the loop to display a shift in preference for one nicotinamide cofactor, preferably NADH, over NADPH. CFMO (Uniprot code I3IEE4) and PSFMO (M2V3J0) were selected on this basis. Each is a putative FMO of a similar size to SMFMO (361 and 358 amino acids, respectively, <italic>versus</italic> 357 for SMFMO) and contains two Rossman domains and the FXGXXXHXXXY FMO motif <xref rid="bib0150" ref-type="bibr">[30]</xref>. A sequence alignment of these targets and SMFMO (<xref rid="fig0005" ref-type="fig">Fig. 1</xref>) revealed 58% and 61% sequence identity between CFMO and SMFMO and PSFMO and SMFMO, respectively. In place of the Arg234Thr235 couple in mFMO or Gln193His194 couple in SMFMO, CFMO possessed Ser202 and Thr203, and PSFMO, Gln194 and Glu195, respectively. The latter was particularly interesting in being one of the only homologs identified from database searches as having either a glutamate or aspartate residue within the putative phosphate recognition region. No homolog was identified yet that possessed a Glu or Asp residue in place of Gln193 in SMFMO. Genes encoding CFMO and PSFMO were synthesised, subcloned and expressed in the soluble fractions of transformed strains of <italic>E. coli</italic> BL21 (DE3). The proteins were readily purified using nickel affinity and size exclusion chromatography as described in Section <xref rid="sec0010" ref-type="sec">2</xref>, each yielding protein solutions of a bright yellow colour, indicative of the presence of bound FAD.<fig id="fig0005"><label>Fig. 1</label><caption><p>Sequence alignment of SMFMO, CFMO and PSFMO. The characteristic Rossman motifs are highlighted in blue; the FMO motif <xref rid="bib0150" ref-type="bibr">[30]</xref> in green, and the two residues proposed to be the closest to the 2&#x02032; ribose hydroxyl phosphate in red. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)</p></caption><graphic xlink:href="gr1"/></fig></p></sec><sec id="sec0055"><label>3.2</label><title>Cofactor promiscuity in the reduction of FAD</title><p id="par0070">The ability of CFMO and PSFMO to use either NADH or NADPH to reduce FAD was assessed using standard UV spectrophotometry assays <xref rid="bib0050" ref-type="bibr">[10]</xref>, <xref rid="bib0105" ref-type="bibr">[21]</xref>, in which the oxidation of NAD(P)H was monitored at 340&#x000a0;nm. Kinetic parameters (<italic>K</italic>_m and <italic>k</italic>_cat) were determined for each enzyme and each cofactor and the results are shown in <xref rid="tbl0010" ref-type="table">Table 2</xref>, with data for SMFMO <xref rid="bib0050" ref-type="bibr">[10]</xref> reproduced for comparison.<table-wrap position="float" id="tbl0010"><label>Table 2</label><caption><p>Kinetic constants for CFMO and PSFMO in the reduction of flavin by either NADH or NADPH. Values for SMFMO, reproduced from Ref. <xref rid="bib0050" ref-type="bibr">[10]</xref> are also given for comparison.</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="center">Enzyme/cofactor</th><th align="center"><italic>K</italic>_M (&#x003bc;M)</th><th align="center"><italic>k</italic>_cat (s^&#x02212;1)</th><th align="center"><italic>k</italic>_cat/<italic>K</italic>_M (M^&#x02212;1&#x000a0;s^&#x02212;1)</th></tr></thead><tbody><tr><td align="center">SMFMO<xref rid="tblfn0005" ref-type="table-fn">*</xref></td><td align="center">24&#x000a0;&#x000b1;&#x000a0;9</td><td align="center">0.03</td><td align="center">1250</td></tr><tr><td align="center">NADH</td><td/><td/><td/></tr><tr><td align="center">SMFMO</td><td align="center">27&#x000a0;&#x000b1;&#x000a0;5</td><td align="center">0.02</td><td align="center">740</td></tr><tr><td align="center">NADPH</td><td/><td/><td/></tr><tr><td align="center">CFMO</td><td align="center">3&#x000a0;&#x000b1;&#x000a0;1</td><td align="center">0.02</td><td align="center">6670</td></tr><tr><td align="center">NADH</td><td/><td/><td/></tr><tr><td align="center">CFMO</td><td align="center">5&#x000a0;&#x000b1;&#x000a0;1</td><td align="center">0.02</td><td align="center">4000</td></tr><tr><td align="center">NADPH</td><td/><td/><td/></tr><tr><td align="center">PSFMO</td><td align="center">16&#x000a0;&#x000b1;&#x000a0;2</td><td align="center">0.03</td><td align="center">1880</td></tr><tr><td align="center">NADH</td><td/><td/><td/></tr><tr><td align="center">PSFMO</td><td align="center">17&#x000a0;&#x000b1;&#x000a0;2.0</td><td align="center">0.04</td><td align="center">2350</td></tr><tr><td align="center">NADPH</td><td/><td/><td/></tr></tbody></table><table-wrap-foot><fn id="tblfn0005"><label>*</label><p id="npar0005">Data taken from Ref. <xref rid="bib0050" ref-type="bibr">[10]</xref>.</p></fn></table-wrap-foot></table-wrap></p><p id="par0075">The first observation was that both CFMO and PSFMO were both able to accept NADH or NADPH for the reduction of FAD. CFMO was observed to bind both NADH and NADPH with increased affinity compared to SMFMO, as illustrated by <italic>K</italic>_m values of 3 and 5&#x000a0;&#x003bc;M compared to values in the range of 24&#x02013;27&#x000a0;&#x003bc;M for NADH and NADPH, respectively. Although <italic>k</italic>_cat values were lower for CFMO than the other two enzymes, the low <italic>K</italic>_m contributed to CFMO having a catalytic activity, as determined by values of <italic>k</italic>_cat/<italic>K</italic>_M, of 5.3 and 5.4 fold greater than SMFMO for NADH and NADPH, respectively. These values also reveal a slight preference for NADH by a factor of 1.7, compared with SMFMO, for which the value was 1.5. For PSFMO, the <italic>K</italic>_M was similar with either NADH (16&#x000a0;&#x003bc;M) or NADPH (17&#x000a0;&#x003bc;M) which were each lower than for SMFMO. <italic>k</italic>_cat/<italic>K</italic>_M was slightly lower with NADH (1880&#x000a0;M^&#x02212;1&#x000a0;s^&#x02212;1) than NADPH (2350&#x000a0;M^&#x02212;1&#x000a0;s^&#x02212;1), indicating a preference for NADPH in this case of 1.3, but each value was again higher than that for SMFMO with either cofactor.</p></sec><sec id="sec0060"><label>3.3</label><title>Biotransformations of prochiral sulfides using CFMO and PSFMO using either NADH or NADPH as cofactor</title><p id="par0080">Both CFMO and PSFMO were then challenged with a range of prochiral sulfide substrates <bold>1</bold>&#x02013;<bold>6</bold> (<xref rid="fig0010" ref-type="fig">Fig. 2</xref>), with the addition of either catalytic NADH or NADPH as cofactor, and in the presence of a suitable cofactor recycling system, as had been used for SMFMO <xref rid="bib0050" ref-type="bibr">[10]</xref>. The results of all biotransformations by CFMO and PSFMO, compared to those obtained with SMFMO are shown in <xref rid="tbl0015" ref-type="table">Table 3</xref>. CFMO oxidised sulfides <bold>1</bold> and <bold>3&#x02013;6</bold> to, for the most part, their corresponding (<italic>R</italic>)-sulfoxides, with either cofactor. However, substrate <bold>1</bold> was transformed to the (<italic>S</italic>)-sulfoxide in the presence of NADPH. The conversions were overall higher with NADH, however, the enantiomeric excess of sulfoxide products was slightly higher with NADPH in general. Substrate <bold>2</bold> was not converted by CFMO in the presence of either cofactor. The highest conversion was seen for substrate <bold>3</bold>, methyl tolyl sulfide, with 65% for NADH (22% e.e. (<italic>R</italic>)-sulfoxide product) and 32% for NADPH (32% e.e.). The most enantioselective transformation was that of substrate <bold>6</bold> with 66% -(<italic>R</italic>) and 77% -(<italic>R</italic>) for achieved with NADH or NADPH, respectively. CFMO was overall less active than SMFMO with sulfide substrates and NADH, although both the conversion and product enantiomeric excess in the oxidation of phenyl methyl sulfide <bold>4</bold> were superior. Interestingly, when NADPH was employed as nicotinamide cofactor CFMO gave higher conversions compared to SMFMO, except for substrate <bold>3,</bold> possibly indicative of its greater activity as shown by the kinetic constants. For PSFMO, the highest conversions were observed for substrate <bold>3</bold> and <bold>4</bold>, but enantioselectivity was poor in each case. PSFMO again gave (<italic>R</italic>)-sulfoxide products for the most part. The most enantioselective reaction was observed for substrate <bold>6</bold>, yielding (<italic>R</italic>)-sulfoxide product of 85% and 57% with NADH and NADPH, respectively. Overall, PSFMO catalysed sulfoxidation reactions gave higher conversions than either SMFMO or CFMO, but with poorer e.e.s overall, save for substrate <bold>6</bold>.<fig id="fig0010"><label>Fig. 2</label><caption><p>Sulfide substrates screened in this study against CFMO and PSFMO.</p></caption><graphic xlink:href="gr2"/></fig><table-wrap position="float" id="tbl0015"><label>Table 3</label><caption><p>Results of biotransformations of prochiral sulfide substrates <bold>1</bold> and <bold>3&#x02013;6</bold> by CFMO or PSFMO using either NADH or NADPH as nicotinamide cofactor. Values for SMFMO, reproduced from Ref. <xref rid="bib0050" ref-type="bibr">[10]</xref> are also given for comparison.</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left">Sulfide</th><th colspan="3" align="left">NADH<hr/></th><th colspan="3" align="left">NADPH<hr/></th></tr><tr><th/><th colspan="3" align="left">Conversion (%); absolute configuration; e.e.<hr/></th><th colspan="3" align="left">Conversion (%); absolute configuration; e.e. (%)<hr/></th></tr><tr><th/><th align="left">SMFMO<xref rid="tblfn0010" ref-type="table-fn">*</xref></th><th align="left">CFMO</th><th align="left">PSFMO</th><th align="left">SMFMO<xref rid="tblfn0010" ref-type="table-fn">*</xref></th><th align="left">CFMO</th><th align="left">PSFMO</th></tr></thead><tbody><tr><td align="left">1</td><td align="left">27, (<italic>R</italic>)-, 71% e.e.</td><td align="left">17, (<italic>R</italic>)-, 54% e.e.</td><td align="left">61, (<italic>R</italic>)-, 30% e.e.</td><td align="left">2, (<italic>R</italic>)-, 57% e.e.</td><td align="left">6, (<italic>S</italic>)-, 43% e.e.</td><td align="left">28, (<italic>R</italic>)-, 24% e.e.</td></tr><tr><td align="left">3</td><td align="left">90, (<italic>R</italic>)-, 25% e.e.</td><td align="left">65, (<italic>R</italic>)-, 22% e.e.</td><td align="left">97, (<italic>R</italic>)-, 47% e.e.</td><td align="left">33, (<italic>R</italic>)-, 44% e.e.</td><td align="left">32, (<italic>R</italic>)-, 32% e.e.</td><td align="left">78, (<italic>R</italic>)-, 32% e.e.</td></tr><tr><td align="left">4</td><td align="left">8, (<italic>R</italic>)-, 21% e.e.</td><td align="left">64, (<italic>R</italic>)-, 58% e.e.</td><td align="left">99, (<italic>R</italic>)-, 14% e.e.</td><td align="left">1, n.d.</td><td align="left">38, (<italic>R</italic>)-, 64% e.e.</td><td align="left">73, (<italic>S</italic>)-, 4% e.e.</td></tr><tr><td align="left">5</td><td align="left">6, (<italic>S</italic>)-, 15% e.e.</td><td align="left">No conversion</td><td align="left">13, (<italic>S</italic>)-, 10% e.e.</td><td align="left">&#x0003c;1, n.d.</td><td align="left">11, (<italic>R</italic>)-, 3% e.e.</td><td align="left">No conversion</td></tr><tr><td align="left">6</td><td align="left">40, (<italic>R</italic>)-, 80% e.e.</td><td align="left">14, (<italic>R</italic>)-, 66% e.e.</td><td align="left">50, (<italic>R</italic>)-, 85% e.e.</td><td align="left">9, (<italic>R</italic>)-, 82% e.e.</td><td align="left">47, (<italic>R</italic>)-, 77% e.e.</td><td align="left">22, (<italic>R</italic>)-, 57% e.e.</td></tr></tbody></table><table-wrap-foot><fn id="tblfn0010"><label>*</label><p id="npar0010">Data taken from Ref. <xref rid="bib0050" ref-type="bibr">[10]</xref>.</p><p id="npar0015">n.d.&#x000a0;=&#x000a0;not determined.</p></fn></table-wrap-foot></table-wrap></p></sec><sec id="sec0065"><label>3.4</label><title>Structures of CFMO and PSFMO</title><p id="par0085">In order to shed light on the nature of the cofactor binding loops in CFMO and PSFMO, the structures of each in complex with the flavin FAD were determined to resolutions of 2.39&#x000a0;&#x0212b; and 1.83&#x000a0;&#x0212b;, respectively. Crystals of CFMO grew in the <italic>C</italic>2 space group, with two molecules &#x02018;A&#x02019; and &#x02018;B&#x02019;, representing one dimer in the asymmetric unit. The CFMO dimer was made up of two monomers (<xref rid="fig0015" ref-type="fig">Fig. 3</xref>b), sharing an interfacial area of approximately 1207&#x000a0;&#x0212b;². Analysis of the CFMO structure using PISA <xref rid="bib0155" ref-type="bibr">[31]</xref> found that the interactions that stabilise the dimer included six hydrogen bonds, including those between the backbone nitrogen of Trp108(A) and the side chain oxygen of Glu124(B) and the side chain N&#x02014;H of Gln320(A) with the backbone carbonyl of Ile145(B). The calculated dissocation energy of the dimer interface (&#x00394;ⁱ<italic>G</italic>) was &#x02212;4.4&#x000a0;kcal&#x000a0;mol^&#x02212;1. The monomer of CFMO superposed with the structure of the SMFMO monomer with an r.m.s.d. of 0.77&#x000a0;&#x0212b; over 317 C&#x003b1; atoms. The secondary elements of the structure are summarised in <xref rid="fig0015" ref-type="fig">Fig. 3</xref>a. Each CFMO monomer consists of two domains, an FAD binding domain and the putative substrate binding domain. Electron density was visible for the majority of the backbone in each monomer, from residue Ser13 to Ala361, with a stretch of missing density corresponding to six amino acids between positions Gln236 and Asp243 (PVGGLG) in subunit A and fifteen between Ala227 and Asp243 (AQEGREIEQPVGGLG) in subunit B that could not be modelled. Side chain density in the region of residues 224&#x02013;236 in subunit A was also missing. It appears that the residues for which there is poor electron density are part of a flexible helix&#x02013;loop&#x02013;helix structure, incorporating helix &#x003b1;8, which is found over the FAD binding pocket, perhaps shielding the active site from bulk solvent. An equivalent region of density was missing in structures of both wild-type SMFMO (4A9W) <xref rid="bib0050" ref-type="bibr">[10]</xref> and the Gln193Arg/His194Thr mutant (4C5O) <xref rid="bib0075" ref-type="bibr">[15]</xref>. There was substantial residual density in the omit map in the putative active site following building and refinement of the protein atoms of CFMO. This was successfully modelled and refined as FAD in the flat, oxidised form.<fig id="fig0015"><label>Fig. 3</label><caption><p>(a) Primary structure of CFMO, with secondary structure assignment, created using DSSP <xref rid="bib0160" ref-type="bibr">[32]</xref>, <xref rid="bib0165" ref-type="bibr">[33]</xref> and represented using ALINE <xref rid="bib0170" ref-type="bibr">[34]</xref>. (b) Structure of dimer of CFMO, in ribbon format, with subunit <bold>A</bold> in green and <bold>B</bold> in coral. One FAD molecule per monomer is shown in cylinder format with carbon atoms in grey. Helix &#x003b1;8, which forms part of the flexible helix&#x02013;loop&#x02013;helix structure that may shield the active site from bulk solvent, is also highlighted. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)</p></caption><graphic xlink:href="gr3"/></fig></p><p id="par0090">Crystals of PSFMO grew in the <italic>P</italic>3₂21 space group, with one molecule in the asymmetric unit, although both size-exclusion studies (not shown), and the relationship of the monomer to its closest crystallographic symmetry partner, are strongly suggestive of an active dimer for PSFMO as with SMFMO and CFMO. Analysis of a dimer pair by PISA <xref rid="bib0155" ref-type="bibr">[31]</xref> revealed an interfacial area of 955&#x000a0;&#x0212b;², with six H-bonds and 8 salt bridges formed between the monomers. The calculated dissociation energy of the dimer interface (&#x00394;ⁱ<italic>G</italic>) was &#x02212;7.1&#x000a0;kcal&#x000a0;mol^&#x02212;1. The secondary structural elements are summarised in <xref rid="fig0020" ref-type="fig">Fig. 4</xref>a and the monomer structure is shown in <xref rid="fig0020" ref-type="fig">Fig. 4</xref>b. In the case of PSFMO, there was electron density for each amino acid in the monomer from residue Met1 to Pro354, which allowed for the first time the observation of the entire backbone of the flexible helix&#x02013;loop&#x02013;helix structure, represented by residues 207&#x02013;234 in PSFMO, and incorporating helices &#x003b1;9 and &#x003b1;10, over FAD that had been missing from SMFMO structures and only partially present in CFMO, although side-chain density for residues 222&#x02013;227 was poor. Electron density was also present for residues Gly-2, Pro-1 and Ala 0 at the N-terminus, which represent the first part of the linker to the His-tag in the protein produced using the YSBLIC-3C construct. The monomer of PSFMO superposed with the structure of the SMFMO with an r.m.s.d. of 0.90&#x000a0;&#x0212b; over 327 C&#x003b1; atoms. FAD was again clearly visible in the omit map after building and refinement of the protein atoms, and was also modelled and refined successfully in the oxidised form, the density revealing no puckering of the tricyclic ring system that might be indicative of the reduced flavin.<fig id="fig0020"><label>Fig. 4</label><caption><p>(a) Primary structure of PSFMO, with secondary structure assignment, created using DSSP <xref rid="bib0160" ref-type="bibr">[32]</xref>, <xref rid="bib0165" ref-type="bibr">[33]</xref> and represented using ALINE <xref rid="bib0170" ref-type="bibr">[34]</xref>. (b) Structure of PSFMO monomer, in ribbon format. One FAD molecule per monomer is shown in cylinder format with carbon atoms in grey. Helix &#x003b1;9, which with helix &#x003b1;10 forms part of the flexible helix&#x02013;loop&#x02013;helix structure over the active site, and for which electron density is continuous in the PSFMO structure, is also highlighted.</p></caption><graphic xlink:href="gr4"/></fig></p></sec><sec id="sec0070"><label>3.5</label><title>The FAD binding region in CFMO and PSFMO</title><p id="par0095">In both CFMO and PSFMO, the tricyclic isoalloxazine ring of FAD was bound at the domain interface in a cavity beneath the flexible helix&#x02013;loop&#x02013;helix region of each enzyme (<xref rid="fig0025" ref-type="fig">Fig. 5</xref>a and b). The oxygen atoms of the pyrimidinedione ring form H-bonds with the backbone N&#x02014;Hs of CFMO Phe62 (PSFMO Phe53) and Leu340 (Leu322). In common with SMFMO, the aromatic side-chain of a phenylalanine residue CFMO Phe 62 (PSFMO Phe53) is found beneath the pyrimidinedione ring in the place where, in mFMO, an asparagine residue is found, which is thought to stabilise the oxygenating species in catalysis. Mutation of this Asn to a serine residue in mFMO led to removal of oxygenase activity <xref rid="bib0065" ref-type="bibr">[13]</xref>, but in SMFMO, mutation of the Phe residue to a valine resulted in a mutant of inverted enantioselectivity for the transformation of <italic>para</italic>-tolyl methyl sulfide <xref rid="bib0075" ref-type="bibr">[15]</xref>. The similarity in the FAD environment in all three homologs is reflected in the broadly similar substrate ranges and enantioselectivities of the cofactor-promiscuous enzymes with (<italic>R</italic>)-selectivity conserved for the most part towards substrates <bold>1</bold>, <bold>3</bold>, <bold>4</bold> and <bold>6</bold> with NADH. The poorer performance of NADPH-dependent biotransformations by SMFMO, CFMO and PSFMO appears at variance with their comparable abilities to reduce FAD, but may be due to more effective participation of the NADH cofactor in stabilisation of the active site during oxygenation, as has been observed for mFMO <xref rid="bib0060" ref-type="bibr">[12]</xref>.<fig id="fig0025"><label>Fig. 5</label><caption><p>FAD environment within the active sites of (a) CFMO and (b) PSFMO. The peptide backbones are shown in with transparency levels of 30%. Amino acid side chains within the active site are shown in cylinder format in coral for CFMO and green for PSFMO. The FAD molecules are shown in cylinder format with carbon atoms in grey, surrounded by the <italic>F_o</italic>&#x000a0;&#x02212;&#x000a0;<italic>F_c</italic> omit map in blue displayed at a level of 3<italic>&#x003c3;</italic> in each case, which was obtained through refinement in the absence of FAD. The maps illustrate the superior quality of the resolution (1.83&#x000a0;&#x0212b; vs. 2.39&#x000a0;&#x0212b;) in the PSFMO structure.</p></caption><graphic xlink:href="gr5"/></fig></p></sec><sec id="sec0075"><label>3.6</label><title>Cofactor binding loops in CFMO and PSFMO</title><p id="par0100">While the structures of SMFMO <xref rid="bib0050" ref-type="bibr">[10]</xref>, <xref rid="bib0075" ref-type="bibr">[15]</xref> have not yet been determined in complex with a nicotinamide cofactor, superimposition with the structure(s) of NADPH-complexes of mFMO (<italic>e.g.</italic> 2XLT) revealed that the phosphate binding sites for NADPH, including the 2&#x02032; hydroxyl ribose phosphate that distinguishes NADPH from NADH, were occupied by sulfate ions that had resulted from the crystallisation conditions in lithium sulfate (<xref rid="fig0030" ref-type="fig">Fig. 6</xref>a). This observation led to the hypothesis that substitution of arginine and threonine residues in mFMO for glutamine and histidine in SMFMO were one of the factors that determined cofactor promiscuity in the latter enzyme. Having acquired the structure of CFMO and PSFMO, these were now superimposed with the structure of mFMO in order to gain insight into the changes in the cofactor binding loops that may occur as a result of having threonine/serine and glutamine/glutamate in these positions. In the case of CFMO the side-chain oxygen atoms of Thr202 and Ser203 are both orientated towards the putative phosphate binding site (<xref rid="fig0030" ref-type="fig">Fig. 6</xref>b). The structure(s) of mFMO had previously demonstrated that threonine is able to make an effective contact with one of the phosphate oxygen atoms of NADPH <xref rid="bib0060" ref-type="bibr">[12]</xref>, <xref rid="bib0065" ref-type="bibr">[13]</xref>, <xref rid="bib0070" ref-type="bibr">[14]</xref>.<fig id="fig0030"><label>Fig. 6</label><caption><p>The putative NADP⁺ ribose 2&#x02032; phosphate recognition site in (a) SMFMO, (b) CFMO and (c) PSFMO, each superimposed with the corresponding site in mFMO, derived from PDB accession code 2XLT. The backbone and side chain and NADP⁺ carbon atoms for mFMO are shown in grey in each case. Glc&#x000a0;=&#x000a0;glycerol in (c). Equivalent features for SMFMO, CFMO and PSFMO are shown in yellow, coral and green in (a), (b) and (c) respectively. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)</p></caption><graphic xlink:href="gr6"/></fig></p><p id="par0105">The nucleotide binding site of PSFMO featured a glycerol molecule, from the cryoprotectant, occupying approximately the same position as the adenine ring of NADPH in mFMO, and with the hydroxyl groups forming hydrogen bonds with the side chain of Tyr140, the backbone NH of Gln194 and a water molecule. Superimposition of the putative phosphate binding sites of SMFMO and PSFMO revealed that the backbone C&#x003b1; atoms of residues thought to be responsible for cofactor promiscuity in SMFMO, Gln193 and His194, are in the same position as those in PSFMO, Gln194 and Glu195 (<xref rid="fig0030" ref-type="fig">Fig. 6</xref>c). However, the side chains of both Gln194 and Glu195 are observed to point away from the putative phosphate binding site into the bulk solvent.</p><p id="par0110">No data on the activity with mFMO and NADH have been reported, but other NADPH-specific FPMOs, such as NADPH-dependent Baeyer&#x02013;Villiger monooxygenases possess the Arg&#x02013;Thr couple in common with mFMO <xref rid="bib0080" ref-type="bibr">[16]</xref>, <xref rid="bib0085" ref-type="bibr">[17]</xref>. The studies with SMFMO and with CFMO and PSFMO reported herein suggest first that a Gln&#x02013;His couple favours NADH binding slightly, but that a Thr&#x02013;Ser couple, as found in CFMO is superior for activity with both cofactors overall. Counter-intuitively, the presence of a glutamate in the second position of the couple, as found in PSFMO, does not prohibit NADPH binding, but rather the Glu side chain is able to project away from the putative phosphate binding site. It is possible that the presence of the glutamate side-chain allows selective recognition of the NADH ribose oxygen atoms in the presence of that cofactor, by rotating into the binding pocket, but is able to rotate away in the presence of NADPH, thus allowing cofactor promiscuity in PSFMO. More detailed analysis awaits the acquisition of structures of these enzymes in the presence of both NADPH and NADH however.</p></sec></sec><sec id="sec0080"><label>4</label><title>Conclusion</title><p id="par0115">CFMO and PSFMO provide new examples of enzymes within the emerging sub-family of cofactor-promiscuous flavin-containing monooxygenases, named &#x02018;Type II FMOs&#x02019; by Fraaije and co-workers <xref rid="bib0035" ref-type="bibr">[7]</xref>, <xref rid="bib0055" ref-type="bibr">[11]</xref>. The study of this sub-class of FMOs is providing new information on structure and evolutionary diversity within this family of enzymes, and also suggesting new avenues for the engineering of related enzymes for cofactor promiscuity with a view to greater suitability for application in biocatalytic processes.</p></sec>

Divergent tissue and sex effects of rapamycin on the proteasome-chaperone network of old mice

<p>Rapamycin, an allosteric inhibitor of the mTOR kinase, increases longevity in mice in a sex-specific manner. In contrast to the widely accepted theory that a loss of proteasome activity is detrimental to both life- and healthspan, biochemical studies <italic>in vitro</italic> reveal that rapamycin inhibits 20S proteasome peptidase activity. We tested if this unexpected finding is also evident after chronic rapamycin treatment <italic>in vivo</italic> by measuring peptidase activities for both the 26S and 20S proteasome in liver, fat, and brain tissues of old, male and female mice fed encapsulated chow containing 2.24 mg/kg (14 ppm) rapamycin for 6 months. Further we assessed if rapamycin altered expression of the chaperone proteins known to interact with the proteasome-mediated degradation system (PMDS), heat shock factor 1 (HSF1), and the levels of key mTOR pathway proteins. Rapamycin had little effect on liver proteasome activity in either gender, but increased proteasome activity in female brain lysates and lowered its activity in female fat tissue. Rapamycin-induced changes in molecular chaperone levels were also more substantial in tissues from female animals. Furthermore, mTOR pathway proteins showed more significant changes in female tissues compared to those from males. These data show collectively that there are divergent tissue and sex effects of rapamycin on the proteasome-chaperone network and that these may be linked to the disparate effects of rapamycin on males and females. Further our findings suggest that rapamycin induces indirect regulation of the PMDS/heat-shock response through its modulation of the mTOR pathway rather than via direct interactions between rapamycin and the proteasome.</p>

<contrib contrib-type="author"><name><surname>Rodriguez</surname><given-names>Karl A.</given-names></name><xref ref-type="aff" rid="aff1">¹</xref><xref ref-type="aff" rid="aff2">²</xref><uri xlink:type="simple" xlink:href="http://community.frontiersin.org/people/u/172951"/></contrib><contrib contrib-type="author"><name><surname>Dodds</surname><given-names>Sherry G.</given-names></name><xref ref-type="aff" rid="aff1">¹</xref><xref ref-type="aff" rid="aff3">³</xref><uri xlink:type="simple" xlink:href="http://community.frontiersin.org/people/u/181805"/></contrib><contrib contrib-type="author"><name><surname>Strong</surname><given-names>Randy</given-names></name><xref ref-type="aff" rid="aff1">¹</xref><xref ref-type="aff" rid="aff4">⁴</xref><uri xlink:type="simple" xlink:href="http://community.frontiersin.org/people/u/53465"/></contrib><contrib contrib-type="author"><name><surname>Galvan</surname><given-names>Veronica</given-names></name><xref ref-type="aff" rid="aff1">¹</xref><xref ref-type="aff" rid="aff2">²</xref><uri xlink:type="simple" xlink:href="http://community.frontiersin.org/people/u/153876"/></contrib><contrib contrib-type="author"><name><surname>Sharp</surname><given-names>Z. D.</given-names></name><xref ref-type="aff" rid="aff1">¹</xref><xref ref-type="aff" rid="aff3">³</xref><uri xlink:type="simple" xlink:href="http://community.frontiersin.org/people/u/190197"/></contrib><contrib contrib-type="author"><name><surname>Buffenstein</surname><given-names>Rochelle</given-names></name><xref ref-type="aff" rid="aff1">¹</xref><xref ref-type="aff" rid="aff2">²</xref><xref ref-type="author-notes" rid="fn001">^*</xref><uri xlink:type="simple" xlink:href="http://community.frontiersin.org/people/u/74073"/></contrib>

Frontiers in Molecular Neuroscience

<sec sec-type="introduction" id="s1"><title>Introduction</title><p>Rapamycin, an allosteric inhibitor of mechanistic (mammalian) target of rapamycin (mTOR) reportedly increases longevity in mice even when given at an advanced age as an encapsulated formulation in chow. This effect is significantly greater in females (Miller et al., <xref rid="B33" ref-type="bibr">2014</xref>). The biological mechanism and relevance of this sex divergence in response to rapamycin is unclear. Surprisingly, rapamycin has also recently been shown to be an allosteric inhibitor of proteasome activity <italic>in vitro</italic> (Osmulski and Gaczynska, <xref rid="B35" ref-type="bibr">2013</xref>). Further, 20S proteasome activity is reduced in liver tissues of elderly (25 month old) mice treated with rapamycin for 6 months when compared to control animals (Zhang et al., <xref rid="B60" ref-type="bibr">2014a</xref>). However, gene expression studies undertaken in tissues harvested from the same animals showed equivocal effects on proteolytic pathways with differentially expressed genes linked to these pathways both up-regulated and also down-regulated (Fok et al., <xref rid="B17" ref-type="bibr">2014</xref>). Down-regulated proteolytic pathways are at variance with the widely accepted theory that the decline in functionality during aging is linked to an accrual of damaged or misfolded proteins (Ross and Poirier, <xref rid="B43" ref-type="bibr">2004</xref>; David et al., <xref rid="B10" ref-type="bibr">2010</xref>). This breakdown in proteostasis during aging is attributed at least in part to an age-associated decline in the efficiency of the proteolytic machinery (Rodriguez et al., <xref rid="B41" ref-type="bibr">2010</xref>; Grimm et al., <xref rid="B20" ref-type="bibr">2012</xref>). This leads to the concomitant accrual of aggregation-prone cytotoxic proteins that underlie many age-associated pathologies (Bucciantini et al., <xref rid="B4" ref-type="bibr">2002</xref>; Ross and Poirier, <xref rid="B43" ref-type="bibr">2004</xref>).</p><p>Age-related changes contributing to compromised proteasome and autophagy function in short-lived species include decreased transcription of some catalytic subunits, altered proteasome subcellular distribution, disruption of lysosomal control, and reduced degradative capacity of both proteolytic machineries (Ferrington et al., <xref rid="B16" ref-type="bibr">2005</xref>; Massey et al., <xref rid="B30" ref-type="bibr">2006</xref>; Rodriguez et al., <xref rid="B41" ref-type="bibr">2010</xref>). In contrast to these declines in proteolytic degradation processes, the proteasome-mediated protein degradation system (PMDS), that includes both ubiquitin-dependent and independent proteasome machinery and associated molecular chaperones, is more robust in naturally long-lived species (Rodriguez et al., <xref rid="B40" ref-type="bibr">2012</xref>; Edrey et al., <xref rid="B14" ref-type="bibr">2014</xref>), long-lived mutants (Kruegel et al., <xref rid="B25" ref-type="bibr">2011</xref>), calorically-restricted animals (Bonelli et al., <xref rid="B3" ref-type="bibr">2008</xref>), and centenarians (Chondrogianni et al., <xref rid="B6" ref-type="bibr">2000</xref>). As such it appears hard to reconcile a decline in proteasome function with rapamycin-induced extended longevity.</p><p>The 20S core when doubly capped by 19S regulatory particles is called the 26S proteasome and is primarily responsible for ubiquitinylated protein degradation and the bulk of PMDS related proteolytic activity (Demartino and Gillette, <xref rid="B12" ref-type="bibr">2007</xref>). While 20S proteasomes can exist un-capped <italic>in vivo</italic>, they are mostly inactive and contribute only 20&#x02013;30% of the total proteasome content in a cell (Babbitt et al., <xref rid="B1" ref-type="bibr">2005</xref>). The 19S regulatory caps through a combination of ATP-dependent (RPT) and ATP-independent (RPN) subunits mediate substrate uptake, by unfolding, deubiquitinylating, and moving proteins through to the 20S core (Demartino and Gillette, <xref rid="B12" ref-type="bibr">2007</xref>). Upstream trafficking of substrates to the proteasome is partially controlled by the heat-shock protein (HSP) family of molecular chaperones. HSP70/72 together with its co-chaperones may have a key regulatory role in PMDS and prolonged efficient function in long-lived species (Grune et al., <xref rid="B21" ref-type="bibr">2011</xref>; Rodriguez et al., <xref rid="B42" ref-type="bibr">2014</xref>). Further, the smaller molecular chaperone, HSP25, has been implicated in the mitigation of protein aggregation in vertebrates (Goldbaum et al., <xref rid="B18" ref-type="bibr">2009</xref>). While chaperone interactions in the PMDS may be critical for maintained protein homeostasis, recently the drug rapamycin has been shown to suppress the heat-shock response in cell culture (Chou et al., <xref rid="B7" ref-type="bibr">2012</xref>), yet nevertheless facilitates prolonged healthspan and extended longevity <italic>in vivo</italic> (Harrison et al., <xref rid="B22" ref-type="bibr">2009</xref>; Miller et al., <xref rid="B32" ref-type="bibr">2011</xref>, <xref rid="B33" ref-type="bibr">2014</xref>; Zhang et al., <xref rid="B60" ref-type="bibr">2014a</xref>).</p><p>We question whether this observation of reduced proteasome activity (Osmulski and Gaczynska, <xref rid="B35" ref-type="bibr">2013</xref>; Zhang et al., <xref rid="B60" ref-type="bibr">2014a</xref>) and suppression of the heat-shock response seen <italic>in vitro</italic> (Chou et al., <xref rid="B7" ref-type="bibr">2012</xref>), is observed <italic>in vivo</italic> especially in light of rapamycin-mediated transcriptional regulation of certain proteasome-related genes (Fok et al., <xref rid="B17" ref-type="bibr">2014</xref>). If this is the case, it may reveal that the mechanisms facilitating the increase in healthspan and longevity seen by improved protein homeostasis are independent of rapamycin-induced longevity and healthspan. We test the hypothesis that rapamycin counters the deleterious effects of aging in C57BL/6 mice through differentially suppressing the PMDS. We also ask if there are sex and/or tissue differences in in the various proteasome activities and molecular chaperone responses to rapamycin and whether changes in the mTOR pathway elucidate a mechanism for this rapamycin induced modulation of proteolytic function and lifespan.</p></sec><sec sec-type="methods" id="s2"><title>Methods</title><sec><title>Animals</title><p>Care of animals followed UT Health Science Center Institutional Animal Care and Use Committees approved procedures. Specific pathogen-free C57BL/6 mice were purchased from the National Institutes of Health colony reared in the Charles River Laboratories at 19 months of age. Mice were maintained under barrier conditions by the UTHSCSA Nathan Shock Center Aging Animal and Longevity Assessment Core and started on the rapamycin/eudragit control diet at ~19 months of age. Six months later at ~25 months of age, the <italic>ad libitum</italic> fed animals were anesthetized with isofluorane, euthanized by cardiac exsanguination and the tissues immediately excised and flash frozen in liquid nitrogen. All the tissues from these animals that were not fasted prior to sacrifice were stored at &#x02212;80&#x000b0;C until used in analyses.</p></sec><sec><title>Diet preparation</title><p>Mice were fed a diet containing either encapsulated rapamycin or empty capsules (eudragit control). Microencapsulated rapamycin or empty microcapsules were incorporated into Purina 5LG6 diet. The rapamycin diet was prepared at 14 ppm using methods described by Harrison et al. (<xref rid="B22" ref-type="bibr">2009</xref>) and blood levels checked to confirm appropriate dosing. Data pertaining to the rapamycin blood levels, lifespan, and healthspan effects of these animals have been previously published (Fok et al., <xref rid="B17" ref-type="bibr">2014</xref>; Zhang et al., <xref rid="B60" ref-type="bibr">2014a</xref>).</p></sec><sec><title>Lysates</title><p>Mouse tissue (~50&#x02013;100 mg fat, ~15&#x02013;30 mg liver, 100&#x02013;200 mg brain) was cryofractured under liquid nitrogen with a mortar and pestle. Powdered fat was lysed by homogenization in 2&#x000d7; dry weight in Protein Homogenization Buffer (50 mM HEPES, pH 7.6, 150 mM sodium chloride, 20 mM sodium pyrophosphate, 20 mM &#x000df;-glycerophosphate, 10 mM sodium fluoride, 2 mM sodium orthovanadate, 2 mM ethylenediaminetetraacetic acid, 1% IGEPAL, 10% glycerol, 1 mM magnesium chloride, 1 mM calcium chloride, 1 mM phenylmethylsulfonyl fluoride, and one tablet/10 ml Complete Mini (EDTA free) protease inhibitor tablets from Roche) for Western blots and in Re-suspension Buffer (RSB) (10 mM HEPES, pH 6.2, 10 mM NaCl, 1.4 mM MgCl₂) without protease inhibitors at a weight per volume ratio of 1 g of tissue to 2 mL of buffer for peptidolytic assays. The RSB buffer was supplemented with the addition of 1 mM ATP, 0.5 mM DTT, 5 mM MgCl₂ to help maintain intact 26S subassemblies (Liu et al., <xref rid="B28" ref-type="bibr">2006</xref>). Powdered liver was lysed by homogenization in 5&#x000d7; dry weight in modified RIPA Buffer (150 mM sodium chloride, 50 mM Tris-HCl, pH 7.4, 1 mM ethylenediaminetetraacetic acid, 1 mM phenylmethylsulfonyl flouride, 1% Triton X-100, 1% sodium deoxycholic acid, 0.1% sodium dodecyl sulfate, 1 &#x003bc;M Bortezomib proteasome inhibitor, and one tablet/10 ml protease and phosphatase inhibitor mini tablets (Thermo Fisher Scientific, Waltham, MA, USA) for Western blots. For peptidolytic assays, liver powder was lysed in RSB without protease inhibitors and with ATP supplemented as described above at a weight per volume ration of 1 g of tissue to 5 mL of buffer. Brain tissue was separated and lysed in RSB containing the protease/phosphatase cocktail tablet described above (Thermo Fisher Scientific) for Western blot analyses or RSB without protease inhibitors containing the ATP supplement for peptidolytic assays. Debris was cleared by centrifugation (3000 g for 12 min). The Bio-Rad Protein Assay (Bio-Rad Life Sciences, Hercules, CA) or BCA Assay (Thermo Fisher Scientific) was used to determine protein concentrations.</p></sec><sec><title>Peptidolytic activity</title><p>In each assay 20 &#x003bc;g of whole tissue lysates were incubated with 100 &#x003bc;M of substrate specific for the type of proteasome activity. A saturating concentration of proteasome inhibitor N-(benzyloxycarbonyl) leucinyl-leucinyl-leucinal (MG132) was added to parallel samples. The difference of the fluorescence released with and without inhibitor was used as a measure of the net peptidolytic activity of proteasome as previously described using model peptide substrates to represent cleavage after hydrophobic (Chymotrypsin-like; ChTL) residues, basic residues (Trypsin-like; TL) and acidic residues (Post-glutamyl, peptide-hydrolizing; PGPH) (Rodriguez et al., <xref rid="B41" ref-type="bibr">2010</xref>).</p></sec><sec><title>Western blots</title><p>Tissue lysates were separated using a 4&#x02013;12% SDS-PAGE (Biorad Life Sciences) and transferred to nitrocellulose membranes (Biorad Life Sciences). The membranes were probed with antibodies against the following proteasome and chaperone proteins: HSP90 (mouse mAb, SPA831, 1:20K), HSF1 (rabbit pAb, SPA901, 1:5K), HSP70/72 (mouse mAb, SPA810, 1:10K), HSP40 (HDJ1) (rabbit pAb, SPA400, 1:2.5K), HSP25/HSPB1 (rabbit pAb, SPA801, 1:10K), &#x003b1;7 (mouse mAb, PW8110, 1:5K,), RPT5 (mouse mAb, PW8310, 1:5K) (Enzo Life Sciences, Plymouth Meeting, PA, USA) and Carboxyl-terminus of HSP70 Interacting Protein (CHIP) [rabbit mAb(C3B6), #2080, 1:5K] (Cell Signaling Technology, Inc., Danvers, MA, USA). Antibodies against GAPDH (mouse mAb, G8795, 1:30K) (Sigma-Aldrich, St. Louis, MO, USA) were used as a loading control. For mTOR pathway proteins, the following antibodies were used all at 1:1K dilutions: mTOR (rabbit pAb, #2972), phospho-mTOR (Ser2448) (rabbit pAb, #2971), AKT (rabbit pAb, #9272), phospho-AKT (Ser473) (rabbit, pAb, #9271), S6 ribosomal protein (mouse mAb (54D2), #2317), phospho-S6 ribosomal protein (Ser240/244) (rabbit pAb, #2215), 4EBP1 (rabbit pAb, #9452) and phospho-4EBP1 (Thr37/46) (236B4) (rabbit mAb, #2855) (Cell Signaling Technology). Either the GAPDH antibody used above or pan-Actin (mouse mAb, MS-1295-P0, 1:20K) (Thermo-Fisher Scientific) was used as a loading control.</p><p>Primary antibodies were detected using anti-mouse IRDye 680LT, or anti-rabbit IR Dye800 CW (Li-Cor, Lincoln, NB, USA) conjugated antibodies. Secondary antibodies were incubated at 1:10K (anti-rabbit) or 1:20K (anti-mouse) for 2 h at room temperature and images were captured and subsequently quantified using the Odyssey Imaging System (Li-Cor) by quantifying fluorescent signals as Integrated Intensities (I.I. K Counts) using the Odyssey Infrared Imaging System, Application Version 3.0 software. We used a local background subtraction method to subtract independent background values from each box, more specifically, the &#x0201c;median&#x0201d; background function with a 3 pixel width border above and below each box was subtracted from individual counts. We calculated ratios for each antibody against the pan-actin or GAPDH loading control using I.I. K Counts. The respective antibody to pan-actin or GAPDH ratio was then used to calculate phosphorylated protein to total protein ratio were applicable.</p></sec><sec><title>Native gel electrophoresis</title><p>In this current study, we use an ATP-reconstituting system to maintain the integrity of the 26S proteasome (Liu et al., <xref rid="B28" ref-type="bibr">2006</xref>), however it cannot be ruled out that the fluorogenic proteasome assay measures total proteasome activity for all the proteasome subassemblies, nor does it directly report ubiquitin-dependent activity of the whole 26S. With that caveat, however, we supplement this reliable indirect measure with in-gel assays on native gels.</p><p>Proteasome ChTL function was also measured using Native Gel Electrophoresis. This technique enabled us to discriminate if 26S or 20S proteasome activity predominated, the relative quantities of both the double-capped and uncapped proteasome and proteasome specific activities. Fifty micrograms of fractionated lysate from each of the sample groups prepared as described in Subcellular Fractionation above (q.v) were run on a 3&#x02013;12% non-denaturing, gradient polyacrylamide gel (Life Technologies, Carlsbad, CA). The gel was run at 30 V for 30 min in a 4&#x000b0;C cold cabinet, thereafter the voltage was increased to 35 V for 1 h, 50 V for 1 h and further increased to 75 V for three more hours (Elasser et al., <xref rid="B15" ref-type="bibr">2005</xref>; Tai et al., <xref rid="B49" ref-type="bibr">2010</xref>).</p><p>Peptidolytic activity of proteasomes was detected after incubating the gels in a Suc-LLVY-MCA substrate dissolved in 50 mM Tris pH 8.0, 5 mM MgCl₂, 1 mM DTT, 2 mM ATP, and 0.02% SDS for 15, 30, and 60 min at 37&#x000b0;C. Proteasome bands were identified by the release of highly fluorescent, free AMC (Elasser et al., <xref rid="B15" ref-type="bibr">2005</xref>; Vernace et al., <xref rid="B52" ref-type="bibr">2007</xref>; Rodriguez et al., <xref rid="B40" ref-type="bibr">2012</xref>). Fluorescence was quantitated using ImageJ software (<ext-link ext-link-type="uri" xlink:href="http://imagej.nih.gov/ij/">http://imagej.nih.gov/ij/</ext-link>). Following the in-gel assay, the protein from the gel was transferred to nitrocellulose using the i-blot transfer apparatus (Life Technologies) and subjected to Western blotting analyses using the &#x003b1;7 antibody described above to determine where the 26S and 20S proteasome complexes lay. The IRDye conjugating antibodies and the Odyssey Imaging System (Li-Cor) were used to quantitate the amounts of &#x003b1;7 signal (See Western Blots above).</p></sec><sec><title>Statistical analyses</title><p>Prism 5 (GraphPad Software, Inc.) was used to analyze and graph the mTOR pathway Western blot data. An unpaired two-tailed <italic>t</italic>-test or Mann Whitney test was used to obtain <italic>p</italic>-values. <italic>P</italic>-values below 0.05 were considered significant. Proteasome and chaperone data was analyzed with Sigma Plot (v.11) using Two-Way ANOVA. Statistical significance was set at the <italic>p</italic> &#x0003c; 0.05 level with Tukey and Holm-Sidak corrections to counteract the probability of false positives. Cluster Analysis was performed using Cluster 3.0 and Treeview v 1.16r2 to generate output (open source code <ext-link ext-link-type="uri" xlink:href="http://bonsai.hgc.jp/~mdehoon/software/cluster/">http://bonsai.hgc.jp/~mdehoon/software/cluster/</ext-link>; University of Tokyo, 2002). Data were log transformed to normalize the activity and protein expression data, and the cluster was generated using the group medians and hierarchical clustering with an uncentered Pearson Correlation to generate a complete linkage to create the most unbiased set of clusters (Yeung and Ruzzo, <xref rid="B56" ref-type="bibr">2001</xref>; De Hoon, <xref rid="B11" ref-type="bibr">2002</xref>).</p></sec></sec><sec sec-type="results" id="s3"><title>Results</title><sec><title>Rapamycin-treated females showed increases in both chymotrypsin-like and trypsin-like 26S proteasome activity in brain, but not in liver or visceral fat</title><p>Peptidolytic activity of the proteasome was measured in old (25 month) male and female, rapamycin-treated and untreated, brain, liver, and visceral fat lysates using model peptide substrates specific for each of the three types of catalytic sites. The core particle contains three cleavage sites that degrade polypeptides or unfolded proteins severing peptide bonds on the carboxyl side of hydrophobic (&#x0201c;chymotrypsin-like&#x0201d;; ChTL), basic (&#x0201c;trypsin-like&#x0201d;; TL), or acidic (&#x0201c;peptidylglutamyl peptide hydrolyzing&#x0201d; or &#x0201c;caspase-like&#x0201d;; PGPH) residues (Demartino and Gillette, <xref rid="B12" ref-type="bibr">2007</xref>). To determine net proteasome activity, assays were run in parallel with and without the proteasome inhibitor MG132. Divergent responses were evident with regards to treatment, sex, and peptidolytic activities (Figure <xref ref-type="fig" rid="F1">1</xref>; summarized in Table <xref ref-type="table" rid="T1">1</xref>). In untreated animals, males showed significantly higher ChTL activity in brain (<italic>p</italic> = 0.01), liver (<italic>p</italic> = 0.002), and fat (<italic>p</italic> = 0.03) (Figure <xref ref-type="fig" rid="F1">1</xref>, top panels) than observed in females. TL activity, while still higher for males in liver (<italic>p</italic> = 0.009), was the same in brain and higher in females visceral fat (<italic>p</italic> = 0.02). PGPH activity was higher in lysates from female animals compared to males in both brain (<italic>p</italic> = 0.03) and fat (<italic>p</italic> = 0.01), but in liver, PGPH activity was higher in untreated males compared to females (<italic>p</italic> = 0.003) (Figure <xref ref-type="fig" rid="F1">1</xref>, bottom panels).</p><fig id="F1" position="float"><label>Figure 1</label><caption><p><bold>Changes in chymotrypsin-like (ChTL) and trypsin-like (TL) 26S proteasome activity in rapamycin-treated lysates from old animals depended upon the tissue type and sex of the animals tested</bold>. Significant increase was observed for both ChTL (3X) and TL (2X) 26S proteasome activities in whole brain lysates from female animals only. No significant effects were evident in male brains. Liver TL activity declined (0.8X) with rapamycin treatment in males. Rapamycin treatment led to a decline in visceral fat ChTL (0.7X) and TL (0.7X) activity in females but not in males. PGPH activity did not show any significant differences between treated or untreated animals in any of the tissues examined. The y-axis indicates net proteasome activity in pmol/min/&#x003bc;g lysate for each of the respective activities measured. Statistically significant differences (Two-Way ANOVA, <italic>p</italic> &#x0003c; 0.05) comparing treatment vs. vehicle (&#x02217;) and/or male vs. female untreated (#) or male vs. female treated ($) are indicated (<italic>n</italic> = 5 brain, fat, <italic>n</italic> = 6 liver).</p></caption><graphic xlink:href="fnmol-07-00083-g0001"/></fig><table-wrap id="T1" position="float"><label>Table 1</label><caption><p><bold>Rapamycin-influenced changes in markers of proteasome, chaperone, and mTOR pathway in old male and female brain, liver, and visceral fat lysates</bold>.</p></caption><graphic xlink:href="fnmol-07-00083-i0001"/><table-wrap-foot><p>&#x0201c;UP&#x0201d; indicates a rapamycin-dependent significant increase. &#x0201c;DN&#x0201d; represents a rapamycin-dependent significant decrease.</p></table-wrap-foot></table-wrap><p>Rapamycin effects were also tissue-specific but in general proteasome activity in females was more responsive to rapamycin treatment. In whole brain tissue lysates males showed no change in proteasome activity, whereas in females, the change in proteasome activity was significant, being nearly 3-fold higher for ChTL (<italic>p</italic> = 0.0009) and 2-fold higher for TL (<italic>p</italic> = 0.0005) substrates in treated compared to untreated samples (Figure <xref ref-type="fig" rid="F1">1</xref>, top left). In the liver lysates only male TL activity changed significantly, with a 20% decline in activity in the rapamycin treated (<italic>p</italic> = 0.02) (Figure <xref ref-type="fig" rid="F1">1</xref>, middle). Finally, proteasome activity from visceral fat lysates showed significant declines in ChTL (<italic>p</italic> = 0.05) and TL (<italic>p</italic> = 0.01) activities of 30% in samples from treated females compared to controls. Activity in male fat lysates did not change with treatment (Figure <xref ref-type="fig" rid="F1">1</xref>, right panels). Rapamycin treatment had no effect on PGPH activity in either sex or in any of the tissues tested (Figure <xref ref-type="fig" rid="F1">1</xref>, left panels). Interestingly, ChTL activity showed the greatest sexual dimorphism with sex-differences evident in the activities of samples from both rapamycin-treated and untreated controls. Activity was higher in male control samples, whereas females showed the higher activity in rapamcyin treated samples (Figure <xref ref-type="fig" rid="F1">1</xref>, top left).</p><p>Native gel electrophoresis has been used to determine if the proteasome remains intact in a higher molecular weight form (i.e., 26S) or exists disassembled (20S) (Elasser et al., <xref rid="B15" ref-type="bibr">2005</xref>; Rodriguez et al., <xref rid="B40" ref-type="bibr">2012</xref>). We used this technique to determine if either the 20S or 26S proteasome assembly, was affected by rapamycin treatment or showed sex-specific differences (Figures <xref ref-type="fig" rid="F2">2</xref>&#x02013;<xref ref-type="fig" rid="F4">4</xref>). In tissue lysates from female mouse brains, the in-gel assay measuring ChTL activity (Figures <xref ref-type="fig" rid="F2">2A,B</xref>) showed a significant increase in 26S proteasome activity in a rapamycin-dependent manner (<italic>p</italic> = 0.02), but not in 20S activity. In contrast ChTL activity in male mouse brains did not change (<italic>p</italic> &#x0003e; 0.05). The immunblot analyses of &#x003b1;7 (Figures <xref ref-type="fig" rid="F2">2C,D</xref>) revealed that the 26S proteasome content of this protein significantly increased in brain lysates of rapamycin treated female mice (<italic>p</italic> = 0.003). Interestingly, the proteasome &#x003b1;7 decreased significantly in rapamycin-treated male mice brain samples compared to those of the male control group (Figures <xref ref-type="fig" rid="F2">2C,D</xref>; <italic>p</italic> = 0.02), though this did not reflect on activity. 20S proteasome content did not change for either sex.</p><fig id="F2" position="float"><label>Figure 2</label><caption><p><bold>Female Brain Lysates showed a rapamycin-dependent increase in 26S proteasome. (A)</bold> Representative zymogram of Chymotrypsin-like (ChTL) proteasome activity after native gel electrophoresis of 50 &#x003bc;g of brain lysate from control and rapamycin-treated old male and female animals. <bold>(B)</bold> Quantitation of zymogram showed that there was a treatement-induced increase in 26S-specific ChTL proteasome activity in female brain lysates (<italic>n</italic> = 4, ^*<italic>p</italic> &#x0003c; 0.05 Two-Way ANOVA). <bold>(C)</bold> Representative immunoblot of &#x003b1;7 proteasome subunit after transfer from native gel. <bold>(D)</bold> Similarly, &#x003b1;7 subunit showed a 26S-specific change (increase in females, decrease in males; ^*<italic>p</italic> &#x0003c; 0.05, ^**<italic>p</italic> &#x0003c; 0.01; <italic>n</italic> = 4, Two-Way ANOVA) in assembled proteasome content. It did not seem that change in male proteasome content was sufficient to influence a change in activity.</p></caption><graphic xlink:href="fnmol-07-00083-g0002"/></fig><p>Zymograms for 26S proteasome activity showed a significant decline in both liver (Figures <xref ref-type="fig" rid="F3">3A,B</xref>) and fat (Figures <xref ref-type="fig" rid="F4">4A,B</xref>) tissue lysates of rapamycin-treated females (<italic>p</italic> = 0.008 brain, <italic>p</italic> = 0.009 fat) with no change in proteasome activity for either sex at the 20S site. A significant decline in &#x003b1;7 protein content occurred in 26S proteasomes measured in liver samples from rapamycin-treated females (<italic>p</italic> = 0.004). Further, 20S &#x003b1;7 protein content declined in both male (<italic>p</italic> = 0.02) and female liver samples (<italic>p</italic> = 0.001) (Figures <xref ref-type="fig" rid="F3">3C,D</xref>). Proteasome content in fat tissue lysates (&#x003b1;7) also decreased significantly for both the 26S (<italic>p</italic> = 0.003) and 20S (<italic>p</italic> = 0.04) sites of treated female sample whereas &#x003b1;7 content did not change in the fat samples from males.</p><fig id="F3" position="float"><label>Figure 3</label><caption><p><bold>Liver Lysates revealed potential rapamycin-dependent declines in proteasome activity. (A)</bold> Representative zymogram of Chymotrypsin-like (ChTL) proteasome activity after native gel electrophoresis of 50 &#x003bc;g of liver lysate from control and rapamycin-treated old male and female animals. <bold>(B)</bold> Quantitation of zymogram showed that there was a treatement-induced decrease in 26S-specific ChTL proteasome activity in female brain lysates (<italic>n</italic> = 4, ^*<italic>p</italic> &#x0003c; 0.05 Two-Way ANOVA). <bold>(C)</bold> Representative immunoblot of &#x003b1;7 proteasome subunit after transfer from native gel. <bold>(D)</bold> The calculated amounts of the &#x003b1;7 subunit showed significant decreases at both the 26S (females) and 20S sites (males and females) (^*<italic>p</italic> &#x0003c; 0.05, ^**<italic>p</italic> &#x0003c; 0.01; <italic>n</italic> = 4, Two-Way ANOVA).</p></caption><graphic xlink:href="fnmol-07-00083-g0003"/></fig><fig id="F4" position="float"><label>Figure 4</label><caption><p><bold>Female Fat Lysates also showed rapamycin-influenced declines in proteasome activity. (A)</bold> Representative zymogram of Chymotrypsin-like (ChTL) proteasome activity after native gel electrophoresis of 50 &#x003bc;g of visceral fat lysate from control and rapamycin-treated old male and female animals. <bold>(B)</bold> Quantitation of the zymogram revealed that there was a treatment-dependent decrease in female 26S ChTL activity (<italic>n</italic> = 4, ^*<italic>p</italic> &#x0003c; 0.05 Two-Way ANOVA). <bold>(C)</bold> Representative immunoblot of &#x003b1;7 proteasome subunit after transfer from native gel. <bold>(D)</bold> The &#x003b1;7 subunit quantitation indicated significant decreases in both female 26S and 20S content (^*<italic>p</italic> &#x0003c; 0.05, ^**<italic>p</italic> &#x0003c; 0.01; <italic>n</italic> = 4, Two-Way ANOVA).</p></caption><graphic xlink:href="fnmol-07-00083-g0004"/></fig></sec><sec><title>Levels of proteasome-related chaperones changed more in treated females than in treated males</title><p>Western blot analyses using a panel of antibodies representing major chaperone families and key proteasome subunits was undertaken in the tissues harvested from rapamycin-fed and Eudragit-(vehicle) fed controls. Tissues from three male and three female 25 month old mice were analyzed. In lysates from male samples, rapamycin had no effect on chaperone levels (Figure <xref ref-type="fig" rid="F5">5A</xref>). In female brain lysates, the levels of HSP70, CHIP, and HSP25 proteins increased in treated samples compared to controls (Figure <xref ref-type="fig" rid="F5">5A</xref>). Further, the 19S cap protein RPT5 also showed a higher protein content in these brain lysates from rapamycin-treated females (Figure <xref ref-type="fig" rid="F5">5A</xref>). Proteins were normalized to GAPDH whose mRNA is eIF4E insensitive and would not change with rapamycin treatment (Livingstone et al., <xref rid="B29" ref-type="bibr">2010</xref>), These data are indicative of an increased 26S population and RPT5 levels correlated with an increase in ChTL and TL proteasome activity in brain lysates (Figure <xref ref-type="fig" rid="F1">1</xref>). In liver, with the exception of HSP25 (which did not change), there was a decline in heat shock proteins in both sexes with treatment (Figure <xref ref-type="fig" rid="F5">5B</xref>). While &#x003b1;7 levels also decreased in males and females, this did not correlate with any loss in proteasome activity (Figures <xref ref-type="fig" rid="F1">1</xref>, <xref ref-type="fig" rid="F5">5B</xref>). In visceral fat, HSF1 as well as HSP90, HSP70, HSP40, and HSP25 decreased significantly in samples from treated female animals (Figure <xref ref-type="fig" rid="F5">5C</xref>). Both RPT5 and &#x003b1;7 also showed protein declines in fat samples from rapamycin-treated females that correlated with a decline in ChTL and TL proteasome activity. Interestingly, samples from treated males showed declines in CHIP, HSP, and HSP25 (Figure <xref ref-type="fig" rid="F5">5C</xref>). Lastly, HSP90, HSP70, and RPT5 showed significant differences between males and females sample in their immuno-detection (Figure <xref ref-type="fig" rid="F5">5C</xref>).</p><fig id="F5" position="float"><label>Figure 5</label><caption><p><bold>Both proteasome subunits and proteasome-related chaperones in brain, liver, and visceral fat lysates revealed tissue and sex dependent treatment-related changes</bold>. Changes in protein expression as measured by Western blots relative to GAPDH, and normalized to old control in brain lysates from old (25 mo) male and female rapamycin-fed mice <bold>(A)</bold> indicated a female-specific, treatment-related increase for HSP72, HSP40, CHIP, HSP25, and RPT5. All these proteins also showed sex-dependent differences as well. In liver lysates <bold>(B)</bold> both sexes showed a treatment-related decline in all proteins tested except for HSP25 and RPT5. There was only a significant sex difference in &#x003b1;7. In visceral fat lysates <bold>(C)</bold> females again showed a treatment-related change most strikingly for the proteasome markers &#x003b1;7 and RPT5. Visceral fat showed significant sex-dependent differences in all proteins tested except for HSP25. Significant differences between male and female pairs (<italic>n</italic> = 3 brain, <italic>n</italic> = 5 fat, <italic>n</italic> = 6 liver; Two-Way ANOVA, <italic>p</italic> &#x0003c; 0.05) are indicated by (&#x02217;) treatment-related change or (#) sex-dependent change.</p></caption><graphic xlink:href="fnmol-07-00083-g0005"/></fig></sec><sec><title>mTOR pathway markers responded to rapamycin treatment in females but these changes were also tissue dependent</title><p>To examine for a correlation between the changes in proteasome activity and chaperone profile with the mTOR pathway, we immunoassayed a subset of mTOR pathway proteins and their phosphorylated forms (p) in brain, liver, and visceral fat lysates from old male and female, rapamycin-fed animals and compared to similarly aged controls relative to Actin (also insensitive to rapamycin) or GAPDH (Livingstone et al., <xref rid="B29" ref-type="bibr">2010</xref>) (Figure <xref ref-type="fig" rid="F6">6B</xref>). Female rapamycin-treated brains had higher levels of p-mTOR (Ser2448), and both total and phosphorylated AKT (Ser473) and rpS6 (Ser240/244). We also observed an increase in total 4EBP1 and unchanged p-4EBP1 (Thr36/47) levels which led to significant decline in the ratio of p-4EBP1/4EBP1 (Figure <xref ref-type="fig" rid="F6">6C</xref>). In male lysates there was no significant difference in the phosphorylated or total protein expression of these four mTOR pathway proteins, with rapamycin-treatment (Figures <xref ref-type="fig" rid="F6">6B,C</xref>).</p><fig id="F6" position="float"><label>Figure 6</label><caption><p><bold>Treatment with rapamycin altered expression of key proteins of the mTOR pathway in female brain lysates. (A)</bold> Basic schematic of the relationship of the proteins tested (highlighted in red) to mTOR (R1 indicates mTORC1-dependent branch while R2 indicates the mTORC2-dependent branch). <bold>(B)</bold> Changes in protein expression of these markers of the mTOR pathway measured in brain lysates from 25 month-old male and female mice (rapamycin-fed vs. control) and quantified Integrated Intensities (I.I. K Counts) relative to GAPDH (I.I. K Count Ratio) <bold>(C)</bold> showed that rapamycin influenced change in mTOR itself and both proteins upstream of mTOR (AKT) as well as its downstream targets (rpS6 and 4EBP1). Both total and phosphorylated (p) forms are presented as well as the ratio of the phosphorylated vs. the total protein. Significant <italic>p</italic>-values as derived by Two-Way ANOVA (<italic>n</italic> = 3) using Prism GraphPad Software is indicated for each set of proteins.</p></caption><graphic xlink:href="fnmol-07-00083-g0006"/></fig><p>Rapamycin had little effect on protein expression of mTOR related proteins in liver tissue (Figures <xref ref-type="fig" rid="F7">7</xref>, <xref ref-type="fig" rid="F8">8</xref>). Only p-4EBP1 (Thr36/47) in liver lysates of females was significantly higher with rapamycin (Figures <xref ref-type="fig" rid="F7">7A,B</xref>) whereas in liver lysates from males, rapamycin induced increased expression of both p-4EBP1 (Thr36/47) and total 4EBP1 (Figures <xref ref-type="fig" rid="F8">8A,B</xref>) as well as total AKT. The latter change led to a lower ratio of p-AKT (Ser473) to total AKT (Figure <xref ref-type="fig" rid="F8">8B</xref>).</p><fig id="F7" position="float"><label>Figure 7</label><caption><p><bold>Liver lysates from female C57BL6 mice did not show a clear pattern of change in mTOR signaling. (A)</bold> Liver lysates from 25 month female rapamycin-treated and control animals were analyzed by Western blot for several mTOR-pathway related proteins and quantified using I.I. K Counts relative to pan-Actin or GAPDH <bold>(B)</bold>. <italic>P</italic>-values (unpaired <italic>t</italic>-test, Prism GraphPad) are indicated above each dot plot (<italic>n</italic> = 6).</p></caption><graphic xlink:href="fnmol-07-00083-g0007"/></fig><fig id="F8" position="float"><label>Figure 8</label><caption><p><bold>Liver lysates from male C57BL6 mice also did not show a clear pattern of change in mTOR signaling. (A)</bold> Western blots from lysates from 25 month-old males and quantitation using I.I. K Counts relative to pan-Actin or GAPDH <bold>(B)</bold>. <italic>P</italic>-values (unpaired <italic>t</italic>-test, Prism GraphPad) are indicated above each dot plot (<italic>n</italic> = 6).</p></caption><graphic xlink:href="fnmol-07-00083-g0008"/></fig><p>Visceral fat lysates from rapamycin-treated females showed a trend toward a decrease in p-AKT (Ser473), total mTOR and downstream effectors p-rp-S6 (Ser240/244) and total 4EBP1 (Figures <xref ref-type="fig" rid="F9">9A,B</xref>). These rapamycin mediated declines in protein expression led to altered ratios of the phosphorylated form to the total proteins (Figure <xref ref-type="fig" rid="F7">7B</xref>). The effects of rapamycin on the expression of these proteins were attenuated in males. Only phosphorylated AKT (Ser473) and the ratios of p-mTOR (Ser2448) to total mTOR significantly increased (Figures <xref ref-type="fig" rid="F10">10A,B</xref>).</p><fig id="F9" position="float"><label>Figure 9</label><caption><p><bold>The phosphorylation state of downstream mTOR pathway markers changed significantly in visceral fat lysates from rapamycin-treated females</bold>. Measuring protein content of several mTOR pathway markers analyzed from visceral fat lysates taken from 25 month rapamycin-fed and control female mice via immunoblot <bold>(A)</bold> quantified using I.I. K Counts relative to pan-Actin <bold>(B)</bold> revealed a decrease in p-rp-S6 which also lowered the ratio (phosphorylation state) of p-rpS6 to total rp-S6. Total 4EBP1 also declined, increasing the phosphorylation state. <italic>P</italic>-values, derived using Prism GraphPad Software, are indicated for each dot plot (<italic>n</italic> = 5).</p></caption><graphic xlink:href="fnmol-07-00083-g0009"/></fig><fig id="F10" position="float"><label>Figure 10</label><caption><p><bold>mTOR pathway proteins changed minimally in visceral fat lysates from rapamycin-treated males</bold>. Lysates created from the visceral fat of 25 month-old rapamycin-fed and control male mice were analyzed via Western blot <bold>(A)</bold> and quantified using I.I. K Counts <bold>(B)</bold> relative to pan-Actin as a loading control. Only phosphorylated AKT showed a significant change. <italic>P</italic>-values (unpaired <italic>t</italic>-test, Prism GraphPad) are shown above each dot plot (<italic>n</italic> = 5).</p></caption><graphic xlink:href="fnmol-07-00083-g0010"/></fig></sec><sec><title>Cluster analyses revealed that brain proteasome activity is most influenced by rapamycin</title><p>Table <xref ref-type="table" rid="T1">1</xref> summarizes the observed rapamycin-mediated changes in proteasome activity, chaperones, and mTOR pathway proteins tested in this study. To determine whether the combined proteasome activities, chaperone, and mTOR pathway data assembled into common patterns of rapamycin-associated changes, cluster analyses were performed (Figure <xref ref-type="fig" rid="F11">11</xref>). First, proteasome activity and proteasome-related chaperones were clustered (Figure <xref ref-type="fig" rid="F11">11A</xref>). Cluster analysis revealed that groups separated by tissue type, weighted toward the higher levels of proteasome activity in liver and brain. These groups were further divided in the analysis by sex, but treatment was indistinguishable (Figure <xref ref-type="fig" rid="F11">11A</xref>). In this comparison, two distinct clusters formed. The first major cluster consisted of proteasome activity, the 19S ATPase RPT5 and HSF1in one sub-cluster with &#x003b1;7 and HSP25 in another. The second major cluster showed that the large chaperones (HSP90 and HSP70) and the HSP70 co-chaperones HSP40 and CHIP sub-divided into sub-clusters (Figure <xref ref-type="fig" rid="F11">11A</xref>).</p><fig id="F11" position="float"><label>Figure 11</label><caption><p><bold>Cluster analysis shows a proteasome-influenced grouping by tissue, sex, and then treatment</bold>. A heat map showing results of the cluster analysis of variables including log-transformed peptidolytic activities and quantities of proteins determined by immuno-blot first with proteasome activity, proteasome subunits, and several proteasome-related chaperones together <bold>(A)</bold>, and then with the addition selected mTOR pathway proteins <bold>(B)</bold>. Groups representing the various lysates collected from animals treated with rapamycin and corresponding controls are organized in columns represent cases. Four letter codes define the groups with the first two in the code representing gender (OM, Old Male; OF, Old Female). The third letter represents the tissue (L, Liver; F, Fat; B, Brain), and the forth letter denotes treatment group (C, Control (Eudragit); R, Rapamycin-treated; i.e., OMFC, Old Male Fat Control). The three proteasome activity and protein tests are organized in rows. Results obtained with Cluster 3.0 analysis are shown. Heat maps were prepared with TreeView v.1.16r2. Color scheme corresponds to the normalized values of variables where bright green represent the lowest (approaching &#x02212;3.0) and bright red the highest (close to +3.0) values. The lengths of tree branches are proportional to a relative similarity between variables and between cases.</p></caption><graphic xlink:href="fnmol-07-00083-g0011"/></fig><p>The mTOR pathway proteins were added into the next cluster analysis (Figure <xref ref-type="fig" rid="F11">11B</xref>). This revealed that rapamycin-treated female brain and treated male liver formed two distinct groups. Three clusters formed in this analysis, with the mTOR pathway proteins 4EBP1 and p-AKT joining the &#x0201c;proteasome activity cluster&#x0201d; with the same proteins associated in the first Chaperone-Proteasome analysis from panel A (RPT5, &#x003b1;7, HSP25, and HSF1). A second cluster containing HSP90 and the co-chaperones HSP40 and CHIP also contained p-4EBP1, AKT, and rpS6 (Figure <xref ref-type="fig" rid="F11">11B</xref>). The last cluster in this analysis was characterized by low p-mTOR and low p-rpS6 (brain), and associated with HSP70 and total mTOR (Figure <xref ref-type="fig" rid="F11">11B</xref>).</p></sec></sec><sec sec-type="discussion" id="s4"><title>Discussion</title><p>In this study we evaluated if rapamycin-mediated changes in mTOR signaling and the ubiquitin proteasome system (i.e., proteasome activity and content and associated protein levels various chaperones) correlated with the sexually dimorphic effects of rapamycin on lifespan. Rapamycin-induced changes in proteasome activity and expression of the various molecular chaperones and mTOR pathway proteins were both sex- and tissue-specific (Figure <xref ref-type="fig" rid="F11">11</xref>). Rapamycin effects are readily apparent from both individual variables and cluster analyses (Figure <xref ref-type="fig" rid="F11">11</xref>). The latter were generated by applying multiple peptidolytic assays and associated chaperone proteins as well as a sampling of mTOR pathway proteins across the three tissues harvested from the same individuals. Cluster analyses reveal there are interactions between the proteasome-chaperone network and the mTOR pathway. Rapamycin treatment led to elevated proteasome activities in the brain of females, potentially promoting the removal of oxidatively damaged and misfolded proteins and creating improved proteostasis in female mice that are likely to contribute to the better protection of their brains against environmentally mediated (e.g., oxidatively damaged/ glucose crosslinking) protein aggregation. Better protection of the female brain may be an evolutionary life-history tradeoff and these traits may possibly vary with reproductive status (non-reproductive pregnant, lactation). Improved proteostasis and concomitant neuroprotection in the brains of rapamycin treated females supports previous studies using genetically engineered mouse models of Alzheimer's disease in which rapamycin firstly improved cognition and when subjected to a chronic high sugar diet prevented the accrual of protein aggregates, plaques and neurofibrillary tangles (Orr et al., <xref rid="B34" ref-type="bibr">2014</xref>).</p><sec><title>Rapamycin effects are tissue-specific</title><p>Contrary to a previous data that examined only the 20S catalytic core ChTL activity (Zhang et al., <xref rid="B60" ref-type="bibr">2014a</xref>), our extensive study revealed that rapamycin treatment did not globally suppress the PMDS or its associated chaperones. Indeed in our current study, declines in proteasome were observed predominantly in visceral fat (Figures <xref ref-type="fig" rid="F1">1</xref>, <xref ref-type="fig" rid="F4">4</xref>) in contrast the previous study using mice in the same cohort showed decreases in both heart and liver 20S ChTL proteasome activity in both sexes of rapamycin-treated mice (Zhang et al., <xref rid="B60" ref-type="bibr">2014a</xref>). The decline in proteasome activity in peripheral tissues both in this study and that of Zhang, is in keeping with <italic>in vitro</italic> findings that rapamycin can allosterically inhibit proteasome activity (Osmulski and Gaczynska, <xref rid="B35" ref-type="bibr">2013</xref>) However, this cannot explain the tissue specific increases in ChTL and TL proteasome activity in the brains of rapamycin treated animals.</p><p>While ChTL is considered the pivotal peptidolytic activity of oxidatively damaged proteins, other peptidase activities such as TL and PGPH activity may be more sensitive to rapamycin. Female mice treated with rapamycin showed the most divergent responses compared to untreated females for both the ChTL and TL catalytic sites in fat (decrease) and brain (increase) (Figure <xref ref-type="fig" rid="F1">1</xref>). PGPH proteasome activity did not change suggesting that rapamycin-mediated effects target specific types of proteins and peptides for cleavage driving changes in proteasome function rather than a restructuring of the proteasome although there is some evidence that mTOR signaling can induce proteasome subunits through both complex 1 or complex 2 (Lamming et al., <xref rid="B26" ref-type="bibr">2014</xref>; Zhang et al., <xref rid="B61" ref-type="bibr">2014b</xref>).</p><p>Given that rapamycin is likely to first reach the liver via the hepatic portal vein and be at its most concentrated there, it is surprising that of the three tissues examined, the liver was least sensitive to rapamycin treatment as shown by the peptide assays (Figure <xref ref-type="fig" rid="F1">1</xref>). It is possible that rapamycin is metabolized in such a way in the liver that it has minimal effects on liver proteasome function. Alternately, as the liver has several pathways involving nutrient signaling, a slight perturbation in the mTOR pathway is likely to be compensated for by other signaling pathways. Unfortunately, published RNA-sequence analyses data revealing differentially expressed genes in the transcriptome with rapamycin were equivocal (Fok et al., <xref rid="B17" ref-type="bibr">2014</xref>) and did not reveal an obvious explanation for a lack of a change in 26S proteasome activity or a decline only in 26S TL activity as was seen in males with treatment (Figure <xref ref-type="fig" rid="F1">1</xref>). However, the in-gel assay on native gels did show a decline ChTL activity in lysates from rapamycin-treated samples from females (Figure <xref ref-type="fig" rid="F3">3</xref>) and a decline in &#x003b1;7 proteasome content suggesting alterations in liver proteasome expression and quantity by rapamycin(Zhang et al., <xref rid="B60" ref-type="bibr">2014a</xref>). Given the low blood titer of C57BL6 compared to the UM-HET3 mice (~3&#x02013;4 ng/L compared to 13.4 ng/L, respectively, after 6 months of treatment) (Harrison et al., <xref rid="B22" ref-type="bibr">2009</xref>; Zhang et al., <xref rid="B60" ref-type="bibr">2014a</xref>), we may have only observed effects in tissues that are particularly sensitive to rapamycin. Also, there definitely seem to be strain-dependent differences in at least rapamycin uptake or metabolism which could dictate the effect of the drug on measured parameters.</p></sec><sec><title>Rapamycin decreases proteasome activity in visceral fat</title><p>We chose to include fat tissue in our analyses because we hypothesized it may be more sensitive to alterations in mTOR signaling, being a nutrient storage tissue, and was previously shown to be responsive male and female mice (Harrison et al., <xref rid="B22" ref-type="bibr">2009</xref>). In C57BL6 mice, proteasome activity declines with age in adipose tissue (Dasuri et al., <xref rid="B9" ref-type="bibr">2011</xref>). Our study shows that in old animals, there is a further reduction in proteolytic activity in visceral fat harvested from female mice after rapamycin-treatment (Figure <xref ref-type="fig" rid="F1">1</xref>) which was supported by a similar observation on native gel zymograms (Figure <xref ref-type="fig" rid="F4">4</xref>). This was accompanied by a treatment-related reduction in the key proteasome subunits &#x003b1;7 and RPT5 (Figures <xref ref-type="fig" rid="F4">4</xref>, <xref ref-type="fig" rid="F5">5</xref>). Thus, rapamycin appears to affect upstream control of critical genes in the PMDS in visceral fat though at this time it is unknown whether this effect is transcriptional or translational. Optimal function of the 26S is dependent on a tightly regulated ratio of ATPases such as RPT5 in the 19S regulatory cap (Smith et al., <xref rid="B45" ref-type="bibr">2011</xref>). Alternatively, or in combination with an upstream regulation of RPT5 and/ or other proteasome ATPases, rapamycyin has been shown to directly block proteasome activity <italic>in vitro</italic> by interfering with the attachment of the 19S cap (Osmulski and Gaczynska, <xref rid="B35" ref-type="bibr">2013</xref>). A similar mechanism may be in place in fat tissue whereby the inhibition of mTOR signaling in adipose tissue may block the binding of the regulatory cap to the catalytic core and impair proteasome function. Reduction of the proteasome activity has been shown to induce cytotoxicity, upregulate cell stress responses, and lead to various pathologies in other tissues (Goldbaum et al., <xref rid="B19" ref-type="bibr">2006</xref>; Grimm et al., <xref rid="B20" ref-type="bibr">2012</xref>). Taken together these data suggest that rapamycin-induced declines in visceral fat PMDS function could be detrimental and contribute to some of the observed peripheral tissue pathologies associated with age (Dasuri et al., <xref rid="B9" ref-type="bibr">2011</xref>). For example, the development of obesity and insulin signaling in type 2 diabetes is influenced by the PMDS as insulin receptor substrate 1 is inactivated by degradation through this system (Sun et al., <xref rid="B48" ref-type="bibr">1999</xref>; Chang et al., <xref rid="B5" ref-type="bibr">2009</xref>). Thus, effects that mimic aging in PMDS function could substantially contribute to age-related insulin resistance in adipose tissue and other deleterious features associated with both fat and liver metabolism (Umemura et al., <xref rid="B51" ref-type="bibr">2014</xref>). While we did not measure changes in insulin signaling in this cohort of animals, others have shown consistently that rapamycin affects the insulin pathway in much the same manner in multiple strains of mice including C57BL6, namely creating glucose intolerance, but causing insulin sensitivity (Lamming et al., <xref rid="B27" ref-type="bibr">2013</xref>; Orr et al., <xref rid="B34" ref-type="bibr">2014</xref>; Yu et al., <xref rid="B58" ref-type="bibr">2014</xref>). PMDS could be a potential pathway to explain this phenomenon.</p></sec><sec><title>Rapamycin increases proteasome activity in the brain</title><p>The most pronounced effects to proteasome activity were evident in the brain suggesting that it is not the rapamycin directly inducing these changes but rather the down-stream signaling it induces in rapamycin-responsive tissues. In contrast to the potential detrimental effects of rapamycin seen in visceral fat proteolytic function, the increase of PMDS in the brain could be beneficial and could suggest that organisms selectively protect certain more vulnerable tissues. Here, we observe that rapamycin induced an increase in proteasome activity, a phenotype commonly observed in long-lived species in the periphery as well as in the brain (Chondrogianni et al., <xref rid="B6" ref-type="bibr">2000</xref>; Rodriguez et al., <xref rid="B40" ref-type="bibr">2012</xref>; Edrey et al., <xref rid="B14" ref-type="bibr">2014</xref>). Brain lysates from treated females showed both an increase in ChTL and TL activity (Figures <xref ref-type="fig" rid="F1">1</xref>, <xref ref-type="fig" rid="F2">2</xref>) as well as significant increases in 26S proteasome content on non-denaturing gels (Figure <xref ref-type="fig" rid="F2">2</xref>). Proteasome-related chaperones HSP70, C-terminal HSP-interacting protein (CHIP), RPT5, and HSP25 also showed significant increases in rapamycin-treated samples compared to control (Figure <xref ref-type="fig" rid="F5">5</xref>). Further, &#x003b1;7, RPT5,HSF1, and HSP25correlated strongly with higher levels of proteasome activity (Figure <xref ref-type="fig" rid="F11">11</xref>) in the cluster analysis. Other studies examining brains from rapamycin-treated mice have seen a similar enhancement of protein homeostasis and/or the HSP system (Spilman et al., <xref rid="B46" ref-type="bibr">2010</xref>; Pierce et al., <xref rid="B36" ref-type="bibr">2013</xref>; Orr et al., <xref rid="B34" ref-type="bibr">2014</xref>). For example, rapamycin-fed mice showed the enhanced expression of the small chaperone gene, alpha-crystallin B chain (<italic>CRYAB</italic>) in the brains of a mouse Alzheimer's model (Pierce et al., <xref rid="B36" ref-type="bibr">2013</xref>). Further, crossing this model with an HSF1-transgenic mouse (Pierce et al., <xref rid="B37" ref-type="bibr">2010</xref>), which showed increased HSF1, HSP90, and CryAB, reduced toxic brain amyloid-&#x003b2; levels and improved cognitive function (Pierce et al., <xref rid="B36" ref-type="bibr">2013</xref>). Interestingly, we did not see an increase in HSF1 in brain tissue with rapamycin treatment though cluster analyses suggested a correlation (Figures <xref ref-type="fig" rid="F5">5</xref>, <xref ref-type="fig" rid="F11">11</xref>). However, several chaperones showed increases in rapamycin-treated female brain samples (i.e., HSP70, CHIP, HSP25) suggesting both HSF1 and non-canonical upregulation of HSPs. One possibility is that if rapamycin indeed inhibits proteasome activity <italic>in vivo</italic> as suggested by <italic>in vitro</italic> studies (Osmulski and Gaczynska, <xref rid="B35" ref-type="bibr">2013</xref>) an alternate heat shock factor such as HSF2 may be triggered inducing the same set of chaperones as HSF1 (Mathew et al., <xref rid="B31" ref-type="bibr">1998</xref>).</p><p>The differences seen between the effects of rapamycin on the brain and peripheral tissues also suggests a tissue-specific decoupling of rapamycin function. In a study in which mice genetically mutated to serve as transgenic mouse models for Alzheimer's disease a tissue-specific decoupling of rapamycin effects were observed when these mice were fed a high sucrose diet and developed insulin resistance. While the exacerbation of plaques were ameliorated by the simultaneous treatment with rapamycin, rapamycin had no effect on peripheral insulin resistance and liver protein levels (Orr et al., <xref rid="B34" ref-type="bibr">2014</xref>).</p><p>In this study, we did not examine markers of autophagy. Counter-intuitively, a previous investigation reports that, autophagy in brain tissue lysates was not increased in rapamycin-treated non-transgenic mice but only was manifest in animals genetically engineered to express high levels of amyloid-&#x003b2; or tau (Spilman et al., <xref rid="B46" ref-type="bibr">2010</xref>; Orr et al., <xref rid="B34" ref-type="bibr">2014</xref>). So while autophagy may have a key role in reducing aggregates from disease pathologies, non-aggregated protein degradation may instead be enhanced by the PMDS. The increase in chaperone-E3 ligase CHIP in the brain and a significant decline in liver tissue (Figure <xref ref-type="fig" rid="F5">5</xref>) may hold a clue. CHIP has been shown to be essential in modulating oxidative load and degradation of oxidized proteins through the PMDS as well as act as the E3 ligase for the degradation of tau, parkin, and polyglutamine expansions in brain tissue (Dickey et al., <xref rid="B13" ref-type="bibr">2007</xref>; Sisoula and Gonos, <xref rid="B44" ref-type="bibr">2011</xref>). As such, in the brain chronic rapamycin treatment could trigger an E3 ligase like CHIP to protect against proteotoxic stress, increasing the translation of chaperone proteins and the proteasome-mediated, protein degradation machinery to maintain protein equilibrium.</p></sec><sec><title>Changes in mTOR signaling correlate with sex and tissue differences</title><p>In both female brain and fat tissues the mTOR pathway proteins, as expected, showed the most changes with rapamycin-treatment (Figures <xref ref-type="fig" rid="F3">3</xref>, <xref ref-type="fig" rid="F7">7</xref>; Table <xref ref-type="table" rid="T1">1</xref>). However, while the changes in female fat samples suggest rapamycin blocks mTOR signaling, rapamycin-mediated changes in the brain did not (Figures <xref ref-type="fig" rid="F6">6</xref>, <xref ref-type="fig" rid="F9">9</xref>; Table <xref ref-type="table" rid="T1">1</xref>) (reviewed in Wullschleger et al., <xref rid="B55" ref-type="bibr">2006</xref>). Rather, these data suggest that mTOR remains active in the brain, and that the inhibitory effects of rapamycin are suppressed in brain tissue. This uncoupling of brain and peripheral effects on mTOR was unexpected but could also explain the lack of increased levels of autophagy in the brain of rapamycin-treated control animals seen in previous studies (Spilman et al., <xref rid="B46" ref-type="bibr">2010</xref>; Orr et al., <xref rid="B34" ref-type="bibr">2014</xref>). Further, the increase in p-rpS6 in the female brain (Figure <xref ref-type="fig" rid="F6">6</xref>) was very different to what was observed in female fat (Figure <xref ref-type="fig" rid="F9">9</xref>) or in the intestine of rapamycin-treated familial adenomatous polyposis mice both which showed a decrease in the phosphorylation state (ratio of phosphorylated to non-phosphorylated protein) of rp-S6 and other mTOR signaling molecules (Hasty et al., <xref rid="B23" ref-type="bibr">2014</xref>). This paradox may be linked to the increase in both total and phosphorylated AKT (Figure <xref ref-type="fig" rid="F6">6</xref>, Table <xref ref-type="table" rid="T1">1</xref>). A similar induced activation of AKT leading to a resistance of rapamycin treatment has been observed in human lung cancer cells and human rhabdomyosarcoma cell lines and in rodent cells overexpressing insulin-like growth factor (IGF) II (Sun et al., <xref rid="B47" ref-type="bibr">2005</xref>; Wan et al., <xref rid="B53" ref-type="bibr">2007</xref>). Unlike these cell systems, whereby the phosphorylation of both downstream mTOR targets S6K1 and 4EBP1 were suppressed, this was not observed in brain lysates from rapamycin-treated old females in this study (Figure <xref ref-type="fig" rid="F6">6</xref>). Instead we observed an increase in phosphorylated ribosomal protein S6, the target of S6K1 (Figure <xref ref-type="fig" rid="F6">6</xref>), responsible for 5&#x02032; terminal oligopyrimidine (TOP)-dependent translation (Wullschleger et al., <xref rid="B55" ref-type="bibr">2006</xref>). Further, there was a decline in the phosphorylation state of 4EBP1 (Figure <xref ref-type="fig" rid="F6">6</xref>) which is responsible for control of cap-dependent translation (Wullschleger et al., <xref rid="B55" ref-type="bibr">2006</xref>) and inhibition of TOP-dependent translation (Thoreen et al., <xref rid="B50" ref-type="bibr">2012</xref>). This may be also be reflected by the cluster analysis correlation of low 4EBP1 with proteasome activity (Figure <xref ref-type="fig" rid="F11">11</xref>). These changes in translational control suggest a focus on TOP-dependent translation, and have further repercussions on the PMDS, as several proteasome-related HSP mRNAs can be preferentially translated through TOP-dependent translation during stress (Cuesta et al., <xref rid="B8" ref-type="bibr">2000</xref>; Pierce et al., <xref rid="B36" ref-type="bibr">2013</xref>). Conversely, in fat we observed a decline in phosphorylated rp-S6 and an increase in the phosphorylation state of 4EBP1 suggesting a decrease in translation (Figure <xref ref-type="fig" rid="F9">9</xref>) (Thoreen et al., <xref rid="B50" ref-type="bibr">2012</xref>) that could in turn influence the observed reduction of proteasome activity (Figure <xref ref-type="fig" rid="F1">1</xref>).</p><p>Phosphorylation of AKT also protects against the apoptotic effects of proteasome inhibition further linking PMDS to the AKT cell survival program (Yu et al., <xref rid="B57" ref-type="bibr">2006</xref>; Zanotto-Filho et al., <xref rid="B59" ref-type="bibr">2012</xref>). As total AKT is also induced by 17-&#x003b2; estradiol (Haynes et al., <xref rid="B24" ref-type="bibr">2000</xref>), this may explain why rapamycin-treated female animals have more robust effects when treated by the drug. Taken together, the sex-specific increase seen in brains of rapamycin treated mice in our comprehensive measures of proteasome activity could be linked to a sex-dependent change in AKT signaling driven by mTOR regulation of the heat-shock pathway through activated rp-s6 (as we observed in treated female brains; Figure <xref ref-type="fig" rid="F3">3</xref>). Interestingly, mTOR signaling through rp-S6 has also been shown to increase proteasome subunits with a dependence on nuclear factor erythroid-derived 2-related factor 1 (NRF1) (Zhang et al., <xref rid="B61" ref-type="bibr">2014b</xref>). NRF1 also mediates the recovery of proteasome activity after stress or proteotoxic insults (Radhakrishnan et al., <xref rid="B39" ref-type="bibr">2010</xref>; Balasubramanian et al., <xref rid="B2" ref-type="bibr">2012</xref>).</p></sec></sec><sec sec-type="conclusions" id="s5"><title>Conclusions</title><p>Our data indicate that the sexually dimorphic effects of lifespan extension induced by rapamycin (Harrison et al., <xref rid="B22" ref-type="bibr">2009</xref>; Fok et al., <xref rid="B17" ref-type="bibr">2014</xref>; Miller et al., <xref rid="B33" ref-type="bibr">2014</xref>; Zhang et al., <xref rid="B60" ref-type="bibr">2014a</xref>) may be linked to sex differences in tissue-specific responsiveness of the regulators and components of mTOR, HSPs and PMDS pathways. Proteolytic activity is augmented in the brain facilitating more efficient removal of damaged or unfolded proteins, and as a consequence thereof, improved brain structural and functional integrity. This protective response to rapamycin treatment in brain tissue is uncoupled from that of the response in peripheral fat tissue. The latter, appears to be left more vulnerable to the potentially detrimental peripheral proteotoxic effects of rapamycin (Wilkinson et al., <xref rid="B54" ref-type="bibr">2012</xref>; Ponticelli, <xref rid="B38" ref-type="bibr">2014</xref>; Zhang et al., <xref rid="B60" ref-type="bibr">2014a</xref>). A proposed summary of what happens to the PMDS in female brain vs. the periphery (specifically visceral fat) with rapamycin treatment is shown in Figure <xref ref-type="fig" rid="F12">12</xref>. Here we show that rapamycin appears to influence up- stream regulators of the proteasome-chaperone network through stimulation of the AKT pathway (Figure <xref ref-type="fig" rid="F12">12</xref>). Defining how the AKT pathway is stimulated or repressed in various tissues in a sex-dependent manner can give us insight on how to better understand rapamycin's effects in a therapeutic context. Upstream regulators of the PMDS including E3 ligases such as CHIP such as HSF1, 2 or NRF1 could be activated or suppressed by the AKT stress response in a sex-dependent manner thereby differentially affecting proteostasis in peripheral and brain tissues. Maintenance of protein homeostasis in the female brain may play a pivotal role in the extended longevity observed only in the rapamycin-treated female C57BL6 mice.</p><fig id="F12" position="float"><label>Figure 12</label><caption><p><bold>Response of the proteasome-chaperone network differs between the brain and periphery. (A)</bold> Rapamycin treatment may stimulate AKT in the brain, increasing proteasome-related chaperones through rp-S6 top-dependent translation and a non-cannonical stress response pathway (red-dashed arrows) ultimately increasing proteasome activity in the female brain. <bold>(B)</bold> In contrast, visceral fat shows a suppression of the mTOR network, reduction of the HSF1 chaperone response, and a decrease in proteasome activity.</p></caption><graphic xlink:href="fnmol-07-00083-g0012"/></fig></sec><sec><title>Author contributions</title><p>Karl A. Rodriguez and Sherry G. Dodds conducted the experiments. Karl A. Rodriguez and Rochelle Buffenstein wrote the initial draft of the paper. Randy Strong, Veronica Galvan, Z. D. Sharp, and Rochelle Buffenstein contributed materials for the study. All authors helped with editing and revising the manuscript.</p><sec><title>Conflict of interest statement</title><p>None of the contributing authors received payment or services from a third party for any aspect of this submission. Z. D. Sharp, Randy Strong, and Veronica Galvan are uncompensated scientific advisors for Rapamycin Holdings, Inc. Z. D. Sharp, Randy Strong, and Veronica Galvan are part holders of a patent (#13/128/800 pending) for the use of encapsulated rapamycin in treating or preventing an age-related disease, condition, or disorder. There are no other relationships that could have influenced or give the appearance of potentially influencing this work. The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.</p></sec></sec>

Altered dynamics between neural systems sub-serving decisions for unhealthy food

<p>Using BOLD functional magnetic resonance imaging (fMRI) techniques, we examined the relationships between activities in the neural systems elicited by the decision stage of the Iowa Gambling Task (IGT), and food choices of either vegetables or snacks high in fat and sugar. Twenty-three healthy normal weight adolescents and young adults, ranging in age from 14 to 21, were studied. Neural systems implicated in decision-making and inhibitory control were engaged by having participants perform the IGT during fMRI scanning. The Youth/Adolescent Questionnaire, a food frequency questionnaire, was used to obtain daily food choices. Higher consumption of vegetables correlated with higher activity in prefrontal cortical regions, namely the left superior frontal gyrus (SFG), and lower activity in sub-cortical regions, namely the right insular cortex. In contrast, higher consumption of fatty and sugary snacks correlated with lower activity in the prefrontal regions, combined with higher activity in the sub-cortical, insular cortex. These results provide preliminary support for our hypotheses that unhealthy food choices in real life are reflected by neuronal changes in key neural systems involved in habits, decision-making and self-control processes. These findings have implications for the creation of decision-making based intervention strategies that promote healthier eating.</p>

<contrib contrib-type="author"><name><surname>He</surname><given-names>Qinghua</given-names></name><xref ref-type="aff" rid="aff1">¹</xref><xref ref-type="aff" rid="aff2">²</xref><uri xlink:type="simple" xlink:href="http://community.frontiersin.org/people/u/184645"/></contrib><contrib contrib-type="author"><name><surname>Xiao</surname><given-names>Lin</given-names></name><xref ref-type="aff" rid="aff2">²</xref><xref ref-type="author-notes" rid="fn001">^*</xref><uri xlink:type="simple" xlink:href="http://community.frontiersin.org/people/u/87164"/></contrib><contrib contrib-type="author"><name><surname>Xue</surname><given-names>Gui</given-names></name><xref ref-type="aff" rid="aff3">³</xref><uri xlink:type="simple" xlink:href="http://community.frontiersin.org/people/u/3405"/></contrib><contrib contrib-type="author"><name><surname>Wong</surname><given-names>Savio</given-names></name><xref ref-type="aff" rid="aff4">⁴</xref><uri xlink:type="simple" xlink:href="http://community.frontiersin.org/people/u/189031"/></contrib><contrib contrib-type="author"><name><surname>Ames</surname><given-names>Susan L.</given-names></name><xref ref-type="aff" rid="aff5">⁵</xref><uri xlink:type="simple" xlink:href="http://community.frontiersin.org/people/u/15455"/></contrib><contrib contrib-type="author"><name><surname>Xie</surname><given-names>Bin</given-names></name><xref ref-type="aff" rid="aff5">⁵</xref><uri xlink:type="simple" xlink:href="http://community.frontiersin.org/people/u/190268"/></contrib><contrib contrib-type="author"><name><surname>Bechara</surname><given-names>Antoine</given-names></name><xref ref-type="aff" rid="aff2">²</xref><uri xlink:type="simple" xlink:href="http://community.frontiersin.org/people/u/8018"/></contrib>

Frontiers in Neuroscience

<sec sec-type="introduction" id="s1"><title>Introduction</title><p>With an increase of abundant and easily accessible high-calorie foods, an important characteristic of human choices in food is the unhealthy consumption of high calorie foods. Such choices can have long-term negative consequences, such as medical problems associated with overweight and obesity. The question is: why do some individuals become insensitive to the future consequences of their unhealthy eating habits and have difficulty making better healthful choices? While some research has found that poorer decision-making capacity may be associated with abnormal eating behaviors, most of these studies have focused on patients with differing forms of eating pathology (Pignatti et al., <xref rid="B68" ref-type="bibr">2006</xref>; Brogan et al., <xref rid="B15" ref-type="bibr">2010</xref>; Danner et al., <xref rid="B22" ref-type="bibr">2012</xref>; Fagundo et al., <xref rid="B34" ref-type="bibr">2012</xref>). In the current study, we evaluate normal individuals who are not medically diagnosed with an eating disorder. We examine the activity of neural systems hypothesized to subserve decision-making, using the Iowa Gabling Task (IGT), as well as the relationship between this neural activity and real life eating behavior.</p><p>Recent work has hypothesized that at least three neural systems influence behaviors involving complex decision-making, especially choices that include conflicts between immediate and long-term consequences (Naqvi and Bechara, <xref rid="B60" ref-type="bibr">2009</xref>; Noel et al., <xref rid="B64" ref-type="bibr">2013</xref>; He et al., <xref rid="B45" ref-type="bibr">2014a</xref>,<xref rid="B46" ref-type="bibr">b</xref>). One neural system is thought to mediate habitual behaviors that are elicited spontaneously or automatically. This neural system has been referred to as the &#x0201c;Impulsive System,&#x0201d; and key neural regions in this (impulsive) system include the amygdala and ventral striatum (and its mesolimbic dopamine link), which has been found to play a key role in the incentive motivational effects of a variety of non-natural rewards (e.g., psychoactive drugs) and natural rewards (e.g., food) (Stewart et al., <xref rid="B78" ref-type="bibr">1984</xref>; Robbins et al., <xref rid="B69" ref-type="bibr">1989</xref>; Wise and Rompre, <xref rid="B92" ref-type="bibr">1989</xref>; Robinson and Berridge, <xref rid="B70" ref-type="bibr">1993</xref>; Di Chiara et al., <xref rid="B31" ref-type="bibr">1999</xref>; Everitt et al., <xref rid="B33" ref-type="bibr">1999</xref>; Balleine and Dickinson, <xref rid="B1" ref-type="bibr">2000</xref>; Koob and Le Moal, <xref rid="B54" ref-type="bibr">2001</xref>; Dagher, <xref rid="B20" ref-type="bibr">2009</xref>; Dagher and Robbins, <xref rid="B21" ref-type="bibr">2009</xref>). Another neural system relates to executive and inhibitory control, referred to as the &#x0201c;Reflective System,&#x0201d; and a critical neural region in the reflective system is the ventromedial prefrontal cortex (VMPFC) region, as well as the medial orbitofrontal cortex (Bechara et al., <xref rid="B9" ref-type="bibr">2000</xref>). However, other neural components, including the dorsolateral prefrontal cortex implicated in working memory capacity and the cingulate cortex are also parts of this neural circuitry, and are essential for the normal operation of the VMPFC (Bechara, <xref rid="B3" ref-type="bibr">2004</xref>; Boorman et al., <xref rid="B11" ref-type="bibr">2013</xref>).</p><p>More recent evidence suggests that there is a third neural system mediated through the insular cortex. This pathway plays a key role in translating interoceptive signals into what one subjectively experiences as a feeling of desire, anticipation, or urge (Naqvi et al., <xref rid="B62" ref-type="bibr">2007</xref>; Naqvi and Bechara, <xref rid="B60" ref-type="bibr">2009</xref>). There is evidence demonstrating that the insular cortex is implicated in drug craving (Garavan, <xref rid="B37" ref-type="bibr">2010</xref>). For example, strokes that damage this region eliminate the urge to smoke in individuals previously addicted to cigarette smoking (Naqvi et al., <xref rid="B62" ref-type="bibr">2007</xref>). Additionally, an increasing number of studies suggest that the insula shows exaggerated responsiveness to drug cues in individuals addicted to drugs, and is hyper-reactive to visual food cues in obese individuals (Killgore et al., <xref rid="B52" ref-type="bibr">2003</xref>; DelParigi et al., <xref rid="B29" ref-type="bibr">2006</xref>; Geliebter et al., <xref rid="B41" ref-type="bibr">2006</xref>; Grill et al., <xref rid="B43" ref-type="bibr">2007</xref>; Rothemund et al., <xref rid="B74" ref-type="bibr">2007</xref>; Stoeckel et al., <xref rid="B80" ref-type="bibr">2008</xref>; Brooks et al., <xref rid="B16" ref-type="bibr">2013</xref>; Garc&#x000ed;a-Garc&#x000ed;a et al., <xref rid="B38" ref-type="bibr">2013</xref>; Tomasi and Volkow, <xref rid="B83" ref-type="bibr">2013</xref>). Finally, a behavioral measure of urgency, defined as an individual's tendency to give in to strong impulses, specifically when accompanied by negative emotions such as depression, anxiety, or anger (Whiteside and Lynam, <xref rid="B90" ref-type="bibr">2001</xref>), has also been shown to positively correlate with insula activity in recent fMRI studies (Joseph et al., <xref rid="B49" ref-type="bibr">2009</xref>; Xue et al., <xref rid="B99" ref-type="bibr">2010</xref>).</p><p>Emerging evidence suggests that overweight and obesity represents a special case of addictive behavior, which involves underlying neural mechanisms similar to other addictions (Kelley and Berridge, <xref rid="B51" ref-type="bibr">2002</xref>; Rolls, <xref rid="B73" ref-type="bibr">2007</xref>; Trinko et al., <xref rid="B84" ref-type="bibr">2007</xref>; Volkow et al., <xref rid="B86" ref-type="bibr">2008</xref>; Johnson and Kenny, <xref rid="B48" ref-type="bibr">2010</xref>). Specifically, a hyper-functioning impulsive system, a hypo-functioning reflective system, and/or an altered insula system were suggested by previous empirical studies as potential candidate mechanisms for the over-eating behavior (He et al., <xref rid="B45" ref-type="bibr">2014a</xref>,<xref rid="B46" ref-type="bibr">b</xref>), thus consistent with proposed theories on behavioral addiction to substances in general (Bechara and Damasio, <xref rid="B5" ref-type="bibr">2005</xref>; Naqvi and Bechara, <xref rid="B60" ref-type="bibr">2009</xref>; Noel et al., <xref rid="B64" ref-type="bibr">2013</xref>). Based on these findings, we hypothesized that a loss of self-control or inability to resist tempting/rewarding foods, and the development of less healthful eating habits (e.g., greater intake of high-calorie foods), may be explained by some alternation in any of these three neural systems.</p><p>The aim of this study was to utilize a laboratory-based task that taps into the functions of the different neural systems involved in affective decision-making, and to use functional imaging to evaluate the activities of these neural systems in relation to food choices in real-life. The most frequently used paradigm to assess affective decision-making is the Iowa Gambling Task (IGT) (Bechara et al., <xref rid="B6" ref-type="bibr">1994</xref>; Bechara and Damasio, <xref rid="B7" ref-type="bibr">2002</xref>; Waters-Wood et al., <xref rid="B88" ref-type="bibr">2012</xref>), which was initially developed to investigate decision-making defects of patients with focal brain lesions. The IGT has been shown to tap into aspects of decision-making that are influenced by affect and emotion (Bechara and Damasio, <xref rid="B5" ref-type="bibr">2005</xref>). Many studies have demonstrated that in comparison to normal controls, a wide range of patients (e.g., substance users, schizophrenia, pathological gamblers, and adolescents with externalizing behavior) show poor behavioral decisions as measured by the IGT (Bechara and Damasio, <xref rid="B7" ref-type="bibr">2002</xref>; Cavedini et al., <xref rid="B18" ref-type="bibr">2002</xref>; Whitney et al., <xref rid="B91" ref-type="bibr">2004</xref>; Sevy et al., <xref rid="B75" ref-type="bibr">2007</xref>; Xiao et al., <xref rid="B97" ref-type="bibr">2009</xref>). The same set of brain regions (i.e., ventral striatum, prefrontal cortex, and insula) linked to decision-making impairments in brain lesion studies have also been shown to be engaged during functional neuroimaging studies in healthy individuals during performance of the IGT (Li et al., <xref rid="B56" ref-type="bibr">2010</xref>; Xiao et al., <xref rid="B96" ref-type="bibr">2013</xref>).</p><p>The present study used Functional Magnetic Resonance Imaging (fMRI) techniques to investigate the relationship between the brain activity underlying decision-making (as elicited by the IGT) and real-life food choices in a group of normal young adults. Specifically, we tested the hypothesis that decision-making during the IGT will activate a neural circuitry that includes the mesial orbitofrontal and VMPFC region, the dorsolateral prefrontal cortex, and the anterior cingulate/SMA (supplementary motor area), which are components of the so-called &#x0201c;reflective system.&#x0201d; The degree of activity in these neural regions was hypothesized to inversely correlate with the degree of self-reported consumption of snacks high in fat and sugar, i.e., higher snack consumption would be associated with lower neural activity. Further, the degree of activity in these neural regions was hypothesized to positively correlate with the degree of self-reported consumption of vegetables, i.e., higher consumption would be associated with higher neural activity. We also tested the hypothesis that decision-making during the IGT would activate a subcortical neural circuitry that includes neural components of the so-called impulsive and urge system, namely the amygdala, the ventral striatum, and the insular cortex. The degree of activity in these neural regions was hypothesized to positively correlate with the degree of self-reported consumption of snacks high in fat and sugar but negatively correlate with the degree of self-reported consumption of vegetables.</p></sec><sec sec-type="methods" id="s2"><title>Methods</title><sec><title>Participants</title><p>Twenty-three (12 female) healthy adolescents and young adults aged 18.01 &#x000b1; 2.61 years were recruited from the University of Southern California (USC) and recreation centers in Los Angeles, California. None of the participants were currently diagnosed with an eating disorder or receiving clinical treatment for obesity. All participants had normal or corrected-to-normal vision. Based on the Structured Clinical Interview for DSM-IV (SCID), all participants were free of neurological or psychiatric history. Adolescents who were under 18 were accompanied to the university by a parent or designated family member. Written informed consents were obtained from the participants and their parent/legal guardians (for participants under 18) prior to participation. Research protocols and instruments were approved by the USC Institutional Review Boards.</p></sec><sec><title>Procedures</title><p>Participants came to the lab for two sessions. During the first session, participants and their parent (for participants under 18) completed and signed the consent form(s) and completed behavioral tasks. During the second session, participants were returned for the fMRI scan session. We asked participants to have their usual meal before they arrived for the fMRI session and eat normally. Therefore, the last meal was roughly equivalent across all the participants. We measured height and weight of participants using standard procedures. We also assessed the hunger level on a scale ranging from 1 (not hungry at all) to 10 (very hungry) and assure the participants were not in a deprived state prior to the fMRI scan.</p></sec><sec><title>Behavioral tests</title><p>Wechsler Abbreviated Scale of Intelligence [WASI, (Wechsler, <xref rid="B89" ref-type="bibr">1999</xref>)]. The WASI was used to measure a participant's Intelligence Quotient (IQ) and basic aspects of cognitive functioning. The WASI is designed for use with a broad age range (from 6 to 89 years of age), is nationally standardized and, similar to other Wechsler scales. It consists of four subtests (Vocabulary, Similarities, Block Design and Matrix Reasoning) chosen based on the high loadings on general intellectual ability (g) and the cognitive skills tapped by each. A combination of the four subtests yields a Full Scale IQ score.</p><p>Youth/Adolescent Eating Questionnaire (YAQ) (Rockett et al., <xref rid="B72" ref-type="bibr">1995</xref>). We used the YAQ to assess eating behavior in real life. The YAQ is a self-report food frequency questionnaire with acceptable validity and reliability (Rockett et al., <xref rid="B72" ref-type="bibr">1995</xref>, <xref rid="B71" ref-type="bibr">1997</xref>). It asks about intake of 132 food items over the past year and food items can be grouped for analysis (Xie et al., <xref rid="B98" ref-type="bibr">2003</xref>; Field et al., <xref rid="B36" ref-type="bibr">2004</xref>). In the present study, we were mainly interested in snack and vegetable food consumption. The YAQ includes 25 questions assessing intake of snack foods. Snack items included the items high in sugar (e.g., fruit rollups, Pop-tarts) and those high in fat/high salt (e.g., potato chips, crackers). Reported consumption to these items was summed to calculate daily servings according to previous studies (Field et al., <xref rid="B36" ref-type="bibr">2004</xref>; Xie et al., <xref rid="B98" ref-type="bibr">2003</xref>). The same calculation was done for vegetable items (e.g., celery, carrot).</p></sec><sec><title>fMRI tasks</title><p>Participants were scanned while performing an event-related IGT. As described in previous studies (Bechara et al., <xref rid="B6" ref-type="bibr">1994</xref>, <xref rid="B8" ref-type="bibr">1999</xref>), the IGT is a computerized version of a gambling task with an automated and computerized method for collecting data. In the IGT, four decks of cards labeled A&#x02032;, B&#x02032;, C&#x02032; or D&#x02032; are displayed on the computer screen. The subject is required to select one card at a time from one of the four decks. When the subject selects a card, a message is displayed on the screen indicating the amount of money the subject has won or lost. Choosing a card can result in an immediate reward (the immediate reward is higher in decks A&#x02032; and B&#x02032; relative to Decks C&#x02032; and D&#x02032;). As the game progresses, there are also unpredictable losses associated with each deck. Total losses are on average higher in decks A&#x02032; and B&#x02032; relative to decks C&#x02032; and D&#x02032;, thus creating a conflict in each choice, i.e., decks A&#x02032; and B&#x02032; are disadvantageous in the long-term (even though they bring higher immediate reward), whereas decks C&#x02032; and D&#x02032; are advantageous in the long-term (i.e., the long-term losses are smaller than the short-term gains, thus yielding a net profit). Net decision-making scores are obtained by subtracting the total number of selections from the disadvantageous decks (A&#x02032; and B&#x02032;) from the total number selections from the advantageous decks (C&#x02032; and D&#x02032;). Thus, positive numbers reflect good decisions, while negative numbers reflect bad decisions.</p></sec><sec><title>fMRI protocol</title><p>Participants lay supine on a scanner bed and viewed visual stimuli back-projected onto a screen through a mirror built into the head coil. The IGT was written in Matlab (Mathworks) based on Psychtoolbox (<ext-link ext-link-type="uri" xlink:href="http://www.psychtoolbox.org">www.psychtoolbox.org</ext-link>). Participants were given instructions on the IGT. Details of these instructions have been published previously (Bechara et al., <xref rid="B9" ref-type="bibr">2000</xref>). We used an event-related design of the IGT which was described in a recent paper (Koritzky et al., <xref rid="B55" ref-type="bibr">2013</xref>). Each trial of the IGT includes two phases: a decision phase and a feedback phase. In the decision phase, participants were requested to select a card from four Decks (A&#x02032;, B&#x02032;, C&#x02032; or D&#x02032;) by pressing the corresponding button when a message (&#x0201c;Pick a Card&#x0201d;) was displayed at the center of screen. In the feedback phase, a message was shown to inform the participants how much money they won or lost based on their choice of cards. The time for the responses to be made in the decision phase was between 3 s and 7 s. The mean was 4 s since this interval varied randomly between trials. At the feedback stage, participants were informed how much money they won or lost by their selected card. The feedback phase last for 3 s. If the trial is a win-only trial (i.e., no loss), the feedback (&#x0201c;you win $X&#x0201d;) was displayed for 1.5 s, followed by a 1.5 s blank screen. If the trial is a win-but-loss trial, the win feedback (&#x0201c;you win $X&#x0201d;) was displayed for 1.5 s, followed by a 1.5 s display of the loss feedback (&#x0201c;but you also lose $X&#x0201d;). The mean length of the inter-trial interval was 2 s with variation from 1.1 s to 3.2 s. The design was optimized with an in-house program to maximize efficiency. There were total 100 trials and lasted for 15 min.</p><p>fMRI was acquired in the Dana and David Dornsife Cognitive Neuroscience Imaging Center at the USC with a 3T Siemens MAGNETOM Tim/Trio scanner. Z-SAGA sequence with PACE (Prospective Acquisition Correction) was used for functional scan to collect blood oxygen level-dependent (BOLD) signals. This specific sequence is dedicated to reduce signal loss in the prefrontal and orbitofrontal areas, with the following scanning parameters: TR/TE = 2000/25 ms; flip angle = 90&#x000b0;; 64 &#x000d7; 64 matrix size with resolution 3 &#x000d7; 3 mm². Thirty-one 3.5-mm axial slices were used to cover the whole cerebral cortex and most of the cerebellum with no gap. The anatomical T1-weighted structural scan was done using an MPRAGE sequence (TR/TE/TI = 2530/3.1/800 ms; flip angle 10&#x000b0;; 208 sagittal slices; 256 &#x000d7; 256 matrix size with spatial resolution as 1 &#x000d7; 1 &#x000d7; 1 mm³).</p></sec><sec><title>fMRI analysis</title><p>FEAT (fMRI Expert Analysis Tool, part of FSL package, <ext-link ext-link-type="uri" xlink:href="http://www.fmrib.ox.ac.uk/fsl">www.fmrib.ox.ac.uk/fsl</ext-link>) was used for image preprocessing and statistical analysis. Standard preprocessing procedures were performed including brain extraction, image realignment, smooth (5 mm FWHM Gaussian kernel), and temporal filtering (100 s cut-off). A two-step registration procedure was used whereby EPI images were first registered to the MPRAGE structural image, and then into standard MNI space, using affine transformations (Jenkinson and Smith, <xref rid="B47" ref-type="bibr">2001</xref>). Registration from MPRAGE structural image to standard space was further refined using FNIRT non-linear registration. Statistical analyses were performed in the native image space, with the statistical maps normalized to the standard space prior to higher-level analysis.</p><p>The data were modeled at the first level using a general linear model within FSL's FILM module. To examine brain activations related to decision making, two types of events were modeled: (1) decision-making stage, and (2) feedback stage. In this paper, we were particularly interested in the BOLD responses related to the decision-making phase (i.e., the deck selection of the IGT). The event onsets were convolved with a canonical hemodynamic response function (HRF, double-gamma) to generate the regressors used in the GLM. Temporal derivatives were included as covariates of no interest to improve statistical sensitivity. Null events were not explicitly modeled, and therefore constituted an implicit baseline. Missing trials were modeled separately as a nuisance variable. The six movement parameters were also included as covariates in the first-level general linear model.</p><p>Higher level random-effect model was tested for group activation in decision making stages (i.e., decision making stage VS baseline) in particular using FMRIB's Local Analysis of Mixed Effect stage 1 only (Beckmann et al., <xref rid="B10" ref-type="bibr">2003</xref>; Woolrich et al., <xref rid="B95" ref-type="bibr">2004</xref>) with automatic outlier detection (Woolrich, <xref rid="B94" ref-type="bibr">2008</xref>). Unless otherwise noted, group images were thresholded using cluster detection statistics, with a height threshold of <italic>Z</italic> &#x0003e; 2.3 and a cluster probability of <italic>p</italic> &#x0003c; 0.05, corrected for whole-brain multiple comparisons based on Gaussian Random Field Theory (GRFT).</p><p>To test the correlation between brain activation in the decision making phase of IGT and dietary intake, region of Interests (ROI) were created from clusters of voxels with significant activation in the voxelwise analyses. Brain activation (% signal change) in these regions when making decisions was extracted using a method suggested by Mumford (<ext-link ext-link-type="uri" xlink:href="http://mumford.fmripower.org/perchange_guide.pdf">http://mumford.fmripower.org/perchange_guide.pdf</ext-link>). Robust regression was used to minimize the impact of outliers in the behavioral data, using iteratively reweighted least squares implemented in the robustfit command in the MATLAB Statistics Toolbox (Tom et al., <xref rid="B82" ref-type="bibr">2007</xref>). Reported <italic>r</italic>-values reflect (non-robust) Pearson product-moment correlation values, whereas the reported <italic>p</italic>-values and regression lines are based on the robust regression results (Tom et al., <xref rid="B82" ref-type="bibr">2007</xref>).</p></sec></sec><sec sec-type="results" id="s3"><title>Results</title><sec><title>Behavior results</title><sec><title>Demographic variables</title><p>Participants in the study fell within the normal range of the body mass index (BMI). Average BMI was 21.88 &#x000b1; 1.62, with a range of 19.1&#x02013;25. IQ scores were all within a normal range (118.29 &#x000b1; 8.6, range = 103&#x02013;132). Participants reported 2.57 &#x000b1; 1.88 on the hunger rating scale (1-not at all hungry; 10-extremely hungry), reflecting the fact that they were being evaluated in a non-food deprived state. With regard to dietary intake, participants reported consuming 2.95 &#x000b1; 2.15 servings/day of vegetables and 1.0 &#x000b1; 0.84 servings/day of fatty and sugary snacks. Participants reported consuming significantly more vegetables than snacks in their daily life [<italic>T</italic>₍₂₃₎ = 3.52, <italic>p</italic> &#x0003c; 0.01]. No age or gender differences were observed on consumption of vegetables or snacks, BMI, IQ, or IGT net scores and hunger ratings.</p></sec><sec><title>Partial correlations</title><p>Table <xref ref-type="table" rid="T1">1</xref> shows partial correlations among the following variable measures: vegetables, snacks, BMI, IQ, the IGT net scores, and hunger ratings after controlling for age and gender. Vegetable consumption did not correlate with consumption of snacks (<italic>r</italic> = &#x02212;0.01, <italic>p</italic> &#x0003e; 0.05). Although these relationships were not statistically significant, vegetable and snack consumption were negatively and positively correlated with BMI (<italic>r</italic> = &#x02212;0.19, <italic>r</italic> = 0.21, respectively). Moreover, none of the variables were significantly correlated with the IGT net scores. Finally, the more vegetables the participants consumed in their daily life, the higher their self-reported hunger rating prior to the fMRI session (<italic>r</italic> = 0.43, <italic>p</italic> &#x0003c; 0.05, corrected for multiple comparison).</p><table-wrap id="T1" position="float"><label>Table 1</label><caption><p><bold>Partial correlations among vegetables, snacks, BMI, IQ, SOPT and the IGT net scores after controlling for age and gender</bold>.</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1"><bold>Measures</bold></th><th align="center" rowspan="1" colspan="1"><bold>2</bold></th><th align="center" rowspan="1" colspan="1"><bold>3</bold></th><th align="center" rowspan="1" colspan="1"><bold>4</bold></th><th align="center" rowspan="1" colspan="1"><bold>5</bold></th><th align="center" rowspan="1" colspan="1"><bold>6</bold></th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">1. Vegetables</td><td align="char" char="." rowspan="1" colspan="1">&#x02212;0.01</td><td align="char" char="." rowspan="1" colspan="1">&#x02212;0.19</td><td align="char" char="." rowspan="1" colspan="1">0.30</td><td align="char" char="." rowspan="1" colspan="1">0.16</td><td align="char" char="." rowspan="1" colspan="1">0.43<xref ref-type="table-fn" rid="TN1">^*</xref></td></tr><tr><td align="left" rowspan="1" colspan="1">2. Snacks</td><td rowspan="1" colspan="1"/><td align="char" char="." rowspan="1" colspan="1">0.21</td><td align="char" char="." rowspan="1" colspan="1">&#x02212;0.15</td><td align="char" char="." rowspan="1" colspan="1">0.13</td><td align="char" char="." rowspan="1" colspan="1">0.03</td></tr><tr><td align="left" rowspan="1" colspan="1">3. BMI</td><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/><td align="char" char="." rowspan="1" colspan="1">&#x02212;0.02</td><td align="char" char="." rowspan="1" colspan="1">0.08</td><td align="char" char="." rowspan="1" colspan="1">&#x02212;0.1</td></tr><tr><td align="left" rowspan="1" colspan="1">4. IQ</td><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/><td align="char" char="." rowspan="1" colspan="1">&#x02212;0.01</td><td align="char" char="." rowspan="1" colspan="1">&#x02212;0.11</td></tr><tr><td align="left" rowspan="1" colspan="1">5. IGT net scores</td><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/><td align="char" char="." rowspan="1" colspan="1">&#x02212;0.22</td></tr><tr><td align="left" rowspan="1" colspan="1">6. Hungry rating</td><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/></tr></tbody></table><table-wrap-foot><p>Results of two-tailed significance tests are denoted by superscripts.</p><fn id="TN1"><label>*</label><p><italic>P &#x0003c; 0.05, IGT = Iowa Gambling Task</italic>.</p></fn></table-wrap-foot></table-wrap></sec><sec><title>IGT performance</title><p>The fMRI optimized version of the IGT task involved 100 trials (or 100 card selections). The trials are divided into five blocks of 20 trials each. In each block, the number of selections from Decks A&#x02032; and B&#x02032; (the disadvantageous decks) and the number of selections from Decks C&#x02032; and D&#x02032; (the advantageous decks) are counted and a net score for each block ((C&#x02032; + D&#x02032;) &#x02013; (A&#x02032; + B&#x02032;)) is obtained. A net score above zero implies that participants are selecting cards advantageously, and a net score below zero implies disadvantageous selection. The behavioral results revealed a significant effect of block after the Greenhouse-Geisser adjustment [<italic>F</italic>_{(3.6, 81.7)} = 5.98; <italic>P</italic> &#x0003c; 0.001]. As shown in Figure <xref ref-type="fig" rid="F1">1</xref>, the participants in this study showed normal learning as the task progressed. They gradually switched their preferences toward the advantageous decks (C&#x02032; and D&#x02032;) and away from the disadvantageous decks (A&#x02032; and B&#x02032;), as reflected by increasingly positive net scores.</p><fig id="F1" position="float"><label>Figure 1</label><caption><p><bold>The Iowa Gambling Task net scores ((C&#x02032; + D&#x02032;) &#x02013; (A&#x02032; + B&#x02032;)) across five blocks of 20 cards expressed as mean &#x000b1; SE</bold>. Positive net scores reflect advantageous (non-impaired performance) while negative net scores reflect disadvantageous (impaired) performance.</p></caption><graphic xlink:href="fnins-08-00350-g0001"/></fig></sec></sec><sec><title>Neuroimaging results</title><sec><title>IGT activity during the decision stage</title><p>As shown in Figure <xref ref-type="fig" rid="F2">2</xref> and Table <xref ref-type="table" rid="T2">2</xref>, during the decision stage, the IGT activated brain regions belonging to both the impulsive system (namely the right amygdala and ventral striatum) and the reflective system (namely the VMPFC and dorsolateral prefrontal cortex (DLPFC), and anterior cingulate cortex (ACC). The IGT also elicited activity in the &#x0201c;urge/craving&#x0201d; system, namely the insular cortex. Activity was also observed in additional neural regions (e.g., temporal cortex, post-central cortex and visual cortex), but there were no <italic>a priori</italic> hypotheses regarding the roles of these brain regions in the behaviors under the current study.</p><fig id="F2" position="float"><label>Figure 2</label><caption><p><bold>fMRI results of the Iowa Gambling Task (IGT) during the decision stage</bold>. Both the impulsive system, including the bilateral putamen/caudate, and the reflective system including the bilateral dorsoateral prefrontal cortex (DLPFC), ventromedial prefrontal cortex (VMPFC), and anterior cingulate cortex (ACC) are involved in the decision stage of the IGT. Activation in IGT also includes insula and visual cortex.</p></caption><graphic xlink:href="fnins-08-00350-g0002"/></fig><table-wrap id="T2" position="float"><label>Table 2</label><caption><p><bold>Brain activity of the Iowa Gambling Task during the decision stage</bold>.</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1"><bold>L/R</bold></th><th align="left" rowspan="1" colspan="1"><bold>Brain regions</bold></th><th align="center" rowspan="1" colspan="1"><bold>N of voxels</bold></th><th align="center" colspan="3" rowspan="1"><bold>MNI coordinates</bold></th><th align="center" rowspan="1" colspan="1"><bold>Z</bold></th></tr><tr><th rowspan="1" colspan="1"/><th rowspan="1" colspan="1"/><th rowspan="1" colspan="1"/><th align="center" rowspan="1" colspan="1"><bold>x</bold></th><th align="center" rowspan="1" colspan="1"><bold>y</bold></th><th align="center" rowspan="1" colspan="1"><bold>z</bold></th><th rowspan="1" colspan="1"/></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">L/R</td><td align="left" rowspan="1" colspan="1">Visual cortex</td><td align="center" rowspan="1" colspan="1">17663</td><td align="center" rowspan="1" colspan="1">&#x02212;6</td><td align="center" rowspan="1" colspan="1">&#x02212;96</td><td align="center" rowspan="1" colspan="1">&#x02212;2</td><td align="center" rowspan="1" colspan="1">6.05</td></tr><tr><td align="left" rowspan="1" colspan="1">R</td><td align="left" rowspan="1" colspan="1">Frontal pole/VMPFC/DLPFC</td><td align="center" rowspan="1" colspan="1">2277</td><td align="center" rowspan="1" colspan="1">42</td><td align="center" rowspan="1" colspan="1">0</td><td align="center" rowspan="1" colspan="1">58</td><td align="center" rowspan="1" colspan="1">5.03</td></tr><tr><td align="left" rowspan="1" colspan="1">L/R</td><td align="left" rowspan="1" colspan="1">Thalamus/Brain stem/Ventral striatum</td><td align="center" rowspan="1" colspan="1">1321</td><td align="center" rowspan="1" colspan="1">&#x02212;6</td><td align="center" rowspan="1" colspan="1">&#x02212;24</td><td align="center" rowspan="1" colspan="1">6</td><td align="center" rowspan="1" colspan="1">5.03</td></tr><tr><td align="left" rowspan="1" colspan="1">L</td><td align="left" rowspan="1" colspan="1">Frontal pole/VMPFC/DLPFC</td><td align="center" rowspan="1" colspan="1">1217</td><td align="center" rowspan="1" colspan="1">&#x02212;36</td><td align="center" rowspan="1" colspan="1">52</td><td align="center" rowspan="1" colspan="1">16</td><td align="center" rowspan="1" colspan="1">4.82</td></tr><tr><td align="left" rowspan="1" colspan="1">L/R</td><td align="left" rowspan="1" colspan="1">ACC</td><td align="center" rowspan="1" colspan="1">1002</td><td align="center" rowspan="1" colspan="1">4</td><td align="center" rowspan="1" colspan="1">26</td><td align="center" rowspan="1" colspan="1">32</td><td align="center" rowspan="1" colspan="1">5.13</td></tr><tr><td align="left" rowspan="1" colspan="1">L</td><td align="left" rowspan="1" colspan="1">SPL/SMG</td><td align="center" rowspan="1" colspan="1">569</td><td align="center" rowspan="1" colspan="1">&#x02212;30</td><td align="center" rowspan="1" colspan="1">&#x02212;54</td><td align="center" rowspan="1" colspan="1">36</td><td align="center" rowspan="1" colspan="1">4.36</td></tr><tr><td align="left" rowspan="1" colspan="1">L/R</td><td align="left" rowspan="1" colspan="1">PCC</td><td align="center" rowspan="1" colspan="1">478</td><td align="center" rowspan="1" colspan="1">&#x02212;2</td><td align="center" rowspan="1" colspan="1">&#x02212;26</td><td align="center" rowspan="1" colspan="1">24</td><td align="center" rowspan="1" colspan="1">5.50</td></tr><tr><td align="left" rowspan="1" colspan="1">L</td><td align="left" rowspan="1" colspan="1">Post-central cortex</td><td align="center" rowspan="1" colspan="1">300</td><td align="center" rowspan="1" colspan="1">&#x02212;58</td><td align="center" rowspan="1" colspan="1">&#x02212;20</td><td align="center" rowspan="1" colspan="1">46</td><td align="center" rowspan="1" colspan="1">4.20</td></tr><tr><td align="left" rowspan="1" colspan="1">L</td><td align="left" rowspan="1" colspan="1">Temporal cortex</td><td align="center" rowspan="1" colspan="1">302</td><td align="center" rowspan="1" colspan="1">&#x02212;62</td><td align="center" rowspan="1" colspan="1">&#x02212;20</td><td align="center" rowspan="1" colspan="1">16</td><td align="center" rowspan="1" colspan="1">4.78</td></tr><tr><td align="left" rowspan="1" colspan="1">R</td><td align="left" rowspan="1" colspan="1">Amygdala/Ventral striatum</td><td align="center" rowspan="1" colspan="1">216</td><td align="center" rowspan="1" colspan="1">28</td><td align="center" rowspan="1" colspan="1">&#x02212;2</td><td align="center" rowspan="1" colspan="1">&#x02212;10</td><td align="center" rowspan="1" colspan="1">3.61</td></tr><tr><td align="left" rowspan="1" colspan="1">L</td><td align="left" rowspan="1" colspan="1">Hippocampus</td><td align="center" rowspan="1" colspan="1">151</td><td align="center" rowspan="1" colspan="1">&#x02212;18</td><td align="center" rowspan="1" colspan="1">&#x02212;28</td><td align="center" rowspan="1" colspan="1">&#x02212;10</td><td align="center" rowspan="1" colspan="1">4.27</td></tr><tr><td align="left" rowspan="1" colspan="1">L</td><td align="left" rowspan="1" colspan="1">Insula</td><td align="center" rowspan="1" colspan="1">110</td><td align="center" rowspan="1" colspan="1">&#x02212;38</td><td align="center" rowspan="1" colspan="1">2</td><td align="center" rowspan="1" colspan="1">0</td><td align="center" rowspan="1" colspan="1">3.45</td></tr><tr><td align="left" rowspan="1" colspan="1">R</td><td align="left" rowspan="1" colspan="1">Insula</td><td align="center" rowspan="1" colspan="1">89</td><td align="center" rowspan="1" colspan="1">42</td><td align="center" rowspan="1" colspan="1">14</td><td align="center" rowspan="1" colspan="1">&#x02212;4</td><td align="center" rowspan="1" colspan="1">3.73</td></tr></tbody></table><table-wrap-foot><p><italic>VMPFC: Ventromedial Prefrontal Cortex; DLPFC: Dorsolateral Prefrontal Cortex; SPL: Superior Parietal Lobe; SMG: Suparamarginal Cortex; ACC: Anterior Cingulate Cortex; PCC: Posterior Cingulate Cortex</italic>.</p></table-wrap-foot></table-wrap></sec><sec><title>Correlations between brain activity and eating behaviors</title><p>We performed a correlation analyses between the consumption of vegetables or snacks, and the BOLD response elicited by IGT performance in the decision stage. The results shown in Figure <xref ref-type="fig" rid="F3">3</xref> reveals that higher consumption of vegetables correlates with higher activity in the left superior frontal gyrus (SFG) (<italic>r</italic> = 0.55, <italic>P</italic> &#x0003c; 0.01), and with lower activity in the right insula (<italic>r</italic> = &#x02212;0.66, <italic>P</italic> &#x0003c; 0.001). Figure <xref ref-type="fig" rid="F4">4</xref> reveals that higher snack consumption correlates with lower activity in the left frontal pole (<italic>r</italic> = &#x02212;0.63, <italic>P</italic> &#x0003c; 0.001), and with higher activity in the right ventral striatum (<italic>r</italic> = 0.60, <italic>P</italic> &#x0003c; 0.01) and right insular cortex (<italic>r</italic> = 0.56, <italic>P</italic> &#x0003c; 0.01).</p><fig id="F3" position="float"><label>Figure 3</label><caption><p><bold>Functional MRI correlation of vegetable consumption and IGT activity during decision stage of the Iowa Gambling Task (IGT). (A)</bold> Regions show a significant positive correlation (red) between vegetable consumption and the left superior frontal gyrus (SFG) activation. <bold>(B)</bold> Regions show significant negative correlation (blue) between vegetable consumption and the right insula. <bold>(C, D)</bold> Scatterplots of correlations between vegetable consumption and the averaged covariance of the parameter estimates in the left SFG and right insular cortex, respectively.</p></caption><graphic xlink:href="fnins-08-00350-g0003"/></fig><fig id="F4" position="float"><label>Figure 4</label><caption><p><bold>Functional MRI correlation of snack consumption and the Iowa Gambling Task (IGT) activity during decision stage. (A)</bold> Regions show significant positive correlation (red) between snack consumption and right insular cortex activation. <bold>(B)</bold> Regions show significant positive correlation (red) between snack consumption and right ventral striatum activation. <bold>(C)</bold> Regions show significant negative correlation (blue) between snack consumption and left frontal pole. <bold>(D, E, F)</bold> Scatterplots of correlations between snack consumption and the averaged covariance of parameter estimates in the right insular cortex, right ventral striatum and left frontal pole, respectively.</p></caption><graphic xlink:href="fnins-08-00350-g0004"/></fig></sec></sec></sec><sec sec-type="discussion" id="s4"><title>Discussion</title><p>In the current study, IGT performance elicited neural activity in neural systems hypothesized to play key roles in complex decision-making: (1) neural regions belonging to the so-called &#x0201c;reflective system&#x0201d; concerned with impulse control and self-control, namely the VMPFC, the DLPFC, as well as the ACC in both hemispheres; (2) neural regions belonging to the so-called &#x0201c;impulsive system&#x0201d; concerned with reward and habit behaviors, namely the striatum in both hemispheres; and (3) neural systems implicated in the processing of interoceptive signals and their translation into what may subjectively become experienced as an urge, namely the insula in both hemispheres. Moreover, higher consumption of vegetables positively correlated with activity in the left superior frontal gyrus (SFG) (i.e., a component of the reflective system), but negatively correlated with activity in the right insular cortex. In contrast, high consumption of snacks negatively correlated with activity in the left frontal pole (a part of the reflective system), but positively correlated with activity in the right ventral striatum and right insula cortex.</p><p>These results are consistent with several behavioral studies showing that poor decision-making scores measured by the IGT are found in obese, patients with binge eating disorders, and overweight adolescents (Pignatti et al., <xref rid="B68" ref-type="bibr">2006</xref>; Brogan et al., <xref rid="B15" ref-type="bibr">2010</xref>; Verdejo-Garcia et al., <xref rid="B85" ref-type="bibr">2010</xref>; Danner et al., <xref rid="B22" ref-type="bibr">2012</xref>; Fagundo et al., <xref rid="B34" ref-type="bibr">2012</xref>). They are also consistent with previous reports that performance on the IGT was related to the magnitude of weight loss in a diet-induced weight loss intervention in overweight women (Witbracht et al., <xref rid="B93" ref-type="bibr">2012</xref>). The brain regions implicated in this study are also consistent with several previous studies on food (high vs. low calorie), weight (obese vs. average weight), and activity in neural regions (Killgore et al., <xref rid="B52" ref-type="bibr">2003</xref>; Pelchat et al., <xref rid="B66" ref-type="bibr">2004</xref>; DelParigi et al., <xref rid="B28" ref-type="bibr">2005</xref>, <xref rid="B29" ref-type="bibr">2006</xref>; Killgore and Yurgelun-Todd, <xref rid="B53" ref-type="bibr">2005</xref>; Davis et al., <xref rid="B25" ref-type="bibr">2007</xref>, <xref rid="B24" ref-type="bibr">2010</xref>; Stice et al., <xref rid="B79" ref-type="bibr">2008</xref>; Small, <xref rid="B76" ref-type="bibr">2009</xref>; Batterink et al., <xref rid="B2" ref-type="bibr">2010</xref>; Ng et al., <xref rid="B63" ref-type="bibr">2011</xref>; He et al., <xref rid="B46" ref-type="bibr">2014b</xref>). The unique contribution of our current study is the use of a neural framework that assigns multiple neural regions to functionally specialized neural systems involved in behavioral decisions (Naqvi and Bechara, <xref rid="B61" ref-type="bibr">2010</xref>; Noel et al., <xref rid="B64" ref-type="bibr">2013</xref>). More importantly, our current study examines the dynamics among these neural systems (i.e., hyperactivity in one system, but hypoactivity in another). The examination of these dynamics is especially significant in terms of devising therapeutic strategies.</p><p>High consumption of high-calorie snacks in real-life correlated with higher activity in the ventral striatum. The ventral striatum has long been known for its role in various types of reward, including food reward (Demos et al., <xref rid="B30" ref-type="bibr">2012</xref>; Mehta et al., <xref rid="B59" ref-type="bibr">2012</xref>). Animal studies indicate that direct pharmacological activation of the ventral striatum increases preferentially the intake of foods high in fat and sugar, even in animals fed beyond apparent satiety (Petrovich et al., <xref rid="B67" ref-type="bibr">2002</xref>; Kelley, <xref rid="B50" ref-type="bibr">2004</xref>). In humans, several lines of evidence suggest that high calorie food may induce greater incentive values in obese individuals compared to normal controls (Volkow et al., <xref rid="B87" ref-type="bibr">2012</xref>; Tomasi and Volkow, <xref rid="B83" ref-type="bibr">2013</xref>). Behavioral studies also show that compared to their normal controls, overweight children indicate that high calorie food (pizza and snack food) is more reinforcing (Temple et al., <xref rid="B81" ref-type="bibr">2008</xref>). Thus, our current findings are consistent with this long line of studies in both animals and humans.</p><p>A unique aspect of the current study is that we used a monetary reward in order to engage the neural systems sub-serving decision-making instead of food reward. The results indicate that the observed changed dynamics between these neural systems apply not only to food, but to reward in general, including monetary reward. This is quite consistent with the conceptualization about a common currency for reward that relates to dopamine, especially that associated with the ventral striatum (McClure et al., <xref rid="B58" ref-type="bibr">2004b</xref>). Many studies have shown that this region is similarly engaged by food as well as monetary cues. For instance, increased ventral striatal activity (reflecting increased dopamine) potentiated the rewarding effects of food as well as the association between food cues and the feeling of pleasure associated with food consumption (Smith and Robbins, <xref rid="B77" ref-type="bibr">2013</xref>). Also the anticipation of food (as opposed the experience of food) is rewarding and it is associated with increased ventral striatal activity (that presumably reflects increased dopamine release) (Smith and Robbins, <xref rid="B77" ref-type="bibr">2013</xref>). Even the numerous behavioral studies in humans that suggested that obese individuals are hyper-responsive to food cues in a wide range of assessments (Braet and Crombez, <xref rid="B12" ref-type="bibr">2003</xref>; Halford et al., <xref rid="B44" ref-type="bibr">2004</xref>), and the behavioral studies in both healthy and overweight populations suggesting that personality traits of reward drive predict food craving, overeating, and relative body weight (Davis and Woodside, <xref rid="B27" ref-type="bibr">2002</xref>; Bulik et al., <xref rid="B17" ref-type="bibr">2003</xref>; Davis et al., <xref rid="B26" ref-type="bibr">2004</xref>). are all considered as consistent with the constructs that increased reward sensitivity is linked to a biologically-based personality trait regulated by mesocorticolimbic dopamine systems (Cohen et al., <xref rid="B19" ref-type="bibr">2005</xref>; Evans et al., <xref rid="B32" ref-type="bibr">2006</xref>). Indeed the increased neuronal activity elicited by fatty food cues in the ventral striatum predicted the macronutrient choice at an <italic>ad libitum</italic> buffet, i.e., greater ventral striatum activity predicted the choice of food items with higher fat content (Mehta et al., <xref rid="B59" ref-type="bibr">2012</xref>). This ventral striatal activity also predicted weight gain 6 months later (Demos et al., <xref rid="B30" ref-type="bibr">2012</xref>). In parallel, these same striatal regions responsive to food cues have also been shown to respond in a similar manner to monetary reward (Breiter and Rosen, <xref rid="B14" ref-type="bibr">1999</xref>; Breiter et al., <xref rid="B13" ref-type="bibr">2001</xref>), thus supporting the notion that altered dynamics between these neural systems may be general, and not specific to food reward.</p><p>Higher right insular activity correlated with more snack, but less vegetable, consumption in real life. Given the hypothesized role of the insular cortex in translating interoceptive signals into what one may subjectively experience as a feeling of desire, anticipation or urge (Naqvi and Bechara, <xref rid="B60" ref-type="bibr">2009</xref>; Noel et al., <xref rid="B64" ref-type="bibr">2013</xref>), we suggest that engaging the insula system increases the urge or craving for high calorie food by (1) exacerbating activity within the striatal (impulsive) system, and (2) weakening activity of the prefrontal (reflective) system [e.g., see (Noel et al., <xref rid="B64" ref-type="bibr">2013</xref>)]. This suggestion is consistent with studies showing that activity within the insular cortex is associated with food craving (Pelchat et al., <xref rid="B66" ref-type="bibr">2004</xref>). Finally, our study revealed a role for prefrontal regions (parts of the reflective system) in the inhibitory control of some high calorie food items, consistent with several previous studies suggesting a role for the SFG in introspection, self-judgments, and the subjective rating of self-awareness (Goldberg et al., <xref rid="B42" ref-type="bibr">2006</xref>). Goldberg et al. proposed that the left SFG is involved in allowing the individual to reflect upon sensory experiences, to judge their possible significance to the self, and to allow the individual to report about the occurrence of his sensory experience to the outside world (Goldberg et al., <xref rid="B42" ref-type="bibr">2006</xref>). Others implicated the frontal pole area (Broadmann 10) in insight into one's future and the planning of future actions (McClure et al., <xref rid="B57" ref-type="bibr">2004a</xref>; Fellows and Farah, <xref rid="B35" ref-type="bibr">2005</xref>; D'Argembeau et al., <xref rid="B23" ref-type="bibr">2008</xref>; Koritzky et al., <xref rid="B55" ref-type="bibr">2013</xref>). These studies are quite consistent with our early conceptualization on the role of these regions in what we called a &#x0201c;reflective&#x0201d; system in the context of other rewards, namely drugs (e.g., Bechara, <xref rid="B4" ref-type="bibr">2005</xref>). However, the novel contribution of the current study is the examination of the dynamics between multiple neural systems (e.g., hypoactivity in the reflective system combined with hyperactivity in the striatal and insula systems in response to high calorie food).</p><p>Although our early conceptualization about an imbalance between an impulsive and reflective system was initially discussed in the context of drug reward (Bechara, <xref rid="B4" ref-type="bibr">2005</xref>), a similar conceptualization argued that eating disorders and obesity may be associated with a mismatch between the impulsive and reflective systems (Gearhardt et al., <xref rid="B40" ref-type="bibr">2011</xref>; Brooks et al., <xref rid="B16" ref-type="bibr">2013</xref>; Garc&#x000ed;a-Garc&#x000ed;a et al., <xref rid="B38" ref-type="bibr">2013</xref>). Our study is very consistent with these earlier reports, except that we now show that this imbalance also applies to normal people who are not necessarily diagnosed with obesity or eating disorder. Since our study was cross-sectional, we are not able to make inferences about whether the differences in the neural substrates of decision-making reflect the cause or effect of real-life food consumption. It is likely that activities of these brain systems mediate the development of our eating behaviors. This is pertinent to the argument made by some researchers that we should emphasize the importance of focusing on high-risk food substances (and their potential to alter specific brain systems) rather than high-risk people, which has tended to be the focus of most research to date (Gearhardt and Brownell, <xref rid="B39" ref-type="bibr">2013</xref>). An emphasis on such future research could provide an insight on the neural basis and related cognitive and behavioral interventions that help weight management and prevent obesity and other eating disorders (Paolini et al., <xref rid="B65" ref-type="bibr">2012</xref>; Gearhardt and Brownell, <xref rid="B39" ref-type="bibr">2013</xref>).</p><p>Finally, we note that the IGT is a task that taps into the brain mechanisms sub-serving decision-making, but it only involves abstract money/points as a reward, as opposed to food reward. Thus, the task itself does not ask subjects to consume real food, nor to view images of food while in the scanner. As such, the current study using the IGT could potentially be deemed as non-ecological valid, and thus limit the generalization of our results. However, we argue the opposite in that the use of the IGT had several important advantages. First, the use of food related executive function tasks (e.g., go/no go tasks with food stimuli) has been reported multiple times in the literature and yielding consistent results (He et al., <xref rid="B46" ref-type="bibr">2014b</xref>). Second, even structural volumetric measures of ROIs within the so-called &#x0201c;reflective system&#x0201d; showed consistent negative correlations with BMI, independent of using any tasks that involve food images (He et al., <xref rid="B45" ref-type="bibr">2014a</xref>). Hence, the current results using the IGT, which is a complex task that taxes the functions of all three neural systems hypothesized to be engaged in addiction (Li et al., <xref rid="B56" ref-type="bibr">2010</xref>; Xiao et al., <xref rid="B96" ref-type="bibr">2013</xref>), suggest that the relatively poor ability to delay gratification from high calorie food reward is not specific to food reward, but it generalizes to other rewards (and in this case it is monetary reward). These findings are significant as they support the notion that the process leading to overweight and obesity is one that is reflected by a relative imbalance in neural systems implicated in addictive behaviors, and also decision-making in general.</p><sec><title>Conflict of interest statement</title><p>The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.</p></sec></sec>

From genes to behavior: placing cognitive models in the context of biological pathways

<p>Connecting neural mechanisms of behavior to their underlying molecular and genetic substrates has important scientific and clinical implications. However, despite rapid growth in our knowledge of the functions and computational properties of neural circuitry underlying behavior in a number of important domains, there has been much less progress in extending this understanding to their molecular and genetic substrates, even in an age marked by exploding availability of genomic data. Here we describe recent advances in analytical strategies that aim to overcome two important challenges associated with studying the complex relationship between genes and behavior: (i) reducing distal behavioral phenotypes to a set of molecular, physiological, and neural processes that render them closer to the actions of genetic forces, and (ii) striking a balance between the competing demands of discovery and interpretability when dealing with genomic data containing up to millions of markers. Our proposed approach involves linking, on one hand, models of neural computations and circuits hypothesized to underlie behavior, and on the other hand, the set of the genes carrying out biochemical processes related to the functioning of these neural systems. In particular, we focus on the specific example of value-based decision-making, and discuss how such a combination allows researchers to leverage existing biological knowledge at both neural and genetic levels to advance our understanding of the neurogenetic mechanisms underlying behavior.</p>

<contrib contrib-type="author"><name><surname>Saez</surname><given-names>Ignacio</given-names></name><xref ref-type="aff" rid="aff1">¹</xref><uri xlink:type="simple" xlink:href="http://community.frontiersin.org/people/u/6511"/></contrib><contrib contrib-type="author"><name><surname>Set</surname><given-names>Eric</given-names></name><xref ref-type="aff" rid="aff1">¹</xref><xref ref-type="aff" rid="aff2">²</xref><uri xlink:type="simple" xlink:href="http://community.frontiersin.org/people/u/185701"/></contrib><contrib contrib-type="author"><name><surname>Hsu</surname><given-names>Ming</given-names></name><xref ref-type="aff" rid="aff1">¹</xref><xref ref-type="author-notes" rid="fn001">^*</xref><uri xlink:type="simple" xlink:href="http://community.frontiersin.org/people/u/17444"/></contrib>

Frontiers in Neuroscience

<sec sec-type="introduction" id="s1"><title>Introduction</title><p>There is widespread interest in the application of formal computational models to connect behavior to its underlying biological substrates (Glimcher and Rustichini, <xref rid="B24" ref-type="bibr">2004</xref>; Sugrue et al., <xref rid="B67" ref-type="bibr">2005</xref>; Landis and Insel, <xref rid="B40" ref-type="bibr">2008</xref>; Rangel et al., <xref rid="B56" ref-type="bibr">2008</xref>; Behrens et al., <xref rid="B1" ref-type="bibr">2009</xref>; Ebstein et al., <xref rid="B17" ref-type="bibr">2010</xref>). At the neural level, we now have substantial knowledge of computational properties underlying a number of important domains of human cognition and behavior, and the set of brain regions that perform these functions (Glimcher and Rustichini, <xref rid="B24" ref-type="bibr">2004</xref>; Landis and Insel, <xref rid="B40" ref-type="bibr">2008</xref>; Rangel et al., <xref rid="B56" ref-type="bibr">2008</xref>; Behrens et al., <xref rid="B1" ref-type="bibr">2009</xref>; Ebstein et al., <xref rid="B17" ref-type="bibr">2010</xref>). An intriguing question that has only recently become possible to address is the extent to which we can extend this understanding to uncover the genetic forces shaping and constraining these systems (Frank and Fossella, <xref rid="B21" ref-type="bibr">2011</xref>; den Ouden et al., <xref rid="B29" ref-type="bibr">2013</xref>).</p><p>This has important scientific and clinical implications. First, identifying mechanisms by which genomic differences lead to variations at cellular and neural circuit levels, resulting in changes in behavior and cognition, is an important step toward informing and improving the diagnosis and treatments of behavioral disorders (Glimcher and Rustichini, <xref rid="B24" ref-type="bibr">2004</xref>; Sugrue et al., <xref rid="B67" ref-type="bibr">2005</xref>; Landis and Insel, <xref rid="B40" ref-type="bibr">2008</xref>; Ebstein et al., <xref rid="B17" ref-type="bibr">2010</xref>; Insel, <xref rid="B34" ref-type="bibr">2010</xref>; Kapur et al., <xref rid="B37" ref-type="bibr">2012</xref>). In addition, the prospect that computational models can uncover not only computations at the circuit level, but also gene variation that influences these circuits, should substantially bolster the prospect that they have clinical utility (Meyer-Lindenberg and Weinberger, <xref rid="B43" ref-type="bibr">2006</xref>; Rangel et al., <xref rid="B56" ref-type="bibr">2008</xref>; Behrens et al., <xref rid="B1" ref-type="bibr">2009</xref>; Montague et al., <xref rid="B46" ref-type="bibr">2012</xref>).</p><p>However, despite the growing number of studies linking gene variation to complex behavioral traits in humans, comparatively few studies have attempted to link genotype data to behavioral phenotypes through the lens of computational models of behavior. This is even so in cases where existing models have shown considerable validity at both neurophysiological and molecular levels, as in the case of reinforcement learning models of reward-guided behavior (Schultz et al., <xref rid="B65" ref-type="bibr">1997</xref>; Dayan and Niv, <xref rid="B14" ref-type="bibr">2008</xref>; Doya, <xref rid="B15" ref-type="bibr">2008</xref>; Frank and Fossella, <xref rid="B21" ref-type="bibr">2011</xref>; den Ouden et al., <xref rid="B29" ref-type="bibr">2013</xref>). One possible reason is these computational models, which are most often used in neuroimaging studies and therefore focus on capturing variation at the circuit level, are simply not well suited for capturing variation that operates on the developmental and evolutionary timescales (Bell and Robinson, <xref rid="B2" ref-type="bibr">2011</xref>).</p><p>Here we argue that, on the contrary, computational models are useful precisely because they provide valuable mechanistic explanations at the intermediate neural levels so often absent in human studies linking genes, and behavior (Frank and Fossella, <xref rid="B21" ref-type="bibr">2011</xref>). That is, because the effects of genetic and molecular mechanisms operating at longer timescales are necessarily mediated by neural mechanisms, computational models provide a framework through which we can unveil the impact of more distal effects of genes and molecules on the intermediate systems (Landis and Insel, <xref rid="B40" ref-type="bibr">2008</xref>; Zhong et al., <xref rid="B76" ref-type="bibr">2009</xref>; Bogdan et al., <xref rid="B5" ref-type="bibr">2012</xref>).</p><p>Perhaps most importantly, when combined with emerging analytical approaches in genomics that enable researchers to focus on specific biological pathways and networks, these models allow behavior across different studies to be unified within a common biological framework. In doing so, this promises to move us beyond accumulating lists of significant gene-behavior pairings, and toward attempting to organize them in a unified and coherent mechanistic framework.</p><p>Here, we review analytical strategies and concepts to enable a biologically informed characterization of neurogenetic mechanisms underlying value-based decision-making in humans, and describe how to integrate them with computational principles that are beginning to emerge from the burgeoning neuroimaging literature tying formal mathematical models to choice behavior at the level of neural circuits. Our goal is to propose a new analytical strategy that combines computational models and gene pathways that can be used to unveil mechanistic relationship between genetic variants and behavior. To this end, we will review the foundations of the approach: (1) computational models of behavior, and how they can be used as cognitive phenotypes, and (2) the use of gene pathways as a strategy to balance the competing demands of interpretability and discovery in the analysis of human genetic data; finally, we will review a prior application of these principles (Set et al., <xref rid="B66" ref-type="bibr">2014</xref>) as a case study that illustrates the fruitful combination of these two approaches.</p></sec><sec><title>Genetics of human behavior</title><p>Two main research strategies exist for the identification of genes associated with heritable traits&#x02014;candidate gene approaches and genome-wide association (GWAS) approaches (Yang et al., <xref rid="B74" ref-type="bibr">2010</xref>; Flint and Munafo, <xref rid="B19" ref-type="bibr">2013</xref>) (see Box <xref ref-type="boxed-text" rid="Box1">1</xref> for glossary of genetic terms). While linkage studies are also available, we focus on association studies in this perspective as they are increasingly the primary tool in the case of human studies (Sabb et al., <xref rid="B63" ref-type="bibr">2009</xref>). First, in candidate gene studies, one or a small number of gene variants with known effects on the protein structure or expression are used to detect genotype-phenotype associations (Flint et al., <xref rid="B18" ref-type="bibr">2001</xref>; Flint and Munafo, <xref rid="B19" ref-type="bibr">2013</xref>). These studies are typically motivated by prior knowledge of biological mechanisms underlying the physiology of a certain trait. In GWAS studies, this goal is achieved using all gene variants across the entire genome, which are independently tested in a hypothesis-free manner (International Schizophrenia Consortium, <xref rid="B35" ref-type="bibr">2009</xref>; Rucker et al., <xref rid="B62" ref-type="bibr">2011</xref>).</p><boxed-text id="Box1" position="float"><label>Box 1</label><caption><title>Some prerequisites for understanding neural and genetic studies of behavior.</title></caption><list list-type="bullet"><list-item><p><bold>Allele</bold>: One of two or more forms of a gene, located on a specific position on a chromosome.</p></list-item><list-item><p><bold>Candidate gene studies</bold>: Studies that focus on association of pre-specified genes of interest, typically based on prior knowledge, and phenotypes.</p></list-item><list-item><p><bold>Genome-wide association studies (GWAS)</bold>: Studies that aims to find associations by scanning common genetic variation in the entire genome in hypothesis-free manner.</p></list-item><list-item><p><bold>Gene pathway</bold>: A group of functionally related genes that mediate a particular biological process, e.g., DA functioning.</p></list-item><list-item><p><bold>Linkage Disequilibrium</bold>: Extent to which alleles are correlated due to common inheritance. Alleles of nearby genes are typically in high linkage disequilibrium.</p></list-item><list-item><p><bold>Minor allele frequency (MAF)</bold>: The frequency at which the <italic>least common</italic> allele occurs in a given population. Typically alleles with MAF below 5% or 10% are excluded from the study.</p></list-item><list-item><p><bold>Single Nucleotide Polymorphism (SNP)</bold>: In genetics, a difference in DNA sequence among individuals. A common form of a genetic polymorphism is a SNP, which occurs when a nucleotide&#x02014;A, T, C or, G&#x02014;differs between individuals. The human genome contains millions of SNPs. Below are a list of common types of polymorphisms.</p><list list-type="simple"><list-item><p>&#x025e6; <bold>Exonic mutation</bold>: Polymorphisms in gene region that remains present within the final mature RNA product.</p></list-item><list-item><p>&#x025e6; <bold>Synonymous mutation</bold>: Exonic mutations that do not modify the protein encoded by the gene. Previously thought to be silent but now known to have potential effects on transcription, splicing, mRNA transport, and translation (Sauna and Kimchi-Sarfaty, <xref rid="B64" ref-type="bibr">2011</xref>).</p></list-item><list-item><p>&#x025e6; <bold>Non-synonymous mutation</bold>: Exonic mutations where the protein encoded by the gene is modified.</p></list-item><list-item><p>&#x025e6; <bold>Intronic mutation</bold>: Region within a gene that is removed by RNA splicing while the final mature RNA product of a gene is being generated. Previously thought to be silent but now known to have potential effects on splicing accuracy and translational efficiency (Cartegni et al., <xref rid="B9" ref-type="bibr">2002</xref>).</p></list-item><list-item><p>&#x025e6; <bold>Untranslated region (UTR)</bold>: Region directly adjacent of coding region of the gene, important for regulation of RNA translation.</p></list-item><list-item><p>&#x025e6; <bold>Intergenic regions</bold>: Stretches of DNA sequences located between genes. Most variants in this region have no currently known function, but some are thought to have regulatory functions. In humans, intergenic regions comprise about 80%&#x02013;90% of the genome.</p></list-item></list></list-item></list></boxed-text><p>Despite the rapid growth of studies based on these approaches, and the accumulation of gene markers implicated in behavior, findings from these studies have been subject to widespread skepticism about their (i) reliability, and (ii) ability to inform us about the genetic architecture underlying behaviors and disorders where they are affected (Figure <xref ref-type="fig" rid="F1">1A</xref>) (Hart et al., <xref rid="B28" ref-type="bibr">2013</xref>). At least in the case of human behavior, many behaviors of interest relate to highly human-specific activities that are the result of complex social, cognitive, and cultural influences. Thus, even in cases where candidate genes are carefully motivated and have clear biological implications, their connection to basic cognitive processes underlying the trait of interest can be unclear (Figure <xref ref-type="fig" rid="F1">1A</xref>) (Flint et al., <xref rid="B18" ref-type="bibr">2001</xref>; Reuter et al., <xref rid="B58" ref-type="bibr">2011</xref>; Flint and Munafo, <xref rid="B19" ref-type="bibr">2013</xref>).</p><fig id="F1" position="float"><label>Figure 1</label><caption><p><bold>Cognitive models as quantitative descriptions of putative intermediate mechanisms</bold>. <bold>(A)</bold> For most human behaviors of interest, the intermediate neural, synaptic, and molecular mechanisms are far from clear. As a result, studies of the genetic basis of these behaviors are forced to directly examine the effects of the chosen genotype onto behavior, without consideration of the ways in which genetic variation propagates through and constrains these intermediate levels. <bold>(B)</bold> Computational models provide a principled way in which complex patterns of behavior can be quantified and reduced to a lower-dimensional space via the set of parameters governing the computations. Variation of the parameters in the population can be related to underlying genetic variation and other inter-individual factors (i.e., environmental) and interactions. In the example, a schematic of a simple reinforcement learning model is presented where the parameter &#x003b1; _i governs the extent to which an individual organism is sensitive to more recent rewards relative to past ones. This parameter in turn can be thought of as an intermediate cognitive phenotype that is under the influence of genes, environment, and their interaction. <bold>(C)</bold> When validated at the neural level, these models can serve as quantitative descriptions of the missing intermediate mechanisms through which genes exert their influence on behavior. In this sense, model parameters are equivalent to cognitive phenotypes and can act as a nexus that mechanistically connects different biological levels underlying behavior.</p></caption><graphic xlink:href="fnins-08-00336-g0001"/></fig><p>To use a concrete example, consider a previous study finding that voting propensity is associated with serotonin gene polymorphisms, specifically alleles in the MAOA and SERT (Fowler and Dawes, <xref rid="B20" ref-type="bibr">2008</xref>) (Box <xref ref-type="boxed-text" rid="Box1">1</xref>). Although such studies provide valuable insights into possible biological substrates of an important feature of modern human civilization, a vast gap exists between the functions of these genes on the one hand, and the act of voting in an election in a modern Western democracy.</p><p>As the authors of the study point out, even taking genetic associations identified in the study as given, the nature of the genetic contribution remains far from clear (Fowler and Dawes, <xref rid="B20" ref-type="bibr">2008</xref>). First, the identified polymorphisms may play a role in promoting prosociality, but it could also be related to aggression. It may increase the sense of satisfaction one derives from fulfilling a civic duty. It may increase the strength of desire for expression. It may be part of a broad constellation of personality traits. This is only a partial list of the possible ways that serotonin genes might influence voting propensity.</p><p>Perhaps more importantly, the lack of mechanistic insights has contributed to a fragmentation that impedes the accumulation of knowledge critical for scientific advancement. A central question, therefore, is whether it is possible for genetic studies of behavior, like those in morphology or simpler types of phenotypes, to trace through the complex biological pathways connecting genes and behavior in a way that makes it possible to integrate diverse behavior-genotype associations in a biologically based framework.</p></sec><sec><title>Cognitive models as candidate mechanisms</title><p>Note that in all the above cases, the key question is how to relate and map diverse behavioral phenotypes to a more constrained set of intermediate cognitive phenotypes (Houle et al., <xref rid="B31" ref-type="bibr">2010</xref>; Rasetti and Weinberger, <xref rid="B57" ref-type="bibr">2011</xref>; Bogdan et al., <xref rid="B5" ref-type="bibr">2012</xref>). That is, a crucial step in overcoming these hurdles is to reduce the distal behavioral phenotype to a set of molecular, physiological, and neural processes that render them closer to the actions of genetic forces. In the case of model organisms we have the ability to interrogate these molecular and neural mechanisms directly, but most are unavailable in humans due to their invasive nature.</p><p>At least in the case of the brain, our understanding has been transformed by recent applications of formal computational models that connect behavior to their underlying neural circuitry (Schultz et al., <xref rid="B65" ref-type="bibr">1997</xref>; Montague et al., <xref rid="B47" ref-type="bibr">2004</xref>; Behrens et al., <xref rid="B1" ref-type="bibr">2009</xref>; Maia and Frank, <xref rid="B41" ref-type="bibr">2011</xref>). In a number of cases, these models have been shown to have considerable validity at both behavioral and neural levels (O'Doherty et al., <xref rid="B50" ref-type="bibr">2007</xref>; Rangel et al., <xref rid="B56" ref-type="bibr">2008</xref>). For example, the basic temporal difference model is able to explain a variety of reward-guided behavior using a single parameter governing the strength of impact of the reward prediction error on future behavior (Figure <xref ref-type="fig" rid="F1">1B</xref>) (Schultz et al., <xref rid="B65" ref-type="bibr">1997</xref>; Montague et al., <xref rid="B47" ref-type="bibr">2004</xref>). At the neural level, although details regarding interpretation remain debated (Berridge, <xref rid="B3" ref-type="bibr">2007</xref>), substantial evidence points to a key role of midbrain dopaminergic neurons in carrying a quantitative signal guiding choice behavior, which can be captured using both neurophysiological evidence in model organisms and neuroimaging evidence in humans (Dayan and Niv, <xref rid="B14" ref-type="bibr">2008</xref>).</p><p>At the genetic level, then, cognitive models provide a principled way in which complex patterns of behavior can be quantified and reduced to a lower-dimensional space via the set of parameters governing the computations. Variation of the parameters in the population can be related to underlying genetic variation, and other inter-individual factors (i.e., environmental), and interactions (Figure <xref ref-type="fig" rid="F1">1C</xref>). This parameter in turn can be thought of as an intermediate cognitive phenotype that is under the influence of genes, environment, and their interaction.</p><p>In an early example of this approach, (Frank et al., <xref rid="B22" ref-type="bibr">2007</xref>) investigated how genetic polymorphisms in candidate genes affected reward and avoidance learning in humans. Using a cognitive model that captures distinct computational components connected to reward and avoidance learning, the authors found that variation in different dopaminergic genes, specifically DARPP-32, DRD2, and COMT, were associated with separate parameters governing reward and avoidance learning. Importantly, these findings can be directly connected to our knowledge of how these genes relate to dopaminergic functioning. For example, both DARPP-32 and DRD2 are thought to affect primarily striatal, as opposed to prefrontal, dopamine (Missale et al., <xref rid="B45" ref-type="bibr">1998</xref>), whereas the reverse is true for COMT (M&#x000e4;nnist&#x000f6; and Kaakkola, <xref rid="B42" ref-type="bibr">1999</xref>). The fact that striatal dopamine genes affected the speed of learning is notable as it is consistent with a broad class of neurophysiological and neuroimaging work in both human and animal studies.</p><p>For example, associations of D2 receptor gene variation to behavior can be linked to its potential effects on striatal D2 receptor density, which are then linked to systems-level changes that translate to changes in behavior. Importantly, the predictions of this working model can be tested using pharmacological manipulation, PET imaging, or via invasive methods using model organisms. In contrast, such a systems approach would be considerably more challenging in distal phenotypes such as voting behavior.</p><p>Taken together, connecting genes to computational models therefore would help to address a key limitation in many studies of genetic basis of behavior (Figure <xref ref-type="fig" rid="F1">1C</xref>) (Frank and Fossella, <xref rid="B21" ref-type="bibr">2011</xref>; den Ouden et al., <xref rid="B29" ref-type="bibr">2013</xref>; Set et al., <xref rid="B66" ref-type="bibr">2014</xref>). Importantly, a focus on mechanisms can advance existing conversation from one focused on &#x0201c;gene-hunting,&#x0201d; with a goal of accumulating highly significant polymorphisms regardless of their functional importance (or &#x0201c;behavior hunting&#x0201d; in the case of candidate genes, where one seeks to accumulate a list of behaviors regardless of their interdependence), to one focused on mechanism and the phenotype of interest.</p></sec><sec><title>Gene pathways</title><p>Despite these promising features, candidate mechanisms are not by themselves sufficient to overcome the formidable challenges arising from the inherent complexity of genomic data. First, the sheer size of modern gene array data have resulted in a situation where it is often the rule rather than the exception that significant gene markers have little direct relationship to plausible biological mechanisms (Figure <xref ref-type="fig" rid="F2">2A</xref>). For example, a recent study (Rietveld et al., <xref rid="B59" ref-type="bibr">2013</xref>) identified a genome-wide significant SNP that is significantly associated with a complex and distal phenotype, academic achievement; however, this SNP is not located in the proximity of any genes which might mediate its biological effect, and so how the effect comes to be is unclear even if we had a precise cognitive model of academic achievement. That is, even when there are candidate mechanisms available, the associated gene markers often have no discernible relationship with the mechanism.</p><fig id="F2" position="float"><label>Figure 2</label><caption><p><bold>(A)</bold> Comparison of genomic analysis approaches illustrated on a DNA schematic, including 22 pairs of autosomal and sex chromosomes. Orange shaded regions indicate genetic materials used for analysis. (Left) GWAS analyses assess association of phenotype of interest with all sequenced SNPs, often hundreds of thousands, independently. Hence all chromosomes are shaded orange. (Right) On the other side of the technical spectrum, candidate gene approaches focus on a single polymorphism, often well motivated by prior biological data. In this example, the non-synonymous rs4680 SNP of the COMT gene is selected. (Middle) A pathway approach offers a compromise, where prior biological information is leveraged to define a set of genes, organized around a biological process. In this example, all genes whose products have an impact on dopaminergic neurotransmission are selected. <bold>(B)</bold> Dopamine metabolic pathway captures the biological process involved in neurotransmission, including dopamine synthesis (blue), dopamine signal transduction (orange), and dopamine transport and clearance (green). In principle, genes that regulate/act on these dopaminergic genes can also be included, although we do not include them here as they have broad functions in the nervous system.</p></caption><graphic xlink:href="fnins-08-00336-g0002"/></fig><p>Second, genes do not function independently but within biological pathways, and they interact within biological networks (Figure <xref ref-type="fig" rid="F2">2B</xref>) (Wang et al., <xref rid="B71" ref-type="bibr">2007</xref>; Ramanan et al., <xref rid="B55" ref-type="bibr">2012</xref>). In particular, the accumulation of weak but coordinated effects arising from multiple alleles within specific biological systems is increasingly thought to be an important source of phenotypic variation. The fact that both GWAS and candidate gene studies focus on individual genotype markers poses a challenge for them to detect subtle effects distributed across the genome (Wang and Abbott, <xref rid="B70" ref-type="bibr">2008</xref>). This point is particularly crucial as it is now widely accepted that common alleles, including those used in candidate gene studies, exhibit modest effect sizes. As such, the statistical approach of treating individual alleles as independent results in a potentially serious loss of power by ignoring the underlying biological structure.</p><p>In recent years, studies that strike a middle ground, using so-called pathway approaches, are becoming increasingly popular (Figure <xref ref-type="fig" rid="F2">2A</xref> and Box <xref ref-type="boxed-text" rid="Box1">1</xref>) (Wang et al., <xref rid="B71" ref-type="bibr">2007</xref>, <xref rid="B72" ref-type="bibr">2010</xref>; Yaspan and Veatch, <xref rid="B75" ref-type="bibr">2011</xref>; Ramanan et al., <xref rid="B55" ref-type="bibr">2012</xref>). A genetic pathway consists of a group of functionally related genes that mediate a particular biological process, e.g., DA functioning (Figure <xref ref-type="fig" rid="F2">2B</xref>). Each gene along the pathway encodes a protein that carries out a specific biological function. For example, the DAT1 gene encodes the dopamine transporter (DAT), whose function is to remove dopamine from the synaptic cleft, thus terminating the signal of the neurotransmitter. Although these pathways are abstractions of complex biological process that have no discrete start or end points, they have been invaluable to researchers as they capture and organize our knowledge in a parsimonious and tractable manner.</p><p>The pathway approach addresses these issues by limiting our search to a set of genes underlying a specific biological process, thereby improving the interpretability of potential results (Wang et al., <xref rid="B71" ref-type="bibr">2007</xref>, <xref rid="B72" ref-type="bibr">2010</xref>; Yaspan and Veatch, <xref rid="B75" ref-type="bibr">2011</xref>; Ramanan et al., <xref rid="B55" ref-type="bibr">2012</xref>). For behavior, there are a number of molecularly defined pathways that are suitable as candidates based on previous anatomical, pharmacological and physiological studies in both humans and animals: neuromodulatory pathways (serotonergic, dopaminergic, noradrenergic, etc.), hormonal and, neuropeptide pathways (oxytocin, vasopressin), synaptic plasticity related pathways, growth factors such as neurotrophins (BDNF, NT-3, NT-4, etc.), and transcription factors, to name a few.</p><p>In particular, because a pathway approach fosters a view centered on biological processes, as opposed to individual polymorphisms, statistical inference can be made at multiple level of analysis, from SNP, to gene, to pathway, in a way that can adapt to the particular question, but without being completely unconstrained as in GWA studies (Chen et al., <xref rid="B11" ref-type="bibr">2010</xref>; Ramanan et al., <xref rid="B55" ref-type="bibr">2012</xref>). For example, compared to previous studies making inferences at the level of individual SNP or VNTR, in Set et al. [36] we considered the combined impact of all common polymorphisms within individual DA genes. With larger sample sizes, it is possible to compare whole pathways with hundreds of variants, as have been done in a number of disease studies.</p></sec><sec><title>Case study: connecting cognitive models to gene pathways</title><p>Given the number of analytical steps involved in our proposed approach, we give in this section a detailed step-by-step guide to conducting pathway studies of cognition and behavior. To fix ideas we will use the specific example of a recent study by Set et al. (<xref rid="B66" ref-type="bibr">2014</xref>) that applied dopaminergic pathways to strategic learning.</p><sec><title>Phenotype</title><p>Strategic learning refers to decisions made in the presence of competitive or cooperative intelligent agents, where, in addition to learning about rewards and punishments available in the environment, agents need to also anticipate and respond to actions of others competing for the same rewards (Figure <xref ref-type="fig" rid="F3">3A</xref>) (Fudenberg, <xref rid="B23" ref-type="bibr">1998</xref>; Hofbauer and Sigmund, <xref rid="B30" ref-type="bibr">1998</xref>). Specifically, Set et al. (<xref rid="B66" ref-type="bibr">2014</xref>), applied the well-established experience weighted attraction (EWA) model to reduce individual variation in competitive winner-take-all paradigm to two key parameters capturing (1) the degree to which players are sensitive to actions of others, captured by &#x003b4;, and (2) learning rate or sensitivity of players to more recent observations relative to past ones, captured by &#x003c1; (Figure <xref ref-type="fig" rid="F3">3B</xref>) (Sutton and Barto, <xref rid="B68" ref-type="bibr">1998</xref>; Camerer, <xref rid="B8" ref-type="bibr">2003</xref>; Zhu et al., <xref rid="B77" ref-type="bibr">2012</xref>).</p><fig id="F3" position="float"><label>Figure 3</label><caption><p><bold>Mapping neural and genetic correlates of strategic learning</bold>. <bold>(A)</bold> Choice behavior in economic games provides basic material to characterize neural and genetic correlates of behavior. In this example, subjects make sequential choices over 240 rounds of a multi-strategy competitive learning paradigm, the patent race. <bold>(B)</bold> Trial-by-trial variation in behavior is captured by a model&#x02014;experience weighted attraction (EWA)&#x02014;containing two parameters governing two distinct aspects of strategic learning. (i) Belief learning parameter &#x003b4; that captures the degree to which participants anticipate and respond to the actions of others, and (ii) learning rate parameter &#x003c1; captures the strength of past experiences on behavior. Individual differences, i.e., person-by-person variation, is captured by different parameter values of &#x003b4;_{<italic>i</italic>} and &#x003c1;_{<italic>i</italic>} for participant i. <bold>(C)</bold> Neural circuits subserving specific computations can be mapped using outputs of the calibrated model outputs at a trial-by-trial level. In the example, belief learning signals were localized to mPFC activity, whereas reinforcement learning signals to striatal activity. Adapted from Zhu et al. (<xref rid="B77" ref-type="bibr">2012</xref>). <bold>(D)</bold> Genetic influence on behavior can similarly be mapped by connecting gene variation in the dopaminergic pathway to intermediate phenotype, captured by parameter variation at the individual level. In the example, variation in belief learning &#x003b4;_{<italic>i</italic>} is significantly associated with variation in genes responsible for dopaminergic degradation (COMT, MAOB, MAOA), which govern dopaminergic levels in the prefrontal cortex but not striatum. In contrast, variation in learning rate &#x003c1;_{<italic>i</italic>} is significantly associated with variation in genes highly expressed in the striatum (DAT1, DRD2), but not prefrontal cortex. Interestingly, COMT variation is also associated with learning rate. Adapted from Zhu et al. (<xref rid="B77" ref-type="bibr">2012</xref>) and Adapted from Set et al. (<xref rid="B66" ref-type="bibr">2014</xref>) should be made consistent.</p></caption><graphic xlink:href="fnins-08-00336-g0003"/></fig><p>Importantly, this computational characterization of behavior was able to capture trial-by-trial variation in fMRI BOLD activity of players during game play (Hsu and Zhu, <xref rid="B32" ref-type="bibr">2012</xref>; Zhu et al., <xref rid="B77" ref-type="bibr">2012</xref>). Specifically, whereas the medial prefrontal cortex was found to respond selectively to belief-based inputs and reflected individual differences in degree of engagement of belief learning, striatal activity was correlated with both reinforcement and belief-based signals, suggesting possible convergence of these signals in the striatum (Figure <xref ref-type="fig" rid="F3">3C</xref>) (Zhu et al., <xref rid="B77" ref-type="bibr">2012</xref>).</p></sec><sec><title>Pathway selection</title><p>First, given the phenotype of interest and candidate cognitive model, one needs to determine the appropriate pathway involved. One option is to select a set of genes that are related to a specific biological function, such as neurotransmission. For many behavioral or cognitive processes, neuromodulatory systems such as dopamine and serotonin are particularly attractive targets (e.g., Figure <xref ref-type="fig" rid="F2">2B</xref>).</p><p>In the case of strategic learning, dopaminergic mechanisms are a natural candidate owing to the involvement of reward learning processes. Moreover, DA transmission is known to exhibit remarkable regional variation in expression levels of genes coding for the set of enzymes, receptors, and transporters involved in DA functioning (Pierce and Kumaresan, <xref rid="B53" ref-type="bibr">2006</xref>; O'Connell and Hofmann, <xref rid="B49" ref-type="bibr">2012</xref>) (Figure <xref ref-type="fig" rid="F2">2B</xref>). In the prefrontal cortex, where DAT1 expression is low, genes regulating enzymatic breakdown, in particular COMT and to a lesser extent isoforms of the MAO genes, are important determinants of DA flux (Nemoda et al., <xref rid="B48" ref-type="bibr">2011</xref>). In contrast, these genes have much less impact on striatal DA levels, where DAT1 expression is high (Frank and Fossella, <xref rid="B21" ref-type="bibr">2011</xref>). On the receptor side, regional variation results from distribution of DA receptor types (Missale et al., <xref rid="B45" ref-type="bibr">1998</xref>). Receptors of the D1 family, D1 and, D5, are expressed throughout the brain. In contrast, receptors in the D2 family exhibit more regional specificity: D2 receptors are expressed primarily in the dorsal striatum, D3 receptors in the ventral striatum, including nucleus accumbens but less so in dorsal striatum, and D4 receptors in the frontal cortex and limbic regions (Missale et al., <xref rid="B45" ref-type="bibr">1998</xref>).</p><p>Another popular technique is to use gene ontology annotations, such as the Gene Ontology (GO) database (Harris et al., <xref rid="B27" ref-type="bibr">2004</xref>). A third option is to select genes that are expressed at a given developmental time in brain areas that are known or suspected to be implicated in said processes. Yet many others are possible, and we are only beginning to appreciate how to best divide the complex set of molecular and cellular processes in ways that shed light on cognitive processes.</p><p>Because the underlying biological processes have no real starting or ending points, the pathway definitions require decisions that trade off between coverage and interpretability. For example, for neurotransmitter-centered pathways, the focus point is the locus of action of the neurotransmitter, i.e., the neurotransmitter-receptor interaction in the synaptic cleft. From that pivot point, sets of genes that are involved in neurotransmitter synthesis, signal transduction, and signal degradation form the core of the pathway, which can then be concentrically expanded to include secondary messengers in the postsynaptic side, regulatory elements such as kinases and phosphatases, transcription factors, etc. The cost of such an expansion is a loss of statistical power and biological interpretability; for example, secondary messengers are promiscuous and are typically activated in response to activation of numerous membrane receptors, a characteristic akin to the pleiotropy of genetic effects.</p></sec><sec><title>Assigning data elements to genes</title><p>Once the gene set underlying the pathway is determined, the set of data elements, whether SNPs, variable number of tandem repeats, or copy number variations (Box <xref ref-type="boxed-text" rid="Box1">1</xref>), must be decided. Due to the current technical capability and low cost of SNP sequencing, the former is by far the most common. All SNPs located in known coding or regulatory regions are typically analyzed, since they offer a straightforward connection to the biological effects of the genetic variation, mediated by changes in protein sequence. However, to capture possible regulatory variations, all SNPs within the coding region of the gene (both exonic and intronic) as defined by current genomic atlases may be included, as was the case in Set et al. (<xref rid="B66" ref-type="bibr">2014</xref>). Furthermore, upstream or downstream SNPs can have regulatory functions such as effect in transcriptional or translational efficiency, and may also be included.</p></sec><sec><title>Dealing with linkage disequilibrium</title><p>Genes often contain multiple SNPs. Due to their physical proximity, they are often co-inherited and thus variation in them is typically correlated, an effect called linkage disequilibrium (LD) (Box <xref ref-type="boxed-text" rid="Box1">1</xref>). Analyzing each of these as an independent factor inflates the multiple comparison problem, and therefore statistical methods have been proposed to deal with this issue, such as principal component regression (PCR) (Wang and Abbott, <xref rid="B70" ref-type="bibr">2008</xref>).</p><p>Specifically, this approach uses the first few principal components (PCs), so-called eigenSNPs, computed from the sample covariance matrix of SNP genotype scores as regressors, and has been used in a number of previous gene expression and SNP marker studies (10). For example, in Set et al. (<xref rid="B66" ref-type="bibr">2014</xref>), 4 eigenSNPs contained 91% of the variation in the COMT gene, from an initial set of 17 SNPs that exceeded an MAF threshold of 0.1.</p><p>Compared to traditional candidate gene approaches, this multilocus approach can be used to detect association between a phenotype and groups of SNPs (genes), and is more efficient when there exists weaker but coordinated effects arising from multiple SNP markers. Other solutions, such as shrinkage methods including LASSO and random forests, have been developed but increase the computational burden substantially (Bridges et al., <xref rid="B6" ref-type="bibr">2011</xref>).</p></sec><sec><title>Combining pathways and models</title><p>Once inter-subject genetic (through pathway analysis) and phenotypic (through computational models) variability have been assessed, they must be mapped onto one another. A multiple linear regression of genetic variation on estimated parameter values offers a simple way of doing this. Effectively, optimal weights for each piece of genetic variation (SNP, eigenSNP, etc.) are assigned to explain as much of the variation in parameter space as possible (Wang and Abbott, <xref rid="B70" ref-type="bibr">2008</xref>).</p><p>For example, in Set et al. (<xref rid="B66" ref-type="bibr">2014</xref>), this involved allowing each parameter (e.g., &#x003b4;) of the model to vary according to the set of associated eigenSNPs of each gene in the DA pathway. In the case of the COMT gene, this included the addition of four additional parameters {&#x003b4;₁, &#x003b4;₂, &#x003b4;₃, &#x003b4;₄} corresponding to the four eigenSNPs of the DAT1 gene, in addition to the population (mean) parameter &#x003b4;. Intuitively, this analysis asks the question of whether inclusion of genetic information can improve statistical fit of the model by capturing individual differences.</p><p>At this stage, nuisance regressors that are known or suspected to impact the behavior under study can be included. For example, inclusion of the first 10&#x02013;20 whole-genome principal components is an effective way of controlling for population stratification (Price et al., <xref rid="B54" ref-type="bibr">2006</xref>).</p></sec><sec><title>Assessment of significance</title><p>Although asymptotic tests are possible in this approach, potential violations of standard assumptions have led to the widespread use of permutation tests, which requires a weaker set of assumptions to be valid (Wang et al., <xref rid="B72" ref-type="bibr">2010</xref>; Winkler et al., <xref rid="B73" ref-type="bibr">2014</xref>). Here, the null distribution is created by shuffle the gene-behavior pairings, such that the observed association has to be significantly higher than that of a &#x0201c;random&#x0201d; genome (Wang et al., <xref rid="B72" ref-type="bibr">2010</xref>; Winkler et al., <xref rid="B73" ref-type="bibr">2014</xref>).</p><p>Alternatively, if one has access to GWAS data, one can compare the association in a particular gene to comparison &#x0201c;null&#x0201d; genes outside of the pathway that possess similar statistical properties (e.g., same number of SNPs that reduce to similar number of eigenSNPs). In Set et al. (<xref rid="B66" ref-type="bibr">2014</xref>), this is referred to as the &#x0201c;empirical <italic>p</italic>-value,&#x0201d; to distinguish from the permutation <italic>p</italic>-value. Importantly, because these genes are selected because of a hypothesized negative relationship (e.g., genes that do not express in the CNS), they provide a highly useful negative control with which to dissociate candidate pathways against null pathways.</p></sec><sec><title>Biological interpretation of results</title><p>In the past, a significant hurdle existed in attempting to connect gene association findings to intermediate neural mechanisms. In the case of Set et al. (<xref rid="B66" ref-type="bibr">2014</xref>), restricting attention to the gene level and pathway alleviated potential interpretational issues considerably. First, the fact that belief learning processes engaged primarily medial prefrontal cortex accord well with the associations between belief learning parameter &#x003b4; and variations in the COMT, MAOB, and, MAOA genes (Figure <xref ref-type="fig" rid="F3">3D</xref>). All three are genes implicated in dopamine catabolism and are responsible for regulating dopaminergic levels in the prefrontal cortex. In contrast, learning rate &#x003c1; was found to be significantly associated with variation in striatal genes DAT1 and DRD2 (Figure <xref ref-type="fig" rid="F3">3D</xref>). Overall, these findings raise a number of interesting questions regarding the anatomical specificity of the genetic effects, which can be tested in imaging genetic studies. For example, an interesting question is whether the COMT effect on learning rate is exerted through prefrontal DA or its indirect effects on striatal dopamine, as has been reported in previous imaging genetics findings (Dreher et al., <xref rid="B16" ref-type="bibr">2009</xref>).</p></sec><sec><title>Validation and followup</title><p>One important drawback of including all polymorphisms is that the functionality of the identified polymorphisms can be obscure. For example, of the 143 common SNPs in Set et al. (<xref rid="B66" ref-type="bibr">2014</xref>), only one, the extremely well-studied rs4680, is associated with a change in protein structure. The rest were either synonymous mutations or resided in intronic or untranslated regions. In recent years, however, there are a growing number of computational methods available to gain further insight into these potential biological functions. They rely on identifying sequences with known biological effects in the DNA sequence including and surrounding SNPs of interest. The SNPInfo web server (<ext-link ext-link-type="uri" xlink:href="http://snpinfo.niehs.nih.gov">http://snpinfo.niehs.nih.gov</ext-link>), for instance, provides a web interface where multiple SNPs can be queried to obtain information about their potential biological effects for SNPs located in coding (protein sequence changes, changes in stop codons) and non-coding (transcription factor binding sites, splicing regulation, miRNA binding sites, etc.) regions.</p><p>In addition to mining existing data, new data can be acquired to gain insight into the nature of the association. For example, in the case of polymorphisms that putatively result in changes in protein concentration, what is the association between protein levels and the behavioral effect? Imaging genetics approaches can be used to gain further insight into the mechanisms whereby a genetic change affects neural mechanisms underlying a cognitive phenotype (Hariri et al., <xref rid="B26" ref-type="bibr">2006</xref>; Klein et al., <xref rid="B38" ref-type="bibr">2007</xref>). Pharmacological manipulations can further be carried out to demonstrate the causal involvement of the identified molecular mechanism. Although not all genes can be targeted, in the case of neural pathways there are a variety of drugs that have been applied to the study of behavior which affect different neurotransmitter systems such as dopamine, serotonin, neuropeptides (e.g., oxytocin) (Kosfeld et al., <xref rid="B39" ref-type="bibr">2005</xref>; Pessiglione et al., <xref rid="B52" ref-type="bibr">2006</xref>; Crockett et al., <xref rid="B13" ref-type="bibr">2008</xref>). For the cases in which a more detailed examination is warranted or for which no pharmacological manipulation is possible, animal models can be used to investigate the impact of a single gene (e.g., gene knockouts, gene knockdowns).</p></sec></sec><sec sec-type="conclusion" id="s2"><title>Conclusion</title><p>In contrast to phenotypes such as morphology, behavior has always presented special challenges for biological studies because of its temporal nature and context dependence (Houle et al., <xref rid="B31" ref-type="bibr">2010</xref>). In the case of human behavior, the situation is even more challenging as many behaviors of interest relate to highly human-specific activities that are the result of complex social, cognitive, and, cultural influences (Bilder et al., <xref rid="B4" ref-type="bibr">2009</xref>; Houle et al., <xref rid="B31" ref-type="bibr">2010</xref>).</p><p>At the neural level, recent applications of functional neuroimaging, combined with formal economic models, have greatly expanded our understanding of the neurocognitive processes underlying complex behaviors, such as decision-making in strategic environments (Behrens et al., <xref rid="B1" ref-type="bibr">2009</xref>; Burke et al., <xref rid="B7" ref-type="bibr">2010</xref>; Hsu and Zhu, <xref rid="B32" ref-type="bibr">2012</xref>; Zhu et al., <xref rid="B77" ref-type="bibr">2012</xref>). At the same time, recent technical advancements have significantly advanced our knowledge of human genetic variation and the location and impact of human genetic polymorphisms.</p><p>Despite such progress, however, there has been surprisingly little attempt to connect and cross-pollinate these different levels in ways that emphasize the relative strengths of each approach while minimizing their weaknesses. In this perspective, we described an approach focusing on specific biological processes in ways that relate systems of functionally-related genes to putative mechanistic models of behavior (Wang et al., <xref rid="B72" ref-type="bibr">2010</xref>; Yaspan and Veatch, <xref rid="B75" ref-type="bibr">2011</xref>; Ramanan et al., <xref rid="B55" ref-type="bibr">2012</xref>). Specifically, this involves linking, on one hand, working models of neural computations carried out by local circuits (Frank and Fossella, <xref rid="B21" ref-type="bibr">2011</xref>), and on the other hand, the set of the biochemical processes that are carried out by genes (Wang et al., <xref rid="B71" ref-type="bibr">2007</xref>; Ramanan et al., <xref rid="B55" ref-type="bibr">2012</xref>).</p><p>Clinically, a better integration of genetic and neural data is an important step toward improving diagnosis and treatment of neuropsychiatric disorders (Gottesman and Gould, <xref rid="B25" ref-type="bibr">2003</xref>; Kapur et al., <xref rid="B37" ref-type="bibr">2012</xref>; Miller and Rockstroh, <xref rid="B44" ref-type="bibr">2013</xref>). Genes involved in dopamine functioning may be directly involved in neuropsychiatric disorders (Gottesman and Gould, <xref rid="B25" ref-type="bibr">2003</xref>; Insel, <xref rid="B34" ref-type="bibr">2010</xref>; Miller and Rockstroh, <xref rid="B44" ref-type="bibr">2013</xref>). In this case, a combined neurogenetic approach would be invaluable in the identification of endophenotypes&#x02014;patterns of brain function that can be linked to a particular genotype (Gottesman and Gould, <xref rid="B25" ref-type="bibr">2003</xref>; Insel, <xref rid="B34" ref-type="bibr">2010</xref>; Miller and Rockstroh, <xref rid="B44" ref-type="bibr">2013</xref>). The elucidation of genetic differences among patients may, for example, lead to improved understanding of diagnostic subtypes or creation of new subtypes (Charney et al., <xref rid="B10" ref-type="bibr">2002</xref>).</p><p>An alternative, and perhaps more likely scenario is that the causative gene resides elsewhere but yet indirectly affect many related systems and circuits, including those mediated by dopaminergic genes (Insel et al., <xref rid="B33" ref-type="bibr">2010</xref>; Papassotiropoulos and de Quervain, <xref rid="B51" ref-type="bibr">2011</xref>). In this case, an understanding of the dopaminergic variation in genetically normal systems is no less valuable by facilitating understanding of therapeutic impacts (Charney et al., <xref rid="B10" ref-type="bibr">2002</xref>). This is in particular if key defective genes identified prove to be difficult to target, in which case downstream genes or pathways affected by the illness that can be repaired constitutes a natural target of intervention (Wang et al., <xref rid="B71" ref-type="bibr">2007</xref>; Chen et al., <xref rid="B11" ref-type="bibr">2010</xref>; Yaspan and Veatch, <xref rid="B75" ref-type="bibr">2011</xref>).</p><p>For some phenotypes of interest to social scientists, such as wealth or the aforementioned education attainment, the phenotype is sufficiently far removed from the underlying biology that little is gained by applying a pathway approach. In these cases, a purely exploratory GWAS approach may well be an appropriate choice. Even in these cases, however, exploratory versions of pathway analyses can be used. For example, &#x0201c;genome-wide pathway analysis&#x0201d; attempts to segment the genome in terms of biological processes and then attempts to find pathways differentially involved in a particular phenotype. This method has proved fruitful in identifying an association between IQ, a complex proxy-phenotype, and heterotrimeric G proteins that are central relay factors that may serve as &#x0201c;signaling bottleneck&#x0201d; for neuronal responses (Ruano et al., <xref rid="B61" ref-type="bibr">2010</xref>). Another set of network-based methods uses graph theory methods to infer networks of genes that are involved in a phenotype, and are particularly useful for dealing with gene-gene interactions (Ramanan et al., <xref rid="B55" ref-type="bibr">2012</xref>).</p><p>However, for a growing class of behavioral and clinical measures, the underlying biologically processes mapping sensory input to behavioral outcomes are increasingly mapped out at both neural and molecular levels. In these cases, pathways represent an important way of capturing our prior knowledge regarding biological processes mediating specific outcomes, and actionable therapeutic targets (Veenstra-VanderWeele and Anderson, <xref rid="B69" ref-type="bibr">2000</xref>). Thus, if we think of a priori pathway selection as a &#x0201c;top-down&#x0201d; approach that generalizes the candidate gene approach, data-driven approaches can be thought of as a &#x0201c;bottom-up&#x0201d; approaches that generalizes the GWAS approach.</p><p>Overall, our approach explicitly acknowledges the inherent tension regarding our current state of knowledge (Robinson et al., <xref rid="B60" ref-type="bibr">2008</xref>; Set et al., <xref rid="B66" ref-type="bibr">2014</xref>). On the one hand, we now have an immense and growing base of knowledge regarding the biological basis of economic behavior, which can explain observation across multiple biological levels and, in some cases, across multiple species (Robinson et al., <xref rid="B60" ref-type="bibr">2008</xref>; Connell and Hofmann, <xref rid="B12" ref-type="bibr">2011</xref>). On the other hand, our knowledge is highly incomplete. For example, we still know little about the precise quantitative relationship between many of the allele variants in DA genes and gene expression levels, nor of their influence on neural circuits (Jia et al., <xref rid="B36" ref-type="bibr">2011</xref>; Set et al., <xref rid="B66" ref-type="bibr">2014</xref>). Finally, and perhaps most importantly, by centering the focus on biological processes as opposed to individual genes, a combined neurogenetic approach allows behavior across different studies to be related to a common set of mathematical principles, thereby moving beyond merely cataloging lists of genes and the myriad of associated behaviors.</p><sec><title>Conflict of interest statement</title><p>The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.</p></sec></sec>

Short-Term Adaptation of Joint Position Sense Occurs during and after Sustained Vibration of Antagonistic Muscle Pairs

<p>Proprioception is critical for the control of many goal-directed activities of daily living. While contributions from skin and joint receptors exist, the muscle spindle is thought to play an important role in allowing accurate judgments of limb position and movement to occur. The discharges elicited from muscle spindles can be degraded by simultaneous agonist-antagonist tendon vibration, causing proprioception to be distorted. Despite this, changes in limb perception that may result from sensory adaptation to this stimulus remain misunderstood. The purpose of this study was, therefore, to investigate short-term proprioceptive adaptation resulting from vibration of antagonistic muscle pairs. We measured elbow joint position sense in 21 healthy young adults while 80&#x02009;Hz vibration was applied simultaneously to the distal tendons of the elbow flexor and extensor muscles. Matching errors were then analyzed during early and late adaptation phases to assess short-term adaptation to the vibration stimuli. Participants committed significant undershoot errors during the early adaptation phase, but were comparable to baseline measurements during the late adaptation phase. When we removed the vibration stimuli and conducted a second joint position matching task, matching variability increased significantly, and participants committed overshoot errors. These results bring into question the efficacy of simultaneous agonist-antagonist tendon vibration to degrade proprioceptive acuity.</p>

<contrib contrib-type="author"><name><surname>Gonzales</surname><given-names>Tomas I.</given-names></name><xref ref-type="aff" rid="aff1">¹</xref><xref ref-type="aff" rid="aff2">²</xref><uri xlink:type="simple" xlink:href="http://frontiersin.org/people/u/172761"/></contrib><contrib contrib-type="author"><name><surname>Goble</surname><given-names>Daniel J.</given-names></name><xref ref-type="aff" rid="aff1">¹</xref><xref ref-type="corresp" rid="cor1">*</xref><uri xlink:type="simple" xlink:href="http://frontiersin.org/people/u/77578"/></contrib>

Frontiers in Human Neuroscience

<sec sec-type="introduction" id="S1"><title>Introduction</title><p>Sensory adaptation is a term that relates to the ability of a sensory system to change responsiveness over time. This process can be predictive in nature, as might be necessary for the case of sensory-motor learning. Alternatively, sensory adaptation can be reactive, serving as a means for allowing equilibrium states to be achieved in the face of external stimuli. In this way, the time course of sensory adaptation is both situation dependent, and dependent on the strength and duration of sensory stimuli.</p><p>While a multitude of studies have investigated sensory adaptation of the traditional five senses (i.e., sight, sound, smell, taste, and touch), few attempts have been made to specifically explore this phenomenon in the proprioceptive system (Desmurget et al., <xref rid="B8" ref-type="bibr">2000</xref>; Seizova-Cajic et al., <xref rid="B34" ref-type="bibr">2007</xref>). Proprioception is a term first coined by Sherrington (<xref rid="B35" ref-type="bibr">1907</xref>) that referred to the set of bodily sensations generated during one&#x02019;s own actions. Over time, the term has come to be defined as the means by which an individual is able to sense and perceive body positions in the absence of vision (for review, see Proske and Gandevia, <xref rid="B25" ref-type="bibr">2009</xref>). Regardless, it has been shown that the underlying neural signals that subserve proprioceptive sense arise from joint, cutaneous, and muscle spindle receptors. Of these &#x0201c;proprioceptors,&#x0201d; it is feedback from muscle spindles that is thought to play a particularly pivotal role in allowing accurate judgments of limb position and movement to be made (Burke et al., <xref rid="B3" ref-type="bibr">1976</xref>; Roll and Vedel, <xref rid="B30" ref-type="bibr">1982</xref>; Roll et al., <xref rid="B31" ref-type="bibr">1989</xref>).</p><p>The importance of muscle spindle signals for proprioceptive sense has most clearly been demonstrated through tendon vibration studies. This experimental paradigm encompasses the application of a mechanical vibration stimulus to the tendon of a target muscle in order to stimulate primary (Ia) muscle spindles. It has been shown using microneurography that the rate of muscle spindle firing increases harmonically in response to tendon vibration (Roll and Vedel, <xref rid="B30" ref-type="bibr">1982</xref>), and that the increased neural signal is perceived as lengthening of the muscle by the brain (Goodwin et al., <xref rid="B14" ref-type="bibr">1972</xref>). Interestingly, this illusory response decreases with prolonged stimulation, and cessation of the vibratory stimulus elicits a transitory (i.e., 30&#x02009;s) kinesthetic aftereffect in which the vibrated limb seems to be moving in the opposite direction of the illusory movement (Seizova-Cajic et al., <xref rid="B34" ref-type="bibr">2007</xref>). This response is believed to correspond to a depression in the firing rate of muscle spindle primary afferents (Ribot-Ciscar et al., <xref rid="B28" ref-type="bibr">1998</xref>), although adaptation at other locations within the nervous system is likely.</p><p>A great deal of evidence now exists supporting the notion that perception of joint movement is based on the central appreciation of primary muscle afferent activity originating from both the shortening and lengthening muscles. More specifically, an imbalance of inputs from agonist and antagonist muscles results in the perception of motion in the corresponding direction (Gilhodes et al., <xref rid="B11" ref-type="bibr">1986</xref>). During natural voluntary movements, this imbalance favors the lengthening muscle, since primary muscle afferents will typically fail to awaken in the shortening (i.e., agonist) muscle (Roll and Vedel, <xref rid="B30" ref-type="bibr">1982</xref>). Mechanical vibration can, therefore, be used to confound information from these channels of proprioceptive information by interfering with the ability of muscle spindles to respond to naturally evoked stimuli (Roll et al., <xref rid="B31" ref-type="bibr">1989</xref>).</p><p>In light of the above findings, several recent attempts have been made to demonstrate the applicability of simultaneous agonist-antagonist mechanical tendon vibration to degrade proprioception in healthy adults (Bock et al., <xref rid="B2" ref-type="bibr">2007</xref>; Vidoni and Boyd, <xref rid="B37" ref-type="bibr">2008</xref>; Ronsse et al., <xref rid="B32" ref-type="bibr">2009</xref>). This dual agonist-antagonist vibration approach might serve as an important research technique, as it could provide a feasible and reversible means for studying the consequences of poor proprioception, known to be characteristic of numerous clinical populations (Smith et al., <xref rid="B36" ref-type="bibr">1983</xref>; Sainburg et al., <xref rid="B33" ref-type="bibr">1993</xref>; Goble et al., <xref rid="B13" ref-type="bibr">2012</xref>). It is yet unknown, however, to what extent the nervous system adapts to mechanical tendon vibration when applied to both agonist and antagonist muscle pairs. This adaptation processes, presumably mediated through changes in muscle spindle activity, may parallel processes observed in other types of sensory receptors following sustained vibratory stimulation. For example, cutaneous mechanoreceptors become desensitized to suprathershold vibration after prolonged periods of stimulation (Bensmaia et al., <xref rid="B1" ref-type="bibr">2005</xref>) and the time-course of adaptation is faster than perceptual measures observed during psychophysical experimentation (Leung et al., <xref rid="B21" ref-type="bibr">2005</xref>). It is reasonable to predict that muscle spindle adaptation processes may be similar to those observed during cutaneous mechanoreceptor stimulation. If so, this would bring into question whether dual agonist-antagonist tendon vibration may be used in lieu of other known reversible methods to reduce proprioceptive acuity, such as ischemic nerve block.</p><p>The aim of the present study was, therefore, to determine whether short-term (i.e., within 10 trials) proprioceptive adaptation occurs in response to simultaneous agonist-antagonist tendon vibration. This was accomplished by measuring proprioceptive bias (i.e., constant error) and variability (i.e., variable error) in an elbow joint position matching task before, during, and after continuous dual vibration of the biceps and triceps muscle tendons for approximately 10&#x02009;min. Proprioceptive bias and variability were analyzed during early and late periods within each phase of the matching experiment to assess perceptual changes to the dual vibration stimuli over time. We hypothesized that only matching variability would increase during the early period of the dual vibration phase of the experiment and diminish during the late period due to proprioceptive adaptation. Alternatively, based on previous work examining changes in proprioceptive bias following dual vibration (Ronsse et al., <xref rid="B32" ref-type="bibr">2009</xref>), it could have been hypothesized that constant error (i.e., bias) would also change appreciably when the vibration stimuli were applied and removed. In this case, adaptation to the perturbation would be expected such that bias was reduced from early to late matching trials.</p></sec><sec sec-type="materials|methods" id="S2"><title>Materials and Methods</title><sec id="S2-1"><title>Participants</title><p>Informed consent was obtained from a convenience sample of 21 healthy young adults (12 males; 9 females; mean&#x02009;&#x000b1;&#x02009;SD, age&#x02009;=&#x02009;26.6&#x02009;&#x000b1;&#x02009;4.6&#x02009;years) prior to their participation in the study. Exclusion criteria for participants were any self-reported history of upper-limb sensorimotor deficits or cognitive impairment, as well as any tendency toward left handedness measured using the Edinburgh Handedness Inventory (Oldfield, <xref rid="B24" ref-type="bibr">1971</xref>). Procedures for this study were approved the institutional review board at San Diego State University.</p></sec><sec id="S2-2"><title>Experimental setup</title><p>For all elbow position matching trials (see <xref ref-type="sec" rid="S2-3">Experimental Procedure</xref> below) participants were seated in a height adjustable chair with their dominant, right forearm resting in a padded cast on a horizontally rotating robotic manipulandum shaft driven by a programmable torque motor (Kollmorgen servomotors, AKM13D 7000RPM @ 160VDC). The height of the chair was adjusted so that the manipulandum shaft was at the level of the xiphoid process. The axis of rotation of the elbow was aligned with the rotational axis of the manipulandum and elbow angle data were digitized and processed using custom software developed in the LabVIEW environment (National Instruments, TX, USA). To minimize the influence of sensory information from the left arm on right elbow proprioception (Izumizaki et al., <xref rid="B19" ref-type="bibr">2010</xref>), the left arm was positioned comfortably on the participant&#x02019;s lap and was not moved during testing. Participants were randomly assigned to two experimental groups (FLEXED or EXTENDED) to counterbalance any differences due to movement of the arm toward the proprioceptive targets. The starting posture of the elbow to be tested was 90&#x000b0; of flexion in the FLEXED group and 0&#x000b0; flexion (i.e., full extension) in the EXTENDED group.</p><p>The triceps and biceps brachii muscles of the testing arm were fitted with cylindrical electromechanical vibrators secured using an elastic arm sleeve (Nike Men&#x02019;s Arm Sleeve). Specifically, the biceps brachii tendon vibrator was positioned perpendicular to biceps tendon about 1&#x02009;cm proximally from the cubital fossa and the triceps brachii tendon vibrator was positioned perpendicular to the distal triceps tendon about 2&#x02009;cm proximally from the olecranon. The vibrators were calibrated to stimulate the muscle spindles at 80&#x02009;Hz with amplitude of ~1&#x02009;mm. In agreement with previous work, participants reported no proprioceptive illusions when both vibrators operated simultaneously at this frequency (Gilhodes et al., <xref rid="B11" ref-type="bibr">1986</xref>). All participants were blindfolded and wore noise canceling headphones during testing, which respectively served to eliminate any visual and/or auditory feedback regarding movement or position of the elbow joint.</p></sec><sec id="S2-3"><title>Experimental procedure</title><p>Each testing session consisted of three experimental conditions under which proprioceptive matching of elbow joint angles was performed. The first condition was a baseline condition (BASE) where participants completed proprioceptive matching with the vibrators turned off to obtain pre-vibration levels of proprioceptive bias and variability. The second condition (VIB) consisted of proprioceptive matching while adapting to vibration applied to the biceps and triceps tendons simultaneously. The last condition (AFTER) was identical to the first condition (i.e., BASE), and was conducted to assess de-adaptation following the VIB condition.</p><p>Elbow matching was performed according to the following protocol for all conditions (illustrated in Figure <xref ref-type="fig" rid="F1">1</xref>). Prior to testing, participants were instructed to completely relax their arm, to not move throughout testing, and to not interfere with any movements imposed by the manipulandum. The manipulandum was programed to shut off when any low level resistance was detected. Once participants were relaxed, the forearm was moved by the manipulandum to an angular target between 35 and 55&#x000b0; from the initial elbow angle. This target was in the direction of extension for the FLEXED group and in the direction of flexion for the EXTENDED group. When the angular target was reached, the manipulandum stopped moving and the forearm was held stationary for 3&#x02009;s while the participant memorized the location based on proprioceptive information. Next, the arm was moved by the manipulandum to 0&#x000b0; of flexion in the FLEXED group and 90&#x000b0; of flexion in the EXTENDED group. The forearm was held stationary at this position for 3&#x02009;s and then was returned to the initial position. While the forearm was being returned, participants indicated when the memorized angular target was achieved by pressing a mouse button with their contralateral hand. Participants were instructed to press the mouse button when they perceived their elbow angle to be equal to the target position. All arm movements were passive and had a constant velocity of 5&#x000b0;/s. Each trial was approximately 40&#x02009;s in duration, and vibratory stimulus was sustained throughout the duration of all vibration trials (~10&#x02009;min). The duration of the vibratory stimulus aligns with previous work indicating that simultaneous agonist-antagonist vibration increases position uncertainty after 20&#x02009;s of sustained vibratory stimulus (Fuentes et al., <xref rid="B10" ref-type="bibr">2012</xref>).</p><fig id="F1" position="float"><label>Figure 1</label><caption><p><bold>Top-down perspective of the experimental setup and protocol in the FLEXED group</bold>. Position of the forearm initially <bold>(A)</bold>, when moving toward and being held stationary at the angular target <bold>(B)</bold>, when moving and being held stationary at an extended posture <bold>(C)</bold>, and when returning to the initial position to match the angular target <bold>(D)</bold>. The shaded rectangle represents the position of the manipulandum throughout testing.</p></caption><graphic xlink:href="fnhum-08-00896-g001"/></fig><p>Five angular targets were used for testing trials: 35, 40, 45, 50, and 55&#x000b0; from the starting position of the elbow. Angular targets with the same amplitude were not presented subsequently in order to prevent their memorization across trials. The inter-trial delay was 6&#x02013;8&#x02009;s as randomly specified by the testing computer, and the three experimental conditions were presented concurrently with no appreciable delay between conditions. An early adaptation and late adaptation phase, each consisting of five trials, were conducted consecutively for the VIB and AFTER conditions in order to assess short-term adaptation/de-adaptation to the vibration stimuli. Trials were grouped in this manner to account for the effects of different movement amplitudes on position sense errors (Goble, <xref rid="B12" ref-type="bibr">2010</xref>).</p></sec><sec id="S2-4"><title>Data analysis</title><p>Constant error and variable error were used to determine proprioceptive accuracy. Constant error is a measure of proprioceptive bias (i.e., underestimation or overestimation of angular targets) and was calculated by subtracting the matching angle from the target angle in degrees. Negative constant error values indicated undershooting of angular targets and positive values indicated overshooting. Variable error is an accepted measure of trial to trial variability for proprioceptive matching tasks and was determined as the standard deviation of constant error trials across each block in each subject. Higher variable error measures are indicative of increased position uncertainty (Goble, <xref rid="B12" ref-type="bibr">2010</xref>).</p></sec><sec id="S2-5"><title>Statistical analysis</title><p>Constant error and variable error measures obtained during the VIB and AFTER conditions were normalized to mean performance during the BASE condition. Mean constant error was calculated for trials conducted during the BASE condition for each participant. This value was subtracted from mean constant error values obtained during the early and late adaptation phases of the VIB and AFTER conditions. The same procedure was used to normalize variable error. Proprioceptive performance was analyzed in this manner to determine the degree of sensory adaptation across each condition, irrespective of performance during the BASE condition.</p><p>Multiple one-sample <italic>t</italic>-tests were conducted to determine whether constant error and variable error during the early and late adaptation phases of the VIB and AFTER conditions were statistically significantly different from the BASE condition. In all, 95% confidence intervals were calculated for each comparison, and effect sizes were computed as Cohen&#x02019;s <italic>d</italic>. Differences were considered significant with respect to an alpha of <italic>p</italic>&#x02009;&#x0003c;&#x02009;0.05.</p></sec></sec><sec id="S3"><title>Results</title><sec id="S3-6"><title>Changes in proprioceptive bias (i.e., constant error)</title><p>As expected, no participant reported having an illusion of movement about the elbow joint. Despite this, as shown in Figure <xref ref-type="fig" rid="F2">2</xref>, participants had greater undershooting during the early VIB adaptation phase, as shown by negative constant errors that were statistically significantly lower than BASE measures of proprioceptive bias [<italic>t</italic>(20)&#x02009;=&#x02009;&#x02212;2.383, <italic>p</italic>&#x02009;=&#x02009;0.027, 95% CI: &#x02212;4.81, &#x02212;0.32, <italic>d</italic>&#x02009;=&#x02009;&#x02212;1.07]. Mean constant error during the late VIB phase (mean&#x02009;=&#x02009;&#x02212;1.58&#x02009;&#x000b1;&#x02009;0.84) was not statistically significantly different from BASE [<italic>t</italic>(20)&#x02009;=&#x02009;&#x02212;1.89, <italic>p</italic>&#x02009;&#x0003e;&#x02009;0.05, 95% CI: &#x02212;3.33, 0.16, <italic>d</italic>&#x02009;=&#x02009;&#x02212;0.85]. When vibration was removed, mean constant error during the early AFTER phase (mean&#x02009;=&#x02009;2.62&#x02009;&#x000b1;&#x02009;0.74) was statistically significantly greater than BASE measures [<italic>t</italic>(20)&#x02009;=&#x02009;3.56, <italic>p</italic>&#x02009;=&#x02009;0.02, 95% CI: 1.09, 4.16, <italic>d</italic>&#x02009;=&#x02009;1.59], indicating greater overshooting. Mean constant error during the late AFTER phase (mean&#x02009;=&#x02009;1.93&#x02009;&#x000b1;&#x02009;0.73) was also statistically significantly greater [<italic>t</italic>(20)&#x02009;=&#x02009;2.66, <italic>p</italic>&#x02009;=&#x02009;0.15, 95% CI: 0.42, 3.45, <italic>d</italic>&#x02009;=&#x02009;1.19], although the size of difference was closer to baseline.</p><fig id="F2" position="float"><label>Figure 2</label><caption><p><bold>Constant error results during the VIB and AFTER conditions</bold>. Significant undershoot errors were observed during the early adaptation phase of the VIB condition. When the vibration stimuli were removed, participants committed overshoot errors during both early and late adaptation phases.</p></caption><graphic xlink:href="fnhum-08-00896-g002"/></fig></sec><sec id="S3-7"><title>Changes in matching variability (i.e., variable error)</title><p>Variable error (mean&#x02009;=&#x02009;0.40&#x02009;&#x000b1;&#x02009;0.58), shown in Figure <xref ref-type="fig" rid="F3">3</xref>, was not statistically significantly different from baseline when vibration was applied during both the early VIB [<italic>t</italic>(20)&#x02009;=&#x02009;0.68, <italic>p</italic>&#x02009;&#x0003e;&#x02009;0.05, 95% CI: &#x02212;0.81, 1.61, <italic>d</italic>&#x02009;=&#x02009;0.31] and late VIB (mean&#x02009;=&#x02009;0.36&#x02009;&#x000b1;&#x02009;0.50) [<italic>t</italic>(20)&#x02009;=&#x02009;0.71, <italic>p</italic>&#x02009;&#x0003e;&#x02009;0.05, 95% CI: &#x02212;0.69, 1.41, <italic>d</italic>&#x02009;=&#x02009;0.32] phases. Interestingly, when vibration was removed, variable error (mean&#x02009;=&#x02009;1.49&#x02009;&#x000b1;&#x02009;0.63) increased significantly during the early AFTER phase [<italic>t</italic>(20)&#x02009;=&#x02009;2.36, <italic>p</italic>&#x02009;=&#x02009;0.029, 95% CI: 0.17, 2.80, <italic>d</italic>&#x02009;=&#x02009;1.05]. Variable error (mean&#x02009;=&#x02009;0.98&#x02009;&#x000b1;&#x02009;0.41) remained statistically significantly greater than baseline [<italic>t</italic>(20)&#x02009;=&#x02009;2.36, <italic>p</italic>&#x02009;=&#x02009;0.028, 95% CI: 0.12, 1.84, <italic>d</italic>&#x02009;=&#x02009;1.06] in the late AFTER phase, and was closer in magnitude to baseline measurements.</p><fig id="F3" position="float"><label>Figure 3</label><caption><p><bold>Variable error results during the VIB and AFTER conditions</bold>. No appreciable increase in variable error was observed during the VIB condition. In contrast, matching variability was significantly elevated following removal of the vibration stimuli. The magnitude of variable errors diminished from early to late adaptation phases in the AFTER condition.</p></caption><graphic xlink:href="fnhum-08-00896-g003"/></fig></sec></sec><sec sec-type="discussion" id="S4"><title>Discussion</title><p>Our sense of limb position and movement (i.e., proprioception) depends on afferent information conveyed to the central nervous system by muscle spindles, known to be sensitive to tendon vibration (Burke et al., <xref rid="B3" ref-type="bibr">1976</xref>; Roll et al., <xref rid="B31" ref-type="bibr">1989</xref>). Several investigations suggest that vibration degrades the quality of this afferent information, causing joint proprioception to be disrupted (Roll et al., <xref rid="B31" ref-type="bibr">1989</xref>; Bock et al., <xref rid="B2" ref-type="bibr">2007</xref>). To investigate how the proprioceptive system adapts in the short-term to perceptual perturbations caused by vibration, we conducted an elbow joint position matching task on healthy young adults while applying vibration to the distal tendons of the biceps and triceps brachii muscles. The vibration stimuli caused participants to commit undershoot errors when replicating reference joint angles. The magnitude of undershoot errors decreased over the course of the task. Following completion of the task, we removed the vibration stimuli and conducted a second joint position matching task. Matching performance on the second matching task revealed an aftereffect consisting of overshoot errors and a significant increase in matching variability.</p><p>Increased matching variability immediately following the removal of the vibration stimuli, but not when the stimuli were being applied, may be best interpreted in light of known changes to muscle spindles following prolonged vibratory stimulation. Ribot-Ciscar et al. (<xref rid="B28" ref-type="bibr">1998</xref>) investigated postvibration effects on the firing properties of a small population of muscle spindles located in the ankle dorsiflexor muscles. The spontaneous firing rate of most muscle spindles in the population decreased, while the rate of a subpopulation increased. Additionally, when the ankle was passively stretched, the mean firing rate of the spindle population was diminished and highly variable compared to previbratory measurements. These results parallel those of the present study, where the increase in matching variability that we observed was likely due to an increase in the variability of the spindle population&#x02019;s response to stretch. Changes in the variability of firing properties of muscle spindles, however, may be dependent on the duration of the proceeding vibration stimuli. Fuentes et al. (<xref rid="B10" ref-type="bibr">2012</xref>) found that variability in the perceived angular position of the wrist joint increased significantly only after 20&#x02009;s of simultaneous agonist-antagonist tendon vibration. The VIB phase of the present study consisted of ten trials, with each trial lasting approximately 40&#x02009;s. This effect, therefore, appears to only emerge after prolonged periods of simultaneous vibration. Collectively, these results provide additional evidence for the theory that sensations of limb position and movement depend on responses from entire spindle populations (i.e., population coding) (Ribot-Ciscar and Roll, <xref rid="B27" ref-type="bibr">1998</xref>; Cordo et al., <xref rid="B7" ref-type="bibr">2002</xref>).</p><p>In light of the observed increase in matching variability following vibration, we hypothesize that the central nervous system cannot integrate proprioceptive information proficiently when the neuronal variability of the muscle spindle population exceeds a certain threshold level. We define this threshold as the amount of neuronal variability that will cause the population to have a widely spread response distribution. Our hypothesis is not unfounded given the roll of the fusimotor system in controlling the stretch sensitivity of muscle spindles (Hulliger, <xref rid="B18" ref-type="bibr">1984</xref>). Further, the proposed hypothesis is supported by our observation that matching variability did not increase when the vibration stimuli were applied. Previous work has demonstrated that muscle spindles respond harmonically to vibratory stimulus that is within the 80&#x02013;120&#x02009;Hz range (Roll and Vedel, <xref rid="B30" ref-type="bibr">1982</xref>). Since the vibrators in the present study operated invariantly at this frequency, it is unlikely that the vibration stimuli were capable of increasing the variability of spindle responses beyond the proposed threshold. Rather, vibrators designed to operate stochastically with sufficient amplitude may achieve this aim. Investigating this would require elements from information theory and is beyond the scope of this manuscript.</p><p>Simultaneous vibration of agonist and antagonist muscles caused participants to commit undershoot errors when replicating joint angles. Similar alterations of proprioceptive bias have been revealed when agonist or antagonist muscle groups are vibrated alone (Capaday and Cooke, <xref rid="B4" ref-type="bibr">1981</xref>, <xref rid="B5" ref-type="bibr">1983</xref>; Cordo et al., <xref rid="B6" ref-type="bibr">1995</xref>). Since we vibrated the biceps and triceps brachii simultaneously, how might these previous investigations corroborate with the results of the present study? Illusions of joint movement can be elicited if the vibration frequency applied to antagonist muscle pairs is different (Gilhodes et al., <xref rid="B11" ref-type="bibr">1986</xref>). The perceived velocity of these illusory movements is proportional to the difference in frequency between vibrators (Roll and Vedel, <xref rid="B30" ref-type="bibr">1982</xref>; Ribot-Ciscar and Roll, <xref rid="B27" ref-type="bibr">1998</xref>). If the same vibration frequency is applied, vibratory afferent information from the opposing muscle groups is negated when integrated by the central nervous system. Therefore, one would expect that agonist-antagonist muscle vibration would cause no appreciable change in joint position bias.</p><p>The changes in proprioceptive bias we observed in the present study may be accounted for by the joint position matching task that we used. We passively rotated the elbow in opposite directions when the reference angle was presented and reproduced. This caused participants to rely on different sources of afferent information when memorizing and matching reference angles. For example, in the FLEXED condition, the reference angle was memorized using afferent feedback from the elbow flexors and reproduced using elbow extensor feedback. The weighting of afferent information from each group of muscle may have been different, resulting in changes in proprioceptive bias when the muscles were vibrated (Mel&#x02019;nichouk et al., <xref rid="B22" ref-type="bibr">2007</xref>). Also, passive joint rotations reduced the effects of alpha-gamma coactivation on intrafusal fiber slack. This intrafusal slack could cause the firing properties of muscle spindles in the shortening muscles to be dependent on the history of the previously imposed stretch (Proske et al., <xref rid="B26" ref-type="bibr">1992</xref>; Kostyukov and Cherkassky, <xref rid="B20" ref-type="bibr">1997</xref>).</p><p>The magnitude of undershooting errors decreased over the duration of the vibration stimuli. Similarly, participants committed overshoot errors when the vibration stimuli were removed. These findings corroborate with results obtained by Seizova-Cajic et al. (<xref rid="B34" ref-type="bibr">2007</xref>) who found that movement illusions wavered when vibratory stimulation was applied to the elbow flexors over an extended period of time. Overshoot errors following vibration have been reported previously as well (Rogers et al., <xref rid="B29" ref-type="bibr">1985</xref>; Gregory et al., <xref rid="B15" ref-type="bibr">1988</xref>). Collectively, these results suggest that the proprioceptive system adapts to sustained afferent feedback caused by tendon vibration. The purpose of proprioceptive adaptation under natural conditions is likely analogous to other sensory systems: to maintain the sensitivity of the sensory system to changes in the surrounding environment (Helson, <xref rid="B17" ref-type="bibr">1948</xref>). We propose that proprioceptive adaptation ensures that the central nervous system is capable of perceiving changes in limb position and movement when proprioceptors are continuously activated by self-generated movements. Supraspinal (Ebner and Pasalar, <xref rid="B9" ref-type="bibr">2008</xref>; Mulliken et al., <xref rid="B23" ref-type="bibr">2008</xref>) and intraspinal (Hantman and Jessell, <xref rid="B16" ref-type="bibr">2010</xref>) mechanisms involved in the integration of corollary and sensory feedback likely play a crucial role in allowing the proprioceptive system to adapt over time.</p><p>There are several limitations in the present study worthy of recognition. First, we did not use electromyography to monitor activity of the biceps and triceps brachii during testing. We therefore can only speculate as to whether participants were able to completely relax. Further, since muscle activity was not monitored, we cannot discount the possibility that the tonic vibration reflex may have influenced our findings. Considerable effort was made to ensure that participants remained relaxed throughout the testing period, and the manipulandum was programed to operate only when it detected minimal resistance to imposed movements. Other sensory receptors, such as cutaneous mechanoreceptors, may have been influenced by their respective adaptive processes (Bensmaia et al., <xref rid="B1" ref-type="bibr">2005</xref>; Leung et al., <xref rid="B21" ref-type="bibr">2005</xref>). Therefore, this study does not explicitly describe the time-course of adaptation in muscle spindles following vibration. Rather, it provides evidence for adaptation processes that may occur throughout the proprioceptive system.</p><p>In conclusion, the results of the present study bring into question the use of dual agonist-antagonist tendon vibration to degrade proprioception about an adjacent joint. While previous investigations have used vibration to temporarily degrade proprioceptive feedback (Bock et al., <xref rid="B2" ref-type="bibr">2007</xref>; Vidoni and Boyd, <xref rid="B37" ref-type="bibr">2008</xref>; Ronsse et al., <xref rid="B32" ref-type="bibr">2009</xref>), our results suggest this method elicits changes in proprioceptive bias that diminish with time and has little effect on variable error, as would be expected in the case of increased proprioceptive noise. Rather, removal of the vibration stimulus caused the most powerful increase in limb position uncertainty. We hypothesize that this may be due to limitations in the capacity of the central nervous system to integrate sensory input that is predominantly stochastic. Future investigations of the time-course underlying increases in limb position uncertainty following sustained tendon vibration are warranted.</p></sec><sec id="S5"><title>Conflict of Interest Statement</title><p>The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.</p></sec>

Neck Proprioception Shapes Body Orientation and Perception of Motion

<p>This review article deals with some effects of neck muscle proprioception on human balance, gait trajectory, subjective straight-ahead (SSA), and self-motion perception. These effects are easily observed during neck muscle vibration, a strong stimulus for the spindle primary afferent fibers. We first remind the early findings on human balance, gait trajectory, SSA, induced by limb, and neck muscle vibration. Then, more recent findings on self-motion perception of vestibular origin are described. The use of a vestibular asymmetric yaw-rotation stimulus for emphasizing the proprioceptive modulation of motion perception from the neck is mentioned. In addition, an attempt has been made to conjointly discuss the effects of unilateral neck proprioception on motion perception, SSA, and walking trajectory. Neck vibration also induces persistent aftereffects on the SSA and on self-motion perception of vestibular origin. These perceptive effects depend on intensity, duration, side of the conditioning vibratory stimulation, and on muscle status. These effects can be maintained for hours when prolonged high-frequency vibration is superimposed on muscle contraction. Overall, this brief outline emphasizes the contribution of neck muscle inflow to the construction and fine-tuning of perception of body orientation and motion. Furthermore, it indicates that tonic neck-proprioceptive input may induce persistent influences on the subject&#x02019;s mental representation of space. These plastic changes might adapt motion sensitiveness to lasting or permanent head positional or motor changes.</p>

<contrib contrib-type="author"><name><surname>Pettorossi</surname><given-names>Vito Enrico</given-names></name><xref ref-type="aff" rid="aff1">¹</xref><uri xlink:type="simple" xlink:href="http://frontiersin.org/people/u/191029"/></contrib><contrib contrib-type="author"><name><surname>Schieppati</surname><given-names>Marco</given-names></name><xref ref-type="aff" rid="aff2">²</xref><xref ref-type="aff" rid="aff3">³</xref><xref ref-type="corresp" rid="cor1">*</xref><uri xlink:type="simple" xlink:href="http://frontiersin.org/people/u/165307"/></contrib>

Frontiers in Human Neuroscience

<sec sec-type="intro" id="S1"><title>Introduction</title><p>To many aged professors of physiology (and young students as well), the term proprioception promptly calls to mind the tendon-tap reflex, i.e., the &#x0201c;monosynaptic&#x0201d; reflex elicited by a tap onto the patellar or Achilles&#x02019; tendon and the consequent leg extension and foot plantarflexion, respectively. This phenomenon has been given such straightforward explanation (Ia spindle afferent fibers &#x02013; motoneurons &#x02013; homonymous muscle contraction) that no-one would have thought to doubt on the vital role of proprioceptive reflexes in all aspects of human movement, and in particular in the control of quiet upright stance or gait. Some doubts, however, should emerge from the simple observation that people devoid of deep muscle reflexes, e.g., patients with Holmes-Adie&#x02019;s syndrome (Adie, <xref rid="B1" ref-type="bibr">1932</xref>) or Charcot-Marie-Tooth type Ia (Nardone et al., <xref rid="B120" ref-type="bibr">2000</xref>), can stand up and walk (Mazzaro et al., <xref rid="B109" ref-type="bibr">2005</xref>) almost as well as if their monosynaptic reflexes were present and brisk. Still more reservations should arise from observing the decreased excitability of the Achilles&#x02019; reflex during stance (Bove et al., <xref rid="B23" ref-type="bibr">2006a</xref>) and locomotion (Crenna and Frigo, <xref rid="B41" ref-type="bibr">1987</xref>) compared to laying down.</p><p>The over simplistic attitude was modified when the function of the spindle intrafusal secondary endings and group II afferent fibers, revealed long ago (Matthews, <xref rid="B106" ref-type="bibr">1972</xref>, <xref rid="B105" ref-type="bibr">2006</xref>), was re-evaluated thanks to reports emphasizing their role in stance and locomotion, based on findings from normal subjects and neuropathic patients (Corna et al., <xref rid="B38" ref-type="bibr">1995</xref>; Mazzaro et al., <xref rid="B109" ref-type="bibr">2005</xref>). The group II fibers are about as numerous as the Ia fibers (Hunt, <xref rid="B68" ref-type="bibr">1954</xref>) and, in spite of their conduction velocity being about half that of the former (about 30 wrt 60&#x02009;m/s, in man), they play a paramount role in the afferent control of quiet and perturbed stance (Schieppati and Nardone, <xref rid="B149" ref-type="bibr">1997</xref>, <xref rid="B150" ref-type="bibr">1999</xref>; Nardone and Schieppati, <xref rid="B119" ref-type="bibr">1998</xref>; Simonetta-Moreau et al., <xref rid="B161" ref-type="bibr">1999</xref>; Bove et al., <xref rid="B22" ref-type="bibr">2003</xref>) and of locomotion (Mazzaro et al., <xref rid="B108" ref-type="bibr">2006</xref>).</p><p>Those findings have led to somewhat re-dimension the role of the spindle primary afferent fibers in the control of posture and gait. If their inflow is down-weighted during stance, in favor of their companion group II fibers, which are more appropriate for transducing slow changes in muscle length (Matthews, <xref rid="B104" ref-type="bibr">1977</xref>) and have more positive reflex effects (e.g., production of a larger EMG burst in the postural muscles, diverging excitation to both legs&#x02019; motoneurons) (Corna et al., <xref rid="B37" ref-type="bibr">1996</xref>; Schieppati and Nardone, <xref rid="B149" ref-type="bibr">1997</xref>), what is left to the primary endings and Ia-afferent fibers to do?</p><p>There are excellent books chapters and review papers that place proprioception in the context of the global control of movement by the human nervous system (e.g., Prochazka et al., <xref rid="B134" ref-type="bibr">2000</xref>; Pierrot-Deseilligny and Burke, <xref rid="B131" ref-type="bibr">2005</xref>). Some address the interaction of proprioceptive information and its modulation with the operation of the spinal circuits (e.g., Windhorst, <xref rid="B180" ref-type="bibr">2007</xref>). Others address the sensing of limb position and limb movement, originating in the spindles, emphasizing the existence of two separate senses, and point to the contribution of centrally generated motor command signals (e.g., Proske and Gandevia, <xref rid="B135" ref-type="bibr">2009</xref>). In a more recent review paper, Proske and Gandevia (<xref rid="B136" ref-type="bibr">2012</xref>) expanded their target to include the senses of position and movement of our limbs and trunk, the sense of effort, the sense of force, and the sense of heaviness, and the effects of exercise and aging on proprioceptive sense. The present short review intends to summarize recent findings on the effects of activation of the Ia spindle afferent fibers, with specific reference to body orientation in space during stance and locomotion and to perception of motion in space. In particular, attention is focused on the neck proprioception and on its activation by muscle vibration or contraction, considering both immediate and long-term effects.</p></sec><sec id="S2"><title>Muscle Vibration is a Powerful Tool for Activating the Primary Endings of Muscle Spindles</title><p>In spite of the wealth of knowledge on proprioception and on the role of the primary afferent spindle fibers, novel information on less obvious but not less important roles, is adding up continuously, also thanks to the use of time honored experimental procedures, like lesion or stimulation. As to the former, mother nature helps by providing us with appropriate models featuring loss of large-diameter sensory fiber function (e.g., peripheral neuropathy, as briefly mentioned above). As to the latter, a rough though harmless and selective way of activating the primary spindle endings is muscle vibration (Bianconi and Van Der Meulen, <xref rid="B11" ref-type="bibr">1963</xref>; Burke et al., <xref rid="B28" ref-type="bibr">1976a</xref>,<xref rid="B29" ref-type="bibr">b</xref>; Roll et al., <xref rid="B141" ref-type="bibr">1989a</xref>).</p><p>Vibration (~100&#x02009;Hz) is a potent stimulus for the primary endings of the muscle spindle, less so for the secondary endings and the Golgi tendon organs (Roll et al., <xref rid="B141" ref-type="bibr">1989a</xref>). For instance, vibration can induce a tonic contraction of the vibrated muscle (De Gail et al., <xref rid="B44" ref-type="bibr">1966</xref>; Schieppati and Crenna, <xref rid="B148" ref-type="bibr">1984</xref>), but, in addition to segmental responses, it also produces global effects. In standing subjects, leg muscle vibration elicits illusions of position that provoke postural reactions dependent on the new illusory postural set. Achilles&#x02019; tendon vibration while standing produces an inclination of the body backwards (Eklund, <xref rid="B50" ref-type="bibr">1972</xref>; Thompson et al., <xref rid="B168" ref-type="bibr">2007</xref>). This may be an automatic postural reaction to the &#x0201c;illusion&#x0201d; of forward displacement (or of triceps lengthening, as conveyed by the vibration-induced increased discharge of Ia fibers), since restrained subjects adjusted their body backward via a joystick when allowed to do so (Ceyte et al., <xref rid="B31" ref-type="bibr">2006</xref>). These and other studies on this topic have led to the proposition that our sense of verticality may depend to a large extent on proprioception (Hlavacka et al., <xref rid="B67" ref-type="bibr">1992</xref>; Barbieri et al., <xref rid="B5" ref-type="bibr">2008</xref>; Barra and P&#x000e9;rennou, <xref rid="B7" ref-type="bibr">2013</xref>).</p></sec><sec id="S3"><title>The Neck is the Functional Link between Head and Body</title><p>Proprioception of the neck, as also of the axial muscles, has a powerful body-orienting effect during quiet stance and locomotion. Such a peculiar influence must have evolved with the neck itself and with the need to counteract gravity, when our ancestors emerged from the water (Jouffroy, <xref rid="B75" ref-type="bibr">1992</xref>). Fish have no neck, and the axial muscles have basically a medio-lateral action during swimming. For their orientation in space, the function of the lateral-line system is enough (Webb, <xref rid="B175" ref-type="bibr">1989</xref>). With terrestrial life and erect bipedal posture and heavy, mobile head, the interaction of neck, and trunk proprioception with the vestibular sense has reached a highly developed grade. Since vestibular signals cannot distinguish whether the head or the whole body is moving when the head moves on a stationary trunk, the neck-proprioceptive input provides the necessary information about head movements relative to the trunk. Accordingly, neck muscles are richly endowed with spindles, which are highly sensitive to head yaw rotation (Chan et al., <xref rid="B32" ref-type="bibr">1987</xref>). With the development of the neck muscles and their function (head yaw rotation, roll inclination, flexion, and extension), a unique mode of control has arisen, whereby rotation is produced by activation of the sternocleidomastoideus (SCM) muscle of one side (opposite to the direction of rotation) and of the dorsal neck (DN) muscle group of the same side as the rotation. For instance, during voluntary head turning to the left, right SCM, and left DN are agonists, as are both SCM during head flexion, and both DN muscles during the usual antigravity tonic action and voluntary head extension (Mazzini and Schieppati, <xref rid="B110" ref-type="bibr">1992</xref>). Such control must rest on the concurrent operation of separate ipsi- and contralateral descending pathways (Zangemeister et al., <xref rid="B184" ref-type="bibr">1982</xref>; Mastaglia et al., <xref rid="B103" ref-type="bibr">1986</xref>; Beimborn and Morrissey, <xref rid="B8" ref-type="bibr">1988</xref>; Gandevia and Applegate, <xref rid="B57" ref-type="bibr">1988</xref>; Conley et al., <xref rid="B36" ref-type="bibr">1995</xref>; Mayoux-Benhamou et al., <xref rid="B107" ref-type="bibr">1997</xref>).</p><p>The functional relevance of the neck as a crucial segment in the head and trunk relationship is attested by the strength of the cervico-collic reflex and the vestibulo-collic reflexes. Activation of proprioceptors in the neck evokes cervico-collic reflex, which works in combination with vestibulo-collic reflex for the head stability and body posture. Signals interact downstream at the level of the spinal cord and upstream at the level of the vestibular and reticular nuclei (Pompeiano, <xref rid="B132" ref-type="bibr">1979</xref>; Wilson and Peterson, <xref rid="B179" ref-type="bibr">1988</xref>). Animal data are available on the operation of these reflexes in the pitch, yaw, and roll plane (Peterson et al., <xref rid="B125" ref-type="bibr">1985</xref>; Dutia and Price, <xref rid="B49" ref-type="bibr">1987</xref>; Zennou-Azogui et al., <xref rid="B185" ref-type="bibr">1993</xref>) and on the effect of limb proprioception on these reflexes (Rosenberg et al., <xref rid="B144" ref-type="bibr">1980</xref>). The two reflexes appear to behave approximately linearly, both individually and in combination (Peterson et al., <xref rid="B125" ref-type="bibr">1985</xref>), whereby the dynamic of these reflexes and their spatial organization assure a correct response to prevent oscillation of the head on a stationary body. In human, Guitton et al. (<xref rid="B63" ref-type="bibr">1986</xref>) found that the contribution to a head stabilization task of the short-latency cervico-collic and vestibulo-collic reflexes may be unimportant, while longer-latency effects can be as powerful as vision. We found no data specifically concerning the effects of neck muscle vibration on cervico-collic and vestibulo-collic reflexes in man. Likely, these reflexes could modulate perception and orientation by way of their effects on head-in-space and head-on-trunk posture. In turn, these effects might be modulated by vibration. Research is needed to get insight in the interaction between reflexes and motion perception.</p><sec id="S3-1"><title>Neck-proprioceptive influence on body orientation</title><p>Simple slow head turns can result in lateral displacements of the body&#x02019;s center of mass toward to the &#x0201c;occipital&#x0201d; side, particularly so in the presence of a tonic level of spindle discharge from leg muscles (Gurfinkel et al., <xref rid="B64" ref-type="bibr">1995</xref>). Continuous vibration of the DN muscles, bilaterally, produces a reactive response in the sagittal plane consisting in a <italic>forward</italic> inclination of the body (Lekhel et al., <xref rid="B89" ref-type="bibr">1988</xref>; Kavounoudias et al., <xref rid="B81" ref-type="bibr">1999</xref>; Ivanenko et al., <xref rid="B71" ref-type="bibr">2000</xref>). Various explanations for this may be entertained. Robust spindle discharge normally ensues when gamma-MNs are active, as during a tonic voluntary neck extension. Illusion thereof would produce forward body inclination, in order to align the head with the vertical again. Other explanations are plausible. Ivanenko et al. (<xref rid="B71" ref-type="bibr">2000</xref>) suggested that, since the vestibular input is constant, the head may well be considered stationary in space and the neck flexed (as if the DN muscles were elongated) on the trunk inclined backwards. The subjects would react to the illusion of the body center of mass being displaced forward, and would be pressed to propel the body forward. Haptic supplementation offered by a touch on a firm surface (Bove et al., <xref rid="B24" ref-type="bibr">2006b</xref>) and vision (Bove et al., <xref rid="B21" ref-type="bibr">2009</xref>) modulate the effects of neck vibration on posture in amplitude and time, indicating a key role of multiple sensorimotor integration for body orientation in space. It could be argued that our nervous system weights the proprioceptive inflow according to its priorities, which may lead to &#x0201c;compensatory reactions&#x0201d; aimed at maintaining the task variable stationary (Lockhart and Ting, <xref rid="B91" ref-type="bibr">2007</xref>).</p><p>Neck vibration also influences the perception of body position in the yaw plane, without necessarily producing postural changes in response to equilibrium challenge. Unilateral vibration influences the subjective straight-ahead (SSA) perception, inducing a disparity between subjective perception and objective position of the body midline, and determines an illusory movement of the head and of the visual target (Biguer et al., <xref rid="B12" ref-type="bibr">1988</xref>; Roll et al., <xref rid="B142" ref-type="bibr">1989b</xref>; Taylor and McCloskey, <xref rid="B167" ref-type="bibr">1991</xref>; Karnath et al., <xref rid="B77" ref-type="bibr">1993</xref>; Lekhel et al., <xref rid="B88" ref-type="bibr">1997</xref>; Ceyte et al., <xref rid="B31" ref-type="bibr">2006</xref>). The SSA (detected by asking the subject to point to a remembered visual target presented before the vibration or to point to their own nose) and the visual target representation are displaced during neck vibration (Taylor and McCloskey, <xref rid="B167" ref-type="bibr">1991</xref>; Seizova-Cajic et al., <xref rid="B158" ref-type="bibr">2006</xref>). The SSA moves toward the same side as the vibrated DN muscle, while the visual target moves toward the opposite side (Biguer et al., <xref rid="B12" ref-type="bibr">1988</xref>). The shifts of SSA and visual target are in the opposite direction when the SCM is vibrated. This is consistent as DN and contralateral SCM act as a synergistic pair during head rotations. Therefore, illusory head movement correlates with the illusory target movement (Taylor and McCloskey, <xref rid="B167" ref-type="bibr">1991</xref>) during neck muscle vibration. Conversely, eye movements and illusory perception do not correlate, since vibration induces little or no change in eye movements (Lackner and Levine, <xref rid="B85" ref-type="bibr">1979</xref>; Seizova-Cajic et al., <xref rid="B158" ref-type="bibr">2006</xref>). Neck muscle vibration seems to influence primarily the relative position of the body with respect to space creating an illusory head deviation. This illusory head deviation corresponds to the real head deviation induced by contraction of the vibrated muscles. Emphasis on the neck proprioception is dictated by the topic of the present review; however, we would note that, among other things, transcutaneous electrical nerve stimulation in the region of the neck (P&#x000e9;rennou et al., <xref rid="B124" ref-type="bibr">2001</xref>), wedge-prism exposure (Rode et al., <xref rid="B140" ref-type="bibr">2003</xref>), or podokinetic stimulation (Scott et al., <xref rid="B155" ref-type="bibr">2011</xref>) can also affect the SSA, possibly through activation of brain functions related to multisensory integration that help restore the body proprioceptive representation.</p></sec><sec id="S3-2"><title>Neck proprioception and self-motion perception</title><p>Self-motion perception depends on the integration of sensory signals about body movement from vestibular, visual, proprioceptive, auditory, and kinesthetic signals. Can neck proprioception interfere with the conscious perception of self-motion? Several studies have examined the convergence and the interaction between neck and vestibular input at level of the vestibular nuclei (Anastasopoulos and Mergner, <xref rid="B3" ref-type="bibr">1982</xref>; Manzoni, <xref rid="B99" ref-type="bibr">1988</xref>), the cerebellum (Manzoni et al., <xref rid="B100" ref-type="bibr">1998</xref>; Brooks and Cullen, <xref rid="B26" ref-type="bibr">2009</xref>; Luan et al., <xref rid="B93" ref-type="bibr">2013</xref>), and the parieto-insular vestibular cortex (Shinder and Newlands, <xref rid="B160" ref-type="bibr">2014</xref>). In addition, the dynamic interactions of neck proprioceptive and vestibular inputs in the perception of body movement have been systematically described (Mergner et al., <xref rid="B115" ref-type="bibr">1991</xref>, <xref rid="B113" ref-type="bibr">1992</xref>, <xref rid="B114" ref-type="bibr">1998</xref>; Mergner and Rosemeier, <xref rid="B112" ref-type="bibr">1998</xref>), and a linear summation mechanism between the two signals has been proposed (Karnath, <xref rid="B76" ref-type="bibr">1994</xref>; Mergner and Rosemeier, <xref rid="B112" ref-type="bibr">1998</xref>; Bottini et al., <xref rid="B16" ref-type="bibr">2001</xref>). While combinations of dynamic vestibular and neck proprioception activity have been widely analyzed, little information is available on the influence of tonic, prolonged proprioceptive signals on the self-motion perception of vestibular origin (Cullen, <xref rid="B42" ref-type="bibr">2012</xref>, <xref rid="B43" ref-type="bibr">2014</xref>; Medrea and Cullen, <xref rid="B111" ref-type="bibr">2013</xref>).</p><sec id="S3-2-1"><title>Neck muscle vibration modulates self-motion perception of vestibular origin</title><p>The brain continuously keeps track of the body movement, in order to establish the instantaneous spatial relationship between self and the world. In man, motion perception can be estimated by having standing subjects oscillate in the yaw plane in the dark, and tracking with a pointer the remembered position of an earth-fixed visual target. Panichi et al. (<xref rid="B123" ref-type="bibr">2011</xref>) use a modified version of this protocol, whereby the cyclic left&#x02013;right rotation was of equal amplitude but had asymmetric velocity. This stimulus causes a strongly biased perception of movement due to the vestibular dynamic properties (Panichi et al., <xref rid="B123" ref-type="bibr">2011</xref>; Pettorossi et al., <xref rid="B128" ref-type="bibr">2013a</xref>), since vestibular signals promptly indicate fast head movements, while they are poor at sensing very slow movements (Goldberg and Fernandez, <xref rid="B59" ref-type="bibr">1971</xref>; Kolev et al., <xref rid="B83" ref-type="bibr">1996</xref>; Massot et al., <xref rid="B102" ref-type="bibr">1999</xref>; Valko et al., <xref rid="B170" ref-type="bibr">2012</xref>; Tremblay et al., <xref rid="B169" ref-type="bibr">2013</xref>). By continuing the asymmetric vestibular stimulation, the bias in motion perception progressively increased, whereby the gain of the tracking response gradually and continuously increased during the fast rotation cycle and decreased during the slow rotation cycle (Pettorossi et al., <xref rid="B128" ref-type="bibr">2013a</xref>). This way of stimulating the vestibular system proved to be appropriate for showing the neck influence on self-motion perception, since symmetric whole-body rotation was not able to disclose sizeable and unambiguous effects of superimposed neck muscle vibration.</p><p>The motion perception bias produced by the asymmetric vestibular stimulation was strongly modified by unilateral neck muscle vibration or contraction or both (Panichi et al., <xref rid="B123" ref-type="bibr">2011</xref>). These maneuvers doubled or annulled the bias, depending on the side of vibration or direction of head active deviation. Vibration of the DN or SCM muscle with the head in primary position differentially influenced the perceived rotation during asymmetric oscillation, coherently with their effect on head yaw voluntary rotation (Figure <xref ref-type="fig" rid="F1">1</xref>). The sign of the influence on the perceptive &#x0201c;bias&#x0201d; was opposite, while its amplitude was comparable. For instance, vibration of the left SCM produced an exaggerated perception of body rotation to the right (the sense of the fast cycle of whole-body rotation), while vibration of the left DN muscles almost canceled the bias in the vestibular-induced perception of rotation. Therefore, by enhancing the spindle firing from the muscles that turn the head, say, to the right, the sensitivity of the brain to whole-body rotation to the right was enhanced. Tonic active (but not passive) head deviation superimposed to the asymmetric whole-body oscillation also enhanced movement perception when the head was turned toward the side of the fast rotation and decreased it with opposite deviation (toward the site of the slow rotation) (Figure <xref ref-type="fig" rid="F1">1</xref>).</p><fig id="F1" position="float"><label>Figure 1</label><caption><p><bold>Influence of unilateral neck muscle vibration on the self-motion perception elicited by whole-body rotation</bold>. <bold>(A)</bold> Representation of the experimental setting. Subjects stood on a computer-controlled rotating platform in the dark (rotating platform, PL; pointer, P; holder, H). Circular lines indicate the fast platform rotation to the right (solid) and the slow rotation to left (dashed). Subjects were asked to manually track with the pointer (P) a remembered light spot (diameter 1&#x02009;cm) presented before the test in front of them. Four consecutive asymmetric oscillation cycles were administered, made of two sinusoidal half cycles of equal amplitude but different frequency. <bold>(B,C)</bold> Tracking task recording during the vestibular stimulation (top, fast rotation to the right) and the effect of the conditioning maneuvers. The traces of platform oscillation (PL) and tracking position (TP) are shown during four cycles of asymmetric rotation. The remembered target position is progressively shifted toward the side of slow rotation (left) due to the properties of the vestibular system (NV, no vibration; 0&#x000b0;H-T, head in the primary position). Right Splenius Capitis (SC) muscle vibration <bold>(B)</bold> or maintaining the head rotated to right <bold>(C)</bold> induced an enhancement of the motion perception to the right during the asymmetric rotation, as shown by the displacement of the target representation. Conversely, left SC vibration and head active rotation to the left reduced motion perception and diminished the error in the target representation. In <bold>(B)</bold>, in the inset, is reported the schematic drawing of the platform rotation during 100&#x02009;Hz frequency, 0.8&#x02009;mm amplitude vibration (V: vibrator device) of right and left Splenius Capitis (rSc, lSC). In <bold>(C)</bold>, the effects of the three conditions of head-trunk positions (H-T angle) are reported: 0&#x000b0;, 45&#x000b0; to the right (r) and to the left (l) [adapted from Panichi et al. (<xref rid="B123" ref-type="bibr">2011</xref>)].</p></caption><graphic xlink:href="fnhum-08-00895-g001"/></fig><p>Therefore, the vestibular-evoked perception of body rotation is enhanced by neck-proprioceptive input as a function of the muscles&#x02019; action (turning the head, or trunk, to the right or to the left) rather than of their anatomical position (right or left of the midline). This effect may be useful for increasing the gain of the perception of motion in the presence of intense active rotation of the body, when the body movement must follow the direction in which the head turns, a condition that may require a superior perception for a better performance of the goal-directed movement (Panichi et al., <xref rid="B123" ref-type="bibr">2011</xref>).</p></sec><sec id="S3-2-2"><title>Does vibration mimic passive muscle lengthening or muscle contraction?</title><p>Vibration at the appropriate frequency (80&#x02013;120&#x02009;Hz) is as good a stimulus for the spindles of the neck muscles as it is for other body muscles. In spite of the uniqueness of the neck muscle spindles (Richmond and Abrahams, <xref rid="B139" ref-type="bibr">1979</xref>; Price and Dutia, <xref rid="B133" ref-type="bibr">1989</xref>), they are endowed with fusimotor fibers as are almost all body muscles. Therefore, the spindle discharge may well be larger during voluntary contraction (even more so for isometric than shortening contractions) than during passive lengthening of the muscles. Hence, the &#x0201c;illusion&#x0201d; of lengthening produced by vibration may as well conceal the illusion of muscle contraction. Accordingly, passive head deviation (without vibration) had no significant effects on the vestibular-evoked self-motion perception (Panichi et al., <xref rid="B123" ref-type="bibr">2011</xref>). This implies that deliberate activation to keep the head deviated is necessary, while neck muscle lengthening induced by passive head rotation may be not sufficient. This must be a different process from that leading to gating of afferent signals to somatosensory cortex during active movement (Williams and Chapman, <xref rid="B178" ref-type="bibr">2002</xref>; Barnett-Cowan and Harris, <xref rid="B6" ref-type="bibr">2011</xref>). Based on others&#x02019; findings (e.g., Inglis et al., <xref rid="B70" ref-type="bibr">1991</xref>), on Panichi et al. (<xref rid="B123" ref-type="bibr">2011</xref>), and Schieppati and Pettorossi (<xref rid="B151" ref-type="bibr">2014</xref>), one can deduce that vibration-induced and contraction-induced effects both depend on a strong discharge of the primary afferent spindle fibers.</p><p>Notably, however, the intensity of the perceptive effects produced by vibration and deliberate muscle contraction can differ due to the motor command (or its &#x0201c;efference copy&#x0201d;) reaching the same centers responsible for the perceptive responses [Panichi et al., <xref rid="B123" ref-type="bibr">2011</xref>; see for a discussion Feldman et al. (<xref rid="B51" ref-type="bibr">2013</xref>)]. The efference copy, by definition, is ahead of the motor performance, and may not correspond to the desired motor effect. It could be argued, based on the ample equivalence of the effects of vibration and contraction on self-motion perception (Schieppati and Pettorossi, <xref rid="B151" ref-type="bibr">2014</xref>) that perception is more driven by real movement than by the intention to move, if it has to have a functional meaning. As a corollary, since the secondary spindle endings are hardly activated by vibration but are certainly activated by the fusimotor discharge, the similarity of the effects of vibration, and contraction suggests that the secondary endings may be not relevant for eliciting the perceptive responses discussed here.</p></sec></sec><sec id="S3-3"><title>Neck proprioception and body orientation during locomotion</title><p>Unilateral vibration of the neck muscles in normal subjects while stepping-in-place or walking produces, in the case of SCM, body turns to the side opposite to vibration, while in the case of DN muscles subjects deviate from the straight-ahead toward the same side as the vibrated muscles (Bove et al., <xref rid="B20" ref-type="bibr">2001</xref>, <xref rid="B19" ref-type="bibr">2002</xref>). The vibration-induced &#x0201c;orienting&#x0201d; effect is also common to other axial muscles, stimulated when the vibrators are in a paraspinal position at the toraco-lumbar junction (Schmid et al., <xref rid="B153" ref-type="bibr">2005</xref>). Among the paraspinal muscles, the multifidus, rotatores, and semispinalis muscles rotate the vertebral column and the trunk to the opposite side [the erectors spinae also receive the vibratory stimulation when the vibrator is placed on the lumbar back, but their role would be that of a stabilizer rather than a rotator, see Kumar et al. (<xref rid="B84" ref-type="bibr">2002</xref>)]. Interestingly, axial muscles have a larger spindle density than other muscles (Voss, <xref rid="B171" ref-type="bibr">1971</xref>; Banks, <xref rid="B4" ref-type="bibr">2006</xref>). Mapping of several muscles within the same subjects during ground locomotion has confirmed the notion that only axial muscles (as opposed to limb muscles) are capable, when vibrated, of producing major, clear-cut deviations of the walking trajectories eyes closed (Courtine et al., <xref rid="B39" ref-type="bibr">2007</xref>).</p><p>It is worth noting that not only neck or trunk muscle vibration but also galvanic vestibular stimulation induces major effects on the trajectory of the walking path (Fitzpatrick et al., <xref rid="B55" ref-type="bibr">1999</xref>). Moreover, changing head posture changes the interpretation of the galvanic vestibular signal for balance and orientation responses (Fransson et al., <xref rid="B56" ref-type="bibr">2000</xref>; Deshpande and Patla, <xref rid="B45" ref-type="bibr">2005</xref>; Fitzpatrick et al., <xref rid="B54" ref-type="bibr">2006</xref>). Thus, vibration-evoked responses from axial muscles might disclose interesting properties of vestibular influences on the control of body orientation. Clearly, the axial muscles are an important source of information about head and trunk orientation in space, and their discharge provides the CNS with cues about body orientation and rotation in space, which are then somehow transmitted to the centers controlling locomotion. Remarkably, though, subjects are not aware of any head or body yaw deviation during walking or rotation during stepping-in-place with vibration, and are always surprised by their unexpected position in space at the end of the trials, indicating that neck proprioception <italic>per se</italic> may not produce strong <italic>conscious</italic> perception of self-motion. In a similar manner people with vestibular dysfunction when asked to step in place with eyes closed are surprised by their change in body orientation.</p><p>The complexity of the underlying mechanisms can be appreciated by the fact that trajectory deviations by vibration are only obtained when locomotion is in progress. If the unilateral vibration starts before subjects initiate stepping, both feet on the ground, no obvious deviation is detected (Schmid et al., <xref rid="B153" ref-type="bibr">2005</xref>). This seems to be in line with the notion that orientation in space is not only the result of an automatic sensory integration process but also depends on awareness of the orientation of the body segments, including the feet (Lyon and Day, <xref rid="B97" ref-type="bibr">2005</xref>), very much as occurs for the sense of verticality (Barra and P&#x000e9;rennou, <xref rid="B7" ref-type="bibr">2013</xref>).</p><p>Interestingly, in cervical dystonia, patients stepping-in-place show non-systematic body rotations during vibration of SCM. In addition, rotations are smaller than in normal subjects, and the confidence intervals in the patient population are about twice as much as those obtained for the normal subjects (Bove et al., <xref rid="B17" ref-type="bibr">2004</xref>). It seems that in many patients the reference system used in the control of body orientation in space is either refractory to the lateralized proprioceptive neck input [also the effects on the standing body orientation are attenuated in cervical dystonia; see Lekhel et al. (<xref rid="B88" ref-type="bibr">1997</xref>) and Bove et al. (<xref rid="B18" ref-type="bibr">2007</xref>)], or modified such that the input from either sides produces small or even &#x0201c;wrong&#x0201d; effects. Note that, in a seated patients, long-lasting vibration of the dystonic muscle produced persistent reorientation of the head, as a sign of the function of segmental circuitry subserving head rotation (Karnath et al., <xref rid="B78" ref-type="bibr">2000</xref>). Perhaps, this relative obliviousness of neck proprioception in the context of whole-body orientation in dystonia is connected to plasticity in the supraspinal circuits and centers integrating the neck input, shaped by the long-term asymmetric spindle inflow from one side of the neck (M&#x000fc;nchau and Bronstein, <xref rid="B116" ref-type="bibr">2001</xref>). Likewise, in Writer&#x02019;s Cramp (Gr&#x000fc;newald et al., <xref rid="B62" ref-type="bibr">1997</xref>), the sensation of movement produced by the vibratory stimulus was not perceived normally in the dystonic patients, as if misinterpretation of Ia-afferent discharges also occurred (Wagner et al., <xref rid="B174" ref-type="bibr">2008</xref>).</p><p>Body orientation during locomotion and stepping-in-place must be instant-by-instant coherent with the SSA. Thus, unilateral vibration of a neck muscle must exert an influence on the centers that produce the gait pattern, not unlikely that exerted by the galvanic stimulation (Fitzpatrick et al., <xref rid="B55" ref-type="bibr">1999</xref>; Iles et al., <xref rid="B69" ref-type="bibr">2007</xref>) or by the voluntary command for turning (Courtine and Schieppati, <xref rid="B40" ref-type="bibr">2003</xref>). Notably, during volitional locomotion along a curved trajectory, head yaw anticipates body yaw (Courtine and Schieppati, <xref rid="B40" ref-type="bibr">2003</xref>). The head turns more than dictated by the heading change, probably as a sign of anticipation: head orientation with respect to the body antecedes the body heading at the next step, and so on for the successive steps. This, however, may be not an obligatory coupling during volitional locomotion, since Cinelli and Warren (<xref rid="B33" ref-type="bibr">2012</xref>) argue that head rotations <italic>per se</italic> are neither necessary nor sufficient to induce changes in the direction of locomotion when walking to a goal.</p><p>The asymmetric vestibular stimulation mentioned in a preceding paragraph influences the SSA because the information associated with the fast rotation to one side largely prevails, while that associated with rotation to the opposite side (of equal amplitude, but slower) weakens the sense of the rotation (Pettorossi et al., <xref rid="B128" ref-type="bibr">2013a</xref>,<xref rid="B129" ref-type="bibr">b</xref>). In turn, the unilateral vibration of neck muscle strongly influences the effect of the asymmetric whole-body rotation (Panichi et al., <xref rid="B123" ref-type="bibr">2011</xref>). There must be some algebraic effect of the two stimulations at some central site (the bias in the self-motion perception of vestibular origin may be either enhanced or annulled depending on the vibrated neck muscle). If this is tenable, one would argue that the rotation during stepping-in-place or the deviation during locomotion induced by unilateral vibration of axial muscles depend on the priority of the moving body, i.e., continuously keeping the current SSA in front of it. Interestingly, blindfolded subjects have a tendency to walk in a large circle. Souman et al. (<xref rid="B163" ref-type="bibr">2009</xref>) suggested that veering from a straight course may result from accumulating noise in the sensorimotor system, without an external directional reference to recalibrate the subjective straight ahead. It is not unlikely that minor but enduring asymmetric proprioceptive input, not periodically checked by vision, may causes people to walk in circles as a result of errors in their SSA.</p></sec></sec><sec id="S4"><title>Aftereffects of Neck Vibratory Stimulation</title><p>Aftereffect is by definition an aspect of adaptation due to the history of stimulation, which persists after the end of the stimulus (Helson, <xref rid="B66" ref-type="bibr">1948</xref>). The aftereffect can be simply a continuation of the effect or it can show responses of opposite sign. Neck-proprioceptive stimulation, especially after prolonged vibration to the muscles, should elicit aftereffects. Other systems, apart from the proprioceptive, are also involved in postural control and space orientation and show aftereffects. The vestibular and the optokinetic systems, after prolonged stimulation, exhibit responses that are initially coherent with those induced by the stimulus (post-rotatory nystagmus, PRN; optokinetic after-nystagmus, OKAN) (Brandt et al., <xref rid="B25" ref-type="bibr">1974</xref>; Waespe and Henn, <xref rid="B173" ref-type="bibr">1978</xref>; Clement et al., <xref rid="B35" ref-type="bibr">1981</xref>; Koenig and Dichgans, <xref rid="B82" ref-type="bibr">1981</xref>; Lisberger et al., <xref rid="B90" ref-type="bibr">1981</xref>; Maioli, <xref rid="B98" ref-type="bibr">1988</xref>; Pettorossi et al., <xref rid="B127" ref-type="bibr">1999</xref>). Shortly afterward, these responses reverse their sign, typically showing after-nystagmus of the opposite sign (PRN II and OKAN II). These responses may be due to habituation taking place in the central optokinetic and vestibular circuitry. Neck muscle proprioception activation can also produce effects on body orientation that outlast the vibration train. These persistent effects would not be produced by reflex adaptation or by proprioceptive receptor post-discharges, as if previously activated spindles continue firing (Ribot-Ciscar et al., <xref rid="B138" ref-type="bibr">1996</xref>). They do not reverse in sign, and are possibly linked to a specific central-integration process.</p><sec id="S4-4"><title>Aftereffect on balance</title><p>The inclination of the body induced by symmetric DN muscle vibration is in the <italic>forward</italic> direction both during and after the end of stimulation (Lund, <xref rid="B94" ref-type="bibr">1980</xref>; Ivanenko et al., <xref rid="B72" ref-type="bibr">1999</xref>, <xref rid="B71" ref-type="bibr">2000</xref>; Kavounoudias et al., <xref rid="B81" ref-type="bibr">1999</xref>; Bove et al., <xref rid="B23" ref-type="bibr">2006a</xref>,<xref rid="B24" ref-type="bibr">b</xref>). The aftereffect on posture can last several minutes (Wierzbicka et al., <xref rid="B177" ref-type="bibr">1998</xref>). As mentioned in Duclos et al. (<xref rid="B48" ref-type="bibr">2004</xref>), similar aftereffects were found not only after prolonged vibrations applied to neck muscles but also after prolonged voluntary contraction of the same muscles.</p><p>Conversely, a different aftereffect on balance displacement has been reported when vibration is applied to non-axial muscles. For instance, backwards body inclination and trunk extension (Thompson et al., <xref rid="B168" ref-type="bibr">2007</xref>), observed during vibration of soleus (Capic&#x000ed;kov&#x000e1; et al., <xref rid="B30" ref-type="bibr">2006</xref>), invert to forward inclination after the end of the stimulus, albeit it with a large variability of the responses. An opposite aftereffect has also been observed for joint movement perception (Seizova-Cajic et al., <xref rid="B159" ref-type="bibr">2007</xref>). Habituation of the illusion of elbow extension occurs during biceps brachii vibration, and after vibration ends a flexion illusion subsides. It appears that the direction of the vibratory aftereffect is coherent with that observed during vibration when neck muscles are vibrated but has an opposite direction when limb muscles are vibrated.</p></sec><sec id="S4-5"><title>Aftereffect on the subjective straight-ahead</title><p>The vibration-induced deviation of the SSA (or the space in front of our nose) (Taylor and McCloskey, <xref rid="B167" ref-type="bibr">1991</xref>; Seizova-Cajic et al., <xref rid="B158" ref-type="bibr">2006</xref>) persists in the same direction as during the stimulus, when vibration stops. Depending on the duration of the stimulus, the aftereffect on SSA can last several hours or days in both normal subjects (Karnath et al., <xref rid="B79" ref-type="bibr">2002</xref>) and neglect patients (Ferber and Karnath, <xref rid="B52" ref-type="bibr">1999</xref>; Schindler et al., <xref rid="B152" ref-type="bibr">2002</xref>; Johannsen et al., <xref rid="B74" ref-type="bibr">2003</xref>).</p><p>On the other hand, the motion of an illusory visual target induced by vibration reverses its direction at the end of the vibration. Therefore, the illusory visual target movement, which is coherent with the SSA displacement during vibration, becomes incoherent when vibration is discontinued (Lackner and Levine, <xref rid="B85" ref-type="bibr">1979</xref>; Biguer et al., <xref rid="B12" ref-type="bibr">1988</xref>; Taylor and McCloskey, <xref rid="B167" ref-type="bibr">1991</xref>). To explain this discrepancy, Seizova-Cajic and Sachtler (<xref rid="B157" ref-type="bibr">2007</xref>) have proposed that the aftereffect inversion of illusory target movement is primarily related to the presence of visual signal, since the inversion is absent when the target is not seen during vibration but only after vibration.</p></sec><sec id="S4-6"><title>Aftereffect on self-motion perception</title><p>The large modulation produced by neck muscle vibration in the movement perception of vestibular origin, mentioned above (see Section &#x0201c;Neck Muscle Vibration Modulates Self-Motion Perception of Vestibular Origin&#x0201d;), is present not only during the on-going vibratory stimulation but also after it (Schieppati and Pettorossi, <xref rid="B151" ref-type="bibr">2014</xref>) (Figure <xref ref-type="fig" rid="F2">2</xref>). The enhancement of the vestibular-elicited motion perception bias (in the direction of the head deviation, or in the direction that the vibrated muscle would induce if contracted), or the reduction of the motion perception bias (with the head deviated in the opposite direction or when an antagonistic muscle was vibrated), both persist at the end of the vibratory stimulus. The aftereffect endures minutes or hours depending on the duration and frequency of vibration and on the status of the vibrated muscle (relaxed or contracted). In passing, persistent aftereffects of proprioceptive origin have been observed not only in motion perception but also in completely different experiments and in other muscle groups. For instance, prolonged vibration of limb muscles induces long-term cortical excitability change (Marconi et al., <xref rid="B101" ref-type="bibr">2008</xref>), enhancement of leg muscle power, and improvement of body balance (Brunetti et al., <xref rid="B27" ref-type="bibr">2006</xref>; Filippi et al., <xref rid="B53" ref-type="bibr">2009</xref>).</p><fig id="F2" position="float"><label>Figure 2</label><caption><p><bold>Long-lasting aftereffect of neck muscle vibration on the self-motion perception</bold>. <bold>(A)</bold> Traces of the tracking of remembered visual target during a single cycle of asymmetric rotation (0.15&#x02009;Hz frequency, asymmetry 80%). Top: platform oscillation trace. Bottom: tracking traces, with (red) and without (blue) left sternocleidomastoideus (SCM) muscle vibration. Note the tracking position error (TPE) at the end of the rotation cycle. The position error between the real position of target and its representation (vertical bars), produced by the asymmetric whole-body rotation, is further increased by the left SCM vibration (red trace). <bold>(B)</bold> The time course of the enhancement of TPE after SCM vibration. In abscissa: time after vibration, at which the asymmetric oscillation cycle is administered; in ordinate: amplitude of TPE. Note that the TPE enhancement and its persistence is influenced by the duration and frequency of the vibration train and by the simultaneous contraction of SCM [adapted from Schieppati and Pettorossi (<xref rid="B151" ref-type="bibr">2014</xref>)].</p></caption><graphic xlink:href="fnhum-08-00895-g002"/></fig><p>Neck vibration aftereffects on self-motion perception could be explained by plastic events occurring in the vestibular networks responsible for motion perception when there is an intense proprioceptive activation, able to drive persistent membrane and genomic synaptic changes (Grassi and Pettorossi, <xref rid="B60" ref-type="bibr">2001</xref>; Wolpaw and Tennissen, <xref rid="B181" ref-type="bibr">2001</xref>; Lynch, <xref rid="B96" ref-type="bibr">2004</xref>; Straka et al., <xref rid="B166" ref-type="bibr">2005</xref>; Pettorossi et al., <xref rid="B126" ref-type="bibr">2011</xref>). A remarkable consolidation of the aftereffect is obtained with frequencies of 80&#x02013;100&#x02009;Hz, while below this range the persistence of the aftereffect is scarce. The greater efficacy of the high-frequency entails a stronger activation of the primary spindle afferents onto the central network. Higher frequencies may be also more apt <italic>per se</italic> to induce synaptic plasticity, since high-frequency stimulation induces learning processes in other afferent systems, while low frequencies tend to reduce such effect [Lynch, <xref rid="B96" ref-type="bibr">2004</xref>; Stanton and Sejnowski, <xref rid="B165" ref-type="bibr">1989</xref>; Bliss and Collingridge, <xref rid="B14" ref-type="bibr">1993</xref>; Nicoll and Malenka, <xref rid="B121" ref-type="bibr">1995</xref>; Pettorossi et al., <xref rid="B129" ref-type="bibr">2013b</xref>; Scarduzio et al., <xref rid="B147" ref-type="bibr">2013</xref>; Beste and Dinse, <xref rid="B9" ref-type="bibr">2013</xref>; Seitz and Dinse (<xref rid="B156" ref-type="bibr">2007</xref>)]. <italic>In vitro and in vivo</italic> experiments with different types of afferent fiber stimulation, tactile (Dinse et al., <xref rid="B47" ref-type="bibr">2003</xref>, <xref rid="B46" ref-type="bibr">2011</xref>; Ragert et al., <xref rid="B137" ref-type="bibr">2008</xref>), visual (Beste et al., <xref rid="B10" ref-type="bibr">2011</xref>; Beste and Dinse, <xref rid="B9" ref-type="bibr">2013</xref>), acoustic (Amitay et al., <xref rid="B2" ref-type="bibr">2006</xref>), and vestibular afferents (Grassi et al., <xref rid="B61" ref-type="bibr">1996</xref>; Grassi and Pettorossi, <xref rid="B60" ref-type="bibr">2001</xref>; Pettorossi et al., <xref rid="B129" ref-type="bibr">2013b</xref>; Scarduzio et al., <xref rid="B147" ref-type="bibr">2013</xref>), suggest that high-frequency afferent fiber stimulation leads to long-term potentiation (LTP, Lynch, <xref rid="B96" ref-type="bibr">2004</xref>) in several regions of the CNS, while low frequency to long-term depression or cancelation of previously induced LTP.</p><sec id="S4-6-3"><title>Muscle status</title><p>The status of the muscle during the vibration is critical for inducing the long-term aftereffect on self-motion perception. Tonic, isometric muscle contraction can increase both the amplitude and the duration of the perceptive aftereffect induced by vibration (Schieppati and Pettorossi, <xref rid="B151" ref-type="bibr">2014</xref>). The enhancement of the aftereffect obtained by concomitant vibration and muscle contraction on self-motion perception is unexpectedly greater than that estimated by adding the aftereffects of both muscle contraction and vibration (Figure <xref ref-type="fig" rid="F2">2</xref>). Post-vibratory effects and post-contraction response show remarkable similarities, as studied in the arm muscles (Gilhodes et al., <xref rid="B58" ref-type="bibr">1992</xref>). In a study based on a different paradigm and addressing the effects of a limb movement on movement orientation, repetitive active arm movements <italic>against a load</italic> induced lasting changes in the space representation when active movement repetition lasted for at least 10&#x02009;min (Ostry et al., <xref rid="B122" ref-type="bibr">2010</xref>). While isometric contraction increases the activation of muscle spindles in response to the vibration either by enhancing spindle sensitiveness through &#x003b3;-motoneuron activity, or by facilitating a better diffusion of the vibration within the muscle thanks to its increased muscle stiffness (Burke et al., <xref rid="B29" ref-type="bibr">1976b</xref>), muscle contraction (and its efference copy) superimposed to the vibratory stimulation would favor the build-up and consolidation of the influences on motion perception (Rymer and D&#x02019;Almeida, <xref rid="B146" ref-type="bibr">1980</xref>; Smith et al., <xref rid="B162" ref-type="bibr">2009</xref>; Luu et al., <xref rid="B95" ref-type="bibr">2011</xref>). In particular, the voluntary activation in concomitance with peripheral proprioceptive stimulation may lead to potentiation of the synaptic responses, where the peripheral input and central drive converge along the perceptive central pathway.</p></sec></sec><sec id="S4-7"><title>Aftereffect on locomotion</title><p>The walking speed increment induced by bilateral vibration of the DN muscles, likely the consequence of the postural illusion mentioned above (Ivanenko et al., <xref rid="B71" ref-type="bibr">2000</xref>), promptly subsides at the end of the vibratory train. Aftereffects on stepping induced by bilateral contraction of DN muscles, performed during stance, are also inconsistent, unless contraction has produced fatigue (Schmid and Schieppati, <xref rid="B154" ref-type="bibr">2005</xref>), under which circumstance the stepping body tends to move backwards. On the other hand, unilateral neck muscle vibration shows non-systematic aftereffects on stepping direction. The body initially rotates toward one side (most often the same side as during the vibration administered during stepping) and rotate afterward toward the opposite side (Bove et al., <xref rid="B19" ref-type="bibr">2002</xref>). Further, when vibration (or contraction, see above) is applied during stance, and stepping follows at the end of the vibration, the poor consistency of the aftereffect on body rotation is likely due on the details of the experimental procedure. Having both feet on the ground during vibration provides a fixed reference that attenuates the effects of the unilateral proprioceptive activation under this circumstance, much as light-touch does on the standing body orientation during vibration (Bove et al., <xref rid="B23" ref-type="bibr">2006a</xref>,<xref rid="B24" ref-type="bibr">b</xref>).</p><p>The information from the foot and leg status must interact with the supraspinal spatial orientation areas that influence spinal-level circuits for locomotion (Figure <xref ref-type="fig" rid="F3">3</xref>). Not unlikely, this is what occurs as a consequence of the so-called podokinetic adaptation, a whole-body yaw rotation during stepping-in-place eyes closed occurring after a period of stepping on a rotating treadmill (Weber et al., <xref rid="B176" ref-type="bibr">1998</xref>). Interestingly, when asked to indicate their SSA with a laser pointer, these subjects demonstrated a significant shift in SSA regardless of whether they were standing or sitting (Scott et al., <xref rid="B155" ref-type="bibr">2011</xref>). This would be in keeping with the notion that prolonged adaptive rotation of the feet may influence the SSA, and with the proposal that subjects track their SSA during the involuntary rotation aftereffect as much as they do with unilateral neck muscle vibration.</p><fig id="F3" position="float"><label>Figure 3</label><caption><p><bold>Outline of the immediate and sustained interaction between neck proprioceptive and vestibular inflow, and of the motor command for shaping the space reference and influencing motor perception and locomotion</bold>. Neck proprioception (blue arrow) and vestibular (red arrow) signals can modify movement and space perception separately. However, these changes can also result from the interaction between neck proprioception and vestibular system. Unilateral neck muscle vibration, mimicking either neck muscle elongation or contraction by &#x003b3;-motoneuron drive, increases the vestibular responsiveness to yaw rotation (black arrow) in the direction corresponding to the head movement that would be produced by the contraction of vibrated muscle, and decrease when the antagonist muscle is being vibrated. Thus, tonic proprioceptive signals elicited by vibration can interact with the dynamic vestibular input and change the perception of whole-body rotation in yaw plane. Space reference and locomotion may be subsequently modified through this interaction (violet arrow). Therefore, neck proprioception directly, or indirectly by vestibular system, contributes to shape the subjective straight-ahead, the direction of walking and the self-motion perception. Prolonged neck muscle vibration also induces aftereffects in the straight ahead and motion perception in the same direction of the immediate effects. The persistence of the aftereffects depends on the intensity and duration of the vibratory stimulation. Descending signals for the head movement may also influence movement and motion perception (green arrow). They can contribute either to the effect and aftereffect by enhancing the peripheral signals from neck muscle (continuous line) or by directly changing the vestibular responsiveness to rotation (dashed line).</p></caption><graphic xlink:href="fnhum-08-00895-g003"/></fig><p>On the other hand, also the post-contraction facilitatory effect (Kohnstamm phenomenon), induced by prolonged and forceful deliberate body torsion, can modify the direction of an intended straight-ahead walking task, such that subjects walk along a curved trajectory in the direction of the preceding torsion (Ivanenko et al., <xref rid="B73" ref-type="bibr">2006</xref>). This further supports the view that the proprioceptive inflow responsible for the orientation effects and aftereffect can be elicited by both vibration and contraction of axial muscles.</p></sec><sec id="S4-8"><title>The direction of the aftereffect</title><p>The aftereffect of neck muscle vibration seems to have the same direction as the effect observed during vibration: this would be true for the SSA displacement, standing balance displacement, and self-motion perception. On the other hand, the direction of the aftereffect would be opposite in the case of vibration of limb muscles. The reason for this divergence between neck and limb muscle vibration is not obvious. Neck vibration seems to consolidate the effect elicited during vibration, so that the effect is maintained even after the end of the stimulus. With limb proprioceptive vibration, on the other hand, the aftereffect of opposite sign could be attributed to the sustained stimulation leading to a habituation of the responses.</p><p>The consolidation of proprioceptive effects after neck muscle vibration such as those mentioned above, as opposed to habituation, would be explained by the different roles played by the proprioceptive system in the neck and limb muscles. It has been suggested that the tendency to habituation in the effect of limb proprioceptive activation is aimed to minimize the response to common environmental stimuli and increase the sensitivity to change (Seizova-Cajic et al., <xref rid="B159" ref-type="bibr">2007</xref>), similarly to what happens in the vestibular and optokinetic system (Brandt et al., <xref rid="B25" ref-type="bibr">1974</xref>; Waespe and Henn, <xref rid="B173" ref-type="bibr">1978</xref>; Clement et al., <xref rid="B35" ref-type="bibr">1981</xref>; Koenig and Dichgans, <xref rid="B82" ref-type="bibr">1981</xref>; Lisberger et al., <xref rid="B90" ref-type="bibr">1981</xref>; Maioli, <xref rid="B98" ref-type="bibr">1988</xref>; Pettorossi et al., <xref rid="B127" ref-type="bibr">1999</xref>). Conversely, the persistence of the effects on orientation and self-motion perception observed upon prolonged neck muscle activation supports the idea that the repeated proprioceptive information can shape a new reference frame for head and body around a new postural set (Karnath et al., <xref rid="B79" ref-type="bibr">2002</xref>; Schieppati and Pettorossi, <xref rid="B151" ref-type="bibr">2014</xref>). This might not be dissimilar from what occurs during motor learning (Lalazar and Vaadia, <xref rid="B87" ref-type="bibr">2008</xref>). It is not unlikely that anomalous persistence of neck effects on motion perception may occur also as a consequence of pathological conditions presumably associated with persistent abnormal spindle discharge.</p></sec></sec><sec id="S5"><title>Neck Muscle Spindle Primary Afferent Fibers Produce Immediate and Long-Term Influences on the Cognitive Body Representation</title><p>All in all, the behavioral and neurophysiological data reported above emphasize that proprioception from neck muscles contributes to the construction of cognitive representation of the body that includes position of limb segments, their hierarchical arrangement, and configuration of the segments in space. This does not normally enter into awareness, and may be primarily used for spatial organization of action (Haggard and Wolpert, <xref rid="B65" ref-type="bibr">2005</xref>). Apparently, the proprioceptive information is processed according to the task performed, the time-interval during which the afferent volley takes place, the body segment from which the sensory inflow arises (neck and trunk), and concurrent stabilizing information. The information conveyed by the Ia fibers is rapidly transmitted to diverse parts of the CNS, and updates the brain on muscle length changes, and thus on movement. Its integration may occur at various level of the central nervous system, known to supervise the formation of reference frames for movement. Most likely, the vestibular nuclei, which receive neck muscle input, are the first stage for the integration of neck muscle vibratory signals and play a crucial role in conscious awareness of motion, spatial orientation, and navigation (Lopez, <xref rid="B92" ref-type="bibr">2013</xref>) (Figure <xref ref-type="fig" rid="F3">3</xref>). The fastigial nucleus of the cerebellum is a site, in which computation of body motion is performed (Brooks and Cullen, <xref rid="B26" ref-type="bibr">2009</xref>). Other anatomical substrates involved in the processing of neck muscle inflow are the motor cortex (Naito, <xref rid="B118" ref-type="bibr">2004</xref>) and the parieto-temporal junction (Bottini et al., <xref rid="B16" ref-type="bibr">2001</xref>). Interestingly, studies based on structural brain imaging [reviewed in Karnath and Rorden (<xref rid="B80" ref-type="bibr">2011</xref>), Blanke (<xref rid="B13" ref-type="bibr">2012</xref>), and Pfeiffer et al. (<xref rid="B130" ref-type="bibr">2014</xref>)] suggest that diverse subcortical (Clark and Taube, <xref rid="B34" ref-type="bibr">2012</xref>) and cortical areas spanning the Sylvian fissure can be a substrate of the integration concerned in cross-modal interactions between somatosensory and vestibular signals (Bottini et al., <xref rid="B15" ref-type="bibr">2013</xref>). These areas can be lesioned in various forms of neglect (Vuilleumier, <xref rid="B172" ref-type="bibr">2013</xref>).</p><p>The effects of intense neck muscle vibration on self-motion perception during whole-body rotation are not restricted to the epoch of the stimulation but persist for an extended period of time. Vibration can lead to a consolidation of new space coordinate system by persistently modifying the self-motion perception of vestibular origin and interacting with the adaptive processes of the vestibular system (St George et al., <xref rid="B164" ref-type="bibr">2011</xref>). This may be useful for adapting the perceptive (and consequently motor) responses to a novel postural set or motor bias, when proprioceptive activation persists for a sufficient period of time, thereby influencing the spatial references, motion perception, and locomotor orientation.</p><p>Admittedly, while vibration is an adequate stimulus for the rapidly adapting primary spindle terminals, vibratory trains are quite an unusual stimulation for the proprioceptive system, not least because it can signify anatomically impossible kinematics (Lackner and Taublieb, <xref rid="B86" ref-type="bibr">1984</xref>; Seizova-Cajic et al., <xref rid="B159" ref-type="bibr">2007</xref>). However, the effect of vibration mimics, at least in part, the effect of the &#x003b3;-motoneuron activation, thereby functionally engaging the same pathway traveled during voluntary movement, and has been shown to have positive effect in various patients. It is on these premises that therapeutic effects of focal vibration may have a role in the armamentarium of the restorative neurology [see for a recent review Murillo et al. (<xref rid="B117" ref-type="bibr">2014</xref>)]. Long-duration trains of Ia firing, as induced by vibration, may disclose the capacity of proprioception to produce adaptive effects in an as yet unnoticed way. Recent findings by Yu et al. (<xref rid="B183" ref-type="bibr">2013</xref>) have shown in the cat that repeated exposure to cross-modal stimulation enhances neuronal sensitivity to the stimuli in the exposure set [see for a review Rowland and Stein (<xref rid="B145" ref-type="bibr">2014</xref>)]. By looking for aftereffects, Wright (<xref rid="B182" ref-type="bibr">2014</xref>) asked whether postural responses seen during discordant virtual-reality and physical vection stimulation involved adaptation, and described an aftereffect in the center of foot pressure, that could even last for a few days. New experiments dedicated to the observation of the effects of persistent activation of proprioceptors could provide novel insight into the plastic changes of our motor processes.</p></sec><sec id="S6"><title>Conclusion</title><p>Neck muscle inflow has prominent immediate and late effects on perception of body orientation and motion. Prolonged, intense proprioceptive input from neck muscles can induce persistent influences on self-motion perception and cognitive body representation (Figure <xref ref-type="fig" rid="F3">3</xref>). These plastic changes might adapt motion sensitiveness to lasting or permanent head positional or motor changes, like those accompanying movement disorders (see above) or those accompanying weightlessness (Roll et al., <xref rid="B143" ref-type="bibr">1998</xref>). New experimental protocols based on these findings could open new avenues in the investigation of the consolidation of motor learning.</p></sec><sec id="S7"><title>Conflict of Interest Statement</title><p>No party having a direct interest in the results of the research supporting this article has or will confer a benefit on the authors or on any organization with which the authors are associated.</p></sec>

Mortality from external causes in Africa and Asia: evidence from INDEPTH Health and Demographic Surveillance System Sites

<sec id="st1"><title>Background</title><p>Mortality from external causes, of all kinds, is an important component of overall mortality on a global basis. However, these deaths, like others in Africa and Asia, are often not counted or documented on an individual basis. Overviews of the state of external cause mortality in Africa and Asia are therefore based on uncertain information. The INDEPTH Network maintains longitudinal surveillance, including cause of death, at population sites across Africa and Asia, which offers important opportunities to document external cause mortality at the population level across a range of settings.</p></sec><sec id="st2"><title>Objective</title><p>To describe patterns of mortality from external causes at INDEPTH Network sites across Africa and Asia, according to the WHO 2012 verbal autopsy (VA) cause categories. </p></sec><sec id="st3"><title>Design</title><p>All deaths at INDEPTH sites are routinely registered and followed up with VA interviews. For this study, VA archives were transformed into the WHO 2012 VA standard format and processed using the InterVA-4 model to assign cause of death. Routine surveillance data also provide person-time denominators for mortality rates.</p></sec><sec id="st4"><title>Results</title><p>A total of 5,884 deaths due to external causes were documented over 11,828,253 person-years. Approximately one-quarter of those deaths were to children younger than 15 years. Causes of death were dominated by childhood drowning in Bangladesh, and by transport-related deaths and intentional injuries elsewhere. Detailed mortality rates are presented by cause of death, age group, and sex.</p></sec><sec id="st5"><title>Conclusions</title><p>The patterns of external cause mortality found here generally corresponded with expectations and other sources of information, but they fill some important gaps in population-based mortality data. They provide an important source of information to inform potentially preventive intervention designs.</p></sec>

<contrib contrib-type="author"><name><surname>Streatfield</surname><given-names>P. Kim</given-names></name><xref ref-type="aff" rid="AF0001">1</xref><xref ref-type="aff" rid="AF0002">2</xref><xref ref-type="aff" rid="AF0003">3</xref></contrib><contrib contrib-type="author"><name><surname>Khan</surname><given-names>Wasif A.</given-names></name><xref ref-type="aff" rid="AF0002">2</xref><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0004">4</xref></contrib><contrib contrib-type="author"><name><surname>Bhuiya</surname><given-names>Abbas</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0005">5</xref><xref ref-type="aff" rid="AF0006">6</xref></contrib><contrib contrib-type="author"><name><surname>Hanifi</surname><given-names>Syed M.A.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0005">5</xref><xref ref-type="aff" rid="AF0006">6</xref></contrib><contrib contrib-type="author"><name><surname>Alam</surname><given-names>Nurul</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0007">7</xref><xref ref-type="aff" rid="AF0008">8</xref></contrib><contrib contrib-type="author"><name><surname>Diboulo</surname><given-names>Eric</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0009">9</xref><xref ref-type="aff" rid="AF0010">10</xref></contrib><contrib contrib-type="author"><name><surname>Niamba</surname><given-names>Louis</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0009">9</xref><xref ref-type="aff" rid="AF0010">10</xref></contrib><contrib contrib-type="author"><name><surname>Si&#x000e9;</surname><given-names>Ali</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0009">9</xref><xref ref-type="aff" rid="AF0010">10</xref></contrib><contrib contrib-type="author"><name><surname>Lankoand&#x000e9;</surname><given-names>Bruno</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0011">11</xref><xref ref-type="aff" rid="AF0012">12</xref></contrib><contrib contrib-type="author"><name><surname>Millogo</surname><given-names>Roch</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0011">11</xref><xref ref-type="aff" rid="AF0012">12</xref></contrib><contrib contrib-type="author"><name><surname>Soura</surname><given-names>Abdramane B.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0011">11</xref><xref ref-type="aff" rid="AF0012">12</xref></contrib><contrib contrib-type="author"><name><surname>Bonfoh</surname><given-names>Bassirou</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0013">13</xref><xref ref-type="aff" rid="AF0014">14</xref></contrib><contrib contrib-type="author"><name><surname>Kone</surname><given-names>Siaka</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0013">13</xref><xref ref-type="aff" rid="AF0014">14</xref></contrib><contrib contrib-type="author"><name><surname>Ngoran</surname><given-names>Eliezer K.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0013">13</xref><xref ref-type="aff" rid="AF0015">15</xref></contrib><contrib contrib-type="author"><name><surname>Utzinger</surname><given-names>Juerg</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0013">13</xref><xref ref-type="aff" rid="AF0016">16</xref></contrib><contrib contrib-type="author"><name><surname>Ashebir</surname><given-names>Yemane</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0017">17</xref><xref ref-type="aff" rid="AF0018">18</xref></contrib><contrib contrib-type="author"><name><surname>Melaku</surname><given-names>Yohannes A.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0017">17</xref><xref ref-type="aff" rid="AF0018">18</xref></contrib><contrib contrib-type="author"><name><surname>Weldearegawi</surname><given-names>Berhe</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0017">17</xref><xref ref-type="aff" rid="AF0018">18</xref></contrib><contrib contrib-type="author"><name><surname>Gomez</surname><given-names>Pierre</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0019">19</xref><xref ref-type="aff" rid="AF0020">20</xref></contrib><contrib contrib-type="author"><name><surname>Jasseh</surname><given-names>Momodou</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0019">19</xref><xref ref-type="aff" rid="AF0020">20</xref></contrib><contrib contrib-type="author"><name><surname>Azongo</surname><given-names>Daniel</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0021">21</xref><xref ref-type="aff" rid="AF0022">22</xref></contrib><contrib contrib-type="author"><name><surname>Oduro</surname><given-names>Abraham</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0021">21</xref><xref ref-type="aff" rid="AF0022">22</xref></contrib><contrib contrib-type="author"><name><surname>Wak</surname><given-names>George</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0021">21</xref><xref ref-type="aff" rid="AF0022">22</xref></contrib><contrib contrib-type="author"><name><surname>Wontuo</surname><given-names>Peter</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0021">21</xref><xref ref-type="aff" rid="AF0022">22</xref></contrib><contrib contrib-type="author"><name><surname>Attaa-Pomaa</surname><given-names>Mary</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0023">23</xref><xref ref-type="aff" rid="AF0024">24</xref></contrib><contrib contrib-type="author"><name><surname>Gyapong</surname><given-names>Margaret</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0023">23</xref><xref ref-type="aff" rid="AF0024">24</xref></contrib><contrib contrib-type="author"><name><surname>Manyeh</surname><given-names>Alfred K.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0023">23</xref><xref ref-type="aff" rid="AF0024">24</xref></contrib><contrib contrib-type="author"><name><surname>Kant</surname><given-names>Shashi</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0025">25</xref><xref ref-type="aff" rid="AF0026">26</xref></contrib><contrib contrib-type="author"><name><surname>Misra</surname><given-names>Puneet</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0025">25</xref><xref ref-type="aff" rid="AF0026">26</xref></contrib><contrib contrib-type="author"><name><surname>Rai</surname><given-names>Sanjay K.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0025">25</xref><xref ref-type="aff" rid="AF0026">26</xref></contrib><contrib contrib-type="author"><name><surname>Juvekar</surname><given-names>Sanjay</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0027">27</xref><xref ref-type="aff" rid="AF0028">28</xref></contrib><contrib contrib-type="author"><name><surname>Patil</surname><given-names>Rutuja</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0027">27</xref><xref ref-type="aff" rid="AF0028">28</xref></contrib><contrib contrib-type="author"><name><surname>Wahab</surname><given-names>Abdul</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0029">29</xref><xref ref-type="aff" rid="AF0030">30</xref></contrib><contrib contrib-type="author"><name><surname>Wilopo</surname><given-names>Siswanto</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0029">29</xref><xref ref-type="aff" rid="AF0030">30</xref></contrib><contrib contrib-type="author"><name><surname>Bauni</surname><given-names>Evasius</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0031">31</xref><xref ref-type="aff" rid="AF0032">32</xref></contrib><contrib contrib-type="author"><name><surname>Mochamah</surname><given-names>George</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0031">31</xref><xref ref-type="aff" rid="AF0032">32</xref></contrib><contrib contrib-type="author"><name><surname>Ndila</surname><given-names>Carolyne</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0031">31</xref><xref ref-type="aff" rid="AF0032">32</xref></contrib><contrib contrib-type="author"><name><surname>Williams</surname><given-names>Thomas N.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0031">31</xref><xref ref-type="aff" rid="AF0032">32</xref><xref ref-type="aff" rid="AF0033">33</xref></contrib><contrib contrib-type="author"><name><surname>Khaggayi</surname><given-names>Christine</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0034">34</xref><xref ref-type="aff" rid="AF0035">35</xref></contrib><contrib contrib-type="author"><name><surname>Nyaguara</surname><given-names>Amek</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0034">34</xref><xref ref-type="aff" rid="AF0035">35</xref></contrib><contrib contrib-type="author"><name><surname>Obor</surname><given-names>David</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0034">34</xref><xref ref-type="aff" rid="AF0035">35</xref></contrib><contrib contrib-type="author"><name><surname>Odhiambo</surname><given-names>Frank O.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0034">34</xref><xref ref-type="aff" rid="AF0035">35</xref></contrib><contrib contrib-type="author"><name><surname>Ezeh</surname><given-names>Alex</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0036">36</xref><xref ref-type="aff" rid="AF0037">37</xref></contrib><contrib contrib-type="author"><name><surname>Oti</surname><given-names>Samuel</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0036">36</xref><xref ref-type="aff" rid="AF0037">37</xref></contrib><contrib contrib-type="author"><name><surname>Wamukoya</surname><given-names>Marylene</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0036">36</xref><xref ref-type="aff" rid="AF0037">37</xref></contrib><contrib contrib-type="author"><name><surname>Chihana</surname><given-names>Menard</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0038">38</xref><xref ref-type="aff" rid="AF0039">39</xref></contrib><contrib contrib-type="author"><name><surname>Crampin</surname><given-names>Amelia</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0038">38</xref><xref ref-type="aff" rid="AF0039">39</xref><xref ref-type="aff" rid="AF0040">40</xref></contrib><contrib contrib-type="author"><name><surname>Collinson</surname><given-names>Mark A.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0041">41</xref><xref ref-type="aff" rid="AF0042">42</xref><xref ref-type="aff" rid="AF0043">43</xref></contrib><contrib contrib-type="author"><name><surname>Kabudula</surname><given-names>Chodziwadziwa W.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0041">41</xref><xref ref-type="aff" rid="AF0042">42</xref></contrib><contrib contrib-type="author"><name><surname>Wagner</surname><given-names>Ryan</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0041">41</xref><xref ref-type="aff" rid="AF0042">42</xref></contrib><contrib contrib-type="author"><name><surname>Herbst</surname><given-names>Kobus</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0043">43</xref><xref ref-type="aff" rid="AF0044">44</xref></contrib><contrib contrib-type="author"><name><surname>Mossong</surname><given-names>Jo&#x000eb;l</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0043">43</xref><xref ref-type="aff" rid="AF0044">44</xref><xref ref-type="aff" rid="AF0045">45</xref></contrib><contrib contrib-type="author"><name><surname>Emina</surname><given-names>Jacques B.O.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref></contrib><contrib contrib-type="author"><name><surname>Sankoh</surname><given-names>Osman A.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0046">46</xref><xref ref-type="aff" rid="AF0047">47</xref><xref ref-type="corresp" rid="cor1">*</xref></contrib><contrib contrib-type="author"><name><surname>Byass</surname><given-names>Peter</given-names></name><xref ref-type="aff" rid="AF0042">42</xref><xref ref-type="aff" rid="AF0048">48</xref></contrib>

Global Health Action

<p>Mortality from external causes &#x02013; whether unintentional (such as transport-related, falls, drowning, fires and burns, venoms, and poisons) or intentional (suicides and assaults) &#x02013; forms a worldwide phenomenon of considerable magnitude. Which cause categories dominate in particular places and which age-sex groups are most affected in particular populations vary widely. Fatalities due to external causes also present a non-trivial measurement issue, since instantaneous deaths in many settings are dealt with differently (e.g. by police and other authorities) as compared to deaths during or following medical treatment for injuries (typically in hospitals).</p><p>The Global Status Report on Road Safety 2013 (<xref rid="CIT0001" ref-type="bibr">1</xref>) reports over 1 million people killed on the world's roads annually, with numbers rising in some countries. Despite technological improvements in vehicles and roads, increasing traffic density can bring increased risks, particularly to pedestrians. The World Health Organization (WHO) African Region is estimated to have the highest rate of road traffic deaths, at 0.24 per 1,000 population, with the South-East Asia Region at 0.18 per 1,000 population.</p><p>Child injuries have also been documented globally in the World Report on Child Injury Prevention (<xref rid="CIT0002" ref-type="bibr">2</xref>). Globally, child injury deaths number close to one million per year, with the majority occurring in low- and middle-income countries. Leading cause categories are road traffic and drowning.</p><p>A review of data on suicide in Africa showed major gaps, making estimates of overall patterns uncertain (<xref rid="CIT0003" ref-type="bibr">3</xref>). Published rates from various African countries ranged from 0.004 to 0.17 per 1,000 population. A global analysis of suicide estimated a rate of 0.06 per 1,000 in the WHO African Region and 0.16 in the WHO South-East Asia Region (<xref rid="CIT0004" ref-type="bibr">4</xref>). The same source estimated rates for violence and war at 0.23 per 1,000 population in Africa and 0.08 per 1,000 in South-East Asia.</p><p> The INDEPTH Network works with Health and Demographic Surveillance Sites (HDSS) across Africa and Asia, which each follow circumscribed populations on a longitudinal basis. Core data collected include person-time at risk, together with deaths and, by means of verbal autopsy (VA), assessment of cause of death (<xref rid="CIT0005" ref-type="bibr">5</xref>). This allows reporting of external cause mortality on the basis of individually documented deaths within defined populations, adding considerably to existing overall estimates, which are often based on health facility data.</p><p>Our aim in this article is to document deaths among entire populations in a dataset from 22 INDEPTH HDSSs covering Africa and Asia, looking particularly at those deaths attributable to external causes. We define external causes here to include all of the WHO 2012 VA standard chapter 12 causes, corresponding to ICD-10 codes S00 to Y98 (<xref rid="CIT0006" ref-type="bibr">6</xref>). Although these 22 sites are not designed to be a representative sample, they enable comparisons to be made over widely differing situations, using standardised methods.</p><sec sec-type="methods" id="S0002"><title>Methods</title><p>The overall INDEPTH data set from which these analyses of external cause mortality are drawn is described in detail elsewhere (<xref rid="CIT0007" ref-type="bibr">7</xref>). Across the 22 participating sites (<xref rid="CIT0008" ref-type="bibr">8</xref>&#x02013;<xref rid="CIT0029" ref-type="bibr">29</xref>) , there is documentation on 111,910 deaths in 12,204,043 person-years of observation. These data are available in a public-domain data set (<xref rid="CIT0030" ref-type="bibr">30</xref>), and the methods used to compile that data set are summarised in <xref ref-type="boxed-text" rid="B0001">Box 1</xref>.</p><p> <italic>Box 1</italic>. Summary of methodology based on the detailed description in the introductory paper (<xref rid="CIT0007" ref-type="bibr">7</xref>)</p><boxed-text id="B0001" position="float"><p> <bold>Age&#x02013;sex&#x02013;time standardisation</bold> </p><p>To avoid effects of differences and changes in age-sex structures of populations, mortality fractions and rates have been adjusted using the INDEPTH 2013 population standard (<xref rid="CIT0031" ref-type="bibr">31</xref>). A weighting factor was calculated for each site, age group, sex, and year category in relation to the standard for the corresponding age group and sex, and incorporated into the overall data set. This is referred to in this article as <italic>age-sex-time standardisation</italic> in the contexts where it is used.</p><p> <bold>Cause of death assignment</bold> </p><p>The InterVA-4 (version 4.02) probabilistic model was used for all of the cause-of-death assignments in the overall data set (<xref rid="CIT0032" ref-type="bibr">32</xref>). InterVA-4 is fully compliant with the WHO 2012 Verbal Autopsy (VA) standard and generates causes of death categorised by ICD-10 groups (<xref rid="CIT0033" ref-type="bibr">33</xref>). The data reported here were collected before the WHO 2012 VA standard was available, but were transformed into the WHO2012 and InterVA-4 format to optimise cross-site standardisation in cause-of-death attribution. For a small proportion of deaths, VA interviews were not successfully completed; a few others contained inadequate information to arrive at a cause of death. InterVA-4 assigns causes of death (a maximum of three) with associated likelihoods; thus, cases for which likely causes did not total 100% were also assigned a residual indeterminate component. This served as a means of encapsulating uncertainty in cause of death at the individual level within the overall data set, as well as accounting for 100% of every death.</p><p> <bold>Overall dataset</bold> </p><p>The overall public-domain data set (<xref rid="CIT0030" ref-type="bibr">30</xref>) thus contains between one and four records for each death, with the sum of likelihoods for each individual being unity. Each record includes a specific cause of death, its likelihood, and its age-sex-time weighting. </p></boxed-text><p>Deaths assigned to any of the WHO 2012 VA cause-of-death categories relating to external causes (VA12.01 to VA12.99) were extracted from the overall database, together with details of site, age group at death, and sex. Person-year denominators corresponding to the same categories were included from the corresponding surveillance data.</p><p>Of the 22 sites reported in the data set, two (FilaBavi, Vietnam; Niakhar, Senegal) reported very few deaths due to external causes, accompanied by little specific information as to cause of death. These did not provide a credible picture of mortality from external causes, and consequently the following analyses are based on data from the remaining 20 sites, relating to 5,884 deaths over 11,828,253 person-years observed. The Karonga, Malawi, site did not contribute VAs for children. Sites reported for different time periods; overall, 5.0% of the person-time observed occurred before 2000, 28.2% from 2000 to 2005, and 66.7% from 2006 to 2012. As each HDSS covers a total population, rather than a sample, uncertainty intervals are not shown.</p><p>In this context, all of these data are secondary data sets derived from primary data collected separately by each participating site. In all cases, the primary data collection was covered by site-level ethical approvals relating to ongoing health and demographic surveillance in those specific locations. No individual identity or household location data were included in the secondary data, and no specific ethical approvals were required for these pooled analyses.</p></sec><sec sec-type="results" id="S0003"><title>Results</title><p> <xref ref-type="table" rid="T0001">Table 1</xref> shows the overall numbers of deaths from external causes and the exposure time for each site, by age group. <xref ref-type="fig" rid="F0001">Figure 1</xref> shows a map of the 20 sites, each one marked with its age-sex-time standardised overall mortality rate for deaths due to external causes, plus a note of the specific WHO 2012 VA external cause category and age group which accounted for the largest proportion of overall deaths from external causes. Approximately one-quarter of deaths due to external causes occurred in the under-15-year age group. External cause mortality at three of the Bangladeshi sites was dominated by drownings among small children, while elsewhere leading cause categories mainly comprised transport-related deaths, suicides, and assaults.</p><fig id="F0001" position="float"><label>Fig. 1</label><caption><p>Map showing overall age-sex-time standardised mortality rates per 1,000 person-years due to external causes, also listing the specific cause category and age group accounting for the largest proportion of deaths due to external causes at each site, for 20 INDEPTH sites.</p></caption><graphic xlink:href="GHA-7-25366-g001"/></fig><table-wrap id="T0001" position="float"><label>Table 1</label><caption><p>Numbers of deaths from external causes and person-years (py) of exposure, by age group, for 20 INDEPTH sites</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="center" rowspan="1" colspan="1"/><th align="center" colspan="2" rowspan="1">Infants</th><th align="center" colspan="2" rowspan="1">1&#x02013;4 years</th><th align="center" colspan="2" rowspan="1">5&#x02013;14 years</th><th align="center" colspan="2" rowspan="1">15&#x02013;49 years</th><th align="center" colspan="2" rowspan="1">50&#x02013;64 years</th><th align="center" colspan="2" rowspan="1">65+years</th></tr><tr><th align="center" rowspan="1" colspan="1"/><th colspan="12" rowspan="1"> <hr/> </th></tr><tr><th align="center" rowspan="1" colspan="1"/><th align="center" rowspan="1" colspan="1">Deaths</th><th align="center" rowspan="1" colspan="1">py</th><th align="center" rowspan="1" colspan="1">Deaths</th><th align="center" rowspan="1" colspan="1">py</th><th align="center" rowspan="1" colspan="1">Deaths</th><th align="center" rowspan="1" colspan="1">py</th><th align="center" rowspan="1" colspan="1">Deaths</th><th align="center" rowspan="1" colspan="1">py</th><th align="center" rowspan="1" colspan="1">Deaths</th><th align="center" rowspan="1" colspan="1">py</th><th align="center" rowspan="1" colspan="1">Deaths</th><th align="center" rowspan="1" colspan="1">py</th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">Bangladesh: Matlab</td><td align="center" rowspan="1" colspan="1">12.4</td><td align="center" rowspan="1" colspan="1">41,792</td><td align="center" rowspan="1" colspan="1">242.5</td><td align="center" rowspan="1" colspan="1">167,334</td><td align="center" rowspan="1" colspan="1">88.7</td><td align="center" rowspan="1" colspan="1">401,272</td><td align="center" rowspan="1" colspan="1">202.3</td><td align="center" rowspan="1" colspan="1">886,951</td><td align="center" rowspan="1" colspan="1">56.4</td><td align="center" rowspan="1" colspan="1">189,069</td><td align="center" rowspan="1" colspan="1">78.7</td><td align="center" rowspan="1" colspan="1">108,061</td></tr><tr><td align="left" rowspan="1" colspan="1">Bangladesh: Bandarban</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1,242</td><td align="center" rowspan="1" colspan="1">2.0</td><td align="center" rowspan="1" colspan="1">5,770</td><td align="center" rowspan="1" colspan="1">2.0</td><td align="center" rowspan="1" colspan="1">13,626</td><td align="center" rowspan="1" colspan="1">10.5</td><td align="center" rowspan="1" colspan="1">30,173</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">5,891</td><td align="center" rowspan="1" colspan="1">2.0</td><td align="center" rowspan="1" colspan="1">2,705</td></tr><tr><td align="left" rowspan="1" colspan="1">Bangladesh: Chakaria</td><td align="center" rowspan="1" colspan="1">4.8</td><td align="center" rowspan="1" colspan="1">5,636</td><td align="center" rowspan="1" colspan="1">29.5</td><td align="center" rowspan="1" colspan="1">21,992</td><td align="center" rowspan="1" colspan="1">20.7</td><td align="center" rowspan="1" colspan="1">60,951</td><td align="center" rowspan="1" colspan="1">15.8</td><td align="center" rowspan="1" colspan="1">104,097</td><td align="center" rowspan="1" colspan="1">5.8</td><td align="center" rowspan="1" colspan="1">16,234</td><td align="center" rowspan="1" colspan="1">9.9</td><td align="center" rowspan="1" colspan="1">8,257</td></tr><tr><td align="left" rowspan="1" colspan="1">Bangladesh: AMK</td><td align="center" rowspan="1" colspan="1">2.4</td><td align="center" rowspan="1" colspan="1">10,558</td><td align="center" rowspan="1" colspan="1">60.9</td><td align="center" rowspan="1" colspan="1">43,236</td><td align="center" rowspan="1" colspan="1">25.5</td><td align="center" rowspan="1" colspan="1">105,701</td><td align="center" rowspan="1" colspan="1">112.1</td><td align="center" rowspan="1" colspan="1">274,129</td><td align="center" rowspan="1" colspan="1">20.2</td><td align="center" rowspan="1" colspan="1">53,184</td><td align="center" rowspan="1" colspan="1">25.4</td><td align="center" rowspan="1" colspan="1">26,927</td></tr><tr><td align="left" rowspan="1" colspan="1">Burkina Faso: Nouna</td><td align="center" rowspan="1" colspan="1">13.7</td><td align="center" rowspan="1" colspan="1">30,362</td><td align="center" rowspan="1" colspan="1">37.8</td><td align="center" rowspan="1" colspan="1">105,185</td><td align="center" rowspan="1" colspan="1">50.0</td><td align="center" rowspan="1" colspan="1">181,699</td><td align="center" rowspan="1" colspan="1">91.8</td><td align="center" rowspan="1" colspan="1">275,936</td><td align="center" rowspan="1" colspan="1">30.0</td><td align="center" rowspan="1" colspan="1">47,682</td><td align="center" rowspan="1" colspan="1">44.6</td><td align="center" rowspan="1" colspan="1">27,722</td></tr><tr><td align="left" rowspan="1" colspan="1">Burkina Faso: Ouagadougou</td><td align="center" rowspan="1" colspan="1">0.9</td><td align="center" rowspan="1" colspan="1">6,943</td><td align="center" rowspan="1" colspan="1">3.0</td><td align="center" rowspan="1" colspan="1">27,941</td><td align="center" rowspan="1" colspan="1">6.5</td><td align="center" rowspan="1" colspan="1">51,217</td><td align="center" rowspan="1" colspan="1">17.1</td><td align="center" rowspan="1" colspan="1">119,468</td><td align="center" rowspan="1" colspan="1">6.6</td><td align="center" rowspan="1" colspan="1">11,459</td><td align="center" rowspan="1" colspan="1">8.3</td><td align="center" rowspan="1" colspan="1">4,149</td></tr><tr><td align="left" rowspan="1" colspan="1">C&#x000f4;te d'Ivoire: Taabo</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">3,962</td><td align="center" rowspan="1" colspan="1">3.0</td><td align="center" rowspan="1" colspan="1">12,951</td><td align="center" rowspan="1" colspan="1">6.9</td><td align="center" rowspan="1" colspan="1">30,967</td><td align="center" rowspan="1" colspan="1">14.1</td><td align="center" rowspan="1" colspan="1">48,484</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">6,967</td><td align="center" rowspan="1" colspan="1">3.4</td><td align="center" rowspan="1" colspan="1">3,173</td></tr><tr><td align="left" rowspan="1" colspan="1">Ethiopia: Kilite Awlaelo</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">3,185</td><td align="center" rowspan="1" colspan="1">1.5</td><td align="center" rowspan="1" colspan="1">13,009</td><td align="center" rowspan="1" colspan="1">11.3</td><td align="center" rowspan="1" colspan="1">39,917</td><td align="center" rowspan="1" colspan="1">16.4</td><td align="center" rowspan="1" colspan="1">59,397</td><td align="center" rowspan="1" colspan="1">4.6</td><td align="center" rowspan="1" colspan="1">11,173</td><td align="center" rowspan="1" colspan="1">6.9</td><td align="center" rowspan="1" colspan="1">7,125</td></tr><tr><td align="left" rowspan="1" colspan="1">The Gambia: Farafenni</td><td align="center" rowspan="1" colspan="1">1.6</td><td align="center" rowspan="1" colspan="1">11,438</td><td align="center" rowspan="1" colspan="1">3.4</td><td align="center" rowspan="1" colspan="1">42,802</td><td align="center" rowspan="1" colspan="1">8.1</td><td align="center" rowspan="1" colspan="1">88,740</td><td align="center" rowspan="1" colspan="1">21.4</td><td align="center" rowspan="1" colspan="1">139,746</td><td align="center" rowspan="1" colspan="1">5.9</td><td align="center" rowspan="1" colspan="1">22,485</td><td align="center" rowspan="1" colspan="1">15.8</td><td align="center" rowspan="1" colspan="1">11,506</td></tr><tr><td align="left" rowspan="1" colspan="1">Ghana: Navrongo</td><td align="center" rowspan="1" colspan="1">19.2</td><td align="center" rowspan="1" colspan="1">30,124</td><td align="center" rowspan="1" colspan="1">52.3</td><td align="center" rowspan="1" colspan="1">116,283</td><td align="center" rowspan="1" colspan="1">119.3</td><td align="center" rowspan="1" colspan="1">296,767</td><td align="center" rowspan="1" colspan="1">314.7</td><td align="center" rowspan="1" colspan="1">534,464</td><td align="center" rowspan="1" colspan="1">140.7</td><td align="center" rowspan="1" colspan="1">128,494</td><td align="center" rowspan="1" colspan="1">226.6</td><td align="center" rowspan="1" colspan="1">70,664</td></tr><tr><td align="left" rowspan="1" colspan="1">Ghana: Dodowa</td><td align="center" rowspan="1" colspan="1">1.9</td><td align="center" rowspan="1" colspan="1">14,120</td><td align="center" rowspan="1" colspan="1">9.9</td><td align="center" rowspan="1" colspan="1">58,318</td><td align="center" rowspan="1" colspan="1">19.9</td><td align="center" rowspan="1" colspan="1">138,762</td><td align="center" rowspan="1" colspan="1">91.6</td><td align="center" rowspan="1" colspan="1">255,677</td><td align="center" rowspan="1" colspan="1">24.8</td><td align="center" rowspan="1" colspan="1">37,001</td><td align="center" rowspan="1" colspan="1">32.7</td><td align="center" rowspan="1" colspan="1">27,227</td></tr><tr><td align="left" rowspan="1" colspan="1">India: Ballabgarh</td><td align="center" rowspan="1" colspan="1">4.0</td><td align="center" rowspan="1" colspan="1">8,405</td><td align="center" rowspan="1" colspan="1">12.9</td><td align="center" rowspan="1" colspan="1">30,478</td><td align="center" rowspan="1" colspan="1">17.3</td><td align="center" rowspan="1" colspan="1">77,584</td><td align="center" rowspan="1" colspan="1">165.0</td><td align="center" rowspan="1" colspan="1">194,902</td><td align="center" rowspan="1" colspan="1">27.8</td><td align="center" rowspan="1" colspan="1">30,823</td><td align="center" rowspan="1" colspan="1">32.0</td><td align="center" rowspan="1" colspan="1">15,597</td></tr><tr><td align="left" rowspan="1" colspan="1">India: Vadu</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">4,285</td><td align="center" rowspan="1" colspan="1">0.0</td><td align="center" rowspan="1" colspan="1">16,484</td><td align="center" rowspan="1" colspan="1">2.0</td><td align="center" rowspan="1" colspan="1">33,973</td><td align="center" rowspan="1" colspan="1">49.7</td><td align="center" rowspan="1" colspan="1">128,387</td><td align="center" rowspan="1" colspan="1">11.4</td><td align="center" rowspan="1" colspan="1">15,518</td><td align="center" rowspan="1" colspan="1">15.8</td><td align="center" rowspan="1" colspan="1">7,469</td></tr><tr><td align="left" rowspan="1" colspan="1">Indonesia: Purworejo</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">2,845</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">14,350</td><td align="center" rowspan="1" colspan="1">2.6</td><td align="center" rowspan="1" colspan="1">44,166</td><td align="center" rowspan="1" colspan="1">16.4</td><td align="center" rowspan="1" colspan="1">136,422</td><td align="center" rowspan="1" colspan="1">6.7</td><td align="center" rowspan="1" colspan="1">27,091</td><td align="center" rowspan="1" colspan="1">3.2</td><td align="center" rowspan="1" colspan="1">21,793</td></tr><tr><td align="left" rowspan="1" colspan="1">Kenya: Kilifi</td><td align="center" rowspan="1" colspan="1">3.0</td><td align="center" rowspan="1" colspan="1">38,526</td><td align="center" rowspan="1" colspan="1">13.5</td><td align="center" rowspan="1" colspan="1">147,331</td><td align="center" rowspan="1" colspan="1">41.8</td><td align="center" rowspan="1" colspan="1">310,584</td><td align="center" rowspan="1" colspan="1">169.2</td><td align="center" rowspan="1" colspan="1">422,507</td><td align="center" rowspan="1" colspan="1">61.6</td><td align="center" rowspan="1" colspan="1">65,606</td><td align="center" rowspan="1" colspan="1">86.1</td><td align="center" rowspan="1" colspan="1">33,092</td></tr><tr><td align="left" rowspan="1" colspan="1">Kenya: Kisumu</td><td align="center" rowspan="1" colspan="1">21.3</td><td align="center" rowspan="1" colspan="1">39,887</td><td align="center" rowspan="1" colspan="1">57.6</td><td align="center" rowspan="1" colspan="1">144,451</td><td align="center" rowspan="1" colspan="1">41.6</td><td align="center" rowspan="1" colspan="1">324,153</td><td align="center" rowspan="1" colspan="1">202.2</td><td align="center" rowspan="1" colspan="1">467,691</td><td align="center" rowspan="1" colspan="1">60.5</td><td align="center" rowspan="1" colspan="1">89,105</td><td align="center" rowspan="1" colspan="1">73.5</td><td align="center" rowspan="1" colspan="1">67,080</td></tr><tr><td align="left" rowspan="1" colspan="1">Kenya: Nairobi</td><td align="center" rowspan="1" colspan="1">11.9</td><td align="center" rowspan="1" colspan="1">14,350</td><td align="center" rowspan="1" colspan="1">22.0</td><td align="center" rowspan="1" colspan="1">62,552</td><td align="center" rowspan="1" colspan="1">22.2</td><td align="center" rowspan="1" colspan="1">108,651</td><td align="center" rowspan="1" colspan="1">354.7</td><td align="center" rowspan="1" colspan="1">383,810</td><td align="center" rowspan="1" colspan="1">23.6</td><td align="center" rowspan="1" colspan="1">24,804</td><td align="center" rowspan="1" colspan="1">10.6</td><td align="center" rowspan="1" colspan="1">5,640</td></tr><tr><td align="left" rowspan="1" colspan="1">Malawi: Karonga</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">41.0</td><td align="center" rowspan="1" colspan="1">117,499</td><td align="center" rowspan="1" colspan="1">11.5</td><td align="center" rowspan="1" colspan="1">14,783</td><td align="center" rowspan="1" colspan="1">15.5</td><td align="center" rowspan="1" colspan="1">11,356</td></tr><tr><td align="left" rowspan="1" colspan="1">South Africa: Agincourt</td><td align="center" rowspan="1" colspan="1">8.4</td><td align="center" rowspan="1" colspan="1">36,811</td><td align="center" rowspan="1" colspan="1">28.3</td><td align="center" rowspan="1" colspan="1">148,961</td><td align="center" rowspan="1" colspan="1">58.3</td><td align="center" rowspan="1" colspan="1">369,285</td><td align="center" rowspan="1" colspan="1">565.5</td><td align="center" rowspan="1" colspan="1">725,431</td><td align="center" rowspan="1" colspan="1">90.4</td><td align="center" rowspan="1" colspan="1">92,519</td><td align="center" rowspan="1" colspan="1">65.3</td><td align="center" rowspan="1" colspan="1">63,187</td></tr><tr><td align="left" rowspan="1" colspan="1">South Africa: Africa Centre</td><td align="center" rowspan="1" colspan="1">7.3</td><td align="center" rowspan="1" colspan="1">22,468</td><td align="center" rowspan="1" colspan="1">34.4</td><td align="center" rowspan="1" colspan="1">91,367</td><td align="center" rowspan="1" colspan="1">69.8</td><td align="center" rowspan="1" colspan="1">232,962</td><td align="center" rowspan="1" colspan="1">544.8</td><td align="center" rowspan="1" colspan="1">374,099</td><td align="center" rowspan="1" colspan="1">92.3</td><td align="center" rowspan="1" colspan="1">54,852</td><td align="center" rowspan="1" colspan="1">87.7</td><td align="center" rowspan="1" colspan="1">39,160</td></tr></tbody></table></table-wrap><p> <xref ref-type="fig" rid="F0002">Figure 2</xref> shows the breakdown of overall external cause mortality age-sex-time standardised rates by cause category and site. At the Nouna, Burkina Faso, site, almost all external cause deaths were attributed to transport-related causes (possibly through the use of an historic VA instrument that did not contain all of the WHO 2012 VA items). Elsewhere, there were similar mixes of cause categories between sites, with some local variations.</p><fig id="F0002" position="float"><label>Fig. 2</label><caption><p>Age-sex-time standardised mortality rates per 1,000 person-years by category of external causes of death, from 20 INDEPTH sites.</p></caption><graphic xlink:href="GHA-7-25366-g002"/></fig><p> <xref ref-type="table" rid="T0002">Table 2</xref> shows age-sex-time standardised cause-specific mortality rates by cause category, sex, and site for adults (aged 15 years and older). Men were at higher risk of transport-related death than women at every site. Suicides were most common in Bangladesh, particularly among women; in Eastern and Southern Africa, they were more common among men. Sites in Western Africa generally recorded low rates of suicide. South African men were subject to high rates of death following assault.</p><table-wrap id="T0002" position="float"><label>Table 2</label><caption><p>Age-sex-time standardised mortality rates per 1,000 person-years for adults (aged 15 years and older), by sex and category of external causes of death, for 20 INDEPTH sites</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1"/><th align="center" colspan="2" rowspan="1">Transport</th><th align="center" colspan="2" rowspan="1">Falls</th><th align="center" colspan="2" rowspan="1">Drowning</th><th align="center" colspan="2" rowspan="1">Fire and burns</th><th align="center" colspan="2" rowspan="1">Venom and poison</th><th align="center" colspan="2" rowspan="1">Suicide</th><th align="center" colspan="2" rowspan="1">Assault</th><th align="center" colspan="2" rowspan="1">Other</th></tr><tr><th align="left" rowspan="1" colspan="1"/><th colspan="16" rowspan="1"> <hr/> </th></tr><tr><th align="left" rowspan="1" colspan="1">Site</th><th align="center" rowspan="1" colspan="1">Male</th><th align="center" rowspan="1" colspan="1">Female</th><th align="center" rowspan="1" colspan="1">Male</th><th align="center" rowspan="1" colspan="1">Female</th><th align="center" rowspan="1" colspan="1">Male</th><th align="center" rowspan="1" colspan="1">Female</th><th align="center" rowspan="1" colspan="1">Male</th><th align="center" rowspan="1" colspan="1">Female</th><th align="center" rowspan="1" colspan="1">Male</th><th align="center" rowspan="1" colspan="1">Female</th><th align="center" rowspan="1" colspan="1">Male</th><th align="center" rowspan="1" colspan="1">Female</th><th align="center" rowspan="1" colspan="1">Male</th><th align="center" rowspan="1" colspan="1">Female</th><th align="center" rowspan="1" colspan="1">Male</th><th align="center" rowspan="1" colspan="1">Female</th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">Bangladesh: Matlab</td><td align="center" rowspan="1" colspan="1">0.12</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.01</td></tr><tr><td align="left" rowspan="1" colspan="1">Bangladesh: Bandarban</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">0.18</td><td align="center" rowspan="1" colspan="1">0.19</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">Bangladesh: Chakaria</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.04</td></tr><tr><td align="left" rowspan="1" colspan="1">Bangladesh: AMK</td><td align="center" rowspan="1" colspan="1">0.19</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.15</td><td align="center" rowspan="1" colspan="1">0.21</td><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">Burkina Faso: Nouna</td><td align="center" rowspan="1" colspan="1">0.62</td><td align="center" rowspan="1" colspan="1">0.38</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.03</td></tr><tr><td align="left" rowspan="1" colspan="1">Burkina Faso: Ouagadougou</td><td align="center" rowspan="1" colspan="1">0.42</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.01</td></tr><tr><td align="left" rowspan="1" colspan="1">C&#x000f4;te d'Ivoire: Taabo</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.15</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.04</td></tr><tr><td align="left" rowspan="1" colspan="1">Ethiopia: Kilite Awlaelo</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.15</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.14</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.05</td></tr><tr><td align="left" rowspan="1" colspan="1">The Gambia: Farafenni</td><td align="center" rowspan="1" colspan="1">0.17</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.01</td></tr><tr><td align="left" rowspan="1" colspan="1">Ghana: Navrongo</td><td align="center" rowspan="1" colspan="1">0.41</td><td align="center" rowspan="1" colspan="1">0.12</td><td align="center" rowspan="1" colspan="1">0.32</td><td align="center" rowspan="1" colspan="1">0.25</td><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.13</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.12</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">0.03</td></tr><tr><td align="left" rowspan="1" colspan="1">Ghana: Dodowa</td><td align="center" rowspan="1" colspan="1">0.39</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.01</td></tr><tr><td align="left" rowspan="1" colspan="1">India: Ballabgarh</td><td align="center" rowspan="1" colspan="1">0.33</td><td align="center" rowspan="1" colspan="1">0.26</td><td align="center" rowspan="1" colspan="1">0.14</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.24</td><td align="center" rowspan="1" colspan="1">0.21</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.02</td></tr><tr><td align="left" rowspan="1" colspan="1">India: Vadu</td><td align="center" rowspan="1" colspan="1">0.36</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">0.12</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">Indonesia: Purworejo</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">Kenya: Kilifi</td><td align="center" rowspan="1" colspan="1">0.33</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.22</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.12</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="left" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.20</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.53</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">Kenya: Kisumu</td><td align="center" rowspan="1" colspan="1">0.21</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.19</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.34</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.01</td></tr><tr><td align="left" rowspan="1" colspan="1">Kenya: Nairobi</td><td align="center" rowspan="1" colspan="1">0.71</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.17</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.26</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.66</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">0.01</td></tr><tr><td align="left" rowspan="1" colspan="1">Malawi: Karonga</td><td align="center" rowspan="1" colspan="1">0.26</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.18</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.23</td><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">0.14</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.01</td></tr><tr><td align="left" rowspan="1" colspan="1">South Africa: Africa Centre</td><td align="center" rowspan="1" colspan="1">0.89</td><td align="center" rowspan="1" colspan="1">0.13</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.39</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">2.01</td><td align="center" rowspan="1" colspan="1">0.30</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.01</td></tr><tr><td align="left" rowspan="1" colspan="1">South Africa: Agincourt</td><td align="center" rowspan="1" colspan="1">0.38</td><td align="center" rowspan="1" colspan="1">0.12</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">0.12</td><td align="center" rowspan="1" colspan="1">0.48</td><td align="center" rowspan="1" colspan="1">0.17</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.02</td></tr></tbody></table></table-wrap><p> <xref ref-type="table" rid="T0003">Table 3</xref> shows, in the same format, age-sex-time standardised cause-specific mortality rates for children. Boys generally experienced higher rates of transport-related mortality than girls, although they were lower rates than for adults. At most sites, drowning occurred at higher rates among boys.</p><table-wrap id="T0003" position="float"><label>Table 3</label><caption><p>Age-sex-time standardised mortality rates per 1,000 person-years for children (aged under 15 years), by sex and category of external causes of death, for 19 INDEPTH sites</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1"/><th align="center" colspan="2" rowspan="1">Transport</th><th align="center" colspan="2" rowspan="1">Falls</th><th align="center" colspan="2" rowspan="1">Drowning</th><th align="center" colspan="2" rowspan="1">Fire and Burns</th><th align="center" colspan="2" rowspan="1">Venom and Poison</th><th align="center" colspan="2" rowspan="1">Suicide</th><th align="center" colspan="2" rowspan="1">Assault</th><th align="center" colspan="2" rowspan="1">Other</th></tr><tr><th align="left" rowspan="1" colspan="1"/><th colspan="16" rowspan="1"> <hr/> </th></tr><tr><th align="left" rowspan="1" colspan="1">Site</th><th align="center" rowspan="1" colspan="1">Male</th><th align="center" rowspan="1" colspan="1">Female</th><th align="center" rowspan="1" colspan="1">Male</th><th align="center" rowspan="1" colspan="1">Female</th><th align="center" rowspan="1" colspan="1">Male</th><th align="center" rowspan="1" colspan="1">Female</th><th align="center" rowspan="1" colspan="1">Male</th><th align="center" rowspan="1" colspan="1">Female</th><th align="center" rowspan="1" colspan="1">Male</th><th align="center" rowspan="1" colspan="1">Female</th><th align="center" rowspan="1" colspan="1">Male</th><th align="center" rowspan="1" colspan="1">Female</th><th align="center" rowspan="1" colspan="1">Male</th><th align="center" rowspan="1" colspan="1">Female</th><th align="center" rowspan="1" colspan="1">Male</th><th align="center" rowspan="1" colspan="1">Female</th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">Bangladesh: Matlab</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.65</td><td align="center" rowspan="1" colspan="1">0.47</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.02</td></tr><tr><td align="left" rowspan="1" colspan="1">Bangladesh: Bandarban</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.25</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.12</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">Bangladesh: Chakaria</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.61</td><td align="center" rowspan="1" colspan="1">0.48</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.06</td></tr><tr><td align="left" rowspan="1" colspan="1">Bangladesh: AMK</td><td align="center" rowspan="1" colspan="1">0.12</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.59</td><td align="center" rowspan="1" colspan="1">0.46</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.01</td></tr><tr><td align="left" rowspan="1" colspan="1">Burkina Faso: Nouna</td><td align="center" rowspan="1" colspan="1">0.36</td><td align="center" rowspan="1" colspan="1">0.17</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.02</td></tr><tr><td align="left" rowspan="1" colspan="1">Burkina Faso: Ouagadougou</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.16</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">C&#x000f4;te d'Ivoire: Taabo</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">Ethiopia: Kilite Awlaelo</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">0.07</td></tr><tr><td align="left" rowspan="1" colspan="1">The Gambia: Farafenni</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.02</td></tr><tr><td align="left" rowspan="1" colspan="1">Ghana: Navrongo</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.13</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.29</td><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.02</td></tr><tr><td align="left" rowspan="1" colspan="1">Ghana: Dodowa</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">India: Ballabgarh</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.13</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.06</td></tr><tr><td align="left" rowspan="1" colspan="1">India: Vadu</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">Indonesia: Purworejo</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">Kenya: Kilifi</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">Kenya: Kisumu</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.01</td></tr><tr><td align="left" rowspan="1" colspan="1">Kenya: Nairobi</td><td align="center" rowspan="1" colspan="1">0.16</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.13</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.09</td></tr><tr><td align="left" rowspan="1" colspan="1">South Africa: Africa Centre</td><td align="center" rowspan="1" colspan="1">0.12</td><td align="center" rowspan="1" colspan="1">0.12</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.02</td></tr><tr><td align="left" rowspan="1" colspan="1">South Africa: Agincourt</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.02</td></tr></tbody></table></table-wrap><p> <xref ref-type="fig" rid="F0003">Figure 3</xref> shows site-specific mortality rates by categories of unintentional external causes and age group, with rates for all sites shown on the same logarithmic scale for ease of comparison. <xref ref-type="fig" rid="F0004">Figure 4</xref> shows intentional external causes on the same basis.</p><fig id="F0003" position="float"><label>Fig. 3</label><caption><p>Site-specific mortality rates per 1,000 person-years by age group and category of unintentional external causes of death.</p></caption><graphic xlink:href="GHA-7-25366-g003"/></fig><fig id="F0004" position="float"><label>Fig. 4</label><caption><p>Site-specific mortality rates per 1,000 person-years by age group and category of intentional external causes of death.</p></caption><graphic xlink:href="GHA-7-25366-g004"/></fig></sec><sec sec-type="discussion" id="S0004"><title>Discussion</title><p>As was clear from the overview of this cause-specific mortality data set (<xref rid="CIT0007" ref-type="bibr">7</xref>), deaths due to external causes form an important component of overall mortality, and in particular account for many premature deaths, in both childhood and early adulthood. The major advantage of addressing external cause mortality from this data set, which included all deaths within circumscribed surveillance populations, is that various biases from attempting to capture injury data alone were avoided. Most obviously, it avoids the difficulties of accounting for both instantaneous fatalities and health facility deaths, which otherwise involves trying to combine diverse reporting mechanisms.</p><p>Patterns of external cause mortality revealed from these analyses were more or less consistent with the relatively few other direct measurements from Sub-Saharan Africa and South-East Asia. It is clear that geographic location, age, and sex are major determinants not only of overall external cause mortality but also of specific cause categories. In some cases, geography appeared to play a direct role, for example in the problematically high rates of child drowning in Bangladesh. At the Bandarban site, some 200 m above sea level near the Myanmar border, drowning rates were appreciably lower than in the flat river delta environments of the other three sites in Bangladesh. Similarly, in the mountainous area covered by the Kilite Awlaelo, Ethiopia, site, falling was the major cause of death. It is important to be clear that the WHO 2012 VA standard and the InterVA-4 model are designed for assigning causes of death, and not mechanisms of injury, which are consequently not discussed here.</p><p> For suicide, rates were high among Bangladeshi women, whereas in Eastern and Southern Africa rates were high among men; suicide overall was much less common in Western Africa. South Africa and Kenya showed appreciably higher rates of assault-related deaths than other countries reporting here. Countries with poorly developed road transport infrastructures, for example in Western Africa, emerged clearly with high rates of transport-related mortality.</p><p>It is sometimes assumed that external causes of death represent a relatively easy option for assigning cause of death via VA. This may be true for a proportion of deaths from external causes, for example instantaneous fatalities with no complicating factors. This assumes, however, that all deaths from external causes have reliable witnesses who can be traced for VA interviews, and who, in the case of inflicted injuries, were not the perpetrators. In this study, based on VA material derived via a variety of antecedents to the WHO 2012 VA standards, there may also have been some difficulties in extracting all the necessary data items correctly, particularly for details of injuries contained in narratives. This probably led to artefacts with road transport deaths in the Nouna, Burkina Faso, site. A few deaths at the Karonga, Malawi, site were incorrectly attributed to burns only on the basis of skin symptoms.</p><p>However, there may also be cases where not all is as it seems at first sight, and the details of these may be difficult to ascertain from VA interviews. It has been suggested that suicide rates are actually correlated with autopsy rates; in other words, methods of assigning cause of death are important, particularly when complex and sensitive issues may be involved (<xref rid="CIT0034" ref-type="bibr">34</xref>). Using VA, it is very likely, for example, that fatal injuries involving a motor vehicle will be attributed to road traffic deaths, even though motor vehicles can be used as weapons of assault or instruments of suicide.</p></sec><sec sec-type="conclusions" id="S0005"><title>Conclusions</title><p>The patterns of external cause mortality presented here generally conform to expectations, but at the same time they provide detail to fill in some of the gaps in knowledge about deaths arising from injuries of various kinds in Africa and Asia. Clearly, many of the specific mortality burdens identified must be considered as in principle being largely avoidable, given that they do not happen uniformly across locations and population groups. However, preventing external cause mortality poses major challenges involving social, behavioural, environmental, and regulatory considerations. Nevertheless, documenting the major targets for prevention is an important prerequisite.</p></sec>

Cause-specific childhood mortality in Africa and Asia: evidence from INDEPTH health and demographic surveillance system sites

<sec id="st1"><title> Background</title><p>Childhood mortality, particularly in the first 5 years of life, is a major global concern and the target of Millennium Development Goal 4. Although the majority of childhood deaths occur in Africa and Asia, these are also the regions where such deaths are least likely to be registered. The INDEPTH Network works to alleviate this problem by collating detailed individual data from defined Health and Demographic Surveillance sites. By registering deaths and carrying out verbal autopsies to determine cause of death across many such sites, using standardised methods, the Network seeks to generate population-based mortality statistics that are not otherwise available.</p></sec><sec id="st2"><title>Objective</title><p>To present a description of cause-specific mortality rates and fractions over the first 15 years of life as documented by INDEPTH Network sites in sub-Saharan Africa and south-east Asia.</p></sec><sec id="st3"><title>Design</title><p>All childhood deaths at INDEPTH sites are routinely registered and followed up with verbal autopsy (VA) interviews. For this study, VA archives were transformed into the WHO 2012 VA standard format and processed using the InterVA-4 model to assign cause of death. Routine surveillance data also provided person-time denominators for mortality rates. Cause-specific mortality rates and cause-specific mortality fractions are presented according to WHO 2012 VA cause groups for neonatal, infant, 1&#x02013;4 year and 5&#x02013;14 year age groups.</p></sec><sec id="st4"><title>Results</title><p>A total of 28,751 childhood deaths were documented during 4,387,824 person-years over 18 sites. Infant mortality ranged from 11 to 78 per 1,000 live births, with under-5 mortality from 15 to 152 per 1,000 live births. Sites in Vietnam and Kenya accounted for the lowest and highest mortality rates reported.</p></sec><sec id="st5"><title>Conclusions</title><p>Many children continue to die from relatively preventable causes, particularly in areas with high rates of malaria and HIV/AIDS. Neonatal mortality persists at relatively high, and perhaps sometimes under-documented, rates. External causes of death are a significant childhood problem in some settings.</p></sec>

<contrib contrib-type="author"><name><surname>Streatfield</surname><given-names>P. Kim</given-names></name><xref ref-type="aff" rid="AF0001">1</xref><xref ref-type="aff" rid="AF0002">2</xref><xref ref-type="aff" rid="AF0003">3</xref></contrib><contrib contrib-type="author"><name><surname>Khan</surname><given-names>Wasif A.</given-names></name><xref ref-type="aff" rid="AF0002">2</xref><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0004">4</xref></contrib><contrib contrib-type="author"><name><surname>Bhuiya</surname><given-names>Abbas</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0005">5</xref><xref ref-type="aff" rid="AF0006">6</xref></contrib><contrib contrib-type="author"><name><surname>Hanifi</surname><given-names>Syed M.A.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0005">5</xref><xref ref-type="aff" rid="AF0006">6</xref></contrib><contrib contrib-type="author"><name><surname>Alam</surname><given-names>Nurul</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0007">7</xref><xref ref-type="aff" rid="AF0008">8</xref></contrib><contrib contrib-type="author"><name><surname>Ouattara</surname><given-names>Mamadou</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0009">9</xref><xref ref-type="aff" rid="AF0010">10</xref></contrib><contrib contrib-type="author"><name><surname>Sanou</surname><given-names>Aboubakary</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0009">9</xref><xref ref-type="aff" rid="AF0010">10</xref></contrib><contrib contrib-type="author"><name><surname>Si&#x000e9;</surname><given-names>Ali</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0009">9</xref><xref ref-type="aff" rid="AF0010">10</xref></contrib><contrib contrib-type="author"><name><surname>Lankoand&#x000e9;</surname><given-names>Bruno</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0011">11</xref><xref ref-type="aff" rid="AF0012">12</xref></contrib><contrib contrib-type="author"><name><surname>Soura</surname><given-names>Abdramane B.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0011">11</xref><xref ref-type="aff" rid="AF0012">12</xref></contrib><contrib contrib-type="author"><name><surname>Bonfoh</surname><given-names>Bassirou</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0013">13</xref><xref ref-type="aff" rid="AF0014">14</xref></contrib><contrib contrib-type="author"><name><surname>Jaeger</surname><given-names>Fabienne</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0013">13</xref><xref ref-type="aff" rid="AF0015">15</xref></contrib><contrib contrib-type="author"><name><surname>Ngoran</surname><given-names>Eliezer K.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0013">13</xref><xref ref-type="aff" rid="AF0016">16</xref></contrib><contrib contrib-type="author"><name><surname>Utzinger</surname><given-names>Juerg</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0013">13</xref><xref ref-type="aff" rid="AF0015">15</xref></contrib><contrib contrib-type="author"><name><surname>Abreha</surname><given-names>Loko</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0017">17</xref><xref ref-type="aff" rid="AF0018">18</xref></contrib><contrib contrib-type="author"><name><surname>Melaku</surname><given-names>Yohannes A.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0017">17</xref><xref ref-type="aff" rid="AF0019">19</xref></contrib><contrib contrib-type="author"><name><surname>Weldearegawi</surname><given-names>Berhe</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0017">17</xref><xref ref-type="aff" rid="AF0019">19</xref></contrib><contrib contrib-type="author"><name><surname>Ansah</surname><given-names>Akosua</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0020">20</xref><xref ref-type="aff" rid="AF0021">21</xref></contrib><contrib contrib-type="author"><name><surname>Hodgson</surname><given-names>Abraham</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0020">20</xref><xref ref-type="aff" rid="AF0021">21</xref></contrib><contrib contrib-type="author"><name><surname>Oduro</surname><given-names>Abraham</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0020">20</xref><xref ref-type="aff" rid="AF0021">21</xref></contrib><contrib contrib-type="author"><name><surname>Welaga</surname><given-names>Paul</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0020">20</xref><xref ref-type="aff" rid="AF0021">21</xref></contrib><contrib contrib-type="author"><name><surname>Gyapong</surname><given-names>Margaret</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0022">22</xref><xref ref-type="aff" rid="AF0023">23</xref></contrib><contrib contrib-type="author"><name><surname>Narh</surname><given-names>Clement T.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0022">22</xref><xref ref-type="aff" rid="AF0023">23</xref></contrib><contrib contrib-type="author"><name><surname>Narh-Bana</surname><given-names>Solomon A.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0022">22</xref><xref ref-type="aff" rid="AF0023">23</xref></contrib><contrib contrib-type="author"><name><surname>Kant</surname><given-names>Shashi</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0024">24</xref><xref ref-type="aff" rid="AF0025">25</xref></contrib><contrib contrib-type="author"><name><surname>Misra</surname><given-names>Puneet</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0024">24</xref><xref ref-type="aff" rid="AF0025">25</xref></contrib><contrib contrib-type="author"><name><surname>Rai</surname><given-names>Sanjay K.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0024">24</xref><xref ref-type="aff" rid="AF0025">25</xref></contrib><contrib contrib-type="author"><name><surname>Bauni</surname><given-names>Evasius</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0026">26</xref><xref ref-type="aff" rid="AF0027">27</xref></contrib><contrib contrib-type="author"><name><surname>Mochamah</surname><given-names>George</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0026">26</xref><xref ref-type="aff" rid="AF0027">27</xref></contrib><contrib contrib-type="author"><name><surname>Ndila</surname><given-names>Carolyne</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0026">26</xref><xref ref-type="aff" rid="AF0027">27</xref></contrib><contrib contrib-type="author"><name><surname>Williams</surname><given-names>Thomas N.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0026">26</xref><xref ref-type="aff" rid="AF0027">27</xref><xref ref-type="aff" rid="AF0028">28</xref></contrib><contrib contrib-type="author"><name><surname>Hamel</surname><given-names>Mary J.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0029">29</xref><xref ref-type="aff" rid="AF0030">30</xref></contrib><contrib contrib-type="author"><name><surname>Ngulukyo</surname><given-names>Emmanuel</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0029">29</xref><xref ref-type="aff" rid="AF0030">30</xref></contrib><contrib contrib-type="author"><name><surname>Odhiambo</surname><given-names>Frank O.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0029">29</xref><xref ref-type="aff" rid="AF0030">30</xref></contrib><contrib contrib-type="author"><name><surname>Sewe</surname><given-names>Maquins</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0029">29</xref><xref ref-type="aff" rid="AF0030">30</xref></contrib><contrib contrib-type="author"><name><surname>Beguy</surname><given-names>Donatien</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0031">31</xref><xref ref-type="aff" rid="AF0032">32</xref></contrib><contrib contrib-type="author"><name><surname>Ezeh</surname><given-names>Alex</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0031">31</xref><xref ref-type="aff" rid="AF0032">32</xref></contrib><contrib contrib-type="author"><name><surname>Oti</surname><given-names>Samuel</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0031">31</xref><xref ref-type="aff" rid="AF0032">32</xref></contrib><contrib contrib-type="author"><name><surname>Diallo</surname><given-names>Aldiouma</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0033">33</xref><xref ref-type="aff" rid="AF0034">34</xref></contrib><contrib contrib-type="author"><name><surname>Douillot</surname><given-names>Laetitia</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0033">33</xref><xref ref-type="aff" rid="AF0034">34</xref></contrib><contrib contrib-type="author"><name><surname>Sokhna</surname><given-names>Cheikh</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0033">33</xref><xref ref-type="aff" rid="AF0034">34</xref></contrib><contrib contrib-type="author"><name><surname>Delaunay</surname><given-names>Val&#x000e9;rie</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0033">33</xref><xref ref-type="aff" rid="AF0034">34</xref></contrib><contrib contrib-type="author"><name><surname>Collinson</surname><given-names>Mark A.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0035">35</xref><xref ref-type="aff" rid="AF0036">36</xref><xref ref-type="aff" rid="AF0037">37</xref></contrib><contrib contrib-type="author"><name><surname>Kabudula</surname><given-names>Chodziwadziwa W.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0035">35</xref><xref ref-type="aff" rid="AF0036">36</xref></contrib><contrib contrib-type="author"><name><surname>Kahn</surname><given-names>Kathleen</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0035">35</xref><xref ref-type="aff" rid="AF0036">36</xref><xref ref-type="aff" rid="AF0037">37</xref></contrib><contrib contrib-type="author"><name><surname>Herbst</surname><given-names>Kobus</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0038">38</xref><xref ref-type="aff" rid="AF0039">39</xref></contrib><contrib contrib-type="author"><name><surname>Mossong</surname><given-names>Jo&#x000eb;l</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0038">38</xref><xref ref-type="aff" rid="AF0039">39</xref><xref ref-type="aff" rid="AF0040">40</xref></contrib><contrib contrib-type="author"><name><surname>Chuc</surname><given-names>Nguyen T.K.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0041">41</xref><xref ref-type="aff" rid="AF0042">42</xref></contrib><contrib contrib-type="author"><name><surname>Bangha</surname><given-names>Martin</given-names></name><xref ref-type="aff" rid="AF0003">3</xref></contrib><contrib contrib-type="author"><name><surname>Sankoh</surname><given-names>Osman A.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0043">43</xref><xref ref-type="aff" rid="AF0044">44</xref><xref ref-type="corresp" rid="cor1">*</xref></contrib><contrib contrib-type="author"><name><surname>Byass</surname><given-names>Peter</given-names></name><xref ref-type="aff" rid="AF0036">36</xref><xref ref-type="aff" rid="AF0045">45</xref></contrib>

Global Health Action

<p>Mortality in childhood, particularly in the first 5 years of life, has been a major global concern in recent years. Additional attention has been given to rates of all-cause mortality reduction within the framework of Millennium Development Goal 4 (<xref rid="CIT0001" ref-type="bibr">1</xref>) and considerable successes are being achieved in various countries. At the same time, some components of child mortality, for example, deaths in the early days of life, are proving less transigent. Overall burdens of childhood mortality can only be clearly understood when causes of death are reliably attributed, and it has to be recognised that some causes may be more susceptible to reduction than others. At the same time, the mix of causes varies considerably between different settings, as well as between age groups.</p><p>Cause-specific childhood mortality in low- and middle-income countries is estimated from a range of sources, including the Child Epidemiology Reference Group (CHERG) (<xref rid="CIT0002" ref-type="bibr">2</xref>), and the Global Burden of Disease study (<xref rid="CIT0003" ref-type="bibr">3</xref>). However, the data underlying these estimates are often sparse and inconsistent, particularly when it comes to understanding mortality patterns on a population basis (<xref rid="CIT0004" ref-type="bibr">4</xref>).</p><p>The INDEPTH Network Health and Demographic Surveillance Sites (HDSS) follow vital events within defined populations continuously, and so they provide a means for documenting mortality on a population-related basis (<xref rid="CIT0005" ref-type="bibr">5</xref>). Furthermore, by undertaking standardised verbal autopsy (VA) enquiries to follow-up deaths, cause-specific mortality can be assessed within specific childhood age groups to see which cause groups account for substantial components of overall mortality (<xref rid="CIT0006" ref-type="bibr">6</xref>).</p><p>Our aim in this paper is to describe childhood cause-specific mortality patterns on the basis of a dataset collected at 22 INDEPTH Network HDSSs across Africa and Asia (<xref rid="CIT0007" ref-type="bibr">7</xref>). We have chosen here to take as &#x02018;childhood&#x02019; the overall age range from birth to 15 years, to give a complete picture of mortality patterns up to adulthood, at the same time providing results separately for the neonatal period, infancy and the under-5 year age group. Although these INDEPTH sites are not constituted as a representative sample, they provide point estimates over a wide range of settings and time periods.</p><sec sec-type="methods" id="S0002"><title>Methods</title><p>The overall INDEPTH dataset (<xref rid="CIT0008" ref-type="bibr">8</xref>) from which these childhood mortality analyses are drawn is described in detail elsewhere (<xref rid="CIT0007" ref-type="bibr">7</xref>). The Karonga, Malawi, site did not contribute VAs for childhood deaths, and the Purworejo, Indonesia; Farafenni, The Gambia; and Vadu, India, sites carried out verbal autopsies for less than half of the childhood deaths that occurred and/or did not report for the period 2006&#x02013;2012. Therefore these sites are not considered further here. This leaves documentation on 28,751 deaths in 4,387,824 person-years of observation across 18 sites. VA interviews were successfully completed on 25,357 (88.2%) of the deaths that occurred. A summary of the detailed methods used in common for this series of multisite papers is shown in <xref ref-type="boxed-text" rid="T0003">Box 1</xref>.</p><p><italic>Box 1</italic>. Summary of methodology based on the detailed description in the introductory paper (<xref rid="CIT0007" ref-type="bibr">7</xref>).</p><boxed-text id="T0003" position="float"><p> <bold>Age&#x02013;sex&#x02013;time standardisation</bold> </p><p>To avoid effects of differences and changes in age&#x02013;sex structures of populations, mortality fractions and rates have been adjusted using the INDEPTH 2013 population standard (<xref rid="CIT0009" ref-type="bibr">9</xref>). A weighting factor was calculated for each site, age group, sex and year category in relation to the standard for the corresponding age group and sex, and incorporated into the overall dataset. This is referred to in this paper as age&#x02013;sex&#x02013;time standardisation in the contexts where it is used.</p><p> <bold>Cause of death assignment</bold> </p><p>The InterVA-4 (version 4.02) probabilistic model was used for all the cause of death assignments in the overall dataset (<xref rid="CIT0010" ref-type="bibr">10</xref>). InterVA-4 is fully compliant with the WHO 2012 Verbal Autopsy standard and generates causes of death categorised by ICD-10 groups (<xref rid="CIT0011" ref-type="bibr">11</xref>). The data reported here were collected before the WHO 2012 VA standard was available, but were transformed into the WHO 2012 and InterVA-4 format to optimise cross-site standardisation in cause of death attribution. For a small proportion of deaths, VA interviews were not successfully completed; a few others contained inadequate information to arrive at a cause of death. InterVA-4 assigns causes of death (maximum 3) with associated likelihoods; thus cases for which likely causes did not total 100% were also assigned a residual indeterminate component. This served as a means of encapsulating uncertainty in cause of death at the individual level within the overall dataset, as well as accounting for 100% of every death.</p><p> <bold>Overall dataset</bold> </p><p>The overall public-domain dataset (<xref rid="CIT0008" ref-type="bibr">8</xref>) thus contains between one and four records for each death, with the sum of likelihoods for each individual being unity. Each record includes a specific cause of death, its likelihood and its age&#x02013;sex&#x02013;time weighting.</p></boxed-text><p>In this context, all of these data are secondary datasets derived from primary data collected separately by each participating site. In all cases the primary data collection was covered by site-level ethical approvals relating to on-going health and demographic surveillance in those specific locations. No individual identity or household location data were included in the secondary data and no specific ethical approvals were required for these pooled analyses.</p></sec><sec id="S0003"><title> Results</title><p>Over the total of 28,751 deaths during 4,387,824 person-years of observation, 5,213 occurred in the neonatal period (first 28 days of life); 8,967 during the remainder of infancy (from one month up to the first birthday); 10,764 in the 1&#x02013;4 year age group and 3,807 in the 5&#x02013;14 year age group. All 18 sites reported mortality during at least part of the period 2006&#x02013;2012, which comprised 68.8% of overall person-time observed; the period 2000&#x02013;2005 accounted for a further 25.7%. The most natural way to analyse these longitudinal population data across sites is to calculate site-specific mortality rates per 1,000 person-years, shown in <xref ref-type="table" rid="T0001">Table 1</xref> by age group, period and site. In the sites that have longer-term data, there are some trends reflecting falling childhood mortality. There are also exceptions, however; at the Agincourt, South Africa site, there are clear indications of mortality rising in the middle period, when the HIV/AIDS epidemic was at its height. For the period 2006&#x02013;12, the highest rates of neonatal mortality were observed in Asian sites, even though they recorded generally lower mortality rates than many African sites in subsequent age groups.</p><table-wrap id="T0001" position="float"><label>Table 1</label><caption><p>Childhood all-cause mortality rates per 1,000 person-years by age group and period for 18 INDEPTH HDSS sites</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1">Age group</th><th align="center" colspan="3" rowspan="1">0&#x02013;28 days</th><th align="center" colspan="3" rowspan="1">1&#x02013;11 months</th><th align="center" colspan="3" rowspan="1">1&#x02013;4 years</th><th align="center" colspan="3" rowspan="1">5&#x02013;14 years</th></tr><tr><th colspan="13" rowspan="1"> <hr/> </th></tr><tr><th align="left" rowspan="1" colspan="1">Period</th><th align="center" rowspan="1" colspan="1">&#x0003c;2000</th><th align="center" rowspan="1" colspan="1">2000&#x02013;05</th><th align="center" rowspan="1" colspan="1">2006&#x02013;12</th><th align="center" rowspan="1" colspan="1">&#x0003c;2000</th><th align="center" rowspan="1" colspan="1">2000&#x02013;05</th><th align="center" rowspan="1" colspan="1">2006&#x02013;12</th><th align="center" rowspan="1" colspan="1">&#x0003c;2000</th><th align="center" rowspan="1" colspan="1">2000&#x02013;05</th><th align="center" rowspan="1" colspan="1">2006&#x02013;12</th><th align="center" rowspan="1" colspan="1">&#x0003c;2000</th><th align="center" rowspan="1" colspan="1">2000&#x02013;05</th><th align="center" rowspan="1" colspan="1">2006&#x02013;12</th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">Bangladesh: Matlab</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">389.8</td><td align="center" rowspan="1" colspan="1">357.6</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">11.8</td><td align="center" rowspan="1" colspan="1">10.6</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">3.2</td><td align="center" rowspan="1" colspan="1">2.3</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.8</td><td align="center" rowspan="1" colspan="1">0.6</td></tr><tr><td align="left" rowspan="1" colspan="1">Bangladesh: Bandarban</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">171.0</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">28.6</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1.9</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1.0</td></tr><tr><td align="left" rowspan="1" colspan="1">Bangladesh: Chakaria</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">458.0</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">16.6</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">4.0</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1.0</td></tr><tr><td align="left" rowspan="1" colspan="1">Bangladesh: AMK</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">444.7</td><td align="center" rowspan="1" colspan="1">326.8</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">11.9</td><td align="center" rowspan="1" colspan="1">8.0</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">3.4</td><td align="center" rowspan="1" colspan="1">2.7</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.8</td><td align="center" rowspan="1" colspan="1">0.6</td></tr><tr><td align="left" rowspan="1" colspan="1">Burkina Faso: Nouna</td><td align="center" rowspan="1" colspan="1">101.2</td><td align="center" rowspan="1" colspan="1">142.2</td><td align="center" rowspan="1" colspan="1">92.9</td><td align="center" rowspan="1" colspan="1">39.3</td><td align="center" rowspan="1" colspan="1">42.5</td><td align="center" rowspan="1" colspan="1">24.4</td><td align="center" rowspan="1" colspan="1">29.8</td><td align="center" rowspan="1" colspan="1">19.2</td><td align="center" rowspan="1" colspan="1">12.7</td><td align="center" rowspan="1" colspan="1">6.0</td><td align="center" rowspan="1" colspan="1">2.6</td><td align="center" rowspan="1" colspan="1">1.6</td></tr><tr><td align="left" rowspan="1" colspan="1">Burkina Faso: Ouagadougou</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">136.4</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">20.8</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">7.8</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1.4</td></tr><tr><td align="left" rowspan="1" colspan="1">C&#x000f4;te d'Ivoire: Taabo</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">200.9</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">32.0</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">15.2</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1.8</td></tr><tr><td align="left" rowspan="1" colspan="1">Ethiopia: Kilite-Awlaelo</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">188.0</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">12.6</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">2.8</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1.1</td></tr><tr><td align="left" rowspan="1" colspan="1">Ghana: Navrongo</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">305.5</td><td align="center" rowspan="1" colspan="1">209.7</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">43.5</td><td align="center" rowspan="1" colspan="1">22.0</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">11.4</td><td align="center" rowspan="1" colspan="1">8.2</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">2.2</td><td align="center" rowspan="1" colspan="1">1.7</td></tr><tr><td align="left" rowspan="1" colspan="1">Ghana: Dodowa</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">90.4</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">8.7</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">4.7</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1.4</td></tr><tr><td align="left" rowspan="1" colspan="1">India: Ballabgarh</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">280.0</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">24.4</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">4.0</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.8</td></tr><tr><td align="left" rowspan="1" colspan="1">Kenya: Kilifi</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">160.0</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">9.6</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">2.5</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.8</td></tr><tr><td align="left" rowspan="1" colspan="1">Kenya: Kisumu</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">302.6</td><td align="center" rowspan="1" colspan="1">243.0</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">111.7</td><td align="center" rowspan="1" colspan="1">74.2</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">31.7</td><td align="center" rowspan="1" colspan="1">22.8</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">2.7</td><td align="center" rowspan="1" colspan="1">2.4</td></tr><tr><td align="left" rowspan="1" colspan="1">Kenya: Nairobi</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">373.3</td><td align="center" rowspan="1" colspan="1">319.8</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">58.0</td><td align="center" rowspan="1" colspan="1">49.8</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">8.4</td><td align="center" rowspan="1" colspan="1">6.4</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">2.1</td><td align="center" rowspan="1" colspan="1">1.1</td></tr><tr><td align="left" rowspan="1" colspan="1">Senegal: Niakhar</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">210.6</td><td align="center" rowspan="1" colspan="1">126.9</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">31.2</td><td align="center" rowspan="1" colspan="1">16.8</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">20.5</td><td align="center" rowspan="1" colspan="1">9.9</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">3.2</td><td align="center" rowspan="1" colspan="1">1.5</td></tr><tr><td align="left" rowspan="1" colspan="1">South Africa: Agincourt</td><td align="center" rowspan="1" colspan="1">81.0</td><td align="center" rowspan="1" colspan="1">119.7</td><td align="center" rowspan="1" colspan="1">154.7</td><td align="center" rowspan="1" colspan="1">13.5</td><td align="center" rowspan="1" colspan="1">30.3</td><td align="center" rowspan="1" colspan="1">30.9</td><td align="center" rowspan="1" colspan="1">4.4</td><td align="center" rowspan="1" colspan="1">7.0</td><td align="center" rowspan="1" colspan="1">5.3</td><td align="center" rowspan="1" colspan="1">0.7</td><td align="center" rowspan="1" colspan="1">1.0</td><td align="center" rowspan="1" colspan="1">1.3</td></tr><tr><td align="left" rowspan="1" colspan="1">South Africa: Africa Centre</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">151.1</td><td align="center" rowspan="1" colspan="1">53.0</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">49.5</td><td align="center" rowspan="1" colspan="1">27.6</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">8.9</td><td align="center" rowspan="1" colspan="1">4.7</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1.7</td><td align="center" rowspan="1" colspan="1">1.2</td></tr><tr><td align="left" rowspan="1" colspan="1">Vietnam: FilaBavi</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">123.3</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">3.0</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1.0</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.4</td></tr></tbody></table></table-wrap><p>The terms &#x02018;infant mortality rate&#x02019; and &#x02018;under-5 mortality rate&#x02019; are frequently used, arguably incorrectly, to refer to numbers of deaths per 1,000 live births, rather than to person-time based rates. However, for the sake of comparability with other sources, <xref ref-type="fig" rid="F0001">Fig. 1</xref> shows these widely used measures of infant and under-5 mortality rates per 1,000 live births for the period 2006&#x02013;2012, during which all 18 sites reported. The FilaBavi site in Vietnam recorded infant mortality of 11 and under-5 mortality of 15 per 1,000 live births, while the Kisumu site on the northern shores of Lake Victoria recorded infant mortality of 78 and under-5 mortality of 152 per 1,000 live births.</p><fig id="F0001" position="float"><label>Fig. 1</label><caption><p>Location of the 18 contributing INDEPTH HDSSs, showing infant mortality rates (deaths in first year of life per 1,000 live births, IMR) and under-5 mortality rates (deaths in first 5 years of life per 1,000 live births, U5MR) for the period 2006&#x02013;2012.</p></caption><graphic xlink:href="GHA-7-25363-g001"/></fig><p> <xref ref-type="table" rid="T0002">Table 2</xref> shows a detailed breakdown of cause-specific mortality rates per 1,000 person-years by site for major causes and cause groups of childhood mortality. More specific considerations of mortality due to malaria, HIV and external causes are given in accompanying papers (<xref rid="CIT0012" ref-type="bibr">12</xref>&#x02013;<xref rid="CIT0014" ref-type="bibr">14</xref>) ; these causes are included here for the sake of completeness rather than for detailed discussion.</p><table-wrap id="T0002" position="float"><label>Table 2</label><caption><p>Childhood mortality rates per 1,000 person-years, by cause group and age group, for 18 INDEPTH HDSS sites from 2006 to 2012</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1">Cause</th><th align="center" rowspan="1" colspan="1">Birth asphyxia</th><th align="center" rowspan="1" colspan="1">Neonatal infections</th><th align="center" rowspan="1" colspan="1">Congenital</th><th align="center" rowspan="1" colspan="1">Prematurity</th><th align="center" rowspan="1" colspan="1">Diarrhoea</th><th align="center" rowspan="1" colspan="1">HIV/AIDS</th><th align="center" rowspan="1" colspan="1">Malaria</th><th align="center" rowspan="1" colspan="1">Pneumonia</th><th align="center" rowspan="1" colspan="1">Other infections</th><th align="center" rowspan="1" colspan="1">External causes</th><th align="center" rowspan="1" colspan="1">NCDs</th><th align="center" rowspan="1" colspan="1">Other causes</th><th align="center" rowspan="1" colspan="1">Indeterminate</th><th align="center" rowspan="1" colspan="1">All causes</th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">0&#x02013;28 days</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Bangladesh: Matlab</td><td align="center" rowspan="1" colspan="1">30.69</td><td align="center" rowspan="1" colspan="1">116.03</td><td align="center" rowspan="1" colspan="1">4.67</td><td align="center" rowspan="1" colspan="1">71.16</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.55</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">35.77</td><td align="center" rowspan="1" colspan="1">67.91</td><td align="center" rowspan="1" colspan="1">326.78</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Bangladesh: Bandarban</td><td align="center" rowspan="1" colspan="1">38.20</td><td align="center" rowspan="1" colspan="1">73.43</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">22.58</td><td align="center" rowspan="1" colspan="1">36.83</td><td align="center" rowspan="1" colspan="1">171.04</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Bangladesh: Chakaria</td><td align="center" rowspan="1" colspan="1">104.36</td><td align="center" rowspan="1" colspan="1">42.35</td><td align="center" rowspan="1" colspan="1">1.97</td><td align="center" rowspan="1" colspan="1">105.09</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1.79</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">73.27</td><td align="center" rowspan="1" colspan="1">129.14</td><td align="center" rowspan="1" colspan="1">457.97</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Bangladesh: AMK</td><td align="center" rowspan="1" colspan="1">53.31</td><td align="center" rowspan="1" colspan="1">126.26</td><td align="center" rowspan="1" colspan="1">9.17</td><td align="center" rowspan="1" colspan="1">25.66</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">44.75</td><td align="center" rowspan="1" colspan="1">98.46</td><td align="center" rowspan="1" colspan="1">357.61</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Burkina Faso: Nouna</td><td align="center" rowspan="1" colspan="1">18.91</td><td align="center" rowspan="1" colspan="1">44.89</td><td align="center" rowspan="1" colspan="1">0.63</td><td align="center" rowspan="1" colspan="1">5.10</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">2.42</td><td align="center" rowspan="1" colspan="1">20.93</td><td align="center" rowspan="1" colspan="1">92.88</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Burkina Faso: Ouagadougou</td><td align="center" rowspan="1" colspan="1">18.42</td><td align="center" rowspan="1" colspan="1">49.15</td><td align="center" rowspan="1" colspan="1">7.80</td><td align="center" rowspan="1" colspan="1">16.90</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">5.83</td><td align="center" rowspan="1" colspan="1">38.33</td><td align="center" rowspan="1" colspan="1">136.43</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;C&#x000f4;te d'Ivoire: Taabo</td><td align="center" rowspan="1" colspan="1">53.32</td><td align="center" rowspan="1" colspan="1">80.20</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">20.39</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">15.25</td><td align="center" rowspan="1" colspan="1">31.75</td><td align="center" rowspan="1" colspan="1">200.91</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Ethiopia: Kilite-Awlaelo</td><td align="center" rowspan="1" colspan="1">11.56</td><td align="center" rowspan="1" colspan="1">79.00</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">4.00</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">22.12</td><td align="center" rowspan="1" colspan="1">71.32</td><td align="center" rowspan="1" colspan="1">188.00</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Ghana: Navrongo</td><td align="center" rowspan="1" colspan="1">52.57</td><td align="center" rowspan="1" colspan="1">43.80</td><td align="center" rowspan="1" colspan="1">0.95</td><td align="center" rowspan="1" colspan="1">57.14</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">32.36</td><td align="center" rowspan="1" colspan="1">22.91</td><td align="center" rowspan="1" colspan="1">209.73</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Ghana: Dodowa</td><td align="center" rowspan="1" colspan="1">11.17</td><td align="center" rowspan="1" colspan="1">15.45</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">5.94</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">16.09</td><td align="center" rowspan="1" colspan="1">41.79</td><td align="center" rowspan="1" colspan="1">90.44</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;India: Ballabgarh</td><td align="center" rowspan="1" colspan="1">32.37</td><td align="center" rowspan="1" colspan="1">68.05</td><td align="center" rowspan="1" colspan="1">2.18</td><td align="center" rowspan="1" colspan="1">77.48</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">28.55</td><td align="center" rowspan="1" colspan="1">71.38</td><td align="center" rowspan="1" colspan="1">280.01</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Kenya: Kilifi</td><td align="center" rowspan="1" colspan="1">37.98</td><td align="center" rowspan="1" colspan="1">50.88</td><td align="center" rowspan="1" colspan="1">3.04</td><td align="center" rowspan="1" colspan="1">9.92</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">7.00</td><td align="center" rowspan="1" colspan="1">51.20</td><td align="center" rowspan="1" colspan="1">160.02</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Kenya: Kisumu</td><td align="center" rowspan="1" colspan="1">52.66</td><td align="center" rowspan="1" colspan="1">65.56</td><td align="center" rowspan="1" colspan="1">2.46</td><td align="center" rowspan="1" colspan="1">9.61</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.49</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">27.37</td><td align="center" rowspan="1" colspan="1">84.88</td><td align="center" rowspan="1" colspan="1">243.03</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Kenya: Nairobi</td><td align="center" rowspan="1" colspan="1">77.53</td><td align="center" rowspan="1" colspan="1">80.27</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">19.41</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1.31</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">37.69</td><td align="center" rowspan="1" colspan="1">103.61</td><td align="center" rowspan="1" colspan="1">319.82</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Senegal: Niakhar</td><td align="center" rowspan="1" colspan="1">5.83</td><td align="center" rowspan="1" colspan="1">51.38</td><td align="center" rowspan="1" colspan="1">1.10</td><td align="center" rowspan="1" colspan="1">6.48</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">13.01</td><td align="center" rowspan="1" colspan="1">49.11</td><td align="center" rowspan="1" colspan="1">126.91</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;South Africa: Agincourt</td><td align="center" rowspan="1" colspan="1">26.50</td><td align="center" rowspan="1" colspan="1">73.77</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">10.99</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1.04</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">13.94</td><td align="center" rowspan="1" colspan="1">28.40</td><td align="center" rowspan="1" colspan="1">154.64</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;South Africa: Africa Centre</td><td align="center" rowspan="1" colspan="1">10.26</td><td align="center" rowspan="1" colspan="1">20.89</td><td align="center" rowspan="1" colspan="1">4.70</td><td align="center" rowspan="1" colspan="1">1.06</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">2.47</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.91</td><td align="center" rowspan="1" colspan="1">12.72</td><td align="center" rowspan="1" colspan="1">53.01</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Vietnam: FilaBavi</td><td align="center" rowspan="1" colspan="1">31.56</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">38.58</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">2.36</td><td align="center" rowspan="1" colspan="1">50.74</td><td align="center" rowspan="1" colspan="1">123.24</td></tr><tr><td align="left" rowspan="1" colspan="1">1&#x02013;11 months</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Bangladesh: Matlab</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.30</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">4.85</td><td align="center" rowspan="1" colspan="1">0.92</td><td align="center" rowspan="1" colspan="1">0.19</td><td align="center" rowspan="1" colspan="1">0.19</td><td align="center" rowspan="1" colspan="1">0.86</td><td align="center" rowspan="1" colspan="1">0.62</td><td align="center" rowspan="1" colspan="1">8.00</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Bangladesh: Bandarban</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1.61</td><td align="center" rowspan="1" colspan="1">1.03</td><td align="center" rowspan="1" colspan="1">0.85</td><td align="center" rowspan="1" colspan="1">8.15</td><td align="center" rowspan="1" colspan="1">0.42</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.77</td><td align="center" rowspan="1" colspan="1">0.36</td><td align="center" rowspan="1" colspan="1">15.40</td><td align="center" rowspan="1" colspan="1">28.59</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Bangladesh: Chakaria</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1.39</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1.55</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">3.46</td><td align="center" rowspan="1" colspan="1">1.86</td><td align="center" rowspan="1" colspan="1">0.76</td><td align="center" rowspan="1" colspan="1">0.59</td><td align="center" rowspan="1" colspan="1">0.38</td><td align="center" rowspan="1" colspan="1">6.44</td><td align="center" rowspan="1" colspan="1">16.58</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Bangladesh: AMK</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1.11</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">7.24</td><td align="center" rowspan="1" colspan="1">0.48</td><td align="center" rowspan="1" colspan="1">0.35</td><td align="center" rowspan="1" colspan="1">0.21</td><td align="center" rowspan="1" colspan="1">0.43</td><td align="center" rowspan="1" colspan="1">0.69</td><td align="center" rowspan="1" colspan="1">10.56</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Burkina Faso: Nouna</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1.17</td><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">14.14</td><td align="center" rowspan="1" colspan="1">3.32</td><td align="center" rowspan="1" colspan="1">0.43</td><td align="center" rowspan="1" colspan="1">0.59</td><td align="center" rowspan="1" colspan="1">0.17</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">4.40</td><td align="center" rowspan="1" colspan="1">24.39</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Burkina Faso: Ouagadougou</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.64</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1.72</td><td align="center" rowspan="1" colspan="1">1.80</td><td align="center" rowspan="1" colspan="1">3.03</td><td align="center" rowspan="1" colspan="1">6.66</td><td align="center" rowspan="1" colspan="1">0.92</td><td align="center" rowspan="1" colspan="1">0.15</td><td align="center" rowspan="1" colspan="1">0.68</td><td align="center" rowspan="1" colspan="1">0.62</td><td align="center" rowspan="1" colspan="1">4.64</td><td align="center" rowspan="1" colspan="1">20.86</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;C&#x000f4;te d'Ivoire: Taabo</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.96</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1.84</td><td align="center" rowspan="1" colspan="1">2.79</td><td align="center" rowspan="1" colspan="1">5.83</td><td align="center" rowspan="1" colspan="1">8.13</td><td align="center" rowspan="1" colspan="1">2.01</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.78</td><td align="center" rowspan="1" colspan="1">0.62</td><td align="center" rowspan="1" colspan="1">9.04</td><td align="center" rowspan="1" colspan="1">32.00</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Ethiopia: Kilite-Awlaelo</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.31</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.65</td><td align="center" rowspan="1" colspan="1">0.32</td><td align="center" rowspan="1" colspan="1">6.14</td><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">5.00</td><td align="center" rowspan="1" colspan="1">12.60</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Ghana: Navrongo</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.69</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">3.14</td><td align="center" rowspan="1" colspan="1">0.99</td><td align="center" rowspan="1" colspan="1">2.59</td><td align="center" rowspan="1" colspan="1">4.93</td><td align="center" rowspan="1" colspan="1">2.09</td><td align="center" rowspan="1" colspan="1">0.44</td><td align="center" rowspan="1" colspan="1">1.11</td><td align="center" rowspan="1" colspan="1">0.25</td><td align="center" rowspan="1" colspan="1">5.76</td><td align="center" rowspan="1" colspan="1">21.99</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Ghana: Dodowa</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.51</td><td align="center" rowspan="1" colspan="1">0.39</td><td align="center" rowspan="1" colspan="1">0.36</td><td align="center" rowspan="1" colspan="1">3.47</td><td align="center" rowspan="1" colspan="1">0.23</td><td align="center" rowspan="1" colspan="1">0.15</td><td align="center" rowspan="1" colspan="1">0.16</td><td align="center" rowspan="1" colspan="1">0.35</td><td align="center" rowspan="1" colspan="1">3.12</td><td align="center" rowspan="1" colspan="1">8.74</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;India: Ballabgarh</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.63</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">4.16</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.71</td><td align="center" rowspan="1" colspan="1">8.88</td><td align="center" rowspan="1" colspan="1">0.99</td><td align="center" rowspan="1" colspan="1">0.53</td><td align="center" rowspan="1" colspan="1">1.20</td><td align="center" rowspan="1" colspan="1">0.39</td><td align="center" rowspan="1" colspan="1">6.94</td><td align="center" rowspan="1" colspan="1">24.43</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Kenya: Kilifi</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.12</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.40</td><td align="center" rowspan="1" colspan="1">1.94</td><td align="center" rowspan="1" colspan="1">1.08</td><td align="center" rowspan="1" colspan="1">2.48</td><td align="center" rowspan="1" colspan="1">0.60</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">0.19</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">2.60</td><td align="center" rowspan="1" colspan="1">9.57</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Kenya: Kisumu</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.34</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">6.55</td><td align="center" rowspan="1" colspan="1">7.54</td><td align="center" rowspan="1" colspan="1">17.99</td><td align="center" rowspan="1" colspan="1">25.49</td><td align="center" rowspan="1" colspan="1">2.74</td><td align="center" rowspan="1" colspan="1">0.49</td><td align="center" rowspan="1" colspan="1">1.70</td><td align="center" rowspan="1" colspan="1">0.44</td><td align="center" rowspan="1" colspan="1">10.96</td><td align="center" rowspan="1" colspan="1">74.24</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Kenya: Nairobi</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">2.90</td><td align="center" rowspan="1" colspan="1">3.84</td><td align="center" rowspan="1" colspan="1">1.16</td><td align="center" rowspan="1" colspan="1">16.71</td><td align="center" rowspan="1" colspan="1">6.48</td><td align="center" rowspan="1" colspan="1">0.99</td><td align="center" rowspan="1" colspan="1">0.17</td><td align="center" rowspan="1" colspan="1">0.23</td><td align="center" rowspan="1" colspan="1">17.26</td><td align="center" rowspan="1" colspan="1">49.78</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Senegal: Niakhar</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">6.10</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">1.86</td><td align="center" rowspan="1" colspan="1">2.03</td><td align="center" rowspan="1" colspan="1">0.55</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.86</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">5.21</td><td align="center" rowspan="1" colspan="1">16.77</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;South Africa: Agincourt</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.23</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">3.32</td><td align="center" rowspan="1" colspan="1">4.42</td><td align="center" rowspan="1" colspan="1">0.75</td><td align="center" rowspan="1" colspan="1">12.90</td><td align="center" rowspan="1" colspan="1">2.79</td><td align="center" rowspan="1" colspan="1">0.26</td><td align="center" rowspan="1" colspan="1">0.55</td><td align="center" rowspan="1" colspan="1">0.21</td><td align="center" rowspan="1" colspan="1">5.47</td><td align="center" rowspan="1" colspan="1">30.90</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;South Africa: Africa Centre</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.29</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1.87</td><td align="center" rowspan="1" colspan="1">3.97</td><td align="center" rowspan="1" colspan="1">0.26</td><td align="center" rowspan="1" colspan="1">15.96</td><td align="center" rowspan="1" colspan="1">0.59</td><td align="center" rowspan="1" colspan="1">0.22</td><td align="center" rowspan="1" colspan="1">0.24</td><td align="center" rowspan="1" colspan="1">0.49</td><td align="center" rowspan="1" colspan="1">3.65</td><td align="center" rowspan="1" colspan="1">27.54</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Vietnam: FilaBavi</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.16</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1.82</td><td align="center" rowspan="1" colspan="1">0.46</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.57</td><td align="center" rowspan="1" colspan="1">3.01</td></tr><tr><td align="left" rowspan="1" colspan="1">1&#x02013;4 years</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Bangladesh: Matlab</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.44</td><td align="center" rowspan="1" colspan="1">0.15</td><td align="center" rowspan="1" colspan="1">1.11</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.62</td><td align="center" rowspan="1" colspan="1">0.22</td><td align="center" rowspan="1" colspan="1">2.65</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Bangladesh: Bandarban</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.17</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.17</td><td align="center" rowspan="1" colspan="1">0.16</td><td align="center" rowspan="1" colspan="1">0.17</td><td align="center" rowspan="1" colspan="1">0.35</td><td align="center" rowspan="1" colspan="1">0.29</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.59</td><td align="center" rowspan="1" colspan="1">1.90</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Bangladesh: Chakaria</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.27</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.78</td><td align="center" rowspan="1" colspan="1">0.27</td><td align="center" rowspan="1" colspan="1">1.34</td><td align="center" rowspan="1" colspan="1">0.28</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.92</td><td align="center" rowspan="1" colspan="1">3.96</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Bangladesh: AMK</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.27</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.49</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">1.37</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">2.25</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Burkina Faso: Nouna</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.98</td><td align="center" rowspan="1" colspan="1">0.12</td><td align="center" rowspan="1" colspan="1">6.91</td><td align="center" rowspan="1" colspan="1">1.46</td><td align="center" rowspan="1" colspan="1">0.16</td><td align="center" rowspan="1" colspan="1">0.28</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">2.64</td><td align="center" rowspan="1" colspan="1">12.73</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Burkina Faso: Ouagadougou</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.48</td><td align="center" rowspan="1" colspan="1">0.58</td><td align="center" rowspan="1" colspan="1">2.43</td><td align="center" rowspan="1" colspan="1">1.03</td><td align="center" rowspan="1" colspan="1">0.39</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">0.29</td><td align="center" rowspan="1" colspan="1">0.64</td><td align="center" rowspan="1" colspan="1">1.78</td><td align="center" rowspan="1" colspan="1">7.76</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;C&#x000f4;te d'Ivoire: Taabo</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.70</td><td align="center" rowspan="1" colspan="1">1.55</td><td align="center" rowspan="1" colspan="1">4.88</td><td align="center" rowspan="1" colspan="1">1.50</td><td align="center" rowspan="1" colspan="1">0.54</td><td align="center" rowspan="1" colspan="1">0.23</td><td align="center" rowspan="1" colspan="1">0.58</td><td align="center" rowspan="1" colspan="1">0.24</td><td align="center" rowspan="1" colspan="1">4.92</td><td align="center" rowspan="1" colspan="1">15.21</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Ethiopia: Kilite-Awlaelo</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.15</td><td align="center" rowspan="1" colspan="1">0.22</td><td align="center" rowspan="1" colspan="1">0.17</td><td align="center" rowspan="1" colspan="1">0.30</td><td align="center" rowspan="1" colspan="1">0.12</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">0.22</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">1.49</td><td align="center" rowspan="1" colspan="1">2.84</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Ghana: Navrongo</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.74</td><td align="center" rowspan="1" colspan="1">0.76</td><td align="center" rowspan="1" colspan="1">2.08</td><td align="center" rowspan="1" colspan="1">0.53</td><td align="center" rowspan="1" colspan="1">0.46</td><td align="center" rowspan="1" colspan="1">0.44</td><td align="center" rowspan="1" colspan="1">0.83</td><td align="center" rowspan="1" colspan="1">0.24</td><td align="center" rowspan="1" colspan="1">2.04</td><td align="center" rowspan="1" colspan="1">8.17</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Ghana: Dodowa</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.13</td><td align="center" rowspan="1" colspan="1">0.19</td><td align="center" rowspan="1" colspan="1">0.85</td><td align="center" rowspan="1" colspan="1">0.98</td><td align="center" rowspan="1" colspan="1">0.18</td><td align="center" rowspan="1" colspan="1">0.17</td><td align="center" rowspan="1" colspan="1">0.21</td><td align="center" rowspan="1" colspan="1">0.14</td><td align="center" rowspan="1" colspan="1">1.87</td><td align="center" rowspan="1" colspan="1">4.72</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;India: Ballabgarh</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.72</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.59</td><td align="center" rowspan="1" colspan="1">0.61</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.42</td><td align="center" rowspan="1" colspan="1">0.12</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">1.30</td><td align="center" rowspan="1" colspan="1">3.97</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Kenya: Kilifi</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">0.48</td><td align="center" rowspan="1" colspan="1">0.61</td><td align="center" rowspan="1" colspan="1">0.33</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.70</td><td align="center" rowspan="1" colspan="1">2.52</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Kenya: Kisumu</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1.26</td><td align="center" rowspan="1" colspan="1">5.18</td><td align="center" rowspan="1" colspan="1">7.61</td><td align="center" rowspan="1" colspan="1">2.46</td><td align="center" rowspan="1" colspan="1">0.60</td><td align="center" rowspan="1" colspan="1">0.41</td><td align="center" rowspan="1" colspan="1">0.75</td><td align="center" rowspan="1" colspan="1">0.71</td><td align="center" rowspan="1" colspan="1">3.82</td><td align="center" rowspan="1" colspan="1">22.80</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Kenya: Nairobi</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.35</td><td align="center" rowspan="1" colspan="1">1.04</td><td align="center" rowspan="1" colspan="1">0.22</td><td align="center" rowspan="1" colspan="1">0.95</td><td align="center" rowspan="1" colspan="1">1.18</td><td align="center" rowspan="1" colspan="1">0.37</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">2.18</td><td align="center" rowspan="1" colspan="1">6.43</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Senegal: Niakhar</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">2.94</td><td align="center" rowspan="1" colspan="1">0.24</td><td align="center" rowspan="1" colspan="1">3.45</td><td align="center" rowspan="1" colspan="1">0.38</td><td align="center" rowspan="1" colspan="1">0.15</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.72</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">1.96</td><td align="center" rowspan="1" colspan="1">9.89</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;South Africa: Agincourt</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.34</td><td align="center" rowspan="1" colspan="1">1.88</td><td align="center" rowspan="1" colspan="1">0.28</td><td align="center" rowspan="1" colspan="1">1.07</td><td align="center" rowspan="1" colspan="1">0.43</td><td align="center" rowspan="1" colspan="1">0.17</td><td align="center" rowspan="1" colspan="1">0.23</td><td align="center" rowspan="1" colspan="1">0.15</td><td align="center" rowspan="1" colspan="1">0.78</td><td align="center" rowspan="1" colspan="1">5.35</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;South Africa: Africa Centre</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">1.25</td><td align="center" rowspan="1" colspan="1">0.12</td><td align="center" rowspan="1" colspan="1">1.46</td><td align="center" rowspan="1" colspan="1">0.25</td><td align="center" rowspan="1" colspan="1">0.39</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.13</td><td align="center" rowspan="1" colspan="1">0.88</td><td align="center" rowspan="1" colspan="1">4.71</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Vietnam: FilaBavi</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.29</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.16</td><td align="center" rowspan="1" colspan="1">0.34</td><td align="center" rowspan="1" colspan="1">0.96</td></tr><tr><td align="left" rowspan="1" colspan="1">5&#x02013;14 years</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Bangladesh: Matlab</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">0.19</td><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.55</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Bangladesh: Bandarban</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.18</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.15</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.33</td><td align="center" rowspan="1" colspan="1">1.03</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Bangladesh: Chakaria</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.15</td><td align="center" rowspan="1" colspan="1">0.34</td><td align="center" rowspan="1" colspan="1">0.17</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.25</td><td align="center" rowspan="1" colspan="1">1.00</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Bangladesh: AMK</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">0.19</td><td align="center" rowspan="1" colspan="1">0.12</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.56</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Burkina Faso: Nouna</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.60</td><td align="center" rowspan="1" colspan="1">0.14</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">0.20</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.49</td><td align="center" rowspan="1" colspan="1">1.63</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Burkina Faso: Ouagadougou</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">0.35</td><td align="center" rowspan="1" colspan="1">0.16</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">0.13</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.41</td><td align="center" rowspan="1" colspan="1">1.35</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;C&#x000f4;te d'Ivoire: Taabo</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.27</td><td align="center" rowspan="1" colspan="1">0.27</td><td align="center" rowspan="1" colspan="1">0.25</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">0.22</td><td align="center" rowspan="1" colspan="1">0.12</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.48</td><td align="center" rowspan="1" colspan="1">1.78</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Ethiopia: Kilite-Awlaelo</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.28</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.45</td><td align="center" rowspan="1" colspan="1">1.05</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Ghana: Navrongo</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.13</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">0.17</td><td align="center" rowspan="1" colspan="1">0.40</td><td align="center" rowspan="1" colspan="1">0.46</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.41</td><td align="center" rowspan="1" colspan="1">1.74</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Ghana: Dodowa</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.21</td><td align="center" rowspan="1" colspan="1">0.20</td><td align="center" rowspan="1" colspan="1">0.13</td><td align="center" rowspan="1" colspan="1">0.14</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.46</td><td align="center" rowspan="1" colspan="1">1.33</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;India: Ballabgarh</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">0.22</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.35</td><td align="center" rowspan="1" colspan="1">0.84</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Kenya: Kilifi</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.20</td><td align="center" rowspan="1" colspan="1">0.12</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.13</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.15</td><td align="center" rowspan="1" colspan="1">0.78</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Kenya: Kisumu</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.44</td><td align="center" rowspan="1" colspan="1">0.59</td><td align="center" rowspan="1" colspan="1">0.30</td><td align="center" rowspan="1" colspan="1">0.18</td><td align="center" rowspan="1" colspan="1">0.14</td><td align="center" rowspan="1" colspan="1">0.23</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.47</td><td align="center" rowspan="1" colspan="1">2.39</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Kenya: Nairobi</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.24</td><td align="center" rowspan="1" colspan="1">0.15</td><td align="center" rowspan="1" colspan="1">0.16</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.31</td><td align="center" rowspan="1" colspan="1">1.07</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Senegal: Niakhar</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.16</td><td align="center" rowspan="1" colspan="1">0.30</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">0.12</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.30</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.53</td><td align="center" rowspan="1" colspan="1">1.50</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;South Africa: Agincourt</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.28</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.23</td><td align="center" rowspan="1" colspan="1">0.24</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.32</td><td align="center" rowspan="1" colspan="1">1.34</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;South Africa: Africa Centre</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.16</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.41</td><td align="center" rowspan="1" colspan="1">0.25</td><td align="center" rowspan="1" colspan="1">0.15</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.16</td><td align="center" rowspan="1" colspan="1">1.22</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Vietnam: FilaBavi</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.25</td><td align="center" rowspan="1" colspan="1">0.36</td></tr></tbody></table></table-wrap><p>Cause-specific mortality fractions (CSMF) for each of the 18 sites (<xref rid="CIT0015" ref-type="bibr">15</xref>&#x02013;<xref rid="CIT0032" ref-type="bibr">32</xref>) are shown separately for each age group (neonates, 1&#x02013;11 months, 1&#x02013;4 years and 5&#x02013;14 years) in <xref ref-type="fig" rid="F0002">Figs. 2</xref>&#x02013;<xref ref-type="fig" rid="F0005">5</xref> respectively, to give a sense of what the dominant causes of mortality are in particular sites and age groups. For most sites, infections accounted for the largest proportion of neonatal deaths, although prematurity was also an important cause in some settings. Pneumonia dominated as the major cause of infant deaths, although malaria was also important in some endemic areas. Local factors dictated major causes in the 1&#x02013;4 year age group, from external causes in Bangladesh to malaria and HIV/AIDS in highly endemic settings. For the 5&#x02013;14 year age group, external causes and, in some places, malaria continued as important causes, while there was also an increased proportion of mortality due to non-communicable diseases in some sites.</p><fig id="F0002" position="float"><label>Fig. 2</label><caption><p>Cause-specific mortality fractions (CSMF) for major cause of death groups for neonates at 18 INDEPTH sites during 2006&#x02013;2012.</p></caption><graphic xlink:href="GHA-7-25363-g002"/></fig><fig id="F0003" position="float"><label>Fig. 3</label><caption><p>Cause-specific mortality fractions (CSMF) for major cause of death groups for infants (1&#x02013;11 months) at 18 INDEPTH sites during 2006&#x02013;2012.</p></caption><graphic xlink:href="GHA-7-25363-g003"/></fig><fig id="F0004" position="float"><label>Fig. 4</label><caption><p>Cause-specific mortality fractions (CSMF) for major cause of death groups for children aged 1&#x02013;4 years at 18 INDEPTH sites during 2006&#x02013;2012.</p></caption><graphic xlink:href="GHA-7-25363-g004"/></fig><fig id="F0005" position="float"><label>Fig. 5</label><caption><p>Cause-specific mortality fractions (CSMF) for major cause of death groups for children aged 5&#x02013;14 years at 18 INDEPTH sites during 2006&#x02013;2012.</p></caption><graphic xlink:href="GHA-7-25363-g005"/></fig></sec><sec sec-type="discussion" id="S0004"><title>Discussion</title><p>This large-scale description of childhood mortality, all based on individually documented deaths within defined populations in Africa and Asia, provides important insights into the continuing causes of deaths that would be regarded as largely preventable in other parts of the world. Despite encouraging progress in reducing childhood mortality in many places, the data presented here reflect an overall situation from these sites in recent years where around 1 out of every 12 children born does not survive into adulthood.</p><p>The causes behind these individual tragedies are multifactorial, including poverty, living conditions and health services, and cannot be explored in full detail from the data presented here. What we do know from these cause of death data is that there are substantial differences in patterns of childhood mortality between countries, and in some cases, such as Kenya, also widespread within-country variation. Certain infectious diseases, particularly malaria and HIV/AIDS, contribute major components of childhood mortality in settings where they occur commonly, and hence account for a substantial part of variation in overall mortality. For example, at the Kisumu, Kenya, site malaria and HIV/AIDS together accounted for 56% of deaths in the 1&#x02013;4 year age group; but other causes did not occur at rates that were markedly different from those in a number of other sites. Conversely, at the FilaBavi, Vietnam site, where overall mortality in the 1&#x02013;4 year age group was less than 10% of the level observed in Kisumu, pneumonia accounted for 29% of the relatively few deaths that did occur. Similarly, it is very obvious from <xref ref-type="fig" rid="F0004">Fig. 4</xref> that external causes of death are a major problem in Bangladesh. This illustrates the importance of considering both population-based cause-specific mortality rates and CSMF when coming to any understanding of significant causes of mortality burdens within particular populations. Both parameters are essential information to have when considering interventions, either for specific diseases or for health promotion in general.</p><p>One of the strengths of maintaining surveillance of all deaths within particular populations, as the INDEPTH HDSS sites do, is that by definition the results for each particular cause constitute a clear component of 100% total mortality. By contrast, when deaths due to particular causes are documented at health facilities, or in vertical disease-oriented programmes, denominators are always unclear. In addition, the InterVA-4 methodology that was used in this dataset captures the uncertainty around cause of death assignment at the individual level, which is presented here as part of the &#x02018;indeterminate&#x02019; category. Although some studies have sought to reclassify so-called &#x02018;garbage&#x02019; cause of death codes into more specific groupings (<xref rid="CIT0033" ref-type="bibr">33</xref>), in reality the assignment of cause of death at the individual level is not something that can proceed with total certainty for every case, irrespective of the methods used, and this may be particularly true for some childhood deaths. We contend therefore that maintaining an &#x02018;indeterminate&#x02019; category that encompasses uncertainty in individual cause of death assignments, as well as accounting for a minority of deaths for which a VA interview was for some reason not possible, is an important and realistic concept in these population-based analyses (<xref rid="CIT0034" ref-type="bibr">34</xref>). Whether or not the &#x02018;indeterminate&#x02019; group actually constitutes a similar mix of causes of death as those that are successfully assigned has to remain a matter for conjecture.</p><p>Because childhood mortality globally is reported as falling, we have concentrated our analyses here on data from the 2006 to 2012 period to reflect a relatively contemporary scenario. A detailed comparative study by country, age group and cause of death between these results and other findings on cause-specific child mortality is beyond the scope of this paper. However, some selected comparisons can be made. The GBD2010 global estimates of child mortality (<xref rid="CIT0035" ref-type="bibr">35</xref>) are approximately contemporaneous with the 2006&#x02013;12 time period presented here, as are the 2010 mortality estimates presented by UNICEF in the State of the World's Children 2012 (<xref rid="CIT0036" ref-type="bibr">36</xref>). The basic rates of all-cause childhood mortality are reasonably congruent between these sources and the findings from the INDEPTH sites, although this is by no means a precise comparison (specific population site measurements versus national estimates). The perennial concern that some early neonatal deaths may have been considered as stillbirths (<xref rid="CIT0037" ref-type="bibr">37</xref>), and therefore not registered as deaths, may have been an issue at some of the INDEPTH sites in Africa that registered fairly low neonatal mortality rates in comparison to infant mortality. More systematic application of the WHO 2012 VA tool (<xref rid="CIT0011" ref-type="bibr">11</xref>) in the future may help to resolve this, since it contains a number of questions specifically aimed at making this distinction. Operationally, it is probably important to consider undertaking a VA interview for all third trimester pregnancies that do not result in a live baby, rather than making <italic>a priori</italic> distinctions between stillbirths and early neonatal deaths, in order to capture all available information. Local cultural and spiritual beliefs around the deaths of babies may also be an important consideration.</p></sec><sec sec-type="conclusions" id="S0005"><title>Conclusions</title><p>Individual deaths in childhood are always causes for great sadness; all the more so if the circumstances and the eventual cause of death mean that survival could have been reasonably possible. These analyses of individual deaths show that large numbers of children in Africa and Asia continue to die of avoidable causes, starting from suboptimal delivery care, through treatable infections and preventable accidents. Despite some countries achieving MDG4 targets, there is still room for further improvement. Documenting the magnitude of the various leading causes of childhood death, across relevant age groups, is a pre-requisite for planning effective survival interventions.</p></sec>

Cause-specific mortality in Africa and Asia: evidence from INDEPTH health and demographic surveillance system sites

<sec id="st1"><title>Background</title><p>Because most deaths in Africa and Asia are not well documented, estimates of mortality are often made using scanty data. The INDEPTH Network works to alleviate this problem by collating detailed individual data from defined Health and Demographic Surveillance sites. By registering all deaths over time and carrying out verbal autopsies to determine cause of death across many such sites, using standardised methods, the Network seeks to generate population-based mortality statistics that are not otherwise available.</p></sec><sec id="st2"><title>Objective</title><p>To build a large standardised mortality database from African and Asian sites, detailing the relevant methods, and use it to describe cause-specific mortality patterns.</p></sec><sec id="st3"><title>Design</title><p>Individual demographic and verbal autopsy (VA) data from 22 INDEPTH sites were collated into a standardised database. The INDEPTH 2013 population was used for standardisation. The WHO 2012 VA standard and the InterVA-4 model were used for assigning cause of death.</p></sec><sec id="st4"><title>Results</title><p>A total of 111,910 deaths occurring over 12,204,043 person-years (accumulated between 1992 and 2012) were registered across the 22 sites, and for 98,429 of these deaths (88.0%) verbal autopsies were successfully completed. There was considerable variation in all-cause mortality between sites, with most of the differences being accounted for by variations in infectious causes as a proportion of all deaths.</p></sec><sec id="st5"><title>Conclusions</title><p>This dataset documents individual deaths across Africa and Asia in a standardised way, and on an unprecedented scale. While INDEPTH sites are not constructed to constitute a representative sample, and VA may not be the ideal method of determining cause of death, nevertheless these findings represent detailed mortality patterns for parts of the world that are severely under-served in terms of measuring mortality. Further papers explore details of mortality patterns among children and specifically for NCDs, external causes, pregnancy-related mortality, malaria, and HIV/AIDS. Comparisons will also be made where possible with other findings on mortality in the same regions. Findings presented here and in accompanying papers support the need for continued work towards much wider implementation of universal civil registration of deaths by cause on a worldwide basis.</p></sec>

<contrib contrib-type="author"><name><surname>Streatfield</surname><given-names>P. Kim</given-names></name><xref ref-type="aff" rid="AF0001">1</xref><xref ref-type="aff" rid="AF0002">2</xref><xref ref-type="aff" rid="AF0003">3</xref></contrib><contrib contrib-type="author"><name><surname>Khan</surname><given-names>Wasif A.</given-names></name><xref ref-type="aff" rid="AF0002">2</xref><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0004">4</xref></contrib><contrib contrib-type="author"><name><surname>Bhuiya</surname><given-names>Abbas</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0005">5</xref><xref ref-type="aff" rid="AF0006">6</xref></contrib><contrib contrib-type="author"><name><surname>Alam</surname><given-names>Nurul</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0007">7</xref><xref ref-type="aff" rid="AF0008">8</xref></contrib><contrib contrib-type="author"><name><surname>Si&#x000e9;</surname><given-names>Ali</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0009">9</xref><xref ref-type="aff" rid="AF0010">10</xref></contrib><contrib contrib-type="author"><name><surname>Soura</surname><given-names>Abdramane B.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0011">11</xref><xref ref-type="aff" rid="AF0012">12</xref></contrib><contrib contrib-type="author"><name><surname>Bonfoh</surname><given-names>Bassirou</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0013">13</xref><xref ref-type="aff" rid="AF0014">14</xref></contrib><contrib contrib-type="author"><name><surname>Ngoran</surname><given-names>Eliezer K.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0013">13</xref><xref ref-type="aff" rid="AF0015">15</xref></contrib><contrib contrib-type="author"><name><surname>Weldearegawi</surname><given-names>Berhe</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0016">16</xref><xref ref-type="aff" rid="AF0017">17</xref></contrib><contrib contrib-type="author"><name><surname>Jasseh</surname><given-names>Momodou</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0018">18</xref><xref ref-type="aff" rid="AF0019">19</xref></contrib><contrib contrib-type="author"><name><surname>Oduro</surname><given-names>Abraham</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0020">20</xref><xref ref-type="aff" rid="AF0021">21</xref></contrib><contrib contrib-type="author"><name><surname>Gyapong</surname><given-names>Margaret</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0022">22</xref><xref ref-type="aff" rid="AF0023">23</xref></contrib><contrib contrib-type="author"><name><surname>Kant</surname><given-names>Shashi</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0024">24</xref><xref ref-type="aff" rid="AF0025">25</xref></contrib><contrib contrib-type="author"><name><surname>Juvekar</surname><given-names>Sanjay</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0026">26</xref><xref ref-type="aff" rid="AF0027">27</xref></contrib><contrib contrib-type="author"><name><surname>Wilopo</surname><given-names>Siswanto</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0028">28</xref><xref ref-type="aff" rid="AF0029">29</xref></contrib><contrib contrib-type="author"><name><surname>Williams</surname><given-names>Thomas N.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0030">30</xref><xref ref-type="aff" rid="AF0031">31</xref><xref ref-type="aff" rid="AF0032">32</xref></contrib><contrib contrib-type="author"><name><surname>Odhiambo</surname><given-names>Frank O.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0033">33</xref><xref ref-type="aff" rid="AF0034">34</xref></contrib><contrib contrib-type="author"><name><surname>Beguy</surname><given-names>Donatien</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0035">35</xref><xref ref-type="aff" rid="AF0036">36</xref></contrib><contrib contrib-type="author"><name><surname>Ezeh</surname><given-names>Alex</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0035">35</xref><xref ref-type="aff" rid="AF0036">36</xref></contrib><contrib contrib-type="author"><name><surname>Kyobutungi</surname><given-names>Catherine</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0035">35</xref><xref ref-type="aff" rid="AF0036">36</xref></contrib><contrib contrib-type="author"><name><surname>Crampin</surname><given-names>Amelia</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0037">37</xref><xref ref-type="aff" rid="AF0038">38</xref><xref ref-type="aff" rid="AF0039">39</xref></contrib><contrib contrib-type="author"><name><surname>Delaunay</surname><given-names>Val&#x000e9;rie</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0040">40</xref><xref ref-type="aff" rid="AF0041">41</xref></contrib><contrib contrib-type="author"><name><surname>Tollman</surname><given-names>Stephen M.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0042">42</xref><xref ref-type="aff" rid="AF0043">43</xref><xref ref-type="aff" rid="AF0044">44</xref></contrib><contrib contrib-type="author"><name><surname>Herbst</surname><given-names>Kobus</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0045">45</xref><xref ref-type="aff" rid="AF0046">46</xref></contrib><contrib contrib-type="author"><name><surname>Chuc</surname><given-names>Nguyen T.K.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0047">47</xref><xref ref-type="aff" rid="AF0048">48</xref></contrib><contrib contrib-type="author"><name><surname>Sankoh</surname><given-names>Osman A.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0049">49</xref><xref ref-type="aff" rid="AF0050">50</xref><xref ref-type="corresp" rid="cor1">*</xref></contrib><contrib contrib-type="author"><name><surname>Tanner</surname><given-names>Marcel</given-names></name><xref ref-type="aff" rid="AF0051">51</xref></contrib><contrib contrib-type="author"><name><surname>Byass</surname><given-names>Peter</given-names></name><xref ref-type="aff" rid="AF0043">43</xref><xref ref-type="aff" rid="AF0052">52</xref></contrib>

Global Health Action

<p>The vast majority of deaths in sub-Saharan Africa and southern Asia are not individually registered nor assigned a cause of death. Data from WHO (<xref rid="CIT0001" ref-type="bibr">1</xref>) show that, apart from in a few countries, the coverage of routine vital registration including cause of death in Africa and Asia is minimal and thus such cause-specific mortality data that do exist generally come from health facility records and <italic>ad hoc</italic> surveys. Therefore, when global estimates of cause-specific mortality are made, the data contributed from Africa and Asia are inevitably patchy and outcomes depend heavily on modelling assumptions that create huge uncertainty (<xref rid="CIT0002" ref-type="bibr">2</xref>). As a result, very little is accurately known about mortality patterns in these regions, but nonetheless policy, practice, and investment decisions are made that supposedly depend on knowledge of death rates and causes.</p><p>The INDEPTH Network (International Network for the Demographic Evaluation of Populations and their Health) is an umbrella organisation for a number of independent centres operating health and demographic surveillance system (HDSS) sites, most of which are located in sub-Saharan Africa and Asia (<xref rid="CIT0003" ref-type="bibr">3</xref>). These HDSS operations were started at various times and cover a range of defined rural and urban populations. Basic requirements in all the sites include registering all deaths occurring within the defined populations, and carrying out verbal autopsy (VA) procedures (interviews with relatives, care-givers, and witnesses after deaths have occurred, the results of which are subsequently interpreted into likely causes of death).</p><p>Any process of attributing cause of death, ranging from pathologists&#x02019; post-mortems through hospital cause of death records, physician certificates, and verbal autopsies, involves a combination of expertise and evidence (<xref rid="CIT0004" ref-type="bibr">4</xref>). Consequently, all causes of death data also incorporate some degree of uncertainty, which may include both systematic and random variations. Undertaking VA interviews and attributing causes of death are complex processes which need to be standardised as far as is possible. A WHO-led process resulted in new standard procedures for VA in 2012 (<xref rid="CIT0005" ref-type="bibr">5</xref>), in terms of defining questions that need to be included in VA interviews and VA cause of death categories corresponding to the International Classification of Diseases 10th Edition (ICD-10) (<xref rid="CIT0006" ref-type="bibr">6</xref>). A detailed review of that process, building on previous VA materials, is available (<xref rid="CIT0007" ref-type="bibr">7</xref>). A new version of the InterVA model for interpreting cause of death from VA data was also released in 2012 (<xref rid="CIT0008" ref-type="bibr">8</xref>), which exactly corresponds to the WHO 2012 VA standard in terms of VA questions and cause of death categories.</p><p>Consideration of the absolute validity of any cause of death data is also complex. A number of studies have made comparisons between pathologists&#x02019; post-mortems and hospital records (<xref rid="CIT0009" ref-type="bibr">9</xref>&#x02013;<xref rid="CIT0011" ref-type="bibr">11</xref>); others have compared validity between hospital records and mortality registers (<xref rid="CIT0012" ref-type="bibr">12</xref>), with varying degrees of concordance. In some cases, VA findings have been compared with hospital records, but this approach has been hampered by the generally small and unrepresentative proportions of deaths actually occurring in hospitals located in populations where VAs are used ( <xref rid="CIT0013" ref-type="bibr">13</xref>&#x02013;<xref rid="CIT0015" ref-type="bibr">15</xref>). Many studies have compared the use of automated VA coding models with results from physicians coding the same VA material (often termed physician-coded verbal autopsy (PCVA)), several of which have involved using InterVA models (<xref rid="CIT0016" ref-type="bibr">16</xref>). However, there can be difficulties in separating differences that arise from possible systematic errors in models and more random inter- or intra-physician variations (<xref rid="CIT0017" ref-type="bibr">17</xref>). For some specific causes of death, there can be absolute standards for comparison, for example, <italic>ante-mortem</italic> HIV or sickle cell status, but this only applies to a minority of causes (<xref rid="CIT0018" ref-type="bibr">18</xref>, <xref rid="CIT0019" ref-type="bibr">19</xref>). Many cause of death processes do not allow the attribution of uncertainty at the individual level. Although assigning a death as being 100% due to a particular cause simplifies further analyses, in reality there is a range of certainty associated with individual cause of death assignments, depending on the extent of available information for a particular case, as well as other factors. The probabilistic modelling used by InterVA-4 facilitates the capture of this uncertainty for each individual case, with the possibility of attributing part of a death as being of indeterminate cause (<xref rid="CIT0008" ref-type="bibr">8</xref>).</p><p>Despite the absence of widespread and reliable cause of death registration across Africa and Asia, much can be learnt about mortality patterns by considering standardised VA findings from sites where such data are routinely collected at the population level on a longitudinal basis. Mortality surveillance within circumscribed populations leads to findings based on every death that occurs. Consequently, cause-specific fractions total 100%, without the difficulties that some modelled estimates have encountered of needing to impose an overall mortality envelope. Furthermore, the advantages of consistency over time and place offered by VA models in assigning cause of death is particularly relevant for large epidemiological studies such as reported here, even though physicians might arguably follow a more nuanced approach in assigning individual causes of death.</p><p>The objectives of this introductory paper are to describe a large VA dataset compiled across a range of INDEPTH HDSS sites in Africa and Asia together with details of the overall methods used, as well as to report key findings on overall patterns of mortality and highlight areas of specific interest which have been examined in more detail in accompanying papers. Specifically, childhood mortality (<xref rid="CIT0020" ref-type="bibr">20</xref>) and adult non-communicable disease (NCD) mortality (<xref rid="CIT0021" ref-type="bibr">21</xref>), plus mortality from external causes (<xref rid="CIT0022" ref-type="bibr">22</xref>) and associated with pregnancy (<xref rid="CIT0023" ref-type="bibr">23</xref>), have been explored in more detail. Malaria (<xref rid="CIT0024" ref-type="bibr">24</xref>) and HIV/AIDS-related (<xref rid="CIT0025" ref-type="bibr">25</xref>) mortality, being two highly significant causes that vary considerably between sites, are also documented separately. Overall, findings and ways forward are brought together in a concluding synthesis (<xref rid="CIT0026" ref-type="bibr">26</xref>).</p><p>The publication of this series of papers coincides with depositing the overall cause of death dataset (<xref rid="CIT0027" ref-type="bibr">27</xref>) into the public domain at the INDEPTH Data Repository (<xref rid="CIT0028" ref-type="bibr">28</xref>). Half of the HDSSs involved are already part of INDEPTH's iSHARE programme (<ext-link ext-link-type="uri" xlink:href="http://www.indepth-ishare.org">www.indepth-ishare.org</ext-link>), and already have other individual-level population surveillance data in the public domain. The remaining sites have aggregated population data publicly available as part of the associated INDEPTHStats programme. As agreed by the INDEPTH Network Board, a separate set of anonymised identifiers have been used for the public domain cause of death data as distinct from other individual-level data, in order to guard against identity disclosure risks. Enquiries relating to specific research plans that would need to link the individual cause of death data with other individual-level data can be made to the INDEPTH Secretariat.</p><sec id="S0002"><title>Populations and methods</title><p>Cause of death data based on VA interviews were contributed by 14 INDEPTH HDSS sites in sub-Saharan Africa and eight sites in Asia, located as shown in <xref ref-type="fig" rid="F0001">Fig. 1</xref>. The principles of the Network and its constituent population surveillance sites have been described generically (<xref rid="CIT0003" ref-type="bibr">3</xref>), and detailed descriptions of the individual sites involved, including local attributes, are available elsewhere (<xref rid="CIT0029" ref-type="bibr">29</xref>&#x02013;<xref rid="CIT0050" ref-type="bibr">50</xref>). Each HDSS site is committed to long-term longitudinal surveillance of circumscribed populations, typically each covering around 50,000&#x02013;100,000 people. Households are registered and visited regularly by lay field-workers, with a frequency varying from once per year to several times per year. All vital events are registered at each such visit, and any deaths recorded are followed up with VA interviews, usually undertaken by specially trained lay interviewers. A few sites were already operational in the 1990s, but in this dataset 95% of the person-time observed related to the period from 2000 onwards, with 68% from 2006 onwards. Two sites, in Nairobi and Ouagadougou, followed urban populations, while the remainder covered areas that were generally more rural in character, although some included local urban centres. Sites covered entire populations, although the Karonga, Malawi, site only contributed VAs for deaths of people aged 12 years and older. Because the sites were not located or designed in a systematic way to be representative of national or regional populations, it is not meaningful to aggregate results over sites. Therefore, site-specific, cause-specific mortality fractions (CSMFs) and mortality rates (CSMRs) were used as the basis for analyses and comparisons. Since each site encompassed an entire non-sampled population, it was not meaningful to consider confidence intervals around site-specific measurements. Uncertainty around individual assignments of cause of death was however incorporated into the dataset as described below.</p><fig id="F0001" position="float"><label>Fig. 1</label><caption><p>Map showing locations of the 22 sites, with numbers of deaths registered, VAs completed, and person-years observed.</p></caption><graphic xlink:href="GHA-7-25362-g001"/></fig><p>Because there are possible inter-site and inter-year variations in the age&#x02013;sex structure of the populations concerned, it was also important, at least for some comparisons, to be able to adjust CSMFs and CSMRs to a standard population. Due to the different age and sex profiles of many causes of death, if this was not done then observed variations in CSMFs and CSMRs could have been due to (or masked by) differences in population structure between sites and over time. For this purpose, we took the INDEPTH 2013 standard population structure for low- and middle-income countries (LMICs) in Africa and Asia (<xref rid="CIT0051" ref-type="bibr">51</xref>), as shown in <xref ref-type="table" rid="T0001">Table 1</xref>. This public-domain standard population has been presented in relation to other global standards such as Segi and WHO, from which it differs in reflecting the higher fertility and younger-age mortality rates commonly seen in LMIC populations (<xref rid="CIT0051" ref-type="bibr">51</xref>). As shown in <xref ref-type="table" rid="T0001">Table 1</xref>, this standard has a very similar structure to the aggregated population from which these VA data came. Using the INDEPTH 2013 standard, encompassing lower income populations in both Africa and Asia, meant that the overall effect of standardisation across the whole mortality dataset was minimal, while resulting in important adjustments for certain sub-groups. Thus, calculating a standardised weight for every combination of site, age group, sex, and year made it possible to compare cause-specific mortality without concern for differences in underlying population structures (referred to hereafter as age&#x02013;sex&#x02013;time standardisation). Using the same INDEPTH 2013 standard, it will be possible to directly compare future work on these same lines with any mortality data from Africa and Asia.</p><table-wrap id="T0001" position="float"><label>Table 1</label><caption><p>Adapted INDEPTH 2013 standard population (51) by age group and sex, as percentage of total population (used for calculating adjusted cause-specific fractions and rates), compared with observed population structures in this dataset, overall and for African sites and Asian sites</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1"/><th align="center" colspan="2" rowspan="1">INDEPTH 2013 standard</th><th align="center" colspan="2" rowspan="1">All sites in this dataset</th><th align="center" colspan="2" rowspan="1">African sites in this dataset</th><th align="center" colspan="2" rowspan="1">Asian sites in this dataset</th></tr><tr><th align="left" rowspan="1" colspan="1"/><th colspan="8" rowspan="1"> <hr/> </th></tr><tr><th align="left" rowspan="1" colspan="1">Age group</th><th align="center" rowspan="1" colspan="1">Male (%)</th><th align="center" rowspan="1" colspan="1">Female (%)</th><th align="center" rowspan="1" colspan="1">Male (%)</th><th align="center" rowspan="1" colspan="1">Female (%)</th><th align="center" rowspan="1" colspan="1">Male (%)</th><th align="center" rowspan="1" colspan="1">Female (%)</th><th align="center" rowspan="1" colspan="1">Male (%)</th><th align="center" rowspan="1" colspan="1">Female (%)</th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">0&#x02013;28 days</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">0.12</td><td align="center" rowspan="1" colspan="1">0.12</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">0.08</td></tr><tr><td align="left" rowspan="1" colspan="1">1&#x02013;11 months</td><td align="center" rowspan="1" colspan="1">1.49</td><td align="center" rowspan="1" colspan="1">1.38</td><td align="center" rowspan="1" colspan="1">1.33</td><td align="center" rowspan="1" colspan="1">1.31</td><td align="center" rowspan="1" colspan="1">1.46</td><td align="center" rowspan="1" colspan="1">1.45</td><td align="center" rowspan="1" colspan="1">1.03</td><td align="center" rowspan="1" colspan="1">0.98</td></tr><tr><td align="left" rowspan="1" colspan="1">1&#x02013;4 years</td><td align="center" rowspan="1" colspan="1">6.01</td><td align="center" rowspan="1" colspan="1">5.57</td><td align="center" rowspan="1" colspan="1">5.62</td><td align="center" rowspan="1" colspan="1">5.48</td><td align="center" rowspan="1" colspan="1">6.07</td><td align="center" rowspan="1" colspan="1">5.99</td><td align="center" rowspan="1" colspan="1">4.51</td><td align="center" rowspan="1" colspan="1">4.24</td></tr><tr><td align="left" rowspan="1" colspan="1">5&#x02013;14 years</td><td align="center" rowspan="1" colspan="1">13.76</td><td align="center" rowspan="1" colspan="1">12.57</td><td align="center" rowspan="1" colspan="1">12.78</td><td align="center" rowspan="1" colspan="1">12.44</td><td align="center" rowspan="1" colspan="1">13.50</td><td align="center" rowspan="1" colspan="1">13.24</td><td align="center" rowspan="1" colspan="1">11.01</td><td align="center" rowspan="1" colspan="1">10.48</td></tr><tr><td align="left" rowspan="1" colspan="1">15&#x02013;49 years</td><td align="center" rowspan="1" colspan="1">22.54</td><td align="center" rowspan="1" colspan="1">23.50</td><td align="center" rowspan="1" colspan="1">22.62</td><td align="center" rowspan="1" colspan="1">25.39</td><td align="center" rowspan="1" colspan="1">21.74</td><td align="center" rowspan="1" colspan="1">24.71</td><td align="center" rowspan="1" colspan="1">24.79</td><td align="center" rowspan="1" colspan="1">27.05</td></tr><tr><td align="left" rowspan="1" colspan="1">50&#x02013;64 years</td><td align="center" rowspan="1" colspan="1">3.87</td><td align="center" rowspan="1" colspan="1">4.36</td><td align="center" rowspan="1" colspan="1">3.61</td><td align="center" rowspan="1" colspan="1">4.40</td><td align="center" rowspan="1" colspan="1">3.11</td><td align="center" rowspan="1" colspan="1">3.11</td><td align="center" rowspan="1" colspan="1">4.84</td><td align="center" rowspan="1" colspan="1">5.16</td></tr><tr><td align="left" rowspan="1" colspan="1">65+ years</td><td align="center" rowspan="1" colspan="1">2.22</td><td align="center" rowspan="1" colspan="1">2.52</td><td align="center" rowspan="1" colspan="1">2.18</td><td align="center" rowspan="1" colspan="1">2.62</td><td align="center" rowspan="1" colspan="1">1.96</td><td align="center" rowspan="1" colspan="1">1.96</td><td align="center" rowspan="1" colspan="1">2.73</td><td align="center" rowspan="1" colspan="1">3.04</td></tr></tbody></table></table-wrap><p>All causes of death assignments in this dataset were made using the InterVA-4 model version 4.02 (<xref rid="CIT0008" ref-type="bibr">8</xref>). InterVA-4 uses probabilistic modelling to arrive at likely cause(s) of death for each VA case, the workings of the model being based on a combination of expert medical opinion and relevant available data. InterVA-4 is the only model currently available that processes VA data according to the WHO 2012 standard and categorises causes of death according to ICD-10. Since the VA data reported here were collected before the WHO 2012 standard was formulated, they were all retrospectively transformed into the WHO 2012 and InterVA-4 input format for processing. The phrase &#x02018;successfully completed VA interview&#x02019; means that a VA interview was undertaken which yielded basic personal characteristics of the deceased (age, sex, etc.) and some symptoms relating to the final illness. The InterVA-4 &#x02018;high&#x02019; malaria setting was used for all the West African sites, plus the East African sites (with the exceptions, on the grounds of high altitude, of Nairobi, Kenya, and Kilite-Awlaelo, Ethiopia); other sites used the &#x02018;low&#x02019; setting. The InterVA-4 &#x02018;high&#x02019; HIV/AIDS setting was used for sites in Kenya, Malawi, and South Africa; for all other sites the &#x02018;low&#x02019; setting was used. These settings were chosen in line with InterVA recommendations and previous experience, and are discussed further in the specific papers on malaria and HIV/AIDS-related mortality (<xref rid="CIT0024" ref-type="bibr">24</xref>, <xref rid="CIT0025" ref-type="bibr">25</xref>).</p><p>The InterVA-4 model was applied to the data from each site, yielding, for each case, up to three possible causes of death or an indeterminate result. In a minority of cases, for example, where symptoms were vague, contradictory or mutually inconsistent, it was impossible for InterVA-4 to determine a cause of death, and these deaths were attributed as entirely indeterminate. For the remaining cases, one to three likely causes and their likelihoods were assigned by InterVA-4, and if the sum of their likelihoods was less than one, the residual component was then assigned as being indeterminate. This was an important process for capturing uncertainty in cause of death outcome(s) from the model at the individual level, thus avoiding over-interpretation of specific causes. As a consequence there were three sources of unattributed cause of death: deaths registered for which VAs were not successfully completed; VAs completed but where the cause was entirely indeterminate; and residual components of deaths attributed as indeterminate.</p><p>An overall dataset (<xref rid="CIT0027" ref-type="bibr">27</xref>) was compiled in which each case had between one and four records, each with its own cause and likelihood. Cases for which VAs were not successfully completed had single records with the cause of death recorded as &#x02018;VA not completed&#x02019; and a likelihood of one. Thus, the overall sum of the likelihoods equated to the total number of deaths. Each record also contained a population weighting factor reflecting the ratio of the population fraction for its site, age group, sex, and year to the corresponding age group and sex fraction in the standard population described in <xref ref-type="table" rid="T0001">Table 1</xref>, for the purposes of standardisation. Then a further factor was calculated for each record as the product of the VA cause likelihood and the population standard weighting (both described above), which could be used as the basis for calculating age&#x02013;sex&#x02013;time standardised CSMFs and CSMRs.</p><p>These descriptions of methods used to construct this multisite dataset (<xref rid="CIT0027" ref-type="bibr">27</xref>) apply to the following series of analytical papers using the dataset (<xref rid="CIT0020" ref-type="bibr">20</xref>&#x02013;<xref rid="CIT0025" ref-type="bibr">25</xref>). A standard Box summarising these methods is included in each of these papers.</p><p>In this context, all of these data are secondary datasets derived from primary data collected separately by each participating site. In all cases, the primary data collection was covered by site-level ethical approvals relating to on-going health and demographic surveillance in those specific locations. No individual identity or household location data were included in the secondary data and no specific ethical approvals were required for these pooled analyses.</p></sec><sec sec-type="results" id="S0003"><title>Results</title><p>A total of 111,910 deaths occurring over 12,204,043 person-years were registered across the 22 sites. For 98,429 of these deaths (88.0%), VAs were successfully completed. <xref ref-type="fig" rid="F0001">Figure 1</xref> includes the numbers of deaths, completed VAs and person-time observed for each site. Among the 98,429 completed VAs, InterVA-4 was unable to reach any conclusive cause of death (i.e. arrived at 100% indeterminate outcome) in 4,680 (4.8%) of cases. Residual indeterminate fractions totalled 7,545.9 (7.7%) of completed VAs. Thus, out of the total of 111,910 deaths recorded, specific causes were successfully assigned to 86,203 deaths (77.0%). Age&#x02013;sex&#x02013;time standardisation made less than 1% difference to the overall dataset (112,653 standardised deaths compared with 111,910 observed deaths) but was particularly important for some sites, for example, the urban slum population in Nairobi, where the population structure differed markedly from the standard.</p><p> <xref ref-type="table" rid="T0002">Table 2</xref> shows mortality rates per 1,000 person-years by age group and time period for each of the 22 participating sites. <xref ref-type="fig" rid="F0002">Figure 2</xref> shows age&#x02013;sex&#x02013;time standardised mortality rates for each site, by major categories of cause of death (infections, neoplasms, NCDs, maternal/neonatal, trauma, and indeterminate). The indeterminate category includes cases where VAs were not done, as well as indeterminate components reflecting individual uncertainty in cause of death assignment. All-cause age&#x02013;sex&#x02013;time standardised mortality rates in individual sites ranged from 18.5 per 1,000 person-years in Kisumu, Kenya, to 3.9 per 1,000 person-years in FilaBavi, Vietnam. A large part of this variation was accounted for by differences in infectious causes of death (10.7 per 1,000 person-years in Kisumu to 0.5 per 1,000 person-years in FilaBavi). A number of sites reflected low overall mortality rates as a consequence of being at stages of demographic transition where life expectancy increases as health standards improve, but where population proportions of elderly people remain relatively low.</p><fig id="F0002" position="float"><label>Fig. 2</label><caption><p>Age&#x02013;sex&#x02013;time-standardised mortality rates per 1,000 person-years by cause group and site for a total of 111,910 deaths over 12,204,043 person-years observed.</p></caption><graphic xlink:href="GHA-7-25362-g002"/></fig><table-wrap id="T0002" position="float"><label>Table 2</label><caption><p>Mortality rates per 1,000 person-years by site, period, and age group for a total of 111,910 deaths observed over 12,204,043 person-years at INDEPTH sites in Africa and Asia</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1"/><th align="left" rowspan="1" colspan="1"/><th align="center" colspan="2" rowspan="1">0&#x02013;28 days</th><th align="center" colspan="2" rowspan="1">1&#x02013;11 months</th><th align="center" colspan="2" rowspan="1">1&#x02013;4 years</th><th align="center" colspan="2" rowspan="1">5&#x02013;14 years</th><th align="center" colspan="2" rowspan="1">15&#x02013;49 years</th><th align="center" colspan="2" rowspan="1">50&#x02013;64 years</th><th align="center" colspan="2" rowspan="1">65+ years</th></tr><tr><th align="left" rowspan="1" colspan="1"/><th align="left" rowspan="1" colspan="1"/><th colspan="14" rowspan="1"> <hr/> </th></tr><tr><th align="left" rowspan="1" colspan="1">Site</th><th align="center" rowspan="1" colspan="1">Period</th><th align="center" rowspan="1" colspan="1">Male</th><th align="center" rowspan="1" colspan="1">Female</th><th align="center" rowspan="1" colspan="1">Male</th><th align="center" rowspan="1" colspan="1">Female</th><th align="center" rowspan="1" colspan="1">Male</th><th align="center" rowspan="1" colspan="1">Female</th><th align="center" rowspan="1" colspan="1">Male</th><th align="center" rowspan="1" colspan="1">Female</th><th align="center" rowspan="1" colspan="1">Male</th><th align="center" rowspan="1" colspan="1">Female</th><th align="center" rowspan="1" colspan="1">Male</th><th align="center" rowspan="1" colspan="1">Female</th><th align="center" rowspan="1" colspan="1">Male</th><th align="center" rowspan="1" colspan="1">Female</th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">Matlab, Bangladesh</td><td align="left" rowspan="1" colspan="1">2000&#x02013;05</td><td align="center" rowspan="1" colspan="1">471.1</td><td align="center" rowspan="1" colspan="1">417.3</td><td align="center" rowspan="1" colspan="1">10.6</td><td align="center" rowspan="1" colspan="1">13.3</td><td align="center" rowspan="1" colspan="1">3.5</td><td align="center" rowspan="1" colspan="1">3.2</td><td align="center" rowspan="1" colspan="1">0.8</td><td align="center" rowspan="1" colspan="1">0.9</td><td align="center" rowspan="1" colspan="1">2.2</td><td align="center" rowspan="1" colspan="1">1.5</td><td align="center" rowspan="1" colspan="1">14.1</td><td align="center" rowspan="1" colspan="1">8.0</td><td align="center" rowspan="1" colspan="1">65.9</td><td align="center" rowspan="1" colspan="1">57.3</td></tr><tr><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1">2006&#x02013;12</td><td align="center" rowspan="1" colspan="1">379.8</td><td align="center" rowspan="1" colspan="1">272.1</td><td align="center" rowspan="1" colspan="1">7.5</td><td align="center" rowspan="1" colspan="1">8.5</td><td align="center" rowspan="1" colspan="1">2.6</td><td align="center" rowspan="1" colspan="1">2.7</td><td align="center" rowspan="1" colspan="1">0.6</td><td align="center" rowspan="1" colspan="1">0.5</td><td align="center" rowspan="1" colspan="1">1.9</td><td align="center" rowspan="1" colspan="1">1.3</td><td align="center" rowspan="1" colspan="1">14.0</td><td align="center" rowspan="1" colspan="1">7.2</td><td align="center" rowspan="1" colspan="1">62.7</td><td align="center" rowspan="1" colspan="1">54.9</td></tr><tr><td align="left" rowspan="1" colspan="1">Bandarban, Bangladesh</td><td align="left" rowspan="1" colspan="1">2006&#x02013;12</td><td align="center" rowspan="1" colspan="1">265.5</td><td align="center" rowspan="1" colspan="1">70.6</td><td align="center" rowspan="1" colspan="1">38.9</td><td align="center" rowspan="1" colspan="1">17.8</td><td align="center" rowspan="1" colspan="1">0.7</td><td align="center" rowspan="1" colspan="1">3.1</td><td align="center" rowspan="1" colspan="1">1.5</td><td align="center" rowspan="1" colspan="1">0.6</td><td align="center" rowspan="1" colspan="1">3.0</td><td align="center" rowspan="1" colspan="1">2.5</td><td align="center" rowspan="1" colspan="1">10.0</td><td align="center" rowspan="1" colspan="1">9.7</td><td align="center" rowspan="1" colspan="1">43.1</td><td align="center" rowspan="1" colspan="1">35.7</td></tr><tr><td align="left" rowspan="1" colspan="1">Chakaria, Bangladesh</td><td align="left" rowspan="1" colspan="1">2006&#x02013;12</td><td align="center" rowspan="1" colspan="1">506.4</td><td align="center" rowspan="1" colspan="1">405.5</td><td align="center" rowspan="1" colspan="1">11.6</td><td align="center" rowspan="1" colspan="1">22.0</td><td align="center" rowspan="1" colspan="1">3.9</td><td align="center" rowspan="1" colspan="1">4.0</td><td align="center" rowspan="1" colspan="1">1.1</td><td align="center" rowspan="1" colspan="1">0.9</td><td align="center" rowspan="1" colspan="1">2.0</td><td align="center" rowspan="1" colspan="1">1.8</td><td align="center" rowspan="1" colspan="1">12.4</td><td align="center" rowspan="1" colspan="1">13.2</td><td align="center" rowspan="1" colspan="1">64.0</td><td align="center" rowspan="1" colspan="1">68.9</td></tr><tr><td align="left" rowspan="1" colspan="1">AMK, Bangladesh</td><td align="left" rowspan="1" colspan="1">2000&#x02013;05</td><td align="center" rowspan="1" colspan="1">454.2</td><td align="center" rowspan="1" colspan="1">325.4</td><td align="center" rowspan="1" colspan="1">9.4</td><td align="center" rowspan="1" colspan="1">14.2</td><td align="center" rowspan="1" colspan="1">3.4</td><td align="center" rowspan="1" colspan="1">3.1</td><td align="center" rowspan="1" colspan="1">0.8</td><td align="center" rowspan="1" colspan="1">0.8</td><td align="center" rowspan="1" colspan="1">1.8</td><td align="center" rowspan="1" colspan="1">1.6</td><td align="center" rowspan="1" colspan="1">11.5</td><td align="center" rowspan="1" colspan="1">7.0</td><td align="center" rowspan="1" colspan="1">64.4</td><td align="center" rowspan="1" colspan="1">59.1</td></tr><tr><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1">2006&#x02013;12</td><td align="center" rowspan="1" colspan="1">413.9</td><td align="center" rowspan="1" colspan="1">298.1</td><td align="center" rowspan="1" colspan="1">8.1</td><td align="center" rowspan="1" colspan="1">13.2</td><td align="center" rowspan="1" colspan="1">2.3</td><td align="center" rowspan="1" colspan="1">2.2</td><td align="center" rowspan="1" colspan="1">0.7</td><td align="center" rowspan="1" colspan="1">0.5</td><td align="center" rowspan="1" colspan="1">2.1</td><td align="center" rowspan="1" colspan="1">1.4</td><td align="center" rowspan="1" colspan="1">11.8</td><td align="center" rowspan="1" colspan="1">8.3</td><td align="center" rowspan="1" colspan="1">68.3</td><td align="center" rowspan="1" colspan="1">55.5</td></tr><tr><td align="left" rowspan="1" colspan="1">Nouna, Burkina Faso</td><td align="left" rowspan="1" colspan="1">1992&#x02013;99</td><td align="center" rowspan="1" colspan="1">114.1</td><td align="center" rowspan="1" colspan="1">87.6</td><td align="center" rowspan="1" colspan="1">39.8</td><td align="center" rowspan="1" colspan="1">38.8</td><td align="center" rowspan="1" colspan="1">28.1</td><td align="center" rowspan="1" colspan="1">31.6</td><td align="center" rowspan="1" colspan="1">8.1</td><td align="center" rowspan="1" colspan="1">4.3</td><td align="center" rowspan="1" colspan="1">8.1</td><td align="center" rowspan="1" colspan="1">5.4</td><td align="center" rowspan="1" colspan="1">21.7</td><td align="center" rowspan="1" colspan="1">28.7</td><td align="center" rowspan="1" colspan="1">92.5</td><td align="center" rowspan="1" colspan="1">56.7</td></tr><tr><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1">2000&#x02013;05</td><td align="center" rowspan="1" colspan="1">148.0</td><td align="center" rowspan="1" colspan="1">136.4</td><td align="center" rowspan="1" colspan="1">43.0</td><td align="center" rowspan="1" colspan="1">41.9</td><td align="center" rowspan="1" colspan="1">20.0</td><td align="center" rowspan="1" colspan="1">18.4</td><td align="center" rowspan="1" colspan="1">2.8</td><td align="center" rowspan="1" colspan="1">2.5</td><td align="center" rowspan="1" colspan="1">4.7</td><td align="center" rowspan="1" colspan="1">4.4</td><td align="center" rowspan="1" colspan="1">18.2</td><td align="center" rowspan="1" colspan="1">13.2</td><td align="center" rowspan="1" colspan="1">65.1</td><td align="center" rowspan="1" colspan="1">64.6</td></tr><tr><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1">2006&#x02013;12</td><td align="center" rowspan="1" colspan="1">101.4</td><td align="center" rowspan="1" colspan="1">84.3</td><td align="center" rowspan="1" colspan="1">24.6</td><td align="center" rowspan="1" colspan="1">24.2</td><td align="center" rowspan="1" colspan="1">14.4</td><td align="center" rowspan="1" colspan="1">11.1</td><td align="center" rowspan="1" colspan="1">1.7</td><td align="center" rowspan="1" colspan="1">1.5</td><td align="center" rowspan="1" colspan="1">3.3</td><td align="center" rowspan="1" colspan="1">2.2</td><td align="center" rowspan="1" colspan="1">15.8</td><td align="center" rowspan="1" colspan="1">10.6</td><td align="center" rowspan="1" colspan="1">60.5</td><td align="center" rowspan="1" colspan="1">43.3</td></tr><tr><td align="left" rowspan="1" colspan="1">Ouagadougou, Burkina Faso</td><td align="left" rowspan="1" colspan="1">2006&#x02013;12</td><td align="center" rowspan="1" colspan="1">167.2</td><td align="center" rowspan="1" colspan="1">106.9</td><td align="center" rowspan="1" colspan="1">20.5</td><td align="center" rowspan="1" colspan="1">21.2</td><td align="center" rowspan="1" colspan="1">8.1</td><td align="center" rowspan="1" colspan="1">7.4</td><td align="center" rowspan="1" colspan="1">1.5</td><td align="center" rowspan="1" colspan="1">1.2</td><td align="center" rowspan="1" colspan="1">2.0</td><td align="center" rowspan="1" colspan="1">1.9</td><td align="center" rowspan="1" colspan="1">15.9</td><td align="center" rowspan="1" colspan="1">7.5</td><td align="center" rowspan="1" colspan="1">51.9</td><td align="center" rowspan="1" colspan="1">32.5</td></tr><tr><td align="left" rowspan="1" colspan="1">Taabo, C&#x000f4;te d'Ivoire</td><td align="left" rowspan="1" colspan="1">2006&#x02013;12</td><td align="center" rowspan="1" colspan="1">202.7</td><td align="center" rowspan="1" colspan="1">198.8</td><td align="center" rowspan="1" colspan="1">29.6</td><td align="center" rowspan="1" colspan="1">34.3</td><td align="center" rowspan="1" colspan="1">15.4</td><td align="center" rowspan="1" colspan="1">15.0</td><td align="center" rowspan="1" colspan="1">1.7</td><td align="center" rowspan="1" colspan="1">1.8</td><td align="center" rowspan="1" colspan="1">5.4</td><td align="center" rowspan="1" colspan="1">4.8</td><td align="center" rowspan="1" colspan="1">16.4</td><td align="center" rowspan="1" colspan="1">11.7</td><td align="center" rowspan="1" colspan="1">61.4</td><td align="center" rowspan="1" colspan="1">45.6</td></tr><tr><td align="left" rowspan="1" colspan="1">Kilite-Awaleo, Ethiopia</td><td align="left" rowspan="1" colspan="1">2006&#x02013;12</td><td align="center" rowspan="1" colspan="1">244.2</td><td align="center" rowspan="1" colspan="1">130.0</td><td align="center" rowspan="1" colspan="1">12.8</td><td align="center" rowspan="1" colspan="1">12.4</td><td align="center" rowspan="1" colspan="1">3.4</td><td align="center" rowspan="1" colspan="1">2.3</td><td align="center" rowspan="1" colspan="1">1.2</td><td align="center" rowspan="1" colspan="1">0.9</td><td align="center" rowspan="1" colspan="1">2.2</td><td align="center" rowspan="1" colspan="1">1.8</td><td align="center" rowspan="1" colspan="1">6.2</td><td align="center" rowspan="1" colspan="1">6.5</td><td align="center" rowspan="1" colspan="1">36.4</td><td align="center" rowspan="1" colspan="1">32.6</td></tr><tr><td align="left" rowspan="1" colspan="1">Farafenni, The Gambia</td><td align="left" rowspan="1" colspan="1">1992&#x02013;99</td><td align="center" rowspan="1" colspan="1">333.3</td><td align="center" rowspan="1" colspan="1">320.0</td><td align="center" rowspan="1" colspan="1">47.5</td><td align="center" rowspan="1" colspan="1">76.2</td><td align="center" rowspan="1" colspan="1">42.2</td><td align="center" rowspan="1" colspan="1">22.4</td><td align="center" rowspan="1" colspan="1">7.3</td><td align="center" rowspan="1" colspan="1">4.0</td><td align="center" rowspan="1" colspan="1">5.9</td><td align="center" rowspan="1" colspan="1">5.2</td><td align="center" rowspan="1" colspan="1">34.7</td><td align="center" rowspan="1" colspan="1">18.6</td><td align="center" rowspan="1" colspan="1">97.0</td><td align="center" rowspan="1" colspan="1">67.2</td></tr><tr><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1">2000&#x02013;05</td><td align="center" rowspan="1" colspan="1">202.2</td><td align="center" rowspan="1" colspan="1">212.9</td><td align="center" rowspan="1" colspan="1">35.1</td><td align="center" rowspan="1" colspan="1">29.0</td><td align="center" rowspan="1" colspan="1">12.9</td><td align="center" rowspan="1" colspan="1">13.1</td><td align="center" rowspan="1" colspan="1">2.1</td><td align="center" rowspan="1" colspan="1">2.1</td><td align="center" rowspan="1" colspan="1">4.9</td><td align="center" rowspan="1" colspan="1">4.4</td><td align="center" rowspan="1" colspan="1">24.8</td><td align="center" rowspan="1" colspan="1">14.9</td><td align="center" rowspan="1" colspan="1">80.3</td><td align="center" rowspan="1" colspan="1">64.9</td></tr><tr><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1">2006&#x02013;12</td><td align="center" rowspan="1" colspan="1">286.8</td><td align="center" rowspan="1" colspan="1">127.8</td><td align="center" rowspan="1" colspan="1">22.5</td><td align="center" rowspan="1" colspan="1">17.8</td><td align="center" rowspan="1" colspan="1">6.4</td><td align="center" rowspan="1" colspan="1">6.0</td><td align="center" rowspan="1" colspan="1">1.8</td><td align="center" rowspan="1" colspan="1">1.2</td><td align="center" rowspan="1" colspan="1">4.5</td><td align="center" rowspan="1" colspan="1">3.7</td><td align="center" rowspan="1" colspan="1">21.7</td><td align="center" rowspan="1" colspan="1">11.7</td><td align="center" rowspan="1" colspan="1">85.7</td><td align="center" rowspan="1" colspan="1">53.5</td></tr><tr><td align="left" rowspan="1" colspan="1">Navrongo, Ghana</td><td align="left" rowspan="1" colspan="1">2000&#x02013;05</td><td align="center" rowspan="1" colspan="1">327.5</td><td align="center" rowspan="1" colspan="1">283.3</td><td align="center" rowspan="1" colspan="1">46.1</td><td align="center" rowspan="1" colspan="1">40.8</td><td align="center" rowspan="1" colspan="1">11.6</td><td align="center" rowspan="1" colspan="1">11.3</td><td align="center" rowspan="1" colspan="1">2.6</td><td align="center" rowspan="1" colspan="1">1.7</td><td align="center" rowspan="1" colspan="1">6.6</td><td align="center" rowspan="1" colspan="1">4.8</td><td align="center" rowspan="1" colspan="1">33.7</td><td align="center" rowspan="1" colspan="1">18.1</td><td align="center" rowspan="1" colspan="1">74.7</td><td align="center" rowspan="1" colspan="1">61.3</td></tr><tr><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1">2006&#x02013;12</td><td align="center" rowspan="1" colspan="1">240.2</td><td align="center" rowspan="1" colspan="1">179.4</td><td align="center" rowspan="1" colspan="1">22.8</td><td align="center" rowspan="1" colspan="1">21.2</td><td align="center" rowspan="1" colspan="1">8.5</td><td align="center" rowspan="1" colspan="1">7.9</td><td align="center" rowspan="1" colspan="1">2.0</td><td align="center" rowspan="1" colspan="1">1.5</td><td align="center" rowspan="1" colspan="1">6.0</td><td align="center" rowspan="1" colspan="1">3.2</td><td align="center" rowspan="1" colspan="1">28.6</td><td align="center" rowspan="1" colspan="1">13.7</td><td align="center" rowspan="1" colspan="1">69.6</td><td align="center" rowspan="1" colspan="1">49.7</td></tr><tr><td align="left" rowspan="1" colspan="1">Dodowa, Ghana</td><td align="left" rowspan="1" colspan="1">2006&#x02013;12</td><td align="center" rowspan="1" colspan="1">115.7</td><td align="center" rowspan="1" colspan="1">63.9</td><td align="center" rowspan="1" colspan="1">7.9</td><td align="center" rowspan="1" colspan="1">9.6</td><td align="center" rowspan="1" colspan="1">4.9</td><td align="center" rowspan="1" colspan="1">4.6</td><td align="center" rowspan="1" colspan="1">1.4</td><td align="center" rowspan="1" colspan="1">1.3</td><td align="center" rowspan="1" colspan="1">4.8</td><td align="center" rowspan="1" colspan="1">4.6</td><td align="center" rowspan="1" colspan="1">20.6</td><td align="center" rowspan="1" colspan="1">14.5</td><td align="center" rowspan="1" colspan="1">62.6</td><td align="center" rowspan="1" colspan="1">50.4</td></tr><tr><td align="left" rowspan="1" colspan="1">Ballabgarh, India</td><td align="left" rowspan="1" colspan="1">2006&#x02013;12</td><td align="center" rowspan="1" colspan="1">274.8</td><td align="center" rowspan="1" colspan="1">286.1</td><td align="center" rowspan="1" colspan="1">21.8</td><td align="center" rowspan="1" colspan="1">27.6</td><td align="center" rowspan="1" colspan="1">3.2</td><td align="center" rowspan="1" colspan="1">4.9</td><td align="center" rowspan="1" colspan="1">0.7</td><td align="center" rowspan="1" colspan="1">1.0</td><td align="center" rowspan="1" colspan="1">3.4</td><td align="center" rowspan="1" colspan="1">1.9</td><td align="center" rowspan="1" colspan="1">19.0</td><td align="center" rowspan="1" colspan="1">9.2</td><td align="center" rowspan="1" colspan="1">75.3</td><td align="center" rowspan="1" colspan="1">54.0</td></tr><tr><td align="left" rowspan="1" colspan="1">Vadu, India</td><td align="left" rowspan="1" colspan="1">2006&#x02013;12</td><td align="center" rowspan="1" colspan="1">88.6</td><td align="center" rowspan="1" colspan="1">117.4</td><td align="center" rowspan="1" colspan="1">3.7</td><td align="center" rowspan="1" colspan="1">4.4</td><td align="center" rowspan="1" colspan="1">1.5</td><td align="center" rowspan="1" colspan="1">1.2</td><td align="center" rowspan="1" colspan="1">0.5</td><td align="center" rowspan="1" colspan="1">0.0</td><td align="center" rowspan="1" colspan="1">1.8</td><td align="center" rowspan="1" colspan="1">1.2</td><td align="center" rowspan="1" colspan="1">12.2</td><td align="center" rowspan="1" colspan="1">8.1</td><td align="center" rowspan="1" colspan="1">43.3</td><td align="center" rowspan="1" colspan="1">43.4</td></tr><tr><td align="left" rowspan="1" colspan="1">Purworejo, Indonesia</td><td align="left" rowspan="1" colspan="1">2000&#x02013;05</td><td align="center" rowspan="1" colspan="1">212.3</td><td align="center" rowspan="1" colspan="1">132.2</td><td align="center" rowspan="1" colspan="1">10.9</td><td align="center" rowspan="1" colspan="1">6.9</td><td align="center" rowspan="1" colspan="1">2.4</td><td align="center" rowspan="1" colspan="1">2.4</td><td align="center" rowspan="1" colspan="1">0.9</td><td align="center" rowspan="1" colspan="1">0.5</td><td align="center" rowspan="1" colspan="1">2.1</td><td align="center" rowspan="1" colspan="1">1.6</td><td align="center" rowspan="1" colspan="1">15.5</td><td align="center" rowspan="1" colspan="1">13.6</td><td align="center" rowspan="1" colspan="1">60.9</td><td align="center" rowspan="1" colspan="1">54.0</td></tr><tr><td align="left" rowspan="1" colspan="1">Kilifi, Kenya</td><td align="left" rowspan="1" colspan="1">2006&#x02013;12</td><td align="center" rowspan="1" colspan="1">176.4</td><td align="center" rowspan="1" colspan="1">143.4</td><td align="center" rowspan="1" colspan="1">10.0</td><td align="center" rowspan="1" colspan="1">9.1</td><td align="center" rowspan="1" colspan="1">2.8</td><td align="center" rowspan="1" colspan="1">2.3</td><td align="center" rowspan="1" colspan="1">0.9</td><td align="center" rowspan="1" colspan="1">0.7</td><td align="center" rowspan="1" colspan="1">3.0</td><td align="center" rowspan="1" colspan="1">2.9</td><td align="center" rowspan="1" colspan="1">16.4</td><td align="center" rowspan="1" colspan="1">10.4</td><td align="center" rowspan="1" colspan="1">61.2</td><td align="center" rowspan="1" colspan="1">42.5</td></tr><tr><td align="left" rowspan="1" colspan="1">Kisumu, Kenya</td><td align="left" rowspan="1" colspan="1">2000&#x02013;05</td><td align="center" rowspan="1" colspan="1">298.4</td><td align="center" rowspan="1" colspan="1">307.0</td><td align="center" rowspan="1" colspan="1">110.7</td><td align="center" rowspan="1" colspan="1">112.7</td><td align="center" rowspan="1" colspan="1">32.3</td><td align="center" rowspan="1" colspan="1">31.1</td><td align="center" rowspan="1" colspan="1">2.5</td><td align="center" rowspan="1" colspan="1">2.8</td><td align="center" rowspan="1" colspan="1">18.0</td><td align="center" rowspan="1" colspan="1">19.0</td><td align="center" rowspan="1" colspan="1">40.6</td><td align="center" rowspan="1" colspan="1">22.7</td><td align="center" rowspan="1" colspan="1">75.4</td><td align="center" rowspan="1" colspan="1">52.7</td></tr><tr><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1">2006&#x02013;12</td><td align="center" rowspan="1" colspan="1">260.9</td><td align="center" rowspan="1" colspan="1">224.9</td><td align="center" rowspan="1" colspan="1">75.6</td><td align="center" rowspan="1" colspan="1">72.9</td><td align="center" rowspan="1" colspan="1">22.6</td><td align="center" rowspan="1" colspan="1">23.0</td><td align="center" rowspan="1" colspan="1">2.6</td><td align="center" rowspan="1" colspan="1">2.2</td><td align="center" rowspan="1" colspan="1">12.0</td><td align="center" rowspan="1" colspan="1">11.4</td><td align="center" rowspan="1" colspan="1">30.3</td><td align="center" rowspan="1" colspan="1">18.7</td><td align="center" rowspan="1" colspan="1">75.5</td><td align="center" rowspan="1" colspan="1">56.3</td></tr><tr><td align="left" rowspan="1" colspan="1">Nairobi, Kenya</td><td align="left" rowspan="1" colspan="1">2000&#x02013;05</td><td align="center" rowspan="1" colspan="1">428.4</td><td align="center" rowspan="1" colspan="1">320.9</td><td align="center" rowspan="1" colspan="1">67.3</td><td align="center" rowspan="1" colspan="1">48.8</td><td align="center" rowspan="1" colspan="1">10.0</td><td align="center" rowspan="1" colspan="1">6.9</td><td align="center" rowspan="1" colspan="1">2.4</td><td align="center" rowspan="1" colspan="1">1.8</td><td align="center" rowspan="1" colspan="1">9.6</td><td align="center" rowspan="1" colspan="1">5.6</td><td align="center" rowspan="1" colspan="1">39.9</td><td align="center" rowspan="1" colspan="1">10.6</td><td align="center" rowspan="1" colspan="1">55.2</td><td align="center" rowspan="1" colspan="1">36.0</td></tr><tr><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1">2006&#x02013;12</td><td align="center" rowspan="1" colspan="1">342.5</td><td align="center" rowspan="1" colspan="1">299.3</td><td align="center" rowspan="1" colspan="1">51.0</td><td align="center" rowspan="1" colspan="1">48.6</td><td align="center" rowspan="1" colspan="1">7.1</td><td align="center" rowspan="1" colspan="1">5.8</td><td align="center" rowspan="1" colspan="1">1.1</td><td align="center" rowspan="1" colspan="1">1.0</td><td align="center" rowspan="1" colspan="1">8.2</td><td align="center" rowspan="1" colspan="1">4.0</td><td align="center" rowspan="1" colspan="1">35.9</td><td align="center" rowspan="1" colspan="1">6.6</td><td align="center" rowspan="1" colspan="1">48.8</td><td align="center" rowspan="1" colspan="1">40.4</td></tr><tr><td align="left" rowspan="1" colspan="1">Karonga, Malawi</td><td align="left" rowspan="1" colspan="1">2000&#x02013;05</td><td align="center" rowspan="1" colspan="1">354.5</td><td align="center" rowspan="1" colspan="1">228.4</td><td align="center" rowspan="1" colspan="1">34.4</td><td align="center" rowspan="1" colspan="1">27.0</td><td align="center" rowspan="1" colspan="1">9.1</td><td align="center" rowspan="1" colspan="1">9.2</td><td align="center" rowspan="1" colspan="1">2.3</td><td align="center" rowspan="1" colspan="1">1.0</td><td align="center" rowspan="1" colspan="1">8.3</td><td align="center" rowspan="1" colspan="1">9.4</td><td align="center" rowspan="1" colspan="1">23.3</td><td align="center" rowspan="1" colspan="1">15.7</td><td align="center" rowspan="1" colspan="1">48.9</td><td align="center" rowspan="1" colspan="1">54.3</td></tr><tr><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1">2006&#x02013;12</td><td align="center" rowspan="1" colspan="1">293.5</td><td align="center" rowspan="1" colspan="1">219.8</td><td align="center" rowspan="1" colspan="1">23.2</td><td align="center" rowspan="1" colspan="1">21.9</td><td align="center" rowspan="1" colspan="1">8.0</td><td align="center" rowspan="1" colspan="1">6.9</td><td align="center" rowspan="1" colspan="1">1.2</td><td align="center" rowspan="1" colspan="1">1.4</td><td align="center" rowspan="1" colspan="1">4.9</td><td align="center" rowspan="1" colspan="1">4.2</td><td align="center" rowspan="1" colspan="1">18.4</td><td align="center" rowspan="1" colspan="1">11.5</td><td align="center" rowspan="1" colspan="1">45.5</td><td align="center" rowspan="1" colspan="1">44.5</td></tr><tr><td align="left" rowspan="1" colspan="1">Niakhar, Senegal</td><td align="left" rowspan="1" colspan="1">2000&#x02013;05</td><td align="center" rowspan="1" colspan="1">235.9</td><td align="center" rowspan="1" colspan="1">184.7</td><td align="center" rowspan="1" colspan="1">37.1</td><td align="center" rowspan="1" colspan="1">25.6</td><td align="center" rowspan="1" colspan="1">18.4</td><td align="center" rowspan="1" colspan="1">22.6</td><td align="center" rowspan="1" colspan="1">3.4</td><td align="center" rowspan="1" colspan="1">3.0</td><td align="center" rowspan="1" colspan="1">5.1</td><td align="center" rowspan="1" colspan="1">5.8</td><td align="center" rowspan="1" colspan="1">13.6</td><td align="center" rowspan="1" colspan="1">15.3</td><td align="center" rowspan="1" colspan="1">99.6</td><td align="center" rowspan="1" colspan="1">79.8</td></tr><tr><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1">2006&#x02013;12</td><td align="center" rowspan="1" colspan="1">148.1</td><td align="center" rowspan="1" colspan="1">105.1</td><td align="center" rowspan="1" colspan="1">17.2</td><td align="center" rowspan="1" colspan="1">16.3</td><td align="center" rowspan="1" colspan="1">9.6</td><td align="center" rowspan="1" colspan="1">10.2</td><td align="center" rowspan="1" colspan="1">1.9</td><td align="center" rowspan="1" colspan="1">1.1</td><td align="center" rowspan="1" colspan="1">3.3</td><td align="center" rowspan="1" colspan="1">2.8</td><td align="center" rowspan="1" colspan="1">13.9</td><td align="center" rowspan="1" colspan="1">8.2</td><td align="center" rowspan="1" colspan="1">67.1</td><td align="center" rowspan="1" colspan="1">49.2</td></tr><tr><td align="left" rowspan="1" colspan="1">Agincourt, South Africa</td><td align="left" rowspan="1" colspan="1">1992&#x02013;99</td><td align="center" rowspan="1" colspan="1">103.6</td><td align="center" rowspan="1" colspan="1">58.0</td><td align="center" rowspan="1" colspan="1">12.0</td><td align="center" rowspan="1" colspan="1">15.1</td><td align="center" rowspan="1" colspan="1">4.4</td><td align="center" rowspan="1" colspan="1">4.4</td><td align="center" rowspan="1" colspan="1">0.8</td><td align="center" rowspan="1" colspan="1">0.7</td><td align="center" rowspan="1" colspan="1">3.8</td><td align="center" rowspan="1" colspan="1">3.0</td><td align="center" rowspan="1" colspan="1">13.0</td><td align="center" rowspan="1" colspan="1">10.6</td><td align="center" rowspan="1" colspan="1">30.6</td><td align="center" rowspan="1" colspan="1">56.7</td></tr><tr><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1">2000&#x02013;05</td><td align="center" rowspan="1" colspan="1">133.3</td><td align="center" rowspan="1" colspan="1">105.9</td><td align="center" rowspan="1" colspan="1">32.5</td><td align="center" rowspan="1" colspan="1">27.9</td><td align="center" rowspan="1" colspan="1">7.7</td><td align="center" rowspan="1" colspan="1">6.3</td><td align="center" rowspan="1" colspan="1">1.0</td><td align="center" rowspan="1" colspan="1">0.9</td><td align="center" rowspan="1" colspan="1">8.7</td><td align="center" rowspan="1" colspan="1">8.9</td><td align="center" rowspan="1" colspan="1">23.7</td><td align="center" rowspan="1" colspan="1">20.9</td><td align="center" rowspan="1" colspan="1">34.4</td><td align="center" rowspan="1" colspan="1">70.7</td></tr><tr><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1">2006&#x02013;12</td><td align="center" rowspan="1" colspan="1">159.1</td><td align="center" rowspan="1" colspan="1">150.2</td><td align="center" rowspan="1" colspan="1">31.7</td><td align="center" rowspan="1" colspan="1">30.1</td><td align="center" rowspan="1" colspan="1">5.7</td><td align="center" rowspan="1" colspan="1">4.9</td><td align="center" rowspan="1" colspan="1">1.5</td><td align="center" rowspan="1" colspan="1">1.2</td><td align="center" rowspan="1" colspan="1">9.6</td><td align="center" rowspan="1" colspan="1">8.6</td><td align="center" rowspan="1" colspan="1">24.4</td><td align="center" rowspan="1" colspan="1">25.1</td><td align="center" rowspan="1" colspan="1">33.2</td><td align="center" rowspan="1" colspan="1">84.0</td></tr><tr><td align="left" rowspan="1" colspan="1">Africa Centre, South Africa</td><td align="left" rowspan="1" colspan="1">2000&#x02013;05</td><td align="center" rowspan="1" colspan="1">169.7</td><td align="center" rowspan="1" colspan="1">132.0</td><td align="center" rowspan="1" colspan="1">49.5</td><td align="center" rowspan="1" colspan="1">49.5</td><td align="center" rowspan="1" colspan="1">9.0</td><td align="center" rowspan="1" colspan="1">8.8</td><td align="center" rowspan="1" colspan="1">1.6</td><td align="center" rowspan="1" colspan="1">1.8</td><td align="center" rowspan="1" colspan="1">18.7</td><td align="center" rowspan="1" colspan="1">16.9</td><td align="center" rowspan="1" colspan="1">46.4</td><td align="center" rowspan="1" colspan="1">22.6</td><td align="center" rowspan="1" colspan="1">88.2</td><td align="center" rowspan="1" colspan="1">49.6</td></tr><tr><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1">2006&#x02013;12</td><td align="center" rowspan="1" colspan="1">56.1</td><td align="center" rowspan="1" colspan="1">49.9</td><td align="center" rowspan="1" colspan="1">28.7</td><td align="center" rowspan="1" colspan="1">26.3</td><td align="center" rowspan="1" colspan="1">5.3</td><td align="center" rowspan="1" colspan="1">4.1</td><td align="center" rowspan="1" colspan="1">1.3</td><td align="center" rowspan="1" colspan="1">1.1</td><td align="center" rowspan="1" colspan="1">12.9</td><td align="center" rowspan="1" colspan="1">11.1</td><td align="center" rowspan="1" colspan="1">46.2</td><td align="center" rowspan="1" colspan="1">21.2</td><td align="center" rowspan="1" colspan="1">82.3</td><td align="center" rowspan="1" colspan="1">52.4</td></tr><tr><td align="left" rowspan="1" colspan="1">FilaBavi, Vietnam</td><td align="left" rowspan="1" colspan="1">2006&#x02013;12</td><td align="center" rowspan="1" colspan="1">144.5</td><td align="center" rowspan="1" colspan="1">92.3</td><td align="center" rowspan="1" colspan="1">1.2</td><td align="center" rowspan="1" colspan="1">4.3</td><td align="center" rowspan="1" colspan="1">1.1</td><td align="center" rowspan="1" colspan="1">0.8</td><td align="center" rowspan="1" colspan="1">0.6</td><td align="center" rowspan="1" colspan="1">0.1</td><td align="center" rowspan="1" colspan="1">3.0</td><td align="center" rowspan="1" colspan="1">1.2</td><td align="center" rowspan="1" colspan="1">9.6</td><td align="center" rowspan="1" colspan="1">3.1</td><td align="center" rowspan="1" colspan="1">50.7</td><td align="center" rowspan="1" colspan="1">33.3</td></tr></tbody></table></table-wrap></sec><sec sec-type="discussion" id="S0004"><title>Discussion</title><p>This dataset documents individual deaths across sub-Saharan Africa and southern Asia on a hitherto unprecedented scale. In addition, because the deaths are recorded in the context of longitudinal surveillance operations, it is also possible to determine population-based mortality rates, overall and by year, cause, age group, and sex. The application of the WHO 2012 VA standard (<xref rid="CIT0007" ref-type="bibr">7</xref>) and the InterVA-4 (<xref rid="CIT0008" ref-type="bibr">8</xref>) model to these data enabled assignment of cause of death in a consistent manner, and, where appropriate, age&#x02013;sex&#x02013;time standardisation of rates enabled systematic comparisons between sites.</p><p>This dataset provided opportunities for a wide range of more detailed analyses, as reflected in the following series of papers. Apart from looking in more detail at causes of death within obvious population sub-groups such as neonates, infants, and under-5s (<xref rid="CIT0020" ref-type="bibr">20</xref>), it is also interesting to consider the detail within some of the cause groups shown in <xref ref-type="fig" rid="F0002">Figure 2</xref>. It would appear that population-based rates of NCD mortality are relatively constant across the sites, compared with the variations in overall mortality, which seem to be more strongly driven by the magnitude of infectious causes (<xref rid="CIT0021" ref-type="bibr">21</xref>). Although there is much reported on so-called epidemics of NCDs in LMICs, this may partly reflect relatively large proportions of NCD mortality in populations also experiencing historically low levels of overall mortality, particularly in Asia. These low overall rates are demographically driven in populations experiencing rapid life expectancy increases, but not yet having accumulated larger proportions of older people. Nevertheless, concerns over current accumulations of NCD risk factors are very valid in relation to future NCD mortality. These results on NCD mortality may therefore constitute an important baseline measurement against which to judge future developments. External causes of death &#x02013; intentional and unintentional, self-inflicted and imposed &#x02013; also constitute an increasingly large component of mortality in various places and age groups, which have been explored further in this dataset (<xref rid="CIT0022" ref-type="bibr">22</xref>). This was also a sufficiently large dataset to look in more detail at some specific causes of death, such as pregnancy-related deaths (<xref rid="CIT0023" ref-type="bibr">23</xref>) and infectious causes such as malaria (<xref rid="CIT0024" ref-type="bibr">24</xref>) and HIV/AIDS (<xref rid="CIT0025" ref-type="bibr">25</xref>). Availability of the dataset in the public domain will facilitate further analyses of specific causes and patterns of mortality.</p><p>Although HDSS sites such as those contributing data here can be critiqued in terms of how representative they may be of surrounding areas, this consideration has to be offset against the unique opportunities HDSSs have of assessing mortality patterns for complete communities, rather than the more commonly used health facility mortality records. Most deaths in Africa and Asia do not occur in health facilities, and it is by no means evident that facility-based deaths are very representative of mortality in general. Cause of death as determined by VA may also be regarded as less than ideal, but it is the only viable method for large-scale cause of death assignment in Africa and Asia. Here, the WHO 2012 VA standard and the InterVA-4 model have been used to ensure, as far as possible, that there is consistency across the sites and time periods involved, which is not guaranteed with physician interpretation of VA. Many sites also undertake physician interpretation of their VAs, which may differ in detail from these results, and be reported separately; we are not making any comparisons with physician findings here. We acknowledge that it was a compromise to have to transform VA data collected using a range of historical instruments, but that was unavoidable. Most of the instruments used had evolved from earlier WHO and INDEPTH VA versions with much common core content, which were themselves the starting point for the development of the WHO 2012 standard. The application of age&#x02013;sex&#x02013;time standardisation, using the INDEPTH 2013 population standard, further enabled comparisons of mortality over time and place to be made objectively.</p><p>A Population Health Metrics Research Consortium (PHMRC) study aimed to collect a &#x02018;gold standard&#x02019; VA dataset from selected tertiary institutions, which has been used both to build VA models and evaluate different approaches to VA cause of death assignment (<xref rid="CIT0052" ref-type="bibr">52</xref>). Although that study concluded that InterVA-4 was less effective than PHMRC methods in assigning cause of death, that conclusion was reached by comparing the internal validity of PHMRC methods within the &#x02018;gold standard&#x02019; dataset against the external validity of InterVA-4 and physician coding (<xref rid="CIT0053" ref-type="bibr">53</xref>). Issues with the quality of the PHMRC VA data, the use of different VA questions and deviations from ICD-10 classifications further compromised those findings, but nevertheless InterVA-4 coding of the PHMRC data still demonstrated an overall concordance correlation of 0.61. Since InterVA-4 is the only available VA model which corresponds to the WHO 2012 VA standard and ICD-10 coded causes of death, it was the preferred choice to use here.</p><p>A number of sites contributing data to the overall dataset have undertaken site-specific analyses of their mortality patterns which are reported separately (<xref rid="CIT0054" ref-type="bibr">54</xref>&#x02013;<xref rid="CIT0066" ref-type="bibr">66</xref>). In some countries (Bangladesh, Ghana, Burkina Faso, Kenya, South Africa) there were multiple sites involved which present interesting opportunities to consider within-country variations. This also facilitates comparisons with other national sources of data such as Demographic and Health Surveys and Global Burden of Disease outputs. To some extent, it was also possible to look at trends in mortality over time, although that was limited by the different time periods over which individual sites have been operating. Since more sites have reported for recent periods, there may be findings here of interest in terms of trends towards the 2015 deadlines for Millennium Development Goals. These may also serve as baseline figures for post-2015 targets.</p><p>Most of the detailed comparisons made between results from this dataset and comparable figures from various other sources of estimates, explored in the accompanying papers, showed a high degree of congruence. Given that the methodologies involved &#x02013; of counting individual deaths at INDEPTH sites and aggregating upwards, contrasted with taking available data sources and constructing global models to derive national estimates &#x02013; are completely different, this congruence in findings adds plausibility to both approaches. Nevertheless, it must still be recognised that moving towards complete civil registration of deaths, including cause of death, is a critical objective yet to be achieved in Africa and Asia (<xref rid="CIT0067" ref-type="bibr">67</xref>).</p></sec>

Who died of what in rural KwaZulu-Natal, South Africa: a cause of death analysis using InterVA-4

<sec id="st1"><title>Background</title><p>For public health purposes, it is important to see whether men and women in different age groups die of the same causes in South Africa.</p></sec><sec id="st2"><title>Objective</title><p>We explored sex- and age-specific patterns of causes of deaths in a rural demographic surveillance site in northern KwaZulu-Natal in South Africa over the period 2000&#x02013;2011.</p></sec><sec id="st3"><title>Design</title><p>Deaths reported through the demographic surveillance were followed up by a verbal autopsy (VA) interview using a standardised questionnaire. Causes of death were assigned likelihoods using the publicly available tool InterVA-4. Cause-specific mortality fractions were determined by age and sex.</p></sec><sec id="st4"><title>Results</title><p>Over the study period, a total of 5,416 (47%) and 6,081 (53%) deaths were recorded in men and women, respectively. Major causes of death proportionally affecting more women than men were (all <italic>p</italic>&#x0003c;0.0001): human immunodeficiency virus/acquired immunodeficiency syndrome (HIV/AIDS) (20.1% vs. 13.6%), other and unspecified cardiac disease (5.9% vs. 3.2%), stroke (4.5% vs. 2.7%), reproductive neoplasms (1.7% vs. 0.4%), diabetes (2.4% vs. 1.2%), and breast neoplasms (0.4% vs. 0%). Major causes of deaths proportionally affecting more men than women were (all <italic>p</italic>&#x0003c;0.0001) assault (6.1% vs. 1.7%), pulmonary tuberculosis (34.5% vs. 30.2%), road traffic accidents (3.0% vs. 1.0%), intentional self-harm (1.3% vs. 0.3%), and respiratory neoplasms (2.5% vs. 1.5%). Causes of death due to communicable diseases predominated in all age groups except in older persons.</p></sec><sec id="st5"><title>Conclusions</title><p>While mortality during the 2000s was dominated by tuberculosis and HIV/AIDS, we found substantial sex-specific differences both for communicable and non-communicable causes of death, some which can be explained by a differing sex-specific age structure. InterVA-4 is likely to be a valuable tool for investigating causes of death patterns in other similar Southern African settings.</p></sec>

<contrib contrib-type="author"><name><surname>Mossong</surname><given-names>Jo&#x000eb;l</given-names></name><xref ref-type="aff" rid="AF0001">1</xref><xref ref-type="aff" rid="AF0002">2</xref><xref ref-type="corresp" rid="cor1">*</xref></contrib><contrib contrib-type="author"><name><surname>Byass</surname><given-names>Peter</given-names></name><xref ref-type="aff" rid="AF0003">3</xref></contrib><contrib contrib-type="author"><name><surname>Herbst</surname><given-names>Kobus</given-names></name><xref ref-type="aff" rid="AF0001">1</xref></contrib>

Global Health Action

<p>Despite the fact that the rationale for any successful health interventions should be informed by the cause of deaths (CODs) that are of the greatest importance locally, many developing countries lack this crucial information (<xref rid="CIT0001" ref-type="bibr">1</xref>). Verbal autopsy (VA) has been shown as a useful tool in such settings to establish a probable COD by interviewing a close caregiver to give details on circumstances and symptoms leading up to the death event (<xref rid="CIT0002" ref-type="bibr">2</xref>).</p><p>The Africa Centre for Health and Population Studies has hosted a longitudinal demographic surveillance system (DSS) in a rural area in South Africa which enables detailed analyses of CODs. Following the introduction in 2004 and massive scale-up thereafter of an antiretroviral treatment (ART) programme, mortality rates in adults and young children have declined significantly (<xref rid="CIT0003" ref-type="bibr">3</xref>, <xref rid="CIT0004" ref-type="bibr">4</xref>). This decline has been essentially driven by a substantial decrease of deaths due to HIV/AIDS (<xref rid="CIT0005" ref-type="bibr">5</xref>). In particular, the cause-specific mortality fraction (CSMF) of HIV-related causes declined from 56% in 2002 to 39% in 2009 and all-cause mortality rates have dropped by 33%. More recently, it was shown that life expectancy increased from 49.2 years in 2004 to 60.5 years in 2011 (<xref rid="CIT0006" ref-type="bibr">6</xref>).</p><p>In 2012, WHO published a revised shortened VA questionnaire (<xref rid="CIT0007" ref-type="bibr">7</xref>) which can be used in combination with a publicly available tool InterVA-4 to assign causes of death (<xref rid="CIT0008" ref-type="bibr">8</xref>). In our and other rural contexts in Africa, the previous version InterVA-3 was shown to give substantial agreement with physician-coded causes of death (<xref rid="CIT0005" ref-type="bibr">5</xref>, <xref rid="CIT0009" ref-type="bibr">9</xref>&#x02013;<xref rid="CIT0011" ref-type="bibr">11</xref>). A comparison between InterVA-3 and InterVA-4 using data from another demographic surveillance site in South Africa has shown continuity of interpretation (<xref rid="CIT0008" ref-type="bibr">8</xref>).</p><p>The objective of this study is to further explore sex- and age-specific patterns of causes of deaths in our setting using the new tool InterVA-4 (<xref rid="CIT0008" ref-type="bibr">8</xref>). We will analyse the major causes of deaths for men and women and focus particularly on sex and age disparities.</p><sec sec-type="methods" id="S0001"><title>Methods</title><sec id="S0001-S20001"><title>Study population</title><p>The DSS, which is located in the rural district of uMkhanyakude in northern KwaZulu-Natal in South Africa, has been running since January 2000. Roughly 11,000 households with approximately 60,000 resident and 30,000 non-resident members were visited twice a year to collect information on births, migrations, and deaths, as well as other demographic, health and socioeconomic data. The population is poor, has high HIV prevalence (<xref rid="CIT0012" ref-type="bibr">12</xref>) and high incidence (<xref rid="CIT0013" ref-type="bibr">13</xref>), although incidence is lower in areas with high ART coverage (<xref rid="CIT0014" ref-type="bibr">14</xref>). The current analysis pertains to deaths of DSS residents reported to have occurred between 2000 and 2011.</p><p>All deaths reported by the demographic surveillance were followed up by a verbal autopsy (VA) interview conducted by a trained nurse with the closest available caregiver of the deceased. The VA interview was conducted at the earliest 3 months after the date of death, but on average 10 months after the death. It included an open narrative of the circumstances leading up to the death, a checklist of signs and symptoms, and a structured questionnaire substantially similar to the INDEPTH/WHO VA questionnaire (<xref rid="CIT0003" ref-type="bibr">3</xref>). The check list and structured questionnaire were entered twice by different data capturers and validated by a third person.</p></sec><sec id="S0001-S20002"><title>COD assignment using InterVA-4</title><p>For assigning causes of death, we downloaded the publicly available tool InterVA-4 (version 4.02) (<xref rid="CIT0015" ref-type="bibr">15</xref>) and ran it on a standard Microsoft Windows 7 PC. InterVA-4 takes as input a text file where each row consists of death cases. The columns in the input file consist of one name and 245 indicator variables in a predefined order. Indicator is the terminology used by InterVA-4 to describe the whole range of items of information about the circumstances of a death, including basic background characteristics, details of any illness (signs and symptoms) leading to death, previous medical history, etc. All indicators are coded as &#x02018;y&#x02019; for yes or &#x02018;n&#x02019; for no. In addition, InterVA-4 requires the user to specify HIV and malaria prevalence, which were set on &#x02018;high&#x02019; and &#x02018;low&#x02019;, respectively, for our setting. The questionnaire and check list data from the local database were extracted using Microsoft SQL and converted into the WHO 2012 standard VA format required by InterVA-4 (<xref rid="CIT0016" ref-type="bibr">16</xref>). Of the 245 input indicators, 56 (23%) indicators were not available in the version of VA questionnaire used in this study.</p><p>InterVA-4 produced an output file containing for each death one to three causes of deaths and their respective likelihoods (<xref rid="CIT0008" ref-type="bibr">8</xref>). One of the interesting features of InterVA-4 is that it explicitly allowed for multiple causes of death and enabled assignment of any residual likelihood for any individual case to be assigned as indeterminate. The residual likelihood is the fraction of the likelihood that could not be attributed to any particular COD, which can be considered as an individual-level expression of uncertainty in the COD assignment.</p><p>The full data set is available via the INDEPTH Network Data Repository (<xref rid="CIT0017" ref-type="bibr">17</xref>).</p></sec><sec id="S0001-S20003"><title>Statistical analysis</title><p>CSMFs were determined overall and by population subgroup (age or sex) by summing likelihoods across all one to three possible causes of death as determined by InterVA-4 and dividing by the sum of the likelihoods for all causes (<xref rid="CIT0018" ref-type="bibr">18</xref>). For each COD, hypothesis testing for the difference of mortality fractions by sex was performed using the immediate form of two-sample test of proportion command &#x02018;prtesti&#x02019; in Stata 12.1 (StataCorp, College Station, USA). Age-dependent mortality rates were calculated by dividing estimated number of deaths for a given cause by the person-years of observation. Hypothesis testing for the difference of mortality rates by sex in each age group was performed using the command &#x02018;iri&#x02019; in Stata 12.1 (StataCorp, College Station, USA).</p></sec></sec><sec sec-type="results" id="S0002"><title>Results</title><p>Following the demographic surveillance reports of 11,497 deaths between 2000 and 2011, 10,958 (95.3%) VA interviews were successfully conducted which were used as input data for InterVA-4. Approximately one in six deaths occurred before the age of 15 years, almost half of deaths occurred in adulthood between the ages of 15 and 49 years, and approximately one third in persons aged 50 years and older (<xref ref-type="table" rid="T0001">Table 1</xref>). There were 5,416 (47%) and 6,081 (53%) deaths recorded in men and women, respectively. Deaths in males outnumbered those in women for neonates, infants, young children aged 1&#x02013;4 years, and adults aged 50&#x02013;64 years. Deaths in women outnumbered those in men for adults aged 15&#x02013;49 years and persons aged 65 years or older.</p><table-wrap id="T0001" position="float"><label>Table 1</label><caption><p>Distribution by age and sex of death counts, person years (PY) of observation and mortality rates of residents at Africa Centre demographic surveillance area in rural KwaZulu-Natal, 2000&#x02013;2011</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1"/><th align="center" colspan="2" rowspan="1">Death counts</th><th align="center" colspan="2" rowspan="1">PY</th><th align="center" colspan="2" rowspan="1">Mortality rate (per 1,000)</th></tr><tr><th align="left" rowspan="1" colspan="1"/><th align="center" colspan="2" rowspan="1"> <hr/> </th><th align="center" colspan="2" rowspan="1"> <hr/> </th><th align="center" colspan="2" rowspan="1"> <hr/> </th></tr><tr><th align="left" rowspan="1" colspan="1">Sex</th><th align="center" rowspan="1" colspan="1">Male</th><th align="center" rowspan="1" colspan="1">Female</th><th align="center" rowspan="1" colspan="1">Male</th><th align="center" rowspan="1" colspan="1">Female</th><th align="center" rowspan="1" colspan="1">Male</th><th align="center" rowspan="1" colspan="1">Female</th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">Neonate (&#x0003c;28 days)</td><td align="center" rowspan="1" colspan="1">96 (56%)</td><td align="center" rowspan="1" colspan="1">75 (44%)</td><td align="center" rowspan="1" colspan="1">840</td><td align="center" rowspan="1" colspan="1">818</td><td align="center" rowspan="1" colspan="1">114.3</td><td align="center" rowspan="1" colspan="1">91.7</td></tr><tr><td align="left" rowspan="1" colspan="1">Infant (1&#x02013;11 months)</td><td align="center" rowspan="1" colspan="1">412 (52%)</td><td align="center" rowspan="1" colspan="1">388 (49%)</td><td align="center" rowspan="1" colspan="1">10,560</td><td align="center" rowspan="1" colspan="1">10,250</td><td align="center" rowspan="1" colspan="1">39.0</td><td align="center" rowspan="1" colspan="1">37.9</td></tr><tr><td align="left" rowspan="1" colspan="1">1&#x02013;4 years</td><td align="center" rowspan="1" colspan="1">327 (53%)</td><td align="center" rowspan="1" colspan="1">290 (47%)</td><td align="center" rowspan="1" colspan="1">46,161</td><td align="center" rowspan="1" colspan="1">45,205</td><td align="center" rowspan="1" colspan="1">7.1</td><td align="center" rowspan="1" colspan="1">6.4</td></tr><tr><td align="left" rowspan="1" colspan="1">5&#x02013;14 years</td><td align="center" rowspan="1" colspan="1">172 (50%)</td><td align="center" rowspan="1" colspan="1">169 (50%)</td><td align="center" rowspan="1" colspan="1">117,615</td><td align="center" rowspan="1" colspan="1">115,347</td><td align="center" rowspan="1" colspan="1">1.5</td><td align="center" rowspan="1" colspan="1">1.5</td></tr><tr><td align="left" rowspan="1" colspan="1">15&#x02013;49 years</td><td align="center" rowspan="1" colspan="1">2,572 (47%)</td><td align="center" rowspan="1" colspan="1">2,945 (53%)</td><td align="center" rowspan="1" colspan="1">163,257</td><td align="center" rowspan="1" colspan="1">210,842</td><td align="center" rowspan="1" colspan="1">15.8</td><td align="center" rowspan="1" colspan="1">14.0</td></tr><tr><td align="left" rowspan="1" colspan="1">50&#x02013;64 years</td><td align="center" rowspan="1" colspan="1">889 (53%)</td><td align="center" rowspan="1" colspan="1">781 (47%)</td><td align="center" rowspan="1" colspan="1">19,202</td><td align="center" rowspan="1" colspan="1">35,650</td><td align="center" rowspan="1" colspan="1">46.3</td><td align="center" rowspan="1" colspan="1">21.9</td></tr><tr><td align="left" rowspan="1" colspan="1">65+ years</td><td align="center" rowspan="1" colspan="1">948 (40%)</td><td align="center" rowspan="1" colspan="1">1,433 (60%)</td><td align="center" rowspan="1" colspan="1">11,122</td><td align="center" rowspan="1" colspan="1">28,038</td><td align="center" rowspan="1" colspan="1">85.2</td><td align="center" rowspan="1" colspan="1">51.1</td></tr><tr><td align="left" rowspan="1" colspan="1">Total</td><td align="center" rowspan="1" colspan="1">5,416 (47%)</td><td align="center" rowspan="1" colspan="1">6,081 (53%)</td><td align="center" rowspan="1" colspan="1">368,758</td><td align="center" rowspan="1" colspan="1">446,150</td><td align="center" rowspan="1" colspan="1">14.7</td><td align="center" rowspan="1" colspan="1">13.6</td></tr></tbody></table></table-wrap><p>The following COD categories, which can potentially be assigned by InterVA-4, were not observed in our population: haemorrhagic fever, sickle cell with crisis, non-road transport accident, obstructed labour, and ruptured uterus. Overall, the fraction of deaths for which no COD could be determined was 10.6% and this did not vary significantly between men and women (10.7 and 10.5%, respectively, <italic>p</italic>&#x0003e;0.05). This fraction of unknown causes of death can be further divided into three categories: for 539 (4.7%) deaths no VA interview had been conducted, so obviously no information was available to assign a probable COD; for a further 192 (1.7%) deaths, although the VA interview had been conducted, the quality and breadth of information collected was insufficient or too unspecific for InterVA-4 to be able to assign any causes of death; and finally, the remaining 4.2% represented the residual likelihood of indeterminate causes of death.</p><p> <xref ref-type="table" rid="T0002">Table 2</xref> shows that CSMFs differed substantially between women and men in our population. Major causes of death proportionally affecting more women than men (ordered by decreasing difference and with a <italic>p</italic>&#x0003c;0.0001) were HIV/AIDS (20.1% vs. 13.6%), other and unspecified cardiac disease (5.9% vs. 3.2%), stroke (4.5% vs. 2.7%), reproductive neoplasms (1.7% vs. 0.4%), diabetes (2.4% vs. 1.2%), and breast neoplasms (0.4% vs. 0%). Major causes of deaths proportionally affecting more men than women (ordered by decreasing difference and with a <italic>p</italic>&#x0003c;0.0001) were assault (6.1% vs. 1.7%), pulmonary tuberculosis (34.5% vs. 30.2%), road traffic accidents (3.0% vs. 1.0%), intentional self-harm (1.3% vs. 0.3%), and respiratory neoplasms (2.5% vs. 1.5%).</p><table-wrap id="T0002" position="float"><label>Table 2</label><caption><p>Cause-specific mortality fractions expressed as percentages in rural KwaZulu-Natal by sex</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1">WHO VA cause of death code</th><th align="center" rowspan="1" colspan="1">Male (<italic>N</italic>=5,416)</th><th align="center" rowspan="1" colspan="1">Female (<italic>N</italic>=6,081)</th><th align="center" rowspan="1" colspan="1">Total (<italic>N</italic>=11,597)</th><th align="center" rowspan="1" colspan="1">F&#x02013;M difference</th><th align="center" rowspan="1" colspan="1"> <italic>p</italic> </th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">01.01 Sepsis (non-obstetric)</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">&#x02212;0.01</td><td align="center" rowspan="1" colspan="1">0.69</td></tr><tr><td align="left" rowspan="1" colspan="1">01.02 Acute resp infect incl. pneumonia</td><td align="center" rowspan="1" colspan="1">7.15</td><td align="center" rowspan="1" colspan="1">6.87</td><td align="center" rowspan="1" colspan="1">7.00</td><td align="center" rowspan="1" colspan="1">&#x02212;0.27</td><td align="center" rowspan="1" colspan="1">0.56</td></tr><tr><td align="left" rowspan="1" colspan="1">01.03 HIV/AIDS related death</td><td align="center" rowspan="1" colspan="1">13.55</td><td align="center" rowspan="1" colspan="1">20.12</td><td align="center" rowspan="1" colspan="1">17.03</td><td align="center" rowspan="1" colspan="1">6.57</td><td align="center" rowspan="1" colspan="1">&#x0003c;0.0001</td></tr><tr><td align="left" rowspan="1" colspan="1">01.04 Diarrhoeal diseases</td><td align="center" rowspan="1" colspan="1">0.65</td><td align="center" rowspan="1" colspan="1">0.55</td><td align="center" rowspan="1" colspan="1">0.60</td><td align="center" rowspan="1" colspan="1">&#x02212;0.11</td><td align="center" rowspan="1" colspan="1">0.46</td></tr><tr><td align="left" rowspan="1" colspan="1">01.05 Malaria</td><td align="center" rowspan="1" colspan="1">0.39</td><td align="center" rowspan="1" colspan="1">0.45</td><td align="center" rowspan="1" colspan="1">0.42</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.61</td></tr><tr><td align="left" rowspan="1" colspan="1">01.06 Measles</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">&#x02212;0.06</td><td align="center" rowspan="1" colspan="1">0.18</td></tr><tr><td align="left" rowspan="1" colspan="1">01.07 Meningitis and encephalitis</td><td align="center" rowspan="1" colspan="1">0.79</td><td align="center" rowspan="1" colspan="1">1.00</td><td align="center" rowspan="1" colspan="1">0.90</td><td align="center" rowspan="1" colspan="1">0.21</td><td align="center" rowspan="1" colspan="1">0.23</td></tr><tr><td align="left" rowspan="1" colspan="1">01.08 &#x00026; 10.05 Tetanus</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">&#x02212;0.01</td><td align="center" rowspan="1" colspan="1">0.35</td></tr><tr><td align="left" rowspan="1" colspan="1">01.09 Pulmonary tuberculosis</td><td align="center" rowspan="1" colspan="1">34.52</td><td align="center" rowspan="1" colspan="1">30.15</td><td align="center" rowspan="1" colspan="1">32.21</td><td align="center" rowspan="1" colspan="1">&#x02212;4.37</td><td align="center" rowspan="1" colspan="1">&#x0003c;0.0001</td></tr><tr><td align="left" rowspan="1" colspan="1">01.10 Pertussis</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">&#x02212;0.01</td><td align="center" rowspan="1" colspan="1">0.71</td></tr><tr><td align="left" rowspan="1" colspan="1">01.99 Other and unspecified infect dis</td><td align="center" rowspan="1" colspan="1">0.22</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">0.16</td><td align="center" rowspan="1" colspan="1">&#x02212;0.12</td><td align="center" rowspan="1" colspan="1">0.11</td></tr><tr><td align="left" rowspan="1" colspan="1">02.01 Oral neoplasms</td><td align="center" rowspan="1" colspan="1">0.12</td><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">&#x02212;0.03</td><td align="center" rowspan="1" colspan="1">0.61</td></tr><tr><td align="left" rowspan="1" colspan="1">02.02 Digestive neoplasms</td><td align="center" rowspan="1" colspan="1">1.82</td><td align="center" rowspan="1" colspan="1">1.20</td><td align="center" rowspan="1" colspan="1">1.49</td><td align="center" rowspan="1" colspan="1">&#x02212;0.62</td><td align="center" rowspan="1" colspan="1">0.006</td></tr><tr><td align="left" rowspan="1" colspan="1">02.03 Respiratory neoplasms</td><td align="center" rowspan="1" colspan="1">2.53</td><td align="center" rowspan="1" colspan="1">1.46</td><td align="center" rowspan="1" colspan="1">1.96</td><td align="center" rowspan="1" colspan="1">&#x02212;1.07</td><td align="center" rowspan="1" colspan="1">&#x0003c;0.0001</td></tr><tr><td align="left" rowspan="1" colspan="1">02.04 Breast neoplasms</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.44</td><td align="center" rowspan="1" colspan="1">0.23</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">02.05 &#x00026; 02.06 Reproductive neoplasms MF</td><td align="center" rowspan="1" colspan="1">0.44</td><td align="center" rowspan="1" colspan="1">1.67</td><td align="center" rowspan="1" colspan="1">1.09</td><td align="center" rowspan="1" colspan="1">1.23</td><td align="center" rowspan="1" colspan="1">&#x0003c;0.0001</td></tr><tr><td align="left" rowspan="1" colspan="1">02.99 Other and unspecified neoplasms</td><td align="center" rowspan="1" colspan="1">1.11</td><td align="center" rowspan="1" colspan="1">0.76</td><td align="center" rowspan="1" colspan="1">0.92</td><td align="center" rowspan="1" colspan="1">&#x02212;0.34</td><td align="center" rowspan="1" colspan="1">0.05</td></tr><tr><td align="left" rowspan="1" colspan="1">03.01 Severe anaemia</td><td align="center" rowspan="1" colspan="1">0.16</td><td align="center" rowspan="1" colspan="1">0.40</td><td align="center" rowspan="1" colspan="1">0.29</td><td align="center" rowspan="1" colspan="1">0.24</td><td align="center" rowspan="1" colspan="1">0.05</td></tr><tr><td align="left" rowspan="1" colspan="1">03.02 Severe malnutrition</td><td align="center" rowspan="1" colspan="1">0.43</td><td align="center" rowspan="1" colspan="1">0.45</td><td align="center" rowspan="1" colspan="1">0.44</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.86</td></tr><tr><td align="left" rowspan="1" colspan="1">03.03 Diabetes mellitus</td><td align="center" rowspan="1" colspan="1">1.21</td><td align="center" rowspan="1" colspan="1">2.35</td><td align="center" rowspan="1" colspan="1">1.81</td><td align="center" rowspan="1" colspan="1">1.13</td><td align="center" rowspan="1" colspan="1">&#x0003c;0.0001</td></tr><tr><td align="left" rowspan="1" colspan="1">04.01 Acute cardiac disease</td><td align="center" rowspan="1" colspan="1">0.64</td><td align="center" rowspan="1" colspan="1">0.38</td><td align="center" rowspan="1" colspan="1">0.50</td><td align="center" rowspan="1" colspan="1">&#x02212;0.26</td><td align="center" rowspan="1" colspan="1">0.05</td></tr><tr><td align="left" rowspan="1" colspan="1">04.02 Stroke</td><td align="center" rowspan="1" colspan="1">2.74</td><td align="center" rowspan="1" colspan="1">4.51</td><td align="center" rowspan="1" colspan="1">3.68</td><td align="center" rowspan="1" colspan="1">1.77</td><td align="center" rowspan="1" colspan="1">&#x0003c;0.0001</td></tr><tr><td align="left" rowspan="1" colspan="1">04.99 Other and unspecified cardiac dis</td><td align="center" rowspan="1" colspan="1">3.23</td><td align="center" rowspan="1" colspan="1">5.94</td><td align="center" rowspan="1" colspan="1">4.66</td><td align="center" rowspan="1" colspan="1">2.71</td><td align="center" rowspan="1" colspan="1">&#x0003c;0.0001</td></tr><tr><td align="left" rowspan="1" colspan="1">05.01 Chronic obstructive pulmonary dis</td><td align="center" rowspan="1" colspan="1">1.18</td><td align="center" rowspan="1" colspan="1">1.85</td><td align="center" rowspan="1" colspan="1">1.54</td><td align="center" rowspan="1" colspan="1">0.67</td><td align="center" rowspan="1" colspan="1">0.004</td></tr><tr><td align="left" rowspan="1" colspan="1">05.02 Asthma</td><td align="center" rowspan="1" colspan="1">0.64</td><td align="center" rowspan="1" colspan="1">0.98</td><td align="center" rowspan="1" colspan="1">0.82</td><td align="center" rowspan="1" colspan="1">0.33</td><td align="center" rowspan="1" colspan="1">0.05</td></tr><tr><td align="left" rowspan="1" colspan="1">06.01 Acute abdomen</td><td align="center" rowspan="1" colspan="1">1.11</td><td align="center" rowspan="1" colspan="1">1.17</td><td align="center" rowspan="1" colspan="1">1.14</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.74</td></tr><tr><td align="left" rowspan="1" colspan="1">06.02 Liver cirrhosis</td><td align="center" rowspan="1" colspan="1">0.26</td><td align="center" rowspan="1" colspan="1">0.26</td><td align="center" rowspan="1" colspan="1">0.26</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.97</td></tr><tr><td align="left" rowspan="1" colspan="1">07.01 Renal failure</td><td align="center" rowspan="1" colspan="1">0.70</td><td align="center" rowspan="1" colspan="1">0.43</td><td align="center" rowspan="1" colspan="1">0.56</td><td align="center" rowspan="1" colspan="1">&#x02212;0.27</td><td align="center" rowspan="1" colspan="1">0.05</td></tr><tr><td align="left" rowspan="1" colspan="1">08.01 Epilepsy</td><td align="center" rowspan="1" colspan="1">0.68</td><td align="center" rowspan="1" colspan="1">0.28</td><td align="center" rowspan="1" colspan="1">0.47</td><td align="center" rowspan="1" colspan="1">&#x02212;0.40</td><td align="center" rowspan="1" colspan="1">0.002</td></tr><tr><td align="left" rowspan="1" colspan="1">09.01 Ectopic pregnancy</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">09.02 Abortion-related death</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">09.03 Pregnancy-induced hypertension</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.22</td><td align="center" rowspan="1" colspan="1">0.12</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">09.04 Obstetric haemorrhage</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.13</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">09.06 Pregnancy-related sepsis</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">09.07 Anaemia of pregnancy</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">09.99 Other and unspecified maternal CoD</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">10.01 Prematurity</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">0.12</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.92</td></tr><tr><td align="left" rowspan="1" colspan="1">10.02 Birth asphyxia</td><td align="center" rowspan="1" colspan="1">0.42</td><td align="center" rowspan="1" colspan="1">0.25</td><td align="center" rowspan="1" colspan="1">0.33</td><td align="center" rowspan="1" colspan="1">&#x02212;0.17</td><td align="center" rowspan="1" colspan="1">0.11</td></tr><tr><td align="left" rowspan="1" colspan="1">10.03 Neonatal pneumonia</td><td align="center" rowspan="1" colspan="1">0.70</td><td align="center" rowspan="1" colspan="1">0.41</td><td align="center" rowspan="1" colspan="1">0.54</td><td align="center" rowspan="1" colspan="1">&#x02212;0.29</td><td align="center" rowspan="1" colspan="1">0.03</td></tr><tr><td align="left" rowspan="1" colspan="1">10.04 Neonatal sepsis</td><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.74</td></tr><tr><td align="left" rowspan="1" colspan="1">10.06 Congenital malformation</td><td align="center" rowspan="1" colspan="1">0.19</td><td align="center" rowspan="1" colspan="1">0.17</td><td align="center" rowspan="1" colspan="1">0.18</td><td align="center" rowspan="1" colspan="1">&#x02212;0.02</td><td align="center" rowspan="1" colspan="1">0.82</td></tr><tr><td align="left" rowspan="1" colspan="1">10.99 Other and unspecified neonatal CoD</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">&#x02212;0.03</td><td align="center" rowspan="1" colspan="1">0.32</td></tr><tr><td align="left" rowspan="1" colspan="1">12.01 Road traffic accident</td><td align="center" rowspan="1" colspan="1">3.00</td><td align="center" rowspan="1" colspan="1">1.05</td><td align="center" rowspan="1" colspan="1">1.97</td><td align="center" rowspan="1" colspan="1">&#x02212;1.95</td><td align="center" rowspan="1" colspan="1">&#x0003c;0.0001</td></tr><tr><td align="left" rowspan="1" colspan="1">12.03 Accid fall</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.86</td></tr><tr><td align="left" rowspan="1" colspan="1">12.04 Accid drowning and submersion</td><td align="center" rowspan="1" colspan="1">0.38</td><td align="center" rowspan="1" colspan="1">0.18</td><td align="center" rowspan="1" colspan="1">0.28</td><td align="center" rowspan="1" colspan="1">&#x02212;0.20</td><td align="center" rowspan="1" colspan="1">0.04</td></tr><tr><td align="left" rowspan="1" colspan="1">12.05 Accid expos to smoke fire &#x00026; flame</td><td align="center" rowspan="1" colspan="1">0.21</td><td align="center" rowspan="1" colspan="1">0.34</td><td align="center" rowspan="1" colspan="1">0.28</td><td align="center" rowspan="1" colspan="1">0.13</td><td align="center" rowspan="1" colspan="1">0.20</td></tr><tr><td align="left" rowspan="1" colspan="1">12.06 Contact with venomous plant/animal</td><td align="center" rowspan="1" colspan="1">0.16</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">&#x02212;0.14</td><td align="center" rowspan="1" colspan="1">0.008</td></tr><tr><td align="left" rowspan="1" colspan="1">12.07 Accid poisoning &#x00026; noxious subs</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.77</td></tr><tr><td align="left" rowspan="1" colspan="1">12.08 Intentional self-harm</td><td align="center" rowspan="1" colspan="1">1.26</td><td align="center" rowspan="1" colspan="1">0.27</td><td align="center" rowspan="1" colspan="1">0.74</td><td align="center" rowspan="1" colspan="1">&#x02212;0.98</td><td align="center" rowspan="1" colspan="1">&#x0003c;0.0001</td></tr><tr><td align="left" rowspan="1" colspan="1">12.09 Assault</td><td align="center" rowspan="1" colspan="1">6.08</td><td align="center" rowspan="1" colspan="1">1.66</td><td align="center" rowspan="1" colspan="1">3.74</td><td align="center" rowspan="1" colspan="1">&#x02212;4.42</td><td align="center" rowspan="1" colspan="1">&#x0003c;0.0001</td></tr><tr><td align="left" rowspan="1" colspan="1">12.99 Other and unspecified external CoD</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.14</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.21</td></tr><tr><td align="left" rowspan="1" colspan="1">98 Other and unspecified NCD</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.31</td></tr><tr><td align="left" rowspan="1" colspan="1">99 Indeterminate</td><td align="center" rowspan="1" colspan="1">5.60</td><td align="center" rowspan="1" colspan="1">6.19</td><td align="center" rowspan="1" colspan="1">5.91</td><td align="center" rowspan="1" colspan="1">0.59</td><td align="center" rowspan="1" colspan="1">0.18</td></tr><tr><td align="left" rowspan="1" colspan="1">XX &#x02013; VA not completed</td><td align="center" rowspan="1" colspan="1">5.10</td><td align="center" rowspan="1" colspan="1">4.32</td><td align="center" rowspan="1" colspan="1">4.69</td><td align="center" rowspan="1" colspan="1">&#x02212;0.77</td><td align="center" rowspan="1" colspan="1">0.05</td></tr></tbody></table></table-wrap><p> <xref ref-type="table" rid="T0003">Table 3</xref> shows that cause-specific morbidity fractions varied substantially by age. The fraction of deaths for which no cause could be determined (codes 99 and XX) was highest in children below the age of 5 years (14.8%) and in persons 65 years or older (14.3%) and lowest in adults aged 15&#x02013;49 years (7.6%). In general, causes of death due to communicable diseases (WHO VA code chapter 01) predominated in all age groups except in persons aged 65 years and older. Acute respiratory infections including pneumonia appeared as one of the principal causes of death for children below 5 years of age, particularly neonates (56.9%) and infants (24.2%). HIV/AIDS was a major COD responsible for more than 10% of deaths in all age groups except neonates (0%) and persons older than 65 years (3.1%). Pulmonary tuberculosis was the predominant COD in children aged 5&#x02013;14 years (26.6%) and in adults aged 15&#x02013;64 years (48.8%). As expected, the mortality fractions of non-communicable causes of death of neoplasms, diabetes, stroke, and unspecified cardiac disease grew in importance beyond 50 years of age.</p><table-wrap id="T0003" position="float"><label>Table 3</label><caption><p>Cause-specific mortality fractions expressed as percentages by age group</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1">WHO VA cause of death code</th><th align="center" rowspan="1" colspan="1">Neonate (<italic>N</italic>=171)</th><th align="center" rowspan="1" colspan="1">Infant (<italic>N</italic>=800)</th><th align="center" rowspan="1" colspan="1">1&#x02013;5 year (<italic>N =</italic>617)</th><th align="center" rowspan="1" colspan="1">5&#x02013;14 year (<italic>N=</italic>341)</th><th align="center" rowspan="1" colspan="1">15&#x02013;49 year (<italic>N</italic>=5,517)</th><th align="center" rowspan="1" colspan="1">50&#x02013;64 year (<italic>N</italic>=1,670)</th><th align="center" rowspan="1" colspan="1">65+ year (<italic>N</italic>=2,381)</th><th align="center" rowspan="1" colspan="1">Total (<italic>N</italic>=11,497)</th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">01.01 Sepsis (non-obstetric)</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.32</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.30</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.04</td></tr><tr><td align="left" rowspan="1" colspan="1">01.02 Acute resp infect incl. pneumonia</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">56.88</td><td align="center" rowspan="1" colspan="1">24.17</td><td align="center" rowspan="1" colspan="1">4.85</td><td align="center" rowspan="1" colspan="1">1.37</td><td align="center" rowspan="1" colspan="1">2.05</td><td align="center" rowspan="1" colspan="1">3.15</td><td align="center" rowspan="1" colspan="1">7.00</td></tr><tr><td align="left" rowspan="1" colspan="1">01.03 HIV/AIDS related death</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">19.66</td><td align="center" rowspan="1" colspan="1">37.62</td><td align="center" rowspan="1" colspan="1">16.24</td><td align="center" rowspan="1" colspan="1">22.25</td><td align="center" rowspan="1" colspan="1">12.62</td><td align="center" rowspan="1" colspan="1">3.13</td><td align="center" rowspan="1" colspan="1">17.03</td></tr><tr><td align="left" rowspan="1" colspan="1">01.04 Diarrhoeal diseases</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">3.68</td><td align="center" rowspan="1" colspan="1">2.16</td><td align="center" rowspan="1" colspan="1">0.51</td><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">0.72</td><td align="center" rowspan="1" colspan="1">0.60</td></tr><tr><td align="left" rowspan="1" colspan="1">01.05 Malaria</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1.08</td><td align="center" rowspan="1" colspan="1">2.24</td><td align="center" rowspan="1" colspan="1">2.16</td><td align="center" rowspan="1" colspan="1">0.17</td><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">0.30</td><td align="center" rowspan="1" colspan="1">0.42</td></tr><tr><td align="left" rowspan="1" colspan="1">01.06 Measles</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.61</td><td align="center" rowspan="1" colspan="1">0.25</td><td align="center" rowspan="1" colspan="1">0.28</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.06</td></tr><tr><td align="left" rowspan="1" colspan="1">01.07 Meningitis and encephalitis</td><td align="center" rowspan="1" colspan="1">0.81</td><td align="center" rowspan="1" colspan="1">0.37</td><td align="center" rowspan="1" colspan="1">0.73</td><td align="center" rowspan="1" colspan="1">4.14</td><td align="center" rowspan="1" colspan="1">1.22</td><td align="center" rowspan="1" colspan="1">0.56</td><td align="center" rowspan="1" colspan="1">0.16</td><td align="center" rowspan="1" colspan="1">0.90</td></tr><tr><td align="left" rowspan="1" colspan="1">01.08 &#x00026; 10.05 Tetanus</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.23</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.01</td></tr><tr><td align="left" rowspan="1" colspan="1">01.09 Pulmonary tuberculosis</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.71</td><td align="center" rowspan="1" colspan="1">5.82</td><td align="center" rowspan="1" colspan="1">26.60</td><td align="center" rowspan="1" colspan="1">48.77</td><td align="center" rowspan="1" colspan="1">33.35</td><td align="center" rowspan="1" colspan="1">13.58</td><td align="center" rowspan="1" colspan="1">32.21</td></tr><tr><td align="left" rowspan="1" colspan="1">01.10 Pertussis</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.18</td><td align="center" rowspan="1" colspan="1">0.28</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.03</td></tr><tr><td align="left" rowspan="1" colspan="1">01.99 Other and unspecified infect dis</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1.03</td><td align="center" rowspan="1" colspan="1">0.14</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.26</td><td align="center" rowspan="1" colspan="1">0.16</td></tr><tr><td align="left" rowspan="1" colspan="1">02.01 Oral neoplasms</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.14</td><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">0.10</td></tr><tr><td align="left" rowspan="1" colspan="1">02.02 Digestive neoplasms</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.97</td><td align="center" rowspan="1" colspan="1">3.69</td><td align="center" rowspan="1" colspan="1">2.35</td><td align="center" rowspan="1" colspan="1">1.49</td></tr><tr><td align="left" rowspan="1" colspan="1">02.03 Respiratory neoplasms</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1.08</td><td align="center" rowspan="1" colspan="1">3.52</td><td align="center" rowspan="1" colspan="1">4.50</td><td align="center" rowspan="1" colspan="1">1.96</td></tr><tr><td align="left" rowspan="1" colspan="1">02.04 Breast neoplasms</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.27</td><td align="center" rowspan="1" colspan="1">0.16</td><td align="center" rowspan="1" colspan="1">0.38</td><td align="center" rowspan="1" colspan="1">0.23</td></tr><tr><td align="left" rowspan="1" colspan="1">02.05 &#x00026; 02.06 Reproductive neoplasms MF</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.73</td><td align="center" rowspan="1" colspan="1">1.60</td><td align="center" rowspan="1" colspan="1">2.46</td><td align="center" rowspan="1" colspan="1">1.09</td></tr><tr><td align="left" rowspan="1" colspan="1">02.99 Other and unspecified neoplasms</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.16</td><td align="center" rowspan="1" colspan="1">0.27</td><td align="center" rowspan="1" colspan="1">0.44</td><td align="center" rowspan="1" colspan="1">1.38</td><td align="center" rowspan="1" colspan="1">2.39</td><td align="center" rowspan="1" colspan="1">0.92</td></tr><tr><td align="left" rowspan="1" colspan="1">03.01 Severe anaemia</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.48</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.19</td><td align="center" rowspan="1" colspan="1">1.09</td><td align="center" rowspan="1" colspan="1">0.29</td></tr><tr><td align="left" rowspan="1" colspan="1">03.02 Severe malnutrition</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.73</td><td align="center" rowspan="1" colspan="1">3.80</td><td align="center" rowspan="1" colspan="1">0.80</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.54</td><td align="center" rowspan="1" colspan="1">0.44</td></tr><tr><td align="left" rowspan="1" colspan="1">03.03 Diabetes mellitus</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.14</td><td align="center" rowspan="1" colspan="1">0.33</td><td align="center" rowspan="1" colspan="1">2.99</td><td align="center" rowspan="1" colspan="1">5.86</td><td align="center" rowspan="1" colspan="1">1.81</td></tr><tr><td align="left" rowspan="1" colspan="1">04.01 Acute cardiac disease</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.25</td><td align="center" rowspan="1" colspan="1">1.29</td><td align="center" rowspan="1" colspan="1">0.93</td><td align="center" rowspan="1" colspan="1">0.50</td></tr><tr><td align="left" rowspan="1" colspan="1">04.02 Stroke</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.43</td><td align="center" rowspan="1" colspan="1">0.51</td><td align="center" rowspan="1" colspan="1">6.28</td><td align="center" rowspan="1" colspan="1">12.13</td><td align="center" rowspan="1" colspan="1">3.68</td></tr><tr><td align="left" rowspan="1" colspan="1">04.99 Other and unsp. cardiac disease</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1.05</td><td align="center" rowspan="1" colspan="1">1.19</td><td align="center" rowspan="1" colspan="1">6.85</td><td align="center" rowspan="1" colspan="1">14.81</td><td align="center" rowspan="1" colspan="1">4.66</td></tr><tr><td align="left" rowspan="1" colspan="1">05.01 Chronic obstr. pulmonary disease</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.13</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">1.57</td><td align="center" rowspan="1" colspan="1">6.18</td><td align="center" rowspan="1" colspan="1">1.54</td></tr><tr><td align="left" rowspan="1" colspan="1">05.02 Asthma</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.32</td><td align="center" rowspan="1" colspan="1">1.61</td><td align="center" rowspan="1" colspan="1">0.21</td><td align="center" rowspan="1" colspan="1">1.37</td><td align="center" rowspan="1" colspan="1">2.21</td><td align="center" rowspan="1" colspan="1">0.82</td></tr><tr><td align="left" rowspan="1" colspan="1">06.01 Acute abdomen</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.41</td><td align="center" rowspan="1" colspan="1">0.16</td><td align="center" rowspan="1" colspan="1">2.66</td><td align="center" rowspan="1" colspan="1">0.71</td><td align="center" rowspan="1" colspan="1">1.78</td><td align="center" rowspan="1" colspan="1">2.08</td><td align="center" rowspan="1" colspan="1">1.14</td></tr><tr><td align="left" rowspan="1" colspan="1">06.02 Liver cirrhosis</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.32</td><td align="center" rowspan="1" colspan="1">0.18</td><td align="center" rowspan="1" colspan="1">0.47</td><td align="center" rowspan="1" colspan="1">0.46</td><td align="center" rowspan="1" colspan="1">0.26</td></tr><tr><td align="left" rowspan="1" colspan="1">07.01 Renal failure</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.27</td><td align="center" rowspan="1" colspan="1">0.68</td><td align="center" rowspan="1" colspan="1">1.60</td><td align="center" rowspan="1" colspan="1">0.56</td></tr><tr><td align="left" rowspan="1" colspan="1">08.01 Epilepsy</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.34</td><td align="center" rowspan="1" colspan="1">0.16</td><td align="center" rowspan="1" colspan="1">3.39</td><td align="center" rowspan="1" colspan="1">0.34</td><td align="center" rowspan="1" colspan="1">0.58</td><td align="center" rowspan="1" colspan="1">0.41</td><td align="center" rowspan="1" colspan="1">0.47</td></tr><tr><td align="left" rowspan="1" colspan="1">09.01 Ectopic pregnancy</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.03</td></tr><tr><td align="left" rowspan="1" colspan="1">09.02 Abortion-related death</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.02</td></tr><tr><td align="left" rowspan="1" colspan="1">09.03 Pregnancy-induced hypertension</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.24</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.12</td></tr><tr><td align="left" rowspan="1" colspan="1">09.04 Obstetric haemorrhage</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.15</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.07</td></tr><tr><td align="left" rowspan="1" colspan="1">09.06 Pregnancy-related sepsis</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.05</td></tr><tr><td align="left" rowspan="1" colspan="1">09.07 Anaemia of pregnancy</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.03</td></tr><tr><td align="left" rowspan="1" colspan="1">09.99 Other and unspecified maternal CoD</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.03</td></tr><tr><td align="left" rowspan="1" colspan="1">10.01 Prematurity</td><td align="center" rowspan="1" colspan="1">7.73</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.11</td></tr><tr><td align="left" rowspan="1" colspan="1">10.02 Birth asphyxia</td><td align="center" rowspan="1" colspan="1">22.35</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.33</td></tr><tr><td align="left" rowspan="1" colspan="1">10.03 Neonatal pneumonia</td><td align="center" rowspan="1" colspan="1">36.60</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.54</td></tr><tr><td align="left" rowspan="1" colspan="1">10.04 Neonatal sepsis</td><td align="center" rowspan="1" colspan="1">6.64</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.10</td></tr><tr><td align="left" rowspan="1" colspan="1">10.06 Congenital malformation</td><td align="center" rowspan="1" colspan="1">5.12</td><td align="center" rowspan="1" colspan="1">0.71</td><td align="center" rowspan="1" colspan="1">1.02</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.18</td></tr><tr><td align="left" rowspan="1" colspan="1">10.99 Other and unspecified neonatal CoD</td><td align="center" rowspan="1" colspan="1">2.29</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.03</td></tr><tr><td align="left" rowspan="1" colspan="1">12.01 Road traffic accident</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.12</td><td align="center" rowspan="1" colspan="1">1.44</td><td align="center" rowspan="1" colspan="1">9.92</td><td align="center" rowspan="1" colspan="1">2.57</td><td align="center" rowspan="1" colspan="1">1.32</td><td align="center" rowspan="1" colspan="1">0.77</td><td align="center" rowspan="1" colspan="1">1.97</td></tr><tr><td align="left" rowspan="1" colspan="1">12.03 Accid fall</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.16</td><td align="center" rowspan="1" colspan="1">0.19</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.12</td><td align="center" rowspan="1" colspan="1">0.04</td></tr><tr><td align="left" rowspan="1" colspan="1">12.04 Accid drowning and submersion</td><td align="center" rowspan="1" colspan="1">0.58</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.97</td><td align="center" rowspan="1" colspan="1">4.08</td><td align="center" rowspan="1" colspan="1">0.18</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.28</td></tr><tr><td align="left" rowspan="1" colspan="1">12.05 Accid expos to smoke fire &#x00026; flame</td><td align="center" rowspan="1" colspan="1">0.58</td><td align="center" rowspan="1" colspan="1">0.37</td><td align="center" rowspan="1" colspan="1">0.80</td><td align="center" rowspan="1" colspan="1">0.29</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">0.17</td><td align="center" rowspan="1" colspan="1">0.62</td><td align="center" rowspan="1" colspan="1">0.28</td></tr><tr><td align="left" rowspan="1" colspan="1">12.06 Contact with venomous plant/animal</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">0.28</td><td align="center" rowspan="1" colspan="1">0.29</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.19</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.08</td></tr><tr><td align="left" rowspan="1" colspan="1">12.07 Accid poisoning &#x00026; noxious subs</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.16</td><td align="center" rowspan="1" colspan="1">0.43</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.05</td></tr><tr><td align="left" rowspan="1" colspan="1">12.08 Intentional self-harm</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1.91</td><td align="center" rowspan="1" colspan="1">1.24</td><td align="center" rowspan="1" colspan="1">0.36</td><td align="center" rowspan="1" colspan="1">0.15</td><td align="center" rowspan="1" colspan="1">0.74</td></tr><tr><td align="left" rowspan="1" colspan="1">12.09 Assault</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1.61</td><td align="center" rowspan="1" colspan="1">2.64</td><td align="center" rowspan="1" colspan="1">5.70</td><td align="center" rowspan="1" colspan="1">3.29</td><td align="center" rowspan="1" colspan="1">1.76</td><td align="center" rowspan="1" colspan="1">3.74</td></tr><tr><td align="left" rowspan="1" colspan="1">12.99 Other and unspecified external CoD</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.15</td><td align="center" rowspan="1" colspan="1">0.72</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.13</td><td align="center" rowspan="1" colspan="1">0.17</td><td align="center" rowspan="1" colspan="1">0.10</td></tr><tr><td align="left" rowspan="1" colspan="1">98 Other and unspecified NCD</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.21</td><td align="center" rowspan="1" colspan="1">0.06</td></tr><tr><td align="left" rowspan="1" colspan="1">99 Indeterminate</td><td align="center" rowspan="1" colspan="1">17.31</td><td align="center" rowspan="1" colspan="1">5.67</td><td align="center" rowspan="1" colspan="1">7.51</td><td align="center" rowspan="1" colspan="1">6.35</td><td align="center" rowspan="1" colspan="1">3.41</td><td align="center" rowspan="1" colspan="1">6.60</td><td align="center" rowspan="1" colspan="1">10.01</td><td align="center" rowspan="1" colspan="1">5.91</td></tr><tr><td align="left" rowspan="1" colspan="1">XX &#x02013; VA not completed</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">8.00</td><td align="center" rowspan="1" colspan="1">7.94</td><td align="center" rowspan="1" colspan="1">5.57</td><td align="center" rowspan="1" colspan="1">4.15</td><td align="center" rowspan="1" colspan="1">4.49</td><td align="center" rowspan="1" colspan="1">4.33</td><td align="center" rowspan="1" colspan="1">4.69</td></tr></tbody></table><table-wrap-foot><fn><p>Empty cells represent no deaths except for maternally-related death codes starting with 09 that are only relevant for the 15&#x02013;49 age group and neonatal death codes 10.01&#x02013;10.04 and 10.99 that are only relevant for neonates.</p></fn></table-wrap-foot></table-wrap><p>The fraction of external causes of death was highest in children aged 5&#x02013;14 years, principally from road traffic accidents (9.9%). Assault was the most common (5.7%) external COD of young adults aged 14&#x02013;49 years, principally affecting men.</p><p>An age-stratified comparison of mortality rates between men and women for the most important causes of death (<xref ref-type="table" rid="T0004">Table 4</xref>) shows that for most causes of death, mortality rates differ in almost all adult age groups. This is particularly the case for HIV/AIDS (except the age group of persons 65 years or older), pulmonary tuberculosis, digestive and respiratory neoplasms in persons 50 years or older, road traffic accidents, intentional self-harm, and assault. Age-dependent mortality rates were similar in all age groups between men and women for diabetes, stroke, and unspecified cardiac diseases.</p><table-wrap id="T0004" position="float"><label>Table 4</label><caption><p>Comparison of mortality rates (per 100,000) between men and women for the most important causes of death in adults, stratified by age group</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1"/><th align="center" colspan="3" rowspan="1">15&#x02013;49 year</th><th align="center" colspan="3" rowspan="1">50&#x02013;64 year</th><th align="center" colspan="3" rowspan="1">65+ year</th></tr><tr><th align="left" rowspan="1" colspan="1"/><th align="center" colspan="3" rowspan="1"> <hr/> </th><th align="center" colspan="3" rowspan="1"> <hr/> </th><th align="center" colspan="3" rowspan="1"> <hr/> </th></tr><tr><th align="left" rowspan="1" colspan="1"/><th align="center" rowspan="1" colspan="1">Male</th><th align="center" rowspan="1" colspan="1">Female</th><th align="center" rowspan="1" colspan="1"> <italic>p</italic> </th><th align="center" rowspan="1" colspan="1">Male</th><th align="center" rowspan="1" colspan="1">Female</th><th align="center" rowspan="1" colspan="1"> <italic>p</italic> </th><th align="center" rowspan="1" colspan="1">Male</th><th align="center" rowspan="1" colspan="1">Female</th><th align="center" rowspan="1" colspan="1"> <italic>p</italic> </th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">01.03 HIV/AIDS related death</td><td align="center" rowspan="1" colspan="1">226</td><td align="center" rowspan="1" colspan="1">407</td><td align="center" rowspan="1" colspan="1">&#x0003c;0.0001</td><td align="center" rowspan="1" colspan="1">522</td><td align="center" rowspan="1" colspan="1">310</td><td align="center" rowspan="1" colspan="1">0.0002</td><td align="center" rowspan="1" colspan="1">229</td><td align="center" rowspan="1" colspan="1">175</td><td align="center" rowspan="1" colspan="1">0.31</td></tr><tr><td align="left" rowspan="1" colspan="1">01.09 Pulmonary tuberculosis</td><td align="center" rowspan="1" colspan="1">783</td><td align="center" rowspan="1" colspan="1">669</td><td align="center" rowspan="1" colspan="1">&#x0003c;0.000</td><td align="center" rowspan="1" colspan="1">1,741</td><td align="center" rowspan="1" colspan="1">625</td><td align="center" rowspan="1" colspan="1">&#x0003c;0.0001</td><td align="center" rowspan="1" colspan="1">1,730</td><td align="center" rowspan="1" colspan="1">467</td><td align="center" rowspan="1" colspan="1">&#x0003c;0.0001</td></tr><tr><td align="left" rowspan="1" colspan="1">02.02 Digestive neoplasms</td><td align="center" rowspan="1" colspan="1">20</td><td align="center" rowspan="1" colspan="1">10</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">191</td><td align="center" rowspan="1" colspan="1">70</td><td align="center" rowspan="1" colspan="1">0.0001</td><td align="center" rowspan="1" colspan="1">267</td><td align="center" rowspan="1" colspan="1">94</td><td align="center" rowspan="1" colspan="1">0.0001</td></tr><tr><td align="left" rowspan="1" colspan="1">02.03 Respiratory neoplasms</td><td align="center" rowspan="1" colspan="1">19</td><td align="center" rowspan="1" colspan="1">13</td><td align="center" rowspan="1" colspan="1">0.17</td><td align="center" rowspan="1" colspan="1">226</td><td align="center" rowspan="1" colspan="1">43</td><td align="center" rowspan="1" colspan="1">&#x0003c;0.0001</td><td align="center" rowspan="1" colspan="1">559</td><td align="center" rowspan="1" colspan="1">161</td><td align="center" rowspan="1" colspan="1">&#x0003c;0.0001</td></tr><tr><td align="left" rowspan="1" colspan="1">02.05 &#x00026; 02.06 Reproductive neoplasms MF</td><td align="center" rowspan="1" colspan="1">2</td><td align="center" rowspan="1" colspan="1">17</td><td align="center" rowspan="1" colspan="1">&#x0003c;0.0001</td><td align="center" rowspan="1" colspan="1">12</td><td align="center" rowspan="1" colspan="1">68</td><td align="center" rowspan="1" colspan="1">0.002</td><td align="center" rowspan="1" colspan="1">162</td><td align="center" rowspan="1" colspan="1">145</td><td align="center" rowspan="1" colspan="1">0.71</td></tr><tr><td align="left" rowspan="1" colspan="1">03.03 Diabetes mellitus</td><td align="center" rowspan="1" colspan="1">7</td><td align="center" rowspan="1" colspan="1">3</td><td align="center" rowspan="1" colspan="1">0.15</td><td align="center" rowspan="1" colspan="1">74</td><td align="center" rowspan="1" colspan="1">101</td><td align="center" rowspan="1" colspan="1">0.30</td><td align="center" rowspan="1" colspan="1">360</td><td align="center" rowspan="1" colspan="1">355</td><td align="center" rowspan="1" colspan="1">0.95</td></tr><tr><td align="left" rowspan="1" colspan="1">04.02 Stroke</td><td align="center" rowspan="1" colspan="1">9</td><td align="center" rowspan="1" colspan="1">6</td><td align="center" rowspan="1" colspan="1">0.30</td><td align="center" rowspan="1" colspan="1">256</td><td align="center" rowspan="1" colspan="1">156</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">748</td><td align="center" rowspan="1" colspan="1">733</td><td align="center" rowspan="1" colspan="1">0.90</td></tr><tr><td align="left" rowspan="1" colspan="1">04.99 Other and unspecified cardiac disease</td><td align="center" rowspan="1" colspan="1">12</td><td align="center" rowspan="1" colspan="1">22</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">225</td><td align="center" rowspan="1" colspan="1">200</td><td align="center" rowspan="1" colspan="1">0.54</td><td align="center" rowspan="1" colspan="1">1,004</td><td align="center" rowspan="1" colspan="1">859</td><td align="center" rowspan="1" colspan="1">0.17</td></tr><tr><td align="left" rowspan="1" colspan="1">12.01 Road traffic accident</td><td align="center" rowspan="1" colspan="1">70</td><td align="center" rowspan="1" colspan="1">13</td><td align="center" rowspan="1" colspan="1">&#x0003c;0.0001</td><td align="center" rowspan="1" colspan="1">90</td><td align="center" rowspan="1" colspan="1">13</td><td align="center" rowspan="1" colspan="1">0.0001</td><td align="center" rowspan="1" colspan="1">94</td><td align="center" rowspan="1" colspan="1">28</td><td align="center" rowspan="1" colspan="1">0.02</td></tr><tr><td align="left" rowspan="1" colspan="1">12.08 Intentional self-harm</td><td align="center" rowspan="1" colspan="1">34</td><td align="center" rowspan="1" colspan="1">6</td><td align="center" rowspan="1" colspan="1">&#x0003c;0.0001</td><td align="center" rowspan="1" colspan="1">31</td><td align="center" rowspan="1" colspan="1">0</td><td align="center" rowspan="1" colspan="1">0.002</td><td align="center" rowspan="1" colspan="1">18</td><td align="center" rowspan="1" colspan="1">6</td><td align="center" rowspan="1" colspan="1">0.39</td></tr><tr><td align="left" rowspan="1" colspan="1">12.09 Assault</td><td align="center" rowspan="1" colspan="1">162</td><td align="center" rowspan="1" colspan="1">24</td><td align="center" rowspan="1" colspan="1">&#x0003c;0.0001</td><td align="center" rowspan="1" colspan="1">182</td><td align="center" rowspan="1" colspan="1">56</td><td align="center" rowspan="1" colspan="1">&#x0003c;0.0001</td><td align="center" rowspan="1" colspan="1">180</td><td align="center" rowspan="1" colspan="1">78</td><td align="center" rowspan="1" colspan="1">0.009</td></tr></tbody></table></table-wrap></sec><sec sec-type="discussion" id="S0003"><title>Discussion</title><p>This is the first report on causes of death in rural KwaZulu-Natal based on the standardised 2012 WHO VA instrument and InterVA-4 tool. The results presented here are in line with previously published work using causes of death based on physician-coded diagnoses or InterVA-3. As reported previously, mortality in our setting during the 2000s was largely dominated by HIV/AIDS and tuberculosis (<xref rid="CIT0003" ref-type="bibr">3</xref>, <xref rid="CIT0005" ref-type="bibr">5</xref>, <xref rid="CIT0019" ref-type="bibr">19</xref>, <xref rid="CIT0020" ref-type="bibr">20</xref>), although HIV-related mortality rates have declined substantially in recent years (<xref rid="CIT0003" ref-type="bibr">3</xref>, <xref rid="CIT0005" ref-type="bibr">5</xref>), resulting in a dramatic increase in life expectancy (<xref rid="CIT0006" ref-type="bibr">6</xref>). Our results are also in broad agreement with the Agincourt DSS in a different rural area in South Africa (<xref rid="CIT0010" ref-type="bibr">10</xref>), although the fraction of undetermined causes of deaths are substantially lower in our setting (<xref rid="CIT0008" ref-type="bibr">8</xref>). This suggests that the InterVA-4 tool for determining causes of death in combination with the 2012 VA instrument is appropriate for wider use in other Southern African settings.</p><p>One finding which warrants further investigation is the differing sex and age patterns of causes of death due to HIV/AIDS and tuberculosis. One the one hand, the observed sex-specific difference could be due to differing prevalence patterns of HIV/AIDS and tuberculosis between sexes. In 2010 in our setting, women were significantly more likely to be infected with HIV than men (27.6% vs. 16.2%) (<xref rid="CIT0012" ref-type="bibr">12</xref>), which might explain the higher fraction of deaths due to HIV/AIDS observed in women. As the age group most affected by HIV/AIDS were women between the ages of 25 and 49 years, this could be a plausible explanation why we observed more deaths in women than men for the age group of 15&#x02013;49 years. Similarly, it is known that in many settings in the world, notification rates of tuberculosis are higher in men than women, due to some extent by higher prevalence of cigarette smoking (<xref rid="CIT0021" ref-type="bibr">21</xref>). This would explain the higher proportion of deaths due to tuberculosis in men compared to women. The high proportion of deaths attributed to tuberculosis in children aged 5&#x02013;14 years is likely to be related to the high prevalence of HIV/AIDS in this population. Given the considerable overlap in mortality from HIV infection and tuberculosis (<xref rid="CIT0022" ref-type="bibr">22</xref>), much previous work has therefore often combined the two categories of HIV and tuberculosis (<xref rid="CIT0003" ref-type="bibr">3</xref>, <xref rid="CIT0005" ref-type="bibr">5</xref>). For future work, it would be interesting to investigate why InterVA-4 produces a higher fraction of deaths due to tuberculosis in comparison to HIV/AIDS, and whether the known uncertainty in COD for this important group is also reflected in the likelihoods.</p><p>An earlier analysis of causes of child mortality also highlighted the importance of lower respiratory tract infections as a COD in children under 1 year of age, which corresponds to our finding of a high CSMF of acute respiratory infections including pneumonia (<xref rid="CIT0020" ref-type="bibr">20</xref>). A recent review indicated that 43% of childhood deaths occurred due to pneumonia in sub-Saharan Africa in 2011 (<xref rid="CIT0023" ref-type="bibr">23</xref>).</p><p>Our study has several limitations. First, overall in our study, approximately 10% of all deaths could not be attributed to any of the causes of death, either because no VA interview was conducted or because the information provided by the informant was too unspecific for InterVA-4 to be able to assign a COD. Interestingly this fraction of unknown causes of death was found to vary by age. For young children between the age of 1 month and 5 years, we observed a higher fraction of VA interviews not being completed. This lower VA completion rate could be due to the parent not wanting to share information, because they are still traumatised by their loss. Although VA interviews in our setting are conducted at least 3 months after the date of death, a longer period might be considered for childhood deaths to give parents more time for the bereavement process. For neonates and persons older than 65 years, the fraction of indeterminate deaths (code 99) is higher than in other age groups and this could be an indication that in these age groups, assigning a COD is more difficult, either because informants do not have enough information or deaths occur with less discernible symptoms.</p><p>Another limitation is that although our overall sample size is large, many causes of deaths occurred infrequently: for 30 causes of death, the CSMF was less than 0.5% such that our study could be underpowered to assess sex-specific differences for these cause categories. Moreover, because of the large number of statistical tests conducted, we acknowledge that there is a potential risk of type I errors. Nevertheless, because our focus was on causes of death differences with <italic>p</italic>-values below 0.0001, a conservative Bonferroni correction for 50 tests would still make these sex-specific differences significant at level <italic>p</italic>&#x0003c;0.005.</p><p>As far as non-communicable diseases are concerned, we found that CSMFs of stroke and diabetes were generally higher in women than in men. As the age-stratified analysis has shown, this can be explained by the longer life expectancy we observe for women in our setting (<xref rid="CIT0006" ref-type="bibr">6</xref>), such that many more women than men attain an age where these underlying conditions can cause death. It is generally assumed that with the epidemiological transition occurring in low- and middle-income countries like South Africa, that non-communicable diseases will grow in importance and that this will eventually reflect in COD patterns. So far, there is little sign in our mortality data of such an increasing trend (data not shown), but it is obviously an area of interest and of global public health concern in the years to come.</p></sec><sec sec-type="conclusions" id="S0004"><title>Conclusions</title><p>Mortality during the 2000s in our DSS continued to be dominated by tuberculosis and HIV/AIDS. For certain communicable and non-communicable causes of death, we identified sex-specific differences of mortality fractions, some of which can be explained by the fact that there is a higher proportion of older women than men living in the DSS. InterVA-4 is a valuable tool for investigating causes of death patterns in Southern African settings.</p></sec>

Adult non-communicable disease mortality in Africa and Asia: evidence from INDEPTH Health and Demographic Surveillance System sites

<sec id="st1"><title>Background</title><p>Mortality from non-communicable diseases (NCDs) is a major global issue, as other categories of mortality have diminished and life expectancy has increased. The World Health Organization's Member States have called for a 25% reduction in premature NCD mortality by 2025, which can only be achieved by substantial reductions in risk factors and improvements in the management of chronic conditions. A high burden of NCD mortality among much older people, who have survived other hazards, is inevitable. The INDEPTH Network collects detailed individual data within defined Health and Demographic Surveillance sites. By registering deaths and carrying out verbal autopsies to determine cause of death across many such sites, using standardised methods, the Network seeks to generate population-based mortality statistics that are not otherwise available.</p></sec><sec id="st2"><title>Objective</title><p>To describe patterns of adult NCD mortality from INDEPTH Network sites across Africa and Asia, according to the WHO 2012 verbal autopsy (VA) cause categories, with separate consideration of premature (15&#x02013;64 years) and older (65+ years) NCD mortality.</p></sec><sec id="st3"><title>Design</title><p>All adult deaths at INDEPTH sites are routinely registered and followed up with VA interviews. For this study, VA archives were transformed into the WHO 2012 VA standard format and processed using the InterVA-4 model to assign cause of death. Routine surveillance data also provide person-time denominators for mortality rates.</p></sec><sec id="st4"><title>Results</title><p>A total of 80,726 adult (over 15 years) deaths were documented over 7,423,497 person-years of observation. NCDs were attributed as the cause for 35.6% of these deaths. Slightly less than half of adult NCD deaths occurred in the 15&#x02013;64 age group. Detailed results are presented by age and sex for leading causes of NCD mortality. Per-site rates of NCD mortality were significantly correlated with rates of HIV/AIDS-related mortality.</p></sec><sec id="st5"><title>Conclusions</title><p>These findings present important evidence on the distribution of NCD mortality across a wide range of African and Asian settings. This comes against a background of global concern about the burden of NCD mortality, especially among adults aged under 70, and provides an important baseline for future work.</p></sec>

<contrib contrib-type="author"><name><surname>Streatfield</surname><given-names>P. Kim</given-names></name><xref ref-type="aff" rid="AF0001">1</xref><xref ref-type="aff" rid="AF0002">2</xref><xref ref-type="aff" rid="AF0003">3</xref></contrib><contrib contrib-type="author"><name><surname>Khan</surname><given-names>Wasif A.</given-names></name><xref ref-type="aff" rid="AF0002">2</xref><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0004">4</xref></contrib><contrib contrib-type="author"><name><surname>Bhuiya</surname><given-names>Abbas</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0005">5</xref><xref ref-type="aff" rid="AF0006">6</xref></contrib><contrib contrib-type="author"><name><surname>Hanifi</surname><given-names>Syed M.A.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0005">5</xref><xref ref-type="aff" rid="AF0006">6</xref></contrib><contrib contrib-type="author"><name><surname>Alam</surname><given-names>Nurul</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0007">7</xref><xref ref-type="aff" rid="AF0008">8</xref></contrib><contrib contrib-type="author"><name><surname>Bagagnan</surname><given-names>Cheik H.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0009">9</xref><xref ref-type="aff" rid="AF0010">10</xref></contrib><contrib contrib-type="author"><name><surname>Si&#x000e9;</surname><given-names>Ali</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0009">9</xref><xref ref-type="aff" rid="AF0010">10</xref></contrib><contrib contrib-type="author"><name><surname>Zabr&#x000e9;</surname><given-names>Pascal</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0009">9</xref><xref ref-type="aff" rid="AF0010">10</xref></contrib><contrib contrib-type="author"><name><surname>Lankoand&#x000e9;</surname><given-names>Bruno</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0011">11</xref><xref ref-type="aff" rid="AF0012">12</xref></contrib><contrib contrib-type="author"><name><surname>Rossier</surname><given-names>Clementine</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0011">11</xref><xref ref-type="aff" rid="AF0012">12</xref></contrib><contrib contrib-type="author"><name><surname>Soura</surname><given-names>Abdramane B.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0011">11</xref><xref ref-type="aff" rid="AF0012">12</xref></contrib><contrib contrib-type="author"><name><surname>Bonfoh</surname><given-names>Bassirou</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0013">13</xref><xref ref-type="aff" rid="AF0014">14</xref></contrib><contrib contrib-type="author"><name><surname>Kone</surname><given-names>Siaka</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0013">13</xref><xref ref-type="aff" rid="AF0014">14</xref></contrib><contrib contrib-type="author"><name><surname>Ngoran</surname><given-names>Eliezer K.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0013">13</xref><xref ref-type="aff" rid="AF0015">15</xref></contrib><contrib contrib-type="author"><name><surname>Utzinger</surname><given-names>Juerg</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0013">13</xref><xref ref-type="aff" rid="AF0016">16</xref></contrib><contrib contrib-type="author"><name><surname>Haile</surname><given-names>Fisaha</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0017">17</xref><xref ref-type="aff" rid="AF0018">18</xref></contrib><contrib contrib-type="author"><name><surname>Melaku</surname><given-names>Yohannes A.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0017">17</xref><xref ref-type="aff" rid="AF0018">18</xref></contrib><contrib contrib-type="author"><name><surname>Weldearegawi</surname><given-names>Berhe</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0017">17</xref><xref ref-type="aff" rid="AF0018">18</xref></contrib><contrib contrib-type="author"><name><surname>Gomez</surname><given-names>Pierre</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0019">19</xref><xref ref-type="aff" rid="AF0020">20</xref></contrib><contrib contrib-type="author"><name><surname>Jasseh</surname><given-names>Momodou</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0019">19</xref><xref ref-type="aff" rid="AF0020">20</xref></contrib><contrib contrib-type="author"><name><surname>Ansah</surname><given-names>Patrick</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0021">21</xref><xref ref-type="aff" rid="AF0022">22</xref></contrib><contrib contrib-type="author"><name><surname>Debpuur</surname><given-names>Cornelius</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0021">21</xref><xref ref-type="aff" rid="AF0022">22</xref></contrib><contrib contrib-type="author"><name><surname>Oduro</surname><given-names>Abraham</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0021">21</xref><xref ref-type="aff" rid="AF0022">22</xref></contrib><contrib contrib-type="author"><name><surname>Wak</surname><given-names>George</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0021">21</xref><xref ref-type="aff" rid="AF0022">22</xref></contrib><contrib contrib-type="author"><name><surname>Adjei</surname><given-names>Alexander</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0023">23</xref><xref ref-type="aff" rid="AF0024">24</xref></contrib><contrib contrib-type="author"><name><surname>Gyapong</surname><given-names>Margaret</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0023">23</xref><xref ref-type="aff" rid="AF0025">25</xref></contrib><contrib contrib-type="author"><name><surname>Sarpong</surname><given-names>Doris</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0023">23</xref><xref ref-type="aff" rid="AF0025">25</xref></contrib><contrib contrib-type="author"><name><surname>Kant</surname><given-names>Shashi</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0025">25</xref><xref ref-type="aff" rid="AF0026">26</xref></contrib><contrib contrib-type="author"><name><surname>Misra</surname><given-names>Puneet</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0025">25</xref><xref ref-type="aff" rid="AF0026">26</xref></contrib><contrib contrib-type="author"><name><surname>Rai</surname><given-names>Sanjay K.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0025">25</xref><xref ref-type="aff" rid="AF0026">26</xref></contrib><contrib contrib-type="author"><name><surname>Juvekar</surname><given-names>Sanjay</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0027">27</xref><xref ref-type="aff" rid="AF0028">28</xref></contrib><contrib contrib-type="author"><name><surname>Lele</surname><given-names>Pallavi</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0027">27</xref><xref ref-type="aff" rid="AF0028">28</xref></contrib><contrib contrib-type="author"><name><surname>Bauni</surname><given-names>Evasius</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0029">29</xref><xref ref-type="aff" rid="AF0030">30</xref></contrib><contrib contrib-type="author"><name><surname>Mochamah</surname><given-names>George</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0029">29</xref><xref ref-type="aff" rid="AF0030">30</xref></contrib><contrib contrib-type="author"><name><surname>Ndila</surname><given-names>Carolyne</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0029">29</xref><xref ref-type="aff" rid="AF0030">30</xref></contrib><contrib contrib-type="author"><name><surname>Williams</surname><given-names>Thomas N.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0029">29</xref><xref ref-type="aff" rid="AF0030">30</xref><xref ref-type="aff" rid="AF0031">31</xref></contrib><contrib contrib-type="author"><name><surname>Laserson</surname><given-names>Kayla F.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0032">32</xref><xref ref-type="aff" rid="AF0033">33</xref></contrib><contrib contrib-type="author"><name><surname>Nyaguara</surname><given-names>Amek</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0032">32</xref><xref ref-type="aff" rid="AF0033">33</xref></contrib><contrib contrib-type="author"><name><surname>Odhiambo</surname><given-names>Frank O.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0032">32</xref><xref ref-type="aff" rid="AF0033">33</xref></contrib><contrib contrib-type="author"><name><surname>Phillips-Howard</surname><given-names>Penelope</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0032">32</xref><xref ref-type="aff" rid="AF0033">33</xref></contrib><contrib contrib-type="author"><name><surname>Ezeh</surname><given-names>Alex</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0034">34</xref><xref ref-type="aff" rid="AF0035">35</xref></contrib><contrib contrib-type="author"><name><surname>Kyobutungi</surname><given-names>Catherine</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0034">34</xref><xref ref-type="aff" rid="AF0035">35</xref></contrib><contrib contrib-type="author"><name><surname>Oti</surname><given-names>Samuel</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0034">34</xref><xref ref-type="aff" rid="AF0035">35</xref></contrib><contrib contrib-type="author"><name><surname>Crampin</surname><given-names>Amelia</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0036">36</xref><xref ref-type="aff" rid="AF0037">37</xref><xref ref-type="aff" rid="AF0038">38</xref></contrib><contrib contrib-type="author"><name><surname>Nyirenda</surname><given-names>Moffat</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0036">36</xref><xref ref-type="aff" rid="AF0037">37</xref></contrib><contrib contrib-type="author"><name><surname>Price</surname><given-names>Alison</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0036">36</xref><xref ref-type="aff" rid="AF0037">37</xref><xref ref-type="aff" rid="AF0038">38</xref></contrib><contrib contrib-type="author"><name><surname>Delaunay</surname><given-names>Val&#x000e9;rie</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0039">39</xref><xref ref-type="aff" rid="AF0040">40</xref></contrib><contrib contrib-type="author"><name><surname>Diallo</surname><given-names>Aldiouma</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0039">39</xref><xref ref-type="aff" rid="AF0040">40</xref></contrib><contrib contrib-type="author"><name><surname>Douillot</surname><given-names>Laetitia</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0039">39</xref><xref ref-type="aff" rid="AF0040">40</xref></contrib><contrib contrib-type="author"><name><surname>Sokhna</surname><given-names>Cheikh</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0039">39</xref><xref ref-type="aff" rid="AF0040">40</xref></contrib><contrib contrib-type="author"><name><surname>G&#x000f3;mez-Oliv&#x000e9;</surname><given-names>F. Xavier</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0041">41</xref><xref ref-type="aff" rid="AF0042">42</xref></contrib><contrib contrib-type="author"><name><surname>Kahn</surname><given-names>Kathleen</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0041">41</xref><xref ref-type="aff" rid="AF0042">42</xref><xref ref-type="aff" rid="AF0043">43</xref></contrib><contrib contrib-type="author"><name><surname>Tollman</surname><given-names>Stephen M.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0041">41</xref><xref ref-type="aff" rid="AF0042">42</xref><xref ref-type="aff" rid="AF0043">43</xref></contrib><contrib contrib-type="author"><name><surname>Herbst</surname><given-names>Kobus</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0044">44</xref><xref ref-type="aff" rid="AF0045">45</xref></contrib><contrib contrib-type="author"><name><surname>Mossong</surname><given-names>Jo&#x000eb;l</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0044">44</xref><xref ref-type="aff" rid="AF0046">46</xref></contrib><contrib contrib-type="author"><name><surname>Chuc</surname><given-names>Nguyen T.K.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0047">47</xref><xref ref-type="aff" rid="AF0048">48</xref></contrib><contrib contrib-type="author"><name><surname>Bangha</surname><given-names>Martin</given-names></name><xref ref-type="aff" rid="AF0003">3</xref></contrib><contrib contrib-type="author"><name><surname>Sankoh</surname><given-names>Osman A.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0049">49</xref><xref ref-type="aff" rid="AF0050">50</xref><xref ref-type="corresp" rid="cor1">*</xref></contrib><contrib contrib-type="author"><name><surname>Byass</surname><given-names>Peter</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0042">42</xref><xref ref-type="aff" rid="AF0051">51</xref></contrib>

Global Health Action

<p>Mortality due to non-communicable diseases (NCDs) of various kinds has become an increasing focus of attention in recent years (<xref rid="CIT0001" ref-type="bibr">1</xref>), including at a special meeting of the UN General Assembly in September 2011 (<xref rid="CIT0002" ref-type="bibr">2</xref>). This has happened both because of rapid increases in major NCD risk factors to which billions of people globally are exposed and also because many populations are ageing more successfully as a result of various reductions in early-life mortality and increases in mid-life survival. Consequently, more people globally are experiencing the consequences of NCD risk factors in middle and later life, according to the classic theories of epidemiological transition (<xref rid="CIT0003" ref-type="bibr">3</xref>). In 2012, WHO Member States at the World Health Assembly made a resolution to &#x02018;adopt a global target of a 25% reduction in premature mortality from non-communicable diseases by 2025&#x02019; (<xref rid="CIT0004" ref-type="bibr">4</xref>), although this has been widely misquoted with the omission of the word &#x02018;premature&#x02019;. The biological reality is that human beings who survive into their 80s and 90s, thereby having avoided death from many other causes, are extremely likely to die of NCDs. The public health imperatives around NCDs therefore relate to preventing or delaying NCD incidence (including risk factor management) and effectively managing chronic conditions in the decades of mid-life.</p><p> There has been much written about epidemics of NCD morbidity and mortality globally, and particularly in low- and middle-income countries (<xref rid="CIT0005" ref-type="bibr">5</xref>). While it is certainly true that there are many more NCD deaths now than there have been previously, it is critical to understand the medical, social, and demographic drivers of these changes, in addition to trends in specific risk factors, diseases, and health systems responses. This is important in terms of positioning the burden of NCDs appropriately in future development agendas (<xref rid="CIT0006" ref-type="bibr">6</xref>).</p><p>The INDEPTH Network works with Health and Demographic Surveillance sites (HDSS) across Africa and Asia, which each follow circumscribed populations on a longitudinal basis. Core data collected include person-time at risk, together with deaths and, by means of verbal autopsy (VA), assessment of cause of death (<xref rid="CIT0007" ref-type="bibr">7</xref>).</p><p>Our aim in this paper is to document deaths among adults (15 years upwards) on the basis of a dataset collected at 22 INDEPTH HDSSs covering Africa and Asia (<xref rid="CIT0008" ref-type="bibr">8</xref>), looking particularly at those deaths attributable to NCDs. Although these 22 sites are not designed to be a representative sample, they enable comparisons to be made over widely differing situations.</p><sec sec-type="methods" id="S0002"><title>Methods</title><p>The overall INDEPTH dataset (<xref rid="CIT0009" ref-type="bibr">9</xref>) from which these adult NCD mortality analyses are drawn is described in detail elsewhere (<xref rid="CIT0008" ref-type="bibr">8</xref>). Across the 22 participating sites (<xref rid="CIT0010" ref-type="bibr">10</xref>&#x02013;<xref rid="CIT0031" ref-type="bibr">31</xref>) , there was documentation on 80,726 deaths in 7,423,497 person-years of observation for people aged 15 and over. VA interviews were successfully completed on 72,330 (89.6%) of the deaths that occurred. A summary of the detailed methods used in common for this series of multisite papers is shown in <xref ref-type="boxed-text" rid="B0001">Box 1</xref>. NCD mortality was defined as neoplasms, metabolic, cardiovascular, respiratory, abdominal, neurological and other NCDs, corresponding to chapters in WHO 2012 VA standard.</p><p><italic>Box 1.</italic> Summary of methodology based on the detailed description in the introductory paper (<xref rid="CIT0008" ref-type="bibr">8</xref>).</p><boxed-text id="B0001" position="float"><p> <bold>Age&#x02013;sex&#x02013;time standardisation</bold> </p><p>To avoid effects of differences and changes in age&#x02013; sex structures of populations, mortality fractions and rates have been adjusted using the INDEPTH 2013 population standard (<xref rid="CIT0032" ref-type="bibr">32</xref>). A weighting factor was calculated for each site, age group, sex, and year category in relation to the standard for the corresponding age group and sex, and incorporated into the overall dataset. This is referred to in this paper as age&#x02013;sex-time standardisation in the contexts where it is used.</p><p> <bold>Cause of death assignment</bold> </p><p>The InterVA-4 (version 4.02) probabilistic model was used for all the cause of death assignments in the overall dataset (<xref rid="CIT0033" ref-type="bibr">33</xref>). InterVA-4 is fully compliant with the WHO 2012 VA standard and generates causes of death categorised by ICD-10 groups (<xref rid="CIT0034" ref-type="bibr">34</xref>). The data reported here were collected before the WHO 2012 VA standard was available, but were transformed into the WHO 2012 and InterVA-4 format to optimise cross-site standardisation in cause of death attribution. For a small proportion of deaths, VA interviews were not successfully completed; a few others contained inadequate information to arrive at a cause of death. InterVA-4 assigns causes of death (maximum 3) with associated likelihoods; thus, cases for which likely causes did not total to 100% were also assigned a residual indeterminate component. This served as a means of encapsulating uncertainty in cause of death at the individual level within the overall dataset, as well as accounting for 100% of every death.</p><p> <bold>Overall dataset</bold> </p><p>The overall public-domain dataset (<xref rid="CIT0009" ref-type="bibr">9</xref>) thus contains between one and four records for each death, with the sum of likelihoods for each individual being unity. Each record includes a specific cause of death, its likelihood, and its age&#x02013;sex&#x02013;time weighting.</p></boxed-text><p>In this context, all of these data are secondary datasets derived from primary data collected separately by each participating site. In all cases, the primary data collection was covered by site-level ethical approvals relating to on-going health and demographic surveillance in those specific locations. No individual identity or household location data were included in the secondary data and no specific ethical approvals were required for these pooled analyses.</p></sec><sec sec-type="results" id="S0003"><title>Results</title><p>After age&#x02013;sex&#x02013;time standardisation, 81814.1 deaths were documented, of which 73288.7 were covered by VA interviews; the InterVA-4 model was unable to assign any cause of death for 3458.3 deaths (4.2%), and residual uncertainties (residual likelihoods in cases where likely cause(s) of death did not total to 100%) accounted for 5864.2 deaths (7.2%). As a consequence of sites returning data over a range of time periods, 4.6% of the person-time observed occurred before 2000, 30.0% from 2000 to 2005, and 69.4% from 2006 to 2012.</p><p>Non-NCD causes were attributed to 35478.5 deaths (43.4%), while 28487.7 deaths (34.8%) were attributed to 20 specific WHO 2012 cause of death categories, plus the residual &#x02018;other and unspecified NCD&#x02019; category. Since the Purworejo, Indonesia, site achieved limited coverage of adult VAs, and did not report for the period 2006&#x02013;2012, that site is excluded from the following analyses (1173.4 deaths in 185,306 person-years).</p><p> <xref ref-type="fig" rid="F0001">Figure 1</xref> maps the 21 INDEPTH sites, showing the age&#x02013;sex&#x02013;time standardised NCD-specific fraction of adult mortality and the age&#x02013;sex&#x02013;time standardised NCD-specific mortality rate per 1,000 person-years for each site. This shows a relatively narrow range of variation across sites, while the sites with the highest proportions of NCD-attributable mortality are not necessarily those with the highest rates. Generally, higher proportions of NCD mortality, but not always higher rates, were found in Asia compared to Africa. The urban site in Ouagadougou, Burkina Faso, returned the highest proportion of NCD mortality in Africa (46.6%), while the Africa Centre, South Africa, site reported the highest rate of NCD mortality (5.79 per 1,000 person-years). The highest proportion and rate of NCD mortality in Asia was 66.9% and 4.39 per 1,000 person-years, respectively, at the AMK, Bangladesh, site.</p><fig id="F0001" position="float"><label>Fig. 1</label><caption><p>Map showing age&#x02013;sex&#x02013;time standardised proportions of mortality due to non-communicable diseases and age&#x02013;sex&#x02013;time standardised non-communicable disease mortality rates per 1,000 person-years, for 21 INDEPTH sites.</p></caption><graphic xlink:href="GHA-7-25365-g001"/></fig><p> <xref ref-type="table" rid="T0001">Table 1</xref> shows mortality rates per 1,000 person-years for NCD causes. Age groupings follow the WHO 2012 categories (15&#x02013;49, 50&#x02013;64, 65+). NCD mortality rates were substantially lower in the 15&#x02013;49 age group (range 0.31 to 2.76 per 1,000 person-years across all sites and periods) compared to the 50&#x02013;64 age group (range 2.81 to 14.06 per 1,000 person-years) and more so compared to the 65-plus age group (range 10.72 to 97.20 per 1,000 person-years).</p><table-wrap id="T0001" position="float"><label>Table 1</label><caption><p>NCD mortality rates per 1,000 person-years by site, age group and period</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="2" valign="middle" colspan="1">Age group</th><th align="center" colspan="3" rowspan="1">15&#x02013;49 years</th><th align="center" colspan="3" rowspan="1">50&#x02013;64 years</th><th align="center" colspan="3" rowspan="1">65+ years</th></tr><tr><th colspan="9" align="left" rowspan="1"><hr/></th></tr><tr><th align="left" rowspan="1" colspan="1">Period</th><th align="center" rowspan="1" colspan="1">&#x0003c;2000</th><th align="center" rowspan="1" colspan="1">2000&#x02013;2005</th><th align="center" rowspan="1" colspan="1">2006&#x02013;2012</th><th align="center" rowspan="1" colspan="1">&#x0003c;2000</th><th align="center" rowspan="1" colspan="1">2000&#x02013;2005</th><th align="center" rowspan="1" colspan="1">2006&#x02013;2012</th><th align="center" rowspan="1" colspan="1">&#x0003c;2000</th><th align="center" rowspan="1" colspan="1">2000&#x02013;2005</th><th align="center" rowspan="1" colspan="1">2006&#x02013;2012</th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">Matlab, Bangladesh</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.81</td><td align="center" rowspan="1" colspan="1">0.83</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">6.33</td><td align="center" rowspan="1" colspan="1">7.41</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">37.55</td><td align="center" rowspan="1" colspan="1">41.27</td></tr><tr><td align="left" rowspan="1" colspan="1">Bandarban, Bangladesh</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.83</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">4.80</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">18.31</td></tr><tr><td align="left" rowspan="1" colspan="1">Chakaria, Bangladesh</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.77</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">7.09</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">35.58</td></tr><tr><td align="left" rowspan="1" colspan="1">AMK, Bangladesh</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.87</td><td align="center" rowspan="1" colspan="1">0.75</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">6.60</td><td align="center" rowspan="1" colspan="1">7.12</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">45.29</td><td align="center" rowspan="1" colspan="1">43.82</td></tr><tr><td align="left" rowspan="1" colspan="1">Nouna, Burkina Faso</td><td align="center" rowspan="1" colspan="1">0.70</td><td align="center" rowspan="1" colspan="1">0.48</td><td align="center" rowspan="1" colspan="1">0.27</td><td align="center" rowspan="1" colspan="1">7.11</td><td align="center" rowspan="1" colspan="1">3.64</td><td align="center" rowspan="1" colspan="1">2.80</td><td align="center" rowspan="1" colspan="1">23.22</td><td align="center" rowspan="1" colspan="1">16.81</td><td align="center" rowspan="1" colspan="1">9.67</td></tr><tr><td align="left" rowspan="1" colspan="1">Ouagadougou, Burkina Faso</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.63</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">5.87</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">21.29</td></tr><tr><td align="left" rowspan="1" colspan="1">Taabo, C&#x000f4;te d'Ivoire</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.67</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">3.77</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">16.91</td></tr><tr><td align="left" rowspan="1" colspan="1">Kilite Awlaelo, Ethiopia</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.49</td><td align="center" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1.52</td><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">12.32</td></tr><tr><td align="left" rowspan="1" colspan="1">Farafenni, The Gambia</td><td align="center" rowspan="1" colspan="1">1.07</td><td align="center" rowspan="1" colspan="1">0.85</td><td align="center" rowspan="1" colspan="1">0.65</td><td align="center" rowspan="1" colspan="1">7.61</td><td align="center" rowspan="1" colspan="1">6.93</td><td align="center" rowspan="1" colspan="1">4.06</td><td align="center" rowspan="1" colspan="1">24.57</td><td align="center" rowspan="1" colspan="1">29.12</td><td align="center" rowspan="1" colspan="1">21.58</td></tr><tr><td align="left" rowspan="1" colspan="1">Navrongo, Ghana</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1.87</td><td align="center" rowspan="1" colspan="1">1.72</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">10.16</td><td align="center" rowspan="1" colspan="1">8.96</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">26.82</td><td align="center" rowspan="1" colspan="1">24.51</td></tr><tr><td align="left" rowspan="1" colspan="1">Dodowa, Ghana</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.84</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">5.55</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">19.07</td></tr><tr><td align="left" rowspan="1" colspan="1">Ballabgarh, India</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.56</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">6.00</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">30.44</td></tr><tr><td align="left" rowspan="1" colspan="1">Vadu, India</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.43</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">3.90</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">17.46</td></tr><tr><td align="left" rowspan="1" colspan="1">Kilifi, Kenya</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.65</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">5.45</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">27.68</td></tr><tr><td align="left" rowspan="1" colspan="1">Kisumu, Kenya</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">2.44</td><td align="center" rowspan="1" colspan="1">1.65</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">7.96</td><td align="center" rowspan="1" colspan="1">7.36</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">25.95</td><td align="center" rowspan="1" colspan="1">32.05</td></tr><tr><td align="left" rowspan="1" colspan="1">Nairobi, Kenya</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.45</td><td align="center" rowspan="1" colspan="1">0.61</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">4.21</td><td align="center" rowspan="1" colspan="1">2.76</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">12.64</td><td align="center" rowspan="1" colspan="1">19.37</td></tr><tr><td align="left" rowspan="1" colspan="1">Karonga, Malawi</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1.63</td><td align="center" rowspan="1" colspan="1">0.79</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">5.70</td><td align="center" rowspan="1" colspan="1">4.70</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">27.81</td><td align="center" rowspan="1" colspan="1">23.31</td></tr><tr><td align="left" rowspan="1" colspan="1">Niakhar, Senegal</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.76</td><td align="center" rowspan="1" colspan="1">0.56</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">3.15</td><td align="center" rowspan="1" colspan="1">2.54</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">24.22</td><td align="center" rowspan="1" colspan="1">16.15</td></tr><tr><td align="left" rowspan="1" colspan="1">Agincourt, South Africa</td><td align="center" rowspan="1" colspan="1">0.53</td><td align="center" rowspan="1" colspan="1">1.02</td><td align="center" rowspan="1" colspan="1">1.25</td><td align="center" rowspan="1" colspan="1">4.01</td><td align="center" rowspan="1" colspan="1">6.88</td><td align="center" rowspan="1" colspan="1">7.79</td><td align="center" rowspan="1" colspan="1">17.10</td><td align="center" rowspan="1" colspan="1">22.48</td><td align="center" rowspan="1" colspan="1">27.26</td></tr><tr><td align="left" rowspan="1" colspan="1">Africa Centre, South Africa</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1.26</td><td align="center" rowspan="1" colspan="1">1.06</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">10.54</td><td align="center" rowspan="1" colspan="1">10.50</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">36.86</td><td align="center" rowspan="1" colspan="1">36.93</td></tr><tr><td align="left" rowspan="1" colspan="1">FilaBavi, Vietnam</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.96</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">3.76</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">28.63</td></tr></tbody></table></table-wrap><p> <xref ref-type="fig" rid="F0002">Figure 2</xref> shows the age&#x02013;sex&#x02013;time standardised mortality rates by site for sub-categories of NCD causes of death (neoplasms, metabolic, cardiovascular, respiratory, abdominal, neurological and other NCDs, according to chapters in WHO 2012 VA standard (34)) for 21 INDEPTH HDSS sites. These categories differ slightly from the four main NCD groups that are commonly considered in the global NCD literature, namely cancer, diabetes, cardiovascular disease, and chronic respiratory disease. Cancer and cardiovascular disease map directly to the WHO 2012 VA neoplasms and cardiovascular disease chapters, respectively. Diabetes accounted for 72% of deaths in the metabolic disease chapter, and chronic obstructive pulmonary disease accounted for 65% of deaths in the chronic respiratory disease chapter.</p><fig id="F0002" position="float"><label>Fig. 2</label><caption><p>Age&#x02013;sex&#x02013;time standardised mortality rates per 1,000 person-years among adults (15 years and over) in 21 INDEPTH HDSS sites in Africa and Asia, by sub-category of non-communicable diseases causing death (according to WHO 2012 VA cause of death chapters).</p></caption><graphic xlink:href="GHA-7-25365-g002"/></fig><p>The population burden of premature NCD mortality has to be considered in terms of relative numbers of deaths between age groups, rather than only cause-specific rates, because of the very different population proportions in various age groups. While the mortality rate ratios for NCD mortality between the 15&#x02013;64 and over-65 year age groups varied from 1:7 (Navrongo, Ghana) to 1:107 (Nairobi, Kenya), with a median of 1:18 (Niakhar, Senegal), this does not mean that the overwhelming burden of NCD mortality in terms of numbers of deaths lay within the over-65 age group. Person-time for the over-65 age group ranged from 1.4% of that for the 15&#x02013;64 age group (Nairobi, Kenya) to 13.8% (FilaBavi, Vietnam), with a median of 8.2% (Chakaria, Bangladesh). NCD mortality rates for the 15&#x02013;64 year age group ranged from 0.12 per 1,000 person-years in the Matlab, Bangladesh, site to 0.77 per 1,000 person-years in the Africa Centre, South Africa, site. Similarly for the over-65 age group, the range was from 1.24 in FilaBavi, Vietnam, to 13.4 in Nairobi, Kenya. Consequently, the numbers of NCD deaths in the 15&#x02013;64 age group, which might be considered as &#x02018;premature&#x02019; NCD mortality, were considerable. <xref ref-type="fig" rid="F0003">Figure 3</xref> presents the percentages of adult NCD deaths for the 15&#x02013;64 and over 65-year age groups by site, using the same cause categories as <xref ref-type="fig" rid="F0002">Fig. 2</xref>, to illustrate the magnitude of the burden of premature NCD mortality. The overall bar for each site represents 100% of NCD deaths, split by cause categories and age group. Consequently, bars relating to higher life expectancy countries, with higher proportions of older people, tend to be shifted further to the right. Slightly more than half of the overall adult NCD deaths occurred in the over-65 age group, to the right; deaths shown to the left of the axis might be considered as &#x02018;premature&#x02019;. Sites are shown in decreasing order of overall adult NCD mortality rate, the same as in <xref ref-type="fig" rid="F0002">Fig. 2</xref>, for ease of comparison.</p><fig id="F0003" position="float"><label>Fig. 3</label><caption><p>Age&#x02013;sex&#x02013;time standardised percentages of adult NCD deaths for the 15&#x02013;64 and over 65 year age groups by site and cause category.</p></caption><graphic xlink:href="GHA-7-25365-g003"/></fig><p> <xref ref-type="table" rid="T0002">Table 2</xref> shows cause-specific rates at each site for the various categories of neoplasms specified in the WHO 2012 VA standard, separately for the under- and over-65 year age groups, and by sex, for the period 2006&#x02013;2012.</p><table-wrap id="T0002" position="float"><label>Table 2</label><caption><p>Cause-specific adult mortality rates for neoplasms (according to WHO 2012 VA cause categories), by site, age group, and sex for 2006&#x02013;2012</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1"/><th align="center" colspan="4" rowspan="1">Oral neoplasms</th><th align="center" colspan="4" rowspan="1">Digestive neoplasms</th><th align="center" colspan="4" rowspan="1">Respiratory neoplasms</th><th align="center" colspan="2" rowspan="1">Breast neoplasms</th><th align="center" colspan="4" rowspan="1">Reproductive neoplasms</th><th align="center" colspan="4" rowspan="1">Other neoplasms</th></tr><tr><th align="left" rowspan="1" colspan="1"/><th colspan="22" align="left" rowspan="1"><hr/></th></tr><tr><th align="left" rowspan="1" colspan="1">Sex</th><th align="center" colspan="2" rowspan="1">Males</th><th align="center" colspan="2" rowspan="1">Females</th><th align="center" colspan="2" rowspan="1">Males</th><th align="center" colspan="2" rowspan="1">Females</th><th align="center" colspan="2" rowspan="1">Males</th><th align="center" colspan="2" rowspan="1">Females</th><th align="center" colspan="2" rowspan="1">Females</th><th align="center" colspan="2" rowspan="1">Males</th><th align="center" colspan="2" rowspan="1">Females</th><th align="center" colspan="2" rowspan="1">Males</th><th align="center" colspan="2" rowspan="1">Females</th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">Age (years)</td><td align="center" rowspan="1" colspan="1">&#x0003c;65</td><td align="center" rowspan="1" colspan="1">65+</td><td align="center" rowspan="1" colspan="1">&#x0003c;65</td><td align="center" rowspan="1" colspan="1">65+</td><td align="center" rowspan="1" colspan="1">&#x0003c;65</td><td align="center" rowspan="1" colspan="1">65+</td><td align="center" rowspan="1" colspan="1">&#x0003c;65</td><td align="center" rowspan="1" colspan="1">65+</td><td align="center" rowspan="1" colspan="1">&#x0003c;65</td><td align="center" rowspan="1" colspan="1">65+</td><td align="center" rowspan="1" colspan="1">&#x0003c;65</td><td align="center" rowspan="1" colspan="1">65+</td><td align="center" rowspan="1" colspan="1">&#x0003c;65</td><td align="center" rowspan="1" colspan="1">65+</td><td align="center" rowspan="1" colspan="1">&#x0003c;65</td><td align="center" rowspan="1" colspan="1">65+</td><td align="center" rowspan="1" colspan="1">&#x0003c;65</td><td align="center" rowspan="1" colspan="1">65+</td><td align="center" rowspan="1" colspan="1">&#x0003c;65</td><td align="center" rowspan="1" colspan="1">65+</td><td align="center" rowspan="1" colspan="1">&#x0003c;65</td><td align="center" rowspan="1" colspan="1">65+</td></tr><tr><td align="left" rowspan="1" colspan="1">Bangladesh: Matlab</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.45</td><td align="center" rowspan="1" colspan="1">3.84</td><td align="center" rowspan="1" colspan="1">0.23</td><td align="center" rowspan="1" colspan="1">2.35</td><td align="center" rowspan="1" colspan="1">0.38</td><td align="center" rowspan="1" colspan="1">5.41</td><td align="center" rowspan="1" colspan="1">0.12</td><td align="center" rowspan="1" colspan="1">2.80</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.56</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">1.34</td><td align="center" rowspan="1" colspan="1">0.19</td><td align="center" rowspan="1" colspan="1">1.95</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">1.89</td></tr><tr><td align="left" rowspan="1" colspan="1">Bangladesh: Bandarban</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.61</td><td align="center" rowspan="1" colspan="1">0.70</td><td align="center" rowspan="1" colspan="1">0.19</td><td align="center" rowspan="1" colspan="1">1.30</td><td align="center" rowspan="1" colspan="1">0.17</td><td align="center" rowspan="1" colspan="1">3.88</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1.97</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.14</td><td align="center" rowspan="1" colspan="1">1.40</td><td align="center" rowspan="1" colspan="1">0.24</td><td align="center" rowspan="1" colspan="1">0.71</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.77</td></tr><tr><td align="left" rowspan="1" colspan="1">Bangladesh: Chakaria</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.2</td><td align="center" rowspan="1" colspan="1">0.16</td><td align="center" rowspan="1" colspan="1">1.82</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">1.38</td><td align="center" rowspan="1" colspan="1">0.16</td><td align="center" rowspan="1" colspan="1">4.47</td><td align="center" rowspan="1" colspan="1">0.15</td><td align="center" rowspan="1" colspan="1">2.36</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.37</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.69</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">1.01</td><td align="center" rowspan="1" colspan="1">0.30</td><td align="center" rowspan="1" colspan="1">2.55</td><td align="center" rowspan="1" colspan="1">0.18</td><td align="center" rowspan="1" colspan="1">1.72</td></tr><tr><td align="left" rowspan="1" colspan="1">Bangladesh: AMK</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.19</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.30</td><td align="center" rowspan="1" colspan="1">2.23</td><td align="center" rowspan="1" colspan="1">0.25</td><td align="center" rowspan="1" colspan="1">3.12</td><td align="center" rowspan="1" colspan="1">0.19</td><td align="center" rowspan="1" colspan="1">5.40</td><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">1.76</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.42</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">1.02</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">1.46</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">1.55</td></tr><tr><td align="left" rowspan="1" colspan="1">Burkina Faso: Nouna</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.35</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.23</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">Burkina Faso: Ouagadougou</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.20</td><td align="center" rowspan="1" colspan="1">3.85</td><td align="center" rowspan="1" colspan="1">0.15</td><td align="center" rowspan="1" colspan="1">0.64</td><td align="center" rowspan="1" colspan="1">0.19</td><td align="center" rowspan="1" colspan="1">2.55</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">1.69</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.76</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.52</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">0.72</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.31</td></tr><tr><td align="left" rowspan="1" colspan="1">Cote d'Ivoire: Taabo</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">1.55</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.23</td><td align="center" rowspan="1" colspan="1">1.52</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">1.69</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.37</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.14</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1.14</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">2.21</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1.25</td></tr><tr><td align="left" rowspan="1" colspan="1">Ethiopia: Kilite Awlaelo</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.80</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">1.27</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">1.88</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">1.27</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.41</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.44</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.87</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1.43</td></tr><tr><td align="left" rowspan="1" colspan="1">The Gambia: Farafenni</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.52</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.32</td><td align="center" rowspan="1" colspan="1">2.59</td><td align="center" rowspan="1" colspan="1">0.15</td><td align="center" rowspan="1" colspan="1">1.52</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">1.40</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">1.03</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.99</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.14</td><td align="center" rowspan="1" colspan="1">2.51</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">1.84</td></tr><tr><td align="left" rowspan="1" colspan="1">Ghana: Navrongo</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.38</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">1.57</td><td align="center" rowspan="1" colspan="1">9.51</td><td align="center" rowspan="1" colspan="1">0.89</td><td align="center" rowspan="1" colspan="1">5.25</td><td align="center" rowspan="1" colspan="1">0.29</td><td align="center" rowspan="1" colspan="1">3.78</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">1.38</td><td align="center" rowspan="1" colspan="1">0.14</td><td align="center" rowspan="1" colspan="1">0.34</td><td align="center" rowspan="1" colspan="1">0.13</td><td align="center" rowspan="1" colspan="1">1.08</td><td align="center" rowspan="1" colspan="1">0.15</td><td align="center" rowspan="1" colspan="1">1.80</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.33</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.29</td></tr><tr><td align="left" rowspan="1" colspan="1">Ghana: Dodowa</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.32</td><td align="center" rowspan="1" colspan="1">3.37</td><td align="center" rowspan="1" colspan="1">0.25</td><td align="center" rowspan="1" colspan="1">1.99</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.56</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.53</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.16</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.2</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.48</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">1.18</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.84</td></tr><tr><td align="left" rowspan="1" colspan="1">India: Ballabgarh</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.26</td><td align="center" rowspan="1" colspan="1">0.27</td><td align="center" rowspan="1" colspan="1">2.82</td><td align="center" rowspan="1" colspan="1">0.19</td><td align="center" rowspan="1" colspan="1">2.13</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">1.16</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.80</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.12</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">2.06</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">1.13</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.31</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.30</td></tr><tr><td align="left" rowspan="1" colspan="1">India: Vadu</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.47</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.27</td><td align="center" rowspan="1" colspan="1">5.95</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">1.17</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">1.80</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">1.08</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">1.43</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.19</td></tr><tr><td align="left" rowspan="1" colspan="1">Kenya: Kilifi</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.45</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.17</td><td align="center" rowspan="1" colspan="1">2.85</td><td align="center" rowspan="1" colspan="1">0.12</td><td align="center" rowspan="1" colspan="1">0.94</td><td align="center" rowspan="1" colspan="1">0.29</td><td align="center" rowspan="1" colspan="1">4.02</td><td align="center" rowspan="1" colspan="1">0.14</td><td align="center" rowspan="1" colspan="1">2.36</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">1.02</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.55</td><td align="center" rowspan="1" colspan="1">0.14</td><td align="center" rowspan="1" colspan="1">3.16</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">1.27</td></tr><tr><td align="left" rowspan="1" colspan="1">Kenya: Kisumu</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">0.39</td><td align="center" rowspan="1" colspan="1">3.33</td><td align="center" rowspan="1" colspan="1">0.31</td><td align="center" rowspan="1" colspan="1">2.01</td><td align="center" rowspan="1" colspan="1">0.36</td><td align="center" rowspan="1" colspan="1">5.11</td><td align="center" rowspan="1" colspan="1">0.27</td><td align="center" rowspan="1" colspan="1">3.19</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.34</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">0.55</td><td align="center" rowspan="1" colspan="1">0.46</td><td align="center" rowspan="1" colspan="1">3.47</td><td align="center" rowspan="1" colspan="1">0.19</td><td align="center" rowspan="1" colspan="1">3.31</td></tr><tr><td align="left" rowspan="1" colspan="1">Kenya: Nairobi</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.14</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.86</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">1.48</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">4.41</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">2.80</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.45</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">1.71</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.84</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.44</td></tr><tr><td align="left" rowspan="1" colspan="1">Malawi: Karonga</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.22</td><td align="center" rowspan="1" colspan="1">2.56</td><td align="center" rowspan="1" colspan="1">0.20</td><td align="center" rowspan="1" colspan="1">1.12</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.25</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.21</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.55</td><td align="center" rowspan="1" colspan="1">0.17</td><td align="center" rowspan="1" colspan="1">1.27</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">Senegal: Niakhar</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">0.55</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.19</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.14</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.13</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.24</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.82</td><td align="center" rowspan="1" colspan="1">0.25</td><td align="center" rowspan="1" colspan="1">4.59</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">2.02</td></tr><tr><td align="left" rowspan="1" colspan="1">South Africa: Agincourt</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">0.24</td><td align="center" rowspan="1" colspan="1">1.29</td><td align="center" rowspan="1" colspan="1">0.17</td><td align="center" rowspan="1" colspan="1">2.69</td><td align="center" rowspan="1" colspan="1">0.26</td><td align="center" rowspan="1" colspan="1">2.70</td><td align="center" rowspan="1" colspan="1">0.19</td><td align="center" rowspan="1" colspan="1">2.68</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">0.13</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.24</td><td align="center" rowspan="1" colspan="1">0.15</td><td align="center" rowspan="1" colspan="1">0.68</td><td align="center" rowspan="1" colspan="1">0.12</td><td align="center" rowspan="1" colspan="1">0.99</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">1.69</td></tr><tr><td align="left" rowspan="1" colspan="1">South Africa: Africa Centre</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.38</td><td align="center" rowspan="1" colspan="1">1.88</td><td align="center" rowspan="1" colspan="1">0.18</td><td align="center" rowspan="1" colspan="1">0.86</td><td align="center" rowspan="1" colspan="1">0.34</td><td align="center" rowspan="1" colspan="1">5.42</td><td align="center" rowspan="1" colspan="1">0.20</td><td align="center" rowspan="1" colspan="1">1.62</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.27</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">1.31</td><td align="center" rowspan="1" colspan="1">0.23</td><td align="center" rowspan="1" colspan="1">1.47</td><td align="center" rowspan="1" colspan="1">0.18</td><td align="center" rowspan="1" colspan="1">2.31</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">1.03</td></tr><tr><td align="left" rowspan="1" colspan="1">Vietnam: FilaBavi</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.24</td><td align="center" rowspan="1" colspan="1">2.85</td><td align="center" rowspan="1" colspan="1">0.17</td><td align="center" rowspan="1" colspan="1">1.37</td><td align="center" rowspan="1" colspan="1">0.38</td><td align="center" rowspan="1" colspan="1">8.61</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">2.74</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.50</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">1.06</td><td align="center" rowspan="1" colspan="1">0.20</td><td align="center" rowspan="1" colspan="1">3.50</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">1.60</td></tr></tbody></table></table-wrap><p> <xref ref-type="table" rid="T0003">Table 3</xref> shows cause-specific rates at each site for other selected categories of NCDs specified in the WHO 2012 VA standard, separately for the under- and over-65 year age groups, and by sex, for the period 2006&#x02013;2012.</p><table-wrap id="T0003" position="float"><label>Table 3</label><caption><p>Cause-specific adult mortality rates for selected NCDs (according to WHO 2012 VA cause categories), by site, age group, and sex for 2006&#x02013;2012</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1"/><th align="center" colspan="4" rowspan="1">Diabetes mellitus</th><th align="center" colspan="4" rowspan="1">Acute cardiac disease</th><th align="center" colspan="4" rowspan="1">Stroke</th><th align="center" colspan="4" rowspan="1">Other cardiac disease</th><th align="center" colspan="4" rowspan="1">Chronic obstructive pulmonary disease</th></tr><tr><th align="left" rowspan="1" colspan="1"/><th colspan="20" align="left" rowspan="1"><hr/></th></tr><tr><th align="left" rowspan="1" colspan="1">Sex</th><th align="center" colspan="2" rowspan="1">Males</th><th align="center" colspan="2" rowspan="1">Females</th><th align="center" colspan="2" rowspan="1">Males</th><th align="center" colspan="2" rowspan="1">Females</th><th align="center" colspan="2" rowspan="1">Males</th><th align="center" colspan="2" rowspan="1">Females</th><th align="center" colspan="2" rowspan="1">Males</th><th align="center" colspan="2" rowspan="1">Females</th><th align="center" colspan="2" rowspan="1">Males</th><th align="center" colspan="2" rowspan="1">Females</th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">Age (years)</td><td align="center" rowspan="1" colspan="1">&#x0003c;65</td><td align="center" rowspan="1" colspan="1">65+</td><td align="center" rowspan="1" colspan="1">&#x0003c;65</td><td align="center" rowspan="1" colspan="1">65+</td><td align="center" rowspan="1" colspan="1">&#x0003c;65</td><td align="center" rowspan="1" colspan="1">65+</td><td align="center" rowspan="1" colspan="1">&#x0003c;65</td><td align="center" rowspan="1" colspan="1">65+</td><td align="center" rowspan="1" colspan="1">&#x0003c;65</td><td align="center" rowspan="1" colspan="1">65+</td><td align="center" rowspan="1" colspan="1">&#x0003c;65</td><td align="center" rowspan="1" colspan="1">65+</td><td align="center" rowspan="1" colspan="1">&#x0003c;65</td><td align="center" rowspan="1" colspan="1">65+</td><td align="center" rowspan="1" colspan="1">&#x0003c;65</td><td align="center" rowspan="1" colspan="1">65+</td><td align="center" rowspan="1" colspan="1">&#x0003c;65</td><td align="center" rowspan="1" colspan="1">65+</td><td align="center" rowspan="1" colspan="1">&#x0003c;65</td><td align="center" rowspan="1" colspan="1">65+</td></tr><tr><td align="left" rowspan="1" colspan="1">Bangladesh: Matlab</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.91</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">1.71</td><td align="center" rowspan="1" colspan="1">0.30</td><td align="center" rowspan="1" colspan="1">2.07</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.47</td><td align="center" rowspan="1" colspan="1">0.60</td><td align="center" rowspan="1" colspan="1">14.14</td><td align="center" rowspan="1" colspan="1">0.39</td><td align="center" rowspan="1" colspan="1">17.03</td><td align="center" rowspan="1" colspan="1">0.27</td><td align="center" rowspan="1" colspan="1">4.74</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">3.44</td><td align="center" rowspan="1" colspan="1">0.13</td><td align="center" rowspan="1" colspan="1">2.92</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">1.71</td></tr><tr><td align="left" rowspan="1" colspan="1">Bangladesh: Bandarban</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.35</td><td align="center" rowspan="1" colspan="1">3.75</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.33</td><td align="center" rowspan="1" colspan="1">0.15</td><td align="center" rowspan="1" colspan="1">3.15</td><td align="center" rowspan="1" colspan="1">0.16</td><td align="center" rowspan="1" colspan="1">2.69</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.35</td><td align="center" rowspan="1" colspan="1">0.13</td><td align="center" rowspan="1" colspan="1">0.73</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.71</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">2.64</td></tr><tr><td align="left" rowspan="1" colspan="1">Bangladesh: Chakaria</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">5.19</td><td align="center" rowspan="1" colspan="1">0.19</td><td align="center" rowspan="1" colspan="1">7.13</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">2.98</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">1.39</td><td align="center" rowspan="1" colspan="1">0.27</td><td align="center" rowspan="1" colspan="1">5.12</td><td align="center" rowspan="1" colspan="1">0.24</td><td align="center" rowspan="1" colspan="1">6.00</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">3.28</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">2.93</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">5.04</td><td align="center" rowspan="1" colspan="1">0.21</td><td align="center" rowspan="1" colspan="1">4.95</td></tr><tr><td align="left" rowspan="1" colspan="1">Bangladesh: AMK</td><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">2.61</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">2.26</td><td align="center" rowspan="1" colspan="1">0.40</td><td align="center" rowspan="1" colspan="1">2.31</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">1.55</td><td align="center" rowspan="1" colspan="1">0.43</td><td align="center" rowspan="1" colspan="1">14.38</td><td align="center" rowspan="1" colspan="1">0.31</td><td align="center" rowspan="1" colspan="1">16.82</td><td align="center" rowspan="1" colspan="1">0.35</td><td align="center" rowspan="1" colspan="1">6.98</td><td align="center" rowspan="1" colspan="1">0.22</td><td align="center" rowspan="1" colspan="1">5.06</td><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">5.63</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">2.65</td></tr><tr><td align="left" rowspan="1" colspan="1">Burkina Faso: Nouna</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">1.76</td><td align="center" rowspan="1" colspan="1">0.12</td><td align="center" rowspan="1" colspan="1">2.22</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">2.26</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.27</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.41</td><td align="center" rowspan="1" colspan="1">0.25</td><td align="center" rowspan="1" colspan="1">2.06</td><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">1.35</td></tr><tr><td align="left" rowspan="1" colspan="1">Burkina Faso: Ouagadougou</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.26</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.15</td><td align="center" rowspan="1" colspan="1">2.86</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">1.25</td><td align="center" rowspan="1" colspan="1">0.29</td><td align="center" rowspan="1" colspan="1">11.66</td><td align="center" rowspan="1" colspan="1">0.21</td><td align="center" rowspan="1" colspan="1">6.54</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">5.89</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.22</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">Cote d'Ivoire: Taabo</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.51</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">1.12</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">2.87</td><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">5.63</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">0.53</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">1.49</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">Ethiopia: Kilite Awlaelo</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.53</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.40</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.49</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.18</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">1.81</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">1.18</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">1.16</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.57</td><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">0.41</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.61</td></tr><tr><td align="left" rowspan="1" colspan="1">The Gambia: Farafenni</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.27</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.14</td><td align="center" rowspan="1" colspan="1">0.34</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.43</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">6.36</td><td align="center" rowspan="1" colspan="1">0.31</td><td align="center" rowspan="1" colspan="1">7.12</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">1.94</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.45</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.92</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">Ghana: Navrongo</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">1.15</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.66</td><td align="center" rowspan="1" colspan="1">0.37</td><td align="center" rowspan="1" colspan="1">2.64</td><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">0.93</td><td align="center" rowspan="1" colspan="1">0.26</td><td align="center" rowspan="1" colspan="1">2.60</td><td align="center" rowspan="1" colspan="1">0.15</td><td align="center" rowspan="1" colspan="1">1.88</td><td align="center" rowspan="1" colspan="1">0.16</td><td align="center" rowspan="1" colspan="1">2.05</td><td align="center" rowspan="1" colspan="1">0.14</td><td align="center" rowspan="1" colspan="1">1.41</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.76</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.54</td></tr><tr><td align="left" rowspan="1" colspan="1">Ghana: Dodowa</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.78</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.39</td><td align="center" rowspan="1" colspan="1">0.24</td><td align="center" rowspan="1" colspan="1">4.30</td><td align="center" rowspan="1" colspan="1">0.17</td><td align="center" rowspan="1" colspan="1">2.60</td><td align="center" rowspan="1" colspan="1">0.27</td><td align="center" rowspan="1" colspan="1">4.93</td><td align="center" rowspan="1" colspan="1">0.30</td><td align="center" rowspan="1" colspan="1">5.09</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">1.48</td><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">1.34</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.08</td></tr><tr><td align="left" rowspan="1" colspan="1">India: Ballabgarh</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">1.67</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.81</td><td align="center" rowspan="1" colspan="1">0.22</td><td align="center" rowspan="1" colspan="1">4.38</td><td align="center" rowspan="1" colspan="1">0.20</td><td align="center" rowspan="1" colspan="1">1.63</td><td align="center" rowspan="1" colspan="1">0.26</td><td align="center" rowspan="1" colspan="1">7.78</td><td align="center" rowspan="1" colspan="1">0.13</td><td align="center" rowspan="1" colspan="1">5.38</td><td align="center" rowspan="1" colspan="1">0.12</td><td align="center" rowspan="1" colspan="1">3.96</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">2.12</td><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">8.54</td><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">4.39</td></tr><tr><td align="left" rowspan="1" colspan="1">India: Vadu</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.24</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">1.64</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">1.02</td><td align="center" rowspan="1" colspan="1">0.14</td><td align="center" rowspan="1" colspan="1">4.12</td><td align="center" rowspan="1" colspan="1">0.18</td><td align="center" rowspan="1" colspan="1">5.43</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">1.41</td><td align="center" rowspan="1" colspan="1">0.15</td><td align="center" rowspan="1" colspan="1">1.17</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">1.19</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.74</td></tr><tr><td align="left" rowspan="1" colspan="1">Kenya: Kilifi</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">1.91</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.73</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.92</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.29</td><td align="center" rowspan="1" colspan="1">0.29</td><td align="center" rowspan="1" colspan="1">7.00</td><td align="center" rowspan="1" colspan="1">0.24</td><td align="center" rowspan="1" colspan="1">7.16</td><td align="center" rowspan="1" colspan="1">0.15</td><td align="center" rowspan="1" colspan="1">5.94</td><td align="center" rowspan="1" colspan="1">0.14</td><td align="center" rowspan="1" colspan="1">3.84</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">1.92</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">1.25</td></tr><tr><td align="left" rowspan="1" colspan="1">Kenya: Kisumu</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">2.04</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.56</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.85</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.49</td><td align="center" rowspan="1" colspan="1">0.13</td><td align="center" rowspan="1" colspan="1">1.27</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">2.18</td><td align="center" rowspan="1" colspan="1">0.39</td><td align="center" rowspan="1" colspan="1">8.84</td><td align="center" rowspan="1" colspan="1">0.38</td><td align="center" rowspan="1" colspan="1">7.97</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">1.89</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">1.64</td></tr><tr><td align="left" rowspan="1" colspan="1">Kenya: Nairobi</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.44</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">1.16</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">1.58</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.50</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">1.81</td><td align="center" rowspan="1" colspan="1">0.27</td><td align="center" rowspan="1" colspan="1">5.61</td><td align="center" rowspan="1" colspan="1">0.27</td><td align="center" rowspan="1" colspan="1">6.25</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">1.23</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">1.39</td></tr><tr><td align="left" rowspan="1" colspan="1">Malawi: Karonga</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">1.97</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">0.58</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">1.06</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.83</td><td align="center" rowspan="1" colspan="1">0.22</td><td align="center" rowspan="1" colspan="1">6.41</td><td align="center" rowspan="1" colspan="1">0.23</td><td align="center" rowspan="1" colspan="1">7.12</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">2.80</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">2.76</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1.47</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">1.72</td></tr><tr><td align="left" rowspan="1" colspan="1">Senegal: Niakhar</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.60</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.38</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1.13</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.59</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.14</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.20</td><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">5.15</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">4.03</td></tr><tr><td align="left" rowspan="1" colspan="1">South Africa: Agincourt</td><td align="center" rowspan="1" colspan="1">0.14</td><td align="center" rowspan="1" colspan="1">1.83</td><td align="center" rowspan="1" colspan="1">0.12</td><td align="center" rowspan="1" colspan="1">6.57</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.29</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.60</td><td align="center" rowspan="1" colspan="1">0.25</td><td align="center" rowspan="1" colspan="1">2.53</td><td align="center" rowspan="1" colspan="1">0.24</td><td align="center" rowspan="1" colspan="1">11.42</td><td align="center" rowspan="1" colspan="1">0.17</td><td align="center" rowspan="1" colspan="1">2.15</td><td align="center" rowspan="1" colspan="1">0.19</td><td align="center" rowspan="1" colspan="1">8.77</td><td align="center" rowspan="1" colspan="1">0.22</td><td align="center" rowspan="1" colspan="1">1.69</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">4.71</td></tr><tr><td align="left" rowspan="1" colspan="1">South Africa: Africa Centre</td><td align="center" rowspan="1" colspan="1">0.18</td><td align="center" rowspan="1" colspan="1">4.22</td><td align="center" rowspan="1" colspan="1">0.25</td><td align="center" rowspan="1" colspan="1">4.20</td><td align="center" rowspan="1" colspan="1">0.12</td><td align="center" rowspan="1" colspan="1">0.75</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.34</td><td align="center" rowspan="1" colspan="1">0.23</td><td align="center" rowspan="1" colspan="1">6.36</td><td align="center" rowspan="1" colspan="1">0.26</td><td align="center" rowspan="1" colspan="1">6.58</td><td align="center" rowspan="1" colspan="1">0.33</td><td align="center" rowspan="1" colspan="1">8.82</td><td align="center" rowspan="1" colspan="1">0.47</td><td align="center" rowspan="1" colspan="1">9.79</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">4.14</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">4.09</td></tr><tr><td align="left" rowspan="1" colspan="1">Vietnam: FilaBavi</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.36</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.28</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">1.40</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.42</td><td align="center" rowspan="1" colspan="1">0.23</td><td align="center" rowspan="1" colspan="1">8.27</td><td align="center" rowspan="1" colspan="1">0.12</td><td align="center" rowspan="1" colspan="1">10.04</td><td align="center" rowspan="1" colspan="1">0.32</td><td align="center" rowspan="1" colspan="1">5.00</td><td align="center" rowspan="1" colspan="1">0.14</td><td align="center" rowspan="1" colspan="1">2.83</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">1.11</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0.69</td></tr></tbody></table></table-wrap><p>Because some of the sites reporting the highest rates of NCD mortality were also those with the greatest burden of HIV/AIDS-related mortality (as explored from the same dataset in a separate paper (<xref rid="CIT0035" ref-type="bibr">35</xref>)), in <xref ref-type="fig" rid="F0004">Fig. 4</xref> we present, separately for the 15&#x02013;64 and over-65 year age groups, the correlations between NCD and HIV/AIDS mortality rates, showing points for each site and least absolute deviations regression lines for both age groups. In both age categories, correlations are significant (p&#x0003c;0.005; R²=0.40 and 0.37, respectively) and suggest that in high HIV settings, around half of the mortality attributed to NCDs may well be associated with HIV. This was also reflected in falling NCD mortality rates among younger adults at most African sites between the 2000&#x02013;05 and 2006&#x02013;12 periods (<xref ref-type="table" rid="T0001">Table 1</xref>), as HIV/AIDS-related mortality reduced sharply (35).</p><fig id="F0004" position="float"><label>Fig. 4</label><caption><p>Age&#x02013;sex&#x02013;time standardised NCD mortality for 15&#x02013;64 and over-65 year age groups in relation to age&#x02013;sex&#x02013;time standardised HIV/AIDS-related mortality in the same populations, all per 1,000 person-years.</p></caption><graphic xlink:href="GHA-7-25365-g004"/></fig></sec><sec sec-type="discussion" id="S0004"><title>Discussion</title><p>This study reports NCD mortality patterns for a large number of adults across a wide range of settings in Africa and Asia. Being part of the overall INDEPTH cause-specific mortality study (<xref rid="CIT0008" ref-type="bibr">8</xref>), these rates for NCDs constitute a component of complete mortality surveillance in each site, as distinct from studies only considering NCDs. The age&#x02013;sex&#x02013;time standardised results presented here enable comparisons to be made between sites allowing for the dynamics of age&#x02013;sex population profiles in different places.</p><p>Although proportions of NCD mortality were generally higher in Asian sites (<xref ref-type="fig" rid="F0001">Fig. 1</xref>), rates of NCD mortality were more consistent across African and Asian sites. This observation is driven less by NCD-specific factors than by demographic transition. Many populations in Asia are experiencing marked increases in mid-life survival, and consequently life expectancy, which lead to historically low levels of crude mortality. Thus, according to UN Population estimates, many Asian countries currently record much lower crude mortality rates than their European and American counterparts, having reduced infection-related mortality but not having yet accumulated substantial population proportions of older people (<xref rid="CIT0036" ref-type="bibr">36</xref>). As populations gain larger numbers of older people, there will be increasing numbers of NCD-related deaths, even if rates remain the same.</p><p>Because not all sites operated population surveillance for the same time period, we have limited evidence on the time trends in NCD mortality rates (<xref ref-type="table" rid="T0001">Table 1</xref>). However, in the sites where trends could be seen, there was little clear evidence of NCD mortality rates increasing over time. As expected for NCD mortality, rates increased very markedly with age; however, the relatively small population proportions of people over 65 years in most of these sites means that the high mortality rates observed in that age group do not yet translate into large numbers of deaths.</p><p> <xref ref-type="fig" rid="F0002">Figure 2</xref> shows that cancers and cardiovascular diseases were the major components of NCD mortality in almost all sites. Similar proportions of most causes were seen across most sites; a notable exception was the lack of cancer-related mortality at the Nouna, Burkina Faso, site, which is possibly a consequence of items missing from the VA instrument in use there. Some cancer deaths at Nouna were probably misclassified into the relatively high rates of respiratory and abdominal NCDs at that site. This is an example of some of the difficulties that may occur as a result of harmonising VA data collected at many sites, over different time periods, using a variety of antecedents to the WHO 2012 standard (<xref rid="CIT0037" ref-type="bibr">37</xref>). This is discussed further in the introductory paper (<xref rid="CIT0008" ref-type="bibr">8</xref>) but there are no simple solutions.</p><p> <xref ref-type="table" rid="T0002">Table 2</xref> shows a breakdown of cancer-related mortality rates. As one of few long-established national cancer registries in sub-Saharan Africa, the Gambian cancer registry has reported incidence rates for various cancers over more than 20 years, partly with the aim of tracking possible reductions in liver cancer incidence following the pilot introduction of hepatitis B vaccination in the 1980s (<xref rid="CIT0038" ref-type="bibr">38</xref>). Although incidence and mortality rates are not directly comparable, there are similar age&#x02013;sex patterns in the Gambian incidence rates and the mortality rates reported here, particularly for several of the West African sites. In eastern and southern Africa, cancer mortality rates may also be influenced by high HIV rates (<xref rid="CIT0039" ref-type="bibr">39</xref>). Respiratory cancer mortality as reported here showed higher rates for men in most sites over both age groups, which probably reflects gendered norms around smoking as a risk factor. Appreciable rates of &#x02018;other neoplasms&#x02019; were recorded from many sites, which may partly reflect a lack of specific information about cancers recalled in VA interviews, or a lack of specific items in some historic VA instruments.</p><p> <xref ref-type="table" rid="T0003">Table 3</xref> shows, in the same format as <xref ref-type="table" rid="T0002">Table 2</xref>, mortality rates for diabetes, cardiovascular disease, and chronic obstructive pulmonary disease. These categories were major contributors to NCD mortality across many sites, although the relative magnitude of these categories of mortality varied between sites. Some sites were using VA instruments designed before NCDs were considered to be major causes of death in the populations concerned. It is likely that the quality of NCD cause assignment achievable by InterVA-4 will increase as primary VA data collected according to the complete WHO 2012 VA specification become more widely available. Nevertheless, the overall consistency observed across these varied sites in many respects suggests that historical inter-site variations did not have a major influence on the overall pattern of NCD mortality findings.</p><p>The importance of separating NCD mortality into premature and less-avoidable groupings by age is clearly illustrated in <xref ref-type="fig" rid="F0003">Fig. 3</xref>. Due consideration must be given to both NCD mortality rates as well as numbers of NCD deaths in considering how disease burdens may be changing. Here, we have somewhat arbitrarily taken 65 years as the cut-off for considering premature mortality, because that fitted with the WHO 2012 VA age groups used in this dataset. In addition, many of the countries represented here have life expectancies at birth in the range of 60 to 70 years, and so 65 years may represent a reasonable dichotomy in terms of &#x02018;prematurity&#x02019;. In the global context, 70 years might be a more reasonable cut-off (<xref rid="CIT0040" ref-type="bibr">40</xref>). As yet, none of the countries represented here has a top-heavy population pyramid, with the consequence that there is a reasonably even balance in proportions of NCD deaths occurring before and after 65 years of age, despite high NCD rates among older people. However, it is to be expected, as time passes, that there will be a gradual rightward shift in the pattern shown in <xref ref-type="fig" rid="F0003">Fig. 3</xref>, unless increased NCD risk factor exposures at younger ages also increase NCD mortality rates among the under-65 age group. Nevertheless, the appreciable number of NCD deaths recorded among the under-65 age group must also reflect a considerable burden of chronic disease morbidity in that otherwise productive age group. To the contrary, modelling results suggest that it could be possible to achieve significant global reductions in premature (under-70 years) NCD deaths if relevant risk factors were appropriately managed (<xref rid="CIT0040" ref-type="bibr">40</xref>); however, the extent to which those risk factors are likely to be effectively managed in the countries reported here remains an important question.</p><p>The contribution of diabetes to the overall picture of NCD mortality in these populations was relatively small compared to cancers and cardiovascular diseases. Whether this reflects situations where risk factors are not yet translating into mortality rates, or whether VA methods led to some misclassification of diabetes mortality into, for example, cardiovascular deaths is uncertain.</p><p>The correlations between HIV-related mortality and NCD mortality across the whole adult age range, as shown in <xref ref-type="fig" rid="F0004">Fig. 4</xref>, show how important it is to take a holistic view of cause-specific mortality across entire populations. The increased NCD mortality rates in higher HIV mortality populations here are consistent with previous work that analysed causes of death by HIV status (<xref rid="CIT0039" ref-type="bibr">39</xref>). This is clearly an important factor to bear in mind when comparing patterns of NCD mortality across sites with varying HIV epidemiology. The ICD-10 system tries to classify almost all deaths in HIV-positive people under the B2 categories related to HIV (<xref rid="CIT0041" ref-type="bibr">41</xref>). In reality, however, particularly if HIV status is not known and for final illnesses that are not indicative of HIV/AIDS, it is very likely (by any means of cause of death assignment) that a proportion of deaths among HIV-positive people will be attributed to NCDs. This is likely to be an increasing issue as higher proportions of HIV-positive people receive effective antiretroviral treatment.</p></sec><sec sec-type="conclusions" id="S0005"><title>Conclusions</title><p>Against a background of considerable concern about the magnitude of the global NCD mortality burden, these population-based findings provide useful geographic examples of the state of adult NCD mortality across a range of locations in Africa and Asia, standardised for ease of comparison. The importance of distinguishing between premature NCD mortality and the inevitably high rates of NCDs that occur late in life is clear from these data. These findings also provide a useful baseline from which to track future trends in NCD mortality, using the same standardised methods. It is also evident that, in areas of high HIV endemicity, NCD mortality rates have to be seen as possibly including some HIV-related deaths, in cases where no clear evidence of HIV status was available.</p></sec>

Cause-specific mortality and socioeconomic status in Chakaria, Bangladesh

<sec id="st1"><title>Background</title><p>Bangladesh has achieved remarkable gains in health indicators during the last four decades despite low levels of economic development. However, the persistence of inequities remains disturbing. This success was also accompanied by health and demographic transitions, which in turn brings new challenges for a nation that has yet to come to terms with pre-transition health challenges. It is therefore important to understand the causes of death and their relationship with socioeconomic status (SES).</p></sec><sec id="st2"><title>Objective</title><p>The paper aims to assess the causes of death by SES based on surveillance data from a rural area of Bangladesh, in order to understand the situation and inform policy makers and programme leaders.</p></sec><sec id="st3"><title>Design</title><p>We analysed population-based mortality data collected from the Chakaria Health and Demographic Surveillance System in Bangladesh. The causes of death were determined by using a Bayesian-based programme for interpreting verbal autopsy findings (InterVA-4). The data included 1,391 deaths in 217,167 person-years of observation between 2010 and 2012. The wealth index constructed using household assets was used to assess the SES, and disease burdens were compared among the wealth quintiles.</p></sec><sec id="st4"><title>Results</title><p>Analysing cause of death (CoD) revealed that non-communicable diseases (NCDs) were the leading causes of deaths (37%), followed by communicable diseases (CDs) (22%), perinatal and neonatal conditions (11%), and injury and accidents (6%); the cause of remaining 24% of deaths could not be determined. Age-specific mortality showed premature birth, respiratory infections, and drowning were the dominant causes of death for childhood mortality (0&#x02013;14 years), which was inversely associated with SES (<italic>p</italic>&#x0003c;0.04). For adult and the elderly (15 years and older), NCDs were the leading cause of death (51%), followed by CDs (23%). For adult and the elderly, NCDs concentrated among the population from higher SES groups (<italic>p</italic>&#x0003c;0.005), and CDs among the lower SES groups (<italic>p</italic>&#x0003c;0.001).</p></sec><sec id="st5"><title>Conclusions</title><p>Epidemiologic transition is taking place with a shift from the dominance of CDs to NCDs. SES inequity in mortality still persists &#x02013; the poor suffer from CDs in all age groups, whereas those better off suffer more from NCDs than CDs. Policy makers thus need to consider the social distribution of diseases before developing any public health action targeted towards reducing mortality and the extent of disease burden in an equitable manner.</p></sec>

<contrib contrib-type="author"><name><surname>Hanifi</surname><given-names>Syed M. A.</given-names></name><xref ref-type="aff" rid="AF0001">1</xref><xref ref-type="aff" rid="AF0002">2</xref><xref ref-type="corresp" rid="cor1">*</xref></contrib><contrib contrib-type="author"><name><surname>Mahmood</surname><given-names>Shehrin S.</given-names></name><xref ref-type="aff" rid="AF0001">1</xref><xref ref-type="aff" rid="AF0002">2</xref></contrib><contrib contrib-type="author"><name><surname>Bhuiya</surname><given-names>Abbas</given-names></name><xref ref-type="aff" rid="AF0001">1</xref><xref ref-type="aff" rid="AF0002">2</xref></contrib>

Global Health Action

<p>Bangladesh has recently been applauded as an &#x02018;exceptional health performer&#x02019; for its achievements in health outcomes during the last four decades despite widespread poverty (<xref rid="CIT0001" ref-type="bibr">1</xref>&#x02013;<xref rid="CIT0006" ref-type="bibr">6</xref>). Increased life expectancy at birth (70 years in 2012, from 47 years in 1971) (<xref rid="CIT0007" ref-type="bibr">7</xref>), reduced total fertility (2.2 in 2012, from 7 in 1971) (<xref rid="CIT0007" ref-type="bibr">7</xref>), and increased infant and child survival are some of the highlights of this achievement (<xref rid="CIT0001" ref-type="bibr">1</xref>&#x02013;<xref rid="CIT0005" ref-type="bibr">5</xref>). However, socioeconomic and geographic inequities in health outcomes, though reduced for some indicators, still persist and continue to be a development challenge (<xref rid="CIT0008" ref-type="bibr">8</xref>&#x02013;<xref rid="CIT0010" ref-type="bibr">10</xref>). Despite the fact that sex-based inequities in life expectancy (higher for male than female) have reversed since the late 1980s, the economically better off still expect to live 71 years at birth, whereas the expectation is only 63 years for the worse off (<xref rid="CIT0011" ref-type="bibr">11</xref>). The under-five mortality rate for the poor is almost two times the rate of the rich, and the overall disease burden is also higher for the poor than the rich (<xref rid="CIT0012" ref-type="bibr">12</xref>). Socioeconomic inequity also exists in utilisation of health care services (<xref rid="CIT0012" ref-type="bibr">12</xref>).</p><p>The country has started to experience an epidemiological transition from burden of acute infectious and nutritional deficiency diseases to chronic non-communicable diseases (NCDs) (<xref rid="CIT0005" ref-type="bibr">5</xref>). Bangladesh is on the way to having NCDs as leading causes of death like many other low- and middle-income countries (<xref rid="CIT0013" ref-type="bibr">13</xref>). With persistent poverty and socioeconomic inequity in disease burdens, health seeking behaviour, and health outcomes in Bangladesh, the emerging epidemiological transition is expected to hit different groups in various ways. Although a number of studies have reported differentials in cause-specific mortality among countries at various income levels, studies on socioeconomic differentials within countries have been rare (<xref rid="CIT0013" ref-type="bibr">13</xref>). The current paper hence aims to examine socioeconomic differentials in causes of death using systematically collected data from a rural area of Bangladesh.</p><sec sec-type="materials|methods" id="S0002"><title>Materials and methods</title><sec id="S0002-S20001"><title>Study area</title><p>The study was carried out in Chakaria <italic>Upazila</italic> (a sub-district) situated in the southeast coastal area of Bangladesh where ICDDR,B has been running a health and demographic surveillance system (HDSS) since 1999. The purpose of the surveillance has been to monitor health outcomes and health care service utilisation with equity focus and to generate relevant health, demographic, and socioeconomic information for policy formation, programme design, and research. The surveillance currently covers 80,166 residents living in 15,000 households (<xref rid="CIT0014" ref-type="bibr">14</xref>). The population density is 782 individuals km^&#x02212;2. The population is comprised mainly of Muslims (93%), and a small number of Hindus (5%) and Buddhists (2%). About 72% of the households consist of nuclear families, and the remainder are extended and joint families (<xref rid="CIT0015" ref-type="bibr">15</xref>).</p><p>The main economic activities in the area have been agriculture, forestry, and sea fishing. Thirty percent of the households are landless and about half of the households depend on income from menial labour. The adult literacy rate is 64% which is higher than 58%, the national average of Bangladesh (<xref rid="CIT0011" ref-type="bibr">11</xref>). Ten percent of the households have a television, 61% of households have a cell phone, and about one-third of the households have electricity. (<xref rid="CIT0011" ref-type="bibr">11</xref>).</p><p>A transition in fertility level is also taking place in the area as in rest of the country. The total fertility rate per woman declined to 2.9 in 2012 from 5.1 in 1999. The life expectancy at birth was 68 years in 2012 compared to 65 years in 1999, with females living longer than males by about 2 years. The under-five mortality rate in the area was 56/1,000 livebirths in 2012 compared to 69/1,000 livebirths in 1999 (<xref rid="CIT0014" ref-type="bibr">14</xref>, <xref rid="CIT0015" ref-type="bibr">15</xref>). However, socioeconomic inequity in under-five mortality, although reducing, is still persistent. Undernutrition among children less than 2 years of age is high with about one-fourth undernourished (<xref rid="CIT0011" ref-type="bibr">11</xref>).</p><p>The health care delivery system in the HDSS area comprises services by public, private, and non-governmental organisations. Private care providers are dominated by informally trained providers practicing by using modern drugs (<xref rid="CIT0016" ref-type="bibr">16</xref>, <xref rid="CIT0017" ref-type="bibr">17</xref>). However, the number of available accredited facilities still falls short of meeting the health care needs of the population in the area. The poor seek care from public health facilities and informal health care providers, and the better off from the private clinics mostly attended by physicians from public facilities after hours. Although the use of antenatal, postnatal, and skilled attendants for delivery (around 80% of the deliveries take place at home) have been on the rise, substantial gaps remain in utilisation between poor and better off. For the lowest quintile, about 8% of deliveries take place at health facilities, whereas for the highest quintile it is 40% (<xref rid="CIT0014" ref-type="bibr">14</xref>).</p><p>Among the 19 millennium development goal (MDG) indicators, which are related to socioeconomic, demographic, health, and water sanitation status, Chakaria lags behind the national level for some of the water sanitation and socioeconomic indicators. For health, the status indicators in the study area are almost similar to that of the national level (<xref rid="CIT0011" ref-type="bibr">11</xref>).</p></sec><sec id="S0002-S20002"><title>Data collection</title><p>We used data from the Chakaria HDSS which is a member of INDEPTH network (<xref rid="CIT0018" ref-type="bibr">18</xref>). A team of trained surveillance workers collected information about the circumstances of a death, including signs and symptoms leading to death, and previous medical history during quarterly household visits, from the immediate next of kin of the deceased. The data used for the current analysis covered the period from 2010 to 2012, which has been subject to INDEPTH data validation. VA data for the year 2010 were collected using earlier standards, which later were converted to the WHO 2012 standard (<xref rid="CIT0019" ref-type="bibr">19</xref>). For 2011 and 2012, data were collected using interVA data collection tools. INDEPTH verbal autopsy questionnaires were used to collect causes of death data and its VA algorithm was used to ascertain causes of death. The cause of death data sets are stored and available at the INDEPTH data repository (<xref rid="CIT0020" ref-type="bibr">20</xref>). Socioeconomic data were linked from the Chakaria HDSS database.</p><table-wrap id="T0001" position="float"><label>Table 1</label><caption><p>Mortality rate per 1,000 person-years observation, 2010&#x02013;2012</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1"/><th align="center" colspan="3" rowspan="1">Person-years</th><th align="center" colspan="3" rowspan="1">Number of deaths</th><th align="center" colspan="3" rowspan="1">Death rates/1,000 person-years</th></tr><tr><th align="left" rowspan="1" colspan="1"/><th colspan="3" rowspan="1"> <hr/> </th><th colspan="3" rowspan="1"> <hr/> </th><th colspan="3" rowspan="1"> <hr/> </th></tr><tr><th align="left" rowspan="1" colspan="1">Age group</th><th align="center" rowspan="1" colspan="1">Male</th><th align="center" rowspan="1" colspan="1">Female</th><th align="center" rowspan="1" colspan="1">Both</th><th align="center" rowspan="1" colspan="1">Male</th><th align="center" rowspan="1" colspan="1">Female</th><th align="center" rowspan="1" colspan="1">Both</th><th align="center" rowspan="1" colspan="1">Male</th><th align="center" rowspan="1" colspan="1">Female</th><th align="center" rowspan="1" colspan="1">Both</th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">&#x0003c;28 days</td><td align="center" rowspan="1" colspan="1">235</td><td align="center" rowspan="1" colspan="1">217</td><td align="center" rowspan="1" colspan="1">452</td><td align="center" rowspan="1" colspan="1">119</td><td align="center" rowspan="1" colspan="1">88</td><td align="center" rowspan="1" colspan="1">207</td><td align="center" rowspan="1" colspan="1">506.4</td><td align="center" rowspan="1" colspan="1">405.5</td><td align="center" rowspan="1" colspan="1">458.0</td></tr><tr><td align="left" rowspan="1" colspan="1">1&#x02013;11 months</td><td align="center" rowspan="1" colspan="1">2,680</td><td align="center" rowspan="1" colspan="1">2,504</td><td align="center" rowspan="1" colspan="1">5,184</td><td align="center" rowspan="1" colspan="1">31</td><td align="center" rowspan="1" colspan="1">55</td><td align="center" rowspan="1" colspan="1">86</td><td align="center" rowspan="1" colspan="1">11.6</td><td align="center" rowspan="1" colspan="1">22.0</td><td align="center" rowspan="1" colspan="1">16.6</td></tr><tr><td align="left" rowspan="1" colspan="1">1&#x02013;4 years</td><td align="center" rowspan="1" colspan="1">11,185</td><td align="center" rowspan="1" colspan="1">10,807</td><td align="center" rowspan="1" colspan="1">21,992</td><td align="center" rowspan="1" colspan="1">44</td><td align="center" rowspan="1" colspan="1">43</td><td align="center" rowspan="1" colspan="1">87</td><td align="center" rowspan="1" colspan="1">3.9</td><td align="center" rowspan="1" colspan="1">4.0</td><td align="center" rowspan="1" colspan="1">4.0</td></tr><tr><td align="left" rowspan="1" colspan="1">5&#x02013;14 years</td><td align="center" rowspan="1" colspan="1">31,076</td><td align="center" rowspan="1" colspan="1">29,875</td><td align="center" rowspan="1" colspan="1">60,951</td><td align="center" rowspan="1" colspan="1">34</td><td align="center" rowspan="1" colspan="1">27</td><td align="center" rowspan="1" colspan="1">61</td><td align="center" rowspan="1" colspan="1">1.1</td><td align="center" rowspan="1" colspan="1">0.9</td><td align="center" rowspan="1" colspan="1">1.0</td></tr><tr><td align="left" rowspan="1" colspan="1">15&#x02013;49 years</td><td align="center" rowspan="1" colspan="1">51,244</td><td align="center" rowspan="1" colspan="1">52,853</td><td align="center" rowspan="1" colspan="1">104,097</td><td align="center" rowspan="1" colspan="1">103</td><td align="center" rowspan="1" colspan="1">93</td><td align="center" rowspan="1" colspan="1">196</td><td align="center" rowspan="1" colspan="1">2.0</td><td align="center" rowspan="1" colspan="1">1.8</td><td align="center" rowspan="1" colspan="1">1.9</td></tr><tr><td align="left" rowspan="1" colspan="1">50&#x02013;64 years</td><td align="center" rowspan="1" colspan="1">8,635</td><td align="center" rowspan="1" colspan="1">7,599</td><td align="center" rowspan="1" colspan="1">16,234</td><td align="center" rowspan="1" colspan="1">107</td><td align="center" rowspan="1" colspan="1">100</td><td align="center" rowspan="1" colspan="1">207</td><td align="center" rowspan="1" colspan="1">12.4</td><td align="center" rowspan="1" colspan="1">13.2</td><td align="center" rowspan="1" colspan="1">12.8</td></tr><tr><td align="left" rowspan="1" colspan="1">65+ years</td><td align="center" rowspan="1" colspan="1">4,468</td><td align="center" rowspan="1" colspan="1">3,789</td><td align="center" rowspan="1" colspan="1">8,257</td><td align="center" rowspan="1" colspan="1">286</td><td align="center" rowspan="1" colspan="1">261</td><td align="center" rowspan="1" colspan="1">547</td><td align="center" rowspan="1" colspan="1">64.0</td><td align="center" rowspan="1" colspan="1">68.9</td><td align="center" rowspan="1" colspan="1">66.2</td></tr><tr><td align="left" rowspan="1" colspan="1">Total</td><td align="center" rowspan="1" colspan="1">109,523</td><td align="center" rowspan="1" colspan="1">107,644</td><td align="center" rowspan="1" colspan="1">217,167</td><td align="center" rowspan="1" colspan="1">724</td><td align="center" rowspan="1" colspan="1">667</td><td align="center" rowspan="1" colspan="1">1,391</td><td align="center" rowspan="1" colspan="1">6.6</td><td align="center" rowspan="1" colspan="1">6.2</td><td align="center" rowspan="1" colspan="1">6.4</td></tr></tbody></table></table-wrap></sec><sec id="S0002-S20003"><title>Definition of variables</title><sec><title>Socioeconomic status</title><p>The wealth index, constructed from household asset data using principal component analysis was used to categorise households into socioeconomic groups. Households were categorised into five socioeconomic quintiles: lowest, second, middle, fourth, and highest. Information on asset ownership was available for 1,328 death cases and thus the analysis of socioeconomic status (SES) inequities is based on these 1,328 cases.</p></sec><sec><title>CDs and NCDs</title><p>The communicable diseases (CDs) included sepsis-non-obstetric, acute respiratory infection including pneumonia, diarrheal diseases, malaria, meningitis and encephalitis, and pulmonary tuberculosis (TB). The NCDs included diabetes mellitus, acute cardiac disease, stroke, chronic obstructive pulmonary disease, asthma, acute abdomen, liver cirrhosis, renal failure, epilepsy, and neoplasm.</p></sec></sec><sec id="S0002-S20004"><title>Data analysis</title><p>Data were transferred into input format for the InterVA-4 Bayesian model for assigning cause of death (CoD) (<xref rid="CIT0021" ref-type="bibr">21</xref>). For each death case, the model gives up to three possible causes of deaths or an indeterminate result. For death cases, where symptoms were contradictory or mutually inconsistent, the cause of death was categorised as indeterminate. For the remaining cases, one to three likely causes were assigned, and if the sum of their likelihoods was less than 100%, the residual component was then assigned as being indeterminate. Stillbirths were excluded from these cause of death analyses. In total 1,391 deaths were registered during 2010&#x02013;2012 and verbal autopsy completed for 1,328 deaths. For 63 of these deaths not enough information was available to assign cause of death.</p></sec></sec><sec sec-type="results" id="S0003"><title>Results</title><sec id="S0003-S20001"><title>Crude mortality</title><p>The analysis presented in this paper is based on follow-up of 217,167 person-years during 2010&#x02013;2012. Forty-one percent of the population observed were children aged less than 15 years, 48% were adult (aged 15&#x02013;49 years), and 11% were elderly (aged 50 years and older). Of the 1,391 deaths, 31% were children, 14% were adults, and 55% were elderly. The crude death rate was 6.4 per 1,000 person-years and the rate was higher for males (6.6 per 1,000 person-years) than for females (6.2 per 1,000 person-years). The mortality rate for age 15 years and older was 7.39 per 1,000 person-years (95% CI: 6.93&#x02013;7.87), and the rate was 4.97 per 1,000 person-years (95% CI: 4.51&#x02013;5.45) among the children under 15 years of age (<xref ref-type="table" rid="T0001">Table 1</xref>).</p></sec><sec id="S0003-S20002"><title>Cause-specific deaths</title><p><xref ref-type="table" rid="T0002">Tables 2</xref> and <xref ref-type="table" rid="T0003">3</xref> present cause-specific death rates by age and sex. Results showed that the highest mortality rates were among the youngest and the oldest age groups. Among neonates (&#x0003c;28 days), prematurity (23%) was the major cause of death, followed closely by birth asphyxia (22%). Death due to prematurity was higher for males (144 per 1,000 person-years) compared to females (62 per 1,000 person-years).</p><table-wrap id="T0002" position="float"><label>Table 2</label><caption><p>Cause-specific mortality rates per 1,000 person-years by age groups (Male)</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1">Causes</th><th align="center" rowspan="1" colspan="1">&#x0003c;28 days</th><th align="center" rowspan="1" colspan="1">1&#x02013;11 months</th><th align="center" rowspan="1" colspan="1">1&#x02013;4 years</th><th align="center" rowspan="1" colspan="1">5&#x02013;14 years</th><th align="center" rowspan="1" colspan="1">15&#x02013;49 years</th><th align="center" rowspan="1" colspan="1">50&#x02013;64 years</th><th align="center" rowspan="1" colspan="1">65+ years</th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">01.01 Sepsis (non-obstetric)</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.18</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.07</td></tr><tr><td align="left" rowspan="1" colspan="1">01.02 Acute respiratory infection including pneumonia</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">2.20</td><td align="center" rowspan="1" colspan="1">0.78</td><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.75</td><td align="center" rowspan="1" colspan="1">3.93</td></tr><tr><td align="left" rowspan="1" colspan="1">01.03 HIV/AIDS-related death</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">01.04 Diarrhoeal diseases</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">1.50</td><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">01.05 Malaria</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.06</td></tr><tr><td align="left" rowspan="1" colspan="1">01.07 Meningitis and encephalitis</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.33</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">01.09 Pulmonary tuberculosis</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.19</td><td align="center" rowspan="1" colspan="1">2.35</td><td align="center" rowspan="1" colspan="1">10.68</td></tr><tr><td align="left" rowspan="1" colspan="1">01.10 Pertussis</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.19</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">01.11 Haemorrhagic fever</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">01.99 Other and unspecified infectious diseases</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.13</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">1.75</td></tr><tr><td align="left" rowspan="1" colspan="1">02.02 Digestive neoplasms</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">0.45</td><td align="center" rowspan="1" colspan="1">1.82</td></tr><tr><td align="left" rowspan="1" colspan="1">02.03 Respiratory neoplasms</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.12</td><td align="center" rowspan="1" colspan="1">0.40</td><td align="center" rowspan="1" colspan="1">4.47</td></tr><tr><td align="left" rowspan="1" colspan="1">02.05 &#x00026; 02.06 Reproductive neoplasms</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.69</td></tr><tr><td align="left" rowspan="1" colspan="1">02.99 Other and unspecified neoplasms</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.19</td><td align="center" rowspan="1" colspan="1">0.99</td><td align="center" rowspan="1" colspan="1">2.55</td></tr><tr><td align="left" rowspan="1" colspan="1">03.01 Severe anaemia</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.63</td></tr><tr><td align="left" rowspan="1" colspan="1">03.02 Severe malnutrition</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.37</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.13</td></tr><tr><td align="left" rowspan="1" colspan="1">03.03 Diabetes mellitus</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.34</td><td align="center" rowspan="1" colspan="1">5.19</td></tr><tr><td align="left" rowspan="1" colspan="1">04.01 Acute cardiac disease</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">2.98</td></tr><tr><td align="left" rowspan="1" colspan="1">04.02 Stroke</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">1.21</td><td align="center" rowspan="1" colspan="1">5.12</td></tr><tr><td align="left" rowspan="1" colspan="1">04.99 Other and unspecified cardiac diseases</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.71</td><td align="center" rowspan="1" colspan="1">3.28</td></tr><tr><td align="left" rowspan="1" colspan="1">05.01 Chronic obstructive pulmonary diseases</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.77</td><td align="center" rowspan="1" colspan="1">5.04</td></tr><tr><td align="left" rowspan="1" colspan="1">05.02 Asthma</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.44</td><td align="center" rowspan="1" colspan="1">1.06</td></tr><tr><td align="left" rowspan="1" colspan="1">06.01 Acute abdomen</td><td align="center" rowspan="1" colspan="1">3.15</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">1.17</td></tr><tr><td align="left" rowspan="1" colspan="1">06.02 Liver cirrhosis</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.44</td><td align="center" rowspan="1" colspan="1">0.74</td></tr><tr><td align="left" rowspan="1" colspan="1">07.01 Renal failure</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.17</td><td align="center" rowspan="1" colspan="1">1.25</td></tr><tr><td align="left" rowspan="1" colspan="1">08.01 Epilepsy</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.66</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.38</td></tr><tr><td align="left" rowspan="1" colspan="1">10.01 Prematurity</td><td align="center" rowspan="1" colspan="1">144.26</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">10.02 Birth asphyxia</td><td align="center" rowspan="1" colspan="1">98.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">10.03 Neonatal pneumonia</td><td align="center" rowspan="1" colspan="1">42.04</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">10.04 Neonatal sepsis</td><td align="center" rowspan="1" colspan="1">6.04</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">10.06 Congenital malformation</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.81</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">10.99 Other and unspecified neonatal CoD</td><td align="center" rowspan="1" colspan="1">74.26</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">12.01 Road traffic accident</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.15</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">12.03 Accidental fall</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">0.17</td><td align="center" rowspan="1" colspan="1">0.22</td></tr><tr><td align="left" rowspan="1" colspan="1">12.04 Accidental drowning and submersion</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.37</td><td align="center" rowspan="1" colspan="1">1.52</td><td align="center" rowspan="1" colspan="1">0.23</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">12.06 Contact with venomous plant/animal</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.20</td></tr><tr><td align="left" rowspan="1" colspan="1">12.07 Accidental poisoning &#x00026; noxious substances</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">12.08 Intentional self-harm</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">0.16</td></tr><tr><td align="left" rowspan="1" colspan="1">12.09 Assault</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">12.99 Other and unspecified external CoD</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">98 Other and unspecified NCD</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.19</td><td align="center" rowspan="1" colspan="1">0.18</td></tr><tr><td align="left" rowspan="1" colspan="1">99 Indeterminate</td><td align="center" rowspan="1" colspan="1">130.13</td><td align="center" rowspan="1" colspan="1">3.10</td><td align="center" rowspan="1" colspan="1">0.40</td><td align="center" rowspan="1" colspan="1">0.25</td><td align="center" rowspan="1" colspan="1">0.58</td><td align="center" rowspan="1" colspan="1">1.94</td><td align="center" rowspan="1" colspan="1">9.38</td></tr><tr><td align="left" rowspan="1" colspan="1">All causes</td><td align="center" rowspan="1" colspan="1">506.38</td><td align="center" rowspan="1" colspan="1">11.57</td><td align="center" rowspan="1" colspan="1">3.93</td><td align="center" rowspan="1" colspan="1">1.09</td><td align="center" rowspan="1" colspan="1">2.01</td><td align="center" rowspan="1" colspan="1">12.39</td><td align="center" rowspan="1" colspan="1">64.01</td></tr></tbody></table></table-wrap><table-wrap id="T0003" position="float"><label>Table 3</label><caption><p>Cause-specific mortality rates per 1,000 person-years by age groups (Female)</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1">Causes</th><th align="center" rowspan="1" colspan="1">&#x0003c;28 days</th><th align="center" rowspan="1" colspan="1">1&#x02013;11 months</th><th align="center" rowspan="1" colspan="1">1&#x02013;4 years</th><th align="center" rowspan="1" colspan="1">5&#x02013;14 years</th><th align="center" rowspan="1" colspan="1">15&#x02013;49 years</th><th align="center" rowspan="1" colspan="1">50&#x02013;64 years</th><th align="center" rowspan="1" colspan="1">65+ years</th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">01.01 Sepsis (non-obstetric)</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.17</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">01.02 Acute respiratory infection including pneumonia</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">4.82</td><td align="center" rowspan="1" colspan="1">0.77</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.59</td><td align="center" rowspan="1" colspan="1">5.79</td></tr><tr><td align="left" rowspan="1" colspan="1">01.03 HIV/AIDS-related death</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.14</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">01.04 Diarrhoeal diseases</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">1.61</td><td align="center" rowspan="1" colspan="1">0.46</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.48</td></tr><tr><td align="left" rowspan="1" colspan="1">01.05 Malaria</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.17</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">01.06 Measles</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.80</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">01.07 Meningitis and encephalitis</td><td align="center" rowspan="1" colspan="1">5.81</td><td align="center" rowspan="1" colspan="1">1.32</td><td align="center" rowspan="1" colspan="1">0.15</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">01.09 Pulmonary tuberculosis</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">2.12</td><td align="center" rowspan="1" colspan="1">10.79</td></tr><tr><td align="left" rowspan="1" colspan="1">01.10 Pertussis</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.59</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">01.99 Other and unspecified infectious diseases</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.21</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">0.13</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">0.74</td></tr><tr><td align="left" rowspan="1" colspan="1">02.01 Oral neoplasms</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.20</td></tr><tr><td align="left" rowspan="1" colspan="1">02.02 Digestive neoplasms</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.49</td><td align="center" rowspan="1" colspan="1">1.38</td></tr><tr><td align="left" rowspan="1" colspan="1">02.03 Respiratory neoplasms</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">0.62</td><td align="center" rowspan="1" colspan="1">2.36</td></tr><tr><td align="left" rowspan="1" colspan="1">02.04 Breast neoplasms</td><td align="center" rowspan="1" colspan="1">na</td><td align="center" rowspan="1" colspan="1">na</td><td align="center" rowspan="1" colspan="1">na</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.37</td></tr><tr><td align="left" rowspan="1" colspan="1">02.05 &#x00026; 02.06 Reproductive neoplasms</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">1.01</td></tr><tr><td align="left" rowspan="1" colspan="1">02.99 Other and unspecified neoplasms</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">1.00</td><td align="center" rowspan="1" colspan="1">1.72</td></tr><tr><td align="left" rowspan="1" colspan="1">03.01 Severe anaemia</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.74</td></tr><tr><td align="left" rowspan="1" colspan="1">03.02 Severe malnutrition</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.40</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.43</td></tr><tr><td align="left" rowspan="1" colspan="1">03.03 Diabetes mellitus</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">0.79</td><td align="center" rowspan="1" colspan="1">7.13</td></tr><tr><td align="left" rowspan="1" colspan="1">04.01 Acute cardiac disease</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.21</td><td align="center" rowspan="1" colspan="1">1.39</td></tr><tr><td align="left" rowspan="1" colspan="1">04.02 Stroke</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">1.16</td><td align="center" rowspan="1" colspan="1">6.00</td></tr><tr><td align="left" rowspan="1" colspan="1">04.99 Other and unspecified cardiac diseases</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.45</td><td align="center" rowspan="1" colspan="1">2.93</td></tr><tr><td align="left" rowspan="1" colspan="1">05.01 Chronic obstructive pulmonary diseases</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">1.65</td><td align="center" rowspan="1" colspan="1">4.95</td></tr><tr><td align="left" rowspan="1" colspan="1">05.02 Asthma</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.57</td><td align="center" rowspan="1" colspan="1">0.94</td></tr><tr><td align="left" rowspan="1" colspan="1">06.01 Acute abdomen</td><td align="center" rowspan="1" colspan="1">1.94</td><td align="center" rowspan="1" colspan="1">0.52</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.13</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.41</td><td align="center" rowspan="1" colspan="1">0.60</td></tr><tr><td align="left" rowspan="1" colspan="1">06.02 Liver cirrhosis</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.34</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">07.01 Renal failure</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">0.37</td><td align="center" rowspan="1" colspan="1">1.47</td></tr><tr><td align="left" rowspan="1" colspan="1">08.01 Epilepsy</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.25</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.48</td></tr><tr><td align="left" rowspan="1" colspan="1">09.03 Pregnancy-induced hypertension</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">10.01 Prematurity</td><td align="center" rowspan="1" colspan="1">62.67</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">10.02 Birth asphyxia</td><td align="center" rowspan="1" colspan="1">111.24</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">10.03 Neonatal pneumonia</td><td align="center" rowspan="1" colspan="1">27.47</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">10.04 Neonatal sepsis</td><td align="center" rowspan="1" colspan="1">2.86</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">10.06 Congenital malformation</td><td align="center" rowspan="1" colspan="1">4.10</td><td align="center" rowspan="1" colspan="1">2.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">10.99 Other and unspecified neonatal CoD</td><td align="center" rowspan="1" colspan="1">66.87</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">12.03 Accidental fall</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.38</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.26</td></tr><tr><td align="left" rowspan="1" colspan="1">12.04 Accidental drowning and submersion</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.80</td><td align="center" rowspan="1" colspan="1">0.93</td><td align="center" rowspan="1" colspan="1">0.26</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.23</td></tr><tr><td align="left" rowspan="1" colspan="1">12.05 Accidental expose to smoke fire &#x00026; flame</td><td align="center" rowspan="1" colspan="1">3.73</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">1.24</td></tr><tr><td align="left" rowspan="1" colspan="1">12.08 Intentional self-harm</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.22</td><td align="center" rowspan="1" colspan="1">0.20</td></tr><tr><td align="left" rowspan="1" colspan="1">12.09 Assault</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">12.99 Other and unspecified external CoD</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.14</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">98 Other and unspecified NCD</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.20</td></tr><tr><td align="left" rowspan="1" colspan="1">99 Indeterminate</td><td align="center" rowspan="1" colspan="1">114.24</td><td align="center" rowspan="1" colspan="1">4.44</td><td align="center" rowspan="1" colspan="1">0.73</td><td align="center" rowspan="1" colspan="1">0.14</td><td align="center" rowspan="1" colspan="1">0.46</td><td align="center" rowspan="1" colspan="1">1.57</td><td align="center" rowspan="1" colspan="1">11.96</td></tr><tr><td align="left" rowspan="1" colspan="1">All causes</td><td align="center" rowspan="1" colspan="1">405.53</td><td align="center" rowspan="1" colspan="1">21.96</td><td align="center" rowspan="1" colspan="1">3.98</td><td align="center" rowspan="1" colspan="1">0.90</td><td align="center" rowspan="1" colspan="1">1.76</td><td align="center" rowspan="1" colspan="1">13.16</td><td align="center" rowspan="1" colspan="1">68.88</td></tr></tbody></table></table-wrap><p>About one-fifth of the deaths among post-neonates (1&#x02013;11 months) were associated with acute respiratory infection including pneumonia, followed by diarrhoeal diseases and congenital malformations. Death due to acute respiratory infection was double for females (4.8 per 1,000 person-years) than for males (2.2 per 1,000 person-years).</p><p>Thirty-one percent of deaths among children aged 1&#x02013;4 years was due to accidental drowning or submersion (1.2/1,000 person-years), followed by acute respiratory infection including pneumonia (0.8/1,000 per years) and diarrhoeal diseases. Deaths due to accidental drowning or submersion were higher for males (1.5 per 1,000 person-years) than for females (1.0 per 1,000 person-years). Accidental drowning or submersion was also the leading cause of death among 5&#x02013;14 year old children. The death rates due to accidental drowning or submersion were similar for males (0.23 per 1,000 person-years) and females (0.26 per 1,000-person-years) in this age group.</p><p>Pulmonary TB was the leading cause of death among age groups 15&#x02013;49, 50&#x02013;64, and 65 years and older. Death due to TB increases as age increases. No TB cases were observed before the age of 15 years. The death rate due to TB was similar for both males and females.</p></sec><sec id="S0003-S20003"><title>Broad pattern of cause of death</title><p>The specific causes were then categorised into four broad groups: NCDs, CDs, perinatal and neonatal conditions, and injury and accidents. When grouped according to causes, NCDs were the leading cause of deaths (37%), followed by CDs (22%), perinatal and neonatal conditions (11%), and injury and accidents (6%). Causes for remaining 24% of the deaths could not be determined. Distribution of cause of death varied by sex. CDs were higher for females (24%) than for males (20%); perinatal and neonatal conditions were higher for males (12%) than for females (10%). No sex differential was observed for NCDs (37% for males and 36% for females), and injury and accidents (6%) (<xref ref-type="fig" rid="F0001">Fig. 1</xref>).</p><fig id="F0001" position="float"><label>Fig. 1</label><caption><p>Distribution of cause of death by sex.</p></caption><graphic xlink:href="GHA-7-25473-g001"/></fig></sec><sec id="S0003-S20004"><title>Socioeconomic differential in cause and age-specific mortality</title><p> <xref ref-type="fig" rid="F0002">Figure 2</xref> shows the distribution of a broad pattern of causes of deaths by SES. It is clear that NCDs concentrated more among the better-off and CDs among the poor.</p><fig id="F0002" position="float"><label>Fig. 2</label><caption><p>Distribution of cause of death by SES.</p></caption><graphic xlink:href="GHA-7-25473-g002"/></fig><p>We further analysed the relationship between SES and age-specific mortality. For this analysis we divided the population into two groups &#x02013; 15 years or less, and above 15 years. This was done because of the finding that the diseases found in the population aged less than 15 years were mostly infectious diseases and that the population above 15 years suffered mostly from NCDs.</p><p>For the population aged 15 years or less, results show a significantly negative relationship between mortality and SES (<italic>p</italic>&#x0003c;0.004) (<xref ref-type="fig" rid="F0003">Fig. 3</xref>). For the age group 15 and above, the effect of SES was not significant. Mortality rates were higher among the poorest and the richest groups, whereas for the other three middle quintiles rates remained similar (<xref ref-type="fig" rid="F0003">Fig. 3</xref>). We carried out further analysis to explain this particular finding. The broad categorisation of cause-specific mortality rate showed that NCDs and CDs are the leading causes of death and these are concentrated among the adult and elderly, that is, the 15 and older age group.</p><fig id="F0003" position="float"><label>Fig. 3</label><caption><p>Mortality rates (per 1,000 person-years) by age groups.</p></caption><graphic xlink:href="GHA-7-25473-g003"/></fig><p>Analysing the socioeconomic differential in mortality due to NCDs and CDs showed that death rate due to NCDs steeply increased with increasing SES (<italic>p</italic>&#x0003c;0.005) for the 15 and older age group (<xref ref-type="fig" rid="F0004">Fig. 4</xref>). Mortality rate was 39% higher for the highest quintile compared to the lowest quintile (OR 1.39 (95% CI: 1.02&#x02013;1.92)). On the other hand, burden of CDs was concentrated among the poor with rates being 92% higher for lowest quintile compared to the highest quintile (<italic>p</italic>&#x0003c;0.001) [OR 1.92 (95% CI: 1.25&#x02013;2.95)] (<xref ref-type="fig" rid="F0004">Fig. 4</xref>). This can be explained by the fact that TB, which is known to be the disease of the poor, occupies the major share in CDs (<xref ref-type="fig" rid="F0005">Fig. 5</xref>). Data indicates that mortality due to TB is also negatively related to the SES of the deceased (<xref ref-type="fig" rid="F0005">Fig. 5</xref>).</p><fig id="F0004" position="float"><label>Fig. 4</label><caption><p>Mortality rates per 1,000 person-years for CDs and NCDs by SES (15 years and above).</p></caption><graphic xlink:href="GHA-7-25473-g004"/></fig><fig id="F0005" position="float"><label>Fig. 5</label><caption><p>Prevalence of communicable diseases per 100,000 person-years (15 years and above).</p></caption><graphic xlink:href="GHA-7-25473-g005"/></fig></sec></sec><sec sec-type="discussion" id="S0004"><title>Discussion</title><p>Two sets of empirical evidence drive the discussion of this paper. One is on the mortality differential among the various socioeconomic groups, concluding that mortality rate in general is higher among the lower socioeconomic groups (<xref rid="CIT0005" ref-type="bibr">5</xref>, <xref rid="CIT0012" ref-type="bibr">12</xref>, <xref rid="CIT0022" ref-type="bibr">22</xref>). The other set of evidence indicates the emergence of an epidemiological transition of disease burden from CDs to NCDs (<xref rid="CIT0005" ref-type="bibr">5</xref>, <xref rid="CIT0023" ref-type="bibr">23</xref>). Data presented in this paper showed that the poor experience higher mortality from CDs, and the better off experience lower mortality from CDs and higher mortality from NCDs. This particular finding is quite expected.</p><p>In Chakaria, pulmonary TB, acute respiratory infection, stroke, diabetes mellitus, and chronic obstructive pulmonary diseases were found to be the leading causes of death in all age groups. Premature birth for neonatal period, respiratory infections for post-neonatal period, and drowning and accidents for ages 1&#x02013;4 years, were the observed cause of death which is in line with the national estimates (<xref rid="CIT0012" ref-type="bibr">12</xref>). Overall childhood (0&#x02013;14 years) mortality was dominated by premature birth, respiratory infection, and drowning. At age 15 years and older, pulmonary TB was the observed leading cause of death in Chakaria. This differs from the findings of Matlab HDSS, where the overall mortality is much lower than Chakaria (<xref rid="CIT0024" ref-type="bibr">24</xref>). The dominance of TB in Chakaria was observed despite the presence of the National TB Control Programme in the area. This may be because of inadequate coverage of the TB programme in the area due to its remoteness and high prevalence of poverty (<xref rid="CIT0016" ref-type="bibr">16</xref>), for TB is known to be a disease of poverty (<xref rid="CIT0025" ref-type="bibr">25</xref>&#x02013;<xref rid="CIT0027" ref-type="bibr">27</xref>). This is particularly important for Bangladesh as it is one of the 22 high TB-burden countries of the world with a prevalence of TB cases of 225 per 100,000 population and this rate has remained more or less constant for the last two decades (<xref rid="CIT0028" ref-type="bibr">28</xref>). In addition, among these 22 countries, the percentage of cases that were bacteriologically confirmed was high in Bangladesh (81%) (<xref rid="CIT0028" ref-type="bibr">28</xref>). Therefore, it is essential to strengthen the TB control programme in the remote areas like Chakaria if equitable progress is to be made.</p><p>Reducing the level of and inequity in child mortality are of global importance. One of the MDGs directly targets child mortality (<xref rid="CIT0029" ref-type="bibr">29</xref>). The current study findings suggest that child mortality (0&#x02013;14 years) in Chakaria was inversely associated with SES. An inverse association between SES and under-five mortality has been observed in several earlier studies conducted elsewhere in Bangladesh (<xref rid="CIT0012" ref-type="bibr">12</xref>, <xref rid="CIT0014" ref-type="bibr">14</xref>). The leading cause of death in this group in Chakaria can mostly be classified into the infectious disease category, which is highly influenced by living conditions including the status of health care delivery, water sanitation, and above all the extent of poverty. In Chakaria, 30% of the households have no land, and half the population depend on menial labour to make a living. Sanitation is poor with half the population not using a sanitary toilet (<xref rid="CIT0011" ref-type="bibr">11</xref>). Low utilisation of maternal and child health (MCH) services and inequity in safe motherhood practices still remain a challenge in Chakaria. Skipping meals during economic fluctuations such as price hikes is common in the area. All these factors together might have contributed to the higher mortality rate observed among the children (0&#x02013;14 years) belonging to the lower SES groups in Chakaria. Increasing child survival and decreasing the gap between rich and poor warrant interventions and approaches targeted towards improving access of the poor to effective health care particularly, to safe motherhood services.</p><p>On the contrary, for the population aged 15 years and older, SES did not have significant impact on mortality rate other than mortality rates were higher among the poorest from CDs and among the richest groups from NCDs. The counter effect of each of the two disease burdens in the two SES groups might have negated the overall effect of SES on mortality rate. The high mortality rate for the poor is expected to be contributed by CDs, mostly by TB, as TB was found to be more prevalent in lower SES groups.</p><p>The findings of our study suggest that a disaggregated investigation of causes of death by age and SES brings in results that indicate existence of variation in risk factors for the different groups of population. Thus a deeper understanding and careful exploration of causes of death is necessary for refining health services. Targeted interventions which deal with diseases specific to each group have the potential to contribute in achieving the global target of 25% reduction in premature mortality from NCDs by 2025 (<xref rid="CIT0030" ref-type="bibr">30</xref>) and the reduction of child mortality by two-thirds between 1990 and 2015 (MDG4) for Bangladesh (<xref rid="CIT0029" ref-type="bibr">29</xref>).</p></sec><sec id="S0005"><title>Strengths and weaknesses</title><p>Cause-of-death data for this study came from a continuous surveillance system in a particular population, which is a major strength of this study. Data were collected through quarterly household visits by experienced field workers. The routine visits by locally recruited trained staff members reduced the chance of missing deaths and collecting detailed information required to have quality VA for ascertaining causes of death. The indeterminate category is a weakness, a usual characteristic of this kind of effort. In the absence of a national system for death reporting and cause of death, these data are still a valuable source for planning health policies and programmes. Data on causes of death by SES is rarely available in settings like Bangladesh. Thus this paper is a significant contribution to research and proves that it is important and possible to analyse these kind of data by SES which is another noteworthy strength of this paper.</p></sec><sec sec-type="conclusion" id="S0006"><title>Conclusion</title><p>The shift towards NCDs both as a burden and mortality is expected and quite consistent with other small-area-based data in Bangladesh such as Matlab (<xref rid="CIT0005" ref-type="bibr">5</xref>, <xref rid="CIT0024" ref-type="bibr">24</xref>). The fact that CD is a major burden and cause of death for poor in all age groups, clearly indicates that the epidemiological transition towards NCDs from CD is yet to set in for Bangladesh. Both the poor and the health system of the country still have a double burden to carry for quite some time. More burden of NCDs among the better-off is not a surprise. However, preventive and curative measures to minimize the risk of NCDs nationwide should be initiated before it is too late. The existence of SES inequity is an alert for the policy makers and programme managers to remain vigilant and not to be carried away by the average gains in mortality reduction.</p></sec>

Malaria mortality in Africa and Asia: evidence from INDEPTH health and demographic surveillance system sites

<sec id="st1"><title>Background</title><p>Malaria continues to be a major cause of infectious disease mortality in tropical regions. However, deaths from malaria are most often not individually documented, and as a result overall understanding of malaria epidemiology is inadequate. INDEPTH Network members maintain population surveillance in Health and Demographic Surveillance System sites across Africa and Asia, in which individual deaths are followed up with verbal autopsies.</p></sec><sec id="st2"><title>Objective</title><p>To present patterns of malaria mortality determined by verbal autopsy from INDEPTH sites across Africa and Asia, comparing these findings with other relevant information on malaria in the same regions.</p></sec><sec id="st3"><title>Design</title><p>From a database covering 111,910 deaths over 12,204,043 person-years in 22 sites, in which verbal autopsy data were handled according to the WHO 2012 standard and processed using the InterVA-4 model, over 6,000 deaths were attributed to malaria. The overall period covered was 1992&#x02013;2012, but two-thirds of the observations related to 2006&#x02013;2012. These deaths were analysed by site, time period, age group and sex to investigate epidemiological differences in malaria mortality.</p></sec><sec id="st4"><title>Results</title><p>Rates of malaria mortality varied by 1:10,000 across the sites, with generally low rates in Asia (one site recording no malaria deaths over 0.5 million person-years) and some of the highest rates in West Africa (Nouna, Burkina Faso: 2.47 per 1,000 person-years). Childhood malaria mortality rates were strongly correlated with Malaria Atlas Project estimates of <italic>Plasmodium falciparum</italic> parasite rates for the same locations. Adult malaria mortality rates, while lower than corresponding childhood rates, were strongly correlated with childhood rates at the site level.</p></sec><sec id="st5"><title>Conclusions</title><p>The wide variations observed in malaria mortality, which were nevertheless consistent with various other estimates, suggest that population-based registration of deaths using verbal autopsy is a useful approach to understanding the details of malaria epidemiology.</p></sec>

<contrib contrib-type="author"><name><surname>Streatfield</surname><given-names>P. Kim</given-names></name><xref ref-type="aff" rid="AF0001">1</xref><xref ref-type="aff" rid="AF0002">2</xref><xref ref-type="aff" rid="AF0003">3</xref></contrib><contrib contrib-type="author"><name><surname>Khan</surname><given-names>Wasif A.</given-names></name><xref ref-type="aff" rid="AF0002">2</xref><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0004">4</xref></contrib><contrib contrib-type="author"><name><surname>Bhuiya</surname><given-names>Abbas</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0005">5</xref><xref ref-type="aff" rid="AF0006">6</xref></contrib><contrib contrib-type="author"><name><surname>Hanifi</surname><given-names>Syed M.A.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0005">5</xref><xref ref-type="aff" rid="AF0006">6</xref></contrib><contrib contrib-type="author"><name><surname>Alam</surname><given-names>Nurul</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0007">7</xref><xref ref-type="aff" rid="AF0008">8</xref></contrib><contrib contrib-type="author"><name><surname>Diboulo</surname><given-names>Eric</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0009">9</xref><xref ref-type="aff" rid="AF0010">10</xref></contrib><contrib contrib-type="author"><name><surname>Si&#x000e9;</surname><given-names>Ali</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0009">9</xref><xref ref-type="aff" rid="AF0010">10</xref></contrib><contrib contrib-type="author"><name><surname>Y&#x000e9;</surname><given-names>Maurice</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0009">9</xref><xref ref-type="aff" rid="AF0010">10</xref></contrib><contrib contrib-type="author"><name><surname>Compaor&#x000e9;</surname><given-names>Yacouba</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0011">11</xref><xref ref-type="aff" rid="AF0012">12</xref></contrib><contrib contrib-type="author"><name><surname>Soura</surname><given-names>Abdramane B.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0011">11</xref><xref ref-type="aff" rid="AF0012">12</xref></contrib><contrib contrib-type="author"><name><surname>Bonfoh</surname><given-names>Bassirou</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0013">13</xref><xref ref-type="aff" rid="AF0014">14</xref></contrib><contrib contrib-type="author"><name><surname>Jaeger</surname><given-names>Fabienne</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0013">13</xref><xref ref-type="aff" rid="AF0015">15</xref></contrib><contrib contrib-type="author"><name><surname>Ngoran</surname><given-names>Eliezer K.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0013">13</xref><xref ref-type="aff" rid="AF0016">16</xref></contrib><contrib contrib-type="author"><name><surname>Utzinger</surname><given-names>Juerg</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0013">13</xref><xref ref-type="aff" rid="AF0015">15</xref></contrib><contrib contrib-type="author"><name><surname>Melaku</surname><given-names>Yohannes A.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0017">17</xref><xref ref-type="aff" rid="AF0018">18</xref></contrib><contrib contrib-type="author"><name><surname>Mulugeta</surname><given-names>Afework</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0017">17</xref><xref ref-type="aff" rid="AF0018">18</xref></contrib><contrib contrib-type="author"><name><surname>Weldearegawi</surname><given-names>Berhe</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0017">17</xref><xref ref-type="aff" rid="AF0018">18</xref></contrib><contrib contrib-type="author"><name><surname>Gomez</surname><given-names>Pierre</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0019">19</xref><xref ref-type="aff" rid="AF0020">20</xref></contrib><contrib contrib-type="author"><name><surname>Jasseh</surname><given-names>Momodou</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0019">19</xref><xref ref-type="aff" rid="AF0020">20</xref></contrib><contrib contrib-type="author"><name><surname>Hodgson</surname><given-names>Abraham</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0021">21</xref><xref ref-type="aff" rid="AF0022">22</xref></contrib><contrib contrib-type="author"><name><surname>Oduro</surname><given-names>Abraham</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0021">21</xref><xref ref-type="aff" rid="AF0022">22</xref></contrib><contrib contrib-type="author"><name><surname>Welaga</surname><given-names>Paul</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0021">21</xref><xref ref-type="aff" rid="AF0022">22</xref></contrib><contrib contrib-type="author"><name><surname>Williams</surname><given-names>John</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0021">21</xref><xref ref-type="aff" rid="AF0022">22</xref></contrib><contrib contrib-type="author"><name><surname>Awini</surname><given-names>Elizabeth</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0023">23</xref><xref ref-type="aff" rid="AF0024">24</xref><xref ref-type="aff" rid="AF0025">25</xref></contrib><contrib contrib-type="author"><name><surname>Binka</surname><given-names>Fred N.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0023">23</xref><xref ref-type="aff" rid="AF0025">25</xref></contrib><contrib contrib-type="author"><name><surname>Gyapong</surname><given-names>Margaret</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0023">23</xref><xref ref-type="aff" rid="AF0025">25</xref></contrib><contrib contrib-type="author"><name><surname>Kant</surname><given-names>Shashi</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0026">26</xref><xref ref-type="aff" rid="AF0027">27</xref></contrib><contrib contrib-type="author"><name><surname>Misra</surname><given-names>Puneet</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0026">26</xref><xref ref-type="aff" rid="AF0027">27</xref></contrib><contrib contrib-type="author"><name><surname>Srivastava</surname><given-names>Rahul</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0026">26</xref><xref ref-type="aff" rid="AF0027">27</xref></contrib><contrib contrib-type="author"><name><surname>Chaudhary</surname><given-names>Bharat</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0028">28</xref><xref ref-type="aff" rid="AF0029">29</xref></contrib><contrib contrib-type="author"><name><surname>Juvekar</surname><given-names>Sanjay</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0028">28</xref><xref ref-type="aff" rid="AF0029">29</xref></contrib><contrib contrib-type="author"><name><surname>Wahab</surname><given-names>Abdul</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0030">30</xref><xref ref-type="aff" rid="AF0031">31</xref></contrib><contrib contrib-type="author"><name><surname>Wilopo</surname><given-names>Siswanto</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0030">30</xref><xref ref-type="aff" rid="AF0031">31</xref></contrib><contrib contrib-type="author"><name><surname>Bauni</surname><given-names>Evasius</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0032">32</xref><xref ref-type="aff" rid="AF0033">33</xref></contrib><contrib contrib-type="author"><name><surname>Mochamah</surname><given-names>George</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0032">32</xref><xref ref-type="aff" rid="AF0033">33</xref></contrib><contrib contrib-type="author"><name><surname>Ndila</surname><given-names>Carolyne</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0032">32</xref><xref ref-type="aff" rid="AF0033">33</xref></contrib><contrib contrib-type="author"><name><surname>Williams</surname><given-names>Thomas N.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0032">32</xref><xref ref-type="aff" rid="AF0033">33</xref><xref ref-type="aff" rid="AF0034">34</xref></contrib><contrib contrib-type="author"><name><surname>Hamel</surname><given-names>Mary J.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0035">35</xref><xref ref-type="aff" rid="AF0036">36</xref></contrib><contrib contrib-type="author"><name><surname>Lindblade</surname><given-names>Kim A.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0035">35</xref><xref ref-type="aff" rid="AF0036">36</xref></contrib><contrib contrib-type="author"><name><surname>Odhiambo</surname><given-names>Frank O.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0035">35</xref><xref ref-type="aff" rid="AF0036">36</xref></contrib><contrib contrib-type="author"><name><surname>Slutsker</surname><given-names>Laurence</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0035">35</xref><xref ref-type="aff" rid="AF0036">36</xref></contrib><contrib contrib-type="author"><name><surname>Ezeh</surname><given-names>Alex</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0037">37</xref><xref ref-type="aff" rid="AF0038">38</xref></contrib><contrib contrib-type="author"><name><surname>Kyobutungi</surname><given-names>Catherine</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0037">37</xref><xref ref-type="aff" rid="AF0038">38</xref></contrib><contrib contrib-type="author"><name><surname>Wamukoya</surname><given-names>Marylene</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0037">37</xref><xref ref-type="aff" rid="AF0038">38</xref></contrib><contrib contrib-type="author"><name><surname>Delaunay</surname><given-names>Val&#x000e9;rie</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0039">39</xref><xref ref-type="aff" rid="AF0040">40</xref></contrib><contrib contrib-type="author"><name><surname>Diallo</surname><given-names>Aldiouma</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0039">39</xref><xref ref-type="aff" rid="AF0040">40</xref></contrib><contrib contrib-type="author"><name><surname>Douillot</surname><given-names>Laetitia</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0039">39</xref><xref ref-type="aff" rid="AF0040">40</xref></contrib><contrib contrib-type="author"><name><surname>Sokhna</surname><given-names>Cheikh</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0039">39</xref><xref ref-type="aff" rid="AF0040">40</xref></contrib><contrib contrib-type="author"><name><surname>G&#x000f3;mez-Oliv&#x000e9;</surname><given-names>F. Xavier</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0041">41</xref><xref ref-type="aff" rid="AF0042">42</xref></contrib><contrib contrib-type="author"><name><surname>Kabudula</surname><given-names>Chodziwadziwa W.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0041">41</xref><xref ref-type="aff" rid="AF0042">42</xref></contrib><contrib contrib-type="author"><name><surname>Mee</surname><given-names>Paul</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0041">41</xref><xref ref-type="aff" rid="AF0042">42</xref></contrib><contrib contrib-type="author"><name><surname>Herbst</surname><given-names>Kobus</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0043">43</xref><xref ref-type="aff" rid="AF0044">44</xref></contrib><contrib contrib-type="author"><name><surname>Mossong</surname><given-names>Jo&#x000eb;l</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0043">43</xref><xref ref-type="aff" rid="AF0044">44</xref><xref ref-type="aff" rid="AF0045">45</xref></contrib><contrib contrib-type="author"><name><surname>Chuc</surname><given-names>Nguyen T.K.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0046">46</xref><xref ref-type="aff" rid="AF0047">47</xref></contrib><contrib contrib-type="author"><name><surname>Arthur</surname><given-names>Samuelina S.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref></contrib><contrib contrib-type="author"><name><surname>Sankoh</surname><given-names>Osman A.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0048">48</xref><xref ref-type="aff" rid="AF0049">49</xref><xref ref-type="corresp" rid="cor1">*</xref></contrib><contrib contrib-type="author"><name><surname>Tanner</surname><given-names>Marcel</given-names></name><xref ref-type="aff" rid="AF0050">50</xref></contrib><contrib contrib-type="author"><name><surname>Byass</surname><given-names>Peter</given-names></name><xref ref-type="aff" rid="AF0051">51</xref></contrib>

Global Health Action

<p>The epidemiology of malaria in Africa and Asia has been extensively, but not always systematically, investigated. Many studies have focused on young children&#x02019;s exposure to the disease (<xref rid="CIT0001" ref-type="bibr">1</xref>), and to some extent the effects on pregnant women (<xref rid="CIT0002" ref-type="bibr">2</xref>), without evaluating the malaria status of other population sub-groups. Few studies have looked specifically at the impact of malaria on older people (<xref rid="CIT0003" ref-type="bibr">3</xref>). Many data have been taken from heath facilities at various levels and may be influenced by patterns of health services utilisation rather than clearly representing malaria patterns within communities (<xref rid="CIT0004" ref-type="bibr">4</xref>). Some work has taken whatever data may be available and sought to generalise patterns of malaria burden using sophisticated modelling techniques (<xref rid="CIT0005" ref-type="bibr">5</xref>). Nevertheless, malaria remains as an important cause of infectious disease mortality in many parts of Africa, and some areas in Asia and Latin America. WHO&#x02019;s World Malaria Report 2013 suggests that malaria mortality rates fell by more than 40% from 2000 to 2012, a period during which there was substantial international investment in malaria control (<xref rid="CIT0006" ref-type="bibr">6</xref>). However, although malaria transmission has successfully been reduced in many former high-incidence settings, few areas have become malaria-free. The need for adequate, reliable evidence on malaria mortality in various populations therefore remains as important as ever, and data at the population level are crucially needed to validate and understand top-down estimates.</p><p>As is the case for deaths from all diseases, malaria deaths are generally poorly verified and documented in Africa and some parts of Asia. Attributing a death to malaria after the event is not easy &#x02013; in highly endemic areas, acute febrile deaths may be likely to be described as malaria and lead to over-attribution, whereas the converse may apply in settings where malaria is uncommon. It has been suggested that over-attribution of malaria as a clinical diagnosis in endemic areas may even be dangerous (<xref rid="CIT0007" ref-type="bibr">7</xref>). Because most malaria deaths occur in areas not covered by routine death certification, verbal autopsy (VA) methods have been used in many settings as the only available source of cause of death data, but their validity in absolute terms for assigning malaria as a cause of death remains open to question. Rapid diagnostic tests (RDTs) are becoming increasingly widely used as a basis for malaria treatment decisions, and, where RDT results are known from an illness leading to death, either positive or negative RDT results may increase the available VA information and hence the accuracy of cause of death attribution. Consequently in the WHO 2012 VA standard, specific items on a recent positive or negative test result were introduced (<xref rid="CIT0008" ref-type="bibr">8</xref>). However, it will be some time before sufficient VAs are collected which include those data items to assess their utility as part of the VA process.</p><p>In this paper, we present malaria-specific mortality rates derived from standardised VA data in 22 INDEPTH Network Health and Demographic Surveillance Sites (HDSS) across Africa and Asia (<xref rid="CIT0009" ref-type="bibr">9</xref>). Although these HDSSs are not designed to form a representative network, each one follows a geographically defined population longitudinally, systematically recording all death events and undertaking VAs on deaths that occur. Sites with longer time-series may therefore be able to measure changes over time effectively. Our aim is to present the malaria mortality patterns at each site, comparing these community-level findings with other information on malaria in Africa and Asia.</p><sec sec-type="methods" id="S0001"><title>Methods</title><p>The overall public-domain INDEPTH dataset (<xref rid="CIT0010" ref-type="bibr">10</xref>) from which these malaria-specific analyses are drawn is described in detail elsewhere (<xref rid="CIT0011" ref-type="bibr">11</xref>), with full details of methods used, which are also summarised here in <xref ref-type="boxed-text" rid="B0001">Box 1</xref>. Briefly, the dataset documents 111,910 deaths in 12,204,043 person-years of observation across 22 sites, all processed in a standardised manner. The Karonga site in Malawi did not contribute VAs for children, and for that reason is excluded from further analyses here. The InterVA-4 &#x02018;high&#x02019; malaria setting was used for all the West African sites, plus the East African sites (with the exceptions, on the grounds of high altitude, of Nairobi, Kenya and Kilite-Awlaelo, Ethiopia), on the basis of local experience. All other sites used the &#x02018;low&#x02019; setting; the &#x02018;very low&#x02019; setting was not used. The InterVA-4 guideline is that the &#x02018;high&#x02019; setting is appropriate for an expected malaria cause-specific mortality fraction (CSMF) higher than about 1%, though the setting chosen does not result in any great dichotomisation of outputs; the clinical equivalent would be a physician&#x02019;s knowledge that his/her current case comes from a setting where malaria is more or less likely, irrespective of particular symptoms.</p><p> <italic>Box 1</italic>. Summary of methodology based on the detailed description in the introductory paper (<xref rid="CIT0011" ref-type="bibr">11</xref>) </p><boxed-text id="B0001" position="float"><p> <bold>Age&#x02013;sex&#x02013;time standardisation</bold> </p><p>To avoid effects of differences and changes in age&#x02013;sex structures of populations, mortality fractions and rates have been adjusted using the INDEPTH 2013 population standard (<xref rid="CIT0012" ref-type="bibr">12</xref>). A weighting factor was calculated for each site, age group, sex and year category in relation to the standard for the corresponding age group and sex, and incorporated into the overall dataset. This is referred to in this paper as age&#x02013;sex&#x02013;time standardisation in the contexts where it is used.</p><p> <bold>Cause of death assignment</bold> </p><p>The InterVA-4 (version 4.02) probabilistic model was used for all the cause of death assignments in the overall dataset (<xref rid="CIT0013" ref-type="bibr">13</xref>). InterVA-4 is fully compliant with the WHO 2012 Verbal Autopsy Standards and generates causes of death categorised by ICD-10 groups (<xref rid="CIT0014" ref-type="bibr">14</xref>). The data reported here were collected before the WHO 2012 VA standard was available, but were transformed into the WHO 2012 and InterVA-4 format to optimise cross-site standardisation in cause of death attribution. For a small proportion of deaths VA interviews were not successfully completed; a few others contained inadequate information to arrive at a cause of death. InterVA-4 assigns causes of death (maximum 3) with associated likelihoods; thus cases for which likely causes did not total 100% were also assigned a residual indeterminate component. This served as a means of encapsulating uncertainty in cause of death at the individual level within the overall dataset, as well as accounting for 100% of every death.</p><p> <bold>Overall dataset</bold> </p><p>The overall public-domain dataset (<xref rid="CIT0010" ref-type="bibr">10</xref>) thus contains between one and four records for each death, with the sum of likelihoods for each individual being unity. Each record includes a specific cause of death, its likelihood and its age-sex-time weighting. </p></boxed-text><p>Deaths assigned to malaria were extracted from the overall data set together with data on person-time exposed by site, year, age and sex. Overall malaria mortality as reflected here amounted to a total of 6,330.8 age&#x02013;sex&#x02013;time standardised deaths, to which 8,076 individually recorded deaths contributed some component of probable malaria mortality. As each HDSS covers a total population, rather than a sample, uncertainty intervals are not shown.</p><p>In this context, all of these data are secondary datasets derived from primary data collected separately by each participating site. In all cases the primary data collection was covered by site-level ethical approvals relating to on-going health and demographic surveillance in those specific locations. No individual identity or household location data were included in the secondary data and no specific ethical approvals were required for these pooled analyses.</p></sec><sec sec-type="results" id="S0002"><title>Results</title><p>The CSMFs for malaria at each site are shown, together with the population-based malaria-specific mortality rate per 1,000 person-years, in <xref ref-type="fig" rid="F0001">Fig. 1</xref>. In West African sites, malaria CSMF ranged from 4.90% to 25.53%, with malaria-specific standardised mortality rates ranging from 0.48 to 2.47 per 1,000 person-years. In Eastern and Southern Africa, CSMFs were 0.40&#x02013;11.61%, with rates from 0.06 to 2.15 per 1,000 person-years. In Asia, CSMFs were 0&#x02013;4.25%, with rates from 0 to 0.24 per 1,000 person-years. One site, AMK in Bangladesh, recorded no malaria deaths in over 0.5 million person-years of observation.</p><fig id="F0001" position="float"><label>Fig. 1</label><caption><p>Map showing participating sites, with age&#x02013;sex&#x02013;time standardised cause-specific mortality fractions and mortality rates for malaria.</p></caption><graphic xlink:href="GHA-7-25369-g001"/></fig><p> <xref ref-type="table" rid="T0001">Table 1</xref> breaks down malaria-specific mortality rates by age group and site. Malaria mortality rates among infants varied considerably, from 0 to 1.4 per 1,000 person-years, with the highest rates not necessarily being in the locations with highest overall malaria mortality. The largest numbers of malaria deaths at most sites occurred in the 1&#x02013;4 year age group, though the highest malaria mortality rate in that age group was 0.43 per 1,000 person-years at Taabo, C&#x000f4;te d&#x02019;Ivoire. Malaria mortality rates in the 5&#x02013;14 year age group were generally lower than the rates for younger children. Similarly, malaria mortality rates among adults were generally lower than those for children, although they tended to increase among the elderly. <xref ref-type="fig" rid="F0002">Figure 2</xref> shows malaria-specific mortality rates for each site by age group, split into time periods (1992&#x02013;1999; 2000&#x02013;2005 and 2006&#x02013;2012), depending on periods when individual sites operated. Logarithmic scales have been used to visualise both high and low levels of malaria mortality while using the same scale for each site. For most sites and most periods there were generally U-shaped relationships between malaria mortality rates and age; naturally more random variation was evident in sites with generally low malaria mortality because of relatively small numbers of cases.</p><fig id="F0002" position="float"><label>Fig. 2</label><caption><p>Malaria mortality rates by site, age group and period at 20 INDEPTH Network sites.</p></caption><graphic xlink:href="GHA-7-25369-g002"/></fig><table-wrap id="T0001" position="float"><label>Table 1</label><caption><p>Malaria-specific deaths and mortality rates per 1,000 person-years, by age group and site</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1"/><th align="center" colspan="6" rowspan="1">Age group at death</th></tr><tr><th align="left" rowspan="1" colspan="1"/><th colspan="6" rowspan="1"> <hr/> </th></tr><tr><th align="left" rowspan="1" colspan="1">Country: Site</th><th align="left" rowspan="1" colspan="1">Infant</th><th align="center" rowspan="1" colspan="1">1&#x02013;4 years</th><th align="center" rowspan="1" colspan="1">5&#x02013;14 years</th><th align="center" rowspan="1" colspan="1">15&#x02013;49 years</th><th align="center" rowspan="1" colspan="1">50&#x02013;64 years</th><th align="center" rowspan="1" colspan="1">65+ years</th></tr></thead><tbody><tr><td colspan="7" align="left" rowspan="1">Bangladesh: Matlab</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Deaths</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.41</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Rate/1,000 py</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td colspan="7" align="left" rowspan="1">Bangladesh: Bandarban</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Deaths</td><td align="center" rowspan="1" colspan="1">0.98</td><td align="center" rowspan="1" colspan="1">1.00</td><td align="center" rowspan="1" colspan="1">2.46</td><td align="center" rowspan="1" colspan="1">3.76</td><td align="center" rowspan="1" colspan="1">1.47</td><td align="center" rowspan="1" colspan="1">3.25</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Rate/1,000 py</td><td align="center" rowspan="1" colspan="1">0.79</td><td align="center" rowspan="1" colspan="1">0.17</td><td align="center" rowspan="1" colspan="1">0.18</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">0.25</td><td align="center" rowspan="1" colspan="1">1.03</td></tr><tr><td colspan="7" align="left" rowspan="1">Bangladesh: Chakaria</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Deaths</td><td align="center" rowspan="1" colspan="1">0.43</td><td align="center" rowspan="1" colspan="1">1.23</td><td align="center" rowspan="1" colspan="1">1.99</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.28</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Rate/1,000 py</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.03</td></tr><tr><td colspan="7" align="left" rowspan="1">Bangladesh: AMK</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Deaths</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Rate/1,000 py</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td colspan="7" align="left" rowspan="1">Burkina Faso: Nouna</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Deaths</td><td align="center" rowspan="1" colspan="1">507.76</td><td align="center" rowspan="1" colspan="1">859.38</td><td align="center" rowspan="1" colspan="1">140.73</td><td align="center" rowspan="1" colspan="1">108.93</td><td align="center" rowspan="1" colspan="1">76.24</td><td align="center" rowspan="1" colspan="1">287.96</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Rate/1,000 py</td><td align="center" rowspan="1" colspan="1">0.75</td><td align="center" rowspan="1" colspan="1">0.20</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.42</td><td align="center" rowspan="1" colspan="1">0.70</td></tr><tr><td colspan="7" align="left" rowspan="1">Burkina Faso: Ouagadougou</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Deaths</td><td align="center" rowspan="1" colspan="1">19.48</td><td align="center" rowspan="1" colspan="1">68.03</td><td align="center" rowspan="1" colspan="1">17.90</td><td align="center" rowspan="1" colspan="1">8.56</td><td align="center" rowspan="1" colspan="1">2.72</td><td align="center" rowspan="1" colspan="1">4.43</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Rate/1,000 py</td><td align="center" rowspan="1" colspan="1">0.72</td><td align="center" rowspan="1" colspan="1">0.19</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.24</td><td align="center" rowspan="1" colspan="1">0.90</td></tr><tr><td colspan="7" align="left" rowspan="1">C&#x000f4;te d&#x02019;Ivoire: Taabo</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Deaths</td><td align="center" rowspan="1" colspan="1">22.74</td><td align="center" rowspan="1" colspan="1">63.22</td><td align="center" rowspan="1" colspan="1">8.24</td><td align="center" rowspan="1" colspan="1">22.79</td><td align="center" rowspan="1" colspan="1">2.99</td><td align="center" rowspan="1" colspan="1">8.56</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Rate/1,000 py</td><td align="center" rowspan="1" colspan="1">1.42</td><td align="center" rowspan="1" colspan="1">0.43</td><td align="center" rowspan="1" colspan="1">0.14</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">0.43</td><td align="center" rowspan="1" colspan="1">1.35</td></tr><tr><td colspan="7" align="left" rowspan="1">Ethiopia: Kilite-Awlaelo</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Deaths</td><td align="center" rowspan="1" colspan="1">1.83</td><td align="center" rowspan="1" colspan="1">2.22</td><td align="center" rowspan="1" colspan="1">1.22</td><td align="center" rowspan="1" colspan="1">1.00</td><td align="center" rowspan="1" colspan="1">0.70</td><td align="center" rowspan="1" colspan="1">4.93</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Rate/1,000 py</td><td align="center" rowspan="1" colspan="1">0.57</td><td align="center" rowspan="1" colspan="1">0.13</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.41</td></tr><tr><td colspan="7" align="left" rowspan="1">The Gambia: Farafenni</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Deaths</td><td align="center" rowspan="1" colspan="1">35.28</td><td align="center" rowspan="1" colspan="1">113.11</td><td align="center" rowspan="1" colspan="1">38.72</td><td align="center" rowspan="1" colspan="1">43.35</td><td align="center" rowspan="1" colspan="1">19.85</td><td align="center" rowspan="1" colspan="1">43.46</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Rate/1,000 py</td><td align="center" rowspan="1" colspan="1">1.06</td><td align="center" rowspan="1" colspan="1">0.33</td><td align="center" rowspan="1" colspan="1">0.15</td><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">0.55</td><td align="center" rowspan="1" colspan="1">1.15</td></tr><tr><td colspan="7" align="left" rowspan="1">Ghana: Navrongo</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Deaths</td><td align="center" rowspan="1" colspan="1">121.42</td><td align="center" rowspan="1" colspan="1">283.42</td><td align="center" rowspan="1" colspan="1">39.50</td><td align="center" rowspan="1" colspan="1">12.34</td><td align="center" rowspan="1" colspan="1">9.45</td><td align="center" rowspan="1" colspan="1">32.61</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Rate/1,000 py</td><td align="center" rowspan="1" colspan="1">0.42</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.14</td></tr><tr><td colspan="7" align="left" rowspan="1">Ghana: Dodowa</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Deaths</td><td align="center" rowspan="1" colspan="1">4.74</td><td align="center" rowspan="1" colspan="1">49.53</td><td align="center" rowspan="1" colspan="1">28.83</td><td align="center" rowspan="1" colspan="1">154.67</td><td align="center" rowspan="1" colspan="1">45.91</td><td align="center" rowspan="1" colspan="1">138.68</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Rate/1,000 py</td><td align="center" rowspan="1" colspan="1">0.28</td><td align="center" rowspan="1" colspan="1">0.14</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.21</td><td align="center" rowspan="1" colspan="1">0.26</td></tr><tr><td colspan="7" align="left" rowspan="1">India: Ballabgarh</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Deaths</td><td align="center" rowspan="1" colspan="1">5.41</td><td align="center" rowspan="1" colspan="1">17.89</td><td align="center" rowspan="1" colspan="1">3.64</td><td align="center" rowspan="1" colspan="1">4.25</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">5.38</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Rate/1,000 py</td><td align="center" rowspan="1" colspan="1">0.45</td><td align="center" rowspan="1" colspan="1">0.20</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.26</td></tr><tr><td colspan="7" align="left" rowspan="1">India: Vadu</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Deaths</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.91</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Rate/1,000 py</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td colspan="7" align="left" rowspan="1">Indonesia: Purworejo</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Deaths</td><td align="center" rowspan="1" colspan="1">2.42</td><td align="center" rowspan="1" colspan="1">3.13</td><td align="center" rowspan="1" colspan="1">2.00</td><td align="center" rowspan="1" colspan="1">4.34</td><td align="center" rowspan="1" colspan="1">5.64</td><td align="center" rowspan="1" colspan="1">13.50</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Rate/1,000 py</td><td align="center" rowspan="1" colspan="1">0.85</td><td align="center" rowspan="1" colspan="1">0.19</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.14</td><td align="center" rowspan="1" colspan="1">0.19</td></tr><tr><td colspan="7" align="left" rowspan="1">Kenya: Kilifi</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Deaths</td><td align="center" rowspan="1" colspan="1">38.53</td><td align="center" rowspan="1" colspan="1">90.21</td><td align="center" rowspan="1" colspan="1">36.03</td><td align="center" rowspan="1" colspan="1">14.84</td><td align="center" rowspan="1" colspan="1">3.97</td><td align="center" rowspan="1" colspan="1">12.72</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Rate/1,000 py</td><td align="center" rowspan="1" colspan="1">0.17</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.18</td></tr><tr><td colspan="7" align="left" rowspan="1">Kenya: Kisumu</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Deaths</td><td align="center" rowspan="1" colspan="1">672.20</td><td align="center" rowspan="1" colspan="1">1177.46</td><td align="center" rowspan="1" colspan="1">177.79</td><td align="center" rowspan="1" colspan="1">321.30</td><td align="center" rowspan="1" colspan="1">99.16</td><td align="center" rowspan="1" colspan="1">181.89</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Rate/1,000 py</td><td align="center" rowspan="1" colspan="1">0.38</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.14</td><td align="center" rowspan="1" colspan="1">0.17</td></tr><tr><td colspan="7" align="left" rowspan="1">Kenya: Nairobi</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Deaths</td><td align="center" rowspan="1" colspan="1">16.42</td><td align="center" rowspan="1" colspan="1">16.50</td><td align="center" rowspan="1" colspan="1">4.59</td><td align="center" rowspan="1" colspan="1">7.23</td><td align="center" rowspan="1" colspan="1">3.91</td><td align="center" rowspan="1" colspan="1">0.26</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Rate/1,000 py</td><td align="center" rowspan="1" colspan="1">0.80</td><td align="center" rowspan="1" colspan="1">0.18</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.15</td><td align="center" rowspan="1" colspan="1">0.05</td></tr><tr><td colspan="7" align="left" rowspan="1">Senegal: Niakhar</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Deaths</td><td align="center" rowspan="1" colspan="1">23.25</td><td align="center" rowspan="1" colspan="1">126.45</td><td align="center" rowspan="1" colspan="1">21.32</td><td align="center" rowspan="1" colspan="1">16.31</td><td align="center" rowspan="1" colspan="1">4.04</td><td align="center" rowspan="1" colspan="1">28.49</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Rate/1,000 py</td><td align="center" rowspan="1" colspan="1">1.05</td><td align="center" rowspan="1" colspan="1">0.33</td><td align="center" rowspan="1" colspan="1">0.15</td><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">0.22</td><td align="center" rowspan="1" colspan="1">0.68</td></tr><tr><td colspan="7" align="left" rowspan="1">South Africa: Africa Centre</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Deaths</td><td align="center" rowspan="1" colspan="1">8.67</td><td align="center" rowspan="1" colspan="1">13.84</td><td align="center" rowspan="1" colspan="1">7.37</td><td align="center" rowspan="1" colspan="1">9.44</td><td align="center" rowspan="1" colspan="1">1.53</td><td align="center" rowspan="1" colspan="1">7.22</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Rate/1,000 py</td><td align="center" rowspan="1" colspan="1">0.33</td><td align="center" rowspan="1" colspan="1">0.12</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.17</td></tr><tr><td colspan="7" align="left" rowspan="1">South Africa: Agincourt</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Deaths</td><td align="center" rowspan="1" colspan="1">12.45</td><td align="center" rowspan="1" colspan="1">29.39</td><td align="center" rowspan="1" colspan="1">19.45</td><td align="center" rowspan="1" colspan="1">54.40</td><td align="center" rowspan="1" colspan="1">7.56</td><td align="center" rowspan="1" colspan="1">4.93</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Rate/1,000 py</td><td align="center" rowspan="1" colspan="1">0.28</td><td align="center" rowspan="1" colspan="1">0.14</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">0.08</td></tr><tr><td colspan="7" align="left" rowspan="1">Vietnam: FilaBavi</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Deaths</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.55</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">2.46</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Rate/1,000 py</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.14</td></tr></tbody></table></table-wrap><p>We undertook a sensitivity analysis to examine the effects of the &#x02018;high&#x02019; and &#x02018;low&#x02019; InterVA-4 malaria settings across this large and diverse dataset. Re-running the InterVA-4 model with the &#x02018;high&#x02019; and &#x02018;low&#x02019; settings reversed at site level gave the results shown in <xref ref-type="fig" rid="F0003">Fig. 3</xref>. Incorrect use of the &#x02018;high&#x02019; setting in low malaria populations appeared to result in high relative rates of falsely attributed malaria, although the numbers involved would still be relatively small at the lowest endemicities. Conversely using the &#x02018;low&#x02019; setting in high malaria populations reduced the number of malaria assignments. Although the rate ratios changed less in high endemicity settings, the numbers of cases involved would be important with increasing rates.</p><fig id="F0003" position="float"><label>Fig. 3</label><caption><p>Sensitivity analysis showing the effect of choosing the &#x02018;wrong&#x02019; malaria endemicity setting (&#x02018;high&#x02019; and &#x02018;low&#x02019; reversed) in processing VA data using the InterVA-4 model, by site.</p></caption><graphic xlink:href="GHA-7-25369-g003"/></fig><p> <xref ref-type="table" rid="T0002">Table 2</xref> shows estimates of malaria-specific mortality rates for the countries with INDEPTH sites reporting here, for the under-5 and 5-plus age groups for comparison with other sources of malaria mortality estimates. INDEPTH estimates for countries with multiple sites were derived as population-weighted average rates.</p><table-wrap id="T0002" position="float"><label>Table 2</label><caption><p>Within-country estimates of malaria-specific mortality rates derived from WHO/CHERG (<xref rid="CIT0042" ref-type="bibr">42</xref>, <xref rid="CIT0043" ref-type="bibr">43</xref>), IHME (<xref rid="CIT0044" ref-type="bibr">44</xref>) compared with population-weighted average country rates from INDEPTH sites</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1"/><th align="center" colspan="2" rowspan="1">WHO/CHERG</th><th align="center" colspan="2" rowspan="1">IHME</th><th align="center" colspan="2" rowspan="1">INDEPTH</th></tr><tr><th align="left" rowspan="1" colspan="1"/><th colspan="6" rowspan="1"> <hr/> </th></tr><tr><th align="left" rowspan="1" colspan="1">Country</th><th align="center" rowspan="1" colspan="1">Under 5 years</th><th align="center" rowspan="1" colspan="1">5 years and over</th><th align="center" rowspan="1" colspan="1">Under 5 years</th><th align="center" rowspan="1" colspan="1">5 years and over</th><th align="center" rowspan="1" colspan="1">Under 5 years</th><th align="center" rowspan="1" colspan="1">5 years and over</th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">Bangladesh</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.004</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.006</td></tr><tr><td align="left" rowspan="1" colspan="1">Burkina Faso</td><td align="center" rowspan="1" colspan="1">9.94</td><td align="center" rowspan="1" colspan="1">0.15</td><td align="center" rowspan="1" colspan="1">8.34</td><td align="center" rowspan="1" colspan="1">1.19</td><td align="center" rowspan="1" colspan="1">6.08</td><td align="center" rowspan="1" colspan="1">1.00</td></tr><tr><td align="left" rowspan="1" colspan="1">C&#x000f4;te d&#x02019;Ivoire</td><td align="center" rowspan="1" colspan="1">6.92</td><td align="center" rowspan="1" colspan="1">0.13</td><td align="center" rowspan="1" colspan="1">5.49</td><td align="center" rowspan="1" colspan="1">0.92</td><td align="center" rowspan="1" colspan="1">5.04</td><td align="center" rowspan="1" colspan="1">0.57</td></tr><tr><td align="left" rowspan="1" colspan="1">Ethiopia</td><td align="center" rowspan="1" colspan="1">0.38</td><td align="center" rowspan="1" colspan="1">?</td><td align="center" rowspan="1" colspan="1">1.86</td><td align="center" rowspan="1" colspan="1">0.36</td><td align="center" rowspan="1" colspan="1">0.32</td><td align="center" rowspan="1" colspan="1">0.06</td></tr><tr><td align="left" rowspan="1" colspan="1">Ghana</td><td align="center" rowspan="1" colspan="1">2.90</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">2.99</td><td align="center" rowspan="1" colspan="1">0.58</td><td align="center" rowspan="1" colspan="1">2.40</td><td align="center" rowspan="1" colspan="1">0.30</td></tr><tr><td align="left" rowspan="1" colspan="1">India</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.53</td><td align="center" rowspan="1" colspan="1">0.03</td></tr><tr><td align="left" rowspan="1" colspan="1">Indonesia</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.80</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.74</td><td align="center" rowspan="1" colspan="1">0.08</td></tr><tr><td align="left" rowspan="1" colspan="1">Kenya</td><td align="center" rowspan="1" colspan="1">0.47</td><td align="center" rowspan="1" colspan="1">?</td><td align="center" rowspan="1" colspan="1">1.86</td><td align="center" rowspan="1" colspan="1">0.44</td><td align="center" rowspan="1" colspan="1">3.35</td><td align="center" rowspan="1" colspan="1">0.31</td></tr><tr><td align="left" rowspan="1" colspan="1">Senegal</td><td align="center" rowspan="1" colspan="1">2.39</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">1.96</td><td align="center" rowspan="1" colspan="1">0.59</td><td align="center" rowspan="1" colspan="1">2.95</td><td align="center" rowspan="1" colspan="1">0.39</td></tr><tr><td align="left" rowspan="1" colspan="1">The Gambia</td><td align="center" rowspan="1" colspan="1">4.31</td><td align="center" rowspan="1" colspan="1">0.14</td><td align="center" rowspan="1" colspan="1">5.55</td><td align="center" rowspan="1" colspan="1">0.46</td><td align="center" rowspan="1" colspan="1">2.34</td><td align="center" rowspan="1" colspan="1">0.61</td></tr><tr><td align="left" rowspan="1" colspan="1">Vietnam</td><td align="center" rowspan="1" colspan="1">0.004</td><td align="center" rowspan="1" colspan="1">0.000</td><td align="center" rowspan="1" colspan="1">0.003</td><td align="center" rowspan="1" colspan="1">0.013</td><td align="center" rowspan="1" colspan="1">0</td><td align="center" rowspan="1" colspan="1">0.015</td></tr></tbody></table></table-wrap><p>The Malaria Atlas Project (MAP) produced geo-referenced estimates of <italic>Plasmodium falciparum</italic> parasite rates (PfPR) across endemic areas for children aged 2&#x02013;10 years in 2010 (<xref rid="CIT0015" ref-type="bibr">15</xref>). Since all the INDEPTH HDSSs cover defined small areas, it was possible to extract a PfPR value for each endemic site from the MAP data. Where sites covered more than one cell of the MAP surface, all the cells relating to the site were averaged. Data were available for 14 sites with childhood malaria mortality data; data were not available for seven sites in Vietnam, India, Bangladesh and Ethiopia, presumably because of very low or uncertain endemicity. <xref ref-type="fig" rid="F0004">Figure 4</xref> shows the correlation between per-site malaria mortality rates for the 1&#x02013;14 year age group as determined by InterVA-4 and the MAP PfPR values for the same geographic locations. The line in <xref ref-type="fig" rid="F0004">Fig. 4</xref> represents a highly significant correlation (<italic>R</italic> ²=0.69, <italic>p</italic>=0.002), fitting the relationship:<disp-formula id="FD1"><mml:math id="UM1"><mml:mrow><mml:mtext>Malaria mortality rate</mml:mtext><mml:mo>=</mml:mo><mml:msup><mml:mtext>e</mml:mtext><mml:mrow><mml:mo>[</mml:mo><mml:mo stretchy="false">(</mml:mo><mml:mtext>PfPR</mml:mtext><mml:mo>&#x000d7;</mml:mo><mml:mn>0.6274</mml:mn><mml:mo stretchy="false">)</mml:mo><mml:mo>+</mml:mo><mml:mn>0.7023</mml:mn><mml:mo stretchy="false">]</mml:mo></mml:mrow></mml:msup></mml:mrow></mml:math></disp-formula> </p><fig id="F0004" position="float"><label>Fig. 4</label><caption><p>Scatter plot of age&#x02013;sex&#x02013;time standardised InterVA malaria mortality rates per 1,000 person-years for children aged 1&#x02013;14 years versus <italic>Plasmodium falciparum</italic> parasite rate data for children aged 2&#x02013;10 years, for 14 INDEPTH HDSS sites reporting malaria mortality which also had geo-referenced parasite rate data for 2010 in the Malaria Atlas Project (<xref rid="CIT0015" ref-type="bibr">15</xref>). Line shows correlation, <italic>R</italic> ²=0.56. (1. Africa Centre, South Africa; 2. Agincourt, South Africa; 3. Nairobi, Kenya; 4. Purworejo, Indonesia; 5. Bandarban, Bangladesh; 6. Kilifi, Kenya; 7. Dodowa, Ghana; 8. Navrongo, Ghana; 9. Farafenni, The Gambia; 10. Ouagadougou, Burkina Faso; 11. Niakhar, Senegal; 12. Taabo, C&#x000f4;te d&#x02019;Ivoire; 13. Kisumu, Kenya; 14. Nouna, Burkina Faso).</p></caption><graphic xlink:href="GHA-7-25369-g004"/></fig><p>An important area of uncertainty in malaria epidemiology is the ratio of malaria-specific mortality rates between children and adults. Seventeen sites recorded malaria deaths in both under-15 and over-15 year age categories. Apart from one outlier (Dodowa, Ghana, where the malaria-specific mortality rate ratio for over-15: under-15 age categories was 2.5), in the remaining 16 sites the malaria-specific mortality rate ratios for over-15:under-15 age categories were in the range 0.05 to 0.82, while overall malaria-specific mortality rates ranged from 0.018 to 2.47 per 1,000 person-years. <xref ref-type="fig" rid="F0005">Figure 5</xref> shows the correlation between adult and child malaria rates for these 17 sites, shown on logarithmic scales for clarity. As expected, the sites from West Africa dominate the top-right quadrant, together with Kisumu, on the shores of Lake Victoria in Kenya. Other African and Asian sites largely occupy the lower-left quadrant, with the Chakaria site in Bangladesh showing very low malaria mortality rates for both adults and children. The per-site correlation (represented by the line in <xref ref-type="fig" rid="F0005">Fig. 5</xref>) between age&#x02013;sex&#x02013;time standardised adult and child malaria mortality rates was highly significant (<italic>R</italic> ²=0.65, <italic>p</italic>=0.0001), fitting the relationship: <disp-formula id="FD2"><mml:math id="UM2"><mml:mrow><mml:mtext>Adult malaria mortality rate</mml:mtext><mml:mo>=</mml:mo><mml:msup><mml:mtext>e</mml:mtext><mml:mrow><mml:mo stretchy="false">[</mml:mo><mml:mo stretchy="false">(</mml:mo><mml:mtext>child malaria mortality rate</mml:mtext><mml:mo>&#x000d7;</mml:mo><mml:mtext>1</mml:mtext><mml:mn>.002</mml:mn><mml:mo stretchy="false">)</mml:mo><mml:mo>-</mml:mo><mml:mn>1.157</mml:mn><mml:mo stretchy="false">]</mml:mo></mml:mrow></mml:msup></mml:mrow></mml:math></disp-formula> </p><fig id="F0005" position="float"><label>Fig. 5</label><caption><p>Scatter plot of age&#x02013;sex&#x02013;time standardised malaria mortality rates per 1,000 person-years for adults (15 years and over) and children (under 15 years), for 17 INDEPTH HDSS sites reporting malaria mortality among adults and children. Line shows correlation, <italic>R</italic> ²=0.65.</p></caption><graphic xlink:href="GHA-7-25369-g005"/></fig></sec><sec sec-type="discussion" id="S0003"><title>Discussion</title><p>These results represent widely-based evidence on malaria mortality, which has not previously been documented at the population level on this scale, using standardised methods. The interpretation of findings at individual sites depends on local characteristics (<xref rid="CIT0016" ref-type="bibr">16</xref>&#x02013;<xref rid="CIT0036" ref-type="bibr">36</xref>). Two sites, Ouagadougou in Burkina Faso and Nairobi in Kenya, followed urban populations and recorded lower levels of malaria than some of their rural neighbours. Bandarban in Bangladesh is located in a frontier zone close to the Myanmar border, which may explain higher rates of malaria compared with other sites in Bangladesh; this is consistent with WHO malaria mapping for Bangladesh (<xref rid="CIT0037" ref-type="bibr">37</xref>). The very low overall levels of malaria mortality in Bangladesh are not only consistent with expectations, but form an important part of these analyses in that they suggest our methods are capable of assigning malaria as a cause of death with high specificity. Kisumu in Kenya is located on the shores of Lake Victoria, in an area known to have higher malaria transmission than most other parts of the country, such as the coastal area around Kilifi (<xref rid="CIT0038" ref-type="bibr">38</xref>). Kilite-Awlaelo is located in the Ethiopian highlands, at an altitude around 2,000 m above sea level, at which malaria is typically unstable and epidemic in nature. The two South African sites are located on the margins of malaria transmission, and some of the relatively few cases that occurred may reflect travel, for example to neighbouring Mozambique (<xref rid="CIT0039" ref-type="bibr">39</xref>).</p><p>The validity of VA cause of death assignment specifically for malaria is difficult to determine precisely. The InterVA model has previously been used in a WHO study of malaria treatment, showing a significant difference in malaria-specific mortality following a treatment delivery intervention (<xref rid="CIT0040" ref-type="bibr">40</xref>). A review of VA methodological validations in relation to hospital data found some examples relating to malaria, but a generalisable formal validation for malaria mortality remains elusive (<xref rid="CIT0041" ref-type="bibr">41</xref>). In principle validity of VA methods for malaria as a cause of death could be established in a large VA dataset from an endemic area which included systematic parasitaemia testing across all age groups. Operationally this could be incorporated in a minimally-invasive autopsy approach (<xref rid="CIT0042" ref-type="bibr">42</xref>). The Population Health Metrics Research Consortium (PHMRC) collected a &#x02018;gold standard&#x02019; VA dataset of 12,530 tertiary facility cases, which contained 216 cases meeting the PHMRC definitions of a malaria death (basically diagnoses based on parasitaemia and fever) (<xref rid="CIT0043" ref-type="bibr">43</xref>, <xref rid="CIT0044" ref-type="bibr">44</xref>). Unfortunately however there were no data on the presence or absence of malaria parasitaemia in cases attributed to other causes, nor on parasite species for the malaria cases. Most (64%) of the adult malaria deaths in this series came from hospitals in India, while the childhood cases were mainly from Dar-es-Salaam city (88%), though it should be noted that this study did not aim to represent any particular population. Only 25% of the malaria deaths mentioned the word &#x02018;malaria&#x02019; in the open-ended part of the subsequent VA interview (which did not contain any specific question on malaria), while 69% of malaria case VAs for adults and 54% for children reported severe respiratory symptoms. This may partly reflect the tertiary facility settings of these cases, where some cases may have progressed to respiratory complications of malaria (<xref rid="CIT0045" ref-type="bibr">45</xref>), or VA respondents may simply have noted hospital treatment for breathing difficulties in the trajectory towards death (<xref rid="CIT0046" ref-type="bibr">46</xref>). Consequently, the PHMRC dataset is not particularly useful in terms of validating VA in general for malaria.</p><p>The WHO 2012 VA standard (<xref rid="CIT0008" ref-type="bibr">8</xref>) includes indicators relating to positive or negative malaria test results during the final illness, as well as other relevant symptomatic parameters. However, because these data were collected before the WHO 2012 standard was directly implemented for data capture, specific responses for these indicators were missing in over 90% of cases. However, a previous sensitivity analysis showed that InterVA-4 was generally relatively robust in relation to missing data items (<xref rid="CIT0046" ref-type="bibr">46</xref>). Nevertheless, the malaria-specific outputs here, using the WHO 2012 standard and the corresponding InterVA-4 model, show huge differences between locations and age groups, as might be expected. These plausible patterns suggest that there may be at least a reasonable degree of validity in terms of InterVA-4&#x02019;s assignment of malaria deaths. The application of a standard probabilistic model such as InterVA-4 at least guarantees that inter-site differences are reflections of variations in the VA source data (<xref rid="CIT0013" ref-type="bibr">13</xref>). If, alternatively, physicians at each site were used to assign cause of death, it would be easy for inter- and intra-physician variations to contribute to apparent differences between sites and over time. This is the first time such a large VA dataset relating to malaria has been compiled that spans complete populations in Africa and Asia, covers a wide spectrum of endemicity, and uses standardised cause of death attribution. The sensitivity analysis reported here is important in justifying the design assumptions in InterVA-4 that require local settings for malaria (and HIV) endemicity. The crossover region between the &#x02018;high&#x02019; and &#x02018;low&#x02019; settings, recommended at 1%, has been seen as a difficult concept by some InterVA-4 users. However, the sensitivity analysis shown in <xref ref-type="fig" rid="F0003">Fig. 3</xref> suggests that this setting is both important and appropriate, and analogous to a clinician&#x02019;s local knowledge of malaria endemicity, irrespective of the history and symptomatology of the next patient.</p><p>There are other major pieces of work describing malaria mortality in Africa and Asia, using totally different methods, with which these findings can be compared and contrasted. The WHO World Malaria Report 2013 (<xref rid="CIT0006" ref-type="bibr">6</xref>) sets out WHO&#x02019;s most recent compilation of malaria reports from its member countries, together with associated data estimates in the WHO Global Health Observatory (<xref rid="CIT0047" ref-type="bibr">47</xref>) and, for children, from the Child Epidemiology Reference Group (CHERG) (<xref rid="CIT0048" ref-type="bibr">48</xref>). The Institute of Health Metrics and Evaluation (IHME) has also published global and country estimates of malaria mortality covering a similar time period, based on mathematical modelling of available data (<xref rid="CIT0049" ref-type="bibr">49</xref>). Both of these sources take the approach of gathering whatever malaria mortality data may be available across all endemic areas (to which this dataset now adds), and then making best estimates to fill in the considerable gaps in the available data.</p><p> <xref ref-type="table" rid="T0002">Table 2</xref> enables comparisons of malaria-specific mortality rates for the countries with INDEPTH sites reporting here, for the under-5 and 5-plus age groups, with other sources of estimates. South Africa is not included because the majority of the country is malaria-free, while the two INDEPTH sites represent marginal transmission areas, making national estimates difficult to interpret. WHO and CHERG publish separate data estimates for all-age malaria deaths and under-5 malaria deaths, respectively; while these are largely congruent, allowing the calculation of 5-plus deaths, for Kenya and Ethiopia the number of under-5 deaths exceeded total deaths, so that no rate could be calculated for the 5-plus age group. Comparisons between these three data sources have to be interpreted with care. The WHO/CHERG and IHME numbers come from estimates based on such data as are available, modelled to represent the national situation as far as is possible, and may include facility and community sources, as well as diverse methods of cause of death assignment. The INDEPTH numbers come from the specific HDSS populations as described above, which are not intended to be nationally representative, but which are collected and processed in a standardised way across the various countries represented. In the case of Kenya, for example, the higher INDEPTH rate for under-5s reflects high malaria mortality in the Kisumu area. While it would be inappropriate to over-interpret comparisons of the rates presented in <xref ref-type="table" rid="T0002">Table 2</xref>, it is clear that there are substantial similarities between all three sources. IHME and INDEPTH figures tend towards higher rates for the 5-plus age group, though the reasons for this are not clear. In INDEPTH&#x02019;s case, InterVA-4 appears to be detecting a number of acute febrile illnesses among older people and attributing them as malaria; but there is absolutely no associated biomedical evidence that these deaths are indeed directly due to malaria.</p><p>However, <xref ref-type="fig" rid="F0004">Fig. 4</xref> showed a strong correlation between InterVA-4 estimates of childhood malaria mortality and geo-referenced parasite prevalence estimates from MAP (<xref rid="CIT0015" ref-type="bibr">15</xref>). There are three possible consequences to consider. Firstly, if one accepts the validity of the parasite prevalence estimates, then the observed correlation suggests that for children (notwithstanding the slightly different age groups of 1&#x02013;14 years for mortality and 2&#x02013;10 years for parasite prevalence), InterVA-4 is capturing a directly related pattern of malaria mortality, across a 100-fold range of endemicity. The second option is to accept the validity of the InterVA-4 malaria mortality findings reported here, in which case they add veracity to the parasite prevalence estimates. Thirdly, if both the InterVA-4 and MAP findings are considered to be reasonably valid, then this correlation establishes an interesting relationship between childhood parasite prevalence and malaria mortality. This relationship seems to hold over a wide range of sites, even though it might be reasonable to presume that local factors such as the effectiveness of treatment and control programmes could also play a part. Previous work (among hospitalised cases) in Tanzania showed relationships between age, transmission intensity and malaria mortality (<xref rid="CIT0050" ref-type="bibr">50</xref>). Another modelling study sought to establish relationships between malaria transmission and mortality, though starting from a rather disparate group of datasets (<xref rid="CIT0051" ref-type="bibr">51</xref>).</p><p> <xref ref-type="fig" rid="F0005">Figure 5</xref> showed a strong correlation between InterVA-4 adult and childhood malaria mortality rates at the site level. If InterVA-4 were generally misclassifying a wide range of acute adult febrile illnesses as malaria, this would not be the expected pattern. If there were appreciable misclassification, the so-called &#x02018;malaria&#x02019; deaths in adults might be expected to occur at a rate largely independent of childhood malaria mortality, in the absence of any hypothesis as to other causes of acute adult febrile mortality that happened to correlate geographically with childhood malaria. However, there were clearly much higher rates of what InterVA-4 was calling &#x02018;malaria&#x02019; among adults in West Africa, where malaria transmission is known to be the highest in the world. A more detailed analysis of malaria mortality by age from the Kisumu site in Kenya showed complex and changing relationships between malaria mortality and age (<xref rid="CIT0052" ref-type="bibr">52</xref>). Because malaria surveillance among older people has generally not been given high priority, there appears to be a need for further population-based research to further resolve this question.</p><p>The public availability of these malaria mortality data creates interesting opportunities for further analyses. Apart from contributing to the overall body of malaria mortality data, there are several other ways in which they may be specifically useful. While one can debate the generalisability of HDSS sites (<xref rid="CIT0053" ref-type="bibr">53</xref>), the cross-site relationships established here between gridded parasite prevalence data and childhood malaria mortality, and between child and adult malaria mortality rates, could well be incorporated into wider estimations of malaria mortality.</p></sec><sec sec-type="conclusions" id="S0004"><title>Conclusions</title><p>Measuring malaria mortality effectively continues to be a global problem. As remarked in the context of malaria transmission modelling (<xref rid="CIT0054" ref-type="bibr">54</xref>), malaria mortality events frequently fall under the radar of health information systems. The data presented here, from a wide range of INDEPTH HDSSs across Africa and Asia, demonstrate the value of detailed longitudinal surveillance in defined populations, rather than relying on more disparate sources. VA may not be an ideal tool for tracking malaria, but nevertheless the malaria-specific mortality rate estimates obtained here using the WHO 2012 standard and the InterVA-4 model closely correspond to other sources of estimates, despite the 1:10,000 range in the magnitude of rates measured using the same methods in different settings. More widespread use of these population-based approaches would add considerably to global understanding of malaria, and thereby inform control and elimination programmes.</p></sec>

HIV/AIDS-related mortality in Africa and Asia: evidence from INDEPTH health and demographic surveillance system sites

<sec id="st1"><title>Background</title><p>As the HIV/AIDS pandemic has evolved over recent decades, Africa has been the most affected region, even though a large proportion of HIV/AIDS deaths have not been documented at the individual level. Systematic application of verbal autopsy (VA) methods in defined populations provides an opportunity to assess the mortality burden of the pandemic from individual data.</p></sec><sec id="st2"><title>Objective</title><p>To present standardised comparisons of HIV/AIDS-related mortality at sites across Africa and Asia, including closely related causes of death such as pulmonary tuberculosis (PTB) and pneumonia.</p></sec><sec id="st3"><title>Design</title><p>Deaths related to HIV/AIDS were extracted from individual demographic and VA data from 22 INDEPTH sites across Africa and Asia. VA data were standardised to WHO 2012 standard causes of death assigned using the InterVA-4 model. Between-site comparisons of mortality rates were standardised using the INDEPTH 2013 standard population.</p></sec><sec id="st4"><title>Results</title><p>The dataset covered a total of 10,773 deaths attributed to HIV/AIDS, observed over 12,204,043 person-years. HIV/AIDS-related mortality fractions and mortality rates varied widely across Africa and Asia, with highest burdens in eastern and southern Africa, and lowest burdens in Asia. There was evidence of rapidly declining rates at the sites with the heaviest burdens. HIV/AIDS mortality was also strongly related to PTB mortality. On a country basis, there were strong similarities between HIV/AIDS mortality rates at INDEPTH sites and those derived from modelled estimates.</p></sec><sec id="st5"><title>Conclusions</title><p>Measuring HIV/AIDS-related mortality continues to be a challenging issue, all the more so as anti-retroviral treatment programmes alleviate mortality risks. The congruence between these results and other estimates adds plausibility to both approaches. These data, covering some of the highest mortality observed during the pandemic, will be an important baseline for understanding the future decline of HIV/AIDS.</p></sec>

<contrib contrib-type="author"><name><surname>Streatfield</surname><given-names>P. Kim</given-names></name><xref ref-type="aff" rid="AF0001">1</xref><xref ref-type="aff" rid="AF0002">2</xref><xref ref-type="aff" rid="AF0003">3</xref></contrib><contrib contrib-type="author"><name><surname>Khan</surname><given-names>Wasif A.</given-names></name><xref ref-type="aff" rid="AF0002">2</xref><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0004">4</xref></contrib><contrib contrib-type="author"><name><surname>Bhuiya</surname><given-names>Abbas</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0005">5</xref><xref ref-type="aff" rid="AF0006">6</xref></contrib><contrib contrib-type="author"><name><surname>Hanifi</surname><given-names>Syed M.A.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0005">5</xref><xref ref-type="aff" rid="AF0006">6</xref></contrib><contrib contrib-type="author"><name><surname>Alam</surname><given-names>Nurul</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0007">7</xref><xref ref-type="aff" rid="AF0008">8</xref></contrib><contrib contrib-type="author"><name><surname>Millogo</surname><given-names>Ourohir&#x000e9;</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0009">9</xref><xref ref-type="aff" rid="AF0010">10</xref></contrib><contrib contrib-type="author"><name><surname>Si&#x000e9;</surname><given-names>Ali</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0009">9</xref><xref ref-type="aff" rid="AF0010">10</xref></contrib><contrib contrib-type="author"><name><surname>Zabr&#x000e9;</surname><given-names>Pascal</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0009">9</xref><xref ref-type="aff" rid="AF0010">10</xref></contrib><contrib contrib-type="author"><name><surname>Rossier</surname><given-names>Clementine</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0011">11</xref><xref ref-type="aff" rid="AF0012">12</xref><xref ref-type="aff" rid="AF0013">13</xref></contrib><contrib contrib-type="author"><name><surname>Soura</surname><given-names>Abdramane B.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0011">11</xref><xref ref-type="aff" rid="AF0012">12</xref></contrib><contrib contrib-type="author"><name><surname>Bonfoh</surname><given-names>Bassirou</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0014">14</xref><xref ref-type="aff" rid="AF0015">15</xref></contrib><contrib contrib-type="author"><name><surname>Kone</surname><given-names>Siaka</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0014">14</xref><xref ref-type="aff" rid="AF0015">15</xref></contrib><contrib contrib-type="author"><name><surname>Ngoran</surname><given-names>Eliezer K.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0014">14</xref><xref ref-type="aff" rid="AF0016">16</xref></contrib><contrib contrib-type="author"><name><surname>Utzinger</surname><given-names>Juerg</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0014">14</xref><xref ref-type="aff" rid="AF0017">17</xref></contrib><contrib contrib-type="author"><name><surname>Abera</surname><given-names>Semaw F.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0018">18</xref><xref ref-type="aff" rid="AF0019">19</xref></contrib><contrib contrib-type="author"><name><surname>Melaku</surname><given-names>Yohannes A.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0018">18</xref><xref ref-type="aff" rid="AF0019">19</xref></contrib><contrib contrib-type="author"><name><surname>Weldearegawi</surname><given-names>Berhe</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0018">18</xref><xref ref-type="aff" rid="AF0019">19</xref></contrib><contrib contrib-type="author"><name><surname>Gomez</surname><given-names>Pierre</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0020">20</xref><xref ref-type="aff" rid="AF0021">21</xref></contrib><contrib contrib-type="author"><name><surname>Jasseh</surname><given-names>Momodou</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0020">20</xref><xref ref-type="aff" rid="AF0021">21</xref></contrib><contrib contrib-type="author"><name><surname>Ansah</surname><given-names>Patrick</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0022">22</xref><xref ref-type="aff" rid="AF0023">23</xref></contrib><contrib contrib-type="author"><name><surname>Azongo</surname><given-names>Daniel</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0022">22</xref><xref ref-type="aff" rid="AF0023">23</xref></contrib><contrib contrib-type="author"><name><surname>Kondayire</surname><given-names>Felix</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0022">22</xref><xref ref-type="aff" rid="AF0023">23</xref></contrib><contrib contrib-type="author"><name><surname>Oduro</surname><given-names>Abraham</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0022">22</xref><xref ref-type="aff" rid="AF0023">23</xref></contrib><contrib contrib-type="author"><name><surname>Amu</surname><given-names>Alberta</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0024">24</xref><xref ref-type="aff" rid="AF0025">25</xref></contrib><contrib contrib-type="author"><name><surname>Gyapong</surname><given-names>Margaret</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0024">24</xref><xref ref-type="aff" rid="AF0025">25</xref></contrib><contrib contrib-type="author"><name><surname>Kwarteng</surname><given-names>Odette</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0024">24</xref><xref ref-type="aff" rid="AF0025">25</xref></contrib><contrib contrib-type="author"><name><surname>Kant</surname><given-names>Shashi</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0026">26</xref><xref ref-type="aff" rid="AF0027">27</xref></contrib><contrib contrib-type="author"><name><surname>Pandav</surname><given-names>Chandrakant S.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0026">26</xref><xref ref-type="aff" rid="AF0027">27</xref></contrib><contrib contrib-type="author"><name><surname>Rai</surname><given-names>Sanjay K.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0026">26</xref><xref ref-type="aff" rid="AF0027">27</xref></contrib><contrib contrib-type="author"><name><surname>Juvekar</surname><given-names>Sanjay</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0028">28</xref><xref ref-type="aff" rid="AF0029">29</xref></contrib><contrib contrib-type="author"><name><surname>Muralidharan</surname><given-names>Veena</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0028">28</xref><xref ref-type="aff" rid="AF0029">29</xref></contrib><contrib contrib-type="author"><name><surname>Wahab</surname><given-names>Abdul</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0030">30</xref><xref ref-type="aff" rid="AF0031">31</xref></contrib><contrib contrib-type="author"><name><surname>Wilopo</surname><given-names>Siswanto</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0030">30</xref><xref ref-type="aff" rid="AF0031">31</xref></contrib><contrib contrib-type="author"><name><surname>Bauni</surname><given-names>Evasius</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0032">32</xref><xref ref-type="aff" rid="AF0033">33</xref></contrib><contrib contrib-type="author"><name><surname>Mochamah</surname><given-names>George</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0032">32</xref><xref ref-type="aff" rid="AF0033">33</xref></contrib><contrib contrib-type="author"><name><surname>Ndila</surname><given-names>Carolyne</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0032">32</xref><xref ref-type="aff" rid="AF0033">33</xref></contrib><contrib contrib-type="author"><name><surname>Williams</surname><given-names>Thomas N.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0032">32</xref><xref ref-type="aff" rid="AF0033">33</xref><xref ref-type="aff" rid="AF0034">34</xref></contrib><contrib contrib-type="author"><name><surname>Khagayi</surname><given-names>Sammy</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0035">35</xref><xref ref-type="aff" rid="AF0036">36</xref></contrib><contrib contrib-type="author"><name><surname>Laserson</surname><given-names>Kayla F.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0035">35</xref><xref ref-type="aff" rid="AF0036">36</xref></contrib><contrib contrib-type="author"><name><surname>Nyaguara</surname><given-names>Amek</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0035">35</xref><xref ref-type="aff" rid="AF0036">36</xref></contrib><contrib contrib-type="author"><name><surname>Van Eijk</surname><given-names>Anna M.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0035">35</xref><xref ref-type="aff" rid="AF0036">36</xref></contrib><contrib contrib-type="author"><name><surname>Ezeh</surname><given-names>Alex</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0037">37</xref><xref ref-type="aff" rid="AF0038">38</xref></contrib><contrib contrib-type="author"><name><surname>Kyobutungi</surname><given-names>Catherine</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0037">37</xref><xref ref-type="aff" rid="AF0038">38</xref></contrib><contrib contrib-type="author"><name><surname>Wamukoya</surname><given-names>Marylene</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0037">37</xref><xref ref-type="aff" rid="AF0038">38</xref></contrib><contrib contrib-type="author"><name><surname>Chihana</surname><given-names>Menard</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0039">39</xref><xref ref-type="aff" rid="AF0038">38</xref></contrib><contrib contrib-type="author"><name><surname>Crampin</surname><given-names>Amelia</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0039">39</xref><xref ref-type="aff" rid="AF0040">40</xref><xref ref-type="aff" rid="AF0041">41</xref></contrib><contrib contrib-type="author"><name><surname>Price</surname><given-names>Alison</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0039">39</xref><xref ref-type="aff" rid="AF0040">40</xref><xref ref-type="aff" rid="AF0041">41</xref></contrib><contrib contrib-type="author"><name><surname>Delaunay</surname><given-names>Val&#x000e9;rie</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0042">42</xref><xref ref-type="aff" rid="AF0043">43</xref></contrib><contrib contrib-type="author"><name><surname>Diallo</surname><given-names>Aldiouma</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0042">42</xref><xref ref-type="aff" rid="AF0043">43</xref></contrib><contrib contrib-type="author"><name><surname>Douillot</surname><given-names>Laetitia</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0042">42</xref><xref ref-type="aff" rid="AF0043">43</xref></contrib><contrib contrib-type="author"><name><surname>Sokhna</surname><given-names>Cheikh</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0042">42</xref><xref ref-type="aff" rid="AF0043">43</xref></contrib><contrib contrib-type="author"><name><surname>G&#x000f3;mez-Oliv&#x000e9;</surname><given-names>F. Xavier</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0044">44</xref><xref ref-type="aff" rid="AF0045">45</xref></contrib><contrib contrib-type="author"><name><surname>Mee</surname><given-names>Paul</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0044">44</xref><xref ref-type="aff" rid="AF0045">45</xref></contrib><contrib contrib-type="author"><name><surname>Tollman</surname><given-names>Stephen M.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0044">44</xref><xref ref-type="aff" rid="AF0043">43</xref><xref ref-type="aff" rid="AF0046">46</xref></contrib><contrib contrib-type="author"><name><surname>Herbst</surname><given-names>Kobus</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0047">47</xref><xref ref-type="aff" rid="AF0048">48</xref></contrib><contrib contrib-type="author"><name><surname>Mossong</surname><given-names>Jo&#x000eb;l</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0047">47</xref><xref ref-type="aff" rid="AF0048">48</xref><xref ref-type="aff" rid="AF0049">49</xref></contrib><contrib contrib-type="author"><name><surname>Chuc</surname><given-names>Nguyen T.K.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0050">50</xref><xref ref-type="aff" rid="AF0051">51</xref></contrib><contrib contrib-type="author"><name><surname>Arthur</surname><given-names>Samuelina S.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref></contrib><contrib contrib-type="author"><name><surname>Sankoh</surname><given-names>Osman A.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="aff" rid="AF0052">52</xref><xref ref-type="aff" rid="AF0053">53</xref><xref ref-type="corresp" rid="cor1">*</xref></contrib><contrib contrib-type="author"><name><surname>Byass</surname><given-names>Peter</given-names></name><xref ref-type="aff" rid="AF0045">45</xref><xref ref-type="aff" rid="AF0054">54</xref></contrib>

Global Health Action

<p>The human immunodeficiency virus (HIV) and the consequent acquired immune deficiency syndrome (AIDS) caused a globally devastating pandemic starting in the late twentieth century. This pandemic caused mortality of such a magnitude as to distort population age&#x02013;sex distributions in the worst affected areas (<xref rid="CIT0001" ref-type="bibr">1</xref>). Now with the advent and roll-out of effective treatment for case management, the situation is improving (<xref rid="CIT0002" ref-type="bibr">2</xref>). However, because the pandemic most affected those areas of the world where reliable health data are scarcest, there remain large uncertainties about measuring the impact of HIV/AIDS, with many assessments relying on modelled estimates (<xref rid="CIT0003" ref-type="bibr">3</xref>). Thus, in reality, many millions of people attributed with HIV infection and/or AIDS mortality over the course of the pandemic were neither tested for the virus, nor had their deaths certified by physicians.</p><p>In the absence of laboratory testing and physician diagnosis, one way of determining the magnitude of HIV/AIDS-related mortality is by using verbal autopsy (VA), involving a structured interview with family or friends after a death (<xref rid="CIT0004" ref-type="bibr">4</xref>). The interview material is then used to assign the cause of death. In many settings, particularly at earlier stages, physicians made such assessments of VA data. Recently, it has become more common to use computerised models to attribute cause of death, which are faster, cheaper, and more consistent. Neither approach can be regarded as absolutely correct. Following a VA interview, assigning a death as due to HIV/AIDS is not entirely straightforward, because HIV-infected people may die of a variety of causes. As well as the wasting syndromes typical of AIDS deaths, other causes of death, including particularly pulmonary tuberculosis (PTB) and pneumonia, occur at higher rates and in different age groups among HIV-infected people.</p><p>In this paper, we present HIV/AIDS-specific mortality rates as determined by computer-interpreted VA from 22 INDEPTH Network Health and Demographic Surveillance Sites (HDSS) across Africa and Asia (<xref rid="CIT0005" ref-type="bibr">5</xref>). These findings are complemented with the corresponding rates for PTB and pneumonia. Although these HDSSs are not designed to form a representative network, each one follows a geographically defined population longitudinally, systematically recording all death events and undertaking verbal autopsies on all deaths that occur. Our aim is to present the HIV/AIDS mortality patterns at each site, comparing these community-level findings with other estimated information on HIV/AIDS in Africa and Asia.</p><sec sec-type="methods" id="S0001"><title>Methods</title><p>The overall INDEPTH dataset (<xref rid="CIT0006" ref-type="bibr">6</xref>) from which these HIV/AIDS-specific analyses are drawn is described in detail elsewhere (<xref rid="CIT0007" ref-type="bibr">7</xref>). The methods used are summarised in <xref ref-type="boxed-text" rid="B0001">Box 1</xref>. Briefly, it documents 111,910 deaths in 12,204,043 person-years of observation across 22 sites. The Karonga site in Malawi did not contribute VAs for children.</p><p> <italic>Box 1</italic>. Summary of methodology based on the detailed description in the introductory paper (<xref rid="CIT0007" ref-type="bibr">7</xref>) </p><boxed-text id="B0001" position="float"><p> <bold>Age&#x02013;sex&#x02013;time standardisation</bold> </p><p>To avoid effects of differences and changes in age&#x02013;sex structures of populations, mortality fractions and rates have been adjusted using the INDEPTH 2013 population standard (<xref rid="CIT0008" ref-type="bibr">8</xref>). A weighting factor was calculated for each site, age group, sex, and year category in relation to the standard for the corresponding age group and sex, and incorporated into the overall dataset. This is referred to in this paper as age&#x02013;sex&#x02013;time standardisation in the contexts where it is used.</p><p> <bold>Cause of death assignment</bold> </p><p>The InterVA-4 (version 4.02) probabilistic model was used for all the cause of death assignments in the overall dataset (<xref rid="CIT0009" ref-type="bibr">9</xref>). InterVA-4 is fully compliant with the WHO 2012 Verbal Autopsy standard and generates causes of death categorised by ICD-10 groups (<xref rid="CIT0010" ref-type="bibr">10</xref>). The data reported here were collected before the WHO 2012 VA standard was available, but were transformed into the WHO 2012 and InterVA-4 format to optimise cross-site standardisation in cause of death attribution. For a small proportion of deaths, VA interviews were not successfully completed; a few others contained inadequate information to arrive at a cause of death. InterVA-4 assigns causes of death (maximum 3) with associated likelihoods; thus cases for which likely causes did not total to 100% were also assigned a residual indeterminate component. This served as a means of encapsulating uncertainty in cause of death at the individual level within the overall dataset, as well as accounting for 100% of every death.</p><p> <bold>Overall dataset</bold> </p><p>The overall public-domain dataset (<xref rid="CIT0006" ref-type="bibr">6</xref>) thus contains between one and four records for each death, with the sum of likelihoods for each individual being unity. Each record includes a specific cause of death, its likelihood and its age&#x02013;sex&#x02013;time weighting. </p></boxed-text><p>The InterVA-4 &#x02018;high&#x02019; HIV/AIDS setting was used for sites in Kenya, Malawi, and South Africa. All other sites used the &#x02018;low&#x02019; setting; the &#x02018;very low&#x02019; setting was not used. The InterVA-4 guideline is that the &#x02018;high&#x02019; setting is appropriate for an expected HIV/AIDS cause-specific mortality fraction (CSMF) higher than about 1%, though it does not result in any great dichotomisation of outputs; the clinical equivalent is a physician&#x02019;s knowledge that his/her current case comes from a setting where HIV/AIDS is more or less likely, irrespective of that current case&#x02019;s particular symptoms. The validity of the InterVA-4 model in assigning HIV/AIDS as a cause of death in relation to HIV sero-status has been extensively explored in conjunction with the ALPHA Network (<xref rid="CIT0011" ref-type="bibr">11</xref>), and found to be over 90% specific. Sensitivity is more difficult to assess, since not all people infected with HIV evidently die of AIDS. The same validation exercise pointed to large numbers of cases of PTB and pneumonia as causes of death among the HIV-positive.</p><p>Deaths assigned to HIV/AIDS, and the closely related causes of PTB and pneumonia, were extracted from the overall data set together with data on person-time exposed by site, year, age, and sex. As each HDSS covers a total population, rather than a sample, uncertainty intervals are not shown.</p><p>For the sake of comparison with other estimates of HIV/AIDS-related mortality, unadjusted data were extracted for all sites for the period 2008&#x02013;2012 (excluding data from the Farafenni, The Gambia; Purworejo, Indonesia; and FilaBavi, Vietnam sites which did not report for that period). These data were grouped into three age bands (0&#x02013;14, 15&#x02013;49, and 50 +) and aggregated by country, to facilitate comparison with contemporaneous national point estimates for 2010.</p><p>In this context, all of these data are secondary datasets derived from primary data collected separately by each participating site. In all cases, the primary data collection was covered by site-level ethical approvals relating to on-going health and demographic surveillance in those specific locations. No individual identity or household location data were included in the secondary data and no specific ethical approvals were required for these pooled analyses.</p></sec><sec sec-type="results" id="S0002"><title>Results</title><p>In the overall dataset, there were 10,455.4 deaths attributed to HIV/AIDS (including fractions of 11,972 individual deaths), with a further 10,363.4 deaths attributed to acute respiratory infections (including pneumonia), and 12,874.8 attributed to PTB.</p><p>The age&#x02013;sex&#x02013;time standardised CSMFs for HIV/AIDS at each site are shown, together with the population-based HIV/AIDS-specific mortality rate per 1,000 person-years, in <xref ref-type="fig" rid="F0001">Fig. 1</xref>. In West African sites, HIV/AIDS CSMF ranged from 2.10 to 8.00%, with HIV/AIDS-specific adjusted mortality rates ranging from 0.16 to 0.77 per 1,000 person-years. In eastern and southern Africa, except Ethiopia, CSMFs were 9.81&#x02013;18.85%, with rates from 0.65 to 3.09 per 1,000 person-years. In Asia, CSMFs were 0.15&#x02013;3.83%, with rates from 0.01 to 0.21 per 1,000 person-years.</p><fig id="F0001" position="float"><label>Fig. 1</label><caption><p>Map showing participating sites, with age&#x02013;sex&#x02013;time adjusted cause-specific mortality fractions and adjusted mortality rates for HIV/AIDS.</p></caption><graphic xlink:href="GHA-7-25370-g001"/></fig><p> <xref ref-type="fig" rid="F0002">Figure 2</xref> shows HIV/AIDS mortality epidemic curves for the five sites where overall HIV/AIDS mortality was at least 1 per 1,000 person-years. Apart from the Agincourt, South Africa, site, for which a more or less complete epidemic curve can be seen, the other sites recorded mortality during a period of mainly declining HIV/AIDS mortality.</p><fig id="F0002" position="float"><label>Fig. 2</label><caption><p>Epidemic curves for HIV/AIDS mortality rates for the five sites where overall HIV/AIDS mortality exceeded 1 per 1,000 person-years.</p></caption><graphic xlink:href="GHA-7-25370-g002"/></fig><p> <xref ref-type="table" rid="T0001">Table 1</xref> gives HIV/AIDS-specific mortality rates by age group and site. During infancy, the highest HIV/AIDS-specific mortality rate was reported from the Africa Centre, South Africa (7.00 per 1,000 person-years), contrasting with a zero rate from several Asian sites. For the 1&#x02013;4 age group, the Kisumu, Kenya, site recorded the highest rate (5.40 per 1,000 person-years). In the 5&#x02013;14 year age group, Asian sites recorded rates from 0 to 0.07 per 1,000 person-years, compared with African sites from 0.02 to 0.40 per 1,000 person-years. In adulthood, the ranges across Asian sites for 15&#x02013;49 years, 50&#x02013;64 years, and 65+ years were 0&#x02013;0.23, 0.02&#x02013;0.66, and 0&#x02013;0.09, respectively. Similarly for African sites, ranges were 0.08&#x02013;3.65, 0.37&#x02013;4.56, and 0&#x02013;2.26, respectively.</p><table-wrap id="T0001" position="float"><label>Table 1</label><caption><p>HIV/AIDS-specific deaths and mortality rates per 1,000 person-years, by age group and site, from 111,910 deaths in 12,204,043 person-years of observation across 22 sites</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1"/><th align="center" colspan="6" rowspan="1">Age group at death</th></tr><tr><th align="left" rowspan="1" colspan="1"/><th colspan="6" rowspan="1"> <hr/> </th></tr><tr><th align="left" rowspan="1" colspan="1">Country: Site</th><th align="center" rowspan="1" colspan="1">Infant</th><th align="center" rowspan="1" colspan="1">1&#x02013;4 years</th><th align="center" rowspan="1" colspan="1">5&#x02013;14 years</th><th align="center" rowspan="1" colspan="1">15&#x02013;49 years</th><th align="center" rowspan="1" colspan="1">50&#x02013;64 years</th><th align="center" rowspan="1" colspan="1">65+ years</th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">Bangladesh: Matlab</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">Adjusted deaths</td><td align="center" rowspan="1" colspan="1">1.71</td><td align="center" rowspan="1" colspan="1">7.23</td><td align="center" rowspan="1" colspan="1">2.18</td><td align="center" rowspan="1" colspan="1">4.16</td><td align="center" rowspan="1" colspan="1">3.17</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">Rate/1,000 py</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">Bangladesh: Bandarban</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">Adjusted deaths</td><td align="center" rowspan="1" colspan="1">1.19</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">1.00</td><td align="center" rowspan="1" colspan="1">7.02</td><td align="center" rowspan="1" colspan="1">3.89</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">Rate/1,000 py</td><td align="center" rowspan="1" colspan="1">0.96</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.23</td><td align="center" rowspan="1" colspan="1">0.66</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">Bangladesh: Chakaria</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">Adjusted deaths</td><td align="center" rowspan="1" colspan="1">0.36</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">1.71</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">Rate/1,000 py</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">Bangladesh: AMK</td><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Adjusted deaths</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">1.44</td><td align="center" rowspan="1" colspan="1">0.87</td><td align="center" rowspan="1" colspan="1">0.74</td><td align="center" rowspan="1" colspan="1">1.00</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">Rate/1,000 py</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">Burkina Faso: Nouna</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">Adjusted deaths</td><td align="center" rowspan="1" colspan="1">7.92</td><td align="center" rowspan="1" colspan="1">51.12</td><td align="center" rowspan="1" colspan="1">8.89</td><td align="center" rowspan="1" colspan="1">88.29</td><td align="center" rowspan="1" colspan="1">20.34</td><td align="center" rowspan="1" colspan="1">3.11</td></tr><tr><td align="left" rowspan="1" colspan="1">Rate/1,000 py</td><td align="center" rowspan="1" colspan="1">0.26</td><td align="center" rowspan="1" colspan="1">0.49</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.32</td><td align="center" rowspan="1" colspan="1">0.43</td><td align="center" rowspan="1" colspan="1">0.11</td></tr><tr><td align="left" rowspan="1" colspan="1">Burkina Faso: Ouagadougou</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">Adjusted deaths</td><td align="center" rowspan="1" colspan="1">11.55</td><td align="center" rowspan="1" colspan="1">16.15</td><td align="center" rowspan="1" colspan="1">4.63</td><td align="center" rowspan="1" colspan="1">19.85</td><td align="center" rowspan="1" colspan="1">7.57</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">Rate/1,000 py</td><td align="center" rowspan="1" colspan="1">1.66</td><td align="center" rowspan="1" colspan="1">0.58</td><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">0.17</td><td align="center" rowspan="1" colspan="1">0.66</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">C&#x000f4;te d&#x02019;Ivoire: Taabo</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">Adjusted deaths</td><td align="center" rowspan="1" colspan="1">10.20</td><td align="center" rowspan="1" colspan="1">20.09</td><td align="center" rowspan="1" colspan="1">8.24</td><td align="center" rowspan="1" colspan="1">28.95</td><td align="center" rowspan="1" colspan="1">7.23</td><td align="center" rowspan="1" colspan="1">5.06</td></tr><tr><td align="left" rowspan="1" colspan="1">Rate/1,000 py</td><td align="center" rowspan="1" colspan="1">2.57</td><td align="center" rowspan="1" colspan="1">1.55</td><td align="center" rowspan="1" colspan="1">0.27</td><td align="center" rowspan="1" colspan="1">0.60</td><td align="center" rowspan="1" colspan="1">1.04</td><td align="center" rowspan="1" colspan="1">1.59</td></tr><tr><td align="left" rowspan="1" colspan="1">Ethiopia: Kilite Awlaelo</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">Adjusted deaths</td><td align="center" rowspan="1" colspan="1">1.91</td><td align="center" rowspan="1" colspan="1">2.91</td><td align="center" rowspan="1" colspan="1">0.87</td><td align="center" rowspan="1" colspan="1">4.85</td><td align="center" rowspan="1" colspan="1">4.13</td><td align="center" rowspan="1" colspan="1">1.89</td></tr><tr><td align="left" rowspan="1" colspan="1">Rate/1,000 py</td><td align="center" rowspan="1" colspan="1">0.60</td><td align="center" rowspan="1" colspan="1">0.22</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">0.37</td><td align="center" rowspan="1" colspan="1">0.27</td></tr><tr><td align="left" rowspan="1" colspan="1">The Gambia: Farafenni</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">Adjusted deaths</td><td align="center" rowspan="1" colspan="1">13.14</td><td align="center" rowspan="1" colspan="1">44.29</td><td align="center" rowspan="1" colspan="1">11.86</td><td align="center" rowspan="1" colspan="1">40.20</td><td align="center" rowspan="1" colspan="1">20.01</td><td align="center" rowspan="1" colspan="1">3.86</td></tr><tr><td align="left" rowspan="1" colspan="1">Rate/1,000 py</td><td align="center" rowspan="1" colspan="1">1.15</td><td align="center" rowspan="1" colspan="1">1.03</td><td align="center" rowspan="1" colspan="1">0.13</td><td align="center" rowspan="1" colspan="1">0.29</td><td align="center" rowspan="1" colspan="1">0.89</td><td align="center" rowspan="1" colspan="1">0.34</td></tr><tr><td align="left" rowspan="1" colspan="1">Ghana: Navrongo</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">Adjusted deaths</td><td align="center" rowspan="1" colspan="1">31.16</td><td align="center" rowspan="1" colspan="1">92.71</td><td align="center" rowspan="1" colspan="1">24.70</td><td align="center" rowspan="1" colspan="1">195.22</td><td align="center" rowspan="1" colspan="1">52.08</td><td align="center" rowspan="1" colspan="1">10.61</td></tr><tr><td align="left" rowspan="1" colspan="1">Rate/1,000 py</td><td align="center" rowspan="1" colspan="1">1.03</td><td align="center" rowspan="1" colspan="1">0.80</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">0.37</td><td align="center" rowspan="1" colspan="1">0.41</td><td align="center" rowspan="1" colspan="1">0.15</td></tr><tr><td align="left" rowspan="1" colspan="1">Ghana: Dodowa</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">Adjusted deaths</td><td align="center" rowspan="1" colspan="1">5.09</td><td align="center" rowspan="1" colspan="1">10.80</td><td align="center" rowspan="1" colspan="1">10.01</td><td align="center" rowspan="1" colspan="1">41.98</td><td align="center" rowspan="1" colspan="1">13.51</td><td align="center" rowspan="1" colspan="1">2.86</td></tr><tr><td align="left" rowspan="1" colspan="1">Rate/1,000 py</td><td align="center" rowspan="1" colspan="1">0.36</td><td align="center" rowspan="1" colspan="1">0.19</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.16</td><td align="center" rowspan="1" colspan="1">0.37</td><td align="center" rowspan="1" colspan="1">0.11</td></tr><tr><td align="left" rowspan="1" colspan="1">India: Ballabgarh</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">Adjusted deaths</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">1.44</td><td align="center" rowspan="1" colspan="1">0.61</td><td align="center" rowspan="1" colspan="1">2.66</td><td align="center" rowspan="1" colspan="1">2.21</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">Rate/1,000 py</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.07</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">India: Vadu</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">Adjusted deaths</td><td align="center" rowspan="1" colspan="1">8.42</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">2.25</td><td align="center" rowspan="1" colspan="1">1.23</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">Rate/1,000 py</td><td align="center" rowspan="1" colspan="1">1.96</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.02</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">Indonesia: Purworejo</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">Adjusted deaths</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">1.34</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">3.35</td><td align="center" rowspan="1" colspan="1">1.98</td></tr><tr><td align="left" rowspan="1" colspan="1">Rate/1,000 py</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.12</td><td align="center" rowspan="1" colspan="1">0.09</td></tr><tr><td align="left" rowspan="1" colspan="1">Kenya: Kilifi</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">Adjusted deaths</td><td align="center" rowspan="1" colspan="1">69.17</td><td align="center" rowspan="1" colspan="1">70.79</td><td align="center" rowspan="1" colspan="1">60.88</td><td align="center" rowspan="1" colspan="1">276.26</td><td align="center" rowspan="1" colspan="1">99.83</td><td align="center" rowspan="1" colspan="1">50.99</td></tr><tr><td align="left" rowspan="1" colspan="1">Rate/1,000 py</td><td align="center" rowspan="1" colspan="1">1.80</td><td align="center" rowspan="1" colspan="1">0.48</td><td align="center" rowspan="1" colspan="1">0.20</td><td align="center" rowspan="1" colspan="1">0.65</td><td align="center" rowspan="1" colspan="1">1.52</td><td align="center" rowspan="1" colspan="1">1.54</td></tr><tr><td align="left" rowspan="1" colspan="1">Kenya: Kisumu</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">Adjusted deaths</td><td align="center" rowspan="1" colspan="1">297.87</td><td align="center" rowspan="1" colspan="1">780.55</td><td align="center" rowspan="1" colspan="1">128.47</td><td align="center" rowspan="1" colspan="1">1708.84</td><td align="center" rowspan="1" colspan="1">406.33</td><td align="center" rowspan="1" colspan="1">132.85</td></tr><tr><td align="left" rowspan="1" colspan="1">Rate/1,000 py</td><td align="center" rowspan="1" colspan="1">7.47</td><td align="center" rowspan="1" colspan="1">5.40</td><td align="center" rowspan="1" colspan="1">0.40</td><td align="center" rowspan="1" colspan="1">3.65</td><td align="center" rowspan="1" colspan="1">4.56</td><td align="center" rowspan="1" colspan="1">1.98</td></tr><tr><td align="left" rowspan="1" colspan="1">Kenya: Nairobi</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Adjusted deaths</td><td align="center" rowspan="1" colspan="1">66.99</td><td align="center" rowspan="1" colspan="1">81.34</td><td align="center" rowspan="1" colspan="1">18.42</td><td align="center" rowspan="1" colspan="1">356.93</td><td align="center" rowspan="1" colspan="1">44.46</td><td align="center" rowspan="1" colspan="1">4.34</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Rate/1,000 py</td><td align="center" rowspan="1" colspan="1">4.67</td><td align="center" rowspan="1" colspan="1">1.30</td><td align="center" rowspan="1" colspan="1">0.17</td><td align="center" rowspan="1" colspan="1">0.93</td><td align="center" rowspan="1" colspan="1">1.79</td><td align="center" rowspan="1" colspan="1">0.77</td></tr><tr><td align="left" rowspan="1" colspan="1">Malawi: Karonga</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Adjusted deaths</td><td align="center" rowspan="1" colspan="1">n/a</td><td align="center" rowspan="1" colspan="1">n/a</td><td align="center" rowspan="1" colspan="1">n/a</td><td align="center" rowspan="1" colspan="1">226.10</td><td align="center" rowspan="1" colspan="1">50.15</td><td align="center" rowspan="1" colspan="1">16.82</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Rate/1,000 py</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1.92</td><td align="center" rowspan="1" colspan="1">3.39</td><td align="center" rowspan="1" colspan="1">1.48</td></tr><tr><td align="left" rowspan="1" colspan="1">Senegal: Niakhar</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Adjusted deaths</td><td align="center" rowspan="1" colspan="1">2.68</td><td align="center" rowspan="1" colspan="1">16.06</td><td align="center" rowspan="1" colspan="1">11.39</td><td align="center" rowspan="1" colspan="1">62.15</td><td align="center" rowspan="1" colspan="1">21.95</td><td align="center" rowspan="1" colspan="1">7.05</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Rate/1,000 py</td><td align="center" rowspan="1" colspan="1">0.30</td><td align="center" rowspan="1" colspan="1">0.50</td><td align="center" rowspan="1" colspan="1">0.19</td><td align="center" rowspan="1" colspan="1">0.64</td><td align="center" rowspan="1" colspan="1">1.37</td><td align="center" rowspan="1" colspan="1">0.66</td></tr><tr><td align="left" rowspan="1" colspan="1">South Africa: Agincourt</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Adjusted deaths</td><td align="center" rowspan="1" colspan="1">140.09</td><td align="center" rowspan="1" colspan="1">302.84</td><td align="center" rowspan="1" colspan="1">67.11</td><td align="center" rowspan="1" colspan="1">1434.89</td><td align="center" rowspan="1" colspan="1">309.61</td><td align="center" rowspan="1" colspan="1">142.86</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Rate/1,000 py</td><td align="center" rowspan="1" colspan="1">3.81</td><td align="center" rowspan="1" colspan="1">2.03</td><td align="center" rowspan="1" colspan="1">0.18</td><td align="center" rowspan="1" colspan="1">1.98</td><td align="center" rowspan="1" colspan="1">3.35</td><td align="center" rowspan="1" colspan="1">2.26</td></tr><tr><td align="left" rowspan="1" colspan="1">South Africa: Africa Centre</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Adjusted deaths</td><td align="center" rowspan="1" colspan="1">157.25</td><td align="center" rowspan="1" colspan="1">232.14</td><td align="center" rowspan="1" colspan="1">55.38</td><td align="center" rowspan="1" colspan="1">1227.58</td><td align="center" rowspan="1" colspan="1">210.72</td><td align="center" rowspan="1" colspan="1">74.44</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Rate/1,000 py</td><td align="center" rowspan="1" colspan="1">7.00</td><td align="center" rowspan="1" colspan="1">2.54</td><td align="center" rowspan="1" colspan="1">0.24</td><td align="center" rowspan="1" colspan="1">3.28</td><td align="center" rowspan="1" colspan="1">3.84</td><td align="center" rowspan="1" colspan="1">1.90</td></tr><tr><td align="left" rowspan="1" colspan="1">Vietnam: FilaBavi</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Adjusted deaths</td><td align="center" rowspan="1" colspan="1">0.32</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">3.02</td><td align="center" rowspan="1" colspan="1">2.80</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Rate/1,000 py</td><td align="center" rowspan="1" colspan="1">0.15</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.17</td><td align="center" rowspan="1" colspan="1">0.00</td></tr></tbody></table></table-wrap><p> <xref ref-type="fig" rid="F0003">Figure 3</xref> shows the relationships between age&#x02013;sex&#x02013;time standardised HIV/AIDS mortality rates and PTB mortality rates for all 22 sites. Seven of the eight sites in Asia had an HIV/AIDS rate below 0.1 per 1,000 person-years, but PTB rates ranged from 0.11 to 0.75 per 1,000 person-years. Conversely, six of the seven sites in eastern and southern Africa had HIV/AIDS rates above 0.5 per 1,000 person-years, with PTB rates ranging from 0.52 to 4.96 per 1,000 person-years. The highest age&#x02013;sex&#x02013;time standardised HIV/AIDS mortality rate ratio was between Kisumu, Kenya, and AMK, Bangladesh, at 343:1.</p><fig id="F0003" position="float"><label>Fig. 3</label><caption><p>Relationship between HIV/AIDS and pulmonary TB age&#x02013;sex&#x02013;time standardised mortality rates for 22 INDEPTH Network sites in Africa and Asia.</p></caption><graphic xlink:href="GHA-7-25370-g003"/></fig><p> <xref ref-type="fig" rid="F0004">Figure 4</xref> shows HIV/AIDS mortality rates for 15 sites which had an overall HIV/AIDS-specific mortality rate over 0.1 per 1,000 person-years, by age group, also showing corresponding data for PTB and pneumonia. Logarithmic scales have been used to visualise both high and low levels of mortality while using the same scale for each site.</p><fig id="F0004" position="float"><label>Fig. 4</label><caption><p>Mortality rates for HIV/AIDS, pulmonary TB, and pneumonia, by site and age group at 15 INDEPTH Network sites for which the overall rate of HIV/AIDS mortality exceeded 0.1/1,000 person-years.</p></caption><graphic xlink:href="GHA-7-25370-g004"/></fig><p> <xref ref-type="table" rid="T0002">Table 2</xref> shows the INDEPTH results in comparison with national estimates from the UNAIDS Spectrum model (<xref rid="CIT0012" ref-type="bibr">12</xref>) and Global Burden of Disease 2010 (<xref rid="CIT0013" ref-type="bibr">13</xref>). Longitudinal INDEPTH data were aggregated over 2008&#x02013;2012 (for the 19 sites reporting for that period) for the purposes of comparison with the Spectrum and Global Burden of Disease 2010 (GBD 2010) estimates, together with corresponding estimates for PTB.</p><table-wrap id="T0002" position="float"><label>Table 2</label><caption><p>Within-country estimates of cause-specific (per 1,000 population) mortality rates for HIV/AIDS and TB for 2010 according to UNAIDS Spectrum and the Global Burden of Disease 2010, compared to the equivalent rates across 19 INDEPTH sites, aggregated within 10 countries, for 2008&#x02013;2012 (per 1,000 person-years)</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1"/><th align="center" colspan="12" rowspan="1">HIV/AIDS mortality rates per 1,000</th><th align="center" rowspan="1" colspan="1"/><th align="left" rowspan="1" colspan="1"/><th align="left" rowspan="1" colspan="1"/></tr><tr><th align="left" rowspan="1" colspan="1"/><th colspan="12" rowspan="1"> <hr/> </th></tr><tr><th align="left" rowspan="1" colspan="1"/><th align="center" colspan="3" rowspan="1">0&#x02013;14 years</th><th align="center" colspan="3" rowspan="1">15&#x02013;49 years</th><th align="center" colspan="3" rowspan="1">50+ years</th><th align="center" colspan="3" rowspan="1">All ages</th><th align="center" colspan="3" rowspan="1">TB mortality rates per 1,000</th></tr><tr><th align="left" rowspan="1" colspan="1"/><th colspan="15" rowspan="1"> <hr/> </th></tr><tr><th align="left" rowspan="1" colspan="1">Country</th><th align="center" rowspan="1" colspan="1">Spectrum</th><th align="center" rowspan="1" colspan="1">GBD 2010</th><th align="center" rowspan="1" colspan="1">INDEPTH</th><th align="center" rowspan="1" colspan="1">Spectrum</th><th align="center" rowspan="1" colspan="1">GBD 2010</th><th align="center" rowspan="1" colspan="1">INDEPTH</th><th align="center" rowspan="1" colspan="1">Spectrum</th><th align="center" rowspan="1" colspan="1">GBD 2010</th><th align="center" rowspan="1" colspan="1">INDEPTH</th><th align="center" rowspan="1" colspan="1">Spectrum</th><th align="center" rowspan="1" colspan="1">GBD 2010</th><th align="center" rowspan="1" colspan="1">INDEPTH</th><th align="center" rowspan="1" colspan="1">Spectrum</th><th align="center" rowspan="1" colspan="1">GBD 2010</th><th align="center" rowspan="1" colspan="1">INDEPTH</th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">Bangladesh</td><td align="center" rowspan="1" colspan="1">0.0004</td><td align="center" rowspan="1" colspan="1">0.0003</td><td align="center" rowspan="1" colspan="1">0.016</td><td align="center" rowspan="1" colspan="1">0.004</td><td align="center" rowspan="1" colspan="1">0.002</td><td align="center" rowspan="1" colspan="1">0.015</td><td align="center" rowspan="1" colspan="1">0.002</td><td align="center" rowspan="1" colspan="1">0.001</td><td align="center" rowspan="1" colspan="1">0.034</td><td align="center" rowspan="1" colspan="1">0.002</td><td align="center" rowspan="1" colspan="1">0.002</td><td align="center" rowspan="1" colspan="1">0.019</td><td align="center" rowspan="1" colspan="1">0.533</td><td align="center" rowspan="1" colspan="1">0.180</td><td align="center" rowspan="1" colspan="1">0.436</td></tr><tr><td align="left" rowspan="1" colspan="1">Burkina Faso</td><td align="center" rowspan="1" colspan="1">0.207</td><td align="center" rowspan="1" colspan="1">0.273</td><td align="center" rowspan="1" colspan="1">0.234</td><td align="center" rowspan="1" colspan="1">0.521</td><td align="center" rowspan="1" colspan="1">0.588</td><td align="center" rowspan="1" colspan="1">0.132</td><td align="center" rowspan="1" colspan="1">0.455</td><td align="center" rowspan="1" colspan="1">0.369</td><td align="center" rowspan="1" colspan="1">0.276</td><td align="center" rowspan="1" colspan="1">0.371</td><td align="center" rowspan="1" colspan="1">0.429</td><td align="center" rowspan="1" colspan="1">0.188</td><td align="center" rowspan="1" colspan="1">0.124</td><td align="center" rowspan="1" colspan="1">0.216</td><td align="center" rowspan="1" colspan="1">0.133</td></tr><tr><td align="left" rowspan="1" colspan="1">C&#x000f4;te d&#x02019;Ivoire</td><td align="center" rowspan="1" colspan="1">0.477</td><td align="center" rowspan="1" colspan="1">0.509</td><td align="center" rowspan="1" colspan="1">0.804</td><td align="center" rowspan="1" colspan="1">2.077</td><td align="center" rowspan="1" colspan="1">2.021</td><td align="center" rowspan="1" colspan="1">0.597</td><td align="center" rowspan="1" colspan="1">1.802</td><td align="center" rowspan="1" colspan="1">1.022</td><td align="center" rowspan="1" colspan="1">1.212</td><td align="center" rowspan="1" colspan="1">1.379</td><td align="center" rowspan="1" colspan="1">1.286</td><td align="center" rowspan="1" colspan="1">0.749</td><td align="center" rowspan="1" colspan="1">n/a</td><td align="center" rowspan="1" colspan="1">0.309</td><td align="center" rowspan="1" colspan="1">0.430</td></tr><tr><td align="left" rowspan="1" colspan="1">Ethiopia</td><td align="center" rowspan="1" colspan="1">0.417</td><td align="center" rowspan="1" colspan="1">0.185</td><td align="center" rowspan="1" colspan="1">0.101</td><td align="center" rowspan="1" colspan="1">1.114</td><td align="center" rowspan="1" colspan="1">0.472</td><td align="center" rowspan="1" colspan="1">0.082</td><td align="center" rowspan="1" colspan="1">0.827</td><td align="center" rowspan="1" colspan="1">0.160</td><td align="center" rowspan="1" colspan="1">0.329</td><td align="center" rowspan="1" colspan="1">0.776</td><td align="center" rowspan="1" colspan="1">0.320</td><td align="center" rowspan="1" colspan="1">0.124</td><td align="center" rowspan="1" colspan="1">0.519</td><td align="center" rowspan="1" colspan="1">0.453</td><td align="center" rowspan="1" colspan="1">0.408</td></tr><tr><td align="left" rowspan="1" colspan="1">Ghana</td><td align="center" rowspan="1" colspan="1">0.227</td><td align="center" rowspan="1" colspan="1">0.293</td><td align="center" rowspan="1" colspan="1">0.212</td><td align="center" rowspan="1" colspan="1">0.913</td><td align="center" rowspan="1" colspan="1">1.116</td><td align="center" rowspan="1" colspan="1">0.217</td><td align="center" rowspan="1" colspan="1">0.737</td><td align="center" rowspan="1" colspan="1">0.569</td><td align="center" rowspan="1" colspan="1">0.329</td><td align="center" rowspan="1" colspan="1">0.626</td><td align="center" rowspan="1" colspan="1">0.735</td><td align="center" rowspan="1" colspan="1">0.232</td><td align="center" rowspan="1" colspan="1">0.141</td><td align="center" rowspan="1" colspan="1">0.157</td><td align="center" rowspan="1" colspan="1">0.479</td></tr><tr><td align="left" rowspan="1" colspan="1">India</td><td align="center" rowspan="1" colspan="1">0.034</td><td align="center" rowspan="1" colspan="1">0.030</td><td align="center" rowspan="1" colspan="1">0.061</td><td align="center" rowspan="1" colspan="1">0.179</td><td align="center" rowspan="1" colspan="1">0.267</td><td align="center" rowspan="1" colspan="1">0.015</td><td align="center" rowspan="1" colspan="1">0.083</td><td align="center" rowspan="1" colspan="1">0.119</td><td align="center" rowspan="1" colspan="1">0.051</td><td align="center" rowspan="1" colspan="1">0.119</td><td align="center" rowspan="1" colspan="1">0.171</td><td align="center" rowspan="1" colspan="1">0.033</td><td align="center" rowspan="1" colspan="1">0.299</td><td align="center" rowspan="1" colspan="1">0.346</td><td align="center" rowspan="1" colspan="1">0.538</td></tr><tr><td align="left" rowspan="1" colspan="1">Kenya</td><td align="center" rowspan="1" colspan="1">0.744</td><td align="center" rowspan="1" colspan="1">0.617</td><td align="center" rowspan="1" colspan="1">0.846</td><td align="center" rowspan="1" colspan="1">2.446</td><td align="center" rowspan="1" colspan="1">2.201</td><td align="center" rowspan="1" colspan="1">1.112</td><td align="center" rowspan="1" colspan="1">1.879</td><td align="center" rowspan="1" colspan="1">1.371</td><td align="center" rowspan="1" colspan="1">2.001</td><td align="center" rowspan="1" colspan="1">1.671</td><td align="center" rowspan="1" colspan="1">1.452</td><td align="center" rowspan="1" colspan="1">1.082</td><td align="center" rowspan="1" colspan="1">0.304</td><td align="center" rowspan="1" colspan="1">0.257</td><td align="center" rowspan="1" colspan="1">0.853</td></tr><tr><td align="left" rowspan="1" colspan="1">Malawi</td><td align="center" rowspan="1" colspan="1">1.584</td><td align="center" rowspan="1" colspan="1">1.694</td><td align="center" rowspan="1" colspan="1">n/a</td><td align="center" rowspan="1" colspan="1">4.910</td><td align="center" rowspan="1" colspan="1">5.656</td><td align="center" rowspan="1" colspan="1">0.854</td><td align="center" rowspan="1" colspan="1">3.763</td><td align="center" rowspan="1" colspan="1">3.343</td><td align="center" rowspan="1" colspan="1">1.140</td><td align="center" rowspan="1" colspan="1">3.279</td><td align="center" rowspan="1" colspan="1">3.614</td><td align="center" rowspan="1" colspan="1">n/a</td><td align="center" rowspan="1" colspan="1">0.469</td><td align="center" rowspan="1" colspan="1">0.431</td><td align="center" rowspan="1" colspan="1">0.268</td></tr><tr><td align="left" rowspan="1" colspan="1">Senegal</td><td align="center" rowspan="1" colspan="1">0.082</td><td align="center" rowspan="1" colspan="1">0.076</td><td align="center" rowspan="1" colspan="1">0.178</td><td align="center" rowspan="1" colspan="1">0.184</td><td align="center" rowspan="1" colspan="1">0.295</td><td align="center" rowspan="1" colspan="1">0.483</td><td align="center" rowspan="1" colspan="1">0.075</td><td align="center" rowspan="1" colspan="1">0.279</td><td align="center" rowspan="1" colspan="1">1.564</td><td align="center" rowspan="1" colspan="1">0.129</td><td align="center" rowspan="1" colspan="1">0.198</td><td align="center" rowspan="1" colspan="1">0.474</td><td align="center" rowspan="1" colspan="1">0.228</td><td align="center" rowspan="1" colspan="1">0.210</td><td align="center" rowspan="1" colspan="1">0.861</td></tr><tr><td align="left" rowspan="1" colspan="1">South Africa</td><td align="center" rowspan="1" colspan="1">1.449</td><td align="center" rowspan="1" colspan="1">1.166</td><td align="center" rowspan="1" colspan="1">0.695</td><td align="center" rowspan="1" colspan="1">10.628</td><td align="center" rowspan="1" colspan="1">8.048</td><td align="center" rowspan="1" colspan="1">1.935</td><td align="center" rowspan="1" colspan="1">9.031</td><td align="center" rowspan="1" colspan="1">2.422</td><td align="center" rowspan="1" colspan="1">2.917</td><td align="center" rowspan="1" colspan="1">7.641</td><td align="center" rowspan="1" colspan="1">5.099</td><td align="center" rowspan="1" colspan="1">1.587</td><td align="center" rowspan="1" colspan="1">1.970</td><td align="center" rowspan="1" colspan="1">0.322</td><td align="center" rowspan="1" colspan="1">2.483</td></tr></tbody></table></table-wrap></sec><sec sec-type="discussion" id="S0003"><title>Discussion</title><p>Against the background of extensive modelling approaches that have been applied to HIV/AIDS mortality, this dataset presents results from individually documented deaths at a range of sites across Africa and Asia. The expected huge differences in HIV/AIDS mortality rates between Africa and Asia were evident from these results, and, to a lesser extent, the substantial differences that occurred within the African continent. The good news is that HIV/AIDS deaths declined in recent years in all the sites with high mortality rates (<xref ref-type="fig" rid="F0002">Fig. 2</xref>), as the effects of prevention and treatment programmes took effect. The interpretation of findings at individual sites depends on local characteristics (<xref rid="CIT0014" ref-type="bibr">14</xref>&#x02013;<xref rid="CIT0035" ref-type="bibr">35</xref>) . Two sites, Ouagadougou in Burkina Faso and Nairobi in Kenya, followed urban populations. Bandarban in Bangladesh is located in a militarised frontier zone close to the Myanmar border, which may be associated with higher rates of HIV/AIDS mortality compared with other sites in Bangladesh.</p><p>The validity of VA cause of death assignment for HIV/AIDS is not straightforward. In these results, the similar and marked changes over time in the high mortality sites (<xref ref-type="fig" rid="F0002">Fig. 2</xref>) added veracity to the InterVA-4 outputs, since the model had no information about the progress of the epidemic over time. Similarly, the extremely low levels of HIV/AIDS-related death assigned as a cause in countries such as Bangladesh and India confirmed the specificity of the methods used. A previous assessment of InterVA-4 validity versus HIV sero-status showed high specificity, but sensitivity was unmeasurable since not all HIV-positive people go on to die from HIV/AIDS (<xref rid="CIT0011" ref-type="bibr">11</xref>). However, the same study also showed high mortality rate ratios for PTB and pneumonia between HIV positive and negative cases. ICD-10 classification (<xref rid="CIT0036" ref-type="bibr">36</xref>) suggests that almost all HIV-related deaths should be classified under the B20-B24 rubrics, but this is easier said than done in practice, either when using VA or when certifying a death, if there is no evidence of HIV status. In view of the apparently complex relationships between HIV/AIDS deaths and PTB deaths in different settings, as evidenced in <xref ref-type="fig" rid="F0003">Fig. 3</xref>, it is not simply a matter of adding together HIV/AIDS and PTB deaths across all settings. However, the total of what InterVA-4 assigns as HIV/AIDS and PTB deaths may provide a better approximation of the overall burden of HIV/AIDS-related mortality for at least the 15&#x02013;49 year age group in high HIV settings. The question of HIV/AIDS-related mortality associated with pregnancy has also been a matter of debate (<xref rid="CIT0037" ref-type="bibr">37</xref>). Another paper in this series analyses pregnancy-related mortality in detail, including the attribution of HIV/AIDS-related deaths between indirect maternal and incidental categories (<xref rid="CIT0038" ref-type="bibr">38</xref>).</p><p>The WHO 2012 VA standard (<xref rid="CIT0039" ref-type="bibr">39</xref>) includes an indicator relating to previous diagnosis of HIV, although the validation study suggested that this was seriously under-reported in VA interviews (<xref rid="CIT0011" ref-type="bibr">11</xref>). The WHO 2012 standard, and therefore InterVA-4, does not yet include any details of anti-retroviral therapy (ART), although that will become a more pressing issue as experience of mortality patterns among HIV positive individuals with long exposure to ART develops. It is as yet a relatively open question as to what the major causes of death among HIV-positive people might be after possible decades of ART.</p><p>There are other major pieces of work describing HIV/AIDS mortality patterns across Africa and Asia, but these largely relied on modelling estimates from whatever specific sources of data were available, and therefore carried large degrees of uncertainty given the sparse nature of the data from many settings. The two major sources of contemporaneous estimates for HIV/AIDS mortality come from the UNAIDS Spectrum model (<xref rid="CIT0012" ref-type="bibr">12</xref>) and the GBD 2010 model (<xref rid="CIT0013" ref-type="bibr">13</xref>). Although our purpose here is not to compare these two models with each other, it is worth noting that there are some major differences. For example, among the countries represented here, the estimates for Ethiopia vary three-fold.</p><p> <xref ref-type="table" rid="T0002">Table 2</xref> shows estimates of HIV/AIDS-related and PTB mortality rates for 12 countries according to Spectrum, GBD 2010, and InterVA-4, which in many cases were very similar, though with differences in places. It must be remembered that these comparisons were compromised by taking INDEPTH sites that are not designed to be nationally representative and putting their findings alongside modelled estimates that are intended to reflect national situations. In South Africa, it appeared that InterVA-4 assigned a substantial amount of HIV/AIDS mortality as PTB, which is perhaps unsurprising in that high-prevalence setting. InterVA-4 arrived at a substantially higher HIV/AIDS mortality estimate than Spectrum for Senegal, and vice-versa for India. There were also many similarities in PTB mortality rates, though differences were evident in Ghana, Kenya, and Senegal.</p><p>Similarly there were relatively few appreciable differences between GBD 2010 and InterVA-4 estimates. The differences may reflect local disparities in rates between sites and national populations, given that the relationships between symptoms and causes would not be expected to vary substantially between countries. It also has to be remembered that, although all these VAs have been processed in a standardised way using the WHO 2012 protocol, they were collected in the field in slightly different ways before 2012, and some observed differences may also reflect that. Overall, however, there was appreciable congruence in mortality rates between these various sources.</p></sec><sec sec-type="conclusions" id="S0004"><title>Conclusions</title><p>Measuring HIV/AIDS mortality continues to be a highly challenging area, particularly in Africa, where rates are high and data are often unavailable. This is the largest single systematic study that has applied common methodologies to HIV/AIDS mortality at the individual level across Africa and Asia, and it largely confirms the corresponding findings coming from modelled estimates. This mutually adds plausibility to both existing estimates and to these population-based findings. The challenges involved in measuring HIV/AIDS mortality will grow as ART coverage and individual duration on treatment increase; in many ways, these results represent an important baseline for future studies of the treated pandemic.</p></sec>

Causes of death in two rural demographic surveillance sites in Bangladesh, 2004&#x02013;2010: automated coding of verbal autopsies using InterVA-4

<sec id="st1"><title>Objective</title><p>Population-based information on causes of death (CoD) by age, sex, and area is critical for countries with limited resources to identify and address key public health issues. This study analysed the demographic surveillance and verbal autopsy (VA) data to estimate age- and sex-specific mortality rates and cause-specific mortality fractions in two well-defined rural populations within the demographic surveillance system in Abhoynagar and Mirsarai subdistricts, located in different climatic zones.</p></sec><sec id="st2"><title>Design</title><p>During 2004&#x02013;2010, the sample demographic surveillance system registered 1,384 deaths in Abhoynagar and 1,847 deaths in Mirsarai. Trained interviewers interviewed the main caretaker of the deceased with standard VA questionnaires to record signs and symptoms of diseases or conditions that led to death and health care experiences before death. The computer-automated InterVA-4 method was used to analyse VAs to determine probable CoD.</p></sec><sec id="st3"><title>Results</title><p>Age- and sex-specific death rates revealed a higher neonatal mortality rate in Abhoynagar than Mirsarai, and death rates and sex ratios of male to female death rates were higher in the ages after infancy. Communicable diseases (CDs) accounted for 16.7% of all deaths in Abhoynagar and 21.2% in Mirsarai &#x02013; the difference was due mostly to more deaths from acute respiratory infections, pneumonia, and tuberculosis in Mirsarai. Non-communicable diseases (NCDs) accounted for 56.2 and 55.3% of deaths in each subdistrict, respectively, with leading causes being stroke (16.5&#x02013;19.3%), neoplasms (13.2% each), cardiac diseases (8.9&#x02013;11.6%), chronic obstructive pulmonary diseases (5.1&#x02013;6.3%), diseases of the digestive system (3.1&#x02013;4.1%), and diabetes (2.8&#x02013;3.5%), together accounting for 49.2&#x02013;51.2% points of the NCD deaths in the two subdistricts. Injury and other external causes accounted for another 7.5&#x02013;7.7% deaths, with self-harm being higher among females in Abhoynagar.</p></sec><sec id="st4"><title>Conclusions</title><p>The computer-automated coding of VA to determine CoD reconfirmed that NCDs were the leading CoD with some differences between the sites. Incorporating VA into the national sample vital registration system can help policy makers to identify the leading CoDs for public health planning.</p></sec>

<contrib contrib-type="author"><name><surname>Alam</surname><given-names>Nurul</given-names></name><xref ref-type="aff" rid="AF0001">1</xref><xref ref-type="corresp" rid="cor1">*</xref></contrib><contrib contrib-type="author"><name><surname>Chowdhury</surname><given-names>Hafizur R.</given-names></name><xref ref-type="aff" rid="AF0002">2</xref></contrib><contrib contrib-type="author"><name><surname>Das</surname><given-names>Subhash C.</given-names></name><xref ref-type="aff" rid="AF0003">3</xref></contrib><contrib contrib-type="author"><name><surname>Ashraf</surname><given-names>Ali</given-names></name><xref ref-type="aff" rid="AF0004">4</xref></contrib><contrib contrib-type="author"><name><surname>Streatfield</surname><given-names>P. Kim</given-names></name><xref ref-type="aff" rid="AF0001">1</xref></contrib>

Global Health Action

<p>Reliable and up-to-date population-based statistics on morbidity and mortality by cause are important for health sector planning to combat ill health and ill-health-induced poverty. Such statistics, to our knowledge, are lacking in Bangladesh, except for children aged under 5 and maternal deaths for which data are available through periodic special health surveys (<xref rid="CIT0001" ref-type="bibr">1</xref>&#x02013;<xref rid="CIT0004" ref-type="bibr">4</xref>). More than 70% of the population in Bangladesh lives in rural areas, where most deaths (&#x0003e;88%) occur at home and death certificates from which one can derive causes of death (CoD) are hardly available (<xref rid="CIT0005" ref-type="bibr">5</xref>&#x02013;<xref rid="CIT0007" ref-type="bibr">7</xref>). The Health Management Information System of the Ministry of Health and Family Welfare publishes service statistics on morbidity and mortality based on mostly public hospital registry. Deaths with physician-certified cause in public hospitals are a tiny non-representative fraction (&#x0003c;4%) of more than 1 million deaths each year (<xref rid="CIT0008" ref-type="bibr">8</xref>). Since 1982, the nationally representative sample vital registration system (SVRS) administered by the Bangladesh Bureau of Statistics records CoD reported by family members of the deceased (<xref rid="CIT0007" ref-type="bibr">7</xref>). This non-scientific lay-reporting of CoD, with a high proportion being classified as either unspecified or ill-defined, seriously limits the utility of such information for health sector planning.</p><p>In settings such as Bangladesh, where civil registration of deaths is incomplete and medical certification of death is not common practice, verbal autopsy (VA) is a scientific, practical, and low-cost approach to generating population-based information on CoD for health sector planning, implementing, and monitoring (<xref rid="CIT0009" ref-type="bibr">9</xref>, <xref rid="CIT0010" ref-type="bibr">10</xref>). Longitudinal demographic surveillance sites, national health surveys, and sample vital registration schemes are increasingly using VA to generate vital statistics with CoD (<xref rid="CIT0001" ref-type="bibr">1</xref>&#x02013;<xref rid="CIT0004" ref-type="bibr">4</xref>, <xref rid="CIT0006" ref-type="bibr">6</xref>, <xref rid="CIT0011" ref-type="bibr">11</xref>, <xref rid="CIT0012" ref-type="bibr">12</xref>).</p><p>The International Centre for Diarrhoeal Disease Research, Bangladesh (icddr,b) introduced the VA questionnaires in 2004 in its longitudinal sample health and demographic surveillance system (HDSS) maintained in two rural sites in Bangladesh. The sites are located in two different climatic zones. The VA data from these sites provide a rare opportunity to compare the pattern of CoD in two rural populations with varying climates and socioeconomic conditions, but comparable public primary health care facilities and service delivery system.</p><p>The objectives of the study are to compare death rates and cause-specific mortality fractions (CSMFs) in males and females between the two rural HDSS sites, as well as the rank-order of CSMFs between age groups. The spatial difference in CSMFs may reveal the importance of such statistics for planning of local public health services.</p><sec sec-type="methods" id="S0002"><title>Methods</title><sec id="S0002-S20001"><title>Location of HDSS sites</title><p>A member of the INDEPTH Network in Bangladesh, AMK HDSS, has two active rural field sites (<xref ref-type="fig" rid="F0001">Fig. 1</xref>). Abhoynagar is 30 km north of Khulna city on the Khulna&#x02013;Jessore axis where industrialisation and urbanisation are taking place. Mirsarai is 60 km north of Chittagong seaport city on the Dhaka&#x02013;Chittagong axis, and is on the coast of the Bay of Bengal. Bangladesh is divided into seven distinct climatic zones (<xref rid="CIT0013" ref-type="bibr">13</xref>). Abhoynagar is located in the southwestern zone, which is characterised by heavy dewfall; and Mirsarai is in the southeastern zone characterised by more frequent severe hail storms, northwesterly winds and tornadoes, and heavy winter dewfall. Temperatures differ considerably, with the southwestern zone showing average temperatures in January and April of 19&#x000b0;C and 29&#x000b0;C, respectively, and the southeastern zone showing temperatures of 20&#x000b0;C and 28&#x000b0;C in the same months, respectively. Annual rainfall is substantially less in the southwestern zone, 150 cm compared to 250 cm in the southeastern zone. Weather may affect health and CoD differently in different zones.</p><fig id="F0001" position="float"><label>Fig. 1</label><caption><p>Map of the two health and demographic surveillance sites in Bangladesh.</p></caption><graphic xlink:href="GHA-7-25511-g001"/></fig></sec><sec id="S0002-S20002"><title>Socioeconomic condition</title><p>The primary and secondary school enrolment rates of boys and girls in 2008 were comparable between the sites (<xref rid="CIT0014" ref-type="bibr">14</xref>). Agriculture is the predominant occupation in each site, with engagement in agriculture being substantially higher in Abhoynagar. Female labour force participation was higher in Mirsarai than Abhoynagar (11% vs. 2%). Households in Mirsarai more frequently reported additional income through remittances than in Abhoynagar (77% vs. 52%). The self-reported average household income and per capita income was higher in Mirsarai compared to Abhoynagar.</p></sec><sec id="S0002-S20003"><title>Local health delivery system</title><p>The Government of Bangladesh provides primary health care services to all Bangladeshis for a nominal fee through a three-tiered health service delivery system: the Community Clinics, each for about 6,000 people; the Health and Family Welfare Centres, each for 25,000 people; and the Upazila (subdistrict) Health Complexes with an out-patient and an emergency department, 50 in-patient beds, and an operating room, each covering around 250,000 people. Both sites have similar public primary health care infrastructures.</p></sec><sec id="S0002-S20004"><title>Sampling in HDSS</title><p>The sample HDSS set up by icddr,b in Abhoynagar, covered 122 villages in 7 out of 17 Unions (each has about 25,000 people) since 1982; and in Mirsarai, covered 119 villages in 7 out of 16 Unions since 1994. A stratified, two-staged, random, systematic sampling design is used in each stratum &#x02013; a subdistrict. The first stage was a random sample of unions and the second stage was a systematic random sample of households (<xref rid="CIT0014" ref-type="bibr">14</xref>). A household listing operation was carried out in selected unions to prepare the sampling frame for selection of households. Each household had an equal probability of selection in each stratum. The sample fraction was every sixth household in the Abhoynagar field site and every fourth household in the Mirsarai field site.</p></sec><sec id="S0002-S20005"><title>Introduction of VA into HDSS</title><p>The standard VA questionnaires developed by the WHO and modified by INDEPTH for neonatal, child, and adult deaths were introduced into AMK HDSS in 2004 to generate population-based vital statistics and CoD of neonates, children, and adults. VA refers to method of interviewing close family members and caretakers of the deceased about the events surrounding the fatal illness episodes or conditions. The interview attempts to unearth what happened during the hours, days, or months preceding the death event. The standard VA contains both open narratives related to death and leading questions to elicit symptoms and signs of illness or conditions of the deceased. For neonatal deaths, a description of the mother's delivery is recorded. These signs and symptoms or conditions alone or in combination are highly indicative of specific disease.</p><p>The VA tools in English were customised to suit local conditions and then translated into Bangla. Customisation includes reducing the number of questions on HIV/AIDs and malaria as the prevalence of these two diseases is very low. The VA tools were revised in 2009 to be comparable with the 2008 WHO-revised VA tools. Collection of vital events including VA was approved by the Ethical Review Committee of icddr,b.</p></sec><sec id="S0002-S20006"><title>Training of field research assistants</title><p>Female field research assistants (FRAs &#x02013; eight in Abhoynagar, and nine in Mirsarai) having at least higher secondary education were provided intensive training on collection of vital events and migration information. A public-health physician and a medical sociologist provided 4 days&#x02019; training to FRAs and field research supervisors (FRSs) of non-medical background on modular VA tools, followed by 2 days of field practice. FRAs with a standard consent form informed the closest caretakers and relatives about the purpose of the study and guaranteed confidentiality of the information they would provide. Willingness to take part was expressed by signature or thumb impression.</p></sec><sec id="S0002-S20007"><title>Identification of deaths and VA collection</title><p>Each FRA visited about 1,200 households quarterly to record vital events: births, deaths, migrations, and marriages and marital disruptions. During quarterly household visits of FRAs for recording vital events, deaths were identified and VA interviews were conducted. With their consent, FRAs interviewed the closest caretakers and relatives who had lived with the deceased in the same household around terminal illness or death using the VA questionnaire within 6&#x02013;12 weeks after the date of death. In absence of the main caretaker or relative for a long period, a member of the same <italic>bari</italic> (a group of households close to each other by relationship) was interviewed. FRSs regularly supervised the fieldwork, and the public-health physician was available to provide technical support, such as clarification of questions when required.</p></sec><sec id="S0002-S20008"><title>Quality control</title><p>Scheduled revisits to 5% of randomly selected households were part of the quality control measures. FRSs visited the FRAs during data collection and reviewed surveillance data collected at household level, including the VA data. Immediate feedback was provided. The research officer completed mandatory checks and edits of all events before sending them for entry to the Surveillance and Data Resources Unit in Dhaka.</p></sec><sec id="S0002-S20009"><title>VA data management</title><p>HDSS data including VA were entered by three staff using R-base (DOS Version 3.0) software, under the supervision of two data management officers. The software was customised and allowed for inconsistency checks of the data during entry. All inconsistencies detected during data entry were resolved by checking the original forms or by returning the forms with errors to the field sites for necessary checking and corrections. Collected vital events were edited for consistencies and added to the longitudinal relational database.</p><sec><title>Assessment of CoD from VA</title><p>For assessment of probable CoD from VA symptoms we used the computer-automated probabilistic model &#x02018;InterVA-4 (version 4.02)&#x02019; in the structure of the 2012 WHO VA instruments (<xref rid="CIT0015" ref-type="bibr">15</xref>). The computer model is relatively fast, low-cost, and produces consistent and comparable CoD in comparison with the physician's (either single or a panel) review of VA for allocating CoD (<xref rid="CIT0016" ref-type="bibr">16</xref>&#x02013;<xref rid="CIT0019" ref-type="bibr">19</xref>). It speeds up VA interviews by eliminating the need for transcribing lengthy narratives related to death history (<xref rid="CIT0019" ref-type="bibr">19</xref>, <xref rid="CIT0020" ref-type="bibr">20</xref>). It processes a range of items of information about the background characteristics and circumstances of a death, details of any illness (signs and symptoms), or conditions leading to death, and previous medical history in a mathematical model based on Bayes&#x02019; theorem, and produced likely CoD (<xref rid="CIT0020" ref-type="bibr">21</xref>).</p></sec></sec><sec id="S0002-S20010"><title>Data analysis</title><p>AMK HDSS contributed to the INDEPTH multisite dataset, providing VA data and person-years of observations during 2004&#x02013;2010 (<xref rid="CIT0020" ref-type="bibr">20</xref>). HDSS data were used to compute percentage and rate (per 1,000 person-years) of deaths by age and sex for each site. VA data on symptoms and signs of illness or conditions collected were converted to the WHO 2012 standard (<xref rid="CIT0021" ref-type="bibr">21</xref>). It may be noted that for 17 deaths no symptom or sign of illness or condition was recorded in VA and these deaths were excluded from the analysis of CoD. We ran InterVA-4 (version 4.02) with options of low prevalence of HIV/AIDS and malaria deaths in the surveillance sites to produce likely biomedical CoDs. The InterVA-4 yields, for each case, up to three possible causes with estimated probabilities or an indeterminate result. The estimated probabilities for the first, second, and third most likely CoD were all summed, and if the sum of their probabilities was less than 100%, the residual component was then assigned as being indeterminate. CoD were broadly grouped into communicable diseases (CDs), non-communicable diseases (NCDs), perinatal and neonatal causes, pregnancy-related deaths, injury and other external causes, or indeterminate. CSMFs per 100 deaths by sex within the sites and by sites were estimated to exhibit sex and areal differentials. The differentials were tested for statistical significance using <italic>Z</italic>-value at <italic>p</italic>&#x0003c;0.05.</p></sec></sec><sec sec-type="results" id="S0003"><title>Results</title><p>The midyear population data in 2010 revealed that Abhoynagar had a smaller proportion of young population (age below 15) and larger proportion of working-age male population (aged 15&#x02013;49) as compared to Mirsarai (<xref ref-type="table" rid="T0001">Table 1</xref>). The proportion of old adults (aged 50&#x02013;64) and elderly were comparable between the sites. The age-specific sex ratios of the male per 100 female person-years were less skewed in Abhoynagar than Mirsarai, with sex ratio being the lowest in the age group 15&#x02013;49 due to higher out-migration of men aged 15&#x02013;49 in the latter site. The average ages of the males and females (29.8 and 29.2, respectively) in Abhoynagar were higher than the average ages of the males and females (27.3 and 28.3, respectively) in Mirsarai.</p><table-wrap id="T0001" position="float"><label>Table 1</label><caption><p>Age distribution (in%) and sex ratio of males per 100 females of the midyear population in 2010 in each site, AMK HDSS 2010</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1"/><th align="center" colspan="3" rowspan="1">Abhoynagar</th><th align="center" colspan="3" rowspan="1">Mirsarai</th></tr><tr><th align="left" rowspan="1" colspan="1"/><th colspan="3" rowspan="1"> <hr/> </th><th colspan="3" rowspan="1"> <hr/> </th></tr><tr><th align="left" rowspan="1" colspan="1">Age group (in years)</th><th align="center" rowspan="1" colspan="1">Male (<italic>n</italic>=17,489)</th><th align="center" rowspan="1" colspan="1">Female (<italic>n</italic>=17,375)</th><th align="center" rowspan="1" colspan="1">Sex ratio (100*M/F)</th><th align="center" rowspan="1" colspan="1">Male (<italic>n</italic>=18,321)</th><th align="center" rowspan="1" colspan="1">Female (<italic>n</italic>=20,929)</th><th align="center" rowspan="1" colspan="1">Sex ratio (100*M/F)</th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">0&#x02013;4</td><td align="center" rowspan="1" colspan="1">8.7</td><td align="center" rowspan="1" colspan="1">8.6</td><td align="center" rowspan="1" colspan="1">100.9</td><td align="center" rowspan="1" colspan="1">10.9</td><td align="center" rowspan="1" colspan="1">9.1</td><td align="center" rowspan="1" colspan="1">105.5</td></tr><tr><td align="left" rowspan="1" colspan="1">5&#x02013;14</td><td align="center" rowspan="1" colspan="1">19.6</td><td align="center" rowspan="1" colspan="1">18.8</td><td align="center" rowspan="1" colspan="1">104.9</td><td align="center" rowspan="1" colspan="1">22.8</td><td align="center" rowspan="1" colspan="1">19.8</td><td align="center" rowspan="1" colspan="1">100.9</td></tr><tr><td align="left" rowspan="1" colspan="1">15&#x02013;49</td><td align="center" rowspan="1" colspan="1">53.2</td><td align="center" rowspan="1" colspan="1">55.6</td><td align="center" rowspan="1" colspan="1">96.3</td><td align="center" rowspan="1" colspan="1">48.6</td><td align="center" rowspan="1" colspan="1">54.9</td><td align="center" rowspan="1" colspan="1">77.5</td></tr><tr><td align="left" rowspan="1" colspan="1">50&#x02013;64</td><td align="center" rowspan="1" colspan="1">13.0</td><td align="center" rowspan="1" colspan="1">11.3</td><td align="center" rowspan="1" colspan="1">115.8</td><td align="center" rowspan="1" colspan="1">12.2</td><td align="center" rowspan="1" colspan="1">10.7</td><td align="center" rowspan="1" colspan="1">99.4</td></tr><tr><td align="left" rowspan="1" colspan="1">65 +</td><td align="center" rowspan="1" colspan="1">5.5</td><td align="center" rowspan="1" colspan="1">5.6</td><td align="center" rowspan="1" colspan="1">98.9</td><td align="center" rowspan="1" colspan="1">5.5</td><td align="center" rowspan="1" colspan="1">5.6</td><td align="center" rowspan="1" colspan="1">86.6</td></tr><tr><td align="left" rowspan="1" colspan="1">Mean age&#x000b1;SD<xref ref-type="table-fn" rid="TF0001">a</xref></td><td align="center" rowspan="1" colspan="1">29.8&#x000b1;20.0</td><td align="center" rowspan="1" colspan="1">29.2&#x000b1;19.6</td><td align="center" rowspan="1" colspan="1">100.7</td><td align="center" rowspan="1" colspan="1">27.3&#x000b1;20.4</td><td align="center" rowspan="1" colspan="1">28.3&#x000b1;19.7</td><td align="center" rowspan="1" colspan="1">87.5</td></tr></tbody></table><table-wrap-foot><fn id="TF0001"><label>a</label><p>Means and SDs (standard deviations) were estimated from individual's age.</p></fn><fn><p><italic>n</italic>=size of the midyear population.</p></fn></table-wrap-foot></table-wrap><p>There were 1,384 deaths in Abhoynagar surveillance site and 1,847 deaths in Mirsarai surveillance site during 2004&#x02013;2010, yielding crude death rates of 5.8 and 6.7 per 1,000 person-years, respectively (<xref ref-type="table" rid="T0002">Table 2</xref>). Annual crude death rates and death rates in all age groups, except in the age group 1&#x02013;14 did not exhibit declining trends in both sites during 2004&#x02013;2010 (data not shown). There was a difference in age and sex patterns of mortality in these two rural sites. Infant, particularly neonatal, mortality rate (<italic>p</italic>&#x0003c;0.01) was critically higher in Abhoynagar than in Mirsarai, where mortality rates were higher in all age groups except in the infancy. The overall sex ratio of the male to female death rates was less skewed in Abhoynagar than Mirsarai, with a marked difference between age groups in each site. The sex ratios were more skewed in the neonatal, post-neonatal, and 1&#x02013;4-year age groups in Abhoynagar than Mirsarai, where sex ratios were skewed in all age groups 5 years and above.</p><table-wrap id="T0002" position="float"><label>Table 2</label><caption><p>Distribution of deaths, death rates,<xref ref-type="table-fn" rid="TF0002">a</xref> and sex ratios<xref ref-type="table-fn" rid="TF0003">b</xref> of death rates by age in Abhoynagar and Mirsarai sites, AMK HDSS 2004&#x02013;2010</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1"/><th align="center" colspan="3" rowspan="1">Abhoynagar</th><th align="center" colspan="3" rowspan="1">Mirsarai</th><th align="center" rowspan="1" colspan="1"/></tr><tr><th align="left" rowspan="1" colspan="1"/><th colspan="3" rowspan="1"> <hr/> </th><th colspan="3" rowspan="1"> <hr/> </th></tr><tr><th align="left" rowspan="1" colspan="1">Age group</th><th align="center" rowspan="1" colspan="1"># Deaths (%)</th><th align="center" rowspan="1" colspan="1">Death rate<xref ref-type="table-fn" rid="TF0002">a</xref> </th><th align="center" rowspan="1" colspan="1">Sex ratio<xref ref-type="table-fn" rid="TF0003">b</xref> </th><th align="center" rowspan="1" colspan="1">% (# Deaths)</th><th align="center" rowspan="1" colspan="1">Death rate<xref ref-type="table-fn" rid="TF0002">a</xref> </th><th align="center" rowspan="1" colspan="1">Sex ratio<xref ref-type="table-fn" rid="TF0003">b</xref> </th><th align="center" rowspan="1" colspan="1"><italic>P</italic>-value for the comparison of the death rates between sites</th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">0&#x02013;28 days<xref ref-type="table-fn" rid="TF0004">c</xref> </td><td align="center" rowspan="1" colspan="1">156 (11.3)</td><td align="center" rowspan="1" colspan="1">34.3</td><td align="center" rowspan="1" colspan="1">174.0</td><td align="center" rowspan="1" colspan="1">131 (7.1)</td><td align="center" rowspan="1" colspan="1">21.8</td><td align="center" rowspan="1" colspan="1">108.6</td><td align="center" rowspan="1" colspan="1"> <italic>p</italic>&#x0003c;0.01</td></tr><tr><td align="left" rowspan="1" colspan="1">29 days&#x02013;11 months<xref ref-type="table-fn" rid="TF0004">c</xref> </td><td align="center" rowspan="1" colspan="1">49 (3.5)</td><td align="center" rowspan="1" colspan="1">10.8</td><td align="center" rowspan="1" colspan="1">48.6</td><td align="center" rowspan="1" colspan="1">57 (3.1)</td><td align="center" rowspan="1" colspan="1">9.5</td><td align="center" rowspan="1" colspan="1">74.0</td><td align="center" rowspan="1" colspan="1"> <italic>p</italic>&#x0003c;0.32</td></tr><tr><td align="left" rowspan="1" colspan="1">1&#x02013;4 years</td><td align="center" rowspan="1" colspan="1">32 (2.3)</td><td align="center" rowspan="1" colspan="1">1.7</td><td align="center" rowspan="1" colspan="1">160.3</td><td align="center" rowspan="1" colspan="1">78 (4.2)</td><td align="center" rowspan="1" colspan="1">3.2</td><td align="center" rowspan="1" colspan="1">92.3</td><td align="center" rowspan="1" colspan="1"> <italic>p</italic>&#x0003c;0.01</td></tr><tr><td align="left" rowspan="1" colspan="1">5&#x02013;14 years</td><td align="center" rowspan="1" colspan="1">21 (1.5)</td><td align="center" rowspan="1" colspan="1">0.4</td><td align="center" rowspan="1" colspan="1">87.2</td><td align="center" rowspan="1" colspan="1">45 (2.4)</td><td align="center" rowspan="1" colspan="1">0.8</td><td align="center" rowspan="1" colspan="1">146.4</td><td align="center" rowspan="1" colspan="1"> <italic>p</italic>&#x0003c;0.05</td></tr><tr><td align="left" rowspan="1" colspan="1">15&#x02013;49 years</td><td align="center" rowspan="1" colspan="1">195 (14.1)</td><td align="center" rowspan="1" colspan="1">1.5</td><td align="center" rowspan="1" colspan="1">107.0</td><td align="center" rowspan="1" colspan="1">282 (15.3)</td><td align="center" rowspan="1" colspan="1">2.0</td><td align="center" rowspan="1" colspan="1">157.9</td><td align="center" rowspan="1" colspan="1"> <italic>p</italic>&#x0003c;0.01</td></tr><tr><td align="left" rowspan="1" colspan="1">50&#x02013;64years</td><td align="center" rowspan="1" colspan="1">199 (14.4)</td><td align="center" rowspan="1" colspan="1">7.8</td><td align="center" rowspan="1" colspan="1">118.6</td><td align="center" rowspan="1" colspan="1">327 (17.7)</td><td align="center" rowspan="1" colspan="1">11.8</td><td align="center" rowspan="1" colspan="1">177.1</td><td align="center" rowspan="1" colspan="1"> <italic>p</italic>&#x0003c;0.01</td></tr><tr><td align="left" rowspan="1" colspan="1">65+years</td><td align="center" rowspan="1" colspan="1">732 (52.9)</td><td align="center" rowspan="1" colspan="1">59.4</td><td align="center" rowspan="1" colspan="1">107.6</td><td align="center" rowspan="1" colspan="1">927 (50.2)</td><td align="center" rowspan="1" colspan="1">63.5</td><td align="center" rowspan="1" colspan="1">128.8</td><td align="center" rowspan="1" colspan="1"> <italic>p</italic>&#x0003c;0.15</td></tr><tr><td align="left" rowspan="1" colspan="1">All ages</td><td align="center" rowspan="1" colspan="1">1,384 (100.0)</td><td align="center" rowspan="1" colspan="1">5.8</td><td align="center" rowspan="1" colspan="1">114.0</td><td align="center" rowspan="1" colspan="1">1,847 (100.0)</td><td align="center" rowspan="1" colspan="1">6.7</td><td align="center" rowspan="1" colspan="1">135.2</td><td align="center" rowspan="1" colspan="1"> <italic>p</italic>&#x0003c;0.01</td></tr><tr><td align="left" rowspan="1" colspan="1">#Person-years</td><td colspan="3" align="center" rowspan="1">237,876</td><td colspan="3" align="center" rowspan="1">275,853</td><td align="left" rowspan="1" colspan="1"/></tr></tbody></table><table-wrap-foot><fn id="TF0002"><label>a</label><p>Per 1,000 person-years of observation</p></fn><fn id="TF0003"><label>b</label><p>ratio of the male to female death rates (100*M/F)</p></fn><fn id="TF0004"><label>c</label><p>per 1,000 person-years in infancy.</p></fn></table-wrap-foot></table-wrap><sec id="S0003-S20001"><title>Differences in CoD in two surveillance sites</title><p>The distribution of CSMFs by site shows the differences in broad CoD categories (<xref ref-type="table" rid="T0003">Table 3</xref>). It may be noted that the InterVA-4 could not assign a specific cause to 10.3% (10.9% in Abhoynagar and 9.8% in Mirsarai) of all deaths. CSMF due to CDs was lower (16.7% vs. 21.2%) in Abhoynagar than Mirsarai. The most common CDs were acute respiratory infections (ARI) including pneumonia and pulmonary tuberculosis with differences between the sites. CSMFs for ARI/pneumonia (9.4% vs. 7.1%) and tuberculosis (9.0% vs. 6.4%) were higher in Mirsarai than Abhoynagar. More than half of the deaths were caused by NCDs with no significant differences between the two sites (55.3 and 56.2% in Mirsarai and Abhoynagar, respectively). However, CSMF due to stroke was higher (19.3% vs. 16.5%) in Abhoynagar than Mirsarai, whereas CSMF due to cardiac diseases was higher in Mirsarai (11.6% vs. 8.9%). CSMF due to malignant neoplasms was 13.2% in each site. Malignancies were more frequent in the digestive system (5.4 and 5.6%, respectively) followed by the respiratory system (3.8 and 5.0%, respectively) in both sites. Maternal causes accounted for 0.8&#x02013;0.9% of the deaths in each site, but perinatal and neonatal causes were higher (7.6% vs. 5.3%) in Abhoynagar than Mirsarai. Though CSMF due to injury and other external causes was comparable (7.5&#x02013;7.7%) between the two sites, accidental drowning (2.3% vs. 1.4%) and road traffic accidents (1.9% vs. 1.0%) were more frequent in Mirsarai than Abhoynagar, whereas intentional self-harm was more frequent in Abhoynagar (3.5% vs. 1.5%).</p><table-wrap id="T0003" position="float"><label>Table 3</label><caption><p>Cause-specific mortality fractions (in%) by sex and site, AMK HDSS 2004&#x02013;2010</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1"/><th align="center" colspan="3" rowspan="1">Abhoynagar</th><th align="center" colspan="3" rowspan="1">Mirsarai</th></tr><tr><th align="left" rowspan="1" colspan="1"/><th colspan="3" rowspan="1"> <hr/> </th><th colspan="3" rowspan="1"> <hr/> </th></tr><tr><th align="left" rowspan="1" colspan="1">Major causes of death</th><th align="center" rowspan="1" colspan="1">Male</th><th align="center" rowspan="1" colspan="1">Female</th><th align="center" rowspan="1" colspan="1">Total</th><th align="center" rowspan="1" colspan="1">Male</th><th align="center" rowspan="1" colspan="1">Female</th><th align="center" rowspan="1" colspan="1">Total</th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">Communicable diseases</td><td align="center" rowspan="1" colspan="1">16.8</td><td align="center" rowspan="1" colspan="1">16.6</td><td align="center" rowspan="1" colspan="1">16.7</td><td align="center" rowspan="1" colspan="1">21.6</td><td align="center" rowspan="1" colspan="1">20.7</td><td align="center" rowspan="1" colspan="1">21.2<xref ref-type="table-fn" rid="TF0007">**</xref></td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;ARI or pneumonia</td><td align="center" rowspan="1" colspan="1">5.5</td><td align="center" rowspan="1" colspan="1">8.8<xref ref-type="table-fn" rid="TF0006">*</xref></td><td align="center" rowspan="1" colspan="1">7.1</td><td align="center" rowspan="1" colspan="1">8.9</td><td align="center" rowspan="1" colspan="1">10.0</td><td align="center" rowspan="1" colspan="1">9.4<xref ref-type="table-fn" rid="TF0006">*</xref></td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Pulmonary tuberculosis</td><td align="center" rowspan="1" colspan="1">8.6</td><td align="center" rowspan="1" colspan="1">3.9<xref ref-type="table-fn" rid="TF0007">**</xref></td><td align="center" rowspan="1" colspan="1">6.4</td><td align="center" rowspan="1" colspan="1">10.4</td><td align="center" rowspan="1" colspan="1">7.3<xref ref-type="table-fn" rid="TF0006">*</xref></td><td align="center" rowspan="1" colspan="1">9.0<xref ref-type="table-fn" rid="TF0007">**</xref></td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Other infections</td><td align="center" rowspan="1" colspan="1">2.7</td><td align="center" rowspan="1" colspan="1">3.9</td><td align="center" rowspan="1" colspan="1">3.2</td><td align="center" rowspan="1" colspan="1">2.3</td><td align="center" rowspan="1" colspan="1">3.4</td><td align="center" rowspan="1" colspan="1">2.8</td></tr><tr><td align="left" rowspan="1" colspan="1">NCD</td><td align="center" rowspan="1" colspan="1">54.6</td><td align="center" rowspan="1" colspan="1">58.1</td><td align="center" rowspan="1" colspan="1">56.2</td><td align="center" rowspan="1" colspan="1">53.4</td><td align="center" rowspan="1" colspan="1">56.9</td><td align="center" rowspan="1" colspan="1">55.3</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Stroke</td><td align="center" rowspan="1" colspan="1">17.8</td><td align="center" rowspan="1" colspan="1">20.9</td><td align="center" rowspan="1" colspan="1">19.3</td><td align="center" rowspan="1" colspan="1">13.8</td><td align="center" rowspan="1" colspan="1">20.0<xref ref-type="table-fn" rid="TF0007">**</xref></td><td align="center" rowspan="1" colspan="1">16.5<xref ref-type="table-fn" rid="TF0006">*</xref></td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Cardiac disease</td><td align="center" rowspan="1" colspan="1">10.2</td><td align="center" rowspan="1" colspan="1">7.3</td><td align="center" rowspan="1" colspan="1">8.9</td><td align="center" rowspan="1" colspan="1">13.5</td><td align="center" rowspan="1" colspan="1">9.2<xref ref-type="table-fn" rid="TF0007">**</xref></td><td align="center" rowspan="1" colspan="1">11.6</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;COPD or asthma</td><td align="center" rowspan="1" colspan="1">6.9</td><td align="center" rowspan="1" colspan="1">5.7</td><td align="center" rowspan="1" colspan="1">6.3</td><td align="center" rowspan="1" colspan="1">5.4</td><td align="center" rowspan="1" colspan="1">4.7</td><td align="center" rowspan="1" colspan="1">5.1</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Neoplasm</td><td align="center" rowspan="1" colspan="1">12.6</td><td align="center" rowspan="1" colspan="1">13.8</td><td align="center" rowspan="1" colspan="1">13.2</td><td align="center" rowspan="1" colspan="1">13.9</td><td align="center" rowspan="1" colspan="1">12.3</td><td align="center" rowspan="1" colspan="1">13.2</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;&#x02003;Digestive system</td><td align="center" rowspan="1" colspan="1">4.1</td><td align="center" rowspan="1" colspan="1">7.5<xref ref-type="table-fn" rid="TF0007">**</xref></td><td align="center" rowspan="1" colspan="1">5.6</td><td align="center" rowspan="1" colspan="1">5.1</td><td align="center" rowspan="1" colspan="1">5.8</td><td align="center" rowspan="1" colspan="1">5.4</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;&#x02003;Respiratory system</td><td align="center" rowspan="1" colspan="1">5.5</td><td align="center" rowspan="1" colspan="1">1.8<xref ref-type="table-fn" rid="TF0007">**</xref></td><td align="center" rowspan="1" colspan="1">3.8</td><td align="center" rowspan="1" colspan="1">6.2</td><td align="center" rowspan="1" colspan="1">3.4<xref ref-type="table-fn" rid="TF0007">**</xref></td><td align="center" rowspan="1" colspan="1">5.0</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;&#x02003;Other neoplasm</td><td align="center" rowspan="1" colspan="1">3.0</td><td align="center" rowspan="1" colspan="1">4.6</td><td align="center" rowspan="1" colspan="1">3.7</td><td align="center" rowspan="1" colspan="1">2.6</td><td align="center" rowspan="1" colspan="1">3.0</td><td align="center" rowspan="1" colspan="1">2.8</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Acute abdomen</td><td align="center" rowspan="1" colspan="1">1.4</td><td align="center" rowspan="1" colspan="1">2.2</td><td align="center" rowspan="1" colspan="1">1.8</td><td align="center" rowspan="1" colspan="1">2.0</td><td align="center" rowspan="1" colspan="1">4.4</td><td align="center" rowspan="1" colspan="1">3.1</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Liver cirrhosis</td><td align="center" rowspan="1" colspan="1">1.1</td><td align="center" rowspan="1" colspan="1">1.5</td><td align="center" rowspan="1" colspan="1">1.3</td><td align="center" rowspan="1" colspan="1">1.2</td><td align="center" rowspan="1" colspan="1">0.7</td><td align="center" rowspan="1" colspan="1">1.0</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Diabetes mellitus</td><td align="center" rowspan="1" colspan="1">3.1</td><td align="center" rowspan="1" colspan="1">3.9</td><td align="center" rowspan="1" colspan="1">3.5</td><td align="center" rowspan="1" colspan="1">2.9</td><td align="center" rowspan="1" colspan="1">2.7</td><td align="center" rowspan="1" colspan="1">2.8</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Severe anaemia/malnutrition</td><td align="center" rowspan="1" colspan="1">0.5</td><td align="center" rowspan="1" colspan="1">1.1</td><td align="center" rowspan="1" colspan="1">0.8</td><td align="center" rowspan="1" colspan="1">0.5</td><td align="center" rowspan="1" colspan="1">1.2</td><td align="center" rowspan="1" colspan="1">0.8</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Other NCD</td><td align="center" rowspan="1" colspan="1">1.0</td><td align="center" rowspan="1" colspan="1">1.6</td><td align="center" rowspan="1" colspan="1">1.2</td><td align="center" rowspan="1" colspan="1">0.7</td><td align="center" rowspan="1" colspan="1">1.7</td><td align="center" rowspan="1" colspan="1">1.2</td></tr><tr><td align="left" rowspan="1" colspan="1">Perinatal and neonatal causes</td><td align="center" rowspan="1" colspan="1">9.1</td><td align="center" rowspan="1" colspan="1">5.8<xref ref-type="table-fn" rid="TF0006">*</xref></td><td align="center" rowspan="1" colspan="1">7.6</td><td align="center" rowspan="1" colspan="1">5.1</td><td align="center" rowspan="1" colspan="1">5.6</td><td align="center" rowspan="1" colspan="1">5.3<xref ref-type="table-fn" rid="TF0007">**</xref></td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Neonatal pneumonia</td><td align="center" rowspan="1" colspan="1">3.5</td><td align="center" rowspan="1" colspan="1">2.6</td><td align="center" rowspan="1" colspan="1">3.1</td><td align="center" rowspan="1" colspan="1">1.9</td><td align="center" rowspan="1" colspan="1">1.5</td><td align="center" rowspan="1" colspan="1">1.7<xref ref-type="table-fn" rid="TF0006">*</xref></td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Birth asphyxia</td><td align="center" rowspan="1" colspan="1">1.4</td><td align="center" rowspan="1" colspan="1">1.4</td><td align="center" rowspan="1" colspan="1">1.4</td><td align="center" rowspan="1" colspan="1">1.7</td><td align="center" rowspan="1" colspan="1">1.3</td><td align="center" rowspan="1" colspan="1">1.5</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Prematurity</td><td align="center" rowspan="1" colspan="1">1.4</td><td align="center" rowspan="1" colspan="1">0.2</td><td align="center" rowspan="1" colspan="1">0.9</td><td align="center" rowspan="1" colspan="1">0.5</td><td align="center" rowspan="1" colspan="1">1.1</td><td align="center" rowspan="1" colspan="1">0.7</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Other neonatal cause</td><td align="center" rowspan="1" colspan="1">2.8</td><td align="center" rowspan="1" colspan="1">1.5</td><td align="center" rowspan="1" colspan="1">2.2</td><td align="center" rowspan="1" colspan="1">1.0</td><td align="center" rowspan="1" colspan="1">1.7</td><td align="center" rowspan="1" colspan="1">1.3</td></tr><tr><td align="left" rowspan="1" colspan="1">Maternal cause</td><td align="center" rowspan="1" colspan="1">NA</td><td align="center" rowspan="1" colspan="1">1.8</td><td align="center" rowspan="1" colspan="1">0.8</td><td align="center" rowspan="1" colspan="1">NA</td><td align="center" rowspan="1" colspan="1">2.0</td><td align="center" rowspan="1" colspan="1">0.9</td></tr><tr><td align="left" rowspan="1" colspan="1">Injury and other external causes</td><td align="center" rowspan="1" colspan="1">7.9</td><td align="center" rowspan="1" colspan="1">7.4</td><td align="center" rowspan="1" colspan="1">7.7</td><td align="center" rowspan="1" colspan="1">9.3</td><td align="center" rowspan="1" colspan="1">5.4</td><td align="center" rowspan="1" colspan="1">7.5</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Accidental drowning</td><td align="center" rowspan="1" colspan="1">1.6</td><td align="center" rowspan="1" colspan="1">1.1</td><td align="center" rowspan="1" colspan="1">1.4</td><td align="center" rowspan="1" colspan="1">2.2</td><td align="center" rowspan="1" colspan="1">2.5</td><td align="center" rowspan="1" colspan="1">2.3</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Road traffic accident</td><td align="center" rowspan="1" colspan="1">1.4</td><td align="center" rowspan="1" colspan="1">0.5</td><td align="center" rowspan="1" colspan="1">1.0</td><td align="center" rowspan="1" colspan="1">3.3</td><td align="center" rowspan="1" colspan="1">0.2<xref ref-type="table-fn" rid="TF0007">**</xref></td><td align="center" rowspan="1" colspan="1">1.9<xref ref-type="table-fn" rid="TF0006">*</xref></td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;International self-harm</td><td align="center" rowspan="1" colspan="1">2.5</td><td align="center" rowspan="1" colspan="1">5.1<xref ref-type="table-fn" rid="TF0007">**</xref></td><td align="center" rowspan="1" colspan="1">3.5</td><td align="center" rowspan="1" colspan="1">1.3</td><td align="center" rowspan="1" colspan="1">1.7</td><td align="center" rowspan="1" colspan="1">1.5<xref ref-type="table-fn" rid="TF0007">**</xref></td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Assault</td><td align="center" rowspan="1" colspan="1">1.3</td><td align="center" rowspan="1" colspan="1">0.2<xref ref-type="table-fn" rid="TF0006">*</xref></td><td align="center" rowspan="1" colspan="1">0.8</td><td align="center" rowspan="1" colspan="1">1.2</td><td align="center" rowspan="1" colspan="1">0.4<xref ref-type="table-fn" rid="TF0006">*</xref></td><td align="center" rowspan="1" colspan="1">0.8</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Other external cause</td><td align="center" rowspan="1" colspan="1">1.4</td><td align="center" rowspan="1" colspan="1">0.7</td><td align="center" rowspan="1" colspan="1">1.1</td><td align="center" rowspan="1" colspan="1">1.3</td><td align="center" rowspan="1" colspan="1">0.6</td><td align="center" rowspan="1" colspan="1">1.0</td></tr><tr><td align="left" rowspan="1" colspan="1">Indeterminate<xref ref-type="table-fn" rid="TF0005">a</xref> </td><td align="center" rowspan="1" colspan="1">11.5</td><td align="center" rowspan="1" colspan="1">10.2</td><td align="center" rowspan="1" colspan="1">10.9</td><td align="center" rowspan="1" colspan="1">10.0</td><td align="center" rowspan="1" colspan="1">9.4</td><td align="center" rowspan="1" colspan="1">9.8</td></tr><tr><td align="left" rowspan="1" colspan="1">Number of deaths</td><td align="center" rowspan="1" colspan="1">743</td><td align="center" rowspan="1" colspan="1">641</td><td align="center" rowspan="1" colspan="1">1,384</td><td align="center" rowspan="1" colspan="1">1,016</td><td align="center" rowspan="1" colspan="1">831</td><td align="center" rowspan="1" colspan="1">1,847</td></tr></tbody></table><table-wrap-foot><fn><p>ARI=acute respiratory infection; COPD=chronic obstetric pulmonary diseases; NA=not applicable; NCD=non-communicable diseases.</p></fn><fn id="TF0005"><label>a</label><p>Excluded 17 deaths for which no symptom or sign of illness was record in VA from determining cause of death, but included in the indeterminate.</p></fn><fn id="TF0006"><label>*</label><p><italic>p</italic>&#x0003c;0.05</p></fn><fn id="TF0007"><label>**</label><p><italic>p</italic>&#x0003c;0.01 (compared between &#x02018;Male&#x02019; and &#x02018;Female&#x02019; within the site or between totals of sites).</p></fn></table-wrap-foot></table-wrap></sec><sec id="S0003-S20002"><title>Sex difference in CoD within the surveillance sites</title><p>The breakdown of CSMFs by sex shows biosocial differences in mortality risks (<xref ref-type="table" rid="T0003">Table 3</xref>). Although the CSMF due to CDs exhibited sex parity in each site, the CSMF for pulmonary tuberculosis was higher for males than females in both sites (8.6% vs. 3.9% in Abhoynagar and 10.4% vs. 7.3% in Mirsarai) and for ARI/pneumonia it was higher for males in Abhoynagar (8.8% vs. 5.5%) only. CSMFs due to NCDs were comparable between females and males within the site. Stroke was more frequent (20.0% vs. 13.8%) among females in Mirsarai and cardiac disease was more frequent among males in both sites (13.5% vs. 9.2% and 10.2% vs. 7.3% in Abhoynagar). There was no sex difference in CSMF due to malignant neoplasms, but malignancies in the digestive system were more frequent among females (7.5% vs. 4.1%) in Abhoynagar, and in the respiratory system were more frequent among males in both sites (5.5% vs. 1.8% in Abhoynagar and 6.2% vs. 3.4% in Mirsarai). No sex difference was noted in CSMF due to diabetes mellitus and liver cirrhosis, but acute abdomen was more frequent among females than males (3.4% vs. 1.8%) in both sites.</p><p>Perinatal and neonatal causes exhibited sex differences in favour of females in Abhoynagar, and conditions relating to pregnancy accounted for 13.3% of the deaths of adult (aged 15&#x02013;49) females. CSMF due to external causes exhibited no sex difference in Abhoynagar, but against males (9.3% vs. 5.4%) in Mirsarai. Although the frequency of accidental drowning did not vary by sex, road traffic accident (1.4% vs. 0.5% in Abhoynagar and 3.3% vs. 0.2% in Mirsarai) and assault (1.3% vs. 0.2% in Abhoynagar and 1.2% vs. 0.4% in Mirsarai) were more frequent among males than females, among whom intentional self-harm was more frequent, particularly in Abhoynagar (5.1% vs. 2.5%).</p></sec><sec id="S0003-S20003"><title>The rank-order of CoD by age</title><p> <xref ref-type="table" rid="T0004">Table 4</xref> shows marked variations in the rank-order (measured with CSMFs) of top 10 CoDs between age groups. It may be noted that the percentage of indeterminate cases varied by age; highest (27.3%) among the neonates, and lowest (1.7%) among children aged 1&#x02013;4. Neonatal deaths were due mostly to ARI/pneumonia (25.7%), followed by birth asphyxia (16.8%), prematurity (9.0%), and sepsis (4.1%), summing up to 55.6% of the deaths. ARI/pneumonia (65.4%) and diarrhoea (14.4%), accounted for 79.8% of the post-neonatal deaths. More than half (55.4%; with accidental drowning accounting for 49%) of the child (aged 1&#x02013;4) deaths was due to injury and other external cause, followed by ARI/pneumonia (26.3%), diarrhoea (9.2%), and malnutrition (4.2%), totalling to 95.1% of the deaths. The leading CoD of older children (aged 5&#x02013;14) were injury (38.7%, drowning accounting for 9.8%), ARI/pneumonia (13.6%), tuberculosis (7%), and acute abdomen (6.9%), totalling to 66.2% of the deaths.</p><table-wrap id="T0004" position="float"><label>Table 4</label><caption><p>Cause-specific mortality fractions [in%] by age group of the deceased (both sites combined), AMK HDSS 2004&#x02013;2010</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1">0&#x02013;28 days (# deaths=287)</th><th align="center" rowspan="1" colspan="1">29 days&#x02013;&#x0003c;1 year (# deaths=106)</th><th align="center" rowspan="1" colspan="1">1&#x02013;4 years (# deaths=110)</th><th align="center" rowspan="1" colspan="1">5&#x02013;14 years (# deaths=66)</th><th align="center" rowspan="1" colspan="1">15&#x02013;49 years (# deaths=477)</th><th align="center" rowspan="1" colspan="1">50&#x02013;64 years (# deaths=526)</th><th align="center" rowspan="1" colspan="1">65+ years (# deaths=1,659)</th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">ARI/pneumonia [25.7%]</td><td align="left" rowspan="1" colspan="1">ARI/pneumonia [65.4%]</td><td align="left" rowspan="1" colspan="1">Injury [55.4%] Drowning [49.0%]</td><td align="left" rowspan="1" colspan="1">Injury [38.7%] Drowning [9.8%]</td><td align="left" rowspan="1" colspan="1">Injury [23.5%] Self-harm [11.9%]</td><td align="left" rowspan="1" colspan="1">Neoplasm [19.7%]</td><td align="left" rowspan="1" colspan="1">Stroke [26.8%]</td></tr><tr><td align="left" rowspan="1" colspan="1">Birth asphyxia [16.8%]</td><td align="left" rowspan="1" colspan="1">Diarrhoea [14.4%]</td><td align="left" rowspan="1" colspan="1">ARI/pneumonia [26.3%]</td><td align="left" rowspan="1" colspan="1">ARI/pneumonia [13.6%]</td><td align="left" rowspan="1" colspan="1">Neoplasm [16.5%]</td><td align="left" rowspan="1" colspan="1">Stroke [17.9%]</td><td align="left" rowspan="1" colspan="1">Neoplasm [14.6%]</td></tr><tr><td align="left" rowspan="1" colspan="1">Prematurity [9.0%]</td><td align="left" rowspan="1" colspan="1">Meningitis/encephalitis [2.6%]</td><td align="left" rowspan="1" colspan="1">Diarrhoea [9.2%]</td><td align="left" rowspan="1" colspan="1">Tuberculosis [7.0%]</td><td align="left" rowspan="1" colspan="1">Cardiac diseases [11.8%]</td><td align="left" rowspan="1" colspan="1">Cardiac diseases [17.4%]</td><td align="left" rowspan="1" colspan="1">Cardiac diseases [11.3%]</td></tr><tr><td align="left" rowspan="1" colspan="1">Sepsis [4.1%],</td><td align="left" rowspan="1" colspan="1">Injury [2.2%]</td><td align="left" rowspan="1" colspan="1">Malnutrition [4.2%]</td><td align="left" rowspan="1" colspan="1">Acute abdomen [6.9%]</td><td align="left" rowspan="1" colspan="1">Tuberculosis [9.8%]</td><td align="left" rowspan="1" colspan="1">Tuberculosis [10.5%]</td><td align="left" rowspan="1" colspan="1">Tuberculosis [8.8%]</td></tr><tr><td align="left" rowspan="1" colspan="1">Malformation [2.4%]</td><td align="left" rowspan="1" colspan="1">Malnutrition [1.8%]</td><td align="left" rowspan="1" colspan="1">Tuberculosis [1.9%]</td><td align="left" rowspan="1" colspan="1">Meningitis/encephalitis [3.1%]</td><td align="left" rowspan="1" colspan="1">Stroke [6.4%]</td><td align="left" rowspan="1" colspan="1">COPD/asthma [5.2%]</td><td align="left" rowspan="1" colspan="1">COPD/asthma [8.7%]</td></tr><tr><td align="left" rowspan="1" colspan="1">Meningitis/encephalitis [2.3%]</td><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1">Stroke [3.0%]</td><td align="left" rowspan="1" colspan="1">Maternal cause [6.0%]</td><td align="left" rowspan="1" colspan="1">Acute abdomen [3.8%]</td><td align="left" rowspan="1" colspan="1">ARI/pneumonia [6.9%]</td></tr><tr><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1">Cardiac disease [2.7%]</td><td align="left" rowspan="1" colspan="1">ARI/pneumonia [4.1%]</td><td align="left" rowspan="1" colspan="1">Injury [3.8%]</td><td align="left" rowspan="1" colspan="1">Diabetes [4.4%]</td></tr><tr><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1">Acute abdomen [3.3%]</td><td align="left" rowspan="1" colspan="1">Diabetes [3.6%]</td><td align="left" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">Other [12.3%]</td><td align="left" rowspan="1" colspan="1">Other [5.5%]</td><td align="left" rowspan="1" colspan="1">Other [3.2%]</td><td align="left" rowspan="1" colspan="1">Other [13.4%]</td><td align="left" rowspan="1" colspan="1">Other [10.2%]</td><td align="left" rowspan="1" colspan="1">Other [10.5%]</td><td align="left" rowspan="1" colspan="1">Other [9.2%]</td></tr><tr><td align="left" rowspan="1" colspan="1">Indeterminate [27.3%]</td><td align="left" rowspan="1" colspan="1">Indeterminate [7.1%]</td><td align="left" rowspan="1" colspan="1">Indeterminate [1.7%]</td><td align="left" rowspan="1" colspan="1">Indeterminate [11.6%]</td><td align="left" rowspan="1" colspan="1">Indeterminate [8.4%]</td><td align="left" rowspan="1" colspan="1">Indeterminate [7.7%]</td><td align="left" rowspan="1" colspan="1">Indeterminate [9.3%]</td></tr></tbody></table></table-wrap><p> The most common CoD for young adults (aged 15&#x02013;49) was injury (23.5%, self-harm accounting for 11.9%), followed by malignant neoplasm (16.5%), cardiac diseases (11.8%), tuberculosis (9.8%), stroke (6.4%), and ARI/pneumonia (4.1%). These accounted for 72.1% of the deaths. Deaths from conditions related to pregnancy accounted for 13.3% of the adult female deaths. Of the older adult (aged 50&#x02013;64) deaths, malignant neoplasms (19.7%), stroke (17.9%), cardiac disease (17.4%), tuberculosis (10.5%), and COPD including asthma (5.2%) accounted for 70.7% of the deaths. The leading cause of elderly (aged 65 and older) deaths was stroke (26.8%), followed by malignant neoplasm (14.6%), cardiac diseases (11.3%), tuberculosis (8.8%), COPD (8.7%), ARI/pneumonia (6.9%), and diabetes (4.4%), totalling to 81.5% of all the deaths.</p></sec></sec><sec sec-type="discussion" id="S0004"><title>Discussion</title><p>The results revealed critically higher mortality, particularly among male neonates in Abhoynagar than Mirsarai, which does not match with lower crude death rate and use of maternal health services. In 2009, in Abhoynagar, 31% of the mothers who gave live births received recommended 4+ antenatal care visits and 64% received 1+ postnatal care visit compared to 24 and 16%, respectively, in Mirsarai (<xref rid="CIT0014" ref-type="bibr">14</xref>). However, the rates of institutional deliveries of live births in 2009 were comparable between these two sites (21% in Abhoynagar and 23% in Mirsarai). Male neonates are biologically more vulnerable, but vulnerability was critically higher in Abhoynagar, which is a surprise and we do not have any reasonable explanation. However, high neonatal mortality rate could be due to high rates of teenage marriage and fertility and colder ambient temperature in the winter season in Abhoynagar compared to Mirsarai (<xref rid="CIT0014" ref-type="bibr">14</xref>, <xref rid="CIT0022" ref-type="bibr">22</xref>). Teenage motherhood is associated with increased risks for pre-term delivery, low birth weight, and neonatal mortality (<xref rid="CIT0023" ref-type="bibr">23</xref>). Moreover, perinatal mortality steeply increased with a decrease in temperature in the winter below the temperature of 23&#x000b0;C (<xref rid="CIT0024" ref-type="bibr">24</xref>).</p><p>The mortality rates in all age groups except in infancy were lower in Abhoynagar than Mirsarai. This difference was due mostly to less frequent deaths from CDs, particularly from ARI/pneumonia and tuberculosis in Abhoynagar. Such site-specific statistics are important because they could help health managers in local-level planning of health services and designing appropriate measures that will save lives and improve economic conditions.</p><p>Sex differences in mortality rates in the age groups 5 years and above were in favour of females and they were much higher in Mirsarai than Abhoynagar. Why males in Mirsarai experienced excess mortality compared to females of the same age groups may be explained by the lower population sex ratio, which is determined by more out-migration (national or international) of males. The overall population sex ratio of males to 100 females in 2010 was 87 in Mirsarai compared to 101 in Abhoynagar, although the sex ratios in the age group 0&#x02013;4 were comparable (101 vs. 105) between sites (<xref ref-type="table" rid="T0001">Table 1</xref>). The sex ratio in the age group 15&#x02013;49 was even lower (77 vs. 96) in Mirsarai, perhaps due to economic and labour migration. In general, healthy individuals are more likely to undertake migration, leaving the less healthy ones at home. The healthy migrant effect in terms of mortality was observed in Germany, comparing Turkish migrants to Germans locals, and it could be due to self-selection at the time of immigration (<xref rid="CIT0025" ref-type="bibr">25</xref>).</p><p>The CoD patterns revealed sex differences in the health burdens of specific NCDs, tuberculosis, and injuries. Parallel sex differentials were observed in the distribution of certain morbidity conditions in the Bangladesh Demographic and Health Survey (BDHS) 2011, which screened for prevalence of hypertension and diabetes in women and men aged 35 and older (<xref rid="CIT0002" ref-type="bibr">2</xref>). More women than men (32% vs. 19%) were hypertensive, but diabetes (11%) was similar in both sexes. More than half of them, however, were not aware that they had the diseases. Another study noted similar sex differentials in the risk factors of NCDs. More men aged 25&#x02013;64 used tobacco products (68.2% vs. 32.7%) than women of similar ages. Women were more often overweight (15.2% vs. 10.8%) compared to men (<xref rid="CIT0026" ref-type="bibr">26</xref>).</p><p>As expected and also noted with the physician-coded CoD in a large-scale national survey BDHS 2011 (<xref rid="CIT0002" ref-type="bibr">2</xref>), the rank-order of the InterVA-4-coded CoD changed markedly by age group despite the methodological differences (in terms of questionnaires used, length of the recall period, assessment and categorisation of CoD, and time period). Except for neonates, the rank-order of the physician-coded CoD of the post-neonates and children (aged 1&#x02013;4) is comparable with the rank-order of the InterVA-4 coded CoD. There is a large variation in the distribution of CoD of neonates assessed by the InterVA-4 and by the physician. The leading physician-coded CoD was &#x02018;possible serious infections&#x02019; (24.3%), for which there is no comparable category in the InterVA-4. However, the sum of the physician-coded deaths due to infections (i.e. possible serious infections, ARI/pneumonia, and diarrhoea) was 34.6%, which is comparable to the sum of the InterVA-4-coded deaths due to ARI/pneumonia, sepsis, meningitis, and diarrhoea, totalling to 32.5%.</p><p>The rank-order of the top two causes of post-neonatal deaths coded by physician and InterVA-4 are comparable; ARI/pneumonia (66% in InterVA-4 and 52.9% in BDHS 2011), followed by diarrhoea (14.6% in InterVA-4 and 7.5% in BDHS 2011) (<xref rid="CIT0002" ref-type="bibr">2</xref>). The InterVA-4 coded causes of child deaths are also comparable in rank-order with the physician-coded causes; drowning (49.0% vs. 42.6%), followed by ARI/pneumonia (26.3% vs. 21.7%). Also the physician-coded top five CoDs of females aged 15&#x02013;49 reported in the Bangladesh Maternal Mortality and Health Care Survey 2010 were found comparable with the InterVA-4 coded top five CoDs of females of the same age group in these two sites. The comparability of the InterVa-4 coded CoDs, despite several methodological differences, with the physician-coded CoDs reveals the potentials of VA in HDSS sites to be used for planning and monitoring of the disease burdens not only in these sites but also in the regions as well as in the country with obvious subtle differences.</p><p>The patterns of CoDs revealed the prominence of NCDs compared to CDs in both sites, which has implications for the public health system to respond. Management of most NCDs is available at tertiary level hospitals, but it is not for reduction and prevention of the risks of developing NCDs. Many NCDs are, however, amenable to prevention through behavioural changes. Lifestyle and behaviours are linked to 20&#x02013;25% of the global burden of disease, which is likely to increase rapidly in poorer countries in the process of rapid urbanisation and demographic transition (<xref rid="CIT0027" ref-type="bibr">27</xref>). In Bangladesh, consumption of vegetables and fruits and regular exercise are at a low level, whereas use of tobacco products, excessive intake of salt, and abuse of substances are considerably high (<xref rid="CIT0026" ref-type="bibr">26</xref>). These risk factors are shared by a number of NCDs, so health-promotion directed towards these risk factors will address most simultaneously (<xref rid="CIT0028" ref-type="bibr">28</xref>). Strengthening behaviour change activities at the community level for promoting risk-reducing behaviour, expanding screening facilities for early detection of NCDs, and increasing compliance with effective medication can lower the disease burden, health expenditure, and loss of productivity and national health expenditure.</p><p>Injury and other external causes are the leading cause of mortality in the age groups 1&#x02013;4, 5&#x02013;14, and 15&#x02013;49 in rural communities. Particularly accidental drowning accounted for 88% of the injury-related deaths in the 1&#x02013;4 age group and 25% in the 5&#x02013;14 age group. Evidence-based interventions and community awareness are needed for lowering such deaths. One-fourth (24%) of the adult deaths were due to injury and other external causes, half (51%) of them were due to self-harm, and another 14% were due to assault. Physical and mental assaults often provoke self-harm, thus may have underestimated the share of assault. Violence against young women is quite high in South Asia including Bangladesh and is often perpetrated by their husbands or his family members (<xref rid="CIT0029" ref-type="bibr">29</xref>&#x02013;<xref rid="CIT0031" ref-type="bibr">31</xref>). Assessment of CoD from VA provided usually by victim's family members cannot divulge true cause &#x02013; a limitation of VA collecting from the deceased's family members.</p><p>Evidence-based planning of health services and logistics requires reliable and up-to-date public health statistics including CoD. In Bangladesh, the civil registration, particularly deaths with medical certification is too incomplete to generate such statistics. The national SVRS administered by the Bangladesh Bureau of Statistics, the Government of Bangladesh, records vital events including deaths from a representative sample of the population. Introduction of VA into the national SVRS and computer-automated coding of VAs of a nationally representative sample of deaths can generate CoD statistics on a regular basis for use in public health planning until the civil registration system well functions.</p><p>Health burdens posed by CDs, NCDs, and external causes are the major challenges to improving population health. The government Health, Population, and Nutrition Sector Development Program for 2011&#x02013;2016 includes a plan for expanding access to health services for controlling conventional (hypertension, diabetes, cancer, COPD, psychiatric illness, etc.) and non-conventional (road safety and injury and violence against women) NCDs (<xref rid="CIT0032" ref-type="bibr">32</xref>). The operation plan includes conducting training on NCD screening and management for health care providers at district and subdistrict levels, organising awareness-building workshops on injuries, and pilot screening and management of selected NCDs at the subdistrict level facilities, gradually expanding to the lower level facilities. The private health sector, particularly pharmacies in urban and rural areas, and workplace based prevention and screening, can play an important role in screening and referral. Prevention, early detection, and compliance with effective medication can save national health expenditure as NCDs require long-term care and bring catastrophic economic consequences for high out-of-pocket payments. Appropriate measures to minimise the catastrophic effects may include, but are not limited to, community-based health insurance, credit to cushion income loss, and social safety net programmes.</p></sec><sec sec-type="conclusions" id="S0005"><title>Conclusions</title><p>In conclusion, analyses of VA symptom data using InterVA-4 model revealed health burdens, with leading causes being stroke, neoplasms, cardiac diseases, ARI/pneumonia, tuberculosis, and COPD. External causes were more frequent among males, but self-harm was higher in Abhoynagar, particularly for females. The primary health care centres, currently equipped to manage CDs which is the outmost concern, must be equipped for prevention, screening, and management NCDs as well.</p></sec>

Multisensory stimuli elicit altered oscillatory brain responses at gamma frequencies in patients with schizophrenia

<p>Deficits in auditory and visual unisensory responses are well documented in patients with schizophrenia; however, potential abnormalities elicited from multisensory audio-visual stimuli are less understood. Further, schizophrenia patients have shown abnormal patterns in task-related and task-independent oscillatory brain activity, particularly in the gamma frequency band. We examined oscillatory responses to basic unisensory and multisensory stimuli in schizophrenia patients (<italic>N</italic> = 46) and healthy controls (<italic>N</italic> = 57) using magnetoencephalography (MEG). Time-frequency decomposition was performed to determine regions of significant changes in gamma band power by group in response to unisensory and multisensory stimuli relative to baseline levels. Results showed significant behavioral differences between groups in response to unisensory and multisensory stimuli. In addition, time-frequency analysis revealed significant decreases and increases in gamma-band power in schizophrenia patients relative to healthy controls, which emerged both early and late over both sensory and frontal regions in response to unisensory and multisensory stimuli. Unisensory gamma-band power predicted multisensory gamma-band power differently by group. Furthermore, gamma-band power in these regions predicted performance in select measures of the Measurement and Treatment Research to Improve Cognition in Schizophrenia (MATRICS) test battery differently by group. These results reveal a unique pattern of task-related gamma-band power in schizophrenia patients relative to controls that may indicate reduced inhibition in combination with impaired oscillatory mechanisms in patients with schizophrenia.</p>

<contrib contrib-type="author"><name><surname>Stone</surname><given-names>David B.</given-names></name><xref ref-type="aff" rid="aff1">¹</xref><uri xlink:type="simple" xlink:href="http://community.frontiersin.org/people/u/66900"/></contrib><contrib contrib-type="author"><name><surname>Coffman</surname><given-names>Brian A.</given-names></name><xref ref-type="aff" rid="aff1">¹</xref><xref ref-type="aff" rid="aff2">²</xref><uri xlink:type="simple" xlink:href="http://community.frontiersin.org/people/u/100078"/></contrib><contrib contrib-type="author"><name><surname>Bustillo</surname><given-names>Juan R.</given-names></name><xref ref-type="aff" rid="aff3">³</xref><uri xlink:type="simple" xlink:href="http://community.frontiersin.org/people/u/17879"/></contrib><contrib contrib-type="author"><name><surname>Aine</surname><given-names>Cheryl J.</given-names></name><xref ref-type="aff" rid="aff4">⁴</xref><uri xlink:type="simple" xlink:href="http://community.frontiersin.org/people/u/14573"/></contrib><contrib contrib-type="author"><name><surname>Stephen</surname><given-names>Julia M.</given-names></name><xref ref-type="aff" rid="aff1">¹</xref><xref ref-type="author-notes" rid="fn001">^*</xref><uri xlink:type="simple" xlink:href="http://community.frontiersin.org/people/u/6151"/></contrib>

Frontiers in Human Neuroscience

<sec sec-type="introduction" id="s1"><title>Introduction</title><p>On a moment-by-moment basis, behaviorally salient information reaches us from multiple senses and seamlessly shapes our perceptions and actions in the world. The capacity to organize the many stimuli arriving from different sensory modalities into a coherent and operable percept remains one of the marvels of the central nervous system. Despite a rich history exploring unisensory processes in the brain, understanding the processes involved in multisensory integration remains a central challenge in systems neuroscience.</p><p>Empirical and computational approaches have revealed some of the neural mechanisms subserving multisensory integration at the cellular and systems level. Based on both human and animal studies, it is clear that a broad network of cortical (sensory and association areas&#x02014;Barth et al., <xref rid="B7" ref-type="bibr">1995</xref>; Schroeder et al., <xref rid="B57" ref-type="bibr">2001</xref>; Falchier et al., <xref rid="B22" ref-type="bibr">2002</xref>; Brett-Green et al., <xref rid="B11" ref-type="bibr">2003</xref>) and subcortical (superior colliculus&#x02014;Meredith and Stein, <xref rid="B46" ref-type="bibr">1986</xref>; Wallace et al., <xref rid="B86" ref-type="bibr">1993</xref>, <xref rid="B87" ref-type="bibr">1998</xref>; Peck, <xref rid="B55" ref-type="bibr">1996</xref>; Bell et al., <xref rid="B9" ref-type="bibr">2001</xref>) areas are involved in multisensory processing (Ghazanfar and Schroeder, <xref rid="B28" ref-type="bibr">2006</xref>). Both non-invasive functional neuroimaging and invasive studies have confirmed either facilitation or suppression of unisensory responses with multisensory stimuli both at the single cell (Meredith and Stein, <xref rid="B46" ref-type="bibr">1986</xref>; Kadunce et al., <xref rid="B35" ref-type="bibr">1997</xref>; Fu et al., <xref rid="B25" ref-type="bibr">2003</xref>) and at the population level (Barth et al., <xref rid="B7" ref-type="bibr">1995</xref>; Molholm et al., <xref rid="B48" ref-type="bibr">2002</xref>, <xref rid="B49" ref-type="bibr">2006</xref>; Murray et al., <xref rid="B50" ref-type="bibr">2005</xref>; Stone et al., <xref rid="B68" ref-type="bibr">2011</xref>) indicating that multisensory responses often do not represent a simple linear summation of the unisensory responses. In conjunction with these neurophysiological results, behavioral studies also confirm that multisensory stimuli lead to facilitation of unisensory reaction times and improvements in accuracy under certain conditions (Calvert et al., <xref rid="B16" ref-type="bibr">2004</xref>). Additional studies provide evidence that multisensory integration is dependent upon both bottom-up sensory features, such as salience of sensory stimuli (Stein and Meredith, <xref rid="B65" ref-type="bibr">1993</xref>), and top-down cognitive processes, including attention (Busse et al., <xref rid="B13" ref-type="bibr">2005</xref>; Talsma et al., <xref rid="B72" ref-type="bibr">2007</xref>; Keitel et al., <xref rid="B37" ref-type="bibr">2013</xref>).</p><p>However, the mechanism by which sensory information from independent sensory modalities is integrated to generate the unified percept is still poorly understood. Certain brain regions contain cells that react to input from multiple sensory modalities (e.g., superior colliculus, association areas). Though studies examining these multisensory cells have played an important role in advancing research in this area, research at the cellular level does not provide a complete view of multisensory integration. Gamma-band oscillations (&#x0003e;30 Hz), in particular, are implicated in aspects of feature binding (e.g., linking objects within a visual scene) by establishing temporal synchrony both within and across cortical regions (Tallon-Baudry et al., <xref rid="B71" ref-type="bibr">1996</xref>; Von Stein and Sarnthein, <xref rid="B85" ref-type="bibr">2000</xref>). For example, conscious perception of stimuli is accompanied by increases in oscillatory synchrony at frequencies above 30 Hz (Melloni et al., <xref rid="B45" ref-type="bibr">2007</xref>). Additional results provide evidence that changes in gamma-band power are associated with perceptual and cognitive processing (Bertrand and Tallon-Baudry, <xref rid="B10" ref-type="bibr">2000</xref>; Tallon-Baudry, <xref rid="B70" ref-type="bibr">2009</xref>). Multisensory studies have further established that gamma-band oscillations play a role in cross-modal feature binding in both animal and human studies (Senkowski et al., <xref rid="B58" ref-type="bibr">2007</xref>; Ghazanfar et al., <xref rid="B27" ref-type="bibr">2008</xref>; Chandrasekaran and Ghazanfar, <xref rid="B18" ref-type="bibr">2009</xref>; Chandrasekaran et al., <xref rid="B19" ref-type="bibr">2011</xref>). For example, Ghazanfar et al. (<xref rid="B27" ref-type="bibr">2008</xref>) identified increased gamma-band coherence between auditory cortex and the superior temporal sulcus in response to multisensory relative to unisensory stimuli in Rhesus monkeys. Furthermore, using local field potential recordings in the superior temporal sulcus of macaques, Chandrasekaran and Ghazanfar (<xref rid="B18" ref-type="bibr">2009</xref>) determined that gamma oscillations demonstrated robust multisensory facilitation whereas multisensory facilitation in other frequency bands was dependent on the timing between auditory and visual stimuli. Finally, both humans and non-human primates demonstrate the same multisensory behavioral facilitation in parallel tasks (Chandrasekaran et al., <xref rid="B19" ref-type="bibr">2011</xref>) implying similar mechanisms are employed across species. The effort to better understand the role of oscillations in multisensory processing in the human brain has been expanded further by the application of functional neuroimaging techniques (for a review, see Stein and Stanford, <xref rid="B66" ref-type="bibr">2008</xref>). Both Kaiser et al. (<xref rid="B36" ref-type="bibr">2005</xref>) and Yuval-Greenberg and Deouell (<xref rid="B93" ref-type="bibr">2007</xref>) identified greater gamma-band power in response to congruent vs. incongruent multisensory stimuli. Furthermore, Mishra et al. (<xref rid="B47" ref-type="bibr">2007</xref>) determined that perception of the AV flash illusion (Shams et al., <xref rid="B60" ref-type="bibr">2002</xref>) was accompanied by bursts of gamma oscillations.</p><p>Schizophrenia is accompanied by both sensory and cognitive deficits in conjunction with the core symptoms associated with the disorder (Adcock et al., <xref rid="B1" ref-type="bibr">2009</xref>; Javitt, <xref rid="B34" ref-type="bibr">2009</xref>). Recent studies have indicated that cognitive factors are more directly related to quality of life than symptom alleviation (Green et al., <xref rid="B29" ref-type="bibr">2004</xref>). Correlations between sensory deficits and cognitive functioning are found in multiple studies (Uhlhaas et al., <xref rid="B80" ref-type="bibr">2008</xref>; Bedwell et al., <xref rid="B8" ref-type="bibr">2011</xref>; Silverstein and Keane, <xref rid="B61" ref-type="bibr">2011</xref>) and impaired processing of sensory information likely contributes to cognitive dysfunction in schizophrenia. Multisensory integration provides a bridge between unisensory and cognitive processing by requiring activation of a broader cortical network without requiring explicit cognitive skills. Our previous study (Stone et al., <xref rid="B68" ref-type="bibr">2011</xref>) indicates that multisensory stimuli may benefit patients with schizophrenia relative to healthy controls with both behavioral and neurophysiological multisensory facilitation observed in patients with schizophrenia in a forced choice paradigm despite unisensory deficits in patients with schizophrenia. Although neurophysiological and behavioral multisensory facilitation were both observed in patients, these effects were not directly correlated. In contrast, Williams et al. (<xref rid="B90" ref-type="bibr">2010</xref>) determined that schizophrenia patients had less behavioral facilitation than controls using a multisensory detection task. These results differ from our previous study employing a forced choice paradigm. Furthermore, De Gelder et al. (<xref rid="B20" ref-type="bibr">2002</xref>) reported no difference in multisensory reaction time between schizophrenia patients and controls, providing consistency with our results. In sum, results are currently mixed in terms of how patients with schizophrenia process multisensory stimuli relative to controls with few multisensory neuroimaging studies focused on this population.</p><p>Changes in gamma-band power have been observed in schizophrenia both at rest and during task execution (for recent reviews, see Gandal et al., <xref rid="B26" ref-type="bibr">2012</xref>; Uhlhaas and Singer, <xref rid="B82" ref-type="bibr">2012</xref>). Many of the gamma-band differences reported in schizophrenia relative to controls have been elicited in response to auditory steady state stimuli showing reduced gamma-band activity and reduced hemispheric asymmetry (Kwon et al., <xref rid="B40" ref-type="bibr">1999</xref>; Hamm et al., <xref rid="B32" ref-type="bibr">2011</xref>; Tsuchimoto et al., <xref rid="B79" ref-type="bibr">2011</xref>). Because perceptual and cognitive deficits are core impairments in schizophrenia, individuals suffering from the disorder may be especially prone to disruptions in multisensory processing (De Jong et al., <xref rid="B21" ref-type="bibr">2010</xref>; Williams et al., <xref rid="B90" ref-type="bibr">2010</xref>; Stone et al., <xref rid="B68" ref-type="bibr">2011</xref>). Evidence suggests that reductions in auditory induced gamma-band power may reflect impaired sensory responsiveness due to increased resting-state gamma-band activity (Teale et al., <xref rid="B76" ref-type="bibr">2008</xref>; Wilson et al., <xref rid="B91" ref-type="bibr">2008</xref>; Spencer, <xref rid="B64" ref-type="bibr">2011</xref>); however, the relationship between multisensory processing and gamma-band oscillations in schizophrenia has yet to be reported. The goal of the current study is to investigate the link between event-related gamma-band oscillations and multisensory integration in patients with schizophrenia and healthy controls.</p><p>To address this question, we recruited schizophrenia patients (SP) and healthy controls (HC) to perform a multisensory integration task while brain activity was recorded using magnetoencephalography (MEG). A simple multisensory paradigm was employed by presenting both unisensory (auditory-A and visual-V) stimuli as well as multisensory (AV) stimuli within a forced choice reaction time task. Based on the previously reported deficits in gamma-band oscillations, we hypothesized that gamma-band power would be reduced in SP relative to HC. The task also required that the participants identify the location of the stimulus within a perspective drawing, thereby presenting participants with near and far stimuli which corresponded to both peripherally- and centrally-presented visual stimuli, respectively. Peripherally-presented visual stimuli preferentially activate the dorsal visual stream (Ungerleider and Desimone, <xref rid="B83" ref-type="bibr">1986</xref>; Livingstone and Hubel, <xref rid="B42" ref-type="bibr">1987</xref>; Stephen et al., <xref rid="B67" ref-type="bibr">2002</xref>) which is impaired in schizophrenia (Butler and Javitt, <xref rid="B14" ref-type="bibr">2005</xref>; Koychev et al., <xref rid="B39" ref-type="bibr">2011</xref>). Therefore, we hypothesized that patients would show greater deficits in gamma-band power in response to peripheral (dorsal stream) visual stimuli relative to central (ventral stream) visual stimuli. Furthermore, multisensory studies have demonstrated that patients with schizophrenia have a wider window of integration than healthy controls when the stimuli are offset in time (Foucher et al., <xref rid="B24" ref-type="bibr">2007</xref>), suggesting that HC can differentiate asynchronous stimuli with smaller delays between auditory and visual stimuli than SP. In summary, we hypothesized that SP and HC would show differential patterns of event-related gamma-band activity and these differences would vary by condition (unisensory/multisensory, near/far, and synchronous/asynchronous).</p></sec><sec sec-type="materials|methods" id="s2"><title>Materials and methods</title><sec><title>Participants</title><p>Participants in the study included 103 individuals (46 SP and 57 age-matched HC). All participants provided written informed consent prior to study procedures. All procedures were performed in accordance with the Declaration of Helsinki and with prior approval from the University of New Mexico Health Sciences Center Human Research Review Committee. All SP met DSM-IV criteria for a diagnosis of schizophrenia or schizoaffective disorder, were stable on their medications for at least 1 month prior to study participation, and were periodically assessed throughout study enrollment to confirm clinical and pharmacological stability. HC and their first-degree relatives possessed no prior history of any psychiatric disorder based on the SCID-NP. None of the participants suffered from substance abuse, prior head trauma, or other neurological disorders, based on a standard neurological exam. Participant demographics are provided in Table <xref ref-type="table" rid="T1">1</xref>. As part of recruitment into the study, participants' neurocognitive abilities were assessed using the MATRICS test battery. This battery is designed to measure neurocognitive impairments, which present as core deficits in schizophrenia (Kern et al., <xref rid="B38" ref-type="bibr">2008</xref>; Nuechterlein et al., <xref rid="B52" ref-type="bibr">2008</xref>).</p><table-wrap id="T1" position="float"><label>Table 1</label><caption><p><bold>Participant demographics</bold>.</p></caption><table frame="hsides" rules="groups"><thead><tr><th rowspan="1" colspan="1"/><th align="center" rowspan="1" colspan="1"><bold>HC</bold></th><th align="center" rowspan="1" colspan="1"><bold>SP</bold></th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">Gender</td><td align="center" rowspan="1" colspan="1">40 males/17 females</td><td align="center" rowspan="1" colspan="1">39 males/7 females</td></tr><tr><td align="left" rowspan="1" colspan="1">Age</td><td align="center" rowspan="1" colspan="1">39.4 (12.7) years</td><td align="center" rowspan="1" colspan="1">39.2 (13.9) years</td></tr><tr><td align="left" rowspan="1" colspan="1">IQ</td><td align="center" rowspan="1" colspan="1">111.9<xref ref-type="table-fn" rid="TN1">^*</xref> (11.3)</td><td align="center" rowspan="1" colspan="1">101.5<xref ref-type="table-fn" rid="TN2">^**</xref> (17.1)</td></tr><tr><td align="left" rowspan="1" colspan="1">Medication</td><td align="center" rowspan="1" colspan="1">-</td><td align="center" rowspan="1" colspan="1">14.5<xref ref-type="table-fn" rid="TN2">^**</xref> (7.5) mg/day</td></tr><tr><td align="left" rowspan="1" colspan="1">Positive symptoms</td><td align="center" rowspan="1" colspan="1">-</td><td align="center" rowspan="1" colspan="1">15.5 (4.9)</td></tr><tr><td align="left" rowspan="1" colspan="1">Negative symptoms</td><td align="center" rowspan="1" colspan="1">-</td><td align="center" rowspan="1" colspan="1">14.6 (5.2)</td></tr></tbody></table><table-wrap-foot><p><italic>IQ based on WASI Neuropsychological Test; Medication is based on olanzapine equivalent dosage; Positive and Negative symptoms are cumulative scores from PANSS symptom scale; Values in parentheses represent standard deviations</italic>.</p><fn id="TN1"><label>*</label><p><italic>Mean based on 56 controls</italic>.</p></fn><fn id="TN2"><label>**</label><p><italic>Means based on 44 patients</italic>.</p></fn></table-wrap-foot></table-wrap></sec><sec><title>Behavioral task</title><p>To assess responses to unisensory and multisensory stimuli, participants performed a simple stimulus discrimination task. During the task, participants were presented with ecologically relevant audio-visual stimuli designed to mimic the image of a soccer ball and the sound of a soccer ball &#x0201c;bounce.&#x0201d; As control conditions, visual-only (the soccer ball image) and auditory-only (the soccer ball bounce sound) stimuli were also presented. During the presentation of all stimuli, participants viewed a static background on a projection screen positioned at a distance of 1 m from nasion. A simplified soccer field with a goalie and net provided a perspective-drawing framework, and participants were asked to fixate upon the goalie during all stimulus presentations (Figure <xref ref-type="fig" rid="F1">1A</xref>). There were two visual-only, two auditory-only, and four audio-visual stimulus presentations yielding eight distinct stimulus conditions:</p><fig id="F1" position="float"><label>Figure 1</label><caption><p><bold>Task stimuli</bold>. The background presented in <bold>(A)</bold> was present throughout the entire task performance. The participants were instructed to fixate on the goalie during the task. Auditory stimuli were presented with the visual background in place. During trials with visual stimuli the only change to the visual scene was the addition of a soccer ball in one of two locations Near <bold>(B)</bold> or Far <bold>(C)</bold>. During multisensory conditions the visual stimulus was presented in combination with the auditory tone.</p></caption><graphic xlink:href="fnhum-08-00788-g0001"/></fig><p><italic>The visual-only near stimulus condition (VN):</italic> During the VN condition, the image of a black and white soccer ball appeared on the static soccer field background for 200 ms. The ball was centered at 8&#x000b0; below fixation and subtended 2.7&#x000b0; of visual angle, giving the appearance that the ball was &#x0201c;downfield&#x0201d; and near to the participant (Figure <xref ref-type="fig" rid="F1">1B</xref>). This condition activates peripheral visual field associated with dorsal stream processing.</p><p><italic>The visual-only far stimulus condition (VF):</italic> During the VF condition, the soccer ball image appeared on the soccer field at 1.8&#x000b0; below fixation and subtended 1&#x000b0; of visual angle, giving the impression that the ball was &#x0201c;upfield&#x0201d; and farther away from the participant. The ball remained on the screen for 200 ms (Figure <xref ref-type="fig" rid="F1">1C</xref>). The VF and VN stimuli were scaled according to Rovamo and Virsu (<xref rid="B56" ref-type="bibr">1979</xref>) to account for the cortical magnification factor. This condition activates central visual field associated with ventral stream processing.</p><p><italic>The auditory-only near stimulus condition (AN):</italic> During the AN condition, a 550 Hz tone was presented binaurally for 200 ms through a set of ear plugs at a volume of 63 dB above hearing threshold. Threshold was determined uniquely for each participant prior to task performance. No other visual stimuli were presented during the AN condition, other than the static soccer field background (Figure <xref ref-type="fig" rid="F1">1A</xref>).</p><p><italic>The auditory-only far stimulus condition (AF):</italic> The AF condition was identical to the AN condition except that the tone was presented at a lower volume of 45 dB above hearing threshold in order to mimic a more distant sound.</p><p><italic>The audio-visual near stimulus condition (AVSN):</italic> The AVSN condition consisted of the presentation of the VN stimulus followed by the AN stimulus with a 5 ms delay.</p><p><italic>The audio-visual far stimulus condition (AVSF):</italic> During the AVSF condition, the VF stimulus was presented and followed by the AF stimulus after a 5 ms delay.</p><p><italic>The audio-visual asynchronous near stimulus condition (AVAN) and The audio-visual asynchronous far stimulus condition (AVAF):</italic> The final two audio-visual conditions were identical to the AVN and AVF conditions except that the delay between the visual and auditory stimuli was increased from 5 to 50 ms to emulate the natural delay between the two stimulus types which occur at a distance from the participant due to the differential speed of light and sound.</p><p>During each trial, one of the eight stimulus conditions was randomly presented. Participants were asked to indicate whether the stimulus was near to them or farther away by pressing one of two buttons on a response device with either their right index finger or right middle finger, respectively. During 20% of the trials, feedback regarding response accuracy was given. If the response was correct on an audio-visual trial, the image of the ball rolled into the soccer goal along with the sound of a cheering crowd. If an incorrect response or no response was given the ball rolled away from the goal accompanied by a &#x0201c;groaning crowd&#x0201d; sound. Unisensory trial feedback was matched to the sensory modality (i.e., only the rolling ball for visual conditions, or the crowd reaction for auditory conditions). Trial duration was between 1500 and 1900 ms and varied randomly from trial to trial, rendering an inter-stimulus interval (ISI) between 1300 and 1700 ms. Trials were organized into six blocks of 200 trials, separated by short breaks, where each condition was randomly presented with equal probability such that there were approximately 150 trials of each condition. Participants received pre-recorded instructions and were given a brief practice run to ensure that they understood the task prior to MEG data collection. The entire task, including auditory threshold determination, instructions, the practice run and the six trial blocks, was performed in a magnetically shielded room while participants were seated in the MEG chair. MEG and behavioral data used in the analyses were collected during the six trial blocks.</p></sec><sec><title>MEG data collection and processing</title><p>MEG data were collected with a whole head, 306-channel Elekta Neuromag system located at the Mind Research Network in Albuquerque, NM and data were acquired at a sampling rate of 1000 Hz with a 0.1 Hz high-pass and 330 Hz low-pass anti-aliasing filter. To permit comparisons between participants and groups, MEG sensor data for each participant were interpolated to the same reference head position using Neuromag Maxfilter software (Taulu et al., <xref rid="B74" ref-type="bibr">2004</xref>; Taulu and Kajola, <xref rid="B73" ref-type="bibr">2005</xref>) during post-processing. The reference head position was chosen based on the average head location across participants within the study. Maxfilter software also eliminates noise from sources that originate from outside of the defined head volume including muscle artifact and non-physiological flux jumps identified by an automated algorithm (Taulu et al., <xref rid="B74" ref-type="bibr">2004</xref>; Taulu and Simola, <xref rid="B75" ref-type="bibr">2006</xref>). Eyeblinks were eliminated from the data using a projector based on the average eye blink for each subject (Uusitalo and Ilmoniemi, <xref rid="B84" ref-type="bibr">1997</xref>). MEG data for each participant were epoched for each condition over an interval from 500 ms preceding stimulus onset until 500 ms after onset. Trials with incorrect responses and trials where the magnetic field exceeded 7 pT in any MEG gradiometer were also rejected. SP had significantly more trials rejected for incorrect responses, so HC trials were culled by randomly removing trials until both groups had an equal number of trials for each condition. Results are based upon an average of 142 &#x000b1; 10 trials/condition for HC and 144 &#x000b1; 8 trials/condition for SP. Preprocessing was performed using the scriptable MNE preprocessing pipeline (<ext-link ext-link-type="uri" xlink:href="http://martinos.org/mne">martinos.org/mne</ext-link>).</p></sec><sec><title>Event-related oscillation analysis</title><p>Linear trends, which spanned the 1000 ms time-window, were removed prior to time-frequency transformation. Each 1000 ms trial was then converted to the time-frequency domain using Morlet wavelets applied to each MEG gradiometer (width = 7 cycles, frequency range = 7&#x02013;50 Hz). Baseline-corrected spectral maps were computed by frequency for each trial in decibels (dB), with average spectral power from &#x02212;100 to 0 ms relative to stimulus onset as the measure of baseline noise. Spectral power at each MEG gradiometer, time point, and frequency was then averaged across trials for each condition and participant. Based on the spatial specificity of planar gradiometers (Ahonen et al., <xref rid="B2" ref-type="bibr">1993</xref>), we only analyzed the planar gradiometer data (magnetometer data is not reported here). To facilitate processing and reduce the number of comparisons, spectral power from the two planar gradiometers that occupied the same MEG sensor location was summed for all sensor pairs using the Fieldtrip function &#x0201c;ft_combineplanar.&#x0201d; Thus, there were 102 combined gradiometers (henceforth referred to simply as sensors) used for analysis. All time-frequency analyses were performed using Fieldtrip (Oostenveld et al., <xref rid="B53" ref-type="bibr">2011</xref>) and custom MATLAB programs (MathWorks, Inc., Natick, MA, USA).</p></sec><sec><title>Statistical analysis</title><sec><title>Behavioral data</title><p>To assess significant differences in accuracy and reaction time (RT), three-way multivariate analysis-of-variance tests (MANOVAs) were performed on the reaction times and accuracy (% correct) for the unisensory and multisensory stimuli. In these tests, stimulus type (auditory-only, visual-only, or audio-visual) and stimulus location (near or far) were treated as within-subject factors, while group identification (SP or HC) was treated as a between-subjects factor. Reaction times were only assessed for correct responses. Separate three-way MANOVAs were performed comparing reaction times and accuracy in which synchrony (synchronous or asynchronous) and stimulus location were treated as within-subjects factors and group identification was treated as a between-subjects factor. When significant main or interaction effects were detected, Bonferroni-corrected <italic>t</italic>-tests were performed to more closely examine the nature of these effects.</p></sec><sec><title>Event-related oscillation data</title><p>Event-related oscillations were determined by identifying significant increases or decreases in baseline-corrected spectral power within subject group in the 0&#x02013;480 ms time window. Significance was determined by one-sample <italic>t</italic>-tests using FDR correction with <italic>q</italic> = 0.05. Group differences in spectral power were confirmed to overlap with the time/frequency windows of significant increases or decreases in power relative to baseline in at least one group, and those which did not overlap were excluded from further analysis. No group differences were found which did not overlap with event-related oscillations showing either significant increases or decreases in power relative to baseline.</p><p>Spectral power was compared between HC and SP across the 30&#x02013;50 Hz (gamma-band) frequency range from 0 to 480 ms post-stimulus for each condition. We limited our time window to 0&#x02013;480 ms to focus on event-related oscillations following stimulus presentation. We focused our analysis on the 30&#x02013;50 Hz gamma-band range based on unisensory and multisensory studies reporting results in this frequency range. Independent sample <italic>t</italic>-tests comparing SP to HC were applied at each sensor, time, and frequency point. The results of these <italic>t</italic>-tests were then used to identify candidate clusters of significant group differences in gamma-band power. Candidate clusters were identified when at least 3 adjacent channels, 3 adjacent frequency points, and 20 adjacent time points were significantly different at the <italic>p</italic> &#x0003c; 0.05 threshold. This criterion alone limits spurious results as described by Guthrie and Buchwald (<xref rid="B30" ref-type="bibr">1991</xref>) when data are highly correlated. Following the recommendation of Maris and Oostenveld (<xref rid="B44" ref-type="bibr">2007</xref>), further testing to limit Type I errors associated with multiple comparisons was performed through permutation testing (repeated analyses with random shuffling of SP and HC) of each candidate sensor-time-frequency cluster to determine if the candidate group differences were statistically unlikely to have been observed by chance (<italic>p</italic> &#x0003c; 0.05). For these tests, participants' data were randomly re-assigned group identification (shuffled) while keeping the number of SP and HC constant. <italic>T</italic>-tests were applied to the data points in the identified clusters using these new group identifications. Group permutations were performed 5253 times (the number of unique permutations of 103 binary numbers) for each cluster and <italic>t</italic>-statistics were calculated and summed for each cluster for each permutation. When this summed <italic>t</italic>-statistic after permutation exceeded the summed <italic>t</italic>-statistic from the candidate cluster in more than 5% of the permutations, the cluster was rejected as not significantly different by group and was excluded from further analysis.</p></sec><sec><title>Regression and correlation analyses</title><p>Stepwise multiple regressions were performed separately for HC and SP to assess the extent to which mean spectral power of unisensory clusters predicted mean spectral power of multisensory clusters while limiting comparisons to the same stimulus location (near or far). Regression analyses were only performed for the clusters, which differed significantly from baseline values within group. Additionally, regressions were performed by group to determine if gamma-band power in each of the clusters predicted multisensory reaction time facilitation and MATRICS composite <italic>t</italic>-scores for subtests of interest (processing speed, attention, verbal learning, visual learning, and working memory). Regression models were considered significant at <italic>p</italic> &#x0003c; 0.025 (0.05 divided by the number of multisensory clusters) to correct for multiple comparisons.</p><p>Finally, Pearson correlation was used to assess the relationships among mean spectral power of unisensory clusters, medication dosage (in Olanzapine equivalent dose), and positive/negative symptoms (as assessed by the Positive and Negative Symptoms [PANS] scale).</p></sec></sec></sec><sec sec-type="results" id="s3"><title>Results</title><sec><title>Behavioral results</title><sec><title>Reaction times</title><p>Table <xref ref-type="table" rid="T2">2</xref> and Figure <xref ref-type="fig" rid="F2">2</xref> display group mean reaction times and accuracy results for each condition. A main effect of stimulus type was detected [<italic>F</italic>_{(2, 99)} = 410.26, <italic>p</italic> &#x0003c; 0.001] in which reaction times to visual-only and audio-visual stimuli were significantly faster when compared to auditory-only stimuli [<italic>t</italic>₍₁₀₁₎ = 17.06; <italic>t</italic>₍₁₀₁₎ = 24.27, respectively; <italic>p</italic> &#x0003c; 0.001, both tests]. Audio-visual stimuli also evoked faster reaction times than visual-only stimuli [<italic>t</italic>₍₁₀₁₎ = 9.06; <italic>p</italic> &#x0003c; 0.001]. By comparing the AV RTs to the fastest unisensory RTs (V for both HC and SP), this result indicates significant multisensory facilitation based on average RT. A significant stimulus type by group interaction was also detected [<italic>F</italic>_{(2, 99)} = 5.25; <italic>p</italic> = 0.007], in which differences between visual-only and audio-visual responses were significantly greater for SP than HC [<italic>t</italic>₍₁₀₀₎ = 3.24; <italic>p</italic> = 0.002; Figure <xref ref-type="fig" rid="F2">2A</xref>]. This result provides evidence of differential multisensory facilitation by group with SP showing greater improvement in AV RTs relative to V RTs.</p><table-wrap id="T2" position="float"><label>Table 2</label><caption><p><bold>Behavioral means</bold>.</p></caption><table frame="hsides" rules="groups"><thead><tr><th rowspan="1" colspan="1"/><th align="center" rowspan="1" colspan="1"><bold>HC</bold></th><th align="center" rowspan="1" colspan="1"><bold>SP<xref ref-type="table-fn" rid="TN3">^*</xref></bold></th><th align="center" rowspan="1" colspan="1"><bold>HC</bold></th><th align="center" rowspan="1" colspan="1"><bold>SP<xref ref-type="table-fn" rid="TN3">^*</xref></bold></th></tr><tr><th rowspan="1" colspan="1"/><th align="center" colspan="2" rowspan="1"><bold>Reaction time</bold></th><th align="center" colspan="2" rowspan="1"><bold>Accuracy (% Correct)</bold></th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">VN</td><td align="center" rowspan="1" colspan="1">428.6 (9.0)</td><td align="center" rowspan="1" colspan="1">439.7 (17.6)</td><td align="center" rowspan="1" colspan="1">95.1 (0.7)</td><td align="center" rowspan="1" colspan="1">88.4 (1.5)</td></tr><tr><td align="left" rowspan="1" colspan="1">VF</td><td align="center" rowspan="1" colspan="1">439.3 (8.6)</td><td align="center" rowspan="1" colspan="1">454.9 (17.8)</td><td align="center" rowspan="1" colspan="1">94.2 (0.8)</td><td align="center" rowspan="1" colspan="1">85.5 (1.8)</td></tr><tr><td align="left" rowspan="1" colspan="1">AN</td><td align="center" rowspan="1" colspan="1">579.3 (12.7)</td><td align="center" rowspan="1" colspan="1">565.1 (21.0)</td><td align="center" rowspan="1" colspan="1">90.4 (0.8)</td><td align="center" rowspan="1" colspan="1">78.0 (1.9)</td></tr><tr><td align="left" rowspan="1" colspan="1">AF</td><td align="center" rowspan="1" colspan="1">578.7 (11.0)</td><td align="center" rowspan="1" colspan="1">569.7 (19.6)</td><td align="center" rowspan="1" colspan="1">92.4 (0.7)</td><td align="center" rowspan="1" colspan="1">78.8 (2.4)</td></tr><tr><td align="left" rowspan="1" colspan="1">AVSN</td><td align="center" rowspan="1" colspan="1">406.8 (8.6)</td><td align="center" rowspan="1" colspan="1">399.2 (14.8)</td><td align="center" rowspan="1" colspan="1">97.2 (0.4)</td><td align="center" rowspan="1" colspan="1">90.5 (1.2)</td></tr><tr><td align="left" rowspan="1" colspan="1">AVSF</td><td align="center" rowspan="1" colspan="1">423.1 (8.4)</td><td align="center" rowspan="1" colspan="1">419.6 (15.4)</td><td align="center" rowspan="1" colspan="1">96.2 (0.5)</td><td align="center" rowspan="1" colspan="1">87.2 (1.8)</td></tr><tr><td align="left" rowspan="1" colspan="1">AVAN</td><td align="center" rowspan="1" colspan="1">425.2 (9.1)</td><td align="center" rowspan="1" colspan="1">415.0 (15.0)</td><td align="center" rowspan="1" colspan="1">97.2 (0.4)</td><td align="center" rowspan="1" colspan="1">91.5 (1.2)</td></tr><tr><td align="left" rowspan="1" colspan="1">AVAF</td><td align="center" rowspan="1" colspan="1">442.2 (8.6)</td><td align="center" rowspan="1" colspan="1">437.9 (16.1)</td><td align="center" rowspan="1" colspan="1">95.8 (0.5)</td><td align="center" rowspan="1" colspan="1">87.7 (1.6)</td></tr></tbody></table><table-wrap-foot><p><italic>Means (s.e.m.) of reaction times and accuracy</italic>.</p><fn id="TN3"><label>*</label><p><italic>Behavioral data missing for one patient. Excluded from table and all behavioral analyses</italic>.</p></fn></table-wrap-foot></table-wrap><fig id="F2" position="float"><label>Figure 2</label><caption><p><bold>Behavioral differences. (A)</bold> Mean reaction times for the SP and HC group to unisensory (A and V) and multisensory (AV) stimuli. Significant main effects and interactions are denoted by asterisks. In addition to a main effect of condition, there was a group by condition interaction showing that SP had greater multisensory facilitation than HC&#x02014;the fastest unisensory response (V) was significantly slower in SP, yet the AV RTs were equivalent by group. <bold>(B)</bold> Accuracy (% correct responses) for each group in response to unisensory (A and V) and multisensory (AV) stimuli. HC had significantly more correct responses to all three stimulus types compared to SP. <bold>(C)</bold> Accuracy for each group in response to near and far stimuli collapsed across conditions. HC had significantly more correct responses to both near and far stimuli compared to SP. SP had a significantly greater difference in near and far response accuracy compared to HC. Error bars represent s.e.m.</p></caption><graphic xlink:href="fnhum-08-00788-g0002"/></fig><p>A significant effect of stimulus location (near vs. far) was also detected [<italic>F</italic>_{(1, 100)} = 19.51, <italic>p</italic> &#x0003c; 0.001], where RTs to near stimuli were significantly faster. Additionally, there was a significant stimulus type by stimulus location interaction [<italic>F</italic>_{(2, 99)} = 8.25; <italic>p</italic> &#x0003c; 0.001], in which RTs to the VN condition were significantly faster than RTs to the VF condition [<italic>t</italic>₍₁₀₁₎ = 4.28; <italic>p</italic> &#x0003c; 0.001]. Likewise AVN RTs were faster than AVF RTs [<italic>t</italic>₍₁₀₁₎ = 6.48; <italic>p</italic> &#x0003c; 0.001].</p><p>A significant main effect of audio-visual synchrony was detected when comparing AVN/AVF RTs to AVAN/AVAF RTs. Responses to synchronous presentations were faster than asynchronous presentations [<italic>F</italic>_{(1, 100)} = 160.75; <italic>p</italic> &#x0003c; 0.001]. No significant group effects were detected based on the synchrony of the AV stimuli (all <italic>p</italic>'s &#x0003e; 0.05).</p></sec><sec><title>Response accuracy</title><p>The accuracy comparisons yielded similar results. There was a significant main effect of stimulus type [<italic>F</italic>_{(2, 99)} = 122.42; <italic>p</italic> &#x0003c; 0.001], where more correct responses occurred in the visual-only and audio-visual conditions compared to the auditory-only condition [<italic>t</italic>₍₁₀₁₎ = 7.50; <italic>t</italic>₍₁₀₁₎ = 13.37, respectively; <italic>p</italic> &#x0003c; 0.001, both cases]. There were also more correct responses to audio-visual presentations when directly compared to visual-only presentations [<italic>t</italic>₍₁₀₁₎ = 3.96; <italic>p</italic> &#x0003c; 0.001]. A significant stimulus type by group interaction was found [<italic>F</italic>_{(2, 99)} = 12.17; <italic>p</italic> &#x0003c; 0.001] such that HC had significantly more correct responses to all stimulus types compared to SP [visual-only <italic>t</italic>₍₁₀₁₎ = 29.62; auditory-only <italic>t</italic>₍₁₀₁₎ = 28.78; audio-visual <italic>t</italic>₍₁₀₁₎ = 41.15; <italic>p</italic> &#x0003c; 0.001, all cases; Figure <xref ref-type="fig" rid="F2">2B</xref>]. This interaction provides evidence that both groups showed improved accuracy for the AV condition relative to the unisensory conditions.</p><p>There was also a significant main effect of stimulus location [<italic>F</italic>_{(1, 100)} = 4.87; <italic>p</italic> = 0.03], where more correct responses were made for near stimuli. A significant stimulus location by group interaction was also detected [<italic>F</italic>_{(2, 99)} = 4.93; <italic>p</italic> = 0.03], in which HC had significantly more correct responses to both near stimuli [<italic>t</italic>₍₁₀₀₎ = 6.20; <italic>p</italic> &#x0003c; 0.001] and far stimuli [<italic>t</italic>₍₁₀₀₎ = 5.81; <italic>p</italic> &#x0003c; 0.001; Figure <xref ref-type="fig" rid="F2">2C</xref>]. Finally, a significant stimulus type by stimulus location interaction was detected [<italic>F</italic>_{(2, 99)} = 14.49; <italic>p</italic> &#x0003c; 0.001], where more correct responses were found for VN than VF stimuli and AVN than AVF stimuli [<italic>t</italic>₍₁₀₁₎ = 4.48; <italic>t</italic>₍₁₀₁₎ = 4.58, respectively; <italic>p</italic> &#x0003c; 0.001, both cases]. This provides evidence that the peripheral visual stimuli and higher volume auditory stimuli lead to improved accuracy in this task and the combined AV condition improved accuracy, especially for the near condition.</p></sec></sec><sec><title>Event-related gamma-band oscillation results</title><p>Permutation testing of the candidate channel-time-frequency clusters revealed regions of significant group differences in gamma-band power, which are summarized in Table <xref ref-type="table" rid="T3">3</xref> and Figures <xref ref-type="fig" rid="F3">3</xref> and <xref ref-type="fig" rid="F4">4</xref>. Figure <xref ref-type="fig" rid="F3">3</xref> depicts regions and time intervals where group differences in gamma-band power were found while Figure <xref ref-type="fig" rid="F4">4</xref> shows the time-frequency power maps for the representative sensor where group differences were identified (denoted by the white asterisk in Figure <xref ref-type="fig" rid="F3">3</xref>). Significant group differences were found only in near unisensory and synchronous multisensory conditions. All clusters demonstrating significant group differences were confirmed to show significant increases or decreases in gamma-band power relative to the prestimulus time period in the FDR-corrected one-sample <italic>t</italic>-test comparisons within one or both diagnostic groups. This provides evidence of task-related increases or decreases in gamma-band power in either one or both groups in each of the group difference clusters (see asterisks in Table <xref ref-type="table" rid="T3">3</xref>). Examples of the full time-frequency maps (7&#x02013;50 Hz, &#x02212;100&#x02013;500 ms) are presented in Figure <xref ref-type="fig" rid="F5">5</xref> with the analysis region (30&#x02013;50 Hz, 0&#x02013;480 ms) for group comparisons highlighted with a black box. A clear event-related response is present in both the alpha- (~10 Hz) and gamma-band (30&#x02013;50 Hz) frequency ranges.</p><table-wrap id="T3" position="float"><label>Table 3</label><caption><p><bold>Clusters of significant group differences in gamma-band Power</bold>.</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1"><bold>Cluster label<xref ref-type="table-fn" rid="TN4">^&#x02020;</xref></bold></th><th align="center" rowspan="1" colspan="1"><bold>MEG channel</bold></th><th align="center" rowspan="1" colspan="1"><bold>Latency (ms)</bold></th><th align="center" rowspan="1" colspan="1"><bold>Frequency (Hz)</bold></th><th align="center" rowspan="1" colspan="1"><bold>Group difference</bold></th><th align="center" rowspan="1" colspan="1"><bold>Cluster <italic>t</italic>-value</bold></th><th align="center" rowspan="1" colspan="1"><bold>Cluster <italic>p</italic>-value</bold></th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">VN-RC</td><td align="center" rowspan="1" colspan="1">1041</td><td align="center" rowspan="1" colspan="1">96&#x02013;280</td><td align="center" rowspan="1" colspan="1">37&#x02013;50</td><td align="center" rowspan="1" colspan="1">SP<xref ref-type="table-fn" rid="TN5">^*</xref> &#x0003c; HC</td><td align="center" rowspan="1" colspan="1">&#x02212;2.71</td><td align="center" rowspan="1" colspan="1">0.019</td></tr><tr><td rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1111</td><td align="center" rowspan="1" colspan="1">98&#x02013;252</td><td align="center" rowspan="1" colspan="1">36&#x02013;50</td><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/></tr><tr><td rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1121</td><td align="center" rowspan="1" colspan="1">0&#x02013;247</td><td align="center" rowspan="1" colspan="1">33&#x02013;50</td><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">AN-RT1</td><td align="center" rowspan="1" colspan="1">1441</td><td align="center" rowspan="1" colspan="1">0&#x02013;274</td><td align="center" rowspan="1" colspan="1">30&#x02013;36</td><td align="center" rowspan="1" colspan="1">SP &#x0003e; HC<xref ref-type="table-fn" rid="TN5">^*</xref></td><td align="center" rowspan="1" colspan="1">2.61</td><td align="center" rowspan="1" colspan="1">0.020</td></tr><tr><td rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">2611</td><td align="center" rowspan="1" colspan="1">15&#x02013;174</td><td align="center" rowspan="1" colspan="1">30&#x02013;37</td><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/></tr><tr><td rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">2621</td><td align="center" rowspan="1" colspan="1">38&#x02013;147</td><td align="center" rowspan="1" colspan="1">30&#x02013;38</td><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">AN-RT2</td><td align="center" rowspan="1" colspan="1">1441</td><td align="center" rowspan="1" colspan="1">349&#x02013;480</td><td align="center" rowspan="1" colspan="1">30&#x02013;36</td><td align="center" rowspan="1" colspan="1">SP<xref ref-type="table-fn" rid="TN5">^*</xref> &#x0003e; HC<xref ref-type="table-fn" rid="TN5">^*</xref></td><td align="center" rowspan="1" colspan="1">3.03</td><td align="center" rowspan="1" colspan="1">0.032</td></tr><tr><td rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">2611</td><td align="center" rowspan="1" colspan="1">233&#x02013;480</td><td align="center" rowspan="1" colspan="1">30&#x02013;45</td><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/></tr><tr><td rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">2621</td><td align="center" rowspan="1" colspan="1">244&#x02013;480</td><td align="center" rowspan="1" colspan="1">30&#x02013;50</td><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/></tr><tr><td rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">2641</td><td align="center" rowspan="1" colspan="1">349&#x02013;480</td><td align="center" rowspan="1" colspan="1">30&#x02013;37</td><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">AN-RC</td><td align="center" rowspan="1" colspan="1">0731</td><td align="center" rowspan="1" colspan="1">248&#x02013;480</td><td align="center" rowspan="1" colspan="1">30&#x02013;37</td><td align="center" rowspan="1" colspan="1">SP &#x0003e; HC<xref ref-type="table-fn" rid="TN5">^*</xref></td><td align="center" rowspan="1" colspan="1">2.51</td><td align="center" rowspan="1" colspan="1">0.024</td></tr><tr><td rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">2211</td><td align="center" rowspan="1" colspan="1">338&#x02013;480</td><td align="center" rowspan="1" colspan="1">30&#x02013;50</td><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/></tr><tr><td rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">2241</td><td align="center" rowspan="1" colspan="1">321&#x02013;480</td><td align="center" rowspan="1" colspan="1">30&#x02013;37</td><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">AN-RF</td><td align="center" rowspan="1" colspan="1">0921</td><td align="center" rowspan="1" colspan="1">279&#x02013;480</td><td align="center" rowspan="1" colspan="1">30&#x02013;50</td><td align="center" rowspan="1" colspan="1">SP &#x0003e; HC<xref ref-type="table-fn" rid="TN5">^*</xref></td><td align="center" rowspan="1" colspan="1">2.98</td><td align="center" rowspan="1" colspan="1">0.022</td></tr><tr><td rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0931</td><td align="center" rowspan="1" colspan="1">312&#x02013;480</td><td align="center" rowspan="1" colspan="1">30&#x02013;41</td><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/></tr><tr><td rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0941</td><td align="center" rowspan="1" colspan="1">367&#x02013;480</td><td align="center" rowspan="1" colspan="1">30&#x02013;39</td><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/></tr><tr><td rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1231</td><td align="center" rowspan="1" colspan="1">367&#x02013;480</td><td align="center" rowspan="1" colspan="1">30&#x02013;37</td><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">AVSN-LF</td><td align="center" rowspan="1" colspan="1">0541</td><td align="center" rowspan="1" colspan="1">232&#x02013;362</td><td align="center" rowspan="1" colspan="1">34&#x02013;50</td><td align="center" rowspan="1" colspan="1">SP<xref ref-type="table-fn" rid="TN5">^*</xref> &#x0003e; HC<xref ref-type="table-fn" rid="TN5">^*</xref></td><td align="center" rowspan="1" colspan="1">2.60</td><td align="center" rowspan="1" colspan="1">0.031</td></tr><tr><td rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0611</td><td align="center" rowspan="1" colspan="1">199&#x02013;480</td><td align="center" rowspan="1" colspan="1">34&#x02013;50</td><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/></tr><tr><td rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">1011</td><td align="center" rowspan="1" colspan="1">222&#x02013;480</td><td align="center" rowspan="1" colspan="1">30&#x02013;46</td><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">AVSN-RF</td><td align="center" rowspan="1" colspan="1">0921</td><td align="center" rowspan="1" colspan="1">229&#x02013;355</td><td align="center" rowspan="1" colspan="1">35&#x02013;46</td><td align="center" rowspan="1" colspan="1">SP &#x0003e; HC<xref ref-type="table-fn" rid="TN5">^*</xref></td><td align="center" rowspan="1" colspan="1">2.41</td><td align="center" rowspan="1" colspan="1">0.031</td></tr><tr><td rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0931</td><td align="center" rowspan="1" colspan="1">242&#x02013;334</td><td align="center" rowspan="1" colspan="1">36&#x02013;43</td><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/></tr><tr><td rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">0941</td><td align="center" rowspan="1" colspan="1">242&#x02013;391</td><td align="center" rowspan="1" colspan="1">30&#x02013;46</td><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/></tr></tbody></table><table-wrap-foot><fn id="TN4"><label>&#x02020;</label><p><italic>Clusters are listed with the stimulus condition, followed by the sensor region where they were detected. LF, left frontal; RF, right frontal; RC, right central; RT, right temporal</italic>;</p></fn><fn id="TN5"><label>*</label><p><italic>indicates gamma power that significantly deviated from baseline using the 1-sample FDR-corrected test</italic>.</p></fn></table-wrap-foot></table-wrap><fig id="F3" position="float"><label>Figure 3</label><caption><p><bold>Regions of significant group differences</bold>. A top down view of the sensor array is presented to show regions of significant group differences in gamma-band power. The top of the plot denotes frontal regions whereas left and right correspond to left and right temporal regions with occipital regions located at the bottom of the plot. The colored ovals denote where group differences in gamma-band power relative to baseline were detected in the VN <bold>(A)</bold>, AN <bold>(B)</bold>, and AVSN <bold>(C)</bold> conditions. Black points within each colored oval mark sensor locations where time-frequency differences emerged. The ovals are color-coded to represent time (in ms) when significant differences occurred in each region. Two regions of significant group differences overlapped spatially; AN-RT1 is shown with red diagonal stripes overlaid on AN-RT2. The white asterisk denotes the time frequency plot that is displayed in Figure <xref ref-type="fig" rid="F4">4</xref>. LF, left frontal; RF, right frontal; RC, right central; RT, right temporal.</p></caption><graphic xlink:href="fnhum-08-00788-g0003"/></fig><fig id="F4" position="float"><label>Figure 4</label><caption><p><bold>Example time-frequency maps for each condition</bold>. Group-averaged baseline-corrected time-frequency power maps (scale in dB) for a representative sensor of 5/7 clusters with significant group differences (Table <xref ref-type="table" rid="T3">3</xref>) are displayed. Group comparisons are displayed by comparing the left (HC) and right (SP) columns. The black boxes denote the overlapping region of significant group differences for the location presented in Figure <xref ref-type="fig" rid="F3">3</xref> across the three regional channels. Dotted boxes denote regions which did not show significant within-group differences from baseline, while solid boxes denote those which do show significant within-group differences relative to baseline. The white outline denotes the region that showed significant group differences within the displayed channel. As expected the overlap across regional channels is smaller than the significant region for the displayed channel due to spatial variation of oscillatory activity. Three of the clusters of significant group differences indicate increased gamma-band power in SP. AN-RT1 (early) and AN-RT2 (late) are both displayed in the AN-RT plots. Group-averaged time frequency plots of gamma band power are displayed for region VN-RC <bold>(A,B)</bold>, AN-RT <bold>(C,D)</bold>, AN-RF <bold>(E,F)</bold>, and AVSN-RF <bold>(G,H)</bold>.</p></caption><graphic xlink:href="fnhum-08-00788-g0004"/></fig><fig id="F5" position="float"><label>Figure 5</label><caption><p><bold>Example time frequency maps showing task-related responses</bold>. The group averaged time frequency maps showed the expected increases in low frequency power demonstrating a clear task-related response to stimuli over the relevant brain regions. To display the broader task-related response we display the time frequency map from 7 to 50 Hz and from &#x02212;100 to 500 ms. For example, right temporal locations showed a response to the AN stimulus <bold>(A,B)</bold>&#x02014;see zoomed in time frequency plot in Figures <xref ref-type="fig" rid="F4">4C,D</xref>. Furthermore, AVSN-RF time frequency plots (<bold>C,D</bold>&#x02014;correspond to Figures <xref ref-type="fig" rid="F4">4G,H</xref>) located over right frontal region are shown. The black box denotes the time-frequency window that was analyzed in this study and displayed in Figure <xref ref-type="fig" rid="F4">4</xref>. Note the change in scale between this figure and Figure <xref ref-type="fig" rid="F4">4</xref> due to the larger changes in power generally observed at lower frequencies. Clear changes from baseline are observed by 100 ms poststimulus in response to the auditory/visual stimuli below 20 Hz.</p></caption><graphic xlink:href="fnhum-08-00788-g0005"/></fig><p>There were five clusters with significant group differences in gamma-band power in response to unisensory stimuli (see Figures <xref ref-type="fig" rid="F3">3A,B</xref>, <xref ref-type="fig" rid="F4">4A&#x02013;F</xref>), which were located over frontal, central and temporal brain regions. The time windows of these group differences in gamma-band power ranged from early (e.g., AN-RT1) to late (e.g., AN-RC) and spanned the 30&#x02013;50 Hz range (see Table <xref ref-type="table" rid="T3">3</xref>).</p><p>During multisensory stimulus presentation there were two clusters showing significant group differences in gamma-band power located over frontal cortex (Figures <xref ref-type="fig" rid="F3">3C</xref>, <xref ref-type="fig" rid="F4">4G,H</xref>). The timing of these differences in gamma-band power was restricted to the latter half of the analysis window (&#x0003e;225 ms poststimulus), and the location, time window and frequency range of one of these clusters (AVSN-RF) overlapped with unisensory differences in the auditory cluster AN-RF.</p></sec><sec><title>Unisensory and multisensory cluster regressions</title><p>Table <xref ref-type="table" rid="T4">4</xref> displays the significant results from multisensory cluster regressions for HC based on clusters that deviated significantly from baseline values. There was a clear difference in predictive power between SP and HC. In HC, gamma-band power in multisensory clusters was predicted by unisensory clusters restricted to the right central sensors, whereas none of the gamma band clusters that deviated significantly from zero for SP predicted multisensory gamma. Finally, gamma-band power did not predict RT facilitation for HC or SP.</p><table-wrap id="T4" position="float"><label>Table 4</label><caption><p><bold>Significant multisensory cluster regressions for HC</bold>.</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1"><bold>Regressands</bold></th><th align="center" rowspan="1" colspan="1"><bold>Regressors</bold></th><th align="center" rowspan="1" colspan="1"><bold>Beta (&#x003b2;)</bold></th><th align="center" rowspan="1" colspan="1"><bold>Partial correlation</bold></th><th align="center" rowspan="1" colspan="1"><bold><italic>R</italic>²</bold></th><th align="center" rowspan="1" colspan="1"><bold><italic>p</italic></bold></th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">AVSN-LF</td><td align="center" rowspan="1" colspan="1">AN-RC</td><td align="center" rowspan="1" colspan="1">0.28</td><td align="center" rowspan="1" colspan="1">0.43</td><td align="center" rowspan="1" colspan="1">0.26</td><td align="center" rowspan="1" colspan="1">&#x0003c; 0.001<xref ref-type="table-fn" rid="TN6">^*</xref></td></tr><tr><td align="left" rowspan="1" colspan="1">AVSN-RF</td><td align="center" rowspan="1" colspan="1">AN-RC</td><td align="center" rowspan="1" colspan="1">0.30</td><td align="center" rowspan="1" colspan="1">0.35</td><td align="center" rowspan="1" colspan="1">0.12</td><td align="center" rowspan="1" colspan="1">0.008<xref ref-type="table-fn" rid="TN6">^*</xref></td></tr><tr><td align="left" rowspan="1" colspan="1">RT-fac. (AVSN)</td><td align="center" rowspan="1" colspan="1">None</td><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/></tr></tbody></table><table-wrap-foot><fn id="TN6"><label>*</label><p><italic>These p-values were significant with correction for multiple comparisons (&#x003b1; = 0.05/2 = 0.025). RT-fac, Reaction Time multisensory facilitation relative to fastest unisensory RT</italic>.</p></fn></table-wrap-foot></table-wrap></sec><sec><title>Neurocognitive regressions</title><p>Table <xref ref-type="table" rid="T5">5</xref> displays group means and standard errors for the MATRICS subtests. As expected, SP performed significantly worse on all measures relative to HC.</p><table-wrap id="T5" position="float"><label>Table 5</label><caption><p><bold>MATRICS scores</bold>.</p></caption><table frame="hsides" rules="groups"><thead><tr><th rowspan="1" colspan="1"/><th align="center" colspan="2" rowspan="1"><bold>HC</bold></th><th align="center" colspan="2" rowspan="1"><bold>SP</bold></th></tr><tr><th rowspan="1" colspan="1"/><th align="center" rowspan="1" colspan="1"><bold>Mean (s.e.m.)</bold></th><th align="center" rowspan="1" colspan="1"><bold><italic>n</italic><xref ref-type="table-fn" rid="TN7">^*</xref></bold></th><th align="center" rowspan="1" colspan="1"><bold>Mean (s.e.m.)</bold></th><th align="center" rowspan="1" colspan="1"><bold><italic>n</italic><xref ref-type="table-fn" rid="TN7">^*</xref></bold></th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">Processing speed<xref ref-type="table-fn" rid="TN8">^**</xref></td><td align="center" rowspan="1" colspan="1">53.5 (1.2)</td><td align="center" rowspan="1" colspan="1">55</td><td align="center" rowspan="1" colspan="1">36.9 (1.9)</td><td align="center" rowspan="1" colspan="1">45</td></tr><tr><td align="left" rowspan="1" colspan="1">Attention/Vigilance<xref ref-type="table-fn" rid="TN8">^**</xref></td><td align="center" rowspan="1" colspan="1">49.0 (1.3)</td><td align="center" rowspan="1" colspan="1">51</td><td align="center" rowspan="1" colspan="1">37.6 (2.1)</td><td align="center" rowspan="1" colspan="1">45</td></tr><tr><td align="left" rowspan="1" colspan="1">Working memory<xref ref-type="table-fn" rid="TN8">^**</xref></td><td align="center" rowspan="1" colspan="1">50.1 (1.3)</td><td align="center" rowspan="1" colspan="1">55</td><td align="center" rowspan="1" colspan="1">42.2 (1.9)</td><td align="center" rowspan="1" colspan="1">45</td></tr><tr><td align="left" rowspan="1" colspan="1">Verbal learning<xref ref-type="table-fn" rid="TN8">^**</xref></td><td align="center" rowspan="1" colspan="1">45.9 (1.2)</td><td align="center" rowspan="1" colspan="1">55</td><td align="center" rowspan="1" colspan="1">39.0 (1.3)</td><td align="center" rowspan="1" colspan="1">45</td></tr><tr><td align="left" rowspan="1" colspan="1">Visual learning<xref ref-type="table-fn" rid="TN8">^**</xref></td><td align="center" rowspan="1" colspan="1">45.7 (1.4)</td><td align="center" rowspan="1" colspan="1">55</td><td align="center" rowspan="1" colspan="1">38.0 (1.7)</td><td align="center" rowspan="1" colspan="1">45</td></tr></tbody></table><table-wrap-foot><p><italic>SEM, Standard Error of the Mean</italic>.</p><fn id="TN7"><label>*</label><p><italic>Not all participants completed each MATRICS domain and their data for that domain were excluded from the table and regression analyses</italic>.</p></fn><fn id="TN8"><label>**</label><p><italic>Significant at p &#x0003c; 0.001</italic>.</p></fn></table-wrap-foot></table-wrap><p>Tables <xref ref-type="table" rid="T6">6</xref> and <xref ref-type="table" rid="T7">7</xref> display results of neurocognitive regressions for HC and SP, respectively. For HC, working memory was predicted by VN-RC, and verbal learning was predicted by AVSN-LF gamma band power, both with negative relationships. For SP, VN-RC predicted attention, working memory, and verbal learning, all with negative relationships. VN-RC was always the first to enter the model when it was predictive of MATRICS scores; however, AN-RT2 predicted working memory when accounting for variance explained by VN-RC. Within SP, medication dosage was positively correlated with AVSN-LF (<italic>r</italic> = 0.33, <italic>p</italic> = 0.031).</p><table-wrap id="T6" position="float"><label>Table 6</label><caption><p><bold>Significant neurocognitive regressions for HC</bold>.</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1"><bold>Regressands</bold></th><th align="center" rowspan="1" colspan="1"><bold>Regressors</bold></th><th align="center" rowspan="1" colspan="1"><bold>Beta (&#x003b2;)</bold></th><th align="center" rowspan="1" colspan="1"><bold>Partial correlation</bold></th><th align="center" rowspan="1" colspan="1"><bold><italic>R</italic>²</bold></th><th align="center" rowspan="1" colspan="1"><bold><italic>p</italic></bold></th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">MATRICS Processing speed</td><td align="center" rowspan="1" colspan="1">None</td><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">MATRICS Attention</td><td align="center" rowspan="1" colspan="1">None</td><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">MATRICS Working memory</td><td align="center" rowspan="1" colspan="1">VN-RC</td><td align="center" rowspan="1" colspan="1">&#x02212;7.1</td><td align="center" rowspan="1" colspan="1">&#x02212;0.27</td><td align="center" rowspan="1" colspan="1">0.71</td><td align="center" rowspan="1" colspan="1">0.049</td></tr><tr><td align="left" rowspan="1" colspan="1">MATRICS Verbal learning</td><td align="center" rowspan="1" colspan="1">AVSN-LF</td><td align="center" rowspan="1" colspan="1">&#x02212;9.9</td><td align="center" rowspan="1" colspan="1">&#x02212;0.27</td><td align="center" rowspan="1" colspan="1">0.74</td><td align="center" rowspan="1" colspan="1">0.044</td></tr><tr><td align="left" rowspan="1" colspan="1">MATRICS Visual learning</td><td align="center" rowspan="1" colspan="1">None</td><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/></tr></tbody></table></table-wrap><table-wrap id="T7" position="float"><label>Table 7</label><caption><p><bold>Significant neurocognitive regressions for SP</bold>.</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="center" rowspan="1" colspan="1"><bold>Regressands</bold></th><th align="center" rowspan="1" colspan="1"><bold>Regressors</bold></th><th align="center" rowspan="1" colspan="1"><bold>Beta (&#x003b2;)</bold></th><th align="center" rowspan="1" colspan="1"><bold>Partial correlation</bold></th><th align="center" rowspan="1" colspan="1"><bold><italic>R</italic>²</bold></th><th align="center" rowspan="1" colspan="1"><bold><italic>p</italic></bold></th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">MATRICS Processing speed</td><td align="center" rowspan="1" colspan="1">None</td><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">MATRICS Attention</td><td align="center" rowspan="1" colspan="1">VN-RC</td><td align="center" rowspan="1" colspan="1">&#x02212;17.8</td><td align="center" rowspan="1" colspan="1">&#x02212;0.39</td><td align="center" rowspan="1" colspan="1">0.26</td><td align="center" rowspan="1" colspan="1">0.002</td></tr><tr><td rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">AN-RT2</td><td align="center" rowspan="1" colspan="1">10.3</td><td align="center" rowspan="1" colspan="1">0.37</td><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">MATRICS Working memory</td><td align="center" rowspan="1" colspan="1">VN-RC</td><td align="center" rowspan="1" colspan="1">&#x02212;16.2</td><td align="center" rowspan="1" colspan="1">&#x02212;0.38</td><td align="center" rowspan="1" colspan="1">0.20</td><td align="center" rowspan="1" colspan="1">0.010</td></tr><tr><td align="left" rowspan="1" colspan="1">MATRICS Verbal learning</td><td align="center" rowspan="1" colspan="1">VN-RC</td><td align="center" rowspan="1" colspan="1">&#x02212;9.4</td><td align="center" rowspan="1" colspan="1">&#x02212;0.32</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="center" rowspan="1" colspan="1">0.032</td></tr><tr><td align="left" rowspan="1" colspan="1">MATRICS Visual learning</td><td align="center" rowspan="1" colspan="1">None</td><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/><td rowspan="1" colspan="1"/></tr></tbody></table></table-wrap></sec></sec><sec sec-type="discussion" id="s4"><title>Discussion</title><p>The aim of this study was to determine if differences in 30&#x02013;50 Hz gamma-band power in response to unisensory and multisensory stimuli in a large cohort of schizophrenia patients relative to age-matched healthy controls explained the multisensory facilitation observed in our previous study. Similar to our previous study, this larger cohort of SP showed greater behavioral facilitation to multisensory stimuli than HC. However, gamma-band power was not directly associated with RT facilitation. Despite this lack of association with RT, gamma-band power predicted performance on MATRICS scores differently by group suggesting that gamma-band power may play a role in cognitive deficits in SP. Also, medication was positively correlated with gamma-band power for only one of the multisensory clusters suggesting that medication alone cannot account for the differences in gamma-band power. Furthermore, SP showed decreases in gamma-band power in the peripheral visual (VN) condition relative to HC, as hypothesized. In contrast to our hypothesis, we identified both <italic>decreases</italic> and <italic>increases</italic> in gamma-band power in SP relative to HC. These group differences in multisensory gamma-band power were not directly accounted for by group differences in unisensory gamma-band power (e.g., the unisensory and multisensory group differences in gamma-band power did not simultaneously overlap in time, frequency, and location). Despite the lack of spatio-temporal overlap, certain unisensory clusters predicted gamma-band power in multisensory clusters indicating a cortical network of local gamma-band power that may influence gamma-band power in other regions. Finally, synchrony of the AV stimuli modulated group differences such that group differences were only obtained for the synchronous multisensory conditions. These results are discussed in more detail below.</p><p>Our results are consistent with the previous literature (Senkowski et al., <xref rid="B59" ref-type="bibr">2008</xref>) suggesting that gamma-band oscillations play a role in multisensory integration by identifying task-related increases and decreases in gamma-band power. Interestingly, despite the greater behavioral improvement in SP relative to HC (more facilitation of multisensory RTs relative to unisensory RTs), gamma-band power did not predict multisensory RT facilitation for either HC or SP. Previous studies indicated that increased gamma was associated with conscious recognition of the multisensory flash illusion (Mishra et al., <xref rid="B47" ref-type="bibr">2007</xref>) and stimuli with greater salience (e.g., looming AV stimuli&#x02014;Maier et al., <xref rid="B43" ref-type="bibr">2008</xref>); whereas other studies only found multisensory effects in other frequency bands (e.g., theta-band Naue et al., <xref rid="B51" ref-type="bibr">2011</xref>). Based on our current results, gamma-band oscillations do not play a direct role in facilitating multisensory RTs in SP. However, there is considerable variability in RTs in both HC and SP and single-trial analysis (not feasible in low SNR non-invasive studies) may be required to confidently conclude that gamma oscillations do not play a role in behavioral RT measures. Finally, Xu et al. (<xref rid="B92" ref-type="bibr">2013</xref>) determined that oscillatory activity across multiple frequency bands including gamma (30&#x02013;50 Hz) provided excellent (91% accuracy) discrimination between SP and HC groups during lexical processing. This may indicate that despite a lack of behavioral correlates, group differences in gamma-band oscillations may provide a means to better differentiate groups.</p><p>Perhaps the most surprising result from the current study is that 6 out of 7 clusters showed greater gamma-band power in SP relative to HC. However, in a number of cases (4/6) this represented a failure to suppress gamma-band power following the stimulus in SP rather than a significant increase from baseline gamma-band power. Only 2 of the clusters (AN-RT2 and AVSN-LF) represented increased gamma-band power from baseline in SP as well as significantly greater gamma-band power than HC. Increased gamma-band power in the AVSN-LF cluster may in part be explained by the positive correlation between gamma-band power and medication level in SP. Significant differences in gamma-band power in schizophrenia have been observed in a number of paradigms, but reports of decreases in gamma-band power in SP are more common than increases (for a review, see Sun et al., <xref rid="B69" ref-type="bibr">2011</xref>). It is important to note that some of these studies limited their analysis to specific regions of interest (e.g., Teale et al., <xref rid="B76" ref-type="bibr">2008</xref>; Oribe et al., <xref rid="B54" ref-type="bibr">2010</xref>), thereby limiting the scope of the study to the region analyzed. Our analysis limited the frequency range to the low gamma-band, yet performed comparisons across the full sensor array providing a broader view of changes in gamma-band power. Also, the regions that showed increased gamma in SP relative to HC (anterior temporal and frontal regions) are consistent with previous reports of increased gamma-band power in frontal regions in SP as summarized in a review by Sun et al. (<xref rid="B69" ref-type="bibr">2011</xref>). Furthermore, Tikka et al. (<xref rid="B77" ref-type="bibr">2013</xref>) reported increased gamma-band power (30&#x02013;50 Hz) in unmedicated schizophrenia patients with minor physical anomalies relative to controls over right frontal, temporal and parietal regions, similar to our findings. The remaining clusters with group differences showing SP &#x0003e; HC were associated with significant decreases in gamma-band power in HC (AN-RT1, AN-RC, AN-RF, and AVSN-RF) in which each cluster included auditory stimuli. Haenschel et al. (<xref rid="B31" ref-type="bibr">2009</xref>) found both increases and decreases in frontal gamma activity in SP and HC, depending on working memory load. While our task was not specifically designed as a working memory task <italic>per se</italic>, it required that participants maintain a representation of &#x0201c;Near&#x0201d; and &#x0201c;Far&#x0201d; stimuli to perform the discrimination in the fully randomized design. In conjunction with Haenschel and colleagues' results, this may indicate that the current task is tantamount to a low-demand working memory task for healthy controls but requires additional effort for SP, which is accompanied by increased gamma in these patients. In support of this hypothesis, SP performed more poorly than HC for the auditory discrimination, which was more difficult than the visual discrimination, and gamma-band power increases in SP were observed during the auditory and multisensory task.</p><p>To further characterize the role of unisensory processing deficits on multisensory gamma responses, we also investigated whether unisensory gamma-band power was predictive of multisensory gamma-band power. The non-linear transformation of the Morlet wavelet eliminates the ability to directly compare gamma-band power between A+V vs. AV (Senkowski et al., <xref rid="B58" ref-type="bibr">2007</xref>), as is commonly performed in multisensory evoked response studies (Calvert and Thesen, <xref rid="B17" ref-type="bibr">2004</xref>). Therefore, we assessed the influence of unisensory gamma-band power on multisensory gamma-band power through regression analyses. Different patterns of unisensory gamma predicting multisensory gamma are demonstrated in Table <xref ref-type="table" rid="T4">4</xref> for HC (relative to no predictors in SP). AN-RC predicted gamma-band power for AVSN-RF and AVSN-LF. In both cases the prediction showed a positive relationship suggesting that increased unisensory gamma-band power predicted increases in multisensory gamma-band power. However, the absence of a relationship between gamma-band power and RT may be related to the variability in the gamma response relative to RT.</p><p>As hypothesized, group differences were noted based on the location of the visual stimulus in the visual field (central-Far vs. peripheral-Near). Our results only reveal group differences in gamma-band power during the peripheral&#x02014;Near visual conditions. These results are consistent with previous studies indicating dorsal stream deficits in SP (Butler and Javitt, <xref rid="B14" ref-type="bibr">2005</xref>; Butler et al., <xref rid="B15" ref-type="bibr">2008</xref>). In this case SP showed significant decreases in gamma-band power relative to baseline and relative to HC; this result may indicate that gamma-band deficits in SP may contribute to impaired peripheral field processing in SP. Additionally, group differences were only found with synchronous presentation of auditory and visual stimuli. This result is contrary to our hypothesis that group differences would be obtained during the asynchronous condition due to differences in sensitivity to the temporal integration window (differences in the response to asynchrony of the AV stimuli&#x02014;Foucher et al., <xref rid="B24" ref-type="bibr">2007</xref>). This may indicate group differences in how gamma-band oscillations bind auditory and visual stimuli. Parametric manipulations of temporal synchrony are needed to better understand this result in relation to schizophrenia.</p><p>Finally, we investigated whether event-related power in the gamma-band clusters was predictive of cognitive outcome on five subtests of the MATRICS, and whether it was related to symptomology in SP. As shown in Tables <xref ref-type="table" rid="T6">6</xref> and <xref ref-type="table" rid="T7">7</xref>, gamma-band power predicted performance on the MATRICS for both HC and SP. Surprisingly, in most cases the relationship indicated that increased gamma-band power negatively correlated with MATRICS scores (in all cases for HC). Furthermore, multisensory gamma-band power only showed a relationship with MATRICS scores (verbal learning) in HC. On the other hand, only unisensory gamma-band power predicted MATRICS scores in SP. Unlike HC, the SP group showed an association between gamma-band power and MATRICS attention scores, with both AN and VN gamma-band power predicting the attention score. Yet, visual gamma-band power negatively predicted attention scores, whereas auditory gamma-band power positively predicted gamma-band power when controlling for visual gamma-band power. These results suggest that alterations in unisensory processing impact cognitive abilities in SP, consistent with previous visual and auditory studies (Butler et al., <xref rid="B15" ref-type="bibr">2008</xref>; Uhlhaas et al., <xref rid="B80" ref-type="bibr">2008</xref>; Smith et al., <xref rid="B63" ref-type="bibr">2010</xref>).</p><p>The underlying pathophysiology responsible for the gamma-band differences observed in this study and others could arise from a number of sources. GABAergic interneurons have been implicated in the generation and modulation of gamma-band oscillations in cortex and hippocampus (Traub et al., <xref rid="B78" ref-type="bibr">1996</xref>; Whittington et al., <xref rid="B89" ref-type="bibr">2000</xref>), and recent evidence suggests a critical role for GABA during audio-visual integration in rodents (Iurilli et al., <xref rid="B33" ref-type="bibr">2012</xref>). Furthermore, GABA synthesis is reduced in individuals with schizophrenia, and reduced GABA transmission in prefrontal cortex has been associated with cognitive deficits including working memory deficits (Akbarian et al., <xref rid="B5" ref-type="bibr">1995a</xref>,<xref rid="B6" ref-type="bibr">b</xref>). Yet another study found indications of decreased GABA_B in schizophrenia (Farzan et al., <xref rid="B23" ref-type="bibr">2010</xref>), which plays a role in modulating gamma oscillations. This alteration in GABA_B may explain increases in gamma-band power in SP as an improper inhibition of cortical oscillations. Another potential source of gamma-band differences is reduced or altered connectivity within and between brain areas in schizophrenia. Disruptions in anatomical and functional brain networks have been observed in individuals with schizophrenia (Burns et al., <xref rid="B12" ref-type="bibr">2003</xref>; Zhou et al., <xref rid="B94" ref-type="bibr">2007</xref>), and direct connectivity between sensory areas and other cortical regions has been implicated in generating the oscillatory changes observed during cross-modal stimulus presentations (Lakatos et al., <xref rid="B41" ref-type="bibr">2007</xref>). Based on the differences in gamma-band power in response to auditory and multisensory stimuli, this may indicate disconnection of auditory cortex in SP, thereby leading to altered local gamma-band power. Therefore, disordered cortical pathways may contribute to the group differences in gamma-band power observed in these results.</p><p>Our results extend previous work characterizing multisensory differences in task-related activity in schizophrenia (Stone et al., <xref rid="B68" ref-type="bibr">2011</xref>), and add to the effort to better understand the causes and consequences of schizophrenia. However, we recognize certain limitations exist in the current study that bear consideration and highlight the direction of future research. In this initial characterization, we have limited our analysis to the 30&#x02013;50 Hz gamma-band range; however, differences in other frequency bands likely exist. Evidence suggests that different frequencies are associated with different cognitive processes (for a review, see Ward, <xref rid="B88" ref-type="bibr">2003</xref>). For example, high gamma (&#x0003e;60 Hz) has been implicated in higher cognitive processes (Uhlhaas et al., <xref rid="B81" ref-type="bibr">2011</xref>). Exploring differences in other frequency bands and their interactions during multisensory processing in schizophrenia will be the focus of future work. Furthermore, unlike EEG, MEG does not suffer from propagation of artifacts from the reference electrode. Yet, it would be beneficial to extend the current results to a source-based analysis of oscillatory activity to better understand the network interplay associated with multisensory processing. Furthermore, directly linking these multisensory gamma differences to unisensory processing through empirical methods suggested by Senkowski et al. (<xref rid="B58" ref-type="bibr">2007</xref>) will allow us to better understand how differences in unisensory processing influence multisensory abilities in both HC and SP. Medication effects and differences in cognitive ability between groups further hamper our ability to fully understand the underlying aspects of schizophrenia independent of these confounds. A recent study by Tikka et al. (<xref rid="B77" ref-type="bibr">2013</xref>) indicated increased gamma-band power in unmedicated patients with schizophrenia relative to controls; however, medication levels only corresponded to gamma-band power in one of the group clusters, suggesting that this alone cannot explain the current results. Finally, Sivarao et al. (<xref rid="B62" ref-type="bibr">2013</xref>) reported that nicotine appears to enhance gamma-band power associated with the auditory steady-state response in rats. Nicotine is a known confound in the current study with decreased access to HC who smoke based on smoking cessation programs, thereby limiting the ability to match on this factor in large studies. As mentioned by Sivarao and others, nicotine may provide a means for SP to self-medicate and normalize brain function.</p><p>In summary, this study demonstrates that group differences between SP relative to HC in gamma-band activity are present in response to both unisensory and multisensory stimuli. Yet, the unisensory deficits do not directly map onto changes in gamma-band power in response to multisensory stimuli. Future work will tease apart the role of sensory parameters and attention in performing unisensory and multisensory tasks. Characterizing oscillations across the frequency spectrum will also likely provide broader insight into multisensory integration in schizophrenia and help provide a link between sensory and cognitive functions.</p><sec><title>Conflict of interest statement</title><p>The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.</p></sec></sec>

Estimating cause of adult (15+ years) death using InterVA-4 in a rural district of southern Ghana

<sec id="st1"><title>Background</title><p>Data needed to estimate causes of death and the pattern of these deaths are scarce in sub-Saharan Africa. Such data are very important for targeting, monitoring, and evaluating health interventions.</p></sec><sec id="st2"><title>Objective</title><p>To estimate the mortality rate and determine causes of death among adults (aged 15 years and older) in a rural district of southern Ghana, using the InterVA-4 model.</p></sec><sec id="st3"><title>Design</title><p>Data used were generated from verbal autopsies conducted for registered adult members of the Dodowa Health and Demographic Surveillance System who died between 2006 and 2010. The InterVA-4 model was used to assign the cause of death.</p></sec><sec id="st4"><title>Results</title><p>Overall, the mortality rate for the period under review was 7.5/1,000 person-years (py) for the general population and 10.4/1,000 py for those aged 15 and older. The leading cause of death was communicable diseases (CDs), with a malaria-specific mortality rate of 1.06/1,000 py. Pulmonary tuberculosis (TB)-specific mortality rate was the next highest (1.01/1,000 py). HIV/AIDS attributed deaths were lower among males than females. Non-communicable diseases (NCDs) contributed to 28.3% of the deaths with cause-specific mortality rate of 2.93/1,000 py. Stroke topped the list with cause-specific mortality rate of 0.69/1,000 py. As expected, young males (15&#x02013;49 years) contributed to more road traffic accident (RTA) deaths; they had a lower RTA cause-specific mortality rate than older males (50&#x02013;64 years).</p></sec><sec id="st5"><title>Conclusions</title><p>Data indicate that CDs (e.g. malaria and TB) remain the major cause of death with NCDs (e.g. stroke) following closely behind. Verbal autopsy data can provide the causes of mortality in poorly resourced settings where access to timely and accurate data is scarce.</p></sec>

<contrib contrib-type="author"><name><surname>Awini</surname><given-names>Elizabeth</given-names></name><xref ref-type="aff" rid="AF0001">1</xref><xref ref-type="aff" rid="AF0002">2</xref><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="corresp" rid="cor1">*</xref></contrib><contrib contrib-type="author"><name><surname>Sarpong</surname><given-names>Doris</given-names></name><xref ref-type="aff" rid="AF0001">1</xref><xref ref-type="aff" rid="AF0002">2</xref></contrib><contrib contrib-type="author"><name><surname>Adjei</surname><given-names>Alexander</given-names></name><xref ref-type="aff" rid="AF0001">1</xref><xref ref-type="aff" rid="AF0002">2</xref></contrib><contrib contrib-type="author"><name><surname>Manyeh</surname><given-names>Alfred Kwesi</given-names></name><xref ref-type="aff" rid="AF0001">1</xref><xref ref-type="aff" rid="AF0002">2</xref></contrib><contrib contrib-type="author"><name><surname>Amu</surname><given-names>Alberta</given-names></name><xref ref-type="aff" rid="AF0001">1</xref><xref ref-type="aff" rid="AF0002">2</xref></contrib><contrib contrib-type="author"><name><surname>Akweongo</surname><given-names>Patricia</given-names></name><xref ref-type="aff" rid="AF0001">1</xref><xref ref-type="aff" rid="AF0003">3</xref></contrib><contrib contrib-type="author"><name><surname>Adongo</surname><given-names>Philip</given-names></name><xref ref-type="aff" rid="AF0001">1</xref><xref ref-type="aff" rid="AF0003">3</xref></contrib><contrib contrib-type="author"><name><surname>Kukula</surname><given-names>Vida</given-names></name><xref ref-type="aff" rid="AF0001">1</xref><xref ref-type="aff" rid="AF0002">2</xref><xref ref-type="aff" rid="AF0003">3</xref></contrib><contrib contrib-type="author"><name><surname>Odonkor</surname><given-names>Gabriel</given-names></name><xref ref-type="aff" rid="AF0001">1</xref><xref ref-type="aff" rid="AF0002">2</xref></contrib><contrib contrib-type="author"><name><surname>Narh</surname><given-names>Solomon</given-names></name><xref ref-type="aff" rid="AF0001">1</xref><xref ref-type="aff" rid="AF0002">2</xref></contrib><contrib contrib-type="author"><name><surname>Gyapong</surname><given-names>Margaret</given-names></name><xref ref-type="aff" rid="AF0001">1</xref><xref ref-type="aff" rid="AF0002">2</xref></contrib>

Global Health Action

<p>The ability to have a healthy long life is an essential part of development. Survival can, thus, be said to be a measure of a country's development (<xref rid="CIT0001" ref-type="bibr">1</xref>). For interventions to be effectively implemented and tailored to the right people, it is important for policy makers and planners to be aware of what causes diseases and death among the population they serve, hence the need for accurate and timely data (<xref rid="CIT0002" ref-type="bibr">2</xref>). This will help in prioritizing interventions, using the most appropriate strategies for their delivery, and monitoring of their effectiveness (<xref rid="CIT0003" ref-type="bibr">3</xref>). Data for the estimation of causes of death and the pattern of these deaths are scarce in sub-Sahara Africa (SSA) (<xref rid="CIT0004" ref-type="bibr">4</xref>) unlike the developed countries where vital registration systems are well-developed and medical certifications of death are available (<xref rid="CIT0005" ref-type="bibr">5</xref>).</p><p>Studies have shown that conducting verbal autopsy (VA) is the best available approach in obtaining empirical information on the cause of many deaths in settings with poor or no routine death certification (<xref rid="CIT0005" ref-type="bibr">5</xref>, <xref rid="CIT0006" ref-type="bibr">6</xref>). The conducting of VA for community-reported death is a common practice among sites belonging to the International Network for Demographic Evaluation of Population and Their Health (INDEPTH) Network (<xref rid="CIT0007" ref-type="bibr">7</xref>, <xref rid="CIT0008" ref-type="bibr">8</xref>). VA is the process used to obtain information on the cause of death by interviewing close relatives or primary caregivers on the signs and symptoms experienced by the deceased and the sequence of events that led to the death of their relative (<xref rid="CIT0006" ref-type="bibr">6</xref>, <xref rid="CIT0009" ref-type="bibr">9</xref>).</p><p>The idea of VAs started in 1956 when WHO encouraged lay reporting of health information and a lay reporting form was developed. Diversity in the use of VA instruments demanded standardization; hence, in 2007 WHO standardized the VA form for use (<xref rid="CIT0010" ref-type="bibr">10</xref>). A detailed review of the process involved in the development of VA tools exists (<xref rid="CIT0011" ref-type="bibr">11</xref>). Traditionally, VAs are coded independently by at least two physicians to determine the cause of death. This has been demanding for sites in low- and middle-income countries where there is a lack of physicians to do clinical work, let alone code VAs. This creates huge backlogs of VAs that require coding. In fact, physician coding has been found to be an expensive and slow process (<xref rid="CIT0009" ref-type="bibr">9</xref>). Besides, physician's reliability and repeatability of interpreting the VAs has been questioned (<xref rid="CIT0012" ref-type="bibr">12</xref>). It is in this light that a probabilistic model has been developed to interpret VAs (InterVA) for the determination of causes of death (<xref rid="CIT0009" ref-type="bibr">9</xref>).</p><p>This model has been tested on data from both demographic surveillance sites and hospital records as well as in a number of studies and has been refined based on previous InterVA models (<xref rid="CIT0005" ref-type="bibr">5</xref>, <xref rid="CIT0009" ref-type="bibr">9</xref>, <xref rid="CIT0012" ref-type="bibr">12</xref>, <xref rid="CIT0013" ref-type="bibr">13</xref>). The revised InterVA-4 brings on a new standard of interpreting VA that fits into the WHO VA instrument in terms of cause of death categories and input indicators (<xref rid="CIT0014" ref-type="bibr">14</xref>). This new tool was used on the Dodowa Health and Demographic Surveillance System (DHDSS) data to estimate causes of deaths of those aged 15 and older for 2006&#x02013;2010.</p><p>This paper therefore focuses on estimating the mortality rate and determining the causes of death among the adult population (15 years and older) in a rural district of southern Ghana using the revised InterVA-4 model.</p><sec sec-type="methods" id="S0002"><title>Methods</title><sec id="S0002-S20001"><title>Study area</title><p>The study area is the DHDSS site. The DHDSS operates within the boundaries of the former Dangme West District (now the Shai-Osudoku and Ningo-Prampram districts); one of the ten districts within the Greater Accra Region of Ghana located in the southeastern part of Ghana, lying between latitude 5&#x000b0;45&#x02032; S and 6&#x000b0;05&#x02032; N and longitude 0&#x000b0;05&#x02032; E and 0&#x000b0;20&#x02032; W. The district covers about 41.5% (1528.9/km²) of the total land size within the region. It is about 40.8 km from the national capital of Ghana, Accra.</p><p>The site conducted its baseline survey in 2005. By the end of 2010, there were 22,767 households with 111,976 residents under surveillance. Persons younger than age 15 formed 40.5% of the population, which is similar to other developing countries (<xref rid="CIT0015" ref-type="bibr">15</xref>) with children younger than 5 years accounting for 15.2%. There were 87 males to every 100 females. Households headed by females constituted 39.1%. The district is fairly rural and the inhabitants are mainly fishermen, petty traders, and artisans, with a handful of civil servants. Detailed description of the study area is available elsewhere (<xref rid="CIT0016" ref-type="bibr">16</xref>). In total, 21 static health facilities delivered services in the district. Many inhabitants live more than 5 km away from government health facilities (<xref rid="CIT0017" ref-type="bibr">17</xref>). Malaria, diarrhea, Acute Respiratory Infection (ARI), hypertension, and skin diseases are the top five most common diseases seen at the outpatient departments in the district, with malaria ranking first (<xref rid="CIT0018" ref-type="bibr">18</xref>). Malaria prevalence in the district was estimated at about 7% in 2011 (<xref rid="CIT0019" ref-type="bibr">19</xref>). The national HIV/AIDS level was 2.0% in 2010, but that of the Greater Accra region was 2.6% (<xref rid="CIT0020" ref-type="bibr">20</xref>).</p></sec><sec id="S0002-S20002"><title>Death registration and VA procedures</title><p>The DHDSS collects vital statistics from all households in its Demographic Surveillance Area (DSA). Between 2006 and 2010, households were visited every 6 months and events such as pregnancies, births, deaths, and migration was registered. Community key informants (CKIs) were trained to pick the events in their communities to supplement those collected by the fieldworkers. Deaths of registered household members picked by fieldworkers and CKIs were followed up by trained field supervisors who conducted VAs using standard VA questionnaires, which are in three categories: neonatal (0&#x02013;27 days), children (28 days to below 12 years), and adult (12 years and above). Once an interview was completed, VA forms were returned to the field office of the HDSS for cross checking of inconsistencies and blanks.</p></sec><sec id="S0002-S20003"><title>The InterVA model</title><p>The InterVA-4 model version 4.02 which was used to estimate the cause of death for this paper is computer-based, and uses the Bayes&#x02019; theorem, in an attempt to overcome the longstanding limitations of alternative methods (<xref rid="CIT0001" ref-type="bibr">1</xref>, <xref rid="CIT0004" ref-type="bibr">4</xref>, <xref rid="CIT0006" ref-type="bibr">6</xref>), such as physicians coding. To use this model, there is the need to categorize the local conditions of malaria and HIV into &#x02018;high&#x02019; or &#x02018;low&#x02019; (<xref rid="CIT0009" ref-type="bibr">9</xref>). Because malaria is persistently number one on the list of top 10 diseases in the study area (<xref rid="CIT0019" ref-type="bibr">19</xref>), malaria was categorized as &#x02018;high&#x02019; and HIV as &#x02018;low&#x02019; in this analysis. More information on the development of this probabilistic model and its robustness are available elsewhere (<xref rid="CIT0009" ref-type="bibr">9</xref>, <xref rid="CIT0022" ref-type="bibr">22</xref>). InterVA-4 generates up to three probable causes of death for each case with their assigned likelihoods or indeterminate result. If the sum of the three likelihoods is less than 1, then the residual component is assigned as indeterminate. For cases in which the information is limited or inconsistent, that case is assigned indeterminate with a likelihood of 1. Registered deaths without VAs were assigned as &#x02018;VA not completed&#x02019;. This group was, however, added to the indeterminate during analysis. The InterVA-4 model was applied to the dataset using the methods previously described (<xref rid="CIT0014" ref-type="bibr">14</xref>) and as described in detail in the introductory paper of this issue (<xref rid="CIT0023" ref-type="bibr">23</xref>).</p></sec></sec><sec sec-type="results" id="S0003"><title>Results</title><sec id="S0003-S20001"><title>General trends in mortality between 2006 and 2010</title><p>Between 2006 and 2010, 3,988 deaths were registered with the DSA. Of these, 3,005 had a VA completed. In total, 3,324 of the registered deaths were aged 15 and older, and of those, 2,547 had a VA completed (76.7%). Among the 15 years and older, 1,158 deaths did not have their causes determined. Of these, 777 did not have VA completed, whereas for 381 of them, InterVA-4 could not assign any cause.</p><p>The overall mortality rate for the period was 7.5/1,000 person-years (py) whereas that of 15 years and older was 10.4/1,000 py. The crude mortality rate declined from 9.8/1,000 in 2006 to 6.6/1,000 py in 2010 (data not shown).</p><p> <xref ref-type="table" rid="T0001">Table 1</xref> presents the number of deaths, person-time observed, and the trends in mortality between 2006 and 2010. Mortality rates consistently declined for the age groups 15&#x02013;49 and 65 years and older for the 5-year period. Males generally had higher mortality rates than females.</p><table-wrap id="T0001" position="float"><label>Table 1</label><caption><p>Number of deaths, person-time observed, and mortality rates by sex, age, and year 2010</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1"/><th align="center" colspan="3" rowspan="1">15&#x02013;49 years</th><th align="center" colspan="3" rowspan="1">50&#x02013;64 years</th><th align="center" colspan="3" rowspan="1">65+ years</th></tr><tr><th align="left" rowspan="1" colspan="1"/><th align="center" colspan="3" rowspan="1"> <hr/> </th><th align="center" colspan="3" rowspan="1"> <hr/> </th><th align="center" colspan="3" rowspan="1"> <hr/> </th></tr><tr><th align="left" rowspan="1" colspan="1"/><th align="center" rowspan="1" colspan="1">Person year</th><th align="center" rowspan="1" colspan="1">Deaths</th><th align="center" rowspan="1" colspan="1">Rate/1,000 py</th><th align="center" rowspan="1" colspan="1">Person year</th><th align="center" rowspan="1" colspan="1">Deaths</th><th align="center" rowspan="1" colspan="1">Rate/1,000 py</th><th align="center" rowspan="1" colspan="1">Person year</th><th align="center" rowspan="1" colspan="1">Deaths</th><th align="center" rowspan="1" colspan="1">Rate/1,000 py</th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">2006</td><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Male</td><td align="center" rowspan="1" colspan="1">20,462</td><td align="center" rowspan="1" colspan="1">124</td><td align="center" rowspan="1" colspan="1">6.06</td><td align="center" rowspan="1" colspan="1">2,771</td><td align="center" rowspan="1" colspan="1">75</td><td align="center" rowspan="1" colspan="1">27.07</td><td align="center" rowspan="1" colspan="1">1,848</td><td align="center" rowspan="1" colspan="1">153</td><td align="center" rowspan="1" colspan="1">82.80</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Female</td><td align="center" rowspan="1" colspan="1">24,327</td><td align="center" rowspan="1" colspan="1">153</td><td align="center" rowspan="1" colspan="1">6.29</td><td align="center" rowspan="1" colspan="1">3,737</td><td align="center" rowspan="1" colspan="1">66</td><td align="center" rowspan="1" colspan="1">17.66</td><td align="center" rowspan="1" colspan="1">3,389</td><td align="center" rowspan="1" colspan="1">207</td><td align="center" rowspan="1" colspan="1">61.08</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Total</td><td align="center" rowspan="1" colspan="1">44,789</td><td align="center" rowspan="1" colspan="1">277</td><td align="center" rowspan="1" colspan="1">6.18</td><td align="center" rowspan="1" colspan="1">6,508</td><td align="center" rowspan="1" colspan="1">141</td><td align="center" rowspan="1" colspan="1">21.67</td><td align="center" rowspan="1" colspan="1">5,237</td><td align="center" rowspan="1" colspan="1">360</td><td align="center" rowspan="1" colspan="1">68.74</td></tr><tr><td align="left" rowspan="1" colspan="1">2007</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Male</td><td align="center" rowspan="1" colspan="1">22,096</td><td align="center" rowspan="1" colspan="1">124</td><td align="center" rowspan="1" colspan="1">5.61</td><td align="center" rowspan="1" colspan="1">3,004</td><td align="center" rowspan="1" colspan="1">50</td><td align="center" rowspan="1" colspan="1">16.64</td><td align="center" rowspan="1" colspan="1">1,879</td><td align="center" rowspan="1" colspan="1">128</td><td align="center" rowspan="1" colspan="1">68.13</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Female</td><td align="center" rowspan="1" colspan="1">26,635</td><td align="center" rowspan="1" colspan="1">130</td><td align="center" rowspan="1" colspan="1">4.88</td><td align="center" rowspan="1" colspan="1">4,015</td><td align="center" rowspan="1" colspan="1">57</td><td align="center" rowspan="1" colspan="1">14.20</td><td align="center" rowspan="1" colspan="1">3,471</td><td align="center" rowspan="1" colspan="1">162</td><td align="center" rowspan="1" colspan="1">46.68</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Total</td><td align="center" rowspan="1" colspan="1">48,731</td><td align="center" rowspan="1" colspan="1">254</td><td align="center" rowspan="1" colspan="1">5.21</td><td align="center" rowspan="1" colspan="1">7,019</td><td align="center" rowspan="1" colspan="1">107</td><td align="center" rowspan="1" colspan="1">15.24</td><td align="center" rowspan="1" colspan="1">5,349</td><td align="center" rowspan="1" colspan="1">290</td><td align="center" rowspan="1" colspan="1">54.21</td></tr><tr><td align="left" rowspan="1" colspan="1">2008</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Male</td><td align="center" rowspan="1" colspan="1">23,876</td><td align="center" rowspan="1" colspan="1">117</td><td align="center" rowspan="1" colspan="1">4.90</td><td align="center" rowspan="1" colspan="1">3,232</td><td align="center" rowspan="1" colspan="1">63</td><td align="center" rowspan="1" colspan="1">19.49</td><td align="center" rowspan="1" colspan="1">1,933</td><td align="center" rowspan="1" colspan="1">106</td><td align="center" rowspan="1" colspan="1">54.83</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Female</td><td align="center" rowspan="1" colspan="1">28,956</td><td align="center" rowspan="1" colspan="1">124</td><td align="center" rowspan="1" colspan="1">4.28</td><td align="center" rowspan="1" colspan="1">4,311</td><td align="center" rowspan="1" colspan="1">63</td><td align="center" rowspan="1" colspan="1">14.61</td><td align="center" rowspan="1" colspan="1">3,576</td><td align="center" rowspan="1" colspan="1">184</td><td align="center" rowspan="1" colspan="1">51.45</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Total</td><td align="center" rowspan="1" colspan="1">52,832</td><td align="center" rowspan="1" colspan="1">241</td><td align="center" rowspan="1" colspan="1">4.56</td><td align="center" rowspan="1" colspan="1">7,543</td><td align="center" rowspan="1" colspan="1">126</td><td align="center" rowspan="1" colspan="1">16.70</td><td align="center" rowspan="1" colspan="1">5,509</td><td align="center" rowspan="1" colspan="1">290</td><td align="center" rowspan="1" colspan="1">52.64</td></tr><tr><td align="left" rowspan="1" colspan="1">2009</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Male</td><td align="center" rowspan="1" colspan="1">24,310</td><td align="center" rowspan="1" colspan="1">102</td><td align="center" rowspan="1" colspan="1">4.20</td><td align="center" rowspan="1" colspan="1">3,299</td><td align="center" rowspan="1" colspan="1">62</td><td align="center" rowspan="1" colspan="1">18.80</td><td align="center" rowspan="1" colspan="1">1,921</td><td align="center" rowspan="1" colspan="1">107</td><td align="center" rowspan="1" colspan="1">55.69</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Female</td><td align="center" rowspan="1" colspan="1">29,668</td><td align="center" rowspan="1" colspan="1">129</td><td align="center" rowspan="1" colspan="1">4.35</td><td align="center" rowspan="1" colspan="1">4,397</td><td align="center" rowspan="1" colspan="1">57</td><td align="center" rowspan="1" colspan="1">12.96</td><td align="center" rowspan="1" colspan="1">3,580</td><td align="center" rowspan="1" colspan="1">170</td><td align="center" rowspan="1" colspan="1">47.48</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Total</td><td align="center" rowspan="1" colspan="1">53,978</td><td align="center" rowspan="1" colspan="1">231</td><td align="center" rowspan="1" colspan="1">4.28</td><td align="center" rowspan="1" colspan="1">7,696</td><td align="center" rowspan="1" colspan="1">119</td><td align="center" rowspan="1" colspan="1">15.46</td><td align="center" rowspan="1" colspan="1">5,502</td><td align="center" rowspan="1" colspan="1">277</td><td align="center" rowspan="1" colspan="1">50.35</td></tr><tr><td align="left" rowspan="1" colspan="1">2010</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Male</td><td align="center" rowspan="1" colspan="1">24,859</td><td align="center" rowspan="1" colspan="1">87</td><td align="center" rowspan="1" colspan="1">3.50</td><td align="center" rowspan="1" colspan="1">3,516</td><td align="center" rowspan="1" colspan="1">76</td><td align="center" rowspan="1" colspan="1">21.61</td><td align="center" rowspan="1" colspan="1">1,986</td><td align="center" rowspan="1" colspan="1">105</td><td align="center" rowspan="1" colspan="1">52.86</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Female</td><td align="center" rowspan="1" colspan="1">30,488</td><td align="center" rowspan="1" colspan="1">111</td><td align="center" rowspan="1" colspan="1">3.64</td><td align="center" rowspan="1" colspan="1">4,719</td><td align="center" rowspan="1" colspan="1">65</td><td align="center" rowspan="1" colspan="1">13.78</td><td align="center" rowspan="1" colspan="1">3,644</td><td align="center" rowspan="1" colspan="1">167</td><td align="center" rowspan="1" colspan="1">45.83</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Total</td><td align="center" rowspan="1" colspan="1">55,346</td><td align="center" rowspan="1" colspan="1">198</td><td align="center" rowspan="1" colspan="1">3.58</td><td align="center" rowspan="1" colspan="1">8,235</td><td align="center" rowspan="1" colspan="1">141</td><td align="center" rowspan="1" colspan="1">17.12</td><td align="center" rowspan="1" colspan="1">5,631</td><td align="center" rowspan="1" colspan="1">272</td><td align="center" rowspan="1" colspan="1">48.31</td></tr></tbody></table></table-wrap></sec><sec id="S0003-S20002"><title>Cause-specific mortality as determined by InterVA-4</title><p>Malaria was found to be the leading cause of death with a cause-specific mortality rate of 1.06/1,000 py, followed by pulmonary tuberculosis (TB), with 1.01/1,000 py. Cause-specific mortality rate for stroke was 0.69/1,000 py, whereas that of ARIs and digestive neoplasms were 0.54/1,000 py and 0.47/1,000 py, respectively. The cause-specific mortality rates for acute cardiac diseases and acute abdomen were 0.46/1,000 py and 0.44/1,000 py, respectively (<xref ref-type="table" rid="T0002">Table 2</xref>).</p><table-wrap id="T0002" position="float"><label>Table 2</label><caption><p>Deaths calculated as sum of fractional likelihoods, then rounded to nearest whole number</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1">Cause of death</th><th align="center" rowspan="1" colspan="1">Deaths<xref ref-type="table-fn" rid="TF0001">*</xref> </th><th align="center" rowspan="1" colspan="1">CSMFs (%)</th><th align="center" rowspan="1" colspan="1">Rate/1,000 py</th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">Malaria</td><td align="center" rowspan="1" colspan="1">339</td><td align="center" rowspan="1" colspan="1">10.21</td><td align="center" rowspan="1" colspan="1">1.06</td></tr><tr><td align="left" rowspan="1" colspan="1">Pulmonary tuberculosis</td><td align="center" rowspan="1" colspan="1">322</td><td align="center" rowspan="1" colspan="1">9.7</td><td align="center" rowspan="1" colspan="1">1.01</td></tr><tr><td align="left" rowspan="1" colspan="1">Stroke</td><td align="center" rowspan="1" colspan="1">220</td><td align="center" rowspan="1" colspan="1">6.62</td><td align="center" rowspan="1" colspan="1">0.69</td></tr><tr><td align="left" rowspan="1" colspan="1">ARIs</td><td align="center" rowspan="1" colspan="1">174</td><td align="center" rowspan="1" colspan="1">5.22</td><td align="center" rowspan="1" colspan="1">0.54</td></tr><tr><td align="left" rowspan="1" colspan="1">Digestive neoplasms</td><td align="center" rowspan="1" colspan="1">150</td><td align="center" rowspan="1" colspan="1">4.52</td><td align="center" rowspan="1" colspan="1">0.47</td></tr><tr><td align="left" rowspan="1" colspan="1">Acute cardiac disease</td><td align="center" rowspan="1" colspan="1">146</td><td align="center" rowspan="1" colspan="1">4.4</td><td align="center" rowspan="1" colspan="1">0.46</td></tr><tr><td align="left" rowspan="1" colspan="1">Acute abdomen</td><td align="center" rowspan="1" colspan="1">141</td><td align="center" rowspan="1" colspan="1">4.25</td><td align="center" rowspan="1" colspan="1">0.44</td></tr><tr><td align="left" rowspan="1" colspan="1">Other and unspecified infect diseases</td><td align="center" rowspan="1" colspan="1">120</td><td align="center" rowspan="1" colspan="1">3.61</td><td align="center" rowspan="1" colspan="1">0.38</td></tr><tr><td align="left" rowspan="1" colspan="1">Road traffic accident</td><td align="center" rowspan="1" colspan="1">70</td><td align="center" rowspan="1" colspan="1">2.11</td><td align="center" rowspan="1" colspan="1">0.22</td></tr><tr><td align="left" rowspan="1" colspan="1">Other and unspecified cardiac diseases</td><td align="center" rowspan="1" colspan="1">61</td><td align="center" rowspan="1" colspan="1">1.82</td><td align="center" rowspan="1" colspan="1">0.19</td></tr><tr><td align="left" rowspan="1" colspan="1">HIV/AIDS-related death</td><td align="center" rowspan="1" colspan="1">59</td><td align="center" rowspan="1" colspan="1">1.76</td><td align="center" rowspan="1" colspan="1">0.18</td></tr><tr><td align="left" rowspan="1" colspan="1">Other and unspecified neoplasms</td><td align="center" rowspan="1" colspan="1">38</td><td align="center" rowspan="1" colspan="1">1.14</td><td align="center" rowspan="1" colspan="1">0.12</td></tr><tr><td align="left" rowspan="1" colspan="1">Accidental fall</td><td align="center" rowspan="1" colspan="1">29</td><td align="center" rowspan="1" colspan="1">0.86</td><td align="center" rowspan="1" colspan="1">0.09</td></tr><tr><td align="left" rowspan="1" colspan="1">Respiratory neoplasms</td><td align="center" rowspan="1" colspan="1">26</td><td align="center" rowspan="1" colspan="1">0.78</td><td align="center" rowspan="1" colspan="1">0.08</td></tr><tr><td align="left" rowspan="1" colspan="1">Indeterminate</td><td align="center" rowspan="1" colspan="1">381</td><td align="center" rowspan="1" colspan="1">11.45</td><td align="center" rowspan="1" colspan="1">1.19</td></tr><tr><td align="left" rowspan="1" colspan="1">VA not completed</td><td align="center" rowspan="1" colspan="1">777</td><td align="center" rowspan="1" colspan="1">23.38</td><td align="center" rowspan="1" colspan="1">2.43</td></tr></tbody></table><table-wrap-foot><fn id="TF0001"><label>*</label><p>Deaths calculated as sum of fractional likelihoods, then rounded to nearest whole number.</p></fn></table-wrap-foot></table-wrap></sec><sec id="S0003-S20003"><title>Causes of deaths and mortality rates for males and females aged 15 and above</title><p> <xref ref-type="table" rid="T0003">Table 3</xref> shows the leading causes of deaths for males and females by mortality rates and age groups. TB was the first leading cause of death for males in all the age groups, whereas malaria was the leading cause of death in females within the 15&#x02013;49 and 65 years and above age groups. However, mortality rate attributable to stroke was the leading cause of death for the females in the 50&#x02013;64 years age group. Mortality rate due to deaths from road traffic accident (RTA) was higher in males than females. HIV/AIDs mortality rate was higher in females than in males. Unexpectedly, digestive neoplasms were among the leading causes of death. Males in the age group 50 years and above had higher mortality rates than females, whereas in the 15&#x02013;49 years age group females had a higher mortality rate.</p><table-wrap id="T0003" position="float"><label>Table 3</label><caption><p>Deaths calculated as sum of fractional likelihoods, then rounded to nearest whole number</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1"/><th align="center" colspan="2" rowspan="1">Male</th><th align="center" rowspan="1" colspan="1"/><th align="center" colspan="2" rowspan="1">Female</th></tr><tr><th align="left" rowspan="1" colspan="1"/><th align="center" colspan="2" rowspan="1"> <hr/> </th><th align="center" rowspan="1" colspan="1"/><th align="center" colspan="2" rowspan="1"> <hr/> </th></tr><tr><th align="left" rowspan="1" colspan="1">Cause of deaths</th><th align="center" rowspan="1" colspan="1">Deaths<xref ref-type="table-fn" rid="TF0002">*</xref> </th><th align="center" rowspan="1" colspan="1">Rate/1,000 py</th><th align="center" rowspan="1" colspan="1">Cause of deaths</th><th align="center" rowspan="1" colspan="1">Deaths<xref ref-type="table-fn" rid="TF0002">*</xref> </th><th align="center" rowspan="1" colspan="1">Rate/1,000 py</th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">15&#x02013;49 years</td><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Pulmonary tuberculosis</td><td align="center" rowspan="1" colspan="1">72</td><td align="center" rowspan="1" colspan="1">0.62</td><td align="left" rowspan="1" colspan="1">Malaria</td><td align="center" rowspan="1" colspan="1">103</td><td align="center" rowspan="1" colspan="1">0.74</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Malaria</td><td align="center" rowspan="1" colspan="1">51</td><td align="center" rowspan="1" colspan="1">0.45</td><td align="left" rowspan="1" colspan="1">Pulmonary tuberculosis</td><td align="center" rowspan="1" colspan="1">64</td><td align="center" rowspan="1" colspan="1">0.45</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Road traffic accident</td><td align="center" rowspan="1" colspan="1">36</td><td align="center" rowspan="1" colspan="1">0.31</td><td align="left" rowspan="1" colspan="1">HIV/AIDS-related death</td><td align="center" rowspan="1" colspan="1">32</td><td align="center" rowspan="1" colspan="1">0.23</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Other unspecified infections</td><td align="center" rowspan="1" colspan="1">28</td><td align="center" rowspan="1" colspan="1">0.24</td><td align="left" rowspan="1" colspan="1">Digestive neoplasms</td><td align="center" rowspan="1" colspan="1">26</td><td align="center" rowspan="1" colspan="1">0.18</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Acute abdomen</td><td align="center" rowspan="1" colspan="1">23</td><td align="center" rowspan="1" colspan="1">0.2</td><td align="left" rowspan="1" colspan="1">Acute abdomen</td><td align="center" rowspan="1" colspan="1">26</td><td align="center" rowspan="1" colspan="1">0.18</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;ARIs</td><td align="center" rowspan="1" colspan="1">22</td><td align="center" rowspan="1" colspan="1">0.19</td><td align="left" rowspan="1" colspan="1">ARIs</td><td align="center" rowspan="1" colspan="1">25</td><td align="center" rowspan="1" colspan="1">0.18</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Digestive neoplasms</td><td align="center" rowspan="1" colspan="1">17</td><td align="center" rowspan="1" colspan="1">0.14</td><td align="left" rowspan="1" colspan="1">Stroke</td><td align="center" rowspan="1" colspan="1">17</td><td align="center" rowspan="1" colspan="1">0.12</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Stroke</td><td align="center" rowspan="1" colspan="1">13</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="left" rowspan="1" colspan="1">Other unspecified infections</td><td align="center" rowspan="1" colspan="1">15</td><td align="center" rowspan="1" colspan="1">0.10</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Accidental drowning</td><td align="center" rowspan="1" colspan="1">12</td><td align="center" rowspan="1" colspan="1">0.10</td><td align="left" rowspan="1" colspan="1">Acute cardiac disease</td><td align="center" rowspan="1" colspan="1">13</td><td align="center" rowspan="1" colspan="1">0.10</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;HIV/AIDS-related death</td><td align="center" rowspan="1" colspan="1">10</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="left" rowspan="1" colspan="1">Obstetric hemorrhage</td><td align="center" rowspan="1" colspan="1">13</td><td align="center" rowspan="1" colspan="1">0.09</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Acute cardiac disease</td><td align="center" rowspan="1" colspan="1">10</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="left" rowspan="1" colspan="1">Road traffic accident</td><td align="center" rowspan="1" colspan="1">10</td><td align="center" rowspan="1" colspan="1">0.07</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Indeterminate</td><td align="center" rowspan="1" colspan="1">62</td><td align="center" rowspan="1" colspan="1">0.54</td><td align="left" rowspan="1" colspan="1">Indeterminate</td><td align="center" rowspan="1" colspan="1">71</td><td align="center" rowspan="1" colspan="1">0.5</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;VA not completed</td><td align="center" rowspan="1" colspan="1">141</td><td align="center" rowspan="1" colspan="1">1.22</td><td align="left" rowspan="1" colspan="1">VA not completed</td><td align="center" rowspan="1" colspan="1">152</td><td align="center" rowspan="1" colspan="1">1.09</td></tr><tr><td align="left" rowspan="1" colspan="1">50&#x02013;64 years</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Pulmonary tuberculosis</td><td align="center" rowspan="1" colspan="1">35</td><td align="center" rowspan="1" colspan="1">2.20</td><td align="left" rowspan="1" colspan="1">Stroke</td><td align="center" rowspan="1" colspan="1">30</td><td align="center" rowspan="1" colspan="1">1.42</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Digestive neoplasms</td><td align="center" rowspan="1" colspan="1">26</td><td align="center" rowspan="1" colspan="1">1.62</td><td align="left" rowspan="1" colspan="1">Pulmonary tuberculosis</td><td align="center" rowspan="1" colspan="1">29</td><td align="center" rowspan="1" colspan="1">1.36</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;ARIs</td><td align="center" rowspan="1" colspan="1">23</td><td align="center" rowspan="1" colspan="1">1.44</td><td align="left" rowspan="1" colspan="1">Malaria</td><td align="center" rowspan="1" colspan="1">25</td><td align="center" rowspan="1" colspan="1">1.18</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Stroke</td><td align="center" rowspan="1" colspan="1">22</td><td align="center" rowspan="1" colspan="1">1.42</td><td align="left" rowspan="1" colspan="1">Acute cardiac disease</td><td align="center" rowspan="1" colspan="1">15</td><td align="center" rowspan="1" colspan="1">0.70</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Acute cardiac disease</td><td align="center" rowspan="1" colspan="1">22</td><td align="center" rowspan="1" colspan="1">1.36</td><td align="left" rowspan="1" colspan="1">Digestive neoplasms</td><td align="center" rowspan="1" colspan="1">15</td><td align="center" rowspan="1" colspan="1">0.70</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Malaria</td><td align="center" rowspan="1" colspan="1">21</td><td align="center" rowspan="1" colspan="1">1.32</td><td align="left" rowspan="1" colspan="1">Other unspecified infections</td><td align="center" rowspan="1" colspan="1">13</td><td align="center" rowspan="1" colspan="1">0.63</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Other unspecified infections</td><td align="center" rowspan="1" colspan="1">15</td><td align="center" rowspan="1" colspan="1">0.93</td><td align="left" rowspan="1" colspan="1">ARIs</td><td align="center" rowspan="1" colspan="1">13</td><td align="center" rowspan="1" colspan="1">0.61</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Acute abdomen</td><td align="center" rowspan="1" colspan="1">14</td><td align="center" rowspan="1" colspan="1">0.86</td><td align="left" rowspan="1" colspan="1">HIV/AIDS-related death</td><td align="center" rowspan="1" colspan="1">9</td><td align="center" rowspan="1" colspan="1">0.41</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Road traffic accident</td><td align="center" rowspan="1" colspan="1">9</td><td align="center" rowspan="1" colspan="1">0.56</td><td align="left" rowspan="1" colspan="1">Acute abdomen</td><td align="center" rowspan="1" colspan="1">8</td><td align="center" rowspan="1" colspan="1">0.39</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Other unspecified cardiac disease</td><td align="center" rowspan="1" colspan="1">7</td><td align="center" rowspan="1" colspan="1">0.43</td><td align="left" rowspan="1" colspan="1">Other unspecified cardiac disease</td><td align="center" rowspan="1" colspan="1">8</td><td align="center" rowspan="1" colspan="1">0.39</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;HIV/AIDS-related death</td><td align="center" rowspan="1" colspan="1">5</td><td align="center" rowspan="1" colspan="1">0.3</td><td align="left" rowspan="1" colspan="1">Road traffic accident</td><td align="center" rowspan="1" colspan="1">7</td><td align="center" rowspan="1" colspan="1">0.32</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Diabetes mellitus</td><td align="center" rowspan="1" colspan="1">4</td><td align="center" rowspan="1" colspan="1">0.26</td><td align="left" rowspan="1" colspan="1">Severe anemia</td><td align="center" rowspan="1" colspan="1">4</td><td align="center" rowspan="1" colspan="1">0.2</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Indeterminate</td><td align="center" rowspan="1" colspan="1">38</td><td align="center" rowspan="1" colspan="1">2.41</td><td align="left" rowspan="1" colspan="1">Indeterminate</td><td align="center" rowspan="1" colspan="1">28</td><td align="center" rowspan="1" colspan="1">1.32</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;VA not completed</td><td align="center" rowspan="1" colspan="1">70</td><td align="center" rowspan="1" colspan="1">4.42</td><td align="left" rowspan="1" colspan="1">VA not completed</td><td align="center" rowspan="1" colspan="1">79</td><td align="center" rowspan="1" colspan="1">3.73</td></tr><tr><td align="left" rowspan="1" colspan="1">65+ years</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Pulmonary tuberculosis</td><td align="center" rowspan="1" colspan="1">59</td><td align="center" rowspan="1" colspan="1">6.17</td><td align="left" rowspan="1" colspan="1">Malaria</td><td align="center" rowspan="1" colspan="1">92</td><td align="center" rowspan="1" colspan="1">5.24</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Stroke</td><td align="center" rowspan="1" colspan="1">47</td><td align="center" rowspan="1" colspan="1">4.93</td><td align="left" rowspan="1" colspan="1">Stroke</td><td align="center" rowspan="1" colspan="1">90</td><td align="center" rowspan="1" colspan="1">5.09</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Malaria</td><td align="center" rowspan="1" colspan="1">46</td><td align="center" rowspan="1" colspan="1">4.83</td><td align="left" rowspan="1" colspan="1">Pulmonary tuberculosis</td><td align="center" rowspan="1" colspan="1">64</td><td align="center" rowspan="1" colspan="1">3.62</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Acute cardiac disease</td><td align="center" rowspan="1" colspan="1">41</td><td align="center" rowspan="1" colspan="1">4.3</td><td align="left" rowspan="1" colspan="1">ARIs</td><td align="center" rowspan="1" colspan="1">53</td><td align="center" rowspan="1" colspan="1">3.00</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;ARIs</td><td align="center" rowspan="1" colspan="1">39</td><td align="center" rowspan="1" colspan="1">4.03</td><td align="left" rowspan="1" colspan="1">Acute cardiac disease</td><td align="center" rowspan="1" colspan="1">46</td><td align="center" rowspan="1" colspan="1">2.60</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Acute abdomen</td><td align="center" rowspan="1" colspan="1">33</td><td align="center" rowspan="1" colspan="1">3.44</td><td align="left" rowspan="1" colspan="1">Acute abdomen</td><td align="center" rowspan="1" colspan="1">38</td><td align="center" rowspan="1" colspan="1">2.14</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Digestive neoplasms</td><td align="center" rowspan="1" colspan="1">32</td><td align="center" rowspan="1" colspan="1">3.37</td><td align="left" rowspan="1" colspan="1">Digestive neoplasms</td><td align="center" rowspan="1" colspan="1">35</td><td align="center" rowspan="1" colspan="1">1.99</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Other unspecified infections</td><td align="center" rowspan="1" colspan="1">21</td><td align="center" rowspan="1" colspan="1">2.24</td><td align="left" rowspan="1" colspan="1">Other unspecified infections</td><td align="center" rowspan="1" colspan="1">28</td><td align="center" rowspan="1" colspan="1">1.60</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Other unspecified cardiac disease</td><td align="center" rowspan="1" colspan="1">14</td><td align="center" rowspan="1" colspan="1">1.48</td><td align="left" rowspan="1" colspan="1">Other unspecified cardiac disease</td><td align="center" rowspan="1" colspan="1">24</td><td align="center" rowspan="1" colspan="1">1.34</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Other unspecified neoplasms</td><td align="center" rowspan="1" colspan="1">11</td><td align="center" rowspan="1" colspan="1">1.18</td><td align="left" rowspan="1" colspan="1">Accidental fall</td><td align="center" rowspan="1" colspan="1">16</td><td align="center" rowspan="1" colspan="1">0.90</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Diabetes mellitus</td><td align="center" rowspan="1" colspan="1">7</td><td align="center" rowspan="1" colspan="1">0.78</td><td align="left" rowspan="1" colspan="1">Other unspecified neoplasms</td><td align="center" rowspan="1" colspan="1">15</td><td align="center" rowspan="1" colspan="1">0.84</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Asthma</td><td align="center" rowspan="1" colspan="1">6</td><td align="center" rowspan="1" colspan="1">0.62</td><td align="left" rowspan="1" colspan="1">Severe malnutrition</td><td align="center" rowspan="1" colspan="1">11</td><td align="center" rowspan="1" colspan="1">0.61</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Respiratory neoplasms</td><td align="center" rowspan="1" colspan="1">5</td><td align="center" rowspan="1" colspan="1">0.56</td><td align="left" rowspan="1" colspan="1">Severe anemia</td><td align="center" rowspan="1" colspan="1">10</td><td align="center" rowspan="1" colspan="1">0.55</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Indeterminate</td><td align="center" rowspan="1" colspan="1">65</td><td align="center" rowspan="1" colspan="1">6.82</td><td align="left" rowspan="1" colspan="1">Indeterminate</td><td align="center" rowspan="1" colspan="1">116</td><td align="center" rowspan="1" colspan="1">6.59</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;VA not completed</td><td align="center" rowspan="1" colspan="1">141</td><td align="center" rowspan="1" colspan="1">14.74</td><td align="left" rowspan="1" colspan="1">VA not completed</td><td align="center" rowspan="1" colspan="1">194</td><td align="center" rowspan="1" colspan="1">10.99</td></tr></tbody></table><table-wrap-foot><fn id="TF0002"><label>*</label><p>Deaths calculated as sum of fractional likelihoods, then rounded to nearest whole number.</p></fn></table-wrap-foot></table-wrap></sec><sec id="S0003-S20004"><title> Distribution of causes of deaths by cause category</title><p>Considering cause categories, communicable diseases (CDs) were the leading causes of death, with a cause-specific mortality rate of 3.29/1,000 py. Non-communicable diseases (NCDs) (hypertension, stroke, acute cardiac disease, diabetes mellitus, severe anemia, sickle cell disease, chronic obstructive pulmonary disease, asthma, acute abdomen, liver cirrhosis, renal failure, epilepsy, neoplasm, severe malnutrition, other unspecified cardiac disease, and other unspecified NCD) had a mortality rate of 2.93/1,000 py. Trauma or injury-specific mortality rate was 0.47/1,000 py, whereas that of maternal-related causes was 0.08/1,000 py (data not shown).</p></sec><sec id="S0003-S20005"><title>Cause of death by cause group by sex</title><p>The mortality rates and the pattern of distribution of causes of death were virtually the same in both males and females except with trauma (<xref ref-type="fig" rid="F0001">Fig. 1</xref>). Mortality rate was highest in CDs, followed by NCDs, and then trauma. The rates of CDs (3.30/1,000 py) and NCDs (2.99/1,000 py) in females (3.28/1,000 py) were slightly higher than in males (2.86/1,000 py). Males had higher mortality rate from trauma (0.68/1,000 py) than females (0.30/1,000 py).</p><fig id="F0001" position="float"><label>Fig. 1</label><caption><p>Mortality rates per 1,000 person-years by cause group by sex (2006&#x02013;2010).</p></caption><graphic xlink:href="GHA-7-25543-g001"/></fig></sec><sec id="S0003-S20006"><title>Cause of death by age</title><p>The pattern of cause of death differed for the various age groups (<xref ref-type="table" rid="T0004">Table 4</xref>). Generally, mortality rates for all cause groups, except maternal mortality, increased with age. Mortality rates from NCDs were higher in males than in females for the 50 and above age groups, whereas for the 15&#x02013;49 years age group, the rate was higher in females. Among the same age group, mortality rate for maternal causes was 0.19/1,000 py.</p><table-wrap id="T0004" position="float"><label>Table 4</label><caption><p>Deaths calculated as sum of fractional likelihoods, then rounded to nearest whole number</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1"/><th align="center" colspan="2" rowspan="1">Male</th><th align="center" colspan="2" rowspan="1">Female</th></tr><tr><th align="left" rowspan="1" colspan="1"/><th align="center" colspan="2" rowspan="1"> <hr/> </th><th align="center" colspan="2" rowspan="1"> <hr/> </th></tr><tr><th align="left" rowspan="1" colspan="1">Cause of deaths</th><th align="center" rowspan="1" colspan="1">Deaths<xref ref-type="table-fn" rid="TF0003">*</xref> </th><th align="center" rowspan="1" colspan="1">Rate/1,000 py</th><th align="center" rowspan="1" colspan="1">CSMFs</th><th align="center" rowspan="1" colspan="1">Rate/1,000 py</th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">15&#x02013;49 years</td><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1"/><td align="left" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Communicable</td><td align="center" rowspan="1" colspan="1">219</td><td align="center" rowspan="1" colspan="1">1.89</td><td align="center" rowspan="1" colspan="1">253</td><td align="center" rowspan="1" colspan="1">1.80</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Non-communicable</td><td align="center" rowspan="1" colspan="1">92</td><td align="center" rowspan="1" colspan="1">0.79</td><td align="center" rowspan="1" colspan="1">122</td><td align="center" rowspan="1" colspan="1">0.87</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Maternal</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">26</td><td align="center" rowspan="1" colspan="1">0.19</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Trauma</td><td align="center" rowspan="1" colspan="1">68</td><td align="center" rowspan="1" colspan="1">0.59</td><td align="center" rowspan="1" colspan="1">24</td><td align="center" rowspan="1" colspan="1">0.17</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Indeterminate</td><td align="center" rowspan="1" colspan="1">203</td><td align="center" rowspan="1" colspan="1">1.76</td><td align="center" rowspan="1" colspan="1">223</td><td align="center" rowspan="1" colspan="1">1.59</td></tr><tr><td align="left" rowspan="1" colspan="1">50&#x02013;64 years</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Communicable</td><td align="center" rowspan="1" colspan="1">98</td><td align="center" rowspan="1" colspan="1">6.22</td><td align="center" rowspan="1" colspan="1">121</td><td align="center" rowspan="1" colspan="1">5.72</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Non-communicable</td><td align="center" rowspan="1" colspan="1">103</td><td align="center" rowspan="1" colspan="1">6.53</td><td align="center" rowspan="1" colspan="1">102</td><td align="center" rowspan="1" colspan="1">4.82</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Maternal</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Trauma</td><td align="center" rowspan="1" colspan="1">16</td><td align="center" rowspan="1" colspan="1">1.02</td><td align="center" rowspan="1" colspan="1">9</td><td align="center" rowspan="1" colspan="1">0.41</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Indeterminate</td><td align="center" rowspan="1" colspan="1">108</td><td align="center" rowspan="1" colspan="1">6.84</td><td align="center" rowspan="1" colspan="1">107</td><td align="center" rowspan="1" colspan="1">5.06</td></tr><tr><td align="left" rowspan="1" colspan="1">65+ years</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Communicable</td><td align="center" rowspan="1" colspan="1">174</td><td align="center" rowspan="1" colspan="1">18.14</td><td align="center" rowspan="1" colspan="1">247</td><td align="center" rowspan="1" colspan="1">13.98</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Non-communicable</td><td align="center" rowspan="1" colspan="1">208</td><td align="center" rowspan="1" colspan="1">21.74</td><td align="center" rowspan="1" colspan="1">311</td><td align="center" rowspan="1" colspan="1">17.62</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Maternal</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Trauma</td><td align="center" rowspan="1" colspan="1">11</td><td align="center" rowspan="1" colspan="1">1.16</td><td align="center" rowspan="1" colspan="1">22</td><td align="center" rowspan="1" colspan="1">1.22</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Indeterminate</td><td align="center" rowspan="1" colspan="1">206</td><td align="center" rowspan="1" colspan="1">21.56</td><td align="center" rowspan="1" colspan="1">310</td><td align="center" rowspan="1" colspan="1">17.57</td></tr></tbody></table><table-wrap-foot><fn id="TF0003"><label>*</label><p>Deaths calculated as sum of fractional likelihoods, then rounded to nearest whole number.</p></fn></table-wrap-foot></table-wrap></sec><sec id="S0003-S20007"><title>Mortality rates and causes of death by cause group by year</title><p>The information in <xref ref-type="fig" rid="F0002">Fig. 2</xref> shows that with the exception of maternal-related causes, mortality rates attributable to all causes of death by cause group recorded the same pattern in the 5-year period. There was no recorded maternal-related death in 2008, whereas the highest mortality rate was recorded in 2006 (0.17/1,000 py). Indeterminate rate was highest in 2007 (4.71/1,000 py) and lowest in 2010 (2.39/1,000 py).</p><fig id="F0002" position="float"><label>Fig. 2</label><caption><p>Mortality rates per 1,000 person-years by cause group by year (2006&#x02013;2010).</p></caption><graphic xlink:href="GHA-7-25543-g002"/></fig></sec></sec><sec sec-type="discussion" id="S0004"><title>Discussion</title><p>Our results show that the major causes of death among adults in the study area from 2006 to 2010 remain CDs followed by NCDs. The main causes of CDs are malaria and TB, and for NCDs, stroke and digestive neoplasms. These are largely preventable and treatable diseases.</p><sec id="S0004-S20001"><title>Mortality rates</title><p>The findings indicate a decline of mortality from 9.8/1,000 py in 2006 to 6.6/1,000 py in 2010. According to the 2012 World Bank Report, the crude death rate for Ghana was 8/1,000 people in 2010. The death rate for the year 2010 is lower than that of the national rate, an indication of low mortality rates in the study area (<xref rid="CIT0024" ref-type="bibr">24</xref>). It must, however, be noted that the methods used for estimating the death rates are slightly different. Whereas, the World Bank used the midyear population as the denominator, this study used person-years contributed. Mortality rates are lower in the 15&#x02013;49 years age group and highest among the 65 years and older. This finding is similar to what was reported by Becher et al. that mortality rates were lowest in the 15&#x02013;49 years age group and highest in the 60 years and above in a malaria endemic area of West Africa.</p><p>Generally, mortality rates in males are higher than in females. The sex differential in mortality rates confirms what has been documented in other studies (<xref rid="CIT0025" ref-type="bibr">25</xref>). Obermeyer et al. also showed higher mortality in males than females for Ghana in their study, which covered 44 countries (<xref rid="CIT0026" ref-type="bibr">26</xref>). There are a number of explanations given for the sex difference in mortality that are based on biological, psychological, and social interpretations (<xref rid="CIT0027" ref-type="bibr">27</xref>).</p></sec><sec id="S0004-S20002"><title>Causes of deaths</title><p>The predominant cause of death is CDs with cause-specific rate of 3.29/1,000 py in the 5-year period. This reflects what pertains in the rural and less industrial settings in many developing countries, including Ghana (<xref rid="CIT0016" ref-type="bibr">16</xref>). The trends over the 5-year period in this study suggest that there has been a decrease in all causes of death by cause group even though the trend has not been progressive. Nevertheless, studies indicate that although the threats of communicable and poverty-related diseases (malaria, infant mortality, cholera, malnutrition) are still in existence (<xref rid="CIT0028" ref-type="bibr">28</xref>, <xref rid="CIT0029" ref-type="bibr">29</xref>) chronic disease prevalence was increasing in SSA countries such as Ghana, Nigeria, and South Africa.</p><p>Although, Group II causes which are NCDs are the leading causes of death worldwide, Group I causes which are CDs, maternal, neonatal, and nutritional causes are the leading causes of death in SSA (<xref rid="CIT0028" ref-type="bibr">28</xref>). Lozano et al. estimated that CDs, maternal, neonatal, and nutritional causes accounted for 76% of deaths in SSA in 2010 and 24.9% of deaths worldwide (<xref rid="CIT0005" ref-type="bibr">5</xref>). In estimating cause-specific mortality rates in SSA, Adjuik et al. found that in most of the countries, including Ghana, deaths were caused mainly by CDs (<xref rid="CIT0007" ref-type="bibr">7</xref>). The double burden of NCDs and CDs in developing countries like Ghana has been reported by other studies (<xref rid="CIT0030" ref-type="bibr">30</xref>, <xref rid="CIT0031" ref-type="bibr">31</xref>) and has a long-term impact on its public health which can lead to the collapse of the health system due to the further stretching of limited resources in terms of infrastructure and finance (<xref rid="CIT0028" ref-type="bibr">28</xref>).</p><p>This study found that mortality rates attributable to NCDs were higher in females than in males. However, BeLue et al. reported that men were more likely to develop NCDs as a result of lifestyle behaviors such as smoking and alcoholism for which those living in low socioeconomic settings are not excluded (<xref rid="CIT0028" ref-type="bibr">28</xref>). This can be attributed to nutritional transitions resulting in obesity or overweight, which has been found to be increasing in some rural areas (<xref rid="CIT0032" ref-type="bibr">32</xref>, <xref rid="CIT0033" ref-type="bibr">33</xref>).</p><p> Malaria in this study is the leading cause of death in the 15&#x02013;49 and the 65 years and above age groups among females. These findings were least expected since malaria is known to be more prevalent in children. However, the same concerns were raised in the findings of a systematic analysis of global malaria mortality from 1989 to 2010. It was found that 20% of malaria deaths in 2010 was contributed by adults aged 15&#x02013;49 years (<xref rid="CIT0034" ref-type="bibr">34</xref>).</p><p>Pulmonary TB was the leading cause of death among males and also among the top three causes of mortality among females. According to the WHO 2013 Global TB report, Africa is one of the regions, which currently is not on track in achieving the mortality and prevalence targets of 50% reduction by 2015 (<xref rid="CIT0035" ref-type="bibr">35</xref>). From the report, TB was the leading cause of deaths among men globally and remains among the top three killers of women; these were confirmed in the study being discussed.</p><p>The proportion of HIV/AIDS deaths in this study is very low and is not different from Ghana's HIV prevalence in 2009 (1.9%), which dropped further to 1.5%, respectively, in 2010 and 2011 (<xref rid="CIT0020" ref-type="bibr">20</xref>). It is possible that some of the HIV cases may be comorbid and diagnosed as TB by the model. This can be attributed to the fact that measuring TB mortality is very difficult among HIV-positive people even in cases where vital registration systems are complete (<xref rid="CIT0035" ref-type="bibr">35</xref>). According to the WHO 2013 Global TB report, causes of death by TB are usually not reliably recorded (<xref rid="CIT0035" ref-type="bibr">35</xref>).</p><p>With RTA as one of the emerging causes of death, the global status report on road safety showed that more than 90% of fatalities on the road occur in low- and middle-income countries where only 48% of the world's registered vehicles can be found (<xref rid="CIT0036" ref-type="bibr">36</xref>). According to the report, RTAs were among the 10 leading causes of death, ranking first and third, respectively, among the age groups 15&#x02013;29 and 30&#x02013;44. In this study, RTA is among the top three causes of death among males aged 15&#x02013;49. This finding is comparable to what was found by Ohene et al., and in the 2012 UNODC report that males contribute more to injury deaths than female (<xref rid="CIT0037" ref-type="bibr">37</xref>, <xref rid="CIT0038" ref-type="bibr">38</xref>).</p></sec><sec id="S0004-S20003"><title>Strengths of the study</title><p>This study covered the whole population in the two districts under study, and data collected at the household level closely following the death event. It is therefore representative of the two districts and can be extrapolated to the districts with similar ecological and demographic characteristics.</p></sec><sec id="S0004-S20004"><title>Study limitation</title><p>About 23% of deaths did not have VA completed because of inability to find a relative or individual who was with or knew the deceased to be interviewed. Additionally, 11.5% of the forms had very limited information or the description of the type of ailment or symptoms made it difficult for a cause to be assigned. The narrative part of VA and open-ended questions are excluded from the InterVA model, but these questions may be more appropriate in settings where there is poor knowledge of symptoms of certain diseases, and especially in cases where more local terms maybe relevant (<xref rid="CIT0022" ref-type="bibr">22</xref>).</p></sec></sec><sec id="S0005"><title>Conclusion</title><p>This work has demonstrated that VA can provide data for estimating causes of death in settings where the civil vital registration system is poor or nonexistent. The leading cause of death among the study population was CDs with malaria topping the list. The findings also indicate variations in the patterns of mortality and causes of death and have provided useful empirical information, which is instrumental in understanding disease burden, health planning, and prioritization of health interventions in resource-poor settings where access to timely and accurate data is scarce. To unravel and understand the sources of differential vulnerability in the distribution and patterns of mortality and causes of death among adult rural dwellers, additional research is needed.</p></sec>

Comparing causes of death between formal and informal neighborhoods in urban Africa: evidence from Ouagadougou Health and Demographic Surveillance System

<sec id="st1"><title>Background</title><p>The probable coexistence of two or more epidemiological profiles in urban Africa is poorly documented. In particular, very few studies have focused on the comparison of cause-specific mortality between two types of neighborhoods that characterize contemporary southern cities: formal neighborhoods, that is, structured or delineated settlements (planned estates) that have full access to public utilities (electricity and water services), and the informal neighborhoods, that is, spontaneous and unplanned peri-urban settlements where people live in slum-like conditions, often with little or no access to public utilities.</p></sec><sec id="st2"><title>Objective</title><p>To compare the causes of death between the formal and informal neighborhoods covered by the Ouagadougou Health and Demographic Surveillance Systems (HDSS).</p></sec><sec id="st3"><title>Design</title><p>The data used come from the INDEPTH pooled dataset which includes the contribution of Ouagadougou HDSS and are compiled for the INDEPTH Network Data repository. The data were collected between 2009 and 2011 using verbal autopsy (VA) questionnaires completed by four fieldworkers well trained in the conduction of VAs. The VA data were then interpreted using the InterVA-4 program (version 4.02) to arrive at the causes of death.</p></sec><sec id="st4"><title>Results</title><p>Communicable diseases are the leading cause of death among children (aged between 29 days and 14 years) in both formal and informal neighborhoods, contributing more than 75% to the mortality rate. Mortality rates from non-communicable diseases (NCDs) are very low before age 15 but are the leading causes from age 50, especially in formal neighborhoods. Mortality from injuries is very low, with no significant difference between the two neighborhoods.</p></sec><sec id="st5"><title>Conclusions</title><p>The fact that mortality from NCDs is higher among adults in formal neighborhoods seems consistent with the idea of a correlation between modern life and epidemiological transition. However, NCDs do affect informal neighborhoods as well. They consist mainly of cardiovascular diseases and neoplasms most of which are preventable and/or manageable through a change in lifestyle. A prevention program would certainly reduce the burden of these chronic diseases among adults and the elderly with a significant economic impact for families.</p></sec>

<contrib contrib-type="author"><name><surname>Soura</surname><given-names>Abdramane Bassiahi</given-names></name><xref ref-type="aff" rid="AF0001">1</xref><xref ref-type="corresp" rid="cor1">*</xref></contrib><contrib contrib-type="author"><name><surname>Lankoande</surname><given-names>Bruno</given-names></name><xref ref-type="aff" rid="AF0001">1</xref></contrib><contrib contrib-type="author"><name><surname>Millogo</surname><given-names>Roch</given-names></name><xref ref-type="aff" rid="AF0001">1</xref></contrib><contrib contrib-type="author"><name><surname>Bangha</surname><given-names>Martin</given-names></name><xref ref-type="aff" rid="AF0002">2</xref></contrib>

Global Health Action

<p>Demographic transition, defined as the decline in mortality and fertility resulting from socio-economic development, is usually followed by epidemiological transition. The latter is the evolution in which infectious and parasitic diseases, as leading causes of morbidity and death, are gradually replaced by chronic conditions. All industrialized countries have gone through this process that began since the 1800s (<xref rid="CIT0001" ref-type="bibr">1</xref>). The developing countries are currently at varying stages in this epidemiological transition (<xref rid="CIT0002" ref-type="bibr">2</xref>). In some regions, such as sub-Saharan Africa, the stages overlap, since the resurgence of some infectious diseases (meningitis, malaria &#x02026;) and HIV/AIDS co-exist with the increase of chronic diseases (<xref rid="CIT0002" ref-type="bibr">2</xref>). Sub-Saharan Africa is therefore facing a &#x02018;double burden&#x02019;, especially at adult ages (<xref rid="CIT0002" ref-type="bibr">2</xref>, <xref rid="CIT0003" ref-type="bibr">3</xref>).</p><p>Determinants of epidemiological transition are intimately associated with those of demographic transition and are generally attributed to &#x02018;modernization&#x02019;. In developing countries, the epidemiological transition is correlated with the urban transition. Indeed, urbanization, depending on its speed and intensity, may result in environmental change, which in turn can influence behaviors such as physical activity (<xref rid="CIT0004" ref-type="bibr">4</xref>). For instance, the mechanization of certain tasks (at home and at work) and the availability of intra-urban means of travelling most often results in some amount of physical inactivity. Furthermore, some work-related physical activities or tasks may lose their value because of e-administration. Moreover, urban transition is usually followed by nutrition transition, which is the transition from traditional (natural) nutritive diets toward western-style diets (particularly fast foods), rich in processed products. Indeed, with industrialization of food, globalization of trade, advertising and marketing of agro-food industries, urban dwellers are increasingly exposed to massive imports of manufactured food products (<xref rid="CIT0005" ref-type="bibr">5</xref>). The nutrition transition in turn leads to a health transition marked by the emergence of overweight and obesity, hypertension, diabetes, hypercholesterolemia, and increased mortality from cardiovascular diseases and some cancers (<xref rid="CIT0005" ref-type="bibr">5</xref>).</p><p>Several studies in developing countries have shown a higher risk of chronic diseases among urban adults compared to their rural counterparts (<xref rid="CIT0006" ref-type="bibr">6</xref>&#x02013;<xref rid="CIT0012" ref-type="bibr">12</xref>). Comparative studies on causes of death are rare due to the absence of population-specific surveys on the topic. In most cases, data on causes of death are from hospital statistics, which are severely affected by selection bias given that most deaths in developing countries occur outside the hospital system and clinical autopsies are almost never performed. One of the few leading works in the field is the study by Walker et al. that compares stroke mortality in rural and urban Tanzania using data from three demographic surveillance systems (<xref rid="CIT0013" ref-type="bibr">13</xref>). They found that stroke mortality, especially among urban women, was significantly higher than in rural areas because of the advanced epidemiological transition in the city (<xref rid="CIT0013" ref-type="bibr">13</xref>).</p><p>In Burkina Faso, as in most sub-Saharan African countries, evidence on the epidemiological transition can be pieced together only for selected parts of the country where Health and Demographic Surveillance Systems (HDSS) exist. Four of the five currently operational HDSSs in the country are located in areas that are predominantly rural. Studies on cause-specific mortality in urban areas, and more importantly within the same city, are thus virtually not found in Burkina Faso as in most developing countries. The probable coexistence of two or more epidemiological profiles in urban areas is thus poorly documented. This paper is a contribution to knowledge in this area by taking advantage of data from the Ouagadougou HDSS to compare the causes of death between two types of neighborhoods that characterize contemporary southern cities: formal neighborhoods and informal neighborhoods (or settlements). Formal neighborhoods are structured or delineated settlements (planned estates), with full access to public utilities (electricity and water services), while informal neighborhoods are mainly spontaneous settlements that sprang up in these cities as a result of rapid and &#x02018;uncontrolled&#x02019; urbanization. Their inhabitants often live in abject poverty with no means to secure adequate housing in formal neighborhoods.</p><p>This paper on cause-specific mortality is mainly descriptive work in which we expect more non-communicable diseases (NCDs), that is, diseases that cannot be passed from person to person as adult causes of death in formal neighborhoods compared to the informal ones. In the latter, we expect a mixed epidemiological profile for adults; NCDs and communicable diseases (CDs), that is, transmitted through direct contact with an infected individual or indirectly through a vector, should be of similar magnitude. Indeed, social and dietary changes that accompany urbanization and expose people to NCDs are much more noticeable in formal neighborhoods. In informal neighborhoods, poor living conditions place the population at a high risk of CDs but at the same time, people, because of their poverty, may quickly adopt &#x02018;deviant&#x02019; behaviors such as excessive drinking and smoking, two NCD risk factors (<xref rid="CIT0014" ref-type="bibr">14</xref>). Urban poor may also consume more energetically dense foods because they seem inexpensive but which, combined with physical inactivity, expose them more to NCDs (<xref rid="CIT0015" ref-type="bibr">15</xref>).</p><p>In the case of children, one will normally not expect any notable difference in the causes of death according to the type of neighborhood. Although all age groups can be affected by NCDs, these diseases are often associated with older age groups. In developing countries, children die more from CDs whatever the place of residence (<xref rid="CIT0016" ref-type="bibr">16</xref>). Even urban Ouagadougou is not different in this regard and so the expectation in this study is for mortality in children to be dominated by CDs regardless of the type of neighborhood.</p><sec id="S0002"><title>Data and methods</title><p>This work contributes to enrich the scientific debate on intra-urban differences in the epidemiological profile through an analysis of causes of death. The latter are compared between the formal and informal neighborhoods covered by the Ouagadougou HDSS.</p><sec id="S0002-S20001"><title>Study population</title><p>Data used come from the INDEPTH pooled dataset which includes the contribution of Ouagadougou HDSS and compiled for the INDEPTH Network Data repository (<xref rid="CIT0017" ref-type="bibr">17</xref>). The Ouagadougou HDSS is a platform for health research and interventions established in 2008 covering five neighborhoods of Ouagadougou (<xref rid="CIT0018" ref-type="bibr">18</xref>). Two of these neighborhoods (Kilwin and Tanghin) are formal neighborhoods with full access to public services, while the other three (Nonghin, Polesgo, and Nioko 2) are spontaneous (such as slums) without access to such services. People living in informal areas are poorer on average, less educated, and born in rural areas in comparison with people living in formal settlements (<xref rid="CIT0019" ref-type="bibr">19</xref>), which highlights the importance of rural outmigration to the growth of informal urban settlements. Households in the informal settlements are usually small, made up of single men or young nuclear families in search of affordable housing (<xref rid="CIT0019" ref-type="bibr">19</xref>).</p><p>After an initial census conducted between October 2008 and March 2009 in the five neighborhoods, fieldworkers make regular household visits for update rounds (with an average periodicity of 7 months), registering vital events (births and deaths, marriages, and migrations). As at November 2012, the population under surveillance by the Ouagadougou HDSS totaled 86,071 residents (defined as individuals present in the zone for at least 6 months). In case of death, a verbal autopsy (VA) questionnaire is completed with the next of kin to determine the circumstances that led to the death, including history of the illness and the specific symptoms that preceded death. It should be noted that although the data used come from the INDEPTH pooled dataset, not all INDEPTH members used the INDEPTH standard VA instrument. In 2012, a group of experts under the auspices of WHO reviewed the existing VA instruments in the world and proceeded to their simplification and their standardization to make the results comparable (<xref rid="CIT0020" ref-type="bibr">20</xref>).</p><p>A revised list of causes of death has been established by grouping all ICD-10 causes of death into 62 broad categories. These categories were chosen on the basis of their public health relevance and their potential for ascertainment from VA. A total of 245 indicators (questions) were included in the revised VA instrument. A matrix of these indicators is the input file for the InterVA-4 model used for processing VA data to produce CoD for analysis in this special issue; all the contributing HDSSs transformed their CoD data into this matrix for use in the version 4.02 of InterVA-4 (<xref rid="CIT0021" ref-type="bibr">21</xref>). This model applies Bayesian probabilistic methods to VA data and arrives at possible causes of death (<xref rid="CIT0021" ref-type="bibr">21</xref>). It generates a maximum of three likely causes of death per case with their associated partial likelihoods (between 0 and 1). For some cases, the input data are insufficient for InterVA-4 to generate any cause of death and such cases are classified by InterVA-4 into the &#x02018;indeterminate&#x02019; cause of death. For each case where the sum of the partial likelihoods does not total 1, the difference between their sum and 1 is assigned to the &#x02018;indeterminate&#x02019; cause. For this paper, all identified causes of death will be considered proportionate to their partial likelihoods in the calculation of the number of deaths from each cause.</p><p>In this INDEPTH pooled dataset, data from Ouagadougou HDSS cover the period 2009&#x02013;2011 and include 1,032 deaths recorded across 221,178 person-years. Of the 1,032 recorded deaths, 870 VAs were completed. These VA data are used to compare formal and informal neighborhoods in terms of causes of death. In the corresponding multisite papers presented in this special issue, the Ouagadougou results are presented as one site.</p></sec><sec id="S0002-S20002"><title>Indicators and methods</title><p>This study examined mortality rates, proportion of deaths due to each cause, and the contribution of each cause to the all-cause mortality rate. Mortality rates are obtained by dividing the number of deaths by the number of person-years. Our estimates will not provide confidence intervals since the HDSS covers an entire non-sampled population. Due to small number of deaths involved, the mortality rates are calculated only for major groups of causes (CDs, NCDs, maternal and neonatal causes, injuries, and unspecified causes). These groups are predefined in the InterVA-4 model (version 4.02) used. CDs include diarrheal diseases, HIV/AIDS, non-obstetric sepsis, malaria, meningitis and encephalitis, respiratory infections, TB, and other infectious diseases. The most common NCDs are anemia, asthma, cardiovascular diseases, neoplasms, diabetes, renal failure, acute abdomen, epilepsy, and severe malnutrition. Maternal and neonatal mortality includes by implication pregnancy-/birth-related causes (pregnancy-induced hypertension, pregnancy-related sepsis, obstetric hemorrhage) and neonatal causes (prematurity, birth asphyxia, neonatal pneumonia, neonatal sepsis, and congenital malformation).</p><p>To better portray the cause-specific mortality by age, we used the seven age groups predefined in InterVA-4 model (version 4.02), which correspond theoretically to different leading causes of death. Thus, children were grouped into four categories with different levels of exposure to various diseases: neonates (less than 28 days), infants (29 days&#x02013;11 months), children between 1 and 5, and those between 5 and 15. Among adults, the elderly (65 and over) have been distinguished from people aged 50&#x02013;64 and from those aged 15&#x02013;49.</p><p> <xref ref-type="table" rid="T0001">Table 1</xref> presents the person-years distribution by age group, sex, and neighborhood, although the small number of deaths here does not allow us to perform mortality analysis by sex. For each sex, formal and informal neighborhoods have close distributions. Regardless of gender and type of neighborhood, people aged 15&#x02013;49 are the most represented, accounting for more than 50%, followed by those aged 5&#x02013;14 representing just over 1 in 5. The proportion of people aged 15&#x02013;49 is slightly higher in formal neighborhoods. There are relatively more children (1&#x02013;11 months and 1&#x02013;4 years) in informal neighborhoods while older people (50 years and older) are slightly more in formal neighborhoods. To control for this slight difference in age structure between formal and informal neighborhoods, we provide standardized mortality rates next to the crude mortality rate (all ages) for each type of neighborhood. For this purpose, we have used the structure of the two types of neighborhoods combined as the standard population.</p><table-wrap id="T0001" position="float"><label>Table 1</label><caption><p>Distribution (%) of person-years by age group, sex, and neighborhood, 2009&#x02013;2011</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1"/><th align="center" colspan="2" rowspan="1">Formal neighborhood</th><th align="center" colspan="2" rowspan="1">Informal neighborhood</th><th align="center" colspan="2" rowspan="1">Total</th></tr><tr><th align="left" rowspan="1" colspan="1"/><th colspan="2" rowspan="1"> <hr/> </th><th colspan="2" rowspan="1"> <hr/> </th><th colspan="2" rowspan="1"> <hr/> </th></tr><tr><th align="left" rowspan="1" colspan="1">Age group</th><th align="center" rowspan="1" colspan="1">Male</th><th align="center" rowspan="1" colspan="1">Female</th><th align="center" rowspan="1" colspan="1">Male</th><th align="center" rowspan="1" colspan="1">Female</th><th align="center" rowspan="1" colspan="1">Male</th><th align="center" rowspan="1" colspan="1">Female</th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">0&#x02013;28 days</td><td align="center" rowspan="1" colspan="1">0.2%</td><td align="center" rowspan="1" colspan="1">0.1%</td><td align="center" rowspan="1" colspan="1">0.2%</td><td align="center" rowspan="1" colspan="1">0.3%</td><td align="center" rowspan="1" colspan="1">0.2%</td><td align="center" rowspan="1" colspan="1">0.2%</td></tr><tr><td align="left" rowspan="1" colspan="1">29 days&#x02013;11 months</td><td align="center" rowspan="1" colspan="1">2.1%</td><td align="center" rowspan="1" colspan="1">2.1%</td><td align="center" rowspan="1" colspan="1">3.7%</td><td align="center" rowspan="1" colspan="1">3.9%</td><td align="center" rowspan="1" colspan="1">2.9%</td><td align="center" rowspan="1" colspan="1">3.0%</td></tr><tr><td align="left" rowspan="1" colspan="1">1&#x02013;4 years</td><td align="center" rowspan="1" colspan="1">10.0%</td><td align="center" rowspan="1" colspan="1">9.8%</td><td align="center" rowspan="1" colspan="1">15.7%</td><td align="center" rowspan="1" colspan="1">15.6%</td><td align="center" rowspan="1" colspan="1">12.7%</td><td align="center" rowspan="1" colspan="1">12.5%</td></tr><tr><td align="left" rowspan="1" colspan="1">5&#x02013;14 years</td><td align="center" rowspan="1" colspan="1">22.4%</td><td align="center" rowspan="1" colspan="1">23.7%</td><td align="center" rowspan="1" colspan="1">22.2%</td><td align="center" rowspan="1" colspan="1">24.4%</td><td align="center" rowspan="1" colspan="1">22.3%</td><td align="center" rowspan="1" colspan="1">24.0%</td></tr><tr><td align="left" rowspan="1" colspan="1">15&#x02013;49 years</td><td align="center" rowspan="1" colspan="1">56.0%</td><td align="center" rowspan="1" colspan="1">56.0%</td><td align="center" rowspan="1" colspan="1">53.2%</td><td align="center" rowspan="1" colspan="1">50.5%</td><td align="center" rowspan="1" colspan="1">54.6%</td><td align="center" rowspan="1" colspan="1">53.4%</td></tr><tr><td align="left" rowspan="1" colspan="1">50&#x02013;64 years</td><td align="center" rowspan="1" colspan="1">7.0%</td><td align="center" rowspan="1" colspan="1">5.9%</td><td align="center" rowspan="1" colspan="1">4.0%</td><td align="center" rowspan="1" colspan="1">3.6%</td><td align="center" rowspan="1" colspan="1">5.6%</td><td align="center" rowspan="1" colspan="1">4.8%</td></tr><tr><td align="left" rowspan="1" colspan="1">65+ years</td><td align="center" rowspan="1" colspan="1">2.3%</td><td align="center" rowspan="1" colspan="1">2.4%</td><td align="center" rowspan="1" colspan="1">1.0%</td><td align="center" rowspan="1" colspan="1">1.7%</td><td align="center" rowspan="1" colspan="1">1.7%</td><td align="center" rowspan="1" colspan="1">2.1%</td></tr><tr><td align="left" rowspan="1" colspan="1">Person-years</td><td align="center" rowspan="1" colspan="1">57500.5</td><td align="center" rowspan="1" colspan="1">57932.8</td><td align="center" rowspan="1" colspan="1">53492.4</td><td align="center" rowspan="1" colspan="1">52252.4</td><td align="center" rowspan="1" colspan="1">110992.9</td><td align="center" rowspan="1" colspan="1">110185.2</td></tr></tbody></table></table-wrap></sec></sec><sec sec-type="results" id="S0003"><title>Results</title><sec id="S0003-S20001"><title>All-age mortality</title><p>In the study area, the overall crude mortality rate is estimated at 4.7 per 1,000 person-years. It is 4.2 per 1,000 person-years in the formal neighborhoods and 5.2 per 1,000 person-years in the informal ones. Despite the standardization to account for the difference in age structure, the excess mortality remains in informal neighborhoods with a mortality rate of 4.6 per 1,000 persons-years against 3.9 for formal neighborhoods. All ages combined, CDs are the leading causes of death in the Ouagadougou HDSS. Indeed, the mortality rate from CDs is estimated at 1.71 per 1,000 person-years against 1.22 for NCDs, 0.19 for injuries (see <xref ref-type="table" rid="T0003">Table 3</xref>). Mortality from CDs is more prevalent in informal neighborhoods with an age-adjusted rate of 2.11 per 1,000 person-years against 1.22 in the formal ones. In the latter, the risk of death associated with NCDs is close to that related to CDs (1.32 per 1,000 person-years against 1.22 per 1,000 person-years as age-adjusted rates). There is little difference between the two types of neighborhoods in terms of injuries-related mortality risk (age-adjusted rate of 0.22 per 1,000 person-years for informal neighborhoods against 0.16 per 1,000 person-years for the formal ones).</p><p>In terms of proportion (<xref ref-type="fig" rid="F0001">Fig. 1</xref>), CDs are responsible for about 27% of all deaths that occurred in formal neighborhoods and NCDs responsible for 37% in the same neighborhoods. The situation is reversed in informal neighborhoods where CDs account for more than 45% of all deaths against about 17% for NCDs.</p><fig id="F0001" position="float"><label>Fig. 1</label><caption><p>Percentage of people dying from each group of causes according to the type of neighborhood (Percentage calculated on 485 deaths in formal neighborhoods and 547 deaths in informal neighborhoods).</p></caption><graphic xlink:href="GHA-7-25523-g001"/></fig><p>Among the CDs, malaria is the top cause of death as in many sub-Saharan African countries. It was responsible for about 11.7% of all-ages deaths in the study site, 7.1% in formal neighborhoods and 15.8% in informal neighborhoods (<xref ref-type="table" rid="T0002">Table 2</xref>). Second to malaria, in order of importance are acute respiratory infections, HIV/AIDS, and diarrheal diseases with 10.0, 5.8 and 3.2% of deaths, respectively. NCDs consist mainly of cardiovascular diseases which accounted for 17.7% of deaths in formal neighborhoods and 5.7% in informal neighborhoods. Mortality from neoplasms, taken separately, is also more prevalent in formal neighborhoods (11.9% of deaths) than in the informal ones (4.7% of deaths) while mortality from asthma, diabetes, and anemia is low, accounting for less than 1%.</p><table-wrap id="T0002" position="float"><label>Table 2</label><caption><p>Cause-specific mortality fraction (%) by neighborhood, 2009&#x02013;2011</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1">Cause of death</th><th align="center" rowspan="1" colspan="1">Formal (<italic>n</italic>=485) (%)</th><th align="center" rowspan="1" colspan="1">Informal (<italic>n</italic>=547) (%)</th><th align="center" rowspan="1" colspan="1">Both (<italic>n</italic>=1,032) (%)</th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1"> <bold>Communicable diseases (CDs)</bold> </td><td align="center" rowspan="1" colspan="1"> <bold>26.6</bold> </td><td align="center" rowspan="1" colspan="1"> <bold>45.5</bold> </td><td align="center" rowspan="1" colspan="1"> <bold>36.6</bold> </td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Sepsis (non-obstetric)</td><td align="center" rowspan="1" colspan="1">0.8</td><td align="center" rowspan="1" colspan="1">0.4</td><td align="center" rowspan="1" colspan="1">0.6</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Acute resp infect incl. pneumonia</td><td align="center" rowspan="1" colspan="1">8.0</td><td align="center" rowspan="1" colspan="1">11.7</td><td align="center" rowspan="1" colspan="1">10.0</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;HIV-/AIDS-related death</td><td align="center" rowspan="1" colspan="1">4.8</td><td align="center" rowspan="1" colspan="1">6.7</td><td align="center" rowspan="1" colspan="1">5.8</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Diarrheal diseases</td><td align="center" rowspan="1" colspan="1">1.2</td><td align="center" rowspan="1" colspan="1">4.9</td><td align="center" rowspan="1" colspan="1">3.2</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Malaria</td><td align="center" rowspan="1" colspan="1">7.1</td><td align="center" rowspan="1" colspan="1">15.8</td><td align="center" rowspan="1" colspan="1">11.7</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Meningitis and encephalitis</td><td align="center" rowspan="1" colspan="1">2.1</td><td align="center" rowspan="1" colspan="1">3.1</td><td align="center" rowspan="1" colspan="1">2.6</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Pulmonary tuberculosis</td><td align="center" rowspan="1" colspan="1">2.4</td><td align="center" rowspan="1" colspan="1">2.1</td><td align="center" rowspan="1" colspan="1">2.2</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Other and unspecified infect dis</td><td align="center" rowspan="1" colspan="1">0.2</td><td align="center" rowspan="1" colspan="1">0.8</td><td align="center" rowspan="1" colspan="1">0.5</td></tr><tr><td align="left" rowspan="1" colspan="1"> <bold>Non-communicable diseases (NCDs)</bold> </td><td align="center" rowspan="1" colspan="1"> <bold>36.7</bold> </td><td align="center" rowspan="1" colspan="1"> <bold>16.9</bold> </td><td align="center" rowspan="1" colspan="1"> <bold>26.2</bold> </td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Neoplasms</td><td align="center" rowspan="1" colspan="1">11.9</td><td align="center" rowspan="1" colspan="1">4.7</td><td align="center" rowspan="1" colspan="1">8.1</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Severe anemia</td><td align="center" rowspan="1" colspan="1">0.3</td><td align="center" rowspan="1" colspan="1">0.3</td><td align="center" rowspan="1" colspan="1">0.3</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Severe malnutrition</td><td align="center" rowspan="1" colspan="1">1.8</td><td align="center" rowspan="1" colspan="1">3.1</td><td align="center" rowspan="1" colspan="1">2.5</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Diabetes</td><td align="center" rowspan="1" colspan="1">0.8</td><td align="center" rowspan="1" colspan="1">0.2</td><td align="center" rowspan="1" colspan="1">0.5</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Cardiovascular diseases</td><td align="center" rowspan="1" colspan="1">17.7</td><td align="center" rowspan="1" colspan="1">5.7</td><td align="center" rowspan="1" colspan="1">11.3</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Asthma</td><td align="center" rowspan="1" colspan="1">0.3</td><td align="center" rowspan="1" colspan="1">0.2</td><td align="center" rowspan="1" colspan="1">0.2</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Acute abdomen</td><td align="center" rowspan="1" colspan="1">3.1</td><td align="center" rowspan="1" colspan="1">2.3</td><td align="center" rowspan="1" colspan="1">2.7</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Renal failure</td><td align="center" rowspan="1" colspan="1">0.0</td><td align="center" rowspan="1" colspan="1">0.1</td><td align="center" rowspan="1" colspan="1">0.0</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Epilepsy</td><td align="center" rowspan="1" colspan="1">0.4</td><td align="center" rowspan="1" colspan="1">0.3</td><td align="center" rowspan="1" colspan="1">0.4</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Other and unspecified NCDs</td><td align="center" rowspan="1" colspan="1">0.4</td><td align="center" rowspan="1" colspan="1">0.0</td><td align="center" rowspan="1" colspan="1">0.2</td></tr><tr><td align="left" rowspan="1" colspan="1"> <bold>Maternal/neonatal</bold> </td><td align="center" rowspan="1" colspan="1"> <bold>2.9</bold> </td><td align="center" rowspan="1" colspan="1"> <bold>7.9</bold> </td><td align="center" rowspan="1" colspan="1"> <bold>5.6</bold> </td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Maternal causes</td><td align="center" rowspan="1" colspan="1">0.8</td><td align="center" rowspan="1" colspan="1">0.4</td><td align="center" rowspan="1" colspan="1">0.6</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Neonatal causes</td><td align="center" rowspan="1" colspan="1">2.1</td><td align="center" rowspan="1" colspan="1">7.5</td><td align="center" rowspan="1" colspan="1">5.0</td></tr><tr><td align="left" rowspan="1" colspan="1"> <bold>Injuries</bold> </td><td align="center" rowspan="1" colspan="1"> <bold>4.3</bold> </td><td align="center" rowspan="1" colspan="1"> <bold>3.9</bold> </td><td align="center" rowspan="1" colspan="1"> <bold>4.1</bold> </td></tr><tr><td align="left" rowspan="1" colspan="1"> <bold>Indeterminate</bold> </td><td align="center" rowspan="1" colspan="1"> <bold>11.1</bold> </td><td align="center" rowspan="1" colspan="1"> <bold>12.5</bold> </td><td align="center" rowspan="1" colspan="1"> <bold>11.8</bold> </td></tr><tr><td align="left" rowspan="1" colspan="1"> <bold>VA not done</bold> </td><td align="center" rowspan="1" colspan="1"> <bold>18.4</bold> </td><td align="center" rowspan="1" colspan="1"> <bold>13.3</bold> </td><td align="center" rowspan="1" colspan="1"> <bold>15.7</bold> </td></tr></tbody></table></table-wrap></sec><sec id="S0003-S20002"><title> Age-specific mortality</title><p>The risk of dying from CDs for children is between 1 and 5 times higher in informal neighborhoods than in the formal ones (<xref ref-type="table" rid="T0003">Table 3</xref>). The child mortality rate from this cause is estimated at 6.18 per 1,000 person-years in the informal neighborhoods against 3.09 per 1,000 person-years in the formal ones. For infants aged between 29 days and 11 months, an excess mortality from CDs is clearly visible in informal neighborhoods with a rate of 17.08 per 1,000 person-years against 9.17 per 1,000 person-years in formal neighborhoods. For older age groups (children aged 5&#x02013;14, adults aged 15&#x02013;49, those aged 50&#x02013;64, or the elderly, i.e., aged 65+), mortality rates due to CDs in formal and informal neighborhoods are close though relatively higher in informal neighborhoods (<xref ref-type="table" rid="T0003">Table 3</xref>).</p><table-wrap id="T0003" position="float"><label>Table 3</label><caption><p>Cause-specific mortality rates (per 1,000 person-years) by neighborhood and age group, 2009&#x02013;2011</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1"/><th align="center" colspan="3" rowspan="1">CDs</th><th align="center" colspan="3" rowspan="1">NCDs</th></tr><tr><th align="left" rowspan="1" colspan="1"/><th colspan="3" rowspan="1"> <hr/> </th><th colspan="3" rowspan="1"> <hr/> </th></tr><tr><th align="left" rowspan="1" colspan="1"/><th align="center" rowspan="1" colspan="1">Both</th><th align="center" rowspan="1" colspan="1">Formal</th><th align="center" rowspan="1" colspan="1">Informal</th><th align="center" rowspan="1" colspan="1">Both</th><th align="center" rowspan="1" colspan="1">Formal</th><th align="center" rowspan="1" colspan="1">Informal</th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">Neonatal<xref ref-type="table-fn" rid="TF0001">a</xref> </td><td align="center" rowspan="1" colspan="1">6.92</td><td align="center" rowspan="1" colspan="1">4.07</td><td align="center" rowspan="1" colspan="1">8.54</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">29 days&#x02013;11 months</td><td align="center" rowspan="1" colspan="1">14.12</td><td align="center" rowspan="1" colspan="1">9.17</td><td align="center" rowspan="1" colspan="1">17.08</td><td align="center" rowspan="1" colspan="1">1.29</td><td align="center" rowspan="1" colspan="1">0.31</td><td align="center" rowspan="1" colspan="1">1.88</td></tr><tr><td align="left" rowspan="1" colspan="1">1&#x02013;4 years</td><td align="center" rowspan="1" colspan="1">4.92</td><td align="center" rowspan="1" colspan="1">3.09</td><td align="center" rowspan="1" colspan="1">6.18</td><td align="center" rowspan="1" colspan="1">0.93</td><td align="center" rowspan="1" colspan="1">0.44</td><td align="center" rowspan="1" colspan="1">1.26</td></tr><tr><td align="left" rowspan="1" colspan="1">5&#x02013;14 years</td><td align="center" rowspan="1" colspan="1">0.70</td><td align="center" rowspan="1" colspan="1">0.51</td><td align="center" rowspan="1" colspan="1">0.91</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">0.08</td><td align="center" rowspan="1" colspan="1">0.13</td></tr><tr><td align="left" rowspan="1" colspan="1">15&#x02013;49 years</td><td align="center" rowspan="1" colspan="1">0.47</td><td align="center" rowspan="1" colspan="1">0.44</td><td align="center" rowspan="1" colspan="1">0.50</td><td align="center" rowspan="1" colspan="1">0.63</td><td align="center" rowspan="1" colspan="1">0.68</td><td align="center" rowspan="1" colspan="1">0.57</td></tr><tr><td align="left" rowspan="1" colspan="1">50&#x02013;64 years</td><td align="center" rowspan="1" colspan="1">2.63</td><td align="center" rowspan="1" colspan="1">2.06</td><td align="center" rowspan="1" colspan="1">3.69</td><td align="center" rowspan="1" colspan="1">5.87</td><td align="center" rowspan="1" colspan="1">7.50</td><td align="center" rowspan="1" colspan="1">2.84</td></tr><tr><td align="left" rowspan="1" colspan="1">65+ years</td><td align="center" rowspan="1" colspan="1">5.79</td><td align="center" rowspan="1" colspan="1">5.02</td><td align="center" rowspan="1" colspan="1">7.27</td><td align="center" rowspan="1" colspan="1">21.29</td><td align="center" rowspan="1" colspan="1">25.76</td><td align="center" rowspan="1" colspan="1">12.70</td></tr><tr><td align="left" rowspan="1" colspan="1"> <bold>All ages crude rate</bold> </td><td align="center" rowspan="1" colspan="1"> <bold>1.71</bold> </td><td align="center" rowspan="1" colspan="1"> <bold>1.12</bold> </td><td align="center" rowspan="1" colspan="1"> <bold>2.35</bold> </td><td align="center" rowspan="1" colspan="1"> <bold>1.22</bold> </td><td align="center" rowspan="1" colspan="1"> <bold>1.54</bold> </td><td align="center" rowspan="1" colspan="1"> <bold>0.87</bold> </td></tr><tr><td align="left" rowspan="1" colspan="1"> <bold>Age-adjusted rate</bold> </td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"> <bold>1.22</bold> </td><td align="center" rowspan="1" colspan="1"> <bold>2.11</bold> </td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"> <bold>1.32</bold> </td><td align="center" rowspan="1" colspan="1"> <bold>0.94</bold> </td></tr><tr><td align="left" rowspan="1" colspan="1"/><td colspan="3" align="center" rowspan="1">Maternal/neonatal</td><td colspan="3" align="center" rowspan="1">Injuries</td></tr><tr><td align="left" rowspan="1" colspan="1"/><td colspan="3" rowspan="1"> <hr/> </td><td colspan="3" rowspan="1"> <hr/> </td></tr><tr><td align="left" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">Both</td><td align="center" rowspan="1" colspan="1">Formal</td><td align="center" rowspan="1" colspan="1">Informal</td><td align="center" rowspan="1" colspan="1">Both</td><td align="center" rowspan="1" colspan="1">Formal</td><td align="center" rowspan="1" colspan="1">Informal</td></tr><tr><th colspan="7" align="left" rowspan="1"><hr/></th></tr><tr><td align="left" rowspan="1" colspan="1">Neonatal<xref ref-type="table-fn" rid="TF0001">a</xref> </td><td align="center" rowspan="1" colspan="1">91.18</td><td align="center" rowspan="1" colspan="1">55.37</td><td align="center" rowspan="1" colspan="1">111.61</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td></tr><tr><td align="left" rowspan="1" colspan="1">29 days&#x02013;11 months</td><td align="center" rowspan="1" colspan="1">0.64</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">1.02</td><td align="center" rowspan="1" colspan="1">0.15</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.24</td></tr><tr><td align="left" rowspan="1" colspan="1">1&#x02013;4 years</td><td align="center" rowspan="1" colspan="1">0.03</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.11</td><td align="center" rowspan="1" colspan="1">0.09</td><td align="center" rowspan="1" colspan="1">0.12</td></tr><tr><td align="left" rowspan="1" colspan="1">5&#x02013;14 years</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.01</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.13</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.22</td></tr><tr><td align="left" rowspan="1" colspan="1">15&#x02013;49 years</td><td align="center" rowspan="1" colspan="1">0.05</td><td align="center" rowspan="1" colspan="1">0.06</td><td align="center" rowspan="1" colspan="1">0.04</td><td align="center" rowspan="1" colspan="1">0.14</td><td align="center" rowspan="1" colspan="1">0.13</td><td align="center" rowspan="1" colspan="1">0.15</td></tr><tr><td align="left" rowspan="1" colspan="1">50&#x02013;64 years</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.57</td><td align="center" rowspan="1" colspan="1">0.67</td><td align="center" rowspan="1" colspan="1">0.40</td></tr><tr><td align="left" rowspan="1" colspan="1">65+ years</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">0.00</td><td align="center" rowspan="1" colspan="1">1.99</td><td align="center" rowspan="1" colspan="1">1.94</td><td align="center" rowspan="1" colspan="1">2.10</td></tr><tr><td align="left" rowspan="1" colspan="1"> <bold>All ages crude rate</bold> </td><td align="center" rowspan="1" colspan="1"> <bold>na</bold> </td><td align="center" rowspan="1" colspan="1"> <bold>na</bold> </td><td align="center" rowspan="1" colspan="1"> <bold>na</bold> </td><td align="center" rowspan="1" colspan="1"> <bold>0.19</bold> </td><td align="center" rowspan="1" colspan="1"> <bold>0.18</bold> </td><td align="center" rowspan="1" colspan="1"> <bold>0.20</bold> </td></tr><tr><td align="left" rowspan="1" colspan="1"> <bold>Age-adjusted rate</bold> </td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"> <bold>na</bold> </td><td align="center" rowspan="1" colspan="1"> <bold>na</bold> </td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1"> <bold>0.16</bold> </td><td align="center" rowspan="1" colspan="1"> <bold>0.22</bold> </td></tr></tbody></table><table-wrap-foot><fn><p>na=non-applicable. Not calculated because all age groups or all sexes are not exposed to the risk of maternal mortality. At the same time, all age groups are not exposed to a risk of death from neonatal causes.</p></fn><fn id="TF0001"><label>a</label><p>Rates in the Table are calculated per 1,000 person-years and hence may appear very high for neonatal mortality since each neonate cannot contribute more than one-twelfth of a person-year each.</p></fn></table-wrap-foot></table-wrap><p>Mortality rates from NCDs are relatively low before age 50, which is logical given that these diseases are strongly associated with age. Beyond this age, the difference between formal and informal neighborhoods is large. The mortality rate is at least two times higher in formal neighborhoods (<xref ref-type="table" rid="T0003">Table 3</xref>). For instance, it is 25.76 per 1,000 person-years among people aged 65 and over against 12.70 per 1,000 person-years for their counterparts in informal neighborhoods. Deaths from injuries are very rare and do not suggest any significant difference between the two types of neighborhoods (<xref ref-type="table" rid="T0003">Table 3</xref>). Neonatal causes are two times higher in informal neighborhoods compared to the formal ones (<xref ref-type="table" rid="T0003">Table 3</xref>).</p><p> <xref ref-type="fig" rid="F0002">Figure 2</xref> provides a better illustration of the contrast between CDs and NCDs in terms of leading cause of death by age group and neighborhood. As can be observed, before age 15, the proportion of deaths attributable to CDs is higher than that of NCD-related deaths whatever the type of neighborhood. The only exception is for the neonates where the two groups of causes have quite similar proportions. Among adults (aged 15 and over), two distinct situations are noticeable depending on the type of neighborhood. In formal neighborhoods, the proportion of deaths due to NCDs is far higher than that of deaths from CDs while in informal neighborhoods, the situation is mixed. Indeed, for informal neighborhoods, the proportion of NCDs and CDs are close before the age 65, but from 65 years, NCDs become more pronounced compared to CDs.</p><fig id="F0002" position="float"><label>Fig. 2</label><caption><p>Mortality fraction from communicable diseases or non-communicable diseases by age group and type of neighborhood.</p></caption><graphic xlink:href="GHA-7-25523-g002"/></fig><p>This analysis of mortality fractions was complemented by calculating the contribution of each group of causes (among known causes) to the total risk of death (<xref ref-type="table" rid="T0004">Table 4</xref>). The results confirm that CDs are the leading causes of death among children aged 29 days to 14 years, regardless of the type of neighborhood. For instance, among the known causes, CDs are responsible for 87.2% of the risk of death for infant aged 29 days&#x02013;11 months. This proportion is 84.5% in informal neighborhoods and 96.7% in the formal ones. Among children 1&#x02013;4 years, this proportion is estimated at 82.2% and does not vary much according to the type of neighborhood; for those aged 5&#x02013;14, it is 79.4% in formal neighborhoods and 71.8% in the informal ones.</p><table-wrap id="T0004" position="float"><label>Table 4</label><caption><p>Contributions (%) of the grouped causes to the risk of death by neighborhood and age group, 2009&#x02013;2011</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1">Group of causes</th><th align="center" rowspan="1" colspan="1">Type of neighborhood</th><th align="center" rowspan="1" colspan="1">Neonatal</th><th align="center" rowspan="1" colspan="1">29 days&#x02013;11 months</th><th align="center" rowspan="1" colspan="1">1&#x02013;4 years</th><th align="center" rowspan="1" colspan="1">5&#x02013;14 years</th><th align="center" rowspan="1" colspan="1">15&#x02013;49 years</th><th align="center" rowspan="1" colspan="1">50&#x02013;64 years</th><th align="center" rowspan="1" colspan="1">65+ years</th><th align="center" rowspan="1" colspan="1">Under 15 years</th><th align="center" rowspan="1" colspan="1">15+ years</th><th align="center" rowspan="1" colspan="1">All ages</th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">Both</td><td align="center" rowspan="1" colspan="1">7.0</td><td align="center" rowspan="1" colspan="1">87.2</td><td align="center" rowspan="1" colspan="1">82.2</td><td align="center" rowspan="1" colspan="1">74.5</td><td align="center" rowspan="1" colspan="1">36.4</td><td align="center" rowspan="1" colspan="1">29.1</td><td align="center" rowspan="1" colspan="1">19.9</td><td align="center" rowspan="1" colspan="1">72.4</td><td align="center" rowspan="1" colspan="1">29.1</td><td align="center" rowspan="1" colspan="1">50.5</td></tr><tr><td align="left" rowspan="1" colspan="1">Communicable</td><td align="center" rowspan="1" colspan="1">Formal</td><td align="center" rowspan="1" colspan="1">6.8</td><td align="center" rowspan="1" colspan="1">96.7</td><td align="center" rowspan="1" colspan="1">85.3</td><td align="center" rowspan="1" colspan="1">79.4</td><td align="center" rowspan="1" colspan="1">33.7</td><td align="center" rowspan="1" colspan="1">20.2</td><td align="center" rowspan="1" colspan="1">15.4</td><td align="center" rowspan="1" colspan="1">77.6</td><td align="center" rowspan="1" colspan="1">23.0</td><td align="center" rowspan="1" colspan="1">37.7</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;diseases</td><td align="center" rowspan="1" colspan="1">Informal</td><td align="center" rowspan="1" colspan="1">7.1</td><td align="center" rowspan="1" colspan="1">84.5</td><td align="center" rowspan="1" colspan="1">81.1</td><td align="center" rowspan="1" colspan="1">71.8</td><td align="center" rowspan="1" colspan="1">39.8</td><td align="center" rowspan="1" colspan="1">53.2</td><td align="center" rowspan="1" colspan="1">32.9</td><td align="center" rowspan="1" colspan="1">70.6</td><td align="center" rowspan="1" colspan="1">41.0</td><td align="center" rowspan="1" colspan="1">61.3</td></tr><tr><td align="left" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">Both</td><td align="center" rowspan="1" colspan="1">0.0</td><td align="center" rowspan="1" colspan="1">8.0</td><td align="center" rowspan="1" colspan="1">15.5</td><td align="center" rowspan="1" colspan="1">11.4</td><td align="center" rowspan="1" colspan="1">48.8</td><td align="center" rowspan="1" colspan="1">64.6</td><td align="center" rowspan="1" colspan="1">73.2</td><td align="center" rowspan="1" colspan="1">10.7</td><td align="center" rowspan="1" colspan="1">61.0</td><td align="center" rowspan="1" colspan="1">36.1</td></tr><tr><td align="left" rowspan="1" colspan="1">Non-communicable</td><td align="center" rowspan="1" colspan="1">Formal</td><td align="center" rowspan="1" colspan="1">0.0</td><td align="center" rowspan="1" colspan="1">3.3</td><td align="center" rowspan="1" colspan="1">12.3</td><td align="center" rowspan="1" colspan="1">12.9</td><td align="center" rowspan="1" colspan="1">51.8</td><td align="center" rowspan="1" colspan="1">73.3</td><td align="center" rowspan="1" colspan="1">78.7</td><td align="center" rowspan="1" colspan="1">8.7</td><td align="center" rowspan="1" colspan="1">68.0</td><td align="center" rowspan="1" colspan="1">52.0</td></tr><tr><td align="left" rowspan="1" colspan="1">diseases</td><td align="center" rowspan="1" colspan="1">Informal</td><td align="center" rowspan="1" colspan="1">0.0</td><td align="center" rowspan="1" colspan="1">9.3</td><td align="center" rowspan="1" colspan="1">16.6</td><td align="center" rowspan="1" colspan="1">10.5</td><td align="center" rowspan="1" colspan="1">45.0</td><td align="center" rowspan="1" colspan="1">41.0</td><td align="center" rowspan="1" colspan="1">57.6</td><td align="center" rowspan="1" colspan="1">11.4</td><td align="center" rowspan="1" colspan="1">47.2</td><td align="center" rowspan="1" colspan="1">22.7</td></tr><tr><td align="left" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">Both</td><td align="center" rowspan="1" colspan="1">93.0</td><td align="center" rowspan="1" colspan="1">3.9</td><td align="center" rowspan="1" colspan="1">0.5</td><td align="center" rowspan="1" colspan="1">0.6</td><td align="center" rowspan="1" colspan="1">3.7</td><td align="center" rowspan="1" colspan="1">0.0</td><td align="center" rowspan="1" colspan="1">0.0</td><td align="center" rowspan="1" colspan="1">14.1</td><td align="center" rowspan="1" colspan="1">1.5</td><td align="center" rowspan="1" colspan="1">7.7</td></tr><tr><td align="left" rowspan="1" colspan="1">Maternal/neonatal</td><td align="center" rowspan="1" colspan="1">Formal</td><td align="center" rowspan="1" colspan="1">93.2</td><td align="center" rowspan="1" colspan="1">0.0</td><td align="center" rowspan="1" colspan="1">0.0</td><td align="center" rowspan="1" colspan="1">1.8</td><td align="center" rowspan="1" colspan="1">4.3</td><td align="center" rowspan="1" colspan="1">0.0</td><td align="center" rowspan="1" colspan="1">0.0</td><td align="center" rowspan="1" colspan="1">11.5</td><td align="center" rowspan="1" colspan="1">1.5</td><td align="center" rowspan="1" colspan="1">4.2</td></tr><tr><td align="left" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">Informal</td><td align="center" rowspan="1" colspan="1">92.9</td><td align="center" rowspan="1" colspan="1">5.0</td><td align="center" rowspan="1" colspan="1">0.7</td><td align="center" rowspan="1" colspan="1">0.0</td><td align="center" rowspan="1" colspan="1">2.9</td><td align="center" rowspan="1" colspan="1">0.0</td><td align="center" rowspan="1" colspan="1">0.0</td><td align="center" rowspan="1" colspan="1">14.9</td><td align="center" rowspan="1" colspan="1">1.6</td><td align="center" rowspan="1" colspan="1">10.7</td></tr><tr><td align="left" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">Both</td><td align="center" rowspan="1" colspan="1">0.0</td><td align="center" rowspan="1" colspan="1">0.9</td><td align="center" rowspan="1" colspan="1">1.8</td><td align="center" rowspan="1" colspan="1">13.5</td><td align="center" rowspan="1" colspan="1">11.1</td><td align="center" rowspan="1" colspan="1">6.3</td><td align="center" rowspan="1" colspan="1">6.9</td><td align="center" rowspan="1" colspan="1">2.8</td><td align="center" rowspan="1" colspan="1">8.4</td><td align="center" rowspan="1" colspan="1">5.7</td></tr><tr><td align="left" rowspan="1" colspan="1">Injuries</td><td align="center" rowspan="1" colspan="1">Formal</td><td align="center" rowspan="1" colspan="1">0.0</td><td align="center" rowspan="1" colspan="1">0.0</td><td align="center" rowspan="1" colspan="1">2.4</td><td align="center" rowspan="1" colspan="1">5.9</td><td align="center" rowspan="1" colspan="1">10.2</td><td align="center" rowspan="1" colspan="1">6.5</td><td align="center" rowspan="1" colspan="1">5.9</td><td align="center" rowspan="1" colspan="1">2.2</td><td align="center" rowspan="1" colspan="1">7.5</td><td align="center" rowspan="1" colspan="1">6.1</td></tr><tr><td align="left" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">Informal</td><td align="center" rowspan="1" colspan="1">0.0</td><td align="center" rowspan="1" colspan="1">1.2</td><td align="center" rowspan="1" colspan="1">1.6</td><td align="center" rowspan="1" colspan="1">17.7</td><td align="center" rowspan="1" colspan="1">12.3</td><td align="center" rowspan="1" colspan="1">5.8</td><td align="center" rowspan="1" colspan="1">9.5</td><td align="center" rowspan="1" colspan="1">3.1</td><td align="center" rowspan="1" colspan="1">10.2</td><td align="center" rowspan="1" colspan="1">5.3</td></tr></tbody></table></table-wrap><p>Among people aged 15 and over, CDs no longer top the chart as the leading cause of death, all neighborhoods combined (<xref ref-type="table" rid="T0004">Table 4</xref>). Indeed with advancing age, CDs increasingly give way to NCDs, especially in formal neighborhoods. For example, in formal neighborhoods, these diseases account for 73.3% of the mortality risk between 50 and 64 years of age against 20.2% for CDs. NCDs have a contribution of 78.7% to the elderly (65 years and over) mortality rate in formal neighborhoods, against 15.4% for CDs. In informal neighborhoods, none of the two groups of causes contributes more than 50% of adult mortality but the contribution of NCDs remains high (47.2% among people aged 15 and over, <xref ref-type="table" rid="T0004">Table 4</xref>). It is specifically estimated at 45.0% between 15 and 49 years, 41.0% between 50 and 64 years, and 57.6% for people aged 65 and over (<xref ref-type="table" rid="T0004">Table 4</xref>).</p></sec></sec><sec sec-type="discussion" id="S0004"><title>Discussion</title><p>In the Ouagadougou HDSS, CDs remain the leading cause of death, accounting for 36.6% of deaths recorded between 2009 and 2011, and contributing 50.5% of the mortality rate. Among these CDs, malaria is the leading cause, followed by acute respiratory infections and HIV/AIDS. For all ages combined, deaths from CDs are more prevalent in informal neighborhoods while deaths from NCDs are more prevalent in formal neighborhoods. Results from the analysis of causes of death by age group are consistent with expectations and confirm our research hypotheses. Indeed, CDs are the leading cause of death among children (aged between 29 days and 14 years) in both formal and informal neighborhoods. There is no apparent difference between formal and informal neighborhoods in the epidemiological profile of children even though childhood mortality is higher in informal neighborhoods. Our results are also consistent with studies in other African urban settings such as Nairobi where Kyobutungi et al. have shown that CDs are the major causes of under-5 mortality in slums, a pattern that may be observable across sub-Saharan Africa (<xref rid="CIT0022" ref-type="bibr">22</xref>).</p><p>The findings also suggest that among adults (15 years and older), the two types of neighborhoods have different epidemiological profiles. For formal neighborhoods, the profile is dominated by NCDs with a contribution higher than 50% that increases with age. This is very visible among the ages 50&#x02013;64 and among the elderly (65 and over), with a contribution far greater than that observed for the same age groups in informal neighborhoods. For the latter, the epidemiological profile is mixed, with globally no group of causes accounting for more than 50%. NCDs which consist mainly of cardiovascular diseases and neoplasms exceed the threshold of 50% contribution only for the elderly (65 and over).</p><p>The fact that mortality from NCDs is higher in formal neighborhoods confirms the idea of a correlation between urbanization and epidemiological transition. Indeed, the formal neighborhoods of Ouagadougou HDSS are inhabited by wealthier and more educated people. Three adults (15 years and over) out of every four are educated with two in five having at least secondary level (<xref rid="CIT0019" ref-type="bibr">19</xref>). These ratios compare to about two in five and one in five in informal neighborhoods (<xref rid="CIT0019" ref-type="bibr">19</xref>). To further elucidate with some examples of facilities linked to the standard of living, it is estimated that 54% of households in formal areas have a TV, 17% a refrigerator, 66% a motorcycle, and 10% a car (<xref rid="CIT0019" ref-type="bibr">19</xref>). The numbers are respectively 13, 0.8, 30, and 0.6% for informal settlements (<xref rid="CIT0019" ref-type="bibr">19</xref>). Moreover, in the formal neighborhoods there are more bars, more shops; in short, this is where one feels the city. A health survey conducted in 2010 in the Ouagadougou HDSS areas showed a considerable prevalence of overweight and hypertension among adults in both neighborhoods, which was significantly higher in formal neighborhoods compared to the informal ones (<xref rid="CIT0023" ref-type="bibr">23</xref>).</p><p>Prevalence of overweight and hypertension was estimated at 32 and 19.6%, respectively, for formal neighborhoods and at 20 and 13% for the informal ones.</p><p> This study is very important for health policy evaluation. Malaria, for instance, is already the subject of a vertical disease control program in Ouagadougou with the distribution of bed nets running for several years. This study shows that malaria remains the leading cause of mortality among CDs. Hence, considerable efforts are still needed to achieve a significant reduction or even eradication of malaria mortality. Meanwhile, NCDs, especially cardiovascular diseases are preventable and/or manageable with a change in lifestyle. A prevention program would probably reduce the burden of these NCDs among adults and the elderly with a significant economic impact for families. In the case of Ouagadougou, such a program should involve both formal and informal neighborhoods as we have seen that both types of neighborhoods are affected by NCDs but with different variants. However, a greater emphasis should be devoted to formal neighborhoods. We believe that a community approach would be effective and useful for the Burkina Faso Health Ministry's new Bureau for NCDs which is currently working on a national communications plan for NCDs. This approach should involve community health workers collaborating with community groups to promote awareness of NCD risk factors and to help individuals make changes in their lives.</p><p>This study has some limitations, one of which is the high proportion (15.7%) of deaths without VA. This is partly explained by the migration of family members of the deceased that may be compounded by refusals from others to talk about their deceased relatives. It should also be noted that informal neighborhoods of Ouagadougou are populated mostly by migrants (<xref rid="CIT0019" ref-type="bibr">19</xref>) and so may be subject to a selection bias by migration. Indeed, due to their poor living conditions and also because of the low level of social ties in cities (<xref rid="CIT0024" ref-type="bibr">24</xref>), many city dwellers of informal neighborhoods may leave the area in the event of chronic conditions in search of care in their villages of origin or in other districts of the city where they may have relations. Other studies have already shown that migrants often tend to return home to die (<xref rid="CIT0025" ref-type="bibr">25</xref>, <xref rid="CIT0026" ref-type="bibr">26</xref>). Their eventual death may not be captured by the HDSS. This can partly exaggerate the differences in NCD mortality between formal and informal neighborhoods. A final limitation of the work is the impossibility of extrapolating our findings to the entire city of Ouagadougou as the HDSS only covers five districts including two formal and three informal.</p></sec><sec sec-type="conclusions" id="S0005"><title>Conclusions</title><p>In sub-Saharan Africa, studies on cause-specific mortality are scarce because of a lack of data on the causes of death. As we have already mentioned, frequently used data from hospital-based statistics on cause-specific mortality are seriously affected by selection bias since most deaths occur outside the formal heath care system with no possibility for clinical autopsy. Only HDSSs that use the VA method can produce &#x02018;reliable&#x02019; data on causes of death. Very few among currently existing HDSSs are located in urban areas. The Ouagadougou HDSS is the only one in Africa currently following both formal and informal settlements within a city. These two types of neighborhoods characterize contemporary African cities. The results presented in this paper are an important step toward understanding the intra-urban differences in the epidemiological profile in sub-Saharan Africa. Notwithstanding its limitations, this study provides credible evidence to stimulate discussion on cause-specific mortality differences between formal and informal neighborhoods in urban Africa.</p></sec>

Applying the InterVA-4 model to determine causes of death in rural Ethiopia

<sec id="st1"><title>Background</title><p>In Ethiopia, most deaths take place at home and routine certification of cause of death by physicians is lacking. As a result, reliable cause of death (CoD) data are often not available. Recently, a computerized method for interpretation of verbal autopsy (VA) data, called InterVA, has been developed and used. It calculates the probability of a set of CoD given the presence of circumstances, signs, and symptoms reported during VA interviews. We applied the InterVA model to describe CoD in a rural population of Ethiopia.</p></sec><sec id="st2"><title>Objective</title><p>VA data for 436/599 (72.7%) deaths that occurred during 2010&#x02013;2011 were included. InterVA-4 was used to interpret the VA data into probable cause of death. Cause-specific mortality fraction was used to describe frequency of occurrence of death from specific causes.</p></sec><sec id="st3"><title>Results</title><p>InterVA-4 was able to give likely cause(s) of death for 401/436 of the cases (92.0%). Overall, 35.0% of the total deaths were attributed to communicable diseases, and 30.7% to chronic non-communicable diseases. Tuberculosis (12.5%) and acute respiratory tract infections (10.4%) were the most frequent causes followed by neoplasms (9.6%) and diseases of circulatory system (7.2%).</p></sec><sec id="st4"><title>Conclusion</title><p>InterVA-4 can produce plausible estimates of the major public health problems that can guide public health interventions. We encourage further validation studies, in local settings, so that InterVA can be integrated into national health surveys.</p></sec>

<contrib contrib-type="author"><name><surname>Weldearegawi</surname><given-names>Berhe</given-names></name><xref ref-type="aff" rid="AF0001">1</xref><xref ref-type="aff" rid="AF0002">2</xref><xref ref-type="corresp" rid="cor1">*</xref></contrib><contrib contrib-type="author"><name><surname>Melaku</surname><given-names>Yohannes Adama</given-names></name><xref ref-type="aff" rid="AF0001">1</xref><xref ref-type="aff" rid="AF0002">2</xref></contrib><contrib contrib-type="author"><name><surname>Spigt</surname><given-names>Mark</given-names></name><xref ref-type="aff" rid="AF0001">1</xref><xref ref-type="aff" rid="AF0003">3</xref></contrib><contrib contrib-type="author"><name><surname>Dinant</surname><given-names>Geert Jan</given-names></name><xref ref-type="aff" rid="AF0003">3</xref></contrib>

Global Health Action

<p>Information about causes of death (CoD) is needed by health managers and policy makers at all levels of governance (<xref rid="CIT0001" ref-type="bibr">1</xref>, <xref rid="CIT0002" ref-type="bibr">2</xref>). In countries where registration of vital events and medical documentation of death are practiced, such information can easily be generated. However, three quarters of the world's total population lives in countries where registration of vital events and CoD certification are not in place (<xref rid="CIT0003" ref-type="bibr">3</xref>). Verbal autopsy (VA) is a technique growing in importance for estimating the CoD in populations without vital registration or other medical death certification and where the proportion of people who die at home is high (<xref rid="CIT0004" ref-type="bibr">4</xref>, <xref rid="CIT0005" ref-type="bibr">5</xref>).</p><p>VA means that trained data collectors interview the caregivers or family of a recently deceased person, asking about signs and symptoms preceding the death, which are then interpreted into a probable cause of death. It is now widely used to estimate cause-specific deaths in research and for routine registration of deaths (<xref rid="CIT0001" ref-type="bibr">1</xref>, <xref rid="CIT0006" ref-type="bibr">6</xref>, <xref rid="CIT0007" ref-type="bibr">7</xref>). Physician review, the commonly used method to derive a probable CoD from VA data, is a costly, slow, and non-reproducible process. Recently, a computerized method for interpretation of VA data (InterVA) has come into use. The InterVA process is comparatively fast and cheap and it is reproducible over time and place (<xref rid="CIT0007" ref-type="bibr">7</xref>). It is also aligned with the WHO 2012 VA standard (<xref rid="CIT0008" ref-type="bibr">8</xref>).</p><p>In Ethiopia, routine registration of vital events is non-existent and death certification is not compulsory. Thus, producing consistent, timely, and reliable CoD data has remained a challenge. Therefore, we used the InterVA method to interpret VA data from a rural population in Ethiopia.</p><sec sec-type="methods" id="S0002"><title>Methods</title><sec id="S0002-S20001"><title>Study setting</title><p>The Kilite Awlaelo Health and Demographic Surveillance System (KA-HDSS) is a longitudinal population-based surveillance site located about 802 km north of Addis Ababa, Ethiopia. The KA-HDSS is a member of the International Network for the Demographic Evaluation of Populations and Their Health (INDEPTH) Network [<ext-link ext-link-type="uri" xlink:href="http://www.indepth-network.org/">http://www.indepth-network.org/</ext-link>]. The KA-HDSS was started in September 2009 with a baseline population of 66,438 individuals living in 14,453 households. Data on household and vital individual events (pregnancy status, birth, death, marital status change, and migrations) and VA data were collected during house-to-house visits twice a year.</p></sec><sec id="S0002-S20002"><title>VA questionnaire and interview</title><p>The VA questionnaire was adapted from the WHO, INDEPTH Network, Sample Vital Registration with Verbal Autopsy (<xref rid="CIT0008" ref-type="bibr">8</xref>, <xref rid="CIT0009" ref-type="bibr">9</xref>). It has three separate questionnaires for the three age groups: neonate, post-neonate and children (29 days to 15 years), and adults (&#x0003e;15 years). Deaths were identified by data collectors during regular visits to the households. An adult relative of the deceased, who was the caregiver during the terminal illness, was interviewed by trained data collectors who completed at least high school. Data were collected after the end of the mourning period (45&#x02013;55 days) using the paper format which on average takes 110 min to fill out.</p></sec><sec id="S0002-S20003"><title>Interpretation of VA data</title><p>The InterVA-4 model (version 4.02) was used to interpret VA data into probable cause(s) of death. As described by Byass et al. (<xref rid="CIT0007" ref-type="bibr">7</xref>), the model is based on Bayes&#x02019; theorem, which calculates the probability of a set of CoD given the presence of indicators (circumstances, signs, and symptoms) reported in VA interviews (<xref rid="CIT0007" ref-type="bibr">7</xref>, <xref rid="CIT0010" ref-type="bibr">10</xref>). The InterVA requires extraction of a defined set of indicators from the VA questionnaire, and then processes these indicators to generate a summary of as many as three possible CoD with their corresponding likelihood (<xref rid="CIT0002" ref-type="bibr">2</xref>, <xref rid="CIT0011" ref-type="bibr">11</xref>). Fractional causes are then aggregated and any residual component (where fractional causes total less than 1) ascribed as indeterminate. This approach thus integrates a measure of the individual uncertainty with which InterVA-4 is able to assign cause(s) of death into the analysis for each case. The InterVA model assigned indeterminate cause, either to a certain fraction of a single case (indeterminate) or as a whole (completely indeterminate). However, if the VA questionnaire did not contain usable data, it was excluded from the analysis.</p><p>Before interpretation of VA data to likely CoD, InterVA-4 requires labeling the incidence of malaria and HIV/AIDS in the study setting as &#x02018;high&#x02019; or &#x02018;low&#x02019;. In Ethiopia, the prevalence of malaria and HIV/AIDS is 1% and 1.5%, respectively (<xref rid="CIT0012" ref-type="bibr">12</xref>, <xref rid="CIT0013" ref-type="bibr">13</xref>). Thus, for the current report, levels for both Malaria and HIV/AIDS were set as &#x0201c;low&#x0201d;. The current report is based on VA for deaths during 2010&#x02013;2011. The dataset used for this study was also contributed to the multisite INDEPTH Network cause-specific mortality dataset (<xref rid="CIT0014" ref-type="bibr">14</xref>).</p></sec><sec id="S0002-S20004"><title>Ethical statement</title><p>The KA-HDSS received ethical clearance from the Ethiopian Science and Technology Agency with identification number &#x02013; IERC 0030. Informed verbal consent was obtained from every respondent. The consent procedure was stated in the proposal which was approved by the ethical review committee.</p></sec></sec><sec sec-type="results" id="S0003"><title>Results</title><p>VA data were collected for a total of 436/599 (72.7%) deaths that occurred in KA-HDSS during 2010&#x02013;2011. These were processed by the InterVA-4 model which assigned cause(s) of death to all except 35 cases (8.0%) which were completely indeterminate. Residual components assigned to indeterminate cause amounted to an additional 33.5 (7.7%) of cases. Ninety percent of the deceased were from rural areas and the median age at death was 58 years (inter quartile range = 33 years). Most deaths, about 89%, took place outside health facilities. Neonates accounted for 6.9% of the cases, post neonates for 5.7%, children of 1&#x02013;4 years for 5.3%, and those 5&#x02013;14 for 6.9%. Among adults, age groups 15&#x02013;49, 50&#x02013;65 years, and 65-plus years accounted for 20.2, 12.2, and 42.9%, respectively.</p><p>Overall, 152.8 deaths (35.0%) were attributed to communicable diseases, 133.4 deaths (30.7%) to chronic non-communicable diseases, and 28.3 deaths (6.5%) to neonatal causes (<xref ref-type="table" rid="T0001">Table 1</xref>). Tuberculosis (TB) and acute respiratory tract infections (ARTI) including pneumonia were frequent communicable CoD, contributing 12.5 and 10.4% of the overall mortality, respectively. Neoplasms and diseases of the circulatory system were major chronic non-communicable causes, contributing 9.6 and 7.2% of the deaths, respectively. Neonatal pneumonia (4.0%) and external causes (9.3%) were the other important components of overall mortality.</p><table-wrap id="T0001" position="float"><label>Table 1</label><caption><p>InterVA-4 based cause of death by sex in KA-HDSS Ethiopia, 2010&#x02013;2011</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1">Cause of death</th><th align="center" rowspan="1" colspan="1">WHO VA code</th><th align="center" rowspan="1" colspan="1">Female <italic>N</italic> (%)</th><th align="center" rowspan="1" colspan="1">Male <italic>N</italic> (%)</th><th align="center" rowspan="1" colspan="1">Total <italic>N</italic> (%)</th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">I. Communicable diseases</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">75.4 (38.5)</td><td align="center" rowspan="1" colspan="1">77.4 (32.2)</td><td align="center" rowspan="1" colspan="1">152.8 (35.0)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Tuberculosis</td><td align="center" rowspan="1" colspan="1">VA-01.09</td><td align="center" rowspan="1" colspan="1">25.9 (13.2)</td><td align="center" rowspan="1" colspan="1">28.7 (12.0)</td><td align="center" rowspan="1" colspan="1">54.6 (12.5)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;ARTI, including pneumonia</td><td align="center" rowspan="1" colspan="1">VA-01.02</td><td align="center" rowspan="1" colspan="1">21.9 (11.2)</td><td align="center" rowspan="1" colspan="1">23.4 (9.7)</td><td align="center" rowspan="1" colspan="1">45.3 (10.4)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;HIV/AIDS related</td><td align="center" rowspan="1" colspan="1">VA-01.03</td><td align="center" rowspan="1" colspan="1">8.8 (4.5)</td><td align="center" rowspan="1" colspan="1">7.8 (3.2)</td><td align="center" rowspan="1" colspan="1">16.6 (3.8)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Malaria</td><td align="center" rowspan="1" colspan="1">VA-01.05</td><td align="center" rowspan="1" colspan="1">6.0 (3.1)</td><td align="center" rowspan="1" colspan="1">5.9 (2.5)</td><td align="center" rowspan="1" colspan="1">11.9 (2.7)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Diarrheal diseases</td><td align="center" rowspan="1" colspan="1">VA-01.04</td><td align="center" rowspan="1" colspan="1">5.8 (2.9)</td><td align="center" rowspan="1" colspan="1">3.3 (1.4)</td><td align="center" rowspan="1" colspan="1">9.0 (2.1)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Others</td><td align="center" rowspan="1" colspan="1">Other VA-01</td><td align="center" rowspan="1" colspan="1">7.0 (3.6)</td><td align="center" rowspan="1" colspan="1">8.3 (3.4)</td><td align="center" rowspan="1" colspan="1">15.4 (3.5)</td></tr><tr><td align="left" rowspan="1" colspan="1">II. Non-communicable diseases:</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">56.8 (29.0)</td><td align="center" rowspan="1" colspan="1">77.0 (32.1)</td><td align="center" rowspan="1" colspan="1">133.4 (30.7)</td></tr><tr><td align="left" rowspan="1" colspan="1">Neoplasms</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">20.4 (10.4)</td><td align="center" rowspan="1" colspan="1">21.6 (9.0)</td><td align="center" rowspan="1" colspan="1">41.9 (9.6)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Respiratory neoplasms</td><td align="center" rowspan="1" colspan="1">VA-02.03</td><td align="center" rowspan="1" colspan="1">6.2 (3.2)</td><td align="center" rowspan="1" colspan="1">9.8 (4.1)</td><td align="center" rowspan="1" colspan="1">16.0 (3.7)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Other neoplasms</td><td align="center" rowspan="1" colspan="1">VA-02.99</td><td align="center" rowspan="1" colspan="1">4.7 (2.4)</td><td align="center" rowspan="1" colspan="1">5.8 (2.4)</td><td align="center" rowspan="1" colspan="1">10.5 (2.4)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Digestive neoplasms</td><td align="center" rowspan="1" colspan="1">VA-02.02</td><td align="center" rowspan="1" colspan="1">6.9 (3.5)</td><td align="center" rowspan="1" colspan="1">4.4 (1.8)</td><td align="center" rowspan="1" colspan="1">11.3 (2.6)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Reproductive neoplasms</td><td align="center" rowspan="1" colspan="1">VA-02.05, 06</td><td align="center" rowspan="1" colspan="1">2.6 (1.3)</td><td align="center" rowspan="1" colspan="1">1.6 (0.7)</td><td align="center" rowspan="1" colspan="1">4.2 (1.0)</td></tr><tr><td align="left" rowspan="1" colspan="1">Diseases of the circulatory system</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">13.1 (6.7)</td><td align="center" rowspan="1" colspan="1">18.5 (7.7)</td><td align="center" rowspan="1" colspan="1">31.4 (7.2)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Stroke</td><td align="center" rowspan="1" colspan="1">VA-04.02</td><td align="center" rowspan="1" colspan="1">8.1 (4.1)</td><td align="center" rowspan="1" colspan="1">8.8 (3.7)</td><td align="center" rowspan="1" colspan="1">16.9 (3.9)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Other cardiac disease</td><td align="center" rowspan="1" colspan="1">VA-04.99</td><td align="center" rowspan="1" colspan="1">4.1 (2.1)</td><td align="center" rowspan="1" colspan="1">6.8 (2.8)</td><td align="center" rowspan="1" colspan="1">10.9 (2.5)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Acute cardiac disease</td><td align="center" rowspan="1" colspan="1">VA-04.01</td><td align="center" rowspan="1" colspan="1">0.9 (0.5)</td><td align="center" rowspan="1" colspan="1">2.9 (1.2)</td><td align="center" rowspan="1" colspan="1">3.8 (0.8)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Gastrointestinal disorders</td><td align="center" rowspan="1" colspan="1">VA-06.01, 02</td><td align="center" rowspan="1" colspan="1">4.2 (2.1)</td><td align="center" rowspan="1" colspan="1">17.1 (7.1)</td><td align="center" rowspan="1" colspan="1">21.3 (4.9)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Diabetes mellitus</td><td align="center" rowspan="1" colspan="1">VA-03.03</td><td align="center" rowspan="1" colspan="1">2.7 (1.4)</td><td align="center" rowspan="1" colspan="1">2.9 (1.2)</td><td align="center" rowspan="1" colspan="1">5.6 (1.3)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Mental disorders: epilepsy</td><td align="center" rowspan="1" colspan="1">VA-08.01</td><td align="center" rowspan="1" colspan="1">3.9 (2.0)</td><td align="center" rowspan="1" colspan="1">7.3 (3.0)</td><td align="center" rowspan="1" colspan="1">11.6 (2.6)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Respiratory disorders<xref ref-type="table-fn" rid="TF0001">a</xref></td><td align="center" rowspan="1" colspan="1">VA-05.01, 02</td><td align="center" rowspan="1" colspan="1">6.6 (3.4)</td><td align="center" rowspan="1" colspan="1">6.0 (2.5)</td><td align="center" rowspan="1" colspan="1">12.6 (2.9)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Renal disorders: renal failure</td><td align="center" rowspan="1" colspan="1">VA-07.01</td><td align="center" rowspan="1" colspan="1">2.5 (1.3)</td><td align="center" rowspan="1" colspan="1">2.1 (0.9)</td><td align="center" rowspan="1" colspan="1">4.5 (1.0)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Other and unspecified NCDs</td><td align="center" rowspan="1" colspan="1">VA-98</td><td align="center" rowspan="1" colspan="1">3.4 (1.7)</td><td align="center" rowspan="1" colspan="1">1.5 (0.6)</td><td align="center" rowspan="1" colspan="1">4.9 (1.1)</td></tr><tr><td align="left" rowspan="1" colspan="1">III. Neonatal causes of death</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">9.6 (4.9)</td><td align="center" rowspan="1" colspan="1">18.7 (7.8)</td><td align="center" rowspan="1" colspan="1">28.3 (6.5)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Neonatal pneumonia</td><td align="center" rowspan="1" colspan="1">VA-10.03</td><td align="center" rowspan="1" colspan="1">7.0 (3.6)</td><td align="center" rowspan="1" colspan="1">10.5 (4.4)</td><td align="center" rowspan="1" colspan="1">17.5 (4.0)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Other neonatal</td><td align="center" rowspan="1" colspan="1">Other VA-10</td><td align="center" rowspan="1" colspan="1">2.6 (1.3)</td><td align="center" rowspan="1" colspan="1">8.2 (3.4)</td><td align="center" rowspan="1" colspan="1">10.8 (2.5)</td></tr><tr><td align="left" rowspan="1" colspan="1">IV. External causes of death</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">14.5 (7.4)</td><td align="center" rowspan="1" colspan="1">26.1 (10.9)</td><td align="center" rowspan="1" colspan="1">40.6 (9.3)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Accidental fall, drowning</td><td align="center" rowspan="1" colspan="1">VA-12.03-04</td><td align="center" rowspan="1" colspan="1">3.8 (1.9)</td><td align="center" rowspan="1" colspan="1">8.5 (3.5)</td><td align="center" rowspan="1" colspan="1">12.3 (2.8)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Self-harm, assault</td><td align="center" rowspan="1" colspan="1">VA-12.08-09</td><td align="center" rowspan="1" colspan="1">2.7 (1.4)</td><td align="center" rowspan="1" colspan="1">9.0 (3.8)</td><td align="center" rowspan="1" colspan="1">11.7 (2.7)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Road traffic accident</td><td align="center" rowspan="1" colspan="1">VA-12.01</td><td align="center" rowspan="1" colspan="1">2.0 (1.0)</td><td align="center" rowspan="1" colspan="1">4.5 (1.9)</td><td align="center" rowspan="1" colspan="1">6.5 (1.5)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Others &#x00026; unspecified</td><td align="center" rowspan="1" colspan="1">Other VA-12</td><td align="center" rowspan="1" colspan="1">6.0 (3.1)</td><td align="center" rowspan="1" colspan="1">4.1 (1.7)</td><td align="center" rowspan="1" colspan="1">10.1 (2.3)</td></tr><tr><td align="left" rowspan="1" colspan="1">V. Malnutrition</td><td align="center" rowspan="1" colspan="1">VA-03.01-02</td><td align="center" rowspan="1" colspan="1">4.9 (2.5)</td><td align="center" rowspan="1" colspan="1">5.0 (2.1)</td><td align="center" rowspan="1" colspan="1">10.0 (2.3)</td></tr><tr><td align="left" rowspan="1" colspan="1">VI. Maternal causes</td><td align="center" rowspan="1" colspan="1">VA-09</td><td align="center" rowspan="1" colspan="1">2.4 (1.2)</td><td align="center" rowspan="1" colspan="1"> <bold>&#x02013;</bold> </td><td align="center" rowspan="1" colspan="1">2.4 (0.6)</td></tr><tr><td align="left" rowspan="1" colspan="1">VII. Indeterminate<xref ref-type="table-fn" rid="TF0002">b</xref> </td><td align="center" rowspan="1" colspan="1">VA-99</td><td align="center" rowspan="1" colspan="1">32.6 (16.6)</td><td align="center" rowspan="1" colspan="1">35.9 (15.0)</td><td align="center" rowspan="1" colspan="1">68.5 (15.7)</td></tr><tr><td align="left" rowspan="1" colspan="1">Total</td><td align="center" rowspan="1" colspan="1"/><td align="center" rowspan="1" colspan="1">196 (100.0)</td><td align="center" rowspan="1" colspan="1">240 (100.0)</td><td align="center" rowspan="1" colspan="1">436 (100.0)</td></tr></tbody></table><table-wrap-foot><fn id="TF0001"><label>a</label><p>Chronic obstructive pulmonary disease, Asthma</p></fn><fn id="TF0002"><label>b</label><p>residual and completely indeterminate.</p></fn></table-wrap-foot></table-wrap><p>Chronic non-communicable diseases and communicable diseases caused comparable proportions of deaths in both sexes. A large proportion of neonatal deaths (58.3%) was attributed to neonatal pneumonia (<xref ref-type="table" rid="T0002">Table 2</xref>). ARTI including pneumonia was the leading cause of death in infants and children, accounting for 72.0 and 17.0%, respectively. Among adults, TB was the leading cause of death in age groups 15&#x02013;49, 50&#x02013;65, and 65-plus years, accounting for 19.9, 24.5, and 12.4%, respectively.</p><table-wrap id="T0002" position="float"><label>Table 2</label><caption><p>Leading InterVA-4 based causes of death by age group, KA-HDSS Ethiopia, 2010&#x02013;2011</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1"/><th align="center" rowspan="1" colspan="1"> <italic>N</italic> (%)</th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">Neonates (<italic>n</italic>=30)</td><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Neonatal pneumonia</td><td align="center" rowspan="1" colspan="1">17.5 (58.3)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Unspecified neonatal CoD</td><td align="center" rowspan="1" colspan="1">5.5 (18.3)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Birth asphyxia</td><td align="center" rowspan="1" colspan="1">2.9 (9.6)</td></tr><tr><td align="left" rowspan="1" colspan="1">Infants (<italic>n</italic>=25)</td><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;ARTI, including pneumonia</td><td align="center" rowspan="1" colspan="1">18.0 (72.0)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Indeterminate</td><td align="center" rowspan="1" colspan="1">2.7 (10.8)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;HIV/AIDS related</td><td align="center" rowspan="1" colspan="1">1.9 (7.6)</td></tr><tr><td align="left" rowspan="1" colspan="1">1&#x02013;4 years (<italic>n</italic>=23)</td><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Indeterminate</td><td align="center" rowspan="1" colspan="1">5.3 (23.0)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;ARTI, including pneumonia</td><td align="center" rowspan="1" colspan="1">3.9 (17.0)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;HIV/AIDS related</td><td align="center" rowspan="1" colspan="1">2.9 (12.6)</td></tr><tr><td align="left" rowspan="1" colspan="1">5&#x02013;14 years (<italic>n</italic>=30)</td><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Indeterminate</td><td align="center" rowspan="1" colspan="1">6.0 (20.0)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Accidental drowning</td><td align="center" rowspan="1" colspan="1">3.0 (10.0)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Road traffic accident</td><td align="center" rowspan="1" colspan="1">2.5 (8.3)</td></tr><tr><td align="left" rowspan="1" colspan="1">15&#x02013;49 years (<italic>n</italic>=88)</td><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Tuberculosis</td><td align="center" rowspan="1" colspan="1">17.5 (19.9)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Indeterminate</td><td align="center" rowspan="1" colspan="1">7.7 (8.8)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Acute abdomen</td><td align="center" rowspan="1" colspan="1">5.0 (5.7)</td></tr><tr><td align="left" rowspan="1" colspan="1">50&#x02013;64 years (<italic>n</italic>=53)</td><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Tuberculosis</td><td align="center" rowspan="1" colspan="1">13.0 (24.5)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Indeterminate</td><td align="center" rowspan="1" colspan="1">8.7 (16.4)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;HIV/AIDS related</td><td align="center" rowspan="1" colspan="1">4.1 (7.7)</td></tr><tr><td align="left" rowspan="1" colspan="1">65-plus years (<italic>n</italic>=187)</td><td align="center" rowspan="1" colspan="1"/></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Indeterminate</td><td align="center" rowspan="1" colspan="1">37.4 (20.0)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;Tuberculosis</td><td align="center" rowspan="1" colspan="1">23.1 (12.4)</td></tr><tr><td align="left" rowspan="1" colspan="1">&#x02003;ARTI, including pneumonia</td><td align="center" rowspan="1" colspan="1">14.6 (7.8)</td></tr></tbody></table></table-wrap></sec><sec sec-type="discussion" id="S0004"><title>Discussion</title><p>In this rural community, where reliable sources of CoD data are absent, the InterVA model generated plausible estimates of the major public health problems. Moreover, InterVA yields CoD which is completely internally consistent, allowing comparisons of data from different countries. It is also less labor intensive as compared to physician review. Despite its computational simplicity, it is certainly true that using any mathematical model for interpreting cause of death may not reflect the subjective subtleties of physician review, barring inconsistent physician reviews.</p><p>The proportion of deaths attributed to chronic non-communicable causes in our study (30.7%) was similar to the 34.5% (cardiac diseases, other non-communicable diseases, diabetes) reported in a similar study from rural north western Ethiopia (<xref rid="CIT0015" ref-type="bibr">15</xref>). Comparable estimates were also reported from other studies in Ethiopia that used hospital records (31.0%) and physician review (28.6%) methods (<xref rid="CIT0016" ref-type="bibr">16</xref>, <xref rid="CIT0017" ref-type="bibr">17</xref>). The preponderance of chronic non-communicable causes in the rural setting is likely to be explained by rapid socioeconomic development and parallel large-scale investments in health care (<xref rid="CIT0012" ref-type="bibr">12</xref>, <xref rid="CIT0018" ref-type="bibr">18</xref>). According to the World Bank, the Ethiopian economy has experienced strong and broad-based growth over the past decade, averaging 9.9% per year in 2004&#x02013;05 to 2011&#x02013;12 compared to the East African average of 5.4% (<xref rid="CIT0019" ref-type="bibr">19</xref>). As a result, there is an increase in life expectancy (<xref rid="CIT0012" ref-type="bibr">12</xref>), as well as exposure to risk factors for chronic non-communicable diseases (<xref rid="CIT0020" ref-type="bibr">20</xref>). A recent survey in rural south western Ethiopia showed that 80% of the population surveyed had at least one risk factor for chronic non-communicable diseases (<xref rid="CIT0020" ref-type="bibr">20</xref>). Several studies from Ethiopia also showed that chronic non-communicable diseases are increasingly becoming more apparent health problems (<xref rid="CIT0020" ref-type="bibr">20</xref>&#x02013;<xref rid="CIT0022" ref-type="bibr">22</xref>).</p><p>The contribution of communicable causes to the overall deaths in our study (35.0%) was lower than the 47.5% (TB, HIV/AIDS, and other infectious disease) reported from northwestern Ethiopia (<xref rid="CIT0015" ref-type="bibr">15</xref>) and much lower than the 58.0% in Kenya (<xref rid="CIT0002" ref-type="bibr">2</xref>). Misganaw et al. also reported that mortality from communicable, maternal, neonatal, and nutritional conditions have decreased from 68.0% in 2002 to 41.0% in 2010 (<xref rid="CIT0016" ref-type="bibr">16</xref>). Despite the variation in the estimates, both studies showed that the burden of communicable diseases in Ethiopia has declined. This could also be explained by the improvements in health and socioeconomic status of the population. Primary health service coverage has now reached 92% (<xref rid="CIT0023" ref-type="bibr">23</xref>). The national health care program, which focuses on health promotion and prevention of common health problems also, is likely to have played a significant role (<xref rid="CIT0012" ref-type="bibr">12</xref>). Deaths from malaria have declined by 50% between 2007&#x02013;8 and 2011, child mortality rate by 28.4%, during 2005&#x02013;2010, and HIV/AIDS prevalence among the adults has dropped to 1.5% in 2010&#x02013;11 (<xref rid="CIT0012" ref-type="bibr">12</xref>).</p><p>In our study, TB and ARTI including pneumonia were frequently diagnosed communicable CoD. TB was also identified as the leading communicable CoD in other similar studies, but the estimates attributed to TB were higher than in our findings (12.5%); 36% in Ethiopia and 31% in Nairobi, Kenya (<xref rid="CIT0002" ref-type="bibr">2</xref>, <xref rid="CIT0015" ref-type="bibr">15</xref>). A lower prevalence of TB than the national estimate was also reported in a recent survey in the region where our study was conducted (<xref rid="CIT0024" ref-type="bibr">24</xref>). Studies from Ethiopia reported comparable mortality estimates attributed to ARTI including pneumonia (<xref rid="CIT0016" ref-type="bibr">16</xref>, <xref rid="CIT0017" ref-type="bibr">17</xref>). The proportion of deaths attributed to HIV/AIDS in our study (3.8%) was much lower than findings from other studies in Ethiopia (7.6%), Nairobi (17.0%), and Kilifi (12.4%) in Kenya (<xref rid="CIT0002" ref-type="bibr">2</xref>, <xref rid="CIT0010" ref-type="bibr">10</xref>, <xref rid="CIT0015" ref-type="bibr">15</xref>). Despite, the geographic variation in the prevalence of HIV/AIDS, in three of these studies (<xref rid="CIT0002" ref-type="bibr">2</xref>, <xref rid="CIT0010" ref-type="bibr">10</xref>, <xref rid="CIT0015" ref-type="bibr">15</xref>), the level of HIV was set to be &#x02018;high&#x02019; in the model which might have affected the estimates.</p><p>Diseases arising during the neonatal period were important CoD next to the two leading groups of CoD. This was comparable to findings from rural south Ethiopia, where 6.5% of total deaths were attributed to neonatal causes and premature deaths (<xref rid="CIT0016" ref-type="bibr">16</xref>). In our findings, neonatal pneumonia was the major killer among neonates, causing more than half of all deaths during the neonatal period. Generally, pneumonia is the major cause of neonatal deaths in developing countries (<xref rid="CIT0025" ref-type="bibr">25</xref>).</p><p>The model also estimated deaths from accidents and injuries consistent to findings from Ethiopia (9.6%), Kenya (8.8%), and the global WHO estimate (9%) (<xref rid="CIT0002" ref-type="bibr">2</xref>, <xref rid="CIT0015" ref-type="bibr">15</xref>, <xref rid="CIT0026" ref-type="bibr">26</xref>). In our study, deaths from accidents and injuries were more prevalent in males, and children aged 5&#x02013;14 were more affected than the other age groups. This was consistent with findings from Uganda and a WHO global report on injuries (<xref rid="CIT0026" ref-type="bibr">26</xref>). The sex difference in burden of deaths from accidents and injuries is explained by variation in the roles men have in most societies. Males often engage in more hazardous and risky jobs than females (<xref rid="CIT0026" ref-type="bibr">26</xref>). Children are also more vulnerable to accidents and injuries as they are less able to predict and prevent accidents than adults (<xref rid="CIT0026" ref-type="bibr">26</xref>).</p><p>This study used standardized data collection tools and trained full-time data collectors. Moreover, the VA data analyzed in this study were collected as part of the routine follow-up of the KA-HDSS, which would have minimized recall bias. However, this study will have limitations inherent to limitations of the VA process.</p></sec><sec sec-type="conclusion" id="S0005"><title>Conclusion</title><p>In general, the major public health problems identified by the InterVA model were comparable to the expected local burden of diseases. Communicable diseases and chronic non-communicable diseases caused similar proportions of deaths. Neoplasms and diseases of the circulatory system were the major chronic non-communicable causes. TB and acute respiratory infections were the leading specific CoD. In countries where death certification is non-existent, the InterVA tool is feasible for generating cause of death data that would be satisfactory to guide public health interventions. We encourage validation studies, in local settings, so that the InterVA can be integrated into the national health surveys to yield nationwide cause of death data.</p></sec>

Microbial contributions to coupled arsenic and sulfur cycling in the acid-sulfide hot spring Champagne Pool, New Zealand

<p>Acid-sulfide hot springs are analogs of early Earth geothermal systems where microbial metal(loid) resistance likely first evolved. Arsenic is a metalloid enriched in the acid-sulfide hot spring Champagne Pool (Waiotapu, New Zealand). Arsenic speciation in Champagne Pool follows reaction paths not yet fully understood with respect to biotic contributions and coupling to biogeochemical sulfur cycling. Here we present quantitative arsenic speciation from Champagne Pool, finding arsenite dominant in the pool, rim and outflow channel (55&#x02013;75% total arsenic), and dithio- and trithioarsenates ubiquitously present as 18&#x02013;25% total arsenic. In the outflow channel, dimethylmonothioarsenate comprised &#x02264;9% total arsenic, while on the outflow terrace thioarsenates were present at 55% total arsenic. We also quantified sulfide, thiosulfate, sulfate and elemental sulfur, finding sulfide and sulfate as major species in the pool and outflow terrace, respectively. Elemental sulfur concentration reached a maximum at the terrace. Phylogenetic analysis of 16S rRNA genes from metagenomic sequencing revealed the dominance of <italic>Sulfurihydrogenibium</italic> at all sites and an increased archaeal population at the rim and outflow channel. Several phylotypes were found closely related to known sulfur- and sulfide-oxidizers, as well as sulfur- and sulfate-reducers. Bioinformatic analysis revealed genes underpinning sulfur redox transformations, consistent with sulfur speciation data, and illustrating a microbial role in sulfur-dependent transformation of arsenite to thioarsenate. Metagenomic analysis also revealed genes encoding for arsenate reductase at all sites, reflecting the ubiquity of thioarsenate and a need for microbial arsenate resistance despite anoxic conditions. Absence of the arsenite oxidase gene, <italic>aio</italic>, at all sites suggests prioritization of arsenite detoxification over coupling to energy conservation. Finally, detection of methyl arsenic in the outflow channel, in conjunction with increased sequences from <italic>Aquificaceae</italic>, supports a role for methyltransferase in thermophilic arsenic resistance. Our study highlights microbial contributions to coupled arsenic and sulfur cycling at Champagne Pool, with implications for understanding the evolution of microbial arsenic resistance in sulfidic geothermal systems.</p>

<contrib contrib-type="author"><name><surname>Hug</surname><given-names>Katrin</given-names></name><xref ref-type="aff" rid="aff1">¹</xref><xref ref-type="author-notes" rid="fn001">^*</xref><uri xlink:type="simple" xlink:href="http://community.frontiersin.org/people/u/164375"/></contrib><contrib contrib-type="author"><name><surname>Maher</surname><given-names>William A.</given-names></name><xref ref-type="aff" rid="aff2">²</xref><uri xlink:type="simple" xlink:href="http://community.frontiersin.org/people/u/188477"/></contrib><contrib contrib-type="author"><name><surname>Stott</surname><given-names>Matthew B.</given-names></name><xref ref-type="aff" rid="aff3">³</xref><uri xlink:type="simple" xlink:href="http://community.frontiersin.org/people/u/24239"/></contrib><contrib contrib-type="author"><name><surname>Krikowa</surname><given-names>Frank</given-names></name><xref ref-type="aff" rid="aff2">²</xref><uri xlink:type="simple" xlink:href="http://community.frontiersin.org/people/u/189487"/></contrib><contrib contrib-type="author"><name><surname>Foster</surname><given-names>Simon</given-names></name><xref ref-type="aff" rid="aff2">²</xref><uri xlink:type="simple" xlink:href="http://community.frontiersin.org/people/u/189467"/></contrib><contrib contrib-type="author"><name><surname>Moreau</surname><given-names>John W.</given-names></name><xref ref-type="aff" rid="aff1">¹</xref><xref ref-type="author-notes" rid="fn001">^*</xref><uri xlink:type="simple" xlink:href="http://community.frontiersin.org/people/u/31175"/></contrib>

Frontiers in Microbiology

<sec sec-type="introduction" id="s1"><title>Introduction</title><p>Active geothermal springs provide a modern analog for environments in which early life on Earth evolved metal(loid) resistance mechanisms (Stetter, <xref rid="B57" ref-type="bibr">2006</xref>; Martin et al., <xref rid="B37" ref-type="bibr">2008</xref>). In addition to high temperatures, high concentrations of dissolved toxic metal(loid)s present a strong selective pressure (Hirner et al., <xref rid="B23" ref-type="bibr">1998</xref>) on extant hot spring microbial communities. Correspondingly, there is evidence to support the evolution of several microbial metal(loid) tolerance mechanisms in geothermal settings (Barkay et al., <xref rid="B4" ref-type="bibr">2003</xref>; Jackson and Dugas, <xref rid="B24" ref-type="bibr">2003</xref>; Maezato and Blum, <xref rid="B35" ref-type="bibr">2012</xref>). In this regard, understanding the structure, diversity and functionality of modern hot spring microbial communities in the context of arsenic speciation may yield insights into the environmental conditions and constraints under which specific arsenic tolerance strategies evolved.</p><p>Arsenic is a metal(loid) that is toxic to microorganisms at elevated concentrations (Ballantyne and Moore, <xref rid="B3" ref-type="bibr">1988</xref>) and can be present as several chemical species including the oxyanions arsenite (AsO^3&#x02212;₃) and arsenate (AsO^3&#x02212;₄) as well as arsenic thioanions. Arsenite has a high affinity to sulfhydryl groups in amino acids, thereby disrupting protein function (Oremland and Stolz, <xref rid="B39" ref-type="bibr">2003</xref>). Arsenate is a phosphate analog, which displaces phosphate ions in enzyme reactions and therefore interferes with the cellular metabolism (Oremland and Stolz, <xref rid="B39" ref-type="bibr">2003</xref>) or leads to mutagenic effects (Li&#x000e8;vremont et al., <xref rid="B30" ref-type="bibr">2009</xref>). Previous work on arsenic speciation in geothermal environments reported the presence of primarily arsenite and arsenate contributing to the bulk arsenic speciation (Ballantyne and Moore, <xref rid="B3" ref-type="bibr">1988</xref>; Yokoyama et al., <xref rid="B68" ref-type="bibr">1993</xref>; Macur et al., <xref rid="B34" ref-type="bibr">2004</xref>). However, improved sample preservation techniques, have revealed significant concentrations of thioarsenate species (Wilkin et al., <xref rid="B64" ref-type="bibr">2003</xref>; Stauder et al., <xref rid="B56" ref-type="bibr">2005</xref>; Planer-Friedrich et al., <xref rid="B42" ref-type="bibr">2007</xref>; Wallschl&#x000e4;ger and Stadey, <xref rid="B62" ref-type="bibr">2007</xref>), which can comprise more than 50% of the total dissolved arsenic in sulfidic waters (Wilkin et al., <xref rid="B64" ref-type="bibr">2003</xref>). The presence of thioarsenates implies a potential dependence for arsenic speciation on sulfur redox cycling.</p><p>Sulfide, elemental sulfur, thiosulfate, and sulfate are common electron donors or acceptors for microorganisms under hydrothermal conditions (Amend and Shock, <xref rid="B71" ref-type="bibr">2001</xref>; Kletzin et al., <xref rid="B28" ref-type="bibr">2004</xref>; Gosh and Dam, <xref rid="B20" ref-type="bibr">2009</xref>; Macur et al., <xref rid="B33" ref-type="bibr">2013</xref>), and sulfide ions are highly reactive with arsenic (Sharma and Sohn, <xref rid="B54" ref-type="bibr">2009</xref>). Thus, microbially-mediated sulfur cycling can exert a profound, although indirect, influence on arsenic speciation, specifically through controlling the relative abundance of thioarsenate species. In comparison to arsenite and arsenate, thioarsenates are considered to be less toxic for microorganisms, as the sulfur-arsenic bond leaves no free electron pair to bind with sulfhydryl-groups in amino acids (Stauder et al., <xref rid="B56" ref-type="bibr">2005</xref>). Comparative genomic studies of the selenocysteine synthesis mechanism suggest thioarsenates may even be a microbial detoxification product in sulfur-rich environments (Couture et al., <xref rid="B9" ref-type="bibr">2012</xref>). However, work by Planer-Friedrich et al. (<xref rid="B41" ref-type="bibr">2008</xref>) identified thioarsenate species as potentially toxic to microorganisms over longer exposure times.</p><p>Microbes employ a range of strategies to detoxify arsenic. The most ubiquitous arsenic resistance mechanism is the <italic>ars</italic> operon gene expression, which requires genes encoding for proteins that identify and transport arsenic (Pa&#x000e9;z-Espino et al., <xref rid="B40" ref-type="bibr">2009</xref>). The gene <italic>arsC</italic> expresses a reductase, which is able to convert arsenate into arsenite (Gladysheva et al., <xref rid="B18" ref-type="bibr">1994</xref>), thereby providing resistance for arsenate. The gene <italic>arsR</italic> encodes for a transcriptional repressor, which controls the expression of the remaining <italic>ars</italic> operon genes <italic>arsA, arsB, arsD, arsH</italic>, and can only be activated by arsenite (Wu and Rosen, <xref rid="B65" ref-type="bibr">1991</xref>). The gene <italic>arsD</italic> encodes for the metallochaperon ArsD that transfers arsenite to ArsA, which is an ATPase encoded by <italic>arsA</italic> and located at the cell membrane (Lin et al., <xref rid="B31" ref-type="bibr">2007</xref>). The allosterically activated ArsA works as a catalytic subunit of ArsB, enhancing the activity of the membrane-located arsenite transporter that excludes arsenite from the cell (Rosen, <xref rid="B51" ref-type="bibr">2002</xref>). In some cases the <italic>ars</italic> operon includes <italic>arsH</italic>, which encodes for an arsenite resistance enhancer ArsH, important at high arsenite concentrations (Branco et al., <xref rid="B6" ref-type="bibr">2008</xref>). The gene <italic>aio</italic> (formerly known as <italic>aox</italic>, <italic>aro</italic> or <italic>aso</italic>) is a well-conserved arsenic resistance gene amongst several species that responds to degenerate primers (Qu&#x000e9;m&#x000e9;neur et al., <xref rid="B46" ref-type="bibr">2008</xref>). It encodes for the arsenite oxidase Aio, which is responsible for the oxidation of arsenite into arsenate. Conversely, the highly diverse <italic>arr</italic> gene encodes for the respiratory arsenate reductase Arr in arsenate respiring microorganisms, which reduces arsenate into arsenite. A recent study by Richey et al. (<xref rid="B50" ref-type="bibr">2009</xref>) identifies a bidirectional enzyme Arr that is able to reduce arsenate as well as oxidize arsenite, implying an ancient origin. Despite the closer evolutionary relationship to Arr, Zargar et al. (<xref rid="B70" ref-type="bibr">2010</xref>, <xref rid="B69" ref-type="bibr">2012</xref>) identify this gene as a new arsenite oxidase encoding gene referred to as <italic>arxA</italic>, because of its known function as an arsenite oxidase. The respiratory arsenate reductase and arsenite oxidase resistance mechanisms are both beneficial for microorganisms since they conserve energy for the cell (Pa&#x000e9;z-Espino et al., <xref rid="B40" ref-type="bibr">2009</xref>). Another arsenic resistance mechanism that microorganisms can apply involves methylation (Bentley and Chasteen, <xref rid="B5" ref-type="bibr">2002</xref>). A study by Wallschl&#x000e4;ger and London (<xref rid="B61" ref-type="bibr">2008</xref>) detected mono- and dimethylated arsenic oxyanions in sulfidic groundwater, linking the presence of arsenic species with the activity of arsenic-methylating microorganisms. The relevance of methylated arsenic species in geothermal waters, and within the context of the evolution of arsenic resistance, is not yet well understood. To date, only a small number of thermophiles with arsenic-methylating activity have been identified (Qin et al., <xref rid="B45" ref-type="bibr">2006</xref>; Takacs-Vesbach et al., <xref rid="B59" ref-type="bibr">2013</xref>).</p><p>This study presents quantitative arsenic and sulfur speciation data, as well as cultivation-independent metagenomic analysis of microbial community structure and functional sulfur and arsenic gene inventories from Champagne Pool, an acid-sulfide hot spring at Waiotapu, New Zealand. The objectives of this work were (1) to determine potential microbial contributions to arsenic speciation in an extreme environment analogous to geothermal sites on the early Earth, (2) to characterize microbial diversity and richness at the 16S ribosomal RNA gene level across the hydrologic gradient of the pool and (3) to elucidate possible environmental constraints on the evolution of microbial arsenic resistance.</p></sec><sec sec-type="materials|methods" id="s2"><title>Materials and methods</title><sec><title>Field site</title><p>The Taup<overline>o</overline> Volcanic Zone (TVZ) consists of a complex group of high temperature geothermal systems in the central North Island of New Zealand. One of the major geothermal fields in the TVZ is Waiotapu, which is characterized by a large number of springs with elevated arsenic concentrations (Hedenquist and Henley, <xref rid="B21" ref-type="bibr">1985</xref>; Mountain et al., <xref rid="B38" ref-type="bibr">2003</xref>). The largest feature at Waiotapu is Champagne Pool, ~65 m in diameter with an estimated volume of ~50,000 m³ (Hedenquist and Henley, <xref rid="B21" ref-type="bibr">1985</xref>), and an arsenic concentration between 2.9 and 4.2 mg l^&#x02212;1 (this study). Champagne Pool is a geothermal surface feature and a source of high dissolved arsenite and sulfide concentrations (Childs et al., <xref rid="B8" ref-type="bibr">2008</xref>). The inner rim of Champagne Pool is characterized by subaqueous orange amorphous As-S precipitate (Jones et al., <xref rid="B26" ref-type="bibr">2001</xref>). The narrow outflow channel (~40 cm wide and 5 cm deep), in a subaerial sinter dam, drains the spring water out across a shallow siliceous sinter terrace. Convection in Champagne Pool stabilizes water temperatures to ~75&#x000b0;C within the pool itself, while on the surrounding silica terrace (&#x0201c;Artist's Palette&#x0201d;), the temperature decreases to ~45&#x000b0;C. Water-rock interactions beneath the pool that lead to silica dissolution and sulfide oxidation (Ellis and Mahon, <xref rid="B14" ref-type="bibr">1964</xref>) provide sources of acidity to Champagne Pool waters. A high bicarbonate concentration, however, buffers the pH to ~5.5 (Hetzer et al., <xref rid="B22" ref-type="bibr">2007</xref>). The precipitation of silica around the rim of Champagne Pool (Mountain et al., <xref rid="B38" ref-type="bibr">2003</xref>) results in increased pH values toward Artist's Palette up to 6.9, favoring the dissolution of arsenic sulfide minerals that were precipitated inside the pool and washed out (Jones et al., <xref rid="B26" ref-type="bibr">2001</xref>).</p><p>Four sampling sites at Champagne Pool were selected on the basis of distinctive physical and chemical characteristics. These sites were located along a natural hydrologic gradient from the inner pool (pool or &#x0201c;CPp&#x0201d;) through the inner rim (rim or &#x0201c;CPr&#x0201d;) and outflow channel (channel or &#x0201c;CPc&#x0201d;) on to an outer silica terrace (Artist's Palette or &#x0201c;AP&#x0201d;) (Figure <xref ref-type="fig" rid="F1">1</xref>). The AP samples were taken at a point immediately adjacent to CPc, where elemental sulfur precipitation was visible (Pope et al., <xref rid="B44" ref-type="bibr">2004</xref>).</p><fig id="F1" position="float"><label>Figure 1</label><caption><p><bold>Sampling sites (with abbreviations) at Champagne Pool, Waiotapu, New Zealand. (A)</bold> Aerial view of Champagne Pool, photo credit: courtesy of GNS Science <bold>(B)</bold> CPp, central pool; CPr, rim of pool; <bold>(C)</bold> CPc, outflow channel (40 cm wide; 5 cm deep); AP, &#x0201c;Artist's Palette&#x0201d; terrace.</p></caption><graphic xlink:href="fmicb-05-00569-g0001"/></fig></sec><sec><title>Physical and chemical parameters</title><p>The pH, temperature, redox potential and DO (dissolved oxygen) saturation were measured <italic>in situ</italic> using a Professional Plus multimeter (YSI, USA). Water samples for dissolved organic carbon (DOC) were frozen at &#x02212;20&#x000b0;C in the field and sent out for commercial analysis (Hills Laboratory, Hamilton, New Zealand), where the samples were filtered through a 0.45 &#x003bc;m nylon HPLC grade membrane filter and analyzed following the American Public Health Association APHA 5310-B Standard Method (Rice et al., <xref rid="B49" ref-type="bibr">2012</xref>). Basic cations were measured using inductively coupled plasma atomic emission spectrometry (ICP-AES) (IRIS Intrepid II XDL, Thermo Corp). Chloride was measured using the potentiometric method following the American Public Health Association APHA 3500-Cl^&#x02212; D Standard Method (Rice et al., <xref rid="B49" ref-type="bibr">2012</xref>), and total bicarbonate was measured using the HCO^&#x02212;₃ titration method following the ASTM Standards D513-82 (<xref rid="B2" ref-type="bibr">1988</xref>).</p></sec><sec><title>Sampling and storage</title><p>Water samples for arsenic speciation analysis were stored in opaque 125 ml high-density polyethylene bottles (Nalgene, USA) that were washed with 1 M HCl and rinsed three times with sterile nano-pure water (Pall Corporation, USA) before a final rinse using the sample water immediately prior to sample collection. Water samples were collected via 50 ml sterile syringes (Terumo, USA), filtered through sterile 0.22 &#x003bc;m pore size Sterivex-GP polyethersulfone syringe filters (Merck Millipore, Germany) into the sample bottles, immediately flash frozen with liquid nitrogen, and placed into anoxic bags (BD Biosciences, USA). Frozen samples were transported on dry ice to the laboratory, where they were stored at &#x02212;80&#x000b0;C until analysis. Immediately prior to arsenic speciation analysis, the samples were thawed under nitrogen in an anaerobic chamber to avoid oxidation.</p><p>Water samples for sulfur speciation and total sulfur analysis were collected via a portable peristaltic pump at 2 ml min^&#x02212;1 (Geopump Series II; Envco, Auckland, NZ). The sterile sample inlet tube made of silicon was placed directly into the sample site and the water was pumped directly from the springs into sterile polypropylene Falcon tubes (BD Biosciences, USA). The tubing was flushed thoroughly with spring water before taking samples. All samples, except those for elemental sulfur, were passed through a 0.45 &#x003bc;m pore size nylon filter (Merck Millipore, Germany) prior to collection in sterile Falcon tubes (BD Biosciences, USA). Additionally, 5% (w/v) zinc acetate (ZnAc) was added to the elemental sulfur samples in a 10:1 (v/v) ratio of sample:ZnAc to induce precipitation of (and thereby remove) zinc-sulfide from the sample. All sulfur samples, except the sulfide and total sulfur samples, were immediately frozen in liquid nitrogen and transported on dry ice to the laboratory, where they were stored at &#x02212;80&#x000b0;C until analysis.</p><p>Sediment and water for DNA sequencing from each sample site except CPp were collected and stored in sterile polypropylene Falcon tubes (BD Biosciences, USA). Water from CPp was collected in a 5 l sterilized polypropylene vessel and immediately transported back to the laboratory with no temperature control. Approximately 500 ml volumes of CPp water were then filtered through a sterile 0.22 &#x003bc;m pore size cellulose membrane filter (Merck Millipore, Germany), collected and dried at room temperature on sterile petri dishes. The Falcon tubes and petri dishes with the sediment samples were stored at &#x02212;80&#x000b0;C until further analysis.</p></sec><sec><title>Preparation of standards</title><p>Stock solutions of arsenite, arsenate, methylarsonic acid (MA) and dimethylarsinic acid (DMA) were prepared as standards (1000 mg l^&#x02212;1) by dissolving sodium arsenite, sodium arsenate heptahydrate (AJAX Laboratory Chemicals), disodium monomethylarsenic and sodium dimethylarsenic (Sigma-Aldrich, Australia), respectively, in deionised water (Sartorius, Germany). Sodium monothioarsenate (Na₃AsO₃S^*7H₂O), sodium dithioarsenate (Na₃AsO₂S₂^*H₂O), sodium trithioarsenate (Na₃AsOS₃^*10H₂O) and sodium tetrathioarsenate (Na₃AsS₄^*8H₂O) were synthesized in the lab using published protocols (Schwedt and Rieckhoff, <xref rid="B52" ref-type="bibr">1996</xref>). Monomethylmonothioarsenate (MTMA) was synthesized by adding a saturated sulfide solution (deionized water purged with H₂S for 1 h) to the monomethylarsenate (MA) standard and reacted for 30 min. Dimethylmonothioarsenate (MTDMA) was synthesized using the protocol by Raml et al. (<xref rid="B47" ref-type="bibr">2006</xref>). All thioarsenate standards were stored under nitrogen at 4&#x000b0;C. For the thiosulfate standard, 0.05 g of sodium thiosulfate (Na₂S₂O₃^*5H₂O) was dissolved in 50 ml deionized water (Sartorius, USA) to obtain a thiosulfate concentration of 1000 mg l^&#x02212;1. For the sulfate standard, 0.1 g of sodium sulfate (NaSO₄^*10H₂O) was dissolved in 200 ml deionized water (Sartorius, USA) to obtain a sulfate concentration of 500 mg l^&#x02212;1.</p></sec><sec><title>Total arsenic and arsenic speciation analysis</title><p>Samples were thawed in a glove box filled with nitrogen gas. N₂-purged deionized water (Sartorius, USA) was used to dilute samples when necessary. Total arsenic concentrations in water samples were measured in triplicate by electrothermal atomic absorption spectroscopy with a Perkin Elmer AAnalyst 600 graphite furnace using a previously published protocol (Deaker and Maher, <xref rid="B10" ref-type="bibr">1999</xref>) with optimum concentrations of Pd/Mg [0.15 &#x003bc;mol (Pd) + 0.4 &#x003bc;mol (Mg)].</p><p>Arsenic speciation was measured using high-performance liquid chromatography coupled with inductively coupled plasma mass spectrometry (HPLC-ICPMS). Arsenic oxyanions were measured using a PEEK PRP-X100 anion exchange column (250 mm &#x000d7; 4.6 mm, 10 &#x003bc;m) (Phenomenex, USA). The mobile phase consisted of 20 mM ammonium phosphate buffer at pH 5.6, a flow rate of 1.5 ml min^&#x02212;1, column temperature of 40&#x000b0;C and injection volume of 40 &#x003bc;l (Kirby et al., <xref rid="B27" ref-type="bibr">2004</xref>). Arsenic thioanions were measured using a 4 mm IonPac AG16 Guard and AS16 Analytical Column (Dionex, Sunnyvale, CA, USA) eluted with a NaOH gradient (1&#x02013;100 mM) at 25&#x000b0;C and using a flow rate of 1 ml min^&#x02212;1 (Maher et al., <xref rid="B36" ref-type="bibr">2013</xref>).</p></sec><sec><title>Total sulfur and sulfur speciation analysis</title><p>Samples for sulfide analysis were fixed in the field, using the methylene blue method following the American Public Health Association APHA 3500-S2-D Standard Method (Rice et al., <xref rid="B49" ref-type="bibr">2012</xref>). A volume of 50 ml of filtered sample from each site was collected and 1 ml of 1% (w/v) ZnAc solution (1 g dissolved in 100 ml degassed water) was added following three drops of 20 mM N,N'-dimethyl-p-phenylenediamine sulfate solution (7.4094 mg dissolved in 1 ml of 7.2 mM HCl). After 3 min incubation, 1.6 ml of 30 mM FeCl₃ solution (4.866 mg dissolved in 1 ml of 1.2 mM HCl) was added. After returning to the laboratory, sulfide samples were measured using an UV/VIS spectrophotometer (Lambda 35 UV-Vis Spectrometer, Perkin Elmer) at extinction of 665 nm and with a detection limit of 0.01 mg sulfide kg^&#x02212;1. In order to be able to measure sulfide, the samples were diluted to within the standards concentration range of 0.04&#x02013;1.5 mg l^&#x02212;1. In the laboratory, total sulfur samples were bubbled with oxygen for 15 min to oxidize all dissolved sulfur species into sulfate. The total sulfur concentration was measured using inductively coupled plasma spectrometry atomic emission spectroscopy (ICP-AES, IRIS Intrepid II XDL, Thermo Corp) using the American Public Health Association APHA 3120-B Standard Method (Rice et al., <xref rid="B49" ref-type="bibr">2012</xref>). Validation of the results was obtained by the use of a certified quality control sample obtained from the New Zealand Accreditation Institute (IANZ). The thiosulfate, sulfate and elemental sulfur samples were thawed under nitrogen before analysis. Sulfate and thiosulfate concentrations were measured using HPLC UV spectrometry at 256 nm under the same conditions as described for the arsenic speciation. Prior to the elemental sulfur analysis, the elemental sulfur was extracted from the sample by shaking 40 ml of the samples with 5 ml toluene for 16 h, which dissolves at least 50 mg l^&#x02212;1 sulfur. After shaking, the toluene was withdrawn with a rubber-free syringe and filtered with a solvent-tolerant 0.2 &#x003bc;m pore size filter into a sterile 50 ml Falcon tube (BD Biosciences, USA) that was sent for commercial analysis (Geoscience Department, Southern Cross University, NSW, Australia). This method only extracts elemental sulfur, with sulfate and sulfide remaining in the water. The elemental sulfur samples were run on a HPLC reversed-phase silica column (Acclaim 120, Dionex). Methanol (95%) was used as the mobile phase at a flow rate of 1.6 ml min^&#x02212;1.</p></sec><sec><title>SEM imaging</title><p>Environmental Scanning Electron Microscope (ESEM) photomicrographs of the precipitates observed at and collected from sites CPr and CPc were obtained with a FEI Quanta Scanning Electron Microscope (Bio21 Institute, University of Melbourne, VIC, Australia). Prior to analysis, the samples were centrifuged at 10,000 rpm for 10 min and excess water was decanted. The samples were stored at &#x02212;20&#x000b0;C until analysis. Thawed samples were attached to sample holders and transferred to the ESEM chamber for microscopy under 0.8 mbar pressure.</p></sec><sec><title>DNA extraction and quantification</title><p>The PowerSoil&#x000ae; DNA Isolation Kit (Mo Bio Laboratories, Carlsbad, USA) was used to extract total genomic DNA (gDNA) from the microbial communities in the sediments according to manufacturer protocol. The extracted DNA was stored at &#x02212;20&#x000b0;C before further use. A NanoDrop 1000 Spectrophotometer (Thermo Fisher Scientific, USA) was used for the DNA quality determination at a wavelength ratio of A260/A280.</p></sec><sec><title>Genomic DNA preparation for metagenomic sequencing</title><p>The gDNA was quantified using the Qubit dsDNA BR (Molecular probes&#x000ae;) assay system following manufacturer's protocol. Samples of sufficient quality were processed using the Illumina's Nextera XT sample preparation kit to generate Clean Amplified Nextera Tagment Amplicons (CAN) following manufacturer's protocol. CAN-DNA concentrations were checked using the Qubit dsDNA High sensitivity kit, while DNA fragment sizes were validated and quantified using the Agilent 2100 Bioanalyzer and Agilent high sensitivity DNA kit. The dilution factors for each sample library to obtain correct concentrations for sequencing on the MiSeq Sequencer were as follows: For library size of 250 bp from bioanalyzer the conversion factor for ng &#x003bc;l^&#x02212;1 to nM is 1 ng &#x003bc;l^&#x02212;1 = 6 nM, for library size of 500 bp from bioanalyzer the conversion factor for ng &#x003bc;l^&#x02212;1 to nM is 1 ng &#x003bc;l^&#x02212;1 = 3 nM, and for library size of 1000&#x02013;1500 bp from bioanalyzer the conversion factor for ng &#x003bc;l^&#x02212;1 to nM is 1 ng &#x003bc;l^&#x02212;1 = 1.5 nM. Samples were diluted using Qiagen's EB (elution buffer) instead of Tris-Cl 10 mM 0.1% Tween 20.</p></sec><sec><title>Illumina Miseq sequencing</title><p>Samples were processed for sequencing using the Illumina MiSeq reagent kit v2 (500 cycle) following a modified manufacturer's protocol. The modifications included: 1% (v/v) spike-in ratio of PhiX, the denatured DNA was diluted to a final concentration of 12 pM with pre-chilled HT1 buffer and Qiagen's EB solution instead of Tris-Cl 10 mM 0.1% (v/v) Tween 20 was used to dilute sequencing libraries and phiX throughout the protocol. Metagenomic sequencing was performed using the Illumina MiSeq machine (Peter Doherty Institute for Infection and Immunity, University of Melbourne, Australia).</p></sec><sec><title>Metagenomic analysis</title><p>Sequence analysis was performed using the rapid annotation subsystems technology for metagenomes (MG-RAST) bioinformatics package, which is publicly available through <ext-link ext-link-type="uri" xlink:href="http://metagenomics.anl.gov">http://metagenomics.anl.gov</ext-link>. Preprocessing steps included the removal of artificial sequences generated by sequencing artifacts (Gomez-Alvarez et al., <xref rid="B19" ref-type="bibr">2009</xref>), and filtering any reads from the library that mapped to the <italic>Homo sapiens</italic> genome using Bowtie (Langmead et al., <xref rid="B29" ref-type="bibr">2009</xref>). Furthermore, sequences were trimmed to contain at most five bases below a Phred score of 15, which was considered to be the lowest quality score included as a high-quality base. The maximum allowed number of ambiguous base pairs per sequence read was set to five. The numbers of sequences obtained by Illumina MiSeq sequencing was 3,843,368 for CPp; 3,146,467 for CPr; 2,926,799 for CPc and 4,623,251 for AP. The number of sequences after MG-RAST quality filtering was 2,461,097 for CPp; 2,392,176 for CPr, 2,176,985 for CPc and 3,969,176 for AP. MG-RAST used the SEED microbial genome annotation platform to determine the protein encoding genes of a metagenome via BLASTX. Sets of sequences were compared by grouping sets of annotations into higher-level functional groups. For taxonomic analysis, 16S rRNA gene sequence data were compared to all accessory databases (e.g., GREENGENES, RDP-II, etc.) by using search criteria specific for each database. Comparative analysis tools integrated into the MG-RAST pipeline were used to build rarefaction curves from 16S rRNA gene sequences detected at the Champagne Pool sites in order to investigate species richness. Graphics were generated with the R graphic program (R Development Core Team, <xref rid="B48" ref-type="bibr">2013</xref>). The Fisher's exact test of independence was applied to functional gene distributions with the purpose of identifying significant differences (<italic>p</italic> &#x0003c; 0.05) in gene proportions from one Champagne Pool site to another.</p></sec></sec><sec sec-type="results" id="s3"><title>Results</title><sec><title>Water chemistry</title><p>The four sample sites at Champagne Pool showed similar physical and chemical conditions (Table <xref ref-type="table" rid="T1">1</xref>). At sites CPp, CPr, and CPc, pH ranged between 5.5 and 5.8; while at site AP, pH increased to 6.9. All sites exhibited Eh values of ~ &#x02212;117 to &#x02212;15 mV (relative to the standard hydrogen electrode). The stable temperature in Champagne Pool (75&#x000b0;C) decreased toward the rim of the pool to 68&#x000b0;C and at Artist's Palette to 45&#x000b0;C. The DO saturation increased toward the margin of the pool to 45% at CPc. Dissolved organic carbon (DOC) concentrations declined below the detection limit of 0.5 mg l^&#x02212;1 at CPc. Dissolved iron concentrations were under the detection limit of 0.08 mg l^&#x02212;1 at all sites, and magnesium and aluminum were detected at concentrations &#x02264;0.061 mg l^&#x02212;1 and &#x02264; 0.24 mg l^&#x02212;1 respectively (Table <xref ref-type="supplementary-material" rid="SM1">S1</xref>). In Champagne Pool the silica (as silicon) concentration was measured at ~490 mg l^&#x02212;1 and the bicarbonate (HCO^&#x02212;₃) concentration was 127 mg l^&#x02212;1 (Table <xref ref-type="supplementary-material" rid="SM1">S1</xref>).</p><table-wrap id="T1" position="float"><label>Table 1</label><caption><p><bold>Temperature, pH, dissolved oxygen (DO) saturation, redox potential (Eh) and dissolved organic carbon (DOC) concentration in Champagne Pool</bold>.</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1"><bold>Site ID</bold></th><th align="center" valign="top" rowspan="1" colspan="1"><bold>CPp</bold></th><th align="center" valign="top" rowspan="1" colspan="1"><bold>CPr</bold></th><th align="center" valign="top" rowspan="1" colspan="1"><bold>CPc</bold></th><th align="center" valign="top" rowspan="1" colspan="1"><bold>AP</bold></th></tr></thead><tbody><tr><td align="left" valign="top" rowspan="1" colspan="1">Site description</td><td align="center" valign="top" rowspan="1" colspan="1">Central Champagne Pool</td><td align="center" valign="top" rowspan="1" colspan="1">Rim Champagne Pool</td><td align="center" valign="top" rowspan="1" colspan="1">Channel Champagne Pool</td><td align="center" valign="top" rowspan="1" colspan="1">Terrace &#x0201c;Artist's Palette&#x0201d;</td></tr><tr><td align="left" rowspan="1" colspan="1">Image</td><td align="center" colspan="2" rowspan="1"><inline-graphic xlink:href="fmicb-05-00569-i0003.jpg"/></td><td align="center" colspan="2" rowspan="1"><inline-graphic xlink:href="fmicb-05-00569-i0004.jpg"/></td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">Temperature (&#x000b0;C) (&#x000b1;0.2&#x000b0;C)</td><td align="center" valign="top" rowspan="1" colspan="1">75</td><td align="center" valign="top" rowspan="1" colspan="1">68</td><td align="center" valign="top" rowspan="1" colspan="1">75</td><td align="center" valign="top" rowspan="1" colspan="1">45</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">pH (&#x000b1;0.2 units)</td><td align="center" valign="top" rowspan="1" colspan="1">5.5</td><td align="center" valign="top" rowspan="1" colspan="1">5.5</td><td align="center" valign="top" rowspan="1" colspan="1">5.8</td><td align="center" valign="top" rowspan="1" colspan="1">6.9</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">Redox potential (mV) (&#x000b1;20 mV)</td><td align="center" valign="top" rowspan="1" colspan="1">&#x02212;117</td><td align="center" valign="top" rowspan="1" colspan="1">&#x02212;75</td><td align="center" valign="top" rowspan="1" colspan="1">&#x02212;74</td><td align="center" valign="top" rowspan="1" colspan="1">&#x02212;15</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">Dissolved oxygen (%) (&#x000b1;2%)</td><td align="center" valign="top" rowspan="1" colspan="1">15</td><td align="center" valign="top" rowspan="1" colspan="1">20</td><td align="center" valign="top" rowspan="1" colspan="1">45</td><td align="center" valign="top" rowspan="1" colspan="1">35</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">Dissolved organic carbon (mg l^&#x02212;1) (&#x000b1;5%)</td><td align="center" valign="top" rowspan="1" colspan="1">2.2</td><td align="center" valign="top" rowspan="1" colspan="1">4.1</td><td align="center" valign="top" rowspan="1" colspan="1">&#x0003c;0.5</td><td align="center" valign="top" rowspan="1" colspan="1">5.9</td></tr></tbody></table></table-wrap></sec><sec><title>Total arsenic concentrations and arsenic speciation</title><p>Total dissolved arsenic concentrations of 3.0, 2.9, 3.6, and 4.2 mg l^&#x02212;1 were measured at sites CPp, CPr, CPc and AP, respectively (Table <xref ref-type="table" rid="T2">2</xref>). The sum of arsenic species concentrations showed &#x02264;10% difference from the total arsenic concentration at each Champagne Pool site (Table <xref ref-type="table" rid="T2">2</xref> and Figure <xref ref-type="fig" rid="F2">2</xref>). Changes in arsenic speciation occurred, however, along the sampling gradient at Champagne Pool (Table <xref ref-type="table" rid="T2">2</xref>, Figure <xref ref-type="fig" rid="F2">2</xref>, and Figure <xref ref-type="supplementary-material" rid="SM1">S1</xref>). At sites CPp, CPr and CPc, arsenite was the major As species present, at between 55 and 75% of the total dissolved arsenic concentration; while at AP, thioarsenates were the primary detected species at 55% of the total dissolved arsenic concentration. CPp and CPr showed very similar proportions of arsenic species (Figure <xref ref-type="fig" rid="F2">2</xref>). A transition in As speciation occurred at CPc; however, where arsenate concentrations were observed to increase, the organic species dimethylmonothioarsenate (MTDMA) was first detected, and trithioarsenate (TriTA) was not detected (Figure <xref ref-type="fig" rid="F2">2</xref>). At Artist's Palette, arsenate concentrations decreased and no MTDMA was detected; however, the proportions of di- and trithioarsenate increased significantly (Figure <xref ref-type="fig" rid="F2">2</xref>). All Champagne Pool sites featured di- and trithioarsenate species; but noticeably, monothioarsenate was absent at all sites.</p><table-wrap id="T2" position="float"><label>Table 2</label><caption><p><bold>Arsenic speciation and total dissolved arsenic concentrations in Champagne Pool</bold>.</p></caption><graphic xlink:href="fmicb-05-00569-i0001"/><table-wrap-foot><p><italic>CPp, central pool; CPr, rim of pool; CPc, outflow channel; AP, Artist's Palette. DTA, dithioarsenate; TriTA, trithioarsenate; MTDMA, dimethylmonothioarsenate. bdl, below detection limit (&#x0003c;0.01 &#x003bc;g l^&#x02212;1)</italic>.</p></table-wrap-foot></table-wrap><fig id="F2" position="float"><label>Figure 2</label><caption><p><bold>Arsenic speciation as a percentage of total dissolved arsenic at Champagne Pool</bold>. CPp, central pool; CPr, rim of pool; CPc, outflow channel; AP, &#x0201c;Artist's Palette&#x0201d; terrace. DTA, dithioarsenate; TriTA, trithioarsenate; MTDMA, dimethylmonothioarsenate.</p></caption><graphic xlink:href="fmicb-05-00569-g0002"/></fig></sec><sec><title>Total sulfur concentrations and sulfur speciation</title><p>All sample sites contained total dissolved sulfur concentrations between 91 and 105 mg l^&#x02212;1 (Table <xref ref-type="table" rid="T3">3</xref>). The sum of sulfur species showed recoveries of 70% (CPp), 80% (CPr and CPc), and 83% (AP) (Table <xref ref-type="table" rid="T3">3</xref> and Figure <xref ref-type="fig" rid="F3">3</xref>). A possible explanation for the observed difference between sulfur speciation and total dissolved sulfur concentrations could have been that some sulfur was present as sulfite (SO^2&#x02212;₃), which was not quantified. Sulfur speciation changed across the sampling transect (Table <xref ref-type="table" rid="T3">3</xref> and Figure <xref ref-type="fig" rid="F3">3</xref>). The highest sulfide concentration was detected at CPp (12.6 mg l^&#x02212;1), before decreasing at the other sites. Thiosulfate showed stable concentrations throughout the sites (~30 mg l^&#x02212;1), whereas sulfate reached a maximum concentration at AP (55.1 mg l^&#x02212;1). Elemental sulfur, not included in the calculation of total (dissolved) sulfur, was detected at concentrations &#x02264;1.7 mg l^&#x02212;1 at CPp, CPr and CPc, and at 34.8 mg l^&#x02212;1 at AP. Further SEM imaging of the precipitate, referred to as floc, at CPr and CPc, revealed a filamentous As-S phase and elemental sulfur colloids in a silica rich floc-sample (Figure <xref ref-type="supplementary-material" rid="SM1">S2</xref>).</p><table-wrap id="T3" position="float"><label>Table 3</label><caption><p><bold>Sulfur speciation and total dissolved sulfur concentrations in Champagne Pool</bold>.</p></caption><graphic xlink:href="fmicb-05-00569-i0002"/><table-wrap-foot><p><italic>CPp, central pool; CPr, rim of pool; CPc, outflow channel; AP, Artist's Palette</italic>.</p></table-wrap-foot></table-wrap><fig id="F3" position="float"><label>Figure 3</label><caption><p><bold>Sulfur speciation as a percentage of total dissolved sulfur concentration at Champagne Pool</bold>. CPp, central pool; CPr, rim of pool; CPc, outflow channel; AP, &#x0201c;Artist's Palette&#x0201d; terrace.</p></caption><graphic xlink:href="fmicb-05-00569-g0003"/></fig></sec><sec><title>Microbial 16S rRNA gene diversity</title><p>The richness of the microbial communities at the sample sites was characterized using bioinformatic analysis (Figure <xref ref-type="supplementary-material" rid="SM1">S3</xref>). The species abundance with increasing rarefaction sequence depths showed the highest species abundance at AP. The lowest species abundance was detected at CPc, with species abundances of CPp and CPr closer to CPp than to AP.</p><p>The archaeal 16S rRNA gene diversity at all sites was analyzed at the genus level (Figure <xref ref-type="fig" rid="F4">4</xref>). At CPp, 12% of the sequences belonged to archaea, consisting almost exclusively of <italic>Thermofilum, Sulfolobus and Pyrobaculum</italic> (together 8%). At CPr and CPc, 16S rRNA gene sequences changed from ~12% (CPp) to ~21&#x02013;28% (CPr, CPc) archaea with their members belonging mostly to genera <italic>Sulfolobus</italic>, <italic>Thermofilum, Pyrobaculum, Desulfurococcus</italic>, <italic>Thermococcus</italic>, and <italic>Staphylothermus</italic>. At AP, only 2% of the sequences belonged to archaea, with no dominant genus present.</p><fig id="F4" position="float"><label>Figure 4</label><caption><p><bold>Distribution of microbial 16S rRNA gene sequences (classified to genus level) at the four sample sites at Champagne Pool</bold>. CPp, central pool; CPr, rim of pool; CPc, outflow channel; AP, &#x0201c;Artist's Palette&#x0201d; terrace. &#x0201c;Others&#x0201d; category in legend signifies genera present at &#x0003c;1% each.</p></caption><graphic xlink:href="fmicb-05-00569-g0004"/></fig><p>The bacterial 16S rRNA gene diversity at all sites was analyzed at the genus level (Figure <xref ref-type="fig" rid="F4">4</xref>). Dominant sequences across all sites were most closely related to the genus <italic>Sulfurihydrogenibium</italic>. <italic>Sulfurihydrogenibium</italic>-related sequence abundance decreased continuously from 19% at CPp to 10% at AP. Additionally, bacterial genera <italic>Anoxybacillus</italic> and <italic>Persephonella</italic> comprised 38% and 3% of the total sequences at CPp, respectively. At CPr, members of the family <italic>Sulfurihydrogenibium</italic> were a dominant clade with 19% of the total sequences. Other major groups of bacteria from genera <italic>Persephonella</italic>, <italic>Thermodesulfovibrio</italic>, <italic>Desulfovibrio</italic>, <italic>Thermotoga</italic>, and <italic>Syntrophobacter</italic> represented 3.0%, 2.1%, 1.8%, 1.1%, and 1.0% of the total sequences, respectively. The dominant bacterial sequences at CPc were still related to <italic>Sulfurihydrogenibium</italic>, with an abundance of 13%. 16S rRNA gene sequences closely related to <italic>Persephonella</italic>, also belonging to the order <italic>Aquificales</italic>, comprised 2.2% of the total sequences. Additionally at CPc, bacteria belonging to the family <italic>Aquificaceae</italic> in genera <italic>Aquifex, Hydrogenobacter, Hydrogenivirga</italic>, and <italic>Thermocrinis</italic> represented 18% of the total sequences. Bacteria from genera <italic>Thermodesulfovibrio</italic> and <italic>Desulfovibrio</italic> comprised 3.6% of total sequences each at CPc. At AP, alongside <italic>Sulfurihydrogenibium</italic> (10% of total sequences), <italic>Thiomonas</italic> and <italic>Thermus</italic> were the most abundant bacterial genera in the community with 9% and 4% of total sequences.</p></sec><sec><title>Relative abundance of functional arsenic resistance, sulfur metabolizing and hydrogenase encoding genes</title><p>Metagenomes analyzed with MG-RAST were searched for arsenic resistance, sulfur metabolism, and H₂ respiration genes. The relative abundances of these genes are illustrated in Table <xref ref-type="table" rid="T4">4</xref>.</p><table-wrap id="T4" position="float"><label>Table 4</label><caption><p><bold>Arsenic and sulfur genes found in the metagenome of Champagne Pool sites</bold>.</p></caption><table frame="hsides" rules="groups"><thead><tr><th rowspan="1" colspan="1"/><th align="center" valign="top" rowspan="1" colspan="1"><bold>CPp</bold></th><th align="center" valign="top" rowspan="1" colspan="1"><bold>CPr</bold></th><th align="center" valign="top" rowspan="1" colspan="1"><bold>CPc</bold></th><th align="center" valign="top" rowspan="1" colspan="1"><bold>AP</bold></th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">% Arsenic resistance genes<xref ref-type="table-fn" rid="TN1">^a</xref></td><td align="center" valign="top" rowspan="1" colspan="1">(0.07)</td><td align="center" valign="top" rowspan="1" colspan="1">(0.09)</td><td align="center" valign="top" rowspan="1" colspan="1">(0.1)</td><td align="center" valign="top" rowspan="1" colspan="1">(0.1)</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1"><italic>arsB</italic></td><td align="center" valign="top" rowspan="1" colspan="1">49 (0.04)</td><td align="center" valign="top" rowspan="1" colspan="1">26 (0.02)</td><td align="center" valign="top" rowspan="1" colspan="1">16 (0.02)</td><td align="center" valign="top" rowspan="1" colspan="1">19 (0.02)</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1"><italic>acr3</italic></td><td align="center" valign="top" rowspan="1" colspan="1">0.6 (0.0004)</td><td align="center" valign="top" rowspan="1" colspan="1">15 (0.01)</td><td align="center" valign="top" rowspan="1" colspan="1">16 (0.02)</td><td align="center" valign="top" rowspan="1" colspan="1">31 (0.03)</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1"><italic>arsA</italic></td><td align="center" valign="top" rowspan="1" colspan="1">14 (0.01)</td><td align="center" valign="top" rowspan="1" colspan="1">34 (0.03)</td><td align="center" valign="top" rowspan="1" colspan="1">44 (0.05)</td><td align="center" valign="top" rowspan="1" colspan="1">15 (0.01)</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1"><italic>arsR</italic></td><td align="center" valign="top" rowspan="1" colspan="1">11 (0.008)</td><td align="center" valign="top" rowspan="1" colspan="1">8 (0.008)</td><td align="center" valign="top" rowspan="1" colspan="1">4 (0.004)</td><td align="center" valign="top" rowspan="1" colspan="1">8 (0.008)</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1"><italic>arsC</italic></td><td align="center" valign="top" rowspan="1" colspan="1">26 (0.02)</td><td align="center" valign="top" rowspan="1" colspan="1">17 (0.02)</td><td align="center" valign="top" rowspan="1" colspan="1">20 (0.02)</td><td align="center" valign="top" rowspan="1" colspan="1">25 (0.02)</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1"><italic>arsH</italic></td><td align="center" valign="top" rowspan="1" colspan="1">0.1 (0.00007)</td><td align="center" valign="top" rowspan="1" colspan="1">&#x02013;</td><td align="center" valign="top" rowspan="1" colspan="1">1 (0.001)</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1"><italic>arsD</italic></td><td align="center" valign="top" rowspan="1" colspan="1">&#x02013;</td><td align="center" valign="top" rowspan="1" colspan="1">&#x02013;</td><td align="center" valign="top" rowspan="1" colspan="1">&#x02013;</td><td align="center" valign="top" rowspan="1" colspan="1">0.5 (0.0005)</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">% Respiration genes<xref ref-type="table-fn" rid="TN2">^b</xref></td><td align="center" valign="top" rowspan="1" colspan="1">(2)</td><td align="center" valign="top" rowspan="1" colspan="1">(3)</td><td align="center" valign="top" rowspan="1" colspan="1">(4)</td><td align="center" valign="top" rowspan="1" colspan="1">(3)</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1"><italic>arr subunit A</italic></td><td align="center" valign="top" rowspan="1" colspan="1">0.002 (0.00004)</td><td align="center" valign="top" rowspan="1" colspan="1">0.005 (0.0002)</td><td align="center" valign="top" rowspan="1" colspan="1">&#x02013;</td><td align="center" valign="top" rowspan="1" colspan="1">0.03 (0.0009)</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1"><italic>arr subunit B</italic></td><td align="center" valign="top" rowspan="1" colspan="1">&#x02013;</td><td align="center" valign="top" rowspan="1" colspan="1">&#x02013;</td><td align="center" valign="top" rowspan="1" colspan="1">&#x02013;</td><td align="center" valign="top" rowspan="1" colspan="1">0.06 (0.002)</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1"><italic>arr subunit C</italic></td><td align="center" valign="top" rowspan="1" colspan="1">&#x02013;</td><td align="center" valign="top" rowspan="1" colspan="1">0.002 (0.00005)</td><td align="center" valign="top" rowspan="1" colspan="1">&#x02013;</td><td align="center" valign="top" rowspan="1" colspan="1">&#x02013;</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">Hydrogenase encoding genes</td><td align="center" valign="top" rowspan="1" colspan="1">2 (0.05)</td><td align="center" valign="top" rowspan="1" colspan="1">7 (0.2)</td><td align="center" valign="top" rowspan="1" colspan="1">7 (0.3)</td><td align="center" valign="top" rowspan="1" colspan="1">7 (0.2)</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">% Sulfur metabolizing genes<xref ref-type="table-fn" rid="TN3">^c</xref></td><td align="center" valign="top" rowspan="1" colspan="1">(0.5)</td><td align="center" valign="top" rowspan="1" colspan="1">(0.6)</td><td align="center" valign="top" rowspan="1" colspan="1">(0.7)</td><td align="center" valign="top" rowspan="1" colspan="1">(0.7)</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">Sulfur oxidation genes</td><td align="center" valign="top" rowspan="1" colspan="1">8 (0.04)</td><td align="center" valign="top" rowspan="1" colspan="1">8 (0.05)</td><td align="center" valign="top" rowspan="1" colspan="1">11 (0.07)</td><td align="center" valign="top" rowspan="1" colspan="1">24 (0.2)</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">Sulfur reduction genes</td><td align="center" valign="top" rowspan="1" colspan="1">2 (0.009)</td><td align="center" valign="top" rowspan="1" colspan="1">24 (0.1)</td><td align="center" valign="top" rowspan="1" colspan="1">24 (0.2)</td><td align="center" valign="top" rowspan="1" colspan="1">13 (0.1)</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">% Sulfur oxidation genes<xref ref-type="table-fn" rid="TN4">^d</xref></td><td align="center" valign="top" rowspan="1" colspan="1">(0.04)</td><td align="center" valign="top" rowspan="1" colspan="1">(0.05)</td><td align="center" valign="top" rowspan="1" colspan="1">(0.07)</td><td align="center" valign="top" rowspan="1" colspan="1">(0.2)</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">CcdA encoding gene</td><td align="center" valign="top" rowspan="1" colspan="1">59 (0.02)</td><td align="center" valign="top" rowspan="1" colspan="1">30 (0.01)</td><td align="center" valign="top" rowspan="1" colspan="1">21 (0.02)</td><td align="center" valign="top" rowspan="1" colspan="1">6.8 (0.01)</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1"><italic>sox</italic> gene complex (without <italic>soxCD</italic>)</td><td align="center" valign="top" rowspan="1" colspan="1">0.3 (0.0001)</td><td align="center" valign="top" rowspan="1" colspan="1">5.7 (0.003)</td><td align="center" valign="top" rowspan="1" colspan="1">63 (0.05)</td><td align="center" valign="top" rowspan="1" colspan="1">47 (0.08)</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1"><italic>soxCD</italic></td><td align="center" valign="top" rowspan="1" colspan="1">&#x02013;</td><td align="center" valign="top" rowspan="1" colspan="1">&#x02013;</td><td align="center" valign="top" rowspan="1" colspan="1">&#x02013;</td><td align="center" valign="top" rowspan="1" colspan="1">15 (0.03)</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">Sulfide dehydrogenase encoding gene</td><td align="center" valign="top" rowspan="1" colspan="1">10 (0.004)</td><td align="center" valign="top" rowspan="1" colspan="1">25 (0.01)</td><td align="center" valign="top" rowspan="1" colspan="1">4 (0.003)</td><td align="center" valign="top" rowspan="1" colspan="1">7 (0.01)</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">Sulfur oxygenase- reductase/Sulfite oxygenase encoding genes</td><td align="center" valign="top" rowspan="1" colspan="1">30 (0.01)</td><td align="center" valign="top" rowspan="1" colspan="1">37 (0.02)</td><td align="center" valign="top" rowspan="1" colspan="1">12 (0.008)</td><td align="center" valign="top" rowspan="1" colspan="1">8 (0.02)</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1">% Sulfur reduction genes<xref ref-type="table-fn" rid="TN5">^e</xref></td><td align="center" valign="top" rowspan="1" colspan="1">(0.009)</td><td align="center" valign="top" rowspan="1" colspan="1">(0.1)</td><td align="center" valign="top" rowspan="1" colspan="1">(0.2)</td><td align="center" valign="top" rowspan="1" colspan="1">(0.1)</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1"><italic>dsrMKJOP</italic> gene complex</td><td align="center" valign="top" rowspan="1" colspan="1">72.8 (0.006)</td><td align="center" valign="top" rowspan="1" colspan="1">62 (0.1)</td><td align="center" valign="top" rowspan="1" colspan="1">63 (0.1)</td><td align="center" valign="top" rowspan="1" colspan="1">59 (0.05)</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1"><italic>dsrC</italic></td><td align="center" valign="top" rowspan="1" colspan="1">2 (0.0001)</td><td align="center" valign="top" rowspan="1" colspan="1">7 (0.01)</td><td align="center" valign="top" rowspan="1" colspan="1">6 (0.01)</td><td align="center" valign="top" rowspan="1" colspan="1">5 (0.005)</td></tr><tr><td align="left" valign="top" rowspan="1" colspan="1"><italic>asrAB</italic></td><td align="center" valign="top" rowspan="1" colspan="1">20 (0.002)</td><td align="center" valign="top" rowspan="1" colspan="1">20 (0.03)</td><td align="center" valign="top" rowspan="1" colspan="1">21 (0.04)</td><td align="center" valign="top" rowspan="1" colspan="1">22 (0.02)</td></tr></tbody></table><table-wrap-foot><p><italic>arsB, arsenic efflux pump encoding gene; acr3, homolog to arsB; arsA, arsenical pump ATPase encoding gene; arsR, arsenic resistance repressor encoding gene; arsH, arsenite resistance enhancer encoding gene; arsD, metallochaperone transferase encoding gene; arr, respiratory arsenate reductase encoding gene; CcdA, protein for biogenesis of cytochrome c-type proteins; sox, sulfur oxidation genes; dsr, dissimilatory sulfite reductase encoding genes; asr, anaerobic sulfite reductase encoding gene. CPp, central pool; CPr, rim of pool; CPc, outflow channel; AP, Artist's Palette. Numbers in parentheses represent the % of the individual functional gene or gene complex on the total number of annotated genes in the microbial community at the given site</italic>.</p><fn id="TN1"><label>a</label><p><italic>The percentage of the individual arsenic resistance gene on the overall arsenic resistance genes</italic>.</p></fn><fn id="TN2"><label>b</label><p><italic>The percentage of the arr arsenate reductase and hydrogenase genes on the overall respiration genes</italic>.</p></fn><fn id="TN3"><label>c</label><p><italic>The percentage of the sulfur oxidation and sulfur reduction genes on the overall sulfur metabolizing genes</italic>.</p></fn><fn id="TN4"><label>d</label><p><italic>The percentage of the individual sulfur oxidation gene or gene complex on the overall sulfur oxidation genes</italic>.</p></fn><fn id="TN5"><label>e</label><p><italic>The percentage of the individual sulfur reduction gene or gene complex on the overall sulfur reduction genes</italic>.</p></fn></table-wrap-foot></table-wrap><p>Fisher's exact tests revealed significant differences in the proportion of <italic>arsB</italic> vs. <italic>acr3</italic> from CPp to all other sites (<italic>p</italic> = 0.00001), and CPr to AP (<italic>p</italic> = 0.02). Similarly, differences in the proportion of sulfur metabolism genes (sulfur oxidation to sulfur reduction or vice versa) were significant for CPp to CPr (<italic>p</italic> = 0.003), CPp to CPc (<italic>p</italic> = 0.01), CPr to AP (<italic>p</italic> = 0.002) and CPc to AP (<italic>p</italic> = 0.005).</p></sec></sec><sec sec-type="discussion" id="s4"><title>Discussion</title><sec><title>Geochemical influences on arsenic speciation</title><p>As a result of low or absent potential adsorbents such as the ox(yhydrox)ides of iron, magnesium and aluminum (Table <xref ref-type="supplementary-material" rid="SM1">S1</xref>), dissolved arsenic can accumulate in Champagne Pool waters to high concentrations (Table <xref ref-type="table" rid="T2">2</xref>). Arsenic speciation in Champagne Pool is influenced by changes in the physiochemical conditions along the hydraulic gradient from the inner pool (CPp) to the Artist's Palette (AP) (Table <xref ref-type="table" rid="T1">1</xref>). Analysis of arsenic and sulfur speciation in the context of Eh-pH stability fields (Lu and Zhu, <xref rid="B32" ref-type="bibr">2011</xref>) predicts a predominance of arsenite at all sites, consistent with the relative percentage of arsenite measured at CPp, CPr, and CPc (Figure <xref ref-type="fig" rid="F2">2</xref>). HPLC-ICPMS analyses, however, showed significant thioarsenate concentrations at all sites, and furthermore revealed an organic arsenic species, dimethylmonothioarsenate, present at CPc. These observations were not consistent with the arsenic speciation predicted from redox potential and pH. Thus, both geochemical and microbial community analyses are required to understand arsenic transformations in Champagne Pool.</p></sec><sec><title>Arsenic and sulfur cycling</title><p>Hot springs with high sulfide or elemental sulfur concentrations contain higher proportions of thioarsenates, as sulfur has a higher affinity than oxygen for arsenic (Sharma and Sohn, <xref rid="B54" ref-type="bibr">2009</xref>). All Champagne Pool sites contained significant amounts of thioarsenates, namely di-, and trithioarsenate (Figure <xref ref-type="fig" rid="F2">2</xref>). High sulfide concentrations at CPp (12.6 mg l^&#x02212;1) promote the transformation of arsenite into monothioarsenate and hydrogen (Figure <xref ref-type="fig" rid="F5">5</xref>). Planer-Friedrich et al. (<xref rid="B43" ref-type="bibr">2010</xref>) describe the arsenite transformation into thioarsenates via sulfide. No monothioarsenate was detected, however, as excess sulfide can further transform monothioarsenate into di- and trithioarsenate (Figure <xref ref-type="fig" rid="F5">5</xref>). The decrease in sulfide concentration from CPp to CPr is consistent with the observed increase in DO % saturation (Table <xref ref-type="table" rid="T1">1</xref>) and loss of sulfide to precipitation (Figure <xref ref-type="supplementary-material" rid="SM1">S2</xref>). Lower trithioarsenate concentrations in CPr compared to CPp can be attributed to a slightly lower sulfide concentration (Table <xref ref-type="table" rid="T3">3</xref>).</p><fig id="F5" position="float"><label>Figure 5</label><caption><p><bold>Sulfide-dependent arsenic cycle at the pool (CPp) of Champagne Pool</bold>. H₃AsO₃S, monothioarsenate; H₃AsO₂S₂, dithioarsenate; H₃AsOS₃, trithioarsenate.</p></caption><graphic xlink:href="fmicb-05-00569-g0005"/></fig><p>The arsenite transformation process in CPr and CPc may involve the activity of observed elemental sulfur (Figure <xref ref-type="supplementary-material" rid="SM1">S2</xref>), which transforms arsenite into monothioarsenate and H⁺ ions instead of H₂ (Figure <xref ref-type="fig" rid="F6">6</xref>). Stauder et al. (<xref rid="B56" ref-type="bibr">2005</xref>) suggested the transformation of arsenite into thioarsenates via elemental sulfur. Further transformation of monothioarsenate into dithioarsenate would presumably be carried out via addition of sulfide, as abundant sulfide was also measured (Table <xref ref-type="table" rid="T3">3</xref>). We estimate, based on the observations that similar sulfide concentrations were measured at CPr and CPc, but no trithioarsenate was detected at CPc, that the transformation from di- to trithioarsenate occured at a threshold of ~10 mg l^&#x02212;1 for dissolved sulfide (Table <xref ref-type="table" rid="T3">3</xref>). Although relatively low elemental sulfur concentrations of 1.1 and 1.7 mg l^&#x02212;1 were detected in the water column of CPr and CPc, respectively, evidence for localized enrichment of S⁰ was found on filamentous precipitate detected at the rim and in the outflow channel (Figure <xref ref-type="supplementary-material" rid="SM1">S2</xref>). The shallow character of CPr and the increased DO saturation of 45% in CPc presumably facilitate the oxidation of sulfide into elemental sulfur (Xu et al., <xref rid="B66" ref-type="bibr">1998</xref>, <xref rid="B67" ref-type="bibr">2000</xref>).</p><fig id="F6" position="float"><label>Figure 6</label><caption><p><bold>Elemental sulfur-dependent arsenic cycle at the rim (CPr) and outflow channel (CPc) of Champagne Pool</bold>. H₃AsO₃S, monothioarsenate; H₃AsO₂S₂, dithioarsenate; H₃AsOS₃, trithioarsenate.</p></caption><graphic xlink:href="fmicb-05-00569-g0006"/></fig><p>Low sulfide concentrations of 2.2 mg l^&#x02212;1, and the increase of pH to 6.9 at AP, resulted in transformation via elemental sulfur to monothioarsenate and H₂O (Figure <xref ref-type="fig" rid="F7">7</xref>). Furthermore, the pH increase at AP favors the dissolution of arsenic-sulfide precipitates detected at CPr and CPp (Figure <xref ref-type="supplementary-material" rid="SM1">S2</xref>), and conversion of As into di- and trithioarsenate (Figure <xref ref-type="fig" rid="F7">7</xref>), which explains their predominance adding to higher total dissolved arsenic concentrations (Table <xref ref-type="table" rid="T2">2</xref>), despite low sulfide concentrations at AP. This interpretation is consistent with the observed higher solubility of arsenic-sulfide precipitates at higher pH (Webster, <xref rid="B63" ref-type="bibr">1990</xref>; Eary, <xref rid="B12" ref-type="bibr">1992</xref>). Elevated sulfur concentration in the water column of AP (Table <xref ref-type="table" rid="T3">3</xref>) could be a result of elemental sulfur release during dissolution of the As-S precipitate (Figure <xref ref-type="supplementary-material" rid="SM1">S2</xref>).</p><fig id="F7" position="float"><label>Figure 7</label><caption><p><bold>Elemental sulfur-dependent arsenic cycle and dissolution of arsenic-sulfide precipitate at Artist's Palette (AP)</bold>. H₃AsO₃S, monothioarsenate; H₃AsO₂S₂, dithioarsenate; H₃AsOS₃, trithioarsenate.</p></caption><graphic xlink:href="fmicb-05-00569-g0007"/></fig></sec><sec><title>Microbial influences on the arsenic and sulfur cycle</title><p>As expected, microbial community analysis revealed increasing richness with decreasing temperatures and increasing pH (Figure <xref ref-type="supplementary-material" rid="SM1">S3</xref>). Community 16S rRNA gene sequences included a large group of sulfur-cycling microorganisms along the sampling transect at Champagne Pool, an observation which presents direct evidence to address previous hypotheses about the potential for indirect biological mediation of arsenic speciation via microbial sulfur cycling (Ulrich et al., <xref rid="B60" ref-type="bibr">2013</xref>).</p><p>Despite reducing conditions at CPp (Table <xref ref-type="table" rid="T1">1</xref>), thiosulfate and sulfate concentrations were elevated relative to sulfide (Table <xref ref-type="table" rid="T3">3</xref>). The metagenome of the pool contained nearly 0.04% sulfur oxidation genes, in contrast with 0.009% sulfur reduction genes of the total annotated genes (Table <xref ref-type="table" rid="T4">4</xref>), potentially explaining this observation. The combination of sulfide dehydrogenase and sulfur oxygenase-reductase encoding genes, detected as major sulfur oxidation genes at CPp (Table <xref ref-type="table" rid="T4">4</xref>), suggests a two-step sulfide oxidation process to sulfite and thiosulfate, also producing sulfide. Thiosulfate could be further oxidized via oxygen at the surface of the pool. The annotation of sulfur oxidation genes detected at CPp (Table <xref ref-type="table" rid="T4">4</xref>) to gene sequences from close relatives of detected 16S rRNA gene sequences (Figure <xref ref-type="fig" rid="F4">4</xref>) revealed a close relationship to genus <italic>Sulfolobus</italic>, which was detected in abundance at CPp (Figure <xref ref-type="fig" rid="F4">4</xref>). This sulfur-oxidizing genus (Brock et al., <xref rid="B7" ref-type="bibr">1972</xref>) enhances the potential for production of thiosulfate and sulfate (Table <xref ref-type="table" rid="T3">3</xref>). Other 16S rRNA gene sequences detected in the pool (CPp) were closely related to members of the order <italic>Aquificales</italic>, primary producers in high temperature ecosystems (Eder and Huber, <xref rid="B13" ref-type="bibr">2002</xref>) and capable of oxidizing H₂ or reduced sulfur species. The chemolithoautotrophic <italic>Sufurihydrogenibium</italic>, a genus of <italic>Aquificales</italic>, was a dominant phylotype detected along the gradient from CPp to AP (Figure <xref ref-type="fig" rid="F4">4</xref>). Although the relative abundance of <italic>Sulfurihydrogenibium</italic> decreased from CPp to AP, the genus remained a major component of the microbial communities at CPr, CPc, and AP (Figure <xref ref-type="fig" rid="F4">4</xref>). With only a few strains capable of oxidizing H₂, the majority of this genus oxidizes S⁰ or S₂O^2&#x02212;₃ with O₂ as the electron acceptor (Flores et al., <xref rid="B15" ref-type="bibr">2008</xref>). <italic>Sulfurihydrogenibium</italic> as well as <italic>Persephonella</italic>, another member of the order <italic>Aquificales</italic>, which were the closest relatives of sequences detected at CPp (Figure <xref ref-type="fig" rid="F4">4</xref>), however, showed similarities in hydrogenase encoding genes (responsible for H₂-oxidation) to those detected at this site (Table <xref ref-type="table" rid="T4">4</xref>). The presence of sulfur reduction genes at CPp (Table <xref ref-type="table" rid="T4">4</xref>), belonging to the <italic>dsr</italic> and <italic>asr</italic> gene complexes, is consistent with the thiosulfate- or elemental sulfur-reducing genus <italic>Pyrobaculum</italic> (Stetter et al., <xref rid="B58" ref-type="bibr">1990</xref>) detected at the 16S rRNA gene level (Figure <xref ref-type="fig" rid="F4">4</xref>). This was further confirmed during the annotation of detected sulfur reduction genes (Table <xref ref-type="table" rid="T4">4</xref>) with genes of close relatives from 16S rRNA gene sequences detected at CPp (Figure <xref ref-type="fig" rid="F4">4</xref>), which revealed a close relationship to representatives of genera <italic>Pyrobaculum</italic> and <italic>Anoxybacillus</italic> for the <italic>dsr</italic> and <italic>asrAB</italic> gene complexes. The resulting biogenic sulfide produced would then be available to transform arsenite to monothioarsenate (also yielding H₂) (Figure <xref ref-type="fig" rid="F5">5</xref>).</p><p>At the rim of Champagne Pool, the increase in sulfur reduction genes (Table <xref ref-type="table" rid="T4">4</xref>), belonging to the <italic>dsr</italic> and <italic>asr</italic> gene complexes, is consistent with the detection of close relatives from the genera <italic>Thermofilum</italic>, <italic>Pyrobaculum</italic>, <italic>Desulforococcus</italic>, <italic>Staphylothermus</italic>, <italic>Thermococcus</italic>, and <italic>Thermotoga</italic>, which have the potential to reduce thiosulfate or elemental sulfur to sulfide (Janssen and Morgan, <xref rid="B25" ref-type="bibr">1992</xref>). Additionally, relatives of the observed genera <italic>Thermodesulfovibrio</italic>, <italic>Desulfovibrio</italic>, and <italic>Syntrophobacter</italic> can reduce sulfate to sulfide (Sekiguchi et al., <xref rid="B53" ref-type="bibr">2008</xref>). The annotation of the detected sulfur reduction genes <italic>dsr</italic> and <italic>asr</italic> at CPr (Table <xref ref-type="table" rid="T4">4</xref>) to genomic data from close relatives of 16S rRNA gene sequences detected at CPr (Figure <xref ref-type="fig" rid="F4">4</xref>) revealed a close relationship with genera <italic>Pyrobaculum</italic>, <italic>Thermotoga</italic>, <italic>Desulfovibrio</italic>, <italic>Thermodesulfovibrio</italic>, and <italic>Syntrophobacter</italic> for <italic>dsr</italic> and <italic>asr</italic>. Since the sulfide concentration at the rim did not increase significantly (Table <xref ref-type="table" rid="T3">3</xref>), biogenic sulfide was probably rapidly reoxidized via microbial sulfur oxidation (e.g., <italic>Sulfolobus</italic>, detected at CPr; Figure <xref ref-type="fig" rid="F4">4</xref>). The annotation of major sulfur oxidation genes detected at CPr (Table <xref ref-type="table" rid="T4">4</xref>), sulfide dehydrogenase and sulfur oxygenase-reductase encoding genes, with genes of close relatives of 16S rRNA genes detected at CPr (Figure <xref ref-type="fig" rid="F4">4</xref>) revealed a close relationship to members of the genus <italic>Sulfolobus</italic>. Alternatively, atmospheric oxygen could be responsible for the reoxidation of sulfide or biogenic sulfide could react with arsenite to form the ubiquitous orange arsenic-sulfide precipitates found around the rim of Champagne Pool (Figure <xref ref-type="supplementary-material" rid="SM1">S2</xref>). 16S rRNA gene sequences closely related to members of genera <italic>Sulfurihydrogenibium</italic> and <italic>Persephonella</italic> were dominant at CPr (Figure <xref ref-type="fig" rid="F4">4</xref>), and these genera most likely use H₂ instead of reduced sulfur species as their electron donor, as supported by annotation of detected functional genes (Table <xref ref-type="table" rid="T4">4</xref>) to genomic data of close relatives of 16S rRNA gene sequences detected at CPr (Figure <xref ref-type="fig" rid="F4">4</xref>).</p><p>At the rim and channel, the microbial community composition changed from ~12% (CPp) to ~21&#x02013;28% archaea, with most archaeal phylotypes related to genera capable of heterotrophic sulfur-respiration, such as <italic>Thermofilum, Pyrobaculum, Desulfurococcus</italic>, <italic>Thermococcus</italic>, and <italic>Staphylothermus</italic> (Stetter et al., <xref rid="B58" ref-type="bibr">1990</xref>) (Figure <xref ref-type="fig" rid="F4">4</xref>). The observed change suggests an increased bioavailability of dissolved organic carbon (DOC) in these locations. However, any DOC present in CPc samples was below the detection limit of 0.5 mg l^&#x02212;1, while CPp, CPr, and AP had measurable DOC concentrations (Table <xref ref-type="table" rid="T1">1</xref>). Presumably, a relative increase in heterotrophs could have contributed to the disappearance of DOC at CPc. This interpretation leads to an interesting hypothesis that the first appearance of methylated arsenic species at Champagne Pool, in the channel site, is related to fundamental changes in microbial carbon utilization. The detection of 16S rRNA gene sequences closely related to <italic>Thermodesulfovibrio</italic> and <italic>Desulfovibrio</italic> (Figure <xref ref-type="fig" rid="F4">4</xref>) also supports the potential for bacterial sulfate reduction at this site, which is consistent with the sulfur reduction genes <italic>dsr</italic> and <italic>asr</italic> detected at CPr annotated to genes from the genera <italic>Thermodesulfovibrio</italic> and <italic>Desulfovibrio</italic>.</p><p>These changes in microbial community structure and sulfur speciation at CPr and CPc would likely result in a small decrease in H₂ relative to background levels produced abiotically by geothermal reactions, i.e., if arsenite transformation to thioarsenate followed the reaction pathway with S⁰ instead of sulfide (Figure <xref ref-type="fig" rid="F6">6</xref>). Detection of a significant increase (greater than two standard deviations above the mean of sites CPp, CPr, and AP) in close relatives of <italic>Aquificaceae</italic> at CPc (Figure <xref ref-type="fig" rid="F4">4</xref>), however, with members <italic>Aquifex, Hydrogenobacter, Hydrogenivirga, Thermocrinis</italic> known for H₂ oxidation (Eder and Huber, <xref rid="B13" ref-type="bibr">2002</xref>), and detected hydrogenase encoding genes (Table <xref ref-type="table" rid="T4">4</xref>) annotated to genes from these genera, confirm the persistence of H₂ utilization as a primary energy source. Additionally, detected hydrogenase encoding genes from CPc (Table <xref ref-type="table" rid="T4">4</xref>) could also be annotated to genes from genera <italic>Sulfurihydrogenibium</italic>, <italic>Persephonella</italic>, and <italic>Thermofilum</italic>, whose closely related sequences were detected at CPc (Figure <xref ref-type="fig" rid="F4">4</xref>).</p><p>The greater microbial species richness at AP (Figure <xref ref-type="supplementary-material" rid="SM1">S3</xref>), likely facilitated by decreased temperature of waters on the outflow terrace, establishes metabolic opportunities in terms of arsenic and sulfur cycling for a wider range of microorganisms. The increased proportion of sulfur oxidation vs. sulfur reduction genes detected at AP compared to CPr and CPc (Table <xref ref-type="table" rid="T4">4</xref>), confirmed by Fisher's exact test, enables the oxidation of reduced sulfur compounds to sulfate. The annotation of the dominantly detected <italic>sox</italic> genes (including <italic>soxCD</italic>) (Table <xref ref-type="table" rid="T4">4</xref>) to gene sequences from close relatives of detected 16S rRNA gene sequences at AP (Figure <xref ref-type="fig" rid="F4">4</xref>) revealed a close relationship to genera <italic>Thermus</italic>, <italic>Thiomonas</italic>, and <italic>Sulfurihydrogenibium</italic> (Figure <xref ref-type="fig" rid="F4">4</xref> and Table <xref ref-type="table" rid="T4">4</xref>). Members of the family <italic>Thermus</italic>, <italic>Thiomonas</italic>, and <italic>Sulfurihydrogenibium</italic> are known to oxidize elemental sulfur (Skirnisdottir et al., <xref rid="B55" ref-type="bibr">2001</xref>; Dopson and Johnson, <xref rid="B11" ref-type="bibr">2012</xref>) and the activity of these groups is consistent with the observed elevated SO^2&#x02212;₄ concentrations at AP (Table <xref ref-type="table" rid="T3">3</xref>). The significance of the change in sulfur oxidation vs. sulfur reduction gene proportion from CPp to CPr and CPc to AP, but not from CPp to AP, can be explained by a change in sulfur metabolism from CPp to CPr and CPc to AP (Table <xref ref-type="table" rid="T4">4</xref>). At CPp, the main proportion of sulfur metabolizing genes belonged to sulfur oxidation genes, whereas at CPr, the dominant component of sulfur metabolizing genes belonged to sulfur reduction genes (Table <xref ref-type="table" rid="T4">4</xref>). From CPc to AP the proportion of sulfur metabolizing genes changed again from sulfur reduction genes at CPr and CPc to sulfur oxidation genes at AP (Table <xref ref-type="table" rid="T4">4</xref>). This trend is reflected by the sulfur species detected at the individual site (Table <xref ref-type="table" rid="T3">3</xref>) and the microbial communities present (Figure <xref ref-type="fig" rid="F4">4</xref>).</p><p>Alongside the indirect impacts on arsenic transformation from microbial sulfur cycling, the metagenomic data for all sites revealed the presence of arsenic resistance genes (Figure <xref ref-type="fig" rid="F8">8</xref>). The dominance of the <italic>ars</italic> operon at all sites (Table <xref ref-type="table" rid="T4">4</xref>) supports the high degree of utility and conservation of this arsenic resistance mechanism. The smallest functional <italic>ars</italic> operon, <italic>arsRB</italic>, detected at all sites (Table <xref ref-type="table" rid="T4">4</xref>), was co-present with the arsenical pump-driving ATPase-encoding gene <italic>arsA</italic> (Table <xref ref-type="table" rid="T4">4</xref>), which may enhance arsenite exclusion from cells. Some microbes at CPp and AP may also benefit from the arsenite resistance enhancer encoding gene <italic>arsH</italic> found in the metagenomes of these sites (Table <xref ref-type="table" rid="T4">4</xref>). At AP, detection of the gene encoding for ArsD, which increases the affinity of ArsA to arsenite, would further enhance cellular arsenite exclusion. Although, statistical analysis did not show a significant change in the proportion of <italic>arsB</italic> to <italic>acr3</italic> from CPc to AP, a high proportion of the gene <italic>acr3</italic> at AP may reflect increasing microbial community diversity and the wider phylogenetic distribution of <italic>acr3</italic> compared to <italic>arsB</italic> (Achour et al., <xref rid="B1" ref-type="bibr">2007</xref>; Fu et al., <xref rid="B16" ref-type="bibr">2009</xref>). The detection of the arsenate reductase gene <italic>arsC</italic> throughout all sites (Table <xref ref-type="table" rid="T4">4</xref>) may reflect the persistence of (thio)arsenate, as confirmed by arsenic speciation analysis (Figure <xref ref-type="fig" rid="F2">2</xref>). At CPp, CPr, and AP, <italic>arsC</italic> is accompanied by the respiratory arsenate reductase gene subunits of <italic>arr</italic> from which only the combination of <italic>arrAB</italic> at AP is functional (Table <xref ref-type="table" rid="T4">4</xref>). In addition to arsenate, the ArsC and Arr enzymes likely also recognize thioarsenate, with arsenic in the same oxidation state as in arsenate (Figure <xref ref-type="fig" rid="F8">8</xref>). From these findings, we speculate an early origin for arsenate reductase, as thioarsenate levels would presumably be elevated in sulfidic hot springs on the early Earth.</p><fig id="F8" position="float"><label>Figure 8</label><caption><p><bold>Arsenic and sulfur cycle influenced by microorganisms in Champagne Pool</bold>. H₃AsO₃, arsenite; H₃AsO₄, arsenate; H₃AsO₃S, monothioarsenate; H₃AsO₂S₂, dithioarsenate; H₃AsOS₃, trithioarsenate; (CH₃)₂AsO(SH), dimethylmonothioarsenate. purple: biotic reactions.</p></caption><graphic xlink:href="fmicb-05-00569-g0008"/></fig><p>The absence of the arsenite oxidase gene <italic>aio</italic> at all sites implies the <italic>ars</italic> operon as the sole arsenite resistance mechanism for coping with high arsenite levels at Champagne Pool (Figure <xref ref-type="fig" rid="F2">2</xref>), although there is a possibility that a newly identified arsenite oxidase encoding <italic>arxA</italic> gene (Zargar et al., <xref rid="B70" ref-type="bibr">2010</xref>, <xref rid="B69" ref-type="bibr">2012</xref>) has not been correctly annotated in MG-RAST. However, the implication of the <italic>ars</italic> operon as the sole arsenite resistance mechanism is an early prevalence of a purely detoxification mechanism over an arsenic resistance pathway coupled to energy conservation (<italic>aio</italic>). This hypothesis is perhaps supported by the detection of abundant sulfur and hydrogen oxidizing microorganisms (Figure <xref ref-type="fig" rid="F4">4</xref>), and sulfur and H₂ respiratory genes, (Table <xref ref-type="table" rid="T4">4</xref>) in the Champagne Pool metagenome for all sites. Either of these respiratory pathways would allow a cell to conserve more energy than could be obtained from arsenite oxidation.</p><p>The detection of the organic arsenic species dimethylmonothioarsenate in the outflow channel of Champagne Pool, CPc, revealed direct microbial control over arsenic speciation (Figure <xref ref-type="fig" rid="F8">8</xref>). The occurrence of methylated arsenic species can be uniquely attributed to arsenic-methylating microorganisms (Bentley and Chasteen, <xref rid="B5" ref-type="bibr">2002</xref>). This resistance mechanism transforms arsenate and arsenite into mono-, di- and trimethylated arsenic, prior to excluding these species from the cell (Bentley and Chasteen, <xref rid="B5" ref-type="bibr">2002</xref>). The trimethylated arsenic species arsine is highly volatile, and was therefore not included in our analyses. Interestingly, the occurrence of the organic arsenic species dimethylmonothioarsenate (Figure <xref ref-type="fig" rid="F2">2</xref>) corresponds in location (CPc) with a significant increase in sequences closely related to the family <italic>Aquificaceae</italic> (Figure <xref ref-type="fig" rid="F4">4</xref>). This family, with genera <italic>Aquifex, Hydrogenobacter</italic>, and <italic>Hydrogenivirga</italic>, all detected in CPc samples, is known for H₂ oxidation. Molecular hydrogen is abundant at Champagne Pool, generated from hydrothermal reactions (Giggenbach et al., <xref rid="B17" ref-type="bibr">1994</xref>), and is oxidized aerobically by microbes with O₂ as the electron acceptor. The abundance of <italic>Aquificaceae</italic> at CPc may therefore be related to increased DO saturation, which reached a maximum at the outflow channel (Table <xref ref-type="table" rid="T1">1</xref>). The family of <italic>Aquificaceae</italic> belongs to the order <italic>Aquificales</italic>, which is known to possess a functional arsenic methylation mechanism (Takacs-Vesbach et al., <xref rid="B59" ref-type="bibr">2013</xref>). Typically, the gene <italic>arsM</italic> is associated with arsenic methylation (Qin et al., <xref rid="B45" ref-type="bibr">2006</xref>), and may be co-expressed with the <italic>ars</italic> operon. However, no <italic>arsM</italic> sequences were detected at CPc. Nonetheless, the co-appearance of methylated arsenic and a significant increase in <italic>Aquificaceae</italic> supports the inference that As-methylation via an alternate mechanism to ArsM can likely be attributed to this group at CPc. These results reveal a potentially interesting role for methylation as an arsenic tolerance strategy in early geothermal springs.</p><p>In summary, quantitative arsenic and sulfur speciation of acid-sulfide hot spring waters provided the biogeochemical context in this study for understanding microbial community composition and functionality at Champagne Pool, Waiotapu, New Zealand. Analyses of community diversity; and arsenic resistance and sulfur cycling genes revealed several potential direct and indirect impacts of various microbial groups on coupled arsenic and sulfur cycling. Principally, the distributions of (thio)arsenate and the arsenate resistance genes <italic>arsC</italic> and <italic>arr</italic> suggest an ancient evolution for arsenate reductase that does not invoke arsenic oxidation by molecular oxygen. Further, our results support evolutionary prioritization of arsenite detoxification via the <italic>ars</italic> operon over oxidation via arsenite oxidase. Finally, the combination of abundant <italic>Aquificaceae</italic> relatives and the unique appearance of dimethylmonothioarsenate in the outflow channel of Champagne Pool suggests an important role for thermophilic arsenic methylation in the evolution of arsenic tolerance strategies.</p></sec><sec><title>Conflict of interest statement</title><p>The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.</p></sec></sec>

Cause-specific mortality at INDEPTH Health and Demographic Surveillance System Sites in Africa and Asia: concluding synthesis

<p>This synthesis brings together findings on cause-specific mortality documented by means of verbal autopsies applied to over 110,000 deaths across Africa and Asia, within INDEPTH Network sites.</p><sec id="st1"><title>Methods</title><p>Developments in computerised methods to assign causes of death on the basis of data from verbal autopsy (VA) interviews have made possible these standardised analyses of over 110,000 deaths from 22 African and Asian Health and Demographic Surveillance System sites in the INDEPTH Network. In addition to previous validations of the InterVA-4 probabilistic model, these wide-ranging analyses provide further evidence of the applicability of this approach to assigning the cause of death. Plausible comparisons with existing knowledge of disease patterns, as well as substantial correlations with out-of-model parameters such as time period, country, and other independent data sources were observed.</p></sec><sec id="st2"><title>Findings</title><p>Substantial variations in mortality between sites, and in some cases within countries, were observed. A number of the mortality burdens revealed clearly constitute grounds for public health actions. At an overall level, these included high maternal and neonatal mortality rates. More specific examples were childhood drowning in Bangladesh and homicide among adult males in eastern and southern Africa. Mortality from non-communicable diseases, particularly in younger adulthood, is an emerging cause for concern. INDEPTH&#x02019;s approach of documenting all deaths in particular populations, and successfully assigning causes to the majority, is important for formulating health policies.</p></sec><sec id="st3"><title>Future directions</title><p>The pooled dataset underlying these analyses is available at the INDEPTH Data Repository for further analysis. INDEPTH will continue to fill cause-specific mortality knowledge gaps across Africa and Asia, which will also serve as a baseline for post-2015 development goals. The more widespread use of similar VA methods within routine civil registration systems is likely to become an important medium-term strategy in many countries.</p></sec>

<contrib contrib-type="author"><name><surname>Sankoh</surname><given-names>Osman</given-names></name><xref ref-type="aff" rid="AF0001">1</xref><xref ref-type="aff" rid="AF0002">2</xref><xref ref-type="aff" rid="AF0003">3</xref><xref ref-type="corresp" rid="cor1">*</xref></contrib><contrib contrib-type="author"><name><surname>Byass</surname><given-names>Peter</given-names></name><xref ref-type="aff" rid="AF0002">2</xref><xref ref-type="aff" rid="AF0004">4</xref></contrib>

Global Health Action

<p>This Special Issue on cause-specific mortality has described an unprecedented dataset &#x02013; more than 110,000 deaths in African and Asian sites, documented over 12 million person-years of population-based surveillance (<xref rid="CIT0001" ref-type="bibr">1</xref>). By way of synthesis across the detailed papers, here we reflect on some of the cross-cutting issues in terms of methods, findings, and future directions.</p><sec sec-type="methods" id="S0001"><title>Methods</title><p>In the early years of the INDEPTH Network, following its establishment in 1998, one of the recurring issues reported by sites was the difficulty of getting physicians to assign causes of death to material from verbal autopsy (VA) interviews in a timely, cost-effective, and consistent manner. There is a reason why the vast majority of deaths in African and Asian communities still pass by unregistered and uncertified: usually when someone dies, there is nobody there to certify the death.</p><p>There have been over 500 PubMed entries for &#x02018;verbal autopsy&#x02019; over the past 25 years, but the past decade has seen rapid developments in methods for handling VA data that do not require a physician being there to give his or her expert assessment of cause of death. Work leading directly to the development of the InterVA-4 model as used here (<xref rid="CIT0002" ref-type="bibr">2</xref>) was first published in 2003 (<xref rid="CIT0003" ref-type="bibr">3</xref>), from the FilaBavi INDEPTH site in Vietnam.</p><p>There have been, and no doubt will continue to be, debates about the advantages and the disadvantages of particular methods and how they relate to what might happen if well-trained and effective physicians were available to assign cause of every death. But at any particular point in time, the relevant question must be &#x02018;what available method can yield adequate information?&#x02019; For these analyses, we chose to use the InterVA-4 model (version 4.02) because it was available in the public domain, and complied exactly with the WHO 2012 VA standard (<xref rid="CIT0004" ref-type="bibr">4</xref>), and also had an appreciable track record in a number of other studies.</p><p>Validation of VA methods is a non-trivial exercise. Physician causes of death assigned to VA material are often considered insufficiently consistent to be regarded as a reference standard for methodological comparisons, so there have been a number of attempts to use harder endpoints such as hospital diagnoses or autopsy findings (<xref rid="CIT0005" ref-type="bibr">5</xref>). A Population Health Metrics Research Consortium study of hospital deaths was probably the most extensive of these (<xref rid="CIT0006" ref-type="bibr">6</xref>), though it still had methodological problems (<xref rid="CIT0007" ref-type="bibr">7</xref>).</p><p>Interestingly, though, the application of a particular VA method to a large and wide-ranging dataset also provides further insights into the validity of its outputs. This is particularly the case when patterns that emerge are observed to be linked to parameters that are not part of the VA data processed by the model. Interesting examples here include time period, site, country, altitude, and other independent sources such as the Malaria Atlas Project parasite prevalence map (<xref rid="CIT0008" ref-type="bibr">8</xref>). Thus, at the Agincourt, South Africa, site, it was possible to show the complete rise and fall of the overwhelming HIV/AIDS epidemic (<xref rid="CIT0009" ref-type="bibr">9</xref>), even though the InterVA-4 model &#x02018;knew&#x02019; nothing about the epidemic dynamics. In addition, modelled estimates of cause-specific mortality such as those from international agencies and the Global Burden of Disease (GBD) study (<xref rid="CIT0010" ref-type="bibr">10</xref>) can be compared and contrasted, even if the validity of those estimates for countries with sparse data may not be absolute.</p><p>InterVA-4 had previously been validated in relation to HIV status in Africa (<xref rid="CIT0011" ref-type="bibr">11</xref>) and so it was unsurprising that rates for high-prevalence countries in the HIV/AIDS results from INDEPTH sites here (<xref rid="CIT0012" ref-type="bibr">12</xref>) were similar in many cases to UNAIDS and GBD national estimates. Likewise, the dynamics of HIV-related mortality over time &#x02013; a complete epidemic curve at the Agincourt, South Africa, site and sharp declines from epidemic peaks in other high-prevalence settings &#x02013; supported the validity with which InterVA-4 was identifying HIV-related deaths. Equally interesting though was the situation in low-prevalence settings. Particularly for the four sites in Bangladesh, very low HIV-related death rates were assigned by InterVA-4 except at the Bandarban site, which covers a militarised high-migration area close to the Myanmar border.</p><p>Similarly, analyses in the malaria paper in this Special Issue (<xref rid="CIT0013" ref-type="bibr">13</xref>) threw up interesting findings in terms of validity. Using VA to identify possible malaria deaths has long been controversial (<xref rid="CIT0014" ref-type="bibr">14</xref>), particularly among adults. Although no biomedical evidence was presented in these analyses, there were significantly higher rates of acute adult febrile deaths (major proportions of which InterVA-4 attributed to malaria) in settings where higher rates of childhood malaria mortality were also attributed, across hundred-fold variations in rates. This is something that can only become clear when VAs are applied systematically and consistently to all deaths in a range of populations, rather than counting local malaria deaths at health facilities or in surveys. The conclusion must either be that there is some proportionality between childhood and adult malaria mortality rates, or, less probably, there is some non-malaria cause of acute adult febrile death that occurs more commonly in high-malaria settings. Equally striking was the significant correlation at the local level between the MAP estimates of parasite prevalence (<xref rid="CIT0008" ref-type="bibr">8</xref>) and childhood malaria mortality rates, again across hundred-fold variations. In this case, there was no ambiguity between local findings and national estimates, because the MAP estimates are available in a high-resolution spatial grid from which site areas could be precisely located. These findings, put together, add a considerable sense of validity to InterVA-4&#x02019;s assignment of malaria as a cause of death.</p><p>In totality, this dataset is based on over 20 million individual data items extracted from the original VA interview material, condensed to findings on around 100,000 deaths, and categorised into 60 causes by 22 sites by 14 age&#x02013;sex groups. Inevitably, there will be both systematic and random errors within this huge amount of material. As with any data processing, the quality of outputs depends on the quality of inputs. Two specific examples of dubious input data led to anomalies: very high rates of digestive neoplasms among adults at the Navrongo, Ghana, site, and almost all deaths from external causes being attributed to transport accidents at the Nouna, Burkina Faso, site. These anomalies may have arisen from issues with historic VA instruments, or with problems in extracting data into the WHO 2012 format. However, in the overall context, these are relatively isolated examples, and certainly do not invalidate the wider findings.</p></sec><sec id="S0002"><title>Findings</title><p>Taking all the findings from these large-scale analyses, there were no major surprises, even though some of the burdens of mortality revealed must be regarded as unacceptably high in terms of population health. Some parameters at certain sites varied from national estimates, but there may be good reasons for that. Particularly in the case of Kenya, a highly heterogeneous country in many respects, it was not surprising to find major differences between the coastal site at Kilifi, the site based in the Nairobi slum population, and the inland Kisumu site. One can interpret this either as evidence of the richness of local detail that population surveillance sites can generate or criticise such sites as being individually unrepresentative of any wider situation. This will be an on-going debate for as long as there is no universal civil registration including VA implemented across whole populations. Nevertheless, it is clear, looking across all the sites reported here, that there is considerable diversity in cause-specific mortality, which is being successfully captured.</p><p>Across all the sites reported in this supplement, there were examples of mortality burdens that clearly constitute grounds for urgent public health actions &#x02013; with the advantage that the same tools used in these assessments would be equally applicable for intervention monitoring and evaluation purposes. Some of these mortality burdens were common across all sites &#x02013; such as high maternal and neonatal mortality rates compared with other parts of the world (<xref rid="CIT0015" ref-type="bibr">15</xref>, <xref rid="CIT0016" ref-type="bibr">16</xref>). Other striking findings related to very specific settings and sub-populations, such as childhood drowning in Bangladesh, and homicide among adult males in some eastern and southern African settings (<xref rid="CIT0017" ref-type="bibr">17</xref>).</p><p>Patterns of non-communicable disease (NCD) mortality are one of the most complex topics explored here (<xref rid="CIT0018" ref-type="bibr">18</xref>). NCDs have rapidly acquired an increased importance following the September 2011 UN special high-level meeting and subsequent review meeting in July 2014. Nevertheless, the complex mixture of risk factors, ageing populations, and multiple causes of mortality can make it difficult to separate facts from hyperbole. Consequently, the NCD mortality data presented here distinguish clearly between the premature mortality burden (taken here as under 65 years of age) and the more inevitable occurrence of NCDs as people get older. At this particular point of demographic transition in the low- and middle-income countries represented here, the population proportions of people aged &#x02265;65 are fairly low, but set to increase rapidly. Various risk factors are moving in parallel transitions, with the result that large &#x02018;healthy-exposed&#x02019; cohorts are developing, which will influence future NCD mortality. The global concern around NCDs, as being a problem for low-, middle-, and high-income countries alike, is therefore well justified, even if the results presented here suggest that the current premature NCD mortality burden in Africa and Asia is not that high.</p><p>An important strength of INDEPTH&#x02019;s approach to measuring cause-specific mortality is that all deaths in complete populations are included in the surveillance operations, irrespective of health care seeking and other factors. Of course, there will always be a proportion of those deaths which are impossible to follow-up with VA interviews &#x02013; perhaps because nobody witnessed a death or because nobody was available to respond to the interview. Similarly, there will always be a proportion of VA interviews that contain little or no useful information relating to the cause of death. Nonetheless, including the not-done and indeterminate cases in the analyses is critical to the overall understanding of the burdens related to specific causes of death, as is having all the cause-specific fractions adding up to the total deaths in the population.</p></sec><sec id="S0003"><title>Future directions</title><p>Making the pooled dataset (<xref rid="CIT0019" ref-type="bibr">19</xref>) publicly available at the INDEPTH Data Repository simultaneously with the publication of this Special Issue is part of the INDEPTH Network&#x02019;s on-going commitment to open-access data (<xref rid="CIT0020" ref-type="bibr">20</xref>). We hope that this will lead to many further analyses based on these important data, beyond the basic descriptions presented in this Special Issue. The INDEPTH Network is also committed to furthering the long-term process of development in VA methods, led by WHO through its Reference Group on Health Statistics and the WHO Collaborating Centre for Verbal Autopsy at Ume&#x000e5; University.</p><p>Apart from providing important insights into cause-specific mortality at 22 INDEPTH sites, these analyses also provide clear proof of principle about the viability of large-scale VA operations as the basis for understanding mortality patterns globally. The INDEPTH Network basically exists because of the extremely poor coverage of vital event data across Africa and Asia (<xref rid="CIT0021" ref-type="bibr">21</xref>) and would absolutely welcome a scenario in which its operations were no longer relevant because all lives were being systematically documented. Routine application of VA as part of universal civil registration seems the most likely way forward towards that goal, but it is likely to take a long time to achieve anywhere near complete coverage. Although the technical problems are more or less solved, through the WHO 2012 VA standard and models for assigning cause of death like InterVA-4, there remain considerable political and financial challenges. Which ministry or government agency in each country should take the responsibility for implementation? Which budget lines should be used for the necessary resources? How can populations develop confidence that their personal details are being collected for the wider good, rather from any ulterior motive? These are all big questions, which will not find easy answers in many countries.</p><p>For the immediate future, therefore, INDEPTH expects to continue its work to fill in some of the information gaps. As the 2015 deadline for the Millennium Development Goals (MDGs) approaches, attention is turning to the post-2015 goals. One of the most problematic areas around evaluating the current MDGs has been the lack of detailed information pertaining to the 1990 starting point. One of the ways in which INDEPTH expects to contribute to the next phase of health-related development goals is by providing a comprehensive set of baseline cause-specific mortality data for the 2012&#x02013;2015 period. Continued monitoring beyond that baseline, using standardised and comparable methods, will be an important on-going activity for the foreseeable future.</p></sec>