article_title
stringlengths 1
1.44k
| abstract
stringlengths 10
22.8k
| contributor
stringlengths 29
38.3k
| journal
stringlengths 3
246
| full_text
stringlengths 22
2.53M
|
|---|---|---|---|---|
A Six-Step Model for Evaluation of Community-based Physical Activity Programs | Could not extract abstract | <contrib contrib-type="author" corresp="yes"><name><surname>Sarah</surname><given-names>Levin Martin</given-names></name><degrees>PhD</degrees><aff>Division of Nutrition and Physical Activity, Physical Activity and Health Branch, National Center for Chronic Disease Prevention and Health Promotion, Centers for Disease Control and Prevention</aff><address><email>sjl2@cdc.gov</email><addr-line>4770 Buford Hwy NE, Mail Stop K-46, Atlanta, GA 30341</addr-line><phone>770-488-5413</phone></address></contrib><contrib contrib-type="author"><name><surname>Heath</surname><given-names>Gregory W</given-names></name><degrees>DHSc, MPH</degrees><aff>Division of Nutrition and Physical Activity, Physical Activity and Health Branch, National Center for Chronic Disease Prevention and Health Promotion, Centers for Disease Control and Prevention, Atlanta, Ga</aff><aff>Dr Heath is now with the Department of Health and Human Performance, University of Tennessee at Chattanooga, Chattanooga, Tenn</aff></contrib> | Preventing Chronic Disease | <sec><title>Introduction</title><p>Physical activity is a leading health indicator (<xref rid="B1" ref-type="bibr">1</xref>) and has numerous benefits, including reduced risk of coronary heart disease, hypertension, colon cancer, and diabetes (<xref rid="B2" ref-type="bibr">2</xref>). Regular physical activity can help prevent the onset of diabetes (<xref rid="B3" ref-type="bibr">3</xref>), even among those at high risk (i.e., those with impaired glucose tolerance [<xref rid="B4" ref-type="bibr">4</xref>]), and is part of diabetes self-management among people with the disease (<xref rid="B5" ref-type="bibr">5</xref>).</p><p>The Diabetes Prevention Program, a study funded by the National Institute of Diabetes & Digestive & Kidney Diseases, found that participants with impaired glucose tolerance who were assigned to an intensive lifestyle intervention reduced their risk of getting type 2 diabetes by 58%. On average, this group maintained their physical activity at 30 minutes per day, usually with walking or other moderate-intensity exercise, and lost 5% to 7% of their body weight (<xref rid="B4" ref-type="bibr">4</xref>).</p><p>It is likely that health educators at local health departments addressing diabetes and other chronic diseases will be asked to evaluate a physical activity program, because evaluation has a central role not only in improving programs but also in satisfying accountability requirements. A carefully planned evaluation can engage community members and build community capacity, and the results can be used to influence policy makers, share what works and what doesn't work with other communities, and help ensure funding and sustainability.</p><p>The Centers for Disease Control and Prevention (CDC) has published the <italic>Framework for Program Evaluation in Public Health</italic> (<xref rid="B6" ref-type="bibr">6</xref>), which recommends six steps for effective program evaluation: 1) engaging stakeholders, 2) describing the program, 3) focusing the evaluation design, 4) gathering credible evidence, 5) justifying conclusions, and 6) ensuring use and sharing lessons learned. In this article, we describe these six steps using a hypothetical example of physical activity programming aimed at diabetes prevention. For this example, we assume that the reader is a community-based health educator at a local health department.</p></sec><sec><title>Selecting an Intervention</title><p>Before planning the evaluation, you should be familiar with strategies or interventions proven to increase physical activity at the population level. <italic>The Guide to Community Preventive Services</italic> (<italic>Community Guide</italic>) (<xref rid="B7" ref-type="bibr">7</xref>), available from <ext-link xlink:href="www.thecommunityguide.org/pa/default.htm" ext-link-type="uri">www.thecommunityguide.org/pa/default.htm</ext-link>, includes eight recommended strategies that fall within three domains: informational approaches to increasing physical activity, behavioral and social approaches to increasing physical activity, and environmental and policy changes to increasing physical activity. Becoming familiar with these strategies is important background work. Because you may not have all the resources needed to carry out a population-based physical activity intervention, it is essential to work with partners. In fact, your role may be to influence others to carry out the program. One possible way to select the intervention strategy is by using the RE-AIM framework (information available from <ext-link xlink:href="www.re-aim.org/" ext-link-type="uri">www.re-aim.org/</ext-link>[<xref rid="B8" ref-type="bibr">8</xref>]), which considers the <italic>reach</italic>, <italic>efficacy</italic>, <italic>adoption</italic>, <italic>implementation</italic>, and <italic>maintenance</italic> of public health interventions.</p><p>For this article, we selected the strategy of creating or enhancing access to places for physical activity, combined with informational outreach activities, from the <italic>Community Guide</italic>. (A description of this strategy is available from <ext-link xlink:href="www.thecommunityguide.org/pa/pa-int-create-access.pdf" ext-link-type="uri">www.thecommunityguide.org/pa/pa-int-create-access.pdf</ext-link>.) This strategy involves the efforts of worksites, coalitions, agencies, and communities to change the local environment to create or improve access to opportunities for physical activity.</p></sec><sec><title>Steps in an Evaluation of a Physical Activity Intervention</title><p>The CDC's six-step <italic>Framework for Program Evaluation in Public Health</italic> will be used to guide this step-by-step example (<xref rid="B6" ref-type="bibr">6</xref>). These six steps have been adapted for use in physical activity programs and published in the <italic>Physical Activity Evaluation Handbook</italic> (<xref rid="B9" ref-type="bibr">9</xref>), available from <ext-link xlink:href="www.cdc.gov/nccdphp/dnpa/physical/handbook/index.htm" ext-link-type="uri">www.cdc.gov/nccdphp/dnpa/physical/handbook/index.htm</ext-link>. Because an evaluation is not worth doing if the information gleaned will not be used, utility is perhaps the most important standard for program evaluation. The other standards are feasibility (you cannot evaluate with resources you do not have), accuracy (you cannot evaluate with poor or invalid measures), and propriety (you cannot evaluate if you are not fair and ethical to everyone involved).</p><sec><title>Step 1: engage stakeholders</title><p>Important stakeholders for you as a health educator at a local health department are <italic>partners</italic>; these partners will carry out the intervention strategy. For an intervention to create or enhance places for physical activity, potential partners might include a city park, a shopping mall, the YMCA, the tourism bureau, and the community college. Also of great importance in terms of meeting the utility standard are <italic>decision makers</italic> — individuals who can use evaluation results to allocate future funds or cut programs. Examples might include the city mayor, the president of the community college, and the county-level director of parks and recreation.</p><p>You should invite all of the partner and decision-maker stakeholders to a meeting to describe the recommended strategy. The group should then discuss its role in making this strategy into a reality in its community and what evaluation resources it can offer.</p><p>Another group of stakeholders is the <italic>participants</italic>, individuals at high risk of developing diabetes. One way to engage participants is to invite them to a focus group or town meeting. At such a gathering, their ideas about the program can be assessed and used to refine the program to meet their needs. For the "creating access" strategy, for example, the participants can reveal what physical activity offerings would interest them and what venues might best reach them with information about these offerings.</p></sec><sec><title>Step 2: describe the program</title><p>The partners should be invited back for a second meeting to work on a logic model to depict graphically the proposed relationship between activities and expected outcomes. In this step, the work pertains to planning both the intervention and its evaluation. On the basis of step 1, the health educator is able to share the list of activities discussed at the first meeting and indicate which ones seem more popular or less popular among participants in the focus groups or town meetings.</p><p>To begin creating a logic model, the partners can divide the activities into two columns, early activities and later activities. Then the group should discuss outcomes they can realistically expect from the proposed activities. One outcome that seems obvious is an increased level of physical activity of residents, but there are more immediate and targeted outcomes that may precede such a behavioral change, such as increased opportunities for physical activity, increased awareness of physical activity offerings, and limiting of the target population to those at risk for developing diabetes, diabetes complications, or both. For diabetes care there may be outcomes even later than increased levels of physical activity, such as decreased levels of hemoglobin A1c (HbA1c) and, eventually, decreased incidence of diabetes morbidity and mortality. It is the role of the health educator to insist that the group set short-term objectives that include measurable outcomes. Objectives should be SMART — that is, <italic>specific</italic>, <italic>measurable</italic>, <italic>achievable</italic>, <italic>relevant</italic>, and <italic>time-bound</italic>. (See Appendix 4 in the <italic>Physical Activity Evaluation Handbook</italic> [<xref rid="B9" ref-type="bibr">9</xref>].) After the activities and the outcomes have been placed in sequence, the logic model begins to take shape. Inputs (i.e., resources to carry out the activities) can be added to the far left, and an overarching goal can be added to follow from long-term outcomes, as shown in the Figure.</p><boxed-text position="float"><fig position="float" id="F1" fig-type="diagram"><label>Figure</label><caption><p>Example of a logic model for an intervention to create or enhance access to physical activity (PA) combined with informational outreach activities.</p></caption><alt-text>Logic model</alt-text><table-wrap position="anchor" id="d95e183"/><alternatives><graphic xlink:href="PCD31A24s01" position="float"/><table frame="hsides" rules="groups"><tbody><tr><td rowspan="1" colspan="1">This figure is a logic model that illustrates a sequence of events for a sample community intervention to create or enhance access to physical activity. In general, the figure reads from left to right. It begins on the left with a box labeled “Inputs,” which lists health educator, YMCA, parks and recreation, shopping mall, community college, and tourism bureau. These inputs lead to two boxes, one labeled “Early Activities” and one labeled “Later Activities.” “Early Activities” include 1) coordinating a meeting and conducting a focus group and 2) developing informational outreach activities via clinics and the city hospital. “Later Activities” include 1) mapping facilities, 2) offering reduced-fee night for use of pool, 3) establishing a women’s volleyball league, 4) marketing new and existing opportunities, and 5) purchasing and placing signage on historical walking tour. To the right of the box labeled “Later Activities” there are three more boxes: “Short-term Outcomes,” “Mid-term Outcomes,” and “Long-term Outcomes.” The box labeled “Short-term Outcomes” includes “increased awareness of physical activity opportunities.” The box labeled “Mid-term Outcomes” includes “increased participation in water aerobics and swimming, volleyball, and walking tour of historical sites.” The box labeled “Long-term Outcomes” includes 1) walking in town more normative, 2) increased social support for physical activity, 3) increased levels of physical activity among target population, and 4) decreased HbA1c. This box leads to the box below labeled “Goal,” which is “better prevention and control of diabetes in the community.”</td></tr><tr><td rowspan="1" colspan="1">A box above this row of boxes lists factors that influence the program’s success, such as politics, other initiatives, socioeconomic factors, staff turnover within each partner site, social norms and conditions, program history, or stage of development. This box has arrows that point down to the boxes in the row described above, indicating that external factors affect every stage of the intervention and its outcomes.</td></tr></tbody></table></alternatives></fig></boxed-text></sec><sec><title>Step 3: focus on evaluation</title><p>As a health educator, you want to be sure that the evaluation is useful not only to your health department but also to the partner organizations that help to implement the program. To ensure buy-in and later use of the evaluation, ask stakeholders to develop questions that they would like to have answered. For example, the YMCA staff may want to know whether their membership increases. Increased membership becomes an outcome in the logic model. Note that the logic model can be made from right to left, that is, by asking, "How will we increase membership?" and then proposing some activities that might lead toward that outcome. Or it can be made from left to right, by asking, "Why are we doing that?" and then stating the expected outcomes of such activities. Either way, a focused evaluation will be one that poses questions based on the program and one that results in answers that serve the purpose of the evaluation. The purpose often will be to improve the program; other purposes may include gaining insight and assessing program effects. Defining your purpose is an important component of this step.</p><p>In our example, the stakeholders have already agreed on a logic model in step 2 (<xref rid="F1" ref-type="fig">Figure</xref>), and so they can use it to focus their questions. They might decide to ask both process and outcome evaluation questions. Process questions relate to the inputs and activities, and outcome questions relate to the expected outcomes. It is possible to generate a long list of possible questions from the logic model, but then the list needs to be prioritized. Evaluating all questions may not be essential or even feasible. The stakeholders should remember the purpose of the evaluation and decide what would be useful for decision makers in prioritizing the list of questions. Examples of process and outcome questions include the following:</p><p>
<bold>Process</bold>
</p><list list-type="bullet"><list-item><p>Are the proposed activities being carried out by the partners? If not, why not?</p></list-item><list-item><p>What seem to be the most popular activities, and why?</p></list-item><list-item><p>Are we reaching individuals at risk for diabetes?</p></list-item></list><p>
<bold>Outcome</bold>
</p><list list-type="bullet"><list-item><p>Did awareness of opportunities increase?</p></list-item><list-item><p>Did participation increase? Why or why not?</p></list-item><list-item><p>Did HbA1c levels decrease among the population with diabetes?</p></list-item></list></sec><sec><title>Step 4: gather credible evidence</title><p>To answer the questions posed in step 3, evidence needs to be collected. How much evidence (quantity) and what kind of evidence (quality) are central to feasibility and accuracy. There must be a balance between collecting enough data and assuring it is of high quality. Sometimes a mix of quantitative and qualitative data will help achieve that balance: quantitative data can provide the numbers you need to answer some questions (e.g., participation rates), and qualitative data can help you understand why you got those numbers (e.g., interview a few who participate and a few who do not to learn why). Data are available from people, documents, observations, and existing information.</p><p>The <xref rid="T1" ref-type="table">Table</xref> provides a guide to collecting data for process and outcome questions by indicators, data sources, and performance measures. Indicators are what answer the question, data sources are the methods by which you collect data about the indicators, and performance measures are the outcomes you would like to achieve. It is helpful to have more than one indicator and more than one data source to answer each evaluation question. Using multiple indicators and data sources is often called <italic>triangulation</italic> and is recommended to increase accuracy. There are many tools available for collecting physical activity data. (See Appendix 5 in the <italic>Physical Activity Evaluation Handbook</italic> [<xref rid="B9" ref-type="bibr">9</xref>].)</p></sec><sec><title>Step 5: justify conclusions</title><p>There are three parts to this step: 1) analyze the data, 2) interpret the results, and 3) make judgments about the program. Having the performance measures helps to justify your conclusions. Perhaps a community college student needs an internship. You can hire him or her (often without financial compensation) to help with the evaluation. With guidance from you and the supervising professor, the student can analyze the data. Analysis for some questions will be easier than for others. For example, the difference between participation rates preintervention and postintervention is simple math, whereas analyzing focus group and interview data takes more time because all of the text must be read and common themes identified to answer the appropriate evaluation question.</p><p>After the analyses, you should convene a meeting of stakeholders to go over the results. Talk about possible alternative explanations to the findings of the evaluation. Discuss the limitations. One common limitation is having no control community; if possible, use a selected community as a basis of comparison in a quasi-experimental design. (For more on experimental designs, see the <italic>Physical Activity Evaluation Handbook</italic>, p. 26 [<xref rid="B9" ref-type="bibr">9</xref>].) Compare the results with the performance measures, and make judgments based primarily on that comparison. If you realize you did not achieve a performance measure, decide if you are willing to say that the program failed. It could be that almost every indicator showed improvement. With the decision makers present, the group can decide which results matter most and use those to summarize their findings to share with the community at large.</p></sec><sec><title>Step 6: ensure use and share lessons learned</title><p>The findings can be printed in the local newspaper, which, in our case, has the beneficial effect of increasing awareness of physical activity opportunities even further. The best features of the program should be highlighted. You can send the findings electronically to the CDC, where they can be widely distributed through the Physical Activity Listserv; in addition, examples of state physical activity programs can be posted on the State-based Physical Activity Directory at <ext-link xlink:href="http://apps.nccd.cdc.gov/DNPAProg/" ext-link-type="uri">http://apps.nccd.cdc.gov/DNPAProg/</ext-link>.</p><p>To ensure use of your evaluation findings, formulate action-oriented recommendations. To help share lessons learned, consider your audience: use appropriate communications strategies and consider the most effective format for information (e.g., report, fact sheet, oral presentation) and venue (e.g., Web site, television, news media, town hall meeting).</p></sec></sec><sec><title>Conclusion</title><p>Although the principles of evaluation transcend topical areas, we hope that this step-by-step guide provides insight and examples for evaluating physical activity programs. There are evidence-based strategies for promoting physical activity in a community setting, and there are diabetes programs across the nation that could implement these strategies by engaging partners and initiating systems change. In our examples for creating access to places for physical activity, systems were engaged, and the strategies depended on these system changes. Partner organizations benefit by learning principles of evaluation that they can use for continuous quality improvement. Program evaluation plays a key role in ensuring success and sustainability of these programs. </p></sec> |
Identifying Best Practices for WISEWOMAN Programs Using a Mixed-Methods Evaluation | Could not extract abstract | <contrib contrib-type="author" corresp="yes"><name><surname>Besculides</surname><given-names>Melanie</given-names></name><degrees>DrPH</degrees><role>Senior Public Health Researcher</role><aff>Mathematica Policy Research, Inc</aff><address><email>mbesculides@mathematica-mpr.com</email><addr-line>955 Massachusetts Ave, Suite 801, Cambridge, MA 02139</addr-line><phone>617-491-7900 ext 273</phone></address></contrib><contrib contrib-type="author"><name><surname>Zaveri</surname><given-names>Heather</given-names></name><aff>Mathematica Policy Research, Inc, Princeton, NJ</aff></contrib><contrib contrib-type="author"><name><surname>Farris</surname><given-names>Rosanne</given-names></name><degrees>PhD</degrees><aff>Centers for Disease Control and Prevention (CDC), Atlanta, Ga</aff></contrib><contrib contrib-type="author"><name><surname>Will</surname><given-names>Julie</given-names></name><degrees>PhD</degrees><aff>CDC, Atlanta, Ga</aff></contrib> | Preventing Chronic Disease | <sec><title>Introduction</title><p>There is no doubt that public health programs should follow best practices. In the programmatic setting, best practices are the processes that lead to the implementation of the most appropriate intervention for a given location and population (<xref rid="B1" ref-type="bibr">1</xref>). Identifying and applying best practices is complex — largely because recommendations on what works are based typically on experimental or other one-of-a-kind settings. As a result, the practices recommended are not likely to be relevant to most other settings. An alternative is to identify best practices by collecting data from existing programs and to use two or more complementary methods, or a mixed-methods approach, to data collection (<xref rid="B2" ref-type="bibr">2</xref>). A mixed-methods approach can be a combination of one or more qualitative methods or a mix of qualitative and quantitative methods.</p><p>A mixed-methods approach strengthens evaluation research, because no single method is without weakness or bias (<xref rid="B3" ref-type="bibr">3</xref>). Quantitative data, for example, may be objective, but they often lack the depth needed to elucidate how and why a program works. Qualitative data can enhance understanding of program implementation and operation, but are considered less objective. By combining the two, research can be both objective and rich.</p><p>There are several qualitative methods, each with strengths and weaknesses. For instance, although interviews with program staff can provide a detailed picture of program operations, they cannot objectively provide the range of participants' perspectives (T.S. Weisner, personal communication, September 2005). Focus group participants can provide information on program experiences and effects, but this information is not generalizable because focus group members typically do not represent all program participants.</p><p>This article describes a mixed-methods approach to identifying best practices in the Centers for Disease Control and Prevention's (CDC's) WISEWOMAN (Well Integrated Screening and Evaluation for Women Across the Nation) program.</p><p>In 1995, the CDC created the WISEWOMAN program, which was authorized by Congress in the Breast and Cervical Cancer Mortality Prevention Act of 1990 (Public Law 101-354). WISEWOMAN is designed to build upon the National Breast and Cervical Cancer Early Detection Program (NBCCEDP) by offering 1) screening for risk factors associated with cardiovascular disease and 2) lifestyle intervention services to women aged 40 to 64 years who participate in NBCCEDP. WISEWOMAN participants must be uninsured and ineligible for Medicaid.</p><p>Although the cardiovascular screening is undoubtedly important, the lifestyle intervention offered through WISEWOMAN is a key service intended to modify the behaviors associated with increased risk for cardiovascular and other chronic diseases. In fact, the intervention is predicated on the notion that obesity, poor diet, physical inactivity, and tobacco use can be modified to reduce high blood pressure and elevated serum cholesterol levels at relatively low cost and with minimal risk to participants. Theoretically, a reduction in risk factors leads to a decreased incidence of cardiovascular events such as myocardial infarction. The literature on lifestyle interventions suggests that a combination of diet and physical activity is most effective in reducing the risk factors for cardiovascular disease in women (<xref rid="B4" ref-type="bibr">4</xref>,<xref rid="B5" ref-type="bibr">5</xref>).</p><p>The CDC not only requires all WISEWOMAN programs to offer a lifestyle intervention but also encourages them to use the national guidelines for heart-healthy eating, physical activity, and tobacco cessation in developing their interventions (<xref rid="B6" ref-type="bibr">6</xref>). Beyond this, the CDC does not prescribe the lifestyle intervention, preferring instead to have projects develop or select a culturally appropriate intervention shown by scientific evidence to be effective either in lowering blood pressure or cholesterol levels or in improving diet and physical activity (<xref rid="B6" ref-type="bibr">6</xref>). Lifestyle interventions therefore vary among states and tribes. Although state or tribal programs often dictate which intervention should be used at their local sites, they sometimes allow flexibility in how sites implement the intervention. To distill a set of best practices from these highly variable interventions, Mathematica Policy Research, Inc (MPR) reviewed the literature on lifestyle interventions and collected qualitative data from sites through interviews, observation, and focus groups. The best practices identified will be disseminated to existing and new WISEWOMAN practitioners through a user-friendly toolkit.</p></sec><sec><title>Methods</title><p>We conducted in-depth case studies of selected WISEWOMAN projects and of high-performing and low-performing local sites within each project. Case studies allowed us to explore how and why projects and local sites used certain practices, providing insight into the relationship between program implementation and program effectiveness (<xref rid="B7" ref-type="bibr">7</xref>). The practices of high-performing sites were compared with practices of low-performing sites to identify if and how they differed. To conduct this mixed-methods evaluation, we applied the following five steps: 1) site selection using quantitative program performance data, 2) development of a conceptual framework for guiding qualitative inquiry, 3) development and refinement of data collection instruments, 4) collection of qualitative data, and 5) analysis of qualitative data to identify best program practices.</p><p>We applied the RE-AIM (Reach, Effectiveness, Implementation, Adoption, and Maintenance) model, developed by Glasgow et al (<xref rid="B8" ref-type="bibr">8</xref>), as an organizing framework for our study. The purpose of the model is to facilitate evaluations of the translatability and overall public health impact of a health promotion intervention (<xref rid="B8" ref-type="bibr">8</xref>). The framework specifies dimensions at the individual and institutional levels. Dimensions are defined as the following: 1) the intervention's <italic>reach</italic> into the target population, 2) its <italic>effectiveness</italic> in modifying risk, 3) its <italic>adoption</italic> by target settings, 4) its consistent <italic>implementation</italic>, and 5) <italic>maintenance</italic> of its effects among participants and target settings (<xref rid="B8" ref-type="bibr">8</xref>). Including both the individual and institutional levels in the framework allows an investigation of whether an intervention has an impact on one level but not the other. Both are essential in determining which interventions will work in other settings and have a strong impact on overall public health.</p><sec><title>Step 1: site selection</title><p>The first step of this evaluation was to select state or tribal projects and identify the highest-performing and lowest-performing local sites within each project using the RE-AIM framework. Fifteen state or tribal WISEWOMAN projects are currently funded; each has been operating for a different length of time. Through a separate contract, the CDC worked with the Research Triangle Institute (RTI) to select five projects for the MPR study. The five projects selected had program data for at least 100 women per local site, which allowed the tracking of participants from the time of program enrollment through a rescreening 10 to 14 months later.</p><p>The CDC and RTI then developed a method for rating performance along RE-AIM dimensions for three local sites — the two highest-performing and one lowest-performing — from each of the five projects, providing a total of 15 sites for case studies (R.P. Farris, J. Will, O. Khavjou, E.A. Finkelstein, unpublished data, 2004). MPR and a WISEWOMAN consultant group provided input on site selection methods. <xref rid="F1" ref-type="fig">Figure 1</xref> identifies the data elements and measures that were used to calculate the RE-AIM scores for site selection. For example, <italic>reach</italic> is measured with 1) total number of screenings, 2) total number of women screened for the first time, 3) percentage of NBCCEDP participants screened for WISEWOMAN, 4) percentage of minority NBCCEDP participants screened for WISEWOMAN, and 5) percentage of women attending at least one intervention session. Effectiveness is measured with 1-year average changes in 1) systolic blood pressure, 2) total cholesterol, 3) body weight, and 4) percentage change in smoking rate.</p><boxed-text position="float"><fig position="float" id="F1" fig-type="diagram"><label>Figure 1</label><caption><p>Site selection method for WISEWOMAN study on best practices using the RE-AIM framework. RE-AIM indicates Reach, Effectiveness, Adoption, Implementation, and Maintenance (<xref rid="B8" ref-type="bibr">8</xref>). [A full-page version of this figure is available in PDF format (166K)]</p></caption><alt-text>Flow chart</alt-text><media xlink:href="05_0133_01.pdf" id="f1m1" position="anchor"/><graphic xlink:href="PCD31A07s01" position="float"/></fig></boxed-text><p>Program data from each project were analyzed separately, and each site's performance was ranked as follows: first, sites were ranked from highest to lowest on each of five RE-AIM dimensions. For every site, the rank scores within each dimension were averaged. Then, within each dimension, the averaged rank score for each site was scaled from 0 to 100 (normalized) to allow for comparisons across dimensions. Normalized scores for each site were then averaged across dimensions to create a composite RE-AIM score that measures the overall public health impact (success) of each site (<xref rid="F2" ref-type="fig">Figure 2</xref>). The figure also illustrates the resulting normalized rank score for each dimension in three local sites from one project. After the one low-performing and two high-performing sites were selected for a project, researchers were given the names of the sites but not their rank. This approach ensured that rank did not influence perception of local sites during qualitative data collection and initial data analysis.</p><boxed-text position="float"><fig position="float" id="F2" fig-type="diagram"><label>Figure 2</label><caption><p>Overall RE-AIM score (an average of five RE-AIM scores) and individual RE-AIM scores for each of three local sites within one WISEWOMAN project. RE-AIM indicates Reach, Effectiveness, Adoption, Implementation and Maintenance (<xref rid="B8" ref-type="bibr">8</xref>)</p></caption><alt-text>Bar chart</alt-text><alternatives><graphic xlink:href="PCD31A07s02" position="float"/><table frame="hsides" rules="groups" border="1" cellpadding="4" cellspacing="1"><thead><tr><th align="left" valign="top" rowspan="1" colspan="1"> </th><td align="center" valign="top" colspan="6" rowspan="1">
<bold>Score</bold>
</td></tr><tr><th scope="col" align="left" valign="bottom" rowspan="1" colspan="1">
<bold>Site</bold>
</th><th scope="col" valign="bottom" rowspan="1" colspan="1">Reach</th><th scope="col" valign="bottom" rowspan="1" colspan="1">Effectiveness</th><th scope="col" valign="bottom" rowspan="1" colspan="1">Adoption</th><th scope="col" valign="bottom" rowspan="1" colspan="1">Implementation</th><th scope="col" valign="bottom" rowspan="1" colspan="1">Maintenance</th><th scope="col" valign="bottom" rowspan="1" colspan="1">Overall</th></tr></thead><tbody><tr><td scope="row" align="left" valign="top" rowspan="1" colspan="1">Local High Performer 1</td><td align="right" valign="top" rowspan="1" colspan="1">100</td><td align="right" valign="top" rowspan="1" colspan="1">100</td><td align="right" valign="top" rowspan="1" colspan="1">25</td><td align="right" valign="top" rowspan="1" colspan="1">100</td><td align="right" valign="top" rowspan="1" colspan="1">57</td><td align="right" valign="top" rowspan="1" colspan="1">76.4</td></tr><tr><td scope="row" align="left" valign="top" rowspan="1" colspan="1">Local High Performer 2</td><td align="right" valign="top" rowspan="1" colspan="1">86</td><td align="right" valign="top" rowspan="1" colspan="1">86</td><td align="right" valign="top" rowspan="1" colspan="1">46</td><td align="right" valign="top" rowspan="1" colspan="1">71</td><td align="right" valign="top" rowspan="1" colspan="1">71</td><td align="right" valign="top" rowspan="1" colspan="1">72.0</td></tr><tr><td scope="row" align="left" valign="top" rowspan="1" colspan="1">Local Low Performer</td><td align="right" valign="top" rowspan="1" colspan="1">15</td><td align="right" valign="top" rowspan="1" colspan="1">29</td><td align="right" valign="top" rowspan="1" colspan="1">75</td><td align="right" valign="top" rowspan="1" colspan="1">29</td><td align="right" valign="top" rowspan="1" colspan="1">29</td><td align="right" valign="top" rowspan="1" colspan="1">35.4</td></tr></tbody></table></alternatives><table-wrap position="anchor" id="d95e255"/></fig></boxed-text></sec><sec><title>Step 2: development of conceptual framework for qualitative inquiry</title><p>The second step was to develop a conceptual framework for guiding qualitative inquiry. The overarching research questions for qualitative data collection were:</p><list list-type="bullet"><list-item><p>What are the most effective practices used in selected state or tribal WISEWOMAN projects and local sites to design and deliver lifestyle interventions?</p></list-item><list-item><p>How have projects and local sites implemented these practices?</p></list-item><list-item><p>What lessons can these practices offer to other projects and local sites?</p></list-item></list><p>
<xref rid="F3" ref-type="fig">Figure 3</xref> depicts the framework that addresses these questions.</p><boxed-text position="float"><fig position="float" id="F3" fig-type="diagram"><label>Figure 3</label><caption><p>Framework for guiding qualitative inquiry for WISEWOMAN study on best practices. [A full-page version of this figure is available in PDF format (454K)]</p></caption><alt-text>Flow chart</alt-text><media xlink:href="05_0133_03.pdf" id="f3m1" position="anchor"/><graphic xlink:href="PCD31A07s03" position="float"/></fig></boxed-text></sec><sec><title>Step 3: development and refinement of data collection instruments </title><p>The next step was to develop a data collection plan and data collection instruments based on the framework for qualitative inquiry. The instruments were piloted in one project to assess how well they captured data. They were then revised to focus more on the lifestyle intervention rather than the broad program-implementation processes.</p><p>The three objectives of the qualitative data collection plan were as follows:</p><list list-type="bullet"><list-item><p>To gather enough information about selected WISEWOMAN projects and local sites to understand the contexts in which programs operate, the design and implementation of different lifestyle interventions, and how each intervention fits into the overall approach to service delivery in each project or site</p></list-item><list-item><p>To gather detailed qualitative information on practices related to the delivery of the lifestyle intervention and the five RE-AIM dimensions</p></list-item><list-item><p>To systematically assess the most effective practices according to each RE-AIM dimension</p></list-item></list><p>To reach these objectives, we developed semistructured interview guides for each type of informant: federal project officers, state or tribal project directors and coordinators, and local coordinators, interventionists, and partners. We also developed a guide for focus groups, which were composed of program participants, and for observation of lifestyle intervention sessions, which also involved program participants. Each guide was organized by a RE-AIM dimension to ensure that data were collected systematically and consistently across projects and local sites. Although the interview guide provided questions, the order of the questions was not prescribed to researchers conducting the interviews, and the questions were not intended to be asked verbatim. <xref rid="F4" ref-type="fig">Figure 4</xref> shows the interview topics that were included in all guides. All methods involving program participants (i.e., focus groups and observations) were approved by the institutional review board (IRB) of Public/Private Ventures (Philadelphia, Pa).</p><boxed-text position="float"><fig position="float" id="F4" fig-type="diagram"><label>Figure 4</label><caption><p>Semistructured interview guide topics for WISEWOMAN study on best practices. Each of the five RE-AIM dimensions are identified in parentheses. RE-AIM indicates Reach, Effectiveness, Adoption, Implementation, and Maintenance (<xref rid="B8" ref-type="bibr">8</xref>). Not all interviews included all questions. Order of questions was not prescribed; order here reflects a logical interview flow.</p></caption><p>1. Background Information on Interview Participant</p><p>2. Overview of the WISEWOMAN Projects</p><list list-type="bullet"><list-item><p>Overall structure at state and local levels</p></list-item><list-item><p>Chronology and development</p></list-item></list><p>3. Local Site Involvement <italic>(Adoption)</italic>
</p><list list-type="bullet"><list-item><p>Recruitment strategies and contracting mechanisms with local sites</p></list-item><list-item><p>Common barriers to site participation and how they have been addressed</p></list-item></list><p>4. Participant Recruitment and Involvement <italic>(Reach)</italic>
</p><list list-type="bullet"><list-item><p>Process for recruiting participants into WISEWOMAN and the lifestyle intervention</p></list-item><list-item><p>Initiatives for targeting specific groups of women for participation</p></list-item></list><p>5. Design and Delivery of the Lifestyle Intervention <italic>(Implementation)</italic>
</p><list list-type="bullet"><list-item><p>Overall design of the intervention</p></list-item><list-item><p>Feasibility and consistency of intervention delivery</p></list-item><list-item><p>Staffing issues, including commitment, availability, training, and others</p></list-item></list><p>6. WISEWOMAN'S Effect on Participants <italic>(Effectiveness)</italic>
</p><list list-type="bullet"><list-item><p>Perceived effectiveness of the lifestyle intervention</p></list-item><list-item><p>Impact on participants' short-term and long-term quality of life</p></list-item></list><p>7. Program Sustainability <italic>(Maintenance)</italic>
</p><list list-type="bullet"><list-item><p>Challenges to ongoing service delivery</p></list-item><list-item><p>Strategies for overcoming challenges both at state and local levels</p></list-item></list><p>8. Additional Critical Themes and Wrap-up</p><list list-type="bullet"><list-item><p>Role of key community and other partners</p></list-item></list></fig></boxed-text></sec><sec><title>Step 4: qualitative data collection</title><p>Qualitative data collection consisted of preliminary data collection and site visits. The goal of preliminary data collection was to understand the context in which selected projects operated and learn how projects or sites deliver the lifestyle intervention to WISEWOMAN participants. Data were derived from a review of project and local site reports and from semistructured telephone interviews with CDC staff members who oversee the study projects, state or tribal project staff, and staff at each local site.</p><p>Site visits were conducted by two researchers at each state or tribal project and at the three selected local sites within each project. The purpose of the visits was to expand on what we learned from the preliminary data about how lifestyle interventions are developed and implemented. Toward this end, we conducted individual and small group interviews with staff and focus groups with current program participants; we also observed lifestyle intervention sessions.</p><sec><title>Interviews</title><p>In-person interviews were conducted with staff members who play a role in developing or delivering the lifestyle intervention at the project and site levels. When possible, we also interviewed staff of local program partners, such as a community swimming pool that offers discounted passes to WISEWOMAN participants as a way to increase their physical activity. The goal of the interviews was to learn about the staff's perspectives on practices used to implement the lifestyle intervention. We also gathered data to <italic>triangulate</italic>, or to combine and compare responses from multiple informants and sources, to develop a consistent understanding of lifestyle intervention implementation.</p></sec><sec><title>Focus groups</title><p>We conducted a focus group with WISEWOMAN participants at each local site to better understand the intervention's reach, effectiveness, implementation, and ability to maintain behavioral change. When recruiting women for the focus group, we attempted to select those who had participated in the lifestyle intervention recently and could address why they chose to participate, as well as women who participated more than a year ago and could address maintenance of the intervention. Each woman was given a $25 incentive to participate in the focus group. We typically invited 10 to 12 women to attend each focus group, which resulted in six to nine participants.</p></sec><sec><title>Observations</title><p>We developed procedures for observing various forms of intervention delivery, including individual in-person, telephone, and group interventions. Before researchers were permitted to observe the sessions, the interventionist solicited the participants' consent — by obtaining either a signature (for in-person and group interventions) or verbal consent (for telephone interventions). For scheduling purposes, we restricted our observations to what was scheduled the day we visited the site, so we did not know in advance the type of intervention we would observe (i.e., an initial visit or a follow-up visit).</p></sec></sec><sec><title>Step 5: Qualitative data analysis to identify best program practices</title><p>This analysis consisted of four steps: writing site reports, developing theme tables, identifying practices of interest, and applying a best-practices algorithm.</p><sec><title>Writing site reports</title><p>Both members of the site visit team worked together to write a report for each state or tribal project, for each local site within a project, and for each data source (e.g., interviews, focus groups). The rationale for joint report writing was to minimize the degree to which one researcher's impressions might influence the findings. In addition, to ensure that the reports were accurate, they were reviewed by interviewees at the state or tribal projects and local sites. To protect confidentiality, focus group data were not shared with program staff.</p></sec><sec><title>Developing theme tables</title><p>Theme tables, developed from the reports, were intended to present data in a simplified manner grouped by theme. The research team used the interview guides and narratives to identify areas, or themes, that were consistently investigated during data collection. The idea was to organize themes according to the dimensions in the RE-AIM framework to make it easier to compare themes across programs. For example, data from one case study indicated that the theme of <italic>adapting intervention materials provided by the state or tribe to local community</italic> is common to several local sites; the theme is categorized as an <italic>implementation</italic> dimension within the RE-AIM framework.</p></sec><sec><title>Identifying practices of interest</title><p>Identifying practices of interest is a multistep process: 1) select the practice within each theme that may be a best practice, 2) determine which sites use the practice of interest, 3) determine the purpose of the practice, 4) determine whether the purpose varies across sites, and 5) develop simple categories of the purpose (if applicable) that may help discern what makes it useful.</p><p>To illustrate this process, we use the example identified above: <italic>adapting intervention materials provided by the state or tribe to local community</italic>. In this example, we determined that some, but not all, local sites within the state or tribal project adapted intervention materials provided by the state or tribe to fit the needs of their community (steps 1 and 2). We found that the purpose of these adaptations varied, as did the actual adaptations (steps 3 and 4). By examining data in this way, we could assess which local sites adapted state or tribal materials and the purpose of the adaptation. We then could categorize the purposes (step 5). We learned that two sites adapted materials provided by the state by simplifying them and making them more accessible to and flexible for participants. The third site made no adaptations.</p></sec><sec><title>Applying a best-practices algorithm</title><p>Before the final step in the analysis, we obtained the rankings for each local site. This was necessary if we were to select the best practices from those identified as a practice of interest in step 3. To make the selection, we developed an algorithm to tabulate the qualitative data by applying standardized decision rules to the information (<xref rid="F5" ref-type="fig">Figure 5</xref>). Following the algorithm, we determined how many times the practice was observed and then assessed the ranking of the sites that used the practice (i.e., high-performing or low-performing). The ranking told us how well sites are reaching desired program outcomes, averaged across multiple indicators.</p><boxed-text position="float"><fig position="float" id="F5" fig-type="diagram"><label>Figure 5</label><caption><p>Algorithm for determining best practices in selected WISEWOMAN programs.</p></caption><alt-text>Flow chart</alt-text><long-desc>This algorithm begins with a box for "Practice 1." An arrow links this box to a second box, "Number of times practice was state across all sites." An arrow leads from the second box to a third, "Ranking of sites who stated practice." Three arrows point from this box to three more boxes. One is "High-performing and low-performing sites = May Be Best Practice." The second is "All high-performing sites = Best Practice." The third is "All low-performing sites = Not Best Practice. The box with "High-performing and low-performing sites = May Be Best Practice" has an arrow that points to a box "Was there a difference in how sites implemented the practice that explains why the practice worked in high-performing but not low-performing sites?" From this box, one arrow points to "Yes = Best Practice" and another points to "No = May Be Best Practice." From "No = May Be Best Practice" an arrow points to a box "Are more than two thirds of the sites that use the practice high-performing?" From this box, one arrow points to "Yes = Best Practice" and another points to "No = Not Best Practice." </long-desc><graphic xlink:href="PCD31A07s04" position="float"/></fig></boxed-text><p>When only high-performing sites were using a practice of interest, it was considered <italic>best</italic>. To allow innovative practices at a single site to be eligible for best practices, and to allow the practices in a small number of sites to be eligible, a practice was considered <italic>best</italic> if all the sites using it were high-performing, regardless of whether one or 10 sites were using it. In contrast, when only low-performing sites were using a practice, it was not considered a best practice.</p><p>It was more difficult to identify a practice as best when it was being used by both high-performing and low-performing sites. In such cases, we attempted to learn how implementation of the practice differed at each site to assess whether the higher-performing sites were doing something different to contribute to their better outcomes. If we found a difference and could identify the factor that made the difference, we considered the practice to be a best practice. If both types of sites were using a practice, and there was no obvious difference in how it was used, then we deemed the practice <italic>best</italic> if more than two thirds of the sites were high performing. This ratio is consistent with the ratio of high-performing and low-performing sites selected for each project area (i.e., two high-performing, one low-performing). If only one high-performing and one low-performing site used a practice, the practice was not considered best because only half who used the practice was high performing.</p><p>This logic should be used with caution for the following reason: a composite score was judged to be the best overall indicator of a site's performance because it does not overemphasize the importance of any one RE-AIM dimension. In the algorithm step, however, we judged the effectiveness of certain site practices on the basis of these composite scores of site performance. For this reason, practices identified as best may not always lead to the desired outcomes. The success of a practice also may depend on the context in which it was implemented. To this end, the toolkit will provide information about the context in which those best practices were implemented.</p></sec></sec></sec><sec><title>Results</title><sec><title>Lessons learned in developing and implementing the methodology</title><p>Piloting the methods in one project before collecting data from the other four projects was extremely helpful in refining our data collection instruments and methods. Data collected during the pilot using the initial protocols was overly process oriented. We realized that we needed to revise our methods to target program practices related to the lifestyle intervention. We also learned that conducting preliminary interviews with program staff provided us with an initial understanding of how the project or local site operated and helped us identify information we needed to collect onsite. The interview helped to build rapport with staff and made time available during site visits to ask more detailed questions about how and why practices were implemented.</p><p>The pilot also revealed the importance of observing a lifestyle intervention session at each local site because it gave us a much better understanding of how the intervention was actually implemented. Following the pilot, observing the lifestyle intervention became a key component of data collection in the remaining case studies. When arranging a visit to a local site, we selected and coordinated the day's activities around the observation, resulting in a greater probability of completing this piece of data collection. Nevertheless, it was a challenge to arrange the observations, particularly given privacy and confidentiality concerns. To address confidentiality concerns, we e-mailed the IRB approval of the observation protocol, the focus group guide, and the consent form to each site. We also allowed at least one month for planning a visit so we could work through any problems that arose. During this planning phase, we e-mailed a detailed list of the case study components to the sites and reviewed the components — including the observation — with them on the telephone.</p><p>Another methodological issue we encountered was difficulty in recruiting women for focus groups. Recruitment of women was especially challenging at sites that served a large geographic area: transportation over long distances was an issue for many women. Some sites recruited participants themselves because they were concerned that providing us with telephone numbers would breach confidentiality. We preferred this method because local site staff typically had a rapport with their participants. However, sites were not always able to help with recruitment because of time constraints. Advance planning was key to successful recruiting. Despite the challenges involved, the focus groups were worth the effort; they were vital for giving a voice to program participants.</p></sec><sec><title>Dissemination</title><p>Disseminating the best practices in a toolkit is a critical final step in this evaluation. The process for developing this toolkit began by conducting a needs assessment with staff from currently funded WISEWOMAN projects during their 2004 annual meeting. Through this needs assessment, we learned not only about the kinds of information that projects would find most useful but also about the formats in which the material would be accessible to projects and local sites. The result will be a toolkit that provides a portfolio of ideas and options that current and newly funded WISEWOMAN projects can use. It will present concrete information on best practices for delivering the lifestyle intervention, including the specific mechanisms used by projects or sites to develop the interventions, the challenges they faced, and other issues relevant to the replication or adaptation of these practices. The toolkit also will describe the context or contexts in which each best practice was implemented; the idea is to acknowledge and let others know that some practices work in some settings but not in others. We expect the toolkit to be completed in spring 2006.</p></sec></sec><sec><title>Discussion</title><p>By using a mix of quantitative and multiple qualitative methods in our study of WISEWOMAN programs, we minimized the weaknesses inherent in each method and improved the completeness, and therefore the applicability and quality, of the data collected. By triangulating data collected from different sources, we were able to develop a more complete picture of how programs implement the lifestyle intervention and to identify best practices in implementation. The data on practices is detailed and complete enough to provide guidance for other WISEWOMAN programs interested in replicating the practices. A mixed-methods approach is clearly a promising strategy for identifying best practices in current programs. We recommend that others interested in using this method consider the strengths and weaknesses associated with each mode of data collection before settling on an approach. We also strongly encourage researchers to pilot their methodology.</p></sec> |
This World Is Yours | Could not extract abstract | <contrib contrib-type="author"><name><surname>Wilcox</surname><given-names>Lynne S.</given-names></name><degrees>MD, MPH</degrees><role>Editor in Chief</role></contrib> | Preventing Chronic Disease | <p>
<italic>This land is your land, this land is my land, from California to the New York Island . . .</italic> (<xref rid="B1" ref-type="bibr">1</xref>). When Woody Guthrie wrote and sang these lyrics in 1940, the song was more than a paean to a beautiful country. It was a call to recognize the inherent dignity of the common man. The song's passion arose from Guthrie's experiences during the Great Depression, when he traveled with other migrants from the poverty of the Oklahoma Dust Bowl to potential employment and a new life in California (<xref rid="B2" ref-type="bibr">2</xref>). The catastrophic economic events occurring in the country, combined with a terrible drought and harsh winds, forced families to abandon their farms and travel hundreds of miles, looking for work and gambling on a new future. Steinbeck's novel <italic>The Grapes of Wrath</italic> (<xref rid="B3" ref-type="bibr">3</xref>) is the epic story of this period in history.</p><p>Songs and stories are essential but not sufficient for transforming desperation into dignity. Public health plays a critical role in the conversion. In the initial issue of <italic>Preventing Chronic Disease (PCD),</italic> we announced that a goal of this journal is to encourage a dialogue between researchers and practitioners (<xref rid="B4" ref-type="bibr">4</xref>). Such communication moves public health forward from investigating the principles for health promotion to implementing those principles in programs for all citizens, including disenfranchised populations. This first issue of our third year illustrates another aspect of that dialogue: evaluating health programs to assess the effectiveness of these principles and practices. We thank Qaiser Mukhtar and Leonard Jack of the Division of Diabetes Translation, National Center for Chronic Disease Prevention and Health Promotion, Centers for Disease Control and Prevention, for serving as guest editors for this issue.</p><p>During the Great Depression, unemployment rates rose as high as 33%, and the Works Progress Administration (WPA) was formed to provide jobs and wages across the country. The WPA included the Federal Art Project, which hired artists to design posters for government programs such as public awareness campaigns supported by federal health agencies. Many of these posters addressed the prevention and control of chronic diseases. "Obey Cancer's Danger Signals" featured an outline of a human body and recommended consulting a physician for symptoms such as irregular bleeding, changes in the appearance of a mole, or lumps in the breast (<xref rid="F1" ref-type="fig">Figure 1</xref>). The "Eat Fruit, Be Healthy" poster highlighted a drawing of grapes and an apple (<xref rid="F2" ref-type="fig">Figure 2</xref>), and "Milk for Health" emphasized that milk consumption would result in good teeth and strong bones (<xref rid="F3" ref-type="fig">Figure 3</xref>).</p><boxed-text position="float"><fig position="float" id="F1" fig-type="diagram"><label>Figure 1</label><caption><p>Cancer prevention poster from the Federal Art Project of the Works Progress Administration. (From Herzog H. Obey cancer's danger signals. 1938. Library of Congress,  Prints and Photographs Division. By the people, for the people: posters from the WPA, 1936-1943 [Internet]. Reproduction number cph 3g03639. Available from: URL: <ext-link xlink:href="http://memory.loc.gov/ammem/ndlpedu/collections/poster/file.html" ext-link-type="uri">http://memory.loc.gov/ammem/ ndlpedu/collections/poster/file.html</ext-link>.)</p></caption><alt-text>Poster showing a human figure with potentially cancerous symptoms indicated. Text reads "Obey cancer's danger signals. Do not wait for pain"</alt-text><graphic xlink:href="PCD31A01s01" position="float"/></fig></boxed-text><boxed-text position="float"><fig position="float" id="F2" fig-type="diagram"><label>Figure 2</label><caption><p>Healthy eating poster from the Federal Art Project of the Works Progress Administration. (From Eat fruit, be healthy. 1938. Library of Congress, Prints and Photographs Division. By the people, for the people: posters from the WPA, 1936-1943 [Internet]. Digital ID cph 3f05301. Available from: URL: <ext-link xlink:href="http://memory.loc.gov/ammem/ndlpedu/collections/poster/file.html" ext-link-type="uri">http://memory.loc.gov/ammem/ ndlpedu/collections/poster/file.html</ext-link>.)</p></caption><alt-text>Poster of stylized fruit illustration. Text reads "Eat Fruit, Be Healthy"</alt-text><graphic xlink:href="PCD31A01s02" position="float"/></fig></boxed-text><boxed-text position="float"><fig position="float" id="F3" fig-type="diagram"><label>Figure 3</label><caption><p>Milk promotion poster from the Federal Art Project of the Works Progress Administration. (From Milk for health. 1940. Library of Congress, Prints and Photographs Division. By the people, for the people: posters from the WPA, 1936-1943 [Internet]. Digital ID cph 3b48986. Available from: URL: <ext-link xlink:href="http://memory.loc.gov/ammem/ndlpedu/collections/poster/file.html" ext-link-type="uri">http://memory.loc.gov/ammem/ ndlpedu/collections/poster/file.html</ext-link>.)</p></caption><alt-text>Poster showing a bottle of milk and people of good health and fitness. Text reads "Milk for health. Good teeth, vitality, endurance, strong bones"</alt-text><graphic xlink:href="PCD31A01s03" position="float"/></fig></boxed-text><p>We know that these poster campaigns increased artists' incomes, but did they improve the health of U.S. citizens? World War II intervened before formal program evaluations could be conducted, although it remains unknown whether such evaluations were ever planned. Today, as Jack et al note, rigorous evaluations require recognition of real-world complexities and an approach that is multifaceted, multidisciplinary, and multidimensional (<xref rid="B5" ref-type="bibr">5</xref>).</p><p>Although we do not offer folk songs, over the past 2 years <italic>PCD</italic> has provided a forum for many voices. Our own evaluation indicates that we have published descriptions of public health programs from 37 states and research from 79 universities, federal or state agencies, and research centers. We are establishing an international presence: we now translate article abstracts into Spanish, French, and Chinese and have received Web site visitors from 67 countries.</p><p>Our hopes for <italic>PCD</italic> parallel Guthrie's aspirations for his music. "I am out to sing songs," Guthrie said, "that will prove to you that this is your world" (<xref rid="B6" ref-type="bibr">6</xref>).</p> |
A Community Coalition Board Creates a Set of Values for Community-based Research | Could not extract abstract | <contrib contrib-type="author" corresp="yes"><name><surname>Blumenthal</surname><given-names>Daniel S</given-names></name><degrees>MD, MPH</degrees><role>Professor and Chair</role><aff>Department of Community Health and Preventive Medicine, Morehouse School of Medicine</aff><address><email>dblumenthal@msm.edu</email><addr-line>720 Westview Dr SW, Atlanta, GA 30310</addr-line><phone>404-752-1854</phone></address></contrib> | Preventing Chronic Disease | <sec><title>Background</title><p>During the last 30 years, researchers have been rethinking the relationship between researchers and research participants. <italic>The</italic>
<italic>Belmont Report</italic> (<xref rid="B1" ref-type="bibr">1</xref>) outlined three ethical principles to help researchers protect the rights of research participants: respect for individuals (autonomy), beneficence, and justice. Subsequent reports have expanded and refined these principles (<xref rid="B2" ref-type="bibr">2</xref>,<xref rid="B3" ref-type="bibr">3</xref>). Simultaneously, community-based research has emerged, so researchers have had to consider the rights of communities participating in research in addition to the rights of individual participants (<xref rid="B4" ref-type="bibr">4</xref>-<xref rid="B6" ref-type="bibr">6</xref>).</p><p>Many approaches to community interactions with institutions and professionals have incorporated components of efforts during the 1970s to develop "maximum feasible community participation" (<xref rid="B7" ref-type="bibr">7</xref>-<xref rid="B9" ref-type="bibr">9</xref>). In 1971, Arnstein created a <italic>ladder of citizen participation</italic> that defined eight levels of participation in service projects (<xref rid="B10" ref-type="bibr">10</xref>). More than 20 years later, Hatch et al defined four levels of community participation in research projects (<xref rid="B11" ref-type="bibr">11</xref>). At Hatch's first level, researchers consult with people who work for human service agencies but usually do not live in the community. At the second level, community leaders are recruited to be project advisors, but the researchers retain control of the projects. At the third level, community leaders are asked to endorse the projects and assist with hiring community residents to serve in roles such as interviewers and outreach workers. At the fourth level, community representatives are <italic>first among equals,</italic> or senior partners. They define the research agenda, identify and analyze the problem to be studied, and propose possible solutions. In Hatch's model, the status of community representatives as first among equals differs from the community control model of the 1970s, in which the governance structure could have consisted entirely of community representatives.</p><p>Building on this background, several writers have defined ethical principles for community participation in research (<xref rid="B12" ref-type="bibr">12</xref>,<xref rid="B13" ref-type="bibr">13</xref>). Community-based research is now widely referred to as <italic>community-based participatory research,</italic> reflecting the general acknowledgement that community participation is desirable (<xref rid="B14" ref-type="bibr">14</xref>). However, the protection of research participants, whether individuals or communities, has traditionally been a topic addressed by researchers and ethicists, not by communities. We report one case study in which community participants created their own set of principles, or <italic>community</italic>
<italic>values,</italic> to guide research being conducted in their community.</p></sec><sec><title>Context</title><p>The Morehouse School of Medicine Prevention Research Center (PRC) was established in 1999 in Atlanta, Ga, with the support of a grant from the Centers for Disease Control and Prevention. The PRC was created in partnership with a low-income, predominantly African American community known as <italic>Neighborhood Planning Unit Y (NPU-Y)</italic>, which is one of 24 NPUs into which the city of Atlanta is divided. NPU-Y has a population of about 25,000 and comprises eight well-defined neighborhoods, each with its own neighborhood organization, and a publicly owned apartment building for senior citizens. Approximately 90% of the community residents are African American, and the median yearly household income is about $17,000. NPU-Y 1) holds a monthly meeting that can be attended by any resident, 2) elects officers, and 3) is recognized by the city (as are the other NPUs) as the advisory body on community matters.</p><p>When establishing an entity to represent the community during its interactions with the PRC, we avoided using the term <italic>advisory</italic> (e.g., Community Advisory Board) because of its implied powerlessness. Instead, we hoped to create a governance model in which the community would serve as the senior partner in its relationship with the medical school and other academic and agency collaborators. We therefore established a Community Coalition Board to which all the partners belong but on which community representatives hold the preponderance of power. Seats on the board were assigned to the medical school, two other academic institutions, six agencies (the health department, the public schools, the public housing authority, the local community health center, the area health education center, and the Empowerment Zone Corporation), and each of the neighborhoods in NPU-Y, as well as some adjoining neighborhoods. (The Atlanta Empowerment Zone was a federally designated inner-city area represented by a nonprofit corporation, the Empowerment Zone Corporation [now defunct], with the authority to award grants and offer tax concessions.) The bylaws of the Community Coalition Board state that community representatives must always hold the majority of the positions, and a community representative must serve as the chairperson.</p><p>The bylaws permit a maximum of 25 board members. At the beginning of fiscal year 2005 (October), 17 board positions had been filled — nine with community representatives, three with academic representatives, and five with agency representatives. The board chairperson was a retired elementary school teacher; the community membership was diverse and included a minister, an acupuncturist, a community center director, a television repairman, a computer analyst, a computer technician, an office worker, and a building contractor. Five of the nine community representatives were charter members of the board from 1998. None of the board members had served fewer than 4 years. (The bylaws do not specify a term of office.) Eight of the nine community members were African American.</p></sec><sec><title>Methods</title><p>The Community Coalition Board established a research agenda for the PRC as well as a set of criteria against which it could review all research protocols. The board did not want to develop a disease-specific research agenda and therefore created a broad agenda called <italic>Research Priorities</italic> that expressed the board's concern with the overall poor state of health among African Americans, particularly African American males (<xref rid="T1" ref-type="table">Table 1</xref>).</p><p>The criteria established by the board for evaluating projects includes the statement, "They [the projects] should not violate community values or standards." One of the board members questioned exactly which community values were being considered. The result of this inquiry was a 6-month board effort to create the <italic>Statement of Community Values</italic> (<xref rid="T2" ref-type="table">Table 2</xref>). Some of the value statements that were created were modified from statements of other organizations, such as the National Association of Black Social Workers; others were generated to address board members' specific concerns. After each monthly meeting, the principal investigator revised each emerging statement, and the draft was presented at the following meeting for additional assessment and refinement. The process was repeated until a final statement was created.</p></sec><sec><title>Consequences</title><p>The PRC's experience applying the values during its first 6 years of operation can be characterized by consideration of each value statement.</p><list list-type="order"><list-item><p>
<bold>Policies and programs should be based on mutual respect and justice for all people, free from any form of discrimination or bias.</bold>
</p><p>Having programs and policies based on mutual respect is the fundamental ground rule that governs all PRC research. It specifically addresses the exploitative and discriminatory experiences that some minority populations have had with research and health care in the past. When the PRC was being organized, some community members openly expressed their lack of trust in the researchers. For instance, during one NPU-Y meeting, the idea of applying for a PRC grant was being presented. A community resident stood and said, "I grew up in Tuskegee, Alabama, and I know what you researchers are about. You will exploit the community for your own purposes." It is unlikely that our reassurances convinced her of our benign intent. Likewise, a community leader who later became a member of the Community Coalition Board told a group of faculty members: "I've seen university faculty in action. You want to gather some data, then go back to your offices and write your papers. I don't want to stand in your way, but don't expect me to help you." It was not immediately clear to the community that our research would be conducted on the basis of mutual respect and justice for all people; trust that we would adhere to this principle had to be developed over time. </p></list-item><list-item><p>
<bold>All people have a right to political, economic, cultural, and environmental self-determination.</bold>
</p><p>A conflict related to the value of self-determination arose over the initial location of the PRC offices. Community members of the board insisted that the PRC should be geographically located within NPU-Y. The other board members shared the view of the staff, which was that NPU-Y did not have a suitable office building. A compromise location was eventually found just outside the NPU-Y boundary, on the other side of an expressway. The principle of self-determination dictated that the community should decide the site of its research center, but the availability of a suitable building was a limiting factor.
</p></list-item><list-item><p>
<bold>The community has the right to participate as an equal partner at every level of decision making, including needs assessment, planning, implementation, enforcement, and evaluation.</bold>
</p><p>Serving as an equal partner means that the community members should not be overruled by the researchers. As previously mentioned, the bylaws established the community as first among equals by declaring that community representatives should hold the majority of seats and that one should serve as the chairperson of the Community Coalition Board. However, resource limitations have restricted the community's ability to make certain decisions. For instance, many of the board's community members were disappointed that the PRC did not have researchers with the expertise to pursue grants in some of their areas of interest, such as complementary and alternative medicine.
</p></list-item><list-item><p>
<bold>Principles of individual and community informed consent should be strictly enforced.</bold>
</p><p>Obtaining individual informed consent has not been an issue for the board, which has deferred to the Morehouse School of Medicine Institutional Review Board (IRB). However, the board has acted on behalf of the community in providing <italic>community informed consent,</italic> which is consent of the community, through appropriate representatives, to serve collectively as the subject of a research project. The consent process has been managed by a board committee consisting entirely of community members. With the assistance of a trusted professional, the committee reviews each research protocol. The committee has commented on every protocol and has recommended changes for most of them. The response of researchers to the recommended changes has usually been to alter the protocol to address community concerns, although one researcher walked out of a meeting and withdrew her proposal after hearing the recommended changes. Since the PRC's first year, criticisms have been relatively minor and have primarily consisted of requests that the projects focus more exclusively on NPU-Y.
</p></list-item><list-item><p>
<bold>The community repudiates the targeting of people of color and lower socioeconomic status for the purpose of testing reproductive and medical procedures and vaccinations.</bold>
</p><p>A perception exists that, historically, dangerous procedures, vaccines, and contraceptives have been tested on minority populations without adequate informed consent. None of the PRC projects has proposed to test reproductive or medical procedures or vaccinations.
</p></list-item><list-item><p>
<bold>Present and future generations should be provided an education that emphasizes social and environmental issues, based on our experience and an appreciation of our diverse cultural perspectives.</bold>
</p><p>The education principle is relevant for public schools and health promotion research. All PRC intervention projects have attempted to design culturally sensitive interventions.
</p></list-item><list-item><p>
<bold>Research processes and outcomes should benefit the community. Community members should be hired and trained whenever possible and appropriate, and the research should help build and enhance community assets.</bold>
</p><p>The community's most immediate needs are for services and jobs, not research, a principle that is emphasized in the PRC's <italic>Research Priorities</italic> (<xref rid="T1" ref-type="table">Table 1</xref>)<italic>,</italic> which states: "They [projects] should have the potential to benefit the community through a health promotion intervention." The PRC requires that all observational study proposals include an explanation of how they may be beneficial for the community. Of the first 14 projects conducted by the PRC, four were observational, eight consisted primarily of intervention testing, and two included both components. The PRC has also responded by hiring a community member as a core staff member and six community members as health workers on specific projects.
</p></list-item><list-item><p>
<bold>Community members should be part of the analysis and interpretation of data and should have input into how the results are distributed. This does not imply censorship of data or of publication, but rather the opportunity to make clear the community's views about the interpretation prior to final publication.</bold>
</p><p>A perception exists that researchers have a tendency to describe low-income minority communities in excessively negative terms, so we attempted to ensure that community members would have the opportunity to provide input on publications. In practice, community members of the board have expressed little interest in reviewing manuscripts before their submission to professional journals. However, this article was reviewed and approved by an ad hoc committee of the board, which consisted of three community members.
</p></list-item><list-item><p>
<bold>Productive partnerships between researchers and community members should be encouraged to last beyond the life of the project. This will make it more likely that research findings will be incorporated into ongoing community programs and therefore provide the greatest possible benefit to the community from research.</bold>
</p><p>The partnership between researchers and the community has indeed outlived individual projects, some of which have ended. The value statement affirms that the partnership itself is important and emphasizes that much of its significance is a direct result of the services provided by the partnership. Unfortunately, the services provided by expired projects often ended with the projects.
</p></list-item><list-item><p>
<bold>Community members should be empowered to initiate their own research projects that address needs they identify themselves.</bold>
</p><p>Through a program of minigrants, we conducted several workshops on writing proposals and then provided grants of up to $5000 to community organizations located in NPU-Y and adjacent communities. Five projects have been funded, and all have been service projects rather than research projects.
</p></list-item></list></sec><sec><title>Interpretation</title><p>During the deliberations of the Community Coalition Board that produced the <italic>Research Priorities </italic>and <italic>Statement of</italic>
<italic>Community Values,</italic> the three principles in <italic>The</italic>
<italic>Belmont Report</italic> (<xref rid="B1" ref-type="bibr">1</xref>) were never mentioned. Nonetheless, the 10 values clearly reflect the principles of beneficence (values 6, 7, 9, and 10), autonomy (values 2, 3, 4, and 8), and justice (values 1 and 5).</p><p>Thus, although the 10 values determined by the board represent universal principles of ethics, they are also tailored to represent the needs of the NPU-Y community. For instance, NPU-Y has many environmental issues. It surrounds an industrial plant that is thought by the community to be a source of air pollution. It is home to a sewage treatment plant, and several polluted creeks flow through the area. Anxiety is mounting about the potential loss of green space to developers. The environmental concerns are reflected in values 2 and 6.</p><p>The <italic>Statement of Community Values</italic> also reflects concerns that are typical of low-income communities but are often of less interest to researchers. The need for jobs is recognized in value 7. The overall need for education — not just on matters related to research projects — is reflected in value 6, whereas political and economic issues are addressed in value 2. Values 7 to 10 all suggest that the research program should benefit the local community; the overall benefits to humankind are not the focus. The value-related issues that arose were often unrelated to research ethics. Instead, they involved community matters such as the location of the PRC or the research topics to be pursued.</p><p>The concept of an equal partnership in community-based research between researchers and community members (reflected in values 3 and 9 and implied in other items) differs from the concept of a more paternalistic relationship between researchers and participants in traditional clinical research. The focus in traditional clinical research is on protecting the participant from the researcher, a concept that is not emphasized in the PRC's <italic>Statement of Community Values.</italic> Rather, the emphasis is on the belief that if properly informed and treated as an equal, the community can protect its own interests. One manifestation of this perspective is the committee that reviews each protocol. Although occasionally referred to as a <italic>community IRB,</italic> it differs from an IRB in that it is composed entirely of community representatives. The committee's ability to review projects is enhanced by the fact that protocols in health promotion research are generally less technical than protocols in traditional clinical research. In addition, in the past, the term "community laboratory" has been used in community-based research. We have discouraged its use because of the implication that community residents are analogous to subjects in a laboratory.</p><p>The PRC was not the first community-based research initiative at Morehouse School of Medicine, and the existence of a relatively robust program of community-based research provided the foundation on which to build the PRC. However, the PRC has provided a home base for community research, and it is now recognized as an important part of the school's research portfolio. In addition, community-based researchers often serve as mentors for public health students who are conducting their master's thesis research.</p><p>A partnership between community members and researchers is most successful if the partners can agree on a set of values to govern research. The set of community values developed by the Morehouse School of Medicine PRC Community Coalition Board reflect generally accepted ethical principles as well as community priorities. Attention to the <italic>Statement of Values</italic> has helped us build a stable relationship that benefits the community and the research enterprise.</p></sec> |
Practical Program Evaluation: Assessing and Improving Planning, Implementation, and Effectiveness | Could not extract abstract | <contrib contrib-type="author"><name><surname>Lavinghouze</surname><given-names>S. René</given-names></name><aff>Division of Oral Health, Centers for Disease Control and Prevention, Atlanta, Ga</aff></contrib> | Preventing Chronic Disease | <fig position="anchor" id="F1" fig-type="diagram"><alt-text>Cover of Practical Program Evaluation: Assessing and Improving Planning, Implementation, and Effectiveness</alt-text><graphic xlink:href="PCD31A25s01" position="float"/></fig><p>
<italic>Practical Program Evaluation: Assessing and Improving Planning, Implementation, and Effectiveness</italic> is an excellent first step toward a much-needed program evaluation taxonomy — one that is particularly useful for those new to evaluation, as well as for seasoned evaluators who would like to encourage understanding of the evaluation process. In this four-part, 11-chapter book, Chen discusses the evaluation process from program planning to outcome assessment. He advocates a theory-driven evaluation approach that supports his taxonomy and provides a thorough review of the theory-driven approach. Chen explains that many evaluation concepts are too vague or ambiguous to apply effectively to actual evaluations. He proposes an evaluation taxonomy to exemplify a holistic approach to evaluation practice. The new taxonomy supports evaluators in their attempts to understand and apply evaluation designs, data collection techniques, and use of evaluation information at a practical program level. This is no small achievement and should be an impetus for additional work in the area of evaluation taxonomy development. The taxonomy and overview of the process are excellent tools for strengthening communication between stakeholders and evaluators. The work is replete with diagrams, examples, and definitions and would be a welcome addition to any evaluation course curriculum.</p><p>Chen states that the intended audience for this book is students who have completed an entry-level evaluation course, as well as seasoned evaluators who would like to expand their knowledge and strengthen their practical skills. He acknowledges that many of the terms and definitions presented are not consistent with terms readily found in the current literature, but he encourages the readers to broaden their understanding of evaluation. For example, Chen discards the logic model in favor of his action-model and change-model approach. Although this could be an interesting challenge for seasoned evaluators, it might prove more of a burden for novice evaluators. In the final chapters, Chen also briefly addresses some criticisms of theory-driven evaluation and challenges evaluation practitioners to think about the politics and contributions of evaluation activities.</p><p>Chen also skillfully demonstrates that stakeholder theory is a legitimate basis for theory-driven programs and needs to be explained and vetted like established theories of behavior change such as the Health Belief Model. He reminds evaluation practitioners that program theory and the program itself belong to the stakeholders. Chen proposes that the role evaluation practitioners should play varies from one of objective observer to one of a clear partner in the development and design of the program. With these varying roles, the evaluation approach will also change.</p><p>As with all the strategies and approaches he addresses, Chen provides a thorough discussion of the positives and negatives of efficiency and effectiveness evaluation. Although the book provides an insightful discussion of evaluation in the program planning, development, and maturation stages, Chen does not discuss evaluation use or discern intended users (program stakeholders). The taxonomy presented can guide evaluation practitioners through the conceptualization and implementation of approaches and methods suited to each stage of a program's development. However, it does not provide adequate guidance for obtaining stakeholder questions and evaluation priorities. Chen does argue for the inclusion of stakeholders throughout the evaluation process and proposes that use of the taxonomy will facilitate discussions between evaluators and stakeholders, but this will be derived from the program theory and stage of development rather than from the intended users and projected uses of the evaluation. Chen's discussion of qualitative, quantitative, and mixed methods still helps students conceptualize potential problems and learn how to address them as they read about the abundance of evaluation approaches available. </p><p>The book is a welcome addition to the expanding literature on evaluation, because it provides an overall conceptualization of the evaluation process from program design to implementation. Although the book is somewhat limited by its lack of discussion about how evaluation results are used, this breakthrough in conceptualization will surely encourage more work in the area.</p> |
Identifying Geographic Disparities in the Early Detection of Breast Cancer Using a Geographic Information System | Could not extract abstract | <contrib contrib-type="author" corresp="yes"><name><surname>McElroy</surname><given-names>Jane A</given-names></name><degrees>PhD</degrees><aff>Comprehensive Cancer Center</aff><address><email>jamcelroy@wisc.edu</email><addr-line>610 Walnut St, 307 WARF, Madison, WI 53726</addr-line><phone>608-265-8780</phone></address></contrib><contrib contrib-type="author" corresp="no"><name><surname>Remington</surname><given-names>Patrick L</given-names></name><degrees>MD</degrees><aff>Comprehensive Cancer Center and Department of Population Health Sciences, University of Wisconsin, Madison, Wis</aff></contrib><contrib contrib-type="author" corresp="no"><name><surname>Gangnon</surname><given-names>Ronald E</given-names></name><degrees>PhD</degrees><aff>Departments of Population Health Sciences and Biostatistics and Medical Informatics, University of Wisconsin, Madison, Wis</aff></contrib><contrib contrib-type="author" corresp="no"><name><surname>Hariharan</surname><given-names>Luxme</given-names></name><aff>Department of Molecular Biology, University of Wisconsin, Madison, Wis</aff></contrib><contrib contrib-type="author" corresp="no"><name><surname>Andersen</surname><given-names>LeAnn D</given-names></name><degrees>MS</degrees><aff>Department of Population Health Sciences, University of Wisconsin, Madison, Wis</aff></contrib> | Preventing Chronic Disease | <sec><title>Introduction</title><p>Geographic differences in health status and use of health services have been reported in the United States and internationally (<xref ref-type="bibr" rid="B1">1</xref>), including stage of breast cancer incidence and mammography screening practices (<xref ref-type="bibr" rid="B2">2</xref>). Early diagnosis of breast cancer through mammography screening improves breast cancer treatment options and may reduce mortality (<xref ref-type="bibr" rid="B3">3</xref>,<xref ref-type="bibr" rid="B4">4</xref>), yet many women in the United States are not routinely screened according to recommended guidelines (<xref ref-type="bibr" rid="B5">5</xref>). </p><p>Needs assessment to account for noncompliance with breast cancer screening recommendations has focused on personal factors related to participation, including the barriers women perceive (<xref ref-type="bibr" rid="B6">6</xref>), the role of physicians (<xref ref-type="bibr" rid="B7">7</xref>), and the role of services such as mobile vans (<xref ref-type="bibr" rid="B8">8</xref>) and insurance coverage (<xref ref-type="bibr" rid="B9">9</xref>). Evaluations of the effectiveness of interventions directed at patients, communities, and special populations have also provided important information about mammography use (<xref ref-type="bibr" rid="B10">10</xref>). However, little attention has been paid to geographic location, except to focus on inner-city and rural disparities in mammography use (<xref ref-type="bibr" rid="B11">11</xref>,<xref ref-type="bibr" rid="B12">12</xref>).</p><p>The purpose of this study was to identify geographic disparities in the early detection of breast cancer using cancer registry data. This information can be used to identify areas where increased mammography screening is needed and to understand the diffusion of innovation in an urban or a rural setting.</p><p>Cancer registry data were used for these analyses. Validity of the use of these data rests on the correlation between the percentage of breast cancer diagnosed in situ and mammography screening rates; breast cancer in situ (BCIS) (excluding lobular carcinoma in situ [13-15]) is the earliest stage of localized breast cancer and is diagnosed almost exclusively by mammography (<xref ref-type="bibr" rid="B16">16</xref>). In the 1970s, before widespread use of mammography, BCIS represented less than 2% of breast cancer cases in the United States (<xref ref-type="bibr" rid="B15">15</xref>). A nationwide community-based breast cancer screening program showed that among populations of women screened regularly, the stage distribution of diagnosed cases was skewed to earlier stages, with BCIS accounting for more than 35% (<xref ref-type="bibr" rid="B17">17</xref>). Trends in the relative frequency of BCIS are closely correlated with trends in mammography use (reflected in data from surveys of mammography providers in Wisconsin) and with trends in self-reported mammography use (reflected in data from the Behavioral Risk Factor Surveillance System) (<xref ref-type="bibr" rid="B18">18</xref>-<xref ref-type="bibr" rid="B20">20</xref>). </p><p>In Wisconsin, either a physician can refer a patient for screening or a woman can self-refer. More than 60% of the mammography imaging facilities in the state accept self-referrals (<xref ref-type="bibr" rid="B21">21</xref>). Since 1989, Wisconsin state law has mandated health insurance coverage for women aged 45 to 65 years, and Medicare covers mammography screening for eligible women (<xref ref-type="bibr" rid="B22">22</xref>). In Wisconsin, the Department of Health and Family Services provides a toll-free number through which women can contact more than 400 service providers (<xref ref-type="bibr" rid="B22">22</xref>). Finally, several programs such as the Wisconsin Well Woman Program, which is funded by the Centers for Disease Control and Prevention, provide free or low-cost screening to underserved women.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>Study population</title><p>All female breast cancer cases diagnosed from 1981 through 2000 were identified by the Wisconsin Cancer Reporting System (WCRS). The WCRS was established in 1976 as mandated by Wisconsin state statute to collect cancer incidence data on Wisconsin residents. In compliance with state law, hospitals and physicians are required to report cancer cases to the WCRS (within 6 months of initial diagnosis for hospitals and within 3 months for physicians, through their clinics). Variables obtained from the WCRS included histology (<italic>International Classification of Diseases for Oncology, 2nd Edition </italic>[ICD-02] codes), stage (0 = in situ, 1 = localized, 2–5 = regional, 7 = distant, and 9 = unstaged), year of diagnosis, county of residence at time of diagnosis, and number of incident cases in 5-year age groups by ZIP code for all breast cancer cases among women. ZIP codes and county of residences, self-reported by the women with diagnosed breast cancer, are provided to the WCRS. Only ZIP codes verified for Dane County by the U.S. Postal Service were included in the data set (n = 37). The ZIP code was the smallest area unit available for WCRS incidence data.</p></sec><sec><title>Study location and characteristics</title><p>Dane County is located in south central Wisconsin. The population of the county in 1990 was 367,085, with 20% of the population living in rural areas (<xref ref-type="bibr" rid="B23">23</xref>); approximately 190,000 people lived in Madison, the second largest city in Wisconsin and home to the University of Wisconsin. The 37 unique ZIP codes in Dane County incorporate 60 cities, villages, and towns (<xref rid="F1" ref-type="fig">Figure 1</xref>).</p><boxed-text position="float"><fig id="F1" fig-type="diagram" position="float"><label>Figure 1</label><caption><p>Map of Dane County, Wisconsin, showing capital city of Madison, major lakes, active mammogram facilities, and percentage of area classified as urban by ZIP code, using 1996 ZIP code boundaries and 1990 census data. Inset map shows location of Dane County within the state.</p></caption><alt-text>Map of Dane County, Wisconsin</alt-text><graphic xlink:href="PCD31A10s01"/></fig></boxed-text></sec><sec><title>Data analysis</title><p>We determined the percentage of breast cancer cases diagnosed as BCIS in Wisconsin and Dane County over time and by ZIP codes for Dane County. For ZIP codes that encompassed areas beyond the borders of Dane County, only women who reported their county of residence as Dane were included in the analysis. The percentage of BCIS by ZIP code was mapped using 1996 ZIP code boundary files. For 17 breast cancer cases in which the women's ZIP codes no longer existed, each ZIP code was reassigned to the ZIP code in the same location.</p><p>We used analytic methods to estimate rates of early breast cancer detection by ZIP code. Because of small numbers of BCIS cases in each ZIP code, a well-characterized statistical method was used to stabilize the prediction of rates by borrowing information from neighboring ZIP codes (<xref ref-type="bibr" rid="B24">24</xref>). This is done by using Bayesian hierarchical logistic regression models to estimate ZIP-code–specific effects on percentage of breast cancer cases diagnosed in situ (excluding lobular carcinoma in situ). ZIP-code–specific effects (log odds ratios) were modeled as a Gaussian conditional autoregression (CAR) (<xref ref-type="bibr" rid="B25">25</xref>). Using the CAR model, one assumes that the log odds ratio for one ZIP code is influenced by the average log odds ratio for its neighbors. The conditional standard deviation of the CAR model, the free parameter which controls the smoothness of the map, was given a uniform prior (<xref ref-type="bibr" rid="B24">24</xref>).</p><p>For each time period, two CAR models were fitted. The first model included age group as the only covariate. Age group effects were modeled using an exchangeable normal prior. The standard deviation of this distribution was given a uniform prior. The second model included additional ZIP-code–level covariates. Potential covariates were urban or rural status, education, median household income, marital status, employment status, and commuting time from the Summary Tape File 3 of the 1990 U.S. Census of Population and Housing (<xref ref-type="bibr" rid="B23">23</xref>). Census data from 1990 were used because 1990 is the midpoint of the years included in these analyses (1981–2000). Urban or rural status was defined as percentage of women living in each of the four census classifications: urban inside urbanized area, urban outside of urbanized area, rural farm, and rural nonfarm for each ZIP code. Education was defined as percentage of women in each ZIP code aged 25 years and older with less than a high school diploma. Median household income for each ZIP code was based on self-reported income. Marital status was defined as women aged 25 years and older in each ZIP code who had never been married. Employment status was defined as percentage of women aged 16 years and older in each ZIP code who worked in 1989. Full-time employment variable was defined as percentage of women 25 years and older in each ZIP code who worked at least 40 hours per week. Commuting time was divided into five categories of percentage of female workers in each ZIP code: worked at home, commuted 1 to 14 minutes, commuted 15 to 29 minutes, commuted 30 to 44 minutes, and commuted 45 minutes or more. Age was defined as age at diagnosis. These potential covariates were initially screened using forward stepwise logistic regression models, which included ZIP code as an exchangeable (nonspatially structured) random effect. Covariates included in the best model selected using Schwarz's Bayesian Information Criterion (BIC) (<xref ref-type="bibr" rid="B26">26</xref>) were used in the second covariate-adjusted model. The covariate effects and the intercept were given posterior priors.</p><p>Posterior estimates of the age-adjusted percentage of BCIS for each ZIP code in each time period were obtained from the CAR model. Posterior medians were used as point estimates of the parameters, and 95% posterior credible intervals were obtained. Analyses were performed using WinBUGS software (<xref ref-type="bibr" rid="B27">27</xref>). Covariate screening was performed using SAS software, version 8 (SAS Institute Inc, Cary, NC). ZIP-code–specific estimates were mapped using ESRI 3.2 ArcView software (Environmental Systems Research Institute, Redwood, Calif) and 1996 ZIP code boundary files to display the data.</p><p>As an empirical check on our mapping, we fitted a regression model to the BCIS rates by ZIP code. The dependent variable was BCIS rates (using the posterior estimates of age-adjusted percentage of BCIS), and the independent variable in the model was linear distance from the University of Wisconsin Comprehensive Cancer Center (UWCCC), located in Madison, to the centroid of each ZIP code.</p></sec></sec><sec sec-type="results"><title>Results</title><p>A total of 4769 breast cancer cases were reported in Dane County from 1981 through 2000: 825 from 1981 through 1985, 1119 from 1986 through 1990, 1239 from 1991 through 1995, and 1586 from 1996 through 2000. Percentage of cases in situ varied by age group from a high of 18% among women aged 45 to 49 years to a low of 0% among women aged 20 to 24 years. From the mid 1980s, the age group most frequently diagnosed with BCIS was women aged 45 to 49. In contrast, women aged 20 to 34 and older than 84 were the least often (≤2%) diagnosed with BCIS (data not shown). Based on the 1990 U.S. census, the total female population (aged 18 years and older) in Dane County was 145,974; 60% of the female population had more than a high school degree, and 15% of the female population aged 25 and older had never married.</p><p>In Dane County, the percentage of BCIS increased from 1.3% in 1981 to 11.9% in 2000. For the state, the percentage of BCIS increased from 1.5% in 1981 to 12.8% in 2000. From 1981 to 1993, Dane County had a higher percentage of BCIS diagnosis than the state as a whole. By the mid-1990s, the percentage of BCIS among breast cancer cases in Dane County was similar to the percentage in the state (<xref rid="F2" ref-type="fig">Figure 2</xref>). Similar results are seen when mapping the observed data (maps not shown).</p><boxed-text position="float"><fig position="float" id="F2" fig-type="diagram"><label>Figure 2</label><caption><p>Smoothed trends in percentage of breast cancer cases diagnosed in situ (excluding lobular carcinoma in situ), Dane County, Wisconsin, and Wisconsin, 1981–2000. Data point for Dane County, 1980, was estimated from Andersen et al (<xref ref-type="bibr" rid="B28">28</xref>).</p></caption><alt-text>Scatter Graph</alt-text><alternatives><graphic xlink:href="PCD31A10s02"/><table frame="hsides" rules="groups"><thead><tr><th scope="col" valign="bottom" align="left" rowspan="1" colspan="1">
<bold>Year</bold>
</th><th scope="col" valign="bottom" align="center" rowspan="1" colspan="1">
<bold>Dane County %</bold>
</th><th scope="col" valign="bottom" align="center" rowspan="1" colspan="1">
<bold>Wisconsin %</bold>
</th></tr></thead><tbody><tr><td valign="top" align="left" rowspan="1" colspan="1">1980</td><td valign="top" align="center" rowspan="1" colspan="1">4.7</td><td valign="top" align="center" rowspan="1" colspan="1">1.5</td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">1981</td><td valign="top" align="center" rowspan="1" colspan="1">1.3</td><td valign="top" align="center" rowspan="1" colspan="1">1.5</td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">1982</td><td valign="top" align="center" rowspan="1" colspan="1">0.0</td><td valign="top" align="center" rowspan="1" colspan="1">0.9</td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">1983</td><td valign="top" align="center" rowspan="1" colspan="1">2.4</td><td valign="top" align="center" rowspan="1" colspan="1">1.9</td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">1984</td><td valign="top" align="center" rowspan="1" colspan="1">4.6</td><td valign="top" align="center" rowspan="1" colspan="1">2.1</td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">1985</td><td valign="top" align="center" rowspan="1" colspan="1">5.5</td><td valign="top" align="center" rowspan="1" colspan="1">2.6</td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">1986</td><td valign="top" align="center" rowspan="1" colspan="1">8.6</td><td valign="top" align="center" rowspan="1" colspan="1">4.3</td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">1987</td><td valign="top" align="center" rowspan="1" colspan="1">7.6</td><td valign="top" align="center" rowspan="1" colspan="1">5.1</td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">1988</td><td valign="top" align="center" rowspan="1" colspan="1">13.9</td><td valign="top" align="center" rowspan="1" colspan="1">6.8</td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">1989</td><td valign="top" align="center" rowspan="1" colspan="1">6.4</td><td valign="top" align="center" rowspan="1" colspan="1">5.8</td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">1990</td><td valign="top" align="center" rowspan="1" colspan="1">11.5</td><td valign="top" align="center" rowspan="1" colspan="1">7.1</td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">1991</td><td valign="top" align="center" rowspan="1" colspan="1">8.6</td><td valign="top" align="center" rowspan="1" colspan="1">7.2</td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">1992</td><td valign="top" align="center" rowspan="1" colspan="1">12.5</td><td valign="top" align="center" rowspan="1" colspan="1">9.8</td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">1993</td><td valign="top" align="center" rowspan="1" colspan="1">9.6</td><td valign="top" align="center" rowspan="1" colspan="1">9.5</td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">1994</td><td valign="top" align="center" rowspan="1" colspan="1">8.6</td><td valign="top" align="center" rowspan="1" colspan="1">8.9</td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">1995</td><td valign="top" align="center" rowspan="1" colspan="1">11.2</td><td valign="top" align="center" rowspan="1" colspan="1">11.9</td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">1996</td><td valign="top" align="center" rowspan="1" colspan="1">12.2</td><td valign="top" align="center" rowspan="1" colspan="1">12.3</td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">1997</td><td valign="top" align="center" rowspan="1" colspan="1">15.2</td><td valign="top" align="center" rowspan="1" colspan="1">14.3</td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">1998</td><td valign="top" align="center" rowspan="1" colspan="1">13.7</td><td valign="top" align="center" rowspan="1" colspan="1">13.4</td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">1999</td><td valign="top" align="center" rowspan="1" colspan="1">13.6</td><td valign="top" align="center" rowspan="1" colspan="1">13.9</td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">2000</td><td valign="top" align="center" rowspan="1" colspan="1">11.9</td><td valign="top" align="center" rowspan="1" colspan="1">12.8</td></tr></tbody></table></alternatives></fig></boxed-text><p>
<xref rid="F3" ref-type="fig">Figure 3</xref> shows model-based estimates of the age-adjusted percentage of BCIS diagnosis by ZIP code in Dane County during four 5-year periods. These maps demonstrate the increase in the percentage of cases diagnosed as BCIS noted in <xref rid="F2" ref-type="fig">Figure 2</xref>. These maps also show that the increase in the percentage of BCIS was not uniform across Dane County. From 1981 through 1985, the entire county had uniformly low rates of BCIS (2%–3%). From 1986 through 1990, urban ZIP codes had markedly higher rates of BCIS (approximately 12%) compared with rural ZIP codes (approximately 5%). From 1991 through 1995, use of mammography screening had begun to increase in the rural ZIP codes (with a 7% rate of BCIS), although the rates of BCIS remained higher in urban ZIP codes (12%). From 1996 through 2000, mammography screening was fairly universal across the county, with BCIS rates of 13% to 14%. Similar patterns were observed from models that adjusted for additional covariates of marital status and education (data not shown).</p><boxed-text position="float"><fig id="F3" fig-type="diagram" position="float"><label>Figure 3</label><caption><p>Model-based estimates of age-adjusted percentage of breast cancer cases diagnosed in situ during four 5-year periods, by ZIP code, Dane County, Wisconsin, 1981–2000. BCIS indicates breast cancer in situ.</p></caption><alt-text>Maps of Dane County, Wisconsin</alt-text><graphic xlink:href="PCD31A10s03"/></fig></boxed-text><p>From 1981 through 1985, there was no significant relationship between distance from UWCCC and the rate of BCIS (<italic>P</italic> = .27). From 1986 through 1990 and from 1991 through 1995, there was strong evidence of an inverse relationship between distance from UWCCC and the rate of BCIS (i.e., the closer to UWCCC, the higher the BCIS rate [<italic>P</italic> < .001] for both periods). From 1996 through 2000, there was a nonsignificant inverse relationship between distance from UWCCC and the rate of BCIS (<italic>P</italic> = .07).</p></sec><sec sec-type="discussion"><title>Discussion</title><p>The frequency of BCIS diagnosis increased substantially in Wisconsin and in Dane County from 1981 through 2000. This increase in percentage of BCIS among diagnosed breast cancer cases is consistent with increases in self-reported mammography use, Wisconsin Medicare claims for mammography, and the number of medical imaging centers in Wisconsin (<xref ref-type="bibr" rid="B21">21</xref>). However, progress in mammography screening was not uniform across Dane County, and this lack of uniformity represents a classic case of diffusion of innovation. Early adopters of mammography use lived in and near the city of Madison. We can speculate that Madison embodies one characteristic that accelerates the diffusion process: namely, a more highly educated population living in a university community with a strong medical presence. One predictor of mammography use is education: women who are more educated are more likely to ask their physician for a referral or to self-refer (<xref ref-type="bibr" rid="B29">29</xref>), and the strongest predictor of mammography use is physician referral (<xref ref-type="bibr" rid="B30">30</xref>). Furthermore, physicians are more likely to have chosen to live in the Madison area instead of a more rural location because they value the opportunity for regular contact with the medical school and the medical community (<xref ref-type="bibr" rid="B31">31</xref>). Consequently, a greater number of interpersonal networks and more information exchange among physicians about adoption of this innovation might have occurred earlier in the Madison medical community than in the more rural areas of the county (<xref ref-type="bibr" rid="B32">32</xref>). </p><p>Although median household income by ZIP code was not a predictor of mammography use in our study, the amount of disposable income by individuals, which is not captured by this variable, might also have been an important factor for early adopters (<xref ref-type="bibr" rid="B33">33</xref>,<xref ref-type="bibr" rid="B34">34</xref>). In a national study of mammography use, income was a significant predictor of repeat screening in 1987 but not in 1990 (<xref ref-type="bibr" rid="B35">35</xref>). In the mid-1980s, few insurance plans covered mammography screening. Therefore, women of higher socioeconomic status (SES) would have been more likely to be able to pay the cost of the mammogram. Efforts to reduce costs, such as a 1987 statewide promotional campaign sponsored by the American Cancer Society, still required a $50 copay from women who were able to self-refer for a mammogram (<xref ref-type="bibr" rid="B36">36</xref>).</p><p>As the use of this technology diffused outward, increasing numbers of women living in suburban and rural areas surrounding Madison elected to get a mammogram. From 1996 through 2000, the geographic disparity in mammography use was muted, although the eastern corridor of Dane County still had slightly lower rates of BCIS than other parts of the county. The reasons for persistent disparity in this region of Dane County are unclear: it is unlikely to be because of proximity to mammography screening facilities, nor are the ZIP-code–level SES measures such as percentage unemployed, household income, percentage below poverty level, or education level statistically different from the western corridor of Dane County.</p><p>Differences in the trends of early detection of breast cancer within Dane County suggest that progress in mammography screening has not been uniform across the county. From 1996 through 2000, while more than 14% of age-adjusted breast cancer cases were diagnosed as BCIS in Madison, fewer than 6% of age-adjusted breast cancer cases were diagnosed as BCIS in a few outlying and more rural areas of Dane County, reflecting lower mammography use by residents in this area. The results of an earlier analysis of these data were shared with local health department staff in rural Dane County who were working to increase early detection efforts through outreach education and referrals to providers. As suggested by Andersen et al, strategies to improve mammography use include improving access to primary care physicians, increasing the number of mammography facilities located in rural areas, and increasing outreach efforts by a network of public health professionals promoting screening in their community (<xref ref-type="bibr" rid="B28">28</xref>). In addition, pointing out the variations in care may lead to improvements, since the first step toward change is identifying a problem. With identification of particular areas of need, resources can be garnered toward alleviating the disparity.</p><p>Persistent disparities in mammography use after adjusting for community level of educational attainment and marital status were found. Other studies have found that patients with cancer living in census tracts with lower median levels of education attainment are diagnosed in later disease stages than are patients in tracts with higher median levels of education (<xref ref-type="bibr" rid="B29">29</xref>). Studies have also shown that one predictor for getting a mammogram is being married (<xref ref-type="bibr" rid="B37">37</xref>).</p><p>This study demonstrates the use of percentage of BCIS as a tool for comparing population-based mammography screening rates in different geographic areas. Using cancer incidence data to monitor population-based rates of breast cancer screening is possible throughout the nation, because data from population-based cancer registries are now widely available, often by ZIP code or census tract. This method permits comparison of mammography screening rates among geographic areas smaller than areas used in many previous studies of geographic variation in the early detection of breast cancer (<xref ref-type="bibr" rid="B2">2</xref>). </p><p>The method described in this article can be used to complement other ways to assess the quality of health care in communities, such as the Health Plan Employer Data and Information Set (HEDIS), created by the National Committee for Quality Assurance. HEDIS addresses overall rates in managed care but does not include the underinsured or fee-for-service populations particularly at risk for inadequate screening (<xref ref-type="bibr" rid="B34">34</xref>). Cancer registry data are population based; therefore, using cancer registry data is not only effective but also economical and efficient for outreach specialists and health providers.</p><p>A potential weakness in this method is the representativeness of the statewide tumor registry. However, the WCRS has been evaluated by the North American Association of Central Cancer Registries and was given its gold standard for quality, completeness, and timeliness in 1995 and 1996, the first 2 years this standard was recognized (<xref ref-type="bibr" rid="B38">38</xref>). Completeness estimates are a general measure of accuracy. The WCRS participated in national audits that measured completeness in 1987, 1992, and 1996 as well as one formal study in 1982. Overall, the quality of the data improved slightly after 1994 when clinics and neighboring state data-sharing agreements were implemented (oral communication, Laura Stephenson, WCRS, July 2005). In addition, the tumor registry has used standard methods for classifying tumor stage (e.g., in situ) throughout the entire period of the study. Incidence data from data sources of lesser quality or completeness than the WCRS would need to be carefully evaluated for use in this type of analysis.</p><p>Another limitation of this type of analysis is our use of BCIS as a proxy for mammography screening practices. Undoubtedly, some diagnoses of BCIS result from diagnostic mammograms, but reported use of screening mammograms by individuals and medical facilities correlates strongly with percentage of BCIS over time, particularly ductal carcinoma in situ (<xref ref-type="bibr" rid="B18">18</xref>-<xref ref-type="bibr" rid="B20">20</xref>). Furthermore, we chose to exclude lobular carcinoma in situ from our BCIS category because this condition is often opportunistic (<xref ref-type="bibr" rid="B13">13</xref>-<xref ref-type="bibr" rid="B15">15</xref>). </p><p>A third limitation, which would be found in any type of geographic analysis, rests on the accuracy of the assignment of participants to the proper location. For area analysis (e.g., ZIP code, county), this legitimate concern is ameliorated by using tools to check ZIP codes and county assignments for correctness. For this study, women diagnosed with breast cancer provided their addresses, including county of residence, to their medical facilities. These addresses were forwarded to the WCRS, where quality-control checks were implemented to validate ZIP code and county assignments. For example, reference tables of ZIP codes and their county codes were cross-referenced to the ZIP codes and county codes of the addresses provided by the women diagnosed with breast cancer. Inaccuracies were corrected by the WCRS (oral communication, Laura Stephenson, WCRS, January 2005).</p><p>Although there has been significant improvement in breast cancer screening across the state and county, this study demonstrates that the improvement has not been uniform. The maps clearly indicate for program directors and policy makers the areas where further outreach and research should be conducted. More specifically, this type of analysis can be used to identify specific areas (such as ZIP codes) within a community (such as a county) with varying rates of early-stage breast cancer. Using this method, public health professionals can provide population-level data to all health care providers to target interventions to improve the early detection of breast cancer in other counties in Wisconsin and other states. Finally, this type of analysis is useful for comprehensive cancer control efforts and can be conducted for other cancers with effective screening methods, such as colorectal cancer.</p></sec> |
Using Focus Groups in the Consumer Research Phase of a Social Marketing Program to Promote Moderate-Intensity Physical Activity and Walking Trail Use in Sumter County, South Carolina | Could not extract abstract | <contrib contrib-type="author" corresp="yes"><name><surname>Burroughs</surname><given-names>Ericka</given-names></name><degrees>MA, MPH</degrees><role>Program Coordinator</role><aff>Prevention Research Center, Arnold School of Public Health, University of South Carolina</aff><address><email>burrouge@gwm.sc.edu</email><addr-line>730 Devine St, Columbia, SC 29208</addr-line><phone>803-777-4253</phone></address></contrib><contrib contrib-type="author"><name><surname>Peck</surname><given-names>Lara E</given-names></name><degrees>MPH, CHES</degrees><aff>Prevention Research Center, Arnold School of Public Health, University of South Carolina, Columbia, SC</aff></contrib><contrib contrib-type="author"><name><surname>Sharpe</surname><given-names>Patricia A</given-names></name><degrees>MPH, PhD</degrees><aff>Prevention Research Center, Arnold School of Public Health, University of South Carolina, Columbia, SCa</aff></contrib><contrib contrib-type="author"><name><surname>Granner</surname><given-names>Michelle L</given-names></name><degrees>PhD</degrees><aff>Prevention Research Center, Arnold School of Public Health, University of South Carolina, Columbia, SC</aff><aff>Dr Granner is now with the School of Public Health, University of Nevada, Reno, NV</aff></contrib><contrib contrib-type="author"><name><surname>Bryant</surname><given-names>Carol A</given-names></name><degrees>PhD</degrees><aff>College of Public Health Arts and Sciences, University of South Florida, Tampa, Fla</aff></contrib><contrib contrib-type="author"><name><surname>Fields</surname><given-names>Regina</given-names></name><degrees>MS, CHES</degrees><aff>South Carolina Healthy Schools, South Carolina Department of Education, Columbia, SC</aff></contrib> | Preventing Chronic Disease | <sec><title>Introduction</title><p>In a community-based participatory research study funded by the Centers for Disease Control and Prevention (CDC), researchers worked with a community advisory council to design a social marketing program promoting moderate-intensity physical activity (specifically walking and trail use) in Sumter County, South Carolina. The county has 104,000 residents and is one of many rural counties located in South Carolina. The population is 50.1% white, 46.7% black, and 3.2% other racial or ethnic groups, mainly Hispanic (<xref rid="B1" ref-type="bibr">1</xref>).</p><p>In 2003, consumer research was conducted by analyzing a year of quantitative and qualitative data. Quantitative data were collected with a countywide random-digit–dialed telephone survey (E.L.B., unpublished data, 2003). Qualitative data were collected through focus groups; this article describes the qualitative findings and illustrates the value of using focus groups as a consumer research tool. </p><p>Qualitative data, such as those collected in focus groups, provide in-depth information necessary to understanding attitudes and motivations that influence consumers' decisions and behavior. Focus groups give participants an opportunity to describe their experiences and preferences without the limitation of preset response categories (<xref rid="B2" ref-type="bibr">2</xref>). Focus groups also are helpful because researchers want participants to interact with each other, sharing ideas in their own words and responding to each other's comments.</p><p>Social marketing is a strategy that employs principles of commercial marketing to influence consumer behavior or decision making. It represents a set of principles comprising an approach to health promotion but not a theory of behavioral change; therefore, social marketing campaigns often include other behavioral-change strategies, such as enhancing social support and promoting behavioral skills. Social marketing was selected to promote physical activity across the county because it can influence individual-level behavior change through community-wide interventions.</p><p>A key feature of social marketing is its consumer-centered approach, meaning that the consumers' perceived needs, values, and preferences are considered before a behavior modification strategy is determined and promoted. Variables known as the "four Ps" (<italic>product, price, place,</italic> and <italic>promotion</italic>) influence consumers' decisions to change their behaviors (<xref rid="B3" ref-type="bibr">3</xref>). Consumer research is the first step to assessing consumer preferences for and opinions about these variables.</p><p>In social marketing,<italic> product</italic> refers to the behavior being promoted and its associated benefits. As Kotler et al (<xref rid="B4" ref-type="bibr">4</xref>) note, marketers distinguish between the actual product (the behavior being promoted) and the core product (the benefits of the desired behavior). Because competition is an essential component of marketing, social marketing programs attempt to identify the benefits that best distinguish the product from its competition (<xref rid="B5" ref-type="bibr">5</xref>). <italic>Positioning</italic> refers to how the product will be aligned relative to competing demands in consumers' minds. To increase demand for the product, it must be positioned to minimize competition from these competing demands. For example, some women may attend to family matters before caring for themselves, or they may choose to forgo physical activity to watch their favorite television show.</p><p>In social marketing, <italic>price</italic> is best conceptualized as the costs people must exchange for product benefits. Price includes emotional, social, and psychological costs in addition to monetary exchanges and includes barriers to adopting or purchasing a behavior or product. According to exchange theory, if the consumer perceives that the price is too high, the product is not likely to be purchased. Conversely, if the consumer feels a product is affordable and the costs are outweighed by the benefits, the product is likely to be purchased. As stated by McCormack, "The objective of the pricing strategy is to lower the perceived costs and/or make them more acceptable to potential consumers" (<xref rid="B3" ref-type="bibr">3</xref>). </p><p>
<italic>Place</italic> can refer to the ideal location where consumers can obtain information about the product (<xref rid="B3" ref-type="bibr">3</xref>). In this study, place refers to when and where consumers can practice the product or behavior, with the goal of making it more accessible (<xref rid="B5" ref-type="bibr">5</xref>). With walking, the place can range from a treadmill at home to a local gym or a trail. If consumers are dissatisfied with the places where the behavior can be practiced, they are less likely to adopt it on a regular basis. Because there are several new walking trails and paths in Sumter County that are not often used, the community advisory council and the local county recreation department wanted to increase trail use in the county.</p><p>
<italic>Promotion</italic> involves a carefully designed set of activities intended to influence change and usually includes a wide range of activities to create demand for the product (<xref rid="B6" ref-type="bibr">6</xref>). The promotional strategy highlights the benefits of the product while minimizing the price and offering attractive places where the consumer may gain access to the product. Promotion also includes any outreach, media, and incentives that would be used to market the product to the consumer (<xref rid="B5" ref-type="bibr">5</xref>). </p><p>Social marketing has been used successfully to increase breast cancer screening (<xref rid="B3" ref-type="bibr">3</xref>) and breastfeeding rates (<xref rid="B7" ref-type="bibr">7</xref>). <italic>The Guide to Community Preventive Services</italic> (<xref rid="B8" ref-type="bibr">8</xref>), a comprehensive review of the effectiveness of interventions to increase physical activity, found strong evidence for community-wide campaigns and social-support interventions.  </p><p>Insufficient evidence is available to assess the efficacy of mass media campaigns (<xref rid="B8" ref-type="bibr">8</xref>); however, mass media interventions have been used to promote physical activity (<xref rid="B9" ref-type="bibr">9</xref>,<xref rid="B10" ref-type="bibr">10</xref>). All three of these approaches — community-wide campaigns, social-support interventions, and mass media campaigns — may be incorporated into a social marketing campaign; however, there have been few published studies of social marketing campaigns to promote physical activity (<xref rid="B11" ref-type="bibr">11</xref>,<xref rid="B12" ref-type="bibr">12</xref>).</p></sec><sec><title>Methods</title><p>Focus groups were conducted in two phases. All participants resided in Sumter County, South Carolina. The first phase of the focus groups included diverse members of the community, both physically active and inactive, to characterize the opinions on and preferences for physical activity promotion at the broadest level and to identify and characterize community subgroups by age, sex, race or ethnicity, activity level, and socioeconomic status (SES). This initial characterization guided a tailored marketing approach and selection of the target audience. </p><p>The University of South Carolina Institutional Review Board approved the study procedures. For each set of focus groups, a project coordinator used a purposive convenience sampling approach (<xref rid="B13" ref-type="bibr">13</xref>,<xref rid="B14" ref-type="bibr">14</xref>), an approach in which participants are selected based on certain characteristics. For this study, these characteristics included age, sex, race or ethnicity, activity level, and SES. The project coordinator first recruited community members who then volunteered to recruit group participants by word of mouth and by distributing flyers about the focus groups in their social and occupational circles. The project coordinator then screened potential participants to determine whether they met the parameters for the focus group and to ensure each group was as homogeneous as possible. Before each group discussion, participants were informed about the purpose of the project and provided an informed consent agreement. Each was offered a $20.00 money order to offset any costs, such as transportation or childcare, associated with participating in the group.</p><p>Each group discussion was held at a local community center, a local church, or the county recreation and parks department. They were moderated by an experienced interviewer, audiotaped, and transcribed verbatim by a professional transcriptionist. Additionally, a recorder took notes to assist with accurate transcription.</p><sec><title>Phase 1</title><p>Twelve groups varying in age, race, and activity level were selected for Phase 1 (<xref rid="T1" ref-type="table">Table 1</xref>). There were 27 men and 63 women for a total of 90 participants. Before the discussion, participants completed an anonymous questionnaire on their demographic characteristics. Fifty seven (63.0%) of the participants were black, 30 (33.0%) were white, 2 (2.2%) were Hispanic or Asian, and 1 (1.1%) was unidentified. The majority of the groups included irregularly active participants. Irregularly active was defined as participating in some physical activity but not enough to meet current physical activity recommendations of 30 minutes of moderate-intensity physical activity per day on 5 or more days of the week or 20 minutes of vigorous activity on 3 or more days per week (<xref rid="B15" ref-type="bibr">15</xref>,<xref rid="B16" ref-type="bibr">16</xref>). One group of men and one group of women interviewed, however, did meet the recommendation. Two special activity groups also participated; these included older adult shopping-mall walkers and female trail users.</p><p>Each focus group discussion lasted 75 to 90 minutes. The moderator asked 17 scripted questions on product, price, place, and promotion and probed further as needed. The research staff constructed the questions with input from the community advisory council, which consisted of county residents with an interest in community health initiatives, after the council members were trained in social marketing strategies. The community advisory council training included an overview of the four principles of social marketing and discussions about the difference between commercial marketing and social marketing, audience segmentation, the role of research results in developing a marketing plan, and choosing a target audience. The Phase 1 and Phase 2 interview guides are available from <ext-link xlink:href="http://prevention.sph.sc.edu/sips/PRPA_focusgroups.pdf" ext-link-type="uri">http://prevention.sph.sc.edu/sips/PRPA_focusgroups.pdf</ext-link>.</p><p>As summarized in the Appendix, Phase 1 focus group questions assessed participants' perceptions of the product, its benefits and positioning, costs or pricing strategy, placement strategies, and promotional strategies in addition to their experiences, preferences, and attitudes concerning physical activity in general and walking in particular. During Phase 1, no specific duration or intensity of physical activity was described.</p><p>Two researchers independently assigned conceptual codes to sections of the transcribed text and met to reach consensus on coding categories. Analysis included both coding based on transcriptions from the focus groups and coding assisted by the NVivo (QSR International Pty Ltd, Victoria, Australia) software package. The results were categorized by the four Ps and by the discussion questions.</p></sec><sec><title>Phase 2</title><p>The second phase consisted of five focus groups of 43 newly recruited women aged 35 to 54. Two of the groups were composed of regularly active women, and three of the groups were composed of irregularly active women. According to the anonymous questionnaire that the women completed, 58% of the 43 women interviewed were black and 40% were white. A description of Phase 2 focus group characteristics is shown in <xref rid="T2" ref-type="table">Table 2</xref>.</p><p>Phase 2 further defined the product (walking) and its benefits; it also identified possible strategies for promoting the product. Participants were first asked to distinguish between <italic>physical activity </italic>and<italic> exercise </italic>and to define<italic> moderate activity</italic>. As noted in the <xref rid="A1" ref-type="app">Appendix</xref>, women also were asked to share their opinions about using pedometers, to describe images that motivated them to be active, and to name or describe convincing spokespersons who would encourage them and their peers to be active. Finally, the women were asked about the radio and television stations they and their peers were most likely to listen to and how they would like to be contacted if they chose to participate in an exercise intervention. </p><p>In Phase 2, two researchers also assigned conceptual codes independently to sections of transcribed text and met to reach consensus on coding categories. Because of the smaller number of focus groups in Phase 2 and the smaller number of transcripts compared with Phase 1, the use of analysis software to manage the transcription data was not essential. As in Phase 1, note-based analysis was used in Phase 2 following the same procedures for identifying and summarizing key themes.</p></sec></sec><sec><title>Results</title><sec><title>Phase 1: participant perceptions related to the principles of social marketing</title><p>On the basis of Phase 1 focus group findings, the community advisory council selected irregularly active women aged 35 to 54 as the target audience. They also identified the behavioral recommendation or actual product (<xref rid="B4" ref-type="bibr">4</xref>) as walking and other moderate-intensity physical activities for at least 30 minutes 5 or more days per week.</p><p>
<bold>Product and product positioning</bold>
</p><p>In this study, walking and gardening were the types of moderate-intensity physical activity preferred by participants in both phases of focus groups. The participants agreed, however, that walking should be the primary activity promoted in a social marketing program to increase physical activity because walking is easily accessible, low impact, and inexpensive. A few, principally those in regularly active groups, felt that walking should be just one of several aerobic activities promoted.</p><p>Participants cited several benefits of walking, including transportation, relieving stress, and spending time with family and friends, but all groups cited health as the primary reason to walk. Walking as a means of transportation was mentioned only in groups that included participants of lower SES; for example, participants who received public assistance or were in blue-collar positions. </p><p>When asked about how they would motivate their friends to choose walking as a daily activity, participants responded with a variety of strategies that would position walking competitively with other preferred activities. Their responses suggest that walking may be positioned effectively by emphasizing body toning, social support opportunities, and decreased health risks that can result from regular physical activity. For example, having a walking companion may appeal to some women, thus combining social opportunities with exercise.</p><p>
<bold>Price</bold>
</p><p>The perceived price of walking varied among participants; however, most believed that their daily schedules were too full to include regular walking. In addition to time constraints, discomfort from the hot, humid climate and insects were commonly cited costs. </p><p>
<bold>Place</bold>
</p><p>When participants were asked to describe their walking routine (i.e., when they walk, with whom, and what time of day), most preferred walking in the morning or early afternoon. All groups included a mix of participants who preferred walking alone or with someone. However, most female participants, regardless of age or activity level, preferred to walk with someone. Participants often cited their neighborhood as their favorite place to walk. The local shopping mall offered a walking program, but many focus group participants incorrectly assumed the program was for older adults only.</p><p>Sumter County has several public, multiuse trails and tracks, both paved and unpaved. Because the presence of trails has been shown to be associated with meeting the CDC's and American College of Sports Medicine's recommendation for physical activity (<xref rid="B17" ref-type="bibr">17</xref>), participants were asked about their perception of local trails. Their responses conflicted. Although most people appreciated trails for their abundance of foliage and tranquility, these same qualities led others to suspect that they were not safe or secure, especially for women. Some participants also found pet waste, litter, and inadequate lighting to be deterrents to using the trails.</p><p>
<bold>Promotion</bold>
</p><p>In this study, all focus groups highly favored walking groups, even though some individuals preferred walking alone. Many acknowledged that walking groups could provide walkers with social support, security, and childcare solutions. Regularly active men clearly stated that walking groups were appealing for women only. Unlike self-paced programs in which they could monitor their own progress using a pedometer, participants were less enthusiastic about community competitions.</p><p>Additionally, the participants suggested incentives that could be incorporated into a social marketing program. The incentives suggested varied with age and activity level. Regularly active participants preferred physical activity supports, such as athletic shoes, hats, or T-shirts; others (primarily irregularly active women) preferred monetary incentives or gift certificates.</p></sec><sec><title>
Phase 2: product refinement and promotional strategies
</title><p>Based on Phase 1 focus groups and survey data, the community advisory council decided that the target audience would be irregularly active women aged 35 to 54. These women were attempting to be active but reported more barriers than men (e.g., responsibilities, lack of time). These women also found walking appealing and could influence their families and peers to walk. Because Phase 2 focus groups included the target audience, the discussion centered on refining the product for promotion.</p><p>When participants were asked to distinguish between <italic>physical activity </italic>and<italic> exercise </italic>and to define <italic>moderate physical activity, </italic>they described exercise as "structured activity" and physical activity as "unstructured activity," even incidental. Exercise, unlike physical activity, was described as "intentional," "purposeful," and "deliberate." Participants overwhelmingly preferred the term exercise to physical activity. Women held various perceptions of the term moderate activity. However, they all concurred that moderate physical activity requires exertion to be beneficial. </p><p>Although most of the women in these groups had never before used a pedometer, they were willing to try one so that they could better monitor their activity. They found it especially appealing that pedometers could be used to set daily goals.</p><p>Participants from several groups stated that images of people exercising, particularly women, would motivate them to be active and expressed interest in the use of spokespersons. They strongly felt that the spokespersons should be women of various ages, races, and body shapes; they preferred spokeswomen with whom they could identify. Celebrities and well-known local women were mentioned as women who could motivate others to exercise. Local women recommended were women who were successful at losing weight or older women who were perceived to be active for their age.</p><p>As expected, the range of preferred radio and television stations was wide. The women admitted that they were most likely to listen to the radio when in the car, especially when traveling to and from work. When asked about how they would like to receive consistent feedback while participating in an intervention, almost all women felt that telephone calls were intrusive. Most preferred either e-mail or direct mail. Some women asked to receive tips that would encourage them to exercise.</p></sec></sec><sec><title>Discussion</title><p>The focus group findings were discussed among the community advisory council, the research team, and two professional advertising and publications consultants. The focus groups enriched our understanding of consumers' perceptions of walking and trail-use benefits and costs. Although walking was a favorite moderate-intensity exercise because it was considered to be an excellent way to improve physical and mental health, neither men nor women believed they had time to walk on a regular basis. Participants reported several priorities that dissuaded them from being active. Furthermore, even though Sumter County has several places where people can walk, some places, such as local trails, were misperceived as unsafe areas. Another misperception was that the local shopping-mall walking program was for older adults only.</p><sec><title>Use of focus group results to design a social marketing campaign</title><sec><title>Product and product positioning</title><p>The focus group research enabled us to identify walking as the product to promote. Walking was perceived as a healthy activity for any age. However, the best way to position walking so it is competitive with other activities is to promote it as a way to spend time with friends and relatives and improve one's appearance while enjoying its well-known health benefits. The subsequent social marketing program focused on walking in an effort to provide an attractive and sustainable physical activity for women. </p></sec><sec><title>Price</title><p>Although several costs were identified, lack of time was the greatest barrier for most participants. With further probing during focus group discussions, ideas for overcoming time limitations were elicited. For example, one could schedule walking time during the day as one would schedule a meeting at work. In an effort to make the time spent walking more acceptable, the social marketing program that was developed promoted walking as an opportunity to spend time with family and friends. </p></sec><sec><title>Place</title><p>Most focus group participants clearly enjoyed walking in their neighborhoods, but some walking resources were unfamiliar to them. For example, the focus groups revealed the need to educate the community about the shopping mall walking program and local trail locations. In response to this need, the social marketing program highlighted a description of the shopping mall walking program and local trails in an exercise resource guide that lists places in the county to walk and be active. All of the women participating in the program received this guide.</p><p>An unanticipated finding was the overwhelming perception that local trails are potential sites for criminal activity. This finding was unexpected because according to the study's random-digit–dialed telephone survey, most people perceived trails as safe. Nevertheless, the program attempted to address safety concerns noted by the focus groups by organizing group walks on the trails. These group walks offered security to the women and provided an opportunity to meet potential walking partners.</p></sec><sec><title>Promotion</title><p>The focus group discussion centered on walking groups as a form of social support and security. Many participants, particularly women, favored self-paced programs with the use of a pedometer. Incentives were also identified; as a result of preferences expressed in the focus groups, the social marketing program incorporated a variety of incentives, including T-shirts, water bottles, and gift certificates. </p><p>The Phase 2 focus groups, which consisted of irregularly active women aged 35 to 54, emphasized that marketing materials should incorporate images of women of diverse ages, body shapes, and races. It was clear that images of thin women were discouraging for this audience. In response to these findings, the resulting social marketing program invited local women of various ages, body shapes, and races to serve as program spokeswomen.</p><p>For this audience, marketing materials should perhaps include pedometers to motivate women to walk because the majority of women liked the idea of being able to measure their progress. All of the women who enrolled in the study received a pedometer and log to chart their progress in addition to receiving other educational materials.</p><p>Because most women participating in Phase 2 focus groups said they were more likely to listen to the radio while driving to and from work, the social marketing program used radio in addition to print media and television to promote walking and trail use and to recruit women for the exercise intervention. The spots were timed to coincide with the morning and afternoon commute.</p><p>Because telephone calls were seen as intrusive, participants received weekly tips on topics such as managing competing demands, goal setting, building social support, self-monitoring, and self-reward through their choice of mail or e-mail. In addition to those messages, the women received print materials, incentive prizes, and invitations to monthly exercise events with other women.</p></sec></sec><sec><title>
Study limitations
</title><p>There are caveats to this study. First, focus groups are not representative of an entire population, and participants' responses may be influenced by the responses of others. As a result, an issue such as crime on trails may be artificially inflated. Second, a focus group is not a tool for testing hypotheses; rather, it should be used to identify issues and themes. For these reasons, the study's focus groups were conducted in conjunction with quantitative data collection methods.</p><p>The essential components of social marketing — product, price, place, and promotion — must be considered and fully addressed for a social marketing program to be successful; this can only be done through consumer research. Because qualitative data are essential components of comprehensive consumer research, focus groups, as a qualitative research tool, can play a valuable role in enhancing consumer research. In this study, focus groups proved to be the key to understanding the target audience. As a result, a social marketing program promoting walking and other moderate-intensity physical activities was developed in Sumter County to meet the needs of irregularly active women.</p></sec></sec> |
The Prevention Research Centers Healthy Aging Research Network | Could not extract abstract | <contrib contrib-type="author" corresp="yes"><name><surname>Lang</surname><given-names>Jason E</given-names></name><degrees>MS, MPH</degrees><aff>Health Care and Aging Studies Branch, Division of Adult and Community Health, National Center for Chronic Disease Prevention and Health Promotion, Centers for Disease Control and Prevention</aff><aff>Mr Lang is a member of the Healthy Aging Research Network Writing Group</aff><address><email>jlang@cdc.gov</email><addr-line>4770 Buford Hwy NE, Mail Stop K-51, Atlanta, GA 30341</addr-line><phone>770-488-5597</phone></address></contrib><contrib contrib-type="author"><collab>The Healthy Aging Research Network Writing Group</collab></contrib><contrib contrib-type="author"><name><surname>Anderson</surname><given-names>Lynda</given-names></name><degrees>PhD</degrees><aff>Division of Adult and Community Health, Centers for Disease Control and Prevention, Atlanta, Ga</aff></contrib><contrib contrib-type="author"><name><surname>James</surname><given-names>LoGerfo</given-names></name><degrees>MD, MPH</degrees><aff>Health Promotion Research Center, University of Washington, Seattle, Wash</aff></contrib><contrib contrib-type="author"><name><surname>Sharkey</surname><given-names>Joseph</given-names></name><degrees>PhD, MPH, RD</degrees><aff>Texas Healthy Aging Research Network Center, Texas A&M University, College Station, Tex</aff></contrib><contrib contrib-type="author"><name><surname>Belansky</surname><given-names>Elaine</given-names></name><degrees>PhD</degrees><aff>Rocky Mountain Prevention Research Center, University of Colorado, Denver, Colo</aff></contrib><contrib contrib-type="author"><name><surname>Bryant</surname><given-names>Lucinda</given-names></name><degrees>PhD, MSHA</degrees><aff>Rocky Mountain Prevention Research Center, University of Colorado, Denver, Colo</aff></contrib><contrib contrib-type="author"><name><surname>Prohaska</surname><given-names>Tom</given-names></name><degrees>PhD</degrees><aff>Center for Research on Health and Aging, University of Illinois at Chicago, Chicago, Ill</aff></contrib><contrib contrib-type="author"><name><surname>Altpeter</surname><given-names>Mary</given-names></name><degrees>PhD</degrees><aff>Institute on Aging, University of North Carolina at Chapel Hill, Chapel Hill, NC</aff></contrib><contrib contrib-type="author"><name><surname>Marshall</surname><given-names>Victor</given-names></name><degrees>PhD</degrees><aff>Institute on Aging, University of North Carolina at Chapel Hill, Chapel Hill, NC</aff></contrib><contrib contrib-type="author"><name><surname>Satariano</surname><given-names>William</given-names></name><degrees>PhD, MPH</degrees><aff>University of California, Berkeley, Berkeley, Calif</aff></contrib><contrib contrib-type="author"><name><surname>Ivey</surname><given-names>Susan</given-names></name><degrees>MD, MSHA</degrees><aff>University of California, Berkeley, Berkeley, Calif</aff></contrib><contrib contrib-type="author"><name><surname>Bayles</surname><given-names>Constance</given-names></name><degrees>PhD</degrees><aff>Center for Healthy Aging, University of Pittsburgh, Pittsburgh, Pa</aff></contrib><contrib contrib-type="author"><name><surname>Pluto</surname><given-names>Delores</given-names></name><degrees>PhD</degrees><aff>Rocky Mountain Prevention Research Center, University of Colorado, Denver, Colo</aff></contrib><contrib contrib-type="author"><name><surname>Wilcox</surname><given-names>Sara</given-names></name><degrees>PhD</degrees><aff>Prevention Research Center, University of South Carolina, Columbia, SC</aff></contrib><contrib contrib-type="author"><name><surname>Goins</surname><given-names>R. Turner</given-names></name><degrees>PhD</degrees><aff>Robert C. Byrd Health Sciences Center School of Medicine and Center on Aging, West Virginia University, Morgantown, WVa</aff></contrib> | Preventing Chronic Disease | <sec><title>Background</title><p>In 2001, the Centers for Disease Control and Prevention's (CDC's) Healthy Aging program issued a request for proposals to create a Prevention Research Centers Healthy Aging Research Network (PRC–HAN). The plan for this network was guided by the principles underlying academic–community partnerships and coalitions (<xref rid="B1" ref-type="bibr">1</xref>,<xref rid="B2" ref-type="bibr">2</xref>) and had several goals. First, the network would foster a strong sense of local program ownership by drawing on the Prevention Research Centers' (PRCs') established community relationships to collaboratively develop a healthy aging research agenda. Second, facilitating partnerships among PRCs would strengthen the focus, cohesiveness, and collaboration of research on aging. Third, by combining the diverse perspectives of network members and establishing new national linkages, the network would have the potential to develop and promote interventions at individual, organizational, environmental, and policy levels. Fourth and finally, the network would link the CDC's Healthy Aging program to additional prevention researchers outside of the CDC. Here we describe the evolution of the PRC–HAN's structure, mission, and research agenda. We then provide several examples of network activities that are helping the CDC and its partners to meet the challenges of an aging U.S. population. We conclude with planned evaluation activities and next steps.</p></sec><sec><title>Context</title><sec><title>Changing demographics in the United States</title><p>Both the number and the percentage of older adults (aged 65 or older) in this country are increasing at a much faster rate than ever before. This population will continue to grow, doubling from 35 million in 2005 to 70 million by 2030, when one in every five Americans will be aged 65 years or older (<xref rid="B3" ref-type="bibr">3</xref>). The looming health and economic impact of an aging society compels the CDC and the public health community to increase emphasis on preventing unnecessary disease, disability, and injury among older Americans. An enhanced focus on prevention is critical to preserving older individuals' independence and reducing long-term care needs. This enhanced focus on prevention is also critical for helping to stem escalating health care costs. Fortunately, our increased understanding of the role of risk factors such as physical inactivity in the onset and progression of chronic illness and disability has made this goal achievable.</p></sec><sec><title>The CDC's Prevention Research Centers (PRC) Program</title><p>The PRC Program, which includes 33 academic centers in 26 states, was established in 1984. (More information on the program is available from <ext-link xlink:href="www.cdc.gov/prc" ext-link-type="uri">www.cdc.gov/prc</ext-link>)The PRC Program, which includes 33 academic centers in 26 states, was established in 1984. (More information on the program is available from www.cdc.gov/prc). The centers are selected through a competitive peer-review process. All funded PRCs share a set of common principles, including 1) a focus on community-based participatory research that contributes to the knowledge base and informs public health programs and policies; 2) the creation of partnerships among communities, health departments, and other groups; 3) the timely collection, synthesis, and dissemination of the results of research and programs; 4) education and training in prevention research for the community, policymakers, health advocacy groups, and others; and 5) strong collaborative ties to the CDC PRC Program office for ongoing network enhancement and evaluation (<xref rid="B4" ref-type="bibr">4</xref>). The research focus of each PRC reflects the scientific expertise of its faculty and the goals of its community partners. Each PRC establishes a committee of community leaders and organizations to build a long-term relationship for engaging communities as partners in research and dissemination efforts. In addition to conducting a core research project within the area of research focus, the centers work with partners and are eligible to apply for Special Interest Projects (SIPs) Competitive Supplements, which are defined by the CDC and other federal agencies.</p></sec></sec><sec><title>Methods</title><sec><title>Creation of the PRC–HAN</title><p>Through the PRC SIPs competitive process in 2001, the University of Washington; the University of California, Berkeley; the University of Colorado; the University of Illinois at Chicago; the University of North Carolina at Chapel Hill; and the University of South Carolina became the first members of the PRC–HAN. The University of Pittsburgh was added to the network in 2002, and Texas A&M University and West Virginia University joined in 2004. All of the centers funded in 2001 competed and were refunded in 2004. For more information on the PRC–HAN, see <ext-link xlink:href="http://depts.washington.edu/harn/" ext-link-type="uri">http://depts.washington.edu/harn/</ext-link>.</p></sec><sec><title>Structure and communications</title><p>The participants from each university bring their unique perspectives and those of their communities to the network. Two of the network's PRCs have healthy aging as a focus of their core programs. The other network members represent various aging centers and programs from the broader universities that house their core PRC programs. Each member of the PRC–HAN has an equal voice in a consensus-based decision-making process that includes the agreement of all funded centers and key partners, such as the CDC, when selecting research topics and identifying the number of network centers that will participate in a given activity. The network structure allows additional research expertise and community capacity to be brought in from outside the network membership to help investigate new topics. The University of Washington Health Promotion Research Center (UWHPRC) is designated as the coordinating center. UWHPRC receives additional funds to support coordinating activities, which include initiating regular network communications, acting as a clearinghouse for information among network members, and serving as the PRC–HAN's primary representative to outside organizations and groups.</p><p>The CDC's Healthy Aging program funds the infrastructure of the PRC–HAN. During the PRC–HAN's first year, its members collaborated with the CDC to formulate a common mission, establish a definition of healthy aging, and develop a framework and agenda for research in public health and aging. The CDC SIPs and external partners continue to support the work of the network.</p></sec><sec><title>Mission</title><p>The mission of the PRC–HAN is to better understand the determinants of healthy aging in older adult populations, to identify interventions that promote healthy aging, and to help translate this knowledge into sustainable community-based programs throughout the nation. The PRC–HAN defines healthy aging as the development and maintenance of optimal physical, mental, and social well-being and function in older adults. It is most easily achieved when physical environments and communities are safe and support the adoption and maintenance of attitudes and behaviors known to promote health and well-being and when health services and community programs are used effectively to prevent or minimize the impact of acute and chronic disease on function. This definition reinforces a social–ecological view of the determinants of health (<xref rid="B5" ref-type="bibr">5</xref>). The PRC–HAN adapted and applied a social–ecological framework first developed by McLeroy et al and later refined by Sallis et al (<xref rid="B5" ref-type="bibr">5</xref>,<xref rid="B6" ref-type="bibr">6</xref>). This model emphasizes the interactions among individual, group, and community levels while also considering the environmental and policy arenas. The model is available from <ext-link xlink:href="http://depts.washington.edu/harn/research%20agenda/research%20agenda.shtml" ext-link-type="uri">http://depts.washington.edu/harn/research%20agenda/research%20agenda.shtml</ext-link>.</p><p>The research agenda developed by the PRC–HAN is intentionally broad. The agenda is built upon the intersections of three areas: disease topics, health-promoting skills, and knowledge domains (<xref rid="T1" ref-type="table">Table</xref>). The diseases and conditions that the agenda addresses are leading causes of illness and death among older adults. The second intersecting area of the agenda consists of six key health-promoting skills and behaviors. Although many of the skills and individual behaviors are analogous to the concept of underlying causes of illness and death, the agenda has expanded the usual list to include skill building for self-management and involvement in social activity (<xref rid="B7" ref-type="bibr">7</xref>). Finally, the agenda points to five primary areas of knowledge required to translate topics from research to policy arenas. These five domains allow the PRC–HAN to determine priority areas for the network by identifying risk factors, their impact on health and well-being, and behavioral, social, and environmental strategies to reduce them. </p></sec><sec><title>Objectives</title><p>Grounded in the philosophy of participatory research and collaboration, the network strives to 1) expand prevention research for older adults and their communities; 2) promote the use of research findings by translating and disseminating them to stakeholders; 3) strengthen PRC–HAN capacity through partnerships and expanded funding; and 4) stimulate the adoption of policies and programs by engaging policymakers, planners, and practitioners in the development process. Consistent with the national PRC Program, each network member works with its own community committee on PRC–HAN activities.</p></sec><sec><title>Research activities</title><p>Through a consensus process, the network chose physical activity for older adults as its initial focus for research. As an example of the ecological framework within which the PRC–HAN works, one research question of particular interest was how barriers within the built environment limit the amount of physical activity that older adults engage in.</p><p>Two PRC–HAN projects have involved all network members. The first of these, led by the University of Illinois at Chicago and modeled after the <italic>Get in Shape Chicago</italic> guide (available from <ext-link xlink:href="http://shapechicago.com" ext-link-type="uri">http://shapechicago.com</ext-link>), was a comprehensive, multisite survey that collected information on the capacity, content, and accessibility of physical activity programs for older adults. The workgroup developed a common survey instrument, refined the survey and analysis methods, tracked field activities, and coordinated data collection and analysis in the PRC–HAN community sites. The PRC–HAN members used the results of this survey to develop local resource guides for their communities. The project will track community distribution and use of the guides, which will soon be available online. The survey results are also being used to develop a publicly available interactive Web guide that can be applied in any community. An example of a draft guide is available from <ext-link xlink:href="www.shapeupkingcounty.org" ext-link-type="uri">www.shapeupkingcounty.org</ext-link>.</p><p>The second networkwide project was the development of a peer-reviewed publication, which is currently under review. This publication describes the role of public health in promoting physical activity among older adults. The in-depth, evidence-based review has become a template for how the PRC–HAN will prepare for new topic areas. Further, the messages emerging from this effort will form the basis for training and educational materials that will be shared with constituents in the public health and aging services network.</p><p>The PRC–HAN conducts many of its activities through workgroups. The environmental factors workgroup is addressing issues associated with the physical environment and its relationship to physical activity among older adults. The University of California, Berkeley serves as lead center for this workgroup. Reaching beyond the expertise of the current PRC–HAN members, the workgroup collaborated with researchers at St Louis University PRC to develop and refine an instrument for assessing environmental factors that affect the walkability of neighborhoods for older adults. Using a revised instrument based on their previous work, St Louis University researchers trained PRC–HAN members on how to conduct audits. These PRC–HAN members then conducted pilot environmental audits in each of their communities. These data are currently being analyzed and will be used to refine the final instrument, along with qualitative data from interviews conducted with older adults to determine their reasons for walking or not walking. This workgroup has developed a directory of Web sites related to environmental assessment and a manuscript assessing the current knowledge base and next steps for research. Finally, members of the network have received funding from The Robert Wood Johnson Foundation Active Living by Design program to assess the environmental correlates of physical activity among older adults. The project will include environmental assessments and interviews with older adults in five areas covered by PRC–HAN member sites.</p><p>Additional workgroups have been formed. A CDC-supported depression workgroup, led by the UWHPRC, is conducting an evidence-based review of community interventions to prevent or alleviate depression. The nutrition interest group, under the leadership of Texas A&M University, is looking at the neighborhood food environment by examining the accessibility, affordability, and availability of healthy foods in communities where older adults live.</p></sec><sec><title>Translation and dissemination</title><p>Led by the University of Colorado and the University of North Carolina at Chapel Hill, the research dissemination and practice workgroup has been the PRC–HAN's direct link to providing training to build state-based programs for healthy aging. The workgroup's goal is to help create local, state, and national community partnerships to build capacity and infrastructure in the public health and aging services networks and to provide more evidence-based health promotion opportunities for older adults. Key partners in this effort include the CDC, the Chronic Disease Directors, the Administration on Aging (AoA), the National Council on the Aging (NCOA), the National Association of State Units on Aging (NASUA), and the Agency for Healthcare Research and Quality (AHRQ). The workgroup plays a central role in developing materials to increase understanding of evidence-based interventions and disseminating these interventions and best practices.</p></sec><sec><title>Capacity</title><p>To strengthen the network's capacity, the PRC–HAN is working to maximize relationships with partners and obtain new sources of funds for PRC–HAN priorities. Developing new partnerships has made it possible to enhance rather than duplicate efforts. For example, PRC–HAN members are supporting NCOA's development of a National Resource Center (<ext-link xlink:href="www.healthyagingprograms.org" ext-link-type="uri">www.healthyagingprograms.org</ext-link>) for AoA's evidence-based grants by developing and disseminating technical assistance topic papers and manuals for the AoA grantees. The PRC–HAN also collaborated with Medstat, the Centers for Medicare & Medicaid Services, NASUA, and NCOA to plan a demonstration of the Senior Risk Reduction Program, which will evaluate health risk appraisal systems in target communities nationwide. The PRC–HAN's contribution to planning the study was to develop a sampling frame that allows researchers to assess the usefulness of community support systems for behavior change. For example, ensuring that a known number of communities with adequate capacity (e.g., physical activity programs or information and assistance capability) are included in the sample would facilitate assessment. The network is also strengthening capacity by pooling its resources with those of its partners to generate new funding for PRC–HAN priorities. For example, the PRC–HAN successfully obtained funding from The Robert Wood Johnson Foundation for a <italic>best practices</italic> project in partnership with NCOA. Finally, fostering the development of workgroups enhances PRC–HAN functioning, strengthens the capacity of the individual members, and promotes synergy among network members. The depression workgroup, for example, has benefited from the contributions of researchers from both within and outside the PRC–HAN.</p></sec><sec><title>Conference planning and communications</title><p>The PRC–HAN will stimulate the adoption of policies and programs by engaging policymakers, planners, and practitioners. The conference support workgroup, lead by UWHPRC, is designing a series of conferences to disseminate the lessons learned from our current work in physical activity and future work on depression and nutrition. The planning committee worked with two HAN-affiliated groups, the PRC–HAN community committees and representatives from the PRC–HAN research dissemination and practice workgroup, to ascertain their interests and needs for the design and content of the conferences. Presenters for the conferences include nationally known scholars in the field of public health, aging, and evidence-based programs in the areas of physical activity, depression, and nutrition. Such efforts are aimed at developing teams of public health and aging services practitioners, community members, and policymakers from states.</p></sec></sec><sec><title>Consequences</title><p>An evaluation of the PRC-HAN is important from several perspectives, including those of the CDC's Healthy Aging and PRC programs, PRC–HAN members, PRC–HAN community committees, other partners, and funders. In 2003, the PRC–HAN initiated an internal evaluation to better define the network's contributions to healthy aging, formalize internal processes, and better equip itself to serve as a model for other PRC thematic networks. The process began with internal interviews that led to a retreat to start developing a PRC–HAN logic model. The PRC–HAN is refining its logic model, developing performance indicators, and conducting a pilot evaluation to ensure that the PRC–HAN evaluation can be integrated into the PRC national evaluation.</p></sec><sec><title>Interpretation</title><p>By making partnership, shared responsibility, and broad expertise key components of its capacity, the PRC–HAN has earned the recognition of other organizations as an important partner in advancing science-based health promotion activities for older adults. Its work in the area of physical activity for older adults brought additional partners to the network, which in turn has increased the diversity of the group's expertise and provided new outlets for its products. Such expansion also enables the network to increase its resources and engage in new project areas. Established trust and mutual respect among participants, strong ties to local communities, and the ability to combine its expertise in healthy aging with that of partners in national, state, and local organizations have positioned the PRC–HAN to continue as a productive and effective research network for years to come.</p></sec> |
The Impact of Inequality: How to Make Sick Societies Healthier | Could not extract abstract | <contrib contrib-type="author"><name><surname>Szaflarski</surname><given-names>Magdalena</given-names></name><degrees>PhD</degrees><aff>Institute for the Study of Health, Department of Family Medicine, University of Cincinnati, Cincinnati, Ohio</aff></contrib> | Preventing Chronic Disease | <fig position="anchor" id="F1" fig-type="diagram"><alt-text>Cover of The Impact of Inequality: How to Make Sick Societies Healthier</alt-text><graphic xlink:href="PCD31A26s01" position="float"/></fig><p>
<italic>The Impact of Inequality: How to Make Sick Societies Healthier</italic>, by Richard G. Wilkinson, is a sequel to his book, <italic>Unhealthy Societies: The Afflictions of Inequality</italic> (<xref rid="B1" ref-type="bibr">1</xref>). The premise of both books is that the structure of social relations determines the health status of populations; that is, the greater the social inequality (e.g., income disparities) within a society, the poorer the health outcomes (e.g., the higher the death rates). Wilkinson's first book focused on documenting the relationship based on an emerging body of evidence; his second book takes the next step by reviewing the current state of knowledge, offering an explanation, and suggesting potential solutions to the problem. With its carefully assembled and weighted evidence and its clear and convincing argument, the second book is an excellent resource for social scientists, epidemiologists, public health officials, policy makers, and students.</p><p>This nine-chapter book addresses the primary questions of <italic>how</italic> and <italic>why</italic> inequality negatively affects individuals and populations. Chapters 2 through 4 describe the patterns of association between inequality and health and social outcomes, underscoring that "differences in inequality as small as those found between different market democracies or different U.S. states produce very substantial social and health effects." Chapter 2 reviews evidence supporting the idea that the quality of social relations is associated with income disparities, and Chapter 3 outlines psychosocial factors that contribute to ill health and premature death, including low social status, poor social affiliations, and negative childhood experiences, all of which can be linked to inequality. The data presented in Chapter 4 build upon previous chapters and strongly indicate that the more unequal a society is, the worse its health: "The pathway runs from inequality, through its effects on social relations and the problems of low social status and family functioning, to its impact on stress and health."</p><p>The remaining chapters explain the causal processes responsible for these relationships. In Chapter 5, the author uses violence as an example of a strong correlate of inequality and discusses the contributions of low social status and self-worth. Chapter 6 moves the theme forward and describes the social processes responsible for social distances and distinctions, including discrimination. In Chapter 7, race and gender inequality are examined, revealing among other things an interesting paradox: men appear to be more harmed by male domination than women are. Chapter 8 examines the pathway from the form of social organization (degree of inequality) through stress and coping mechanisms to physiological factors (e.g., cardiovascular, immune) that shape health status. Finally, in Chapter 9, the author frames the problem of inequality and health in terms of ideology and political objectives, revisiting the traditional democratic values: liberty, equality, and fraternity.</p><p>A systematic review of the existing research is one of the book's most important strengths. Wilkinson carefully reviews studies to date citing supporting evidence as well as negative findings. When comparing studies, he points out methodological variations and offers explanations for differences in findings. He chooses his examples with care, focusing on the most robust and convincing research; for example, he examines homicide as one type of violence and death rates as indicators of population health, definitions of which are similar across societies, allowing straightforward comparisons. The book also highlights current and high-priority social issues such as obesity and their links to social class throughout history.</p><p>In addition to the content, Wilkinson's book is well-organized and written in language that can reach a more general audience. What is somewhat unconventional is that some background information (e.g., evolution of human societies and social inequality) appears later in the book instead of at the beginning. However, this strategy seems to work well for this book — by presenting his argument and much of the supporting evidence in the first half of the book, Wilkinson succeeds in getting his point across more powerfully. There is one weakness to this scheme, however: Chapters 5 and 6 contain information and examples already introduced in previous sections, making these chapters somewhat redundant and less effective.</p><p>The success of the book is based in part on the author's obvious passion for the topic and his sincere concern about the social issue. To make his point, Wilkinson sometimes offers an extreme opinion or example, such as "perhaps we should liken the injustice of health inequalities to that of a government that executed a significant portion of its population each year without cause." Wilkinson's speculations provoke thought. He wonders, for example, how different the government's response to health disparities would be if the income gradient in health were opposite to the existing one: that is, if it were the higher income groups experiencing the worst health. Although written from a particular viewpoint — that of social justice and reform — the book does not attack capitalism per se, and it does not impose an extreme ideology. Instead, the author's view of social progress exists within the framework of modern society, where the market seems to be an inevitable element. Within that framework, inequality can be reduced, quality of social relations improved, and social stress decreased, all leading to improvements in health and well-being.</p><p>Wilkinson ends the book on an optimistic note: although differences in health inequalities across societies and history exist, change is possible, and inequalities can be reduced. As our moral universe expands along the line of democratic values and as we become more sensitive to the suffering and pain of others, the reduction of inequality and the improvement of well-being across the social strata should strengthen as political goals. This book is highly recommended to anyone interested in health and its related social issues.</p> |
Evaluating Progress Toward Healthy People 2010 National Diabetes Objectives | Could not extract abstract | <contrib contrib-type="author" corresp="yes"><name><surname>Mukhtar</surname><given-names>Qaiser</given-names></name><degrees>PhD</degrees><role>Epidemiologist</role><aff>Applied Behavioral Research, Epidemiology, Surveillance, and Program Evaluation Team, Program Development Branch, Division of Diabetes Translation, Centers for Disease Control and Prevention</aff><address><email>qam2@cdc.gov</email><addr-line>4770 Buford Hwy NE, Mail Stop K-10, Atlanta, GA 30341</addr-line><phone>770-488-5505</phone></address></contrib><contrib contrib-type="author" corresp="no"><name><surname>Jack</surname><given-names>Leonard</given-names></name><degrees>PhD, MSc</degrees><role>Team Lead</role><aff>Applied Behavioral Research, Epidemiology, Surveillance, and Program Evaluation, Program Development Branch, Division of Diabetes Translation, Centers for Disease Control and Prevention, Atlanta, Ga</aff></contrib><contrib contrib-type="author" corresp="no"><name><surname>Martin</surname><given-names>Maurice</given-names></name><degrees>PhD</degrees><role>Research and Evaluation Scientist</role><aff>Applied Behavioral Research, Epidemiology, Surveillance, and Program Evaluation, Program Development Branch, Division of Diabetes Translation, Centers for Disease Control and Prevention, Atlanta, Ga</aff></contrib><contrib contrib-type="author" corresp="no"><name><surname>Rivera</surname><given-names>Mark</given-names></name><degrees>PhD</degrees><role>Research and Evaluation Scientist</role><aff>Applied Behavioral Research, Epidemiology, Surveillance, and Program Evaluation, Program Development Branch, Division of Diabetes Translation, Centers for Disease Control and Prevention, Atlanta, Ga</aff></contrib><contrib contrib-type="author" corresp="no"><name><surname>Murphy</surname><given-names>Dara</given-names></name><degrees>MPH</degrees><role>Branch Chief</role><aff>Division of Diabetes Translation, Centers for Disease Control and Prevention, Atlanta, Ga</aff></contrib> | Preventing Chronic Disease | <sec><title>Introduction</title><p>Diabetes and its associated burden are continuously evolving (<xref ref-type="bibr" rid="B1">1</xref>). More than 18 million adults in the United States have diabetes (<xref ref-type="bibr" rid="B2">2</xref>). People with diabetes are at increased risk of developing complications such as blindness, lower extremity amputations, end-stage renal disease, and cardiovascular disease (<xref ref-type="bibr" rid="B2">2</xref>). Although recommended preventive care services can prevent or delay the development of these serious health complications (<xref ref-type="bibr" rid="B3">3</xref>-<xref ref-type="bibr" rid="B6">6</xref>), the proportion of adults with diabetes who report receiving preventive care services is less than what we hope for (<xref ref-type="bibr" rid="B7">7</xref>-<xref ref-type="bibr" rid="B8">8</xref>).</p><p>Increasing the rates of preventive care services among people with diabetes is an objective of <italic>Healthy People 2010</italic> (<italic>HP2010</italic>), a program established by the U.S. Department of Health and Human Services (<xref ref-type="bibr" rid="B9">9</xref>). In 1999, the National Diabetes Prevention and Control Program (NDPCP), established by the Centers for Disease Control and Prevention (CDC), adopted three of the 17 <italic>HP2010</italic> diabetes objectives and two of the <italic>HP2010</italic> immunization objectives. The five objectives adopted include percentage targets for annual hemoglobin A1c (HbA1c) tests, annual dilated eye examinations, annual foot examinations, and influenza and pneumococcal vaccinations (<xref rid="T1" ref-type="table">Table 1</xref>). Hereafter, we refer to these objectives as <italic>national diabetes objectives</italic>. This report focuses on the following three national diabetes objectives and their percentage targets for the U.S. population with diabetes: 1) 50% receive an HbA1c measurement at least twice each year, 2) 75% receive an annual dilated eye examination, and 3) 75% receive an annual foot examination. Implementation strategies used to help achieve these objectives are described elsewhere (<xref ref-type="bibr" rid="B10">10</xref>).</p><p>The NDPCP supports a Diabetes Prevention and Control Program (DPCP) in each of the 50 U.S. states, 8 territories, and the District of Columbia (<xref ref-type="bibr" rid="B10">10</xref>), or <italic>jurisdictions</italic>. The 59 DPCPs strive to reduce the preventable burden of diabetes and its associated complications in their jurisdictions by building collaborations and working closely with their partners. These partners include managed care organizations, policy makers, and the American Diabetes Association and its local affiliates (<xref ref-type="bibr" rid="B10">10</xref>). When the NDPCP adopted the five national diabetes objectives, it asked the DPCPs to increase target percentage rates for each objective within their jurisdiction. The NDPCP itself did not set these targets because it recognized that DPCPs vary in capacity and funding.</p></sec><sec sec-type="methods"><title>Methods</title><p>To evaluate progress toward national diabetes objectives, we analyzed data on diabetes preventive care services from the Behavioral Risk Factor Surveillance System (BRFSS). The BRFSS is a state-based, randomized telephone survey of the noninstitutionalized U.S. population aged 18 years and older. BRFSS surveys are conducted in each of 50 U.S. states, the District of Columbia, and three U.S. territories (the U.S. Virgin Islands, Puerto Rico, and Guam). The optional diabetes module contains 12 diabetes-specific questions; these questions are asked of people who identify themselves as having diabetes (<xref ref-type="bibr" rid="B11">11</xref>). People with diabetes were defined as those who answered yes to the core BRFSS question, "Has a doctor ever told you that you have diabetes?" Women who were told that they had diabetes only during pregnancy were excluded from the analysis. Individuals who refused to answer or had missing or unknown values for any variable were also excluded from the analysis. In addition, data from Guam and the U.S. Virgin Islands were excluded from the analysis.</p><p>We analyzed responses to three preventive care practice questions from the diabetes module: 1) "When was the last time you had an eye examination in which the pupils were dilated?" 2) "About how many times in the last year has a health professional checked your feet for any sores or irritations?" and 3) "A test for hemoglobin 'A one C' measures the average level of blood sugar over the past 3 months. About how many times in the past 12 months has a doctor, nurse, or other health professional checked you for hemoglobin 'A one C'?" People with diabetes who reported receipt of at least one dilated eye examination, one foot examination, or two or more HbA1c tests in the past year were defined as having received these preventive care services.</p><p>Forty-four jurisdictions used the diabetes module in both 2000 and 2003 (<xref ref-type="bibr" rid="B10">10</xref>); 47 used the diabetes module in 2003. First, we calculated the percentage of people with diabetes who reported receipt of the three selected preventive care services for the 44 jurisdictions. Second, we compared the percentages for 47 jurisdictions that used the module in 2003 with <italic>HP2010</italic> percentage targets. All rates were age-adjusted to the 2000 U.S. standard population. T-tests were used to determine whether differences between 2000 and 2003 were statistically significant. No correction was made for multiple testing of the three outcomes. We used simple linear regression to assess the association between baseline rates in 2000 and the magnitude of change between 2000 and 2003 rates. Analyses were conducted using SAS 9.1 (SAS Institute Inc, Cary, NC) and SUDAAN 9.0 (RTI International, Research Triangle Park, NC).</p></sec><sec sec-type="results"><title>Results</title><p>From 2000 to 2003, the aggregate age-adjusted rate for foot examinations increased from 63.7% to 69.3% (<xref rid="T2" ref-type="table">Table 2</xref>); the rate for HbA1c tests increased from 68.3% to 69.5% (<xref rid="T3" ref-type="table">Table 3</xref>), and the rate for dilated eye examinations decreased from 67.7% to 65.2%(<xref rid="T4" ref-type="table">Table 4</xref>). Only the increase in the rate for foot examinations was statistically significant (<italic>P</italic> < .001).</p><p>Age-adjusted rates for foot examinations increased in 26 (59%) jurisdictions (<xref rid="T2" ref-type="table">Table 2</xref>), rates for HbA1c tests increased in 33 (75%) jurisdictions (<xref rid="T3" ref-type="table">Table 3</xref>), and rates for dilated eye examinations increased in 15 (34%) jurisdictions (<xref rid="T4" ref-type="table">Table 4</xref>). Compared with interquartile ranges for all three tests in 2000, interquartile ranges are smaller in 2003 (<xref rid="F1" ref-type="fig">Figure 1</xref>). However, overall ranges are large for both 2000 and 2003. In 2003, rates for foot examinations ranged from 47.0% to 82.4%, rates for HbA1c tests ranged from 53.6% to 85.5%, and rates for eye examinations ranged from 45.8% to 81.6% (<xref rid="F1" ref-type="fig">Figure 1</xref>). These ranges are similar to ranges for 2000. The ranges for 2000 were 42.1% to 85.1% for foot examinations, 40.5% to 80.5% for HbA1c tests, and 51.0% to 89.0% for eye examinations (<xref rid="F1" ref-type="fig">Figure 1</xref>).</p><boxed-text position="float"><fig position="float" id="F1" fig-type="diagram"><label>Figure 1</label><caption><p>Percentage rates for annual foot examination, two or more annual hemoglobin A1c (HbA1c) tests, and annual dilated eye examination for 47 jurisdictions, Behavioral Risk Factor Surveillance System, 2000 and 2003. Overall range, interquartile range, and median are indicated.</p></caption><alt-text>Graph</alt-text><alternatives><graphic xlink:href="PCD31A11s01"/><table frame="hsides" rules="groups"><thead><tr><th scope="col" align="left" valign="bottom" rowspan="1" colspan="1">Year</th><th scope="col" align="center" valign="bottom" rowspan="1" colspan="1">Test</th><th scope="col" align="center" valign="bottom" rowspan="1" colspan="1">Minimum %</th><th scope="col" align="center" valign="bottom" rowspan="1" colspan="1">Inter-quartile Range Minimum %</th><th scope="col" align="center" valign="bottom" rowspan="1" colspan="1">Median %</th><th scope="col" align="center" valign="bottom" rowspan="1" colspan="1">Inter-quartile Range Maximum %</th><th scope="col" align="center" valign="bottom" rowspan="1" colspan="1">Maximum %</th></tr></thead><tbody><tr><th scope="row" align="left" valign="top" rowspan="1" colspan="1">2000</th><td align="left" valign="top" rowspan="1" colspan="1">Foot</td><td align="right" valign="top" rowspan="1" colspan="1">42.1</td><td align="right" valign="top" rowspan="1" colspan="1">61.7</td><td align="right" valign="top" rowspan="1" colspan="1">68.6</td><td align="right" valign="top" rowspan="1" colspan="1">73.3</td><td align="right" valign="top" rowspan="1" colspan="1">85.1</td></tr><tr><th scope="row" align="left" valign="top" rowspan="1" colspan="1">2003</th><td align="left" valign="top" rowspan="1" colspan="1">Foot</td><td align="right" valign="top" rowspan="1" colspan="1">47.0</td><td align="right" valign="top" rowspan="1" colspan="1">67.0</td><td align="right" valign="top" rowspan="1" colspan="1">71.1</td><td align="right" valign="top" rowspan="1" colspan="1">74.1</td><td align="right" valign="top" rowspan="1" colspan="1">82.4</td></tr><tr><th scope="col" align="left" valign="bottom" rowspan="1" colspan="1">2000</th><th scope="col" align="left" valign="bottom" rowspan="1" colspan="1">HbA1c</th><th scope="col" align="right" valign="bottom" rowspan="1" colspan="1">40.5</th><th scope="col" align="right" valign="bottom" rowspan="1" colspan="1">60.8</th><th scope="col" align="right" valign="bottom" rowspan="1" colspan="1">65.4</th><th scope="col" align="right" valign="bottom" rowspan="1" colspan="1">71.7</th><th scope="col" align="right" valign="bottom" rowspan="1" colspan="1">80.5</th></tr><tr><th scope="row" align="left" valign="top" rowspan="1" colspan="1">2003</th><td align="left" valign="top" rowspan="1" colspan="1">HbA1c</td><td align="right" valign="top" rowspan="1" colspan="1">53.6</td><td align="right" valign="top" rowspan="1" colspan="1">67.5</td><td align="right" valign="top" rowspan="1" colspan="1">72.2</td><td align="right" valign="top" rowspan="1" colspan="1">76.1</td><td align="right" valign="top" rowspan="1" colspan="1">85.5</td></tr><tr><th scope="row" align="left" valign="top" rowspan="1" colspan="1">2000</th><td align="left" valign="top" rowspan="1" colspan="1">Eye</td><td align="right" valign="top" rowspan="1" colspan="1">51.0</td><td align="right" valign="top" rowspan="1" colspan="1">63.2</td><td align="right" valign="top" rowspan="1" colspan="1">68.7</td><td align="right" valign="top" rowspan="1" colspan="1">76.0</td><td align="right" valign="top" rowspan="1" colspan="1">89.0</td></tr><tr><th scope="row" align="left" valign="top" rowspan="1" colspan="1">2003</th><td align="left" valign="top" rowspan="1" colspan="1">Eye</td><td align="right" valign="top" rowspan="1" colspan="1">45.8</td><td align="right" valign="top" rowspan="1" colspan="1">62.4</td><td align="right" valign="top" rowspan="1" colspan="1">66.2</td><td align="right" valign="top" rowspan="1" colspan="1">72.7</td><td align="right" valign="top" rowspan="1" colspan="1">81.6</td></tr></tbody></table></alternatives></fig></boxed-text><p>Compared with <italic>HP2010</italic> targets, the 2003 data from 47 jurisdictions showed that 11 (23%) jurisdictions met the 75% target for foot examination (<xref rid="T2" ref-type="table">Table 2</xref>), nine (19%) met the 75% target for dilated eye examination (<xref rid="T4" ref-type="table">Table 4</xref>), and all jurisdictions met the 50% target of at least two HbA1c tests per year (<xref rid="T3" ref-type="table">Table 3</xref>).The following five states were at or above the <italic>HP2010</italic> targets for all three objectives: Delaware, Hawaii, Massachusetts, Minnesota, and New Hampshire (<xref rid="T2" ref-type="table">Tables 2</xref>-<xref rid="T4" ref-type="table">4</xref>).</p><p>An inverse association was found between the baseline rate in 2000 and the magnitude of change between 2000 and 2003. Jurisdictions that had relatively low baseline rates in 2000 were more likely to show an increase in 2003 or to maintain their baseline rates. Jurisdictions with relatively high baseline rates in 2000 were more likely to show a decrease in 2003. For every unit increase in the baseline percentage rate in 2000, the predicted 2003 rate decreased as follows: for foot examination by 0.66% (<xref rid="F2" ref-type="fig">Figure 2</xref>; F value = 52.4; <italic>P</italic> < .001); for HbA1c tests by 0.69% (<xref rid="F3" ref-type="fig">Figure 3</xref>; F value = 34.7; <italic>P</italic> < .001), and for eye examination by 0.46% (<xref rid="F4" ref-type="fig">Figure 4</xref>; F value = 9.7; <italic>P</italic> = .003).</p><boxed-text position="float"><fig id="F2" fig-type="diagram" position="float"><label>Figure 2</label><caption><p>Regression of change in 2003 rates for annual foot examinations on 2000 baseline rates for 44 jurisdictions, Behavioral Risk Factor Surveillance System data.</p></caption><alt-text>Graph that shows regression line for predicting 2003 rates of obtaining annual foot examination based on rates for 2000. For every unit increase in the baseline percentage rate in 2000, the predicted 2003 rate decreased by 0.66%.</alt-text><graphic xlink:href="PCD31A11s02"/></fig></boxed-text><boxed-text position="float"><fig id="F3" fig-type="diagram" position="float"><label>Figure 3</label><caption><p>Regression of change in 2003 rates for two or more annual hemoglobin A1c tests on 2000 baseline rates for 44 jurisdictions, Behavioral Risk Factor Surveillance System data.</p></caption><alt-text>Graph that shows regression line for predicting 2003 rates of obtaining two or more HbA1c tests based on rates for 2000. For every unit increase in the baseline percentage rate in 2000, the predicted 2003 rate decreased by 0.69%.</alt-text><graphic xlink:href="PCD31A11s03"/></fig></boxed-text><boxed-text position="float"><fig id="F4" fig-type="diagram" position="float"><label>Figure 4</label><caption><p>Regression of change in 2003 rates for annual dilated eye examinations on 2000 baseline rates for 44 jurisdictions, Behavioral Risk Factor Surveillance System data.</p></caption><alt-text>Graph that shows regression line for predicting 2003 rates of obtaining annual eye examination based on rates for 2000. For every unit increase in the baseline percentage rate in 2000, the predicted 2003 rate decreased by 0.46%.</alt-text><graphic xlink:href="PCD31A11s04"/></fig></boxed-text></sec><sec sec-type="discussion"><title>Discussion</title><p>Factors affecting the rate of receipt of preventive care services are numerous and include an individual's knowledge about the importance of obtaining the recommended tests and examinations, access to diabetes care, availability of health insurance, extent of health insurance coverage for diabetes care services, and the adherence of health care providers to recommended diabetes care guidelines (<xref ref-type="bibr" rid="B10">10</xref>).</p><p>Our findings show an inverse association between the baseline rates of the three selected preventive care services and the magnitude of change in rates from 2000 to 2003. This negative association may result from the statistical phenomenon known as <italic>regression toward the mean</italic>. In regression toward the mean, rates that are initially high or initially low eventually move toward the average because the initial performance occurred by chance (e.g., as a result of measurement error) or was caused by a factor unrelated to the intervention (<xref ref-type="bibr" rid="B12">12</xref>). A <italic>ceiling effect</italic> could also explain the negative association. With a ceiling effect, the initial impact of an intervention on rates may be large, but once rates reach a certain threshold, the intervention does not yield the same rate increases; rates may then reach a plateau or even decrease. The negative association between baseline rates and magnitude in change should be studied further.</p><p>The inverse association between baseline rates in 2000 and the predicted rates for 2003 confirms that it would not be appropriate for the NDPCP to evaluate the performance of a jurisdiction exclusively on the basis of rate increases or decreases. Moreover, 20 (42%) of 47 jurisdictions have already reached the <italic>HP2010</italic> target of 75% for foot or eye examinations, and an additional 27 (57%) jurisdictions have rates between 70% and 75%. These jurisdictions are likely to experience a ceiling effect. Additionally, other revisions to the <italic>HP2010</italic> targets are being considered. The proposed targets are as follows: 91% of people with diagnosed diabetes receive an annual foot examination, 76% of people with diagnosed diabetes receive an annual dilated eye examination, and 65% of people with diagnosed diabetes receive at least two HbA1c tests annually (<xref ref-type="bibr" rid="B13">13</xref>). In 2003, the highest rate for foot examinations was 82.5%; jurisdictions are likely to struggle to reach the 91% target.  </p><p>Regardless of whether new <italic>HP2010</italic> revisions are adopted, the NDPCP should consider setting multilevel objectives instead of expecting all jurisdictions to show continuous rate increases. For example, jurisdictions below or at certain rates could aim to increase rates while jurisdictions with high rates (e.g., above 75%) could strive to maintain these rates and adopt new objectives. The idea of setting multilevel objectives is supported by evidence of the disparity in rates for national diabetes objectives among jurisdictions (<xref rid="F1" ref-type="fig">Figure 1</xref>).</p><p>In addition to considering multilevel targets, the NDPCP should consider moving from the process-focused objective of increasing the percentage of people with diabetes who receive at least two HbA1c tests to an outcome-focused objective such as reducing the percentage of people with poor glycemic control assessed through measuring HbA1c values. Currently, all 47 jurisdictions analyzed in this study have achieved the <italic>HP2010</italic> objective of 50% for at least two HbA1c tests, and 32 (68%) jurisdictions  have achieved a rate of more than 70% (<xref rid="T3" ref-type="table">Table 3</xref>). Strong evidence supports a significant correlation between glycemic control and microvascular and cardiovascular disease risk and mortality rates (<xref ref-type="bibr" rid="B5">5</xref>). For example, in the U.K. Prospective Diabetes Study, each 1% reduction in HbA1c levels was associated with a 37% reduction in risk for diabetic peripheral neuropathy, diabetic retinopathy, and diabetic nephropathy, conditions responsible for lower-extremity amputation, blindness, and end-stage renal disease (<xref ref-type="bibr" rid="B5">5</xref>).</p><p>State-level data on HbA1c values are available through the Health Plan Employer Data and Information Set (HEDIS). Currently, few jurisdictions use this data source; possible barriers to use include lack of access to the data, cost, not having an epidemiologist on staff, and complexity of data analysis (<xref ref-type="bibr" rid="B14">14</xref>).</p><p>Findings from this study have important implications for the NDPCP and its partners, especially as the CDC moves to align its priorities toward health promotion and prevention of disease, injury, and disability outcomes through improving quality of health by stages of life (<xref ref-type="bibr" rid="B15">15</xref>). The NDPCP and its partners will have opportunities to revise and add new national diabetes objectives to reduce the preventable burden of diabetes. Jurisdictions that are close to meeting or have already met the <italic>HP2010</italic> objectives will be good candidates to pioneer new objectives. New national diabetes objectives could be selected from existing <italic>HP2010</italic> diabetes objectives, or entirely new objectives could be developed and adopted.</p><p>Program-related information should be linked with traditional data sources such as the BRFSS so that we can understand the role of environmental factors as we evaluate the progress of jurisdictions toward national diabetes objectives. For program evaluation, it is imperative to understand why some jurisdictions perform better than others. It is especially important to examine jurisdictions that had high baseline rates in 2000 and showed rate increases in 2003. A range of environmental factors are likely to influence jurisdiction performance. These factors include budget, resources, leadership, partnerships, public health infrastructure, policies, and specific interventions. We need to understand how the adoption and sustainability of clinical and community-based interventions are effective in increasing preventive care services. To evaluate the impact of the effort of a jurisdiction on national diabetes objectives, we must identify and possibly modify existing data sources that contain program-related information.</p><p>For several years, the NDPCP has been building capacity to capture program-related information in its own electronic management information system (MIS). The MIS is a Web-based system designed to facilitate the collection and sharing of program-related information among state-based diabetes prevention and control programs and the NDPCP. The NDPCP is refining the MIS to link it with traditional data sources such as the BRFSS. When it is completed, the MIS will enable NDPCP to measure the association between the effort of each jurisdiction and its impact on rates of preventive care services. Enhancing the MIS will allow the NDPCP and jurisdictions to gain a better understanding of the influence of environmental factors.</p><p>Our findings are subject to several limitations. The BRFSS collects data through telephone surveys that exclude institutionalized people (e.g., nursing home residents) and people without telephones. The exclusion of these populations, particularly people without telephones, could result in overestimation of people receiving the three preventive care services, assuming that such people are likely to be less educated, poor, and therefore less likely to receive preventive care (<xref ref-type="bibr" rid="B16">16</xref>,<xref ref-type="bibr" rid="B17">17</xref>). The BRFSS data are also self-reported and subject to recall bias.</p><p>Additionally, the BRFSS data reflect statewide changes; they do not reflect changes that are taking place in special populations. Additional data sources such as Medicare and Medicaid records are needed for a more robust analysis of progress in achieving national diabetes objectives for such populations. Finally, comparing single-year data (2000 with 2003) could be a possible limitation, but if we had pooled years (i.e., 2000 and 2001 compared with 2002 and 2003), we would not have had a sufficient time lag for comparison.</p></sec><sec sec-type="conclusions"><title>Conclusion</title><p>The progress of jurisdictions toward achieving national diabetes objectives cannot be evaluated based on simple increases or decreases in their rates of preventive care services. Baseline rates should be considered in evaluating their progress toward achieving national diabetes objectives. There is a need for developing and adopting new objectives, especially for the jurisdictions that have already met or are close to reaching <italic>HP2010</italic> objectives. Gaining better understanding of the environmental factors that affect each jurisdiction — such as budget, policy, and health care coverage — is critical for appreciating and evaluating progress. To evaluate the impact of the effort of a jurisdiction on rates of national diabetes objectives, we must identify and modify existing data sources (e.g., the MIS) that contain program-related information. Linking this system with traditional data sets such as the BRFSS could help identify and quantify the underlying factors vital to progress toward national diabetes objectives.</p></sec> |
Development of the Diabetes Indicators and Data Sources Internet Tool (DIDIT) | Could not extract abstract | <contrib contrib-type="author" corresp="yes"><name><surname>Mukhtar</surname><given-names>Qaiser</given-names></name><degrees>PhD</degrees><aff>Division of Diabetes Translation, Centers for Disease Control and Prevention</aff><address><email>qam2@cdc.gov</email><addr-line>4770 Buford Hwy, Mail Stop K-10, Atlanta, GA 30341</addr-line><phone>770-488-5505</phone></address></contrib><contrib contrib-type="author"><name><surname>Mehta</surname><given-names>Prachi</given-names></name><degrees>DrPH</degrees><aff>Northrop Grumman Mission Systems, Atlanta, Ga</aff></contrib><contrib contrib-type="author"><name><surname>Brody</surname><given-names>Erica R</given-names></name><degrees>MPH</degrees><aff>RTI International, Research Triangle Park, NC</aff></contrib><contrib contrib-type="author"><name><surname>Camponeschi</surname><given-names>Jenny</given-names></name><degrees>MS</degrees><aff>Wisconsin Department of Health and Family Services, Division of Public Health Diabetes Prevention and Control Program, Madison, Wis</aff></contrib><contrib contrib-type="author"><name><surname>Friedrichs</surname><given-names>Michael</given-names></name><degrees>MS</degrees><aff>Diabetes Prevention and Control Program, Utah Bureau of Health Promotion, Utah Department of Health, Salt Lake City, Utah</aff></contrib><contrib contrib-type="author"><name><surname>Ralls</surname><given-names>Brenda</given-names></name><degrees>PhD</degrees><aff>Diabetes Prevention and Control Program, Utah Bureau of Health Promotion, Utah Department of Health, Salt Lake City, Utah</aff></contrib><contrib contrib-type="author"><name><surname>Kemple</surname><given-names>Angela M</given-names></name><degrees>MS</degrees><aff>Oregon Department of Human Services, Diabetes Prevention and Control Program, Portland, Ore</aff></contrib> | Preventing Chronic Disease | <sec><title>Introduction</title><p>The Diabetes Indicators and Data Sources Internet Tool (DIDIT) is a Web-based resource designed to strengthen the capacity of the staff and partners of state-based Diabetes Prevention and Control Programs (DPCPs) and staff of the Centers for Disease Control and Prevention (CDC) to conduct diabetes surveillance and program evaluation. The tool contains detailed information on 38 diabetes indicators (measures of health or factors associated with health) and their associated data sources (e.g., Behavioral Risk Factor Surveillance System [BRFSS]). The content, design, and function of the DIDIT have been described elsewhere (<xref rid="B1" ref-type="bibr">1</xref>). This article describes the process of developing the DIDIT, beginning with the conceptual phase and proceeding through the content and systems phases (<xref rid="T1" ref-type="table">Table 1</xref>). In so doing, we provide an example for other agencies and organizations as well as other entities within the CDC that are interested in developing a similar tool.</p></sec><sec><title>Developing the Concept and Content</title><p>The DIDIT was developed in response to a request from the DPCPs for technical assistance in surveillance and program evaluation (<xref rid="B2" ref-type="bibr">2</xref>). In August 2001, a work group comprised of representatives from the CDC’s Division of Diabetes Translation (DDT) and eight DPCPs (representing Alabama, Minnesota, Montana, New Mexico, New York, Oregon, Utah, and Wisconsin) was convened to develop a tool that would provide comprehensive information about 1) diabetes indicators (e.g., foot examinations, diabetes-related hospitalizations) and 2) their associated data sources (e.g., BRFSS, Medicare) in one centralized place. The work group generated a list of 55 indicators (later reduced to 38) and developed an outline describing the desired contents and format of a tool called the <italic>diabetes indicator tool</italic>.</p><p>In October 2001, an overview and vision of the DIDIT was presented to six focus groups at the annual meeting of DPCP directors. Three key themes emerged: </p><list list-type="bullet"><list-item><p>The DIDIT should be a Web-based application to allow for content updates and easy accessibility (in contrast to a CD-ROM).</p></list-item><list-item><p>The DIDIT should be a reference tool that promotes consistency and standardization of data analysis required for diabetes surveillance.</p></list-item><list-item><p>The tool's development should continue to be informed by DPCP representatives (the intended user group) to ensure that it meets the needs of program staff involved in diabetes surveillance.</p></list-item></list><p>Input from these initial focus groups and subsequent feedback from the DPCPs served as the basis for developing the concept, content, and Web application for the DIDIT.</p><sec><title>Development of content</title><p>Content development took place in two stages. During the first stage, the work group selected 10 of the originally identified 55 indicators to develop a prototype. The 10 indicators were as follows: 1) diabetes prevalence, 2) annual hemoglobin A1c test, 3) annual influenza vaccination, 4) pneumococcal vaccination, 5) level of diabetes education, 6) diabetes-related hospitalizations, 7) prevalence of end-stage renal disease, 8) hospitalization for lower extremity amputations, 9) physical inactivity, and 10) overweight. Because members of the work group lived in different states, discussions were conducted through a series of telephone conferences and two in-person meetings.</p><p>During the second stage, the work group selected an additional 28 indicators from the original 55 through a two-round modified Delphi process. Indicators were ranked in priority according to the following four criteria:</p><list list-type="bullet"><list-item><p>Relationship to a national policy objective (such as the DDT's national diabetes objectives [<xref rid="B2" ref-type="bibr">2</xref>] or <italic>Healthy People 2010</italic> objectives [<xref rid="B3" ref-type="bibr">3</xref>])</p></list-item><list-item><p>Alignment with current practice guidelines, such as those from the American Diabetes Association (<xref rid="B4" ref-type="bibr">4</xref>) </p></list-item><list-item><p>Responsiveness to efforts of the DPCPs</p></list-item><list-item><p>Measurability through public data sources, particularly state-level data such as the BRFSS (<xref rid="B5" ref-type="bibr">5</xref>)</p></list-item></list><p>The typical reason for excluding an indicator was that no state-level data source could be identified to measure it. A list of indicators that were excluded and the rationale for excluding them can be found on the DIDIT (available from <ext-link xlink:href="www.cdc.gov/diabetes/statistics/index.htm" ext-link-type="uri">www.cdc.gov/diabetes/statistics/index.htm</ext-link>). All 10 indicators used to develop the prototype as well as all 28 selected during the second stage were retained, with a total of 38 indicators selected for inclusion. At this stage, the selection of fields to describe each indicator (e.g., definitions of indicators) and data source (e.g., its methodology for data collection, data access) was also finalized with input from the DIDIT work group.</p><p>After selecting indicators, associated data sources, and their related fields of information, the DIDIT team, including work group members, contractors for both content and Web site development, and DDT leaders, met in January 2003 (<xref rid="T1" ref-type="table">Table 1</xref>). The major focus of the meeting was to develop a process to ensure that the information included in the DIDIT was accurate and complete. The work group viewed the Web-based prototype with 10 indicators and eight associated data sources. Each work group member was assigned a set of indicators from the group of 38 to review and revise.</p><p>During content development, a contractor researched information on selected indicators and their associated data sources, and the work group provided feedback on its accuracy and completeness. Comments and suggestions on content were discussed during monthly conference calls, and the content was approved by all work group members before it was finalized. A second in-person meeting was held to review and refine content and to ensure that the process of reviewing content was efficient and effective. The protocol for developing and refining content resulted in valuable end-user feedback. Development of the Web-based application took place at the same time as development of content.</p></sec></sec><sec><title>Developing the Web-based Application</title><p>The parallel development of the application and its content relied on the software development life cycle (<xref rid="T2" ref-type="table">Table 2</xref>). This process included the following four major phases: 1) planning, 2) analysis, 3) design, and 4) implementation (<xref rid="B6" ref-type="bibr">6</xref>). We adopted an iterative and incremental approach, with overlap between analysis, design, and implementation.</p><sec><title>Phase 1: planning</title><p>The primary purpose of <italic>project planning</italic> was to articulate the objectives and scope of the DIDIT and ensure the technical feasibility of the system. We noted previously that the purpose and vision for the DIDIT were defined early during the concept development phase. During the project planning phase, the work group articulated the purpose and vision to the technical contractors, and they produced a document that outlined the definition and scope of the project; this document served as a blueprint for system development efforts. Additional activities in this phase included confirming the project's operational and organizational feasibility, developing a schedule, estimating costs, and allocating resources. The project planning phase culminated in the development of a final project plan, which was reviewed by the project lead and management staff. After the plan was approved, technical development efforts, or <italic>systems analysis</italic>, for the DIDIT were initiated.</p></sec><sec><title>Phase 2: analysis</title><p>The <italic>analysis</italic> phase defined in detail what the information system needed to accomplish to provide users with the benefits they desired. Several storyboards were created to display preliminary design options for Web page content and format. The storyboards were uploaded onto a secure Internet site to allow sharing among and feedback from a geographically dispersed user group. As design options were presented, users were quickly able to provide comments. An iterative feedback process allowed further revisions to DIDIT storyboards. System requirements were prioritized as they were identified. The analysis phase culminated with the development of model diagrams, which were used to drive the next phase, <italic>system design</italic>.</p></sec><sec><title>Phase 3: design</title><p>While the <italic>analysis</italic> phase focused on what the system should do, the <italic>design</italic> phase focused on how the system should function. Information from the analysis phase was used to design the application, the database, the user interface, and the operating environment. The application and database were designed in parallel with the user interface. The user interface is a critical component because it ensures ease of use. It was designed with stakeholder input to ensure that the final product would reflect stakeholder needs.</p></sec><sec><title>Phase 4: implementation</title><p>During the <italic>implementation</italic> phase, a demonstration model was built, tested, and released with information on the 10 pilot indicators and associated data sources. The model included core functionalities such as the ability to browse, sort, and search, and it was demonstrated at the first in-person work group meeting in early 2003. Input was solicited on additional features, including the addition of DPCP-specific data sources, system-searching functions, and report formats.</p><sec><title>Phase 4a: Merging the content and the Web system</title><p>After the demonstration model was developed, the development of the content and the Web system converged. The work group began using the DIDIT to review and refine the content of indicators and data sources. The contractor responsible for researching and developing the content delivered a data set on indicators and data sources to the systems developers for upload into the DIDIT. Once uploaded, the content was reviewed by the work group, and appropriate modifications were made. This process took place iteratively between February and April 2003, with a total of three uploads, until all 38 indicators and 12 data sources had been uploaded, reviewed, and finalized. An unexpected positive outcome from this process was that as the work group reviewed DIDIT content, it also tested and evaluated system functionality and design, leading to several important changes.</p><table-wrap position="float" id="TS"><table frame="hsides" rules="groups"><thead><tr><th valign="top" align="left" scope="col" rowspan="1" colspan="1">
<bold>Rationale for Conducting Pilot Tests of the Diabetes Indicators and Data Sources Internet Tool (DIDIT)</bold>
</th></tr></thead><tbody><tr><td valign="top" align="left" rowspan="1" colspan="1">1. Provides validation that system function, design, and content are consistent with the responses elicited from users during the processes of requirements gathering and usability testing. This validation closes the information loop and confirms earlier assumptions.</td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">2. Enables exploration of requirements or ideas suggested by users after the processes of requirements gathering and usability testing are complete. Although it might be too late to include these features in the first release, they can be incorporated into later phases.</td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">3. Allows users to work with a real-life model, permitting them to visualize and respond to more advanced requirements they may find difficult to comprehend without such a model. Users also understand more advanced requirements when they can work with a system designed for fundamental needs and functions. In addition, requirements often build on one another.</td></tr><tr><td valign="top" align="left" rowspan="1" colspan="1">4. Permits testing among a small subset of a large population of users, preferably subsets that differ from those selected in earlier development phases. This ensures a more representative sampling throughout the development process, and it ensures that feedback is well-rounded and unbiased. Although not all suggestions made during the pilot-testing phase are ultimately incorporated, the process often sparks ideas for future enhancements and provides insights for training and user support.</td></tr></tbody></table></table-wrap><p>Before the DIDIT was formally implemented in September 2003, both usability testing and pilot testing were conducted to obtain user feedback about system design and usage to further refine the new tool. Usability testing was conducted in April 2003 at the DDT national conference, with pilot testing in July and August 2003 through conference calls and NetMeeting (Microsoft Corp, Redmond, Wash). NetMeeting allowed participants throughout the nation to view the DIDIT as it was being demonstrated at the CDC in Atlanta, Ga. Pilot testing is a critical and often overlooked component of the software life cycle, and there are important reasons for conducting it (<xref rid="TS" ref-type="table">Sidebar</xref>). The objective of usability testing and pilot testing was to formally validate that system function, design, and content were consistent with the needs of users as determined during earlier phases. Feedback from these processes was evaluated and used to make further refinements to the system before its final release in September 2003.</p></sec><sec><title>Phase 4b: System maintenance and training</title><p>Shortly after the release of the DIDIT, the project lead conducted a national training session for DPCPs and CDC staff using NetMeeting. A team of DDT professionals was then assigned the responsibility of providing ongoing user support and training for technical and functional aspects of the DIDIT. The project lead's responsibilities included providing support on questions and issues related to the content and application of the DIDIT in the context of DPCP programs.</p></sec></sec></sec><sec><title>Elements That Facilitated the Development of the DIDIT</title><p>Several factors were critical to successfully developing and implementing the DIDIT. The factors have practical implications for other agencies that want to undertake a similar effort.</p><p>The work group members had extensive knowledge and experience in diabetes surveillance and epidemiology, which proved essential in guiding the content and technical contractors during the development process. DIDIT team members were a motivated, dedicated, enthusiastic, and knowledgeable group of DPCP representatives and DDT staff. In addition, the knowledge and skills of the contractor were critical to researching and developing content on indicators and data sources.</p><sec><title>Buy-in of management</title><p>The project lead effectively solicited the interest and support of DDT management to ensure that financial and staff resources were available to develop the new tool. To sustain interest and support of management, the project lead presented draft content and DIDIT prototypes at various CDC and national public health meetings throughout the development process (<xref rid="T1" ref-type="table">Table 1</xref>). These presentations allowed management to realize the high level of interest among prospective users and the potential for the DIDIT as an important tool for diabetes surveillance. Updates were also shared with DDT management on an ongoing basis.</p></sec><sec><title>Commitment of time and resources</title><p>Development of a comprehensive reference tool such as the DIDIT requires a commitment of time and resources. The management of the DDT supported allocation of resources and time needed to create the DIDIT.</p></sec><sec><title>Strong leadership and clear vision</title><p>The DIDIT project lead had a clear vision of the type of tool that would fulfill the surveillance needs of the DPCP and the DDT. A strength of the project lead was her ability to communicate the vision of the DIDIT to the project team and stakeholders throughout the development process.</p></sec><sec><title>Collaboration among stakeholders and contractors</title><p>Development of the DIDIT involved input from stakeholders across the country. Clear and ongoing communication among stakeholders was essential to the development process. During the first in-person work group meeting, we learned that face-to-face interactions were highly appreciated by work group members and that these interactions helped build rapport among members. In-person meetings were arranged at national conferences to avoid issues of travel approval and costs. Timelines and other defined plans facilitated collaboration. A contractor who was skillful at organizing materials, facilitating meetings, motivating work group members, and responding to the needs of work group members was also essential. Because the work group volunteered its time to create the DIDIT, efforts were made to minimize the burden placed on its members. Minimizing this burden helped to maintain a core group of members who have actively participated for more than 3 years.</p></sec><sec><title>Iterative development process</title><p>Development of both content and Web application took place incrementally and iteratively. Members of the work group reviewed the content in phases, allowing the content contractor to apply feedback to subsequent phases. Similarly, because an incremental process was conducted that involved analysis, design, and implementation at the same time, the contractors were able to make a demonstration model of the DIDIT during the early phases of development, which facilitated refinements to its content and design. Working with an actual tool triggered ideas among users for additional functions and alternative designs that may have been overlooked at the prototyping stage. A model also allowed us to obtain user input on database-driven features such as system searches.</p></sec></sec><sec><title>Implications for Public Health Practice</title><p>The ability to assess the status of the public's health in a timely, consistent, and accurate manner satisfies the first two of the 10 essential public health services as defined by the Institute of Medicine: 1) "monitor health status to identify community health problems" and 2) "diagnose and investigate health problems and health hazards in the community" (<xref rid="B7" ref-type="bibr">7</xref>).</p><p>The DIDIT represents an innovative approach to enhancing the capacity of state and federal agencies to perform public health surveillance. As one user has described, "The DIDIT offers a one-stop shop that is available 24 hours a day." It empowers users by providing them easy access to information that has been reviewed by DIDIT work group members for accuracy and content. In addition to providing a road map for development, this article highlights components that were critical to the successful development of the DIDIT. These components synergistically influenced the development process. Having adequate time, expertise, and commitment of resources, for example, would not have been sufficient for success without the clear communication and rapport among the project team members or buy-in and involvement of all stakeholders. Because these critical factors enhance one another, it is difficult to prioritize them. Other entities that wish to undertake a similar effort of systems development can use these requirements as guiding principles and customize them for their own needs and circumstances.</p><p>A major benefit of sharing these elements is to prevent other agencies from having to "reinvent the wheel" when they can draw directly on the experiences of the DIDIT team. While the technology is available to develop information technology solutions for addressing public health problems, it is vital to have effective processes and methods in place to successfully identify the needs of users and harness and customize appropriate technology to meet those needs.</p></sec> |
Hereditary profiles of disorderly transcription? | <sec><title>Background</title><p>Microscopic examination of living cells often reveals that cells from some cell strains appear to be in a permanent state of disarray without obvious reason. In all probability such a disorderly state affects cell functioning.</p><p>The aim of this study was to establish whether a disorderly state could occur that adversely affects gene expression profiles and whether such a state might have biomedical consequences. To this end, the expression profiles of the 14 genes of the proteasome derived from the GEO SAGE database were utilized as a model system.</p></sec><sec><title>Results</title><p>By adopting the overall expression profile as the standard for normal expression, deviation in transcription was frequently observed. Each deviating tissue exhibited its own characteristic profile of over-expressed and under-expressed genes. Moreover such a specific deviating profile appeared to be epigenetic in origin and could be stably transmitted to a clonal derivative e.g. from a precancerous normal tissue to its tumor. A significantly greater degree of deviation was observed in the expression profiles from the tumor tissues.</p><p>The changes in the expression of different genes display a network of interdependencies. Therefore our hypothesis is that deviating profiles reflect disorder in the localization of genes within the nucleus</p><p>The underlying cause(s) for these disorderly states remain obscure; it could be noise and/or deterministic chaos. Presence of mutational damage does not appear to be predominantly involved.</p></sec><sec><title>Conclusion</title><p>As disturbances in expression profiles frequently occur and have biomedical consequences, its determination could prove of value in several fields of biomedical research.</p></sec><sec><title>Reviewers</title><p>This article was reviewed by Trey Ideker, Itai Yanai and Stephan Beck</p></sec> | <contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Simons</surname><given-names>Johannes WIM</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>j.w.i.m.simons@lumc.nl</email></contrib> | Biology Direct | <sec><title>Open peer review</title><p>Reviewed by Trey Ideker, Itai Yanai and Stephan Beck. For the full reviews, please go to the Reviewers' comments section.</p></sec><sec><title>Background</title><p>Within a living cell there will always be "spontaneous" variation in functioning. The origin of this variation could be presence of mutational damage, random fluctuations (also known as noise) or deterministic chaos. Noise has been shown to affect cell functioning [<xref ref-type="bibr" rid="B1">1</xref>] and consequently it can be assumed that some degree of disorder will always be present in a cell. In theory, since extensive disorder could affect the health of the cell and thus ultimately the health of the individual, it would seem prudent to investigate whether a phenomenon like cellular disorder can be demonstrated and analyzed. Databases on gene transcription are now available that facilitate such an investigation.</p><p>In this study the word "disorder" is used as an inclusive term to describe "excessive variation in transcription irrespective of cause", thus random variation, deterministic chaos or presence of mutational damage could all be causally involved.</p><p>The questions we want to address are: 1) can excessive variation in transcription be demonstrated, 2) does such excessive variation have a degree of permanence and 3) does it play any role in health and disease.</p><p>That disorder in gene expression does occur and could be of relevance for understanding carcinogenesis is suggested by the following observations. Firstly, exposure of cells to carcinogens may lead to a state of "delayed, persistent genomic instability" [<xref ref-type="bibr" rid="B2">2</xref>] that can affect each aspect of cell structure and function and predisposes the cell to immortalization. Such genomic instability can also be transmitted to neighboring cells via the medium (bystander effect) [<xref ref-type="bibr" rid="B3">3</xref>,<xref ref-type="bibr" rid="B4">4</xref>] and thus does not depend on the presence of mutational damage within the cell. Persistent disorder in transcription might be implied from this unstable state.</p><p>Secondly, it is generally accepted that although mutations in oncogenes and/or tumor suppressor genes predispose cells to carcinogenesis, epigenetic and non-genetic mechanisms also play a role [<xref ref-type="bibr" rid="B5">5</xref>]. These processes ultimately lead to the cell acquiring a transformed phenotype with concomitant alterations in the life span or eventual immortalization [<xref ref-type="bibr" rid="B6">6</xref>]. Although telomerase is known to be expressed during this process of cell transformation, the way in which this gene becomes activated is still unknown [<xref ref-type="bibr" rid="B7">7</xref>,<xref ref-type="bibr" rid="B8">8</xref>]. Since the process of transformation appears to involve progressive deregulation of cell functioning, increasing disorder in transcription profiles may play a role.</p><p>Thirdly, since predisposition to tumorigenesis can be related to only a small change in the expression level of a single gene [<xref ref-type="bibr" rid="B9">9</xref>], even minor fluctuations in gene expression could have major consequences. Moreover it has recently been shown that noise in gene expression is biologically relevant as it is detrimental to organismal fitness [<xref ref-type="bibr" rid="B10">10</xref>].</p><p>Thus perturbations in the transcriptome could be present in cells that are in the process of malignant transformation and such disorder in the transcriptome might, in itself, be a driving factor in carcinogenesis.</p><p>The availability of the Cancer Genome Anatomy Project's (CGAP) SAGE (Serial Analysis of Gene Expression) database of human gene expression levels in a wide variety of cells [<xref ref-type="bibr" rid="B11">11</xref>] has enabled us to check our hypothesis and to establish that excessive variation in transcription can be constitutive and hereditary. These findings could prove to be of importance in various fields of biomedical research.</p></sec><sec><title>Results and discussion</title><sec><title>The proteasome as a model system to investigate disorder in transcription</title><p>Although the ultimate aim of this investigation is to establish the existence of a state of cellular disorder that affects all transcription profiles, the first step involves the choice of only one expression profile that could serve as a model for all profiles in a complex system although it is evident that one expression profile will not represent the whole human transcriptome. The transcription profile of the genes that code for a cellular organelle, the 20S proteasome, was chosen to fulfill this role. Since a cellular organelle has a well-defined structure, a prerequisite for its assembly would be that the products of the genes involved be available in, ideally, the correct amounts. Therefore we assume that an optimal expression pattern for the transcription of the genes in question exists although, of course, it might turn out that the expression pattern can be influenced by factors like tissue type or response to stimuli. Sampling errors will affect the number of transcripts of the proteosomal genes found in a library. However, If the degree of variation turns out to be greater than that expected due to sampling, then this could be indicative of the existence of transcriptional disorder.</p><p>The 20S proteasome, a structure 15 nm in length with a diameter of 11–12 nm, is organized as four stacked rings with a central channel. This architecture is highly conserved from bacteria to man. Each rings consists of 7 different subunits, each located at a defined position [<xref ref-type="bibr" rid="B12">12</xref>], Alpha-type subunits and beta-type subunits form the two outer and two inner rings respectively. The 14 genes coding for these subunits are all regulated independently from each other and are located on different chromosomes. Thus 14 gene products are needed in equal amounts to build the proteasome. For our calculations we assume the existence of a preferential expression profile for these 14 genes and deviation from this preferential expression profile could indicate "disorder in transcription".</p><p>The tags (listed in Table <xref ref-type="table" rid="T1">1</xref>) used to establish the degrees of expression of the proteasome genes were derived from the NCBI website [<xref ref-type="bibr" rid="B13">13</xref>] Unique cDNA tags are available for 6 of the 14 genes, whilst the remaining tags also detect expression of genes other than the proteasome genes. Therefore, additional tags specific for the expression of these non-proteasome genes were utilized to check their expression. However, in the available libraries, the expression of these non-proteasome genes appeared so rare that the counts obtained with the proteasome tags of Table <xref ref-type="table" rid="T1">1</xref> are reliable.</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>List of tags used to identify the degree of expression of the proteasome genes.</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left">Symbol</td><td align="left">tag</td><td align="left">UID</td><td align="left">gene description</td><td align="left">control tag</td></tr></thead><tbody><tr><td align="left">PSMA1</td><td align="left">GTCTGCGTGC</td><td align="left">Hs.82159</td><td align="left">proteasome subunit, alpha type, 1</td><td></td></tr><tr><td></td><td></td><td align="left">Hs.169942</td><td align="left">ESTs</td><td align="left">TCTAAGAGAA</td></tr><tr><td align="left">PSMA2</td><td align="left">GTTTAAATCG</td><td align="left">Hs.181309</td><td align="left">proteasome subunit, alpha type, 2</td><td></td></tr><tr><td></td><td></td><td align="left">Hs.1290</td><td align="left">complement component 9</td><td align="left">TGTCCAAGGG</td></tr><tr><td align="left">PSMA3</td><td align="left">AAATTGTTCC</td><td align="left">Hs.346918</td><td align="left">proteasome subunit, alpha type, 3</td><td></td></tr><tr><td></td><td></td><td align="left">Hs.270791</td><td align="left">FLJ11437 fis</td><td align="left">TGTAAATGAA</td></tr><tr><td></td><td></td><td align="left">Hs.60293</td><td align="left">FLJ10883</td><td align="left">TTTTGCCTGA</td></tr><tr><td align="left">PSMA4</td><td align="left">GACGTCTTAA</td><td align="left">Hs.251531</td><td align="left">proteasome subunit, alpha type, 4</td><td></td></tr><tr><td align="left">PSMA5</td><td align="left">TTCACAAAGG</td><td align="left">Hs.76913</td><td align="left">proteasome subunit, alpha type, 5</td><td></td></tr><tr><td></td><td></td><td align="left">Hs.11223</td><td align="left">isocitrate dehydrogenase</td><td align="left">ACCAAGGACT</td></tr><tr><td align="left">PSMA6</td><td align="left">GAGGTCCCTG</td><td align="left">Hs.74077</td><td align="left">proteasome subunit, alpha type, 6</td><td></td></tr><tr><td></td><td></td><td align="left">Hs.121516</td><td align="left">Rho GDP dissociation inhibitor</td><td align="left">TGCCCAAGAG</td></tr><tr><td align="left">PSMA7</td><td align="left">AGGCGAGATC</td><td align="left">Hs.233952</td><td align="left">proteasome subunit, alpha type, 7</td><td></td></tr><tr><td></td><td></td><td align="left">Hs.195464</td><td align="left">filamin A, alpha</td><td align="left">AGGCCGAGAT</td></tr><tr><td align="left">PSMB1</td><td align="left">CGGCTGGTGA</td><td align="left">Hs.75748</td><td align="left">proteasome subunit, beta type, 1</td><td></td></tr><tr><td></td><td></td><td align="left">Hs.334775</td><td align="left">MGC20255</td><td align="left">AGAAAGTGGC</td></tr><tr><td align="left">PSMB2</td><td align="left">TCCTCCCTCC</td><td align="left">Hs.1390</td><td align="left">proteasome subunit, beta type, 2</td><td></td></tr><tr><td align="left">PSMB3</td><td align="left">GGAGTCATTG</td><td align="left">Hs.82793</td><td align="left">proteasome subunit, beta type, 3</td><td></td></tr><tr><td align="left">PSMB4</td><td align="left">AAGGAATCGG</td><td align="left">Hs.89545</td><td align="left">proteasome subunit, beta type, 4</td><td></td></tr><tr><td align="left">PSMB5</td><td align="left">AGAAGTATAG</td><td align="left">Hs.78596</td><td align="left">proteasome subunit, beta type, 5</td><td></td></tr><tr><td align="left">PSMB6</td><td align="left">GAGCGGGATG</td><td align="left">Hs.77060</td><td align="left">proteasome subunit, beta type, 6</td><td></td></tr><tr><td align="left">PSMB7</td><td align="left">TGGCTAGTGT</td><td align="left">Hs.118065</td><td align="left">proteasome subunit, beta type, 7</td><td></td></tr><tr><td></td><td></td><td align="left">Hs.283429</td><td align="left">Smcx homolog</td><td align="left">GGCGGTGTGT</td></tr></tbody></table></table-wrap><p>In order to obtain data suitable for statistical analysis, only those libraries that have a total tag count of at least 24 for the expression of all the proteasome genes together were used. Additionally, only libraries derived from biopsies were used in order to circumvent any possible effects associated with tissue culture conditions. At the beginning of this study 60 libraries were available that met these criteria, 30 of these were derived from normal tissues and 30 from cancer tissues. A possible disadvantage of using these datasets is that they are derived from very different tissues. However, this was unavoidable due to the limited number of available libraries. Nevertheless, no significant difference between normal and tumor tissues existed for the total tag count per library (Wilcoxon: P = 0,274) or for the counted number of proteasome tags per library (Wilcoxon: P = 0,504). Therefore the groups are homogeneous with respect to these characteristics.</p></sec><sec><title>Excessive variation in transcription of proteasomal genes in libraries derived from normal and tumor tissues</title><p>Expected frequencies of tags for the 14 genes of a library were obtained with the overall expression profile. This expression profile was derived from the sums of the tag counts in all 60 libraries (Table <xref ref-type="table" rid="T2">2</xref> A). The observed tag counts are shown in Table <xref ref-type="table" rid="T3">3</xref>. The expected and observed tag counts were compared by chi-square. Of the 30 normal tissues, 13 deviated (P < 0,05) from the expected values, compared to 23 of the 30 tumor tissues. As the prerequisites for the application of the chi-square test were not met the probabilities of the chi-squares of the two groups were compared using Wilcoxon's test. The two groups proved to be significantly different from each other (P = 1.73 × 10<sup>-6</sup>), with tumor tissues being more disorderly. In conclusion it is apparent that excessive variation in transcription does occur and that as a group, tumor tissues show a significantly greater degree of variation in transcription than do normal tissues. However, the data also indicates that some normal tissues show excessive variation whilst some of the tumor tissues might still have a normal expression profile.</p><table-wrap position="float" id="T2"><label>Table 2</label><caption><p>Mean expression profiles of the proteasomal genes used for the calculation of the expected frequencies of tags.</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td align="center" colspan="2">A</td><td align="center" colspan="2">B</td><td align="center" colspan="2">C</td><td align="center" colspan="2">D</td><td align="center" colspan="2">E</td><td align="center" colspan="2">F</td></tr><tr><td></td><td align="left" colspan="2">60 libraries</td><td align="left" colspan="2">80 libraries</td><td align="left" colspan="2">All 140 libraries</td><td align="left" colspan="2">30 breast libraries</td><td align="left" colspan="2">37 most disorderly</td><td align="center" colspan="2">37 most orderly</td></tr><tr><td align="left">genes</td><td align="left">tags</td><td align="left">rel. freq.</td><td align="left">tags</td><td align="left">rel. freq.</td><td align="left">tags</td><td align="left">rel. freq.</td><td align="left">tags</td><td align="left">re. freq.</td><td align="left">tags</td><td align="left">rel. freq.</td><td align="left">tags</td><td align="left">rel. freq.</td></tr></thead><tbody><tr><td align="center">PSMA1</td><td align="center">204</td><td align="center">0,044</td><td align="center">372</td><td align="center">0,047</td><td align="center">576</td><td align="center">0,046</td><td align="center">104</td><td align="center">0,046</td><td align="center">216</td><td align="center">0,049</td><td align="center">118</td><td align="center">0,038</td></tr><tr><td align="center">PSMA2</td><td align="center">397</td><td align="center">0,086</td><td align="center">800</td><td align="center">0,100</td><td align="center">1197</td><td align="center">0,095</td><td align="center">188</td><td align="center">0,084</td><td align="center">406</td><td align="center">0,091</td><td align="center">290</td><td align="center">0,093</td></tr><tr><td align="center">PSMA3</td><td align="center">81</td><td align="center">0,018</td><td align="center">191</td><td align="center">0,024</td><td align="center">272</td><td align="center">0,022</td><td align="center">40</td><td align="center">0,018</td><td align="center">73</td><td align="center">0,016</td><td align="center">78</td><td align="center">0,025</td></tr><tr><td align="center">PSMA4</td><td align="center">321</td><td align="center">0,070</td><td align="center">517</td><td align="center">0,065</td><td align="center">838</td><td align="center">0,067</td><td align="center">160</td><td align="center">0,071</td><td align="center">265</td><td align="center">0,060</td><td align="center">263</td><td align="center">0,084</td></tr><tr><td align="center">PSMA5</td><td align="center">169</td><td align="center">0,037</td><td align="center">263</td><td align="center">0,033</td><td align="center">432</td><td align="center">0,034</td><td align="center">44</td><td align="center">0,020</td><td align="center">143</td><td align="center">0,032</td><td align="center">109</td><td align="center">0,035</td></tr><tr><td align="center">PSMA6</td><td align="center">578</td><td align="center">0,125</td><td align="center">1159</td><td align="center">0,146</td><td align="center">1737</td><td align="center">0,138</td><td align="center">327</td><td align="center">0,146</td><td align="center">574</td><td align="center">0,129</td><td align="center">449</td><td align="center">0,144</td></tr><tr><td align="center">PSMA7</td><td align="center">335</td><td align="center">0,073</td><td align="center">499</td><td align="center">0,063</td><td align="center">834</td><td align="center">0,066</td><td align="center">189</td><td align="center">0,084</td><td align="center">323</td><td align="center">0,073</td><td align="center">210</td><td align="center">0,067</td></tr><tr><td align="center">PSMB1</td><td align="center">410</td><td align="center">0,089</td><td align="center">729</td><td align="center">0,092</td><td align="center">1139</td><td align="center">0,091</td><td align="center">237</td><td align="center">0,106</td><td align="center">391</td><td align="center">0,088</td><td align="center">271</td><td align="center">0,087</td></tr><tr><td align="center">PSMB2</td><td align="center">185</td><td align="center">0,040</td><td align="center">357</td><td align="center">0,045</td><td align="center">542</td><td align="center">0,043</td><td align="center">87</td><td align="center">0,039</td><td align="center">170</td><td align="center">0,038</td><td align="center">115</td><td align="center">0,037</td></tr><tr><td align="center">PSMB3</td><td align="center">652</td><td align="center">0,141</td><td align="center">761</td><td align="center">0,096</td><td align="center">1413</td><td align="center">0,112</td><td align="center">409</td><td align="center">0,182</td><td align="center">523</td><td align="center">0,118</td><td align="center">326</td><td align="center">0,104</td></tr><tr><td align="center">PSMB4</td><td align="center">197</td><td align="center">0,043</td><td align="center">230</td><td align="center">0,029</td><td align="center">427</td><td align="center">0,034</td><td align="center">67</td><td align="center">0,030</td><td align="center">175</td><td align="center">0,039</td><td align="center">99</td><td align="center">0,032</td></tr><tr><td align="center">PSMB5</td><td align="center">121</td><td align="center">0,026</td><td align="center">219</td><td align="center">0,028</td><td align="center">340</td><td align="center">0,027</td><td align="center">25</td><td align="center">0,011</td><td align="center">128</td><td align="center">0,029</td><td align="center">67</td><td align="center">0,021</td></tr><tr><td align="center">PSMB6</td><td align="center">370</td><td align="center">0,080</td><td align="center">594</td><td align="center">0,075</td><td align="center">964</td><td align="center">0,077</td><td align="center">141</td><td align="center">0,063</td><td align="center">344</td><td align="center">0,077</td><td align="center">230</td><td align="center">0,074</td></tr><tr><td align="center">PSMB7</td><td align="center">589</td><td align="center">0,128</td><td align="center">1270</td><td align="center">0,160</td><td align="center">1859</td><td align="center">0,148</td><td align="center">224</td><td align="center">0,100</td><td align="center">708</td><td align="center">0,159</td><td align="center">468</td><td align="center">0,150</td></tr><tr><td></td><td align="center">4609</td><td align="center">1.000</td><td align="center">7961</td><td align="center">1.000</td><td align="center">12570</td><td align="center">1.000</td><td align="center">2242</td><td align="center">1,000</td><td align="center">4439</td><td align="center">1,000</td><td align="center">3125</td><td align="center">1,000</td></tr></tbody></table></table-wrap><table-wrap position="float" id="T3"><label>Table 3</label><caption><p>Expression of the proteasome genes in the selected SAGE libraries.</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td align="left" colspan="2">deviation index</td></tr><tr><td align="left">GEO</td><td align="left">total tags</td><td align="left">tissue</td><td align="left">PSMA1</td><td align="left">PSMA2</td><td align="left">PSMA3</td><td align="left">PSMA4</td><td align="left">PSMA5</td><td align="left">PSMA6</td><td align="left">PSMA7</td><td align="left">PSMB1</td><td align="left">PSMB2</td><td align="left">PSMB3</td><td align="left">PSMB4</td><td align="left">PSMB5</td><td align="left">PSMB6</td><td align="left">PSMB7</td><td align="left">Sum</td><td></td><td align="left">log ratio</td><td align="left">z-score</td></tr></thead><tbody><tr><td align="left" colspan="21">30 normal libraries</td></tr><tr><td align="left">676</td><td align="left">94876</td><td align="left">brain</td><td align="left">6</td><td align="left">4</td><td align="left">0</td><td align="left">0</td><td align="left">6</td><td align="left">3</td><td align="left">6</td><td align="left">5</td><td align="left">7</td><td align="left">1</td><td align="left">3</td><td align="left">1</td><td align="left">4</td><td align="left">8</td><td align="left">54</td><td></td><td align="left">0,377</td><td align="left">1,698</td></tr><tr><td align="left">677</td><td align="left">37642</td><td align="left">breast</td><td align="left">4</td><td align="left">6</td><td align="left">2</td><td align="left">4</td><td align="left">3</td><td align="left">14</td><td align="left">7</td><td align="left">12</td><td align="left">1</td><td align="left">7</td><td align="left">4</td><td align="left">0</td><td align="left">3</td><td align="left">6</td><td align="left">73</td><td></td><td align="left">0,224</td><td align="left">1,087</td></tr><tr><td align="left">780</td><td align="left">63227</td><td align="left">breast</td><td align="left">2</td><td align="left">4</td><td align="left">2</td><td align="left">4</td><td align="left">4</td><td align="left">4</td><td align="left">9</td><td align="left">4</td><td align="left">2</td><td align="left">6</td><td align="left">1</td><td align="left">1</td><td align="left">2</td><td align="left">10</td><td align="left">55</td><td></td><td align="left">0,221</td><td align="left">1,083</td></tr><tr><td align="left">781</td><td align="left">58444</td><td align="left">breast</td><td align="left">1</td><td align="left">10</td><td align="left">0</td><td align="left">0</td><td align="left">0</td><td align="left">3</td><td align="left">8</td><td align="left">1</td><td align="left">4</td><td align="left">4</td><td align="left">1</td><td align="left">2</td><td align="left">6</td><td align="left">5</td><td align="left">45</td><td></td><td align="left">0,326</td><td align="left">1,556</td></tr><tr><td align="left">685</td><td align="left">66483</td><td align="left">prostate</td><td align="left">8</td><td align="left">15</td><td align="left">2</td><td align="left">8</td><td align="left">6</td><td align="left">11</td><td align="left">20</td><td align="left">19</td><td align="left">4</td><td align="left">28</td><td align="left">29</td><td align="left">9</td><td align="left">9</td><td align="left">21</td><td align="left">189</td><td></td><td align="left">0,274</td><td align="left">2,856</td></tr><tr><td align="left">688</td><td align="left">28950</td><td align="left">breast</td><td align="left">1</td><td align="left">3</td><td align="left">3</td><td align="left">8</td><td align="left">1</td><td align="left">9</td><td align="left">5</td><td align="left">4</td><td align="left">2</td><td align="left">7</td><td align="left">1</td><td align="left">0</td><td align="left">4</td><td align="left">6</td><td align="left">54</td><td></td><td align="left">0,232</td><td align="left">1,037</td></tr><tr><td align="left">691</td><td align="left">7165</td><td align="left">breast</td><td align="left">1</td><td align="left">5</td><td align="left">1</td><td align="left">3</td><td align="left">1</td><td align="left">7</td><td align="left">1</td><td align="left">1</td><td align="left">0</td><td align="left">4</td><td align="left">1</td><td align="left">0</td><td align="left">1</td><td align="left">2</td><td align="left">28</td><td></td><td align="left">0,235</td><td align="left">0,911</td></tr><tr><td align="left">692</td><td align="left">12142</td><td align="left">breast</td><td align="left">1</td><td align="left">7</td><td align="left">0</td><td align="left">2</td><td align="left">0</td><td align="left">2</td><td align="left">0</td><td align="left">2</td><td align="left">1</td><td align="left">3</td><td align="left">0</td><td align="left">0</td><td align="left">2</td><td align="left">4</td><td align="left">24</td><td></td><td align="left">0,181</td><td align="left">1,063</td></tr><tr><td align="left">695</td><td align="left">58826</td><td align="left">brain</td><td align="left">2</td><td align="left">2</td><td align="left">1</td><td align="left">1</td><td align="left">5</td><td align="left">5</td><td align="left">1</td><td align="left">4</td><td align="left">0</td><td align="left">2</td><td align="left">0</td><td align="left">1</td><td align="left">5</td><td align="left">7</td><td align="left">36</td><td></td><td align="left">0,270</td><td align="left">1,244</td></tr><tr><td align="left">708</td><td align="left">41857</td><td align="left">kidney</td><td align="left">2</td><td align="left">2</td><td align="left">3</td><td align="left">4</td><td align="left">1</td><td align="left">6</td><td align="left">7</td><td align="left">1</td><td align="left">0</td><td align="left">4</td><td align="left">2</td><td align="left">1</td><td align="left">4</td><td align="left">5</td><td align="left">42</td><td></td><td align="left">0,280</td><td align="left">1,156</td></tr><tr><td align="left">713</td><td align="left">48548</td><td align="left">brain</td><td align="left">1</td><td align="left">0</td><td align="left">0</td><td align="left">1</td><td align="left">1</td><td align="left">5</td><td align="left">0</td><td align="left">4</td><td align="left">1</td><td align="left">4</td><td align="left">0</td><td align="left">2</td><td align="left">5</td><td align="left">5</td><td align="left">29</td><td></td><td align="left">0,243</td><td align="left">1,017</td></tr><tr><td align="left">574</td><td align="left">102359</td><td align="left">retina</td><td align="left">4</td><td align="left">13</td><td align="left">2</td><td align="left">19</td><td align="left">0</td><td align="left">17</td><td align="left">7</td><td align="left">5</td><td align="left">0</td><td align="left">16</td><td align="left">5</td><td align="left">2</td><td align="left">8</td><td align="left">36</td><td align="left">134</td><td></td><td align="left">0,305</td><td align="left">1,784</td></tr><tr><td align="left">719</td><td align="left">48552</td><td align="left">ovary</td><td align="left">6</td><td align="left">14</td><td align="left">0</td><td align="left">10</td><td align="left">3</td><td align="left">7</td><td align="left">4</td><td align="left">13</td><td align="left">9</td><td align="left">9</td><td align="left">8</td><td align="left">2</td><td align="left">7</td><td align="left">8</td><td align="left">100</td><td></td><td align="left">0,230</td><td align="left">1,453</td></tr><tr><td align="left">572</td><td align="left">59661</td><td align="left">retina</td><td align="left">3</td><td align="left">1</td><td align="left">1</td><td align="left">5</td><td align="left">2</td><td align="left">6</td><td align="left">6</td><td align="left">6</td><td align="left">2</td><td align="left">8</td><td align="left">2</td><td align="left">0</td><td align="left">9</td><td align="left">16</td><td align="left">67</td><td></td><td align="left">0,256</td><td align="left">1,081</td></tr><tr><td align="left">573</td><td align="left">105312</td><td align="left">retina</td><td align="left">2</td><td align="left">9</td><td align="left">4</td><td align="left">14</td><td align="left">3</td><td align="left">8</td><td align="left">5</td><td align="left">5</td><td align="left">3</td><td align="left">14</td><td align="left">5</td><td align="left">1</td><td align="left">8</td><td align="left">29</td><td align="left">110</td><td></td><td align="left">0,243</td><td align="left">1,481</td></tr><tr><td align="left">760</td><td align="left">49281</td><td align="left">ovary</td><td align="left">11</td><td align="left">8</td><td align="left">1</td><td align="left">8</td><td align="left">2</td><td align="left">27</td><td align="left">10</td><td align="left">21</td><td align="left">3</td><td align="left">3</td><td align="left">6</td><td align="left">0</td><td align="left">10</td><td align="left">8</td><td align="left">118</td><td></td><td align="left">0,310</td><td align="left">1,849</td></tr><tr><td align="left">728</td><td align="left">50179</td><td align="left">colon</td><td align="left">5</td><td align="left">1</td><td align="left">2</td><td align="left">1</td><td align="left">4</td><td align="left">1</td><td align="left">6</td><td align="left">11</td><td align="left">2</td><td align="left">0</td><td align="left">1</td><td align="left">1</td><td align="left">9</td><td align="left">2</td><td align="left">46</td><td></td><td align="left">0,468</td><td align="left">1,918</td></tr><tr><td align="left">729</td><td align="left">49593</td><td align="left">colon</td><td align="left">5</td><td align="left">0</td><td align="left">1</td><td align="left">2</td><td align="left">1</td><td align="left">4</td><td align="left">5</td><td align="left">7</td><td align="left">3</td><td align="left">7</td><td align="left">1</td><td align="left">0</td><td align="left">6</td><td align="left">6</td><td align="left">48</td><td></td><td align="left">0,265</td><td align="left">1,130</td></tr><tr><td align="left">761</td><td align="left">51280</td><td align="left">brain</td><td align="left">4</td><td align="left">1</td><td align="left">2</td><td align="left">4</td><td align="left">0</td><td align="left">5</td><td align="left">4</td><td align="left">1</td><td align="left">2</td><td align="left">1</td><td align="left">0</td><td align="left">1</td><td align="left">6</td><td align="left">6</td><td align="left">37</td><td></td><td align="left">0,337</td><td align="left">1,204</td></tr><tr><td align="left">1499</td><td align="left">84357</td><td align="left">heart</td><td align="left">6</td><td align="left">16</td><td align="left">4</td><td align="left">7</td><td align="left">7</td><td align="left">28</td><td align="left">11</td><td align="left">14</td><td align="left">1</td><td align="left">17</td><td align="left">1</td><td align="left">5</td><td align="left">14</td><td align="left">27</td><td align="left">158</td><td></td><td align="left">0,302</td><td align="left">1,005</td></tr><tr><td align="left">819</td><td align="left">53853</td><td align="left">muscle</td><td align="left">2</td><td align="left">6</td><td align="left">0</td><td align="left">3</td><td align="left">1</td><td align="left">9</td><td align="left">10</td><td align="left">2</td><td align="left">3</td><td align="left">2</td><td align="left">5</td><td align="left">3</td><td align="left">6</td><td align="left">7</td><td align="left">59</td><td></td><td align="left">0,279</td><td align="left">1,348</td></tr><tr><td align="left">824</td><td align="left">53875</td><td align="left">muscle</td><td align="left">5</td><td align="left">3</td><td align="left">2</td><td align="left">2</td><td align="left">3</td><td align="left">13</td><td align="left">9</td><td align="left">3</td><td align="left">2</td><td align="left">4</td><td align="left">0</td><td align="left">6</td><td align="left">10</td><td align="left">9</td><td align="left">71</td><td></td><td align="left">0,291</td><td align="left">1,492</td></tr><tr><td align="left">785</td><td align="left">66861</td><td align="left">liver</td><td align="left">2</td><td align="left">14</td><td align="left">2</td><td align="left">9</td><td align="left">12</td><td align="left">14</td><td align="left">2</td><td align="left">14</td><td align="left">3</td><td align="left">15</td><td align="left">2</td><td align="left">4</td><td align="left">7</td><td align="left">27</td><td align="left">127</td><td></td><td align="left">0,273</td><td align="left">1,504</td></tr><tr><td align="left">762</td><td align="left">89143</td><td align="left">lung</td><td align="left">1</td><td align="left">15</td><td align="left">6</td><td align="left">21</td><td align="left">7</td><td align="left">27</td><td align="left">2</td><td align="left">10</td><td align="left">6</td><td align="left">22</td><td align="left">1</td><td align="left">4</td><td align="left">10</td><td align="left">23</td><td align="left">155</td><td></td><td align="left">0,369</td><td align="left">1,597</td></tr><tr><td align="left">786</td><td align="left">77986</td><td align="left">brain</td><td align="left">3</td><td align="left">4</td><td align="left">0</td><td align="left">4</td><td align="left">1</td><td align="left">8</td><td align="left">1</td><td align="left">3</td><td align="left">2</td><td align="left">4</td><td align="left">0</td><td align="left">5</td><td align="left">4</td><td align="left">24</td><td align="left">63</td><td></td><td align="left">0,284</td><td align="left">1,744</td></tr><tr><td align="left">763</td><td align="left">63208</td><td align="left">brain</td><td align="left">4</td><td align="left">0</td><td align="left">0</td><td align="left">5</td><td align="left">0</td><td align="left">2</td><td align="left">4</td><td align="left">5</td><td align="left">3</td><td align="left">7</td><td align="left">4</td><td align="left">0</td><td align="left">5</td><td align="left">5</td><td align="left">44</td><td></td><td align="left">0,292</td><td align="left">1,284</td></tr><tr><td align="left">3242</td><td align="left">37292</td><td align="left">skin</td><td align="left">0</td><td align="left">5</td><td align="left">0</td><td align="left">6</td><td align="left">1</td><td align="left">5</td><td align="left">8</td><td align="left">3</td><td align="left">1</td><td align="left">4</td><td align="left">0</td><td align="left">0</td><td align="left">3</td><td align="left">4</td><td align="left">40</td><td></td><td align="left">0,210</td><td align="left">1,289</td></tr><tr><td align="left">2386</td><td align="left">55422</td><td align="left">spinal cord</td><td align="left">2</td><td align="left">7</td><td align="left">3</td><td align="left">3</td><td align="left">1</td><td align="left">6</td><td align="left">3</td><td align="left">8</td><td align="left">2</td><td align="left">9</td><td align="left">2</td><td align="left">2</td><td align="left">5</td><td align="left">9</td><td align="left">62</td><td></td><td align="left">0,170</td><td align="left">0,698</td></tr><tr><td align="left">738</td><td align="left">54096</td><td align="left">peritoneum</td><td align="left">2</td><td align="left">1</td><td align="left">0</td><td align="left">0</td><td align="left">1</td><td align="left">2</td><td align="left">5</td><td align="left">4</td><td align="left">3</td><td align="left">4</td><td align="left">1</td><td align="left">0</td><td align="left">4</td><td align="left">4</td><td align="left">31</td><td></td><td align="left">0,253</td><td align="left">1,051</td></tr><tr><td align="left">739</td><td align="left">59553</td><td align="left">prostate</td><td align="left">6</td><td align="left">7</td><td align="left">3</td><td align="left">8</td><td align="left">7</td><td align="left">8</td><td align="left">0</td><td align="left">6</td><td align="left">3</td><td align="left">11</td><td align="left">0</td><td align="left">8</td><td align="left">9</td><td align="left">10</td><td align="left">86</td><td></td><td align="left">0,328</td><td align="left">1,611</td></tr><tr><td align="left" colspan="21">30 tumor libraries</td></tr><tr><td align="left">2451</td><td align="left">38634</td><td align="left">brain</td><td align="left">0</td><td align="left">3</td><td align="left">0</td><td align="left">3</td><td align="left">2</td><td align="left">3</td><td align="left">3</td><td align="left">4</td><td align="left">1</td><td align="left">4</td><td align="left">2</td><td align="left">3</td><td align="left">5</td><td align="left">7</td><td align="left">40</td><td></td><td align="left">0,199</td><td align="left">0,872</td></tr><tr><td align="left">2443</td><td align="left">80265</td><td align="left">brain</td><td align="left">2</td><td align="left">15</td><td align="left">0</td><td align="left">10</td><td align="left">2</td><td align="left">15</td><td align="left">8</td><td align="left">10</td><td align="left">7</td><td align="left">13</td><td align="left">1</td><td align="left">3</td><td align="left">8</td><td align="left">12</td><td align="left">106</td><td></td><td align="left">0,229</td><td align="left">0,981</td></tr><tr><td align="left">2578</td><td align="left">69513</td><td align="left">brain</td><td align="left">0</td><td align="left">10</td><td align="left">3</td><td align="left">0</td><td align="left">6</td><td align="left">12</td><td align="left">1</td><td align="left">9</td><td align="left">6</td><td align="left">9</td><td align="left">4</td><td align="left">3</td><td align="left">5</td><td align="left">11</td><td align="left">79</td><td></td><td align="left">0,313</td><td align="left">1,246</td></tr><tr><td align="left">699</td><td align="left">28159</td><td align="left">brain</td><td align="left">1</td><td align="left">3</td><td align="left">1</td><td align="left">0</td><td align="left">3</td><td align="left">7</td><td align="left">13</td><td align="left">2</td><td align="left">4</td><td align="left">1</td><td align="left">12</td><td align="left">2</td><td align="left">10</td><td align="left">9</td><td align="left">68</td><td></td><td align="left">0,441</td><td align="left">2,451</td></tr><tr><td align="left">698</td><td align="left">77004</td><td align="left">brain</td><td align="left">4</td><td align="left">5</td><td align="left">1</td><td align="left">3</td><td align="left">2</td><td align="left">10</td><td align="left">7</td><td align="left">5</td><td align="left">1</td><td align="left">3</td><td align="left">6</td><td align="left">4</td><td align="left">11</td><td align="left">2</td><td align="left">64</td><td></td><td align="left">0,323</td><td align="left">1,504</td></tr><tr><td align="left">697</td><td align="left">52479</td><td align="left">brain</td><td align="left">0</td><td align="left">4</td><td align="left">2</td><td align="left">2</td><td align="left">2</td><td align="left">4</td><td align="left">6</td><td align="left">4</td><td align="left">0</td><td align="left">4</td><td align="left">1</td><td align="left">0</td><td align="left">0</td><td align="left">5</td><td align="left">34</td><td></td><td align="left">0,230</td><td align="left">1,115</td></tr><tr><td align="left">672</td><td align="left">67386</td><td align="left">breast</td><td align="left">2</td><td align="left">9</td><td align="left">3</td><td align="left">9</td><td align="left">5</td><td align="left">15</td><td align="left">9</td><td align="left">9</td><td align="left">3</td><td align="left">62</td><td align="left">1</td><td align="left">1</td><td align="left">8</td><td align="left">6</td><td align="left">142</td><td></td><td align="left">0,327</td><td align="left">3,481</td></tr><tr><td align="left">687</td><td align="left">41378</td><td align="left">breast</td><td align="left">2</td><td align="left">9</td><td align="left">1</td><td align="left">1</td><td align="left">1</td><td align="left">25</td><td align="left">1</td><td align="left">15</td><td align="left">0</td><td align="left">1</td><td align="left">0</td><td align="left">0</td><td align="left">3</td><td align="left">5</td><td align="left">64</td><td></td><td align="left">0,375</td><td align="left">2,245</td></tr><tr><td align="left">670</td><td align="left">40223</td><td align="left">breast</td><td align="left">3</td><td align="left">3</td><td align="left">0</td><td align="left">14</td><td align="left">1</td><td align="left">12</td><td align="left">6</td><td align="left">4</td><td align="left">4</td><td align="left">65</td><td align="left">1</td><td align="left">1</td><td align="left">2</td><td align="left">8</td><td align="left">124</td><td></td><td align="left">0,375</td><td align="left">4,170</td></tr><tr><td align="left">755</td><td align="left">57686</td><td align="left">colon</td><td align="left">5</td><td align="left">3</td><td align="left">2</td><td align="left">2</td><td align="left">2</td><td align="left">4</td><td align="left">3</td><td align="left">4</td><td align="left">0</td><td align="left">3</td><td align="left">4</td><td align="left">0</td><td align="left">2</td><td align="left">3</td><td align="left">37</td><td></td><td align="left">0,252</td><td align="left">1,216</td></tr><tr><td align="left">756</td><td align="left">49064</td><td align="left">colon</td><td align="left">7</td><td align="left">6</td><td align="left">1</td><td align="left">1</td><td align="left">5</td><td align="left">5</td><td align="left">2</td><td align="left">7</td><td align="left">5</td><td align="left">9</td><td align="left">0</td><td align="left">0</td><td align="left">2</td><td align="left">6</td><td align="left">56</td><td></td><td align="left">0,300</td><td align="left">1,423</td></tr><tr><td align="left">792</td><td align="left">34537</td><td align="left">brain</td><td align="left">2</td><td align="left">4</td><td align="left">2</td><td align="left">3</td><td align="left">2</td><td align="left">9</td><td align="left">0</td><td align="left">15</td><td align="left">2</td><td align="left">5</td><td align="left">0</td><td align="left">0</td><td align="left">3</td><td align="left">5</td><td align="left">52</td><td></td><td align="left">0,239</td><td align="left">1,565</td></tr><tr><td align="left">1497</td><td align="left">46928</td><td align="left">brain</td><td align="left">6</td><td align="left">8</td><td align="left">0</td><td align="left">2</td><td align="left">5</td><td align="left">6</td><td align="left">0</td><td align="left">4</td><td align="left">9</td><td align="left">9</td><td align="left">0</td><td align="left">5</td><td align="left">6</td><td align="left">9</td><td align="left">69</td><td></td><td align="left">0,331</td><td align="left">1,651</td></tr><tr><td align="left">793</td><td align="left">56871</td><td align="left">brain</td><td align="left">9</td><td align="left">9</td><td align="left">3</td><td align="left">8</td><td align="left">3</td><td align="left">13</td><td align="left">2</td><td align="left">17</td><td align="left">0</td><td align="left">7</td><td align="left">2</td><td align="left">0</td><td align="left">4</td><td align="left">13</td><td align="left">90</td><td></td><td align="left">0,281</td><td align="left">1,471</td></tr><tr><td align="left">696</td><td align="left">70087</td><td align="left">brain</td><td align="left">3</td><td align="left">11</td><td align="left">0</td><td align="left">3</td><td align="left">4</td><td align="left">5</td><td align="left">11</td><td align="left">13</td><td align="left">11</td><td align="left">10</td><td align="left">6</td><td align="left">3</td><td align="left">15</td><td align="left">13</td><td align="left">108</td><td></td><td align="left">0,257</td><td align="left">1,561</td></tr><tr><td align="left">765</td><td align="left">61886</td><td align="left">brain</td><td align="left">4</td><td align="left">3</td><td align="left">0</td><td align="left">2</td><td align="left">2</td><td align="left">5</td><td align="left">13</td><td align="left">9</td><td align="left">8</td><td align="left">10</td><td align="left">0</td><td align="left">0</td><td align="left">11</td><td align="left">10</td><td align="left">77</td><td></td><td align="left">0,291</td><td align="left">1,714</td></tr><tr><td align="left">745</td><td align="left">60069</td><td align="left">brain</td><td align="left">2</td><td align="left">6</td><td align="left">0</td><td align="left">7</td><td align="left">1</td><td align="left">29</td><td align="left">6</td><td align="left">3</td><td align="left">0</td><td align="left">5</td><td align="left">0</td><td align="left">0</td><td align="left">12</td><td align="left">15</td><td align="left">86</td><td></td><td align="left">0,295</td><td align="left">1,908</td></tr><tr><td align="left">1516</td><td align="left">76168</td><td align="left">brain</td><td align="left">6</td><td align="left">18</td><td align="left">1</td><td align="left">14</td><td align="left">2</td><td align="left">20</td><td align="left">4</td><td align="left">11</td><td align="left">12</td><td align="left">11</td><td align="left">1</td><td align="left">2</td><td align="left">7</td><td align="left">19</td><td align="left">128</td><td></td><td align="left">0,271</td><td align="left">1,351</td></tr><tr><td align="left">693</td><td align="left">19572</td><td align="left">brain</td><td align="left">2</td><td align="left">1</td><td align="left">0</td><td align="left">2</td><td align="left">3</td><td align="left">8</td><td align="left">2</td><td align="left">3</td><td align="left">5</td><td align="left">0</td><td align="left">4</td><td align="left">4</td><td align="left">2</td><td align="left">3</td><td align="left">39</td><td></td><td align="left">0,375</td><td align="left">1,568</td></tr><tr><td align="left">690</td><td align="left">38933</td><td align="left">brain</td><td align="left">1</td><td align="left">4</td><td align="left">1</td><td align="left">3</td><td align="left">2</td><td align="left">18</td><td align="left">3</td><td align="left">9</td><td align="left">5</td><td align="left">7</td><td align="left">1</td><td align="left">1</td><td align="left">1</td><td align="left">13</td><td align="left">69</td><td></td><td align="left">0,280</td><td align="left">1,212</td></tr><tr><td align="left">727</td><td align="left">35181</td><td align="left">brain</td><td align="left">2</td><td align="left">4</td><td align="left">0</td><td align="left">0</td><td align="left">2</td><td align="left">2</td><td align="left">9</td><td align="left">4</td><td align="left">2</td><td align="left">3</td><td align="left">1</td><td align="left">1</td><td align="left">1</td><td align="left">6</td><td align="left">37</td><td></td><td align="left">0,269</td><td align="left">1,386</td></tr><tr><td align="left">671</td><td align="left">45673</td><td align="left">breast</td><td align="left">0</td><td align="left">9</td><td align="left">1</td><td align="left">21</td><td align="left">1</td><td align="left">11</td><td align="left">10</td><td align="left">7</td><td align="left">4</td><td align="left">121</td><td align="left">0</td><td align="left">0</td><td align="left">5</td><td align="left">10</td><td align="left">200</td><td></td><td align="left">0,447</td><td align="left">6,252</td></tr><tr><td align="left">673</td><td align="left">61040</td><td align="left">breast</td><td align="left">3</td><td align="left">12</td><td align="left">1</td><td align="left">12</td><td align="left">2</td><td align="left">24</td><td align="left">6</td><td align="left">6</td><td align="left">2</td><td align="left">9</td><td align="left">4</td><td align="left">2</td><td align="left">11</td><td align="left">7</td><td align="left">101</td><td></td><td align="left">0,208</td><td align="left">1,315</td></tr><tr><td align="left">689</td><td align="left">28133</td><td align="left">brain</td><td align="left">5</td><td align="left">3</td><td align="left">0</td><td align="left">1</td><td align="left">2</td><td align="left">4</td><td align="left">6</td><td align="left">5</td><td align="left">1</td><td align="left">2</td><td align="left">0</td><td align="left">0</td><td align="left">0</td><td align="left">1</td><td align="left">30</td><td></td><td align="left">0,354</td><td align="left">1,481</td></tr><tr><td align="left">731</td><td align="left">17485</td><td align="left">ovary</td><td align="left">0</td><td align="left">1</td><td align="left">2</td><td align="left">2</td><td align="left">2</td><td align="left">2</td><td align="left">1</td><td align="left">1</td><td align="left">1</td><td align="left">3</td><td align="left">0</td><td align="left">2</td><td align="left">3</td><td align="left">6</td><td align="left">26</td><td></td><td align="left">0,293</td><td align="left">1,031</td></tr><tr><td align="left">735</td><td align="left">42445</td><td align="left">ovary</td><td align="left">2</td><td align="left">5</td><td align="left">0</td><td align="left">0</td><td align="left">4</td><td align="left">3</td><td align="left">9</td><td align="left">4</td><td align="left">5</td><td align="left">4</td><td align="left">3</td><td align="left">5</td><td align="left">5</td><td align="left">2</td><td align="left">51</td><td></td><td align="left">0,313</td><td align="left">1,544</td></tr><tr><td align="left">737</td><td align="left">33675</td><td align="left">ovary</td><td align="left">3</td><td align="left">2</td><td align="left">0</td><td align="left">0</td><td align="left">2</td><td align="left">0</td><td align="left">3</td><td align="left">2</td><td align="left">3</td><td align="left">0</td><td align="left">1</td><td align="left">2</td><td align="left">4</td><td align="left">4</td><td align="left">26</td><td></td><td align="left">0,328</td><td align="left">1,275</td></tr><tr><td align="left">736</td><td align="left">55002</td><td align="left">ovary</td><td align="left">13</td><td align="left">12</td><td align="left">2</td><td align="left">7</td><td align="left">1</td><td align="left">3</td><td align="left">10</td><td align="left">6</td><td align="left">5</td><td align="left">5</td><td align="left">9</td><td align="left">4</td><td align="left">14</td><td align="left">11</td><td align="left">102</td><td></td><td align="left">0,331</td><td align="left">1,897</td></tr><tr><td align="left">740</td><td align="left">65351</td><td align="left">prostate</td><td align="left">5</td><td align="left">14</td><td align="left">5</td><td align="left">6</td><td align="left">9</td><td align="left">20</td><td align="left">1</td><td align="left">7</td><td align="left">1</td><td align="left">12</td><td align="left">1</td><td align="left">2</td><td align="left">16</td><td align="left">14</td><td align="left">113</td><td></td><td align="left">0,373</td><td align="left">1,540</td></tr><tr><td align="left">686</td><td align="left">68626</td><td align="left">prostate</td><td align="left">8</td><td align="left">18</td><td align="left">2</td><td align="left">1</td><td align="left">5</td><td align="left">18</td><td align="left">14</td><td align="left">13</td><td align="left">3</td><td align="left">28</td><td align="left">43</td><td align="left">10</td><td align="left">9</td><td align="left">15</td><td align="left">187</td><td></td><td align="left">0,420</td><td align="left">4,321</td></tr><tr><td align="left" colspan="21">12 additional normal libraries</td></tr><tr><td align="left">14799</td><td align="left">308589</td><td align="left">brain</td><td align="left">35</td><td align="left">44</td><td align="left">5</td><td align="left">12</td><td align="left">8</td><td align="left">44</td><td align="left">11</td><td align="left">9</td><td align="left">15</td><td align="left">11</td><td align="left">9</td><td align="left">33</td><td align="left">39</td><td align="left">55</td><td align="left">330</td><td></td><td align="left">0,310</td><td align="left">3,377</td></tr><tr><td align="left">14796</td><td align="left">42498</td><td align="left">brain</td><td align="left">0</td><td align="left">6</td><td align="left">0</td><td align="left">4</td><td align="left">0</td><td align="left">3</td><td align="left">3</td><td align="left">1</td><td align="left">2</td><td align="left">3</td><td align="left">0</td><td align="left">1</td><td align="left">1</td><td align="left">11</td><td align="left">35</td><td></td><td align="left">0,249</td><td align="left">1,183</td></tr><tr><td align="left">14754</td><td align="left">33115</td><td align="left">breast</td><td align="left">5</td><td align="left">6</td><td align="left">1</td><td align="left">1</td><td align="left">0</td><td align="left">2</td><td align="left">3</td><td align="left">6</td><td align="left">1</td><td align="left">0</td><td align="left">0</td><td align="left">0</td><td align="left">4</td><td align="left">1</td><td align="left">30</td><td></td><td align="left">0,348</td><td align="left">1,505</td></tr><tr><td align="left">14756</td><td align="left">58181</td><td align="left">breast</td><td align="left">1</td><td align="left">7</td><td align="left">1</td><td align="left">1</td><td align="left">4</td><td align="left">7</td><td align="left">7</td><td align="left">4</td><td align="left">5</td><td align="left">11</td><td align="left">2</td><td align="left">1</td><td align="left">3</td><td align="left">11</td><td align="left">65</td><td></td><td align="left">0,268</td><td align="left">1,005</td></tr><tr><td align="left">14801</td><td align="left">59327</td><td align="left">breast</td><td align="left">1</td><td align="left">7</td><td align="left">1</td><td align="left">1</td><td align="left">4</td><td align="left">7</td><td align="left">7</td><td align="left">4</td><td align="left">5</td><td align="left">11</td><td align="left">2</td><td align="left">1</td><td align="left">3</td><td align="left">11</td><td align="left">65</td><td></td><td align="left">0,268</td><td align="left">1,005</td></tr><tr><td align="left">14757</td><td align="left">79152</td><td align="left">breast</td><td align="left">2</td><td align="left">4</td><td align="left">6</td><td align="left">6</td><td align="left">0</td><td align="left">8</td><td align="left">6</td><td align="left">12</td><td align="left">2</td><td align="left">2</td><td align="left">0</td><td align="left">0</td><td align="left">9</td><td align="left">12</td><td align="left">69</td><td></td><td align="left">0,316</td><td align="left">1,653</td></tr><tr><td align="left">14798</td><td align="left">78288</td><td align="left">liver</td><td align="left">8</td><td align="left">12</td><td align="left">9</td><td align="left">6</td><td align="left">1</td><td align="left">17</td><td align="left">5</td><td align="left">10</td><td align="left">7</td><td align="left">9</td><td align="left">6</td><td align="left">3</td><td align="left">15</td><td align="left">17</td><td align="left">125</td><td></td><td align="left">0,260</td><td align="left">1,417</td></tr><tr><td align="left">14749</td><td align="left">89265</td><td align="left">placenta</td><td align="left">2</td><td align="left">23</td><td align="left">6</td><td align="left">8</td><td align="left">0</td><td align="left">27</td><td align="left">5</td><td align="left">22</td><td align="left">6</td><td align="left">14</td><td align="left">3</td><td align="left">4</td><td align="left">21</td><td align="left">29</td><td align="left">170</td><td></td><td align="left">0,299</td><td align="left">1,530</td></tr><tr><td align="left">14750</td><td align="left">118083</td><td align="left">placenta</td><td align="left">2</td><td align="left">20</td><td align="left">6</td><td align="left">15</td><td align="left">1</td><td align="left">26</td><td align="left">2</td><td align="left">15</td><td align="left">9</td><td align="left">12</td><td align="left">3</td><td align="left">4</td><td align="left">18</td><td align="left">22</td><td align="left">155</td><td></td><td align="left">0,332</td><td align="left">1,456</td></tr><tr><td align="left">14780</td><td align="left">26653</td><td align="left">stomach</td><td align="left">0</td><td align="left">6</td><td align="left">1</td><td align="left">8</td><td align="left">1</td><td align="left">7</td><td align="left">14</td><td align="left">3</td><td align="left">1</td><td align="left">5</td><td align="left">0</td><td align="left">0</td><td align="left">5</td><td align="left">9</td><td align="left">60</td><td></td><td align="left">0,272</td><td align="left">1,730</td></tr><tr><td align="left">14771</td><td align="left">101677</td><td align="left">placenta</td><td align="left">2</td><td align="left">5</td><td align="left">4</td><td align="left">5</td><td align="left">0</td><td align="left">12</td><td align="left">0</td><td align="left">5</td><td align="left">5</td><td align="left">10</td><td align="left">0</td><td align="left">1</td><td align="left">1</td><td align="left">14</td><td align="left">64</td><td></td><td align="left">0,331</td><td align="left">1,354</td></tr><tr><td align="left">7498</td><td align="left">31538</td><td align="left">brain</td><td align="left">6</td><td align="left">2</td><td align="left">1</td><td align="left">4</td><td align="left">1</td><td align="left">4</td><td align="left">4</td><td align="left">0</td><td align="left">1</td><td align="left">1</td><td align="left">3</td><td align="left">0</td><td align="left">2</td><td align="left">4</td><td align="left">33</td><td></td><td align="left">0,321</td><td align="left">1,415</td></tr><tr><td align="left" colspan="21">68 additional tumor libraries</td></tr><tr><td align="left">14763</td><td align="left">106982</td><td align="left">brain</td><td align="left">8</td><td align="left">15</td><td align="left">6</td><td align="left">7</td><td align="left">6</td><td align="left">16</td><td align="left">7</td><td align="left">9</td><td align="left">3</td><td align="left">13</td><td align="left">2</td><td align="left">12</td><td align="left">16</td><td align="left">22</td><td align="left">142</td><td></td><td align="left">0,231</td><td align="left">1,515</td></tr><tr><td align="left">14737</td><td align="left">105764</td><td align="left">brain</td><td align="left">4</td><td align="left">10</td><td align="left">7</td><td align="left">14</td><td align="left">10</td><td align="left">29</td><td align="left">2</td><td align="left">8</td><td align="left">8</td><td align="left">22</td><td align="left">3</td><td align="left">6</td><td align="left">13</td><td align="left">20</td><td align="left">156</td><td></td><td align="left">0,272</td><td align="left">1,458</td></tr><tr><td align="left">14765</td><td align="left">102439</td><td align="left">brain</td><td align="left">1</td><td align="left">14</td><td align="left">1</td><td align="left">6</td><td align="left">3</td><td align="left">14</td><td align="left">5</td><td align="left">11</td><td align="left">6</td><td align="left">14</td><td align="left">1</td><td align="left">6</td><td align="left">11</td><td align="left">16</td><td align="left">109</td><td></td><td align="left">0,281</td><td align="left">1,001</td></tr><tr><td align="left">14739</td><td align="left">88568</td><td align="left">brain</td><td align="left">3</td><td align="left">4</td><td align="left">5</td><td align="left">7</td><td align="left">4</td><td align="left">17</td><td align="left">7</td><td align="left">10</td><td align="left">4</td><td align="left">14</td><td align="left">5</td><td align="left">1</td><td align="left">8</td><td align="left">23</td><td align="left">112</td><td></td><td align="left">0,212</td><td align="left">0,983</td></tr><tr><td align="left">14773</td><td align="left">118733</td><td align="left">brain</td><td align="left">10</td><td align="left">18</td><td align="left">10</td><td align="left">16</td><td align="left">12</td><td align="left">28</td><td align="left">0</td><td align="left">8</td><td align="left">6</td><td align="left">17</td><td align="left">4</td><td align="left">6</td><td align="left">12</td><td align="left">25</td><td align="left">172</td><td></td><td align="left">0,182</td><td align="left">1,320</td></tr><tr><td align="left">14766</td><td align="left">107344</td><td align="left">brain</td><td align="left">5</td><td align="left">35</td><td align="left">2</td><td align="left">9</td><td align="left">7</td><td align="left">29</td><td align="left">0</td><td align="left">6</td><td align="left">14</td><td align="left">31</td><td align="left">2</td><td align="left">2</td><td align="left">21</td><td align="left">61</td><td align="left">224</td><td></td><td align="left">0,398</td><td align="left">2,380</td></tr><tr><td align="left">1732</td><td align="left">81495</td><td align="left">brain</td><td align="left">0</td><td align="left">18</td><td align="left">3</td><td align="left">0</td><td align="left">1</td><td align="left">28</td><td align="left">5</td><td align="left">6</td><td align="left">3</td><td align="left">7</td><td align="left">1</td><td align="left">1</td><td align="left">6</td><td align="left">20</td><td align="left">99</td><td></td><td align="left">0,303</td><td align="left">1,594</td></tr><tr><td align="left">14753</td><td align="left">49794</td><td align="left">breast</td><td align="left">0</td><td align="left">3</td><td align="left">2</td><td align="left">2</td><td align="left">1</td><td align="left">19</td><td align="left">6</td><td align="left">6</td><td align="left">2</td><td align="left">3</td><td align="left">2</td><td align="left">0</td><td align="left">4</td><td align="left">5</td><td align="left">55</td><td></td><td align="left">0,237</td><td align="left">1,476</td></tr><tr><td align="left">14748</td><td align="left">65314</td><td align="left">breast</td><td align="left">6</td><td align="left">4</td><td align="left">2</td><td align="left">2</td><td align="left">2</td><td align="left">7</td><td align="left">8</td><td align="left">8</td><td align="left">2</td><td align="left">7</td><td align="left">7</td><td align="left">1</td><td align="left">7</td><td align="left">5</td><td align="left">68</td><td></td><td align="left">0,248</td><td align="left">1,307</td></tr><tr><td align="left">14743</td><td align="left">72857</td><td align="left">breast</td><td align="left">8</td><td align="left">1</td><td align="left">0</td><td align="left">3</td><td align="left">1</td><td align="left">10</td><td align="left">4</td><td align="left">8</td><td align="left">1</td><td align="left">9</td><td align="left">3</td><td align="left">0</td><td align="left">5</td><td align="left">5</td><td align="left">58</td><td></td><td align="left">0,301</td><td align="left">1,354</td></tr><tr><td align="left">14745</td><td align="left">81452</td><td align="left">breast</td><td align="left">1</td><td align="left">8</td><td align="left">1</td><td align="left">1</td><td align="left">1</td><td align="left">8</td><td align="left">1</td><td align="left">15</td><td align="left">2</td><td align="left">8</td><td align="left">8</td><td align="left">0</td><td align="left">7</td><td align="left">4</td><td align="left">65</td><td></td><td align="left">0,360</td><td align="left">1,825</td></tr><tr><td align="left">14747</td><td align="left">37435</td><td align="left">breast</td><td align="left">10</td><td align="left">2</td><td align="left">0</td><td align="left">2</td><td align="left">0</td><td align="left">6</td><td align="left">2</td><td align="left">8</td><td align="left">0</td><td align="left">0</td><td align="left">1</td><td align="left">0</td><td align="left">6</td><td align="left">0</td><td align="left">37</td><td></td><td align="left">0,382</td><td align="left">2,252</td></tr><tr><td align="left">14746</td><td align="left">89184</td><td align="left">breast</td><td align="left">15</td><td align="left">7</td><td align="left">1</td><td align="left">4</td><td align="left">0</td><td align="left">29</td><td align="left">45</td><td align="left">42</td><td align="left">7</td><td align="left">16</td><td align="left">12</td><td align="left">4</td><td align="left">6</td><td align="left">9</td><td align="left">197</td><td></td><td align="left">0,409</td><td align="left">3,547</td></tr><tr><td align="left">14800</td><td align="left">50875</td><td align="left">breast</td><td align="left">3</td><td align="left">1</td><td align="left">1</td><td align="left">4</td><td align="left">2</td><td align="left">1</td><td align="left">1</td><td align="left">3</td><td align="left">6</td><td align="left">1</td><td align="left">1</td><td align="left">1</td><td align="left">1</td><td align="left">1</td><td align="left">27</td><td></td><td align="left">0,413</td><td align="left">1,606</td></tr><tr><td align="left">14797</td><td align="left">21951</td><td align="left">breast</td><td align="left">2</td><td align="left">5</td><td align="left">0</td><td align="left">0</td><td align="left">3</td><td align="left">3</td><td align="left">3</td><td align="left">2</td><td align="left">4</td><td align="left">0</td><td align="left">0</td><td align="left">0</td><td align="left">0</td><td align="left">5</td><td align="left">27</td><td></td><td align="left">0,297</td><td align="left">1,378</td></tr><tr><td align="left">2383</td><td align="left">61480</td><td align="left">breast</td><td align="left">3</td><td align="left">11</td><td align="left">1</td><td align="left">14</td><td align="left">2</td><td align="left">19</td><td align="left">2</td><td align="left">7</td><td align="left">4</td><td align="left">9</td><td align="left">1</td><td align="left">1</td><td align="left">6</td><td align="left">13</td><td align="left">93</td><td></td><td align="left">0,246</td><td align="left">1,248</td></tr><tr><td align="left">1733</td><td align="left">70099</td><td align="left">breast</td><td align="left">1</td><td align="left">8</td><td align="left">2</td><td align="left">10</td><td align="left">1</td><td align="left">16</td><td align="left">4</td><td align="left">6</td><td align="left">6</td><td align="left">13</td><td align="left">1</td><td align="left">2</td><td align="left">6</td><td align="left">15</td><td align="left">91</td><td></td><td align="left">0,262</td><td align="left">0,968</td></tr><tr><td align="left">2382</td><td align="left">65045</td><td align="left">breast</td><td align="left">8</td><td align="left">8</td><td align="left">2</td><td align="left">12</td><td align="left">0</td><td align="left">6</td><td align="left">9</td><td align="left">10</td><td align="left">13</td><td align="left">12</td><td align="left">1</td><td align="left">3</td><td align="left">11</td><td align="left">20</td><td align="left">115</td><td></td><td align="left">0,294</td><td align="left">1,555</td></tr><tr><td align="left">2389</td><td align="left">58801</td><td align="left">breast</td><td align="left">3</td><td align="left">1</td><td align="left">0</td><td align="left">0</td><td align="left">0</td><td align="left">2</td><td align="left">1</td><td align="left">3</td><td align="left">0</td><td align="left">0</td><td align="left">0</td><td align="left">0</td><td align="left">1</td><td align="left">1</td><td align="left">12</td><td></td><td align="left">0,289</td><td align="left">1,204</td></tr><tr><td align="left">1730</td><td align="left">61367</td><td align="left">breast</td><td align="left">3</td><td align="left">12</td><td align="left">0</td><td align="left">7</td><td align="left">2</td><td align="left">9</td><td align="left">7</td><td align="left">3</td><td align="left">1</td><td align="left">6</td><td align="left">5</td><td align="left">3</td><td align="left">3</td><td align="left">13</td><td align="left">74</td><td></td><td align="left">0,237</td><td align="left">1,108</td></tr><tr><td align="left">1731</td><td align="left">43902</td><td align="left">breast</td><td align="left">1</td><td align="left">8</td><td align="left">4</td><td align="left">4</td><td align="left">0</td><td align="left">15</td><td align="left">1</td><td align="left">4</td><td align="left">0</td><td align="left">9</td><td align="left">1</td><td align="left">1</td><td align="left">2</td><td align="left">16</td><td align="left">66</td><td></td><td align="left">0,323</td><td align="left">1,399</td></tr><tr><td align="left">14776</td><td align="left">75379</td><td align="left">brain</td><td align="left">3</td><td align="left">16</td><td align="left">4</td><td align="left">9</td><td align="left">4</td><td align="left">11</td><td align="left">0</td><td align="left">10</td><td align="left">7</td><td align="left">7</td><td align="left">4</td><td align="left">0</td><td align="left">9</td><td align="left">13</td><td align="left">97</td><td></td><td align="left">0,293</td><td align="left">1,295</td></tr><tr><td align="left">1735</td><td align="left">74499</td><td align="left">brain</td><td align="left">1</td><td align="left">4</td><td align="left">4</td><td align="left">4</td><td align="left">6</td><td align="left">18</td><td align="left">0</td><td align="left">9</td><td align="left">2</td><td align="left">5</td><td align="left">0</td><td align="left">1</td><td align="left">7</td><td align="left">6</td><td align="left">67</td><td></td><td align="left">0,329</td><td align="left">1,590</td></tr><tr><td align="left">2408</td><td align="left">52659</td><td align="left">brain</td><td align="left">5</td><td align="left">13</td><td align="left">2</td><td align="left">10</td><td align="left">5</td><td align="left">15</td><td align="left">0</td><td align="left">14</td><td align="left">5</td><td align="left">7</td><td align="left">1</td><td align="left">3</td><td align="left">6</td><td align="left">15</td><td align="left">101</td><td></td><td align="left">0,288</td><td align="left">1,135</td></tr><tr><td align="left">2384</td><td align="left">52934</td><td align="left">brain</td><td align="left">5</td><td align="left">7</td><td align="left">5</td><td align="left">4</td><td align="left">5</td><td align="left">8</td><td align="left">4</td><td align="left">14</td><td align="left">3</td><td align="left">14</td><td align="left">0</td><td align="left">1</td><td align="left">10</td><td align="left">16</td><td align="left">96</td><td></td><td align="left">0,252</td><td align="left">1,220</td></tr><tr><td align="left">14762</td><td align="left">68614</td><td align="left">brain</td><td align="left">3</td><td align="left">11</td><td align="left">5</td><td align="left">11</td><td align="left">4</td><td align="left">14</td><td align="left">0</td><td align="left">10</td><td align="left">6</td><td align="left">10</td><td align="left">1</td><td align="left">4</td><td align="left">7</td><td align="left">10</td><td align="left">96</td><td></td><td align="left">0,307</td><td align="left">1,220</td></tr><tr><td align="left">14740</td><td align="left">122690</td><td align="left">brain</td><td align="left">10</td><td align="left">18</td><td align="left">4</td><td align="left">4</td><td align="left">8</td><td align="left">28</td><td align="left">1</td><td align="left">19</td><td align="left">5</td><td align="left">8</td><td align="left">13</td><td align="left">0</td><td align="left">23</td><td align="left">16</td><td align="left">157</td><td></td><td align="left">0,388</td><td align="left">2,004</td></tr><tr><td align="left">14786</td><td align="left">84073</td><td align="left">brain</td><td align="left">2</td><td align="left">17</td><td align="left">1</td><td align="left">7</td><td align="left">10</td><td align="left">11</td><td align="left">0</td><td align="left">11</td><td align="left">3</td><td align="left">6</td><td align="left">0</td><td align="left">2</td><td align="left">4</td><td align="left">12</td><td align="left">86</td><td></td><td align="left">0,320</td><td align="left">1,766</td></tr><tr><td align="left">14741</td><td align="left">120431</td><td align="left">brain</td><td align="left">14</td><td align="left">12</td><td align="left">2</td><td align="left">6</td><td align="left">1</td><td align="left">21</td><td align="left">2</td><td align="left">14</td><td align="left">7</td><td align="left">7</td><td align="left">3</td><td align="left">4</td><td align="left">15</td><td align="left">11</td><td align="left">119</td><td></td><td align="left">0,290</td><td align="left">1,603</td></tr><tr><td align="left">9103</td><td align="left">11582</td><td align="left">stomach</td><td align="left">1</td><td align="left">3</td><td align="left">1</td><td align="left">3</td><td align="left">0</td><td align="left">3</td><td align="left">6</td><td align="left">0</td><td align="left">1</td><td align="left">1</td><td align="left">0</td><td align="left">1</td><td align="left">0</td><td align="left">3</td><td align="left">23</td><td></td><td align="left">0,266</td><td align="left">1,300</td></tr><tr><td align="left">758</td><td align="left">70433</td><td align="left">stomach</td><td align="left">4</td><td align="left">8</td><td align="left">2</td><td align="left">5</td><td align="left">7</td><td align="left">14</td><td align="left">20</td><td align="left">11</td><td align="left">6</td><td align="left">15</td><td align="left">5</td><td align="left">0</td><td align="left">6</td><td align="left">11</td><td align="left">114</td><td></td><td align="left">0,220</td><td align="left">1,569</td></tr><tr><td align="left">2385</td><td align="left">64102</td><td align="left">stomach</td><td align="left">8</td><td align="left">4</td><td align="left">1</td><td align="left">12</td><td align="left">6</td><td align="left">13</td><td align="left">31</td><td align="left">13</td><td align="left">8</td><td align="left">8</td><td align="left">0</td><td align="left">0</td><td align="left">8</td><td align="left">14</td><td align="left">126</td><td></td><td align="left">0,316</td><td align="left">2,559</td></tr><tr><td align="left">757</td><td align="left">66032</td><td align="left">stomach</td><td align="left">9</td><td align="left">10</td><td align="left">3</td><td align="left">11</td><td align="left">6</td><td align="left">13</td><td align="left">21</td><td align="left">9</td><td align="left">11</td><td align="left">23</td><td align="left">5</td><td align="left">1</td><td align="left">16</td><td align="left">22</td><td align="left">160</td><td></td><td align="left">0,229</td><td align="left">1,412</td></tr><tr><td align="left">9104</td><td align="left">15382</td><td align="left">stomach</td><td align="left">0</td><td align="left">5</td><td align="left">1</td><td align="left">1</td><td align="left">6</td><td align="left">9</td><td align="left">17</td><td align="left">1</td><td align="left">1</td><td align="left">2</td><td align="left">2</td><td align="left">0</td><td align="left">0</td><td align="left">1</td><td align="left">46</td><td></td><td align="left">0,450</td><td align="left">2,707</td></tr><tr><td align="left">14767</td><td align="left">100600</td><td align="left">brain</td><td align="left">3</td><td align="left">12</td><td align="left">1</td><td align="left">8</td><td align="left">5</td><td align="left">22</td><td align="left">2</td><td align="left">15</td><td align="left">11</td><td align="left">12</td><td align="left">2</td><td align="left">6</td><td align="left">13</td><td align="left">20</td><td align="left">132</td><td></td><td align="left">0,267</td><td align="left">1,209</td></tr><tr><td align="left">14768</td><td align="left">102322</td><td align="left">brain</td><td align="left">2</td><td align="left">2</td><td align="left">2</td><td align="left">4</td><td align="left">6</td><td align="left">24</td><td align="left">0</td><td align="left">10</td><td align="left">0</td><td align="left">12</td><td align="left">1</td><td align="left">0</td><td align="left">4</td><td align="left">19</td><td align="left">86</td><td></td><td align="left">0,351</td><td align="left">1,738</td></tr><tr><td align="left">14769</td><td align="left">99099</td><td align="left">brain</td><td align="left">7</td><td align="left">13</td><td align="left">4</td><td align="left">16</td><td align="left">9</td><td align="left">43</td><td align="left">2</td><td align="left">10</td><td align="left">10</td><td align="left">13</td><td align="left">9</td><td align="left">5</td><td align="left">16</td><td align="left">25</td><td align="left">182</td><td></td><td align="left">0,253</td><td align="left">1,579</td></tr><tr><td align="left">1498</td><td align="left">62675</td><td align="left">brain</td><td align="left">9</td><td align="left">11</td><td align="left">3</td><td align="left">8</td><td align="left">3</td><td align="left">33</td><td align="left">2</td><td align="left">14</td><td align="left">2</td><td align="left">17</td><td align="left">0</td><td align="left">4</td><td align="left">8</td><td align="left">13</td><td align="left">127</td><td></td><td align="left">0,281</td><td align="left">1,527</td></tr><tr><td align="left">14807</td><td align="left">86887</td><td align="left">lung</td><td align="left">21</td><td align="left">12</td><td align="left">4</td><td align="left">5</td><td align="left">3</td><td align="left">18</td><td align="left">23</td><td align="left">35</td><td align="left">9</td><td align="left">4</td><td align="left">13</td><td align="left">0</td><td align="left">14</td><td align="left">18</td><td align="left">179</td><td></td><td align="left">0,312</td><td align="left">2,505</td></tr><tr><td align="left">14806</td><td align="left">35916</td><td align="left">lung</td><td align="left">17</td><td align="left">4</td><td align="left">0</td><td align="left">0</td><td align="left">3</td><td align="left">6</td><td align="left">8</td><td align="left">14</td><td align="left">2</td><td align="left">6</td><td align="left">4</td><td align="left">0</td><td align="left">6</td><td align="left">4</td><td align="left">74</td><td></td><td align="left">0,351</td><td align="left">2,516</td></tr><tr><td align="left">14731</td><td align="left">52645</td><td align="left">brain</td><td align="left">1</td><td align="left">5</td><td align="left">0</td><td align="left">11</td><td align="left">1</td><td align="left">15</td><td align="left">5</td><td align="left">9</td><td align="left">4</td><td align="left">7</td><td align="left">0</td><td align="left">1</td><td align="left">7</td><td align="left">10</td><td align="left">76</td><td></td><td align="left">0,254</td><td align="left">1,186</td></tr><tr><td align="left">14788</td><td align="left">74295</td><td align="left">brain</td><td align="left">6</td><td align="left">10</td><td align="left">3</td><td align="left">6</td><td align="left">6</td><td align="left">11</td><td align="left">3</td><td align="left">6</td><td align="left">1</td><td align="left">11</td><td align="left">7</td><td align="left">3</td><td align="left">11</td><td align="left">19</td><td align="left">103</td><td></td><td align="left">0,247</td><td align="left">1,067</td></tr><tr><td align="left">14791</td><td align="left">32570</td><td align="left">brain</td><td align="left">2</td><td align="left">5</td><td align="left">1</td><td align="left">3</td><td align="left">1</td><td align="left">5</td><td align="left">0</td><td align="left">4</td><td align="left">1</td><td align="left">2</td><td align="left">1</td><td align="left">2</td><td align="left">1</td><td align="left">5</td><td align="left">33</td><td></td><td align="left">0,217</td><td align="left">0,758</td></tr><tr><td align="left">14793</td><td align="left">89258</td><td align="left">brain</td><td align="left">4</td><td align="left">25</td><td align="left">8</td><td align="left">14</td><td align="left">4</td><td align="left">29</td><td align="left">5</td><td align="left">10</td><td align="left">2</td><td align="left">20</td><td align="left">4</td><td align="left">10</td><td align="left">8</td><td align="left">33</td><td align="left">176</td><td></td><td align="left">0,265</td><td align="left">1,603</td></tr><tr><td align="left">14774</td><td align="left">85984</td><td align="left">brain</td><td align="left">7</td><td align="left">13</td><td align="left">3</td><td align="left">18</td><td align="left">4</td><td align="left">28</td><td align="left">10</td><td align="left">14</td><td align="left">4</td><td align="left">15</td><td align="left">5</td><td align="left">7</td><td align="left">8</td><td align="left">37</td><td align="left">173</td><td></td><td align="left">0,144</td><td align="left">1,154</td></tr><tr><td align="left">14790</td><td align="left">45342</td><td align="left">brain</td><td align="left">4</td><td align="left">10</td><td align="left">0</td><td align="left">10</td><td align="left">3</td><td align="left">33</td><td align="left">3</td><td align="left">16</td><td align="left">4</td><td align="left">12</td><td align="left">2</td><td align="left">0</td><td align="left">14</td><td align="left">16</td><td align="left">127</td><td></td><td align="left">0,244</td><td align="left">1,521</td></tr><tr><td align="left">14794</td><td align="left">74612</td><td align="left">brain</td><td align="left">3</td><td align="left">11</td><td align="left">3</td><td align="left">1</td><td align="left">6</td><td align="left">13</td><td align="left">6</td><td align="left">9</td><td align="left">2</td><td align="left">21</td><td align="left">5</td><td align="left">6</td><td align="left">0</td><td align="left">14</td><td align="left">100</td><td></td><td align="left">0,375</td><td align="left">1,522</td></tr><tr><td align="left">14781</td><td align="left">83671</td><td align="left">brain</td><td align="left">2</td><td align="left">8</td><td align="left">1</td><td align="left">2</td><td align="left">4</td><td align="left">10</td><td align="left">6</td><td align="left">11</td><td align="left">5</td><td align="left">25</td><td align="left">8</td><td align="left">16</td><td align="left">22</td><td align="left">30</td><td align="left">150</td><td></td><td align="left">0,356</td><td align="left">2,407</td></tr><tr><td align="left">14782</td><td align="left">68392</td><td align="left">brain</td><td align="left">1</td><td align="left">4</td><td align="left">3</td><td align="left">3</td><td align="left">4</td><td align="left">18</td><td align="left">4</td><td align="left">20</td><td align="left">4</td><td align="left">37</td><td align="left">10</td><td align="left">7</td><td align="left">8</td><td align="left">21</td><td align="left">144</td><td></td><td align="left">0,346</td><td align="left">2,142</td></tr><tr><td align="left">14792</td><td align="left">59498</td><td align="left">brain</td><td align="left">2</td><td align="left">10</td><td align="left">1</td><td align="left">9</td><td align="left">7</td><td align="left">8</td><td align="left">3</td><td align="left">5</td><td align="left">0</td><td align="left">3</td><td align="left">0</td><td align="left">0</td><td align="left">2</td><td align="left">23</td><td align="left">73</td><td></td><td align="left">0,308</td><td align="left">1,796</td></tr><tr><td align="left">14783</td><td align="left">61853</td><td align="left">brain</td><td align="left">0</td><td align="left">12</td><td align="left">0</td><td align="left">7</td><td align="left">1</td><td align="left">8</td><td align="left">5</td><td align="left">3</td><td align="left">4</td><td align="left">2</td><td align="left">0</td><td align="left">1</td><td align="left">2</td><td align="left">29</td><td align="left">74</td><td></td><td align="left">0,324</td><td align="left">2,003</td></tr><tr><td align="left">14779</td><td align="left">72318</td><td align="left">brain</td><td align="left">1</td><td align="left">16</td><td align="left">0</td><td align="left">5</td><td align="left">2</td><td align="left">12</td><td align="left">1</td><td align="left">4</td><td align="left">8</td><td align="left">15</td><td align="left">1</td><td align="left">0</td><td align="left">5</td><td align="left">22</td><td align="left">92</td><td></td><td align="left">0,337</td><td align="left">1,588</td></tr><tr><td align="left">14772</td><td align="left">60454</td><td align="left">brain</td><td align="left">5</td><td align="left">10</td><td align="left">3</td><td align="left">4</td><td align="left">0</td><td align="left">10</td><td align="left">0</td><td align="left">9</td><td align="left">3</td><td align="left">6</td><td align="left">0</td><td align="left">2</td><td align="left">3</td><td align="left">13</td><td align="left">68</td><td></td><td align="left">0,270</td><td align="left">1,155</td></tr><tr><td align="left">14787</td><td align="left">57469</td><td align="left">brain</td><td align="left">2</td><td align="left">17</td><td align="left">2</td><td align="left">6</td><td align="left">4</td><td align="left">20</td><td align="left">0</td><td align="left">14</td><td align="left">8</td><td align="left">17</td><td align="left">2</td><td align="left">4</td><td align="left">5</td><td align="left">27</td><td align="left">128</td><td></td><td align="left">0,307</td><td align="left">1,334</td></tr><tr><td align="left">14795</td><td align="left">67404</td><td align="left">brain</td><td align="left">1</td><td align="left">9</td><td align="left">0</td><td align="left">3</td><td align="left">3</td><td align="left">8</td><td align="left">0</td><td align="left">17</td><td align="left">5</td><td align="left">4</td><td align="left">0</td><td align="left">3</td><td align="left">2</td><td align="left">6</td><td align="left">61</td><td></td><td align="left">0,315</td><td align="left">1,793</td></tr><tr><td align="left">14761</td><td align="left">85376</td><td align="left">brain</td><td align="left">1</td><td align="left">31</td><td align="left">3</td><td align="left">16</td><td align="left">3</td><td align="left">31</td><td align="left">1</td><td align="left">3</td><td align="left">7</td><td align="left">5</td><td align="left">2</td><td align="left">5</td><td align="left">3</td><td align="left">86</td><td align="left">197</td><td></td><td align="left">0,475</td><td align="left">3,693</td></tr><tr><td align="left">14734</td><td align="left">69971</td><td align="left">brain</td><td align="left">2</td><td align="left">12</td><td align="left">3</td><td align="left">10</td><td align="left">3</td><td align="left">19</td><td align="left">4</td><td align="left">7</td><td align="left">2</td><td align="left">11</td><td align="left">0</td><td align="left">2</td><td align="left">6</td><td align="left">20</td><td align="left">101</td><td></td><td align="left">0,217</td><td align="left">1,043</td></tr><tr><td align="left">14733</td><td align="left">43068</td><td align="left">brain</td><td align="left">0</td><td align="left">14</td><td align="left">1</td><td align="left">9</td><td align="left">1</td><td align="left">8</td><td align="left">2</td><td align="left">5</td><td align="left">4</td><td align="left">8</td><td align="left">1</td><td align="left">2</td><td align="left">3</td><td align="left">7</td><td align="left">65</td><td></td><td align="left">0,252</td><td align="left">1,327</td></tr><tr><td align="left">14732</td><td align="left">48451</td><td align="left">brain</td><td align="left">5</td><td align="left">2</td><td align="left">0</td><td align="left">2</td><td align="left">4</td><td align="left">9</td><td align="left">1</td><td align="left">15</td><td align="left">6</td><td align="left">12</td><td align="left">0</td><td align="left">5</td><td align="left">4</td><td align="left">11</td><td align="left">76</td><td></td><td align="left">0,345</td><td align="left">1,587</td></tr><tr><td align="left">14775</td><td align="left">41338</td><td align="left">brain</td><td align="left">4</td><td align="left">4</td><td align="left">0</td><td align="left">1</td><td align="left">1</td><td align="left">1</td><td align="left">27</td><td align="left">1</td><td align="left">15</td><td align="left">4</td><td align="left">0</td><td align="left">2</td><td align="left">9</td><td align="left">29</td><td align="left">98</td><td></td><td align="left">0,530</td><td align="left">3,340</td></tr><tr><td align="left">14742</td><td align="left">32442</td><td align="left">brain</td><td align="left">2</td><td align="left">2</td><td align="left">1</td><td align="left">1</td><td align="left">0</td><td align="left">1</td><td align="left">1</td><td align="left">4</td><td align="left">1</td><td align="left">3</td><td align="left">1</td><td align="left">0</td><td align="left">2</td><td align="left">7</td><td align="left">26</td><td></td><td align="left">0,231</td><td align="left">0,823</td></tr><tr><td align="left">743</td><td align="left">33957</td><td align="left">pancreas</td><td align="left">3</td><td align="left">2</td><td align="left">1</td><td align="left">0</td><td align="left">1</td><td align="left">3</td><td align="left">4</td><td align="left">1</td><td align="left">3</td><td align="left">6</td><td align="left">3</td><td align="left">1</td><td align="left">2</td><td align="left">0</td><td align="left">30</td><td></td><td align="left">0,329</td><td align="left">1,241</td></tr><tr><td align="left">744</td><td align="left">35750</td><td align="left">pancreas</td><td align="left">6</td><td align="left">2</td><td align="left">0</td><td align="left">0</td><td align="left">2</td><td align="left">3</td><td align="left">3</td><td align="left">5</td><td align="left">0</td><td align="left">4</td><td align="left">2</td><td align="left">0</td><td align="left">1</td><td align="left">1</td><td align="left">29</td><td></td><td align="left">0,330</td><td align="left">1,507</td></tr><tr><td align="left">7800</td><td align="left">34660</td><td align="left">stomach</td><td align="left">9</td><td align="left">8</td><td align="left">2</td><td align="left">2</td><td align="left">3</td><td align="left">9</td><td align="left">19</td><td align="left">4</td><td align="left">1</td><td align="left">1</td><td align="left">9</td><td align="left">0</td><td align="left">4</td><td align="left">3</td><td align="left">74</td><td></td><td align="left">0,435</td><td align="left">2,559</td></tr><tr><td align="left">8867</td><td align="left">43908</td><td align="left">stomach</td><td align="left">6</td><td align="left">10</td><td align="left">2</td><td align="left">10</td><td align="left">10</td><td align="left">10</td><td align="left">6</td><td align="left">2</td><td align="left">3</td><td align="left">5</td><td align="left">4</td><td align="left">0</td><td align="left">4</td><td align="left">3</td><td align="left">75</td><td></td><td align="left">0,322</td><td align="left">1,822</td></tr><tr><td align="left">8505</td><td align="left">32174</td><td align="left">stomach</td><td align="left">5</td><td align="left">10</td><td align="left">3</td><td align="left">3</td><td align="left">6</td><td align="left">7</td><td align="left">11</td><td align="left">1</td><td align="left">1</td><td align="left">3</td><td align="left">5</td><td align="left">0</td><td align="left">2</td><td align="left">1</td><td align="left">58</td><td></td><td align="left">0,453</td><td align="left">1,949</td></tr><tr><td align="left">14760</td><td align="left">51620</td><td align="left">stomach</td><td align="left">1</td><td align="left">10</td><td align="left">1</td><td align="left">9</td><td align="left">0</td><td align="left">11</td><td align="left">16</td><td align="left">7</td><td align="left">2</td><td align="left">7</td><td align="left">4</td><td align="left">0</td><td align="left">4</td><td align="left">13</td><td align="left">85</td><td></td><td align="left">0,281</td><td align="left">1,570</td></tr><tr><td align="left">1734</td><td align="left">51973</td><td align="left">universal</td><td align="left">3</td><td align="left">11</td><td align="left">4</td><td align="left">24</td><td align="left">4</td><td align="left">20</td><td align="left">15</td><td align="left">16</td><td align="left">7</td><td align="left">18</td><td align="left">2</td><td align="left">3</td><td align="left">7</td><td align="left">40</td><td align="left">174</td><td></td><td align="left">0,221</td><td align="left">1,633</td></tr><tr><td></td><td></td><td align="left">total</td><td align="left">575</td><td align="left">1197</td><td align="left">270</td><td align="left">817</td><td align="left">432</td><td align="left">1737</td><td align="left">822</td><td align="left">1139</td><td align="left">542</td><td align="left">1413</td><td align="left">424</td><td align="left">333</td><td align="left">964</td><td align="left">1859</td><td align="left">12524</td><td align="left">mean</td><td align="left">0,299</td><td align="left">1,633</td></tr><tr><td></td><td></td><td align="left">relative freq.</td><td align="left">0,046</td><td align="left">0,096</td><td align="left">0,022</td><td align="left">0,065</td><td align="left">0,034</td><td align="left">0,139</td><td align="left">0,066</td><td align="left">0,091</td><td align="left">0,043</td><td align="left">0,113</td><td align="left">0,034</td><td align="left">0,027</td><td align="left">0,077</td><td align="left">0,148</td><td align="left">1,000</td><td align="left">stand deviation</td><td align="left">0,066</td><td align="left">0,747</td></tr></tbody></table></table-wrap><p>At the start of this study the GEO database (GPL4) consisted out of 154 libraries, of which 60 met our criteria (derived from tissue biopsies and at least 24 tags). At present 254 libraries are available of which an additional 80 meet our criteria. These extra libraries have therefore been used to check whether our previous findings are reproducible. However, since only 12 of the additional 80 libraries were derived from normal tissues a new calculation was performed using all 140 libraries (the 60 original and 80 new libraries), giving totals of 42 derived from normal tissues and 98 derived from tumor tissues. The observed tag counts are shown in Table <xref ref-type="table" rid="T3">3</xref>. Expected frequencies for the tag counts of the 14 genes of a library were obtained with the overall expression profile derived from the sums of the tag counts in all 140 libraries (Table <xref ref-type="table" rid="T2">2</xref> C). The expected and observed tag counts were again compared by chi-square. A comparison of the chi-squares of the normal and tumor tissues by Wilcoxon's test revealed a much greater significant difference between the two groups then previously observed (P = 1.65 × 10<sup>-8 </sup>instead of P = 1.73 × 10<sup>-6</sup>). This shows that the 80 new libraries display a similar difference between normal and tumor tissues to that observed with the 60 libraries used previously.</p><p>Of these 140 libraries, 30 were derived from breast tissue, 11 of these originating from normal breast. These normal and tumor breast tissues were compared separately to deal with the potential disadvantage posed by tissue heterogeneity. Expected frequencies for the tag counts were again obtained with the overall expression profile derived from the sums of the tag counts in these 30 libraries (Table <xref ref-type="table" rid="T2">2</xref> D). The expected and observed tag counts were again compared by chi-square. Comparison of the chi-squares of the normal and tumor breast tissues by Wilcoxon's test (P = 0,0033) revealed a similar difference between normal and tumor to that found for all tissues.</p></sec><sec><title>Variation in transcription specific or unspecific?</title><p>Comparison of the expression profile of the 37 most orderly libraries with that of the 37 most disorderly libraries shows that these profiles are rather similar (Table <xref ref-type="table" rid="T2">2</xref>E, F). This indicates that excessive variation in transcription is multi-directional and does not lead to a specific and systematic change in the expression profiles of all tissues. However for individual tissues this still could be the case.</p></sec><sec><title>Heritability of variant profiles</title><p>That the observed variation in transcription is not just due to momentary fluctuations in transcription rates follows from an analysis of pairs of libraries present in the database; i.e. pairs from both tumor and normal tissue or from both tumor and metastatic tissue, each pair being derived from the same individual. Of the six available library pairs, four have enough tags for a detailed analysis of deviation in the expression of each individual proteasome gene. By taking the log of the ratio of "observed tags"/"expected tags" for each gene, a profile is obtained that shows the degree of aberration in expression for each proteasome gene. The data presented in Figure <xref ref-type="fig" rid="F1">1A</xref> reveal a high correlation between the abnormal expression pattern of a normal prostate and its tumor (R = 0,77 and P = 0,00135). This indicates that a deviating expression profile can be extremely stable and can be transmitted to a clonal derivative as a constitutive trait.</p><fig position="float" id="F1"><label>Figure 1</label><caption><p>Comparison of deviation profiles of clonal derivatives from normal (A and B) or tumor tissues (C and D). Significance of the correlations: for A: R = 0,77 and P = .0,00135, for B: R = 0,73 and P = 0,00329, for C: R = 0,84 and P = 0,00016 and for D: R = 0,04 and P = 0,887</p></caption><graphic xlink:href="1745-6150-1-9-1"/></fig><p>Essentially the same picture is seen in Figure <xref ref-type="fig" rid="F1">1B</xref>, where a high correlation exists between the deviant profile of another prostate tumor and the normal prostate tissue from which the tumor was derived (R = 0,73 and P = 0,00329). The expression profiles of the two prostates clearly differ from each other, which suggests that a specific expression profile for the prostate does not exist and that these two profiles are deviating dissimilar from the mean expression profile.</p><p>The similarity in deviating expression profiles of a normal tissue and its tumor suggests the possibility of a disorderly condition in the normal tissue being a predisposing factor in the eventual appearance of the tumor.</p><p>The same persistence of a deviating expression profile is observed for the correlation of a breast tumor and its metastasis as shown in Figure <xref ref-type="fig" rid="F1">1C</xref> (R = 0,84 and P = 0,00016).</p><p>Although heritability of deviating profiles is indicated for these three instances, a direct proof for disorder as cause for these deviations is absent.</p></sec><sec><title>Both progression and regression in degree of variation occurs</title><p>Inspection of the three profiles of the clonal derivatives that correlated significantly with the three profiles of the tissues of origin (Figure <xref ref-type="fig" rid="F1">1A, B</xref> and <xref ref-type="fig" rid="F1">1C</xref>) indicates that 11 genes progressed to further deviation and 3 to less variation (only those genes were counted that had a log obs/exp that is larger than 0.5 or -0.5 which represents a 3.2-fold over-expression or under-expression respectively). As this difference is significant (P = 0,032) it suggests that the deviation in expression is progressing in clonal derivatives. The figure further indicates that progression of a deviating expression profile generally occurs in small steps, but that larger jumps may also occur (PSMA4 in Figure <xref ref-type="fig" rid="F1">1A</xref>).</p><p>That progression does not always occur is suggested by Figure <xref ref-type="fig" rid="F1">1D</xref> in which no significant correlation is observed between a tumor and its metastasis (R = 0,04 and P = 0,887) and in which a decrease in deviation is apparent in the metastasis. This figure therefore indicates that the deviating profile does not always persist in a clonal derivative. The metastasis actually exhibits an expression profile that is almost indistinguishable from normal, thereby suggesting that the changes in expression were non-genetic in origin.</p><p>Apparently, spontaneous epigenetic modifications that interfere with normal gene expression patterns can occur. At present, the cause of these modifications remains obscure. Changes in transcription factors, DNA methylation patterns [<xref ref-type="bibr" rid="B14">14</xref>], unusual DNA structures [<xref ref-type="bibr" rid="B15">15</xref>], alterations in nuclear organization [<xref ref-type="bibr" rid="B16">16</xref>], interference by noncoding RNAs [<xref ref-type="bibr" rid="B17">17</xref>] or changes in the macromolecular transcriptional apparatus [<xref ref-type="bibr" rid="B18">18</xref>] might be involved. Possibly all factors known to influence transcription rates could be involved. Among these factors alteration in nuclear organization is very attractive as it easily explains the simultaneous changes in expression in a number of genes and as numerous nuclear constituents could be involved. If so these findings would be of importance to the rapidly developing field of "spatial nuclear organization as a structural component in gene expression" [<xref ref-type="bibr" rid="B19">19</xref>,<xref ref-type="bibr" rid="B20">20</xref>].</p></sec><sec><title>Under-expression and over-expression in deviating libraries</title><p>In the first set of 60 libraries, 15 variant libraries were identified that deviated strongly (P < 0,01) from the overall profile. The changes in the degree of transcription of the individual genes in these libraries were expressed as the log of the ratio of observed and expected tags (Table <xref ref-type="table" rid="T4">4</xref> A). In the second set of 80 libraries, 22 variant libraries were identified (Table <xref ref-type="table" rid="T4">4</xref> B).</p><table-wrap position="float" id="T4"><label>Table 4</label><caption><p>Deviation in transcription (log obs/exp) in the individual genes of the 37 most disorderly libraries.</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left" colspan="17">A. The 15 most disorderly libraries identified in the first set of 60 libraries.</td></tr><tr><td></td><td></td><td align="left" colspan="15">log obs/exp</td></tr><tr><td></td><td align="center">gene</td><td align="center">PSMA1</td><td align="center">PSMA2</td><td align="center">PSMA3</td><td align="center">PSMA4</td><td align="center">PSMA5</td><td align="center">PSMA6</td><td align="center">PSMA7</td><td align="center">PSMB1</td><td align="center">PSMB2</td><td align="center">PSMB3</td><td align="center">PSMB4</td><td align="center">PSMB5</td><td align="center">PSMB6</td><td align="center">PSMB7</td><td align="center">deviation index</td></tr></thead><tbody><tr><td align="center">GSM</td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="center">671</td><td></td><td align="center">-0,954</td><td align="center">-0,288</td><td align="center">-0,552</td><td align="center">0,172</td><td align="center">-0,872</td><td align="center">-0,364</td><td align="center">-0,169</td><td align="center">-0,412</td><td align="center">-0,309</td><td align="center">0,625</td><td align="center">-0,938</td><td align="center">-0,727</td><td align="center">-0,513</td><td align="center">-0,414</td><td align="center">0,431</td></tr><tr><td align="center">686</td><td></td><td align="center">-0,015</td><td align="center">0,048</td><td align="center">-0,216</td><td align="center">-1,115</td><td align="center">-0,137</td><td align="center">-0,115</td><td align="center">0,013</td><td align="center">-0,107</td><td align="center">-0,398</td><td align="center">0,025</td><td align="center">0,731</td><td align="center">0,309</td><td align="center">-0,222</td><td align="center">-0,202</td><td align="center">0,399</td></tr><tr><td align="center">672</td><td></td><td align="center">-0,497</td><td align="center">-0,133</td><td align="center">0,080</td><td align="center">-0,041</td><td align="center">-0,018</td><td align="center">-0,075</td><td align="center">-0,059</td><td align="center">-0,147</td><td align="center">-0,279</td><td align="center">0,489</td><td align="center">-0,783</td><td align="center">-0,571</td><td align="center">-0,154</td><td align="center">-0,481</td><td align="center">0,317</td></tr><tr><td align="center">687</td><td></td><td align="center">-0,171</td><td align="center">0,193</td><td align="center">-0,071</td><td align="center">-0,669</td><td align="center">-0,390</td><td align="center">0,474</td><td align="center">-0,688</td><td align="center">0,401</td><td align="center">-0,430</td><td align="center">-0,977</td><td align="center">-0,457</td><td align="center">-0,245</td><td align="center">-0,254</td><td align="center">-0,234</td><td align="center">0,409</td></tr><tr><td align="center">699</td><td></td><td align="center">-0,485</td><td align="center">-0,297</td><td align="center">-0,084</td><td align="center">-0,682</td><td align="center">0,074</td><td align="center">-0,092</td><td align="center">0,414</td><td align="center">-0,487</td><td align="center">0,160</td><td align="center">-0,989</td><td align="center">0,609</td><td align="center">0,043</td><td align="center">0,257</td><td align="center">0,009</td><td align="center">0,438</td></tr><tr><td align="center">745</td><td></td><td align="center">-0,299</td><td align="center">-0,111</td><td align="center">-0,199</td><td align="center">0,048</td><td align="center">-0,519</td><td align="center">0,410</td><td align="center">-0,038</td><td align="center">-0,426</td><td align="center">-0,558</td><td align="center">-0,406</td><td align="center">-0,585</td><td align="center">-0,373</td><td align="center">0,220</td><td align="center">0,115</td><td align="center">0,311</td></tr><tr><td align="center">786</td><td></td><td align="center">0,032</td><td align="center">-0,132</td><td align="center">-0,044</td><td align="center">-0,040</td><td align="center">-0,364</td><td align="center">0,005</td><td align="center">-0,661</td><td align="center">-0,271</td><td align="center">-0,102</td><td align="center">-0,348</td><td align="center">-0,430</td><td align="center">0,480</td><td align="center">-0,102</td><td align="center">0,474</td><td align="center">0,315</td></tr><tr><td align="center">574</td><td></td><td align="center">-0,171</td><td align="center">0,052</td><td align="center">-0,071</td><td align="center">0,309</td><td align="center">-0,691</td><td align="center">0,005</td><td align="center">-0,143</td><td align="center">-0,377</td><td align="center">-0,731</td><td align="center">-0,074</td><td align="center">-0,059</td><td align="center">-0,245</td><td align="center">-0,129</td><td align="center">0,323</td><td align="center">0,305</td></tr><tr><td align="center">760</td><td></td><td align="center">0,323</td><td align="center">-0,104</td><td align="center">-0,317</td><td align="center">-0,012</td><td align="center">-0,335</td><td align="center">0,261</td><td align="center">0,067</td><td align="center">0,301</td><td align="center">-0,198</td><td align="center">-0,745</td><td align="center">0,075</td><td align="center">-0,491</td><td align="center">0,024</td><td align="center">-0,275</td><td align="center">0,312</td></tr><tr><td align="center">736</td><td></td><td align="center">0,459</td><td align="center">0,135</td><td align="center">0,048</td><td align="center">-0,006</td><td align="center">-0,573</td><td align="center">-0,630</td><td align="center">0,130</td><td align="center">-0,180</td><td align="center">0,087</td><td align="center">-0,460</td><td align="center">0,315</td><td align="center">0,174</td><td align="center">0,233</td><td align="center">-0,074</td><td align="center">0,329</td></tr><tr><td align="center">1497</td><td></td><td align="center">0,275</td><td align="center">0,111</td><td align="center">-0,102</td><td align="center">-0,399</td><td align="center">0,277</td><td align="center">-0,178</td><td align="center">-0,719</td><td align="center">-0,204</td><td align="center">0,493</td><td align="center">-0,054</td><td align="center">-0,488</td><td align="center">0,422</td><td align="center">0,016</td><td align="center">-0,010</td><td align="center">0,347</td></tr><tr><td align="center">762</td><td></td><td align="center">-0,836</td><td align="center">0,051</td><td align="center">0,343</td><td align="center">0,289</td><td align="center">0,090</td><td align="center">0,143</td><td align="center">-0,751</td><td align="center">-0,140</td><td align="center">-0,016</td><td align="center">0,001</td><td align="center">-0,821</td><td align="center">-0,007</td><td align="center">-0,095</td><td align="center">0,065</td><td align="center">0,392</td></tr><tr><td align="center">765</td><td></td><td align="center">0,053</td><td align="center">-0,361</td><td align="center">-0,148</td><td align="center">-0,445</td><td align="center">-0,166</td><td align="center">-0,302</td><td align="center">0,349</td><td align="center">0,102</td><td align="center">0,396</td><td align="center">-0,054</td><td align="center">-0,534</td><td align="center">-0,322</td><td align="center">0,234</td><td align="center">-0,010</td><td align="center">0,291</td></tr><tr><td align="center">740</td><td></td><td align="center">0,000</td><td align="center">0,158</td><td align="center">0,401</td><td align="center">-0,118</td><td align="center">0,337</td><td align="center">0,150</td><td align="center">-0,915</td><td align="center">-0,157</td><td align="center">-0,657</td><td align="center">-0,125</td><td align="center">-0,684</td><td align="center">-0,171</td><td align="center">0,246</td><td align="center">-0,013</td><td align="center">0,395</td></tr><tr><td align="center">739</td><td></td><td align="center">0,198</td><td align="center">-0,025</td><td align="center">0,298</td><td align="center">0,126</td><td align="center">0,346</td><td align="center">-0,130</td><td align="center">-0,796</td><td align="center">-0,106</td><td align="center">-0,061</td><td align="center">-0,044</td><td align="center">-0,565</td><td align="center">0,549</td><td align="center">0,115</td><td align="center">-0,041</td><td align="center">0,347</td></tr><tr><td align="left">mean</td><td></td><td align="center">-0,139</td><td align="center">-0,047</td><td align="center">-0,042</td><td align="center">-0,172</td><td align="center">-0,196</td><td align="center">-0,029</td><td align="center">-0,264</td><td align="center">-0,147</td><td align="center">-0,173</td><td align="center">-0,209</td><td align="center">-0,308</td><td align="center">-0,078</td><td align="center">-0,008</td><td align="center">-0,051</td><td></td></tr><tr><td align="left">standard deviation</td><td></td><td align="center">0,413</td><td align="center">0,175</td><td align="center">0,252</td><td align="center">0,406</td><td align="center">0,380</td><td align="center">0,291</td><td align="center">0,446</td><td align="center">0,254</td><td align="center">0,361</td><td align="center">0,462</td><td align="center">0,522</td><td align="center">0,400</td><td align="center">0,229</td><td align="center">0,252</td><td></td></tr><tr><td align="left" colspan="17">B. The 22 most disorderly libraries identified in the second set of 80 libraries.</td></tr><tr><td></td><td align="center">gene</td><td align="center">PSMA1</td><td align="center">PSMA2</td><td align="center">PSMA3</td><td align="center">PSMA4</td><td align="center">PSMA5</td><td align="center">PSMA6</td><td align="center">PSMA7</td><td align="center">PSMB1</td><td align="center">PSMB2</td><td align="center">PSMB3</td><td align="center">PSMB4</td><td align="center">PSMB5</td><td align="center">PSMB6</td><td align="center">PSMB7</td><td></td></tr><tr><td align="center">GSM</td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="center">14761</td><td></td><td align="center">-0,956</td><td align="center">0,218</td><td align="center">-0,153</td><td align="center">0,086</td><td align="center">-0,353</td><td align="center">0,056</td><td align="center">-1,116</td><td align="center">-0,775</td><td align="center">-0,084</td><td align="center">-0,646</td><td align="center">-0,525</td><td align="center">-0,028</td><td align="center">-0,702</td><td align="center">0,470</td><td align="center">0,475</td></tr><tr><td align="center">14746</td><td></td><td align="center">0,218</td><td align="center">-0,430</td><td align="center">-0,632</td><td align="center">-0,519</td><td align="center">-0,833</td><td align="center">0,025</td><td align="center">0,535</td><td align="center">0,369</td><td align="center">-0,086</td><td align="center">-0,143</td><td align="center">0,251</td><td align="center">-0,127</td><td align="center">-0,403</td><td align="center">-0,512</td><td align="center">0,409</td></tr><tr><td align="center">14799</td><td></td><td align="center">0,364</td><td align="center">0,146</td><td align="center">-0,155</td><td align="center">-0,263</td><td align="center">-0,152</td><td align="center">-0,016</td><td align="center">-0,299</td><td align="center">-0,521</td><td align="center">0,023</td><td align="center">-0,528</td><td align="center">-0,095</td><td align="center">0,568</td><td align="center">0,188</td><td align="center">0,052</td><td align="center">0,310</td></tr><tr><td align="center">14775</td><td></td><td align="center">-0,059</td><td align="center">-0,377</td><td align="center">-0,335</td><td align="center">-0,824</td><td align="center">-0,536</td><td align="center">-1,140</td><td align="center">0,610</td><td align="center">-0,957</td><td align="center">0,541</td><td align="center">-0,449</td><td align="center">-0,531</td><td align="center">-0,131</td><td align="center">0,069</td><td align="center">0,292</td><td align="center">0,530</td></tr><tr><td align="center">7800</td><td></td><td align="center">0,418</td><td align="center">0,049</td><td align="center">0,091</td><td align="center">-0,398</td><td align="center">0,066</td><td align="center">-0,061</td><td align="center">0,582</td><td align="center">-0,230</td><td align="center">-0,510</td><td align="center">-0,926</td><td align="center">0,548</td><td align="center">-0,307</td><td align="center">-0,158</td><td align="center">-0,568</td><td align="center">0,435</td></tr><tr><td align="center">14806</td><td></td><td align="center">0,683</td><td align="center">-0,263</td><td align="center">-0,222</td><td align="center">-0,710</td><td align="center">0,054</td><td align="center">-0,249</td><td align="center">0,195</td><td align="center">0,302</td><td align="center">-0,220</td><td align="center">-0,159</td><td align="center">0,184</td><td align="center">-0,319</td><td align="center">0,007</td><td align="center">-0,454</td><td align="center">0,351</td></tr><tr><td align="center">14807</td><td></td><td align="center">0,396</td><td align="center">-0,164</td><td align="center">0,002</td><td align="center">-0,390</td><td align="center">-0,324</td><td align="center">-0,150</td><td align="center">0,275</td><td align="center">0,322</td><td align="center">0,055</td><td align="center">-0,714</td><td align="center">0,318</td><td align="center">0,002</td><td align="center">-0,003</td><td align="center">-0,179</td><td align="center">0,312</td></tr><tr><td align="center">2385</td><td></td><td align="center">0,131</td><td align="center">-0,487</td><td align="center">-0,446</td><td align="center">0,145</td><td align="center">0,131</td><td align="center">-0,137</td><td align="center">0,559</td><td align="center">0,046</td><td align="center">0,158</td><td align="center">-0,258</td><td align="center">-0,642</td><td align="center">-0,242</td><td align="center">-0,092</td><td align="center">-0,134</td><td align="center">0,316</td></tr><tr><td align="center">14781</td><td></td><td align="center">-0,536</td><td align="center">-0,252</td><td align="center">-0,511</td><td align="center">-0,699</td><td align="center">-0,110</td><td align="center">-0,317</td><td align="center">-0,220</td><td align="center">-0,092</td><td align="center">-0,112</td><td align="center">0,171</td><td align="center">0,196</td><td align="center">0,596</td><td align="center">0,282</td><td align="center">0,131</td><td align="center">0,356</td></tr><tr><td align="center">14766</td><td></td><td align="center">-0,314</td><td align="center">0,213</td><td align="center">-0,386</td><td align="center">-0,222</td><td align="center">-0,043</td><td align="center">-0,030</td><td align="center">-1,174</td><td align="center">-0,531</td><td align="center">0,159</td><td align="center">0,088</td><td align="center">-0,582</td><td align="center">-0,483</td><td align="center">0,085</td><td align="center">0,263</td><td align="center">0,398</td></tr><tr><td align="center">14782</td><td></td><td align="center">-0,819</td><td align="center">-0,535</td><td align="center">-0,016</td><td align="center">-0,505</td><td align="center">-0,092</td><td align="center">-0,044</td><td align="center">-0,378</td><td align="center">0,185</td><td align="center">-0,191</td><td align="center">0,359</td><td align="center">0,311</td><td align="center">0,255</td><td align="center">-0,140</td><td align="center">-0,006</td><td align="center">0,346</td></tr><tr><td align="center">14740</td><td></td><td align="center">0,160</td><td align="center">0,098</td><td align="center">0,088</td><td align="center">-0,401</td><td align="center">0,188</td><td align="center">0,128</td><td align="center">-1,001</td><td align="center">0,143</td><td align="center">-0,115</td><td align="center">-1,230</td><td align="center">0,404</td><td align="center">-0,611</td><td align="center">0,298</td><td align="center">-0,145</td><td align="center">0,493</td></tr><tr><td align="center">14783</td><td></td><td align="center">-0,553</td><td align="center">0,208</td><td align="center">0,074</td><td align="center">0,129</td><td align="center">-0,428</td><td align="center">-0,129</td><td align="center">-0,015</td><td align="center">-0,372</td><td align="center">0,075</td><td align="center">-0,642</td><td align="center">-0,423</td><td align="center">-0,324</td><td align="center">-0,476</td><td align="center">0,400</td><td align="center">0,324</td></tr><tr><td align="center">8867</td><td></td><td align="center">0,236</td><td align="center">0,140</td><td align="center">0,085</td><td align="center">0,295</td><td align="center">0,583</td><td align="center">-0,021</td><td align="center">0,076</td><td align="center">-0,537</td><td align="center">-0,038</td><td align="center">-0,233</td><td align="center">0,190</td><td align="center">-0,313</td><td align="center">-0,163</td><td align="center">-0,574</td><td align="center">0,322</td></tr><tr><td align="center">14745</td><td></td><td align="center">-0,481</td><td align="center">0,105</td><td align="center">-0,155</td><td align="center">-0,643</td><td align="center">-0,356</td><td align="center">-0,057</td><td align="center">-0,641</td><td align="center">0,399</td><td align="center">-0,153</td><td align="center">0,033</td><td align="center">0,552</td><td align="center">-0,252</td><td align="center">0,141</td><td align="center">-0,387</td><td align="center">0,360</td></tr><tr><td align="center">14792</td><td></td><td align="center">-0,241</td><td align="center">0,140</td><td align="center">-0,216</td><td align="center">0,250</td><td align="center">0,428</td><td align="center">-0,118</td><td align="center">-0,226</td><td align="center">-0,139</td><td align="center">-0,515</td><td align="center">-0,455</td><td align="center">-0,412</td><td align="center">-0,313</td><td align="center">-0,465</td><td align="center">0,311</td><td align="center">0,308</td></tr><tr><td align="center">14795</td><td></td><td align="center">-0,467</td><td align="center">0,169</td><td align="center">-0,141</td><td align="center">-0,153</td><td align="center">0,135</td><td align="center">-0,044</td><td align="center">-0,628</td><td align="center">0,467</td><td align="center">0,258</td><td align="center">-0,255</td><td align="center">-0,337</td><td align="center">0,239</td><td align="center">-0,390</td><td align="center">-0,198</td><td align="center">0,315</td></tr><tr><td align="center">14786</td><td></td><td align="center">-0,305</td><td align="center">0,307</td><td align="center">-0,280</td><td align="center">0,077</td><td align="center">0,519</td><td align="center">-0,044</td><td align="center">-0,766</td><td align="center">0,140</td><td align="center">-0,102</td><td align="center">-0,217</td><td align="center">-0,476</td><td align="center">-0,076</td><td align="center">-0,227</td><td align="center">-0,035</td><td align="center">0,320</td></tr><tr><td align="center">14768</td><td></td><td align="center">-0,309</td><td align="center">-0,627</td><td align="center">0,016</td><td align="center">-0,171</td><td align="center">0,293</td><td align="center">0,290</td><td align="center">-0,771</td><td align="center">0,093</td><td align="center">-0,584</td><td align="center">0,079</td><td align="center">-0,480</td><td align="center">-0,382</td><td align="center">-0,232</td><td align="center">0,159</td><td align="center">0,351</td></tr><tr><td align="center">14780</td><td></td><td align="center">-0,460</td><td align="center">0,000</td><td align="center">-0,135</td><td align="center">0,280</td><td align="center">-0,335</td><td align="center">-0,095</td><td align="center">0,525</td><td align="center">-0,279</td><td align="center">-0,434</td><td align="center">-0,151</td><td align="center">-0,330</td><td align="center">-0,231</td><td align="center">0,015</td><td align="center">-0,015</td><td align="center">0,272</td></tr><tr><td align="center">14757</td><td></td><td align="center">-0,217</td><td align="center">-0,234</td><td align="center">0,586</td><td align="center">0,097</td><td align="center">-0,393</td><td align="center">-0,095</td><td align="center">0,099</td><td align="center">0,265</td><td align="center">-0,191</td><td align="center">-0,607</td><td align="center">-0,388</td><td align="center">-0,289</td><td align="center">0,212</td><td align="center">0,052</td><td align="center">0,316</td></tr><tr><td align="center">1734</td><td></td><td align="center">-0,425</td><td align="center">-0,178</td><td align="center">0,026</td><td align="center">0,316</td><td align="center">-0,175</td><td align="center">-0,080</td><td align="center">0,114</td><td align="center">0,006</td><td align="center">-0,030</td><td align="center">-0,036</td><td align="center">-0,471</td><td align="center">-0,196</td><td align="center">-0,280</td><td align="center">0,192</td><td align="center">0,221</td></tr><tr><td align="left">mean</td><td></td><td align="center">-0,161</td><td align="center">-0,080</td><td align="center">-0,128</td><td align="center">-0,192</td><td align="center">-0,079</td><td align="center">-0,106</td><td align="center">-0,167</td><td align="center">-0,077</td><td align="center">-0,095</td><td align="center">-0,314</td><td align="center">-0,124</td><td align="center">-0,135</td><td align="center">-0,111</td><td align="center">-0,040</td><td></td></tr><tr><td align="left">standard deviation</td><td></td><td align="center">0,432</td><td align="center">0,284</td><td align="center">0,260</td><td align="center">0,367</td><td align="center">0,358</td><td align="center">0,261</td><td align="center">0,580</td><td align="center">0,402</td><td align="center">0,265</td><td align="center">0,382</td><td align="center">0,409</td><td align="center">0,308</td><td align="center">0,269</td><td align="center">0,314</td><td></td></tr></tbody></table></table-wrap><p>If one considers the data In Table <xref ref-type="table" rid="T4">4</xref> as a whole, 331 under-expressions are observed against 187 over-expressions. Thus under-expression predominates (P = 2,5 × 10<sup>-10</sup>) in deviating libraries. There is no evidence that differences exist between the genes in their frequencies of under- and over-expression, since the chi-square for heterogeneity is not significant (P = 0,539). However differences between the genes in the degree of under- and over-expressions do exist. The 14 standard deviations of the log obs/exp values, calculated for each individual gene from the 15 most abnormal libraries as shown in Table <xref ref-type="table" rid="T4">4</xref> A, were found to be heterogeneous when compared by the test of Bartlett (P = 0,002). The standard deviations of the 22 most abnormal libraries in Table <xref ref-type="table" rid="T4">4</xref> B correlate with those of Table <xref ref-type="table" rid="T4">4</xref> A (P = 0,022) thus showing a similar pattern of differences between genes. For the whole set of 37 deviating libraries the 3 most variable genes are PSMA7, PSMB4 and PSMA1, while the 3 least variable genes are PSMA2, PSMB6 and PSMA3.</p></sec><sec><title>Epigenetic origin of excessive variation in transcription</title><p>At first sight it does not seem surprising that cancer tissues show greater variation in transcript abundances, as tumor cells are usually aneuploid. However excessive variation in expression profiles is observed in normal tissues as well. This holds not only for the two normal prostates mentioned previously but also for other normal tissues e.g. cortex (GSM 786), normal retina (574), normal breast (760), normal colon (728), normal brain (676) and normal lung (762). Moreover tumors with a normal expression profile are not rare although tumors, as a rule, are known to be aneuploid and carry mutational damage in oncogenes. In addition the similarity in deviating expression profiles of pairs of normal and tumor prostate tissue (supposedly the first being diploid and the second aneuploid) also indicates a non-genetic origin of the deviating profiles Loss of deviation in a metastasis is, similarly, also suggestive of a non-genetic origin. Therefore it would appear that aneuploidy, as such, or mutated oncogenes do not play predominant roles in the emergence of excessive variation in transcription and thus that there is an epigenetic origin.</p></sec><sec><title>An index for the degree of deviation in transcription</title><p>The degree of deviation can be quantified in a deviation index by taking the standard deviation of the log obs/exp values from the 14 genes in a library (log ratio deviation index) or the standard deviation of the z-scores (z-score deviation index). The log ratio deviation index will be suitable to reflect fold-changes in expression, while the z-score deviation index wiil be more suitable to reflect percentual changes in expression.</p><p>If the log ratio deviation index and the z-score deviation index were calculated for all 140 libraries (shown in table <xref ref-type="table" rid="T3">3</xref>) and then normal and tumor libraries were compared with the test of Wilcoxon, tumor libraries were again found to be more deviating than normal libraries. For the log ratio deviation index the significance is considerable (P = 3.6 × 10<sup>-6</sup>) while for the z-score deviation index the significance is much less (P = 0,0177) indicating that the changes in expression reflect fold-changes rather than percentual changes.</p><p>This deviation index provides a means by which to test whether differences in degree of deviation exist between tumors derived from different tissues. To this end a table was prepared with the log ratio deviation index of 5 tumors (astrocytoma, breast cancer, ependymoma, gastric cancer and medulloblastoma), each tumor represented by 9 libraries (Table <xref ref-type="table" rid="T5">5</xref>). When compared with ANOVA (Analysis Of VAriance), the degree of deviation was not influenced by tumor type (P = 0,387). Therefore tumors do not appear to differ systematically in degree of deviation.</p><table-wrap position="float" id="T5"><label>Table 5</label><caption><p>No heterogeneity in deviation index in 5 types of tumors, 9 libraries per type of tumor (P = 0,387).</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td align="center" colspan="9">replication</td></tr><tr><td align="left">tumor type</td><td align="center">1</td><td align="center">2</td><td align="center">3</td><td align="center">4</td><td align="center">5</td><td align="center">6</td><td align="center">7</td><td align="center">8</td><td align="center">9</td></tr></thead><tbody><tr><td align="left">Astrocytoma</td><td align="center">0,231</td><td align="center">0,229</td><td align="center">0,313</td><td align="center">0,272</td><td align="center">0,281</td><td align="center">0,212</td><td align="center">0,182</td><td align="center">0,398</td><td align="center">0,303</td></tr><tr><td align="left">breast carcinoma</td><td align="center">0,248</td><td align="center">0,409</td><td align="center">0,246</td><td align="center">0,262</td><td align="center">0,294</td><td align="center">0,327</td><td align="center">0,375</td><td align="center">0,237</td><td align="center">0,323</td></tr><tr><td align="left">Ependymoma</td><td align="center">0,293</td><td align="center">0,331</td><td align="center">0,288</td><td align="center">0,252</td><td align="center">0,307</td><td align="center">0,388</td><td align="center">0,320</td><td align="center">0,290</td><td align="center">0,281</td></tr><tr><td align="left">gastric cancer</td><td align="center">0,266</td><td align="center">0,220</td><td align="center">0,316</td><td align="center">0,229</td><td align="center">0,450</td><td align="center">0,435</td><td align="center">0,322</td><td align="center">0,453</td><td align="center">0,281</td></tr><tr><td align="left">Medulloblastoma</td><td align="center">0,254</td><td align="center">0,247</td><td align="center">0,265</td><td align="center">0,144</td><td align="center">0,244</td><td align="center">0,375</td><td align="center">0,356</td><td align="center">0,346</td><td align="center">0,308</td></tr></tbody></table></table-wrap></sec><sec><title>Deviation of individual genes is often not independent</title><p>Correlations between the values of log obs/exp in the 140 libraries were calculated for each pair of genes to determine whether the deviation of the individual genes is independent from each other. Of the 91 correlations, 18 significant correlations (P < 0,01) were observed. These significant correlations are shown In Figure <xref ref-type="fig" rid="F2">2</xref>. This figure shows a rather simple network of links in deviation of individual genes. Both positive and negative correlations were seen. The existence of these interactions suggests that there will be patterns in the emergence of deviant profiles.</p><fig position="float" id="F2"><label>Figure 2</label><caption><p>Network of significant correlations between the deviations in expression (log obs/exp) of the 14 proteasomal genes.</p></caption><graphic xlink:href="1745-6150-1-9-2"/></fig><p>The 3 most unstable genes are involved in significant correlations on 17 occasions as compared to 5 occasions for the 3 most stable genes. This indicates that instability in expression and involvement in correlation are somewhat related (P = 0,0105). The degree of expression of the individual genes (see Table <xref ref-type="table" rid="T2">2</xref>) does not seem to be involved in this pattern of interactions. Changes in the spatial architecture of the nucleus could well be responsible for the observed dependence, as any change in architecture will affect many genes.</p></sec><sec><title>Possible significance of the observed variation in transcription</title><p>As a whole, tumor tissues demonstrate a much larger variation in transcription profiles than normal tissues. This suggests, therefore, that the observed excessive variation in expression of proteasome genes is due to disorder and is not a consequence of an orderly regulation. However one cannot exclude that alternate hypotheses are possible. If due to disorder it is at present still to early to characterize the sources of the disorder observed. Although mutational damage could, in theory, be involved, this is not supported by the data as discussed above. Consequently, noise and/or chaotic processes might be causally involved but it is still too early to make any decisive statement in relation to this.</p><p>The observed deviation appears to be due to stable structural epigenetic changes. If our assumption is correct; that all tissues initially have an approximately similar expression profile (which is supported by the grosso modo similar expression profile of orderly and disorderly libraries, see Table <xref ref-type="table" rid="T2">2</xref> E and F), then the findings suggest that this expression profile can be altered in a progressive and unpredictable way resulting in widely different expression profiles. In addition, degree of progression of this deviation seems to have some unpredictability, as it was observed to exert a small effect on a number of genes simultaneously or a major effect for just one gene in particular (Figure <xref ref-type="fig" rid="F1">1</xref>). Therefore a deterministic epigenetic process could well be the cause for the observed deviations. Only time will tell.</p><p>Although the deviations in transcription, as described in this paper, relate only to the transcription of the proteasomal genes, it is not illogical to suppose that similar deviations will exist in transcription profiles of genes involved in other functions of the cell and that the corresponding deviation indices could provide information on the degree of order in transcription.</p><p>Our present working hypothesis for the structural aspect of the observed variation is that the patterns of over- and under-expression are a reflection of the localization of the genes within the nucleus. If this hypothesis were correct, then one would expect that future research would reveal correlations between genes that are completely unrelated in function in terms of their degree of over- or under-expression. Consequently, the next desirable step will be to investigate whether similar deviating profiles can be found in other organelles and/or pathways and whether a deviation in one profile corresponds with deviation in another profile.</p><p>This phenomenon of deviation in transcription could provide a new method to study genetic dysfunction. Apart from the field of gene expression, research into this phenomenon could turn out to be of value in other fields:</p><sec><title>1. Genomic instability</title><p>Any decrease in the ability of a cell to carry out normal cellular functions could lead to a less efficient DNA replication and to increased production of free radicals, resulting in a greater degree of spontaneous DNA damage. As DNA repair could also be less efficient a higher spontaneous and induced mutation rate might result. Therefore this new phenomenon of variation might well underlie the hitherto unexplained phenomenon of "persistent delayed genomic instability " [<xref ref-type="bibr" rid="B2">2</xref>] and might also provide an explanation for the trans-generational effects of parental irradiation [<xref ref-type="bibr" rid="B21">21</xref>].</p></sec><sec><title>2. Carcinogenesis</title><p>Mutations in oncogenes predispose a cell to developing a transformed phenotype. The switching on of the telomerase gene and other genes involved in immortalization is another prerequisite step in the process of carcinogenesis. In fact the acquisition of an immortal phenotype is the rate-limiting step in carcinogenesis [<xref ref-type="bibr" rid="B6">6</xref>,<xref ref-type="bibr" rid="B22">22</xref>]. The course of these epigenetic events, ultimately resulting in malignant transformation, is still not understood. Progressive disorder in transcription in pre-cancerous lesions could be involved in the rare switching on of genes involved in immortalization.</p></sec><sec><title>3. Cellular aging</title><p>The current database does not contain enough information to determine whether disorder plays any role in cellular aging. However as progressive epigenetic changes could well be at the core of cellular aging it is not inconceivable that the aging process will be reflected in the degree of deviation in expression profiles. Whether this will be seen as an increase or a decrease in variation is a fascinating question. As cellular aging is still largely a black box, investigation of the role of disorder during aging might contribute to an improved understanding of this process.</p></sec><sec><title>4. Cell dynamics</title><p>The cell is a complex system. It is a mystery as to how all of the cellular subsystems of the cells interact and function collectively as a complex whole. In complex systems 'spontaneous" processes like pattern formation, oscillation, bifurcation and chaotization occur, These processes might depend on very simple rules[<xref ref-type="bibr" rid="B23">23</xref>]. The observed deviations in transcription might reflect chaotic processes. So far however there is no direct evidence that the observed deviations have anything to do with chaos, only time will tell. Nevertheless, whether chaotic in nature or not, the study of variation in the regulation of gene expression might contribute to a better insight into the cell as a complex system, especially if it reflects changes in the spatial organization of the nucleus.</p></sec><sec><title>5. Practical implications</title><p>One direct consequence arising from this study is that determination of the degree of deviation can serve as a control for the quality of libraries, which are to be used for the identification of genes involved in cellular processes. Deviating libraries will be less suitable for the identification of the genes involved.</p><p>A deviation index might further prove to be of prognostic value in predicting the probability of progression of neoplastic and possibly of pre-neoplastic lesions and likewise could be used as an indicator of health.</p><p>In addition, since deviation in gene expression can either increase or decrease, it would be useful to determine the effects of medication, promoters and anti-promoters on the degree of deviation.</p><p>In many respects, the data presented in this paper should be considered as very provisional. Ideally instead of only the 14 genes studied here, one would like to see comparative data for a few hundred genes as well as the use of greater numbers of large libraries from both healthy and unhealthy donors. Although such information is not yet available, this could soon be the case. Nevertheless the data obtained so far does indicate that the study of variation in transcription in the cell could provide new clues in biology and biomedicine.</p></sec></sec></sec><sec sec-type="methods"><title>Methods</title><sec><title>Gene expression</title><p>The tags used to determine the degree of expression of the proteasome genes in SAGE libraries were derived from the GEO database [<xref ref-type="bibr" rid="B24">24</xref>] the proteasome tags in SAGE libraries were found by importing both the tags of table <xref ref-type="table" rid="T1">1</xref> as well as the library tag count file into Microsoft Access. A query that joins both tag fields results in a table showing the abundance of tags for each proteasome gene.</p><p>The SAGE libraries were obtained from the "Gene Expression Omnibus" (GEO) [<xref ref-type="bibr" rid="B25">25</xref>].</p><p>The 30 libraries from normal tissues are: GSM number 572, 573, 574, 676, 677, 685, 688, 691, 692, 695, 708, 713, 719, 728, 729, 738, 739, 760, 761, 762, 763, 780, 781, 785, 786, 819, 824, 1499, 2386, 3242. The 30 libraries from cancer tissues are: 670, 671, 672, 673, 686, 687, 689, 690, 693, 696, 697, 698, 699, 727, 731, 735, 736, 737, 740, 745, 755, 756, 765, 792, 793, 1497, 1516, 2443, 2451, 2578 (Table <xref ref-type="table" rid="T3">3</xref>). Only one library (GSM 709, leukocytes) was excluded since libraries derived from blood and blood-forming tissues might express the immunoproteasome that might then interfere with expression of the 20S proteasome. The 80 additional libraries are: 743; 744; 757; 758; 1498; 1730; 1731; 1732; 1733; 1734; 1735; 2382; 2383; 2384; 2385; 2389; 2408; 7498; 7800; 8505; 8867; 9103; 9104; 14731; 14732; 14733; 14734; 14737; 14739; 14740; 14741; 14742; 14743; 14745; 14746; 14747; 14748; 14749; 14750; 14753; 14754; 14756; 14757; 14760; 14761; 14762; 14763; 14765; 14766; 14767; 14768; 14769; 14771; 14772; 14773; 14774; 14775; 14776; 14779; 14780; 14781; 14782; 14783, 14786, 14787, 14788, 14790, 14791, 14792, 14793, 14794, 14795, 14796, 14797, 14798, 14799, 14800, 14801, 14806 and 14807</p></sec><sec><title>Statistics</title><p>The observed frequencies of the expression profiles were compared with their expected frequencies by chi-square. As the lower limit of the proteasomal tag count per library was 24 and the number of genes was 14, the expected number of tags was often less than 5 which is the lower limit of reliability for the application of chi-square. Therefore the outcome of the chi-square test was only used in a parameter free test (Wilcoxon) to compare the group of normal tissues with the group of cancer tissues.</p><p>For the calculation of the deviation index the standard deviation of the log of observed tags/expected tags was used. In those cases where the observed number of tags was 0, it was assumed that there was 1 tag. This index has as disadvantage that it is only symmetric if the changes in expression occur as fold-change. If the variation in expression occurs in percentages this index is not symmetric. Therefore a second index was calculated based on z-scores (z-score = (obs - exp)/√exp). Both deviation indices, noted as log ratio deviation index and z-score deviation index, have been applied.</p><p>The deviation index of 5 types of tumors was compared with ANOVA. In order to have 9 tumors per group some libraries had to be omitted from the analysis. This was achieved by leaving out the libraries with the lowest number of counts.</p></sec></sec><sec><title>Reviewers' comments</title><sec><title>Reviewer's report 1</title><p>Trey Ideker, University of California San Diego, La Jolla, California, United States</p><p>This manuscript by JWIM Simons examines the mRNA levels of proteasomal proteins across publicly-available SAGE data from both cancer and normal tissues. It reports that proteasomal mRNAs show more variance away from their average levels when looking in cancer tissues versus in normal tissues. It also presents a corollary finding, that normal and cancer tissues taken from the same patient tend to have protesomal expression levels that are very similar. Finally, the manuscript makes speculative remarks about the possible interpretation and impact of these findings. In this regard, the main claim is that high variance in proteasomal RNA levels is indicative that the cell is in a "disorderly state" and that this disorderly state is likely a cause, not an effect, of cancer.</p><p>The basic finding, that proteasome mRNA levels as measured by SAGE have higher variance from the mean when looking in cancer cells, is interesting and, as far as this reviewer can tell, arrived at through reasonable use of statistical methods. The corollary finding, that mRNA levels from the same patient are correlated in cancer versus normal cells, is also interesting and is nicely controlled by comparing the correlation within versus between patients.</p><p>On the other hand, framing these findings within an argument that cellular transcription can be "ordered" or "disordered" is speculative at best and much less compelling. In order to support the "disorder" argument (which is not a small suggestion in the discussion, but also appears in the manuscript title, abstract, introduction, and results) a larger body of evidence would need to be examined and presented. For instance, perhaps a more likely null hypothesis is simply that cancer cells are proliferating and thus have more protein turnover. And there are other equally plausible ideas that do not relate to a global order vs. disorder phenomenon. Without examining such alternate hypotheses and addressing them, the article reads much more like a "commmentary" or "opinion" article than a primary research paper.</p><sec><title>Author response</title><p><italic>Surely so far there is only an indication and not yet a proof that cellular transcription can be disordered. At present there is also not yet an overview of all possible alternate hypotheses let alone to examine them. To make sure that disorderly transcription is only one possible explanation a sentence has been added that states that alternative explanations are possible</italic>.</p><p>There is also a semantic problem with the use of the term "disorder". Ordered versus disorder has a concrete (and very different) meaning in the field of information theory, which attempts to measure it through quantities such as entropy. In fact, entropy might have been a much more natural metric to use for the proposed disorder index.</p></sec><sec><title>Author response</title><p><italic>In order to have a better separation between findings and speculation the term disorder has been replaced by a more descriptive term (e.g. excessive deviation) in the case of findings. This, I hope, dissolves the "semantic problem" and improves the distinction between fact and speculative interpretation. The changes have been made throughout the text. Your remark that "entropy might have been a much more natural metric to use for the proposed disorder index" has raised my interest greatly but I have no idea how such a thing could be accomplished</italic>.</p><p>A key assumption of the paper is that for organelles such as the proteasome, "the products of the genes involved [are] available in the correct amounts" and "therefore we can assume an optimal expression pattern of the transcription of the genes in question exists" (Results and discussion section, first paragraph). This may not need be the case. For instance, for the ribosome it has been shown that, while RNA levels can fluctuate in response to stimuli, the overall ribosomal protein levels are buffered from change. Depending on the function/component, such buffering can be due to differential RNA degradation, protein translation, and so on.</p></sec><sec><title>Author response</title><p><italic>Probably there will be many mechanisms within the cell that can buffer undesirable fluctuations. This does not bring down the assumption that optimal functioning will depend on optimal conditions and to some extent this also should hold for an expression profile</italic>.</p><p>Moreover, given the assumption that correct amounts of each proteasomal subunit are needed, then why are the average or "expected" amounts of each subunit so different from one another in the SAGE data? Since the main result is measuring deviations from these expectations in individual patients, this point is particularly important. I would be interested to see how the results are impacted if the expected amounts are equal to each other across all subunits.</p></sec><sec><title>Author response</title><p><italic>You rightly put the finger on the remarkable differences in tag counts between genes in the mean expression profile while all subunits are equally important for building the proteasome and thus similar frequencies are expected. For this there is not yet a definite answer: remarkable is that when another tag is used (for some genes an additional tag is available) the absolute frequency can be consistently much lower. Explanations could be the length of the cDNA or the efficiency of the cutting enzyme, maybe also splicing or RNA degradation?? For this point I have been looking to microarray data and found that the mean profile is very different from the SAGE profile, the two profiles did not even correlate. Apparently the technique to obtain expression data affects the outcome and it is not clear to me whether this heterogeneity is biological. Taking the mean of the two expression profiles of microarray and SAGE produces an expression profile that is much closer to the "equal amounts" concept. Also this is an issue for the future</italic>.</p><p>In conclusion, my recommendation for this article would be to (1) remove many of the speculative remarks, or at least leave them for the discussion, including any interpretation of the results as "disorderly"; and to also (2) perform and present a more comprehensive body of findings which support the points in the discussion section that remain. For this second point, at minimum it would be nice to see a survey of all organelles/functions in the cell and whether they show greater variance in cancer than normal. Otherwise there is no evidence that the specific anecdote of the proteasome can be abstracted to some general principle of the cell.</p></sec><sec><title>Author response</title><p><italic>these recommendations have been met by the clearer distinction between findings and possible interpretations as described above</italic>. We <italic>surely agree that the findings with the proteasome cannot yet be abstracted to a general principle. Therefore </italic>w<italic>ith respect to your recommendation to make a similar analysis for all organelles/functions in the cell, it is obvious that such would be our wish and this was as such also stated in the manuscript. However this is physically impossible, my lifespan would not be long enough. The only thing I can do is to point to this new phenomenon and to contribute further to its interpretation in the hope that also other scientists will study this new phenomenon</italic>.</p></sec></sec><sec><title>Reviewer's report 2</title><p>Itai Yanai, Harvard University, Cambridge, Massachusetts, United States</p><p>In this paper, Simons uses public SAGE data to quantify changes in gene expression of the set of 14 genes that compose the proteasome. First, the overall relative frequencies of these genes are calculated. A SAGE library is then described as disorderly if the standard deviation of the genes' observed to overall differences is high. It is noteworthy, that the frequencies of the most disorderly libraries are remarkably similar to those of the most orderly libraries, suggesting there is no characteristic state of disorder but instead that each mess is unique.</p><p>Simons then shows that the overall tumor libraries are significantly more disorderly – as evidenced by Wilcoxon's test – as a group than the normal libraries. Since this is an important point, I believe it would be helpful to visualize this difference with a principal components type analysis. The 14 dimensions (genes) can be reduced to 2 or 3 and plotted for both the normal and tumor samples. This method is further called for since the author shows in Figure <xref ref-type="fig" rid="F2">2</xref> that the genes expression are correlated.</p><sec><title>Author response</title><p><italic>No doubt this could be worthwhile. However, the analysis is unfamiliar to me (I had even never heard of the method). In a first trial with PCA it was found that the first three factors count for only 41,46% of the variability and thus principal components do not appear to be present. In the future, when I am more familiar with this method, I certainly will try to perform a PCA</italic>.</p><p>An issue is next raised about whether disorderly profiles represent "momentary fluctuations in transcription rates" or heritable states. The author shows that deviations of the proteasome genes' expression in a tumor correlates with those in non-tumor from the same patient. Based upon this evidence, the author states that "a disorderly expression profile can be extremely stable and can be transmitted to a clonal derivative as a constitutive trait." However, since this result was observed in three of only four instances, it would be prudent not to draw too strong of a conclusion about the stability of gene disorder based upon this dataset alone. Furthermore, one could argue that difficulties associated with exclusively dissecting tumor vs. non-tumor samples from a given tissue, compromise our ability to meaningfully compare them; i.e. the two may be similar simply because of impure sample isolations.</p></sec><sec><title>Author response</title><p><italic>Hereditability of deviating expression profiles was observed for three of the four cases. To me it seems unlikely to be due to admixtures of normal tissue in three tumor samples. Of course so far there are only three cases. To stress that this hereditability is not necessarily connected with disorder in transcription a sentence has been added and a question remark has been added</italic>.</p><p>The author presents an index for disorderliness: standard deviation of the log obs/exp values. I would advise against this formulation because it is not symmetric, biasing in favor of reduced expression. For example, an increase of expression by 10% would be log(1.1/1) = 0.0952, while a decrease by 10% results in log(0.9/1) = -0.1054. Since, the author makes the point that disorderliness tends to occur in terms of under-expression, the lack of symmetry in the index is a clear confounding effect. This can be easily fixed by taking the log of the absolute difference, and adding a negative sign if obs is less than exp. However, it may be the best to convert to Z-scores, to explicitly take into account the variation of each gene's expression.</p></sec><sec><title>Author response</title><p><italic>Your remark on the possible absence of symmetry in the log ratio deviation index did initially worry me. According to your suggestion z-score values have been determined and compared to the log ratio. It turns out that the z-scores discriminate less between normal and tumor indicating that the changes in expression reflect fold-changes rather than percentual changes. This has been added to the text</italic>.</p></sec></sec><sec><title>Reviewer's report 3</title><p>Stephan Beck, The Wellcome Trust Sanger Institute, Hinxton, United Kingdom</p><p>I accepted to review this manuscript on the premise that experts in SAGE expression analysis and statistics will be secured as additional reviewers to assess the methodologies and tests carried out in this study.</p><p>The manuscript by JWIM Simons aims to address several fundamental questions, listed on page 2:</p><p>1) Can disorder in transcription be demonstrated?</p><p>2) Does such disorder have a degree of permanence?</p><p>3) Does it play any role in health and disease?</p><p>While I commend the author for tackling such complex questions, I do not agree with many of the conclusions and believe the study is compromised by inadequate assumptions and data selection. My main concerns are:</p><p>1) The selection of only 14 genes (20S proteasome complex) is not nearly enough to represent the human transcriptome. In addition, I do not agree with the rationale for some of the additional stratifications of the libraries/data given in the 'Results and Discussion' (first section), and in the 'Methods'. For example, leukocyte libraries were excluded from the analysis on the basis that "these cells might express the immunoproteasome PSMB8 and PSMB9 genes that might then interfere with expression of the 20S proteasome". According to the GNF gene expression database <ext-link ext-link-type="uri" xlink:href="http://expression.gnf.org/"/>, PSMB9 for instance is not only expressed in blood, but also in lung, thymus, spleen and heart. Therefore, why were these tissues not excluded as well if this is the right thing to do in the first place?</p><sec><title>Author response</title><p><italic>I agree completely with your remark that a sample of 14 genes is not nearly enough to represent the human transcriptome. This has also been stated in the paper. This sample is just the first start to investigate whether such an approach is possible and could be useful. With respect to selection criteria, a selection has to be made, as the immunoproteasome is another organelle than the proteasome. The point therefore is where to draw a line. Whether the immunoproteasome could be normally present in some tissues that are not involved in blood-formation I really do not know. As these expressions, if present, are very low the best policy, in my view, is to draw a line between blood-forming tissues and others. Therefore for the selection of the libraries all libraries from blood forming tissues would have excluded, thus also spleen, thymus and tonsils</italic>.</p><p>2) I could not work out how the author defines disorderly expression profiles, except for the definition on page 2 where 'disorder' is defined as "excessive variation in transcription irrespective of cause". However, the author does not seem to take into account that natural variation in gene expression can be quite high (up to 14.13%) in unrelated individuals as compared to e.g. monozygotic twins (up to 1.76%) (see e.g. Sharma et al. Physiol. Genomics 2005 21:117-23). If the above ~10-fold difference falls within the definition used here for 'excessive', then perfectly 'normal' expression profiles would be classified as 'disorderly'.</p></sec><sec><title>Author response</title><p><italic>I also agree that at present it is not possible to conclude with certainty what to classify as disorderly. Therefore in order to have a better separation between findings and speculation I have been replacing the term disorder by more descriptive terms (e.g. excessive deviation) in the case it concerns findings. The changes have been made throughout the text and a question remark has been put in the title. Your argument about natural variation in unrelated individuals is strictly speaking not fully valid as in that case the quantitative expression of genes was compared and no use was made of an expression profile, which shows the relative expression</italic>.</p><p>3) I do not agree with the 'one-fits-all' assumption made on in the 'Results and Discussion' (first paragraph), that "to establish the existence of a state of cellular disorder that affects all transcription profiles, the first step requires the choice of only one expression profile that could serve as a model for all profiles in a complex system."</p></sec><sec><title>Author response</title><p><italic>It has not been my intention to suggest that the use of only one profile would be sufficient to establish the existence of a state of disorder in a tissue (the "one fits all assumption"). This was also discussed in the text. To avoid such a misinterpretation the remark referenced here has been improved</italic>.</p><p>4) There are numerous statements throughout the manuscript which are unsubstantiated and I cannot not agree with. For instance, the statement in the 'Results and Discussion' (section: Heritability of variant profiles): "The expression profiles of the two prostates clearly differ from each other, which suggests that a specific expression profile for the prostate does not exist and that these two profiles are disorderly. The similarity in disorderly expression profiles of a normal tissue and its tumor suggests the possibility of the disorderly condition in the normal tissue being a predisposing factor in the eventual appearance of the tumor."</p></sec><sec><title>Author response</title><p><italic>After a better separation of findings and speculations I assume that it is evident that the numerous unsubstantiated sayings throughout the manuscript are of a speculative nature</italic>.</p><p>5) On several occasions the author suggests epigenetic changes to be responsible for 'disorderly' profiles. Yet, no supporting evidence is provided.</p></sec><sec><title>Author response</title><p><italic>"Epigenetic changes". The subheading 'Disorder in transcription can occur in euploid cells" has been changed into "Epigenetic origin of excessive variation in transcription". In this section three arguments were already given as supportive evidence for an epigenetic origin of the changes in expression profiles</italic>.</p><p>6) The final 5-point conclusion/outlook in 'Results and Discussion' is pure speculation.</p></sec><sec><title>Author response</title><p><italic>The 5-point reflection should indeed be read as speculation. The subheading of this section has been altered according to this</italic>.</p><p>7) Details of the additional 80 libraries mentioned in 'Methods' (section: Gene expression) should be included in Table <xref ref-type="table" rid="T3">3</xref>.</p></sec><sec><title>Author response</title><p><italic>Details of the additional 80 libraries have been included in </italic>table <xref ref-type="table" rid="T3">3</xref>.</p><p>For the reasons outlined above, I find the manuscript not acceptable as Research or Review article and even questionable as Commentary/Hypothesis article. I declare that I have no competing interests.</p></sec><sec><title>Author response</title><p><italic>for the reasons outlined above and after the improvements made we trust that the manuscript is a valuable contribution. In our view this is clearly a research article</italic>.</p></sec></sec></sec> |
Pathogenesis and vertical transmission of a transplacental rat cytomegalovirus | <sec><title>Background</title><p>Cytomegalovirus (CMV) congenital infection is the major viral cause of well-documented birth defects in human. Because CMV is species-specific, the main obstacle to developing animal models for congenital infection is the difference in placental architecture, which preludes virus transmission across the placenta. The rat placenta, resembling histologically to that of human, could therefore facilitate the study of CMV congenital infection in human.</p></sec><sec><title>Results</title><p>In this report, we present clear evidences of the transplacental property of a new rat CMV (RCMV), namely ALL-03, which had been isolated from placenta and uterus of the house rat. Our study signifies the detection of infectious virus, virus particles, viral protein and DNA as well as immune response to demonstrate a natural model of acute CMV infection including the immunocompetent and immunocompromised host associated with or without pregnancy. It is characterized by a full range of CMV related clinical signs; lesions and anatomical virus distribution to uterus, placenta, embryo, fetus, neonate, lung, kidney, spleen, liver and salivary gland of the infected rats in addition to the virus-specific seroconversion. The preference of the virus for different organs mimics the situation in immunocompromised man. Most interestingly, the placenta was observed to be involved in the maternofetal infection and hence confirmed the hypothesis that the RCMV strain ALL-03 is capable to cross the placenta and infect the offsprings congenitally.</p></sec><sec><title>Conclusion</title><p>The maternal viremia leading to uterine infection which subsequently infecting to the fetus through the placenta is the most likely phenomenon of CMV vertical transmission in our study.</p></sec> | <contrib id="A1" contrib-type="author"><name><surname>Loh</surname><given-names>Hwei-San</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>sandylhs@mail2world.com</email></contrib><contrib id="A2" corresp="yes" contrib-type="author"><name><surname>Mohd-Lila</surname><given-names>Mohd-Azmi</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>azmi@ibs.upm.edu.my</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Abdul-Rahman</surname><given-names>Sheikh-Omar</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>sheikh@vet.upm.edu.my</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Kiew</surname><given-names>Lik-Jun</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>likjunk@hotmail.com</email></contrib> | Virology Journal | <sec><title>Background</title><p>Cytomegalovirus (CMV) infection is the most frequent congenital infection in humans worldwide, with an incidence of 0.2–2.2% of live births [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B2">2</xref>]. One major concern of CMV congenital infection is birth defects including mental retardation, microcephaly, epilepsy, and blindness. However, little is known on how the virus is transmitted to the fetus during pregnancy [<xref ref-type="bibr" rid="B3">3</xref>]. The possible routes of transmission of human CMV (HCMV) to the offsprings are vertical via germ line cells or transplacentally; perinatally and postnatally. There are several reports strongly supporting the hypothesis that placental infection precedes viral transmission to the fetus [<xref ref-type="bibr" rid="B3">3</xref>-<xref ref-type="bibr" rid="B6">6</xref>].</p><p>Due to the strict species-specificity of HCMV, it has not generally been possible to study this virus in experimental animals. A number of natural CMV infections in various animal species have been utilized for modeling HCMV infection. Among the animal CMVs, transplacental transmission has been reported for rhesus macaque CMV [<xref ref-type="bibr" rid="B7">7</xref>], porcine CMV [<xref ref-type="bibr" rid="B8">8</xref>] and guinea pig CMV (GPCMV) [<xref ref-type="bibr" rid="B9">9</xref>]. However, the expenses of the primates and pigs, as well as the rarity of their CMV seronegative animals make these models impractical for large-scale vaccine studies. For these reasons, rats, mice, and guinea pigs came into favor because of their small size, low cost, short life span, ease of handling and high reproductive rate. More importantly, these CMVs (RCMV, MCMV and GPCMV) closely resemble HCMV. For studying the transplacental hypothesis, it is important to consider the great diversity in the placental structures among human and model. Favorably, these three animals have similar discoidal hemochorial placentation to that of human [<xref ref-type="bibr" rid="B10">10</xref>]. However, none of the existing MCMVs and RCMVs demonstrated a clear involvement of the placenta in vertical transmission [<xref ref-type="bibr" rid="B11">11</xref>,<xref ref-type="bibr" rid="B12">12</xref>] and are therefore, less suitable for the study of CMV congenital infection [<xref ref-type="bibr" rid="B13">13</xref>,<xref ref-type="bibr" rid="B14">14</xref>]. Although GPCMV provides a well-characterized model of transplacental viral infection, studies in this system have been hampered by a lack of genetic knowledge of the animal itself. In addition, the cost of guinea pigs is less practical for large-scale vaccine and long-term maintenance studies as compared to mice and rats. Meanwhile, the desirable features of rat biology include more human-like physiological responses for disease process, an extensive behavioral database, and larger size (better suited to surgical manipulation and repeated blood sampling) are the major advantages of the rat model over the mouse model. Besides, following human [<xref ref-type="bibr" rid="B15">15</xref>,<xref ref-type="bibr" rid="B16">16</xref>] and mouse [<xref ref-type="bibr" rid="B17">17</xref>], rat is the third mammalian for which the complete genome has been determined. Almost all human genes noted to be associated with disease have known counterparts in the rat genome [<xref ref-type="bibr" rid="B18">18</xref>]. This genetic explorer for the rat provides an unprecedented opportunity to take advantage of the rich and robust history of experimental studies utilizing this species to study HCMV disease. Hence, the rat system is a significant advance on the guinea pig or mouse model for studying various aspects of viral pathogenesis, the effect of therapeutic intervention as well as the evaluation of vaccine candidates for CMV congenital infection in humans.</p><p>In our previous study, we have discovered a new RCMV isolate (ALL-03) obtained from placenta and uterus of the house rat, <italic>Rattus rattus diardii </italic>[<xref ref-type="bibr" rid="B19">19</xref>]. The involvement of the placental and uterine tissues during virus isolation indicates that the virus has the ability to cross the placenta and infect the fetus. Therefore, an attempt was made to study the maternofetal involvement in the pathogenicity of RCMV infection. In this report, we demonstrate a natural model of acute RCMV infection, which includes the characteristic organ distribution of RCMV in male rats and female rats with or without pregnancy as well as the immune response to the infection. More importantly, this is the first RCMV infection study capable of presenting a clear evidence of transplacental transmission in pregnant rats.</p></sec><sec><title>Results</title><p>The rats were challenged with RCMV and sampled at different time point, i.e. day 21 p.i. for Experiment A, B and D, meanwhile, day 13–14 p.i. for Experiment C. The presences of infectious virus, viral DNA and antigen, virus particles as well as seroconversion were assessed by employing techniques such as histological and immunohistological stainings including H&E, IIP and IIF; virus assay; protein blotting; PCR; TEM and indirect ELISA.</p><sec><title>Clinical observation</title><p>The animals in the four experiments were observed twice daily until the time for sampling. No abnormality was observed in all control groups throughout the study. All treatment groups showed no clinical signs from day 1 to day 5 p.i. After an incubation period of 6 to 21 days, the RCMV infection became symptomatic especially the immunocompromised groups. The infected rats of all immunocompromised groups in Experiment A, B, C and D as well as immunocompetent groups in Experiment C and D became less active. The clinical signs such as hemorrhages at the extremities of the limbs and tails, and ruffling of hair coat were obvious. There were absences of abortion and mortality in rats up to day 21 p.i. The post-partum neonates in Experiment C did not show any apparent abnormality as compared to the control groups except the litter size in treatment groups (7–8 pups) was slightly smaller than that of control groups (8–9 pups).</p></sec><sec><title>Gross pathology</title><p>No abnormalities were observed in the organs of all control animals in the four experiments. The lesions such as congestion of renal cortex and corticomedullary junction, generalized hemorrhage of the lung and marked splenomegaly were common and observed mostly in immunosuppressed and pregnant rats. Mild hemorrhage was found in the uterus serosal surface of an infected immunosuppressed dam (Experiment D) carrying seven conceptuses.</p></sec><sec><title>Histological and immunohistological pathology</title><p>The presences of the characteristic histopathological changes in the organs of animals in the four experiments were determined by H&E staining and further confirmed by IIP test. No specific lesions caused by RCMV disease were observed in all control groups. The organs that appeared normal histologically and did not show characteristics of infection in all treatment groups were brain, heart, testes and ovary. The immunoreactivity of IIP test of the treatment groups is presented in Table <xref ref-type="table" rid="T1">1</xref>. The histopathological and immunopathological findings are described in the following:</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>Positive immunoreactivity of IIP test on different tissue sections of treatment groups.</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="center"><bold>Experiment/Organ</bold></td><td align="center" colspan="2"><bold>A (Day 21 p.i.)</bold></td><td align="center" colspan="2"><bold>B (Day 21 p.i.)</bold></td><td align="center" colspan="2"><bold>C (Day 13–14 p.i.)</bold></td><td align="center" colspan="2"><bold>D (Day 21 p.i.)</bold></td></tr><tr><td></td><td colspan="8"><hr></hr></td></tr><tr><td></td><td align="center" colspan="8"><bold>Group</bold></td></tr><tr><td></td><td colspan="8"><hr></hr></td></tr><tr><td></td><td align="center"><bold>v</bold></td><td align="center"><bold>pv</bold></td><td align="center"><bold>v</bold></td><td align="center"><bold>pv</bold></td><td align="center"><bold>v</bold></td><td align="center"><bold>pv</bold></td><td align="center"><bold>v</bold></td><td align="center"><bold>pv</bold></td></tr></thead><tbody><tr><td align="center">Brain</td><td align="center">0/3</td><td align="center">0/3</td><td align="center">0/3</td><td align="center">0/3</td><td align="center">0/3</td><td align="center">0/3</td><td align="center">0/3</td><td align="center">0/3</td></tr><tr><td align="center">Salivary gland</td><td align="center">0/3</td><td align="center">1/3</td><td align="center">0/3</td><td align="center">1/3</td><td align="center">0/3</td><td align="center">0/3</td><td align="center">1/3</td><td align="center">2/3</td></tr><tr><td align="center">Heart</td><td align="center">0/3</td><td align="center">0/3</td><td align="center">0/3</td><td align="center">0/3</td><td align="center">0/3</td><td align="center">0/3</td><td align="center">0/3</td><td align="center">0/3</td></tr><tr><td align="center">Lung</td><td align="center">0/3</td><td align="center">2/3</td><td align="center">1/3</td><td align="center">3/3</td><td align="center">2/3</td><td align="center">3/3</td><td align="center">3/3</td><td align="center">3/3</td></tr><tr><td align="center">Spleen</td><td align="center">0/3</td><td align="center">1/3</td><td align="center">0/3</td><td align="center">2/3</td><td align="center">2/3</td><td align="center">3/3</td><td align="center">2/3</td><td align="center">3/3</td></tr><tr><td align="center">Liver</td><td align="center">0/3</td><td align="center">1/3</td><td align="center">0/3</td><td align="center">2/3</td><td align="center">1/3</td><td align="center">3/3</td><td align="center">1/3</td><td align="center">3/3</td></tr><tr><td align="center">Kidney</td><td align="center">0/3</td><td align="center">2/3</td><td align="center">0/3</td><td align="center">2/3</td><td align="center">2/3</td><td align="center">3/3</td><td align="center">2/3</td><td align="center">3/3</td></tr><tr><td align="center">Testes</td><td align="center">0/3</td><td align="center">0/3</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">-</td></tr><tr><td align="center">Ovary</td><td align="center">-</td><td align="center">-</td><td align="center">0/3</td><td align="center">0/3</td><td align="center">0/3</td><td align="center">0/3</td><td align="center">0/3</td><td align="center">0/3</td></tr><tr><td align="center">Uterus</td><td align="center">-</td><td align="center">-</td><td align="center">1/3</td><td align="center">3/3</td><td align="center">2/3</td><td align="center">3/3</td><td align="center">3/3</td><td align="center">3/3</td></tr><tr><td align="center">Neonate</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">6/15</td><td align="center">12/15</td><td align="center">-</td><td align="center">-</td></tr><tr><td align="center">Placenta</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">12/15</td><td align="center">8/10</td></tr><tr><td align="center">Fetus</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">9/15</td><td align="center">6/10</td></tr><tr><td align="center">Embryo*</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">5/5</td></tr></tbody></table><table-wrap-foot><p>Note: Abbreviations: v = virus-infected and pv = virus-infected with immunosuppression.</p><p>* = eroded placenta and developing embryo in uterus at ≤ 7 days of pregnancy.</p><p>0/3 = no detectable positive result over triplicate sample trials in all three rats.</p></table-wrap-foot></table-wrap><sec><title>Salivary gland</title><p>Localization of RCMV infection in all salivary glands, i.e. parotid, submandibular and sublingual glands was observed. However, the submandibular gland was stained more frequently than the other types of salivary glands. The positive findings were established in immunosuppressed rats in Experiment A and B; in pregnant rats of both treatment groups in Experiment D. No positive features of RCMV infection in all groups of Experiment C were evident. The RCMV infection in the salivary glands was confined to the striated ducts, secretory acini (Figure <xref ref-type="fig" rid="F1">1a</xref>) and trabeculae connective tissues. The histological abnormalities such as the swollen and enlarged mucous cells and acinar cells were evident but not frequently.</p><fig position="float" id="F1"><label>Figure 1</label><caption><p><bold>Positive IIP-stained tissue sections of infected immunosuppressed rats</bold>. (a) secretory acinar cells (arrows) of sublingual gland (D; day 21 p.i.; × 400), (b) bronchioles (arrows) and lung parenchyma (D; day 21 p.i.; × 200), (c) splenic cells (arrow; D; day 21 p.i.; × 400), (d) nucleus (arrow) and cytoplasm (arrowhead) of hepatocytes (C; day 13 p.i.; × 400), (e) renal tubules (arrows; D; day 21 p.i.; × 400), (f) stratum basalis (arrows) of endometrium (C; day 13 p.i.; × 200).</p></caption><graphic xlink:href="1743-422X-3-42-1"/></fig></sec><sec><title>Lung</title><p>The parenchyma particularly the bronchioles and alveoli was solely permissive for CMV infection (Figure <xref ref-type="fig" rid="F1">1b</xref>). Intranuclear and intracytoplasmic inclusion bodies stained extensively by IIP were found in the swollen bronchiolar and alveolar cells. The macrophages and occasional pneumocytes in alveolar wall as well as ciliated bronchiolar epithelia were immunoreactive to CMV. The common pathological features included the congested and hemorrhagic interstitium, accumulation of proteinaceous fluid with infected and uninfected monocytes and macrophages in alveoli and bronchioles, thickened alveolar septa, perivascular inflammatory cell cuffings and lymphocytic hyperplasia.</p></sec><sec><title>Spleen</title><p>Some of the infected immunocompetent animals showed reactive hyperplasia of spleens though IIP test did not show positive staining. In contrast, the splenic tissue of immunosuppressed animals especially those with splenomegaly was notably stained by IIP (Figure <xref ref-type="fig" rid="F1">1c</xref>). Most of the infected areas were less extensive and often scattered at a distance in red pulps. The periarterior lymphocyte sheaths of immunosuppressed animals had shrunk to some extent. The splenic sinusoids were infiltrated with numerous macrophages, many of which contained viral antigens. Numerous lymphocytes and plasma cells were often present in both white and red pulps.</p></sec><sec><title>Liver</title><p>The intensity of immunostaining was marked in liver tissues of immunosuppressed animals in Experiment C, which involved almost entirely the tested sections (two cases; Figure <xref ref-type="fig" rid="F1">1d</xref>). Most of the immunoreactive cells were located in the liver lobules adjacent to the capsule. Numerous hepatocytes showed characteristic inclusion bodies. The hepatocytes and many Kupffer cells contained viral antigens. The cytoplasm of hepatocytes stained more frequently than the nucleus (Figure <xref ref-type="fig" rid="F1">1d</xref>). The parenchyma showed patchy necrosis and degeneration. Hepatitis seen as infiltration of inflammatory cells in the parenchyma was one of the lesions found.</p></sec><sec><title>Kidney</title><p>Almost all treatment groups had animal(s) with signs of infection except the immunocompetent groups in Experiment A and B. In the kidney, infected cells were seen in both the cortex and medulla regions whereby the cortex region adjacent to the renal capsule was predominantly infected. Viral antigens were profound in the proximal and distal tubules, loop of Henle, and collecting tubules (Figure <xref ref-type="fig" rid="F1">1e</xref>), but less intensive in the renal corpuscles. The infection was predominant in cytoplasm rather than the nucleus. The mesangial cells were swollen and displayed characteristic nuclear inclusions, which contained the viral antigens. Tubulonephrosis in the form of ballooning degeneration was evident. Hypercellularity of the glomerulus was one of the lesions showing adhesion between the glomerular tuft and Bowman's capsule.</p></sec><sec><title>Uterus</title><p>All immunosuppressed female rats in the three experiments (B, C and D) regardless of presence or absence of pregnancy demonstrated signs of infection in particularly the endometrium. The immunoreactive cells were found majority in the stroma and surface epithelia, i.e. stratum basalis and stratum functionalis. The predominant localization of viral antigen was slightly different from one rat to another even within a group receiving identical treatment. Two rats in Experiment B and one in Experiment C had positive stromal cells for the immunostaining but not epithelial cells of glands. Meanwhile, three pregnant rats in Experiment D had viral tropism in epithelial cells only. Nevertheless, the majority of the rats showed immunoreactivity in the two regions and with more extensive staining in the stratum functionalis and stratum basalis (Figure <xref ref-type="fig" rid="F1">1f</xref>).</p></sec><sec><title>Placenta</title><p>Both immunocompetent and immunosuppressed groups in Experiment D gave 80% of positivity in IIP staining. Meanwhile, the placenta sections (categorized as Embryo* in Table <xref ref-type="table" rid="T1">1</xref>) of the two dams with about 7-day pregnancy, gave the most intensive stains i.e. 100% of positivity, which far surpassed those with pregnancy length greater than 14 days. The immunoreactive sites of the placenta were mostly at the decidual basalis, junctional zone and labyrinth zone (Figure <xref ref-type="fig" rid="F2">2a, 2b, 2c, 2d</xref>) but scarcely in the embryonic sites. However, the placenta with shorter gestation period showed more signs of infection in decidual basalis and junctional zone as compared to those with longer gestation period by which infections were found in the labyrinth zone predominantly. The chorionic villi anchoring to the decidual basalis concomitantly passing infection to junctional zone of placenta was observed (Figure <xref ref-type="fig" rid="F2">2c</xref>). These cells of maternal (decidual basalis) and fetal (chorionic villi and junctional zone) portions of placenta, were confirmed to be infected. The infected regions were found to be associated with intranuclear and intracytoplasmic inclusion bodies mostly of trophoblast cells in junctional and labyrinth zones (Figure <xref ref-type="fig" rid="F2">2b, 2d</xref>).</p><fig position="float" id="F2"><label>Figure 2</label><caption><p><bold>Positive IIP-stained placental and fetal tissue sections of infected immunosuppressed dams</bold>. Seven-day old placenta (D; day 21 p.i.): (a) decidual epithelia (arrows; × 200), (b) junctional zone (arrows; × 200), (c) chorionic villi (arrow) anchored to the decidual basalis concomitantly passing infection to junctional zone (arrowhead; × 400), and (d) trophoblast cells (arrows) in labyrinth zone (× 400); (e) fetal renal tubules (arrows) of 17-day pregnancy (D; day 21 p.i.; × 200), (f) fetal liver (arrow) of 18-day pregnancy (D; day 21 p.i.; × 400).</p></caption><graphic xlink:href="1743-422X-3-42-2"/></fig></sec><sec><title>Neonate and fetus</title><p>The fetal tissues of those dams beyond 14 days of pregnancy in Experiment D, especially liver and kidney showed a significant presence of viral antigen (Figure <xref ref-type="fig" rid="F2">2e, 2f</xref>). For neonatal rats, no immunoreactivity was observed in salivary gland, however, positive results were found in the kidney and liver. The renal tubules were stained more frequently than the glomeruli. The proportion of immunoreactivity in a tissue was found generally greater in fetus rather than neonate.</p></sec></sec><sec><title>Virus assay</title><p>Virus was isolated from tissues of animals in Experiment C and D, namely the uterus, placenta, embryo, neonate and fetus; examined by culture in rat embryonic fibroblasts (REF). The virus produced typical herpesvirus-like CPE in REF inoculated with infected tissue homogenates beginning from 3 days p.i. and was identified as RCMV infection by IIP technique at day 5 p.i. The CPE and IIP results were similar as previously mentioned in Loh <italic>et al </italic>[<xref ref-type="bibr" rid="B19">19</xref>]. However, these features were not observed in mock-infected REF cells. The quantity of positive observations in different tissues is tabulated in Table <xref ref-type="table" rid="T2">2</xref>.</p><table-wrap position="float" id="T2"><label>Table 2</label><caption><p>Positivity of CPE development and protein blotting of treatment groups in Experiment C and D.</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="center"><bold>Test</bold></td><td align="center" colspan="4"><bold>Virus assay (CPE)</bold></td><td align="center" colspan="4"><bold>Protein blotting</bold></td></tr></thead><tbody><tr><td align="center"><bold>Experiment/Organ</bold></td><td align="center" colspan="2"><bold>C (Day 13–14 p.i.)</bold></td><td align="center" colspan="2"><bold>D (Day 21 p.i.)</bold></td><td align="center" colspan="2"><bold>C (Day 13–14 p.i.)</bold></td><td align="center" colspan="2"><bold>D (Day 21 p.i.)</bold></td></tr><tr><td></td><td colspan="8"><hr></hr></td></tr><tr><td></td><td align="center" colspan="4"><bold>Group</bold></td><td align="center" colspan="4"><bold>Group</bold></td></tr><tr><td></td><td colspan="8"><hr></hr></td></tr><tr><td></td><td align="center"><bold>v</bold></td><td align="center"><bold>pv</bold></td><td align="center"><bold>v</bold></td><td align="center"><bold>pv</bold></td><td align="center"><bold>v</bold></td><td align="center"><bold>pv</bold></td><td align="center"><bold>v</bold></td><td align="center"><bold>pv</bold></td></tr><tr><td colspan="9"><hr></hr></td></tr><tr><td align="center">Uterus</td><td align="center">5/5</td><td align="center">5/5</td><td align="center">5/5</td><td align="center">5/5</td><td align="center">5/5</td><td align="center">5/5</td><td align="center">5/5</td><td align="center">5/5</td></tr><tr><td align="center">Neonate</td><td align="center">11/18</td><td align="center">16/18</td><td align="center">-</td><td align="center">-</td><td align="center">12/18</td><td align="center">15/18</td><td align="center">-</td><td align="center">-</td></tr><tr><td align="center">Placenta</td><td align="center">-</td><td align="center">-</td><td align="center">16/20</td><td align="center">10/10</td><td align="center">-</td><td align="center">-</td><td align="center">14/20</td><td align="center">8/10</td></tr><tr><td align="center">Fetus</td><td align="center">-</td><td align="center">-</td><td align="center">14/20</td><td align="center">7/10</td><td align="center">-</td><td align="center">-</td><td align="center">12/20</td><td align="center">6/10</td></tr><tr><td align="center">Embryo*</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">8/8</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">8/8</td></tr></tbody></table><table-wrap-foot><p>Note: Abbreviations: v = virus-infected and pv = virus-infected with immunosuppression.</p><p>* = eroded placenta and developing embryo in uterus at ≤ 7 days of pregnancy.</p></table-wrap-foot></table-wrap></sec><sec><title>Protein blotting</title><p>In the system, we used RCMV-infected cell lysate and mock-infected cell lysate, respectively for the positive and negative controls. The system was employed on the same samples for virus assay i.e. uterus and neonatal tissues collected from Experiment C as well as uterus, placenta and fetal tissues collected from Experiment D. The purified virus protein blots of uterus, placenta, embryo, neonate and fetus reacted positively in different frequency with HIS raised against RCMV (Table <xref ref-type="table" rid="T2">2</xref>).</p></sec><sec><title>PCR detection of <italic>IE1 </italic>gene</title><p>Similar samples tested in protein blotting were transversely analyzed by PCR amplification of viral DNA. Pure RCMV DNA serving as the positive control showed a distinct band of 569 bp in molecular size. Significant positive results in uterine, placental, neonatal and fetal samples were obtained (Figure <xref ref-type="fig" rid="F3">3</xref>). One heart sample, which had no immunostain in IIP test showed positive result in PCR. In contrast, no similar DNA band was detected in any tissue samples of control rats. The magnitude of positive observations is shown in Table <xref ref-type="table" rid="T3">3</xref>.</p><fig position="float" id="F3"><label>Figure 3</label><caption><p><bold>PCR profile of <italic>IE1</italic>-specific products</bold>. Viral DNA extracted from (i) infected immunosuppressed rats: uterus (C; day 14 p.i.; lane 2), 17-day old placenta (D; day 21 p.i.; lane 3), one-day post-partum neonatal tissues (C; day 14 p.i.; lane 4) and 17-day old fetal tissues (D; day 21 p.i.; lane 5); (ii) mock-infected immunosuppressed rats: uterus (C; day 13 p.i.; lane 6) and 17-day old placenta (D; day 21 p.i.; lane 7). Lane 1: GeneRuler™ 1 kb DNA ladder (Fermentas).</p></caption><graphic xlink:href="1743-422X-3-42-3"/></fig><table-wrap position="float" id="T3"><label>Table 3</label><caption><p>Positivity of PCR amplification of <italic>IE1 </italic>gene on viral DNA of treatment groups.</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="center"><bold>Experiment/Organ</bold></td><td align="center" colspan="2"><bold>A (Day 21 p.i.)</bold></td><td align="center" colspan="2"><bold>B (Day 21 p.i.)</bold></td><td align="center" colspan="2"><bold>C (Day 13–14 p.i.)</bold></td><td align="center" colspan="2"><bold>D (Day 21 p.i.)</bold></td></tr><tr><td></td><td colspan="8"><hr></hr></td></tr><tr><td></td><td align="center" colspan="8"><bold>Group</bold></td></tr><tr><td></td><td colspan="8"><hr></hr></td></tr><tr><td></td><td align="center"><bold>v</bold></td><td align="center"><bold>pv</bold></td><td align="center"><bold>v</bold></td><td align="center"><bold>pv</bold></td><td align="center"><bold>v</bold></td><td align="center"><bold>pv</bold></td><td align="center"><bold>v</bold></td><td align="center"><bold>pv</bold></td></tr></thead><tbody><tr><td align="center">Brain</td><td align="center">0/3</td><td align="center">0/3</td><td align="center">0/3</td><td align="center">0/3</td><td align="center">0/3</td><td align="center">0/3</td><td align="center">0/3</td><td align="center">0/3</td></tr><tr><td align="center">Heart</td><td align="center">0/3</td><td align="center">0/3</td><td align="center">0/3</td><td align="center">0/3</td><td align="center">0/3</td><td align="center">1/3</td><td align="center">0/3</td><td align="center">0/3</td></tr><tr><td align="center">Testes</td><td align="center">0/3</td><td align="center">0/3</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">-</td></tr><tr><td align="center">Ovary</td><td align="center">-</td><td align="center">-</td><td align="center">0/3</td><td align="center">0/3</td><td align="center">0/3</td><td align="center">0/3</td><td align="center">0/3</td><td align="center">0/3</td></tr><tr><td align="center">Uterus</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">5/5</td><td align="center">5/5</td><td align="center">5/5</td><td align="center">5/5</td></tr><tr><td align="center">Neonate</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">12/18</td><td align="center">15/18</td><td align="center">-</td><td align="center">-</td></tr><tr><td align="center">Placenta</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">16/20</td><td align="center">10/10</td></tr><tr><td align="center">Fetus</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">14/20</td><td align="center">8/10</td></tr><tr><td align="center">Embryo*</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">8/8</td></tr></tbody></table><table-wrap-foot><p>Note: Abbreviations: v = virus-infected and pv = virus-infected with immunosuppression.</p><p>* = eroded placenta and developing embryo in uterus at ≤ 7 days of pregnancy.</p><p>0/3 = no detectable positive result over triplicate sample trials in all three rats.</p></table-wrap-foot></table-wrap></sec><sec><title>TEM examination</title><p>TEM revealed virions exhibiting typical herpesvirus morphology in the placenta samples of the infected rats in Experiment D. None of the control groups established similar findings. Figure <xref ref-type="fig" rid="F4">4a</xref> shows the negatively stained naked virion with a size of about 106 nm. The virions were found either naked or enveloped (Figure <xref ref-type="fig" rid="F4">4b</xref>) in ultrathin section and mostly assembled near the mitochondria, golgi apparatus and endoplasm reticulum. The enveloped virions with a size of larger than 200 nm were found in a dense or light and sometime coreless capsid form.</p><fig position="float" id="F4"><label>Figure 4</label><caption><p><bold>Electron micrographs</bold>. (a) negatively-stained herpesvirus-like naked nucleocapsid isolated from placenta sample of an infected immunosuppressed rat of 17-day pregnancy (D; day 21 p.i., × 168k), and (b) ultrathin sectioned placenta of the same rat (D; day 21 p.i.) showing enveloped virions with light capsid (thick arrow) and hollow core (thin arrow) present adjacently to nucleus and mitochondria (× 63k). All bar markers represent 100 nm.</p></caption><graphic xlink:href="1743-422X-3-42-4"/></fig></sec><sec><title>ELISA for antibody detection</title><p>The humoral response of the animals at the end of the study is presented in Figure <xref ref-type="fig" rid="F5">5</xref>. The control groups of all experiments were devoid of RCMV-specific antibody. However, all the infected immunocompetent and immunosuppressed rats seroconverted and their antibody titers were significantly (<italic>p </italic>< 0.05) different to those of control groups. Meanwhile, the immunocompetent groups had significantly (<italic>p </italic>< 0.05) higher mean antibody titers than those of immunosuppressed groups.</p><fig position="float" id="F5"><label>Figure 5</label><caption><p><bold>The mean antibody titers of control and treatment groups in all experiments</bold>. Abbreviations: c = mock-infected; v = virus-infected; pc = mock-infected with immunosuppression and pv = virus-infected with immunosuppression in Experiment A, B and D (day 21 p.i.); C (day 13–14 p.i.).</p></caption><graphic xlink:href="1743-422X-3-42-5"/></fig></sec><sec><title>Fluorescent-antibody technique on buffy coat cells</title><p>The buffy coat cells of the two infected groups of rats in Experiment D were stained positively when observed under fluorescence microscope. Three categories of cells were differentiated based on their sizes, i.e. leukocytes, red blood cells and platelets in a descending order. The positive fluorescence-stained cells were the leukocytes of the infected rats especially those with immunosuppression.</p></sec></sec><sec><title>Discussion</title><p>The RCMV strain ALL-03 was first isolated from placenta and uterus of rats [<xref ref-type="bibr" rid="B19">19</xref>]. There was an urgent need to investigate and confirm the virus capability to infect the fetus. An attempt was made by Priscott and Tyrrell [<xref ref-type="bibr" rid="B12">12</xref>] to isolate RCMV from wild conceptuses. The failure of CPE observation during two weeks of culture concluded no evidence of transplacental infection in the single pregnancy of a naturally infected female [<xref ref-type="bibr" rid="B12">12</xref>]. However, in our study, an analogous procedure using conceptuses from Experiment D (about 7-day of gestation) was carried out. Interestingly, a delayed type CPE resembling characteristics that previously mentioned in our previous study [<xref ref-type="bibr" rid="B19">19</xref>] was observed.</p><p>Like HCMV, RCMV is poorly pathogenic in the immunocompetent host. The transient suppression in host immunity induced by cyclophosphamide is necessary for the induction of disease and the severity of disease always reflects the level of virus localization in the organs. The incubation time of symptomatic infection varied but commonly started at day 6 and onwards. This was similar to a previous study, which reported the emergence of clinical signs and absence of mortality in the immunocompromised groups [<xref ref-type="bibr" rid="B13">13</xref>]. The pregnant rats (Experiment C and D) seem to have partial immunosuppressive effect similar to that of other groups receiving cyclophosphamide as they were more permissive to RCMV infection than non-pregnant rats. Gould and Mims [<xref ref-type="bibr" rid="B20">20</xref>] showed that the virus could be reactivated during pregnancy. As a result of immunosuppression caused by the pregnancy alone or in conjunction with RCMV, the virus may have a better conducive environment for growth. In fact, one characteristic of CMVs is that the infection may have an immunosuppressive effect to the host during the acute phase. This has been observed in man, mice and rats [<xref ref-type="bibr" rid="B13">13</xref>,<xref ref-type="bibr" rid="B21">21</xref>,<xref ref-type="bibr" rid="B22">22</xref>].</p><p>Disease symptoms correlated well with the presences of infectious virus, viral antigen and DNA, which were found highest concentration in uterus, placenta, embryo and fetus; abundantly in lung, kidney, spleen and liver; less in salivary gland; even rare in heart (one case) but none in brain, ovary and testes. The detection of the RCMV in the spleen and liver was consistent with that of many previous studies [<xref ref-type="bibr" rid="B12">12</xref>,<xref ref-type="bibr" rid="B13">13</xref>,<xref ref-type="bibr" rid="B23">23</xref>,<xref ref-type="bibr" rid="B24">24</xref>]. The incidence of splenomegaly coincided with detection of RCMV in spleen. The finding is similar to that of mouse model [<xref ref-type="bibr" rid="B25">25</xref>]. The occurrence of RCMV immunoreactive monocytes and macrophages with characteristic inclusions in the spleen is consistent with the symptomatic infection. This parallels the situation in man where the involvement of the spleen is common in CMV infections [<xref ref-type="bibr" rid="B26">26</xref>]. The finding of RCMV particles in the liver parenchyma of immunocompromised rats is similar to that observed in HCMV infections, whereby the occurrence of hepatitis in immunocompromised patient is frequent [<xref ref-type="bibr" rid="B27">27</xref>]. The findings of the present study do closely resemble the pathological changes in the HCMV hepatitis, for example, the extensive liver damage with numerous inclusion bodies in hepatocytes, Kupffer cells as well as focal liver cell necrosis [<xref ref-type="bibr" rid="B28">28</xref>,<xref ref-type="bibr" rid="B29">29</xref>]. In our study, more viral antigens detected in the tubular epithelia than the glomeruli contrast to a previous study of RCMV strain Maastricht which localized predominantly in glomeruli and hardly ever in the tubular epithelia [<xref ref-type="bibr" rid="B24">24</xref>]. The finding that the renal capsule contained immunoreactive cells mimics that of the CMV infection in humans and rats [<xref ref-type="bibr" rid="B24">24</xref>].</p><p>Pneumonitis is the leading cause of death in CMV-infected transplant patients [<xref ref-type="bibr" rid="B14">14</xref>]. In RCMV-infected rats numerous immunoreactive cells were found in the lungs, including alveolar macrophages and interstitial mononuclear cells, resembling the histopathology of HCMV induced pneumonitis. Such damages caused by extensive virus replication in rats injected with cyclophosphamide are similar to that observed in the mouse model [<xref ref-type="bibr" rid="B30">30</xref>]. The virus persistence in the salivary glands resembles the typical characteristic of CMV in rat [<xref ref-type="bibr" rid="B23">23</xref>], mouse [<xref ref-type="bibr" rid="B25">25</xref>], guinea pig [<xref ref-type="bibr" rid="B31">31</xref>] and human [<xref ref-type="bibr" rid="B32">32</xref>]. The salivary gland is believed to be the principal route by which the virus is spread within the population of susceptible hosts [<xref ref-type="bibr" rid="B33">33</xref>]. The absence of a case in Experiment C may due to the fact that infectious RCMV (Maastricht strain) in salivary glands is detected at a later time than in all other organs, starting at day 14 p.i. [<xref ref-type="bibr" rid="B33">33</xref>]. In addition, the subcutaneous route and duration of infection (13–14 days) carried out in Experiment C would most probably decrease the severity of the disease. The submandibular gland was the preferred organ for tropism of the virus. These characteristics conformed to the previous study of Kloover <italic>et al </italic>[<xref ref-type="bibr" rid="B33">33</xref>].</p><p>The detection of viral antigen was not success in brain, heart, testes and ovary. Only one heart sample was found to contain viral DNA. This positive result was, most likely, due to contamination from infected blood cells. These four organs were reported to be involved in CMV infection in previous studies. A similar work studying acute infection of RCMV conducted previously [<xref ref-type="bibr" rid="B24">24</xref>] showed the brain tissue was negative for RCMV antigen. In contrast, a significant infection in brain was demonstrated in mouse model [<xref ref-type="bibr" rid="B34">34</xref>]. In fact, CNS involvement is a frequent feature of congenital infection [<xref ref-type="bibr" rid="B35">35</xref>]. MCMV infections were reported to be associated in the development of myopericarditis and dystrophic cardiac calcification [<xref ref-type="bibr" rid="B36">36</xref>] but cardiac infection in rat model was transient [<xref ref-type="bibr" rid="B13">13</xref>]. The recovery of infectious virus from sperm [<xref ref-type="bibr" rid="B37">37</xref>] and the detection of latent viral genomes in the prostate gland, testes, and spermatogonia of infected mice suggested that transmission of virus was by sexual contact [<xref ref-type="bibr" rid="B38">38</xref>,<xref ref-type="bibr" rid="B39">39</xref>]. With the congenital infection, inclusion-bearing cells are found also in testes and ovary after reactivation of latent infection. Nevertheless, the tropism of CMV in these germ line organs was in more chronic phase than the visceral organs [<xref ref-type="bibr" rid="B40">40</xref>]. Thus, it is reasonable to argue that the viral antigen as well as DNA of these germ line organs was untraceable.</p><p>The presence of RCMV infection in the endometrium of uterus regardless of pregnancy or different stages of pregnancy suggested that the uterus is one of the target organs. The current finding showed RCMV infection localized in different sites of uterus of different rats treated identically. One explanation might be that the different degree of susceptibility of an individual to the infection by which is largely affected by the host's physiology and immune response. Besides, CMV is evident by its asynchronous development <italic>in vitro </italic>[<xref ref-type="bibr" rid="B41">41</xref>]; it might also happen <italic>in vivo</italic>. The uterine infection extends to adjacent cell type during more advanced dissemination, i.e. from stromal cells to epithelial cells. This observation is similar to CMV infection in human and contiguous endometrial cells dissemination plays an important role in congenital infection where HCMV can establish active and latent infection to the placenta subsequently [<xref ref-type="bibr" rid="B3">3</xref>].</p><p>High un-natural dosage of infection at titer 10<sup>6 </sup>TCID<sub>50 </sub>per rat has no effect on abortion and severe fetus wastage as observed in Experiment C and D. These findings contrast to guinea pig CMV infection by which the highest rates of fetus resorption/abortion and mortality are correlated well with the increase of infection dosage [<xref ref-type="bibr" rid="B42">42</xref>]. This might suggest that RCMV strain ALL-03 is either a benign virus for the offsprings naturally or somewhat attenuated throughout the subsequent tissue culture passages or when infecting a different rat strain. If the attenuation of tissue culture passage is the case, it can be reversed by a few <italic>in vivo </italic>passages and the pathogenicity of this 'virulent' virus can be determined in future investigation. On the other hand, one explanation, which is more fascinating, might be that the current experiments performed using a virus isolated from the black rat, <italic>Rattus rattus diardii</italic>, in a laboratory rat, <italic>R. norvegicus</italic>. This different host strain may contribute to the mild effects of the fetal and neonatal infections. Nevertheless, a definite answer for this speculation cannot be given presently since we realize that there is no SPF colony of <italic>R. rattus </italic>available for the moment.</p><p>Although virus infection in Experiment C was conducted via s.c. route (less infective than i.p. route) and in shorter incubation period (about 13–14 days), the signs of infection were closely resembling those of Experiment D. These indicate maternal virus dissemination had started earlier than 2 weeks time. The <italic>in utero </italic>virus transmission was more promising when one-day old neonates and conceptuses (fetuses and embryos) had already harbored the virus. In fact, there was no probable virus transmission from the female rats to them perinatally or postnatally by close contact. This is due to the slow growth of RCMV which is normally detected in organs such as kidney and salivary gland starting on day 4 and 10 p.i., respectively [<xref ref-type="bibr" rid="B12">12</xref>,<xref ref-type="bibr" rid="B13">13</xref>]. Therefore, it is believed that the virus transmitted either by direct passage of the virus across the placenta to the fetus or through germ cells as proposed by Brautigam and Oldsone [<xref ref-type="bibr" rid="B43">43</xref>], Chantler <italic>et al </italic>[<xref ref-type="bibr" rid="B44">44</xref>], and Osborn [<xref ref-type="bibr" rid="B45">45</xref>]. However, the precise localization of the virus in tissue section for IIP test had elucidated that the infections occurred in placenta, uterus, embryo, fetus and neonate, but not in testes and ovary. The presence of infectious viruses in the aforementioned sites suggests the RCMV infection was successive and responsible for the vertical transmission. Furthermore, electron microscopy showing visible typical herpesvirus-like particles in infected placenta, had further confirmed the transplacental transmission route of RCMV strain ALL-03 without doubt. Generally, the frequency and concentration of virus infection were predominantly in the uterus, placenta and offspring differing from those reported previously in other RCMVs. It is believed that this unique infection preference was indeed the nature of ALL-03 virus.</p><p>The presence of CMV infection in the placental parenchyma and membrane had been confirmed in a previous study [<xref ref-type="bibr" rid="B5">5</xref>]. It is likely that CMV or CMV DNA could be detected in the villi, including the mesenchyme and trophoblasts, extravillous trophoblast, and decidual cells. Consistent with their study [<xref ref-type="bibr" rid="B5">5</xref>], the IIP staining in our study showed immunogenic sites containing RCMV antigen were the decidual basalis, junctional and labyrinth zones. The likely cells involved were the trophoblast and decidual cells. The placenta of earlier gestation (about 7 days) showed more signs and intensities of infection than that with lengthier gestation period. Furthermore, the intensity of infection in placenta surpassed that in fetal tissues. However, at later stage (> 14 days) of pregnancy, the fetal tissues such as liver and kidney showed a more significant infection. The most likely explanation of the events might be the differences in the degree of permissiveness to RCMV in various tissues during development. The virus may subsequently infect the fetus following direct crossing of the labyrinth zone of placenta after a successive virus replication period.</p><p>The exact mechanism of how RCMV crosses the placenta to infect fetus has yet to be elucidated. It is either caused by viremia, transportation of the virus by maternal leukocytes entering the placenta, direct passage of the virus from uterus into the placenta or direct invasion of placenta and fetus. The preliminary study employing immunofluorescence staining on buffy coat cells of rats in Experiment D found that the infected rats suffered a leukocyte associated-viremia. The circulation of infected leukocytes in the blood had most probably promoted the spread of the virus throughout the animal body. This finding was in agreement with that of Bruggeman <italic>et al </italic>[<xref ref-type="bibr" rid="B13">13</xref>]. Similar observation had also been made for HCMV [<xref ref-type="bibr" rid="B46">46</xref>]. During the viremic phase, the virus circulates and disseminates as it has been carried in leukocytes [<xref ref-type="bibr" rid="B47">47</xref>].</p><p>The findings obtained from Experiment D with absolute uterine infection in relation to 70–100% of placental infection illuminate the important intersection of maternal uterus for the congenital infection. Indeed, the earlier <italic>in vitro </italic>study in which leukocytes infected with a clinical HCMV strain VR1814 (thus reproducing the <italic>in vivo </italic>phase of acute viremia) was used to infect either explants of floating and anchoring villi or differentiating cytotrophoblast cells, no infection was observed [<xref ref-type="bibr" rid="B6">6</xref>]. On the other hand, the same study showed that HCMV-infected leukocytes could productively infect uterine endothelial cells, which in turn, were able to transmit the infection to cytotrophoblast cells. In this context, the infected anchoring villi, which extended into the uterine wall passing the infection to the placenta, were well demonstrated in our study. Concurrent to the aforementioned <italic>in vitro </italic>model and our findings, congenital infection is acquired only during primary maternal infection whereby uterine infection must take place preconceptionally or periconceptionally.</p><p>Another intriguing aspect of the natural history of HCMV infection during pregnancy concerns the transmission rate during different gestation period. In particular, while primary HCMV infection acquired either before or around conception carries the lowest risk of transmission [<xref ref-type="bibr" rid="B48">48</xref>], maternal infections acquired during the first and second trimester of gestation can be transmitted at a similar rate (approximately 45%). On the other hand, during the third trimester, maternal infection has the highest probability of being transmitted to the fetus (78.6%). These data clearly indicate that: (i) the virus is transmitted efficiently from mother to fetus despite the presence of an innate barrier; (ii) mechanisms of protection are more effective during the first two-thirds of gestation, becoming less effective in late pregnancy [<xref ref-type="bibr" rid="B49">49</xref>]. In parallel to these reports, the dams of Experiment D and Experiment C by which infections occurred in preconception and during the midterm (about 10 days) showed transmission to offsprings in 65–81% and 59–84%, respectively. It agreed that the infection to offsprings was more effective in dams of Experiment C occurring in earlier time interval (day 13–14 p.i.). Since placental infection has been detected either in the presence or absence of fetal infection, the placenta is considered as the most important site of either protection (by sheltering the fetus from CMV infection) or transmission (by acting as a viral reservoir and allowing the infection to reach the fetal compartment). Nevertheless, whenever an infection of the fetus occurred, virus could be found in the associated placenta at different degree of infection. Moreover, the discrepancy of the number of positive virus infection in placenta to that of fetus was only 10–30%. Hence, it is suggested that the placenta more likely serves as a reservoir rather than protective barrier in which the virus replicates first prior reaching the embryo or fetus in our study. As discussed earlier, the human placenta is not an effective barrier to HCMV transmission in the same way [<xref ref-type="bibr" rid="B3">3</xref>].</p></sec><sec><title>Conclusions</title><p>The current study exhibits a widespread systemic RCMV infection. The maternal viremia, uterine infection, placental infection and direct dissemination to the fetus are the most likely sequence of events leading to congenital infection after a primary maternal infection mimicking the features of congenital CMV infection in human. We believe that RCMV strain ALL-03 has the potentials to provide predictable information on the pathogenesis and manifestations of congenital CMV infection, rational designs of new antiviral therapies as well as <italic>in utero </italic>vaccine to specifically prevent prenatal infection in future investigations.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>Preparation of virus working stock and hyperimmune serum (HIS)</title><p>Virus stock of RCMV strain ALL-03 was prepared and titrated by mean of TCID<sub>50</sub>/ml prior to animal inoculation. Hyperimmune serum (HIS) was prepared in mice according to standard procedure using heat-inactivated purified RCMV suspension (10<sup>7 </sup>TCID<sub>50</sub>/ml). The antibody titers were determined using indirect enzyme-linked immunosorbent assay (ELISA).</p></sec><sec><title>Design of the experiments</title><p>Two-month old SPF Sprague-Dawley rats were assigned into four different experiments (A, B, C and D). The rats from each experiment were subdivided into immunocompetent and immunosuppressed groups. Each immunosuppressed rat was induced by subcutaneous (s.c.) injection with cyclophosphamide at a dosage of 40 mg, a day before the virus inoculation. All treatment groups were infected with 10<sup>6 </sup>TCID<sub>50 </sub>RCMV suspensions in either intraperitoneal (i.p.) or s.c. route. Five rats were allotted for each treatment group, whereas, three rats which inoculated with PBS in similar route served as the control group. Blood samples were collected before and after the experiments for antibody titration. The animals were observed twice daily for clinical signs and mortality.</p><p>The male and non-pregnant female rats employed in Experiment A and B respectively, were inoculated i.p. route with RCMV suspensions. These rats were sampled at day 21 p.i. The brain, salivary gland, heart, lung, spleen, liver, kidney, testes and uterus were processed for hematoxylin and eosin (H&E) staining, and indirect immunoperoxidase (IIP) test.</p><p>In Experiment C, female rats of about 10-day pregnancy (determined by vaginal plug observation) were inoculated via s.c. route. The inoculation was carried out in s.c. route rather than i.p. in order to prevent abortion that may be caused by the injection. The sampling was carried out at one day post-parturition of the neonates, i.e. day 13–14 p.i. of the dams. The brain, salivary gland, heart, lung, spleen, liver, kidney, uterus, ovary and the one-day old neonatal tissues (salivary gland, liver and kidney) were subjected to H&E staining and IIP test. Additional neonatal tissues and uterus were assigned for virus assay to isolate the infectious virus as determined by cytopathic effect (CPE) development (as described in Loh <italic>et al </italic>[<xref ref-type="bibr" rid="B19">19</xref>]); protein blotting as well as polymerase chain reaction (PCR) amplification.</p><p>In Experiment D, non-pregnant female rats were inoculated via i.p. route. The rats of each group were housed together with a male rat and observed for pregnancy. The pregnant rats were sacrificed at day 21 p.i., i.e. just before delivery. The salivary gland, heart, lung, spleen, liver, kidney, uterus, ovary, placenta and fetal tissues (liver and kidney) were prepared for H&E staining and IIP test. Additional uterus, placenta and fetal tissues were further tested by virus assay, protein blotting and PCR analyses. The remaining placenta was processed for transmission electron microscopy (TEM) examination.</p></sec><sec><title>Indirect immunoperoxidase (IIP) test</title><p>After deparaffinization, the sections were blocked for endogenous peroxidase and covered with 1% SDS/PBS for 5 minutes. Following washing thrice with PBS containing 1% Triton X-100 (PBSTx), the sections were immersed in 5% BSA/PBSTx for 1 hour. The diluted mouse HIS (1:200) with additional 2% normal rat serum (only for neonatal tissues) was added and incubated for 1 hour at 37°C. The sections were washed and incubated with diluted peroxidase-conjugated goat anti-mouse IgG (1:250). After stopping the stain development of DAB substrate (KPL), the sections were counterstained with hematoxylin, washed in dH<sub>2</sub>O, dehydrated and then mounted.</p></sec><sec><title>Protein blotting</title><p>The test strips (Millipore) pre-treated with transfer buffer, were blotted with purified intracellular virus (from tissue homogenates). The air-dried test strips were immersed in 5% BSA/PBS and then incubated with diluted mouse HIS (1:500) for 1 hour. After washing in PBST (0.2% Tween 20), the test trips were incubated with diluted peroxidase-conjugated goat anti-mouse IgG (1:2000) for 1 hour and then washed again. DAB substrate was added. The test strips were rinsed in dH<sub>2</sub>O and then air-dried.</p></sec><sec><title>Polymerase chain reaction (PCR)</title><p>The sequences of the gene-specific primers flanking on immediate-early 1 (<italic>IE1</italic>) gene region of RCMV strain ALL-03 were 5'-CACAGAGATCTCACTAACCTGCCACCTATAACCAC-3' (Forward) and 5'-TCCAGCAGACTTCTGTATCCTGATTCAAG-3' (Reverse). The PCR reaction contained 100 ng DNA extracted from each tissue sample, 0.5 μM of each primer, 1X optimized buffer, 0.2 mM dNTP mix, 2 unit of DyNAzyme™ II DNA polymerase (Finnzymes) and nuclease-free H<sub>2</sub>O. The protocol included an initial denaturation step at 95°C for 5 minutes, 40 cycles of 1-minute denaturation at 94°C, 30-second annealing at 69°C and 1-minute extension at 72°C. This was followed by a final extension step at 72°C for 1 minute.</p></sec><sec><title>Transmission electron microscopy (TEM) examination</title><p>Intracellular virus from placenta was purified and subjected to negative staining. For ultrathin sectioning, the placenta was processed accordingly to the procedures described in Loh <italic>et al </italic>[<xref ref-type="bibr" rid="B19">19</xref>] and subjected to TEM examination.</p></sec><sec><title>Indirect enzyme-linked immunosorbent assay (ELISA)</title><p>Pre-immune and hyperimmune sera were used as negative and positive controls, respectively. Microtiter plates (Dynatech) were coated with purified virus (3.2 μg/ml). Reaction wells were rinsed thrice with PBST (0.05% Tween 20 in PBS) and blocked with 5% BSA/PBST. After incubation with diluted test sera (1:50) at 37°C for 2 hours, the bound antibodies were reacted with diluted peroxidase-conjugated goat anti-rat IgG (1:2000) for another 2 hours. Following washings, TMB substrate (KPL) was added. The absorbance of a sample was determined using an ELISA reader.</p></sec><sec><title>Statistical analysis</title><p>Data were expressed as mean ± SD, and statistical analysis was performed using two-tailed Student's <italic>t</italic>-test. Differences between groups were considered statistically significant at <italic>P </italic>< 0.05.</p></sec><sec><title>Fluorescent-antibody technique on buffy coat cells</title><p>A test to assess cell-associated viremia was conducted on buffy coat cells of animals in Experiment D. The buffy coat cells were fixed on a chamber slide and subjected to an indirect immunofluorescence (IIF) procedure as mentioned in Loh <italic>et al </italic>[<xref ref-type="bibr" rid="B19">19</xref>] with a few modifications, i.e. using mouse HIS at dilution 1:200 (in 1% BSA/PBS) and FITC-conjugated goat anti-mouse IgG at dilution 1:250. The normal mouse sera were used as negative controls.</p></sec></sec><sec><title>Abbreviations</title><p>cytomegalovirus (CMV), cytopathic effect (CPE), enzyme-linked immunosorbent assay (ELISA), hematoxylin and eosin (H&E), hyperimmune serum (HIS), immediate-early 1 (<italic>IE1</italic>), indirect immunofluorescence (IIF), indirect immunoperoxidase (IIP), intraperitoneal (i.p.), murine cytomegalovirus (MCMV), polymerase chain reaction (PCR), post-infection (p.i.), rat cytomegalovirus (RCMV), rat embryonic fibroblast (REF), subcutaneous (s.c.), transmission electron microscopy (TEM).</p></sec><sec><title>Competing interests</title><p>The author(s) declare that they have no competing interests.</p></sec><sec><title>Authors' contributions</title><p>HSL participated in the experimental design, performed all experiments and drafted the manuscript. MAML participated in the experimental design and coordination and helped to draft the manuscript. SOAR conceived of the study and participated in its design and interpretation of data. LJK participated in part of the experiments and assisted in post-mortem investigation. All authors read and approved the final manuscript.</p></sec> |
Role of CD8+ cells in controlling replication of nonpathogenic Simian Immunodeficiency Virus SIVmac1A11 | <p>Infection of macaques with the avirulent molecular clone SIVmac1A11 results in transient low viremia and no disease. To investigate if this low viremia is solely due to intrinsic poor replication fitness or is mediated by efficient immune-mediated control, 5 macaques were inoculated intravenously with SIVmac1A11. Three animals that were depleted of CD8+ cells at the start of infection had more prolonged viremia with peak virus levels 1 to 2 logs higher than those of 2 animals that received a non-depleting control antibody. Thus, CD8+ cell-mediated immune responses play an important role in controlling SIVmac1A11 replication during acute viremia.</p> | <contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Van Rompay</surname><given-names>Koen KA</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>kkvanrompay@ucdavis.edu</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Blackwood</surname><given-names>Emily J</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>ejblackwood@ucdavis.edu</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Landucci</surname><given-names>Gary</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>glanducc@uci.edu</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Forthal</surname><given-names>Don</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>dnfortha@uci.edu</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Marthas</surname><given-names>Marta L</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I3">3</xref><email>mlmarthas@ucdavis.edu</email></contrib> | Virology Journal | <sec><title/><p>Simian immunodeficiency virus (SIV) infection of macaques has proven useful for modeling HIV disease pathogenesis and intervention strategies [<xref ref-type="bibr" rid="B1">1</xref>-<xref ref-type="bibr" rid="B3">3</xref>]. While infection of macaques with most SIV isolates results eventually in an AIDS-like disease, there are also attenuated isolates and clones. SIVmac1A11 is a molecular clone originally derived from a virus isolate from an SIV-infected macaque that was also the source of virulent uncloned SIVmac251 isolates [<xref ref-type="bibr" rid="B4">4</xref>,<xref ref-type="bibr" rid="B5">5</xref>]. Although the kinetics are slower than for other isolates, SIVmac1A11 replicates well <italic>in vitro </italic>and is highly cytopathogenic (with induction of syncytia) in T-cell lines and rhesus macaque peripheral blood mononuclear cells (PBMC); SIVmac1A11 replicates well in macrophage cultures [<xref ref-type="bibr" rid="B6">6</xref>]. In early studies, it was observed that SIVmac1A11 inoculation of juvenile macaques resulted in transient viremia and no disease, even after prolonged follow-up for more than 12 years ([<xref ref-type="bibr" rid="B4">4</xref>]; unpublished observations). Subsequent studies documented that SIVmac1A11 inoculation of fetal and newborn macaques also resulted in transient viremia and no disease [<xref ref-type="bibr" rid="B7">7</xref>,<xref ref-type="bibr" rid="B8">8</xref>]. SIVmac1A11 has a tissue distribution distinct from that of virulent isolates [<xref ref-type="bibr" rid="B9">9</xref>].</p><p>Because of these unique properties, SIVmac1A11 has proven useful to study determinants of viral virulence. The genome of SIVmac1A11 has been sequenced, and recombination experiments revealed that differences in more than one region of the viral genome were responsible for the lack of virulence [<xref ref-type="bibr" rid="B5">5</xref>,<xref ref-type="bibr" rid="B10">10</xref>]. SIVmac1A11 has also shown promise as a live-attenuated vaccine in both infant and juvenile/adult macaques [<xref ref-type="bibr" rid="B10">10</xref>-<xref ref-type="bibr" rid="B13">13</xref>].</p><p>The transient low-level viremia (peak levels ≤ 4 to 5 log RNA copies per ml plasma) that results from SIVmac1A11 infection suggests either poor intrinsic replication fitness <italic>in vivo </italic>and/or relatively effective immune control. CD8+ cell depletion experiments (via administration of monoclonal antibody) have demonstrated the important role of CD8+ cell-mediated immune responses in controlling acute and chronic viremia with virulent SIV isolates (such as SIVmac251; [<xref ref-type="bibr" rid="B14">14</xref>]) and chronic viremia with the attenuated clone SIVmac239Δnef [<xref ref-type="bibr" rid="B15">15</xref>]; however, CD8+ cell depletion had no detectable effect on viremia in animals chronically infected with the more attenuated clone SIVmac239Δ3 [<xref ref-type="bibr" rid="B16">16</xref>] or with SIVmac1A11 (unpublished data). To our knowledge, no CD8+ cell depletion experiments have been performed during acute infection with nonpathogenic SIV isolates.</p><p>Accordingly, we sought to determine the role of CD8+ cell-mediated immune responses on acute SIVmac1A11 viremia. Animals in this study were juvenile rhesus macaques (Macaca mulatta; ~1 year of age), housed in accordance with American Association for Accreditation of Laboratory Animal Care Standards with strict adherence to the "Guide for the Care and Use of Laboratory Animals" [<xref ref-type="bibr" rid="B17">17</xref>]. When necessary, the animals were immobilized with ketamine HCL (Parke-Davis, Morris Plains, New Jersey) 10 mg/kg injected intramuscularly.</p><p>All 5 macaques were inoculated intravenously with a high dose of SIVmac1A11 (5 × 10<sup>5 </sup>50% tissue culture infectious doses, grown in CEMx174 cells). Immediately before virus inoculation, 3 animals were depleted of CD8+ cells via administration of the anti-CD8 antibody cM-T807 at a dose of 50 mg/kg body weight (administered slowly intravenously); the same dose was repeated 3 weeks later. This dosage regimen, which is higher than the regimen used in previous CD8+ cell depletion studies [<xref ref-type="bibr" rid="B14">14</xref>,<xref ref-type="bibr" rid="B18">18</xref>], was selected because it gives more prolonged depletion of CD8+ cells (K. Reimann, personal communication). In the current study, CD8+ cells (both CD8+CD3+ T lymphocytes and CD8+CD3- NK cells) in peripheral blood were undetectable or low (< 1% of lymphocytes; ≤ 40 cells per μl blood) for 21 to 35 days after treatment (Fig. <xref ref-type="fig" rid="F1">1B,C</xref>). The remaining 2 animals received a control (i.e., non-depleting) human immunoglobulin preparation (Aventis Gammar-P I.V.) at the same dosage regimen (50 mg/kg at 0 and 3 weeks).</p><fig position="float" id="F1"><label>Figure 1</label><caption><p><bold>Effect of CD8+ cell depletion on SIVmac1A11 infection: viral and immunologic parameters</bold>. Five animals were inoculated with SIVmac1A11 at time zero. Three animals were CD8+ cell depleted via administration of cM-T807 while the other 2 animals received control antibody. (A) Viral RNA levels in plasma (measured by bDNA assay, with a limit of detection of 125 copies/ml; [18]). Results from virus isolation from 1 million PBMC, using CEMx174 cells and p27 measurement [34] are given as positive (+) or negative (-). The absolute counts of CD8+CD3+ T lymphocytes, CD8+CD3- NK cells and CD4+CD3+ T lymphocytes were measured according to flow cytometry techniques described previously [18], and are presented in graphs B through D, respectively. (E) SIV-specific IgG titers measured by a whole SIV ELISA [29]; although the CD8+ cell depleted animals made a faster response than the undepleted animals, from week 6 onwards, both animal groups had similar antiviral IgG titers (1: 6,400 to 1: 25,600). (F) Antiviral activity of plasma collected at 17 days after SIVmac1A11 inoculation as measured in a ADCVI assay, described in detail elsewhere (Forthal <italic>et al</italic>., submitted for publication). Briefly, CEMx174 cells were infected with SIVmac1A11 at a MOI of 0.01; 48 hours later, cells were plated in 96-well plates at 50,000 cells per well. Plasma samples (including negative and positive control samples) were added at a 1:100 dilution and human PBMC effector cells were added to obtain an effector:target cell ratio of 10:1. Five days later, SIV p27 was measured in supernatant fluid using a commercially available ELISA (Zeptometrix Corporation, Buffalo, NY). Percent inhibition by the plasma samples collected on day 17 was calculated relative to the level of virus replication in the presence of plasma collected on day zero (before SIVmac1A11 inoculation); the presented values represent mean +/- SEM of 4 separate assays (with effector PBMC of 4 different donors). In the absence of effector cells, no significant inhibition (≤ 11%) was observed (data not shown).</p></caption><graphic xlink:href="1743-422X-3-22-1"/></fig><p>The 2 control-antibody treated animals had peak plasma viral RNA levels of 4 to 7 × 10<sup>3 </sup>copies/ml at 3 days after virus inoculation (Fig. <xref ref-type="fig" rid="F1">1A</xref>). For one animal (number 35391), a second smaller peak of viremia was observed on day 17. The levels of viremia in these 2 control animals are thus similar to those described previously for SIVmac1A11-infected juvenile macaques [<xref ref-type="bibr" rid="B19">19</xref>]. The 3 CD8+ cell-depleted animals had viral RNA levels during the first 7 days that were indistinguishable from those of the control animals, suggesting that during these early stages, CD8+ cells had no detectable role in controlling SIVmac1A11 replication. However, after an initial decline, viral RNA levels in the CD8+ cell-depleted animals increased from day 10 onwards and reached peak levels of 45,000 to 790,000 on day 17; these values were 1–2 log higher than those of the control animals (p = 0.015, two-tailed t-test comparing day 17 values, and area-under-the curve values for day 0 to 35) and only ~1–2 log lower than peak viremia levels observed with the pathogenic molecular clone SIVmac239 [<xref ref-type="bibr" rid="B20">20</xref>-<xref ref-type="bibr" rid="B22">22</xref>]. Despite this higher viremia in the CD8+ cell-depleted animals, there were no significant changes in CD4+CD3+ T lymphocyte counts in peripheral blood (Fig. <xref ref-type="fig" rid="F1">1D</xref>); this study was not designed to monitor CD4+CD3+ T lymphocyte levels in gut-associated lymphoid tissue. Plasma viral RNA levels declined upon the return of CD8+ cells and became undetectable from 28 to 35 days of infection onwards throughout the rest of the observation period (> 6 months). These results indicate that CD8+ cells play a major role in controlling SIVmac1A11 replication because in their absence, peak viremia was higher and the acute viremia phase was significantly prolonged.</p><p>Because the cM-T807 antibody depletes both CD8+ T lymphocytes as well as NK cells, the relative contribution of each cell type could not be determined. CD8+ T cells and NK cells inhibit virus replication <italic>in vitro </italic>through a variety of mechanisms, including cytolytic and non-cytolytic pathways [<xref ref-type="bibr" rid="B23">23</xref>-<xref ref-type="bibr" rid="B25">25</xref>]. Most NK cells also have the low-affinity Fc-gamma III receptor (CD16), which triggers antibody-dependent cellular cytotoxicity (ADCC) and antibody-dependent cell-mediated virus inhibition (ADCVI). ADCVI is similar to ADCC but is a measure of virus inhibition, rather than target cell cytotoxicity. ADCVI has been observed <italic>ex vivo </italic>with serum and effector cells from HIV-infected humans and SIV-infected macaques [<xref ref-type="bibr" rid="B26">26</xref>](Forthal <italic>et al</italic>, manuscript submitted). Forthal <italic>et al</italic>. have also demonstrated that ADCVI-mediating antibodies can be found early during HIV-1 infection and reduce HIV-1 yield both by lysis of infected target cells and by the release of beta-chemokines from NK effector cells [<xref ref-type="bibr" rid="B27">27</xref>,<xref ref-type="bibr" rid="B28">28</xref>]. In the current study, the SIVmac1A11-infected animals had detectable antiviral IgG antibodies (as measured by whole SIV ELISA techniques; [<xref ref-type="bibr" rid="B29">29</xref>]) at ~2 weeks of infection, and the CD8+ cell-depleted animals had a more rapid increase in antibody titers, possibly due to more antigenic stimulation (Fig. <xref ref-type="fig" rid="F1">1E</xref>). Early plasma samples were also tested for ADCVI activity; pronounced inhibition (> 70%) was observed in plasma collected at 17 days of infection at a 1:100 dilution in all animals (Fig. <xref ref-type="fig" rid="F1">1F</xref>). Thus, some of the loss of viremia control following CD8+ cell depletion could be due to the loss of CD8+ NK cells that would likely serve as ADCVI effector cells.</p><p>Levels of interleukin-12 and interferon-α were measured in plasma using commercial ELISA-kits (monkey IL-12 ELISA, U-CyTech, Utrecht, the Netherlands; human interferon-α ELISA kit, PBL Biomedical Laboratories, Piscataway, NJ). Although variable levels were detected for both cytokines, there was no correlation with virus levels (data not shown).</p><p>In conclusion, this experiment demonstrated that the acute low-level viremia of SIVmac1A11 which is observed following inoculation of untreated animals cannot be explained solely by poor intrinsic replication fitness of the virus; instead, immune responses that are dependent on CD8+ cells limit the magnitude and duration of acute viremia. Viremia of SIVmac1A11 has always been observed to be transient (~2–6 weeks), even following inoculation of fetal and newborn macaques [<xref ref-type="bibr" rid="B7">7</xref>,<xref ref-type="bibr" rid="B8">8</xref>]. This indicates that these antiviral immune responses are not abrogated or prevented from emerging during acute SIVmac1A11 viremia and suggests that there is relatively little or no virus-induced immunosuppression. Rather, the anti-SIVmac1A11 immune responses appear able to induce a long-term asymptomatic infection [<xref ref-type="bibr" rid="B10">10</xref>]. This is in contrast to infection with virulent SIV isolates, for which irreversible damage to the immune system appears to occur early during the course of infection [<xref ref-type="bibr" rid="B9">9</xref>,<xref ref-type="bibr" rid="B30">30</xref>-<xref ref-type="bibr" rid="B32">32</xref>]. Accordingly, further experiments that combine avirulent strains such as SIVmac1A11 infection with selective depletions of immune cell populations may prove to be a useful and sensitive model to further unravel precisely the immune responses that are important to control viremia, but that may be difficult to detect during infection with virulent isolates. Attempts to boost or preserve such immune responses may lead to immunotherapeutic strategies that are more effective in achieving long-term control on viremia of virulent virus isolates, including HIV-1.</p></sec><sec><title>Competing interests</title><p>The author(s) declare that they have no competing interest.</p></sec><sec><title>Authors' contributions</title><p>KVR designed and coordinated the study, and drafted the manuscript; EB performed and analyzed viral and immunological assays; GL and DF performed and analyzed the ADCVI assays; MM participated in the design and interpretation of the study. All authors helped with and approved the final manuscript.</p></sec> |
Prevalence of visual impairment in relation to the number of ophthalmologists in a given area: a nationwide approach | <sec><title>Background</title><p>Sociological and economic risk factors of visual impairment have never been described in France at a national level as the association between the number of ophthalmologists per inhabitant and visual impairment prevalence.</p></sec><sec sec-type="methods"><title>Methods</title><p>Two national surveys were pooled. First, 2075 institutions were selected at random from the French Health Ministry files. Second, a random, stratified sample of 356,208 citizens living in the community was selected. Blindness and low vision (LV) prevalence rates were estimated by age and gender. Geographical equities were estimated by logistic regression adjusted on age and occupational category. The association between ophthalmologist density and visual impairment prevalence rate was estimated <italic>per </italic>region. Interviews were completed with 14,603 (94.9%) of 15,403 randomly selected subjects in institutions, and 16,945 (77.8%) of 21,760 randomly selected subjects in the community. Three groups were defined from the interviews: low vision, blind, and control.</p></sec><sec><title>Results</title><p>Prevalence rates were LV 2.08% and blindness 0.12%. Both rates increased exponentially with age. No major difference was found with gender. Injury was the declared reason for both LV (12%) and blindness (12%). Large regional differences in prevalence persisted for LV after adjustment on age and occupation (ORs: 0.35 to 2.10), but not for blindness. Regions with ophthalmologists below the national <italic>per capita </italic>average were usually those with higher LV prevalence.</p></sec><sec><title>Conclusion</title><p>An inverse correlation was found between ophthalmologist number and LV prevalence rates for subjects of similar age and socio-professional category. This denoted possible inequity in the provision of healthcare.</p></sec> | <contrib id="A1" contrib-type="author"><name><surname>Lafuma</surname><given-names>Antoine J</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>antoine.lafuma@cemka.fr</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Brézin</surname><given-names>Antoine P</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>antoine.brezin@cch.ap-hop-paris.fr</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Fagnani</surname><given-names>Francis L</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>francis.fagnani@cemka.fr</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Mesbah</surname><given-names>Mounir</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>mesbah@ccr.jussieu.fr</email></contrib><contrib id="A5" corresp="yes" contrib-type="author"><name><surname>Berdeaux</surname><given-names>Gilles H</given-names></name><xref ref-type="aff" rid="I4">4</xref><xref ref-type="aff" rid="I5">5</xref><email>gilles.berdeaux@alconlabs.com</email></contrib> | Health and Quality of Life Outcomes | <sec><title>Background</title><p>Visual impairment was responsible for 2,286,000 'disability adjusted life years' in the high income countries in 2001 [<xref ref-type="bibr" rid="B1">1</xref>]. The cost of blindness to the Australian government and community was estimated at between AUS$ 9,749 and AUS$ 26,720 per patient per year [<xref ref-type="bibr" rid="B2">2</xref>]. The mean cost per blind person per year was US$ 11,896 in 1990 in the USA and totalled US$ 4 billion [<xref ref-type="bibr" rid="B3">3</xref>]. Therefore, it is crucial to obtain nationwide estimates of low vision and blindness prevalence to allocate the right amount of resources especially when life expectancy is predicted to increase [<xref ref-type="bibr" rid="B4">4</xref>,<xref ref-type="bibr" rid="B5">5</xref>]. It is also important to understand the causes of visual impairment, in order to implement adequate preventative activities.</p><p>According to the International Statistical Classification of Diseases, Injuries and Causes of Death, visual impairment includes both low vision and blindness. Low vision is defined as visual acuity less than 6/18, but equal to or better than 3/60, or a corresponding visual field loss to less than 20 degrees in the better eye using best possible correction. Blindness is defined as visual acuity less than 3/60, or a corresponding visual field loss to less than 10 degrees in the better eye using best possible correction [<xref ref-type="bibr" rid="B6">6</xref>].</p><p>Primary open-angle glaucoma and age-related macular degeneration are the main two diseases leading to blindness in Western developed countries. Apart from cataract surgery, treatments are available which at best maintain vision, or otherwise postpone visual acuity deterioration. A significant portion of the burden caused by visual impairment is borne by families and includes rehabilitation, medical devices, dedicated software, home modifications, caring time, loss of family revenue, etc.. Nationwide extrapolation has shown that the non-medical costs of visual impairment were comparable to the nationally reimbursed drug budget [<xref ref-type="bibr" rid="B7">7</xref>]. It is therefore crucial to obtain nationwide estimates of low vision and blindness prevalence rates so that sufficient resources are allocated appropriately (medical and non-medical), especially when increasing life expectancy is predicted to continue [<xref ref-type="bibr" rid="B8">8</xref>].</p><p>The use of registers to estimate the prevalence of blindness is controversial, since a high proportion of visually impaired subjects do not register [<xref ref-type="bibr" rid="B9">9</xref>-<xref ref-type="bibr" rid="B12">12</xref>]. According to a WHO review on the prevalence of blindness, ten surveys were conducted in Europe up to 1994 [<xref ref-type="bibr" rid="B13">13</xref>]. An update was performed in 2002 [<xref ref-type="bibr" rid="B14">14</xref>]. Most studies were conducted at a local level, using direct standardisation to derive national estimates. This technique was used by the Eye Diseases Prevalence Research Group [<xref ref-type="bibr" rid="B15">15</xref>]. However, local surveys do not estimate disparities in prevalence rate amongst different geographical areas.</p><p>Healthcare expenditure has increased substantially in all Western industrialised countries during the last decades [<xref ref-type="bibr" rid="B16">16</xref>]. As a result, efficiency in resource allocation has become a major issue in public health decisions, but equity is very important, too, as stated by the National Institute for Clinical Excellence [<xref ref-type="bibr" rid="B17">17</xref>]. Equity is necessary to ensure that two patients, suffering from a similar disease, have access to the same quality of care, and experience the same clinical outcome. However, equity and efficiency (cost per unit of production) are incompatible [<xref ref-type="bibr" rid="B18">18</xref>], so political decisions must be made. Such decisions should be based on studies aimed at quantifying acceptable levels of <italic>in</italic>equity, in order to accommodate fixed budgets. Little has been published on equity and eye care delivery [<xref ref-type="bibr" rid="B19">19</xref>].</p><p>The issue of equity might differ according to healthcare systems, e.g., as between France and the United Kingdom. Some econometric surveys confirm the existence of 'physician-induced demand' in the French system of ambulatory care, which causes healthcare expenditure to increase [<xref ref-type="bibr" rid="B20">20</xref>]. This relationship has been used for decades to justify limits on the number of students entering medical schools. It is contrary to the idealistic theory that an optimal number of physicians would maximise efficient healthcare provision. In this context, a link between the demography of ophthalmologists and the prevalence of both low vision and blindness has never been studied.</p><p>The present survey had three aims: (1) to identify patient demographic risk factors of visual impairment; (2) to compare a visual impairment index across the different regions of France; and (3) to study the relationship between this index and ophthalmologist demographics.</p></sec><sec sec-type="methods"><title>Methods</title><p>Data were gathered in two surveys by the Institut National de la Statistique et des Etudes Economiques (INSEE) [<xref ref-type="bibr" rid="B21">21</xref>]. The databases were subsequently made available to researchers for secondary analyses. The methodology of these two surveys has already been described [<xref ref-type="bibr" rid="B22">22</xref>-<xref ref-type="bibr" rid="B24">24</xref>] elsewhere. The following is a condensed description which should help readers to understand and interpret the results.</p><sec><title>Experimental design: the community survey</title><p>A national census survey is performed every ten years in France. Each household is visited by an interviewer and data are collected on each member of the family. Information was provided by one person of the household. All French people (no age limits) are questioned and answering is compulsory.</p><p>A "Handicap-Dependency" survey documented "handicap", incapacity and dependency of French citizens, living in the community, at a national level. It was decided to piggy-back on the 1999 national census survey.</p><p>The survey followed guidelines and principles for developing disability statistics, published by the United Nations [<xref ref-type="bibr" rid="B25">25</xref>]. The sample was selected by a two-step process [<xref ref-type="bibr" rid="B26">26</xref>,<xref ref-type="bibr" rid="B27">27</xref>].</p><p>1. In 1990, 57,831,816 citizens were documented and statistics on geographical area were available. During the 1999 national census a filtering survey called "Everyday Life and Health" was added. A total of 2,275 geographical areas were picked at random from the 1990 survey, stratified by departments nested within regions, by family, and by socio-professional statistics. The survey consisted of a self-administered 18-item questionnaire that collected information on activities of daily living. Ultimately, 2,223 of the 2,275 geographical areas (97.7%) collaborated in the Everyday Life and Health survey. From the 399,784 questionnaires distributed, 359,010 were completed and returned (89.1%). Questionnaires were to be answered by (or for) all members of a household. This survey did not check the validity of proxy respondents. Non-French speakers having no translation support were unable to answer the questionnaire.</p><p>2. Subjects from the Everyday Life and Health survey were clustered into six impairment groups ranging from no impairment (group 1) to severe impairment (group 6), based upon an impairment severity score [<xref ref-type="bibr" rid="B27">27</xref>]. Subjects in the severe impairment group had a higher probability of being detected by the Handicap-Dependency survey than did those in the Everyday Life and Health survey [<xref ref-type="bibr" rid="B22">22</xref>]. This over-sampling method made it possible to describe the consequences of impairments in detail, since subjects with impairments were over-represented in the Handicap-Dependency survey. Face-to-face interviews were available for 16,945 (77.8%) of 21,760 subjects selected at random from the 'Everyday Life and Health' respondents.</p></sec><sec><title>Experimental design: the institution survey</title><p>Institutions were selected at random from the French Health Ministry files; day-care centres were not included. The sample was stratified according to eighteen strata [<xref ref-type="bibr" rid="B24">24</xref>]. The probability of selecting an institution was inversely proportional to the number of institutions in its stratum and proportional to its number of beds. Eight subjects were picked at random by the interviewers from each resident list.</p><p>In 1998, 2,075 institutions were selected and 155 of them (7.5%) refused to participate. The three most frequent reasons for refusal were lack of time (22.7%), the non-compulsory character of the INSEE survey (10.7%), and lack of staff to help the interviewer (7.3%). In total, 14,611 interviews (94.9%) were performed with 15,403 randomly selected subjects. Analyses were performed on 14,603 subjects with documented impairments, except for eight cases where interviews were stopped before impairments could be documented.</p></sec><sec><title>Data collected</title><p>The survey documented blindness and low vision as declared by subjects, with no medical input. Three formal questions specific to vision were asked during the interview: (1) "Do you have trouble reading newspapers, books, etc ... with spectacles, if you use them?" (2) "Do you have trouble recognising the features of someone standing four meters away from you (with spectacles or contact lenses, if you usually use them)?" (3) "Would you say you are completely blind (light perception at the best), partially blind (still-form perception), or visually impaired?". Data were collected descriptively and experts in medical coding performed <italic>post hoc </italic>classifications of declared diseases. Thus, subjects were classified as belonging to one of the following groups: (1) blind; (2) low visual acuity; or (3) control (i.e. neither blind nor low vision). The cause of impairment was elicited by an open-ended question: "What is the cause of the stated impairment?" The free text was then coded by the interviewer under one of four broad categories: disease, birth-related, injuries, others.</p><p>Ophthalmologists' demography was derived from national statistics [<xref ref-type="bibr" rid="B21">21</xref>,<xref ref-type="bibr" rid="B28">28</xref>] published by the French Ministry of Health (<italic>Ministère de la Santé et des Solidarités</italic>). The <italic>Direction de la Recherche, des Études, de l'Évaluation et des Statistiques </italic>is in charge of up-dating the ophthalmologists' demography, amongst others statistics. We used 2002 data as proxy for regional eye-care services.</p></sec><sec><title>Statistical analysis</title><p>Analyses were conducted with SAS Institute (North Carolina) software release 8.2. Weights for extrapolating data to the entire population were estimated by INSEE from the 1999 national census. These weights were applied to the Everyday Life and Health survey of impairment severity, refusal to participate in the Handicap-Dependency survey, and age, gender, size of household, type of household and geographical area size based on the latter survey. For the institution survey, weights included size of strata, the institution occupation rate (number of subjects in the institution/number of available beds), and the answer refusal rate (higher in psychiatric centres).</p><p>A weighted logistic regression was used to identify risk factors. One regression identified the risk factors for blindness (blind <italic>versus </italic>no visual impairment) and another the risk factors for low vision (low vision <italic>versus </italic>no visual impairment). The reference state was "no visual impairment". Risk factors included in the models were age (continuous variable), job classification (Reference farmer) and national region (Reference <italic>Ile de France</italic>). Odds-ratios with 95% confidence limits are presented. Age and job classification factors were chosen to adjust on the socio-economic variability amongst French regions (e.g., people living on the French Riviera are older and richer than those living in the North).</p></sec></sec><sec><title>Results</title><p>Altogether, 16,945 questionnaires were collected by the community survey and 14,603 by the institution survey. 2,703 subjects declared low vision and 350 blindness. Extrapolation to the national level predicted that 664,253 of 58,096,060 subjects (1.14%) lived in institutions. People living in institutions were older and less frequently male than those in the community. Additional descriptive information (socio-demographics, co-morbidity, etc ...) of this population can be found elsewhere [<xref ref-type="bibr" rid="B22">22</xref>-<xref ref-type="bibr" rid="B24">24</xref>].</p><p>Prevalences of low vision and blindness increased exponentially with age (Table <xref ref-type="table" rid="T1">1</xref>). More than one-quarter of patients older than 90 years declared a visual impairment. Responses of centenarians were few and should be interpreted with caution.</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>Prevalence of low vision in all populations (persons living at home or in institutions). 95% confidence interval. n.e. not estimable</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left"><bold>Age (years) </bold>Persons living in institution and at home (n = 58,096,060)</td><td align="center"><bold>Prevalence of low vision</bold></td><td align="center"><bold>Prevalence of blindness</bold></td></tr></thead><tbody><tr><td align="left">0–9</td><td align="center">0.62% [0.32%,1.20%]</td><td align="center"><0.001% [0.00%,2.37%]</td></tr><tr><td align="left">10–19</td><td align="center">0.27% [0.12%,0.60%]</td><td align="center">0.03% [0.00%,0.45%]</td></tr><tr><td align="left">20–29</td><td align="center">1.34% [0.77%,2.33%]</td><td align="center">0.02% [0.00%,0.42%]</td></tr><tr><td align="left">30–39</td><td align="center">0.29% [0.14%,0.63%]</td><td align="center">0.02% [0.00%,0.39%]</td></tr><tr><td align="left">40–49</td><td align="center">1.91% [1.13%,3.21%]</td><td align="center">0.06% [0.01%,0.59%]</td></tr><tr><td align="left">50–59</td><td align="center">1.30% [0.73%,2.31%]</td><td align="center">0.11% [0.01%,0.99%]</td></tr><tr><td align="left">60–69</td><td align="center">3.06% [1.82%,5.11%]</td><td align="center">0.21% [0.03%,1.65%]</td></tr><tr><td align="left">70–79</td><td align="center">5.92% [3.64%,9.48%]</td><td align="center">0.09% [0.01%,0.97%]</td></tr><tr><td align="left">80–89</td><td align="center">14.10% [8.90%,21.62%]</td><td align="center">0.91% [0.12%,6.4%]</td></tr><tr><td align="left">90–99</td><td align="center">23.13% [14.18%,35.41%]</td><td align="center">4.73% [0.68%,26.43%]</td></tr><tr><td align="left">100 +</td><td align="center">33.71% [n.e.]</td><td align="center">3.27% [n.e.]</td></tr></tbody></table></table-wrap><p>The major cause of blindness declared by subjects was acquired diseases (Table <xref ref-type="table" rid="T2">2</xref>). 21,600 blindness were congenital and 35,000 were acquired. The figures for LV were 179,000 and 660,000, respectively. About 160,800 instances of blindness and low vision could be avoided in France by accident prevention.</p><table-wrap position="float" id="T2"><label>Table 2</label><caption><p>Causes of blindness and low vision declared by the respondents.</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left"><bold>Cause declared by respondent</bold></td><td align="center"><bold>Low vision</bold></td><td align="center"><bold>Blind</bold></td></tr></thead><tbody><tr><td align="left">Pregnancy and/or birth complications, congenital or hereditary disease</td><td align="center">14.9%</td><td align="center">30.9%</td></tr><tr><td align="left">Acquired diseases</td><td align="center">55.3%</td><td align="center">50.2%</td></tr><tr><td align="left">Injuries</td><td align="center">12.4%</td><td align="center">12.0%</td></tr><tr><td align="left">Other or unknown</td><td align="center">17.4%</td><td align="center">6.9%</td></tr></tbody></table></table-wrap><p>The prevalence of visual impairment was similar between the sexes. After adjustment on age and region, people exempted from employment, and those working as artisans, shopkeepers or business-owners, had significantly less risk of developing low vision (odds-ratios 2.10 and 1.51, respectively) than did farmers (Table <xref ref-type="table" rid="T3">3</xref>). Persons with no professional activity had a higher probability (odds-ratio 0.28) of developing blindness than farmers.</p><table-wrap position="float" id="T3"><label>Table 3</label><caption><p>Probability of developing low vision or blindness according to job classification, adjusted on age and region. An OR greater than 1 means less risk of visual impairment. Reference category is 'Farmer'.</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td align="center" colspan="2"><bold>Low vision</bold></td><td align="center" colspan="2"><bold>Blindness</bold></td></tr><tr><td></td><td align="center"><bold>OR</bold></td><td align="center"><bold>95% CL</bold></td><td align="center"><bold>OR</bold></td><td align="center"><bold>95% CL</bold></td></tr></thead><tbody><tr><td align="left">Farmer (Reference)</td><td align="center">1.00</td><td align="center">na</td><td align="center">1.00</td><td align="center">na</td></tr><tr><td align="left">Artisan, shopkeeper, business owner</td><td align="center">1.51</td><td align="center">1.12–2.03</td><td align="center">0.91</td><td align="center">0.33–2.51</td></tr><tr><td align="left">Exempt</td><td align="center">2.10</td><td align="center">1.34–3.28</td><td align="center">1.40</td><td align="center">0.29–6.71</td></tr><tr><td align="left">Part-time worker</td><td align="center">1.08</td><td align="center">0.79–1.47</td><td align="center">3.60</td><td align="center">0.42–31.09</td></tr><tr><td align="left">Employee</td><td align="center">0.96</td><td align="center">0.71–1.32</td><td align="center">0.91</td><td align="center">0.27–3.03</td></tr><tr><td align="left">Unskilled worker</td><td align="center">0.97</td><td align="center">0.77–1.21</td><td align="center">1.38</td><td align="center">0.53–3.62</td></tr><tr><td align="left">No professional activity</td><td align="center">1.11</td><td align="center">0.66–1.86</td><td align="center">0.28</td><td align="center">0.08–0.96</td></tr><tr><td align="left">Unclassified</td><td align="center">1.34</td><td align="center">0.81–2.20</td><td align="center">1.05</td><td align="center">0.18–6.29</td></tr></tbody></table></table-wrap><p>After adjustment on age (more old people live in the South of France) and job classification (GDP in northern regions is lower), the prevalence of blindness did not differ significantly (95% CL) between regions (Table <xref ref-type="table" rid="T4">4</xref>). The picture was different for low vision. In seven regions the probability of developing low vision (odds-ratios between 0.35 and 0.59) was significantly higher than in Ile-de-France, and only one region had a statistically significant lower probability (odds-ratio 2.10). On comparing range extremes, people of the same age and job classification had 6.0 more risk of low vision if they lived in Poitou-Charentes than in Haute-Normandie.</p><table-wrap position="float" id="T4"><label>Table 4</label><caption><p>Probability of developing low vision or blindness according to region, adjusted on age and job categories. An OR greater than 1 means less risk of develop low vision. Reference category is "Ile de France", i.e. Paris and its suburb.</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td align="center" colspan="2"><bold>Low vision</bold></td><td align="center" colspan="2"><bold>Blindness</bold></td></tr><tr><td></td><td align="center"><bold>OR</bold></td><td align="center"><bold>95% CL</bold></td><td align="center"><bold>OR</bold></td><td align="center"><bold>95% CL</bold></td></tr></thead><tbody><tr><td align="left">Ile de France (Reference)</td><td align="center">1</td><td align="center">na</td><td align="center">1</td><td align="center">Na</td></tr><tr><td align="left">Champagne Ardennes</td><td align="center">0.82</td><td align="center">0.48–1.43</td><td align="center">2.38</td><td align="center">0.10–58.00</td></tr><tr><td align="left">Picardie</td><td align="center">0.59</td><td align="center">0.37–0.96</td><td align="center">0.90</td><td align="center">0.11–7.15</td></tr><tr><td align="left">Haute-Normandie</td><td align="center">0.35</td><td align="center">0.23–0.53</td><td align="center">1.27</td><td align="center">0.13–12.01</td></tr><tr><td align="left">Centre</td><td align="center">0.52</td><td align="center">0.35–0.77</td><td align="center">1.36</td><td align="center">0.22–8.55</td></tr><tr><td align="left">Basse-Normandie</td><td align="center">1.57</td><td align="center">0.72–3.44</td><td align="center">2.79</td><td align="center">0.09–88.38</td></tr><tr><td align="left">Bourgogne</td><td align="center">0.41</td><td align="center">0.27–0.62</td><td align="center">2.17</td><td align="center">0.11–43.46</td></tr><tr><td align="left">Nord-Pas-de-Calais</td><td align="center">0.63</td><td align="center">0.43–0.91</td><td align="center">1.34</td><td align="center">0.24–7.52</td></tr><tr><td align="left">Lorraine</td><td align="center">0.73</td><td align="center">0.48–1.13</td><td align="center">1.82</td><td align="center">0.20–16.67</td></tr><tr><td align="left">Alsace</td><td align="center">1.62</td><td align="center">0.88–2.97</td><td align="center">1.35</td><td align="center">0.16–11.11</td></tr><tr><td align="left">France-Comté</td><td align="center">1.95</td><td align="center">0.86–4.45</td><td align="center">0.87</td><td align="center">0.10–7.45</td></tr><tr><td align="left">Pays de Loire</td><td align="center">0.45</td><td align="center">0.31–0.64</td><td align="center">1.65</td><td align="center">0.24–11.38</td></tr><tr><td align="left">Bretagne</td><td align="center">1.38</td><td align="center">0.85–2.26</td><td align="center">1.16</td><td align="center">0.23–5.75</td></tr><tr><td align="left">Poitou-Charentes</td><td align="center">2.10</td><td align="center">1.04–4.23</td><td align="center">1.39</td><td align="center">0.19–10.34</td></tr><tr><td align="left">Aquitaine</td><td align="center">0.49</td><td align="center">0.34–0.73</td><td align="center">0.76</td><td align="center">0.19–3.10</td></tr><tr><td align="left">Midy-Pyrénées</td><td align="center">1.20</td><td align="center">0.70–2.07</td><td align="center">1.47</td><td align="center">0.21–10.37</td></tr><tr><td align="left">Limousin</td><td align="center">0.47</td><td align="center">0.27–0.81</td><td align="center">3.88</td><td align="center">0.03–442.19</td></tr><tr><td align="left">Rhônes-Alpes</td><td align="center">0.74</td><td align="center">0.52–1.06</td><td align="center">0.87</td><td align="center">0.23–3.22</td></tr><tr><td align="left">Auvergne</td><td align="center">0.62</td><td align="center">0.36–1.04</td><td align="center">2.24</td><td align="center">0.08–59.30</td></tr><tr><td align="left">Languedoc-Roussillon</td><td align="center">0.75</td><td align="center">0.47–1.20</td><td align="center">0.64</td><td align="center">0.15–2.77</td></tr><tr><td align="left">Provence-Alpes-Côtes d'Azur</td><td align="center">0.77</td><td align="center">0.54–1.12</td><td align="center">0.64</td><td align="center">0.20–2.03</td></tr></tbody></table></table-wrap><p>A [non-significant] linear trend in Figure <xref ref-type="fig" rid="F1">1</xref> indicates that the probability of low vision decreases as the density of ophthalmologists (number of ophthalmologists per 100,000 inhabitants) increases, after adjustment on age and job classification. Also, six of the seven regions with significantly higher prevalence rates of low vision had ophthalmologist densities below the national average.</p><fig position="float" id="F1"><label>Figure 1</label><caption><p>Relationship between the number of ophthalmologists per 100,000 inhabitants and the OR (adjusted on age and socio-professional categories) for low vision per region. Reference: "Ile de France". Dotted line: linear regression. Region ORs that differed significantly from 1 are underlined. An OR greater than 1 means less risk of developing low vision. PACA: Provence-Alpes-Côtes d'Azur.</p></caption><graphic xlink:href="1477-7525-4-34-1"/></fig></sec><sec><title>Discussion</title><p>The surveys analysed shared two limitations: (1) their cross-sectional design did not allow an analysis of possible causalities between blindness, or low vision, and risk factors; and (2) the actual visual acuity of subjects who responded was not measured by ophthalmologists. Subjects classified as blind self-declared that they could not perceive shapes. This may be a serious limitation to our analyses, although our prevalence figures are close to the only French report in the international literature [<xref ref-type="bibr" rid="B29">29</xref>]. On the other hand, we did study representative samples of subjects from both the community and institutions. Another issue concerns the small number of subjects who declared themselves blind, which resulted in large OR confidence intervals.</p><p>The different relationships between age and the prevalence rates of low vision and blindness may be explained by the different reasons given by subjects for the impairments. A considerable proportion of blindness was related to pregnancy and childbirth, whereas the main cause of low vision was attributed to acquired diseases. In other words, a significant proportion of blindness is not managed by ophthalmologists, which might explain the lack of association between ophthalmologist density and the blindness prevalence rate Lastly, most diseases affecting vision in developed countries do not make patients immediately blind since treatments are available and costs reimbursed. Therefore, most patients have had low vision before becoming blind. However, since the cause of visual impairment was self-declared and was not medically certified, apparent differences between the causes of low vision and blindness might be explained by recall bias.</p><p>It should be noted that one-in-eight visual impairments were related to injury. Therefore, preventative measures would have avoided some cases of low vision and blindness, which totalled 152,400 and 8,400 total persons, respectively, for a country with 58,000,000 inhabitants.</p><p>Persons with higher educational achievement were less at risk for low vision, but this was not so for blindness. Higher education enables people to become better informed about potential diseases related to ageing, and gives them more effective access to healthcare.</p><p>When the present data were collected, access to ophthalmologists in France did not require referral by general practitioners. In addition, more than 95% of French people have private insurance supplementing their national sick fund protection [<xref ref-type="bibr" rid="B30">30</xref>]. Insurance policies cover hospitalisation costs and all out-patient care: drugs, visits, examinations, etc ... Average patient co-payment in 2001 was 11.1% of total expenditure [<xref ref-type="bibr" rid="B31">31</xref>]. These financial provisions were supposed to ensure excellent equity. What we found, however, was inequity.</p><p>It could be expected that people with greater economic means or greater educational levels might be much more aggressive in seeking out eye care and some of them might even be seeking eye cares outside their area. This is why it was very important to get prevalence rates adjusted on job description to control for the above effect. In France, most of the vision is under the control of ophthalmologists: visual acuity, diagnosis, treatments, etc ... There is no limitation to access them, outside their availability. The role of optometrists is very low, almost inexistent. Therefore, the ophthalmologist density could be considered as a good indicator of resources available to preserve vision at a national level.</p><p>After adjusting on age and job classification, our analysis showed that differences existed between geographic regions with respect to the prevalence of low vision. Subjects living in Haute-Normandie had a 2.86 greater chance of developing low vision than people in the Paris area, whereas persons in Poitou-Charentes had a 2.10 lower chance than Parisians. In contrast, an association was found between ophthalmologist density (number/100,000 inhabitants) and the regional distribution of low vision. Thus, seven of eight regions (85.7%) with a significantly higher prevalence of low vision had ophthalmologist densities below the national average. This suggests that the supply of vision-related services may be a determinant of eye morbidity at a national level.</p><p>To confirm these findings, it would be worthwhile to study the relationship between regional visual impairment rates and indicators of other eye-care activities, such as number of visits/inhabitant, glaucoma diagnosis campaign, etc.. Inequality of quality of care could also be a factor explaining the prevalence rate differences across the regions. Unfortunately, such aspects of ophthalmological activity or quality estimates are unavailable in France at a regional level Lastly, the same HID surveys showed that visual impairment impacted dramatically on activities of daily living [<xref ref-type="bibr" rid="B22">22</xref>-<xref ref-type="bibr" rid="B24">24</xref>] and had economic consequences on the family revenue [<xref ref-type="bibr" rid="B7">7</xref>]. The latter, alone, might reduce access to eye-care. Hence, to conclude that a similar density of ophthalmologists should be provided in all regions is premature.</p><p>However, the causality of the association between the prevalence of visual impairment in relation to the number of ophthalmologists in a given area might be confounded by some factors that were not collected in our surveys. This encompasses, for example, population genetic factors distribution across the different areas, other health care resource supply (access to hospital is more difficult in rural area), or eating habits (south part of France people used to eat more fresh fruits and vegetables which is known to protect against acquired visual impairment). These are strong limitations to the analyses we conducted and additional data should be collected to confirm our findings.</p><p>It is interesting that a recent national survey of the UK system for delivering care to low vision subjects, involving a wide range of service providers, also found regional inequity, as in France [<xref ref-type="bibr" rid="B32">32</xref>]. The number of service providers was lowest in areas where the general population was small, but the prevalence of low vision was highest. Conversely, the number of service providers was highest in cities where the general population was large, despite the prevalence of low vision being only moderate.</p><p>It is evident that where practitioner remuneration is based on a fee-for-service, as in France, measures are needed to control physician-induced demand. However, on a broader scale, irrespective of the healthcare system, there is some evidences to justify including a minimum level of equity in plans to reorganise eye-care services. For example, the prevalence of visual impairment in the Auvergne does not differ significantly from the Ile-de France, yet the density of ophthalmologists is below the national average. It would be equitable if such standard were applied to all regions.</p><p>It was not the intention of this paper to demonstrate or claim the need for a fixed ratio of ophthalmologists to inhabitants. However, investment in healthcare is supposed to be effective, as resources are limited. Ultimately, the daily work of ophthalmologists is to preserve vision, so maintenance of vision or reduction of low vision prevalence rates is a legitimate public health aim. We found some weak associations. This suggests that a minimum ophthalmologist density might be an aspect to consider when allocating resources for the preservation of vision.</p></sec><sec><title>Conclusion</title><p>An association was found between the number of ophthalmologists/inhabitants and the prevalence of low vision, in France. These data suggests that ophthalmologist density could be one of the drivers of good vision at a population level.</p></sec><sec><title>Competing interests</title><p>The author(s) declare that they have no competing interests.</p></sec><sec><title>Authors' contributions</title><p>AL and FF retrieved the data bases. The analyses were performed GB. All authors contributed to the writing of the manuscript.</p></sec> |
Quality of life in patients with psoriasis | <p>Psoriasis is one of the prevalent skin conditions in the United States. This chronic condition has a significant negative impact on patients' quality of life. Psoriasis has been linked to the depression and suicidal tendencies in the patients. The costs associated with decrements in quality of life, lost productivity, and work absenteeism may be enormous, increasing overall costs associated with the disease management. This review attempts to outline different quality of life measures available for psoriasis and describes their use in studies examining patient reported outcomes associated with pharmacological interventions for psoriasis. Factors associated with quality of life in psoriasis patients are described. It further describes physician's role in the psoriasis management to improve patients' overall well-being.</p> | <contrib id="A1" contrib-type="author"><name><surname>Bhosle</surname><given-names>Monali J</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>bhosle.1@osu.edu</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Kulkarni</surname><given-names>Amit</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>kulkarni.40@osu.edu</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Feldman</surname><given-names>Steven R</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>sfeldman@wfubmc.edu</email></contrib><contrib id="A4" corresp="yes" contrib-type="author"><name><surname>Balkrishnan</surname><given-names>Rajesh</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>balkrishnan.1@osu.edu</email></contrib> | Health and Quality of Life Outcomes | <sec><title>Review</title><sec><title>Psoriasis: a growing problem</title><p>Psoriasis affects approximately 2% of the world's population, with men and women being equally affected [<xref ref-type="bibr" rid="B1">1</xref>]. In the United States (US) about 250,000 new cases of psoriasis are observed annually, affecting almost 2.2% of the US population [<xref ref-type="bibr" rid="B2">2</xref>,<xref ref-type="bibr" rid="B3">3</xref>]. Psoriasis accounted for nearly 2.25 million visits to ambulatory care centers during 1996 in the US [<xref ref-type="bibr" rid="B4">4</xref>].</p><p>Psoriasis is a serious condition strongly affecting the way a person sees himself and the way he is seen by others. It has tremendous economic and financial ramifications. Total annual cost for treating psoriasis is estimated to be in the range of $1.6 billion to $4.3 billion dollars [<xref ref-type="bibr" rid="B5">5</xref>-<xref ref-type="bibr" rid="B7">7</xref>]. Psoriasis is linked with social stigmatization, pain, discomfort, physical disability and psychological distress [<xref ref-type="bibr" rid="B8">8</xref>].</p></sec><sec><title>Impact of psoriasis on patients' quality of life</title><p>Psoriasis has a significant negative impact on patients' health related quality of life (HRQoL). In a survey by the National Psoriasis Foundation almost 75% of patients believed that psoriasis had moderate to large negative impact on their quality of life (QoL), with alterations in their daily activities [<xref ref-type="bibr" rid="B9">9</xref>]. Another study reported that at least 20% of psoriasis patients had contemplated suicide [<xref ref-type="bibr" rid="B10">10</xref>]. Furthermore, physical and emotional effects of psoriasis were found to have a significant negative impact at patients' workplace as measured by the validated scales including Work Productivity Assessment Index (WPAI), SF-8, Hospital Anxiety and Depression (HADS) and past medical/psoriasis history [<xref ref-type="bibr" rid="B11">11</xref>]. Absenteeism is a greater concern for people suffering from psoriasis than their co-workers without psoriasis with nearly 60% patients reporting missing an average of 26 days a year directly related to their psoriasis [<xref ref-type="bibr" rid="B12">12</xref>]. Patients with psoriasis have a higher financial burden due to absenteeism in addition to the cost of caring for their disease [<xref ref-type="bibr" rid="B13">13</xref>,<xref ref-type="bibr" rid="B14">14</xref>].</p><p>Psoriasis patients often experience difficulties like maladaptive coping responses, problems in body image, self esteem, self concept and also have feelings of stigma, shame and embarrassment regarding their appearance [<xref ref-type="bibr" rid="B8">8</xref>]. This is often times accompanied by a perception of being evaluated by others based on their disfigurement [<xref ref-type="bibr" rid="B8">8</xref>]. Individuals with psoriasis commonly engage in coping strategies to avoid unwanted and unpleasant social consequences. However, most of these strategies fail to improve patients' QoL [<xref ref-type="bibr" rid="B14">14</xref>-<xref ref-type="bibr" rid="B16">16</xref>]. Discussing their skin condition, covering their lesions, and avoiding contact with people are significantly associated with negative impact on life [<xref ref-type="bibr" rid="B11">11</xref>-<xref ref-type="bibr" rid="B16">16</xref>]. Studies have indicated that talking to others regarding the non-contagious nature of psoriasis lessens the negative impact on the QoL and thereby reduces social discomfort [<xref ref-type="bibr" rid="B17">17</xref>].</p><p>Studies that have probed the link between psoriasis and depression hint towards a reciprocal relation between them. Psoriasis patients were more likely to be depressed than the general population with patients' age, education and disease severity being important predictors of psychological distress in the patient cohort [<xref ref-type="bibr" rid="B18">18</xref>,<xref ref-type="bibr" rid="B19">19</xref>]. Gupta et al. in their study of 127 psoriasis patients found that 9.7% of patients reported their wish to be dead, and 5.5% reported active suicidal ideation at the time of the study [<xref ref-type="bibr" rid="B20">20</xref>]. These studies have highlighted the need for psychosocial strategies in treating patients with psoriasis and helping them to improve their overall QoL.</p></sec><sec><title>Determinants of QoL in psoriasis patients</title><p>Various factors may be attributed to the lower QoL in psoriasis patients. The chronic and recurring nature of this disease often brings about a feeling of hopelessness in terms of cure for the condition [<xref ref-type="bibr" rid="B21">21</xref>]. Patients are constantly concerned with the interference with future plans due to an unexpected outbreak of symptoms. This possibly intensifies due to their lack of control over the disease [<xref ref-type="bibr" rid="B21">21</xref>]. Lack of control is one of the most bothersome aspects in psoriasis patients [<xref ref-type="bibr" rid="B22">22</xref>].</p><p>Many psoriasis patients experience social and psychological difficulties created by their environment [<xref ref-type="bibr" rid="B23">23</xref>]. Psoriasis patients may feel humiliated when they need to expose their bodies during swimming, intimate relationships, using public showers, or living in conditions that do not provide appropriate privacy [<xref ref-type="bibr" rid="B24">24</xref>]. Many of the patients suffering from psoriasis often feel the need to hide their disease, thus severely affecting their self confidence [<xref ref-type="bibr" rid="B25">25</xref>].</p><p>People suffering from psoriasis feel that the general public and sometimes their own physicians fail to appreciate the negative impact of psoraisis on their life [<xref ref-type="bibr" rid="B10">10</xref>,<xref ref-type="bibr" rid="B15">15</xref>]. Psoriasis has an immense impact on social life, with patients frequently complaining of social difficulties and friction with family members [<xref ref-type="bibr" rid="B26">26</xref>]. Psoriasis patients frequently feel ashamed and embarrassed about their condition and considered this to be the worst aspect of their disease [<xref ref-type="bibr" rid="B27">27</xref>]. High levels of stress in this population may often result from other people reacting to their disease or anticipation of the same [<xref ref-type="bibr" rid="B27">27</xref>].</p><p>Psoriasis is also associated with limitations in daily activities, occupational, and sexual functioning [<xref ref-type="bibr" rid="B14">14</xref>,<xref ref-type="bibr" rid="B28">28</xref>,<xref ref-type="bibr" rid="B29">29</xref>]. Patients with psoriasis suffer comparable disability as other patients with chronic illnesses [<xref ref-type="bibr" rid="B16">16</xref>]. All these factors may have detrimental effect on the patients' QoL [<xref ref-type="bibr" rid="B30">30</xref>]. In one qualitative study carried out to assess the determinants of QoL in the US population with psoriasis, body surface area showed the strongest association with decrements in QoL (Spearman's ρ = 0.50; p < 0.0001), among other factors including patients' age, gender, income, duration of psoriasis, and number of physicians seen in last two years. Increasing psoriasis severity was significantly associated with seeking care from multiple physicians and having decrements in income in this population [<xref ref-type="bibr" rid="B31">31</xref>]. While measures of body surface area affected by psoriasis are commonly used in clinical trials to assess severity of the disease, there is a movement which argues that QoL standard would be a better method of determining the severity of psoriasis [<xref ref-type="bibr" rid="B10">10</xref>]. In the clinic setting, treatment judgments may be largely guided by QoL issues.</p></sec><sec><title>Tools to measure QoL</title><p>Various measures have been used to assess QoL in psoriasis patients. These measures may be categorized as psoriasis-specific, skin specific, generic QoL measures, and "mixed" measures. Following section attempts to describe each of these categories with related examples. Psoriasis specific measures are the most sensitive, however the more general measures facilitate comparisons across diseases. Often, studies measuring QoL in psoriasis patients utilize more than one of the available measures mentioned here. Use of these measures varied across different clinical controlled trials examining effect of different pharmacological treatments on QoL of psoriasis patients (Table <xref ref-type="table" rid="T1">1</xref>).</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>Table Quality of Life Measures and Outcomes in Randomized Controlled Trials for Psoriasis Treatments</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left"><bold>Study Title (Year)</bold></td><td align="left"><bold>Treatment</bold></td><td align="left"><bold>QoL Measures</bold></td><td align="left"><bold>Outcomes</bold></td></tr></thead><tbody><tr><td align="left">"Impact of efalizumab on patient-reported outcomes in high-need psoriasis patients: results of the international, randomized, placebo-controlled Phase III Clinical Experience Acquired with Raptiva (CLEAR) trial" (2005) 40</td><td align="left">Efalizumab 1 mg/kg/wk (n = 529) or placebo (n = 264) for 12 weeks</td><td align="left">1. SF-36<break/>2. DLQI<break/>3. PSA<break/>4. VAS for itching<break/>5. PGPA</td><td align="left">QOL, measured using all QOL measures, was significantly higher among the Efalizumab group as compared with the placebo group (p < 0.001)</td></tr><tr><td align="left">"Patient-reported outcomes of psoriasis improvement with etanercept therapy: results of a randomized phase III trial" (2005) [41]</td><td align="left">Etanercept 50 mg twice weekly (n = 194) Placebo (n = 193), etanercept 50 mg per week (n = 196) during the initial 12-week, double-blind period.</td><td align="left">1.DLQI<break/>2.SF-36<break/>3. PGPA</td><td align="left">DLQI total score improved by 65–70% in etanercept group compared with 6% in placebo group (P < 0.0001). Significant improvement in etanercept group as measured by other SF-36 and PGPA</td></tr><tr><td align="left">"Alefacept in the treatment of psoriasis in patients for whom conventional therapies are inadequate" (2005) [42]</td><td align="left">Alefacept (Amevive<sup>®</sup>)</td><td align="left">1. PASI<break/>2. DLQI</td><td align="left">The QOL effects of alefacept in patients who were not candidates for conventional systemic psoriasis therapies or phototherapy were similar to those reported previously for the overall alefacept-treated population in the phase III studies (p = 0.001).</td></tr><tr><td align="left">"Infliximab treatment results in significant improvement in the quality of life of patients with severe psoriasis: a double-blind placebo-controlled trial" (2005) [43]</td><td align="left">Intravenous infusions of 3 or 5 mg kg(-1) of infliximab or placebo</td><td align="left">DLQI</td><td align="left">Infliximab induction therapy resulted in a substantial improvement in HRQOL. At week 10, patients in the infliximab 3- and 5-mg kg(-1) groups showed a median percentage improvement in DLQI scores of 84.0% and 91.0%, respectively, compared with 0% in the placebo group (P < 0.001)</td></tr><tr><td align="left">"The efficacy and tolerability of clobetasol propionate foam 0.05% in the treatment of mild to moderate plaque-type psoriasis of nonscalp regions" (2003) [44]</td><td align="left">Clobetasol propionate foam (clobetasol foam) 0.05%</td><td align="left">PGA</td><td align="left">Clobestasol propionate foam 0.05% had greater improvement in QoL as compared to other topical therapies reported by patients.</td></tr><tr><td align="left">"Quality of life and clinical outcome in psoriasis patients using intermittent cyclosporine" (2001) [45]</td><td align="left">Cyclosporin (Neoral<sup>®</sup>) (n = 255)</td><td align="left">1. DLQI<break/>2. PASI</td><td align="left">Intermittent short courses of cyclosporin significantly improved the QoL of the patients and decreases the extent and severity of disease and itch</td></tr><tr><td align="left">"The impact of a two-compound product containing calcipotriol and betamethasone dipropionate (Daivobet/Dovobet) on the quality of life in patients with psoriasis vulgaris: a randomized controlled trial" (2004) [46]</td><td align="left">Combination therapy with topical vitamin D analogue calcipotriol (50 microg g(-1)) and corticosteroid betamethasone dipropionate (0.5 mg g(-1)) vs. calcipotriol monotherapy</td><td align="left">1. PDI<break/>2. EuroQoL 5D<break/>3. VAS</td><td align="left">Once-daily application of the combination product was found to be superior to calcipotriol twice daily terms of QoL</td></tr><tr><td align="left">"The effect of treatment on quality of life in psoriasis patients" (2005) [47]</td><td align="left">Treatment with short contact dithranol treatment, UVB phototherapy or inpatient dithranol</td><td align="left">1. Dutch short form of the SIP<break/>2. PDI</td><td align="left">Comparable improvement in HRQoL with short contact dithranol treatment and UVB phototherapy, inpatients experienced a more impaired HRQoL and showed no significant improvement in HRQoL directly following treatment</td></tr><tr><td align="left">"Methotrexate versus cyclosporine in moderate-to-severe chronic plaque psoriasis" (2003) [48]</td><td align="left">Methotrexate (n= 44; initial dose, 15 mg per week) or cyclosporine (n= 44; initial dose, 3 mg per kilogram of body weight per day)</td><td align="left">1. PASI<break/>2. PGA</td><td align="left">The difference in the QOL for both the treatment arms was statistically insignificant</td></tr><tr><td align="left">"Calcipotriol vs. tazarotene as combination therapy with narrowband ultraviolet B (311 nm): efficacy in patients with severe psoriasis" (2000) [49]</td><td align="left">Combination of UVB (311 nm) and tazarotene vs. UVB (311 nm) plus calcipotriol or vice versa</td><td align="left">PASI</td><td align="left">No significant differences in QoL of patients in both the regimens</td></tr><tr><td align="left">"A comparison of treatment with dithranol and calcipotriol on the clinical severity and quality of life in patients with psoriasis" (1998) [50]</td><td align="left">Calcipotriol ointment (50 micrograms/g) twice daily or Dithrocream (short-contact dithranol) 0.1–2%</td><td align="left">1. PDI<break/>2. SIP</td><td align="left">Significant improvement in patients' QoL as assessed by the PDI and the SIP were seen in both treatment groups, with greater improvement in calcipotriol group</td></tr></tbody></table></table-wrap></sec><sec><title>Psoriasis-specific measures</title><sec><title>Psoriasis Index of Quality of Life (PSORIQoL)</title><p>The PSORIQoL is based on a "needs-based" approach. This instrument is based on the theory that "life gains its quality from the ability and capacity of individuals to satisfy their needs" [<xref ref-type="bibr" rid="B30">30</xref>]. This 25 dichotomous item instrument was developed through interviews conducted in three countries in Europe. An American version has also been recently developed. It has the advantage of being based on theory and measuring the impact of the disease on QoL rather than assessing impairment or disability. Moreover, it is expected to work in a uniform manner across patient samples, irrespective of age and gender [<xref ref-type="bibr" rid="B30">30</xref>].</p></sec><sec><title>Psoriasis Life Stress Inventory (PLSI)</title><p>The PLSI is a 15-item questionnaire that provides a measure of the daily hassles of psychosocial stress associated with having to cope with everyday events in living with psoriasis. Scores on this scale range from 0 to 45. The PLSI also permits patients to be classified as a function of their distribution of scores into two groups: those patients who react significantly to the stress associated with having psoriasis (score of > 10); and those patients who are not significantly affected with having psoriasis-related stress (score of < 10). The PLSI is scored by having the respondent rate the absolute frequency with which each item has been experienced in the last 4 weeks from (scoring 0) to a great deal (scoring 3) [<xref ref-type="bibr" rid="B27">27</xref>].</p></sec><sec><title>Psoriasis Disability Index (PDI)</title><p>The PDI is a 15-item scale that specifically addresses self-reported disability in areas of daily activities, employment, personal relationships, leisure and treatment effects. The items are concerned with the practical effects of psoriasis in every day life [<xref ref-type="bibr" rid="B27">27</xref>].</p></sec><sec><title>Psoriasis Area and Severity Index (PASI) and Simplified PASI (SAPASI)</title><p>Four main areas were assessed for calculation of the PASI scores: the head, the trunk, the upper extremities, and the lower extremities, corresponding to 10%,20%,30%, and 40% of the total body area, respectively. The maximum score for PASI is 72. The SAPASI is a self-assessed, using the same criteria as the PASI, but presented in non-professional terminology. Scores on the SAPASI range from 0 to 72 [<xref ref-type="bibr" rid="B25">25</xref>]. Though in essence the PASI and the SAPASI are measures for severity of psoriasis, they provide an adequate picture of the impact of the disease on patients' QoL. Studies have indicated an inverse relationship between QoL and severity of psoriasis. Moreover, PASI is the most widely used measure of severity in the research as well as the clinical setting. This makes it an important tool in gauging the impact of the disease on QoL, though other instruments to measure QoL are encouraged. Since PASI or SAPASI do not measure the impact of psoriasis on patients' QoL directly, use of other QoL scales is recommended.</p></sec></sec><sec><title>Skin-specific measures</title><sec><title>Questionnaire on Experience with Skin Complaints (QES)</title><p>The short form of the QES with 23 items is a valid instrument for examination of social and psychic burdens of psoriasis. The recording of stigmatization feeling and of quality of life determines different supplementary aspects of the illness-related stress of patients with chronic skin diseases [<xref ref-type="bibr" rid="B32">32</xref>].</p></sec><sec><title>Dermatology Life Quality Index (DLQI)</title><p>The DLQI is a compact self-reported questionnaire to measure HRQoL over the previous week in patients with skin diseases. It consists of 10 items covering symptoms and feelings (items 1 and 2), daily activities (items 3 and 4), leisure (items 5 and 6), work and school (item 7), personal relationships (items 8 and 9) and treatment (item 10). Each item is scored on a four point scale, with higher scores indicating greater impairment in HRQoL [<xref ref-type="bibr" rid="B33">33</xref>].</p></sec></sec><sec><title>Generic QOL measures</title><sec><title>Short Form 36 (SF-36)</title><p>The SF-36 health survey is a widely used generic, 36-item, self-reported health status questionnaire assessing 8 domains of health status(1) physical activities;(2) social activities;(3) usual physical role activities;(4) bodily pain;(5) general mental health (6) usual emotional role activities; (7) vitality ;(8)general health perceptions. A score from 0 to 100 is calculated for each subscale, with higher scores indicating better HRQL [<xref ref-type="bibr" rid="B25">25</xref>]. The SF-36 may be the best characterized measure for comparing QoL differences across different diseases. The SF-36 was used to show that the impact of psoriasis is as great as that of other major medical disorders [<xref ref-type="bibr" rid="B16">16</xref>].</p></sec><sec><title>Subjective Well Being Scale (SWLS)</title><p>The SWLS is a short 5-item instrument designed to measure global life satisfaction. The scale has been validated and correlates with other measures of subjective well-being (SWB). The SWLS was developed to assess satisfaction with the respondent's life as a whole, without assessing satisfaction with specific life event [<xref ref-type="bibr" rid="B25">25</xref>].</p></sec><sec><title>EuroQoL 5D (EQ-5D)</title><p>The EQ-5D is a standardized generic instrument developed for describing and valuing health states. The EQ-5D was created for use in population health surveys or in conjunction with a condition-targeted instrument for assessment of outcomes related to specific health conditions or their treatment. It specifically refers to health status at the time of questioning. The first of 2 parts records a patient's health state along 5 dimensions (mobility, self-care, usual activities, pain/discomfort, and anxiety/depression). Each dimension has 3 levels reflecting no problem, some problem, and extreme problem. Respondents are asked to indicate one of the 3 levels along each of the 5 dimensions. This classifies respondents into 1 of 243 distinct health states. The second part of the EQ-5D is a 20-cm VAS that has the end points: "best imaginable health state" (100) and "worst imaginable health state"(0). Respondents are asked to illustrate how they rate their own health stat by drawing a line to that best represents their own health state on that day [<xref ref-type="bibr" rid="B25">25</xref>].</p></sec></sec><sec><title>Mixed QoL measures</title><sec><title>Salford Psoriasis Index (SPI)</title><p>The SPI is derived from combining a score of current severity of psoriasis based on the PASI, a score indicating psychosocial disability, and a score based on historical information. The resultant three- figure SPI (signs, psychosocial disability, interventions) is a similar paradigm to the TNM (tumor, nodes, metastasis) classification used for cancer staging [<xref ref-type="bibr" rid="B34">34</xref>].</p></sec><sec><title>Koo-Menter Psoriasis Instrument (KMPI)</title><p>The Koo-Menter Psoriasis Instrument (KMPI) is a diagnostic algorithm and a formal measure, to aid in identifying patients with significant impact on QoL warranting systemic therapy. In addition, the KMPI can be used to document and justify treatment decisions for health care payers. While the decision to undertake systemic treatment and the choice of specific treatment plan must ultimately be made mutually by the patient and the physician, these tools are designed to provide information that will be valuable in making informed decisions regarding treatments [<xref ref-type="bibr" rid="B35">35</xref>].</p></sec></sec><sec><title>Pharmacological treatments and their impact on QoL</title><p>Topical corticosteroids remain the mainstay of psoriasis therapy in the US. Steroid potencies range from class 7 steroids, such as 1% hydrocortisone, which is available in drug- without prescription, to superpotent class 1 corticosteroids such as clobestasol propionate, halobetasol propionate, betamethasone dipropionate [<xref ref-type="bibr" rid="B36">36</xref>]. The side effects of topical potent corticosteroids limit their use to an extent, and they are prescribed less frequently outside the US.</p><p>Primary treatments for severe psoriasis are phototherapy, systemic retinoids, methotrexate, cyclosporine and newer biological therapies. Ultraviolet B (UVB) phototherapy is an effective treatment for psoriasis and has been the safest way to maintain control of extensive psoriasis over the long-term. If UVB phototherapy is not sufficient to control a patient's psoriasis, then a combination of UVB plus the oral retinoid acitretin is often effective. These therapies often result in prolonged remission of varying duration; however, they are inconvenient for the patient. Advances in our understanding of the immune system in psoriasis have seen the development of biological agents which target molecular and cellular events leading to the disease. Alefacept was approved by the FDA for the treatment of psoriasis in January 2003. It inhibits the activation of T cells and reduces the number of activated memory T cells through apoptosis. As it causes reduction in activated memory T cells, T cells counts are performed to monitor for toxicity. Etanercept which antagonizes Tumor Necrosis Factor (TNF) activity was approved by the FDA in 2002. Studies using infliximab, an anti-TNF anti-body, also have shown efficacy in the treatment of psoriasis. TNF inhibitors are not FDA- approved for psoriasis, however, they are approved for other indications and physicians can prescribe them for psoriasis. Adalimumab is also other TNF inhibitor [<xref ref-type="bibr" rid="B37">37</xref>].</p><p>Different pharmacological treatments available for psoriasis are found to have varying effects on patients' QoL. Table <xref ref-type="table" rid="T1">1</xref> illustrates outcomes reported in various randomized controlled trials examining the effect of pharmacological intervention on patients QoL. Few studies have reported head to head comparisons of effect of different psoriasis treatments in improving QoL. More recently, studies are focused on understanding the effect of biologic agents on patients QoL, however none of them reported direct comparison of these agents. Clinical trials most often used DLQI or PDI to measure QoL. While very few trials utilized single QoL instrument, others used multiple instruments for QoL measurement. Although the exact reason for utilizing more than one measure for QOL was not mentioned in these studies, this may potentially be done to either compare measures or because no single measure was considered to be superior over the other with each offering a different aspect to the measure of QoL.</p></sec><sec><title>Physicians' role: what we need to know</title><p>Counseling patients with psoriasis may improve their mental and psychological condition. Such treatment should be aimed at increasing personal control, encouraging active coping strategies, restructuring negative thoughts about the disease, and encouraging patients to express emotions, seek social support and distract themselves [<xref ref-type="bibr" rid="B36">36</xref>]. Inducing remission and achieving reduction in severity of psoriasis (reduce area affected by psoriasis) may not be enough. Pharmacologic interventions should be accompanied by patient education and reassurance by family and social interventions [<xref ref-type="bibr" rid="B21">21</xref>].</p><p>As in treatment of other medical conditions, establishing a strong physician-patient relationship is the foundation of effective psoriasis treatment. Due to the recurring nature of the disease, patients are not only frustrated with the disease, but also with the care they receive or have received in the past. Physicians have to be empathetic and work with the patient to effectively manage their disease. Establishing this bond and trust between patient and physician will encourage patients to be more complaint to their physician's recommendations concerning treatment and will potentially improve treatment compliance and outcomes [<xref ref-type="bibr" rid="B38">38</xref>,<xref ref-type="bibr" rid="B39">39</xref>].</p><p>The physician may find many opportunities during the patient interview to establish this bond. To start, the physician should sit within touching distance of the patient and palpate the lesions as a part of physical examination. This act helps the patient overcome inhibitions regarding social interactions. Touching communicates that the patient is not untouchable and that the psoriasis does not have to impede intimate social relationships. Physicians should also ask a few leading questions about how psoriasis influences the patient's life. This communicates a sense of empathy and understanding that will assure the patient of the physician's competence in managing psoriasis.</p></sec></sec><sec><title>Conclusion</title><p>Psoriasis is a serious condition and is associated with significantly lower QoL. Studies have utilized different measures available to assess QoL of psoriasis patients. Most commonly used measures were psoriasis specific such as PASI and DLQI followed by generic measures such as SF-36. Pharmacological interventions along with patient counseling and education may be an effective strategy to improve QoL among psoriasis patients. Lack of head to head comparisons of available treatment options limits conclusions regarding superiority of one agent over another in improving QoL in psoriasis patients.</p></sec><sec><title>Competing interests</title><p>The Center for Dermatology Research is supported by an unrestricted educational grant from Galderma Laboratories. Dr. Feldman has received research, speaking and/or consulting support from Abbott, Amgen, Biogenidec, Centocor, Connetics, Genentech, and Roche.</p></sec><sec><title>Authors' contributions</title><p>Ms. Bhosle and Mr. Kulkarni were responsible for the literature review and writing of the manuscript. Dr. Feldman and Dr. Balkrishnan critically revised the manuscript.</p></sec> |
Publishing proteomic data | <p>Scientific publications should provide sufficient detail in terms of methodology and presented data to enable the community to reproduce the methodology to generate similar data and arrive at the same conclusion, if an identical sample is provided for analysis. The advent of high-throughput methods in biological experimentation impose some unique challenges both in data presentation in classical print format, as well as in describing methodology and data analysis in sufficient detail to conform to good publication practice. To facilitate this process, Proteome Science is adopting a set of methodology and data presentation guidelines to enable both peer reviewers, as well as the scientific community, to better evaluate high-throughput proteomic studies.</p> | <contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Latterich</surname><given-names>Martin</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>martin.latterich@mcgill.ca</email></contrib> | Proteome Science | <sec><title>Editorial</title><p>Historically, most scientific publications included a detailed methodology section that provided details on source of reagents, information, such as batch or lot numbers, and a description of methodology that would enable another research group to follow the same procedures. Given the same starting material, this practice would allow arriving at identical or very similar data. At the very least, methodology sections should refer to prior publications that provide sufficient experimental detail to allow the reproduction of scientific experiments. Most publications would then display "typical" results, such as photographs or micrographs of the experimental subject, images of detected molecules, or minimally processed data, such as statistically evaluated graphs or tables. These results were displayed together with negative and often positive controls that validate the experiment and reagents. The printed media was mostly adequate to publish these studies, because most studies investigated individual phenomena or molecules.</p><p>The advent of high-throughput methods in biological experimentation have imposed some unique challenges both in data presentation in classical print format, as well as in describing the methodology and data analysis workflow in sufficient detail to conform to good publication practice. This especially is an issue with proteomic analyses conducted by mass spectrometry [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B2">2</xref>]. Electronic media and public repositories are addressing the need for publishing uninterpreted data sets [<xref ref-type="bibr" rid="B3">3</xref>-<xref ref-type="bibr" rid="B5">5</xref>], such as raw or minimally processed mass spectrometer data, as well as lists of identified peptides. The remaining challenge is in the generation of ontologies and common experimental descriptions that capture the wealth of information that has both gone into the design and the analysis of proteomic experiments. This ultimately is needed when directly comparing multi-centre studies.</p><p>Much progress has been made by the community to propose data format standards that are compatible with most if not all analytical platforms [<xref ref-type="bibr" rid="B5">5</xref>,<xref ref-type="bibr" rid="B6">6</xref>]. However, there appears to be less conformity in the community when deciding what are minimal publication standards for such proteomic datasets [<xref ref-type="bibr" rid="B3">3</xref>]. The peer review process is normally rigorous enough to weed out submissions that are considered poor quality due to study design, choice of methods, or overall data quality. Unfortunately, dependent on the expertise of the peer reviewer, methodological detail is sometimes not considered as much as it should. While in the short term these studies will have their place in the community, in the long term they may not be considered valid because of lacking descriptive detail.</p><p>To ensure that studies of this nature withstand the test of time, Proteome Science has adopted a set of methodology and data presentation guidelines to enable both peer reviewers, as well as the scientific community, to better evaluate high-throughput proteomic studies leading to peptide and protein identification. These guidelines are by no-means top-down guidelines to restrict publication; instead, they are meant to reflect the accepted community standards in the field. As always with community guidelines, the publication guidelines proposed by the HUPO Proteomics Standards Initiative [<xref ref-type="bibr" rid="B7">7</xref>] will hopefully help to enable our authors to withstand the critique by the proteomics community over time. They are in no way intended to impose a standardized method to conduct experiments, which would be counterproductive to this still emerging and exciting field. We expect our authors to adhere to good scientific practise, such as listing source materials, methods of sample processing, the precise conditions to which samples were exposed prior to sampling, and the number of times an experiment has been conducted. In addition, if the authors use mass spectrometry to identify proteins in their samples, we recommend the adherence to the following guidelines to allow re-interpretation of the experimental data and comparison to other studies.</p><p>The following publication guidelines for the reporting and documentation of mass spectrometer-based peptide and protein identifications have in part been proposed by the HUPO Proteomics Standards Initiative [<xref ref-type="bibr" rid="B7">7</xref>,<xref ref-type="bibr" rid="B8">8</xref>]. They were heavily consulted during the development of publication recommendations for Proteome Science:</p><p>(a) Supporting information to be included in submitted manuscript:</p><p>1. Make, model, and version number of mass spectrometer, version of operating software, detailed acquisition parameters, and performance specs, such as resolution, sensitivity and dynamic range. If LC-MS/MS was performed, make, model, and version of HPLC system, operating specs on flow rate, gradients and columns used. Details on ionization source and conditions. Number of times experiment has been performed, and concordance between experiments.</p><p>2. The method(s), software (including version number) used to create the peak list from raw spectra, and the pertinent parameters used in the creation of the peak list. If custom algorithms or software were used to compile the list, these need to be listed in detail. Examples include parameters, such as smoothing, signal-to-noise ratio, whether charge states were calculated or peaks de-isotoped. In cases where additional customized processing of peak lists have been performed, such as clustering or filtering, the algorithm or software (including version) must be referenced or described.</p><p>3. The application and version number used for database searching, as well as the search parameters. Examples include precursor-ion mass tolerance, fragment-ion mass tolerance, fixed and variable modifications allowed for, number of missed cleavages, protein cleavage agents, isotopic or isobaric tagging chemistry, and so on.</p><p>4. The name and version of the sequence database and sequence space searched, including details on taxonomy and other search restrictions. If the database was custom compiled, a complete description of the sequence source is needed, and if not easily reproducible, a provision for making the database publicly available is needed. The number of entries actually searched from each database should be included. Authors should justify the use of very small databases, since this may generate misleading assignments. Common contaminants (keratins, trypsin) should be included in the database.</p><p>5. Methods used to interpret MS/MS data, thresholds and values specific to judging probability of identification, statistical methods used, and description of how analysis was validated, need to be described.</p><p>6. For large projects (e.g. mapping or comparison between complex fractions), additional statistical details should be listed, as they are pertinent to identification certainty, determination of false-positive rate, randomized database validation, or other computational approaches.</p><p>(b) When compiling information for protein identifications, the following information should be included:</p><p>1. Accession number and database source.</p><p>2. Score(s) and any statistical information for searches conducted.</p><p>3. Sequence coverage, expressed as the number of amino acids spanned by the assigned peptides by the intact protein's length.</p><p>4. The total number of peptides assigned to the protein. To compute this number, different forms of the same peptide are to be counted as one peptide.</p><p>(c) When presenting relative and absolute quantification data, the following information should be listed in addition to the information in section (b):</p><p>1. Quantification methods used and labelling conditions, if any.</p><p>2. Calibration standards used, if any.</p><p>3. Algorithms, software, and method details on how quantitative data was obtained.</p><p>4. Concordance and variance for different peptides from same protein, and method used to average individual peptides for a given protein.</p><p>5. Additional statistical information, such as p-values of quantified proteins, and common variance between experimental repeats.</p></sec><sec><title>Competing interests</title><p>The author(s) declares that they have no competing interests.</p></sec><sec><title>Authors' contributions</title><p>ML contributed 100 % to the Editorial Commentary.</p></sec> |
End of life content in geriatric textbooks: what is the current situation? | <sec><title>Background</title><p>Physicians caring for elderly people encounter death and dying more frequently than their colleagues in most other disciplines. Therefore we sought to examine the end-of-life content in popular geriatric textbooks and determine their usefulness in helping geriatricians manage patients at the end of their lives.</p></sec><sec sec-type="methods"><title>Methods</title><p>Five popular geriatric textbooks were chosen. Chapters on Alzheimer's disease, stroke, chronic heart failure, chronic obstructive pulmonary disease and lung cancer were examined because of their high mortality rates among the elderly patients. Text relevant to end-of-life care was highlighted. Two reviewers independently coded text into 10 pre specified domains and rated them for the presence of end-of-life information. Content was rated as absent, minimally helpful, or helpful. The proportion of helpful information was calculated.</p></sec><sec><title>Results</title><p>The textbook with the best end-of-life coverage contained 38% helpful information, the worst had only 15% helpful information. Minimally helpful information ranged from 24% to 50%. As much as 61% of the content in one textbook contained no helpful information at all. Of the ten domains, epidemiology, disease progression and prognostic factors were fairly well covered. Information on advance care planning, ethical issues, decision making and effects of death and dying on patient's family were generally lacking under the individual diseases though they were covered as general topics in other parts of the textbooks. All except one textbook dedicated a chapter to the care of the dying.</p></sec><sec><title>Conclusion</title><p>This study showed that end-of-life content in geriatric textbooks differed significantly. Most of the textbooks lack good coverage on end-of-life care and more can be done to improve on this.</p></sec> | <contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Wu</surname><given-names>Huei Yaw</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>huei_yaw_wu@ttsh.com.sg</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Malik</surname><given-names>Farida A</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>farida.malik@kcl.ac.uk</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Higginson</surname><given-names>Irene J</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>irene.higginson@kcl.ac.uk</email></contrib> | BMC Palliative Care | <sec><title>Background</title><p>Over the past three decades, end-of-life care has become an essential aspect of medical care in many parts of the world. Debates on physician-assisted suicide and euthanasia have further increased interest in palliative care in recent years.</p><p>Despite greater awareness amongst medical professionals on death and dying, the results of the Study to Understand Prognoses and Preferences for Outcomes and Risks of Treatment (SUPPORT) [<xref ref-type="bibr" rid="B1">1</xref>] in the US showed little evidence of a trend towards improved palliation. Studies in different countries have raised concerns regarding physicians' lack of knowledge on symptom control and national guidelines for such care, as well as poor communication between physicians and patients or their families, are barriers to the provision of good care at the end of life [<xref ref-type="bibr" rid="B2">2</xref>-<xref ref-type="bibr" rid="B4">4</xref>].</p><p>Geriatric Medicine emphasises the holistic care of elderly people. With a rapidly ageing global population, more people are currently dying from chronic diseases as compared to acute illnesses. This trend is projected to continue [<xref ref-type="bibr" rid="B5">5</xref>]. Most people who die are elderly. Given that most geriatric patients live out the last stage of life under the care of their doctors, there is an obligation for care-of-the-elderly physicians to be competent in the management of terminally-ill patients. However, improving the end-of-life experience of the dying older persons is complex. Many elderly people still suffer from difficult physical and emotional symptoms in their dying moments [<xref ref-type="bibr" rid="B6">6</xref>-<xref ref-type="bibr" rid="B8">8</xref>]. As many as one in three conscious elderly patients were found to experience severe pain in the last days of their lives [<xref ref-type="bibr" rid="B9">9</xref>].</p><p>The American Geriatrics Society's (AGS) position statement on "The care of dying patients" [<xref ref-type="bibr" rid="B10">10</xref>] reinforces the need to improve knowledge at all levels of training in end-of-life care amongst physicians and other health care professionals. Although technological advances have enabled easy access to medical information on the internet, textbooks still play an important role in medical education. Three studies conducted during the 1990's examining the end-of-life content in medical textbooks found that coverage was lacking in many disciplines of medicine [<xref ref-type="bibr" rid="B11">11</xref>-<xref ref-type="bibr" rid="B13">13</xref>]. Geriatric medicine textbooks published in the mid-nineties fared slightly better than other medical textbooks (12). Since then, some publishers and editors have responded positively to the call to improve the end-of-life content in their medical textbooks [<xref ref-type="bibr" rid="B14">14</xref>-<xref ref-type="bibr" rid="B16">16</xref>]. With so much recent emphasis on quality care for the terminally-ill and dying patients, we wished to examine the latest editions of some of the popular geriatric textbooks published in both the UK and the US, and to determine if the end-of-life content in these textbooks were up to the standard expected of them.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>Textbook selection</title><p>Five geriatric textbooks: "<italic>Brocklehurst's Textbook of Geriatric Medicine and Gerontology</italic>" (Tallis et al, sixth edition), "<italic>Oxford Textbook of Geriatric Medicine</italic>" (Evans et al, second edition), "<italic>Principles of Geriatric Medicine and Gerontology</italic>" (Hazzard et al, fifth edition), "<italic>The Merck Manual of Geriatrics</italic>" (Beers et al, third edition) and "<italic>Essentials of Clinical Geriatrics</italic>" (Kane et al, fifth edition), were reviewed. The first two were published in the UK, the latter three in the US. All the textbooks reviewed were the latest editions in order to ensure that the most current status of their end-of-life content was being assessed. The UK-published textbooks were chosen based on the results of a local survey conducted to determine the most commonly recommended geriatric medicine textbooks. This survey was carried out by doctors, of a group of physicians specialising in care of the elderly in a South London hospital. The three US-published textbooks were chosen from the list of top ten best-selling geriatric textbooks listed in the <italic>Amazon </italic>website [<xref ref-type="bibr" rid="B17">17</xref>].</p></sec><sec><title>Disease and content selection</title><p>Five medical conditions were examined. Four of these are predicted to be leading causes of death in 2020- stroke, chronic obstructive pulmonary disease (COPD), lung cancer and congestive heart failure [<xref ref-type="bibr" rid="B5">5</xref>]. We added Alzheimer's disease to our review because of its high incidence and prevalence among older people and the emotional and financial burden presented to both the patient's family [<xref ref-type="bibr" rid="B18">18</xref>,<xref ref-type="bibr" rid="B19">19</xref>] and society. For each disease, we assessed ten domains identified as important in end-of-life care: epidemiology (disease prevalence), disease progression, prognostic factors, symptom management (which encompasses both <italic>symptoms present at the end of life or advanced stage of the disease </italic>and <italic>medical intervention that improves the symptoms of the disease</italic>), mode of death, emotional/social/spiritual issues, advance care planning, ethical issues, decision making, and effect of death and dying on patient's family [see table <xref ref-type="table" rid="T1">1</xref>]. Eight of these domains were adapted from an earlier study by Carron et al [<xref ref-type="bibr" rid="B11">11</xref>]. Two domains, ethical and emotional/social/spiritual issues, were added because of their importance and relevance to palliative care.</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>Textbook coverage of end-of-life care</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td align="left"><bold>Brocklehurst's Textbook of Geriatric Medicine and Gerontology</bold></td><td align="left"><bold>Oxford Textbook of Geriatric Medicine</bold></td><td align="left"><bold>Principles of Geriatric Medicine and Gerontology-Hazzard et al.</bold></td><td align="left"><bold>The Merck Manual of Geriatrics</bold></td><td align="left"><bold>Essentials of Clinical Geriatrics- Kane</bold></td></tr></thead><tbody><tr><td align="left"><bold>Alzheimer's Disease</bold></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left"><italic>Epidemiology</italic></td><td align="center">++</td><td align="center">++</td><td align="center">++</td><td align="center">++</td><td align="center">++</td></tr><tr><td align="left"><italic>Disease progression</italic></td><td align="center">++</td><td align="center">+</td><td align="center">++</td><td align="center">++</td><td align="center">+</td></tr><tr><td align="left"><italic>Prognostic factors</italic></td><td align="center">0</td><td align="center">0</td><td align="center">0</td><td align="center">++</td><td align="center">0</td></tr><tr><td align="left"><italic>Symptom management</italic>:<break/><italic>a) Symptom present at eol/advanced stage</italic><break/><italic>b) Medical intervention that improves symptom</italic></td><td align="center">+<break/>+</td><td align="center">+<break/>+</td><td align="center">+<break/>+</td><td align="center">++<break/>0</td><td align="center">+<break/>+</td></tr><tr><td align="left"><italic>Mode of death</italic></td><td align="center">0</td><td align="center">0</td><td align="center">0</td><td align="center">+</td><td align="center">0</td></tr><tr><td align="left"><italic>Emotional, social and spiritual</italic></td><td align="center">++</td><td align="center">+</td><td align="center">0</td><td align="center">+</td><td align="center">++</td></tr><tr><td align="left"><italic>Advance care planning</italic></td><td align="center">+</td><td align="center">+</td><td align="center">0</td><td align="center">++</td><td align="center">++</td></tr><tr><td align="left"><italic>Ethics</italic></td><td align="center">+</td><td align="center">0</td><td align="center">0</td><td align="center">+</td><td align="center">+</td></tr><tr><td align="left"><italic>Decision making</italic></td><td align="center">+</td><td align="center">0</td><td align="center">0</td><td align="center">+</td><td align="center">+</td></tr><tr><td align="left"><italic>Effect of death and dying on family</italic></td><td align="center">++</td><td align="center">++</td><td align="center">+</td><td align="center">++</td><td align="center">++</td></tr><tr><td align="left"><bold>Stroke</bold></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left"><italic>Epidemiology</italic></td><td align="center">++</td><td align="center">++</td><td align="center">+</td><td align="center">++</td><td align="center">++</td></tr><tr><td align="left"><italic>Disease progression</italic></td><td align="center">++</td><td align="center">+</td><td align="center">++</td><td align="center">+</td><td align="center">+</td></tr><tr><td align="left"><italic>Prognostic factors</italic></td><td align="center">+</td><td align="center">++</td><td align="center">++</td><td align="center">++</td><td align="center">0</td></tr><tr><td align="left"><italic>Symptom management</italic>:<break/><italic>a) Symptom present at eol/advanced stage</italic><break/><italic>b) Medical intervention that improves symptom</italic></td><td align="center">+<break/>0</td><td align="center">++<break/>++</td><td align="center">+<break/>0</td><td align="center">++<break/>+</td><td align="center">+<break/>0</td></tr><tr><td align="left"><italic>Mode of death</italic></td><td align="center">+</td><td align="center">+</td><td align="center">+</td><td align="center">0</td><td align="center">0</td></tr><tr><td align="left"><italic>Emotional, social and spiritual</italic></td><td align="center">++</td><td align="center">+</td><td align="center">0</td><td align="center">+</td><td align="center">0</td></tr><tr><td align="left"><italic>Advance care planning</italic></td><td align="center">0</td><td align="center">0</td><td align="center">0</td><td align="center">++</td><td align="center">0</td></tr><tr><td align="left"><italic>Ethics</italic></td><td align="center">+</td><td align="center">++</td><td align="center">0</td><td align="center">+</td><td align="center">0</td></tr><tr><td align="left"><italic>Decision making</italic></td><td align="center">+</td><td align="center">+</td><td align="center">0</td><td align="center">+</td><td align="center">0</td></tr><tr><td align="left"><italic>Effect of death and dying on family</italic></td><td align="center">++</td><td align="center">++</td><td align="center">0</td><td align="center">0</td><td align="center">0</td></tr><tr><td align="left"><bold>Chronic Heart Failure</bold></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left"><italic>Epidemiology</italic></td><td align="center">++</td><td align="center">++</td><td align="center">++</td><td align="center">+</td><td align="center">0</td></tr><tr><td align="left"><italic>Disease progression</italic></td><td align="center">+</td><td align="center">0</td><td align="center">+</td><td align="center">+</td><td align="center">0</td></tr><tr><td align="left"><italic>Prognostic factors</italic></td><td align="center">+</td><td align="center">+</td><td align="center">++</td><td align="center">++</td><td align="center">0</td></tr><tr><td align="left"><italic>Symptom management</italic>:<break/><italic>a) Symptom present at eol/advanced stage</italic><break/><italic>b) Medical intervention that improves symptom</italic></td><td align="center">+<break/>+</td><td align="center">+<break/>+</td><td align="center">+<break/>+</td><td align="center">+<break/>+</td><td align="center">0<break/>+</td></tr><tr><td align="left"><italic>Mode of death</italic></td><td align="center">0</td><td align="center">0</td><td align="center">+</td><td align="center">0</td><td align="center">0</td></tr><tr><td align="left"><italic>Emotional, social and spiritual</italic></td><td align="center">0</td><td align="center">0</td><td align="center">+</td><td align="center">+</td><td align="center">0</td></tr><tr><td align="left"><italic>Advance care planning</italic></td><td align="center">0</td><td align="center">0</td><td align="center">++</td><td align="center">+</td><td align="center">0</td></tr><tr><td align="left"><italic>Ethics</italic></td><td align="center">0</td><td align="center">0</td><td align="center">++</td><td align="center">+</td><td align="center">0</td></tr><tr><td align="left"><italic>Decision making</italic></td><td align="center">0</td><td align="center">0</td><td align="center">++</td><td align="center">+</td><td align="center">0</td></tr><tr><td align="left"><italic>Effect of death and dying on family</italic></td><td align="center">0</td><td align="center">0</td><td align="center">+</td><td align="center">0</td><td align="center">0</td></tr><tr><td align="left"><bold>COPD</bold></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left"><italic>Epidemiology</italic></td><td align="center">++</td><td align="center">++</td><td align="center">++</td><td align="center">++</td><td align="center">-</td></tr><tr><td align="left"><italic>Disease progression</italic></td><td align="center">++</td><td align="center">+</td><td align="center">++</td><td align="center">++</td><td align="center">-</td></tr><tr><td align="left"><italic>Prognostic factors</italic></td><td align="center">++</td><td align="center">++</td><td align="center">+</td><td align="center">++</td><td align="center">-</td></tr><tr><td align="left"><italic>Symptom management</italic>:<break/><italic>a) Symptom present at eol/advanced stage</italic><break/><italic>b) Medical intervention that improves symptom</italic></td><td align="center">++<break/>+</td><td align="center">+<break/>+</td><td align="center">+<break/>+</td><td align="center">++<break/>++</td><td align="center">-<break/>-</td></tr><tr><td align="left"><italic>Mode of death</italic></td><td align="center">+</td><td align="center">+</td><td align="center">0</td><td align="center">+</td><td align="center">-</td></tr><tr><td align="left"><italic>Emotional, social and spiritual</italic></td><td align="center">++</td><td align="center">+</td><td align="center">+</td><td align="center">+</td><td align="center">-</td></tr><tr><td align="left"><italic>Advance care planning</italic></td><td align="center">0</td><td align="center">+</td><td align="center">0</td><td align="center">++</td><td align="center">-</td></tr><tr><td align="left"><italic>Ethics</italic></td><td align="center">0</td><td align="center">+</td><td align="center">0</td><td align="center">+</td><td align="center">-</td></tr><tr><td align="left"><italic>Decision making</italic></td><td align="center">0</td><td align="center">+</td><td align="center">0</td><td align="center">++</td><td align="center">-</td></tr><tr><td align="left"><italic>Effect of death and dying on family</italic></td><td align="center">0</td><td align="center">0</td><td align="center">0</td><td align="center">+</td><td align="center">-</td></tr><tr><td align="left"><bold>Lung Cancer</bold></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left"><italic>Epidemiology</italic></td><td align="center">++</td><td align="center">+</td><td align="center">++</td><td align="center">++</td><td align="center">-</td></tr><tr><td align="left"><italic>Disease progression</italic></td><td align="center">+</td><td align="center">+</td><td align="center">+</td><td align="center">+</td><td align="center">-</td></tr><tr><td align="left"><italic>Prognostic factors</italic></td><td align="center">0</td><td align="center">++</td><td align="center">++</td><td align="center">+</td><td align="center">-</td></tr><tr><td align="left"><italic>Symptom management</italic>:<break/><italic>a) Symptom present at eol/advanced stage</italic><break/><italic>b) Medical intervention that improves symptom</italic></td><td align="center">+<break/>+</td><td align="center">0<break/>+</td><td align="center">++<break/>++</td><td align="center">++<break/>++</td><td align="center">-<break/>-</td></tr><tr><td align="left"><italic>Mode of death</italic></td><td align="center">0</td><td align="center">0</td><td align="center">0</td><td align="center">0</td><td align="center">-</td></tr><tr><td align="left"><italic>Emotional, social and spiritual</italic></td><td align="center">+</td><td align="center">+</td><td align="center">+</td><td align="center">+</td><td align="center">-</td></tr><tr><td align="left"><italic>Advance care planning</italic></td><td align="center">0</td><td align="center">0</td><td align="center">0</td><td align="center">+</td><td align="center">-</td></tr><tr><td align="left"><italic>Ethics</italic></td><td align="center">0</td><td align="center">+</td><td align="center">0</td><td align="center">0</td><td align="center">-</td></tr><tr><td align="left"><italic>Decision making</italic></td><td align="center">0</td><td align="center">++</td><td align="center">0</td><td align="center">+</td><td align="center">-</td></tr><tr><td align="left"><italic>Effect of death and dying on family</italic></td><td align="center">0</td><td align="center">+</td><td align="center">+</td><td align="center">+</td><td align="center">-</td></tr></tbody></table><table-wrap-foot><p>'0'-no information on death and dying; '+' -score 1, includes minimally helpful information on death and dying; '++' -score 2, includes helpful information on care of the dying; '-' -disease not covered in textbook</p></table-wrap-foot></table-wrap></sec><sec><title>Review and rating procedures</title><p>Two clinicians, a palliative care physician (FM) and a care-of-the-elderly physician (WHY), independently reviewed the textbooks. For each textbook and each disease, any text mentioning or related to death and dying, the fatal nature of the condition or the course of the disease near death was identified. Text was included for assessment if it was identified by either reviewer, or of course, both.</p><p>The reviewers then independently categorised each piece of text according to the domain(s) covered (there could be more than one), and rated it for content. Ratings were '0' (no relevant information, score 0), '+' (minimally helpful information, score 1), and '++' (helpful information, score 2). For example, in the evaluation of the domain on "disease progression", a rating of '+' was given if there was any mention that the disease might lead to death and '++' if each stage of the illness leading to death was being discussed. For "symptom management", '+' was given if the symptoms that might be present at the advanced stage of the disease or the end of life were mentioned anywhere (however briefly) in the text, and '++' if the symptoms present at the end of life were specifically discussed. Any discrepancies or disagreements between the two reviewers in the coding of the domains were documented and referred to a third reviewer (IH).</p></sec><sec><title>Rating consistency</title><p>To ensure consistency of the rating process, prior to the assessments, the two reviewers coded and rated a different chapter in one of the five selected textbooks. This process identified and resolved potential differences.</p></sec><sec><title>Analysis</title><p>Content scores were available for each domain, for each condition. For the domain on symptom management, ratings for the subcategories were considered separately. The content score for each disease was calculated by summing the scores for the ten domains and expressing it as a ratio over a total score of 22 to allow comparison of the five diseases. The "helpful", "minimally helpful", and "absent" content scores were further analysed and expressed as a percentage across all ten domains. For instance, if the topic on lung cancer in a textbook had two ratings of '++', five of '+' and four of '0', the "helpful" content score is 18.2%, "minimally helpful" content score is 45.5% and "absent" content score is 36.3%.</p><p>Content rating ratios were calculated for each textbook, and for individual diseases within these textbooks, using the formula:</p><p><inline-formula><mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="M1" name="1472-684X-5-5-i1" overflow="scroll"><mml:semantics definitionURL="" encoding=""><mml:mrow><mml:mfrac><mml:mrow><mml:mtext>Score achieved</mml:mtext></mml:mrow><mml:mrow><mml:mtext>Total possible score</mml:mtext></mml:mrow></mml:mfrac></mml:mrow><mml:annotation encoding="MathType-MTEF">
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaadaWcaaqaaiabbofatjabbogaJjabb+gaVjabbkhaYjabbwgaLjabbccaGiabbggaHjabbogaJjabbIgaOjabbMgaPjabbwgaLjabbAha2jabbwgaLjabbsgaKbqaaiabbsfaujabb+gaVjabbsha0jabbggaHjabbYgaSjabbccaGiabbchaWjabb+gaVjabbohaZjabbohaZjabbMgaPjabbkgaIjabbYgaSjabbwgaLjabbccaGiabbohaZjabbogaJjabb+gaVjabbkhaYjabbwgaLbaaaaa@58D6@</mml:annotation></mml:semantics></mml:math></inline-formula></p><p>For example, a disease with a content score of 8 out of a possible score of 22 will have a content rating of 0.36. Similarly, a textbook with a total content score of 55 out of a possible score of 110 will be given a rating of 0.5.</p></sec></sec><sec><title>Results</title><p>Table <xref ref-type="table" rid="T1">1</xref> shows the coding and ratings of the 10 domains [see table <xref ref-type="table" rid="T1">1</xref>], and the appendix shows the chapters assessed [see <xref ref-type="supplementary-material" rid="S1">additional file 1</xref>]. The rating process was fairly consistent between the two reviewers, with an inter-rater variability of less than 6%. Ratings can be assessed for both the individual textbooks and the specific diseases [see table <xref ref-type="table" rid="T1">1</xref>].</p><sec><title>Analysis by textbook</title><p><italic>The Merck Manual of Geriatrics </italic>has the best coverage on end-of-life care on the whole, with a chapter for every disease and COPD containing the most helpful information across the ten domains studied. <italic>Oxford Textbook of Geriatric Medicine</italic>, <italic>Brocklehurst's Textbook of Geriatric Medicine and Gerontology</italic>, and <italic>Principles of Geriatric Medicine and Gerontology </italic>have relatively similar depth of coverage. Each textbook places different emphasis on different diseases. For instance, <italic>Oxford Textbook of Geriatric Medicine </italic>has good end-of-life content in the chapters on stroke and COPD whereas <italic>Principles of Geriatric Medicine and Gerontology </italic>covers the end-of-life care in chronic heart failure in greater depth. In contrast, <italic>Essentials of Clinical Geriatrics </italic>has little to offer on end-of-life care. It also does not have dedicated chapters on COPD and lung cancer.</p><p>The textbooks were ranked according to their content rating from the most to least end-of-life content: <italic>The Merck Manual of Geriatrics </italic>(0.63), <italic>Oxford Textbook of Geriatric Medicine </italic>(0.47), <italic>Brocklehurst's Textbook of Geriatric Medicine and Gerontology </italic>(0.45), <italic>Principles of Geriatric Medicine and Gerontology </italic>(0.45) and <italic>Essentials of Clinical Geriatrics </italic>(0.27) [see table <xref ref-type="table" rid="T2">2</xref>].</p><table-wrap position="float" id="T2"><label>Table 2</label><caption><p>Content rating of end-of-life care across the ten domains</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="center"><bold>Textbook</bold></td><td align="center"><bold>Disease</bold></td><td align="center"><bold>No information (%)</bold></td><td align="center"><bold>Minimally helpful information (%)</bold></td><td align="center"><bold>Helpful information present (%)</bold></td><td align="center"><bold>Content rating by disease</bold></td><td align="center"><bold>Content rating by textbook</bold></td></tr></thead><tbody><tr><td align="center">Brocklehurst's Textbook of Geriatric Medicine and Gerontology (6<sup>th </sup>edition, 2003)</td><td align="center">Alzheimer's Disease</td><td align="center">18</td><td align="center">45</td><td align="center">36</td><td align="center">0.59</td><td></td></tr><tr><td></td><td align="center">Stroke</td><td align="center">18</td><td align="center">45</td><td align="center">36</td><td align="center">0.59</td><td></td></tr><tr><td></td><td align="center">Chronic Heart Failure</td><td align="center">55</td><td align="center">36</td><td align="center">9</td><td align="center">0.27</td><td align="center">0.45</td></tr><tr><td></td><td align="center">COPD</td><td align="center">36</td><td align="center">18</td><td align="center">45</td><td align="center">0.55</td><td></td></tr><tr><td></td><td align="center">Lung Cancer</td><td align="center">55</td><td align="center">36</td><td align="center">9</td><td align="center">0.27</td><td></td></tr><tr><td colspan="7"><hr></hr></td></tr><tr><td align="center">Oxford Textbook of Geriatric Medicine (2<sup>nd </sup>edition, 2000)</td><td align="center">Alzheimer's Disease</td><td align="center">36</td><td align="center">45</td><td align="center">18</td><td align="center">0.41</td><td></td></tr><tr><td></td><td align="center">Stroke</td><td align="center">9</td><td align="center">36</td><td align="center">55</td><td align="center">0.73</td><td></td></tr><tr><td></td><td align="center">Chronic Heart Failure</td><td align="center">63</td><td align="center">27</td><td align="center">9</td><td align="center">0.23</td><td align="center">0.47</td></tr><tr><td></td><td align="center">COPD</td><td align="center">9</td><td align="center">73</td><td align="center">18</td><td align="center">0.55</td><td></td></tr><tr><td></td><td align="center">Lung Cancer</td><td align="center">27</td><td align="center">55</td><td align="center">18</td><td align="center">0.45</td><td></td></tr><tr><td colspan="7"><hr></hr></td></tr><tr><td align="center">Principles of Geriatric Medicine and Gerontology-Hazzard et al (5<sup>th </sup>edition, 2003)</td><td align="center">Alzheimer's Disease</td><td align="center">55</td><td align="center">27</td><td align="center">18</td><td align="center">0.32</td><td></td></tr><tr><td></td><td align="center">Stroke</td><td align="center">55</td><td align="center">27</td><td align="center">18</td><td align="center">0.32</td><td></td></tr><tr><td></td><td align="center">Chronic Heart Failure</td><td align="center">0</td><td align="center">55</td><td align="center">45</td><td align="center">0.73</td><td align="center">0.45</td></tr><tr><td></td><td align="center">COPD</td><td align="center">45</td><td align="center">36</td><td align="center">18</td><td align="center">0.36</td><td></td></tr><tr><td></td><td align="center">Lung Cancer</td><td align="center">36</td><td align="center">27</td><td align="center">36</td><td align="center">0.50</td><td></td></tr><tr><td colspan="7"><hr></hr></td></tr><tr><td align="center">The Merck Manual of Geriatrics (3<sup>rd </sup>edition, 2000)</td><td align="center">Alzheimer's Disease</td><td align="center">9</td><td align="center">36</td><td align="center">55</td><td align="center">0.73</td><td></td></tr><tr><td></td><td align="center">Stroke</td><td align="center">18</td><td align="center">45</td><td align="center">36</td><td align="center">0.59</td><td></td></tr><tr><td></td><td align="center">Chronic Heart Failure</td><td align="center">18</td><td align="center">73</td><td align="center">9</td><td align="center">0.45</td><td align="center">0.63</td></tr><tr><td></td><td align="center">COPD</td><td align="center">0</td><td align="center">36</td><td align="center">63</td><td align="center">0.82</td><td></td></tr><tr><td></td><td align="center">Lung Cancer</td><td align="center">18</td><td align="center">55</td><td align="center">27</td><td align="center">0.55</td><td></td></tr><tr><td colspan="7"><hr></hr></td></tr><tr><td align="center">Essentials of Clinical Geriatrics-Kane (5<sup>th </sup>edition, 2003)</td><td align="center">Alzheimer's Disease</td><td align="center">18</td><td align="center">45</td><td align="center">36</td><td align="center">0.59</td><td></td></tr><tr><td></td><td align="center">Stroke</td><td align="center">73</td><td align="center">18</td><td align="center">9</td><td align="center">0.18</td><td></td></tr><tr><td></td><td align="center">Chronic Heart Failure</td><td align="center">91</td><td align="center">9</td><td align="center">0</td><td align="center">0.05</td><td align="center">0.27</td></tr><tr><td></td><td align="center">COPD</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td></td></tr><tr><td></td><td align="center">Lung Cancer</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td></td></tr></tbody></table></table-wrap></sec><sec><title>Analysis by disease</title><p><italic>The Merck Manual of Geriatrics </italic>placed fairly equal emphasis on end-of-life content for all the five diseases. Beyond this, contact by disease varied greatly. An end of life content score of greater than 0.5 was achieved in <italic>Brocklehurst's Textbook of Geriatric Medicine and Gerontology </italic>for Alzheimer's disease, stroke and COPD; in the <italic>Oxford Textbook of Geriatric Medicine </italic>for stroke and COPD; in the <italic>Principles of Geriatric Medicine and Gerontology </italic>for chronic heart failure; and in <italic>Essentials of Clinical Geriatrics </italic>for Alzheimer's disease. Chronic heart failure faired fairly poorly in three of the five textbooks, with a score of 0.27 or less [see table <xref ref-type="table" rid="T2">2</xref>].</p></sec><sec><title>Analysis by domain</title><p>All the textbooks contained a fair amount of information on the domains of epidemiology, disease progression and prognostic factors of the diseases. The domains most lacking in helpful information were "mode of death", "advance care planning", "ethical issues", "decision making" and "effects of death and dying on patient's family" [see table <xref ref-type="table" rid="T1">1</xref>]. It was noteworthy that all the textbooks, except for <italic>Essentials of Clinical Geriatrics</italic>, had dedicated a chapter to the care of dying patients [see table <xref ref-type="table" rid="T3">3</xref> and <xref ref-type="supplementary-material" rid="S1">additional file 1</xref>]. Symptom management at the end-of-life was well covered in these chapters. Ethical issues, advance care planning, psychosocial issues, spiritual issues, end-of-life decision making and bereavement support were also covered either briefly in the same chapters or in greater detail in other parts of the textbooks. Two good examples were <italic>Principles of Geriatric Medicine and Gerontology </italic>and <italic>The Merck Manual of Geriatrics</italic>, both of which had in-depth coverage on ethical issues and spirituality in the elderly in two separate chapters [see table <xref ref-type="table" rid="T3">3</xref>]. It was interesting to find that only <italic>The Merck Manual of Geriatrics </italic>made cross-references to these chapters when discussing end-of-life issues under the individual diseases.</p></sec></sec><sec><title>Discussion</title><p>The end-of-life content varied substantially between different geriatric textbooks. Four textbooks contained a fair amount of end-of-life content but less than half of the information available was helpful. Of the five textbooks reviewed, <italic>The Merck Manual of Geriatrics </italic>contained the most end-of-life content. It also dedicated a small section at the end of each chapter on end-of-life issues. This is a reflection of the editorial staff's effort to improve the manual's coverage on end-of-life care following the feedback they received from Carron et al's study [<xref ref-type="bibr" rid="B11">11</xref>]. At the other end of the spectrum, <italic>Essentials of Clinical Geriatrics </italic>provided little information that would help the physician care for the dying patient. Information on end-of-life care was hardly available in chronic heart failure and stroke, the most common causes of death in older people. Life-challenging diseases commonly affecting the elderly population like COPD and lung cancer were not covered in this textbook.</p><p>Four out of five textbooks dedicated a chapter to the care of the dying patient. However, most of them did not make cross-references to this chapter elsewhere in the text. Including a chapter on palliative care cannot completely make up for the textbooks' lack of end-of-life content under the individual diseases. Most busy clinicians and students who refer to textbooks will probably be selective in their reading. Therefore, the only way for them to learn more about the end-of-life care of a disease is when such information is available in the same chapter or when cross-references to end-of-life content elsewhere in the textbook are made. In our review, <italic>The Merck Manual of Geriatrics </italic>was the only textbook that made cross-references to chapters on care of the dying and ethical issues at the end of life.</p><p>Many clinicians associate palliative care with cancer and so might expect textbooks to provide more end-of-life information in the chapters on cancer. This was not supported by our findings. The end-of-life content in lung cancer was generally poor in all the textbooks reviewed. In contrast, the chapters on stroke and COPD contained more end-of-life content in three of the textbooks (<italic>Brocklehurst's Textbook of Geriatric Medicine and Gerontology</italic>, <italic>Oxford Textbook of Geriatric Medicine </italic>and <italic>The Merck Manual of Geriatrics</italic>) while Alzheimer's disease had better coverage on end-of-life care in <italic>Brocklehurst's Textbook of Geriatric Medicine and Gerontology </italic>and <italic>The Merck Manual of Geriatrics </italic>compared to lung cancer. This is encouraging as we now know that non-cancer conditions such as congestive cardiac failure and COPD have the same palliative care needs as cancers [<xref ref-type="bibr" rid="B20">20</xref>,<xref ref-type="bibr" rid="B21">21</xref>]. By placing more emphasis on end-of-life care in these non-cancer conditions, textbooks can help to highlight the needs of older persons dying from these diseases.</p><p>Our study showed that epidemiology of diseases, disease progression and prognostic factors were better covered in the textbooks compared to the other aspects of end-of-life care. Information on advance care planning, decision-making and effect of death and dying on family were often minimally helpful or absent. Even if these issues were discussed, they were usually found in a separate chapter, with no cross-references. Ethical issues at the end of life and bereavement support were usually given insufficient emphasis at the undergraduate level [<xref ref-type="bibr" rid="B19">19</xref>] even though they were important measures of quality care at the end of life [<xref ref-type="bibr" rid="B20">20</xref>]. Sadly, these areas were often overlooked in clinical practice as well. The SUPPORT study showed that more than half of the patients who died in hospitals were given aggressive treatment in their last days although they would have preferred comfort measures [<xref ref-type="bibr" rid="B8">8</xref>]. With advance care planning and end-of-life decision-making receiving more attention in the recent years, it is encouraging to note that many geriatric textbooks have special coverage on these topics.</p><p>In their study on end-of-life care content in textbooks from multiple disciplines [<xref ref-type="bibr" rid="B12">12</xref>], Rabow et al. reviewed the previous editions of <italic>The Merck Manual of Geriatrics </italic>and <italic>Principles of Geriatric Medicine and Gerontology </italic>(Hazzard et al). As a point of interest, we noted that the end-of-life content in <italic>The Merck Manual of Geriatrics </italic>had improved significantly in the latest edition (third edition). Almost half of the text reviewed in the previous edition (second edition) contained no helpful information on end-of-life care while our results showed that only 13% of the text in the latest edition had no helpful information. <italic>Principles of Geriatric Medicine and Gerontology </italic>also appeared to contain more 'minimally helpful' information and less 'not helpful' information in the fifth edition as compared to the fourth edition. However, we need to be cautious in interpreting and drawing conclusions from these results as the diseases and domains studied were different. Also, the reviewers in the two studies may not have rated the contents similarly.</p><p>In the course of our study, we found that <italic>Oxford Textbook of Geriatric Medicine </italic>and <italic>The Merck Manual of Geriatrics </italic>were available electronically on the internet. Clinicians can now have easy access to these textbooks and will be able to refer to them in the wards when queries arise in the middle of their ward rounds, provided computer terminals are available. We hope that the publishers of the other geriatric textbooks will also consider making theirs available online as this will certainly facilitate learning among clinicians and medical students.</p><p>This study was limited in that the domain ratings assigned may be unreliable. We tried to overcome this by setting clear definitions for each domain and agreeing on how each piece of text should be rated in the respective domains. This study was also limited by the subjective nature of the methodology. However, with any assessment of medical textbooks, the methodology will nearly always be subjective. Indeed when using textbooks to guide clinical practice, healthcare professionals read the content subjectively. In order to improve inter-rater consistency, we carried out an exercise to improve reliability prior to the actual text review. However, as the two reviewers came from different backgrounds and training, it was interesting to note how the two reviewers viewed some aspects of end-of-life care. For instance, the palliative care physician felt that discussion of the "multidisciplinary approach" of management at the end of life in the text could be considered in the end-of-life content. The care of the elderly physician thought otherwise as he felt that this approach was widely adopted in geriatric medicine and not specific to palliative care. Another example was in the rating of the domains. The care of the elderly physician was less stringent in the rating of symptom management as compared to the palliative care physician who expected more in-depth discussion in this area. In spite of these initial differences in opinion, there was only 6% discrepancy, which was resolved by the third reviewer. Having two reviewers from different medical disciplines was a benefit to this study, as it provided different perspectives and helped to reduce potential blind spots that would have otherwise gone undetected.</p><p>Future research might include similar studies assessing current end of life content in major journals and how this has changed over recent years, or assessing the end of life content in textbooks and journals since the inception of educational resource sites online such as the EPEC project [<xref ref-type="bibr" rid="B24">24</xref>] (Education in Palliative and End of Life Care) and EPERC (End of Life/Palliative Education Resource Centre) [<xref ref-type="bibr" rid="B25">25</xref>]. It might have been interesting to interview textbook editors or chapter authors to understand the differences in the end of life content of their text books; however this would not have necessarily been representable or generalisable to the population of health care professionals who would read the textbooks and would also be subjective.</p></sec><sec><title>Conclusion</title><p>The end-of-life content differed significantly in different geriatric textbooks. Important aspects of end-of-life care such as advance care planning, decision-making and effects of death and dying on the family/carers were not given enough emphasis under the individual diseases. Four out of five textbooks provided specific chapters on care of the dying and end-of-life ethical issues, though cross-references were usually not made to these chapters. It was noteworthy that non-cancer conditions received similar, if not better end-of-life coverage, as cancer in geriatric textbooks. It is hoped that the results of this study will encourage the editors and publishers to improve on the end-of-life content in their textbooks, and in so doing, indirectly influence the quality of the care older persons are going receive in the last days of their lives.</p></sec><sec><title>Competing interests</title><p>The author(s) declare that they have no competing interests.</p></sec><sec><title>Authors' contributions</title><p>WHY participated in design of the study, the rating of the contents of the textbooks, interpreting the results and preparing the manuscript. FAM participated in the rating of the contents of the textbooks, reading, amending and approving the manuscript, corresponding with the editor and answering the reviewers' queries. IJH participated in the design of the study, was the third independent rater when there were conflicting opinions regarding the EOL content rating, participated in reading, amending and approving the manuscript. All authors read and approved the final manuscript.</p><table-wrap position="float" id="T3"><label>Table 3</label><caption><p>Chapters dedicated to end-of-life care or to issues related to the end of life</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left"><bold>Textbook</bold></td><td align="left"><bold>Chapter</bold></td><td align="center"><bold>End of Life Issues Covered</bold></td></tr></thead><tbody><tr><td align="left">Brocklehurst's Textbook of Geriatric Medicine and Gerontology (6<sup>th </sup>Edition)</td><td align="left">1)Palliative Care<break/>2)Ethical Issues in Geriatric Medicine</td><td align="left">-Symptom management<break/>-Advance care planning<break/>-Decision making</td></tr><tr><td align="left">Oxford Textbook of Geriatric Medicine (2<sup>nd </sup>Edition)</td><td align="left">1)Symptom Management and Palliative Care<break/>2) Legal and Ethical Issues in Geriatric Medicine</td><td align="left">-Symptom management<break/>-Ethical issues<break/>-Advance care planning<break/>-Psychological and spiritual issues<break/>-Bereavement support<break/>-Advance care planning<break/>-Decision making</td></tr><tr><td align="left">Principles of Geriatric Medicine and Gerontology (5<sup>th </sup>Edition)</td><td align="left">1)Care of the Dying Patient<break/>2)Ethical issues in Geriatric Medicine<break/>3)Spirituality and the Elderly</td><td align="left">-Symptom management<break/>-Decision making<break/>-Bereavement support<break/>-Decision making<break/>-Advance care planning<break/>-Withdrawing and withholding treatment<break/>-Euthanasia and assisted suicide<break/>-Spiritual issues</td></tr><tr><td align="left">The Merck Manual of Geriatrics (3<sup>rd </sup>Edition)</td><td align="left">1)Care of the Dying Patient<break/>2)Legal and ethical issues<break/>3)Social issues</td><td align="left">-Symptom management<break/>-Emotional, social and spiritual issues<break/>-Ethical issues<break/>-Advance care planning<break/>-Bereavement support<break/>-Advance directives<break/>-Decision making<break/>-Euthanasia and assisted suicide<break/>-Spiritual issues</td></tr><tr><td align="left">Essentials of Clinical Geriatrics (5<sup>th </sup>Edition)</td><td align="left">1)Terminal Care*<break/>2)Ethical issues in the care of older persons</td><td align="left">-Psychological issues<break/>-Advance directives<break/>-Decision making</td></tr></tbody></table><table-wrap-foot><p>* a section within the chapter "Developing Clinical Expectations"</p></table-wrap-foot></table-wrap></sec><sec><title>Pre-publication history</title><p>The pre-publication history for this paper can be accessed here:</p><p><ext-link ext-link-type="uri" xlink:href="http://www.biomedcentral.com/1472-684X/5/5/prepub"/></p></sec><sec sec-type="supplementary-material"><title>Supplementary Material</title><supplementary-material content-type="local-data" id="S1"><caption><title>Additional File 1</title><p>1. Appendix : An overview of the chapters covered in the study. The appendix provides an overview of the chapters covered in the study.</p></caption><media xlink:href="1472-684X-5-5-S1.doc" mimetype="application" mime-subtype="msword"><caption><p>Click here for file</p></caption></media></supplementary-material></sec> |
Protective effect of EDTA preadministration on renal ischemia | <sec><title>Background</title><p>Chelation therapy with sodium edetate (EDTA) improved renal function and slowed the progression of renal insufficiency in patients subjected to lead intoxication. This study was performed to identify the underlying mechanism of the ability of EDTA treatment to protect kidneys from damage.</p></sec><sec sec-type="methods"><title>Methods</title><p>The effects of EDTA administration were studied in a rat model of acute renal failure induced by 60 minutes ischemia followed or not by 60 minutes reperfusion. Renal ischemic damage was evaluated by histological studies and by functional studies, namely serum creatinine and blood urea nitrogen levels. Treatment with EDTA was performed 30 minutes before the induction of ischemia. Polymorphonuclear cell (PMN) adhesion capability, plasmatic nitric oxide (NO) levels and endothelial NO synthase (eNOS) renal expression were studied as well as the EDTA protection from the TNFα-induced vascular leakage in the kidneys. Data was compared by two-way analysis of variance followed by a post hoc test.</p></sec><sec><title>Results</title><p>EDTA administration resulted in the preservation of both functional and histological parameters of rat kidneys. PMN obtained from peripheral blood of EDTA-treated ischemized rats, displayed a significant reduction in the expression of the adhesion molecule Mac-1 with respect to controls. NO was significantly increased by EDTA administration and eNOS expression was higher and more diffuse in kidneys of rats treated with EDTA than in the controls. Finally, EDTA administration was able to prevent in vivo the TNFα-induced vascular leakage in the kidneys.</p></sec><sec><title>Conclusion</title><p>This data provides evidence that EDTA treatment is able to protect rat kidneys from ischemic damage possibly through the stimulation of NO production.</p></sec> | <contrib id="A1" contrib-type="author"><name><surname>Foglieni</surname><given-names>Chiara</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>foglieni.chiara@hsr.it</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Fulgenzi</surname><given-names>Alessandro</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>alessandro.fulgenzi@unimi.it</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Ticozzi</surname><given-names>Paolo</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>paolo.ticozzi@libero.it</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Pellegatta</surname><given-names>Fabio</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>fabio.pellegatta@unimi.it</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Sciorati</surname><given-names>Clara</given-names></name><xref ref-type="aff" rid="I4">4</xref><email>sciorati.clara@hsr.it</email></contrib><contrib id="A6" contrib-type="author"><name><surname>Belloni</surname><given-names>Daniela</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>daniela.belloni@unimi.it</email></contrib><contrib id="A7" contrib-type="author"><name><surname>Ferrero</surname><given-names>Elisabetta</given-names></name><xref ref-type="aff" rid="I4">4</xref><email>ferrero.elisabetta@hsr.it</email></contrib><contrib id="A8" corresp="yes" contrib-type="author"><name><surname>Ferrero</surname><given-names>Maria Elena</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>mariaelena.ferrero@unimi.it</email></contrib> | BMC Nephrology | <sec><title>Background</title><p>Chelation therapy with sodium edetate (EDTA) has been successfully used to treat chronic lead intoxication [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B2">2</xref>]. More specifically, in patients affected by chronic renal insufficiency due to environmental lead exposure, EDTA chelation therapy improved renal function and slowed the progression of renal insufficiency [<xref ref-type="bibr" rid="B3">3</xref>]. The mechanism by which lead-chelation therapy with EDTA delayed renal damage is unknown. Chelation with another chelating agent, the dimercaptosuccinic acid (DMSA) improved renal function and was efficacious in treating nephropathy [<xref ref-type="bibr" rid="B4">4</xref>] and hypertension [<xref ref-type="bibr" rid="B5">5</xref>], both induced in animals by long-term exposure to low-levels of lead. It has been proposed that chronic, low-level lead exposure may increase the levels of reactive oxygen species (ROS), responsible for nitric oxide (NO) inactivation [<xref ref-type="bibr" rid="B6">6</xref>]. Indeed, lead-chelation therapy might reduce the levels of ROS, associated to NO inactivation, and thus enhance the availability of vascular NO, potentially improving renal function and reducing hypertension [<xref ref-type="bibr" rid="B4">4</xref>-<xref ref-type="bibr" rid="B6">6</xref>]. Moreover, a multifunctional antioxidant activity has been shown for an iron chelating agent, the N,N'-bis (2-hydroxybenzyl) ethylendiamine-N,N'-diacetic acid (HBED) [<xref ref-type="bibr" rid="B7">7</xref>]. We asked if EDTA treatment in rats was able to reduce the renal damage, when not provoked by lead exposure. Indeed, in the present work we have studied the effect of EDTA treatment in preventing rat kidney acute damage following ischemia (Isc) or ischemia/reperfusion (Isc/R) [<xref ref-type="bibr" rid="B8">8</xref>,<xref ref-type="bibr" rid="B9">9</xref>].</p><p>We assessed the effect of EDTA systemically administered in rats, before the induction of renal Isc or Isc/R. Functional and histological kidney alterations and rat plasmatic levels of NO were evaluated, given that NO availability has been found to be responsible for the increased renal function [<xref ref-type="bibr" rid="B4">4</xref>,<xref ref-type="bibr" rid="B6">6</xref>]. In addition, being NO able to control leukocyte adhesion [<xref ref-type="bibr" rid="B10">10</xref>], we determined the expression of the adhesion molecule Mac-1 (monocyte chemoattractant protein-1) (CD18/CD11b) on polymorphonuclear cells (PMN) isolated from control and EDTA-treated rats. In this context, it has been shown that PMN are able to play an important role as mediators of reperfusion injury [<xref ref-type="bibr" rid="B11">11</xref>,<xref ref-type="bibr" rid="B12">12</xref>]. Finally, since endothelial NO production is an indicator of well functioning endothelium [<xref ref-type="bibr" rid="B10">10</xref>], we have evaluated the effect of EDTA in TNFα-induced vascular leakage in rat kidneys.</p><p>Herein we show that a single administration of EDTA results in the preservation of renal function and in the prevention of tissue damage induced by ischemic injury. In addition, we demonstrate that the preventive block of NO synthesis abrogate the protective effect of EDTA against renal ischemic damage.</p></sec><sec sec-type="methods"><title>Methods</title><p>The investigation conforms with the <italic>Guide for the Care and Use of Laboratory Animals </italic>published by the US National Institute of Health (NIH publication NO.85-23, revised 1996), according to the animal welfare regulations of the Italian local authorities.</p><sec><title>Animals</title><p>Male Sprague-Dawley rats weighing about 200 g were used (Charles River Italia, Lecco, Italy) and were allowed water and standard rat chow ad libitum. All the rats were maintained at 22 ± 1°C with a 12/12 hours light/dark cycle.</p></sec><sec><title>Ischemia/Reperfusion (Isc/R) model</title><p>The rats were anesthetized with an inhaled anesthesia mixture of halothane 2% (Hoechst, Milano, Italy) and oxygen. They were placed on a temperature-regulated table (38°C) (Ugo Basile, Comerio, Lecco, Italy) to maintain body temperature. Kidney ischemia (Isc) was induced by clamping the right renal artery and the right renal vein for 60 minutes with a microsurgical clamp. In the Isc/R group, at the end of the I period, the vascular clamp was removed and reperfusion of 60 minutes was performed. During the surgical procedure the heart rate and the mean arterial blood pressure (MABP) were monitored.</p><p>At the end of Isc or of Isc/R, blood samples were obtained by exanguination of rats at the aorta bifurcation level and kidneys were collected and processed for different studies. Blood and kidneys from EDTA-treated-not-ischemized rats were collected 90 minutes after EDTA administration (corresponding to 30 minutes EDTA pre-treatment+60 min Isc).</p></sec><sec><title>Measurement of mean arterial blood pressure</title><p>The right femoral artery was cannulated through a polyethylene catheter and connected to a pressure transducer for the measurement of MABP [<xref ref-type="bibr" rid="B15">15</xref>,<xref ref-type="bibr" rid="B16">16</xref>]. The data was collected continuously by means of a computer and were calculated at baseline, at the end of EDTA pre-treatment (e.g 30 minutes after EDTA intravenous injection), at the end of Isc and at the end of postischemic R. In sham-operated rats the values were calculated 90 minutes after EDTA pre-treatment.</p></sec><sec><title>EDTA treatment</title><p>EDTA (calcium disodium EDTA) (Collalto, Brescia, Italy) used in human therapy was employed [<xref ref-type="bibr" rid="B3">3</xref>], and at the same dosage (e.g. 40 mg/kg body weight). The sterile drug solution of 2 g/10 ml was opportunely diluted in physiological saline and administered by left intrafemoral vein slow infusion.</p></sec><sec><title>L-NAME treatment</title><p>The inhibitor of NO synthases L-NAME [N(omega)-nitro-L-arginine methyl ester], when required, was injected simultaneously with the EDTA through the intrafemoral vein at the dose of 30 mg/kg body weight, 30 minutes before the induction of Isc or Isc/R.</p></sec><sec><title>Experimental groups</title><p>The rats were randomly allocated to 4 study groups, each composed of 15 rats: group 1, controls; group 2, sham operated: the rats underwent the same surgical procedure, except that the clamp was not applied; group 3, Isc: ischemia was induced for 60 min; group 4, Isc/R: ischemia was induced for 60 min, followed by 60 min reperfusion at room temperature. Other identical 4 groups were studied, in which EDTA treatment was performed. In groups 3 and 4 intrafemoral injection of physiological saline 30 minutes before clamping was performed. The 3 and 4 EDTA-treated groups received a single intravenous injection of EDTA 30 minutes before clamping. In groups 1 and 2 intrafemoral injection of physiological saline or EDTA was performed 90 minutes before kidney removal (= 30 minutes EDTA pre-treatment+60 min Isc).</p><p>To take in consideration that EDTA could lead to increase in NO plasmatic levels through increase in eNOS expression, we further performed histological evaluations on two additional groups of rats, to verify whether the eNOS inhibitor L-NAME was able to block the protective effect of EDTA in renal ischemic injury. In such groups the animals were simultaneously treated with EDTA and L-NAME 30 minutes before the induction of Isc (group 5) and 30 minutes before the induction of Isc/R (group 6).</p></sec><sec><title>Functional studies</title><p>Serum creatinine was measured using a modified Jaffe's reaction, and blood urea nitrogen was measured on the AEROSET system (Abbott Laboratories, Abbott Park, IL) [<xref ref-type="bibr" rid="B17">17</xref>].</p></sec><sec><title>Histopathology and immunofluorescence microscopy</title><p>Kidneys were excised, decapsulated, dissected into 4 pieces along the major ax, fixed by immersion in 4% paraformaldehyde in Dulbecco's PBS (DPBS) overnight at 4°C, cryo-protected in 10% sucrose in DPBS, then embedded in Tissue-Tek medium and frozen in liquid nitrogen. Cryostat-cut four sections/animal (5 μm thick) were submitted to Hematoxylin/Eosin stain; renal damage was evaluated as tubular epithelial cell necrosis, tubular dilation, protein casts and medullary congestion (18). The alterations were semi-quantitatively graded by a pathologist blind to the nature of the experiments. The grading was performed by the following criteria: - =absent, + = barely present, ++ = moderate, +++ = severe. Expression of eNOS, e.g. the endothelial form of the constitutive NO synthase, was assessed on serial sections, with the use of a specific monoclonal antibody (BD Pharmingen, Franklin Lakes, NJ), followed by a Rabbit-anti-Mouse IgG- AlexaFluor488 (Molecular Probes, Eugene, OR). Observations were performed by using an Eclipse 55i microscope (Nikon, Tokyo, Japan), digital images acquired with DS-L1 camera and LUCIA G software (all from Nikon) and mounted using AdobePhotoshop CS software.</p></sec><sec><title>Cytofluorimetry</title><p>The expression of Mac-1 was evaluated by following FACS analysis. Whole blood was incubated with 0.5 μg of FITC-conjugated CD11b monoclonal antibody (clone WT5, isotype mouse IgA, K) (Pharmingen, San Diego, CA) for 20 minutes in ice. After erythrocyte lysis, samples were run on a FACscan (Becton-Dickinson, Mountain View, CA) and gated on PMN parameters. Results are expressed as arbitrary units of mean fluorescence intensity (MFI, a.u.).</p></sec><sec><title>Nitrite/Nitrate (NO2<sup>-</sup>/NO3<sup>-</sup>) determination</title><p>The rats were bled off at the aorta bifurcation level. Blood was collected in the presence of 0.065 mM citric acid (Riedel, Hannover, Germany), 0.085 mM sodium citrate (Farmitalia, Milan, Italy) and 2% glucose monohydrate (Riedel) in the blood: anticoagulant ratio of: 7:1. Samples were obtained from rats immediately after the end of each treatment or surgical procedure.</p><p>NO release was determined spectrophotometrically [<xref ref-type="bibr" rid="B19">19</xref>] by measuring the nitrate/nitrite (NO<sub>2</sub><sup>-</sup>/NO<sub>3</sub><sup>-</sup>) concentration in plasma samples from arterial non coagulated blood. Briefly, whole blood was centrifuged and plasma samples were collected, incubated for 30 min at 37°C in the presence of 0.2 U/ml <italic>Aspergillus </italic>nitrate reductase (Boehringer-Mannheim, Milan, Italy), 50 mM HEPES buffer (pH 7.4), 5 μM flavin adenine dinucleotide (Sigma Aldrich), and 0.1 mM NADPH (Sigma Aldrich). Then, lactate dehydrogenase (Boehringer Mannheim) and sodium pyruvate (Sigma Aldrich) were added to a final concentration of 10 U/ml and 10 mM, respectively, and the samples were incubated for 10 minutes at 37°. The Griess reagent (Sigma Aldrich) was added to the samples (100 μl), and absorbance was measured at 540 nm after 15 minutes incubation at room temperature. Standard curves with increasing concentrations of sodium nitrate and sodium nitrite were run in parallel.</p></sec><sec><title>In vivo permeability assay</title><p>The assay was performed as described [<xref ref-type="bibr" rid="B20">20</xref>]. Briefly, the exit of albumin from vessels into the parenchyma of rat kidneys was assayed. The dye solution contained 0.4% albumin (Sigma Aldrich) and 0.5% trypan blue (Sigma Aldrich) in saline. Following laparatomy, animals were perfused with 5 ml dye-solution through the right renal artery for 10 minutes. The perfusate was drawn from the right renal vein. The right kidney was washed with saline <italic>in vivo</italic>, removed, weighted, suspended and homogenized in buffered phosphate solution at pH 7.4 (1 g tissue dissolved in 3 ml buffer). In treated animals, after halothane anesthesia, EDTA (40 mg/kg) was injected intravenously (through the femoral vein), followed by rat TNFα (R&D System, Abingdon, UK) (0.1 ng/g). TNFα and EDTA, alone or together, were injected 30 minutes before kidney dye perfusion.</p><p>Tissue extracts were centrifuged, the supernatants recovered and treated with 10% deoxycholic acid (sodium salt monohydrate, Sigma Aldrich) in saline, to remove lipid interference. Dye was evaluated by spectrophotometer analysis (Pye Unicam SP6-550, Cambridge, United Kingdom) at 540 nm.</p></sec><sec><title>Statistics</title><p>The results are expressed as the mean ± SEM of 15 animals in each group. They were analyzed using a two way analysis of variance followed by Bonferroni t-test. The results were considered statistically significant when p < 0.05.</p></sec></sec><sec><title>Results</title><sec><title>Systemic hemodynamic data</title><p>In the rats studied the heart rate did not vary significantly during the experimental procedure (data not shown). To establish whether EDTA could maintain vascular homeostasis, we measured MABP in both untreated and EDTA-treated rats. EDTA treated rats displayed MABP values lower than those of untreated control and sham operated animals. To note, the increase of MABP due to Isc was significantly prevented by EDTA pre-administration (Table <xref ref-type="table" rid="T1">1</xref>).</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>Measure of mean arterial blood pressure (MABP) in rats</p></caption><table frame="hsides" rules="groups"><tbody><tr><td align="left"><bold>CONTROLS SHAM</bold></td><td align="center" colspan="2"><bold>UNTREATMENT (mmHg)</bold></td><td align="center" colspan="2"><bold>EDTA pre-treatment (mmHg)</bold></td></tr><tr><td></td><td colspan="4"><hr></hr></td></tr><tr><td></td><td align="center" colspan="2">100 ± 8</td><td align="center" colspan="2">85 ± 3*</td></tr><tr><td></td><td align="center" colspan="2">108 ± 11</td><td align="center" colspan="2">93 ± 2*</td></tr><tr><td></td><td colspan="4"><hr></hr></td></tr><tr><td></td><td align="center"><bold><italic>Before clamping</italic></bold></td><td align="center"><bold><italic>End Isc or Isc/R</italic></bold></td><td align="center"><bold><italic>Before clamping</italic></bold></td><td align="center"><bold><italic>End Isc or Isc/R</italic></bold></td></tr><tr><td></td><td colspan="4"><hr></hr></td></tr><tr><td align="left"><bold>Isc</bold></td><td align="center">104 ± 6</td><td align="center">130 ± 5**</td><td align="center">90 ± 2*</td><td align="center">98 ± 7*</td></tr><tr><td align="left"><bold>Isc/R</bold></td><td align="center">105 ± 6</td><td align="center">115 ± 9</td><td align="center">90 ± 3*</td><td align="center">88 ± 8*</td></tr></tbody></table><table-wrap-foot><p>EDTA pre-administration (30 min) is able to avoid the increase of MABP induced by kidney Isc. EDTA administration reduces MABP in controls and in sham-operated rats. lsc = ischemia; lsc/R = ischemia/reperfusion.</p><p>*p < 0.05 vs. corresponding untreatment; **p < 0.05 vs Isc before clamping</p></table-wrap-foot></table-wrap></sec><sec><title>EDTA administration preserved kidneys from ischemic damage</title><p>Rats undergoing either renal Isc, obtained by clamping the right renal artery and the right renal vein for 60 minutes, or Isc followed by 60 minutes reperfusion (Isc/R), obtained by removing the clamp, were evaluated for the levels of serum creatinine and blood urea nitrogen (Fig. <xref ref-type="fig" rid="F1">1</xref>), two parameters routinely used to assess renal function. Both creatinine and urea had a significant increase after the induction of Isc and Isc/R, clearly indicating an impairment of the renal filter function. Interestingly, the administration of EDTA before Isc and Isc/R, maintained both parameters at physiological levels (Fig. <xref ref-type="fig" rid="F1">1</xref>), thus suggesting a protective role of EDTA toward the renal filter capacity.</p><fig position="float" id="F1"><label>Figure 1</label><caption><p><bold>Effect of EDTA administration on renal function after Isc and Isc/R</bold>. Serum creatinine and blood urea nitrogen levels were measured. Rats that received intravenous injection of EDTA; 30 minutes before Isc or Isc/R induction; showed reduced levels of serum creatinine and blood urea nitrogen as compared with control rats (controls = C); lsc = ischemia; lsc/R = 60 minutes kidney ischemia followed by 60 min reperfusion. *p < 0.05.</p></caption><graphic xlink:href="1471-2369-7-5-1"/></fig></sec><sec><title>EDTA administration protected kidney from renal structural alterations</title><p>To assess whether EDTA, administered before renal Isc or Isc/R induction, protected kidney not only from functional damage but also from structural alterations, we performed histological evaluations, aimed to determine the presence of tubular epithelial cell necrosis, tubular dilation, protein casts and medullary congestion (Fig. <xref ref-type="fig" rid="F2">2</xref>). For this, kidneys from treated rats were excised and sections were stained with Hematoxylin/Eosin, to compare their architecture with that of control kidneys (Fig. <xref ref-type="fig" rid="F2">2a</xref>). Picture relative to kidneys from EDTA-treated rats (Fig. <xref ref-type="fig" rid="F2">2b</xref>) was similar to that of control kidneys (Fig. <xref ref-type="fig" rid="F2">2a</xref>): indeed, interstitial spaces were maintained and proximal tubule as well as cortical distal segments were preserved. Kidneys from sham-operated rats did not show evidence of important modifications with respect to the controls (data not shown). Kidneys from rats undergoing Isc (Fig. <xref ref-type="fig" rid="F2">2c</xref>) showed severe renal lesions, mainly tubular, such as dilation and focal engulfment by protein casts. Glomerular and interstitial hemorrhage were also present. Some tubular cells were necrotic, whereas other appeared vacuolized. This picture worsened when kidneys were obtained from rats submitted to R (60 min) after Isc, displaying (Fig. <xref ref-type="fig" rid="F2">2e</xref>) tubular cast increase and glomerular hypertrophy. Noteworthy, kidneys from animals pre-treated with EDTA before the induction of Isc, (Fig. <xref ref-type="fig" rid="F2">2d</xref>) failed to show important renal lesions. EDTA pretreatment preserved also the architecture of kidneys submitted to Isc/R (Fig. <xref ref-type="fig" rid="F2">2f</xref>). No significant differences were evident by comparing panel d and f of Fig. <xref ref-type="fig" rid="F2">2</xref>. Pictures related to the ascending thick limb in the kidney medulla displayed interstitial hemorrhage at the end of Isc in control kidneys. On the contrary, interstitial hemorrhage was absent in kidneys from EDTA-treated ischemized rats (data not shown). The semiquantitative analysis of renal damage, which represents the mean features for each group of animals, is summarized in Table <xref ref-type="table" rid="T2">2</xref>. The use of the eNOS inhibitor L-NAME, simultaneously injected with EDTA before the induction of Isc and Isc/R was able to block the beneficial effects induced by EDTA.</p><fig position="float" id="F2"><label>Figure 2</label><caption><p><bold>Renal morphology</bold>. Hematoxylin/Eosin images of differently treated rats. lsc = ischemia; lsc/R = ischemia/reperfusion. Representative cortical areas are shown. Notice the abundance of red blood cells and tubular protein casts in c and e panels in comparison with d and f (original magnification × 200).</p></caption><graphic xlink:href="1471-2369-7-5-2"/></fig><table-wrap position="float" id="T2"><label>Table 2</label><caption><p>Histologic evaluations of renal injury</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left"><bold>Rat treatment</bold></td><td align="center"><bold>Tubular necorsis</bold></td><td align="center"><bold>Tubular dilation</bold></td><td align="center"><bold>Protein casts</bold></td><td align="center"><bold>Medullary congestion</bold></td><td align="center"><bold>Glomerular damages</bold></td><td align="center"><bold>Interstitial stasis</bold></td></tr></thead><tbody><tr><td align="left"><bold>Controls</bold></td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">-</td></tr><tr><td align="left"><bold>Sham-operated</bold></td><td align="center">+ *</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">+ *</td></tr><tr><td align="left"><bold>EDTA</bold></td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">+</td></tr><tr><td align="left"><bold>Isc</bold></td><td align="center">-</td><td align="center">+</td><td align="center">+ **</td><td align="center">+/++</td><td align="center">-</td><td align="center">++</td></tr><tr><td align="left"><bold>EDTA+Isc</bold></td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">-</td><td align="center">+</td></tr><tr><td align="left"><bold>Isc/R</bold></td><td align="center">+</td><td align="center">++</td><td align="center">+++</td><td align="center">+</td><td align="center">-</td><td align="center">+</td></tr><tr><td align="left"><bold>EDTA+lsc/R</bold></td><td align="center">+ *</td><td align="center">++ *</td><td align="center">+ *</td><td align="center">-</td><td align="center">-</td><td align="center">+ *</td></tr><tr><td align="left"><bold>EDTA+L-NAME+Isc</bold></td><td align="center">-</td><td align="center">+</td><td align="center">+</td><td align="center">++</td><td align="center">-</td><td align="center">++</td></tr><tr><td align="left"><bold>EDTA+L-NAME+lsc/R</bold></td><td align="center">+</td><td align="center">++</td><td align="center">+++</td><td align="center">++</td><td align="center">-</td><td align="center">+</td></tr></tbody></table><table-wrap-foot><p>*Focal; **Big, but focal</p><p>- = absent; + = barely present; ++ = moderate; +++ = severe</p><p>Semiquantitative analysis of renal damage representative of mean features, obtained for each group of rats.</p></table-wrap-foot></table-wrap></sec><sec><title>Effect of EDTA on Mac-1 expression by PMN</title><p>To investigate a putative mechanism of action of EDTA, we considered its effect on PMN, which are largely involved in the damage associated with Isc/R [<xref ref-type="bibr" rid="B11">11</xref>,<xref ref-type="bibr" rid="B12">12</xref>].</p><p>For this purpose, PMN, isolated from peripheral rat blood, were analyzed for the expression of the pro-adhesive molecule Mac-1 (Fig. <xref ref-type="fig" rid="F3">3</xref>); the existence of Mac-1 up-regulation is suggestive of PMN activation [<xref ref-type="bibr" rid="B21">21</xref>]. Mac-1 expression by PMN obtained from control rats increased significantly after Isc and Isc/R. Following EDTA pretreatment, the increase was significantly impaired in rats submitted to Isc and, at lower extent, to Isc/R.</p><fig position="float" id="F3"><label>Figure 3</label><caption><p><bold>Expression of Mac-1 by PMN recovered from rat blood</bold>. The data represents the values; expressed as mean fluorescence intensity (MFI) (obtained by subtracting the respective value of negative control from each intensity value). lsc = ischemia; lsc/R = ischemia/reperfusion. *p < 0.05.</p></caption><graphic xlink:href="1471-2369-7-5-3"/></fig></sec><sec><title>EDTA administration strongly influenced NO production in vivo and renal eNOS expression</title><p>Being the expression of adhesion molecules, the adhesive and migratory pattern of leukocytes finely regulated by NO, both in physiologic and pathologic conditions [<xref ref-type="bibr" rid="B10">10</xref>,<xref ref-type="bibr" rid="B22">22</xref>-<xref ref-type="bibr" rid="B24">24</xref>], we then measured rat NO plasmatic levels. EDTA pre-treatment significantly increased the levels of circulating NO (Fig. <xref ref-type="fig" rid="F4">4</xref>) both in control and in ischemic rats. Conversely, post-ischemic reperfusion impaired dramatically the production of NO but was not insensitive to EDTA pre-administration: in fact NO production following Isc/R in EDTA-pre-treated rats was similar to that measured in control rats.</p><fig position="float" id="F4"><label>Figure 4</label><caption><p><bold>Plasmatic NO levels</bold>. They are expressed in μM. Rats that received intravenous injection of EDTA showed increased levels of NO as compared with controls (C). Sham = sham-operated rats. lsc = lschemia; lsc/R = ischemia/reperfusion. *p < 0.05 vs C; **p < 0.05 vs. Isc; ***p < 0.05 vs. Isc/R.</p></caption><graphic xlink:href="1471-2369-7-5-4"/></fig><p>As NO in vascular endothelial cells is synthesized primarily by the endothelial form of the constitutive NO-producing enzyme (eNOS), we examined the possibility that a decrease in NO bioavailability might be related to a change in the rate of expression of eNOS. The renal expression of eNOS (Fig. <xref ref-type="fig" rid="F5">5</xref>), observed in glomerular and interstitial capillaries, was slightly higher and diffuse in animals treated with EDTA (b), as compared to untreated control rats (a). Induction of short time (60 min) Isc, in control rats, produced a loss in the glomerular eNOS and an increase of its interstitial expression (c). When Isc followed EDTA treatment (d), eNOS expression was prevalently assessed inside glomerular capillaries. Kidney sections obtained following Isc/R in controls showed very low expression of eNOS both at glomerular and interstitial levels (e). Kidney sections from rats treated with EDTA before Isc/R (f) displayed fluorescence findings comparable to that of controls (a). The use of L-NAME together with EDTA before the induction of Isc and Isc/R abrogated the increase in eNOS expression due to EDTA treatment alone (data not shown).</p><fig position="float" id="F5"><label>Figure 5</label><caption><p><bold>Immunofluorescence microscopy of eNOS</bold>. Localization of eNOS (green) on differently treated rats (lsc = ischemia; lsc/R = ischemia/reperfusion); arrows pointed to positive glomeruli; and arrowheads to negative. Nuclei were counterstained with DAPI (original magnification × 200).</p></caption><graphic xlink:href="1471-2369-7-5-5"/></fig></sec><sec><title>EDTA regulated the vascular permeability in vivo</title><p>It has been recently demonstrated that eNOS has a critical role in regulating the microcirculatory endothelial barrier function <italic>in vivo </italic>[<xref ref-type="bibr" rid="B25">25</xref>]. We investigated whether EDTA influenced the TNFα-induced vascular leakage in kidneys. Vascular leakage values (expressed as μg dye/g fresh kidney and mean ± SEM of 8 rats) are reported in Table <xref ref-type="table" rid="T2">2</xref>. A significant increase in dye retention has been shown by kidneys of rats treated with TNFα with respect to kidneys of untreated animals (controls). EDTA treatment alone did not alter the endothelial barrier function. The concomitant administration of EDTA and TNFα resulted in the significant reduction of TNFα-induced leakage, indicating the existence of tights links among EDTA-NO-vascular protection.</p></sec></sec><sec><title>Discussion</title><p>EDTA, used in patients affected by chronic lead intoxication, improved renal function [<xref ref-type="bibr" rid="B3">3</xref>]. We investigated whether EDTA exerted its protective effect also toward kidneys affected by Isc or Isc/R. For this purpose, we administered intravenously EDTA 30 min before the induction of renal Isc, obtained by clamping the right renal artery and the right renal vein.</p><p>The severe renal injury induced by Isc or Isc/R was assessed both as functional impairment, through the serum creatinine and blood urea nitrogen dosages (Figure <xref ref-type="fig" rid="F1">1</xref>), and as structural alteration of tubules and glomeruli (Fig. <xref ref-type="fig" rid="F2">2</xref>). It should be noted that EDTA administration was efficient in significantly preserving renal function and in preventing structural alterations and necrotic lesions.</p><p>NO plays an important role in regulating vascular tone and improving renal blood flow [<xref ref-type="bibr" rid="B26">26</xref>]. We show that circulating levels of NO are increased after EDTA injection, followed or not by Isc or Isc/R (Fig. <xref ref-type="fig" rid="F4">4</xref>). NO administration could act by scavenging the ROS [<xref ref-type="bibr" rid="B6">6</xref>]. Indeed, the improvement of NO induced by EDTA treatment could be responsible for a reduced endothelial damage mediated by ROS. In the present study the increase of circulating NO well correlates with the expression of eNOS in kidneys from EDTA-treated rats, also when Isc or Isc/R occurred. Recent data indicates that the renal protective effects due to ischemic preconditioning are attributable to eNOS-mediated NO production [<xref ref-type="bibr" rid="B27">27</xref>]. In fact, it has been found that ischemic preconditioning (e.g. three cycles of 2 minutes Isc followed by 5 minutes reperfusion) was able to protect against the Isc/R-induced acute renal failure [<xref ref-type="bibr" rid="B27">27</xref>]. Congruously with the finding that pharmacological inhibition of NO synthesis- or disruption of the eNOS gene- significantly increases blood pressure [<xref ref-type="bibr" rid="B10">10</xref>,<xref ref-type="bibr" rid="B25">25</xref>], EDTA pretreatment has been demonstrated able to prevent the ischemic increase of MABP (Table <xref ref-type="table" rid="T1">1</xref>).</p><p>NO modulates leukocyte adhesion in the microcirculation by decreasing the binding of PMN to the adhesion molecules E-selectin and ICAM-1 [<xref ref-type="bibr" rid="B22">22</xref>,<xref ref-type="bibr" rid="B24">24</xref>]. PMN are involved in the tissue damage due to Isc/R injury: their activation and migration in ischemic tissues is followed by release of lytic enzymes and production of ROS [<xref ref-type="bibr" rid="B11">11</xref>,<xref ref-type="bibr" rid="B28">28</xref>]. We show that Mac-1 expression, widely considered a sensitive marker of PMN activation [<xref ref-type="bibr" rid="B21">21</xref>], is up-regulated in rats submitted to kidney Isc and Isc/R. Treatment with EDTA prevents PMN activation in both ischemized and undergoing postischemic reperfusion rats (Figure <xref ref-type="fig" rid="F3">3</xref>). The efficacy of EDTA treatment in protecting PMN from activation is possibly mediated by the increase in NO production (Fig. <xref ref-type="fig" rid="F2">2</xref>), given that NO inhibits the increase of adhesion molecule expression [<xref ref-type="bibr" rid="B22">22</xref>,<xref ref-type="bibr" rid="B24">24</xref>]. Moreover, it has been shown that during the acute myocardial Isc/R the low level of NO increased PMN adhesion to the endothelium [<xref ref-type="bibr" rid="B23">23</xref>].</p><p>It is known that NO derived from eNOS is a powerful vasodilator and possesses vasoprotective effects [<xref ref-type="bibr" rid="B29">29</xref>]. Here we show that EDTA is able to maintain the expression of eNOS on the glomerular and interstitial capillaries after Isc and Isc/R. Several divalent cations (Mn<sup>++</sup>, Zn<sup>++ </sup>and Fe<sup>++</sup>) suppressed eNOS activity in crude cell extracts and intact cells whereas Cu<sup>++ </sup>increased eNOS activation [<xref ref-type="bibr" rid="B30">30</xref>]. So, we could argue that the removal of some divalent cations by EDTA may improve eNOS levels. In this context, the in vivo use of a divalent cation, the Cd<sup>++</sup>, was responsible for decreased NO concentration in rat serum [<xref ref-type="bibr" rid="B31">31</xref>]. Some clinical evidences support our results. Recently, chelation therapy with EDTA (also associated with vitamin B) in subjects with coronary artery disease showed a significant NO-related endothelial function improvement [<xref ref-type="bibr" rid="B32">32</xref>]. Analogously, iron chelation with deferoxamine infusion in cardiomyopathy patients improved NO-mediated endothelium dependent vasodilation, suggesting that iron availability contributes to impair NO action in atherosclerosis [<xref ref-type="bibr" rid="B33">33</xref>]. Moreover, cardiovascular protection obtained by the use of high-dose corticosteroids has been shown to be mediated by non-transcriptional activation of eNOS [<xref ref-type="bibr" rid="B34">34</xref>]. The role of eNOS as a trigger and mediator of isoflurane-induced delayed preconditioning in vivo has been recently reported [<xref ref-type="bibr" rid="B35">35</xref>].</p><p>We propose that EDTA may act through an enhancement of endothelial NO production, as previously reported for corticosteroid [<xref ref-type="bibr" rid="B34">34</xref>] and also for desflurane, a preconditioning agent able to protect myocardium against Isc/R injury, by favouring NO release [<xref ref-type="bibr" rid="B36">36</xref>].</p><p>New data suggests for EDTA the favorable antioxidant mechanism of action previously described for other chelating agents [<xref ref-type="bibr" rid="B4">4</xref>,<xref ref-type="bibr" rid="B7">7</xref>]. In fact the use of EDTA complexes with metal ions as Fe<sup>++ </sup>and Cu<sup>++ </sup>suppressed superoxide and hydrogen peroxide activity [<xref ref-type="bibr" rid="B37">37</xref>]. In addition, recently, Hininger et al. [<xref ref-type="bibr" rid="B38">38</xref>] showed the beneficial antioxidant effects of EDTA chelating therapy. Since oxidative stress contributes to the pathogenesis of many diseases, including cardiovascular diseases, the protection exerted by EDTA against ischemic damage could be reconducible also to its antioxidant ability.</p></sec><sec><title>Conclusion</title><p>The data shows that functional and histological parameters of rat kidneys are preserved from damage due to Isc and Isc/R by EDTA treatment. These results suggest the existence of a tight loop EDTA/eNOS/NO, which on the one hand results in the loss of PMN activation and on the other hand in the maintenance of the endothelial barrier function.</p></sec><sec><title>List of abbreviations used</title><p>EDTA Sodium edetate</p><p>NO Nitric oxide</p><p>eNOS Endothelial NO synthase</p><p>Isc Ischemia</p><p>Isc/R Ischemia/reperfusion</p><p>PMN Polymorphonucelar cells</p><p>L-NAME N(omega)-nitro-L-arginine methyl ester</p><p>ROS Reactive oxygen species</p><p>MABP Mean arterial blood pressure</p><p>SEM Standard error of the mean</p><p>NO<sub>2</sub><sup>- </sup>Nitrate</p><p>NO<sub>3</sub><sup>- </sup>Nitrite</p><p>Mac-1 Monocyte chemoattractant protein-1</p><p>TNFα Tumor necrosis factor</p></sec><sec><title>Competing interests</title><p>The author(s) declare that they have no competing interests.</p></sec><sec><title>Authors' contributions</title><p>CF performed histological and immunohisochemical analyses. AF and PT performed animal studies and collected samples. FP measured Mac-1 and performed statistical analyses. CS measured NO<sub>2</sub><sup>-</sup>/NO<sub>3</sub><sup>-</sup>levels. DB performed spectrophotometrical analyses and measured MABP. EF coordinated the in vitro studies. MEF coordinated the in vivo studies and wrote and edited the manuscript.</p><table-wrap position="float" id="T3"><label>Table 3</label><caption><p>Dye solution retention by rat kidneys</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left">Treatment</td><td align="left"><bold>μg/g</bold></td></tr></thead><tbody><tr><td align="left">Controls</td><td align="left">163 ± 8.3</td></tr><tr><td align="left">+TNFα</td><td align="left">456 ± 41.8*</td></tr><tr><td align="left">+EDTA</td><td align="left">178 ± 7.4</td></tr><tr><td align="left">+TNFα+EDTA</td><td align="left">298 ± 14.5*§</td></tr></tbody></table><table-wrap-foot><p>The table reports the modification of vascular permeability; following in vivo treatment with EDTA and TNFα (see Methods section). The rat right kidney was in vivo perfused with trypan blue solution; washed with saline; removed; homogenized and centrifuged. The supernatants were run on a spectrophotometer at 540 nm wavelength. The data was then expressed as μg dye retained per weight (g) of fresh kidney. Maximum dye retention (dye perfusion without washing) yielded a value of 619 ± 24.7.</p><p>*p < 0.05 vs. controls; <sup>§</sup>p < 0.05 vs. TNFα</p></table-wrap-foot></table-wrap></sec><sec><title>Pre-publication history</title><p>The pre-publication history for this paper can be accessed here:</p><p><ext-link ext-link-type="uri" xlink:href="http://www.biomedcentral.com/1471-2369/7/5/prepub"/></p></sec> |
Posttransplantation malignancy in a patient presenting with weight loss and changed bowel habits: a case report | <sec><title>Backround</title><p>Advancements in immunosuppressive therapy have significantly improved patient and graft survival following renal transplantation. This is paralleled by an increasing occurrence of posttransplantation malignancy.</p></sec><sec><title>Case presentation</title><p>We report on a patient who presented with a history reminding of colon cancer seven years after receiving a kidney transplant. Initial diagnostic imaging seemed to confirm this diagnosis showing a constricting colonic lesion. To our surprise, colonoscopy findings were unremarkable. Review of the imaging studies revealed that the tumor-like picture was caused by the renal graft impressing the intestine. The following search for malignancy in other locations resulted in the diagnosis of glioblastoma multiforme of which the patient died several weeks later.</p></sec><sec><title>Conclusion</title><p>Follow-up of renal transplant patients must include screening tests directed at tumor detection. Imaging studies and other tests in this patient group should be interpreted by physicians who are familiar with transplant related peculiarities.</p></sec> | <contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Schmitt</surname><given-names>Roland</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>roland.schmitt@yale.edu</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Kettritz</surname><given-names>Ute</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>ukettritz@berlin.helios-kliniken.de</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Luft</surname><given-names>Friedrich C</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>luft@fvk-berlin.de</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Kettritz</surname><given-names>Ralph</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>kettritz@fvk-berlin.de</email></contrib> | BMC Nephrology | <sec><title>Backround</title><p>Kidney transplantation is the treatment of choice for appropriate patients with chronic renal failure. With advancements in immunosuppressive therapy patient and graft survival have significantly improved over the last decades. This achievement of long-term organ acceptance, however, coincides with an increased occurrence of posttransplantation malignancy [<xref ref-type="bibr" rid="B1">1</xref>]. We will present a renal transplant patient who was referred to us because of suspected colon cancer.</p></sec><sec><title>Case presentation</title><p>A 66 year old male patient who had received a renal transplant seven years earlier because of IgA nephropathy in his native kidneys was referred to us by his nephrologist. His graft had functioned well and his creatinine was stable at 130 μmol/l with an immunosuppressive regimen of cyclosporine, azathioprine, and prednisolone that he had received since his surgery. Over the past 6 months he reported a 6 kg weight loss and increasing nausea. He had also noted loose, darkish stools several times daily and significant night sweats but no fever. He denied having travelled outside of Germany in the past year. His nephrologists had noted small contracted native kidneys, an unremarkable graft and normal remaining abdominal structures in an ultrasound study. A gastroscopy, chest x-ray and urological evaluation were unremarkable. All tests conducted for infectious diseases were negative. The haemoglobin and hematocrit, liver enzymes, thyroid hormone levels and serum proteins were normal. His creatinine was unchanged and the C-reactive protein was 4 mg/dl. His nephrologists, who suspected an intestinal malignancy in this transplanted patient, referred him to our service for colonoscopy and further evaluation.</p><p>The physical examination, including neurological examination was unremarkable. The blood pressure was 120/80 mmHg, the heart rate was 70/min. Our laboratory studies confirmed the ambulatory results. The cyclosporine level was in the therapeutic range. Serologic testing showed no signs of acute infection with adenovirus, Epstein-Barr Virus, Chlamydia pneumoniae, or Borrelia. PCR for CMV-DNA was negative. A colonoscopy was performed but was only successful to the level of the proximal sigmoid colon. The examination was terminated because of pain and definite resistance the gastroenterologist could not pass. Other than scattered diverticuli, the mucosa was normal. We next scheduled a radiocontrast to visualize the colon. In this study a constricting lesion was encountered in the cecum which the radiologist identified as a typical sign of colon cancer (figure <xref ref-type="fig" rid="F1">1</xref>). A computerized tomography was performed which showed a concentric wall thickening of the ascending colon about 3 cm above the cecum, which according to the radiologist, was consistent with the diagnosis of colon cancer (figure <xref ref-type="fig" rid="F2">2</xref>, panel a). All other abdominal organs were unremarkable otherwise or consistent with the earlier ultrasound examination. Before surgery another attempt at colonoscopy was scheduled. To our surprise, the second colonoscopy, performed with additional analgesia and sedation, revealed, a completely normal examination including the cecum. To address the obvious discrepancy between this result and the prior radiological findings we consulted another radiologist. She noticed that the radiocontrast study was performed with the patient lying prone (figure <xref ref-type="fig" rid="F1">1</xref>). Moreover, she pointed out the proximity of the ascending colon to the renal graft as shown in the computerized tomography, performed in the supine position (figure <xref ref-type="fig" rid="F2">2</xref>, panel b). She suggested that with the patient in this position, the graft could squash the colon in such a fashion that the instilled radiocontrast agent would produce a constricting tumor-like image. With these new insights we were forced to start from scratch and to look for a different diagnosis explaining the patient's symptoms. During his hospital stay we had observed that the patient, although unremarkable in the general neurological examination, appeared slow and sometimes inappropriate. In a mini-mental examination we found a significant reduction in cognitive function. The neurologist also found no localizing signs, but additional psychomotor testing verified rapidly progressive early dementia. A magnetic resonance imaging study of the brain was performed, revealing an infiltrating, destructive, multilocular lesion in the frontal lobe, (figure <xref ref-type="fig" rid="F3">3</xref>). The process was consistent with a malignant tumor and less suggestive of an infectious process. The patient was transferred to the neurosurgical section where a brain biopsy secured the diagnosis of glioblastoma multiforme. Radiation therapy was started but over the course of the next few days the patient's general status deteriorated rapidly and he became progressively more somnolent. Radiation had to be stopped and 6 weeks after initial admission the patient died.</p><fig position="float" id="F1"><label>Figure 1</label><caption><p>Water-soluble contrast study performed after instillation. The patient is lying prone. The cecum appears filled with a space-occupying lesion that the radiologist interpreted as highly suspicious of tumor.</p></caption><graphic xlink:href="1471-2369-7-9-1"/></fig><fig position="float" id="F2"><label>Figure 2</label><caption><p>CT in the supine position. Panel A: The wall of the ascending colon was interpreted as 'thickened'. Panel B: CT at the level of the cecum showing the renal transplant kidney functioning well. The kidney is adjacent to the ascending colon and cecum.</p></caption><graphic xlink:href="1471-2369-7-9-2"/></fig><fig position="float" id="F3"><label>Figure 3</label><caption><p>Magnetic resonance imaging of the brain revealed a left frontal, invasive, multilocular, infiltrating process.</p></caption><graphic xlink:href="1471-2369-7-9-3"/></fig></sec><sec><title>Conclusion</title><p>A significantly increased incidence of malignancy is well documented in renal transplant recipients if compared to the general population [<xref ref-type="bibr" rid="B1">1</xref>]. With improvements in immunosuppressive regimens that prolong patient and graft survival, neoplastic disease has developed into a frequent long-term complication in kidney recipients. One study found that 40 percent of renal graft recipients had cancer after 20 years of immunosuppression [<xref ref-type="bibr" rid="B2">2</xref>]. Many types of posttransplant cancers display a more malignant course than the same type of cancer in non-transplant patients. This aggressive behaviour of posttransplant malignancy contributes to the associated mortality causing 26 percent of deaths in patients surviving transplantation for at least 10 years [<xref ref-type="bibr" rid="B3">3</xref>].</p><p>One of the contributing factors to posttransplant cancer is the immunosuppressive therapy [<xref ref-type="bibr" rid="B1">1</xref>]. The association of immunosuppressive therapy and cancer development is also known from non-transplant patients but the precise role of immunosuppressive drugs in carcinogenesis is not fully understood. There is experimental evidence that the inhibition of immunological responses, especially after giving anti-T cell antibodies, increases cancer susceptibility by interfering with the host's anti-cancer surveillance system [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B4">4</xref>]. Accordingly, the amount of immunosuppression would be directly correlated to the risk of cancer. While this applies clinically to renal graft recipients on high-dose versus low-dose cyclosporine [<xref ref-type="bibr" rid="B5">5</xref>] it has also been demonstrated that cyclosporine is still carcinogenic in mice without an immune system [<xref ref-type="bibr" rid="B6">6</xref>]. This is partly ascribed to a TGF-beta dependent mechanism as cancer progression could be inhibited by TGF-beta blockade [<xref ref-type="bibr" rid="B6">6</xref>]. Similar effects have been reported for tacrolimus which also promoted cancer growth in immunodeficient mice [<xref ref-type="bibr" rid="B7">7</xref>]. A direct carcinogenic potential has recently been highlighted for azathioprine which increases the risk of skin cancer through higher UVA radiation sensitivity and subsequent DNA mutations [<xref ref-type="bibr" rid="B8">8</xref>]. In contrast to these cancer promoting side effects of calcineurin inhibitors and azathioprine, inhibitors of target-of-rapamycin (TOR), such as sirolimus and everolimus, seem to inhibit primary and metastatic tumor growth [<xref ref-type="bibr" rid="B9">9</xref>,<xref ref-type="bibr" rid="B10">10</xref>]. TOR inhibitor treatment was associated with reduced TGF-beta levels and with attenuated VEGF signalling [<xref ref-type="bibr" rid="B9">9</xref>,<xref ref-type="bibr" rid="B10">10</xref>]. A recent multivariate analysis including more than 30.000 kidney recipients showed a three times higher incidence rate of posttransplant malignancy in patients on calcineurin inhibitors as compared to patients on sirolimus or everolimus [<xref ref-type="bibr" rid="B11">11</xref>]. Another study followed 15 renal transplant patients with Kaposi's sarcoma while they were switched from cyclosporine to sirolimus [<xref ref-type="bibr" rid="B12">12</xref>]. Three months after the switch all Kaposi's sarcoma lesions had disappeared while no graft rejection occurred and the serum creatinine stayed stable [<xref ref-type="bibr" rid="B12">12</xref>]. These and other studies indicate that TOR inhibitors have an equivalent immunosuppressive potency but a reduced risk of malignancy.</p><p>Guided by our patient's history we agreed with the referring nephrologist and suspected post-transplant malignancy. Considering the negative results of outpatient gastroscopy, x-ray and ultrasound together with reported changes in bowel habits, we wrongly assumed the colon as the most likely tumor location. This assumption seemed to be confirmed by the misinterpreted radiological findings. While the incidence of rectal cancer is decreased in transplant patients, the risk of colon cancer is significantly increased [<xref ref-type="bibr" rid="B13">13</xref>]. A general odds ratio hierarchy of tumor risk in kidney transplanted patients shows in decreasing order: Non-melanoma skin cancer, thyroid and other endocrine tumors, oral cavity cancers, cervix and vaginal cancers, non-Hodgkin lymphoma, renal, ureteral, and bladder cancer, colorectal cancer, lung cancer, and brain tumors [<xref ref-type="bibr" rid="B14">14</xref>]. Our patient's final diagnosis was glioblastoma multiforme, which together with other brain tumors has an odds ratio of 2.5 in transplant patients. Although we were not able to help our patient this case illustrates several aspects:</p><p>1. Transplant patients need a regular history, physical examination and screening tests directed at tumor detection (reviewed in reference 1).</p><p>2. The prevalence and the behaviour of specific types of cancers in transplant recipients are significantly different from the general population.</p><p>3. Post-transplant malignancy can only in part be explained by immunosuppression itself.</p><p>4. Imaging studies in transplant patients should be interpreted by physicians familiar with transplant related peculiarities.</p></sec><sec><title>Competing interests</title><p>The author(s) declare that they have no competing interests.</p></sec><sec><title>Pre-publication history</title><p>The pre-publication history for this paper can be accessed here:</p><p><ext-link ext-link-type="uri" xlink:href="http://www.biomedcentral.com/1471-2369/7/9/prepub"/></p></sec> |
E-selectin S128R polymorphism and severe coronary artery disease in Arabs | <sec><title>Background</title><p>The E-selectin p. S128R (g. A561C) polymorphism has been associated with the presence of angiographic coronary artery disease (CAD) in some populations, but no data is currently available on its association with CAD in Arabs.</p></sec><sec sec-type="methods"><title>Methods</title><p>In the present study, we determined the potential relevance of the E-selectin S128R polymorphism for severe CAD and its associated risk factors among Arabs. We genotyped Saudi Arabs for this polymorphism by PCR, followed by restriction enzyme digestion.</p></sec><sec><title>Results</title><p>The polymorphism was determined in 556 angiographically confirmed severe CAD patients and 237 control subjects with no CAD as established angiographically (CON). Frequencies of the S/S, S/R and R/R genotypes were found as 81.1%, 16.6% and 2.3% in CAD patients and 87.8%, 11.8%, and 0.4% in CON subjects, respectively. The frequency of the mutant 128R allele was higher among CAD patients compared to CON group (11% vs. 6%; odds ratio = 1.76; 95% CI 1.14 – 2.72; <italic>p </italic>= .007), thus indicating a significant association of the 128R allele with CAD among our population. However, the stepwise logistic regression for the 128R allele and different CAD risk factors showed no significant association.</p></sec><sec><title>Conclusion</title><p>Among the Saudi population, The E-selectin p. S128R (g. A561C) polymorphism was associated with angiographic CAD in Univariate analysis, but lost its association in multivariate analysis.</p></sec> | <contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Abu-Amero</surname><given-names>Khaled K</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>kamero@kfshrc.edu.sa</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Al-Boudari</surname><given-names>Olayan M</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>olayan@kfshrc.edu.sa</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Mohamed</surname><given-names>Gamal H</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>gmohamed@kfshrc.edu.sa</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Dzimiri</surname><given-names>Nduna</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>zimiri@kfshrc.edu.sa</email></contrib> | BMC Medical Genetics | <sec><title>Background</title><p>E-selectin (endothelial leukocyte adhesion molecule; ELAM1) is an 11-kD cell surface glycoprotein expressed on endothelial cells after activation by cytokines, and mediates adhesion of circulating monocytes and lymphocytes to endothelial cells. This adherence to activated arterial endothelium is one of the earliest detectable events in the pathogenesis of atherosclerosis [<xref ref-type="bibr" rid="B1">1</xref>]. Double-knockout mouse experiments suggested that E-selectin plays an essential role in both early and advanced stages of atherosclerotic lesion development and that mutations in cellular adhesion molecules like E-selectin may act as genetic risk factors for coronary atherosclerosis [<xref ref-type="bibr" rid="B2">2</xref>,<xref ref-type="bibr" rid="B3">3</xref>]. Additionally, the involvement of E-selectin in cardiovascular diseases is suggested by the fact that it is expressed only in activated endothelial cells. Amino acid change from serine (S) to arginine (R) at codon 128 (S128R), which corresponds to A > C nucleotide change at position 561 (A561C), in the epidermal growth factor-like domain of the E-selectin gene has been implicated in the pathogenesis of CAD in several ethnic groups, including Germans, Japanese, Americans, Chinese and Africans [<xref ref-type="bibr" rid="B4">4</xref>-<xref ref-type="bibr" rid="B10">10</xref>]. The 128R mutant allele was significantly higher in the CAD patients than in controls (12.6% versus 6.7%, 17.4% versus 7.1%, and 19.5% versus 10.6%) in Japanese, [<xref ref-type="bibr" rid="B4">4</xref>] German [<xref ref-type="bibr" rid="B11">11</xref>] and white American [<xref ref-type="bibr" rid="B6">6</xref>] populations, respectively. However, no previous studies are available on possible association of this polymorphism with CAD among Arabs. Furthermore, apart from a study, which did not find a link between this mutation and CAD in Austrian patients with diabetes mellitus [<xref ref-type="bibr" rid="B12">12</xref>], there is hardly any data in the literature pertaining to the possible association of this mutation for the different CAD risk factors. Therefore, the aim of this investigation was to evaluate the potential relevance of E-selectin 128R polymorphism for angiographic CAD and its risk factors in Arabs, using the Saudi population as a study model.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>Study population</title><p>Two groups of Saudi individuals were recruited for the present study. The patient group comprised 556 candidates (396 males and 160 females; mean age 50 ± 16 yr) of Saudi Arabian descent with angiographically documented severe CAD. The inclusion criterion for CAD was the presence of angiographically determined narrowing of the coronary vessels by at least 70%, which we define as having severe disease. Exclusion criteria for CAD were major cardiac rhythm disturbances, incapacitating or life-threatening illness, major psychiatric illness or substance abuse, history of cerebral vascular disease, neurological disorder, and administration of psychotropic medication. A second group of 237 individuals (105 males and 132 females, mean age 50 ± 17 yr) undergoing surgery for heart valvular diseases and those who reported with chest pain, but were established to have no significant coronary stenosis by angiography, were recruited as angiographed controls (CON). Exclusion criteria for this group included among others diseases such as cancer, autoimmune disease, or any other disorders likely to interact with variables under investigation. This study was performed in accordance with the regulations laid down by the Hospital Ethics Committee and all participants signed an informed consent.</p></sec><sec><title>DNA preparation</title><p>Five ml of peripheral blood were collected in EDTA tubes from all participating individuals after obtaining their written consent. DNA was extracted using the PURGENE kit from Gentra Systems (Minneapolis, MN, USA), and stored at -20°C in aliquots until required.</p></sec><sec><title>Determination of CAD risk factors</title><p>Serum cholesterol and triglyceride levels were measured as routine in the main Hospital Pathology Laboratory. Triglyceride levels >1.8 mmol/L and total cholesterol levels > 5.2 mmol/L were considered elevated. Diabetic patients either had a known history of diabetes mellitus or were diagnosed as such according to the American Diabetes Association criteria [<xref ref-type="bibr" rid="B13">13</xref>]. Diagnosis of myocardial infarction was based on the consensus specified by the European Society of Cardiology and the American College of Cardiology [<xref ref-type="bibr" rid="B14">14</xref>]. Body mass index (BMI) was determined for all participants, and individuals with BMI ≥ 30 were considered obese in accordance with the Center for Disease Control and Prevention (Atlanta, GA, USA). Information about all other risk factors was procured either through patient interviews or by referring to their medical records.</p></sec><sec><title>Detection of the S128R (A561C) polymorphism</title><p>This was carried out by polymerase chain reaction (PCR) amplification followed by <italic>Pst</italic>I restriction enzyme digestion. For DNA amplification, we used the forward primer 5'- AGT AAT AGT CCT CCT CAT CAT G -3' and reverse primer 5'- ACC ATC TCA AGT GAA GAA AGA G-3', designed to amplify a 186 bp fragment of the E-selectin gene [<xref ref-type="bibr" rid="B6">6</xref>,<xref ref-type="bibr" rid="B11">11</xref>]. Each 25 μl PCR reaction contained 2.5 μl of 10X reaction buffer with MgCl<sub>2 </sub>(Amersham Pharmacia Biotech, Piscataway, NJ, USA), 10 ρmol of each primer, 100 ρmol/μl each of deoxynucleoside triphophates (deoxyadenosine triphosphate, deoxyguanosine triphosphate, deoxycytidine triphosphate and deoxythymidine triphosphate) (Perkin-Elmer Corporation, Foster City, CA, USA) in Tris HCl buffer, 1 unit Taq DNA polymerase (Amersham Pharmacia Biotech, Piscataway, NJ, USA) and 50 ng genomic DNA template. The mixture was denatured at 95°C for 5 min and the PCR reaction was carried out for 35 cycles, in a GeneAmp 9600 PCR system (Perkin-Elmer Corporation, Foster City, CA, USA), under the following conditions: denaturation at 95°C for 1 min, annealing at 54°C for 45 sec, extension at 72°C for 1 min and final extension cycle of 72°C for 7 min. The PCR products were electrophoresed on a 1% agarose gel and detected with 0.5 μg/ml ethidium bromide to confirm the correct amplicon size. The products were digested using the <italic>Pst</italic>I restriction enzyme (Stratagene, LaJolla, Calif, USA) and the resultant fragments resolved on a 4% MetaPhor agarose gel (FMC Bio-products, Rockland, Maine, USA) in TE buffer containing 0.5 μg/ml ethidium bromide. The sizes of the digested amplicons were determined using the 50-bp ladder (Amersham Pharmacia Biotech, Piscataway, NJ, USA). As a quality control, we confirmed by direct sequencing the genotype status of 384 random samples representing the three different genotypes.</p></sec><sec><title>Statistical analysis</title><p>Genotype frequencies in various groups were compared by Chi-Square test. Multivariable logistic regression was used to study the effect of the E-selectin 561 C allele (128R allele) on CAD status, incorporating other variables (coronary risk factors) into the model. Additionally, we tried multiple logistic regression models involving the inclusion of interaction of the genotype × different CAD risk factors. All analyses were performed using SPSS v.10 (SPSS Inc., Chicago, USA) statistical analysis software. A two-tailed <italic>p </italic>value < .05 was considered statistically significant. We did power analysis employing the nQuery Adviser version.4 using two scenarios described in the results section and we concluded from this analysis that the study was adequately powered.</p></sec></sec><sec><title>Results</title><p>In the angiographically confirmed CAD patients (n = 556), 451 (81.1%) were homozygous S/S genotype, 92 (16.6%) were heterozygous S/R and 13 (2.3%) were homozygous for the R/R (Table <xref ref-type="table" rid="T1">1</xref>). Among the CON group (n = 237), 208 (87.8%) were S/S, 28 (11.8%) were S/R and 1 (0.4%) was R/R genotype. When we compared the prevalence of the different genotypes between the CAD and CON groups, the odds ratio for the S/R genotype was 1.52 (95% CI: 0.94 – 2.45; <italic>p </italic>= .07) and that for the R/R genotype was 6.0 (95% CI: 0.82 – 123.6; <italic>p </italic>= .05).</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>Crude odds ratio for all CAD risk factors among CAD and CON groups</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="center"><bold>Risk factor</bold></td><td align="center"><bold><italic>Status</italic></bold></td><td align="center"><bold>Total</bold></td><td align="center"><bold>CAD</bold></td><td align="center"><bold>CON</bold></td><td align="center"><bold>Odds ratio</bold></td><td align="center"><bold>95% C.I</bold></td><td align="center"><bold><italic>P </italic>value</bold></td></tr></thead><tbody><tr><td align="center">Genotype</td><td align="center">S/S</td><td align="center">659</td><td align="center">451</td><td align="center">208</td><td align="center">-</td><td align="center">-</td><td align="center">-</td></tr><tr><td></td><td align="center">S/R</td><td align="center">120</td><td align="center">92</td><td align="center">28</td><td align="center">1.52</td><td align="center">0.94 – 2.45</td><td align="center">.07</td></tr><tr><td></td><td align="center">R/R</td><td align="center">14</td><td align="center">13</td><td align="center">1</td><td align="center">6.0</td><td align="center">0.82 – 123.6</td><td align="center">.05</td></tr><tr><td align="center">Alleles</td><td align="center">S</td><td align="center">1438</td><td align="center">994</td><td align="center">444</td><td align="center">-</td><td align="center">-</td><td align="center">-</td></tr><tr><td></td><td align="center">R</td><td align="center">148</td><td align="center">118</td><td align="center">30</td><td align="center">1.76</td><td align="center">1.14 – 2.72</td><td align="center">.007</td></tr><tr><td align="center">Age</td><td align="center">< 40</td><td align="center">207</td><td align="center">139</td><td align="center">68</td><td align="center">-</td><td align="center">-</td><td align="center">-</td></tr><tr><td></td><td align="center">≥ 40</td><td align="center">586</td><td align="center">417</td><td align="center">169</td><td align="center">1.20</td><td align="center">0.86 – 1.69</td><td align="center">.29</td></tr><tr><td align="center">Hypercholesterolemia</td><td align="center">No</td><td align="center">252</td><td align="center">73</td><td align="center">179</td><td align="center">-</td><td align="center">-</td><td align="center">-</td></tr><tr><td></td><td align="center">Yes</td><td align="center">541</td><td align="center">483</td><td align="center">58</td><td align="center">20.4</td><td align="center">13.8 – 30.0</td><td align="center"><.001</td></tr><tr><td align="center">DM</td><td align="center">No</td><td align="center">382</td><td align="center">194</td><td align="center">188</td><td align="center">-</td><td align="center">-</td><td align="center">-</td></tr><tr><td></td><td align="center">Yes</td><td align="center">411</td><td align="center">362</td><td align="center">49</td><td align="center">7.16</td><td align="center">4.9 – 10.2</td><td align="center"><.001</td></tr><tr><td align="center">FH</td><td align="center">No</td><td align="center">544</td><td align="center">376</td><td align="center">168</td><td align="center">-</td><td align="center">-</td><td align="center">-</td></tr><tr><td></td><td align="center">Yes</td><td align="center">249</td><td align="center">180</td><td align="center">69</td><td align="center">1.1</td><td align="center">0.84 – 1.62</td><td align="center">.40</td></tr><tr><td align="center">Gender</td><td align="center">F</td><td align="center">292</td><td align="center">160</td><td align="center">132</td><td align="center">-</td><td align="center">-</td><td align="center">-</td></tr><tr><td></td><td align="center">M</td><td align="center">501</td><td align="center">396</td><td align="center">105</td><td align="center">3.11</td><td align="center">2.3 – 4.3</td><td align="center"><.001</td></tr><tr><td align="center">Hypertension</td><td align="center">No</td><td align="center">120</td><td align="center">49</td><td align="center">71</td><td align="center">-</td><td align="center">-</td><td align="center">-</td></tr><tr><td></td><td align="center">Yes</td><td align="center">673</td><td align="center">507</td><td align="center">166</td><td align="center">4.42</td><td align="center">2.95 – 6.62</td><td align="center"><.001</td></tr><tr><td align="center">MI</td><td align="center">No</td><td align="center">599</td><td align="center">368</td><td align="center">231</td><td align="center">-</td><td align="center">-</td><td align="center">-</td></tr><tr><td></td><td align="center">Yes</td><td align="center">194</td><td align="center">188</td><td align="center">6</td><td align="center">19.7</td><td align="center">8.5 – 45.1</td><td align="center"><.001</td></tr><tr><td align="center">Obesity</td><td align="center">No</td><td align="center">73</td><td align="center">45</td><td align="center">28</td><td align="center">-</td><td align="center">-</td><td align="center">-</td></tr><tr><td></td><td align="center">Yes</td><td align="center">720</td><td align="center">511</td><td align="center">209</td><td align="center">0.65</td><td align="center">0.39 – 1.08</td><td align="center">.07</td></tr><tr><td align="center">Smoking</td><td align="center">No</td><td align="center">667</td><td align="center">462</td><td align="center">205</td><td align="center">-</td><td align="center">-</td><td align="center">-</td></tr><tr><td></td><td align="center">Yes</td><td align="center">126</td><td align="center">94</td><td align="center">32</td><td align="center">1.30</td><td align="center">0.84 – 2.01</td><td align="center">.25</td></tr><tr><td align="center">Hypertriglyceridemia</td><td align="center">No</td><td align="center">247</td><td align="center">90</td><td align="center">157</td><td align="center">-</td><td align="center">-</td><td align="center">-</td></tr><tr><td></td><td align="center">Yes</td><td align="center">546</td><td align="center">466</td><td align="center">80</td><td align="center">10.2</td><td align="center">7.1 – 14.4</td><td align="center"><.001</td></tr></tbody></table><table-wrap-foot><p>DM, diabetes Mellitus; FH, family history of CAD; MI, myocardial infarction on admission; hypertriglyceridemia (>1.8 mmol/L); hypercholesterolemia (> 5.2 mmol/L) and obesity (Body Mass Index ≥ 30).</p></table-wrap-foot></table-wrap><p>We found that the mutant 128R allele accounted for 11% in the CAD group, which was significantly higher than that in the CON group (6%). The odds ratio for the risk of CAD associated with the 128R allele was 1.76 (95% CI 1.14 – 2.72; <italic>p </italic>= .007), thus indicating a significant association of this allele with CAD in our population.</p><p>The variables showing an association (<italic>p </italic>=< .05) from Table <xref ref-type="table" rid="T1">1</xref>, were then put into a stepwise logistic regression, in order to study the possible combined effect of the mutant 128R allele with other risk factors on angiographic CAD. The variables retained in the model were hypertension (<italic>p </italic>= .03), diabetes mellitus (DM) (<italic>p </italic>=< .001), hypercholesterolemia (<italic>p </italic>=< .001), hypertriglyceridemia (<italic>p </italic>= .008), MI (<italic>p </italic>=< .001) and gender (<italic>p </italic>=< .001) (Table <xref ref-type="table" rid="T2">2</xref>).</p><table-wrap position="float" id="T2"><label>Table 2</label><caption><p>Results of multiple logistic regression analysis: final significant variables in equation</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="center"><bold>Risk Factor (variables)</bold></td><td align="center"><bold>Odds ratio</bold></td><td align="center"><bold>[95% C.I]</bold></td><td align="center"><bold><italic>p </italic>value</bold></td></tr></thead><tbody><tr><td align="center">Hypertension</td><td align="center">1.53</td><td align="center">1.03 – 2.26</td><td align="center">.03</td></tr><tr><td align="center">DM</td><td align="center">3.95</td><td align="center">2.85 – 5.48</td><td align="center"><.001</td></tr><tr><td align="center">Hypercholesterolemia</td><td align="center">7.96</td><td align="center">5.30 – 11.9</td><td align="center"><.001</td></tr><tr><td align="center">Hypertriglyceridemia</td><td align="center">1.72</td><td align="center">1.15 – 2.58</td><td align="center">.008</td></tr><tr><td align="center">MI</td><td align="center">14.1</td><td align="center">7.45 – 26.7</td><td align="center"><.001</td></tr><tr><td align="center">Gender</td><td align="center">2.36</td><td align="center">1.73 – 3.23</td><td align="center"><.001</td></tr></tbody></table></table-wrap><p>As for the power calculation, we did power analysis employing nQuery Adviser version 4 using two scenarios as follow: First, In our study, when we compared the risk of CAD among (S/R and R/R) to S/S, we got an odds ratio of 1.67 (p = .023). When α =.05, using 2 sided test, proportion of S/R and R/R among CAD is .189 and among controls is .122. Number of controls = 237, number of cases = 556 (we used the average, which is 397). Using this information we have calculated the power in nQuery as 74%. Second, In our study, when we analyzed the risk of CAD among (R alleles) compared to S alleles, we got an odds ratio of 1.7 (p = .007). When α = .05, using 2 sided test, proportion of C alleles among CAD is .106 and among controls is .063. Number of controls = 474, number of cases = 1112 (we used the average, which is 793). Using this information, we have calculated the power in nQuery as 86%. From this power calculation, we concluded that our study is adequately powered.</p></sec><sec><title>Discussion</title><p>In the past 3 decades, the incidence of CAD is on the rise in Saudi Arabia. According to the biggest study conducted so far, the overall prevalence of this disease in Saudi Arabia is 5.5% [<xref ref-type="bibr" rid="B15">15</xref>]. The rise in the incidence of CAD has been attributed to the major changes in the life-style of the Saudi population. High-fat diets, obesity, diabetes, and smoking, all of which are considered CAD risk factors, have become more prevalent, and people are leading a more sedentary lifestyle. Two independent studies reported predicted an incremental scale for the development of CAD in the Saudi population because of sharp increases in CAD risk factors such as obesity, hypercholesterolemia, diabetes, hypertriglyceridemia, and high blood pressure [<xref ref-type="bibr" rid="B16">16</xref>,<xref ref-type="bibr" rid="B17">17</xref>]. Early detection of individuals genetically susceptible to CAD can lead to early intervention, and knowledge of genetic susceptibility to CAD has value in providing risk information and guiding decision-making. To our knowledge, this study is the first to evaluate the prevalence of E-selectin polymorphism (S128R) and its potential relevance for angiographic CAD and its associated risk factors in the Arab population. The first step was to determine the prevalence of different E-selectin genotypes among our general population of Saudi Arabs. We found that the S/S is the most abundant and R/R the least common genotype among Arabs. The prevalence of the 128R allele was 6% in our controls, which is almost similar to the rate observed in the Germans (7.1%) and Japanese (6.7%) and slightly higher that in Africans (3.7%) [<xref ref-type="bibr" rid="B9">9</xref>] and Chinese (0%) [<xref ref-type="bibr" rid="B10">10</xref>]. When we compared the genotype frequencies in CAD versus CON group, the odds ratio was significant, indicating an association between these genotypes and CAD.</p><p>However, possibly because of the relatively small number of individuals with the S/R and R/R genotype, this analysis did not attain significance see Table <xref ref-type="table" rid="T1">1</xref>. On the other hand, a significant odds ratio and <italic>p </italic>value was obtained when we compared the frequency of the 128R mutant allele in CAD and CON groups, pointing to an association of this allele with CAD among our Saudi Arab population. These results are comparable to the findings in the Japanese [<xref ref-type="bibr" rid="B4">4</xref>], German [<xref ref-type="bibr" rid="B11">11</xref>] white American [<xref ref-type="bibr" rid="B6">6</xref>] and Chinese [<xref ref-type="bibr" rid="B10">10</xref>] populations, but in contrast to a previous study in CAD patients with type 2 DM [<xref ref-type="bibr" rid="B12">12</xref>]. Although the frequencies of several classical risk factors for CAD, including elevated cholesterol and triglycerides levels, DM, age, hypertension, gender and MI were higher in the CAD patients compared to controls, when we entered these risk factors with the mutant 128R allele into a multiple variable logistic regression, the association was no longer significant.</p></sec><sec><title>Conclusion</title><p>In summary, the mutant 128R allele of the E-selectin gene is associated with angiographic severe CAD in Saudi Arabs. This association is lost after adjustment for traditional CAD risk factors.</p></sec><sec><title>Abbreviations</title><p><bold>BMI </bold>Body Mass Index</p><p><bold>CAD </bold>Coronary artery disease</p><p><bold>CI </bold>Confidence interval</p><p><bold>CON </bold>Control group</p><p><bold>DM </bold>Diabetes mellitus</p><p><bold>FH </bold>Family history of CAD</p><p><bold>MI </bold>Myocardial infarction</p><p><bold>PCR </bold>Polymerase chain reaction</p><p><bold>R </bold>Arginine</p><p><bold>S </bold>Serine</p><p><bold>TG </bold>Triglycerdies</p></sec><sec><title>Competing interests</title><p>The author(s) declare that they have no competing interests.</p></sec><sec><title>Authors' contributions</title><p>KKA was in charge of design and analysis of data, OMA performed the technical aspect of the study, PCR and genotyping, GHM performed the statistical analysis and ND was responsible for recruiting patients and overall supervision of the study.</p></sec><sec><title>Pre-publication history</title><p>The pre-publication history for this paper can be accessed here:</p><p><ext-link ext-link-type="uri" xlink:href="http://www.biomedcentral.com/1471-2350/7/52/prepub"/></p></sec> |
Dioxin-like activities in serum across European and Inuit populations | <sec><title>Background</title><p>Persistent organic pollutants (POPs) such as polychlorinated dibenzo-<italic>p</italic>-dioxins/furans, polychlorinated biphenyls (PCBs) and organochlorine pesticides can cause a series of adverse effects on e.g. reproduction in animals and humans, many of which involve the aryl hydrocarbon receptor (AhR). The aim of the present study was to compare the integrated serum level of AhR mediated activity among European and Inuit populations, and evaluate whether the activity was associated to the selected POP markers, 2,2',4,4',5,5'-hexachlorobiphenyl (CB-153) and 1,1-dichloro-2,2-bis(p-chlorophenyl)-ethylene (<italic>p,p'</italic>-DDE).</p></sec><sec sec-type="methods"><title>Methods</title><p>The study included 338 males from Greenland (Inuit's), Sweden, Warsaw (Poland) and Kharkiv (Ukraine). The AhR transactivity of serum extracts alone (AhRag) and competitive AhR activity (AhRcomp) upon co-exposure with 2,3,7,8-tetrachlorodibenzo-<italic>p</italic>-dioxin (TCDD) were determined in the lipophilic serum fraction containing the POPs using the AhR mediated luciferase reporter Hepa1.12cR cell assay.</p></sec><sec><title>Results</title><p>The European groups showed higher median level of AhR-TEQ (TCDD toxic equivalents) compared to the Inuit's, whereas higher incidence of Inuits sample further induced AhRcomp activity. Neither AhRag nor AhR-TEQ were correlated to CB-153 or <italic>p,p'</italic>-DDE for any of the study groups. Multiple regressions showed a significant heterogeneity of association between the CB-153 and the AhRcomp across the study groups, and accordingly a negative association between AhRcomp and CB-153 was found for the Kharkiv group.</p></sec><sec><title>Conclusion</title><p>No consistent correlation between AhR activities and two POP markers was found. Although the difference of AhRag between European and Inuit men could not be explained by CB-153 or <italic>p,p'</italic>-DDE levels alone, we believe that the variation of AhR serum activity reflects different pattern of POP exposure, genetics and/or life style factors.</p></sec> | <contrib id="A1" contrib-type="author"><name><surname>Long</surname><given-names>Manhai</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>ml@mil.au.dk</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Andersen</surname><given-names>Birgitte S</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>bsa@mil.au.dk</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Lindh</surname><given-names>Christian H</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>christian.lindh@med.lu.se</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Hagmar</surname><given-names>Lars</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>lars.hagmar@med.lu.se</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Giwercman</surname><given-names>Aleksander</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>aleksander.giwercman@med.lu.se</email></contrib><contrib id="A6" contrib-type="author"><name><surname>Manicardi</surname><given-names>Gian-Carlo</given-names></name><xref ref-type="aff" rid="I4">4</xref><email>manicardi.giancarlo@unimore.it</email></contrib><contrib id="A7" contrib-type="author"><name><surname>Bizzaro</surname><given-names>Davide</given-names></name><xref ref-type="aff" rid="I5">5</xref><email>d.bizzaro@univpm.it</email></contrib><contrib id="A8" contrib-type="author"><name><surname>Spanò</surname><given-names>Marcello</given-names></name><xref ref-type="aff" rid="I6">6</xref><email>spanomrc@casaccia.enea.it</email></contrib><contrib id="A9" contrib-type="author"><name><surname>Toft</surname><given-names>Gunnar</given-names></name><xref ref-type="aff" rid="I7">7</xref><email>gutof@as.aaa.dk</email></contrib><contrib id="A10" contrib-type="author"><name><surname>Pedersen</surname><given-names>Henning S</given-names></name><xref ref-type="aff" rid="I8">8</xref><email>hsp@gh.gl</email></contrib><contrib id="A11" contrib-type="author"><name><surname>Zvyezday</surname><given-names>Valentyna</given-names></name><xref ref-type="aff" rid="I9">9</xref><email>dimusic@ic.kharkov.ua</email></contrib><contrib id="A12" contrib-type="author"><name><surname>Bonde</surname><given-names>Jens Peter</given-names></name><xref ref-type="aff" rid="I7">7</xref><email>jpbon@as.aaa.dk</email></contrib><contrib id="A13" corresp="yes" contrib-type="author"><name><surname>Bonefeld-Jorgensen</surname><given-names>Eva C</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>ebj@mil.au.dk</email></contrib> | Environmental Health | <sec><title>1. Background</title><p>The polychlorinated dibenzo-<italic>p</italic>-dioxins/furans (PCDDs/PCDFs), polychlorinated biphenyls (PCBs) and organochlorine pesticides, such as 2,2-bis(p-chlorophenyl)-1,1,1-trichloroethane (DDT), are prominent among the persistent organic pollutants (POPs). Owing to their negative effects on wildlife and human health, PCBs and DDT were restricted or totally banned in most countries during the 1970s. However, PCBs can still be released into the environment from poorly maintained hazardous waste site and illegal or improper dumping of PCB wastes like leaking from old electrical transformers, and DDT is still used in some developing countries [<xref ref-type="bibr" rid="B1">1</xref>]. Being resistant to both biotic and abiotic degradation, DDT (mainly as its major metabolite, 1,1-dichloro-2.2-bis (p-chlorophenyl)-ethylene (<italic>p,p'</italic>-DDE)) and PCBs bioaccumulate and magnify in animals and humans [<xref ref-type="bibr" rid="B2">2</xref>,<xref ref-type="bibr" rid="B3">3</xref>]. Residues have been detected in various food substances and in human adipose tissue, milk, and serum [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B4">4</xref>]. While PCBs and DDT contamination is ubiquitous globally, a high intake of fish and sea mammal food in the Arctic regions is associated with extraordinary high exposure [<xref ref-type="bibr" rid="B5">5</xref>,<xref ref-type="bibr" rid="B6">6</xref>]. Also Swedish fisherman's families of the Baltic Sea with a high consumption of herring and salmon being contaminated with POPs, constitute a highly exposed group [<xref ref-type="bibr" rid="B7">7</xref>,<xref ref-type="bibr" rid="B8">8</xref>]. For the general populations in Eastern Europe, the burden of POPs has been less systematically examined [<xref ref-type="bibr" rid="B9">9</xref>,<xref ref-type="bibr" rid="B10">10</xref>].</p><p>It has been documented that exposure to POPs such as dioxins (e.g. 2,3,7,8-tetrachlorodibenzo-<italic>p</italic>-dioxin, TCDD) and dioxin-like compounds (DLCs) such as <italic>non-ortho </italic>and <italic>mono-ortho </italic>PCBs may cause a series of negative effects both in animal experiments and in human epidemiologic studies including carcinogenicity [<xref ref-type="bibr" rid="B11">11</xref>], immunotoxicity and adverse effects on reproductive, neurobehavioral [<xref ref-type="bibr" rid="B12">12</xref>]. The toxicity of dioxins and DLCs is mediated mainly through binding to the aryl hydrocarbon receptor (AhR), which is an intracellular ligand-dependent transcriptional factor expressed in most tissues of mammals [<xref ref-type="bibr" rid="B13">13</xref>]. Upon receptor-ligand binding and translocation to the nucleus, the complex with the AhR nuclear translocator binds to the DNA dioxin-responsive elements, causing induction of gene transcription, for instance, encoding for metabolic enzymes [<xref ref-type="bibr" rid="B14">14</xref>]. More recently, interference of POPs or their metabolites with hormone receptors has also been observed [<xref ref-type="bibr" rid="B15">15</xref>,<xref ref-type="bibr" rid="B16">16</xref>]. Previous studies demonstrated the presence of a two-way cross talk between the intracellular signalling pathways involving the estrogen- (ER), androgen- (AR) and the Ah- receptor [<xref ref-type="bibr" rid="B17">17</xref>]. Several studies on wildlife and laboratory animals showed that exposure to PCBs and <italic>p,p'</italic>-DDE can affect reproductive and endocrine functions [<xref ref-type="bibr" rid="B5">5</xref>,<xref ref-type="bibr" rid="B18">18</xref>]. However, human epidemiologic data are limited and major gaps in knowledge continue to preclude evidence based risk assessment.</p><p>Since dioxins and DLCs exist as complex mixtures of various congeners throughout the environment, the concept of TEQ (TCDD toxic equivalent) has been introduced to simplify risk assessment and regulatory control [<xref ref-type="bibr" rid="B3">3</xref>]. The classical TEQs are calculated by multiplying the concentration of individual PCDDs/PCDFs/PCBs by their respective Toxic Equivalency Factors (TEFs), which correspond to the relative potency of the congener to generate AhR-mediated effects in relation to TCDD, the most potent AhR ligand. Previous studies emphasize that assessment of the toxicological potential of a chemical mixture is much more complex than can be deduced by a given calculated TEQ value [<xref ref-type="bibr" rid="B5">5</xref>,<xref ref-type="bibr" rid="B19">19</xref>]. There are several drawbacks using the TEF concept for risk assessment of mixtures of POPs such as expensive and time consuming gas chromatography mass spectrometry (GC-MS) determinations, small concentrations of individual congeners, presence of compounds not routinely measured or unknown substances with AhR affinity, the lack of TEF values for several POPs, and possible antagonistic or synergistic interactions between POPs [<xref ref-type="bibr" rid="B20">20</xref>-<xref ref-type="bibr" rid="B22">22</xref>]. Thus there is a need for an integrated risk assessment of dioxins and DLCs. The <italic>in vitro </italic>AhR mediated <underline>c</underline>hemical <underline>a</underline>ctivated <underline>lu</underline>ciferase gene e<underline>x</underline>pression (CALUX) bioassay has proven to be a quick and sensitive assay to detect the AhR mediated potential of pure chemicals [<xref ref-type="bibr" rid="B20">20</xref>-<xref ref-type="bibr" rid="B23">23</xref>], extracts of environmental and biological matrices and thus the integrated TEQ value (CALUX-TEQ) of complex mixtures as found in sediment, pore water, bovine and human milk, human serum and follicular fluid [<xref ref-type="bibr" rid="B24">24</xref>].</p><p>This study was a part of the EU supported research project Inuedo [<xref ref-type="bibr" rid="B25">25</xref>] with the main objective to elucidate the fertility in European and Inuit groups with different intake of POPs [<xref ref-type="bibr" rid="B26">26</xref>,<xref ref-type="bibr" rid="B27">27</xref>]. The 2,2',4,4',5,5'-hexachlorobiphenyl (CB-153) and <italic>p,p'</italic>-DDE were selected as proxy biomarkers of POPs exposure because CB-153 generally correlates with serum total PCB concentration and chemical derived TEQ [<xref ref-type="bibr" rid="B28">28</xref>,<xref ref-type="bibr" rid="B29">29</xref>], and <italic>p,p'</italic>-DDE was considered as a relevant marker of POPs [<xref ref-type="bibr" rid="B30">30</xref>]. The specific aim of the present study was to compare the actual level of AhR mediated dioxin-like activity in the lipophilic serum fraction between European and Inuit study groups, and to evaluate whether the tested dioxin-like activity was correlated to CB-153 or <italic>p,p'</italic>-DDE.</p></sec><sec><title>2. Methods</title><sec><title>2. 1. Study groups and sampling</title><p>The Inuedo source populations encompassed women and their male spouses who had antenatal care visits from May 2002 through February 2004 at the local hospitals in Greenland, Warsaw, Poland and Kharkiv, Ukraine [<xref ref-type="bibr" rid="B26">26</xref>]. An established cohort of Swedish fishermen was also included [<xref ref-type="bibr" rid="B31">31</xref>]. The study was approved by the local ethical committees representing all participating populations and all subjects signed an informed consent. The subjects of the present study were adult males randomly selected from the source populations. Demographic and lifestyle factors such as age, body mass index (BMI), alcohol consumption, intake of seafood, coffee and smoking habits were collected by questionnaires (Table <xref ref-type="table" rid="T1">1</xref>) [<xref ref-type="bibr" rid="B26">26</xref>,<xref ref-type="bibr" rid="B27">27</xref>].</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>Characteristics of the men in the study groups</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td></td><td align="center"><bold><italic>Greenland n </italic>= 75</bold></td><td align="center"><bold><italic>Warsaw n </italic>= 99</bold></td><td align="center"><bold><italic>Sweden n </italic>= 78</bold></td><td align="center"><bold><italic>Kharkiv n </italic>= 86</bold></td><td align="center"><bold><italic>All n </italic>= 338</bold></td></tr></thead><tbody><tr><td align="left"><bold>Age </bold>(years)</td><td align="left">median</td><td align="center">30</td><td align="center">30</td><td align="center">46</td><td align="center">26</td><td align="center">32</td></tr><tr><td></td><td align="left">min-max</td><td align="center">23–47</td><td align="center">18–46</td><td align="center">24–67</td><td align="center">16–45</td><td align="center">16–67</td></tr><tr><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left"><bold>BMI </bold>(Kg/m<sup>2</sup>)</td><td align="left">median</td><td align="center">26</td><td align="center">26</td><td align="center">26</td><td align="center">24</td><td align="center">25</td></tr><tr><td></td><td align="left">min-max</td><td align="center">19–38</td><td align="center">12–58</td><td align="center">22–37</td><td align="center">19–36</td><td align="center">12–58</td></tr><tr><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left"><bold>Alcohol </bold>(drink/week)</td><td align="left">median</td><td align="center">2.0</td><td align="center">3.5</td><td align="center">n.a</td><td align="center">2.5</td><td align="center">3.00</td></tr><tr><td></td><td align="left">min-max</td><td align="center">0–35</td><td align="center">0–30</td><td></td><td align="center">0.2–15</td><td align="center">0–35</td></tr><tr><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left"><bold>Smoking </bold>(ever)</td><td align="left">%</td><td align="center">87</td><td align="center">49</td><td align="center">40</td><td align="center">82</td><td align="center">68</td></tr><tr><td align="left"><bold>Seafood </bold>(days/week)</td><td align="left">median</td><td align="center">2.0</td><td align="center">1.0</td><td align="center">n.a</td><td align="center">4.0</td><td align="center">2.0</td></tr><tr><td></td><td align="left">min-max</td><td align="center">0–9.0</td><td align="center">0–9.0</td><td></td><td align="center">1.0–9.0</td><td align="center">0–9.0</td></tr><tr><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left"><bold>Coffee </bold>(cups/day)</td><td align="left">median</td><td align="center">3.0</td><td align="center">2.0</td><td align="center">n.a</td><td align="center">2.0</td><td align="center">2.0</td></tr><tr><td></td><td align="left">min-max</td><td align="center">0–20</td><td align="center">0–6.0</td><td></td><td align="center">1.0–7.0</td><td align="center">0–20</td></tr><tr><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left"><bold>Total testosterone </bold>(nmol/l)</td><td align="left">median</td><td align="center">14</td><td align="center">13</td><td align="center">12</td><td align="center">18</td><td align="center">14</td></tr><tr><td></td><td align="left">min-max</td><td align="center">3.2–75</td><td align="center">6.5–23</td><td align="center">4.2–28</td><td align="center">8.4–32</td><td align="center">3.2–32</td></tr><tr><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left"><bold>Estradiol </bold>(nmol/l)</td><td align="left">median</td><td align="center">59</td><td align="center">72</td><td align="center">67</td><td align="center">81</td><td align="center">71</td></tr><tr><td></td><td align="left">min-max</td><td align="center">31–88</td><td align="center">45–296</td><td align="center">25–155</td><td align="center">33–144</td><td align="center">25–296</td></tr></tbody></table></table-wrap><p>This study included in total 338 blood samples taken from males from Greenland (75), Sweden (78), Warsaw (99) and Kharkiv (86). Venous blood samples were collected in 10 ml vacuum tubes and after centrifugation the serum was transferred to Nunc tubes and stored at -80°C until analyzed [<xref ref-type="bibr" rid="B27">27</xref>].</p></sec><sec><title>2. 2. Determination of CB-153 and p,p'-DDE in serum</title><p>Serum concentrations of CB-153 and <italic>p,p'</italic>-DDE were determined using GC-MS after solid phase extraction and on-column degradation of lipids. CB-153 and <italic>p,p'-</italic>DDE levels were adjusted for serum lipids analyzed by enzymatic methods [<xref ref-type="bibr" rid="B27">27</xref>,<xref ref-type="bibr" rid="B31">31</xref>]. Levels of detection, coefficients of variation (CV) and participation in quality control programs have been described in detail elsewhere [<xref ref-type="bibr" rid="B27">27</xref>]. All analysis of CB-153 and <italic>p,p'</italic>-DDE were performed at the Department of Occupational and Environmental Medicine, University of Lund, Sweden.</p></sec><sec><title>2. 3. AhR-CALUX assay</title><sec><title>2. 3.1. Sample preparation</title><p>The extraction of serum sample to obtain the fraction containing lipophilic POPs for AhR-CALUX activity measurements was performed at a certified laboratory, Le Centre de Toxicologie, Sainte Foy, Quebec, Canada. Serum samples (2 ml) were mixed with an aqueous solution of ammonium sulfate and ethanol (1:1) and then extracted with hexane. Extracts were concentrated and cleaned by elution through two columns containing Florisil. The details of extraction has been described elsewhere [<xref ref-type="bibr" rid="B32">32</xref>]. The extracts dissolved in 500 μl hexane were stored at -80°C until analyzed.</p></sec><sec><title>2.3.2. Dissolving of samples</title><p>The serum extracts were thawed and evaporated to near dryness at 30°C under the gentle stream of nitrogen. The sample solvent, 10 μl DMSO: H<sub>2</sub>O (5:5, v/v), was added to each sample vial and stored overnight at room temperature. After giving the samples a quick spin (1000 rpm, 25 sec.), 500 μl α-minimal essential medium (α-MEM, GibcoBRL, UK) was added, mixed completely and transferred to two new test tubes (250 μl/tube) each containing 417 μl supplemented α-MEM (α-MEM plus 10% fetal calf serum (FCS, GibcoBRL, UK), 64 μg/ml garamycin (Schering-Plough, Brussels, Belgium)) with or without the 60 pM (EC<sub>50</sub>) (see 2.3.3) TCDD (98%, Cambridge Isotopes Laboratories Inc., USA), respectively. The final serum extract was equal to 150 μl serum per well (96- well plate) which was shown to be in the linear range of the AhR mediated luciferease activity (Bonefeld Jorgensen and Long, manuscript in prep.). All the processes were protected from light.</p></sec><sec><title>2.3.3. AhR-CALUX bioassay</title><p>The stable transfected mouse hepatoma cell line Hepa1.12cR carrying the AhR-luciferase reporter gene (kindly provided by M.S. Denison (University of California, USA)), induces luciferase in an AhR-, time- and dose-dependent manner [<xref ref-type="bibr" rid="B23">23</xref>]. In each independent experiment, a dose-response of TCDD (dissolved in DMSO and subsequently diluted in the supplemented α-MEM) was performed at concentrations ranging from 2 × 10<sup>-12 </sup>to 5 × 10<sup>-9 </sup>M (see <xref ref-type="supplementary-material" rid="S1">additional file 1</xref>: TCDD dose-response curve for AhR mediated luciferase activity). The maximum effect concentration was 1000 pM, and the half maximum effect concentration (EC<sub>50(TCDD)</sub>) was calculated to be 60 pM by fitting the dose-response data into a three parameter sigmoidal Hill curve using Sigma Plot (SPSS, Chicago, IL, USA). The minimal detection limit was 64 fg/well with an intra CV of 5–10% and an inter CV of 10–20%. The sample solvent controls (+/- EC<sub>50(TCDD)</sub>) consisted of sample solvent treated like the serum extract but without the extract, and the EC<sub>50(TCDD) </sub>was used as parallel positive control in each assay on each plate.</p><p>AhR agonistic effect (AhRag) was determined by exposure of the cells to serum extracts alone, and the competitive AhR effect (AhRcomp) was determined upon co-exposure with serum extract and 60 pM TCDD (EC<sub>50</sub>). The AhR-CALUX assay can be described shortly as follows: The Hepa1.12cR cells were seeded into sterile 96-well white CulturPlate™ (Packard Instruments) at 6 × 10<sup>4 </sup>cells per well and cultured in supplemented α-MEM containing 400 μg/ml geneticin (G418, Sigma-Aldrich) at 37°C, 5% CO<sub>2 </sub>in 95% humidified air for 24 h, allowing cells to reach 90–100% confluence. Then media were removed and the cells were in parallel exposed to the serum extract, serum extract plus 60 pM TCDD, and sample solvent controls in a total volume of 100 μl per well in triplicate. After exposure for 4 h, cells were washed with phosphate-buffered saline (PBS, pH 7.4) followed by addition of cell lyses buffer. Luciferase activity and cell protein were determined as described [<xref ref-type="bibr" rid="B33">33</xref>]. The luciferase activity was expressed in relative light units per microgram protein (RLUs/μg protein). The average intra-sample CV was 11% and the inter CV of solvent control was 19%.</p><p>No cell toxicity on Hepa1.12cR cells was determined by the CellTiter 96 assay from Promega (Madison. WI, US) [<xref ref-type="bibr" rid="B33">33</xref>] after exposure to the tested serum extract.</p><p>The determinations of AhR activity was in good inter-lab precision, as determined from the results of interlaboratory comparison program (Second round of interlaboratory comparison of dioxin-like compounds in food using bioassay, Orebro, Sweden).</p></sec></sec><sec><title>2. 4. Calculation and statistical analysis</title><p>In the independent assays the activity differences between the triple serum extract determinations and their respective solvent controls (% agonistic, % antagonistic and % additive/synergistic, Table <xref ref-type="table" rid="T2">2</xref>) were evaluated using the Student t-test (Microsoft Excel).The data was given as RLU per ml serum and the value of the solvent controls was 6.67 RLU/ml serum.</p><table-wrap position="float" id="T2"><label>Table 2</label><caption><p>AhR-mediated serum activities, TCDD equivalents and lipid adjusted POP markers in serum</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td></td><td align="center"><bold>Greenland</bold></td><td align="center"><bold>Warsaw</bold></td><td align="center"><bold>Sweden</bold></td><td align="center"><bold>Kharkiv</bold></td><td align="center"><bold>All</bold></td></tr></thead><tbody><tr><td align="center"><bold>AhRag*</bold><sup>1 </sup>(RLU/ml serum)</td><td align="center">N</td><td align="center">75</td><td align="center">99</td><td align="center">78</td><td align="center">86</td><td align="center">338</td></tr><tr><td></td><td align="center"><bold>Median</bold></td><td align="center"><bold>24</bold></td><td align="center"><bold>34</bold></td><td align="center"><bold>33</bold></td><td align="center"><bold>26</bold></td><td align="center"><bold>29</bold></td></tr><tr><td></td><td align="center">Min</td><td align="center">6.6</td><td align="center">11</td><td align="center">8.0</td><td align="center">8.8</td><td align="center">6.6</td></tr><tr><td></td><td align="center">Max</td><td align="center">257</td><td align="center">118</td><td align="center">103</td><td align="center">56</td><td align="center">257</td></tr><tr><td></td><td align="center">% agonist</td><td align="center">92</td><td align="center">100</td><td align="center">95</td><td align="center">100</td><td align="center">97</td></tr><tr><td align="center"><bold>AhR-TEQ*</bold><sup>2 </sup>(pg/g lipid)</td><td align="center">N</td><td align="center">70</td><td align="center">99</td><td align="center">76</td><td align="center">80</td><td align="center">325</td></tr><tr><td></td><td align="center"><bold>Median</bold></td><td align="center"><bold>197</bold></td><td align="center"><bold>312</bold></td><td align="center"><bold>428</bold></td><td align="center"><bold>337</bold></td><td align="center"><bold>310</bold></td></tr><tr><td></td><td align="center">Min</td><td align="center">38</td><td align="center">72</td><td align="center">104</td><td align="center">110</td><td align="center">38</td></tr><tr><td></td><td align="center">Max</td><td align="center">1188</td><td align="center">1054</td><td align="center">1261</td><td align="center">781</td><td align="center">1261</td></tr><tr><td align="center"><bold>AhRcomp*</bold><sup>3 </sup>(RLU/ml serum)</td><td align="center">N</td><td align="center">75</td><td align="center">99</td><td align="center">78</td><td align="center">86</td><td align="center">339</td></tr><tr><td></td><td align="center"><bold>Median</bold></td><td align="center"><bold>8.3</bold></td><td align="center"><bold>6.4</bold></td><td align="center"><bold>6.2</bold></td><td align="center"><bold>5.5</bold></td><td align="center"><bold>6.8</bold></td></tr><tr><td></td><td align="center">Min</td><td align="center">3.9</td><td align="center">3.2</td><td align="center">1.6</td><td align="center">1.5</td><td align="center">1.5</td></tr><tr><td></td><td align="center">Max</td><td align="center">16</td><td align="center">9.1</td><td align="center">10</td><td align="center">11</td><td align="center">16</td></tr><tr><td></td><td align="center">% add/syn</td><td align="center">41</td><td align="center">3.0</td><td align="center">6.4</td><td align="center">18</td><td align="center">16</td></tr><tr><td></td><td align="center">%antagonist</td><td align="center">2.7</td><td align="center">8.0</td><td align="center">12</td><td align="center">34</td><td align="center">14</td></tr><tr><td align="center"><bold>CB-153 </bold>(ng/g lipid)</td><td align="center">N</td><td align="center">74</td><td align="center">100</td><td align="center">98</td><td align="center">82</td><td align="center">355</td></tr><tr><td></td><td align="center"><bold>Median</bold></td><td align="center"><bold>220</bold></td><td align="center"><bold>16</bold></td><td align="center"><bold>210</bold></td><td align="center"><bold>47</bold></td><td align="center"><bold>78</bold></td></tr><tr><td></td><td align="center">Min</td><td align="center">5.1</td><td align="center">3.3</td><td align="center">41</td><td align="center">5.5</td><td align="center">3.3</td></tr><tr><td></td><td align="center">Max</td><td align="center">5500</td><td align="center">130</td><td align="center">1500</td><td align="center">200</td><td align="center">5500</td></tr><tr><td align="center"><bold><italic>p,p'</italic>-DDE </bold>(ng/g lipid)</td><td align="center">N</td><td align="center">74</td><td align="center">100</td><td align="center">98</td><td align="center">82</td><td align="center">355</td></tr><tr><td></td><td align="center"><bold>Median</bold></td><td align="center"><bold>630</bold></td><td align="center"><bold>570</bold></td><td align="center"><bold>240</bold></td><td align="center"><bold>880</bold></td><td align="center"><bold>560</bold></td></tr><tr><td></td><td align="center">Min</td><td align="center">66</td><td align="center">240</td><td align="center">55</td><td align="center">320</td><td align="center">55</td></tr><tr><td></td><td align="center">Max</td><td align="center">13000</td><td align="center">2100</td><td align="center">2300</td><td align="center">12000</td><td align="center">13000</td></tr></tbody></table><table-wrap-foot><p>*<sup>1</sup>: AhRag: AhR activity of serum extract alone determined as the relative luciferase activity (RLU) per ml serum; Solvent background control = 6.67 RLU/ml serum. The % agonistic indicates the % of samples eliciting a significant increase in AhR activity compared to the solvent control. *<sup>2 </sup>AhR-TEQ (TCDD equivalents): The samples eliciting significantly agonistic activity was calculated by interpolation to the TCDD dose-response curve using the Sigmaplot program, given as pg/g serum lipid. *<sup>3 </sup>AhRcomp: AhR competitive activity of s<italic>erum extract </italic>+ 60 pM TCDD (EC<sub>50</sub>) given as RLU/ml serum. EC<sub>50 </sub>solvent control = 6.67 RLU/ml serum; % add/syn and %antagonistic indicates the % of samples responding with a further increase or decrease of the EC<sub>50(TCDD) </sub>induced activity, respectively.</p></table-wrap-foot></table-wrap><p>The CALUX-based AhR-TEQs values of serum extract were obtained by interpolation of AhRag values onto the TCDD dose-response sigmodal Hill curve (see <xref ref-type="supplementary-material" rid="S1">additional file 1</xref>: TCDD dose-response curve for AhR mediated luciferase activity). Only the AhRag values being significantly higher than the solvent control and in the linear range of the TCDD dose-response curve were used to calculate AhR-TEQ.</p><p>The natural logarithmic transformed AhR-mediated activities and markers of POPs improved the normality (checked by Q-Q plots) and homogeneity of variance, and the statistical analysis was performed on the ln-transformed data. The comparisons of means between the different variables (POP markers, AhRag, AhR-TEQ and AhRcomp) were performed with One-way ANOVA test. When ANOVA showed statistical significant difference complementary multiple comparison <italic>ad hoc </italic>tests was performed. Test for equal variances was performed with Levene's test. The least-significant difference (LSD) test was used if the variables have equal variance; otherwise Dunett T3 test was used.</p><p>Bivariate correlations were evaluated by Spearman's rank correlation test. The overall association between the POP markers and AhR-mediated activities across the study groups (combined data) was assessed by comparing the regression lines for each study group by using multiple regression analysis.</p><p>Up to date few studies on dioxin-like activities in human serum have been reported [<xref ref-type="bibr" rid="B34">34</xref>-<xref ref-type="bibr" rid="B37">37</xref>], and thus the knowledge is limited about which dietary or other life-style determinants might affect serum dioxin-like activity. Our hypothesis is that potential determinants of POP bioaccumulation might also be potential determinants for serum dioxin-like activity. Guided by the literature [<xref ref-type="bibr" rid="B38">38</xref>] and also from the assessment of the main Inuendo study populations [<xref ref-type="bibr" rid="B26">26</xref>,<xref ref-type="bibr" rid="B27">27</xref>], age and seafood are known determinants affecting the POP serum level. Moreover, lifestyle characteristics (Table <xref ref-type="table" rid="T1">1</xref>) were evaluated as potential determinants of AhRag, AhR-TEQ and AhRcomp levels. Multiple linear regression model was used to assess the impact of the POP biomarkers on AhRag, AhR-TEQ and AhRcomp. The impact of potential confounders were evaluated by entering blocks of variables together with either CB-153 or <italic>p,p'</italic>-DDE as follows: in the first step, age and seafood intake were included in the model, and in the second step additionally smoking status, BMI, coffee intake and alcohol consumption were included in the model. Due to many missing values on the potential confounders the number of available observations in the confounder analyses are much smaller than in the unadjusted analysis on the full dataset (full dataset: n = 338, first step of confounders: n = 232, second step: n = 164). A reduction of the number of observations with more than 50% might introduce serious selection problems, and hence the confounder analyses might lack greater validity.</p><p>The AhR activities were tested in protein free serum extract, which may still contain endogenous steroids, thus testosterone (total and free) and estradiol [<xref ref-type="bibr" rid="B39">39</xref>] were further included in the linear regression model on the combined study group data.</p><p>The inter-population variations in POP markers, AhRag, AhR-TEQ and AhRcomp serum level were also assessed by linear regression models. In these models age was considered as a potential confounder to make age-adjusted comparisons.</p><p>The statistical analysis was performed on SPSS 13.0 (SPSS Inc. Chicago, IL, USA). The statistical significant level was set to p ≤ 0.05.</p></sec></sec><sec><title>3. Results</title><sec><title>3.1. The basic characteristics and serum levels of CB-153 and p,p'-DDE</title><p>The distributions of demographic and lifestyle factors (Table <xref ref-type="table" rid="T1">1</xref>) and the serum CB-153 and <italic>p,p'</italic>-DDE levels (Table <xref ref-type="table" rid="T2">2</xref>) of the 338 adult males in this study were similar to that of the main Inuendo study population [<xref ref-type="bibr" rid="B26">26</xref>,<xref ref-type="bibr" rid="B27">27</xref>]. The Greenland and Swedish groups elicited higher CB-153 level than the Kharkiv and Warsaw group, and the level of CB-153 of Warsaw group was significantly lower than the other groups (Table <xref ref-type="table" rid="T2">2</xref> and see <xref ref-type="supplementary-material" rid="S2">additional file 2A</xref>: Multiple comparisons of variables). The <italic>p,p'</italic>-DDE level in Kharkiv was the highest, followed by Greenland and Warsaw, and the significant lowest level was found for the Swedish fishermen (Table <xref ref-type="table" rid="T2">2</xref> and see <xref ref-type="supplementary-material" rid="S2">additional file 2A</xref>: Multiple comparisons of variables). As for the main study [<xref ref-type="bibr" rid="B27">27</xref>], age adjustment did not change the pattern of difference for the POP markers between the study groups.</p><p>A higher correlation between serum concentration of CB-153 and <italic>p,p'</italic>-DDE was found for the Greenlandic (r<sub>s </sub>= 0.94, p < 0.001) and Swedish groups (r<sub>s </sub>= 0.75, p < 0.001), and lower correlations were observed for the Kharkiv (r<sub>s </sub>= 0.45, p < 0.001) and the Warsaw (r<sub>s </sub>= 0.27, p < 0.01) study groups (see <xref ref-type="supplementary-material" rid="S2">additional file 2B</xref>: Spearman's correlation between serum AhR activities and the levels of CB-153 and <italic>p.p'</italic>-DDE).</p></sec><sec><title>3.2. Agonistic and competitive AhR activity in the four study groups</title><p>Almost all serum samples (97%) showed AhRag activity significantly higher than the solvent control (Table <xref ref-type="table" rid="T2">2</xref>). The AhRag activity and AhR-TEQ differed significantly among the study groups (Table <xref ref-type="table" rid="T2">2</xref> and see <xref ref-type="supplementary-material" rid="S2">additional file 2A</xref>: Multiple comparisons of variables): the Greenland and Kharkiv groups had lower AhRag medians than Sweden and Warsaw groups (Fig. <xref ref-type="fig" rid="F1">1A</xref> and see <xref ref-type="supplementary-material" rid="S2">additional file 2A</xref>: Multiple comparison of variables), and the AhR-TEQ median level of Greenland group was significantly lower than that of the European study groups (Table <xref ref-type="table" rid="T2">2</xref>, Fig <xref ref-type="fig" rid="F1">1B</xref> and see <xref ref-type="supplementary-material" rid="S2">additional file 2A</xref>: Multiple comparisons of variables).</p><fig position="float" id="F1"><label>Figure 1</label><caption><p><bold>AhR-CALUX activities of the study groups</bold>. (A) Agonistic activity of serum extracts alone (AhRag), (B) AhR-TEQ (AhR-CALUX- TCDD toxic equivalent) and (C) competitive AhR activity upon cotreatment with 60pMTCDD (EC<sub>50</sub>) and serum extract (AhRcomp). For the AhRag the outliers ranging from 86.01 to 111.28 RLU/ml serum and extreme values (117.87 – 257.13 RLU/ml serum) are not shown. The reference lines of the respective solvent controls ± SD (6.67 ± 0.74) are given as dashed lines.</p></caption><graphic xlink:href="1476-069X-5-14-1"/></fig><p>The Greenlandic AhRcomp activity, eliciting the highest incidence (41%) of sample with further increasing TCDD induced AhR activity, was significantly higher than for the European groups (Table <xref ref-type="table" rid="T2">2</xref>, Fig. <xref ref-type="fig" rid="F1">1C</xref> and see <xref ref-type="supplementary-material" rid="S2">additional file 2A</xref>: Multiple comparisons of variables), whereas the Kharkiv group showed the highest frequency of sample with antagonistic effect on TCDD induced AhR activity (Table <xref ref-type="table" rid="T2">2</xref>).</p><p>Compared to the crude data, the pattern of differences of AhRag, AhR-TEQ and AhRcomp between the study groups did not change after adjustment for age (data not shown).</p></sec><sec><title>3.3. Associations between AhRag, AhR-TEQ, AhRcomp and the POP markers</title><p>Significant inverse correlation (r<sub>s </sub>= -0.30, p < 0.01) between AhRcomp and CB-153 was observed for the Kharkiv group (Fig. <xref ref-type="fig" rid="F2">2</xref> and see <xref ref-type="supplementary-material" rid="S2">additional file 2B</xref>: Spearman's correlation between serum AhR activities and the levels of CB-153 and <italic>p.p'</italic>-DDE). Neither AhRag nor AhR-TEQ was found to correlate to the two POP markers for any of the study groups (see <xref ref-type="supplementary-material" rid="S2">additional file 2B</xref>: Spearman's correlation between serum AhR activities and the levels of CB-153 and <italic>p.p'</italic>-DDE).</p><fig position="float" id="F2"><label>Figure 2</label><caption><p><bold>The relationship between AhRcomp activity and CB-153 for the Kharkiv group</bold>. Scatterplot of correlation between serum AhRcomp and CB-153. For definitions of AhRcomp see legend to Table 2. Ln-transformed data was used. RLU: relative luciferase units.</p></caption><graphic xlink:href="1476-069X-5-14-2"/></fig><p>Adjustment for potential confounders in the multiple regression models did not change the strength of association between exposure variables (POP markers) and outcome variables (AhR-mediate activities) when compared with the unadjusted models. Further adjustment for endogenous testosterone and estradiol neither changed this pattern (data not shown).</p></sec><sec><title>3. 4. Multiple regressions of AhR-mediated activities on POP markers across the study groups</title><p>Scatter plots of AhR-mediated activities against POP markers for the study groups are shown in Figure <xref ref-type="fig" rid="F3">3</xref>. Multiple regression analysis showed homogeneity of associations between CB-153 or <italic>p,p' </italic>-DDE and AhRag and AhR-TEQ as well as AhRcomp and <italic>p,p' </italic>-DDE across the study groups (Table <xref ref-type="table" rid="T3">3</xref>), i.e. parallel regression lines among study groups. Furthermore, a model with parallel regression lines showed a significant differences between the intercepts of the study groups (Table <xref ref-type="table" rid="T3">3</xref>), thus the differences in AhRag/AhR-TEQ/AhRcomp between study groups still exist after adjustment for CB-153 or <italic>p,p' </italic>-DDE. However, heterogenetic associations between serum CB-153 and the AhRcomp across the study groups were found (Table <xref ref-type="table" rid="T3">3</xref>). Thus the difference of AhRcomp between the study groups was correlated to the CB-153 level, which supports the negative correlation between AhRcomp and CB-153 for the Kharkiv group.</p><fig position="float" id="F3"><label>Figure 3</label><caption><p><bold>The association between AhR activity and POP markers in the study groups</bold>. The AhR activities are given for the four country based study groups as relation between (A) AhRag and CB-153, (B) AhRcomp and CB-153, (C) AhRag and <italic>p,p'</italic>-DDE, (D) AhRcomp and <italic>p,p'</italic>-DDE. Ln-transformed data was used. For the definition of AhRag and AhRcomp, see the legend to Table 2. RLU: relative luciferase units. <inline-graphic xlink:href="1476-069X-5-14-i1.gif"/></p></caption><graphic xlink:href="1476-069X-5-14-3"/></fig><table-wrap position="float" id="T3"><label>Table 3</label><caption><p>Multiple regressions of the combined study groups</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left">Response variable</td><td align="center">Homogeneity of slope (p value)</td><td align="center">Common slope, Estimate (SE), p value</td><td align="center">Common intercept (p value)</td><td align="center">Adjusted R square</td></tr></thead><tbody><tr><td align="left"><bold>AhRag (n = 327)</bold></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">CB-153</td><td align="center">0.95</td><td align="center">-0.03 (0.04), 0.43</td><td align="center"><bold>0.001</bold></td><td align="center">0.05</td></tr><tr><td align="left"><italic>p,p'</italic>-DDE</td><td align="center">0.28</td><td align="center">-0.02 (0.05), 0.70</td><td align="center"><bold>0.001</bold></td><td align="center">0.05</td></tr><tr><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left"><bold>AhR-TEQ (n = 324)</bold></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">CB-153</td><td align="center">0.54</td><td align="center">0.11(0.04), 0.77</td><td align="center"><bold>< 0.001</bold></td><td align="center">0.20</td></tr><tr><td align="left"><italic>p,p'</italic>-DDE</td><td align="center">0.99</td><td align="center">0.08(0.05), 0.07</td><td align="center"><bold>< 0.001</bold></td><td align="center">0.20</td></tr><tr><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left"><bold>AhRcomp (n = 327)</bold></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">CB-153</td><td align="center"><bold>0.01</bold></td><td align="center">-*</td><td align="center">-*</td><td align="center">-*</td></tr><tr><td align="left"><italic>p,p'</italic>-DDE</td><td align="center">0.80</td><td align="center">-0.03 (0.03), 0.24</td><td align="center"><bold>< 0.001</bold></td><td align="center">0.18</td></tr></tbody></table><table-wrap-foot><p>Both AhR activities and POP markers are ln transformed. Homogeneity of slope: test for homogeneity of association between exposure variables and outcome variables across the study group (p > 0.05, accept the hypotheses of homogeneity of slope). Common slope: the estimated common slope across study groups assuming homogeneity (p > 0.05, accept the hypotheses that slope equals to zero). Common intercept: test of a common intercept across study groups assuming a common slope (p > 0.05, accept the hypotheses having common intercept across the study groups). Adjusted R square assumes a common slope. *: Since heterogeneity of slope exists between CB-153 and AhRcomp across the study groups, i.e. there were country differences in the association of CB-153 and AhRcomp, no further evaluation was performed.</p></table-wrap-foot></table-wrap></sec></sec><sec><title>4. Discussion</title><p>In the present study we measured the integrated AhR mediated activity in the lipophilic serum fraction containing POPs using the mechanistically based AhR-CALUX bioassay. No consistent correlation between the POP markers (CB-153 and <italic>p,p'</italic>-DDE) and AhR mediated activities were found. However, in accordance with its high incidence of antagonistic AhRcomp activity, a negative correlation between AhRcomp and CB-153 was found in Kharkiv group. This finding is supported by earlier reports that CB-153 and associated compounds antagonized TCDD induced AhR action [<xref ref-type="bibr" rid="B40">40</xref>,<xref ref-type="bibr" rid="B41">41</xref>], as well as observation in our laboratory (Long and Bonefeld-Jorgensen, manuscript in prep.).</p><p>The study showed that 97% of the serum samples across the study groups elicited significant agonistic AhR activity. The median of AhRag activity in the Warsaw and Swedish groups reached higher level compared to the Kharkiv and the Greenlandic groups. Moreover, the median levels of AhR-TEQ of the European study groups were significantly higher than that of the Greenlandic Inuit's having high burden of both CB-153 and <italic>p,p'</italic>-DDE. We do not exactly know the dietary habit of the Warsaw study group. However, relatively high level of dioxins and DLCs were reported to be produced and emitted to the Polish environment [<xref ref-type="bibr" rid="B42">42</xref>,<xref ref-type="bibr" rid="B43">43</xref>]. The relatively higher AhRag and AhR-TEQ in the Swedish study group may be caused by high exposure to dioxins and/or dioxin-like PCBs (DL-PCBs) [<xref ref-type="bibr" rid="B8">8</xref>].</p><p>Except for a recent study of Slovakia males [<xref ref-type="bibr" rid="B44">44</xref>], the AhR-TEQ in the present study was relatively higher (2~8 fold) than other similar studies [<xref ref-type="bibr" rid="B35">35</xref>,<xref ref-type="bibr" rid="B37">37</xref>], which can be related to the difference in exposure time (4 hours versus 24 hours). During the optimal exposure time (4 h) of Hepa1.12cR cell line used in this study [<xref ref-type="bibr" rid="B45">45</xref>,<xref ref-type="bibr" rid="B46">46</xref>], other active compounds in the lipophilic serum fraction such as polycyclic aromatic hydrocarbons (PAHs), polybrominated biphenyls (PBBs), polyhalogenated naphthalenes may interact and activate the AhR in the CALUX- bioassay. In other studies e.g. using rat H4IIE cells, the labile AhR agonists in the serum such as some PAHs and other chemicals may be degraded or metabolized during the 24 hours of exposure resulting in a relatively lower AhR response [<xref ref-type="bibr" rid="B47">47</xref>]. PAHs elicite high AhR response after short exposure [<xref ref-type="bibr" rid="B47">47</xref>,<xref ref-type="bibr" rid="B48">48</xref>], nevertheless, in the present study the PAHs level would be very low in the serum extract because the purification method was not built specifically for these molecules. We do not assume that the endogenous compounds play a major role in our AhR response data because fatty acids and other nonclassical AhR ligands may be removed from the crude extract in the cleanup procedure [<xref ref-type="bibr" rid="B24">24</xref>,<xref ref-type="bibr" rid="B32">32</xref>]. However, we can not exclude that some organic endogenous AhR inducers possibly passed through the purification columns and contribute to the AhR response [<xref ref-type="bibr" rid="B49">49</xref>]. Owing to the possible existence of cross-talk between AhR and ER and AR, it can not be excluded whether the xenohormones and/or endogenous sex hormones (estradiol and testosterone) influence the observed AhR response. Although the influence is expected to be of minor importance since no correlation between the sex hormones and AhR-mediated activity across the study groups were found. The extract in this study contains most organochlorine compounds including organochlorine pesticides like hexachlorobenzene (HCB) that can contribute to the AhR response [<xref ref-type="bibr" rid="B50">50</xref>]. Future investigations are required to elucidate the profile of serum compounds contributing to the AhR-CALUX response. Possibly, the gender might also influence the lipid adjusted AhR-TEQ level since the participants in other studies were female [<xref ref-type="bibr" rid="B35">35</xref>,<xref ref-type="bibr" rid="B36">36</xref>].</p><p>The differences in AhRag/AhRcomp activities among the study groups suggest that there are regional differences in profiles of POPs, PAHs and/or other lipophilic AhR activating compounds. The higher frequency of samples further increasing the TCDD induced AhR activity in Inuit's indicated the presence of compounds which can enhance the effect of the TCDD, the most potent AhR ligand. Considering the further increase of AhRcomp using the TCDD dose-response curve (see <xref ref-type="supplementary-material" rid="S2">additional file 2</xref>: TCDD dose-response curve for AhR mediated luciferase activity), the median AhR-TEQ determined in serum of the corresponding subgroup (n = 180) was calculated to be 1.3 ng/g lipid, suggesting an increasing risk from DLCs when a strong AhR ligand exists simultaneously in the body.</p><p>Studies of Inuit populations in Canada support in general the use of CB-153 as a surrogate marker of exposure to non-DL-PCBs present in the Arctic food-chain [<xref ref-type="bibr" rid="B51">51</xref>]. However, the level of coplanar and <italic>non-coplanar </italic>PCBs was shown to differ between Canadian Inuits and Caucasian [<xref ref-type="bibr" rid="B32">32</xref>,<xref ref-type="bibr" rid="B52">52</xref>]. For non-occupational exposed Inuit populations in the Arctic Quebec the ratio between coplanar PCBs (e.g. CB-126 or CB-169) and the <italic>non-coplanar </italic>PCB (e.g. CB-153) of Inuits was lower than that of the Caucasian reference group [<xref ref-type="bibr" rid="B52">52</xref>]. Moreover, DL-PCBs contributed to a larger extent to the chemical calculated TEQ than PCDD and PCDFs, with the <italic>mono-ortho </italic>coplanar CB-118 as major contributor for the total toxicity [<xref ref-type="bibr" rid="B52">52</xref>]. Similar difference may also exist for the populations included in this study, supporting the higher AhRag or AhR-TEQ level of Europeans.</p><p>Previously, CB-153 was reported to be highly correlated with calculated chemical-derived total TEQs and/or PCDD/PCDFs TEQs and/or <italic>non-ortho </italic>PCBs TEQs [<xref ref-type="bibr" rid="B29">29</xref>]. It should be noted that the chemical derived TEQ was calculated according to analytical chemistry data of some congeners under the assumption of additivity. Even though additive effects of PCDDs/PCDFs/DL-PCBs are predominating, non-additive effects such as antagonism and synergism can apply to interactions between individual DLCs in a complex mixture [<xref ref-type="bibr" rid="B53">53</xref>]. The TEQ based on the AhR-CALUX bioassay represents the integrated sum of dioxin-like activities including additive, synergistic and antagonistic effect. Few epidemiology studies have reported analysis of association between <italic>di-ortho </italic>PCB (including CB-153) and CALUX-TEQ, and the results were contradictory. A positive correlation was reported by Pauwels et al [<xref ref-type="bibr" rid="B35">35</xref>]. However, a recent report of the potential of the CALUX bioassay to estimate TEQ in plasma of Italian women with background exposure to dioxin and DLCs showed no significant correlation between CALUX-TEQ and the sum of four major PCB congeners (CB-118, CB-138, CB-153, CB-180) [<xref ref-type="bibr" rid="B36">36</xref>]. Similar result was reported in a Belgium study [<xref ref-type="bibr" rid="B34">34</xref>]. The contradictory results may be related to the differences in study design, sample selection (gender and age of subjects, serum or plasma), composition and concentrations of bio-accumulated compounds in the blood and/or different sample extraction methods used in the CALUX bioassay as discussed [<xref ref-type="bibr" rid="B36">36</xref>,<xref ref-type="bibr" rid="B49">49</xref>,<xref ref-type="bibr" rid="B54">54</xref>]. Moreover, an important part of dioxin-like activity is elicited especially by PCDD/Fs when their concentration is high in the sample [<xref ref-type="bibr" rid="B55">55</xref>]. One can find some correlations between PCB levels and dioxin-like activity only in the case that the concentrations of PCDD/Fs are low or comparable in different cohorts [<xref ref-type="bibr" rid="B34">34</xref>,<xref ref-type="bibr" rid="B56">56</xref>].</p></sec><sec><title>5. Conclusion</title><p>97% of analyzed samples elicited significant agonistic AhR induced activity. The level of AhR mediated activities differed among the study groups. European groups elicited higher AhR-TEQ than the Greenland Inuits, suggesting a higher exposure to DLCs. In addition, dietary habits/life style factors and the genetic difference between Inuits and Caucasians [<xref ref-type="bibr" rid="B57">57</xref>] may also be taken into account.</p><p>No consistent significant correlation between CB-153 and <italic>p,p'</italic>-DDE and AhR activities was observed and these two selected POP markers cannot alone predict the contribution of POPs, PAHs, and other lipophilic xenobitics to serum dioxin-like activity. Other more sensitive and specific tentative markers such as DL-PCBs (e.g. the mono-ortho congener CB-118 and CB-156) should be included in future epidemiology studies.</p><p>Since some less persistent compounds might contribute to the AhR-CALUX response due to the selected fractionation technique and that <italic>in vitro </italic>AhR-CALUX detects the overall dioxin-like response, it is not clear which compounds contribute to the observed dioxin-like activity. Thus the serum dioxin-like activity determined in this study must be interpreted as an independent parameter, complementary to chemical data. It should be kept in mind that the AhR-CALUX bioassay is not a substitute of actual chemical analysis by GC-MS techniques but provide biologically relevant results and normally is utilized as first tier screening tool followed by the chemical analysis to identify specific response compounds [<xref ref-type="bibr" rid="B24">24</xref>]. AhR-CALUX bioassay provides an overall biological response/potency of mixture, whereas chemical analysis provides the concentration of specific compounds in the mixture.</p></sec><sec><title>Abbreviations</title><p><bold>PCDDs/PCDFs </bold>polychlorinated dibenzo-p-dioxins/furans</p><p><bold>PCBs </bold>polychlorinated biphenyls</p><p><bold>DDT </bold>2,2-bis(p-chlorophenyl)-1,1,1-trichloroethane</p><p><bold>POPs </bold>persistent organochlorine pollutants</p><p><bold>CB-153 </bold>2,2',4,4',5,5'-hexachlorobiphenyl</p><p><bold><italic>p,p'</italic>-DDE </bold>1,1-dichloro-2.2-bis (p-chlorophenyl)-ethylene</p><p><bold>TCDD </bold>2,3,7,8-tetrachlorodibenzo-p-dioxin</p><p><bold>AhR </bold>aryl hydrocarbon receptor</p><p><bold>TEQs </bold>TCDD toxic equivalents</p><p><bold>CALUX </bold>Chemical activated luciferase gene expression</p><p><bold>TEFs </bold>Toxic Equivalency Factors</p><p><bold>DLCs </bold>Dioxin-like compounds</p><p><bold>DL-PCBs </bold>dioxin-like PCBs</p><p><bold>GC-MS </bold>gas chromatography mass spectrometry</p><p><bold>AhRag </bold>agonistic AhR activity</p><p><bold>AhRcomp </bold>competitive AhR activity</p></sec><sec><title>Competing interests</title><p>The author(s) declare that they have no competing interest</p></sec><sec><title>Authors' contributions</title><p>ML and ECB-J drafted the work and were the main responsible for design, performance, data evaluation and statistical analyses of the specific project; ML and BSA performed the mechanistic work; CL performed POP determinations in blood; JPB was main responsible for raising funding for the project. GT, HSP, VZ have been responsible for collecting blood samples and for obtaining the interview data. JPB, AG and LH initiated and designed the Inuedo project. JPB and GT coordinated the execution of project and GT had main responsibility for creating the joint database. All authors participated in the design of the study, comment on the draft and have read and approved the final 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</title><p><bold>TCDD dose-response curve for AhR mediated luciferase activity</bold>. The 96-well plates containing Hepa1.12cR cells at 90–100% confluence were incubated with TCDD at the indicated concentration for 4 hours. Luciferase activity in cell lysates were determined and corrected to cell protein. The data is expressed as the luciferase activity above solvent control which was set to 1. Values represent the mean ± S.D. (n ≥ 5).The EC<sub>50 </sub>is the half maximum effect concentration.</p></caption><media xlink:href="1476-069X-5-14-S1.doc" mimetype="application" mime-subtype="msword"><caption><p>Click here for file</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="S2"><caption><title>Additional file 2A</title><p><bold>Multiple comparisons of variables</bold>. Multiple comparisons were performed on ln-transformed data. The values given are p values. <bold>B</bold> Spearman's correlation between serum AhR activities and the levels of CB-153 and <italic>p,p' </italic>-DDE. Continuous data was used. Spearman's correlation data is given. For definition of AhRag, AhRcomp and AhR-TEQ see legend to Table <xref ref-type="table" rid="T2">2</xref>. Statistical significant data is given in bold.</p></caption><media xlink:href="1476-069X-5-14-S2.doc" mimetype="application" mime-subtype="msword"><caption><p>Click here for file</p></caption></media></supplementary-material></sec> |
Influence of gold nanoparticles on collagen fibril morphology quantified using transmission electron microscopy and image analysis | <sec><title>Background</title><p>Development of implantable biosensors for disease detection is challenging because of poor biocompatibility of synthetic materials. A possible solution involves engineering interface materials that promote selfassembly and adhesion of autologous cells on sensor surfaces. Crosslinked type-I collagen is an acceptable material for developing engineered basement membranes. In this study, we used functionalized gold nanoparticles as the crosslinking agent. Functionalized nanoparticles provide sites for crosslinking collagen as well as sites to deliver signaling compounds that direct selfassembly and reduce inflammation. The goal of this study was to obtain a quantitative parameter to objectively determine the presence of crosslinks.</p></sec><sec sec-type="methods"><title>Methods</title><p>We analyzed TEM images of collagen fibrils by two methods: Run length analysis and topology analysis after medial axis transform.</p></sec><sec><title>Results</title><p>Run length analysis showed a significant reduction of the interfibril spaces in the presence of nanoparticles (change of 40%, P < 0.05), whereas the fibril thickness remained unchanged. In the topological network, the number of elements, number of branches and number of sides increased significantly in the presence of nanoparticles (P < 0.05). Other parameters, especially the number of loops showed only a minimal and nonsignificant change. We chose a ratiometric parameter of the number of branches normalized by the number of loops to achieve independence from gross fibril density. This parameter is lower by a factor of 2.8 in the presence of nanoparticles (P < 0.05).</p></sec><sec><title>Conclusion</title><p>The numerical parameters presented herein allow not only to quantify fibril mesh complexity and crosslinking, but also to help quantitatively compare cell growth and adhesion on collagen matrices of different degree of crosslinking in further studies.</p></sec> | <contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Haidekker</surname><given-names>Mark A</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>HaidekkerM@missouri.edu</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Boettcher</surname><given-names>Lisa W</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>lwb6c9@mizzou.edu</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Suter</surname><given-names>Jonathan D</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>jds5y7@mizzou.edu</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Rone</surname><given-names>Rebecca</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>RebeccaRone@mizzou.edu</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Grant</surname><given-names>Sheila A</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>GrantSA@missouri.edu</email></contrib> | BMC Medical Imaging | <sec><title>Background</title><p>Implantable biosensors suffer from biofouling and fibrous encapsulation. The key to success with long term implantable sensors lies in integrating the sensor with the surrounding tissue to avoid the immune response. For example, intravascular devices could be coated with autologous endothelial cells to hide the material from the immune system [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B2">2</xref>]. Since these cells would not easily adhere to the biosensor material, a collagen layer would facilitate cell adhesion [<xref ref-type="bibr" rid="B3">3</xref>]. Furthermore, covalent crosslinking of the collagen increases stability of the collagen layer [<xref ref-type="bibr" rid="B4">4</xref>] and improves cell adhesion and proliferation [<xref ref-type="bibr" rid="B5">5</xref>], likely aided by the nanoscale substrate texture [<xref ref-type="bibr" rid="B6">6</xref>]. Recent work has focused on functionalized gold nanoparticles as crosslinking agents for collagen and as a facilitator for cell migration and ingrowth [<xref ref-type="bibr" rid="B7">7</xref>]. In addition, some studies indicate that increased substrate stiffness influences cell adhesion [<xref ref-type="bibr" rid="B8">8</xref>,<xref ref-type="bibr" rid="B9">9</xref>]. It is conceivable that enhanced crosslinking may stiffen the network and enhance cell ingrowth through the link between crosslinking and biomechanical properties.</p><p>The purpose of this study was to develop an objective image analysis method to quantify the degree of crosslinking of type-I collagen fibrils in the presence or absence of gold nanoparticles. On visual inspection of transmission electron images (TEM), differences could be seen where fibrils in the presence of nanoparticles subjectively appeared disordered and shaped in a more complex manner. We wanted to find a numerical parameter that describes the differences in morphology, therefore allowing to differentiate fibrils in the presence of nanoparticles from control fibrils in the absence of nanoparticles, and eventually promoting further studies where cell growth and adhesion may be related to this parameter.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>Collagen fibril formation</title><p>Gold nanoparticles of 10 nm average diameter, functionalized with a coating of gum arabic, were generously provided by Dr. Katish Katti, University of Missouri-Columbia. Gum arabic coating provides reaction sites for crosslinking with collagen [<xref ref-type="bibr" rid="B10">10</xref>]. Control collagen matrix was formed by mixing 160 μl PureCol™ (3.2 mg/ml bovine collagen solution, Inamed) with 650 μl water and 184 μl PBS. Collagen matrices in the presence of gold nanoparticles were created from 160 μl PureCol, 600 μl water, 240 μl PBS and 250 μl of aqueous gold nanoparticle suspension. Actual fibril formation was initiated by gradually raising both the temperature and pH of the collagen solution simultaneously. The temperature of the solution was increased to 37°C in a water bath. The pH was increased from 2.5 to 7.4 by adding phosphate buffer solution (PBS). This process of fibril formation was carried out over 4 hours. Final collagen was at a concentration of 0.5 mg/ml. The suspension of functionalized gold nanoparticles was added in a 1+4 volumetric ratio (one part gold, 4 parts collagen fibril solution).</p></sec><sec><title>Transmission Electron Microscopy (TEM) sample preparation</title><p>Samples of both the collagen control and collagen-nanoparticle solution were prepared by placing 10 μm of the sample on a TEM carbon grid, which was then allowed to air dry for 5 minutes. The sample grid was lightly rinsed with distilled water and dried with filter paper. A drop of 5% uracyl acetate solution was then placed on the grid, and the grid was slightly dried along the edge using filter paper after 10 minutes. This method ensured that the water content of the collagen fibers remained as constant as possible. Images of the matrix were acquired with a JEOL 1200EX transmission electron microscope at 80 kV. Images were printed on Film and scanned on an Epson Expression 800 scanner at 3600dpi and 16 bit/pixel to obtain digital images for analysis.</p></sec><sec><title>Image preprocessing and segmentation</title><p>The resulting images were scaled to a size of approximately 1000 by 1000 pixels while retaining the aspect ratio. Binary images were created through a filtering and segmentation process as follows: The scanned image was inverted to obtain bright features on a darker background. Four iterations of a grayscale rank filter, where each pixels is replaced by the median of its 3 by 3 neighborhood, were applied to reduce noise. Background inhomogeneities were then removed by creating a severely blurred image (convolution with a circular kernel of 120 pixels diameter) and by subtracting this blurred image from the rank-filtered image. The image was then binarized by applying Otsu's method [<xref ref-type="bibr" rid="B11">11</xref>]. The resulting binary image now contained the fibrils as well as the nanoparticles as white features over a black background. The nanoparticles were then eliminated with a feature size filter, which removed any connected region of less than 150 pixels from the image. The remaining features were finally filtered through morphological closing, so that any small interior holes or fissures disappeared.</p></sec><sec><title>Image analysis</title><p>Two different methods were devised to quantify fibril shape complexity. The first method is based on the analysis of run lengths. In this method, the image is scanned line by line in a similar manner as a fax is being scanned for transmission. In a fax, it is much more efficient to transmit the information "100 black dots" than actually transmitting 100 black dots. This information is called a run. Probability histograms of black and white run lengths were created. Our hypothesis was that curved and self-intersecting fibrils (higher feature complexity due to higher degree of crosslinking in the presence of nanoparticles) lead to a higher probability of shorter black runs, while not influencing the white runs (corresponding to unchanged fibril width).</p><p>The second method was based on the medial axis transform (MAT) [<xref ref-type="bibr" rid="B12">12</xref>]. The MAT erodes features to their central ridges, i.e. their skeletons. The resulting skeletonized image can be analyzed topologically, i.e. in terms of the number of nodes, endpoints, links (connections between nodes), branches (connection between a node and an endpoint), and loops. Our hypothesis is that a more complex shape, caused by curled and self-intersecting fibrils, results in a higher number of nodes and a higher number of loops.</p></sec><sec><title>Statistical analysis</title><p>Statistical analysis was performed by using Graphpad Prism 4.01 (Graphpad, San Diego). The images were divided into two equally-sized groups (with and without nanoparticles, n = 10 each). The parameters obtained for each group were tested for normal distribution (Kolmogorov-Smirnov-test) and, after passing the normality test, subjected to the t-test to test the hypothesis that the group means are different.</p></sec></sec><sec><title>Results</title><p>Typical images with fibrils in the presence and absence of nanoparticles with their corresponding segmentation results are shown in Figures <xref ref-type="fig" rid="F1">1</xref> and <xref ref-type="fig" rid="F2">2</xref>, respectively.</p><p>Figure <xref ref-type="fig" rid="F3">3</xref> shows the runs along one horizontal scanline close to the top of the segmented image in Figure <xref ref-type="fig" rid="F1">1</xref>. Black runs (value of zero) in Figure <xref ref-type="fig" rid="F3">3</xref> correspond to background runs, while white runs (value of 1) correspond to fibril runs. Runs were grouped into 16 equally-sized bins (1 pixel to the maximum run in the respective image), and collected over all scanlines. Two resulting typical run length histograms for a nanoparticle image and a control are displayed in Figure <xref ref-type="fig" rid="F4">4</xref>.</p><p>One possible metric for the quantitative description of the fibril complexity is the overall probability per image that an arbitrary run lies in the upper 75% of possible run lengths (i.e. falls into bins 4 to 16). The comparison of this metric over all 20 images (10 images with nanoparticles, 10 controls) showed that a background run was 40% more likely to be in the longer 75% of the runs in the control image than in the nanoparticle image. This difference was statistically significant (t-test, P < 0.05). A foreground (fibril) run was 11% more likely to be in the longer 75% of the runs in the control image than in the nanoparticle image, but this difference was not statistically significant (Figure <xref ref-type="fig" rid="F5">5</xref>).</p><p>Consistent with our hypothesis, the more complex shape of the fibrils in the presence of nanoparticles leads to shorter background runs. The difference can be quantified by computing the probability of the occurrence of long runs. The selected percentile of 75% was not chosen arbitrarily, but it was the percentile of highest specificity: The relative difference at very low percentiles was low, because the overall probability of very long runs was low. The difference increased with increasing percentile, but random influences increased as well (Figure <xref ref-type="fig" rid="F6">6</xref>).</p><sec><title>Skeleton analysis</title><p>The medial axis transform of the segmented image in Figure <xref ref-type="fig" rid="F1">1</xref> is displayed in Figure <xref ref-type="fig" rid="F7">7</xref>. Out of several topological parameters, only the number of elements, the number of branches, and the number of sides (branches and links) were significantly different between control images and nanoparticle images (Table <xref ref-type="table" rid="T1">1</xref>). In addition, the average length of the branches was significantly shorter in the nanoparticle images (34 ± 34 pixels) than in the control images (91 ± 76 pixels, P < 0.05). However, this metric is not scaling-independent.</p><p>The number of loops showed the smallest difference between nanoparticle and control images. A ratiometric number was devised by dividing the number of loops by the number of branches. This value was approximately 2.8 times higher in the control group (P < 0.05).</p></sec></sec><sec><title>Discussion</title><p>In recent years, efforts focused on creating functionalized surfaces for implants (see e.g. [<xref ref-type="bibr" rid="B13">13</xref>] for review). Collagen plays a key role in this process as a substrate to grow cells on. The gold nanoparticles used in this experiment were synthesized and stabilized with a thin film of gum arabic. Gum arabic is a mixture of branched polysaccharides andf glycoproteins that contains many functional sites for crosslinking or drug delivery [<xref ref-type="bibr" rid="B10">10</xref>]. Polysaccharides have been shown to be a useful class of biomaterials and typically have good biocompatibility [<xref ref-type="bibr" rid="B10">10</xref>]. Collagen fibrils contain amine and carboxyl sites, and are capable of nonspecific hydrogen bonding [<xref ref-type="bibr" rid="B14">14</xref>]. Due to the complex nature and variable structure of the gum arabic, the exact binding mechanism is still under investigation. However, we speculate that the nanoparticles bind either by nonspecific hydrogen absorption to the fibrils or by NH<sub>2</sub>/COOH cross-links. Objective, quantitative image analysis methods to determine the degree of crosslinking may be a helpful tool in this process.</p><p>Two different unsupervised methods to quantify fibril shape complexity and therefore fibril crosslinking in TEM images were devised in this study. The first was based on the analysis of black (background) and white (fibril) runs in the segmented image. It was hypothesized that the white runs do not show any statistical difference between nanoparticle and control images as the fibril width remains unchanged. The black runs, however, would be shorter in average because curling and self-intersection reduces the average size of the gaps between the fibrils. This hypothesis was supported by our data. A single metric to describe the overall fibril shape in one image was presented, which showed statistically significant differences between nanoparticle and control images. The second analysis method was based on the medial axis transform which provides topological parameters. Several topological parameters were found where a statistically significant difference between the two groups existed. Consistent with our initial hypothesis, those parameters which showed a significant difference (number of elements, number of branches, number of sides) were higher in the nanoparticle images compared to the control images. This indicates a higher complexity of the skeletonized shape.</p><p>The medial axis transform is a popular method to quantify network-like structures. For example, confocal images of fluorescently stained DNA and fibrin networks were analyzed for differences of topological and size parameters associated with cystic fibrosis [<xref ref-type="bibr" rid="B15">15</xref>]. Shah et al introduce a novel segmentation method for rod-like structures, which was specially developed for very thin networks at the high contrast typical for fluorescence. At higher magnifications, as provided by electron microscopy, the contrast is lower, and the structures become thicker. Morphological analysis has been found to provide significantly different values for segment length, number of segments, and segment orientation when cancer cells were treated with transforming growth factor α [<xref ref-type="bibr" rid="B16">16</xref>], a study based on scanning electron microscopy images of the cell's keratin filament network. The most crucial step for the analysis of this type of images is the segmentation step, where the features (fibers) are separated from background. The algorithm presented by Shah <italic>et al </italic>[<xref ref-type="bibr" rid="B15">15</xref>] inherently performs the segmentation, whereas no segmentation details are given by Beil et al [<xref ref-type="bibr" rid="B16">16</xref>]. Performance comparison with the methods that we presented in this study is therefore difficult. On the basis of our data, however, we conclude that our relatively straightforward approach to fibril segmentation yields sufficiently reliable results.</p><p>Shape analysis needs to be robust against various transformations. A curved shape remains – for the human eye – the same shape even if it is rotated and scaled. It is important that any metric reflects this tolerance of human perception against affine transformations. In addition, changes in magnification have a strong impact on length metrics. Therefore, studies using size-based parameters (such as e.g. [<xref ref-type="bibr" rid="B15">15</xref>] and [<xref ref-type="bibr" rid="B16">16</xref>]) must be particularly carefully designed to ensure unchanged imaging parameters throughout the study.</p><p>In this study, it was also important to ensure that the metric was not be affected by the absolute number of fibrils in the image. This is particularly important as fibrils in the presence of nanoparticles tend to cluster. For example, determining the relative segmented fibril area normalized by the total image area yields significantly higher values in the nanoparticle group (P < 0.0001). However, the segmented relative fibril area is not a shape measurement, and therefore not acceptable. Run-length histograms are affected by the total number of fibrils as well. The presence of a higher number of fibrils leads to shorter runs in average. For this reason, we devised a relative parameter – the probability of a run to be in the larger 75% of all runs. This parameter now becomes independent of affine transformations and even hypothetical tiling of the same image (simulation of more fibrils). The same consideration gave us the idea of using a ratiometric parameter in the analysis of fibril topology. While topological parameters are by nature invariant over affine transformations, the addition of more fibrils would change the parameters. Since the average number of loops is almost the same between nanoparticle and control images (less than 14% difference, and with a reverse trend), normalization of a strongly varying parameter (in this case, the number of sides) by the number of loops would yield a parameter that is independent from the overall fibril density. The number of sides divided by the number of loops is therefore our preferred metric to describe overall fibril shape.</p><p>The use of shape analysis instead of the analysis of size or density is particularly important, since water content of the collagen fibers may change because of evaporation in spite of the fibril treatment. Water content affects fibril length and density distribution. However, since the amount of drying is the same for control and nanoparticle networks, our comparative measurements will not be affected. In addition, we were careful to use size-independent metrics instead of fibril length, for example. Consequently, our proposed metrics would be robust against some change of water content.</p><p>The robustness of the parameters that we analyzed in this study is further supported by the wide variation of the fibril density in the images even within one group. The electron microscope images were taken by eye without consideration of future image analysis. As a consequence of this large intra-group variation, values between the groups overlap. We believe that a complete separation of the values between nanoparticle and control groups would be possible provided that the microscope field of view was selected to provide approximately matched fibril density. However, considering the huge variation of the fibril image, the parameters that we presented in this study are very robust and therefore well suited in the analysis of collagen fibril shape.</p><p>The presented image analysis methods rely on standard image operations. It is therefore possible to easily reproduce the image processing steps with most image analysis programs, even freely available ones such as the popular ImageJ <ext-link ext-link-type="uri" xlink:href="http://rsb.info.nih.gov/ij/"/>. On the long run, the parameters presented in this study may serve as a tool to further quantify the ability of crosslinked collagen to increase cell ingrowth and proliferation. For example, numerical parameters such as cell density or cell morphology could be correlated with the fibril crosslinking parameter. This would allow even more accurate design of collagen matrices to support stable biocompatible cell layers.</p></sec><sec><title>Conclusion</title><p>We presented two numerical parameters to quantitatively describe the complexity of collagen fibril morphology. The complexity parameters showed significant differences between TEM images of collagen fibrils crossinked with functionalized gold nanoparticles and control images without gold nanoparticles. Since it has been hypothesized that a complex collagen type-I morphology may provide some of the functionality of collagen type IV, this parameter may be used to quantitatively relate cell growth and proliferation on the collagen substrate to the collagen morphology, thus enabling to optimize the collagen substrate in further studies.</p></sec><sec><title>Competing interests</title><p>The authors declare that they have no competing interests.</p></sec><sec><title>Authors' contributions</title><p>M.H. refined the image analysis process, performed the actual image analysis and wrote the major part of the manuscript. L.W.B. was responsible for providing fibril samples and performing electron microscopy. In the course of a graduate class project in a biomedical imaging class, R.R. developed the basic algorithm for run length quantification, and J.D.S. developed the basic algorithm for morphological quantification. S.G. provided the methods to create collagen fibers and use functionalized nanoparticles to enhance crosslinking.</p></sec><sec><title>Pre-publication history</title><p>The pre-publication history for this paper can be accessed here:</p><p><ext-link ext-link-type="uri" xlink:href="http://www.biomedcentral.com/1471-2342/6/4/prepub"/></p></sec> |
The effect of web based depression interventions on self reported help seeking: randomised controlled trial [ISRCTN77824516] | <sec><title>Background</title><p>To date, there has been very little work investigating behaviour changes induced by interventions that are designed to increase help seeking. The present paper examines the effects of two Internet depression websites on help seeking.</p></sec><sec sec-type="methods"><title>Methods</title><p>414 individuals with elevated scores on a depression assessment scale were randomly allocated to a depression information website, a cognitive-behavioural skills training website (CBT) or an attention control condition. Reports of help seeking for specific treatments, from specific sources and for categories of treatments were assessed.</p></sec><sec><title>Results</title><p>Relative to the control, the depression information site was associated with decreases in seeking support from friends and family, the use of music and of everyday treatments and no increase in seeking evidence based interventions. The CBT site was associated with the report of help seeking for CBT, massage and exercise.</p></sec><sec><title>Conclusion</title><p>Methods to encourage the use of evidence-based treatments need further research to determine whether the assistance sought is evidence based and whether there are unintended effects.</p></sec> | <contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Christensen</surname><given-names>Helen</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>Helen.Christensen@anu.edu.au</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Leach</surname><given-names>Liana S</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>Liana.Leach@anu.edu.au</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Barney</surname><given-names>Lisa</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>Lisa.Barney@anu.edu.au</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Mackinnon</surname><given-names>Andrew J</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>Andrew.Mackinnon@anu.edu.au</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Griffiths</surname><given-names>Kathy M</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>Kathy.Griffiths@anu.edu.au</email></contrib> | BMC Psychiatry | <sec><title>Background</title><p>There is recognition from the results of representative epidemiological surveys that most adults with mental health problems do not receive treatment [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B2">2</xref>].</p><p>Despite this, there has been very little research investigating the effect of interventions designed to increase help seeking. Although the effects of educational pamphlets [<xref ref-type="bibr" rid="B3">3</xref>], voluntary screening [<xref ref-type="bibr" rid="B4">4</xref>] and educational programs in schools [<xref ref-type="bibr" rid="B5">5</xref>] on help seeking have been reported, these studies typically neither used randomised controlled designs nor investigated the types of treatment sought. One randomised controlled study [<xref ref-type="bibr" rid="B6">6</xref>,<xref ref-type="bibr" rid="B7">7</xref>] did examine the effect on help seeking attitudes of a consumer guide to depression. This study provided information about a range of treatments with demonstrated support for efficacy based on a systematic review of the evidence [<xref ref-type="bibr" rid="B8">8</xref>]. A major weakness of this study was that the short brochure used in the control condition also provided information about evidence based treatments for depression.</p><p>Using a randomised controlled trial design in which the efficacy of the active intervention arms had been established [<xref ref-type="bibr" rid="B9">9</xref>], the current study examined the effect of a depression information website (BluePages) on self reported help seeking. The effects were compared to those produced by a cognitive-behavioural skills training website (CBT) (MoodGYM) and an attention control condition. The inclusion of results from the MoodGYM arm of this trial constitutes an additional active comparison group. MoodGYM has been shown to be of equal effectiveness to BluePages in reducing depression but it does not directly encourage participants to seek additional help.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>Participants</title><p>A screening questionnaire was posted to 27,000 individuals aged 18 to 52 randomly selected from the Canberra electoral roll. 6130 people (22.7% response rate; 24.6% excluding those returned to sender) returned the screening questionnaire and 657 respondents met inclusion criteria. The screening questionnaire included a brief questionnaire (about two printed pages) assessing basic demographic information, education level, depression symptoms, help-seeking and internet use. Respondents were not eligible for inclusion in the trial if they; (i) scored below 12 on the Kessler Psychological Distress Scale (K10)[<xref ref-type="bibr" rid="B10">10</xref>], (ii) indicated they did not have Internet access at home or at work, (iii) indicated they did not wish to participate in an intervention, or (iv) were under psychiatric or psychological care. Of those who responded, 525 participants met these criteria, returned a letter of consent, and were randomised to the interventions. 414 participants provided post intervention depression scores and these individuals formed the sample analyzed here. The sample comprised 114 men and 300 women. The mean respondent age was 36.8 years (SD = 9.3), and the mean Kessler score was 17.5 (SD = 5.0) at screening.</p></sec><sec><title>Design</title><p>The present study used data from a recently completed randomised controlled trial of two Internet sites [<xref ref-type="bibr" rid="B9">9</xref>,<xref ref-type="bibr" rid="B11">11</xref>]. One of these Internet sites, BluePages [<xref ref-type="bibr" rid="B12">12</xref>], delivers information about a range of psychological, medical and alternative treatments and recommends those supported by scientific evidence. At the time of the trial, the BluePages website reviewed 5 medical, 7 psychological and 36 alternative treatments, recommending 3, 5 and 13 of these, respectively. The recommended therapies were based on systematic reviews of treatments [<xref ref-type="bibr" rid="B8">8</xref>,<xref ref-type="bibr" rid="B13">13</xref>]. The second site, MoodGYM, provides online cognitive behaviour therapy (CBT). The site consists of five modules that provide cognitive behaviour therapy, strategies to manage emotional distress, relaxation and problem solving. The site provides downloadable summary pages at the end of each module, which provide individualized symptom scores and homework completion outcomes. The site offers a user profile which tracks changes in anxiety, depression and dysfunctional thinking across modules. A control condition involved telephone contact with an interviewer but no access to the websites. Participants were randomly assigned to receive either; a) BluePages (n = 165, at endpoint n = 136), b) MoodGYM (n = 182; at endpoint n = 121) or c) a control condition (n = 178, at endpoint n = 157) in which an interviewer asked open-ended questions about factors that may influence depression. Participants were randomised to interventions using the SPSS SELECT CASES random selection option. A flow chart of participation in the trial is included in a previous publication [<xref ref-type="bibr" rid="B9">9</xref>].</p></sec><sec sec-type="methods"><title>Procedures</title><p>Participants completed questionnaires delivered through the mail at screening, pre-test, endpoint and at six months after the trial. Those in the BluePages and MoodGYM conditions were provided with a login identification number and a manual containing information about MoodGYM or BluePages. This manual outlined which sections or modules of the sites participants were to complete for each of 5 weeks. Interviewers maintained weekly telephone contact with participants in all conditions over the period of the intervention (a total of 6 contacts of approximately 10 minutes each).</p><p>Interviewers were provided with separate instruction booklets for each participant containing verbatim instructions for each of the weekly contacts. Participants randomised to the control intervention were phoned weekly for 10 minutes to discuss lifestyle and environmental factors that may influence depression. A different topic was discussed each week; 1) physical and artistic activities, 2) education and hobbies, 3) social financial and family roles, 4) work habits and stress, 5) physical health, medication and pain, and 6) nutrition and alcohol. At the end of the intervention period and at six months all participants were sent a questionnaire. The study was approved by the Human Ethics Committee of the Australian National University. Participants provided written informed consent.</p></sec><sec><title>Measures</title><sec><title>Demographic</title><p>Age, sex, marital status, years of education and previous history of depression were assessed.</p></sec><sec><title>Depression</title><p>Severity of depression at pre and post intervention was assessed using the 20 item Centre for Epidemiological Studies Depression Scale (CES-D) scale (range 0 to 60) [<xref ref-type="bibr" rid="B14">14</xref>]. Higher scores represent greater psychological distress with scores 16 or higher indicating clinical depression [<xref ref-type="bibr" rid="B14">14</xref>]. To compare to Jorm et al., [<xref ref-type="bibr" rid="B6">6</xref>] CES-D depression scores at pre-intervention were then grouped into five intervals 'Low' (0–9), 'Mild' (10–19), 'Moderate' (20–29), 'High' (30–39) and 'Severe' (40–60).</p></sec><sec><title>Help-seeking</title><p>(i) <italic>Treatments sought by participants</italic>. At post intervention and at six months respondents were asked 'We would now like you to tell us which of the following treatments or activities (if any) you have used in the past 2 months to cope with depression.' Respondents were then given a list of items and answered either 'yes' or 'no' to each item. This list of self-help and standard professional therapies was compiled from a systematic review of treatments for depression [<xref ref-type="bibr" rid="B8">8</xref>,<xref ref-type="bibr" rid="B13">13</xref>]. Only items with more than 10% endorsement rate were analysed. The list of these items contained nine evidence-based treatments: one medical professional treatment (use of anti-depressant medication), two psychological treatments (CBT and self-help books) and six alternative treatments (exercise, yoga, massage, relaxation, cut out alcohol, use of vitamins); and eight non-evidence based treatments: one psychological treatment (counseling), and seven alternative treatments (eating chocolate, music, being with pets, doing enjoyable things, meditation, avoiding caffeine, use of alcohol).</p><p>(ii) <italic>Seeking help from particular sour</italic>ces was also assessed. Participants were asked whether they had sought help from a 'general practitioner', 'counsellor or psychologist' or from 'friends or family'.</p><p>(iii) <italic>Help seeking treatments by category </italic>using the categorization introduced by Jorm et al [<xref ref-type="bibr" rid="B6">6</xref>] were also assessed. The treatments were grouped into four categories: Everyday treatments included family and friends, exercise, eating chocolate, listening to music, being with pets and doing more things you enjoy. Everyday actions were so named because they reflected activities which are not necessarily seen to be therapeutic but which are frequently undertaken or experienced in daily life. Complementary treatments included yoga, massage, relaxation therapy and meditation. Complementary treatments were those identified through a systematic search of the literature, but which were not commonly identified in clinical practice guidelines. Dietary treatments included avoiding caffeine and drinking alcohol. Professional treatments included seeking the help of general practitioners, counselors and psychologists, anti-depressants, cognitive-behavioural therapy, counseling and reading self-help books. A previously used category – non-prescription treatments (which includes painkillers, St John's wort, fish oils, vitamins and 'cut down on alcohol') – was not able to be used in the present analysis because most items were too infrequently endorsed. Two components within this category, 'vitamins' and 'cut out alcohol'- were included in the dietary category. Thus the everyday category had six treatments, the complementary category four treatments, the dietary category four treatments and the professional category six treatments. These categories were originally derived from a principal components analysis of a range of treatments found in the literature [<xref ref-type="bibr" rid="B6">6</xref>].</p></sec></sec><sec><title>Statistical analysis</title><p>The first set of analyses addressed the question of whether the interventions were associated with reports of help seeking. It consisted of a series of separate logistic regressions assessing the impact of BluePages compared to the control condition, and BluePages compared to the MoodGYM site on help seeking for each of the specific treatments used (e.g., antidepressant medication). A second analysis assessed the effect of these the two interventions and the control condition on help seeking from different sources of help (professionals/family). A further set of four linear regressions examined whether those in the BluePages intervention compared to the two other interventions used different categories of treatments.</p><p>BluePages was selected as the comparison condition in preference to the control condition because we hypothesized it would lead to greater help seeking for evidence-based treatments than the other two conditions. Supplementary analyses addressed the question of whether the reported seeking of help from a specific treatment or source was associated with a reduction in depressive symptoms. Separate linear regressions were used to examine the reported use of each specific treatment and its association with depression (CES-D scores) at post intervention and at six months. A further four linear regressions examined whether the category types of treatment predicted depression. A second set of supplementary analyses investigated the relationship between the severity of depression and the reported uptake of specific treatments or sources, controlling for intervention type. A series of separate logistic regressions were used to assess the effect of the severity of depression (CES-D scores) at pre-intervention on the use of each treatment. Four linear regressions were used to examine the impact of depression on the number of every day, complementary, dietary and professional treatments used. No formal adjustment was undertaken for the number of comparisons that were tested. However, the majority of effects reported would remain significant even if correct using a extremely conservative approach such as the Bonferroni method. Nevertheless, the possibility of inflated Type I error arising from evaluating multiple treatments should be borne in mind.</p></sec></sec><sec><title>Results</title><sec><title>Demographic characteristics</title><p>There were no significant differences in age, sex and CES-D score between the intervention conditions at baseline [<xref ref-type="bibr" rid="B9">9</xref>,<xref ref-type="bibr" rid="B11">11</xref>]. Participants who completed the post-intervention questionnaire did not differ in age, sex or years of education or the CES-D at endpoint in comparison to those who did not return the questionnaire, although the dropouts had higher depression scores at baseline (F (1,523) = 5.15, p = 0.24).</p></sec><sec><title>Effects of the sites on depression symptoms</title><p>Both BluePages and MoodGYM significantly reduced depression symptoms at post intervention relative to the control condition. Scores reduced 4.5 points (1.8 to 7.3) (95% confidence interval) for MoodGYM and 3.6 points (1.0 to 6.3) for BluePages relative to the control condition.</p></sec><sec><title>Effects of the interventions on help seeking</title><p>Frequencies for the reported use of each treatment are shown in Table <xref ref-type="table" rid="T1">1</xref>. The results for the logistic and linear regressions are also displayed. Compared to BluePages, MoodGYM participants were more likely to report using cognitive behavioural therapy, exercise and massage. Both MoodGYM and control participants were more likely to report listening to music than were BluePages participants. Compared to BluePages, the control group was more likely to report seeking the help of family and friends. Both MoodGYM and the control condition were associated with the reported use of more everyday treatments. MoodGYM was also associated with the seeking of more complementary and professional treatments. These analyses were repeated as six months (Table <xref ref-type="table" rid="T2">2</xref>). A number of the differences between the intervention groups were no longer significant. However, MoodGYM participants were still more likely to say they had used cognitive-behavioural therapy than those in BluePages (OR = 2.73; p = 0.011), and to use everyday treatments (B = .48, p = .044). After six months those in the control group were still more likely to have listened to music than those in the BluePages intervention (OR = 2.10, p = .005), and were now more likely to report the use of pets as a form of treatment than those in BluePages (OR = 1.86, p = .018). Both MoodGYM participants and the control group were more likely to use alcohol as a form of treatment at six months, than those in BluePages (OR = 2.26, p = .031, OR = 2.13, p = .037).</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>Treatment use percentages for each condition and summary of regressions predicting treatment use from intervention type</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td align="center">BluePages</td><td align="center">Control</td><td align="center">MoodGym</td><td align="center" colspan="3">Control – BluePages</td><td align="center" colspan="3">MoodGym – BluePages</td></tr></thead><tbody><tr><td align="center">Specific Treatments</td><td align="center">(N = 136)</td><td align="center">(N = 157)</td><td align="center">(N = 121)</td><td align="center">OR</td><td align="center">p</td><td align="center">95%CI</td><td align="center">OR</td><td align="center">p</td><td align="center">95%CI</td></tr><tr><td colspan="10"><hr></hr></td></tr><tr><td align="center"><italic>Evidence based</italic></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">Antidepressants</td><td align="center">27.9</td><td align="center">25.5</td><td align="center">24.8</td><td align="center">.88</td><td align="center">.634</td><td align="center">.53 – 1.48</td><td align="center">.85</td><td align="center">.568</td><td align="center">.49 – 1.48</td></tr><tr><td align="left">CBT</td><td align="center">11.0</td><td align="center">7.6</td><td align="center">28.9</td><td align="center">.67</td><td align="center">.320</td><td align="center">.30 – 1.48</td><td align="center">3.28</td><td align="center">.000**</td><td align="center">1.69 – 6.38</td></tr><tr><td align="left">Self-help books</td><td align="center">23.5</td><td align="center">19.7</td><td align="center">27.3</td><td align="center">.80</td><td align="center">.432</td><td align="center">.46 – 1.40</td><td align="center">1.22</td><td align="center">.491</td><td align="center">.69 – 2.14</td></tr><tr><td align="left">Exercise</td><td align="center">56.6</td><td align="center">63.7</td><td align="center">68.6</td><td align="center">1.34</td><td align="center">.217</td><td align="center">.84 – 2.15</td><td align="center">1.67</td><td align="center">.049*</td><td align="center">1.01 – 2.79</td></tr><tr><td align="left">Yoga</td><td align="center">9.6</td><td align="center">9.6</td><td align="center">13.2</td><td align="center">1.00</td><td align="center">.999</td><td align="center">.46 – 2.18</td><td align="center">1.44</td><td align="center">.356</td><td align="center">.66 – 3.14</td></tr><tr><td align="left">Massage</td><td align="center">13.2</td><td align="center">15.9</td><td align="center">24.0</td><td align="center">1.24</td><td align="center">.517</td><td align="center">.65 – 2.39</td><td align="center">2.07</td><td align="center">.028*</td><td align="center">1.08 – 3.95</td></tr><tr><td align="left">Relaxation therapy</td><td align="center">12.5</td><td align="center">7.6</td><td align="center">18.2</td><td align="center">.58</td><td align="center">.169</td><td align="center">.27 – 1.26</td><td align="center">1.57</td><td align="center">.207</td><td align="center">.78 – 3.09</td></tr><tr><td align="left">Cut out alcohol</td><td align="center">14.0</td><td align="center">19.0</td><td align="center">19.0</td><td align="center">1.22</td><td align="center">.540</td><td align="center">.64 – 2.32</td><td align="center">1.45</td><td align="center">.277</td><td align="center">.74 – 2.81</td></tr><tr><td align="left">Use vitamins</td><td align="center">10.3</td><td align="center">15.9</td><td align="center">14.9</td><td align="center">1.65</td><td align="center">.160</td><td align="center">.82 – 3.32</td><td align="center">1.52</td><td align="center">.269</td><td align="center">.72 – 3.21</td></tr><tr><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left" colspan="10"><italic>Non-evidence based</italic></td></tr><tr><td align="left">Counselling</td><td align="center">8.1</td><td align="center">14.0</td><td align="center">11.6</td><td align="center">1.85</td><td align="center">.114</td><td align="center">.86 – 30.97</td><td align="center">1.49</td><td align="center">.349</td><td align="center">.65 – 30.14</td></tr><tr><td align="left">Eating chocolate</td><td align="center">19.9</td><td align="center">22.9</td><td align="center">27.3</td><td align="center">1.20</td><td align="center">.523</td><td align="center">.69 – 2.11</td><td align="center">1.51</td><td align="center">.162</td><td align="center">.85 – 2.71</td></tr><tr><td align="left">Listening to music</td><td align="center">42.6</td><td align="center">59.9</td><td align="center">56.2</td><td align="center">2.01</td><td align="center">.003**</td><td align="center">1.26 – 3.20</td><td align="center">1.73</td><td align="center">.031*</td><td align="center">1.05 – 2.83</td></tr><tr><td align="left">Being with pets</td><td align="center">41.9</td><td align="center">52.9</td><td align="center">50.4</td><td align="center">1.56</td><td align="center">.062</td><td align="center">.98 – 2.47</td><td align="center">1.41</td><td align="center">.173</td><td align="center">.86 – 2.31</td></tr><tr><td align="left">Doing things you enjoy</td><td align="center">57.4</td><td align="center">60.5</td><td align="center">59.5</td><td align="center">1.14</td><td align="center">.584</td><td align="center">.71 – 1.82</td><td align="center">1.09</td><td align="center">.727</td><td align="center">.67 – 1.80</td></tr><tr><td align="left">Meditation</td><td align="center">13.2</td><td align="center">12.7</td><td align="center">19.0</td><td align="center">.96</td><td align="center">.900</td><td align="center">.48–1.89</td><td align="center">1.54</td><td align="center">.209</td><td align="center">.79–3.01</td></tr><tr><td align="left">Avoiding caffeine</td><td align="center">11.0</td><td align="center">10.2</td><td align="center">11.6</td><td align="center">.92</td><td align="center">.816</td><td align="center">.44–1.93</td><td align="center">1.06</td><td align="center">.891</td><td align="center">.49–2.29</td></tr><tr><td align="left">Use alcohol</td><td align="center">18.4</td><td align="center">23.6</td><td align="center">19.0</td><td align="center">1.37</td><td align="center">.280</td><td align="center">.78-.242</td><td align="center">1.04</td><td align="center">.898</td><td align="center">.56–1.95</td></tr><tr><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left" colspan="10">Sources of Help</td></tr><tr><td align="left">GP</td><td align="center">22.1</td><td align="center">22.3</td><td align="center">28.1</td><td align="center">1.01</td><td align="center">.962</td><td align="center">.58 – 1.76</td><td align="center">1.38</td><td align="center">.265</td><td align="center">.78 – 2.43</td></tr><tr><td align="left">Counsellor/psychologist</td><td align="center">10.3</td><td align="center">15.9</td><td align="center">18.2</td><td align="center">1.65</td><td align="center">.160</td><td align="center">.82 – 3.32</td><td align="center">1.93</td><td align="center">.072</td><td align="center">.94 – 3.98</td></tr><tr><td align="left">Family and friends</td><td align="center">56.6</td><td align="center">71.3</td><td align="center">64.5</td><td align="center">1.91</td><td align="center">.009**</td><td align="center">1.18 – 3.10</td><td align="center">1.39</td><td align="center">.200</td><td align="center">.84 – 2.30</td></tr><tr><td colspan="10"><hr></hr></td></tr><tr><td align="left">Categories of Treatments</td><td align="center">Mean (SD)</td><td align="center">Mean (SD)</td><td align="center">Mean (SD)</td><td align="center">B</td><td align="center">p</td><td align="center">95% CI</td><td align="center">B</td><td align="center">p</td><td align="center">95% CI</td></tr><tr><td colspan="10"><hr></hr></td></tr><tr><td></td><td colspan="1"><hr></hr></td><td colspan="1"><hr></hr></td><td colspan="1"><hr></hr></td><td colspan="1"><hr></hr></td><td colspan="1"><hr></hr></td><td colspan="1"><hr></hr></td><td colspan="1"><hr></hr></td><td colspan="1"><hr></hr></td><td colspan="1"><hr></hr></td></tr><tr><td align="left">Professional treatments (Scale: 0–6)</td><td align="center">1.03 (1.25)</td><td align="center">1.05 (1.37)</td><td align="center">1.39 (1.47)</td><td align="center">.16</td><td align="center">.893</td><td align="center">-.29 – .43</td><td align="center">.36</td><td align="center">.036*</td><td align="center">.02 – .69</td></tr><tr><td align="left">Everyday treatments (Scale: 0–4)</td><td align="center">2.75 (1.65)</td><td align="center">3.31 (1.68)</td><td align="center">3.26 (1.60)</td><td align="center">.19</td><td align="center">.004**</td><td align="center">.18 – .94</td><td align="center">.51</td><td align="center">.013*</td><td align="center">.11 – .92</td></tr><tr><td align="left">Complementary treatments (Scale: 0–4)</td><td align="center">0.49 (0.81)</td><td align="center">0.46 (0.77)</td><td align="center">0.74 (0.94)</td><td align="center">.10</td><td align="center">.786</td><td align="center">-.22 – .17</td><td align="center">.26</td><td align="center">.014*</td><td align="center">.05 – .46</td></tr><tr><td align="left">Dietary treatments (Scale: 0–6)</td><td align="center">0.54 (0.78)</td><td align="center">0.66 (0.83)</td><td align="center">0.64 (0.89)</td><td align="center">.10</td><td align="center">.198</td><td align="center">-.07 – .32</td><td align="center">.11</td><td align="center">.360</td><td align="center">-.10 – .31</td></tr></tbody></table><table-wrap-foot><p>Notes: Professional treatments included; GP, Counsellors and psychologists, anti-depressants, cognitive-behavioural therapy, counselling and reading self-help books. Everyday treatments included; family and friends, exercise, eating chocolate, listening to music, being with pets and doing more things you enjoy. Complimentary treatments included; yoga, massage, relaxation therapy and meditation. Dietary treatments included; cutting out alcohol, taking vitamins, avoiding caffeine and drinking alcohol.</p></table-wrap-foot></table-wrap><table-wrap position="float" id="T2"><label>Table 2</label><caption><p>Treatment use percentages at six month follow-up for each condition and summary of regressions predicting treatment use from intervention type</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td align="center">BluePages</td><td align="center">Control</td><td align="center">MoodGym</td><td align="center" colspan="3">Control – BluePages</td><td align="center" colspan="3">MoodGym – BluePages</td></tr></thead><tbody><tr><td align="center">Specific Treatments</td><td align="center">(N = 114)</td><td align="center">(N = 130)</td><td align="center">(N = 102)</td><td align="center">OR</td><td align="center">p</td><td align="center">95%CI</td><td align="center">OR</td><td align="center">p</td><td align="center">95%CI</td></tr><tr><td colspan="10"><hr></hr></td></tr><tr><td align="center"><italic>Evidence based</italic></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">Antidepressants</td><td align="center">25.4</td><td align="center">26.2</td><td align="center">28.4</td><td align="center">1.04</td><td align="center">.899</td><td align="center">.58–1.85</td><td align="center">1.16</td><td align="center">.620</td><td align="center">.64–2.13</td></tr><tr><td align="left">CBT</td><td align="center">9.6</td><td align="center">6.9</td><td align="center">22.5</td><td align="center">.70</td><td align="center">.441</td><td align="center">.28–1.75</td><td align="center">2.73</td><td align="center">.011*</td><td align="center">1.26–5.92</td></tr><tr><td align="left">Self-help books</td><td align="center">24.6</td><td align="center">22.3</td><td align="center">15.7</td><td align="center">.88</td><td align="center">.678</td><td align="center">.49–1.60</td><td align="center">.57</td><td align="center">.108</td><td align="center">.29–1.13</td></tr><tr><td align="left">Exercise</td><td align="center">54.4</td><td align="center">56.9</td><td align="center">64.7</td><td align="center">1.11</td><td align="center">.691</td><td align="center">.67–1.84</td><td align="center">1.54</td><td align="center">.124</td><td align="center">.90–2.66</td></tr><tr><td align="left">Yoga</td><td align="center">13.2</td><td align="center">10.8</td><td align="center">8.8</td><td align="center">.80</td><td align="center">.566</td><td align="center">.37–1.73</td><td align="center">.64</td><td align="center">.314</td><td align="center">.27–1.53</td></tr><tr><td align="left">Massage</td><td align="center">16.7</td><td align="center">15.4</td><td align="center">18.6</td><td align="center">.91</td><td align="center">.785</td><td align="center">.46–1.80</td><td align="center">1.15</td><td align="center">.706</td><td align="center">.57–2.31</td></tr><tr><td align="left">Relaxation therapy</td><td align="center">12.3</td><td align="center">6.2</td><td align="center">10.8</td><td align="center">.47</td><td align="center">.102</td><td align="center">.19–1.16</td><td align="center">.86</td><td align="center">.732</td><td align="center">.37–2.00</td></tr><tr><td align="left">Cut out alcohol</td><td align="center">13.2</td><td align="center">15.4</td><td align="center">15.7</td><td align="center">1.20</td><td align="center">.621</td><td align="center">.58–2.47</td><td align="center">1.23</td><td align="center">.597</td><td align="center">.57–2.63</td></tr><tr><td align="left">Use vitamins</td><td align="center">11.4</td><td align="center">15.4</td><td align="center">12.7</td><td align="center">1.41</td><td align="center">.366</td><td align="center">.67–2.99</td><td align="center">1.14</td><td align="center">.762</td><td align="center">.50–2.58</td></tr><tr><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left" colspan="10"><italic>Non-evidence based</italic></td></tr><tr><td align="left">Counselling</td><td align="center">11.4</td><td align="center">12.3</td><td align="center">11.8</td><td align="center">1.09</td><td align="center">.828</td><td align="center">.50–2.38</td><td align="center">1.04</td><td align="center">.934</td><td align="center">.45–2.39</td></tr><tr><td align="left">Eating chocolate</td><td align="center">21.1</td><td align="center">23.8</td><td align="center">30.4</td><td align="center">1.17</td><td align="center">.60</td><td align="center">.64–2.15</td><td align="center">1.64</td><td align="center">.117</td><td align="center">.88–3.04</td></tr><tr><td align="left">Listening to music</td><td align="center">47.4</td><td align="center">65.4</td><td align="center">53.9</td><td align="center">2.10</td><td align="center">.005**</td><td align="center">1.25–3.51</td><td align="center">1.300</td><td align="center">.337</td><td align="center">.76–2.22</td></tr><tr><td align="left">Being with pets</td><td align="center">38.6</td><td align="center">53.8</td><td align="center">45.1</td><td align="center">1.86</td><td align="center">.018*</td><td align="center">1.11–3.09</td><td align="center">1.31</td><td align="center">.334</td><td align="center">.76–2.25</td></tr><tr><td align="left">Doing things you enjoy</td><td align="center">56.1</td><td align="center">55.4</td><td align="center">67.6</td><td align="center">.94</td><td align="center">.906</td><td align="center">.58–1.61</td><td align="center">1.63</td><td align="center">.084</td><td align="center">.94–2.85</td></tr><tr><td align="left">Meditation</td><td align="center">12.3</td><td align="center">8.5</td><td align="center">16.7</td><td align="center">.66</td><td align="center">.329</td><td align="center">.29–1.52</td><td align="center">1.43</td><td align="center">.360</td><td align="center">.67–3.07</td></tr><tr><td align="left">Avoiding caffeine</td><td align="center">8.8</td><td align="center">12.3</td><td align="center">8.8</td><td align="center">1.46</td><td align="center">.374</td><td align="center">.63–3.36</td><td align="center">1.01</td><td align="center">.989</td><td align="center">.39–2.58</td></tr><tr><td align="left">Use alcohol</td><td align="center">11.4</td><td align="center">21.5</td><td align="center">22.5</td><td align="center">2.13</td><td align="center">.037*</td><td align="center">1.05–4.35</td><td align="center">2.26</td><td align="center">.031*</td><td align="center">1.08–4.75</td></tr><tr><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left" colspan="10">Sources of Help</td></tr><tr><td align="left">GP</td><td align="center">26.3</td><td align="center">21.5</td><td align="center">29.4</td><td align="center">.77</td><td align="center">.382</td><td align="center">.43–1.39</td><td align="center">1.17</td><td align="center">.612</td><td align="center">.64–2.12</td></tr><tr><td align="left">Counsellor/psychologist</td><td align="center">11.4</td><td align="center">16.9</td><td align="center">17.6</td><td align="center">1.58</td><td align="center">.222</td><td align="center">.76–3.31</td><td align="center">1.67</td><td align="center">.194</td><td align="center">.77–3.60</td></tr><tr><td align="left">Family and friends</td><td></td><td></td><td></td><td align="center">.92</td><td align="center">.757</td><td align="center">.55–1.54</td><td align="center">1.62</td><td align="center">.094</td><td align="center">.92–2.86</td></tr><tr><td colspan="10"><hr></hr></td></tr><tr><td align="left">Categories of Treatments</td><td align="center">Mean (SD)</td><td align="center">Mean (SD)</td><td align="center">Mean (SD)</td><td align="center">B</td><td align="center">p</td><td align="center">95% CI</td><td align="center">B</td><td align="center">p</td><td align="center">95% CI</td></tr><tr><td colspan="10"><hr></hr></td></tr><tr><td></td><td colspan="1"><hr></hr></td><td colspan="1"><hr></hr></td><td colspan="1"><hr></hr></td><td colspan="1"><hr></hr></td><td colspan="1"><hr></hr></td><td colspan="1"><hr></hr></td><td colspan="1"><hr></hr></td><td colspan="1"><hr></hr></td><td colspan="1"><hr></hr></td></tr><tr><td align="left">Professional treatments (Scale: 0–6)</td><td align="center">1.09 (1.25)</td><td align="center">1.06 (1.36)</td><td align="center">1.25 (1.49)</td><td align="center">-.03</td><td align="center">.832</td><td align="center">-.34-.27</td><td align="center">.15</td><td align="center">.374</td><td align="center">-.18-.47</td></tr><tr><td align="left">Everyday treatments (Scale: 0–4)</td><td align="center">2.77 (1.66)</td><td align="center">3.13 (1.75)</td><td align="center">3.32 (1.50)</td><td align="center">.27</td><td align="center">.226</td><td align="center">-.17-.71</td><td align="center">.48</td><td align="center">.044*</td><td align="center">.01-.94</td></tr><tr><td align="left">Complementary treatments (Scale: 0–4)</td><td align="center">.54 (.86)</td><td align="center">.41 (.76)</td><td align="center">.55 (.87)</td><td align="center">-.12</td><td align="center">.197</td><td align="center">-.30-.06</td><td align="center">.01</td><td align="center">.943</td><td align="center">-.19-.20</td></tr><tr><td align="left">Dietary treatments (Scale: 0–6)</td><td align="center">.45 (.70)</td><td align="center">.65 (.87)</td><td align="center">.60 (.73)</td><td align="center">.16</td><td align="center">.064</td><td align="center">-.01-.33</td><td align="center">.13</td><td align="center">.162</td><td align="center">-.05-.31</td></tr></tbody></table><table-wrap-foot><p>Notes: Professional treatments included; GP, Counsellors and psychologists, anti-depressants, cognitive-behavioural therapy, counselling and reading self-help books. Everyday treatments included; family and friends, exercise, eating chocolate, listening to music, being with pets and doing more things you enjoy. Complimentary treatments included; yoga, massage, relaxation therapy and meditation. Dietary treatments included; cutting out alcohol, taking vitamins, avoiding caffeine and drinking alcohol.</p></table-wrap-foot></table-wrap></sec><sec><title>Associations between actions taken and symptom reduction</title><p>Table <xref ref-type="table" rid="T3">3</xref> shows the mean depression scores at post intervention for the users and non-users of each individual treatment. These analyses include individuals from the three intervention conditions combined. Table <xref ref-type="table" rid="T3">3</xref> also lists the results of the linear regressions used to assess the effect of the reported use of each specific treatment, sources of help and treatment categories on depression. Intervention condition and depression at pre-test were controlled for in these analyses. Users of relaxation and those reporting "doing things that they enjoyed' were less depressed. Use of alcohol and the number of professional treatments used were associated with higher levels of depression. These analyses were repeated at six months. There were no major differences in findings at 6 months compared to immediately post-intervention for most of the relationships (Table <xref ref-type="table" rid="T4">4</xref>). However, the number of professional treatments used, doing enjoyable things and drinking alcohol, were not longer significant predictors of CES-D scores after six months. Relaxation therapy continued to be associated with lower CES-D scores (B = -.378, p = 0.015). New associations after six months were that CBT was now significantly associated with higher CES-D scores (B = 3.61, p = 0.015), seeing a GP was significantly associated with lower CES-D scores (B = -.2.77, p = .026), and being with pets was associated with lower CES-D scores (B = -3.19, p = .002).</p><table-wrap position="float" id="T3"><label>Table 3</label><caption><p>CES-D scores at post intervention as a function of treatment use and summary of regressions predicting CES-D scores at post intervention</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td align="center" colspan="3">Treatment Users</td><td align="center" colspan="3">Treatment Non Users</td><td align="center" colspan="4">Regression Statistics</td></tr></thead><tbody><tr><td align="center">Specific Treatments</td><td align="center">Mean<sup>b</sup></td><td align="center">SE</td><td align="center">N</td><td align="center">Mean<sup>b</sup></td><td align="center">SE</td><td align="center">N</td><td align="center">B</td><td align="center">SE</td><td align="center">p</td><td align="center">95%CI</td></tr><tr><td colspan="11"><hr></hr></td></tr><tr><td align="center"><italic>Evidence Based</italic></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">Antidepressants</td><td align="center">18.46</td><td align="center">.86</td><td align="center">108</td><td align="center">16.90</td><td align="center">.50</td><td align="center">306</td><td align="center">1.58</td><td align="center">.98</td><td align="center">.106</td><td align="center">-.34 – 3.50</td></tr><tr><td align="left">CBT</td><td align="center">17.99</td><td align="center">1.12</td><td align="center">62</td><td align="center">17.18</td><td align="center">.47</td><td align="center">352</td><td align="center">1.97</td><td align="center">1.23</td><td align="center">.108</td><td align="center">-.44 – 4.38</td></tr><tr><td align="left">Self-help books</td><td align="center">17.25</td><td align="center">.91</td><td align="center">96</td><td align="center">17.32</td><td align="center">.50</td><td align="center">318</td><td align="center">.24</td><td align="center">1.02</td><td align="center">.812</td><td align="center">-1.77 – 2.25</td></tr><tr><td align="left">Exercise</td><td align="center">17.30</td><td align="center">.55</td><td align="center">260</td><td align="center">17.31</td><td align="center">.71</td><td align="center">154</td><td align="center">.00</td><td align="center">.88</td><td align="center">.998</td><td align="center">-1.73 – 1.73</td></tr><tr><td align="left">Yoga</td><td align="center">17.48</td><td align="center">1.33</td><td align="center">44</td><td align="center">17.28</td><td align="center">.46</td><td align="center">370</td><td align="center">.41</td><td align="center">1.38</td><td align="center">.766</td><td align="center">-2.29 – 3.11</td></tr><tr><td align="left">Massage</td><td align="center">16.19</td><td align="center">1.04</td><td align="center">72</td><td align="center">17.54</td><td align="center">.48</td><td align="center">342</td><td align="center">-1.12</td><td align="center">1.12</td><td align="center">.997</td><td align="center">-3.33 – 1.09</td></tr><tr><td align="left">Relaxation therapy</td><td align="center">13.38</td><td align="center">1.22</td><td align="center">51</td><td align="center">17.86</td><td align="center">.46</td><td align="center">356</td><td align="center">-3.83</td><td align="center">1.29</td><td align="center">.003*</td><td align="center">-6.35 – -.30</td></tr><tr><td align="left">Cut out alcohol</td><td align="center">17.68</td><td align="center">1.07</td><td align="center">68</td><td align="center">17.23</td><td align="center">.47</td><td align="center">346</td><td align="center">.48</td><td align="center">1.15</td><td align="center">.677</td><td align="center">-1.78 – 2.73</td></tr><tr><td align="left">Use vitamins</td><td align="center">16.57</td><td align="center">1.17</td><td align="center">57</td><td align="center">17.42</td><td align="center">.47</td><td align="center">357</td><td align="center">-1.10</td><td align="center">1.24</td><td align="center">.378</td><td align="center">-3.53 – 1.34</td></tr><tr><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left" colspan="11"><italic>Non-evidence based</italic></td></tr><tr><td align="left">Counselling</td><td align="center">19.54</td><td align="center">1.28</td><td align="center">47</td><td align="center">17.02</td><td align="center">.46</td><td align="center">367</td><td align="center">2.17</td><td align="center">1.34</td><td align="center">.105</td><td align="center">-.46 – 4.81</td></tr><tr><td align="left">Eating chocolate</td><td align="center">16.91</td><td align="center">.90</td><td align="center">96</td><td align="center">17.42</td><td align="center">.49</td><td align="center">318</td><td align="center">-.45</td><td align="center">1.01</td><td align="center">.659</td><td align="center">-2.43 – 1.54</td></tr><tr><td align="left">Listening to music</td><td align="center">17.61</td><td align="center">.59</td><td align="center">220</td><td align="center">16.95</td><td align="center">.63</td><td align="center">194</td><td align="center">.33</td><td align="center">.86</td><td align="center">.703</td><td align="center">-1.36 – 2.02</td></tr><tr><td align="left">Being with pets</td><td align="center">16.73</td><td align="center">.62</td><td align="center">201</td><td align="center">17.85</td><td align="center">.60</td><td align="center">213</td><td align="center">-1.35</td><td align="center">.85</td><td align="center">.114</td><td align="center">-3.02 – .32</td></tr><tr><td align="left">Doing things you enjoy</td><td align="center">15.72</td><td align="center">.55</td><td align="center">245</td><td align="center">19.6</td><td align="center">.66</td><td align="center">169</td><td align="center">-3.95</td><td align="center">.84</td><td align="center">.000**</td><td align="center">-5.60 – -2.31</td></tr><tr><td align="left">Meditation</td><td align="center">16.42</td><td align="center">1.13</td><td align="center">61</td><td align="center">17.46</td><td align="center">.47</td><td align="center">353</td><td align="center">-.75</td><td align="center">1.20</td><td align="center">.531</td><td align="center">-3.10 – 1.60</td></tr><tr><td align="left">Avoiding caffeine</td><td align="center">15.88</td><td align="center">1.31</td><td align="center">45</td><td align="center">17.48</td><td align="center">.46</td><td align="center">369</td><td align="center">-1.49</td><td align="center">1.36</td><td align="center">.273</td><td align="center">-4.16 – 1.18</td></tr><tr><td align="left">Use alcohol</td><td align="center">19.35</td><td align="center">.96</td><td align="center">85</td><td align="center">16.78</td><td align="center">.48</td><td align="center">329</td><td align="center">2.29</td><td align="center">1.05</td><td align="center">.030*</td><td align="center">.22 – 4.37</td></tr><tr><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left" colspan="11">Sources of Help</td></tr><tr><td align="left">GP</td><td align="center">18.20</td><td align="center">.90</td><td align="center">99</td><td align="center">17.02</td><td align="center">.50</td><td align="center">315</td><td align="center">1.35</td><td align="center">1.01</td><td align="center">.183</td><td align="center">-.64 – 3.33</td></tr><tr><td align="left">Counsellor/psychologist</td><td align="center">19.14</td><td align="center">1.13</td><td align="center">61</td><td align="center">16.07</td><td align="center">.47</td><td align="center">353</td><td align="center">2.09</td><td align="center">1.20</td><td align="center">.083</td><td align="center">-.28 – 4.46</td></tr><tr><td align="left">Family and friends</td><td align="center">17.83</td><td align="center">.54</td><td align="center">267</td><td align="center">16.35</td><td align="center">.72</td><td align="center">147</td><td align="center">1.10</td><td align="center">.89</td><td align="center">.281</td><td align="center">-.65 – 2.85</td></tr><tr><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">Categories of Treatments</td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">Professional treatments (0–6)</td><td></td><td></td><td></td><td></td><td></td><td></td><td align="center">.74</td><td align="center">.32</td><td align="center">.021*</td><td align="center">.11 – 1.37</td></tr><tr><td align="left">Everyday treatments (0–4)</td><td></td><td></td><td></td><td></td><td></td><td></td><td align="center">-.39</td><td align="center">.26</td><td align="center">.133</td><td align="center">-.87 – .12</td></tr><tr><td align="left">Complementary treatments (0–4)</td><td></td><td></td><td></td><td></td><td></td><td></td><td align="center">-.89</td><td align="center">.51</td><td align="center">.079</td><td align="center">-1.88 – .10</td></tr><tr><td align="left">Dietary treatments (0–6)</td><td></td><td></td><td></td><td></td><td></td><td></td><td align="center">.24</td><td align="center">.52</td><td align="center">.646</td><td align="center">-.78 – 1.23</td></tr></tbody></table><table-wrap-foot><p>Notes: a: Linear regression was used for all analyses adjusting for condition and pre-intervention depression.</p><p>b: Estimated marginal means adjusted for mean CESD at pre-intervention (21.05).</p></table-wrap-foot></table-wrap><table-wrap position="float" id="T4"><label>Table 4</label><caption><p>CES-D scores at six month follow-up as a function of treatment use and summary of regressions predicting CES-D at six month follow-up</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td align="center" colspan="3">Treatment Users</td><td align="center" colspan="3">Treatment Non Users</td><td align="center" colspan="4">Regression Statistics</td></tr></thead><tbody><tr><td align="center">Specific Treatments</td><td align="center">Mean<sup>b</sup></td><td align="center">SE</td><td align="center">N</td><td align="center">Mean<sup>b</sup></td><td align="center">SE</td><td align="center">N</td><td align="center">B</td><td align="center">SE</td><td align="center">p</td><td align="center">95%CI</td></tr><tr><td colspan="11"><hr></hr></td></tr><tr><td align="center"><italic>Evidence Based</italic></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">Antidepressants</td><td align="center">16.65</td><td align="center">1.03</td><td align="center">93</td><td align="center">16.10</td><td align="center">.62</td><td align="center">253</td><td align="center">.50</td><td align="center">1.20</td><td align="center">.679</td><td align="center">-1.86–2.85</td></tr><tr><td align="left">CBT</td><td align="center">18.62</td><td align="center">1.32</td><td align="center">55</td><td align="center">15.80</td><td align="center">.57</td><td align="center">291</td><td align="center">3.61</td><td align="center">1.48</td><td align="center">.015*</td><td align="center">.71–6.51</td></tr><tr><td align="left">Self-help books</td><td align="center">15.29</td><td align="center">1.10</td><td align="center">80</td><td align="center">16.54</td><td align="center">.60</td><td align="center">302</td><td align="center">-1.11</td><td align="center">1.26</td><td align="center">.377</td><td align="center">-3.59–1.36</td></tr><tr><td align="left">Exercise</td><td align="center">16.28</td><td align="center">.66</td><td align="center">223</td><td align="center">16.20</td><td align="center">.89</td><td align="center">123</td><td align="center">.13</td><td align="center">1.11</td><td align="center">.907</td><td align="center">-2.05–2.31</td></tr><tr><td align="left">Yoga</td><td align="center">15.37</td><td align="center">1.61</td><td align="center">37</td><td align="center">16.35</td><td align="center">.56</td><td align="center">309</td><td align="center">-.82</td><td align="center">1.70</td><td align="center">.630</td><td align="center">-4.17–2.52</td></tr><tr><td align="left">Massage</td><td align="center">14.54</td><td align="center">1.29</td><td align="center">58</td><td align="center">16.59</td><td align="center">.58</td><td align="center">288</td><td align="center">-1.80</td><td align="center">1.42</td><td align="center">.206</td><td align="center">-4.58-.99</td></tr><tr><td align="left">Relaxation therapy</td><td align="center">12.59</td><td align="center">1.43</td><td align="center">46</td><td align="center">16.81</td><td align="center">.56</td><td align="center">300</td><td align="center">-3.78</td><td align="center">1.55</td><td align="center">.015*</td><td align="center">-6.82–-.74</td></tr><tr><td align="left">Cut out alcohol</td><td align="center">15.95</td><td align="center">1.30</td><td align="center">57</td><td align="center">16.31</td><td align="center">.58</td><td align="center">289</td><td align="center">-.27</td><td align="center">1.42</td><td align="center">.848</td><td align="center">-3.07–2.52</td></tr><tr><td align="left">Use vitamins</td><td align="center">13.91</td><td align="center">1.46</td><td align="center">46</td><td align="center">16.61</td><td align="center">.57</td><td align="center">300</td><td align="center">-2.75</td><td align="center">1.56</td><td align="center">.079</td><td align="center">-5.82-.32</td></tr><tr><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left" colspan="11"><italic>Non-evidence based</italic></td></tr><tr><td align="left">Counselling</td><td align="center">16.22</td><td align="center">1.48</td><td align="center">44</td><td align="center">16.25</td><td align="center">.57</td><td align="center">302</td><td align="center">-.35</td><td align="center">1.59</td><td align="center">.828</td><td align="center">-3.47–2.78</td></tr><tr><td align="left">Eating chocolate</td><td align="center">15.75</td><td align="center">1.07</td><td align="center">85</td><td align="center">16.41</td><td align="center">.61</td><td align="center">261</td><td align="center">-.71</td><td align="center">1.23</td><td align="center">.565</td><td align="center">-3.13–1.71</td></tr><tr><td align="left">Listening to music</td><td align="center">16.09</td><td align="center">.72</td><td align="center">187</td><td align="center">16.44</td><td align="center">.78</td><td align="center">159</td><td align="center">-.63</td><td align="center">1.07</td><td align="center">.554</td><td align="center">-2.73–1.47</td></tr><tr><td align="left">Being with pets</td><td align="center">14.80</td><td align="center">.74</td><td align="center">172</td><td align="center">17.68</td><td align="center">.74</td><td align="center">174</td><td align="center">-3.19</td><td align="center">1.05</td><td align="center">.002**</td><td align="center">-5.25–-1.14</td></tr><tr><td align="left">Doing things you enjoy</td><td align="center">15.82</td><td align="center">.67</td><td align="center">211</td><td align="center">16.93</td><td align="center">.84</td><td align="center">135</td><td align="center">-1.13</td><td align="center">1.07</td><td align="center">.292</td><td align="center">-3.24-.98</td></tr><tr><td align="left">Meditation</td><td align="center">15.51</td><td align="center">1.39</td><td align="center">50</td><td align="center">16.23</td><td align="center">.57</td><td align="center">296</td><td align="center">-.26</td><td align="center">1.50</td><td align="center">.863</td><td align="center">-3.21–2.69</td></tr><tr><td align="left">Avoiding caffeine</td><td align="center">14.03</td><td align="center">1.59</td><td align="center">38</td><td align="center">16.52</td><td align="center">.56</td><td align="center">308</td><td align="center">-2.47</td><td align="center">1.67</td><td align="center">.140</td><td align="center">-5.76-.82</td></tr><tr><td align="left">Use alcohol</td><td align="center">16.50</td><td align="center">1.15</td><td align="center">74</td><td align="center">16.18</td><td align="center">.60</td><td align="center">272</td><td align="center">.063</td><td align="center">1.30</td><td align="center">.961</td><td align="center">-2.49–2.61</td></tr><tr><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left" colspan="11">Sources of Help</td></tr><tr><td align="left">GP</td><td align="center">14.17</td><td align="center">1.08</td><td align="center">84</td><td align="center">16.91</td><td align="center">.60</td><td align="center">262</td><td align="center">-2.77</td><td align="center">1.24</td><td align="center">.026*</td><td align="center">-5.20–-.34</td></tr><tr><td align="left">Counsellor/psychologist</td><td align="center">15.56</td><td align="center">1.37</td><td align="center">52</td><td align="center">16.32</td><td align="center">.57</td><td align="center">294</td><td align="center">-.68</td><td align="center">1.49</td><td align="center">.650</td><td align="center">-3.60–2.25</td></tr><tr><td align="left">Family and friends</td><td align="center">16.59</td><td align="center">.65</td><td align="center">227</td><td align="center">15360</td><td align="center">.90</td><td align="center">119</td><td align="center">.81</td><td align="center">1.11</td><td align="center">.465</td><td align="center">-1.37–3.00</td></tr><tr><td align="left">Categories of Treatments</td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">Professional treatments (0–6)</td><td></td><td></td><td></td><td></td><td></td><td></td><td align="center">-.15</td><td align="center">.39</td><td align="center">.707</td><td align="center">-.92-.63</td></tr><tr><td align="left">Everyday treatments (0–4)</td><td></td><td></td><td></td><td></td><td></td><td></td><td align="center">-.44</td><td align="center">.33</td><td align="center">.179</td><td align="center">-1.09-.20</td></tr><tr><td align="left">Complementary treatments (0–4)</td><td></td><td></td><td></td><td></td><td></td><td></td><td align="center">-1.13</td><td align="center">.63</td><td align="center">.073</td><td align="center">-2.36-.11</td></tr><tr><td align="left">Dietary treatments (0–6)</td><td></td><td></td><td></td><td></td><td></td><td></td><td align="center">-.85</td><td align="center">.61</td><td align="center">.181</td><td align="center">-2.10-.40</td></tr></tbody></table><table-wrap-foot><p>Notes: a: Linear regression was used for all analyses adjusting for condition and pre-intervention depression.</p><p>b: Estimated marginal means adjusted for mean CESD at pre-intervention (20.86).</p></table-wrap-foot></table-wrap></sec><sec><title>Severity of depression and its impact on actions taken</title><p>Table <xref ref-type="table" rid="T5">5</xref> shows the mean depression scores at pre-intervention for the users and non-users of each treatment, including non-evidence based treatments. Table <xref ref-type="table" rid="T5">5</xref> lists the results for the logistic and linear regressions used to assess the effect of depression severity at pre-intervention on the likelihood of using each individual treatment, and on the number of every day, complementary, dietary and professional treatments used. Frequently adopted actions were seeking out family and friends, exercise, listening to music, engaging in pleasant events and being with pets. Taking antidepressants, reading self-help books for depression, using vitamins and using alcohol were associated with higher pretest depression scores. Seeking help from counsellors/psychologists and general practitioners was also associated with higher baseline depression scores, as was the total number of professional treatments. Following Jorm et al. [<xref ref-type="bibr" rid="B6">6</xref>] the mean number of treatments used pre-intervention in each category was examined across depression severity (Figure <xref ref-type="fig" rid="F1">1</xref>). While there was little change in the use of every day, complementary and dietary treatments across severity, the number of professional treatments used increased with increasing severity.</p><table-wrap position="float" id="T5"><label>Table 5</label><caption><p>CES-D scores at pretest as a function of treatment use and summary of regressions predicting treatment use from CES-D at pre-intervention</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left">Specific Treatments<sup>a</sup></td><td align="center">Mean<sup>b</sup></td><td align="center">SE</td><td align="center">N</td><td align="center">Mean<sup>b</sup></td><td align="center">SE</td><td align="center">N</td><td align="center">OR</td><td align="center">p</td><td align="center">95% CI</td></tr></thead><tbody><tr><td align="center"><italic>Evidence Based</italic></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">Antidepressants</td><td align="center">24.08</td><td align="center">11.25</td><td align="center">108</td><td align="center">19.97</td><td align="center">9.35</td><td align="center">306</td><td align="center">1.04</td><td align="center">.000**</td><td align="center">1.02–1.07</td></tr><tr><td align="left">VBT</td><td align="center">22.94</td><td align="center">10.09</td><td align="center">62</td><td align="center">20.71</td><td align="center">10.00</td><td align="center">352</td><td align="center">1.02</td><td align="center">.108</td><td align="center">1.00–1.05</td></tr><tr><td align="left">Self-help books</td><td align="center">24.44</td><td align="center">9.40</td><td align="center">96</td><td align="center">20.02</td><td align="center">10.00</td><td align="center">318</td><td align="center">1.05</td><td align="center">.000**</td><td align="center">1.02–1.07</td></tr><tr><td align="left">Exercise</td><td align="center">21.27</td><td align="center">9.83</td><td align="center">260</td><td align="center">20.66</td><td align="center">10.37</td><td align="center">154</td><td align="center">1.01</td><td align="center">.557</td><td align="center">.99–1.03</td></tr><tr><td align="left">Yoga</td><td align="center">22.11</td><td align="center">10.15</td><td align="center">44</td><td align="center">20.92</td><td align="center">10.02</td><td align="center">370</td><td align="center">1.01</td><td align="center">.462</td><td align="center">.98–1.04</td></tr><tr><td align="left">Massage</td><td align="center">21.33</td><td align="center">9.21</td><td align="center">72</td><td align="center">20.99</td><td align="center">10.21</td><td align="center">342</td><td align="center">1.00</td><td align="center">.808</td><td align="center">.98–1.03</td></tr><tr><td align="left">Relaxation therapy</td><td align="center">21.94</td><td align="center">9.84</td><td align="center">51</td><td align="center">20.92</td><td align="center">10.06</td><td align="center">363</td><td align="center">1.01</td><td align="center">.513</td><td align="center">.98–1.04</td></tr><tr><td align="left">Cut out alcohol</td><td align="center">22.72</td><td align="center">10.20</td><td align="center">68</td><td align="center">20.72</td><td align="center">9.98</td><td align="center">346</td><td align="center">1.02</td><td align="center">.134</td><td align="center">.99–1.05</td></tr><tr><td align="left">Use vitamins</td><td align="center">24.28</td><td align="center">9.57</td><td align="center">57</td><td align="center">20.53</td><td align="center">10.02</td><td align="center">357</td><td align="center">1.04</td><td align="center">.009*</td><td align="center">1.01–1.07</td></tr><tr><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left" colspan="10"><italic>Non-evidence based</italic></td></tr><tr><td align="left">Counselling</td><td align="center">23.49</td><td align="center">9.72</td><td align="center">47</td><td align="center">20.73</td><td align="center">10.04</td><td align="center">367</td><td align="center">1.03</td><td align="center">.077</td><td align="center">1.00–1.06</td></tr><tr><td align="left">Eating chocolate</td><td align="center">22.68</td><td align="center">9.67</td><td align="center">96</td><td align="center">22.68</td><td align="center">10.12</td><td align="center">318</td><td align="center">1.02</td><td align="center">.071</td><td align="center">1.00–1.04</td></tr><tr><td align="left">Listening to music</td><td align="center">21.44</td><td align="center">9.81</td><td align="center">220</td><td align="center">20.60</td><td align="center">10.29</td><td align="center">194</td><td align="center">1.01</td><td align="center">.389</td><td align="center">.99–1.03</td></tr><tr><td align="left">Being with pets</td><td align="center">21.78</td><td align="center">10.47</td><td align="center">201</td><td align="center">20.36</td><td align="center">9.58</td><td align="center">213</td><td align="center">1.01</td><td align="center">.147</td><td align="center">.10–1.03</td></tr><tr><td align="left">Doing things you enjoy</td><td align="center">21.05</td><td align="center">10.04</td><td align="center">245</td><td align="center">21.04</td><td align="center">10.04</td><td align="center">169</td><td align="center">1.00</td><td align="center">.992</td><td align="center">.98–1.02</td></tr><tr><td align="left">Meditation</td><td align="center">20.62</td><td align="center">10.44</td><td align="center">61</td><td align="center">21.12</td><td align="center">9.97</td><td align="center">353</td><td align="center">1.00</td><td align="center">.702</td><td align="center">.97–1.02</td></tr><tr><td align="left">Avoiding caffeine</td><td align="center">21.96</td><td align="center">9.68</td><td align="center">45</td><td align="center">20.94</td><td align="center">10.08</td><td align="center">369</td><td align="center">1.01</td><td align="center">.522</td><td align="center">.98–1.04</td></tr><tr><td align="left">Use alcohol</td><td align="center">23.85</td><td align="center">9.99</td><td align="center">85</td><td align="center">20.32</td><td align="center">9.93</td><td align="center">329</td><td align="center">1.04</td><td align="center">.004*</td><td align="center">1.01–1.06</td></tr><tr><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left" colspan="10">Sources of Help</td></tr><tr><td align="left">GP</td><td align="center">24.41</td><td align="center">9.96</td><td align="center">99</td><td align="center">19.99</td><td align="center">9.83</td><td align="center">315</td><td align="center">1.05</td><td align="center">.000**</td><td align="center">1.02–1.07</td></tr><tr><td align="left">Counsellor/psychologist</td><td align="center">24.11</td><td align="center">10.79</td><td align="center">61</td><td align="center">20.52</td><td align="center">9.81</td><td align="center">353</td><td align="center">1.01</td><td align="center">.522</td><td align="center">.98–1.04</td></tr><tr><td align="left">Family and friends</td><td align="center">21.42</td><td align="center">10.22</td><td align="center">267</td><td align="center">20.36</td><td align="center">9.67</td><td align="center">147</td><td align="center">1.01</td><td align="center">.289</td><td align="center">.99–1.03</td></tr><tr><td align="left">Categories of Treatments<sup>c</sup></td><td></td><td></td><td></td><td></td><td></td><td></td><td align="center">B</td><td align="center">p</td><td align="center">95% CI</td></tr><tr><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td colspan="3"><hr></hr></td></tr><tr><td align="left">Professional treatments (0–6)</td><td></td><td></td><td></td><td></td><td></td><td></td><td align="center">.03</td><td align="center">.000*</td><td align="center">.02-.05</td></tr><tr><td align="left">Everyday treatments (0–4)</td><td></td><td></td><td></td><td></td><td></td><td></td><td align="center">.01</td><td align="center">.101</td><td align="center">.00-.03</td></tr><tr><td align="left">Complementary treatments (0–4)</td><td></td><td></td><td></td><td></td><td></td><td></td><td align="center">.00</td><td align="center">.640</td><td align="center">-.01-.01</td></tr><tr><td align="left">Dietary treatments (0–6)</td><td></td><td></td><td></td><td></td><td></td><td></td><td align="center">.00</td><td align="center">.187</td><td align="center">.00-.01</td></tr></tbody></table><table-wrap-foot><p>Notes: a: Logistic regressions adjusting for condition.</p><p>b: Estimated marginal means adjusted for mean CESD at pre-intervention (21.05).</p><p>c: Linear regressions adjusting for condition.</p></table-wrap-foot></table-wrap><fig position="float" id="F1"><label>Figure 1</label><caption><p>Mean treatments used across severity of depression at pre-intervention.</p></caption><graphic xlink:href="1471-244X-6-13-1"/></fig></sec></sec><sec><title>Discussion</title><p>The present study found that use of BluePages was associated with different reports of help seeking behaviours relative to the control condition and MoodGYM. For specific treatment types, compared to MoodGYM, Bluepages use was associated with fewer reports of having taken up CBT, exercise, massage or music. Participants also had lower use of professional treatments, and everyday and complementary categories of treatment. BluePages was associated with decreased use of music, decreased help seeking from friends and family and decreased use of everyday treatments compared to the control condition. Specific self reported actions of participants over the trial period were associated with reduced depression symptoms at endpoint for the use of relaxation, and doing more enjoyable things, but not for other evidenced based and non-evidenced based treatments. Taking antidepressants, reading self-help books for depression, using alcohol and seeking professional treatments were associated with higher baseline depression scores.</p><sec><title>Frequency with which treatments were sought and types of treatments sought</title><p>In some respects, these findings are similar to those reported in the recent help seeking literature. First, there is a low level of seeking help from mental health professionals and a preference for everyday treatments, rather than evidence-based treatments. Although figures might be better gauged from representative samples of the population rather than from a sample with high symptoms, only approximately 24% of participants in this study with elevated depression reported seeking the help of a GP. This concords closely with findings by Oliver et al., in the UK that 28% of individuals with high GHQ scores sought help from a general practitioner, and findings in Australia, that 33% of adults with an affective disorder and 56% of those with anxiety disorder did not seek help for their disorder in the last 12 months. A recent study [<xref ref-type="bibr" rid="B1">1</xref>] reported that 49% of individuals with a mood disorder perceived the need for care, but only 13% sought help from a mental health professional. Our findings also accord with others that demonstrate that many individuals seek complementary and non-traditional treatments [<xref ref-type="bibr" rid="B15">15</xref>,<xref ref-type="bibr" rid="B16">16</xref>] rather than treatment through formal mental health services. The most preferred 'treatments' were exercise, doing things you enjoy, listening to music, and being with pets.</p></sec><sec><title>The effect of the specific interventions on reports of help seeking</title><p>Jorm et al., [<xref ref-type="bibr" rid="B7">7</xref>] demonstrated that the provision of an evidence-based guide led to greater attitudinal change than a short brochure, with those reading the guide more likely to affirm that the endorsed treatments are more likely to be helpful. Those receiving the self help guide were more likely to report that they had tried a self help treatment, and to give advice to someone about it. However, the present study has failed to find strong evidence that the information website BluePages led to the reported initiation of evidence-based treatments either during the intervention or at six months. Compared to the control condition, BluePages participants were no more likely to report having sought any evidence-based specific treatment. Nevertheless, depression improved in the BluePages intervention relative to the control condition at post test [<xref ref-type="bibr" rid="B9">9</xref>] and at 6 months. BluePages was associated with a reduction in the use of music, the use of family and friends for support and the use of everyday treatments. We postulated that BluePages might exert its effect through the initiation of self help for evidence based treatments. However, the findings of the study suggest that the help seeking influence of BluePages may be to decrease the use of non-evidence based interventions. For example, seeking support from friends and relatives may be unnecessary, unhelpful, even toxic under certain circumstances and the website may have served to reduce inappropriate help seeking actions. This accords with some adolescent literature that reports that help seeking from younger friends may not be helpful. Moreover, it is possible that BluePages served as a partial substitute for the support and reassurance that might normally be sought from family and friends. The mechanism by which the psychoeducation website exerts it influence on depression symptoms needs further investigation.</p><p>The increase in the reported use of CBT in the MoodGYM condition is to be expected since CBT formed the core of the intervention and participants knew CBT was offered in their intervention. The reasons for the increased reporting of massage and exercise are less clear. Neither exercise nor massage was directly encouraged by the MoodGYM site although one of 29 self help exercises mentioned the use of physical activities as a means of rewarding positive thinking. One explanation is that massage and exercise arose a consequence of a reduction in depression symptoms. More generally, this explanation predicts that help seeking is linked to depression improvement across both MoodGYM and BluePages sites. However, although both MoodGYM and BluePages were equally effective in reducing depression symptoms, only MoodGYM was associated with reported increases in these activities.</p></sec><sec><title>Relationships between help sought and a reduction in depression symptoms</title><p>The present study found that relaxation therapy, doing enjoyable things, and not using alcohol were associated with lower depression symptoms at post test. These relationships were found controlling for both baseline levels of depression and intervention type. Taking specific evidenced based treatments such as anti-depressant medication (reported by one quarter of the participants) did not predict improvement in depression outcomes. Interpretations of these associations must consider the correlational nature of these relationships as well as the complexity of tracking the uptake of a range of treatments. For example, there may be a number of reasons for failing to find a relationship for anti-depressant medication. First, improvement in mood might be associated itself with the discontinuation of depression medication or the choice not to investigate this option. Other possible explanations are that participants might be unlikely to take antidepressants if they have engaged in online treatment, or the antidepressant effect for those already on antidepressants at commencement of the trial may have stablized at the time of enrolment in the trial. Conversely, the anti-depressant effect may be too small to detect over a six week period if this intervention is taken up later in the course of the trial. Finally, the medication may have been discontinued or used inappropriately and the use of medication might be combined with a number of positive or negative other treatments. These findings however, do also accord with recent data from the UK [<xref ref-type="bibr" rid="B3">3</xref>,<xref ref-type="bibr" rid="B17">17</xref>] which show that increases in the use of anti-depressant medication for community depression is not associated with improvement in mental health outcomes. A range of methodological problems may be responsible for the absence of these effects, which require tracking longitudinally over a number of data points.</p><p>The findings are also interesting in view of the recent NICE guidelines for the management of depression in primary and secondary care which promote the use of guided self help, computerized CBT, and exercise in mild depression, but the use of medication and combined treatments in those with treatment resistant or severe depression [<xref ref-type="bibr" rid="B18">18</xref>].</p></sec><sec><title>Depression severity and outcomes</title><p>A set of analyses controlling for treatment intervention examined the effect of depression severity on different types of reported help seeking. These findings accord to some degree with those of Jorm et al., [<xref ref-type="bibr" rid="B6">6</xref>] who reported that some actions such as taking antidepressants and seeing a doctor became increasingly prevalent with greater severity of symptoms while other actions were less prevalent at higher levels of severity, such as undertaking every day activities, for example, engaging in physical activity.</p></sec><sec><title>Limitations of the findings</title><p>There are a number of limitations of the present analysis and design of the study. The major limitation is that the help seeking data are based on self-report. We do not have data on the actual activities undertaken. Secondly, the sample was restricted to that community subgroup of participants who was prepared to participate in a randomised controlled trial of Internet websites. These individuals may have specific preferences for types of treatment not shared by the general community. By virtue of the population from which they were selected, they are likely to be from higher socio-economic groups. Evidence from other work indicates that higher education is associated with the uptake of complementary and alternative treatments [<xref ref-type="bibr" rid="B18">18</xref>]. The screening questionnaire was completed by 22.7% of those approached. This level of response is lower than that achieved by US based surveys using the telephone (40%) [<xref ref-type="bibr" rid="B16">16</xref>] or surveys with reminders or multiple reminders, or those incorporating participants from established cohort studies. Nevertheless, the rate was consistent with earlier postal surveys of mental health using the electoral roll in Canberra [<xref ref-type="bibr" rid="B7">7</xref>]. This lower rate may be due to the lack of personal approach as our broader epidemiological studies of the region yield higher response rates (approximately 60% when telephoned or home visited) [<xref ref-type="bibr" rid="B19">19</xref>]. The analyses investigating depression severity and treatments sought are correlational and may reflect differences in the uptake of treatments as a function of changes in depression rather than the influence of particular interventions on the improvement of depression symptoms. The drop out rate was greater in the MoodGYM condition. The increased use of professional, everyday and complementary treatments in the MoodGYM condition may be due to the additive effect of increased use of a number of the individual specific treatments.</p><p>Despite these shortcomings the present paper has a number of strengths over previous reports of help seeking in mental health. It is one of the first trials to examine the effect of interventions on both evidence-based and non-evidence based help seeking using adequate comparison groups. The significant findings are that exposure to one internet site results in significantly greater increased help seeking of evidence based treatments (CBT, massage and exercise), and an increase in a range of help seeking. Exposure to BluePages reduces what may be unhelpful responses to mental health problems – seeking support from family and friends. It also confirms severity of depression as a risk factor for help seeking, and reveals the complexity of the relationship between self reported help and treatment outcomes.</p></sec></sec><sec><title>Conclusion</title><p>Models of 'pathways to care' are often limited because they fail to account for the large numbers of consumers who do not have access to health care or those who have preferences not to use traditional health services. These models typically consider only the health services that are within the embrace of service providers, excluding self-care and care provided by the not for profit sector and other community agencies. There is clearly the opportunity to pursue broader models of how help seeking may result in better outcomes of those with mental health problems. More needs to be understood about the preferences the community has for various types of treatments (see for example [<xref ref-type="bibr" rid="B20">20</xref>,<xref ref-type="bibr" rid="B21">21</xref>]) the ways in which people with mental health problems seek both traditional and non-traditional help [<xref ref-type="bibr" rid="B10">10</xref>], and the methods which may prove effective in improving the rate and type of help seeking. Further randomised controlled trials are required which examine the effect of various interventions on the actual uptake of help seeking for evidence-based treatments.</p></sec><sec><title>Competing interests</title><p>The author(s) declare they have no competing interests.</p></sec><sec><title>Authors' contributions</title><p>HC and KG conceived and designed the study. HC drafted the article and KG revised it critically for content. LL undertook the data analyses, drafted method and results sections and contributed comments to the final draft. AM critically revised the content of the paper and its analyses, and LB contributed to intellectual content and data interpretation. All gave final approval of the final manuscript to be published.</p></sec><sec><title>Pre-publication history</title><p>The pre-publication history for this paper can be accessed here:</p><p><ext-link ext-link-type="uri" xlink:href="http://www.biomedcentral.com/1471-244X/6/13/prepub"/></p></sec> |
Genetic analysis of male reproductive success in relation to density in the zebrafish, <italic>Danio rerio</italic> | <sec><title>Background</title><p>We used behavioural and genetic data to investigate the effects of density on male reproductive success in the zebrafish, <italic>Danio rerio</italic>. Based on previous measurements of aggression and courtship behaviour by territorial males, we predicted that they would sire more offspring than non-territorial males.</p></sec><sec><title>Results</title><p>Microsatellite analysis of paternity showed that at low densities territorial males had higher reproductive success than non-territorial males. However, at high density territorial males were no more successful than non-territorials and the sex difference in the opportunity for sexual selection, based on the parameter <italic>I</italic><sub>mates</sub>, was low.</p></sec><sec><title>Conclusion</title><p>Male zebrafish exhibit two distinct mating tactics; territoriality and active pursuit of females. Male reproductive success is density dependent and the opportunity for sexual selection appears to be weak in this species.</p></sec> | <contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Spence</surname><given-names>Rowena</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>rs153@le.ac.uk</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Jordan</surname><given-names>William C</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>bill.jordan@ioz.ac.uk</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Smith</surname><given-names>Carl</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>cs152@le.ac.uk</email></contrib> | Frontiers in Zoology | <sec><title>Background</title><p>The advent of genetic parentage analysis has had a substantial impact on our understanding of animal mating systems. Many socially monogamous species have proven to be genetically polygamous [<xref ref-type="bibr" rid="B1">1</xref>], while territorial or harem-holding males have frequently been shown to be cuckolded [<xref ref-type="bibr" rid="B2">2</xref>]. Moreover, due to the operation of sperm competition [<xref ref-type="bibr" rid="B3">3</xref>] and cryptic female choice [<xref ref-type="bibr" rid="B4">4</xref>], mating success is not equivalent to reproductive success. It is now recognised that genetic analyses are crucial for measuring parentage, and thereby in understanding mating system evolution and the strength of sexual selection [<xref ref-type="bibr" rid="B5">5</xref>-<xref ref-type="bibr" rid="B7">7</xref>].</p><p>The strength of sexual selection depends on the sex difference in the degree of variance in reproductive success for each sex; the greater the difference between the sexes, the more opportunity there is for selection to operate [<xref ref-type="bibr" rid="B8">8</xref>]. In most species the variation in male mating success, as defined by mate number and offspring number, exceeds that of female mating success, as defined by clutch size and number of clutches. While females tend to produce similar numbers of offspring, variance in the number of offspring fathered among males can be high, with successful males monopolising females and some males failing to reproduce. The sex difference in the opportunity for sexual selection can be quantified using the parameter <italic>I</italic><sub>mates </sub>[<xref ref-type="bibr" rid="B8">8</xref>]; the difference in male and female variance in reproductive success as a function of the squared mean reproductive success for each sex:</p><p><italic>I</italic><sub>mates </sub>= <italic>I</italic><sub>males </sub>- <italic>I</italic><sub>females</sub></p><p>where</p><p><inline-graphic xlink:href="1742-9994-3-5-i1.gif"/></p><p>and</p><p><inline-graphic xlink:href="1742-9994-3-5-i2.gif"/></p><p>if <italic>V </italic>= variance in reproductive success and <italic>X</italic><sup><italic>2 </italic></sup>= squared mean reproductive success.</p><p>The greater the value of <italic>I</italic><sub>mates</sub>, the greater the opportunity for sexual selection, which is typically reflected in the degree of sexual dimorphism. However, variance in male success can be eroded through alternative male mating tactics, sperm competition and cryptic female mate choice [<xref ref-type="bibr" rid="B9">9</xref>,<xref ref-type="bibr" rid="B10">10</xref>].</p><p>The intensity of competition for mates is influenced both by the number of competitors and the temporal and spatial distribution of fertilisations [<xref ref-type="bibr" rid="B11">11</xref>,<xref ref-type="bibr" rid="B8">8</xref>]. When sexually receptive females are spatially clustered, it is easier for a single dominant male to monopolise matings to the exclusion of other males. In addition, if fertilisations are distributed evenly through time, rather than clustered, a small number of males may be able to monopolise matings. Variance in male reproductive success also correlates negatively with male density. As male density increases, rates of cuckoldry increase and the fitness payoff of territoriality relative to alternative mating tactics declines [<xref ref-type="bibr" rid="B2">2</xref>,<xref ref-type="bibr" rid="B12">12</xref>,<xref ref-type="bibr" rid="B13">13</xref>]. In this study we used behavioural and molecular analysis to investigate the effects of density on male mating success in the zebrafish, <italic>Danio rerio</italic>.</p><p>The zebrafish is a small (30–40 mm body length) tropical cyprinid fish, native to the floodplains of North Eastern India and Bangladesh [<xref ref-type="bibr" rid="B14">14</xref>]. It is an abundant fish in this region, inhabiting ditches and ponds, where it occurs in small shoals of between 2–30 individuals (R. Spence, personal observation). Males and females are similar in size and colouration. Under laboratory conditions zebrafish breed all year; females spawn every 2–3 days, with all mature ova released during a single spawning bout [<xref ref-type="bibr" rid="B15">15</xref>]. Spawning is influenced by photoperiod; mating activity commences within a few minutes of exposure to light following darkness and continues for about one hour [<xref ref-type="bibr" rid="B16">16</xref>]. Female zebrafish will release eggs directly onto a bare substrate, but when provided with an artificial spawning site, such as a plastic box filled with gravel or marbles, will preferentially use it for oviposition. Some male zebrafish are territorial during mating and a single male will aggressively attempt to control access of rivals to a spawning site [<xref ref-type="bibr" rid="B17">17</xref>]. Courtship in zebrafish involves the male swimming quickly in association with the female, often touching her flanks with his snout, circling tightly in front of her while attempting to lead her to a spawning site. Once over the spawning site, the male swims alongside the female, in close contact but slightly behind her, sometimes oscillating his body at high frequency and low amplitude. Both territorial and non-territorial males show the same courtship behaviour but whereas non-territorial males pursue females all around the aquarium, territorial males confine their activities to within a few body lengths of the spawning site and chase other males away when they try to approach.</p><p>In a previous study we manipulated density and sex ratio and measured aggressive and courtship behaviour by territorial males under each treatment [<xref ref-type="bibr" rid="B17">17</xref>]. We showed that rates of aggression increased as a function of density, although courtship did not [<xref ref-type="bibr" rid="B17">17</xref>]. Here we investigate the effects of density on territorial male reproductive success. We predicted that territorial males would have lower reproductive success at higher densities, due to increased competition from rival males. In addition, in the high-density treatment, we measured the sex difference in the opportunity for sexual selection. For zebrafish, which show little sexual dimorphism, we predicted a low estimate of <italic>I</italic><sub>mates</sub>.</p></sec><sec><title>Results</title><p>In each replicate, a single male established a territory around the artificial spawning site and remained in possession of it throughout the 4 days of the experiment. In the low-density treatment territorial males sired a mean ± SD of 56.3% ± 7.58 of the offspring, a significantly greater proportion than non-territorial males at 43.8% ± 7.58 (paired <italic>t</italic>-test: <italic>t</italic><sub>5 </sub>= 3.05, <italic>P </italic>= 0.028, <italic>d </italic>= 1.76) (Figure <xref ref-type="fig" rid="F1">1</xref>). In the high-density treatments there was no effect of either territoriality (ANOVA: <italic>F</italic><sub>1,20 </sub>= 6.39, <italic>P </italic>= 0.304) or sex bias (ANOVA: <italic>F</italic><sub>1,20 </sub>= 12.3, <italic>P </italic>= 0.159) on the number of offspring sired per male, nor was there an interaction (ANOVA: <italic>F</italic><sub>1,20 </sub>= 5.21, <italic>P </italic>= 0.352). In the high-density male-biased treatment territorial males sired a mean ± SD of 17.1% ± 12.91 of the offspring, compared to 83.0% ± 12.91 by non-territorial males, a mean of 9.21% ± 9.79 offspring per non-territorial male. In the high-density female-biased treatment the territorial male sired a mean ± SD of 16.9% ± 11.02 of the offspring, compared to 83.3% ± 10.72 by non-territorial males, a mean of 20.8% ± 14.82 per non-territorial male. We then combined both high-density treatments, thereby doubling the sample size; there was still no significant difference in reproductive success between territorial and non-territorial males; territorial males sired a mean ± SD of 17% ± 11.4 of the offspring, compared to 83% ± 11.4 by non-territorial males, a mean of 15% ± 6.39 per non-territorial male (paired <italic>t</italic>-test: <italic>t</italic><sub>10 </sub>= 0.48, <italic>P </italic>= 0.639, <italic>d </italic>= 0.223) (Figure <xref ref-type="fig" rid="F1">1</xref>).</p><fig position="float" id="F1"><label>Figure 1</label><caption><p><bold>Territoriality and reproductive success</bold>. The mean proportion (%) + SE of offspring sired by territorial male zebrafish (black bars) and the mean proportion sired per non-territorial male (white bars) under two density treatments.</p></caption><graphic xlink:href="1742-9994-3-5-1"/></fig><p>In the high-density male biased treatment the majority of experimental fish of each sex were represented in the parentage analysis. In total, 87% of males sired at least one of the 30 genotyped offspring (range 0–12, fig. <xref ref-type="fig" rid="F2">2</xref>). Females produced a mean ± SD of 14.8% ± 3.57 of the offspring each and 89% of the females in each replicate produced at least one of the 30 genotyped offspring. The estimated opportunity for sexual selection was low; <italic>I</italic><sub>males </sub>= 0.83 ± 0.22 and <italic>I</italic><sub>females </sub>= 0.80 ± 0.25, giving an estimate of 0.03 for <italic>I</italic><sub>mates</sub>. In the low density treatment where there was only one female, <italic>I</italic><sub>females </sub>= 0, so <italic>I</italic><sub>males </sub>= <italic>I</italic><sub>mates </sub>= 0.07 ± 0.01.</p><fig position="float" id="F2"><label>Figure 2</label><caption><p><bold>Distribution of paternity</bold>. The frequency distribution of genotyped offspring among territorial (black bars) and non-territorial (white bars) males in the high-density male-biased treatment.</p></caption><graphic xlink:href="1742-9994-3-5-2"/></fig></sec><sec><title>Discussion</title><p>Despite the wealth of genetic studies conducted on zebrafish, this is the first time a genetic parentage analysis has been applied to understanding their mating system. This study also serves as an example of how genetic analyses can provide insights that are not obvious from behavioural studies alone. We showed that at a low density territorial males sired significantly more offspring than non-territorial males (Fig. <xref ref-type="fig" rid="F1">1</xref>). While this result is consistent with our previous observation that at low densities territorial males monopolise spawning sites [<xref ref-type="bibr" rid="B17">17</xref>], non-territorial males were still able to achieve a relatively high reproductive success. The overall rate of paternity assignment was high (92%) and while it is possible that the paternity of the remaining unassigned 8% could erode the significance of this result, it is unlikely that the pattern of paternity would differ significantly from that observed in the majority that were assigned, given that non-assignment of offspring is likely to be random across potential sires.</p><p>At higher densities we detected no significant difference in the reproductive success of territorial and non-territorial males in either sex ratio treatment. We also analysed the combined data from both high-density treatments, thereby doubling the number of replicates, which increased the statistical power of the test and also yielded a non-significant result. While we had expected aggressive territoriality to confer some fitness advantage at higher densities, we had predicted that the advantage would be negatively correlated with density. As the fitness advantage at low densities is not great it is unsurprising that it should be eroded at higher densities. However, this result does raise the question of how territorial behaviour is maintained in the population. Both density levels used here are within the range of densities at which zebrafish occur in nature and territorial defence also occurs under natural conditions (R. Spence & C. Smith, unpublished data). Given that territoriality only occurs during the brief daily spawning period, the possible fitness cost associated with energy expenditure on territoriality may not exceed that for courtship. Consequently, while territorial defence confers a fitness advantage at low densities, it may not always do so at high densities, though the reproductive fitness of territorial males may be at least equal to that of non-territorial males. In addition, the adoption of one or other tactic, territorial defence or active pursuit of females, may be frequency dependent, in which case they would be predicted to confer equal fitness payoffs.</p><p>Alternative mating tactics are common in fish, partly because of the prevalence of external fertilisation, which makes it hard to exclude rivals [<xref ref-type="bibr" rid="B18">18</xref>]. In genetic studies of several nest-tending species with paternal care Dewoody & Avise [<xref ref-type="bibr" rid="B2">2</xref>] found that between 5–30% of embryos in nests were not sired by the nest owner. In some species, such as Atlantic salmon, <italic>Salmo salar</italic>, and bluegill sunfish, <italic>Lepomis macrochirus</italic>, territorial and sneaker males are morphologically distinct [<xref ref-type="bibr" rid="B19">19</xref>]. In other species, such as the three-spined stickleback <italic>Gasterosteus aculeatus </italic>[<xref ref-type="bibr" rid="B20">20</xref>] and the European bitterling, <italic>Rhodeus sericeus </italic>[<xref ref-type="bibr" rid="B13">13</xref>,<xref ref-type="bibr" rid="B21">21</xref>] the adoption of one or other role is flexible and probably frequency dependent. Although each territorial male in this study maintained his territorial role throughout the 4 days of the experiment, territorial male zebrafish are not morphologically distinct. Further, on the basis of laboratory and field observations, the frequency of territorial behaviour is influenced by the availability of high quality spawning sites, a factor that was not manipulated in the current study (R. Spence & C. Smith, unpublished data).</p><p>It should be noted that fertilisation rates in zebrafish may not always reach 100%, though this was not recorded in the current study. Thus, our results reflect the reproductive success of males as determined by differential embryo survival to hatching (three days post-fertilisation) rather than fertilisation success. Our genetic analysis also showed that the majority of females in each replicate produced offspring. We previously showed that females produce fewer eggs <italic>per capita </italic>at high densities [<xref ref-type="bibr" rid="B17">17</xref>] but in the absence of paternity analysis it was not clear whether this was because each female produced fewer eggs, or whether some females were excluded from spawning. This analysis suggests that females spawn smaller clutches at higher densities.</p><p>Our estimate of <italic>I</italic><sub>mates </sub>indicates that there is a low opportunity for sexual selection in zebrafish; there was no significant difference in the variance in reproductive success between males and females. Because our estimate is based on genetic parentage data, it takes account of both behavioural differences (such as territoriality) and post-mating forms of sexual selection (sperm competition and cryptic female choice) which can impact on variance in reproductive success. This result is consistent with the lack of a marked sexual dimorphism in this species. Variance in male mating success is dependent on the temporal and spatial clustering of females; the opportunity for selection is predicted to be highest when receptive females show low temporal but high spatial clustering [<xref ref-type="bibr" rid="B8">8</xref>]. Although zebrafish spawn almost daily under laboratory conditions, spawning is confined to an approximately 1-hour period each day; i.e. matings are highly temporally clustered. In nature, spawning is more seasonal, but is similarly largely confined to a brief period at dawn (R Spence & C Smith, personal observation). Spatial clustering of females may occur where sites for oviposition are limiting. However, although both females and males discriminate among spawning locations (Spence & Smith, unpublished data) they use a broad range of oviposition sites. Consequently, under natural conditions males may not always be able to monopolise either receptive females or sites of reproduction, with the outcome that the opportunity for sexual selection is weak.</p></sec><sec><title>Conclusion</title><p>The results of this study suggest that zebrafish have a promiscuous mating system with a low opportunity for sexual selection. Parentage analysis indicates that territoriality confers a fitness benefit at low densities, but at higher densities has an equivalent fitness payoff to non-territorial behaviour.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>Experiment</title><p>We carried out a behavioural experiment in February 2004, using 216 zebrafish obtained from a commercial supplier. The experimental design and results of behavioural analyses are presented in [<xref ref-type="bibr" rid="B17">17</xref>], though no results are repeated here. The original experimental design comprised two factors: density and sex ratio. However, only density effects were investigated in the present study. The low-density treatment consisted of three fish, one female and two males (six replicates) and the high-density treatment of fifteen fish, either ten males and five females or five males and ten females (six replicates of each).</p><p>Fish were housed in an environmentally controlled room with a 14:10 h hour light: dark cycle. Experiments were conducted in 60 l glass aquaria (60 × 35 cm and 40 cm high) with rested tap water at 26.5 – 28.5°C. Water quality in aquaria was maintained using filters and aquaria water was aerated with an air stone. Males and females were randomly assigned to experimental treatments, females being distinguished by the presence of a genital papilla. We placed opaque dividers between aquaria to prevent visual interactions between neighbouring fish. Fish were fed three times each day with a mixture of frozen bloodworm and commercial salmon smolt pellets. A single plastic box (150 × 100 mm and 40 mm deep) filled with 150 small glass marbles, was placed in the front right hand corner of every experimental aquarium as a spawning site. Zebrafish readily use boxes of marbles for oviposition; the eggs fall among the marbles, which prevent egg cannibalism.</p><p>Each replicate lasted for four days. We observed each aquarium daily during the spawning period to determine whether there was a single, dominant male defending the spawning site, and whether the same male remained territorial throughout. Individual territorial males were identified from unique features of their colour pattern. A subset of the replicates was videotaped for 5 min each morning and the frequency of aggressive and courtship behaviours by the territorial male were scored from video footage. Spawning sites were removed daily after the end of the spawning period and all the eggs deposited in them were carefully removed, counted and incubated until hatching, whereupon the embryos were preserved in 95% ethanol. Spawning sites were replaced 3–4 h after removal.</p><p>On the fifth day following the start of the experiment, all experimental fish were removed from aquaria and killed with an overdose of anaesthetic. Fish were measured to the nearest 0.1 mm, dissected to confirm sex and fin clips taken for parentage analysis.</p></sec><sec><title>Parentage analysis</title><p>We genotyped all the adults and a subset of 31–32 offspring from each replicate. The mean (± SD) daily number of eggs per replicate was 52 ± 8.84 in the low density treatment and 270 ± 28.5 in the high density treatments. A total of 198 adults and 560 offspring were genotyped. Offspring from a single day's spawning were selected haphazardly from each replicate. DNA was extracted from fin samples or embryos with the yolk sac removed, using Promega Wizard SV 96 Genomic DNA purification system.</p><p>Individuals were screened across seven microsatellite loci (Z10914, Z1213, Z1233, Z1496, Z6454, Z669 and Z851) belonging to six different linkage groups, with a mean of 7 alleles per locus. Primer sequences were taken from the zebrafish genetic map website [<xref ref-type="bibr" rid="B22">22</xref>-<xref ref-type="bibr" rid="B24">24</xref>]. DNA was amplified in a 6 μl reaction volume containing 1 μl genomic DNA; 2 pmol dye-labelled forward primer; 2 pmol reverse primer and 4 μl Qiagen Multiplex PCR Master Mix. PCR cycling consisted of an initial 12 min denaturation at 95°C, 30 cycles of 30 s at 94°C, 2 min at 58°C and 1 min at 72°C and a final extension at 60°C for 30 s. 1 μl PCR product was mixed with 10 μl dionized formamide and 15 μl LIZ 500 size standard and run on an automated sequencer, Applied Biosystems 3100 Genetic Analyzer. Fragment length was determined using Genemapper 3.5 software. All adults were genotyped 3 times in order to control for the effects of null alleles and allelic dropout [<xref ref-type="bibr" rid="B25">25</xref>]. The error rate (the ratio of observed allelic differences to total allelic comparisons among repeated amplifications) was estimated to be 3.4%.</p><p>Parentage was assigned using CERVUS 2.0 software [<xref ref-type="bibr" rid="B26">26</xref>]. Observed heterozygosities ranged from 0.20 to 0.71. In the low-density treatment, where maternal genotype was known, paternity was assigned with 95% confidence in 92% of the offspring sampled, (range 78%–100% between replicates). In the high-density treatments where maternity was not known, CERVUS was used in a stepwise manner, determining parentage first for the sex with the smallest number of candidate parents in a given replicate (i.e. females in the male-biased treatment and males in the female-biased treatment) and using those data where parentage was assigned with 95% confidence to determine parentage for the other sex [<xref ref-type="bibr" rid="B26">26</xref>]. Paternity was assigned with 95% confidence to 244 of 380 offspring genotyped and with 80% confidence to a further 107, (range 60%–100% between replicates). Maternity was assigned with 95% confidence to 205 of the offspring and with 80% confidence to a further, 137, (range 78%–100% between replicates). The combined exclusionary power for the high-density treatments was thus 83%.</p></sec><sec><title>Data analysis</title><p>We tested all data for normality using a Kolmogorov-Smirnov test and for equality of variance using Bartlett's test. The reproductive success of males was calculated as a percentage of offspring sired within a replicate, paternity being assigned for each individual male, both territorial and non-territorial. Only data where paternity could be assigned with at least 80% confidence were used in the analysis. A two factor ANOVA was used to test for the effects of territoriality and sex bias on male reproductive success in the high density treatments. We used paired <italic>t</italic>-tests to test for a difference in reproductive success between territorial and non-territorial males (mean value per male) within each density level. Following the recommendation of Nakagawa [<xref ref-type="bibr" rid="B27">27</xref>], Bonferroni corrections were not applied and instead a measure of effect size (Cohen's <italic>d</italic>) was estimated. This index measures the magnitude of a treatment effect as the standardised difference between two means by comparing the overlap in the distribution between the two data sets independently of sample size. An effect size of 0.8 is defined as large [<xref ref-type="bibr" rid="B28">28</xref>].</p></sec></sec><sec><title>Authors' contributions</title><p>RS and CS designed the study, RS conducted the experiment and parentage analysis, analysed the data and wrote the paper, WCJ oversaw the genetic work.</p></sec> |
Outcome assessment of the VADO approach in psychiatric rehabilitation: a partially randomised multicentric trial | <sec><title>Background</title><p>Recent studies on representative samples of psychiatric services have shown that low proportions of cases received effective rehabilitation interventions. The following are likely to be the most important causes: the scarcity of mental health workers trained in social and work skills strategies and the absence of a structured framework to formulate rehabilitation practices.</p><p>The aim of this study was to assess if a specific structured planning and evaluation manual, called VADO (<italic>Valutazione delle Abilità e Definizione degli Obiettivi </italic>– in english: <italic>Skills Assessment and Definition of Goals)</italic>, is more effective than routine interventions in reducing disability in patients with schizophrenia.</p></sec><sec><title>Method</title><p>Each of 10 mental health services were invited to recruit 10 patients with a schizophrenic disorder. Altogether 98 patients were recruited. Of these, 62 patients were randomly allocated to the intervention/experimental or a control group. The remaining group of 36 patients was not randomised and it was considered as a parallel effectiveness study. Assessment measures at the beginning of the study and at the one-year follow-up included the FPS scale of social functioning and the BPRS 4.0. Between group (VADO vs. Routine) and time effects were examined with ANOVA, Chi-square or Fisher exact. Clinical "improvement" was defined as an increase of at least ten points on the FPS or a decrease of at least 20% on BPRS scores.</p></sec><sec><title>Results</title><p>31 of the 62 randomized patients received the experimental interventions, while 31 followed the routine ones. At follow-up, the experimental group showed statistically and clinically greater improvements in psychopathology and social functioning.</p><p>Better outcomes of both social functioning and symptom severity were observed in non randomised patients (parallel effectiveness study).</p></sec><sec><title>Conclusion</title><p>The results suggest that setting personalised and measurable objectives, as recommended by the manual, can improve the outcome of rehabilitation of severe mental disorders. Better outcomes in the parallel effectiveness study could be attributed to the greater confidence and enthusiasm of staff in centres where the VADO approach originated.</p></sec> | <contrib contrib-type="author" id="A1"><name><surname>Pioli</surname><given-names>Rosaria</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>rpioli@fbf.it</email></contrib><contrib contrib-type="author" id="A2"><name><surname>Vittorielli</surname><given-names>Michela</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>mvittorielli@fbf.it</email></contrib><contrib contrib-type="author" corresp="yes" id="A3"><name><surname>Gigantesco</surname><given-names>Antonella</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>gigantesco@iss.it</email></contrib><contrib contrib-type="author" id="A4"><name><surname>Rossi</surname><given-names>Giuseppe</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>grossi@fbf.it</email></contrib><contrib contrib-type="author" id="A5"><name><surname>Basso</surname><given-names>Luigi</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>crp.grieserhof@asbz.it</email></contrib><contrib contrib-type="author" id="A6"><name><surname>Caprioli</surname><given-names>Chiara</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>wpa.irccs@fbf.it</email></contrib><contrib contrib-type="author" id="A7"><name><surname>Buizza</surname><given-names>Chiara</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>wpa.irccs@fbf.it</email></contrib><contrib contrib-type="author" id="A8"><name><surname>Corradi</surname><given-names>Angela</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>ricerca.psichiatria@fbf.it</email></contrib><contrib contrib-type="author" id="A9"><name><surname>Mirabella</surname><given-names>Fiorino</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>mirabella@iss.it</email></contrib><contrib contrib-type="author" id="A10"><name><surname>Morosini</surname><given-names>Pierluigi</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>morosini@iss.it</email></contrib><contrib contrib-type="author" id="A11"><name><surname>Falloon</surname><given-names>Ian RH</given-names></name><xref ref-type="aff" rid="I4">4</xref><email>ifalloon@tiscali.it</email></contrib> | Clinical Practice and Epidemiology in Mental Health : CP & EMH | <sec><title>Background</title><p>The results of a recent study on a representative sample of Italian psychiatric services [<xref ref-type="bibr" rid="B1">1</xref>] have shown that only 35% of patients with schizophrenia who live with their own family follow a rehabilitation programme and that such rehabilitation programmes include the setting of personalised objectives in only 66% of cases. Therefore, only 23% of patients may receive effective rehabilitation interventions. Similar low proportions of cases receiving evidence-based psychosocial interventions have been observed in the US [<xref ref-type="bibr" rid="B2">2</xref>]. Among the causes, the following are likely to be the most important: psychodynamic orientation of many rehabilitation managers, the scarcity of mental health workers trained in social and work skills strategies and the absence of a structured framework to formulate rehabilitation practices [<xref ref-type="bibr" rid="B3">3</xref>].</p><p>In 1998 the rehabilitation centre of the Fatebenefratelli Institute of Brescia, in collaboration with the Italian National Institute of Health and the Institute of Psychiatry of the University of Napoli, published a short manual for the planning and evaluation of rehabilitative interventions in psychiatric facilities [<xref ref-type="bibr" rid="B4">4</xref>]. The principles were derived from the Boston University approach [<xref ref-type="bibr" rid="B5">5</xref>], introduced in Italy by Marianne Farkas. The handbook was called VADO (<italic>Valutazione delle Abilità e Definizione degli Obiettivi </italic>– in english: <italic>Skills Assessment and Definition of Goals)</italic>.</p><p>The effectiveness of the VADO approach was assessed in a pilot study of 45 patients with schizophrenia, all treated in Brescia. Encouraging results in this study over a 6-month follow-up [<xref ref-type="bibr" rid="B6">6</xref>] raised the question of its efficacy and a multicentric controlled study was then planned.</p><sec><title>Aims of the study</title><p>The aim of this report is to give a full description of the controlled study and the one-year follow-up results. The main outcome criterion was social functioning; however, since the rehabilitation strategy also focussed on self-management of mental disorders, psychopathological symptoms were also considered.</p></sec></sec><sec sec-type="methods"><title>Methods</title><sec><title>Procedure</title><p>The study was conducted in 10 day or residential rehabilitation centres; two in Brescia, Rome and Mantova, and one in Bolzano, Bologna, Brugherio, Soresina.</p><p>In the 6 months following a VADO training course (see later), each centre recruited at least 10 patients fulfilling the following criteria: a) aged between 18 and 65 years; b) diagnosis of schizophrenic, schizoaffective or delusional disorder according to ICD-10; c) global functional score < 70 on the FPS (see later; this score corresponds to manifest difficulties in at least one key functioning area); d) absence of disabling physical diseases, psycho-organic syndromes or of mental retardation of medium to severe degree, according to the ICD-10 criteria; e) stable medication regimen.</p><p>The patients were allocated to the intervention/experimental or the control group according to a randomised sequence. The randomisation protocol was held in the Brescia central unit and the centres had no possibility to influence the allocation. Altogether 98 patients were recruited, 57 in the intervention group and 41 in the control group. However, only 62 patients were randomised (31 to the intervention group and 31 to the control group). The remaining 36 patients were not randomised because: a) 18 patients recruited by the two residential facilities of Brescia and 8 from the rehabilitation day centre of Monza were all allocated to the intervention group, in order to avoid a contamination bias, because the rehabilitation workers there were already all implementing the VADO approach; b) all 10 patients of the Bologna centre that had been originally allocated to the VADO group had to be included in the control group, because all the VADO trained workers were suddenly transferred before beginning the VADO-oriented rehabilitation process. This further subgroup of 36 non-randomised patients was considered as a parallel effectiveness study.</p><p>Patients signed an informed consent form after the study had been explained to them.</p></sec><sec><title>Assessments</title><p>At the beginning of the study (T0) the socio-demographic characteristics of the patients, illness duration and hospital admissions during the last year were recorded. At T0, after 6 months (T1) and after 1 year (T2), psychopathology was measured with the Italian standardized BPRS 4.0 [<xref ref-type="bibr" rid="B7">7</xref>] and social functioning with the Personal and Social Functioning Scale, FPS (see later).</p><p>BPRS and FPS assessments were carried out by the VADO-trained rehabilitation workers. The assessments were checked by one independent research assistant at each centre, who interviewed the rehabilitation workers after each assessment. The research assistants were instructed not to ask anything that could reveal the treatment received by the patients. In the few cases of disagreement between the research assistant and the rehabilitation workers, the rating was made by another independent research assistant. The VADO Personal and Social Functioning Scale (FPS) is a modified version of the Social and Occupational Functioning Assessment Scale (SOFAS) [<xref ref-type="bibr" rid="B8">8</xref>]. As with the SOFAS, the FPS scores range from 100 (excellent functioning) to 1 (extremely severe impairment with risk for survival). The instructions for scoring the 10 points within each level are more detailed. The rater is instructed to take into account four main areas: <italic>work and/or socially useful activities</italic>; <italic>family, personal and social relationships; self-care</italic>; <italic>aggressive and destructive behaviours</italic>. Suicide risk is considered in the score only as much as suicidal ruminations may interfere with social-functioning. The FPS requires a brief and simple training, that is described in the VADO manual. The FPS can be easily scored by rehabilitation workers, including those with a limited psychiatric experience. It is, therefore, an instrument that may be useful to assess severity and outcome in routine practice [<xref ref-type="bibr" rid="B9">9</xref>]. Thirty patients were rated independently by two professionals: the non-weighted 10 levels k = 0.75.</p><p>BPRS was analysed according to the following factors: <italic>positive (items 9–12, 14–15, and 24) and negative (items 13, 16–18, and 20) symptoms, anxiety and depression (item 1–5, and 19), mania and hostility (items 6–8. and 21–23)</italic>.</p></sec><sec><title>Approach groups</title><sec><title>a) VADO approach</title><p>The aim of the VADO approach is to help the rehabilitative team to define individual rehabilitative programs, to focus on negotiating realistic specific (measurable) goals, and to routinely evaluate the attainment of those specific goals, first, within the rehabilitation unit, and subsequently, in real life. The VADO approach comprises five components: 1) assessment of 28 area of social functioning, including an area of skills, and choice of priority areas aspects compatible with so called 'global objective' (how the patient would like to live); 2) global evaluation, 3) negotiation of realistic, attainable, specific measurable objectives, 4) subdivision of specific objectives in skills and tasks; 5) maintenance and generalisation. The VADO handbook [<xref ref-type="bibr" rid="B4">4</xref>] provides detailed instructions on how to assess patients' disabilities and residual strengths and how to negotiate the rehabilitation program with him/her, as well as worksheets and forms. These include the Functioning Assessment (FA) to collect information covering 28 domains of patient's functioning. On the basis of the FA, rehabilitation workers score the 28 domains on a 6-point scale on the Rehabilitation Areas Form (RAF). The RAF assesses the need for rehabilitation in each domain, and whether a rehabilitation intervention is planned to meet that need.</p><p>The VADO manual does not detail rehabilitation techniques, but explicitly recommends modelling, non-verbal and verbal prompting, role-playing [<xref ref-type="bibr" rid="B10">10</xref>], and structured problem solving. In the self-management of mental disorders it outlines the stress-vulnerability model, identification of early signs of relapses and functional coping strategies enhancement, according to the models of Falloon [<xref ref-type="bibr" rid="B11">11</xref>], Fowler et al. [<xref ref-type="bibr" rid="B12">12</xref>], and Chadwick et al. [<xref ref-type="bibr" rid="B13">13</xref>]. Motivational interviewing strategies are recommended to correct misinformation about medication and to improve adherence.</p></sec><sec><title>b) Routine approach</title><p>The control treatment consisted of the usual rehabilitation activities of each participating centre. The ongoing activities were described according to the classification of "<italic>Glossary of interventions and activities of Mental Health Departments" </italic>[<xref ref-type="bibr" rid="B14">14</xref>]. Two thirds of participating centres stated that they had already a policy of personalised programs in interpersonal and social skills and all stated that they were organising some kind of social activity (group discussions, newspapers reading with comments, movie watching, and so on); two thirds of centres were also organising parties, excursions and supported holidays. The rehabilitation workers of the control patients were not trained in the VADO approach.</p></sec></sec><sec><title>Psychopharmacological treatment</title><p>No change in psychopharmacological treatment was planned for patients in either group. The VADO approach recommended providing detailed information to patients about the purpose, nature and side effects of pharmacological treatments in order to improve not only adherence, but also active participation in medication management.</p></sec><sec><title>Training in the VADO approach</title><p>Before beginning the study, at least two staff members from each unit (one physician and one or more rehabilitation workers) attended a 32-hour training course (8 hours a day for 4 days). The course was based on the small groups methodology, with extensive role playing and case discussion. Participants were shown video-recorded cases and were asked to give the FPS score (see later) and fill the RAF forms independently and then to discuss any inconsistencies in their ratings. Participants did not receive no further training in rehabilitation techniques. However, because of its features, the VADO training induced to apply or consolidated the application of social skills training techniques in the program. For example, the trainer showed the group segments of videocassettes that provided examples on how to evaluate the progress toward the achievement of a specific measurable patient goal after using techniques based on social learning techniques such as demonstration, role playing, positive feedback, shaping and generalisation.</p><p>After the course, participants in the study attended one day per month supervision sessions for six months. During these sessions they discussed problems and doubts about the clinical management of cases.</p></sec><sec><title>Statistical analysis</title><p>Data were analysed with SPSS 12.0. Between group (VADO vs. Routine) and time effects (T0 vs. T1; T0 vs. T2) were examined with ANOVA, Chi-square or Fisher exact tests depending on the nature of the data. Clinical "improvement" was defined as an increase of at least ten points on the FPS or a decrease of at least 20% on BPRS scores.</p><p>To identify variables associated with improvement at T2, two different multiple logistic regression analyses were performed. In the first, the FPS score was the dependent variable and we entered treatment group, age, sex, length of disease, and FPS T0 scores as independent variables. In the second, the BPRS total score was the dependent variable and treatment group, age, sex, length of disease, and BPRS total score at T0 were entered as independent variables.</p></sec></sec><sec><title>Results</title><sec><title>1. Patient characteristics (see Table <xref ref-type="table" rid="T1">1</xref>)</title><table-wrap id="T1" position="float"><label>Table 1</label><caption><p>T0. Characteristics, living conditions, previous admissions of patients initially recruited in the study. Between brackets, standard deviations.</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left">Variables</th><th align="center">Experimental group N = 57</th><th align="center">Control group N = 41</th><th align="center">Total N = 98</th></tr></thead><tbody><tr><td align="left"><bold>Male sex (%)</bold></td><td align="center">60</td><td align="center">70</td><td align="center">64</td></tr><tr><td align="left"><bold>Age average</bold></td><td align="center">39 (10)</td><td align="center">42 (10)</td><td align="center">40 (10)</td></tr><tr><td align="left"><bold>Years of completed education cycles (%)</bold></td><td/><td/><td/></tr><tr><td align="left"> 5 years or less</td><td align="center">13</td><td align="center">35</td><td align="center">22</td></tr><tr><td align="left"> 8 years</td><td align="center">60</td><td align="center">44</td><td align="center">53</td></tr><tr><td align="left"> 12–13 years or more</td><td align="center">27</td><td align="center">21</td><td align="center">25</td></tr><tr><td align="left"><bold>Occupied (%)</bold></td><td align="center">29</td><td align="center">33</td><td align="center">31</td></tr><tr><td align="left"><bold>Usual living conditions</bold></td><td/><td/><td/></tr><tr><td align="left"> Alone</td><td align="center">6</td><td align="center">15</td><td align="center">10</td></tr><tr><td align="left"> With relatives</td><td align="center">59</td><td align="center">54</td><td align="center">57</td></tr><tr><td align="left"> in residential centres</td><td align="center">35</td><td align="center">31</td><td align="center">33</td></tr><tr><td align="left"><bold>Duration of disorder in years *</bold></td><td align="center">17 (9)</td><td align="center">21 (10)</td><td align="center">19 (9)</td></tr><tr><td align="left"><bold>Voluntary admissions in last year</bold></td><td/><td/><td/></tr><tr><td align="left"> 0</td><td align="center">55</td><td align="center">70</td><td align="center">61</td></tr><tr><td align="left"> 1</td><td align="center">20</td><td align="center">10</td><td align="center">21</td></tr><tr><td align="left"> 2 or more</td><td align="center">15</td><td align="center">20</td><td align="center">18</td></tr><tr><td align="left"><bold>Compulsory admissions in last year</bold></td><td/><td/><td/></tr><tr><td align="left"> 0</td><td align="center">86</td><td align="center">93</td><td align="center">89</td></tr><tr><td align="left"> 1</td><td align="center">12</td><td align="center">5</td><td align="center">9</td></tr><tr><td align="left"> 2 or more</td><td align="center">2</td><td align="center">2</td><td align="center">2</td></tr></tbody></table><table-wrap-foot><p>statistically significant differences (P < 0.05) between the two groups.</p></table-wrap-foot></table-wrap><p>Patients were mostly unemployed single males with a long duration of illness; two-thirds had limited education. The two groups had similar socio-demographic and clinical variables. However, duration of illness was significantly greater in the control group, which also showed a trend towards higher BPRS and lower FPS scores. The RAF domains in which problems were present in more than 75% of the patients were social activities, family life, friends and supporting relationships, self management of mental health, work and socially useful activities. At T0, the presence of problems was ascertained in 69% of the domains in the experimental group and in 73% in the control group.</p></sec><sec><title>2. Setting and achievement of rehabilitation objectives</title><p>The percentage of problem areas in which rehabilitative programs were implemented was 20% in the experimental and 13% in the control group.</p><p>In the 57 experimental group patients, 174 objectives were planned of which 122 were achieved. All experimental group patients agreed on at least one objective. 24% of objectives involved household chores; 31% participation in residential or day centre life; and 25% aimed at self management of mental health. The areas in which a significantly greater between-group difference in the decreased proportion of patients with problems were: participation in residential or day centre life, self care, self management of mental health, work and socially useful activities, and coping with emergencies.</p></sec><sec><title>3. Randomised controlled study (N = 61)</title><sec><title>3a. Patient discharges and drop-outs</title><p>During the first 6 months no patients in the experimental group and only 3 from the control group dropped out after having psychotic exacerbations. All were single women with a duration of the illness of more than 10 years. Therefore, at 6 months, T1, the study included 59 patients, 31 in the intervention group and 28 in the control group.</p><p>From 6–12 months, 2 patients were discharged from the experimental group, and did not attend the T2 assessment. In the control group, one patient was discharged and one moved to another city; they did not attend T2 assessment. Therefore, the one-year assessment, T2, included 55 of 62 original patients (89%), 29 in the intervention group and 26 in the control group.</p></sec><sec><title>3b. Change in social functioning (see Table <xref ref-type="table" rid="T2">2</xref> and Figure <xref ref-type="fig" rid="F1">1</xref>)</title><table-wrap id="T2" position="float"><label>Table 2</label><caption><p>Percentages of improved patients at 12-month (T2) follow-up versus T0 in the randomised experimental and control groups (N = 55)</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left">Variables</th><th/><th/><th/></tr></thead><tbody><tr><td/><td align="center"><bold>Experimental (N = 29)</bold></td><td align="center"><bold>Control (N = 26)</bold></td><td align="center"><bold><italic>P value </italic></bold></td></tr><tr><td colspan="4"><hr/></td></tr><tr><td align="left"><bold>BPRS positive symptoms</bold></td><td align="center">34.5</td><td align="center">3.8</td><td align="center">0.00</td></tr><tr><td align="left"><bold>BPRS negative symptoms</bold></td><td align="center">37.9</td><td align="center">19.2</td><td align="center">0.13</td></tr><tr><td align="left"><bold>BPRS mania/hostility</bold></td><td align="center">31.0</td><td align="center">7.7</td><td align="center">0.04</td></tr><tr><td align="left"><bold>BPRS anxiety/depression</bold></td><td align="center">34.5</td><td align="center">23.1</td><td align="center">0.35</td></tr><tr><td align="left"><bold>BPRS overall</bold></td><td align="center">31.0</td><td align="center">7.7</td><td align="center">0.04</td></tr><tr><td align="left"><bold>FPS</bold></td><td align="center">51.7</td><td align="center">42.3</td><td align="center">0.48</td></tr></tbody></table></table-wrap><fig id="F1" position="float"><label>Figure 1</label><caption><p>Average FPS scores at baseline, after 6 months, and after 12 months in the treated (T) and control (C) groups.</p></caption><graphic xlink:href="1745-0179-2-5-1"/></fig><p>From the 31 experimental group patients that completed T1 assessments, 7 (22.6%) improved clinically compared to 3 (10.7%) of the 28 control patients (chi square = 1.472, p = 0.22). At T2, 15 of the 29 experimental group (51.7%) and 11 of 26 (42.3%) controls were improved (chi square = 0.488) (Table <xref ref-type="table" rid="T2">2</xref>).</p><p>At T0, mean FPS scores in the experimental group were 41.1 ± 12.8 and in control group were 37.8 ± 13.4 (t = -0.984, df = 57, p = 0.33). At T1, the FPS increased in both (46.6 ± 13.1 versus 39.8 ± 14.5; t = -1.887, df = 57, p = 0.06). At T2, a the experimental group showed further improvement compared to no further improvement in the control group (49.0 ± 14.3 versus 39.9 ± 14.3; t = -2.360, df = 53, p < 0.05) (Figure <xref ref-type="fig" rid="F1">1</xref>).</p></sec><sec><title>3c. Changes in psychopathology (see Table <xref ref-type="table" rid="T2">2</xref> and Figure <xref ref-type="fig" rid="F2">2</xref>)</title><fig id="F2" position="float"><label>Figure 2</label><caption><p>Average total BPRS scores at baseline, after 6 months, and after 12 months in the treated (T) and control (C) groups.</p></caption><graphic xlink:href="1745-0179-2-5-2"/></fig></sec></sec><sec><title>BPRS global score</title><p>Significant differences between groups were found at T1 (chi square = 3.641, p = 0.06) and T2 (Table <xref ref-type="table" rid="T2">2</xref>).</p><p>At T0, BPRS total scores in the experimental and control groups were respectively 57.8 ± 21.2 and 55.4 ± 20.3 (t = 0.445, df = 57, p = 0.66). At T1, the BPRS decreased in the experimental group (50.3 ± 21.1) and increased in the controls (58.4 ± 21.1) (t = 1.439, df = 57, p = 0.16). At T2, both groups had improved (49.8 ± 20.7 versus 55.2 ± 21.0; t = 0.965, df = 53, p = 0.34) (Figure <xref ref-type="fig" rid="F2">2</xref>).</p><p>In the multiple logistic regression analysis, a test of the full model with all five independent variables was significant (chi square = 11.5, df = 5, p < 0.05). According to the Wald criterion, the only strong association with improvement was being an experimental group patient (odds ratio: 9.67, 95% CI: 1.39 -67.08).</p></sec><sec><title>BPRS positive symptoms</title><p>Significant differences between groups were found at T1 (chi square = 4.647, p < 0.05), and T2 (chi square = 8.042, p <0.01).</p></sec><sec><title>BPRS negative symptoms</title><p>At T1, difference between the groups approached significance (chi square = 3.487, p = 0.06). But at T2, no significant difference was observed.</p></sec><sec><title>BPRS mania/hostility</title><p>Significant difference between groups was found at T2 (chi square = 4.668, p < 0.05), but not at T1.</p></sec><sec><title>BPRS anxiety/depression</title><p>No statistically significant differences were found between the groups at follow-up.</p></sec><sec><title>4. Not randomised controlled study (N = 36)</title><sec><title>4a. Patient discharges and drop-outs</title><p>During the first 6 months no drop-outs occurred in either group. At 6 months, T1, the study included 36 patients, 26 in the intervention group and 10 in the control group.</p><p>From 6–12 months, 2 patients were discharged from the experimental group, and did not attend the T2 assessment; another dropped-out after a psychotic exacerbation. In the control group, one patient had a psychotic exacerbation. Therefore, the one-year assessment, T2, included 32 of the 36 original patients (89%), 23 in the intervention and 9 in the control group.</p></sec><sec><title>4b.Change in social functioning (see Table <xref ref-type="table" rid="T3">3</xref>)</title><table-wrap id="T3" position="float"><label>Table 3</label><caption><p>Percentages of improved patients at 12-month (T2) follow-up versus T0 in the non randomised experimental and control groups (N = 32)</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left">Variables</th><th/><th/><th/></tr></thead><tbody><tr><td/><td align="center"><bold>Experimental (N = 23)</bold></td><td align="center"><bold>Control (N = 9)</bold></td><td align="center"><bold><italic>P value </italic></bold></td></tr><tr><td colspan="4"><hr/></td></tr><tr><td align="left"><bold>BPRS positive symptoms</bold></td><td align="center">43.5</td><td align="center">0.0</td><td align="center">0.03</td></tr><tr><td align="left"><bold>BPRS negative symptoms</bold></td><td align="center">60.9</td><td align="center">11.1</td><td align="center">0.02</td></tr><tr><td align="left"><bold>BPRS mania/hostility</bold></td><td align="center">34.8</td><td align="center">11.1</td><td align="center">0.38</td></tr><tr><td align="left"><bold>BPRS anxiety/depression</bold></td><td align="center">34.8</td><td align="center">0.0</td><td align="center">0.07</td></tr><tr><td align="left"><bold>BPRS overall</bold></td><td align="center">52.2</td><td align="center">0.0</td><td align="center">0.01</td></tr><tr><td align="left"><bold>FPS</bold></td><td align="center">73.9</td><td align="center">33.3</td><td align="center">0.05</td></tr></tbody></table></table-wrap><p>Of the 26 experimental group patients completing T1 assessments, 16 (61.5%) improved clinically compared to 0 of the 10 control patients (chi square = 11.077, p < 0.01). At T2, 17 of the 23 experimental group (73.9%) and 3 of 9 (33.3%) controls were improved (chi square = 4.545, p = 0.05).</p></sec><sec><title>4c. Changes in psychopathology (see Table <xref ref-type="table" rid="T3">3</xref>)</title><sec><title>BPRS global score</title><p>Significant differences between groups were found at T1 (chi square = 11.077, p < 0.01) and T2 (chi square = 7.513, p < 0.05).</p></sec></sec><sec><title>BPRS positive symptoms</title><p>Significant differences between groups were found at T1 (chi square = 9.890, p < 0.01), and T2 (chi square = 5.692, p <0.05).</p></sec><sec><title>BPRS negative symptoms</title><p>Significant differences between groups were found at T1 (chi square = 5.713, p < 0.05), and T2 (chi square = 6.432, p <0.05).</p></sec><sec><title>BPRS mania/hostility</title><p>Significant difference between groups was found at T1 (chi square = 9.890, p < 0.01), but not at T2.</p></sec><sec><title>BPRS anxiety/depression</title><p>Significant difference between groups was found at T1 (chi square = 6.092, p < 0.05), and at T2, difference approached significance (chi square = 4.174, p = 0.07).</p></sec></sec></sec><sec><title>Discussion</title><p>This study suggests that multi-centred controlled studies of complex psychosocial interventions in routine rehabilitation settings is challenging, but feasible. The present study suggests that a structured approach to the assessment of rehabilitation needs, with specific goal setting and accountability to motivate workers to follow through with rehabilitation plans may encourage the application of evidence-based treatment approaches, and lead to improved social and clinical outcomes.</p><p>This study has several strengths:</p><p>1) It evaluated the efficacy of a structured rehabilitation intervention in clinical practice of common centres, with a low consumption of professional and financial resources and after a very brief, even if intensive, training. Most studies of the efficacy of specific rehabilitation interventions have been performed by highly specialised personnel and on highly selected patients [<xref ref-type="bibr" rid="B5">5</xref>,<xref ref-type="bibr" rid="B15">15</xref>-<xref ref-type="bibr" rid="B22">22</xref>].</p><p>2) Outcomes were investigated also for the patients that were discharged. All patients discharged in the first six months from the experimental group and most of those discharged in the second six months came back to follow-up assessments.</p><p>3) This was the first controlled partially randomised psychiatric rehabilitation study to be carried out in Italy, and it has stimulated further similar studies.</p><p>4) The VADO approach, which is based on the negotiation of measurable objectives that are relevant to patient's quality of life, seems to be very promising from a common sense point of view. This study seems to suggest that it is effective, at least when compared with less structured and more limited rehabilitation strategies. It should be noted that control group patients also improved their social functioning. This indicates that the control patients were not neglected, even if they usually received less attention from staff.</p><p>5) The lack of difference between randomised and non-randomised control groups shows that contamination cannot be ruled out. However, there is some evidence against it. Better outcomes of both social functioning and symptom severity in non-randomised patients could be attributed to the greater confidence and enthusiasm of staff in centres where the VADO approach originated.</p><p>6) The VADO approach is simple, requires only brief training and does not cost extra to apply. Thus, it is likely to prove cost-effective.</p><p>On the other hand, this study had a number of serious weaknesses. The optimal size of the study was not estimated a priori. The outcome measures were applied by the same rehabilitation workers who were assessing needs and setting goals and were neither independent or blind to treatment allocation. An effort was made to check the validity of each assessment by independent research assistants, but bias was highly likely. Future studies are planned with improved financial support to enable blind, independent assessment, to avoid contamination when different strategies are implemented in the same setting, by the same workers, and to improve the standardization of the psychosocial and pharmacological strategies implemented.</p><p>In the experimental group, the decrease in disability was associated with a marked improvement in psychopathological symptoms and a low rate of psychotic exacerbation. This result was somewhat unexpected because these clinical domains were not the main targets for rehabilitation. In the absence of any substantial changes in pharmacotherapy this finding may be partially explained with the inclusion of a symptom self-management module or it could be also a reflection of more attention paid by workers that used the VADO approach.</p><p>Patients treated with the VADO approach were also more likely to be discharged home or to sheltered apartments. In the one year of the study 40% were discharged in this manner and showed further improvement in their new environments. The VADO approach helps the patients decide where and how to live and attempts to provide them with the skills and confidence needed to achieve their independence as advocated by Marshall [<xref ref-type="bibr" rid="B23">23</xref>] and Mueser [<xref ref-type="bibr" rid="B24">24</xref>]. In contrast, a nationwide Italian study on the state of the Italian psychiatric residential facilities [<xref ref-type="bibr" rid="B25">25</xref>] showed that, although 32% of the patients were younger than 40 years of age, discharge rates were extremely low.</p><p>Finally, it may be noted that despite the comprehensive needs assessment and goal setting only a small proportion of problem issues were addressed during the one-year study period, and less than half the patients were discharged successfully. This may be seen as a limitation of the VADO approach, but perhaps may be also a reflection of the need for long-term rehabilitation programmes that continue to address the priority needs of patients for many years, albeit prioritising those needs that most impede the progress of patients towards their desired life styles. Research studies of rehabilitation programmes may need to extend for much longer periods to examine the full benefits of such approaches [<xref ref-type="bibr" rid="B26">26</xref>].</p></sec><sec><title>Competing interests</title><p>The author(s) declare that they have no competing interests.</p></sec> |
Work, identity and health | Could not extract abstract | <contrib contrib-type="author" corresp="yes" id="A1"><name><surname>Fryers</surname><given-names>Tom</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>yanyak@doctors.org.uk</email></contrib> | Clinical Practice and Epidemiology in Mental Health : CP & EMH | <sec><title>Personal introduction</title><p>Having free-lanced as a public health physician for over a decade, with the inevitable variation in both the avaiIability and type of work, 'retirement' is, perhaps, a vaguer concept than for those with conventional full time jobs. In my middle sixties, I was drifting towards retirement, thinking that I should soon refuse any new work, when an incident occurred quite unconnected with work. Many have experienced being treated by someone with utter contempt, but a dramatic confrontation in which I was made to feel a worthless 'nothing' forced me to reflect on my immediate future, my social status in retirement, and the importance of work. I abandoned thoughts of full retirement and took up new research commitments, having concluded that, for me, it was important to be a 'something', whatever that 'something' was.</p><p>Since then, talking to many others about retirement, I have found both resonances and differences in their concerns and strategies approaching retirement, but many other issues concerning the status of work in our society have arisen. This paper was an attempt to explore these issues for the 2004 meeting of the European Association for Medicine of the Person, Drubeck, Germany. The Association derives historically from the work of Paul Tournier a Swiss general practitioner, whose approach as a Christian physician encompassed physical, psychological, spiritual and social elements [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B2">2</xref>].</p></sec><sec><title>A. Work</title><sec><title>1. The work ethic</title><p>Many 'Western' societies are very work-oriented cultures. Work is usually necessary to earn a living, and there are strong moral pressures to work deriving from our Christian heritage. Luther saw daily work as a vocation which we are bound to undertake under God, sharing in His work on earth. Calvinists believed that hard and successful work was a sign of salvation [<xref ref-type="bibr" rid="B3">3</xref>]. John Wesley, founder of Methodism, said, in 1746 [<xref ref-type="bibr" rid="B4">4</xref>] that we should "make our daily employment a sacrifice to God; to buy and sell, to eat and drink, to His glory." These attitudes became very powerful, especially in protestant communities, leading to the 'protestant work ethic', which both Max Weber and RH Tawney considered to underpin the rise of modern capitalism. Marx analysed the way in which work reflects social structures and argued that the way it is organised shapes both society and individual lives, but, whatever the context, the importance of work was emphasised [<xref ref-type="bibr" rid="B3">3</xref>].</p><p>Paul Tournier in 'Learning to Grow Old' [<xref ref-type="bibr" rid="B5">5</xref>] wrote; "In the West, almost everyone agrees in proclaiming in words, in theory, the supreme value of the human person. We quote ... 'Man is the measure of all things'. It is man, we say, who gives work its dignity, and not work which gives man his value. But in reality things are quite different. The general social atmosphere in which we are immersed from childhood, and which influences us without even our thinking clearly about it, teaches us the superior importance of work. Work, as a duty, is seen to be full of dignity, even if it is tedious or inhuman." We might note, in the UK parliamentary election of May 2005, the insistence of all three major parties that they serve the interests of 'hard-working families'!</p><p>Thomas Carlyle and John Ruskin, sharing that protestant inheritance, nevertheless gave a different emphasis, ideological and romantic. Ruskin believed that men should and could be happy in work "God intends no man to live in this world without working, but it seems to me no less self-evident that He intends every man to be happy in his work" [<xref ref-type="bibr" rid="B6">6</xref>]; Carlyle said "Labour is life" [<xref ref-type="bibr" rid="B7">7</xref>]. William Morris commended Ruskin for demonstrating "that it is possible for men to rejoice in their work" [<xref ref-type="bibr" rid="B8">8</xref>].</p></sec><sec><title>2. Work in society</title><p>The eminent sociologist Peter Worsley [<xref ref-type="bibr" rid="B9">9</xref>] wrote: "Work is central to our culture. When someone asks "What do you do?" they really mean "What work do you do?". When a woman is asked "Do you work?", what is meant is "Are you doing a paid job?" Yet many people without a paid job work at other kinds of productive activities." So we note that the comments above referred largely to paid work, and applied almost entirely to men; issues concerning women's work have changed greatly and are both varied and complex; they will be addressed later.</p><p>There are cultures which do not value work as highly as in Europe and America. Lee [<xref ref-type="bibr" rid="B10">10</xref>] quotes Kalahari Bushmen saying "Why should we work, when God has provided so many mongongo nuts?" But it has been generally held in 'western' societies that people should work hard and conscientiously at whatever job they have, both to earn money for their families, and to serve the community. However, it was also a common assumption that material improvement, accumulation of wealth, increased social status and influence, that is, worldly success, was a sign of God's blessing on you and your work. One still hears echos of this, alongside individualistic ideologies and wide disparities in wealth.</p><p>Preparation for a society in which work-values predominate has become the principle ground of education. The voices of those who argue for classical education, or education for life, or education for increased leisure, or even education for the 'global village', are lost in the clamour of those who say that education should be for 'the world of work'. Because it does not do this very well in a constantly changing 'world of work', governments frequently change the system, and teachers are unclear what is expected of them. Children are expected to find their adult identity in work, but are not well prepared to do so. And their more fundamental needs – 'education for life' perhaps, will not have been met. Perhaps this is part of the explanation of anti-social behaviour in the young; having no clear identity, no clear status and no clear role in the adult world?</p><p>Work is perceived, therefore, as not only providing an income, but giving social legitimacy to our lives. For many, it may be the principle source of personal identity, mediating the sense of being a valued person necessary for self-esteem. A profession or trade gives us an identity – as a doctor, a teacher, an engineer, a motor mechanic, a secretary, an electrician, and so on. A particular job, independent of profession or trade, may give us an identity – as a consultant, a manager, a director, a foreman, a work-team leader. Or we may derive a sense of identity simply from being a worker, a participant and contributor; people who cannot work, or cannot find work in our society may understandably envy even this identity.</p></sec></sec><sec><title>B. Attitudes to work</title><sec><title>1. The value of work</title><p>Our societies have changed substantially in recent decades, and continue to change, not least in attitudes to work, but attitudes are also very varied. Some see work as an unpleasant chore, to be undertaken as little as possible only because you need money, and to be given up as soon as possible. Some will complain about aspects of their particular job, but see work in general as a good thing, without which life would lose a lot of its meaning. Others may see work – at least, their work – as the main source of life-satisfaction, deeply rewarding, important for the community, and hard to relinquish.</p><p>Most of us will have shared these attitudes at different periods in our careers, or even concurrently about different aspects of our work. Whatever the justified complaints of doctors, nurses, teachers and other professions, we are privileged to have such professions, providing jobs with intrinsic intellectual, emotional, personal, even spiritual satisfactions, with social respect, and, usually, more than adequate pay.</p><p>For some jobs there is little satisfying in the tasks to be done, but work has other things to offer, without which people can feel lost, useless, un-valued, of no clear identity.</p><p>Whatever the job, it can give a sense of belonging, of being a contributor; an important part, however menial, of an organisation with a bigger purpose, a valued part of society. A nice example is the care-taker/cleaner portrayed by Gervase Phinn in his fictionalised accounts of a Yorkshire education service [<xref ref-type="bibr" rid="B11">11</xref>].</p><p>Work can provide a structure for the day, week and year without which life just drifts by, a commitment to yourself and others, a purpose for getting up in the morning and going to do something. Our local sheep farmers, in spite of the current dis-satisfactions with farming, are time-tabled by their flocks.</p><p>Work can provide the satisfaction of doing something well, whatever it is, of pride in your work, of being a 'craftsman', of seeing a good end-product. We have recently experienced this with a local builder, joiner plumber and electrician, all of whom had a deep commitment to craftsmanship.</p><p>Work can provide friends, a social group to belong to, companionship, shared lives in joys and sorrows, support when suffering or in need. My father, with a job well below his potential, found this in the 'Sports and Social Club'.</p><p>Any or all of these can play a major role in fostering self-esteem, an essential ingredient of good mental health, and closely tied up with our sense of being a valued person, of having a personal identity. Not everyone finds this in work; there are other sources of a sense of self-worth, but for many individuals, and for society in general, work is one very important source of self-esteem and personal identity. We need, therefore, to consider the situation of those who do not or cannot work, but first, we will look briefly at those who work to excess.</p></sec><sec><title>2. Workers to excess</title><p>There are several types of excessive worker, in all of whom there may be substantial stresses related both to working and to not working. There are those who work for themselves, running their own business, who work very long hours to make the business a success, to earn more, or because they see the business as dependent on them and they cannot let go. They may be identified almost wholly by their work and may have little other life. Sickness or retirement may be a huge loss, for they have no other investment for their lives. PG Wodehouse portraits several examples of retired business men, very successful, very wealthy, desperately taking up a hobby with the same obsessive approach that characterised their business life [eg [<xref ref-type="bibr" rid="B12">12</xref>]].</p><p>There are those who are employed but are so involved in and committed to their job that they cannot keep away from their work. Some academics and some health workers are like this. In extreme cases this may create stresses in both the worker and their family, and may make sickness or retirement appear catastrophic. However, it may be possible to perpetuate their identity, for example by writing.</p><p>These groups elect to work excessively; a third group does not really choose to do so. They work very long hours, far more than they are officially paid for, because it is expected of them by the culture of the firm or wider society, and because they share a perception that keeping their job, gaining promotion, or moving to a better job, demands 150% commitment to work, and nothing else ranks of similar importance in their lives.</p><p>This attitude seems to be increasing in the UK, and perhaps elsewhere in Europe. I am told by my youngest son in Washington DC that it is very common, even characteristic of many jobs in the USA [<xref ref-type="bibr" rid="B13">13</xref>]. In this situation, people may be forced by societal pressures to identify as a 'worker', not necessarily in any particular trade or profession, equating long hours with importance and success. Perhaps for such workers, if financially secure, retirement (and even sickness), and relinquishing their worker identity may be a great relief.</p><p>A fourth group, with commitments beyond their means feel forced to work extra hours or additional jobs in order to cope financially. They may be on low wages, supporting uneconomic family members, or a family life-style incurring substantial debts. Both can be very stressful situations, and work may seem little more than slavery.</p></sec></sec><sec><title>C. People not in paid work</title><sec><title>1. Those who will not work</title><p>Society exhibits ambivalent attitudes towards the few who are wealthy enough not to work – they are both envied and resented, even though some fill their lives with 'good works' and are highly respected. But those without wealth who are perceived as unwilling to work are denigrated as 'loungers', 'loafers', 'work-shy'. There are, no doubt, adults who are fit and well and lazy, who will not work if they can get away with it, though I believe they are few. However, it is not always easy to distinguish these from those who have genuine reasons for not working – physical ill-health, mental ill-health or serious social stresses – or who cannot find work.</p></sec><sec><title>2. Those who are unemployed</title><p>Those who cannot find work, the unemployed seeking work, are put under increasing pressure. Many, for example in the North-East of England, have had to leave their families to find work in other regions; they cannot afford to move house, and, in any case, jobs have little security. These men (mostly) are working, but the stresses on them and their families are huge. It is not surprising if marriages break down and families split up. But for many men, unemployment is the worst option, bringing low income, low status, loss of purpose, loss of dignity, loss of identity.</p><p>The relationship of unemployment with illness is well established, though physical illness is likely to be more commonly cause than effect. It is likely to be a mixed picture with respect to depression and anxiety, the 'common mental disorders'. Following a recent review of all large-scale population studies since 1980, we can be confident that there is a substantial excess of these disorders associated with poor educational background and low material resources, as well as unemployment [<xref ref-type="bibr" rid="B14">14</xref>,<xref ref-type="bibr" rid="B15">15</xref>]. Many Family Doctors will be well aware of these things in their daily practice.</p></sec><sec><title>3. Sickness absence from work</title><p>The confusion of refusal to work, inability to find work and inability to work is very damaging, because there are many who can't easily work on account of acute illness, long-standing sickness, or serious disability, not all of which is obvious to the casual observer. At this point doctors play a major part in helping to distinguish reasons for not working, as, in many societies, they have a legitimising role for sickness absence from work or receipt of sick pay and disability pensions. The numbers involved are large. For example, in Sweden, in 2000, of 5.6 million adults of working age, 62.4% were in work, and only 4.7% were unemployed; but 11.8% were on sick-leave, and 21.1% were in receipt of disability pensions [<xref ref-type="bibr" rid="B16">16</xref>]. In the UK, in the three months March to May 2004, 1.7 scheduled working days were lost to sickness absence, involving 2.9% of all UK employees [<xref ref-type="bibr" rid="B17">17</xref>].</p><p>Sickness certification is the means by which society formally permits people to be not working without loss of dignity or loss of identity, but they have to fulfil certain conditions as part of the 'contract' with the community [<xref ref-type="bibr" rid="B18">18</xref>]. They must allow their claim of illness to be verified by a doctor, must refrain from work, and must behave appropriately. As all doctors know, certification is important but inherantly problematic. Short periods are usually excused verification, and specific industrial diseases (accidental injury, certain cancers, pneumoconioses, etc) raise different issues which will not be dealt with here.</p><p>But, sick-leave from work is often a complex clinical, personal and social phenomenon, the presenting complaint not necessarily being the sole or principle problem. We know for example, that sickness absence increases with adverse conditions at work [<xref ref-type="bibr" rid="B19">19</xref>]. These include pressure to increase the pace of work, conflicting demands, and low control over your own work. (It is interesting to note Ruskin's three conditions for being happy at work: they must be fit for it; they must not do too much of it; they must have a sense of success in it [<xref ref-type="bibr" rid="B6">6</xref>]). Problems outside work, in families and other relationships, may be important. However, if approached holistically, in the manner of Paul Tournier, these situations have positive potential for medical practice, requiring careful physical, psychological and social diagnosis, and prescription of a wide range of appropriate therapies, and personal and social interventions.</p></sec><sec><title>4. People with disabilities</title><p>Very few people with disabilities cannot work in any way; most, of course, have significant abilities, and want to work as far as they are able. But their ability to work often depends as much upon the environment, practical circumstances and attitudes of others, as upon their own limitations. We have gone some way to adapting work situations to permit people with limited sight, hearing or mobility to work, but there is much more to do.</p><p>Identity is a key issue, with two facets. First, they may suffer discrimination as non-workers, lacking the possibilities of identity from a profession, trade, job or worker-status. Second, they are labelled with an identity as 'disabled', or even as the type of disease or disability they suffer from. The worst stigmatising terms are often used in society as terms of general abuse, but it is also common to refer to people as 'a diabetic', 'an epileptic', 'a schizophrenic', and so on. It must be very hard to live with your identity as a disease category, though, of course, there are many individuals who have overcome all these things with heroic fortitude and great success.</p></sec><sec><title>5. Women</title><p>I pointed out earlier that most of the literature relating to work applies to men. From a current view-point, it is amazing how strong was the historical assumption that, in relation to paid work and economic activity, women could largely be ignored. Even Tournier's book of 1971 [<xref ref-type="bibr" rid="B4">4</xref>] seems to assume 'traditional' women's' roles. But worker-identity remains problematic for many women. Except where paid to undertake it for someone else's family, the traditional house-wife and mother role, whilst arguably the most important in all society, has never been awarded the status of a job, even less a career or profession, in spite of recent attempts to do so. Research classifications still cannot cope with it; traditionally married women were classified by their husband's job.</p><p>In my parents' generation, it was probably still 'normal', though by no means universal, for most women to expect no profession or career except that of housewife and mother. My mother was withdrawn from school by her father to look after the family on the grounds that a girl did not need education; she never 'worked', and always felt something important had been lost. This attitude towards women's work can still be found among men. Among women it is often implied: how often have you heard "I'm only a housewife"; no job, no status, low self-esteem. Yet, raising a family and running a household may be very demanding and, especially with larger families, an excellent training in organisational and managerial skills. We might do better to recruit some of these experienced 'housewives' into business and public service management, than many of the men we employ.</p><p>Of course, many of my generation have worked when their children were older, but, in spite of equal opportunities legislation, women get less top jobs, often get less pay than men doing the same job, and predominate in part-time jobs, which often carry poverty wages. Can this partially explain why women have more of the common mental disorders than men? In the British National Psychiatric Survey of 2000 [<xref ref-type="bibr" rid="B20">20</xref>], women were twice as likely to have obsessions, somatic symptoms, compulsions and phobias, and almost one and a half times more likely to suffer symptoms of fatigue, to have problems with sleep, and to have a total of neurotic symptoms above case-threshold. Mixed anxiety and depressive disorders in the age group 16–64 were 19% for women, 13% for men.</p><p>Of course, nowadays, many women are training and pursuing careers of all sorts. As we well know, many are postponing (where not eschewing altogether) having children until well into their thirties, with long-term consequences which may not be entirely satisfactory. Their 'worker-identity' may be a real gain, with improved self-esteem, but surely we still need to find a way of recognising the 'housewife and mother' role, the 'domestic manager' role, and investing it with a clear, high-value identity and due status, especially as many women find themselves doing 'double duty' – working both outside and inside the home!</p></sec><sec><title>6. People who have retired</title><p>The idea of retirement is relatively modern and is characteristic of industrial societies [<xref ref-type="bibr" rid="B21">21</xref>]. Pre-industrial societies always had a category of people called 'old', for whom the role expectations were different from those of younger adults. But new roles, usually requiring less physical strength, were aquired gradually and at no specific age, a transition to different but equally valued roles – men might no longer be hunters and farmers, but elders and priests. Although retired people in modern industrialised societies have many roles, and undertake many responsibilities, such as looking after grand-children, these may be considered rather trivial, and not highly valued except, perhaps, by the parents of those children.</p><p>The situation is likely to change. Where unemployment is low, as in the UK, we are short of workers, short of skills and experience in many trades and professions, and need people to stay in work longer. People over 65 are now much fitter and healthier, and are mostly able to work; many want to work, though not necessarily full-time. Yet ageist attitudes persist; there is widespread prejudice against employing older workers, even in their fifties. With the increase in older people, the financial burden of pensions has become a big political issue, and many private sector pension schemes have failed, with workers losing their pensions.</p><p>Although the traditions and social arrangements for retirement vary between countries, the main issues are generally the same. Many perceptions of retirement for men are very negative; "a first step towards social dependence" [<xref ref-type="bibr" rid="B22">22</xref>], of which the pension is a symbol. Media reports referring merely to 'a pensioner', with no name, no personal identity, tend to reinforce negative images. Like 'the elderly', it is not really pejorative, but it does imply dependence. It is a non-identity, symbolic of low personal value.</p><p>In Townsend's classic 1954–55 study of working class Bethnal Green [<xref ref-type="bibr" rid="B23">23</xref>], although poor pensions were a problem, men complained just as much of boredom and a sense of uselessness – of being unwanted. He reported "Among the retired, there was scarcely a single person in favour of retirement." I think that this would not be as true now in any group; many have good pensions, and opportunities for alternative activity have increased dramatically for most people. But for many others, it remains an experience of loss, including a profound loss of identity.</p><p>Tournier [<xref ref-type="bibr" rid="B4">4</xref>] recognised this common sense of loss with its consequences for ill-health and unhappiness in older age. He advocated deliberate preparation for retirement, now more common, and the attention at earlier ages to interests outside work which will sustain life in retirement. He thought that unstructured hobbies did not go far enough, and proposed the idea of a 'second career', by which interests and activities in retirement should be structured, should have "coherence and continuity", like a previous career, so that it resembles 'work'. Nowadays, it might be a third, fourth or fifth career, but the principle is the same.</p><p>Some people do obtain paid work after retirement, though often part-time, and often of a lower status than their previous work. Some can use their skills and experience in consultancy, teaching or writing related to their main career, and retain their work-related identity. Some, especially retiring early, re-train and take up a new paid career in a more satisfying field than their previous one. But many, assuming an adequate pension, do not look for paid work, but, very much as Tournier suggested, pursue a new career in voluntary work with NGOs. In the UK at least, NGOs are largely dependent upon an army of retired people. To be chairman or a trustee of a major charitable organisation requires abilities, skills and experience akin to those needed for senior positions in private business or public service, and carries similar roles, responsibilities, commitments and risks.</p><p>These positions carry some status in society, but I do not think that they confer a recognisable work-type identity. Perhaps those that undertake them for a long period are people who do not crave work-related identity, but find their personal identity from other sources. They include many women, who may have been denied work and career identity in their main adult years, but have achieved a satisfying sense of personhood at a deeper level in terms of relationships, family, and community service.</p><p>Since each country has an official 'retirement age' when pensions will normally begin, and when not-working is accepted without stigma, 'early retirement' is likely to have some special features. It became common in the UK about twenty-five years ago, partly in response to high levels of unemployment, and the rate of technological change. It was sometimes a matter of individual choice, encouraged by generous early pension arrangements. But it was often forced on people through redundancy, and 'ageism' among employers, which precluded people in their fifties getting new jobs. Some retired early in response to poor health; in some cases a partial reason, but conveying legitimation. For society there were huge losses in skills and experience, now openly regretted in many fields. Also regretted is the huge extra financial burden of pensions on companies and public institutions, for no productivity. Retirement at 50 incurs an expected pensioned life of 30 years or more.</p><p>Physical health problems may mostly precede early retirement, but mental health problems may be either causes or consequences of retirement. There is evidence of causes – 20% of early retirements from the UK National Health Service are for psychiatric reasons [<xref ref-type="bibr" rid="B24">24</xref>]. There is also evidence of consequences – involuntary early retirement appears to increase the prevalence of the common mental disorders, but planned, voluntary retirement appears to reduce them [<xref ref-type="bibr" rid="B25">25</xref>,<xref ref-type="bibr" rid="B26">26</xref>].</p><p>Data from the British National Psychiatric Survey of 2000 are very interesting [<xref ref-type="bibr" rid="B27">27</xref>,<xref ref-type="bibr" rid="B28">28</xref>]. They showed a significant drop in prevalence of the common mental disorders (mostly depression, anxiety or both) at the age of 65 for men, but not for women. For men, total symptom counts of 18 on the Clinical Interview Schedule (CIS) were 7.3% at age 60–64 and 1.5% at age 65–69. 'Depression', was 3% at 60–64, and 0.3% at 65–69. It is not a cohort effect (from analysis of the 1993 and 2000 surveys together), and no variables explain it statistically other than age. Men who retired early had high rates, indeed higher rates than employed men, until 65, when they suddenly achieved similar low rates. Men still employed after retirement age not only had low rates then, but had relatively low rates in their fifties and early sixties. But the lowest rates were for those who retired at 65. For women, the rates were higher than men at all ages, peaked at age 50–54, then slowly diminished with no large reduction at either age 60 or 65.</p><p>These are British figures; it would be interesting to have this replicated in other countries, but few have appropriate surveys. It seems that early retirement is bad for men, but retirement at the age legitimated by society is very good for men. Why? Perhaps legitimation is the secret; at age 65 perhaps, the generality of men can relax, no longer needing to justify themselves as workers; can, perhaps, 'be' more than 'do'. And, perhaps, people like me, who make such a fuss about retirement, are out of step, and need to learn something about personal identity which has nothing to do with being 'something' in work.</p></sec></sec><sec><title>D. Conclusion: identity – what am I? – who am I?</title><p>Who we are and what we are, are very important questions for all of us, with social, psychological and spiritual dimensions. Without a clear sense of personal identity, it is difficult to have the self-esteem we need to function well as independent people in inter-dependent society. Without a clear sense of personal identity we are vulnerable to psychological injury, at risk of anxiety and depression, and social disengagement. Without a clear sense of personal identity we cannot easily respond to love with love, to accept forgiveness, to start again after failure.</p><p>For many people, work, a job, a profession or a trade, provide an important source of personal identity; for some it may be the only significant source. Even for those for whom this is minimal, being a worker is important for the other benefits work mediates. Both before and after retirement age, continuing some work, paid or not, perhaps part-time, but definitely 'work', may be important both for individuals and the community. And for those with specialist skills and experience, is there not also a moral duty to continue to make a contribution for as long as it is needed and remains appropriate, whether financially rewarded or not?</p><p>However, work as the source of our identity is fundamentally inadequate, because few of us can claim that identity for ever. Sickness, disability, redundancy, retirement, age, all threaten an identity built upon work. But it is not the only source; we all do have other identities. We are sons and daughters, mothers and fathers, grandmothers and grandfathers and great-aunts, friends and neighbours; identities, built upon personal relationships, and surely more fundamentally important.</p><p>Ultimately personal identity and self-esteem are closely bound up together, and derive from a sense of personal value, of personal worth, of being needed, of being loved for what you are, not just for what you do. This is true health and wholeness, and, no doubt, much depends upon our experience of parenting as children. It is also a spiritual issue. Christians, and others of faith, will claim the ultimate value of a human being loved by God and therefore of infinite worth; others will be satisfied with a non-religious belief that each individual human being has intrinsic equal and great value.</p><p>Personally, perhaps I should think again about retirement, and accept my rapidly developing identity as a grandfather; and maybe mongongo nuts will grow in the greenhouse.</p></sec><sec><title>Note</title><p>The first version of this paper was given to the 56th 'réunion internationale de Médecine de la Personne' in Drübeck, Saxony, Germany in August 2004.</p></sec> |
Muscle dissatisfaction in young adult men | <sec><title>Backround</title><p>Appearance concerns are of increasing importance in young men's lives. We investigated whether muscle dissatisfaction is associated with psychological symptoms, dietary supplement or anabolic steroid use, or physical activity in young men.</p></sec><sec><title>Methods</title><p>As a part of a questionnaire assessment of health-related behaviors in the population-based FinnTwin16 study, we assessed factors associated with muscle dissatisfaction in 1245 men aged 22–27 using logistic regression models.</p></sec><sec><title>Results</title><p>Of men, 30% experienced high muscle dissatisfaction, while 12% used supplements/steroids. Of highly muscle-dissatisfied men, 21.5% used supplements/steroids. Mean body mass index, waist circumference, or leisure aerobic activity index did not differ between individuals with high/low muscle dissatisfaction. Muscle dissatisfaction was significantly associated with a psychological and psychosomatic problems, alcohol and drug use, lower height satisfaction, sedentary lifestyle, poor subjective physical fitness, and lower life satisfaction.</p></sec><sec><title>Conclusion</title><p>Muscle dissatisfaction and supplement/steroid use are relatively common, and are associated with psychological distress and markers of sedentary lifestyle.</p></sec> | <contrib contrib-type="author" corresp="yes" id="A1"><name><surname>Raevuori</surname><given-names>Anu</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>anu.raevuori@helsinki.fi</email></contrib><contrib contrib-type="author" id="A2"><name><surname>Keski-Rahkonen</surname><given-names>Anna</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I2">2</xref><xref ref-type="aff" rid="I3">3</xref><email>anna.keski-rahkonen@helsinki.fi</email></contrib><contrib contrib-type="author" id="A3"><name><surname>Bulik</surname><given-names>Cynthia M</given-names></name><xref ref-type="aff" rid="I4">4</xref><email>cbulik@med.unc.edu</email></contrib><contrib contrib-type="author" id="A4"><name><surname>Rose</surname><given-names>Richard J</given-names></name><xref ref-type="aff" rid="I5">5</xref><email>rose@indiana.edu</email></contrib><contrib contrib-type="author" id="A5"><name><surname>Rissanen</surname><given-names>Aila</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>aila.rissanen@medi.inet.fi</email></contrib><contrib contrib-type="author" id="A6"><name><surname>Kaprio</surname><given-names>Jaakko</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I6">6</xref><email>jaakko.kaprio@helsinki.fi</email></contrib> | Clinical Practice and Epidemiology in Mental Health : CP & EMH | <sec><title>Background</title><p>Men are increasingly concerned about the way they look: a moderately or extremely muscular body is widely accepted as an ideal body shape for young men, creating discrepancy between the actual and desired body size and shape [<xref ref-type="bibr" rid="B1">1</xref>-<xref ref-type="bibr" rid="B5">5</xref>].</p><p>Young men who are dissatisfied with their body shape and musculature may be more likely to turn to bodybuilding, dietary supplements, and anabolic steroids to shape their bodies [<xref ref-type="bibr" rid="B6">6</xref>,<xref ref-type="bibr" rid="B7">7</xref>]. The pursuit of muscularity has been associated with significant behavioural and psychological problems [<xref ref-type="bibr" rid="B4">4</xref>]. However, it is also plausible that exercise, weight training, and even competitive bodybuilding may actually improve men's body image [<xref ref-type="bibr" rid="B8">8</xref>-<xref ref-type="bibr" rid="B10">10</xref>].</p><p>Most of the existing studies in this area have focused on defined groups of young men, such as athletes and college students: it is unclear how generalizable their results are. Very little is known about the association of muscle dissatisfaction with physical activity, dietary supplement or anabolic steroid use, or psychological symptoms in young men from the general population. The aim of this study was to describe the psychological and behavioral correlates of muscle dissatisfaction in a population study of young Finnish males.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>Sample</title><p>The data reported are from a population-based twin study, FinnTwin16, and comprised 1245 adult male twins from the 1977 to 1979 birth cohorts, for whom a questionnaire included items on muscle dissatisfaction (MD) and supplement and steroid use as described below. Data collection was approved by the Ethics committees of the University of Helsinki and Indiana University. The present study is based on the fourth wave of the study (response rate 83%) that assessed mental and physical health, weight control, dietary and exercise habits, social relationships, and alcohol and drug use. Mean age at response was 23.8 y (SD 23.7–23.9), mean BMI was 23.9 (SD 3.1), mean waist circumference 85.3 cm (SD 9.4) and mean height 179.4 cm (SD 6.6). A total of 70% of men studied were at normal body weight (BMI 20.0–24.9), while 26% were overweight (BMI 25.0–29.9) and 4% were obese (BMI 30.0–34.9). Of the study subjects 30% of men were 180–184 cm tall, 28% were 175–179 cm, 17% were 170–174 cm, 15% were 185–189 cm, and 6% were 190 cm or more while 4% were 169 cm or less. At the time of the study, 53% of men were not studying, while 21% attended a college or university, 18% attended a polytechnic, 4% attended high school, vocational school or vocational college and 4% studied in some other school not mentioned above. By this age, 5% of men had reached a higher academic degree while 8% had a degree from polytechnic and 42% from a vocational school.</p></sec></sec><sec><title>Measures</title><sec><title>Muscle dissatisfaction</title><p>Muscle dissatisfaction (MD) was assessed based on the following questionnaire item: "I would like to be more muscular". The six original responses were recategorized into "Always or Commonly" (High MD), "Often" (Intermediate MD), and "Sometimes, Rarely, or Never" (Low MD).</p></sec><sec><title>Supplement use</title><p>Supplement and/or steroid use (from here on, referred to as supplement use) was assessed based on a question: "Have you ever used hormone preparations, dietary supplements, or other special preparations in order to increase your muscle mass or to maximise the effects of gym training?". The answering alternatives were: "Yes, continuously during the last three months", "Yes, continuously at least three months sometimes earlier", "I have sometimes tried", and "I have never tried nor used". To differentiate established use, supplement use was dichotomized: men who had engaged in supplement use for ≥3 months either recently or previously were considered supplement users, while others were not considered as users.</p></sec><sec><title>Height satisfaction</title><p>Height satisfaction was assessed based on a following questionnaire item: "I am satisfied with my height". The answering alternatives were: "Always", "Commonly", "Often", "Sometimes", "Rarely", or "Never".</p></sec><sec><title>Anthropometrics</title><p>Values for height (cm), waist circumference (cm) and weight (kg) were based on self-reported data, from which body mass indices (BMI) were calculated. The correlations between self-reported and measured values were excellent, >0.9 for height and BMI, and 0.84 for waist circumference in a subsample of these data [<xref ref-type="bibr" rid="B11">11</xref>,<xref ref-type="bibr" rid="B12">12</xref>].</p></sec><sec><title>Eating disorder inventory</title><p>The questionnaire included the original items of three subscales of the Eating Disorder Inventory-1 (EDI) [<xref ref-type="bibr" rid="B13">13</xref>,<xref ref-type="bibr" rid="B14">14</xref>]: Body dissatisfaction (BD, 9 items), Drive for thinness (DT, 7 items), and Bulimia (B, 7 items). A Finnish version of this instrument has been translated and validated (Charpentier, in preparation). Scoring for EDI used a range of 1–6 for responses in the questionnaire. DT and B were reasonably normally distributed and were used as continuous variables. However, the distribution of BD was highly skewed, and therefore categorized based on tertiles of the distribution.</p></sec><sec><title>Physical activity and fitness</title><p>The participants were asked to characterize their subjective physical fitness using the alternatives: "very good", "fairly good", "satisfactory", "fairly poor" and "very poor". For the analyses, the variable was recategorized into three classes: "Fairly/very good", "satisfactory", "fairly/very poor".</p><p>Leisure aerobic activity index (Metabolic Equivalent, MET h × d -1, work metabolic rate divided by resting metabolic rate) was calculated from the product of self reported intensity × duration × days/year frequency. We used MET-values 4 (light aerobic activity, walking), 6 (moderate aerobic activity, brisk walking), 10 (intense aerobic activity, jogging), 13 (very intense activity, running).</p><p>We also asked for daily time spent on sedentary leisure interests, for example driving, TV viewing, computer game and Internet use.</p></sec><sec><title>Illicit drug use</title><p>Drug use was assessed based on answers to a question "Have you ever used hash, marijuana or other drugs, or have you for example sniffed glue?". The response alternatives were: "Never", "1–3 times", "4–9 times", "10–19 times", "Over 20 times". For the analyses, the variable was recategorized into three classes: "Never used", "Experimental or mild drug use (1–19 trials)", and "Established drug use (over 20 trials)".</p></sec><sec><title>Other psychological measures</title><p>The General Health Questionnaire (GHQ-12) is a screening instrument designed to detect psychiatric disorders in community settings and non-psychiatric clinical settings [<xref ref-type="bibr" rid="B15">15</xref>]. It has been validated for the Finnish population [<xref ref-type="bibr" rid="B16">16</xref>,<xref ref-type="bibr" rid="B17">17</xref>]. Rutgers Alcohol Problem Index (RAPI) is a 23-item, unidimensional self-administered screening tool for assessing adolescent problem drinking [<xref ref-type="bibr" rid="B18">18</xref>]. Psychosomatic symptom frequencies were assessed based on answers to six items: stomach aches, tension and nervousness, difficulties in falling asleep or nightly awakenings, headache, lower back pain, and pain in the neck and shoulder area [<xref ref-type="bibr" rid="B19">19</xref>]. Each item contained four response alternatives which were: "Rarely or never", "About once a month", "About once a week, and "Almost every day".</p><p>The questionnaire also included the Life Satisfaction Scale, which covers four items: interest in life, happiness, general ease of living, and loneliness [<xref ref-type="bibr" rid="B20">20</xref>]. The scale correlates highly with the Beck depression inventory (r > 0.6) [<xref ref-type="bibr" rid="B21">21</xref>].</p></sec><sec><title>Sociodemographic factors</title><p>We assessed place of residence (rural vs. urban) using five response categories based on population density. Rural areas were considered to include both rural villages and countryside.</p><p>Because the men studied were relatively young and many were still studying, we assessed current educational level in addition to highest educational attainment. The question asked was: "Do you attend a school or college at the moment?" The response alternatives were "I do not attend a school at the moment", "I attend a high school, vocational school or in vocational college", "I attend a polytechnic", "I attend a college or university" and "Other". About educational attainment we asked: "Which schools/degrees have you completed?", using similar response categories as in the previous question.</p></sec><sec><title>Analysis</title><p>We investigated associations between independent and dependent variables using cross-tabulations, Pearson's chi-squared test of independence and logistic regression. For continuous variables, odds ratios were calculated for both a unit increase in the dependent variable, and by using a median split to form a dichotomized variable. All analyses were controlled for clustered sampling [<xref ref-type="bibr" rid="B22">22</xref>] within the twin pair using Stata (Version 8.0) [<xref ref-type="bibr" rid="B23">23</xref>].</p></sec></sec><sec><title>Results</title><sec><title>Characteristics of study sample by muscle dissatisfaction status</title><p>Of all respondents, 29.9% experienced high MD, 14.3% intermediate MD, and 55.8% low MD (see Table <xref ref-type="table" rid="T1">1</xref>). The subjects of our study were relatively similar in age, height, BMI and waist circumference irrespective of their MD and supplement use status (Table <xref ref-type="table" rid="T1">1</xref>). BMI and waist were correlated at r = 0.79 among low MD subjects, r = 0.80 among intermediate MD subjects and r = 0.75 among high MD subjects.</p><table-wrap id="T1" position="float"><label>Table 1</label><caption><p>Characteristics of the study sample by muscle dissatisfaction (MD) status. Means and 95% CI are given for continuous traits</p></caption><table frame="hsides" rules="groups"><thead><tr><th/><th align="center">High MD</th><th align="center">Intermediate MD</th><th align="center">Low MD</th></tr></thead><tbody><tr><td align="left">n(% of population)</td><td align="center">372 (29.9%)</td><td align="center">178 (14.3%)</td><td align="center">695 (55.8%)</td></tr><tr><td align="left">SU, n (Proportion of MD subgroup subjects)</td><td align="center">80 (21.5%)</td><td align="center">20 (11.2%)</td><td align="center">50 (7.2%)</td></tr><tr><td align="left">Mean age, years</td><td align="center">23.8 (23.8–23.9)</td><td align="center">23.9 (23.8–23.9)</td><td align="center">23.8 (23.8–23.9)</td></tr><tr><td align="left">Mean BMI, kg/m<sup>2</sup></td><td align="center">23.5 (23.2–23.8)</td><td align="center">23.7. (23.2–24.2)</td><td align="center">23.9 (23.6–24.1)</td></tr><tr><td align="left">Waist circumference, cm</td><td align="center">84.1 (83.2–85.1)</td><td align="center">85.0 (83.4–86.5)</td><td align="center">84.5 (83.7–85.2)</td></tr><tr><td align="left">Mean height, cm</td><td align="center">180.0 (179.2–180.7)</td><td align="center">180.4 (179.3–181.5)</td><td align="center">179.0 (178.4–179.5)</td></tr></tbody></table><table-wrap-foot><p>Note: No statistically differences between high, intermediate and low muscle disatisfied participants for any traits</p></table-wrap-foot></table-wrap></sec><sec><title>Height satisfaction</title><p>Of respondents, 52.5% reported being always satisfied with their height as 28.5% reported this commonly and 8.4% often. A total of 4.7% of men were height satisfied only sometimes, 3.5% rarely and 2.4% never.</p><p>Height satisfaction decreased significantly from low to high MD even when adjusted with actual height (OR for always versus commonly 1.6, 95% CI 1.3–2.1, OR for always versus often 1.9, 95% CI 1.4–2.7, OR for always versus sometimes 2.9, 95% CI 1.8–4.8, OR for always versus rarely 3.9, 95% CI 2.3–6.9 and OR for always versus never height satisfied 3.2, 95% CI 1.4–7.6). Those whose actual height was 174 cm or less were significantly more often height dissatisfied (OR 3.9, 95% CI 2.9–5.2), but no more dissatisfied with their musculature (<italic>x</italic><sup>2 </sup>= 1.78, p = 0.41) than their taller peers.</p></sec><sec><title>Sociodemographic correlates of muscle dissatisfaction</title><p>The place of residence was not associated with MD, but living in a rural area (rural village or countryside) was associated with less supplement use (OR 0.4, 95% CI 0.2–0.8). Current school attendance or a level of educational attainment at this age were not significantly associated with MD.</p></sec><sec><title>Psychological correlates of muscle dissatisfaction</title><p>The basic psychometric characteristics of MD subgroup subjects are given in Table <xref ref-type="table" rid="T2">2</xref>. The correlates of MD in the univariate models are presented in Table <xref ref-type="table" rid="T5">5</xref>. Among young men from the general population, psychological distress as measured by GHQ, psychosomatic symptoms, life satisfaction, drive for thinness and body dissatisfaction were statistically significantly associated with muscle dissatisfaction. These findings were independent of the place of residence and current educational status. Body Dissatisfaction had the strongest association with MD of the three EDI subscales. Eating Disorder Inventory's Drive for thinness and Bulimia were also significantly associated with MD. All three EDI subscales remained significantly associated with MD, even when controlling in regression models for measures of depressive symptoms, i.e. GHQ or life satisfaction.</p><table-wrap id="T2" position="float"><label>Table 2</label><caption><p>Basic psychological characteristics of high/intermediate/low muscle dissatisfied (MD) men: mean scores and SDs.</p></caption><table frame="hsides" rules="groups"><thead><tr><th/><th align="left">high MD</th><th align="left">Intermediate MD</th><th align="left">Low MD</th></tr></thead><tbody><tr><td align="left">n</td><td align="left">372</td><td align="left">178</td><td align="left">695</td></tr><tr><td align="left">Body dissatisfaction<sup>1</sup></td><td align="left">22.2 (6.4)</td><td align="left">21.1 (6.4)</td><td align="left">18.6 (5.4)</td></tr><tr><td align="left">Drive for thinness<sup>1</sup></td><td align="left">17.4 (4.5)</td><td align="left">16.9 (4.3)</td><td align="left">16.2 (4.1)</td></tr><tr><td align="left">Bulimia<sup>1</sup></td><td align="left">10.2 (2.6)</td><td align="left">10.0 (3.0)</td><td align="left">9.5 (2.5)</td></tr><tr><td align="left">RAPI<sup>2</sup></td><td align="left">30.9 (9.8)</td><td align="left">29.0 (8.5)</td><td align="left">28.2 (8.1)</td></tr><tr><td align="left">Life satisfaction<sup>3</sup></td><td align="left">9.2 (3.2)</td><td align="left">9.1 (3.0)</td><td align="left">8.1 (2.8)</td></tr><tr><td align="left">Psychosomatic symptom score</td><td align="left">10.2 (3.1)</td><td align="left">9.9 (2.9)</td><td align="left">9.2 (2.8)</td></tr><tr><td align="left">General Health Questionnaire<sup>4</sup></td><td align="left">2.0 (2.9)</td><td align="left">1.7 (2.5)</td><td align="left">1.0 (1.9)</td></tr></tbody></table><table-wrap-foot><p><sup>1 </sup>Eating Disorder Inventory[13]</p><p><sup>2 </sup>Rutgers Alcohol Problem Index[18]</p><p><sup>3 </sup>higher score → lower life satisfaction</p><p><sup>4 </sup>higher score → poorer mental health</p></table-wrap-foot></table-wrap></sec><sec><title>Physical activity and fitness</title><p>Higher leisure aerobic activity was not statistically significantly associated with MD (OR 1.0, 95% CI 0.99–1.04) even though the leisure aerobic activity increased categorically with muscle dissatisfaction (Table <xref ref-type="table" rid="T4">4</xref>). Instead, MD was associated with longer daily duration (>2 hours/day) of sedentary leisure interests like reading, Internet and TV viewing (OR 1.9, 95% CI 1.12–3.33). MD was also associated with poorer subjective physical fitness, the OR for satisfactory versus good/very good physical fitness being 1.1 (95% CI 0.8–1.3), and the OR for poor/very poor versus good/very good physical fitness being 2.5 (95% CI 1.7–3.9).</p><table-wrap id="T3" position="float"><label>Table 3</label><caption><p>Drug use and supplement use of high/intermediate/low muscle dissatisfied (MD) men based on categoric measures.</p></caption><table frame="hsides" rules="groups"><thead><tr><th/><th align="left">High MD</th><th align="left">Intermediate MD</th><th align="left">Low MD</th></tr></thead><tbody><tr><td align="left">Drug use (n, % of MD subgroup)</td><td/><td/><td/></tr><tr><td align="left"> Never</td><td align="left">258 (69.4%)</td><td align="left">122 (68.5%)</td><td align="left">524 (75.4%)</td></tr><tr><td align="left"> Experimental or mild use<sup>1</sup></td><td align="left">82 (22.0%)</td><td align="left">46 (25.8%)</td><td align="left">130 (18.7%)</td></tr><tr><td align="left"> Established use<sup>2</sup></td><td align="left">32 (8.6%)</td><td align="left">9 (5.1%)</td><td align="left">40 (5.8%)</td></tr><tr><td align="left">Supplement Use (n, % of MD subgroup)</td><td/><td/><td/></tr><tr><td align="left"> SU</td><td align="left">80 (21.5%)</td><td align="left">20 (11.2%)</td><td align="left">50 (7.2%)</td></tr><tr><td align="left"> Non-SU</td><td align="left">292 (78.5%)</td><td align="left">158 (88.8%)</td><td align="left">645 (92.8%)</td></tr></tbody></table><table-wrap-foot><p><sup>1 </sup>1–19 trials</p><p><sup>2 </sup>20 or more trials</p></table-wrap-foot></table-wrap><table-wrap id="T4" position="float"><label>Table 4</label><caption><p>Self-reported leisure aerobic activity and subjective physical fitness of high/intermediate/low muscle dissatisfied men</p></caption><table frame="hsides" rules="groups"><thead><tr><th/><th align="left">High MD</th><th align="left">Intermediate MD</th><th align="left">Low MD</th></tr></thead><tbody><tr><td align="left">Leisure aerobic activity (MET) mean value (95% CI)</td><td align="left">5.3 (4.7–5.9)</td><td align="left">5.2 (4.4–5.9)</td><td align="left">4.9 (4.5–5.3)</td></tr><tr><td align="left">Subjective physical fitness n (% of MD subgroup)</td><td/><td/><td/></tr><tr><td align="left"> Fairly/very good</td><td align="left">234 (62.9%)</td><td align="left">109 (61.2%)</td><td align="left">474 (68.2%)</td></tr><tr><td align="left"> Satisfactory</td><td align="left">96 (25.8%)</td><td align="left">52 (29.2%)</td><td align="left">189 (27.2%)</td></tr><tr><td align="left"> Fairly/very bad</td><td align="left">38 (10.2%)</td><td align="left">14 (7.9%)</td><td align="left">26 (3.7%)</td></tr></tbody></table></table-wrap><table-wrap id="T5" position="float"><label>Table 5</label><caption><p>Psychological correlates of muscle dissatisfaction from univariate models. Odds ratios were calculated both per unit increase and on median split (high vs. low) increase.</p></caption><table frame="hsides" rules="groups"><thead><tr><th/><th align="center">OR (per unit)</th><th align="center">95% CI</th><th align="center">OR, high vs low</th><th align="center">95% CI</th><th align="center">p-value</th></tr></thead><tbody><tr><td align="left"><italic>General psychological measures</italic></td><td/><td/><td/><td/><td/></tr><tr><td align="left">GHQ<sup>1</sup></td><td align="center">1.18</td><td align="center">1.12–1.23</td><td align="center">-</td><td align="center">-</td><td align="center"><0.0001</td></tr><tr><td align="left">Psychosomatic symptom score<sup>1</sup></td><td align="center">1.11</td><td align="center">1.07–1.16</td><td align="center">2.65</td><td align="center">1.89–3.72</td><td align="center"><0.0001</td></tr><tr><td align="left">Life satisfaction score<sup>1</sup></td><td align="center">1.12</td><td align="center">1.08–1.17</td><td align="center">2.55</td><td align="center">1.90–3.44</td><td align="center"><0.0001</td></tr><tr><td align="left"><italic>Substance use and abuse</italic></td><td/><td/><td/><td/><td/></tr><tr><td align="left">RAPI<sup>1,2</sup></td><td align="center">1.03</td><td align="center">1.02–1.04</td><td align="center">2.26</td><td align="center">1.59–3.22</td><td align="center"><0.0001</td></tr><tr><td align="left">Drug use</td><td/><td/><td/><td/><td/></tr><tr><td align="left"> Never</td><td align="center">1.0 (reference)</td><td/><td/><td/><td/></tr><tr><td align="left"> Experimental or mild (1–19 trials)</td><td align="center">1.28</td><td align="center">0.99–1.66</td><td align="center">-</td><td align="center">-</td><td align="center">0.06</td></tr><tr><td align="left"> Established use (≥20 trials)</td><td align="center">1.51</td><td align="center">0.95–2.39</td><td/><td/><td align="center">0.08</td></tr><tr><td align="left"><italic>Eating disorder inventory</italic></td><td/><td/><td/><td/><td/></tr><tr><td align="left">Drive for thinness<sup>1</sup></td><td align="center">1.06</td><td align="center">1.03–1.09</td><td align="center">2.60</td><td align="center">1.70–3.95</td><td align="center"><0.0001</td></tr><tr><td align="left">Bulimia<sup>1</sup></td><td align="center">1.09</td><td align="center">1.04–1.13</td><td align="center">2.09</td><td align="center">1.41–3.08</td><td align="center">0.0002</td></tr><tr><td align="left">Body dissatisfaction</td><td/><td/><td/><td/><td/></tr><tr><td align="left">1<sup>st </sup>tertile</td><td align="center">1.0 (reference)</td><td/><td/><td/><td/></tr><tr><td align="left">2nd tertile</td><td align="center">1.99</td><td align="center">1.46–2.68</td><td align="center">-</td><td align="center">-</td><td align="center"><0.0001</td></tr><tr><td align="left">3rd tertile</td><td align="center">3.89</td><td align="center">2.94–5.16</td><td/><td/><td align="center"><0.0001</td></tr></tbody></table><table-wrap-foot><p><sup>1 </sup>Dichotomy based on median split</p><p><sup>2 </sup>Rutgers Alcohol Problem Index[18]</p></table-wrap-foot></table-wrap></sec><sec><title>Substance and alcohol abuse and supplement use</title><p>Association with MD and alcohol use measured by RAPI was highly significant (Table <xref ref-type="table" rid="T5">5</xref>) (p < 0.0001). Association with MD and drug use was marginally significant, with the OR for never used nor tried versus experimental or mild use being 1.3 (95% CI 0.99–1.66) and OR for never used nor tried versus established use being 1.5 (95% CI 0.95–2.39), see Table <xref ref-type="table" rid="T5">5</xref>.</p><p>The proportion of supplement users was 12.0% of all respondents: supplement use increased monotonically with muscle dissatisfaction (Table <xref ref-type="table" rid="T3">3</xref>). Of non-supplement users, 26.7% experienced high MD compared to 53.0% of supplement users (<italic>x</italic><sup>2 </sup>= 47.0, p < 0.001).</p></sec><sec><title>Multivariable analyses of muscle dissatisfaction</title><p>In a multivariable model where the same variables as in the univariate model excluding the EDI subscale of Body Dissatisfaction were entered, GHQ score (OR 1.09, 95% CI 1.02–1.16), life satisfaction score (OR 1.65, 95% CI 1.15–2.39), and EDI Drive for thinness (OR 2.05, 95% CI 1.21–3.47) remained statistically significantly associated with MD (Table <xref ref-type="table" rid="T6">6</xref>). Psychosomatic symptom score remained marginally significantly associated with MD in a multivariate model (OR 1.5, 95% CI 0.99–2.21).</p><table-wrap id="T6" position="float"><label>Table 6</label><caption><p>Correlates of muscle dissatisfaction from multivariate models. Variables significantly associated with muscle dissatisfaction in univariate models excluding Eating Disorder Inventory's subscale Body Dissatisfaction were entered in the multivariable model. Variables presented in this table were adjusted for all other variables in the model.</p></caption><table frame="hsides" rules="groups"><thead><tr><th/><th align="center">OR</th><th align="center">95% CI</th><th align="center">p value</th></tr></thead><tbody><tr><td align="left"><italic>General psychological measures </italic>GHQ score</td><td align="center">1.09</td><td align="center">1.02–1.16</td><td align="center">0.006</td></tr><tr><td align="left">Psychosomatic symptom score<sup>1</sup></td><td align="center">1.48</td><td align="center">0.99–2.21</td><td align="center">0.06</td></tr><tr><td align="left">Life Satisfaction Score<sup>1</sup></td><td align="center">1.65</td><td align="center">1.15–1.2.39</td><td align="center">0.007</td></tr><tr><td align="left"><italic>Substance use and abuse</italic></td><td/><td/><td/></tr><tr><td align="left">RAPI<sup>1,2</sup></td><td align="center">1.33</td><td align="center">0.86–2.08</td><td align="center">0.20</td></tr><tr><td align="left">Drug use</td><td/><td/><td/></tr><tr><td align="left"> Never</td><td align="center">1.0 (reference)</td><td align="center">-</td><td align="center">-</td></tr><tr><td align="left"> Experimental or mild use (1–19 trials)</td><td align="center">1.01</td><td align="center">0.77–1.33</td><td align="center">0.93</td></tr><tr><td align="left"> Established use (≥20 trials)</td><td align="center">0.98</td><td align="center">0.57–1.69</td><td align="center">0.94</td></tr><tr><td align="left"><italic>Eating disorder inventory</italic></td><td/><td/><td/></tr><tr><td align="left">EDI subscales</td><td/><td/><td/></tr><tr><td align="left"> Drive for thinness<sup>1</sup></td><td align="center">2.05</td><td align="center">1.21–3.47</td><td align="center">0.007</td></tr><tr><td align="left"> Bulimia<sup>1</sup></td><td align="center">1.03</td><td align="center">0.65–1.64</td><td align="center">0.89</td></tr></tbody></table><table-wrap-foot><p><sup>1 </sup>These explanatory variables in the multivariable model were dichotomised using a median split.</p><p><sup>2 </sup>Rutgers Alcohol Problem Index[18]</p></table-wrap-foot></table-wrap></sec></sec><sec><title>Discussion</title><p>Muscle dissatisfaction was strongly associated with psychological distress such as depression and anxiety symptoms, as was hypothesized based on previous studies on body dissatisfaction and related disorders among both men and women [<xref ref-type="bibr" rid="B2">2</xref>,<xref ref-type="bibr" rid="B4">4</xref>,<xref ref-type="bibr" rid="B24">24</xref>]. In our study, lower general life satisfaction, psychosomatic symptoms, lower subjective physical fitness, sedentary lifestyles, lower height satisfaction and problems with alcohol and illicit drug use were also associated with muscle dissatisfaction. In addition, the correlation between muscle dissatisfaction and supplement use was strong. Compared to others, muscle dissatisfied men scored higher on all three EDI subscales referring to potential eating problems and dissatisfaction with one's body shape and weight.</p><p>In our current Western culture, appearance concerns are ubiquitous, particularly among young people [<xref ref-type="bibr" rid="B25">25</xref>,<xref ref-type="bibr" rid="B26">26</xref>]. It is generally accepted that popular culture, advertisement and even toys do not only reflect, but also foster appearance concerns. Very little is known about how these concerns manifest in young men. Our study demonstrates that male appearance concerns are potential indicators of other health problems, such as depressive and anxiety symptoms, and substance use and sedentary lifestyles.</p><p>In this population based study we wanted to focus on muscle dissatisfaction since the current young males' body ideal is perhaps primarily based on muscularity. Nevertheless, we do not wish to belittle other aspects of males' body dissatisfaction such as weight and body shape concerns and welcome more sophisticated study designs in the future. In our sample, those more dissatisfied with their musculature were also more dissatisfied with their height even though actual height did not differ between the groups and was adjusted in the regression model. This further signals how high dissatisfaction with one's body easily has more than a single dimension and thus perhaps merely reflects the overall discontent with the self. We also analysed whether relative shortness (height 174 cm or less; 21% of the sample) would be associated with muscle dissatisfaction, but that was not the case.</p><p>Muscle dissatisfaction does not appear to be more common among young men who exercise frequently than those who exercise little. There is previous evidence that even demanding training and fitness regimes are not associated with poor body image [<xref ref-type="bibr" rid="B8">8</xref>,<xref ref-type="bibr" rid="B27">27</xref>]. Rather, muscle dissatisfaction seems to be common in men who are also dissatisfied with many other aspects of their lives, and also tend to rate their subjective physical fitness less favorably than other men. There is evidence that regular exercise is associated with positive psychological health [<xref ref-type="bibr" rid="B28">28</xref>,<xref ref-type="bibr" rid="B29">29</xref>] and body satisfaction especially in males [<xref ref-type="bibr" rid="B30">30</xref>]. Although our study is not able to establish causality, future studies should address whether exercise interventions may also have positive effects on muscle dissatisfaction.</p><p>This study explored correlates of muscle dissatisfaction among men in the largest population sample yet reported. The strengths of our sample include its relatively large size, good population coverage, and high response rate. The battery we used was fairly extensive, and included many standardised and validated questionnaires.</p><p>Limitations of this study include self-reporting bias: however, an inherent bias in studies of this type. We also measured muscle dissatisfaction using only a single item. Although this item has not undergone extensive psychometric validation, the psychological associations of muscle dissatisfaction were robust and not sensitive to choice of cut-off points. Muscle dissatisfaction was highly correlated with the Body for Dissatisfaction subscale of the EDI, and is probably more appropriate for the measurement of body dissatisfaction in young men, because the EDI mainly focuses on female-specific areas of appearance concerns. Conceptually, muscle dissatisfaction can be regarded to be part of body dissatisfaction, which was therefore not entered into multivariate model like other variables significantly associated with the muscle dissatisfaction in the univariate model. We regret that for practical reasons, the inclusion of longer and more comprehensive measurement of male-specific appearance concerns, like the Muscle Appearance Satisfaction Scale [<xref ref-type="bibr" rid="B31">31</xref>], Swansea Muscularity Attitudes Questionnaire [<xref ref-type="bibr" rid="B32">32</xref>] or the Yale Brown Obsessive Compulsive Scale Modified for Body Dysmorphic Disorder (BDD-YBOCS) [<xref ref-type="bibr" rid="B33">33</xref>] was not an option in this large population study: we highly recommend assessing their use in future studies.</p><p>Further limitations of our study include that we had little information about our participants' body composition. By itself, BMI cannot separate lean and total fat body mass, but including waist circumference to these measures was an attempt to compensate for this limitation. Also, the question concerning supplement use did not separate the illegal use of anabolic steroids, other hormones and medicine-like substances from more widely used and legal substances like amino acids or creatinine. However, there is some evidence that creatinine and other nutritional supplements marketed as muscle mass enhancers in Finland have contained small amounts of anabolic steroids [<xref ref-type="bibr" rid="B34">34</xref>]. The same phenomenon has been noted internationally [<xref ref-type="bibr" rid="B35">35</xref>]. Thus supplement users would not always be aware of the true nature and actual content of the muscle enhancers of their choice.</p><p>Finally, we used a large twin sample to address muscle dissatisfaction in the general population. Although twins are in many respects similar to singletons [<xref ref-type="bibr" rid="B36">36</xref>] in particular for most psychiatric disorders [<xref ref-type="bibr" rid="B37">37</xref>,<xref ref-type="bibr" rid="B38">38</xref>], some previous studies suggest that adolescent twins may also be more physically active and physically fit than non-twins [<xref ref-type="bibr" rid="B39">39</xref>,<xref ref-type="bibr" rid="B40">40</xref>]. It is not known whether these small differences persist into adulthood. Further, twins may differ from non-twins in body size, which in childhood and adolescence is on average slightly smaller compared to singletons. Pietiläinen et al. [<xref ref-type="bibr" rid="B41">41</xref>] compared mean body height and mean body weight in the twin cohort used in this study (FT16) when the participants were aged 16–17 y to that of Finnish singletons at the same age and found that by the age of 17 y, the twin boys had reached the height of singleton boys, but still had lower BMIs than the 16.5 y singletons. Our sample suggests that this difference also disappears by young adulthood [<xref ref-type="bibr" rid="B14">14</xref>].</p></sec><sec><title>Conclusion</title><p>In summary, in young men, muscle dissatisfaction is a broad indicator of poor psychological well-being. It is associated with psychological symptoms and substance use. It was also associated with poor self-rated physical fitness and sedentary lifestyles. Future studies should assess prospectively factors that can improve young men's body image and muscle satisfaction and thus mitigate their psychological suffering.</p></sec><sec><title>Authors' contributions</title><p>ARa, AKR, ARi, JK, and RJR participated in the design of the study.</p><p>ARa, AKR, and JK carried out the statistical analyses and interpretation of the data.</p><p>ARa, AKR, JK, and CMB carried out the drafting of the manuscript</p><p>ARa, AKR, CMB, RJR, ARi, and JK carried out the critical revision of the manuscript for the important intellectual content.</p><p>ARa, AKR, ARi, and JK obtained funding for the study.</p><p>All authors read and approved the final manuscript.</p></sec> |
The effect of a worksite based walking programme on cardiovascular risk in previously sedentary civil servants [NCT00284479] | <sec><title>Background</title><p>A significant proportion of Europeans do not meet the recommendations for 30 mins of physical activity 5 times per week. Whether lower frequency, moderate intensity exercise alters cardiovascular disease (CVD) risk has received little attention. This study examined the effects of 45 minutes self-paced walking, 2 d· wk<sup>-1 </sup>on aerobic fitness, blood pressure (BP), body composition, lipids and C-Reactive Protein (CRP) in previously sedentary civil servants.</p></sec><sec sec-type="methods"><title>Methods</title><p>37 subjects (24 women) aged 41.5 ± 9.3 years were randomly assigned to either two 45 minute walks per week (walking group) or no training (control group). Aerobic fitness, body composition, blood pressure (BP), CRP and lipoprotein variables were measured at baseline and following 8 weeks. Steps counts were measured at baseline and during weeks 4 and 8 of the intervention.</p></sec><sec><title>Results</title><p>Compared to the control group, the walking group showed a significant reduction in systolic BP and maintained body fat levels (<italic>P </italic>< 0.05). There were no changes other risk factors. Subjects took significantly more steps on the days when prescribed walking was performed (9303 ± 2665) compared to rest days (5803 ± 2749; <italic>P </italic>< 0.001).</p></sec><sec><title>Conclusion</title><p>These findings suggest that walking twice per week for 45 minutes at ~ 62% HR<sub>max</sub>, improves activity levels, reduces systolic BP and prevents an increase in body fat in previously sedentary adults. This walking prescription, however, failed to induce significant improvements in other markers of cardiovascular disease risk following eight weeks of training.</p></sec> | <contrib id="A1" equal-contrib="yes" corresp="yes" contrib-type="author"><name><surname>Murphy</surname><given-names>Marie H</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>mh.murphy@ulster.ac.uk</email></contrib><contrib id="A2" equal-contrib="yes" contrib-type="author"><name><surname>Murtagh</surname><given-names>Elaine M</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>e_murtagh@yahoo.ie</email></contrib><contrib id="A3" equal-contrib="yes" contrib-type="author"><name><surname>Boreham</surname><given-names>Colin AG</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>ca.boreham@ulster.ac.uk</email></contrib><contrib id="A4" equal-contrib="yes" contrib-type="author"><name><surname>Hare</surname><given-names>Lesley G</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>l.hare@qub.ac.uk</email></contrib><contrib id="A5" equal-contrib="yes" contrib-type="author"><name><surname>Nevill</surname><given-names>Alan M</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>a.m.nevill@wlv.ac.uk</email></contrib> | BMC Public Health | <sec><title>Background</title><p>It has become increasingly clear that many of the chronic diseases we face today are associated fundamentally with the pervasive sedentariness of modern lifestyle. In Northern Ireland at least 2000 people die each year due to an inactive lifestyle and the avoidable cost of inactivity to the health service is estimated at £0.62 million each year [<xref ref-type="bibr" rid="B1">1</xref>]. Despite the proven benefits of regular exercise (Kohl, 2001) and published activity guidelines [<xref ref-type="bibr" rid="B2">2</xref>], the majority of westernised societies do not undertake enough physical activity to confer a health protective benefit. It is estimated that 66% of Europeans are physically active for less than 30 minutes daily [<xref ref-type="bibr" rid="B3">3</xref>]. Given the strong association between physical inactivity and cardiovascular disease[<xref ref-type="bibr" rid="B4">4</xref>], and the prevalence of inactivity in today's population, increasing levels of physical activity represents great potential for public health gain.</p><p>In the past decade, physical activity research has been characterised by the development of exercise prescriptions that are palatable to sedentary populations in the western world. First advocated by the American College of Sports Medicine [<xref ref-type="bibr" rid="B5">5</xref>] and subsequently adopted by numerous consensus panels [<xref ref-type="bibr" rid="B6">6</xref>,<xref ref-type="bibr" rid="B7">7</xref>] the recommendation that "every adult should accumulate 30 minutes of moderate intensity activity on most, preferably all, days of the week" is now widely accepted. However, compliance with these guidelines requires considerable commitment in terms of time spent exercising per week (≥ 150 minutes) and this may deter individuals from starting an exercise programme. Accordingly, randomised controlled trials are needed to evaluate the effects of smaller volumes of exercise on health.</p><p>Walking is eminently suited to population exercise prescription as it is easy to do, requires no special skills or facilities, and is achievable by virtually all age groups with little risk of injury [<xref ref-type="bibr" rid="B8">8</xref>]. There is some evidence that a training frequency of as low as two days per week may elicit improvements in cardiorespiratory fitness in the lower fitness categories [<xref ref-type="bibr" rid="B9">9</xref>]. To the authors' knowledge, only two studies have examined the effects of a twice-weekly training intervention using walking as the exercise mode. A twice-weekly, six-month progressive walking programme (45 minutes per session) improved indices of aerobic fitness, but failed to alter blood pressure (BP) in 60–70 year old women [<xref ref-type="bibr" rid="B10">10</xref>]. Similarly, walking 25 minutes, 2 d· wk<sup>-1 </sup>for six months, improved fitness in 79–91 year old females [<xref ref-type="bibr" rid="B11">11</xref>]. However, the effect of a low frequency walking prescription on fitness in younger adults remains to be investigated. Furthermore, the effects of low frequency, moderate intensity training on other markers of cardiovascular disease (CVD) risk have received very little attention. There is increasing recognition that traditional risk factors, such as hypercholesterolemia, hypertension and obesity, do not fully account for the occurrence of CVD. For example, almost one-half of the 1.3 million individuals who develop myocardial infarction in the United States each year have either normal or only moderately increased cholesterol concentrations [<xref ref-type="bibr" rid="B12">12</xref>]. Recently, C-Reactive Protein (CRP) has emerged as a novel risk factor for CVD. It has a strong association with cardiovascular risk, which is consistent, dose-related and independent. This is the first study to examine the effects of walking, without dietary intervention, on CRP in previously sedentary adults.</p><p>The present study evaluates the efficacy of a workplace walking intervention on physical activity levels and cardiovascular risk factors. An examination of the effects of self-paced walking twice per week will reveal whether individuals can achieve health benefits from outdoor walking performed with minimal investment in time. The purpose of the present study was to examine the effects of 45 minutes walking, 2 d· wk<sup>-1 </sup>on fitness, BP, body composition, lipids and CRP.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>Study design</title><p>This study was a randomised controlled trial, with subjects assigned to a walking or control group. The Research Ethics Committee at Queen's University, Belfast approved the study. Measurements were made at baseline (pre-intervention) and following eight weeks (post-intervention).</p></sec><sec><title>Subjects</title><p>Subjects were recruited from staff at the Northern Ireland Civil Service via internal email. All subjects completed a Health History questionnaire. Exclusionary criteria were a physically active lifestyle (defined as compliance with current physical activity guidelines (Pate, 1995), age > 65 years, resting BP > 159/99 mm Hg, total cholesterol > 6.2 mmol· L<sup>-1</sup>, fasting blood glucose > 7.0 mmol· L<sup>-1</sup>, body mass index (BMI) > 34.9 kg· m<sup>-2</sup>, current cigarette smokers, individuals with cardiovascular, pulmonary or metabolic disease, pain or discomfort in the chest, dizziness or heart murmur. In addition, individuals taking medication known to interfere with lipid metabolism, and females who were pregnant or planning to become pregnant in the following five months were excluded from taking part in the study. Thirty seven subjects (24 women) aged 41.5 ± 9.3 years were randomised to either a walking (n = 23; 16 women) or control group (n = 14; 8 women) on a 3 to 2 basis (3 walkers for every 2 controls). All subjects gave their written informed consent.</p></sec><sec><title>Measures</title><p>Height and body mass were recorded using a stadiometer (Seca model 770, Vogel & Halke; Hamburg, Germany) and scales (Seca model 707 digital physicians scale; Vogel & Halke, Hamburg, Germany) respectively. BMI was calculated by dividing body mass (kg) by height (m<sup>2</sup>). Waist measurements were made at the level of the trunk where the girth is minimal, i.e. the location where there was a noticeable indentation of the trunk. If there was no noticeable indentation the tape was located at the umbilicus. Hip girth was the horizontal circumference at the broadest part of the lower body, usually at the level of the trochanters [<xref ref-type="bibr" rid="B13">13</xref>]. Body fat percentage was assessed using bioelectrical impedance analysis following recommended procedures [<xref ref-type="bibr" rid="B14">14</xref>].</p><p>Duplicate measurements of BP were taken two minutes apart using an automated device (Omron 705 CP; Omron Matsusaka Co. Ltd., Japan) after the subject had rested in a seated position for five minutes. The average of the readings was used. If the first two readings differed by more than 5 mm Hg, an additional reading was obtained and all three readings averaged.</p><p>To monitor cardiovascular adaptations to training, a submaximal, graded exercise test was conducted pre- and post-intervention. In the week prior to the treadmill test, subjects were familiarised with walking on the treadmill (Vision Fitness HRC T8600; Lake Mills, U.S.) at various speeds and slopes. Following a three minute warm-up, subjects were instructed to walk on the treadmill for four minutes at each of four gradients selected to elicit 40, 50, 60, and 70% of age-predicted (220 – age) maximum heart rate (HR<sub>max</sub>). They were encouraged to complete all four stages, but the test was terminated if heart rate (HR) reached 85% HR<sub>max</sub>. HR was measured continually by short-range telemetry. During the last 30 seconds of each test stage, a capillary sample of blood was obtained and immediately analysed for lactate using a Lactate Pro Test Meter (Arkray Inc.; Kyoto, Japan). Ratings of perceived exertion (RPE) using the Borg 15-grade scale (Borg, 1982) were obtained during the last minute of each test stage, in accordance with scripted instructions [<xref ref-type="bibr" rid="B14">14</xref>]</p><p>Whole blood glucose and total cholesterol were determined from a fresh sample of capillary blood using dry chemistry methods (Reflotron analyser, Boehinger Mannheim Ltd.; UK) for the purpose of initial screening criteria. Blood samples were obtained by venepuncture after a 10 hour fast and with subjects in a seated position. Subjects were instructed to refrain from physical activity on the previous day. Samples were separated, frozen at -20°C, and analysed within 8 months. Total cholesterol, HDL and triglycerides were determined by automated enzyme assay using a Cobas FARA bioanalyser (Roche Products Ltd; Herts, UK). Commercial enzyme assay kits were purchased from Randox Laboratories Ltd. (Crumlin, UK). The concentration of LDL cholesterol was calculated using the Friedewald formula [<xref ref-type="bibr" rid="B15">15</xref>]. CRP was measured using an automated high sensitivity immunoturbidimetric assay (Randox Laboratories Ltd; Crumlin, UK) on a Cobas FARA bioanalyser (Roche Products Ltd; Herts, UK). Where CRP levels were below detectable levels (0.5 mg.L<sup>-1</sup>) a value of 0.5 mg.L<sup>-1 </sup>was recorded. All samples for each given assay were analysed on the same day in a single batch at the Institute of Clinical Science at Queen's University, Belfast.</p><p>A sample of whole blood was analysed within 12 hours for haematocrit and haemoglobin. Haematocrit was corrected by 1.5% to account for plasma trapped between erythrocytes. Haematocrit and haemoglobin values were subsequently used to correct for changes in plasma volume [<xref ref-type="bibr" rid="B16">16</xref>]. As female hormonal changes throughout the menstrual cycle may influence lipid profiles [<xref ref-type="bibr" rid="B17">17</xref>], pre- and post-intervention samples were taken during the same phase of the menstrual cycle. In order that lipid parameters were not influenced by the acute effects of previous exercise [<xref ref-type="bibr" rid="B18">18</xref>], the post-intervention sample was obtained on day 3 after the last walking session.</p></sec><sec><title>Walking programme</title><p>It has previously been reported that adults intuitively self-select a pace concordant with cardiorespiratory benefits when walking for exercise [<xref ref-type="bibr" rid="B19">19</xref>]. Therefore, in the present study, subjects were allowed to choose their own walking speed. The progressive walking programme lasted eight weeks. During week one, subjects completed a 25 minute walk on two days. During week two, subjects walked for 35 minutes on two days. From week three to week eight, all walkers completed two 45 minute walks per week. All walking sessions were performed outdoors. Those assigned to the walking group were given a training diary to record their walks and note the day, time of day and duration of the walk. Subjects were also required to rate their perceived exertion during the walk on the Borg 15-grade scale [<xref ref-type="bibr" rid="B20">20</xref>]. In addition, subjects noted their HR immediately following the walk, having previously been instructed in carotid and radial artery palpation. Nine of the walkers, who had difficulty with artery palpation, were given heart rate monitors to monitor their heart rate. Each of these subjects were trained in the use of the monitor and given written instructions. All subjects were instructed regarding the importance of maintaining their usual activity and dietary habits throughout the study.</p></sec><sec><title>Step counts</title><p>Twelve walkers and all of the controls wore a pedometer during weeks 0, 4 and 8 of the intervention (Yamax SW-200, Yamasa Corporation; Tokyo, Japan). They were instructed to wear the pedometer during all waking hours and record the number of steps taken at the end of each day. In addition, during weeks 4 and 8, those assigned to the walking group recorded their step count before and after each walking session in their training diary. Each subject was trained in the correct use of the pedometer and given written guidelines to follow.</p></sec><sec><title>Statistical analysis</title><p>Physiological differences between groups at baseline were compared using independent <italic>t</italic>-tests. Non-parametric Kruskall-Wallis tests were used to compare the step counts of groups at baseline. Non-parametric Wilcoxon tests were used to compare paired step count data of the walking group (Walk-days vs. Rest-days). Physiological data were analysed using a 2-way ANOVA with repeated measures, with one factor between subjects (walkers vs. controls) and one factor within subjects (pre- vs. post-intervention).</p></sec></sec><sec><title>Results</title><p>Four individuals dropped out of the study due to: illness (1 control), moving job (1 control), family circumstances (1 walker) and lack of interest (1 walker). Due to equipment error at the end of the walking intervention, body fat measurement was only possible in 12 walkers and 12 controls. Due to problems in blood sampling we were only able to determine blood lipids (TC, HDL-C, LDL-C, TG) in 11 walkers and 9 controls. In addition one control subject, who had CRP values > 10 mg· L<sup>-1</sup>, was excluded from the analysis as CRP values of this order may be indicative of infection or trauma (Ridker, 2003). CRP data is therefore presented for 11 walkers and 8 controls. For all other parameters measured data are presented for the 21 walkers and 12 controls who completed the study. Physiological characteristics of the participants at baseline are shown in Table <xref ref-type="table" rid="T1">1</xref>. There were no significant differences between groups at baseline for any variable (<italic>P </italic>> 0.05).</p><p>During week 0 (i.e. the week prior to commencing the intervention) daily step counts for the walking and control groups averaged 6437 ± 2285 and 6831 ± 2727 respectively. There was no significant difference in the week 0 step counts between groups (<italic>P </italic>> 0.05). Daily step counts for the walking group on days when prescribed walking was performed (Walk-days), on days when no prescribed walking was performed (Rest-days) and all days during the programme are shown in Table <xref ref-type="table" rid="T2">2</xref>. Walkers took significantly more steps on Walk-days compared to Rest-days (<italic>P </italic>< 0.001). Walkers undertook more voluntary steps (steps per day not including any accrued from prescribed walking) on Rest-days (5803 ± 2749) than on Walk-days (4567 ± 2639) (<italic>P </italic>< 0.05). During the intervention, mean step counts for the control group averaged 6470 ± 1709.</p><p>Subjects assigned to the walking group completed a 45 minute walk on two days of the week, at approximately 62.0 ± 7.1% predicted HR<sub>max</sub>. The walks elicited a mean RPE of 12.6 ± 0.9 and consisted of 4736.4 ± 539.2 steps. Subjects completed 83.9 ± 18.9% of prescribed sessions.</p><p>Table <xref ref-type="table" rid="T3">3</xref> shows measurements made at baseline and post-intervention. There were significant differences in the change in systolic BP and body fat percentage between groups from pre- to post-intervention as identified by the group-by-time interaction (<italic>P </italic>< 0.05). Systolic BP for the walking group decreased from 120.4 ± 19.7 mm Hg at baseline to 115.4 ± 17.7 mm Hg at post intervention. Body fat percentage of the walking group was 28.0 ± 5.8 and 27.9 ± 5.6 at pre- and post-intervention respectively. No significant changes were observed in body mass, waist and hip circumference, diastolic BP or lipid variables.</p><p>Mean values for CRP in the walking group at pre- and post-intervention were 1.9 ± 1.7 and 1.6 ± 1.5 mg· L<sup>-1 </sup>respectively (n = 11). Corresponding values for controls were 1.5 ± 1.5 and 1.5 ± 1.3 mg· L<sup>-1 </sup>at pre- and post-intervention respectively (n = 8). There were no significant effects observed. One subject, who had CRP values > 10 mg· L<sup>-1</sup>, was excluded from the analysis as CRP values of this order may be indicative of infection or trauma (Ridker, 2003).</p><p>Mean HR for the walking group during the treadmill test was 116.5 ± 8.5 beats· min<sup>-1 </sup>and 112.0 ± 9.3 beats· min<sup>-1 </sup>at pre- and post-intervention respectively. Corresponding values for the control group were 121.2 ± 14.7 and 118.1 ± 13.0 beats· min<sup>-1 </sup>at pre- and post-intervention respectively. Blood lactate values for walkers at pre-and post-intervention were 1.6 ± 0.5 and 1.5 ± 0.5 mmol· L<sup>-1 </sup>respectively. Pre- and post-intervention values for the control group were 1.5 ± 0.5 and 1.7 ± 0.5 mmol· L<sup>-1</sup>. Following training RPE during the treadmill test reduced from 11.2 ± 1.4 to 10.8 ± 1.4 in the walking group. A similar reduction was observed in controls, who demonstrated RPE values of 11.3 ± 2.1 and 10.8 ± 1.8 at pre- and post-intervention respectively. No significant differences were observed between groups in HR, blood lactate or RPE.</p></sec><sec><title>Discussion</title><p>The main finding of the present study is that 45 minutes self-paced walking, two days per week, decreases systolic BP, but has no discernable effects on fitness, body mass, waist/hip circumferences, diastolic BP, CRP or lipoproteins, in previously sedentary employees. The novel aspect of this investigation is that it is the first to use a low frequency walking programme with adults < 60 years of age, and the first to examine the effects of walking, without dietary intervention, on CRP levels. The study augments the limited body of evidence on the efficacy of worksite physical activity interventions.</p><p>A significant effect was observed between groups for the change in systolic BP from pre- to post-intervention. Systolic BP was reduced by 5.0 ± 7.9 mm Hg in the walking group following eight weeks of training. A reduction of this magnitude is in general agreement with the findings of a meta-analysis on the effect of aerobic exercise on BP in normotensive adults [<xref ref-type="bibr" rid="B21">21</xref>]. In such individuals a modest reduction in BP confers prophylaxis against CVD morbidity and mortality as CVD risk is manifest on an approximately linear continuum with BP [<xref ref-type="bibr" rid="B22">22</xref>]. The benefit of such a reduction is supported by cross-sectional evidence showing that individuals in the normal range (120–128 systolic BP, and 81–84 diastolic BP) have almost a three-fold increase in risk over subjects with lower baseline levels [<xref ref-type="bibr" rid="B23">23</xref>]. Reductions in BP of this order on a population level may offer significant public health gains. A downward shift of about 2 mm Hg in the BP distribution of the general population should result in an annual reduction in stroke, coronary heart disease and all-cause mortality of about 6%, 4%, and 3% respectively [<xref ref-type="bibr" rid="B24">24</xref>].</p><p>Few other studies [<xref ref-type="bibr" rid="B25">25</xref>,<xref ref-type="bibr" rid="B26">26</xref>] have investigated the BP-lowering effects of walking in adults with mean baseline SBP values of less than 130 mm Hg. The present study is the first to show positive changes in SBP in normotensive adults following regular walking on just two days per week. Given the findings of the present study, the health benefits of lowered blood pressure, and the fact that walking is the most popular form of physical activity in the European Union [<xref ref-type="bibr" rid="B27">27</xref>], it is reasonable to suggest that continued and increased emphasis should be placed on increasing the proportion of adults who participate in this form of activity [<xref ref-type="bibr" rid="B21">21</xref>].</p><p>A significant effect was observed between groups for change in body fat percentage from pre- to post-intervention. Relative to an increase in controls, the body fat percentage of walkers remained stable. The increases in body fat percentage of the control group may, however, be the result of over-cautious control subjects. It is possible that a desire 'not to be active' may have resulted in marginally lower than normal physical activity levels thus contributing to worsened adiposity.</p><p>Participants in the present study are classified as being at moderate risk of future cardiovascular events according to their baseline CRP levels [<xref ref-type="bibr" rid="B28">28</xref>]. In the present investigation, although not statistically significant, CRP decreased by 15.8% following training in walkers, compared to 6.7% in controls. Furthermore, CRP values tended to decrease more notably in those with higher baseline values. Several cross-sectional studies have observed an inverse association between physical activity and CRP levels [<xref ref-type="bibr" rid="B29">29</xref>-<xref ref-type="bibr" rid="B33">33</xref>]. However, intervention studies examining the effects of training on CRP levels are less plentiful. There are only four exercise training interventions identified in the published literature [<xref ref-type="bibr" rid="B34">34</xref>-<xref ref-type="bibr" rid="B37">37</xref>], which reported reductions in CRP following training of 17, 35, 31 and 45 % respectively. Although the magnitude of CRP reduction observed in the present study (15.8%) is less than reported in the other exercise interventions, the exercise prescribed in the present study was considerably less (90 min· wk<sup>-1 </sup>vs. 150–420 min· wk<sup>-1</sup>). Given the varied training regimes in the aforementioned studies, and the paucity of data in the area, comparisons are difficult to draw. Further research is required to elucidate the relationship between exercise levels (frequency, intensity, volume) and changes in CRP, and the effects of training in various population groups. The present study augments the very limited body of evidence on the effects of training on resting CRP concentrations.</p><p>Heart rate and blood lactate responses to submaximal workloads were used in the present study to monitor changes in aerobic fitness. No significant effects were observed for changes in heart rate or lactate response to submaximal workloads. The authors know of one other study in the published literature that has used the same walking prescription as the present investigation. Eighty women, aged 60–70 years progressed to walking 45 minutes 2 d· wk<sup>-1 </sup>[<xref ref-type="bibr" rid="B10">10</xref>]. Following 26 weeks of training at approximately 67 – 78% HR<sub>max</sub>, heart rate during a submaximal exercise test was reduced. Failure to evoke changes in aerobic fitness in the present study, despite the same volume of exercise, may be due to differences in training intensity (67–78% HR<sub>max </sub>vs. ~ 62% HR<sub>max</sub>) or differences in the length of the intervention (26 weeks vs. 8 weeks) [<xref ref-type="bibr" rid="B38">38</xref>]. It has been postulated that the minimal intensity for improving cardiorespiratory fitness is 30% maximal oxygen uptake reserve/~60% HR<sub>max </sub>for lower-fit subjects [<xref ref-type="bibr" rid="B39">39</xref>]. The fact that the subjects in the present study exercised extremely close to the minimum threshold, or that they exercised for only eight weeks, could explain the lack of change in indices of aerobic fitness.</p><p>A systematic review of 32 cross-sectional, prospective observational and intervention studies, suggested that for adults, 7000 – 13,000 steps per day can be expected [<xref ref-type="bibr" rid="B40">40</xref>]. Baseline daily step counts from the present study are below this range (steps for the walking and control groups averaged 6437 ± 2285 and 6831 ± 2727 respectively), suggesting a comparatively sedentary sample for this age group. During the intervention, walkers significantly increased the total number of steps taken on Walk-days compared to Rest-days, indicating greater amounts of activity as a result of the walking programme. However, in the present study, there was a significant difference between the number of voluntary steps taken on Walk-days compared to the number of steps taken on Rest-days. This suggests that employees decreased their non-programme activity during Walk-days. A compensatory decline in both energy expenditure and steps in physical activity during the reminder of the day, in individuals who begin an exercise programme has been reported previously [<xref ref-type="bibr" rid="B41">41</xref>,<xref ref-type="bibr" rid="B42">42</xref>]. This may be due to a decrease in spontaneous physical activity and/or a reduction in voluntary physical activities [<xref ref-type="bibr" rid="B41">41</xref>].</p><p>The walking intervention was well tolerated by the employees in the present study. No injuries were reported during the programme, most likely due to the mode of exercise and that fact that low injury rates are associated with moderate intensity activity [<xref ref-type="bibr" rid="B43">43</xref>]. This finding is important because fear of sustaining an injury, and stopping activity because of an injury, have both been associated with failure to start or maintain a physically active lifestyle [<xref ref-type="bibr" rid="B44">44</xref>] Compliance was high with walkers completing 83.9 ± 18.9% of prescribed sessions, while attrition was low with only two walkers and two controls dropping out of the study. In addition, only one subject cited lack of interest as the reason for discontinued participation. Allowing people to exercise at their own pace may enhance comfort and enjoyment during exercise and may have contributed to the high compliance and low attrition levels in the present study. Additionally, existing channels of communication and opportunities for support may contribute to the use of the workplace as a positive setting for the promotion of physical activity [<xref ref-type="bibr" rid="B45">45</xref>]. These characteristics should be taken into account when public health professionals are designing programmes and prescribing physical activity within the workplace.</p><p>The limitations of this study need to be considered when interpreting findings. The CRP assay used in the present study did not detect CRP in the blood samples of 6 subject (3 walkers 3 controls) as they were below the detectable level of < 0.5 mg· L<sup>-1</sup>. However, apart from one subject, the same individuals were below detectable limits at both sampling times, thus limiting the resulting effect on observed changes over time.</p></sec><sec><title>Conclusion</title><p>In summary, the findings of the present study suggest that self-paced walking 45 min, 2 d· wk<sup>-1 </sup>for eight weeks at ~ 62% HR<sub>max</sub>, reduces systolic BP and prevents an increase in body fat, in previously sedentary employees. This walking prescription, however, failed to induce significant improvements in fitness, diastolic BP, body mass, serum lipids and CRP levels. The walking programme was associated with high adherence. These findings support the use of a twice-weekly, self-paced, worksite based, walking programme, to improve activity levels and systolic BP in previously sedentary employees. There is little evidence however to support the use of this exercise prescription for improvements in other markers of CVD risk. This walking prescription may therefore be useful as a stepping-stone to further increase levels of exercise, which may then provide greater benefits.</p></sec><sec><title>Competing interests</title><p>The author(s) declare that they have no competing interests.</p></sec><sec><title>Authors' contributions</title><p>MHM, EMM and CAGB were responsible for the design of the study</p><p>EMM was responsible for data collection</p><p>LGH was responsible for blood profiling</p><p>AMN was responsible for statistical analysis</p><p>MHM and EMM were responsible for preparing the manuscript</p><p>All authors read and approved the final manuscript</p></sec><sec><title>Pre-publication history</title><p>The pre-publication history for this paper can be accessed here:</p><p><ext-link ext-link-type="uri" xlink:href="http://www.biomedcentral.com/1471-2458/6/136/prepub"/></p></sec> |
Health seeking behaviour, health system experience and tuberculosis case finding in Gambians with cough | <sec><title>Background</title><p>Studies in Africa investigating health-seeking behaviour by interviewing tuberculosis patients have revealed patient knowledge issues and significant delays to diagnosis. We aimed to study health-seeking behaviour and experience of those with cough in The Gambia and to identify whether they had tuberculosis.</p></sec><sec sec-type="methods"><title>Methods</title><p>During a round of a population under 3-monthly demographic surveillance, we identified people >10 years old who had been coughing ≥ 3 weeks. A questionnaire was administered concerning demographic data, cough, knowledge, health seeking, and experience at health facilities. Case finding utilised sputum smear and chest X-ray.</p></sec><sec><title>Results</title><p>122/29,871 coughing individuals were identified. Of 115 interviewed, 93 (81%) had sought treatment; 76 (81.7%) from the health system. Those that visited an alternative health provider first were significantly older than those who visited the health system first (p = 0.03). The median time to seek treatment was 2 weeks (range 0 – 106). 54 (58.1%) made their choice of provider because they believed it was right. Of those who left the health system to an alternative provider (n = 13): 7 believed it was the best place, 3 cited cost and 2 failure to improve. 3 cases were identified by sputum analysis, 11 more by X-ray; all had visited the health system first. Total 'excess' cough time was 1079 person weeks.</p></sec><sec><title>Conclusion</title><p>The majority of people with cough in this population seek appropriate help early. Improved case detection might be achieved through the use of chest X-ray in addition to sputum smear.</p></sec> | <contrib id="A1" contrib-type="author"><name><surname>Kasse</surname><given-names>Yaya</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>yayak25@hotmail.com</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Jasseh</surname><given-names>Momodou</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>mjasseh@mrc.gm</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Corrah</surname><given-names>Tumani</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>tcorrah@mrc.gm</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Donkor</surname><given-names>Simon A</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>sdonkor@mrc.gm</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Antonnio</surname><given-names>Martin</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>mantonio@mrc.gm</email></contrib><contrib id="A6" contrib-type="author"><name><surname>Jallow</surname><given-names>Adama</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>adamahaddyjatou@hotmail.com</email></contrib><contrib id="A7" contrib-type="author"><name><surname>Adegbola</surname><given-names>Richard A</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>radegbola@mrc.gm</email></contrib><contrib id="A8" corresp="yes" contrib-type="author"><name><surname>Hill</surname><given-names>Philip C</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>phill@mrc.gm</email></contrib> | BMC Public Health | <sec><title>Background</title><p>Tuberculosis (TB) causes approximately 2 million deaths per year[<xref ref-type="bibr" rid="B1">1</xref>]. 98% occur in low-income countries[<xref ref-type="bibr" rid="B2">2</xref>]. Directly observed therapy (DOTS), the main strategy for TB control globally, relies on self-presentation of adults from the community and sputum smear for diagnosis. In certain populations, even in the presence of substantial drug-resistance, it is highly effective at reducing <italic>M. tuberculosis </italic>transmissio[<xref ref-type="bibr" rid="B3">3</xref>]. However, in Africa, despite considerable DOTS expansion, the incidence rate is rising by approximately 6% per year[<xref ref-type="bibr" rid="B4">4</xref>]. While high rates of HIV infection have contributed to this[<xref ref-type="bibr" rid="B5">5</xref>], other factors should be considered, such as the level of understanding of TB in the community, health-seeking behaviour and health system performance.</p><p>In The Gambia, which has implemented DOTS and has an HIV seroprevalence of 2%[<xref ref-type="bibr" rid="B6">6</xref>], the TB case notification rose from 82/100,000 in 1994 to 140/100,000 in 2004[<xref ref-type="bibr" rid="B7">7</xref>]. As in many African countries, there is thought to be considerable under-reporting of TB cases. We engaged the quarterly round of a population under demographic surveillance to identify individuals who were coughing for 3 weeks or more, investigate their health seeking behaviour and related experiences and screen them for TB.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>Recruitment</title><p>The study was approved by the combined Gambia government/MRC ethics committee and took place on the North bank, the River Gambia, approximately 100 kilometres inland. The Demographic Surveillance System (DSS) monitors 43,548 people in Farafenni town (26,592) and surrounding villages (16,956), recording: births, deaths, nuptial events, and movement in and out of each of over 5500 households. The population is characterised by youthfulness and a high growth rate. Health services are provided by a 155-bed hospital and clinics in larger villages. There are private dispensaries and pharmacies plus traditional health providers. Diagnosis of TB is at the main hospital. Confirmed cases receive free treatment; 61 were treated in 2004, from a source population of approximately 80,000 people in the DSS and surrounding district.</p><p>During a 3-monthly DSS round from 17 January to 16 April 2005, the head of each household, or an appropriate substitute, was asked if there was any household member over the age of 10 years, coughing for 3 weeks or more. The person who was identified by the head of household with cough of 3 weeks or more was asked to confirm this and advised to go to the nearest health centre. If they asked why, the field worker explained that a cough of 3 weeks or more is not normal. He/she did not mention any possible causes of a cough.</p></sec><sec><title>Investigation of health seeking behaviour</title><p>A semi-structured questionnaire was developed. Questions concerned demographic data (age, sex, occupation, marital status), duration of cough, knowledge of TB, time taken to seek healthcare, type of care sought, assistance available and utilised, quality of the treatment received and associated cost. The questionnaire was administered by trained field workers after informed consent. Participants were divided into those coughing for 3 to 7 weeks and those coughing for 8 weeks and more. The second group were interviewed immediately and the first 6 weeks later.</p></sec><sec><title>Case finding for tuberculosis</title><p>After interview, case finding was undertaken. Three sputum samples were collected per person over three days and transported to the Farafenni hospital laboratory. Sputum smears were prepared and stained with Ziel-Neehlsen for acid-fast bacilli. For a positive result, acid-fast bacilli were seen in at least one sputum sample. Each person interviewed was offered a Farafenni hospital chest X-ray, reported by consensus by two specialist physicians, blinded to the sputum results. Those with equivocal X-rays were offered a repeat X-ray after 1 month. Those diagnosed with TB were referred to the National TB control programme.</p><p>For quality control, 9 individuals with negative sputum smears and a normal chest X-ray had 2 sputum samples taken to the MRC laboratory. They were prepared and stained with auramine-phenol[<xref ref-type="bibr" rid="B8">8</xref>]. positive smears were confirmed by Ziel-Neehlsen (ZN) staining. All were AFB negative. Three sputum smear negative cases each provided 2 repeat sputum samples for smear and culture. Decontaminated specimens were inoculated into BACTEC 9000 MB medium and positive samples were confirmed by ZN stain[<xref ref-type="bibr" rid="B9">9</xref>]. All 3 were smear negative but culture positive.</p></sec><sec><title>Data management and analysis</title><p>All data were double-entered into an ACCESS database and checked for errors. The data were analysed using STATA (version 8; Stata Corp, College Station, TX). Categorical variables were compared using Chi-Square and quantitative variables were compared using t-test and the sign rank sum test where appropriate.</p></sec></sec><sec><title>Results</title><p>The study population from the DSS round was 29,871 over 10 years old. The survey identified 122 individuals (0.4%) coughing for 3 weeks or more; 115 (93.5%) agreed to be interviewed, 8 had travelled, 2 refused, and one had died. Those interviewed were more likely to be villagers, were older than, and had a different ethnic mix to, their source population (table <xref ref-type="table" rid="T1">1</xref>). The majority were illiterate (83%), the most common occupation was farming (49%), 63% were married, 19% were single, 14% were widowed and 4% divorced. No significant differences in basic characteristics or health seeking behaviour were found between those interviewed immediately and those interviewed after 6 weeks.</p><p>Approximately 80% of interviewees had heard of TB, that it is transmissible and curable (table <xref ref-type="table" rid="T2">2</xref>). Radio was the most common source of knowledge. Only 10% knew that a germ caused TB and fewer than 50% could name any key TB signs or symptoms. A large majority (93; 81%; Figure <xref ref-type="fig" rid="F1">1</xref>) had sought treatment for their cough either at health facilities (hospital, health centres, private physicians) or alternative health services (Pharmacists, traditional healers, drug shop keepers). Only 2 individuals said their decision to seek treatment was influenced by the DSS team. Those who initially went to an alternative provider were significantly older than those who attended the health system first (mean median 55 years, range 30–90 years; versus median 44 years, range 10–80 years; p = 0.03). However there were no significant gender or ethnic differences. The pattern of behaviour of villagers was similar to that of urban residents, except that they were more likely not to seek help at all: 14 (25%) villagers did not seek treatment versus 8 (13%) urban residents (p = 0.1).</p><p>Fifty-two (55.9%) participants decided to seek help and where to go by themselves, while a close relative intervened in 25 (26.9%) cases and another compound or village resident in 14 (15.1%) cases. Over half (54; 58.1%) made their choice of provider because they believed it was most appropriate; 19 (20.4%) cited proximity to where they lived and only 4 stated that cost was the primary factor. There were no significant differences in these decision processes between those who initially sought help at the health system and those who first visited an alternative provider, although this analysis was limited by the small sample size. The main reasons given for not seeking treatment were lack of knowledge about the importance of a chronic cough (8) and inability to get to a health centre through lack of resource or transport (10).</p><p>The median duration of cough was 10 weeks (mean 37; range 3–572). The median time to seek treatment after the start of coughing was 2 weeks, both for those who visited the health system first (range 0 – 106) and those who visited an alternative provider (range 1 – 52); 81% sought treatment between 1 to 4 weeks. The majority took over half an hour to get to a health provider (51; 67%); the average cost of the trip was 26 Dalasis (1 Euro), for which 7% had to borrow money.</p><p>Of those who sought treatment at a health facility, 17 (22%) were asked to provide a sputum specimen for analysis at their first visit, including 9 (19%) of those coughing for less than 3 weeks at the time; one was diagnosed as a case and 3 were later diagnosed through case finding. Eight individuals (18%) who visited a health facility a second time were asked to provide a sputum specimen; all were negative but 3 were later diagnosed as a case in our study. Of those who visited a health facility a third time, 3 (9%) were asked to provide a sputum specimen; all were negative. Two (13%) of those who visited a health facility a fourth time were asked to provide a sputum sample; both were negative. Nine individuals were asked to have a chest x-ray, including 2 who were later diagnosed as a case in our study. None of the X-rays were available for inspection.</p><p>The reasons why people left the health system to an alternative provider (n = 13, figure <xref ref-type="fig" rid="F1">1</xref>) were varied: 7 believed that it might be the best place, 3 cited cost as the main reason and 2 cited failure to improve after treatment. Seventy-five (80.6%) of those who initially sought treatment had stopped seeking help by the interview; 71 gave a reason: 30 (42%) were feeling like they were getting better, 8 were not sure what they were going to do, 7 were still completing a course of treatment and 5 were 'too busy'. However 8 (11%) cited cost as the main reason and 6 (9%) had lost confidence in health providers.</p><p>Ninety seven (84 %) participants agreed to produce 3 sputum samples for analysis; 12 refused to participate, 4 had travelled, 1 died, and 1 had been diagnosed with TB. Two were found to be sputum smear positive. Eighty-nine individuals agreed to have a chest X-ray; 2 were pregnant, 4 refused, 1 had died and 1 had been hospitalised. There were 27 individuals with an abnormal X-ray. Thirteen of these (14.6% of all X-rays) were highly suggestive of TB, warranting anti-tuberculosis treatment. The chest-X-rays of four other individuals had infiltrates; they were offered a repeat chest X-ray after 1 month: three agreed and repeat X-rays were normal. Other diagnoses were chronic obstructive airways disease (4), congestive hearth failure (2), old scarring (1), interstitial lung disease (1), pleural effusion (1), and diaphragmatic hernia (1).</p><p>The details of the 13 individuals with TB, plus the person who was diagnosed already, are presented in Table <xref ref-type="table" rid="T3">3</xref>. The median age of the cases was 57.5 years (mean 55; range 14–80) and almost 2/3 were women. Their health seeking behaviour prior to case finding is shown in figure <xref ref-type="fig" rid="F2">2</xref>. All 14 visited the health system first. The median time to this visit was 2 weeks (mean 5.7; range 0–52), significantly shorter than the median time to diagnosis (median 21.5; mean 82.8; range 3–572; p = 0.001). The total 'extra' cough time for those eventually diagnosed was 1079 person weeks.</p></sec><sec><title>Discussion</title><p>In this study, most coughing individuals sought help from the 'health system' promptly, while 1/3 either sought assistance from an alternative provider or no assistance at all. They were largely guided by a desire to go to the most appropriate place, but after the initial visit health seeking behaviour was complex: those who first sought help within the health system tended to stay in that system, while few people visited an alternative provider more than once and tended to not seek help again. Only a minority of individuals who visited a health facility were asked to provide a sputum sample for analysis and very few were asked to have a chest X-ray. Our case finding identified TB disease in 13 of 97 people investigated, making these individuals a group worthy of attention.</p><p>Most studies have investigated health-seeking behaviour with respect to TB by interviewing TB patients. They have revealed patient knowledge issues and significant delays to diagnosis [<xref ref-type="bibr" rid="B10">10</xref>-<xref ref-type="bibr" rid="B12">12</xref>]. In this community-based study of coughing individuals, a key finding is that the majority did seek appropriate attention early. Is health system failure the main culprit in the delay to diagnosis and treatment that we observe? While the health system cannot be blamed for not performing sputum smear assessment in those with less than 3 weeks of cough (if they have no other key symptoms or signs) and this study did not ask whether such individuals were told to return if their cough persisted, it does appear that health system delay was important.</p><p>This study supports the findings of gender bias in passive case detection [<xref ref-type="bibr" rid="B13">13</xref>]: 9 (64%) of the cases identified were women, while approximately 70% of cases treated in the Farafenni area, and in The Gambia overall, are men [<xref ref-type="bibr" rid="B14">14</xref>]. Also, alternative case finding techniques may identify a sub-group of patients who are subject to increased delay. In the Gambian community there is a median delay from onset of the symptoms to treatment of approximately 8 weeks [<xref ref-type="bibr" rid="B15">15</xref>,<xref ref-type="bibr" rid="B16">16</xref>]. Median delays of up to 16 weeks have been reported elsewhere in Africa [<xref ref-type="bibr" rid="B17">17</xref>,<xref ref-type="bibr" rid="B18">18</xref>]. We observed even longer delays in the present study and in a previous alternative case finding study with traditional healers [<xref ref-type="bibr" rid="B19">19</xref>]. In the latter study we also found, consistent with the present study, that those visiting alternative health providers tend to be older than those who visit the health system.</p><p>In Kenya a single cough question first to household heads was the most successful method to identify TB cases in the community [<xref ref-type="bibr" rid="B20">20</xref>]. However, it is possible that we had a relatively low yield in The Gambia with this approach. In rural Uganda, 289 (5.2%) of 5574 people reported a cough of longer than 1 month's duration [<xref ref-type="bibr" rid="B21">21</xref>] Pronyk et a [<xref ref-type="bibr" rid="B22">22</xref>] identified 366 people with a chronic cough from a demographic surveillance population of 38,251. However, the yield from case finding was similar to our study: of 340 investigated, 6 sputum smear positive cases were identified.</p><p>Because case addresses are not routinely documented at the district hospital we do not know the number of sputum smear positive cases that were missed by our study. However, assuming approximately 30 cases per year are identified in the DSS population, with an average duration of cough (in those coughing for at least 3 weeks) of 4–9 weeks, one would expect to identify 1–4 smear positive cases in a cross-sectional survey such as ours. This is consistent with the 3 smear positive case cases that we found. Therefore, DSS patients diagnosed separately by the health system are likely to be small in number.</p><p>There has been considerable debate regarding the place of chest X-ray in the diagnosis of TB. Gothi et al [<xref ref-type="bibr" rid="B23">23</xref>] found in India, using sputum culture as the gold standard, that screening with X-ray at the population level offered little benefit over symptom screening. Harries et al [<xref ref-type="bibr" rid="B24">24</xref>] showed in Malawi, that when sputum culture is available, initial screening with chest X-ray followed by sputum analysis is not advantageous and may miss smear positive cases. However, sputum smear alone would have diagnosed less than half of those identified overall if chest X-ray had been offered to smear negative individuals. Furthermore, approximately one quarter of chest X-ray positive individuals would be expected to be negative on culture [<xref ref-type="bibr" rid="B24">24</xref>]. Therefore, in settings such as ours where sputum culture is not routinely available, it is advised to make use of the option of a chest X-ray if available and affordable. However, it is important that the X-rays are of good quality, that they are read by a radiologist or a physician experienced in X-ray diagnosis of TB, and only those considered to have an X-ray highly suggestive of TB are sent for treatment in the first instance [<xref ref-type="bibr" rid="B25">25</xref>].</p></sec><sec><title>Conclusion</title><p>This study has identified possible ways to improve TB control in The Gambia. Firstly, the large majority of individuals who do seek appropriate help early need to be adequately investigated. Where chest X-ray is available and affordable (the cost is less than 2 Euro in The Gambia), especially in the absence of sputum culture facilities, it is recommended in conjunction with sputum smear. Secondly, educational efforts should encourage those with ongoing symptoms after an initial visit to the health system to seek help again. Community based case finding of itself may not offer added benefit: it is likely that the extra burden of disease in this study would have been identified through enhanced hospital case detection, since all the cases that were eventually diagnosed had sought help appropriately. However it may have a role in certain situations, such as where pockets of increased transmission are clearly identified. Enhanced case detection is a major priority for Global TB control. This study reinforces this urgent need and the importance of related health system delay [<xref ref-type="bibr" rid="B26">26</xref>,<xref ref-type="bibr" rid="B27">27</xref>]. While new diagnostic tools are required to realise this goal, some benefit can be gained in certain settings from the intelligent use of chest X-ray.</p></sec><sec><title>Competing interests</title><p>The author(s) declare that they have no competing interests.</p></sec><sec><title>Authors' contributions</title><p>YK was involved in the design of the study, supervised the field work and drafted the manuscript. MJ was involved in the design of the study, assisted in supervising the fieldwork and in the data analysis and write-up. TC was involved in the diagnosis of the TB cases by chest xray and was involved in the writeup. SD managed the data entry and verification and quality and contributed to the write-up. MA and RA supervised the microbiological aspects of the study and contributed to the write-up. AJ was involved in the design of the study, supervised the government diagnostic work and assisted in the writeup. PH conceived the study was involved in the design, analysis and write-up. All authors read and approved the final manuscript.</p></sec><sec><title>Pre-publication history</title><p>The pre-publication history for this paper can be accessed here:</p><p><ext-link ext-link-type="uri" xlink:href="http://www.biomedcentral.com/1471-2458/6/143/prepub"/></p></sec> |
Self-reported drug utilization, health, and lifestyle factors among 70–74 year old community dwelling individuals in Western Norway. The Hordaland Health Study (HUSK) | <sec><title>Background</title><p>To examine the level and patterns of self-reported medication use (prescription and non-prescription drugs) among 70–74 year old individuals living in the community, and to explore self-reported indications for use, and factors possibly predictive of drug use.</p></sec><sec sec-type="methods"><title>Methods</title><p>A health survey carried out in 1997–99 in the county of Hordaland (Western Norway) in the setting of a population study. A self-administered questionnaire was mailed to 4338 persons born in 1925–27, and a health check-up was offered. Drug use the previous day was reported (point prevalence). 3341 (77.0%) persons who responded, comprise the material for the analyses.</p></sec><sec><title>Results</title><p>Between one third (males) and one quarter (females) did not take any drug the previous day. Mean number of drugs among users was 2.8 (men and women). 32% used three or more drugs and 11.5% five or more. Hypertension and other cardiovascular problems were by far the most common reasons for drug use, followed by respiratory, musculoskeletal and mental health problems. Self-reported poor health, a high Body Mass Index (BMI), and being an ex-smoker (but not currently a smoker) correlated with increasing number of drugs taken.</p></sec><sec><title>Conclusion</title><p>Among 70–74-year old individuals living in the community no use of medication was more common than major polypharmacy (5+ drugs). Persons who had fallen ill and were put on regular medication, probably tended to quit smoking, while those who remained healthy, continued to smoke.</p></sec> | <contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Brekke</surname><given-names>Mette</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>mette.brekke@medisin.uio.no</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Hunskaar</surname><given-names>Steinar</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>steinar.hunskar@isf.uib.no</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Straand</surname><given-names>Jørund</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>jorund.straand@medisin.uio.no</email></contrib> | BMC Public Health | <sec><title>Background</title><p>Elderly people are the major drug users in the community [<xref ref-type="bibr" rid="B1">1</xref>]. In old age, complex medical conditions are relatively frequent, which may indicate multiple drug therapy. In Sweden, drug use among elderly probands has been surveyed over many years [<xref ref-type="bibr" rid="B2">2</xref>]. In Norway remarkably little is known about who use drugs, and for which reasons, although whole sales statistics are available [<xref ref-type="bibr" rid="B3">3</xref>]. Previous Norwegian studies have been carried out on general practitioners' prescription habits [<xref ref-type="bibr" rid="B4">4</xref>], on drug use in nursing homes [<xref ref-type="bibr" rid="B5">5</xref>,<xref ref-type="bibr" rid="B6">6</xref>], or on the use of particular drugs, e.g. antihypertensives, psychotropics and diuretics [<xref ref-type="bibr" rid="B7">7</xref>].</p><p>Although appropriate drug therapy can alleviate symptoms and reduce both morbidity and mortality, the use of numerous drugs also implies increased risks of side-effects, drug-drug and drug-disease interactions, inappropriate use, and non-compliance. Physiological changes, such as decreased renal function and mental impairment, make the elderly more vulnerable to drug related harm [<xref ref-type="bibr" rid="B8">8</xref>]. Principally there are three major categories of sub-optimal drug use: over-use or polypharmacy, inappropriate use, and under-use [<xref ref-type="bibr" rid="B9">9</xref>,<xref ref-type="bibr" rid="B10">10</xref>]. Under-utilization may imply under-prescribing of potentially beneficial therapies because of high age [<xref ref-type="bibr" rid="B10">10</xref>]. Polypharmacy can be defined either as the concomitant use of multiple drugs, or the use of more medications than are clinically indicated [<xref ref-type="bibr" rid="B11">11</xref>]. In epidemiological studies, polypharmacy can only be assessed according to the first definition, by a cut-off value based on a simple count of medications. Although no specific number of drugs can be established to define polypharmacy, cut-off-points from three to five daily drugs per patient have been commonly used [<xref ref-type="bibr" rid="B11">11</xref>,<xref ref-type="bibr" rid="B9">9</xref>].</p><p>In the present study we explore the level and patterns of medication use (prescribed and over the counter – OTC – drugs) among 70–74 year old individuals living in the community, in relation to self-reported reasons for use, sociodemographic factors, and various health status and lifestyle parameters.</p></sec><sec sec-type="methods"><title>Methods</title><p>In 1997–1999 a population based health survey was carried out in Hordaland County in western Norway (with capital city Bergen) [<xref ref-type="bibr" rid="B12">12</xref>]. In total, 44 342 persons were invited, among whom 4338 were born between 1925 and 1927 and were living in Bergen and three nearby communities. The survey was carried out by the National Health Screening Service, Oslo (now The Norwegian Institute of Public Health) in cooperation with the University of Bergen.</p><p>The study was approved by the Regional Committee for Medical Research Ethics and the Norwegian data Inspectorate.</p><p>A questionnaire was mailed to the invited persons, along with an invitation to a health check-up. The filled-in questionnaires were collected at the screening station, where measurements of blood pressure, height and weight, as well as blood-tests were performed. A written consent was given at the screening station.</p><p>The questionnaire included the following questions regarding drug consumption: Did you take any drug yesterday? (yes/no). If yes: report the name of the drug, the reason for taking it, and whether you usually take this drug every day (yes/no). The questionnaire provided space for reporting up to nine different drugs, and participants were asked to continue on a separate sheet if needed. It was specified that "drugs" meant all kinds of medicines: with and without prescription, herbal drugs, vitamins and minerals. Tablets, inhalators, injections, suppositories, ointments as well as droplets were included.</p><p>The drug names written by respondents were subsequently coded according to the Anatomical Therapeutic Chemical (ATC) classification system [<xref ref-type="bibr" rid="B13">13</xref>] and were classified to the substance (5<sup>th</sup>) level of the ATC codes. The reported reasons for drug use were coded according to the International Classification for Primary Care (ICPC) [<xref ref-type="bibr" rid="B14">14</xref>].</p><p>The questionnaire also addressed physical exercise, smoking and alcohol use, general health and depressive symptoms Table (<xref ref-type="table" rid="T1">1</xref>), as well as duration of formal education. For analysis purpose, the answers on questions on past and present smoking and alcohol use were converted to categories.</p><p>In the present paper we focus on drugs with an ATC-code. The use of non ATC-drugs will be reported elsewhere. Data from the 3341 individuals aged 70–74 who responded to the invitation (response rate 77.0%) comprise the material for this article.</p><p>Statistical analyses were performed using SPSS version 11. Bivariate comparisons were examined by chi-square test (categorical variables) and independent samples t-test (continuous variables). Bivariate relationships were computed by means of Pearson's correlation coefficient (continuous variables). Binary logistic regression analyses (enter) were carried out to examine the relationships between drug use and health status and lifestyle measures.</p></sec><sec><title>Results</title><p>Out of the 3341 respondents, 948 (28.4%) reported they did not use any drugs the previous day (men: 33.1%, women: 24.6%). 2393 had taken at least one drug, in total 6 590 drugs (2719 drugs used by 985 men, 3871 drugs used by 1408 women). This corresponds to an average use of 1.97 drugs per person (men: 1.85, women: 2.07, p < 0.01) or 2.75 drugs per individual on treatment (both genders).</p><p>The maximum number of different drugs taken by one person on the previous day was 17. Among women, 33.0% used three or more drugs, 19.4% four or more, and 11.7% five or more drugs (Figure <xref ref-type="fig" rid="F1">1</xref>). Corresponding figures for men were 30.1, 17.6 and 11.4%, respectively. The majority reported they used the medications every day, as 84% of the 6590 drugs were taken on a daily basis.</p><p>The "top five" drug groups comprised close to one third of all medication used. These drugs were from the top: low dose aspirin (used by 15.5% of respondents), beta-blockers (15.5%), statins (13.0%), calcium channel blockers (11.7%), and ACE-antagonists (9.1%).</p><p>Reasons for use were given for 4 970 (75.4%) of the 6 590 drugs. The "top five" diagnostic reasons were: Elevated blood pressure (men: 409, women: 566, corresponding to 27.8 and 30.3% of respondents, respectively), cardiovascular problems other than hypertension [men: 514 (34.9%), women: 365, (19.5%)], respiratory tract problems, incl. allergy [men: 261(17.7%), women: 298 (16.0%)], musculo-skeletal problems [men: 68 (4.6%), women: 322 (17.2%)], and mental health problems [men: 75 (5.1%), women: 200 (10.7%)]. Table <xref ref-type="table" rid="T3">3</xref> shows the most frequently used drugs for the most common diagnostic categories.</p><sec><title>Potentially influencing factors</title><p>In bivariate analyses, female sex, low education, a poor self-reported health and feeling depressed correlated positively with increase drug use Table (<xref ref-type="table" rid="T4">4</xref>). Table <xref ref-type="table" rid="T5">5</xref> shows the association between lifestyle factors and drug use: Non-smokers used more drugs compared to daily cigarette smokers. However, those who had quit smoking used generally more drugs compared to all others. Abstinence from alcohol and performing no physical exercise correlated with increasing drug use, as was being overweight or obese.</p><p>An increasing use of drugs correlated positively with number of daily cigarettes smoked (Pearson's correlation coefficient: 0.058 p = 0.01), with the use of non ATC-drugs like herbal drugs, fish oil etc (Pearson's correlation coefficient: 0.074, p < 0.01), and with decreasing systolic blood pressure (Pearson's correlation coefficient: -0.041, p = 0.02), and diastolic blood pressure (Pearson's correlation coefficient: -0.090, p < 0.01) at the health check-up.</p><p>In binary logistic regression analyses, female sex (OR 1.6 {95% CI 1.04–2.5}), a self-reported poor health (OR 3.0 {1.8–5.0}), and a BMI < 27 kg/m<sup>2 </sup>(OR 1.9 {1.2–3.0}) correlated positively with medication use (data not shown in tables).</p></sec></sec><sec><title>Discussion</title><p>The validity of elderly persons' self-reported drug use and reasons for use may be questioned. However, asking which drugs were taken <italic>the previous day </italic>(point prevalence) probably minimized recall bias [<xref ref-type="bibr" rid="B15">15</xref>]. Asking which drugs were actually taken also bypassed the problem of non-compliance, as we know that around 25% of prescribed drugs may not be used [<xref ref-type="bibr" rid="B16">16</xref>]. Data quality in our study was considered good, as drug names were generally correctly written and only few questionnaires had to be rejected because of inconsistencies. The high response rate (77%) also contributes to strengthen the validity of the results. We may assume, however, that the 23% who did not participate, would have been found to use more drugs than the respondents, as people with poor health, poor social status and unhealthy lifestyle are known to be generally over-represented among non-respondents in health surveys [<xref ref-type="bibr" rid="B17">17</xref>].</p><p>75% of those who had taken any drug gave an explicit reason for using it. Probably, the remaining quarter did not know or was not sure about the indications (low dose ASA, beta-blockers and estrogens were the main drug groups taken by those who gave no reason for taking them). Our figures, however, correspond well with the results in a comprehensive British interview study among people aged 65 +, where 76% knew the reason for taking their medications [<xref ref-type="bibr" rid="B16">16</xref>].</p><p>In our study, 28% (men: 33%, women: 25%) reported no use of ATC-drugs on the day previous to the survey. Direct comparison with other studies is complicated by differences in data collection, population sample and kind of drugs included. The previously cited British interview study, gave results almost identical with ours, as roughly one third had taken no drugs during the last 24 hours, and mean number of drugs for those on treatment was 2.8 [<xref ref-type="bibr" rid="B16">16</xref>]. In a Swedish study among 70-year old individuals performed in 2000, only one in five men and one in ten women reported to use no drugs at all [<xref ref-type="bibr" rid="B2">2</xref>]. Average number of drugs among those on treatment in Sweden was 3.3 in men and 4.0 in women, which is substantially higher than our figures. This is, however, not surprising, as the overall drug consumption in Norway (measured in daily doses per inhabitants) only reached 60% of the Swedish level in 2000 [<xref ref-type="bibr" rid="B3">3</xref>]. A corresponding study carried out by the Swedish authors back in 1972, showed that 40% of men and 24% of women did not use any drug. The proportion who took four or more drugs increased from 21% (1972) to 28% (2000) in men and from 30% to 41% in women [<xref ref-type="bibr" rid="B2">2</xref>].</p><p>A Danish study among 75-year old persons living in the community revealed that only 3% of the subjects did not take any drug, and that the average total number of drugs – prescribed and OTC – was as high as 5.4 [<xref ref-type="bibr" rid="B18">18</xref>]. The higher level was probably partly caused by examination of drug storages in patients' homes and thereby including drugs that were in fact not used, as total drug consumption in Denmark is slightly lower compared to Norway (close to 90%) [<xref ref-type="bibr" rid="B3">3</xref>]. In the Danish study, CNS drugs were the most commonly used category, compared to cardiovascular drugs in our and other studies [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B2">2</xref>,<xref ref-type="bibr" rid="B16">16</xref>,<xref ref-type="bibr" rid="B19">19</xref>]. 26% of the elderly Danes had taken a benzodiazepine the previous day, reflecting the relatively high use of benzodiazepines in Denmark in the late 1990s [<xref ref-type="bibr" rid="B20">20</xref>]. In our study, only 5% reported having taken a benzodiazepine, corresponding to the average level of use in the population in Hordaland County [<xref ref-type="bibr" rid="B3">3</xref>]. Other studies have shown that use of benzodiazepines increases by age [<xref ref-type="bibr" rid="B21">21</xref>], and we can only speculate if our results are influenced by selection bias or possible underreporting.</p><p>The lower use of drugs in our study compared to neighbouring countries probably reflects habits and preferences in the population as well as among physicians. In a population traditionally founded on fishing and agriculture like on the Western coast of Norway, it is often considered a benefit to be independent of regular medication, and to cope with occasional illnesses without drugs. Norwegian family physicians may also be more cautious than colleagues elsewhere in prescribing regular medications for elderly people, especially drugs with potential CNS side-effects. On the other hand, we may simply lag behind our neighbours in an ongoing process towards increased drug use. As the Swedish study shows, drug use among 70 year old persons increased remarkably during three decades, despite generally improved health status [<xref ref-type="bibr" rid="B2">2</xref>]. However, over that period, a range of new drugs have been introduced, especially for the treatment of cardiovascular risk factors [<xref ref-type="bibr" rid="B22">22</xref>,<xref ref-type="bibr" rid="B23">23</xref>]. Recently, focus has switched from concerns of unnecessary treatment to possible under-prescribing of potentially beneficial therapies to seniors [<xref ref-type="bibr" rid="B10">10</xref>]. Low dose aspirin, beta-blockers and lipid lowering agents were already the dominant drug groups in our sample (13% used a statin). The sales of lipid reducing agents have doubled in Norway from 1999 to 2004 [<xref ref-type="bibr" rid="B3">3</xref>], and recently, 13% of presumed healthy, low-risk 75-year old women in Oslo were found to use statins [<xref ref-type="bibr" rid="B24">24</xref>].</p><p>In our study, 31.7% had used at least three drugs, while 11.5% had taken five or more daily drugs. This is less compared to the Swedish [<xref ref-type="bibr" rid="B2">2</xref>] and the Danish [<xref ref-type="bibr" rid="B18">18</xref>] studies cited above, but equals the results of the British interview study [<xref ref-type="bibr" rid="B16">16</xref>]. Based on data from a Danish prescription database, Bjerrum found that on a random day, close to 5% of 70–74 year old persons used five or more prescribed drugs [<xref ref-type="bibr" rid="B11">11</xref>]. He also found a six-fold variation between general practices regarding the prevalence of major polypharmacy (5+ drugs), correlating – among other things – with the doctors' workload, as increasing workload implied increased prescription.</p><sec><title>Potentially influencing factors</title><p>Many of the factors which correlated positively with drug use in our study Tables (<xref ref-type="table" rid="T4">4</xref> and <xref ref-type="table" rid="T5">5</xref>) have also been identified by others: female sex, a relatively low education (especially for respiratory and cardiovascular drugs) [<xref ref-type="bibr" rid="B1">1</xref>] depressed mood, and a poor self-reported general health [<xref ref-type="bibr" rid="B25">25</xref>]. Not surprisingly, we found (in bivariate analyses) that reporting poor general health, feeling depressed, and performing no regular exercise were all strongly correlated with drug use, both with the use of any drug, with an increasing number of drugs used, as well as with the concomitant use of five or more different drugs.</p><p>It may be surprising that being a non-smoker and being abstinent from alcohol both correlated with increased use of drugs. However, having quit smoking strongly correlated to increased drug use. In the group who had smoked more than 10 cigarettes a day for 30 years or more, and who now had quit smoking, (n = 237), one out of six used five or more drugs. That former smokers were most likely to be on medication, corresponds to what has been found in other studies [<xref ref-type="bibr" rid="B25">25</xref>,<xref ref-type="bibr" rid="B26">26</xref>]. But how can we explain that heavy present smokers (more than 10 cigarettes a day for 30 years or more, n = 114) in fact used less drugs than the average? One explanation may be that people who had fallen ill and were put on regular medication, quitted smoking, while those who remained healthy in spite of their smoking habit, were more likely to continue smoking. Another probable explanation may be that a substantial proportion of heavy smokers died before reaching the age of 70–74 years [<xref ref-type="bibr" rid="B27">27</xref>].</p><p>A corresponding relationship may also be valid for alcohol use, but because we did not ask about quitting drinking, we could not explore this. The negative correlation between use of alcohol and medications was nevertheless weaker than for smoking. Even regular drinkers (> 10 drinking days a month or > 20 glasses of alcohol a month) used (non-significantly) less drugs compared to the rest of the sample. This corresponds with epidemiological data supporting that regular moderate alcohol consumption is associated with less morbidity and mortality [<xref ref-type="bibr" rid="B28">28</xref>]. An American study found no significant relationship between current alcohol use and current medication use among persons 65 and over [<xref ref-type="bibr" rid="B25">25</xref>]. However, in our study under-reporting of actual drinking and a selection effect implying that heavy smokers and drinkers with poor health may be overrepresented among the non-responders can not be ruled out.</p></sec></sec><sec><title>Conclusion</title><p>No use of medication was more common than major polypharmacy (5+ drugs). Among 70–74 year old individuals living in the community, between one third (males) and one fourth (females) did not use prescription or OTC drugs. Mean number of drugs among those on treatment was 2.8, and 11.5% had taken five or more drugs. Drugs for hypertension and other cardiovascular risk-factors were the most frequently used. In bivariate analyses being an ex-smoker (but not currently a smoker) correlated with drug use. Female sex (OR 1.6), being overweight (OR 1.9), and reporting a poor general health (OR 3.0), were factors that independently predicted medication use in this elderly population.</p></sec><sec><title>Competing interests</title><p>The author(s) declare they have no competing interests.</p></sec><sec><title>Authors' contributions</title><p>MB carried out the data analyses and drafted the article.</p><p>SH planned the study, drafted the questionnaire, and carried out the data collection.</p><p>JS planned the study, drafted the questionnaire, carried out the data collection, and participated in drafting the article.</p></sec><sec><title>Pre-publication history</title><p>The pre-publication history for this paper can be accessed here:</p><p><ext-link ext-link-type="uri" xlink:href="http://www.biomedcentral.com/1471-2458/6/121/prepub"/></p></sec> |
Views of new internal medicine faculty of their preparedness and competence in physician-patient communication | <sec><title>Background</title><p>We sought to assess self-rated importance of the medical interview to clinical practice and competence in physician-patient communication among new internal medicine faculty at an academic medical center.</p></sec><sec sec-type="methods"><title>Methods</title><p>Since 2001, new internal medicine faculty at the Mayo Clinic College of Medicine (Rochester, Minnesota) have completed a survey on physician-patient communication. The survey asks the new faculty to rate their overall competence in medical interviewing, the importance of the medical interview to their practice, their confidence and adequacy of previous training in handling eight frequently encountered challenging communication scenarios, and whether they would benefit from additional communication training.</p></sec><sec><title>Results</title><p>Between 2001 and 2004, 75 general internists and internal medicine subspecialists were appointed to the faculty, and of these, 58 (77%) completed the survey. The faculty rated (on a 10-point scale) the importance of the medical interview higher than their competence in interviewing; this difference was significant (average ± SD, 9.4 ± 1.0 vs 7.7 ± 1.2, <italic>P </italic>< .001). Similar results were obtained by sex, age, specialty, years since residency or fellowship training, and perceived benefit of training. Experienced faculty rated their competence in medical interviewing and the importance of the medical interview higher than recent graduates (ie, less than one year since training). For each challenging communication scenario, the new faculty rated the adequacy of their previous training in handling the scenario relatively low. A majority (57%) said they would benefit from additional communication training.</p></sec><sec><title>Conclusion</title><p>Although new internal medicine faculty rate high the importance of the medical interview, they rate their competence and adequacy of previous training in medical interviewing relatively low, and many indicate that they would benefit from additional communication training. These results should encourage academic medical centers to make curricula in physician-patient communication available to their faculty members because many of them not only care for patients, but also teach clinical skills, including communication skills, to trainees.</p></sec> | <contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Mueller</surname><given-names>Paul S</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I2">2</xref><email>mueller.pauls@mayo.edu</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Barrier</surname><given-names>Patricia A</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I2">2</xref><email>barrier.patricia@mayo.edu</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Call</surname><given-names>Timothy G</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I2">2</xref><email>call.timothy@mayo.edu</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Duncan</surname><given-names>Alan K</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I2">2</xref><email>duncan.alan@mayo.edu</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Hurley</surname><given-names>Daniel L</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I2">2</xref><email>hurley.daniel@mayo.edu</email></contrib><contrib id="A6" contrib-type="author"><name><surname>Multari</surname><given-names>Adamarie</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I2">2</xref><email>multari.adamarie@mayo.edu</email></contrib><contrib id="A7" contrib-type="author"><name><surname>Rabatin</surname><given-names>Jeffrey T</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I2">2</xref><email>rabatin.jeffrey@mayo.edu</email></contrib><contrib id="A8" contrib-type="author"><name><surname>Li</surname><given-names>James TC</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I2">2</xref><email>li.james@mayo.edu</email></contrib> | BMC Medical Education | <sec><title>Background</title><p>Medical interviewing is one of the most important skills in medicine and effective physician-patient communication is essential for optimal medical care. Effective physician-patient communication is associated with improved patient satisfaction [<xref ref-type="bibr" rid="B1">1</xref>-<xref ref-type="bibr" rid="B6">6</xref>], increased patient adherence with medications and recommendations [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B6">6</xref>-<xref ref-type="bibr" rid="B8">8</xref>], improved medical outcomes [<xref ref-type="bibr" rid="B6">6</xref>,<xref ref-type="bibr" rid="B7">7</xref>,<xref ref-type="bibr" rid="B9">9</xref>], and less malpractice risk [<xref ref-type="bibr" rid="B10">10</xref>]. Indeed, consensus statements [<xref ref-type="bibr" rid="B11">11</xref>,<xref ref-type="bibr" rid="B12">12</xref>] have concluded that effective physician-patient communication is an integral part of medical practice. The Association of American Medical Colleges (AAMC) [<xref ref-type="bibr" rid="B13">13</xref>] and the American College of Physicians [<xref ref-type="bibr" rid="B14">14</xref>] regard effective communication as a key attribute of professionalism, and the Accreditation Council for Graduate Medical Education (ACGME) [<xref ref-type="bibr" rid="B15">15</xref>] describes effective interpersonal and communication skills as a general competency common to all specialties.</p><p>Although previous surveys [<xref ref-type="bibr" rid="B16">16</xref>-<xref ref-type="bibr" rid="B20">20</xref>] assessed preparedness and competence in communication among internists, they did so only partially (eg, one or two questions). In addition, these surveys did not assess the preparedness for specific physician-patient communication scenarios. Herein, we describe the results of a survey that assessed self-rated preparedness, competence, and perceived benefit of additional training in physician-patient communication among new internal medicine faculty, regardless of their age or previous experience, at an academic medical center.</p></sec><sec sec-type="methods"><title>Methods</title><p>Since 2001, the Department of Internal Medicine of the Mayo Clinic College of Medicine (Rochester, Minnesota) has conducted a physician-patient communication program for new faculty members. In an effort to assess self-rated importance of the medical interview to clinical practice and competence in physician-patient communication and perceived benefit of additional physician-patient communication training, the new faculty members were asked to complete a survey. The survey asked new faculty to report sex, age, specialty, years since completing training, and whether they would benefit from additional communication training ("yes," "no," or "maybe"). Using 10-point Likert scale questions, the survey also asked the faculty to rate their overall competence in medical interviewing, the importance of the medical interview to their practices, and their confidence and adequacy of previous training (eg, during medical school and residency) in handling eight frequently encountered challenging physician-patient communication scenarios. The survey questions are listed in Table <xref ref-type="table" rid="T1">1</xref>. The survey was conducted during the years 2001 through 2004. The results of the survey comprise the data set of this study.</p><sec><title>Statistical analysis</title><p>For within-participant comparisons of continuous outcomes, Wilcoxon signed rank tests were used. The relation between continuous outcomes was modeled using an ordinary least squares fit of a linear model, and <italic>F</italic>-tests were used for a test of association. For between-participant group comparisons of continuous outcomes, Wilcoxon rank sum tests were used. To test for independence of variables (eg, sex, age, specialty, and years since completing training), an analysis of variance (ANOVA) was done where feasible and appropriate. Significance was a <italic>P </italic>value of .05 or less. All analyses were conducted using JMP 4.0.4 (SAS Institute, Inc., Cary, North Carolina). Permission to perform an analysis of the surveys was granted by the Mayo Foundation Institutional Review Board in accordance with federal regulations.</p></sec></sec><sec><title>Results</title><p>Between 2001 and 2004, 75 general internists and internal medicine subspecialists were appointed to the faculty of the Mayo Clinic College of Medicine (Rochester, Minnesota). Of these, 58 (77%) completed the survey. Of the 58 survey respondents, 37 (64%) were men, 43 (75%) were 40 years old or younger, 40 (69%) were subspecialists, and 22 (39%) were recent (within 1 year) graduates of a residency or fellowship training program. Thirty-three participants (57%) said they would benefit from additional communication training, 24 (41%) said they might benefit, and one (2%) said he or she would not benefit (Table <xref ref-type="table" rid="T2">2</xref>).</p><p>On a 10-point scale, the average ± standard deviation of self-rated competence (1 = not competent, 10 = highly competent) in medical interviewing by the new faculty was 7.7 ± 1.2. General internists rated themselves more competent in medical interviewing than subspecialists did (8.1 ± 1.1 vs 7.5 ± 1.1, <italic>P </italic>= .042), as did faculty who had completed their residency or fellowship training more than one year before beginning the physician-patient communication curriculum compared with recent graduates (7.9 ± 1.0 vs 7.3 ± 1.4, <italic>P </italic>= .032). There were no differences in self-rated competence in medical interviewing according to sex, age, or perception of benefiting from the curriculum (Table <xref ref-type="table" rid="T2">2</xref>). On the basis of these results and the limited sample size, we performed ANOVA using the variables sex, specialty, and years since training. After accounting for sex and years since training, a statistically significant association was not found between specialty (general internist vs subspecialist) and self-rated competence in medical interviewing (<italic>P </italic>= .162), and a marginally significant association was found between years since training (ie, more than one year) and self-rated competence (<italic>P </italic>= .055).</p><p>The average self-rated importance of the medical interview in the practices of the new faculty was 9.4 ± 1.0. Faculty who had completed their residency or fellowship training more than one year before beginning the curriculum rated the importance of the medical interview higher than did recent graduates (9.6 ± 0.6 vs 9.0 ± 1.3, <italic>P </italic>= .039). There were no statistically significant differences in self-rated importance of the medical interview according to sex, age, specialty, or perception of benefiting from the curriculum (Table <xref ref-type="table" rid="T2">2</xref>). ANOVA was then performed using the variables sex, specialty, and years since training. After accounting for sex and specialty, a statistically significant association between years since training (ie, more than one year) and self-rated importance of the medical interview was present (<italic>P </italic>= .012).</p><p>New faculty rated their overall competence in interviewing lower than the importance of the medical interview to practice. Matched-pairs testing revealed a significant difference when comparing how faculty rated their overall competence in medical interviewing and how they rated the importance of the medical interview in practice (difference 1.7 [95% CI, 1.4–2.0], <italic>P </italic>< .001). Similar results were obtained by sex, age, generalist versus specialist, years since training, and perceived benefit of training (Table <xref ref-type="table" rid="T2">2</xref>).</p><p>The new faculty were asked to rate their confidence and adequacy of their previous training in handling eight frequently encountered challenging physician-patient communication scenarios. For every scenario, the new faculty rated the adequacy of their previous training in handling the scenarios relatively low (range, 4.5 ± 2.4 to 6.1 ± 2.4). Furthermore, for each scenario, new faculty rated their confidence in handling the scenario higher than their adequacy of training. Indeed, matched-pairs testing showed significant differences in self-rated confidence in handling and self-rated adequacy of training for every scenario. For each scenario, a significant percentage of variability of self-rated confidence in handling the scenario was associated with self-rated adequacy of previous training (Table <xref ref-type="table" rid="T3">3</xref>). Similar results were obtained when the data were analyzed by subgroup (ie, sex, age, generalist vs specialist, and years since training) (data not shown).</p><p>The new faculty who said they would benefit from additional communication training rated the adequacy of their previous training in handling six of the challenging communication scenarios significantly lower than the faculty who said they might or would not benefit from communication training (Table <xref ref-type="table" rid="T4">4</xref>).</p></sec><sec><title>Discussion</title><p>Several previous surveys have partially assessed preparedness and competence in physician-patient communication. For example, a survey of 210 graduates of an internal medicine training program rated the adequacy of their communication training significantly lower than the importance of those skills to clinical practice [<xref ref-type="bibr" rid="B16">16</xref>]. Several other surveys of graduates of internal medicine training programs had similar results [<xref ref-type="bibr" rid="B17">17</xref>-<xref ref-type="bibr" rid="B19">19</xref>]. Finally, in a telephone survey of 300 physicians, 92% rated effective communication as important; yet, 28% rated the adequacy of their training in physician-patient communication as fair or poor [<xref ref-type="bibr" rid="B20">20</xref>].</p><p>The results of our survey add to those of previous surveys in several important ways. First, whereas the previous surveys partially assessed physician-patient communication (eg, only one or two questions), our survey was focused entirely on this topic. Second, unlike previous surveys, ours focused on new internal medicine faculty at an academic medical center regardless of their age or previous experience. Third, we not only assessed self-rated importance of the medical interview to clinical practice and competence in medical interviewing among the new faculty, we also assessed self-rated confidence and perceived adequacy of previous training in 8 challenging physician-patient communication scenarios. Fourth, we specifically asked the new faculty whether they would benefit from additional training in physician-patient communication. Finally, unlike previous surveys, we analyzed our results according to sex, age, years in practice, and generalist versus subspecialty focus.</p><p>Our new internal medicine faculty rated the importance of medical interviewing relatively high. This high regard for medical interviewing is consistent with consensus statements [<xref ref-type="bibr" rid="B11">11</xref>,<xref ref-type="bibr" rid="B12">12</xref>] that concluded effective physician-patient communication is an integral part of medical practice and the position of the ACGME that effective communication is a general competency important to all specialties [<xref ref-type="bibr" rid="B15">15</xref>]. However, our faculty also rated the importance of medical interviewing to clinical practice significantly higher than their competence in medical interviewing. There are several possible reasons for this difference. On the one hand, the difference may be true. The statistical analysis (matched pairs) compared the means of 2 related medical interviewing topics – self-rated importance of and competence in medical interviewing – from the same group of physicians. A higher rating for importance than for competence suggests a perceived educational gap. On the other hand, the difference may not be true, but rather reflect different scales of judgment. Although a 10-point scale was used for all the survey questions, directly comparing self-rated importance of with self-rated competence in medical interviewing assumes the scales themselves are comparable (ie, each particular value on the importance scale means the same on the competence scale). In other words, physicians may naturally rate the importance of medical interviewing higher than their competence in interviewing not because of a true difference but because the items being rated are different (ie, akin to comparing an apple to an orange).</p><p>Older (age >40 years) and more experienced new faculty (completed residency or fellowship training more than one year before beginning the communication curriculum) rated the importance of medical interviewing higher than younger new faculty and those who were recent graduates. One reason for these differences may be the relative lack of clinical experience (compared with that of more experienced physicians) among younger faculty and recent graduates for whom the value and centrality of the medical interview to clinical practice has not been fully realized. Another reason may be that physician-patient communication is not part of the formal curriculum, and trainees may graduate from their program with the conception that effective physician-patient communication is not as important as other skills. Indeed, training in communication skills is underemphasized during internal medicine residency training [<xref ref-type="bibr" rid="B16">16</xref>,<xref ref-type="bibr" rid="B17">17</xref>,<xref ref-type="bibr" rid="B19">19</xref>]. Finally, we found that more experienced new faculty rated their competence in medical interviewing higher than less experienced faculty. Over time, it is likely that physicians expand their repertoires of communication skills and competencies in using them. Indeed, communication skills may improve with time and experience alone [<xref ref-type="bibr" rid="B21">21</xref>].</p><p>Our new faculty rated the adequacy of their previous training in handling 8 challenging physician-patient communication scenarios relatively low. For each scenario, a significant percentage of variability of confidence in handling the scenario was associated with self-rated adequacy of previous training (Table <xref ref-type="table" rid="T3">3</xref>). Although it is plausible that confidence in handling the scenarios is influenced by previous training, our study does not establish a causal relationship. An alternative explanation may be that confidence in handling the scenarios is influenced by learning and that the less confident faculty attributed their confidence to their previous training programs rather than their own learning.</p><p>Importantly, a majority (57%) of our new faculty specifically said they would benefit from additional training in physician-patient communication. These faculty also rated the adequacy of their training in handling six of the challenging physician-patient communication scenarios significantly lower than those who said they might or would not benefit from additional training (Table <xref ref-type="table" rid="T4">4</xref>). These findings are important because many internal medicine faculty at academic medical centers not only care for patients, but also teach clinical skills, including communication skills, to medical students, residents, and other trainees [<xref ref-type="bibr" rid="B22">22</xref>]. Teaching these skills, however, requires qualified and willing faculty members [<xref ref-type="bibr" rid="B23">23</xref>].</p><p>Our results also suggest that many new internal medicine faculty may perceive a need for additional communication training. Such training may enhance clinical practice and teaching and role-modeling communication skills to trainees. In fact, evidence suggests that effective physician-patient communication skills can be taught and learned [<xref ref-type="bibr" rid="B24">24</xref>]. Furthermore, training may improve medical outcomes (eg, improving patient satisfaction and reducing patient emotional distress) [<xref ref-type="bibr" rid="B25">25</xref>-<xref ref-type="bibr" rid="B27">27</xref>]. A review of specific curricula that might be used to teach such skills is beyond the scope of this paper.</p><p>Our survey has a number of limitations. First, the results may not apply to institutions unlike ours and to physicians who are not general internists or internal medicine specialists. Second, the survey used data derived from self-reported ratings of physician competence, confidence, and adequacy of previous training in communication (not measures of actual competency such as assessments of actual interactions between physicians and patients). Also, they may be influenced by recall bias [<xref ref-type="bibr" rid="B16">16</xref>] and experiences during the time since training [<xref ref-type="bibr" rid="B18">18</xref>]. However, a previous study of physicians found that performance correlated with self-reported preparedness for clinical practice [<xref ref-type="bibr" rid="B28">28</xref>]. Third, the size of our cohort was relatively small. However, this size was similar to that of some of the studies discussed above.</p></sec><sec><title>Conclusion</title><p>Although new internal medicine faculty regard medical interviewing as important to clinical practice, they also rate the adequacy of their training in handling challenging communication scenarios relatively low. In fact, many faculty specifically stated they would benefit from additional training in physician-patient communication. These results should encourage academic medical centers to make curricula (ie, faculty development) in physician-patient communication available to their faculty because many of them not only care for patients, but also teach clinical skills, including communication skills, to trainees.</p></sec><sec><title>Competing interests</title><p>The author(s) declare that they have no competing interests.</p></sec><sec><title>Authors' contributions</title><p>All authors have given final approval of the submitted manuscript.</p><p>The following authors have participated in the work and take responsibility for part of the content (P.A.B., T.G.C., A.K.D., D.L.H., A.M., J.T.R., and J.T.C.L.), the whole content (P.S.M.), contributed to conception and design (P.S.M., P.A.B., T.G.C., A.K.D., D.L.H., A.M., J.T.R., and J.T.C.L.), acquisition of data (P.S.M., P.A.B., T.G.C., A.K.D., D.L.H., A.M., J.T.R., and J.T.C.L.), analysis and interpretation of data (P.S.M., J.T.R., and J.T.C.L.), drafting of the manuscript (P.S.M.), and critical revision of the manuscript for important intellectual content (P.S.M., P.A.B., T.G.C., A.K.D., D.L.H., A.M., J.T.R., and J.T.C.L). The Mayo Clinic (Rochester, Minnesota) Center for Patient-Oriented Research provided statistical support.</p></sec><sec><title>Pre-publication history</title><p>The pre-publication history for this paper can be accessed here:</p><p><ext-link ext-link-type="uri" xlink:href="http://www.biomedcentral.com/1472-6920/6/30/prepub"/></p></sec> |
Role of apoptosis in the pathogenesis of COPD and pulmonary emphysema | <p>Chronic obstructive pulmonary disease (COPD) is characterised by chronic inflammation of the airways and progressive destruction of lung parenchyma, a process that in most cases is initiated by cigarette smoking. Several mechanisms are involved in the development of the disease: influx of inflammatory cells into the lung (leading to chronic inflammation of the airways), imbalance between proteolytic and anti-proteolytic activity (resulting in the destruction of healthy lung tissue) and oxidative stress. Recently, an increasing number of data suggest a fourth important mechanism involved in the development of COPD: apoptosis of structural cells in the lung might possibly be an important upstream event in the pathogenesis of COPD. There is an increase in apoptotic alveolar epithelial and endothelial cells in the lungs of COPD patients. Since this is not counterbalanced by an increase in proliferation of these structural cells, the net result is destruction of lung tissue and the development of emphysema. Data from animal models suggest a role for Vascular Endothelial Growth Factor (VEGF) in the induction of apoptosis of structural cells in the lung. Other mediators of apoptosis, such as caspase-3 and ceramide, could be interesting targets to prevent apoptosis and the development of emphysema.</p><p>In this review, recent data on the role of apoptosis in COPD from both animal models as well as from studies on human subjects will be discussed. The aim is to provide an up to date summary on the increasing knowledge on the role of apoptosis in COPD and pulmonary emphysema.</p> | <contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Demedts</surname><given-names>Ingel K</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>M.DemedtsIngelK@UGent.be</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Demoor</surname><given-names>Tine</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>tine.demoor@UGent.be</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Bracke</surname><given-names>Ken R</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>ken.bracke@UGent.be</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Joos</surname><given-names>Guy F</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>guy.joos@UGent.be</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Brusselle</surname><given-names>Guy G</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>Guy.Brusselle@UGent.be</email></contrib> | Respiratory Research | <sec><title>Review</title><p>Chronic Obstructive Pulmonary Disease (COPD) is a chronic respiratory disease that is associated with an abnormal inflammatory response of the lungs to noxious particles or gases (mainly cigarette smoke). This leads to chronic bronchitis-bronchiolitis (small airways disease) and/or emphysema that cause airflow limitation that is not fully reversible. [<xref ref-type="bibr" rid="B1">1</xref>]. COPD is the fifth leading cause of death worldwide, accounting for more than 2 500 000 deaths every year (WHO world health report 2002). Moreover, the prevalence and mortality of COPD are expected to increase in the coming decades [<xref ref-type="bibr" rid="B2">2</xref>].</p><p>Several mechanisms contribute to the pathogenesis of COPD [<xref ref-type="bibr" rid="B3">3</xref>]. First, the inhalation of noxious particles such as cigarette smoke causes the influx of inflammatory cells into the airways and lungs, leading to chronic inflammation. Different kinds of inflammatory cells (macrophages, neutrophils, CD8+ T lymphocytes) have been described to participate in the inflammatory response in the airways of COPD patients.</p><p>Second, there is a disruption of the balance between proteolytic and anti-proteolytic molecules in the lungs of COPD patients, resulting in an increased proteolytic activity [<xref ref-type="bibr" rid="B4">4</xref>]. This causes the destruction of healthy lung parenchyma, which leads to the development of emphysema. This increase in proteolytic activity may be a consequence of inflammation (release of proteolytic enzymes by inflammatory cells such as macrophages and neutrophils) or may arise from genetic factors (eg alpha-1 antitrypsin deficiency).</p><p>A third mechanism involved in the pathogenesis of COPD is oxidative stress, which occurs when reactive oxygen species are produced in excess of the antioxidant defence mechanisms [<xref ref-type="bibr" rid="B3">3</xref>]. Oxidants are generated in the airways by cigarette smoking or are released from inflammatory leukocytes and epithelial cells. Oxidative stress can lead to cell dysfunction or cell death and can induce damage to the lung extracellular matrix. Moreover, oxidative stress influences the proteinase-antiproteinase imbalance by activating proteases and inactivating antiproteinases. Additionally, oxidants contribute to the inflammatory reaction by activating the transcription factor NF-κB and thus inducing the transcription of pro-inflammatory genes. In conclusion, it is clear that these three processes (chronic inflammation, proteinase/anti-proteinase imbalance and oxidative stress) involved in the pathogenesis of COPD are not independent mechanisms and several interactions between these processes occur during the development of the disease.</p><p>Recent data from both animal models of COPD (Figure <xref ref-type="fig" rid="F1">1</xref>, own unpublished observations) as well as from studies in human subjects suggest that a fourth mechanism might be involved in the pathogenesis of COPD: disruption of the balance between apoptosis and replenishment of structural cells in the lung might contribute to the destruction of lung tissue in response to cigarette smoke, leading to emphysema.</p><fig position="float" id="F1"><label>Figure 1</label><caption><p><bold>Identification of apoptotic cells in the lung in an animal model of COPD</bold>. TUNEL-staining demonstrating an increase in apoptotic cells (dark brown nuclei) in the lungs of mice exposed to cigarette smoke (Figure 1B-C) compared to air-exposed animals (Figure 1A). Double stainings against pro-surfactant-protein C (Figure 1B, alveolar epithelial cells identified as blue pro-surfactant-protein C+ cells) or CD45 (Figure 1C, inflammatory cells identified as blue CD45+ cells) and TUNEL-staining (Figure 1B-C, dark brown nuclei) demonstrating apoptosis of both structural and inflammatory cells (own unpublished data).</p></caption><graphic xlink:href="1465-9921-7-53-1"/></fig><sec><title>Apoptosis</title><p>Apoptosis is a tightly regulated mechanism of cell death. This programmed cell death allows the elimination of unwanted, damaged or infected cells. At present, three different pathways that are involved in the regulation of apoptosis have been described (Figure <xref ref-type="fig" rid="F2">2</xref>). Different caspases (these are proteases with an important function in the regulation of apoptosis) are involved in these different pathways [<xref ref-type="bibr" rid="B5">5</xref>]. A first pathway is activated in response to extracellular signals and is mediated by binding of members of the tumour necrosis factor family (e.g. Fas ligand, TNF-α) to death receptors on the cell surface (e.g. Fas, TNFR). This results in the multimerization of the death receptor and the formation of the death inducing signalling complex (DISC), containing multiple adaptor molecules such as the Fas associated death domain (FADD). This FADD interacts with caspase-8. [<xref ref-type="bibr" rid="B6">6</xref>,<xref ref-type="bibr" rid="B7">7</xref>], leading to the autolytic activation from pro-caspase-8 to caspase-8. Caspase-8 then activates caspase-3 [<xref ref-type="bibr" rid="B8">8</xref>,<xref ref-type="bibr" rid="B9">9</xref>], which finally executes apoptosis by releasing caspase-activated DNAse (CAD) from its inhibitor (ICAD) with DNA fragmentation as a consequence [<xref ref-type="bibr" rid="B10">10</xref>]. This pathway is called the receptor mediated extrinsic pathway (Figure <xref ref-type="fig" rid="F2">2</xref>). Importantly, caspase-8 can also cleave the pro-apoptotic Bid [<xref ref-type="bibr" rid="B11">11</xref>], which then, through interaction with Bax and Bak, translocates to the mitochondria and causes the release of cytochrome C (see below). [<xref ref-type="bibr" rid="B12">12</xref>].</p><fig position="float" id="F2"><label>Figure 2</label><caption><p><bold>Schematic representation of different pathways involved in apoptosis</bold>. <bold>Extrinsic pathway: </bold><bold>1. </bold><inline-graphic xlink:href="1465-9921-7-53-i1.gif"/><bold>Ligand-death-receptor pathway (green): </bold>death factors such as Fas ligand (FasL) and tumour necrosis factor (TNF) trigger apoptosis by binding on 'death receptors' such as Fas and Tumour Necrosis Factor Receptor 1 (TNFR1). FasL may be solubilized to sFasL by matrix metalloproteinases (MMP's). The death receptors recruit procaspase-8 by means of an adaptor protein, Fas associated death domain protein (FADD). After cleavage the mature caspase-8 then directly activates caspase-3 or cleaves Bid. Truncated Bid (tBid) interacts with Bax and Bak. A pore is formed in the outer mitochondrial membrane through which cytochrome c (Cyt C) is released. <bold>2. </bold><inline-graphic xlink:href="1465-9921-7-53-i2.gif"/><bold>Cytolytic effector cell pathway (orange): </bold>cytotoxic T cells can release granzyme B and perforin, a pore-forming protein. Granzyme B activates caspase-3 through cleavage. It can also cleave caspase-8. <bold>3. </bold><inline-graphic xlink:href="1465-9921-7-53-i3.gif"/><bold>Growth factor depletion pathway (red): </bold>deprivation of survival factors triggers Cyt C release through activation of Bax and Bak. <bold>Intrinsic pathway: </bold><bold>4. </bold><inline-graphic xlink:href="1465-9921-7-53-i4.gif"/><bold>Mitochondrial pathway (grey): </bold>mitochondria release cytochrome c (Cyt C) in response to stress. Together with apoptotic protease activating factor-1 (Apaf-1) and procaspase-9, Cyt C will form the apoptosome complex. This results in the proteolytic activation of the procaspase. Mature caspase-9 can then proteolytically activate caspase-3 and other executioner caspases. <bold>5. </bold><inline-graphic xlink:href="1465-9921-7-53-i5.gif"/><bold>Endoplasmatic reticulum pathway (blue): </bold>the ER can also induce apoptosis as a reaction to stress. It might do so by stimulating the mitochondrial pathway or by directly targeting the nucleus. In mice both caspase-7 and -12 are linked to this pathway. These different initiation pathways converge further downstream into activation of caspase-3. The effector caspase-3 cleaves ICAD (inhibitor of CAD) and releases it from CAD (caspase-activated DNAase). CAD translocates from the cytoplasm to the nucleus and can now act as active endonuclease and fragment DNA.</p></caption><graphic xlink:href="1465-9921-7-53-2"/></fig><p>A second pathway, the mitochondrial intrinsic pathway, responds to physical and chemical stress signals by the release of cytochrome C from mitochondria. As a consequence, cytochrome C, apoptotic protease activating factor-1 (Apaf-1) and caspase-9 form the apoptosome. [<xref ref-type="bibr" rid="B13">13</xref>-<xref ref-type="bibr" rid="B15">15</xref>], resulting in the activation of caspase-9 which then activates caspase-3 and initiates the execution of apoptosis [<xref ref-type="bibr" rid="B16">16</xref>].</p><p>Finally, in the endoplasmatic reticulum pathway, caspase-12 is activated in response to stress signals such as hypoxia. [<xref ref-type="bibr" rid="B17">17</xref>,<xref ref-type="bibr" rid="B18">18</xref>]. In addition to these caspase-dependent pathways, it has been shown that noncaspase proteases can process and activate caspases directly (eg activation of caspase-3 by granzyme B) [<xref ref-type="bibr" rid="B19">19</xref>]. Moreover, the deprivation of survival signals such as growth factors can also induce apoptosis by mitochondrial release of cytochrome C. [<xref ref-type="bibr" rid="B20">20</xref>,<xref ref-type="bibr" rid="B21">21</xref>].</p><p>While apoptosis is a tightly regulated active mechanism leading to programmed cell death, cell necrosis is a rather uncontrolled process, which is accidental and in which the cell has no active role [<xref ref-type="bibr" rid="B22">22</xref>]. As a consequence, apoptosis seems to be a process that can be modulated to maintain cell viability, while necrosis is difficult to prevent.</p><p>Apoptosis is critical for the maintenance of normal tissue homeostasis and is in equilibrium with proliferation and differentiation. There is increasing evidence that disturbance of the balance between apoptosis and proliferation in lung tissue contributes to the pathogenesis of COPD. A limited number of descriptive studies in human subjects suggest a possible role for apoptosis in COPD, while an increasing number of experimental studies in animal models of COPD provides more insight into the association between cigarette smoking, apoptosis and the development of emphysema.</p></sec><sec><title>Studies on human subjects</title><sec><title>Apoptosis in human lung: ex vivo/in vitro</title><p>Several groups studied the role of apoptosis in the pathogenesis of COPD in human subjects, mostly by using lung tissue sections from COPD patients and controls (Table <xref ref-type="table" rid="T1">1</xref>). Segura-Valdez et al described an increase in endothelial cell apoptosis in lung tissue sections from COPD patients compared to controls. Although less frequently, apoptotic alveolar epithelial cells, interstitial cells and inflammatory cells (neutrophils and lymphocytes) were also described in the lungs of COPD patients, while this was not the case in control subjects [<xref ref-type="bibr" rid="B23">23</xref>].</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>Overview of studies on apoptosis in human lung.</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="center"><bold>Reference</bold></td><td align="center"><bold>Patients</bold></td><td align="center"><bold>COPD defined by</bold></td><td align="center"><bold>Controls</bold></td><td align="center"><bold>Increased Apoptosis in patients</bold></td><td align="center"><bold>Increased Proliferation in patients</bold></td></tr></thead><tbody><tr><td align="left">Segura-Valdez [23]</td><td align="left">Chronic Bronchitis Emphysema</td><td align="left">-Medical history<break/>-Pulmonary Function<break/>-Histology/CT</td><td align="left">-Male individuals who died from other causes than lung diseases smoking status unknown</td><td align="left">-endothelial cells<break/>-alveolar epithelial cells<break/>-interstitial cells<break/>-inflammatory cells</td><td align="left">NA</td></tr><tr><td align="left">Majo [27]</td><td align="left">Emphysema</td><td align="left">-Pulmonary Function<break/>-Histology</td><td align="left">-Never smokers<break/>-Smokers without emphysema</td><td align="left">-No difference between groups</td><td align="left">NA</td></tr><tr><td align="left">Kasahara [28]</td><td align="left">Emphysema</td><td align="left">-Pulmonary Function</td><td align="left">Non smokers</td><td align="left">-alveolar epithelial cells</td><td align="left">NA</td></tr><tr><td></td><td></td><td align="left">-Histology/CT</td><td align="left">-Smokers without airway obstruction</td><td align="left">-endothelial cells</td><td></td></tr><tr><td align="left">Yokohori [25]</td><td align="left">Emphysema</td><td align="left">-Pulmonary Function<break/>-Histology</td><td align="left">-Asymptomatic smokers<break/>-Asymptomatic nonsmokers</td><td align="left">-alveolar epithelial cells type II</td><td align="left">alveolar epithelial cells type II</td></tr><tr><td align="left">Imai [24]</td><td align="left">Emphysema</td><td align="left">-Pulmonary Function<break/>-Histology</td><td align="left">-accidental death victims unused donor lungs for LTX<break/>-Smoking status unknown</td><td align="left">-alveolar epithelial cells<break/>-endothelial cells<break/>-mesenchymal cells</td><td align="left">Increased (cell type not specified)</td></tr><tr><td align="left">Hodge [26]</td><td align="left">COPD</td><td align="left">-Medical history<break/>-Pulmonary Function</td><td align="left">-Never smokers</td><td align="left">-airway epithelial cells (obtained by BAL)<break/>-BAL T-cells</td><td align="left">NA</td></tr></tbody></table><table-wrap-foot><p>BAL = bronchoalveolar lavage; CT = computed tomography; LTX = lung transplantation; NA = not available</p></table-wrap-foot></table-wrap><p>Imai and colleagues described an increase in apoptotic cells (alveolar epithelial cells, endothelial cells and mesenchymal cells) in emphysematous lung tissue, as well as an increase in the activated subunits of caspase-3 (an important caspase in the execution of downstream events in apoptosis). Moreover, expression of the pro-apoptotic proteins Bax and Bad was detected in emphysema patients, while this was not the case in controls. The anti-apoptotic protein Bcl-2 was not detected in either normal or emphysematous lung tissue. Interestingly, increased cell proliferation was found in emphysematous lungs [<xref ref-type="bibr" rid="B24">24</xref>].</p><p>Other groups described similar findings, with an increase in both apoptosis and proliferation of alveolar wall cells in patients with emphysema compared to smokers without COPD and non-smokers [<xref ref-type="bibr" rid="B25">25</xref>]. Hodge et al described an increase in apoptosis of alveolar epithelial cells and T-cells from bronchial brushings and bronchoalveolar lavage in COPD patients compared to non-smoking controls. [<xref ref-type="bibr" rid="B26">26</xref>]. This increase in apoptosis in COPD patients persisted despite smoking cessation.</p><p>Others, on the contrary, did not find a significant difference in apoptotic alveolar wall cells in the lungs from smokers without emphysema compared to smokers with emphysema. [<xref ref-type="bibr" rid="B27">27</xref>]. However, in this study, apoptosis in smokers showed a bilinear relationship with the amount smoked: the apoptotic index decreased in smokers without emphysema to a minimum at 40 pack year, then increasing sharply as the pack year increased in smokers with emphysema. Interesting findings were obtained by Kasahara et al. These authors demonstrated an increase in apoptotic epithelial and endothelial alveolar septal cells in emphysematous lungs compared to non-smokers, smokers and primary pulmonary hypertension patients. [<xref ref-type="bibr" rid="B28">28</xref>]. Moreover, expression of VEGF and VEGF R2 protein and mRNA was significantly reduced in emphysema. The authors hypothesized that this decrease of endothelial cell maintenance factors, leading to endothelial alveolar septal death, may be part of the pathogenesis of emphysema. Recent data from other groups support this finding, by demonstrating that VEGF levels in induced sputum from COPD patients decreased with severity of COPD. [<xref ref-type="bibr" rid="B29">29</xref>]. However, while VEGF signalling may be required for the maintenance of the alveolar structures, the 936 C/T polymorphism of the VEGF gene (associated with lower VEGF plasma levels) was not associated with the development of COPD [<xref ref-type="bibr" rid="B30">30</xref>].</p><p>Altogether, several studies in human COPD patients describe an increase in apoptosis, especially in structural cells in the lung (Table <xref ref-type="table" rid="T1">1</xref>). However, some points should be taken into consideration when interpreting these data. First, not all studies have studied changes in lung cell proliferation in addition to apoptosis. As mentioned above, in physiologic circumstances, apoptosis is in balance with processes such as proliferation and differentiation. As a consequence, when studying the role of apoptosis in diseases such as COPD, it is recommendable to evaluate changes in proliferation as well. By doing so, this will allow to discriminate between a net increase in apoptosis (not counterbalanced by an increase in proliferation and leading to the loss of structural lung cells and tissue) and an equal increase in both apoptosis and proliferation (where the loss of structural cells by apoptosis is prevented by the regeneration of structural cells in the lung). By using this approach, Calabrese and colleagues recently demonstrated that there was a significant increase in apoptotic alveolar epithelial cells in end-stage emphysema (particularly in emphysema due to α1-antitrypsin deficiency), while there was no difference in proliferation of alveolar septal cells between emphysema patients and controls [<xref ref-type="bibr" rid="B31">31</xref>].</p><p>Second, in some of these studies, the control groups used consisted of non-smokers or of smokers with significantly less pack years of cigarette smoking compared to the COPD groups. Strictly speaking, one cannot exclude the possibility that the increase in apoptosis of structural cells is only related to cigarette smoking per se, rather than being an event that is specifically associated with the development of COPD. The ideal situation would be to compare COPD patients with heavy smokers who did not develop COPD. Another confounding factor could be the difference in treatment between patients. Most of the studies discussed above do not discriminate between COPD patients that are treated with inhaled corticosteroids and those who are not. It has been demonstrated that corticosteroids induce apoptosis of airway epithelial cells and eosinophils in asthma. [<xref ref-type="bibr" rid="B32">32</xref>]. No such data are available for COPD, but these findings underscore the importance of taking into account the use of inhaled steroids when examining apoptosis in the airways of COPD patients.</p><p>Thirdly, there are important differences between the different studies regarding the patient population: some groups identified COPD patients by pulmonary function tests, while others studied mainly emphysema patients, as defined by the use of radiologic or histological data. From these studies, it is unclear whether apoptosis is an underlying disease mechanism only for the development of emphysema, or on the contrary, if it is also involved in the disease process of COPD patients without emphysema (and with predominantly bronchiolitis).</p><p>Finally, while most groups studied apoptosis of structural cells, it would be interesting to evaluate changes in apoptosis of inflammatory cells in the lungs of COPD patients as well. It has been suggested that chronic inflammation in the airways might result from reduced apoptosis of inflammatory cells, with accumulation of inflammatory cells and sustained inflammation as a consequence [<xref ref-type="bibr" rid="B33">33</xref>].</p></sec><sec><title>Apoptosis outside the lungs</title><p>COPD is currently regarded as a multi-component disease with systemic manifestations in addition to local pulmonary inflammation [<xref ref-type="bibr" rid="B34">34</xref>]. The lung is of course the principal organ affected by the disease, but the pulmonary manifestations of the disease are often accompanied by systemic abnormalities. This seems to be the case for the disturbance of the balance between apoptosis and regeneration too: while apoptosis of structural lung cells has been demonstrated in COPD patients, several groups described alterations in apoptosis or apoptotic signals in the systemic circulation or in skeletal muscles from COPD patients.</p><p>An increased propensity of peripheral blood T cells in COPD to undergo apoptosis has been described [<xref ref-type="bibr" rid="B35">35</xref>]. This was accompanied by upregulation of several mediators involved in the induction of T cell apoptosis, such as TNF-α/TNFR1, Fas and TGFR. The authors hypothesized that increased rates of T cell apoptosis result in unbalanced homeostasis, defective clearance mechanisms and perpetuation of the inflammatory response.</p><p>Takabatake and colleagues described significantly higher TNF-α and sTNF-R55 and R75 levels in the circulation of COPD patients, while serum levels of soluble Fas ligand (sFas-L), an inducer of apoptosis, and plasma levels of the soluble Fas receptor (sFas), an inhibitor of apoptosis, were not increased in COPD patients. [<xref ref-type="bibr" rid="B36">36</xref>].</p><p>Others described a significant increase in sFas in plasma from severe COPD patients compared to patients with mild or moderate COPD, while sFas-L was within normal limits in all groups. [<xref ref-type="bibr" rid="B37">37</xref>].</p><p>Peripheral muscle weakness, due to muscle atrophy, is commonly observed in COPD patients. [<xref ref-type="bibr" rid="B38">38</xref>,<xref ref-type="bibr" rid="B39">39</xref>]. A possible mechanism of this muscle wasting could be a decrease in the number of muscle fibres resulting from activation of apoptotic pathways. It has been reported that skeletal muscle apoptosis is increased in patients with COPD having a low body mass index (BMI) compared to COPD patients with normal BMI and to healthy volunteers and is associated with a lower exercise capacity [<xref ref-type="bibr" rid="B40">40</xref>]. Osteoporosis is another systemic manifestation of COPD [<xref ref-type="bibr" rid="B41">41</xref>]. The precise mechanisms involved are unknown and it is unclear if apoptosis contributes to the development of osteoporosis in COPD patients. In summary, a limited number of studies investigated changes in apoptosis outside the lung in COPD patients. The relevance of these findings in the development of COPD is unknown. Future studies will need to investigate in more detail the relation between apoptosis in- and outside the lung in COPD and the importance of apoptosis in the development of systemic manifestations in the course of the disease.</p></sec></sec><sec><title>Animal models of COPD and emphysema</title><p>When interpreting data obtained from animal models of COPD, it is important to keep in mind that the development of emphysematous lesions in the lungs of cigarette smoke exposed mice is strain dependent. [<xref ref-type="bibr" rid="B42">42</xref>,<xref ref-type="bibr" rid="B43">43</xref>]. DBA/2 mice, for example, develop patchy emphysema in response to cigarette smoke faster than C57Bl/6J mice. Additionally, in DBA/2 mice, the development of emphysema is preceded by the appearance of apoptotic cells in areas with a low signal for VEGF-R2 [<xref ref-type="bibr" rid="B43">43</xref>], while this is much less the case for C57Bl/6J mice.</p><p>Moreover, animal models of COPD often do not represent all characteristics of the disease as it occurs in humans. The current view on the pathogenesis of COPD is that cigarette smoke induces the recruitment of inflammatory cells, which then release reactive oxygen species and proteolytic enzymes, causing the degradation of lung matrix and the death of structural cells. However, in several animal models of COPD, development of emphysema was observed despite a remarkable lack of pulmonary inflammation. These studies demonstrated that, at least in animal models, apoptosis of alveolar wall or endothelial cells is sufficient to cause pulmonary emphysema, even without the accumulation of inflammatory cells. In those studies, emphysema was induced by directly targeting the alveolar cells [<xref ref-type="bibr" rid="B44">44</xref>], or by inactivating VEGFR [<xref ref-type="bibr" rid="B45">45</xref>-<xref ref-type="bibr" rid="B47">47</xref>] or VEGF [<xref ref-type="bibr" rid="B48">48</xref>].</p><p>The relationship between VEGF, endothelial cell apoptosis and emphysema has first been described by Kasahara and colleagues [<xref ref-type="bibr" rid="B45">45</xref>]. These authors demonstrated in a rat model that blocking of the VEGF receptor induced alveolar cell apoptosis and led to enlargement of the airspaces. Moreover, treatment with a caspase inhibitor prevented septal cell apoptosis and emphysema development in response to the blocking of VEGF receptor.</p><p>Other groups showed that cathepsin S-dependent epithelial cell apoptosis is a critical event in the pathogenesis of IFN-γ induced emphysema [<xref ref-type="bibr" rid="B49">49</xref>]. Using this model, it was recently shown that IFN-γ is a potent activator of the extrinsic/death receptor and intrinsic/mitochondrial apoptosis pathways and that these activation events are partially CCR5 dependent [<xref ref-type="bibr" rid="B50">50</xref>].</p><p>Aoshiba and colleagues demonstrated that intratracheal administration of active caspase-3 resulted in epithelial apoptosis, enhanced elastolytic activity in BAL and the development of emphysematous changes in mice [<xref ref-type="bibr" rid="B44">44</xref>].</p><p>Recently, Petrache et al reported that intratracheal instillation of ceramide, a highly regulated sphingolipid second messenger, triggers apoptosis of alveolar epithelial and endothelial cells and induced airspace enlargement in mice [<xref ref-type="bibr" rid="B47">47</xref>]. Moreover, increased lung ceramide levels were detected in the lungs of emphysema patients, suggesting that ceramide upregulation might be an important pathogenetic element in the development of emphysema. This was the first study to describe the involvement of a non-protein mediator of apoptosis in the pathogenesis of emphysema.</p><p>Others have suggested that humoral- and CD4+ cell- dependent mechanisms can lead to alveolar septal cell apoptosis and the development of emphysema [<xref ref-type="bibr" rid="B51">51</xref>]. Intraperitoneal injection of endothelial cells led to the production of antibodies against endothelial cells, influx of CD4+ T cells in the lung, alveolar septal apoptosis, activation of matrix metalloproteinases and the development of emphysema.</p></sec><sec><title>Apoptosis and interaction with other pathogenetic mechanisms in COPD</title><p>As mentioned before, several disease mechanisms are involved in the development of COPD: inflammation, proteinase/anti-proteinase imbalance and oxidative stress. Apoptosis interacts with all of these pathways, adding to the complexity of the disease (Figure <xref ref-type="fig" rid="F3">3</xref>).</p><fig position="float" id="F3"><label>Figure 3</label><caption><p><bold>Interaction of apoptosis with other pathogenetic mechanisms in COPD, including inflammation, oxidative stress and protease/anti-proteinase imbalance</bold>. 1. Neutrophil elastase (NE) cleaves the phosphatidylserine receptor on macrophages, resulting in impaired clearance of apoptotic cells and sustained inflammation [54]. 2. Cytotoxic CD8+ T-cells cause apoptosis of alveolar epithelial cells through the release of perforins and granzyme-B [57,58]. 3. Degradation of the basement membrane (BM) by matrix metalloproteinases (MMPs) leads to loss of survival signals and induces apoptosis of epithelial cells [60]. 4. Apoptosis may also be affected by direct proteolysis of death-inducing signals. It has been shown that MMP-7 sheds and activates Fas ligand (FasL) that is produced by epithelial cells, thereby mediating apoptosis [62]. 5. Oxidative stress could lead to a reduction of Vascular Endothelial Growth Factor (VEGF) levels, resulting in apoptosis of alveolar endothelial cells [29]. NE: neutrophil elastase; BM: basement membrane; MMPs: matrix metalloproteinases; TIMP: tissue inhibitor of metalloproteinase; α1-AT: α1-anti-trypsin; FasL: Fas ligand; VEGF: vascular endothelial growth factor.</p></caption><graphic xlink:href="1465-9921-7-53-3"/></fig><sec><title>Apoptosis and inflammation</title><p>Several groups demonstrated that, at least in animal models, apoptosis of alveolar wall or endothelial cells is sufficient to cause pulmonary emphysema, even without the accumulation of inflammatory cells. However, in the lungs of COPD patients, there is an impressive influx of inflammatory cells and interactions between inflammatory and apoptotic mechanisms most probably take place.</p><p>Alveolar macrophages from patients with COPD are less effective in phagocytosing apoptotic airway epithelial cells compared to controls [<xref ref-type="bibr" rid="B51">51</xref>]. This might be mediated by the presence of activated numbers of neutrophils in COPD. [<xref ref-type="bibr" rid="B52">52</xref>,<xref ref-type="bibr" rid="B53">53</xref>]. It has been shown that neutrophil elastase cleaves the phosphatidylserine receptor on macrophages, resulting in impaired clearance of apoptotic cells and sustained inflammation [<xref ref-type="bibr" rid="B54">54</xref>].</p><p>In addition to neutrophils and macrophages, there is also an increase in CD8+ T-lymphocytes in the lungs of COPD patients. [<xref ref-type="bibr" rid="B55">55</xref>,<xref ref-type="bibr" rid="B56">56</xref>]. These cytotoxic CD8+ T-cells could cause apoptosis of alveolar epithelial cells through the release of perforins, granzyme-B and TNF-α. [<xref ref-type="bibr" rid="B57">57</xref>,<xref ref-type="bibr" rid="B58">58</xref>].</p></sec><sec><title>Apoptosis and proteinase-antiproteinase imbalance</title><p>The increase in proteolytic activity in the lungs of COPD patients might interfere with apoptosis in several ways. [<xref ref-type="bibr" rid="B59">59</xref>]. The basal membrane contains signals for cell survival and loss of these survival signals (as a consequence of degradation of the basement membrane by matrix metalloproteinases) can induce apoptosis. This process of apoptosis induced by loss of appropriate cell-matrix contacts (called anoikis) is involved in tissue homeostasis by maintaining the correct cell number of high turnover epithelial tissues. [<xref ref-type="bibr" rid="B60">60</xref>]. Aoshiba and colleagues already suggested that cell-extracellular matrix interactions modulate apoptosis in bronchial epithelium. [<xref ref-type="bibr" rid="B61">61</xref>]. Moreover, they recently reported elastolytic activity in apoptotic lung epithelial cells in a mouse model of emphysema [<xref ref-type="bibr" rid="B44">44</xref>]. Apoptosis may also be affected by direct proteolysis of death-inducing signals. It has been shown that MMP-7 sheds and activates Fas ligand that is produced by epithelial cells, thereby mediating apoptosis [<xref ref-type="bibr" rid="B62">62</xref>]. Finally, recent data suggest that MMP-8 has anti-inflammatory effects on airway inflammation due to a regulation of inflammatory cell apoptosis [<xref ref-type="bibr" rid="B63">63</xref>].</p></sec><sec><title>Apoptosis and oxidative stress</title><p>In a rat model of emphysema induced by VEGFR blockade, Tuder et al demonstrated that apoptosis predominated in the lung in areas of oxidative stress and that experimental blockade of apoptosis markedly reduced the expression of markers of oxidative stress [<xref ref-type="bibr" rid="B46">46</xref>]. The administration of a compound with antioxidant activity prevented the development of alveolar cell apoptosis and airspace enlargement, suggesting a positive feedback interaction between oxidative stress and apoptosis. Other groups have shown that mice with impaired expression of antioxidant genes have increased numbers of apoptotic alveolar septal cells (predominantly endothelial and type II epithelial cells) and develop early and extensive emphysema in response to cigarette smoke. [<xref ref-type="bibr" rid="B64">64</xref>]. Recent work from Kanazawa and colleagues nicely demonstrated elevated oxidative stress levels and a reciprocal reduction of VEGF levels in induced sputum from COPD patients. [<xref ref-type="bibr" rid="B29">29</xref>]. These changes increased with severity of the disease. These findings confirm the relationship between oxidant-antioxidant imbalance and VEGF-dependent homeostasis of alveolar walls in the lungs of COPD patients. The authors hypothesize that epithelial cell injury mediated by oxidative stress induces a decrease in lung VEGF levels, resulting in the development of COPD. These data clearly point out that apoptosis is not an isolated event in the development of COPD and that it interferes with other underlying disease mechanisms.</p></sec></sec></sec><sec><title>Conclusion</title><p>An increasing number of data, both from animal models and studies on human subjects, supports an important role for apoptosis in the pathogenesis of COPD.</p><p>More studies are needed to identify the most important apoptotic pathways and the caspases involved in the development of COPD. Moreover, it has to be evaluated if apoptosis can be used as a therapeutic target to prevent further deterioration of the disease, which occurs even after the patients have quit smoking. Finally, as disturbance of the balance between apoptosis and regeneration of structural lung cells seems important in the destruction of healthy lung tissue, it would be interesting to evaluate the potential use of stem cell therapy for emphysema.</p></sec><sec><title>Abbreviations</title><p>CCR5: chemokine receptor 5</p><p>COPD: Chronic Obstructive Pulmonary Disease</p><p>IFN-γ: interferon-γ</p><p>MMP: Matrixmetalloproteinase</p><p>NE: Neutrophil Elastase</p><p>SFas: soluble Fas receptor</p><p>sFas-L: soluble Fas ligand</p><p>STNF-R: Soluble TNF receptor</p><p>TGFR: Transforming Growth Factor Receptor</p><p>TIMP: Tissue Inhibitor of Metalloproteinases</p><p>TNF: Tumour Necrosis Factor</p><p>TNFR: Tumour Necrosis Factor Receptor</p><p>VEGF: Vascular Endothelial Growth Factor</p><p>VEGF R2: Vascular Endothelial Growth Factor Receptor 2</p><p>WHO: World Health Organisation</p></sec><sec><title>Competing interests</title><p>The author(s) declare that they have no competing interests.</p></sec> |
An analysis of intra array repeats: the good, the bad and the non informative | <sec><title>Background</title><p>On most common microarray platforms many genes are represented by multiple probes. Although this is quite common no one has systematically explored the concordance between probes mapped to the same gene.</p></sec><sec><title>Results</title><p>Here we present an analysis of all the cases of multiple probe sets measuring the same gene on the Affymetrix U133a GeneChip and found that although in the majority of cases both measurements tend to agree there are a significant number of cases in which the two measurements differ from each other. In these cases the measurements can not be simply averaged but rather should be handled individually.</p></sec><sec><title>Conclusion</title><p>Our analysis allows us to provide a comprehensive list of the correlation between all pairs of probe sets that are mapped to the same gene and thus allows microarray users to sort out the cases that deserve further analysis. Comparison between the set of highly correlated pairs and the set of pairs that tend to differ from each other reveals potential factors that may affect it.</p></sec> | <contrib id="A1" contrib-type="author"><name><surname>Elbez</surname><given-names>Yedid</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>yedid.e@gmail.com</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Farkash-Amar</surname><given-names>Shlomit</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>shlomitf@md.huji.ac.il</email></contrib><contrib id="A3" corresp="yes" contrib-type="author"><name><surname>Simon</surname><given-names>Itamar</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>itamars@md.huji.ac.il</email></contrib> | BMC Genomics | <sec><title>Background</title><p>High density DNA microarrays are extensively used by biologists in order to gain information about the expression levels of thousands of genes simultaneously. The advantage of the genomic approach is that it allows the study of cell behavior on the systems level rather than on the gene specific level [<xref ref-type="bibr" rid="B1">1</xref>]. One of the hurdles of this approach is that the confidence one has in each measurement is lower in comparison to the confidence gained by classical molecular biology techniques[<xref ref-type="bibr" rid="B2">2</xref>], thus extensive statistical analysis is needed in order to define the significance of the results. One of the approaches used to asses the quality of results is comparison between replicates [<xref ref-type="bibr" rid="B3">3</xref>]. This is accomplished by both printing each DNA fragment several times on the array (internal replicates) and repeating the experiment multiple times (experimental replicates). While internal replicates usually give very similar results, experimental replicates may differ from one another due to variation in the biological samples, exact experimental conditions and array processing [<xref ref-type="bibr" rid="B4">4</xref>]. There is yet another type of replicates – multiple array spots that are mapped to different regions of the same gene, thus allowing multiple measurements of the RNA levels in the exact same biological sample on the same array. We call these "intra array repeats" and not "intra array replicates" in order to highlight the fact that they are not identical. Although this is quite common, for example 35% of the genes on Affymetrix U133a array are represented by multiple probe sets, no systematic evaluation of the concordance between multiple representations of the same gene has been reported, thus leaving microarray users without clear guidelines how to deal with this problem.</p><p>A related issue which has been studied extensively in recent years is the question of cross platform comparisons in which researchers try to compare results obtained for the same gene on different microarray platforms. Overall there is only limited agreement between experiments performed on different platforms (reviewed by [<xref ref-type="bibr" rid="B5">5</xref>,<xref ref-type="bibr" rid="B6">6</xref>]). Many factors can contribute to this cross platform inconsistency, including differences in RNA processing, noise levels [<xref ref-type="bibr" rid="B7">7</xref>], the statistical analysis required by each platform [<xref ref-type="bibr" rid="B8">8</xref>] and the sequences chosen to represent the gene by the different manufacturers [<xref ref-type="bibr" rid="B9">9</xref>,<xref ref-type="bibr" rid="B10">10</xref>]. Recent studies suggest that the latter issue is very important since there is much more agreement between platforms when only spots with overlapping sequences are used [<xref ref-type="bibr" rid="B11">11</xref>-<xref ref-type="bibr" rid="B13">13</xref>]. In contrast to cross platform comparisons which are affected by many variables, intra platform comparisons may be affected only by the sequences chosen to represent the gene and thus may serve as an excellent test case for assessing the quality of microarrays results. In this study we identified all groups of probe sets mapped to the same gene that are present on the Affymetrix U133a GeneChip and examined the degree of agreement between those pairs. In most cases we observed significant correlations between the pairs suggesting that both probe sets are reporting similar results. However, we did find a large subset that differs significantly from one another. The identification of these pairs is important for analyzing expression profile results, since in these cases the investigator can not rely on the expression levels reported by those probes.</p></sec><sec><title>Results</title><sec><title>Gene expression data</title><p>In order to compare RNA measurements from two different regions of a gene on the same array, we chose to explore the Affymetrix U133a platform since it is a very popular platform with thousands of publicly available datasets. The U133a platform contains 20,267 probe sets which are mapped to 12,942 distinct genes with 4,552 of them represented by multiple probe sets (figure <xref ref-type="fig" rid="F1">1</xref>). Overall this creates 11,722 pairs of probe sets in which both members are mapped to the same gene. For our analysis we chose four publicly available data sets each containing at least 29 experiments (table <xref ref-type="table" rid="T1">1</xref>) and downloaded the data from the Gene Expression Omnibus database. Each of these datasets were acquired in different laboratories, using different cells and growth conditions, thus we analyzed each of them separately.</p><p>Since we are interested in comparing the changes in transcriptional level of both members of each pair of probe sets, we defined as informative only probe sets that showed changes in transcription along the different measurements in a dataset. Only those probe sets were included in our analysis (see methods).</p></sec><sec><title>Correlation coefficient</title><p>In order to asses the overall agreement within pairs of probe sets, we decided to look at the Pearson correlation between the set of measurements for each of the probe sets in each dataset. We chose the Pearson correlation as our metric since it is not sensitive to differences in scaling and average expression levels but rather compares the degree of linear relationship between the two probe-sets which is a better indication for the overall similarity. We calculated the correlation coefficient of all the informative pairs in each dataset. Histograms of these correlations and correlations obtained by a random list of pairs are shown in figure <xref ref-type="fig" rid="F2">2a</xref>. The statistical significance of the deviation of the correlation of each pair from the underlying distribution was calculated, and the results are summarized in <xref ref-type="supplementary-material" rid="S1">additional file 1</xref>.</p><p>Comparison of the correlation values obtained in the different datasets revealed a high degree of correlation between them (figure <xref ref-type="fig" rid="F2">2b,c</xref>), suggesting that the agreement between pairs of probe sets is characteristic of the probe sets and is not heavily dependent on the particular dataset from which the data was derived.</p></sec><sec><title>Statistical assessment of the results</title><p>One of the main purposes of our analysis was to guide microarray users in cases of genes that are represented by multiple probe sets. The users would like to know when the multiple measurements can be combined and when they should be handled with caution. In order to achieve this goal, we wanted to define a list of pairs which show overall good agreement between them. This list should be based on the combination of the correlation values calculated in each dataset. Thus for each pair, we combined the individual P values obtained for each dataset in which it was informative, resulting in a single combined P value for each pair. A list of all the pairs along with their combined P values is included in <xref ref-type="supplementary-material" rid="S1">additional file 1</xref>. This list includes 6536 pairs of probe sets that were informative at least in one of the data sets and 77% of them showed significant correlation (FDR = 0.01).</p><p>Microarrays users frequently ask themselves what to do when a gene is represented by multiple probes. Our analysis provides a partial solution to this problem by mapping all the pairs of probe sets on the Affymetrix U133a platform and providing a significant measurement (P value) as to whether the transcription levels reported by the two different probe sets correlates. In order to demonstrate that the combined P value may be useful in analyzing microarray data, we calculated the number of cases in which averaging the intensity reported by a pair of probe sets is misleading. The clear dependency between the error rate level and the P value (figure <xref ref-type="fig" rid="F3">3</xref>) suggests that the P value is a useful guideline for dealing with intra array repeats.</p></sec><sec><title>Exploring the "good" and "bad" pairs</title><p>Recent studies have shown that when performing cross platform comparisons between Affymetrix arrays and cDNA arrays, restricting the analysis to sequence-matched probes produces a higher level of consistency between the platforms [<xref ref-type="bibr" rid="B11">11</xref>-<xref ref-type="bibr" rid="B14">14</xref>]. These studies suggest that one of the main reasons for inconsistencies between different microarray platforms is the differences in the gene regions chosen to be printed for each platform. Different sequences may report different values for a variety of reasons that can be sorted into two categories – the two sequences are reporting the levels of different transcripts (due to annotation mistakes or different splice variants) alternatively, both sequences may report the transcription of the same transcript at a different accuracy (due to the well documented 3'-5' degradation of the mRNA or due to differences in hybridization efficiency of the two regions).</p><p>Intra array repeats are printed on the same array and therefore they share all the external conditions (including the biological sample, growth conditions, probe labeling and data handling) and differ from one another only in the particular sequence chosen to be printed on the array. Thus multiple probe sets printed on the same array serve as an excellent case for exploring the contribution of sequence effects to the overall deviation of the transcription values reported by each probe set. To this end we defined two large sub groups of pairs (each containing approximately 1500 pairs) – pairs that were highly correlated (P < 10<sup>-15</sup>; herein "good") and pairs which did not show a significant concordance between them (P > 0.1; herein "bad").</p><p>First we wanted to test the possibility that incorrect annotation contributes to the differences measured between pairs of probe sets. Recently, several reports claimed that Affymetrix annotation is not ideal and that using the sequences to reannotate the probe sets gives better results [<xref ref-type="bibr" rid="B15">15</xref>-<xref ref-type="bibr" rid="B17">17</xref>]. If mistakes in annotation were a major cause for the differences between pairs of probe sets we should expect to see a higher degree of misannotation in the "bad" pairs than in the "good" set of pairs. We calculated the percentage of cases in which Affymetrix gene assignment was supported by the GeneAnnot gene assignment. The results show that indeed misannotation has a big contribution to the disconcordance between the "bad" pairs, actually >25% of these pairs have no sequence support to the fact that they are reporting the transcript of the same gene (figure <xref ref-type="fig" rid="F4">4a</xref>).</p><p>Next we analyzed the affect of alternative splicing on probe accordance. If two probe sets are reporting the expression of two different splice variants of a gene, they may differ in the transcription values they are measuring. Thus another possible explanation for disagreement between pairs of probe sets is that each is reporting the transcription level of a different transcript of the same gene. In order to explore this idea, we looked at the number of known variant transcripts for each gene. Calculation of the average number of transcript variants in the "good" and in the "bad" pairs revealed that there are on average approximately 1.7 transcripts per gene and there is almost no difference in the distribution between the two groups (figure <xref ref-type="fig" rid="F4">4b</xref>). Moreover the frequency of the "bad" and of the "good" pairs is almost identical even in the group of genes with only a single known transcript, strongly suggesting that according to the current knowledge about alternative splicing, it can not be the reason for the deviation between pairs of probe sets in most cases.</p><p>Finally, we wanted to explore the contribution of cross hybridization between probes to the agreement between probe sets. To this end we used the Affymetrix suffix to the probe name as a measurement of the probe uniqueness. _at suffix designates a unique probe set, whereas _s_at and _x_at suffixes designate probe sets that can cross hybridize with multiple genes. To our surprise we found a very significant reverse correlation between the probe uniqueness and its agreement with another probe set. We found significant enrichment (P = 0.001) for the s_at and x_at suffixes in the "good" pairs (figure <xref ref-type="fig" rid="F4">4c</xref>), suggesting that cross hybridization between the genes is not the reason for the disagreement between probe sets.</p></sec></sec><sec><title>Discussion</title><p>Microarray users frequently encounter the problem of different values for the same gene resulting from multiple probe sets. Without clear guidelines from the manufacturers, each investigator deals with this problem in a different way, some are not aware of it and therefore either use the first appearance of the gene or use the probe set that fits their expectations better. Alternatively, users may use the average of all probe sets without validating that all indeed report similar transcription rates. In order to help microarray users to chose the right approach for multiple measurements of a gene in the Affymetrix U133a GeneChip we examined all such cases and developed a statistical method to asses the agreement between the measurements in each of these cases. Our method is based on the calculation of the Pearson correlation between all the measured values of a pair of probe sets along all the experiments in a data set. The Pearson correlation gives a good estimation of the overall agreement between such pairs without being sensitive to variability in the average expression level reported by the probe sets. We excluded from the analysis probes that were non informative either because their expression level was below detection or because their expression level did not change (>2 folds) along the experiments. As expected the correlation values obtained between probe sets from the same gene were much higher than those obtained from randomly chosen pairs of probe sets with 77% of the pairs reporting similar values. The pairs of probe sets where sorted according to the statistical significance of the Pearson correlation between them and a list of all the pairs together with their P value is provided [see <xref ref-type="supplementary-material" rid="S2">additional file 2</xref>]. This list should serve as a guideline to the user: the lower the P value the higher chances that the two probes are reporting the same information whereas high P values suggest that the two observations should be treated individually (figure <xref ref-type="fig" rid="F3">3</xref>).</p><p>The division of the pairs of probe sets into those that tend to agree ("good") or disagree ("bad") allows us to examine the factors affecting it. The deviation between two probes that are mapped to the same gene may be due to several reasons including alternative splicing, mistakes in the annotation and differences in the efficiency each probe set reports the transcription. We found no evidence for the contribution of alternative splicing to the disagreement between probe sets (figure <xref ref-type="fig" rid="F4">4b</xref>). On the other hand by comparing the annotation in the sets of the "good" and the "bad" pairs we have shown that misannotation is obviously a main issue, with >25% of the "bad" pairs being misannotated (figure <xref ref-type="fig" rid="F4">4a</xref>) thus in cases of disagreement between probe sets one should first validate that both probe sets are indeed annotated correctly.</p><p>Recent studies about cross platform comparisons found that the best results are obtained if the comparison is done on the probe level rather than on the probe set level [<xref ref-type="bibr" rid="B12">12</xref>,<xref ref-type="bibr" rid="B18">18</xref>,<xref ref-type="bibr" rid="B19">19</xref>], suggesting that the actual sequence has a big contribution to the array results. In accordance with this idea we found that pairs of probe sets that share at least one common probe show almost always significant agreement (97.6% of 970 such cases passed the FDR of 0.01).</p><p>The observation that non unique probe sets perform better than unique ones (figure <xref ref-type="fig" rid="F4">4c</xref>) is surprising because two probe sets each cross hybridize with a different set of genes may cause disagreement. However our results suggest the contrary – probes that were designed to recognize multiple genes, are much more abundant in the "good" pairs suggesting that the cross hybridization may have a smoothing affect that reduces the difference between the probes. In other words if the reason for the dis accordance between pairs is that each recognizes a unique transcript, less specific probe sets will mask this difference and thus will be more abundant among the "good" pairs of probe sets.</p><p>Our analysis suggests that in most cases (77%) pairs of informative probe sets report similar transcription levels. However, it should be noted that our decision to concentrate only on informative probe sets may influence those results. By excluding from the analysis the most problematic cases in which one probe set is informative and the other is not we elevated the percentage of cases with good correlation [see <xref ref-type="supplementary-material" rid="S3">additional file 3</xref>]. These cases were omitted from the analysis because of the limited usefulness of the Pearson correlation on non informative probe sets (see methods). Data about the concordance between these pairs is included in <xref ref-type="supplementary-material" rid="S1">additional file 1</xref>.</p><p>The analysis of the differences between the "good" and the "bad" pairs suggests that biological differences such as alternative splicing have a very limited contribution to the disagreement between probe sets. This conclusion was derived from the current knowledge of the genome sequence and of splicing variants (using fairly strict criteria). Updates of the genome sequence and advancements in the technology of measuring splice variants (such as the use of exon microarrays) may change this conclusion. Although those changes may affect the reasons for the disagreements between probe sets it will not change the definition of "good" and "bad" pairs. Only updating the probe set definition according to the new sequence information [<xref ref-type="bibr" rid="B18">18</xref>] may solve this problem but it is beyond the scope of the current study.</p></sec><sec><title>Conclusion</title><p>An important internal control for microarray experiments is multiple array spots that are mapped to different regions of the same genes. In the ideal world these different probe sets should give similar measurements for expression of the gene. Our analysis tested whether this is true using four data sets that used the Affymetrix U133a platform. We found that in the majority of such cases both measurements agree and thus using the average value of the probe sets would be appropriate. However in a significant number of cases this approach is not suitable and our results can guide the user in deciding about each case. Further analysis is needed to define guidelines for choosing better regions in the gene that will represent the transcription level more accurately.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>Data sets</title><p>For our analysis we downloaded the data of four large experiment series from the gene expression omnibus database [<xref ref-type="bibr" rid="B20">20</xref>]. We chose datasets with at least 29 experiments in order to have multiple measurements in each dataset. The GEO accession number, a short description of the experiments and the number of arrays used in each series are shown in table <xref ref-type="table" rid="T1">1</xref>. All the expression values were log transformed.</p></sec><sec><title>Pearson correlation</title><p>In order to define intra array repeats, we used Affymetrix annotation and defined 11,722 pairs of probe sets in which both members were mapped to the same gene. In order to compare the RNA levels reported by the two probes in each pair we calculated the Pearson correlation between the expression values (log transformed) reported by each probe set in all the experiments in each dataset. We also performed Spearman correlation and the results [see <xref ref-type="supplementary-material" rid="S4">additional file 4</xref>] are very similar [see <xref ref-type="supplementary-material" rid="S5">additional file 5</xref>].</p><p>Since we were interested only in the correlations between probe sets of active genes we looked only on probe sets that passed two filtering criteria – they got a present call (P) in at least 10% of the experiments in a dataset and they showed at least 2 fold changes in their expression level over the average in at least 10% of the experiments. This filtering resulted in a different number of informative pair of probe sets for each dataset (table <xref ref-type="table" rid="T1">1</xref>).</p><p>Our decision to include in the analysis only cases were both probe sets in a pair are informative was based on the observation that the transcription values reported by non informative probe sets frequently reflect random noise fluctuations and thus the meaning of the correlation between those values and transcription values from informative probes is not clear. Nevertheless this decision introduced a bias in our analysis since cases in which the two probe sets have clear different characteristics were excluded. Calculating the correlation of the group of pairs in which only one of the probe sets was informative revealed that in all data sets the average correlation of this group of pairs was much closer to the average correlation between random pairs than to the average correlation of the pairs in which both probe sets were informative [see <xref ref-type="supplementary-material" rid="S3">additional file 3</xref>].</p></sec><sec><title>Statistical significance</title><p>In order to calculate the statistical significance of the correlation between the probe sets in each pair we transformed the values into Fisher Z values and compared those values to the average correlation values obtained for random pairs of informative probe sets using a Z test. The P values obtained for each data set were combined using the chi square test for combined q probabilities [<xref ref-type="bibr" rid="B21">21</xref>].</p><p>In order to define a P value threshold while accounting for the multiple hypothesis problem we used the FDR approach [<xref ref-type="bibr" rid="B22">22</xref>].</p></sec><sec><title>Comparison between datasets</title><p>The fisher Z values obtained for each informative pair of probe sets in each dataset was used for comparing the results obtained in the different datasets. The comparison was done by calculating the Pearson correlation of those values. This procedure was repeated after shuffling the order of the pairs in order to obtain an estimation of the random correlation between the datasets.</p></sec><sec><title>Error rate estimation</title><p>Transcription values were converted to ratios by dividing the value reported in each experiment by the probe set mean value in all experiments in a dataset. Cases with ratios greater than |2| were identified. This was done for individual probe sets and for the average values of two probe sets mapped to the same gene. Cases without full agreement were counted.</p></sec><sec><title>Probe sets analysis</title><p>The accuracy of Affymetrix assignments of genes to probe sets was confirmed by comparing their assignments to those of the GeneAnnot database, which remapped Affymetrix probes to genes by performing blat searches with all individual probes forming the Affymetrix array [<xref ref-type="bibr" rid="B15">15</xref>]. This allowed us to divide the probe sets pairs into two categories – those which were mapped to the same gene using updated sequence information (85%) and those in which the current sequence information does not support affymetrix original annotation (15%). The number of transcripts mapped to each probe set was taken from the U133a annotation file provided by Affymetrix using the "RefSeq Transcript ID" column. Sequence homology between probe sets was determined by counting the number of identical probes shared by both probe sets in a pair. For the probe set uniqueness analysis we used the Affymetrix suffix. Affymetrix annotate each probe set with a suffix depicting whether it is unique to a single gene (_at), all its probes cross hybridize with multiple genes (s_at) or there is inconsistency between the probes forming the probe set in terms of cross hybridization with different genes (x_at) <ext-link ext-link-type="uri" xlink:href="http://www.affymetrix.com/support/technical/technotes/hgu133_design_technote.pdf"/>. We limited our analysis only to cases in which both probe sets in a pair have the same suffix.</p><p>The statistic significance of the difference between the distribution of "bad" and "good" probes was assessed with the chi square test or Fisher exact test.</p></sec></sec><sec><title>Authors' contributions</title><p>YE carried out most of the study and SF performed the Spearman correlation analysis. IS conceived the study, participated in its design and coordination and draft the manuscript. All authors read and approved the final 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</title><p>Excel file containing information about all the pairs of probe sets in all the four datasets. The information provided includes the Pearson correlation values, the individual P values and the metrics used to filter non informative probe sets.</p></caption><media xlink:href="1471-2164-7-136-S1.zip" mimetype="application" mime-subtype="x-zip-compressed"><caption><p>Click here for file</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="S2"><caption><title>Additional File 2</title><p>Excel file containing all the pairs of probe sets along with their combined P values.</p></caption><media xlink:href="1471-2164-7-136-S2.xls" mimetype="application" mime-subtype="vnd.ms-excel"><caption><p>Click here for file</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="S3"><caption><title>Additional File 3</title><p>Average Pearson correlation values between random pairs of probe sets (blue), pairs with only one informative probe set (red) and pairs in which both probe sets are informative (red). See methods for the definition of informative probe set.</p></caption><media xlink:href="1471-2164-7-136-S3.PDF" mimetype="text" mime-subtype="plain"><caption><p>Click here for file</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="S4"><caption><title>Additional File 4</title><p>Excel file containing information about the Spearman correlation values of all pairs of probe sets in all the four datasets.</p></caption><media xlink:href="1471-2164-7-136-S4.xls" mimetype="application" mime-subtype="vnd.ms-excel"><caption><p>Click here for file</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="S5"><caption><title>Additional File 5</title><p>Scatter plots showing the Pearson correlation values (X axis) against the Spearman correlation values (y axis) of all the informative pairs in the 4 data sets. Note that the spots are arranged along the diagonal with high R<sup>2 </sup>values. The only dataset with spots that are not along the diagonal is GSE1133 in which ~5% of the spots deviate from the diagonal by >2 folds, the remaining spots gave R<sup>2 </sup>= 0.93.</p></caption><media xlink:href="1471-2164-7-136-S5.pdf" mimetype="application" mime-subtype="pdf"><caption><p>Click here for file</p></caption></media></supplementary-material></sec> |
Microsatellite polymorphism across the <italic>M. tuberculosis </italic>and <italic>M. bovis </italic>genomes: Implications on genome evolution and plasticity | <sec><title>Background</title><p>Microsatellites are the tandem repeats of nucleotide motifs of size 1–6 bp observed in all known genomes. These repeats show length polymorphism characterized by either insertion or deletion (indels) of the repeat units, which in and around the coding regions affect transcription and translation of genes.</p></sec><sec><title>Results</title><p>Systematic comparison of all the equivalent microsatellites in the coding regions of the three mycobacterial genomes, viz. <italic>Mycobacterium tuberculosis </italic>H37Rv, <italic>Mycobacterium tuberculosis </italic>CDC1551 and <italic>Mycobacterium bovis</italic>, revealed for the first time the presence of several polymorphic microsatellites. The coding regions affected by frame-shifts owing to microsatellite indels have undergone changes indicative of gene fission/fusion, premature termination and length variation. Interestingly, the genes affected by frame-shift mutations code for membrane proteins, transporters, PPE, PE_PGRS, cell-wall synthesis proteins and hypothetical proteins.</p></sec><sec><title>Conclusion</title><p>This study has revealed the role of microsatellite indel mutations in imparting novel functions and a certain degree of plasticity to the mycobacterial genomes. There seems to be some correlation between microsatellite polymorphism and the variations in virulence, host-pathogen interactions mediated by surface antigen variations, and adaptation of the pathogens. Several of the polymorphic microsatellites reported in this study can be tested for their polymorphic nature by screening clinical isolates and various mycobacterial strains, for establishing correlations between microsatellite polymorphism and the phenotypic variations among these pathogens.</p></sec> | <contrib id="A1" equal-contrib="yes" contrib-type="author"><name><surname>Sreenu</surname><given-names>Vattipally B</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>sreenu@cdfd.org.in</email></contrib><contrib id="A2" equal-contrib="yes" contrib-type="author"><name><surname>Kumar</surname><given-names>Pankaj</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>pankaj@cdfd.org.in</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Nagaraju</surname><given-names>Javaregowda</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>jnagaraju@cdfd.org.in</email></contrib><contrib id="A4" corresp="yes" contrib-type="author"><name><surname>Nagarajaram</surname><given-names>Hampapathalu A</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>han@cdfd.org.in</email></contrib> | BMC Genomics | <sec><title>Background</title><p>Microsatellites, also known as simple sequence repeats, are the short nucleotide segments comprising tandem repeating motifs of length 1–6 bp [<xref ref-type="bibr" rid="B1">1</xref>]. They are present in all genomes known to date [<xref ref-type="bibr" rid="B2">2</xref>-<xref ref-type="bibr" rid="B4">4</xref>], and are known to be polymorphic [<xref ref-type="bibr" rid="B5">5</xref>]characterized by high rates of indels of repeat units [<xref ref-type="bibr" rid="B1">1</xref>]. Microsatellites provide a framework for crucial genetic rearrangements with their reversible frame-shift mutations that can confer a certain degree of selective advantage on pathogenic bacteria. Microsatellite mutations are known to affect expression levels [<xref ref-type="bibr" rid="B6">6</xref>], switching on/off of genes [<xref ref-type="bibr" rid="B6">6</xref>] and even alteration of gene functions [<xref ref-type="bibr" rid="B7">7</xref>]. The primary cause of microsatellite polymorphism is thought to be strand slippage during DNA replication [<xref ref-type="bibr" rid="B8">8</xref>]. Usually errors owing to strand slippage are repaired by a three-enzyme system comprising the enzymes mutL, mutS and mutH. However, some genomes like those of the mycobacterial species lack these enzymes [<xref ref-type="bibr" rid="B9">9</xref>]. Hence, such genomes serve as interesting systems to investigate the rates of mutations in microsatellites and the existence of regulatory mechanisms that govern microsatellite mutations. Furthermore, these genomes present challenging and exciting systems to understand the role of microsatellite mutations in conferring genome plasticity, and in aiding the pathogens in their adaptation and evolution.</p><p>Previous reports on genomic changes in <italic>M. tuberculosis</italic>, were mainly concerned with single nucleotide polymorphisms (SNPs) and large-sequence polymorphisms (LSPs) (>10 bp) [<xref ref-type="bibr" rid="B10">10</xref>]. While the involvement of SNPs in drug resistance has been shown [<xref ref-type="bibr" rid="B11">11</xref>], most of the LSPs are thought to be deleterious [<xref ref-type="bibr" rid="B12">12</xref>]. In the present study, we show for the first time that the coding regions of the three genomes of mycobacteria (<italic>M. tuberculosis </italic>H37Rv [<xref ref-type="bibr" rid="B13">13</xref>], <italic>M. tuberculosis </italic>CDC1551 [<xref ref-type="bibr" rid="B10">10</xref>] and <italic>M. bovis </italic>[<xref ref-type="bibr" rid="B14">14</xref>]) harbor a number of polymorphic microsatellite loci associated with remarkable changes in the coding regions.</p></sec><sec><title>Results and discussion</title><p>All the three mycobacterial genomes, <italic>M. tuberculosis </italic>H37Rv (MTH), <italic>M. tuberculosis </italic>CDC1551 (MTC) and <italic>M. bovis </italic>(MB) harbor about a million microsatellite tracts each, comprising of mono to hexa repeats (Sreenu, Pankaj Kumar, Nagaraju and Nagarajaram, manuscript communicated). Systematic comparison of all the equivalent microsatellites and the equivalent coding regions harboring them, in all the three genomes revealed several examples of microsatellites exhibiting length polymorphism characterized by indels of the repeat units. Frame-shifts in the coding regions owing to indels in microsatellites, were also observed. While some frame-shifts caused ORFs to split (fission) (see methods), others seemed to bring about fusion of two adjacent ORFs (with or without overlap) giving rise to a single ORF. Our study also revealed several ORFs eliminated as a result of premature termination by stop codons, and numerous other ORFs exhibiting length changes (Fig. <xref ref-type="fig" rid="F1">1</xref>). The complete list of polymorphic microsatellites along with the ORFs in which they are present is given in Table <xref ref-type="table" rid="T1">1</xref> (see <xref ref-type="supplementary-material" rid="S1">Additional File 1</xref> for details of the tracts, microsatellite polymorphism and outcomes). Illustrated below are some examples of microsatellites and their polymorphic effects on the coding regions.</p><p>In the MTH genome, two ORFs annotated as <italic>gmhA </italic>(Rv0113) and <italic>gmhB </italic>(Rv0114) have been identified as sedoheptulose-7-phosphate isomerase and D-α-β-D-heptose-7-biphosphate phosphatase, respectively (the TB structural genomics consortium [<xref ref-type="bibr" rid="B15">15</xref>]). These enzymes are known to be involved in the biosynthesis pathway of nucleotide activated glycerol-manno-heptose precursors of bacterial glycoproteins and cell surface polysaccharides [<xref ref-type="bibr" rid="B16">16</xref>]. Our study indicates that the ORF Rv0113 annotated as <italic>gmhA </italic>harbors the microsatellite (T)<sub>4 </sub>in MTH<sub>,</sub>while it is expanded to (T)<sub>5</sub>in the MTC genome. This expansion has resulted in a frame-shift owing to which the reading frame extends and fuses with that of the <italic>gmhB</italic>, thus giving rise to a fused ORF. Although it is hard to speculate the possible roles of the gmhA-gmhB fused protein in MTC, there exists a high probability of it forming a bi-functional protein with two domains.</p><p>Similarly, two adjacent ORFs viz., Rv0192A and Rv0192 in the MTH genome are observed to have fused into a single ORF (Mb0198) in the MB genome, owing to a frame-shift caused by the expansion of the microsatellite (G)<sub>4 </sub>to (G)<sub>5</sub>. Previous PhoA fusion screening studies have shown Rv0192A in MTH to act as a signal peptide [<xref ref-type="bibr" rid="B17">17</xref>], and in light of this it is reasonable to speculate the fused gene product in MB to be a secretory protein that may act as a surface antigen.</p><p>The ORF MT1966 in MTC encoding a functional isocitrate lyase [<xref ref-type="bibr" rid="B18">18</xref>], is observed to have split into two ORFs (Rv1915 and Rv1916) in MTH due to a single nucleotide deletion in the mononucleotide tract (T)<sub>5</sub>. The failure of these two ORFs to complement isocitrate lyase activity in MTH has been demonstrated [<xref ref-type="bibr" rid="B19">19</xref>]. Immunoblotting studies were unable to detect AceAa or AceAb products [<xref ref-type="bibr" rid="B18">18</xref>]. Subsequent studies by Betts and co-workers (2002) enabled detection of only the mRNA of AceAa, indicating the lack of expression of AceAb [<xref ref-type="bibr" rid="B20">20</xref>]. It is interesting to note that both the MTC and MTH genomes possess another copy of isocitrate lyase. This indicates the existence of two functional copies of the enzyme in MTC, and only a single copy in MTH. In MTC the activity of isocitrate lyase increases during the latent phase when the pathogen utilizes lipid as the energy source [<xref ref-type="bibr" rid="B21">21</xref>]. Redundancy in isocitrate lyase in MTC can therefore be beneficial to the pathogen, providing a greater chance of its survival in the host cell debris where lipid is used as a carbon source. However, in MTH which is cultured under laboratory conditions with no dependence on lipids as the carbon source, the duplication of the isocitrate lyase enzyme is not required. Therefore, the removal of one copy of the enzyme in MTH may not pose as a constraint for the growth of the pathogen.</p><p>On comparison, the highest number (18 ORFs) of split events is observed in the MB genome (Table <xref ref-type="table" rid="T1">1</xref>). The expression of both parts of split genes in the MB genome, imply a favorable situation for versatile protein-protein interactions. However, it is to be noted in the cases of split ORF, the expression of the second part of the ORF is entirely dependent on the availability of regulatory signals (Shine-Dalgarno sequence) for that ORF. In the absence of a regulatory mechanism, the second part of the ORF is unexpressed. As given in Table <xref ref-type="table" rid="T1">1</xref>, section III, the second part of all the four examples, has been annotated as psuedogene because of the absence of the Shine-Dalgarno sequence. If both the parts of the split ORFs are expressing the split subunits can act together [<xref ref-type="bibr" rid="B22">22</xref>,<xref ref-type="bibr" rid="B23">23</xref>] or in isolation resulting in different protein-protein interactions, that can be instrumental in the creation of alternate/new pathways, which in turn may eventually render greater adaptation mechanisms to the bacteria. This may well be the one of the underlying reasons for MB to have a wider host range as compared to <italic>M. tuberculosis</italic>.</p><p>The split ORFs encode membrane proteins, transporters, PE_PGRS, cell-wall synthesis proteins and hypothetical proteins. The membrane proteins are known to play an important role in host-pathogen interactions [<xref ref-type="bibr" rid="B24">24</xref>]. The majority of bacteria are thought to modify their membrane protein structures in order to escape the host immune defense system and promote colonization at various places within the host [<xref ref-type="bibr" rid="B6">6</xref>,<xref ref-type="bibr" rid="B24">24</xref>]. The PE-PGRS proteins are specific to mycobacteria and are speculated to function as surface antigens [<xref ref-type="bibr" rid="B25">25</xref>,<xref ref-type="bibr" rid="B26">26</xref>]. Truncation with respect to the second part can potentially give rise to an antigenic variant.</p><p>MTC as compared to the other genomes exhibits a greater number of cases of premature terminations (10 ORFs) (Table <xref ref-type="table" rid="T1">1</xref>), confined to the PE_PGRS, <italic>umaA1</italic>, <italic>pks5 </italic>and some hypothetical proteins. Of these, the ORF <italic>umaA1 </italic>codes for a mycolic acid methyl transferase that modifies the lipids of the mycobacterial cell wall [<xref ref-type="bibr" rid="B27">27</xref>]. The <italic>umaA1 </italic>deletion mutant of MTH is observed to be more virulent than the wild-type, in the severe combined immune deficiency (SCID) mouse model [<xref ref-type="bibr" rid="B28">28</xref>]. However, it is difficult to categorically stress the importance of <italic>umaA1 </italic>in the virulence of the pathogen. This is because MTC has been shown to be less virulent in the immunocompetent mice as compared to other clinical isolates [<xref ref-type="bibr" rid="B29">29</xref>]. Study on an umaA1 deletion mutant of MTH in immunocompetent mice would provide clues to the role of umaA1 in virulence. In addition, it is equally possible for the other prematurely terminated ORFs to also be responsible for the less virulent nature of MTC. However, such correlations require further studies.</p><p>We also observe an appreciable number of ORFs (43 examples) in all the three genomes exhibiting length variations due to indels of repeat units in microsatellites. Many proteins in this category have been annotated as hypothetical proteins, PPE and mammalian cell entry (mce) family virulence proteins. While the length variation in some ORFs produce no effect on the function of the translated protein with the functional domains being well conserved; in others, drastic changes are observed. For example, Rv2732c in MTH as well as Mb2791c in MB code for a membrane anchoring protein of length 204aa. The equivalent ORF MT2802.1 in MTC is a shorter ORF encoding only 180aa, owing to a frame-shift caused by a single G insertion in the microsatellite tract (G)<sub>2</sub>. <italic>In silico </italic>analysis of these proteins, reveals a greater probability (0.959) of the N-terminal deleted short protein in MTC to act as a signal peptide and secrete outside, than its longer counterparts in MB and MTH that possess negligible propensities of being signal peptides and therefore for external secretion.</p><p>Although the primary focus of this communication is on microsatellite polymorphism in the coding regions, we have also examined the upstream promoter regions of the ORFs and obtained some ORFs harboring polymorphic microsatellites (data not shown). It should be noted that genes are located very close to each other in a prokaryotic genome; at times without any long intergenic region between two adjacent genes. It is probable that the coding sequence of a gene may act as a regulatory sequence for its neighboring genes. In addition to bringing about changes in the coding regions, the observed microsatellite variations may also influence regulation of regions downstream of coding sequences.</p><p>We have referred the Stanford microarray database [<xref ref-type="bibr" rid="B30">30</xref>], Tuberculist [<xref ref-type="bibr" rid="B31">31</xref>], ArrayExpress [<xref ref-type="bibr" rid="B32">32</xref>] and available literature on microarray analysis of mycobacterium [<xref ref-type="bibr" rid="B20">20</xref>,<xref ref-type="bibr" rid="B33">33</xref>-<xref ref-type="bibr" rid="B37">37</xref>] for the expression profiles of all ORFs of MTH listed in Table <xref ref-type="table" rid="T1">1</xref>. Almost 85% of the ORFs (indicated by * in the table) display high expression profiles, including those that have undergone fission. However, further studies are necessary to verify and complement the function of these split gene products with their cognate wild-type/unsplit proteins.</p><p>It is evident from Table <xref ref-type="table" rid="T1">1</xref> that microsatellites with as few as two repeats display polymorphism (i.e., indels of their repeat units). This appears to contradict earlier observations of the requirement of a microsatellite length threshold for repeat expansions or contractions due to strand slippage [<xref ref-type="bibr" rid="B38">38</xref>,<xref ref-type="bibr" rid="B39">39</xref>]. Our study therefore indicates the non-dependence of strand slippage on microsatellite tract lengths. However, one should bear in mind the possibility of random mutational events leading to the observed length variation in microsatellites. For example, the genomes of <italic>M. canetti </italic>and <italic>M. tuberculosis </italic>contain the (GGGCCGC)<sub>2 </sub>tract in the ORF that encodes for <italic>pks15/1</italic>. However, the equivalent regions in the MTC and MTH genomes have a 7 bp deletion of (GGGCCGC) and in the MB genome a 6 bp deletion of (GGCCGC) [<xref ref-type="bibr" rid="B40">40</xref>]. Although the deletion events are independent, the resultant sequences when compared give an impression of the G tract expansion. Alternatively, it can be argued that all three genomes MB, MTC and MTH may have possessed an initial 7 bp deletion (GGGCCGC) similar to <italic>M. canetti</italic>, giving rise to the microsatellite tract (G)<sub>5 </sub>that may have subsequently expanded to (G)<sub>6 </sub>in MB. It is still unclear as to which of the models depict the correct picture of events for the observed microsatellite polymorphism. This is largely because of the unavailability of detailed evolutionary information of the mycobacterial pathogen. Although <italic>M. canetti </italic>is believed to be the root from which the other mycobacterial strains evolved, a clear understanding of the evolutionary relationship between <italic>M. tuberculosis </italic>and <italic>M. bovis </italic>is absent [<xref ref-type="bibr" rid="B41">41</xref>-<xref ref-type="bibr" rid="B44">44</xref>]. Owing to this, it is difficult to put forward precisely the path of microsatellite evolution, although several possibilities can be suggested.</p><p>The rate at which microsatellites mutate is much higher than the single-base substitutions [<xref ref-type="bibr" rid="B45">45</xref>,<xref ref-type="bibr" rid="B46">46</xref>], therefore greater variations are expected in the polymorphic loci than other regions of the genomes. Though mycobacterial genomes are enriched with microsatellite tracts (Sreenu, Pankaj, Nagaraju and Nagarajaram, manuscript communicated), surprisingly there is yet no report available on the microsatellite mediated phase variation in these bacteria. The majority of microsatellite mediated phase variations reported in pathogenic bacteria are changes in the pili [<xref ref-type="bibr" rid="B47">47</xref>,<xref ref-type="bibr" rid="B48">48</xref>], capsule [<xref ref-type="bibr" rid="B49">49</xref>,<xref ref-type="bibr" rid="B50">50</xref>] and flagella [<xref ref-type="bibr" rid="B51">51</xref>,<xref ref-type="bibr" rid="B52">52</xref>] and the mycobacteria do not possess any of these structures. According to Hallet, phase variation is "an adaptive process through which bacteria undergo frequent and reversible phenotypic changes resulting in genetic alterations in their genomes" [<xref ref-type="bibr" rid="B53">53</xref>]. In light of this point it is highly interesting that this work presents several polymorphic microsatellite loci that seem to have been evolutionarily 'selected' and are involved in bringing about phenotypic alterations in the coding regions namely, antigenic variation, virulence and modified host-pathogen interactions for presumably better adaptation of the pathogen.</p><p>It is tempting to speculate that some of the polymorphic microsatellites discovered in this study are those that have undergone mutations at some point of time during microbe evolution, perhaps during speciation, and thereafter remained frozen as the 'molecular fossils'. If this model is correct, then such tracts can be used as markers for species/strain identification. In any case all the loci form a good starting set to screen several isolates and strains. This would enable to study correlation between microsatellite polymorphism and the observed phenotypic variations among different isolates and strains.</p><p>An important point to be noted in connection with microsatellite polymorphism in the mycobacterial genomes is the absence of the post replicative DNA mismatch repair system mediated by <italic>mutS, mutL </italic>and <italic>mutH </italic>genes [<xref ref-type="bibr" rid="B9">9</xref>]. Impairment of these enzymes destabilizes mono, di and trinucleotide repeats [<xref ref-type="bibr" rid="B54">54</xref>]. This probably accounts for the prevalence of mono and dinucleotide microsatellite variations in mycobacterial genomes. Moreover, the absence of these enzymes appears advantageous to these pathogens, resulting in the generation of polymorphic microsatellites, thereby imparting a certain degree of plasticity to the genomes. However, the total number of microsatellites that exhibit polymorphism, and their significance in the context of pathogen adaptability, virulence and survival remains to be tested.</p></sec><sec><title>Conclusion</title><p>The coding regions in the mycobacterial genomes, viz. <italic>M. tuberculosis </italic>H37Rv, <italic>M. tuberculosis </italic>CDC1551 and <italic>M. bovis</italic>, harbor a number of polymorphic microsatellites. The observed indel mutations in microsatellites have brought out some interesting changes in the coding regions indicative of gene fusion/fission, loss, and functional variation. From this study, it can be concluded that microsatellites form an important set of genomic elements, mutations of which are beneficial to the pathogens.</p></sec><sec sec-type="methods"><title>Methods</title><p>Complete genome sequences of <italic>M. tuberculosis </italic>(H37Rv and CDC1551) and <italic>M. bovis </italic>were downloaded from the NCBI ftp site [<xref ref-type="bibr" rid="B55">55</xref>]. Functional annotations of the coding regions were referred to the Tuberculist website [<xref ref-type="bibr" rid="B31">31</xref>] and the TB structural genomics consortium site [<xref ref-type="bibr" rid="B15">15</xref>]. The various microsatellites in the three genomes were identified using SSRF [<xref ref-type="bibr" rid="B56">56</xref>]. SSRF scans a given nucleotide sequence and extracts all microsatellite tracts of motif length 1–6 bp. The extracted information includes genomic location of the tracts, repeating motifs, repeat numbers and regions (coding or non-coding or partial) in which the tracts are present. The program utilizes the GenBank annotation file "xxx.ffn" (where xxx = genome name) that has exon boundary information, using which the location of microsatellites relative to the protein coding regions is subsequently recorded. In addition the internal motif redundancy is taken care of; where a sequence of the type (AAAAGCAAAAGCAAAAGC) is represented as (AAAAGC)<sub>3 </sub>with the internal "A"s (<underline>AAAA</underline>GC) not considered as a separate (A)<sub>4 </sub>tract.</p><p>The ORFs harboring microsatellites of one genome were used as queries to search against the other two complete mycobacterial genome sequences using the BLASTN program (version 2.2.6) [<xref ref-type="bibr" rid="B57">57</xref>] without the repeat masking filter. The alignment hits with queried sequences comprising only indels in the microsatellites were selected for further analysis. The Tuberculist database (for H37Rv and <italic>M. bovis</italic>) and the NCBI (for CDC1551) were checked and confirmed to ensure that the indels in microsatellites especially those of the mononucleotide tracts were indeed authentic mutations and not the results of sequencing errors (however one can not rule out some remote possibility of sequencing artifact). Subsequently, the ORFs and their equivalent sequences were realigned using CLUSTALW [<xref ref-type="bibr" rid="B58">58</xref>] to reconfirm the alignment as well as the INDELS in the microsatellites. As the phylogenetic relation of these genomes is still ambiguous, a consensus of the three genomes for microsatellite categorization into premature terminations, gene fusion/fission and ORF premature termination was used.</p></sec><sec><title>Authors' contributions</title><p>VBS: Computational analysis of microsatellite polymorphisms across the mycobacterial genomes and initial drafting of the manuscript</p><p>PK: Comparative analysis of functions of coding regions harbouring polymorphic microsatellites across the mycobacterial genomes</p><p>JN: Provided suggestions during the initial stages of the manuscript preparation</p><p>HAN: Project leader, project guide and in-charge of final manuscript corrections and submission</p></sec><sec sec-type="supplementary-material"><title>Supplementary Material</title><supplementary-material content-type="local-data" id="S1"><caption><title>Additional File 1</title><p>List of ORFs from <italic>M. tuberculosis </italic>H37Rv (MTH), <italic>M. tuberculosis </italic>CDC1551 (MTC) and <italic>M. bovis </italic>(MB) harboring polymorphic microsatellite tracts. The complete list of the polymorphic microsatellites from the mycobacterial genomes, <italic>M.tuberculosis </italic>H37Rv, <italic>M.tuberculosis </italic>CDC1551 and <italic>M.bovis</italic>, along with the alignments of microsatellite tracts and flanking sequences. This list provides locations of microsatellite in the genomes, microsatellite variation, details of microsatellite position in protein with respect to amino acid sequence, local sequence of the of the microsatellite tract, start and end positions of the ORF, which contains the microsatellite, coding strand information (same strand:'+', template strand:'-'), GenBank ID of a protein, function of protein and protein length</p></caption><media xlink:href="1471-2164-7-78-S1.doc" mimetype="application" mime-subtype="msword"><caption><p>Click here for file</p></caption></media></supplementary-material></sec> |
Production and utilization of a high-density oligonucleotide microarray in channel catfish, <italic>Ictalurus punctatus</italic> | <sec><title>Background</title><p>Functional analysis of the catfish genome will be useful for the identification of genes controlling traits of economic importance, especially innate disease resistance. However, this species lacks a platform for global gene expression profiling, so we designed a first generation high-density oligonucleotide microarray platform based on channel catfish EST sequences. This platform was used to profile gene expression in catfish spleens 2 h, 4 h, 8 h and 24 h after injection of lipopolysaccharide (LPS).</p></sec><sec><title>Results</title><p>In the spleen samples, 138 genes were significantly induced or repressed greater than 2-fold by LPS treatment. Real-time RT-PCR was used to verify the microarray results for nine selected genes representing different expression levels. The results from real-time RT-PCR were positively correlated (R<sup>2 </sup>= 0.87) with the results from the microarray.</p></sec><sec><title>Conclusion</title><p>The first generation channel catfish microarray provided several candidate genes useful for further evaluation of immune response mechanisms in this species. This research will help us to better understand recognition of LPS by host cells and the LPS-signalling pathway in fish.</p></sec> | <contrib id="A1" contrib-type="author"><name><surname>Li</surname><given-names>Robert W</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I2">2</xref><email>rli@anri.barc.usda.gov</email></contrib><contrib id="A2" corresp="yes" contrib-type="author"><name><surname>Waldbieser</surname><given-names>Geoffrey C</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>gwaldbieser@msa-stoneville.ars.usda.gov</email></contrib> | BMC Genomics | <sec><title>Background</title><p>The channel catfish, <italic>Ictalurus punctatus</italic>, is native to North America and is from the order Siluriformes (superorder Ostariophysi) which is considered more primitive among teleosts [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B2">2</xref>]. Commercial production of catfish as dietary protein is the leading industry of North American aquaculture, with more than 600 million pounds of catfish produced annually in the United States [<xref ref-type="bibr" rid="B3">3</xref>]. One of the most significant factors limiting production is loss to disease, and one of our goals is to improve non-specific immunity through selective breeding. The channel catfish immune system is the best characterized for any fish species, and it is the only fish species where clonal functionally distinct lymphocyte lines can be readily established [<xref ref-type="bibr" rid="B4">4</xref>]. However, there is a need to better understand the physiological and immunological pathways controlling host-pathogen interactions <italic>in vivo</italic>.</p><p>Improvement of disease resistance in catfish populations depends on an understanding of the genetic control of immune-related pathways, response to pathogens, and correlations with other production traits. Recent developments in genomic technology, particularly high-throughput cDNA sequencing and development of expressed sequence microarrays, have made possible the profiling of global gene expression in experimental fish tissues.</p><p>Microarray experiments have been utilized to determine regulation of gene expression during developmental and adaptive processes in several fish species. Using this technology, groups of coordinately expressed genes were identified in zebrafish embryos during development [<xref ref-type="bibr" rid="B5">5</xref>,<xref ref-type="bibr" rid="B6">6</xref>]. Microarrays were used to identify changes in gene expression patterns in zebrafish and goby exposed to differing levels of available oxygen [<xref ref-type="bibr" rid="B7">7</xref>,<xref ref-type="bibr" rid="B8">8</xref>], in common carp exposed to cold temperature [<xref ref-type="bibr" rid="B9">9</xref>], and in rainbow trout exposed to handling stress [<xref ref-type="bibr" rid="B10">10</xref>].</p><p>Recently, expression microarrays have been used to identify genes involved in host-pathogen relationships. Experimental infection of cells <italic>in vitro </italic>revealed differential gene expression in flounder kidney cells [<xref ref-type="bibr" rid="B11">11</xref>], and in rainbow trout monocytes and macrophages [<xref ref-type="bibr" rid="B12">12</xref>]. Transcriptional analysis was performed on kidney cells after experimental <italic>in vivo </italic>injection of flounder with a DNA vector expressing the viral hemorrhagic septicemia virus G protein [<xref ref-type="bibr" rid="B13">13</xref>], in Atlantic salmon macrophages and anterior kidney tissues after injection of fish with <italic>P. salmonis </italic>[<xref ref-type="bibr" rid="B14">14</xref>], on Atlantic salmon liver, spleen, and anterior kidney tissues after cohabitation with <italic>A. salmonicida</italic>-infected fish [<xref ref-type="bibr" rid="B15">15</xref>], and from whole zebrafish after infection with <italic>M. marinum </italic>[<xref ref-type="bibr" rid="B16">16</xref>]. By measuring changes in gene expression after pathogen challenge, researchers may identify gene expression fingerprints that provide clues to molecular pathways involved in pathogen neutralization and/or removal, identify candidate genes controlling pathogen-specific immunity, and identify heritable differences in gene expression levels that correlate with disease resistance/susceptibility. This data would help in the formulation of a selection index to identify broodstock with superior genetic potential for resistance to disease.</p><p>While microarray-based transcriptional analysis is a useful tool for functional genomics in fish, to date there has been no microarray platform available for catfish species. Thus, we developed a catfish microarray utilizing existing data from channel catfish expressed sequences obtained in cDNA cloning experiments from several tissues (reviewed in [<xref ref-type="bibr" rid="B17">17</xref>][<xref ref-type="bibr" rid="B18">18</xref>]). These sequences have been clustered and annotated in the Catfish Gene Index [<xref ref-type="bibr" rid="B19">19</xref>]. The catfish microarray was used to measure differential gene expression in the spleen of lipopolysaccharide-injected and sham-injected catfish.</p></sec><sec><title>Results</title><p>A high density oligonucleotide microarray was produced that contained 18,989 catfish expressed sequences, each represented by ten perfectly matched and ten mismatched 24-mer oligonucleotides (GEO Database Accession GSE3261). This microarray contained 8057 sequences that matched an annotated sequence in GenBank. Discounting 875 that matched transposons or sequences that were labelled as unknown, unnamed, or hypothetical, there were 7182 annotated sequences (38%) on the microarray. This microarray was used to demonstrate differences in global gene expression in the spleen of LPS-treated catfish. Using Significance Analysis of Microarrays (SAM), we identified 138 sequences that were differentially expressed after LPS treatment [see <xref ref-type="supplementary-material" rid="S1">Additional file 1</xref> and <xref ref-type="supplementary-material" rid="S2">Additional file 2</xref>].</p><p>There were 64 genes up-regulated by LPS exposure of which only 26 were annotated based on sequence similarity [see <xref ref-type="supplementary-material" rid="S1">Additional file 1</xref>]. Among these were cytokines and chemokines, such as IL-1β, CCL4, a small CXC/IL 8-like chemokine, and chondromodulin II. Transcriptional factors such as NF-κB p100 subunit and NF-κB inhibitor alpha-like proteins A and B, interferon regulatory factor I, and AP-1 were also upregulated at least 2-fold. Expression of Toll-like receptor 5 was upregulated at 2–4 h post-exposure. Within the experimental timeframe, almost all induced genes were induced by 4 h post-exposure.</p><p>There were 74 genes that were significantly down-regulated in response to LPS, of which only 29 were annotated based on sequence similarity [see <xref ref-type="supplementary-material" rid="S2">Additional file 2</xref>]. Among these genes were the immunoglobulin light and heavy chains, MHC class II antigens, invariant chain-like protein 2, an NK lysin-like protein, complement C3, hemoglobin alpha, and a CXC chemokine receptor.</p><p>In the present research, we selected nine expressed sequences representing varying levels of expression to verify the microarray results using real-time RT-PCR [see <xref ref-type="supplementary-material" rid="S3">Additional file 3</xref>]. Linear regression of log transformed expression data demonstrated a strong positive correlation (R<sup>2 </sup>= 0.87) between the two technologies (Fig. <xref ref-type="fig" rid="F1">1</xref>). The microarray results for TNFα failed to pass our quality control criteria, thus no expression ratios were reported (detailed below).</p><fig position="float" id="F1"><label>Figure 1</label><caption><p><bold>Correlation between levels of gene expression measured by microarray and real-time RT-PCR</bold>. The correlation (R<sup>2 </sup>= 0.87) was calculated using log-transformed values of the fold change obtained for the nine selected genes. Four fish were used per treatment/time combination.</p></caption><graphic xlink:href="1471-2164-7-134-1"/></fig></sec><sec><title>Discussion</title><sec><title>Microarray performance</title><p>In our first attempt to monitor the global gene expression profile after LPS stimulation in channel catfish spleen, we have constructed a high-density oligonucleotide microarray containing almost 19,000 catfish unique sequences. In the present research, we used real-time RT-PCR to verify the microarray results. The analysis revealed a strong positive correlation between the microarray and real-time RT-PCR results. The TNFα gene was selected based on its importance in the LPS pathway and also to represent genes that did not make the final "hit" list. The TNFα microarray results failed to pass our quality control criteria, so no ratios were reported. After examination of the oligo sequences for TNFα on the microarray, we found only one of the ten probe sets demonstrated a significant difference in signal intensity between the perfectly matched and mismatched oligos. Five of the TNFα probe sets contained higher signal intensities in the mismatched than in the perfectly matched oligo. The automated oligo design software extracted all ten oligo sets from the most 5' end of the 255 bp TNFα cDNA sequence, so the discrepancy between real-time RT-PCR and microarray results for TNFα was likely due to undesirable oligo design. In fact, the one "good" oligo pair would have identified up-regulation of TNFα by LPS. This observation highlights the importance of high quality oligo design software, and that the ability to identify and eventually remove problematic oligo probes is critical to the collection of robust microarray data and minimization of discordant gene expression data from different technologies.</p></sec><sec><title>Differential expression of catfish genes</title><p>The evolutionary divergence between mammalian and fish species is considerable, and we do not know if LPS can induce the IKK-NF-κB pathway and MAPK pathways (the ERK, JNK, and p38 pathways) in fish as in mammals. Although LPS-responsive cis-elements have been identified in the 5'-flanking region of many mammalian genes [<xref ref-type="bibr" rid="B20">20</xref>], none have been reported in catfish to date. The present research demonstrated increased expression of the p100 subunit of NF-κB at 4 h post-LPS exposure, but a sequence matching NF-κB inhibitor alpha-like proteins A and B were also significantly induced at 2 h post-LPS exposure. NF-κB inhibitor gene expression was also increased in mycobacterium-infected zebrafish [<xref ref-type="bibr" rid="B16">16</xref>] and bacterially infected macrophage cells from Atlantic salmon and rainbow trout [<xref ref-type="bibr" rid="B12">12</xref>,<xref ref-type="bibr" rid="B14">14</xref>]. There are instances in biological systems that a compound can up-regulate both an enzyme and its inhibitor. For example, the short-chained fatty acid butyrate up-regulated both IGF2 and IGF binding proteins IGFBP2 and IGFBP3 in bovine kidney epithelial cells (R.L., unpublished). An alternative explanation to the present observation is that the inhibitor encoded a protein with a function different from that which was annotated due to sequence similarity.</p><p>Toll-like receptor 5 demonstrated significant levels of induction at 4 h post-LPS treatment. In mammals, TLR4 is the extracellular LPS receptor whereas TLR5 recognizes flagellin [<xref ref-type="bibr" rid="B21">21</xref>,<xref ref-type="bibr" rid="B22">22</xref>]. The up-regulation of TLR5 by Gram-negative, virulent <italic>Edwardsiella ictaluri </italic>has been independently observed in channel catfish [[<xref ref-type="bibr" rid="B23">23</xref>]; Li, unpublished]. This gene was also shown to increase in the liver of Atlantic salmon exposed to the bacterium <italic>A. salmonicida </italic>[<xref ref-type="bibr" rid="B15">15</xref>], and in zebrafish exposed to the mycobacterium <italic>M. marinum </italic>[<xref ref-type="bibr" rid="B24">24</xref>]. Toll-like receptor 3 was also included on the catfish microarray, but expression was not significantly altered due to LPS exposure. The present data add evidence to the hypothesis that, in fish, microbial products may induce other TLRs than those specific for their own recognition [<xref ref-type="bibr" rid="B24">24</xref>].</p><p>Interferon regulatory factor 1 (IRF-1) is one of two DNA-binding transcription factors mediating Type 1 interferon gene expression. In rainbow trout gonadal cells, IRF-1 is expressed constitutively and up-regulated by poly I:C but not by LPS [<xref ref-type="bibr" rid="B25">25</xref>]. However, microarray experiments using Japanese flounder have demonstrated stimulation of IRF-1 expression in cultured kidney cells after exposure to LPS [<xref ref-type="bibr" rid="B11">11</xref>] and in kidney tissue after exposure of fish to the viral hemorrhagic septicemia virus G protein [<xref ref-type="bibr" rid="B13">13</xref>].</p><p>Apolipoprotein E (ApoE) has been suggested to have a regulatory role in maintaining a critical balance of various proinflammatory cytokines in mice [<xref ref-type="bibr" rid="B26">26</xref>] in addition to its key role in lipid metabolism. Expression of ApoE decreases in CaCo-2 cells 24 h after LPS exposure [<xref ref-type="bibr" rid="B27">27</xref>], and this was consistent with our observations.</p><p>The chemokine receptor CXCR4 gene selected for analysis by quantitative real-time PCR (IpCG06125) was first up-regulated by LPS at 2 h, rapidly down-regulated at 4 h, and remained down-regulated at 8 h and 24 h [see <xref ref-type="supplementary-material" rid="S3">Additional file 3</xref>]. A CD14-independent model for LPS recognition has been proposed that includes a LPS activation cluster complex containing CXCR4, heat shock proteins 70 and 90, and growth differentiation factor 5 [<xref ref-type="bibr" rid="B28">28</xref>,<xref ref-type="bibr" rid="B29">29</xref>]. In addition, CXCR4 is a functional co-receptor for HIV-1 infection of human macrophages. LPS down-regulates the expression of CXCR4 on monocyte-derived macrophages [<xref ref-type="bibr" rid="B30">30</xref>], and the present research is consistent with this observation.</p><p>While the majority of the catfish sequences could not be annotated due to sequence identify, several catfish genes could be compared with other fish experiments utilizing microarrays to measure gene expression after experimental infection. The interleukin 1β gene demonstrated the greatest increase in LPS-induced expression in catfish spleen, and this gene was highly induced in flounder kidney cells exposed to LPS [<xref ref-type="bibr" rid="B11">11</xref>]. Expression of the immune responsive protein 1 gene increased in catfish spleen, and bacterially infected salmon kidneys and macrophages [<xref ref-type="bibr" rid="B14">14</xref>]. Catfish MHC class II gene expression decreased after LPS exposure, and this was also observed in mycobacterium-infected zebrafish [<xref ref-type="bibr" rid="B16">16</xref>] and salmon macrophage cells exposed to LPS [<xref ref-type="bibr" rid="B12">12</xref>], but LPS-treated flounder kidney cells demonstrated no change in MHC class II gene expression [<xref ref-type="bibr" rid="B11">11</xref>]. Immunoglobulin gene expression decreased in catfish spleen after LPS exposure and in bacterially infected Atlantic salmon macrophage cells [<xref ref-type="bibr" rid="B14">14</xref>], but it increased in zebrafish after mycobacterial infection [<xref ref-type="bibr" rid="B16">16</xref>] and in Atlantic salmon liver and kidney after bacterial infection [<xref ref-type="bibr" rid="B14">14</xref>,<xref ref-type="bibr" rid="B15">15</xref>]. Likewise, complement C3 gene expression was decreased in catfish spleen and salmon macrophages [<xref ref-type="bibr" rid="B14">14</xref>], but increased in kidneys from infected salmon [<xref ref-type="bibr" rid="B14">14</xref>] and in infected zebrafish [<xref ref-type="bibr" rid="B16">16</xref>]. While the present experiment demonstrated similarities and differences in species-specific gene expression after bacterial or LPS exposure, the lack of fully annotated sequences on the catfish microarray hindered a more comprehensive comparative analysis. Evaluation of immune responses across species would also enhanced by a reference set of microarray probes on each species-specific microarray.</p><p>Overall, there was no correlation between genes that were more abundant in catfish spleen (measured as number of cDNA transcripts in [<xref ref-type="bibr" rid="B31">31</xref>]) and genes that demonstrated significant changes in level of expression after LPS stimulation in the present research. While a 2-fold difference in expression is an arbitrary value, there are, undoubtedly, significant physiological and immunological consequences to differences in gene expression less than 2-fold. This differential reflects the inability of current microarray technology to significantly resolve smaller differences in levels of gene expression. The present research exposed difficulties inherent in performing microarray analyses from tissues <italic>in vivo</italic>. Differences observed between individuals within each treatment, due to asynchronized sampling and/or genetic variation, led to averaged effects which could have masked significant responses in some individuals. Nonetheless, microarrays are powerful tools that permit the simultaneous screening of thousands of genes to provide candidates for further study. While the current catfish microarray is limited by the large number of "unknown" candidate sequences, both functional experimentation and comparative genomic analyses should help improve and identification of the large number of catfish ESTs currently without annotation.</p><p>From the patterns of gene expression in this experiment, we hypothesize the observed gene expression profiles in LPS-stimulated catfish spleen result mainly from macrophage activation by LPS. However, contributions from other cell types in spleen cannot be discounted. Since macrophages play a key role in a variety of inflammatory diseases and in host defense such as pathogen phagocytosis, antigen presentation to immunocytes, and production of numerous inflammatory mediators, our future work should focus on homogeneous cell populations derived either from cultured cell lines or from macrophages isolated directly from the tissues with techniques such as laser-capture microdissection.</p></sec></sec><sec><title>Conclusion</title><p>The oligonucleotide microarray produced in the present research contained genes expressed in seventeen juvenile tissues and all neonatal tissues and provided a platform for global analysis of catfish gene expression. Levels of expression measured on the microarray were significantly correlated with levels measured by quantitative RT-PCR. Only 38% of the catfish sequences could be presently annotated, but <italic>de novo </italic>synthesis of oligonucleotides on the platform provides flexibility for adding annotated catfish sequences to the microarray as they become available.</p><p>The present experiment focused on the identification of differential gene expression in spleen in response to LPS injection. The innate immune response to Gram-negative bacteria is influenced by the ability of the host to respond to the LPS component, the major virulence factor of these bacteria. The present research identified 138 differentially expressed sequences that can provide clues as to the response of the catfish immune system to LPS from gram-negative bacteria, and contribute to discerning the NF-κB pathway in teleosts. Dissecting the LPS signal pathway is important in expanding our knowledge of molecular interactions between invading bacteria and host cells. Understanding the difference in the LPS pathway between fish and mammalian species will also facilitate understanding of the evolution of innate immune systems. Knowledge of the catfish immune system will help researchers identify key pathways in pathogen-induced diseases in catfish.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>Experimental fish and LPS treatment</title><p>Channel catfish (USDA103 strain, average weight ~200 g) were maintained in a single 160-liter tank at the Catfish Genetics Research Unit. Fish were fed a commercial diet to apparent satiation daily until 24 h prior to treatment with lipopolysaccharide (LPS). Thirty-two fish were anesthetized with tricaine methanesulfonate (MS222, 100 mg/L) and randomly divided into two groups, control and LPS. There were four individuals in each of time-matched control and LPS treatment time points. Lipopolysaccharide derived from E. coli 0127:B8 (Sigma-Aldrich, St. Louis, MO) was resuspended at 1.0 mg/ml in phosphate buffered saline (PBS, pH 7.0). Fish in the LPS group each received a 0.5 ml intraperitoneal (IP) injection of LPS at a dosage of 2.5 mg/kg body weight. Fish in the control group each received a 0.5 ml IP injection of PBS. Fish were then killed by anesthesia overdose (300 mg/L) at 2, 4, 8, or 24 h post-injection and the spleen was dissected and placed in 1.0 ml Trizol (Invitrogen Corporation, Carlsbad, CA). Samples were snap-frozen in liquid nitrogen and stored at -80°C until RNA isolation. Due to expense, only two fish per treatment/time combination were used for microarray analysis, but all four fish per treatment/time combination were used for real-time RT-PCR analysis. All experimental procedures were conducted in accordance with the principles and procedures approved by the Institutional Animal Care and Use Committee, USDA-ARS Catfish Genetics Research Unit.</p></sec><sec><title>Total RNA purification and generation of biotin-labelled complementary RNA</title><p>Total RNA was extracted using manufacturer's recommendations (Invitrogen), treated with 4–10 units DNase I (Ambion Inc., Austin, TX) per 100 μg total RNA at 37°C for 30 min, then further purified using an RNeasy Mini kit (Qiagen Inc., Valencia, CA). Equal masses of total RNA were pooled from the eight individuals in the control group (two fish in each of four time-matched controls). This pool of RNA was labelled as described below and served as a control in the microarray hybridization. The total RNA from the eight LPS-stimulated fish (two in each of four time points) was not pooled.</p><p>Biotin-labelled complementary RNA (cRNA) was generated using a modified procedure for the Superscript Choice System (Invitrogen) for double-stranded (ds) cDNA synthesis followed by <italic>in vitro </italic>transcription. First strand cDNA was synthesized by SuperScript II reverse transcriptase from 2.5 μg of total RNA using 100 pmoles of T7 promoter Oligo dT primer. After ds cDNA synthesis, the cDNA was purified with DNA Clean & Concentrator-5 (Zymo Research, Orange, CA), eluted with 8–16 μl of ddH<sub>2</sub>O, and concentrated to 3 μl by vacuum centrifugation. Complementary RNA was then synthesized with a MEGAscript <italic>in vitro </italic>Transcription kit (Ambion). The 23.0 μl <italic>in vitro </italic>transcription reaction included the entire 3.0 μl of ds cDNA, 2.3 μl of 10× Ambion reaction buffer, 2.3 μl of 10× Ambion T7 enzyme mix, and 15.4 μl of NTP labelling mix (7.5 mM ATP, 7.5 mM GTP, 5.625 mM UTP, 5.625 mM CTP and 1.875 mM biotin-16-UTP and 1.875 mM biotin-11 CTP). The <italic>in vitro </italic>transcription reaction was incubated at 37°C for 16 hours in a thermal cycler. Transcribed cRNA was then purified with an RNeasy Mini kit. Generally, 40–60 μg of cRNA was obtained from 2.5 μg original RNA. The sample integrity of the cRNA was verified with an Agilent 2100 Bioanalyzer (Agilent Technologies, Inc., Palo Alto, CA). The biotinylated cRNA products were then fragmented to 50–200 bp prior to hybridization by heating cRNA in 1× fragmentation buffer (40 mM Tris-acetate, pH.8.1, 100 mM potassium acetate, 30 mM magnesium acetate) at 95°C for 35 min.</p></sec><sec><title>Catfish oligonucleotide microarray production, hybridization, and image acquisition</title><p>Published sequences from cDNA libraries were obtained from the TIGR catfish gene index release 5.0 [<xref ref-type="bibr" rid="B19">19</xref>]. These ESTs were derived from brain, gill, intestine, liver, kidney (anterior and posterior), macrophage, muscle, olfactory tissue, ovary, peripheral blood leukocytes, pituitary, skin, spleen, stomach, taste tissue, and testis. An additional 9,821 novel sequences were obtained by single-pass directional sequencing of normalized cDNA libraries produced from mRNA obtained from 5 d, 9 d, and 14 d whole fry. The fry and TIGR gene index sequences were clustered using CAP3 software [<xref ref-type="bibr" rid="B32">32</xref>] and the resulting contigs and singlets were assigned an IpCG designation (<italic>Ictalurus punctatus </italic>Catfish Genetics). Subsequent to the production of the microarray, the sequence data from the fry cDNA libraries were submitted to GenBank as accessions <ext-link ext-link-type="gen" xlink:href="CV987367">CV987367</ext-link> – <ext-link ext-link-type="gen" xlink:href="CV997187">CV997187</ext-link> and were incorporated into release 6.0 of the TIGR gene index. The tables were updated to include the release 6.0 Tentative Consensus sequence (TC) designation of these sequences where applicable. The catfish ESTs were compared against the GenBank non-redundant peptide database using BlastX, and only those with E value = 0.005 (minimum bit score = 100) were annotated.</p><p>Microarray production, hybridization, and imaging were performed by NimbleGen Systems Inc. (Madison, WI). A total of 18,989 unique catfish ESTs were used to construct the high-density catfish oligonucleotide DNA microarrays. The microarrays were synthesized <italic>in situ </italic>using photo deprotection chemistry with the Maskless Array Synthesizer system [<xref ref-type="bibr" rid="B33">33</xref>]. Twenty 24-mer oligonucleotides (features) were designed to represent each EST; ten were perfectly matched oligos positioned throughout the sequence, and these ten were duplicated but with two mismatched bases at positions #13 and #19. The feature size for the microarray was 16 μM × 16 μM, and there were 382,409 features within the 17 mm × 13 mm array area. Of these, 379,652 were catfish-specific, and the remaining 2,757 were used for quality control of oligonucleotide synthesis and hybridization, and for signal normalization.</p><p>The microarrays were pre-hybridized with 1X MES hybridization buffer (100 mM MES, 1.0 M Na+, 20 mM EDTA, 0.01% Tween20) plus 40 μg of herring sperm DNA and 200 μg of acetylated BSA at 45°C for 15 min, followed by hybridization in the same buffer with 10 μg of denatured and fragmented cRNA per microarray at 45°C for 16–20 h with constant rotation. After hybridization, the microarrays were immediately washed extensively with non-stringent wash buffer (6X SSPE, 0.01% Tween20) at room temperature, then with stringent wash buffer (100 mM MES salt and free acid solution, 0.1 M Na+, 0.01% Tween20) at 45°C. After final rinsing with non-stringent wash buffer, the microarrays were stained with 1X Stain buffer (100 mM MES, 1 M Na+, 0.05% Tween20, 50 mg/ml of BSA, and 1 mg/ml of Cy3-streptavidin) at room temperature for 25 min. After the stain buffer was removed, the microarrays were rinsed with non-stringent wash buffer and immediately dried under argon gas. The microarrays were scanned with an Axon GenePix 4000B scanner (Molecular Devices Corp., Union City, CA) at 5 μM resolution. The data were extracted from the raw images with NimbleScan software (Nimblegen, Inc.). A total of ten microarrays were used for this project: two replicates for the control pool and one microarray for each of the eight LPS-stimulated individuals.</p></sec><sec><title>Data analysis, statistics, and bioinformatics</title><p>Relative signal intensity (log2 transformed) for each feature was generated using the Robust Multi-Array Average (RMA) algorithm [<xref ref-type="bibr" rid="B34">34</xref>]. The signal intensity was background corrected based on the quantile normalization process [<xref ref-type="bibr" rid="B35">35</xref>,<xref ref-type="bibr" rid="B36">36</xref>] for all microarrays from the entire experiment (under uniform conditions) using data from only perfectly matched oligos. Only oligos with a ratio of signal to global average background greater than 2.0, with a perfectly matched signal 150% of the mismatched signal, and with a correlation > 0.9 between biological replicates were included. The normalized data was analyzed using Significance Analysis of Microarrays [<xref ref-type="bibr" rid="B37">37</xref>] from the TIGR Multiexperiment Viewer analysis software package [<xref ref-type="bibr" rid="B38">38</xref>] with two-class unpaired design for each time point. The RMA value from perfectly matched oligos was used to calculate fold differences as the average signal intensity of treated animals divided by the average signal intensity of control animals. The list of significant genes that changed at least 2-fold was generated with an estimated global false discovery rate <10%. Differentially expressed sequences were compared to the Ensembl peptide databases derived from the zebrafish (v.5) and Tetraodon (v.7) genomes [<xref ref-type="bibr" rid="B39">39</xref>] using BlastX (E-value cutoff 0.005, minimum bit score = 100) and annotated accordingly [see <xref ref-type="supplementary-material" rid="S1">Additional file 1</xref> and <xref ref-type="supplementary-material" rid="S2">Additional file 2</xref>].</p></sec><sec><title>Real-time RT-PCR</title><p>Real-time RT-PCR analysis was carried out with the TaqMan One-Step RT-PCR kit (Applied Biosystems, Foster City, CA) in a 25 μL reaction volume containing 200 nM each amplification primer, 100 nM dual-labelled probe [see <xref ref-type="supplementary-material" rid="S2">Additional file 2</xref>], and 200 ng of total RNA. Products were amplified in an iCycler iQ™ Real Time PCR Detection System (Bio-Rad Laboratories, Hercules, CA) with the following profile: 48°C for 30 min; 95°C for 10 min; 40 cycles of 95°C for 30s, 60°C for 1 min. Expression levels of β-actin were used as endogenous controls within each sample [see <xref ref-type="supplementary-material" rid="S2">Additional file 2</xref>]. Relative levels of gene expression were calculated using the 2<sup>ΔΔCT</sup> method [<xref ref-type="bibr" rid="B40">40</xref>].</p></sec></sec><sec><title>Authors' contributions</title><p>RWL conceived the study, carried out the experiments, provided initial data analysis, and drafted the manuscript. GCW designed the microarray, participated in data analysis, and co-wrote the manuscript. All authors read and approved the final 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</title><p>Genes up-regulated at least 2-fold after LPS exposure.</p></caption><media xlink:href="1471-2164-7-134-S1.doc" mimetype="application" mime-subtype="msword"><caption><p>Click here for file</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="S2"><caption><title>Additional File 2</title><p>Genes down-regulated at least 2-fold after LPS exposure.</p></caption><media xlink:href="1471-2164-7-134-S2.doc" mimetype="application" mime-subtype="msword"><caption><p>Click here for file</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="S3"><caption><title>Additional File 3</title><p>Comparison of relative gene expression levels between microarray and real-time RT-PCR platforms.</p></caption><media xlink:href="1471-2164-7-134-S3.doc" mimetype="application" mime-subtype="msword"><caption><p>Click here for file</p></caption></media></supplementary-material></sec> |
LINE FUSION GENES: a database of LINE expression in human genes | <sec><title>Background</title><p>Long Interspersed Nuclear Elements (LINEs) are the most abundant retrotransposons in humans. About 79% of human genes are estimated to contain at least one segment of LINE per transcription unit. Recent studies have shown that LINE elements can affect protein sequences, splicing patterns and expression of human genes.</p></sec><sec><title>Description</title><p>We have developed a database, LINE FUSION GENES, for elucidating LINE expression throughout the human gene database. We searched the 28,171 genes listed in the NCBI database for LINE elements and analyzed their structures and expression patterns. The results show that the mRNA sequences of 1,329 genes were affected by LINE expression. The LINE expression types were classified on the basis of LINEs in the 5' UTR, exon or 3' UTR sequences of the mRNAs. Our database provides further information, such as the tissue distribution and chromosomal location of the genes, and the domain structure that is changed by LINE integration. We have linked all the accession numbers to the NCBI data bank to provide mRNA sequences for subsequent users.</p></sec><sec><title>Conclusion</title><p>We believe that our work will interest genome scientists and might help them to gain insight into the implications of LINE expression for human evolution and disease.</p></sec><sec><title>Availability</title><p><ext-link ext-link-type="uri" xlink:href="http://www.primate.or.kr/line"/></p></sec> | <contrib id="A1" contrib-type="author"><name><surname>Kim</surname><given-names>Dae-Soo</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>kds2465@pusan.ac.kr</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Kim</surname><given-names>Tae-Hyung</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>kth2001@pusan.ac.kr</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Huh</surname><given-names>Jae-Won</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>primate@pusan.ac.kr</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Kim</surname><given-names>Il-Chul</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>ilchulkim2000@yahoo.co.kr</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Kim</surname><given-names>Seok-Won</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>javamint@kribb.re.kr</email></contrib><contrib id="A6" corresp="yes" contrib-type="author"><name><surname>Park</surname><given-names>Hong-Seog</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>hspark@kribb.re.kr</email></contrib><contrib id="A7" corresp="yes" contrib-type="author"><name><surname>Kim</surname><given-names>Heui-Soo</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I2">2</xref><email>khs307@pusan.ac.kr</email></contrib> | BMC Genomics | <sec><title>Background</title><p>Most retroelements have been considered harmful because they cause accumulation of insertion and deletion mutations in the host genome [<xref ref-type="bibr" rid="B1">1</xref>]. Mutation of retroelements could affect gene transcription and translation. However, recent investigations have shown that HERV and Alu elements in the intron or flanking regions of functional human genes provide alternative promoters, splicing sites and polyadenylation signals [<xref ref-type="bibr" rid="B2">2</xref>,<xref ref-type="bibr" rid="B3">3</xref>]. Unlike HERV and Alu, LINE elements tend to contain multiple potential splice sites (ESE) [<xref ref-type="bibr" rid="B4">4</xref>] and polyadenylation signals [<xref ref-type="bibr" rid="B5">5</xref>] in their sequences. There are four types of transposable elements in the human genome: long interspersed nuclear elements (LINEs or L1s) or non-long terminal repeat retrotransposons, short interspersed nuclear elements (SINEs), LTR retrotransposons (endogenous retroviruses) and DNA transposons [<xref ref-type="bibr" rid="B1">1</xref>], which together constitute 45% of the total genome. Most of these elements are inactive. However, a few LTR elements have been shown to contain intact open reading frames (ORFs) [<xref ref-type="bibr" rid="B6">6</xref>], and LINE elements also have the capacity for autonomous retrotransposition [<xref ref-type="bibr" rid="B7">7</xref>,<xref ref-type="bibr" rid="B8">8</xref>]. SINE elements cannot be expressed by themselves and depend on L1 elements for active mobility [<xref ref-type="bibr" rid="B9">9</xref>]. The L1 elements constitute about 17% of the human genome and are present in an estimated 79% of human genes in at least one copy [<xref ref-type="bibr" rid="B10">10</xref>].</p><p>The full length of L1 is about 6 kb. It consists of a 5' untranslated region (5'UTR); two nonoverlapping open reading frames (ORF1 and ORF2) encoding an RNA binding protein [<xref ref-type="bibr" rid="B11">11</xref>], an endonuclease [<xref ref-type="bibr" rid="B12">12</xref>] and a reverse transcriptase [<xref ref-type="bibr" rid="B13">13</xref>]; and a 3'UTR that ends in an AATAAA polyadenylation signal and a polyA tail [<xref ref-type="bibr" rid="B9">9</xref>]. The Alu and SVA transposable elements and processed pseudogenes are believed to have been inserted into the genome by borrowing the endonuclease and reverse transcriptase from L1 elements [<xref ref-type="bibr" rid="B14">14</xref>-<xref ref-type="bibr" rid="B16">16</xref>]. The L1 element itself has also been inserted into new genomic locations during mammalian evolution. Such elements are mostly truncated and rearranged to form inactive copies of their progenitors. These insertional mutations are reported to be associated with twelve genetic diseases [<xref ref-type="bibr" rid="B17">17</xref>] and also contribute to protein variability or versatility [<xref ref-type="bibr" rid="B18">18</xref>].</p><p>Active or functional L1 elements, which are involved in shaping the human genome, are differentiated into three types depending on where they are inserted into the genome. First, a 6 kb-long full-length or variable-length 5'-truncated L1 element is inserted into the 5'UTR or introns of a gene, affecting its expression. In this process, LINE elements are probably reverse transcribed and integrated in the new location by target-primed reverse transcription (TPRT) [<xref ref-type="bibr" rid="B19">19</xref>]. LINE elements have provided not only many internal promoters at new genomic locations, but also 5'-UTR-located internal promoters, which could guide the transcription of many adjacent genes [<xref ref-type="bibr" rid="B20">20</xref>]. Second, retrotransposition of the L1 element results in the transduction of a 3'-UTR flanking fragment to a new genomic location; this is due to the effect of the ambiguous L1 polyadenylation signal [<xref ref-type="bibr" rid="B21">21</xref>]. Third, the L1 components are shuffled into exons, affecting the splicing site at transcription and consequently leading to the production of alternative mRNA transcripts [<xref ref-type="bibr" rid="B22">22</xref>].</p><p>Assembling genomic information and constructing a web-database of genome annotations and genes with particular functions is generally useful for implementing functional studies and for understanding evolutionary genomic organization. Representative web-databases of transposable elements in the human genome have been reported: a database of Alu elements incorporated within protein-coding gene [<xref ref-type="bibr" rid="B2">2</xref>], an HERV expression and structure analysis system [<xref ref-type="bibr" rid="B3">3</xref>] and a system for extrapolating functional annotation to the prediction of active LINE-1 elements [<xref ref-type="bibr" rid="B23">23</xref>]. Although it is well established that information about the structure and position of LINE elements in genes is important for functional studies of genetic diseases, such data are limited and are not included in any database that allows large amounts of scattered information to be searched easily. To address this deficiency, we developed a database for LINE expression and structure in the human genome, LINE FUSION GENES. Our database provides the structures and expression patterns of LINE elements including their relative positions in the genes, and additional information such as the tissue distribution and chromosomal location of the genes and their domain structures. To enhance ease of access for subsequent users, we linked all of the accession numbers to the NCBI data bank to provide mRNA sequences.</p></sec><sec><title>Construction and content</title><sec><title>Identification of transcript variants by LINE insertion (LINE FUSION GENES)</title><p>First, 28,171 mRNA human-gene sequences and human expressed sequence tags (EST) were downloaded from the NCBI database Build 35 (INSDC, <ext-link ext-link-type="uri" xlink:href="http://insdc.org"/>) and aligned with genomic assembly sequences (Build 35) using the SIM4 program [<xref ref-type="bibr" rid="B24">24</xref>]. Only alignments showing >97% sequence identity were used for further stages. As a result, we extracted positional information about the exon and genome sequences to be matched. On the basis of this information we collected contiguous sequences from 5 kb upstream of the 5'UTR end to the same distance downstream of the 3' UTR end. All the sequences were stored as mapping data for each gene. In addition, the DNA sequences of the LINE elements (LINE-1, LINE-2, LINE-3) were downloaded from Repbase Update [<xref ref-type="bibr" rid="B25">25</xref>]. We constructed a LINE component library, using BLASTX, from these 205 downloaded sequences, which included 5'UTR, ORF1, ORF2 and 3'UTR.</p><p>We used RepeatMasker <ext-link ext-link-type="uri" xlink:href="http://repeatmasker.genome.washington.edu"/> to search for LINE sequences in the contiguous segments. For each gene entry, LINE locations on the contig, orientation and sequence were stored in the database. The locations of LINEs and exons on each contig were calculated from their positions. We then merged them on the basis of their positions and found that 4,489 LINEs were fused on 5' UTR (1,392), 3'UTR (2,167) and exonization (930). Finally, we constructed the LINE FUSION GENES database for chimeric transcripts containing L1-5'UTR heads and cellular sequence tails (102) and L1-3'UTR incorporated within transcripts tails (676), and the LINE elements that led to novel splice variants (632). Information about tissue expression and pathogenic LINE fusion transcripts was obtained by gene expression vocabulary (eVOC) annotations of cDNA library sources [<xref ref-type="bibr" rid="B26">26</xref>].</p></sec><sec><title>Classification of the LINE FUSION GENES</title><p>As shown Figure <xref ref-type="fig" rid="F1">1</xref>, we classified the LINE FUSION GENES into three types, alternative promoter, alternative polyadenylation signal and exonization, on the basis of the effects of their insertion in the genes. These effects of LINE insertion depend on position and sequence.</p><sec><title>Type I. Alternative promoter</title><p>LINE FUSION GENES of Type I involve insertion near the 5'UTR of the gene or in an intron. LINEs have their own sense and antisense promoters in their 5'UTRs. Consequently, Type I genes might be transcribed from the promoters of the inserted LINE rather than from the cellular promoter. Previously, several cases of Type I LINE FUSION GENES have been reported [<xref ref-type="bibr" rid="B27">27</xref>].</p></sec><sec><title>Type II. Alternative polyadenylation signal</title><p>If LINE elements have a polyadenylation signal within the 3' UTR gene flanking region, they could be responsible for a transduction event [<xref ref-type="bibr" rid="B8">8</xref>]. Such LINE expression occurs occasionally in human genes; the transcript is stopped by the LINE polyadenylation signal rather than the one endogenous to the gene. When the LINE is incorporated into the intron behind the 3'UTR, transcription is again occasionally stopped by the LINE polyadenylation signal rather than that of the gene. We classified such genes as Type II LINE FUSION GENES. In other words, Type II LINE FUSION GENES are LINE fusion genes with LINE polyadenylation signals on their 3' UTRs.</p></sec><sec><title>Type III. Exonization</title><p>Generally, the intron sequences are spliced out by the spliceosome, which recognizes the splicing site (AG-GT) between the intron and the exon. Most LINEs inserted into introns are spliced out and do not affect target gene expression. However, recent studies have shown that some LINEs can be recognized as splicing sites (AG-GT) or as intact exons by the spliceosome [<xref ref-type="bibr" rid="B28">28</xref>]. Consequently, the LINE sequences are fused to mRNA coding sequences. We classified these genes as Type III LINE FUSION GENES.</p></sec></sec></sec><sec><title>Utility and discussion</title><p>LINE FUSION GENES uses JSP technology; the data come from a primary database. Users can efficiently retrieve three modes of information concerning LINE expression within genes. First, they can search LINE expression within a gene by typing a gene ID or clicking on the gene name listed on the view page according to its chromosomal location. Second, the database provides type information in which LINE expression is classified into three types (alternative promoter, alternative polyadenylation signal and exonization). The type information can help users to speculate more readily about the effects of LINE expression within interesting genes. Third, users can search interesting genes using accession numbers from the NCBI data bank or from the HUGO symbol name provided on the view page, and even acquire mRNA sequences from the NCBI data bank for further study.</p><p>The result pages are listed in a tabular format that provides the evidence for and information about LINE expression within genes. As shown in Figure <xref ref-type="fig" rid="F2">2</xref>, the LINEs are visualized by colors: red (5' UTR elements), blue (3' UTR elements) and green (ORF1 and ORF2). LINE fusion regions within mRNAs are indicated in red. Moreover, detailed information about the LINE fusion regions are displayed in the table on the result page. Occasionally, LINE incorporation results in domain changes in a protein. In order to speculate about these domain changes, users can check the domain description on the page. The domain information includes the results obtained from searching queries about genes with LINEs by RPS-BLAST [<xref ref-type="bibr" rid="B29">29</xref>].</p></sec><sec><title>Conclusion</title><p>From our in silico analysis of the human genome, 1,329 genes were identified as being affected by LINE elements during expression. LINE FUSION GENES is continually supplemented with new human gene data from the available sources. We are planning to update the database with full length human cDNA data obtained from various clinical samples representing human diseases. Through this update, we will be able to profile the patterns of LINE expression in various diseases and to identify LINEs that affect the expression of functional human genes. We will also supplement the database with LINE fusion genes from other mammalian species and compare them with those of humans. We also envision the integration of our HESAS [<xref ref-type="bibr" rid="B3">3</xref>] and LINE FUSION GENES databases, intended for release in 2007. We believe that our work will help us to gain insight into the implications of LINE expression for human evolution and disease.</p></sec><sec><title>Availability and requirements</title><p>LINE FUSION GENES is publicly available at the URL <ext-link ext-link-type="uri" xlink:href="http://www.primate.or.kr/line"/>. Questions and comments are welcomed through the site.</p></sec><sec><title>Abbreviations</title><p>LINE – Long Interspersed Element</p><p>HERV – Human Endogenous Retrovirus</p><p>SINE – Short Interspersed Nucleotide Element</p><p>ORF – Open Reading Frame</p><p>BLAST – Basic Local Alignment Search Tool</p><p>JSP – Java Server Pages</p><p>RPS-BLAST – Reversed Position Specific Blast</p><p>HESAS – HERVs Expression and Structure Analysis System</p><p>EST – Expressed Sequence Tag</p><p>UTR – Untranslated Regions</p><p>NCBI – National Center for Biotechnology Information</p><p>HUGO – Human Genome Organisation</p><p>INSDC – International Nucleotide Sequence Databases</p></sec><sec><title>Authors' contributions</title><p>DS Kim analyzed the contents of the paper and wrote the manuscript. HS Kim participated in the analysis and provided essential direction. TH Kim provided biological context and guidance during the initial phase of the bioinformatics analysis. HS Park and IC Kim contributed the manuscript correction and continuous discussions. SW Kim helped in the general design of the database and the user interface. JW Huh provided biological direction. All authors read and approved the final manuscript.</p></sec> |
Differential evolutionary conservation of motif modes in the yeast protein interaction network | <sec><title>Background</title><p>The importance of a network motif (a recurring interconnected pattern of special topology which is over-represented in a biological network) lies in its position in the hierarchy between the protein molecule and the module in a protein-protein interaction network. Until now, however, the methods available have greatly restricted the scope of research. While they have focused on the analysis in the resolution of a motif topology, they have not been able to distinguish particular motifs of the same topology in a protein-protein interaction network.</p></sec><sec><title>Results</title><p>We have been able to assign the molecular function annotations of Gene Ontology to each protein in the protein-protein interactions of <italic>Saccharomyces cerevisiae</italic>. For various motif topologies, we have developed an algorithm, enabling us to unveil one million "motif modes", each of which features a unique topological combination of molecular functions. To our surprise, the conservation ratio, i.e., the extent of the evolutionary constraints upon the motif modes of the same motif topology, varies significantly, clearly indicative of distinct differences in the evolutionary constraints upon motifs of the same motif topology. Equally important, for all motif modes, we have found a power-law distribution of the motif counts on each motif mode. We postulate that motif modes may very well represent the evolutionary-conserved topological units of a protein interaction network.</p></sec><sec><title>Conclusion</title><p>For the first time, the motifs of a protein interaction network have been investigated beyond the scope of motif topology. The motif modes determined in this study have not only enabled us to differentiate among different evolutionary constraints on motifs of the same topology but have also opened up new avenues through which protein interaction networks can be analyzed.</p></sec> | <contrib id="A1" contrib-type="author"><name><surname>Lee</surname><given-names>Wei-Po</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>wplee@nuk.edu.tw</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Jeng</surname><given-names>Bing-Chiang</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>jeng@mail.nsysu.edu.tw</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Pai</surname><given-names>Tun-Wen</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>twp@mail.ntou.edu.tw</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Tsai</surname><given-names>Chin-Pei</given-names></name><xref ref-type="aff" rid="I4">4</xref><email>cptsai@pu.edu.tw</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Yu</surname><given-names>Chang-Yung</given-names></name><xref ref-type="aff" rid="I4">4</xref><email>cyyu@pu.edu.tw</email></contrib><contrib id="A6" corresp="yes" contrib-type="author"><name><surname>Tzou</surname><given-names>Wen-Shyong</given-names></name><xref ref-type="aff" rid="I5">5</xref><xref ref-type="aff" rid="I6">6</xref><email>wstzou@ntou.edu.tw</email></contrib> | BMC Genomics | <sec><title>Background</title><p>In the post-genomic era, one major goal of functional genomics has been to identify and analyze molecular interactions in a cellular context to better understand the mechanisms according to which biological molecules interact and function. A protein-protein interaction (PPI) is a specific type of molecular interaction that plays the central role in relaying signals, in building molecular machines, in engaging in enzyme reactions and in decision-making vis-à-vis multiple biological processes. Advancements in PPI-detection technology have unquestionably led to the rapid accumulation of PPI data [<xref ref-type="bibr" rid="B1">1</xref>-<xref ref-type="bibr" rid="B5">5</xref>]. To accurately conceptualize a PPI by determining its precise description is to immediately allow for the utilization of the various tools currently accessible in network science [<xref ref-type="bibr" rid="B6">6</xref>]. A PPI network, for instance, has been found to have a scale-free structure; i.e., the link count (interactions) of a protein (node) follows a broad-tailed distribution that is approximated as a power-law,<italic>P(k)~k</italic><italic><sup>-γ</sup></italic>, where <italic>k </italic>is the link count and γ is the degree exponent [<xref ref-type="bibr" rid="B6">6</xref>-<xref ref-type="bibr" rid="B9">9</xref>]. A network diameter, defined as the average minimal path between two nodes, is reportedly small (~6–7) in a PPI network [<xref ref-type="bibr" rid="B9">9</xref>,<xref ref-type="bibr" rid="B10">10</xref>].</p><p>The contribution made by research on network motifs, where a network motif is the specific topology of a combination of nodes that occurs repeatedly at different positions in that network, is evidenced in the results obtained from the initial search for cause of the topological properties. As the first step, one of the properties of a PPI network, the scale-free feature, was thoroughly examined [<xref ref-type="bibr" rid="B11">11</xref>,<xref ref-type="bibr" rid="B12">12</xref>]. At issue was whether the selection force is the cause of one of the scale-free network features: a random mutation does not harm a network, as a whole, but can cause it to collapse, but only with a deliberate attack on the hubs (nodes that contain many immediate neighbors) [<xref ref-type="bibr" rid="B13">13</xref>]. It was, therefore, reasonably inferred and later clearly observed that hubs are more likely to evolve at a slower rate than non-hub nodes [<xref ref-type="bibr" rid="B14">14</xref>,<xref ref-type="bibr" rid="B15">15</xref>]. However, different conclusions have been drawn regarding the mutation rate of a hub [<xref ref-type="bibr" rid="B12">12</xref>,<xref ref-type="bibr" rid="B16">16</xref>]. In addition to this, recent research into gene regulatory networks in yeast has uncovered both a power-law distribution for the number of regulated genes per regulating protein [<xref ref-type="bibr" rid="B17">17</xref>] and, at the same time, has yielded invaluable information concerning the presence of typical patterns of motifs. Feed forward motifs, single-input motifs and dense overlapping regulons [<xref ref-type="bibr" rid="B18">18</xref>-<xref ref-type="bibr" rid="B21">21</xref>] reportedly occur in the gene regulatory network (transcription factor vs. the regulated gene) of <italic>Saccharomyces cerevisiae </italic>and <italic>Escherichia coli </italic>with greater frequency than they would based on chance alone. Network motifs have also been analyzed in the PPI network of <italic>Saccharomyces cerevisiae</italic>, and from this, it has been concluded that motif topology is correlated with the conservation of motif proteins; besides this, it has been inferred that motifs probably represent the evolutionary-conserved topological units of cellular networks [<xref ref-type="bibr" rid="B22">22</xref>].</p><p>True that the representation of a gene regulatory network where the transcriptional regulator is considered the master and the regulated gene the slave [<xref ref-type="bibr" rid="B18">18</xref>,<xref ref-type="bibr" rid="B19">19</xref>,<xref ref-type="bibr" rid="B22">22</xref>-<xref ref-type="bibr" rid="B24">24</xref>] is appropriate, but such a dichotomous (master-slave) representation cannot be used for a PPI since proteins have multiple functions. In this study, we labeled the protein nodes on the basis of their functional attributes of Gene Ontology (GO) [<xref ref-type="bibr" rid="B25">25</xref>]. We then pursued any recurring patterns of the functional attributes of protein interactions. With this new representation of a PPI network and by categorizing the repertoire of network motifs of the same topology into "motif modes," with each motif mode featuring a special topological combination of molecular functions, we have been able to move one step ahead of what was accomplished by Barabási's group [<xref ref-type="bibr" rid="B22">22</xref>]. This, in turn, led us to the finding that the evolutionary constraints on the motifs of the same topology are certainly not the same when considering their functional attributes; on the contrary, they vary a great deal.</p></sec><sec><title>Results</title><sec><title>Motif topology</title><p>The PPI data for <italic>Saccharomyces cerevisiae </italic>contain a total of 15,129 physical interactions formed by 4,738 proteins (see Methods). We analyzed the occurrence of eight types of three-node and four-node motifs, referred to hereafter as the "motif topology" (Figure <xref ref-type="fig" rid="F1">1</xref>). The motif counts (i.e., the number of instances) of eight motif topologies varies, from 5.3 thousand (four-node motif topology 5, #4–5 hereafter) to an overwhelming 10.6 million (four-node motif topology 0, #4–0 hereafter) (Figure <xref ref-type="fig" rid="F2">2</xref>; see <xref ref-type="supplementary-material" rid="S1">additional file 1</xref>-Table S1).</p></sec><sec><title>Motif mode</title><p>Since we focused on the recurring patterns of the functional attributes of protein interactions, we used the molecular function ontology. On these grounds, we used the molecular function ontology (hereafter GO terms) to annotate each node of a PPI network (Figure <xref ref-type="fig" rid="F3">3</xref>).</p><p>We developed a new algorithm (see Methods) to categorize network motifs into "motif modes" on the basis of the topology of the GO terms annotated for motif nodes (Figure <xref ref-type="fig" rid="F1">1</xref>). Here a "motif mode" is defined as a special combination of GO terms in a motif. In fact, a motif mode illustrates a grouping of network motifs that contain the same topological combination of functional attributes of proteins (Figure <xref ref-type="fig" rid="F1">1</xref>). When we employed the GO terms at depths five and six, we found a total of nearly 1.3 and 1.7 million motif modes, respectively (Figure <xref ref-type="fig" rid="F2">2</xref>; see <xref ref-type="supplementary-material" rid="S1">additional file 1</xref>-Table S1). In general, the higher motif count a particular motif topology has, the more motif modes there are in that motif topology (correlation coefficient = 0.96, <italic>p </italic>< 0.001). However, #4–5 has the largest counts per motif (on average, around eight motif counts per motif mode; Figure <xref ref-type="fig" rid="F4">4</xref>; see <xref ref-type="supplementary-material" rid="S1">additional file 1</xref>-Table S1).</p><p>Noteworthy too is that the motif count for each motif mode can vary considerably among a million motif modes. We calculated the distribution of the motif count for each motif mode based on the logarithmic binning. When we employed the GO terms at depths five and six, nearly 49% and 58% of the motif modes respectively occur less than 2 times in a PPI network (Figure <xref ref-type="fig" rid="F5">5</xref>; see <xref ref-type="supplementary-material" rid="S1">additional file 1</xref>-Table S2). It is particularly interesting that, for all motif modes, there is a power-law distribution of the motif counts, <italic>P(m)~m</italic><sup><italic>-φ</italic></sup>, where <italic>m </italic>is the motif count and <italic>φ </italic>is the degree exponent ~1.6 (R square = 0.99, <italic>p </italic>< 0.001) (Figure <xref ref-type="fig" rid="F5">5</xref>). In fact, quite a few motif modes contain more than one hundred motif counts (1,577 and 1,449 motif modes when we employed the GO terms at depths five and six, respectively).</p><p>In that motif modes allow for an analysis that extends beyond motif topology, we were able to look at the properties of motifs of the same topology at a higher resolution than had ever been done before [<xref ref-type="bibr" rid="B22">22</xref>].</p></sec><sec><title>Annotation transfer within a protein interaction network</title><p>Since we were interested in finding the evolutionary trend of the proteins contained in a motif mode, we built an orthology gene list among the genes of the six organisms (<italic>Saccharomyces cerevisiae</italic>, <italic>Arabidopsis thaliana</italic>, <italic>Caenorhabditis elegans</italic>, <italic>Drosophila melanogaster</italic>, <italic>Mus musculus</italic>, and <italic>Homo sapiens</italic>). When the orthology gene list was used to compute the evolutionary constraint of motifs of different motif topology and motif modes, it was assumed that orthologous proteins also interact with each other. It has been shown that the reliability of the annotation of protein interactions from one organism to another depends on the level of sequence identity for the two pairs of interacting proteins [<xref ref-type="bibr" rid="B26">26</xref>]. The joint E-value (the geometric means of the BLAST E-values for the two pairs of interacting proteins) smaller than <10<sup>-70 </sup>was used as the threshold to evaluate whether such a transfer of the annotation of interactions is reliable [<xref ref-type="bibr" rid="B26">26</xref>]. In order to investigate the validity of the orthology gene list by which the interaction annotations were transferred between organisms, we conducted a sequence comparison between orthologs of yeast and five other species. We found 61% of the orthologous protein pairs have an E-value lower than 10<sup>-70 </sup>(see <xref ref-type="supplementary-material" rid="S2">additional file 2</xref>-Sup2.doc, Figure S1). About 40% of the annotation transfers are less reliable; just the same, we have to emphasize that the orthology approach still remains important in terms of decoding many important biological phenomena [<xref ref-type="bibr" rid="B22">22</xref>,<xref ref-type="bibr" rid="B27">27</xref>].</p></sec><sec><title>Different evolutionary constraints on the motifs of the same topology</title><p>To discover the evolutionary trend of the proteins contained in a motif mode, we computed the "conservation fraction" for each motif topology. The "conservation fraction" is the counts of the evolutionary fully conserved motif divided by total motif counts. Motifs #3–1, #4–4, and #4–5 have higher conservation fractions (0.19 ~0.26) than do the others (Figure <xref ref-type="fig" rid="F6">6</xref>; see <xref ref-type="supplementary-material" rid="S1">additional file 1</xref>-Table S3). Important to note, these values are very similar to those reported earlier (0.22 vs. 0.21 for #3–1, 0.19 vs. 0.19 for #4–4, 0.26 vs. 0.33 for #4–5) [<xref ref-type="bibr" rid="B22">22</xref>]. We also calculated the distribution of the conservation fractions of each motif mode using 0.1 as the bin. When we employed the GO terms at depths five and six, the conservation fraction of nearly 93% and 94% of all motif modes respectively was less than 0.1 (Figure <xref ref-type="fig" rid="F7">7</xref>; see <xref ref-type="supplementary-material" rid="S1">additional file 1</xref>-Table S4). This is due to the fact that the majority of motif counts occur for #4–0 and #4–1 (Figure <xref ref-type="fig" rid="F2">2</xref>), with the average conservation fraction of ~0.05 (Figure <xref ref-type="fig" rid="F6">6</xref>). However, we also noted that the conservation fraction for 3.4% and 3.6% of motif modes, when we employed the GO terms at depths five and six respectively, reaches 1.0. The general trend stated above still holds if the distribution is shown on the basis of each motif topology (see <xref ref-type="supplementary-material" rid="S2">additional file 2</xref>-Sup2.doc, Figure S2).</p><p>We defined the "conservation ratio", a degree of evolutionary constraint, as the value of the conservation fraction stated above divided by the same value but computed after the random assignment of the orthology data to the proteins (see Methods). It is apparent that the more connected to each other the motif nodes are, the higher is the conservation ratio (e.g., #4–5 > #4–4 > #4–3 > #4–2 > #4–1 > #4–0; #3–1 > #3–0) (correlation coefficient = 0.71, <italic>p </italic>= 0.05; Figure <xref ref-type="fig" rid="F8">8</xref>; see <xref ref-type="supplementary-material" rid="S1">additional file 1</xref>-Table S3). This observation has been previously reported [<xref ref-type="bibr" rid="B22">22</xref>].</p><p>To determine if the conservation ratio is the same for motif modes of the same topology, we calculated the conservation ratio for one million motif modes. We calculated the distribution of the conservation ratio for each motif mode based on the logarithmic binning. What we found is that when we employed the GO terms at depths five and six, the conservation ratio for more than 93% and 94% of the motif modes is respectively lower than 2. Surprisingly, we also noted that the conservation ratio for 3,510 (GO terms at depth five) and 3,096 (GO terms at depth six) motif modes exceeds 50, which is approximately the magnitude of the highest ratio observed for motif #4–5 (Figures <xref ref-type="fig" rid="F8">8</xref>, <xref ref-type="fig" rid="F9">9</xref>; see <xref ref-type="supplementary-material" rid="S1">additional file 1</xref>-Table S5). Support for the notion that the evolutionary constraints on motifs of the same topology are not the same gains considerable ground, and this suggests the presence of differential evolutionary constraints upon motif modes of the same motif topology. We can fit the distribution of the conservation ratio for all motif modes on a logarithmic scale by employing a quadratic function if the value of the first bin (0~2) is not taken into account (R square = 0.97, <italic>p </italic>< 0.001). The above observations still hold if the distribution is shown on the basis of each motif topology (see <xref ref-type="supplementary-material" rid="S2">additional file 2</xref>-Sup2.doc, Figure S3).</p></sec><sec><title>A motif mode by chance alone?</title><p>As a unique combination of molecular function descriptions of GO, a motif mode could possibly be the result of a special classification that occurs by chance. We randomized the GO annotations (molecular function) on the nodes and re-calculated three features of each motif mode: the motif counts, the conservation fraction and the conservation ratio. We found a significant difference in the motif counts and the conservation ratio of the motif modes with and without the randomization of the GO annotations (<italic>p </italic><= 0.0001; see <xref ref-type="supplementary-material" rid="S1">additional file 1</xref>-Table S6). Therefore, the motif modes categorized by the current GO annotations and the properties derived from the motif modes (e.g., the level of differential evolutionary conservation) do indeed bear a greater significance than would normally be expected by chance alone.</p></sec></sec><sec><title>Discussion</title><p>We used nearly 400 and 650 GO molecular function descriptions at different depths to annotate nearly 5,000 protein nodes in the <italic>Saccharomyces cerevisiae </italic>PPI network. We employed a motif mode to represent any probable combination of GO annotations with a three-node and four-node topology. To this effect, we collected all of the existing million motif modes and examined the level of evolutionary constraints on their motif constituents (the conservation ratio). We found two interesting distributions of the properties of the motif modes. The first is the distribution of motif counts a motif mode consists of, <italic>P(m)~m</italic><sup><italic>-φ</italic></sup>, where <italic>m </italic>is the motif count and φ is the degree exponent ~1.6 (Figure <xref ref-type="fig" rid="F5">5</xref>). This observation may be related to a recent finding that, in a complex network, the large-scale topological organization (characterized by the degree exponents of the scale-free and hierarchical modularity) and the variable counts of the different motif topologies can define each other [<xref ref-type="bibr" rid="B28">28</xref>,<xref ref-type="bibr" rid="B29">29</xref>]. Whether there is a direct correlation between the large-scale topological organization and the motif mode-motif count dependence has yet to be determined. The second is the distribution that characterizes the dependence of the number of motif modes on the conservation ratio of that motif mode (Figure <xref ref-type="fig" rid="F9">9</xref>, Figure S3). Reportedly, the more interconnected the nodes of a motif are with each other, the more conserved the protein constituents of the motif are [<xref ref-type="bibr" rid="B22">22</xref>]. If a motif, rather than a single protein (e.g. hub), represents evolutionary-conserved topological units in the tapestry of a PPI network [<xref ref-type="bibr" rid="B22">22</xref>], our study further shows that motifs belonging to different motif modes of the same topology are not under the same level of evolutionary constraints. Fewer motif modes are under higher evolutionary constraints, and the level of differences spans the order of three.</p><sec><title>Differences between this and previous studies</title><p>The use of GO annotations has provided us with increased insight into protein interactions through the coloring of protein nodes in the interaction map [<xref ref-type="bibr" rid="B30">30</xref>]. In this study, we categorized motifs on the basis of the topological combinations of GO terms (molecular function), and this has resulted in our discovery of motif modes. It should not be overlooked that motifs in biological networks have previously been analyzed [<xref ref-type="bibr" rid="B18">18</xref>,<xref ref-type="bibr" rid="B22">22</xref>]. The Alon group, for example, developed Mfinder to calculate the motif counts in gene-regulatory networks [<xref ref-type="bibr" rid="B18">18</xref>]. Our algorithm is capable not only of conducting motif counts but also of managing the grouping of motif modes based on GO terms; as mentioned earlier, this analysis, in fact, reaches a higher resolution than any other reported before. We recorded the protein identities and GO terms of each motif using our algorithm for the computation of the conservation ratio. We recorded all motif instances (five million) of the motif modes (one million), and they can be used in future analyses though this requires large computer memory. For this very reason, we focused on the motif modes of three nodes and four nodes in this study.</p><p>Until now, the evolutionary constraints have only been reported on the level of motif topology, and this by computing the evolutionary constraints of all motifs of the same motif topology [<xref ref-type="bibr" rid="B22">22</xref>]. In this study, for the first time, we have been able to distinguish the evolutionary constraints of different motifs of the same topology. A motif topology contains many motif modes which are under different levels of evolutionary constraints. In proposing that motif modes may represent the evolutionary-conserved topological units of a protein interaction network, we have clearly progressed one step farther beyond what has been accomplished before.</p></sec></sec><sec><title>Future direction</title><p>It is possible to re-examine the motifs and motif modes of PPIs of model organisms without relying on the orthology information. One can take several real PPI networks of different organisms and, for each network, look independently for the motif they contain and use the GO annotations to define motif modes, then compare if the same motif modes are used in these different organisms. Apart from this, it would be most enlightening to investigate the temporal patterns of various motif modes when gene expression data are used. This would be just like a PPI network visualized within the context of cell-cycles [<xref ref-type="bibr" rid="B31">31</xref>], hubs categorized as "party hubs," or "date hubs," [<xref ref-type="bibr" rid="B32">32</xref>] and topological changes in transcription regulatory networks observed under environmental or physiological conditions [<xref ref-type="bibr" rid="B33">33</xref>]. Furthermore, it would be most worthwhile enhancing the precision of module-detection methods [<xref ref-type="bibr" rid="B34">34</xref>-<xref ref-type="bibr" rid="B40">40</xref>] and developing specialized tools designed to search for motifs with similar functional annotations. This would greatly assist biologists interested in mining protein interaction networks. PathBlast [<xref ref-type="bibr" rid="B41">41</xref>], TopNet [<xref ref-type="bibr" rid="B42">42</xref>], MAVisto [<xref ref-type="bibr" rid="B43">43</xref>], and FANMOD [<xref ref-type="bibr" rid="B44">44</xref>] are examples of work in these areas.</p></sec><sec><title>Conclusion</title><p>We have reported on using the molecular function vocabulary of GO to annotate a yeast protein interaction network. The motif mode, i.e., the special topological combination of the molecular functions of interacting proteins, was extracted from the yeast protein interaction network and employed in the analysis of the protein interaction network at a higher resolution than ever before. The distribution of the motif counts for all the motif modes follows a scale-free like fashion. The differential evolutionary constraints on the million motif modes are an indication that motif modes may very well represent the evolutionary-conserved topological units of the protein interaction network.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>PPI dataset</title><p>The yeast (<italic>Saccharomyces cerevisiae</italic>) PPI dataset is from the DIP database [<xref ref-type="bibr" rid="B45">45</xref>] (July 2004). There are 15,409 interactions in the original data, and after self-interacting protein interactions were removed, 15,129 interactions composed of 4,738 genes (proteins) remained.</p></sec><sec><title>Gene Ontology associations</title><p>The molecular function of the open reading frame (ORF) name of each gene is based on GO annotations (July 8, 2004). Of 4,738 genes, 857 GO terms are used to describe the molecular functions of genes. No GO term (molecular function) can be found for 289 open reading frames, and these account for 499 interactions. An ORF-GO correspondence table is provided in the <xref ref-type="supplementary-material" rid="S1">additional file 1</xref>-Table S7.</p></sec><sec><title>Molecular functions of GO terms at different depths</title><p>GO terms are organized in structures referred to as 'directed acyclic graphs,' and each term can be traced to different depths in the hierarchies. The molecular function of the GO terms for each protein in our dataset was traced to depths five and six backward to the root. (The root is located at depth 0; "molecular function", GO:0003674, at depth one; "antioxidant activity", GO:0016209, at depth 2). If the GO term for a protein is located at a depth lower than six in the GO tree, the GO term remains unchanged. Our reason for choosing depths five and six is that, based on the statistics, the average depth of the GO terms in our dataset is six (data not shown). The 857 GO terms used to describe a gene's molecular function were then reduced to 398 and 648 terms at depths five and six, respectively.</p></sec><sec><title>Algorithm to detect motif modes</title><p>We considered a PPI network as a graph and used the method of an adjacency list to represent the graph. In this representation, we labeled the nodes, and for each node, we created a linked list to record its immediate neighbors. To count the occurrences of each motif mode, we modified and employed the graph-searching algorithm, Depth-First-Search (DFS). To start the search, we used each graph node sequentially as the initial node, with a depth limit corresponding to the number of motif nodes. We performed the DFS-based search in a recursive way and implemented a checking procedure to ensure that we did not repeatedly count any identical structure during the search. When we recognized a special topological structure, we proceeded to map the nodes of that structure to those of the motif based on the links through which the nodes were connected to each other. We then recorded the identity of the nodes of the newly found structure and added them to the set of this specific motif mode.</p></sec><sec><title>Ortholog assignment</title><p>We retrieved the orthologous sequence pairs from InParanoid [<xref ref-type="bibr" rid="B46">46</xref>] (Version 3.0; updated 15, August 2004) for a bootstrap value of 100% and a score of 1.00 in each cluster. We found 1,247 genes of <italic>Saccharomyces cerevisiae </italic>that have high-confidence orthologs with 1,440 genes from <italic>Arabidopsis thaliana</italic>, 1,286 genes from <italic>Caenorhabditis elegans</italic>, 1,397 genes from <italic>Drosophila melanogaster</italic>, 1,501 genes from <italic>Mus musculus</italic>, and 1,439 genes from <italic>Homo sapiens</italic>. We built an orthology gene list using 1,247 genes of <italic>Saccharomyces cerevisiae </italic>for the protein sequences of the genes orthologous to the five species. Out of 15,129 interactions, 1,247 genes from yeast constitute 2,629 protein interactions.</p></sec><sec><title>Conservation fraction and conservation ratio</title><p>We then had to find all the constituents of the evolutionary fully conserved motif in the orthology gene list. For each motif topology and each motif mode, we computed the "conservation fraction", i.e., the counts of the evolutionary fully conserved motifs divided by the total motif counts of the motif topology.</p><p>To identify the degree to which a motif is under higher evolutionary constraints than would normally be expected by chance alone, we computed "conservation ratio". We randomly assigned the orthology data to the proteins and re-calculated the conservation fraction, as stated above. We repeated this procedure twenty times and obtained an average value which we termed the "random conservation fraction" (see <xref ref-type="supplementary-material" rid="S1">additional file 1</xref>-Table S3). The "conservation ratio" of a motif topology or a motif mode is the ratio of the conservation fraction over the "random conservation fraction" of topology or mode, respectively. If the "random conservation fraction" was zero, we identified the "conservation ratio" as non-determined value (NDV).</p></sec></sec><sec><title>Abbreviations</title><p>PPI, Protein-Protein Interaction</p><p>GO, Gene Ontology</p></sec><sec><title>Authors' contributions</title><p>Wei-Po Lee built the entire algorithm and was charged with its implementation. Bing-Chiang Jeng annotated the protein nodes based on GO. Tun-Wen Pai analyzed the GO terms at different depths. Chin-Pei Tsai and Chang-Yung Yu were responsible for the statistical analysis. Wen-Shyong Tzou conceived the study, participated in its design and coordination and drafted 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</title><p>Table S1. Motif counts and motif modes of each motif topology. Table S2. Distribution of motif counts (logarithmic binning) for the motif modes of each motif topology. Table S3. Conservation fraction and conservation ratio of each motif topology. Table S4. Distribution of conservation fractions binned in the interval of 0.1 for the motif modes of each motif topology. Table S5. Distribution of conservation ratios (logarithmic binning) for the motif modes of each motif topology. Table S6. Average value and standard deviations of the distribution of the motif counts, conservation fraction and conservation ratios for the motif modes of each motif topology. Table S7. List of the protein-protein interactions (link) and the GO terms (molecular function) of the protein nodes.</p></caption><media xlink:href="1471-2164-7-89-S1.xls" mimetype="application" mime-subtype="vnd.ms-excel"><caption><p>Click here for file</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="S2"><caption><title>Additional File 2</title><p>Figure S1. Distribution of the joint E value for the amino acid sequence comparison between orthologs of yeast and five other species. Figure S2. Distribution of the conservation fraction of motif modes in the yeast PPI network on the basis of each motif topology. Figure S3. Distribution of the conservation ratio of the motif modes in the yeast PPI network on the basis of each motif topology.</p></caption><media xlink:href="1471-2164-7-89-S2.pdf" mimetype="application" mime-subtype="pdf"><caption><p>Click here for file</p></caption></media></supplementary-material></sec> |
The subplacenta of the red-rumped agouti <italic>(Dasyprocta leporina L)</italic> | <sec><title>Background</title><p>Hystricognath rodents have a lobed placenta, comprising labyrinthine exchange areas and interlobular trophoblast. These correspond to the labyrinthine and spongy zones of other rodent placentae. Beneath them, however, is a structure unique to hystricognath rodents called the subplacenta. We here describe the subplacenta of the red-rumped agouti and examine the possible functional correlates of this structure.</p></sec><sec sec-type="methods"><title>Methods</title><p>Placentae were collected from early in midgestation to near term of pregnancy and examined by standard histological techniques, immunohistochemistry and transmission electron microscopy. In addition, to study the microvasculature of the subplacenta, vessel casts were inspected by scanning electron microscopy</p></sec><sec><title>Results</title><p>In the subplacenta, lamellae of connective tissue support a layer of mononuclear cytotrophoblast cells. Beneath this is found syncytiotrophoblast. Clusters of multinuclear giant cells occur in the transition zone between the subplacenta and decidua. There are prominent intercellular spaces between the cytotrophoblast cells. The basal membrane of these cells is often close to fetal blood vessels. The syncytiotrophoblast surrounds an extensive system of lacunae. Microvilli project into these lacunae from the plasma membrane of the syncytiotrophoblast. The syncytial cytoplasm contains electron-dense granules. This is probably the amylase-resistant PAS-positive material identified by histochemistry. The subplacenta is supplied entirely from the fetal circulation. Within it the vessels pursue a tortuous course with sinusoidal dilatations and constrictions.</p></sec><sec><title>Conclusion</title><p>The functions that have been attributed to the subplacenta include hormone production. Our findings are consistent with this interpretation, but suggest that hormone secretion is directed towards the fetal circulation rather than the maternal tissues.</p></sec> | <contrib id="A1" contrib-type="author"><name><surname>Rodrigues</surname><given-names>Rosangela Felipe</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>rosangelafelipe@uol.com.br</email></contrib><contrib id="A2" corresp="yes" contrib-type="author"><name><surname>Carter</surname><given-names>Anthony M</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>acarter@health.sdu.dk</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Ambrosio</surname><given-names>Carlos Eduardo</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>ceambrosio@usp.br</email></contrib><contrib id="A4" contrib-type="author"><name><surname>dos Santos</surname><given-names>Tatiana Carlesso</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>tcsantos@usp.br</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Miglino</surname><given-names>Maria Angelica</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>miglino@usp.br</email></contrib> | Reproductive Biology and Endocrinology | <sec><title>Introduction</title><p>The hystricognath rodents (Suborder Hystricomorpha, Infraorder Hystricognathi [<xref ref-type="bibr" rid="B1">1</xref>]) appeared in the Eocene and underwent an extensive radiation in the Miocene. At this time they were able to capitalize upon the emergence of grasslands for which they were well adapted in a number of ways [<xref ref-type="bibr" rid="B2">2</xref>]. They differ from other rodents in giving birth to precocial young. The newborn is well developed with open eyes and a full coat of hair [<xref ref-type="bibr" rid="B3">3</xref>]. This reproductive strategy requires a lower rate of energy consumption and is well suited to an herbivorous diet [<xref ref-type="bibr" rid="B2">2</xref>]. The hystricognath placenta has a number of distinctive features [<xref ref-type="bibr" rid="B4">4</xref>,<xref ref-type="bibr" rid="B5">5</xref>]. The exchange area or labyrinth is lobulated, an adaptation that allows an increase in the total exchange area and helps to support the larger fetus at the end of gestation [<xref ref-type="bibr" rid="B6">6</xref>]. The lobules are separated by interlobular trophoblast that is the counterpart of the spongy layer found in the placenta of other rodents. Beneath this is a structure known as the subplacenta that is unique to the hystricognath rodents.</p><p>In this paper we review the structure of the subplacenta and examine possible functional correlates. The analysis is based on a description of the placenta of the red-rumped agouti (<italic>Dasyprocta leporina</italic>). This is a medium sized rodent, larger than a guinea pig and with longer legs. It is found throughout the forest, where it lives mainly on fallen fruits and nuts. Agoutis usually sit erect to eat, holding the food in their hands. They bury excess nuts and fruits for use when food is scarce. Because they bury them singly (scatter hoarding) rather than many in a cache, they are important seed dispersers for a number of tree species [<xref ref-type="bibr" rid="B7">7</xref>,<xref ref-type="bibr" rid="B8">8</xref>].</p><p>The functions that have been ascribed to the subplacenta include hormone production [<xref ref-type="bibr" rid="B9">9</xref>,<xref ref-type="bibr" rid="B10">10</xref>]. Our findings are consistent with this interpretation, but suggest that hormone secretion is directed towards the fetal circulation. Trophoblast giant cells occur close to the margin of the subplacenta and we included these in our analysis. We note that, unlike the trophoblast of the subplacenta, the giant cells often occur in close proximity to maternal blood vessels.</p></sec><sec sec-type="materials|methods"><title>Materials and methods</title><p>The study was based on 9 placentae from 6 red-rumped agoutis collected from early in midgestation to near term of pregnancy (Table <xref ref-type="table" rid="T1">1</xref>). The research was authorized by the Brazilian Institute of Environment and Renewable Natural Resources (IBAMA). The experimental protocol was approved by the bioethics committee of the School of Veterinary Medicine, University of São Paulo. The samples were collected at an agouti breeding facility at São José do Rio Preto, São Paulo. Pregnant females were submitted to hemihysterectomy. Details of anaesthesia and surgical procedures are given elsewhere [<xref ref-type="bibr" rid="B11">11</xref>].</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>Measurements of fetus, umbilical cord and placenta in the specimens of red-rumped agouti used for the present study.</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="center">Animal No.</td><td align="center">Stage of gestation</td><td align="center">Uterine horn</td><td align="center">Fetal length (cm)</td><td align="center">Fetal weight (g)</td><td align="center">Umbilical cord length (cm)</td><td align="center">Placental weight (g)</td></tr></thead><tbody><tr><td align="center">1</td><td align="center">Early to mid-gestation</td><td align="center">Left</td><td align="center">---</td><td align="center">---</td><td align="center">---</td><td align="center">15</td></tr><tr><td></td><td></td><td align="center">Left</td><td align="center">2</td><td align="center">15</td><td align="center">2</td><td align="center">13</td></tr><tr><td align="center">2</td><td align="center">Early to mid-gestation</td><td align="center">Right</td><td align="center">2</td><td align="center">15.5</td><td align="center">2</td><td align="center">12</td></tr><tr><td align="center">3</td><td align="center">Midgestation</td><td align="center">Right</td><td align="center">10</td><td align="center">42</td><td align="center">9.5</td><td align="center">37</td></tr><tr><td></td><td></td><td align="center">Left</td><td align="center">10</td><td align="center">42</td><td align="center">9</td><td align="center">36</td></tr><tr><td align="center">4</td><td align="center">Midgestation</td><td align="center">Left</td><td align="center">10</td><td align="center">40</td><td align="center">9</td><td align="center">34</td></tr><tr><td align="center">5</td><td align="center">Midgestation</td><td align="center">Right</td><td align="center">10.5</td><td align="center">41</td><td align="center">9.5</td><td align="center">36</td></tr><tr><td align="center">6</td><td align="center">Term gestation</td><td align="center">Right</td><td align="center">16</td><td align="center">100</td><td align="center">13</td><td align="center">78</td></tr><tr><td></td><td></td><td align="center">Left</td><td align="center">12.5</td><td align="center">76</td><td align="center">9</td><td align="center">50</td></tr></tbody></table></table-wrap><p>Placental fragments were fixed in 10% formalin in 0.1 M phosphate buffer and processed by standard histological procedures for embedding in paraplast, and then sectioned at 5 μm (automatic microtome, Model RM2155, Leica, Germany). Sections were stained with haematoxylin and eosin, Masson's trichrome stain or by the periodic acid Schiff (PAS) reaction with and without pretreatment with 1% amylase (Sigma, St Louis, Missouri, U.S.A.) at 37°C for 30 min. The latter sections were counterstained with haematoxylin.</p><p>Immunohistochemistry was performed for cytokeratin (to identify epithelial cells and trophoblasts) and vimentin (to identify mesenchymal cells and stromal decidua). For this purpose, 5 μm sections were transferred to poly-L-lysine coated slides. Endogenous peroxidase was blocked with 0.1% hydrogen peroxide. To improve antigen retrieval, the sections were then treated in a microwave oven in 0.1 M Tris-HCl buffer, pH 7.4. To prevent non-specific binding, slides were blocked with 2% milk powder in phosphate-buffered saline for 20 min. They were incubated overnight with primary antibody at 4°C; either a rabbit polyclonal antibody against cytokeratin (1:500; PU071-UP, Biogenex, San Ramon, California, U.S.A.) or a goat polyclonal antibody against vimentin (1:500; SC -1226, Santa Cruz Biotechnology, Santa Cruz, California, U.S.A.). Immunostaining was then performed using a secondary antibody from a kit (LSAB-HRP Peroxidase, Dako, Carpinteria, California, U.S.A.) with diaminobenzidine in Tris-HCl buffer, pH 8.2, as the chromogen. The sections were counterstained with Harris's haematoxylin.</p><p>The description of the ultrastructure of agouti subplacenta is based on material from the middle of gestation. For transmission electron microscopy, small samples were fixed in 2.5% glutaraldehyde in 0.1 M phosphate buffer, pH 7.4, washed in buffer and post-fixed in phosphate-buffered osmium tetroxide, pH 7.4 (Polysciences, Warrington, PA, USA). They were then dehydrated, washed with propylene oxide and embedded in Spurr's resin (Spurr's Kit, Electron Microscopy Sciences, Fort Washington, PA, U.S.A.). Sections were made at 60 nm and stained with 2% uranyl acetate (5 minutes) and 0.5% lead citrate (10 minutes). The ultrastructural observations were made with a transmission electron microscope (JEOL 1010, Peabody, MA, U.S.A).</p><p>To study the microvasculature of the subplacenta, an umbilical artery was injected with Mercox™ CL-2R (Okenshoji Co., Ltd, Tokyo, Japan) as previously described [<xref ref-type="bibr" rid="B11">11</xref>]. Tissues were digested by immersion of the preparation in several changes of 20% NaOH solution at 50–60°C. The casts were rinsed thoroughly in distilled water and dried in an oven at 37°C. They were then refrigerated in 20% gelatin. For scanning electron microscopy, pieces of the casts were rinsed in distilled water to remove the gelatin, dried, and mounted on stubs with conductive carbon cement (Neubauer, Münster, Germany). They were then coated with gold using a sputter coater (Model K550, Emitech Products Inc., Houston TX, USA) and examined in a scanning electron microscope (Model 435 VP, Leo Electron Microscopy Ltd, Cambridge, UK).</p><p>As a further aid to understanding vessel distribution, some placentae were injected with coloured latex (Neoprene 650, DuPont, Brazil; Latex Stain, Suvinil, Glassurit do Brazil S/A, São Bernardo do Campo, S.P., Brazil). Different colours were injected in a uterine vein, a uterine artery, and the umbilical vein. The placentae were fixed in 10% formalin in 0.1 M phosphate buffer.</p></sec><sec><title>Results</title><sec><title>General structure</title><p>The placenta is found on the mesometrial side of the uterus. In transverse section the main placenta and the subplacenta can be distinguished by their colour and form (Figure <xref ref-type="fig" rid="F1">1A</xref>). The placenta is spherical and reddish and on closer inspection is seen to be divided into lobes separated by interlobar trophoblast (Figure <xref ref-type="fig" rid="F1">1C</xref>). The subplacenta is whitish and formed like a chalice (Figure <xref ref-type="fig" rid="F1">1A</xref>). Beneath it is maternal decidua. Giant cells occur in the transition zone between the subplacenta and decidua. Maternal tissue forms a capsule that encloses the placenta (Figure <xref ref-type="fig" rid="F1">1C</xref>).</p><fig position="float" id="F1"><label>Figure 1</label><caption><p>Gross structure of the agouti placenta. (A) Transverse section of the placenta in midpregnancy showing the labyrinth (lab) of the main placenta and the subplacenta (sub pl). Fetal vessels run from the umbilical cord to the subplacenta through the central excavation (arrow). Scale bar: 1 cm. (B) Placenta injected with Neoprene latex to show the vascularization of the subplacenta. Fetal arteries are in white, while maternal arteries in the decidua (arrow) are in red; maternal venous channels were filled with green latex. (C) Schematic to show location of the subplacenta in relation to the yolk sac placenta, main placenta and decidua. The labyrinth occurs in lobes separated by interlobar trophoblast. In addition to the decidua beneath the subplacenta, maternal tissue forms a capsule that houses the placenta. The trajectory of the fetal vessels is also indicated (arrow).</p></caption><graphic xlink:href="1477-7827-4-31-1"/></fig><p>In midgestation the subplacenta is supplied entirely by fetal blood vessels. Thus a single large artery runs through the central excavation from the umbilical cord to the subplacenta (Figure <xref ref-type="fig" rid="F1">1A, C</xref>). When Neoprene latex is injected into the umbilical artery, it fills the vessels of the subplacenta (white colour in Figure <xref ref-type="fig" rid="F1">1B</xref>). In contrast, latex injected through the uterine arteries does not reach the subplacenta, although it fills large maternal vessels peripheral to it (red colour in Figure <xref ref-type="fig" rid="F1">1B</xref>).</p></sec><sec><title>Histology and immunohistochemistry</title><p>The subplacenta is separated from the main placenta by a layer of connective tissue (fetal mesenchyme), and lamellae of connective tissue support the trophoblast. This is clear early in gestation when there is a relatively open structure with plenty of connective tissue (Figure <xref ref-type="fig" rid="F2">2A</xref>). Later the structure becomes more compact and lobular with thinning of the connective tissue (Figure <xref ref-type="fig" rid="F2">2B</xref>). The subplacenta does persist until the end of gestation, but extensive degenerative changes occur towards term (not shown).</p><fig position="float" id="F2"><label>Figure 2</label><caption><p>Subplacenta of the agouti. (A) Early in gestation. Note the lamellae of fetal mesenchyme (mes) and the relationship of the subplacenta to the decidua (dec). (B) Later in gestation. The lamellae of fetal mesenchyme are much thinner; lab, labyrinth. Haematoxylin and eosin. Scale bars: 0.5 mm.</p></caption><graphic xlink:href="1477-7827-4-31-2"/></fig><p>As might be expected the trophoblast can be immunostained for cytokeratin (Figure <xref ref-type="fig" rid="F3">3A</xref>). The connective tissue carries the fetal blood vessels. The latter are cytokeratin negative (Figure <xref ref-type="fig" rid="F3">3A</xref>), but they can be immunostained for vimentin (Figure <xref ref-type="fig" rid="F3">3B</xref>). The connective tissue lamellae are bordered by a layer of mononuclear cytotrophoblast with clearly marked cell boundaries (Figure <xref ref-type="fig" rid="F4">4</xref>). Beneath this is syncytiotrophoblast with a completely different morphology: multinucleate and without cell boundaries. The cytotrophoblasts have basophilic cytoplasm, round nuclei and rest on a basal membrane. The cytoplasm of the syncytiotrophoblast is eosinophilic, with basophilic granules, and the nuclei are irregular.</p><fig position="float" id="F3"><label>Figure 3</label><caption><p>Immunohistochemistry and PAS reaction of agouti subplacenta in midgestation. (A) Cytotrophoblast (cyt tr) and syncytiotrophoblast (syn tr) immunostain for cytokeratin. Fetal mesenchyme (mes) and wall of blood vessel (bv) are cytokeratin negative. (B) Fetal mesenchyme and wall of blood vessel immunostain for vimentin. Trophoblast is vimentin negative. (C) PAS reaction in the absence of amylase is strong in syncytiotrophoblast and negative in cytotrophoblasts (although their basal lamina gives a positive reaction). (D) Following treatment with amylase there is only a moderate decrease in the positive reaction of syncytiotrophoblast. Scale bars: 40 μm.</p></caption><graphic xlink:href="1477-7827-4-31-3"/></fig><fig position="float" id="F4"><label>Figure 4</label><caption><p>Organization of the agouti subplacenta. The cytotrophoblast (cyt tr) abuts the mesenchyme and fetal capillaries (cap). Behind it is found syncytiotrophoblast (syn tr). (A) Early in gestation. (B) Later in gestation. Semithin sections, toluidine blue. Scale bars: 20 μm.</p></caption><graphic xlink:href="1477-7827-4-31-4"/></fig><p>The syncytiotrophoblast is PAS-positive and PAS-positive granules are present in the cytoplasm (Figure <xref ref-type="fig" rid="F3">3C</xref>). The PAS reaction persists after amylase treatment (Figure <xref ref-type="fig" rid="F3">3D</xref>). The most intense staining is likely related to the lacunae that occur in the syncytiotrophoblast, as described below.</p><p>With the progress of gestation some initial characteristics of the subplacenta change. Thus there is a reduction in the number of the layers of cytotrophoblast, which initially is multi-layered (Figure <xref ref-type="fig" rid="F4">4A</xref>) and later forms a single layer in most places (Figure <xref ref-type="fig" rid="F4">4B</xref>). In the syncytiotrophoblast, the advance of gestation is marked by the appearance of extensive lacunae. Towards term, as the subplacenta starts to degenerate, one sees vacuolization of the syncytiotrophoblast.</p><p>Clusters of multinucleated giant cells are found near the margin of the subplacenta in the transition zone between it and the decidua (Figure <xref ref-type="fig" rid="F5">5A</xref>). They often occur in proximity to maternal blood vessels. The giant cells have clear and irregular cytoplasm with round nuclei and are surrounded by eosinophilic and PAS-positive extracellular material. They can be immunostained for cytokeratin (Figure <xref ref-type="fig" rid="F5">5B</xref>).</p><fig position="float" id="F5"><label>Figure 5</label><caption><p>Trophoblast giant cells of agouti placenta in midgestation. (A) Clusters of multinucleated giant cells (GC) are found in the transition zone between the subplacenta (sub pl) and decidua (dec). (B) The giant cells can be immunostained for cytokeratin. Scale bars: 50 μm (A); 40 μm (B).</p></caption><graphic xlink:href="1477-7827-4-31-5"/></fig></sec><sec><title>Ultrastructure</title><p>The nuclei of the cytotrophoblast cells are large in relation to the amount of cytoplasm and have euchromatin with an evident nucleolus (Figure <xref ref-type="fig" rid="F6">6A</xref>). The basal membrane of the cytotrophoblast layer is in contact with the connective tissue lamellae and often close to the fetal vessels (Figure <xref ref-type="fig" rid="F6">6A</xref>). The lateral membrane shows microvilli in some places. There are often large intercellular spaces (Figure <xref ref-type="fig" rid="F6">6B</xref>). The cytoplasm contains rounded mitochondria, rough endoplasmic reticulum and small electron dense inclusions (Figure <xref ref-type="fig" rid="F6">6C</xref>). Desmosomes are seen between cytotrophoblast cells as well as between cytotrophoblast and the underlying syncytiotrophoblast.</p><fig position="float" id="F6"><label>Figure 6</label><caption><p>Ultrastructure of cytotrophoblast cells of the agouti subplacenta in midgestation. (A) Cytotrophoblasts (cyt tr) rest on a basal lamina (bl) that is closely apposed to the endothelium of a fetal capillary. On the other side they are in contact with the syncytiotrophoblast (syn tr). (B) Large intercellular spaces (ics) occur between the cytotrophoblasts. (C) Detail showing abundant rough endoplasmic reticulum and numerous electron dense inclusions. Scale bars: 2 μm (A); 3 μm (B); (C) 200 nm.</p></caption><graphic xlink:href="1477-7827-4-31-6"/></fig><p>The syncytiotrophoblast encloses many lacunae into which microvilli project from the plasma membrane (Figure <xref ref-type="fig" rid="F7">7A</xref>). As gestation advances, these lacunae come to form an extensive system. The syncytiotrophoblast has many irregular nuclei with euchromatin and evident nucleoli (Figure <xref ref-type="fig" rid="F7">7B</xref>). The cytoplasm contains few mitochondria, some rough endoplasmic reticulum, and an occasional Golgi apparatus. Throughout the cytoplasm it is possible to identify electron dense granules.</p><fig position="float" id="F7"><label>Figure 7</label><caption><p>Ultrastructure of syncytiotrophoblast of agouti subplacenta in midgestation. (A) Syncytiotrophoblast sends microvillous projections into an extensive system of lacunae (lac). (B) Detail showing irregularly shaped nucleus, rough endoplasmic reticulum (rer) and vesicles near the plasma membrane. Scale bars: 5 μm (A); 1 μm (B).</p></caption><graphic xlink:href="1477-7827-4-31-7"/></fig><p>The giant cells have irregular nuclei and their cytoplasm appears vacuolated (Figure <xref ref-type="fig" rid="F8">8A</xref>). They contain few mitochondria and the rough endoplasmic reticulum is rather sparse, but there are many granules of moderate electron density. Microvilli extend from the cell surface into the surrounding matrix. Vesicles seen within this extracellular material seem to have been extruded from the cell (Figure <xref ref-type="fig" rid="F8">8B</xref>).</p><fig position="float" id="F8"><label>Figure 8</label><caption><p>Ultrastructure of a placental giant cell in the agouti in midgestation. (A) The nucleus is irregular and the cytoplasm appears vacuolated. There are few mitochondria, some rough endoplasmic reticulum (rer) and many granules (gr) of moderate electron density. Microvilli extend from the cell surface. (B) At higher power vesicles (arrows) can be seen within the extracellular material. They seem to have been extruded from the cell. Scale bars: 4 μm (A); 1 μm (B).</p></caption><graphic xlink:href="1477-7827-4-31-8"/></fig></sec><sec><title>Microvasculature</title><p>In mid to late gestation, the subplacenta is supplied exclusively by fetal vessels (Figure <xref ref-type="fig" rid="F1">1B</xref>). A large branch of the umbilical artery follows the central band of fetal mesenchyme to the base of the main placenta and then branches to supply the subplacenta (Figure <xref ref-type="fig" rid="F9">9A</xref>). The subplacental vessels pursue a tortuous course with sinusoidal dilatations and constrictions (Figure <xref ref-type="fig" rid="F9">9B</xref>).</p><fig position="float" id="F9"><label>Figure 9</label><caption><p>Microvasculature of agouti subplacenta in midgestation. (A) The large vessel is a branch of the umbilical artery. (B) A skein of vessel branches. Note the sinusoidal dilatations and constrictions of the capillaries. Scale bars: 100 μm.</p></caption><graphic xlink:href="1477-7827-4-31-9"/></fig></sec></sec><sec><title>Discussion</title><p>The subplacenta is a unique structure that distinguishes hystricognath rodents from all other mammals [<xref ref-type="bibr" rid="B4">4</xref>,<xref ref-type="bibr" rid="B5">5</xref>]. The basic arrangement of cytotrophoblast and syncytiotrophoblast, supported by lamellae of connective tissue, is well conserved. The ultrastructure of these elements is also similar across species, including the guinea pig [<xref ref-type="bibr" rid="B9">9</xref>,<xref ref-type="bibr" rid="B10">10</xref>,<xref ref-type="bibr" rid="B12">12</xref>], chinchilla [<xref ref-type="bibr" rid="B13">13</xref>], cane rat [<xref ref-type="bibr" rid="B14">14</xref>], paca [<xref ref-type="bibr" rid="B15">15</xref>], rock cavy [<xref ref-type="bibr" rid="B16">16</xref>] and agouti (this study). The function of the subplacenta is poorly understood.</p><p>In the guinea pig, fetal trophoblast invades the walls of the uterine arteries and this is associated with extensive remodelling of the vessel wall [<xref ref-type="bibr" rid="B17">17</xref>,<xref ref-type="bibr" rid="B18">18</xref>]. Recently, it was proposed that the source of the invasive trophoblast was the cytotrophoblast layer of the subplacenta [<xref ref-type="bibr" rid="B19">19</xref>]. There is experimental evidence from the guinea pig and degu to support this hypothesis [<xref ref-type="bibr" rid="B20">20</xref>,<xref ref-type="bibr" rid="B21">21</xref>]. It does not, however, address the function of the syncytiotrophoblast. An earlier proposal that the subplacenta might play a role in maternal to fetal transfer of high molecular weight molecules [<xref ref-type="bibr" rid="B22">22</xref>] receives no support from studies of its ultrastructure [<xref ref-type="bibr" rid="B12">12</xref>,<xref ref-type="bibr" rid="B13">13</xref>].</p><p>The placenta is a source of steroid hormones. In the guinea pig ovarian progesterone is not required after day 20 of gestation, implying that placental progesterone is adequate for pregnancy maintenance after this time [<xref ref-type="bibr" rid="B23">23</xref>]. It has been shown for a variety of hystricognath rodents that the hormone circulates bound to progesterone-binding protein [<xref ref-type="bibr" rid="B24">24</xref>]. The principal site of placental progesterone synthesis is the interlobar syncytiotrophoblast [<xref ref-type="bibr" rid="B25">25</xref>]. This is supported by the presence there of smooth endoplasmic reticulum and mitochondria containing tubular cristae [<xref ref-type="bibr" rid="B26">26</xref>]. In contrast Wolfer and Kaufmann [<xref ref-type="bibr" rid="B12">12</xref>] argued against a role for the subplacenta in steroid synthesis. Attention has therefore been focussed on the subplacenta as a source of glycoprotein.</p><p>Davies et al. [<xref ref-type="bibr" rid="B9">9</xref>,<xref ref-type="bibr" rid="B10">10</xref>] remarked upon the strongly positive PAS reaction of the guinea pig subplacenta. This is in part due to the presence of glycogen, but there is a diastase-resistant component, as demonstrated here for the agouti. It was suggested that this was glycoprotein and that it was secreted into the lacunae [<xref ref-type="bibr" rid="B9">9</xref>,<xref ref-type="bibr" rid="B10">10</xref>]. Heap and Illingworth [<xref ref-type="bibr" rid="B27">27</xref>] proposed that it might be the progesterone-binding protein, as this is a carbohydrate-rich transport protein. The association of the subplacenta with a transport protein likewise unique to hystricognath rodents constituted an attractive hypothesis. However, the cellular location of progesterone-binding protein was later shown to be the interlobar and marginal trophoblast of the main placenta [<xref ref-type="bibr" rid="B28">28</xref>].</p><p>Davies et al. [<xref ref-type="bibr" rid="B9">9</xref>,<xref ref-type="bibr" rid="B10">10</xref>] had implied that the glycoprotein might be chorionic gonadotropin (CG), but the evidence for this is equivocal. Indeed, recent work suggests that CG is restricted to primates and equids [<xref ref-type="bibr" rid="B29">29</xref>,<xref ref-type="bibr" rid="B30">30</xref>]. There is some evidence for production of luteinising hormone, possibly CG, in guinea pig placenta [<xref ref-type="bibr" rid="B31">31</xref>], but no detectable signal for the corresponding messenger RNA was seen in Northern blot assays of placental RNA [<xref ref-type="bibr" rid="B32">32</xref>]. In primates, CG is secreted directly into the intervillous space from the maternal-facing syncytiotrophoblast, whilst in equids CG is secreted to maternal blood from the endometrial cup cells. Thus, if there is a guinea pig CG, it is more likely to be produced by the syncytiotrophoblast of the main placenta than the subplacenta as it can then be secreted to the maternal blood channels. As we have shown here, the blood supply to the subplacenta is derived from the fetal circulation. A longitudinal study in the guinea pig showed that maternal vessels were present only early in development and had disappeared by day 20 of gestation [<xref ref-type="bibr" rid="B12">12</xref>].</p><p>The subplacental syncytiotrophoblast contains electron dense inclusions that likely are secretion granules as argued by Wolfer and Kaufmann [<xref ref-type="bibr" rid="B12">12</xref>]. Their product is released into the elaborate network of lacunae. These do not have access to the maternal circulation, but King and Tibbitts [<xref ref-type="bibr" rid="B13">13</xref>] suggested that products secreted to the extracellular spaces might be able to reach the fetal capillaries. As we show here, the fetal vessels within the subplacenta pursue a tortuous course with dilatations and constrictions as in an endocrine gland.</p><p>Secretion of most hormones produced by the human placenta is unidirectional: they are released to the maternal circulation and affect maternal physiology. However, some placental products do reach the fetal circulation. In sheep they include prostaglandin E<sub>2 </sub>[<xref ref-type="bibr" rid="B33">33</xref>], adenosine [<xref ref-type="bibr" rid="B34">34</xref>] and steroids [<xref ref-type="bibr" rid="B35">35</xref>]. As an example, fetal breathing movements occur throughout most of gestation, but are influenced by prostaglandin E<sub>2 </sub>from the placenta, which tends to suppress them. The removal of the placenta and the placental prostaglandin E<sub>2 </sub>is critical for the initiation of continuous breathing at birth [<xref ref-type="bibr" rid="B33">33</xref>]. Thus, if the subplacenta is an endocrine organ, the function of its hormones may be to influence fetal physiology.</p><p>The origin of the trophoblastic giant cells was explored by Mossman [<xref ref-type="bibr" rid="B36">36</xref>]. They tend to occur in clusters surrounded by amorphous extracellular material, as described here for the agouti and elsewhere for the chinchilla [<xref ref-type="bibr" rid="B13">13</xref>]. Electron dense granules were seen both within the cells and in the extracellular matrix. This is consistent with a secretory function as suggested in a study of chinchilla giant cells by Tibbitts and Birge [<xref ref-type="bibr" rid="B37">37</xref>]. The giant cell clusters sometimes occur close to maternal vessels. Thus, in contrast to those of the subplacenta, their secretory products may have access to the maternal circulation. On the other hand, many giant cells are vacuolated; this has been considered to be a degenerative change [<xref ref-type="bibr" rid="B37">37</xref>].</p><p>In conclusion, the function of the subplacenta remains elusive. The syncytiotrophoblast secretes PAS-positive material, probably glycoprotein, to an extensive system of lacunae. These lacunae seem to be connected to the intercellular spaces of the cytotrophoblast layer, which lie directly beneath fetal capillaries. Indeed, the subplacenta is plentifully supplied with blood vessels from the umbilical circulation, but lacks direct access to the maternal circulation. Perhaps the function of the subplacenta is to secrete growth factors, hormones or cytokines to the fetal circulation. As in the sheep, they might function to shut down physiological functions that are not needed in fetal life. Placental signals disappear once the cord is severed, allowing immediate resumption of suppressed functions in the neonate. Like sheep, hystricognath rodents give birth to precocial young and we suggest that placental control of fetal function is an important feature of this reproductive strategy.</p></sec> |
Causal thinking and causal language in epidemiology: a cause by any other name is still a cause: <italic>response to Lipton and Ødegaard</italic> | Could not extract abstract | <contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Tam</surname><given-names>Clarence C</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>clarence.tam@lshtm.ac.uk</email></contrib> | Epidemiologic Perspectives & Innovations | <sec><title/><p>I have great sympathy with the thoughts of Lipton and Ødegaard [<xref ref-type="bibr" rid="B1">1</xref>] – the assessment and communication of "causal" associations is a source of continual frustration for epidemiologists. The authors' lucid account of the use of causal language in epidemiology can essentially (if rather unflatteringly) be simplified to the following: it is impossible to prove that <italic>X </italic>causes <italic>Y</italic>; the statement <italic>"Smoking causes lung cancer" </italic>is thus no more informative than the statement <italic>"Smoking two packs a day for </italic>N <italic>years increases your risk of lung cancer ten-fold"</italic>. In fact, it is less informative and even misleading. The authors argue that such causal statements are redundant, logically indefensible and should be avoided in favour of more detailed descriptions of the process by which such associations are established (the "story", as the authors put it). The latter are, in themselves, sufficient causal statements (the notion of "letting the data speak for themselves") and nothing is gained by making subjective attributions of causality.</p><p>It is, of course, difficult to argue with such reasoning. Sensible epidemiologists would shy away from stating that a particular <italic>X </italic>causes a particular <italic>Y</italic>, because they know that, in purely statistical terms, there is always a possibility that they could be wrong. There is, in fact, much with which I agree in Lipton and Ødegaard's paper. To their call for "telling a good story", which they argue lies "not in the naming of something as causal, but in the actual rigor of the analysis" [<xref ref-type="bibr" rid="B1">1</xref>] (p8), I would add that it is not just rigour in the analysis that is required, but also rigour in asking good questions involving testable hypotheses, rigour in study design and execution, and rigour in the description of results. There are, however, notable points on which I disagree. The authors begin with a short caveat, claiming not to attempt "to revisit the long-standing debate between realism and pragmatism in science". I find this somewhat strange as, to me, this issue is at the crux of the argument. The common description of epidemiology as a "pragmatic" science is rather unfortunate. The term "pragmatic" suggests some form of compromise between objective "realism" and the vicissitudes of everyday life (see Appendix Footnote 1). This is somewhat ironic; I see no compromise in epidemiology as a field that applies scientific methods, however imperfect, to study everyday problems with the ultimate aim of improving the health of individuals and populations. Acceptance of its imperfections is, on the other hand, a very "realist" attitude. In a very "pragmatic" sense then, the epidemiologist can say that, having assessed the available evidence, smoking is a cause of lung cancer. The philosopher, however (and I say this meaning no disrespect to the field), can say nothing about any causal link between smoking and lung cancer – either in the "pragmatic" or objective senses – other than the fact that they can say nothing about any causal link between smoking and lung cancer (and, indeed, generally). Lipton and Ødegaard do seem to accept this "pragmatic" role of epidemiology, as they assert that they are interested in "the ability to manipulate the world, to predict and intervene" [<xref ref-type="bibr" rid="B1">1</xref>] (p3), but do not feel that this is relevant to their discussion of causation in epidemiology.</p><p>My main difficulty with Lipton and Ødegaard's position perhaps stems from my love of the written language. The ability to string words together on a piece of paper to inspire a sense of knowledge, excitement, wonder, sadness or depression in someone perhaps sitting thousands of miles away whom you may never have met is without doubt one of the most honourable and worthwhile of human activities. A whisper in someone's ear, a casual conversation, a speech at a convention are all eventually lost to dispersing airwaves or adulterated by time and memory. A well-constructed written sentence is, at least in principle, timeless.</p><p>It is from this rather romantic perspective that I find myself pained by the awkwardness with which epidemiologists must describe associations between exposures and health outcomes. The term "cause" is steeped in great history and philosophy; it is used to describe an objective ideal by which one event invariably leads to another, and as such is often thought to have limited use in our everyday world.</p><p>Imagine the Aristotelian notions of an "ideal" universe, made up of perfect geometrical bodies, and the observable "real" universe, an imperfect approximation of the ideal in which, for example, spherical bodies appear to be spheres, but upon closer examination are shown, by nature of their imperfection, to be made up of flat surfaces. This (admittedly somewhat naïve) scenario provides us with a visual analogy of how epidemiologists might view the world. There is, we suspect, an ideal, true, but unobservable "causal" association between <italic>X </italic>and <italic>Y</italic>. As with other scientific fields, we aim to model this ideal universe using our imperfect methods and, from our ensuing observations, make some sort of inference about the true causal association. The question then becomes how closely our imperfect models resemble the ideal universe they are intended to reflect – if you like, how small the area of those flat surfaces that make up those spherical bodies is. The smaller the surfaces, the smaller the error in our observations, and the closer our imperfect geometrical representations will resemble perfect spherical bodies.</p><p>Lipton and Ødegaard argue that the concept of causation in the ideal universe, by nature of its being unobservable, is irrelevant. One might even wonder whether causation even exists at all in the ideal universe; maybe things just "happen". Consider two identical twins who share in common not just their genetic make-up, but everything else in their lives. They live in the same house, eat exactly the same foods, have the same jobs, think the same thoughts at the same time; in fact, at any given point in time, the two are completely interchangeable. This is perhaps the closest we could ever get to a real interpretation of our ideal universe. Suppose both twins take up smoking two packs of cigarettes a day at the age of 20, which they smoke at exactly the same times. Both then go on to develop lung cancer at the age of 50. The fact that they have been smoking two packs of cigarettes every day for the past 30 years would give us no information as to whether their lung cancer was caused by smoking, in much the same way as their age, sex, occupation, genetic make-up or any of the other factors that the twins share in common (which is everything) would tell us nothing about the cause of their lung cancer. Suppose, however, that only one of the twins had taken up smoking. Only the smoker goes on to develop lung cancer at the age of 50 (I ignore the possible effects of passive smoking here). I imagine that most people would be prepared to bet good money on smoking being the cause of the smoker's cancer in this case, since it is the only factor that distinguishes the twins. But what if both twins had taken up smoking and only one subsequently developed lung cancer? Again, we would be in a quagmire as to what to say about smoking as a cause of lung cancer. We might venture to say that, all else being equal, smoking induces cancer with a certain probability. Of course, in the ideal universe nothing is ever equal, for that would require two instances of the same type of event occurring in the same place at the same time, and all the preceding events for the two events to have occurred in exactly the same point in space-time, apart from the one that was subsequently termed the "cause". Asserting that this one cause effects the outcome with a certain probability would be tantamount to saying that God does indeed play dice, a question perhaps best left to quantum physicists. I imagine, however, that most epidemiologists would not be prepared to accept such an interpretation. After all, our observations tell us that things do not occur at random; health outcomes cluster among individuals with certain characteristics, and it seems unreasonable to suggest that if smoking were not a true cause of lung cancer, we should repeatedly observe this association simply by chance. Our idea of causation is based on a belief that we can assign individuals into groups with certain common, relevant characteristics, disregarding factors that we judge to be uninformative about a given association and, therefore, ignorable. Otherwise, the notion of causation does not make sense. That we, as epidemiologists, have this belief attests to its being relevant and important, regardless of whether it is unobservable.</p><p>It is here that I find a frank contradiction in Lipton and Ødegaard's argument. They agree that the concept of causation is important in epidemiology, and even agree that terms commonly used to imply causal relationships, such as "X <italic>increases the risk of </italic>Y <italic>by</italic>" can be used in much the same sense as "X <italic>causes </italic>Y" [<xref ref-type="bibr" rid="B1">1</xref>] (p5). The authors do not seem to be objecting to the use of the term "cause" on the grounds it describes something that is qualitatively different. It is thus hard to see what they find so disagreeable about its use. Perhaps they simply do not like the word "cause", preferring instead more descriptive associational statements that they nevertheless seem to agree amount to qualitatively the same thing.</p><p>Lipton and Ødegaard's assertion that statements of association between exposures and outcomes are in themselves sufficient causal statements is based on the strong assumption that, for any two given associational statements, the degree of evidence is equivalent – that the flat surfaces making up those imperfect spheres are of equal area. I doubt, however, whether most epidemiologists would lend equal weight to any pair of causal statements, such as <italic>"Smoking </italic>X <italic>packs a day for </italic>N <italic>years increases your risk of lung cancer by </italic>Y <italic>times" </italic>and <italic>"Living within </italic>X <italic>miles of a mobile phone pylon for </italic>N <italic>years increases your risk of brain cancer by </italic>Y <italic>times"</italic>, for any given range of <italic>X </italic>and <italic>Y </italic>values. Lipton and Ødegaard thus argue for better stories, better accounts of study designs, data collection and analysis and more detailed discussion of confounding and potential biases. This is, of course, to be welcomed, as any such increase in rigour can only serve to improve accuracy and precision in describing our observations (see Appendix Footnote 2). However, this still leaves us with the problem of which of these observations are actually true (in the objective, ideal sense). At best, these discussions involve a considerable degree of subjectivity, and decisions regarding which associations are really believable and warrant some form of intervention are reached in a manner resembling an informal and rather undemocratic consensus. Even conceptually more appealing approaches such as multiple bias modelling rely on some consensus about the prior distributions assumed for bias corrections. Thus, when asked whether smoking causes lung cancer, an honest epidemiologist is left merely with the following as a viable response: <italic>"I can tell you that smoking two packs a day for </italic>N <italic>years increases your risk of lung cancer by 10 times"</italic>. In fact, this is clearly an oversimplification, and they would actually be better off saying: <italic>"I can tell you with 95 percent certainty that smoking two packs a day for </italic>N <italic>years increases your risk of lung cancer by between </italic>A <italic>and </italic>B <italic>times"</italic>. But even this is not satisfactory, and one of these two would be preferable: <italic>"I can tell you with 95 percent certainty that smoking two packs a day for </italic>N <italic>years increases your risk of lung cancer by between </italic>A <italic>and </italic>B <italic>times, assuming that there is no systematic error in my observations" </italic>or <italic>"I can tell you with 95 percent certainty that smoking two packs a day for </italic>N <italic>years increases your risk of lung cancer by between </italic>A <italic>and </italic>B <italic>times, and I have tried to correct for biases </italic>C, D <italic>and </italic>E <italic>using prior distributions </italic>S, T <italic>and </italic>U, <italic>which I believe (though I cannot be certain) are rational and exhaustive"</italic>. By this time, the enquirer will probably have regretted asking the question in the first place and, while puffing away at their cigarette, poured themselves a stiff whisky too.</p><p>The problem is that to describe such processes accurately requires the use of very unwieldy language that is not only linguistically unappealing, but also unfamiliar to most of those who would have an interest in whether smoking causes lung cancer. This is well known to any student of introductory statistics who has grappled with statements such as <italic>"There is insufficient evidence to reject the null hypothesis"</italic>. All the goodwill in the world and desire for transparency will not change this fact. It is rather ironic that the field that is best placed to identify adverse effects on people's health is probably the worst-suited for communicating them. This is hardly surprising, for the perspective from which we deliver such statements differs markedly from the perspective of those receiving them. Thus, when the honest epidemiologist asserts that <italic>"I can tell you that smoking two packs a day for </italic>N <italic>years increases your risk of lung cancer by 10 times"</italic>, they are reporting an association obtained within a very specific context that has to do with the way in which their study was designed, the individuals selected to participate in the study, how data were collected and analyzed, issues of bias and confounding, and our knowledge of statistical concepts. When confronted with such a statement, however, an individual might wish to raise some more pointed questions, such as, for example, <italic>"How likely is it that this statement is true for <underline>me</underline>"</italic>. The honest epidemiologist now has no recourse, for this is not a question that they can answer. We can, of course, say that for some individuals the probability may be exactly 0 and that for others it may be exactly 1 (see Appendix Footnote 3). But here is the crux: while Lipton and Ødegaard favour such statements because they are more accurate, this is only the case within a very specific context, that of epidemiologists conducting the science of epidemiology. To the end user of such information, however, it is this sort of statement that can be misleading, because there is no way in which we can say that it is true for any individual, at least not without calling on some abstract idea of an "average" individual. Yet while epidemiologists like to elevate the term "cause" to a philosophical and mythical realm, as used in everyday life, it has many appealing linguistic qualities. If I were to walk ten minutes away from my office (to get away from the concentration of epidemiologists), randomly stop people on the street and ask them if they think smoking "causes" lung cancer, I expect that most of them would say yes. I also expect that most would interpret the term "cause" in a probabilistic manner, recognizing that not everybody who smokes develops lung cancer, and that not everybody who develops lung cancer is a smoker. I imagine that some will even recognize that among smokers who develop lung cancer, it is not necessarily true to say that it was smoking that led to their illness. The common use of the word "cause" thus encapsulates all the oft-recited qualities of partiality, necessity and sufficiency, while imparting at least some qualitative notions of statistical uncertainty.</p><p>By maintaining abstract ideals of objective and unattainable causes, we do a great disservice to our field and those whose interests we aim to protect. The idea of the honest (if not necessarily objective) epidemiologist entrusting others with their carefully observed and qualified associational statements is, to my mind, not entirely satisfactory, because those who can most benefit from that information are not necessarily well equipped to interpret it. The statement <italic>"</italic>X <italic>causes </italic>Y" imparts a sense of conviction – that considering all the available observations, with all their qualifications, our most reasonable interpretation is that, in a real if not necessarily ideal sense, <italic>X </italic>does indeed cause <italic>Y</italic>, to an extent that something should be done about it. To shy behind associational statements, though accurate and transparent, is to shun this sense of social responsibility – that we, as gatherers, processers and interpreters of data, should be compelled to act based upon our observations, and not merely leave them to the interpretations of others whose judgement is perhaps no better than our own. The problem still remains of when and by what criteria we consider evidence to be sufficient to warrant action, and this will be an on-going debate, one that will additionally involve politics, economics and social values, and in which epidemiologists should undoubtedly be increasingly involved. Whether we call them "causes" or "roses" is, to a large extent, a moot point. There is, however, no word in the English language, or in any of the languages with which I am familiar, to describe an association for which there is sufficient evidence to warrant some form of intervention. Perhaps epidemiologists should invent one. For those bourgeoning students of introductory statistics, however, it will be depressing to hear that we can never really know that we are right in ascribing causal associations. Despite our best intentions and qualifications (and at the risk of sounding like a refutationist), at best all we can say is that so far we haven't been not right. But in the greater scheme of things, epidemiologists can only ever be wrong.</p></sec><sec><title>Competing interests</title><p>I declare that I have no competing interests (other than being an epidemiologist).</p></sec><sec><title>Appendix</title><sec><title>Footnote 1</title><p>Lipton and Ødegaard use the term "real" to describe metaphysical, objective truth. I use the term here throughout the text to describe our everyday world, referring to objective truths as "ideals".</p></sec><sec><title>Footnote 2</title><p>I refer to "accuracy and precision in describing our observations" in the sense of our statements about associations being more accurate and precise descriptions of our observations, rather than in the sense of them being necessarily true (in the objective, ideal sense).</p></sec><sec><title>Footnote 3</title><p>So as not to exclude the stochasticists among us, I should also add here that for some individuals, perhaps even for most individuals, the probability might be somewhere between zero and one.</p></sec></sec> |
Maximum static inspiratory and expiratory pressures with different lung volumes | <sec><title>Background</title><p>Maximum pressures developed by the respiratory muscles can indicate the health of the respiratory system, help to determine maximum respiratory flow rates, and contribute to respiratory power development. Past measurements of maximum pressures have been found to be inadequate for inclusion in some exercise models involving respiration.</p></sec><sec sec-type="methods"><title>Methods</title><p>Maximum inspiratory and expiratory airway pressures were measured over a range of lung volumes in 29 female and 19 male adults. A commercial bell spirometry system was programmed to occlude airflow at nine target lung volumes ranging from 10% to 90% of vital capacity.</p></sec><sec><title>Results</title><p>In women, maximum expiratory pressure increased with volume from 39 to 61 cmH<sub>2</sub>O and maximum inspiratory pressure decreased with volume from 66 to 28 cmH<sub>2</sub>O. In men, maximum expiratory pressure increased with volume from 63 to 97 cmH<sub>2</sub>O and maximum inspiratory pressure decreased with volume from 97 to 39 cmH<sub>2</sub>O. Equations describing pressures for both sexes are:</p><p>P<sub>e</sub>/P<sub>max </sub>= 0.1426 Ln( %VC) + 0.3402 R<sup>2 </sup>= 0.95</p><p>P<sub>i</sub>/P<sub>max </sub>= 0.234 Ln(100 - %VC) - 0.0828 R<sup>2 = </sup>0.96</p></sec><sec><title>Conclusion</title><p>These results were found to be consistent with values and trends obtained by other authors. Regression equations may be suitable for respiratory mechanics models.</p></sec> | <contrib id="A1" contrib-type="author"><name><surname>Lausted</surname><given-names>Christopher G</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>clausted@systemsbiology.org</email></contrib><contrib id="A2" corresp="yes" contrib-type="author"><name><surname>Johnson</surname><given-names>Arthur T</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>artjohns@umd.edu</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Scott</surname><given-names>William H</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>enduranc@umd.edu</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Johnson</surname><given-names>Monique M</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>monique.marquardt@med.ge.com</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Coyne</surname><given-names>Karen M</given-names></name><xref ref-type="aff" rid="I4">4</xref><email>karen.coyne@us.army.mil</email></contrib><contrib id="A6" contrib-type="author"><name><surname>Coursey</surname><given-names>Derya C</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>derya@mail.umd.edu</email></contrib> | BioMedical Engineering OnLine | <sec><title>Background</title><p>While maximum respiratory pressures at the mouth have been measured in numerous subjects, less data exists to characterize maximum pressures as they vary with lung volume. Maximum pressure is volume dependent because muscle tension is length dependent, because muscle tension produces higher pressure with a smaller radius of curvature, and because respiratory tissue is elastic. Rahn et al. [<xref ref-type="bibr" rid="B1">1</xref>] first produced static pressure-volume diagrams from a group of adult men, and later, Cook et al. [<xref ref-type="bibr" rid="B2">2</xref>] produced pressure-volume diagrams from a larger group of subjects including women and children. These diagrams were useful in modeling the energetics of respiration [<xref ref-type="bibr" rid="B3">3</xref>] and in monitoring the progress of respiratory muscle training [<xref ref-type="bibr" rid="B4">4</xref>]. Yet the total number of subjects tested remained small, particularly regarding females. The present paper provides additional static pressure-volume data obtained from adult volunteers, both women and men.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>Subjects</title><p>Forty-eight normal subjects agreed to participate in the study. The subjects were recruited from students and staff at the University of Maryland. The study was approved by the Institutional Review Board and all subjects gave informed consent. The subjects' characteristics are shown in Table <xref ref-type="table" rid="T1">1</xref>.</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>Age, height, weight, and lung volume data in the two groups tested shown with standard deviations.</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="center">Group</td><td align="center">Number of Subjects</td><td align="center">Average Age (Range)</td><td align="center">Average Height, cm</td><td align="center">Average Mass kg</td><td align="center">Average Vital Capacity, L</td></tr></thead><tbody><tr><td align="center">Females</td><td align="center">29</td><td align="center">25.5 ± 3.7 (18 to 31)</td><td align="center">164 ± 6</td><td align="center">60.5 ± 13.3</td><td align="center">3.58 ± 0.63</td></tr><tr><td align="center">Males</td><td align="center">19</td><td align="center">26.5 ± 4.3 (19 to 34)</td><td align="center">177 ± 7</td><td align="center">74.1 ± 8.8</td><td align="center">4.72 ± 8.4</td></tr></tbody></table></table-wrap></sec><sec><title>Protocol</title><p>Subjects were first acquainted with the spirometer and the test protocol. They were instructed in the definition of functional residual capacity (FRC) as the resting volume of the lung and given time to practice finding FRC. The subjects were then measured for inspiratory capacity (IC) and expiratory reserve volume (ERV) relative to FRC. Volume measurements were repeated until three consecutive maneuvers produced volumes within a 100 ml range. The average of the three volumes was recorded. Vital capacity (VC) was calculated as the sum of IC and ERV.</p><p>Maximum pressure measurements were taken from occlusions occurring at nine predetermined target volumes. The volumes were randomly ordered and ranged from 10% to 90% of VC by 10% increments. Subjects began each maneuver at FRC. Subjects were instructed to inhale or exhale, as necessary, to the desired volume. When the target volume was obtained, the occlusion valve automatically closed. The subject was told to inhale or exhale, as necessary, for two seconds. After two seconds of effort, the valve was released. This gave the subject access to fresh air for at least one minute of rest. More time was given, if desired. After all 18 measurements were taken, the test was repeated. The stronger effort, or higher maximum pressure, at each measurement was saved.</p></sec><sec><title>Measurements</title><p>A commercial spirometery system (Collins™, Braintree, MA) was used for all the measurements. Lung volumes were monitored by the dry-seal bell spirometer. The occlusion valve and pressure transducer utilized were those located in what Collins refers to as its "universal breathing valve." Collins "Research Assistant" (RA) software controlled the occlusion valve and collected pressure and volume measurements. For the experiment, a supervisory program, "PV", was authored in Microsoft Visual Basic for Applications<sup>™ </sup>to configure RA and provide feedback to the experimenters. PV was designed to fill the spirometer with an appropriate volume of fresh air prior to each measurement. It then calculated the spirometer volume corresponding to the target lung volume, taking thermal expansion into account. Occlusion was triggered automatically when the subject reached the target lung volume. As target lung volumes were randomly ordered, a randomization feature was built into PV. After each occlusion, RA returned pressure and volume data, which PV analyzed and saved to disk. The spirometer was calibrated each test day with a three-liter syringe and the pressure transducer was calibrated each day with a 10 cmH<sub>2</sub>O manometer.</p></sec><sec><title>Data analysis</title><p>The maximum inhalation or exhalation pressure magnitude (P<sub>i </sub>or P<sub>e</sub>) at each lung volume (V<sub>L</sub>) was recorded by the computer. The average pressure for the last one second of each effort was calculated. This pressure was then used to correct V<sub>L </sub>using the method by Cook et al. [<xref ref-type="bibr" rid="B2">2</xref>]. Absolute lung volumes were not measured and volumes were calculated based on the assumption that residual volume was 26% that of TLC.</p></sec></sec><sec><title>Results</title><p>Average maximum pressure values for all of the female subjects tested appear in Table <xref ref-type="table" rid="T2">2</xref> and values for all of the male subjects appear in Table <xref ref-type="table" rid="T3">3</xref>. Observations were grouped according to the lung volumes at which occlusion occurred and actual volumes within each group were averaged to produce the tabled values. The pressures produced by the men were typically 64% higher than the women in expiration, and 53% higher in inspiration.</p><table-wrap position="float" id="T2"><label>Table 2</label><caption><p>Maximal inspiratory and expiratory static pressures at different lung volumes for the female subjects.</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="center">Expiratory</td><td align="center" colspan="2">Pressure</td><td align="center">Inspiratory</td><td align="center" colspan="2">Pressure</td></tr></thead><tbody><tr><td align="center">Volume (%VC)</td><td align="center">Positive Pressure (cmH<sub>2</sub>O)</td><td align="center">P<sub>e</sub>/P<sub>max</sub></td><td align="center">Volume (%VC)</td><td align="center">Negative Pressure (cmH<sub>2</sub>O)</td><td align="center">P<sub>i</sub>/P<sub>max</sub></td></tr><tr><td colspan="6"><hr></hr></td></tr><tr><td align="center">8.9 ± 0.7</td><td align="center">38.7 ± 25.0</td><td align="center">0.5864</td><td align="center">12.1 ± 0.9</td><td align="center">65.9 ± 31.6</td><td align="center">0.9985</td></tr><tr><td align="center">18.4 ± 1.0</td><td align="center">44.2 ± 25.9</td><td align="center">0.6697</td><td align="center">22.5 ± 1.3</td><td align="center">65.0 ± 31.6</td><td align="center">0.9848</td></tr><tr><td align="center">27.6 ± 1.3</td><td align="center">53.4 ± 30.7</td><td align="center">0.8091</td><td align="center">33.0 ± 1.8</td><td align="center">59.6 ± 32.0</td><td align="center">0.9030</td></tr><tr><td align="center">37.3 ± 1.5</td><td align="center">53.7 ± 28.8</td><td align="center">0.8136</td><td align="center">43.3 ± 2.0</td><td align="center">55.0 ± 32.0</td><td align="center">0.8333</td></tr><tr><td align="center">46.8 ± 1.9</td><td align="center">53.3 ± 28.9</td><td align="center">0.8076</td><td align="center">53.5 ± 2.2</td><td align="center">53.0 ± 32.9</td><td align="center">0.8030</td></tr><tr><td align="center">56.2 ± 2.2</td><td align="center">55.4 ± 29.6</td><td align="center">0.8394</td><td align="center">63.8 ± 2.6</td><td align="center">47.6 ± 28.3</td><td align="center">0.7212</td></tr><tr><td align="center">65.7 ± 2.6</td><td align="center">57.3 ± 33.2</td><td align="center">0.8682</td><td align="center">73.6 ± 2.6</td><td align="center">42.3 ± 28.6</td><td align="center">0.6409</td></tr><tr><td align="center">74.7 ± 3.2</td><td align="center">61.7 ± 38.5</td><td align="center">0.9398</td><td align="center">83.1 ± 2.7</td><td align="center">34.2 ± 27.4</td><td align="center">0.5182</td></tr><tr><td align="center">84.4 ± 3.2</td><td align="center">61.2 ± 39.0</td><td align="center">0.9273</td><td align="center">92.6 ± 2.4</td><td align="center">28.0 ± 29.0</td><td align="center">0.4242</td></tr></tbody></table><table-wrap-foot><p>Volumes are expressed in percent of vital capacity at ambient pressure. All values are shown with standard deviations. P<sub>max </sub>for females was determined to be 66 cmH<sub>2</sub>O.</p></table-wrap-foot></table-wrap><table-wrap position="float" id="T3"><label>Table 3</label><caption><p>Maximal inspiratory and expiratory static pressure at different lung volumes for the male subjects.</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="center">Expiratory</td><td align="center" colspan="2">Pressure</td><td align="center">Inspiratory</td><td align="center" colspan="2">Pressure</td></tr></thead><tbody><tr><td align="center">Volume (%VC)</td><td align="center">Positive Pressure (cmH<sub>2</sub>O)</td><td align="center">P<sub>e</sub>/P<sub>max</sub></td><td align="center">Volume (%VC)</td><td align="center">Negative Pressure (cmH<sub>2</sub>O)</td><td align="center">P<sub>i</sub>/P<sub>max</sub></td></tr><tr><td colspan="6"><hr></hr></td></tr><tr><td align="center">8.2 ± 1.0</td><td align="center">63.2 ± 32.6</td><td align="center">0.6196</td><td align="center">13.5 ± 2.0</td><td align="center">96.7 ± 45.5</td><td align="center">0.9480</td></tr><tr><td align="center">17.2 ± 1.3</td><td align="center">75.3 ± 39.0</td><td align="center">0.7382</td><td align="center">24.2 ± 2.8</td><td align="center">95.3 ± 49.8</td><td align="center">0.9343</td></tr><tr><td align="center">26.2 ± 1.6</td><td align="center">84.1 ± 39.9</td><td align="center">0.8245</td><td align="center">35.2 ± 3.2</td><td align="center">97.3 ± 52.2</td><td align="center">0.9539</td></tr><tr><td align="center">35.2 ± 2.2</td><td align="center">91.6 ± 38.3</td><td align="center">0.8980</td><td align="center">45.9 ± 3.6</td><td align="center">87.2 ± 42.3</td><td align="center">0.8549</td></tr><tr><td align="center">44.8 ± 2.6</td><td align="center">89.6 ± 41.3</td><td align="center">0.8784</td><td align="center">55.7 ± 3.2</td><td align="center">77.9 ± 36.9</td><td align="center">0.7637</td></tr><tr><td align="center">53.5 ± 2.7</td><td align="center">93.3 ± 36.3</td><td align="center">0.9147</td><td align="center">66.7 ± 4.2</td><td align="center">75.4 ± 42.2</td><td align="center">0.7392</td></tr><tr><td align="center">62.6 ± 3.4</td><td align="center">94.6 ± 44.3</td><td align="center">0.9275</td><td align="center">76.5 ± 4.5</td><td align="center">69.4 ± 40.4</td><td align="center">0.6804</td></tr><tr><td align="center">72.1 ± 3.8</td><td align="center">94.2 ± 43.3</td><td align="center">0.9235</td><td align="center">85.5 ± 4.9</td><td align="center">51.4 ± 38.4</td><td align="center">0.5039</td></tr><tr><td align="center">80.8 ± 4.2</td><td align="center">97.2 ± 41.5</td><td align="center">0.9529</td><td align="center">94.2 ± 5.1</td><td align="center">38.9 ± 38.8</td><td align="center">0.3814</td></tr></tbody></table><table-wrap-foot><p>Volumes are expressed in percent of vital capacity at ambient pressure. All values are shown with standard deviations. P<sub>max </sub>for males was determined to be 102 cmH<sub>2</sub>O.</p></table-wrap-foot></table-wrap><p>Data were then scrutinized in an exploratory manner to see if they could be easily and universally fit by a simple mathematical expression [<xref ref-type="bibr" rid="B5">5</xref>]. It was found that both men's and women's data could be described by an expression of the form:</p><p>P<sub>e</sub>/P<sub>max </sub>= A Ln (%VC) + B for exhalation</p><p>and</p><p>P<sub>i</sub>/P<sub>max </sub>= C Ln (100 - %VC) + D for inhalation</p><p>Here, P<sub>max </sub>is the asymptotically maximum pressure that could be developed by the respiratory muscles at any lung volume and P<sub>i </sub>is the maximum inspiratory pressure that can be developed at specific lung volumes. The average value of P<sub>max </sub>found by determining the limit of the nonlinear P-V curve for the group of subjects was found to be 102 cmH<sub>2</sub>O for males and 66 cmH<sub>2</sub>O for females, and was found to be the same for both inhalation and exhalation directions. Least squares regression using Microsoft Excel yielded the following two equations:</p><p>P<sub>e</sub>/P<sub>max </sub>= 0.1426 Ln (%VC) + 0.3402 R<sup>2 </sup>= 0.9549</p><p>and</p><p>P<sub>i</sub>/P<sub>max </sub>= 0.234 Ln (100% - %VC) - 0.0828 R<sup>2 </sup>= 0.9642</p><p>These equations are graphed in Figure <xref ref-type="fig" rid="F1">1</xref>.</p><fig position="float" id="F1"><label>Figure 1</label><caption><p>Graph of empirical equations determined to describe relative expiratory and inspiratory pressures for men and women. Some data points are coincident. Symbols: F = female; M = male; other symbols as in text.</p></caption><graphic xlink:href="1475-925X-5-29-1"/></fig></sec><sec><title>Discussion</title><p>Greater (more positive) expiratory pressures were developed at higher V<sub>L</sub>, while greater (more negative) inspiratory pressures were developed at lower V<sub>L</sub>. In women P<sub>e </sub>increased with volume from 39 to 61 cmH<sub>2</sub>O and P<sub>i </sub>decreased with volume from 66 to 28 cmH<sub>2</sub>O. In men, P<sub>e </sub>increased with volume from 63 to 97 cmH<sub>2</sub>O and P<sub>1 </sub>decreased with volume from 97 to 39 cmH<sub>2</sub>O. These trends occur primarily for two reasons. First, respiratory muscles work both with and against respiratory tissue elastance to produce pressure. Expiratory efforts are aided by tissue elastance (lung recoil effects) at high V<sub>L </sub>and inhibited at low V<sub>L</sub>. Inspiratory efforts are inhibited by tissue elastance at high V<sub>L </sub>and aided at low V<sub>L</sub>. Second, respiratory muscles exert greater tension when they are stretched to greater lengths. Expiratory muscles are stretched when the lung is inflated, while inspiratory muscles are stretched when the lung is deflated. Both of these factors describe the general trend of the data.</p><p>The volume dependence of P<sub>i </sub>was much more pronounced than the volume dependence of P<sub>e </sub>in both women and men. This can be seen as a higher slope of the inspiratory equation compared to the expiratory equation in Figure <xref ref-type="fig" rid="F1">1</xref>. This may reflect a combination of strength differences between diaphragm (largely responsible for inhalation) and abdominal muscles (largely responsible for exhalation) recruitment of intercoastal muscle (largely responsible for posture maintenance), and different mechanical advantages of each type of muscle as the lung volume varies.</p><p>The Laplace equation may be relevant here. This equation states that enclosed pressure is proportional to the product of wall tension and wall thickness and inversely proportional to the radius of curvature. The Laplace equation for a sphere differs from that of a cylinder by a factor of two. Pressure in a sphere (P = <inline-graphic xlink:href="1475-925X-5-29-i1.gif"/>) is twice that of a cylinder (P = <inline-graphic xlink:href="1475-925X-5-29-i2.gif"/>), all other things being equal.</p><p>The diaphragm is positioned under the lungs and curves upward in a somewhat spherical shape. As it contracts, it becomes flatter, meaning that its radius of curvature increases. Lung volume increases as the diaphragm contracts. If the Laplace equation can be applied to the respiratory system, then it would show that inspiratory pressure should decrease as radius, and thus lung volume, increases (P<sub>i </sub>∝ <inline-graphic xlink:href="1475-925X-5-29-i3.gif"/>, as long as wall tension and thickness remain steady).</p><p>The abdominal muscles are arranged differently, more like wrapping around a cylinder. The abdominal muscles flatten at smaller lung volumes instead of larger lung volumes, and the Laplace equation indicates that higher pressures should be developed at larger lung volumes (P ∝ V).</p><p>Both effects have been observed. Inspiratory pressures increase as lung volume decreases and expiratory pressure increases as lung volume increases. There is roughly a factor of two between the dependence of pressures upon lung volumes for inspiration and expiration. This could well be related to the difference in the Laplace equation for a sphere and a cylinder.</p><p>The pressure-volume data obtained in this study are of the same general magnitudes as those previously reported [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B2">2</xref>,<xref ref-type="bibr" rid="B6">6</xref>]. Rahn et al. [<xref ref-type="bibr" rid="B1">1</xref>] studied P<sub>i </sub>in 11 men and P<sub>e </sub>in 12 men using similar methods. The highest pressures from three efforts at each of six starting volumes were recorded. Measurements were read from a mercury manometer connected to the subjects' noses. Pressure-volume data closely match the results of the present study. Craig [<xref ref-type="bibr" rid="B6">6</xref>] produced pressure-volume diagrams from 10 men using methods similar to the present study. Pressures were taken from a mercury manometer connected to the subjects' mouths. These data also closely match the results of the present study. Cook et al. [<xref ref-type="bibr" rid="B2">2</xref>] studied 17 males and 9 females using two techniques. One technique was a conventional occlusion maneuver. The other technique involved subjects breathing into or out of large, fixed volumes. The compressibility of the air in differently sized containers provided for various ultimate lung volumes. The volumes were calculated from Boyle's Law using peak pressures that could be sustained for 1–2 seconds. Five volumes were used. It was concluded that the results of the occlusion method and the compression method were the same. In women, the compression-method P<sub>i </sub>values were similar to those of the present study at high volumes, but slightly higher at lower volumes. The P<sub>e </sub>values were similar at low volumes, but much higher at higher lung volumes. In men, the compression-method P<sub>i </sub>values agree well with those of the present study. However, the P<sub>e </sub>values are much higher than those of the present study at the higher volumes. Cook et al. [<xref ref-type="bibr" rid="B2">2</xref>] suggested that their P<sub>e </sub>values might have been higher than the Rahn et al. values because of the use of mouth pressure measurements rather than nose pressure measurements. It was also hypothesized that these P<sub>e </sub>values exceeded Craig's values due to better mouthpiece sealing.</p><p>As the results of this study are more in agreement with work of Rahn et al. [<xref ref-type="bibr" rid="B1">1</xref>] and Craig [<xref ref-type="bibr" rid="B6">6</xref>], it is more likely that there is another reason for the discrepancy. Aside from muscle strength alone, P<sub>e </sub>and P<sub>i </sub>are highly effort dependent. Subjects may limit their maximum pressures due to factors such as pain in the ear or general discomfort. During some maximum pressure maneuvers, researchers have observed changes in hemodynamics leading to loss of consciousness [<xref ref-type="bibr" rid="B7">7</xref>]. It is possible that the subjects of the Cook et al. [<xref ref-type="bibr" rid="B2">2</xref>] study were more highly motivated. It is also possible that these subjects were of above average strength.</p><p>Numerous authors have collected maximal pressures at a single V<sub>L</sub>. Most recently, Wilson et al. [<xref ref-type="bibr" rid="B8">8</xref>] measured maximal P<sub>e </sub>and P<sub>i </sub>in 87 women and 48 men using partial occlusion and Bourdon gauges. The women were found to have P<sub>e </sub>= 93 ± 17 cmH<sub>2</sub>O and P<sub>i </sub>= 73 ± 22 cmH<sub>2</sub>O and the men were found to have P<sub>e </sub>= 148 ± 34 cmH<sub>2</sub>O and P<sub>i </sub>= 106 ± 31 cmH<sub>2</sub>O. It could be expected that the P<sub>i </sub>and P<sub>e </sub>values from a single volume study would exceed the values from a multiple volume study because more efforts are made at the optimal V<sub>L </sub>in the single volume study, while muscle fatigue can be a factor in the multiple volume study.</p><p>Judging from inspiration values, this does not appear to be the case. In the present study, women were found to have P<sub>i </sub>= 66 ± 32 cmH<sub>2</sub>O at V<sub>L </sub>= 12%VC and men were found to have P<sub>i </sub>= 97 ± 46 cmH<sub>2</sub>O at V<sub>L </sub>= 14%VC. These values are virtually identical to the Wilson et al. [<xref ref-type="bibr" rid="B8">8</xref>] data. On the other hand, women in this study were found to have P<sub>e </sub>= 61 ± 39 cmH<sub>2</sub>O at V<sub>L </sub>= 84%VC and men were found to have P<sub>e </sub>97 ± 42 cmH<sub>2</sub>O at V<sub>L </sub>= 81%VC. These values are considerably smaller than the Wilson et al. data. The P<sub>e </sub>values of the single volume study fall in between the maximum P<sub>e </sub>values of the present study and the maximum P<sub>e </sub>values of the Cook et al. [<xref ref-type="bibr" rid="B2">2</xref>] study.</p><p>Satisfactorily describing maximum lung pressures with mathematical expressions can be helpful for respiratory mechanical modeling [<xref ref-type="bibr" rid="B3">3</xref>]. It is not likely that maximal pressures would be developed in young, healthy adults during quiet breathing. During exercise, and especially during expiratory flow limitation, however, maximum pressures may well be developed. For example, modeling the effects of respiratory masks during hard work could use these equations to calculate respiratory work rate. These equation forms are good because pressures and lung volumes both appear as relative rather than absolute values. That way, both men's and women's pressures could be determined with the same equations despite large differences in absolute pressures developed. Respiratory models for those conditions could well use the equations developed here.</p><p>Although we have no data to support the notion, it is possible, if their respiratory mechanics changed proportionally, that maximum pressures developed by patients with respiratory impairments could be described by the same equations as developed here. That is because these equations are in relative pressure and volume form. One would expect that P<sub>max </sub>could be much lower in diseased patients, but P/P<sub>max </sub>could be scaled the same. If this were so, then equations developed here could have more universal value.</p></sec><sec><title>Conclusion</title><p>Maximum pressures at the mouth have been determined to depend on lung volumes. Equations to describe these pressures have been developed, and these are in a form that may be useful for modeling and predictive purposes.</p></sec><sec><title>Abbreviations</title><p>ERV expiratory reserve volume of the lung, L</p><p>FRC functional residual capacity of the lung, L</p><p>IC inspiratory capacity of the lung, L</p><p>P<sub>e </sub>volume-dependent maximum expiratory pressure, cmH<sub>2</sub>O</p><p>P<sub>i </sub>volume-dependent maximum inspiratory pressure, cmH<sub>2</sub>O</p><p>P<sub>max </sub>volume-independent maximum pressure, cmH<sub>2</sub>O</p><p>PV name of supervisory computer program</p><p>RA proprietary data acquisition and analysis program from Collins</p><p>TLC total lung capacity, L</p><p>V<sub>L </sub>lung volume, LVC</p><p>VC vital capacity of the lung, L</p></sec><sec><title>Competing interests</title><p>The author(s) declare that they have no competing interests.</p></sec><sec><title>Authors' contributions</title><p>CGL wrote software and procedures, and conducted many of the tests. ATJ suggested the experiment, wrote much of the paper, and determined the form of equations. WHS provided support, including planning, protocol, and supplies. MMJ assisted with planning and conducted many of the tests. KMC assisted with planning and conducted many of the tests. DCC provided data analysis support.</p></sec> |
Antenna design for microwave hepatic ablation using an axisymmetric electromagnetic model | <sec><title>Background</title><p>An axisymmetric finite element method (FEM) model was employed to demonstrate important techniques used in the design of antennas for hepatic microwave ablation (MWA). To effectively treat deep-seated hepatic tumors, these antennas should produce a highly localized specific absorption rate (SAR) pattern and be efficient radiators at approved generator frequencies.</p></sec><sec sec-type="methods"><title>Methods and results</title><p>As an example, a double slot choked antenna for hepatic MWA was designed and implemented using FEMLAB™ 3.0.</p></sec><sec><title>Discussion</title><p>This paper emphasizes the importance of factors that can affect simulation accuracy, which include boundary conditions, the dielectric properties of liver tissue, and mesh resolution.</p></sec> | <contrib id="A1" contrib-type="author"><name><surname>Bertram</surname><given-names>John M</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>jmbertram@gmail.com</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Yang</surname><given-names>Deshan</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>dyangchicago@yahoo.com</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Converse</surname><given-names>Mark C</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>converse@cae.wisc.edu</email></contrib><contrib id="A4" corresp="yes" contrib-type="author"><name><surname>Webster</surname><given-names>John G</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>webster@engr.wisc.edu</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Mahvi</surname><given-names>David M</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>mahvi@surgadmin.surgery.wisc.edu</email></contrib> | BioMedical Engineering OnLine | <sec><title>Introduction</title><p>Liver cancer is a significant worldwide public health issue. The disease has a mortality rate of 100% at 5 years in untreated cases [<xref ref-type="bibr" rid="B1">1</xref>] and results in the deaths of more than one million people each year worldwide [<xref ref-type="bibr" rid="B2">2</xref>-<xref ref-type="bibr" rid="B5">5</xref>]. Although liver cancer can be treated successfully by surgical resection of the malignant tissue, approximately 90% of patients with the disease are ineligible for the procedure due to factors such as insufficient hepatic reserve and the close proximity of tumors to blood vessels [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B6">6</xref>]. One promising alternative for these patients is hepatic microwave ablation (MWA), an experimental procedure in which an antenna is inserted percutaneously or during surgery [<xref ref-type="bibr" rid="B7">7</xref>] to induce cell necrosis through the heating of deep-seated tumors. Unlike other alternatives to resection such as RF ablation and cryoablation, MWA systems are able to produce large lesions in the presence of blood perfusion and are not restricted by tissue charring [<xref ref-type="bibr" rid="B5">5</xref>,<xref ref-type="bibr" rid="B7">7</xref>]. MWA also has favorable one, two, and three year survival rates of 96%, 83%, and 73% [<xref ref-type="bibr" rid="B8">8</xref>].</p><p>The many perceived advantages of microwave ablation have driven researchers to develop innovative antennas to effectively treat deep-seated, nonresectable hepatic tumors. These designs have focused largely on thin, coaxial-based interstitial antennas [<xref ref-type="bibr" rid="B9">9</xref>-<xref ref-type="bibr" rid="B11">11</xref>], which are minimally invasive and capable of delivering a large amount of electromagnetic power. These antennas can usually be classified as one of three types (dipole, slot, or monopole) based on their physical features and radiative properties.</p><p>To assist in antenna design for MWA, many researchers have employed the use of mathematical models rooted in computational electromagnetics (CEM), a discipline that employs numerical methods to describe propagation of electromagnetic waves. These methods center around the formulation of discrete solutions to the fundamental electromagnetic equations collectively referred to as Maxwell's equations. Three principal techniques exist within CEM. The first of these, the finite-difference time-domain (FDTD) method, is based on the Yee algorithm [<xref ref-type="bibr" rid="B12">12</xref>] and uses finite difference approximations of the time and space derivatives of Maxwell's curl equations to create a discrete three-dimensional representation of the electric and magnetic fields. This method has been widely used to numerically evaluate the electromagnetic radiation patterns of antennas in tissue [<xref ref-type="bibr" rid="B13">13</xref>-<xref ref-type="bibr" rid="B17">17</xref>], although long computation times are generally required. Another commonly used CEM technique is method of moments (MoM) [<xref ref-type="bibr" rid="B18">18</xref>,<xref ref-type="bibr" rid="B19">19</xref>], in which approximate numerical solutions to integral equations are formulated in the frequency domain to determine an unknown current distribution for an antenna. This distribution can then be subsequently extended to yield the antenna's radiation pattern.</p><p>An alternative to the two techniques above is the finite element method (FEM), which has been extensively used in simulations of cardiac and hepatic radio-frequency (RF) ablation [<xref ref-type="bibr" rid="B20">20</xref>,<xref ref-type="bibr" rid="B21">21</xref>]. FEM models can provide users with quick, accurate solutions to multiple systems of differential equations and as such, are well suited to heat transfer problems like ablation. In this study, we have adapted an existing axisymmetric electromagnetic FEM model that was implemented using FEMLAB™ 3.0 [<xref ref-type="bibr" rid="B22">22</xref>] to demonstrate proper techniques for the design of an antenna for hepatic MWA.</p></sec><sec><title>Background</title><p>The FEM model used in this study was adapted from a coaxial slot antenna general model, developed by COMSOL for microwave cancer therapy [<xref ref-type="bibr" rid="B22">22</xref>]. In this model, the electric and magnetic fields associated with the time-varying transverse electromagnetic (TEM) wave generated by the microwave source propagating in a coaxial cable in the z-direction was expressed in 2D axially symmetric cylindrical coordinates as</p><p><inline-graphic xlink:href="1475-925X-5-15-i1.gif"/></p><p><inline-graphic xlink:href="1475-925X-5-15-i2.gif"/></p><p>with</p><p><inline-graphic xlink:href="1475-925X-5-15-i3.gif"/></p><p>where <italic>r</italic><sub>o </sub>and <italic>r</italic><sub>1</sub>are the outer and inner radii of the coaxial cable (m), <italic>P</italic><sub>in </sub>is the input power (W), <italic>ε</italic><sub>rd </sub>is the relative permittivity of the dielectric, <inline-graphic xlink:href="1475-925X-5-15-i4.gif"/> is the wave impedance in the dielectric of the coaxial cable (Ω), <inline-graphic xlink:href="1475-925X-5-15-i5.gif"/> is the intrinsic impedance (Ω), <italic>ε</italic><sub>0 </sub>= 8.854 × 10<sup>-12 </sup>is the permittivity of free space (F/m), <italic>μ</italic><sub>0 </sub>= 4<italic>π </italic>× 10<sup>-7 </sup>is the permeability of free space (H/m), ω = 2<italic>π f </italic>is the angular frequency (rad/s), <italic>f </italic>is the frequency (Hz), <italic>k </italic>= 2<italic>π/λ </italic>is the propagation constant (m<sup>-1</sup>), and <italic>λ </italic>is the wavelength (m). For interstitial coaxial-based antennas during MWA, the magnetic field is purely azimuthal. The electric field is in the radial direction only inside the coaxial cable and in both radial and the axial direction inside the tissue. This allows for the antenna to be modeled using an axisymmetric transverse magnetic (TM) wave formulation, in which the source was modeled as a low reflection boundary</p><p><inline-graphic xlink:href="1475-925X-5-15-i6.gif"/></p><p>with excitation magnetic field H<sub><italic>φ</italic>0 </sub>= <italic>C</italic>/<italic>Zr</italic>. Such axisymmetric FEM models are highly desirable as they dramatically reduce computation time.</p><p>Careful examinations of both the antenna's specific absorption rate (SAR) pattern and frequency-dependent reflection coefficient in tissue are essential for the optimization of antennas for hepatic MWA. SAR represents the electromagnetic power deposited per unit mass in tissue (W/kg) and can be defined mathematically as</p><p><inline-graphic xlink:href="1475-925X-5-15-i7.gif"/></p><p>where <italic>σ </italic>is tissue conductivity (S/m) and <italic>ρ </italic>is tissue density (kg/m<sup>3</sup>) [<xref ref-type="bibr" rid="B23">23</xref>]. For the treatment of deep-seated hepatic tumors, the SAR pattern of an interstitial antenna should be highly localized near the distal tip of the antenna. Antenna efficiency can be quantified using the frequency-dependent reflection coefficient, which can be expressed logarithmically as</p><p><inline-graphic xlink:href="1475-925X-5-15-i8.gif"/></p><p>where <italic>P</italic><sub>r </sub>indicates reflected power (W). The frequency where the reflection coefficient is minimum is commonly referred to as the resonant frequency and should be approximately the same as the operating frequency of the generator used. Antennas operating with high reflection coefficients (especially at higher power levels) can cause overheating of the feedline possibly leading to damage to the coaxial line or due to the thin outer conductor damage to the tissue [<xref ref-type="bibr" rid="B24">24</xref>].</p><p>Critical to the development of electromagnetic models of hepatic MWA is an accurate knowledge of the dielectric properties of liver tissue. Fig. <xref ref-type="fig" rid="F1">1</xref> shows the dielectric properties of fresh bovine liver that was obtained from a local slaughterhouse and measured approximately 1 h postslaughter. Measurements were obtained using the dielectric spectroscopy procedure described by Popovic <italic>et al </italic>[<xref ref-type="bibr" rid="B25">25</xref>], using a custom designed, stainless steel and glass borosilicate open-ended coaxial probe. These measurements were subsequently converted to complex permittivity using a rational function model [<xref ref-type="bibr" rid="B26">26</xref>,<xref ref-type="bibr" rid="B27">27</xref>].</p><fig position="float" id="F1"><label>Figure 1</label><caption><p>Dielectric properties of liver tissue measured ex-vivo on freshly excised bovine liver.</p></caption><graphic xlink:href="1475-925X-5-15-1"/></fig></sec><sec sec-type="materials|methods"><title>Materials and methods</title><sec><title>Antenna design</title><p>Several antenna designs have been shown to be effective for the treatment of deep-seated hepatic tumors. Recently, Saito <italic>et al </italic>[<xref ref-type="bibr" rid="B28">28</xref>] presented a design for a double slot antenna that is capable of achieving a higher degree of SAR localization than a standard single slot antenna and has been used in initial clinical trials. Another popular design, the cap-choke antenna [<xref ref-type="bibr" rid="B29">29</xref>-<xref ref-type="bibr" rid="B31">31</xref>], was originally designed for cardiac MWA applications but has been also found through independent investigation to be an effective applicator for hepatic MWA. In this design, the choke acts as a balun (which isolates the antenna and feedline to provide a balanced output) to prevent current backflow along the axial length of the antenna, resulting in high SAR localization near the distal tip of the antenna [<xref ref-type="bibr" rid="B32">32</xref>-<xref ref-type="bibr" rid="B34">34</xref>].</p><p>Fig. <xref ref-type="fig" rid="F2">2</xref> shows the general schematic for a double slot choked antenna. Antenna geometry parameters, the slot spacing, choke offset, choke length, etc, were chosen based on the effective wavelength in bovine liver tissue at 2.45 GHz, which was calculated using</p><fig position="float" id="F2"><label>Figure 2</label><caption><p>Axial and radial schematics of a coaxial-based double slot choked antenna designed for hepatic microwave ablation (MWA). All units are in mm.</p></caption><graphic xlink:href="1475-925X-5-15-2"/></fig><p><inline-graphic xlink:href="1475-925X-5-15-i9.gif"/></p><p>where <italic>c </italic>is the speed of light in free space (m/s), <italic>f </italic>is the operating frequency of the microwave generator (2.45 GHz), and <italic>ε</italic><sub>r </sub>= 44.4 is the relative permittivity of bovine liver tissue at the operating frequency; this yielded 18.4 mm for the effective wavelength. However, because the catheter and its thickness also affect the optimal geometry and performance of the antenna, equation 7 only provides a very crude approximation for the design. Fig. <xref ref-type="fig" rid="F2">2</xref> shows that slot spacing, choke offset, and choke length correspond to 0.25<italic>λ</italic><sub>eff</sub>, 0.5<italic>λ</italic><sub>eff</sub>, and <italic>λ</italic><sub>eff </sub>respectively, which were chosen to achieve localized power deposition near the distal tip of the antenna. Tip length and catheter thickness were adjusted to achieve resonance.</p></sec><sec><title>Model development</title><p>For simplicity and to eliminate numerical error, the inner and outer conductors of the antenna were modeled using perfect electric conductor (PEC) boundary conditions. Low-reflecting boundary conditions were used along the model boundaries to prevent reflection artifacts and axial symmetry boundary conditions were also employed along the axis of rotation. Table <xref ref-type="table" rid="T1">1</xref> shows the frequency-independent material parameters used in this model. A modeling domain with radial and axial dimensions of 0.03 and 0.08 m was used for the simulation, and calculated results were plotted in terms of normalized SAR.</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>Frequency-independent material parameters.</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left">FEM Region</td><td align="center">Material</td><td align="center">Conductivity (S/m)</td><td align="center">Relative Permittivity</td></tr></thead><tbody><tr><td align="left">Catheter</td><td align="center">Teflon</td><td align="center">0</td><td align="center">2.1</td></tr><tr><td align="left">Coaxial cable dielectric</td><td align="center">PTFE</td><td align="center">0</td><td align="center">2.03</td></tr></tbody></table></table-wrap><p>To determine the frequency response of the antenna, the model was simulated in FEMLAB™ at multiple discrete frequencies between 0.8 and 10 GHz using 50 MHz increments and the dielectric properties for bovine liver tissue shown in Fig. <xref ref-type="fig" rid="F1">1</xref>. At each frequency, a boundary integration of power outflow along the source was performed using FEMLAB™ to determine the net power delivered to the antenna. Subsequently, (6) was used to calculate the reflected power <italic>P</italic><sub>r </sub>and frequency dependent reflection coefficient Γ(<italic>f</italic>).</p><p>To determine the optimal mesh resolution of the model and maximize computational efficiency, a convergence study was conducted in FEMLAB™ using the mesh parameters shown in Table <xref ref-type="table" rid="T2">2</xref>. This was performed by gradually increasing mesh resolution along the effective source, catheter, and outer conductor boundaries of the antenna, as well as within the coaxial cable dielectric. Resolution in each of these regions was adjusted individually and numerical convergence resulted when a uniform change of less than 0.1% in the reflected power and normalized SAR was reached. As a result of the convergence study, maximum element size was set to 0.001 m along the catheter, outer conductor, and axial symmetry boundaries. For the source boundary and within the coaxial cable dielectric, maximum element size was also set to 0.0001 and 0.0002 m, respectively.</p><table-wrap position="float" id="T2"><label>Table 2</label><caption><p>FEMLAB™ mesh parameters.</p></caption><table frame="hsides" rules="groups"><tbody><tr><td align="left">Element growth rate</td><td align="center">1.25</td></tr><tr><td align="left">Maximum element size scaling factor</td><td align="center">0.55</td></tr><tr><td align="left">Mesh curvature cut off</td><td align="center">0.0005</td></tr><tr><td align="left">Mesh curvature factor</td><td align="center">0.25</td></tr></tbody></table></table-wrap></sec><sec><title>Model validation</title><p>To validate the performance of the FEMLAB™ antenna model, a double slot choked antenna was constructed from UT-85 semirigid coaxial cable and gold-plated SMA connectors obtained from Microstock, Inc. This antenna was physically immersed in saline without a catheter and connected to an Agilent E8364A performance network analyzer (PNA) to measure the frequency-dependent reflection coefficient of the antenna. The dielectric properties of saline were measured using the same dielectric spectroscopy procedure mentioned above for liver tissue. These properties were subsequently used in a separate FEMLAB™ model to simulate the reflection coefficient of the antenna in saline without a catheter and were compared against measured results. A catheter was not used for the antenna in both measurement and FEMLAB model in order to avoid the possible effects by the catheter. Antenna performance is quite sensitive to the catheter thickness, but for practical reasons, the thickness could not be controlled exactly during the course of antenna construction. The comparison with the antenna without a catheter would still well serve the purpose to validate the computer model to experimental measurements.</p></sec></sec><sec><title>Results</title><p>Fig. <xref ref-type="fig" rid="F3">3(a)</xref> shows the axisymmetric finite element mesh, which was generated by FEMLAB™ using the mesh parameters in Table <xref ref-type="table" rid="T1">1</xref> and the maximum element sizes determined by the convergence study. This mesh consists of 8920 triangular elements. Fig. <xref ref-type="fig" rid="F3">3(b)</xref> shows the corresponding two-dimensional normalized SAR distribution produced by FEMLAB™ for a double slot choked antenna. Calculation time for a single simulation on a personal computer with 2.8 GHz Pentium™ 4 processor and 1.5 GB memory required approximately 8 s. This same simulation required 75 s on a laptop computer with 1.6 GHz Pentium™ 4 processor and 512 MB memory.</p><fig position="float" id="F3"><label>Figure 3</label><caption><p>An axisymmetric electromagnetic model of the double slot choked antenna used for hepatic microwave ablation (MWA). (a) FEM model. (b) Logarithmic SAR distribution. Results were normalized using the maximum SAR obtained.</p></caption><graphic xlink:href="1475-925X-5-15-3"/></fig><p>Fig. <xref ref-type="fig" rid="F4">4</xref> shows the simulated antenna reflection coefficient in liver expressed logarithmically as a function of frequency with the dashed line indicating the commonly used MWA frequency of 2.45 GHz. This frequency response was calculated using the procedure described above and required approximately 30 min to complete. Fig. <xref ref-type="fig" rid="F5">5</xref> shows the double slot choked antenna that was built to validate these results. Fig. <xref ref-type="fig" rid="F6">6</xref> shows the simulated and measured reflection coefficient of this constructed antenna in saline.</p><fig position="float" id="F4"><label>Figure 4</label><caption><p>The simulated reflection coefficient of the double slot choked antenna, expressed logarithmically. The dashed line represents the commonly used MWA frequency of 2.45 GHz.</p></caption><graphic xlink:href="1475-925X-5-15-4"/></fig><fig position="float" id="F5"><label>Figure 5</label><caption><p>Photograph of the double slot choked antenna, shown without a catheter.</p></caption><graphic xlink:href="1475-925X-5-15-5"/></fig><fig position="float" id="F6"><label>Figure 6</label><caption><p>Comparison of simulated and experimental reflection coefficient of the double slot choked antenna, expressed logarithmically. The antenna was simulated and measured in saline rather than liver to ensure a homogenous medium for improved validation.</p></caption><graphic xlink:href="1475-925X-5-15-6"/></fig></sec><sec><title>Discussion</title><p>Fig. <xref ref-type="fig" rid="F3">3(b)</xref> shows that the double slot choked antenna is able to provide a degree of power localization similar to those produced by other recently designed antennas. Volumes for 50% and 90% energy dissipation are 7.55 cm<sup>3 </sup>and 99.3 cm<sup>3 </sup>respectively. This initial analysis also shows that it may be capable of providing more localized power deposition than the earlier double slot antenna and standard cap-choke antenna, and should therefore be considered as a viable design for future research. Fig. <xref ref-type="fig" rid="F4">4</xref> shows that the antenna is resonant at 2.45 GHz with a low reflection coefficient of approximately -22 dB. Reducing catheter thickness by 0.1 mm slightly improved localization but decreased the resonant frequency of the antenna by approximately 400 MHz. Such effects were expected because the catheter is an important part of the antenna structure. However more studies are apparently necessary to learn the exact effects of the catheter and the antenna geometry with catheter should be optimized together for the desired resonant frequency and SAR localization.</p><p>Several factors should be considered when designing and simulating antennas for hepatic MWA. Due to phase considerations, both antenna dimensions and the thickness of the catheter should be chosen and optimized based on the effective wavelength in tissue. As this quantity is highly dependent on the relative permittivity of the medium, it is also essential that electromagnetic models of antennas for hepatic MWA employ accurate dielectric properties of liver tissue. Great care should be taken when measuring these properties or using previously published data, as the dielectric properties of liver tissue are highly dispersive and change with tissue type [<xref ref-type="bibr" rid="B35">35</xref>], water content [<xref ref-type="bibr" rid="B36">36</xref>], and temperature [<xref ref-type="bibr" rid="B37">37</xref>].</p><p>In addition, tradeoffs that are not well understood exist between performance metrics such as SAR and the reflection coefficient. From our experience, it is usually more desirable to achieve a localized power deposition than a highly resonant antenna since external impedance matching networks can be used to improve antenna efficiency [<xref ref-type="bibr" rid="B38">38</xref>]. However, our additional experiments have shown that due to thermal conduction, a localized SAR pattern will only result in controlled tissue heating when treatment duration is minimized. Not only through the tissue, heat can also be effectively conducted by the antenna metal body and result in uncontrolled tissue heating along the antenna body for long treatment durations even with localized SAR pattern of the antenna.</p><p>Compared with other CEM software packages such as XFDTD™ and FEKO™, FEMLAB™ provides a user-friendly graphical interface for quick and reliable antenna design using axisymmetric electromagnetic models that can be integrated with MATLAB™ for increased functionality. Also, as the software was designed for multiphysics applications, it should also be ideal for future thermal simulations of hepatic MWA that are based on the Pennes bio-heat equation [<xref ref-type="bibr" rid="B39">39</xref>] and are dependent on both the electromagnetic and thermal properties of tissue. Initial work has also shown that electromagnetic models implemented using FEMLAB™ can be used in genetic algorithms [<xref ref-type="bibr" rid="B40">40</xref>] to better optimize antennas.</p></sec><sec><title>Conclusion</title><p>An antenna for hepatic MWA was quickly and accurately simulated using an axisymmetric electromagnetic model implemented in FEMLAB™ 3.0. This model is ideal for analyzing the SAR patterns and reflection coefficient of antennas used in hepatic MWA, two metrics which are commonly used for the evaluation of such antennas. Using this model, a double slot choked antenna with performance comparable to current antennas was designed and evaluated.</p></sec><sec><title>Authors' contributions</title><p>JB performed the research. All authors revised, read and approved the final manuscript.</p></sec> |
A defined medium to investigate sliding motility in a <italic>Bacillus subtilis </italic>flagella-less mutant | <sec><title>Background</title><p>We have recently shown that undomesticated strains of <italic>Bacillus subtilis </italic>can extensively colonize the surfaces of rich, semi-solid media, by a flagellum-independent mechanism and suggested that sliding motility is responsible for surface migration. Here we have used a flagella-less <italic>hag </italic>null mutant to examine and confirm sliding motility.</p></sec><sec><title>Results</title><p>Using a defined semi-solid medium we determined that a <italic>B. subtilis hag </italic>mutant colonized the surface in two stages, first as tendril-like clusters of cells followed by a profuse pellicle-like film. We determined the levels of macro- and micro-nutrients required for the tendril-to-film transition. Sufficient levels of each of the macronutrients, glycerol, Na-glutamate, and Na-phosphate, and inorganic nutrients, K<sup>+</sup>, Mg<sup>2+</sup>, Fe<sup>2+ </sup>and Mn<sup>2+</sup>, were required for robust film formation. The K<sup>+ </sup>requirement was quantified in more detail, and the thresholds for complete tendril coverage (50 μM KCl) or film coverage (2–3 mM KCl) were determined. In addition, disruption of the genes for the higher affinity K<sup>+ </sup>transporter (KtrAB), but not the lower affinity K<sup>+ </sup>transporter (KtrCD), strongly inhibited the formation of both tendrils and films, and could be partially overcome by high levels of KCl. Examination of <italic>hag </italic>tendrils by confocal scanning laser microscopy revealed that tendrils are multicellular structures, but that the cells are not as highly organized as cells in wild-type <italic>B. subtilis </italic>pellicles.</p></sec><sec><title>Conclusion</title><p>These results suggest that <italic>B. subtilis </italic>can use sliding motility to colonize surfaces, using a tendril-like growth mode when various macronutrients or micronutrients are limiting. If nutrients are balanced and sufficient, the surfaces between tendrils can be colonized by robust surface films. Sliding motility may represent a strategy for nutrient-deprived cells to colonize surfaces in natural environments, such as plant roots, and the media described here may be useful in investigations of this growth phenotype.</p></sec> | <contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Fall</surname><given-names>Ray</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>fall@colorado.edu</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Kearns</surname><given-names>Daniel B</given-names></name><xref ref-type="aff" rid="I2">2</xref><xref ref-type="aff" rid="I3">3</xref><email>dbkearns@indiana.edu</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Nguyen</surname><given-names>Tam</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>tam.nguyen@colorado.edu</email></contrib> | BMC Microbiology | <sec><title>Background</title><p>Bacteria use a variety of motility mechanisms to colonize environments, including flagella-dependent swimming and swarming, and flagella-independent, twitching, gliding, and sliding (reviewed in [<xref ref-type="bibr" rid="B1">1</xref>]). Of these motility mechanisms, the least investigated is sliding motility, which Henrichsen [<xref ref-type="bibr" rid="B2">2</xref>] defined as surface translocation produced by expansive forces in the growing colony combined with special surface properties to lower the friction between the cells and substrate. Harshey [<xref ref-type="bibr" rid="B1">1</xref>] points out that sliding motility is a passive mode of translocation for spreading over surfaces. Sliding motility has been genetically studied in <italic>Mycobacterium smegmatis </italic>and found to require the formation of acetylated glycopeptidolipids (GPLs) in the outermost layer of the cell envelope [<xref ref-type="bibr" rid="B3">3</xref>,<xref ref-type="bibr" rid="B4">4</xref>]. It has been proposed that the hydrophobic fatty acyl tails of GPLs on the cell surface lower the friction to produce sliding motility on the surface of the medium (i.e. agarose). Additional examples of sliding-type motility have been reported in other bacteria, but not investigated in detail (reviewed in [<xref ref-type="bibr" rid="B1">1</xref>]).</p><p>Wild strains of <italic>Bacillus subtilis </italic>are known to translocate over solid surfaces by a mechanism of swarming motility [<xref ref-type="bibr" rid="B5">5</xref>,<xref ref-type="bibr" rid="B6">6</xref>]. Swarming cells secrete a lipopeptide surfactant, called surfactin, to reduce surface tension and motility is powered by rotating flagella [<xref ref-type="bibr" rid="B5">5</xref>,<xref ref-type="bibr" rid="B7">7</xref>,<xref ref-type="bibr" rid="B8">8</xref>]. We have recently reported that wild type <italic>B. subtilis </italic>strains can rapidly colonize the surface of semi-solid media in a flagellum-independent manner, and suggested that sliding motility might play a role [<xref ref-type="bibr" rid="B9">9</xref>]. It was shown that such surface colonization was also dependent on the secretion of surfactin, but microscopic examination of the edges and interior cells of sliding surface colonies did not reveal abundant flagella. We suggest that <italic>B. subtilis </italic>has two distinct modes of surface translocation, swarming and sliding, which are presumably advantageous under different environmental conditions.</p><p>Here we have developed an experimental system in <italic>B. subtilis </italic>to study sliding motility. Two key components of this approach are the use of i) a flagella-less mutant to rule out the contribution of swarming in surface migration, and ii) a defined growth medium that was earlier developed to study the formation of floating pellicles and fruiting bodies in <italic>B. subtilis </italic>[<xref ref-type="bibr" rid="B10">10</xref>]. The defined medium allowed us to manipulate the essential macro- and micronutrients needed for sliding motility and colony spreading. As with some gram-negative bacteria, such as <italic>Pseudomonas aeruginosa</italic>, which can swim, swarm or show twitching motility [<xref ref-type="bibr" rid="B11">11</xref>], the work reported here suggests that <italic>B. subtilis </italic>also has multiple means of colonizing surfaces.</p></sec><sec><title>Results</title><sec><title>A defined medium to visualize K<sup>+</sup>-dependent sliding motility in a hag mutant</title><p>As mentioned above, we have presented evidence that undomesticated <italic>B. subtilis </italic>strains, such as the Marburg strain 3610, can colonize the surfaces of semi-soft media using flagellar-dependent swimming and swarming as well as flagellar-independent sliding motility. To focus on sliding motility, a <italic>hag </italic>null mutant (defective in the coding gene for flagellin, an essential subunit in flagellum assembly [<xref ref-type="bibr" rid="B12">12</xref>]) of the undomesticated 3610 strain was used to eliminate any contribution of flagellar-dependent motility. Furthermore, a defined medium (MSgg) that has been used to study pellicle and fruiting body formation in <italic>B. subtilis </italic>[<xref ref-type="bibr" rid="B10">10</xref>] was modified to control the level of potassium ion (K<sup>+</sup>), as this monovalent cation is essential for flagellum-independent surface colonization by undomesticated <italic>B. subtilis </italic>[<xref ref-type="bibr" rid="B9">9</xref>]. For the modified medium, termed MSggN, the potassium phosphate component was substituted by equimolar sodium phosphate, and potassium ion levels were determined by the amount of KCl added.</p><p>As shown in Fig. <xref ref-type="fig" rid="F1">1</xref>, growth of the <italic>hag </italic>mutant on MSggN agarose plates from the central point of inoculation occurred by means of long tendril-like arms if the KCl level was low (i.e. 100 μM). In contrast, if plates contained 5 mM KCl the surface was completely colonized after 18 h of growth (37°C); and the surface had a wrinkled, web-like appearance that is similar to that seen in <italic>B. subtilis </italic>pellicles that form on the surface of MSgg liquid medium [<xref ref-type="bibr" rid="B10">10</xref>,<xref ref-type="bibr" rid="B13">13</xref>]. In contrast, the parent 3610 strain swarmed over the surface with a more dendritic central colony (low KCl) and produced less robust surface films (high KCl). The rate of surface migration of the swarming 3610 cells was about twice that of the sliding <italic>hag </italic>cells. Notably, the sliding motility seen in the <italic>hag </italic>mutant was dependent on the secretion of surfactin, as a <italic>hag </italic>mutant containing a surfactin gene deletion (strain RFH1) did not spread significantly from the point of inoculation, but the colony could spread if authenthic surfactin (50–200 μg) was first added to the center of the plate.</p><fig position="float" id="F1"><label>Figure 1</label><caption><p><bold>Surface colonization by a <italic>B. subtilis hag </italic>mutant DS64 and its wild type parent (3610) on defined semi-solid media is dependent on the level of potassium ion and agarose concentration</bold>. Semi-solid MSggN agarose plates (0.3 to 0.6% w/v agarose) were inoculated in the center with sharpened toothpicks with both <italic>hag </italic>and 3610 strains (in triplicate). After growth for 18 h at 37°C, typical plates were photographed. Abbreviations: Mk, MSggN medium with 100 μM KCl; MK, MSggN medium with 5 mM KCl.</p></caption><graphic xlink:href="1471-2180-6-31-1"/></fig><p>Other investigations of bacterial surface motility [<xref ref-type="bibr" rid="B1">1</xref>] have demonstrated that the degree of colony spreading is often highly dependent on the concentration of the solidifying agent, and can be dependent on the method of inoculation. We examined these issues, and, as shown in Fig. <xref ref-type="fig" rid="F1">1</xref>, as the concentration of agarose was increased the degree of surface colonization – either in <italic>hag </italic>or 3610 cells – decreased. To test the effect of the method of inoculation on surface growth of the <italic>hag </italic>mutant, agarose plates (in triplicate) were inoculated either by toothpick from LB (MLS) agar plates or by centrally spotting with 2.5 μl of various cell densities from cells grown in LB (MLS) broth. Qualitatively, the appearance of tendrils or films on each of these plates was independent of inoculation method or cell density, if at least 5 × 10<sup>3 </sup>viable cells were spotted (data not shown), and the toothpick method of inoculation was used in all subsequent experiments.</p><p>To further investigate the K<sup>+ </sup>requirements for the tendril and film growth phases, MSggN media were prepared with varying levels of KCl and the colony expansion on surfaces was compared to the growth rate in broth. As seen in Fig. <xref ref-type="fig" rid="F2">2</xref>, the KCl level in MSggN broth had dramatic consequences on <italic>hag </italic>growth such that growth yield was proportional to the amount of potassium ion in the medium; cell doubling times were also dependent on KCl concentration. No growth occurred if KCl was omitted. Similar KCl-dependent growth yields and cell doubling times were obtained in MSggN broth with the parent <italic>B. subtilis </italic>3610 strain (data not shown). The surface growth behavior of <italic>hag </italic>on MSggN agarose was similarly affected, in which the formation of tendrils that reached the edges of the plates was correlated with low KCl concentrations (up to 50 μM), whereas complete film coverage of the agarose surface required about 40 times that amount. <italic>B. subtilis hag </italic>sliding motility for the tendrils was substantial, typically 0.4–0.6 μm s<sup>-1 </sup>depending on the level of K<sup>+ </sup>ion and growth temperature. From these observations, we conclude that the different modes of surface colonization are correlated with K<sup>+</sup>-dependent growth rate. The stimulation of surface film growth by KCl could not be replaced by addition of other monovalent cation sources, including LiCl, RbCl, CsCl, or the relatively high level of Na<sup>+ </sup>ion in the basal MSggN medium (data not shown).</p><fig position="float" id="F2"><label>Figure 2</label><caption><p><bold>Quantification of the K<sup>+ </sup>requirement for <italic>hag </italic>growth on MSggN liquid and semisolid media</bold>. Cultures of <italic>hag </italic>were grown in MSggN semisolid or liquid media with the indicated levels of KCl, each in triplicate (semisolid) or duplicate (liquid), at 37°C, and the results were averaged. For the semisolid medium, the effect of KCl concentration on the maximum tendril length as measured at 18 h (<inline-graphic xlink:href="1471-2180-6-31-i1.gif"/>) is shown; KCl thresholds for tendril growth and formation of films that covered the semisolid media (50 mm plate surface) are indicated. For liquid MSggN medium the growth yield (<inline-graphic xlink:href="1471-2180-6-31-i2.gif"/>) and doubling times (three KCl levels from the tendril threshold to above the full film threshold) are shown. The experiment was replicated in a separate experiment with essentially the same results.</p></caption><graphic xlink:href="1471-2180-6-31-2"/></fig></sec><sec><title>The role of high and low affinity potassium ion transporters in the hag sliding phenotypes</title><p>The finding that very low levels of K<sup>+ </sup>ion are sufficient for the tendril growth seen in the <italic>hag </italic>strain suggested that this motility is dependent on the high affinity K<sup>+ </sup>transporter (KtrAB), but not the low affinity K<sup>+ </sup>transporter (KtrCD), present in <italic>B. subtilis </italic>[<xref ref-type="bibr" rid="B14">14</xref>]. To test this idea, we constructed the <italic>hag ktrAB </italic>triple mutant and as a control the <italic>hag ktrCD </italic>triple mutant. Each was grown on MSggN media with varying levels of KCl, with the results shown in Fig. <xref ref-type="fig" rid="F3">3</xref>. The results with the <italic>hag ktrAB </italic>mutant clearly indicate that disruption of the high affinity KtrAB transporter leads to inhibition of surface growth beyond the central point of inoculation of plates with 100 μM KCl, and weak central colony spreading at 1 to 14 mM KCl. In contrast, disruption of the low affinity KtrCD transporter resulted in a surface growth phenotype on MSggN plates that was indistinguishable from the parent <italic>hag </italic>strain. We have reported similar results with <italic>ktrAB </italic>and <italic>ktrCD </italic>mutations in <italic>B. subtilis </italic>3610, the wild-type parent of the <italic>hag </italic>strain examined here [<xref ref-type="bibr" rid="B15">15</xref>]. With respect to possible polarity effects, the <italic>ktrAB </italic>genes constitute a dicistronic operon ending in a putative rho-independent transcriptional terminator [<xref ref-type="bibr" rid="B16">16</xref>]. We therefore predict that the <italic>ktrAB </italic>insertion deletion construct is unlikely to have polar effects. While we have not tested complementation of the <italic>ktrAB </italic>alleles, these experiments are consistent with the importance of the KtrAB transporter for the growth phenotypes observed here, and further suggest that if tendril growth is disrupted by insufficient K<sup>+ </sup>uptake surface films cannot form.</p><fig position="float" id="F3"><label>Figure 3</label><caption><p><bold>The K<sup>+</sup>-dependence of <italic>hag </italic>growth on the defined medium requires the high affinity potassium transporter (KtrAB), but not the low affinity transporter (KtrCD)</bold>. MSggN plates were prepared with the indicated levels of KCl, inoculated in triplicate with either the <italic>hag </italic>strain, or <italic>hag ktrAB or hag ktrCD </italic>triple mutants (Table 1), and photographed after 24 h growth at 37°C.</p></caption><graphic xlink:href="1471-2180-6-31-3"/></fig></sec><sec><title>The requirement of other media components for surface sliding motility</title><p>The dramatic effects of manipulation of K<sup>+ </sup>ion levels on surface colonization in the defined medium led us to question if other components of the MSggN medium are essential for this type of sliding motility. Using the defined MSggN medium with KCl levels (5 mM) to satisfy the K<sup>+</sup>requirement, we investigated the effect of decreasing each macro- or micronutrient on the two phases of surface colony formation. These results are summarized in Table <xref ref-type="table" rid="T2">2</xref>. First with respect to macronutrients, decreasing glycerol, Na-glutamate or Na-phosphate by factors of 10 each resulted in disruption of surface film growth, and tendril-like colonies formed. A further 10-fold decrease of these macronutrients led to tendril growth for lowered glycerol or Na-phosphate but poor growth for lowered Na glutamate; these results are indicative of thresholds for the nitrogen (glutamate), carbon (glycerol and glutamate) and phosphate sources of the MSggN medium. Substitution of MOPS buffer with either HEPES or PIPES buffer (all at pH 7.0) had no discernable effects on the surface growth patterns.</p><p>When micronutrient levels were manipulated, it was seen that decreasing Mg<sup>2+</sup>, Fe<sup>2+ </sup>and Mn<sup>2+ </sup>eliminated the surface films but supported tendril growth, each with different thresholds: 20 μM MgCl<sub>2</sub>; 0.05 μM FeSO<sub>4</sub>; or 0.005 μM MnSO<sub>4 </sub>(Table <xref ref-type="table" rid="T2">2</xref>). Even when diluted 10<sup>-5</sup>-fold, there was no effect of lowering CaCl<sub>2 </sub>or ZnSO<sub>4 </sub>(Table <xref ref-type="table" rid="T2">2</xref>); surface films formed in each case. Some of the most striking visual results were obtained with lowered levels of Fe<sup>2+ </sup>and Mn<sup>2+ </sup>as shown in Fig. <xref ref-type="fig" rid="F4">4</xref>. These images show that at the threshold levels of these metals distinctly different tendril-like patterns of colony growth occur instead of complete surface film growth. Thus, as with K<sup>+ </sup>ion limitation (Fig. <xref ref-type="fig" rid="F1">1</xref>), lowering other metals in the medium can produce generally similar surface growth phenotypes. In each case, tendril rather than film growth could be due to decreased growth rate produced by the limiting component of the medium. However, in the examples shown in Fig. <xref ref-type="fig" rid="F4">4</xref> incubation of plates for up to 72 h led to persistence of tendrils without film formation (even when plates were incubated in a humid chamber to prevent drying).</p><fig position="float" id="F4"><label>Figure 4</label><caption><p><bold>Inhibition of surface film growth, but not tendril growth, occurs if either FeSO<sub>4 </sub>or MnSO<sub>4 </sub>is limiting</bold>. In this experiment, replicated independently, MSggN plates (5 mM KCl) were prepared with the normal levels of FeSO<sub>4 </sub>or MnSO<sub>4 </sub>(50 μM), or with the decreased levels of FeSO<sub>4 </sub>or MnSO<sub>4 </sub>indicated. Plates were inoculated with the hag strain and photographed after 24 h growth at 37°C. The level of agarose in the MSggN plates with MnSO<sub>4 </sub>was raised to 0.4% w/v (normally 0.3% w/v) to accentuate the details of the tendrils.</p></caption><graphic xlink:href="1471-2180-6-31-4"/></fig></sec><sec><title>Examination of tendril morphology by confocal scanning laser microscopy (CSLM)</title><p>To determine the organization of cells in during sliding of the <italic>hag </italic>mutant, we examined tendrils growing on MSggN agarose with low and high KCl, respectively, by CSLM using LIVE/DEAD BacLight straining methodology (Molecular Probes, Inc.; [<xref ref-type="bibr" rid="B17">17</xref>]). As shown in Fig. <xref ref-type="fig" rid="F5">5A</xref>, the staining of a tendril directly on the plate (under a cover slip) led to some dispersion of the tendril and an array of both live (green fluorescence) and dead (red fluorescence) cells. Nonetheless, the cells within the tendril did display cellular organization but were not as closely appressed or found in long chains as seen in <italic>B. subtilis </italic>pellicles [<xref ref-type="bibr" rid="B10">10</xref>]. The relatively high numbers of dead cells were likely due to the length of time the sample was examined in the CSLM instrument. In Fig. <xref ref-type="fig" rid="F5">5B</xref>, when tendril material was transferred to a coverslip and then quickly examined on a microscope slide, the tendril was much more intact and the majority of the cells imaged on the surface were live as evidenced by green fluorescence. Again, the cells did not appear to be forming closely-appressed, long chains. Attempts to image surface film colonies by these methods were complicated by the high density and thickness of the surface film. Clearly, more work is needed to understand the multicellular nature and cell viability of <italic>hag </italic>tendrils and films, but the CSLM images presented here support the idea that tendrils are indeed multicellular structures.</p><fig position="float" id="F5"><label>Figure 5</label><caption><p><bold>Confocal scanning laser microscopy (CSLM) of <italic>hag </italic>tendrils</bold>. In this experiment hag tendrils growing on MSggN plates (with 100 μM KCl) at 37°C were analyzed by CSLM using LIVE/DEAD BAcLight straining as described in the Methods. A) Tendril staining directly on a plate, where a 10 μl drop of stain (SYTO 9 and propidium iodide) was applied to a tendril and then covered with a cover slip (lightly pressed). B) Staining of a tendril that was removed from a plate by transfer to a cover slip that was then examined after contact with a 5 μl drop of stain on a glass slide.</p></caption><graphic xlink:href="1471-2180-6-31-5"/></fig></sec></sec><sec><title>Discussion</title><p>As is evident from many recent reviews (e.g. [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B18">18</xref>-<xref ref-type="bibr" rid="B20">20</xref>]), there is considerable interest in the mechanisms of bacterial motility on surfaces. Considering that <italic>B. subtilis </italic>is found in the soil, it is likely that cell populations of these bacteria utilize motility mechanisms to colonize surfaces such as plant roots [<xref ref-type="bibr" rid="B21">21</xref>]. Investigations of <italic>B. subtilis </italic>motility have primarily focused on flagellar-dependent mechanisms, including swimming and swarming behavior [<xref ref-type="bibr" rid="B22">22</xref>]. It is also notable that there are extensive investigations of pattern formation by <italic>B. subtilis </italic>on nutrient-limited agar media, a relatively slow and passive-type of motility supported by lubrication of the surface [[<xref ref-type="bibr" rid="B23">23</xref>,<xref ref-type="bibr" rid="B24">24</xref>], and references therein]. Here, we highlight the ability of <italic>B. subtilis </italic>to rapidly colonize surfaces using a flagellar-independent mechanism, sliding motility, and detail a defined medium to investigate this mode of surface migration.</p><p>We previously obtained evidence consistent with the phenomenon of sliding motility in wild type <italic>B. subtilis </italic>strains that produce the lipopeptide surfactin [<xref ref-type="bibr" rid="B9">9</xref>]. However, the conclusion that surface colonization was primarily due to flagella-independent motility rested on negative evidence, i.e. the failure to detect substantial numbers of flagellated cells at the moving boundary of spreading colonies. Here we have addressed this issue by studying surface spreading in a flagella-less mutant, in which the <italic>hag </italic>gene (encoding flagellin) was replaced by an antibiotic resistance cassette. As a result, any contribution of flagellar-dependent swarming motility can be eliminated. Notable features of the sliding motility of the <italic>hag </italic>mutant on MSggN agarose plates are i) that surface behavior of the <italic>hag </italic>mutant is comparable to the wild type, ii) colony spreading requires surfactin, iii) selective manipulation of nutrients alters growth and surface behavior, and iv) sliding migration is correlated with cellular growth rate.</p><p>Very few rates of sliding motility in bacteria have been published: 0.03–0.04 μm s<sup>-1 </sup>for <italic>M. smegmatis </italic>[<xref ref-type="bibr" rid="B25">25</xref>] to rates as high as 6 μm s<sup>-1 </sup>for <italic>Serratia marcescens </italic>[<xref ref-type="bibr" rid="B1">1</xref>]. <italic>B. subtilis </italic>sliding motility was substantial, typically 0.4–0.6 μm s<sup>-1 </sup>depending on the level of K<sup>+ </sup>ion and growth temperature. Because the rate of colony expansion depends on the rate of cell growth, the sliding behavior of different bacteria under different media conditions may not be directly comparable. Nonetheless, we note that the rates of <italic>B. subtilis </italic>sliding measured here fall within the published range for this type of surface behavior.</p><p><italic>B. subtilis </italic>sliding proceeds in two distinct growth phases, tendril and film growth. Tendrils first radiate from the point of inoculation, and if nutrient levels are sufficiently high, film growth begins and fills in the voids between the tendril arms. The two phases of sliding are analogous to similar behaviors seen in <italic>M. smegmatis </italic>[<xref ref-type="bibr" rid="B25">25</xref>]. In <italic>M. smegmatis</italic>, growth on a semi-solid medium with low levels of nutrients can lead to the appearance of tendril-like extensions that spread outwards to the edges of the plate, whereas growth on semi-solid nutrient media is seen as a distinct spreading film. Unlike <italic>M. smegmatis </italic>however, <italic>B. subtilis </italic>exhibits both behaviors, tendril and film-like growth, sequentially on the same medium. The question remains, why, in the presence of high levels of nutrients, does <italic>B. subtilis </italic>produce tendrils first rather than initiating film migration immediately. We suggest that the two behaviors may have to do with the fact that surfactin, an extracellular lipopeptide required for both tendril and film formation, is regulated by quorum-sensing. Surfactin production is known to be activated by a pheromone called ComX [<xref ref-type="bibr" rid="B26">26</xref>]. As <italic>B. subtilis </italic>grows, it secretes the ComX peptide, extracellular ComX concentration increases proportionally with cell density, and at a critical cell density, ComX activates surfactin production. Therefore, tendril formation may be the manifestation of a population at subcritical cell-density that secretes surfactin at levels insufficient to permit film spreading. The tendril continues to grow, ComX accumulates, surfactin production increases, and film formation initiates. Consistent with the idea that tendrils form when surfactin is limiting, we have found that when small amounts of surfactin (50–200 μg) are added to the center of agarose plates prior to inoculation, <italic>B. subtilis hag </italic>cells bypass the tendril spreading phase and proceed directly to film formation (unpublished observations). The tendrils are multicellular assemblies of cells, as shown by confocal scanning laser microscopy (CLSM; Fig. <xref ref-type="fig" rid="F5">5</xref>), but the control(s) over their initiation at the point of inoculation are unknown.</p><p>The continued growth of the spreading surface colonies produced a web-like appearance (e.g. Fig. <xref ref-type="fig" rid="F1">1</xref>), with the formation of sheets of cells that often collapsed onto the agarose surface. It is notable that the web-like surface films seen here are similar in appearance to floating pellicles formed by undomesticated <italic>B. subtilis </italic>strains [<xref ref-type="bibr" rid="B10">10</xref>,<xref ref-type="bibr" rid="B13">13</xref>], and it is possible that the spreading of multicellular structures on semi-soft MSggN medium shares features with multicellular organization in pellicles. However, further microscopic examination is needed to investigate this point.</p><p>Lowering the levels of essential nutrients can block the transition from tendril growth to film growth or abolish sliding motility entirely. The mechanism of these effects was examined in more detail for the K<sup>+ </sup>requirement, and we found that the K<sup>+ </sup>threshold for tendril growth was very low, only 50 μM, while that for full surface film formation was in the range of 2–3 mM. Both of these thresholds can be explained by the uptake of K<sup>+ </sup>ion by the KtrAB potassium transport system, which Holtmann et al. [<xref ref-type="bibr" rid="B14">14</xref>] have shown to be the major, higher affinity K<sup>+ </sup>transporter in <italic>B. subtilis</italic>. The KtrAB system and a second lower affinity KtrCD system have K<sub>m </sub>values of approximately 1 mM K<sup>+ </sup>and 10 mM K<sup>+</sup>, respectively [<xref ref-type="bibr" rid="B14">14</xref>]. The essential role of the KtrAB transporter in sliding motility was confirmed by the use of a <italic>hag ktrAB </italic>triple mutant. The model that emerges is that at K<sup>+ </sup>concentrations far below the K<sub>m </sub>of the KtrAB transporter, <italic>hag </italic>cells can grow only in clusters in the tendril form, and the growth rate is greatly reduced. In contrast, at levels of K<sup>+ </sup>ion above the K<sub>m </sub>of the KtrAB transporter full film growth can occur since these higher K<sup>+ </sup>levels allow much more rapid metabolism and growth. This data supports the idea that potassium ion has an intracellular role in sliding motility and that growth rate and sliding motility are highly correlated. It is also likely that limitations of other essential metals, Mg<sup>2+</sup>, Fe<sup>2+ </sup>and Mn<sup>2+ </sup>(Table <xref ref-type="table" rid="T2">2</xref>), inhibit growth rate, also resulting in the tendril phenotype. In addition, it is known that surfactin production in <italic>B. subtilis </italic>is greatly enhanced by the presence of iron sulfate [<xref ref-type="bibr" rid="B27">27</xref>], and perhaps at lowered levels of iron the surfactin formation needed to support robust film growth is insufficient.</p></sec><sec><title>Conclusion</title><p>Overall, the results presented here describe a defined medium and flagella-less <italic>hag </italic>mutant to investigate sliding motility in <italic>B. subtilis</italic>. When presented with a balanced MSggN medium the <italic>hag </italic>mutant can rapidly colonize the entire surface – likely by conditioning the surface with surfactin and using the expansive force of growth to form robust films. Perhaps of greater interest is the finding that nutritional limitations, especially for metals such as K<sup>+</sup>, Fe<sup>2+ </sup>or Mn<sup>2+</sup>, cause the surface growth pattern to revert to the tendril form, where clusters of cells exhibit organized sliding motility in distinctive patterns reminiscent of those described by many other investigators, but under very different and slow growth conditions (reviewed in [<xref ref-type="bibr" rid="B1">1</xref>]). We suggest that formation of the tendrils may represent a growth strategy for nutrient-deprived cells to colonize surfaces in natural environments, such as plant roots where large populations of <italic>B. subtilis </italic>and its relatives have been found (reviewed in [<xref ref-type="bibr" rid="B21">21</xref>]). It will be of interest to examine the controls on tendril formation and determine why the tendril phase persists and precedes the formation of surface films. These future experiments should also shed more light on the phenomenon of bacterial sliding motility.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>Strains and media for the study of sliding motility</title><p>Table <xref ref-type="table" rid="T1">1</xref> summarizes the strains used in this work. Most of the experiments described here used <italic>B. subtilis </italic>DS64, a <italic>hag </italic>mutant (<italic>hag</italic>::MLS) constructed in wild type strain 3610, using gene replacement methods described elsewhere [<xref ref-type="bibr" rid="B5">5</xref>]. DS64 was routinely maintained on Luria-Bertani (LB) medium containing (per L) 10 g tryptone, 5 g yeast extract, and 10 g NaCl, solidified with 1.5% w/v agar (Difco) and supplemented with erythromycin (1 μg ml<sup>-1</sup>) and lincomycin (25 μg ml<sup>-1</sup>; referred to as MLS for macrolide-lincosamide-streptogamin B resistance). Verification of the lack of flagella in the <italic>hag </italic>mutant included examination by flagellar staining [28] and its inability to swim on LB agar plates (0.3% w/v agar; [<xref ref-type="bibr" rid="B5">5</xref>]). For surface sliding motility, the pellicle medium MSgg described by Branda et al. [<xref ref-type="bibr" rid="B10">10</xref>] was used with some modifications: 1) the 5 mM KH<sub>2</sub>PO<sub>4 </sub>component was replaced by 5 mM NaH<sub>2</sub>PO<sub>4 </sub>and K<sup>+ </sup>ion was provided by addition of KCl as described below; 2) the metals solution, prepared as a separate 100× filter sterilized stock and stored at 4°C, was supplemented with 5 mM ascorbic acid to keep the iron component (5 mM FeSO<sub>4</sub>) from oxidizing and precipitating; and 3) thiamine, tryptophan and phenylalanine were deleted. This modified medium, MSggN, supported good planktonic growth (37°C) if 100 μM KCl was added, and weak pellicle formation in stationary cultures (30°C, 48–60 h) if 500 μM KCl was added; when supplemented with 5 mM KCl pellicle formation of strain 3610 and its <italic>hag </italic>mutant was robust with a thick wrinkled pellicle as in Branda et al. [<xref ref-type="bibr" rid="B13">13</xref>]. For sliding motility the MSggN medium, prepared with varying amounts of KCl, was solidified with 0.3% w/v agarose (Fisher Scientific; low EEO high gel strength grade). The 60 mm plates were allowed to air dry on a leveling table in a laminar flow hood overnight, and then routinely inoculated in the center with a sharpened toothpick from overnight cultures of DS 64 on LB (MLS) agar, or LB agar for its wild type parent strain 3610. In some experiments, plates were inoculated in the center from LB (MLS) broth cultures so that the inoculum (2.5 μl) contained 5 × 10<sup>3</sup>, 5 × 10<sup>4</sup>, or 5 × 10<sup>5 </sup>viable cells. Following growth at 37°C for 16–24 h the MSggN plates were photographed. Except for pellicle growth (30°C), all other growth experiments were conducted at 37°C.</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p><italic>B. subtilis </italic>strains used in this work</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left">Strain</td><td align="left">Genotype/phenotype</td><td align="left">Reference/source</td></tr></thead><tbody><tr><td align="left">3610</td><td align="left">Undomesticated wild type</td><td align="left">NCIB 3610; [10]</td></tr><tr><td align="left">JH642</td><td align="left"><italic>trpC2 pheA1</italic></td><td align="left">J. Hoch</td></tr><tr><td align="left">DS64</td><td align="left"><italic>hag</italic>::MLS</td><td align="left">[5]</td></tr><tr><td align="left">GHB1</td><td align="left">Δ(<italic>ktrAB</italic>::neo)</td><td align="left">E. Bremer</td></tr><tr><td align="left">GHB6</td><td align="left"><italic>ktrC</italic>::spec</td><td align="left">E. Bremer</td></tr><tr><td align="left">GHB12</td><td align="left"><italic>ktrD</italic>::tet</td><td align="left">E. Bremer</td></tr><tr><td align="left">M1</td><td align="left"><italic>srfAA</italic>::cat</td><td align="left">[9]</td></tr><tr><td align="left">RFH1</td><td align="left"><italic>hag</italic>::MLS <italic>srfAA</italic>::cat</td><td align="left">This work</td></tr><tr><td align="left">RFH3</td><td align="left"><italic>hag</italic>::MLS Δ(<italic>ktrAB</italic>::neo)</td><td align="left">This work</td></tr><tr><td align="left">RFH7</td><td align="left"><italic>hag</italic>:: MLS <italic>ktrC</italic>::spec <italic>ktrD</italic>::tet</td><td align="left">This work</td></tr></tbody></table></table-wrap><table-wrap position="float" id="T2"><label>Table 2</label><caption><p>Consequences of selective manipulation of MSggN media composition on surface colony formation of the <italic>hag </italic>null strain. This data shows how systematically lowering concentrations of macronutrients and micronutrients affects the growth pattern on MSggN agarose plates. In each case 5 mM KCl was present, except where KCl was varied, and plates were inoculated in triplicate to reveal the surface colony pattern (film or tendrils) after 24 h growth at 37°C.</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left">Media component</td><td align="left">1X conc.</td><td align="left">0.1X conc.</td><td align="left">0.01X conc.</td></tr></thead><tbody><tr><td align="left"><italic>Macronutrients</italic></td><td></td><td></td><td></td></tr><tr><td align="left">54 mM Glycerol</td><td align="left">Film</td><td align="left">Tendril</td><td align="left">Tendril</td></tr><tr><td align="left">30 mM Na glutamate</td><td align="left">Film</td><td align="left">Tendril</td><td align="left">Weak growth</td></tr><tr><td align="left">5 mM NaH<sub>2</sub>PO<sub>4</sub></td><td align="left">Film</td><td align="left">Tendril</td><td align="left">Tendril</td></tr><tr><td align="left"><italic>Micronutrients</italic></td><td></td><td></td><td></td></tr><tr><td align="left">5 mM KCl</td><td align="left">Film</td><td align="left">Tendril</td><td align="left">Tendril</td></tr><tr><td align="left">2 mM MgCl<sub>2</sub></td><td align="left">Film</td><td align="left">Film</td><td align="left">Tendril<sup>a</sup></td></tr><tr><td align="left">700 μM CaCl<sub>2</sub></td><td align="left">Film</td><td align="left">Film</td><td align="left">Film<sup>b</sup></td></tr><tr><td align="left">50 μM MnSO<sub>4</sub></td><td align="left">Film</td><td align="left">Film</td><td align="left">Film<sup>c</sup></td></tr><tr><td align="left">50 μM FeSO<sub>4</sub></td><td align="left">Film</td><td align="left">Film</td><td align="left">Film<sup>d</sup></td></tr><tr><td align="left">1 μM ZnSO<sub>4</sub></td><td align="left">Film</td><td align="left">Film</td><td align="left">Film<sup>b</sup></td></tr></tbody></table><table-wrap-foot><p><sup>a</sup>Tendrils were also formed at the 10<sup>-3 </sup>to 10<sup>-5 </sup>dilutions.</p><p><sup>b</sup>For these micronutrients, films were also formed at the 10<sup>-3 </sup>to 10<sup>-5 </sup>dilutions.</p><p><sup>c</sup>Films were formed at the 10<sup>-3 </sup>dilution and tendrils were formed at the 10<sup>-4 </sup>and 10<sup>-5 </sup>dilutions.</p><p><sup>d</sup>Tendrils were formed at the 10<sup>-3 </sup>to 10<sup>-5 </sup>dilutions.</p></table-wrap-foot></table-wrap><p>In some experiments, the levels of agarose, macro- or micronutrients were varied as described in the text (and Table <xref ref-type="table" rid="T2">2</xref>). To examine the monobasic cation specificity of sliding motility, the KCl component was replaced with equimolar LiCl, RbCl or CsCl. For planktonic growth of the hag mutant, MSggN broth with varying levels of KCl were used, and cell density was determined from the optical density at 600 nm.</p></sec><sec><title>Construction of srfAA, ktrAB and ktrCD mutants in the hag strain</title><p>Strains (<italic>hag </italic>genetic background) with disruptions of surfactin formation or high and low affinity K<sup>+ </sup>uptake systems, KtrAB and KtrCD, were constructed, and are listed in Table <xref ref-type="table" rid="T1">1</xref>. The surfactin mutant was derived from a <italic>srfAA </italic>mutant derived by gene disruption [<xref ref-type="bibr" rid="B9">9</xref>]. KtrAB and KtrCD, were constructed from mutants obtained from E. Bremer each in the laboratory strain JH642 [<xref ref-type="bibr" rid="B14">14</xref>]. Each allele was transferred to the <italic>hag </italic>strain (DS 64; <italic>hag</italic>::MLS) by SPP1-mediated phage transduction to obtain strains RFH1(<italic>hag</italic>::MLS <italic>srfAA</italic>::cat), RFH3 (<italic>hag</italic>::MLS Δ<italic>ktrAB</italic>::neo) and RFH7 (<italic>hag</italic>::MLS <italic>ktrC</italic>::spec <italic>ktrD</italic>::tet). Strains were typically maintained on LB agar supplemented with the appropriate antibiotics, 5 μg ml<sup>-1 </sup>chloramphenicol, 10 μg ml<sup>-1 </sup>neomycin, 10 μg ml<sup>-1</sup>tetracycline, or 100 μg ml<sup>-1 </sup>spectinomycin. Strains were periodically checked for the MLS marker of the <italic>hag </italic>mutant, and stored for longer periods frozen in 7% (v/v) DMSO at -70°C.</p></sec><sec><title>Confocal Scanning Laser Microscopy</title><p>To examine the colony morphology in both tendrils and surface films, confocal scanning laser microscopy (CSLM) was used. Surface colonies on plates or colonies transferred directly to a glass cover slip were stained with Live/Dead BacLight stains from Molecular Probes (Eugene, OR) following the general protocol provided by the manufacturer. In this case a kit 1:1 mixture of green-fluorescent SYTO 9 and red-fluorescent propidium iodide was diluted 5-fold with phosphate-buffered saline and used for staining. Stained cells were examined and images processed with a Leica TCS SP2 AOB5 microscope as described in detail in Lopez et al. [<xref ref-type="bibr" rid="B17">17</xref>]).</p></sec></sec><sec><title>Authors' contributions</title><p>RF carried out many of the aspects of the work and drafted the manuscript. DBK provided strains, guidance on further strain construction, and contributed to the final version of manuscript. TN conducted many of the experiments including the construction of the <italic>hag ktrAB </italic>and <italic>hag ktrCD </italic>triple mutants.</p></sec> |
Investigating the role of small, acid-soluble spore proteins (SASPs) in the resistance of <italic>Clostridium perfringens </italic>spores to heat | <sec><title>Background</title><p><italic>Clostridium perfringens </italic>type A food poisoning is caused by enterotoxigenic <italic>C. perfringens </italic>type A isolates that typically possess high spore heat-resistance. The molecular basis for <italic>C. perfringens </italic>spore heat-resistance remains unknown. In the current study, we investigated the role of small, acid-soluble spore proteins (SASPs) in heat-resistance of spores produced by <italic>C. perfringens </italic>food poisoning isolates.</p></sec><sec><title>Results</title><p>Our current study demonstrated the presence of all three SASP-encoding genes (<italic>ssp1</italic>, <italic>2 </italic>and <italic>3</italic>) in five surveyed <italic>C. perfringens </italic>clinical food poisoning isolates. β-Glucuronidase assay showed that these <italic>ssp </italic>genes are expressed specifically during sporulation. Consistent with these expression results, our study also demonstrated the production of SASPs by <italic>C. perfringens </italic>food poisoning isolates. When the heat sensitivities of spores produced by a <italic>ssp3 </italic>knock-out mutant of a <italic>C. perfringens </italic>food poisoning isolate was compared with that of spores of the wild-type strain, spores of the <italic>ssp3 </italic>mutant were found to exhibit a lower decimal reduction value (<italic>D </italic>value) at 100°C than exhibited by the spores of wild-type strain. This effect was restored by complementing the <italic>ssp3 </italic>mutant with a recombinant plasmid carrying wild-type <italic>ssp3</italic>, suggesting that the observed differences in <italic>D </italic>values between spores of wild-type versus <italic>ssp3 </italic>mutant was due to the specific inactivation of <italic>ssp3</italic>. Furthermore, our DNA protection assay demonstrated that <italic>C. perfringens </italic>SASPs can protect DNA from DNase I digestion.</p></sec><sec><title>Conclusion</title><p>The results from our current study provide evidences that SASPs produced by <italic>C. perfringens </italic>food poisoning isolates play a role in protecting their spores from heat-damage, which is highly significant and relevant from a food safety perspective. Further detailed studies on mechanism of action of SASPs from <italic>C. perfringens </italic>should help in understanding the mechanism of protection of <italic>C. perfringens </italic>spores from heat-damage.</p></sec> | <contrib id="A1" contrib-type="author"><name><surname>Raju</surname><given-names>Deepa</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I2">2</xref><email>rajud@onid.orst.edu</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Waters</surname><given-names>Michael</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I2">2</xref><email>mswaters@usc.edu</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Setlow</surname><given-names>Peter</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>setlow@nso2.uchc.edu</email></contrib><contrib id="A4" corresp="yes" contrib-type="author"><name><surname>Sarker</surname><given-names>Mahfuzur R</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I2">2</xref><email>sarkerm@oregonstate.edu</email></contrib> | BMC Microbiology | <sec><title>Background</title><p><italic>Clostridium perfringens </italic>is a gram-positive, spore-forming, anaerobic bacterium that has long been recognized as a significant cause of histotoxic and gastrointestinal (GI) diseases in both humans and animals [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B2">2</xref>]. <italic>C. perfringens </italic>isolates can be classified into one of five types (A through E) based upon their ability to produce alpha-, beta-, epsilon- and iota-toxin [<xref ref-type="bibr" rid="B2">2</xref>]. Although enterotoxin (CPE)-producing <italic>C. perfringens </italic>type A isolates represent <5% of global <italic>C. perfringens </italic>population, these bacteria are very important human GI pathogens, causing <italic>C. perfringens </italic>type A food poisoning and non-food-borne GI diseases [<xref ref-type="bibr" rid="B1">1</xref>]. Substantial experimental and epidemiological evidence [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B3">3</xref>] now indicates that CPE plays a major role in the development of GI symptoms in cases of <italic>C. perfringens</italic>-associated food poisoning and non-food-borne GI diseases. In addition to producing CPE, <italic>C. perfringens </italic>food poisoning isolates have the ability to form heat-resistant spores, which can survive boiling for an hour or longer [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B4">4</xref>,<xref ref-type="bibr" rid="B5">5</xref>]. The possession of high spore heat-resistance facilitates the survival of <italic>C. perfringens </italic>spores in primary food vehicles (such as meat and poultry products) where, in the presence of appropriate nutrients, dormant spores are reverted to vegetative cells and multiply, and then cause food-borne illness after consumption of these contaminated foods [<xref ref-type="bibr" rid="B1">1</xref>].</p><p>The molecular basis for <italic>C. perfringens </italic>spore heat-resistance remains unknown. However, since small, acid-soluble spore proteins (SASPs) from <italic>Bacillus subtilis </italic>can protect their spores from damage by heat and other environmental factors [<xref ref-type="bibr" rid="B6">6</xref>-<xref ref-type="bibr" rid="B8">8</xref>], it is very likely that <italic>C. perfringens </italic>SASPs play a similar role. Spores of <italic>Bacillus </italic>and <italic>Clostridium </italic>species contain a number of SASPs of molecular weight 5–7 kDa, which comprise 10–20% of the total spore protein [<xref ref-type="bibr" rid="B8">8</xref>]. These proteins are classified into two groups based on their primary sequence, an α/β-type and a ϒ-type [<xref ref-type="bibr" rid="B8">8</xref>]. The <italic>B. subtilis </italic>α/β-type SASPs are encoded by multiple genes and comprise a large protein family whose amino acid sequences are very highly conserved within and between species [<xref ref-type="bibr" rid="B8">8</xref>]. The <italic>B. subtilis </italic>α/β-type SASPs are non-specific DNA binding proteins which saturate the spore chromosome and protect spore DNA from damage caused by heat, UV radiation, and peroxidase [<xref ref-type="bibr" rid="B9">9</xref>-<xref ref-type="bibr" rid="B12">12</xref>]. Although an enormous number of studies on the role of SASPs in <italic>B. subtilis </italic>spore resistance have been carried out, nothing is known about the role of <italic>C. perfringens </italic>SASPs. However, the advance towards understanding the role of SASPs in <italic>C. perfringens </italic>spore resistance is emerging from some old studies [<xref ref-type="bibr" rid="B13">13</xref>-<xref ref-type="bibr" rid="B15">15</xref>] which revealed the existence of multiple α/β-type, but not ϒ-type, SASPs in <italic>Clostridium </italic>species. Three genes (<italic>ssp1</italic>, <italic>2 </italic>and <italic>3</italic>) coding for α/β-type SASPs have been cloned and nucleotide sequenced from <italic>C. perfringens </italic>old strains [<xref ref-type="bibr" rid="B16">16</xref>,<xref ref-type="bibr" rid="B17">17</xref>]. Interestingly, only these three <italic>ssp </italic>genes, compared to at-least 7 <italic>ssp </italic>genes in <italic>B. subtilis </italic>[<xref ref-type="bibr" rid="B8">8</xref>], were identified in the genome of <italic>C. perfringens </italic>strain 13 [<xref ref-type="bibr" rid="B18">18</xref>]. However, the presence and expression of these <italic>ssp </italic>genes in recent clinical <italic>C. perfringens </italic>food poisoning isolates and their role in spore heat-resistance remains unknown.</p><p>In the current study, we investigated the genetics and expression of <italic>ssp </italic>genes in <italic>C. perfringens </italic>clinical food poisoning isolates. Our study also prepared an isogenic <italic>ssp3 </italic>knock-out mutant of a <italic>C. perfringens </italic>type A food poisoning isolate. The heat sensitivities of spores produced by <italic>ssp3 </italic>knock-out mutant was compared with that of spores of the wild type strain. The results from these experiments now provide evidence that <italic>C perfringens </italic>SASPs play a role in spore heat-resistance.</p></sec><sec><title>Results</title><sec><title>PCR screening and comparison of <italic>ssp </italic>ORFs in <italic>C. perfringens </italic>food poisoning isolates</title><p>We first subjected our collection of <italic>C. perfringens </italic>food poisoning isolates to <italic>ssp</italic>-specific PCR analysis to determine whether most, or all, <italic>cpe</italic>-positive food poisoning isolates carry all three <italic>ssp </italic>genes (<italic>ssp1</italic>, <italic>2 </italic>and <italic>3</italic>). Control PCRs were run using template DNA prepared from a known <italic>ssp</italic>-positive <italic>C. perfringens </italic>strain 13 [<xref ref-type="bibr" rid="B18">18</xref>]. When template DNA isolated from each of our five surveyed clinical <italic>C. perfringens </italic>food poisoning isolates was subjected to this same <italic>ssp </italic>PCR analysis, PCR products of 239-, 306-, and 522-bp were invariably obtained (see Fig. <xref ref-type="fig" rid="F1">1</xref> for representative results). These results are in consistent with the known presence (Fig. <xref ref-type="fig" rid="F1">1</xref>) of all three <italic>ssp </italic>ORFs in the genome of the <italic>C. perfringens </italic>strain 13 [<xref ref-type="bibr" rid="B18">18</xref>].</p><p>The nucleotide sequencing analyzes revealed (data not shown) that no mutations or termination codons were found in the <italic>ssp1</italic>, <italic>2 </italic>or <italic>3 </italic>ORF sequences of our five surveyed isolates that encode proteins of 60-, 59- or 60-aa, respectively. The <italic>ssp1</italic>, <italic>2 </italic>or <italic>3 </italic>ORF sequences and also the upstream promoter sequences are highly conserved in all our surveyed food poisoning isolates and matches exactly with previously published sequences [<xref ref-type="bibr" rid="B16">16</xref>,<xref ref-type="bibr" rid="B18">18</xref>].</p></sec><sec><title>Evaluation of expression of <italic>ssp </italic>genes in <italic>C. perfringens </italic>food poisoning isolates</title><p>To examine the expression of <italic>ssp </italic>genes, the putative promoter regions of <italic>ssp1</italic>, <italic>2</italic>, or <italic>3 </italic>from <italic>C. perfringens </italic>food poisoning isolate SM101 was fused to <italic>E. coli gusA</italic>. These <italic>ssp-gusA </italic>fusion constructs were introduced into wild-type SM101 and β-glucuronidase (GUS) activity was measured during vegetative and sporulation growth. A promoter-less <italic>gusA </italic>construct was used as a negative control to ensure the specificity and reliability of the assay (Data not shown).</p><p>When GUS assay was performed on SM101 carrying pSG12 (<italic>ssp1</italic>-<italic>gusA</italic>), pSG22 (<italic>ssp2-gusA</italic>) or pSG32 (<italic>ssp3</italic>-<italic>gusA</italic>), no significant expression of <italic>ssp1</italic>, <italic>2 </italic>or <italic>3 </italic>was observed in vegetative cells (Fig. <xref ref-type="fig" rid="F2">2</xref>). However, SM101 carrying <italic>ssp1-</italic>, <italic>2</italic>-or <italic>3</italic>-<italic>gusA </italic>showed GUS activity during sporulation. The <italic>ssp1</italic>-<italic>gusA </italic>expression began at ~4 h after induction of sporulation, with maximum specific activity attained at ~12 h after induction of sporulation (Fig. <xref ref-type="fig" rid="F2">2</xref>). Both <italic>ssp2-gusA </italic>and <italic>ssp3-gusA </italic>fusions were expressed beginning at ~2 h after induction of sporulation, with maximum specific activity attained ~6 h after induction of sporulation (Fig. <xref ref-type="fig" rid="F2">2</xref>). Collectively, these results suggest that all three <italic>ssp </italic>genes are expressed during sporulation, but not during vegetative growth.</p><p>To analyze the sporulation-dependent expression of <italic>ssp </italic>genes further, we examined the expression of <italic>ssp1</italic>-, <italic>2</italic>- or <italic>3</italic>-<italic>gusA </italic>fusion in a <italic>spo0A </italic>mutant of <italic>C. perfringens</italic>. The rationale for using <italic>spo0A </italic>mutant is that, if <italic>ssp </italic>expression is truly dependent on sporulation, <italic>ssp-gusA </italic>fusions would not be expressed in sporulation-deficient <italic>spo0A </italic>mutant of <italic>C. perfringens </italic>[<xref ref-type="bibr" rid="B19">19</xref>]. The <italic>ssp</italic>-<italic>gusA </italic>fusion constructs pSG12 (<italic>ssp1</italic>-<italic>gusA</italic>), pSG22 (<italic>ssp2-gusA</italic>) and pSG32 (<italic>ssp3</italic>-<italic>gusA</italic>) were introduced into <italic>C. perfringens spo0A </italic>mutant strain IH101 and GUS activity was measured during sporulation growth. As shown in Fig <xref ref-type="fig" rid="F2">2</xref>, no detectable GUS activity was obtained with sporulating cultures of IH101 carrying <italic>ssp1</italic>-, <italic>2</italic>, and <italic>3</italic>-<italic>gusA </italic>fusions, confirming that the expression of <italic>ssp </italic>genes in <italic>C. perfringens </italic>are dependent on <italic>spo0A </italic>expression and sporulation.</p></sec><sec><title>SASP production by <italic>C. perfringens </italic>food poisoning isolates</title><p>Having obtained evidence that all three <italic>ssp </italic>genes are expressed in <italic>C. perfringens </italic>food poisoning isolates, we next examined whether these isolates can, in fact, produce SASPs. When acid-extracts of spores from representative <italic>C. perfringens </italic>food poisoning isolates SM101 (Fig. <xref ref-type="fig" rid="F3">3</xref>) and NCTC8239 (data not shown) were analyzed by polyacrylamide gel electrophoresis at low pH, multiple protein bands were observed. Western blot analysis, using antibodies against SspC of <italic>B. subtilis</italic>, detected SspC-specific immunoreactivity in acid-extracts of spores from SM101 and NCTC8239 (Fig. <xref ref-type="fig" rid="F3">3</xref> and data not shown). These results indicated that <italic>C. perfringens </italic>food poisoning isolates can produce SASPs.</p></sec><sec><title>Creation of <italic>ssp </italic>knock-out mutants</title><p>To evaluate the role of SASPs in spore heat-resistance, we attempted to introduce knock-out mutations into each <italic>ssp </italic>gene. The mutator plasmid pDR63, pDR27 or pMRS62, carrying mutated allele Δ<italic>ssp1</italic>::<italic>catP</italic>, Δ<italic>ssp2</italic>::<italic>catP </italic>and Δ<italic>ssp3</italic>::<italic>catP</italic>, respectively, was introduced into <italic>C. perfringens </italic>strain SM101 by electroporation, and Em<sup>r</sup>Cm<sup>r </sup>transformants were selected. As a positive control of transformation, the <italic>C. perfringens</italic>-<italic>E. coli </italic>shuttle plasmid pJIR751 encoding Em<sup>r </sup>was included. The Em<sup>r</sup>Cm<sup>r </sup>transformants were obtained at ~1000-fold less frequencies than that of pJIR751-derived Em<sup>r </sup>transformants, which was expected because mutator plasmids pDR63, pDR27 and pMRS62 have no origin of replication for <italic>C. perfringens</italic>. The single cross-over event of homologous recombination was confirmed by PCR using each <italic>ssp-</italic>specific primers (data not shown). One transformant for each mutator plasmid was grown in the absence of both Em and Cm, and a double crossover event between each wild-type <italic>ssp </italic>and the respective mutated Δ<italic>ssp</italic>::<italic>catP </italic>allele was screened by selecting colonies for Cm<sup>r </sup>and Em<sup>s </sup>phenotypes. After screening of ~2,500 colonies from SM101(pMRS62) culture, we obtained one putative clone showing a Cm<sup>r </sup>and Em<sup>s </sup>phenotype. This putative <italic>ssp3 </italic>mutant was designated as DR101. However, no Cm<sup>r </sup>and Em<sup>s </sup>clone was obtained after screening of >3,000 colonies from each of SM101(pDR27) and SM101(pDR63) culture. Our three repeated attempts using SM101(pDR63) and SM101(pDR27) transformants from three independent electroporations were failed to isolate <italic>spp1 </italic>and <italic>ssp2 </italic>mutant, respectively.</p></sec><sec><title>Confirmation of DR101 as <italic>ssp3 </italic>knock-out mutant</title><p>Inactivation of <italic>ssp3 </italic>in DR101 was first demonstrated by PCR analysis (Fig. <xref ref-type="fig" rid="F4">4A</xref>) using <italic>ssp3</italic>-specific primers. An ~4.1-kb fragment carrying <italic>ssp3 </italic>was amplified from DNA of wild-type strain SM101. In contrast, an ~5.4-kb PCR product was obtained from DNA of mutant strain DR101. These PCR results are consistent with the wild-type <italic>ssp3 </italic>gene in DR101 having been replaced with the mutated allele, carrying an extra ~1.3-kb <italic>catP</italic>-containing fragment.</p><p>Southern blot analyses showed that a single <italic>Hpa</italic>I-digested DNA fragment from wild-type strain SM101 hybridized with <italic>ssp3</italic>-specific probe (Fig. <xref ref-type="fig" rid="F4">4B</xref>). However, two hybridizing bands were observed with DNA from mutant strain, DR101 (Fig. <xref ref-type="fig" rid="F4">4B</xref>). This profile is consistent with results expected since the ~1.3-kb <italic>catP</italic>-containing fragment has an internal <italic>Hpa</italic>I site. The <italic>catP</italic>-specific probe hybridized with an ~15-kb <italic>Hpa</italic>I fragment of DR101 DNA, but as expected, no hybridizing band was observed with DNA from wild type SM101 (Fig. <xref ref-type="fig" rid="F4">4B</xref>). These results further confirmed that wild-type <italic>ssp3 </italic>gene in DR101 having been replaced with Δ<italic>ssp3</italic>::<italic>catP </italic>allele.</p></sec><sec><title>Sporulation and CPE production by <italic>ssp3 </italic>mutant</title><p>We first compared the sporulation capability of the <italic>ssp3 </italic>mutant against that of its wild-type. Both the mutant strain DR101 and wild-type strain SM101 exhibited significant sporulation, i.e., refractile endospores were visualized by phase-contrast microscopy after 8 h growth in DS medium (data not shown). When we compared the CPE producing capabilities of the wild-type strain SM101 and <italic>ssp3 </italic>knock-out mutant DR101, an ~35-kDa CPE-specific immunoreactive band was detected in Western blots of lysates prepared from sporulating cultures of both SM101 and DR101 (data not shown). Quantitative analyses demonstrated similar level of CPE production in both SM101 and DR101 (data not shown).</p></sec><sec><title>Comparison of the heat sensitivities of spores produced by <italic>C. perfringens </italic>food poisoning isolate carrying wild-type versus knock-out <italic>ssp3</italic></title><p>In order to determine whether <italic>ssp3 </italic>has any role in spore heat-resistance, we performed experiments to evaluate the heat sensitivities of spores produced by <italic>C. perfringens </italic>wild-type and <italic>ssp3 </italic>mutant. Representative thermal death curves obtained at 100°C for heat-shocked sporulating cultures of wild-type SM101 and <italic>ssp3 </italic>mutant DR101 are shown in Fig. <xref ref-type="fig" rid="F5">5</xref>. From the Fig. <xref ref-type="fig" rid="F5">5</xref> results, we calculated that spores of SM101, carrying the wild-type <italic>ssp3 </italic>gene, had a <italic>D </italic>value of ~90 min at 100°C, while spores of DR101, an isogenic <italic>ssp3 </italic>mutant, had a <italic>D </italic>value of ~60 min. To further confirm our results, three additional <italic>D </italic>values at 100°C for heat-shocked sporulating cultures of DR101 and SM101 were determined using three independent thermal death curves (data not shown). These results demonstrated that spores produced by the <italic>ssp3 </italic>mutant had, on average, approximately 0.5-fold-lower <italic>D </italic>value at 100°C than spores produced by the wild-type strain (the difference was statistically significant at <italic>P </italic>= 0.001).</p><p>In order to determine whether the observed differences in <italic>D </italic>values between spores of SM101 and spores of DR101 was due to the specific inactivation of <italic>ssp3</italic>, <italic>D </italic>value was determined at 100°C for heat-shocked sporulating cultures of complemented strain DR101(pDR18). As calculated from Fig. <xref ref-type="fig" rid="F5">5</xref>, spores produced by DR101(pDR18) had a wild-type level <italic>D </italic>value i.e., ~90 min at 100°C. This finding was confirmed by three independent experiments. Collectively, our results indicated that the reduced heat-resistance of spores produced by the <italic>ssp3 </italic>knock-out mutant was caused by the specific inactivation of <italic>ssp3</italic>.</p></sec><sec><title>Protection of plasmid DNA against DNase I</title><p>Since <italic>B. subtilis </italic>SASPs have been shown to bind double-stranded DNA and protect against DNase cleavage [<xref ref-type="bibr" rid="B20">20</xref>], we wondered whether <italic>C. perfringens </italic>SASPs can do the same. We purified total SASPs from <italic>C. perfringens </italic>SM101 (Fig. <xref ref-type="fig" rid="F3">3</xref>) and performed SASP-DNA binding assay by measuring the ability of SASPs to protect plasmid (pJIR751) DNA from DNase I digestion. As shown in Fig. <xref ref-type="fig" rid="F6">6</xref>., when pJIR751 DNA was incubated with DNase I, no DNA band was observed after agarose gel elctrophoresis (lanes 2 and 4), indicating that DNase I is enzymatically active and completely digested the pJIR751 DNA. However, when pJIR751 DNA was incubated with <italic>C. perfringens </italic>SASPs prior to DNase I treatment, discrete DNA fragments remained after DNase I treatment (lane 3). Furthermore, the incomplete digestion of SASP-treated DNA was not due to the inhibition of DNase I activity by urea because the enzyme was completely functional in the presence of 8 M urea (Fig. <xref ref-type="fig" rid="F6">6</xref>, lane 4) that was used to dissolve SASPs. Collectively, these results indicated that SASPs protected plasmid DNA against DNase I digestion. Similar results were obtained in our three independent experiments (see Fig. <xref ref-type="fig" rid="F6">6</xref> for representative results). This observation indicated that the function of SASPs in <italic>C. perfringens </italic>food poisoning isolates is similar to that in <italic>B. subtilis</italic>, i.e., they protect spores from heat-damage by binding to spore DNA.</p></sec></sec><sec><title>Discussion</title><p><italic>C. perfringens </italic>type A food poisoning, which currently ranks as the third most commonly reported food-borne disease in the United States [<xref ref-type="bibr" rid="B1">1</xref>], is caused by enterotoxigenic <italic>C. perfringens </italic>type A isolates that typically possess high spore heat-resistance [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B4">4</xref>]. The possession of high spore heat-resistance favors the survival of <italic>C. perfringens </italic>food poisoning isolates during incomplete cooking or inadequate holding of foods, which are the two major factors contributing to <italic>C. perfringens </italic>type A food-borne illness [<xref ref-type="bibr" rid="B1">1</xref>]. However, the molecular basis for <italic>C. perfringens </italic>spore heat-resistance remains unknown. In the current study, we hypothesized that SASPs produced by <italic>C. perfringens </italic>food poisoning isolates can play a role in the resistance of their spores to heat.</p><p>Our current study demonstrated the presence of all three <italic>ssp </italic>genes in a large number of clinical <italic>C. perfringens </italic>food poisoning isolates. Nucleotide sequencing revealed that the <italic>ssp </italic>ORFs in our surveyed isolates are intact, i.e., no mutations or premature termination codons were detected in the ORFs. Although previous studies [<xref ref-type="bibr" rid="B16">16</xref>,<xref ref-type="bibr" rid="B17">17</xref>] reported the presence of <italic>ssp1</italic>, <italic>2 </italic>and <italic>3 </italic>in <italic>C. perfringens </italic>laboratory strains, to our knowledge, our study is the first to compare all three <italic>ssp </italic>ORFs between a large number of clinical <italic>cpe</italic>-positive <italic>C. perfringens </italic>food poisoning isolates. The deduced amino acid sequence of SASPs from our surveyed <italic>C. perfringens </italic>food poisoning isolates are identical to that of the <italic>C. perfringens </italic>SASPs published earlier [<xref ref-type="bibr" rid="B16">16</xref>,<xref ref-type="bibr" rid="B17">17</xref>] and homologous to α/β-type SASPs of <italic>B. subtilis </italic>[<xref ref-type="bibr" rid="B8">8</xref>] indicating that the <italic>ssp </italic>ORFs present in food poisoning isolates are indeed <italic>ssp </italic>genes.</p><p>This study report evidences that the <italic>ssp </italic>genes present in <italic>C. perfringens </italic>food poisoning isolates are functional. Our GUS assay showed that <italic>ssp1</italic>, <italic>2 </italic>and <italic>3 </italic>from a food poisoning isolate SM101 were expressed during sporulation, but not during vegetative growth, and expression begins ~3–4 h after induction of sporulation. These results suggest that the mechanism of regulation of <italic>C</italic>. <italic>perfringens </italic>SASPs may be similar as that of <italic>B. subtilis</italic>. In <italic>B. subtilis</italic>, SASPs synthesis begins 3–4 h into sporulation, when all three major SASPs are synthesized in parallel [<xref ref-type="bibr" rid="B8">8</xref>]. The differential expression of <italic>ssp </italic>genes (Fig. <xref ref-type="fig" rid="F2">2</xref>) can not be explained by the differences in ribosome binding sites (rbs) because the putative rbs is highly conserved among the three <italic>ssp </italic>genes (data not shown). Further support for the sporulation-dependent expression of <italic>C. perfringens ssp </italic>genes came from our observations that no GUS activity was detected in <italic>spo0A </italic>mutant IH101 carrying <italic>ssp1</italic>-, <italic>2</italic>- or <italic>3</italic>-<italic>gusA </italic>fusion. Consistent with these expression results, our study also provides evidence that <italic>C. perfringens </italic>food poisoning isolates can, in fact, produce SASPs. The multiple protein bands obtained in our study from acid extracts of spores produced by food poisoning isolates (Fig. <xref ref-type="fig" rid="F3">3</xref>) were also observed previously [<xref ref-type="bibr" rid="B13">13</xref>] in acid extracts of spores produced by <italic>C. perfringens </italic>NCTC9268. The identity of these acid-soluble spore proteins as SASPs was confirmed by Western blot analyses using <italic>B. subtilis </italic>SspC antibody (Fig. <xref ref-type="fig" rid="F3">3</xref>). The single immunoreactive band observed in Western blot can be explained by the fact that <italic>C. perfringens ssp1</italic>, <italic>2 </italic>and <italic>3 </italic>are highly homologous (>90%) to each other and encode similar sized proteins (59–60 aa). Collectively, the presence and expression of <italic>ssp </italic>genes in <italic>C. perfringens </italic>food poisoning isolates, which possess high spore heat-resistance as indicated in a previous study [<xref ref-type="bibr" rid="B4">4</xref>], significantly strengthen the hypothesis that SASPs are associated with heat-resistance of spores produced by these isolates.</p><p>The present study's most significant finding is the presentation of the first genetic evidence that <italic>C. perfringens </italic>SASPs play a role in heat-resistance of spores produced by <italic>C. perfringens</italic>. The inactivation of <italic>ssp3 </italic>significantly affected the heat-resistance of spores produced by a food poisoning isolate SM101. Our findings that i) spores of <italic>ssp3 </italic>knock-out mutant exhibited lower heat-resistance than that of the spores of wild type and ii) reversion of this effect by complementing the mutant with a recombinant plasmid carrying wild type <italic>ssp3</italic>, provided direct genetic evidence supporting the strong linkage between the production of SASPs and the resistance of spores to heat. The slight reduction of <italic>D </italic>value in spores of <italic>ssp3 </italic>mutant compared to the <italic>D </italic>value of spores of wild-type can be explained by the presence of the functional <italic>ssp1 </italic>and <italic>ssp2 </italic>genes in <italic>ssp3 </italic>mutant strain DR101. Our DNA protection assay also supports the role of <italic>C. perfringens </italic>SASPs in spore heat-resistance by demonstrating that <italic>C. perfringens </italic>SASPs, like <italic>B. subtilis </italic>SASPs [<xref ref-type="bibr" rid="B20">20</xref>], can protect plasmid DNA from DNase I digestion. Further studies on characterization of SASP-DNA binding and the effect of this binding on plasmid topology should help in understanding the mechanism of interaction between DNA and SASPs from <italic>C. perfringens</italic>.</p><p>To our knowledge, this report represents the first successful study involving the construction of a <italic>C. perfringens ssp </italic>knock-out mutant. The greatest challenge faced in our <italic>ssp </italic>knock-out study was the lack of an easy screening method for the second cross-over event. Although in our first attempt, using the double-antibiotic selection strategy [<xref ref-type="bibr" rid="B3">3</xref>], we were able to isolate <italic>ssp3 </italic>knock-out mutant, our three similar independent attempts were failed to isolate <italic>ssp1 </italic>and <italic>ssp2 </italic>mutant. The reasons for these failures can not be explained by the lack of sufficient amount of homologous DNA in mutator plasmids [<xref ref-type="bibr" rid="B21">21</xref>] because, both pDR62 and pDR27 carry at-least 1.4-kb homologous DNA located on either side of the insertionally inactivated <italic>ssp </italic>gene that should be sufficient to allow double-reciprocal crossover event [<xref ref-type="bibr" rid="B19">19</xref>]. It is also unlikely that incorporation of <italic>catP </italic>in <italic>ssp1 </italic>or <italic>ssp2 </italic>can cause polar effect on the downstream gene whose expression is essential for the survival of <italic>C. perfringens </italic>cells because our nucleotide sequencing analyses (data not shown) demonstrated that neither <italic>ssp1 </italic>nor <italic>ssp2 </italic>forms an operon with any downstream gene in the genome of SM101. Therefore, further research is needed to identify the relevant obstacles for isolating <italic>ssp1 </italic>and <italic>ssp2 </italic>mutants.</p></sec><sec><title>Conclusion</title><p>The current study demonstrated that i) all three <italic>ssp </italic>genes are present and expressed in a large number of clinical <italic>C. perfringens </italic>food poisoning isolates, ii) <italic>ssp3 </italic>knock-out mutant of <italic>C. perfringens </italic>food poisoning isolate possess lower spore-heat resistance compared to that of its parent strain and this effect could be restored by complementing the mutant with wild-type <italic>ssp3 </italic>gene, and iii) SASPs from <italic>C. perfringens </italic>food poisoning isolate can protect DNA from DNase I digestion. Collectively, these results support our initial hypothesis that SASPs produced by <italic>C. perfringens </italic>food poisoning isolates can play a role in resistance of their spores to heat. These findings are highly significant and relevant from a food safety perspective because the possession of high spore heat-resistance favors the survival of <italic>C. perfringens </italic>food poisoning isolates in primary food vehicles (such as, meat and poultry products) contributing to <italic>C. perfringens </italic>food-borne illnesses [<xref ref-type="bibr" rid="B1">1</xref>]. Further detailed studies on mechanism of action of SASPs from <italic>C. perfringens </italic>should help in understanding the mechanism of protection of <italic>C. perfringens </italic>spores from damage caused by heat and other environmental stresses.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>Bacterial strains and growth conditions</title><p><italic>C. perfringens </italic>strains and plasmids used in this study are listed and described in Table <xref ref-type="table" rid="T1">1</xref>. Starter culture (10 ml) of each <italic>C. perfringens </italic>isolate was prepared by overnight growth at 37°C in fluid thioglycollate broth (FTG) (Difco) as described previously [<xref ref-type="bibr" rid="B22">22</xref>]. For DNA isolation, an aliquot (0.2 ml) of each FTG culture was inoculated into 10 ml of TGY broth (3% Trypticase, 2% glucose, 1% yeast extract, 0.1% cysteine [<xref ref-type="bibr" rid="B22">22</xref>]) which was then incubated at 37°C for 18 h without shaking. For selecting <italic>C. perfringens </italic>transformants carrying recombinant plasmids, <italic>C. perfringens </italic>cultures were plated on Brain Heart Infusion agar plate containing erythromycin (50 μg/ml) or chloramphenicol (20 μg/ml) and incubated at 37°C. Sporulating cultures of <italic>C. perfringens </italic>were prepared by inoculating 0.2 ml of starter FTG medium culture into 10 ml of Duncan-Strong (DS) sporulating medium [<xref ref-type="bibr" rid="B22">22</xref>], which was then incubated for 24 h at 37°C. The presence of sporulating cells in each DS medium culture was confirmed by phase-contrast microscopy [<xref ref-type="bibr" rid="B19">19</xref>,<xref ref-type="bibr" rid="B23">23</xref>].</p></sec><sec><title><italic>ssp</italic>-specific PCR analysis</title><p>Total <italic>C. perfringens </italic>DNA was isolated from the overnight TGY medium cultures using a previously described protocol [<xref ref-type="bibr" rid="B4">4</xref>,<xref ref-type="bibr" rid="B24">24</xref>]. The isolated DNA was then subjected to <italic>ssp </italic>PCR analysis using primers specific for each of the <italic>ssp1</italic>, <italic>2 </italic>or <italic>3 </italic>genes as designed based on <italic>C. perfringens </italic>strain 13 genome sequence [<xref ref-type="bibr" rid="B18">18</xref>] (Table <xref ref-type="table" rid="T2">2</xref>). Although the SASPs are multigene family proteins and the coding regions of three <italic>C. perfringens ssp </italic>are very similar, no sequence similarities were found between the flanking regions of these genes [<xref ref-type="bibr" rid="B16">16</xref>,<xref ref-type="bibr" rid="B18">18</xref>]. Therefore, each PCR primer pairs (Table <xref ref-type="table" rid="T2">2</xref>) designed from the flanking region of each of the <italic>ssp </italic>gene amplified the respective <italic>ssp </italic>ORF sequences. These PCRs utilized 100 ng of template DNA, 25 pM of each primer, 200 μM deoxynucleoside triphosphates (dNTPs) (Roche), 2.5 mM MgCl<sub>2</sub>, and 1 U of <italic>Taq </italic>DNA polymerase (Fermentas) in a total volume of 50 μl. The reaction mixture was placed in a thermal cycler (Techne) for an initial period of 5 min at 94°C, then 28 cycles, each consisting of 1 min at 94°C, 1 min at 43°C (for CPP7/CPP8 and CPP9/CPP10) or 44°C (for CPP11/CPP34), 1 min at 72°C, and followed by an additional period of extension for 10 min at 72°C. After PCR, the presence of an amplified product was analyzed by subjecting an aliquot of each PCR sample to agarose (1.5%) gel electrophoresis.</p></sec><sec><title>Cloning and sequencing of the <italic>ssp</italic>-containing fragments from various <italic>cpe</italic>-positive isolates</title><p>The DNA fragment containing <italic>ssp </italic>ORFs and an ~200 bp upstream sequence from each of five <italic>C. perfringens </italic>chromosomal <italic>cpe </italic>isolates was amplified by PCR as described above. These PCR products were then cloned into pCR<sup>®</sup>-XL-TOPO<sup>® </sup>vector using the TOPO<sup>® </sup>XL cloning kit (Invitrogen). Both strands of the <italic>ssp-</italic>containing DNA insert present in recombinant pCR<sup>®</sup>-XL-TOPO<sup>® </sup>plasmid was then sequenced using M13 forward and reverse primers.</p></sec><sec><title>Construction of <italic>gusA</italic>-fusion plasmids and β-glucuronidase assay</title><p>The <italic>gusA </italic>reporter vector pSM242 (obtained from Dr. Melville as a gift), is a derivative of an <italic>E. coli-C. perfringens </italic>shuttle vector pJIR750 [<xref ref-type="bibr" rid="B25">25</xref>], encoding chloramphenicol resistance (Cm<sup>r</sup>), contains the following features: i) four tandem terminators to minimize vector-based transcription, ii) multicloning sites located upstream of a promoterless <italic>cpe</italic>-<italic>gusA </italic>fusion, iii) the ribosome binding site and the first 13 amino acids of the <italic>cpe </italic>gene coding region to provide efficient translation [<xref ref-type="bibr" rid="B26">26</xref>] and iv) the <italic>E. coli gusA </italic>[<xref ref-type="bibr" rid="B27">27</xref>] as a transcriptional reporter element. Since our <italic>C. perfringens spo0A </italic>mutant already contained Cm<sup>r </sup>marker, for our study we first constructed an erythromycin resistant (Em<sup>r</sup>) derivative of pSM242. An ~2.5-kb <italic>EcoR</italic>I-<italic>Hin</italic>dIII fragment of pSM242 was cloned into <italic>Eco</italic>RI/<italic>Hin</italic>dIII sites of an <italic>E. coli-C. perfringens </italic>shuttle vector pJIR751 [<xref ref-type="bibr" rid="B25">25</xref>], which encodes Em<sup>r</sup>, to create pMRS127.</p><p>The PCR amplified product (~200-bp) carrying the upstream region of each <italic>ssp </italic>was first cloned into pCR<sup>®</sup>-XL-TOPO<sup>® </sup>vector using TOPO<sup>®</sup>-XL cloning kit (Invitrogen). Briefly, the DNA fragment carrying the promoter region of each of <italic>ssp1</italic>, <italic>2 </italic>or <italic>3 </italic>from SM101 (a food poisoning isolate carrying <italic>cpe </italic>on the chromosome) was amplified by PCR using primers CPP63/CPP64, CPP65/CPP66 and CPP57/CPP58, respectively (Table <xref ref-type="table" rid="T2">2</xref>). The <italic>Sal</italic>I site was incorporated in the forward and <italic>Pst</italic>I site in the reverse primers of each primer pairs. These PCR products were then cloned into pCR<sup>®</sup>-XL-TOPO<sup>® </sup>vector. The <italic>Sal</italic>I-<italic>Pst</italic>I fragments carrying the promoter regions of <italic>ssp1</italic>, <italic>2 </italic>or <italic>3 </italic>from pCR<sup>®</sup>-XL-TOPO<sup>® </sup>clones were then re-cloned into the <italic>Sal</italic>I/<italic>Pst</italic>I sites of pMRS127 to create <italic>ssp1-</italic>, <italic>2</italic>- or <italic>3-gusA </italic>fusion constructs, pSG12, pSG22 or pSG32, respectively. As a negative control, pMRS130 carrying promoter-less <italic>gusA </italic>was constructed by cloning of <italic>Xba</italic>I-<italic>Sac</italic>I fragment from pSM104 [<xref ref-type="bibr" rid="B26">26</xref>] into the <italic>Xba</italic>I/<italic>Sac</italic>I sites of pJIR751. These plasmids were then introduced into <italic>C. perfringens </italic>wild-type SM101 or <italic>spo0A </italic>mutant IH101 by electroporation [<xref ref-type="bibr" rid="B24">24</xref>] and Em<sup>r </sup>transformants were selected. The SM101 and IH101 transformants carrying <italic>ssp-gusA </italic>fusions or promoter-less <italic>gusA </italic>were grown in vegetative and sporulation conditions and tested for β-glucuronidase (GUS) activity as previously described [<xref ref-type="bibr" rid="B26">26</xref>,<xref ref-type="bibr" rid="B28">28</xref>].</p></sec><sec><title>Extraction of SASPs and Western blotting</title><p><italic>C. perfringens </italic>SASP was extracted using the protocol as previously described [<xref ref-type="bibr" rid="B29">29</xref>]. Briefly, sporulating cells of <italic>C. perfringens </italic>were sonicated and then centrifuged. The spores were washed several times with distilled water and lyophilized. The dried spores were subjected to dry rupture in a dental amalgamator (Wig-L-Bug) using 0.1 g of glass beads as an abrasive. The spore powder was resuspended in 3% acetic acid solution and then subjected to dialysis (Spectr/por<sup>®</sup>3, MWCO- 3,500, Spectrum laboratories) against 1% acetic acid solution at 4°C for at least 24 h. The sample was lyophilized and used for analysis by polyacrylamide gel electrophoresis at low pH as previously described [<xref ref-type="bibr" rid="B29">29</xref>]. The gel was run at 20 mA with the appropriate electrode polarity since SASPs are positively charged and hence run towards the cathode. The gel was stained with commassie brilliant blue (Bio-Rad) or transferred to a nitrocellulose membrane for Western blotting. The blot was probed with antibodies against <italic>B. subtilis </italic>SspC [<xref ref-type="bibr" rid="B13">13</xref>,<xref ref-type="bibr" rid="B29">29</xref>] and developed for chemiluminescence detection (Pierce) to identify immunoreactive species.</p></sec><sec><title>Construction of mutator plasmids for <italic>ssp1</italic>, <italic>2 </italic>and <italic>3</italic></title><p><italic>ssp1 </italic>mutator plasmid: A 3306-bp DNA fragment containing <italic>ssp1 </italic>ORF and ~1.5-kb each upstream and downstream region was PCR amplified from genomic DNA of SM101 using primers CPP45 and CPP139 (Table <xref ref-type="table" rid="T2">2</xref>) and cloned into pCR<sup>®</sup>-XL-TOPO<sup>® </sup>(Invitrogen) to create pDR31. The 15-bp (nucleotides 80–94 from ATG) internal deletion as well as a unique <italic>Nru</italic>I site was incorporated in <italic>ssp1 </italic>in pDR31 by site-directed mutagenesis using Quick Change Mutagenesis System (Qiagen) using primers CPP104 and CPP105 to create pDR61. Plasmid pDR62 was construced by cloning a 1.3-kb <italic>Sma</italic>I<italic>-Nae</italic>I fragment of pJIR418, carrying chloramphenicol resistance marker <italic>catP</italic>, into <italic>Nru</italic>I site of pDR61. The <italic>ssp1 </italic>mutator plasmid, pDR63, was then constructed by re-cloning the <italic>BamH</italic>I-<italic>Xho</italic>I fragment containing Δ<italic>ssp1</italic>::<italic>catP </italic>allele, from pDR62 into <italic>Bam</italic>HI/<italic>Sal</italic>I sites of suicidal plasmid pMRS104.</p><p><italic>ssp2 </italic>mutator plasmid: A 3081-bp DNA fragment containing <italic>ssp2 </italic>ORF and ~1.4-kb each upstream and downstream region was PCR amplified from genomic DNA of SM101 using primers CPP37 and CPP38 (Table <xref ref-type="table" rid="T2">2</xref>) and cloned into pCR<sup>®</sup>-XL-TOPO<sup>® </sup>(Invitrogen) to create pDR13. Plasmid pDR14 was constructed by deleting 15-bp (nucleotides 102–115 from ATG) internal <italic>ssp2 </italic>fragment as well as incorporating a unique <italic>Nru</italic>I site in <italic>ssp2 </italic>present in pDR13 by site-directed mutagenesis using Quick Change Mutagenesis System (Qiagen) using primers CPP47 and CPP48. A 1.3-kb <italic>Sma</italic>I<italic>-Nae</italic>I fragment of pJIR418, carrying chloramphenicol resistance marker <italic>catP</italic>, was cloned into <italic>Nru</italic>I site of pDR14 to create plasmid pDR26. The <italic>ssp2 </italic>mutator plasmid pDR27 was then constructed by re-cloning the <italic>Bam</italic>HI-<italic>Xho</italic>I fragment containing Δ<italic>ssp2</italic>::<italic>catP </italic>allele, from pDR62 into <italic>Bam</italic>HI/<italic>Sal</italic>I sites of suicidal plasmid pMRS104.</p><p><italic>ssp3 </italic>mutator plasmid: A 4129-bp DNA fragment, carrying the <italic>ssp3 </italic>ORF and ~1.9-kb each upstream and downstream region, was PCR amplified from genomic DNA of SM101 using primers CPP13 and CPP16 (Table <xref ref-type="table" rid="T2">2</xref>). This PCR fragment was cloned into pCR<sup>®</sup>-XL-TOPO<sup>® </sup>(Invitrogen) to create the plasmid pBH2. A unique <italic>Hpa</italic>I restriction site was incorporated into the <italic>ssp3 </italic>ORF (at position 117 from ATG) present in pBH2 by site-directed mutagenesis using Quick Change Mutagenesis System (Qiagen) using primers CPP21 and CPP22 to create pMRS60. The <italic>catP </italic>gene was then inserted into the unique <italic>Hpa</italic>I site located within the <italic>ssp3 </italic>ORF in pMRS60 by digesting pMRS60 with <italic>Hpa</italic>I and ligating with a 1.3-kb <italic>Sma</italic>I-<italic>Nae</italic>I fragment containing the <italic>catP </italic>gene from pJIR418, to create pMRS61. The <italic>ssp3 </italic>mutator plasmid pMRS62 was then constructed by re-cloning the <italic>Kpn</italic>I-<italic>Xho</italic>I fragment of pMRS61 into <italic>Kpn</italic>I/<italic>Sal</italic>I sites of suicidal plasmid pMRS104.</p></sec><sec><title>Isolation of <italic>ssp </italic>knock-out mutants</title><p>The mutator plasmids pDR63, pDR27 and pMRS62 were used to transform, by electroporation [<xref ref-type="bibr" rid="B24">24</xref>], <italic>C. perfringens </italic>isolate SM101 to Erythromycin (Em) (50 μg/ml) and Chloramphenicol (Cm) (20 μg/ml) resistance. One transformant for each plasmid was grown in TGY broth without any antibiotics and colonies sensitive to Em, but resistant to Cm were selected as previously described [<xref ref-type="bibr" rid="B3">3</xref>,<xref ref-type="bibr" rid="B19">19</xref>].</p></sec><sec><title>Southern blot analysis</title><p>The <italic>ssp3</italic>-specific DNA probe was prepared using a 4129-bp DNA fragment, carrying <italic>ssp3 </italic>and ~1.9-kb each upstream and downstream region. This DNA fragment was amplified by PCR from genomic DNA of SM101 using primers CPP13 and CPP16 (Table <xref ref-type="table" rid="T2">2</xref>). The <italic>catP </italic>probe was produced using a 517-bp <italic>Eco</italic>RV-<italic>Hpa</italic>I fragment, containing internal <italic>catP </italic>gene sequences, from pJIR418. These <italic>ssp3</italic>- and <italic>catP</italic>-containing DNA fragments were labeled using a Random Primed DNA Labeling system (Roche). Total DNA from wild type and <italic>ssp3 </italic>mutant strains was digested with <italic>Hpa</italic>I and two identical Southern blots were prepared using this digested DNA and hybridized, separately, with probes specific for the <italic>ssp3 </italic>or <italic>catP</italic>. The hybridized probe was detected using a DIG-chemiluminescence detection system utilizing CSPD<sup>® </sup>[disodium 3-(4-methoxyspiro[1]-4-yl)phenyl phosphate] ready-to-use substrate (Roche) as previously described [<xref ref-type="bibr" rid="B3">3</xref>,<xref ref-type="bibr" rid="B4">4</xref>].</p></sec><sec><title>CPE Western blot analysis</title><p><italic>C. perfringens </italic>strains grown in DS or FTG medium were sonicated until >95% of all cells were lysed (lysis was continuously monitored by phase-contrast microscopy). After sonication, each culture lysate was analyzed for the presence of CPE by Western blot analysis using a CPE antibody as previously described [<xref ref-type="bibr" rid="B4">4</xref>,<xref ref-type="bibr" rid="B22">22</xref>].</p></sec><sec><title>Determination of <italic>D </italic>values for <italic>C. perfringens </italic>spores</title><p>The heat sensitivities of <italic>C. perfringens </italic>spores was determined as described previously [<xref ref-type="bibr" rid="B4">4</xref>]. Briefly, DS medium cultures prepared and grown for 24 h as described above were heat shocked at 75°C for 20 min which killed the remaining vegetative cells and facilitated spore germination. A 0.1 ml aliquot of each heat-shocked DS medium culture was then serially diluted with sterile FTG medium and each dilution was plated onto Brain Heart Infusion (BHI) agar plates to establish the number of viable spores per milliliter of DS medium culture at the start of heating (i.e., the zero time point of the experiment).</p><p>The remainder of each heat-shocked DS medium culture was then heated at 100°C for time period ranging from 30 min to 90 min. At each time point, a 0.1-ml aliquot was withdrawn and diluted with FTG broth. The dilutions were then plated on BHI agar plates, which were incubated anaerobically at 37°C for 24 h. Colonies which developed from germinated spores that survived heating were counted to determine the number of viable colony forming unit (CFU) that were present per milliliter of each heated DS medium culture at each time point. The CFU values were then graphed to determine decimal reduction value (<italic>D </italic>value) (i.e., the time that a culture had to be kept at a given temperature to obtain a 90% reduction in viable cell numbers) for spores of each isolate tested.</p></sec><sec><title>Assay of SASP-nucleic acid binding</title><p>The binding of <italic>C. perfringens </italic>SASPs to nucleic acid was assessed by measuring the ability of SASP to protect nucleic acid from nuclease digestion as previously described [<xref ref-type="bibr" rid="B20">20</xref>]. Briefly, SASPs was prepared from <italic>C. perfringens </italic>SM101 spores as described above and quantified using the Bradford protein assay kit (Bio-Rad). 3 μg of plasmid pJIR751 DNA was incubated in 25 μl of 10 mM Tris-acetate (pH 7.0) -1 mM EDTA with 3 μg of SASPs (1:1 ratio). After 1 h at 37°C, 3 μl of DNase I buffer was added, followed by 2 μl of DNase I (Farmentas). The solution was incubated a further 10 min at 37°C, and then 100 μl of 1.25% sodium dodecyl sulfate (SDS)-25 mM EDTA was added, followed by 13 μl of 5 M NaCl and 350 μl of ethanol to precipitate the DNA. The precipitated DNA was dissolved in 10 μl of water and subjected to agarose (1.5%) gel electrophoresis. The presence or absence of DNA was observed under UV illumination after staining with ethidium bromide.</p></sec><sec><title>Statistical analyses</title><p>Statistical analyses were performed with student's <italic>t </italic>test.</p></sec></sec><sec><title>Authors' contributions</title><p>DR carried out most of the experiments, and participated in the discussions on the study design, analyses and interpretation of the data, and in the writing of the manuscript. MW carried out experiments related to isolation of SASPs from <italic>C. perfringens</italic>. PS participated as a consultant and hosted MW in his laboratory to learn basic techniques for isolation of SASPs. MRS, as a principal author of this manuscript, participated in the planning and designing of the experiments and writing of the manuscript.</p></sec> |
Status of insecticide susceptibility in <italic>Anopheles arabiensis </italic>from Mwea rice irrigation scheme, Central Kenya | <sec><title>Background</title><p>Control of the Anopheline mosquito vectors of malaria by use of insecticides has been shown to impact on both morbidity and mortality due to this disease. Evidence of insecticide resistance in different settings necessitates surveillance studies to allow prompt detection of resistance should it arise and thus enable its management. Possible resistance by <italic>Anopheles arabiensis </italic>mosquitoes from Mwea rice irrigation scheme in Central Kenya to insecticides in the four classes of insecticides approved by WHO for indoor residual spraying was investigated.</p></sec><sec sec-type="methods"><title>Methods</title><p>Susceptibility to DDT (an organochlorine), fenitrothion (an organophosphate), bendiocarb (a carbamate), lambdacyhalothrin and permethrin (both pyrethroids) was tested using standard WHO diagnostic bioassay kits. Bioassays were performed on non-blood fed mosquitoes one- to three-day old. Knockdown was recorded every 10 min and mortality 24 h post-exposure was noted.</p></sec><sec><title>Results</title><p>Mortality 24 h post-exposure was 100% for all insecticides except for lambdacyhalothrin, which averaged 99.46%. Knockdown rates at 10 min intervals were not significantly different between the Mwea population and the susceptible KISUMU strain of <italic>Anopheles gambiae </italic>sensu stricto control. The KDT<sub>50 </sub>and KDT<sub>95 </sub>values for the Mwea population were either lower than those for the control or higher by factors of no more than 2 for most comparisons and compared well with those of <italic>An. gambiae </italic>sensu lato categorized as susceptible in other studies.</p></sec><sec><title>Conclusion</title><p>These results suggest that the Mwea population of <italic>An. arabiensis </italic>is susceptible to all the insecticides tested. This implies that vector control measures employing any of these insecticides would not be hampered by resistance.</p></sec> | <contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Kamau</surname><given-names>Luna</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>lkamau@ke.cdc.gov</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Vulule</surname><given-names>John M</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>JVulule@kisian.mimcom.net</email></contrib> | Malaria Journal | <sec><title>Background</title><p><italic>Anopheles gambiae </italic>sensu stricto, <italic>Anopheles arabiensis </italic>and <italic>Anopheles funestus </italic>are the most important vectors of malaria in sub-Saharan African and occur in sympatry across most of their range [<xref ref-type="bibr" rid="B1">1</xref>]. Studies show that the use of insecticides both for Indoor Residual Spraying (IRS) programmes and in the treatment of bed nets has resulted not only in a reduction in vector population densities but also in morbidity and mortality due to malaria [<xref ref-type="bibr" rid="B2">2</xref>-<xref ref-type="bibr" rid="B4">4</xref>]. There is, however, evidence that malaria vectors are developing resistance to commonly used insecticides [<xref ref-type="bibr" rid="B5">5</xref>]. In Western Kenya, resistance was first reported in the context of Insecticide-Treated Net (ITN) use [<xref ref-type="bibr" rid="B6">6</xref>]. Although more recent studies indicate that resistance levels have increased only marginally [<xref ref-type="bibr" rid="B7">7</xref>], there is concern that continued and/or increased use of insecticides may result in increased resistance that would threaten the sustainability of this vector control strategy. Insecticide resistance is more widespread in West Africa where it has been associated with use of insecticides in public health for mosquito control and in agriculture for pesticide control [<xref ref-type="bibr" rid="B8">8</xref>-<xref ref-type="bibr" rid="B11">11</xref>]. Levels of insecticide resistance have been shown to vary even within relatively small geographical scales and during different seasons [<xref ref-type="bibr" rid="B9">9</xref>,<xref ref-type="bibr" rid="B10">10</xref>]. The dominant resistance mechanisms also vary as was observed in Guatemalan populations of <italic>Anopheles albimanus</italic>, where both insecticide resistance levels and mechanism varied within short distances [<xref ref-type="bibr" rid="B12">12</xref>]. These observations suggest the shifting nature of insecticide resistance and imply therefore that extrapolations from one circumstance to another may be misleading. Studies in Haitian populations of <italic>An. albimanus </italic>found resistance frequencies to fenitrothin to increase from 20 to 60% over a period of six months [<xref ref-type="bibr" rid="B13">13</xref>] and underscore the need for continuous insecticide resistance monitoring, even where no evidence of resistance has previously been found.</p><p>The current study presents the first report on the status of insecticide resistance/susceptibility in a rice-irrigation scheme in Central Kenya. Resistance was tested against insecticides in each of the four classes that have been approved for IRS by WHO. The results of this study will enable informed selection of insecticides for vector control programmes as well as provide baseline information essential in the monitoring of the development of insecticide resistance.</p></sec><sec sec-type="materials|methods"><title>Materials and methods</title><sec><title>Study area and insecticide use patterns</title><p>The study was carried out in Mwea area (00° 67'S, 37° 35'E) of Central Kenya. This is predominantly a rice-growing area although other crops such as beans, maize and green vegetables are grown for subsistence. Previously, rice was grown during a single growing season that extended from June to December but in recent years, different paddies are flooded intermittently during the year due to water shortages associated with the prevailing drought, thus maintaining almost all-year-round rice growing although the main growing season is still from June to December.</p><p>A survey to establish insecticide/pesticide-use patterns in the study area was conducted. This was done by administering a simple questionnaire on the pesticides used in agriculture and their concentrations and whether residents used bed nets and if they did, whether the bed nets were insecticide-treated. A total of 42 households were surveyed.</p></sec><sec><title>Specimen collection, identification and rearing</title><p>Specimens were collected both as larvae from rice paddies using standard dippers and as adults by aspiration from walls inside human dwellings. Collections were made on 4<sup>th </sup>and 5<sup>th </sup>August 2004 and again on 9<sup>th </sup>and 10<sup>th </sup>September 2004 during the dry season, which coincided with the main rice growing season and most paddies were flooded, and then again during the rainy season between 3<sup>rd </sup>– 5<sup>th </sup>May 2005. Specimens were identified as <italic>An. gambiae s.l</italic>. based on morphological characteristics [<xref ref-type="bibr" rid="B1">1</xref>]. Larvae from the different paddies were preserved live in separate bottles and transported to the insectary for rearing. The larvae were then reared into adults as follows: a single larval specimen was picked from each of the transportation bottles and placed in a rearing pan so that each pan contained just one specimen from each rice paddy. This was done to limit the chances that siblings were included in individual bioassay runs and thus obtain better estimates of population variability in insecticide susceptibility. Six pans were constituted in this manner and the resulting adults used for each of the five insecticide bioassays that were run and the sixth for the control test using untreated test paper. For the specimens collected as adults, individual field-collected females were allowed to oviposit and F1 families raised separately. Only one specimen from each of the families was used in each of the bioassays. Specimens were identified further to sibling species of the <italic>An. gambiae </italic>complex using species-specific Polymerase Chain Reaction technique [<xref ref-type="bibr" rid="B14">14</xref>] after DNA extraction by the alcohol precipitation method [<xref ref-type="bibr" rid="B15">15</xref>]. Field-collected adults were identified after they had oviposited while specimens collected as larvae were identified after the insecticide resistance bioassays were performed.</p></sec><sec><title>Insecticide susceptibility bioassays</title><p>Insecticide susceptibility assays were performed on adult non-blood fed mosquitoes one- to three-day old that were reared from field-collected larvae as described above or on F1s of field-collected adult mosquitoes. The tests were carried out using 4% DDT, 1% fenitrothion, 0.1% bendiocarb, 0.05% lambdacyhalothrin and 0.75% permethrin, the diagnostic doses recommended by WHO. The Bioassay kit, Mosquito (Adult) Diagnostic test kit WHO/VBC/81.806, was supplied by Universiti Sains Malaysia (USM), Penang, Malaysia and the assay carried out according to the accompanying instructions. Briefly, for each of the insecticides tested, mosquitoes were divided into batches between 15–25 mosquitoes and exposed to insecticide-treated papers for 1 h for DDT, bendiocarb and permethrin and for 2 h fenitrothion and lamdacyhalothrin. Insecticide knockdown effects were recorded every 10 min until 100% knockdown was observed. At the end of the exposure period, mosquitoes were transferred into tubes with untreated papers and allowed a 24 h recovery period after which mortality was recorded. Tests were accompanied by control tests where mosquitoes were exposed to papers treated only with silicone oil for 1 h or 2 h depending on the insecticide that was being tested against. Bioassays were also carried out on the <italic>An. gambiae </italic>KISUMU susceptible strain (KSM Strain). Mortality was noted 24 h post exposure as defined in the criteria for determining resistance or susceptibility to diagnostic doses of insecticide. All mosquitoes were supplied with a 6% glucose meal during the 24 h recovery period.</p></sec><sec><title>Statistical analyses</title><p>Mean mortality was determined across all batches of mosquitoes tested for a particular insecticide and the WHO [<xref ref-type="bibr" rid="B5">5</xref>] criteria used to evaluate the resistance/susceptibility status of the mosquito tested. By the said criteria, resistance is indicated by mortality rates of less than 80% 24 h after exposure to insecticide while mortality rates greater than 98% are indicative of susceptibility.</p><p>Mortality rates between 80–90% suggest the possibility of resistance that needs to be clarified. Knockdown rates at 10 min intervals for the Mwea larval and adult collections for each of the insecticides tested and for the dry and rainy season collections were compared using the paired <italic>t-</italic>test. Knockdown rates at 10 min intervals were also compared between the Mwea mosquito collections and the KSM strain using the paired <italic>t</italic>-test. Fifty and 95% knockdown times (KDT<sub>50 </sub>and KDT<sub>95 </sub>respectively) for both the Mwea collection and the KSM strain were estimated by the log-time probit model using the <italic>LdP Line</italic><sup>R </sup>software [<xref ref-type="bibr" rid="B16">16</xref>]. The fit of the probit model was assessed using chi-square distribution analysis and the Bonferroni Procedure used to determine the overall significance of multiple tests.</p></sec></sec><sec><title>Results</title><p>All households interviewed said that they had used fenitrothion as a pesticide in rice growing for at least the last ten years but did not know the concentration at which it was used. An interview with a manager at the Mwea Rice Growers Multipurpose Co-operative Society, the organization that supplies the pesticides to the farmers and through which the farmers sell their produce revealed that fenitrothion alongside carbofuran have been the pesticides in use for agricultural spraying but the use carbofuran was stopped two years prior to the study due to cost factors. Fenitrothion 50 EC is used at a concetration of 0.5% and is sprayed onto two-week old rice seedlings in the nursery and again 21–28 days after transplanting. The survey also revealed that no organized vector control programmes are available in the study area but that approximately 93% of the 42 households surveyed used bed nets. Of the total number of bed nets used, 39% were pyrethroid (deltamethrin)-treated but bed nets were not retreated after purchase. Approximately 55% of household also use either pyrethroid aerosol sprays or mosquito coils. The use of the aerosols and mosquito coils is higher during the rainy and the rice-growing season when the residents perceive that mosquito densities are high and thus also the threat of malaria.</p><p>The total numbers of field-collected specimens that were tested for each of the five insecticides are shown in the table. In addition, a total of 821 mosquitoes belonging to 49 families (family size 10–73 mosquitoes) were tested for susceptibility to lambdacyhalothrin after initial results indicated recovery after the 24 h period. All specimens tested were <italic>An. arabiensis </italic>by the specific-specific PCR assay.</p><p>Mortality, after the 24 h recovery period, was 100% for DDT, fenitrothion, bendiocarb and permethrin and for lambdacyhalothrin for the adult collection. Mortality was however slightly reduced for the lambdacyhalothrin assay with the larval collection and with the single-family samples, mortality being 99.1% ± 0.63 S.E.(for a total of 221 mosquitoes tested in 11 batches) and 99.3% ± 0.36 S.E. (for a total of 821 mosquitoes tested in 34 batches) respectively. Mortality on the KSM strain control was 100% for all the insecticides tested except for DDT (see Table <xref ref-type="table" rid="T1">1</xref>).</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>Percentage mortality and (in brackets) total number of mosquitoes test and KDT<sub>50 </sub>and KDT<sub>95 </sub>values for the different Insecticides tested</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left">Insecticide</td><td align="left">Sample</td><td align="left">% mortality(n)</td><td align="left">KDT<sub>50 </sub>(95% CI)</td><td align="left">KDT<sub>95 </sub>(95% CI)</td><td align="left">χ<sup>2</sup></td></tr></thead><tbody><tr><td align="left">DDT</td><td align="left">Mwea</td><td align="left">100 (411)</td><td align="left">25.51(23.95–27.0)</td><td align="left">49.89(46.44–54.32)</td><td align="left">3.73<sup>ns</sup></td></tr><tr><td align="left">(4%)</td><td align="left">KSM Strain</td><td align="left">99.23 (130)</td><td align="left">62.51(56.94–70.45)</td><td align="left">210.56(161.84–309.73)</td><td align="left">6.84<sup>ns</sup></td></tr><tr><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">Fenitrothion</td><td align="left">Mwea</td><td align="left">100 (405)</td><td align="left">55.02(46.81–62.39)</td><td align="left">95.1(88.89–121.83)</td><td align="left">91.28*</td></tr><tr><td align="left">(1%)</td><td align="left">KSM Strain</td><td align="left">100 (131)</td><td align="left">88.90(83.68–95.20)</td><td align="left">130.40(123.56–151.10</td><td align="left">25.38*</td></tr><tr><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">Bendiocarb</td><td align="left">Mwea</td><td align="left">100 (366)</td><td align="left">21.31(19.95–22.61)</td><td align="left">37.85(35.06–41.65)</td><td align="left">2.4<sup>ns</sup></td></tr><tr><td align="left">(0.1%)</td><td align="left">KSM Strain</td><td align="left">100 (127)</td><td align="left">14.10(13.19–15.05)</td><td align="left">22.87(20.93–25.66)</td><td align="left">0.23<sup>ns</sup></td></tr><tr><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">Lambda-C<sup>a</sup></td><td align="left">Mwea</td><td align="left">99.61(525)</td><td align="left">21.58(20.09–23.01)</td><td align="left">43.00(39.53–47.57)</td><td align="left">12.55<sup>ns</sup></td></tr><tr><td align="left">(0.05%)</td><td align="left">KSM Strain</td><td align="left">100 (119)</td><td align="left">19.78(15.61–23.39)</td><td align="left">57.42(48.96–78.73)</td><td align="left">19.78*</td></tr><tr><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">Permethrin</td><td align="left">Mwea</td><td align="left">100 (429)</td><td align="left">17.24(15.94–18.49)</td><td align="left">34.77(31.75–38.90)</td><td align="left">2.01<sup>ns</sup></td></tr><tr><td align="left">(0.75%)</td><td align="left">KSM Strain</td><td align="left">100 (123)</td><td align="left">27.75(22.31–32.88)</td><td align="left">66.96(58.66–98.11)</td><td align="left">14.5*</td></tr></tbody></table><table-wrap-foot><p><sup>a</sup>Lambdacyhalothrin</p><p>χ2 values are for the test of fit of the log-time probit model used to estimate the KDT<sub>50 </sub>and KDT<sub>95</sub>values; ns = deviations not significant; * = deviations significant, P < 0.05</p></table-wrap-foot></table-wrap><p>Percentage knockdown at 10 min intervals was not significantly different between the Mwea larval and adult collections for all insecticides tested except for bendiocarb (paired t = 2.9896, df = 4, P = 0.0404). Data for the larval and adult assays were therefore merged for all comparisons except for this insecticide for all subsequent analyses. Percentage knockdown rates at 10 min intervals were also not significantly different between the dry and the rainy season collections nor between the Mwea collections and the KSM Strain for each of the insecticides tested (paired t, P > 0.05 in all cases). Figure <xref ref-type="fig" rid="F1">1</xref> shows percentage knockdown versus exposure time for the Mwea collections and for the KSM Strain for each of the insecticides tested. The log-time probit model used to estimate KDT<sub>50 </sub>and KDT<sub>95 </sub>values did not fit the distribution of percentage knockdown with time for the fenitrothion assay for the Mwea collections or for the KSM Strain for the fenitrothion, lambdacyhalothrin and permethrin bioassays (P values for the chi-square test of heterogeneity <0.05 in each of these cases; P values were also significant for the global test). The KDT<sub>50 </sub>and KDT<sub>95 </sub>estimates in these cases were not therefore included in the comparisons as these would be unreliable although they are given in the table. All other KDT<sub>50 </sub>and KDT<sub>95 </sub>values for the Mwea population were either lower than those for the KSM Strain or only increased slightly, by factors of less than two.</p><fig position="float" id="F1"><label>Figure 1</label><caption><p><bold>Percentage knockdown against time for the Mwea <italic>An.arabiensis </italic>population and the <italic>An. gambiae </italic>KISUMU strain</bold>. The figure shows the results of insecticide resistance bioassays using diagnostic doses of each of the insecticides. Results are for mean knockdown across all batches of mosquitoes that were tested for each of the different seasons.</p></caption><graphic xlink:href="1475-2875-5-46-1"/></fig></sec><sec><title>Discussion</title><p>Overall, the results obtained in this study suggest good susceptibility of <italic>An. arabiensis </italic>in the study area to all the five insecticides tested. This means that vector control programmes employing any of these compounds either in the treatment of bed nets or other materials or for indoor residual spraying would achieve satisfactory success rates. This is especially important as <italic>An. arabiensis </italic>was the only member of the <italic>An. gambiae </italic>complex found in the study area, a finding consistent with earlier studies in the area by researchers who found this species to constitute 87.3% of all Anopheline mosquitoes collected [<xref ref-type="bibr" rid="B17">17</xref>].</p><p>Based on the WHO criteria for characterizing insecticide resistance/susceptibility, where susceptibility is defined by mortality rates greater than 98% 24 h post-exposure, no evidence for resistance to any of the insecticides tested was found. Knockdown rates at 10 mins intervals were not significantly different between the Mwea collections and the KSM susceptible strain. In addition, KDT<sub>50 </sub>and KDT<sub>95 </sub>observed in the present study compare well with those from other studies for <italic>An. gambiae s.l</italic>. populations that are categorized as susceptible [<xref ref-type="bibr" rid="B9">9</xref>,<xref ref-type="bibr" rid="B18">18</xref>,<xref ref-type="bibr" rid="B19">19</xref>]. It is interesting to note that despite the high level of compliance and long-term use of fenitrothion, the <italic>An. arabiensis </italic>mosquito population has not developed resistance to this chemical. A possible explanation is that its levels in agricultural use are below what would select for possible naturally occurring resistance in this species.</p><p>The zero or near-zero levels of insecticide resistance in <italic>An. arabiensis </italic>that were observed in the present study are similar to those recently reported from an area of long-term ITN use in Western Kenya based on the presence of the knockdown resistance (<italic>kdr</italic>) gene [<xref ref-type="bibr" rid="B7">7</xref>]. The <italic>kdr </italic>mechanism results from mutations in the voltage-gated sodium channel, the target-site for DDT and pyrethroids and is one of the two most important forms of biochemical resistance mechanisms, the other being metabolic resistance, which occurs when levels of insecticide-detoxifying enzymes are elevated or their activity modified [<xref ref-type="bibr" rid="B20">20</xref>]. Similarly low or no resistance to pyrethroid insecticides and DDT caused by the <italic>kdr </italic>mutation has been observed within the M form of <italic>An. gambiae s.s</italic>. and <italic>An. arabiensis </italic>in several West African countries despite significant levels of resistance being found within the S form of <italic>An. gambiae s.s</italic>. [<xref ref-type="bibr" rid="B8">8</xref>,<xref ref-type="bibr" rid="B9">9</xref>,<xref ref-type="bibr" rid="B18">18</xref>-<xref ref-type="bibr" rid="B21">21</xref>]. The situation was however found to be different in South Africa where significant levels of resistance to DDT in <italic>An. arabiensis</italic>, by the WHO [<xref ref-type="bibr" rid="B5">5</xref>] criteria, were observed [<xref ref-type="bibr" rid="B22">22</xref>]. Earlier studies in the Sudan also found significantly high resistance levels to malathion in <italic>An. arabiensis </italic>[<xref ref-type="bibr" rid="B23">23</xref>], suggesting that this species is not immune to the development of resistance. These differences re-emphasize the focal nature of insecticide resistance and the need to carry out situation analyses and monitoring for individual settings. In Western Kenya, for example, the frequency of the <italic>kdr </italic>gene was found to increase albeit marginally four years after the introduction of ITNs but remained unchanged in villages 20 km away [<xref ref-type="bibr" rid="B7">7</xref>]. Studies to assess the effect of longer-term use of the ITNs on resistance in this area are crucial. In Burkina Faso, resistance levels were found to vary not only between villages within 100 km of each other and between different seasons but also to different insecticides, with resistance being seen to DDT but not to permethrin [<xref ref-type="bibr" rid="B9">9</xref>]. The Western Kenya and most of the West African studies, however, assayed only for the presence of the <italic>kdr </italic>gene to the exclusion of other possible resistance mechanisms. It would be interesting to obtain data on the levels of phenotypic resistance comparison. Brogdon and McAllister [<xref ref-type="bibr" rid="B20">20</xref>] have however argued that for insecticide resistance to be a concern, the level of resistance must be high enough to compromise the efficacy of intervention programmes employing the insecticides for vector control. It is controversial though what such a level would be given that studies in Côte d'Ivoir, for example, found nets impregnated with permethrin or deltamethrin to provide good levels of protection where the frequency of the kdr allele was 94% kdr [<xref ref-type="bibr" rid="B24">24</xref>].</p></sec><sec><title>Conclusion</title><p>These findings suggest that the <italic>An. arabiensis </italic>populations from Mwea are susceptible to all the insecticides that were tested against and therefore that vector control effort utilizing any of these insecticides would not be compromised by resistance. Thus, the results obtained in this study will enable informed choice of insecticides for use in vector control programmes in the area. In addition, the data obtained will provide baseline information needed in the monitoring of the development of resistance to the insecticides arising either due to selective pressure from the use of insecticides and pesticides or through migration to the area of mosquitoes with insecticide resistance genes.</p></sec><sec><title>Authors' contributions</title><p>LK conceived and designed the study, carried out the insecticide resistance bioassays, data analysis and interpretation and prepared the manuscript. JV participated in the development of the study design, carried out interpretation of the data and provided a critical review of the manuscript.</p></sec> |
Interpreting household survey data intended to measure insecticide-treated bednet coverage: results from two surveys in Eritrea | <sec><title>Background</title><p>As efforts are currently underway to roll-out insecticide-treated bednets (ITNs) to populations within malarious areas in Africa, there is an unprecedented need for data to measure the effectiveness of such programmes in terms of population coverage. This paper examines methodological issues to using household surveys to measure core Roll Back Malaria coverage indicators of ITN possession and use.</p></sec><sec sec-type="methods"><title>Methods</title><p>ITN coverage estimates within Anseba and Gash Barka Provinces from the 2002 Eritrean Demographic and Health Survey, implemented just prior to a large-scale ITN distribution programme, are compared to estimates from the same area from a sub-national Bednet Survey implemented 18 months later in 2003 after the roll-out of the ITN programme.</p></sec><sec><title>Results</title><p>Measures of bednet possession were dramatically higher in 2003 compared to 2002. In 2003, 82.2% (95% confidence interval (CI) 77.4–87.0) of households in Anseba and Gash Barka possessed at least one ITN. RBM coverage indicators for ITN use were also dramatically higher in 2003 as compared to 2002, with 76.1% (95% CI 69.9–82.2) of children under five years old and 52.4% (95% CI 38.2–66.6) of pregnant women sleeping under ITNs. The ITN distribution programme resulted in a gross increase in ITN use among children and pregnant women of 68.3% and 48% respectively.</p></sec><sec><title>Conclusion</title><p>Eritrea has exceeded the Abuja targets of 60% coverage for ITN household possession and use among children under five years old within two malarious provinces. Results point to several important potential sources of bias that must be considered when interpreting data for ITN coverage over time, including: disparate survey universes and target populations that may include non-malarious areas; poor date recall of bednet procurement and treatment; and differences in timing of surveys with respect to malaria season.</p></sec> | <contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Eisele</surname><given-names>Thomas P</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>teisele@tulane.edu</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Macintyre</surname><given-names>Kate</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>kmacint@tulane.edu</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Yukich</surname><given-names>Josh</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>jyukich@tulane.edu</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Ghebremeskel</surname><given-names>Tewolde</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>tewolali@gemel.com.er</email></contrib> | Malaria Journal | <sec><title>Background</title><p>Insecticide-treated bednets (ITNs) have been shown to significantly reduce malaria transmission and concomitant malaria related morbidity and all-cause child mortality across a range of transmission settings, even with sustained use [<xref ref-type="bibr" rid="B1">1</xref>-<xref ref-type="bibr" rid="B6">6</xref>]. ITN use during pregnancy has also been shown to significantly reduce the prevalence of low birth weight deliveries and malaria related morbidity among pregnant women [<xref ref-type="bibr" rid="B7">7</xref>].</p><p>Efforts are currently underway to roll-out ITNs to populations across sub-Saharan Africa with increased funding from the Global Fund to Fight AIDS, Tuberculosis and Malaria, the World Bank Global Strategy and Booster Program, the President's Malaria Initiative (PMI) and the Bill and Melinda Gates Foundation among others. Accordingly, there is an unprecedented need for data to measure the effectiveness of ITN distribution programmes in terms of population coverage [<xref ref-type="bibr" rid="B8">8</xref>]. The Abuja Summit set targets for ITN coverage by 2005 of 60% among vulnerable populations at risk for malaria [<xref ref-type="bibr" rid="B9">9</xref>], while the World Health Assembly and the PMI have recently set loftier targets of 80% and 85%, respectively, for coverage among vulnerable populations by the end of the decade [<xref ref-type="bibr" rid="B10">10</xref>,<xref ref-type="bibr" rid="B11">11</xref>]. To measure progress towards such targets, the Roll Back Malaria (RBM) partnership has established a set of core indicators to measure ITN coverage at the household and individual level [<xref ref-type="bibr" rid="B12">12</xref>]. Data for these indicators will rely almost exclusively on probability household surveys such as the Demographic and Health Survey (DHS), UNICEF's Multiple Indicator Cluster Survey (MICS), the Malaria Indicator Survey, as well as other local surveys being implemented within malarious areas. Ideally, data on ITN coverage will be collected in a way that allows for longitudinal assessment of trends over time. As such, there is a need for surveys to be reliable and comparable over time and between countries.</p><p>Two household surveys implemented 18 months apart that measured ITN coverage within two malarious zobas (zones) of Eritrea are compared. Issues of interpreting cross-sectional survey data for measuring core RBM coverage indicators of ITN possession and use among children less than five years old and pregnant women are explored. Additionally, we demonstrate how an evaluation of a national programme is possible using these data from two cross sectional surveys, and present gross programme effects useful for managers and public health strategists as they focus on moving forward with the global agenda to protect populations most at risk of malaria.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>Study area</title><p>Eritrea, a country of 3.5 million, is divided into 6 administrative provinces or <italic>zobas</italic>. Malaria transmission is seasonal, concentrated primarily in the low-lying areas of three zobas, Anseba, Gash-Barka and Debub from August through October [<xref ref-type="bibr" rid="B13">13</xref>]. While transmission is moderate with population estimates of malaria parasitemia ranging from 0.4% to 6.5% [<xref ref-type="bibr" rid="B13">13</xref>], malaria has historically represented a significant public health problem, accounting for 28% of all hospital admissions and 30% of all outpatient morbidity [<xref ref-type="bibr" rid="B13">13</xref>,<xref ref-type="bibr" rid="B14">14</xref>].</p><p>The Eritrean National Malaria Control Programme (NMCP) began a scaled-up ITN distribution and re-impregnation programme at the end of 2002, focused primarily within malarious zobas. The programme was designed to provide each household with two ITNs free of charge, while providing insecticide treatment annually for all existing bednets. As part of this programme, ITNs were also distributed free of charge to all pregnant women in the programme area through antenatal clinics. It is estimated that by the end of the 2003 over 500,000 ITNs had been distributed within the programme area.</p><p>The scope of this comparison between the 2002 Eritrea DHS (EDHS) and the 2003 Bednet Survey has been limited to Anseba and Gash Barka because the sampling frames used in these two survey domains were similar (Figure <xref ref-type="fig" rid="F1">1</xref>). While nearly all sub-zobas of Debub were eligible for inclusion in the EDHS, 6 sub-zobas of Debub were excluded from the sampling frame in the Bednet Survey for security and logistical issues [<xref ref-type="bibr" rid="B15">15</xref>].</p><fig position="float" id="F1"><label>Figure 1</label><caption><p>Map of Eritrea showing the two zobas used in comparing the Eritrean Demographic and Health Survey (2002) and the Bednet Survey (2003): Anseba and Gash Barka.</p></caption><graphic xlink:href="1475-2875-5-36-1"/></fig></sec><sec><title>Eritrean demographic and health survey (2002)</title><p>The EDHS was carried out in the dry season just prior to the roll-out of the NMCP ITN distribution programme from April through June 2002 in all zobas of Eritrea. Details of the sampling design are described elsewhere [<xref ref-type="bibr" rid="B16">16</xref>]. Briefly, a two-stage cluster design, with first stage probability proportional to estimated cluster size (PPES), was used to obtain a nationally-representative probability sample of households. Within selected primary sampling units (PSUs), households were selected at second stage using systematic sampling from a complete enumeration of households. In total, 9,389 households were successfully interviewed in Eritrea in the EDHS, of which 3,138 were from Anseba and Gash Barka. Information at the household level, including the number of bednets, was obtained from the head of household. Information on bednet use among women and their children was obtained from interviews with all women of reproductive age within selected households (Anseba n = 1,450 women; Gash Barka n = 1,446 women).</p></sec><sec><title>Bednet survey (2003)</title><p>The Bednet Survey was carried out during the rainy season of September and October 2003, approximately one year after the roll-out of the NMCP ITN distribution programme. As recommended by the RBM guidelines for measuring core coverage indicators for ITN possession and use, the Bednet Survey was intentionally conducted during the rains to capture the period when members of the respective household were most likely to be <italic>using </italic>the nets [<xref ref-type="bibr" rid="B12">12</xref>]. Details of the sampling strategy have been described elsewhere [<xref ref-type="bibr" rid="B15">15</xref>]. Briefly, a modified two-stage cluster design, with PPES first stage, was used to obtain a probability sample of households within Anseba, Gash-Barka and Debub. The three zobas were treated as separate survey domains and sampled independently using equal allocation. Second stage selection was based on a modified segmentation design where PSUs were divided into segments of equal size of approximately 26 households, from which one segment per PSU was selected and all households included in the sample [<xref ref-type="bibr" rid="B17">17</xref>]. The survey yielded a total sample size for Anseba and Gash Barka of 1,559 households from 60 villages. All information for this survey, including bednet use among women and children, was obtained from an interview with the head of household. Information on all children within the household was obtained, regardless of whether their mother was a resident of the household. It should be noted that data for socio-economic status of the household was not collected as part of the Bednet Survey as it was not hypothesized to be a determinant of household ITN possession or use in this area of Eritrea because of the extremely low variability of wealth in this target population.</p></sec><sec><title>Data analysis</title><p>SAS 8.2 was used for all data management and statistical analysis. Data from the EDHS were weighted based on relative strata size and corrections to the relative size of PSUs within strata. Data from the Bednet Survey were weighted based on relative strata size only. To account for the effect of inter-cluster correlations due to the two stage cluster designs, all standard errors were estimated using the Taylor expansion method [<xref ref-type="bibr" rid="B18">18</xref>].</p><p>Design effect is used to compare the EDHS and Bednet Survey in terms of the relative loss of precision as a result of each cluster sampling method, calculated as the ratio between the standard error of each sampling design, using empirically estimated standard errors, and the standard error that would result had simple random sampling been used (SE<sub>2-stage design</sub>/SE<sub>simple random sampling</sub>). Relative standard errors are used to compare the relative precision for each point estimate between the EDHS and the Bednet Survey, calculated as the point estimate divided by its standard error.</p><p>All select demographic variables were ascertained from household respondents and are presented at the household level. A bednet is defined as any mosquito net used for sleeping, treated or untreated. To remain consistent with the RBM coverage indicators, an ITN is defined as any bednet that was either procured or dipped in insecticide within the previous 12 months (≤ 11 months) [<xref ref-type="bibr" rid="B12">12</xref>]. All bednets procured within the previous 12 months were assumed to be treated. This was rationalized as almost all new bednets distributed in the area are either pretreated or permanently treated with insecticide. It should be noted that the EDHS did not collect information on bednet treatment at the household level (only at the individual level for children and women of reproductive age), thus the proportion of households with at least one ITN could not be calculated directly. The number of months since bednet procurement and insecticide treatment were smoothed due to significant date heaping at 12 months. For all months since procurement and treatment equal to 12, half were randomly assigned to 11. This procedure was followed because for all respondents who answered "one year" to the length of time since bednet procurement or treatment, it is deemed equally likely the exact number of months to these events would have actually been greater than or less than 12 months.</p><p>Estimates for the proportion of children under five years old and pregnant women who slept under an ITN the previous night for the EDHS and Bednet Survey were constructed based on the recommended RBM coverage indicators for these target populations [<xref ref-type="bibr" rid="B12">12</xref>].</p></sec><sec><title>Ethical procedures</title><p>Ethical clearance for the study protocol was obtained from the institutional review board of Tulane University and the Ministry of Health in Eritrea.</p></sec></sec><sec><title>Results</title><p>There was little difference between the EDHS and the Bednet Survey in terms of their relative precision resulting from their different cluster sampling designs, as measured by design effect (Table <xref ref-type="table" rid="T1">1</xref>). The design effects of the two-stage cluster design used in the EDHS for select demographic and ITN coverage variables within Anseba and Gash-Barka ranged from 1.27 to 4.46, with a mean of 2.46, while the design effects of the modified two-stage cluster design with segmentation used by the Bednet Survey for these demographic variables ranged from 1.17 to 4.69, with a mean of 2.59. The relative standard errors for these demographic and ITN coverage variables ranged from 1.8%–37.5% for the EDHS and 2.3%–20.2% for the Bednet Survey.</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>Survey parameters of select demographic and ITN coverage indicators for the Eritrean Demographic and Health Survey (2002) and the Bednet Survey (2003)</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left"><bold>Indicator</bold></td><td align="left"><bold>Survey</bold></td><td align="right"><bold>Sample size</bold></td><td align="right"><bold>Proportion (unweighted)</bold></td><td align="right"><bold>Proportion (weighted)</bold></td><td align="right"><bold>Standard error</bold></td><td align="right"><bold>Design effect</bold></td><td align="center"><bold>Relative standard error (%)*</bold></td></tr></thead><tbody><tr><td align="left">Proportion households with at least 1 child ≤ 5 years old</td><td align="left">EDHS</td><td align="right">3,138</td><td align="right">0.5102</td><td align="right">0.5063</td><td align="right">0.0114</td><td align="right">1.27</td><td align="center">2.2</td></tr><tr><td></td><td align="left">Bednet Survey</td><td align="right">1,559</td><td align="right">0.5710</td><td align="right">0.5745</td><td align="right">0.0153</td><td align="right">1.17</td><td align="center">2.6</td></tr><tr><td align="left">Proportion households with ≥ 4 members</td><td align="left">EDHS</td><td align="right">3,138</td><td align="right">0.6673</td><td align="right">0.6598</td><td align="right">0.0117</td><td align="right">1.39</td><td align="center">1.8</td></tr><tr><td></td><td align="left">Bednet Survey</td><td align="right">1,559</td><td align="right">0.6459</td><td align="right">0.6469</td><td align="right">0.0152</td><td align="right">1.25</td><td align="center">2.3</td></tr><tr><td align="left">Proportion households respondent Muslim<sup>†</sup></td><td align="left">DHS</td><td align="right">2,832</td><td align="right">0.6207</td><td align="right">0.6199</td><td align="right">0.0381</td><td align="right">4.18</td><td align="center">6.2</td></tr><tr><td></td><td align="left">Bednet Survey</td><td align="right">1,559</td><td align="right">0.6421</td><td align="right">0.6385</td><td align="right">0.0570</td><td align="right">4.69</td><td align="center">8.9</td></tr><tr><td align="left">Proportion household respondents Tigrinya</td><td align="left">DHS</td><td align="right">3,138</td><td align="right">0.2954</td><td align="right">0.3059</td><td align="right">0.0364</td><td align="right">4.46</td><td align="center">11.9</td></tr><tr><td></td><td align="left">Bednet Survey</td><td align="right">1,559</td><td align="right">0.2348</td><td align="right">0.2433</td><td align="right">0.0491</td><td align="right">4.57</td><td align="center">20.2</td></tr><tr><td colspan="8"><hr></hr></td></tr><tr><td align="left">Proportion households with at least 1 bednet</td><td align="left">EDHS</td><td align="right">3,138</td><td align="right">0.4783</td><td align="right">0.4982</td><td align="right">0.0246</td><td align="right">2.76</td><td align="center">4.9</td></tr><tr><td></td><td align="left">Bednet Survey</td><td align="right">1,559</td><td align="right">0.9268</td><td align="right">0.9265</td><td align="right">0.0153</td><td align="right">2.31</td><td align="center">1.6</td></tr><tr><td align="left">Proportion households with at least 1 ITN<sup>§</sup></td><td align="left">EDHS<sup>‡</sup></td><td align="right">-</td><td align="right">-</td><td align="right">-</td><td align="right">-</td><td align="right">-</td><td align="center">-</td></tr><tr><td></td><td align="left">Bednet Survey</td><td align="right">1,559</td><td align="right">0.8287</td><td align="right">0.8217</td><td align="right">0.0243</td><td align="right">2.55</td><td align="center">3.0</td></tr><tr><td align="left">Proportion children under 5 who slept under ITN previous night<sup>§</sup></td><td align="left">EDHS</td><td align="right">2,089</td><td align="right">0.0732</td><td align="right">0.0781</td><td align="right">0.0107</td><td align="right">1.88</td><td align="center">13.7</td></tr><tr><td></td><td align="left">Bednet Survey</td><td align="right">1,488</td><td align="right">0.7655</td><td align="right">0.7605</td><td align="right">0.0314</td><td align="right">2.86</td><td align="center">4.1</td></tr><tr><td align="left">Proportion pregnant women who slept under ITN previous night<sup>§</sup></td><td align="left">EDHS</td><td align="right">279</td><td align="right">0.0466</td><td align="right">0.0435</td><td align="right">0.0163</td><td align="right">1.29</td><td align="center">37.5</td></tr><tr><td></td><td align="left">Bednet Survey</td><td align="right">78</td><td align="right">0.5256</td><td align="right">0.5238</td><td align="right">0.0723</td><td align="right">1.27</td><td align="center">13.8</td></tr></tbody></table><table-wrap-foot><p>*Standard error/proportion (weighted)</p><p><sup>†</sup>Proportion Muslim for EDHS based on women of reproductive age; proportion Muslim for Bednet Survey based on household respondent.</p><p><sup>‡</sup>No data available on bednet treatment status at household level from EDHS</p><p><sup>§</sup>Data smoothed for date heaping at 12 months for defining ITN.</p></table-wrap-foot></table-wrap><p>Results of select household demographic variables were compared to assess the overall comparability of the two surveys (Figure <xref ref-type="fig" rid="F2">2</xref>). All such variables were very similar between the EDHS and the Bednet Survey. The only variable with 95% confidence intervals that did not overlap was the proportion of households with at least one of the members a child under five years old, which was 50.6% (95% confidence interval (CI) 48.4–52.8) for the EDHS and 57.5% (95% CI 54.4–60.4) for the Bednet Survey.</p><fig position="float" id="F2"><label>Figure 2</label><caption><p>Comparison between select demographic indicators measured by the Eritrean Demographic and Health Survey (EDHS) (2002) and the Bednet Survey (Bednet Survey) (2003), Anseba and Gash Barka. Sample size (n) for all indicators from EDHS except proportion Muslim is equal to households (3,138); sample size (n) for all indicators from Bednet Survey is equal to households (1,559). Proportion Muslim for EDHS based on women of reproductive age (n = 2,832) while proportion Muslim for Bednet Survey based on household respondents.</p></caption><graphic xlink:href="1475-2875-5-36-2"/></fig><p>Measures of bednet possession were dramatically higher on the Bednet Survey (2003) compared to the EDHS (2002) within Anseba and Gash-Barka (Figure <xref ref-type="fig" rid="F3">3</xref>). The proportion of households with at least one bednet estimated by the Bednet Survey was nearly double (92.7%; 95% CI 89.7–95.6) that estimated by the EDHS (49.8%; 95% CI 45.0–54.6). Accordingly, the mean number of bednets per household estimated by the Bednet Survey (2.048; 95% CI 1.890–2.205) was more than double that of the EDHS (0.847; 95% CI 0.738–0.957), while the proportion of households with two or more bednets increased from 23.19% on the EDHS (95% CI 19.23–27.14) to 65.41% (59.94–70.88) on the Bednet Survey.</p><fig position="float" id="F3"><label>Figure 3</label><caption><p>Comparison between ITN coverage indicators measured by the Eritrean Demographic and Health Survey (2002) and the Bednet Survey (2003), Anseba and Gash Barka.</p></caption><graphic xlink:href="1475-2875-5-36-3"/></fig><p>While the RBM coverage indicator for household ITN possession was not measured on the EDHS, over three-quarters (82.2%; 95% CI 77.4–87.0) of households were estimated to possess at least one ITN from the Bednet Survey. Additionally, over half (55.16%; 95% CI 49.56–60.76) of the households sampled on the Bednet Survey owned at least two ITNs, with the mean number of ITNs per household equal to 1.752 (95% CI 1.593–1.911).</p><p>RBM coverage indicators for ITN use among children and pregnant women within Anseba and Gash-Barka were dramatically higher on the Bednet Survey in 2003 as compared to the EDHS in 2002 (Figure <xref ref-type="fig" rid="F3">3</xref>). Over three-quarters (76.1%; 95% CI 69.9–82.2) of children under five years old were estimated to have slept under an ITN the previous night on the Bednet Survey, as compared to only 7.8% (95% CI 5.7–9.9) on the EDHS, resulting in gross programme effect of 68.3% [76.1% (Bednet Survey) – 7.8% (EDHS)]. Over half (52.4%; 95% CI 38.2–66.6) of pregnant women were estimated to have slept under an ITN the previous night on the Bednet Survey, as compared to only 4.4% (95% CI 1.2–7.5) on the EDHS, resulting in a gross programme effect of 48.0% [52.4% (Bednet Survey) – 4.4% (EDHS)].</p></sec><sec><title>Discussion</title><p>This paper examines the results of two household surveys on ITN coverage indicators for household possession and use among children under five years old and pregnant women within two malarious areas of Eritrea pre and post roll-out of a full-coverage ITN distribution programme.</p><p>Based on results from the 2003 Bednet Survey, Eritrea has exceeded the Abuja targets of 60% coverage for ITN household possession (82.2%) and use among children under five years old (76.1%) within two malarious provinces. Use among pregnant women in 2003 (52.4%) increased dramatically from 2002 (5.2%) and approaches the Abuja target of 60%. However, the World Health Assembly passed a resolution in 2005 urging member states to increase coverage of such malaria control interventions among vulnerable populations to 80% by the end of the decade, thus there is still work to be done [<xref ref-type="bibr" rid="B10">10</xref>].</p><p>Household selection using segmentation at second stage appears to be a suitable alternative to the gold-standard two-stage cluster design employed by the EDHS in terms of design effect and relative precision for obtaining data for ITN coverage indicators. This may be important when survey resources are scarce, as the field implementation of the Bednet Survey was conducted for approximately US$ 50,000 for all three zobas (US$ 21 per house for total sample size of 2,340 for three zobas), compared to approximately US$ 500,000 for the EDHS (US$ 53 per house for sample size of 9,389), resulting in a savings in cost per house sampled of $US 32. However, the latter covered the entire country and included a much broader spectrum of health parameters.</p><p>Unfortunately, no data on the timing of bednet procurement, bednet type or treatment status was collected at the household level on the 2002 EDHS. This precluded the calculation of the recommended RBM coverage indicator for household ITN possession. However, of the children that slept under a bednet on the 2002 EDHS, 49.9% of the bednets had either been treated or procured within the last 12 months. Thus using this figure in conjunction with the proportion of households on the EDHS that possessed a bednet (49.8%), it can be extrapolated that 24.9% of households on the EDHS likely possessed at least one ITN in 2002. It is recognized this may represent an overestimate of household ITN possession as a bednet within a house with a child sleeping under it may be more likely to have been treated compared to a bednet within a household without a child sleeping under it. Using this estimate from 2002, the gross effect of the NMCP distribution programme is estimated to have contributed to a 57.3% increase in household ITN possession between pre and post roll-out [82.2% (Bednet Survey) – 24.9% (EDHS)].</p><p>These results demonstrate that the NMCP ITN distribution programme was effective in increasing ITN household possession and use among children under five and pregnant women in Anseba and Gash Barka. However, better estimates of the net programme effect are required to determine the exact level of the ITN coverage increase that can be attributed directly to the distribution programme. Unfortunately, multilevel analyses that account for non-programme effects on the outcomes of ITN coverage between the two surveys were not possible due to the limitations of the cross-sectional pre-post evaluation design and lack of a suitable control group due to the nature of the full coverage programme. While the two surveys were relatively comparable over time, results point to several important factors that must be considered when interpreting household survey data for assessing progress towards global targets.</p><p>It is expected that at least some of the gross programme effect observed between these surveys can be attributed to the fact that individuals would have been more likely to use ITNs during the Bednet Survey which was implemented during the rainy season when mosquito nuisance would have been more common as compared to the EDHS that was implemented in the dry season [<xref ref-type="bibr" rid="B19">19</xref>-<xref ref-type="bibr" rid="B22">22</xref>]. This season effect is at least partially illustrated by the fact that while almost twice as many houses were found to own a bednet on the Bednet Survey compared to the EDHS (92.7% compared to 49.8%, respectively), ITN use among children was nearly tenfold higher on the Bednet Survey (76.0%) compared to the EDHS (7.8%). This suggests that among houses with bednets, their members are more likely to use them during the rainy season. It is therefore recommended that household survey designs account for the effect of season on ITN use.</p><p>As global targets are intended to be measured among those at risk for malaria, inclusion of populations not at risk for malaria, such as those within heavily urban areas, highland areas or within extremely arid regions, in the numerator and denominator will likely yield biased estimates of RBM ITN coverage indicators [<xref ref-type="bibr" rid="B19">19</xref>]. The two surveys presented here cover the same geographical areas and target population and thus point estimates are relatively comparable over time. Both urban and rural populations were included in our analysis because Anseba and Gash Barka are overwhelmingly rural (> 80% within rural areas on EDHS). However, when interpreting national-level estimates of ITN coverage indicators, we recommend estimates be disaggregated by urban and rural strata at a minimum, and if possible by highland (> 300 meters) versus lowland areas where applicable. The importance of disaggregating between urban and rural populations is illustrated by the finding that in many instances, urban populations, which are typically at a much lower risk for malaria, routinely have higher bednet coverage, likely due to better access to ITN outlets and socioeconomic factors [<xref ref-type="bibr" rid="B22">22</xref>].</p><p>Recall bias of respondents within both the EDHS and Bednet Survey resulted in substantial date heaping at 12 months, or "one year", for the time since a bednet was procured and treated. This is best illustrated by the extent of date heaping on the Bednet Survey where of the 1439 households with a bednet that reported the number of months since it was procured, nearly half (47.5%) reported "one year" or 12 months, with no households stating 11 or 13 months. Accordingly, date heaping at 12 months had a significant effect on differentiating between a bednet and an ITN as defined by the RBM indicators [<xref ref-type="bibr" rid="B12">12</xref>]. Differences in resulting point estimates adjusted and unadjusted for date heaping are most dramatically illustrated by results from the Bednet Survey (Figure <xref ref-type="fig" rid="F4">4</xref>). These results demonstrate that ignoring date heaping at 12 months may lead to underestimates of ITN coverage, which may prove significant as countries attempt to reach the Abuja targets set at 60%.</p><fig position="float" id="F4"><label>Figure 4</label><caption><p>ITN coverage point estimates adjusted and unadjusted for date heaping at 12 months: Bednet Survey (2003), Anseba and Gash Barka. For adjusted estimates, for all months since procurement and treatment equal to 12, half were randomly assigned to 11.</p></caption><graphic xlink:href="1475-2875-5-36-4"/></fig></sec><sec><title>Authors' contributions</title><p>TE helped conceive this research, performed the statistical analysis and led the drafting of this manuscript. KM conceived this research and helped with drafting of this manuscript; JY helped with statistical analysis and drafting of the manuscript; TG helped conceive this research and reviewed the manuscript. All authors read and approved the final manuscript.</p></sec><sec><title>Financial support</title><p>The work presented here was funded through two projects: 1) the American Schools of Public Health/Centers for Disease Control and Prevention cooperative agreement (S1942-21/21), Enhancing National Malaria Control in Eritrea; and 2) the Environmental Health Project Contract # HRN-1-00-99-00011-00. Support also came through the Ministry of Health, Eritrea and Tulane University.</p></sec> |
A steep decline of malaria morbidity and mortality trends in Eritrea between 2000 and 2004: the effect of combination of control methods | <sec><title>Background</title><p>Malaria is a huge public health problem in Africa that is responsible for more than one million deaths annually. In line with the Roll Back Malaria initiative and the Abuja Declaration, Eritrea and other African countries have intensified their fight against malaria. This study examines the impact of Eritrea's Roll Back Malaria Programme: 2000–2004 and the effects and possible interactions between the public health interventions in use.</p></sec><sec sec-type="methods"><title>Methods</title><p>This study employed cross-sectional survey to collect data from households, community and health facilities on coverage and usage of Insecticide-Treated Nets (ITNs), Indoor Residual Spraying (IRS), larvicidal activities and malaria case management. Comparative data was obtained from a similar survey carried out in 2001. Data from the Health Management Information System (HMIS) and reports of the annual assessments by the National Malaria Control Programme was used to assess impact. Time series model (ARIMA) was used to assess association.</p></sec><sec><title>Results</title><p>In the period 2000–2004, approximately 874,000 ITNs were distributed and 13,109 health workers and community health agents were trained on malaria case management. In 2004, approximately 81% households owned at least one net, of which 73% were ITNs and 58.6% of children 0–5 years slept under a net. The proportion of malaria cases managed by community health agents rose from 50% in 1999 to 78% in 2004. IRS coverage increased with the combined amount of DDT and Malathion used rising from 6,444 kg, in 2000 to 43,491 kg, in 2004, increasing the population protected from 117,017 to 259,420. Drug resistance necessitated regimen change to chloroquine plus sulfadoxine-pyrimethamine. During the period, there was a steep decline in malaria morbidity and case fatality by 84% and 40% respectively. Malaria morbidity was strongly correlated to the numbers of ITNs distributed (β = -0.125, p < 0.005) and the amount (kg) of DDT and Malathion used for IRS (β = -2.352, p < 0.05). The correlation between malaria case fatality and ITNs, IRS, population protected and annual rainfall was not statistically significant.</p></sec><sec><title>Conclusion</title><p>Eritrea has within 5 years attained key Roll Back Malaria targets. ITNs and IRS contributed most to reducing malaria morbidity.</p></sec> | <contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Nyarango</surname><given-names>Peter M</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>mnyarango@yahoo.com</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Gebremeskel</surname><given-names>Tewolde</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>tewoldeg@moh.gov.er</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Mebrahtu</surname><given-names>Goitom</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>mebratg@moh.gov.er</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Mufunda</surname><given-names>Jacob</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>mufunda@yahoo.com</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Abdulmumini</surname><given-names>Usman</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>usmanaeri@yahoo.com</email></contrib><contrib id="A6" contrib-type="author"><name><surname>Ogbamariam</surname><given-names>Andom</given-names></name><xref ref-type="aff" rid="I4">4</xref><email>ogbamariama@moh.gov.er</email></contrib><contrib id="A7" contrib-type="author"><name><surname>Kosia</surname><given-names>Andrew</given-names></name><xref ref-type="aff" rid="I5">5</xref><email>kosiaa@afro.who.int</email></contrib><contrib id="A8" contrib-type="author"><name><surname>Gebremichael</surname><given-names>Andemariam</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>michaelghn@gemel.com.er</email></contrib><contrib id="A9" contrib-type="author"><name><surname>Gunawardena</surname><given-names>Disanayike</given-names></name><xref ref-type="aff" rid="I2">2</xref><xref ref-type="aff" rid="I6">6</xref><email>gunewarden@urctasc2.com.er</email></contrib><contrib id="A10" contrib-type="author"><name><surname>Ghebrat</surname><given-names>Yohannes</given-names></name><xref ref-type="aff" rid="I5">5</xref><email>ghebraty@afro.who.int</email></contrib><contrib id="A11" contrib-type="author"><name><surname>Okbaldet</surname><given-names>Yahannes</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>okbldety@moh.gov.er</email></contrib> | Malaria Journal | <sec><title>Background</title><p>The success in malaria eradication achieved in Europe and North America during the 19<sup>th </sup>and 20<sup>th </sup>centuries has not been replicated anywhere in Sub-Saharan Africa and most tropical countries. This is despite better scientific understanding of the biology of the vector, treatment methods and other means of malaria prevention and control. Malaria causes one million deaths annually in Africa especially among vulnerable groups of pregnant women and children under five years of age [<xref ref-type="bibr" rid="B1">1</xref>]. This can be attributed to a number of factors including inadequate preventive measures for the groups at the highest risk of contracting malaria such as pregnant women and children under five years of age and drug resistance [<xref ref-type="bibr" rid="B2">2</xref>].</p><p>While, malaria eradication has eluded most tropical countries the use of conventional information sources often underestimates the true incidence [<xref ref-type="bibr" rid="B3">3</xref>]. Resurgences of severe malaria and in recent years, recurrent epidemics invariably involving falciparum malaria have been reported from many tropical countries [<xref ref-type="bibr" rid="B4">4</xref>].</p><p>Overtime, there have been several global initiatives to control malaria. The Roll Back Malaria [<xref ref-type="bibr" rid="B5">5</xref>] and the Abuja Declaration [<xref ref-type="bibr" rid="B6">6</xref>] are the recent attempts to coordinate efforts and provide more resources to reduce the malaria burden in the world. Mostly, the strategies used aimed at primary prevention through vector control or use of personal preventive methods such as bed nets, mosquito repellants, chemoprophylaxis and finally, through effective case management and medication.</p><p>In many developing countries there is variable success in vector control using ITNs, IRS with DDT and other agents [<xref ref-type="bibr" rid="B7">7</xref>]. The factors which influence the effectiveness of malaria prevention and control include national policies, community and personal prevention, community awareness, quality of health care, facility and health personnel competence as well as effective monitoring of anti-malarial drug resistance and timely change of drug regimen when resistance occurs [<xref ref-type="bibr" rid="B8">8</xref>].</p><p>In Eritrea, 67% of the population live in malaria endemic areas. Of this population 18% are children aged five years and below and 22% are women aged between 15 and 45 years. Falciparum malaria is predominant in Eritrea and is mainly transmitted by <italic>Anopheles arabiensis </italic>[<xref ref-type="bibr" rid="B9">9</xref>], which is known to be endophilic. Eritrea is inhabited by more than 13 different species of anopheline mosquitoes all capable of spreading the disease and with varying geophysical habitats [<xref ref-type="bibr" rid="B10">10</xref>]. Also, inoculation rates have a high seasonal variability, with peak inoculation rates during the rainy season and minimal or no transmission during the dry season [<xref ref-type="bibr" rid="B11">11</xref>].</p><p>Malaria is known to negatively impact on socio-economic development of Eritrea. About 7 to 12 days are lost per episode of malaria, thus having an enormous impact on the productive labour force [<xref ref-type="bibr" rid="B12">12</xref>]. The average cost for treating an episode of uncomplicated malaria is about 2.00 USD and about 7.00 USD for severe cases [<xref ref-type="bibr" rid="B12">12</xref>]. These treatment costs are significant for a country with a per capita GDP below 200 USD. Also, in 1999, malaria accounted for 31.5% of the total outpatient morbidity and 28.4% of all admissions. Malaria was responsible for 19.6% of hospital admissions among children under five years of age.</p><p>In view of the public health importance of the burden of malaria in Eritrea, in 1999, the Ministry of Health organized a national workshop on Roll Back Malaria to develop control strategies and to launch a 5-year Malaria Control Programme. The outcome of the workshop was a national resolve to control malaria as contained in the Mandefera Declaration and the plan of action for the period 2000–2004. The core objectives of the plan were to reduce malaria morbidity and mortality by 80% from the 1999 levels [<xref ref-type="bibr" rid="B13">13</xref>]. During the period 1995–1998, malaria control activities had succeeded in developing policies and guidelines, and the training of health professionals on malaria control and treatment resulting into 1200 active community agents, distribution of 81,036 ITNs and 76,209 houses were subjected to IRS [<xref ref-type="bibr" rid="B13">13</xref>].</p><p>Implementation of the plan of action commenced immediately at the beginning of 2000 with financing from the Eritrean Government, World Bank, WHO, UNICEF, USAID and later Global Funds. During the same year, the country adopted the Abuja Declaration targets and goals for the purposes of programme management. The Abuja targets aimed at reducing malaria burden by at least 60%, and ensuring that at least 60% persons suffering from malaria had access to prompt treatment using anti-malarial drugs. Further, at least 60% of persons at risk of malaria, particularly children under five and pregnant women will benefit from a suitable combination of personal and community protective measures such as insecticide treated nets (ITNs). Finally, the Abuja declaration requires that at least 60% of pregnant women who are at risk of malaria would have access to intermittent preventive treatment through use of SP [<xref ref-type="bibr" rid="B6">6</xref>].</p><p>This study is aimed at assessing the 5-year achievements of the Roll Back Malaria Programme in Eritrea, 2000–2004. The specific objectives of this study were to assess trends in malaria morbidity and mortality rates in the country, and the effectiveness of the various public health measures used in Eritrea's Malaria Control Programme. The assessment indicators were derived from Roll Back Malaria Initiative, the Abuja Declaration and Eritrea's Mandefera Declaration [<xref ref-type="bibr" rid="B12">12</xref>-<xref ref-type="bibr" rid="B14">14</xref>].</p></sec><sec sec-type="methods"><title>Methodology</title><sec><title>Study population</title><p>Eritrea is geographically divided into six regions or <italic>zobas</italic>. There are four lowland zobas, two of them coastal and to the east of the country (altitude: 0–1,000 m) and the other two in the west of the country (altitude: 600–1,000 m). The eastern and western lowlands stride the remaining two highland <italic>zobas </italic>that are centrally located in the country (altitude: 1,500–2,000 m). The country is stratified on the basis of malaria risk based on the underlying geo-physical stratification and seasonality, as well as yearly probability of cases [<xref ref-type="bibr" rid="B9">9</xref>] (incidence). Four <italic>zobas </italic>(Anseba, Debub, Gash Barka and Northern Red Sea) meet the classification for high malaria risk regions, representing 67% of the population and almost four-fifths of the landmass.</p></sec><sec><title>Programmeme strategies for NMCP 1999–2004</title><p>The foci of the Eritrea malaria control plan were: primary prevention through selective vector control and behavioural practices, mortality and morbidity reduction through effective case management and epidemic control, health systems strengthening, and implementation of an effective information and communication strategy. Primary prevention focused on vector control and personal behaviour change specifically through the consistent use of bed nets. Vector control was done at two levels of source reduction through larval control and through adult mosquito control using indoor residual spraying (IRS). The changes in behavioural practices to prevent mosquito bites and to control adult mosquitoes consisted of distribution, re-treatment and use of ITNs. For the high-risk <italic>zobas</italic>, the national policy applies equity weightings so as to bias resource allocation in the favour of these <italic>zobas</italic>. For example one free ITN is to be issued to every pregnant woman or child who is resident in any of the high-risk zobas. This is to target interventions and to guarantee access of an essential preventive tool to the most vulnerable groups who, while being at risk of infection, would otherwise not afford this intervention.</p><p>To strengthen effective case management, training of health personnel for community and health facility care, including technicians, continued throughout the plan period. Integrated Management of Childhood Illnesses (IMCI) was subsequently adopted as a means of addressing malaria for children [<xref ref-type="bibr" rid="B15">15</xref>].</p><p>Malaria surveillance monitored four factors of monthly rainfall, the weekly number of new malaria cases in relation to the set epidemic thresholds, and resistance to insecticides commonly used for spraying or treatment of nets as well as resistance to chloroquine and the newly introduced first line combination treatment.</p></sec><sec><title>Case definition</title><p>Malaria is diagnosed using direct microscopy in hospitals or health centres and by use of Rapid Diagnostic Test Kit in lower level facilities. Clinical diagnosis is the method used for case definition by community agents, drug vendors and community members.</p></sec><sec><title>Study design</title><p>The study design was composed of a retrospective component and a cross-sectional survey. In the retrospective component, data was obtained from Health Management Information System (HMIS) of the Ministry of Health, quarterly and annual reports of the National Malaria Control Programme (NMCP) and reports of the annual assessment workshops by NMCP. The Ministry collects HMIS data from all health facilities in the country, and annually the HMIS data achieved over 90% completeness for the period 2000–2004 [<xref ref-type="bibr" rid="B16">16</xref>]. In addition, data was obtained from published reports of the midterm evaluation of the Roll Back Malaria programme 1999–2004 that was carried out in 2001. The mid-term survey and the end of programme survey used the same data collecting instruments and study design. Rainfall data was obtained from the Ministry of Agriculture, Eritrea. The monthly rainfall data was collected from 22 meteorological stations located in all the six zobas of the country. Lastly, interviews on programme design and implementation process were carried out and information obtained from the NMCP offices and <italic>zoba </italic>malaria coordinators.</p><p>The study was a three-part one composed of desk review, health facility and community survey. The health facility survey generated data on training, availability of equipment, drugs and other supplies essential for diagnosis and treatment of malaria as well as quality of care. Data for the cross sectional study was obtained during the final evaluation of the 5-year malaria control programme conducted during the period September-December 2004, which coincided with the peak malaria transmission period.</p><p>The desk review component assessed programme management and impact of interventions using data from the HMIS and the NMCP. In the community survey the information gathered included ownership and use of ITNs, health seeking behaviour and community participation in IRS and ecological interventions.</p><p>The methodology for the health facility and community survey is described below.</p></sec><sec><title>Sampling</title><sec><title>Health facility survey</title><p>The health facility survey was conducted in 15% of health facilities in all the four malaria endemic zobas of the country (Gash-Barka, Debub, Anseba and Nothern Red Sea). A total of 28 facilities were randomly selected and included in the facility survey. Considering the seasonal nature of malaria, the survey was conducted over a period of two weeks during the months of October and November, which is the peak malaria transmission season in the four <italic>zobas</italic>. The target population was the general population with a bias towards children under-five years of age, pregnant women and children over five who report to health facilities with fever/malaria during the survey period. All children under five years of age coming to the health facility with fever, during the survey period were included in the sample. The sample frame was made up of a list of all health facilities obtained from the HMIS. Facilities were selected by type, i.e hospitals, health centers and health stations within each of the 4 <italic>zobas</italic>. At least one hospital in each <italic>zoba </italic>was included in the sample.</p><p>The object of the health survey was to assess the quality of care in the health facilities of the endemic zones through direct observation and exit interview. All individuals who presented to the selected health facilities with malaria or fever during the two-week period of the survey were to be observed at the stage of consultation and interviewed at exit time. During the observational stage trained research assistants, using a checklist, recorded the practice of clinical skills: whether clinicians asked caregivers or patients about fever, took temperature measurements, requested for blood tests for malaria, the drugs that were prescribed and if patients or caregivers received counseling.</p></sec><sec><title>Community survey</title><p>The community component of the survey was designed to cover the population residing in the same four high malaria endemicity <italic>zobas</italic>. The sample size calculation was based on input from the National Statistics and Evaluation Office (NSEO) that aimed to generate reliable estimates on malaria indicators for the different segments of the population, (children under 5 years of age, persons aged 5 years and above, and pregnant women). The proportion of pregnant women who slept under the mosquito bed net on the previous night (7.6%) rounded to 8.0% (p = 0.08), based on the 2002 Eritrea Demographic and Health Survey (EDHS)[<xref ref-type="bibr" rid="B17">17</xref>], was used as the starting point of the calculations, with an assumed relative error of 20% (C.V = 0.20).</p><p><bold>n = q/((C.V)</bold><sup>2</sup><bold>*p); where q = 1-p</bold></p><p>n = 0.92/(0.2*0.2*0.08) = 288</p><p>This implies that the survey required a sample size of 288 pregnant women for the four zobas combined together. In the EDHS 2002, a household had on an average of about 0.2 pregnant women. This made the number of households to be selected in this sample to be about 1440 (i.e., 288/0.2).</p><p>Moreover, a response rate of 95% was obtained in the 2002 EDHS for the household survey. Adjustment for this response rate made the sample of households to be covered about 1,516 (1440/0.95) and rounded to 1,520 households. The total of 1,520 households (HHs) was allocated equally among the 4 <italic>zobas </italic>to get equal precision for each <italic>zoba</italic>.</p><p>The sample design adopted for the survey was a two stage stratified cluster design. At the first stage, clusters (rural/urban) were selected as Primary Sampling Units (PSU) and at the second stage households were selected as Secondary Sampling Units (SSU). The domain of the study was all the four zobas combined together.</p><p>In each <italic>zoba </italic>a total of 38 (i.e. 380/10) clusters (villages or urban areas) were selected. The survey aimed to reach 10 households in each cluster. The frame used for this purpose was the list of villages and urban clusters prepared by the respective <italic>zoba </italic>administration offices in 2004.</p></sec></sec><sec><title>Statistical analysis</title><p>Malaria incidence rates and case fatality rates in the country were computed covering the period 2000–2004. The Time Series analysis was used to test for any association between morbidity and/or mortality being dependent variables and the intervention measures as independent variables. The aggregated data from all six <italic>zobas </italic>of the country was used in testing for the impact of the various public health interventions.</p></sec><sec><title>Limitations</title><p>The data set on interventions, malaria morbidity and mortality covers only a five-year period, which is too short to generalise on long-term trends and provide adequate statistical power in all instances of regression analysis and for each of the variables of interest. However, on behaviour change, these shortcomings are corrected for through comparative analysis of cross-sectional data from the same study population and using similar tools, collected at two intervals in 2001 and 2004. The sample size calculation for the community and facility surveys was designed realise validity at the national level and not the zoba level.</p></sec></sec><sec><title>Results</title><sec><title>Sample size coverage</title><p>All the selected health facilities were reached and a total of 231 patients or caretakers were interviewed. Direct observations during clinical management at the health facility were made on the same number of patients. Out of the estimated sample of 288 pregnant women, 238 (82.6%) were identified during the community survey (Table <xref ref-type="table" rid="T1">1</xref>).</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>Expected sample sizes and samples size achieved per <italic>zoba</italic></p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left"><bold>Zone (Zoba)</bold></td><td align="center"><bold>Number of households</bold></td><td align="center"><bold>Sick children under five years</bold></td><td align="center"><bold>Pregnant women</bold></td><td></td></tr><tr><td></td><td colspan="4"><hr></hr></td></tr><tr><td></td><td align="center"><bold>Minimum sample size expected</bold></td><td align="center"><bold>Sample size reached</bold></td><td align="center"><bold>Minimum sample size expected</bold></td><td align="center"><bold>Sample size reached</bold></td></tr></thead><tbody><tr><td align="left">Anseba</td><td align="center">380</td><td align="center">71</td><td align="center">72</td><td align="center">64</td></tr><tr><td align="left">Debub</td><td align="center">380</td><td align="center">48</td><td align="center">72</td><td align="center">61</td></tr><tr><td align="left">Gash-Barka</td><td align="center">380</td><td align="center">42</td><td align="center">72</td><td align="center">55</td></tr><tr><td align="left">Northern Red Sea</td><td align="center">380</td><td align="center">35</td><td align="center">72</td><td align="center">58</td></tr><tr><td colspan="5"><hr></hr></td></tr><tr><td align="left"><bold>TOTAL</bold></td><td align="center"><bold>1520</bold></td><td align="center"><bold>196</bold></td><td align="center"><bold>288</bold></td><td align="center"><bold>238</bold></td></tr></tbody></table></table-wrap></sec><sec><title>Findings from review of HMIS data</title><p>Analysis of data from the HMIS, revealed that during the period 2000–2004, the incidence rate of malaria and case fatality rate in Eritrea declined precipitously. The overall outpatient malaria incidence rate dropped by 83.33% while the malaria case fatality rate decreased from 0.21% to 0.14% (Figure <xref ref-type="fig" rid="F1">1</xref>). An increase in case fatality rate trends during the period 2000–2002 was subsequently followed by marked decline to below the 1999 levels.</p><fig position="float" id="F1"><label>Figure 1</label><caption><p>Malaria morbidity and case fatality rates 1999–2004.</p></caption><graphic xlink:href="1475-2875-5-33-1"/></fig><p>The reduction in morbidity and case fatality rate varied across the different zobas with the greatest reduction occurring in Anseba (Table <xref ref-type="table" rid="T2">2</xref>).</p><table-wrap position="float" id="T2"><label>Table 2</label><caption><p>Trends in malaria incidence rate (IR) and case fatality rate (CFR) by zoba</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="center">Year</td><td align="center" colspan="2">Anseba</td><td align="center" colspan="2">Debub</td><td align="center" colspan="2">G. Barka</td><td align="center" colspan="2">Maekel</td><td align="center" colspan="2">Northern Red Sea</td><td align="center" colspan="2">Southern Red Sea</td></tr></thead><tbody><tr><td align="left">IR</td><td align="left">CFR</td><td align="left">IR</td><td align="left">CFR</td><td align="left">IR</td><td align="left">CFR</td><td align="left">IR</td><td align="left">CFR</td><td align="left">IR</td><td align="left">CFR</td><td align="left">IR</td><td align="left">CFR</td><td></td></tr><tr><td colspan="13"><hr></hr></td></tr><tr><td align="center">1999</td><td align="right">1938.7</td><td align="right">0.3</td><td align="right">2198.0</td><td align="right">0.2</td><td align="right">7696.5</td><td align="right">0.2</td><td align="right">461.6</td><td align="right">0.1</td><td align="right">2270.5</td><td align="right">0.3</td><td align="right">940.7</td><td align="right">1.3</td></tr><tr><td align="center">2000</td><td align="right">830.1</td><td align="right">0.4</td><td align="right">1223.2</td><td align="right">0.2</td><td align="right">4462.1</td><td align="right">0.2</td><td align="right">190.8</td><td align="right">0.4</td><td align="right">1717.9</td><td align="right">0.2</td><td align="right">1059.8</td><td align="right">0.9</td></tr><tr><td align="center">2001</td><td align="right">686.4</td><td align="right">0.7</td><td align="right">1196.2</td><td align="right">0.2</td><td align="right">4636.6</td><td align="right">0.2</td><td align="right">360.2</td><td align="right">0.2</td><td align="right">2523.4</td><td align="right">0.2</td><td align="right">1082.2</td><td align="right">0.4</td></tr><tr><td align="center">2002</td><td align="right">327.9</td><td align="right">0.4</td><td align="right">754.3</td><td align="right">0.2</td><td align="right">2546.8</td><td align="right">0.4</td><td align="right">272.4</td><td align="right">0.1</td><td align="right">1053.8</td><td align="right">0.1</td><td align="right">1018.5</td><td align="right">0.1</td></tr><tr><td align="center">2003</td><td align="right">332.7</td><td align="right">0.7</td><td align="right">1031.8</td><td align="right">0.3</td><td align="right">2691.3</td><td align="right">0.2</td><td align="right">253.7</td><td align="right">0.3</td><td align="right">820.9</td><td align="right">0.1</td><td align="right">825.6</td><td align="right">0.2</td></tr><tr><td align="center">2004</td><td align="right">57.0</td><td align="right">0</td><td align="right">503.2</td><td align="right">0.1</td><td align="right">1318.0</td><td align="right">0.1</td><td align="right">124.5</td><td align="right">0.3</td><td align="right">328.1</td><td align="right">0.4</td><td align="right">490.0</td><td align="right">0</td></tr></tbody></table></table-wrap><p>According to the HMIS data, the number of breeding sites in the country eliminated annually by filling water pools increased by 70% (Table <xref ref-type="table" rid="T3">3</xref>). Similarly, countrywide larvicidal control efforts were carried out with the proportion of sites treated annually rising more than fivefold. The number of houses covered through IRS trebled with the estimated number of people protected using this method rising to twice the number in 1999. Studies on vector resistance to DDT and other commonly used insecticides revealed high efficacy throughout the period. No resistance was reported.</p><table-wrap position="float" id="T3"><label>Table 3</label><caption><p>Malaria control activities in all <italic>zobas </italic>of Eritrea, 2000–2004</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td align="left"><bold>2000</bold></td><td align="left"><bold>2001</bold></td><td align="left"><bold>2002</bold></td><td align="left"><bold>2003</bold></td><td align="left"><bold>2004</bold></td></tr></thead><tbody><tr><td align="left">Number of houses sprayed</td><td align="left">39,838</td><td align="left">76,754</td><td align="left">60,433</td><td align="left">97,069</td><td align="left">92,107</td></tr><tr><td align="left">Malathion used (kg)</td><td align="left">2,399</td><td align="left">7,904</td><td align="left">5,555</td><td align="left">21,890</td><td align="left">30,388</td></tr><tr><td align="left">DDT used (kg)</td><td align="left">4,045</td><td align="left">8,362</td><td align="left">8,500</td><td align="left">17,423</td><td align="left">13,103</td></tr><tr><td align="left">Population protected by IRS</td><td align="left">117,017</td><td align="left">202,652</td><td align="left">159,551</td><td align="left">227,675</td><td align="left">259,420</td></tr><tr><td align="left">Breeding sites filled (pools)</td><td align="left">15,988</td><td align="left">23,810</td><td align="left">25,355</td><td align="left">22615</td><td align="left">27494</td></tr><tr><td align="left">Breeding sites treated (pools)</td><td align="left">11,691</td><td align="left">7,690</td><td align="left">12,547</td><td align="left">67,684</td><td align="left">33,442</td></tr><tr><td align="left">Abate (Temephos) used (litres)</td><td align="left">14.9</td><td align="left">18.5</td><td align="left">145.0</td><td align="left">90.5</td><td align="left">80.2</td></tr><tr><td align="left">Population participating in treating and filling breeding site</td><td align="left">54,219</td><td align="left">72,824</td><td align="left">51,666</td><td align="left">48,948</td><td align="left">111,494</td></tr><tr><td align="left">ITNs distributed</td><td align="left">127,863</td><td align="left">67,708</td><td align="left">276,038</td><td align="left">187,709</td><td align="left">214,752</td></tr></tbody></table></table-wrap><p>During the period 2000–2004, approximately 874,070 free ITNs were distributed to pregnant women and children in the four high malaria transmission regions. Results from the community survey data revealed that 23.7% of the ITNs in the community were obtained through out of pocket household purchase. The rate of the annual increase in the distribution of ITNs is strongly correlated to the declining trend in malaria morbidity. The largest decrease in the incidence rate of malaria was during the year 1999. Thereafter, the decrease in malaria morbidity was a mirror image of the annual distribution of ITNs (Figure <xref ref-type="fig" rid="F2">2</xref>).</p><fig position="float" id="F2"><label>Figure 2</label><caption><p>Distribution of ITNs and trends in malaria incidence rate.</p></caption><graphic xlink:href="1475-2875-5-33-2"/></fig><p>During the period 2000–2004, there was a steady increase in the total number of personnel trained annually (Table <xref ref-type="table" rid="T4">4</xref>). The role of community health agents played in case management gained prominence during the study period. The proportion of cases managed by CHAs rose from approximately 50% in 1999 to 78% in 2004.</p><table-wrap position="float" id="T4"><label>Table 4</label><caption><p>Number of personnel trained in case management</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="center">Year</td><td align="left">Community Health Agents</td><td align="left">Health workers</td><td align="left">Laboratory Technicians</td><td align="left">Rural drug Vendors</td><td align="left">Military health personnel</td><td align="left">Community members</td></tr></thead><tbody><tr><td align="center">2000</td><td align="left">936</td><td align="left">370</td><td align="left">15</td><td align="left">0</td><td align="left">593</td><td align="left">1121</td></tr><tr><td align="center">2001</td><td align="left">1419</td><td align="left">497</td><td align="left">0</td><td align="left">0</td><td align="left">314</td><td align="left">673</td></tr><tr><td align="center">2002</td><td align="left">1077</td><td align="left">274</td><td align="left">15</td><td align="left">66</td><td align="left">0</td><td align="left">666</td></tr><tr><td align="center">2003</td><td align="left">1382</td><td align="left">160</td><td align="left">41</td><td align="left">0</td><td align="left">0</td><td align="left">1176</td></tr><tr><td align="center">2004</td><td align="left">1446</td><td align="left">80</td><td align="left">62</td><td align="left">37</td><td align="left">0</td><td align="left">689</td></tr><tr><td colspan="7"><hr></hr></td></tr><tr><td align="center"><bold>Total</bold></td><td align="left"><bold>6260</bold></td><td align="left"><bold>1381</bold></td><td align="left"><bold>133</bold></td><td align="left"><bold>103</bold></td><td align="left"><bold>907</bold></td><td align="left"><bold>4325</bold></td></tr></tbody></table></table-wrap><p>Between 1992 and 2004, there was increase in rainfall until 1997. This was followed by a decline in the annual rainfall dropping during the period 2000–2004 to a pattern similar to the one of 1992–1995 (Figure <xref ref-type="fig" rid="F3">3</xref>). No epidemics were detected during the study period. The last reported epidemic was in 1998.</p><fig position="float" id="F3"><label>Figure 3</label><caption><p>Average annual rainfall (mm) in Eritrea 1999–2004.</p></caption><graphic xlink:href="1475-2875-5-33-3"/></fig><p>In 2002, resistance to chloroquine as the first line single drug-treatment regimen reached 6% necessitating introduction of combination first line therapy using chloroquine and sulfadoxine-pyrimethamine. In 2004 resistance to this combination of therapy rose to 4%, which was less than the threshold for changing this combination therapy. There was no resistance detected to DDT, Malathion or any of the insecticides used for the control of adult mosquitoes.</p></sec><sec><title>Findings from community surveys</title><p>In 2001, 80% of surveyed households owned at least one net of which 67.9% were ITNs and 87.7% of the ITNs had been re-treated. The corresponding proportions of ITNs in 2004 were 79%, 73% and 62% respectively revealing a significant reduction in the proportion of recently treated nets i.e. within 6 months of the survey. The use of nets was slightly lower than ownership. Although the overall use of ITNs was relatively high, utilization had slightly declined in 2004 compared to the practice in 2001(Figure <xref ref-type="fig" rid="F4">4</xref>). The ownership and use of ITNs varied by zoba with Anseba having the highest coverage and utilization rates (Table <xref ref-type="table" rid="T5">5</xref>)</p><fig position="float" id="F4"><label>Figure 4</label><caption><p>Proportion of children, adults and pregnant women sleeping under ITN.</p></caption><graphic xlink:href="1475-2875-5-33-4"/></fig><table-wrap position="float" id="T5"><label>Table 5</label><caption><p>Availability of ITNs in households by <italic>zoba</italic></p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td></td><td align="center" colspan="2"><bold>% Households own ITN</bold></td><td align="center" colspan="2"><bold>% ITN Re-treatment</bold></td></tr><tr><td></td><td></td><td colspan="4"><hr></hr></td></tr><tr><td align="left"><bold>Zoba</bold></td><td align="center"><bold>N =</bold></td><td align="center"><bold>>1 ITN declared</bold></td><td align="center"><bold>>1 ITN seen</bold></td><td align="center"><bold>>1 ITN seen</bold></td><td align="center"><bold>> 1 ITN in last 6 months</bold></td></tr></thead><tbody><tr><td align="left">Anseba</td><td align="center">386</td><td align="center">98%</td><td align="center">97%</td><td align="center">96%</td><td align="center">94%</td></tr><tr><td align="left">Debub</td><td align="center">380</td><td align="center">84%</td><td align="center">82%</td><td align="center">75%</td><td align="center">69%</td></tr><tr><td align="left">Gash-Barka</td><td align="center">377</td><td align="center">77%</td><td align="center">75%</td><td align="center">68%</td><td align="center">67%</td></tr><tr><td align="left">Northern Red Sea</td><td align="center">380</td><td align="center">64%</td><td align="center">62%</td><td align="center">52%</td><td align="center">18%</td></tr><tr><td colspan="6"><hr></hr></td></tr><tr><td align="left"><bold>TOTAL</bold></td><td align="center"><bold>1523</bold></td><td align="center"><bold>81%</bold></td><td align="center"><bold>79%</bold></td><td align="center"><bold>73%</bold></td><td align="center"><bold>62%</bold></td></tr></tbody></table></table-wrap><p>There was wide geographical variation in the usage of ITNs with Anseba recording the highest rates of use (Table <xref ref-type="table" rid="T6">6</xref>). Children under 5 years of age attained higher usage of ITNs than that amongst older persons.</p><table-wrap position="float" id="T6"><label>Table 6</label><caption><p>Proportion of persons sleeping under Net/ITN in previous night by <italic>zoba </italic>and age</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td align="center" colspan="3"><bold>Age < 5 years</bold></td><td align="center" colspan="3"><bold>Age > 5 years</bold></td><td align="center" colspan="3"><bold>All ages</bold></td></tr></thead><tbody><tr><td align="left"><bold>Zoba</bold></td><td align="center"><bold>N =</bold></td><td align="center"><bold>Any Net</bold></td><td align="center"><bold>ITN</bold></td><td align="center"><bold>N =</bold></td><td align="center"><bold>Any Net</bold></td><td align="center"><bold>ITN</bold></td><td align="center"><bold>N =</bold></td><td align="center"><bold>Any Net</bold></td><td align="center"><bold>ITN</bold></td></tr><tr><td colspan="10"><hr></hr></td></tr><tr><td align="left">Anseba</td><td align="center"><bold>317</bold></td><td align="center">79.8%</td><td align="center">77.6%</td><td align="center"><bold>1713</bold></td><td align="center">72.3%</td><td align="center">71.0%</td><td align="center"><bold>2035</bold></td><td align="center">73.6%</td><td align="center">72.1%</td></tr><tr><td align="left">Debub</td><td align="center"><bold>304</bold></td><td align="center">56.3%</td><td align="center">51.6%</td><td align="center"><bold>1654</bold></td><td align="center">40.3%</td><td align="center">35.2%</td><td align="center"><bold>1965</bold></td><td align="center">42.8%</td><td align="center">37.8%</td></tr><tr><td align="left">Gash-Barka</td><td align="center"><bold>271</bold></td><td align="center">57.2%</td><td align="center">51.3%</td><td align="center"><bold>1653</bold></td><td align="center">43.9%</td><td align="center">39.8%</td><td align="center"><bold>1943</bold></td><td align="center">45.8%</td><td align="center">41.5%</td></tr><tr><td align="left">Northern Red Sea</td><td align="center"><bold>334</bold></td><td align="center">41.6%</td><td align="center">15.0%</td><td align="center"><bold>1544</bold></td><td align="center">27.0%</td><td align="center">7.9%</td><td align="center"><bold>1878</bold></td><td align="center">29.6%</td><td align="center">9.2%</td></tr><tr><td colspan="10"><hr></hr></td></tr><tr><td align="left"><bold>Total</bold></td><td align="center"><bold>1226</bold></td><td align="center"><bold>58.6%</bold></td><td align="center"><bold>48.3%</bold></td><td align="center"><bold>6564</bold></td><td align="center"><bold>46.4%</bold></td><td align="center"><bold>39.3%</bold></td><td align="center"><bold>7821</bold></td><td align="center"><bold>48.4%</bold></td><td align="center"><bold>40.8%</bold></td></tr></tbody></table></table-wrap><p>The use of malaria prevention methods by pregnant women was high in Anseba. About 81.3% of pregnant women slept under ITNs with 76.6% having slept under an ITN during the previous night (Table <xref ref-type="table" rid="T7">7</xref>). In Anseba, attendance of antenatal clinic was also high for the first and second visits only. In all <italic>zobas </italic>the use of chemoprophylaxis for malaria prevention during the present or last pregnancy was very low reaching 18.2% in Gash Barka. This is compared to the use of drugs to prevent other illnesses reported by 54.5% of all pregnant women in the same <italic>zoba</italic>.</p><table-wrap position="float" id="T7"><label>Table 7</label><caption><p>Malaria prevention indicators for pregnant women by <italic>zoba</italic></p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td></td><td align="center" colspan="3"><bold>Sleeping under Net/ITN</bold></td><td align="center" colspan="3"><bold>Receiving Antenatal care</bold></td><td align="center" colspan="3"><bold>Receiving specific chemoprophylaxis</bold></td></tr><tr><td></td><td></td><td colspan="9"><hr></hr></td></tr><tr><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td align="center" colspan="2">Anti-malarial</td><td></td></tr><tr><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td colspan="2"><hr></hr></td><td></td></tr><tr><td align="left"><bold>Zoba</bold></td><td align="center"><bold>N =</bold></td><td align="center">Any Net</td><td align="center">ITN</td><td align="center">ITN in last 6 months</td><td align="center">At least 1 visit</td><td align="center">1–2 visits</td><td align="center">3–6 visits</td><td align="center">Used</td><td align="center">Correct dose</td><td align="center">Other illness</td></tr></thead><tbody><tr><td align="left">Anseba</td><td align="center">64</td><td align="center">81.3%</td><td align="center">81.3%</td><td align="center">76.6%</td><td align="center">78.1%</td><td align="center">65.6%</td><td align="center">12.5%</td><td align="center">3.1%</td><td align="center">3.1%</td><td align="center">23.4%</td></tr><tr><td align="left">Debub</td><td align="center">61</td><td align="center">59.0%</td><td align="center">59.0%</td><td align="center">59.0%</td><td align="center">73.8%</td><td align="center">41.0%</td><td align="center">32.8%</td><td align="center">0.0%</td><td align="center">0.0%</td><td align="center">19.7%</td></tr><tr><td align="left">Gash-Barka</td><td align="center">55</td><td align="center">60.0%</td><td align="center">56.4%</td><td align="center">56.4%</td><td align="center">83.6%</td><td align="center">63.6%</td><td align="center">20.0%</td><td align="center">18.2%</td><td align="center">14.5%</td><td align="center">54.5%</td></tr><tr><td align="left">Northern Red Sea</td><td align="center">58</td><td align="center">29.3%</td><td align="center">13.8%</td><td align="center">6.9%</td><td align="center">77.6%</td><td align="center">58.6%</td><td align="center">19.0%</td><td align="center">0.0%</td><td align="center">0.0%</td><td align="center">31.0%</td></tr><tr><td colspan="11"><hr></hr></td></tr><tr><td align="left"><bold>TOTAL</bold></td><td align="center"><bold>238</bold></td><td align="center"><bold>58.0%</bold></td><td align="center"><bold>53.4%</bold></td><td align="center"><bold>50.4%</bold></td><td align="center"><bold>78.2%</bold></td><td align="center"><bold>57.1%</bold></td><td align="center"><bold>21.0%</bold></td><td align="center"><bold>5.0%</bold></td><td align="center"><bold>4.2%</bold></td><td align="center"><bold>31.5%</bold></td></tr></tbody></table></table-wrap><p>During the previous six months, IRS was carried out to a small extent mostly in Gash Barka where 42.2% of the households were sprayed. This was followed by Debub (25.8%) and Northern Red Sea (3.2%) giving a national average of 18.1%.</p><p>Community members actively participated in malaria control activities in the endemic zobas. More than 80% of the households in Anseba participated in such activities (Figure <xref ref-type="fig" rid="F5">5</xref>).</p><fig position="float" id="F5"><label>Figure 5</label><caption><p>Proportion of households participating in ecological management by zoba.</p></caption><graphic xlink:href="1475-2875-5-33-5"/></fig></sec><sec><title>Findings from facility surveys</title><p>At the facilities, health workers requested laboratory tests in 81% of the patients suspected to have malaria. The diagnostic skills of laboratory staff were high as they realised blood slide examination sensitivity of 99.1% and specificity of 95% when cross-checked for quality control by the national reference laboratory.</p><p>The quality of patient care did not show significant improvement within the facilities. Data from health facility survey shows that only 52% of patients with suspected malaria had a temperature measurement taken, 75% were checked for pallor while only 58% were managed according to the national guidelines.</p><p>Although all facilities surveyed had adequate supplies of the recommended combination therapy of chloroquine and sulfadoxine-pyrimethamine (fansidar) 7% of the patients surveyed were prescribed chloroquine alone. At the same time, in 2001, only 7.2% of infected persons sought and obtained effective treatment within the recommended time of 24 hours. There was little change in this proportion, rising only to 7.5% in 2004.</p></sec><sec><title>Linear regression analysis results</title><p>Results of Time series model analysis (ARIMA) (Table <xref ref-type="table" rid="T8">8</xref>) show strong negative correlation between the numbers of ITNs distributed (β = -0.125, p <0.005), and the amount of DDT and Malathion (in kg) used in IRS (β = -2.352, p < 0.05) with the overall reduction in malaria morbidity. The effect is sustained in model fit where these two factors are included. However this joint effect is not statistically significant. The data was inadequate to compute a model to which rainfall was added to ITNs, and IRS.</p><table-wrap position="float" id="T8"><label>Table 8</label><caption><p>Time series analysis (ARIMA) of cases and deaths, against malaria control interventions</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left">Univariate analysis</td><td></td><td></td></tr></thead><tbody><tr><td align="left"><bold>Effect on malaria morbidity (number of cases)</bold></td><td></td><td></td></tr><tr><td colspan="3"><hr></hr></td></tr><tr><td></td><td align="left">β coefficient</td><td align="left">Probability</td></tr><tr><td align="left">ITNs (number distributed)</td><td align="left">-0.125</td><td align="left">0.005</td></tr><tr><td align="left">Number of ITNs retreated</td><td align="left">-0.016</td><td align="left">0.02</td></tr><tr><td align="left">IRS (kg of DDT & Malathion)</td><td align="left">-2.352</td><td align="left">0.05</td></tr><tr><td align="left">Malathion (kg)</td><td align="left">-3.270</td><td align="left">0.05</td></tr><tr><td align="left">CHAs trained (number)</td><td align="left">-55.483</td><td align="left">0.6</td></tr><tr><td align="left">Rainfall</td><td align="left">275.95</td><td align="left">0.4</td></tr><tr><td align="left">Population protected</td><td align="left">-0.728</td><td align="left">0.11</td></tr><tr><td colspan="3"><hr></hr></td></tr><tr><td></td><td></td><td></td></tr><tr><td align="left"><bold>Effect on case fatality (number of deaths)</bold></td><td></td><td></td></tr><tr><td colspan="3"><hr></hr></td></tr><tr><td></td><td align="left">β coefficient</td><td align="left">Probability</td></tr><tr><td align="left">IRS (kg of DDT & Malathion)</td><td align="left">-0.002</td><td align="left">0.08</td></tr><tr><td align="left">Health workers trained</td><td align="left">0.226</td><td align="left">0.03</td></tr><tr><td align="left">Abate (litres)</td><td align="left">-0.468</td><td align="left">0.3</td></tr><tr><td></td><td></td><td></td></tr><tr><td align="left" colspan="3">Multivariate</td></tr><tr><td align="left" colspan="3"><underline>Effect on malaria morbidity (number of cases)</underline></td></tr><tr><td></td><td align="left">β coefficient</td><td align="left">Probability</td></tr><tr><td align="left">ITNs</td><td align="left">-0.1663</td><td align="left">0.13</td></tr><tr><td align="left">IRS</td><td align="left">0.832</td><td align="left">0.45</td></tr><tr><td></td><td></td><td></td></tr><tr><td align="left" colspan="3"><underline>Case fatality</underline></td></tr><tr><td></td><td align="left">β coefficient</td><td align="left">Probability</td></tr><tr><td align="left">ITNs</td><td align="left">-0.00011</td><td align="left">0.6</td></tr><tr><td align="left">IRS</td><td align="left">0.0006</td><td align="left">0.9</td></tr></tbody></table></table-wrap><p>The correlation between the total number of health personnel trained and the trends in malaria morbidity and case fatality rate decline was statistically significant (p < 0.028).</p><p>There was an overall reduction in the average annual rainfall in the country during the study period. The observed association between the decline in rainfall for the period 1999–2004, and the reduction in case fatality and malaria incidence rates were not statistically significant (p < 0.441).</p><p>The negative association between case fatality and IRS, population protected, number of health workers trained and larvicidal activities was not statistically significant.</p></sec></sec><sec><title>Discussion</title><p>HMIS and cross-sectional survey data were used to assess the effects and impact of Eritrea's Roll Back Malaria Programme for the period 2000–2004. The goal of the programme was to reduce morbidity and mortality due to malaria to such low levels that malaria was no longer a public health problem in the country [<xref ref-type="bibr" rid="B12">12</xref>]. In five years, the programme exceeded the national targets of 80% reduction in malaria morbidity and mortality and surpassing the 60% objective of households owning ITNs. The achievements are also well above the Abuja targets for 2010.</p><p>The thrust of malaria prevention is to reduce human mosquito bites. This can be achieved through use of ITNs or reduction in either or both of the larvae and adult mosquitoes [<xref ref-type="bibr" rid="B1">1</xref>]. In Eritrea, the use of ITNs contributed most to the reduction in malaria morbidity and mortality. The number of ITNs distributed as a proxy for the reduction in human bites was strongly correlated to the observed steep decline in both morbidity and mortality for the period 2000–2004. Access to ITNs was rapidly increased by targeting pregnant women using the antenatal services and through freely availing ITNs to vulnerable groups of women and children below the age of five. Antenatal attendance in the country is estimated to be 70% for at least one visit during pregnancy. Free ITNs were distributed to the vulnerable groups and internally displaced populations, initially as a pilot project [<xref ref-type="bibr" rid="B18">18</xref>]. By 1999, free ITNs distribution had been extended to all the high-risk malaria zones and there was a sustained effort with a target of ensuring that in at least 70% of the population at risk, each household had a least two ITNs. The reported ITN coverage was in excess of 80% in most regions especially the high malaria endemic areas. In Senegal and Nigeria where the thrust is to encourage commercial supply of ITNs and creating demand among consumers and reaching vulnerable populations though targeted subsidy programmes, household coverage of ITNs was reported to be comparatively low at 43% and 10%, respectively [<xref ref-type="bibr" rid="B19">19</xref>].</p><p>Arguably the most cost effective tool in malaria prevention is the use of ITNs. Randomised studies have documented up to 30% reduction in the number of under-5 deaths through ITN use alone [<xref ref-type="bibr" rid="B20">20</xref>]. In a related report it was concluded that 6 deaths are averted for every 1,000 children age 1–59 months that sleep under ITN [<xref ref-type="bibr" rid="B4">4</xref>]. In Eritrea, ITNs use as a single intervention was strongly correlated to the 84% decline in morbidity and mortality. In view of the high effectiveness of this method three related issues needed to be assessed: usage of nets by children and pregnant women, re-treatment of ITNs, and sustainability.</p><p>It is also clear that the population is getting motivated to use the ITNs regularly. The proportion of children sleeping under ITNs in Eritrea is high compared to Senegal or Nigeria where only 25% and less than 5% children below the age of 5 years sleep under ITNs [<xref ref-type="bibr" rid="B18">18</xref>]. In other settings it has been shown that children are disadvantaged in resource allocation [<xref ref-type="bibr" rid="B21">21</xref>]. In Eritrea this does not appear to be the case. Although the proportion of pregnant women sleeping under ITN is only 50%, this is probably the highest in the continent.</p><p>ITNs re-treatment was equally high at 62% (done within the previous six months of the survey). This can be attributed to community involvement and token monetary incentives given to community health agents in return for increased ITN re-impregnation rates [<xref ref-type="bibr" rid="B22">22</xref>]. For each ITN re-treated, the CHA received 40 cents-Nakfa (3 cents US). The observed decline in the ITNs re-treatment may not affect the programme as the country is in the process of introducing Long Lasting Insecticide-Treated Nets.</p><p>One important finding of the study in respect to sustainability was that in Eritrea, 23.7% of the households surveyed in 2004 had purchased the ITNs [<xref ref-type="bibr" rid="B22">22</xref>]. It is plausible that the perceived beneficial effect of ITNs freely distributed in the community has positively influenced some households to invest in ITNs. Moreover, in Eritrea one free ITN is considered cost-effective given the lifetime cost of treating malarial infections and the life-saving effect of ITNs. Therefore, the government has annual running budget-line for procurement of ITNs and their free distribution to high-risk population groups.</p><p>IRS was the next most important vector control method in the country. Approximately 13% of the population in malaria risk areas of the country benefited from IRS. DDT and Malathion were the two chemicals commonly in use. Effective community mobilization and involvement contributed to the observed increasing IRS coverage. Controversies surrounding the use of DDT, which was the mainstay of eradication and vector control, have tended to undermine success in the tropics [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B23">23</xref>]. Recent shifts in favour of controlled indoor use of DDT have supported renewed interest leading to its re-introduction in some countries including Eritrea. The Roll Back Malaria programme in Eritrea advocates for the use of DDT for IRS alongside experimental preparations. Although the use of DDT is contestable given the environmental risks it poses, many of the poor countries cannot afford the alternative chemicals currently being tested, as their cost is prohibitive [<xref ref-type="bibr" rid="B24">24</xref>]. In addition tests in Eritrea on resistance to the commonly used insecticides do not show any evidence of resistance to DDT.</p><p>Although ecological management activities were statistically not associated with significant effect on either morbidity or mortality due to malaria, other studies have shown that trampled pools, rain pools, ponds, dams, swamps, drainage channels and communal water supply points are favorable larval habitats for <italic>anopheline </italic>mosquito in Eritrea [<xref ref-type="bibr" rid="B25">25</xref>]. In addition, there is strong correlation between adult mosquito density and larval density. Elimination of the known habitats was therefore an important primary preventive measure for malaria. The environmental control measures included covering of breeding sites for mosquitoes and using larvicides for water bodies that were not amenable to covering. The relatively large water collections tended to be reservoirs for the vector during the dry season. In 2004, more than 80% of the breeding sites in Southern Red Sea, Northern Red Sea and Anseba were covered through active community participation [<xref ref-type="bibr" rid="B18">18</xref>]. The remainder was subjected to temephos, a larvicidal chemical. In areas where it was not possible to cover the sites, use of larvicides was found to be a favorable alternative because the water bodies are relatively limited in Eritrea, especially in Gash Barka, Debub and Maekel. The role of the community was central to the success. Nearly 60% of the adult population in the four high-risk zobas surveyed, participate in environmental management activities.</p><p>Provision of early clinical diagnosis and laboratory confirmation, are other critical aspects of effective and efficient case management. The training given to laboratory technicians working in hospitals at the <italic>zoba </italic>referral hospitals and <italic>sub-zoba </italic>facilities on microscopy equips them with adequate skills. Malaria diagnostic acumen at the facility level is very high since at least 80% of the diagnoses are confirmed by laboratory tests carried out by well-trained technicians. The National Laboratory quality assurance check confirmed high sensitivity (99.1%) and specificity (95%) of these malaria tests. The thick and thin peripheral blood smear with Giemsa stain is considered the gold standard for the diagnosis of malaria, achieving overall 90% sensitivity and specificity [<xref ref-type="bibr" rid="B26">26</xref>].</p><p>The quality of clinical case management did not change and could not have contributed to the fall in case fatality rate. These findings are consistent with earlier surveys in Eritrea [<xref ref-type="bibr" rid="B27">27</xref>]. This is probably related to the high attrition rates resulting from internal transfers of cadres of staff and insufficient training on integrated management of childhood illnesses (IMCI). In addition the health seeking behaviour of the population showed little change during the study period. In 2004 only 7.5% of the sampled sick children received treatment within 24-hours of the onset of fever reflecting 6% rise from 2001 level. Access to health facilities and information are some of the contributing factors to the slow change in health seeking behaviour. The NMCP developed a malaria communication strategy following KAP survey conducted in 2002 but implementation of the strategy only commenced in 2004 [<xref ref-type="bibr" rid="B22">22</xref>]. The observed decline in case fatality after 2002 following an initial rise between 2000 and 2002, may be due to the change of policy on the first line treatment from chloroquine alone to a combination of chloroquine and suphadoxine-pyrimethamine.</p><p>CHAs are increasingly diagnosing and managing malaria cases. The 50% rise in their contribution to case management in Eritrea is remarkable. CHA are locally based and trained people who are easily accessible and culturally accepted in the locality [<xref ref-type="bibr" rid="B18">18</xref>]. This cadre became a revelation and asset of the NMCP in terms of effectiveness through early diagnosis and prescribing of relatively cheap and safe management regimens [<xref ref-type="bibr" rid="B13">13</xref>,<xref ref-type="bibr" rid="B18">18</xref>]. Training of CHAs and those at other health facilities was an ongoing exercise improving capacity and competence to manage simple cases of malaria and enable recognition of early warning signs for severe malaria for referral to better-equipped centres [<xref ref-type="bibr" rid="B13">13</xref>]. The CHA programme currently suffers from lack of standardization since the CHAs rely only on clinical diagnosis. The availability of reliable Rapid Diagnostic Kit, which is already in use in the country, is expected to become a useful tool in the services the CHAs provide within the community. The Rapid Diagnostic Test Kit has been found to be easy and cheap for on-site use and particularly so in countries like Brazil that have adopted aggressive active case detection [<xref ref-type="bibr" rid="B28">28</xref>,<xref ref-type="bibr" rid="B29">29</xref>].</p><p>Surveillance for emergence of resistance to first line anti-malaria drugs particularly chloroquine and first line combination drugs (chloroquine plus sulphadoxine-pyrimethamine), is important for programme success and reducing case fatality and an adjuvant to improving quality of care. Drug resistance surveillance is also essential in order to increase the useful therapeutic life of a constituent drug [<xref ref-type="bibr" rid="B30">30</xref>]. However, for this to be effective strong systems and decisive management are required for timely and effective response. Eritrea has established twenty centres, which in 2002 enabled the NMCP to change the first line treatment to the combination of chloroquine and sulphadoxine-pyrimethamine. Chloroquine, until recently, the mainstay of malaria treatment, precluded its use as a single drug in Eastern Africa [<xref ref-type="bibr" rid="B4">4</xref>]. In Eritrea, resistance to the new first line regimen is still low.</p><p>Surveillance and epidemic preparedness systems are well entrenched in Eritrea's malaria control programme as a means of reducing morbidity rates and case fatality. The NMCP collects rainfall data daily, continuously analyses it and the results are fed into the national and the <italic>zoba </italic>coordinating office databases. During the decade, the initial decline in the amount of the annual rainfall in the country stabilised at between 300 and 400 mm. Although the extent to which this influenced malaria morbidity and mortality trends was not statistically significant, this finding is inconclusive, as the study period coincided with a persistently lower than normal rainfall averages for Eritrea. In many countries rainfall and temperature data are predictors of impending epidemic and serve as early warning system. In East Africa, climate variability has been shown to contribute to the likelihood of an epidemic [<xref ref-type="bibr" rid="B31">31</xref>]. The precision of rainfall and temperature data and lead-time can be enhanced by satellite based meteorological data [<xref ref-type="bibr" rid="B32">32</xref>]. Eritrea is not accessing satellite data as a routine source but uses data from weather stations spread all over the country.</p><p>For epidemic preparedness, Eritrea monitors the number of new malaria cases during each month. An impending epidemic is suspected once the reported number of cases per month in a health facility first rises higher than the third quartile of the number estimated for that facility. The last epidemic reported in the country was 1998 [<xref ref-type="bibr" rid="B9">9</xref>]. The absence of epidemics in the intervening period cannot be explained on the basis of climatic factors alone. In the case of Eritrea the resultant control of epidemics using the ecological management and household protection may be the mechanism through which epidemics were prevented. The sensitivity of the epidemic thresholds has been questioned leading to the proposed replacement of the quartile measure by the use of weekly averages. The occurrence of severe epidemics is to be expected with the dramatic decline in the malaria incidence in the country and consequent to the declining immunity in the population. Severe <italic>Plasmodium falciparum </italic>malaria is known to occur in low transmission intensity [<xref ref-type="bibr" rid="B33">33</xref>]. Seasonal and unstable malaria transmission, attend to similar phenomena [<xref ref-type="bibr" rid="B34">34</xref>].</p><p>The final question to be explored was the role of each of the interventional measures in reducing morbidity or mortality. Within the limitations of the current study design it is evident that combining ITN use with IRS or other vector control measures did not confer added value to the outcome in malaria mortality or morbidity. This is not surprising since <italic>An. arabiensis </italic>is endophilic, and both methods act at the point of breaking the vector-human contact. This is supported by observations from elsewhere that DDT spray can eliminate up to 93–95% indoor resting density of a vulnerable vector [<xref ref-type="bibr" rid="B23">23</xref>].</p><p>In summary therefore, the Abuja Declaration targets for the Roll Back Malaria initiative were met on schedule because the government had set even higher targets for itself. There was extensive community and personal prevention measures which started as donations of ITNs, initially targeted at pregnant women and children and sustainable through community awareness. Both morbidity and mortality have declined to a point where malaria ceases to be a major infectious disease in this small African country. This is remarkable as no other country in the continent has similar achievements.</p></sec> |
Locked nucleoside analogues expand the potential of DNAzymes to cleave structured RNA targets | <sec><title>Background</title><p>DNAzymes cleave at predetermined sequences within RNA. A prerequisite for cleavage is that the DNAzyme can gain access to its target, and thus the DNAzyme must be capable of unfolding higher-order structures that are present in the RNA substrate. However, in many cases the RNA target sequence is hidden in a region that is too tightly structured to be accessed under physiological conditions by DNAzymes.</p></sec><sec><title>Results</title><p>We investigated how incorporation of LNA (locked nucleic acid) monomers into DNAzymes improves their ability to gain access and cleave at highly-structured RNA targets. The binding arms of DNAzymes were varied in length and were substituted with up to three LNA and α-L-LNA monomers (forming LNAzymes). For one DNAzyme, the overall cleavage reaction proceeded fifty times faster after incorporation of two α-L-LNA monomers per binding arm (<italic>k</italic><sub>obs </sub>increased from 0.014 min<sup>-1 </sup>to 0.78 min<sup>-1</sup>).</p></sec><sec><title>Conclusion</title><p>The data demonstrate how hydrolytic performance can be enhanced by design of LNAzymes, and indicate that there are optimal lengths for the binding arms and for the number of modified LNA monomers.</p></sec> | <contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Vester</surname><given-names>Birte</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>b.vester@bmb.sdu.dk</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Hansen</surname><given-names>Lykke H</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>lhh@bmb.sdu.dk</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Bo Lundberg</surname><given-names>Lars</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>lundberg_larsbo@msn.com</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Babu</surname><given-names>B Ravindra</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>ravi.babu@leo-pharma.com</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Sørensen</surname><given-names>Mads D</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>mds@crc.dk</email></contrib><contrib id="A6" contrib-type="author"><name><surname>Wengel</surname><given-names>Jesper</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>jwe@chem.sdu.dk</email></contrib><contrib id="A7" contrib-type="author"><name><surname>Douthwaite</surname><given-names>Stephen</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>srd@bmb.sdu.dk</email></contrib> | BMC Molecular Biology | <sec><title>Background</title><p>DNAzymes function as specific endonucleases by binding to predetermined sequences in RNA and cleaving its phosphodiester backbone. The discovery that RNA-hydrolytic properties could be encoded within a DNA oligonucleotide indicated potential biotechnological applications in gene silencing. These applications might even surpass those of other oligonucleotide-based gene silencing approaches, such as antisense and RNAi technologies, that require the complicity of the cell's own nuclease systems in order to cleave RNA. However, the use of DNAzymes has been restricted by several limitations, some of which are shared with the other oligonucleotide-based technologies. For any of these approaches to be of value, the oligonucleotide must be capable of transversing the cellular membrane and avoid being inactivated by cellular nucleases long enough to find the appropriate cellular compartment where it can bind and induce cleavage at the target RNA. Modification of DNAzymes to improve their stability against cellular nucleases and their ability to bind and cleave RNA molecules would go a long way towards increasing their general applicability.</p><p>The 10–23 DNAzyme isolated by Santoro and Joyce [<xref ref-type="bibr" rid="B1">1</xref>] (Fig. <xref ref-type="fig" rid="F1">1A</xref>) provides a suitable starting point for such modification and improvement. DNAzymes of this type cleave purine-uracil targets, and to a lesser extent purine-cytosine targets, and have been subjected to numerous investigations to alter target preference [<xref ref-type="bibr" rid="B2">2</xref>], to follow the folding of the catalytic region [<xref ref-type="bibr" rid="B3">3</xref>], and to induce cleavage at different targets [<xref ref-type="bibr" rid="B4">4</xref>,<xref ref-type="bibr" rid="B5">5</xref>]. Quite conclusive is that most DNAzymes of this type fail to cleave their RNA targets <italic>in vitro </italic>[<xref ref-type="bibr" rid="B4">4</xref>,<xref ref-type="bibr" rid="B5">5</xref>], and this does not bode well for <italic>in vivo </italic>application with the additional complications of cellular nucleases and substrate compartmentalization. Poor cleavage <italic>in vitro </italic>is generally linked to the inability of a DNAzyme to recognize its target, which involves unravelling the higher-order structure of the RNA and hybridizing via the two binding arms of the DNAzyme (Fig. <xref ref-type="fig" rid="F1">1</xref>). The binding arms can be constructed to include modified nucleosides, and such modifications have improved resistance against cellular nucleases [<xref ref-type="bibr" rid="B6">6</xref>-<xref ref-type="bibr" rid="B10">10</xref>]. In principle, the inclusion of modified nucleosides at specific sites in the binding arms could also greatly enhance the hybridization potential of DNAzymes at their RNA targets.</p><p>Modified nucleosides in the form of conformationally locked analogues (Locked Nucleic Acid, LNA) are capable of targeting complementary RNA and DNA with high affinity [<xref ref-type="bibr" rid="B11">11</xref>,<xref ref-type="bibr" rid="B12">12</xref>], and are thus potentially interesting as components in DNAzymes. Two commercially available variants of these modified nucleosides (LNA and α-L-LNA, Fig. <xref ref-type="fig" rid="F1">1B</xref>) can be incorporated into oligonucleotides to increase their hybridization capacity [<xref ref-type="bibr" rid="B13">13</xref>-<xref ref-type="bibr" rid="B15">15</xref>]. Recent reports [<xref ref-type="bibr" rid="B10">10</xref>,<xref ref-type="bibr" rid="B16">16</xref>-<xref ref-type="bibr" rid="B19">19</xref>] have shown that inclusion of LNA monomers into the binding arms of DNAzymes (forming what we term LNAzymes) enables structured RNA targets to be cleaved, even in cases where the targets are intractable to unmodified DNAzymes. These studies indicate that the potential of LNAzymes might be optimized by rational design to cleave at targets that are located in highly structured RNA substrates.</p><p>Here, we have conducted a systematic approach geared towards the rational design of LNAzymes based on the 10–23 DNAzyme structure. Targets were displayed in RNA substrates with progressively more stable structures to challenge the cleavage capacity of DNAzymes. The ability of the DNAzymes to unravel the RNA structures was investigated by varying the lengths and the modification content of the hybridization arms. RNA substrates that are essentially inaccessible to unmodified DNAzymes were cleaved with markedly higher efficiency when LNA or α-L-LNA monomers were incorporated into the binding arms. For each RNA target, there appear to be optima for lengths of the binding arms, as well as for the type and number of modifications they contain. While the improved ability of LNAzymes to gain access to structured targets is the main reason for their greater capacity to cleave RNAs, the modifications also appear to favourably affect other kinetic parameters.</p></sec><sec><title>Results</title><sec><title>RNA substrates for cleavage</title><p>RNA substrates contain two consecutive sequences that are complementary to the binding arms of the DNAzymes and bracket the cleavage site between a purine and a pyrimidine (Fig. <xref ref-type="fig" rid="F1">1A</xref>). The cleavage sites studied here are displayed in RNA substrates that are based on sequences from the <italic>Escherichia coli </italic>23S ribosomal RNA, and possess a range of structural complexity (Fig. <xref ref-type="fig" rid="F2">2</xref>). The substrates consist of a 17n minimal RNA (largely unstructured), a 33n with metastable secondary structure, 58n and 74n RNAs with stable secondary structures, and a 2904n RNA (the intact <italic>E. coli </italic>23S rRNA) with extensive secondary and tertiary structure. Within these RNA substrates, three cleavage sites were studied: the 745-site (in the 74n and 2904n RNAs); the 1093-site (in the 17n, 33n, 58n and 2904n RNAs); and the 1096-site (in the 33n, 58n and 2904n RNAs).</p></sec><sec><title>DNAzymes and LNAzymes</title><p>The nucleic acid enzymes used in this study range from unmodified DNAzymes to LNAzymes containing up to three LNA or α-L-LNA monomers per binding arm (Table <xref ref-type="table" rid="T1">1</xref>). The lengths of the binding arms were also varied from five up to eight nucleotides; the catalytic loop remained unchanged in all the enzymes. A single DNAzyme and two LNAzymes were designed against the 745-site; one DNAzyme and one LNAzyme were made to cleave the 1096-site. The cleavage site at 1093 was studied most systematically, and two DNAzymes plus five LNAzymes were designed against this target. In the initial tests, the nucleic acid enzymes were used in molar excess relative to the RNA substrates, enabling us to study their performances under single-turnover conditions. The kinetic parameters of the most effective enzymes were then studied under multiple-turnover conditions with an excess of substrate.</p></sec><sec><title>Cleavage at the 1093-site</title><p>Cleavage at the 1093-site in the 17n, 33n, 58n and 2904n RNAs was carried out under single-turnover conditions with a 5- to 50-fold excess of DNAzyme or LNAzyme (Fig. <xref ref-type="fig" rid="F3">3</xref>). The unstructured 17n RNA substrate was efficiently cleaved by the unmodified DNAzyme (Dz1-1093), and also by the corresponding LNAzymes with arms of eight nucleotides (Lz1-1093, Lz2-1093 and Lz5-1093) (Fig. <xref ref-type="fig" rid="F3">3A</xref>). However, when the 1093-site was displayed in the more tightly structured 58n RNA distinct differences were seen: cleavage by the unmodified DNAzyme was only marginal even at high concentration, whereas the LNAzymes still cut as effectively as in the 17n RNA (Fig. <xref ref-type="fig" rid="F3">3B</xref>). This picture was intensified in the more highly structured 2904n RNA, which the DNAzyme Dz1-1093 failed to cleave, whereas Lz1-1093 and Lz2-1093 still cut the 1093-site effectively (Fig. <xref ref-type="fig" rid="F3">3D</xref>). Incorporation of additional α-L-LNA modifications in Lz5-1093 resulted in less cleavage of 2904n RNA than with Lz1-1093 (Fig. <xref ref-type="fig" rid="F3">3D</xref>).</p><p>Upon reduction of the length of the binding arms to five and six nucleotides, the unmodified DNAzyme (Dz2-1093) lost all measurable hydrolytic activity and, even in large molar excess, was unable to cleave the minimal 17n substrate (Fig. <xref ref-type="fig" rid="F3">3A</xref>). The shorter arm size was counteracted to some extent by incorporating two α-L-LNA monomers into each arm (in Lz4-1093), resulting in good cleavage of 17n and moderate cleavage of 58n RNA (Fig. <xref ref-type="fig" rid="F3">3B</xref>). Further reduction of both arms to four nucleotides led to further loss of activity, and at 50-fold excess the LNAzyme Lz3-1093 could only weakly cut the 17n substrate despite having two α-L-LNA monomers per binding arm.</p></sec><sec><title>Cleavage at the 745- and 1096-sites</title><p>These two additional cleavage sites were investigated to establish whether the observations made for the 1093-site could be extrapolated to other RNA targets. The 1096-site is particularly interesting in this context as it is located only three nucleotides away from the 1093-site and can thus be displayed in the same hairpin loop of the RNA substrates (Fig. <xref ref-type="fig" rid="F2">2</xref>). Despite this, the 1096-site proved more difficult to cleave. The DNAzyme Dz3-1096 failed to cut even when the 1096 target was presented in the relatively unstructured 33n RNA (and consequently also failed to cut the more complexly structured substrates, such as 58n RNA in Fig. <xref ref-type="fig" rid="F3">3B</xref>). Modifying the binding arms with two LNA monomers to form the LNAzyme Lz6-1096 led to effective cleavage at the 1096-site in the 33n and 58n RNA substrates (Fig. <xref ref-type="fig" rid="F3">3B</xref>).</p><p>The 745-site was also resilient to cleavage, and the unmodified DNAzyme Dz4-745 was incapable of cutting this target in either the 74n (Fig. <xref ref-type="fig" rid="F3">3C</xref>) or the 2904n RNAs (Fig. <xref ref-type="fig" rid="F3">3E</xref>). Some cleavage was achieved by modifying the enzyme to include two α-L-LNA monomers (Lz7-745) or two LNA monomers per binding arm (Lz8-745), but remained incomplete at 50-fold molar excess of the enzymes. The LNAzyme with LNA monomers was marginally, but consistently, better at cleaving the RNA substrates than its counterpart containing α-L-LNA.</p></sec><sec><title>Kinetic parameters for cleavage at the 1093-site</title><p>The two LNAzymes Lz1-1093 and Lz2-1093 performed almost identically under single turnover conditions; this is illustrated in Figs. <xref ref-type="fig" rid="F4">4A</xref> and <xref ref-type="fig" rid="F4">4B</xref> and is consistent with a previous report [<xref ref-type="bibr" rid="B16">16</xref>]. The rate constant (<italic>k</italic><sub>obs</sub>) for the unmodified Dz1-1093 and the LNAzyme Lz1-1093 was determined by fitting the time-course data for cleavage of the 17n RNA (Fig. <xref ref-type="fig" rid="F4">4A</xref>) into an exponential curve. This gave a <italic>k</italic><sub>obs </sub>of 0.014 min<sup>-1 </sup>for the Dz1-1093 and a <italic>k</italic><sub>obs </sub>of 0.78 min<sup>-1 </sup>for the Lz1-1093, showing that the overall reaction proceeded fifty times faster for the LNAzyme. It should be noted, however, that whereas the Dz1-1093 data fit comfortably into the exponential curve, the Lz1-1093 data fit much better into an equation describing a two step exponential decay. This suggests that the LNAzyme catalyzes a biphasic reaction, which has two <italic>k</italic><sub>obs </sub>values of 1.37 min<sup>-1 </sup>and 0.09 min<sup>-1</sup>.</p><p>The two most effective enzymes studied here, Lz1-1093 and Lz2-1093, cleave under a wide range of multiple-turnover conditions, and thus function in a truly enzymatic manner. Comparison of cleavage by the two LNAzymes in a 20-fold molar excess of the 17n substrate revealed that Lz1-1093 functions better than Lz2-1093 (Fig. <xref ref-type="fig" rid="F4">4C</xref>). Cleavage by Lz1-1093 under multiple turnover conditions was investigated further, and the results of time-course experiments for cleavage of the 17n RNA substrate at 5 nM to 1.2 μM are plotted in Fig. <xref ref-type="fig" rid="F4">4D</xref>. The initial reaction velocities (v) were measured from the time curves of cleavage at each substrate concentration; these were used in double reciprocal plots to estimate K<sub>m </sub>and the maximal velocity V<sub>max</sub>. Cleavage of the 17n RNA by Lz1-1093 proceeded with a K<sub>m </sub>of 33 nM and V<sub>max </sub>of 1.2 nM min<sup>-1</sup>; the Lz1-1093 LNAzyme concentration was 1 nM, giving a turnover number of 1.2 min<sup>-1</sup>. The unmodified Dz1-1093 did not cleave the substrate under these conditions.</p></sec></sec><sec><title>Discussion</title><p>Hybridization studies on oligonucleotides containing LNA and α-L-LNA monomers have shown that helical thermostability is greater than for unmodified nucleic acid chains [<xref ref-type="bibr" rid="B14">14</xref>]. Thus, we surmised that reconstructing DNAzymes to include LNA and α-L-LNA (forming LNAzymes) would improve their hybridization properties, and facilitate cleavage of obstinate RNA substrates. Consistent with this idea, incorporation of LNA and α-L-LNA monomers into the enzyme binding arms invariably improves cleavage performance, and this proved to be the case irrespective of whether the target was presented in an unstructured or complexly structured RNA substrate. This is consistent with other studies of LNA modified DNAzymes containing blocks of three to five LNA monomers at the ends of the binding arms [<xref ref-type="bibr" rid="B10">10</xref>,<xref ref-type="bibr" rid="B17">17</xref>]. Better hybridization of LNAzymes at the RNA target does indeed play a major role in enhancing cleavage, although other changes in kinetic parameters probably also contribute to the greater effectiveness of LNAzymes.</p><p>DNAzymes with eight nucleotides per binding arm were chosen as the starting point for our studies, as this arm length has been shown to provide a good balance between substrate recognition and product release [<xref ref-type="bibr" rid="B5">5</xref>,<xref ref-type="bibr" rid="B20">20</xref>,<xref ref-type="bibr" rid="B21">21</xref>]. Previous observations that only a small fraction of the potential purine-uridine sites can be cleaved by DNAzymes [<xref ref-type="bibr" rid="B4">4</xref>,<xref ref-type="bibr" rid="B5">5</xref>] were reflected in our choice of sites: one site (1093) is readily cleaved, whereas other two sites (745 and 1096) are intractable to DNAzymes.</p><p>Incorporation of LNA and α-L-LNA monomers into the binding arms of the DNAzymes improved cleavage at all three of the sites. Two LNA or α-L-LNA monomers per eight nucleotide binding arm worked best at the 1093-site, and increasing this number to three α-L-LNA monomers per binding arm (in Lz5-1093) caused a slight loss of activity (Fig. <xref ref-type="fig" rid="F3">3D</xref>), which was particularly noticeable on the 2904n RNA substrate. Heavily modified arms could adversely affect parameters such as nonspecific hybridization at other sites (especially in large substrates such as 2904n) and release of the product fragments after cleavage. However, product release effects would not be expected to come into play under single-turnover conditions with a large enzyme excess, so the extra modifications in Lz5-1093 must also interfere with other processes such as the accessibility of the LNAzyme's binding arms and/or the folding of its catalytic loop.</p><p>Reducing the lengths of the DNAzyme binding arms to five nucleotides led to loss of activity at the 1093-site (Dz2-1093). Activity was re-established by incorporating two α-L-LNA monomers into each of the arms (Lz4-1093). An LNAzyme with four nucleotides per binding arm, two of which were α-L-LNA monomers, also showed some cleavage activity (Lz3-1093). However, cleavage was only evident when the accessible 1093-site was displayed in the least structured substrate (17n) and challenged with a large molar excess of enzyme. Thus, the structure of the Lz3-1093 LNAzyme with its tetranucleotide arms is probably approaching the minimal size that can be expected to have hydrolytic activity.</p><p>LNA and α-L-LNA monomers in a DNA chain have markedly different effects on the helical geometry. Helices containing LNA monomers adopt a local A-form structure [<xref ref-type="bibr" rid="B13">13</xref>], whereas helices containing α-L-LNA tend more towards the B-form structure [<xref ref-type="bibr" rid="B15">15</xref>] with less pronounced stability [<xref ref-type="bibr" rid="B14">14</xref>]. In some cases it appeared more advantageous to use one type of modified monomer – for instance, at the 745-site, the LNA-enzyme was better than its α-L-LNA counterpart. At the more amenable 1093-site, however, no difference was seen in the single turnover cleavage by comparable LNA and α-L-LNA enzymes (Lz1-1093 versus Lz2-1093), nor did any differences appear when the 1093-site accessibility was reduced in the structured RNA substrates (58n and 2904n). The superior performance of Lz1-1093 compared to Lz2-1093 under multiple turnover conditions possibly reflects a difference in hybridisation energy that allows a faster product release from the α-L-LNA containing Lz1-1093 relative to Lz2-1093 with LNA monomers.</p><p>From the data presented here it can be seen that more highly structured RNA substrates have a general tendency to be more resilient to cleavage by nucleic acid enzymes, and that tighter structure can be counteracted by improving the hybridization capacity of the enzyme. However, several exceptions to this rule indicate that factors additional to substrate structure and enzyme hybridization play important roles in the cleavage process. For instance, the 1096-site was more difficult to cleave than the 1093-site despite being only three nucleotides away and being displayed in identical substrates. DNAzymes and LNAzymes designed against the 1096-site had binding arms that were expected to hybridize at least as strongly as the enzymes against the 1093-site (Dz3 and Lz6 are calculated to have Tm values when bound to the RNA that are 4°C higher than Dz1 and Lz2, respectively). Cleavage at the 1096-site occurs between A and U whereas cleavage at the 1093-site is between G and U. The identity of the purine can affect cleavage [<xref ref-type="bibr" rid="B22">22</xref>], although this does not fully explain the large difference observed in Figure <xref ref-type="fig" rid="F3">3B</xref>.</p><p>The kinetic measurements at the 1093-site indicate that several parameters are involved in determining the rate of the cleavage reaction. The apparent rate constant <italic>k</italic><sub>obs </sub>for cleavage of the 17n substrate by the LNAzyme Lz1-1093 was fifty times greater than for the DNAzyme Dz1-1093 under single-turnover conditions with five times excess of enzyme. Although the excess of Dz1-1093 was probably not enough to saturate the target, such large differences in <italic>k</italic><sub>obs </sub>would not be expected if this were the only factor limiting the reaction. Thus, the LNA residues in the binding arms probably influence parameters other than hybridisation. One such parameter could involve the folding of the catalytic loop into an active conformation. The biphasic reaction of Lz1-1093 in itself suggests the existence of two conformations of the LNAzyme and/or two conformations of the LNAzyme/substrate complex.</p></sec><sec><title>Conclusion</title><p>In conclusion, we concur with Breaker's assertion that there is potential for improving the catalytic capacity of 10–23-like DNAzymes [<xref ref-type="bibr" rid="B23">23</xref>]. We show here that the ability of 10–23 DNAzymes to hybridize to their RNA substrates is an important factor in determining whether the target is cleaved, and in doing so we have improved the hybridization potential of several DNAzymes by the incorporation of locked nucleoside monomers into their binding arms. However, hybridization capacity is not the only parameter that can be altered in this way, and the apparent increased rate constant of the LNAzymes indicate that subsequent steps such as folding of the enzyme's catalytic loop have also, if somewhat fortuitously, been enhanced. The K<sub>m </sub>of 33 nM and the turnover number of 1.2 min<sup>-1 </sup>measured for Lz1-1093 show that the binding affinity and <italic>k</italic>cat of this LNAzyme are approaching values that are suitable for practical applications.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>DNAzymes and LNAzymes, 17n RNA, templates and primers</title><p>DNAzymes composed of unmodified deoxyoligonucleotide were obtained commercially. Modified DNAzymes with α-L-LNA- or LNA thymine monomers in the binding arms (LNAzymes) were synthesized using published procedures [<xref ref-type="bibr" rid="B11">11</xref>,<xref ref-type="bibr" rid="B14">14</xref>].</p></sec><sec><title>RNA substrates</title><p>The 17n RNA was chemically synthesized (MWG Biotech AG). The 33n, 58n and 74n RNAs were transcribed using T7 RNA polymerase (Promega) using buffer conditions recommended by the supplier, and 1 mM nucleoside triphosphates. The following deoxyoligonucleotides were used as transcription templates: 33n, AATGATGGCTGCTTCTAAGCCAACATCCTGGCCTATAGTGAGTCGTATTA; 58n, CCAGTGAGCTATTACGCTTTCTTTAAATGATGGCTGCTTCTAAGCC AACATCCTGGCCTATAGTGAGTCGTATTA; 74n, GGCGCATCCGCTAATTTTTCAACATTAGTCCGTTCGGTCCTCCAGTAGTGTTACCCAACCTTCAACCTGCTCCTATAGTGAGTCGTATTA. All sequences were hybridized to the complementary deoxyoligonucleotide TAATACGACTCACTATAGG to form the double-stranded T7 promoter.</p><p>After transcription, the RNA were extracted and precipitated and the full-length RNA was isolated on a 13% denaturing polyacrylamide-7M urea gel. RNAs were eluted with H<sub>2</sub>O from gel bands, were extracted with phenol/chloroform, and were recovered by ethanol precipitation before being redissolved in H<sub>2</sub>O. RNA transcripts were generally 5'-radiolabelled by dephosphorylating using Shrimp phosphatase (USB) followed by incubation with T4 polynucleotide kinase (New England Biolab) and [γ-<sup>32</sup>P]-ATP. Alternatively, RNAs were labelled in the chain by inclusion of [α-<sup>32</sup>P]-CTP in the transcription reaction. The 2904n RNA (23S rRNA) was extracted from ribosomes isolated from <italic>Escherichia coli </italic>strain IB10 [<xref ref-type="bibr" rid="B24">24</xref>] using previously published methods [<xref ref-type="bibr" rid="B25">25</xref>]; the 2904n was not radiolabelled.</p></sec><sec><title>RNA cleavage conditions for single and multiple-turnover analyses</title><p>The enzymes and RNA substrates were allowed to pre-equilibrate in the reaction buffer containing 150 mM NaCl and 50 mM Tris. HCl, pH 7.5 at 37°C for 5 minutes. The reaction was then initiated by addition of MgCl<sub>2 </sub>to a final concentration of 10 mM. In the single-turnover experiments in Figure <xref ref-type="fig" rid="F3">3</xref>, the concentration of the substrate RNA was 1 pmol in a reaction volume of 8 μl (125 nM). The deoxyribozyme was in excess at either 5 pmol or 50 pmol (625 nM or 6.25 μM) in experiments with the small RNAs, and at 10 or 20 pmol (1.25/2.5 μM) in experiments with the 2904n RNA. Single-turnover reactions with minimal RNAs were stopped after 1 h by quenching with one-third volume of ice-cold 90% formamide/20 mM EDTA. Samples were heated at 80°C for 2 min prior to loading onto 13% acrylamide polyacrylamide-7M gels.</p><p>The 23S rRNA was heated at 70°C for 30 sec and slow cooled to 37°C to enable refolding to its natural conformation before addition of the DNAzymes. Single-turnover reactions on the 2904n RNA were stopped by the addition of 20 mM EDTA at 0°C, and analyzed by primer extension as described below. RNA cleavage was detected by scanning the gels using phosphor imaging.</p><p>In the single-turnover experiments in Figure <xref ref-type="fig" rid="F4">4A</xref> and <xref ref-type="fig" rid="F4">4B</xref>, the concentration of the substrate RNA was 0.8 pmol in a reaction volume of 8 μl (100 nM) and the deoxyribozyme was at 4 pmol. In the multiple-turnover experiments shown in Figure <xref ref-type="fig" rid="F4">4C</xref>, the enzyme concentration was 30 nM while the substrate concentration was 0.6 μM. In the multiple-turnover experiments shown in Figure <xref ref-type="fig" rid="F4">4D</xref>, the enzyme concentration was kept constant at 1 nM while the substrate concentration was varied between 5 nM and 1.2 μM. Enzymes were pre-equilibrated with the <sup>32</sup>P-labelled 17n RNA substrate in 150 mM NaCl, 50 mM Tris. HCl, pH 7.5 and 0.01% SDS at 50°C for 2 min followed by a further 2 min at 37°C, before the reaction was initiated by addition of MgCl<sub>2 </sub>to 10 mM. Sample aliquots were removed during the reaction and quenched prior to gel analysis (as described above).</p></sec><sec><title>Analysis of cleavage in the 2904n RNA</title><p>Due to the large size of this RNA substrate, reverse transcriptase primer extension [<xref ref-type="bibr" rid="B26">26</xref>] was used to analyze cleavage. Briefly, AMV reverse transcriptase (Finnzymes) was used to extend two 5'-<sup>32</sup>P-end-labelled primers, 5'-CAAGTCATCCGCTAATTTT and 5'-GCCGACTCGACCAGTGAGC, complementary to 23S rRNA nucleotides 750 to 768 and 1099 to 1117, respectively, that are located immediately downstream from the cleavage sites. Transcripts stop either at the site of cleavage or immediately after on uncleaved RNA molecules due to incorporation of a dideoxynucleotide [<xref ref-type="bibr" rid="B27">27</xref>]. Extension products were analyzed on 13% polyacrylamide/7M urea gels alongside the sequencing reactions performed on uncleaved RNA [<xref ref-type="bibr" rid="B25">25</xref>].</p></sec></sec><sec><title>Authors' contributions</title><p>MDS and BRB performed the chemical synthesis of the LNAzymes. JW helped design and supervised the synthesis of the LNAzymes. LBL and LHH carried out the cleavage experiments. BV and SD conceived and designed the study, and wrote the manuscript. All authors read and approved the final manuscript.</p></sec> |
Applying the Theory of Planned Behavior to healthy eating behaviors in urban Native American youth | <sec><title>Background</title><p>To investigate the efficacy of the Theory of Planned Behavior (TPB) to predict healthy eating behavior in a group of urban Native American youth.</p></sec><sec sec-type="methods"><title>Methods</title><p>Native American boys and girls (n = 139), ages 9–18 years old, were given a self-administered survey to assess eating behavior using the TBP constructs (intention, attitude, subjective norm, barriers, self-efficacy, and perceived behavioral control). Youth were also measured for height and weight and body mass index (BMI) was calculated. Bivariate correlations and stepwise regression analyses of TBP model were performed with SPSS software.</p></sec><sec><title>Results</title><p>No association was found between intention and healthy eating behavior. However, independently healthy eating behavior was correlated with barriers (0.46), attitude (0.44), perceived behavioral control (0.35), and subjective norm (0.34). The most predictive barriers to eating healthy included the availability and taste of foods. Boys' eating behavior was most predicted by subjective norm, while girls' eating behavior was most predicted by barriers.</p></sec><sec><title>Conclusion</title><p>Lack of association between intention and healthy eating behavior suggests that factors other than intentions may drive healthy eating behaviors in urban Native American youth. Results indicate that programs promoting healthy eating to youth might focus on collaborating with families to make healthy foods more appealing to youth.</p></sec> | <contrib id="A1" contrib-type="author"><name><surname>Fila</surname><given-names>Stefanie A</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>fila@aps.edu</email></contrib><contrib id="A2" corresp="yes" contrib-type="author"><name><surname>Smith</surname><given-names>Chery</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>csmith@umn.edu</email></contrib> | The International Journal of Behavioral Nutrition and Physical Activity | <sec><title>Background</title><p>Several studies suggest that Native American youth have a higher prevalence of obesity than the general United States population [<xref ref-type="bibr" rid="B1">1</xref>-<xref ref-type="bibr" rid="B7">7</xref>]. The problem is not restricted to just Native American youth, increases in pediatric obesity have been noted globally [<xref ref-type="bibr" rid="B8">8</xref>]. Additionally, research suggests that obesity may persist into adulthood and increase the risk of chronic diseases including heart disease, increased blood pressure, and type 2 diabetes, thus making it a major public health concern [<xref ref-type="bibr" rid="B9">9</xref>,<xref ref-type="bibr" rid="B10">10</xref>]. Because of the known association between dietary intake and obesity [<xref ref-type="bibr" rid="B11">11</xref>,<xref ref-type="bibr" rid="B12">12</xref>], promoting healthy eating behaviors in youth could help decrease the elevated prevalence of obesity. While little is know about the dietary habits of urban Native American youth, research by Ballew et al. [<xref ref-type="bibr" rid="B13">13</xref>] found that 12–19 year-old rural youth participating in the Navajo Health and Nutrition Survey consumed fruits, vegetables, and dairy products less than once per day and mean intakes of several vitamins and minerals were below the sex-and-age specific recommended dietary allowances. Because preliminary research indicates poor dietary intake and a high prevalence of obesity among Native American youth, it seems prudent to examine whether youth have an interest in or intend to eat healthfully and to identify the factors that influence those intentions so that appropriate nutrition intervention programs could be developed to assist youth in changing their dietary behavior. Therefore, the purpose of this project was to learn more about why urban Native American youth eat the way that they do. Specifically, we were interested in identifying attitudes that promote, or create barriers, to healthful eating; identifying who or what promotes healthful dietary behavior; and examining to what extent the youth perceive control over their dietary behavior. We utilized the Theory of Planned Behavior (TPB) as the theoretical framework to accomplish this.</p><p>The TPB is often used to study health related decision making behavior in youth [<xref ref-type="bibr" rid="B14">14</xref>-<xref ref-type="bibr" rid="B19">19</xref>]. The TPB is an extension of the Theory of Reasoned Action (TRA) [<xref ref-type="bibr" rid="B20">20</xref>], but incorporates a third construct known as perceived behavioral control (PBC). The TPB model suggests that intention is directly driven by three major constructs attitude, subjective norm, and PBC and the stronger the intention, the more likely an individual will perform the behavior [<xref ref-type="bibr" rid="B21">21</xref>]. Attitude is known as the degree to which an individual has a favorable or unfavorable evaluation of the behavior, subjective norm measures the importance others hold about performing or not performing a behavior and one's willingness to comply to those referents, and PBC describes the perceived ease or difficulty an individual has for performing a behavior. In addition, PBC is thought to directly affect behavior by accounting for factors outside an individual's control and especially for behaviors not under volitional control [<xref ref-type="bibr" rid="B22">22</xref>]. Like many health behaviors, healthy eating is not under complete volitional control, as a result, perceived behavioral control becomes a more important determinant of behavior [<xref ref-type="bibr" rid="B23">23</xref>].</p><p>Self-efficacy is another term that has sometimes been used to define PBC. Self-efficacy is a component of Bandura's [<xref ref-type="bibr" rid="B24">24</xref>] Social Cognitive Theory and is defined as an individual's perceived ability to perform a behavior. Ajzen [<xref ref-type="bibr" rid="B23">23</xref>] considers the PBC construct of TPB identical to self-efficacy. However, researchers have not yet come to agreement about this. Self-efficacy has been shown to be an independent contributor to eating behavior [<xref ref-type="bibr" rid="B25">25</xref>] and several studies support a distinction between self-efficacy and PBC when applying the TPB to health behaviors [<xref ref-type="bibr" rid="B26">26</xref>-<xref ref-type="bibr" rid="B28">28</xref>]. Armitage and Conner [<xref ref-type="bibr" rid="B27">27</xref>] applied the TPB to eating a low-fat diet in a group of undergraduate students. They differentiated self-efficacy, the individual's internal motivation to eat a low fat diet, from PBC, the extent to which an individual has control over external factors related to eating a low fat diet [<xref ref-type="bibr" rid="B27">27</xref>]. In another study, Armitage and Conner [<xref ref-type="bibr" rid="B27">27</xref>] found that when self-efficacy was added to the TPB model, it was not only an important contributor but it was often the most important predictor of both behavior and intention. Giles et al. [<xref ref-type="bibr" rid="B29">29</xref>] suggested that the predictive utility of the TPB may be enhanced by replacing PBC with self-efficacy. Alternatively, some research indicates that self-efficacy is not a useful addition to the theory. For example, Terry and O'Leary [<xref ref-type="bibr" rid="B28">28</xref>] suggest that self-efficacy may not be an important predictor of behavior. Based on previous research findings, this current study examined self-efficacy and PBC as separate constructs that could indirectly, through intention, or directly influence healthy eating behavior in urban Native American youth.</p><p>Previous research has also investigated adolescents' knowledge and perceptions of healthy eating and found that while adolescents are informed about healthy eating practices and recommendations, they find it difficult to follow recommendations and often ate unhealthy foods because they perceived too many barriers to eat healthfully [<xref ref-type="bibr" rid="B30">30</xref>,<xref ref-type="bibr" rid="B31">31</xref>]. According to the Health Belief Model, perceived barriers are an individual's opinion of the tangible costs of an action or behavior [<xref ref-type="bibr" rid="B32">32</xref>]. In addition to the Health Belief Model, perceived barriers were also used as a construct to determine how well the TBP predicted fruit and vegetable consumption in adolescents [<xref ref-type="bibr" rid="B15">15</xref>]. Since perceived barriers appear to be a determinant of healthy eating behavior, and may also indirectly affect intention, this study also included barriers as a construct in the TBP model.</p><p>This current study used an expanded TBP model which incorporates the original constructs of attitude, subjective norm, and PBC, as well as two additional constructs, barriers and self-efficacy, to investigate healthy eating behaviors in urban Native American youth (Figure <xref ref-type="fig" rid="F1">1</xref>). Because of the high prevalence of obesity among Native American youth, and the association between diet and weight gain, results from this study are important for the development of intervention strategies that promote healthy eating behaviors in urban Native American youth who are overweight or at risk for becoming overweight.</p><fig position="float" id="F1"><label>Figure 1</label><caption><p>Study model of healthy eating based on the TPB.</p></caption><graphic xlink:href="1479-5868-3-11-1"/></fig></sec><sec sec-type="methods"><title>Methods</title><sec><title>Participants</title><p>Participants were 139 urban Native American youth, predominately from Ojibwe and Lakota tribes, attending an after school program in Minneapolis, Minnesota. The goal of the program was to promote cultural identity, academic success, physical well being, and chemical dependency awareness to Native American youth ages 5 to 18 years old. For this study, surveys were self-administered only to youth 9 to 18 years. This sample included 58 boys (mean age: 12.5 + 2.6 yrs) and 81 girls (mean age: 12.4 + 2.4). Parental consent was given for all activities at the time of program enrollment, and youth were asked to complete the survey voluntarily. Youth who completed the survey and anthropometric measurements received a modest monetary incentive for their participation. Of the approximately 150 youth (ages 9 to 18 yrs) attending the program at the time of this project, 93% volunteered to participate. The project was approved by the Institutional Review Board of Human Subjects at the University of Minnesota.</p></sec><sec><title>Measures</title><p>The survey was developed according to procedures defined by Ajzen and Fishbein [<xref ref-type="bibr" rid="B20">20</xref>]. Formative assessment included a review of literature and six focus group discussions with 39 youth, ages 9 to 18 years, to determine common beliefs about eating healthy, advantages and barriers to eating healthy, and important people who may influence behavior. Focus groups were audio taped and transcribed verbatim and analyzed for common themes. The themes were used to develop a survey using the constructs of the TPB to investigate healthy eating behavior. These include the original constructs of attitudes, subjective norm, PBC, as well as two additional constructs, self-efficacy and barriers (Figure <xref ref-type="fig" rid="F1">1</xref>). Self-efficacy was included to assess whether it was more predictive of intention and/or behavior than PBC. The construct 'barriers' was included in the model because the staff of the program were particularly interested in identifying factors that prevent youth from eating healthfully. The survey included 90 questions pertaining to eating behavior and measured all constructs. It was evaluated by an expert panel of nutritionists accomplished in the areas of survey development and behavioral theory for face validity and breadth of coverage and then pilot tested in a group of urban Native American youth (n = 32) attending an alternative high school for ease of comprehension and readability. The final version of the survey contained slight revisions in wording and question ordering based on recommendations from professionals and findings from the pilot study. Scales for the final survey were assessed using the Cronbach alpha coefficient, an index of inter-item homogeneity (internal consistency.) The reliability levels for the: attitude scale was 0.66, subjective norm scale was 0.66, PBC scale was 0.80, intention scale was 0.84, behavior scale was 0.72, barriers scale was 0.89, and self-efficacy scale was 0.85. These scores indicate a substantial (0.61 – 0.80) to almost perfect (0.81–1.0) range of reliability [<xref ref-type="bibr" rid="B32">32</xref>].</p><p>Survey questions were assessed using a five-point Likert response scale, youth had to choose one of the following responses "strongly agree", "agree", "unsure", "disagree", or "strongly disagree". For positively scaled questions responses were coded from 2 to -2 (strongly agree to strongly disagree) and for negatively scaled questions responses were coded from -2 to 2 (strongly agree to strongly disagree). Attitudes to eating healthy were measured by using the responses to eighteen questions about the importance of eating healthy foods, fruits, vegetables, regular pop, junk food, and fast food, and perceptions of eating healthy, being under or overweight, and diabetes (e.g. <italic>"It is important to me to eat healthy foods everyday"</italic>). Subjective norm was measured by using the responses to eight questions asking if parents, friends, elders, community programs, or television told youth to eat healthy everyday (e.g.<italic>"My parents tell me it is important to eat healthy everyday"</italic>). PBC measured external factors that may directly or indirectly affect healthy eating behavior. Responses to eight questions concerning youth's perceived control over eating healthy, eating junk food, drinking regular pop, eating fast food, eating in front of the TV, and getting diabetes, as well as having fruits and vegetables available (e.g. "<italic>I have control over whether or not I eat healthy"</italic>) were used to determine perceived behavior control. Intention to eat healthy was measured by using the responses to eight questions regarding youth's plans for the next week to eat healthy, eat vegetables, eat fruit, not eat junk food, not eat fast food, not drink regular pop, eat healthy foods in front of the TV, and eat healthy foods to keep a healthy weight everyday (e.g. "<italic>For the next week I plan to eat healthy everyday"</italic>). Eating behavior was measured by using the responses to questions assessing dietary intake of vegetables, fruits, soft drinks, and fast food consumption, along with eleven additional behavior questions using the Likert-scale. Eating behaviors questions included general questions such as, "I mostly eat healthy foods," "I eat healthy to keep me from getting diabetes," and "I eat junk food when I watch TV," but also used specific foods such as "fruits" and "vegetables" because youth defined foods, especially fruits and vegetables, as being healthy. Both healthy and junk foods were defined on the survey using terminology that the youth used during the focus group discussions. The assessment and behavioral question scales were were averaged, recoded, and the sum of the two were calculated to measure youth's behavior.</p><p>The two additional constructs, barriers and self-efficacy, were included in the expanded model. Barriers to eating healthy were measured by using the responses to fourteen questions concerning youth's perceptions about the taste of fruits, vegetables, regular pop, and junk food; the ease of eating healthy away from home, with friends, with family, in front of the TV and to keep a healthy weight; and availability of healthy foods (e.g. <italic>"It is hard for me to eat healthy foods at fast food restaurants"</italic>). Self-efficacy measured internal factors that may directly or indirectly affect healthy eating behavior. Responses to fifteen questions determining youth's ability to eat healthy foods and choose specific healthy foods (fruit, vegetables, chocolate milk, juice, white milk, low-fat milk, salads) over specific unhealthy foods (junk food, chips/cheetos, regular pop, chocolate milk, whole milk, hamburgers) (e.g. <italic>"I can eat healthy foods everyday"</italic>) were used to assess self-efficacy to eat healthy.</p></sec><sec><title>Survey administration</title><p>Youth agreeing to participate in the study were seated in a quiet area and given a survey and writing instrument. Researchers explained to each youth that the purpose of the survey was to find out about how they feel about healthy eating and stressed that there were no right or wrong answers. Definitions for healthy eating, junk food, and fast food were provided on each survey and read aloud to each youth. These terms were identified and defined by youth attending the focus group discussions. Healthy eating was defined as eating different types of food from all food groups like bread, grains, cereals, fruit, vegetables, milk, and meat while limiting sugary and fatty foods; junk food was defined as foods such as pop, chips, candy, donuts, cakes, cookies, and sweets; and fast food was defined as burgers, fries, shakes and foods from McDonald's, Burger King, Taco Bell, KFC etc. Youth were asked to carefully read each question and mark only one response. For youth who had problems answering questions or understanding questions, several researchers were available to aid with completion. The time taken to complete each survey averaged 20 minutes with a range between 10 and 30 minutes. All surveys were rechecked to identify omitted or multiple response questions and youth were immediately instructed to finish or clarify unfinished or unclear responses. Four surveys were removed from analysis because of missing data.</p></sec><sec><title>Anthropometric measurements</title><p>After completing the survey each respondent was measured in light clothing with their shoes removed. Height was measured to the nearest one-tenth centimeter with a GPM anthropometer (Switzerland). Weight was measured to the nearest one-tenth kilogram with a high-quality electronic scale (Seca, France). Body mass index (BMI) was calculated by dividing weight in kilograms by height in meters squared and the BMI for each youth was plotted on age-sex-specific growth charts developed by the Centers for Disease Control and Prevention [<xref ref-type="bibr" rid="B33">33</xref>]. Youth were classified according to their BMI percentile as either underweight (<5<sup>th </sup>percentile), normal weight (≥ 5<sup>th </sup>to <85<sup>th </sup>percentile), at risk for overweight (≥ 85<sup>th </sup>to <95<sup>th </sup>percentile), or overweight (≥ 95<sup>th </sup>percentile) [<xref ref-type="bibr" rid="B34">34</xref>].</p></sec><sec><title>Data analysis</title><p>Data were analyzed using the Statistical Package for Social Sciences for Windows (SSPS, v. 11, Chicago, IL, 1999). Descriptive statistics were used to determine means and standard deviations of all constructs as well as age, gender, and BMI. Mean values for the constructs for boys and girls were compared using independent t-tests to identify any gender or age differences. No differences were found based on age.</p><p>Because gender differences were observed all analyses were conducted by gender. Pearson correlations were conducted to examine associations between the psychosocial variables (constructs) for both the original TPB model, then for the expanded TPB model. Because the PBC construct was only weakly associated to girl's behavior and not associated to girl's intention or boy's behavior or intention, the expanded version of TPB model was used for further analyses. In order to examine constructs (attitude, subjective norm, PBC, self-efficacy, and barriers) most predictive of intention and behavior, stepwise regression analyses were performed respectively. Additionally, further stepwise regression analyses were conducted with the construct that was most predictive of behavioral change for boys (subjective norms) and girls (barriers). Both of these issues will need to be addressed in future interventions if behavioral change is to be achieved. The level of significance was set at <italic>p </italic>< 0.05 for all statistical tests.</p></sec></sec><sec><title>Results</title><p>The mean age and grade of boys and girls was not significantly different, 12.4 years old and 7<sup>th </sup>grade respectively. According to CDC BMI percentiles 38% of girls and 38% of boys were classified as normal weight, 20% of girls and 26% of boys as at risk for overweight, and 42% girls and 36% of boys as overweight.</p><sec><title>Boys</title><p>No association was found between intention and healthy eating behavior. Furthermore, the grade, age, and BMI were not associated with intention or behavior. However, the constructs of attitude, subjective norm, PBC, and barriers were all associated with behavior, and the constructs of attitude, subjective norm, and self-efficacy were all associated with intention (see Table <xref ref-type="table" rid="T1">1</xref>). The mean value of barriers for boys (-0.006) was significantly higher than the mean value of barriers for girls (-0.032) (Table <xref ref-type="table" rid="T1">1</xref>).</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>Correlations between TBP model constructs. Means and standard deviations (SD) for each construct for the total sample and by gender.</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td align="left">1</td><td align="left">2</td><td align="left">3</td><td align="left">4</td><td align="left">5</td><td align="left">6</td><td align="left">7</td></tr></thead><tbody><tr><td align="left" colspan="8">Total sample (n = 139)</td></tr><tr><td colspan="8"><hr></hr></td></tr><tr><td align="left">1. Behavior</td><td align="left">1</td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">2. Intention</td><td align="left">0.05</td><td align="left">1</td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">3. Attitude</td><td align="left">0.44**</td><td align="left">0.45**</td><td align="left">1</td><td></td><td></td><td></td><td></td></tr><tr><td align="left">4. Subjective norm</td><td align="left">0.34**</td><td align="left">0.39**</td><td align="left">0.48**</td><td align="left">1</td><td></td><td></td><td></td></tr><tr><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">5. Barriers</td><td align="left">0.46**</td><td align="left">-0.14</td><td align="left">0.40**</td><td align="left">0.15</td><td align="left">1</td><td></td><td></td></tr><tr><td align="left">6. Self-efficacy</td><td align="left">-0.12</td><td align="left">0.56**</td><td align="left">0.38**</td><td align="left">0.37**</td><td align="left">-0.31**</td><td align="left">1</td><td></td></tr><tr><td align="left">7. Perceived behavioral control</td><td align="left">0.35**</td><td align="left">-0.01</td><td align="left">0.35**</td><td align="left">0.23**</td><td align="left">0.57**</td><td align="left">-0.04</td><td align="left">1</td></tr><tr><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">Mean</td><td align="left">0.048</td><td align="left">0.062</td><td align="left">0.023</td><td align="left">0.067</td><td align="left">-0.021</td><td align="left">0.034</td><td align="left">0.003</td></tr><tr><td align="left">SD</td><td align="left">0.034</td><td align="left">0.093</td><td align="left">0.023</td><td align="left">0.079</td><td align="left">0.054</td><td align="left">0.043</td><td align="left">0.097</td></tr><tr><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">Boys (n = 58)</td><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">1. Behavior</td><td align="left">1</td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">2. Intention</td><td align="left">0.15</td><td align="left">1</td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">3. Attitude</td><td align="left">0.45**</td><td align="left">0.60**</td><td align="left">1</td><td></td><td></td><td></td><td></td></tr><tr><td align="left">4. Subjective norm</td><td align="left">0.46**</td><td align="left">0.38**</td><td align="left">0.58**</td><td align="left">1</td><td></td><td></td><td></td></tr><tr><td align="left">5. Barriers</td><td align="left">0.37**</td><td align="left">0.13</td><td align="left">0.47**</td><td align="left">0.37**</td><td align="left">1</td><td></td><td></td></tr><tr><td align="left">6. Self-efficacy</td><td align="left">0.2</td><td align="left">0.68**</td><td align="left">0.65**</td><td align="left">0.38**</td><td align="left">0</td><td align="left">1</td><td></td></tr><tr><td align="left">7. Perceived behavioral control</td><td align="left">0.43**</td><td align="left">0.22</td><td align="left">0.46**</td><td align="left">0.33*</td><td align="left">0.59**</td><td align="left">0.18</td><td align="left">1</td></tr><tr><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">Mean</td><td align="left">0.05</td><td align="left">0.032†</td><td align="left">0.023</td><td align="left">0.066</td><td align="left">-0.006†</td><td align="left">0.030†</td><td align="left">0.019</td></tr><tr><td align="left">SD</td><td align="left">0.032</td><td align="left">0.09</td><td align="left">0.028</td><td align="left">0.086</td><td align="left">0.056</td><td align="left">0.045</td><td align="left">0.096</td></tr><tr><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">Girls (n = 81)</td><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">1. Behavior</td><td align="left">1</td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">2. Intention</td><td align="left">0.01</td><td align="left">1</td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">3. Attitude</td><td align="left">0.46**</td><td align="left">0.34**</td><td align="left">1</td><td></td><td></td><td></td><td></td></tr><tr><td align="left">4. Subjective norm</td><td align="left">0.25*</td><td align="left">0.41**</td><td align="left">0.37**</td><td align="left">1</td><td></td><td></td><td></td></tr><tr><td align="left">5. Barriers</td><td align="left">0.52**</td><td align="left">-0.26*</td><td align="left">0.35**</td><td align="left">-0.5</td><td align="left">1</td><td></td><td></td></tr><tr><td align="left">6. Self-efficacy</td><td align="left">-0.34**</td><td align="left">0.41**</td><td align="left">0.08</td><td align="left">0.38**</td><td align="left">-0.54**</td><td align="left">1</td><td></td></tr><tr><td align="left">7. Perceived behavioral control</td><td align="left">0.28*</td><td align="left">-0.11</td><td align="left">0.25*</td><td align="left">0.16</td><td align="left">0.53**</td><td align="left">-0.16</td><td align="left">1</td></tr><tr><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">Mean</td><td align="left">0.046</td><td align="left">0.084†</td><td align="left">0.023</td><td align="left">0.067</td><td align="left">-0.032†</td><td align="left">0.046†</td><td align="left">-0.009</td></tr><tr><td align="left">SD</td><td align="left">0.036</td><td align="left">0.089</td><td align="left">0.02</td><td align="left">0.075</td><td align="left">0.05</td><td align="left">0.039</td><td align="left">0.096</td></tr></tbody></table><table-wrap-foot><p>* <italic>p </italic>< 0.05 ** <italic>p </italic>< 0.01 † Mean value significantly different than mean value of boys/girls at <italic>p </italic>< 0.05.</p></table-wrap-foot></table-wrap><p>Stepwise regression of behavior as the dependent variable and constructs (attitudes, subjective norms, PBC, barriers, and self-efficacy) as independent variables, showed that subjective norm (R<sup>2 </sup>= 0.21) and perceived behavioral control (R<sup>2 </sup>= 0.30) predicted 30% of the model (Table <xref ref-type="table" rid="T2">2</xref>). Whereas when intention was regressed with the constructs as independent variables, self-efficacy accounted for 46% of the variance and attitude added another 4%. Furthermore, because subjective norm was most predictive of healthy eating behaviors we ran it as a dependent variable with all survey questions measuring subjective norm as independent variables to see which of the social norms were most predictive of that construct. Three questions explained 86% of the variance in behavior (Table <xref ref-type="table" rid="T2">2</xref>). The three questions included <italic>"My family tells me I should eat healthy everyday to help keep me at a healthy weight" </italic>(R<sup>2 </sup>= 0.57), <italic>"I get hungry for foods I see on TV" </italic>(R<sup>2 </sup>= 0.74), and <italic>"The after school program I attend says it is important to eat healthy everyday" </italic>(R<sup>2 </sup>= 0.86).</p><table-wrap position="float" id="T2"><label>Table 2</label><caption><p>Stepwise regression analyses of healthy eating behavior.</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left"><bold>Variable</bold></td><td align="center"><bold>B</bold></td><td align="center"><bold>Standard error</bold></td><td align="center"><bold>R Square</bold></td><td align="center"><bold>P Value</bold></td></tr></thead><tbody><tr><td align="left"><bold>Boys and Girls</bold></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">Dependent variable: Behavior<sup>a</sup></td><td></td><td></td><td></td><td></td></tr><tr><td align="left"> Barriers</td><td align="center">0.121</td><td align="center">0.058</td><td align="center">0.207</td><td align="center">0.038</td></tr><tr><td align="left"> Attitude</td><td align="center">0.519</td><td align="center">0.142</td><td align="center">0.286</td><td align="center">0.000</td></tr><tr><td align="left"> Self-efficacy</td><td align="center">-0.229</td><td align="center">0.074</td><td align="center">0.315</td><td align="center">0.002</td></tr><tr><td align="left"> Subjective norm</td><td align="center">0.105</td><td align="center">0.035</td><td align="center">0.357</td><td align="center">0.004</td></tr><tr><td align="left">Dependent variable: Barriers<sup>b</sup></td><td></td><td></td><td></td><td></td></tr><tr><td align="left"> Healthy foods not around</td><td align="center">0.006</td><td align="center">0.000</td><td align="center">0.614</td><td align="center">0.000</td></tr><tr><td align="left"> Healthy foods don't fill you up</td><td align="center">0.005</td><td align="center">0.000</td><td align="center">0.762</td><td align="center">0.000</td></tr><tr><td align="left"> Junk food taste better than healthy food</td><td align="center">0.005</td><td align="center">0.000</td><td align="center">0.830</td><td align="center">0.000</td></tr><tr><td align="left"> Friends make it hard to eat healthy</td><td align="center">0.006</td><td align="center">0.000</td><td align="center">0.882</td><td align="center">0.000</td></tr><tr><td align="left"> Fruits don't taste good</td><td align="center">0.005</td><td align="center">0.000</td><td align="center">0.913</td><td align="center">0.000</td></tr><tr><td align="left"><bold>Boys</bold></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">Dependent variable: Behavior<sup>a</sup></td><td></td><td></td><td></td><td></td></tr><tr><td align="left"> Subjective norm</td><td align="center">0.133</td><td align="center">0.045</td><td align="center">0.212</td><td align="center">0.004</td></tr><tr><td align="left"> Perceived behavioral control</td><td align="center">0.104</td><td align="center">0.040</td><td align="center">0.299</td><td align="center">0.011</td></tr><tr><td align="left">Dependent variable: Subjective norm<sup>c</sup></td><td></td><td></td><td></td><td></td></tr><tr><td align="left"> Family says eat healthy to keep weight</td><td align="center">3.125E-02</td><td align="center">0.000</td><td align="center">0.565</td><td align="center">0.000</td></tr><tr><td align="left"> I get hungry for foods on TV</td><td align="center">1.563E-02</td><td align="center">0.000</td><td align="center">0.735</td><td align="center">0.000</td></tr><tr><td align="left"> After school program says eat healthy</td><td align="center">1.562E-02</td><td align="center">0.000</td><td align="center">0.862</td><td align="center">0.000</td></tr><tr><td align="left"> Friends say eat healthy</td><td align="center">1.563E-02</td><td align="center">0.000</td><td align="center">0.927</td><td align="center">0.000</td></tr><tr><td align="left"><bold>Girls</bold></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">Dependent variable: Behavior<sup>a</sup></td><td></td><td></td><td></td><td></td></tr><tr><td align="left"> Barriers</td><td align="center">0.173</td><td align="center">0.082</td><td align="center">0.267</td><td align="center">0.037</td></tr><tr><td align="left"> Attitude</td><td align="center">0.538</td><td align="center">0.187</td><td align="center">0.354</td><td align="center">0.005</td></tr><tr><td align="left"> Self-efficacy</td><td align="center">-0.310</td><td align="center">0.104</td><td align="center">0.388</td><td align="center">0.004</td></tr><tr><td align="left"> Subjective norm</td><td align="center">0.134</td><td align="center">0.047</td><td align="center">0.446</td><td align="center">0.006</td></tr><tr><td align="left">Dependent variable: Barriers<sup>b</sup></td><td></td><td></td><td></td><td></td></tr><tr><td align="left"> Healthy foods are not around</td><td align="center">5.102E-03</td><td align="center">0.000</td><td align="center">0.573</td><td align="center">0.000</td></tr><tr><td align="left"> Parent's don't buy healthy foods</td><td align="center">5.102E-03</td><td align="center">0.000</td><td align="center">0.738</td><td align="center">0.000</td></tr><tr><td align="left"> Junk food taste better than healthy food</td><td align="center">5.102E-03</td><td align="center">0.000</td><td align="center">0.819</td><td align="center">0.000</td></tr><tr><td align="left"> Healthy foods don't fill you up</td><td align="center">5.102E-03</td><td align="center">0.000</td><td align="center">0.868</td><td align="center">0.000</td></tr><tr><td align="left"> Hard to eat healthy away from home</td><td align="center">5.102E-03</td><td align="center">0.000</td><td align="center">0.907</td><td align="center">0.000</td></tr></tbody></table><table-wrap-foot><p><sup>a </sup>Candidate variables: intention, attitude, subjective norm, barriers, self-efficacy, perceived behavioral control.</p><p><sup>b </sup>Candidate variables: fruits don't taste good, vegetables don't taste good, regular pop tastes better than diet, healthy foods don't fill you up, hard to eat healthy at fast food restaurants, hard to eat healthy away from home, hard to eat healthy because junk food tastes better, I am not use to eating healthy, healthy foods are not around, parent's don't buy healthy food, family makes it hard to eat healthy, friends make it hard to eat healthy, TV makes it hard to eat healthy, hard to eat healthy to keep at a healthy weight.</p><p><sup>c </sup>Parents say eat healthy, after school program says eat healthy, friends say eat healthy, Elders say eat healthy, no one tells me not to eat junk food, I get hungry for foods on TV, family says eat healthy to keep weight, family says eat healthy to prevent diabetes.</p></table-wrap-foot></table-wrap></sec><sec><title>Girls</title><p>No association was found between intention and behavior, therefore factors directly affecting behavior and intention were investigated. Healthy eating behavior was positively correlated with attitude, subjective norm, perceived behavior control, and barriers, while self-efficacy had a negative correlation (Table <xref ref-type="table" rid="T1">1</xref>). While the constructs of attitude, subjective norm, and self-efficacy were all positively associated with intention, and barriers was negatively associated with it (see Table <xref ref-type="table" rid="T1">1</xref>). The mean value of healthy eating intention (0.084) and self-efficacy (0.046) for girls was significantly higher than the mean value of intention (0.032) and self-efficacy (0.030) for boys (Table <xref ref-type="table" rid="T1">1</xref>).</p><p>Stepwise regression of behavior as the dependent variable and constructs (attitudes, subjective norms, PBC, barriers, and self-efficacy) as independent variables, showed that barriers (R<sup>2 </sup>= 0.27), attitude (R<sup>2 </sup>= 0.35), self-efficacy (R<sup>2 </sup>= 0.39), and subjective norm (R<sup>2 </sup>= 0.45) predicted 45% of the model (Table <xref ref-type="table" rid="T2">2</xref>). Whereas when intention was regressed with the constructs as independent variables, subjective norm (R<sup>2 </sup>= 0.18), attitude (R<sup>2 </sup>= 0.30), and barriers (R<sup>2 </sup>= 0.35) predicted 35% of the variance (Table <xref ref-type="table" rid="T2">2</xref>). Furthermore, because barriers was most predictive of healthy eating behaviors, we ran it as a dependent variable with all survey questions measuring subjective norm as independent variables to see which of the barriers were most predictive of that construct. Four questions were identified as explaining 82% of the variance in barriers to eat healthy (Table <xref ref-type="table" rid="T2">2</xref>). The four questions included <italic>"I don't eat healthy because healthy foods are not around" </italic>(R<sup>2 </sup>= 0.57), <italic>"My parent's don't buy healthy food" </italic>(R<sup>2 </sup>= 0.74), <italic>"It is hard for me to eat healthy foods because junk foods taste better" </italic>(R<sup>2 </sup>= 0.82), and <italic>"Healthy foods don't fill you up" </italic>(R<sup>2 </sup>= 0.87).</p></sec></sec><sec><title>Discussion</title><p>In this investigation of the healthy eating behaviors in urban Native American youth, TPB was found to be predictive of factors affecting healthy eating intention and behavior independently, but found no direct association between intention to eat healthfully and eating behavior. Furthermore, gender, but not age, grade, or BMI, was found to be a significant factor in youths' responses. Among boys, intention and behavior were predicted by different constructs, with self-efficacy accounting for 46% of the variance for intention. Among the girls, three of the constructs (subjective norms, attitude, and barriers) entered the predictive equation in different ranking order for both intention and behavior and account for much of the variance.</p><p>The lack of association between intention and behavior might be explained by the concept of intention instability. Conner et al. found that intentions were stronger predictors of behavior when intentions were stable in adults eating a low-fat diet [<xref ref-type="bibr" rid="B35">35</xref>]. In the present study, youths' intention to eat healthy may be driven more by external cues and therefore constantly changing. As a result, behavior is affected to a greater extent by other factors and not intention. The TPB is based on the concept that the stronger the intention to perform a given behavior, the greater the likelihood that the person will perform that behavior [<xref ref-type="bibr" rid="B36">36</xref>]. In the current study, intention could not be considered strong because mean intention for boys was found to be 0.062 and girls was found to be 0.32 on a scale from -2 (low intention) to 2 (high intention). Forming strong intentions to eat healthy may not be a priority in youth and therefore do not affect eating behavior.</p><p>In comparing boys and girls, healthy eating behavior among boy was predicted by subjective norm and PBC, while among girls barriers, attitude, self-efficacy, and subjective norms predicted behavior. In addition, behavior in girls was positively associated with age and negatively associated with self-efficacy. It appears that as girls become older they are more likely to engage in healthy eating behavior, however a low self-efficacy also seems to be associated with healthy eating behavior. Results are puzzling and may be explained by girls' body dissatisfaction and distortions of a healthy body size. A study examining body perceptions among urban Native American youth found 61% of girls expressed a desire to be thinner compared to 41% of boys and 26% of girls selected the thinnest silhouette as the healthiest from range of eight varying sized silhouettes [<xref ref-type="bibr" rid="B7">7</xref>]. Compared to boys, girls had significantly higher mean values for intention and self-efficacy, and lower mean values for barriers to eat healthy. In another study examining food choice in youth, girls also reported greater intentions to eat fruit than boys [<xref ref-type="bibr" rid="B19">19</xref>]. Healthy eating promotion programs may benefit by offering gender separate activities that aim to enhance intention and self-efficacy in boys and reduce perceived barriers in girls.</p><p>Subjective norm and PBC explained 30% of the variance to eat healthy in boys while barriers, attitude, subjective norm, and self-efficacy explained 45% of the variance to eat healthy in girls. It appears that girls' behaviors are influenced by more factors and to a greater extent than boys. Subjective norm was the best predictor of healthy eating behavior in boys with family, television, after school programs, and friends being most instrumental in influencing eating behaviors. According to a review of the TPB's application to health related behaviors, the subjective norm construct often did not reach significance and had less influence on behavior than attitude and PBC constructs [<xref ref-type="bibr" rid="B37">37</xref>]. This was not found to be true in the sample of urban Native American boys in the present study. The primary social unit of Native Americans is the extended family and their culture is based on respect for elders and strong community ties [<xref ref-type="bibr" rid="B38">38</xref>]. A strong sense of family and community support may explain why youths' eating behaviors were directly affected by the subjective norm construct. For girls, similar to data from the total sample, barriers were most predictive of eating behavior with unavailability of healthy foods and taste cited as the greatest barriers. Including the entire Native American community in promotion strategies appears to be an appropriate method to enhance healthy eating behaviors in urban Native American youth, especially in boys and could also be effective in reducing perceived barriers regarding food availability in girls.</p><p>This study assessed self-reported eating behavior as part of the survey and as a result could have introduced inaccuracies. However, the authors feel that the this group of Native American youth, who are accustom to completing surveys during the after school program, were attentive when answering survey questions, thus minimizing potential problems. Additionally, youth were shown food models to increase the accuracy of their intake estimates. Measuring actual food intake is very labor intensive and dietary self-report is the method primarily used in studies assessing the TPB and eating practices in youth [<xref ref-type="bibr" rid="B15">15</xref>,<xref ref-type="bibr" rid="B16">16</xref>,<xref ref-type="bibr" rid="B18">18</xref>,<xref ref-type="bibr" rid="B19">19</xref>,<xref ref-type="bibr" rid="B39">39</xref>]. In addition, self-reported low-fat eating behaviors were better predicted by TPB than behavior measured by dietary observation in adults [<xref ref-type="bibr" rid="B27">27</xref>] and prediction was also found superior than observed behavior in a meta-analysis of the TPB [<xref ref-type="bibr" rid="B26">26</xref>]. Future studies may benefit from using multiple measurements of eating behavior to avoid potential inaccuracies from self-report.</p><p>This study only examined a sample of urban Native American youth and findings may not be true for all urban Native American youth. Tribe identity and regional location could alter the factors influencing healthy eating behavior in Native American youth. However, authors believe that findings are appropriate for basing future research and practice in populations of urban Native American youth.</p></sec><sec><title>Conclusion</title><p>Findings indicate that TPB is useful for predicting factors directly related to healthy eating behavior but not for predicting the indirect effect of intention in a sample of urban Native American youth. This suggests that other factors besides intention are driving healthy eating behavior and can be used to develop intervention strategies to promote healthy eating practices in youth who overweight or at risk for overweight. Since barriers, attitude, subjective norm, and self-efficacy appear to be factors affecting healthy eating behaviors they should be incorporated in program design. Nutrition professionals should work with Native American community leaders and elders to provide sound nutritional knowledge to the entire community. This data also suggests that working through the family is important. Encouraging parents and caretakers to purchase and make healthy foods regularly available to their children could reduce youths' perceived barriers to healthy eating. Because the extended family and community are a valuable component of Native American culture, youth appear more willing to accept and follow dietary advice from members of their community. Gender differences may require the need for separate programs or at least special considerations for boys and girls. Boys who have lower self-efficacy but seem receptive to subjective norms, would benefit more from activities involving family and peers to increase healthy eating behaviors. Girls who are most affected by barriers to healthy eating, would benefit more from programs designed to increase the availability of healthy foods and promote the awareness that healthy foods are also tasty. Because of the alarming prevalence of obesity in urban Native American youth, future studies should continue to investigate the factors influencing obesity, such as eating and activity behavior, to identify the most effective way to solve this problem.</p></sec><sec><title>Competing interests</title><p>The author(s) declare that they have no competing interests.</p></sec> |
A trial platform to develop a tailored theory-based intervention to improve professional practice in the disclosure of a diagnosis of dementia: Study protocol [ISRCTN15871014] | <sec><title>Background</title><p>For people with dementia, care should include an explanation of the diagnosis to individuals and their carers, and information about the likely prognosis and possible packages of care. However, this is neither routine nor inevitable, and there is wide variation in the practice of disclosure. The aim of this study is to develop a tailored theory-based intervention to promote appropriate disclosure of diagnosis of dementia.</p></sec><sec sec-type="methods"><title>Methods</title><p>There are three objectives. Objective 1 is to define and develop an appropriate model of disclosure; this will be addressed using a multidisciplinary consensus development process. Objective 2 is to identify factors that influence disclosure of diagnosis; a questionnaire based upon theoretical constructs from a range of behavioural theories will be developed and members of old age mental health teams will be surveyed. The analysis will identify those factors that best predict intention to disclose a diagnosis to a person with dementia. Objective 3 is to develop and pilot test a theory-based intervention to promote disclosure of diagnosis that targets attitudes, beliefs and actions most amenable to change. Objective 3 will use the results of Objectives 1&2 to design and pilot test an intervention to improve the process of and increase the proportion of individuals receiving a diagnosis of dementia, for members of old age mental health teams. This work will lead to a proposal for a randomised controlled trial of the intervention.</p></sec> | <contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Eccles</surname><given-names>Martin P</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>martin.eccles@ncl.ac.uk</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Foy</surname><given-names>Robbie</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>r.c.foy@ncl.ac.uk</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Bamford</surname><given-names>Claire H</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>c.h.bamford@ncl.ac.uk</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Hughes</surname><given-names>Julian C</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>j.c.hughes@ncl.ac.uk</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Johnston</surname><given-names>Marie</given-names></name><xref ref-type="aff" rid="I4">4</xref><email>m.johnston@abdn.ac.uk</email></contrib><contrib id="A6" contrib-type="author"><name><surname>Whitty</surname><given-names>Paula M</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>p.m.whitty@ncl.ac.uk</email></contrib><contrib id="A7" contrib-type="author"><name><surname>Steen</surname><given-names>Nick</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>nick.steen@ncl.ac.uk</email></contrib><contrib id="A8" contrib-type="author"><name><surname>Grimshaw</surname><given-names>Jeremy G</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>jgrimshaw@ohri.ca</email></contrib> | Implementation Science | <sec><title>Background</title><p>Approximately 600,000 people in the UK have dementia. This is 5% of the population aged 65 and older, and 20% of those aged 80 and older [<xref ref-type="bibr" rid="B1">1</xref>]. By 2026 this figure is predicted to be 840,000. Dementia is associated with major social and economic costs, including those to families and carers. The UK National Service Framework (NSF) for Older People states that the improved care of people with dementia depends on early recognition and management [<xref ref-type="bibr" rid="B1">1</xref>]. Such care should involve a sensitive and accurate explanation of the diagnosis to individuals and carers, and information about the likely prognosis and possible packages of care.</p><p>Appropriate disclosure of a diagnosis to individuals with dementia is important for three reasons. First, from an ethical perspective, people with dementia, like other patients, have a right to know their diagnosis. At present, most carers are told the diagnosis [<xref ref-type="bibr" rid="B2">2</xref>], but this is not the case for people with dementia themselves [<xref ref-type="bibr" rid="B3">3</xref>]. Indeed, disclosure is less likely in dementia than in other terminal conditions, such as cancer. Earlier disclosure, supported by advocacy groups, allows the opportunity to plan family, financial, legal and long-term care arrangements. Second, many people with dementia want to know their diagnosis or receive more information about their illness [<xref ref-type="bibr" rid="B4">4</xref>-<xref ref-type="bibr" rid="B6">6</xref>]. Third, disclosure can facilitate decisions about treatment. Whilst this is increasingly important in the advent of therapies to slow disease progression, anecdotal evidence suggests that patients prescribed anti-dementia medication are not inevitably told their diagnosis. The UK Alzheimer's Society Consumer Network has identified issues around early diagnosis and care as research priorities.</p><p>It is a consistent finding that changing clinical practice is unpredictable and can be a slow and haphazard process. Over the last decade a considerable body of literature has been published suggesting that a range of interventions (e.g. reminder systems, interactive education) can be effective in changing health care professionals' behaviour [<xref ref-type="bibr" rid="B7">7</xref>]. However, studies have substantial heterogeneity of interventions used, targeted behaviours, and study settings that make generalising their findings to routine healthcare settings problematic. Moreover, there is no underlying generalisable taxonomy for either research or service settings by which to characterise individuals, settings and interventions. The assumption that clinical practice is a form of human behaviour and can be described in terms of general theories relating to human behaviour offers the basis for a taxonomy for Implementation Research. For example, the effectiveness of interventions may be influenced by factors such as health professionals' <italic>beliefs </italic>or <italic>perceived control over their practice </italic>– generalisable concepts that can be used across different contexts. Two steps are necessary to design a theory-based intervention for a behaviour change trial [<xref ref-type="bibr" rid="B8">8</xref>]. One is to identify modifiable factors underlying professional behaviour in order to identify those processes to target with an intervention (process modelling), and the second is to understand how interventions might work and be optimised (intervention modelling). These respectively correspond to the theoretical phase and the modelling and exploratory trial phases of the UK Medical Research Council (MRC) Framework for the development and evaluation of complex interventions [<xref ref-type="bibr" rid="B9">9</xref>,<xref ref-type="bibr" rid="B10">10</xref>].</p><sec><title>Work conducted to date</title><p>In the clinical area of dementia we have conducted a systematic review that indicates wide variation in the reported practice of disclosure of dementia among health professionals [<xref ref-type="bibr" rid="B3">3</xref>]. Four main factors appear to influence disclosure: (1) patient characteristics (e.g. age, ability to retain the diagnosis); (2) nature of the dementia (e.g. severity, diagnostic uncertainty, availability of disease-slowing therapies); (3) structural factors (e.g. time); and (4) clinician factors (e.g. perceived value of disclosure) [<xref ref-type="bibr" rid="B11">11</xref>-<xref ref-type="bibr" rid="B19">19</xref>]. We have conducted a detailed primary and secondary care case note review to look for symptoms that would allow the earlier diagnosis of dementia. We also have conducted focus groups and in-depth interviews with a range of health professionals, carers, and people with dementia [<xref ref-type="bibr" rid="B20">20</xref>]. These suggest potential ways of improving current practice and highlight the importance of the process of referral and testing in preparing people with dementia and their carers for a diagnosis.</p><p>In the area of Implementation Research we have conducted several pragmatic RCTs of behaviour change strategies and have conducted both process and intervention modelling studies. Although these methods are familiar to psychology, their use with healthcare professionals and their integration into implementation trials is novel. We will use these methods to develop an optimised, theory based intervention – targeting modifiable factors – to increase diagnostic disclosure of dementia. This work will lead to a randomised controlled trial of the intervention.</p></sec><sec><title>Aim</title><p>To develop a tailored theory-based intervention to promote appropriate disclosure of diagnosis of dementia.</p></sec><sec><title>Objectives</title><p>(1) To define and develop an appropriate model of disclosure for dementia; (2) To identify, within a theoretical framework, factors that influence disclosure of a diagnosis of dementia by members of old age mental health teams (MHTs); (3) To develop a theory-based intervention that promotes appropriate disclosure by targeting those factors identified in (2) that are amenable to change.</p></sec></sec><sec sec-type="methods"><title>Methods</title><sec><title>Objective 1: Defining an appropriate model of disclosure for dementia</title><p>Disclosure is ideally a process tailored to individuals' receptiveness and needs for information. A group of ten relevant stake-holders (i.e., psychiatrists, community psychiatric nurses, patient group representatives, general practitioners) will be convened and will use a structured consensus method [<xref ref-type="bibr" rid="B20">20</xref>] to define an appropriate model of disclosure, including consideration of both positive and negative effects. This will be informed by: available documents (e.g. NSF for Older People [<xref ref-type="bibr" rid="B1">1</xref>]), findings from our qualitative work [<xref ref-type="bibr" rid="B18">18</xref>], further interviews with people with dementia and two carer focus groups, and a request for examples of good practice in the Alzheimer's Society national newsletter.</p></sec><sec><title>Objective 2: To identify, within a theoretical framework, factors that influence old age mental health teams' disclosure of a diagnosis of dementia</title><sec><title>Design</title><p>Postal questionnaire survey.</p></sec><sec><title>Study sample</title><p>Disclosing a diagnosis of dementia is predominantly the responsibility of a consultant old age psychiatrist, although a range of other team members contribute to the process. We will identify involvement in disclosure within the postal questionnaire survey (below) of old age MHT members. This survey will also permit assessment of the feasibility of identifying and surveying MHT members. There are about 420 old age MHTs in the UK, and they will have differing structures and working patterns. For Objective 2, we will sample the 60 MHTs in the North of England and Yorkshire, and another 60 randomly selected from the rest of the UK. The former will be subsequently approached for recruitment to the RCT; the latter will provide data about the generalisability of the planned trial participants.</p></sec></sec><sec><title>Theory selection</title><p>The theories (Theory of Planned Behaviour (TPB), Social Cognitive Theory (SCT) and Implementation Intentions (II)) have been chosen for three reasons. First, they have all been rigorously evaluated in other settings. Second, they all explain behaviour in terms of factors amenable to change (e.g., beliefs, attitudes, and perceived external constraints). Third, they all include non-volitional components that assume individuals do not always have complete control over their actions. According to TPB, the strength of a behavioural intention is determined by attitudes towards the behaviour (in this case disclosure), subjective norms based on the perceived views of other individuals or groups (i.e. perceived social pressure); and perceived behavioural control, encompassing beliefs about self-efficacy (the ability to perform an action) and wider environmental factors that facilitate or inhibit performance [<xref ref-type="bibr" rid="B21">21</xref>]. SCT considers self-efficacy and individuals' goals in explaining behaviour [<xref ref-type="bibr" rid="B22">22</xref>]. Self-efficacy is highly predictive of a wide range of behaviours and can be enhanced by experience of success, observation of others' performance, or persuasive communications. II suggests that motivated individuals with a clear action plan are more likely to act. IIs are both predictive and a means of changing behaviour [<xref ref-type="bibr" rid="B23">23</xref>].</p></sec><sec><title>Questionnaire design</title><p>We will develop questions that explore constructs within the theories. Table <xref ref-type="table" rid="T1">1</xref> provides illustrative examples of variables, measures and items for each.</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>Illustrative examples of theories, variables, measures and items</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left"><bold>Theory</bold></td><td align="left"><bold>Predictor variable(s)</bold></td><td align="left"><bold>Measures</bold></td><td align="left"><bold>Illustrative items</bold></td></tr></thead><tbody><tr><td align="left">Theory of Planned Behaviour</td><td align="left">Attitude towards disclosure; subjective norms perceived behavioural control; intention</td><td align="left">Items developed from qualitative work [18] using standard question formats</td><td align="left">Attitudes – outcome beliefs: <italic>Being given a diagnosis of dementia will be beneficial to the patient</italic>.</td></tr><tr><td align="left">Social-Cognitive Theory</td><td align="left">Self-efficacy about disclosure; goals relevant to disclosure</td><td align="left">Items developed from qualitative work [18] using standard question formats</td><td align="left">Self-efficacy: <italic>I am confident in my ability to disclose a diagnosis of dementia sensitively</italic>.</td></tr><tr><td align="left">Implementation Intentions</td><td align="left">Action plans for disclosure</td><td align="left">Open questions with simple coding for presence of action plans</td><td align="left"><italic>Have you thought about increasing disclosure of dementia diagnosis? How will you go about this?</italic></td></tr></tbody></table></table-wrap><p>Questions for TPB and SCT will be assessed using multi-item scales, with items rated on seven-point Likert scales. The <italic>content </italic>of questions about control beliefs (those that influence the ability of professionals to disclose a diagnosis) relating to TPB, self-efficacy and goal beliefs relating to SCT will be drawn from existing qualitative data. Evidence of II will be ascertained using open questions about how the respondent normally discloses, which will be coded for evidence of "action plans" to disclose.</p><sec><title>Outcomes</title><p>We will use two measures of outcome: behavioural intention and behavioural simulation. We will measure behavioural intention using standard methods, i.e. rating scales of likelihood, frequency or agreement with statements or questions about intention. Six clinical scenarios that vary combinations of relevant items of patients and diagnosis will be used to measure behaviour simulation.</p></sec><sec><title>Administration</title><p>Following piloting, the questionnaire will be distributed with reminders at two and four weeks. Surveys of UK mental health professionals with an interest in older age in the last decade have achieved acceptable response rates of between 73% and 88% [<xref ref-type="bibr" rid="B14">14</xref>,<xref ref-type="bibr" rid="B19">19</xref>,<xref ref-type="bibr" rid="B24">24</xref>-<xref ref-type="bibr" rid="B26">26</xref>]. Based on previous experience with theory-based questionnaires such as this, we anticipate the response rates will be lower. Therefore, we will offer a financial incentive of GB£20 for each returned completed questionnaire. MHT members will be asked to complete questionnaires independently (i.e. not together in teams). To conform with the Data Protection Act, we intend to identify and seek participation in the postal survey of professionals as follows: we shall contact the appropriate NHS Trusts and seek demographic profiles of local old age MHTs; we shall then send Trusts the appropriate numbers of introductory letters (not addressed to named individuals) and participant information sheets for distribution to members of local old age MHTs; when members of old age MHTs receive the letters, they can decide whether to opt in to the survey and return their names and preferred contact details to us in a pre-supplied stamped addressed envelope; we shall then send out survey questionnaires to those who have opted in. Under this plan, we should be able to identify our sample denominator (to allow calculation of response rates and sample representativeness), and send reminders and financial incentives to those who have opted in (so as to enhance our response rate).</p></sec><sec><title>Sample size and analysis</title><p>The analysis will allow us to explain variation in both team and individual level behaviour. The surveys will generate at least ordinal level quantitative data. The relationships between predictor (i.e. theoretical constructs and clinical discipline) and outcome measures (behavioural intention and simulation) will be assessed primarily using multiple regression analysis and structural equation modelling – a procedure that utilises the observed covariance matrix. In both cases, the analysis will take into account the hierarchical structure of the data. Power calculations for multiple regression analysis depend on the number of cases per predictor variable. A minimum sample size of 50 + 8 m, where m is the number of predictor variables, is recommended for testing the multiple correlation, and 104 + m for testing individual predictors [<xref ref-type="bibr" rid="B27">27</xref>,<xref ref-type="bibr" rid="B28">28</xref>]. We have approximately 10 predictor variables, requiring a minimum sample size of 130 per survey to test the multiple correlation, or 114 to test individual predictors. We intend to approach approximately 420 individuals that make up approximately 120 mental health care teams. This allows for a (worst case) response rate of around 50% – and for the lack of independence of responses from individuals within a team.</p></sec></sec><sec><title>Objective 3: Development and pilot test of a theory-based intervention to promote disclosure of diagnosis by old age mental health teams (ISRCTN 15871014)</title><sec><title>Design and study sample</title><p>We will use the products of Objective 2 to identify the most promising elements of a potential intervention and then evaluate them using a randomised controlled design with MHTs. We will use members of MHTs that have previously responded in Objective 2, and augment the sample to reach the required size by randomly sampling within those MHTs not previously approached in Objective 2.</p></sec><sec><title>Identifying potential interventions</title><p>Objective 2 will identify the factors that are (a) modifiable and (b) the best predictors of behavioural intention. However, there will be a range of factors with these characteristics, and we will need to choose those factors which, in addition, can be modified by an intervention that is feasible. Therefore, we need to choose the two or three "best bets" (dependent on the results of objective 2) and simulate delivering them within a trial and examine their effects [<xref ref-type="bibr" rid="B29">29</xref>-<xref ref-type="bibr" rid="B31">31</xref>]. In these modelling experiments, elements of an intervention are manipulated within a randomised controlled design in a manner that simulates a real situation as much as possible. Interim endpoints (stated behavioural intention and simulation) are measured rather than changes in professional behaviour or healthcare outcome. This novel approach offers experimental control and the opportunity to vary elements of an intervention in order to better understand intervening variables and the effect on different outcomes. Behavioural intention has been incorporated into virtually all models of health behaviour as the single best predictor of subsequent health behaviour. In a review of 10 meta-analyses Sheeran demonstrated a consistent relationship between behavioural intention and subsequent behaviour, with intention explaining 28% of the variance in behaviour [<xref ref-type="bibr" rid="B32">32</xref>]. Members of MHTs will receive (or will have received) an initial survey as in Objective 2. Responders to this will be randomly allocated by team to receive one of up to three simulated interventions or a no intervention control. They will be instructed to open the intervention materials and then to complete a further theory-based survey as in Objective 2. Thus, we will identify the method that is most effective at changing professionals' beliefs and intentions. The sample size for a four-armed trial, powered to detect a difference between any two arms, is based upon the following assumptions: MHT as the unit of analysis, the outcome variable in the form of a score for the team, 80% power, and a type 1 error rate of 2.5% (rather than 5% to allow for multiple comparisons). Furthermore, we are aiming to detect a relatively large effect size of 0.8 on the basis that the modelling experiment eliminates some of the sources of variability associated with a 'definitive trial' (e.g. patient characteristics), and any smaller modelling effect size is unlikely to translate into a worthwhile effect in the definitive trial. Hence, we require four groups of 30 teams (120 in total). We will survey 240 teams (excluding those from North of England and Yorkshire) to allow for a 50% response rate in order to achieve the required sample size.</p></sec><sec><title>Pilot of outcome measurement</title><p>We will undertake a small scale postal survey of carers and people with dementia to help develop and pilot the feasibility of collecting outcome measures using this method in a definitive trial.</p></sec></sec><sec><title>Partnerships</title><p>We will explore the scope for maximal 'buy in' by a range of potential stakeholders. Representatives of the Alzheimer's Society have been consulted over the development of this proposal, and we aim to involve the Alzheimer's Society directly in the development of the intervention. The Faculty of Old Age Psychiatrists of the Royal College of Psychiatrists has agreed to appoint a member to liaise with us in regard to the design and conduct of the study.</p></sec><sec><title>Predicted outcomes and follow-on opportunities</title><p>We propose to evaluate the intervention developed within a cluster randomised controlled trial. We will seek the participation of MHTs in the North of England. This platform work represents the first stage in engaging MHTs and will provide necessary information about the feasibility and appropriateness of the trial. Although we propose to randomise MHTs, Objective 2 will help demonstrate whether targeting MHTs, as opposed to individual psychiatrists only, is appropriate and feasible. The trial outcomes will include the proportion of people with dementia (and carers) aware of their diagnosis, quality of information provided, use of medical treatments, and the economic consequences of the intervention.</p></sec><sec><title>Ethical and other implications</title><p>The study has MREC approval. The University of Newcastle operates a <italic>Good Practice in Research Code </italic>to ensure highest standards of integrity in research.</p></sec></sec><sec><title>Competing interests</title><p>The author(s) declare that they have no competing interests.</p></sec><sec><title>Authors' contributions</title><p>All authors contributed to the ideas and writing of this paper. They have all seen and approved the final draft.</p></sec> |
Activation of MAPK ERK in peripheral nerve after injury | <sec><title>Background</title><p>Activation of extracellular signal-regulated protein kinase (ERK), a member of mitogen-activated protein kinase (MAPK) family, has been proposed to mediate neurite outgrowth-promoting effects of several neurotrophic factors <italic>in vitro</italic>. However, the precise activity of ERK during axonal regeneration <italic>in vivo </italic>remains unclear. Peripheral axotomy has been shown to activate ERK in the cell bodies of primary afferent neurons and associated satellite cells. Nevertheless, whether ERK is also activated in the axons and surrounded Schwann cells which also play a key role in the regeneration process has not been clarified.</p></sec><sec><title>Results</title><p>Phosphorylation of ERK in the sciatic nerve in several time-points after crush injury has been examined. Higher phosphorylation of ERK was observed in the proximal and distal nerve stumps compared to the contralateral intact nerve from one day to one month after crush. The activation of ERK was mainly localized in the axons of the proximal segments. In the distal segments, however, active ERK was predominantly found in Schwann cells forming Bungner's bands.</p></sec><sec><title>Conclusion</title><p>The findings indicate that ERK is activated in both the proximal and distal nerve stumps following nerve injury. The role of activated ERK in Wallerian degeneration and subsequent regeneration <italic>in vivo </italic>remains to be elucidated.</p></sec> | <contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Agthong</surname><given-names>S</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>fmedsat@md.chula.ac.th</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Kaewsema</surname><given-names>A</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>atkaewsema@yahoo.com</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Tanomsridejchai</surname><given-names>N</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>jamjan_bio@yahoo.com</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Chentanez</surname><given-names>V</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>fmedvct@md.chula.ac.th</email></contrib> | BMC Neuroscience | <sec><title>Background</title><p>Peripheral axotomy can activate several signaling pathways in neurons and associated glial cells leading to two opposing consequences: cell death or adaptation to regenerate neurites. The principal signaling pathways that have been demonstrated to be involved in axonal regeneration are PI3K-Akt [<xref ref-type="bibr" rid="B1">1</xref>-<xref ref-type="bibr" rid="B3">3</xref>] and JAK/STAT [<xref ref-type="bibr" rid="B4">4</xref>,<xref ref-type="bibr" rid="B5">5</xref>].</p><p>Accumulated evidence has shown the participation of mitogen-activated protein kinases (MAPKs) during axonal injuries. MAPKs are a family of serine/threonine specific kinases that transduce extracellular stimuli to altered gene expression and have been shown to play a role in diverse cellular events ranging from proliferation, differentiation to apoptosis. So far, three subfamilies of MAPKs have been identified, namely: extracellular signal-regulated kinase (ERK), c-Jun NH<sub>2</sub>-terminal kinase (JNK) and p38 kinase (p38). JNK along with its main transcription factor, c-Jun, are activated in the dorsal root ganglia (DRG) after sciatic nerve transection [<xref ref-type="bibr" rid="B6">6</xref>-<xref ref-type="bibr" rid="B8">8</xref>]. Similarly, it has been reported that p38 was activated in axotomized neurons and, in case of spinal root ligation, this activation has been linked to the development of mechanical allodynia [<xref ref-type="bibr" rid="B9">9</xref>-<xref ref-type="bibr" rid="B11">11</xref>]. Moreover, inhibition of p38 leads to enhanced axonal regeneration, suggesting that p38 is likely involved in this process [<xref ref-type="bibr" rid="B12">12</xref>].</p><p>As for ERK, its role in neurite extension in response to growth factors is well-recognized. Following axotomy, ERK activation has been observed in transected sciatic nerve and ipsilateral DRG [<xref ref-type="bibr" rid="B4">4</xref>,<xref ref-type="bibr" rid="B13">13</xref>]. Evidence suggests that this increased activity of ERK is mediated by endogenous neurotrophic factors known to stimulate axonal regeneration. Nerve growth factor (NGF) which is known to be up-regulated shortly after nerve injury, requires ERK to promote neurite outgrowth <italic>in vitro </italic>[<xref ref-type="bibr" rid="B14">14</xref>,<xref ref-type="bibr" rid="B15">15</xref>]. Moreover, other regeneration-promoting molecules, such as glial-derived neurotrophic factor (GDNF) and FK506 also mediate their effects via the ERK pathway [<xref ref-type="bibr" rid="B15">15</xref>,<xref ref-type="bibr" rid="B16">16</xref>]. Therefore, it is possible that activation of ERK is essential for axons to regenerate in response either to endogenous growth factors or exogenous molecules. Although the ERK activation has been linked to mechanical allodynia in the models of neuropathic pain [<xref ref-type="bibr" rid="B18">18</xref>,<xref ref-type="bibr" rid="B25">25</xref>], the precise function of activated ERK after nerve injury <italic>in vivo</italic>, especially in term of axonal regeneration, is still unclear. In fact, spatial and temporal changes of ERK activity in the peripheral nerve in response to injuries have not yet been studied. Hence, the objective of this work was to investigate the time-dependent changes in the activation of ERK in crushed sciatic nerve with correlation to the cell types.</p></sec><sec><title>Results</title><sec><title>Activation of ERK in sciatic nerve after crush</title><p>Levels of phosphorylated ERK (ERK-P) relative to those of total ERK (ERK-T) were elevated in the proximal nerve segment compared to the intact nerve in all time-points with statistically significant changes observed at week1 after crush (Figure <xref ref-type="fig" rid="F1">1</xref>). In the distal stump, ERK phosphorylation was even higher relative to the proximal nerve, especially at post-crush 1 week and 1 month. No differences in the expression of ERK (ERK-T) were found among various nerve segments from any time point (quantitative data not shown).</p></sec><sec><title>ERK activation in axons and Schwann cells</title><p>No positive immunoreactivity was observed when primary antibodies were omitted in the presence of either secondary antibodies conjugated with FITC or rhodamine (Figure <xref ref-type="fig" rid="F2">2A</xref> and <xref ref-type="fig" rid="F2">2B</xref>). In the intact nerve, the ERK-P immunoreactivity was found in the axons identified by its co-localization with the pan-neurofilament immunoreactivity (Figure <xref ref-type="fig" rid="F2">2C–2E</xref>). In addition, ERK-P was also located in Schwann cell cytoplasm which appeared as crescent structures encapsulating the axons (Figure <xref ref-type="fig" rid="F2">2E</xref>). However, ERK-P was present mainly in the axons in the proximal nerve stump from all time points (Figure <xref ref-type="fig" rid="F2">2F–2H</xref>). In the distal nerve stump at day 1 after crush, ERK-P immunoreactivity was found in axons and with higher frequency in Schwann cells compared to the intact nerve (Figure <xref ref-type="fig" rid="F3">3A</xref> and <xref ref-type="fig" rid="F3">3B</xref>). In contrast, from week 1 until month 1, ERK-P was predominantly located between clumps of degenerated axons whose unremoved neurofilament proteins were stained red at week 1 (Figure <xref ref-type="fig" rid="F3">3C–3E</xref>) and between the regenerated axons at 2 weeks and 1 month post-crush (Figure <xref ref-type="fig" rid="F3">3F–3G</xref> and <xref ref-type="fig" rid="F3">3J–3K</xref>, respectively). This location of ERK-P was likely the Bungner's band, the tube forming by Schwann cells, since the co-locatization of ERK-P and S-100 (a Schwann cell marker) immunoreactivities was demonstrated (Figure <xref ref-type="fig" rid="F3">3H–3I</xref> and <xref ref-type="fig" rid="F3">3L–3M</xref>).</p></sec></sec><sec><title>Discussion</title><p>This study has demonstrated the activation of ERK in the proximal and distal nerve stumps from one day to one month after crush. Immunohistochemistry has shown that this activation occurred mainly in the axons of the proximal nerve, whereas the activation was more prominent in Schwann cells forming Bungner's bands in the distal nerve. Increased phosphorylation of ERK in the injured sciatic nerve has been previously reported by Sheu and co-workers [<xref ref-type="bibr" rid="B4">4</xref>]. They found that ERK-P levels were increased in the proximal nerve segment adjacent to the transection site as well as in the distal segment starting from one day until at least 16 days after operation. However, they have not demonstrated in which cell types this activation occurred. In accordance with that study, we found the prolonged activation of ERK from one day to one month after crush with slightly higher degree in the distal stumps compared to the proximal stumps. The expression of ERK as determined by the levels of ERK-T appears to be unaffected by the injury although the high variations can be observed. These high variations of ERK-T along with those of ERK-P may explain the insignificant increase in the phosphorylation ratio of ERK in the crushed nerve compared to the intact nerve at week 2. In addition, for the first time, this study showed that this ERK activation occurred mainly in the axons of proximal nerve. Early in the distal nerve at post-crush 1 day, the active ERK was observed both in the axons and Schwann cells. Furthermore, from week 2 to month 1, ERK-P was almost exclusively expressed in the Bungner's bands of Schwann cells.</p><p>The ERK activation in the axons of the proximal stump might be related to its activation in the neuronal cell bodies. The previous study has shown that ERK was activated in spinal ganglion neurons and satellite cells 7 and 14 days after sciatic nerve injury [<xref ref-type="bibr" rid="B13">13</xref>]. Interrupted anterograde axonal transport at the crush site resulting in the accumulation of ERK-P in the proximal nerve segment may account for the earlier increase in the expression of ERK-P in this part at post-crush day 1 than in the DRG (7 days). This possibility is supported by the previous report of ERK axonal transport in DRG neurons [<xref ref-type="bibr" rid="B23">23</xref>]. At later time points, 1, 2 and 4 weeks post-lesion, elevated ERK-P in the proximal stump may be due to the upregulated ERK phosphorylation in the cell bodies. Whether anterograde transport of ERK-P is involved in the up-regulation of ERK-P in the proximal stump remains to be clarified. It is also worth noting that the ERK-P immunoreactivity was observed in many Schwann cells in the intact nerve, whereas most positive signals in the proximal stump were localized to the axons. This may suggest the down-regulation of ERK-P in Schwann cells in the proximal segment after injury. Nevertheless, the significance of this change needs to be clarified.</p><p>The underlying mechanisms that stimulate the ERK pathway are unknown. It is unlikely that ERK was activated by the upregulated cytokines, such as, interleukin-1 beta (IL-1β), interleukin-6 (IL-6) since the patterns of upregulation of these cytokines were not correlated with that of ERK-P [<xref ref-type="bibr" rid="B4">4</xref>]. Some evidence shows that the upregulated growth factors after nerve injury may be responsible for this ERK activation in DRG. Obata and colleagues have administered NGF either intrathecally or intraneurally and found an increase in the ERK phosphorylation in L4/5 DRG [<xref ref-type="bibr" rid="B13">13</xref>,<xref ref-type="bibr" rid="B18">18</xref>]. Furthermore, GDNF, another growth factor whose expression is also upregulated after nerve injury can activate ERK in cultured DRG neurons [<xref ref-type="bibr" rid="B15">15</xref>]. It is also noteworthy that ERK-P was observed in satellite cells around neuronal cell bodies. The importance of this finding is not known but may emphasize the role of glial cells or the glial-neuronal interaction in the DRG after nerve injury.</p><p>Although the downstream events in the process of axonal regeneration stimulated by the active ERK remain unknown, the importance of ERK activation in nerve regeneration is increasingly evident. In the neuronal cell bodies, at least one study has demonstrated that ERK was required for an upregulated expression of brain-derived neurotrophic factor (BDNF) following axotomy [<xref ref-type="bibr" rid="B13">13</xref>] and BDNF can accelerate axonal regeneration [<xref ref-type="bibr" rid="B21">21</xref>,<xref ref-type="bibr" rid="B22">22</xref>]. Moreover, ERK is likely to mediate the regeneration-promoting effects of NGF <italic>in vitro </italic>[<xref ref-type="bibr" rid="B14">14</xref>,<xref ref-type="bibr" rid="B15">15</xref>] and <italic>in vivo </italic>[<xref ref-type="bibr" rid="B19">19</xref>,<xref ref-type="bibr" rid="B20">20</xref>]. Similarly, ERK was required for the neurite outgrowth stimulated by GDNF [<xref ref-type="bibr" rid="B15">15</xref>]. Therefore, it appears that the ERK pathway is essential for nerve regeneration.</p><p>In the distal nerve stump, Schwann cells proliferate to form Bungner's bands following Wallerian degeneration. This activity may need the ERK activation as one study has found that ERK was required for Schwann cell proliferation induced by ascorbate in the co-culture of DRG neurons and Schwann cells [<xref ref-type="bibr" rid="B24">24</xref>]. ERK has also been shown to be involved in Schwann cell proliferation triggered by leprosy bacilli [<xref ref-type="bibr" rid="B26">26</xref>]. Furthermore, ERK has been shown to play a role in maintaining Schwann cells in an immature state which can still proliferate by counteracting the effect of PI3K [<xref ref-type="bibr" rid="B3">3</xref>]. The main localization of ERK-P in Bungner's bands observed in this study is in accordance with these findings. The responsible mechanisms that stimulate the ERK pathway in these proliferating Schwann cells are not clarified. Nevertheless, Sheu and colleagues have reported that the sustained pattern of ERK-P up-regulation in the distal stump was correlated with three sequential peaks of increased expression of growth factors: NGF, GDNF and BDNF, in chronological order [<xref ref-type="bibr" rid="B4">4</xref>]. Whether there is a relationship between these growth factors and ERK phosphorylation during Schwann cell proliferation remains to be studied. Besides the possible role of ERK in Schwann cell proliferation, it has been proposed that ERK may be also involved in the elimination of supernumerary Schwann cells through up-regulated p75<sup>NTR </sup>in the advanced stage of regeneration [<xref ref-type="bibr" rid="B27">27</xref>,<xref ref-type="bibr" rid="B28">28</xref>]. Taken together, the above evidence may explain the sustained and more pronounced activation of ERK in the distal nerve.</p></sec><sec><title>Conclusion</title><p>The higher phosphorylation of ERK was found in the proximal and distal stumps of sciatic nerve from one day until one month after crush. In the proximal segments, active ERK was mainly localized to the axons, whereas it was exclusively expressed in Schwann cells forming the Bungner's bands in the distal segment. These findings indicate that ERK is also activated in the injured peripheral nerve in addition to in the DRG. To elucidate the precise role of ERK in peripheral nerve regeneration, strategies to inhibit the ERK pathway must be employed and various components participating in the regeneration should be carefully examined.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>Animal surgery and sacrifice</title><p>Forty male Wistar rats weighing 200–250 g were anesthetized using halothane and underwent unilateral sciatic nerve crush. The left sciatic nerve of each animal was exposed at the mid-thigh level and crush was induced by the use of a fine arterial clamp with firm pressure against the nerve for 30 s. Epineurial suture with Ethilon<sup>® </sup>6/0 was done to mark the crush site. Following the surgery, the wound was closed and sutured with Ethilon<sup>® </sup>4/0. The animals were allowed to recover and housed in an animal care unit until sacrifice. All experimental procedures were approved by the institutional ethics committee and were performed according to the guidelines of the National Research Council of Thailand.</p><p>Eight rats per each time-point were sacrificed 1 day, 1 week, 2 weeks and 1 month after nerve crush. The sciatic nerves were removed bilaterally and snap-frozen on dry ice. These tissues were later transferred to -70°C and kept there until use for Western blot analysis. Additional 2 rats underwent transcardial perfusion at each time-point with 200 ml of 0.9% NaCl followed by 500 ml of 4% paraformaldehyde (PFA). The sciatic nerves were removed and post-fixed in 4% PFA for 6 hours at 4°C. After fixation, the tissues were washed several times and kept in 20% sucrose in 0.1 M phosphate buffer pH 7.4 until the next process of immunohistochemistry.</p></sec><sec><title>Western blot analysis</title><p>The sciatic nerves on the crush side were divided at the crush location marked by the suture into the proximal and distal stumps, whereas the whole contralateral (intact) nerves were used. All nerve segments were homogenized in homogenization buffer [0.1 mmol/l PIPES pH 6.9, 5 mmol/l magnesium chloride, 5 mmol/l EGTA, 0.5% Triton X-100, 20% glycerol, 10 mmol/l sodium fluoride plus 1 mmol/l PMSF, 2 mmol/l sodium orthovanadate and protease inhibitor cocktail (1 μg/ml pepstatin A, 1 μg/ml leupeptin, 10 μg/ml benzoyl-L-arginine methyl ester, 10 μg/ml p-tosyl-L-arginine methyl ester, 10 μg/ml L-1-tosylamide-2-phenylethylchloromethyl ketone, 10 μg/ml trypsin inhibitor and 10 μg/ml aprotinin); all from Sigma]. Sample buffer (0.25 M Tris pH 6.8, 10% glycerol, 0.01% bromophenol blue, 10 mM dithiothreitol, 2% SDS and 2% β-mercaptoethanol; all from Sigma) was added to the samples before boiling for 5 min. These homogenized samples were stored at -20°C until use.</p><p>Concentration of protein in each sample was determined using Bramhall protein assay [<xref ref-type="bibr" rid="B17">17</xref>]. SDS-PAGE was performed on 10 μg protein in 10% acrylamide and proteins were transferred to nitrocellulose membrane (Hybond ECL, Amersham Biosciences) using a semi-dry electroblotter (Trans-Blot SD semi-dry transfer cell, Bio-Rad). Prevention of non-specific binding on the membrane was achieved by incubating with 10% dried skim milk in Tween buffer (0.05% Tween20, Sigma). The membranes were then incubated overnight at 4°C in primary antibodies [rabbit antibodies to total and phosphospecific ERK1 and 2 (1:500 and 1:5000, respectively, Santa Cruz Biotechnology)]. Different membranes were probed for total and phosphospecific ERK. In the following day, the membranes were washed and incubated in the secondary antibody conjugated with horse-radish peroxidase (HRP) (anti-rabbit-HRP 1:5000, Cell Signaling Technology) for 2 hours at room temperature. The membranes were washed and the immune complex was detected by enhanced chemiluminescence (LumiGLO, Cell Signaling Technology). Hyperfilms (Hyperfilm ECL, Amersham Biosciences) were exposed to the membranes and scanned with a flat-bed scanner. All scanned digital images were imported to microcomputer in tiff format and the densities of specific bands were analyzed with image analysis program (Image ProPlus 4.5). Results from different blots were combined using standards present in every blot in triplicate and are expressed as a ratio of phosphorylated to total protein. The data of 2 isoforms of ERK (ERK1 and ERK2 at 44 and 42 kDa, respectively) were combined and shown in a bar chart.</p></sec><sec><title>Statistical analysis</title><p>The data were imported to SPSS for Windows version 10 and checked for normal distribution and homogeneity of variances. If these assumptions were met, one-way analysis of variance (ANOVA) was employed to compare means from different time-points. However, if the measurements were extremely skewed from normal distribution and/or had a markedly significant difference in variances, a non-parametric test (Kruskal-Wallis test) was used instead. Where statistically significant differences were observed in ANOVA or Kruskal-Wallis test, pair-wise or post-hoc comparisons were achieved by using Tukey's HSD or Mann-Whitney U test, respectively. Statistically significant differences were considered when p < 0.05 unless otherwise stated.</p></sec><sec><title>Immunohistochemistry</title><p>Sciatic nerves kept in 20% sucrose in PBS were embedded in OCT medium and 9 μm-thick slices were cut by cryostat section. Proximal and distal segments of sciatic nerves were sectioned either transversely or longitudinally. The slides with sections were blocked in 10% normal serum (Sigma) and incubated in the primary antibody to phosphorylated ERK (1:200, Cell Signaling Technology) for 48 hours at 4°C. After washing, the slides were incubated in the secondary antibody conjugated with fluorescein isothiocyanate (FITC) (anti-rabbit-FITC 1:200, Santa Cruz Biotechnology) for 2 hours at room temperature. The slides were then mounted with anti-fading mounting medium (Vectashield, Vector Laboratories), cover-slipped and examined under fluorescence microscope. In some sections, double-staining with either anti-pan-neurofilament (1:50, Zymed) or anti-S-100 (1:100, Chemicon) antibodies overnight at 4°C followed by secondary antibody conjugated with rhodamine (1:200, Santa Cruz Biotechnology) was done to locate the axons or Schwann cells, respectively.</p></sec></sec><sec><title>Authors' contributions</title><p>SA carried out the Western blot and immunohistochemical experiments, analysis of the data and drafting the manuscript.</p><p>AK and NT carried out the animal experiment and collection of tissues.</p><p>VC participated in the design of the study, gave advise during the experiments and correcting the manuscript.</p></sec> |
Place preference induced by nucleus accumbens amphetamine is impaired by local blockade of Group II metabotropic glutamate receptors in rats | <sec><title>Background</title><p>The nucleus accumbens (NAc) plays a critical role in amphetamine-produced conditioned place preference (CPP). In previous studies, NAc basal and amphetamine-produced DA transmission was altered by Group II mGluR agents. We tested whether NAc amphetamine CPP depends on Group II mGluR transmission.</p></sec><sec><title>Results</title><p>NAc injections (0.5 μl/side) of the Group II mGluR antagonist (2 <italic>S</italic>)- <italic>a</italic>-ethylglutamic acid (EGLU: 0.01–0.8 μg but not 0.001 μg) impaired CPP. The drug did not block the acute locomotor effect of amphetamine.</p></sec><sec><title>Conclusion</title><p>Results suggest that Group II mGluRs may be necessary for the establishment of NAc amphetamine-produced CPP. These receptors may also mediate other forms of reward-related learning dependent on this structure.</p></sec> | <contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Gerdjikov</surname><given-names>Todor V</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>todor@psyc.queensu.ca</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Beninger</surname><given-names>Richard J</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I2">2</xref><email>beninger@post.queensu.ca</email></contrib> | BMC Neuroscience | <sec><title>Background</title><p>Metabotropic glutamate receptors (mGluRs) are implicated in learning and synaptic plasticity [<xref ref-type="bibr" rid="B1">1</xref>]. mGluRs have been classified into three groups: Group I, including mGluR1 and mGluR5; Group II, including mGluR2 and mGluR3; and Group III, including mGluR 4 and mGluR 6–8. Group I stimulates phospholipase C (PLC) and phosphoinositide hydrolysis, whereas Group II and Group III inhibit cyclic adenosine monophosphate (cAMP) formation most likely through a G<sub>i</sub>-type protein [<xref ref-type="bibr" rid="B2">2</xref>,<xref ref-type="bibr" rid="B3">3</xref>]. Both cAMP and PLC are widely implicated in synaptic plasticity [<xref ref-type="bibr" rid="B4">4</xref>]. Through their activity on G-proteins and other second messengers, mGluRs modulate ion channel conductances, transmission through ligand-gated channels, as well as the activation of immediate early genes. Therefore, mGluRs are well suited to provide a means through which glutamate can induce synaptic changes at the same synapses where it elicits fast responses. The role of Group I mGluRs in learning and plasticity has been characterized extensively. Group II mGluRs have received less attention [<xref ref-type="bibr" rid="B1">1</xref>].</p><p>There is evidence suggesting a role for Group II in synaptic plasticity in learning. Group II is involved in corticostriatal long term depression (LTD) in the nucleus accumbens (NAc) [<xref ref-type="bibr" rid="B5">5</xref>]. Behavioral work implicates Group II receptors in olfactory and fear learning [<xref ref-type="bibr" rid="B6">6</xref>,<xref ref-type="bibr" rid="B7">7</xref>] and in lever pressing for food [<xref ref-type="bibr" rid="B8">8</xref>,<xref ref-type="bibr" rid="B9">9</xref>]. The reported Group II mGluR modulation of reward-related learning is consistent with the role of these receptors in downregulating the cAMP/PKA cascade[<xref ref-type="bibr" rid="B10">10</xref>]. cAMP-dependent protein kinase (PKA) activation mediates the acquisition of learning [<xref ref-type="bibr" rid="B11">11</xref>] and of reward-related learning in particular [<xref ref-type="bibr" rid="B12">12</xref>].</p><p>Both reward-related learning and addiction to psychostimulants critically involve NAc dopamine (DA) and share many of the same intracellular signals [<xref ref-type="bibr" rid="B12">12</xref>-<xref ref-type="bibr" rid="B14">14</xref>]. Glutamate release is necessary for amphetamine- and cocaine-produced conditioned place preference (CPP) [<xref ref-type="bibr" rid="B15">15</xref>,<xref ref-type="bibr" rid="B16">16</xref>] and systemic mGluR antagonists impair cocaine self-administration in rats [<xref ref-type="bibr" rid="B17">17</xref>]. The role of Group II mGluRs in the acquisition of psychostimulant reward has not been addressed in pharmacological studies.</p><p>Group II mGluRs modulate DA transmission. Locally administered agonists reduce, whereas locally administered antagonists increase NAc DA levels [<xref ref-type="bibr" rid="B18">18</xref>,<xref ref-type="bibr" rid="B19">19</xref>]. Group II mGluR agonists also modulate amphetamine-produced DA release, enhancing it in drug-naïve baboons [<xref ref-type="bibr" rid="B20">20</xref>] and impairing it in amphetamine-sensitized rats [<xref ref-type="bibr" rid="B21">21</xref>]. In a recent study mGluR2 receptor knockout mice showed enhanced cocaine-produced CPP [<xref ref-type="bibr" rid="B22">22</xref>]. Results showing that Group II blockade enhanced basal DA release [<xref ref-type="bibr" rid="B19">19</xref>] suggest that mGluR2<sup>-</sup><sup>/</sup><sup>- </sup>mutants may exhibit behaviors related to psychostimulant sensitization [<xref ref-type="bibr" rid="B23">23</xref>], explaining the hyperlocomotion in a novel environment and enhanced cocaine CPP observed in these mice. The acute role of Group II mGluRs in the acquisition of NAc psychostimulant-produced CPP has not been investigated.</p><p>In the present studies, we used CPP [<xref ref-type="bibr" rid="B24">24</xref>] to test the hypothesis that NAc DA-mediated learning depends on Group II mGluRs. A Group II mGluR antagonist was administered directly into NAc and the acquisition of CPP based on NAc amphetamine was assessed. We found that CPP was antagonized by the Group II antagonist. Part of this research has been presented in abstract form [<xref ref-type="bibr" rid="B25">25</xref>].</p></sec><sec><title>Results</title><sec><title>Histology</title><p>A total of 97 rats completed testing. Three rats failed to complete the study due to illness or technical problems. There was no relationship between the type and dose of drug and illness observed in these animals. Cannula placements were assessed for the remaining rats. A total of 24 rats was excluded leaving 73 rats for subsequent analyses. Figure <xref ref-type="fig" rid="F1">1</xref> shows the location of cannula tips for all rats included in the analyses. Animals were classified as hits if the tips of both cannulae were located in the core or shell region of NAc.</p></sec><sec><title>Time spent on each side during pre-exposure</title><p>The interpretation of CPP results is not straightforward if animals have a natural avoidance of the to-be-drug-paired side. In such a case, an apparent increase in time spent on that side after conditioning may be the result of decreased avoidance of the drug-paired side or simply habituation [<xref ref-type="bibr" rid="B24">24</xref>]. To check for bias, we averaged time spent on the side that would be paired with drug across the 3 habituation days and compared it to time spent on the side that would be paired with vehicle for each group. Paired samples t-tests revealed non-significant differences for all groups (see Table <xref ref-type="table" rid="T1">1</xref>). Thus, rats did not avoid the to-be-drug-paired side during habituation and the CPP paradigm was unbiased.</p></sec><sec><title>Tunnel time</title><p>A change in the time spent in the drug-paired side from habituation to test cannot be interpreted unambiguously as a change in place preference if time spent in the tunnel also changes. Thus, additional analyses were performed to compare tunnel time before and after conditioning (see Table <xref ref-type="table" rid="T1">1</xref>). No significant differences were found.</p></sec><sec><title>Place conditioning</title><p>To induce CPP NAc amphetamine was paired with one side of the apparatus over four conditioning days. EGLU was injected into NAc approximately ten minutes before amphetamine. Vehicle alone was paired with the other side of the apparatus on alternate days. CPP was analyzed using a dose (0.0, .001, .01, .4, .8 μg, or .01 alone) × phase (habituation vs. test) mixed ANOVA with phase as the repeated factor and time spent on the drug-paired side as the dependent variable. The ANOVA yielded a main effect of phase [<italic>F</italic>(1, 67) = 5.362, <italic>p </italic>< .05] and a dose × phase interaction [<italic>F</italic>(5, 67) = 2.77, <italic>p </italic>< .05]. A one-way simple effects between-subjects ANOVA testing the effect of dose during the habituation session was not significant. However, a one-way simple effects ANOVA testing the effect of dose during the test session produced a main effect of dose [<italic>F</italic>(5, 67) = 2.99, <italic>p </italic>< .05]. Pairwise comparisons revealed that on the test day rats conditioned with amphetamine alone spent more time on the drug-paired side than rats conditioned with amphetamine plus 0.4 μg of EGLU. Rats conditioned with amphetamine plus the lowest EGLU dose (0.001 μg) spent more time on the drug-paired side compared to all other groups except the group conditioned with amphetamine alone (Fisher LSD <italic>posthoc </italic>tests; all <italic>p</italic>s < .05). This confirmed the hypothesis that EGLU impairs CPP acquisition. Rats receiving 0.01 μg of EGLU but no amphetamine showed no significant preference or avoidance (Figure <xref ref-type="fig" rid="F2">2</xref>).</p></sec><sec><title>Locomotor activity</title><p>Locomotor activity during conditioning sessions was analyzed for EGLU doses on the four amphetamine and four saline conditioning days yielding a dose (0.0, 0.001, 0.01, 0.4, 0.8 μg, 0.01 μg alone) × phase (amphetamine vs. saline) × day (4 days per phase) 3-way mixed ANOVA with phase and day as the within-subjects factors and total number of beam breaks per session as the dependent variable. The analysis revealed main effects of phase [<italic>F</italic>(1, 67) = 100.72, <italic>p </italic>< .001] and dose [<italic>F</italic>(5, 67) = 12.50, <italic>p </italic>< .001] and dose × phase [<italic>F</italic>(5, 67) = 5.47, <italic>p </italic>< .001] and dose × phase × day [<italic>F</italic>(15, 201) = 2.31, <italic>p </italic>< .01] interactions. In the drug phase, there was only a main effect of EGLU dose [<italic>F</italic>(5, 67) = 8.69, <italic>p </italic>< .001]. Animals receiving amphetamine or amphetamine plus EGLU (0.001, 0.01, 0.4 or 0.8 μg) showed higher locomotor activity than animals conditioned with 0.01 μg of EGLU alone (all <italic>p</italic>s < .05, Fisher LSD <italic>posthoc </italic>tests). In the saline phase, there was a main effect of EGLU dose [<italic>F</italic>(5, 67) = 5.35, <italic>p </italic>< .001]. The amphetamine group showed higher activity than the 0.01, 0.01 alone and 0.8 groups. The 0.001 μg group showed higher activity than the higher EGLU doses (0.01, 0.4, 0.8 and 0.01 alone; all <italic>p</italic>s < .05, Fisher LSD <italic>posthoc </italic>tests). Thus amphetamine increased locomotor activity both in groups that showed a CPP (0 and 0.001 doses) and in groups that did not (0.01, 0.4 and 0.8 doses). Although activity was significantly lower in all groups on vehicle days, the two groups that showed a place preference had higher locomotion in the drug-free phase (Figure <xref ref-type="fig" rid="F3">3</xref>).</p></sec></sec><sec><title>Discussion</title><p>In the present studies NAc amphetamine injections produced CPP. This effect was impaired by a Group II mGluR antagonist. The CPP paradigm was unbiased; rats did not show a systematic preference for one side of the apparatus over the other before conditioning. Also, the time rats spent in the tunnel before and after conditioning did not change significantly showing that increases in time spent on the drug-paired side were in fact due to changes in preference. If EGLU produced conditioned place avoidance on its own then simple additivity of this putative effect and the positive effect of amphetamine might explain the observed pattern of results. To test this, we injected one group of rats with EGLU alone on what would normally have been an amphetamine plus EGLU day using an EGLU dose that was found to affect amphetamine CPP (0.01 μg). On test day these rats did not show a preference or avoidance for the chamber normally paired with amphetamine. Thus putative place conditioning properties of EGLU do not appear to account for the observed results. Amphetamine stimulated locomotor activity. Activity among groups differed in drug-free vehicle sessions, with the groups that had received amphetamine alone or amphetamine plus 0.001 μg EGLU into NAc during conditioning being more active.</p><p>Groups that showed the CPP effect also showed higher activity in the vehicle phase. It is difficult to explain this observation. We know of no reports of a relationship between activity on <italic>vehicle </italic>conditioning days and the presence of a CPP effect. In our work we have repeatedly observed locomotor stimulation on <italic>drug </italic>conditioning days in groups that subsequently showed a CPP and groups that did not. Further studies are needed to assess the reliability of this finding and of activity differences among groups on vehicle conditioning days.</p><p>EGLU administration did not affect the acute increases in locomotor activity produced by amphetamine in the drug phase. EGLU has been reported to impair NAc amphetamine-produced locomotion using an amphetamine dose of 2.5 μg/0.5 μl/side [<xref ref-type="bibr" rid="B26">26</xref>]. We have previously reported that NAc amphetamine doses ranging from 5.0–20.0 μg produced similar levels of locomotor stimulation but a dose-dependent CPP effect (Beninger et al., 2003). For the present study we chose a dose of amphetamine (20 μg) that produced a significant CPP effect; this dose was eight times higher than the dose needed to produce motor stimulation in the study of Kim et al. (2000). The doses of EGLU (0.09 and 0.88 μg) that they used to reduce the locomotor stimulatory effect of amphetamine were comparable to the higher doses used here. Our failure to replicate their findings probably relates to our use of a substantially higher dose of amphetamine.</p><p>The NAc can be subdivided into core and shell subregions [<xref ref-type="bibr" rid="B27">27</xref>] and a number of studies suggest that these subregions may be differentially involved in reward-related learning and locomotion [<xref ref-type="bibr" rid="B28">28</xref>-<xref ref-type="bibr" rid="B32">32</xref>]. Our data did not permit a direct investigation of the relative contribution of the two subregions. An attempt was made to assess the possible role of the core and shell by separating the groups into core and shell placements and performing post hoc analyses. However only a few rats had both cannulae located within the core subregion. These rats were distributed unevenly among the groups making it impossible to statistically analyze the effect of region. It should be noted however that we have previously reported significant amphetamine CPP using the same vehicle volume and injection parameters with placements in both the core and shell [<xref ref-type="bibr" rid="B33">33</xref>]. In other work, lower injection volumes have been used to study the differential role of each subregion [<xref ref-type="bibr" rid="B34">34</xref>]. A more thorough investigation of the role of NAc subregions in CPP produced by local injections of amphetamine is awaited.</p><p>Dopaminergic innervation of the NAc, among other structures, has been widely implicated in natural appetitive behaviors and addiction [<xref ref-type="bibr" rid="B35">35</xref>]. NAc plays a critical role in the acquisition of amphetamine-produced CPP [<xref ref-type="bibr" rid="B36">36</xref>,<xref ref-type="bibr" rid="B37">37</xref>]. DA modulation of corticostriatal glutamatergic projections is believed to underlie reward-related learning [<xref ref-type="bibr" rid="B38">38</xref>-<xref ref-type="bibr" rid="B40">40</xref>]. mGluRs may be involved in this modulation. Group I blockade impaired corticostriatal synaptic plasticity [<xref ref-type="bibr" rid="B41">41</xref>]. Group I mGluRs have been implicated in the acquisition of drug reward [<xref ref-type="bibr" rid="B42">42</xref>,<xref ref-type="bibr" rid="B43">43</xref>] and in mediating DA agonist-produced changes in intracellular cascades widely implicated in memory and plasticity [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B44">44</xref>-<xref ref-type="bibr" rid="B46">46</xref>]. While Group II mGluRs have not been investigated as thoroughly in this context, several behavioral studies have demonstrated a role for Group II mGluRs in learning. Group II mGluR activation rescued a nonselective mGluR antagonist-produced impairment of the spontaneous improvement in lever-pressing for food in mice [<xref ref-type="bibr" rid="B9">9</xref>]. Group II activation also impaired lever-pressing for food in rats although this effect was schedule-dependent [<xref ref-type="bibr" rid="B8">8</xref>]. In a mouse study, systemic cocaine-produced CPP was enhanced in Group II mutants [<xref ref-type="bibr" rid="B22">22</xref>]. In contrast, we found that acute administrations of the Group II mGluR antagonist EGLU impairs the acquisition of NAc amphetamine CPP. Our results are not in disagreement with this study. The mutants showed a behavioral spectrum consistent with psychostimulant sensitization, including enhanced cocaine behavioral sensitization, hyperlocomotion in a novel environment and enhanced DA release after a systemic cocaine injection. This observation is supported by work showing that a locally administered mGluR II antagonist increased basal DA levels [<xref ref-type="bibr" rid="B19">19</xref>]. The observed phenotype may therefore reflect long term adaptations produced by changes in striatal DA rather than acute effects of Group II mGluR blockade.</p><p>Little is known about the mechanism though which Group II mGluRs modulate the acquisition of reward-related learning. Group II mGluR antagonist-produced impairments in the acquisition of amphetamine CPP may be mediated by effects on PKA. The acquisition of reward-related learning involves DA receptor-mediated PKA activation [<xref ref-type="bibr" rid="B12">12</xref>]. Both the inhibition and direct activation of PKA impair reward-related learning and learning in other tasks [<xref ref-type="bibr" rid="B6">6</xref>,<xref ref-type="bibr" rid="B33">33</xref>,<xref ref-type="bibr" rid="B47">47</xref>]. In an in vitro study, a group II antagonist enhanced the stimulation of cAMP formation by forskolin [<xref ref-type="bibr" rid="B48">48</xref>]. Both a Group II agonist and antagonist impaired learning in a step-down avoidance task [<xref ref-type="bibr" rid="B6">6</xref>]. The agonist-produced impairment was rescued by coinfusion of forskolin, suggesting that mGluR II may act in synchrony with other receptors including mGluR I and DA<sub>1</sub>-like receptors to produce an optimal level of PKA activation. In fact, work from our lab [<xref ref-type="bibr" rid="B33">33</xref>,<xref ref-type="bibr" rid="B49">49</xref>] has shown that activation of the cAMP/PKA cascade impairs NAc amphetamine-produced CPP – a finding that may parallel the effects of a mGluR antagonist on a tonically active Group II mGluR signal.</p><p>Group II mGluRs may modulate plasticity by altering NMDA-DA receptor interactions in NAc. NMDA receptors are necessary for DA-mediated reward-related learning [<xref ref-type="bibr" rid="B50">50</xref>]. Group II activation alters NMDA-produced intracellular signaling in striatal neurons [<xref ref-type="bibr" rid="B51">51</xref>]. Group II inhibition impairs NMDA-dependent hippocampal LTD [<xref ref-type="bibr" rid="B52">52</xref>] and corticostriatal NAc LTD [<xref ref-type="bibr" rid="B5">5</xref>]. NAc LTD may mediate NAc-dependent learning and psychostimulant-produced plasticity [<xref ref-type="bibr" rid="B53">53</xref>,<xref ref-type="bibr" rid="B54">54</xref>]. Reward-related learning is thought to be mediated by the selective strengthening of behaviorally relevant corticostriatal inputs to striatum [<xref ref-type="bibr" rid="B38">38</xref>-<xref ref-type="bibr" rid="B40">40</xref>]. Perhaps Group II mGluRs are involved in the selective depression of synapses not related to current input in effect adjusting the gain of behaviorally relevant glutamatergic efferents to the NAc [<xref ref-type="bibr" rid="B54">54</xref>]. Further work will have to test this intriguing possibility.</p></sec><sec><title>Conclusion</title><p>We investigated the role of Group II mGluRs in reward-related learning using the CPP paradigm. This report is the first to test the effects of a Group II mGluR antagonist applied locally in the brain on the acquisition of CPP produced by NAc injections of amphetamine. We showed that the antagonist impaired NAc amphetamine-produced CPP. The results are consistent with previous research showing a role for mGluRs in memory and plasticity and DA-mediated reward-related learning. Group II mGluR modulation of DA effects in the NAc may involve changes in PKA activation, changes in NMDA signaling or both.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>Animals and surgery</title><p>Male Wistar rats (Charles River, St. Constant, Quebec) weighing between 200–250 g on arrival were housed in pairs on a 12-hr reversed light-dark cycle (lights on at 19:00 hr) at an average temperature of 21°C, humidity 40–70%. Water and food (LabDiet 5001, PMI Nutrition Intl, Brentwood, MO) were freely available. Rats were handled for about 1 min on each of 5 consecutive days after arrival. The experimental protocol was approved by the Animal Care Committee at Queen's University. All animals were treated in full compliance with the Animals for Research Act and relevant guidelines set by the Canadian Council on Animal Care.</p><p>Approximately one wk after arrival at the colony, rats were anesthetized in an induction chamber using an inhalable anesthetic (5% isoflurane; Bimeda, Cambridge, ON) mixed with oxygen in a vaporizer system (Benson, Merkham, ON) and administered at 1.0 l/min. Anesthetized animals were fitted to a stereotaxic apparatus and isoflurane was administered at a concentration of 2% or as needed to maintain anesthesia. The head was adjusted so that lambda and bregma were on the same horizontal plane. For analgesia, buprenorphine hydrochloride in solution (0.15 mg/kg; Reckitt & Colman, Richmond, VA) was injected subcutaneously preoperatively. Ketoprofen (1.5 mg/kg; Merial, Baie d'Urfé, QC) was injected immediately after surgery and on three subsequent days postoperatively. The experimenter shaved the rat's head and applied betadine solution with a cotton tip applicator before incising the skin. Holes were drilled into the skull and 23 gauge (0.64 mm diameter) stainless-steel guide cannulae were chronically implanted bilaterally into the NAc, with coordinates: 1.6 mm anterior to bregma, 1.4 mm lateral to the midline and 6.7 mm ventral from the skull surface [<xref ref-type="bibr" rid="B55">55</xref>]. The guide cannulae were held in place by four stainless-steel screws and dental acrylic. Stainless steel wire stylets (0.31 mm diameter) flush with the end of the guide cannulae were put in place to prevent occlusion. Rats were allowed approximately one wk to recover before the start of behavioral testing. Behavioral data were analyzed if the tips of the cannulae were located in the core or shell region of NAc.</p></sec><sec><title>Drug infusion</title><p>The Group II mGluR antagonist (2 <italic>S</italic>)- <italic>a</italic>-ethylglutamic acid (EGLU; Tocris, Ellisville, MO) was dissolved in saline before the beginning of the experiment and stored at -20°C. The pharmacological properties of this agent have been well documented. EGLU impaired presynaptic mGluR agonist-produced motoneuron depolarization believed to be mediated by Group II mGluRs and inhibited mGluR agonist binding in mGluR2-transfected cell membranes [<xref ref-type="bibr" rid="B56">56</xref>].</p><p>Amphetamine sulfate (USP, Rockvill, MD) was dissolved in saline daily before each set of injections. Central injections into the NAc were made with a microinfusion pump (KD Scientific, Holliston, MA). Injectors were glued to polyethylene tubing (0.75 mm o.d.) filled with distilled water. The tubing was connected to two 10 μl microsyringes (Microliter #701; Hamilton, Reno, NV) mounted on the microinfusion pump. Drugs were backloaded into the injectors by aspiration with the two syringes. Rats were hand-held as the experimenter removed the stylets from the guide cannulae and inserted the two injectors (0.31 mm o.d.). The injectors projected 1.2 mm beyond the guide cannulae.</p><p>EGLU (0.001, 0.01, 0.4 or 0.8 μg/0.5 μl/side) was injected bilaterally over 30 sec. The highest EGLU dose was similar to a dose previously reported to impair NAc amphetamine-produced locomotion [<xref ref-type="bibr" rid="B26">26</xref>]. After the drug was delivered, the injectors were left in place for another 30 sec to facilitate diffusion, after which they were slowly retracted from the guide cannulae. Amphetamine (20.0 μg/0.5 μl/side) was injected approximately 10 min later using the same procedure. This amphetamine dose was much higher than the one reported by Kim et al. (2000) and has been found to produce reliable CPP in previous studies from our lab (e.g. [<xref ref-type="bibr" rid="B49">49</xref>,<xref ref-type="bibr" rid="B57">57</xref>]). Place conditioning began immediately after amphetamine injection. Groups were tested drug-free.</p></sec><sec><title>Apparatus</title><p>The 4 testing chambers each consisted of two rectangular compartments (38 × 27 × 34 cm) connected with a tunnel (8 × 8 × 8 cm). Two different spatial features were varied across the 4 testing chambers. The compartment walls were either urethane-sealed wood or alternating 1 cm-wide black and white vertical stripes and were covered with clear Plexiglas. The floor was either wire bars 1 cm apart running perpendicular to the tunnel, or a wire grid with openings of 1 cm<sup>2</sup>. This resulted in 4 possible compartment types, distributed as left or right compartment across 4 different testing chambers. Each compartment had a Plexiglas top with a number of circular ventilation holes. The tunnel was fitted with guillotine-type doors, which could be closed to prevent movement from one compartment to the other. Two infrared emitters and detectors (height 5.0 cm) in each compartment and 2 in the tunnel (height 5.0 cm) were used to monitor movement between and within compartments and to record time spent in each compartment and the tunnel. The number of beam breaks during conditioning sessions was used as a measure of locomotion. Each of the 4 testing chambers was housed in a dimly lit (7.5 Watt), sound-attenuated and ventilated wooden box. Indirect light reached the rat though the Plexiglas tops of the compartments. Data from the sensors were collected on a 6809 microcontroller using custom made software and transferred to a Macintosh computer for analyses. For further details of the apparatus see [<xref ref-type="bibr" rid="B58">58</xref>].</p></sec><sec><title>Behavioral procedure</title><p>Training and testing occurred during the day (7:00–19:00 hr). Rats were tested in groups of 4 using a different testing chamber for each rat. The experimental protocol consisted of 3 habituation sessions, 8 conditioning sessions, and 1 test session. Each rat completed one session per day for a total of 12 days.</p><sec><title>Habituation</title><p>At the start of each 15-min session the rat was placed in one compartment of the box – left compartment for half the rats and right compartment for the other half. Tunnel doors were open allowing the animals to move freely between the two compartments. Activity sensors recorded the amount of time spent on each side and in the tunnel.</p></sec><sec><title>Conditioning</title><p>Over 8 30-min sessions carried out on 8 consecutive days, mGluR antagonists and amphetamine were injected into the NAc on days 1, 3, 5 and 7, and vehicle on days 2, 4, 6 and 8. Injection of the mGluR antagonist was followed by an injection of amphetamine approximately 10 min later. Rats received only one injection on vehicle days. Immediately after amphetamine or vehicle injection, depending on conditioning day, the rat was placed in one of the two compartments with the tunnel doors closed preventing movement into the other compartment or the tunnel. Half the rats were confined to the left side on drug days and to the right side on vehicle days. The other half were confined to the right side on drug days and to the left side on vehicle days. In this way, any given compartment was paired with drug for some animals and with vehicle for others. Number of beam breaks was recorded for each rat to assess locomotion.</p><p>Six groups were tested. Five groups received amphetamine (20 μg/0.5 μl/side) immediately before placement into the conditioning compartment; these injections were preceded 10 min earlier by saline or EGLU doses of 0.001, 0.01, 0.4 or 0.8 μg/0.5 μl/side. One additional group received EGLU (0.01 μg) 10 min before and saline immediately before placement into the conditioning compartment. This group assessed possible effects of EGLU alone on side preference.</p></sec><sec><title>Testing</title><p>Testing occurred on the day immediately following conditioning. The session lasted 15 min and was identical to habituation sessions. The start side was the same for each rat on test and habituation days. Time spent on each side and in the tunnel was recorded.</p></sec></sec><sec><title>Data analysis</title><p>Paired samples <italic>t</italic>-tests were used to compare tunnel time before and after conditioning. To check for side bias, another set of <italic>t</italic>-tests compared time spent on the drug-paired side to time spent on the vehicle-paired side before conditioning. The amount of time spent on the drug-paired side was averaged across the 3 habituation days. CPP was analyzed using a mixed dose × phase (habituation vs. test) analysis of variance (ANOVA) followed by restricted ANOVA and pair-wise comparisons. In addition, the number of beam breaks for each rat on each of the 8 conditioning days was summed and used as an index of motor activity. Activity data were analyzed using a dose × session type × day mixed ANOVA.</p></sec><sec><title>Histology</title><p>After completion of the experiment, rats were placed in an airtight chamber and euthanized with CO<sub>2</sub>. Brains were removed and preserved in a 10% formalin solution for at least 72 hrs. Coronal sections 60 μm thick from throughout the cannulated region were obtained by slicing the brains on a cryostat at -20°C. The sections were mounted on gelatin-coated glass slides and stained with cresyl violet. Judgments about the cannulae placements in the NAc were made by an observer who was blind to the results for individual animals. Brains with cannula tips in the shell and/or core regions of the NAc were included in subsequent analyses.</p></sec></sec><sec><title>Authors' contributions</title><p>TVG carried out the behavioral studies, participated in the design of the studies and the statistical analysis of the data, and drafted the manuscript. RJB conceived the study, participated in its design and coordination, and in revising the manuscript. Both authors read and approved the final manuscript.</p></sec> |
Performances and microbial features of an aerobic packed-bed biofilm reactor developed to post-treat an olive mill effluent from an anaerobic GAC reactor | <sec><title>Background</title><p>Olive mill wastewater (OMW) is the aqueous effluent of olive oil producing processes. Given its high COD and content of phenols, it has to be decontaminated before being discharged. Anaerobic digestion is one of the most promising treatment process for such an effluent, as it combines high decontamination efficiency with methane production. The large scale anaerobic digestion of OMWs is normally conducted in dispersed-growth reactors, where however are generally achieved unsatisfactory COD removal and methane production yields. The possibility of intensifying the performance of the process using a packed bed biofilm reactor, as anaerobic treatment alternative, was demonstrated. Even in this case, however, a post-treatment step is required to further reduce the COD. In this work, a biological post-treatment, consisting of an aerobic biological "Manville" silica bead-packed bed aerobic reactor, was developed, tested for its ability to complete COD removal from the anaerobic digestion effluents, and characterized biologically through molecular tools.</p></sec><sec><title>Results</title><p>The aerobic post-treatment was assessed through a 2 month-continuous feeding with the digested effluent at 50.42 and 2.04 gl<sup>-1</sup>day<sup>-1 </sup>of COD and phenol loading rates, respectively. It was found to be a stable process, able to remove 24 and 39% of such organic loads, respectively, and to account for 1/4 of the overall decontamination efficiency displayed by the anaerobic-aerobic integrated system when fed with an amended OMW at 31.74 and 1.70 gl<sup>-1</sup>day<sup>-1 </sup>of COD and phenol loading rates, respectively. Analysis of 16S rRNA gene sequences of biomass samples from the aerobic reactor biofilm revealed that it was colonized by <italic>Rhodobacterales</italic>, <italic>Bacteroidales</italic>, <italic>Pseudomonadales</italic>, <italic>Enterobacteriales</italic>, <italic>Rhodocyclales </italic>and genera incertae sedis TM7. Some taxons occurring in the influent were not detected in the biofilm, whereas others, such as <italic>Paracoccus</italic>, <italic>Pseudomonas, Acinetobacter </italic>and <italic>Enterobacter</italic>, enriched significantly in the biofilter throughout the treatment.</p></sec><sec><title>Conclusion</title><p>The silica-bead packed bed biofilm reactor developed and characterized in this study was able to significantly decontaminate anaerobically digested OMWs. Therefore, the application of an integrated anaerobic-aerobic process resulted in an improved system for valorization and decontamination of OMWs.</p></sec> | <contrib id="A1" equal-contrib="yes" contrib-type="author"><name><surname>Bertin</surname><given-names>Lorenzo</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>lorenzo.bertin@unibo.it</email></contrib><contrib id="A2" equal-contrib="yes" contrib-type="author"><name><surname>Colao</surname><given-names>Maria Chiara</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>colao@unitus.it</email></contrib><contrib id="A3" equal-contrib="yes" contrib-type="author"><name><surname>Ruzzi</surname><given-names>Maurizio</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>ruzzi@unitus.it</email></contrib><contrib id="A4" equal-contrib="yes" contrib-type="author"><name><surname>Marchetti</surname><given-names>Leonardo</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>leonardo.marchetti@unibo.it</email></contrib><contrib id="A5" equal-contrib="yes" corresp="yes" contrib-type="author"><name><surname>Fava</surname><given-names>Fabio</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>fabio.fava@unibo.it</email></contrib> | Microbial Cell Factories | <sec><title>Background</title><p>Olive mill wastewater (OMW) is the effluent resulting from olive oil producing processes. Due to their high COD loading rates and content of toxic phenolic compounds, OMWs have to be decontaminated before being discharged [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B2">2</xref>]. Among the treatment methods currently available for this effluent, anaerobic digestion is generally considered the most promising because of its ability to combine a marked OMW decontamination potential with the ability to generate biogas rich of CH<sub>4 </sub>[<xref ref-type="bibr" rid="B3">3</xref>,<xref ref-type="bibr" rid="B4">4</xref>]. The large scale anaerobic digestion of OMWs is normally conducted in dispersed-growth reactors, where however the removal of the toxic phenolic fraction is often unsatisfactory [<xref ref-type="bibr" rid="B5">5</xref>,<xref ref-type="bibr" rid="B6">6</xref>]. The possibility of intensifying the dephenolisation potential of the process by performing it in anaerobic column reactor packed with granular activated carbon (GAC) was recently demonstrated [<xref ref-type="bibr" rid="B7">7</xref>,<xref ref-type="bibr" rid="B8">8</xref>]. GAC-digestor resulted to be a reproducible and stable OMW digesting process capable of a tolerance to high OMW organic loads and methanogenic performances significantly higher than those of the other bench-scale up-flow packed-bed biofilm OMW digestors described so far in the literature [<xref ref-type="bibr" rid="B3">3</xref>,<xref ref-type="bibr" rid="B9">9</xref>-<xref ref-type="bibr" rid="B11">11</xref>] and of those of the dispersed-growth digestors previously developed with the same microbial inoculum [<xref ref-type="bibr" rid="B7">7</xref>,<xref ref-type="bibr" rid="B12">12</xref>,<xref ref-type="bibr" rid="B13">13</xref>]. Molecular analysis of GAC-packed bed biofilm showed that it was heterogeneously composed by a large number of <italic>Proteobacteria</italic>, bacteria of the <italic>Flexibacter-Cytophaga-Bacteroides </italic>group, sulfate-reducing bacteria and low G+C gram-positive bacteria along with a limited number of highly abundant <italic>Archaea </italic>taxons mostly due to <italic>Methanobacterium formicicum </italic>[<xref ref-type="bibr" rid="B8">8</xref>].</p><p>Despite of its improved OMW decontamination performances, GAC-digestor generated effluents with a COD generally too high to allow their release in the environment or their use in agriculture [<xref ref-type="bibr" rid="B14">14</xref>]. Thus, the opportunity to implement the GAC-digestor decontamination potential by integrating it with an aerobic biological post-treatment was explored in this study. A "Manville" silica bead (SB)-packed bed aerobic reactor was developed, hydraulically connected to the GAC-digestor and employed to post-treat OMW digested effluents. The integrated anaerobic-aerobic process was operated in continuous mode and assessed through a 2 month experiment. In this paper the decontamination potential and main microbial features (e.g., structure and spatial distribution of reactor microbial community) of the aerobic post-treatment are reported and discussed. Despite the large number of studies already published on the biotreatment of OMWs [<xref ref-type="bibr" rid="B2">2</xref>-<xref ref-type="bibr" rid="B4">4</xref>,<xref ref-type="bibr" rid="B7">7</xref>,<xref ref-type="bibr" rid="B8">8</xref>,<xref ref-type="bibr" rid="B11">11</xref>-<xref ref-type="bibr" rid="B13">13</xref>,<xref ref-type="bibr" rid="B15">15</xref>], this is the first work in which the use of an integrated anaerobic-aerobic continuously operating biofilter system is proposed for such a purpose. Furthermore, this is also the first work in which an aerobic biofilter colonized by the native microflora of an anaerobically digested OMW is assessed under a technological and biological point of view.</p></sec><sec><title>Results</title><sec><title>Performances of the integrated anaerobic-aerobic biofilter system</title><p>The integrated anaerobic-aerobic system was fed with AOMW at high and constant organic loading rate for a 2 month period (Table <xref ref-type="table" rid="T1">1</xref>). Fig. <xref ref-type="fig" rid="F2">2</xref> shows the evolution of COD and total phenol concentration in the influent of the integrated system and in the influent and the effluent of the aerobic biofilter throughout the whole experiment. The pollutant removal attained with the anaerobic digestion (expressed as COD or phenol removal efficiency, and calculated by dividing the amount of pollutant removed by the amount of pollutant occurring in the reactor influent) was of about 45 and 60% of initial AOMW COD and phenolic compounds, respectively. These removal efficiency were improved by about 30% through the aerobic post-treatment, thus permitting an overall removal of AOMW COD and phenol loading rates applied to the sequential anaerobic-aerobic treatment of about 59 and 76%, respectively (Table <xref ref-type="table" rid="T1">1</xref>). No HPLC-detectable aromatic metabolites accumulated in the aerobic SB-biofilter throughout the whole experiment. A large array of volatile fatty acids (acetate, propionic acid, iso-butyric acid, butyric acid and valeric acid) occurred in the SB-reactor influent, where they were responsible for about 30% of the influent COD, but they were not detected in the reactor effluent. pH and redox potential of the SB-reactor influent were 5.2 ± 0.2 and -290 mV, respectively, but they increased markedly (up to 6.5 and 150 mV, respectively) as a result of the aerobic treatment.</p><fig position="float" id="F1"><label>Figure 1</label><caption><p><bold>Integrated anaerobic-aerobic biofilter system</bold>. Scheme of the anaerobic-aerobic packed bed reactor system developed in the study.</p></caption><graphic xlink:href="1475-2859-5-16-1"/></fig><fig position="float" id="F2"><label>Figure 2</label><caption><p><bold>COD and phenol removals</bold>. Profiles of COD (a) and total phenol concentration (b) in the influent (▲) of the integrated anaerobic-aerobic process, in the influent (■) and the effluent (□) of the aerobic SB biofilter continuously fed with AOMW throughout the 56 days of treatment. Data provided were obtained through double measurements.</p></caption><graphic xlink:href="1475-2859-5-16-2"/></fig><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>COD and phenol removal efficiency. COD and total phenol loading rates with which the reactors were fed along with removal efficiency (%) of COD and total phenol biodegradation occurred both in the sole reactors and in the whole integrated anaerobic-aerobic system in steady state conditions.</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td align="center">Experiment duration</td><td align="center">Influent COD</td><td align="center">Influent Total phenol concentration</td><td align="center">D</td><td align="center">COD Loading rate</td><td align="center">Total phenol Loading rate</td><td align="center">Effluent COD</td><td align="center">Effluent Total phenol concentration</td><td align="center">COD Removal efficiency (%)</td><td align="center">Total phenol Removal efficiency (%)</td></tr><tr><td></td><td align="center">(day)</td><td align="center">(gl<sup>-1</sup>)</td><td align="center">(gl<sup>-1</sup>)</td><td align="center">(day<sup>-1</sup>)</td><td align="center">(gl<sup>-1</sup>day<sup>-1</sup>)</td><td align="center">(gl<sup>-1</sup>day<sup>-1</sup>)</td><td align="center">(gl<sup>-1</sup>)</td><td align="center">(gl<sup>-1</sup>)</td><td></td><td></td></tr></thead><tbody><tr><td align="center">GAC-digestor</td><td align="center">55</td><td align="center">15.28 ± 1.45</td><td align="center">0.82 ± 0.13</td><td align="center">2.077</td><td align="center">31.74 ± 3.01</td><td align="center">1.70 ± 0.28</td><td align="center">8.35 ± 0.75</td><td align="center">0.33 ± 0.04</td><td align="center">45</td><td align="center">60</td></tr><tr><td align="center">SB-aerobic reactor</td><td align="center">55</td><td align="center">8.17 ± 0.77</td><td align="center">0.33 ± 0.04</td><td align="center">6.171</td><td align="center">50.42 ± 4.78</td><td align="center">2.04 ± 0.27</td><td align="center">6.24 ± 0.69</td><td align="center">0.20 ± 0.03</td><td align="center">24</td><td align="center">39</td></tr><tr><td colspan="11"><hr></hr></td></tr><tr><td align="center">Integrated System</td><td align="center">55</td><td align="center">15.28 ± 1.45</td><td align="center">0.82 ± 0.13</td><td></td><td></td><td></td><td align="center">6.24 ± 0.69</td><td align="center">0.20 ± 0.03</td><td align="center">59</td><td align="center">76</td></tr></tbody></table></table-wrap></sec><sec><title>Biological features of the SB-biofilter</title><p>The biological properties of the SB aerobic biofilter were investigated at the end of the 2 months of operation by determining its content of immobilized biomass and the structure and spatial distribution of its microbial community. The amount of immobilized biomass detected at 5, 20 and 38 cm of height (from the bottom) of the reactor packed-bed was (in mg of dried biomass/g of dried support) 12.74 ± 0.85, 11.38 ± 1.01 and 27.51 ± 1.58, respectively. Running the average of such values (17.21 mg/g) and considering that the reactor was packed with 0.213 kg of dried support, it can be estimated that the SB-reactor harbored a total immobilized biomass of 3.67 g (on dry weight basis).</p><p>The structure of the microbial community occurring at different regions of the reactor packed-bed as well as in the influent and effluent of the reactor, was investigated through T-RFLP analysis. Fluorescent amplifications of 16S rRNA genes with universal eubacterial primers were successful for all of the samples examined. T-RFLP analysis of biofilm consortia obtained at 5, 20 and 38 cm height of the reactor packed bed yielded the same three major T-RFs of 81, 280 and 392 bp length with <italic>Rsa</italic>I restriction enzyme digestion. The three biofilm samples displayed similar T-RFLP patterns with a predominant T-RF with 81 bp of length (Fig. <xref ref-type="fig" rid="F3">3</xref>). Also the abundance of each bacterial population (represented as the peak height of each T-RF in relation to the total peak height of all T-RFs detected) was almost identical in the three biofilm samples. The major T-RFs detected in the influent after <italic>Rsa</italic>I digestion were 98, 280 and 386 bp in length (Table <xref ref-type="table" rid="T2">2</xref>). These peaks, which were detected and characterized in the effluent of the GAC reactor [<xref ref-type="bibr" rid="B8">8</xref>], were affiliated with <italic>Synergistes </italic>(clone B12, T-RF of 98 bp), <italic>Bacteroides </italic>(clone B25, T-RF of 280 bp) and γ-<italic>Proteobacteria </italic>(clone B1, T-RF of 386 bp), respectively. Only T-RF of 386 bp in length was detected in the AOMW introduced in the integrated biofilters. A microbial community displaying T-RFLP patterns very similar to those obtained from biofilm was found to occur in the SB-reactor effluents (data not shown).</p><fig position="float" id="F3"><label>Figure 3</label><caption><p><bold>Biofilm electropherograms</bold>. Electropherogram of the 5' T-RFs derived from <italic>Rsa</italic>I digestion of the <italic>Bacterial </italic>community 16S rDNA of the biofilm samples taken at different region of the reactor (a – 5 cm, b – 20 cm, c – 38 cm height).</p></caption><graphic xlink:href="1475-2859-5-16-3"/></fig><table-wrap position="float" id="T2"><label>Table 2</label><caption><p>Results of T-RF analyses. Schematic representation of T-RFs obtained after RsaI digestion of 16S rRNA genes amplified from DNA of samples of OMW, anaerobically-treated OMW (GAC-reactor effluents) and biofilm from the SB-aerobic biofilter.</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td align="center" colspan="15"><bold>T-RF length (bp)</bold></td></tr><tr><td align="left"><bold>Sample</bold></td><td align="center">68</td><td align="center">81</td><td align="center">98</td><td align="center">168</td><td align="center">183</td><td align="center">274</td><td align="center">280</td><td align="center">282</td><td align="center">386</td><td align="center">392</td><td align="center">404</td><td align="center">414</td><td align="center">428</td><td align="center">436</td><td align="center">441</td></tr></thead><tbody><tr><td align="center">OMW</td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td align="center"><bold>94</bold></td><td></td><td></td><td align="center">4</td><td align="center">2</td><td></td><td></td></tr><tr><td align="center">SB-reactor influent (GAC-digestor effluent)</td><td></td><td></td><td align="center"><bold>80</bold></td><td align="center">2</td><td></td><td align="center">1</td><td align="center">7</td><td align="center">3</td><td align="center">4</td><td></td><td align="center">1</td><td></td><td></td><td align="center">1</td><td align="center">1</td></tr><tr><td align="center">SB-reactor biofilm</td><td align="center">2.6</td><td align="center"><bold>37.7</bold></td><td></td><td></td><td align="center">1.7</td><td></td><td align="center">26.4</td><td></td><td></td><td align="center">13.5</td><td></td><td></td><td align="center">3.6</td><td align="center">7.3</td><td align="center">4.5</td></tr></tbody></table><table-wrap-foot><p>The dominant (height) T-RF peak in each profile is indicated in boldface. Individual clones having a corresponding peak in the T-RF profiles are indicated below. The numbers indicate the relative abundance of individual T-RF. These values were calculated based on the peak height of individual T-RF in relation to the total peak height of all T-RFs detected in the respective community fingerprint pattern. The peak heights were automatically quantified by GeneScan software (PE Applied Biosystems), performing the analysis with a peak height threshold of 50 fluorescent units.</p></table-wrap-foot></table-wrap><p>A total of 34 clones from 16S rRNA gene clonal libraries of biofilm consortia were randomly picked for PCR amplification and the amplified products digested with <italic>Eco</italic>RI, <italic>Eco</italic>RI plus <italic>Kpn</italic>I and <italic>Rsa</italic>I restriction enzymes. Eight different types of RFs (C81, C183, C280, C392, C610, H172, PS1 and PS156) were obtained (data not shown). Sequencing analysis of these RFs showed that clones from biofilm samples taken at different regions of the reactor having the same RF pattern were identical. The corresponding rDNA inserts were considered to belong to the same sequence type, resulting in the identification of 8 operational taxonomic units (OTUs). The sequences retrieved from the biofilm libraries were compared with the 16S rRNA reference sequences of the Ribosomal Database Project II database and were found to be > 80% identical to known rDNA sequences. Clone C280 was identical to clone B25 (<italic>Bacteroides </italic>group; Accession number AJ608923), which was detected in the effluent of the GAC reactor [<xref ref-type="bibr" rid="B8">8</xref>]. Clone C81 was closely related to <italic>Paracoccus versutus </italic>type strains ATCC 25364 (Accession number Y16962) and DSM 582 (Accession number Y16931) with 97.2% similarity (Fig. <xref ref-type="fig" rid="F4">4</xref> panel A) and to <italic>Paracoccus </italic>spp. strains isolated from a solid-phase denitrification process using poly(ε-caprolactone) as carbon and energy source [<xref ref-type="bibr" rid="B16">16</xref>] with 97.6% similarity. The remaining six clonal sequences loosely related to <italic>Pseudomonadales </italic>(clone H172 with 80.7% similarity to <italic>Acinetobacter schindleri </italic>type strain LUH5832; clone PS1 with 80.7% similarity to <italic>Pseudomonas nitroreducens </italic>type strain LMG 1224; clone PS156 with 96.4% similarity to <italic>Pseudomonas alcaliphila </italic>type strain AL15–21), <italic>Enterobacteriales </italic>(clone C392 with 75% similarity to <italic>Enterobacter cloacae</italic>), <italic>Rhodocyclales </italic>(clone C610 with 94.0% similarity to <italic>Dechlorosoma </italic>spp. strain PCC) and genera incertae sedis TM7 (clone C183 with 89.5% similarity to uncultured bacterium TM7 LH21, Fig. <xref ref-type="fig" rid="F4">4</xref> panel B). Analysis of clone distribution indicated that <italic>Paracoccus </italic>and <italic>Bacteroides </italic>were the major groups in the biofilm consortia obtained from the different portions of the reactor packed-bed (70%, Fig. <xref ref-type="fig" rid="F5">5</xref>).</p><fig position="float" id="F4"><label>Figure 4</label><caption><p><bold>Phylogenetic trees of 16S rRNA genes</bold>. Phylogenetic trees of 16S rRNA genes belonging to <italic>Paracoccus </italic>(T-RF 81 bp; panel a) and genera incertae sedis TM7 (T-RF 183 bp; panel b) analyzed using bootstrap and the neighbor-joining methods as distance measures. The cloned sequences are indicated in boldface and the GeneBank accession numbers of sequences are in bracket. The distance bar is shown under the tree, and bootstrap values (1000 replicates) are given for selected nodes.</p></caption><graphic xlink:href="1475-2859-5-16-4"/></fig><fig position="float" id="F5"><label>Figure 5</label><caption><p><bold>Clone distribution</bold>. Clone distribution based on the sequence of 16S rRNA genes recovered from clone libraries.</p></caption><graphic xlink:href="1475-2859-5-16-5"/></fig><p>T-RFLP analysis of each RF showed that clone C81 could be matched with the 81 bp length fragment after <italic>Rsa</italic>I digestion and clone C280 could be matched with the 280 bp length <italic>Rsa</italic>I-fragment. In addition, clone C81 generated a small T-RF with the size of 385 bp after <italic>Rsa</italic>I digestion due to incomplete digestion. Clone C183 generated a 183 bp length fragment which was a low abundant T-RF in the T-RFLP patterns of biofilm consortia. Clone C392 generated T-RF with the size of 392 bp, while clone H172 did not generate a <italic>Rsa</italic>I T-RF.</p></sec></sec><sec><title>Discussion</title><p>An intensified OMW anaerobic digesting process consisting of a GAC-packed bed biofilm reactor was recently developed and assessed [<xref ref-type="bibr" rid="B8">8</xref>]. In order to further reduce COD of the effluents resulting from the process, the possibility of implementing the GAC-digestor with a tailored aerobic post-treatment was explored in this study. Integrated anaerobic-aerobic treatments have been often proposed for a high rate, cost effective and sustainable disposal of agroindustrial effluents [<xref ref-type="bibr" rid="B17">17</xref>-<xref ref-type="bibr" rid="B22">22</xref>]. However, a little is known [<xref ref-type="bibr" rid="B13">13</xref>,<xref ref-type="bibr" rid="B23">23</xref>] about the feasibility of this approach in the management of OMWs. In the present study, a SB-packed bed biofilter was selected to post-treat digested OMW as the same technology was recently found to be very effective in the final decontamination of a similar anaerobically digested wastewater [<xref ref-type="bibr" rid="B23">23</xref>]. However, differently from this previous work, where it was inoculated with a defined co-culture of specialised bacteria and applied in discontinuous mode [<xref ref-type="bibr" rid="B23">23</xref>], SB-biofilm reactor was here hydraulically connected to the OMW digestor and allowed to be colonized by the influent native microflora able to grow on the SB surface under the aerobic conditions provided. To the very best of our knowledge, this is the first integrated anaerobic-aerobic biofilm system developed so far for an improved, continuous biomethanization and decontamination of OMWs.</p><p>The integrated biofilter system was assessed through a 2 month experiment performed under open mode of operation by feeding it with a high and constant AOMW organic loading rates. It was found to remove about 59 and 76% of COD and phenol loading rates, respectively. The aerobic post-treatment contributed for about 1/4 of such removal efficiency, and these performances were comparable to those displayed on a similar anaerobically digested effluents by the aerobic conventional activated sludge post-treatment developed by Beccari et al. [<xref ref-type="bibr" rid="B13">13</xref>] and lower than those observed by Bertin et al. [<xref ref-type="bibr" rid="B8">8</xref>] with a similar aerobic SB-biofilm reactor. However, it has to be pointed out that the latter two processes were operated at lower COD and phenol loading rates and under batch mode.</p><p>SB-reactor biofilm was found to consist of <italic>Rhodobacterales</italic>, <italic>Bacteroidales</italic>, <italic>Pseudomonadales</italic>, <italic>Enterobacteriales</italic>, <italic>Rhodocyclales </italic>and genera incertae sedis TM7. T-RFLP analysis and 16S rRNA gene cloning indicated that <italic>Paracoccus </italic>and uncultured strain B25 were the major groups occurring at different depths of the biofilter. Strain B25, an anaerobic bacterium of the <italic>Bacteroides </italic>group, was one of the major strains colonizing the GAC anaerobic digestor [<xref ref-type="bibr" rid="B8">8</xref>], whose effluents were used to feed the SB-aerobic biofilter. Instead, other members of the same group occurring in the anaerobic GAC biofilm digestor (i.e., uncultured bacteria B23 and B24) [<xref ref-type="bibr" rid="B8">8</xref>], were not found in the SB-biofilter. The presence of anaerobic <italic>Bacteroides </italic>in aerobic reactors has been already reported in the literature. The occurrence of these bacteria has been documented in conventional aerobic reactors treating municipal wastewaters [<xref ref-type="bibr" rid="B24">24</xref>] or anaerobically pre-treated tert-butyl alcohol-contaminated wastewaters [<xref ref-type="bibr" rid="B25">25</xref>], as well as in packed-bed biofilm reactors developed for the oxidization of sulfide-containing effluents [<xref ref-type="bibr" rid="B26">26</xref>]. Members of the <italic>Bacteroides </italic>group are often able to degrade various refractory biomacromolecules, such as cellulose, chitin, DNA, lipids, and proteins, which generally are abundant in a biofilm, in which dead microorganisms are trapped. These species are also known for their ability to produce exopolysaccharide slime, that generally has a primarily role in biofilm formation and development [<xref ref-type="bibr" rid="B27">27</xref>]. Thus, as suggested by Ferrera et al. [<xref ref-type="bibr" rid="B26">26</xref>], the presence of <italic>Bacteroides </italic>strains in aerobic biofilm reactors might be correlated to their role in the biofilm development and in the long-term functioning and versatility of the process. The abundant occurrence of the <italic>Paracoccus </italic>sp. T-RF clone C81 in the biofilm is also interesting because <italic>Paracoccus </italic>is a quite biochemical versatile genus, able to display a wide range of degradative capabilities. Some <italic>Paracoccus </italic>strains are capable of aerobic denitrification (simultaneous reduction of oxygen and nitrate) and heterotrophic nitrification (oxidation of ammonium to nitrite during heterotrophic growth), whereas other strains are capable of a) aerobic growth on formate, b) aerobic chemolithoautotrophic growth using carbon disulfide as energy sources, c) methylotrophic growth on methanol or d) heterotrophic growth on diethyl sulfide, thioethanol, thioacetic acid or substituted thiophenes. Some other strains of this group can also grow anaerobically using thiosulfate, carbon disulfide, methanol or formate as energy sources and nitrate as final electron acceptor [<xref ref-type="bibr" rid="B28">28</xref>]. The classification of the genus <italic>Paracoccus </italic>(alpha subgroup of the <italic>Proteobacteria</italic>) has undergone serious changes during the past decade [<xref ref-type="bibr" rid="B29">29</xref>]. Several new species have been isolated and, currently, the genus consists of 17 species, which can be found in different environments, including soil [<xref ref-type="bibr" rid="B30">30</xref>], contaminated groundwater [<xref ref-type="bibr" rid="B31">31</xref>], biofilters [<xref ref-type="bibr" rid="B32">32</xref>], sewage sludge [<xref ref-type="bibr" rid="B33">33</xref>], denitryfing reactors [<xref ref-type="bibr" rid="B34">34</xref>] and industrial wastewaters [<xref ref-type="bibr" rid="B35">35</xref>]. Bacteria belonging to the genus <italic>Paracoccus </italic>are important components of many wastewater treatment system communities [<xref ref-type="bibr" rid="B36">36</xref>]. Most species in the genus can use nitrate and its reduction products as an alternative electron acceptor to oxygen during anaerobic respiratory growth [<xref ref-type="bibr" rid="B37">37</xref>,<xref ref-type="bibr" rid="B38">38</xref>], and, therefore, can survive and proliferate in ecosystems with fluctuating aerobic/anaerobic conditions. In our study, 16S rDNA sequences belonging to <italic>Paracoccus </italic>populations (T-RF 81 bp) were not detectable in the AOMW that entered the integrated process and in the SB-reactor influent. (Table <xref ref-type="table" rid="T2">2</xref>). These bacteria were therefore massively enriched in the aerobic biofilm throughout the treatment to become the dominant eubacterial group. This finding might be ascribed to the ability of some <italic>Paracoccus </italic>species to denitrify in the presence of oxygen up to levels of 90% of air saturation [<xref ref-type="bibr" rid="B39">39</xref>,<xref ref-type="bibr" rid="B40">40</xref>], ability that might allow the same strains to use both oxygen and nitrate as terminal electron acceptors [<xref ref-type="bibr" rid="B39">39</xref>] and therefore to have nutritional advantages that, in turn, might have allowed them to extensively colonize the reactor system developed in this study exposed to varying oxygen concentrations. The clonal analysis of the 16S rRNA genes suggests that the microbial consortia that inhabit the aerobic biofilter also included minor members belonging to the uncultivated bacterial division TM7 (Fig. <xref ref-type="fig" rid="F4">4</xref>, panel B). Candidate division TM7 has no cultivated representatives and has been exclusively characterized by environmental sequence data. This division takes its name from the German peat bog from which the first sequence was obtained [<xref ref-type="bibr" rid="B41">41</xref>], but additional TM7 sequences deposited by several other investigators have demonstrated that members of this division are present in extremely diverse environments, including soil, freshwater, seawater, hot springs, mouse feces, termite guts, activated sludges and in human subgingival plaque samples [<xref ref-type="bibr" rid="B42">42</xref>-<xref ref-type="bibr" rid="B44">44</xref>]. In situ analysis revealed that members of the uncultivated TM7 division are capable of surviving and growing under a wide range of conditions. Ouverney et al. [<xref ref-type="bibr" rid="B45">45</xref>] suggested that TM7 members may be involved in the formation of a scaffold or biofilm, which could support the development of a disease-associated microbial community in human plaque. Thus, the evidence that TM7-like 16S rDNA could be amplified from the samples taken at different regions of the reactor let to speculate that TM7-related bacteria have had a role in the SB-biofilm formation. Unfortunately, the current limited knowledge on this uncultivated bacterial division does not allow us to speculate on the contribution that TM7 bacteria might have had on the aerobic decontamination of the anaerobic digested AOMW.</p><p>Other bacteria occurring in the biofilm consortia consisted of phylotypes affiliated with <italic>Pseudomonas, Acinetobacter </italic>and <italic>Enterobacter </italic>genera, which are common members of aerobic microbial consortia involved in the biodegradation of biogenic and xenobiotic compounds [<xref ref-type="bibr" rid="B46">46</xref>-<xref ref-type="bibr" rid="B48">48</xref>]. These members of <italic>Pseudomonadales </italic>and <italic>Enterobacteriales </italic>were not detected in the effluents of the anaerobic digestor and therefore enriched in the reactor throughout the 2 months-experiment. This suggests that they were significantly involved in removing organic compounds occurring in the anaerobically digested influent of the SB-aerobic reactor.</p></sec><sec><title>Conclusion</title><p>In conclusion, an effective aerobic biofilm technology able to significantly decontaminate anaerobically digested OMWs was developed, integrated with the anaerobic digestor and assessed in this study. To the very best of our knowledge, this is the first report in which a similar biofilm process is proposed for such a purpose and assessed through an integrated chemical and molecular biotechnology monitoring.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>Chemicals and OMWs employed</title><p>Chemicals employed in the analysis of COD, total phenols as well as solvents used for HPLC and ion chromatography were obtained from Sigma-Aldrich (Milan, Italy) and Baker Italia (Milan, Italy). "Manville" silica spherical beads (diameter: 5 mm) (SB) were supplied by Manville Filtration and Minerals (Denver, CO, USA).</p><p>An industrial OMW containing about 30 g/l of COD and 2.0 g/l of total phenols was collected from an Italian olive oil producing plant, stored in filled and sealed plastic jars at 4°C and employed to prepare an amended OMW, AOMW, that was then used in the study. AOMW was prepared from the industrial OMW by a) diluting the latter with an equal volume of tap water, b) amending the obtained wastewater with Ca(OH)<sub>2 </sub>(up to have its pH equal to 6.5), urea (0.45 g/l) and then 1 N NaOH (to adjust its pH to 7.8 ± 0.2). AOMW was placed in 4 l glass jars, where it was vigorously mixed (through a magnetic stirrer) and purged with 0.22 μm filter (Millipore, MO, USA)-sterilized O<sub>2</sub>-free N<sub>2 </sub>at room temperature for 3 h, before being employed in the experiments. AOMW COD and total phenol concentration were about 15 and 0.8 g/l, respectively.</p></sec><sec><title>Bioreactors, their inoculation, working conditions and sampling</title><p>The bioreactor system employed in the study was composed by a GAC-packed bed anaerobic biofilm reactor hydraulically connected to an aerobic SB-packed bed biofilm reactor. AOMWs were continuously introduced in the first reactor and allowed to undergo sequential anaerobic biomethanisation and aerobic decontamination (Fig. <xref ref-type="fig" rid="F1">1</xref>).</p><p>Configuration, development procedure and working conditions of the GAC-anaerobic biofilter are reported in Bertin et al. [<xref ref-type="bibr" rid="B8">8</xref>], where this innovative bioreactor system has been described in detail. In brief, it was a 2.400 l, hermetically closed and thermostated glass column reactor equipped with a recycle line, an AOMW inlet line at the bottom and an outlet line (for treated wastewater and produced biogas) departing from the top and reaching a closed reservoir hydraulically connected to a 4 l "Mariotte" bottle. After its packing with GAC (its working volume became 1.032 l) it was inoculated with the anaerobic, OMW-digesting microbial consortium developed by Beccari et al. [<xref ref-type="bibr" rid="B12">12</xref>,<xref ref-type="bibr" rid="B13">13</xref>] and then employed under strictly anaerobic conditions for 9 months of experiments (see Bertin et al. [<xref ref-type="bibr" rid="B8">8</xref>] for more details). Biofilm occurring in the reactor at the end of the study was composed by <italic>Proteobacteria</italic>, bacteria of the <italic>Flexibacter-Cytophaga-Bacteroides </italic>group, sulfate-reducing bacteria, low G+C gram-positive bacteria and, in a minor extent, <italic>Methanobacterium formicicum </italic>[<xref ref-type="bibr" rid="B8">8</xref>].</p><p>The aerobic reactor consisted of a 0.7 l glass column reactor with an external jacket in which water at 30°C was continuously recycled (Fig. <xref ref-type="fig" rid="F1">1</xref>). The inlet line and the line for supplying 0.22 <italic>μ</italic>m filter-sterilized air were at the bottom of the column, whereas the outlet lines for exhaust air and treated wastewater were placed on a small reservoir located at the top of the reactor. A recycle line continuously carried wastewater from such a reservoir to the bottom of the reactor. In the same reservoir, a probe for dissolved oxygen (97–08 model, ATI-Orion, Boston, MA) and a probe for pH (81–04 model, ATI-Orion, Boston, MA) were also placed. The bioreactor system, sterilised by recycling an aqueous ethanol solution (70 % v/v) containing HCl (1% v/v) for 2 days, was washed with sterile water and then packed with 213 g (dry weight) of SB previously sterilised in autoclave (110°C per 30 min). The developed reactor was made aerobic by supplying sterile air at 60 ml/min and then fed with the effluent of anaerobic digestor. Considering the medium displacement due to the support (0.330 l) and the supplied air (0.020 l), the actual reactor working volume was 0.350 l. The reactor was allowed to operate in batch mode at high recycling flow rate (upflow; at 0.030 l/min) for 2 weeks to permit a preliminary native biomass adhesion on SB surface. Then, it was forced to operate under continuous mode, and this by feeding it with the anaerobic digested wastewater at the same rate at which it was produced by the anaerobic digestor.</p><p>The sequential anaerobic-aerobic biofilter system was fed with AOMW at a high and constant organic load (calculated by multiplying COD or total phenol content of the influent by the dilution rate at which each reactor operated) for a 2-month period (Table <xref ref-type="table" rid="T1">1</xref>). In particular, the anaerobic GAC-digestor was fed at a dilution rate (D, expressed as the ratio between wastewater influent flow rate and the reactor reaction volume) of 2.077 day<sup>-1 </sup>with COD and total phenol loading rates of about 31.74 and 1.70 gl<sup>-1</sup>day<sup>-1</sup>, respectively. Given the AOMW decontamination efficiency achieved in the anaerobic digestor, such operative conditions imposed to the aerobic biofilter to operate at a D of 6.171 day<sup>-1 </sup>and with COD and total phenol loading rates of 50.42 and 2.04 gl<sup>-1</sup>day<sup>-1</sup>, respectively (Table <xref ref-type="table" rid="T1">1</xref>). The recycle flow rate of each reactor was set up as a function of D to have a recycle ratio (defined as the ratio of the returned flow rate to the influent flow rate) of 77 for both reactors.</p><p>Six ml samples of wastewater were taken daily through sampling ports placed along the inlet line of the integrated reactor system, the inlet and outlet lines of the aerobic reactor (Fig. <xref ref-type="fig" rid="F1">1</xref>). The collected samples were filtered on 0.22 μm cellulose-nitrate filters (Millipore, MO, USA) and then analysed for COD, the concentration of total phenols and that of volatile fatty acids as detailed in previous papers [<xref ref-type="bibr" rid="B7">7</xref>,<xref ref-type="bibr" rid="B8">8</xref>]. An aliquot of each sample was also analysed for low molecular weight phenols and aromatic compounds by HPLC [<xref ref-type="bibr" rid="B23">23</xref>]. At the end of the 2 month-experiment, the aerobic reactor was opened and triplicate 3 g-samples of SB carrier were collected at 5, 20 and 38 cm of height (from the bottom) of the reactor packed-bed and subjected to gravimetric measurements of immobilized biomass. To this aim, the carriers were gently washed with distilled water, dried at 105°C for 16 h, weighted and then suspended in a 1 M NaOH solution at 90°C for 20 min to induce the releasing of the attached biomass. Again, the biomass free carriers were washed with distilled water, dried at 105°C for 16 h and weighted. The biomass weight was calculated as the weight difference of the dried carrier before and after the NaOH treatment. A second set of SB samples (of about 20 g each) collected from the same regions of the reactor were washed and subjected to DNA extraction as described below.</p></sec><sec><title>DNA extraction</title><p>DNA was extracted from the biofilm, influent and effluent samples by using the DNeasy tissue kit (Qiagen, Italy) as described previously [<xref ref-type="bibr" rid="B8">8</xref>]. The amount and quality of nucleic acids were checked by electrophoresis on an ethidium bromide-stained 1% agarose gel.</p></sec><sec><title>16S rRNA amplification and T-RFLP analysis</title><p>16S rRNA genes from the extracted DNA samples were amplified with universal eubacterial primers 63F and 1389R [<xref ref-type="bibr" rid="B49">49</xref>]. The primer 63F was labeled with 6-FAM (5-[<xref ref-type="bibr" rid="B6">6</xref>]-carboxy-fluorescein) on the 5'-end (Applied Biosystems, Italy). Fluorescently labeled PCR products (100 ng) were digested with 10 U of restriction enzyme (Invitrogen, Italy) at 37°C for at least 4 h. T-RFLP profiles were generated using the restriction enzyme <italic>Rsa</italic>I. Additional profiles were generated using the restriction enzyme <italic>Hha</italic>I in order to confirm results obtained with <italic>Rsa</italic>I, and to assist in the assignment of tentative phylogenetic affiliations to T-RFs. The digested samples were run on an ABI Prism 310 Genetic Analyzer (Applied Biosystems) operating in a GeneScan mode with filter set D and the sizes of fragments were compared with internal standards. Replicate T-RF profiles gave reproducible fingerprints.</p></sec><sec><title>DNA sequencing and phylogenetic analysis</title><p>Partial clone libraries of 16S rRNA genes were generated from community samples. Unlabeled PCR products, purified as described above, were cloned using the pGEM-T easy vector system (Promega) and <italic>Escherichia coli </italic>JM109 according to the manufacturer's instructions. From each library randomly selected clones were screened for positive inserts and by T-RF analysis using <italic>Rsa</italic>I and <italic>Hha</italic>I restriction enzymes for digestion. Thirtyfour clones from the bacterial libraries were subjected to cycle sequencing using the M13 primers and the BigDye terminator cycle sequencing ready reaction kit (Applied Biosystems). The DNA sequences were bi-directionally resolved on an ABI Prism 310 in a sequencing mode. Nucleotide sequences were checked for potential chimeric sequences using the CHIMERA-CHECK software, and compared with the sequences in the Ribosomal Database Project (RDP) database to identify the closest relatives. The phylogenetic analysis was carried out according to the maximum likelihood method and neighbor-joining topology using the appropriate tools of the RDP program package. Bootstrapping using 1000 replicates was performed to test reliability of the branches of the trees.</p></sec></sec><sec><title>Competing interests</title><p>The author(s) declare that they have no competing interests.</p></sec><sec><title>Authors' contributions</title><p>LB carried out the experimental work concerning the performances of the integrated anaerobic-aerobic biofilter system, MCC carried out the experimental work concerning the molecular characterization of biofilm developed in the aerobic post-treatment, MR coordinated the latter research activity as well as the manuscript preparation, LM coordinated the biotechnological process research activity and FF coordinated the biotechnological process research activity as well as the manuscript preparation. All authors read and approved the final manuscript.</p></sec> |
Monthly intravenous methylprednisolone in relapsing-remitting multiple sclerosis - reduction of enhancing lesions, T2 lesion volume and plasma prolactin concentrations | <sec><title>Background</title><p>Intravenous methylprednisolone (IV-MP) is an established treatment for multiple sclerosis (MS) relapses, accompanied by rapid, though transient reduction of gadolinium enhancing (Gd+) lesions on brain MRI. Intermittent IV-MP, alone or with immunomodulators, has been suggested but insufficiently studied as a strategy to prevent relapses.</p></sec><sec sec-type="methods"><title>Methods</title><p>In an open, single-cross-over study, nine patients with relapsing-remitting MS (RR-MS) underwent cranial Gd-MRI once monthly for twelve months. From month six on, they received a single i.v.-infusion of 500 mg methylprednisolone (and oral tapering for three days) after the MRI. Primary outcome measure was the mean number of Gd+ lesions during treatment vs. baseline periods; T2 lesion volume and monthly plasma concentrations of cortisol, ACTH and prolactin were secondary outcome measures. Safety was assessed clinically, by routine laboratory and bone mineral density measurements. Soluble immune parameters (sTNF-RI, sTNF-RII, IL1-ra and sVCAM-1) and neuroendocrine tests (ACTH test, combined dexamethasone/CRH test) were additionally analyzed.</p></sec><sec><title>Results</title><p>Comparing treatment to baseline periods, the number of Gd+ lesions/scan was reduced in eight of the nine patients, by a median of 43.8% (p = 0.013, Wilcoxon). In comparison, a pooled dataset of 83 untreated RR-MS patients from several studies, selected by the same clinical and MRI criteria, showed a non-significant decrease by a median of 14% (p = 0.32). T2 lesion volume decreased by 21% during treatment (p = 0.001). Monthly plasma prolactin showed a parallel decline (p = 0.027), with significant cross-correlation with the number of Gd+ lesions. Other hormones and immune system variables were unchanged, as were ACTH test and dexamethasone-CRH test. Treatment was well tolerated; routine laboratory and bone mineral density were unchanged.</p></sec><sec><title>Conclusion</title><p>Monthly IV-MP reduces inflammatory activity and T2 lesion volume in RR-MS.</p></sec> | <contrib id="A1" equal-contrib="yes" corresp="yes" contrib-type="author"><name><surname>Then Bergh</surname><given-names>Florian </given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I8">8</xref><email>ThenBerF@medizin.uni-leipzig.de</email></contrib><contrib id="A2" equal-contrib="yes" contrib-type="author"><name><surname>Kümpfel</surname><given-names>Tania</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I9">9</xref><email>Tania.Kuempfel@med.uni-muenchen.de</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Schumann</surname><given-names>Erina</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>schumann@radiologie-friedberg.de</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Held</surname><given-names>Ulrike</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>held@slcmsr.org</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Schwan</surname><given-names>Michaela</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>michaela.schwan@arcor.de</email></contrib><contrib id="A6" contrib-type="author"><name><surname>Blazevic</surname><given-names>Mirjana</given-names></name><xref ref-type="aff" rid="I4">4</xref><email>mai8@gmx.de</email></contrib><contrib id="A7" contrib-type="author"><name><surname>Wismüller</surname><given-names>Axel</given-names></name><xref ref-type="aff" rid="I4">4</xref><email>Axel@Wismueller.de</email></contrib><contrib id="A8" contrib-type="author"><name><surname>Holsboer</surname><given-names>Florian</given-names></name><xref ref-type="aff" rid="I5">5</xref><email>holsboer@mpipsykl.mpg.de</email></contrib><contrib id="A9" contrib-type="author"><name><surname>Yassouridis</surname><given-names>Alexander</given-names></name><xref ref-type="aff" rid="I6">6</xref><email>yassou@mpipsykl.mpg.de</email></contrib><contrib id="A10" contrib-type="author"><name><surname>Uhr</surname><given-names>Manfred</given-names></name><xref ref-type="aff" rid="I7">7</xref><email>uhr@mpipsykl.mpg.de</email></contrib><contrib id="A11" contrib-type="author"><name><surname>Weber</surname><given-names>Frank</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>fweber@mpipsykl.mpg.de</email></contrib><contrib id="A12" contrib-type="author"><name><surname>Daumer</surname><given-names>Martin</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>daumer@slcmsr.org</email></contrib><contrib id="A13" contrib-type="author"><name><surname>Trenkwalder</surname><given-names>Claudia</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I10">10</xref><email>ctrenkwalder@gmx.de</email></contrib><contrib id="A14" contrib-type="author"><name><surname>Auer</surname><given-names>Dorothee P</given-names></name><xref ref-type="aff" rid="I2">2</xref><xref ref-type="aff" rid="I11">11</xref><email>Dorothee.Auer@nottingham.ac.uk</email></contrib> | BMC Neurology | <sec><title>Background</title><p>Intravenous methylprednisolone (IV-MP), 500–1000 mg for 3–5 days, is an established treatment for acute relapses of multiple sclerosis, speeding recovery [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B2">2</xref>]. It does not prevent future relapses. However, a study in isolated optic neuritis suggested that a single course of IV-MP may delay the occurrence of another episode of inflammatory demyelination and thus conversion to definite MS according to Poser's criteria [<xref ref-type="bibr" rid="B3">3</xref>,<xref ref-type="bibr" rid="B4">4</xref>]. Repeated administration of IV-MP in regular intervals may extend these beneficial effects, but this approach has not been fully evaluated and requires further study [<xref ref-type="bibr" rid="B5">5</xref>]. In secondary progressive MS, IV-MP every two months showed partial, dose-dependant benefit in several secondary endpoints [<xref ref-type="bibr" rid="B6">6</xref>]. In relapsing-remitting MS, a five-day course of IV-MP every four to six months slowed progression of MRI measures of brain atrophy, as compared to treatment for relapses only [<xref ref-type="bibr" rid="B7">7</xref>].</p><p>The mechanisms underlying the beneficial effects of corticosteroids in MS remain incompletely understood. Rapid reduction of inflammatory blood-brain-barrier (BBB) disruption, as shown by disappearance of gadolinium (Gd) enhancement after few days of treatment [<xref ref-type="bibr" rid="B8">8</xref>-<xref ref-type="bibr" rid="B10">10</xref>], is a prominent effect. Corticosteroids are also potent immunomodulatory drugs, inhibiting the expression of inflammatory mediators (including TNF and leukocyte-endothelial cell adhesion molecules) and cellular immunity (reviewed in [<xref ref-type="bibr" rid="B11">11</xref>]). Further, the endocrine and immune systems closely interact (reviewed in [<xref ref-type="bibr" rid="B12">12</xref>]), with ACTH, cortisol, prolactin (PRL) and growth hormone (hGH) being most often implicated in the regulation of the immune system. We [<xref ref-type="bibr" rid="B13">13</xref>,<xref ref-type="bibr" rid="B14">14</xref>] and others [<xref ref-type="bibr" rid="B15">15</xref>] have described a disturbed regulation of the hypothalamo-pituitary-adrenal (HPA) axis in MS, as well as hyperprolactinemia during acute attacks [<xref ref-type="bibr" rid="B16">16</xref>]. While causal relations are at present difficult to establish, corticosteroids do affect the secretion of several other hormones, and may thus have indirect effects on the immune system.</p><p>It is unknown if these effects remain active with regular administration. In order to study the effect of pulsed corticosteroids on several of these aspects, we administered a single infusion of MP, once per month, to RR-MS patients. Based on guidelines derived from natural history and treatment trial data [<xref ref-type="bibr" rid="B17">17</xref>,<xref ref-type="bibr" rid="B18">18</xref>], we designed a single-cross-over study with a six-month baseline and a six-month treatment period, and determined the impact on</p><p>1. inflammatory disease activity as shown by BBB disruption (frequency of Gd-enhancing lesions on brain MRI),</p><p>2. total T2 lesion load,</p><p>3. plasma hormone concentrations and dynamic measures of neuroendocrine regulation, and</p><p>4. selected plasma markers of immune function.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>Study design</title><p>Participation in this open-label, single-cross-over study was offered to patients with clinically definite [<xref ref-type="bibr" rid="B19">19</xref>], relapsing-remitting MS, followed in a natural history study with monthly Gd-enhanced MRI. Patients could enter that study no earlier than three months after the last relapse or course of IV-MP. In each individual patient, scans were performed at the same time of the day, and blood was drawn from an antecubital vein before each scan. We first selected those patients who had at least one Gd-enhancing lesion on the first three scans, and continued monthly visits for another three months. Final inclusion in the treatment trial required that the first six monthly scans showed at least three Gd-enhancing lesions (i.e. a mean of 0.5 Gd+ lesions per scan). Further inclusion criteria were: one or more relapses during the two years before the first MRI scan, contraindication to immunosuppressive or immunomodulatory therapy or decline of these therapies by the patient, and no contraindications to glucocorticoid treatment. Upon inclusion, monthly MRI and blood sampling was continued until month 12. Treatment started after the MRI at month 6 (see below).</p><p>The study was performed and funded by the participating institutions, with no external funding. It was approved by the ethics committee of the Bavarian State Board of Physicians (Bayerische Landesärztekammer). Patients gave written informed consent.</p></sec><sec><title>Patients</title><p>Twenty-five patients were screened by three monthly MRI scans. Ten patients fulfilled the inclusion criteria, but we excluded one patient with a very high baseline activity to reduce the risk of regression to the mean. See table <xref ref-type="table" rid="T1">1</xref> for clinical and demographic characteristics of the finally included nine patients, and figure <xref ref-type="fig" rid="F1">1</xref> for a study flow chart. None of the patients had ever been treated with immunosuppressive or immunomodulatory drugs; mean annualized relapse rate during the entire disease duration was 0.68, and during the two years prior to study inclusion, 0.8.</p></sec><sec><title>Magnetic resonance imaging</title><p>Once every month, brain MRI was performed on the same 1.5T Signa Echospeed (GE Medical Systems), including axial T1-weighted sequences (TR = 640 ms, TE = 14 ms, 4 mm thick, 1 mm gap) as well as a proton density and T2 weighted dual fast spin echo sequence (TR = 3600 ms, TE = 13 and 91 ms, resp.) and a FLAIR sequence (TR = 10 s, TE = 133 ms, TI = 2200 ms) inherently coregistered with the T1 axial scans. Axial and coronal (3 mm thick, 1.5 mm gap) T1-weighted sequences were acquired after i.v. injection of 0.1 mmol/kg gadolinium chelate (Omniscan(R)), with the interval between injection and first axial scan standardized at 6 minutes.</p><p>The number of Gd-enhancing lesions on brain MRI was determined by a blinded radiologist (E.S.) on the scans presented in random order, with the date of acquisition masked. For each patient, the mean number of active lesions per scan was calculated for baseline and treatment period. To determine lesion volume and BFV, a neural network based image analysis system was used as described [<xref ref-type="bibr" rid="B20">20</xref>]. Briefly, data preprocessing included manual delineation of the intracranial cavity contents, automatic grey value correction (coil inhomogeneity and interslice differences) and extraction of a small training data subset by interactive contour tracing of regions representing the structure classes: normal appearing grey matter (GM), normal appearing white matter (WM), cerebro-spinal fluid (CSF), and MS lesion. A hierarchical vector quantization approach was then deployed to segment the multi-dimensional image data into five classes: GM, WM, CSF, lesions and other (comprising meninges, vessels). The BFV was determined as the ratio of pixels belonging to either brain, i.e. GM, WM or MS lesions, to those belonging to brain or CSF. CSF segmentation was based on the image information obtained by all acquired MRI sequences; MS-lesions were identified based on T2, PD, and FLAIR data. The segmentation performance compared favorably (higher retest stability) with the more standard region growing segmentation algorithm for T2 lesion load (T2-LL) and angle image [<xref ref-type="bibr" rid="B21">21</xref>] for BFV [<xref ref-type="bibr" rid="B20">20</xref>]. T2-LL and BFV were calculated for four time points (start and end of baseline: months 1 and 6; and four weeks after the first and the last MP infusion: months 7 and 12).</p></sec><sec><title>Treatment</title><p>From month six to eleven, patients received 500 mg MP i.v. after each MRI scan, followed by an oral tapering dose of 40, 20 and 10 mg MP for one day each to prevent any potential temporary glucocorticoid deficiency.</p></sec><sec><title>Serum cytokine and hormone concentrations</title><p>Blood was collected, between 15:00 and 18:00 in the afternoon and always before the MRI scan, into tubes containing EDTA and trasylol, and plasma was stored at -80 degC until processing. sTNF-RI, sTNF-RII, IL-1ra and sVCAM1 were quantified by commercial enzyme-linked immunosorbent assay (ELISA) kits (s-TNF-RI and RII: Biosource Europe, Belgium; IL-1ra and VCAM-1: R&D Systems, Minneapolis, USA). Hormone concentrations were measured using commercial radioimmuno (RIA) or chemiluminescence assays (cortisol: RIA, DRG Instruments, Germany; ACTH: RIA, and prolactin: Chemiluminescence Immunoassay, both Nichols Institute Diagnostics, San Juan Capistrano, CA, USA).</p></sec><sec><title>Dexamethasone-CRH test</title><p>The dexamethasone-CRH test (Dex-CRH test) was performed, and values were derived, as described [<xref ref-type="bibr" rid="B13">13</xref>], once before the start of treatment, and once between months 11 and 12.</p></sec><sec><title>Safety</title><p>In addition to a full physical and neurological examination including EDSS [<xref ref-type="bibr" rid="B22">22</xref>], blood chemistry (including calcium, phosphate and vitamine D3), blood count, thyroid function tests, urinanalysis and pyrrolidine cross-links in urine were determined monthly. At baseline and after six months of treatment, patients underwent the short ACTH test (i.v. injection of 200 microgram tetracosactid [Synacthen(R), Novartis] at 08:00, with determination of plasma cortisol concentration before and 60 minutes after injection) and CT densitometry of the lumbar spine. All procedures were performed on an outpatient schedule.</p></sec><sec><title>Outcome measures</title><p>The mean number of Gd enhancing lesions per monthly MRI scan was defined as the primary outcome measure. In accordance with published guidelines and sample size estimates [<xref ref-type="bibr" rid="B17">17</xref>,<xref ref-type="bibr" rid="B18">18</xref>], we planned to enroll ten patients. Secondary outcome variables were T2 lesion volume and monthly plasma hormone concentrations. Serum immune variables and results of neuroendocrine tests were analysed in an exploratory fashion. Number of relapses, EDSS, routine laboratory results and bone mineral density measures were safety variables, with study termination criteria predefined for the laboratory results. Brain fractional volume (BFV) was determined to control for global changes in brain volume, potentially resulting in false results of T2 volume analyses; BFV was not an outcome measure.</p></sec><sec><title>Data analysis</title><p>According to the data structure, the effect of treatment on the number of active lesions, on hormone and cytokine receptor concentrations was analyzed using the non-parametric Wilcoxon matched-pairs signed-ranks test and a one-factorial univariate analyses of variance (ANOVA) with repeated measures design. The continuous dependent variables were transformed with the ln-transformation [Y = ln (x+1)] before ANOVAs. Associations over time between the number of active lesions to plasma hormone concentrations and immunological parameters were analysed by cross-correlation analysis. Changes of T2-LL and BFV, and all other laboratory results were analyzed by MANOVA, and followed by univariate F-tests where appropriate. As nominal level of significance, p = 0.05 was accepted; it was corrected (according to Bonferroni correction procedure) for all post-hoc tests, in order to keep the type I error less or equal to 0.05. Data are given as mean +/- standard error of the mean (SEM).</p></sec><sec><title>Comparison group</title><p>The comparison group was extracted from the database maintained at the Sylvia Lawry Center for Multiple Sclerosis Research. This database includes clinical and imaging information on patients examined according to clinical trial protocols at separate centers, including untreated control groups from several large clinical trials and natural history studies. Patients were selected based on the following criteria: A diagnosis of relapsing-remitting multiple sclerosis, no current immunosuppressive or immunomodulatory treatment, longitudinal Gd-MRI follow-up over twelve months, presence of at least 1 new Gd+ lesion during the first three months, and at least three new Gd+ lesions during the first six months. This yielded a group of 83 patients (55 females, mean age 34.4 +/- 0.8 years, mean duration of disease of 7.5 +/- 0.6 years). The MRI protocols applied 0.1 mmol/kg of gadolinium chelate. The mean interval between injection and post-contrast T1w imaging was 5.7 minutes; the delay was 5 minutes in 69 of the selected 83 patients, one minute in one patient, and between 7.5 and 15 minutes in the remaining patients. Monthly scans over the entire twelve-month period were available for seven patients; among the remaining patients, four to six scans over the first six-month-period were available for 75 patients, and for both six-month-periods for 57 patients. The mean number of new Gd+ lesions was calculated for each individual patient for the first and the second six-month-period and expressed as mean number of Gd+ lesions per scan.</p></sec></sec><sec><title>Results</title><p>During the treatment period, the mean number of active lesions per scan was reduced in eight of the nine patients, by 14 to 100% compared to the baseline period; the number increased by 21% in one patient (table <xref ref-type="table" rid="T1">1</xref>). The median rate of reduction was 43.8%, and the mean rate of reduction was 46.4%. The mean number of active lesions averaged over the group was reduced from 3.55 per month (baseline period) to 2.15 (treatment period, a reduction by 39.4%), statistically significant by Wilcoxon matched-pairs signed-ranks test (two-tailed p = 0.013). Analysis of variance also revealed a significant effect of time [Wilk's aver. multivariate test of significance; effect of time: F(11,88) = 2.15, p = 0.025]. The anti-inflammatory effect was apparent from the first scan on treatment (T1 in figure <xref ref-type="fig" rid="F2">2</xref>). Tests with polynomial contrasts in ANOVA revealed that the mean number of enhancing lesions can be optimally fitted by a polynomial curve of third order (effect of polynomial contrast of 3rd order: F(1,8) = 6.35, sig of F = 0.036).</p><p>The comparison group had a mean of 2.95 Gd+ lesions per scan during the first six months of observation. Over the next six months, the frequency of new Gd+ lesions per scan increased or decreased over a wide range in indidvidual patients, yielding a median rate of change of -14.3% (reduction), but a mean rate of change of +15.2% (increase). When analyzing the group averages, the mean number of new Gd+ lesions was reduced to 2.76 Gd+ lesions per scan (a reduction by 6.4%). These changes were not significant (Wilcoxon, p = 0.32).</p><p>In the study population, there was also a significant time effect on T2 lesion load (Wilk's multivariate test for effect of time, p = 0.002; table <xref ref-type="table" rid="T2">2</xref>). Over the five-month baseline period, T2-LL increased by 13% (p < 0.001). The first MP infusion led to a decrease in T2-LL (by 8%, p = 0.021 for comparison of month 7 to month 6, the last baseline scan), and repeated MP infusions further decreased T2-LL until month 12 (p = 0.001 for comparison to month 7 or to month 6), representing a total decrease of 21% during the treatment period. The changes in BFV were not significant (p = 0.086; table <xref ref-type="table" rid="T2">2</xref>).</p><p>IV-MP treatment exhibited a significant effect on mean plasma prolactin concentrations during baseline vs. treatment period, with lower concentrations during the treatment period (Table <xref ref-type="table" rid="T3">3</xref>; Wilks multivariate test in MANOVA (F(1,8) = 7.32, p = 0.027). Cross correlation coefficients revealed significant associations of lag -1 between plasma prolactin concentrations and the number of active lesions on MRI (Cross correlation coefficient: r = 0.728, p < 0.05). This means that prolactin levels increase or decrease one month before a corresponding change in the number of active lesions is observed on MRI. Cortisol and ACTH did not differ significantly between the baseline and treatment period (table <xref ref-type="table" rid="T3">3</xref>). Also, results of the dexamethasone-CRH test were unchanged (table <xref ref-type="table" rid="T3">3</xref>). There was no significant effect of treatment on soluble TNF receptors, IL-1 receptor antagonist or sVCAM1 (which increased slightly; table <xref ref-type="table" rid="T4">4</xref>).</p><p>Four relapses occurred during the baseline period (three treated by three-day courses of IV-MP, and MRI performed no earlier than ten days afterwards), and two mild relapses during the treatment period (no additional steroids administered). The relapses occurred in individual patients, and all recovered to the previous level of impairment. The mean EDSS score was almost unchanged between baseline (1.45 +/- 0.30) and end of study (1.40 +/- 0.37). Apart from occasional sleep disturbance on the day of infusion and mild flush-like sensations immediately following the infusion, treatment was well tolerated. Results of routine laboratory investigations were unchanged (not shown). All patients had an adequate rise of plasma cortisol after i.v. tetracosactid ("short ACTH test"), both at baseline and after six months of treatment (table <xref ref-type="table" rid="T3">3</xref>). Biochemical measures of bone metabolism (calcium, phosphate, vitamin D3 in serum, and pyrrolidine cross-links in urine, data not shown) as well as bone mineral density (CT densitometry of the lumbar spine: 140.34 +/- 15.79 before, 135.50 +/- 15.93 after treatment) were not significantly changed.</p></sec><sec><title>Discussion</title><p>We show here that a single monthly infusion of 500 mg methylprednisolone with a three-day oral taper can reduce inflammatory disease activity in patients with relapsing-remitting MS, without clinically relevant side effects. Our data support and extend two previous studies of pulsed steroid treatment in MS. In RR-MS patients, regular courses of IV-MP every four to six months, as compared to treatment for clinical relapses only, delayed accumulation of T1 hypointensities and brain atrophy [<xref ref-type="bibr" rid="B7">7</xref>]. Only a non-significant effect was seen on T2 lesion load, and the frequency of gadolinium-enhancing lesions was not assessed. In secondary analyses, disease progression appeared to be delayed, while there was no effect on the relapse rate. In secondary-progressive MS, repeated courses of IV-MP every eight weeks delayed the time to confirmed disease progression in a dose-dependant fashion, while failing with respect to the proportion of patients with disease progression (the primary outcome measure) [<xref ref-type="bibr" rid="B6">6</xref>]. MRI was not performed in that trial. IV-MP has previously been shown to rapidly, though transiently reduce gadolinium enhancement in MS, when given for several days for acute relapses [<xref ref-type="bibr" rid="B8">8</xref>-<xref ref-type="bibr" rid="B10">10</xref>,<xref ref-type="bibr" rid="B23">23</xref>]. Our study suggests that in the absence of relapse, a single infusion has a similar effect, and that the reduction of inflammatory activity can be extended by repeat infusion.</p><p>With the single cross-over study protocol and the selection of patients based on a minimum number of enhancing lesions during the baseline period, a potential alternative explanation for the reduction of enhancing lesions would be "regression to the mean". The mean number of active lesions during the baseline period is comparable to some earlier studies using the same single cross-over design as in this study, while it was lower in others [<xref ref-type="bibr" rid="B24">24</xref>-<xref ref-type="bibr" rid="B27">27</xref>]. Comparison to a pooled dataset showed that inflammatory activity in our patients was high, but not exceptional (see below). In addition, the sharpest decline in the slope of the regression curve occurs between months 6 and 7, concomitant with the start of treatment. These factors make pure "regression to the mean" highly unlikely.</p><p>In an effort to estimate the magnitude of regression to the mean, we extracted a comparison group from a large database of patients collected from various centers, who had undergone regular contrast enhanced MRI over twelve months as placebo patients of clinical trials. This comparison group of 83 patients, selected by clinical and imaging criteria equivalent to ours, had a mean frequency of Gd-enhancing lesions which is higher than would be expected for an unselected population of MS patients [<xref ref-type="bibr" rid="B28">28</xref>]. Of note, this number tends to even underestimate the true frequency of new lesions, since not all patients had strictly monthly scans, and some new lesions with a short duration of enhancement may have been missed. By natural history alone, the mean number of new enhancing lesions in the comparison group changed over a wide range during the following six months, resulting in a non-significant reduction by 6.4% for the entire group. Thus, many patients selected for a minimum frequency of active lesions appear to continue on that above-average inflammatory activity, and at least over a subsequent six-month period, "regression" to mean values as described in unselected MS patient cohorts [<xref ref-type="bibr" rid="B28">28</xref>] applies to the selected group only on a relatively limited scale. Importantly, even if we conservatively assume a more pronounced effect of regression to the mean in our patients, we can still conclude that monthly methylprednisolone contributed far more to the reduction of inflammatory activity.</p><sec><title>Corticosteroids and T2 lesion load</title><p>This study is the first to show a favorable effect of methylprednisolone on T2 lesion volume. A potentially confounding effect of the dehydrating properties of high-dose glucocorticoids, leading to pseudoatrophy [<xref ref-type="bibr" rid="B29">29</xref>,<xref ref-type="bibr" rid="B30">30</xref>], can be largely excluded because (i) no accompanying brain volume decrease was seen and (ii) serial MP infusions led to a cumulative decrease of T2 lesion volume. Previous studies over short [<xref ref-type="bibr" rid="B23">23</xref>,<xref ref-type="bibr" rid="B30">30</xref>,<xref ref-type="bibr" rid="B31">31</xref>] and longer periods [<xref ref-type="bibr" rid="B7">7</xref>] failed to demonstrate changes in T2 lesion load. In short-term studies of MP in acute exacerbations [<xref ref-type="bibr" rid="B23">23</xref>,<xref ref-type="bibr" rid="B30">30</xref>,<xref ref-type="bibr" rid="B31">31</xref>], an anti-edematous effect can be expected to be particularly prominent, while our patients received a single infusion in the absence of acute relapse. Edema resolution thus appears to contribute less to this change in T2 signal than other histological changes (extent of gliosis, macrophage infiltration or remyelination).</p></sec><sec><title>The role of prolactin during corticosteroid treatment</title><p>The parallel time-course of prolactin and Gd-enhancing lesions during both baseline and treatment with repeated steroid infusions is intriguing with respect to a possible role of prolactin in the autoimmune process. Elevated plasma levels of prolactin have been described in MS relapse, with return to normal concentrations upon recovery [<xref ref-type="bibr" rid="B16">16</xref>]. In other studies, plasma levels were higher in MS patients than in controls, albeit in the normal range [<xref ref-type="bibr" rid="B32">32</xref>], or normal [<xref ref-type="bibr" rid="B33">33</xref>,<xref ref-type="bibr" rid="B34">34</xref>]. While the role of prolactin in human immune function is a matter of debate (e.g., [<xref ref-type="bibr" rid="B35">35</xref>,<xref ref-type="bibr" rid="B36">36</xref>]), experimental data appear to support an immunostimulatory effect of prolactin [<xref ref-type="bibr" rid="B37">37</xref>-<xref ref-type="bibr" rid="B39">39</xref>]. Elevated levels of endogenous or therapeutic glucocorticoids are associated with clinically relevant immunosuppression [<xref ref-type="bibr" rid="B40">40</xref>] through several mechanisms, and at the same time reduce expression and secretion of prolactin [<xref ref-type="bibr" rid="B41">41</xref>,<xref ref-type="bibr" rid="B42">42</xref>]. A mere side-effect of steroid administration is unlikely given the significant cross-correlation with enhancing lesions during the baseline period. It remains to be determined if lower prolactin during treatment contributes to the reduction of inflammatory disease activity, or if blunted inflammation leads to a reduced level of general stressors which non-specifically stimulate prolactin secretion.</p></sec><sec><title>Measures of immune function</title><p>The panel of immunological analyses performed in this study was chosen based on specific hypotheses rather than being designed as a comprehensive immunological assessment of pulsed corticosteroid treatment. Focusing on these selected parameters, we did not observe significant changes associated with monthly IV-MP. However, this does not argue against a lasting immunological effect of pulsed steroids, which potentially could be measured by additional markers.</p></sec><sec><title>Clinical prospects for pulsed glucocorticoids in MS</title><p>We are aware that the small number of patients studied and lack of long-term data prohibit direct clinical consequences from this study and we do not suggest pulsed corticosteroids as routine treatment. However, our study demonstrates an anti-inflammatory effect with almost no relevant short-term side effects, and no indication for adrenal insufficiency or osteopenia, in line with recently published data [<xref ref-type="bibr" rid="B43">43</xref>]. Our results thus support further research [<xref ref-type="bibr" rid="B17">17</xref>].</p></sec></sec><sec><title>Conclusion</title><p>In relapsing-remitting multiple sclerosis, once-monthly IV methylprednisolone reduces inflammatory activity as measured by the best currently available surrogate measure (high-frequency contrast enhanced brain MRI). The effect is sufficient to support further study of this approach in certain groups of patients.</p></sec><sec><title>Competing interests</title><p>The author(s) declare that they have no competing interests.</p></sec><sec><title>Authors' contributions</title><p>FTB, TK: Trial design, drafting of ethics submission; general organization; data collection and analysis; initial drafting of manuscript; DPA: Trial design, MRI data acquisition and analysis; participated in initial drafting of manuscript; FTB, TK, MS, CT: patient selection, information, enrollment, clinical follow-up, endocrinological testing, oversight of patient safety; ES: Assessment of MRI imaging quality, blinded MRI data analysis; MB, AW: Volumetric MRI analysis (T2-LL and BFV); AY: Trial design (Statistics), data analysis; UH, MD: Extraction, verification and statistical analysis of comparison group data (SLC database); FH: Assessment and interpretation of neuroendocrine testing; MU, FW: Laboratory analyses and interpretation of hormone, cytokine and adhesion molecule concentrations. All authors read and approved the final manuscript.</p></sec><sec><title>Pre-publication history</title><p>The pre-publication history for this paper can be accessed here:</p><p><ext-link ext-link-type="uri" xlink:href="http://www.biomedcentral.com/1471-2377/6/19/prepub"/></p></sec> |
Intraluminal instillation of urokinase and autologous plasma: a method to unblock occluded central venous ports | <sec><title>Background</title><p>Therapeutic use and effective function of recombinant urokinase (r-UK) for occluded ports need the presence of plasminogen.</p></sec><sec sec-type="methods"><title>Methods</title><p>As a therapeutic proof of principle, we demonstrate that the use of r-UK and autologous plasma effectively reestablishes the function of occluded central venous ports (CVP) resistant to routine management of catheter occlusion. Five patients with occluded ports resistant to the routine management were treated.</p></sec><sec><title>Results</title><p>All patients were successfully treated with thrombolytic therapy using intraluminal instillation of r-UK and autologous plasma.</p></sec><sec><title>Conclusion</title><p>Instillation of r-UK and autologous plasma is a safe and effective method for management of CVP occlusion.</p></sec> | <contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Seifert</surname><given-names>Georg</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>georg.seifert@charite.de</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Riess</surname><given-names>Hanno</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>hanno.riess@charite.de</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Seeger</surname><given-names>Karl</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>karl.seeger@charite.de</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Henze</surname><given-names>Guenter</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>guenter.henze@charite.de</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Borgmann</surname><given-names>Anja</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>anja.borgmann@charite.de</email></contrib> | BMC Cancer | <sec><title>Background</title><p>Regular application of plasma products or chemotherapy is often complicated by catheter-related problems such as blockage. Especially in children, an effective and simple reestablishment of catheter patency is important. Maintenance of catheter function is usually achieved with flushing (sodium chloride and/or heparin) or urokinase instillation [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B2">2</xref>]. In the majority of cases low-dose urokinase therapy is a safe and efficacious treatment for catheter occlusion, frequently obviating the need for catheter removal. However, in some cases of occluded central venous ports (CVP), r-UK instillation is not effective. Because effective therapeutic function of r-UK requires the presence of plasminogen (see Figure <xref ref-type="fig" rid="F1">1</xref>) contained in human plasma, we combined autologous plasma and r-UK for treatment of occluded CVP in adult and paediatric patients.</p></sec><sec sec-type="methods"><title>Methods</title><p>Five patients with occluded CVP were treated. Patient characteristics are given in Table <xref ref-type="table" rid="T1">1</xref>. After up to eight attempts every 15 minutes with drawing and pressing to remove the CVP occlusion with sodium chloride and/or heparin and subsequently with 10 mg recombinant tissue plasminogen activator (r-tPA) and 5.000 IU r-UK, we used a combination of 2 ml of the patient's plasma and 10.000 IU r-UK (Medac GmbH, Germany) dissolved in 1 ml sodium chloride. The plasma was obtained by centrifugation at 1100 g for 10 minutes in a sterile syringe. Next, the combination of autologous plasma and r-UK were slowly mixed for 2 minutes within a 5 ml- syringe before application. The combination was slowly instilled, whereby a volume slightly above the CVP volume was injected. Thirty minutes after instillation, success was assessed by attempting to aspirate the combination and residual clot with an empty, sterile syringe. Up to three further attempts of aspiration were made. If the aspiration was successful, the catheter lumen was immediately flushed with 10 ml sodium chloride and/or heparin. Informed consent was obtained from patients, and, if necessary, from parents, as approved by the Institutional Ethical Committee.</p></sec><sec><title>Results</title><p>We report on five patients with an CVP occlusion that could not be managed by routine procedures. All patients had internal jugular CVPs, which had all been implanted in a surgical unit. The indications for CVP placement were regular infusion of plasma products, chemotherapy and antibiotics. We tested the new method described in this paper as a last resort before catheter replacement. All instillations resulted in successful reestablishment of catheter function after approximately 60 minutes of urokinase/plasma instillation. No patient experienced adverse effects during or after instillation. No bleeding complications occurred. The entire management procedure was easy to perform and well tolerated. In one case the instillation of urokinase and plasma failed. The explanted CVP showed that a silicon particle, which had originated during puncture, had occluded the CVP.</p></sec><sec><title>Discussion</title><p>Catheter occlusion is a common complication of long-term CVP placement. Therapeutic fibrinolytic agents like r-UK or r-tPA are used successfully to restore function of occluded CVP [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B2">2</xref>] Generally one or two instillationsof r-UK are required in the majority of patients. A 98.1% patency rate is reported when treating catheter occlusions in paediatric patients with r-UK at a concentration of 5,000 IU/mL [<xref ref-type="bibr" rid="B4">4</xref>] However, after repeated applications of plasminogen activators, a clot may become deprived of plasminogen. This will result in the failure of further attempts.</p></sec><sec><title>Conclusion</title><p>The method we have described in this paper is useful in such situations. We were able to avoid catheter replacement in all five patients with thrombotic catheter occlusions that were treated with r-UK and autologous plasma. This is especially important for children.</p></sec><sec><title>Competing interests</title><p>The author(s) declare that they have no competing interests.</p></sec><sec><title>Authors' contributions</title><p><underline>Georg Seifert</underline> was involved in acquisition of data, drafted the manuscript and has given final approval of the version to be published.</p><p><underline>Hanno Riess</underline> was involved in acquisition of data and made substantial contributions to conception and design and has given final approval of the version to be published.</p><p><underline>Karl Seeger</underline> was involved in acquisition of data and critically revised the manuscript and has given final approval of the version to be published.</p><p><underline>Günter Henze</underline> critically revised the manuscript and has given final approval of the version to be published.</p><p><underline>Anja Borgmann</underline> was involved in acquisition of data and has given final approval of the version to be published.</p></sec><sec><title>Pre-publication history</title><p>The pre-publication history for this paper can be accessed here:</p><p><ext-link ext-link-type="uri" xlink:href="http://www.biomedcentral.com/1471-2407/6/103/prepub"/></p></sec> |
Over-expression of Eph and ephrin genes in advanced ovarian cancer: ephrin gene expression correlates with shortened survival | <sec><title>Background</title><p>Increased expression of Eph receptor tyrosine kinases and their ephrin ligands has been implicated in tumor progression in a number of malignancies. This report describes aberrant expression of these genes in ovarian cancer, the commonest cause of death amongst gynaecological malignancies.</p></sec><sec sec-type="methods"><title>Methods</title><p>Eph and ephrin expression was determined using quantitative real time RT-PCR. Correlation of gene expression was measured using Spearman's rho statistic. Survival was analysed using log-rank analysis and (was visualised by) Kaplan-Meier survival curves.</p></sec><sec><title>Results</title><p>Greater than 10 fold over-expression of EphA1 and a more modest over-expression of EphA2 were observed in partially overlapping subsets of tumors. Over-expression of EphA1 strongly correlated (r = 0.801; p < 0.01) with the high affinity ligand ephrin A1. A similar trend was observed between EphA2 and ephrin A1 (r = 0.387; p = 0.06). A striking correlation of both ephrin A1 and ephrin A5 expression with poor survival (r = -0.470; p = 0.02 and r = -0.562; p < 0.01) was observed. Intriguingly, there was no correlation between survival and other clinical parameters or Eph expression.</p></sec><sec><title>Conclusion</title><p>These data imply that increased levels of ephrins A1 and A5 in the presence of high expression of Ephs A1 and A2 lead to a more aggressive tumor phenotype. The known functions of Eph/ephrin signalling in cell de-adhesion and movement may explain the observed correlation of ephrin expression with poor prognosis.</p></sec> | <contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Herath</surname><given-names>Nirmitha I</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>Nirmitha.Herath@qimr.edu.au</email></contrib><contrib id="A2" corresp="yes" contrib-type="author"><name><surname>Spanevello</surname><given-names>Mark D</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>Mark.Spanevello@qimr.edu.au</email></contrib><contrib id="A3" corresp="yes" contrib-type="author"><name><surname>Sabesan</surname><given-names>Sabe</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>Sabe_Sabesan@qld.health.gov.edu.au</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Newton</surname><given-names>Tanya</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>Tanya.Newton@qimre.edu.au</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Cummings</surname><given-names>Margaret</given-names></name><xref ref-type="aff" rid="I4">4</xref><email>mcummings@somc.uq.edu.au</email></contrib><contrib id="A6" contrib-type="author"><name><surname>Duffy</surname><given-names>Shannon</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>Shannon.Duffy@qimr.edu.au</email></contrib><contrib id="A7" contrib-type="author"><name><surname>Lincoln</surname><given-names>Douglas</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>Douglas.Lincoln@qimr.edu.au</email></contrib><contrib id="A8" contrib-type="author"><name><surname>Boyle</surname><given-names>Glen</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>Glen.Boyle@qimr.edu.au</email></contrib><contrib id="A9" contrib-type="author"><name><surname>Parsons</surname><given-names>Peter G</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>Peter.Parsons@qimr.edu.au</email></contrib><contrib id="A10" contrib-type="author"><name><surname>Boyd</surname><given-names>Andrew W</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I4">4</xref><email>Andrew.Boyd@qimr.edu.au</email></contrib> | BMC Cancer | <sec><title>Background</title><p>The sixteen vertebrate Eph receptors form the largest subfamily of receptor tyrosine kinase (RTK) proteins. Activation and signalling by these receptors is mediated by interaction with nine cell-surface counter receptors known as ephrins. The ephrins are subdivided into an A group which is glycosylphosphatidyl-inositol (GPI) anchored and a B group of type I trans-membrane proteins [<xref ref-type="bibr" rid="B1">1</xref>]. Eph proteins are also classified into A and B groups depending on structural features and preferential binding to either A or B type ephrins [<xref ref-type="bibr" rid="B2">2</xref>]. Eph and ephrin proteins have important roles in facilitating de-adhesion and cell movement, thereby playing critical roles in many developmental processes [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B3">3</xref>,<xref ref-type="bibr" rid="B4">4</xref>]. Dysregulation of cell adhesion and cell motility mechanisms have emerged as key elements in tumor progression and metastasis and it is notable that a large body of evidence details re-expression of both Eph and ephrin proteins at high levels in malignancies including melanoma and colon, gastric, breast, endometrial, and lung carcinomas [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B5">5</xref>].</p><p>EphB and ephrin B proteins have been implicated in both normal and malignant epithelial tissues. EphB2 and EphB3/ephrin B signalling regulates cell sorting in the mature gut epithelium [<xref ref-type="bibr" rid="B6">6</xref>], counter gradients of EphB and ephrin-B proteins regulating cell migration through contact-mediated cell repulsion [<xref ref-type="bibr" rid="B7">7</xref>]. Over-expression of EphB2, EphB4 and ephrin B1 is described in gastric, colon and breast cancers [<xref ref-type="bibr" rid="B8">8</xref>-<xref ref-type="bibr" rid="B12">12</xref>].</p><p>The EphA/ephrin A system is also expressed in normal and neoplastic epithelial tissues. Human EphA1 protein was isolated from the hepatocellular carcinoma cell line ETL-1 [<xref ref-type="bibr" rid="B13">13</xref>]. EphA1 is expressed in normal epithelial organs [<xref ref-type="bibr" rid="B14">14</xref>]. Over-expression of EphA1 has been described in prostate cancer [<xref ref-type="bibr" rid="B15">15</xref>], gastric cancers [<xref ref-type="bibr" rid="B16">16</xref>] and a subset of colon, lung, liver and mammary carcinomas [<xref ref-type="bibr" rid="B13">13</xref>,<xref ref-type="bibr" rid="B17">17</xref>]. EphA2 over-expression has been found in oesophageal, breast and prostate cancers [<xref ref-type="bibr" rid="B18">18</xref>,<xref ref-type="bibr" rid="B19">19</xref>]. In non-small cell lung cancer, high levels of EphA2 predicted metastasis [<xref ref-type="bibr" rid="B20">20</xref>]. Ephrin A1 is the high affinity ligand for both EphA1 [<xref ref-type="bibr" rid="B14">14</xref>] and EphA2. A study of EphA2 and ephrin A1 in the CaCo2 colon cancer line suggests that their interaction may be of importance in colon epithelial structure and function [<xref ref-type="bibr" rid="B21">21</xref>].</p><p>We describe the expression of Eph and ephrin genes in a series of ovarian cancers and non-malignant tissues using quantitative real time RT-PCR. We show that elements of both EphA/ephrin A and EphB/ephrin B signalling systems are over-expressed. A significant proportion of ovarian tumors showed a > 5-fold increase in expression of some Eph and/or ephrin proteins compared with non-malignant tissues. EphA1 and EphA2 over-expression was correlated with ephrin A1 suggesting that their interaction may have a role in ovarian cancer progression. Intriguingly, over-expression of ephrin A1 and ephrin A5 mRNA correlated with shortened survival whereas EphA1 and EphA2 expression did not. The potential consequences of this on ovarian cancer cell biology and clinical behaviour are discussed.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>Patient samples</title><p>Cancer specimens from twenty-four patients with advanced ovarian cancer or primary peritoneal carcinoma (PPC) were obtained after de-bulking surgery at the Royal Brisbane Hospital (Queensland, Australia) after obtaining informed consent and approval by the ethics committees of the Queensland Institute of Medical Research, the University of Queensland and the Royal Brisbane Hospital. These specimens were reviewed by Dr M Cummings, Pathologist at the Royal Brisbane Hospital. The patient's progress was monitored through follow-up visits with their oncologists and CA 125 levels. Two normal ovaries, one benign ovarian adenoma and one ovarian endometriotic cyst were tested as non-malignant controls.</p></sec><sec><title>RNA extraction and DNase digestion</title><p>Total RNA was isolated using the Qiagen RNeasy <sup>® </sup>Midi Kit (Qiagen Pty Ltd, Australia), according to the manufacturer's instructions. The initial part of the protocol was modified so that approximately 200 mg of tissue was crushed with a mortar and pestle in liquid nitrogen. RLT buffer containing 10μl/mlβ-mercaptoethanol (βME) was added and ground into the powdered tissue. Tissue was transferred to a conical tube and homogenized with a rotor/stator homogeniser (Polytron PT 1200, Kinematica, Switzerland) for 2 × 30 second bursts. Samples were digested with proteinase K (Qiagen Pty Ltd, Australia) and extraction continued as recommended by the manufacturer.</p><p>Prior to cDNA synthesis, the optical density of all RNA samples was measured and RNA integrity was assessed by gel electrophoresis. These samples were then subjected to DNase I treatment using RQ1 RNase-free DNase I (Promega Pty Ltd, Australia) following manufacturer's instructions.</p></sec><sec><title>cDNA synthesis</title><p>First strand cDNA was synthesized by reverse transcription using Superscript III Reverse Transcriptase (Invitrogen Pty Ltd, Australia). Briefly, DNase I-digested RNA was incubated with 1.5 μl of 500 μg/ml oligo-dT<sub>30, </sub>1.5 μl of 10 mM dNTPs and 7 μl of diethylpyrocarbonate treated ddH<sub>2</sub>O (DEPC-ddH<sub>2</sub>O) for 5 minutes at 65°C, chilled on ice and spun quickly. This mixture was incubated with 6 μl of 5x modified RT buffer (195 mM Tris-Cl pH 8.3, 375 mM KCl), 1.5 μl of 0.1 M dithiothreitol (DTT), 40 U RNasin (Promega Corp., Australia) and 1.5 μl Superscript III for 60 minutes at 50°C. The reaction was inactivated by heating at 70°C for 15 minutes. cDNA was diluted 20-fold prior to quantitative PCR.</p></sec><sec><title>Relative quantitation by real time PCR</title><p>Real-time PCR was carried out using Quantitect™ SYBR<sup>® </sup>Green PCR Master Mix (Qiagen Pty Ltd, Australia) following manufacturer's instructions. Briefly, 5 μl of diluted cDNA was added to Quantitect™ SYBR<sup>® </sup>Green PCR Master Mix. Forward and reverse primers were added to a final concentration of 0.3 μM. Primer sequences are listed in Table <xref ref-type="table" rid="T1">1</xref>. A standard curve using cDNA serially diluted further to 10<sup>-2</sup>, 10<sup>-4 </sup>and 10<sup>-6 </sup>and 18S rRNA primers was created. All reactions were performed in duplicate. Real-time PCR was carried out in a Corbett Research Rotor-Gene 3000™ (Corbett Research Pty Ltd, Australia). The PCR cycling conditions included activation for 15 minutes at 95°C and 45 cycles of 30 seconds at 95°C, 30 seconds at 55°C, and 30 seconds at 72°C. Fluorescence data was recorded at the end of each 72°C step. A DNA melt profile was run subsequently from 72°C to 95°C with a ramp of 1°C/5 seconds. Fluorescence data was recorded continuously during the melt profile. The relative expression levels of EphA1, EphA2, EphB2, EphB3, EphB6, ephrin A1, ephrin A5, and ephrin B1 were calculated using the standard curve generated from the 18S rRNA gene data.</p></sec><sec><title>Statistical methods</title><p>Correlation of gene expression was measured using Spearman's rho statistic. Receiver operating curve analysis was used to determine the best cut-off value for survival classification. This cut-off value was then used to analyse survival by log-rank analysis and (was visualised by) Kaplan-Meier survival curves. Hypotheses were tested using a 5% significance level. All tests were two-sided.</p></sec></sec><sec><title>Results</title><sec><title>Patient characteristics</title><p>Twenty-three surgical samples were obtained from operative specimens removed at the first presentation of the patient's disease and one additional sample was taken from a patient with recurrent disease. All patients underwent de-bulking surgery including total abdominal hysterectomy (TAH), bilateral salpingo-oophorectomy (BSO) and omentectomy followed by carboplatin/paclitaxel-based adjuvant chemotherapy. All patients in this study presented with stage III/IV disease with poorly differentiated serous papillary adenocarcinoma being the predominant histological type. Seven patients presented with primary peritoneal carcinoma. Median survival was 32 months with a range of 1–57 months. Individual clinical details are presented in Table <xref ref-type="table" rid="T2">2</xref>. In this series there was no significant correlation between clinical parameters (ascites, degree of differentiation, residual tumor, tumor type or age) and survival.</p></sec><sec><title>Gene expression</title><p>Initially, the expression of 20 different Eph and ephrin genes (EphA1-A8, EphB1-4 and 6, ephrin A1-5 and ephrin B1-2) were screened by quantitative real time reverse transcriptase polymerase chain reaction (Q-PCR) on two normal ovaries and ten ovarian cancer specimens. The expression levels of EphA3, A4, A5, A7, A8, B1 and B4 and of ephrin A2, A3, A4, B2 were relatively low in both normal and tumour samples and these genes were not considered further. Hence EphA1, EphA2, EphB2, EphB3, EphB6, ephrin A1, ephrin A5 and ephrin B1 were selected for further analysis in an additional fourteen tumor samples, one ovarian cyst and one ovarian adenoma. The expression levels of the two normal ovaries and two non-malignant pathological specimens were averaged and the tumor expression data are presented relative to this average (hereafter referred to as non-malignant control).</p></sec><sec><title>Expression of EphA1, EphA2 and ephrin A1</title><p>Both EphA1 and EphA2 preferentially interact with ephrin A1 and it was notable that all three were strongly expressed in a proportion of tumors. As shown in Figure <xref ref-type="fig" rid="F1">1</xref>, EphA1 was the most dramatically over-expressed gene. All tumors showed considerably higher EphA1 expression levels than those in the normal control tissues. In 22/24 tumor samples, EphA1 expression was more than five times the level seen in non-malignant controls.</p><p>EphA2 expression in all tumors was comparable to or higher than that observed in non-malignant control samples. More than half (15/23) of the tumors showed expression levels of at least two-fold greater than those seen in non-malignant control tissues, and five tumors showed a level of over-expression of greater than 5-fold.</p><p>More than half (13/24) of the tumor specimens showed greater than five-fold over-expression of ephrin A1 compared to non-malignant control samples. However, it is interesting to note that ephrin A1 correlated strongly with EphA1 (r = 0.581; p = 0.005) but only weakly with EphA2 (r = 0.387; p = 0.063).</p></sec><sec><title>Expression analysis of EphB2, EphB3, EphB6 and ephrin B1</title><p>The B type Ephs and ephrins (shown in Fig <xref ref-type="fig" rid="F1">1</xref>) with higher levels of expression in this series were EphB2, EphB3, EphB6 and ephrin B1. In most samples, EphB2 expression was comparable to or only slightly higher than the average levels found in non-malignant control samples. Only three tumors showed a greater than five-fold increase in expression. A similar trend was also observed with EphB3. In these cases some tumors appeared to have reduced expression whilst others expressed higher levels including five tumors with greater than five-fold increase in expression. EphB6 was also significantly over-expressed in eleven tumors, these showing greater than five-fold over-expression compared to control tissues.</p><p>Ephrin B1 was the most highly expressed ephrin in these samples. Expression was also relatively high in non-malignant samples with no tumors showing more than two-fold over-expression. A significant correlation of expression was found between both EphB2 and EphB3 with their ligand ephrin B1 (r = 0.411; p = 0.046 and r = 0.434; p = 0.034 respectively).</p></sec><sec><title>Ephrin over-expression in ovarian cancer</title><p>Whilst neither EphA1 nor EphA2 over-expression correlated significantly with patient survival, increased expression of both ephrin A1 and ephrin A5 correlated with decreased survival. Ephrin A1, the preferential ligand for EphA1 and a high affinity ligand for EphA2, was the highest expressed ephrin in these tumors and as shown in Fig <xref ref-type="fig" rid="F2">2A</xref>, the level of expression was strongly correlated with poor survival (r = -0.470; p = 0.024). Both benign and malignant specimens showed variable ephrin A5 expression with four and eight cancer specimens demonstrating more than 5-fold and 2-fold over-expression of ephrin A5 respectively. Ephrin A5 preferentially interacts with EphA2, A3, A5, A6, A7 and B2, of which only EphA2 and EphB2 were expressed to any significant degree in the tested normal or tumor samples. Notably, ephrin A5 expression was strongly correlated with shorter overall survival (r = -0.562; p = 0.007).</p><p>Comparison of the Kaplan-Meier survival curves (Figs <xref ref-type="fig" rid="F2">2A-C</xref>) of ephrin A5 with the next best candidate gene for poor survival, ephrin A1, in conjunction with log-rank analysis, shows that ephrin A5 is the dominating influence on overall survival. Notably, survival analysis of EphB and ephrin B gene expression showed no significant correlations with survival. Survival also did not correlate with pathological categorization of tumour grade or cellularity.</p></sec></sec><sec><title>Discussion</title><p>In this study we profiled Eph and ephrin gene expression in twenty-four cases of ovarian carcinoma. Amongst the Eph proteins the striking finding was the significant over-expression of EphA1 and, to a lesser degree, EphA2 relative to the non-malignant controls. Whilst the increased proportion of epithelial cells to stromal elements in tumors explain a small increase in expression of EphA1, the very high levels of expression in a significant proportion of tumors indicates neoplastic over-expression of EphA1. We assessed cellularity of the tumours as a possible correlate, and although such analysis is limited by sampling error, we found no association of the Eph or ephrin expression levels with tumour cellularity.</p><p>Interestingly, in a proportion of these cases, there was also over-expression of ephrin A1, the preferred ligand for both EphA1 and EphA2. Notably, expression of ephrin A1 and ephrin A5, high affinity ligands for EphA2, were increased in the tumors with poorest survival. Whilst not as dramatic, there was also statistically significant co-expression of EphB2 or EphB3 with the EphB ligand, ephrin B1. The co-expression of Ephs and ephrins on the same cells suggests that tumor-tumor cell contact, could autonomously activate the Eph/ephrin system thus promoting tumour plasticity through the well described capacity for Eph/ephrin signals to induce cell-cell repulsion [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B22">22</xref>,<xref ref-type="bibr" rid="B23">23</xref>]. Whilst all cases of high ephrin expression also had high levels of Eph expression, the converse was not true. This may suggest that high Eph expression precedes ephrin over-expression, which might explain why ephrin expression is most highly correlated with tumor aggressiveness. In keeping with Eph over-expression being a priming event it is notable that EphA2 over-expression causes malignant transformation in a normal mammary epithelial cell line (MCF-10A) [<xref ref-type="bibr" rid="B19">19</xref>]. Interestingly, Han <italic>et al </italic>previously reported that EphA2 over-expression is associated with poor prognosis in ovarian cancer [<xref ref-type="bibr" rid="B24">24</xref>]. This did not emerge in our series, perhaps reflecting differences in sample size. Notably, both ephrin A5 and ephrin A1 are high affinity ligands for EphA2 perhaps indicating a link between the two studies.</p><p>EphA1 has also been shown to be oncogenic in the classical 3T3 fibroblast assay [<xref ref-type="bibr" rid="B25">25</xref>], and co-expression of an ephrin ligand could generate an autocrine loop. This has been previously described where co-expression of ephrin A1 and EphA2 was shown to have an autocrine effect [<xref ref-type="bibr" rid="B26">26</xref>,<xref ref-type="bibr" rid="B27">27</xref>].</p><p>The finding that increased expression of ephrin A5 predicted poorer prognosis was an intriguing result as this gene is generally not expressed at significant levels in epithelial tissues (unpublished observations). Ephrin A5 does not bind significantly to EphA1 [<xref ref-type="bibr" rid="B14">14</xref>] but has been shown to activate EphA2 in human carcinoma cells [<xref ref-type="bibr" rid="B28">28</xref>].</p><p>In the current study we could not correlate mRNA levels with protein expression on the biopsy samples, principally due to the lack of antibodies appropriate for immunohistochemistry. The available antibodies are generally suitable for Western blot analysis but not for immunohistochemical applications. Using the ovarian cancer lines OVCAR4 and OVCAR5 we were able to show that protein expression (Western blot analysis) correlated precisely with mRNA analysis. In the absence of suitable antibodies, future studies should include extraction of tissue lysates for Western blot analysis.</p></sec><sec><title>Conclusion</title><p>Our results provide evidence of increased expression of Ephs and ephrins in ovarian cancer. This study provides evidence that ephrin, but not Eph, expression predicts poor prognosis in a group of ovarian cancer patients for whom clinical parameters were not informative. This suggests that co-expression of Eph and ephrin proteins may be a significant event in tumor progression. Whilst this study has yielded surprisingly strong correlations, a prospective study of a much larger cohort is warranted to further assess the use of ephrin expression as a useful predictor of clinical outcomes in this disease.</p></sec><sec><title>Competing interests</title><p>The author(s) declare that they have no competing interests.</p></sec><sec><title>Authors' contributions</title><p>NH – Participated in its design and coordination and helped to interpret and draft the manuscript, MS – Participated in its design and coordination and helped to interpret and draft the manuscript, SS – Carried out Q-PCR assays and was involved in the draft of the manuscript, TN – Extracted RNA and fixed tissue sections for histological analysis, MC – Examined all the pathological specimens, SD – Extracted RNA and was involved in the editing of the manuscript, DL – Performed statistical analyses for the manuscript, GB – Extracted RNA and fixed tissue sections for histological analysis, PP – Revisited the manuscript critically for important intellectual content, AB – Conceived the study, and participated in its design and coordination and helped to interpret and draft the manuscript.</p><p>All authors read and approved the final manuscript.</p></sec><sec><title>Pre-publication history</title><p>The pre-publication history for this paper can be accessed here:</p><p><ext-link ext-link-type="uri" xlink:href="http://www.biomedcentral.com/1471-2407/6/144/prepub"/></p></sec> |
Rapid chemokinetic movement and the invasive potential of lung cancer cells; a functional molecular study | <sec><title>Background</title><p>Non-small cell lung cancer is the most common cause of early casualty from malignant disease in western countries. The heterogeneous nature of these cells has been identified by histochemical and microarray biomarker analyses. Unfortunately, the morphological, molecular and biological variation within cell lines used as models for invasion and metastasis are not well understood. In this study, we test the hypothesis that heterogeneous cancer cells exhibit variable motility responses such as chemokinesis and chemotaxis that can be characterized molecularly.</p></sec><sec sec-type="methods"><title>Methods</title><p>A subpopulation of H460 lung cancer cells called KINE that migrated under chemokinetic (no gradient) conditions was harvested from Boyden chambers and cultured. Time-lapsed microscopy, immunofluorescence microscopy and microarray analyses were then carried out comparing chemokinetic KINE cells with the unselected CON cell population.</p></sec><sec><title>Results</title><p>Time-lapsed microscopy and analysis showed that KINE cells moved faster but less directionally than the unselected control population (CON), confirming their chemokinetic character. Of note was that chemokinetic KINE cells also chemotaxed efficiently. KINE cells were less adhesive to substrate than CON cells and demonstrated loss of mature focal adhesions at the leading edge and the presence of non-focalized cortical actin. These characteristics are common in highly motile amoeboid cells that may favour faster motility speeds. KINE cells were also significantly more invasive compared to CON. Gene array studies and real-time PCR showed the downregulation of a gene called, ROM, in highly chemokinetic KINE compared to mainly chemotactic CON cells. ROM was also reduced in expression in a panel of lung cancer cell lines compared to normal lung cells.</p></sec><sec><title>Conclusion</title><p>This study shows that cancer cells that are efficient in both chemokinesis and chemotaxis demonstrate high invasion levels. These cells possess different morphological, cytoskeletal and adhesive properties from another population that are only efficient at chemotaxis, indicating a loss in polarity. Understanding the regulation of polarity in the context of cell motility is important in order to improve control and inhibition of invasion and metastasis.</p></sec> | <contrib id="A1" contrib-type="author"><name><surname>Fok</surname><given-names>Sandra YY</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>sfok5959@mail.usyd.edu.au</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Rubin</surname><given-names>Jeffrey S</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>rubinj@mail.nih.gov</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Pixley</surname><given-names>Fiona</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>pixley@aecom.yu.edu</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Condeelis</surname><given-names>John</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>condeeli@aecom.yu.edu</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Braet</surname><given-names>Filip</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>filip.braet@emu.usyd.edu.au</email></contrib><contrib id="A6" corresp="yes" contrib-type="author"><name><surname>Soon</surname><given-names>Lilian L</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>lilian.soon@emu.usyd.edu.au</email></contrib> | BMC Cancer | <sec><title>Background</title><p>Non-small cell lung cancer (NSCLC) is the most common cause of early casualty from malignant disease in western countries and is classified into three main histological subtypes: adenocarcinomas (AC), squamous cell carcinomas (SCC), and large cell carcinomas [<xref ref-type="bibr" rid="B1">1</xref>]. The heterogeneous nature of NSCLC has been documented by both histochemical and microarray analyses. Gene expression profiling indicated that ACs are the most heterogeneous type and can be further separated into three [<xref ref-type="bibr" rid="B2">2</xref>] or four subgroups [<xref ref-type="bibr" rid="B3">3</xref>]. Ultrastructural and immunocytochemical analyses showed that 92% of NSCLC had undergone at least one differentiation event while 27% demonstrated double differentiation events: adenosquamous and adeno-neuroendocrine [<xref ref-type="bibr" rid="B4">4</xref>]. Lung cancers are therefore, subject to differentiation and are morphologically and molecularly heterogeneous. In all likelihood, lung cancer cell lines derived from tumor specimens are also heterogeneous in nature. There is, however, little known about the morphological, molecular and biological variation within cell lines used as models for invasion and metastasis.</p><p>Cancer cells disseminate from primary tumor sites utilizing different modes of motility and invasion, including protease-independent amoeboid crawling, and integrin- and protease-dependent mesenchymal migration [<xref ref-type="bibr" rid="B5">5</xref>-<xref ref-type="bibr" rid="B7">7</xref>]. Cell motility is distinguishable by response to external factors. Chemotaxis in the context of cancer is directed cell motility towards diffusible factors. Chemokinesis however, is motility in response to soluble factors in the absence of a gradient, involving a change in speed or turning behaviour. Furthermore, cell motility is generally studied in two ways; using end-point Boyden chambers [<xref ref-type="bibr" rid="B8">8</xref>] and time-lapse microscopy that allows recording and analysis of cell motility [<xref ref-type="bibr" rid="B9">9</xref>].</p><p>The response of cells to growth factors incorporates both chemokinesis and chemotaxis [<xref ref-type="bibr" rid="B10">10</xref>]. Chemokinesis may play a role during the process of epithelial-mesenchymal transition (EMT), facilitating the separation of tumor cells from the tumor mass through autocrine signals [<xref ref-type="bibr" rid="B11">11</xref>]. Chemotaxis, on the other hand, has an important role in homing mechanisms [<xref ref-type="bibr" rid="B12">12</xref>]. In these studies, however, it was unclear whether the observations were representative of a homogenous population or whether they reflected a heterogeneous population where, for instance, some cells were more chemotactic and others more chemokinetic. We hypothesize that some cancer cell lines are heterogeneous exhibiting different types motility based on the premises that (1) the type of motility depends on the cellular architecture regulated by the cytoskeleton and, (2) some cells lines demonstrate variability in cell shape and polarity, indicating differences in cytoskeletal organization, matrix adhesion, and likely cell motility.</p><p>In this study, we isolated and characterized a population of lung cancer cells designated KINE cells that showed enhanced chemokinesis compared to unselected CON cells. The two populations were evaluated for motility using Boyden chambers and time-lapse microscopy, which confirmed that KINE cells were chemokinetic. They were also less adhesive to substrate and were more invasive compared to CON cells. Evaluation of gene expression by microarray studies indicated differential gene expression patterns between the two populations with implications in chemokinesis and invasiveness.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>Cell culture</title><p>H460 lung cancer cells (ATCC) were cultured in Dulbecco's modified Eagle's medium (DMEM) (Invitrogen Corp., Carlsbad, California) supplemented with 10% fetal bovine serum (FBS) (Invitrogen), penicillin and streptomycin (Invitrogen).</p></sec><sec><title>Migration and invasion assays</title><p>Migration assays were carried out as previously described [<xref ref-type="bibr" rid="B8">8</xref>]. In brief, 2 × 10<sup>4 </sup>cells from overnight cultures were placed in upper migration chambers (Beckton Dickinson, Bedford, MA). In chemokinetic assays, both upper and lower chambers were filled with 1.5% serum in DMEM whereas for chemotactic assays, only the lower chamber contained 1.5% serum. Various combinations of serum in upper and lower chambers were used for checkerboard analyses (Table <xref ref-type="table" rid="T1">1</xref>). Incubation was carried out for 8 h in a humidified CO<sub>2 </sub>chamber and cells on the lower surface of the filter were then stained with methylene blue. The number of migrated cells were counted from a total of nine regions on the filter and calculated as numbers/cm<sup>2</sup>. Invasion assays were carried out in chambers coated with a synthetic matrix layer, ECMatrix (Chemicon). A total of 1 × 10<sup>5 </sup>cells in 1.5% FBS were placed in the upper chambers and the lower chamber was filled with 10% FBS in DMEM. The assay was carried out for 48 h and then processed as described above.</p></sec><sec><title>Isolation of chemokinetic cells</title><p>A chemokinetic assay as described above was used to isolate chemokinetic cells. Instead of staining with methylene blue, the cells on the bottom of the filter were trypsinized and allowed to expand in numbers under viable (full serum), growth conditions for two days (Figure <xref ref-type="fig" rid="F1">1A</xref>). This process of chemokinesis, cell collection and expansion was repeated a total of seven times resulting in a population of cells that were highly chemokinetic.</p></sec><sec><title>Digital microscopy and computer-assisted analysis of cell behavior</title><p>Cells were seeded on glass-bottom dishes (MatTek Corp., Ashland, MA) at a density of 0.5 × 10<sup>4 </sup>cells per dish for 16 h [<xref ref-type="bibr" rid="B13">13</xref>]. The motility of cells was recorded for the duration of 1 h by time-lapse microscopy using a ×20 objective on an inverted microscope, a charge-coupled device camera (Cooke Corp., MI, USA), and IP Lab Spectrum software. Each sequence consisting of 60 frames, taken at 1 min intervals, was imported into IMAGEJ for reconstruction into movies. The cells were traced and each traced sequence was analyzed using Dynamic Image Analysis Software (DIAS) (Soll Technologies, Inc., Iowa City, IA USA), to obtain a series of motility parameters.</p></sec><sec><title>Fluorescence microscopy</title><p>Immunofluorescence microscopy was carried out as previously described [<xref ref-type="bibr" rid="B14">14</xref>]. Cells were fixed for 5 min in 3.7% paraformaldehyde (Fisher Scientific, Springfield, NJ) and permeabilized with 0.5% Triton X-100 for 20 min. Following washes with Tris-buffered saline, pH 8.0 (TBS), the samples were treated with a solution containing Alexa-488-phalloidin conjugates (Molecular Probes, Eugene, OR), 1% BSA and 1% FBS in TBS. Anti-tyrosine-188 paxillin antibody was used to bind paxillin and a goat anti-mouse-Alexa-568 conjugate (Molecular Probes) was used as the secondary antibody for the localization and detection of the primary label. Image acquisition was conducted using a 60× objective on an Olympus microscope with filter sets in the excitation-emission range of the fluorophores (Olympus America Inc., Melville, NY).</p></sec><sec><title>Cell adhesion assay</title><p>Cells were trypsinized, washed with Dulbecco's phosphate buffered saline (DPBS) and seeded at a density of 4 × 10<sup>4 </sup>cells/ml in 500 μl of complete medium in 24-well plates for 24 h. A stock trypsin-EDTA solution (.05%) (Invitrogen Corp.) was use undiluted (100%) or diluted to 66%, 50% and 40% of the stock concentration. The cells were washed once in the diluted trypsin and incubated in fresh trypsin for 10 min at 37°C. All detached cells were collected following one wash with DPBS. The remaining cells were lifted by trypsinization and counted. The percentage of adherent cells was then calculated and graphed.</p></sec><sec><title>Microarray analysis and real-time PCR</title><p>Microarray analyses were performed as previously described [<xref ref-type="bibr" rid="B8">8</xref>]. In brief, CON and KINE H460 cells seeded at equal density for 24 h were processed for RNA precipitation using RNeasy columns (Qiagen, Valencia, CA). The RNA was converted into double-stranded cDNA and finally biotinylated cRNA in a series of reactions (Affymetrix, Santa Clara, CA). The cRNA was fragmented and used in a hybridization reaction on U133A Affymetrix chips that were subsequently processed at a fluidics station. The probe array was scanned and the data analyzed using the Microarray Suite Software (Affymetrix). A rank-based statistical test and computation of the log ratio (2) of probe signal intensities identified genes that were differentially expressed between the two populations of cells.</p><p>The microarray data were verified by real-time PCR using a set of RNA preparations different from that used in the array. In real-time PCR studies, first-strand cDNA was generated by oligo-dT priming of RNA from KINE and CON cell populations (Roche). cDNA template derived from various lung cancer cell lines was also used in a reaction to determine relative levels of expression of ROM. The template and gene-specific primers were added to a PCR mix containing the SYBR green reporter molecule (Applied Biosystems, Foster City, CA) and the PCR reaction was carried out in the ABI Prism 7000 Sequence Detector (Applied Biosystems). The comparative cycle threshold (C<sub>T</sub>) method was used to analyze the data by generating relative values of the amount of target cDNA. C<sub>T </sub>values indicate the number of cycles during amplification of target to reach a fixed threshold and correlate with the amount of starting material present. To obtain relative values, the following arithmetic formula was used: 2<sup>-ΔΔCT</sup>, where ΔCT = difference between the threshold cycles of the target and an endogenous reference (actin), and -ΔΔCT = difference between ΔCT of the target sample and a designated calibrator (vector control). The calculated result represents the amount of normalized target relative to the calibrator.</p></sec><sec><title>Microarray data</title><p>The data discussed in this publication have been deposited in NCBIs Gene Expression Omnibus [<xref ref-type="bibr" rid="B15">15</xref>] and are accessible through GEO accession numbers <ext-link ext-link-type="gen" xlink:href="GSE4869">GSE4869</ext-link>, <ext-link ext-link-type="gen" xlink:href="GSM109527">GSM109527</ext-link> and <ext-link ext-link-type="gen" xlink:href="GSM109528">GSM109528</ext-link>.</p></sec></sec><sec><title>Results</title><sec><title>Chemokinesis and chemotaxis of H460 lung cancer cells</title><p>The motility characteristics of the non-small cell lung cancer (NSCLC) cell line, H460, were investigated by checkerboard analysis using Boyden chambers. A spectrum of serum concentrations in both upper and lower chambers was used to generate positive or negative gradients across the filter as shown in Table <xref ref-type="table" rid="T1">1</xref>. Analysis of the behavior of H460 cells indicated that they were highly chemotactic and migrated well in the presence of a serum gradient (Table <xref ref-type="table" rid="T2">2</xref>); high levels of migration across the filter in the presence of serum within the lower chamber (Conditions B-D) illustrated efficient chemotaxis of these cells. Low levels of chemokinesis by a subpopulation of cells were also detected where cells migrated in the absence of a serum gradient (Conditions E, F, I-K and M-P) (Table <xref ref-type="table" rid="T2">2</xref>).</p></sec><sec><title>Enrichment for chemokinetic H460 cells</title><p>An iterative process of selection and culturing was used to isolate a subpopulation of H460 cells demonstrating chemokinesis. We recovered cells that exhibited chemokinesis under conditions where 1.5% serum was present in both upper and lower chambers. Chemokinetic conditions in the presence of serum were chosen over non-serum conditions to avoid stress induced by serum-starvation. A concentration of 1.5% was chosen because this serum level was sufficient to maintain cell survival and was shown to be effective for the assay. Chemokinetic cells were removed from the underside of the filter by trypsinization and were allowed to expand in numbers by culturing for two days in full serum (Figure <xref ref-type="fig" rid="F1">1A</xref>). This procedure (chemokinetic assay, cell collection and cell expansion) was repeated seven times to enrich for chemokinetic cells called KINE. The control population (CON) consisted of the original population of unselected cells. KINE and CON cells were then reanalysed in Boyden chamber experiments. The chemotactic environment was created by placing 1.5 % FBS in the lower chamber (+ve gradient). A chemokinetic environment was generated using 1.5 % FBS in both upper and lower chambers (-ve gradient). Figure <xref ref-type="fig" rid="F1">1B</xref> showed that KINE cells actively migrated across the filter irrespective of whether a 1.5 % serum gradient was present (+) or absent (-). CON cells, on the other hand, were more migratory in the presence of a serum gradient. KINE cells, therefore, were significantly more chemokinetic compared to CON cells but both populations of cells were efficient in their ability to chemotax (Figure <xref ref-type="fig" rid="F1">1B</xref>).</p></sec><sec><title>Cell morphology and actin organization</title><p>The morphology and structural organization of KINE and CON cells were investigated by phase contrast microscopy and fluorescence staining for actin. Cells were fixed, stained with phalloidin-Alexa-488 conjugates, and processed for epifluorescence microscopy. CON cells appeared more spread and possessed polarized shapes whether as tight epitheloid-sheets (Figure <xref ref-type="fig" rid="F2">2A</xref>) or singularly whereby a leading edge and retracting uropod were present (Figure <xref ref-type="fig" rid="F3">3A</xref>). KINE cells, on the other hand, were less adherent to neighboring cells, appeared unpolarized without distinctive uropods, and possessed actin-rich membrane ruffles at the leading edge (Figures <xref ref-type="fig" rid="F2">2D</xref>, <xref ref-type="fig" rid="F3">3E</xref>).</p></sec><sec><title>Cell adhesion, focal complexes and focal adhesions</title><p>Cell attachment to substrate is characterized by the formation of focal adhesions consisting of clusters of membrane-spanning integrins and intracellular molecules such as paxillin, FAK, talin and vinculin. The matrix adhesion structures of the lung cancer cells were investigated by immunofluorescence localization of a phosphospecific antibody to tyrosine-118 of paxillin. In CON cells, y118-paxillin localized to distinct focal adhesions at the retracting uropod (arrowhead) and at the leading edges of lamellae (arrows) (Figure <xref ref-type="fig" rid="F3">3A,C</xref>). Actin cables terminate at the leading edge where they superimpose with focal adhesions (Figure <xref ref-type="fig" rid="F3">3B,C</xref> and inset). In contrast, focal adhesions were absent at the leading edge of KINE cells. Instead, smaller, less distinct focal complexes were present in the lamellipodia of these cells (arrowhead) (Figure <xref ref-type="fig" rid="F3">3D,F</xref>). In addition, the leading edge of KINE cells was active with actin-rich ruffles (arrows) (Figure <xref ref-type="fig" rid="F3">3E,F</xref>).</p></sec><sec><title>Adhesion to the substratum</title><p>The rounded appearance of KINE cells and the lack of a distinct uropod (Figure <xref ref-type="fig" rid="F3">3D–F</xref>) is indicative of low polarity, cell tension and adhesiveness to the substratum. To investigate whether KINE cells differed from CON cells in their adhesive properties, an adhesion assay was conducted. This study investigated adhesion as an inherent property of cells that included modulation of their own adhesive environment through, for example, secretion of matrix proteins; for this reason the cell culture plates were left uncoated. Cells were grown for 24 h on 24-well plates and trypsinized for 10 min using a range of concentrations of trypsin. At low trypsin concentrations, CON cells were resistant to detachment from the culture plate whereas KINE cells detached in significantly greater numbers than CON cells (Figure <xref ref-type="fig" rid="F4">4</xref>).</p></sec><sec><title>Digital image analysis of chemokinesis</title><p>The parameters of cell motility under chemokinetic/isotropic conditions in 2-dimensional cultures were analyzed for both KINE and CON cells. The cells were plated onto tissue culture dishes for 16 h and monitored under time-lapsed microscopy. Frames were obtained 1 min apart for 1 h and analyzed using the DIAS software. Student's t-test results reject the null hypothesis that there was no difference in the total path length, directionality, speed and area but not the net path length between KINE and CON cells. KINE cells achieved significantly greater total path length compared to CON cells, whereas the net path length was not significantly different between the cell lines (Table <xref ref-type="table" rid="T3">3</xref>). The resulting calculation for directionality, a function of total path length divided by net path length, showed significant difference between KINE and CON cells. This indicated that CON cells moved more linearly than KINE cells, which made more random turns during the course of cell motility. The two cell types also showed a large difference in their migration speeds where KINE cells migrated approximately 30% faster than CON cells (Table <xref ref-type="table" rid="T3">3</xref>). The greater total path length for KINE cells in isotropic medium reflected the results for Boyden chamber assays showing that they were more motile than CON cells in the absence of a gradient (Figure <xref ref-type="fig" rid="F1">1B</xref>).</p></sec><sec><title><italic>In vitro </italic>cell invasion</title><p>To evaluate the correlation between increased chemokinesis and invasion, the invasive capacity of KINE and CON cells was examined in basement membrane-coated chamber wells. KINE and CON cells were trypsinized, washed in buffer, resuspended in 1.5% FBS containing DMEM, and placed in the upper chamber. The lower chamber was filled with 10% FBS in DMEM. The number of cells able to invade through the basement membrane was scored. KINE cells were significantly more invasive than CON, demonstrating a positive correlation between chemokinesis and cell invasion (Figure <xref ref-type="fig" rid="F5">5</xref>).</p></sec><sec><title>Identification of ROM as a gene downregulated in chemokinetic cells</title><p>The two population of cells were analyzed in microarray hybridization studies to evaluate gene expression levels. Total RNA extracted from the cells was used for the generation of double-stranded cDNA followed by the production of biotinylated cRNA. The biotinylated probes were hybridized to U133A Affymetrix gene chips that were processed and scanned, and the data subjected to binary comparative analyses using the Microarray Suite Software. A rank-based statistical test and computation of the log ratio (2) of probe signal intensities identified genes that were differentially expressed between the two populations of cells.</p><p>A gene, known by the Affymetrix target, 209606, was lower in expression by approximately four-fold in KINE cells compared to the control. This gene, named reduced on-random motile (ROM) was further analyzed in subsequent experiments. Forward and reverse primers specific to ROM were generated and used to amplify the target from first-strand cDNA in real-time PCR analysis. The cycle threshold method was used to deduce and confirm a reduction in the expression of this gene in KINE compared to CON cells (Figure <xref ref-type="fig" rid="F6">6</xref>).</p></sec><sec><title>ROM expression is reduced in lung cancer cell lines</title><p>Since the expression of ROM was downregulated in KINE which were significantly more invasive than CON cells, we investigated whether the expression of ROM may be modified in lung cancer cell lines. To test this, three non-small cell lung cancer cell lines, H1299, Calu-1, A549, two small cell lung cancer cell lines, H526 and H187, as well as two samples from normal lung tissue were analyzed by real-time PCR for the expression of ROM. All cancer cell lines apart from H526 were found to have reduced expression of ROM compared to the normal tissue samples indicating that ROM may play a role in the process of cell transformation (Figure <xref ref-type="fig" rid="F7">7</xref>).</p></sec></sec><sec><title>Discussion</title><p>The heterogeneity of the H460 large cell lung cancer cell line was investigated by selecting for chemokinetic cells from a CON population that demonstrated both chemokinesis and chemotaxis. Using Boyden chambers, cells that migrated under chemokinetic conditions were collected and their numbers expanded. Time-lapsed microscopy under isotropic conditions showed that KINE cells moved faster and changed directions more frequently than CON confirming their chemokinetic character. KINE cells which lacked stable focal adhesion were also less adhesive to culture plates compared to CON cells which had focal adhesions at the leading edge shown by phospho-Paxillin-tyr118 antibody labeling. Weak substrate adhesion in KINE cells may account for motile characteristics of rapid and random movement [<xref ref-type="bibr" rid="B16">16</xref>-<xref ref-type="bibr" rid="B19">19</xref>]. Furthermore, the selection for increased chemokinesis did not compromise the ability of KINE cells to chemotax. KINE cells were also significantly more invasive compared to CON.</p><p>These results underscore the importance of chemokinesis in cancer cell invasion suggesting that both chemokinetic and chemotactic abilities of cells should be evaluated when assessing invasive potential. Usually, chemotaxis alone is used to correlate with invasion data from <italic>in vivo </italic>and <italic>in vitro </italic>experiments. In checkerboard analyses, chemokinesis may constitute as high as 35% of total cell motility (both chemotaxis and chemokinesis) but is frequently eliminated in the subsequent evaluation of the molecular regulation of invasion [<xref ref-type="bibr" rid="B20">20</xref>]. This may introduce bias in the assessment of the relative importance of the different types of cell motility and cause attribution of molecular signals to the wrong motility type. It is possible that the chemokinetic cells are also chemotactic and represent the main cell cohort that migrated in the assays used.</p><p>Understanding the molecular process of chemotaxis is a work-in-progress in cancer, and there is even less understood about the molecular pathways and underlying processes in chemokinesis. Cancer cell chemotaxis downstream of the epidermal growth factor receptor (EGFR) pathway, involves the activation of cofilin by phospholipase Cγ (PLCγ). Cofilin, in turn, generates barbed ends on branched actin filaments and is sufficient to determine directionality of cell protrusion [<xref ref-type="bibr" rid="B21">21</xref>,<xref ref-type="bibr" rid="B22">22</xref>]. Activation of PI3K by EGFR is also relevant in motility inducing the phosphorylation of phosphtidylinositol-4, 5-biphosphate (PIP2) and catalyzing the formation of phosphtidylinositol-3, 4, 5-triphosphate (PIP3) at the plasma membrane. These inositol lipids are targets for pleckstrin-homology (PH) domain-containing proteins including nucleotide-exchange factors [<xref ref-type="bibr" rid="B23">23</xref>]. The exchange factors for Rho-like GTPases induce the recruitment and activation of Rac and Cdc24, which remodel the actin cytoskeleton [<xref ref-type="bibr" rid="B24">24</xref>-<xref ref-type="bibr" rid="B26">26</xref>] through effectors such as members of the neural Wiskott-Aldrich protein (N-WASP) family. These proteins activate Arp2/3-mediated dendritic nucleation and actin bundling to drive the formation of lamellae and filopodia [<xref ref-type="bibr" rid="B27">27</xref>]. Subsequent adhesion of these protrusions to the extracellular matrix generates both tensile and contractile forces; net forces favour the retraction of the trailing edge resulting in the forward movement of the cell [<xref ref-type="bibr" rid="B28">28</xref>,<xref ref-type="bibr" rid="B29">29</xref>].</p><p>Chemokinesis may involve a degree of stochasticity analogous to the symmetry breaking behaviour of <italic>Saccharomyces cereviesiae </italic>[<xref ref-type="bibr" rid="B30">30</xref>]. Yeast cells lacking the positional landmark Ras-family GTPase, Rsr1p, which interacts with Cdc42, are able to polarize, albeit randomly, towards a cortical site and undergo normal proliferation. The latter process, termed symmetry breaking is reliant upon the scaffold protein Bem1p, which recruits Cdc42 and promotes the local generation of more GTP-Cdc42 through a positive feedback mechanism. This process is thought to be promoted by stochastic fluctuations in the local concentrations of molecules. A feedback loop is initiated when local levels at a random cortical site exceed a threshold value. Similar stochastic processes may govern chemokinesis of cancer cells resulting in random turning behaviour.</p><p>This study shows that cancer cells capable of both chemotaxis and chemokinesis represent the highly motile cohort (faster speeds) that correlate well with invasiveness. We speculate chemokinesis in these cells may operate through an efficient stochastic process under isotropic conditions but in the presence of a strong gradient, their motility is convertible into chemotaxis. Therefore, the development of chemokinesis in highly motile cells complements the molecular machinery required for chemotaxis and does not necessarily inhibit the latter process. It follows that there may exist unique molecular determinants of chemokinesis that are separate from those that regulate chemotaxis. These determinants may also dictate hyperinvasiveness associated with the development of chemokinesis.</p><p>To investigate this further, we conducted gene chip microarray studies, which together with and real-time PCR analyses showed that a gene, reduced on-random Motile (ROM), was downregulated in chemokinetic cells. ROM encodes an adaptor protein [<xref ref-type="bibr" rid="B31">31</xref>] with several functional domains including a PDZ domain, a leucine-rich (LEU) and a PDZ domain binding region at the C-terminus. It has a short history in the literature; it was isolated on two occassions as a gene that was upregulated in the TH1 subset of cells and was called cytohesin-binding protein (CBP) [<xref ref-type="bibr" rid="B32">32</xref>] and cytohesin-binder and regulator (CYBR). The LEU region of CBP/CYBR/ROM was shown to bind cytohesin-1 and enhance the activity of ADP ribosylation factor-1 (ARF1) GTPase [<xref ref-type="bibr" rid="B33">33</xref>].</p><p>ROM has a closely related homologue known as GRP-1-associated-scaffold-protein (GRASP) sharing an overall 50% sequence homology and the presence of similar domains. GRASP has been shown to bind exchange factors for ARF GTPases, cytohesin-2 and GRP1 [<xref ref-type="bibr" rid="B34">34</xref>]. Tamalin, the mouse homologue of GRASP, was demonstrated to target glutamate receptors to the synapses and to bind proteins involved in establishing polarity in cells [<xref ref-type="bibr" rid="B35">35</xref>]. The functions of GRASP and Tamalin in plasma membrane targeting and polarization are highly suggestive that ROM may have similar roles. The expression of ROM was investigated in lung cancer cells and it was found that several cell lines were downregulated in the transcription of ROM compared to normal bronchial epithelial cells. A small cell lung cancer cell line, H187, did not have detectable levels of ROM transcripts.</p></sec><sec><title>Conclusion</title><p>In summary, we tested the hypothesis that a cancer cell line may consist of subpopulations of cells that differed not only morphologically but also in their motility, invasiveness and molecular characteristics. Boyden chambers were used to isolate a subpopulation named KINE cells that were efficient at chemokinesis. However, we discovered that these cells were also chemotactic similar to the unselected CON population. The ability of the highly chemokinetic cells to chemotax is not surprising given that chemotaxis is not an efficient process in cancer cells, requiring steep gradients of chemoattractant [<xref ref-type="bibr" rid="B36">36</xref>]. KINE cells were less adhesive, more rounded in shape (less polarized), moved faster and were more invasive than CON cells. Microarray analysis and real-time PCR demonstrated that a gene, ROM, that was downregulated in KINE cells was also repressed in several lung cancer cell lines. This is suggestive of a role in cancer cell motility and invasion.</p></sec><sec><title>Competing interests</title><p>The author(s) declare that they have no competing interests.</p></sec><sec><title>Authors' contributions</title><p>LS, SF and FB participated in the design, funding, coordination, molecular genetic studies, statistical analyses and writing of the manuscript. FP, JC, JR participated in the design and experiments of the cell adhesion studies including localization of actin and paxillin. All authors have read and approved the final manuscript.</p></sec><sec><title>Pre-publication history</title><p>The pre-publication history for this paper can be accessed here:</p><p><ext-link ext-link-type="uri" xlink:href="http://www.biomedcentral.com/1471-2407/6/151/prepub"/></p></sec> |
Hyperfibrinogenemia is associated with lymphatic as well as hematogenous metastasis and worse clinical outcome in T2 gastric cancer | <sec><title>Background</title><p>Abnormal hemostasis in cancer patients has previously been described, however the correlation between the plasma fibrinogen level and cancer metastasis and prognosis has not been reported in a large-scale clinical study.</p></sec><sec sec-type="methods"><title>Methods</title><p>Preoperative plasma fibrinogen levels were retrospectively examined in 405 patients who underwent surgery for advanced gastric cancer. The association of fibrinogen levels with clinical/pathological findings and clinical outcome was evaluated.</p></sec><sec><title>Results</title><p>There was a positive correlation between plasma fibrinogen levels and the depth of invasion (p < 0.05). Hyperfibrinogenemia (>310 mg/dl) was independently associated with lymph node (Odds Ratio; 2.342, P = 0.0032) and liver (Odds Ratio; 2.933, P = 0.0147) metastasis, not with peritoneal metastasis in this series. Patients with hyperfibrinogenemia showed worse clinical outcome in T2 gastric cancer, however, there was no correlation of plasma fibrinogen level with prognosis in T3/T4 gastric cancer.</p></sec><sec><title>Conclusion</title><p>Our results might support the idea that hyperfibrinogenemia can augment lymphatic and hematogeneous metastasis of advanced gastric cancer, which is major determinant of the prognosis in T2 gastric cancer. Therefore, in the situation without peritoneal involvement, hyperfibrinogenemia is a useful biomarker to predict the possible metastasis and worse clinical outcome in T2 gastric cancer.</p></sec> | <contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Yamashita</surname><given-names>Hiroharu</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>hyamashi-tky@umin.ac.jp</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Kitayama</surname><given-names>Joji</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>kitayama-1su@h.u-tokyo.ac.jp</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Kanno</surname><given-names>Nobuko</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>kannon-lab@h.u-tokyo.ac.jp</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Yatomi</surname><given-names>Yutaka</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>yatomiy-lab@h.u-tokyo.ac.jp</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Nagawa</surname><given-names>Hirokazu</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>nagawa-1su@h.u-tokyo.ac.jp</email></contrib> | BMC Cancer | <sec><title>Background</title><p>An increased frequency of thrombosis in patients with gastrointestinal cancer was first documented in 1865 [<xref ref-type="bibr" rid="B1">1</xref>]. Since then, a number of studies have focused on the relationship between various cancers and hemostatic factors. Abnormalities in blood coagulation are detectable in patients with malignancy, including thrombocytosis and elevated markers of coagulation activation such as fragment 1+2, thrombin-antithrombin III complexes (TAT), fibrinopeptide A (FPA), and D-Dimer [<xref ref-type="bibr" rid="B2">2</xref>-<xref ref-type="bibr" rid="B4">4</xref>]. More recently, a correlation between these factors and the prognosis of malignancies was documented. Thrombocytosis is thought to be associated with poor prognosis in gastric cancer [<xref ref-type="bibr" rid="B5">5</xref>], as well as esophageal cancer [<xref ref-type="bibr" rid="B6">6</xref>], lung cancer [<xref ref-type="bibr" rid="B7">7</xref>], colon cancer [<xref ref-type="bibr" rid="B7">7</xref>], renal cell carcinoma [<xref ref-type="bibr" rid="B8">8</xref>], and gynecological malignancies [<xref ref-type="bibr" rid="B9">9</xref>], D-Dimer is also reported to be associated with poor prognosis in patients with lung cancer [<xref ref-type="bibr" rid="B10">10</xref>,<xref ref-type="bibr" rid="B11">11</xref>] and colorectal cancer [<xref ref-type="bibr" rid="B12">12</xref>,<xref ref-type="bibr" rid="B13">13</xref>] as well as being a good predictor of survival and disease progression.</p><p>Fibrinogen, an essential hemostatic factor, is converted to fibrin (a final product of the hemostatic pathway) by activated thrombin. Dvorak suggested that elevated fibrinogen concentrations were frequently observed in cancer patients with malignant disease [<xref ref-type="bibr" rid="B14">14</xref>]. Moreover, in gastric cancer, preoperative plasma fibrinogen levels correlate with extent of tumor [<xref ref-type="bibr" rid="B15">15</xref>]. In our previous study, we found that preoperative plasma fibrinogen level was a useful predictor for lymph node metastasis in patients with gastric cancer [<xref ref-type="bibr" rid="B16">16</xref>]. Tumor marker carcinoembryonic antigen (CEA) as well as inflammatory biomarker C-reactive protein (CRP) did not show the independent association with lymph node metastasis, suggesting that hyperfibrinogenemia was not a simple result of cancer progression and might support the metastatic process by providing beneficial microenvironment around the tumor. This idea can be partially supported by recent studies in fibrinogen-deficient mice that revealed hematogeneous and lymphatic metastases were greatly reduced, indicating the positive roles of fibrinogen in the metastatic progression of cancer [<xref ref-type="bibr" rid="B17">17</xref>-<xref ref-type="bibr" rid="B19">19</xref>]. Taken together with our previous results, we hypothesized that fibrinogen might augment metastasis in human as well, and a higher level of plasma fibrinogen might be a good clinical marker of metastatic disease and worse clinical outcome. Based on this assumption, we focused on the possible impact of preoperative plasma fibrinogen level on gastric cancer prognosis. We retrospectively examined the preoperative plasma fibrinogen level of patients with advanced gastric cancer who underwent gastrectomy, and evaluated an association of these findings with clinical and pathological factors as well as clinical outcome.</p></sec><sec sec-type="methods"><title>Methods</title><p>A total of 442 patients with advanced gastric cancer underwent gastrectomy between January 1985 and December 1999 at the First Department of Surgery, University of Tokyo Hospital, Tokyo. Oral consent was obtained from each patient for blood test. Among them, plasma fibrinogen levels were evaluated before surgery in 424 patients. Because the plasma fibrinogen level is critically affected by the presence of inflammation, liver cirrhosis and chronic renal failure, five patients with apparent acute inflammatory disease (two with acute cholecystitis, three with acute peritonitis due to stomach perforation), 10 patients with liver cirrhosis and four patients with chronic renal failure were excluded from the study to minimize confounding factors. The remaining 405 patients were enrolled in this study. We also evaluated fibrinogen levels in 387 patients with early gastric cancer and 124 patients with non-inflammatory benign disease who underwent surgery during the same period. Our protocol was approved by the ethics committee of faculty of medicine, the University of Tokyo.</p><p>In this study we referred to the classifications established by the Japanese Research Society for Gastric Cancer [<xref ref-type="bibr" rid="B20">20</xref>] which define T1 as a lesion confined to the mucosal or submucosal layer; T2 as a tumor invading the proper muscle layer or subserosa; T3 as a tumor penetrating serosal without invasion of adjacent structures; and T4 as a tumor invading adjacent structures. Histological classifications were defined as follows: differentiated carcinoma – well and moderately differentiated tubular adenocarcinoma and papillary adenocarcinoma; undifferentiated carcinoma – poorly differentiated adenocarcinoma, signet ring cell carcinoma and mucinous carcinoma.</p><p>The preoperative plasma fibrinogen level was measured from early morning samples taken before breakfast five to ten days before surgery. It had been determined by Clauss clotting method using DADE Thrombin Reagent™ and Coagrex-700™ automated coagulometer (both from Sysmex, Kobe, Japan). The reference range of plasma fibrinogen level was defined as between 210 and 310 mg/dl according to the mean value ± 2 SD of healthy volunteers; plasma fibrinogen levels above 310 mg/dl were defined as hyperfibrinogenemia tentatively in this study.</p><p>Statistical analysis was carried out using JMP 5.1 (SAS Institute, Cary, NC). Plasma fibrinogen levels were compared with one-way ANOVA followed by the SNK test. The association of fibrinogen levels with clinicopathological factors was assessed with Fisher's exact test. A multivariate stepwise logistic regression analysis was performed to identify independent variables that were correlated with hematogenous and lymphatic metastasis. The Kaplan-Meier method was used to estimate the distribution of survival curve, and log-rank test was used to compare the distributions between the groups with or without hyperfibrinogenemia. P < 0.05 was considered significant for all statistical analyses.</p></sec><sec><title>Results</title><sec><title>The association of plasma fibrinogen level with the depth of invasion and metastatic disease in gastric cancer</title><p>The mean ± standard deviation (SD) plasma fibrinogen level in the 792 patients studied was 289.5 ± 81.3 mg/dl, which was not statistically different from patients with benign disease (278.1 ± 43.5) (Figure <xref ref-type="fig" rid="F1">1</xref>). However, when patients were classified into 3 groups according to the T classification of gastric cancer, patients with T2 and T3/T4 cancer showed significantly higher fibrinogen levels than those with benign disease or those with T1 cancer (T2 304.3 ± 85.0, p < 0.01; T3/T4 327.8 ± 101.2, p < 0.01). Interestingly, an increase in the plasma fibrinogen level correlated with an increase in the depth of invasion, showing a statistically significant difference among patients when evaluated by ANOVA (p < 0.0001). The fibrinogen levels in patients with localized disease, lymph node metastases, liver metastasis and peritoneal metastasis were summarized in Table <xref ref-type="table" rid="T1">1</xref>. In patients with localized disease, preoperative fibrinogen level (mean value; 266.6) was significantly lower than that in patients with lymph node metastasis (320.5; P < 0.0001), liver metastasis (356.6; P < 0.0001), peritoneal metastasis (345.0; P < 0.0001).</p></sec><sec><title>The association of hyperfibrinogenemia with lymph node, liver, and peritoneal metastasis in advanced gastric cancer</title><p>Of 405 patients with advanced cancer, lymph node metastases in 301 patients, liver metastases in 28, and peritoneal metastases in 41 were identified. Univariate analysis revealed that lymph node metastases were significantly associated with many pathologic factors such as size of the tumor, serosal invasion, lymphatic and venous involvement as well as plasma fibrinogen level (Table <xref ref-type="table" rid="T2">2</xref>). Multivariate analysis indicated that hyperfibrinogenemia showed an independent association with lymph node metastases with an odds ratio of 2.342 (p < 0.01) (Table <xref ref-type="table" rid="T3">3</xref>). Subsequently, we focused on distant hematogeneous and peritoneal metastases. In individuals with advanced gastric cancer, hyperfibrinogenemia showed a positive association with liver metastasis (p < 0.01), not with peritoneal dissemination, by univariate analysis (Table <xref ref-type="table" rid="T2">2</xref>). Multivariate analysis showed that hyperfibrinogenemia, in addition to differentiated histology and venous involvement, remained independently associated with liver metastasis with an odds ratio of 2.933 (p < 0.05) (Table <xref ref-type="table" rid="T3">3</xref>).</p></sec><sec><title>The clinical outcome in T2 and T3/T4 gastric cancer with or without hyperfibrinogenemia</title><p>In T2 gastric cancer, patients without hyperfibrinogenemia showed an extremely good outcome even in the population with advanced cancer, and the survival rate was significantly lower for patients with hyperfibrinogenemia (P = 0.001; log-rank test) (Figure <xref ref-type="fig" rid="F2">2A</xref>). In marked contrast, hyperfibrinogenemia did not show any correlation with 5-year survival rate in T3/T4 gastric cancer, i.e. cancer invading beyond the serosal (Figure <xref ref-type="fig" rid="F2">2B</xref>).</p></sec></sec><sec><title>Discussion</title><p>A link between hyperfibrinogenemia and cardiovascular diseases such as coronary heart disease, stroke and peripheral vascular disease was previously revealed [<xref ref-type="bibr" rid="B21">21</xref>]. Similarly, a recent study by Preston and colleagues indicates that fibrinogen production is upregulated in patients with pancreatic adenocarcinoma, although the tumor stage was not determined [<xref ref-type="bibr" rid="B22">22</xref>]. This suggests a positive role for fibrinogen in the progression of malignant diseases. In the present study we found that fibrinogen levels in patients with gastric cancer did not show statistically significant difference from those in individuals with benign diseases. However, the plasma fibrinogen level gradually increased with increasing depth of tumor, and the fibrinogen levels of patients with advanced gastric cancer were significantly higher than patients with benign diseases or early tumors. Moreover, the fibrinogen levels were also significantly higher in patients with metastatic disease, which essentially supports the previous study of another group [<xref ref-type="bibr" rid="B15">15</xref>]. Hyperfibrinogenemia is a clinically relevant event in advanced stage and our results strongly suggest that fibrinogen is involved in the progression of gastric cancer during the latter phase of the disease.</p><p>Recent studies in fibrinogen deficient mice provide clear evidence that fibrinogen plays a crucial role in hematogenous and lymphatic metastasis of cancer cells [<xref ref-type="bibr" rid="B18">18</xref>]. They showed that fibrinogen (Aα-chain)-deficient mice with intravenously transferred Lewis lung carcinoma (LLC) or B16 melanoma had a significantly reduced incidence of lung metastasis compared with wild type mice [<xref ref-type="bibr" rid="B17">17</xref>]. Also, the number of metastases in regional lymph nodes and the lungs of these transgenic mice was markedly reduced when LLC was subcutaneously inoculated [<xref ref-type="bibr" rid="B18">18</xref>]. These results raised the possibility that hyperfibrinogenemia might function to enhance metastasis formation as compared to the low fibrinogen level. As a fact, hyperfibrinogenemia showed an independent association with lymph node metastasis in human gastric cancer [<xref ref-type="bibr" rid="B16">16</xref>]. Interestingly, there was no remarkable difference in the growth of subcutaneously transplanted tumors between fibrinogen-deficient and wild type mice [<xref ref-type="bibr" rid="B18">18</xref>], indicating that fibrinogen plays a major role in the development of metastases but not in the growth of the primary tumor. This is exactly consistent with our human study in which patients with early gastric cancer had a similar plasma fibrinogen level to benign subjects.</p><p>In our study, we found that plasma fibrinogen level was associated with worse prognosis in T2 gastric cancer as well as lymphatic and hematogenous metastasis. The impact of hyperfibrinogenemia, however, was not found in T3/T4 gastric cancer. T3/T4 gastric cancers are serosal-infiltrating tumors, which enhances the potential to have occult metastatic foci on the peritoneum and/or form the overt peritoneal metastasis. As a fact, peritoneum is the major site of recurrence in T3/T4 gastric cancers [<xref ref-type="bibr" rid="B23">23</xref>-<xref ref-type="bibr" rid="B25">25</xref>], and the prognosis of T3/T4 cancers mainly determined by this type of recurrence. In contrast, T2 cancer has a less frequency of peritoneal metastasis as compared with T3/T4 cancer and liver or lymph node relapse seems to be major determinant of clinical outcome in T2 cancer. In our study, however, we found no association of hyperfibrinogenemia with peritoneal metastasis, although hyperfibrinogenemia was an independently associated factor with lymph node and liver metastasis. We speculate that peritoneal metastasis has a strong impact on the clinical outcome in T3/T4 gastric cancer and therefore plasma fibrinogen level was not a prognostic marker in this population. These results suggested that some patients with hyperfibrinogenemia could not be cured by surgical modalities in T2 gastric cancer and chemotherapy might be required for the improvement of survival in this population. The clinical benefit of neoadjuvant and/or adjuvant chemotherapy has not been clearly determined in T2 gastric cancer. From our data, we can conclude that patients with hyperfibrinogenemia might be good candidates for neoadjuvant and/or adjuvant chemotherapy in T2 gastric cancer. Patients with T3/T4 gastric cancer will be candidates regardless of the plasma fibrinogen level.</p><p>Fibrinogen may enhance metastasis through several possible mechanisms. Firstly, the soluble form of fibrinogen could serve as the bridging molecule between tumor cells and host cells. Fibrinogen is a dimeric molecule with multiple integrin or non-integrin binding motifs, and malignant cells often express high levels of fibrinogen receptors, such as α5β1, αvβ3 integrins or ICAM-1 molecule. If fibrinogen were to bind to ICAM-1 on endothelial cells, it might promote stable adhesion of tumor cells to the endothelium of target organs. In addition, tumor cells and platelets can form large aggregates through the binding of fibrinogen, because platelet αIIbβ3 integrin receptors have a high affinity for fibrinogen. These aggregates effectively form microemboli in target organs, which can protect tumor cells from the innate immune system [<xref ref-type="bibr" rid="B26">26</xref>,<xref ref-type="bibr" rid="B27">27</xref>]. A study by Palumbo and colleagues found that the number of tumor cells located in the lung was markedly decreased in fibrinogen deficient mice after 4–24 hours of tumor inoculation, but at no earlier time points. This suggests that fibrinogen is essential for the sustained adherence of tumor cells to the endothelia of target organs [<xref ref-type="bibr" rid="B17">17</xref>].</p><p>Recent studies have shown that thrombin might also be an important modulator of cancer metastasis <italic>in vivo</italic>. During the conversion of prothrombin to thrombin by prothrombinase, the polypeptide prothrombin fragment 1+2 (F1+2) is released. The plasma F1+2 level is therefore a useful marker of thrombin generation. A number of studies have shown that F1+2 plasma levels are elevated in patients with specific types of malignancies [<xref ref-type="bibr" rid="B4">4</xref>,<xref ref-type="bibr" rid="B28">28</xref>,<xref ref-type="bibr" rid="B29">29</xref>] but not in other types [<xref ref-type="bibr" rid="B30">30</xref>-<xref ref-type="bibr" rid="B32">32</xref>]. Rahr and coworkers reported no difference in plasma F1+2 levels between patients with and without gastric cancer [<xref ref-type="bibr" rid="B32">32</xref>]. In the present study, prothrombin time (PT) was examined in each patient and no significant association with the presence of cancer, metastasis, tumor stage or plasma fibrinogen level was found (data not shown). Moreover, Kerlin and co-workers recently demonstrated that elevated fibrinogen levels in hyperfibrinogenemia transgenic mice with suppressed thrombin activity do not alter the incidence or extent of thrombus formation [<xref ref-type="bibr" rid="B33">33</xref>]. It therefore seems unlikely that high fibrinogen levels are associated with systemic thrombin activation and subsequent enhancement of fibrin formation in gastric cancer.</p><p>Although there is sufficient evidence to suggest that elevated fibrinogen levels might aid the development of metastatic lesions, we cannot ignore the possibility that this elevation is simply the result of the tumor mass spreading. Fibrinogen, which is one of the major acute phase proteins produced by the liver, is greatly enhanced in response to infection or other inflammatory disorders. Indeed, the fibrinogen level of all of the patients with preoperative acute inflammatory disorders was high. Inflammatory proteins, such as IL-6 or CRP, are reported to be higher in cancer patients compared to non-cancer patients. In our previous study, we found that CRP level did not show an independent association with lymphatic metastasis although plasma fibrinogen level did [<xref ref-type="bibr" rid="B16">16</xref>], suggesting that hyperfibrinogenemia, different from high serum CRP, may not be a simple by-product of inflammatory response caused by tumor progression but might have some etiologic relevance for tumor metastasis and accordingly relationship with clinical outcome in gastric cancer. In this retrospective study as well as our previous study, however, we could not show the exact causal relationship between hyperfibrinogenemia and cancer metastasis.</p></sec><sec><title>Conclusion</title><p>Metastasis is the unequivocal hallmark of cancer. The acquisition of metastatic ability leads to clinically incurable disease and resultingly worse clinical outcome for most cancer cell types. Our data, together with the results of basic experiments in previous studies, raises the idea that hyperfibrinogenemia causally imparts the hematogenous and lymphatic metastatic progression in patients with advanced gastric cancer, and is not simply the result of tumor progression. It also has a prognostic value in T2 gastric cancer, which might suggest hyperfibrinogenemia has the potential to be prognostic marker in other cancer type with less frequent peritoneal metastasis. Routinely examined plasma fibrinogen level is not only the factor to assess the perioperative hemorrhagic risk but also a useful biomarker to predict the possible metastasis and worse prognosis in T2 gastric cancer.</p></sec><sec><title>Abbreviations</title><p>thrombin-antithrombin III complexes: TAT, fibrinopeptide A: FPA, carcinoembryonic antigen: CEA, C-reactive protein: CRP, Lewis lung carcinoma: LLC, prothrombin fragment 1+2: F1+2, prothrombin time: PT.</p></sec><sec><title>Competing interests</title><p>The author(s) declare that they have no competing interests.</p></sec><sec><title>Authors' contributions</title><p><bold>HY </bold>have been involved in preparing the manuscript for submission and made substantial contributions to conception and design, acquisition of data, and analyses and interpretation of data;</p><p><bold>JK </bold>contributed to analyses and interpretation of data and their interpretation and was involved in revising the manuscript for important intellectual content;</p><p><bold>NK </bold>and <bold>YY </bold>have been involved in plasma fibrinogen measurement and data analyses;</p><p><bold>HN </bold>participated in the design of the study and was involved in discussion of the results.</p></sec><sec><title>Pre-publication history</title><p>The pre-publication history for this paper can be accessed here:</p><p><ext-link ext-link-type="uri" xlink:href="http://www.biomedcentral.com/1471-2407/6/147/prepub"/></p></sec> |
Characteristic abnormalities in cerebrospinal fluid biochemistry in children with cerebral malaria compared to viral encephalitis | <sec><title>Background</title><p>In developing countries where <italic>Plasmodium falciparum </italic>malaria is endemic, viral encephalitis and cerebral malaria are found in the same population, and parasitemia with <italic>Plasmodium falciparum </italic>is common in asymptomatic children. The objective of this study was to investigate the cerebrospinal fluid (CSF) biochemistry in children with cerebral malaria compared to those with presumed viral encephalitis.</p></sec><sec sec-type="methods"><title>Methods</title><p>We studied the following CSF parameters: cell count, glucose, protein, lactic dehydrogenase (LDH) and adenosine deaminase (ADA) levels, in children with cerebral malaria, with presumed viral encephalitis, and in control subjects who had a lumbar puncture after a febrile convulsion with postictal coma.</p></sec><sec><title>Results</title><p>We recruited 12 children with cerebral malaria, 14 children with presumed viral encephalitis and 20 controls prospectively, over 2 years in the Government General Hospital in Kakinada, India. Patients with cerebral malaria had significantly lower CSF glucose, and higher protein, LDH, CSF/blood LDH ratio and CSF ADA levels but a lower CSF/serum ADA ratio compared to controls (<italic>p </italic>< 0.01). Patients with cerebral malaria had lower CSF white cell count, glucose, protein, LDH levels and CSF/serum ADA ratio compared to patients with presumed viral encephalitis. CSF/serum ADA ratio was lower in patients with cerebral malaria due to the fact that serum ADA levels were significantly higher in patients with cerebral malaria compared to the other two groups. A CSF/serum ADA ratio of <0.38 and a CSF glucose level of <3.4 mmol/l were selected as the cut-off values with the highest sensitivities and specificities for comparing the two conditions.</p></sec><sec><title>Conclusion</title><p>CSF/serum ADA ratio and CSF glucose levels were the best discriminators of cerebral malaria from presumed viral encephalitis in our study. Further studies are needed to explore their usefulness in epidemiological studies.</p></sec> | <contrib id="A1" contrib-type="author"><name><surname>Jakka</surname><given-names>SR</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>jakkasrinu@yahoo.com</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Veena</surname><given-names>S</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>jakkasrinu@yahoo.com</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Atmakuri</surname><given-names>RM</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>jakkasrinu@yahoo.com</email></contrib><contrib id="A4" corresp="yes" contrib-type="author"><name><surname>Eisenhut</surname><given-names>M</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>michael.eisenhut@talk21.com</email></contrib> | Cerebrospinal Fluid Research | <sec><title>Background</title><p>Previous studies have compared cerebrospinal fluid (CSF) features of bacterial meningitis and cerebral malaria [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B2">2</xref>]. This distinction is in most cases not difficult, because the majority of patients with bacterial meningitis have marked CSF pleocytosis, in contrast to cerebral malaria where this is uncommon. More difficult is the distinction of cerebral malaria from viral encephalitis on clinical grounds or from a CSF cell count. Rapid virological diagnostic methods are not available in most developing countries. With the benefit of acyclovir in herpes encephalitis for prevention of death or neurological sequelae [<xref ref-type="bibr" rid="B3">3</xref>], and the need to recognize the presence of possible arbovirus encephalitis with its public health implications [<xref ref-type="bibr" rid="B4">4</xref>], it has become more important to differentiate viral encephalitis from cerebral malaria. The objective of our study was, therefore, to investigate the distinguishing features of a variety of CSF parameters in patients with cerebral malaria as opposed to those with presumed viral encephalitis. We have chosen for our investigation CSF cell count, glucose and protein levels, together with CSF and serum lactic dehydrogenase (LDH) and adenosine deaminase (ADA). Previous investigations have found that these parameters may be useful in discrimination of infectious diseases of the central nervous system such as meningitis, encephalitis and cerebral malaria [<xref ref-type="bibr" rid="B5">5</xref>,<xref ref-type="bibr" rid="B6">6</xref>]. Lactic dehydrogenase is an intracellular enzyme that is released from damaged cells. Its level in the CSF reflects the degree of damage to cells in the central nervous system. CSF- adenosine deaminase, an enzyme mainly produced by developing immature T-lymphocytes, is increased in the body fluids of patients with conditions associated with stimulation of cellular immunity and was evaluated in this study for its use in differentiation of parasitic and viral infection of the brain.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>Patients</title><p>In a 2-year prospective study at Government General hospital in Kakinada, India we recruited with ethical approval and informed consent by parents or guardians, children with a clinical diagnosis of presumed viral encephalitis, cerebral malaria and controls. The definition of presumed viral encephalitis was a reduced level of consciousness and pyrexia that could not be explained by a metabolic abnormality, dehydration or shock, negative blood and CSF cultures, and a negative blood slide for malaria parasites with recovery without antibiotics or lack of response to broad-spectrum antibiotics [<xref ref-type="bibr" rid="B7">7</xref>]. Cerebral malaria was defined as coma and pyrexia with a positive thick film for asexual <italic>P. falciparum </italic>blood stages and no other identified cause of an encephalopathy following the WHO definition [<xref ref-type="bibr" rid="B8">8</xref>]. We recruited control subjects from a population of patients who had lumbar punctures to exclude meningitis in the context of reduced consciousness following febrile convulsions and who recovered from the postictal state without further signs of central nervous system illness or septicaemia.</p></sec><sec><title>Sample collection and analysis</title><p>CSF was obtained by lumbar puncture as soon as possible after admission if there were no contraindications, and after informed consent by the parents. CSF cell count, glucose, protein and CSF and serum LDH levels were determined by standard methods as in a previous study published by this group [<xref ref-type="bibr" rid="B9">9</xref>]. ADA levels were determined using the Berthelot reaction, through the ammonia released when adenosine is broken down to inosine. After incubation of plasma or CSF with a buffered solution of adenosine, the ammonia is reacted with a Berthelot reagent to form a blue colour, which is proportional to the amount of enzyme activity [<xref ref-type="bibr" rid="B10">10</xref>].</p></sec><sec><title>Data analysis</title><p>For multiple comparisons of data with an approximately normal distribution the one way analysis of variance (ANOVA) was used. Tukey's HSD test was used for post hoc analysis of these data. For non-parametric data the Kruskal-Wallis test was used. The statistical method used for univariate comparisons of continuous data was the Mann-Whitney test, and for categorical data the chi-square test. For determination of a cutoff value for a CSF parameter for possible discrimination between presumed viral encephalitis and cerebral malaria, the co-ordinate of a receiver operating characteristic (ROC) curve indicating the greatest sensitivity and specificity was chosen (for explanation of the statistical methodology see [<xref ref-type="bibr" rid="B11">11</xref>]). The ROC curve is a computer-generated curve of data from the malaria and viral encephalitis patients with sensitivity on the vertical axis plotted against 1-specificity (1-true negative rate) on the horizontal axis. Sensitivity is the true positive rate in percent and it was calculated from the ratio of the number of true positive over the sum of true positive and false negative patients. True positive in the context of this study, was the presence of cerebral malaria in any patient below a certain value of a parameter, the 'cut-off'. Specificity is the true negative rate in percent and it was calculated as the ratio of the number of true negative over the sum of false positive and true negative patients. True negative meant the patient has presumed viral encephalitis if the parameter is above the "cut-off" chosen (Table <xref ref-type="table" rid="T2">2</xref>). The positive predictive value is the post-test probability of a positive test and negative predictive value the post-test probability of a negative test [<xref ref-type="bibr" rid="B12">12</xref>]. The positive predictive value is calculated as the ratio of the number of true positive over the sum of true positive and false positive patients. The negative predictive value is the ratio of the number of true negative over the sum of false negative and true negative patients. The probability for the positive predictive value was the probability of having cerebral malaria in any patient and for the negative predictive value, the probability of having presumed viral encephalitis. A <italic>p</italic>-value < 0.05 was taken as indicator for a statistically significant difference. Statistical calculations were performed with SPSS version 13.0 and Epi Info 6.04b (CDC, Atlanta). Confidence intervals for parameters were calculated by Confidence Interval Analysis software.</p></sec></sec><sec><title>Results</title><p>46 children were recruited over a 2-year period. Twelve patients had cerebral malaria, 14 patients presumed viral encephalitis and 20 patients were controls. Demographic and laboratory parameters including LDH and ADA CSF/serum ratios are listed in Table <xref ref-type="table" rid="T1">1</xref>. There was no significant difference in age between patients with cerebral malaria and presumed viral encephalitis, although both groups were both significantly older than controls. The CSF white cell count was significantly higher in the group with presumed viral encephalitis compared to patients with cerebral malaria or controls. The mean CSF glucose level was significantly lower in patients with cerebral malaria compared to the other two groups and the range for glucose levels in patients with cerebral malaria did not overlap with that for patients with presumed viral encephalitis (Figure <xref ref-type="fig" rid="F1">1</xref>). Patients with cerebral malaria had significantly higher CSF protein, LDH and ADA levels and CSF/serum LDH ratio but a lower CSF/serum ADA ratio, compared to controls (Table <xref ref-type="table" rid="T1">1</xref>). Patients with cerebral malaria had significantly lower CSF glucose, protein and LDH levels and CSF/serum ADA ratios compared to patients with presumed viral encephalitis (Table <xref ref-type="table" rid="T1">1</xref>). Patients with cerebral malaria had a significantly higher serum ADA level compared to patients with presumed viral encephalitis and to controls (Mann-Whitney test, p < 0.05, Figure <xref ref-type="fig" rid="F2">2</xref>). Serum ADA levels in patients with presumed viral encephalitis were not significantly different from controls (<italic>p </italic>> 0.05). Serum LDH levels were not significantly different between groups (<italic>p </italic>> 0.05). However, CSF/Serum LDH ratios for both patients with cerebral malaria and presumed viral encephalitis were significantly higher than controls.</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>Comparison of demographic and cerebrospinal fluid parameters between patients with cerebral malaria, presumed viral encephalitis and controls.</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td align="left">Cerebral malaria n = 12</td><td align="left">Presumed viral encephalitis n = 14</td><td align="left">Controls n = 20</td><td align="left"><italic>p</italic>-value for multiple comparisons</td></tr></thead><tbody><tr><td align="left">Age (years)</td><td align="left">7.0 (2.7)</td><td align="left">7.0 (3.5)</td><td align="left">1.7 (1.2)</td><td align="left"><0.01<sup>a</sup></td></tr><tr><td align="left">Gender (male)</td><td align="left">7</td><td align="left">6</td><td align="left">9</td><td align="left">ns<sup>b</sup></td></tr><tr><td align="left">CSF-parameters:</td><td></td><td></td><td></td><td></td></tr><tr><td align="left"> White cells (cells/μl)<sup>c</sup></td><td align="left">0 (0–3)</td><td align="left">4 (0–9)</td><td align="left">0 (0–4)</td><td align="left"><0.01<sup>d</sup></td></tr><tr><td align="left"> Glucose (mmol/l)</td><td align="left">2.7 (0.3)</td><td align="left">4.2 (0.4)</td><td align="left">3.3 (0.6)</td><td align="left"><0.01<sup>a</sup></td></tr><tr><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left"> Protein (g/l)</td><td align="left">0.4 (0.04)</td><td align="left">0.5 (0.06)</td><td align="left">0.3 (0.10)</td><td align="left"><0.01<sup>a</sup></td></tr><tr><td align="left"> CSF-ADA (IU/l)</td><td align="left">4.6 (0.9)</td><td align="left">3.8 (1.0)</td><td align="left">3.3 (0.8)</td><td align="left"><0.01<sup>a</sup></td></tr><tr><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left"> Serum ADA (IU/l)<sup>c</sup></td><td align="left">14.4 (11.6–17.5)</td><td align="left">6.5 (4.8–10.1)</td><td align="left">6.1 (4.4–8.9)</td><td align="left"><0.01<sup>d</sup></td></tr><tr><td align="left"> Ratio CSF/serum ADA</td><td align="left">0.31 (0.05)</td><td align="left">0.54 (0.11)</td><td align="left">0.53 (0.13)</td><td align="left"><0.01<sup>a</sup></td></tr><tr><td align="left"> CSF-LDH (IU/ml)</td><td align="left">20.7 (2.8)</td><td align="left">26.5 (5.8)</td><td align="left">15.9 (3.8)</td><td align="left"><0.01<sup>a</sup></td></tr><tr><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left"> Serum LDH (IU/ml)</td><td align="left">214.1 (25.7)</td><td align="left">247.5 (46.8)</td><td align="left">244.0 (57.7)</td><td align="left">ns <sup>a</sup></td></tr><tr><td align="left"> Ratio CSF/serum LDH</td><td align="left">0.09 (0.02)</td><td align="left">0.11 (0.02)</td><td align="left">0.07 (0.01)</td><td align="left"><0.01<sup>a</sup></td></tr></tbody></table><table-wrap-foot><p>LDH = Lactic dehydrogenase, ADA = Adenosine deaminase. Results where applicable are given as mean (SD). <sup>a </sup>Analysis of variance: Post hoc analysis results with p < 0.05 (Tukey HSD test): <italic>Age</italic>: Cerebral malaria and presumed viral encephalitis versus controls. <italic>CSF white cell count</italic>: Cerebral malaria versus presumed viral encephalitis and presumed viral encephalitis versus controls. <italic>Glucose and protein</italic>: Each group versus the 2 other groups. CSF-<italic>ADA</italic>: Cerebral malaria versus controls. <italic>Serum ADA</italic>: Cerebral malaria versus the 2 other groups. <italic>Ratio CSF/serum ADA</italic>: Cerebral malaria versus the two 2 groups. <italic>CSF-LDH</italic>: Each group versus the other 2 groups. <italic>Ratio CSF/serum LDH</italic>: Cerebral malaria and presumed viral encephalitis versus controls. <sup>b </sup>ns = not significant by chi-square test. <sup>c </sup>Median and range. <sup>d </sup>Kruskal-Wallis test.</p></table-wrap-foot></table-wrap><table-wrap position="float" id="T2"><label>Table 2</label><caption><p>Cerebrospinal fluid parameters as discriminators between cerebral malaria and presumed viral encephalitis</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left">Cut-off</td><td align="left">Sensitivity<sup>1 </sup>(%)</td><td align="left">Specificity<sup>2 </sup>(%)</td><td align="left">Negative predictive Value<sup>3 </sup>(%)</td><td align="left">Positive predictive Value<sup>4 </sup>(%)</td></tr><tr><td></td><td align="left">(95% CI)</td><td align="left">(95% CI)</td><td align="left">(95% CI)</td><td align="left">(95% CI)</td></tr></thead><tbody><tr><td align="left">CSF white cell count <4 cells/microl</td><td align="left">100 (76–100)</td><td align="left">57 (32–78)</td><td align="left">100 (67–100)</td><td align="left">66 (43–83)</td></tr><tr><td align="left">CSF protein <0.43 g/l</td><td align="left">92 (64–98)</td><td align="left">78 (52–92)</td><td align="left">91 (64–98)</td><td align="left">78 (52–92)</td></tr><tr><td align="left">CSF glucose <3.4 mmol/l</td><td align="left">100 (78–100)</td><td align="left">100 (75–100)</td><td align="left">100 (75–100)</td><td align="left">100 (78–100)</td></tr><tr><td align="left">Ratio CSF/serum ADA <0.385</td><td align="left">91 (64–98)</td><td align="left">100 (78–100)</td><td align="left">93 (70–100)</td><td align="left">100 (74–100)</td></tr><tr><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">CSF LDH < 22.5 IU/ml</td><td align="left">75 (46–91)</td><td align="left">71 (44–88)</td><td align="left">76 (49–92)</td><td align="left">69 (42–87)</td></tr></tbody></table><table-wrap-foot><p><sup>1</sup>Sensitivity is the true positive rate in percent. It is calculated as the ratio of the number of true positive over the sum of true positive and false negative patients. True positive meant in the context of this study the presence of cerebral malaria in a group of patients with presumed viral encephalitis or cerebral malaria below a certain value of a parameter, the "cut-off". <sup>2</sup>Specificity is the true negative rate in percent. It is calculated as the ratio of the number of true negative over the sum of false positive and true negative patients. True negative meant the patient has presumed viral encephalitis if the parameter is above the "cut-off" chosen. <sup>3</sup>The negative predictive value is the ratio of the number of true negative over the sum of false negative and true negative patients. Positive predictive value is the post-test probability of a positive test and negative predictive value the post-test probability of a negative test. <sup>4</sup>The positive predictive value is calculated as the ratio of the number of true positive over the sum of true positive and false positive patients.</p></table-wrap-foot></table-wrap><fig position="float" id="F1"><label>Figure 1</label><caption><p>Scatter plot of CSF glucose levels in patients with cerebral malaria, presumed viral encephalitis and controls.</p></caption><graphic xlink:href="1743-8454-3-8-1"/></fig><fig position="float" id="F2"><label>Figure 2</label><caption><p>Box plot of serum adenosine deaminase levels in patients with cerebral malaria, presumed viral encephalitis and controls. The line inside the box represents the median, the box the quartiles, and the whiskers the extreme values.</p></caption><graphic xlink:href="1743-8454-3-8-2"/></fig><p>For markers with a significant difference between cerebral malaria and the presumed viral encephalitis groups, we determined cut-off values of CSF parameters with maximum sensitivities and specificities and positive and negative predictive values. The best cut-off values taken from the co-ordinates of the ROC curve (see methods section) below which cerebral malaria was more likely than presumed viral encephalitis were: CSF white cell count, 4 cells/μl; CSF protein level, 0.43 g/l; CSF glucose, 3.4 mmol/l; CSF LDH, 22.5 IU/ml; CSF/serum ADA ratio, 0.385 (Table <xref ref-type="table" rid="T2">2</xref>).</p><p>CSF white cell count, protein and glucose levels and CSF/serum ADA ratio below these cut-off values indicated cerebral malaria with a sensitivity of > 90%. This resulted in the probability of having presumed viral encephalitis in our group of children with cerebral malaria or viral encephalitis (or negative predictive value for having cerebral malaria) being above 90% if the values of these parameters were above this cut-off. Specificities for the parameters were 100% for CSF glucose and 100% for CSF/serum ADA ratio and better than for other parameters, i.e. presumed viral encephalitis was always present in any patient with a value above the cut-off. This resulted in a probability (positive predictive value) of 100% of having cerebral malaria in children with parameters below this cut-off (Table <xref ref-type="table" rid="T2">2</xref>).</p></sec><sec><title>Discussion</title><p>This is to our knowledge the first study to investigate differences in CSF parameters between presumed viral encephalitis and cerebral malaria. Asymptomatic parasitaemia with <italic>Plasmodium falciparum </italic>is common in regions hyperendemic for this parasite. In many of these areas, particularly in South East Asia, viral encephalitis is also a public health problem. It is therefore important to investigate discriminating CSF features between encephalitis and cerebral malaria. Our study is the most detailed study on CSF parameters in cerebral malaria reported so far. We could not, however, exclude the possibility that patients labeled as having cerebral malaria had a combination of encephalitis and parasitaemia with <italic>P. falciparum</italic>, but the lack of white cells in their cerebrospinal fluid make this possibility seem very unlikely.</p><sec><title>CSF-glucose</title><p>A CSF glucose level below 3.4 mmol/l was the best discriminator of cerebral malaria from presumed viral encephalitis. This was partly due to the fact that in cerebral malaria CSF glucose levels were below normal range. Low CSF glucose is a well-known phenomenon in cerebral malaria and CSF glucose has been found to be lower in patients with fatal cerebral malaria as compared to survivors [<xref ref-type="bibr" rid="B13">13</xref>]. However, the low CSF glucose may be partly due to low plasma glucose levels not measured in this study, but previously found in falciparum malaria [<xref ref-type="bibr" rid="B14">14</xref>]. The previous study found CSF glucose levels of up to 7 mmol/l in patients with cerebral malaria and mean levels of 4.3 mmol/l in survivors, which indicates that a discriminator of 3.4 mmol/l between cerebral malaria and encephalitis may not be universally applicable, but dependent on disease severity [<xref ref-type="bibr" rid="B13">13</xref>]. Febrile convulsions were an indication for lumbar puncture in the control patients. Consequently the control patients in our study were significantly younger than the disease cases, due to the young age at which febrile convulsions usually occur. CSF glucose levels are independent of age above 2 months of age with lower levels found in infants below this age [<xref ref-type="bibr" rid="B15">15</xref>]. All the children in our study were older than 3 months and therefore the observed differences were not age related. Elevated CSF glucose has been described in cases of viral encephalitis [<xref ref-type="bibr" rid="B16">16</xref>] and this may have contributed to the significant difference found here between the two groups.</p></sec><sec><title>CSF-protein</title><p>Previous studies with data on CSF in cerebral malaria also found increased total protein levels [<xref ref-type="bibr" rid="B17">17</xref>,<xref ref-type="bibr" rid="B18">18</xref>]. Our result of increased cerebrospinal fluid protein levels in presumed viral encephalitis has also been found in patients with herpes encephalitis [<xref ref-type="bibr" rid="B19">19</xref>,<xref ref-type="bibr" rid="B20">20</xref>].</p></sec><sec><title>Adenosine deaminase</title><p>We found that a CSF/serum ADA ratio of <0.38 was the best discriminator of cerebral malaria from presumed viral encephalitis. CSF ADA levels were measured in 3 patients with cerebral malaria previously, and found to have a mean (SD) of 6.6 (1.03) IU/l [<xref ref-type="bibr" rid="B5">5</xref>], slightly higher than the levels in our study (4.6 (0.9) IU/l). There has been no previous investigation of serum ADA levels in <italic>P. falciparum </italic>malaria but serum ADA levels were found to be elevated to more than twice the level of controls in a previous study on patients with <italic>Plasmodium vivax </italic>malaria [<xref ref-type="bibr" rid="B21">21</xref>]. CSF ADA levels in a previous study containing data on 10 patients with viral encephalitis were found to have a mean (SD) of 6.15 (2.93) [<xref ref-type="bibr" rid="B6">6</xref>], which is higher than the levels reported here (3.8 (1.0) IU/l). Although the etiological agent of the presumed viral encephalitis was not determined, it was most likely to be a heterogeneous group over the 2-year period during which the patients were recruited [<xref ref-type="bibr" rid="B22">22</xref>].</p></sec><sec><title>Lactic dehydrogenase</title><p>Mean CSF LDH levels of 26.5 U/ml in our patients with encephalitis were similar to the means of 22.6 and 22.3 U/ml found in two previous studies, respectively, [<xref ref-type="bibr" rid="B23">23</xref>,<xref ref-type="bibr" rid="B24">24</xref>]. Our controls had a slightly lower mean CSF LDH level than the 20 controls (adults and children without neurological illness) of a previous study with a mean level of 21.8 U/ml [<xref ref-type="bibr" rid="B24">24</xref>]. There is to our knowledge no previous study on CSF LDH levels in cerebral malaria.</p><p>CSF lactate, which is characteristically elevated in cerebral malaria [<xref ref-type="bibr" rid="B13">13</xref>], is another parameter for future studies looking at discriminating features between cerebral malaria and encephalitis.</p><p>Patients with clinical features of encephalitis and a positive malaria blood slide should still be treated as suffering from cerebral malaria regardless of the CSF findings, as our study cannot exclude a possible overlap in CSF/serum ADA ratios or CSF glucose levels between patients with encephalitis and cerebral malaria in other settings. The same applies to epidemiological surveillance pending further studies.</p></sec></sec><sec><title>Conclusion</title><p>CSF adenosine deaminase alone was not a useful discriminator between encephalitis and cerebral malaria. However, the CSF/serum ADA ratio was lower in patients with cerebral malaria due to the high levels of ADA in peripheral blood. This ratio should be evaluated in future, larger-scale, investigations into discriminatory factors between encephalitis and cerebral malaria. Although CSF glucose was significantly reduced in malaria patients, the results should be interpreted cautiously and require confirmation in a larger study. It would be preferable to use CSF/serum glucose ratios for future studies. Large-scale prospective studies in areas where <italic>P. falciparum </italic>and for example, Japanese encephalitis viruses, are both endemic are needed to validate our findings in different settings.</p></sec><sec><title>Competing interests</title><p>The author(s) declare that they have no competing interests.</p></sec><sec><title>Authors' contributions</title><p>SRJ participated in the design of the study, recruitment of patients, collection of samples and data-analysis. SV participated in the laboratory analysis and data analysis. RMA participated in design of the study. ME participated in data-analysis, interpreted the data and wrote the paper.</p></sec> |
A qualitative study of an integrated maternity, drugs and social care service for drug-using women | <sec><title>Background</title><p>The care of drug-using pregnant women is a growing health and social care concern in many countries. A specialist clinic was established offering multidisciplinary care and advice to pregnant drug users in and around Aberdeen (UK) in 1997. The majority of women stabilise and reduce their drug use. By determining the needs and views of the women more appropriate services and prevention strategies may be developed. There has been little research conducted in this area and none in Scotland.</p></sec><sec sec-type="methods"><title>Methods</title><p>This is a qualitative study that aimed to gain an understanding of the experiences of women drug users, seeking and receiving prenatal care and drug services from a specialist clinic. Twelve women participated in semi-structured one-to-one interviews.</p></sec><sec><title>Results</title><p>The women preferred the multidisciplinary clinic (one-stop shop) to traditional prenatal care centred within General Practice. The relationships of the clients to the range of Clinic professionals and in hospital were explored as well as attitudes to Clinic care. The study participants attributed success in reducing their drug use to the combination of different aspects of care of the multi-agency clinic, especially the high level prenatal support. It is this arrangement of all aspects of care together that seem to produce better outcomes for mother and child than single care elements delivered separately. Some women reported that their pregnancy encouraged them to rapidly detoxify due to the guilt experienced. The most important aspects of the Clinic care were found to be non-judgemental attitude of staff, consistent staff, high level of support, reliable information and multi-agency integrated care.</p></sec><sec><title>Conclusion</title><p>There is an impetus for women drug users to change lifestyle during pregnancy. The study highlighted a need for women to have access to reliable information on the effects of drugs on the baby.</p><p>Further research is required to determine whether positive outcomes related to clinic attendance in the prenatal period are sustained in the postnatal period. Early referral to a specialist clinic is of benefit to the women, as they reported to receive more appropriate care, especially in relation to their drug use. A greater awareness of needs of the pregnant drug user could help the design of more effective prevention strategies.</p></sec> | <contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Hall</surname><given-names>Jennifer L</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>jennifer.hall@nhs.net</email></contrib><contrib id="A2" contrib-type="author"><name><surname>van Teijlingen</surname><given-names>Edwin R</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>van.teijlingen@abdn.ac.uk</email></contrib> | BMC Pregnancy and Childbirth | <sec><title>Background</title><p>The number of new drug users in contact with services in Scotland [<xref ref-type="bibr" rid="B1">1</xref>] is increasing and over the last five years has risen from 9,128 to 11,472 (total population of Scotland is just over 5 million). The majority of females reported to the national database are of childbearing age. Opiates/opioids, especially, heroin, are the main illegal drugs in the UK [<xref ref-type="bibr" rid="B2">2</xref>], and heroin is also the main drug of pregnant drug users. Many pregnant drug users stabilise on the synthetic heroin substitute, methadone. Polydrug use is common. Other drugs commonly used by this group are (crack) cocaine, amphetamines, benzodiazepines and cannabis [<xref ref-type="bibr" rid="B3">3</xref>].</p><p>In Scotland, the rate per 1000 discharges of maternities recording drug use has risen from 3.6 in 1998/9 to 6.7 in 2002/03 [<xref ref-type="bibr" rid="B1">1</xref>]. The number of babies withdrawing from opioids in Aberdeen Maternity Hospital was 1 or 2 per year in the 1980s, rising to 8 in 1995 and 63 in 2000 [<xref ref-type="bibr" rid="B3">3</xref>]. In May 2005 a local newspaper reported that the "number of pregnant women with drug problems giving birth at the hospital has soared ... to nearly 100" [<xref ref-type="bibr" rid="B4">4</xref>].</p><p>The foetal effects of opiate use include premature birth, low birth weight, incidence of sudden infant death syndrome (SIDS or cot death), small head circumference, and neonatal withdrawal symptoms [<xref ref-type="bibr" rid="B5">5</xref>-<xref ref-type="bibr" rid="B9">9</xref>]. The association of drug use with socio-economic deprivation prevents any single symptom in the foetus being attributed to drug use. The only specific effect of maternal drug use on the foetus is neonatal withdrawal, or neonatal abstinence syndrome (NAS). [<xref ref-type="bibr" rid="B8">8</xref>]</p><p>The stigma of drug use is traditionally greater for women than for men [<xref ref-type="bibr" rid="B10">10</xref>], and is even more pronounced for pregnant drug users. They may suffer anxiety and guilt over the effect of the drugs on the baby, and fear of losing the child into care by Social Work (SW) [<xref ref-type="bibr" rid="B11">11</xref>]. In the UK, SW have a legal obligation towards protection of children, this has led to children being taken into care against the will of the parents. This obligation has given SW a 'bad' name in certain parts of society, especially among drug users. It is stipulated that drug use alone should not be the sole reason for separating a mother from her child [<xref ref-type="bibr" rid="B12">12</xref>]. Women may avoid facing reality and avoid contact with all services, especially prenatal services [<xref ref-type="bibr" rid="B13">13</xref>]. The importance of providing appropriate services to meet the needs of pregnant drug users is acknowledged. [<xref ref-type="bibr" rid="B14">14</xref>]</p><p>There is evidence that drug-using women attend antenatal care late and/or conceal their drugs use to maternity care providers [<xref ref-type="bibr" rid="B6">6</xref>,<xref ref-type="bibr" rid="B15">15</xref>]. To be effective, prenatal care should be appropriate to women's needs, easily accessible and the woman should be involved in the planning of care [<xref ref-type="bibr" rid="B16">16</xref>,<xref ref-type="bibr" rid="B17">17</xref>]. A higher frequency of prenatal care and social support for this group has been associated with improved outcomes in terms of infant birth weight [<xref ref-type="bibr" rid="B18">18</xref>,<xref ref-type="bibr" rid="B19">19</xref>].</p><p>Women should not be deterred from seeking prenatal care [<xref ref-type="bibr" rid="B20">20</xref>]. It has been recognised that providing special services to attract and retain pregnant women drug users into prenatal care can help to address their complex problems [<xref ref-type="bibr" rid="B2">2</xref>]. Multidisciplinary services can address drug use at the same time as providing prenatal care [<xref ref-type="bibr" rid="B21">21</xref>]. A non-judgemental approach has been advocated to encourage women into services [<xref ref-type="bibr" rid="B8">8</xref>].</p><p>Specialised services for pregnant drug users were established in the UK from the late 1980s onwards. In Scotland, only two such services exist, in Glasgow and Aberdeen. Since 1990, the Glasgow Women's Reproductive Service has provided a service for women with any type of social problem. Half of all its clients are drug users [<xref ref-type="bibr" rid="B22">22</xref>].</p><p>The Golden Square Special Clinic in Aberdeen (henceforth referred to as the Aberdeen Clinic) opened in September 1997 within a family planning clinic in the city centre. The Aberdeen Clinic is community-based multi-disciplinary agency, and it also offers a greater frequency of prenatal care than available to the average pregnant woman in the area. The multi-disciplinary team consists of an obstetrician; a community midwife; a community psychiatric nurse from the statutory drugs service called the Substance Misuse Service (SMS); a social worker and a drugs worker from Drugs Action (DA), a voluntary drugs service, which provides counselling and social support. The local hospital offers detoxification and designated space is always available [<xref ref-type="bibr" rid="B3">3</xref>]. However, only a minority of drugs misusing pregnant women use this facility. The overwhelming majority of the clients attending Aberdeen Clinic were prescribed substitute drugs, mainly methadone to stabilise their drug use.</p><p>Identification of pregnant drug users by the services and stabilisation of their drug habit before labour can simplify their clinical management [<xref ref-type="bibr" rid="B23">23</xref>]. The desire to have a healthy baby can give an impetus to change drug use [<xref ref-type="bibr" rid="B15">15</xref>] and pregnancy may be regarded as a window of opportunity for the services as well as for the client, with possible long-term benefits [<xref ref-type="bibr" rid="B24">24</xref>]. This is especially relevant when child protection issues are considered [<xref ref-type="bibr" rid="B25">25</xref>]. A mother is more likely to retain contact with the services after the birth if a good relationship with a service has been established before the birth with positive implications for the future well-being of the child [<xref ref-type="bibr" rid="B26">26</xref>].</p><p>Taking account of client views is regarded as one of the core principles of effective practice in the provision of drug misuse services [<xref ref-type="bibr" rid="B27">27</xref>]. User-friendly services may be developed by determining women's views [<xref ref-type="bibr" rid="B28">28</xref>].</p><p>Much research has covered the effect of maternal drug misuse on the baby, but there has been little work exploring the experiences of pregnant drug users and attempting to understand their needs [<xref ref-type="bibr" rid="B11">11</xref>,<xref ref-type="bibr" rid="B15">15</xref>]. Hepburn and Elliott [<xref ref-type="bibr" rid="B22">22</xref>] showed that staff attitudes were more important to clients than medical aspects of care; in all cases pregnancy provided a strong motivation for change. Klee and Jackson [<xref ref-type="bibr" rid="B10">10</xref>] studied a cohort of 51 female drug users from pregnancy through childbirth and into motherhood. Most believed pregnancy was a good time to reduce their drug use, but many had reverted to street drugs or increased their methadone dosage by the end of the study. Overall, satisfaction with treatment was highest if it was an integrated service and especially if a specialist person, such as a drugs-liaison midwife, was involved.</p><p>This study explores the experiences of women drug users who have received their prenatal care from a specialised multi-agency clinic.</p></sec><sec sec-type="methods"><title>Methods</title><p>A short questionnaire determining background details was administered before a semi-structured taped interview. The questionnaire was based on previous reported studies with drug users [<xref ref-type="bibr" rid="B10">10</xref>,<xref ref-type="bibr" rid="B15">15</xref>]. Administering the questionnaire before the interview allowed a relationship to develop between the researcher and the interviewee thus increasing the likelihood of responses being elicited to further qualitative questions [<xref ref-type="bibr" rid="B29">29</xref>]. The quantitative method was employed in this way to maximise the potential of the qualitative method [<xref ref-type="bibr" rid="B30">30</xref>].</p><sec><title>Sample</title><p>Women, who were pregnant or who had had a baby in the past three years, and who were recent or current drug users, and who had had contact with the Aberdeen Clinic were sought for a research interview (this constitutes approximately 40 potential interviewees [<xref ref-type="bibr" rid="B3">3</xref>]). Due to UK ethics regulations we could not approach women directly, therefore recruitment was facilitated by service providers, who offered all eligible women a leaflet about the study. The following service providers helped in the recruitment process: the Aberdeen Clinic; the local voluntary drugs service (DA); and Social Work (SW). We asked these services to approach both women currently pregnant and using the Aberdeen Clinic as well as those women who were postpartum and recently used the service. Full ethical approval was granted by the regional ethics committee (Grampian Research Ethics Committee) and confidentiality and anonymity has been maintained.</p></sec><sec><title>Semi-structured interview</title><p>The interviews were conducted as informally as possible; in the interviewee's home or at the clinic or service they were attending [<xref ref-type="bibr" rid="B31">31</xref>]. The interview schedule (Table <xref ref-type="table" rid="T1">1</xref>) was developed such that the order could be varied if required [<xref ref-type="bibr" rid="B32">32</xref>].</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>Interview schedule</p></caption><table frame="hsides" rules="groups"><tbody><tr><td align="left">• Experiences of seeking care – route of referral to the Clinic; feelings; ease of access;</td></tr><tr><td align="left">• Range of care received</td></tr><tr><td align="left">• Experiences of receiving care – feelings about the care; care at the Clinic compared to other settings: positive and negative experiences;</td></tr><tr><td align="left">• Relationship with health care professionals – feelings about the staff; consistency of staff; relationships with staff from other services; type of relationship; hospital relationships;</td></tr><tr><td align="left">• Drug use in pregnancy – feelings about drug use during pregnancy; change of drug use; reasons for change;</td></tr><tr><td align="left">• Perceptions of available service options and suggestions</td></tr></tbody></table></table-wrap></sec><sec><title>Data analysis</title><p>Interviews were tape-recorded with permission and transcribed, with field notes written directly after each interview. Key points in the transcripts were categorised using Content Analysis [<xref ref-type="bibr" rid="B33">33</xref>,<xref ref-type="bibr" rid="B34">34</xref>]. The transcripts were independently analysed by two researchers to ensure inter-rater reliability [<xref ref-type="bibr" rid="B35">35</xref>].</p><p>Care was taken to ensure that all selected quotations, following a particular statement of text, are from different individuals. This policy was adopted instead of noting the origin of each quotation in the text, as, due to the small size of the sample there is the potential of an individual interviewee being identified [<xref ref-type="bibr" rid="B36">36</xref>].</p></sec></sec><sec><title>Results</title><sec><title>Quantitative questionnaire</title><p>Twelve women participated who had all received care from the Aberdeen Clinic: four interviewees were pregnant at the time and all were aged between 19 and 36. Six had previous children and five had children who were adopted or being cared for by another family member (Table <xref ref-type="table" rid="T2">2</xref>).</p><table-wrap position="float" id="T2"><label>Table 2</label><caption><p>Characteristics of interviewees</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="center"><bold>Interviewee</bold></td><td align="center"><bold>status: weeks/gestation (g) or postnatal (p)</bold></td><td align="center"><bold>Age (yrs)</bold></td><td align="center"><bold>use of problem drug (yrs)</bold></td><td align="center"><bold>children (number & age)</bold></td></tr></thead><tbody><tr><td align="center">1</td><td align="center">23 (g)</td><td align="center">29</td><td align="center">5</td><td align="center">2 : 11 yr, 5 yr</td></tr><tr><td align="center">2</td><td align="center">12 (p)</td><td align="center">27</td><td align="center">5</td><td align="center">2 : 9 yr, 3 mth</td></tr><tr><td align="center">3</td><td align="center">11 (p)</td><td align="center">19</td><td align="center">3</td><td align="center">1 : 11 wks</td></tr><tr><td align="center">4</td><td align="center">10 (p)</td><td align="center">23</td><td align="center">8</td><td align="center">1 : 10 wks</td></tr><tr><td align="center">5</td><td align="center">20 (p)</td><td align="center">19</td><td align="center">1.5</td><td align="center">1 : 10 wks</td></tr><tr><td align="center">6</td><td align="center">10 (p)</td><td align="center">22</td><td align="center">5</td><td align="center">1 : 10 wks</td></tr><tr><td align="center">7</td><td align="center">52 (p)</td><td align="center">22</td><td align="center">4</td><td align="center">2 : 6 yr, 1 yr</td></tr><tr><td align="center">8</td><td align="center">28 (p)</td><td align="center">26</td><td align="center">12</td><td align="center">1 : 7 mths</td></tr><tr><td align="center">9</td><td align="center">35 (g)</td><td align="center">36</td><td align="center">2</td><td align="center">0</td></tr><tr><td align="center">10</td><td align="center">8 (g)</td><td align="center">26</td><td align="center">5</td><td align="center">2 : 10 yr, 4 yr</td></tr><tr><td align="center">11</td><td align="center">>100 (p)</td><td align="center">31</td><td align="center">6</td><td align="center">2 : 8 yr, 2 yr</td></tr><tr><td align="center">12</td><td align="center">28 (g)</td><td align="center">30</td><td align="center">7</td><td align="center">2 : 13 yr, 8 yr</td></tr></tbody></table></table-wrap><p>Most had a long-term (one year or more) partner at the time of the interview, who was the father of the recent pregnancy. Nearly all interviewees had previously been daily intravenous heroin users and several had also regularly used crack cocaine. All were on a methadone prescription at the time of the interview (variation of 6 – 90 mls daily). Most were stable and successfully reducing but two reported to be also using illicit drugs.</p><p>Some pregnancies had not been identified until approximately 22 weeks gestation. The possibility of receiving early prenatal care and support had been reduced in some due to late identification of the pregnancy. The first visit to the Clinic ranged from being the same week up to 12 weeks after first attending General Practice for prenatal care (Table <xref ref-type="table" rid="T3">3</xref>).</p><table-wrap position="float" id="T3"><label>Table 3</label><caption><p>Prenatal care of interviewees</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="center"><bold>Interviewee</bold></td><td align="center"><bold>Identification of pregnancy (wks gestation)***</bold></td><td align="center"><bold>1st prenatal appointment (wks gestation)</bold></td><td align="center"><bold>1st Aberdeen Clinic appointment (wks gestation)</bold></td></tr></thead><tbody><tr><td align="center">1</td><td align="center">8</td><td align="center">8</td><td align="center">15</td></tr><tr><td align="center">2</td><td align="center">8</td><td align="center">8</td><td align="center">20</td></tr><tr><td align="center">3</td><td align="center">9</td><td align="center">10</td><td align="center">12</td></tr><tr><td align="center">4</td><td align="center">6</td><td align="center">12</td><td align="center">12</td></tr><tr><td align="center">5</td><td align="center">7</td><td align="center">10</td><td align="center">10</td></tr><tr><td align="center">6</td><td align="center">12</td><td align="center">12</td><td align="center">24</td></tr><tr><td align="center">7</td><td align="center">20</td><td align="center">20</td><td align="center">22</td></tr><tr><td align="center">8</td><td align="center">12</td><td align="center">28*</td><td align="center">none</td></tr><tr><td align="center">9</td><td align="center">8</td><td align="center">10</td><td align="center">14</td></tr><tr><td align="center">10</td><td align="center">6</td><td align="center">6</td><td align="center">referred**</td></tr><tr><td align="center">11</td><td align="center">22</td><td align="center">22</td><td align="center">22</td></tr><tr><td align="center">12</td><td align="center">8</td><td align="center">8</td><td align="center">10</td></tr></tbody></table><table-wrap-foot><p>* admitted to maternity hospital by Aberdeen Clinic staff</p><p>** awaiting appointment at Aberdeen Clinic at time of interview</p><p>*** refers to the moment in the pregnancy the woman realised that she was pregnant.</p></table-wrap-foot></table-wrap></sec><sec><title>Qualitative interviews</title><sec><title>Referral to the special Clinic</title><p>In several cases, referral to the Clinic had been delayed since the family physician/General Practitioner (GP) or the community midwife had either not suggested the Clinic or had not provided appropriate information about the service. Consequently, women did not realise that a high level of multi-agency support was available at the Clinic. Several women only attended the Clinic after they had been admitted to hospital. The midwives in hospital had then referred them:</p><p>"I think it was anger. I was five and a half months pregnant. My doctor didn't tell me soon enough really, for me to get right off my methadone."</p><p>And</p><p><italic>"I think he </italic>(GP) <italic>could have.... it was when I came here (</italic>hospital) <italic>that they explained it would be better (</italic>to go to the Clinic)."</p><p>For some there was a lack of explanation and/or choice:</p><p><italic>"He </italic>(GP) <italic>basically said I had to go there for my treatment."</italic></p><p>This resulted in a feeling of rejection for one woman:</p><p>" <italic>It was just like we don't want you up here kind of thing ... passed off because you are an ex junkie."</italic></p><p>And</p><p>"I just felt a bit pushed off; I was told they didn't have the services for me .... and I thought well, what do you need to deal with me?"</p></sec><sec><title>Barriers to accessing care</title><p>Most were very wary of going to the Clinic and feared adverse judgements:</p><p>"I wasn't very sure ...I thought maybe I would be labelled a junkie. You know you wouldn't be treated the same... at the hospital you are quite anonymous."</p><p>And</p><p>"I didn't really want to go along. I thought people would look at you and judge you because you were using heroin."</p><p>Some avoided the Clinic because they were scared of SW knowing that SW had the power to take a baby away from the mother and into care:</p><p>"<italic>I was just too frightened, too scared, thinking they were going to get me into trouble and take her off me the minute she was born."</italic></p><p>Some feared having to go back on methadone either because of its side effects or they felt it was easier to withdraw from heroin, e.g.:</p><p>"<italic>I was back using heroin but very little, but I thought it was better using very little than going back on methadone"</italic></p><p>One woman had encountered a very negative attitude from the community midwife and this made her apprehensive to attend the Clinic:</p><p>"<italic>You feel really bad because after the first midwife I wasn't too sure what the midwives would be like and whether you could trust them..."</italic></p></sec><sec><title>Receiving prenatal care – special Clinic</title><p>The overwhelming majority of the women were reassured by the Clinic, they felt in "safe hands" and often expressed more confidence of the treatment there than at their family physician. The Clinic resolved problems where the family physician had been unable to help. Several expressed real enjoyment at going to the Clinic:</p><p>"Sometimes I couldn't wait to go because it is like a really familyish atmosphere, they make you feel welcome..."</p></sec><sec><title>Care provision</title><p>Some aspects of the service were important to all interviewees: (a) attitude of staff; (b) consistency of staff; (c) high level of support; (d) reliable information; and (e) integrated care from different services (Table <xref ref-type="table" rid="T4">4</xref>). All those interviewed postnatally were positive about the support received and attributed their reduction in drug use to it.</p><table-wrap position="float" id="T4"><label>Table 4</label><caption><p>Themes related to care emerging from interviews</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left"><bold>Theme</bold></td><td align="left"><bold>Example</bold></td></tr></thead><tbody><tr><td align="left">Attitude of staff</td><td align="left"><italic>"They didn't make you feel like it was something to be ashamed of. They made you feel worth something that you were here, at least you are trying you know..."</italic><break/><italic>"They had time for me...they wouldn't shove me out the door, they waited till I had finished"</italic></td></tr><tr><td colspan="2"><hr></hr></td></tr><tr><td align="left">Continuity of carer</td><td align="left"><italic>"... having the same people because they know you and you know them...you don't have to keep repeating yourself, telling your story over and over again."</italic></td></tr><tr><td colspan="2"><hr></hr></td></tr><tr><td align="left">High level of support</td><td align="left"><italic>"I think it's important to drug users to go more often. I mean if you are going to relapse then there is somebody there to talk to."</italic></td></tr><tr><td colspan="2"><hr></hr></td></tr><tr><td align="left">Information</td><td align="left"><italic>"They told you all about the drugs. I was quite far on and I didn't know that crack cocaine can kill the baby and I was a regular user. That's extremely important. They don't tell you that anywhere else"</italic></td></tr><tr><td colspan="2"><hr></hr></td></tr><tr><td align="left">Integrated care</td><td align="left"><italic>"They all interlink with each other, you know....that was really, really handy...it saves you having to travel....it's all in one building, so it was really good."</italic><break/><italic>"... you don't want to have to keep telling people that you are a user, you only want to say it once for everybody to know."</italic></td></tr></tbody></table></table-wrap><p>"I couldn't see myself stopping without the support, it would have been a disaster."</p><p>And</p><p>"I think that the Clinic is brilliant and I know a lot of other people that wouldn't have gotten down so far as they have gotten down without the Clinic."</p><p>Also</p><p>"It was excellent, I don't think I would be clean now without it."</p></sec><sec><title>Range of care at the special Clinic</title><p>The women were asked about the different services they had accessed at the Clinic. Everybody attending the Clinic sees a midwife and a member of staff from the SMS. The use of DA and SW is optional although a woman may be strongly encouraged to talk to the social worker if there are child protection concerns for the baby.</p><sec><title>(a) Social work</title><p>Some women did not want contact with SW because of the stigma involved. Several tried to avoid contact with SW due to previous bad experiences:</p><p>" there can be a stigma with them, you know what I mean."</p><p>And</p><p>"I didn't trust them as a result of that (previous incident). They said it was confidential and turned out it wasn't at all, just made things worse for me."</p><p>And</p><p>"I didn't want a social worker because I have had them before and I dinnae (=didn't) like their twisting ways they have got. They will tell you one thing but mean another, I dinnae (=didn't) like that. I like somebody to be honest."</p><p>Others had been initially reluctant to have contact with SW, but acknowledged that SW had been helpful:</p><p>"I was angry for a while, they (SW) sneaked into my life .... but I didn't realise how supportive she would be...I think, now, looking back, I think it was probably the best thing that could have happened because it made me waken up, opened my eyes. I thought this is serious..."</p><p>And:</p><p>"Excellent help getting a house, she's been excellent really."</p></sec><sec><title>(b) Voluntary Drugs Agency</title><p>Not all of the women had accessed DA. Whilst several relied heavily on the drug worker and valued being able to talk about all their problems, not just those related to the pregnancy. Some felt it was more confidential than with the midwife:</p><p>"She doesn't just speak about my drug use, she speaks about everything ...she's really good."</p><p>One woman had avoided contact, knowing that the worker was able to supply needles:</p><p>"I was scared to talk to her in case I asked her for needles...I just thought if I go speaking to her I might just start asking for needles and then it would make it easier (to relapse)."</p></sec></sec><sec><title>Receiving prenatal care: hospital</title><p>The majority appreciated their hospital care and found the staff friendly and helpful:</p><p>"... in hospital there was somebody with me all the time. It's a lot easier to stay off drugs if you're in hospital...the staff were absolutely amazing."</p><p>Also:</p><p>"They looked after me properly...they took my dinner to me because I couldna (=could not) walk..."</p><p>Several appreciated the staff's discreteness regarding their drug use:</p><p>"They didn't let anybody else in the ward know why you were there and that you were a drug user...they were discrete..."</p><p>Several appreciated the safe environment:</p><p><italic>"You felt sort of safe....you were looked after, it was really nice"</italic>,</p><p>An environment that also gave a shelter from home life:</p><p>"They took me in for a couple of weeks just to sort of get away from everybody...which worked...I haven't used since then..."</p><p>After discharge, the women were able to contact the hospital at any time, if they felt in danger of relapse. Many found this reassuring:</p><p>"Any time you feel you are going to use you can always phone them...I haven't actually called them but it's always in my mind that I could."</p></sec><sec><title>Relationships with professionals</title><p>The nature of the relationship between the health care professionals and the women was very variable and depended on both the professional and the client. However, we found several themes.</p></sec><sec><title>Special clinic</title><p>The relationships between the women and the Clinic staff, the midwife and the SMS staff, were reported as very good and friendly. All felt respected and genuinely cared for:</p><p>"They're very professional and very friendly."</p><p>And</p><p>"I felt like I knew them you know.... they were like a little Mother Hubbard, ready to sort you out."</p><p>Staff was regarded as supportive even when they saw clients away from the clinic:</p><p>"She was really supportive and cuddles and everything. Not like a professional but like a friend, I thought."</p><p>And</p><p>"They weren't just there for you, they went to see the babies afterwards."</p><p>It was reported that the Clinic staff were quite forceful on occasions but this was appreciated, for example:</p><p>"They were basically giving me an ultimatum and I had a choice that I had to make and the doctor can't really do that..."</p><p>And:</p><p><italic>"The doctor was very understanding, maybe too understanding...it feels good (</italic>now, at the Clinic <italic>) to be pushed in that direction."</italic></p></sec><sec><title>Hospital ward</title><p>There were positive feelings expressed towards most of the hospital staff, but not to the same extent as at the Clinic. Some were very positive:</p><p>"They're very friendly, unbelievably friendly. Non-judgemental and to my partner..."</p><p>And</p><p>"Oh, they were brilliant ... Some of them was a bit ...stand off, aye, probably because I was on drugs and covered in abscesses ...but some of them was really good."</p><p>However, several women trusted and liked them less than the Clinic staff:</p><p>"I'm more scared of their reaction on the ward."</p><p>And</p><p>"Some of them just, they looked at you like shite (=shit)...they just didna (=didn't) want to know you sort of thing, because you're a junkie or whatever."</p></sec><sec><title>Social work</title><p>Generally, women were wary of having any contact with SW but there were examples of valued long-standing relationships. In some cases there was a heavy reliance on the social worker:</p><p>"Well I would have probably lost my baby, if it hadn't been for her. She really sort of stuck her neck out for me, because nobody really wanted to give me a chance..."</p></sec><sec><title>Voluntary Drugs Agency</title><p>A few had long-term relationships with a DA worker before falling pregnant. One relationship was obviously close:</p><p>"I've been seeing her a long time now, about three years. I like her, yeah. She doesn't just speak about my drug use, she speaks about everything. Any sort of pressures in your life can make you go back using again, she's very aware of that.... she's really good."</p><p>One woman had greater trust in her DA worker than her family physician. She had immediately sought advice from her DA worker when she realised about the pregnancy:</p><p>"The first thing I did was to phone DA, to tell her I was pregnant...and find out about (effect of) the heroin."</p></sec><sec><title>Confidentiality</title><p>Women are very sensitive to their situation and needed confidentiality. They did not want everyone to know about their drug use and were anxious to hide details about any case conference, a multidisciplinary hearing to further investigate the nature of the drug use and implications for the child. Sometimes the breach of confidentiality had demoralising consequences for the woman drug user:</p><p>"Well the woman in the bed next to me, she seemed really nice then she never really spoke to me after the social worker came in and said something about the case conference"</p><p>And</p><p>"I don't think the social workers give you much privacy. They just came in and started speaking about my case conference in front of like the rest of the ward...."</p></sec><sec><title>Attitudes</title><p>Several themes emerged regarding attitudes to drug use and pregnancy: (a) great anxiety and guilt; and (b) a strong desire to undergo detoxification immediately after discovering the pregnancy.</p></sec><sec><title>Reaction to pregnancy</title><p>The women had experienced great anxiety, guilt and panic:</p><p>"It was sheer panic, where do we go, what are the effects on the baby."</p><p>And:</p><p>"I was really sort of horrified....I honestly thought that if there was a baby inside me there was no possible way that it could be alive."</p><p>Many wanted to give up drugs immediately:</p><p>"I was using and taking methadone and I thought right, that's it, I've got to stop. But when I stopped using heroin, the methadone wasn't enough and I got very ill."</p><p>Some knew that to stop taking drugs could harm the fetus, but still did so:</p><p>"When I found out I was pregnant, I know I shouldn't have done it but I stopped taking all the tablets I was on..."</p></sec><sec><title>Pregnancy as impetus to change</title><p>The majority regarded pregnancy as an impetus to change their drug use and their lives. Only those who had suffered from sickness and nausea during their pregnancy did not hold this view:</p><p>".... I just thought I felt I couldn't do that, to an unborn child that is so innocent, you're actually giving that baby a habit."</p><p>And:</p><p>"In a way it was a real opportunity, but I wouldn't say that to other young lassies in case it didn't work like that."</p><p>The pregnancy made them see how they had been living before:</p><p>"I think when he was born I realised that it's all or nothing – it's all in the past now – it was all shite basically."</p><p>And:</p><p>"It was when I actually had him that I thought, no, I definitely don't want that life anymore."</p><p>The women expressed great determination not to go back to using drugs:</p><p>"It dawned on me when he was born; I thought all I have to do is not use. The whole future depends on me not anyone else."</p><p>And:</p><p>"I don't want to destroy myself on drugs ever again."</p><p>The baby had provided a real focus for their lives:</p><p>"I know I've wanted to come off drugs for the last four years but there is more of a reason now.... I don't want to go back now."</p><p>And:</p><p>"I was clean for ten months. As soon as I came back up here I went straight back on it. It was because I never had a goal, but now I have."</p><p>The women accepted that it would be more difficult to stay off after the birth, for example:</p><p>"It's more of a scarier issue once I have had the baby, because then I'll be on my own. It'll be just me."</p><p>In some cases the real focus not to use drugs had been imposed by SW:</p><p>"... said they were going to take her into care so that fried my head a bit."</p><p>Coming off drugs and having to change friends was an important issue:</p><p>"I could count my friends on my hand now whereas before I could have named hundreds of people but now I am very, very cautious about who I choose to be friends with."</p></sec><sec><title>Guilt</title><p>To many it was impossible to face up to being pregnant because they felt so guilty about their drug use and the effect on the baby. For several this encouraged their drug use because it acted as a release:</p><p>"I blanked it out.... I didn't want to face up to it...then I got into more drugs to forget it."</p><p>And:</p><p>"I think it was just so much easier to deny. It was just a bump in my tummy, you couldn't actually see the little person."</p></sec><sec><title>Failure to identify pregnancy</title><p>One woman had not identified her pregnancy until week 22:</p><p>"I didn't think that I could fall pregnant while I was on so many drugs. I thought that was impossible."</p></sec><sec><title>Trends in views</title><p>No differences in attitude towards the Clinic or the hospital care could be attributed to the interviewee being either pregnant or postnatal or having positive or negative experiences with the other service providers, DA or SW.</p><p>Those who were supported by a partner or parent in addition to clinic care, were more successful in reducing their drug use.</p></sec></sec></sec><sec><title>Discussion</title><p>Most women benefited from the multi-disciplinary Clinic as a result of prompt referral from their family physicians, however, for some women an earlier referral would have been an advantage. The majority would have benefited from more information about the Aberdeen Clinic before referral.</p><p>All women interviewed preferred the service provided at the Clinic to prenatal care at the GP practice and expressed confidence in the information and treatment provided. Aspects of care which were most highly valued were non-judgemental attitude of staff, reassurance and provision of reliable information, consistency of staff and high level of support in terms of frequency of visit and time given to each client (see Table <xref ref-type="table" rid="T4">4</xref>). The women attributed their more stable lifestyle and reduction in drug use almost entirely to the Clinic. The importance of non-judgemental attitudes and consistent, reliable information is recognised [<xref ref-type="bibr" rid="B21">21</xref>,<xref ref-type="bibr" rid="B37">37</xref>]. The availability of 24-hour support at the hospital was felt to be highly reassuring to clients.</p><p>Relationships with professionals varied but were generally very positive with the Clinic staff. The professionals varied individually but also because they were from different agencies. The results indicated the importance of multi-agency care available in a service as a "one-stop shop" as each client had usually developed a close relationship with only one professional. It is reported that staff's attitudes and relationships with clients are often perceived by clients as more important than the quality of the medical care [<xref ref-type="bibr" rid="B22">22</xref>,<xref ref-type="bibr" rid="B38">38</xref>].</p><p>There was strong appreciation when confidentiality was kept and this facilitated the development of trusting relationships between the professionals and clients. Some breaches of confidentiality were noted in this study. Confidentiality is highly important to this client group [<xref ref-type="bibr" rid="B39">39</xref>].</p><p>Some had tried self-detoxification, not realising the dangers to the baby. However, by contrast, many had tried to block out the pregnancy initially. Their drug use continued and even increased to help to forget everything.</p><p>Nearly all thought the pregnancy was a tremendous impetus to change their drug use and their lives and all were determined to stay off illicit drugs and to achieve coming off methadone postnatally. Pregnancy seems to provide a window of opportunity to reduce the drug use. However, there is a difficulty in sustaining a change after the birth and, in this study, those with support from a mother had fared better. Another study highlighted support from a mother or partner as an indicator of good outcomes [<xref ref-type="bibr" rid="B40">40</xref>]. There is much emphasis now on child protection [<xref ref-type="bibr" rid="B14">14</xref>] and the opportunity to change lifestyle could have important implications for long-term child protection issues.</p><p>This study is open to bias, as the women are self-selected due to the nature of the research ethics constraints. Although most clients of the Aberdeen Clinic were on methadone, it is likely that women not coping with their drug use during their pregnancy were less likely to participate in our study. Future studies would benefit from accessing a larger, and, therefore, more representative sample of drug misusing pregnant women. Also drug use is self-reported, which brings with it the possibility of underreporting of illegal drugs in such studies. However, useful findings for service provision as well as ideas for questions to be addressed in a large-scale survey emerged from the data.</p></sec><sec><title>Conclusion</title><p>The results of the study highlight the importance for staff who are involved with pregnant women drug users to be aware of the need for sensitivity especially in relation to confidentiality. A breach of confidentiality can be demoralising with possible negative effect on the success of the client in reducing their drug use.</p><p>Women drug users commonly experience extreme guilt when they find themselves pregnant and may embark on rapid self-detoxification. There should be more reliable information available for drug-using women on the effects of drug use on the baby as many are unaware or ill informed of the dangers of different types of drugs and the effects of rapid detoxification on the baby.</p><p>Pregnancy presents a significant impetus to change lifestyle for the woman drug user and the high level multidisciplinary support provided was a clear aid to achieving such change. Multi-agency integrated care in the prenatal period has been shown to be preferred by this client group to care by the local family physician. It appears that the combination of different aspects of health and social care of the multi-agency clinic (one-stop shop) was contributory to the success of the clients in reducing drug use and stabilising lifestyles, as reported by the interviewees. It has been recognised that properly coordinated services help avoid providing a confusing array of services and appointments for pregnant drug users [<xref ref-type="bibr" rid="B41">41</xref>]. In order that this multi-agency approach may work successfully, family physicians should be better informed about it, refer early and provide relevant information to their clients.</p><p>Further research is required to establish whether the support provided by the Clinic produced sustained changes in lifestyle. As previously suggested we also need more research into the long-term outcomes for the child [<xref ref-type="bibr" rid="B42">42</xref>].</p><p>The risk to women drug users and their children is an increasing public health problem, which is in line with the increasing use of drugs in society. A greater awareness of the needs and feelings of the pregnant drug user can allow more effective prevention strategies to be implemented.</p></sec><sec><title>Competing interests</title><p>The author(s) declare that they have no competing interests.</p></sec><sec><title>Authors' contributions</title><p>JH carried out the data collection and analysis for this study as part of her M.Sc. in Health Services and Public Health Research. EvT supervised this research project and participated at all stages of the study. Both authors have written several drafts and approved the final manuscript.</p></sec><sec><title>Pre-publication history</title><p>The pre-publication history for this paper can be accessed here:</p><p><ext-link ext-link-type="uri" xlink:href="http://www.biomedcentral.com/1471-2393/6/19/prepub"/></p></sec> |
Under- treatment and under diagnosis of hypertension: a serious problem in the United Arab Emirates | <sec><title>Background</title><p>Hypertension, notably untreated or uncontrolled, is a major risk factor for cardiovascular diseases (CVD) morbidity and mortality. In countries in transition, little is known about the epidemiology of hypertension, and its biochemical correlates. This study was carried out in Al Ain, United Arab Emirates, to characterize self-reported (SR) normotensives and hypertensives in terms of actual hypertension status, demographic variables, CVD risk factors, treatment, and sequalae.</p></sec><sec sec-type="methods"><title>Methods</title><p>A sample, stratified by SR hypertensive status, of 349 SR hypertensives (Mean age ± SD; 50.8 ± 9.2 yrs; Male: 226) and 640 SR normotensives (42.9 ± 9.3 yrs, Male: 444) among nationals and expatriates was used. Hypertensives and normotensive subjects were recruited from various outpatient clinics and government organizations in Al-Ain city, United Arab Emirates (UAE) respectively. Anthropometric and demographic variables were measured by conventional methods.</p></sec><sec><title>Results</title><p>Both under-diagnosis of hypertension (33%) and under-treatment (76%) were common. Characteristics of undiagnosed hypertensives were intermediate between normotensives and SR hypertensives. Under-diagnosis of hypertension was more common among foreigners than among nationals. Risk factors for CVD were more prevalent among SR hypertensives. Obesity, lack of exercise and smoking were found as major risk factors for CVD among hypertensives in this population.</p></sec><sec><title>Conclusion</title><p>Hypertension, even severe, is commonly under-diagnosed and under-treated in the UAE. Preventive strategies, better diagnosis and proper treatment compliance should be emphasized to reduce incidence of CVD in this population.</p></sec> | <contrib id="A1" equal-contrib="yes" contrib-type="author"><name><surname>Abdulle</surname><given-names>Abdishakur M</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>shakur.a@uaeu.ac.ae</email></contrib><contrib id="A2" equal-contrib="yes" corresp="yes" contrib-type="author"><name><surname>Nagelkerke</surname><given-names>Nico JD</given-names></name><xref ref-type="aff" rid="I2">2</xref><xref ref-type="aff" rid="I3">3</xref><email>nico.nagelkerke@uaeu.ac.ae</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Abouchacra</surname><given-names>Samra</given-names></name><xref ref-type="aff" rid="I4">4</xref><email>sabouchacra@tawam-hosp.gov.ae</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Pathan</surname><given-names>Javed Y</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>javed.pathan@uaeu.ac.ae</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Adem</surname><given-names>Abdu</given-names></name><xref ref-type="aff" rid="I5">5</xref><email>abdu.adem@uaeu.ac.ae</email></contrib><contrib id="A6" contrib-type="author"><name><surname>Obineche</surname><given-names>Enyioma N</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>eobineche@uaeu.ac.ae</email></contrib> | BMC Cardiovascular Disorders | <sec><title>Background</title><p>Hypertension, notably untreated or poorly controlled, is a major risk factor for cardiovascular diseases (CVD) in almost every population [<xref ref-type="bibr" rid="B1">1</xref>-<xref ref-type="bibr" rid="B4">4</xref>]. Although, often controllable by pharmacotherapy [<xref ref-type="bibr" rid="B5">5</xref>,<xref ref-type="bibr" rid="B6">6</xref>], many patients are unaware of their condition, and thus are untreated, both in Western societies such as the USA [<xref ref-type="bibr" rid="B2">2</xref>,<xref ref-type="bibr" rid="B7">7</xref>], but particularly among populations in transition [<xref ref-type="bibr" rid="B8">8</xref>-<xref ref-type="bibr" rid="B10">10</xref>]. Also, many individuals who are aware of their condition are not adequately treated [<xref ref-type="bibr" rid="B11">11</xref>]. Many factors contribute to these problems including poor treatment compliance, lack of access to health care and lack of physician adherence to therapeutic guidelines [<xref ref-type="bibr" rid="B2">2</xref>,<xref ref-type="bibr" rid="B3">3</xref>,<xref ref-type="bibr" rid="B5">5</xref>]. The high prevalence of undiagnosed hypertension is emphasized by the large number of cases discovered incidentally e.g. during surveys, or when patients are under going treatment for other diseases, especially in developing countries [<xref ref-type="bibr" rid="B12">12</xref>,<xref ref-type="bibr" rid="B13">13</xref>].</p><p>It appears that the average blood pressure (BP) level and its age-related increase differs across populations and may explain much of the observed disparities in stroke mortality [<xref ref-type="bibr" rid="B14">14</xref>]. To date the bulk of the available knowledge about hypertension comes from people of Caucasian origin in developed countries [<xref ref-type="bibr" rid="B15">15</xref>]. Little information is available for populations in transition, some of which may have a particularly high genetic risk for the development of diseases of affluence, e.g. the "thrifty gene" hypothesis for diabetes [<xref ref-type="bibr" rid="B16">16</xref>,<xref ref-type="bibr" rid="B17">17</xref>].</p><p>A specific example of a nation in transition is the United Arab Emirates (UAE). As late as the 1960s the population consisted largely of nomadic Bedouin Arabs. However, the discovery of oil in 1970s has changed their life style dramatically and the UAE is now a modern, wealthy society, heavily influenced by Western living patterns, including a sedentary life style with high CVD risk profiles [<xref ref-type="bibr" rid="B18">18</xref>]. Indeed, CVDs are known to be the leading causes of morbidity and mortality in the UAE among both the nationals and the expatriates, who constitute the majority (approximately 80%) of the population [<xref ref-type="bibr" rid="B19">19</xref>]. Of particular concern is the prevalence of obesity, which reaches 35% among young adolescent Bedouin Arab women [<xref ref-type="bibr" rid="B20">20</xref>]; and about 24% among medical students [<xref ref-type="bibr" rid="B21">21</xref>]. These students also reported high stress levels (65%), unhealthy diets (50%), and low levels of physical activity (77%) – Perhaps attributable to cultural and climatic restrictions [<xref ref-type="bibr" rid="B22">22</xref>]. Also, smoking has increased among men [<xref ref-type="bibr" rid="B23">23</xref>-<xref ref-type="bibr" rid="B25">25</xref>]. Hypertension is also common with prevalence reported as 19–25% [<xref ref-type="bibr" rid="B26">26</xref>], and 15.3% (urban populations) and 10.6% (rural populations) [<xref ref-type="bibr" rid="B27">27</xref>]. However, in these studies no information was collected from the expatriate communities. Neither did they obtain any data on treatment and treatment compliance.</p><p>The present study specifically addresses the characteristics of self reported hypertensives, and self reported normotensives in terms of demographic variables, the prevalence of life-style risk factors, under-diagnosis, and treatment outcome.</p></sec><sec sec-type="methods"><title>Methods</title><p>The study was approved by the Ethics Committee of the Faculty of Medicine and Health Sciences, United Arab Emirates University, Al Ain, UAE.</p><p>Subjects were enrolled over the period February 2001 to May 2005. The sample was stratified by SR hypertension (yes/no) and consisted of 349 SR hypertensives, and 640 SR normotensives (controls) among Arab nationals and expatriates (non-Gulf Arabs, Africans, Asians, Europeans and Far East Asians). There were 670 males and 319 females. Hypertensive subjects were recruited from various outpatients' clinics in Al-Ain city, UAE. Controls were enrolled from among University staff, teachers of secondary schools, factory workers, and various other institutions, to obtain a broad and representative cross-section of the adult population of Al-Ain.</p><p>Individuals between the age of 20 and 75 years with SR hypertension were referred to a special project physician in the hypertension clinic in Tawam Hospital, Al-Ain, to exclude secondary hypertension and known kidney disease or heart problems (no patients, however, had to be excluded for these conditions). Both patients and controls were then informed, in either Arabic or English, about the project's objectives and methods. Informed, written consent was obtained in all cases. Each individual was interviewed by a healthcare professional, using a one-page questionnaire, in Arabic or English, regarding demographic variables including ethnic background (country of origin), history of hypertension, diabetes, history of dyslipidaemia, current smoking (smoking cigarettes every day), current exercise (exercising at least for one hour/three times per week), and current alcohol intake (drinking alcohol at least once/three times per week). The questionnaire was piloted among primary health care patients and adapted accordingly. Although all patients had been prescribed anti-hypertensive treatment, no details on the exact medication were elicited. Besides to their prescribed conventional anti hypertensive drugs, many SR hypertensives were on lipid-lowering treatment if they had concomitant diabetes or dyslipidaemia.</p><p>All subjects were instructed to relax in a sitting position for at least 15 minutes prior to BP measurements. A mercury sphygmomanometer with appropriate cuff size was used for BP measurements. A specially trained nurse took two BP measurements for each individual in two different positions (sitting and standing) the average of which was used for analyses. Individuals with high BP levels among the SR normotensive group (incident of hypertension) were referred to a physician.</p><p>Hypertension was defined according to the Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC-7) definition [<xref ref-type="bibr" rid="B3">3</xref>]: normal (systolic < 120 mm Hg and diastolic < 80 mm Hg), pre-hypertension (systolic = 120–139 mm Hg or diastolic = 80–89 mm Hg), stage I (systolic = 140–159 mm Hg or diastolic = 90–99 mm Hg), stage II (systolic ≥ 160 mm Hg or diastolic ≥ 100 mm Hg). Incident (new, undiagnosed) hypertensives were defined as control subjects with a systolic blood pressure (SBP) ≥ 140 mm Hg and/or a diastolic blood pressure (DBP) ≥ 90 mm Hg. Under-treated hypertensives were defined as patients having BP levels in the above mentioned range.</p><p>In this study, weight, height, waist and hip circumference were all measured according to the WHO MONICA protocol [<xref ref-type="bibr" rid="B28">28</xref>]. Body mass index (BMI) was calculated as weight/height<sup>2 </sup>(kg/m<sup>2</sup>), and waist to hip ratio (WHR) was also calculated. Obesity was defined as BMI ≥ 30 kg/m<sup>2 </sup>and central body fat distribution was defined as a WHR ≥ 0.9 for men and ≥ 0.8 for women. These variables were measured by trained staff nurses in either the outpatient hypertension clinic in the case of hypertensives or in a special room in other organizations. Measurements were taken without shoes, and no heavy clothing was worn at the time of the investigations.</p><sec><title>Statistical analysis</title><p>For statistical analysis of the total data set, the Statistical Package for the Social Sciences (SPSS) Version 13.0 for Windows (SPSS, Chicago, U.S.A.) was used. Associations between categorical (nominal) variables were analyzed using cross tabulations and Chi-square tests, and the strength of the association was expressed in terms of Odds Ratios (OR), adjustment for confounding variables was carried out using the Mantel-Haenszel estimate of the OR. Continuous variables are presented as means +/- SD. Differences, adjusted for sex, among groups in continuous variables were analyzed using Analysis of Variance, and the Tukey method was used for post-hoc tests (i.e. p-values adjusted for multiple comparisons). P-values < 0.05 were considered statistically significant.</p></sec></sec><sec><title>Results</title><p>The number of individuals in each of the three BP groups (SR normotensives with normal blood pressure, incident hypertensives, [i.e. SR normotensives who actually had hypertension], and SR hypertensives) broken down by sex and nationality is shown in Table <xref ref-type="table" rid="T1">1</xref>. For each of these groups, the prevalence of regular exercise, smoking and alcohol consumption are also shown. The level of under-diagnosis of hypertension (216/640; 33%) and under-treatment (266/349; 76%) were very high. Many SR hypertensives even had blood pressures in the stage II hypertensive range (Table <xref ref-type="table" rid="T2">2</xref>). Interestingly, Emiratis, as compared to the immigrants (combined), had a much higher likelihood of having their hypertension diagnosed (Mantel-Haenszel OR, adjusted for sex 2.56 (95% CI: 1.65–3.97).</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>Demographic data by blood pressure group and by nationality and gender.</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left"><bold>Ethnicity</bold></td><td align="left"><bold> Numbers (N)</bold></td><td align="center" colspan="6"><bold>Blood pressure groups</bold></td></tr></thead><tbody><tr><td></td><td align="left">Life style risk factors</td><td align="center" colspan="2">Normotensives</td><td align="center" colspan="2">SR Hypertensives</td><td align="center" colspan="2">Incident Hypertensives</td></tr><tr><td></td><td></td><td colspan="6"><hr></hr></td></tr><tr><td></td><td></td><td align="center">M</td><td align="center">F</td><td align="center">M</td><td align="center">F</td><td align="center">M</td><td align="center">F</td></tr><tr><td></td><td colspan="7"><hr></hr></td></tr><tr><td align="left">Emirates</td><td align="left">N</td><td align="center">36</td><td align="center">34</td><td align="center">56</td><td align="center">60</td><td align="center">20</td><td align="center">13</td></tr><tr><td></td><td align="left">Regular Exercise (%)</td><td align="center">36</td><td align="center">44</td><td align="center">32</td><td align="center">38</td><td align="center">80</td><td align="center">77</td></tr><tr><td></td><td align="left">Cigarette Smoking (%)</td><td align="center">3</td><td align="center">0</td><td align="center">18</td><td align="center">2</td><td align="center">25</td><td align="center">0</td></tr><tr><td></td><td align="left">Alcohol Consumption (%)</td><td align="center">3</td><td align="center">0</td><td align="center">5</td><td align="center">3</td><td align="center">5</td><td align="center">0</td></tr><tr><td align="left">Non Gulf Arabs</td><td align="left">N</td><td align="center">82</td><td align="center">33</td><td align="center">29</td><td align="center">12</td><td align="center">48</td><td align="center">22</td></tr><tr><td></td><td align="left">Regular Exercise (%)</td><td align="center">60</td><td align="center">75</td><td align="center">48</td><td align="center">50</td><td align="center">56</td><td align="center">59</td></tr><tr><td></td><td align="left">Cigarette Smoking (%)</td><td align="center">22</td><td align="center">13</td><td align="center">24</td><td align="center">0</td><td align="center">27</td><td align="center">0</td></tr><tr><td></td><td align="left">Alcohol Consumption (%)</td><td align="center">0</td><td align="center">9</td><td align="center">3</td><td align="center">0</td><td align="center">0</td><td align="center">0</td></tr><tr><td align="left">Africans</td><td align="left">N</td><td align="center">44</td><td align="center">10</td><td align="center">54</td><td align="center">15</td><td align="center">22</td><td align="center">6</td></tr><tr><td></td><td align="left">Regular Exercise (%)</td><td align="center">73</td><td align="center">80</td><td align="center">63</td><td align="center">33</td><td align="center">68</td><td align="center">67</td></tr><tr><td></td><td align="left">Cigarette Smoking (%)</td><td align="center">20</td><td align="center">0</td><td align="center">20</td><td align="center">0</td><td align="center">14</td><td align="center">0</td></tr><tr><td></td><td align="left">Alcohol Consumption (%)</td><td align="center">2</td><td align="center">0</td><td align="center">2</td><td align="center">0</td><td align="center">9</td><td align="center">0</td></tr><tr><td align="left">Asians</td><td align="left">N</td><td align="center">109</td><td align="center">52</td><td align="center">82</td><td align="center">29</td><td align="center">61</td><td align="center">9</td></tr><tr><td></td><td align="left">Regular Exercise (%)</td><td align="center">68</td><td align="center">65</td><td align="center">57</td><td align="center">41</td><td align="center">67</td><td align="center">56</td></tr><tr><td></td><td align="left">Cigarette Smoking (%)</td><td align="center">18</td><td align="center">0</td><td align="center">17</td><td align="center">0</td><td align="center">21</td><td align="center">0</td></tr><tr><td></td><td align="left">Alcohol Consumption (%)</td><td align="center">17</td><td align="center">0</td><td align="center">17</td><td align="center">3</td><td align="center">23</td><td align="center">0</td></tr><tr><td align="left">Caucasians</td><td align="left">N</td><td align="center">10</td><td align="center">9</td><td align="center">1</td><td align="center">2</td><td align="center">10</td><td align="center">3</td></tr><tr><td></td><td align="left">Regular Exercise (%)</td><td align="center">90</td><td align="center">100</td><td align="center">0</td><td align="center">50</td><td align="center">100</td><td align="center">100</td></tr><tr><td></td><td align="left">Cigarette Smoking (%)</td><td align="center">20</td><td align="center">22</td><td align="center">100</td><td align="center">0</td><td align="center">0</td><td align="center">33</td></tr><tr><td></td><td align="left">Alcohol Consumption (%)</td><td align="center">90</td><td align="center">67</td><td align="center">100</td><td align="center">50</td><td align="center">100</td><td align="center">33</td></tr><tr><td align="left">Far East Asians</td><td align="left">N</td><td align="center">0</td><td align="center">5</td><td align="center">4</td><td align="center">5</td><td align="center">2</td><td align="center">0</td></tr><tr><td></td><td align="left">Regular Exercise (%)</td><td align="center">0</td><td align="center">80</td><td align="center">75</td><td align="center">60</td><td align="center">100</td><td align="center">0</td></tr><tr><td></td><td align="left">Cigarette Smoking (%)</td><td align="center">0</td><td align="center">0</td><td align="center">25</td><td align="center">20</td><td align="center">0</td><td align="center">0</td></tr><tr><td></td><td align="left">Alcohol Consumption (%)</td><td align="center">0</td><td align="center">20</td><td align="center">25</td><td align="center">40</td><td align="center">50</td><td align="center">0</td></tr></tbody></table><table-wrap-foot><p>Note: Incident hypertensives are self-reported normotensives with high blood pressure.</p></table-wrap-foot></table-wrap><table-wrap position="float" id="T2"><label>Table 2</label><caption><p>Blood pressure categories according to the JNC seventh report<sup>3</sup>.</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td></td><td></td><td align="center" colspan="2">SR Normotensives</td><td align="center" colspan="2">SR Hypertensives</td></tr></thead><tbody><tr><td align="left">BP Category</td><td align="center">SBP (mmHg)</td><td align="center">DBP (mmHg)</td><td align="center">Female (%)</td><td align="center">Male (%)</td><td align="center">Female (%)</td><td align="center">Male (%)</td></tr><tr><td colspan="7"><hr></hr></td></tr><tr><td align="left">Normal</td><td align="center">< 130</td><td align="center">< 85</td><td align="center">59.7</td><td align="center">45.0</td><td align="center">8.1</td><td align="center">13.4</td></tr><tr><td align="left">Pre-hypertension</td><td align="center">130 – 139</td><td align="center">85 – 89</td><td align="center">13.3</td><td align="center">18.2</td><td align="center">17.9</td><td align="center">8.5</td></tr><tr><td align="left">Stage I</td><td align="center">140 – 159</td><td align="center">90 – 99</td><td align="center">24.0</td><td align="center">30.4</td><td align="center">40.7</td><td align="center">37.5</td></tr><tr><td align="left">Stage II</td><td align="center">≥160</td><td align="center">≥100</td><td align="center">3.1</td><td align="center">6.3</td><td align="center">33.3</td><td align="center">40.6</td></tr><tr><td colspan="7"><hr></hr></td></tr><tr><td align="left">Total (N)</td><td></td><td></td><td align="center">196</td><td align="center">444</td><td align="center">123</td><td align="center">224</td></tr></tbody></table><table-wrap-foot><p>BP: blood pressure; SBP: systolic blood pressure; DBP: diastolic blood pressure; N: numbers; SR: self-reported. Incident Hypertensives are SR normotensives with high blood pressure.</p></table-wrap-foot></table-wrap><p>Table <xref ref-type="table" rid="T3">3</xref> shows the physical measurements of the three BP groups, broken down by gender. These results clearly show a pattern of high levels of CVD risk, with age, BMI and WHR generally increasing from the normotensive group, via the incident hypertensive group, to the SR hypertensives group.</p><table-wrap position="float" id="T3"><label>Table 3</label><caption><p>Demographic and anthropometric characteristics of the three blood pressure groups, by sex.</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td align="center" colspan="4"><bold>Normotensives </bold>(1)</td><td align="center"><bold>P Value </bold>(1 vs. 2)</td><td align="center" colspan="4"><bold>Incident Hypertensives </bold>(2)</td><td align="center"><bold>P Value </bold>(1 vs. 3)</td><td align="center" colspan="4"><bold>Self-Reported Hypertensives </bold>(3)</td><td align="center"><bold>P Value </bold>(2 vs. 3)</td></tr></thead><tbody><tr><td></td><td align="center" colspan="2">F (n = 142)</td><td align="center" colspan="2">M (n = 281)</td><td></td><td align="center" colspan="2">F (n = 53)</td><td align="center" colspan="2">M (n = 163)</td><td></td><td align="center" colspan="2">F (n = 123)</td><td align="center" colspan="2">M (n = 226)</td><td></td></tr><tr><td colspan="16"><hr></hr></td></tr><tr><td></td><td align="center">Mean</td><td align="center">SD</td><td align="center">Mean</td><td align="center">SD</td><td></td><td align="center">Mean</td><td align="center">SD</td><td align="center">Mean</td><td align="center">SD</td><td></td><td align="center">Mean</td><td align="center">SD</td><td align="center">Mean</td><td align="center">SD</td><td></td></tr><tr><td colspan="16"><hr></hr></td></tr><tr><td align="left">Age (yrs)</td><td align="center">38.5</td><td align="center">8.2</td><td align="center">44.2</td><td align="center">9.6</td><td align="center">0.049</td><td align="center">39.7</td><td align="center">8.8</td><td align="center">45.5</td><td align="center">8.2</td><td align="center">0.001</td><td align="center">49.8</td><td align="center">10.3</td><td align="center">51.4</td><td align="center">8.5</td><td align="center">0.001</td></tr><tr><td align="left">BMI (Kg/m<sup>2</sup>)</td><td align="center">27.8</td><td align="center">6.1</td><td align="center">26.5</td><td align="center">3.7</td><td align="center">0.019</td><td align="center">29.3</td><td align="center">5.4</td><td align="center">27.7</td><td align="center">5.1</td><td align="center">0.000</td><td align="center">30.8</td><td align="center">6.7</td><td align="center">27.9</td><td align="center">4.8</td><td align="center">NS</td></tr><tr><td align="left">WHR (× 100)</td><td align="center">84.9</td><td align="center">8.7</td><td align="center">92.5</td><td align="center">5.7</td><td align="center">NS</td><td align="center">83.4</td><td align="center">9.4</td><td align="center">92.3</td><td align="center">5.7</td><td align="center">0.000</td><td align="center">91.9</td><td align="center">8.6</td><td align="center">95.7</td><td align="center">4.9</td><td align="center">0.000</td></tr><tr><td align="left">Waist Circ. (cm)</td><td align="center">88.5</td><td align="center">13.5</td><td align="center">94.8</td><td align="center">10.6</td><td align="center">.004</td><td align="center">92.5</td><td align="center">15.6</td><td align="center">97.3</td><td align="center">12.1</td><td align="center">0.000</td><td align="center">98.7</td><td align="center">12.9</td><td align="center">99.2</td><td align="center">10.5</td><td align="center">0.018</td></tr><tr><td align="left">DBP (mm Hg)</td><td align="center">77.2</td><td align="center">5.5</td><td align="center">77.4</td><td align="center">5.6</td><td align="center">0.000</td><td align="center">89.8</td><td align="center">6</td><td align="center">91.4</td><td align="center">7.2</td><td align="center">0.000</td><td align="center">89.6</td><td align="center">10.5</td><td align="center">89.8</td><td align="center">10.5</td><td align="center">NS</td></tr><tr><td align="left">Pulse (B/Min)</td><td align="center">74.1</td><td align="center">7.8</td><td align="center">72.9</td><td align="center">6.5</td><td align="center">NS</td><td align="center">75.9</td><td align="center">7.8</td><td align="center">73.1</td><td align="center">7.8</td><td align="center">0.030</td><td align="center">75.8</td><td align="center">9.4</td><td align="center">74.2</td><td align="center">8.8</td><td align="center">NS</td></tr></tbody></table><table-wrap-foot><p>BMI: body mass index; WHR: waist to hip ratio; SBP: systolic blood pressure; DBP: diastolic blood pressure; NS: not significant (p > 0.05); Incident Hypertensives are SR normotensives with high blood pressure. The differences between the three methods are tested using Tukey's method, adjusted for sex.</p></table-wrap-foot></table-wrap></sec><sec><title>Discussion</title><p>In view of the uncertainty about the pathophysiology of many cardiovascular disorders, their association with life-style in different societies is of particular interest. This was one of the motivating factors for the 7 countries study by Ancel Keys [<xref ref-type="bibr" rid="B29">29</xref>]. The rapid life style changes that have taken place in the Gulf Region offer a unique opportunity to study the effect of such changes on CVD morbidity patterns and risk factors.</p><p>This study shows that in one of these countries (the UAE), despite a modern and accessible health infra-structure, the levels of both under-diagnosis and under-treatment of hypertension are high, as a large percentage of the SR normotensives and the SR hypertensives actually had BP in the hypertensive range. This situation is similar to that found in many other places in the world, both developed and developing. Even in the USA most hypertension is not properly controlled and among those with untreated or treated but uncontrolled hypertension a majority are aware of their condition [<xref ref-type="bibr" rid="B2">2</xref>]. Although our conclusions are based on a convenience sample, our cases appear to be broadly representative for the population of hypertensives in Al Ain, the second largest city in Abu Dhabi Emirate. Although the situation may differ somewhat in the other Emirates, such as Dubai, we do not believe a qualitatively different situation in other parts of the UAE.</p><p>The undiagnosed, "incident", hypertensives in out study are in many ways an intermediate group (Table <xref ref-type="table" rid="T3">3</xref>), perhaps because the onset of their hypertension is rather recent, which is also suggested by their generally younger age. This may also be due to their better general health than the SR hypertensives, with less concomitant morbidity and symptomatology, and have therefore not been diagnosed as hypertensives before. In addition, a few of them may not be real hypertensives, but only borderline hypertensives, who only occasionally or under special circumstances have BP values in the hypertensive range, and may require regular monitoring.</p><p>Among the Emirati nationals, most hypertension had been diagnosed previously and presumably had been treated. It is noteworthy that under-diagnosis of hypertension among non-Emiratis was much more common than among nationals. This may reflect poorer access to health care for immigrants. A large percentage of immigrants are low-income workers, e.g. employed in agriculture or construction. Until recently, the vast majority of such workers had access to (practically) free health care. Currently, the health system is in transition, and it is unclear what the future situation will be. The fact that until recently, their health care had been free, suggests that under-diagnosis is probably not due to financial reasons, but may reflect a different attitude to health and health care among these foreign workers, and perhaps a problem of communication between these workers, who often speak only Urdu or Pashtu, and health care professionals. One possible way to address this situation is to include a more extensive cardiovascular screening component in the periodic health check up (every 2–3 years) that is required for residence permit extensions.</p><p>Of special concern is the high level of under-treatment of hypertension. Despite the fact that the SR hypertensives were, by definition, aware of their condition, almost half of them did not have their BP adequately controlled (Table <xref ref-type="table" rid="T2">2</xref>). It is unclear whether this is due to inadequate treatment regimens, or lack of compliance e.g. because of asymptomatic nature of hypertension as well as medication side effects. Anecdotically, many individuals diagnosed with hypertension in this region do not properly adhere to treatment, because they partially mistrust the health care providers, and often they have wrong perception of the potential complications from their condition. Clearly, health care providers should better interact with patients to ensure proper compliance. Perhaps, the training of both doctors and nurses should focus more on communicating with patients in this highly multi-cultural environment.</p><p>In this population, the problem of under-diagnosis and under-treatment of hypertension is compounded by a high prevalence of other CVD life-style risk factors including smoking, lack of exercise, and obesity. In this context, the levels of obesity, for example, were extremely high with average BMIs in the overweight to obese range (Table <xref ref-type="table" rid="T3">3</xref>).</p><p>In summary, obesity, lack of exercise and smoking will clearly have a negative impact on the population's health status. Clearly, non- pharmacologic (as well as pharmacologic) approaches to the prevention of CVD among hypertensives and somewhat among normotensives are also needed in this population. The focus therefore should be on prevention, and efforts should address modifiable life style risk factors, such as diet, smoking and lack of exercise. Health authorities in the UAE have recognized smoking of both cigarettes and sheesha (water pipe), as a major health problem and steps have been taken to curb the sale of tobacco to minors and to raise awareness about the dangers of smoking. However, stimulation of regular exercise, known to be associated with substantial reductions in the incidence of coronary events [<xref ref-type="bibr" rid="B30">30</xref>], has received less attention, perhaps due to the fact that the extremely hot summer (June – September) constitutes a clear obstacle to outdoor activities. But, even during the pleasant winter months (October – March), few people exercise on regular basis. Overall, it is estimated that slightly over 50% of males, and 39% of females, exercise regularly [<xref ref-type="bibr" rid="B31">31</xref>]. There are also other, less clear, obstacles such as cultural perceptions and attitudes, and ad-hoc urban planning marked by lack of community based sport facilities. Despite the fact that some citizens will not change their life style whatever interventions are put in place [<xref ref-type="bibr" rid="B32">32</xref>], many others are likely to respond positively to more appealing interventions such as attractive recreational facilities.</p></sec><sec><title>Conclusion</title><p>Hypertension in the UAE should be approached more aggressively in terms of prevention strategies, diagnosis and treatment. <italic>Inter alia</italic>, provision of community sports facilities (children play grounds, safe cycling pathways, and indoor swimming pools), adapted to local culture and climate is recommended. Also, there is a pressing need to raise awareness among physicians and patients of the importance of adherence to treatment guidelines and recommendations, and to the potential complications of hypertension.</p></sec><sec><title>Competing interests</title><p>The author(s) declare that they have no competing interest.</p></sec><sec><title>Authors' contributions</title><p>AA designed the project, collected and interpreted the data, conducted the laboratory analysis and participated in the statistical analysis of the data, and drafted the manuscript.</p><p>NN performed the statistical analysis, and further coauthored the manuscript.</p><p>SA participated in the patient and control recruitment, carried out clinical examinations for the subjects, and critically reviewed the manuscript.</p><p>JP, participated in the measurement of the demographic variables, participated in the recruitment of the subjects and collected blood samples.</p><p>AA, participated in the design of the project, and commented on the manuscript.</p><p>EO, participated in its design and coordination, provided funds for the project, and critically reviewed the manuscript.</p><p>All authors read and approved the final manuscript.</p></sec><sec><title>Pre-publication history</title><p>The pre-publication history for this paper can be accessed here:</p><p><ext-link ext-link-type="uri" xlink:href="http://www.biomedcentral.com/1471-2261/6/24/prepub"/></p></sec> |
AceTree: a tool for visual analysis of <italic>Caenorhabditis elegans </italic>embryogenesis | <sec><title>Background</title><p>The invariant lineage of the nematode <italic>Caenorhabditis elegans </italic>has potential as a powerful tool for the description of mutant phenotypes and gene expression patterns. We previously described procedures for the imaging and automatic extraction of the cell lineage from <italic>C. elegans </italic>embryos. That method uses time-lapse confocal imaging of a strain expressing histone-GFP fusions and a software package, StarryNite, processes the thousands of images and produces output files that describe the location and lineage relationship of each nucleus at each time point.</p></sec><sec><title>Results</title><p>We have developed a companion software package, AceTree, which links the images and the annotations using tree representations of the lineage. This facilitates curation and editing of the lineage. AceTree also contains powerful visualization and interpretive tools, such as space filling models and tree-based expression patterning, that can be used to extract biological significance from the data.</p></sec><sec><title>Conclusion</title><p>By pairing a fast lineaging program written in C with a user interface program written in Java we have produced a powerful software suite for exploring embryonic development.</p></sec> | <contrib id="A1" contrib-type="author"><name><surname>Boyle</surname><given-names>Thomas J</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>biowolp@u.washington.edu</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Bao</surname><given-names>Zhirong</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>bao@gs.washington.edu</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Murray</surname><given-names>John I</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>murray@gs.washington.edu</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Araya</surname><given-names>Carlos L</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>araya@u.washington.edu</email></contrib><contrib id="A5" corresp="yes" contrib-type="author"><name><surname>Waterston</surname><given-names>Robert H</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>waterston@gs.washington.edu</email></contrib> | BMC Bioinformatics | <sec><title>Background</title><p>The invariant lineage of the nematode <italic>C. elegans </italic>[<xref ref-type="bibr" rid="B1">1</xref>] can potentially be exploited to capture detailed information on the location and timing of expression for the genes expressed in the early embryo. In addition, changes in the lineage resulting from mutations or RNAi knockdowns of gene function can provide functional information about genes. To use the lineage with high throughput, one must capture images in sufficient detail and subject those images to automated lineaging.</p><p>Bao et al[<xref ref-type="bibr" rid="B2">2</xref>] described a procedure for generating lineages automatically. A ubiquitously expressed histone-GFP fusion protein is used to label nuclei in the developing embryo. Images sets are captured once a minute by a confocal microscope, with each set containing up to 35 focal planes through the full depth of the embryo. The program StarryNite analyzes the images to locate all the nuclei at each time point and to establish the linkage of nuclei from time point to time point. The resulting annotation, which implicitly establishes the lineage of the embryo, is written to a series of files, called the nuclei files, one for each time point.</p><p>A separate program, AceTree, was written to facilitate viewing, editing and interpretation of the StarryNite output and is described here. This program is separate from StarryNite, with distinct requirements. StarryNite was written in C with a minimal user interface because of the computationally intensive character of automated lineage extraction. The editing and interpretation tools required for AceTree require robust and portable user interfaces and the ability to develop and test new tools rapidly, making Java a logical choice. In addition, AceTree can be used without StarryNite output as an image viewer and manual lineaging tool, although it is not currently optimized for the latter task.</p><p>Three other software packages exist that link image series and lineage trees. The most complete of these is SIMI BioCell, which is optimized for manually lineaging of 4D differential-interference-contrast (DIC) image series and has been used effectively to demonstrate the striking insights that can be obtained by lineage analysis[<xref ref-type="bibr" rid="B3">3</xref>]. Angler was developed by the developers of AceDB and is directed mainly at viewing and interpreting lineaged series with more reference to information stored in WormBase[<xref ref-type="bibr" rid="B4">4</xref>]. Virtual Wormbase has both an educational goal and a research goal and embeds the idea of simulating the development process[<xref ref-type="bibr" rid="B5">5</xref>]. These were all designed to deal with 4D DIC image series and thus are not optimized for the specifics of GFP-histone image series.</p><p>We chose to develop a new program, AceTree, rather than attempt to adapt an existing program for several reasons. The new program could be optimized for viewing fluorescence images, including "second color" images used to track gene expression throughout embryogenesis. The key data structure of AceTree is identical with the nuclei files produced by StarryNite, which facilitates the expected co-evolution of the programs as the project matures. In addition, AceTree is being made available as an open source package supporting all major PC operating systems.</p></sec><sec><title>Implementation</title><p>Since computation speed is unlikely to be an issue in the user interface oriented AceTree, a high-level language solution was sought. Many biological analysis programs with user interface aspects are written in Java and that language received the top consideration. Two components were needed: a package for handling tiff images and a way to produce interactive trees. ImageJ[<xref ref-type="bibr" rid="B6">6</xref>] met the first requirement and provided extra features some of which have since been used to assist in handling the image series as they come from the microscope and to produce and view movies developed from images arising in AceTree. The Java class known as DefaultMutableTreeNode contains all the normal features of a tree data structure and the JTree class provides a convenient graphical user interface for it.</p><p>AceTree is written in Java version 1.4.2. For 3D representations the Java3D module is required. Image operations of AceTree are derived from ImageJ and the ij.jar file is required. Development is carried out in the open source Eclipse Platform [<xref ref-type="bibr" rid="B7">7</xref>,<xref ref-type="bibr" rid="B8">8</xref>]. For source distribution an Ant [<xref ref-type="bibr" rid="B9">9</xref>] build.xml file is provided. The program is packaged as a jar file and has been tested on conventional workstations using three different operating systems: linux; Mac OS X; windows.</p></sec><sec><title>Results and discussion</title><sec><title>Program components</title><p>The normal operating mode of AceTree is shown in Figure <xref ref-type="fig" rid="F1">1</xref> where the process of opening a set of data is complete and the user is viewing annotated images in the ImageWindow and navigating among them using the main control. The title bar of the ImageWindow contains a coded part, "t001-p17" which tells the user that this image is from time point 1 plane 17 of the image series under study, "081505". The underlying tiff image contains only a single gray scale "haze", with strong (whiter) clusters corresponding to the green fluorescent protein histone fusions. AceTree shows this in the green plane of an RGB image and has added annotations.</p><p>The imaging protocol has in this case captured the four-cell stage of the embryo in the first time point. The basic image annotation consists of blue circles around the nuclei, which have been modelled as spheres. At any time, one of the cells will have the special designation of "current cell" and will have a white circle, in this case the ABa cell. Cell names are a basic element of the annotation scheme and AceTree offers the user several ways to control which cells carry their names on the displayed images.</p><p>The main control, shown on the right in Figure <xref ref-type="fig" rid="F1">1</xref>, contains the following seven parts:</p><p>1. A menu bar some of whose application specific functions will be addressed herein.</p><p>2. A "JTree" representation of the lineage where ABa is highlighted.</p><p>3. A text display window with information about the current cell being displayed.</p><p>4. A control that enables the user to bring up a particular cell at a particular time point.</p><p>5. A "movie" control that can be used to automatically sequence through the images tracking the current cell and its descendents.</p><p>6. A pad with 12 buttons. These buttons provide considerable control over navigation and the kinds of information shown with the image. The main navigation keys – next, prev, up, down – are mapped to the arrow keys on the keyboard.</p><p>7. A text window displaying the position of the mouse when it is within the image.</p><p>Although it is not routinely needed during the study of an image series, AceTree via its Analyze..NucleiView menu presents the automated lineaging data structure on which all of the interaction and annotation is based. A sample of this display is shown in Figure <xref ref-type="fig" rid="F2">2</xref>.</p><p>The figure shows the annotations from times 4 and 5 corresponding to the division of ABa and ABp, the results of which are shown in Figure <xref ref-type="fig" rid="F3">3</xref>. The identity, x, y, z, and size columns are readily understood as the name, position, and diameter in pixels of the identified nucleus. The pred, succ1, and succ2 columns describe how nuclei at one time point relate to those at adjacent time points. In the absence of a division, succ2 is negative and succ1 contains the index of the continuing nucleus in the next time point. Where a division was detected, succ1 and succ2 point to the two daughter cells in the next time point. In this case, the division of ABa can thus be followed: the two daughters are located at indices 3 and 7 in the nuclei structure at time 5. The propagation of names to daughters follows from the linkage using rules based on the relative locations of the daughters. This is the basis on which AceTree has built its tree representation. The weight and rweight columns contain the summed intensity within the spherical representation of the nucleus for the green and red channel images, respectively.</p></sec><sec><title>Cell tracking</title><p>Tracking cells from image to image is a common need in analyzing lineaged image data. The "Next" key on the keypad moves to the next image in time while tracking the "current cell". If the nucleus is moving in the z plane (here the left/right embryonic axis), then AceTree adjusts, choosing the z plane image which is closest to the center of the nucleus. Figure <xref ref-type="fig" rid="F3">3</xref> shows the image window when tracking has reached time 5. Tracking can be dismissed permitting the user to follow a sequence of images where the plane of the image remains fixed.</p><p>In the first time point following a division, the current cell designation will move to one of the daughters of the divided cell. Here, the new "current cell" is ABal: AceTree has tracked the left daughter of the division of the ABa cell. However, it also shows in cyan the sister cell, ABar. The annotation in cyan advises that the nucleus of ABar is centered in plane 19. Since we are looking at the plane 13 image, the cyan annotation is a "ghost nucleus" and in this case reassures us that the embryo has followed the left/right division pattern suggested by its canonical name. Notice also that in the position previously occupied by ABp we have ABpl: both daughters of AB have divided at the same time,</p><p>There are multiple ways to select the current cell while studying an image series. The tree representation is "live". One can expand the tree to a cell of interest and left click on it in the tree to bring up the image at its "birth". Alternatively, one can right click on it to bring it up just before its division or death. The cell selection panel can be used to bring up a given time point. Any circle displayed in the image can be right clicked to make that cell the current cell and the tree display is updated correspondingly.</p></sec><sec><title>Curating the automated lineage</title><p>AceTree can be used to curate and edit lineaged data series such as those arising from StarryNite. AceTree accepts as its inputs the image series and the annotation files whose content was shown in Figure <xref ref-type="fig" rid="F2">2</xref>. If the series is edited, AceTree creates a revised set of annotation files which would be used in subsequent studies of the data. The tools of the Edit menu are provided for these purposes.</p><p>Figure <xref ref-type="fig" rid="F4">4</xref> shows the EditTraverse tool from that menu along with the ImageWindow when the tool is in use. EditTraverse builds a list of all cell divisions and deaths starting from the cell named in the edit box. As the highlight is moved, the ImageWindow receives the image showing the cell in question just before its division. The user can quickly examine the sequence of events in the time leading up to and following the division to verify that the lineage has correctly tracked this important event. If problems are detected, the EditImage tool can be brought up. That tool makes it possible to add, remove, reposition, resize, and relink cells to correct any errors.</p><p>The specific situation shown in Figure <xref ref-type="fig" rid="F4">4</xref> shows cell Epr just prior to its division into Epra and Eprp. The image clearly shows the metaphase plate and its orientation for the division along the anterior/posterior axis.</p><p>When a lineage is edited in such a way that cell linkage is changed from that in the original nuclei files, the tree representation must be rebuilt and cell naming must be redone. Therefore, AceTree contains the algorithms described by Bao[<xref ref-type="bibr" rid="B2">2</xref>] to determine cell names upon division. These include the algorithm for determining the embryonic axes from the four founder cells and the subsequent divisions of ABa and ABp which provide the essential starting point for cell naming. Therefore, it is possible to use AceTree to manually lineage an image series and develop annotation files to be used in subsequent studies of the data although the editing tools are not optimal for such a purpose.</p></sec><sec><title>Viewing embryonic lineage and morphology</title><p>The ability to generate automated lineages opens up the possibility of using the lineage as a phenotype in considering developmental changes induced by altered gene function. For this purpose, a tree that tracks the timing of cell divisions is necessary: we call it an ancestral tree. AceTree provides this on its Trees..Ancestral Tree menu. Figure <xref ref-type="fig" rid="F5">5</xref> shows two examples of such a tree where the EMS sublineage is shown from a wild type embryo on the left and the corresponding lineage for an embryo where RNAi has been used to inhibit the expression of gene <italic>lit-1 </italic>on the right. The time points from the two different data series were chosen at a point where the total number of cells present is stable for several time points and the AB sublineage consists of 64 cells.</p><p>The modified embryo has 4 cells more than the wild type due to an extra division in the E lineage which would normally form the gut. With <italic>lit-1 </italic>function inhibited, the E cells adopt the fate of the MS lineage and the gut fails to form resulting in embryonic lethality[<xref ref-type="bibr" rid="B10">10</xref>]. Notice that the E and MS subtrees for the modified embryo are similar to each other and similar to the wild type MS sublineage but not the wild type E sublineage.</p><p>Besides tracking the lineage of cells, the annotation scheme used here (see Figure <xref ref-type="fig" rid="F2">2</xref>) describes the location and size of each nucleus. This raises the possibility of studying the morphology of the embryo in a 3D space filling model representation. AceTree provides this in its View..3D View menu. Figure <xref ref-type="fig" rid="F6">6</xref> shows the results of that tool from each of the series discussed above: wild type embryo on the left and <italic>lit-1 </italic>inhibited embryo on the right. Both views have been rotated so one sees the ventral side of the embryo with the anterior on the left. The nuclei are color coded so the main lineages can be distinguished: ABa daughters in gray; ABp in white; E in red; MS in blue; C in yellow; D in pink; germ line in green.</p><p>At this stage the four E cells (red) of the wild type embryo have migrated to the interior beginning the process of gut formation. The space filling model of the <italic>lit-1 </italic>inhibited embryo shows the E cells remaining on the outside of the embryo and matching in number the MS cells. Aside from the abnormal gastrulation, the positioning of other sublineages has also been altered. AceTree offers considerable freedom to select the colors of different sublineages and the orientation of the model to explore morphology.</p></sec><sec><title>Tracking gene expression</title><p>Tracking the expression of individual genes in time and cellular location is of major interest to developmental biologists. The automated lineaging technique can be used to track expression of genes of interest by introducing a nuclear localized red fluorescent reporter of a gene's promoter activity. The imaging system then collects a parallel set of images in the red channel and the location and size information in the GFP-derived annotations is used to compute the red intensity of each nucleus.</p><p>In such cases, AceTree offers several ways to view such information. In the normal viewing of images the red data is added to the red plane of the RGB ImageWindow so non-expressing nuclei show up in green as usual and expressing nuclei show up in yellow. The user can choose to view only green or only red as well. To make the red expression available in a more interpretable way, the lineage tree can be colored according to the extent of red expression. Figure <xref ref-type="fig" rid="F7">7</xref> shows such a presentation for a data series where the tagged gene, <italic>pha-4</italic>, is known to be expressed in the pharynx and gut of the developing embryo including the E lineage and sublineages of MS and AB[<xref ref-type="bibr" rid="B11">11</xref>] (Murray, JI and Waterston, RH, unpublished).</p><p>Here, the tree branches are color coded to show the extent of red expression: if the cell is not expressing above the minimum threshold, the branch is shown in gray. For cells that are above the threshold the color starts at light green, getting darker, then black, then shades of red as the expression increases. In Figure <xref ref-type="fig" rid="F7">7</xref>, the entire lineage is captured and the overall character of gene expression is seen at a glance. The cell annotations are shown for every tenth leaf in the lineage. The user can adjust the spacing between lines and line thickness to explore things in more detail, but for detailed review, the tree version shown in Figure <xref ref-type="fig" rid="F8">8</xref> may be more effective. Here, the MSa sublineage is shown and all cells are named. For further exploration, these trees are "live" like the JTree so a click on a branch brings up the image corresponding to that cell and time and makes it the "current cell" in the main control where the red expression value is reported numerically.</p><p>Figure <xref ref-type="fig" rid="F9">9</xref> demonstrates how the space-filling model can be used to study gene expression in this situation. Two alternative views of the embryo at time 250, when there are about 500 cells present, are shown. The view on the left emphasizes the observed expression pattern; that on the right the expression pattern expected from the literature on the gene of interest. Both views are rotated so we see the ventral side with the anterior to the left and both use more advanced coloring features to bring out information.</p><p>On the left, each nucleus is colored according to the degree of red expression using the same scheme as in Figure <xref ref-type="fig" rid="F7">7</xref> except that nuclei that are not expressing are shown in transparent white. Thus according to the data, all cells shown clearly are expressing the tagged gene, with the expression being especially high in those shown in red.</p><p>In the view on the right transparent white is again used, in this case for cells where, according to the literature, no expression is expected. Within the groups of cells that are reported to express the gene of interest, the lineage color scheme is: ABal in pink; ABar in blue; E in yellow; MSaa in magenta; MSap in cyan. Pairing the two 3D representations permits the investigator to consider the validity of the hypotheses about the location of gene expression in a very detailed way.</p></sec><sec><title>Embryonic rotation</title><p>Aspects of embryo morphology that are difficult to show in other ways can be vividly illustrated using tools available in AceTree. The pattern of rotation of the developing embryo illustrates this. During normal development, the embryo has been reported to undergo two rigid-body rotations around its a-p axis[<xref ref-type="bibr" rid="B1">1</xref>]. Variability in the direction and magnitude of this rotation was reported by Schnabel et al [<xref ref-type="bibr" rid="B3">3</xref>]. In Figure <xref ref-type="fig" rid="F10">10</xref> a series of 3D images drawn from the annotations is shown as viewed from the anterior/posterior axis with the dorsal on top. Only cells from the ABp lineage are shown and the time points – 18, 35, 58, 84, 115, 151 – are just prior to the next round of divisions in the series. Cells from the ABpl sublineage are shown in red; ABpr in blue. A clockwise rotation of about 45 degrees can be visualized as development proceeds.</p><p>One method of quantifying the angle is to visualize a line drawn between the center of mass of the ABpr cells and that of the ABpl cells. The angle of this line in the plane can then be plotted as shown in the left graph of Figure <xref ref-type="fig" rid="F11">11</xref>. The right side of the figure shows the total cell count in the embryo over time. The calculated angle increases clockwise relatively smoothly after the 15 cell stage of the embryo which occurred at about time 30 in the series.</p></sec></sec><sec><title>Conclusion</title><p>Bao et al [<xref ref-type="bibr" rid="B2">2</xref>] described the lineaging protocol and presented the results of a series of 20 lineaged embryos, using the low error rate observed to establish the validity of the overall approach. The minimum requirement for AceTree was that it facilitate curation and editing of the automated lineages. Bao <italic>et al </italic>reported that an earlier version of AceTree allowed curation and editing of a series through the 194 cell stage in about 2 hours and that an additional 2–8 hours were required to edit through the 350 cell stage. Editing time is a function of both the quality of the StarryNite output and the utility of AceTree. A goal for future versions of AceTree and StarryNite should be to further expedite editing beyond the 194 cell stage.</p><p>The longer-term goal of enhancing the researcher's capability of extracting biological significance from the data is more difficult to quantify. In this regard, AceTree is in a position similar to a number of software tools reported in the literature, especially those supporting the analysis of microarray data [<xref ref-type="bibr" rid="B12">12</xref>,<xref ref-type="bibr" rid="B13">13</xref>]. There, as here, the tools permit a combination of data clustering and tree representations, which are meant to guide the researcher through a complex dataset. In the final analysis, the clues obtained by using such tools form only a part of the evidence that leads to new knowledge. We are encouraged by the character of the examples shown here: the work with the <italic>lit-1 </italic>RNAi treated embryo; the visualizations of the tagged gene; the embryonic rotation study. We remain open to suggestions for additional features deemed helpful by practicing researchers.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>User-interface classes</title><p>The AceTree class provides the main control and is the heart of the application. The link to StarryNite is established through the reading of the nuclei files, which are the main output of that program and by associating them with the same tiff images that were used in the automated lineaging. While this is essentially a user interface class, it is in control of cell and image navigation and as such contains code that embodies the essence of our application. The user navigates by a variety of methods: mouse clicks on the JTree, use of the cell selection panel, use of the button pad and its keyboard equivalents, mouse clicks on the AncestralTree trees, and the movie control. The "current cell" feature is controlled in AceTree. AceTree calls into play about 40 classes that handle various aspects of the user interface. These in turn make use of the javax.swing base classes.</p><p>The ImageWindow class makes the raw data of the process accessible and guides the eye by added annotation: circles locating the cells and the current cell, cell names, sister display. ImageWindow relies on classes of the ImageJ package for rendering the tiff images and annotating them. The EditImage class is derived from ImageWindow.</p><p>The Image3D class is built upon Java3D and this package is required in addition to Java 1.4.2 for the 3D feature to be available.</p></sec><sec><title>Application data structures</title><p>The technical underpinnings of the application are embodied in five classes: Nucleus; NucleiMgr; Identity; Cell; AncesTree.</p><p>The Nucleus class contains the elements used to describe each nucleus at each time point. The member variables are all public and are directly accessed wherever needed in the AceTree code. The specific variables are those shown in Figure <xref ref-type="fig" rid="F2">2</xref>: identity, index, status, predecessor, successor1, successor2, x, y, z, size, green weight, red weight.</p><p>The NucleiMgr class holds the nuclei_record: a Vector of Vectors where the final elements are instances of the Nucleus class. Thus one Vector of the nuclei_record represents all nuclei at a time point. NucleiMgr has code for reading the nuclei files and creating its internal representation as well as code for recreating the successor indices after an edit operation (the "rebuild" action). The various edit operations act directly on the nuclei_record. AceTree supports saving an edited nuclei_record to nuclei files in the same form as the original files.</p><p>The Identity class implements the code to assign cell names. Identity is a singleton class and contains the key public function identityAssignment(), which generates names. Two naming methods described by Bao[<xref ref-type="bibr" rid="B2">2</xref>] are implemented. The main loop of STANDARD naming calls the private sisterID() function, which examines the parent/daughter orientation and attaches the implied character to the growing name. Canonically, the plane of division was determined by the orientation of the spindle axis at the initiation of division, with a preference given to the a/p division. Most divisions were assigned the addition of 'a' to one sister and 'p' to the other. In the current images the spindle is invisible. Instead, we rely on the axis of the daughter cells. When the x position of one sister can clearly be seen to be less than that of the other (say more than one quarter of a cell diameter), the daughter is assigned 'a'. Failing that, the code looks for a y axis difference and if that too is insufficiently large the z direction is used. Naturally there is some special case code in Identity to handle the germ line, polar bodies, and the EMS division.</p><p>Naming can proceed only when the orientation of the embryo is known. The imaging protocol guarantees that the x image axis will align with the anterior-posterior axis but cannot guarantee that the positive x direction is posterior. Likewise the dorsal-ventral and left-right axes may variously align with the y and z axes. The Identity module determines the axes by considering the 4 and 8 cell stages. ABa and ABp are distinguished from EMS and P2 by their earlier divisions and the initial axes can be unambiguously determined. If axes are not determined, all names are stylized: sublineages are called "Nuc" with a number added to distinguish the sublineage and directional letters to track subsequent divisions.</p><p>CANONICAL naming relies on a rule table based on the Sulston[<xref ref-type="bibr" rid="B1">1</xref>] lineage. The rule tells for each canonical parent what axis to use to examine the division and how the daughters are to be named to stay within the canonical list. Finally, there is "manual" naming for use in manual lineaging. The choice of naming method is a user option: canonical is the default.</p><p>The Cell and AncesTree classes enable AceTree to offer tree representations based on the annotation information in the nuclei files. The Cell class derives from the Java DefaultMutableTreeNode class. That class offers all the normal data structure features of trees such as the various traversals. AceTree calls the constructor of AncesTree giving it a NucleiMgr object. The constructor calls the private processEntries() function which proceeds through the nuclei_record from time point to time point. In the general case, the presence of a new cell is detected by examination of the successor indices and new Cell objects are created for each sister. When AceTree needs to refer to the tree as a whole, it refers to the root Cell: the properties of the DefaultMutableTreeNode take care of the details.</p><p>Access to cells within AceTree is facilitated by two hashtables kept by AncesTree and made available via access functions: iCells and iCellsByName. The member variables of the Cell object include: name, starting time, ending time, ending fate (alive, divided, died) and a hashKey which is constructed from the time and index of its Nucleus at birth. The Cell object also contains the drawing code needed to render it in Ancestral Tree style trees.</p></sec></sec><sec><title>Availability and requirements</title><p>AceTree is available as part of the StarryNite-AceTree package from <ext-link ext-link-type="uri" xlink:href="http://waterston.gs.washington.edu"/>. The AceTree.jar file contains the source code. It requires Java 1.4.1 and the ImageJ jar file ij.jar; Java3D must be installed to use the space filling model features. AceTree has been tested on linux, macintosh, and windows. The <underline>AceTreeDemo download</underline> contains the program and a data set in a single zip file of about 200 MB. The AceTree Help menu provides a tutorial in html called AceTreeDemo, which can be followed to learn the program features, the character of the data, and the power of the annotations developed by StarryNite. AceTreeDemo.html can be extracted from the jar file and printed from a browser to provide an offline manual. The AceTree.jar file in the demo package is the full featured program and can be used with other data available on the site.</p></sec><sec><title>Authors' contributions</title><p>TJB designed the software with input from all authors. TJB programmed the software. Experiments were conceived and performed by ZB, JIM, and CA. TJB wrote the manuscript with significant input from ZB, JIM, and RHW.</p></sec> |
ROKU: a novel method for identification of tissue-specific genes | <sec><title>Background</title><p>One of the important goals of microarray research is the identification of genes whose expression is considerably higher or lower in some tissues than in others. We would like to have ways of identifying such tissue-specific genes.</p></sec><sec><title>Results</title><p>We describe a method, ROKU, which selects tissue-specific patterns from gene expression data for many tissues and thousands of genes. ROKU ranks genes according to their overall tissue specificity using Shannon entropy and detects tissues specific to each gene if any exist using an outlier detection method. We evaluated the capacity for the detection of various specific expression patterns using synthetic and real data. We observed that ROKU was superior to a conventional entropy-based method in its ability to rank genes according to overall tissue specificity and to detect genes whose expression pattern are specific only to objective tissues.</p></sec><sec><title>Conclusion</title><p>ROKU is useful for the detection of various tissue-specific expression patterns. The framework is also directly applicable to the selection of diagnostic markers for molecular classification of multiple classes.</p></sec> | <contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Kadota</surname><given-names>Koji</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>kadota@iu.a.u-tokyo.ac.jp</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Ye</surname><given-names>Jiazhen</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>ye@bi.a.u-tokyo.ac.jp</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Nakai</surname><given-names>Yuji</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>yunakai@iu.a.u-tokyo.ac.jp</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Terada</surname><given-names>Tohru</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>tterada@iu.a.u-tokyo.ac.jp</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Shimizu</surname><given-names>Kentaro</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>shimizu@bi.a.u-tokyo.ac.jp</email></contrib> | BMC Bioinformatics | <sec><title>Background</title><p>A major challenge of microarray analysis is to detect genes whose expression in a single or small number of tissues is significantly different than in other tissues. Accurate identification of such tissue-specific genes can allow researchers to deduce the function of their tissues and organs at the molecular level [<xref ref-type="bibr" rid="B1">1</xref>].</p><p>Several methods have been used for this purpose [<xref ref-type="bibr" rid="B1">1</xref>-<xref ref-type="bibr" rid="B5">5</xref>]. Of these, Schug et al. [<xref ref-type="bibr" rid="B4">4</xref>] demonstrated the effectiveness of using Shannon information theoretic entropy for ranking genes according to their tissue-specificity, from restricted (tissue-specific) expression to average (ubiquitous/housekeeping) expression. However, there is also a severe disadvantage. The entropy does not explain to which tissue a gene is tissue-specific, only measuring the degree of overall tissue specificity of the gene. Hence further analysis to identify specific tissues is needed. Although Schug et al. [<xref ref-type="bibr" rid="B4">4</xref>] proposed a new statistic (<italic>Q</italic>) based on entropy to estimate the degree of a gene's specificity on a particular tissue, the issue of redundancies remains where top-ranked genes as specific to tissue <italic>A </italic>are also top-ranked as specific to tissue <italic>B</italic>. We assert such genes are not specific to <italic>A </italic>or <italic>B</italic>, but rather are genes specific to both <italic>A </italic>and <italic>B</italic>. For example, we observed that two of the top five probesets specific to liver were also found in the top five probesets specific to gall bladder [<xref ref-type="bibr" rid="B4">4</xref>]. The issue of such redundancies is a concern with any ranking-based method, such as pattern-matching [<xref ref-type="bibr" rid="B2">2</xref>], when the number of interrogated tissues increases. Methods of identifying genes specific only to objective tissues are needed.</p><p>Unlike ranking-based methods, methods based on outlier detection are free from the issue of redundancies because they identify tissues corresponding to both over- and under-expressed outliers for each gene [<xref ref-type="bibr" rid="B3">3</xref>,<xref ref-type="bibr" rid="B5">5</xref>]. Therefore, these methods can treat equally various types of tissue-specific genes: (1) 'up-type' genes selectively over-expressed in a single or small number of tissues compared to the others, (2) 'down-type' genes selectively under-expressed, and (3) 'mixed-type' genes selectively over- and under-expressed in some tissues. Although the mixed-type is possible, the first two types (up-type and down-type) of expression patterns are particularly important because those genes may be associated with fundamental biological phenomena and may contain particular tissue-specific diagnostic markers. Using outlier-detection-based methods, however, ranking genes according to their degree of overall tissue-specificity is difficult.</p><p>This complementary relationship between ranking-based and outlier-based methods led us to develop a combined approach, ROKU. ROKU analyzes any type of tissue-specific genes (up-, down-, and mixed-type) in two steps. First, it ranks genes according to overall tissue-specificity using Shannon entropy, and second, for each gene, it identifies specific tissues whose observations are regarded as outliers using a method of Kadota et al. [<xref ref-type="bibr" rid="B3">3</xref>]. We applied the method to both synthetic and real gene expression data and demonstrated its utility by comparison with other methods.</p></sec><sec><title>Results and discussion</title><sec><title>Definition of tissue-specific genes</title><p>We first show typical examples of various types of gene expression patterns. We here divided tissue-specific genes into two levels, a narrow sense and a broad sense. Genes over-expressed in a small number of tissues but unexpressed or slightly expressed in others, such as those shown in Figs. <xref ref-type="fig" rid="F1">1a</xref> and <xref ref-type="fig" rid="F1">1c</xref>, are defined as tissue-specific genes in a narrow sense, while genes over- and/or under-expressed in a small number of tissues compared to other tissues are defined as tissue-specific in a broad sense (the latter group includes the former). We focused here on the latter case and wanted to identify such expression patterns (see black scatter plots in Figs. <xref ref-type="fig" rid="F1">1d–f</xref>). We use two terms ("genes" and "probesets") interchangeably throughout this paper.</p></sec><sec><title>Data processing and its effect on Shannon entropy calculation</title><p>When one gene vector <bold><italic>x </italic></bold>= (<italic>x</italic><sub><italic>1</italic></sub>, <italic>x</italic><sub><italic>2</italic></sub>, ..., <italic>x</italic><sub><italic>N</italic></sub>) is given, the entropy <italic>H</italic>(<bold><italic>x</italic></bold>) can be calculated by equation 1 (See Methods). The range of <italic>H </italic>is from 0 whose gene expression is perfectly restricted in a single tissue (Fig. <xref ref-type="fig" rid="F1">1a</xref>) to log<sub>2</sub>(<italic>N</italic>) whose gene expression pattern is flat in all the interrogated tissues (Fig. <xref ref-type="fig" rid="F1">1b</xref>). We therefore rely on the low entropy score for the identification of tissue-specific genes. The black scatter plots in Fig. <xref ref-type="fig" rid="F1">1</xref> are synthetic expression observations for <italic>N </italic>tissues (i.e., <italic>N </italic>= 10 in this case). The entropy <italic>H </italic>for each gene vector <bold><italic>x </italic></bold>is given by the number in black above the figures. Clearly, direct calculation of the entropy for raw gene vector <bold><italic>x </italic></bold>works well only for detecting tissue-specific genes in a narrow sense (Figs. <xref ref-type="fig" rid="F1">1a</xref> and <xref ref-type="fig" rid="F1">1c</xref>) but not for those in a broad sense (Figs. <xref ref-type="fig" rid="F1">1d–f</xref>). The <italic>H </italic>scores (3.22, 3.29, 3.23 for Figs. <xref ref-type="fig" rid="F1">1d–f</xref>, respectively) of tissue-specific genes in a broad sense are close to the maximum value (log<sub>2</sub>10 = 3.32) and cannot identify those genes as 'tissue-specific'.</p><p>To detect tissue-specific genes in a broad sense, we introduce a simple method that processes a given gene vector <bold><italic>x </italic></bold>and makes a new vector <bold><italic>x</italic></bold>'. Data processing is done by subtracting the one-step Tukey biweight and by taking the absolute value of equation 2 (see Methods). The Tukey biweight yields a robust weighted mean able to resist 50% of outliers [<xref ref-type="bibr" rid="B6">6</xref>]. The scatter plots of processed vectors are shown in red in Fig. <xref ref-type="fig" rid="F1">1</xref>. The entropy scores, <italic>H</italic>(<bold><italic>x</italic></bold>'), for the processed vectors to obvious tissue-specific genes in a broad sense (Figs. <xref ref-type="fig" rid="F1">1d–f</xref>) are considerably lower than those for <bold><italic>x</italic></bold>. This is because the relative expression levels for specific tissues (highlighted tissues) become high after data processing. For example, the value (0.04) for tissue 3 in Fig. <xref ref-type="fig" rid="F1">1e</xref> becomes 0.75 after data processing. Since the base-line value is 0.1 (1/<italic>N</italic>, <italic>N </italic>= 10) in this case, such high values decisively contribute low entropy to the gene expression pattern. Also, entropy scores, <italic>H</italic>(<bold><italic>x</italic></bold>') and <italic>H</italic>(<bold><italic>x</italic></bold>), to non-specific (or randomly expressed) genes are quite similar and close to the maximum (3.32) (Figs. <xref ref-type="fig" rid="F1">1g</xref> and <xref ref-type="fig" rid="F1">1h</xref>). These results demonstrate the adequacy for our strategy for detecting tissue-specific genes in a broad sense at least on typical/hypothetical expression data.</p></sec><sec><title>Analysis of real data</title><p>To further investigate the validity of our method (ROKU), we applied the method to a public gene expression matrix consisting of 36 normal human tissues and 22,283 probesets [<xref ref-type="bibr" rid="B5">5</xref>]. Briefly, ROKU (1) processes each probeset expression vector and makes a processed vector <bold><italic>x</italic></bold>', (2) calculates the entropy <italic>H</italic>(<bold><italic>x</italic></bold>'), and (3) assigns specific tissues to each probeset whose observations are detected to be 'outliers' (see Methods). We compared the performance of ROKU to that of Schug's method, which directly uses the original/non-processed vector <bold><italic>x </italic></bold>for measuring the entropy <italic>H</italic>(<bold><italic>x</italic></bold>) [<xref ref-type="bibr" rid="B4">4</xref>]. The two entropy scores (<italic>H</italic>(<bold><italic>x</italic></bold>') and <italic>H</italic>(<bold><italic>x</italic></bold>)) for all probesets are available in the additional file [see <xref ref-type="supplementary-material" rid="S1">Additional file 1</xref>].</p><p>To compare the agreement of top-ranked probesets between ROKU and Schug's method we analyzed the percentage of common probesets in a top-ranked set of ~22,283 probesets. About 80% of ~3,000 top-ranked probesets are common, indicating that ROKU does not change the rank of probesets drastically (data not shown). One way to compare the effect of the data processing used in ROKU to that used in Schug's method is to sort probesets in order of increasing magnitude by the difference between the two entropy scores (<italic>H</italic>(<bold><italic>x</italic></bold>') - <italic>H</italic>(<bold><italic>x</italic></bold>)) calculated by the two methods. Since ROKU outputs relatively low entropy to each probeset compared to Schug's method as a whole [see <xref ref-type="supplementary-material" rid="S1">Additional file 1</xref>], the average value of (<italic>H</italic>(<bold><italic>x</italic></bold>') - <italic>H</italic>(<bold><italic>x</italic></bold>)) tends to be negative: -0.425 (4.314 for ROKU; 4.739 for Schug's method).</p><p>Table <xref ref-type="table" rid="T1">1</xref> lists the ten lowest- and ten highest (<italic>H</italic>(<bold><italic>x</italic></bold>') - <italic>H</italic>(<bold><italic>x</italic></bold>)) valued probesets and Fig. <xref ref-type="fig" rid="F2">2</xref> shows expression profiles for the two lowest- and two highest probesets listed in Table <xref ref-type="table" rid="T1">1</xref>. The difference is greatest for the probeset '206319_s_at'. This is mainly because the relative expression for the testis changes from 0.35 to 0.75 by virtue of data processing. ROKU gives a low entropy (<italic>H</italic>(<bold><italic>x</italic></bold>') = 1.950 and <italic>H</italic>(<bold><italic>x</italic></bold>') <<italic>H</italic>(<bold><italic>x</italic></bold>)) for the probeset '206319_s_at' and a high entropy (<italic>H</italic>(<bold><italic>x</italic></bold>') = 4.729 and <italic>H</italic>(<bold><italic>x</italic></bold>') > <italic>H</italic>(<bold><italic>x</italic></bold>)) for the probeset '201131_s_at'. This is quite reasonable because visual evaluation admits the former to be tissue-specific and the latter to be non-specific. Schug's method, however, gives quite similar values (4.235 for the former and 4.228 for the latter) for the two probesets: the entropy for the former is higher than that for the latter.</p><p>There are 858 probesets satisfying <italic>H</italic>(<bold><italic>x</italic></bold>') > <italic>H</italic>(<bold><italic>x</italic></bold>): processed expression vectors are less tissue-specific than the original vectors. Visual evaluation for those probesets showed no probeset exists whose entropy score is improperly assigned, i.e., no obvious tissue-specific probesets exist. These results demonstrate the data processing strategy used in ROKU successfully estimated/ranked probesets by their overall tissue specificity on real data. We verified such trends in other microarray datasets (data not shown).</p><p>Note that ROKU is inferior to Schug's method (i.e., direct application of entropy to measuring tissue specificity) in rare cases. For example, consider a gene expression pattern of constant high expression in <italic>N</italic>/2 tissues and low expression in other tissues. ROKU gives the processed expression pattern as 'flat' and <italic>H</italic>(<bold><italic>x</italic></bold>') = log<sub>2</sub>(<italic>N</italic>). Accordingly, ROKU cannot distinguish such differential expression patterns from constant expression patterns because it gives the same entropy scores for the two patterns. In other words, <italic>H</italic>(<bold><italic>x</italic></bold>') is not useful for identifying non-specific genes. Nevertheless, this disadvantage is not a problem for detecting the tissue-specific expression patterns we focused on. We also observed that there was no probeset suffer from this disadvantage in the real data set.</p></sec><sec><title>Detection of specific tissues as outliers</title><p>As mentioned earlier, the entropy does not indicate which tissues are specific though it can rank genes according to their degrees of overall tissue specificity. To identify such specific tissue when they exist, ROKU employs an outlier-detection-based method proposed by Kadota et al. [<xref ref-type="bibr" rid="B3">3</xref>] (see Methods for details). Regardless of over- and/or under-expressed outliers, it can return specific tissues corresponding to outliers for each gene. Accordingly, an outlier matrix can be constructed (consisting of 1 for over-expressed outliers, -1 for under-expressed outliers, and 0 for non-outliers) that corresponds to the original gene expression matrix by applying the method. Genes with any expression pattern of interest can be detected using the outlier matrix. The outlier matrix is also available in the additional file [see <xref ref-type="supplementary-material" rid="S1">Additional file 1</xref>].</p><p>For example, ROKU identifies 59 probesets specific to lung and 291 probesets specific to fetal lung and of course no redundancies exist between the two sets by virtue of the advantage of the original method [<xref ref-type="bibr" rid="B3">3</xref>]. Since ROKU is a combined method consisting of calculation of an entropy and assignment of specific tissues to each gene, ROKU can compensate for the disadvantage of the original method [<xref ref-type="bibr" rid="B3">3</xref>] by assigning an entropy score <italic>H</italic>(<bold><italic>x</italic></bold>'): ROKU can rank genes with particular tissue-specific patterns by their overall tissue specificity. We compared the performance of ROKU to that of Schug's <italic>Q</italic><sub><italic>t</italic></sub>(<bold><italic>x</italic></bold>) statistic [<xref ref-type="bibr" rid="B4">4</xref>] which can also rank genes specific to a tissue <italic>t</italic>.</p><p>Fig. <xref ref-type="fig" rid="F3">3</xref> shows the top-ranked gene expression profiles specific to (a) lung and (b) fetal lung identified by ROKU's <italic>H</italic>(<bold><italic>x</italic></bold>') statistic and Schug's <italic>Q</italic><sub><italic>t</italic></sub>(<bold><italic>x</italic></bold>) statistic [<xref ref-type="bibr" rid="B4">4</xref>]. The <italic>Q</italic><sub><italic>t</italic></sub>(<bold><italic>x</italic></bold>) statistic for a tissue <italic>t </italic>in a gene expression vector <bold><italic>x </italic></bold>is defined as <italic>Q</italic><sub><italic>t</italic></sub>(<bold><italic>x</italic></bold>) = <italic>H</italic>(<bold><italic>x</italic></bold>) - log<sub>2</sub>(<italic>p</italic><sub><italic>t</italic></sub>) (see Methods for details). Clearly, ROKU can detect probesets whose expression patterns are specific only to each of the objective tissues (lung or fetal lung) while Schug's <italic>Q </italic>statistic cannot. This is because a low <italic>Q</italic><sub><italic>t</italic></sub>(<bold><italic>x</italic></bold>) statistic indicates that gene <bold><italic>x </italic></bold>is relatively highly expressed in a small number of tissues including tissue <italic>t</italic>, but does not always indicate whether the expression pattern of <bold><italic>x </italic></bold>is specific only to the tissue <italic>t</italic>. Indeed, both probesets ('215454_x_at' detected as specific to lung and '205982_x_at' specific to fetal lung) identified by Schug's <italic>Q</italic><sub><italic>t</italic></sub>(<bold><italic>x</italic></bold>) statistic include another tissue in addition to the objective tissue. We analyzed this trend in the top-ranking probesets (Table <xref ref-type="table" rid="T2">2</xref>). We assert that these probesets are not specific to lung (or fetal lung) but are specific to both lung and fetal lung. Although the choice of which method should be used is, of course, dependent on individual research purposes, our method (ROKU) is superior to Schug's <italic>Q </italic>statistic for detecting genes specific only to tissues of interest.</p><p>Of 22,283 probesets analyzed, 16,072 exhibit one or more specific tissues. We observed that most of them consist of specific up-expression patterns, such as Figs. <xref ref-type="fig" rid="F1">1c</xref> and <xref ref-type="fig" rid="F1">1d</xref> [see <xref ref-type="supplementary-material" rid="S1">Additional file 1</xref>]. This is probably because the distribution of gene expression levels from the dataset we used here roughly follows an exponential distribution in which the probability of a gene's expression observation decays exponentially (data not shown). Still we appreciate the merit of ROKU being able to detect genes with various types of tissue-specific expression patterns, as shown in Fig. <xref ref-type="fig" rid="F1">1</xref>.</p></sec><sec><title>Effect of different quantification algorithms on gene ranking</title><p>As discussed in Grant et al. [<xref ref-type="bibr" rid="B6">6</xref>], a serious issue regarding any method is the choice of quantification algorithms, such as MAS5 [<xref ref-type="bibr" rid="B7">7</xref>] or RMA [<xref ref-type="bibr" rid="B8">8</xref>]; different choices can output different subsets of top-ranked genes. We compared the influence on gene ranking when the same raw data are MAS5-quantified and RMA-quantified. Fig. <xref ref-type="fig" rid="F4">4</xref> shows the percentages of common probesets in a top-ranked set of ~22,283 probesets between MAS5 data and RMA data, by gene ranking using ROKU (red circle) and Schug's method (black circle). Although both methods (ROKU and Schug's method) output relatively low percentages of common probesets, especially in the 100 top-ranked probesets (about 31% for ROKU; about 3% for Schug's method), the percentages for ROKU were consistently higher than those for Schug's method. This result indicates gene ranking based on ROKU is more robust against data transformation than Schug's method.</p><p>There are some ways for extending this work. First, we used an outlier-detection-based method [<xref ref-type="bibr" rid="B3">3</xref>] for the detection of specific tissues. Some other methods, such as Sprent's non-parametric method [<xref ref-type="bibr" rid="B5">5</xref>] and its derivative, could be applicable. The outputs of these methods vary with the selected parameters. A comparative study of these methods is the next task. Second, the current work did not discuss the statistical significance of observed differences in gene expression. We plan to combine the significance analysis, such as a method of Sharov et al. [<xref ref-type="bibr" rid="B9">9</xref>], with the current method.</p></sec></sec><sec><title>Conclusion</title><p>In this work, we propose a novel method (ROKU) for the detection of genes with tissue-specific expression patterns. ROKU was developed to compensate for the disadvantages of two conventional methods [<xref ref-type="bibr" rid="B3">3</xref>,<xref ref-type="bibr" rid="B4">4</xref>] by combining the advantages of the two. Using synthetic expression data, we demonstrated its potential applicability for the detection of various types of specific expression patterns. Although most of the detected tissue-specific genes in real microarray data exhibit one type of expression pattern (i.e., 'up-type' genes selectively over-expressed in a single or small number of tissues compared to the others), the entropy-based gene ranking by ROKU outperforms the two original methods. ROKU can be a powerful tool for selecting genes specific to tissues of interest.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>Microarray data</title><p>Publicly available Affymetrix U133A oligonucleotide microarray data for 22,283 genes in 36 various normal human tissues [<xref ref-type="bibr" rid="B5">5</xref>] were downloaded from the author's website [<xref ref-type="bibr" rid="B10">10</xref>]. For the most part we used the data quantified using MAS5 (Micro Array Suite 5 from Affymetrix) software. Other quantified data using the RMA algorithm [<xref ref-type="bibr" rid="B8">8</xref>] were also analyzed to compare the effects of different quantification algorithms. RMA quantification was performed by the justRMA() function in R [<xref ref-type="bibr" rid="B11">11</xref>] using raw data (Affymetrix CEL files).</p></sec><sec><title>Gene ranking by Shannon entropy</title><p>The use of Shannon entropy [<xref ref-type="bibr" rid="B12">12</xref>] to rank genes by their overall tissue specificity here is the same as described in Schug et al. [<xref ref-type="bibr" rid="B4">4</xref>]. Consider one gene's expression vector <bold><italic>x </italic></bold>= (<italic>x</italic><sub><italic>1</italic></sub>, <italic>x</italic><sub><italic>2</italic></sub>, ..., <italic>x</italic><sub><italic>N</italic></sub>) for <italic>N </italic>tissues and an observation <italic>x</italic><sub><italic>t </italic></sub>for tissue <italic>t</italic>. The entropy of the gene is calculated as</p><p><inline-formula><mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="M1" name="1471-2105-7-294-i1" overflow="scroll">
<mml:semantics definitionURL="" encoding="">
<mml:mrow>
<mml:mi>H</mml:mi>
<mml:mo>=</mml:mo>
<mml:mo>−</mml:mo>
<mml:mstyle displaystyle="true">
<mml:msubsup>
<mml:mo>∑</mml:mo>
<mml:mrow>
<mml:mi>t</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:mi>p</mml:mi>
<mml:mi>t</mml:mi>
</mml:msub>
</mml:mrow>
</mml:mstyle>
<mml:msub>
<mml:mrow>
<mml:mi>log</mml:mi>
<mml:mo>⁡</mml:mo>
</mml:mrow>
<mml:mn>2</mml:mn>
</mml:msub>
<mml:mo stretchy="false">(</mml:mo>
<mml:msub>
<mml:mi>p</mml:mi>
<mml:mi>t</mml:mi>
</mml:msub>
<mml:mo stretchy="false">)</mml:mo>
<mml:mo>,</mml:mo>
<mml:mtext>     </mml:mtext>
<mml:mrow>
<mml:mo>(</mml:mo>
<mml:mn>1</mml:mn>
<mml:mo>)</mml:mo>
</mml:mrow>
</mml:mrow>
<mml:annotation encoding="MathType-MTEF">
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGibascqGH9aqpcqGHsisldaaeWaqaaiabdchaWnaaBaaaleaacqWG0baDaeqaaaqaaiabdsha0jabg2da9iabigdaXaqaaiabd6eaobqdcqGHris5aOGagiiBaWMaei4Ba8Maei4zaC2aaSbaaSqaaiabikdaYaqabaGccqGGOaakcqWGWbaCdaWgaaWcbaGaemiDaqhabeaakiabcMcaPiabcYcaSiaaxMaacaWLjaWaaeWaaeaacqaIXaqmaiaawIcacaGLPaaaaaa@47F0@</mml:annotation>
</mml:semantics>
</mml:math></inline-formula></p><p>where <italic>p</italic><sub><italic>t </italic></sub>is the relative expression of <italic>x</italic><sub><italic>t </italic></sub>for tissue <italic>t </italic>defined as <inline-formula><mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="M2" name="1471-2105-7-294-i2" overflow="scroll"><mml:semantics definitionURL="" encoding=""><mml:mrow><mml:msub><mml:mi>p</mml:mi><mml:mi>t</mml:mi></mml:msub><mml:mo>=</mml:mo><mml:msub><mml:mi>x</mml:mi><mml:mi>t</mml:mi></mml:msub><mml:mo>/</mml:mo><mml:mstyle displaystyle="true"><mml:msubsup><mml:mo>∑</mml:mo><mml:mrow><mml:mi>t</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:mi>x</mml:mi><mml:mi>t</mml:mi></mml:msub></mml:mrow></mml:mstyle></mml:mrow><mml:annotation encoding="MathType-MTEF">
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGWbaCdaWgaaWcbaGaemiDaqhabeaakiabg2da9iabdIha4naaBaaaleaacqWG0baDaeqaaOGaei4la8YaaabmaeaacqWG4baEdaWgaaWcbaGaemiDaqhabeaaaeaacqWG0baDcqGH9aqpcqaIXaqmaeaacqWGobGta0GaeyyeIuoaaaa@3E61@</mml:annotation></mml:semantics></mml:math></inline-formula>. <italic>H </italic>ranges from zero to log<sub>2</sub>(<italic>N</italic>), with the value 0 for genes expressed in a single tissue (Fig. <xref ref-type="fig" rid="F1">1a</xref>) and log<sub>2</sub>(<italic>N</italic>) for genes expressed uniformly in all the interrogated tissues (Fig. <xref ref-type="fig" rid="F1">1b</xref>).</p><p>To equally identify down- and mixed-types of tissue-specific genes as well as up-type genes, we processed the original vector <bold><italic>x</italic></bold>. The processed observation <italic>x</italic><sub><italic>t</italic></sub>' for tissue <italic>t </italic>is defined as</p><p><inline-formula><mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="M3" name="1471-2105-7-294-i3" overflow="scroll"><mml:semantics definitionURL="" encoding=""><mml:mrow><mml:msub><mml:msup><mml:mi>x</mml:mi><mml:mo>′</mml:mo></mml:msup><mml:mi>t</mml:mi></mml:msub></mml:mrow><mml:annotation encoding="MathType-MTEF">
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacuWG4baEgaqbamaaBaaaleaacqWG0baDaeqaaaaa@2FCE@</mml:annotation></mml:semantics></mml:math></inline-formula> = |<italic>x</italic><sub><italic>t </italic></sub>- <italic>T</italic><sub><italic>bw</italic></sub>|,     (2)</p><p>where <italic>T</italic><sub><italic>bw </italic></sub>is the one-step Tukey biweight, a popular statistic robust against outliers. It provides as much robustness as a median and is also used to estimate the expression signal from each probe set in the Affymetrix Micro Array Suite (MAS 5.0) software package [<xref ref-type="bibr" rid="B7">7</xref>,<xref ref-type="bibr" rid="B13">13</xref>]. The parameters for the calculation of <italic>T</italic><sub><italic>bw </italic></sub>are the same as those adopted in the tukey.biweight() function in R package 'affy' (i.e., <italic>c </italic>= 5, ε = 0.0001) [<xref ref-type="bibr" rid="B11">11</xref>]. Our method (ROKU) uses the processed expression vector <bold><italic>x</italic></bold>' of a gene, while Schug et al. [<xref ref-type="bibr" rid="B4">4</xref>] uses the original vector <bold><italic>x</italic></bold>, to calculate the gene's entropy (<italic>H</italic>(<bold><italic>x</italic></bold>') and <italic>H</italic>(<bold><italic>x</italic></bold>)) as a measure of the overall tissue specificity.</p></sec><sec><title>Detecting specific tissues as outliers</title><p>As mentioned above, the entropy does not indicate which tissues are specific, but is a measure of the overall tissue specificity of a gene. We imagine observations in specific tissues to be easily visualized as outliers on the over- and/or under-expressed side if any exist. We used an outlier-detection-based method proposed by Kadota et al. [<xref ref-type="bibr" rid="B3">3</xref>] to detect tissues with specific expression patterns. According to Kadota et al. [<xref ref-type="bibr" rid="B3">3</xref>], the statistic <italic>U </italic>for identifying outliers is defined as</p><p><inline-formula><mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="M4" name="1471-2105-7-294-i4" overflow="scroll">
<mml:semantics definitionURL="" encoding="">
<mml:mrow>
<mml:mi>U</mml:mi>
<mml:mo>=</mml:mo>
<mml:mi>n</mml:mi>
<mml:mi>log</mml:mi>
<mml:mo>⁡</mml:mo>
<mml:mi>σ</mml:mi>
<mml:mo>+</mml:mo>
<mml:msqrt>
<mml:mn>2</mml:mn>
</mml:msqrt>
<mml:mo>×</mml:mo>
<mml:mi>s</mml:mi>
<mml:mo>×</mml:mo>
<mml:mfrac>
<mml:mrow>
<mml:mi>log</mml:mi>
<mml:mo>⁡</mml:mo>
<mml:mi>n</mml:mi>
<mml:mo>!</mml:mo>
</mml:mrow>
<mml:mi>n</mml:mi>
</mml:mfrac>
<mml:mo>,</mml:mo>
<mml:mtext>     </mml:mtext>
<mml:mrow>
<mml:mo>(</mml:mo>
<mml:mn>3</mml:mn>
<mml:mo>)</mml:mo>
</mml:mrow>
</mml:mrow>
<mml:annotation encoding="MathType-MTEF">
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGvbqvcqGH9aqpcqWGUbGBcyGGSbaBcqGGVbWBcqGGNbWzcqaHdpWCcqGHRaWkdaGcaaqaaiabikdaYaWcbeaakiabgEna0kabdohaZjabgEna0oaalaaabaGagiiBaWMaei4Ba8Maei4zaCMaemOBa4MaeiyiaecabaGaemOBa4gaaiabcYcaSiaaxMaacaWLjaWaaeWaaeaacqaIZaWmaiaawIcacaGLPaaaaaa@4A24@</mml:annotation>
</mml:semantics>
</mml:math></inline-formula></p><p>where <italic>n </italic>and <italic>s </italic>denote the numbers of non-outlier and outlier candidates, and σ denotes the standard deviation (SD) of the observations of the <italic>n </italic>non-outlier candidates. The procedure is first, normalize the gene vector <bold><italic>x </italic></bold>= (<italic>x</italic><sub>1</sub>, <italic>x</italic><sub>2</sub>, ..., <italic>x</italic><sub><italic>N</italic></sub>) for <italic>N </italic>(=<italic>n</italic>+<italic>s</italic>) tissues by subtracting the mean and dividing by the SD; second, sort the normalized values (i.e., <italic>Z</italic>-scores) by order; third, calculate the statistics <italic>U </italic>for various combinations of outlier candidates starting from both sides of the values; finally, regard tissues corresponding to outliers detected in the combination of the minimum <italic>U </italic>as 'specific'.</p></sec></sec><sec><title>Authors' contributions</title><p>KK invented the method and wrote the paper. JY made critical comments in light of the current algorithm. YN, TT, and KS provided critical comments and led the project.</p></sec><sec sec-type="supplementary-material"><title>Supplementary Material</title><supplementary-material content-type="local-data" id="S1"><caption><title>Additional file 1</title><p><bold>Full information analyzed by ROKU for dataset of Ge et al. (2005)</bold>. For the original gene expression matrix, an outlier matrix (consisting of 1 for over-expressed outliers, -1 for under-expressed outliers, and 0 for non-outliers) is provided. It also contains two entropy scores measured by ROKU and Schug's method and their ranks.</p></caption><media xlink:href="1471-2105-7-294-S1.xls" mimetype="application" mime-subtype="vnd.ms-excel"><caption><p>Click here for file</p></caption></media></supplementary-material></sec> |
Extraction of phylogenetic network modules from the metabolic network | <sec><title>Background</title><p>In bio-systems, genes, proteins and compounds are related to each other, thus forming complex networks. Although each organism has its individual network, some organisms contain common sub-networks based on function. Given a certain sub-network, the distribution of organisms common to it represents the diversity of its function.</p></sec><sec><title>Results</title><p>We extracted such "common" sub-networks, defined as "phylogenetic network modules," using phylogenetic profiles and cluster analysis. The enzymes in the same "phylogenetic network module" have similar phylogenetic profiles and related functions. These modules are shown to be phylogenetic building blocks. Furthermore, the network of the modules illustrated hierarchical feature as well as the network of enzymes involved in the metabolism.</p></sec><sec><title>Conclusion</title><p>We conclude that phylogenetic network modules are evolutionary conserved functional units in the metabolic network. We claim that our concept of phylogenetic modules provides a more accurate understanding of the evolution of biological networks.</p></sec> | <contrib id="A1" contrib-type="author"><name><surname>Yamada</surname><given-names>Takuji</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>takuji@kuicr.kyoto-u.ac.jp</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Kanehisa</surname><given-names>Minoru</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>kanehisa@kuicr.kyoto-u.ac.jp</email></contrib><contrib id="A3" corresp="yes" contrib-type="author"><name><surname>Goto</surname><given-names>Susumu</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>goto@kuicr.kyoto-u.ac.jp</email></contrib> | BMC Bioinformatics | <sec><title>Background</title><p>Bio-systems are constructed by various relationships between genes, compounds and proteins. These elements are intertwined, thus forming a complex network. Such genetic or physical associations contribute to the complicated cellular functions.</p><p>In order to characterize whole networks or to find unknown interactions between elements (gene/compound/protein), many research groups have applied the integration of qualitatively different interactions. For example, Rison et al attempted to find unknown functions of genes using genome location, sequence similarity, and relative position in the metabolic network [<xref ref-type="bibr" rid="B1">1</xref>], and Yamanishi <italic>et al </italic>similarly attempted to apply a kernel method [<xref ref-type="bibr" rid="B2">2</xref>]. Such integrative methods infer common features and new categories of genes. Classically, such new categories were determined manually using expert knowledge. The KEGG metabolic pathway [<xref ref-type="bibr" rid="B3">3</xref>] is one such example. On the other hand, recent studies have attempted to define new categories automatically using comprehensive data sets such as microarray data and metabolic networks [<xref ref-type="bibr" rid="B4">4</xref>]. Such analyses can classify elements by globally assessing many relationships. Many research groups advocate various functional categories of genes and proteins, such as transcriptional clusters [<xref ref-type="bibr" rid="B5">5</xref>,<xref ref-type="bibr" rid="B6">6</xref>], protein complexes [<xref ref-type="bibr" rid="B7">7</xref>], metabolic network [<xref ref-type="bibr" rid="B3">3</xref>,<xref ref-type="bibr" rid="B8">8</xref>], and others [<xref ref-type="bibr" rid="B9">9</xref>,<xref ref-type="bibr" rid="B10">10</xref>].</p><p>From the perspective of graph theory, it is suggested that networks composed of genes and proteins should have modularity. Ravasz <italic>et al </italic>suggested a hierarchical network model incorporating modularity in the metabolic network [<xref ref-type="bibr" rid="B11">11</xref>]. Following this work, studies were attempted using the modularity in various networks and topological features [<xref ref-type="bibr" rid="B12">12</xref>]. However, these topological investigations were not based on phylogeny; Snel <italic>et al </italic>pointed out problems in which they found conflicts between evolution and function [<xref ref-type="bibr" rid="B13">13</xref>].</p><p>In order to resolve these conflicts, the main purpose of our research is the extraction of the phylogenetic primary units of the metabolic network for understanding the evolutionary process. To our knowledge, we were the first to attempt to extract evolution-based enzyme modules. In our previous work, we attempted to extract phylogenetic units using prokaryotic genomes [<xref ref-type="bibr" rid="B14">14</xref>]. Here, we expanded the analysis by adding eukaryotic data, and by considering the topological properties of the network of phylogenetic units.</p><p>The metabolic network was considered not only to be composed of several particular categories, but also to be a large network consisting of chemical compounds connected by enzymes. For example, KEGG/PATHWAY is a collection of manually drawn pathway maps. However, we can regard all of these pathway maps as a "metabolic network." Our analysis is based on the KEGG/PATHWAY database, which integrates the metabolic network from many organisms. Therefore, in this paper, we use the term "metabolic network" to refer to this integrated metabolic network. Therefore, some parts of the metabolic network are found only in particular organisms, and the number of organisms varies depending on the sub-networks. The phylogenetic profile is very useful to deal with such diversity of enzymes. The phylogenetic profile of an enzyme is the string that encodes the presence or absence of the enzyme in the fully sequenced genome [<xref ref-type="bibr" rid="B15">15</xref>].</p><p>Using the enzyme connectivity in the metabolic network and the similarity between the phylogenetic profiles of the enzymes, we extracted enzyme modules and define them as "phylogenetic network modules." The enzymes in the same phylogenetic network module have similar phylogenetic profiles and are located close to one another in the metabolic network. In other words, these enzymes behave in a similar way in the evolutionary process of the metabolic network. Furthermore, we found a hierarchy of phylogenetic network modules. The method overview is summarized in Figure <xref ref-type="fig" rid="F1">1</xref>. Our result fits the concept of the network model of Ravasz <italic>et al </italic>[<xref ref-type="bibr" rid="B11">11</xref>]. Our phylogenetic network modules are based on the simultaneous behaviour of multiple enzymes in the evolutionary process of the metabolic network. This allows for the detailed understanding of metabolic network evolution.</p></sec><sec><title>Results</title><sec><title>Pathway distance and the similarity of phylogenetic profiles</title><p>Figure <xref ref-type="fig" rid="F2">2-A</xref> illustrates the negative correlation between pathway distance and the similarity of phylogenetic profiles (white points). We use the Jaccard coefficient (JC) as the similarity score of phylogenetic profiles of enzymes, and pathway distance is defined as the smallest number of steps between two enzymes in the metabolic network. The figure shows that the longer the pathway distance, the smaller the average similarity. In fact, two adjacent enzymes in the network have the highest average similarity score. We generated randomly relabeled networks 100 times, and plotted their average for comparison (see Method section). Contrary to the actual metabolic network, no pattern was observed in the JC average of relabeled networks (black points in Fig. <xref ref-type="fig" rid="F2">2-A</xref>). The JC average of relabeled networks is constant approximately 0.14, which is equal to the average similarity score when comparing all profiles against each other. We also performed the same analysis using correlation coefficient (CC) as a different similarity measure, which resulted in a similar trend to that of JC (see <xref ref-type="supplementary-material" rid="S1">additional file 1</xref>).</p></sec><sec><title>Distribution of the number of enzymes included in a module</title><p>The white points in Figure <xref ref-type="fig" rid="F2">2-B</xref> indicate the frequency of the "phylogenetic network modules" according to their sizes (the number of enzymes). The horizontal axis represents the number of enzymes in a module, and the vertical axis represents its frequency. In total, 1179 modules were extracted, but over 900 modules contain only one enzyme. There are a few large modules, and many small modules. Similar results were obtained using different distance measures such as correlation coefficient and Hamming distance (data not shown).</p><p>Furthermore, in order to prove that the enzymes form module structures in the metabolic network, we compared the frequency with the average of 100 randomly relabeled networks, shown as the black points in Figure <xref ref-type="fig" rid="F2">2-B</xref>. The difference of slope indicates that larger modules tend to appear more frequently in actual networks; that is, they hardly appear in the average of relabeled networks. This implies that enzymes with similar phylogenetic profiles tend to aggregate in the network, and that phylogenetic module structures are indeed present in the metabolic network.</p></sec><sec><title>Comparing modules with the categories in the KEGG/PATHWAY database</title><p>KEGG/PATHWAY is a collection of manually drawn diagrams called the KEGG reference pathway diagrams (maps), each corresponding to a known network of functional significance. These diagrams are drawn manually representing particular functions of biological processes. We mapped our phylogenetic network modules onto these diagrams.</p><p>The upper portion of Figure <xref ref-type="fig" rid="F3">3</xref> illustrates an example of mapping phylogenetic network modules onto a diagram (Lysine biosynthesis). Gray colored enzymes are contained in the organisms in the KEGG database. Enzymes surrounded by a solid line represent a part of a particular phylogenetic network module, each corresponding to different phylogenetic and functional features.</p><p>Module 1 is contained in many organisms, spreading to several amino acid biosynthesis pathways such as histidine metabolism (lower part of Figure <xref ref-type="fig" rid="F3">3</xref>). Module 2 is specialized to prokaryotes, so it links to the prokaryote-specific Peptidoglycan biosynthesis pathway. Module 3 is also contained in relatively many organisms, and module 4 is specialized for eukaryotes. Almost all of the diagrams in KEGG/PATHWAY include multiple phylogenetic network modules, and some modules spread to many diagrams. In fact, the lysine biosynthesis pathway contains several modules, and module 1 and 2 spread to other diagrams. Thus, using our phylogenetic network modules, we could detect phylogenetic relationships between known functional categories (across pathways).</p></sec><sec><title>Global network of phylogenetic network modules</title><p>Figure <xref ref-type="fig" rid="F4">4</xref> illustrates the largest component of our global module network which includes 1130 of the 1179 modules. This is the reconstruction of a whole metabolic network by our phylogenetic network modules, so each node corresponds to a module, and each edge corresponds to the relationship between them. The size of a node indicates the number of enzymes included in the module, and its color indicates the number of organisms associated with the module.</p><p>Three features are conspicuous in this figure. The first is a large central module. All of the modules seem to gather around the largest red module as a network core. The big module contains 51 enzymes that spread to a large part of nucleotide metabolism, and a part of amino acid metabolism. Furthermore, this is contained in almost all organisms so that it is not surprising that the module constructs a core of the metabolic network. This module connects other modules with each other and sustains the whole network. Additionally, relatively large modules are also observed in parts of the metabolic network: some glycan related pathway diagrams for eukaryotes and a part of peptidoglycan biosynthesis for bacteria.</p><p>As a second feature, modules also aggregate. In Figure <xref ref-type="fig" rid="F4">4</xref>, some of the modules tend to assemble together into clusters. The dotted circles in the figure indicate the locations of what we call "super-modules," which are modules that are relatively aggregated. The relationship between modules in this figure is based on the connectivity of existing enzymes in the metabolic network, so the "super-modules" represent modules of related functions even though some spread to multiple pathway diagrams (only representative diagrams are annotated in Figure <xref ref-type="fig" rid="F4">4</xref>). For example, the super module structure just above the network core (the biggest red module) in Figure <xref ref-type="fig" rid="F4">4</xref> is composed of many diagrams of Amino acid metabolism, and it contains many links to other modules. Basically, the enzymes in this super-module are highly conserved and are closely related to those in the network core. Thus, we claim that enzymes in these modules use many metabolites synthesized in the core structure and that they also supply many metabolites to the other modules. The third feature is that "linker" modules are scattered around the network. As a topological feature, they have relatively low clustering coefficients compared to the other modules with the same degree (see Methods section). Therefore they do not belong to any particular module cluster, and instead, link module clusters to each other. We found 30 linker modules by the criterion defined in the Methods section. Biologically, linker modules tend to be intermediates for the input and output compounds of functional modules. For example, the linker module indicated by the right arrow in Figure <xref ref-type="fig" rid="F4">4</xref> connects Amino Acid Metabolism, Lipid Metabolism, as well as a few others. This module is composed of three enzymes (3-hydoroxyacyl-CoA dehydrogenase, acyl-CoA dehydorogenase and enoyl-CoA hydrase) that catalyze reactions between acetoacetyl-CoA and crotonyl-CoA. This reaction chain produces acetyl-CoA in Amino Acid Metabolism and consumes it in Lipid Metabolism, thus linking these two pathways together. As another example, the module indicated by the left arrow in Figure <xref ref-type="fig" rid="F4">4</xref> contains three enzymes, glucose-6-phosphate isomerase, phosphoglucomutase and glucose-6-phosphate 1-dehydrogenase. Although these enzymes play roles in Glycolysis, they connect compounds to other Carbohydrate Metabolism pathways such as Aminosugars metabolism and Fructose and mannose metabolism.</p></sec></sec><sec><title>Discussion</title><sec><title>Functionally related phylogenetic module</title><p>Known functional modules do not completely coincide with evolutionary modules. Snel <italic>et al</italic>. investigated whether known functional modules are also evolutionary modules and suggested that all the members of the same functional module do not have co-evolutionary tendencies [<xref ref-type="bibr" rid="B13">13</xref>]. This means that evolutionary modules are not in complete agreement with functional modules.</p><p>Basically, we agree with Snel's opinion because of the difference of enzyme distributions in the phylogenetic network modules due to the addition of enzyme connectivity in the metabolic network. Before adding information of enzyme connectivity, enzyme clusters were constructed using only the similarity between phylogenetic profiles (Fig. <xref ref-type="fig" rid="F1">1B</xref>). After that, enzymes were re-clustered within each cluster using connectivity in the metabolic network (Fig. <xref ref-type="fig" rid="F1">1C,D,E</xref>). We call these sub-clusters phylogenetic network modules. Obviously, the number of enzymes in a phylogenetic network module was less than those in the original clusters. It is clear that enzyme connectivity in the metabolic network subdivides the enzyme clusters. This explains precisely the conflict between functional modules and phylogenetic modules. By definition, the phylogenetic network modules are the evolutionarily conserved and functionally related enzyme modules in the metabolic pathway. As a result, we claim that the phylogenetic network module is the basic functional unit in the metabolic pathway.</p></sec><sec><title>Modularity and hierarchy in the evolutionary process of the metabolic network</title><p>Historically, there have been many analyses of network evolution [<xref ref-type="bibr" rid="B16">16</xref>], which has led to advanced theories on network evolution. Two major assumptions are generally thought to be the main contenders. One is a retrograde model [<xref ref-type="bibr" rid="B17">17</xref>], and the other is a patchwork model [<xref ref-type="bibr" rid="B18">18</xref>]. In the retrograde model, network evolves "backwards" from a key metabolite. This model expands the network due to the acquirement of new enzymes, which synthesize a molecule used up in the environment from other molecules. On the other hand, in the patchwork model, network evolution is based on the concept that enzymes exhibit broad substrate specificity and catalyze multiple reactions. Such enzymes with broad specificity form reaction chains to a key metabolite. The important thing is that those two models are not mutually exclusive, and they are reviewed in [<xref ref-type="bibr" rid="B19">19</xref>].</p><p>In any case, these evolutionary models are based on the relationship between enzymes and substrates (metabolites). In this paper, phylogenetic network module corresponds to one function which is a group of chemical reactions catalyzing metabolites into others. Correspondingly, we consider the enzyme module as an extended enzyme function in these models. We claim that the concept of representing multiple nodes as a single node, the enzyme module in this case, is important for understanding the evolutionary process of metabolic network. A similar concept for the network integrating other types of relations was proposed in [<xref ref-type="bibr" rid="B20">20</xref>]. Our methodology is based on this concept, and our results support the utility of it.</p><p>It is well known that the metabolic network is a hierarchical network. When a particular network has a hierarchical feature, its plot (where the vertical axis is the clustering coefficient (C(k)) and the horizontal axis is the node's degree) gives a power-law distribution (γ = -1) [<xref ref-type="bibr" rid="B11">11</xref>]. In this paper, we constructed a network of phylogenetic network modules, which we plotted. Interestingly, it had just such modular and hierarchical features as illustrated in Figure <xref ref-type="fig" rid="F5">5</xref>. This indicates that our notion of evolutionarily and functionally conserved modules explains the hierarchical features suggested by Ravasz <italic>et al </italic>[<xref ref-type="bibr" rid="B11">11</xref>], who illustrated the hierarchical structure of the metabolic network using topological properties. Our result suggests that the relationship between enzymes based on the similarity of phylogenetic profiles is one of the factors forming the hierarchical structure.</p></sec><sec><title>Future direction and perspective on network modularity</title><p>We hierarchically clustered enzymes to extract phylogenetic network modules. According to this method, enzymes are allocated to a particular module. However, there are some cases where enzymes could belong to multiple functional modules. For example, there are many enzymes catalyzing reactions related with pyruvate or acetyle-CoA. These enzymes have numerous relationships to other enzymes in the metabolic network. In such cases, the phylogenetic relationships between them are complicated, and enzyme allocation to a particular module is difficult. While these enzymes play an important role, it is difficult to divide into a particular module. Given this problem, it may be important to somehow allow enzyme redundancy in the modules, or remove these enzymes from the network. In a module network, linker modules correspond to this case. These modules have a characteristic feature of connecting many modules (module groups), so it will be difficult to determine particular module groups for linker modules.</p><p>Although our work focused on enzymes, the components of the metabolic network consist of both enzymes and chemical compounds. There have been a few attempts to investigate the relationships between chemical compounds and phylogeny. Hattori <italic>et al </italic>defined the similarity between compounds in the metabolic network, and generated clusters according to this similarity measure [<xref ref-type="bibr" rid="B21">21</xref>]. They also attempted to compare these clusters with operon like structures. However, since the operon data they used was very limited, it was insufficient to attempt phylogenetic analysis. Our phylogenetic network modules are suitable for the analysis, and comprehensive analysis of chemical structures and their evolution is our next research focus.</p></sec></sec><sec><title>Conclusion</title><p>We extracted "phylogenetic network modules" from the metabolic network. We claim that these modules are the evolutionary building blocks as well as the basic functional units of the metabolic network. Furthermore, we showed that the module network has the hierarchical character, which is also conserved in the enzyme network of metabolism.</p><p>Barabasi <italic>et al </italic>illustrated the hierarchical structure of the metabolic network using topological property. Our result suggests that the relationship between enzymes based on the similarity of phylogenetic profiles is one of the factors forming the hierarchical structure. Other biological systems, such as protein-protein interaction networks, have been reported to have the similar topological property. Thus, the phylogenetic relationship may be a foundation of network evolution including other biological systems.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>Construction of the phylogenetic profiles and the network of enzymes</title><p>A phylogenetic profile is a bit string that encodes the absence (0) or presence (1) of an enzyme in fully sequenced genomes. We utilized KEGG Orthology (KO) for constructing the phylogenetic profiles. KO is a database of ortholog groups that are defined manually according to the similarity of amino acid sequences, as well as bidirectional best hit information in pairwise genome comparison, and annotated functions of genes in KEGG pathways [<xref ref-type="bibr" rid="B3">3</xref>]. Therefore, some proteins with low sequence similarity may be put into the same ortholog group (the same node in the KEGG pathway). In this study, we used KO entries for enzyme genes annotated in the KEGG/PATHWAY database. Phylogenetic profiles were calculated using the KO entries constructed from 174 fully sequenced genomes. To reduce the effect of bias in the organism distribution, these 174 genomes were merged into 59 taxa (11 eukaryotes, 36 bacteria and 12 archaea) according to the NCBI taxonomy [<xref ref-type="bibr" rid="B22">22</xref>]. That is, logical OR was applied on the phylogenetic profiles of the organisms of the same taxa. As a result, we obtained 1672 phylogenetic profiles consisting of 59 bits.</p><p>The KEGG/PATHWAY database stores metabolic and regulatory pathway information with their functional classification as XML files [<xref ref-type="bibr" rid="B3">3</xref>]. The pathway information (the network of enzymes) is described as a collection of binary relationships of enzymes. The binary relationship is defined when the two enzymes are adjacently located on the pathway diagram. We extracted the binary relationships of the enzymes from the KEGG/PATHWAY database and treated them as a network or a graph of enzymes. All the data is accessible from the KEGG website [<xref ref-type="bibr" rid="B23">23</xref>].</p></sec><sec><title>Similarity measure between phylogenetic profiles</title><p>Similarity measures between phylogenetic profiles are required in cluster analysis. We adopted the Jaccard coefficient (JC) [<xref ref-type="bibr" rid="B24">24</xref>] and the correlation coefficient (CC) as the similarity measures of phylogenetic profiles. The Jaccard coefficient between profile <italic>A </italic>and <italic>B </italic>is defined as <italic>A </italic>∩ <italic>B/A </italic>∪ <italic>B. A </italic>∩ <italic>B </italic>is the number of organisms which include enzyme <italic>A </italic>and <italic>B</italic>, and <italic>A </italic>∪ <italic>B </italic>is the number of organisms which include either enzyme A or B. The correlation coefficient used was Pearson's product moment correlation coefficient using a phylogenetic profile as a vector.</p></sec><sec><title>Pathway distance and the corresponding average of Jaccard coefficient</title><p>Pathway distance is defined as the smallest number of steps between two enzymes in the metabolic network. We calculated the pathway distance of all against all enzymes in KEGG/PATHWAY and further calculated the JC average for each pathway distance. We also generated 100 relabeled enzyme networks as a control. In this paper, we define a relabeled network as one where the phylogenetic profiles are randomly re-assigned to enzymes in the network with the topology unchanged. The JC average of each pathway distance was calculated for each relabeled network. The average was then taken of all the JC averages of each relabeled network.</p></sec><sec><title>Extraction of the "phylogenetic network module"</title><p>An overview of the extraction of the "phylogenetic network module" is illustrated in Figure <xref ref-type="fig" rid="F1">1</xref>. We performed a complete linkage clustering of the enzymes (KO entries) in our dataset (Fig. <xref ref-type="fig" rid="F1">1B</xref>) based on their phylogenetic profile similarities. To determine the threshold of this clustering, we utilized the distribution of all-against-all similarities between the phylogenetic profiles of enzymes. From the distribution, we estimated the top 1, 2.5 and 5 percentile points of the similarities as the significant point (each JC corresponds to 0.76, 0.64 and 0.51 respectively, see additional files <xref ref-type="supplementary-material" rid="S2">2</xref>-<xref ref-type="supplementary-material" rid="S3">3</xref>). After that, by linking the enzymes according to their connection in the metabolic pathway (Fig. <xref ref-type="fig" rid="F1">1C</xref>), some "phylogenetic network modules" were extracted in each cluster (Fig. <xref ref-type="fig" rid="F1">1D</xref>). That is, each group of linked enzymes within each cluster was extracted as a single module. There was no difference in the distribution shape of the module sizes for different significant points. All figures in this paper are based on the 5 percentile point. The same operations were applied to the 100 relabeled networks, and their distribution of the average number of enzymes per module was obtained.</p></sec><sec><title>Network of modules</title><p>Finally, we constructed a network of modules, where each node represents a module, and each edge represents the chemical compounds between the enzymes contained in the corresponding modules (Fig. <xref ref-type="fig" rid="F1">1E</xref>). To infer the structure of the module network, we investigated the relationship between each node's degree and its clustering coefficient. The degree of node <italic>i </italic>is the number of connected nodes. The clustering coefficient of node <italic>i </italic>is defined as <italic>C</italic><sub><italic>i </italic></sub>= 2<italic>n</italic><sub><italic>i</italic></sub>/<italic>k</italic><sub><italic>i </italic></sub>(<italic>k</italic><sub><italic>i</italic></sub><italic>-l</italic>), where <italic>n</italic><sub><italic>i </italic></sub>denotes the number of links connecting the <italic>k</italic><sub><italic>i </italic></sub>neighbors of node <italic>i </italic>to each other.</p><p>Based on the clustering coefficient and the degree, linker modules were selected. These linker modules tend to have high degree and relatively low clustering coefficients compared with other modules of the same degree. We defined the condition of the linker as having degree > 35, and the deviation of the clustering coefficient of the same degree being < -0.1.</p></sec></sec><sec><title>Authors' contributions</title><p>TY and SG contributed to develop methodology and to assess the biological significance of the results. MK supervised the project. All authors read and approved the final 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</title><p>The comparison of Jaccard coefficient with correlation coefficient as the similarity measure of the phylogenetic profile.</p></caption><media xlink:href="1471-2105-7-130-S1.pdf" mimetype="application" mime-subtype="pdf"><caption><p>Click here for file</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="S2"><caption><title>Additional File 2</title><p>The distribution of the number of enzymes in a phylogenetic network module using Jaccard coefficient with three different thresholds.</p></caption><media xlink:href="1471-2105-7-130-S2.pdf" mimetype="application" mime-subtype="pdf"><caption><p>Click here for file</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="S3"><caption><title>Additional File 3</title><p>The distribution of the number of enzymes in a phylogenetic network module using Correlation coefficient with three different thresholds.</p></caption><media xlink:href="1471-2105-7-130-S3.pdf" mimetype="application" mime-subtype="pdf"><caption><p>Click here for file</p></caption></media></supplementary-material></sec> |
Bio++: a set of C++ libraries for sequence analysis, phylogenetics, molecular evolution and population genetics | <sec><title>Background</title><p>A large number of bioinformatics applications in the fields of bio-sequence analysis, molecular evolution and population genetics typically share input/ouput methods, data storage requirements and data analysis algorithms. Such common features may be conveniently bundled into re-usable libraries, which enable the rapid development of new methods and robust applications.</p></sec><sec><title>Results</title><p>We present Bio++, a set of Object Oriented libraries written in C++. Available components include classes for data storage and handling (nucleotide/amino-acid/codon sequences, trees, distance matrices, population genetics datasets), various input/output formats, basic sequence manipulation (concatenation, transcription, translation, etc.), phylogenetic analysis (maximum parsimony, markov models, distance methods, likelihood computation and maximization), population genetics/genomics (diversity statistics, neutrality tests, various multi-locus analyses) and various algorithms for numerical calculus.</p></sec><sec><title>Conclusion</title><p>Implementation of methods aims at being both efficient and user-friendly. A special concern was given to the library design to enable easy extension and new methods development. We defined a general hierarchy of classes that allow the developer to implement its own algorithms while remaining compatible with the rest of the libraries. Bio++ source code is distributed free of charge under the CeCILL general public licence from its website <ext-link ext-link-type="uri" xlink:href="http://kimura.univ-montp2.fr/BioPP"/>.</p></sec> | <contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Dutheil</surname><given-names>Julien</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>Julien.Dutheil@univ-montp2.fr</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Gaillard</surname><given-names>Sylvain</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>syl.gaillard@gmail.com</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Bazin</surname><given-names>Eric</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>bazin@univ-montp2.fr</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Glémin</surname><given-names>Sylvain</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>glemin@univ-montp2.fr</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Ranwez</surname><given-names>Vincent</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>ranwez@isem.univ-montp2.fr</email></contrib><contrib id="A6" contrib-type="author"><name><surname>Galtier</surname><given-names>Nicolas</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>galtier@univ-montp2.fr</email></contrib><contrib id="A7" contrib-type="author"><name><surname>Belkhir</surname><given-names>Khalid</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>belkhir@univ-montp2.fr</email></contrib> | BMC Bioinformatics | <sec><title>Background</title><p>The design of re-usable software components into libraries has proved to be useful to the rapid development of bioinformatics applications, as witnesses the success of the ever growing open source libraries known as the Bio{*} projects. These libraries cover many fields of bioinformatics, but rarely offer tools for intensive calculus, as required for phylogenetic analyses for instance. Such computer-expensive applications are usually written in C, a compiled low-level language. Programming time-consuming tasks often requires the implementation of specific algorithms optimized for each particular problem. However, despite these specifities, there are some common points (<italic>e.g. </italic>likelihood calculation, minimization), and developers will appreciate to use modules pre-programmed in a flexible manner.</p><p>The use of the C++ language combines both the computer efficiency of the C language and the convenience of a Object-Oriented (OO) approach. We present Bio++, a set of C++ libraries dedicated to sequence analysis, molecular evolution and population genetics. Other libraries taking these advantages of C++ have been developed, including the libsequence library [<xref ref-type="bibr" rid="B1">1</xref>] for population genetics analysis (single nucleotides polymorphism and coalescent), PDBlib for manipulating tridimensional structures [<xref ref-type="bibr" rid="B2">2</xref>] or libcov [<xref ref-type="bibr" rid="B3">3</xref>] for protein phylogenetics and molecular evolution. The Bioinformatics Template Library [<xref ref-type="bibr" rid="B4">4</xref>] also provides algorithms for manipulating and analysing three dimensional structures, using a generic programming approach.</p></sec><sec><title>Implementation</title><p>We did a collaborative effort to design a hierarchy of useful classes for the rapid development of efficient applications in the fields of sequence analysis, phylogenetics, molecular evolution and population genetics. These libraries combine tools for data handling and numerical calculations, providing an easy-to-use, powerful and general development environment. The three main objectives of Bio++ are (i) to speed up the implementation of a new method in order to test it, (ii) to make easier the development of robust applications for distribution and (iii) to provide a high level programming language for biologists.</p><sec><title>Library design</title><p>We used object-oriented programming paradigms to develop Bio++. We defined for each class an abstract basal class that contains a number of purely virtual functions (dummy functions with an empty boddy). Such a basal class is called an interface and can be seen has a kind of contract that a class must follow in order to be used by other methods. Such a contract only specifies what an object is at least able to do, and not how it does it. Most classes in Bio++ are derived from interfaces, and some common instanciations are proposed for each interface. The user may, however, write its own implementation of a given method, and still remain compatible with the rest of the library. For instance, one may easily design a new alphabet to handle a new kind of sequences, or write a new substitution model and perform likelihood calculation without writing any likelihood computation function. Conversely, one may implement a new likelihood computation algorithm without re-developping substitution model classes. Several interfaces and abstract classes (partial implementations) linked by inheritance specify different levels of specialization. Depending on their objectives and programming level, users might simply use fully-specified objects, re-implement specific methods, or even design new classes.</p><p>The sequence container hierarchy is a representative example of the class hierarchies defined in Bio++ (see figure <xref ref-type="fig" rid="F1">1</xref>). The SequenceContainer basal interface only specifies that sequences can be accessed by their name. The OrderedSequenceContainer has the additional requirement that sequences can be accessed by index, which is generally the case, even if ordering has no biological meaning. The SiteContainer interface requires that sequences have the same length (<italic>i.e. </italic>are aligned) and hence may also be accessed by site (= column in the alignment). Several implementations are proposed: in the VectorSequenceContainer class, data are stored as a vector of sequences. Each sequence can hence be accessed in <italic>O</italic>(1). If there are <italic>N </italic>sequences, they are hence accessed in <italic>O</italic>(<italic>N</italic>). The AlignedSequenceContainer is derived from the VectorSequenceContainer class, and adds site access. Since data are stored as sequences, the access time for a site is in <italic>O</italic>(<italic>N</italic>) and the complete set of <italic>L </italic>sites is acessed in <italic>O</italic>(<italic>N </italic>× <italic>L</italic>). The VectorSiteContainer proposes an alternative implementation by storing sequences as sites instead of sequence objects. The sequence access is hence achieved in <italic>O</italic>(<italic>L</italic>), and the site access in <italic>O</italic>(1). All methods working on containers only deal with the SequenceContainer or the OrderedSequenceContainer interface, whith no assumption about the implementation. The execution time, however, may vary depending on the implementation used.</p><p>Development is facilitated by the use of the code documentation generated by the doxygen program [<xref ref-type="bibr" rid="B5">5</xref>]. Full class documentation can be consulted online or downloaded in navigable HTML format. A short tutorial is also available.</p></sec><sec><title>Library content</title><p>We split Bio++ classes into five libraries: three main biological libraries (SeqLib, PhylLib and PopGen-Lib) and two utility libraries (Utils and NumCalc), see table <xref ref-type="table" rid="T1">1</xref> for there content and dependencies:</p><p>• the Utils library contains core classes and utilitary functions;</p><p>• the NumCalc library contains classes for numerical calculus, including several optimization tools, random number generators, probability distributions and statistical functions;</p><p>• the SeqLib library is dedicated to sequence analysis: sequence storage and manipulation, input/output toward several file formats, alphabet translation, etc.</p><p>• the PhylLib library deals with phylogenetic trees: Markov models, likelihood calculation, etc. The library is more focused on molecular evolution (model fitting, ancestral state reconstruction, across-site rate variation) than pure phylogenetic reconstruction, although several topology reconstruction methods are implemented (distance methods, nearest-neighbor interchanges (NNI) search for maximum likelihood (ML) and parsimony scores);</p><p>• the PopGenLib library is dedicated to population genetics, with particular sequence and codominant markers storage. Methods are provided for polymorphism analysis, population divergence estimation and neutrality tests.</p></sec></sec><sec><title>Results</title><p>Figure <xref ref-type="fig" rid="F2">2</xref> shows a full program example. This application builds a Neighbor-Joining tree from a sequence file in Phylip format, and re-estimates parameters (branch lengths and shape of the rate across-site distribution) using maximum likelihood. The final tree is then written to a file in Newick format. Additional output files are also created, providing detailed information about the estimation procedure. The program begins with the creation of a ProteicAlphabet object (line 3):</p><p>const ProteicAlphabet * alphabet = new ProteicAlphabet();</p><p>This object specifies the type of sequence the program will use. The Alphabet object family contains classes for nucleic alphabets (DNA and RNA), proteins or codons (nuclear and mitochondrial, for vertebrates, echinoderms/nematodes and other invertebrates). A dataset is then readed from a Phylip sequence file. A sequence reader is created using</p><p>Phylip * seqReader = new Phylip(false, false);</p><p>Other file formats for reading and writing are supported, including the commonly used Fasta and Clustal. Sequences are then read and stored in a container:</p><p>SiteContainer * sites = seqReader->read("Myoglobin.phy", alphabet);</p><p>Since sequences are aligned, they are actually stored in a SiteContainer object (see figure <xref ref-type="fig" rid="F1">1</xref>). Line 7 to 10 display on screen some properties of the container, namely the number of sequences and sites. In this example we need to restrict our analysis to sites without any gap, so at line 11 a new container called "completeSites" is created from the former:</p><p>SiteContainer * completeSites = SiteContainerTools: :getSitesWithoutGaps(* sites);</p><p>Several {*}Tools classes providing utilitary functions are available. For instance, the SiteContainerTools class contains static functions that work on SiteContainer objects.</p><p>A JTT proteic SubstitutionModel object is created [<xref ref-type="bibr" rid="B6">6</xref>,<xref ref-type="bibr" rid="B7">7</xref>]:</p><p>SubstitutionModel * model = new JTT92(alphabet);</p><p>Here again, several commonly used models are available. We also use a discrete gamma rate across sites distribution with 4 rate classes and a shape parameter of 0.5 [<xref ref-type="bibr" rid="B8">8</xref>]:</p><p>DiscreteDistribution * rateDist = new GammaDiscreteDistribution(4, 0.5);</p><p>This model is used first to estimate the distance matrix from the data. A DistanceEstimation object is then created, by giving the model and data as parameters:</p><p>DistanceEstimation distEstimation(model, rateDist, completeSites);</p><p>DistanceMatrix * matrix = distEstimation.getMatrix();</p><p>Distances will be estimated using a maximum likelihood (ML) procedure. The NumCalc library provides an object-oriented implementation of several general optimization procedures (Optimizer objects). For phylogenetic optimization, specific optimizers are also available. The program then computes a neighbor-joining tree [<xref ref-type="bibr" rid="B9">9</xref>]. A Neighbor Joining object is instantiated and the resulting tree stored in a variable:</p><p>NeighborJoining nj(* matrix, false);</p><p>Tree<Node> * tree = nj.getTree();</p><p>This variable is passed to a TreeLikelihood object, together with the data set, model and rate distribution. An additional parameter tells the object to display a few informations. The DRHomogeneousTreeLikelihood object is for homogeneous models, with rate across sites variations.</p><p>DiscreteRatesAcrossSitesTreeLikelihood * likFunction = new DRHomogeneousTreeLikelihood(*tree, *completeSites, model, rateDist, true);</p><p>Several TreeLikelihood implementations are available. The DRHomogeneousTreeLikelihood object uses one 'view' by neighbor node for each node, <italic>i.e. </italic>three views by node in case of bifurcating trees [<xref ref-type="bibr" rid="B10">10</xref>]. This implementation saves computational time during NNI-mediated topology search, and is also more convenient to compute branch-lengths derivatives. An alternative implementation is proposed: HomogeneousTreeLikelihood, and others could be easily added if required. For parameter optimization, we used a utilitary function from the OptimizationTools class. This function automatically instanciates the appropriate optimizer object. It receives as argument a pointer toward the likelihood function, a tolerance number and a maximum number of function evaluations. Two optional log-files are provided to monitor the optimization process ("profiler" and "messenger", which contain several detailed informations, like for instance all parameters and function values at each step of the process), and the last parameter specifies the verbose level:</p><p>OptimizationTools:: optimizeNumericalParameters(likFunction, 0.000001, 1000000, messenger, profiler, 3);</p><p>Finally, the optimized tree is writen to a file named output. dnd in Newick file format:</p><p>Newick newick;</p><p>newick.write(*tree, "output.dnd");</p><p>This program creates three files: two for the ML tree estimation and one for the final tree.</p></sec><sec><title>Conclusion</title><p>Bio++ is a mature project which has been used in previous works like molecular coevolution analysis [<xref ref-type="bibr" rid="B11">11</xref>] or codon analysis [<xref ref-type="bibr" rid="B12">12</xref>]. However it is an active project still receiving new methods and improvements.</p><p>Development snapshots may be accessed by anonymous CVS on the library website. The website also provides on-line documentation of classes, tutorial and several code examples. Several programs have been developped with the Bio++ libraries, including bppML for ML likelihood tree estimation and bppSG for sequence generation by simulation under different kinds of models. Any contribution will be welcome, as specific functions or as additional libraries compatible with the present ones.</p></sec><sec><title>Availability and requirements</title><p>C++ libraries are not organized in a tree-like hierarchy as java packages or perl modules. They are bundled in a non-nested way, and may be compiled in two flavours, either static or shared (= dynamic). Dynamic libraries are loaded during the program execution and hence can be shared by several applications, while static libraries are hard-coded into the executable, which no longer requires the library to be installed.</p><p>Each Bio++ library is complient with the GNU standards and uses the autotools suite for compilation and installation. Bio++ can hence be built with these GNU tools on any unix-like system. Alternatively, sources can be imported and compiled in any C++ development environment (IDE). It has been successfully installed on Linux, MacOS X and Windows using the Cygwin port and MinGW. Bio++ is distributed freely under the CeCILL public license (the French free software license, compatible with the GNU General Public License). It is available from its website at <ext-link ext-link-type="uri" xlink:href="http://kimura.univ-montp2.fr/BioPP"/>.</p></sec><sec><title>Authors' contributions</title><p>JD developped the Utils, NumCalc, SeqLib and PhylLib libraries and drafted the manuscript. SGa developped the PopGenLib library, provided algorithms for the NumCalc library and helped with the GNU configuration tools. EB developped tools for the PopGenLib library. SG1 provided tools for the PopGenLib and SeqLib libraries. VR helped in the development of the PhylLib library. NG participated in the design of the library. KB suppervised the whole project. SG1, VR, NG and KB helped to draft the manuscript. All authors read and approved the final 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</title><p>Sources and data files of the example as a zipped archive.</p></caption><media xlink:href="1471-2105-7-188-S1.zip" mimetype="application" mime-subtype="x-zip-compressed"><caption><p>Click here for file</p></caption></media></supplementary-material></sec> |
Mining gene expression data by interpreting principal components | <sec><title>Background</title><p>There are many methods for analyzing microarray data that group together genes having similar patterns of expression over all conditions tested. However, in many instances the biologically important goal is to identify relatively small sets of genes that share coherent expression across only some conditions, rather than all or most conditions as required in traditional clustering; e.g. genes that are highly up-regulated and/or down-regulated similarly across only a subset of conditions. Equally important is the need to learn which conditions are the decisive ones in forming such gene sets of interest, and how they relate to diverse conditional covariates, such as disease diagnosis or prognosis.</p></sec><sec><title>Results</title><p>We present a method for automatically identifying such candidate sets of biologically relevant genes using a combination of principal components analysis and information theoretic metrics. To enable easy use of our methods, we have developed a data analysis package that facilitates visualization and subsequent data mining of the independent sources of significant variation present in gene microarray expression datasets (or in any other similarly structured high-dimensional dataset). We applied these tools to two public datasets, and highlight sets of genes most affected by specific subsets of conditions (e.g. tissues, treatments, samples, etc.). Statistically significant associations for highlighted gene sets were shown via global analysis for Gene Ontology term enrichment. Together with covariate associations, the tool provides a basis for building testable hypotheses about the biological or experimental causes of observed variation.</p></sec><sec><title>Conclusion</title><p>We provide an unsupervised data mining technique for diverse microarray expression datasets that is distinct from major methods now in routine use. In test uses, this method, based on publicly available gene annotations, appears to identify numerous sets of biologically relevant genes. It has proven especially valuable in instances where there are many diverse conditions (10's to hundreds of different tissues or cell types), a situation in which many clustering and ordering algorithms become problematic. This approach also shows promise in other topic domains such as multi-spectral imaging datasets.</p></sec> | <contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Roden</surname><given-names>Joseph C</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>joe.roden@jpl.nasa.gov</email></contrib><contrib id="A2" contrib-type="author"><name><surname>King</surname><given-names>Brandon W</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>kingb@caltech.edu</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Trout</surname><given-names>Diane</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>diane@caltech.edu</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Mortazavi</surname><given-names>Ali</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>alim@caltech.edu</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Wold</surname><given-names>Barbara J</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>woldb@caltech.edu</email></contrib><contrib id="A6" contrib-type="author"><name><surname>Hart</surname><given-names>Christopher E</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>hart@caltech.edu</email></contrib> | BMC Bioinformatics | <sec><title>Background</title><p>Bioinformatics has placed much emphasis on using various unsupervised clustering techniques as a means to understand the information present in gene microarray expression datasets. Clustering techniques produce a rich taxonomy of results by defining groups of genes that act more or less similarly across a number of experimental conditions. The diverse approaches to clustering genes by expression levels include k-means [<xref ref-type="bibr" rid="B1">1</xref>], self-organizing maps [<xref ref-type="bibr" rid="B2">2</xref>], hierarchical algorithms [<xref ref-type="bibr" rid="B3">3</xref>,<xref ref-type="bibr" rid="B4">4</xref>] and probabilistic models [<xref ref-type="bibr" rid="B5">5</xref>]. Some approaches permit clustering of the conditions as well [<xref ref-type="bibr" rid="B6">6</xref>-<xref ref-type="bibr" rid="B8">8</xref>]. Based on co-expression, genes that comprise individual expression clusters are often postulated to be co-regulated, and to the extent that this hypothesis is correct in any specific biological situation, the gene cluster definitions can offer key insights into gene regulatory network (GRN) structure and function.</p><p>Another common data mining task is to try to identify small sets of genes that can serve as effective predictors of disease diagnosis or prognosis. While clustering at its best is good at finding sets of genes that are similarly expressed across all conditions within a dataset, many issues (e.g. selection of K, stochastic effects, and "noise" from large numbers of genes that change little over most of the conditions) can prevent clustering from successfully highlighting small groups of interestingly co-expressed genes [<xref ref-type="bibr" rid="B9">9</xref>,<xref ref-type="bibr" rid="B10">10</xref>]. This often encountered problem is addressed in part by hierarchical phylogenetic ordering algorithms such as average linkage in Mike Eisen's cluster program [<xref ref-type="bibr" rid="B3">3</xref>], but the information biologists seek regarding shared sub-patterns of co-expression can be obscured by both algorithmic and visualization constraints. The algorithmic limitations in hierarchical clustering confound and "cover" the presence and organization of smaller and more specific gene groups that are similar across only a subset of conditions within the larger dataset. In any case, biologists generally subjectively define a cluster of genes from such phylogenetic trees based largely on human pattern recognition. Finally, nothing inherent in the clustering approach helps to direct a biologist to which cluster is interesting or relevant. Instead, biologists generally take the path of focussing on a group of genes exhibiting a pattern of expression that supports a specific hypothesis, or search for a known gene or genes of interest within a cluster to form an explanation for others in the cluster.</p><p>Support vector machines have been shown to be useful for identifying small sets of related and predictive genes [<xref ref-type="bibr" rid="B11">11</xref>-<xref ref-type="bibr" rid="B14">14</xref>], but represent a supervised learning approach which requires one to first define a set of known positive examples, a set of known negative examples, and a specific covariate to predict. We wanted an unsupervised algorithm that would help us to find relationships and structure in the data that is more specific than what clustering algorithms usually deliver, yet is hypothesis independent. We found it efficient and useful to use as an independent starting tool a very direct approach based on principal components analysis (PCA, see Methods section). A virtue is that this approach is computationally efficient for very large datasets, especially compared with most clustering algorithms, but is also applicable to much smaller ones. It allows one to directly explore each of the independent and diverse sources of variation present within a gene expression dataset and to subsequently identify the specific genes that vary the most, together with the conditions in which they vary.</p><p>Unlike conventional clustering and ordering algorithms, this PCA based approach permits a gene to be highlighted and grouped as influential in multiple condition sets, whereas in cluster membership a gene is typically assigned to one unique cluster. The "single cluster assignment" quality of traditional clustering and ordering algorithms is problematical because it tends to hide commonality of expression that is restricted to a small, interesting, and often entirely unpredicted subgroups of tissues, cell types, treatments or other condition types. This situation, perhaps because of inherent properties of gene network structure, will arise increasingly as the number and diversity of conditions represented in expression datasets increases.</p><p>The clustering method of Barkai et al. [<xref ref-type="bibr" rid="B15">15</xref>,<xref ref-type="bibr" rid="B16">16</xref>] addresses this issue of multiple membership in a different way, by using randomly-selected gene sets to iteratively search for and refine self-consistent groups. Their approach, which is related to PCA through singular value decomposition (SVD), also permits genes to be assigned to multiple "expression modules." In contrast to the method presented here, there is no provision for correlating modules with covariate data.</p><p>The use of principal components analysis presented here differs from other recent applications in gene expression analysis. PCA is most commonly used in as a means of dimensionality reduction prior to clustering [<xref ref-type="bibr" rid="B7">7</xref>,<xref ref-type="bibr" rid="B17">17</xref>] or prior to classification [<xref ref-type="bibr" rid="B18">18</xref>,<xref ref-type="bibr" rid="B19">19</xref>]. It is also used to visualize or confirm clustering results [<xref ref-type="bibr" rid="B19">19</xref>-<xref ref-type="bibr" rid="B21">21</xref>]. In contrast, our use of PCA aims to find, then examine, and where possible, generate hypotheses to explain individual principal components. In this manner, we build on the observations by Hilsenbeck [<xref ref-type="bibr" rid="B22">22</xref>] Raychaudhuri et. al. [<xref ref-type="bibr" rid="B23">23</xref>] who used PCA to gain insight into the underlying factors present in the Chu et. al. yeast sporulation experiments [<xref ref-type="bibr" rid="B24">24</xref>]. Wall et. al. [<xref ref-type="bibr" rid="B25">25</xref>] introduced a novel use of singular value decomposition (SVD) for gene expression analysis that identifies non-exclusive gene groups, and Selaru et. al. [<xref ref-type="bibr" rid="B26">26</xref>] illustrated the potential of PCA to detect molecular phenotypic bases that correspond to relevant clinical or biological features of human tumors. Their approach identifies a subset of principal components that correlate well with known covariates. Here we introduce methods that extend beyond producing gene groups and observing a few principal components. Our methods provide a path for systematically analyzing each principal component by identifying the genes most influential in defining a particular principal component and the conditions in which those influential genes vary significantly. Finally, we describe methods which aim to explain each principal component's observed variance in terms of the condition variables deemed most likely to be driving the variance. We also introduce a software package that implements these methods. These methods and our software package provide automated and objective way of doing what a biologist naturally tries to do through inspection and pattern recognition.</p></sec><sec><title>Implementation</title><p>We have developed a Python package to implement the PCA interpretation capability described in detail in the Methodology section. This PCA analysis package has been added to CompClust developed previously [<xref ref-type="bibr" rid="B9">9</xref>,<xref ref-type="bibr" rid="B27">27</xref>]. The combined packages allow one to cluster, classify and visualize numeric datasets that have discrete or numeric annotations (referred to as labelings, or labelled datasets), and to compare labelings with confusion matrices and metrics such as normalized mutual information (NMI) [<xref ref-type="bibr" rid="B28">28</xref>]. This PCA analysis tool (including the complete results for the dataset analysis described in the Results section) has also been made accessible through the CompClustWeb web-based interface [<xref ref-type="bibr" rid="B29">29</xref>]. Our software makes use of data manipulation and graphical plotting using the matplotlib package [<xref ref-type="bibr" rid="B30">30</xref>], and the statistics are generated using the rPy package [<xref ref-type="bibr" rid="B31">31</xref>] and Gary Strangman's Python stats package [<xref ref-type="bibr" rid="B32">32</xref>].</p><p>The web-based front-end permits users to get a complete report on the interpretation of each principal component, including interactive PCA projection plots with the principal component's extreme genes (PCEGs) highlighted; ranked lists of the PCEGs with detailed annotations; interactive significance-ordered gene set trajectory plots that permit users to drill down to the individual gene level; similarly ordered condition reports ordered by expression difference and grouped by significance, including covariate info (with significantly correlated covariates highlighted); and finally a report of any suggestive covariates that are well correlated with the significant column grouping, including the confusion matrices and/or plots of statistics scores to back up the conclusions. All principal component analysis and results generation is implemented in a Python package so that analyses of large datasets can be executed in a batch mode rather than through the graphical interface. Further, the software that implements the CompClustWeb interface is provided within the CompClust package, so a software developer can create his or her own CompClustWeb server to review results of their PCA interpretation.</p></sec><sec><title>Results</title><sec><title>Application to microarray expression data, Case 1: GNF human data</title><p>We obtained gene expression data from the Genomics Institute of the Novartis Research Foundation ("GNF") Gene Expression Database via their SymAtlas web site [<xref ref-type="bibr" rid="B33">33</xref>,<xref ref-type="bibr" rid="B34">34</xref>]. The dataset is a challenge for most clustering algorithms because it contains 158 tissue samples hybridized to two Affymetrix microarray chips: U133A and GNF1H. The dataset combines the measurements of these chips to provide a total of 33,689 unique probe identities across the 158 tissue samples. Expression data are signal intensities estimated by Affymetrix Microarray Suite v5. For our analysis we used the log base 2 of the expression signal, and included data for all tissues and probes (noting that absent and present calls were not provided with the signal intensities). We applied our principal components analysis tool to generate interpretations for each of this dataset's 158 principal components.</p><p>As detailed in the Methodology section, for each principal component we identified a set of gene probes occupying the high and low extremes of that principal component's axis (we refer to these as principal component extreme genes, or PCEGs). One can adjust parameters to recover smaller and larger numbers of PCEGs per component by specifying either a likelihood threshold or an explicit number of PCEGs. The PCEGs are those probes having the most highly weighted values for that principal component, selected because they stand out from the others, they influence the principal component's direction, and thus they warrant further investigation. We selected probes with likelihoods less than <italic>extremeThresh </italic>= 0.00001, which yielded on average 20 low and 20 high extreme genes per principal component, though the sets sizes do vary considerably (μ = 18.9, σ = 17.2). Next we identified the tissues in which the high PCEGs showed significantly different expression than the low PCEGs. Visualizations produced include scatter plots of the extreme genes in PCA sub-spaces (PC <italic>N</italic>-1vs. PC <italic>N</italic>), and extreme gene trajectories in original tissue order as well as with tissues ordered by decreasing difference of mean of high PCEGs and mean of low PCEGs. The latter trajectory plot emphasizes how the extreme genes for a principal component show a pattern of expression that imposes a partitioning of tissues. It is left to human interpretation to examine the extreme genes and the tissue partitioning exposed by each principal component, and thus to build hypotheses that attach meaning to the sources of variation. The percentage variance explained by the top 50 principal components is provided in Table <xref ref-type="table" rid="T1">1</xref>. Example results for two illustrative principal components, PC7 and PC21, are shown in Figures <xref ref-type="fig" rid="F1">1</xref> through <xref ref-type="fig" rid="F4">4</xref> and in Tables <xref ref-type="table" rid="T2">2</xref> through <xref ref-type="table" rid="T5">5</xref>, and are discussed below. The complete analysis results generated for PC7 are provided as an example in the supplemental files: [see <xref ref-type="supplementary-material" rid="S1">Additional file 1</xref>], [see <xref ref-type="supplementary-material" rid="S2">Additional file 2</xref>], [see <xref ref-type="supplementary-material" rid="S3">Additional file 3</xref>], [see <xref ref-type="supplementary-material" rid="S4">Additional file 4</xref>], [see <xref ref-type="supplementary-material" rid="S5">Additional file 5</xref>], [see <xref ref-type="supplementary-material" rid="S6">Additional file 6</xref>]. Our supplemental materials web site [<xref ref-type="bibr" rid="B35">35</xref>] contains the complete collection of PCA interpretation results generated for these dataset for all principal components, as well as results of a comparable analysis done at <italic>extremeThresh </italic>= 0.001 which yielded larger PCEG sets.</p><p>We addressed the question of biological and statistical significance of PCs and the sets of extreme genes identified. Each set of high and low extreme genes from each principal component was tested for Gene Ontology (GO) statistical enrichment when compared to the human GO annotations from NCBI's loc2go dataset using the hypergeometric to calculate the p-value of each GO term. Terms that were enriched for a particular PC at 1% significance threshold and that were still significant following a Bonferroni correction for multiple hypothesis testing as described in [<xref ref-type="bibr" rid="B36">36</xref>] are reported as enriched (see Table <xref ref-type="table" rid="T6">6</xref>). 26 of the top 42 PCEG lists, derived using a stringent cut-off of 0.00001, produced significant GO enrichments; no PCEG sets beyond PC42 showed significant enrichment. As discussed below, many of the significant results showed obvious biological coherence and relationships to the specific samples associated with the PC of origin. This argues that PCs containing less than 1% of the total variation in this dataset are still relevant and point to coherent and important gene sets and their related samples.</p><p>Relationships between extreme gene sets and the corresponding sets of driving samples could be discerned for many PCs. The Methodology section presents a way to additionally correlate each principal component's sample partitioning with any available sample covariates. Although some human sample covariate information is provided in our test case, the GNF human expression dataset is not amenable to this additional layer of analysis because multiple subject's RNA samples were pooled prior to amplification and array hybridization. However, a second publicly available dataset with rich covariate information is presented below.</p></sec><sec><title>Application to microarray expression data, case 2: human diabetes</title><p>We acquired a Human diabetes expression dataset [<xref ref-type="bibr" rid="B37">37</xref>] from the Broad Institute Cancer Program dataset repository [<xref ref-type="bibr" rid="B38">38</xref>] along with the corresponding phenotype covariate data, and applied the filtering step as they described to produce a set containing 10,983 probes across 43 samples. Tissue samples were skeletal muscle biopsies from 3 diagnosis groups: normal glucose tolerance (NGT, <italic>n </italic>= 17); impaired glucose tolerance (IGT, <italic>n </italic>= 8); and Type 2 diabetes mellitus (DM2, <italic>n </italic>= 18). We used our PCA interpretation software to perform an unsupervised analysis of the DM2 vs. NGT subset (as that subset is comparable to the previous published result). As described in the Methodology section, PCEG sets were determined using an <italic>extremeThresh </italic>likelihood threshold of 0.001, which yielded about 50 high and 50 low extreme genes per principal component. For each principal component <italic>N</italic>, samples were partitioned in to <italic>UP</italic><sub><italic>N</italic></sub>, <italic>FLAT</italic><sub><italic>N </italic></sub>and <italic>DOWN</italic><sub><italic>N </italic></sub>sample sets on the basis of PC-<italic>N </italic>extreme high and extreme low expression differences. The supplemental materials contain PCA interpretation results for all 35 principal components, as well as results of a comparable analysis done at <italic>extremeThresh </italic>= 0.0001 which yielded smaller PCEG sets (see [<xref ref-type="bibr" rid="B35">35</xref>]).</p><p>This dataset contains more than 50 covariates, which provides the opportunity to interpret each principal component by searching for covariates that correlate well with expression patterns in the PCEG sets. As described in the Methodology section, we asked if any of the covariate annotations are well correlated with the partitioning of samples into <italic>UP</italic><sub><italic>N</italic></sub>, <italic>FLAT</italic><sub><italic>N </italic></sub>and <italic>DOWN</italic><sub><italic>N </italic></sub>sets. Covariate distributions were compared across different partitions (when sufficient data was available) and any significant trends identified were recorded (see Table <xref ref-type="table" rid="T7">7</xref>). For covariates identified as significantly correlated with a principal component's sample partitioning, covariate distribution plots were generated to further investigate and evaluate the apparent relationship. For example, Figure <xref ref-type="fig" rid="F5">5</xref> illustrates that PC14's <italic>UP<sub>14</sub></italic>, <italic>FLAT<sub>14 </sub></italic>and <italic>DOWN<sub>14 </sub></italic>sample partitions appear to be significantly related to two covariate measurements: Insulin_0 (<italic>sig </italic>= 0.0010); and Type2b_(%) (<italic>sig</italic>= 0.0077). The Pearson's correlation between the mean expression for the PC14EG-high set and Insulin_0 and Type2b_(%) covariates are <italic>r </italic>= 0.411 and <italic>r </italic>= 0.467 respectively.</p></sec></sec><sec><title>Discussion</title><p>This PCA-based data-mining tool highlights specific patterns of expression and associates them in a convenient way with the genes and samples responsible for those patterns. Some associations in the first few principal components (PCs) of the GNF set reflect major features in the data that are expected. This includes the global high and low constitutive expression profiles of PC1 (67% of variance in the GNF dataset). A component similar to this is often the first or second PC in Affymetrix array datasets. GNF PC3, in contrast, highlighted brain/neuronal tissues, which we expected in this dataset because there are many more samples from brain regions than from any other tissue, and there are thousands of genes that are expressed in a general brain pattern. The GO enrichment analysis associated with PC3lowEGs confirmed this impression by identifying neurogenesis, central nervous system, and synaptic terms as significantly enriched for PC3EG (Table <xref ref-type="table" rid="T6">6</xref>).</p><p>We asked if principal components that account individually for small fractions of variation in the data are likely to be significant. Conventional practice generally ignores principal components accounting for a few percent, or less, of total variation, on the assumption that such minor components are most likely dominated by noise. We believe this assumption, that all of the PC's accounting for small fractions of data should be ignored because they are artefact, to be wrong in the context of our analysis. We believe this, in part, because in our analysis we find many PCEGs for minor components are statistically enriched (Table <xref ref-type="table" rid="T6">6</xref>). Further, computational experiments using randomized data fail to produce any significant enrichments. This GO enrichment analysis, it should be noted, tends to underestimate the fraction of gene sets that are significant. This is a known artefact in instances where the input gene number (here those passing the selected rank sum threshold) is small, and its effects are exacerbated by the fact that GO annotations for human are still very much in a building phase. Many genes that will eventually be associated with GO terms are not yet entered. This means that reducing the threshold modestly, and therefore increasing the gene number, can uncover additional significant GO term enrichment in some of these PCs. For complete results of GNF analysis at <italic>p </italic>< = 0.001 see our supplemental materials web site [<xref ref-type="bibr" rid="B35">35</xref>].</p><p>Viewed from a biological perspective, this PCA mining revealed several different classes of relationships. GNF PC21 is a good example of a component that highlights a coherent gene set and its corresponding tissues, many of which would also be grouped together by conventional clustering algorithms. This is true even though PC21 accounts for just 0.25% of total variation. The PC21EG-low set (defined at <italic>p </italic>< = 0.00001) was enriched in a statistically significant way for five different GO terms (Table <xref ref-type="table" rid="T6">6</xref>), and these terms (myogenesis, muscle contraction, etc.) tell a simple and internally consistent story about muscle development and function. The top tissue samples driving PC21-low are skeletal muscle, tongue (also composed largely of striated muscle), heart and thyroid (which includes a population of myoid cells). Most top-ranked genes in this PCEG set are so specific for striated muscle that they would also appear together in conventional clusterings, although many clustering approaches become technically problematic with datasets of this size. However the PC21EG-low list differs from a conventional muscle cluster because it also includes some genes that are partly associated with muscle and partly associated with other tissues. PP1R1A, a regulatory subunit of protein phosphatase1, is such a gene. A role for it in striated muscle is suggested based on its coherent presence in tongue, skeletal and cardiac samples, even though it might well not have been seen in this light by standard clustering.</p><p>The second example is GNF PC7, which accounts for 0.81% of variation. It illustrates a different kind of biological relationship that more strongly distinguishes results of PCA mining from classical clustering. Top extreme genes associated with PC7 by inspection turn out to be a "who's who" of extracellular matrix components (a specific subset of fibronectins, collagens, laminins plus matrix associated proteins like <italic>MFAP5</italic>, <italic>MGP</italic>, <italic>LUM</italic>; regulatory molecules that mediate stability and function of those matrix components (thrombospondin, <italic>SPARC</italic>, <italic>ADAMTS1</italic>, <italic>Plod2</italic>); and matrix associated signalling and matrix associated signal modulators (insulin like growth factor binding proteins 7, 8 and 10; <italic>Sema3c</italic>). GO analysis confirms what inspection of the top PC7EGs suggested: namely that a set of extracellular matrix components are expressed in these driving tissues. It is instructive to look at the individual expression profiles for these genes directly at the GNF website and also in aggregate, as represented in the tissue (conditions) list in Table <xref ref-type="table" rid="T2">2</xref>. The most prominent contributing tissues associated with high expression of these genes are informative specifically because they <underline>do not</underline> constitute a group that would have been selected <italic>a priori </italic>as a coherent set based on known tissue function or shared developmental origin.</p><p>This is useful because a biologist interrogating the GNF database would not likely have constructed a query combining adipocytes, smooth muscle, bronchial epithelium, nor would one expect traditional clustering algorithms to place these genes so close to each other as to catalyze the same observation. Similarly, conventional ordering algorithms would not have placed them adjacent to each other because other parts of their expression profiles, containing different genes than those in PC7, would dominate their positions. And concerning genes, they have, in addition to commonalities of expression highlighted by PC7, differences from each other in additional diverse tissues not highlighted by PC7. The PCA grouping gives impetus and a necessary starting gene list to search for one or more factors or regulatory RNAs with a similar expression pattern, or to search for a shared and perhaps evolutionarily conserved cis-acting DNA sequence motifs. It is unlikely that these working hypotheses would have been arrived at easily by widely used methods of gene expression analysis.</p><p>The diabetes dataset offered us an opportunity to add more value to principal component interpretations by searching for covariates that appear correlated at some significance level. The relationships highlighted between covariates and principal components are suggestive, but not conclusive by themselves. Rather, they provide hypotheses that a researcher may wish to further investigate. While we have not delved deeply into this dataset, we believe that a number of the principal components are highlighting a number of meaningful sources of variation present. It is not clear whether this set should exhibit the same proportion of meaningful principal components as the GNF dataset, as by design the NGT vs. GM2 dataset does not contain substantial diversity of samples. Likewise, the selection of covariates was focused on a narrow set of measurements selected to be indicative of diabetes status, and so many covariates are redundant. We anticipate this tool will be maximally useful in cases where datasets are rich in both sample complexity and diversity of covariates.</p></sec><sec><title>Conclusion</title><p>Results presented above show that this PCA-mining approach can guide a user to biologically significant observations that both complement and reinforce those from conventional clustering analysis. The software package and web interface make this style of microarray analysis straightforward and accessible.</p><p>We have applied this to four additional microarray datasets (as yet unpublished) and to one multi-spectral imaging dataset. In each case we found the interpretations that the tool presented to be useful. In general, it seems that the top few principal components identify very broad characteristics of the data. Digging to the deeper components that comprise smaller but more particular substructure in the data leads to more subtle but often meaningful observations, many being complementary to standard clustering.</p><p>With respect to the top few components, PC1 is usually the approximate diagonal through the sample/condition space, explaining the overall variation in absolute expression level. For some other datasets we have noticed that the top few PCs can also highlight effects of preprocessing normalization steps or global data quality issues. This means they do not necessarily expose the most important biological variation. Thus, in one microarray dataset not shown here, PC2 was found to be extremely well correlated with a measure of quality of samples, as reflected by the percent of Affymetrix probes called present. Given this evidence of data quality effects comprising a major source of variation over the entire dataset, one might be motivated to remove the major offending conditions, and then repeat the PCA interpretation on the remaining conditions (columns). The idea is that an independent source of variation might be obscured by more dominant signals or noise present in the data from the offending condition.</p><p>Our experience thus far leads us to think that this PCA interpretation method will contribute to microarray expression analysis, as one part of a panel of methods that are sensitive to different features in a dataset, such as sample number, gene number, and distribution of variation across the samples. The PCA method should be especially useful for large, complex datasets that offer rich variation among many samples. What is certain is that there are almost always multiple sources of variation in a dataset and that in any specific study their nature and relative strength is informative, whether the origin is an easily-understood biologic one, a technical one, or a poorly-understood but nonetheless biologically pertinent one.</p><p>We are continuing to explore ways to improve our methodology and software package. We anticipate further advances will come with software infrastructure improvements to permit covariate analyses of both column (sample) covariates and row (gene) covariates. The CompClust dataset labeling capability [<xref ref-type="bibr" rid="B27">27</xref>] allows a user to attach diverse and numerous labelings to rows or columns. For example we can pull in additional row (gene probe) annotations such as Gene Ontology (GO) functional groups. Beyond explicitly comparing the NMI significance of specific row partitionings for discrete covariates, we plan to add routines to CompClust to automatically indicate when a group of genes are found to be enriched in specific GO categories (as was done in our analyses above), and more generally to handle large, discrete, multi-valued distributions of values. Our use of NMI treats discrete covariates as discrete random variables that can have at most a single value per condition, and so does not optimally address issue of multi-valued discrete random variables (e.g. GNF data has covariate "concomitant medications" with values like "aspirin", "tylenol", and "aspirin & tylenol"). We are considering more elaborate extensions of mutual information or alternatives that might be able to take further advantage of such multi-valued entries.</p></sec><sec><title>Availability and requirements</title><p>The PCA interpretation software is implemented as one component of the CompClust Python package [<xref ref-type="bibr" rid="B9">9</xref>,<xref ref-type="bibr" rid="B27">27</xref>], which is freely available for non-commercial use. The software capability is also accessible through the CompClustWeb web-based interface [<xref ref-type="bibr" rid="B29">29</xref>]. The software that implements the web application is also included within the CompClust software distribution.</p><p>Project name: the PCA interpretation component of CompClust</p><p>Project home page: <ext-link ext-link-type="uri" xlink:href="http://woldlab.caltech.edu/compclust"/></p><p>Operating system(s): platform independent (Windows, Linux, Mac OS X)</p><p>Programming language: Python</p><p>Other requirements: Python 2.3 or higher (and some free Python packages)</p><p>License: MLX Public License 1.0 (non-commercial use allowed)</p><p>Use by non-academics: licence needed</p><p>We recommend that interested researchers use the web-based application, CompClustWeb, from any platform to review the PCA interpretation results for the GNF human gene expression and Broad Institute human diabetes expression data sets. We have written a CompClust PCA interpretation tutorial that demonstrates how to use CompClust's programming interface to generate PCA interpretations. Following the tutorial requires that CompClust and its Python prerequisites be installed. We created an easy-to-use CompClustShell installer for Windows that provides everything needed. For other operating systems (e.g. Linux & OS X) we recommend that a software developer or system administrator help with the CompClust source code installation. We are working to simplify installation and plan to provide user-friendly installers for other operating systems in the near future.</p></sec><sec sec-type="methods"><title>Methodology</title><p>We have developed the following algorithm for identifying and analyzing multiple independent sources of variance present within multi-dimensional sample datasets, in particular those that are produced by gene microarray expression experiments. The overall approach can be summarized as follows: 1) perform principal components analysis of the dataset; for each principal component: 2) identify the most extreme gene probes (those with the highest or lowest weighting) for that principal component; 3) identify and group any conditions in which those extreme probes vary significantly; 4) identify any condition covariates that correlate well with the condition grouping. By extending the interpretation of each principal component from extreme genes (rows) to ordered groups of significant conditions (columns) and further to identifying statistically significant correlations with column covariates, we attempt to make full use of the available data, in an objective and data-driven way, to analyze and provide meaningful interpretations of the diverse sources of variation present within the dataset.</p><sec><title>Determine the principal components of the dataset</title><p>Our dataset <italic>D </italic>consists of <italic>nc </italic>columns (e.g. tissue samples or conditions) and <italic>nr </italic>row vectors (e.g. gene probes), each row vector <italic>x<sub>i </sub></italic>∈ ℜ<sup><italic>nc </italic></sup>where <italic>i </italic>∈ [1, <italic>nr</italic>]. Such a dataset is usually represented as a two-dimensional <italic>nr </italic>× <italic>nc </italic>matrix (where <italic>nr </italic>> <italic>nc</italic>). The dataset may optionally have <italic>nk </italic>supplemental covariate annotations <italic>C </italic>associated with each row or column. Each annotation <italic>C</italic><sub><italic>k </italic></sub>where <italic>k </italic>∈ [1, <italic>nk</italic>] can be either discrete (e.g. sex) or continuous (e.g. age), to permit the association of one discrete value per column (e.g. values male or female), or one continuous value per column (e.g. values 12, 16, or 42).</p><p>Our procedure starts by employing principal components analysis (PCA) to sequentially identify a series of new basis vectors or axes <italic>PC</italic><sub>1</sub>, <italic>PC</italic><sub>2</sub>, ...<italic>PC</italic><sub><italic>nc </italic></sub>in the high-dimensional column space ℜ<sup><italic>n </italic></sup>that are each aligned sequentially to capture the most as-yet unexplained variance. This is accomplished by applying the numeric procedure singular value decomposition (SVD) to the covariance matrix of <italic>D</italic>, <italic>cov(D)</italic>, to produce the decomposition <italic>cov</italic>(<italic>D</italic>) = <italic>USV<sup>T </sup></italic>that contains the eigenvectors of <italic>cov</italic>(<italic>D</italic>) in the columns of <italic>U </italic>and eigenvalues in the diagonal of <italic>S </italic>such that the eigenvalues are sorted by descending size. Each covariance eigenvector, or principal component <italic>PC</italic><sub>1</sub>, <italic>PC</italic><sub>2</sub>, ...<italic>PC</italic><sub><italic>nc</italic></sub>, explains a fraction of the total variance contained in the dataset, and each principal component <italic>PC</italic><sub><italic>n+1 </italic></sub>is orthogonal to the previous principal component <italic>PC</italic><sub><italic>n</italic></sub>. such that they define the basis of a new vector space <italic>P</italic>. These results are made available to the users in the form of <italic>nc </italic>plots, one for each of the principal component vectors, as well as a plot of the singular values contained in the diagonal of <italic>S </italic>to indicate the relative amount of variance each component explains.</p></sec><sec><title>Identify extreme gene probes for each principal component</title><p>Next, we project each data point <italic>x<sub>i </sub></italic>(corresponding to a gene probe, or row vector) into the new coordinate system by <italic>P </italic>= <italic>DU</italic>, effectively rotating the entire data point set <italic>D </italic>into the new principal component axes space, producing the rotated data set <italic>P</italic>. Each data point <italic>p<sub>i </sub></italic>in the rows of <italic>P </italic>corresponding to <italic>x<sub>i </sub></italic>has a coordinate for each principal component axis that describes where the data point <italic>p<sub>i </sub></italic>lies when projected along each axis <italic>PC</italic><sub>1</sub>, <italic>PC</italic><sub>1</sub>, ...<italic>PC</italic><sub><italic>nc</italic></sub>. For each principal component <italic>PC</italic><sub><italic>n </italic></sub>(<italic>n </italic>∈ [1, <italic>nc</italic>]) we select a set of data points from each end of that principal component axis- these are the extreme points for <italic>PC</italic><sub><italic>n</italic></sub>, called the principal component extreme genes, or PCEGs for convenience. The PCEGs can be identified and ranked in one of two ways: by identifying points having a low probability (<italic>p </italic>< = <italic>extremeThresh</italic>) of belonging to a Gaussian fit to the distribution of points along the <italic>PC</italic><sub><italic>n </italic></sub>axis, or by taking a fixed number of <italic>nExtreme </italic>points at each tail of the distribution. <italic>H</italic><sub><italic>n </italic></sub>is the resulting set of data points having the highest coordinate values for <italic>PC</italic><sub><italic>n</italic></sub>, and <italic>L</italic><sub><italic>n </italic></sub>is the resulting set of data points having the lowest coordinate values for <italic>PC</italic><sub><italic>n</italic></sub>. These high and low extreme gene point sets are informative in and of themselves because they represent the most extreme of the data points along a principal axis of variation. As such, the high and low PCEG sets are some of the primary outputs generated by our procedure. We use the term "extreme" in a very general sense, in that the points stand out because they are far from the main distribution. We do not mean to imply that such points are either biologically relevant or nuisance data that should be removed; rather we are interested in these points in an unbiased way. By further analyzing their pattern of expression in the original axes we hope to gain a better understanding of their possible biologic significance.</p></sec><sec><title>Identify significant conditions for each principal component</title><p>The extreme gene points comprising <italic>H</italic><sub><italic>n </italic></sub>are located near one edge of the high-dimensional cloud of points, and points <italic>L</italic><sub><italic>n </italic></sub>are near the opposite edge. Thus, points <italic>H</italic><sub><italic>n </italic></sub>are likely to have coordinate values that are maximally different from points <italic>L</italic><sub><italic>n </italic></sub>in a subset of the original column space coordinate system. Our procedure next seeks to identify in which of the original columns (the original axes or dimensions) we find the greatest difference of values for points <italic>H</italic><sub><italic>n </italic></sub>versus points <italic>L</italic><sub><italic>n</italic></sub>. We do this by comparing distributions of values in <italic>H</italic><sub><italic>nj </italic></sub>versus <italic>L</italic><sub><italic>nj </italic></sub>for each of the original columns <italic>j </italic>(<italic>j </italic>∈ [1, <italic>nc</italic>]). A two-sided Wilcoxon rank sum test is used to estimate the likelihood that these two sets of values are drawn from the same distribution [<xref ref-type="bibr" rid="B39">39</xref>]; the resulting p-values for each column are used to rank order and group the columns, rather than as actual probabilities. Columns having a likelihood less than a user-defined significance level <italic>test1Thresh </italic>are identified and placed into one of two column sets: <italic>UP</italic><sub><italic>n </italic></sub>for those where column <italic>j </italic>has <italic>mean</italic>(<italic>H</italic><sub><italic>nj</italic></sub>) > <italic>mean</italic>(<italic>L</italic><sub><italic>nj</italic></sub>), and <italic>DOWN</italic><sub><italic>n </italic></sub>for those columns <italic>j </italic>where <italic>mean</italic>(<italic>H</italic><sub><italic>nj</italic></sub>) <<italic>mean</italic>(<italic>L</italic><sub><italic>nj</italic></sub>). Remaining columns that do not show significant variation are placed in the column set <italic>FLAT</italic><sub><italic>n</italic></sub>. The column sets are also meaningful outputs of our procedure, as <italic>UP</italic><sub><italic>n </italic></sub>and <italic>DOWN</italic><sub><italic>n </italic></sub>describe the groups of columns in which the extreme genes <italic>H</italic><sub><italic>n </italic></sub>and <italic>L</italic><sub><italic>n </italic></sub>vary significantly. Our procedure can output these columns and column sets in various orders simply as an aid to human interpretation, including: original column order; grouped by set and within set ordered by the Wilcoxon p-value significance; ordered by mean difference, <italic>mean</italic>(<italic>H</italic><sub><italic>nj</italic></sub>) – <italic>mean</italic>(<italic>L</italic><sub><italic>nj</italic></sub>); or ordered by the eigenvector column loading. Taken together, the PCEG point sets and significant column sets should provide valuable insight to researchers wishing to interpret each of the sources of variation identified by the principal components procedure.</p></sec><sec><title>Interpret each principal component using covariate annotations</title><p>When provided additional covariate annotations <italic>C</italic>, the procedure seeks to determine which, if any of the annotations <italic>C</italic><sub><italic>k </italic></sub>are well correlated with the partitioning of columns into the sets {<italic>UP</italic><sub><italic>n</italic></sub>, <italic>FLAT</italic><sub><italic>n</italic></sub>, <italic>DOWN</italic><sub><italic>n</italic></sub>}. A discrete annotation <italic>C</italic><sub><italic>k </italic></sub>containing <italic>m </italic>unique values <italic>V<sub>1</sub></italic>, <italic>V<sub>2</sub></italic>, ... <italic>V</italic><sub><italic>m </italic></sub>also defines a partitioning of the columns {<italic>KV<sub>1</sub></italic>, <italic>KV<sub>2</sub></italic>, ..., <italic>KV</italic><sub><italic>m</italic></sub>} where <italic>KV<sub>1 </sub></italic>is the set of columns that share the value <italic>V<sub>1</sub></italic>, <italic>KV<sub>2 </sub></italic>are those that share value <italic>V<sub>2</sub></italic>, and so on. An information theoretic measure known as normalized mutual information (NMI) [<xref ref-type="bibr" rid="B28">28</xref>] describes the degree to which two discrete random variables share information. When there is high mutual information, knowing the value of one of the variables should be useful predictor of the other variable. (See [<xref ref-type="bibr" rid="B9">9</xref>] for a description of the merits of NMI in terms of clustering and understanding the relationships between clusterings.) We construct the <italic>3 x m </italic>confusion matrix to compare the {<italic>UP</italic><sub><italic>n</italic></sub>, <italic>FLAT</italic><sub><italic>n</italic></sub>, <italic>DOWN</italic><sub><italic>n</italic></sub>} column partitioning with the {<italic>KV<sub>1</sub></italic>, <italic>KV<sub>2</sub></italic>, ..., <italic>KV</italic><sub><italic>m</italic></sub>} partitioning and calculate an NMI score between the partitionings. Because the usual NMI score is not symmetric (i.e. NMI(<italic>r</italic>,<italic>c</italic>) ≠ NMI(<italic>c</italic>,<italic>r</italic>)), we use a variant that we refer to as the average NMI score, which is simply the average of the NMI of the confusion matrix and the NMI of the transpose of the confusion matrix. Those covariates <italic>C</italic><sub><italic>k </italic></sub>having an average NMI greater than a user defined threshold <italic>nmiThresh </italic>are added to the set of significant covariate annotations <italic>A</italic><sub><italic>n</italic></sub>.</p><p>We apply a different approach when evaluating the <italic>C</italic><sub><italic>k </italic></sub>that are continuous covariates. We need to assess whether each principal component's {<italic>UP</italic><sub><italic>n</italic></sub>, <italic>FLAT</italic><sub><italic>n</italic></sub>, <italic>DOWN</italic><sub><italic>n</italic></sub>} column partitioning correlates with each <italic>C</italic><sub><italic>k </italic></sub>distribution of values. We can separately score three different partitioning schemes, <italic>UP</italic><sub><italic>n </italic></sub>vs. <italic>DOWN</italic><sub><italic>n</italic></sub>, {<italic>UP</italic><sub><italic>n</italic></sub>+<italic>FLAT</italic><sub><italic>n</italic></sub>} vs. <italic>DOWN</italic><sub><italic>n</italic></sub>, and <italic>UP</italic><sub><italic>n </italic></sub>vs.{<italic>FLAT</italic><sub><italic>n</italic></sub>+<italic>DOWN</italic><sub><italic>n</italic></sub>}, by determining if <italic>C</italic><sub><italic>k</italic></sub>'s value distributions differ significantly across the partition. E.g. does <italic>C</italic><sub><italic>k </italic></sub>within <italic>UP</italic><sub><italic>n </italic></sub>have a different distribution than <italic>C</italic><sub><italic>k </italic></sub>within <italic>DOWN</italic><sub><italic>n</italic></sub>? Next, does <italic>C</italic><sub><italic>k </italic></sub>within {<italic>UP</italic><sub><italic>n </italic></sub>+ <italic>FLAT</italic><sub><italic>n</italic></sub>} have a different distribution than <italic>C</italic><sub><italic>k </italic></sub>within <italic>DOWN</italic><sub><italic>n</italic></sub>? Finally, does <italic>C</italic><sub><italic>k </italic></sub>within <italic>UP</italic><sub><italic>n </italic></sub>have a different distribution than <italic>C</italic><sub><italic>k </italic></sub>within {<italic>FLAT</italic><sub><italic>n</italic></sub>+<italic>DOWN</italic><sub><italic>n</italic></sub>}? For each partition scheme we again use a two-tailed Wilcoxon rank sum test, including the small sample adjustments when sample size is less than 10, to determine whether the covariate's value distributions on each side of the partitioning differ significantly from each other. A <italic>minSetSize </italic>parameter can be specified as desired to reduce false positives when set sizes are very small, e.g. when comparing a distribution of 2 values vs. 7 values. Thus we calculate the Wilcoxon p-value for three partitionings of columns: <italic>UP</italic><sub><italic>n </italic></sub>vs. <italic>DOWN</italic><sub><italic>n</italic></sub>, {<italic>UP</italic><sub><italic>n</italic></sub>+<italic>FLAT</italic><sub><italic>n</italic></sub>} vs. <italic>DOWN</italic><sub><italic>n</italic></sub>, and <italic>UP</italic><sub><italic>n </italic></sub>vs. {<italic>FLAT</italic><sub><italic>n</italic></sub>+<italic>DOWN</italic><sub><italic>n</italic></sub>}. Those covariates having a p-value less than the user defined threshold <italic>test2Thresh </italic>for any of the three partitionings are added to the set of significant covariate annotations <italic>A</italic><sub><italic>n</italic></sub>.</p><p>Upon completion of the covariate analysis, covariates in the set <italic>A</italic><sub><italic>n</italic></sub>, that previously met user-controlled significance thresholds are reported by the software. Covariate reports provide the following information: For discrete-valued covariates, corresponding confusion matrices and average NMI scores are reported; for continuous-valued covariates, the three <italic>Wilcoxon </italic>p-values are reported together with supporting plots illustrating the covariate distributions. There is no guarantee that any covariates will be significantly related to a principal component. Conversely, spurious relationships might be reported, especially in the case of small numbers of samples due to small column partitions. The tool simply points to those covariates related to a principal component that also satisfy a user-controlled significance threshold. It is up to the investigator to consider these hypotheses and to confirm the interesting ones through further investigation.</p></sec><sec><title>Terminating condition</title><p>We have shown that some large-scale expression datasets have biologically pertinent structure that is revealed by deep PC analysis that goes well beyond the first few principal components. However there are limits to the depth of mining and these limits depend on both size and character of the dataset. In all cases, the last principal component is not free to seek a source of variation because it must be orthogonal to all prior <italic>nc-1 </italic>components. To some degree that is also true of some portion of latter principal components that explain ever-diminishing fractions of the variance. We suggest that a natural terminating condition exists: When a principal component cannot find any columns in which the extreme gene sets show significant differences, there is no need to proceed to subsequent principal components. We observe that this condition is often not met because the extreme genes are typically differentially expressed in at least a few of the original columns (the original axes or dimensions), even for the most minor principal components. We also observe that variants of a dataset (e.g. representative column subsets) can affect the relative ordering, but not the existence, of multiple factors or sources of variation that are reflected in the minor principal component regime. We may therefore choose to investigate all principal components, but do so with the expectation that minor principal components will describe increasingly subtle sources of variation, which can, and often do, include noisy processes inherent in the data source.</p></sec></sec><sec><title>Authors' contributions</title><p>JR and CH conceived the methodology of exhaustively analyzing and interpreting principal components, in particular how to identify extreme genes and significant conditions, and how to automate correlating these conditions with covariates to aid interpretation. JR, BK, DT and CH carried out the software development. JR carried out the initial PCA interpretation studies and drafted the manuscript. BK and DT performed additional PCA analyses and results interpretation. AM performed the gene set GO term enrichment analysis. BW conceived of the GNF dataset interpretation study, participated in its design and results interpretation, and helped to draft the manuscript. All authors read and approved the final 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</title><p>The supplemental files provided with this publication are only a representative set of those generated by the PCA interpretation software. The complete collections of PCA interpretation results for both the GNF and diabetes datasets are provided as a supplement to this publication at [<xref ref-type="bibr" rid="B35">35</xref>].</p><p>Tab-delimited text file listing the high and low extreme genes for PC<sub>7</sub>, including PC<sub>7 </sub>coefficient and additional GNF gene annotations: ProbeId, Name, Aliases, Description, Function and Protein Families.</p></caption><media xlink:href="1471-2105-7-194-S1.txt" mimetype="text" mime-subtype="plain"><caption><p>Click here for file</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="S2"><caption><title>Additional File 2</title><p>The supplemental files provided with this publication are only a representative set of those generated by the PCA interpretation software. The complete collections of PCA interpretation results for both the GNF and diabetes datasets are provided as a supplement to this publication at [<xref ref-type="bibr" rid="B35">35</xref>].</p><p>Tab-delimited text file listing the conditions that are up, flat and down for PC<sub>7</sub>, ordered by decreasing difference of means.</p></caption><media xlink:href="1471-2105-7-194-S2.txt" mimetype="text" mime-subtype="plain"><caption><p>Click here for file</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="S3"><caption><title>Additional File 3</title><p>The supplemental files provided with this publication are only a representative set of those generated by the PCA interpretation software. The complete collections of PCA interpretation results for both the GNF and diabetes datasets are provided as a supplement to this publication at [<xref ref-type="bibr" rid="B35">35</xref>].</p><p>Trajectory plot of the PC<sub>7 </sub>eigenvector, or "eigen-condition".</p></caption><media xlink:href="1471-2105-7-194-S3.png" mimetype="text" mime-subtype="plain"><caption><p>Click here for file</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="S4"><caption><title>Additional File 4</title><p>The supplemental files provided with this publication are only a representative set of those generated by the PCA interpretation software. The complete collections of PCA interpretation results for both the GNF and diabetes datasets are provided as a supplement to this publication at [<xref ref-type="bibr" rid="B35">35</xref>].</p><p>Scatter plot of gene probe expression levels projected onto PC<sub>6 </sub>vs. PC<sub>7 </sub>space. The PC<sub>7 </sub>high and low extreme gene sets are highlighted in red and blue colors, respectively.</p></caption><media xlink:href="1471-2105-7-194-S4.png" mimetype="text" mime-subtype="plain"><caption><p>Click here for file</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="S5"><caption><title>Additional File 5</title><p>The supplemental files provided with this publication are only a representative set of those generated by the PCA interpretation software. The complete collections of PCA interpretation results for both the GNF and diabetes datasets are provided as a supplement to this publication at [<xref ref-type="bibr" rid="B35">35</xref>].</p><p>Gene trajectory plots for PC<sub>7 </sub>high and low extreme gene sets with tissues in the order in which the original data were provided.</p></caption><media xlink:href="1471-2105-7-194-S5.png" mimetype="text" mime-subtype="plain"><caption><p>Click here for file</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="S6"><caption><title>Additional File 6</title><p>The supplemental files provided with this publication are only a representative set of those generated by the PCA interpretation software. The complete collections of PCA interpretation results for both the GNF and diabetes datasets are provided as a supplement to this publication at [<xref ref-type="bibr" rid="B35">35</xref>].</p><p>Gene trajectory plots for PC<sub>7 </sub>high and low extreme gene sets with tissues ordered by decreasing mean differences, and thus grouped by significance (up group at left, flat group in middle and low group at right).</p></caption><media xlink:href="1471-2105-7-194-S6.png" mimetype="text" mime-subtype="plain"><caption><p>Click here for file</p></caption></media></supplementary-material></sec> |
Automated in-silico detection of cell populations in flow cytometry readouts and its application to leukemia disease monitoring | <sec><title>Background</title><p>Identification of minor cell populations, e.g. leukemic blasts within blood samples, has become increasingly important in therapeutic disease monitoring. Modern flow cytometers enable researchers to reliably measure six and more variables, describing cellular size, granularity and expression of cell-surface and intracellular proteins, for thousands of cells per second. Currently, analysis of cytometry readouts relies on visual inspection and manual gating of one- or two-dimensional projections of the data. This procedure, however, is labor-intensive and misses potential characteristic patterns in higher dimensions.</p></sec><sec><title>Results</title><p>Leukemic samples from patients with acute lymphoblastic leukemia at initial diagnosis and during induction therapy have been investigated by 4-color flow cytometry. We have utilized multivariate classification techniques, <italic>Support Vector Machines </italic>(SVM), to automate leukemic cell detection in cytometry. Classifiers were built on conventionally diagnosed training data. We assessed the detection accuracy on independent test data and analyzed marker expression of incongruently classified cells. SVM classification can recover manually gated leukemic cells with 99.78% sensitivity and 98.87% specificity.</p></sec><sec><title>Conclusion</title><p>Multivariate classification techniques allow for automating cell population detection in cytometry readouts for diagnostic purposes. They potentially reduce time, costs and arbitrariness associated with these procedures. Due to their multivariate classification rules, they also allow for the reliable detection of small cell populations.</p></sec> | <contrib id="A1" contrib-type="author"><name><surname>Toedling</surname><given-names>Joern</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I3">3</xref><email>toedling@ebi.ac.uk</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Rhein</surname><given-names>Peter</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>rhein@rrk.charite-buch.de</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Ratei</surname><given-names>Richard</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>rratei@berlin.helios-kliniken.de</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Karawajew</surname><given-names>Leonid</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>karawajew@rrk.charite-buch.de</email></contrib><contrib id="A5" corresp="yes" contrib-type="author"><name><surname>Spang</surname><given-names>Rainer</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>spang@molgen.mpg.de</email></contrib> | BMC Bioinformatics | <sec><title>Background</title><p>Flow cytometry has evolved to an indispensable tool in biology and medicine, with a significant impact on hematology. To date, diagnosis and classification of acute lymphocytic leukemia (ALL), depend on the flow-cytometric description of the leukemic cell clone. Recently, flow cytometry has also become an attractive approach for evaluation of therapy response and especially detection of minimal residual disease (MRD) [<xref ref-type="bibr" rid="B1">1</xref>]. Flow cytometry provides a quantitative cell description by a number of variables, including cell size, granularity and expression of cell-surface and intracellular proteins. Due to the continuous development of flow cytometric techniques, their readouts have become increasingly complex and require adequate analysis methods.</p><p>Current diagnostic evaluation of flow cytometry readouts relies on simplistic two-dimensional analysis techniques. The basis is a labor-intensive <italic>gating </italic>procedure. In a series of two-dimensional dot plots, leukemic cells are manually flagged by drawing polygons around regions, which are known to contain mostly leukemic cells. A large number of two-dimensional plots need to be inspected and several regions need to be defined manually. Finally, candidates for leukemic cells are those inside a Boolean combination of drawn regions, called <italic>gate </italic>[<xref ref-type="bibr" rid="B2">2</xref>,<xref ref-type="bibr" rid="B3">3</xref>]. From the data analysis perspective, gating lymphoblastic cells is a problem of supervised statistical learning. One starts with a training set of flow cytometry readouts, which are already gated by an expert. The challenge is to derive a multivariate classification model from this data, which is able to produce accurate gatings on different readouts that have not been pre-gated by an expert. The objects of classification are single cells that can either be leukemic or physiological lymphocytes. Typically, each cell is described by a 4–9 dimensional vector of flow cytometry measurements. With only a small number of pre-gated readouts, one already has several thousand training points. Problems are posed by the non-linear shape of the regions containing leukemic lymphocytes and the patient-to-patient variability of these regions.</p><p>To our knowledge, replacing the manual gating process by a computer-based automated multivariate analysis has not been described previously. Some cytometer software suites contain tools for automated <italic>walk-away </italic>analyses once the gates have been defined. These tools, however, are also restricted to two-dimensional decision rules. In addition, some methods to make use of cytometry readouts in a multivariate setting have been proposed. Here, we briefly review three of these methods and explain the differences to our concept.</p><p>Valet et al. [<xref ref-type="bibr" rid="B4">4</xref>] introduced a classification method for blood samples in flow cytometry, called <italic>algorithmic data sieving</italic>. For each class of samples, a discretized representative is derived from training data. New samples are then classified according to their similarity to these representatives. De Zen et al. [<xref ref-type="bibr" rid="B5">5</xref>] investigated the feasibility to classify acute-leukemia subtypes on flow-cytometry readouts. First, they determined the leukemic cells by a conventional, manual gating procedure and discarded all other cells from the data. For each sample, they summarized the measurements over all leukemic cells for each variable and used these summary values for classifying samples with linear discriminant analysis [<xref ref-type="bibr" rid="B6">6</xref>]. Roederer and Hardy [<xref ref-type="bibr" rid="B7">7</xref>] proposed an algorithm for sample comparison based on cytometry readouts. Their algorithm identifies multi-dimensional hyper-rectangular bins that significantly differ in the proportion of cells contained between a test sample and a control sample. The union of all these regions comprises a <italic>frequency difference gate</italic>. This gate may be used to assign new samples to test or control group, as well as to find differences between similar types of cells under different conditions. While their approach could be modified to search for regions in multivariate space, which differ between leukemic and non-leukemic samples, they restrict these regions to be rectangles, which is not the case in conventional gating [<xref ref-type="bibr" rid="B3">3</xref>]. Moreover, in this case regions are not required to contain the same proportion of cells but rather to contain mostly cells of the same class. In contrast to these approaches, we are not interested in classifying blood samples based on their cytometry readouts, but rather in automated identification of cell populations within the samples. We report on the applicability of statistical learning methodology, for achieving automated, reliable in-silico gatings on flow cytometry readouts. To this aim, we employ supervised classification with <italic>Support Vector Machines </italic>[<xref ref-type="bibr" rid="B8">8</xref>].</p></sec><sec><title>Results</title><sec><title>Algorithm</title><p>For supervised classification of the leukemic status of cells, we employ a Support Vector Machine based algorithm that allows for non-linear decision boundaries in the input space spanned by the cells' measured characteristics and protein expression levels. Our algorithm takes into account outstanding properties of flow-cytometry readout data, namely</p><p>• samples consisting of tens of thousands of individual observations</p><p>• large inter-sample variation due to non-standardized methods of obtaining measurements.</p><sec><title>Support Vector Machines</title><p><italic>Support Vector Machines </italic>(SVM) [<xref ref-type="bibr" rid="B9">9</xref>,<xref ref-type="bibr" rid="B10">10</xref>] are a class of regularized multivariate classification models that are widely used for predictive modelling of multidimensional data. We provide a quick review of SVM here. Let <bold><italic>X </italic></bold>be the data matrix holding <italic>n </italic>observations <bold><italic>x</italic></bold><sub><italic>i</italic></sub>, with <italic>i </italic>∈ 1,..., <italic>n</italic>, in columns and <italic>p </italic>variables in rows. The observations <bold><italic>x</italic></bold><sub><italic>i </italic></sub>are said to reside in a <italic>p</italic>-dimensional input space. For each observation <bold><italic>x</italic></bold><sub><italic>i </italic></sub>its class (clinical phenotype) <italic>y</italic><sub><italic>i </italic></sub>∈ {± 1} is known beforehand in case of the training set or to be predicted in case of the test set. SVM fit a maximal (soft) margin hyperplane between the two classes. With high-dimensional problems, there may be several perfectly separating hyperplanes (the maximum likelihood approach leads to an ill-posed problem). There is, however, only one separating hyperplane with maximal distance to the nearest training points of either class.</p><p>More formally, among all hyperplanes of the form</p><p><bold>w</bold>·<bold><italic>x </italic></bold>+ <italic>b </italic>= 0 | <bold>w </bold>∈ ℝ<sup><italic>N</italic></sup>, <italic>b </italic>∈ ℝ</p><p>corresponding to linear decision functions</p><p><italic>c</italic>(<bold><italic>x</italic></bold><sub><italic>j</italic></sub>) = sgn(<bold>w</bold>·<bold><italic>x</italic></bold><sub><italic>j </italic></sub>+ <italic>b</italic>)</p><p>there exists one that maximizes the distance of each input vector to the hyperplane. It can be shown, that this <italic>optimal hyperplane </italic>can be empirically obtained from data <bold><italic>X </italic></bold>by solving</p><p><inline-formula><mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="M1" name="1471-2105-7-282-i1" overflow="scroll">
<mml:semantics definitionURL="" encoding="">
<mml:mrow>
<mml:mtable>
<mml:mtr>
<mml:mtd>
<mml:mrow>
<mml:munder>
<mml:mrow>
<mml:mi>min</mml:mi>
<mml:mo>⁡</mml:mo>
</mml:mrow>
<mml:mrow>
<mml:mi>w</mml:mi>
<mml:mo>,</mml:mo>
<mml:mi>b</mml:mi>
</mml:mrow>
</mml:munder>
<mml:mrow>
<mml:mo>{</mml:mo>
<mml:mrow>
<mml:mfrac>
<mml:mn>1</mml:mn>
<mml:mn>2</mml:mn>
</mml:mfrac>
<mml:msup>
<mml:mrow>
<mml:mrow>
<mml:mo>‖</mml:mo>
<mml:mi>w</mml:mi>
<mml:mo>‖</mml:mo>
</mml:mrow>
</mml:mrow>
<mml:mn>2</mml:mn>
</mml:msup>
</mml:mrow>
<mml:mo>}</mml:mo>
</mml:mrow>
</mml:mrow>
</mml:mtd>
<mml:mtd>
<mml:mrow>
<mml:mtext>subject to</mml:mtext>
</mml:mrow>
</mml:mtd>
<mml:mtd>
<mml:mrow>
<mml:msub>
<mml:mi>y</mml:mi>
<mml:mi>i</mml:mi>
</mml:msub>
<mml:mo>⋅</mml:mo>
<mml:mo stretchy="false">(</mml:mo>
<mml:mi>w</mml:mi>
<mml:mo>⋅</mml:mo>
<mml:msub>
<mml:mi>x</mml:mi>
<mml:mi>i</mml:mi>
</mml:msub>
<mml:mo>+</mml:mo>
<mml:mi>b</mml:mi>
<mml:mo stretchy="false">)</mml:mo>
<mml:mo>≥</mml:mo>
<mml:mn>1</mml:mn>
<mml:mo>,</mml:mo>
<mml:mtext> </mml:mtext>
<mml:mi>i</mml:mi>
<mml:mo>=</mml:mo>
<mml:mn>1</mml:mn>
<mml:mo>,</mml:mo>
<mml:mo>…</mml:mo>
<mml:mo>,</mml:mo>
<mml:mi>n</mml:mi>
</mml:mrow>
</mml:mtd>
</mml:mtr>
</mml:mtable>
</mml:mrow>
<mml:annotation encoding="MathType-MTEF">
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaafaqabeqadaaabaWaaCbeaeaacyGGTbqBcqGGPbqAcqGGUbGBaSqaaGqabiab=Dha3jabcYcaSiabdkgaIbqabaGcdaGadaqaamaalaaabaGaeGymaedabaGaeGOmaidaamaafmaabaGae83DaChacaGLjWUaayPcSdWaaWbaaSqabeaacqaIYaGmaaaakiaawUhacaGL9baaaeaacqqGZbWCcqqG1bqDcqqGIbGycqqGQbGAcqqGLbqzcqqGJbWycqqG0baDcqqGGaaicqqG0baDcqqGVbWBaeaacqWG5bqEdaWgaaWcbaGaemyAaKgabeaakiabgwSixlabcIcaOiab=Dha3jabgwSixJqadiab+Hha4naaBaaaleaacqWGPbqAaeqaaOGaey4kaSIaemOyaiMaeiykaKIaeyyzImRaeGymaeJaeiilaWIaeeiiaaIaemyAaKMaeyypa0JaeGymaeJaeiilaWIaeSOjGSKaeiilaWIaemOBa4gaaaaa@67D6@</mml:annotation>
</mml:semantics>
</mml:math></inline-formula></p><p>where</p><p><inline-formula><mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="M2" name="1471-2105-7-282-i2" overflow="scroll">
<mml:semantics definitionURL="" encoding="">
<mml:mrow>
<mml:mi>w</mml:mi>
<mml:mo>=</mml:mo>
<mml:mstyle displaystyle="true">
<mml:munderover>
<mml:mo>∑</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:munderover>
<mml:mrow>
<mml:msub>
<mml:mi>α</mml:mi>
<mml:mi>i</mml:mi>
</mml:msub>
<mml:msub>
<mml:mi>y</mml:mi>
<mml:mi>i</mml:mi>
</mml:msub>
<mml:msub>
<mml:mi>x</mml:mi>
<mml:mi>i</mml:mi>
</mml:msub>
</mml:mrow>
</mml:mstyle>
<mml:mo>|</mml:mo>
<mml:msub>
<mml:mi>α</mml:mi>
<mml:mi>i</mml:mi>
</mml:msub>
<mml:mo>∈</mml:mo>
<mml:mi>ℝ</mml:mi>
<mml:mo>.</mml:mo>
</mml:mrow>
<mml:annotation encoding="MathType-MTEF">
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaaieqacqWF3bWDcqGH9aqpdaaeWbqaaGGaciab+f7aHnaaBaaaleaacqWGPbqAaeqaaOGaemyEaK3aaSbaaSqaaiabdMgaPbqabaacbmGccqqF4baEdaWgaaWcbaGaemyAaKgabeaaaeaacqWGPbqAcqGH9aqpcqaIXaqmaeaacqWGUbGBa0GaeyyeIuoakiabcYha8jab+f7aHnaaBaaaleaacqWGPbqAaeqaaOGaeyicI48efv3ySLgznfgDOjdaryqr1ngBPrginfgDObcv39gaiqaacqaFDeIuieaacqWEUaGlaaa@5136@</mml:annotation>
</mml:semantics>
</mml:math></inline-formula></p><p>In practice, <italic>α</italic><sub><italic>i </italic></sub>will often be zero for many <italic>i </italic>and different from zero for only a limited number of <italic>m </italic>observations with <italic>m </italic>≤ <italic>n</italic>. These <italic>m </italic>observations solely define the separating hyperplane and are called <italic>support vectors</italic>. The maximal-margin concept is typically combined with the <italic>kernel trick </italic>to allow for flexible non-linear classification boundaries. The kernel trick is applicable to classification algorithms that can be expressed in terms of inner products of the inputs, as it is the case for the maximum-margin hyperplane. The inner products are substituted by a <italic>kernel function k</italic>(<bold><italic>x</italic></bold><sub><italic>i</italic></sub>, <bold><italic>x</italic></bold><sub><italic>j</italic></sub>), which corresponds to a <italic>feature map </italic>Φ that maps the profiles from the input space into a <italic>feature space </italic><inline-formula><mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="M3" name="1471-2105-7-282-i3" overflow="scroll"><mml:semantics definitionURL="" encoding=""><mml:mi>ℋ</mml:mi><mml:annotation encoding="MathType-MTEF">
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBamrtHrhAL1wy0L2yHvtyaeHbnfgDOvwBHrxAJfwnaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaWaaeGaeaaakeaaimaacqWFlecsaaa@3763@</mml:annotation></mml:semantics></mml:math></inline-formula>:</p><p>Φ: ℝ<sup><italic>p </italic></sup>→ <inline-formula><mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="M4" name="1471-2105-7-282-i3" overflow="scroll"><mml:semantics definitionURL="" encoding=""><mml:mi>ℋ</mml:mi><mml:annotation encoding="MathType-MTEF">
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBamrtHrhAL1wy0L2yHvtyaeHbnfgDOvwBHrxAJfwnaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaWaaeGaeaaakeaaimaacqWFlecsaaa@3763@</mml:annotation></mml:semantics></mml:math></inline-formula></p><p><bold><italic>x </italic></bold>↦ Φ(<italic>x</italic>).</p><p>This results in the original algorithm being carried out in <inline-formula><mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="M5" name="1471-2105-7-282-i3" overflow="scroll"><mml:semantics definitionURL="" encoding=""><mml:mi>ℋ</mml:mi><mml:annotation encoding="MathType-MTEF">
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBamrtHrhAL1wy0L2yHvtyaeHbnfgDOvwBHrxAJfwnaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaWaaeGaeaaakeaaimaacqWFlecsaaa@3763@</mml:annotation></mml:semantics></mml:math></inline-formula> now and leads to non-linear decision boundaries in the input space. After application of the kernel trick, the decision functions are of the more general form</p><p><inline-formula><mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="M6" name="1471-2105-7-282-i4" overflow="scroll">
<mml:semantics definitionURL="" encoding="">
<mml:mrow>
<mml:mi>c</mml:mi>
<mml:mo stretchy="false">(</mml:mo>
<mml:msub>
<mml:mi>x</mml:mi>
<mml:mi>j</mml:mi>
</mml:msub>
<mml:mo stretchy="false">)</mml:mo>
<mml:mo>=</mml:mo>
<mml:mi>sgn</mml:mi>
<mml:mo>⁡</mml:mo>
<mml:mrow>
<mml:mo>(</mml:mo>
<mml:mrow>
<mml:mstyle displaystyle="true">
<mml:munderover>
<mml:mo>∑</mml:mo>
<mml:mi>i</mml:mi>
<mml:mi>n</mml:mi>
</mml:munderover>
<mml:mrow>
<mml:msub>
<mml:mi>y</mml:mi>
<mml:mi>i</mml:mi>
</mml:msub>
<mml:msub>
<mml:mi>α</mml:mi>
<mml:mi>i</mml:mi>
</mml:msub>
<mml:mi>k</mml:mi>
<mml:mo stretchy="false">(</mml:mo>
<mml:msub>
<mml:mi>x</mml:mi>
<mml:mi>j</mml:mi>
</mml:msub>
<mml:mo>,</mml:mo>
<mml:msub>
<mml:mi>x</mml:mi>
<mml:mi>i</mml:mi>
</mml:msub>
<mml:mo stretchy="false">)</mml:mo>
<mml:mo>+</mml:mo>
<mml:mi>b</mml:mi>
</mml:mrow>
</mml:mstyle>
</mml:mrow>
<mml:mo>)</mml:mo>
</mml:mrow>
<mml:mo>.</mml:mo>
</mml:mrow>
<mml:annotation encoding="MathType-MTEF">
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGJbWycqGGOaakieWacqWF4baEdaWgaaWcbaGaemOAaOgabeaakiabcMcaPiabg2da9iGbcohaZjabcEgaNjabc6gaUnaabmaabaWaaabCaeaacqWG5bqEdaWgaaWcbaGaemyAaKgabeaaiiGakiab+f7aHnaaBaaaleaacqWGPbqAaeqaaOGaem4AaSMaeiikaGIae8hEaG3aaSbaaSqaaiabdQgaQbqabaGccqGGSaalcqWF4baEdaWgaaWcbaGaemyAaKgabeaakiabcMcaPiabgUcaRiabdkgaIbWcbaGaemyAaKgabaGaemOBa4ganiabggHiLdaakiaawIcacaGLPaaacqGGUaGlaaa@51DA@</mml:annotation>
</mml:semantics>
</mml:math></inline-formula></p></sec><sec><title>Soft margin classification</title><p>Noise may cause large overlap of the classes even in the feature space, such that a perfectly separating hyperplane may not exist. In this case, one can allow for misclassifications (margin violations) by relaxing the optimization constraints to</p><p><italic>y</italic><sub><italic>i</italic></sub>·(<bold>w</bold>·<italic>x</italic><sub><italic>i </italic></sub>+ <italic>b</italic>) ≥ 1 - <italic>ξ</italic><sub><italic>i</italic></sub>, <italic>i </italic>= 1,..., <italic>n</italic>.</p><p><italic>ξ</italic><sub><italic>i </italic></sub>≥ 0 are commonly called <italic>slack variables</italic>. The optimal <italic>soft </italic>margin classifier is then found by minimizing the regularized risk function</p><p><inline-formula><mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="M7" name="1471-2105-7-282-i5" overflow="scroll">
<mml:semantics definitionURL="" encoding="">
<mml:mrow>
<mml:mi>r</mml:mi>
<mml:mo stretchy="false">(</mml:mo>
<mml:mi>w</mml:mi>
<mml:mo>,</mml:mo>
<mml:mi>ξ</mml:mi>
<mml:mo stretchy="false">)</mml:mo>
<mml:mo>=</mml:mo>
<mml:mfrac>
<mml:mn>1</mml:mn>
<mml:mn>2</mml:mn>
</mml:mfrac>
<mml:msup>
<mml:mrow>
<mml:mrow>
<mml:mo>‖</mml:mo>
<mml:mi>w</mml:mi>
<mml:mo>‖</mml:mo>
</mml:mrow>
</mml:mrow>
<mml:mn>2</mml:mn>
</mml:msup>
<mml:mo>+</mml:mo>
<mml:mi>C</mml:mi>
<mml:mstyle displaystyle="true">
<mml:munderover>
<mml:mo>∑</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:munderover>
<mml:mrow>
<mml:msub>
<mml:mi>ξ</mml:mi>
<mml:mi>i</mml:mi>
</mml:msub>
</mml:mrow>
</mml:mstyle>
<mml:mo>.</mml:mo>
</mml:mrow>
<mml:annotation encoding="MathType-MTEF">
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGYbGCcqGGOaakieqacqWF3bWDcqGGSaaliiGacqGF+oaEcqGGPaqkcqGH9aqpdaWcaaqaaiabigdaXaqaaiabikdaYaaadaqbdaqaaiab=Dha3bGaayzcSlaawQa7amaaCaaaleqabaGaeGOmaidaaOGaey4kaSIaem4qam0aaabCaeaacqGF+oaEdaWgaaWcbaGaemyAaKgabeaaaeaacqWGPbqAcqGH9aqpcqaIXaqmaeaacqWGUbGBa0GaeyyeIuoakiabc6caUaaa@49BC@</mml:annotation>
</mml:semantics>
</mml:math></inline-formula></p><p>In this formulation, perfect separability of the two classes is not required and margin violations are allowed. The trade off between margin violations and margin size is reflected by the regularization (cost) parameter <italic>C</italic>.</p><p>Regularization is essential to counter the additional flexibility acquired by use of the kernel trick.</p><p>In recent years, SVM have proven to be a powerful and robust classification method that can handle various kinds of input data [<xref ref-type="bibr" rid="B11">11</xref>]. For a more extensive introduction to SVM see, e.g, [<xref ref-type="bibr" rid="B8">8</xref>].</p><p>We chose a radial-basis kernel function, which can be defined as:</p><p><italic>k</italic><sub><italic>γ</italic></sub>(<bold><italic>x</italic></bold><sub><italic>i</italic></sub>, <bold><italic>x</italic></bold><sub><italic>j</italic></sub>) = exp(-<italic>γ</italic>·||<bold><italic>x</italic></bold><sub><italic>i </italic></sub>- <bold><italic>x</italic></bold><sub><italic>j</italic></sub>||<sup>2</sup>)</p><p>where <italic>x</italic><sub><italic>i </italic></sub>and <bold><italic>x</italic></bold><sub><italic>j </italic></sub>are two input data points and <italic>γ </italic>is the inverse band width of the smoothing kernel.</p></sec><sec><title>Building the classifier</title><p>With our data, we applied the SVM to separate leukemic from non-leukemic cells. The optimal settings for the SVM parameters <italic>C </italic>and <italic>γ </italic>were determined on set-aside calibration data, while the actual performance of the SVM classifier was analyzed on another set-aside test data set. Since our samples each contained up to 300,000 cells, requiring highly demanding computations, we had to think about methods of data reduction. Since with SVM classification, the decision boundary is only determined by the <italic>support vectors </italic>[[<xref ref-type="bibr" rid="B8">8</xref>], chap. 1], we can discard all observations that are not support vectors from the training data. We split the data into subsets, keeping only the support vectors from each subset and build the final SVM classifier on the sets of support vectors, similar to Boser et al. [<xref ref-type="bibr" rid="B12">12</xref>]. In detail, we used the following procedure to estimate the test error of the learned classifier:</p><p>1. Split data into 50% training set, 25% calibration set, and 25% test set.</p><p>2. For a reasonable number of possible parameter settings Θ<sub><italic>k </italic></sub>= (<italic>γ</italic><sub><italic>k</italic></sub>, <italic>C</italic><sub><italic>k</italic></sub>):</p><p>(a) initialize an empty set of training vectors <inline-formula><mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="M8" name="1471-2105-7-282-i6" overflow="scroll"><mml:semantics definitionURL="" encoding=""><mml:mi mathvariant="script">T</mml:mi><mml:annotation encoding="MathType-MTEF">
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBamrtHrhAL1wy0L2yHvtyaeHbnfgDOvwBHrxAJfwnaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaWaaeGaeaaakeaaimaacqWFtepvaaa@3847@</mml:annotation></mml:semantics></mml:math></inline-formula><sub><italic>k </italic></sub>= { }</p><p>(b) split training set into computationally feasible subsets</p><p>(c) for each of these subsets</p><p>• learn SVM-classifier for leukemic versus non-leukemic cells</p><p>• identify support vectors <italic>S </italic>on this subset</p><p>• include these support vectors into the set of training vectors <inline-formula><mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="M9" name="1471-2105-7-282-i6" overflow="scroll"><mml:semantics definitionURL="" encoding=""><mml:mi mathvariant="script">T</mml:mi><mml:annotation encoding="MathType-MTEF">
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBamrtHrhAL1wy0L2yHvtyaeHbnfgDOvwBHrxAJfwnaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaWaaeGaeaaakeaaimaacqWFtepvaaa@3847@</mml:annotation></mml:semantics></mml:math></inline-formula><sub><italic>k </italic></sub>= <inline-formula><mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="M10" name="1471-2105-7-282-i6" overflow="scroll"><mml:semantics definitionURL="" encoding=""><mml:mi mathvariant="script">T</mml:mi><mml:annotation encoding="MathType-MTEF">
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBamrtHrhAL1wy0L2yHvtyaeHbnfgDOvwBHrxAJfwnaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaWaaeGaeaaakeaaimaacqWFtepvaaa@3847@</mml:annotation></mml:semantics></mml:math></inline-formula><sub><italic>k </italic></sub>∪ <italic>S</italic></p><p>(d) learn SVM-classifier on <inline-formula><mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="M11" name="1471-2105-7-282-i6" overflow="scroll"><mml:semantics definitionURL="" encoding=""><mml:mi mathvariant="script">T</mml:mi><mml:annotation encoding="MathType-MTEF">
MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBamrtHrhAL1wy0L2yHvtyaeHbnfgDOvwBHrxAJfwnaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaWaaeGaeaaakeaaimaacqWFtepvaaa@3847@</mml:annotation></mml:semantics></mml:math></inline-formula><sub><italic>k</italic></sub></p><p>(e) use learned classifier on calibration set to compute the <italic>calibration error </italic><italic>ϕ</italic><sub><italic>k</italic></sub></p><p>3. Keep learned classifier with lowest calibration error.</p><p>4. Use this classifier to predict test data.</p><p>5. Return prediction error on test set as <italic>test error</italic>.</p><p>Keeping only the support vectors from each data subset reduces the amounts of involved data and renders the computations feasible on common present-day computers.</p><p>The final test error provides an unbiased estimate of the classifier's prediction error (generalization error) on new data [13, chap. 7].</p></sec><sec><title>Artificial noise</title><p>We are interested in building classifiers that are insensitive to minor noise induced into experimental measurements by the experimental setup or measuring device. By generating artificial noise and applying it to training data, one tests the ability to learn a "correct" concept in the presence of noise [<xref ref-type="bibr" rid="B14">14</xref>]. Here, we add artificial noise to our training data to simulate such minor experimental variations. As noise, we take random normally distributed numbers. Classification rules that are built on noisy versions of the training data and apply to artificial-noise-free test data as well, have the potential to generalize better to new test samples [<xref ref-type="bibr" rid="B15">15</xref>].</p></sec></sec><sec><title>Application on patient data</title><p>In each of 37 patient samples (Table <xref ref-type="table" rid="T1">1</xref>), the leukemic cells were identified by manual gating beforehand. We then randomly assigned 19 patient samples to a training set, while the remaining 18 patient samples made up the test set. We also included two other samples, for which the proportion of leukemic cells was predefined, into the training set. One of these samples originated from a non-leukemic blood sample (sample 38 in Table <xref ref-type="table" rid="T1">1</xref>) while the other one was enriched with leukemic cells by Ficoll gradient-density centrifugation and subsequent manual filtering of the flow cytometry data upon visual inspection (sample 39 in Table <xref ref-type="table" rid="T1">1</xref>).</p><p>From each sample of the training set, we randomly drew 10,000 cells and discarded all other cells. Thus, our training data consisted of 210,000 cells and the associated labels, either "leukemic" or "non-leukemic".</p><p>To simulate minor experiment-induced variations, we added random noise to the data. For each variable, we determined its standard deviation across all cells of the training data. We then drew random numbers from a normal distribution with mean 0 and standard deviation equal to 10% of the variable's observed standard deviation. These random numbers were added to the values of the training data.</p><p>On the noisy training data, we learned the optimal SVM classifier and evaluated its classification performance on the test set as well as on the two samples with predefined leukemic proportion.</p><p>For building the SVM classifier, we again split the training data into an actual training set, a calibration set and a test set to select the optimal parameter settings and to avoid overfitting.</p><p>The SVM defines a region in six-dimensional space containing the leukemic cells. Due to the kernel trick, the classification boundary between points in the six-dimensional space, at which cells would be classified as being leukemic, and the other points, at which cells would be classified as physiological blood cells, is a non-linear structure. To illustrate this, we presenl a projection of this region on the three-dimensional subspace spanned by the variables SSC, CD34 and CD10 (Figure <xref ref-type="fig" rid="F1">1</xref>).</p><p>The SVM classifier employs a radial-basis kernel function with parameters <italic>σ </italic>= 0.5 and <italic>C </italic>= 4. However, changing each parameter selling by up to 25% had no effect on the classification performance, underlining the robustness of the method. The SVM classifier is based 958 support vectors, of which 478 are leukemic cells. On the training data, the learned radial-basis SVM classifier achieves a classification accuracy of 99.6%.</p><p>For comparison, we also evaluated the classification performance of an SVM classifier utilizing a linear kernel function instead of a radial basis one. We observed a slightly worse performance of this classifier on the training data, namely an error of 0.7%.</p><p>We used the SVM classifier to predict leukemic cells in the two samples, either without or enriched with leukemic cells, respectively. Cells from both samples had been used for learning the SVM as well, but we made sure that the same cells would not be used for testing the SVM performance. From the non-leukemic sample, we drew a subsample of 20,000 cells at random to avoid exceeding available RAM during computations for the SVM prediction. Of these 20,000 cells, only 2 (0.01%) were misclassified as being leukemic (sample 38, Table <xref ref-type="table" rid="T1">1</xref>). From the enriched leukemic sample (containing approx. 94% leukemic blasts), we also drew a random subsample of 20,000 cells. Of these, the SVM classified 576 (2.88%) as being non-leukemic (sample 39, Table <xref ref-type="table" rid="T1">1</xref>).</p><p>We utilized the learned SVM to predict the leukemic status of the cells in the test set, which consisted of 18 patient samples that had not been involved in training the classifier. We compared the predicted leukemia status with the one determined by gating beforehand. On this independent test set, the SVM achieved a <italic>sensitivity </italic>of 99.78% and a <italic>specificity </italic>of 98.87% for predicting the leukemic status of cells. In summary, for 0.94% of the cells of the test data, the leukemia status differed between SVM prediction and gating assignment. The total number of true and wrong predictions can be seen in Table <xref ref-type="table" rid="T2">2</xref>.</p><p>To evaluate the robustness of the observed classification performance, we repeated the random splitting of patient samples into training and test set 500 times, and reran the full analysis for each split. The mean sensitivity was 98.06% (95% confidence interval [CI] 88.7% to 99.8%) and the mean specificity was 99.27% (95% CI 98.6% to 99.9%), which confirms the stability of our results. Across the 500 random splittings on average, we observed 2006 support vectors (sd: 496), which is more than for the original splitting.</p><p>To further assess the built classifier's precision, we took another independent peripheral blood sample, which was taken on the initial day of treatment, and separately measured aliquots from this sample with intervals of several minutes in between. Cytometer settings were not changed in between measurements. On each of the six readouts, we applied the classifier to predict the percentage of leukemic cells included. The predicted values ranged from 21.19% to 21.99% percent, similar to the manual-gating assigned percentage range that extented from 20.55% to 21.94%. We compared the spatial distribution of cells being classified as leukemic by the SVM to that of cells deemed leukemic due to manual gating and to the spatial distribution of all cells in the test data (see Figure <xref ref-type="fig" rid="F2">2</xref>). The distributions of cells deemed leukemic by manual gating and by SVM classification are nearly identical. However, the area in multivariate space, in which the SVM would assign cells to the leukemic class, is slightly larger than that defined by manual gating.</p><p>The large majority (96.8%) of the incongruently classified cells are deemed to be physiological blood cells by manual gating but predicted as being leukemic by the SVM. Most of these stem from samples taken before initial treatment (see Table <xref ref-type="table" rid="T2">2</xref>). Their scatter and fluorescence measurements, compared to those cells deemed leukemic by both methods, can be seen in Figure <xref ref-type="fig" rid="F3">3</xref>.</p></sec></sec><sec><title>Discussion</title><p>Modern flow cytometers enable researchers to reliably measure six and more variables, such as size, shape and expression of cell-surface and intracellular proteins, for thousands of cells per second. In leukemia research, one is interested in the identification of leukemic cells, which are characterized by abnormal patterns of surface marker expression. The physiological co-expression patterns of these proteins during blood-cell development hints to a tight regulation of the expression of these markers. Searching for abnormal expression patterns with analysis techniques employing at most two markers at the same time has been successfully established in clinical leukemia research [<xref ref-type="bibr" rid="B3">3</xref>]. However, these techniques, such as gating, are labor-intensive, subjective and not able to capture higher-order dependencies between measured variables.</p><p>Some existing methods make use of the multivariate setting of cytometry readouts [<xref ref-type="bibr" rid="B4">4</xref>,<xref ref-type="bibr" rid="B5">5</xref>,<xref ref-type="bibr" rid="B7">7</xref>], but these methods aim at sample classification based on the readouts rather than on cell population detection within samples.</p><p>Here, we have shown the potential of well-established multivariate-analysis techniques, such as classification via SVM, to automate detection of leukemic cells in flow cytometry readouts from patients' bone marrow and peripheral blood samples. The SVM operates in the space spanned by all variables and even augments it by the use of the kernel trick [<xref ref-type="bibr" rid="B8">8</xref>]. Classification in this complex space takes into account higher-order dependencies between the variables, which are disregarded when restricting oneself to one- or two-dimensional decision rules. With flow cytometry, there is no denying that dependencies between the measured variables do exist, due to properties of utilized materials and biological, superordinate regulatory mechanisms. Developing blood cells are characterized by combinations of interacting surface markers [<xref ref-type="bibr" rid="B16">16</xref>]. Also, measured fluorescence intensities cannot be considered independent from each other because of overlaps between the fluorochromes' emission spectra [<xref ref-type="bibr" rid="B17">17</xref>]. Most cytometers can be set to compensate for these overlaps. While this compensation removes part of the influence of each measured variable on the others, one cannot expect it to remove every dependence between them.</p><p>We built a SVM classifier on the training data, containing approx. 50% of the available data. We, again, split the training data into separate sets for building the classifier, selecting the optimal parameter settings, and assessing the training error. This procedure and the artificial noise added to the training data prevent overfitting of the learned SVM classifier. The learned SVM had a very low training error of 0.4%. Remarkably, this error did not increase, when modifying the SVM's parameter sellings by up to 25%, indicating a very clear separation of leukemic and non-leukemic cells in the enhanced feature space.</p><p>We tested the learned radial-basis SVM classifier on independent test data, generated from a non-leukemic blood sample, an enriched leukemic sample and 18 patient samples not used for training the classifier. On the non-leukemic sample, only 0.01% of the physiological blood cells were misclassified as being leukemic. These few misclassified cells display physical properties and surface marker expression similar to physiological B-lymphocyte precursors (data not shown). While such immature cells are usually restricted to the bone marrow, they have been described to appear in peripheral blood in small quantities [<xref ref-type="bibr" rid="B18">18</xref>]. On the enriched leukemic sample, only 2.88% of the cells were predicted to be non-leukemic cells, a percentage of remaining physiological cells to be expected with the density-gradient centrifugation method for leukemic cell enrichment.</p><p>We applied the classifier on separately measured aliquots of one single sample to evaluate the classifier's precision. We observed a maximal difference of 0.8% between the predicted leukemic-cell proportion in any two of these aliquots, underlining the precision of SVM classification on cytometry readouts.</p><p>The built SVM classifier was applied to identify leukemic cells in independent patient test samples. In these samples, leukemic cells had been pinpointed beforehand by conventional gating [<xref ref-type="bibr" rid="B3">3</xref>]. By SVM-classification, we could recover these leukemic cells with a sensitivity of 99.78% and a specificity of 98.87% (see Table <xref ref-type="table" rid="T2">2</xref>). A comparison of the spatial distribution of cells deemed leukemic by manual gating with that of cells classified as leukemic by the SVM shows that both distributions are highly similar, but SVM-predicted leukemic cells encompass a slightly larger area than gated ones (Figure <xref ref-type="fig" rid="F2">2</xref>).</p><p>Nearly all (99.06%) cells were classified congruently by both methods. Importantly, in the day-8 and day-15 samples, taken after the first and second treatment phase, only a small number of incongruently classified cells were observed. The SVM approach successfully recovered the small leukemic cell populations remaining at this stage, thus demonstrating its promising potential in the identification and monitoring of small leukemic subpopulations during leukemia therapy.</p><p>Most of the cells that were incongruently classified by manual gating and SVM prediction are deemed non-leukemic by gating but leukemic by the SVM (see Table <xref ref-type="table" rid="T2">2</xref>). These cells generally display a light scattering typical of leukemic lymphocytes, and the majority of them show a CD19 and CD34 expression similar to that of leukemic cells detected by manual gating (Figure <xref ref-type="fig" rid="F3">3</xref>). Their main population is also characterized by a low CD20- and intermediate CD10 expression compatible with a leukemic immunophenotype. However, since their CD10 expression tends to be lower than that of the gating-identified leukemic cells, these cells were not included in the leukemic population by conventional gating.</p><p>As such staining variations can arise, e.g., from incomplete staining of cells in the experiment, they decrease the sensitivity of leukemic cell detection by low-dimensional gating. In contrast, the SVM classification is based on all variables at once, and slight variations in only one variable do not hinder the detection of cell populations as long as the remaining variables are characteristic for the sought-after populations. This highlights the strength of the multivariate approach described here.</p><p>Also in Figure <xref ref-type="fig" rid="F3">3</xref>, it can be seen that a small cell subpopulation with low to intermediate CD19 expression was classified as being leukemic by the SVM. Although the CD19 expression of these cells may have been considered below borderline in conventional gating for leukemic B-cells, the artificial noise added to the training data shifted the SVM's decision boundary to include these cells.</p><p>The artificial noise seems to be advantageous for learning classification rules in the flow cytometry setting. Various sources of variability arising in the experimental procedure, such as sample contamination, incomplete staining, and instrument instability, can induce shifts in fluorescence and scattering measurements [<xref ref-type="bibr" rid="B3">3</xref>]. Classification rules that apply to noisy and noise-free cytometry readouts may be insensitive to such shifts.</p><p>Compared to related approaches [<xref ref-type="bibr" rid="B4">4</xref>,<xref ref-type="bibr" rid="B7">7</xref>], the SVM approach has the advantage that it does not require any control sample group. Thus, it obviates the need to take blood samples from healthy persons. Instead, it is based on a given cell classification, gained from established diagnostic procedures. The SVM approach also does not require a discretization of the numerical data, which would reduce the data's information content, but allows for stable event classification in the high-dimensional space spanned by all measured variables. It does not aim at assigning samples to classes, but rather at assigning single cells to predefined groups. Therefore, no summarization of a variable's distribution over all cells is required.</p></sec><sec><title>Conclusion</title><p>The SVM's high classification accuracy is promising, given the fact that the classifier has been build and tested on independent data sets and the training data had been artificially contaminated. Automating the gating for leukemic cells in flow cytometry readouts from blood and bone marrow samples seems highly feasible, even with moderate variations in the experimental procedure.</p><p>Furthermore, the SVM automation is applicable to any gating-like procedure for identifying, even small, subgroups of cells in flow cytometry readouts. One of these applications could be the identification of MRD blast cells and monitoring of response to therapy in ALL.</p><p>Multivariate classification allows for reliable automation of current diagnostic procedures, taking into account biological dependencies that provide obstacles to simplistic methods. It has the potential to greatly reduce the time, costs and arbitrariness associated with these procedures and allows for shifting efforts to potential research extensions.</p><p>In addition, our results show that classification techniques, whose use is already well established on common biological data types, such as gene expression data, can give rise to new algorithms for the analysis of various other existing and upcoming kinds of biological high-throughput data.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>Cells and cell staining</title><p>Leukemic samples (n = 37) from a series of patients with childhood precursor B-cell ALL (PBC-ALL) collected at initial diagnosis (d0-samples from bone marrow, BM, or peripheral blood, PB), after seven days of initial therapy (d8-samples from PB) and after two weeks of initial therapy (d15 samples from BM) were investigated. In addition to the leukemic samples, an artificial dilution series (n = 6 custom-built mix samples) was generated by mixing blood cells, which had been enriched in leukemic cells by Ficoll density gradient centrifugation, with whole peripheral blood cells from a healthy donor. The sample data on all 43 samples investigated, including estimated percentage of leukemic cells, are shown in Table <xref ref-type="table" rid="T1">1</xref>. The cell samples were processed using a BD FACS Lysing Solution (Becton Dickinson, San Jose, CA), according to the manufacturer's instructions, and subsequently stained with a 4-color combination of fluorochrome-conjugated monoclonal antibodies: anti-CD10 phycoerythrin (Dako, Glostrup, Denmark), anti-CD20 fluorescein (Becton Dickinson), anti-CD34 phycoerythrin-cyanin 5.1, and anti-CD19 phycoerythrin-cyanin 7 (Coulter-Immunotech, Hialeah, FL). Measurements of antigen expression were performed by multi-parameter flow cytometry using a FC500 flow cytometer equipped with the Cytomics RXP Analysis Version 1.0 software (Beckman Coulter, Miami, FL). Instrument setup as well as calibration and compensation procedures have been performed according to the recommendations given in [<xref ref-type="bibr" rid="B19">19</xref>] and as described in [<xref ref-type="bibr" rid="B1">1</xref>]. Analog signals were digitized at 1024-channel resolution. For each sample, the two light-scattering variables FSC and SSC plus four surface markers, CD20, CD10, CD34, and CD 19, were quantified for (50 – 300) 10<sup>3 </sup>cells. Readouts from the flow cytometry experiments were provided in a standardized file format, called Flow Cytometry Standard (FCS), version 3.0 [<xref ref-type="bibr" rid="B20">20</xref>]. The readouts were compensated to reduce effects stemming from overlapping emission spectra of utilized fluorochromes.</p></sec><sec><title>Manual gating</title><p>Upon inspection of cells in a two-dimensional dot-plot, cells within a region of interest can be marked, by manually drawing a polygon around them. Many algorithms exist to determine which cells are within the drawn polygon. Once determined, these cells can be highlighted in color, and/or separated from the other cells for further visualization or computations. Currently in leukemia research, cells, which are possibly leukemic, are pinpointed by researchers manually drawing polygons around cells in a number of two-variable dot plots. Shape and position of such defined regions depend on the subjective expertise of the researcher and are not restricted to rectangles at fixed positions. Finally, candidates for leukemic cells are those inside a Boolean combination of drawn regions, called <italic>gate </italic>[<xref ref-type="bibr" rid="B2">2</xref>,<xref ref-type="bibr" rid="B3">3</xref>].</p></sec><sec><title>Implementation</title><p>All computational methods were implemented in the statistical programming language R [<xref ref-type="bibr" rid="B21">21</xref>]. Analyses were conducted using custom functions, which depend on released R packages. We used the SVM implementation of the package <italic>e1071 </italic>[<xref ref-type="bibr" rid="B22">22</xref>]. Our custom functions have been assembled into a new R-package, called <italic>cytomics</italic>, which is available from the authors upon request.</p></sec></sec><sec><title>Authors' contributions</title><p>JT developed the algorithm, performed the computational analysis and wrote the manuscript. PR performed the flow cytometric analysis and contributed to the manuscript. RR provided patient samples and acquired the list mode data. LK designed the project and assisted in writing of the manuscript. RS designed the project, developed the algorithm and assisted in writing of the manuscript. All authors read and approved the final manuscript.</p></sec> |
Subaortic and mid-ventricular obstructive hypertrophic cardiomyopathy with an apical Aneurysm: a case report | <sec><title>Background</title><p>Most patients with hypertrophic cardiomyopathy (HCM) have asymmetric septal hypertrophy and among them, 25% present dynamic subaortic obstruction. Apical HCM is unusual and mid-ventricular HCM is the most infrequent presentation, but both variants may be associated to an apical aneurysm. An even more rare presentation is the coexistece mid-ventricular and apical HCM. This case is a combination of obstructive HCM with mid-ventricular HCM and an apical aneurysm, which to date, has not been reported in the literature.</p></sec><sec><title>Case presentation</title><p>The patient is a 49 year-old lady who presents a combination of septal asymmetric hypertrophic cardiomyopathy (HCM) and midventricular HCM, a subaortic gradient of 65 mm Hg and a midventricular gradient of 20 mm Hg, plus an apical aneurysm. Her clinical presentation was an acute myocardial infarction in June 2005. One month after hospital discharge, the electrocardiogram (ECG) showed a right bundle branch block (RBBB) with no Q waves or ST segment elevation. Coronary angiography revealed normal coronary arteries, left ventricular hypertrophy and an apical aneurysm.</p></sec><sec><title>Conclusion</title><p>This case is a rare example of an asymptomatic patient with subaortic and mid-ventricular hypertrophic cardiomyopathy, who presents a myocardial infarction and normal coronary arteries, and during the course of her disease develops an apical aneurysm.</p></sec> | <contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Cianciulli</surname><given-names>Tomás Francisco</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I3">3</xref><email>tcianciulli@fibertel.com.ar</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Saccheri</surname><given-names>María Cristina</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I3">3</xref><email>mcsaccheri@fibertel.com.ar</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Konopka</surname><given-names>Isabel Victoria</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I3">3</xref><email>isy7773@yahoo.com.ar</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Serans</surname><given-names>Dora Faustina</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>lytton1932@yahoo.com.ar</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Acunzo</surname><given-names>Rafael Salvador</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>rsaacunzo@fibertel.com.ar</email></contrib><contrib id="A6" contrib-type="author"><name><surname>Escudero</surname><given-names>Alejandro Mario García</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>agescudero@yahoo.com</email></contrib><contrib id="A7" contrib-type="author"><name><surname>Masoli</surname><given-names>Osvaldo Horacio</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I3">3</xref><email>ohmasoli@fibertel.com.ar</email></contrib><contrib id="A8" contrib-type="author"><name><surname>Prezioso</surname><given-names>Horacio Alberto</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>hprezioso@intramed.net.ar</email></contrib> | Cardiovascular Ultrasound | <sec><title>Background</title><p>HCM has a prevalence of 0.2 % (1 in 500) in the general population [<xref ref-type="bibr" rid="B1">1</xref>] and presents with marked clinical polymorphism, given by the extent and degree of hypertrophy of the myocardium involved and by the location and magnitude of the intraventricular gradient. Doppler-echocardiography plays an important role in the diagnosis, identifying the distribution of hypertrophy and the site of intraventricular obstruction.</p><p>Most patients with HCM (95%) have asymmetric septal hypertrophy and among them, 25% present with dynamic subaortic obstruction [<xref ref-type="bibr" rid="B2">2</xref>]. Apical HCM is unusual and mid-ventricular HCM is the most infrequent presentation, but both variants may be associated to an apical aneurysm. An even more rare presentation is that of coexistent mid-ventricular and apical HCM. This case represents a combination of asymmetric septal and mid-ventricular HCM with an apical aneurysm, which to date, has not been reported in the literature.</p></sec><sec><title>Case presentation</title><p>The patient is a 49 year-old lady, with a family history of HCM and no coronary risk factors, who in June 2005 was admitted to the coronary care unit due to an episode of oppressive and prolonged chest pain. Her ECG at admission showed sinus rhythm, left ventricular hypertrophy and ST segment elevation in leads V4 to V6, which later normalized, without developing Q waves (Figure <xref ref-type="fig" rid="F1">1A</xref>). A fourth sound and a 4/6 systolic murmur along the mesocardium were noted in the physical exam. Twelve hours after the beginning of chest pain, creatin phosphokinase (CPK) 2.989 U/l and the CPK-MB fraction was 340 U/l.</p><fig position="float" id="F1"><label>Figure 1</label><caption><p>A: The electrocardiogram (ECG) shows sinus rhythm, biatrial enlargement, and left ventricular hypertrophy. B and C: ECG and vectorcardiogram (VCG) one month after acute myocardial infarction. The ECG (B) shows right bundle branch block, right ventricular hypertrophy and/or posterior necrosis. The VCG (C) shows biatrial enlargement with prevalence of the left atrium left ventricular hypertrophic and apical necrosis.</p></caption><graphic xlink:href="1476-7120-4-15-1"/></fig><p>The Doppler-echocardiogram showed a combination of asymmetric septal and mid-ventricular HCM (Figure <xref ref-type="fig" rid="F2">2</xref>), and the anterobasal septum measured 21 mm. There was systolic anterior motion of the mitral valve with a subaortic gradient of 65 mm Hg (Figure <xref ref-type="fig" rid="F3">3</xref>), anteromedial septal thickness was 19 mm, there was a mid-ventricular systolic gradient of 20 mm Hg and an apical aneurysm. Ejection fraction was 55%. The presence of a restrictive mitral flow indicated an increase in left ventricle end-diastolic pressure due to diastolic dysfunction.</p><fig position="float" id="F2"><label>Figure 2</label><caption><p>Transthoracic echocardiogram. A: Parasternal long axis view showing asymmetric septal hypertrophic cardiomyopathy in end diastole. B: A four-chamber view showing a combination of basal and mid-ventricular hypertrophic cardiomyopathy and an apical aneurysm (*). (LA = left atrium, RA = right atrium, RV = right ventricle, LV = left ventricle, Ao = aorta).</p></caption><graphic xlink:href="1476-7120-4-15-2"/></fig><fig position="float" id="F3"><label>Figure 3</label><caption><p>Continuous Doppler image demonstrates that the predominant gradient was subaortic and not simply midventricular.</p></caption><graphic xlink:href="1476-7120-4-15-3"/></fig><p>The ECG at hospital discharge showed sinus rhythm, right bundle branch block (RBBB) with no Q waves, right ventricular hypertrophy and/or posterior necrosis (Figure <xref ref-type="fig" rid="F1">1B</xref>). The vectorcardiogram was consistent with apical necrosis (Figure <xref ref-type="fig" rid="F1">1C</xref>). The chest X-ray showed a mild increase in cardiothoracic index and the lower left arch. The 24-hour Holter showed sinus rhythm, permanent RBBB and 234 monomorphic premature ventricular contractions (PVC's), without complex arrhythmias.</p><p>Myocardial perfusion was assessed with gated-SPECT with Tc99m-Sestamibi at rest and with exercise, and showed a severe fixed apical perfusion defect, compatible with apical necrosis and without residual ischemia (Figure <xref ref-type="fig" rid="F4">4</xref>). The SPECT-cold pressor test was negative for coronary vasospasm.</p><fig position="float" id="F4"><label>Figure 4</label><caption><p>Gated-SPECT horizontal long axis end-diastolic image at rest showing a severe apical perfusion defect.</p></caption><graphic xlink:href="1476-7120-4-15-4"/></fig><p>Hemodynamic study revealed septal and mid-ventricular thickening (Figure <xref ref-type="fig" rid="F5">5</xref>) with normal epicardial coronary arteries, and apical aneurysm. Coronary angiogram ruled out systolic myocardial bridging of the left anterior descending artery and septal perforation branches.</p><fig position="float" id="F5"><label>Figure 5</label><caption><p>Left ventriculogram in right anterior oblique projection (30°), during diastole (A) and systole (B), demonstrating the angiographic appearances of midventricular obliteration with apical aneurysm.</p></caption><graphic xlink:href="1476-7120-4-15-5"/></fig><p>Laboratory tests and serology for Chagas disease were normal.</p><p>During follow-up, the patient was treated with beta-blockers and remained clinically stable.</p></sec><sec><title>Discussion</title><p>Echocardiogram and left ventriculography of our patient showed an apical aneurysm (although the epicardial coronary arteries were angiographically normal), asymmetric septal and mid-ventricular hypertrophy, a moderate subaortic gradient and a mild mid-ventricular gradient.</p><p>Apical infarction is not a rare finding in patients with mid-ventricular HCM. Several reports have shown an association of mid-ventricular and apical HCM with apical aneurysms [<xref ref-type="bibr" rid="B3">3</xref>-<xref ref-type="bibr" rid="B5">5</xref>]. The mechanisms responsible for an apical infarction are not completely understood, but such patients are known to have normal epicardial coronary arteries, usually present with microvascular dysfunction and a decrease in coronary reserve, attributed to narrowing of the small intramyocardial coronary arteries. The combination of an increase in oxygen demand due to increased ventricular thickness and the decrease in oxygen supply due to a decrease in the capillary network predispose to ischemia. Additionally, the apical chamber is subject to greater and sustained systolic stress due to the high intraventricular gradient.</p><p>The apical aneurysm observed in our patient, in the presence of normal epicardial coronary arteries and in the absence of coronary spasm, could be due to microvascular dysfunction and increased stress in the apical chamber, consequent to the combination of subaortic and mid-ventricular obstruction [<xref ref-type="bibr" rid="B6">6</xref>].</p><p>To the best of our knowledge, no cases of combined subaortic and mid-ventricular obstructive HCM, associated with apical aneurysm have been described.</p><p>Treatment of symptomatic HCM is aimed at reducing the intraventricular gradient which, in most cases, is achieved with beta-blockers and calcium antagonists (diltiazem or verapamil).</p><p>Cases of HCM with an apical aneurysm may be complicated by serious ventricular arrhythmias, which usually respond to antiarrhythmic drugs (amiodarone); however, when they are refractory to medical treatment, the insertion of an implantable defibrillator may be warranted [<xref ref-type="bibr" rid="B7">7</xref>], to prevent sudden death. In some patients with an apical aneurysm and very thin walls, the size of the aneurysm may increase rapidly, which entails a high risk of spontaneous rupture and, hence the surgical resection of the aneurysm may be required [<xref ref-type="bibr" rid="B3">3</xref>]. On rare occasions, a thrombus may form inside the aneurysm, which may cause embolization [<xref ref-type="bibr" rid="B8">8</xref>] and thus, in such cases, anticoagulation is indicated.</p></sec><sec><title>Conclusion</title><p>Several reports have shown the association between mid-ventricular and apical HCM and apical aneurysms, but this is the first case presenting a special phenotypic expression of HCM, characterized by a combination of asymmetric septal and mid-ventricular HCM that presents clinically as an infarction with normal coronary arteries and develops an apical aneurysm.</p></sec><sec><title>Competing interests</title><p>The authors declare that they have no competing interests. The manuscript has not been published and is not being considered for publication elsewhere in whole or in part in any language.</p></sec><sec><title>Authors' contributions</title><p>TFC performed the echocardiographic images and participated in the manuscript description. OHM acquired the Gated-SPECT images. AMGE acquired the hemodynamic study. IVK, DFS and RSA attended the patient and prepared the manuscript and figures. HAP and MCS participared in the design and review of the manuscript. All authors read and approved the final manuscript.</p></sec> |
The characteristics of the spectra of superior venae cavae in patients with right heart failure | <sec><title>Background</title><p>Aimed to elucidate the characteristics of the spectra of superior venae cavae (SVC) in respiratory cycles in patients with right heart failure.</p></sec><sec sec-type="methods"><title>Methods</title><p>The spectra of SVC of 30 patients with right heart failure and 30 paired healthy subjects were recorded through right supraclavicular fossa view. The profiles of spectra of superior venae cavae were observed, and peak velocity and velocity time integral (VTI) of every wave of SVC under spontaneous respiration were measured for statistical analysis.</p></sec><sec><title>Results</title><p>In healthy subjects, the peak velocities and VTI of S wave and D wave increased in inspiratory phase and diminished in expiratory phase, and which of S wave were larger than which of D wave in whole respiratory cycle. In patients with right heart failure, spectral variations of SVC could be classified into three patterns: Pattern I: peak velocities and VTI of S wave were larger than that of D wave in early inspiratory phase, but peak velocities and VTI of D wave were larger than those of S wave in late inspiratory phase and early expiratory phase [Pattern I-1], even in whole respiratory cycle [Pattern I-2]; Pattern II: the S wave disappeared and was substituted by inverse wave with low amplitude in whole respiratory cycle. Pattern III: the profiles of the spectra of SVC in patients were similar to those of healthy subjects. In the whole, the respiratory variation ratios of peak velocities and VTI of S wave and D wave were diminished in patients compared with those in healthy subjects.</p></sec><sec><title>Conclusion</title><p>The spectra of superior venae cavae in patients with right heart failure were abnormal, and these characteristics could be used as signs in evaluating right heart failure.</p></sec> | <contrib id="A1" contrib-type="author"><name><surname>Jia</surname><given-names>Hua-Ping</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>jiahuaping@yahoo.com.cn</email></contrib><contrib id="A2" corresp="yes" contrib-type="author"><name><surname>Duan</surname><given-names>Yun-You</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>duanyy@fmmu.edu.cn</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Cao</surname><given-names>Tie-Sheng</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>caots@fmmu.edu.cn</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Yuan</surname><given-names>Li-Jun</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>yuanlijun@fmmu.edu.cn</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Li</surname><given-names>Juan</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>lijuan@fmmu.edu.cn</email></contrib> | Cardiovascular Ultrasound | <sec><title>Background</title><p>The profiles of spectra of superior venae cavae of healthy subjects have been described, and the factors which affected them have been recognized, such as age, sex, respiration, and so on [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B2">2</xref>]. Among all these factors, respiration was the most important one. Some scholars have described the respiratory variations about the spectra of superior venae cavae of healthy subjects, and the respiratory variations of the spectra of superior venae cavae in some pathological conditions have been studying, but few are about the respiratory variations of the spectra of superior venae cavae in patients with right heart failure.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>Clinical data</title><p>30 consecutive patients with right heart failure and their paired healthy subjects were enrolled in this study. Because the functional grading standard of right heart has not been established as yet, NYHA cardiac functional grading standard was adopted instead. The patients were selected according to these standards as follows: 1. NYHA cardiac functional grading standard III-IV; 2. accompanying definite congestion of systemic circulation, which embraced distension of jugular veins and hepatic veins, edema of lower extremity and dilatation of right heart. Patients would be excluded if they had the diseases as follows: 1. severe arrhythmia; 2. moderate-above pericardial effusion; 3. chronic pulmonary diseases; 4. obstructive diseases of superior vena cava. According to these standards above, 18 male and 12 female patients were enrolled, including 12 patients with dilated cardiomyopathy, 8 patients with ischemic cardiomyopathy and 10 patients with valve diseases. The average age was 48 years old. Thirty healthy subjects in control group were paired with patients according to age and sex.</p></sec><sec><title>Examination methods</title><p>Applying sequoia-512 diasonography (transducer 3V2c or 7V3c, 3.5 MHz), the spectra of superior venae cavae of 30 patients and 30 paired healthy subjects were recorded through right supraclavicular fossa. Electrocardiogram and respiratory curve were recorded simultaneously. S wave and D wave were defined as the major waves, and the spectra of superior vena cava with maximum and minimum peak velocities of the major waves in respiratory cycle were selected for analysis. Peak velocities and velocity time integral (VTI) of every wave of superior venae cavae were measured, and their respiratory variation ratios were calculated. The respiratory variation ratios of peak velocities and VTI embraced inspiratory variation ratio and expiratory variation ratio. Inspiratory variation ratio was defined as the peak velocities (or VTI) in inspiratory phase subtracted it in expiratory phase, and then divided by it in expiratory phase, and expiratory variation ratio was defined as the peak velocities (or VTI) in inspiratory phase subtracted it in expiratory phase, and then divided by it in inspiratory phase.</p></sec><sec><title>Statistical analysis</title><p>Data were processed by paired-samples T test with Statistical Package for Social Sciences (SPSS) 11.0, and p < 0.05 was considered statistically significant.</p></sec></sec><sec><title>Results</title><p>The spectra of superior venae cavae of healthy subjects consisted of four waves typically, which were S wave and VR wave in systolic phase, D wave in diastolic phase, and AR wave in atrial systolic phase. The spectra of superior venae cavae of healthy subjects varied regularly in respiratory cycle. Peak velocities and velocity time integrals (VTI) of S wave and D wave increased in inspiratory phase and diminished in expiratory phase, and the VR wave and AR wave varied inversely. Peak velocities and VTI of S wave were larger than those of D wave in whole respiratory cycle in healthy subjects (Figure <xref ref-type="fig" rid="F1">1</xref>).</p><fig position="float" id="F1"><label>Figure 1</label><caption><p>The spectra of SVC of healthy subjects in the respiratory cycle [Pattern III].</p></caption><graphic xlink:href="1476-7120-4-21-1"/></fig><p>The spectra of superior venae cavae in patients with right heart failure could be classified into three patterns: Pattern I: peak velocities and VTI of S wave were larger than those of D wave in early inspiratory phase, but peak velocities and VTI of D wave were larger than those of wave S in late inspiratory phase and early expiratory phase (Figure <xref ref-type="fig" rid="F2">2</xref>), even in whole respiratory cycle (Figure <xref ref-type="fig" rid="F3">3</xref>). Twelve patients belonged to this pattern; Pattern II: the systolic anterior S wave disappeared and was substituted by inverse wave with low amplitude (Figure <xref ref-type="fig" rid="F4">4</xref>), and 8 patients belonged to this pattern. Pattern III: the profiles of the spectra of SVC in patients were similar to those of health subjects, and 10 patients belonged to this pattern. Peak velocities and VTI of every wave and their respiratory variation ratios in two groups were listed in Table <xref ref-type="table" rid="T1">1</xref> to Table <xref ref-type="table" rid="T4">4</xref>, and the results of statistical analysis were as follows: there were no differences in the peak velocities of S wave in respiratory cycle and its expiratory variation ratios between the patient group and the control group (<italic>P </italic>> 0.05), but the inspiratory variation ratios of S wave were larger in the control group than in the patient group (<italic>P </italic>< 0.05); there were no differences in VTI of S wave in respiratory cycle and its inspiratory variation ratios between the patient group and the control group (<italic>P </italic>> 0.05), but the expiratory variation ratios of VTI of S wave were larger in the control group than in the patient group (<italic>P </italic>< 0.05). There were no differences in the peak velocities of D wave in inspiratory phase between the patient group and the control group (<italic>P </italic>> 0.05), but the peak velocities of D wave in expiratory phase were larger in the patient group than in the control group (<italic>P </italic>< 0.05), and the respiratory variation ratios of peak velocities of D wave were larger in the control group than in the patient group (<italic>P </italic>< 0.05); there were no differences in the VTI of D wave in respiratory cycle and its expiratory variation ratios between the patient group and the control group (<italic>P </italic>> 0.05), but the inspiratory variation ratios of VTI of D wave were larger in the control group than in the patient group (<italic>P </italic>< 0.05). The peak velocities and VTI of VR wave were larger in the patient group than in the control group in whole respiratory cycle (<italic>P </italic>< 0.05), but there were no differences in their respiratory variation ratios. The peak velocities and VTI of AR wave were larger in the patient group than in the control group in expiratory phase (<italic>P </italic>< 0.05), but there were no differences in inspiratory cycle; the respiratory variation ratios of peak velocities and the expiratory variation ratios of VTI of AR wave were larger in the patient group than in the control group (<italic>P </italic>< 0.05), but there were no differences in the inspiratory variation ratios of VTI of AR wave between the two groups (<italic>P </italic>> 0.05).</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>Peak velocities and VTI of S wave and their respiratory variation ratios (<graphic xlink:href="1476-7120-4-21-i1.gif"/> ± <italic>s</italic>).</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td align="left" colspan="2">Patients</td><td align="left" colspan="2">Health subjects</td></tr><tr><td></td><td colspan="2"><hr></hr></td><td colspan="2"><hr></hr></td></tr><tr><td></td><td align="left">Peak velocity (cm/s)</td><td align="left">VTI(mm)</td><td align="left">Peak velocity (cm/s)</td><td align="left">VTI(mm)</td></tr></thead><tbody><tr><td align="left">Inspiration</td><td align="left">53.00 ± 20.19</td><td align="left">0.1038 ± 0.0622</td><td align="left">63.30 ± 17.07</td><td align="left">0.1647 ± 0.0513</td></tr><tr><td align="left">Expiration</td><td align="left">39.63 ± 13.48</td><td align="left">0.0724 ± 0.0359</td><td align="left">43.80 ± 13.88</td><td align="left">0.1028 ± 0.0291</td></tr><tr><td align="left">Inspiratory variation ratio (%)</td><td align="left">32.26 ± 16.49</td><td align="left">39.06 ± 34.04</td><td align="left">49.67 ± 36.30</td><td align="left">68.79 ± 39.64</td></tr><tr><td align="left">Expiratory variation ratio (%)</td><td align="left">23.32 ± 9.13</td><td align="left">24.80 ± 14.57</td><td align="left">29.98 ± 14.02</td><td align="left">34.71 ± 17.12</td></tr></tbody></table><table-wrap-foot><p>There were no differences in the peak velocities of S wave in respiratory cycle and its expiratory variation ratios between the patient group and the control group (<italic>P </italic>> 0.05), but the inspiratory variation ratios of S wave were larger in the control group than in the patient group(<italic>P </italic>< 0.05); there were no differences in VTI of S wave in respiratory cycle and its inspiratory variation ratios between the patient group and the control group (<italic>P </italic>> 0.05), but the expiratory variation ratios of VTI of S wave were larger in the control group than in the patient group(<italic>P </italic>< 0.05).</p></table-wrap-foot></table-wrap><fig position="float" id="F2"><label>Figure 2</label><caption><p>The spectra of SVC in the respiratory cycle in patients with right heart failure [Pattern I-1].</p></caption><graphic xlink:href="1476-7120-4-21-2"/></fig><fig position="float" id="F3"><label>Figure 3</label><caption><p>The spectra of SVC in the respiratory cycle in patients with right heart failure [Pattern I-2].</p></caption><graphic xlink:href="1476-7120-4-21-3"/></fig><fig position="float" id="F4"><label>Figure 4</label><caption><p>The spectra of SVC in the respiratory cycle in patients with right heart failure [Pattern II].</p></caption><graphic xlink:href="1476-7120-4-21-4"/></fig></sec><sec><title>Discussion</title><p>The normal pattern of spectrum of superior vena cava and its respiratory variation have been accepted by most scholars, which were characterized of predominant systolic wave, namely peak velocity of S wave were larger than that of D wave in whole respiration cycle, and it was consistent with our study [<xref ref-type="bibr" rid="B3">3</xref>]. Up to now, the variations of spectrum of superior vena cava under most pathological conditions have not been elucidated, and recent studies highlighted the prospect of examination of spectrum of superior vena cava. The spectrum of superior vena cava could be affected by many physiological factors, and respiration was the most apparent one. Recent studies proved the significance of respiratory variation in the examination of superior vena cava. Izumi and his colleagues [<xref ref-type="bibr" rid="B4">4</xref>-<xref ref-type="bibr" rid="B6">6</xref>] made serial studies on the spectra of superior venae cavae in chronic pulmonary diseases, and categorized the spectra of superior venae cavae of the patients with chronic pulmonary diseases into two patterns: one took on the normal mode like that of the health subjects, and the other with disappearance of D wave or both S wave and D wave in the expiratory phase.</p><p>The pattern of venous return in patients with pulmonary disease varied depending on the mode of the ventilation disturbance and the presence of right ventricular pressure overload. The abnormal patterns were observed in some of the patients with obstructive ventilation disturbance and most of the patients with combined ventilation disturbance, and the right atrial pressure exceeded the subclavian vein pressure in the expiratory phase in cases with abnormal patterns. One study of Byrd [<xref ref-type="bibr" rid="B7">7</xref>] indicated respiratory variation in flow velocities of superior vena cava increased in patients with hemodynamically insignificant pericardial effusions and greatest in patients with cardiac tamponade. In obstructive diseases of superior vena cava, respiratory variation of the spectra of superior venae cavae diminished or even vanished, which was a significant sign in diagnosing these diseases [<xref ref-type="bibr" rid="B8">8</xref>]. Because the spectra of superior vena cava were affected in these diseases above, these patients were excluded in our study.</p><p>In this study, abnormal patterns of spectra of superior venae cavae were observed in patients with right heart failure, which had some differences from those of the diseases above. One study on the patients with congestive heart failure found the similar abnormal spectral pattern of superior vena cava, and indicated the pattern with predominant diastolic wave identified patients with a reduced right ventricular ejection fraction and elevated right atrial pressure (>8 mmHg), so the pattern of superior vena cava was a useful tool to estimate the extent of the right circulatory impairment in patients with congestive heart failure [<xref ref-type="bibr" rid="B9">9</xref>].</p><p>Though the results of our study indicated that the spectral patterns of superior vena cava had their own characteristics in some patients with right heart failure, others had similar spectral pattern to that of health people, in this case, the difference of respiratory variations between two groups could be used as additional features. Respiratory variations of every wave and its VTI were complicated. The peak velocities and VTI of S wave and D wave overlapped in two groups, but their respiratory variation ratios mostly diminished significantly in patient group, and it could be the evidence of right heart failure that had not been reported before. Although the respiratory variation ratios of S wave and D wave could not be the individual index, they should be an important additional index to evaluate right heart failure. There were some differences in respiratory variations of VR wave and AR wave between two groups, particularly in AR wave, and their respiratory variations were contrary to those of S wave and D wave. These two waves were of low amplitude, and their respiratory variations were not apparent as those of S wave and D wave, their clinical values needed further study.</p><p>At present, there was no functional grading standard of right heart available, and NYHA cardiac functional grading standard was adopted instead in this study. In order to confirm right heart failure, some additional conditions were added and they might bring up some deflections. On the other hand, right heart failure often accompanied with pulmonary artery hypertension, tricuspid insufficiency, and elevated right heart pressure, and these factors were often not parallel in one patient and could affect the spectrum of superior vena cava respectively, so there were complicated factors which could affect the results. Even so, this study indicated that the spectra of superior venae cavae in patients with right heart failure had their own characteristics, and these characteristics could be used to evaluate right heart failure. Further studies needed to be done in detail.</p></sec><sec><title>Conclusion</title><p>In patients with right heart failure, the spectra of superior vena cava were abnormal, and it could be a new view to evaluate right heart function through observing the spectrum of superior vena cava.</p></sec><sec><title>Competing interests</title><p>The author(s) declare that they have no competing interests.</p></sec><sec><title>Authors' contributions</title><p>JIA carried out the clinical investigation, performed the statistical analysis, and drafted the manuscript. DUAN designed the experiment and proofread the paper. CAO participated in the design of the study. YUAN and LI participated in the clinical investigation, and helped to draft the manuscript. All authors read and approved the final manuscript.</p><table-wrap position="float" id="T2"><label>Table 2</label><caption><p>Peak velocities and VTI of D wave and their respiratory variation ratios (<graphic xlink:href="1476-7120-4-21-i1.gif"/> ± <italic>s</italic>).</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td align="left" colspan="2">Patients</td><td align="left" colspan="2">Health subjects</td></tr><tr><td></td><td colspan="2"><hr></hr></td><td colspan="2"><hr></hr></td></tr><tr><td></td><td align="left">Peak velocity (cm/s)</td><td align="left">VTI(mm)</td><td align="left">Peak velocity (cm/s)</td><td align="left">VTI(mm)</td></tr></thead><tbody><tr><td align="left">Inspiration</td><td align="left">39.67 ± 26.52</td><td align="left">0.0684 ± 0.0509</td><td align="left">38.00 ± 9.15</td><td align="left">0.0747 ± 0.0371</td></tr><tr><td align="left">Expiration</td><td align="left">34.43 ± 19.21</td><td align="left">0.0580 ± 0.0354</td><td align="left">24.73 ± 9.12</td><td align="left">0.0454 ± 0.0243</td></tr><tr><td align="left">Inspiratory variation ratio (%)</td><td align="left">27.03 ± 16.34</td><td align="left">33.72 ± 30.33</td><td align="left">65.34 ± 44.38</td><td align="left">86.18 ± 66.74</td></tr><tr><td align="left">Expiratory variation ratio (%)</td><td align="left">20.39 ± 13.40</td><td align="left">29.98 ± 24.82</td><td align="left">35.00 ± 17.96</td><td align="left">45.16 ± 35.49</td></tr></tbody></table><table-wrap-foot><p>There were no differences in the peak velocities of D wave in inspiratory phase between the patient group and the control group (<italic>P </italic>> 0.05), but the peak velocities of D wave in expiratory phase were larger in the patient group than in the control group (<italic>P </italic>< 0.05), and the respiratory variation ratios of peak velocities of D wave were larger in the control group than in the patient group (<italic>P </italic>< 0.05); there were no differences in the VTI of D wave in respiratory cycle and its expiratory variation ratios between the patient group and the control group (<italic>P </italic>> 0.05), but the inspiratory variation ratios of VTI of D wave were larger in the control group than in the patient group (<italic>P </italic>< 0.05).</p></table-wrap-foot></table-wrap><table-wrap position="float" id="T3"><label>Table 3</label><caption><p>Peak velocities and VTI of VR wave and their respiratory variation ratios (<graphic xlink:href="1476-7120-4-21-i1.gif"/> ± <italic>s</italic>).</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td align="left" colspan="2">Patients</td><td align="left" colspan="2">Health subjects</td></tr><tr><td></td><td colspan="2"><hr></hr></td><td colspan="2"><hr></hr></td></tr><tr><td></td><td align="left">Peak velocity (cm/s)</td><td align="left">VTI(mm)</td><td align="center">Peak velocity (cm/s)</td><td align="left">VTI(mm)</td></tr></thead><tbody><tr><td align="left">Inspiration</td><td align="left">17.15 ± 6.27</td><td align="left">0.0159 ± 0.0110</td><td align="center">12.77 ± 3.34</td><td align="left">0.0080 ± 0.0043</td></tr><tr><td align="left">Expiration</td><td align="left">19.74 ± 6.71</td><td align="left">0.0192 ± 0.0157</td><td align="center">15.13 ± 3.90</td><td align="left">0.0098 ± 0.0049</td></tr><tr><td align="left">Inspiratory variation ratio (%)</td><td align="left">16.47 ± 5.51</td><td align="left">21.55 ± 12.79</td><td align="center">15.75 ± 8.90</td><td align="left">17.69 ± 17.08</td></tr><tr><td align="left">Expiratory variation ratio (%)</td><td align="left">20.21 ± 7.94</td><td align="left">27.68 ± 18.19</td><td align="center">20.10 ± 14.45</td><td align="left">28.25 ± 34.71</td></tr></tbody></table><table-wrap-foot><p>The peak velocities and VTI of VR wave were larger in the patient group than in the control group in whole respiratory cycle (<italic>P </italic>< 0.05), but there were no differences in their respiratory variation ratios.</p></table-wrap-foot></table-wrap><table-wrap position="float" id="T4"><label>Table 4</label><caption><p>Peak velocities and VTI of AR wave and their respiratory variation ratios (<graphic xlink:href="1476-7120-4-21-i1.gif"/> ± <italic>s</italic>).</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td align="left" colspan="2">Patients</td><td align="left" colspan="2">Health subjects</td></tr><tr><td></td><td colspan="2"><hr></hr></td><td colspan="2"><hr></hr></td></tr><tr><td></td><td align="left">Peak velocity (cm/s)</td><td align="left">VTI(mm)</td><td align="left">Peak velocity (cm/s)</td><td align="left">VTI(mm)</td></tr></thead><tbody><tr><td align="left">Inspiration</td><td align="left">18.33 ± 5.29</td><td align="left">0.0173 ± 0.0103</td><td align="left">17.40 ± 5.82</td><td align="left">0.0134 ± 0.0069</td></tr><tr><td align="left">Expiration</td><td align="left">23.08 ± 7.78</td><td align="left">0.0228 ± 0.0119</td><td align="left">19.23 ± 5.98</td><td align="left">0.0154 ± 0.0075</td></tr><tr><td align="left">Inspiratory variation ratio (%)</td><td align="left">21.71 ± 16.02</td><td align="left">29.79 ± 22.95</td><td align="left">9.83 ± 8.81</td><td align="left">27.83 ± 40.52</td></tr><tr><td align="left">Expiratory variation ratio (%)</td><td align="left">29.65 ± 24.91</td><td align="left">47.21 ± 46.41</td><td align="left">12.11 ± 12.92</td><td align="left">34.71 ± 17.12</td></tr></tbody></table><table-wrap-foot><p>The peak values and VTI of AR wave were larger in the patient group than in the control group in expiratory phase (<italic>P </italic>< 0.05), but there were no differences in inspiratory cycle; the respiratory variation ratios of peak values and the expiratory variation ratios of VTI of AR wave were larger in the patient group than in the control group (<italic>P </italic>< 0.05), but there were no differences in the inspiratory variation ratios of VTI of AR wave between the two groups (<italic>P </italic>> 0.05).</p></table-wrap-foot></table-wrap></sec> |
Hyeropic shift after LASIK induced Diffuse lamellar keratitis | <sec><title>Background</title><p>Diffuse lamellar keratitis (DLK) is a relatively new syndrome that is increasingly being reported after LASIK. We have observed that a hyperopic shift may be associated with the occurrence of this diffuse lamellar keratitis.</p></sec><sec><title>Case presentation</title><p>A 26 year old man developed bilateral diffuse lamellar keratitis (DLK) following myopic LASIK. The residual refractive error was +0.5D OD and +0.25D OS at the end of the first week. The sterile infiltrates resolved over a period of 4–6 weeks on topical steroid therapy. A progressive hyperopic shift was noted in the right eye with an error +4.25Dsph/+0.25Dcyl 20 at the final follow up 6 months post surgery.</p></sec><sec><title>Conclusion</title><p>Diffuse lamellar keratitis after LASIK may be associated with a significant hyperopic shift.</p></sec> | <contrib id="A1" contrib-type="author"><name><surname>Dada</surname><given-names>Tanuj</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>tanujdada@hotmail.com</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Pangtey</surname><given-names>Mayank S</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>mpangtey@yahoo.com</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Sharma</surname><given-names>Namrata</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>namrata123@hotmail.com</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Vajpayee</surname><given-names>Rasik B</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>rasikvajvayee@rediffmail.com</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Jhanji</surname><given-names>Vishal</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>vishaljhanji@gmail.com</email></contrib><contrib id="A6" corresp="yes" contrib-type="author"><name><surname>Sethi</surname><given-names>Harinder Singh</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>sethi2211@yahoo.com</email></contrib> | BMC Ophthalmology | <sec><title>Background</title><p>Laser in situ keratomileusis (LASIK) has become the refractive procedure of choice for myopia [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B2">2</xref>] and hyperopia [<xref ref-type="bibr" rid="B3">3</xref>] due to its inherent advantages of a quick recovery time, preservation of bowman's layer, and a clearer more organized anterior stroma [<xref ref-type="bibr" rid="B4">4</xref>]. Diffuse lamellar keratitis (DLK) is a relatively new syndrome that is increasingly being reported after LASIK [<xref ref-type="bibr" rid="B5">5</xref>]. It is characterized by the appearance of diffuse, multi focal, polymorphonuclear infiltration in the flap interface, between 1–5 days after the surgery. The exact etiology of this noninfectious syndrome is still obscure. We have observed that a hyperopic shift may be associated with the occurrence of this diffuse lamellar keratitis.</p></sec><sec><title>Case presentation</title><p>A 26 year old patient underwent LASIK for myopia of - 10.0 Dsph OD and -10.25 Dsph/-1.50 Dcyl @160 OS. The informed consent was obtained from the patient. He had not undergone previous keratorefractive surgery and there was no evidence of any other ocular or systemic disease. The best-corrected visual acuity was 20/20 in both eyes. The preoperative keratometry was 43.1 @ 179 degree and 44.1@ 89 degree (Orbscan) OD and 43.7@ 160 degree & 44.5@ 70 degree OS. The preoperative central corneal thickness (optical pachymetry, Orbscan) was 511 μOD and 516 μOS respectively.</p><p>The surgery was performed using the Chiron Technolas 217 eximer laser machine and the hansatome was used to create a 9.5 mm diameter flap, 160 μ in thickness. The optic zone diameter was kept at 5 mm for both eyes. In the right eye 130 μ of the stroma was ablated and in the left eye 148 μ was ablated.</p><p>The patient was noted to have a diffuse infiltration confined to the flap interface, with multiple foci of increased density in the center of the flap on the first postoperative day and was started on one hourly, 1% Prednisolone acetate eye drops. The uncorrected visual acuity was 20/200 OD and 20/400 OS. The patient was advised to undergo irrigation under the flap but refused any further surgical intervention. The infiltrates resolved gradually on topical steroid therapy over a period of 4 weeks. The best corrected visual acuity improved to 20/30 in both eyes but the end result was a significant hyperopic error. The visual acuity, refraction, corneal pachymetry and corneal topography were recorded preoperatively and during the follow up visits at 1 week, 1 month, 3 months and 6 months (Table <xref ref-type="table" rid="T1">1</xref>).</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>Pre & Postoperative visual acuity, Refraction, Pachymetry & Corneal Topography</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td></td><td></td><td align="left"><bold>BCVA</bold></td><td align="left"><bold>Refraction</bold></td><td align="left"><bold>Pachymetry(μ)</bold></td><td align="left"><bold>Corneal Topography</bold></td></tr></thead><tbody><tr><td align="left"><bold>Pre Operative</bold></td><td align="left"><bold>Right Eye</bold></td><td></td><td align="left">20/20</td><td align="left">-10.0Dsph</td><td align="left">511</td><td align="left">44.1@89°/43.1@179°</td></tr><tr><td></td><td align="left"><bold>Left Eye</bold></td><td></td><td align="left">20/20</td><td align="left">-10.25Dsph/-2.5Dcyl@165°</td><td align="left">506</td><td align="left">44.6@70°/43.7@160°</td></tr><tr><td align="left"><bold>Post Operative</bold></td><td align="left"><bold>1</bold><sup>st </sup><bold>Week</bold></td><td align="left">Right</td><td align="left">20/80</td><td align="left">+0.50Dsph</td><td align="left">289</td><td align="left">41@79°/39.8@169°</td></tr><tr><td></td><td></td><td align="left">Left</td><td align="left">20/60</td><td align="left">+0.25Dsph</td><td align="left">325</td><td align="left">41.7@67°/39.6@157°</td></tr><tr><td></td><td align="left"><bold>1</bold><sup>st </sup><bold>Month</bold></td><td align="left">Right</td><td align="left">20/30</td><td align="left">+0.75Dsph</td><td align="left">314</td><td align="left">40.9@73°/39.9@163°</td></tr><tr><td></td><td></td><td align="left">Left</td><td align="left">20/30</td><td align="left">+0.25Dsph</td><td align="left">333</td><td align="left">41.4@70°/39.7@160°</td></tr><tr><td></td><td align="left"><bold>3</bold><sup>rd </sup><bold>Month</bold></td><td align="left">Right</td><td align="left">20/30</td><td align="left">+4.0Dsph/+1.75Dcyl@20°</td><td align="left">323</td><td align="left">41.1@84°/39.3@174°</td></tr><tr><td></td><td></td><td align="left">Left</td><td align="left">20/30</td><td align="left">+0.75Dcyl@180°</td><td align="left">347</td><td align="left">41.2@76°/39.7@166°</td></tr><tr><td></td><td align="left"><bold>6</bold><sup>th </sup><bold>Month</bold></td><td align="left">Right</td><td align="left">20/30</td><td align="left">+4.25Dsph/+0.75@20°</td><td align="left">361</td><td align="left">41.1@90°/39.7@0°</td></tr><tr><td></td><td></td><td align="left">Left</td><td align="left">20/30</td><td align="left">+0.24Dcyl@155°</td><td align="left">343</td><td align="left">41.2@75°/39.7@165°</td></tr></tbody></table></table-wrap></sec><sec><title>Discussion</title><p>Diffuse Lamellar Keratitis (DLK) or the Sands of Sahara syndrome is a well recognized complication of lamellar surgery. It is a self perpetuating sterile inflammation of the cornea following any intervention where a lamellar incision has created an interface through stromal tissue. The incidence has been reported to be 0.2% but is believed to be much higher<sup>6</sup>. Various factors that have been implicated include toxic insult due to organic esters and lubricants from lubricants and machine oil, particulate matter on the microkeratome blade, release of endotoxin from bacterial growth on the reusable instruments after autoclaving, traumatic insult from microkeratome, hemorrhage from a micropannus, heating effect of laser on the cornea, meibomian gland secretion, povidone iodine, debris from gloves, absorbent sponges and particles from an adhesive catheter dressing used as a drape [<xref ref-type="bibr" rid="B7">7</xref>-<xref ref-type="bibr" rid="B12">12</xref>]. The present case represents a stage 3 DLK, which may be seen on day 1, and is characterized by a severe inflammation with a decrease in visual acuity and must be managed by manually lifting the flap and irrigating under it. [<xref ref-type="bibr" rid="B15">15</xref>,<xref ref-type="bibr" rid="B16">16</xref>]</p><p>The unique feature in this case was the excessive hyperopic shift after the episode of lamellar keratitis. Previously Smith et al [<xref ref-type="bibr" rid="B5">5</xref>] reported 11 cases of diffuse lamellar keratitis of which 3 cases developed hyperopic shift. Of these three cases a 44 years old man who underwent LASIK with astigmatic keratotomy developed the highest shift of +0.75 Dsph/+1.25 Dcyl @ 86 degree. Peters et al [<xref ref-type="bibr" rid="B13">13</xref>] has reported 18 eyes with DLK which had a mean postoperative refractive error of -0.07 ± 0.48D. in another group of same study, 7 eyes developed DLK with a mean refractive error of +0.04 ± 0.32D.</p><p>Possible reasons for this shift could be overcorrection, scarring of the stroma with irregular astigmatism and flattening, stromal tissue loss due to destructive neutrophilic enzymes [<xref ref-type="bibr" rid="B15">15</xref>,<xref ref-type="bibr" rid="B16">16</xref>]and deposition of material with an abnormal refractive index. In the first week the refractive error was +0.5D sph in the right eye and +0.25 Dsph in left eye (Table <xref ref-type="table" rid="T1">1</xref>). So the excessive hyperopic shift due to an overcorrection is ruled out. Corneal edema could be the cause of a transient hyperopia but the likely effect would resolve over a fortnight. Tissue loss cannot be the implicating factor because there was an increase in the central corneal pachymetry on subsequent follow up (Table <xref ref-type="table" rid="T1">1</xref>). If the inflammation is not controlled at an early stage, permanent stromal scarring and irregular astigmatism could lead to a hyperopic error [<xref ref-type="bibr" rid="B15">15</xref>,<xref ref-type="bibr" rid="B16">16</xref>].</p><p>Another hypothesis for the hyperopic shift is the hyperactive neutrophilic response, which along with the injured keratocytes degranulates, releasing collagenase as well as protease enzymes [<xref ref-type="bibr" rid="B14">14</xref>], and lead to excessive enzymatic thinning of stroma. This may act as a stimulus for compensatory stromal synthesis by the keratocytes and deposition of tissue with an abnormal refractive index, a factor which can only be proven by histochemical/electron microscopic analysis of the tissue.</p><p>Early intervention after the onset of DLK with copious irrigation of the stromal bed after lifting the flap may limit the inflammatory debris and help to prevent the occurrence of this complication although all steps should be taken for a primary prevention of this syndrome. [<xref ref-type="bibr" rid="B15">15</xref>,<xref ref-type="bibr" rid="B16">16</xref>]</p></sec><sec><title>Conclusion</title><p>Diffuse lamellar keratitis after Laser in situ keratomileusis may be associated with a hyperopic shift.</p></sec><sec><title>Competing interests</title><p>The author(s) declare that they have no competing interests.</p></sec><sec><title>Authors' contributions</title><p>TD: Performed the surgery and prepared a manuscript</p><p>MP: Performed the literature search and helped in the documentation of the case</p><p>RBV: Helped in making the clinical diagnosis and management of the case</p><p>NS: Helped in making the clinical diagnosis and management of the case</p><p>VJ: Helped in maintaining the follow up and data of the case</p><p>HSS: Helped in Preparation and submission of the manuscript</p><p>All the authors read and approved the final script.</p></sec><sec><title>Pre-publication history</title><p>The pre-publication history for this paper can be accessed here:</p><p><ext-link ext-link-type="uri" xlink:href="http://www.biomedcentral.com/1471-2415/6/19/prepub"/></p></sec> |
Comparison of silicon oil removal with various viscosities after complex retinal detachment surgery | <sec><title>Background</title><p>Despite the progress in vitreoretinal surgery and the importance of silicone oil as an adjunct for the treatment of complex forms of retinal detachment, controversy still surrounds the issue of selecting the proper oil viscosity for clinical use. Herein, we evaluate the outcomes of retinal detachment (RD) surgery after removing silicone oils of different viscosities.</p></sec><sec sec-type="methods"><title>Methods</title><p>In this retropsective cohort study, eighty-two eyes with surgically re-attached retinas, of which 53 were filled with 5000cs silicone oil and 29 with 1000cs silicone oil were enrolled. We evaluated the outcomes and complications following silicone oil removal. Final anatomic success (stable re-attachment), final visual acuity (VA) and intraocular pressure (IOP)were recorded and analysed.</p></sec><sec><title>Results</title><p>Of 82 eyes, 41 had proliferative vitreoretinopathy (PVR), 24 were associated with intraocular foreign bodies, 10 had endophthalmitis and 7 had proliferative diabetic retinopathy with tractional retinal detachment. Prior to silicone oil removal, the retina was attached in all eyes, 29% had VA ≥ 6/120 and 52% had IOP ≥ 21 mmHg. After silicone oil removal, the retina remained attached in 59(72%) of the eyes, 34% had VA ≥ 6/120 and 9% had IOP ≥ 21 mmHg. Comparing 1000cs and 5000cs silicone oil filled eyes, redetachment occurred more frequently in the latter group especially in cases with associated PVR. Final VA worse than 6/120 was associated with initial VA < 6/120 (OR = 32.2 95%CI 7.4–140.2) and use of 5000cs silicone oil (OR = 7.9 95%CI 1.9–32.2). No factor was significantly associated with final IOP ≥ 21 mmHg.</p></sec><sec><title>Conclusion</title><p>In complicated retinal detachment surgery, use of 5000cs silicone oil may be associated with a poorer anatomic and visual outcome compared with 1000cs silicone oil. However there was no difference between the two viscosities in IOP elevation. A randomized controlled study is necessary to further evaluate such a possibility.</p></sec> | <contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Soheilian</surname><given-names>Masoud</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>masoud_soheilian@yahoo.com</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Mazareei</surname><given-names>Mohammad</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>masoud_soheilian@yahoo.com</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Mohammadpour</surname><given-names>Mehrdad</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>mahammadpour@yahoo.com</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Rahmani</surname><given-names>Bahram</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>masoud_soheilian@yahoo.com</email></contrib> | BMC Ophthalmology | <sec><title>Background</title><p>Despite the progress in vitreoretinal surgery and the importance of silicone oil as an adjunct for the treatment of complex forms of retinal detachment, controversy still surrounds the issue of selecting the proper oil viscosity for clinical use[<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B2">2</xref>]. There is no difference in the tamponading force effected by silicone oils of different viscosities. The tamponading force relates to interfacial surface tension between the oil and the surrounding tissue [<xref ref-type="bibr" rid="B3">3</xref>]. Furthermore, new experiments have disclosed that the optical transmittance, penetration depth and the absorption spectra of mid-infrared cutting lasers through silicone oils of different viscosities are almost identical [<xref ref-type="bibr" rid="B4">4</xref>]. However, different studies have shown that lower viscosity silicone oil is prone to earlier emulsification. Therefore, it is a common assumption that side effects such as cataract, glaucoma and keratopathy may increase with the use of lower viscosity oils [<xref ref-type="bibr" rid="B2">2</xref>,<xref ref-type="bibr" rid="B5">5</xref>,<xref ref-type="bibr" rid="B6">6</xref>].</p><p>The tamponading force of silicone oil continues until emulsification occurs, therefore, to reduce the side effects it is important to use the most suitable viscosity of silicone oil and to remove it once its function for achieving stable reattachment of the retina has been accomplished. In this study, we analyzed and compared the visual outcome, redetachment rate, and complications of two currently available silicone oils with different viscosities (1000cs & 5000cs) after removal in eyes with a stable appearing retina following vitreoretinal procedures for detachment.</p></sec><sec sec-type="methods"><title>Methods</title><p>This study includes 82 eyes of 82 consecutive patients who underwent silicone oil removal from 1986–1997 at a tertiary referral eye center in Tehran. The study was performed in compliance with the Helsinki Declaration after the approval by institutional review board of Labbafinejad Medical Center . Selection of silicone oil viscosity was mostly based on the availability of the oil in the operating room rather than surgeons' choice: throughout the study period, (prior to FDA approval), Labbafinejad Medical Center carried only one silicone viscosity (1000cs or 5000cs) at a given time. Hence, there was no selection bias in terms of the viscosity selected for each case. Before silicone oil removal, all patients had developed different degrees of emulsification, from an early appearance of fish eggs in the superior retina visualized only by indirect ophthalmoscopy, to overt droplets of oil which could be seen in the anterior chamber by slit lamp. All eyes had completely stable, reattached retina prior to silicone oil removal. Oil removal was as complete as possible. All eyes were aphakic and the best-corrected visual acuity was based on performance with a new postoperative correction. Indications for silicone oil removal were any of the following alone or in combination: oil emulsification, high intraocular pressure and keratopathy.</p><p>Twenty-nine eyes had been filled with 1000 centistoke silicone oil and 53 with 5000cs silicone oil. The volume of injected oil ranged from 3.5cc to a maximum of 4.5cc. A single surgeon performed the operations on every patient, including all vitreoretinal procedures and subsequent silicone oil removals.</p><p>Patients had a complete ophthalmologic examination including refraction and best corrected visual acuity (if possible), slit lamp examination, applanation tonometry, gonioscopy and indirect ophthalmoscopy both prior to and after silicone oil removal. Patient charts were also reviewed for the pre-silicone oil status of the retina including the extent of retinal detachment, macular involvement, grading of PVR, cause of endophthalmitis, location, number and type of retinal breaks in traumatic eye injuries and presence and location of neovascularization in PDR cases. All cases were carefully followed for development of possible complications and maitenance of retinal stability.</p><sec><title>Surgical technique</title><p>Since all patients were aphakic, after a localized peritomy, a small superior limbal incision was made and kept open with an iris spatula. An infusion cannula connected to an infusion bottle was placed through the inferior temporal pars plana into the midvitreous cavity. Silicone is light and floats on water, thereby, by allowing the infusion fluid to flow inside the eye, silicone oil was expelled through the superior limbal incision. No other membrane was removed. However, small intraocular silicone bubbles remained postoperatively in all cases. The sclerotomy and the limbal incision were then closed and the conjunctiva was re-approximated. At the conclusion, a subconjunctival injection of antibiotic and steroid was performed.</p></sec><sec><title>Statistical methods</title><p>Statistical analysis of patients' data was performed by SAS statistical software. The data was analyzed using chi-square, Fisher's exact tests, Student t-test and multiple logistic regression. Main outcome measures included final visual acuity, redetachment rate and final intraocular pressure following silicone oil removal.</p></sec></sec><sec><title>Results</title><p>Eighty-two eyes of 82 patients were studied. Forty-four of the operated eyes (53.6%) were right and 38 (46.4%) were left. Fifty-two of the patients were male (63.4%) and 30 were female (36.6%). All patients were caucasians. Based on surgical indications for silicone oil injection, patients were categorized into 4 groups: complex retinal detachment (RD) associated with PVR (Forty-one eyes), RD associated with intraocular foreign body (IOFB) (24 eyes), RD associated with endophthalmitis (10 eyes) and RD due to complications of proliferative diabetic retinopathy (7 eyes). Pre-silicone oil status of the studied eyes was as follows.</p><p>In eyes treated for retinal detachment associated with PVR (41 eyes), the macula was detached in all cases and the extent of retinal detachment ranged from three to four quadrants. PVR grade C (anterior or posterior) was observed in all these cases. Anterior PVR was present in all eyes and involved 12 clock hours in 15 eyes. Posterior PVR was present in 30 eyes involving 6 to 12 clock hours.</p><p>In eyes with retinal detachment associated with IOFB (24 eyes), PVR was present in eight eyes preoperatively. Six eyes had multiple large peripheral breaks and three eyes developed giant retinal dialyses during foreign body removal. Seven eyes developed posterior retinal breaks while deeply embedded intraretinal foreign bodies were being dislodged.</p><p>In the endophthalmitis group (10 eyes) there were 6 cases of post-traumatic bacterial endophthalmitis; 2 acute postoperative bacterial infections and 2 cases of bleb- associated endophthalmitis. Coagulase-negative Staphylococci accounted for two cases, Streptococci, Haemophilus Influenzae and Bacillus Cereus were isolated each in one case, the rest (5 eyes) were culture-negative.</p><p>In the group with complications of proliferative diabetic retinopathy (seven eyes), all had tractional retinal detachment involving the macula and all had been treated with argon laser panretinal photocoagulation. In addition to presence of tractional RD, all eyes displayed different degrees of non-regressed neovascularization of the disc and elsewhere (NVD and NVE).</p><p>Each surgical indication was found in approximately equal proportions in the 1000 and 5000 centistoke oil categories [Table <xref ref-type="table" rid="T1">1</xref>]. All patients who underwent silicone oil removal had an stable attached retina before the operation for at least 50 days. Indications for silicone oil removal were silicone emulsification alone in 8 eyes (9.7%), emulsification associated with high IOP in 43 eyes (52.4%) and emulsification associated with varying degrees of keratopathy in 31 eyes (37.8%). The mean time to oil removal was 10 ± 6.5 months for eyes filled with 1000cs silicone oil and 23.9 ± 15.1 months for eyes filled with 5000cs silicone oil [table <xref ref-type="table" rid="T2">2</xref>]. After silicone oil removal patients were followed from 4 to 120 months. The follow-up period was almost equal for silicone 1000cs and 5000cs filled groups.</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>Relation between surgical indication, viscosity of silicone oil and final retinal status</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left">Diagnosis</td><td align="center" colspan="4">Type of silicone oil</td><td align="center" colspan="8">Final retinal status</td></tr><tr><td/><td colspan="4"><hr></hr></td><td colspan="8"><hr></hr></td></tr><tr><td></td><td align="left" colspan="2">1000 cs</td><td align="left" colspan="2">5000 cs</td><td align="left" colspan="4">Attached 1000cs 5000cs</td><td align="left" colspan="4">Detached 1000cs 5000cs</td></tr><tr><td></td><td colspan="12"><hr></hr></td></tr><tr><td></td><td align="left">No</td><td align="left">%</td><td align="left">No</td><td align="left">%</td><td align="left">No</td><td align="left">%</td><td align="left">No</td><td align="left">%</td><td align="left">No</td><td align="left">%</td><td align="left">No</td><td align="left">%</td></tr></thead><tbody><tr><td align="left">RD* + PDR†</td><td align="left">5</td><td align="left">17</td><td align="left">2</td><td align="left">4</td><td align="left">4</td><td align="left">15.3</td><td align="left">2</td><td align="left">6</td><td align="left">1</td><td align="left">33.3</td><td align="left">0</td><td align="left">0</td></tr><tr><td align="left">RD + endophthalmitis</td><td align="left">4</td><td align="left">14</td><td align="left">6</td><td align="left">11</td><td align="left">4</td><td align="left">15.3</td><td align="left">5</td><td align="left">15</td><td align="left">0</td><td align="left">0</td><td align="left">1</td><td align="left">5</td></tr><tr><td align="left">RD + IOFB‡</td><td align="left">8</td><td align="left">28</td><td align="left">16</td><td align="left">30</td><td align="left">7</td><td align="left">27</td><td align="left">10</td><td align="left">30.3</td><td align="left">1</td><td align="left">33.3</td><td align="left">6</td><td align="left">30</td></tr><tr><td align="left">RD + PVR**</td><td align="left">12</td><td align="left">42</td><td align="left">29</td><td align="left">55</td><td align="left">11</td><td align="left">42.3</td><td align="left">16</td><td align="left">48.5</td><td align="left">1</td><td align="left">33.3</td><td align="left">13</td><td align="left">65</td></tr><tr><td align="left">Total</td><td align="left">29</td><td align="left">100</td><td align="left">53</td><td align="left">100</td><td align="left">26</td><td align="left">100</td><td align="left">33</td><td align="left">100</td><td align="left">3</td><td align="left">100</td><td align="left">20</td><td align="left">100</td></tr></tbody></table><table-wrap-foot><p>*RD = Retinal detachment – †PDR = Proliferative diabetic retinopathy ‡IOFB = Intraocular foreign body – **PVR = Proliferative vitreoretinopathy</p></table-wrap-foot></table-wrap><table-wrap position="float" id="T2"><label>Table 2</label><caption><p>Duration of silicone oil retention before its removal</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td align="left" colspan="2">Silicone 1000cs</td><td align="left" colspan="2">Silicone 5000cs</td></tr><tr><td></td><td colspan="4"><hr></hr></td></tr><tr><td></td><td align="left">No</td><td align="left">%</td><td align="left">No</td><td align="left">%</td></tr></thead><tbody><tr><td align="left">< 2 months</td><td align="left">15</td><td align="left">51.7</td><td align="left">9</td><td align="left">17</td></tr><tr><td align="left">2–6 months</td><td align="left">13</td><td align="left">44.8</td><td align="left">24</td><td align="left">45.3</td></tr><tr><td align="left">6–12 months</td><td align="left">1</td><td align="left">3.4</td><td align="left">18</td><td align="left">34</td></tr><tr><td align="left">> 12 months</td><td align="left">0</td><td align="left">0</td><td align="left">2</td><td align="left">3.7</td></tr><tr><td align="left">Total</td><td align="left">29</td><td align="left">100</td><td align="left">53</td><td align="left">100</td></tr></tbody></table></table-wrap><sec><title>Anatomic status of the retina following silicone oil removal</title><p>Before silicone oil removal all eyes had a reattached and stable retina. Postoperatively and at the last follow-up, out of 82 eyes, the retina remained attached in 59 eyes (72%); recurrent RD occurred in 23 eyes (28%) [Table <xref ref-type="table" rid="T1">1</xref>]. Re-detachment occurred in 3 (10.3%) out of 29 eyes filled with 1000cs silicone oil, however 20 (37.7%) out of 53 eyes filled with 5000cs silicone oil developed re-detachment, this difference is statistically significant (P = 0.008). There was no statistically significant difference among various surgical indications in terms of re-detachment, however the overall re-detachment rate for eyes with RD and PVR was higher (34%). In this category, the re-detachment rate in eyes filled with 5000cs silicone oil and 1000cs silicone oil was 45% and 8.3% respectively (P = 0.0002). After silicone oil removal, the chance of having an attached retina was higher for eyes filled with 1000cs silicone oil than those filled with 5000cs silicone oil (relative risk = 5.25).</p></sec><sec><title>Visual outcome following silicone oil removal</title><p>Overall, 28 eyes (34.2%) had improvement of visual acuity following silicone oil removal. Of these, 16 eyes belonged to the 1000cs silicone oil group [16/29 (55%)] and 12 eyes belonged to the 5000cs silicone oil group [12/53 (22.6%)]. Preoperative visual acuity of counting finger or less was seen in 65% of cases with 1000cs silicone oil and 73% of cases with 5000cs silicone oil. After silicone oil removal, visual acuity of counting fingers or less was observed in 45% and 77% respectively. Table <xref ref-type="table" rid="T3">3</xref> details patients' preoperative and postoperative visual acuity.</p><table-wrap position="float" id="T3"><label>Table 3</label><caption><p>Visual outcome before and after silicone oil removal.</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left">Visual acuity</td><td align="left" colspan="4">Before oil removal</td><td align="left" colspan="4">After oil removal</td></tr><tr><td></td><td colspan="8"><hr></hr></td></tr><tr><td></td><td align="left" colspan="2">Silicone 1000</td><td align="left" colspan="2">Silicone 5000</td><td align="left" colspan="2">Silicone 1000</td><td align="left" colspan="2">Silicone 5000</td></tr></thead><tbody><tr><td></td><td align="left">No</td><td align="left">%</td><td align="left">No</td><td align="left">%</td><td align="left">No</td><td align="left">%</td><td align="left">No</td><td align="left">%</td></tr><tr><td></td><td colspan="8"><hr></hr></td></tr><tr><td align="left">LP*</td><td align="left">6</td><td align="left">20.7</td><td align="left">19</td><td align="left">35.9</td><td align="left">5</td><td align="left">17.2</td><td align="left">27</td><td align="left">50.9</td></tr><tr><td align="left">FC†</td><td align="left">13</td><td align="left">44.9</td><td align="left">20</td><td align="left">37.7</td><td align="left">8</td><td align="left">27.6</td><td align="left">14</td><td align="left">26.4</td></tr><tr><td align="left">5/20–6/60</td><td align="left">8</td><td align="left">27.6</td><td align="left">13</td><td align="left">24.5</td><td align="left">12</td><td align="left">44.4</td><td align="left">7</td><td align="left">13.2</td></tr><tr><td align="left">6/45–6/15</td><td align="left">2</td><td align="left">6.9</td><td align="left">1</td><td align="left">1.9</td><td align="left">4</td><td align="left">13.8</td><td align="left">5</td><td align="left">9.4</td></tr><tr><td align="left">Total</td><td align="left">29</td><td align="left">100</td><td align="left">53</td><td align="left">100</td><td align="left">29</td><td align="left">100</td><td align="left">53</td><td align="left">100</td></tr></tbody></table><table-wrap-foot><p>*LP = Light perception – †FC = Finger counting</p></table-wrap-foot></table-wrap><p>Visual acuity of 6/120 or less was used as an outcome in a multivariate analysis and the association of factors such as initial visual acuity, intraocular pressure, type of oil viscosity, anatomic status of retina, age and the surgical indication were evaluated, while controlling for potential confounders. This analysis showed that only two factors were associated with visual acuity of 6/120 or less: initial visual acuity and type of silicone oil viscosity. Patients with initial visual acuity of 6/120 or less had a higher chance of having final visual acuity of 6/120 or less (OR = 32.2, 95%CI 7.4–140.2). Patients who had retained 5000cs silicone oil also had a higher chance of final visual acuity of 6/120 or less (OR = 7.9, 95%CI 1.9–32.2). Visual acuity varied to some extent with the basic underlying disorder that dictated silicone oil injection, however, these differences were not statistically significant.</p></sec><sec><title>Intraocular pressure following silicone oil removal (IOP)</title><p>Overall, prior to silicone oil removal 43 of 82 eyes (52.4%) had elevated IOP (i.e. ≥21mmHg). After silicone oil removal, IOP remained high in only 8 eyes (9.8%), in 69 eyes (84.1%) it was within the normal range (21 mmHg > IOP > 5 mmHg) and 5 eyes (6%) had severe hypotony (≤ 4 mmHg) [Table <xref ref-type="table" rid="T4">4</xref>]. There was no significant difference among the 4 categories of underlying disorders in regard to IOP, both before and after silicone oil removal.</p><table-wrap position="float" id="T4"><label>Table 4</label><caption><p>Intraocular pressure before and after silicone oil removal</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left">IOP*</td><td align="left" colspan="4">Before oil removal</td><td align="left" colspan="4">After oil removal</td></tr><tr><td></td><td align="left" colspan="2">Silicone 1000</td><td align="left" colspan="2">Silicone 5000</td><td align="left" colspan="2">Silicone 1000</td><td align="left" colspan="2">Silicone 5000</td></tr></thead><tbody><tr><td></td><td align="left">No</td><td align="left">%</td><td align="left">No</td><td align="left">%</td><td align="left">No</td><td align="left">%</td><td align="left">No</td><td align="left">%</td></tr><tr><td align="left"><21 mmHg</td><td align="left">14</td><td align="left">48.3</td><td align="left">25</td><td align="left">47.2</td><td align="left">26</td><td align="left">89.7</td><td align="left">43</td><td align="left">81.2</td></tr><tr><td align="left">>21 mmHg</td><td align="left">15</td><td align="left">51.79</td><td align="left">28</td><td align="left">52.8</td><td align="left">2</td><td align="left">6.9</td><td align="left">6</td><td align="left">11.3</td></tr><tr><td align="left">Phthisical</td><td align="left">-</td><td align="left">-</td><td align="left">-</td><td align="left">-</td><td align="left">1</td><td align="left">3.4</td><td align="left">4</td><td align="left">7.5</td></tr><tr><td align="left">Total</td><td align="left">29</td><td align="left">100</td><td align="left">53</td><td align="left">100</td><td align="left">29</td><td align="left">100</td><td align="left">53</td><td align="left">100</td></tr></tbody></table><table-wrap-foot><p>*IOP = Intraocular pressure</p></table-wrap-foot></table-wrap><p>After oil removal, elevated IOP was present in 2 eyes (6.9%) in the 1000cs silicone oil group and 6 eyes (11.3%) in the 5000cs silicone oil group. Excluding eyes that developed retinal re-detachment following silicone oil removal, only 8 out of 59 eyes (13.5%) had elevated IOP of more than 21 mmHg, 50 eyes (84.7%) had IOP within normal range and only one eye (1.6%) was phthisical (≤4 mmHg).</p><p>Postoperative IOP ≥ 21 mmHg was used as an outcome in a regression model and association of factors such as initial visual acuity, preoperative IOP > 21 mmHg, type of silicone oil viscosity, anatomic status of the retina, age and type of underlying disorder were evaluated. This analysis showed that only the anatomic status of the retina had significant association with IOP and expectedly, eyes with re-detachment following silicone oil removal had an average of 8.97 mmHg lower IOP than those with an attached retina.</p></sec></sec><sec><title>Discussion</title><p>Since the invention of the vitrectomy instrument, the role of silicone oil as a vitreous substitute and retinal tamponade has expanded. More recently, the beneficial effects of silicone oil have been re-confirmed in a multicenter clinical trial by the silicone oil study group[<xref ref-type="bibr" rid="B7">7</xref>,<xref ref-type="bibr" rid="B8">8</xref>]. Even though silicone oil has proved to be a very useful tool and adjunct in the treatment of complicated retinal detachments, the question of the preferred silicone oil viscosity for use in clinical settings still remains unanswered.</p><p>Earlier studies have shown no difference in the tamponading force among purified silicone oils of various viscosities[<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B3">3</xref>]. The most frequently used silicone oils are highly purified polydimethylsiloxanes with viscosities as low as 100cs to a maximum of 12500cs, however 1000cs and 5000cs define the viscosity range of currently used silicone oils in most vitreoretinal surgeries.</p><p>Low viscosity silicone oils are preferred by some surgeons because of easier surgical handling and removal from the vitreous cavity[<xref ref-type="bibr" rid="B2">2</xref>,<xref ref-type="bibr" rid="B5">5</xref>,<xref ref-type="bibr" rid="B6">6</xref>]. On the other hand, higher viscosity silicone oils are subject to decreased and delayed emulsification, so that the tamponading force lasts longer, which may provide better tamponade for some complex forms of retinal detachment that need a longer effect[<xref ref-type="bibr" rid="B2">2</xref>]. However, silicone oils of various viscosities have similar tamponading effects as long as emulsification of the oil has not occurred[<xref ref-type="bibr" rid="B3">3</xref>].</p><p>Despite these facts and the relative agreement on indications of silicone oil use and removal, the issue of preferred viscosity and time for removal deserve further investigation[<xref ref-type="bibr" rid="B9">9</xref>,<xref ref-type="bibr" rid="B10">10</xref>].</p><p>In this non-randomized study we evaluated the outcomes of successful complex retinal detachment surgery following removing silicone oils of two different viscosities, 1000cs and 5000cs. Prior to oil removal all eyes had a stable attached retina. In our patients, silicone oil emulsification occurred in all eyes to some degree. We may thus conclude that the tamponading force of silicone oil had been lost before removal. However, with this fact in mind, we observed an overall redetachment rate of 28%, which is almost similar to some other published reports on silicone oil removal before emulsification [<xref ref-type="bibr" rid="B10">10</xref>,<xref ref-type="bibr" rid="B11">11</xref>,<xref ref-type="bibr" rid="B16">16</xref>]. The majority of the recurrent retinal detachments occurred within 3 months of oil removal. Re-detachment rates in eyes with RD associated with PVR slightly exceeded other indications and interestingly the re-detachment rate in the 5000cs silicone oil group was significantly higher than the 1000cs group. The cause of this re-detachment following silicone oil removal was mostly residual traction and redevelopment of proliferative vitreoretinopathy that had led to reopening of preexisting retinal breaks, or formation of new retinal breaks as a result of surgical manipulations. These findings suggest that in making the decision to remove silicone oil from the eye, not only retinal stability, but also other factors that may aggravate PVR formation should be considered. In this study the re-detachment rate was higher for eyes filled with 5000cs silicone oil, especially eyes with RD and PVR. One possible explanation for this observation could be the use of 5000cs oil in more complex cases. Another explanation may be the length of time since primary vitrectomy surgery unrelated to the properties of the oil. It is possible that all retinas had an increased tendency to re-detach with extended follow-up and eyes that had been filled with 5000cs oil were seen later in the disease process due to longer retention period of the oil, therefore increasing the detection of re-detachment.</p><p>Overall, 34.2% of the eyes in our study experienced improvement of visual acuity following silicone oil removal. Elimination of the variability in refraction induced by the anterior curve of the silicone oil bubble as well as light diffraction induced by droplets of emulsified oil may have rendered the eye more amenable to optical correction.</p><p>Our data suggest that visual acuity prior to silicone oil removal and viscosity of silicone oil are both associated with final visual acuity: eyes with preoperative visual acuity of 6/120 or less and those filled with 5000cs silicone oil had less chance of obtaining a final visual acuity ≥ 6/120. There is no bias involved in the selection between the two varieties of silicone oil and that the selection was purely based on the availability of the same at the given point of time. Aside from the above-mentioned presumed selection bias, the longer duration of 5000cs silicone oil retainment (mean, 23.9 months) compared to 1000cs silicone oil (mean, 10 months) with its attendant pressure effect on the retina may somehow cause retinal damage. The other possible mechanism includes subclinical emulsification with gradual penetration and migration of oil droplets into the retina causing damage or toxicity and diminishing the of chance obtaining good visual outcome [<xref ref-type="bibr" rid="B17">17</xref>-<xref ref-type="bibr" rid="B20">20</xref>]. The difference between the retention time of 1000cs and 5000cs oil groups was due to earlier emulsification of silicone oil 1000cs and development of complications such as increased IOP and keratopathy. Indications for silicone removal in this study were the early detection of these complications such as emulsification alone or associated with increased IOP or keratopathy prior to removal of oil. This policy of postponing silicone oil removal until appearance of complications could skew our results toward a poorer outcome.</p><p>When elevated IOP (≥ 21 mmHg) following silicone oil removal was used as an outcome, in eyes with re-attached retina and controlling for potential confounders none of the factors in the multivariate regression analysis showed significant association with IOP. This may indicate that once silicone oil emulsifies, there should be no difference between the two different viscosities regarding IOP elevation.</p></sec><sec><title>Conclusion</title><p>Accomplishing a stable, re-attached retina is the final goal of all retinal surgeons. The possibility of achieving this goal could be increased with the use of higher viscosity silicone oil for longer periods especially in complex forms of retinal detachment. However, based on our present experience we observed a poorer anatomic and functional outcome with the use of higher viscosity silicone oil (5000cs) as compared to lower viscosity silicone oil (1000cs). It should be emphesized that due to the limitations of retrospective studies the validity of all these observations should be reconfirmed by a randomized clinical trial.</p></sec><sec><title>Abbreviations</title><p>*RD = Retinal detachment – †PDR = Proliferative diabetic retinopathy ‡IOFB = Intraocular foreign body – **PVR = Proliferative vitreoretinopathy.</p></sec><sec><title>Competing interests</title><p>The author(s) declare that they have no competing interests.</p></sec><sec><title>Authors' contributions</title><p>MS is the main researcher and the corresponding author. MMa has scheduled the cases and has recorded the data of the patients praoperative and follow-up sheets. MMo has revised the article and prepared the manuscript to be in the scientific format for publication. BR has helped in collecting the data and in statistics procedures.</p></sec><sec><title>Pre-publication history</title><p>The pre-publication history for this paper can be accessed here:</p><p><ext-link ext-link-type="uri" xlink:href="http://www.biomedcentral.com/1471-2415/6/21/prepub"/></p></sec> |
Accounting for the increase in NSAID expenditure: substitution or leakage? | <sec><title>Background</title><p>National Institute of Health and Clinical Excellence (NICE) guidance stated that a new form of non-steroidal anti-inflammatory drug (NSAID) (selective COX-2 inhibitors) should only be an option for arthritis patients at high risk of a gastro-intestinal (GI) event. Total expenditure on NSAIDs has risen by 57% over five years, to £247 million in 2004. We assess whether this expenditure increase can be accounted for by substitution – an increased prescribing of two (more expensive) selective COX-2 inhibitors (celecoxib and rofecoxib) and a simultaneous equivalent reduction in the prescribing volume of three (cheaper) older NSAIDs (diclofenac, ibuprofen and naproxen).</p></sec><sec sec-type="methods"><title>Methods</title><p>Quarterly prescription data was collated from January 1999 to September 2004. Over this period, the level of correlation between the total prescribing volumes for i) celecoxib and rofecoxib, and ii) diclofenac, ibuprofen, and naproxen were compared, the change in total expenditure on the five NSAIDs was also estimated. The latter was apportioned into that which was estimated to have arisen due to i) substitution, and ii) increased NSAID prescription volume.</p></sec><sec><title>Results</title><p>Total prescription volumes for the two NSAID groups were negatively correlated (r = -0.97, p < 0.001). In the last quarter there were 1.23 million prescriptions for celecoxib and rofecoxib, and 0.46 million fewer prescriptions for naproxen, diclofenac, and ibuprofen (than in the first quarter, when celecoxib and rofecoxib were not prescribed). Total expenditure for the five NSAIDs was £32.7 million higher in the last quarter, than the first, £12.2 million of which was estimated to be due to substitution, and £20.4 million due to increased volume.</p></sec><sec><title>Conclusion</title><p>The introduction of celecoxib and rofecoxib was associated with a reduction in the prescription volume for naproxen, diclofenac, and ibuprofen. However, overall quarterly prescription volume for these five NSAIDs increased by 0.76 million, and we estimate that quarterly expenditure increased by £20.4 million more than would have been expected if overall NSAID volume had remained constant. This suggests that the prescription of both celecoxib and rofecoxib may have 'leaked' to population groups who would not previously have received an older NSAID.</p></sec> | <contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Barton</surname><given-names>Garry R</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>mcxgrb@nottingham.ac.uk</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Avery</surname><given-names>Anthony J</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>tony.avery@nottingham.ac.uk</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Whynes</surname><given-names>David K</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>david.whynes@nottingham.ac.uk</email></contrib> | Cost Effectiveness and Resource Allocation | <sec><title>Background</title><p>Non-steroidal anti-inflammatory drugs (NSAIDs) are used extensively in the health service to relieve conditions that have an inflammatory component, and to relieve pain. In England, over 20 million NSAID treatments were prescribed in 2004, and NSAID expenditure amounted to £247 million (3.1% of annual prescription expenditure) [<xref ref-type="bibr" rid="B1">1</xref>].</p><p>NSAIDs can however induce adverse events – in the United Kingdom (UK) it has been estimated that each year they cause 3,500 adverse gastro-intestinal (GI) events (perforations, ulcers, bleeds, etc.) which require hospitalisation, and 400 deaths [<xref ref-type="bibr" rid="B2">2</xref>]. A gastro-protective agent (GPA) can be co-prescribed in an attempt to reduce the risk of a GI event, however, more recently, a new form of NSAID (the selective cyclooxygenase-2 (COX-2) inhibitor) has been developed. Selective COX-2 inhibitors aim to inhibit the COX-2 enzyme (which is responsible for inflammation), without inhibiting the COX-1 enzyme (which helps to protect the mucosa lining of the stomach and other parts of the gastro-intestinal tract) [<xref ref-type="bibr" rid="B3">3</xref>]. In contrast, older NSAIDs inhibit both the COX-2 and COX-1 enzymes, thus creating increased risk of adverse GI events.</p><p>In 2000, the National Institute of Clinical and Health Excellence (NICE) undertook an assessment of the available evidence on the health benefits and costs of selective COX-2 inhibitors for arthritis patients [<xref ref-type="bibr" rid="B4">4</xref>]. Though the assessment found no evidence that any of the (four) COX-2 inhibitors were clinically superior to one another, it did conclude that selective COX-2 inhibitors had equivalent efficacy to NSAIDs (in terms of their ability to reduce pain and improve physical functioning) and that selective COX-2 inhibitors were associated with fewer GI events than other NSAIDs [<xref ref-type="bibr" rid="B4">4</xref>]. On the basis of this evidence, NICE recommended that selective COX-2 inhibitors should not be routinely used (in preference to an older NSAID) by patients with arthritis, but that they should be an option for those who are at high risk of a GI event [<xref ref-type="bibr" rid="B5">5</xref>]. By December 2004 annual expenditure on selective COX-2 inhibitors had grown to over £150 million in England, and overall NSAID expenditure was approximately £65 million higher than in 2001 [<xref ref-type="bibr" rid="B1">1</xref>].</p><p>In this paper we seek to determine whether the aforementioned increase in overall NSAID expenditure can be accounted for by substitution i.e. the increased prescribing of (more expensive) selective COX-2 inhibitors and a simultaneous equivalent reduction in the prescribing volume of (cheaper) older NSAIDs. An alternative result of increased overall NSAID prescription volume – where the number of selective COX-2 inhibitor prescriptions has increased by a greater amount than the associated reduction in the number of older NSAID prescriptions – might indicate that selective COX-2 inhibitors were being prescribed to certain population groups who would not previously have been prescribed an older NSAID.</p><p>O'Brien [<xref ref-type="bibr" rid="B6">6</xref>] has used the term 'leakage' to refer to the situation where once an intervention is provided for a specific indication and population group (for whom there is evidence of cost-effectiveness) it can 'leak' to other groups for whom it was not originally intended (and for whom it may also be less cost-effective). An example of leakage was given by Lopert [<xref ref-type="bibr" rid="B7">7</xref>], who pointed out that ACE inhibitors are more cost-effective for cardiac failure than for hypertension (they provide no clear benefit over beta blockers, but are considerably more expensive). Similarly, a retrospective examination of the appropriateness of proton pump inhibitor (PPI) prescribing found that in 49.8% of the cases examined patients who were prescribed a PPI did not meet the Australian government Pharmaceutical Benefits Scheme (PBS) prescribing criteria, criteria which was drawn up to restrict the use of PPIs on the grounds of cost-effectiveness [<xref ref-type="bibr" rid="B8">8</xref>]. More recently, NICE has recommended that herceptin be an option for women with advanced stage breast cancer [<xref ref-type="bibr" rid="B9">9</xref>], but women with early stage breast cancer are now also receiving the drug, even though herceptin is yet to be licensed for use, or evaluated in terms of cost-effectiveness, in this population group [<xref ref-type="bibr" rid="B10">10</xref>].</p><p>As well as recommending that selective COX-2 inhibitors only be an option for patients with arthritis who are at high risk of an adverse GI event, NICE also estimated that switching high-risk arthritis patients to selective COX-2 inhibitors would lead to an annual incremental expenditure of approximately £25 million to the National Health Service (NHS) [<xref ref-type="bibr" rid="B5">5</xref>]. This NICE guidance is consistent with other studies which have estimated that, compared to older NSAIDs, selective COX-2 inhibitors have a higher incremental cost, and that their provision is only cost-effective for groups at high risk of a GI event [<xref ref-type="bibr" rid="B11">11</xref>,<xref ref-type="bibr" rid="B12">12</xref>]. In this paper, we compare actual increases in overall NSAID expenditure to the £25 million per annum predicted by NICE. Evidence of expenditure increases beyond those that were predicted may also suggest that the provision of selective COX-2 inhibitors had leaked to population groups for whom its provision was not originally intended.</p><sec><title>Specific approach</title><p>In this paper we use data on the number of prescriptions, and their associated expenditure, as calculated by the Prescription Pricing Authority (PPA) (the PPA is responsible for processing all NHS prescriptions that are dispensed by any community pharmacy or dispensing doctor in England). Data for two selective COX-2 inhibitors (celecoxib and rofecoxib) and three older NSAIDs (diclofenac, ibuprofen, and naproxen) were collated. These two particular selective COX-2 inhibitors were chosen as they accounted for 78% of all selective COX-2 inhibitor prescription items in 2004 [<xref ref-type="bibr" rid="B1">1</xref>]. Similarly, diclofenac, ibuprofen, and naproxen were chosen as in 1998 (prior to the use of celecoxib and rofecoxib in England) they accounted for 77% of all NSAID prescription items [<xref ref-type="bibr" rid="B13">13</xref>]. In addition, diclofenac and ibuprofen were the comparators used in the largest randomised controlled trial (N = 8059) designed to assess the efficacy of celecoxib [<xref ref-type="bibr" rid="B14">14</xref>], and naproxen was the comparator used in the largest randomised controlled trial (N = 8076) designed to assess the efficacy of rofecoxib [<xref ref-type="bibr" rid="B15">15</xref>]. Data from both these clinical trials were submitted to the United States (US) Federal Drug Administration (FDA) for the purposes of licensing celecoxib and rofecoxib, respectively.</p></sec></sec><sec sec-type="methods"><title>Methods</title><sec><title>NSAID prescription volume</title><p>PPA data, for England, on the number of prescriptions for each individual NSAID were collated for quarterly (3 month) periods from January 1<sup>st </sup>1999 (prior to the use of celecoxib and rofecoxib) to September 30<sup>th </sup>2004: 23 quarters in total. Total quarterly prescription volumes for i) the two selective COX-2 inhibitors (celecoxib and rofecoxib), and ii) the three older NSAIDs (diclofenac, ibuprofen, and naproxen) were calculated for each of the 23 quarters. Bi-variate (Pearson) correlations were undertaken to compare total prescriptions volumes for the two selective COX-2 inhibitors to total prescription volumes for the three older NSAIDs over the 23 quarter period. Changes in the absolute prescription volume for both these groups were also calculated by comparing prescription volumes in the first quarter (ending March 1999) to those in the last quarter (ending September 2004).</p></sec><sec><title>NSAID expenditure</title><p>PPA total prescription expenditure data, for England, was collated for the two selective COX-2 inhibitors (celecoxib and rofecoxib) and three older NSAIDs (diclofenac, ibuprofen, and naproxen) for the 23 quarters between January 1999 and September 2004. Similarly, the total increase in expenditure for the five NSAIDs, between the first and last quarter, was calculated. To account for the fact that, over time, higher NSAID prescription costs could have accounted for some of the increase in NSAID expenditure we also compared the average prescription cost for diclofenac, ibuprofen and naproxen in the last quarter to that in the first quarter. Average prescription costs, for each NSAID, were estimated by dividing the actual total prescription expenditure for the particular NSAID (in a particular time period) by the actual number of prescriptions made for the same NSAID (in the same time period).</p><p>Using the following methods, we also estimated the approximate increase in NSAID expenditure that arose due to i) substitution, and ii) increased NSAID prescription volume. To estimate the former, the reduction in the number of prescriptions for the three older NSAIDs (between the first and last quarter) was identified, and we assumed that the total number of selective COX-2 inhibitor prescriptions in the final quarter was equivalent to this reduction (i.e. that the total number of prescriptions for the five NSAIDs had remained constant between the first and last quarter), and that an equivalent number of prescriptions for celecoxib and rofecoxib were made. These quarterly prescription volumes for celecoxib and rofecoxib were then multiplied by their respective average prescription costs (for the last quarter), and summed together, in order to estimate what the total expenditure on the two selective COX-2 inhibitors (in the last quarter) would have been if the total number of prescriptions for the five NSAIDs had remained constant. This estimate of the expenditure on the two selective COX-2 inhibitors (based on constant NAID volume) was then added to the actual expenditure on diclofenac, ibuprofen and naproxen (in the last quarter) in order to estimate what the total expenditure on these the five NSAIDs would have been (in the last quarter), if the total prescription volume for these five NSAIDs had remained constant. Thus, the approximate increase in NSAID expenditure that arose due to substitution was then calculated by deducting the actual expenditure on the five NSAIDs in the first quarter, from the estimated expenditure on the five NSAIDs in the last quarter, based on a constant volume of prescriptions for these five NSAIDs.</p><p>The approximate increase in NSAID expenditure that arose due to increased NSAID prescription volume was estimated by deducting the previous (substitution) estimate of what the total expenditure on the five NSAIDs would have been in the final quarter (based on a constant nsaid volume) from the actual total expenditure on the five NSAIDs (in the final quarter). The increase in NSAID expenditure was thereby divided into that which was estimated to be accounted for by substitution (falling total diclofenac, ibuprofen, and naproxen volume), and that which could not (i.e. that which was a result of increasing overall NSAID volume). Finally, we compared these estimates to the NICE estimate that switching high-risk arthritis patients to selective COX-2 inhibitors would lead to an annual incremental NSAID expenditure of approximately £25 million.</p></sec></sec><sec><title>Results</title><sec><title>NSAID prescription volume</title><p>By the last quarter (ending September 2004) prescription volumes for celecoxib and rofecoxib had reached levels of 606,409 and 620,790 items per quarter respectively (Figure <xref ref-type="fig" rid="F1">1</xref>). Figure <xref ref-type="fig" rid="F1">1</xref> also shows that between the first quarter (ending March 1999) and the last quarter the number of prescriptions for ibuprofen and naproxen fell, but that the number of prescriptions for diclofenac remained relatively stable. Indeed between the first and last quarter total prescription volume for the two selective COX-2 inhibitors (celecoxib and rofecoxib) increased to 1.23 million per quarter, whilst prescription volumes for the three older NSAIDs (diclofenac, ibuprofen and naproxen) fell from 3.55 million to 3.09 million (Figure <xref ref-type="fig" rid="F2">2</xref>).</p><fig position="float" id="F1"><label>Figure 1</label><caption><p>Individual NSAID quarterly NHS prescription volume England for the five NSAIDs (celecoxib, rofecoxib, diclofenac, ibuprofen and naproxen) between January 1999 and September 2004.</p></caption><graphic xlink:href="1478-7547-4-9-1"/></fig><fig position="float" id="F2"><label>Figure 2</label><caption><p>Total NSAID quarterly NHS prescription volume for England for i) celecoxib and rofecoxib, ii) diclofenac, ibuprofen, and naproxen, and iii) all five NSAIDs between January 1999 and September 2004.</p></caption><graphic xlink:href="1478-7547-4-9-2"/></fig><p>Total prescription volumes for the two selective COX-2 inhibitors, and three older NSAIDs were highly negatively correlated (r = -0.97, p < 0.001). Despite this, prescription volumes for three older NSAIDs fell by only 460,084 items per quarter between the first and last quarter, whereas the total number of prescriptions for two selective COX-2 inhibitors grew to 1,227,199 per quarter over the same period (Figure <xref ref-type="fig" rid="F2">2</xref>). As a consequence the cumulative number of prescriptions for the five NSAIDs grew from 3,550,249 in the first quarter to 4,317,364 in the last quarter – an increase of over 0.76 million items per quarter.</p></sec><sec><title>NSAID prescription expenditure</title><p>By the end of September 2004, expenditure on celecoxib and rofecoxib amounted to £15,430,892 per quarter and £17,221,195 per quarter, respectively, (Figure <xref ref-type="fig" rid="F3">3</xref>). Conversely, total expenditure on the three older NSAIDs fell from £26,583,346 in the first period to £18,671,535 in the final quarter (Figure <xref ref-type="fig" rid="F4">4</xref>). As such, over the 23 quarter period, overall expenditure on the five NSAIDs increased from £26,061,182 per quarter to £51,323,622 per quarter (Figure <xref ref-type="fig" rid="F4">4</xref>), an increase of £25.3 million per quarter. In the first quarter the average prescription costs for diclofenac, ibuprofen and naproxen were £11.53, £2.33 and £8.11, respectively, compared to £7.89, £2.71 and £7.02, respectively, in the last quarter. As these average prescription costs have generally fallen the increase in total NSAID expenditure can not be accounted for by an increase in average prescription costs (though the possibility that lower costs led to a higher volume than might otherwise have been the case can not be discounted). The average prescription costs for celecoxib and rofecoxib were £17.65 and £22.47 when they were first prescribed (in the quarters ending June 2000, and June 1999, respectively), compared to £25.45 and £27.74 in the final quarter.</p><fig position="float" id="F3"><label>Figure 3</label><caption><p>Individual NSAID quarterly NHS expenditure for England for the five NSAIDs (celecoxib, rofecoxib, diclofenac, ibuprofen and naproxen) between January 1999 and September 2004.</p></caption><graphic xlink:href="1478-7547-4-9-3"/></fig><fig position="float" id="F4"><label>Figure 4</label><caption><p>Total NSAID quarterly NHS expenditure for England for i) celecoxib and rofecoxib, ii) diclofenac, ibuprofen, and naproxen, and iii) all five NSAIDs between January 1999 and September 2004.</p></caption><graphic xlink:href="1478-7547-4-9-4"/></fig><p>To estimate the approximate increase in total NSAID expenditure that arose due to substitution we assumed that 230,042 prescriptions for both celecoxib and rofecoxib would have been made in the last quarter (prescription volume for diclofenac, ibuprofen and naproxen fell by 460,084 per quarter between the first and last quarter). By multiplying these volumes by the average prescription costs of £25.45 and £27.74, the total expenditure on celecoxib and rofecoxib would have been estimated to be £5,853,728 and £6,381,543, respectively (a combined total expenditure of £12,235,271 for the two selective COX-2 inhibitors). By adding this combined total expenditure to the actual expenditure on the three older NSAIDs in the final quarter (£18,671,535) we estimated that the expenditure on the five NSAIDs would have been £30,906,806 (if NSAID volume had remained constant between the first and last quarter). The actual expenditure on the five NSAIDs was £26,061,182 in the first quarter, thus we estimate that the approximate increase in NSAID expenditure that arose due to substitution was £4,845,625 per quarter (equivalent to £19,382,498 per annum).</p><p>To estimate of the approximate increase in NSAID expenditure that arose due to increased NSAID prescription volume, between the first and last quarter, we deducted the value of £30,906,806 (the estimated expenditure for the five NSAIDs based on constant volume) from the actual expenditure on the five NSAIDs in the last quarter (£51,323,622). Thus we estimate that the actual expenditure on the five NSAIDs was £20,416,816 per quarter higher than we estimate the expenditure on the five NSAIDs would have been if the last quarter total prescription volume for the five NSAIDs had been equivalent to that in the first quarter, and that increased volume accounted for 81% of the increase in NSAID expenditure. This largely arose because, based on a constant volume, total expenditure on the two selective COX-2 inhibitors would have been estimated to be £12,235,271 in the last quarter, whereas it actually amounted to £32,652,087 in the last quarter.</p></sec></sec><sec><title>Discussion</title><p>It has been shown that the introduction of two newer NSAIDs (celecoxib and rofecoxib) was associated with a fall in total prescription volume for three older NSAIDs (diclofenac, ibuprofen, and naproxen). However, for the five NSAIDs overall there were 0.76 million more prescriptions in the last quarter, compared to the first, and expenditure rose by £25.3 million per quarter (equivalent to an increase of over £100 million per annum).</p><p>Had the volume of prescriptions for the five NSAIDs remained constant over the 23 quarter period we estimate that due to substitution (from cheaper NSAIDs to more expensive selective COX-2 inhibitors) expenditure would have been £30.9 million in the last quarter, compared to £26.1 million in the first quarter (the difference of £4.85 million per quarter is equivalent to an annual increase of approximately £19.4 million). NICE estimated that switching high-risk arthritis patients to selective COX-2 inhibitors would increase NSAID expenditure by £25 million per annum. This figure is comparable to the increase in NSAID expenditure that we estimated was due to substitution (£19.4 million). The higher increase estimate by NICE may be partially explained by the fact that expenditure on the five NSAIDs in the quarter prior to the NICE report (quarter ending June 2001) was £23.1 million, compared to £26.1 million in the first quarter of our data set, and by the fact that our expenditure estimates were based on only two selective COX-2 inhibitors, compared to the four used by NICE [<xref ref-type="bibr" rid="B5">5</xref>]. Either way, the NICE estimate that switching patients to selective COX-2 inhibitors would increase NSAID expenditure by £25 million per annum is substantially below the actual increase in NSAID expenditure, which we estimated to be equivalent to over £100 million (for the five NSAIDs, between January 1999 and September 2004). Indeed we estimated that 81% of the increase in NSAID expenditure was due to increased NSAID volume.</p><sec><title>Explanations</title><p>In randomised trials the estimated number of adverse GI events for people taking the NSAIDs of naproxen or ibuprofen has been estimated to be between two and four times higher than the number for people taking rofecoxib [<xref ref-type="bibr" rid="B15">15</xref>-<xref ref-type="bibr" rid="B18">18</xref>]. Similarly, people taking diclofenac or ibuprofen were estimated to suffer between two and three times the number of adverse GI events as those taking celecoxib [<xref ref-type="bibr" rid="B14">14</xref>]. Consequently, it is not surprising that the release of celecoxib and rofecoxib coincided with a reduction in the number of prescriptions for naproxen and ibuprofen. We can not readily explain why prescriptions levels for diclofenac remained largely unchanged over the six year period, though the fact that the average prescription cost of diclofenac fell by 32% during this time may be a contributory factor. As overall volume for the five NSAIDs has increased by 0.76 million items per quarter this may however suggest that both celecoxib and rofecoxib could have leaked to population groups who would not previously have received a NSAID. People may not have been prescribed NSAIDs previously because, even though they were in pain, the potential benefits were judged not to outweigh the risk of an adverse GI event – a situation which presumably changed with the arrival of celecoxib and rofecoxib.</p></sec><sec><title>Comparisons with other studies</title><p>Our results are in line with those from two recent studies. Joshua et al. [<xref ref-type="bibr" rid="B19">19</xref>] found that use of selective COX-2 inhibitors by rheumatology patients increased from 18% (3 months after their release) to 57% (16 months after their release). During the same period prescription rates for other NSAIDs fell from 43% to 20%. Moreover, Joshua et al. [<xref ref-type="bibr" rid="B19">19</xref>] found that prescribing patterns for selective COX-2 inhibitors were largely unrelated to the patient's estimated risk of an adverse GI event. Similarly, Dai et al. [<xref ref-type="bibr" rid="B20">20</xref>] found that 35% of patients at the lowest risk for adverse events from NSAIDs received a COX-2 inhibitor in 2002, compared to just 12% in 1999. In addition, when discussing the use of selective COX-2 inhibitors, the US Food and Drug Administration (FDA) concluded that they had been "...prescribed for indications and patients far beyond their original intent" [<xref ref-type="bibr" rid="B21">21</xref>].</p></sec><sec><title>Study weaknesses</title><p>One of the main weaknesses of this study is that we have looked at prescription items, and not the number of patients taking each individual NSAID. Moreover, the number of prescription items does not take account of the dosage or quantity of the drug prescribed [<xref ref-type="bibr" rid="B22">22</xref>]. Thus though we know that the number of prescriptions for the five NSAIDs has increased by 0.76 million items per quarter we can not be sure that more patients have been prescribed a NSAID (it may be that the same patients are receiving more prescriptions than previously). Moreover this increased volume could be due to confounding factors that have not been controlled for in this 6 year study period. For example, it has been estimated by Harkness et al. [<xref ref-type="bibr" rid="B23">23</xref>] that over the past 40 years the prevalence of musculoskeletal pain has increased by between 2 and 4 fold, though it was acknowledged that this could be due to increased awareness and reporting. Either way, this means that some other factor (e.g. increased prevalence), rather than the release of celecoxib and rofecoxib, could partially account for the increased volume of NSAID prescriptions. That said, weight is added to the argument that there is a substitution effect between selective COX-2 inhibitors and older NSAIDs, by the fact that prescription volumes for diclofenac, ibuprofen and naproxen were increasing prior to the fall that occurred after the release of celecoxib and rofecoxib (in total, there were 14.42 million prescriptions for these three older NSAIDs in 1999 [<xref ref-type="bibr" rid="B24">24</xref>], compared to 14.15 million in 1998 [<xref ref-type="bibr" rid="B13">13</xref>]).</p><p>One of the five NSAIDs (ibuprofen) is available without prescription (often referred to as over-the-counter (OTC)), but we did not monitor how OTC sales of ibuprofen have changed since the release of celecoxib and rofecoxib. Were OTC sales of ibuprofen to have increased over the 6 year period then this, rather than the release of celecoxib and rofecoxib, could account for some of the reduction in ibuprofen prescriptions. This in turn would however mean that the level of NSAID volume would have increased by an even greater amount than that estimated in this study.</p><p>Finally, readers should be aware that rofecoxib has recently been withdrawn from the market by its manufacturer Merck [<xref ref-type="bibr" rid="B25">25</xref>]. This was due to evidence of an increased risk of serious coronary heart disease when taking rofecoxib, particularly at high dosages and for prolonged periods [<xref ref-type="bibr" rid="B26">26</xref>,<xref ref-type="bibr" rid="B27">27</xref>]. Some studies (e.g. [<xref ref-type="bibr" rid="B28">28</xref>]), but not others (e.g. [<xref ref-type="bibr" rid="B29">29</xref>]), have also found that the use of celecoxib increases a person's risk of coronary heart disease. However, after considering all available evidence, the US FDA decided to not to withdraw celecoxib and other selective COX-2 inhibitors from the market [<xref ref-type="bibr" rid="B30">30</xref>]. Given that rofecoxib was withdrawn after the period for which data were collected within this study we do not believe that this impinges on the results of this study. Indeed celecoxib and rofecoxib were the main two selective COX-2 inhibitors in England in 2004, accounting for 74% of NHS expenditure on selective COX-2 inhibitors [<xref ref-type="bibr" rid="B1">1</xref>]. Moreover, it was the evidence of increased risk of coronary heart disease in a group for which the rofecoxib was not originally intended (bowel cancer patients) that led to its withdrawal [<xref ref-type="bibr" rid="B26">26</xref>].</p></sec></sec><sec><title>Conclusion</title><p>It has been shown that the introduction of celecoxib and rofecoxib coincided with a reduction in the total number of prescriptions for diclofenac, ibuprofen and naproxen. However, between January 1999 and September 2004, overall prescription volume for the five NSAIDs increased by 0.76 million items, and expenditure was estimated to have increased by £20.4 million per quarter more than would have been expected if overall NSAID volume had remained constant. This suggests that celecoxib and rofecoxib may have leaked to population groups who would not previously have received a NSAID.</p></sec><sec><title>Competing interests</title><p>The author(s) declare that they have no competing interests.</p></sec><sec><title>Authors' contributions</title><p>GRB had the original idea for the study, undertook the statistical analyses and drafted the paper. Both AJA and DKW made revisions to the paper. All authors have given final approval for this version of the paper to be published.</p></sec> |
Innate immune functions of microglia isolated from human glioma patients | <sec><title>Background</title><p>Innate immunity is considered the first line of host defense and microglia presumably play a critical role in mediating potent innate immune responses to traumatic and infectious challenges in the human brain. Fundamental impairments of the adaptive immune system in glioma patients have been investigated; however, it is unknown whether microglia are capable of innate immunity and subsequent adaptive anti-tumor immune responses within the immunosuppressive tumor micro-environment of human glioma patients. We therefore undertook a novel characterization of the innate immune phenotype and function of freshly isolated human glioma-infiltrating microglia (GIM).</p></sec><sec sec-type="methods"><title>Methods</title><p>GIM were isolated by sequential Percoll purification from patient tumors immediately after surgical resection. Flow cytometry, phagocytosis and tumor cytotoxicity assays were used to analyze the phenotype and function of these cells.</p></sec><sec><title>Results</title><p>GIM expressed significant levels of Toll-like receptors (TLRs), however they do not secrete any of the cytokines (IL-1β, IL-6, TNF-α) critical in developing effective innate immune responses. Similar to innate macrophage functions, GIM can mediate phagocytosis and non-MHC restricted cytotoxicity. However, they were statistically less able to mediate tumor cytotoxicity compared to microglia isolated from normal brain. In addition, the expression of Fas ligand (FasL) was low to absent, indicating that apoptosis of the incoming lymphocyte population may not be a predominant mode of immunosuppression by microglia.</p></sec><sec><title>Conclusion</title><p>We show for the first time that despite the immunosuppressive environment of human gliomas, GIM are capable of innate immune responses such as phagocytosis, cytotoxicity and TLR expression but yet are not competent in secreting key cytokines. Further understanding of these innate immune functions could play a critical role in understanding and developing effective immunotherapies to malignant human gliomas.</p></sec> | <contrib id="A1" contrib-type="author"><name><surname>Hussain</surname><given-names>S Farzana</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>sfhussai@mdanderson.org</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Yang</surname><given-names>David</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>dsyang@mdanderson.org</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Suki</surname><given-names>Dima</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>dsuki@mdanderson.org</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Grimm</surname><given-names>Elizabeth</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>egrimm@mdanderson.org</email></contrib><contrib id="A5" corresp="yes" contrib-type="author"><name><surname>Heimberger</surname><given-names>Amy B</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>aheimber@mdanderson.org</email></contrib> | Journal of Translational Medicine | <sec><title>Background</title><p>Malignant gliomas are the most common type of primary brain tumors and glioblastoma multiforme accounts for more than 50% of all intracranial gliomas [<xref ref-type="bibr" rid="B1">1</xref>]. These tumors are extremely aggressive and are characterized by diffuse infiltration of the brain parenchyma, recurrent growth, and an extremely poor prognosis for survival. Although glioblastoma multiforme is immunogenic [<xref ref-type="bibr" rid="B2">2</xref>-<xref ref-type="bibr" rid="B4">4</xref>], immune-mediated eradication does not occur, and attempts at immunotherapy directed against brain tumors have been minimally successful thus far [<xref ref-type="bibr" rid="B5">5</xref>-<xref ref-type="bibr" rid="B7">7</xref>]. Previously characterized impairments in glioma immunity have included low peripheral lymphocyte counts, reduced delayed type hypersensitivity reactions to recall antigens, impaired mitogen-induced blastogenic responses by peripheral blood mononuclear cells (PBMCs), and increased CD8<sup>+ </sup>suppressor T cells [<xref ref-type="bibr" rid="B8">8</xref>]. Adaptive immune responses are noticeably deficient, with diminished responsiveness of peripheral T cells associated with impaired early transmembrane signaling through the T-cell receptor/CD3 complex [<xref ref-type="bibr" rid="B9">9</xref>]. In addition, diminished induction of immunoglobulin synthesis by B cells <italic>in vitro </italic>from the peripheral blood of patients with intracranial tumors appears to be related to diminished T-helper activity [<xref ref-type="bibr" rid="B10">10</xref>]. Although immune impairments have been identified within the adaptive arm, few studies have examined for potential deficits in the innate arm, especially within the context of immunosuppressed glioma patients.</p><p>Innate immunity is the initial antigen-nonspecific response that results in the rapid production of effector cytokines and is one of the prerequisites for triggering effective adaptive antitumor immune responses. Soluble factors produced by gliomas, such as immunosuppressive cytokines (e.g., transforming growth factor [TGF-β ] and interleukin [IL-10]), presumably impair other cells participating in innate immunologic responses. Microglia are the most prominent immune cell within the CNS, however, it is not known whether microglia, within the immunosuppressive tumor environment, are capable of activated or functional innate immune responses. Microglia are unique to the central nervous system (CNS) and account for as much as 20% of the non-neuronal cell population [<xref ref-type="bibr" rid="B11">11</xref>]. Microglia are defined as being CD11b/c<sup>+</sup>CD45<sup>low</sup>, macrophages as CD11b/c<sup>+</sup>CD45<sup>high</sup>, and lymphocytes as CD11b/c<sup>-</sup>CD45<sup>high</sup>[<xref ref-type="bibr" rid="B12">12</xref>]. There is not a distinct universally accepted histological marker to distinguish macrophages from microglia. However, macrophages are believed to be activated microglia within the CNS. Rodent studies have shown that microglia play a critical effector role in rapidly responding to certain autoimmune and viral infections [<xref ref-type="bibr" rid="B13">13</xref>]. Microglia are thought to be capable of phagocytosis, cytotoxicity, and other pro-inflammatory innate effector functions [<xref ref-type="bibr" rid="B14">14</xref>]. The cells of the innate immune system, especially macrophages, use Toll-like receptors (TLRs) to recognize microbial or non-self factors, such as pathogen-associated molecular patterns. TLRs are not constitutively expressed in the brain parenchyma [<xref ref-type="bibr" rid="B15">15</xref>], but cultured primary murine microglia cells express mRNA encoding TLRs, which are strongly activated after stimulation with their specific agonists (lipopolysaccharide (LPS) (TLR-4), peptidoglycan (TLR-2), dsRNA (TLR-3), or CpG motifs (TLR-9)) [<xref ref-type="bibr" rid="B16">16</xref>]. Signaling through these receptors is coupled to gene transcription processes and has powerful immunomodulatory effects that can result in the activation of anti-tumor immune responses. Murine microglia have been shown to express the mRNA for all TLRs (TLR1-9) [<xref ref-type="bibr" rid="B13">13</xref>]. However, these microglia cell lines have been propagated extensively in culture and thus their phenotype and function could be markedly altered from the characteristics that they would normally have <italic>in situ</italic>.</p><p>The soluble expression of FasL fails to induce inflammatory responses. In contrast, when FasL is expressed on the surface of tumor cells, neutrophil-mediated inflammation is triggered initiating a vigorous innate immune response [<xref ref-type="bibr" rid="B17">17</xref>,<xref ref-type="bibr" rid="B18">18</xref>]. The complex immunological role of FasL has been confounded by a recent report that tumor expression of FasL impairs NK activation [<xref ref-type="bibr" rid="B19">19</xref>]. Expression of FasL has been shown to maintain immune privilege though inducing apoptosis of infiltrating Fas positive effector T cells. In the murine G26 model system, 50% of the total FasL-positive expression in intracranial tumors was accounted for by microglia, and neutralization of FasL resulted in a significant increase in the number of tumor-infiltrating lymphocytes [<xref ref-type="bibr" rid="B20">20</xref>]. This would indicate that microglia may play a role in down regulating innate immune responses within the CNS.</p><p>In this study, for the first time, we have characterized the innate immune phenotype and function of microglia isolated from human glioma tissue immediately after surgical resection. Using sequential Percoll density gradients, the purity of GIM was determined on the basis of previously established CD11b<sup>+</sup>CD45<sup>low </sup>parameters [<xref ref-type="bibr" rid="B21">21</xref>]. The ability of GIM to participate in innate immune function was assessed by analyzing the surface expression of TLRs, their ability to mediate tumor cytotoxicity, phagocytosis and expression of FasL. Similar to findings in the murine systems, GIM did express TLR, however GIM did not elaborate cytokines reflective of innate activation. In marked contrast to findings within the murine system, GIM do not express FasL. Furthermore, although GIM could mediate non-MHC restricted cytotoxicity, microglia isolated from normal brain were statistically more efficient at tumor cytotoxicity. By characterizing the activation state and functioning of GIM, we can not only comprehensively identify their role in immune responses to human gliomas, but we can also begin to develop ways in which they can be manipulated to successfully design and optimize immunotherapeutic strategies against malignant brain tumors.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>Human subjects</title><p>Patients' tumors (n = 50) were graded pathologically as glioblastoma multiforme by a neuropathologist according to the WHO classification [<xref ref-type="bibr" rid="B22">22</xref>,<xref ref-type="bibr" rid="B23">23</xref>]. Fresh normal brain tissue (n = 6) can only be rarely obtained since for most patients efforts are made by the neurosurgeon to preserve this tissue. On occasion, normal brain tissue was obtained during an approach to a benign tumor such as a sphenoid wing meningioma, a well-circumscribed metastasis or a low-grade lesion such as an oligodendroglioma. Additionally, peritumoral GBM tissue was also obtained and separately analyzed, however because this tissue could harbor microscopic tumor it was not characterized as normal for the purposes of this study. The resected normal tissue demonstrated no evidence of increased signal on flair MRI pre-operative imaging. The brain architecture was inspected microscopically intraoperatively by an experienced neurosurgeon (A.B.H.) and found to be consistent with normal brain tissue prior to submission to the laboratory. Peripheral blood was drawn from the patients intraoperatively. This study was conducted under protocol # LAB03-0687, which was approved by the institutional review board of The University of Texas M.D. Anderson Cancer Center, and informed consent was obtained.</p></sec><sec><title>Prioritization of characterization</title><p>Microglia could never be isolated in sufficient numbers from a single GBM specimen to perform all of the assays concurrently. Furthermore, there is significant variability in the numbers of microglia obtained from each clinical specimen that could not be predicted preoperatively. A minimum of 5 grams of tissue was necessary in order to isolate sufficient microglia for the basic phenotypic characterizations. Purity was established on all specimens regardless of the assay priority. Fresh surgical tissues were analyzed as follows: 1) If a normal and GBM specimen could be obtained simultaneously then priority went to performing the cytotoxicity assays. In two years, this scenario was very rare even at a hospital specializing in neuro-oncological neurosurgery. Routinely drawn intraoperative peripheral blood was accessed for the controls. 2) For specimens > 5 grams, the microglia were phenotypically and functionally characterized.</p></sec><sec><title>Isolation of microglia from human brain tumor tissue samples</title><p>Microglia were purified using modifications to an isolation technique previously described [<xref ref-type="bibr" rid="B21">21</xref>] after evaluation of the phenotype of each interphase. This technique minimized artificial activation of the microglia and these cells were isolated usually within three hours of surgical resection. Briefly, after surgical resection, freshly isolated tumor, peritumoral, or normal brain tissue was mechanically dissociated through a stainless steel sieve. The dissociated material was centrifuged and the pellet was washed. Cells were layered onto an isotonic Percoll (Amersham Biosciences, Uppsala, Sweden) gradient diluted to 1.095 g/mL and overlaid with 1.03 g/mL Percoll. After centrifugation, the visible cell layer between the 1.095 g/mL and 1.03 g/mL layers was removed, washed, layered on top of a second gradient (2-mL steps of isotonic Percoll diluted to 1.12, 1.088, 1.072, 1.065, and 1.03 g/mL densities) and centrifuged. Microglia were collected from the interface between the 1.065 g/mL and 1.03 g/mL layers and washed. Their viability was then determined by the Trypan blue dye-exclusion method.</p></sec><sec><title>Antibodies</title><p>Cell surface staining was performed with phycoerythrin- or fluorescein isothiocyanate-labeled antibodies against the following proteins: CD11b, CD11c, CD16, CD32, CD45, FasL (Pharmingen, San Diego, CA); and TLR-1, -2, -3, and -4 (eBiosciences, San Diego, CA). For intracellular cytokine staining, we used phycoerythrin- or fluorescein isothiocyanate- labeled antibodies against TNF-α, IL-1β and IL-6 (R&D Systems, Minneapolis, MN). Appropriate isotype controls were used for every antibody.</p></sec><sec><title>Cell surface and intracellular cytokine marker staining</title><p>Microglia were Fc blocked for 20 minutes using purified anti-CD16 antibody (Pharmingen). After washing, the cells were incubated with the fluorescent-labeled primary antibody or isotype control for 1 hour at 4°C. For the intracellular cytokine analysis, microglia were fixed with Cytofix/Cytoperm (BD Biosciences, San Jose, CA), washed in PermWash (BD Biosciences), and then stained with fluorescence-labeled monoclonal antibodies or isotype controls for 30 min at 4°C. Positive controls for all of the monoclonal antibodies consisted of A375 cells (E. G.). Approximately 1 × 10<sup>4 </sup>live, gated events were assessed during fluorescence-activated cell sorting using an Epics XL-MCL cytometer (Beckman Coulter, Mountain View, CA) and analyzed using IsoContour software (Verity Software House, Topsham, ME).</p></sec><sec><title>Phagocytosis assay</title><p>Fluorescent-labeled polystyrene microparticles (0.99 μm; Polysciences, 10 μL/tube) were coated with fetal bovine serum (FBS) diluted to 50% in phosphate-buffered saline and incubated with microglia (1 × 10<sup>6</sup>/tube) for 30 minutes at 37°C. The reaction was stopped by the addition of ice-cold phosphate-buffered saline. Cells were washed and analyzed by fluorescence microscopy (Nikon E400, Lewisville, TX) for internalized particles.</p></sec><sec><title>Tumor cytotoxicity assay</title><p>1 to 2 × 10<sup>6 </sup>U-87 MG (human malignant glioma) cells in 1 mL phosphate-buffered saline were labeled with 10 μL of 100× carboxy-fluorescein diacetate succinimidyl ester (CFSE, Cell Technology Inc., Minneapolis, MN) for 15 minutes at 24°C. The U-87 MG target cells were washed, resuspended in phosphate-buffered saline with 10% FBS, and incubated for 30 minutes at 37°C. Effector cells (microglia or PBMCs) were isolated and added at varying effector ratios to 1 × 10<sup>5</sup>target cells. Cultures were incubated for 4 hours or 24 hours, and 10 μL of 50 μg/mL propidium iodide was added to each culture before analysis by FACSCalibur Instrumentation and CellQuest Pro software (BD Biosciences, San Jose, CA).</p></sec><sec><title>Statistics</title><p>The intracellular cytokine data was analyzed by Exact McNemars test for dichotomous categorical and matched data. The <italic>in vitro </italic>cytotoxicity data were analyzed using the Chi square test. The cutoff for statistical significance was <italic>P </italic>< 0.05.</p></sec></sec><sec><title>Results</title><sec><title>Purity of microglia isolated from human gliomas</title><p>To determine the purity of our GIM isolation, cells were stained with fluorescent-labeled antibodies to CD11b, CD11c, and CD45, and the percentage of cells that were CD11b/c<sup>+</sup>CD45<sup>low </sup>was determined using flow cytometry (Fig. <xref ref-type="fig" rid="F1">1</xref>). We can isolate a CD45<sup>high </sup>population from peripheral blood, however, upon examining tumor homogenates before Ficoll purification we fail to see a CD11b/c+CD45<sup>high </sup>seen in the previous murine studies. We observed the largest population of CD11b/c<sup>+</sup>CD45<sup>+ </sup>cells at the interphase between the 1.03 and 1.065, unlike in the rodent studies where the interphase between the 1.065 and 1.072 gradients was identified to contain the largest fraction of CD11b/c<sup>+</sup>CD45<sup>+ </sup>cells. The expression levels of CD45 and CD11b/c were not significantly different in the cells isolated from each interphase. Thus, in contrast to the rodent purification of microglia/macrophages in which distinct CD11b/c<sup>+</sup>CD45<sup>high </sup>and CD11b/c<sup>+</sup>CD45<sup>low </sup>cells were identified, we did not observe this heterogeneity in any of our GBM or normal brain tumor specimens. Our purified preparation was stained with antibodies to CD3, CD4, CD8, and CD56 and was negative for contaminating cell populations, including lymphocytes and NK cells (data not shown).</p><fig position="float" id="F1"><label>Figure 1</label><caption><p>Isolation of GIM and expression of surface markers demonstrating the purity of the GIM population (CD11b vs. CD45) at each isolated fraction. Each interphase from the second percoll gradient was analyzed for expression of the surface markers CD11b and CD45. The gated cells in the upper right quadrant indicate the percentage of CD45<sup>+ </sup>gated cells that are also positive for CD11b and represent GIM. An autofluorescent population was identified in both normal and GBM tissue samples during flow cytometry analysis and is also consistently present in the respective isotype controls. This population was excluded when calculating the percentages of positive fluorescing cells.</p></caption><graphic xlink:href="1479-5876-4-15-1"/></fig><p>The average yield of microglia isolated from human GBM was 3.2 × 10<sup>6 </sup>cells per gram of tumor (range: 1.8 × 10<sup>5 </sup>- 3.7 × 10<sup>7 </sup>cell/gram; n = 15) but was only 1.4 × 10<sup>5 </sup>cells per gram of normal brain (range 4.6 × 10<sup>4 </sup>- 3 × 10<sup>5</sup>; n = 5). This represented an average, although highly variable, 30% recovery and a 60 fold-enrichment of the microglia. The purity of GIM isolated from human glioma specimens can vary due to disparities in each excised individual tissue sample.</p></sec><sec><title>TLR expression on GIM</title><p>To determine whether GIM expressed the TLRs, we stained our purified GIM with fluorescent-labeled antibodies to TLR-1, -2, -3, and -4. Although TLR-1 was not expressed in significant quantities, TLR-2, -3, and -4 were highly expressed on GIM (n = 5) (Fig. <xref ref-type="fig" rid="F2">2</xref>). The TLR expression profile was no different on microglia isolated from normal brain (n = 3) (data not shown).</p><fig position="float" id="F2"><label>Figure 2</label><caption><p>GIM express TLRs. GIM purity was determined as in Fig. 1 and the cells were double-stained for CD45 and TLRs. All CD45<sup>+ </sup>expressing cells were gated and observed for TLR expression. In each histogram, the data plot on the left indicates the isotype control, and the second plot represents CD45<sup>+ </sup>gated cells that expressed the respective TLR. These data are from one tumor specimen but are representative of at least five tumor tissue samples from human glioma patients.</p></caption><graphic xlink:href="1479-5876-4-15-2"/></fig></sec><sec><title>Cytokine expression by GIM</title><p>To determine if the microglia are activated within malignant gliomas (n = 7) and participating in innate responses, intracellular cytokine analysis was performed for IL-1β, IL-6, and TNF-α. GIM failed to produce any significant levels of these cytokines when compared to control cells (A375) and was no different from microglia isolated from normal brain tissue (n = 3) (Fig. <xref ref-type="fig" rid="F3">3</xref>, Table <xref ref-type="table" rid="T1">1</xref>).</p><fig position="float" id="F3"><label>Figure 3</label><caption><p>GIM do not secret innate cytokines. GIM purity was determined as in Fig. 1 and the cells were stained for intracellular cytokines IFN-α, IL-6, TNF-α. In each histogram, the data plot on the left indicates the isotype control, and the second plot represents respective cytokine producing cells. These data are from one tumor specimen but are representative of seven tumor tissue samples from human glioma patients and A375 control cells.</p></caption><graphic xlink:href="1479-5876-4-15-3"/></fig><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>Distribution of cytokine production by glioma infiltrating microglia from 7 representative human glioma patients. Positive intracellular staining was confirmed for all cytokines by either a reporter cell line (A375).</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left"><bold>Pathology</bold></td><td align="center"><bold>IL-1β </bold></td><td align="center"><bold>IL-6</bold></td><td align="center"><bold>TNF-α </bold></td></tr></thead><tbody><tr><td align="left">GBM</td><td align="center">0%</td><td align="center">0%</td><td align="center">0%</td></tr><tr><td align="left">GBM</td><td align="center">0%</td><td align="center">0%</td><td align="center">0%</td></tr><tr><td align="left">GBM</td><td align="center">0%</td><td align="center">0%</td><td align="center">0%</td></tr><tr><td align="left">GBM</td><td align="center">0%</td><td align="center">0%</td><td align="center">0%</td></tr><tr><td align="left">GBM</td><td align="center">0%</td><td align="center">0%</td><td align="center">0%</td></tr><tr><td align="left">AA</td><td align="center">0%</td><td align="center">0%</td><td align="center">0%</td></tr><tr><td align="left">AA</td><td align="center">0%</td><td align="center">0%</td><td align="center">0%</td></tr><tr><td align="left">Normal</td><td align="center">0%</td><td align="center">0%</td><td align="center">0%</td></tr><tr><td align="left">Normal</td><td align="center">0%</td><td align="center">0%</td><td align="center">0%</td></tr><tr><td align="left">Normal</td><td align="center">0%</td><td align="center">2%</td><td align="center">0%</td></tr><tr><td align="left">A375</td><td align="center">58%</td><td align="center">3%</td><td align="center">5%</td></tr></tbody></table><table-wrap-foot><p>Percentages indicate CD11b<sup>+</sup>CD45<sup>+ </sup>GIM gated on all CD45<sup>+ </sup>cells and are adjusted to that of corresponding isotype controls.</p><p>Abbreviations: IL, interleukin; TNF, tumor necrosis factor; GBM, glioblastoma multiforme; AA, anaplastic astrocytoma.</p></table-wrap-foot></table-wrap></sec><sec><title>GIM are functionally capable of phagocytosis</title><p>To determine whether GIM isolated from malignant gliomas were capable of mediating phagocytosis, GIM were incubated with opsonized fluorescent beads. GIM showed marked intracellular staining by the beads, indicating that phagocytosis was not impaired in GIM despite the immunosuppressive environment of gliomas (Fig. <xref ref-type="fig" rid="F4">4A</xref>). In contrast, control tumor cells bound the beads only on the surface of the cell (Fig. <xref ref-type="fig" rid="F4">4B</xref>).</p><fig position="float" id="F4"><label>Figure 4</label><caption><p>GIM can mediate phagocytosis. Cells were incubated with opsonized fluorescent beads and analyzed using fluorescence microscopy at 20× magnification. <italic>A</italic>, GIM containing phagocytosed beads. <italic>B</italic>, Control tumor cell that has fluorescent beads sticking to the surface but not internalized. Cells were analyzed through a z-stack to determine whether the beads were internalized or merely on the surface.</p></caption><graphic xlink:href="1479-5876-4-15-4"/></fig></sec><sec><title>GIM can mediate non-MHC restricted anti-tumor cellular cytotoxicity</title><p>To ascertain whether GIM can mediate anti-tumor cytotoxicity, freshly isolated GIM were incubated as effectors with U-87 MG target cells (a cell line derived from malignant human glioma) at various effector-to-target ratios. As controls, microglia from normal brain tissue and PBMCs from the patient were incubated with U-87 MG cells in similar ratios. All effector cells were unstimulated before the assay and did not show significant cytotoxic activity after 4 hours (data not shown). After 24 hours, even at a 1:10 effector:target ratio, GIM were functional in their tumor cytotoxic activity (48.3% cells; 95% CI, 46.8-49.9) and comparable to PBMCs (46.9% cells; 95%CI, 45.5-48.2), however cytotoxic activity of GIM was significantly (p < 0.0001) lower than that of microglia isolated from normal brain tissue (72.4% cells; 95% CI, 69.7-75.0)(Fig. <xref ref-type="fig" rid="F5">5</xref>). We were not able to generate sufficient numbers of microglia from normal brain tissue to test higher effector:target ratios.</p><fig position="float" id="F5"><label>Figure 5</label><caption><p>GIM can mediate non-MHC-restricted tumor cytotoxicity. GIM, peripheral blood mononuclear cells, and normal (5 × 10<sup>6</sup>, 1 × 10<sup>6</sup>, or 5 × 10<sup>5</sup>) microglia were each incubated as an effector population with 1 × 10<sup>5 </sup>carboxy-fluorescein diacetate succinimidyl ester-labeled U-87 target cells, for an effector-to-target ratio of 50:1, 10:1, or 5:1, respectively. Propidium iodide was added after 24 hours incubation, and cells were analyzed by flow cytometry. Cells were first gated only on the carboxy-fluorescein diacetate succinimidyl ester<sup>+ </sup>target population (excluding all other cells in the assay), and the propidium iodide expression on these cells was determined. The percentages of cytolytic activity of the effector cells were calculated (<italic>Y </italic>axis)(dead targets in upper right quadrant/[dead targets in upper right quadrant + live targets in lower right quadrant]) and plotted against each respective effector-to-target ratio (<italic>X </italic>axis). These data are from one tumor specimen and are representative of experiments from cells isolated from 3 different glioma patients. Bars at each data point represent 95% confidence intervals for each proportion. After 24 hours, GIM were functional in their tumor cytotoxic activity and comparable to PBMCs, however cytotoxic activity of GIM was significantly (*p < 0.0001) lower than that of normal microglia.</p></caption><graphic xlink:href="1479-5876-4-15-5"/></fig></sec><sec><title>Microglia express fas ligand at low levels</title><p>To ascertain whether human GIM were preventing antitumor T cell activity via Fas-FasL-mediated apoptosis of T cells, the GIM were stained with an anti-FasL antibody. In contrast to the murine findings, human GIM did not express FasL or expressed it at very low levels (n = 11)(Fig. <xref ref-type="fig" rid="F6">6</xref>), indicating that Fas-FasL-mediated apoptosis is not a predominant mechanism of immune evasion by GIM in humans.</p><fig position="float" id="F6"><label>Figure 6</label><caption><p>GIM cannot initiate Fas-mediated apoptosis. The histogram is depicted similar to that in Fig. 2. GIM were gated on CD45<sup>+ </sup>expression and were analyzed for expression of FasL. These data are from one tumor specimen and are representative of at least 11 tumor tissue samples from human glioma patients.</p></caption><graphic xlink:href="1479-5876-4-15-6"/></fig></sec></sec><sec><title>Discussion and conclusion</title><p>This study evaluates for the first time the innate immunologic phenotype and function of human microglia isolated from brain tumors. Activation of innate immunity triggers the subsequent activation of adaptive immune response [<xref ref-type="bibr" rid="B24">24</xref>,<xref ref-type="bibr" rid="B25">25</xref>] that are essential for tumor eradication/suppression. An ongoing debate is whether the immune system can recognize the presence of a tumor. Although it has been argued that tumors fail to provide sufficient pro-inflammatory responses, others have recently suggested otherwise [<xref ref-type="bibr" rid="B26">26</xref>,<xref ref-type="bibr" rid="B27">27</xref>]. Mechanisms that induce innate immune responses similar to microbial pathogens include dying cells producing uric acid [<xref ref-type="bibr" rid="B28">28</xref>]; heat shock proteins [<xref ref-type="bibr" rid="B29">29</xref>-<xref ref-type="bibr" rid="B31">31</xref>]; and extracellular matrix derivatives, such as hyaluronic acid [<xref ref-type="bibr" rid="B32">32</xref>] or heparin sulfates [<xref ref-type="bibr" rid="B33">33</xref>], which are all TLR-4 agonists. All these materials are prodigiously produced by gliomas [<xref ref-type="bibr" rid="B34">34</xref>-<xref ref-type="bibr" rid="B37">37</xref>] indicating that innate immunity can be potentially activated by these glioma factors. Our data shows that similar to the murine system [<xref ref-type="bibr" rid="B16">16</xref>], there is TLR expression on the GIM. The interaction of the microglia with TLR agonists should lead to the induction of a plethora of inflammatory mediators, such as TNF-α, IL-1 and IL-6. These cytokines subsequently induce the local inflammatory response. However, none of these cytokines were produced by the microglia within the immunosuppressive microenvironment of the malignant glioma. It is important to note that, unlike the phagocytosis of pathogens that is followed by the induction of inflammatory mediators, phagocytosis of apoptotic and senescent cells is immunologically silent and does not lead to the induction of inflammatory responses. Thus, despite the elaboration of factors such as heat shock proteins and hyaluronic acid that could activate innate immunity on microglia, the "balance" of micro-environmental influences including apoptotic glioma cells and immunosuppressive cytokines impede innate immune activation as reflected by cytokine production.</p><p>GIM can participate in the initial steps of innate immunity since they are capable of phagocytosis. However, due to the limitations in obtaining normal brain tissue, we cannot discount the possibility that this phagocytic ability of GIM might be impaired when compared to that of microglia isolated from normal brain. Indeed we find that GIM, while capable of non-MHC mediated cytotoxicity, they are significantly less able to participate in tumor cytotoxic activity compared to normal microglia. The direct cytotoxicity of GIM against tumor cells is tempered by the fact that immunosuppressive cytokines/factors that influence the function of GIM are withdrawn for 24 hours. However, the U-87 glioma cell line secretes the immunosuppressive cytokine TGF-β, in part recapitulating the <italic>in vivo </italic>scenario. <italic>In vivo </italic>cytokines such as IL-10 prevent macrophages from becoming tumoricidal by inhibiting the production of TNF and nitric oxide. Because the assay requires 24 hours, the effects of the tumor microenvironment are in part withdrawn while the assay is in progress. Thus, the relative ability of GIM compared to normal microglia to mediate cytotoxicity may be even more profound <italic>in vivo</italic>.</p><p>The role of FasL within innate immunologic responses is not straightforward. For example, when tumors express soluble FasL, immunologic responses appear to be suppressed. However, when FasL is expressed in a membrane-bound form, neutrophil-mediated inflammation is induced, and both the innate and adaptive immune systems are activated [<xref ref-type="bibr" rid="B17">17</xref>]. Contradictory results have demonstrated that high levels of expression of FasL can inactivate neutrophils [<xref ref-type="bibr" rid="B19">19</xref>]. Murine microglia have previously been reported as expressing FasL [<xref ref-type="bibr" rid="B20">20</xref>]. We wanted to determine whether FasL was similarly a potential immunotherapy target in human GIM; however, we did not find any significant expression of FasL on microglia across a wide array of grades of gliomas and metastatic cancers. Therefore, modulating techniques directed at FasL on human GIM are not likely to be successful. However, this does not rule out a potential immunological benefit of modulating FasL on the tumor cells.</p><p>This study is the first to our knowledge to characterize immune activation at the level of the innate arm on microglia in glioma patients. Of course, we cannot exclude the possibilities that activation of the innate immune system could occur on another CNS cells beside microglia or that innate immune activation could occur via draining TLR agonists to the cervical lymph nodes. The activation of innate immunity on GIM through select TLRs may be a future approach to enhancing the incorporation and activation of adaptive immune responses, but a careful screening through the panel of agonists will be necessary because not all TLR agonists may cause activation, especially in the presence of immunosuppressive cytokines/factors. While it is important to further characterize these innate functions and understand the role they play in anti-tumor immune responses, our data appears to demonstrate that some of the innate immune functions remain unaffected despite the immunosuppressive glioma environment and can perhaps be manipulated to develop effective immunotherapeutic responses to gliomas.</p></sec><sec><title>Competing interests</title><p>The author(s) declare that they have no competing interests.</p></sec><sec><title>Authors' contributions</title><p>SFH participated in the design of the study, carried out the flow cytometry, phagocytosis and cytotoxicity assays and helped draft the manuscript, DY carried out the microglia purification from tumor specimens and provided technical assistance for the assays, DS carried out the statistical analyses, EG provided input in the design of the study and ABH conceived of the study, participated in the design and coordination and helped draft the manuscript.</p></sec> |
Association between adolescent idiopathic scoliosis prevalence and age at menarche in different geographic latitudes | <sec><title>Background</title><p>Age at menarche is considered a reliable prognostic factor for idiopathic scoliosis and varies in different geographic latitudes. Adolescent idiopathic scoliosis prevalence has also been reported to be different in various latitudes and demonstrates higher values in northern countries. A study on epidemiological reports from the literature was conducted to investigate a possible association between prevalence of adolescent idiopathic scoliosis and age at menarche among normal girls in various geographic latitudes. An attempt is also made to implicate a possible role of melatonin in the above association.</p></sec><sec><title>Material-methods</title><p>20 peer-reviewed published papers reporting adolescent idiopathic scoliosis prevalence and 33 peer-reviewed papers reporting age at menarche in normal girls from most geographic areas of the northern hemisphere were retrieved from the literature. The geographic latitude of each centre where a particular study was originated was documented. The statistical analysis included regression of the adolescent idiopathic scoliosis prevalence and age at menarche by latitude.</p></sec><sec><title>Results</title><p>The regression of prevalence of adolescent idiopathic scoliosis and age at menarche by latitude is statistically significant (p < 0.001) and are following a parallel declining course of their regression curves, especially in latitudes northern than 25 degrees.</p></sec><sec><title>Conclusion</title><p>Late age at menarche is parallel with higher prevalence of adolescent idiopathic scoliosis. Pubarche appears later in girls that live in northern latitudes and thus prolongs the period of spine vulnerability while other pre-existing or aetiological factors are contributing to the development of adolescent idiopathic scoliosis. A possible role of geography in the pathogenesis of idiopathic scoliosis is discussed, as it appears that latitude which differentiates the sunlight influences melatonin secretion and modifies age at menarche, which is associated to the prevalence of idiopathic scoliosis.</p></sec> | <contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Grivas</surname><given-names>Theodoros B</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>grivastb@panafonet.gr</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Vasiliadis</surname><given-names>Elias</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>nadel@otenet.gr</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Mouzakis</surname><given-names>Vasilios</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>bmouzakis@donti.gr</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Mihas</surname><given-names>Constantinos</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>gas521@yahoo.co.uk</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Koufopoulos</surname><given-names>Georgios</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>gkfpls@yahoo.com</email></contrib> | Scoliosis | <sec><title>Background</title><p>A wide range of Adolescent Idiopathic Scoliosis (AIS) prevalence in different countries is demonstrated by the various reports in the literature. (1–20) The significance of this specific observation may not be obvious but its evaluation is important because it could be related to a possible contributory factor of AIS pathogenesis.</p><p>In studying variations on the rate of sexual development across the world, a similar observation is recorded for the age at menarche, as well. (21–53) The influence of the geography of a specific region on human biology is determined by socioeconomic and environmental factors such as temperature, humidity and lighting that are transferred and expressed in human cells by specific mediators. (54) Age at menarche is definitely a biologic event and is considered a reliable prognostic factor of AIS. (55, 56)</p><p>The aim of this report is the study of IS prevalence and age at menarche as it is reported on published papers from different countries in various geographic latitudes and the investigation of a possible association between them that may reveal their possible role on AIS pathogenesis.</p></sec><sec><title>Methods and material</title><p>The inclusion criteria for the epidemiological studies were clearly defined before performing a search of the literature on scoliosis prevalence and on age at menarche by browsing the Medline database. The geographic latitude of each centre where a particular study was originated was documented. The included studies cover the whole spectrum of geographic latitudes in the northern hemisphere.</p><sec><title>The scoliosis prevalence</title><p>A paper was considered eligible for inclusion when the study was reporting the prevalence of AIS among normal girls, was age matched, involving girls between 10 and 14 years old, the curves were detected through screening programs and a cut-off point of 10 degrees of Cobb angle was used for AIS definition. Twenty peer-reviewed papers met those criteria and were included in the study [<xref ref-type="bibr" rid="B1">1</xref>-<xref ref-type="bibr" rid="B20">20</xref>].</p></sec><sec><title>The age at menarche</title><p>Thirty three peer-reviewed papers reporting on age at menarche in normal girls at certain geographic regions were found and were included in the study [<xref ref-type="bibr" rid="B21">21</xref>-<xref ref-type="bibr" rid="B53">53</xref>]. All the included papers are reporting the age at menarche of normal girls and not of specific groups.</p></sec><sec><title>The statistical analysis</title><p>The prevalence of idiopathic scoliosis and age at menarche were treated as the dependent variables in a linear regression forward modeling procedure. The geographic latitude of each location where the study had taken place was the candidate independent variable. Because of the different sample size of each study, frequency weights were used in the regression model, controlling for the impact of each latitude value, according to their sample size. The <italic>F</italic>-test of significance of overall regression and Type I partial <italic>F</italic>-test were calculated at a significance level less than 0.05, testing for the significance of overall regression and for each variable added during the modeling. A graphical analysis of the residuals of the regression was performed in order to detect potential problems with the model. Data were analysed using STATA™ (Version 8.0, Stata Corporation, College station, TX 77845, 800-782-8272).</p></sec></sec><sec><title>Results</title><sec><title>The scoliosis prevalence</title><p>The reported prevalence of AIS in the literature increases in the northern geographic latitudes and decreases as the latitude is approaching the equator (Table <xref ref-type="table" rid="T1">1</xref>, Figure <xref ref-type="fig" rid="F1">1</xref>).</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>Demonstrates data on AIS prevalence (%) according to the city or geographical periphery that the epidemiological study was performed, the size of the sample of the examined girls, the geographic latitude in degrees and the author and the year of publication.</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="center"><bold>City or Geographical Periphery</bold></td><td align="center"><bold>Geographic Latitude in degrees</bold></td><td align="center"><bold>No of girls screened</bold></td><td align="center"><bold>Prevalence of IS (%)</bold></td><td align="center"><bold>Author and year of publication</bold></td></tr></thead><tbody><tr><td align="left">Helsinki (Finland)</td><td align="center">65,00</td><td align="center">401</td><td align="center"><bold>12,0</bold></td><td align="center">Nissinen M et al 1993 [1]</td></tr><tr><td align="left">Malmo (Sweden)</td><td align="center">57,50</td><td align="center">8469</td><td align="center"><bold>3,21</bold></td><td align="center">Willner S & Uden A 1982 [2]</td></tr><tr><td align="left">Esbjerg (Denmark)</td><td align="center">55,30</td><td align="center">1000</td><td align="center"><bold>4,10</bold></td><td align="center">Laulund T et al 1982 [3]</td></tr><tr><td align="left">Oxford (UK)</td><td align="center">52,50</td><td align="center">2613</td><td align="center"><bold>3,20</bold></td><td align="center">Dickson RA et al 1983 [4]</td></tr><tr><td align="left">Nottingham(UK)</td><td align="center">52,00</td><td align="center">321</td><td align="center"><bold>1,90</bold></td><td align="center">Burwell RG et al 1983 [5]</td></tr><tr><td align="left">Quebec (Canada)</td><td align="center">47,50</td><td align="center">14701</td><td align="center"><bold>2,37</bold></td><td align="center">Morais Tet al 1985 [6]</td></tr><tr><td align="left">Rochester (Minn, USA)</td><td align="center">47,50</td><td align="center">1212</td><td align="center"><bold>2,70</bold></td><td align="center">Yawn BP et al 1999 [7]</td></tr><tr><td align="left">Wisconsin (USA)</td><td align="center">47,50</td><td align="center">7462</td><td align="center"><bold>2,03</bold></td><td align="center">Gore DR et al 1981 [8]</td></tr><tr><td align="left">Slovenia</td><td align="center">45,00</td><td align="center">70200</td><td align="center"><bold>2,89</bold></td><td align="center">Plenicar-Cucek M et al 1995 [9]</td></tr><tr><td align="left">Montreal (Canada)</td><td align="center">45,30</td><td align="center">13500</td><td align="center"><bold>2,70</bold></td><td align="center">Rogala EJ et al 1978 [10]</td></tr><tr><td align="left">Cape Cod (USA)</td><td align="center">42,00</td><td align="center">928</td><td align="center"><bold>5,00</bold></td><td align="center">Strayer LM et al, 1973 [11]</td></tr><tr><td align="left">Delaware (USA)</td><td align="center">38,50</td><td align="center">25550</td><td align="center"><bold>2,80</bold></td><td align="center">Shands AR et al 1955 [12]</td></tr><tr><td align="left">Epirus (Greece)</td><td align="center">38,00</td><td align="center">40962</td><td align="center"><bold>2,60</bold></td><td align="center">Soucacos PN et al 1997 [13]</td></tr><tr><td align="left">California (USA)</td><td align="center">36,00</td><td align="center">1940</td><td align="center"><bold>7,70</bold></td><td align="center">Brooks HL et al 1975 [14]</td></tr><tr><td align="left">Crete (Greece)</td><td align="center">35,00</td><td align="center">10278</td><td align="center"><bold>1,80</bold></td><td align="center">Koukourakis I et al 1997 [15]</td></tr><tr><td align="left">Wakayaka (Japan)</td><td align="center">34,00</td><td align="center">1702</td><td align="center"><bold>4,90</bold></td><td align="center">Sugita K 2000 [16]</td></tr><tr><td align="left">Hu Guan (China)</td><td align="center">25,30</td><td align="center">12000</td><td align="center"><bold>1,60</bold></td><td align="center">Ma X et al 1995 [17]</td></tr><tr><td align="left">Changsha (China)</td><td align="center">28,00</td><td align="center">3963</td><td align="center"><bold>2,45</bold></td><td align="center">Pin LH et al 1985 [18]</td></tr><tr><td align="left">Taipei (Taiwan)</td><td align="center">27,50</td><td align="center">33596</td><td align="center"><bold>1,00</bold></td><td align="center">Huang SC 1997 [19]</td></tr><tr><td align="left">Singapoure</td><td align="center">5,00</td><td align="center">37141</td><td align="center"><bold>0,93</bold></td><td align="center">Wong HK et al 2005 [20]</td></tr></tbody></table></table-wrap><fig position="float" id="F1"><label>Figure 1</label><caption><p>Frequency weighted linear regression plot. The circle size represents the sample size of each study.</p></caption><graphic xlink:href="1748-7161-1-9-1"/></fig><p>The final regression model is shown in Table <xref ref-type="table" rid="T2">2</xref>. The inclusion of the quadratic term of latitude contributed to the explanatory power of the model according to partial <italic>F </italic>test and overall <italic>F </italic>test. According to the modelling of the data, a significant positive association between prevalence of IS and latitude was found (overall F p-value < 0.001), following a rather curvilinear trend (Figure <xref ref-type="fig" rid="F1">1</xref>).</p><table-wrap position="float" id="T2"><label>Table 2</label><caption><p>Linear regression results of AIS prevalence by the geographic latitude. The inclusion of the (Latitude)<sup>2</sup> term implies a possible quadratic relationship.</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left"><bold>Number of obs=</bold></td><td align="center">287939</td><td align="left"><bold>R-squared=</bold></td><td align="center">0.509</td><td></td></tr><tr><td align="left"><bold><italic>F (</italic></bold><bold>2,288036)=</bold></td><td align="center">149257.39</td><td align="left"><bold>Adj R-squared=</bold></td><td align="center">0.509</td><td></td></tr><tr><td align="left"><bold>(Prob ></bold><bold><italic>F</italic></bold><bold>)</bold></td><td align="center"><0.001</td><td align="left"><bold>Root MSE=</bold></td><td align="center">0.7122</td><td></td></tr></thead><tbody><tr><td align="center"><bold>Prevalence</bold></td><td align="center"><bold>Unstandardized coefficient (B)</bold></td><td align="center"><bold>P > t</bold></td><td align="center"><bold>95% Conf.</bold></td><td align="center"><bold>Interval</bold></td></tr><tr><td align="left">Latitude</td><td align="center">0.024758</td><td align="center">< 0.001</td><td align="center">0.024045</td><td align="center">0.0254713</td></tr><tr><td align="left">Latitude^2</td><td align="center">0.000484</td><td align="center"><0.001</td><td align="center">0.000472</td><td align="center">0.0004965</td></tr><tr><td align="left">Constant (Intercept)</td><td align="center">0.712796</td><td align="center"><0.001</td><td align="center">0.703129</td><td align="center">0.7224631</td></tr></tbody></table></table-wrap></sec><sec><title>The age at menarch</title><p>Age at menarche shows a decreasing trend as the geographic latitude approaches approximately the 25–30 degrees and then increases again toward 0 degrees (near the equator) (Table <xref ref-type="table" rid="T3">3</xref>, Figure <xref ref-type="fig" rid="F2">2a</xref>).</p><table-wrap position="float" id="T3"><label>Table 3</label><caption><p>Demonstrates the city or geographical periphery where each study was performed, the geographic latitude in degrees, the age at menarche in years, and the author and year of publication.</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="center"><bold>City or Geographical Periphery</bold></td><td align="center"><bold>Geographic Latitude in degrees</bold></td><td align="center"><bold>Age at menarche in years</bold></td><td align="center"><bold>Author and year of publication</bold></td></tr></thead><tbody><tr><td align="left">Oslo (Norway)</td><td align="center">59,93</td><td align="center"><bold>16,00</bold></td><td align="center">Rosenberg 1991 [21]</td></tr><tr><td align="left">Copenhagen (Denmark)</td><td align="center">55,43</td><td align="center"><bold>13,40</bold></td><td align="center">Helm P & Grolund L 1998 [22]</td></tr><tr><td align="left">Rotterdam (Netherlands)</td><td align="center">51,55</td><td align="center"><bold>13,15</bold></td><td align="center">Mul et al 2001 [23]</td></tr><tr><td align="left">Brussels (Belgium)</td><td align="center">50,51</td><td align="center"><bold>13,00</bold></td><td align="center">Vercauteren M & Susanne C 1984 [24]</td></tr><tr><td align="left">Switzerland</td><td align="center">47,00</td><td align="center"><bold>13,40</bold></td><td align="center">Largo RH & Prader A 1983 [25]</td></tr><tr><td align="left">Pecs (Hungary)</td><td align="center">46,00</td><td align="center"><bold>12,90</bold></td><td align="center">Dober I & Kiralyfalvi L 1993 [26]</td></tr><tr><td align="left">Zagreb (Croatia)</td><td align="center">45,48</td><td align="center"><bold>13,13</bold></td><td align="center">Prebeg Z & Bralic I 2000 [27]</td></tr><tr><td align="left">L' Aquila (Italy)</td><td align="center">42,36</td><td align="center"><bold>12,55</bold></td><td align="center">Danubio et al 2004 [28]</td></tr><tr><td align="left">Madrid (Spain)</td><td align="center">40,26</td><td align="center"><bold>12,79</bold></td><td align="center">Marroban MD, Mesa MS 2000 [29]</td></tr><tr><td align="left">Coimbra (Portugal)</td><td align="center">40,15</td><td align="center"><bold>12,53</bold></td><td align="center">Padez C & Rocha MA 2003 [30]</td></tr><tr><td align="left">Dayton (Ohio, USA)</td><td align="center">40,13</td><td align="center"><bold>12,43</bold></td><td align="center">Chumlea et al 2003 [31]</td></tr><tr><td align="left">Sardinia (Italy)</td><td align="center">40,00</td><td align="center"><bold>12,78</bold></td><td align="center">Floris G et al 1987 [32]</td></tr><tr><td align="left">Athens (Greece)</td><td align="center">38,00</td><td align="center"><bold>12,58</bold></td><td align="center">Dacou-Voutetakis et al 1983 [33]</td></tr><tr><td align="left">Thriasion Pedion, (Greece)</td><td align="center">37,50</td><td align="center"><bold>12,07</bold></td><td align="center">Grivas et al 2002 [34]</td></tr><tr><td align="left">Columbia (USA)</td><td align="center">38,53</td><td align="center"><bold>13,20</bold></td><td align="center">Vadocz EA et al 2002 [35]</td></tr><tr><td align="left">Atlanta (Georgia, USA)</td><td align="center">33,46</td><td align="center"><bold>12,60</bold></td><td align="center">Freedman et al 2002 [36]</td></tr><tr><td align="left">Marrakech (Morocco)</td><td align="center">31,49</td><td align="center"><bold>13,04</bold></td><td align="center">Montero P et al 1999 [37]</td></tr><tr><td align="left">Texas (USA)</td><td align="center">31,14</td><td align="center"><bold>13,10</bold></td><td align="center">Malina RM et al 1994 [38]</td></tr><tr><td align="left">Chandigarh (India)</td><td align="center">30,72</td><td align="center"><bold>13,20</bold></td><td align="center">Sharma K 1990 [39]</td></tr><tr><td align="left">Patiala (India)</td><td align="center">30,35</td><td align="center"><bold>12,54</bold></td><td align="center">Singh SP & Malhotra P 1988 [40]</td></tr><tr><td align="left">Cairo (Egypt)</td><td align="center">30,05</td><td align="center"><bold>12,59</bold></td><td align="center">Attallah NL 1978 [41]</td></tr><tr><td align="left">Shiraz (Iran)</td><td align="center">29,38</td><td align="center"><bold>12,84</bold></td><td align="center">Ayatollahi SM et al 2002 [42]</td></tr><tr><td align="left">Zhejiang (China)</td><td align="center">28,45</td><td align="center"><bold>12,80</bold></td><td align="center">Hesketh T et al 2002 [43]</td></tr><tr><td align="left">Hong Kong (China)</td><td align="center">22,45</td><td align="center"><bold>12,38</bold></td><td align="center">Huen KF et al 1997 [44]</td></tr><tr><td align="left">Dominican Republic</td><td align="center">19,00</td><td align="center"><bold>13,10</bold></td><td align="center">Mancebo P et al 1990 [45]</td></tr><tr><td align="left">Khartoum (Sudan)</td><td align="center">15,55</td><td align="center"><bold>13,35</bold></td><td align="center">Attallah NL et al 1983 [46]</td></tr><tr><td align="left">Bangkok (Thailand)</td><td align="center">13,73</td><td align="center"><bold>12,35</bold></td><td align="center">Chompootaweep S et al 1997 [47]</td></tr><tr><td align="left">Hausa (Nigeria)</td><td align="center">12,24</td><td align="center"><bold>13,50</bold></td><td align="center">Rehan N 1994 [48]</td></tr><tr><td align="left">Carabobo (Venezuela)</td><td align="center">10,23</td><td align="center"><bold>12,86</bold></td><td align="center">Farid-Coupal N et al 1981 [49]</td></tr><tr><td align="left">Nigeria</td><td align="center">10,00</td><td align="center"><bold>13,70</bold></td><td align="center">Oduntan SO et al 1976 [50]</td></tr><tr><td align="left">Jaffna (Sri Lanka)</td><td align="center">9,66</td><td align="center"><bold>13,78</bold></td><td align="center">Prakash S & Pathmanathan G 1984 [51]</td></tr><tr><td align="left">Kumasi (Ghana)</td><td align="center">6,75</td><td align="center"><bold>13,98</bold></td><td align="center">Adadevoh SW et al 1989 [52]</td></tr><tr><td align="left">Yaunde (Cameroon)</td><td align="center">3,85</td><td align="center"><bold>13,98</bold></td><td align="center">Pasquet P et al 1999 [53]</td></tr></tbody></table></table-wrap><fig position="float" id="F2"><label>Figure 2</label><caption><p><bold>A</bold>. Illustration of the regression of age at menarche (in years) by latitude (in degrees) (p < 0.001). <bold>B</bold>. The linear correlation between age at menarche and geographic latitude is better shown in the figure where all observations below latitude of 25 degrees are not described. </p></caption><graphic xlink:href="1748-7161-1-9-2"/></fig><p>The linear correlation between age at menarche and geographic latitude is better shown when all observations of lower latitude than 25 degrees are not described, (Figure <xref ref-type="fig" rid="F2">2b</xref>).</p><p>The age at menarche in healthy population regressed by latitude showed that there is a statistical significant correlation, (p < 0.001) (Table <xref ref-type="table" rid="T4">4</xref>).</p><table-wrap position="float" id="T4"><label>Table 4</label><caption><p>Linear regression results of age at menarche by the geographic latitude.</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left"><bold>Number of obs=</bold></td><td align="center">68763</td><td align="left"><bold>R-squared=</bold></td><td align="center">0.342</td><td></td></tr><tr><td align="left"><bold><italic>F (</italic></bold><bold>2,288036)=</bold></td><td align="center">17855.41</td><td align="left"><bold>Adj R-squared=</bold></td><td align="center">0.342</td><td></td></tr><tr><td align="left"><bold>(Prob ></bold><bold><italic>F</italic></bold><bold>)</bold></td><td align="center"><0.001</td><td align="left"><bold>Root MSE=</bold></td><td align="center">0.409</td><td></td></tr></thead><tbody><tr><td align="center"><bold>Age at menarche</bold></td><td align="center"><bold>Unstandardized coefficient (B)</bold></td><td align="center"><bold>P > t</bold></td><td align="center"><bold>95% Conf.</bold></td><td align="center"><bold>Interval</bold></td></tr><tr><td align="left">Latitude</td><td align="center">-0.1123787</td><td align="center"><0.001</td><td align="center">-0.1135448</td><td align="center">-0.1112126</td></tr><tr><td align="left">Latitude^2</td><td align="center">0.0018579</td><td align="center"><0.001</td><td align="center">0.0018381</td><td align="center">0.0018776</td></tr><tr><td align="left">Constant (Intercept)</td><td align="center">14.07617</td><td align="center"><0.001</td><td align="center">14.06203</td><td align="center">14.09031</td></tr></tbody></table></table-wrap></sec></sec><sec><title>Discussion</title><p>The regression curves of AIS prevalence and age at menarche by latitude are following a parallel decrement, especially in latitudes northern than 30°, as it is shown in figures <xref ref-type="fig" rid="F1">1</xref> and <xref ref-type="fig" rid="F2">2</xref> respectively.</p><sec><title>The prevalence of AIS</title><p>On reviewing the literature, prevalence figures quoted by various authors, serve to emphasize an apparent divergence in different parts of the world. The figures may be reflecting differences in the definition of a scoliotic curve, the methods of clinical examination, the thresholds for referral, the age group screened and as to whether the studies are based on random sampling or a longitudinal survey of individual children over some years. On the other hand recorded figures may represent real environmental, geographical, genetic or racial influences [<xref ref-type="bibr" rid="B54">54</xref>].</p><p>In the present study the AIS prevalence decreases as the geographic latitude approaches the equator. The parallel decrement of their regression curves in northern hemisphere implicates the role of numerous factors related to latitude in the pathogenesis of AIS.</p></sec><sec><title>Geographic latitude and age at menarche</title><p>Although the age at menarche is to some extent influenced by family heredity, body weight, photic input and season, it seems more susceptible to modification by certain socioeconomic level and by specific disorders (such as diabetes, obesity and blindness) [<xref ref-type="bibr" rid="B55">55</xref>].</p><p>The influence of the geography should be distinguished between the effects of actual geographic factors (latitude, longitude, altitude, humidity and lighting) and those of the socioeconomic circumstances [<xref ref-type="bibr" rid="B55">55</xref>].</p><p>According to the critical weight hypothesis [<xref ref-type="bibr" rid="B56">56</xref>], there is a cut-off level for Body Mass Index (BMI) in relation to pubertal development. Beyond such degree of weight or BMI, there is no influence on age at menarche. The improved nutritional status among black girls in the recent years is implicated for an 8-month decrease of the median menarcheal age in black girls in a 20-year period, in contrast to a smaller decrease of 2 months in white girls at Bogalusa, a semi rural community near New Orleans, in the United States [<xref ref-type="bibr" rid="B36">36</xref>]. A stabilization of a previously decreasing trend of menarcheal age was recorded in studies from Norway [<xref ref-type="bibr" rid="B57">57</xref>] and from the Netherlands, [<xref ref-type="bibr" rid="B23">23</xref>] signifying the role of the nutritional status in sexual maturation.</p><p>Racial differences may also exist in other characteristics that have been suggested to influence pubertal development, such as the secretion of hormones by the hypothalamus, anterior pituitary, and ovary [<xref ref-type="bibr" rid="B58">58</xref>] or the social stress [<xref ref-type="bibr" rid="B59">59</xref>]. Recent studies on leptin, a protein which appears at higher levels among black girls [<xref ref-type="bibr" rid="B60">60</xref>] have suggested that it could act as a link between fat tissue and the central activation of the hypothalamus [<xref ref-type="bibr" rid="B61">61</xref>].</p><p>Earlier reports have not supported the belief once widely held, that sexual development occurs at an earlier age in the tropics than in temperate zones. It was reported that climate in itself has little or no effect on menarche [<xref ref-type="bibr" rid="B55">55</xref>,<xref ref-type="bibr" rid="B62">62</xref>].</p><p>In countries with geographic latitude less than 30° there is different climate. In this region, the so called climatologically moderately favourable belt, the sunshine is estimated at 2 500 h/year [<xref ref-type="bibr" rid="B63">63</xref>].</p><p>In this area apart from the different climate, there is a low socioeconomic status that critically influences the anthropometric dimensions of adolescents. Poverty and children malnutrition results in delay of skeletal maturation. This delay reflects a situation in which the environmental conditions, in terms of nutrition, do not allow the child to reach the optimal genetic potential. Thus the reported age at menarche from those countries is rather confusing. Additionally, there are no reports on AIS prevalence from those countries. A possible association between age at menarche and AIS prevalence in geographic latitudes less than 30° is therefore not realistic.</p></sec><sec><title>Light and age at menarche</title><p>The effect of light on human biology is an issue that has rather not received much attention.</p><p>Environmental lighting exerts important effects on the age at which sexual maturation occurs in birds, [<xref ref-type="bibr" rid="B64">64</xref>,<xref ref-type="bibr" rid="B65">65</xref>] and in monestrous mammals and polyestrous rodents [<xref ref-type="bibr" rid="B66">66</xref>].</p><p>A retinal response to environmental lighting mediates an expanding list of neuroendocrine effects, including control of pubescence, ovulation, and a large number of daily rhythms [<xref ref-type="bibr" rid="B67">67</xref>,<xref ref-type="bibr" rid="B68">68</xref>].</p><p>The pattern of decrease of age at menarche with the shown specific trend as the geographic latitude approaches approximately down to 30 northern degrees could be attributed to daylight and cloudiness duration. The amount of sunlight and the quality of light (degrading relative irradiance and wavelength) may play a major role for the different initiation of menses in above-mentioned latitudes. It is useful for our study to cite the world distribution of solar radiation.</p></sec><sec><title>World distribution of solar radiation and quality of light [<xref ref-type="bibr" rid="B63">63</xref>]</title><p>Solar radiation is unevenly distributed throughout the world because of such variables as solar altitude, which is associated with latitude and season, and atmospheric conditions, which are determined by cloud coverage and degree of pollution. The reported guidelines for the broad identification of the geographic areas with favourable solar energy conditions in the Northern Hemisphere based on the collection of the direct component of sunlight are described below. Similar conditions apply for the Southern Hemisphere [<xref ref-type="bibr" rid="B69">69</xref>].</p><p>The most favourable belt is (15–35°N). <italic>It has over 3000 h/year of sunshine and limited cloud coverage. </italic>More than 90% of the incident solar radiation comes as direct radiation.</p><p>The moderately favourable belt (0–15°N), or equatorial belt, has high atmospheric humidity and cloudiness that tend to increase the proportion of the scattered radiation. The global solar intensity is almost uniform throughout the year as the seasonal variations are only slight. <italic>Sunshine is estimated at 2 500 h/year</italic>.</p><p>In the less favourable belt (35–45°N), the scattering of the solar radiation is significantly increased because of the higher latitudes and lower solar altitude. In addition, cloudiness and atmospheric pollution are important factors that tend to reduce sharply the solar radiation intensity. However, regions beyond 45°N have less favourable conditions for the use of direct solar radiation. This is because almost half of it is in the form of scattered radiation, which is more difficult to collect for use [<xref ref-type="bibr" rid="B63">63</xref>]. As it is also reported the ultraviolet irradiance is different in different latitudes, and a regression model of the form: lnE = a<sub>0 </sub>+ a<sub>1</sub>cosΘ + a<sub>2</sub>Z, (where E is the seasonally integrated irradiance, Θ is the latitude and Z is the elevation of the observing site above sea level) explains 98.4% of the variance in the specific data sets of the related report [<xref ref-type="bibr" rid="B70">70</xref>]. The variation in human illumination exposure at different latitudes has also been elsewhere reported [<xref ref-type="bibr" rid="B71">71</xref>].</p><p>Moreover the different quality of light (wavelength) is probably responsible for the later age at menarche presented by girls living in higher altitudes, as it is reported in the available literature [<xref ref-type="bibr" rid="B39">39</xref>,<xref ref-type="bibr" rid="B72">72</xref>], affecting probably the melatonin rhythm. This issue is in accordance with the reported increase in 6-Hydroxymelatonin excretion in humans during ascent to high altitudes [<xref ref-type="bibr" rid="B73">73</xref>]</p></sec><sec><title>Melatonin and age at menarche</title><p>At the onset of puberty, the hypothalamus, after being quiescent, resumes a marked pulsatile secretion of Gonadotrophine Releasing Hormone (GnRH), leading to an increased secretion of pituitary gonadotropines (Luteinizing Hormone [LH], and Follicle Stimulating Hormone [FSH]), especially during night, which in turn stimulates the gonadal functions. Cerebral adrenergic and/or dopamine neurotransmitters, endogenous opioids, and melatonin from the pineal gland are some of the neuroendocrine factors thought to be involved in the onset of puberty [<xref ref-type="bibr" rid="B74">74</xref>]</p><p>Melatonin is a hormone derived from the amino acid tryptophan and is secreted by the pineal gland. The enzyme Hydroxy-Indolo-O-Methyl-Transferase (HIOMT) converts the N-acetyl-serotonin to melatonin (Figure <xref ref-type="fig" rid="F3">3</xref>).</p><fig position="float" id="F3"><label>Figure 3</label><caption><p>Biosynthesis of melatonin.</p></caption><graphic xlink:href="1748-7161-1-9-3"/></fig><p>Melatonin production is stimulated by darkness. The lack of light on the retina, through the optic nerve and the pre-ganglionic sympathetic fibres of the upper cervical ganglion of the sympathetic trunk reaches firstly the centre of vision and then through the post-ganglionic fibres reaches the pineal gland and provokes the release of norepinephrine from the end plates of the sympathetic system. Norepinephrine mediates the entrance of tryptophan into the pineal gland and controls the activity of many enzymes and mainly that of HIOMT, which is important for the synthesis of melatonin [<xref ref-type="bibr" rid="B75">75</xref>]. The pineal gland concentration of HIOMT reduces during day time and increases during night time. Thus the diurnal variation of melatonin is due to environmental light conditions. Consequently darkness leads to melatonin over production and light reduces the melatonin production [<xref ref-type="bibr" rid="B76">76</xref>,<xref ref-type="bibr" rid="B77">77</xref>]. Melatonin is the main mediator that transfers the changes of the environmental light to the human cells [<xref ref-type="bibr" rid="B78">78</xref>].</p><p>Melatonin acts on the gonads indirectly, reducing the secretion of gonadotropines and mainly that of LH. This finding may explains the inhibition of ovulation in the Eskimos during the months of winter night period, and the increased rates of melatonin in normal women during night- time.</p><p>A controversy exists whether blind girls with no light stimuli experience delayed puberty. Zacharias and Wurtman in 1964 reported earlier age at menarche of girls suffering retroretineal fibroplasia than normal girls [<xref ref-type="bibr" rid="B77">77</xref>], while Thomas and Pizzarello in 1967 reported no difference [<xref ref-type="bibr" rid="B79">79</xref>] and Segos in 1995 reported that blind girls present late age at menarche [<xref ref-type="bibr" rid="B80">80</xref>] Jafarey et al in 1970 and 1971 reported that artificial lighting results in decreasing of age at menarche [<xref ref-type="bibr" rid="B81">81</xref>,<xref ref-type="bibr" rid="B82">82</xref>]. A more thorough research is required in this issue.</p><p>Mean night time serum melatonin concentration presents an increased peak value at 1–3 yr of age (329.5 ± 42.0 pg/mL), and lessens thereafter, averaging 62.5 ± 9.0 pg/mL in individuals aged 15–20 yr and 29.2 ± 6.1 pg/mL in old age (70–90 yr) [<xref ref-type="bibr" rid="B83">83</xref>] The decrease in nocturnal serum melatonin in children and adolescents correlated with body weight and body surface area, whereas no such correlation was found at a later age [<xref ref-type="bibr" rid="B83">83</xref>].</p><p>There are numerous limitations to interpret studies of melatonin in human subjects because of methodological considerations, such as the use of single blood samples collected during the day or the night, failure to include age related characteristics of melatonin secretion, lack of control of the actual duration and intensity of light exposure, and use of broad clinical features without hormonal markers to define puberty [<xref ref-type="bibr" rid="B84">84</xref>].</p><p>Sexual maturation can be delayed in experimental animals by exogenous melatonin administration or by short day exposure [<xref ref-type="bibr" rid="B85">85</xref>,<xref ref-type="bibr" rid="B88">88</xref>]. A rapid decrease in melatonin has also been observed during successful treatment of patients with delayed puberty [<xref ref-type="bibr" rid="B85">85</xref>].</p></sec><sec><title>AIS and age at menarche</title><p>In this study, two linear regression procedures conducted, one between the prevalence of idiopathic scoliosis and the geographic latitude, and the other between the age at menarche and the geographic latitude. The reader could ask the reasonable question why a direct correlation between the prevalence of idiopathic scoliosis and the age at menarche was not performed? The answer is that due to the study design, the data were collected from two different groups of publications, the first consisting of publications regarding the prevalence of scoliosis whereas the second describing the different age at menarche in various geographic latitudes. As a result, the statistical analysis was done with two different samples. Since the data for the prevalence of IS and the age at menarche referred to different latitudes, a direct correlation could not be carried out. Moreover, no articles which had data for age at menarche and the prevalence of IS for a specific latitude could be found. At long last the results of this study suggest a possible correlation between age at menarche and prevalence of IS based on statistical analysis.</p><p>AIS is associated with pubertal growth spurt and its progression decelerates after completion of skeletal maturity. The age at onset of menarche is indicative of the remaining growth potential of girls. Late onset of menses correlates with delayed skeletal maturity and it implies that there is a potential for progression of a scoliotic curve. In a scoliotic girl, it is very important to predict the evolution of the curve and to advise accordingly the patient and her family. Goldberg et al [<xref ref-type="bibr" rid="B86">86</xref>,<xref ref-type="bibr" rid="B87">87</xref>] have shown that menarcheal status is a more meaningful predictor of curve stability than Risser sign, because at menarche peak growth velocity is already past [<xref ref-type="bibr" rid="B88">88</xref>] and although Risser stage 1 is still, on average, 8 months away [<xref ref-type="bibr" rid="B89">89</xref>,<xref ref-type="bibr" rid="B90">90</xref>], the possibility for more growth and significant progression, especially for smaller curves is declining rapidly. Furthermore, girls who are pre-menarcheal at diagnosis have a higher prevalence of surgery, as the menarcheal status alone will divide a female patient group into those who are at significant risk of surgery and those who are not [<xref ref-type="bibr" rid="B91">91</xref>]. A clear association exists between the deterioration of a scoliotic curve and periods of rapid growth, such those occurring before pubarche.</p></sec><sec><title>Melatonin and AIS</title><p>The role of melatonin deficiency in AIS pathogenesis has been proposed by Machida et al [<xref ref-type="bibr" rid="B92">92</xref>] who produced scoliosis similar to those of human adolescent idiopathic scoliosis in pinealectomized chickens. When pinealectomized chickens were administered melatonin, the scoliosis was prevented. They proposed that a defect in melatonin production might be related to the aetiology of human idiopathic scoliosis.</p><p>There is a controversy whether lower animal models are appropriate for studying scoliosis. Chickens present extrapineal sites of melatonin production [<xref ref-type="bibr" rid="B93">93</xref>,<xref ref-type="bibr" rid="B94">94</xref>] that contribute to circulating melatonin levels, in contrast to humans that no extrapineal sources affect the circadian rhythm of melatonin [<xref ref-type="bibr" rid="B95">95</xref>]. There is a loss of the nighttime melatonin peak and a drop in basal levels below detection after pinealectomy in humans [<xref ref-type="bibr" rid="B96">96</xref>,<xref ref-type="bibr" rid="B97">97</xref>], whereas in chickens secretion is preserved with elimination of the nighttime peak [<xref ref-type="bibr" rid="B98">98</xref>]. Melatonin's actions appear to differ between humans, other mammals, and other vertebrates [<xref ref-type="bibr" rid="B93">93</xref>,<xref ref-type="bibr" rid="B95">95</xref>-<xref ref-type="bibr" rid="B97">97</xref>,<xref ref-type="bibr" rid="B99">99</xref>]. In humans melatonin may act by modulating calcium-activated calmodulin [<xref ref-type="bibr" rid="B100">100</xref>].</p><p>In a prospective study on pinealectomy in bipedal nonhuman primates, Cheung et al reported that none of the 18 monkeys developed scoliosis in a mean follow up period of 28 months. This study strongly suggests that the possible etiologic factors producing idiopathic scoliosis in lower animals are different from primates, and findings in lower animals cannot necessarily be extrapolated to human beings [<xref ref-type="bibr" rid="B101">101</xref>].</p><p>Studies on melatonin circadian secretion in humans report contradictory results. Machida et al reported that the integrated melatonin concentration throughout the 24-hour period in the patients who had a progressive curve was significantly lower than the level in the patients who had a stable curve or in the control group [<xref ref-type="bibr" rid="B102">102</xref>].</p><p>Hilibrand et al in a study that was conducted to test the suggestion of Machida found, in contrary to their research hypothesis, that morning urine melatonin levels were higher in patients with scoliosis than in the control subjects, but the differences between these values were not statistically significant [<xref ref-type="bibr" rid="B103">103</xref>]. They also reported a non-statistically significant difference of morning urine melatonin levels between patients with progressive and stable curves. The authors concluded that there was no difference in melatonin levels, as reflected in morning and evening urine collections, between patients with AIS and control subjects.</p><p>Fegan et al, in a case-control study of 24-hour urinary melatonin production in patients with adolescent idiopathic scoliosis reported that in adolescent idiopathic scoliosis, neither the presentation with a stable spinal deformity, nor presentation with a severe deformity requiring surgery is associated with melatonin deficiency [<xref ref-type="bibr" rid="B104">104</xref>].</p><p>No difference was also found by Bagnall et al in single day- and night-time measurements of the serum melatonin level in a control group and 7 patients with progressive adolescent idiopathic scoliosis [<xref ref-type="bibr" rid="B105">105</xref>]. In addition, no difference was found by Brodner et al in the urinary excretion of 6-sufatoxyl-melatonin in a control group and patients with progressive adolescent idiopathic scoliosis [<xref ref-type="bibr" rid="B106">106</xref>].</p><p>The argument against the alleged role of melatonin deficiency in AIS pathogenesis is contained in reports highlighting that an increased incidence of scoliosis has not been observed in children after pinealectomy or pineal irradiation because of pineal neoplasias, although they have a lack of serum melatonin [<xref ref-type="bibr" rid="B107">107</xref>-<xref ref-type="bibr" rid="B109">109</xref>].</p></sec><sec><title>Geographic latitude, lifestyle differences and scoliosis</title><p>The different prevalence of IS by latitude could be probably attributed also to different lifestyle of people at deferent geographical latitudes. The seating people in the northern hemisphere societies are reducing their lumbar lordosis contrary to the physiologic sagittal profile exhibited in photographs of the African people of the tribe Nuba, as seen in the book of German photographer Leni Riefenstahl. Reduced lumbar lordosis (the seating effect in the northern hemisphere societies) frequently occurs in correlation with the lateral spinal curvature. Loss of lumbar lordosis seems to be a fundamental problem leading to a destabilization of the spine also in frontal and coronal plane and vice versa correction of lumbar lordosis seems to correct spinal deformities also in fontal and coronal plane [<xref ref-type="bibr" rid="B110">110</xref>]. This mechanistic approach to scoliotic deformity prevalence which is quite different to the biological approach discussed in this study probably needs to be addressed with pertinent research in human groups living in different latitudes, which is lacking so far from the available literature.</p></sec><sec><title>A hypothesis</title><p>It has been assumed that there are two types of pathogenetic factors for AIS, the initiating and those that cause progression. Initiating factors that can meaningfully be distinguished from progressing factors would eventually faint or disappear, while progressive factors, which are generally thought to be a mechanical process, are associating with curve magnitude [<xref ref-type="bibr" rid="B111">111</xref>].</p><p>A possible preservation of high levels of melatonin secretion during the pre-menarcheal period in scoliotic girls due to light insufficiency in northern countries is associated with delay of the age at menarche. These high levels of melatonin are possibly identifiable before presentation of AIS, but would not be apparent at the time of clinical presentation of AIS in the vast majority of cases. The pre-menarcheal elevated levels of melatonin could be considered as a possible initiating factor of idiopathic scoliosis and it does not correlate with the severity and the site of the curve. It alters growth by lengthening the period of spine vulnerability while other pre-existing or aetiological factors are contributing to the development of AIS. Longitudinal studies on melatonin secretion in pre-pubertal girls that are at risk to develop AIS (i.e. with trunk asymmetry but no radiographic evidence of AIS) could be undertaken in order to test this hypothesis.</p><p>The clinical relevance and therapeutic implication that could also be derived from this study is that in northern latitudes, in girls with anticipated progressive scoliosis with no menarche, hormonal treatment in order to commence it might be of potential value to stop progression.</p></sec></sec><sec><title>Conclusion</title><p>In this survey it appears that late age at menarche is parallel with higher prevalence of AIS, especially in latitudes northern than 30 degrees. Pubarche appears later in girls that live in northern latitudes and thus prolongs the period of spine vulnerability while other pre-existing or aetiological factors are contributing to the development of AIS.</p></sec><sec><title>Abbreviations</title><p>Adolescent Idiopathic Scoliosis (AIS)</p><p>Hydroxy-Indolo-O-Methyl-Transferase (HIOMT)</p><p>Gonadotrophine Releasing Hormone (GnRH),</p><p>Luteinizing Hormone (LH),</p><p>Follicle Stimulating Hormone (FSH)</p></sec><sec><title>Authors' contributions</title><p>T BG conceived the idea of the presented study, was responsible for the methodological setting of the study, performed the major part of literature review and has written the manuscript. EV contributed in the statistical analysis of the study, performed part of literature review and also contributed by reviewing, text editing and adding certain parts of the manuscript. VM performed part of literature review, documented the latitude of the centre where a particular retrieved from Medline paper was originated. CM contributed in the statistical analysis of the study. GK Performed a part of literature review. All authors have read and approved the final manuscript.</p></sec> |
Protease nexin-1 expression is altered in human breast cancer | <sec><title>Background</title><p>Urokinase-type Plasminogen Activator (uPA), a serine protease, plays a pivotal role in human breast cancer metastasis by mediating the degradation of extracellular matrix proteins and promoting cell motility. In more advanced breast cancers, uPA activity is significantly up regulated and serves as a prognostic indicator of poor patient outcome. Classically, regulation of uPA activity, especially in breast cancers, is thought to be mediated by Type 1 Plasminogen Activator Inhibitor (PAI-1). However, we have recently found that a lesser known natural inhibitor of uPA, Protease Nexin 1 (PN-1), is expressed in normal human mammary tissue. Based on this observation, we investigated if PN-1 is also expressed in human breast cancers where it may contribute to the regulation of uPA and participate in the development of a metastatic phenotype.</p></sec><sec><title>Results</title><p>Using quantitative real-time PCR analysis, we measured PN-1 mRNA expression in tissues obtained from 26 human breast tumor biopsies and compared these values with those obtained from 10 normal breast tissue samples. Since both PAI-1 and uPA expression levels are known to be elevated in metastatic breast cancer, we also measured their levels in our 26 tumor samples for direct comparison with PN-1 expression. We found that PN-1 expression was elevated over that found in normal mammary tissue; an increase of 1.5- to 3.5-fold in 21 of 26 human breast tumors examined. As anticipated, both PAI-1 and uPA mRNA levels were significantly higher in the majority of breast tumors; 19 of 26 tumors for PAI-1 and 22 of 26 tumors for uPA. A quantile box plot of these data demonstrates that the elevated PN-1 expression in breast tumor tissues directly correlates with the increased expression levels found for PAI-1 and uPA.</p></sec><sec><title>Conclusion</title><p>The fact that PN-1 expression is elevated in human breast cancer, and that its increased expression is directly correlated with increases measured for PAI-1 and uPA, suggests that PN-1 may contribute to the regulation of uPA-mediate tumor cell motility and metastatic spread.</p></sec> | <contrib id="A1" contrib-type="author"><name><surname>Candia</surname><given-names>Britny J</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>bcandia@unm.edu</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Hines</surname><given-names>William C</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>Curthines@aol.com</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Heaphy</surname><given-names>Christopher M</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>cheaphy@salud.unm.edu</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Griffith</surname><given-names>Jeffrey K</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>jkgriffith@salud.unm.edu</email></contrib><contrib id="A5" corresp="yes" contrib-type="author"><name><surname>Orlando</surname><given-names>Robert A</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>rorlando@salud.unm.edu</email></contrib> | Cancer Cell International | <sec><title>Background</title><p>An important characteristic of highly invasive tumor cells is an elevated capacity to degrade the surrounding extracellular matrix (ECM). To achieve this elevated degradative capacity, tumor cells express a variety of proteases to digest ECM proteins that typically encapsulate growing, benign tumors [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B2">2</xref>]. It is now well established that proteins of the plasminogen activation (PA) system are elevated in breast cancer and serve as the primary functional players in ECM degradation [<xref ref-type="bibr" rid="B3">3</xref>,<xref ref-type="bibr" rid="B4">4</xref>]. Expression of one member of the PA system, the serine protease urokinase (uPA), is significantly upregulated in tumor cells and catalyzes the conversion of extracellular plasminogen to plasmin [<xref ref-type="bibr" rid="B5">5</xref>]. Plasmin is a broad-spectrum protease that cleaves many ECM proteins, as well as activates certain matrix metalloproteinases [<xref ref-type="bibr" rid="B6">6</xref>]. This proteolytic cascade enables highly migratory tumor cells to efficiently degrade their surrounding matrices, exit the primary site of tumor growth and colonize distant secondary sites [<xref ref-type="bibr" rid="B7">7</xref>]. In addition to its protease activity which augments breast tumor cell motility, high expression levels of uPA is also a well-established prognostic indicator of poor patient outcome during the course of breast cancer [<xref ref-type="bibr" rid="B8">8</xref>,<xref ref-type="bibr" rid="B9">9</xref>].</p><p>Regulation of extracellular uPA activity is known to occur through the inhibitory properties of type I plasminogen activator inhibitor (PAI-1), a serine protease inhibitor (SERPIN) that is synthesized and secreted often by the same cells that secrete uPA [<xref ref-type="bibr" rid="B10">10</xref>]. Because of the close functional relationship between uPA proteolytic activity and PAI-1 inhibitory function, it is thought that a well-controlled balance of uPA and PAI-1 dictates the extent of cell motility. Protease Nexin-1 (PN-1), another member of the SERPIN family [<xref ref-type="bibr" rid="B11">11</xref>], is highly expressed by stromal cells [<xref ref-type="bibr" rid="B12">12</xref>] and a potent inhibitor of uPA [<xref ref-type="bibr" rid="B10">10</xref>,<xref ref-type="bibr" rid="B13">13</xref>]. Interestingly, although PN-1 activity has been extensively studied within the context of neural development, few studies have been reported examining its expression in cancerous tissues and its potential role in cancer progression. PN-1 is expressed by astrocytes and glial cells [<xref ref-type="bibr" rid="B14">14</xref>], as well as neuroblastoma cells [<xref ref-type="bibr" rid="B15">15</xref>] where it is thought to promote neuronal cell survival [<xref ref-type="bibr" rid="B16">16</xref>] and modulate neurite outgrowth [<xref ref-type="bibr" rid="B17">17</xref>]. In addition, PN-1 inhibits thrombin-stimulated cell division [<xref ref-type="bibr" rid="B18">18</xref>], migration of cerebrellar granular cells [<xref ref-type="bibr" rid="B19">19</xref>], and uPA-dependent ECM degradation [<xref ref-type="bibr" rid="B20">20</xref>]. Thus, based on findings in other cell types, we hypothesize that PN-1 may contribute to tumor cell motility in advanced stage breast cancer by playing a role in the regulation of uPA activity. To address this hypothesis, we examined the expression of PN-1 in advanced stage human breast cancer tissues to determine if its expression is altered when compared to normal mammary tissue and to directly compare its expression level to those of PAI-1 and uPA. To accomplish this goal, we used quantitative real-time reverse transcription-PCR (QRT-PCR) to measure PN-1, PAI-1 and uPA expression levels within a set of breast tumor and normal breast tissue samples.</p></sec><sec sec-type="materials|methods"><title>Materials and methods</title><sec><title>Breast tissue samples</title><p>Frozen breast tumor specimens from anonymous patients (n = 26) were obtained from the University of New Mexico Cancer Research and Treatment Center Solid Tumor Facility, Albuquerque, New Mexico. In 25 out of 26 cases, tumor grade, tumor size, lymph node status, and the fraction of cells in S phase (based on flow cytometry cell cycle analysis) were included within the clinical history provided with each specimen. Anonymous normal breast mRNA (n = 10) originating from female patients where cause of death was unrelated to cancer, were purchased from Ambion (Austin, TX). The normal, control samples were supplied as two equal pools by the company.</p></sec><sec><title>Cell culture</title><p>MCF-7 human mammary epithelial cells, were provided by Dr. Steven Abcouwer, Hershey Medical Center, Hershey, Pennsylvania. MDA-MB-231 metastatic human mammary epithelial cells were obtained from American Type Culture Collection (Rockland, MD). Both cell lines were propagated in Dulbecco modified Eagle's medium (DMEM, Life Technologies/Invitrogen, Carlsbad, CA) supplemented with 10% fetal calf serum (Irvine Scientific, Santa Ana, CA), 0.1 mM non-essential amino acids, 1 mM sodium pyruvate, 0.01 mg/ml bovine insulin, and 100 U/ml penicillin G. Cells were cultured at 37°C with 5% CO<sub>2 </sub>and passaged once a week.</p></sec><sec><title>Preparation of tissue sections and RNA isolation</title><p>Serial frozen sections of breast samples, 10 μm in width, were mounted on Colorfrost slides (VWR, West Chester, PA) and stored at -70°C. Specimens were stained with hematoxylin/eosin and examined by a board-certified surgical pathologist, who assigned a histopathologic grade to the tumor and analyzed the normal tissue. Total RNA from cultured cells and frozen tumor tissue was isolated using silica-based spin-column extraction kits (RNeasy/DNeasy mini kits, Qiagen, Valencia, CA) according to the manufacturer's protocol. Total RNA was treated with RNase-free DNase I (Ambion, Austin, TX) to eliminate contaminating DNA. RNA integrity was evaluated by agarose gel electrophoresis.</p></sec><sec><title>Quantitative real-time RT-PCR</title><p>cDNA was synthesized by random decamer-primed reverse transcription of RNA (1 μg) using a TaqMan<sup>® </sup>Reverse Transcription kit (Applied Biosystems, Foster City, CA) according to the manufacturer's standard protocol. Negative controls contained RNase-free water substituted for reverse transcriptase. The mRNA levels of PN-1, PAI-1, uPA and TATA-binding protein (TBP) were measured in breast specimens, the MCF-7 mammary epithelial cell line, and the MDA-MB-231 metastatic mammary epithelial cell line using the ABgene Absolute SYBR Green QRT-PCR assay (Fisher Scientific, Hampton, NH). PN-1 primers were selected to amplify an 81 bp sequence spanning the intron located between exons 2 and 3. Primer sequences used for PN-1 were 5'-GAAGCAGCTCGCCATGGT-3' (forward), 5'-AGACGATGGCCTTGTTGATC-3' (reverse). TBP primer sequences used were 5'-CACGAACCACGGCACTGATT-3' (forward), 5'-TTTTCTTGCTGCCAGTCTGGAC-3' (reverse). Primer sequences used for PAI-1 were 5'-TGCTGGTGAATGCCCTCTACT-3' (forward), 5'-CGG TCA TTC CCA GGT TCT CTA-3' (reverse). uPA primer sequences used were 5'-CAC GCA AGG GGA GAT GAA-3' (forward), 5'-CA GCA TTT TGG TGG TGA CTT-3' (reverse) [<xref ref-type="bibr" rid="B21">21</xref>]. Final concentration of PN-1, PAI-1 and uPA primers used for amplification was 600 nmol/L forward, 600 nmol/L reverse; 600 nmol/L forward, 900 nmol/L reverse was used for TBP primers. Amplification of PN-1, PAI-1, uPA and TBP cDNA was performed using the MiniOpticon Real-Time PCR Detection System (Biorad, Hercules, CA). The cycling parameters used were as follows: 1 cycle, 95°C for 10 min; 50 cycles, 95°C for 15 sec and 60°C for 1 min; 1 cycle, 40°C for 3 min. The PN-1, PAI-1 and uPA mRNA levels were normalized to TBP mRNA levels using the Comparative C<sub>T </sub>method and are reported in the figures as fold difference compared to levels found in normal mammary tissue. Melting curve analyses were performed for all amplifications to verify that only single products were generated from the reactions. Amplicons were sequenced to verify authentic PN-1. The cDNA for human PN-1 was obtained from the I.M.A.G.E. Consortium (ID: 4824856; Genbank: <ext-link ext-link-type="gen" xlink:href="BC042628">BC042628</ext-link>; Genbank: <ext-link ext-link-type="gen" xlink:href="BC042628">BC042628</ext-link>).</p></sec></sec><sec><title>Results</title><sec><title>Quantitation of PN-1 expression in human breast tumors</title><p>For QRT-PCR analysis, we designed primers to amplify an 81 bp sequence of PN-1 spanning the splice junction between exons 2 and 3. Spanning a splice junction ensures that amplified products are derived solely from mRNA and not from genomic DNA that might remain in our preparation. In order to test the specificity of these novel primers, we amplified the 81 bp PN-I sequence by straight RT-PCR using RNA purified from a normal human fibroblast cell line (HuFb) and compared the product to that obtained from amplification using the human PN-1 cDNA. We chose to use human fibroblasts since they synthesize and secrete active PN-1 at levels corresponding to ~1% of all secreted proteins [<xref ref-type="bibr" rid="B22">22</xref>,<xref ref-type="bibr" rid="B23">23</xref>]. As anticipated, we found that our newly designed primers amplified only the expected 81 bp sequence (data not shown).</p><p>To quantify PN-1 expression in human breast cancers, we obtained 26 samples of breast tumor tissue, purified RNA and generated cDNA from this material. The cDNA were then analyzed by quantitative PCR and the results were compared directly to quantitative PCR values obtained using cDNA generated from normal human mammary tissue. We found that PN-1 expression was elevated over that found in normal mammary tissue; an increase of 1.5- to 3.5-fold in 21 of 26 human breast tumors examined (Fig. <xref ref-type="fig" rid="F1">1A</xref>). Since both PAI-1 [<xref ref-type="bibr" rid="B24">24</xref>-<xref ref-type="bibr" rid="B26">26</xref>] and uPA [<xref ref-type="bibr" rid="B24">24</xref>,<xref ref-type="bibr" rid="B26">26</xref>] expression levels are known to be elevated in metastatic breast cancer, we measured their levels in our 26 tumor samples for direct comparison with PN-1 expression. As anticipated, both PAI-1 (Fig. <xref ref-type="fig" rid="F1">1B</xref>) and uPA (Fig. <xref ref-type="fig" rid="F1">1C</xref>) mRNA levels were significantly higher in the majority of breast tumors; 19 of 26 tumors for PAI-1 and 22 of 26 tumors for uPA. A quantile box plot of the data shown in Figure <xref ref-type="fig" rid="F1">1</xref> permits a direct comparison of expression levels for PN-1, PAI-1 and uPA (Fig. <xref ref-type="fig" rid="F2">2</xref>). These data clearly show that PN-1 expression is elevated in the majority of human breast cancers examined and that this elevated expression directly correlates with the expected higher expression levels found for PAI-1 and uPA. Since the majority of our tumor samples represent advanced stage, grade 2 and 3 breast cancers, we are unable to determine at this time if PN-1 expression levels correlate with tumor grade, lymph node status or patient reoccurrence.</p><fig position="float" id="F1"><label>Figure 1</label><caption><p><bold>PN-1, PAI-1 and uPA mRNA expression in human breast tumor tissues and normal human mammary tissue</bold>. RNA was isolated from 26 breast tumors (T1–T26) and 10 normal breast samples. Normal samples were pooled into two equal groups (N1 and N2). mRNA levels for each gene were evaluated by QRT-PCR. Relative levels of PN-1, PAI-1 and uPA mRNA were normalized to TATA binding protein mRNA levels. Comparative C<sub>T </sub>method was used to calculate fold difference of PN-1 (A), PAI-1 (B) and uPA (C) expression in breast tumor tissue as compared to levels measured in normal breast tissue. The mean value of the two pooled normal samples was calculated and assigned a value of one in order to determine relative fold change of expression within the tumor samples. The standard deviation of the normal samples was 0.263. The box represents one standard deviation of the mean values obtained from normal mammary tissues.</p></caption><graphic xlink:href="1475-2867-6-16-1"/></fig><fig position="float" id="F2"><label>Figure 2</label><caption><p><bold>Comparison of PN-1, PAI-1 and uPA mRNA expression levels in 26 breast tumor samples</bold>. The box for each gene represents the interquartile range (25–75th percentile) and the line within this box is the median value. Bottom and top bars of the whisker indicate the 10th and 90th percentiles, respectively. Outlier values are indicated (closed squares).</p></caption><graphic xlink:href="1475-2867-6-16-2"/></fig></sec><sec><title>PN-1 expression in MCF-7 and MDA-MB-231 breast cancer cells</title><p>We plan to characterize the mechanism responsible for increased expression of PN-1 in breast cancer and determine its functional role in breast cancer metastasis. To accomplish this goal, we will require a cultured model system for accurate, controlled assessment of PN-1 promoter status, transcription factor requirements, and tumor cell invasive capacity. MCF-7 and MDA-MB-231 cells are well established cultured lines used extensively to study molecular details of breast cancer progression [<xref ref-type="bibr" rid="B27">27</xref>]. Hormone-responsive MCF-7 cells have a low invasive capacity and represent earlier stages of breast cancer, while hormone-independent MDA-MB-231 cells are highly invasive and represent more advanced stage breast cancer. To determine if these cell models will be useful for examining PN-1 function in breast cancer, we performed QRT-PCR analysis to identify if PN-1 expression is disregulated in a comparable manner to that seen in human breast cancer tissues. We found that MDA-MB-231 cells express 3.5-fold greater levels of PN-1 than MCF-7 cells (Fig. <xref ref-type="fig" rid="F3">3</xref>). Moreover, increased expression of PAI-1 and uPA were also found in MDA-MB-231 cells as compared to MCF-7.</p><fig position="float" id="F3"><label>Figure 3</label><caption><p><bold>QRT-PCR analysis of PN-1, PAI-1 and uPA mRNA expression in MCF-7 and MDA-MB-231 breast cancer cell lines</bold>. PN-1, PAI-1 and uPA message levels were quantitated in MCF-7 and MDA-MB-231 cells by QRT-PCR. Expression levels for PN-1, PAI-1 and uPA were normalized to values obtained for TATA binding protein. C<sub>T </sub>values for each gene obtained from MCF-7 cells were averaged and assigned a value of one to assess relative fold increase in expression in MDA-MB-231 cells.</p></caption><graphic xlink:href="1475-2867-6-16-3"/></fig></sec></sec><sec><title>Discussion</title><p>Advanced stage breast cancer is accompanied by a dramatic increase in metastatic potential of epithelial-derived tumor cells. The observed increase in tumor cell motility is aided by increased expression and activity of uPA [<xref ref-type="bibr" rid="B4">4</xref>]. For effective tumor cell migration, the proteolytic activity of uPA is thought to be balanced by the inhibitory activity of PAI-1 [<xref ref-type="bibr" rid="B28">28</xref>,<xref ref-type="bibr" rid="B29">29</xref>]. The cycling activities of proteolysis and protease inhibition lead to sequential rounds of cell detachment-reattachment, which in turn leads to an increase in cell motility. Indeed, elevated expression levels of both uPA and PAI-1 are characteristic of advanced stage breast cancers [<xref ref-type="bibr" rid="B30">30</xref>]. Interestingly, although PN-1 is structurally and functionally related to PAI-1, there have been no studies to date investigating if PN-1 contributes to breast cancer progression in a manner similar to that of PAI-1. To address this gap in knowledge, we examined if PN-1 expression is altered in human breast cancer by quantitating levels of PN-1 expression in human tissue samples obtained from tumor biopsies. In these same samples, we also quantitated PAI-1 and uPA expression levels for direct comparison to PN-1. Our findings indicate that PN-1 expression is elevated in the majority of human breast tumor tissues examined and that its expression levels are directly correlated with increases measured for PAI-1 and uPA. We also found that the highly metastatic, MDA-MB-231 breast cancer line expresses 3.5-fold greater levels of PN-1 compared to the non-tumorgenic, MCF-7 breast cancer cell line. This increase in PN-1 is also correlated to increases seen for PAI-1 and uPA in MDA-MB-231 cells. The elevated expression of all three genes is consistent with our measurements in human breast tumor samples. The significant differences in PN-1 expression between non-tumorigenic MCF-7 cells and highly invasive MDA-MB-231 cells should provide us with a good basis for identifying the mechanism responsible for altered PN-1 expression seen in breast tissues and allow us to examine PN-1 function in the context of elevated PAI-1 and uPA levels. Taken together, these data indicate that PN-1 expression is increased during breast cancer tumorigenesis and may contribute, along with PAI-1, to uPA-mediated tumor cell motility and a more advanced metastatic phenotype.</p><p>Although studies in the literature investigating a role for PN-1 in cancer progression are limited, our results complement and extend data presented in a recent report by Buchholz and colleagues [<xref ref-type="bibr" rid="B31">31</xref>]. Their study demonstrated that a highly metastatic pancreatic cancer line overexpressed PN-1, while a less metastatic subclone showed little PN-1 expression. The authors also noted that stable PN-1 overexpression in the less metastatic subclone greatly enhanced its local invasiveness in <italic>in vivo </italic>studies. Our studies expand on these observations by demonstrating an increase in PN-1 expression in human breast cancer tissues.</p><p>Quantitating expression of PAI-1 and uPA is of high prognostic value for assessing breast cancer survival outcome [<xref ref-type="bibr" rid="B24">24</xref>,<xref ref-type="bibr" rid="B25">25</xref>]. Numerous independent studies have shown that patients with low levels of PAI-1 and uPA in their primary tumor tissue have a significantly better survival rate than patients with high levels of either factor alone. Recently, the prognostic value of PAI-1 and uPA has been verified by a pooled analysis consisting of >8,000 breast cancer patients [<xref ref-type="bibr" rid="B26">26</xref>]. In light of the overlapping protease specificities of PAI-1 and PN-1 [<xref ref-type="bibr" rid="B32">32</xref>], together with the established role of PN-1 in neuronal cell regulation and motility [<xref ref-type="bibr" rid="B33">33</xref>,<xref ref-type="bibr" rid="B34">34</xref>], we believe it is likely that PN-1 also plays a role in breast cancer progression by contributing to events necessary for increased tumor cell motility. Increased expression of PN-1 by tumor cells may serve to modulate their adhesiveness or motility [<xref ref-type="bibr" rid="B35">35</xref>]. Alternatively, tumor cell activity may be influenced by tumor-stromal tissue crosstalk [<xref ref-type="bibr" rid="B36">36</xref>]. The breast neoplastic stroma contains a heterogeneous cell population composed of fibroblasts, myofibroblasts, and endothelial cells, which are all known to synthesize and secrete significant amounts of PN-1 [<xref ref-type="bibr" rid="B37">37</xref>]. Although it remains to be determined the precise mechanism by which PN-1 contributes to breast cancer tumor progression, the results of the present study establish a rationale for further investigation of PN-1 as a modulator of uPA activity in breast tumor cell motility. Future studies will be focused on identifying transcriptional and/or translational mechanisms controlling PN-1 expression by cancer cells and determining if PN-1 serves as an independent prognostic indicator of breast cancer staging by using a more widely defined sample of tumor tissues, including earlier stage cancers as well as late-stage carcinomas. In addition, the use of laser capture dissection technology will confine our QRT-PCR measurements to tumor tissue and eliminate contributions from surrounding normal mammary tissue that are likely to occur when using surgical specimens.</p></sec><sec><title>Conclusion</title><p>We quantitated PN-1 expression in samples obtained from biopsies of human breast tumors and from normal mammary tissues by QRT-PCR analysis and compared these results to those obtained for PAI-1 and uPA. Our findings indicate that PN-1 expression is elevated in a majority of human breast tumor tissues examined when compared to normal human mammary tissue. In addition, the elevated PN-1 expression in tumor tissues directly correlates with increased expression measured for PAI-1 and uPA. We also found that the highly metastatic, MDA-MB-231 breast cancer cell line expresses greater levels of PN-1 compared to the non-tumorgenic, MCF-7 breast cancer cell line. Consistent with observations obtained from tumor biopsies, PN-1 expression levels in MDA-MB-231 directly correlate with increases found for PAI-1 and uPA. These data indicate that PN-1 expression is increased during breast cancer tumorigenesis and may contribute, along with PAI-1, to uPA-mediated tumor cell motility and a more advanced metastatic phenotype.</p></sec><sec><title>Abbreviations</title><p>uPA, urokinase-type plasminogen activator; PAI-1, plasminogen activator inhibitor; ECM, extracellular matrix; PA, plasminogen activation; SERPIN, serine protease inhibitor; PN-1, protease nexin-1.</p></sec><sec><title>Competing interests</title><p>The author(s) declare that they have no competing interests.</p></sec><sec><title>Authors' contributions</title><p>BJC carried out the majority of studies and drafted the manuscript. WCH and CMH provided confirmed breast tumor and normal mammary samples. JKG assisted with data interpretation. RAO provided the original conceptual framework for the study, participated in the experimental design and finalized the manuscript for submission. Authors read and approved the final version.</p></sec> |
Enhanced molecular analyses by combination of the HOPE-technique and laser microdissection | <p>As part of an investigation aimed at illuminating the possibilities and limits of the HOPE-fixation and paraffin-embedding technique we here describe a novel procedure which was developed in order to combine the benefits of the HOPE-technique with the capabilities of laser microdissection. The presented procedure avoids the need for amplification of template-RNA thus facilitating reliable and reproducible results. The excellent preservation of nucleic acids, proteins, and morphology in HOPE-fixed, paraffin-embedded tissues enhances the molecular applications available to date with materials acquired by laser microdissection when compared to formalin fixed, paraffin-embedded tissues, thus substantially extending the methodological panel in tissue based research.</p> | <contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Goldmann</surname><given-names>Torsten</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>tgoldmann@fz-borstel.de</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Burgemeister</surname><given-names>Renate</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>renate.burgemeister@palm-microlaser.com</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Sauer</surname><given-names>Ulrich</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>Ulrich.Sauer@palm-microlaser.com</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Loeschke</surname><given-names>Siegfried</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>sloeschke@fz-borstel.de</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Lang</surname><given-names>Dagmar Silvia</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>dlang@fz-borstel.de</email></contrib><contrib id="A6" contrib-type="author"><name><surname>Branscheid</surname><given-names>Detlev</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>branscheid@kh-grosshansdorf.de</email></contrib><contrib id="A7" contrib-type="author"><name><surname>Zabel</surname><given-names>Peter</given-names></name><xref ref-type="aff" rid="I4">4</xref><email>pzabel@fz-borstel.de</email></contrib><contrib id="A8" contrib-type="author"><name><surname>Vollmer</surname><given-names>Ekkehard</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>evollmer@fz-borstel.de</email></contrib> | Diagnostic Pathology | <sec><title>Introduction</title><p>Laser microdissection is a valuable tool for analysis of molecular parameters in purified cell populations or even single cells taken out of their native environment within tissues. This technology requires both an acceptable preservation of morphological details which is needed for exact dissection and the preservation of the nucleic acids which are analyzed. Fixation of tissues with formalin results in well preserved morphology but – to a high degree – leads to degradation of nucleic acids which substantially constricts the spectrum of applicable molecular techniques [<xref ref-type="bibr" rid="B4">4</xref>]. The novel HOPE-technique with subsequent paraffin-embedding, as an alternative to formalin, has been shown to result in a morphological preservation comparable to formalin-fixed, paraffin-embedded specimens [<xref ref-type="bibr" rid="B6">6</xref>]. Moreover, we described procedures which permit successful application of all common molecular techniques such as <italic>in situ </italic>hybridization [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B4">4</xref>,<xref ref-type="bibr" rid="B7">7</xref>], immunohistochemistry without antigen retrieval and for formalin-refractory antigens [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B3">3</xref>], PCR [<xref ref-type="bibr" rid="B8">8</xref>,<xref ref-type="bibr" rid="B11">11</xref>], RT-PCR [<xref ref-type="bibr" rid="B6">6</xref>,<xref ref-type="bibr" rid="B11">11</xref>], Western blot [<xref ref-type="bibr" rid="B9">9</xref>], Northern blot, and transcription microarrays [<xref ref-type="bibr" rid="B2">2</xref>] to HOPE-fixed, paraffin-embedded tissues. HOPE-fixed tissues can be used for preparation of tissue microarrays for enhanced high throughput analyses on the molecular level [<xref ref-type="bibr" rid="B5">5</xref>,<xref ref-type="bibr" rid="B12">12</xref>]. Using the HOPE-technique as its crucial methodological base, <italic>ex vivo </italic>model systems could be established, e.g. for the simulation of early events in human infections and detection of chemotherapy resistances in human cancer [<xref ref-type="bibr" rid="B7">7</xref>,<xref ref-type="bibr" rid="B13">13</xref>]. In addition to tissues, cell-culture preparations have been prepared utilizing the HOPE-technique, which were then successfully applied to in situ hybridization targeting mRNA or immunocytochemistry with excellent preservation of morphological details [<xref ref-type="bibr" rid="B10">10</xref>]. In this study we describe the use of HOPE-fixed, paraffin-embedded tissues for laser microdissection and subsequent molecular analysis of RNA-transcripts by real time RT-PCR. Results are set in relation to those obtained with formalin fixed, paraffin-embedded tissues from the same lesions. This real time RT-PCR was unambiguously performed without any amplification of RNA or total cDNA, which eliminates the need for such hardly controllable steps necessary in established procedures with formalin fixed tissues.</p></sec><sec sec-type="materials|methods"><title>Materials and methods</title><sec><title>Sample preparation</title><p>The specimens used were tumor-bearing or tumor-free materials (at least 5 cm away from the tumor front) from lobectomy or pneumectomy because of lung cancer. For means of comparison, tissue samples from the same organs were conventionally formalin-fixed and paraffin-embedded or treated according to the HOPE-technique.</p><p>Fixation of tissues by application of the HOPE-technique was carried out like previously described (6), starting with an incubation of fresh tissue specimens in an aqueous protection-solution (containing a mixture of different amino-acids) overnight at low temperatures of 0–4°C (DCS, Hamburg, Germany). Incubation in acetone followed at 0–4°C. Tissues were then directly embedded with paraffin.</p><p>Tissue sections from HOPE- or formalin fixed and paraffin embedded (FFPE) human lung cancer tissue samples were prepared on membrane mounted slides (PALM MembraneSlides, P.A.L.M., Bernried, Germany).</p><p>HOPE-sections were deparaffinized either with isopropanol (2 × 10 min at 60°C) or like the normal FFPE-sections with xylene washes (2 × 10 min at room temperature).</p><p>All sections were stained with cresyl violet (1% w/v cresyl violet acetate in 100% ethanol; Aldrich #86,098-0) for 1 min at room temperature, washed shortly in 70% and 100% ethanol and subsequently air dried.</p></sec><sec><title>Laser microdissection and pressure catapulting (LMPC)</title><p>LMPC was performed using a PALM MicroBeam system with a pulsed UV-A nitrogen laser (337 nm). Using the 10× objective areas of interest were marked, cut and catapulted into the cap of 0.5 ml microfuge tubes with adhesive filling (PALM AdhesiveCaps, P.A.L.M., Bernried, Germany). Smaller areas were pooled to reach average sample sizes of 0.5–1 million square microns. After completed microdissection the remaining part of each tissue section was cut out with a scalpel and collected in regular 0.5 ml microfuge tube (cut out section).</p></sec><sec><title>RNA extraction</title><p>Tissue samples were extracted using the RNeasy Micro Kit (Qiagen, Hilden, Germany, #74004) following the manufacturer's instruction manual. Briefly, the cells were dissolved in 350 μl of RLT lysis buffer, treated with DNaseI after the first washing step according to the manual and finally the purified RNA was eluted in 12 μl of RNase free water.</p><p>Spectrophotometry was performed for testing concentration and purity of RNA, which revealed well-preserved RNA of high concentration throughout the samples.</p><p>Of each RNA sample 1 μl was tested in an Agilent Bioanalyzer 2100 using the RNA 6000 Pico LabChip kit (Agilent, Waldbronn, Germany, #5065-4473) to determine the RNA integrity.</p></sec><sec><title>Reverse transcription and RT-PCR</title><p>The synthesis of cDNA was performed using the 1st strand cDNA synthesis kit for RT-PCR (AMV) (Roche, Penzberg, Germany, #1483-188) following the manufacturer's instructions. Briefly, 8.2 μl of RNA solution were reverse transcribed in a final volume of 20 μl using random primers from the kit to ensure efficient transcription also of fragmented or partially degraded RNA.</p><p>RT-PCR was performed using primers for the low abundant reference gene human Hypoxanthine phosphoribosyltransferase (huHPRT, GenBank: <ext-link ext-link-type="gen" xlink:href="M31642">M31642</ext-link>, pos. 383–411 and pos. 613-591) producing a specific amplicon of 231 bp.</p><p>The PCR reaction was done in a LightCycler Instrument using the Fast Start Master Plus SYBR Green system (Roche, Penzberg, Germany) applying the following conditions: 95°C for 10 min; 50 cycles at 95°C for 10 sec, 67°C for 10 sec, 72°C for 10 sec followed by a melting curve analysis from 70° to 99°C in 0.1°C steps to control specificity. The amount of RT-PCR product was automatically assessed by real time fluorescence detection with the LightCycler software package.</p></sec></sec><sec><title>Results</title><p>HOPE-fixed, paraffin-embedded tissues displayed excellent 'formalin-like' preservation of morphological details after sectioning and cresyl violet staining with good adhesion to the membrane mounted slides (PALM MembraneSlides) used for laser microdissection (Fig. <xref ref-type="fig" rid="F1">1</xref>). Laser microdissection was smoothly possible without any modification to the PALM MicroBeam system; the same holds true for the pressure catapulting of the dissected material into the collection tubes (Fig. <xref ref-type="fig" rid="F1">1</xref>). Extraction of RNA according to the protocol described above was achieved quickly. RNA-quality of the HOPE-fixed samples analyzed by using the Agilent Bioanalyzer was clearly superior to the formalin-fixed materials, which is exemplified in figure <xref ref-type="fig" rid="F2">2</xref>. In the real time RT-PCR (Fig. <xref ref-type="fig" rid="F3">3</xref>) similar results were obtained with deparaffinization utilizing xylene or isopropanol for the samples which have undergone LMPC and the cut out sections. Differences between cut out sections and LMPC can be explained by the differing amounts of starting material. FFPE materials – either LMPC or cut out sections – did not show any amplification of the targeted human Hypoxanthine phosphoribosyltransferase fragment. This is in good agreement with the results of RNA-quality analysis in FFPE materials obtained with the Bioanalyzer, which does not show any visualized RNA in these blocks (Fig. <xref ref-type="fig" rid="F2">2</xref>).</p><fig position="float" id="F1"><label>Figure 1</label><caption><p>Photomicrographs of HOPE-fixed tissues, deparaffinized and stained with cresyl violet (both 100× magnification). Sections were subjected to LMPC utilizing a PALM MicroBeam system. A: Pulmonary carcinoma with the area subjected for LMPC marked. B: Pulmonary carcinoma with the area subjected for LMPC microdissected and transferred to a microfuge tube by pressure catapulting.</p></caption><graphic xlink:href="1746-1596-1-2-1"/></fig><fig position="float" id="F2"><label>Figure 2</label><caption><p>Results of RNA-integrity-testing by a Agilent Bioanalyzer 2100 using the RNA 6000 Pico LabChip kit. Lane 1 and 2 are HOPE-fixed materials and lane 3 is FFPE.</p></caption><graphic xlink:href="1746-1596-1-2-2"/></fig><fig position="float" id="F3"><label>Figure 3</label><caption><p>Real time RT-PCR targeting a 231 bp fragment of the RNA of human Hypoxanthine phosphoribosyltransferase. Green: HOPE, deparaffinized with xylene, cut out section. Black drawn through: HOPE, deparaffinized with isopropanol, cut out section. Black with dots: HOPE, isopropanol, whole slice before mounting on slide directly processed. Blue: HOPE, deparaffinized with isopropanol, LMPC. Red: HOPE, deparaffinized with xylene, LMPC. Black noncontinuous: two superposing samples, FFPE, both deparaffinized with xylene, cut out section and LMPC</p></caption><graphic xlink:href="1746-1596-1-2-3"/></fig></sec><sec><title>Discussion</title><p>Laser microdissection for analysis of molecular parameters in cell populations or single cells out of tissues requires acceptable preservation of morphology and nucleic acids. This cannot be achieved to a sufficient degree by the use of FFPE, which display degradation of nucleic acids if compared to fresh materials, which prompted us to this investigation of HOPE-fixed materials.</p><p>We showed that HOPE-fixed, paraffin-embedded tissues are well suited for laser microdissection. The well preserved morphology is comparable to formalin-fixed specimens and superior when compared to frozen sections. RNA of high quality can be extracted from the microdissected samples and subsequently being used for successful analysis by real time RT-PCR. These RT-PCR analyses can be performed without the need for any RNA amplification procedure. This results in higher specificity and reproducibility if compared to protocols which have to use such techniques due to degradation within the (usually formalin-fixed) specimens. Therefore results of real time RT-PCR with HOPE-fixed tissues are clearly superior to those obtained in formalin-fixed materials.</p><p>The combination of the enhanced molecular possibilities provided by the HOPE-technique with laser microdissection represents a novel tool for future tissue-based studies. Appropriate studies are underway to further extend these promising initial results.</p></sec> |
A high density of ancient spliceosomal introns in oxymonad excavates | <sec><title>Background</title><p>Certain eukaryotic genomes, such as those of the amitochondriate parasites <italic>Giardia </italic>and <italic>Trichomonas</italic>, have very low intron densities, so low that canonical spliceosomal introns have only recently been discovered through genome sequencing. These organisms were formerly thought to be ancient eukaryotes that diverged before introns originated, or at least became common. Now however, they are thought to be members of a supergroup known as excavates, whose members generally appear to have low densities of canonical introns. Here we have used environmental expressed sequence tag (EST) sequencing to identify 17 genes from the uncultivable oxymonad <italic>Streblomastix strix</italic>, to survey intron densities in this most poorly studied excavate group.</p></sec><sec><title>Results</title><p>We find that <italic>Streblomastix </italic>genes contain an unexpectedly high intron density of about 1.1 introns per gene. Moreover, over 50% of these are at positions shared between a broad spectrum of eukaryotes, suggesting theyare very ancient introns, potentially present in the last common ancestor of eukaryotes.</p></sec><sec><title>Conclusion</title><p>The <italic>Streblomastix </italic>data show that the genome of the ancestor of excavates likely contained many introns and the subsequent evolution of introns has proceeded very differently in different excavate lineages: in <italic>Streblomastix </italic>there has been much stasis while in <italic>Trichomonas </italic>and <italic>Giardia </italic>most introns have been lost.</p></sec> | <contrib id="A1" contrib-type="author"><name><surname>Slamovits</surname><given-names>Claudio H</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>claudio@mail.botany.ubc.ca</email></contrib><contrib id="A2" corresp="yes" contrib-type="author"><name><surname>Keeling</surname><given-names>Patrick J</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>pkeeling@interchange.ubc.ca</email></contrib> | BMC Evolutionary Biology | <sec><title>Background</title><p>One of the prominent features that distinguishes eukaryotic genomes from those of prokaryotes is the presence of spliceosomal introns. Introns are intervening sequences that are removed from expressed RNAs, in the case of spliceosomal introns through a series of transesterfications mediated by a large riboprotein complex called the spliceosome [<xref ref-type="bibr" rid="B1">1</xref>]. Spliceosomal introns are only known from eukaryotic nuclear genomes, and were the subject of intense controversy over their potential role in early gene origins and evolution, the so-called introns early versus late debate [<xref ref-type="bibr" rid="B2">2</xref>-<xref ref-type="bibr" rid="B4">4</xref>]. One of the interesting features of intron evolution that came to light during this debate was the large range in intron density. At one extreme, introns appeared to be lacking in several protist lineages that were, at the time, thought to be the earliest-branching eukaryotes. These lineages included diplomonads (e.g., <italic>Giardia</italic>) and parabasalia (e.g. <italic>Trichomonas</italic>).</p><p>The early-branching status of these organisms has since been undermined by a variety of data, and now diplomonads and parabasalia are thought to be part of a large assemblage of protists called excavates, which also includes trypanosomes, euglenids, and a number of parasitic and free living flagellate or amoeboflagellate lineages [<xref ref-type="bibr" rid="B5">5</xref>]. However, despite the accumulation of a considerable quantity of molecular data from both <italic>Giardia </italic>and <italic>Trichomonas</italic>, as well as the identification of proteins involving splicing in <italic>Trichomonas </italic>[<xref ref-type="bibr" rid="B6">6</xref>], evidence for introns in their genomes remained intriguingly elusive. Indeed, only recently were introns finally characterized in these organisms [<xref ref-type="bibr" rid="B7">7</xref>-<xref ref-type="bibr" rid="B9">9</xref>], and remain extremely rare. Only three introns have been found in <italic>G. intestinalis </italic>among thousands of known genes [<xref ref-type="bibr" rid="B8">8</xref>,<xref ref-type="bibr" rid="B9">9</xref>] and forty-one introns were identified in the <italic>T. vaginalis </italic>genome after exhaustive searches [<xref ref-type="bibr" rid="B7">7</xref>]. Information from excavates other than <italic>Trichomonas </italic>and <italic>Giardia </italic>is scarce, but overall there seems to be a generally low density of introns (with the possible exception of Jakobid flagellates based on one family of genes[<xref ref-type="bibr" rid="B10">10</xref>]). Moreover, other instances of non-canonical introns and splicing are known in excavates [<xref ref-type="bibr" rid="B11">11</xref>-<xref ref-type="bibr" rid="B13">13</xref>], as are systems where splicing machinery is put to a slightly different use such as trans-splicing [<xref ref-type="bibr" rid="B14">14</xref>-<xref ref-type="bibr" rid="B16">16</xref>].</p><p>One of the excavate groups about which we know very little are the oxymonads. Oxymonads are anaerobic flagellates found almost exclusively in association with animals, many in the guts of termites and wood-eating roaches [<xref ref-type="bibr" rid="B17">17</xref>]. This is the only group of amitochondriates for which secondary loss of mitochondria has not been yet demonstrated, but they are closely related to the flagellate <italic>Trimastix</italic>, which has a vestigial organelle, so a primary lack of mitochondria in oxymonads is unlikely. Mostoxymonads are not available in culture because they live in complex communities with other protists and prokaryotes. As a result, there are few molecular data available from any oxymonad, and no introns have been identified [<xref ref-type="bibr" rid="B18">18</xref>]. The oxymonad <italic>Streblomastix strix </italic>is asymbiont of the dampwood termite <italic>Zootermopsis angusticollis </italic>from North American Pacific coastal region. This species has a number of unusual morphological characters, including a peculiar long slender cell shape with deep longitudinal vanes which is apparently maintained by intimate association with epibiotic bacteria [<xref ref-type="bibr" rid="B19">19</xref>], So far, many copies of four genes (alpha-tubulin, beta-tubulin, HSP90, and elongation factor-1 alpha) have been characterized from <italic>S. strix </italic>[<xref ref-type="bibr" rid="B18">18</xref>], and the complete absence of introns from all sequences (a total of 19,888 bp) suggests the oxymonads might share low intron densities apparently common to excavates. Here, we have used the recent documentation of a rare non-canonical genetic code in <italic>Streblomastix </italic>[<xref ref-type="bibr" rid="B18">18</xref>] to identify 17 oxymonad genes from an environmental expressed sequence tag (EST) pool from the hindgut of <italic>Zootermopsis</italic>. The genomic DNA sequence for each mRNA was determined and we found that, in contrast to other amitochondriate protists and the limited data previously available for <italic>Streblomastix</italic>, a relatively high density of canonical spliceosomal introns. Moreover, a large proportion of these introns are shared in position with other distantly related eukaryotes, suggesting that they are ancient intron positions retained in oxymonads but lost in other excavates such as <italic>Giardia </italic>and <italic>Trichomonas</italic>.</p></sec><sec><title>Results and discussion</title><sec><title>Identification of oxymonad sequences from ESTs</title><p>A total of 5,337 ESTs from a <italic>Z. angusticollis </italic>termite hindgut cDNAlibrary were sequenced and found to form 2,595 clusters of unique sequences. Overall, the sample was dominated by sequences of parabasalian origin (transcripts encoding parabasalian actin and actin-related proteins alone represented 32% of all ESTs). Moreover, there are few oxymonad sequences known outside this sample, so <italic>Streblomastix </italic>cDNAs could not be identified based on similarity to known genes (only 2 ESTs, corresponding to known <italic>Streblomastix </italic>alpha- and beta-tubulin sequences, were identified by BLASTX searches). Accordingly, we used the presence of a rare non-canonical genetic code in <italic>Streblomastix </italic>as a filter to identify at least those genes where non-canonical codons were sampled. In <italic>Streblomastix</italic>, TAA and TAG encode glutamine (Q) rather than stop as in the universal code [<xref ref-type="bibr" rid="B18">18</xref>], so all clusters were compared to public databases using BLASTX and examined individually for in frame stop codons, in particular at positions normally encoding glutamine. No other protist known to exist in <italic>Z. angusticollis </italic>has been shown to possess a non-canonical genetic code. The other prominent protists in this insect are parabasalia, which are not known to deviate from the universal genetic code and whose sequences are also easy to identify with BLASTX searches given their high similarity with <italic>T. vaginalis </italic>genomic sequences.</p><p>Using the non-canonical code as a filter, we were able to identify 17 protein-coding genes (Table <xref ref-type="table" rid="T1">1</xref>), representing a major increase in the available sequence data from oxymonads. Formerly partial sequences of 4 protein coding genes were known from <italic>Streblomastix</italic>, and a handful of cDNAs were known from other species [<xref ref-type="bibr" rid="B18">18</xref>,<xref ref-type="bibr" rid="B20">20</xref>,<xref ref-type="bibr" rid="B21">21</xref>]. From this sample we recovered 8 complete protein-coding genes, an additional 5 genes missing only 1 to 30 codons at the N-terminus, and another three lacking from 100 to 160 codons at the N-terminus. In addition, a short fragment encoding 258 codons of the large protein UPF1 was severely truncated, but we failed to obtain more sequence. Complete or near-complete sequences included five ribosomal proteins (RPS7 and 9, RPL4, 18 and 21), alpha- and beta-tubulin, the nuclear transporter Ntf2, cyclophilin, a peptidyl-isomerase involved in assisting protein folding, and NAD-dependent glutamate dehydrogenase. Also, two versions of the cystein-protease Cathepsin B were obtained. Although related, these sequences exhibited several differences at the amino acid level, so they are likely to represent multiples copies of the gene. We also identified two copies of the carbon metabolism enzyme pyruvate phosphate dikinase (PPDK), the functional and evolutionary significance of which are discussed elsewhere [<xref ref-type="bibr" rid="B22">22</xref>]. One conserved hypothetical protein was also found to use the <italic>Streblomastix </italic>genetic code. This protein has homologues in diverse eukaryotes (e.g. <italic>Arabidopsis thaliana </italic>AAM67532), buthas no assigned function. UPF1 is a key member of nonsense-mediated decay (NMD). This protein may be of interest in <italic>Streblomastix </italic>because it is involved in a mechanism of mRNA surveillance devoted to eliminating defective transcripts, such as those carrying premature stop codons [<xref ref-type="bibr" rid="B23">23</xref>]. NMD has been described and studied in animals and yeasts, but not yet found in protists [<xref ref-type="bibr" rid="B24">24</xref>]. The presence of UPF1 in <italic>Streblomastix </italic>suggests NMD is used by oxymonads, and in organisms where stop codons are reassigned to encode amino acids. Finally, UAP56 is a member of the DEAD box family of RNA helicases that is associated with the spliceosome and intervenes in early steps of pre-mRNA splicing in mammals and yeasts, but is also linked to mRNA export [<xref ref-type="bibr" rid="B25">25</xref>,<xref ref-type="bibr" rid="B26">26</xref>]. Even in the absence of introns, the presence of UAP56 indicates the likely presence of the spliceosome in oxymonads, and therefore by extension introns as well.</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p><italic>S. strix </italic>genes identified in this study. <italic>Streblomastix </italic>genes recovered from the <italic>Z. angusticollis </italic>hindgut RNA sample. For incomplete sequences, the number of missing amino acids were estimated from homologues from <italic>Giardia </italic>and/or <italic>Trichomonas</italic>. UPF1 shows extensive size variation among eukaryotic lineages (between 800 and 1600 amino acids, approximately), so it is difficult to determine how much sequence this fragment is lacking. ND: not determined.</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left"><bold>Protien name</bold></td><td align="center"><bold>Length(AA)</bold></td><td align="center"><bold>Introns</bold></td></tr></thead><tbody><tr><td align="left">RPS7</td><td align="center">189</td><td align="center">0</td></tr><tr><td align="left">RPS9</td><td align="center">188</td><td align="center">3</td></tr><tr><td align="left">RPL4</td><td align="center">445</td><td align="center">ND</td></tr><tr><td align="left">RPL18</td><td align="center">184</td><td align="center">1</td></tr><tr><td align="left">RPL21</td><td align="center">146</td><td align="center">2</td></tr><tr><td align="left">Alpha-tubulin</td><td align="center">450</td><td align="center">0</td></tr><tr><td align="left">Beta-tubulin</td><td align="center">447</td><td align="center">0</td></tr><tr><td align="left">Cyclophillin</td><td align="center">167</td><td align="center">0</td></tr><tr><td align="left">Cathespin B (1)</td><td align="center">312</td><td align="center">1</td></tr><tr><td align="left">Cathespin B (2)</td><td align="center">283</td><td align="center">3</td></tr><tr><td align="left">NAD-dependent glutamate dehydrogenase</td><td align="center">446</td><td align="center">1</td></tr><tr><td align="left">Pyruvate phosphate dikinase (1)</td><td align="center">779</td><td align="center">1</td></tr><tr><td align="left">Pyruvate phosphate dikinase (2)</td><td align="center">784</td><td align="center">0</td></tr><tr><td align="left">UPF1</td><td align="center">258</td><td align="center">ND</td></tr><tr><td align="left">UAP56/BAT1</td><td align="center">272</td><td align="center">2</td></tr><tr><td align="left">Nuclear transport factor 2</td><td align="center">123</td><td align="center">2</td></tr><tr><td align="left">Conserved hypothetical protein</td><td align="center">203</td><td align="center">5</td></tr></tbody></table></table-wrap></sec><sec><title>Introns in <italic>Streblomastix </italic>genes</title><p>Genomic DNA sequences were obtained for all <italic>Streblomastix </italic>coding regions identified from the cDNA library. Despite the fact that many alleles and loci representing four proteins were previously found to contain no introns, we found that most of the genes encoding these transcripts were interrupted by introns. In total, we found 21 introns in our sample of 17 genes with genes having as many as 5 introns (Table <xref ref-type="table" rid="T1">1</xref>). Including previously known intronless EF-1 alpha and HSP90 genes (alpha and beta tubulin are included in our sample) [<xref ref-type="bibr" rid="B18">18</xref>], the overall density is 1.1 introns per gene. However this is likely to be an underestimation since some of our sequences are truncated and could contain further introns, and there is a bias favouring genes that are more often intronless (e.g. HSP90). This density is less than that observed in the relatively intron-rich mammals and plants, but comparable to many other eukaryotic genomes, and certainly much higher than <italic>Giardia </italic>and <italic>Trichomonas </italic>where only 3 and 41 introns have been detected despite very large quantities of genomic data [<xref ref-type="bibr" rid="B7">7</xref>,<xref ref-type="bibr" rid="B9">9</xref>].</p><p>Overall, the <italic>Streblomastix </italic>introns were found to exhibit characteristics typical of eukaryotic spliceosomal introns. Introns ranged from 46 to 229 bases (Table <xref ref-type="table" rid="T2">2</xref>), but most were between 60 and 100 bases long, and the AT content was markedly higher than that of the coding sequence (Table <xref ref-type="table" rid="T2">2</xref>). Spliceosomal introns are flanked by GT and AG dinucleotides in the vast majority of known introns, while about 0.1% are U12 AT-AC introns[<xref ref-type="bibr" rid="B27">27</xref>] and a very small proportion of known introns use other non-canonical splice boundaries. Interestingly, however, the first of only three introns from <italic>G. intestinalis </italic>to be discovered has CU-AG boundaries [<xref ref-type="bibr" rid="B8">8</xref>]. Of the twenty-one introns from <italic>Streblomastix</italic>, 20 featured canonical GT-AG boundaries, but one intron in <italic>rps9 </italic>was flanked by AC-AG splice sites. However, the <italic>Streblomastix </italic>intron is located very close to the start of the transcript, so we cannot exclude the possibility that this intron sequence is incomplete and a canonical boundary lies upstream.</p><table-wrap position="float" id="T2"><label>Table 2</label><caption><p>Basic features of the <italic>S. strix </italic>introns. Characteristics of 21 introns found in 17 <italic>Streblomastix </italic>genes analysed. Size and base composition are shown. GC% mRNA shows base composition of the coding sequence (excluding introns).</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left"><bold>Protien name</bold></td><td align="center"><bold>Intron</bold></td><td align="center"><bold>Size(bp)</bold></td><td align="center"><bold>%GC(Intron)</bold></td><td align="center"><bold>%GC(coding)</bold></td></tr></thead><tbody><tr><td align="left">RPS9</td><td align="center">1</td><td align="center">57</td><td align="center">0.29</td><td align="center">0.46</td></tr><tr><td></td><td align="center">2</td><td align="center">88</td><td align="center">0.38</td><td></td></tr><tr><td></td><td align="center">3</td><td align="center">57</td><td align="center">0.3</td><td></td></tr><tr><td align="left">RPL18</td><td align="center">1</td><td align="center">60</td><td align="center">0.25</td><td align="center">0.49</td></tr><tr><td align="left">RPL21</td><td align="center">1</td><td align="center">46</td><td align="center">0.11</td><td align="center">0.37</td></tr><tr><td></td><td align="center">2</td><td align="center">66</td><td align="center">0.09</td><td></td></tr><tr><td align="left">Cathespin B (1)</td><td align="center">1</td><td align="center">122</td><td align="center">0.16</td><td align="center">0.35</td></tr><tr><td align="left">Cathespin B (2)</td><td align="center">1</td><td align="center">63</td><td align="center">0.22</td><td align="center">0.47</td></tr><tr><td></td><td align="center">2</td><td align="center">66</td><td align="center">0.18</td><td></td></tr><tr><td></td><td align="center">3</td><td align="center">91</td><td align="center">0.15</td><td></td></tr><tr><td align="left">Glutamate Dehydrogenase</td><td align="center">1</td><td align="center">100</td><td align="center">0.36</td><td align="center">0.38</td></tr><tr><td align="left">Pyruvate phosphate dikinase (1)</td><td align="center">1</td><td align="center">229</td><td align="center">0.21</td><td align="center">0.44</td></tr><tr><td align="left">UAP56/BAT1</td><td align="center">1</td><td align="center">105</td><td align="center">0.22</td><td align="center">0.37</td></tr><tr><td></td><td align="center">2</td><td align="center">168</td><td align="center">0.15</td><td></td></tr><tr><td align="left">Nuclear transport factor 2</td><td align="center">1</td><td align="center">56</td><td align="center">0.27</td><td align="center">0.41</td></tr><tr><td></td><td align="center">2</td><td align="center">58</td><td align="center">0.2</td><td></td></tr><tr><td align="left">Conserved hypothetical</td><td align="center">1</td><td align="center">64</td><td align="center">0.27</td><td align="center">0.44</td></tr><tr><td></td><td align="center">2</td><td align="center">61</td><td align="center">0.31</td><td></td></tr><tr><td></td><td align="center">3</td><td align="center">102</td><td align="center">0.23</td><td></td></tr><tr><td></td><td align="center">4</td><td align="center">54</td><td align="center">0.3</td><td></td></tr><tr><td></td><td align="center">5</td><td align="center">69</td><td align="center">0.22</td><td></td></tr><tr><td colspan="5"><hr></hr></td></tr><tr><td align="left"><bold>Average</bold></td><td></td><td align="center"><bold>85</bold></td><td align="center"><bold>0.23</bold></td><td align="center"><bold>0.42</bold></td></tr></tbody></table></table-wrap><p>We also inspected intron sequences to look for conserved features that may correspond to functional motifs. Although signals important for intron recognition and removal are not very well understood, some have been studied in certain detail, especially in mammals and yeasts. The branch-point is a sequence element required for lariat formation during splicing [<xref ref-type="bibr" rid="B28">28</xref>]. The mammalian branch point consensus sequence has been determined to be CURAY, where the A corresponds to the actual branching point. In yeast, the branch point sequence is more strictly defined as UACUAAC [<xref ref-type="bibr" rid="B29">29</xref>]. The plant branch point appears to be similar to that of mammals [<xref ref-type="bibr" rid="B30">30</xref>]. In all cases, the branch point is located near the 3' splice site, but the exact location varies. In contrast, the putative branch point found in the three introns of <italic>Giardia </italic>(ACURAC) is located directly adjacent to the 3' splice site [<xref ref-type="bibr" rid="B9">9</xref>]. Likewise, the potential branch points in <italic>Trichomonas </italic>are invariably ACUAAC and are also adjacent to the 3' splice site [<xref ref-type="bibr" rid="B7">7</xref>]. The apparently strict requirement for proximity between the branch point and 3' splice site is rare in metazoa and yeast, but common to <italic>Trichomonas </italic>and <italic>Giardia</italic>. This led to the suggestion that the branch point and 3' splice site recognition could be combined in these species [<xref ref-type="bibr" rid="B7">7</xref>]. Aligning the regions around the 5' splice site of all <italic>Streblomastix </italic>introns (Figure <xref ref-type="fig" rid="F1">1</xref>) reveals highly conserved A, U and G residues at positions +3 to +5, respectively. This is in good agreement with the first 5 positions of the yeast 5' splice site (typically GUAUGU), suggesting that interaction with U1 snRNA is conserved. At the 3' splice site no branch point motifs like those of <italic>Giardia </italic>or <italic>Trichomonas </italic>were observed, although the -1 position (adjacent to the AG dinucleotide) was invariably a pyrimidine and the region is T-rich. Overall, branch point specification in <italic>Streblomastix </italic>introns is probably different from that of <italic>Giardia </italic>or <italic>Trichomonas</italic>. Under the assumption that these lineages are related it is possible that the peculiar features observed in <italic>Giardia </italic>and <italic>Trichomonas </italic>may be a consequence of secondary implification in their spliceosomal apparatus.</p><fig position="float" id="F1"><label>Figure 1</label><caption><p>Examples of conserved intron positions between <italic>Streblomastix</italic> and other eukaryotes. In each case a section of the gene is shown aligned at the amino acid level, and the position of the intron found in all aligned sequences is indicated above by a triangle with a number indicating the phase (0, 1, or 2). Aligned sequences are from three unikont groups, animals (<italic>H. sapiens</italic> and <italic>P. troglodytes</italic>), fungi (<italic>S. pombe</italic>, <italic>U. maydis</italic> and <italic>A. fumigatus</italic>), and slime molds (<italic>D. discoideum</italic>), from one chromalveolate group, the ciliate (<italic>P. tetraurelia</italic>), and from three plantae groups, land plants (<italic>A. thaliana</italic>), green algae (<italic>Bigelowiella natans</italic> nucleomorph), and red algae (<italic>Guillardia thet</italic><italic>a</italic> nucleomorph).</p></caption><graphic xlink:href="1471-2148-6-34-1"/></fig></sec><sec><title>Conservation of intron positions in oxymonad genes</title><p><italic>Streblomastix </italic>intron-containing genes were compared to homologues from other eukaryotes, and surprisingly more than half of the <italic>Streblomastix </italic>intron positions were shared with members of at least two different eukaryotic supergroups, unikonts and plants (where the best sampling of intron-containing genes exists), and in one case also with a chromalveolate (for six examples, see Figure <xref ref-type="fig" rid="F2">2</xref>). This suggests these are relatively ancient introns and perhaps date back to the last common ancestor of all eukaryotes. This degree of conservation is high, taking into account data such as those of Rogozin et al., who calculated that approximately 20% of the introns in <italic>Plasmodium </italic>are shared by at least one of the other genomes analysed (Human, <italic>Anopheles</italic>, <italic>Drosophila</italic>, <italic>Caenorhabditis, Arabidopsis</italic>, <italic>Schizosaccharomyces </italic>and <italic>Saccharomyces</italic>), and that 25% of the human introns are shared by <italic>Arabidopsis </italic>[<xref ref-type="bibr" rid="B31">31</xref>]. It is also possible that shared intron positions are due to independent gains, but it is very unlikely that the observed level of shared positions (about 50%) resulted from parallel gains, in particular in the many cases where the intron is found in several of the major lineages of eukaryotes. Whether intron gains or losses predominate in eukaryotic evolution is still a subject of controversy. Recently, several studies using different analytic approaches and datasets addressed this question with varied results, but in all cases, they show that ancestral conservation accounts for the large majority of shared positions [<xref ref-type="bibr" rid="B32">32</xref>-<xref ref-type="bibr" rid="B36">36</xref>]. The degree of conservation observed in <italic>Streblomastix </italic>intron positions suggests two things. First, it suggests that the ancestor of excavates was relatively intron rich and retained a large number of ancient introns, many of which were subsequently lost in the genomes where we have the most information, such as trypanosomes, <italic>Giardia </italic>and <italic>Trichomonas</italic>. This assumes the relationship between oxymonads and other hypothesized excavates is correct, but this is not certain and oxymonads lack the morphological trait used to define excavates (the ventral groove). However, other ultrastructural characters [<xref ref-type="bibr" rid="B37">37</xref>] as well as molecular phylogenies have shown a close affiliation between oxymonads and <italic>Trimastix </italic>[<xref ref-type="bibr" rid="B38">38</xref>], a free living flagellate that does have excavate characteristics [<xref ref-type="bibr" rid="B5">5</xref>,<xref ref-type="bibr" rid="B39">39</xref>]. Multi-gene phylogenies also lend additional support for a common origin of the lineages leading to oxymonads, diplomonads and parabasalia [<xref ref-type="bibr" rid="B21">21</xref>]. The second implication of this data is that intron gain and loss have taken place very slowly in the lineage leading up to <italic>Streblomastix</italic>: if intron turnover were rapid, then we would expect a low proportion of ancient introns to remain unless ancient intron positions were under some selection to be retained. While this is probably true in a few individual cases where introns have acquired some function in the control of gene expression, there is presently no evidence either for or against this as a common feature of ancient introns. None of these shared introns are known from either <italic>Giardia </italic>or <italic>Trichomonas</italic>, so any potential function is clearly dispensable, although it is interesting to note that the <italic>rps9 </italic>intron 1 has been retained by the <italic>G. theta </italic>nucleomorph, which is very intron poor, having kept a total of only 17 introns [<xref ref-type="bibr" rid="B40">40</xref>].</p><fig position="float" id="F2"><label>Figure 2</label><caption><p>Sequence logos showing conservation at intron borders. Top: 5’ splice site (position 1) and surrounding sequence. Bottom: 3’ splice site (-1) and surrounding sequence. Logos were made using Weblogo (<ext-link ext-link-type="uri" xlink:href="http://weblogo.berkeley.edu"/>).</p></caption><graphic xlink:href="1471-2148-6-34-2"/></fig></sec></sec><sec><title>Conclusion</title><p>The present sampling of protein-coding gene sequences from <italic>Streblomastix </italic>suggests that oxymonad genomes contain a relatively large number of canonical splicesomal introns, many of which are at ancient conserved positions. This is in contrast to the better studied excavate genomes such as those of kinetoplastids, <italic>Giardia </italic>and <italic>Trichomonas </italic>where canonical spliceosomal introns are either rare or have been co-opted in specific ways, such as the spliced leaders in euglenozoa. The fact that many <italic>Streblomastix </italic>introns are ancient shows that the genome of the ancestor of these organisms, and indeed probably all extant eukaryotes, contained many introns and that the intron-poor state found in <italic>Giardia </italic>and <italic>Trichomonas </italic>is more likely independently derived.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>cDNA library construction and EST sequencing</title><p>Termites were collected from a rotten log in Point Grey, Vancouver, Canada. The whole hindgut content of about 60 individuals of <italic>Zootermopsis angusticollis </italic>from a single colony was collected and total RNA was extracted using TRIZOL (Invitrogen). A directionally cloned cDNA library was constructed (Amplicon Express) and 5,337 clones were sequenced from the 5' end. ESTs were trimmed for vector and quality, and assembled into clusters by PEPdb <ext-link ext-link-type="uri" xlink:href="http://amoebidia.bcm.umontreal.ca/public/pepdb/agrm.php"/>.</p></sec><sec><title>Identification and genomic characterisation of <italic>Streblomastix </italic>genes</title><p><italic>Streblomastix </italic>sequences were recovered from EST data by identifying protein coding sequences containing in-frame TAA and TAG stop codons. Putatively stop-coding containing mRNAs were re-sequenced in both strands. In cases where cDNA clones were truncated, the sequences were extended by means of 3' and 5' RACE (Ambion) using total termite hindgut RNA. The genomic sequence for each mRNA was amplified using specific primers corresponding to the ends of each complete or partial cDNA and PCR-amplified using genomic DNA purified from the termite hindgut content. All PCR products were cloned using TOPO and sequenced both strands. Accession numbers for new sequences are [genbank<ext-link ext-link-type="gen" xlink:href="DQ363664">DQ363664</ext-link>, genbank<ext-link ext-link-type="gen" xlink:href="DQ363665">DQ363665</ext-link>, genbank<ext-link ext-link-type="gen" xlink:href="DQ363666">DQ363666</ext-link>, genbank<ext-link ext-link-type="gen" xlink:href="DQ363667">DQ363667</ext-link>, genbank<ext-link ext-link-type="gen" xlink:href="DQ363668">DQ363668</ext-link>, genbank<ext-link ext-link-type="gen" xlink:href="DQ363669">DQ363669</ext-link>, genbank<ext-link ext-link-type="gen" xlink:href="DQ363670">DQ363670</ext-link>, genbank<ext-link ext-link-type="gen" xlink:href="DQ363671">DQ363671</ext-link>, genbank<ext-link ext-link-type="gen" xlink:href="DQ363672">DQ363672</ext-link>, genbank<ext-link ext-link-type="gen" xlink:href="DQ363673">DQ363673</ext-link>, genbank<ext-link ext-link-type="gen" xlink:href="DQ363674">DQ363674</ext-link>, genbank<ext-link ext-link-type="gen" xlink:href="DQ363675">DQ363675</ext-link>, genbank<ext-link ext-link-type="gen" xlink:href="DQ363676">DQ363676</ext-link>, genbank<ext-link ext-link-type="gen" xlink:href="DQ363677">DQ363677</ext-link>, genbank<ext-link ext-link-type="gen" xlink:href="DQ363678">DQ363678</ext-link>, genbank<ext-link ext-link-type="gen" xlink:href="DQ363679">DQ363679</ext-link>].</p></sec></sec><sec><title>Authors' contributions</title><p>CHS analysed the EST data, performed PCR and sequencing, and examined conservation of intron positions in other organisms. PJK collected the termites and purified RNA for library construction. Both authors participated in the writing and editing of the manuscript. All authors read and approved the final manuscript.</p></sec> |
Evolution of Curative Therapies For Atrial Fibrillation | Could not extract abstract | <contrib contrib-type="author"><name><surname>Khasnis</surname><given-names>Atul</given-names></name><degrees>MD</degrees></contrib><contrib contrib-type="author"><name><surname>Veerareddy</surname><given-names>Srikar</given-names></name><degrees>MD</degrees></contrib><contrib contrib-type="author"><name><surname>Jongnarangsin</surname><given-names>Krit</given-names></name><degrees>MD</degrees></contrib><contrib contrib-type="author"><name><surname>Ip</surname><given-names>John H</given-names></name><degrees>MD</degrees></contrib><contrib contrib-type="author"><name><surname>Abela</surname><given-names>George S</given-names></name><degrees>MD</degrees></contrib><contrib contrib-type="author"><name><surname>Thakur</surname><given-names>Ranjan K</given-names></name><degrees>MD</degrees></contrib> | Indian Pacing and Electrophysiology Journal | <sec sec-type="" id="s1"><title>Introduction</title><p>Atrial fibrillation (AF) is one of the most common arrhythmias seen in clinical practice. The incidence of AF increases with age and is seen in up to 8.9% of individuals greater than 80 years of age [<xref ref-type="bibr" rid="R1">1</xref>]. Although many treatment modalities are available for AF, curative therapy has recently become possible for some patients with atrial fibrillaton. This is partly due to the complexity and incomplete understanding of this arrhythmia. Pharmacological therapies for the management of AF have many limitations such as side effects, cost, inconvenience, breakthrough arrhythmias, drug-drug interactions and continued need for anticoagulation. An improved understanding of the pathophysiology and electrophysiologic basis of AF coupled with the availability of better mapping techniques as well as advances in catheter technology offer promise for curative treatment. This article will briefly review the understanding of the mechanisms of AF and then the evolution of catheter techniques for control and elimination of AF.</p></sec><sec sec-type="" id="s2"><title>Mechanisms of AF</title><p>The understanding of mechanisms underlying the initiation and maintenance of AF has evolved over the last many decades. The earliest concept of reentry was proposed by Winterberg in 1906 [<xref ref-type="bibr" rid="R2">2</xref>] and Lewis in 1912 advocated that rapid focal activity from one or more centers accounted for AF [<xref ref-type="bibr" rid="R3">3</xref>]. Mines in 1913 showing that the mechanism of reentry was an impulse circling a large anatomical obstacle [<xref ref-type="bibr" rid="R4">4</xref>]. Scherf in 1947 revived the theory of focal trigger in AF [<xref ref-type="bibr" rid="R5">5</xref>]. The work of Moe in 1960s supported the theory of randomly propagating multiple wavelets as the main mechanism underlying AF [<xref ref-type="bibr" rid="R6">6</xref>]. The reentrant wavelet hypothesis required the concept of “wavelength” of the arrhythmia circuit to be introduced. The minimum circuit wavelength as defined by the tachycardia wavelength is the product of the tissue refractory period and wavefront conduction velocity. Any circuit wavelength smaller than the minimum will cause the “head” of the wavefront to catch up with its “tail” thus terminating the reentry. In the 1970s, Allessie introduced the concept of “leading circle reentry”. In a goat model of AF, he demonstrated that the average circuit diameter was 20 - 30 mm and that a minimum of 5-8 random wavelets were required to sustain AF at any given time. Two types of reentry were seen during the study: “leading circle” reentry and “random” reentry [<xref ref-type="bibr" rid="R7">7</xref>]. All these theories of random reentry wavelets explained the sustenance of AF once it occurred, but an explanation of its initiation was also warranted. Allessie et al. offered several possible explanations: a “stable background circuit” capable of initiating new AF when the earlier episode dies out, abnormal focal trigger sites in the atria and the possibility of an echo beat from the AV node or from an accessory pathway. The present understanding is that AF requires a “critical atrial mass” needed to maintain the arrhythmia and that there is a critical rate above which organized atrial activity cannot continue [<xref ref-type="bibr" rid="R8">8</xref>]. Thus, at a certain rate, organized atrial activity can disintegrate into atrial fibrillation provided the critical tissue mass is available to sustain it. Recent studies in isolated human atrial preparations showed that a single meandering functional reentrant wavefront produced atrial fibrillation [<xref ref-type="bibr" rid="R9">9</xref>]. Recent work by Jalife and colleagues questions the randomness of atrial activity in atrial fibrillation [<xref ref-type="bibr" rid="R10">10</xref>]. Their study suggests the presence of a possible “mother circuit” that serves as a periodic background focus; the presence of anatomical obstacles (scar, orifices) serve to break up the wavefront from the “mother circuit” into multiple wavelets that spread in various directions. Wu et al. have proposed the role of pectinate muscles as obstacles that break the activation wave thus promoting reentry. They may also serve as an anchoring site for the wave leading to rotor like activity [<xref ref-type="bibr" rid="R11">11</xref>]. The likelihood that focal activation plays some role in atrial fibrillation is now well accepted. Haissaguerre et al. reported arrhythmogenicity of the pulmonary veins as possible focal triggers in some cases of atrial fibrillation [<xref ref-type="bibr" rid="R12">12</xref>]. The myocardial sleeves that extend from the left atrium onto the pulmonary veins appear to be the pathological correlate of the arrhythmogenic focus. Thus there appears to be a role for both reentry and focal activity in the electrophysiology of AF. The question of randomness of AF is also not settled at present. Treatment modalities aim at elimination of both mechanisms underlying AF. The following sections discuss how the above discoveries have channeled catheter therapy for AF.</p></sec><sec sec-type="" id="s3"><title>Mapping AF</title><p>The discussion of evolution of catheter techniques would be impossible without inclusion of the development of mapping techniques [<xref ref-type="bibr" rid="R13">13</xref>]. However a detailed description is beyond the scope of this article. There has been a tremendous impact of mapping technology on diagnostic and therapeutic electrophysiology. The mapping of AF has helped clarify its mechanism and localize possible anatomical sites for effective radiofrequency ablation. Conventionally, this has been done by careful correlation of 12-lead surface ECG with intracardiac data detected by catheters lying in contact with the endocardium in various cardiac chambers. However, these methods do not cover a vast area of the endocardial surface and spot-by-spot maneuvering of the catheter is required to trace a focus and determine the sequence of impulse spread to surrounding areas. Another drawback of this technique is inability to identify several sites and then return to the most optimal site. Hence it is important to achieve a three dimensional orientation of the focus and activation sequence to be able to localize therapy. Electroanatomic or CARTO™ mapping is a nonfluoroscopic mapping system that uses magnetic technology to accurately determine the location and orientation of the mapping and ablation catheter while simultaneously recording local electrograms from the catheter tip. The 3-D geometry of the mapped chamber is reconstructed in real-time and the optimal sites for ablation are analyzed. The timing of the reference electrogram is used to determine the activation timing in relation to the acquired points and therefore collection of data during the corresponding parts of the cardiac cycle is vital to the performance of the system. Noncontact Mapping using the EnSite3000™ (Endocardial Solutions, St Paul, MN) mapping system consisting of balloon or multielectrode array (MEA) has been studied extensively. This technique is based on the principle that endocardial activation produces a chamber voltage field, which obeys Laplace’s equation. Noncontact intracavitary electrodes are used to detect the potential field on the endocardial surface. An inverse Laplace’s equation is used to improve accurate and stable reconstructions of the recorded potentials. The activation points are displayed as computed electrograms or isopotential maps. A three-step process of establishing geometry, identifying the area of interest and navigating the ablation catheter to this area is used to map and treat arrhythmias. Focal atrial fibrillation arising from both right atrium and pulmonary veins has been successfully ablated using this system [<xref ref-type="bibr" rid="R14">14</xref>]. Other techniques used include the Basket and amplification technique. In order to characterize the local electrical activity of the heart, a "basket" catheter is used with 8 splines carrying 8 electrodes each (EPT, Boston Scientific). The electrodes are coupled two by two to achieve bipolar recordings thus producing 32 bipolar recordings. Each electrode couple is then amplified and filtered separately for every channel (CardioLab System, Prucka Engineering, Huston, USA). Amplification gain is normally set between 500 and 2500 to achieve best signal reading; filters are set between 30 and 500 Hz. The signals are then sampled at 1 kHz and stored on magneto-optical device. Intracardiac echocardiography (ICE) can be a valuable tool in localizing anatomical areas for ablation. It allows for assessment of wall contact of ablation catheters for creation of long linear lesions for catheter ablative treatment of atrial fibrillation. Epstein et al. demonstrated the superiority of ICE over fluoroscopy in treating atrial fibrillation [<xref ref-type="bibr" rid="R15">15</xref>]. Inverse electrocardiography, an established formulation is the imaging of the activation time map on the entire surface of the heart from ECG mapping data [<xref ref-type="bibr" rid="R16">16</xref>,<xref ref-type="bibr" rid="R17">17</xref>]. This enables reconstruction of unifocal, multifocal and more distributed activation patterns. It can distinguish between areas with early and late activation. This novel approach is presently under clinical evaluation and development. A new non-invasive technique for the characterization of time-dependent spectral properties of atrial fibrillation in the surface ECG using the Wigner-Ville distribution for time-frequency analysis and the cross Wigner-Ville to compute trends describing instantaneous frequency of the atrial activity has also been described. Preliminary results indicate that short-term variations exist in the fibrillation cycle lengths and that the variations can exhibit similar behavior in the leads V1-V3 [<xref ref-type="bibr" rid="R18">18</xref>]. Magnetic resonance imaging has also shown promise in demonstrating pulmonary venous anatomy which is central to the technique of radiofrequency ablation of focal AF [<xref ref-type="bibr" rid="R19">19</xref>]. The last three techniques are non-invasive.</p></sec><sec sec-type="" id="s4"><title>Radiofrequency catheter ablation of AF</title><p>Treatment of AF using radiofrequency techniques can be curative or palliative. The curative strategies will be discussed in this review. The palliative procedures will be briefly be touched upon.</p><sec id="s4a"><title>Curative techniques: The Maze Operation - ablation of reentrant AF</title><sec id="s4a1"><title>Surgical Maze procedure</title><p>The understanding of the electrophysiology of AF as consisting of multiple random circulating reentrant wavelets was the basis for the development of the Maze operation. The originally devised Maze operation was an open heart surgical procedure that involved creating multiple surgical incisions (the maze) in both atria aimed at interruption of reentry of random wavefronts while permitting the impulse to travel in a channeled direction to the AV node. This operation underwent serial modifications beginning with the left atrial isolation procedure in 1980 and progressing to the Maze III operation which resulted in higher incidence of postoperative sinus rhythm, improved long term sinus node function, decreased need for pacing, less arrhythmia recurrence and improved long-term atrial transport function. The Maze III procedure can now be successfully performed as a minimally invasive procedure through a submammary incision. The epicardial application of radiofrequency energy for production of lesions without cardiopulmonary bypass was an attempt at reducing morbidity and complications associated with the Maze procedure.</p></sec><sec id="s4a2"><title>Maze procedure using radiofrequency energy</title><p>The use of radiofrequency catheter techniques for selective right or left atrial modification involves the creation of linear transmural atrial lesions by endocardial radiofrequency current application. The right atrial modification technique has met with good success in cases where the AF is triggered by a short run of atrial flutter or in cases when a single macro-reentrant circuit is located in the right atrium and has an unstable cycle length. The site of application of radiofrequency current in the right atrium varies from a single midseptal line to the region of the Bachmann’s bundle to the posterior intercaval right atrium in experimental models. On the other hand, left atrial modification has not been rewarding because of faulty technique and increased risk of stroke. In experimental studies, left posterior atrial modification with rapid atrial pacing and left atrial lines (in mitral regurgitation model) has been reported successful. The left atrium is now frequently the target for curative AF ablation since the reporting of pulmonary veins as triggers for AF (pulmonary vein ablation is discussed later in the article). A number of early investigators reported success with use of radiofrequency energy to create long linear lesions in both animals and humans. Elvan et al. reported successful cure of AF by application of radiofrequency current to five epicardial sites and one endocardial site (coronary sinus) in a canine model of AF induced by burst atrial pacing [<xref ref-type="bibr" rid="R20">20</xref>]. The problem with earlier catheter design was the lack of good tissue contact leading to the production of discontinuous lesions which could themselves be proarrhythmic. Development in catheter technology and mapping has overcome this shortcoming. Olgin et al. reported the use of intracardiac echocardiography in the pig heart for anatomical localization aiding placement of intraatrial catheters for ablation [<xref ref-type="bibr" rid="R21">21</xref>]. This also helped achieve good tissue contact and successful transmural ablation as confirmed by histologic evaluation. Kalman et al. showed that the production of linear atrial lesions reduced the energy requirement for cardioversion of AF and in some cases to its termination [<xref ref-type="bibr" rid="R22">22</xref>]. Swartz et al. reported creating linear atrial lesions in humans with chronic AF using the drag technique [<xref ref-type="bibr" rid="R23">23</xref>]. The main disadvantages were long procedure and fluoroscopy time. In 1997, Patwardhan et al [<xref ref-type="bibr" rid="R24">24</xref>]. reported success of the Maze procedure using radiofrequency bipolar coagulation as a means of energy to produce atrial lesions. Their study population included patients with rheumatic heart disease and atrial fibrillation. They used radiofrequency microbipolar coagulation to produce conduction blocks along incision lines in keeping with the surgical Maze III procedure as an adjunct to valve surgery in 18 patients in atrial fibrillation undergoing surgery for rheumatic valvular disease. The control population consisted of those 26 patients who underwent corrective valve surgery alone. The energy was delivered by a bayonet type bipolar forceps with an active tip length of 7 mm drawing current from a microbipolar port of Valleylab Force 4 electrosurgical unit (Valleylab, Boulder, CO). A 3-mm retinal handheld cryoprobe using nitrous oxide gas was used for cryoablation. 15/18 patients in the MAZE group were followed-up. 12/15 (80%) patients converted to normal sinus rhythm. Pulsed wave Doppler evaluation at follow-up showed return of atrial transport function, presence of a wave in all these patients in tricuspid valve flow and in nine (75%) patients in mitral valve flow. The procedure took 11.62 +/- 3.86 min of elective cardioplegic arrest time for ablation line production in the left atrium and 18.71 +/- 4.25 min of cardiopulmonary bypass time during reperfusion for ablation line production in the right atrium. Only 1/23 patients in the control (4.3%) converted to normal sinus rhythm. The investigators concluded that their modification considerably shortened the procedure for the Maze procedure and was effective in restoring normal sinus rhythm in 80% of the patients. Haissaguerre et al. reported successful ablation of AF using a 7-F specially designed 14-polar catheter with interelectrode distance of 3 mm to create linear lesions in the right atrium using radiofrequency energy [<xref ref-type="bibr" rid="R25">25</xref>]. Calkins et al. performed a Maze-like procedure using the Guidant Heart Rhythm Technologies Linear Ablation System to create long transmural lesions [<xref ref-type="bibr" rid="R26">26</xref>]. The unique features of this system included the availability of different pre-shaped multi-electrode steerable ablation catheters, the use of multi-phased radiofrequency energy and control of RF output by varying the duty cycle. Fourteen out of 15 patients enrolled in this study had an acute successful outcome.</p></sec><sec id="s4a3"><title>Ablation of focal atrial fibrillation</title><p>Since the early part of this century, the role of focal sources of automaticity in sustaining AF was deemed significant. The application of aconitine to the right atrial appendage resulting in the generation of AF and its ligation resulting in restoration of sinus rhythm was amongst the earliest experimental models for this arrhythmia. Degeneration of rapid atrial tachycardia into AF or onset of AF following multiple atrial ectopic beats offers evidence for this hypothesis. The entity flutter-fibrillation may also possibly result from disorganization of regular atrial activity. However, the pulmonary veins were not clearly implicated in the pathogenesis.</p><p>The hypothesis that pulmonary veins and especially the pulmonary vein-left atrial junction as a focal source of AF is supported by anatomical and pathology studies. In 1966, Nathan and Eliakim reported that the proximal portion of the PV has a sleeve of myocardium that is a direct extension from the adjacent atrial tissue and that is electrically coupled to the atrium in an anatomic study of the left atrium-pulmonary vein junction in human hearts [<xref ref-type="bibr" rid="R27">27</xref>]. Cabrera et al. studied pulmonary venous anatomy using high-frequency intravascular ultrasound (IVUS) [<xref ref-type="bibr" rid="R28">28</xref>]. They obtained cross-sectional IVUS images with a 3.2F, 30-MHz ultrasound catheter at intervals on each vein. Histological cross-sections were compared with ultrasonic images. The pulmonary venous wall at the venoatrial junction revealed a 3-layered ultrasonic pattern consisting of inner echogenic layer representing both endothelium and connective tissue of the media, the middle hypoechogenic stratum corresponded to the sleeves of left atrial myocardium surrounding the external aspect of the venous media and an outer echodense layer of fibro-fatty adventitial tissue. The middle layer was thickest at the venoatrial junction and decreased toward the lung hilum. They found close agreement among the IVUS and histological measurements for maximal luminal diameter and maximal muscular thickness. Ho et al. explored the characteristics of normal pulmonary veins so as to provide more information relevant to radiofrequency ablation [<xref ref-type="bibr" rid="R29">29</xref>]. They grossly examined 20 structurally normal heart specimens. Histological sections were made from 65 pulmonary veins. In this study, the longest myocardial sleeves were found in the superior veins. The sleeves were thickest at the venoatrial junction in the left superior pulmonary veins. For the superior veins, the sleeves were thickest along the inferior walls and thinnest superiorly and composed mainly of circularly or spirally oriented bundles of myocytes with additional bundles that were longitudinally or obliquely oriented, sometimes forming mesh-like arrangements. Fibrotic changes estimated at between 5% and 70% across three transverse sections were seen in 17 veins. The longest myocardial sleeves in the superior pulmonary veins in this study correlated with the observation reported by Haissaguerre et al. that the commonest source of the ectopic beats triggering AF were in these veins [<xref ref-type="bibr" rid="R12">12</xref>]. The sites mapped during electrophysiology study were 2-4 cm inside the veins, which is longer than the extent of sleeves found in the specimens. These differences were explained as probably due to the indiscrete nature of the venoatrial junction in the specimens, and fixation causing shrinkage of tissues, thus reducing the dimensions. Saito et al. studied thirty-nine human autopsy hearts; 22 with AF and 17 without atrial arrhythmias with intent to provide a detailed anatomy of the extensions of left atrial myocardium onto the pulmonary veins [<xref ref-type="bibr" rid="R30">30</xref>]. They observed that the peripheral zones of myocardial sleeves were associated with increasing connective tissue deposition between myocardial muscle groups and suggested that this was a degenerative change that, from the histologic viewpoint, fitted with progressive ischemia. They also suggested that these changes could provide a basis for microreentry and atrial arrhythmias. Lin et al. studied the structure of the pulmonary veins in patients with paroxysmal AF that is initiated by PV ectopic beats [<xref ref-type="bibr" rid="R31">31</xref>]. They reported nonspecific dilatation of the ostia and proximal portion of superior pulmonary veins in these patients. They found that the superior PVs were significantly dilated and were the commonest source of the ectopic beats initiating paroxysmal AF. However, site-specific dilatation of the PVs was not evident. Although there was simultaneous dilatation of the PVs and LA, the nonlinear relationship between PV diameter and LA size suggested that the extent of the dilation of the LA and PV ostia was different; it was proposed to be due to the different compliance of the PV vascular wall and LA wall among individuals. The electrical activity in the PVs was presumed to be a result of this extension of cardiac musculature. It is possible that dyssynchronous contraction of the muscle at the atriopulmonary venous junction during rapid and chaotic firing of the ectopic focus may account for an increase in the dimensions of the atriopulmonary venous junction. This study suggested that the ectopic beats might be initiated by the stretch mechanism. Satoh et al. demonstrated that increased atrial stretch in dogs might induce triggered activity resulting in arrhythmia [<xref ref-type="bibr" rid="R32">32</xref>]. This dilatation of the pulmonary vein ostia has implications for catheter ablation techniques. An embryological explanation has also been provided for the origin of ectopic foci. Blom et al. studied the development of the cardiac conduction system with the use of HNK-1 immunohistochemistry in human embryos ranging in age from 42 to 54 days of gestation [<xref ref-type="bibr" rid="R33">33</xref>]. They hypothesized that in patients with abnormal atrial automaticity, the ectopic pacemaker sites correspond to areas of embryonic myocardium with an early phenotypic differentiation, as indicated by differences in antigen expression during normal cardiac development.</p><p>In 1994, Konings et al. classified AF based on the patterns of atrial activation during AF [<xref ref-type="bibr" rid="R34">34</xref>]. The Type I AF in their classification, single broad wave fronts propagated uniformly across the RA. This suggests a possible focal origin of AF. In 1997, Jais et al. reported nine patients with paroxysmal focal AF [<xref ref-type="bibr" rid="R35">35</xref>]. All were free of structural heart disease and had frequent episodes of AF despite the use of antiarrhythmic drugs. AF was associated with runs of irregular atrial tachycardia or monomorphic extrasystoles. Electrophysiological study demonstrated that all the atrial arrhythmias were due to the same focus firing irregularly and exhibiting a consistent and centrifugal pattern of activation. Three foci were found to be located in the right atrium, two near the sinus node and one in the ostium of the coronary sinus. Six others were located in the left atrium at the ostium of the right pulmonary veins and at the ostium of the left superior pulmonary vein. Haissaguerre et al. in 1998 demonstrated that the pulmonary veins are an important source of spontaneous ectopic beats, initiating frequent paroxysms of AF [<xref ref-type="bibr" rid="R12">12</xref>]. The study population consisted of 45 consecutive patients with AF resistant to more than two drugs, at least one episode of AF every two days, receiving anticoagulant treatment, and frequent isolated atrial ectopic beats (more than 700 per 24 hours). Antiarrhythmic medications were discontinued before hospitalization. In 37 patients, at least one episode of sustained AF initiation lasting more than one minute was documented: the ectopic beat initiating AF had a short coupling interval (a P-on-T pattern) and morphologic features similar to those of isolated ectopic beats. Their likely identical origin was confirmed later by intracardiac mapping data. Three multielectrode catheters were placed for the electrophysiology study: one quadripolar roving ablation catheter with a thermocouple, one in the right atrial appendage (to map right atrial and right-pulmonary-vein foci) or coronary sinus (for left-pulmonary-vein foci) to provide stable reference electrograms during mapping, and one for stimulation. If the arrhythmia did not spontaneously develop during electrophysiologic study or was not sufficiently sustained, physiologic procedures (e.g., Valsalva's maneuver or carotid-sinus massage), atrial pacing, pharmacologic agents or all three methods were tried. The preliminary study involved mapping of isolated ectopic beats. The ectopic focus was localized at the earliest atrial activity relative to the reference electrogram or the onset of the ectopic P wave on the surface ECG. Mechanically produced beats were excluded from the analysis by comparing the electrocardiographic pattern and intracardiac sequence with the confirmed spontaneous ectopic beats. If no sharp bipolar right atrial activity was recorded less than 10 ms before the onset of the ectopic P wave, the beats were considered to have originated in the left atrium. Direct mapping of the left atrium and pulmonary veins was then performed. The role of ectopic beats in the initiation of AF was confirmed by on-site recording of a paroxysm of AF. Ablation was performed at the site with the earliest recorded ectopic activity. Telemetry and 24-hour Holter monitoring were performed to identify the cumulative duration of AF and its frequency. Patients were discharged and given oral anticoagulants for at least three months but no antiarrhythmic drugs were prescribed. Late follow-up consisted of out-patient visits and Holter recordings every three months. A single origin of ectopic beats was identified in 29 patients, two in 9 patients, three in 6 patients, and four in 1 patient. Ectopic beats originated in atrial muscle in 4 patients (in the right atrium in 3 and the posterior left atrium in 1) and in the pulmonary veins in 41 patients [94 percent]): 31 foci in the left superior, 17 in the right superior, 11 in the left inferior, and 6 in the right inferior pulmonary vein. The venous origin of the earliest ectopic activity was demonstrated in 23 patients by the radiographic position of the mapping catheter, which was confirmed by angiographic visualization. The earliest local activity was traced to a point 2 to 4 cm within the main pulmonary vein or one of its proximal branches as evidenced by earliest activation deep in the vein and progressively later toward the ostium and the left atrial exit, resulting in distal-to-proximal venous activation during multipolar recordings. AF was initiated by a single focal discharge in 3 patients, short burst of two or more repetitive focal discharges in 40, and both mechanisms in 2 patients. During sinus rhythm, the activation sequence occurred first in the left atrium and progressively later inside the vein, which was the opposite of the sequence during ectopy. Successful ablation of ectopic foci in the hospital was achieved in 38 patients. Pappone et al. developed an anatomic approach aimed at isolating each PV from the left atrium (LA) by circumferential radiofrequency (RF) lesions around their ostia [<xref ref-type="bibr" rid="R36">36</xref>]. They selected 26 patients with resistant AF, either paroxysmal or permanent. A nonfluoroscopic mapping system was used to generate 3D-electroanatomic LA maps and deliver RF energy. Two maps were acquired during coronary sinus and right atrial pacing to validate the lateral and septal PV lesions, respectively. Patients were followed up closely for 6 months. Among 14 patients in AF at the beginning of the procedure, 64% had sinus rhythm restoration during ablation. PV isolation was demonstrated in 76% of 104 PVs treated by low peak-to-peak electrogram amplitude (0.08±0.02 mV) inside the circular line and by disparity in activation times across the lesion. After 9±3 months, 22 patients (85%) were AF-free, including 62% not taking and 23% taking antiarrhythmic drugs, with no difference (P=NS) between paroxysmal and permanent AF.</p><p>However, despite all the technological advances, the technique is still far from perfect and is fraught with complications. Ablation of the pulmonary vein focus for AF is a long-lasting procedure with possible complications, such as pulmonary vein stenosis, thromboembolism, air embolism, hemopericardium, and possible damage to such adjacent structures such as bronchioles, the right pulmonary artery, and lung tissue. Paroxysmal AF can also recur. According to Wellens, the following are the problems with the current technique [<xref ref-type="bibr" rid="R37">37</xref>]. First, the ectopic foci, most commonly found in the superior pulmonary veins and are characterized by a sharp electrogram, which may or may not be conducted to the atrium. A rapidly firing focus may be the mechanism for paroxysmal AF, but 1 or 2 pulmonary vein ectopic beats may also initiate AF in the presence of an additional substrate in the left or right atrium. These potentials have to be differentiated from those of the ligament of Marshall and the right atrium, which can also be foci of paroxysmal AF. Accurate mapping is helped by use of multielectrode catheters or electroanatomic mapping systems and will hopefully be facilitated by clear demonstration of the premature P wave polarity. Unfortunately, ectopic pulmonary vein ectopy may not be present during the electrophysiological study, even with provocative pharmacological or pacing procedures. The second problem is about the energy source for ablation. Currently, most experience is with heat with radiofrequency (RF) energy. Studies are ongoing on the appropriate RF power, heat limits, pulsed versus continuous RF energy delivery, and the use of irrigated-tip RF ablation. This has led to the development of ultrasound delivered through a balloon in the pulmonary vein, laser, and cryoablation. More knowledge is needed about the histopathological consequences of these approaches. Third, much activity is going on in the development of the appropriate shape of the catheter at the site where the energy has to be delivered. The purpose of ablation is to eliminate the high-frequency pulmonary vein potential and the creation of bidirectional block in the pulmonary vein. To prevent ectopic pulmonary vein activity from entering the left atrium, a circular lesion in or around the orifice of the pulmonary vein seems a logical approach. Obviously, technical developments have to occur to speed up localization and isolation or eradication of ectopic activity. The last problem concerns the long-term results of pulmonary vein ablation in AF. The short-term results of pulmonary vein ablation are promising but uncertainty exists about long-term results with regard to arrhythmia recurrence and mechanical complications such as development of pulmonary vein stenosis. Also, the best way to evaluate the possible development of pulmonary vein stenosis needs to be established: echocardiography, CT, MRI, or angiography. There also remains the question of reversibility of electrophysiological changes in the atrium induced by AF (“electrical remodeling”). Wijffels et al. have demonstrated this phenomenon in the goat heart and this has been confirmed by other investigators in the human heart [<xref ref-type="bibr" rid="R38">38</xref>]. It is unknown up to what time point in the natural history of AF those changes are still reversible and the patient can be helped by removal of the initiating trigger.</p><p>The coronary sinus has also been suggested to have a role in the origin and sustenance of AF. Antz et al. studied the presence of electrical connections between the coronary sinus musculature and the right and left atrial myocardium, forming a RA-LA connection [<xref ref-type="bibr" rid="R39">39</xref>]. Gerlis et al. reported two cases of ventricular pre-excitation due to accessory pathways histologically identified as being associated with coronary sinus aneurysms [<xref ref-type="bibr" rid="R40">40</xref>]. They noted that the proximal coronary sinus is surrounded with a spiral myocardial sheath that stops abruptly at or shortly beyond the orifices of the entering coronary veins, including the great cardiac vein, and is continuous with that of the morphological right atrium. This is a remnant of sinus venosus musculature. Recently, Katritsis et al. recorded double potentials within the CS, particularly the distal superoposterior part, near the left superior pulmonary vein [<xref ref-type="bibr" rid="R41">41</xref>]. The prevalence was higher in patients with PAF than in subjects with other or no arrhythmia. Hence, they proposed that their presence denotes possible sources or substrate for atrial arrhythmia.</p><p>The ligament of Marshall is another focus of AF that has been described in the literature. It consists of multiple sympathetic nerve fibers, ganglia, blood vessels and multiple myocardial tracts (Marshall Bundles) insulated by fibrofatty tissue. Kim et al. studied seven hearts at postmortem to study the anatomy of the ligament of Marshall [<xref ref-type="bibr" rid="R42">42</xref>]. They concluded that the ligament of Marshall in human hearts is innervated by sympathetic nerve fibers and has multiple myocardial tract insertions into the left atrial free wall and CS, forming a substrate of reentry. Polymeropoulos et al recently reported a 66-year-old woman with a history of typical atrial flutter and atrial fibrillation in whom endocardial recordings during tachycardia from this region showed a discrete electrical potential (Marshall potential) preceding the atrial electrogram [<xref ref-type="bibr" rid="R43">43</xref>]. Radiofrequency ablation was successfully performed.</p></sec></sec><sec id="s4b"><title>Palliative radiofrequency procedures for AF</title><p>Permanent AF is usually not amenable to ablation procedures and may require multiple pharmacological agents for management. In some of these patients, medical therapy is poorly tolerated or unsuccessful. In these cases, atrioventricular node (AV node) ablation combined with permanent pacemaker implantation is a possible treatment strategy with rate control as the goal. The pacemaker may be single chamber (VVI) for chronic AF or dual chamber (DDD) for paroxysmal AF. This procedure is successful in almost 100% cases and late recovery of AV conduction is rare. However, this procedure still mandates anticoagulation and possibly requires anti-arrhythmic therapy. Olgin and Scheinman compared high-energy direct current and radiofrequency catheter ablation for ablation of the atrioventricular junction [<xref ref-type="bibr" rid="R44">44</xref>]. They concluded that radiofrequency energy is as efficacious as and safer then high energy current. Radiofrequency ablation of the AV junction may be performed using either a left or right-sided approach. Kalbfleisch et al. prospectively compared the left sided approach with persistent attempts from the right side in patients in whom initial radiofrequency applications on the right side were unsuccessful [<xref ref-type="bibr" rid="R45">45</xref>]. The left sided approach required significantly fewer radiofrequency applications after randomization than the right-sided approach. The Ablate and Pace Trial studied the prevalence and characteristics of escape rhythms after radiofrequency ablation of the AV junction [<xref ref-type="bibr" rid="R46">46</xref>]. This was a prospective, multicenter registry of AF patients undergoing AV node ablation and pacing. Before discharge from the hospital, the patients underwent a systematic analysis of the rate and morphologic features of the escape rhythm, if any that was present when the pacing rate was gradually decreased. This study reported that majority of the patients who undergo radiofrequency ablation of the AV node are pacemaker dependent after the procedure, as defined by lack of an escape rhythm that is < 40 bpm. Reports of ventricular arrhythmias and sudden death after ablation have raised concerns about safety. Polymorphic ventricular arrhythmias are related to electrical instability due to an initial prolongation and then slow adaptation of repolarization caused by change in the heart rate and activation sequence. These stabilize during the first week after the procedure. Routine pacing at 80 bpm during this phase is recommended as well as in hospital monitoring for at least 48 hours. Patients with high risk features for arrhythmias such as congestive heart failure or impaired left ventricular function, may require pacing at higher rates. Adjustment of the pacing rate, although rarely below 70 bpm is usually undertaken after a week in most patients, preferably after an electrocardiographic evaluation for repolarization abnormalities at the lower rate [<xref ref-type="bibr" rid="R47">47</xref>].</p><p>AV node modification has been attempted with the intent to obviate the need for permanent pacemaker implantation. AV node modification can be done by the anterior or posterior approach. To compare the safety and efficacy of the anterior versus the posterior approach for AV node modification, Lee et al. randomly assigned 40 patients with medically refractory paroxysmal AF to receive AV junction modification with an anterior or posterior approach [<xref ref-type="bibr" rid="R48">48</xref>]. They concluded that (1) the two techniques had similar efficacies, (2) if one approach was ineffective, switching to the other approach might be safe, (3) combining these two approaches resulted in overall improvement in the success rate of this procedure, and (4) the posterior approach needed more radiofrequency pulses, longer procedural and fluoroscopy exposure time. Rokas et al. evaluated the role of RR interval distribution pattern as an outcome predictor of radiofrequency modification of the AV node in chronic AF and the likely mechanism of rate control [<xref ref-type="bibr" rid="R49">49</xref>]. The RR interval distribution pattern was measured from 24-hour ECG recordings obtained before and after the procedure. The preablation pattern was described as bimodal (B) or unimodal (U). Bimodality was defined as existence of 2 RR populations separated distinctly by a visually estimated intersection point, the value of which must be the same in >2 consecutive heart rate measurements. This criterion was described earlier by the same investigators in an earlier study using RR interval distribution analysis as a noninvasive method for detecting dual AV nodal physiology. They concluded that AV node modification is expected to be more effective, safe and expeditious in patients with chronic AF and a ‘B’ type RR interval distribution pattern whereas ‘U’ pattern patients may benefit from partial injury to the AV node and that posterior atrionodal input ablation may be the mechanism for rate control in these patients. Lee et al. compared the long-term effects of complete AV junction ablation with those of AV junction modification in patients with medically refractory AF [<xref ref-type="bibr" rid="R50">50</xref>]. They concluded that AV junction ablation with permanent pacing had a significantly greater ability to decrease the frequency of attacks and the extent of symptoms of AF, and more satisfaction with their general well-being among these patients.</p><p>Transcatheter radiofrequency AV node ablation followed by ventricular pacing has been shown to improve symptoms and quality of life of the patients with AF. Twidale et al. reported a modest and progressive improvement in cardiac performance following improved control of rapid heart rate after successful AV node ablation [<xref ref-type="bibr" rid="R51">51</xref>]. Edner et al. reported that in patients with left ventricular dysfunction long term improvement of systolic and diastolic left ventricular function was seen after ablation of the atrioventricular junction for rate control of atrial fibrillation [<xref ref-type="bibr" rid="R52">52</xref>]. This procedure had no adverse effects on normal left ventricular function. Natale et al. showed that a chronic irregular heart rate alone could produce an overall reduction in cardiac function that can be reversed by AV node ablation and pacemaker implant [<xref ref-type="bibr" rid="R53">53</xref>]. Therefore, ablation of the AV junction and permanent pacing could represent a more appropriate therapeutic modality over procedures or treatments targeting rate control, particularly in patients with left ventricular dysfunction. Brown et al. hypothesized that the observed improvement in ejection fraction with control of the persistently rapid ventricular response to AF would account for the improved functional class in these patients [<xref ref-type="bibr" rid="R54">54</xref>]. They also hypothesized that heart rate control would lead to an improvement in papillary muscle function, with a relation between a reduction in mitral regurgitation and improved functional class. They reported 15 patients with atrial fibrillation and severe heart failure who had marked functional improvement after AV nodal ablation with VVIR pacemaker implantation. According to their study, although the ejection fraction improved in 53% and mitral regurgitation decreased in 57 %, the improvement could not be explained entirely by either a change in ejection fraction or reduction of mitral regurgitation. However, in a recent study to assess the long-term effect of AV node ablation and ventricular pacing on left ventricular ejection fraction (LVEF) in patients with permanent AF, Szili-Torok et al. concluded that restoration of a regular ventricular rhythm following AV node ablation for patients in permanent AF does not result in improvement in left ventricular function [<xref ref-type="bibr" rid="R55">55</xref>]. It is worthwhile considering these palliative approaches in patients with chronic AF with rapid ventricular rates where drug therapy is not tolerated or unsuccessful. Possible improvement in left ventricular function following AV node ablation also makes this a consideration to subjecting the patients to toxic and potentially proarrhythmic pharmacological therapy.</p></sec></sec><sec sec-type="" id="s5"><title>Summary</title><p>From the above review, it is evident that strides in evolution of intrventional therapies for AF have been extensively guided by improved understanding of pathophysiology, electrical basis, mapping and catheter technology. The question of the electrophysiological mechanism of AF continues to haunt us: re-entry or automaticity? Probably both. The mechanism may have to be deciphered on an individual case basis. Valuable clues can be obtained by looking at multiple temporally separated ECG strips (conventional 12 lead ECG, Holter monitor or event monitor) from patients with AF which can reveal frequent unifocal or multifocal atrial ectopy initiating AF, or AF triggered by sustained supraventricular tachycardia or atrial flutter amidst AF. Other clues may surface during electrophysiology study by analysis of intracardiac ECGs, especially the mechanism of onset of AF. Although some of the sites described as focal AF triggers are only anecdotal reports, they are important to consider in an individual patient with AF. These reports also make it possible that other yet unidentified areas are AF triggers. It is also important to determine whether these foci are producing AF by themselves or merely helping sustain it. Further advances in mapping, localization and ablation techniques and technology will determine the role of these foci and pave the way for curative radiofrequency ablation of AF.</p></sec> |
Osborn Waves: History and Significance | Could not extract abstract | <contrib contrib-type="author"><name><surname>Maruyama</surname><given-names>Mitsunori</given-names></name><degrees>MD</degrees></contrib><contrib contrib-type="author"><name><surname>Kobayashi</surname><given-names>Yoshinori</given-names></name><degrees>MD</degrees></contrib><contrib contrib-type="author"><name><surname>Kodani</surname><given-names>Eitaroh</given-names></name><degrees>MD</degrees></contrib><contrib contrib-type="author"><name><surname>Hirayama</surname><given-names>Yoshiyuki</given-names></name><degrees>MD</degrees></contrib><contrib contrib-type="author"><name><surname>Atarashi</surname><given-names>Hirotsugu</given-names></name><degrees>MD</degrees></contrib><contrib contrib-type="author"><name><surname>Katoh</surname><given-names>Takao</given-names></name><degrees>MD</degrees></contrib><contrib contrib-type="author"><name><surname>Takano</surname><given-names>Teruo</given-names></name><degrees>MD</degrees></contrib> | Indian Pacing and Electrophysiology Journal | <sec sec-type="" id="s1"><title>Introduction</title><p>The Osborn wave is a deflection with a dome or hump configuration occurring at the R-ST junction (J point) on the ECG (<xref ref-type="fig" rid="F1">Figure 1</xref>). In the historical view, different names have been used for this wave in the medical literature, such as “camel-hump sign”, “late delta wave”, “hathook junction”, “hypothermic wave”, “J point wave”, “K wave”, “H wave” and “current of injury” [<xref ref-type="bibr" rid="R1">1</xref>]. Although there is no definite consensus about terminology of this wave, either “Osborn wave” or “J wave” are the most commonly used names for this wave in the current clinical and experimental cardiology. The Osborn wave can be generally observed in hypothermic patients [<xref ref-type="bibr" rid="R1">1</xref>-<xref ref-type="bibr" rid="R4">4</xref>], however, other conditions have been reported to cause Osborn waves, such as hypercalcemia [<xref ref-type="bibr" rid="R5">5</xref>], brain injury [<xref ref-type="bibr" rid="R6">6</xref>], subarachnoid hemorrhage [<xref ref-type="bibr" rid="R7">7</xref>], cardiopulmonary arrest from oversedation [<xref ref-type="bibr" rid="R8">8</xref>], vasospastic angina [<xref ref-type="bibr" rid="R9">9</xref>], or idiopathic ventricular fibrillation [<xref ref-type="bibr" rid="R10">10</xref>-<xref ref-type="bibr" rid="R12">12</xref>]. Our knowledge about the link between the Osborn waves and cardiac arrhythmias remains sparse and the arrhythmogenic potential of the Osborn waves is not fully understood. In this paper, we present a historic review of Osborn waves and discuss their clinical significance in the various clinical settings.</p></sec><sec sec-type="" id="s2"><title>Historic Overview of the Osborn Waves</title><p>In 1953, Osborn [<xref ref-type="bibr" rid="R13">13</xref>] studied the effect of hypothermia on the respiratory and cardiac function in dogs. Experimentally-induced hypothermia caused the development of a distinct deflection at the J point on the ECG, which he called “current of injury”. Earlier than the Osborn’s description, similar deflections on the ECG had already been described in 1920 and 1922 by Kraus [<xref ref-type="bibr" rid="R14">14</xref>,<xref ref-type="bibr" rid="R15">15</xref>] in hypercalcemic conditions, and in 1938 by Tomashewski [<xref ref-type="bibr" rid="R16">16</xref>] in a hypothermic patient. Although there had been several reports regarding an alternation in the ECG at the J point prior to Osborn’s article, this deflection came to be called the “Osborn wave” in honor of his systematic and excellent work. Osborn considered acidemia induced by hypothermia as a primary cause of the Osborn wave, because it disappeared if the arterial pH was normalized by hyperventilation during the same degree of cooling [<xref ref-type="bibr" rid="R13">13</xref>]. In 1959, Emslie-Smith et al [<xref ref-type="bibr" rid="R17">17</xref>]. found differences in the endocardial and epicardial responses of the ventricular myocardium to cold, and the Osborn wave was more prominent in the epicardial than endocardial leads. Also, they questioned the participation of acidosis in the genesis of the Osborn wave on the basis of their observation that the Osborn waves appeared in hypothermic dogs irrespective of the blood pH. In the same year, West et al [<xref ref-type="bibr" rid="R18">18</xref>]. confirmed that the spike and dome pattern could be recorded by a microelectrode technique in the canine epicardial action potential. The resulting notch in the action potential was rate sensitive and markedly accentuated under hypothermic conditions. Earlier studies attributed the Osborn waves to a variety of factors, including anoxia, injury current, acidosis, delayed ventricular depolarization and early ventricular repolarization [<xref ref-type="bibr" rid="R1">1</xref>,<xref ref-type="bibr" rid="R13">13</xref>,<xref ref-type="bibr" rid="R17">17</xref>,<xref ref-type="bibr" rid="R19">19</xref>,<xref ref-type="bibr" rid="R20">20</xref>].</p><p>In 1988, Litovsky and Antzelevitch [<xref ref-type="bibr" rid="R21">21</xref>] proposed a difference in the electrophysiology of the ventricular epicardium and endocardium as the basis for the Osborn waves. The 4-aminopyridine sensitive transient outward current (I<sub>to</sub>) was shown to be prominent in canine ventricular epicardium, but not in the endocardium. The more conspicuous notched configuration of the epicardial action potential was supposed to produce a transmural voltage gradient during ventricular activation that manifested as the Osborn wave in the ECG. In 1996, Yan and Antzelevitch [<xref ref-type="bibr" rid="R22">22</xref>] elegantly clarified their hypothesis using an arterially perfused canine ventricular wedge model, which made it possible to simultaneously record transmembrane action potentials from several sites across the ventricular wall together with a transmural ECG. A highly significant correlation was shown between the amplitude of the epicardial notch and the amplitude of the Osborn wave recorded during several interventions, including hypothermia, premature stimulation, and block of I<sub>to</sub> by 4-amionopyridine. In other studies, they demonstrated that a hypercalcemic [<xref ref-type="bibr" rid="R23">23</xref>] or ischemic condition [<xref ref-type="bibr" rid="R24">24</xref>], that had been reported to trigger the appearance of the Osborn waves [<xref ref-type="bibr" rid="R5">5</xref>,<xref ref-type="bibr" rid="R9">9</xref>], accentuated the epicardial action potential notch. They had also reported that there was a difference in the electrophysiological response of the epicardium and endocardium to acetylcholine and isoproterenol [<xref ref-type="bibr" rid="R25">25</xref>], which might explain the occurrence of the Osborn waves in patients with neurological disorders [<xref ref-type="bibr" rid="R6">6</xref>,<xref ref-type="bibr" rid="R7">7</xref>].</p><p>Although they noted that the conduction time across the ventricular wall and the sequence of the ventricular activation were important determining factors in the manifestation of the Osborn waves [<xref ref-type="bibr" rid="R22">22</xref>], the results from their recent studies indicated that the primary cause of the Osborn waves seemed to result from the transmural voltage gradient associated with the heterogeneous expression of I<sub>to</sub> in the ventricle.</p></sec><sec sec-type="" id="s3"><title>Clinical Significance of the Osborn Waves: Arrhythmogenic Considerations</title><p>In Osborn’s report, the presence of a deflection at the J point which he called “current of injury” heralded ventricular fibrillation and was a very bad prognostic sign in hypothermic dogs [<xref ref-type="bibr" rid="R13">13</xref>]. Fleming and Muir [<xref ref-type="bibr" rid="R26">26</xref>] confirmed the association of the Osborn waves with ventricular fibrillation in hypothermic patients. On the other hand, some reports clarified there was no correlation between the Osborn waves and ventricular fibrillation in hypothermic conditions, thus the value of the Osborn wave in hypothermic patients as a warning sign of life-threatening arrhythmias is controversial [<xref ref-type="bibr" rid="R1">1</xref>,<xref ref-type="bibr" rid="R17">17</xref>]. As a matter of fact, the Osborn waves observed in patients with hypercalcemia and neurological disorders are not usually accompanied by rhythm disturbances [<xref ref-type="bibr" rid="R5">5</xref>-<xref ref-type="bibr" rid="R8">8</xref>]. However, the Osborn waves observed in other situations have been shown to be linked to ventricular fibrillation. Aizawa et al [<xref ref-type="bibr" rid="R10">10</xref>]. reported a small series of patients with ventricular fibrillation of unknown origin whose ECGs showed unusual notches at the J point that was accentuated by longer preceding cycles. Although they attributed the notch to bradycardia-dependent intraventricular block, the characteristics of the notch in regard to the morphology and its rate-dependence, were consistent with the Osborn waves, which were referred later to as Osborn waves [<xref ref-type="bibr" rid="R27">27</xref>]. The occurrence of ventricular fibrillation seemed to be related to the augmentation of the Osborn waves in their report; a similar phenomenon was recently documented in a patient with a non-Q wave myocardial infarction due to severe coronary vasospasms [<xref ref-type="bibr" rid="R9">9</xref>]. The accentuation of the Osborn waves occurred immediately before the episodes of ventricular fibrillation (<xref ref-type="fig" rid="F2">Figure 2</xref>).</p><p>Brugada syndrome, which has been shown to be due in part to a genetic disorder in the sodium channels, is characterized by right precordial non-ischemic R-ST segment elevation on the ECG and sudden cardiac death due to life-threatening arrhythmias such as ventricular fibrillation. Formerly, the RS-T segment elevation in the right precordial leads on the ECG had been considered a normal variant [<xref ref-type="bibr" rid="R20">20</xref>], and this unique ECG change has now been recognized to be associated with sudden cardiac death [<xref ref-type="bibr" rid="R28">28</xref>,<xref ref-type="bibr" rid="R29">29</xref>]. The RS-T segment elevation in Brugada syndrome could also be regarded as a prominent J wave in the right precordial leads, and the cellular mechanism of which has been explained in the same manner as the Osborn waves. The vector of the Osborn wave tends to be toward the left and posterior, as a result, the Osborn waves can usually be seen best in the inferior and lateral precordial leads (<xref ref-type="fig" rid="F1">Figure 1</xref>,<xref ref-type="fig" rid="F2">2</xref>), while in the right precordial leads in Brugada syndrome. Even though the substrates generating these deflections seem to be somewhat different, the Osborn waves might have some arrhythmogenic potentials similar to Brugada syndrome.</p><p>Several mechanisms for the occurrence of ventricular arrhythmias in association with the Osborn waves have been proposed. The Osborn waves provide an index of the presence of a prominent notch in the ventricular epicardium, with a more negative potential at the end of phase 1 of the action potential. As the termination of phase 1 shifts to negative, the availability of ICa is diminished, and outward currents may overwhelm the active inward currents, resulting in a loss of the action potential dome. Heterogeneous loss of the epicardial action potential dome induces a marked increase in the dispersion of repolarization and phase 2 reentry, which can be responsible for sustained ventricular arrhythmias. Accentuation of the epicardial action potential notch, which can lead to phase 2 reentry, has been demonstrated in canine epicardium exposed to hypothermia [<xref ref-type="bibr" rid="R23">23</xref>], increased [Ca<sup>2+</sup>]<sub>o</sub>, simulated ischemia [<xref ref-type="bibr" rid="R24">24</xref>], and sodium channel blockers known to augment the J wave in Brugada syndrome [<xref ref-type="bibr" rid="R30">30</xref>,;<xref ref-type="bibr" rid="R31">31</xref>]. Triggered automaticity is the other proposed mechanism for ventricular arrhythmias in patients with Osborn waves. Intracellular Ca<sup>2+</sup> overload develops in several conditions which can cause Osborn waves such as hypothermia, myocardial ischemia and hypercalcemia. Early or delayed afterdepolarizations are likely to occur and form the basis for triggered activity due to the transient inward oscillatory current in Ca<sup>2+</sup> overloaded cells [<xref ref-type="bibr" rid="R1">1</xref>,<xref ref-type="bibr" rid="R9">9</xref>,<xref ref-type="bibr" rid="R23">23</xref>]. An autonomic imbalance which could attend myocardial ischemia as well as neurological disorders may be another precipitating factor of ventricular arrhythmias.</p><p>Although the arrhythmogenic implications of the Osborn waves are not fully understood, the existence of this characteristic deflection may represent some underlying critical conditions. The risks of the Osborn waves for ventricular arrhythmias may vary with the different background of each patient and should be considered individually. Further studies are needed to determine the true significance of the Osborn waves under various conditions in which they can be observed.</p></sec> |
Role of pharmacotherapy in Brugada syndrome | Could not extract abstract | <contrib contrib-type="author"><name><surname>Tsuchiya</surname><given-names>Takeshi</given-names></name><degrees>MD</degrees></contrib> | Indian Pacing and Electrophysiology Journal | <p>Brugada syndrome is a unique form of idiopathic ventricular fibrillation (VF), and is characterized by an ECG pattern with right bundle branch block (J wave), ST-segment elevation and terminal T-wave inversion in leads V1 through V3 [<xref ref-type="bibr" rid="R1">1</xref>-<xref ref-type="bibr" rid="R10">10</xref>]. It accounts for approximately 20% of all cases of sudden cardiac death in patients with structurally normal hearts and appears to be more frequent in Southeast Asia than in other regions [<xref ref-type="bibr" rid="R5">5</xref>]. To date, the only established therapy for preventing sudden cardiac death due to VF in this disease is an implantation of an implantable cardioverter defibrillator (ICD) [<xref ref-type="bibr" rid="R1">1</xref>-<xref ref-type="bibr" rid="R4">4</xref>]. Although there is no room for discussion regarding the excellent and uniform efficacy of ICDs for terminating VF, the effect is confined to the termination of VF, and ICDs cannot contribute to the prevention of VF. Therefore, there are some concerns regarding ICD therapy. The first is electrical storm associated with VF or polymorphic ventricular tachycardia, which is defined as frequent appropriate ICD shock deliveries for ventricular tachyarrhythmias of ≥3 times over 24 hours. Chalvidan et al [<xref ref-type="bibr" rid="R10">10</xref>], reported a patient who suffered from incessant VF episodes (electrical storm) and became near-fatal, but subsequently fully recovered. The second is ICD related complications including lead dislodgment, inappropriate shock delivery, infection requiring ICD removal and so on) [<xref ref-type="bibr" rid="R11">11</xref>-<xref ref-type="bibr" rid="R13">13</xref>]. The third is the indication of ICDs for infant cases of Brugada syndrome, in which the size of the ICD is too large for implantation [<xref ref-type="bibr" rid="R11">11</xref>-<xref ref-type="bibr" rid="R13">13</xref>]. A new therapy other than ICD implantation to overcome these concerns seems to be urgent.</p><sec sec-type="" id="s1"><title>Cellular mechanism</title><p>Brugada syndrome is suggested to be a primary electrical disease, and two dozen mutations of the cardiac sodium channel gene “SCN5A”, which lead to a loss of the channel function, have been found in about 15-20 % of patients with Brugada syndrome, and an autosomal mode of transmission has been demonstrated [<xref ref-type="bibr" rid="R4">4</xref>-<xref ref-type="bibr" rid="R8">8</xref>].</p><p>The ionic and cellular basis for the ST-segment elevation in Brugada syndrome is hypothesized to be due to an outward shift in the balance of the membrane ionic currents at the end of phase 1 and phase 2 of the action potential, in which an outward current is mainly due to the activation of the transient outward current (I<sub>to</sub>), and the inward current is mainly due to the activation of an inward calcium current (I<sub>Ca</sub>) and an inactivating component of an inward sodium current (I<sub>Na</sub>). The net outward shift of this current balance leads to a loss of the dome or phase 2 of the action potential. Such changes may affect the right ventricular epicardium more markedly than the endocardium and produce a marked voltage gradient in the membrane potential between the end- and epi-cardial sides of the right ventricular muscle fibers during the phase 2 (a change corresponding to the characteristic ST-segment elevation in V1 through V3 in the ECGs of Brugada Syndrome [<xref ref-type="bibr" rid="R4">4</xref>,<xref ref-type="bibr" rid="R6">6</xref>-<xref ref-type="bibr" rid="R8">8</xref>].</p><p>The loss of the action potential dome in the epicardium but not in the endocardium results in the development of a marked transmural dispersion of repolarization and refractoriness, responsible for the development of a vulnerable window during which a premature impulse or extrasystole can induce a reentrant arrhythmia. Because the loss of the action potential dome in the epicardium is usually heterogeneous, it leads to the development of epicardial dispersion of repolarization and refractoriness. Conduction of the action potential dome from sites at which it is maintained to sites at which it is lost causes local re-excitation via a phase 2 reentry mechanism, leading to the development of a very closely coupled extrasystole, which captures the vulnerable window across the wall, thus triggering a circus movement reentry in the form of VT or VF [<xref ref-type="bibr" rid="R4">4</xref>,<xref ref-type="bibr" rid="R6">6</xref>-<xref ref-type="bibr" rid="R8">8</xref>].</p><p>In patients with Brugada syndrome, an abnormal spike-and-dome configuration of the epicardial monophasic action potentials (MAPs) obtained from the right ventricular outflow tract, may create a prominent J point elevation [<xref ref-type="bibr" rid="R14">14</xref>]. Furthermore, a MAP and repolarization of longer duration in the epicardium than in the endocardium, resulting from a delayed dome formation, suggests a reversal current flow during phase 3 of the action potential. The longer duration of the action potential in the epicardium may have created the typical terminal T-wave inversion observed in Brugada syndrome [<xref ref-type="bibr" rid="R14">14</xref>]. Antelevitch et al., [<xref ref-type="bibr" rid="R4">4</xref>,<xref ref-type="bibr" rid="R7">7</xref>,<xref ref-type="bibr" rid="R8">8</xref>] suggested that the arrhythmogenic substrate arises when a further shift in the balance of the current leads to a loss of the action potential dome at some epicardial sites but not others. Actually, Nagase et al [<xref ref-type="bibr" rid="R15">15</xref>]., reported that a delayed potential was observed in the epicardial electrogram of the right ventricular outflow tract but not others, which was considered to be the second upstroke of the epicardial action potential (epicardial echo beat) generated by local and concealed phase 2 reentry [<xref ref-type="bibr" rid="R16">16</xref>].</p></sec><sec sec-type="" id="s2"><title>Pharmacological therapy based on the cellular mechanism</title><p>A variety of pathophysiologic conditions and pharmacologic interventions, which either increase the membrane outward currents or reduce the inward currents should produce a loss of the action potential dome in the right ventricular epicardium and facilitate the occurrence of VF. Phase 2 of the action potential could be abolished by reducing the I<sub>Na</sub> as predicted by the ST-segment elevation after the application of Class IC agents and restored by increasing the inward current, e.g., by enhancing the ICa or by decreasing the outward I<sub>to</sub> current, e.g., by the application of quinidine as reported [<xref ref-type="bibr" rid="R4">4</xref>,<xref ref-type="bibr" rid="R7">7</xref>,<xref ref-type="bibr" rid="R8">8</xref>].</p><p> There are many pharmacologic therapies that have been clinically examined. Class IC agents are clearly contraindicated because theses drugs reduce the INa as mentioned above [<xref ref-type="bibr" rid="R17">17</xref>]. Class IA agents, including procainamide and disopyramide, are contraindicated since these drugs strongly depress the INa but only mildly affect Ito, and are well known to exacerbate the ST elevation and occurrence of VF [<xref ref-type="bibr" rid="R4">4</xref>,<xref ref-type="bibr" rid="R7">7</xref>-<xref ref-type="bibr" rid="R9">9</xref>]. Class IB agents, which lack an Ito blocking effect and display little use-dependence owing to a rapid dissociation from the sodium channel, are not at all effective in the normalization of the ST-segment elevation [<xref ref-type="bibr" rid="R13">13</xref>]. Both amiodarone and beta-blockers are also shown to be ineffective [<xref ref-type="bibr" rid="R4">4</xref>,<xref ref-type="bibr" rid="R7">7</xref>-<xref ref-type="bibr" rid="R9">9</xref>].</p><p>Because the presence of a prominent Ito is at the heart of the mechanism underlying Brugada syndrome, any agent that blocks this current in the heart is likely to be protective [<xref ref-type="bibr" rid="R4">4</xref>,<xref ref-type="bibr" rid="R7">7</xref>,<xref ref-type="bibr" rid="R8">8</xref>]. At the time of writing this review, there was no drug developed that could cardio-selectively block the Ito. An isolated case report showed that DDD-pacing at a relatively high rate is effective in decreasing the I<sub>to</sub> (and hence suppressing VF), possibly because the repriming time of the I<sub>to</sub> channels is reported to be fairly long [<xref ref-type="bibr" rid="R18">18</xref>]. There is, however, no systematic study to examine the effect of relatively high rate pacing on the prevention of VF. Quinidine, another class IA agent, has been shown to directly inhibit Ito in addition to a secondary inhibition due to an increase in the heart rate by a vagolytic action [<xref ref-type="bibr" rid="R4">4</xref>,<xref ref-type="bibr" rid="R7">7</xref>,<xref ref-type="bibr" rid="R8">8</xref>]. Experimental studies have shown quinidine to be effective in restoring the epicardial action potential dome, thus normalizing the ST-segment and preventing phase 2 reentry and polymorphic VT in experimental models of Brugada syndrome [<xref ref-type="bibr" rid="R7">7</xref>,<xref ref-type="bibr" rid="R8">8</xref>].</p><p>Belhassen et al. [<xref ref-type="bibr" rid="R11">11</xref>], reported very important and provocative observations regarding the effect of quinidine on the prevention of VF in 34 consecutive patents with idiopathic VF, 5 of whom met the diagnostic criterion of Brugada syndrome. In the study, electrophysiologic study (EP)-guided therapy with class IA agents (mainly quinidine) was performed. These agents effectively prevented the induction of sustained polymorphic ventricular tachycardia (SPVT) and/or VF in 26 of 27 patients (96%), in all of whom SPVT or VF was induced in the baseline EP studies. Of the 23 patients treated with these medications, no patients died or had a sustained ventricular arrhythmia during a mean follow-up period of 9.1±5.6 years. They concluded that EP-guided therapy with class IA agents (mainly quinidine) was a reasonable, safe and effective approach for the long-term management of patients with idiopathic VF including Brugada syndrome. Although there have been no randomized placebo-controlled prospective studies of EP-guided quinidine therapy, I think that quinidine is a promising antiarrhythmic agent. The experimental antiarrhythmic agent, tedisamil, with its potent Ito, among other outward current, blocking action, has been suggested as a therapeutic candidate [<xref ref-type="bibr" rid="R8">8</xref>]. This drug may be more potent than quinidine because it lacks the relatively strong inward current blocking actions of quinidine.</p><p>There are, however, some limitations in using quinidine for the prevention of VF in patients with Brugada syndrome. There are some risks with antiarrhythmic agents with regard to the potential for a loss of efficacy or proarrhythmic effects under a number of conditions including hypokalemia, hypomagnesemia, bradycardia, therapy with other agents that alter repolarization, metabolic inhibitors, and changes in myocardial substrates. Furthermore, quinidine is known to have other non-cardiac side effects including abdominal cramping, diarrhea, cinchonism, which consists of decreased hearing, tinnitus, and blurred vision, thrombocytopenia, lupus syndrome, and anticholinergic side effects including a dry mouth, urinary retention and so on. These side effects may affect the compliance with the drug [<xref ref-type="bibr" rid="R11">11</xref>]. The anticholinergic action on the heart might contribute to the anti-VF effect of the drug by increasing the heart rate [<xref ref-type="bibr" rid="R13">13</xref>].</p><p>Another approach for the prevention of VF in patients with Brugada syndrome is to increase the I<sub>Ca</sub> by promoting an intracellular level of cyclic-AMP. Intravenous isoproterenol administration is especially known to be effective in suppressing ST elevation in leads V1 through V3 in Brugada syndrome patients, and to restore the action potential dome in models of this syndrome [<xref ref-type="bibr" rid="R4">4</xref>,<xref ref-type="bibr" rid="R7">7</xref>,<xref ref-type="bibr" rid="R8">8</xref>] and patients with Brugada syndrome [<xref ref-type="bibr" rid="R9">9</xref>,<xref ref-type="bibr" rid="R19">19</xref>], because it markedly increases the ICa secondary to an elevation in the intracellular level of cyclic-AMP.</p><p>Cilostazol, a new oral phosphodiesterase type-III inhibitor is primarily used as a strong antiplatelet agent, but has been shown to increase the I<sub>Ca</sub> and modestly increase the heart rate as well as contractile function. All these effects are secondary to an increased level of the cyclic-AMP caused by an inhibition of phosphodiesterase type-III activity, in guinea-pig ventricular myocytes and papillary muscles [<xref ref-type="bibr" rid="R20">20</xref>]. From these findings, it is reasonable to expect that this drug has an antiarrhythmic efficacy for preventing the VF often observed in patients with Brugada syndrome, because it increases the I<sub>Ca</sub> due to inhibition of the phosphodiesterase activity in the ventricular myocytes and decreases the I<sub>to</sub> due to an increased heart rate which again is secondary to an increased I<sub>Ca</sub> in the sinus node. I recently presented a case of a 67-year-old man with Brugada syndrome, in whom daily episodes of ventricular fibrillation occurred early in the morning for 4 days straight and the VF was completely prevented by an oral administration of cilostazol [<xref ref-type="bibr" rid="R21">21</xref>]. This effect was confirmed by the on-and-off challenge test, in which the discontinuation of this drug resulted in recurrence of VF, and the resumption of this drug again prevented the VF. This effect may be related to the suppression of the I<sub>to</sub> secondary to an increase in the heart rate and/or increase in the I<sub>Ca</sub> due to an elevation in the intracellular cyclic-AMP concentration via inhibition of phosphodiesterase activity. I performed a long-term follow-up with cilostazol in three patients (the follow-up periods were 12, 26 and 32 months, respectively), and in all an ICD was implanted for the termination of spontaneous VF. In all the patients several VF episodes were observed over a one year period prior to the cilostazol, no VF was recorded by the ICD memory and the patients were free from symptoms under 100-200 mg/day of cilostazol. Further study is needed to examine the effectiveness of the drug.</p></sec><sec sec-type="" id="s3"><title>Practical way to use anti-VF drugs</title><p>In patients who undergo aborted sudden cardiac death or syncope of unknown origin (symptomatic Brugada syndrome), no one argues that the implantation of an ICD is the first-line therapy regardless of the findings of the EP study. For those patients, drug therapy plays not a contradictory but a complimentary role to the ICD by reducing the number of ICD shock deliveries. Prevention of VF contributes to the improvement in the quality of life of the patients by avoiding uncomfortable ICD shock deliveries.</p><p>As mentioned above, Belhassen et al. [<xref ref-type="bibr" rid="R11">11</xref>], performed EP-guided medical therapy in 34 patients with idiopathic VF, in 5 of whom the criterion of Brugada syndrome were fulfilled, and reported excellent long-term results. Although medical therapy requires markers that can accurately predict the preventive effect of VF over a long-term period, there, however, seems to be no reliable marker. An EP study is usually used to examine the preventive effects of antiarrhythmic agents on sustained ventricular tachycardia in patients with structural heart disease, whereas the prognostic value of the EP study in predicting life-threatening events in Brugada syndrome is still controversial. Brugada et al., [<xref ref-type="bibr" rid="R3">3</xref>] suggested that among the asymptomatic patients, the inducibility of VT during the EP study might be a prognostic marker of risk. Studies by Priori et al., [<xref ref-type="bibr" rid="R22">22</xref>] and Kanda et al., [<xref ref-type="bibr" rid="R23">23</xref>] failed to find an association between the inducibility and recurrence of VT/VF in patients with Brugada syndrome regardless of whether it was symptomatic or asymptomatic. As Belhassen suggested [<xref ref-type="bibr" rid="R13">13</xref>], the difference might be due to the VF induction protocol in the EP study, in which Belhassen et al., used a stimulus current intensity of five times the diastolic threshold along with the use of repetition of double and triple extrastimulation at the shortest coupling intervals that resulted in ventricular capture. Further appropriate clinical trials are needed to clarify this issue.</p><p>In patients with asymptomatic Brugada syndrome who are family members of symptomatic Brugada syndrome patients, the same strategy as that for the symptomatic Brugada syndrome patients should be considered [<xref ref-type="bibr" rid="R2">2</xref>-<xref ref-type="bibr" rid="R4">4</xref>]. In another asymptomatic-patient group in whom an ECG that discloses the Brugada sign is performed for routine reasons such as a workup prior to surgery or sport license or screening for insurance, risk stratification to find the patients at high risk is needed because in most cases these patients have a benign prognosis [<xref ref-type="bibr" rid="R3">3</xref>,<xref ref-type="bibr" rid="R22">22</xref>,<xref ref-type="bibr" rid="R23">23</xref>]. Antzelevitch et al. [<xref ref-type="bibr" rid="R4">4</xref>], recommended that all asymptomatic patients with the Brugada sign should undergo an EP study for risk stratification, and, if inducible, an ICD should be implanted since Brugada et al. [<xref ref-type="bibr" rid="R3">3</xref>], reported that an overall 8% life-threatening event rate was found in initially asymptomatic patients. Belhassen et al., [<xref ref-type="bibr" rid="R13">13</xref>] suggested that EP study-guided quinidine therapy might become an alternative to ICD therapy for prophylaxis of arrhythmic events in these patients. Further appropriate clinical trials are needed.</p></sec> |
A Case of Atrial Fibrillation from Cyclosporine Toxicity | <p>We describe the unusual occurrence of atrial fibrillation immediately after a seizure in a young patient with cyclosporine toxicity. The new onset atrial fibrillation was triggered by high levels of cyclosporine and possibly facilitated by the electrolyte imbalances post seizure and the presence of underlying mild left atrial enlargement.</p> | <contrib contrib-type="author"><name><surname>Sanghi</surname><given-names>Pramod</given-names></name><degrees>MD</degrees></contrib><contrib contrib-type="author"><name><surname>Ahmad</surname><given-names>Masood</given-names></name><degrees>MD</degrees></contrib> | Indian Pacing and Electrophysiology Journal | <p>A twenty-four year old obese (100 kg) man with a past medical history of pure red cell aplasia diagnosed 8 months ago was admitted for elective chemotherapy after having failed outpatient therapy with high dose prednisone and azathioprine. Upon admission, the patient was placed on IV cyclosporine (12 mg/kg/day divided BID) and oral prednisone (1 mg/kg/day). On the 3rd hospital day, IV anti-thymocyte globulin (4000 mg/day) infused with IV methylprednisolone (125 mg/day) was initiated. On the 4th hospital day a serum cyclosporine level returned elevated at 995 ng/ml. At this point the cyclosporine dose was decreased to 8 mg/kg/day. He was also transfused 2 units of leukocyte reduced packed red blood cells for his chronically depressed blood counts (Hemoglobin 6.2 gm/dl). The patient had an uncomplicated hospital course until the 8th hospital day when he was found in his room having a tonic-clonic seizure that lasted approximately 1-2 minutes. Shortly afterwards, he was found to be in atrial fibrillation (<xref ref-type="fig" rid="F1">Figure 1</xref>) with a ventricular rate of 160 bpm and hypotensive with a blood pressure of 90/50 mmHg. After a brief postictal period, he denied any complaints but was aware of his tachycardia. The patient was initially given a loading dose of dilantin, and his heart rate was controlled with boluses of IV diltiazem. Laboratory values taken immediately after his seizure showed a K<sup>+</sup> of 3.0 mmol/l, Ca<sup>2+</sup> of 6.7 mg/dl, Mg<sup>2+</sup> of 1.0 mg/dl, and CO<sub>2</sub> of 9 mmol/l, but on repeat labs taken 2 hours later, the electrolyte imbalances had all normalized. The patient’s heart rhythm spontaneously converted to normal sinus (<xref ref-type="fig" rid="F2">Figure2</xref>) about 5 hours after he was initially found to be in atrial fibrillation. His cyclosporine level that had been drawn just a few hours prior to his seizure returned as 1360 ng/ml, which is over 4 times the therapeutic dose for this treatment. The cyclosporine was held and the following day returned as 264 ng/nl. Echocardiogram performed the following day revealed normal systolic function (LVEF 55 - 60%) with a mildly enlarged left atrium (4.2 cm) and mild to moderate tricuspid regurgitation. His estimated pulmonary artery pressure was 40 mmHg.</p><sec sec-type="" id="s1"><title>Discussion</title><p>Although this is not the first time atrial fibrillation has been reported in the literature from an overdose of cyclosporine [<xref ref-type="bibr" rid="R1">1</xref>], our case is unusual for the following reasons. First of all, our patient received an appropriate dose of cyclosporine but was still found to have toxic levels of the drug. While he was given the correct dosages for his weight (12 kg/mg/day), the patient was obese and some studies suggest that the dosages of cyclosporine in these patients not be based on their actual body weight [<xref ref-type="bibr" rid="R2">2</xref>]. Other potential causes of high blood cyclosporine levels in this patient are increased gastrointestinal absorption secondary to high dietary fat intake while in the hospital or an idiosyncratic drug reaction [<xref ref-type="bibr" rid="R3">3</xref>].</p><p>What is also unique about this case is that the patient suffered a seizure prior to the onset of atrial fibrillation. Seizure is a well described complication of cyclosporine toxicity [<xref ref-type="bibr" rid="R4">4</xref>]. The electrolyte imbalances assumed to be secondary to the seizure may have predisposed the patient to atrial fibrillation in the face of cyclosporine toxicity. Although atrial fibrillation is a known complication of cyclosporine toxicity independent of seizure activity or electrolyte imbalances, the atrial fibrillation seen in this case occurred at a relatively low level of cyclosporine toxicity as compared to other case reports in the literature [<xref ref-type="bibr" rid="R1">1</xref>]. In addition, the patient may have been at further risk of developing atrial fibrillation due to the mild left atrial enlargement seen on his echocardiogram.</p><p> Although the mechanism by which cyclosporine causes atrial fibrillation is unknown, cyclosporine is known to have some deleterious cardiac effects including coronary vasospasm [<xref ref-type="bibr" rid="R5">5</xref>] and ventricular hypertrophy [<xref ref-type="bibr" rid="R6">6</xref>]. Further research is needed to determine the mechanisms by which these potentially harmful cardiac effects occur.</p></sec> |
Re-use of Explanted DDD Pacemakers as VDD- Clinical Utility and Cost Effectiveness | <p>Re-use of DDD pulse generators explanted from patients died of unrelated causes is associated with an additional cost of two transvenous leads if implanted as DDD itself, and high rate of infection according to some studies. We studied the clinical and economical aspects of reutilization of explanted DDD pacemakers programmed to VDD mode. Out of 28 patients who received VDD pacemaker during the period, October 2000- September 2001 in the Department of Cardiology, PGIMER, Chandigarh, 5 poor patients were implanted with explanted DDD pulse generators programmed to VDD mode. Each implantation was planned and carried out according to a standard protocol. The age ranged from 45 to 75 (mean-61) years. The indications for pacing were complete heart block (4) and second degree AV block (1). The clinical profile, costs and complications, if any were noted and followed up at regular intervals. The results were compared with patients who received new DDD pulse generators during this period. The additional cost for the atrial lead was not required in these patients. None of these patients had any local site infection. Compared to the two-lead system, the single lead system provided more rapid implantation and minimized complications associated with placement of an atrial lead. The explanted DDD pacemaker can be safely reused as VDD mode with same efficacy in selected patient population. This is associated with lower cost and complications compared to reimplantation as DDD itself.</p> | <contrib contrib-type="author"><name><surname>Namboodiri</surname><given-names>KKN</given-names></name><role>Senior Resident</role></contrib><contrib contrib-type="author"><name><surname>Sharma</surname><given-names>YP</given-names></name><role>Assistant Professor</role></contrib><contrib contrib-type="author"><name><surname>Bali</surname><given-names>HK</given-names></name><role>Additional Professor</role></contrib><contrib contrib-type="author"><name><surname>Grover</surname><given-names>A</given-names></name><role>Additional Professor & Head</role></contrib> | Indian Pacing and Electrophysiology Journal | <sec id="s1"><title>Introduction</title><p>Technology is outpacing the financial resources available for health care all over the world. Pacemakers are among the expensive devices, and when indicated properly they have no alternative. In many instances pacemakers outlive the patients and it would be a waste to dispose these pacemakers that are still in good condition having considerable battery life left without allowing others to benefit. However the reuse of pacemakers is associated with widespread technical, medical, ethical and legal considerations.</p><p>Reuse of DDD pulse generators explanted from patients dying of unrelated causes is associated with an additional cost of two transvenous leads, if reimplanted in DDD mode itself. There are also reports of unacceptably high rates of infection associated with reuse of explanted pacemakers. We studied the clinical and economical aspects of utilization of explanted DDD pacemakers programmed to VDD mode.</p></sec><sec sec-type="materials|methods" id="s2"><title>Materials and methods</title><p>During the period October 2000 to December 2001 (15 months) the patients who received pacemakers in PGIMER, Chandigarh were studied. Out of these patients, those who received generators explanted from patients who had died of unrelated causes were noted. We used explanted dual chamber pulse generators (Medtronic, Prodigy DR, model 7860, bipolar) after programming to VDD mode, with the use of new VDD lead (58-13.5-9F) in poor patients who could not afford a new pulse generator. Appropriate preprocedural evaluation was done in these patients to achieve optimal patient selection for VDD mode. The pulse generators had been explanted postmortem from patients who died of causes other than pacemaker failure after informed consent from the nearest relative. The generators were properly cleansed in antiseptic solution and sterilized with ethylene oxide and reliably tested for function and battery life. The recipient’s consent was also sought after proper explanation of the risks and benefits.</p><p>These patients were followed up after one month of implantation initially and three monthly thereafter. They were compared with those patients who received DDD pacing during the same period regarding cost, complications and procedural time. To assess the quality of life (QOL), a standard questionnaire (<xref ref-type="table" rid="T1">Table 1</xref>) was given to the patients during their last follow up visit of the study period. The questionnaire contained 7 questions, 3 related to the general QOL and 4 specific to the pacemaker related symptoms. Each question contained 4 responses, the individual score varying from 1-4 according to severity. The total score in each patient was calculated.</p></sec><sec id="s3"><title>Results</title><p>During the study period 136 pacemakers were implanted in same number of patients in this hospital, which included 53 dual chamber (DDD 25, VDD 28) and 83 single chamber (VVI/ R 78, AAI/R 5) pacemakers. Indications for the implantation included sick sinus syndrome (31), high-grade AV block (94), symptomatic trifascicular block (9) and cardiomyopathy (2). 21 cases involved reuse of pacemakers, which included 9 cases of reimplantation in the same patient (pacemaker infection-2, lead fracture or displacement-5, and pacemaker extrusion-2) and 12 cases of reuse of pacemakers explanted postmortem (dual chamber-8, single chamber-4).</p><p>Out of 8 dual chamber pacemaker generators explanted, 5 programmable DDD pulse generators were reimplanted in VDD mode. These 5 patients included one female and their age varied from 45 to 75 (mean 61) years (<xref ref-type="table" rid="T2">Table 2</xref>). Four of them had complete heart block while one had second-degree Mobitz-type II heart block. The battery life assessed prior to reuse varied from 43 to 124 (mean 89.4) months. The average battery voltage at implantation was 2.77 volts. A new VDD lead (58-13.5-9F), which costs about 20,000/-, was used in all. This reduced the extra amount spent on the two transvenous leads (2 x 18, 000/- = Rs 36,000/-) by Rs 16, 000/- per patient. As the VDD lead required only a single lead introducer for its insertion, an extra amount of Rs 2,000/- was also avoided making a net gain of Rs 18,000/-, i.e., 50% of the total expenditure of refurbishment per patient.</p><p>The fluoroscopy time spent for the procedure was also less compared to that used for DDD insertion. None of these 5 patients had any local site infection. The duration of postprocedural hospital stay was not prolonged in these patients (5.2 vs. 7.4 days; NS). These patients on follow up (mean follow up 19.2 months) showed normal atrial sensing and ventricular pacing. The quality of life, assessed based on a standard questionnaire did not reveal any significant difference between those received explanted pulse generator and newer one (22.2 vs. 24.4 points; NS). There was no incidence of pacemaker related tachycardia, pacemaker syndrome or lead displacement in any of these (<xref ref-type="table" rid="T3">Table 3</xref>).</p></sec><sec id="s4"><title>Discussion</title><p>We studied 5 patients who received explanted DDD programmable pulse generators, which were programmed to VDD mode prior to implantation. Analysis of cost effectiveness and safety in these poor and unaffordable patients revealed that these generators can be safely and effectively used at a significantly lower cost.</p><p>The rationale for the reuse of pacemakers is based on the following facts. The current lithium battery pacemakers have an expected life greater than 10 yrs. In high-risk subgroups of patients with coronary artery disease and atrioventricular block, the 3-year mortality approaches 60% [<xref ref-type="bibr" rid="R1">1</xref>] and thus many pacemakers would have more than 5 years of life left when the recipients die. Pringle et al reported a mortality rate of 58% within 2 years of last generator implant based on a retrospective examination of 169 consecutive pacemaker patient-deaths [<xref ref-type="bibr" rid="R2">2</xref>]. It should also be noted that in patients with severe cardiac diseases such as heart failure and cardiomyopathy, more sophisticated pacemakers are often implanted, and these patients do die earlier from their original disease. Thus the waste of pacemaker life is aggravated.</p><p>Mugica et al reviewed over 3,500 patients who had a 10-year follow up and reported no significant difference in the actuarial survival of those patients given explanted generator compared to those received newer one [<xref ref-type="bibr" rid="R3">3</xref>]. Rosengarten et al observed similar incidence of pacemaker related complications and survival among new and refurbished pacemakers in a prospective comparative study over a mean follow up period of 36 months [<xref ref-type="bibr" rid="R4">4</xref>]. Such experiences have repeatedly confirmed that when properly carried out, re-use of pacemakers does not pose any additional risk with considerable reduction in cost. The lack of widespread acceptance of pacemaker re-use is, therefore, not alone due to technical or medical considerations, but is related to ethical and legal problems associated with this approach.</p><p>The ethical and legal issues involve both the retrieval of a still usable pacemaker from a deceased patient and the selection of the patient to receive such a pacemaker. In countries such as Sweden, pacemakers are considered to be the property of the state and can be removed routinely without the need for consent from the families of patients who have died. In many other countries including Canada, US and India, once the pacemaker has been implanted regardless of the source of funding for this procedure, the device is considered to belong to the patient. Retrieval of such a device therefore requires the consent of the next of kin or the living will of the patient. In view of theoretical risks of reimplantation of cadaveric explants, informed consent of the recipient also should be sought, after proper explanation of possible risks involved in the re-use.</p><p>The clinical problem of reuse of pacemaker in the recipient is mainly based on the rate of infection and battery life. The risk of re-use is that an instrument might be improperly selected due to an inaccurate history of use, or improperly cleansed, tested or sterilized. Explanted pacemakers should be considered for re-use only when the reliable clinical record indicates that the instrument has had no malfunction and has an arbitrarily set minimum battery life of 5 years. After proper electronic testing, the pulse generators are to be washed under sterile conditions in distilled water and then gas sterilized with ethylene oxide for two hours at 55ºC and 60% humidity. The device is released for implantation after essential aeration for 48 hours at 55ºC in an aeration device or in the appropriate sterilizer. Mond et al described a high rate of infection with the reuse of explanted pacemakers [<xref ref-type="bibr" rid="R5">5</xref>]. Based on a retrospective case control study, which involved 100 patients who received re-used pacemakers, Linde et al concluded that the re-use of pacemakers could be carried out without increased risk to the patient provided a proper routine for technical control and sterilization is followed [<xref ref-type="bibr" rid="R6">6</xref>]. Experiences of other authors have also confirmed the short- and long-term safety of re-use of cadaveric explants [<xref ref-type="bibr" rid="R7">7</xref>-<xref ref-type="bibr" rid="R9">9</xref>].</p><p>None of our patients had local site infection. Battery life of the pulse generators was properly assessed prior to implantation by a pacing system analyzer in all. The average battery life of these 5 pulse generators at the time of implantation was 89.4 months (at 60 bpm, output 4 V, pulse width 0.4 ms and VDD mode). The re-use as VDD system resulted in avoiding the need of atrial lead placement leading to reduction in fluoroscopy time and virtual nonexistence of complications related to atrial lead, apart from reducing the cost significantly. The cost reduction by Rs 16,000/- per patient was significant considering the baseline economic status of these poor patients. There was no significant difference in the duration of hospital stay, need of antibiotics or additional cost to the patients. Proper atrial sensing, the most essential component of VDD pacing was also unaltered on follow up. Similarly, none of these patients had pacemaker syndrome, pacemaker-mediated tachycardia or atrial arrhythmias on follow up. There was no significant difference in the quality of life score between those received explanted pulse generator and newer one.</p><p>This study has some limitation in that the patient sample is small due to obvious problems mentioned. We didn’t assess the incidence of VVI pacing by Holter monitoring on follow up. However, with good sinus node function, programmed lower rate of 50 bpm and good atrial tracking noted at each follow up visit, significant VVI pacing in between is unlikely, though we cannot categorically deny such an occurrence. However, this issue is largely overwhelmed by the fact that most recipients were extremely poor and they were not in a position to afford anything more than the cost of the new electrode.</p><p>The re-use of pacemakers is not carried out routinely, as many physicians and affordable patients do not encourage it usually. The manufacturers are also understandably reluctant to participate in refurbishing used pacemakers and the pacemaker warranty disclaims liability to anyone except the first recipient. Yokoyama reported that re-use of explanted pacemaker constituted only 0.1% of all implanted pacemaker units in Japan in 1992 [<xref ref-type="bibr" rid="R10">10</xref>]. The corresponding figures from India are not available.</p><p>We could not find many publications on the re-use of explanted pacemakers from our country. Panja et al noted that the efficacy of re-used pacemaker was highly comparable to that of newly implanted one in terms of morbidity and mortality. Re-use in the same patient, but not reuse of cadaveric explants was associated with high rate of infection in their study [<xref ref-type="bibr" rid="R7">7</xref>]. Sethi et al re-used 13 explanted generators out of 42 refurbished pacemakers. There was no pulse generator failure or adverse reactions among these during a mean duration follow up of 34 months [<xref ref-type="bibr" rid="R8">8</xref>]. Balachander et al also reported the safety and efficacy of refurbished pacemakers in 140 implantations [<xref ref-type="bibr" rid="R9">9</xref>]. Hence, the pacemaker reuse needs to be encouraged in developing countries and rationale for reuse should be purely economical.</p><p>Concluding, the explanted programmable dual chamber pacemakers can be safely re-used in VDD mode with same efficacy in selected patient population, at significantly lower cost.</p></sec> |
Refurbishing Pacemakers: A Viable Approach | Could not extract abstract | <contrib contrib-type="author"><name><surname>Anilkumar</surname><given-names>R</given-names></name></contrib><contrib contrib-type="author"><name><surname>Balachander</surname><given-names>J</given-names></name></contrib> | Indian Pacing and Electrophysiology Journal | <p>Cardiologists implant permanent pacemakers widely for indications like sick sinus syndrome and complete heart block. The guidelines for such implantations are well established [<xref ref-type="bibr" rid="R1">1</xref>]. However, in developing countries like India, all patients who need pacemakers do not receive them because of financial constraints. Even when such patients get a pacemaker, it is often a more affordable VVI pacemaker rather than the costly DDD pacemaker. The lack of a health insurance scheme and improper social support programs prevent the more widespread implantation of appropriate pacemakers.</p><p>However, in the developed countries and in affluent pockets of developing countries like India, the pacemaker implantation rates are quite high. Often permanent pacemakers are implanted in the very old and people with predicted brief longevities, due to medico-legal and other social reasons. There are quite a few instances when pacemakers are explanted within a year or even within a few months. This is often due to the unfortunate death of the patient due to unrelated causes. Such pacemakers have battery lives, which are near normal. These can be explanted from the dead patient after taking consent from the relatives and “refurbished” for use in another needy patient. Refurbishing involves proper re-sterilization, checking of battery life, pacing mode and other parameters and re-labelling with the current parameters including predicted battery life. These refurbished pacemakers are a suitable alternative for the financially ‡no option’ group of patients who otherwise would not afford a pacemaker. These can last nearly as long as the original pacemakers. Even pulse generators whose shelf lives have expired can also be resterilised and used gainfully for the economically deprived.</p><p>Nampoothiri et al, in this issue of the journal, have dwelt upon this option [<xref ref-type="bibr" rid="R2">2</xref>]. They have explanted DDD pacemakers from patients who have expired and implanted as VDD pacemaker in a new patient with a VDD lead. The authors have stressed on the use of VDD Medtronic lead being more economical than two Medtronic leads. But two leads, one atrial and one ventricular, available from an Indian company, can be even cheaper than the VDD lead. The article however clearly shows the effective use of refurbished pacemakers.</p><p>Refurbishment of pacemakers requires the active co-operation of medical and cardiology units and the emergency services department of various hospitals. The pacemakers need to be explanted from patients who expired, after consent from the relatives. Then, these pacemakers need to be checked for their integrity and battery life, sterilized re-labeled and kept ready for use in the needy patient.</p><p>STIMUBANK is one such organization in Nancy, France, which collects pacemakers that are prematurely removed, or shelf expired pacemakers from various parts of France. Pulse generators having a potential life of more than four years, i.e. the guarantee period of pacemakers in France, are shipped to needy centers outside France. Legislation in France does not permit the re use of implanted devices.</p><p>JIPMER, Pondicherry, has been the largest beneficiary of STIMUBANK and has a large experience of implanting refurbished pacemakers [<xref ref-type="bibr" rid="R3">3</xref>]. This center has followed up over five hundred patients with refurbished pacemakers over the last twenty years. The longevity, performance and complications compare reasonably well with the newly implanted pulse generators. There are smaller series of refurbished pacemakers from other centers as well [<xref ref-type="bibr" rid="R4">4</xref>,<xref ref-type="bibr" rid="R5">5</xref>]</p><p>Specific consent has to be sought from the patient and his relatives regarding the implantation of a refurbished pacemaker. The parameters of the pacemaker also have to be checked thoroughly before implant. This requires that programmers of all makes of refurbished pacemakers be available at the center implanting such pacemakers. Inadvertently, implantation of a pacemaker programmed for bipolar pacing with a unipolar pacing lead can result in failure of pacing. Hence it is very essential that proper assessment of pacing parameters be done very judiciously and adequate battery life ensured before embarking on implantation of a refurbished pacemaker.</p><p>A good number of pacemakers with reasonable battery lives can be salvaged with some planning and foresight. This could even be extended to some defibrillators and triple chamber pacemakers whose batteries outlive the patients in whom they are implanted. A central registry or bank, akin to the STIMUBANK, can be formed in India where explanted pacemakers from anywhere in the country can be sent. This organization can refurbish the pacemakers at a very minimal cost and supply, on request, to the needy patients. Thus, with just the cost of one or two pacemaker leads, a single or the costlier dual chamber pacemaker, as indicated, can benefit the economically deprived needy patient. This option needs to be seriously considered.</p> |
Current concepts on ventricular fibrillation: A Vicious Circle of Cardiomyocyte Calcium Overload in the Initiation, Maintenance, and Termination of Ventricular Fibrillation | <p>Based on recent experimental studies, this review article introduces the novel concept that cardiomyocyte Ca<sup>2+</sup> and ventricular fibrillation (VF) are mutually related, forming a self-maintaining vicious circle in the initiation, maintenance, and termination of VF. On the one hand, elevated myocyte Ca<sup>2+</sup> can cause delayed afterdepolarizations, triggered activity, and consequently life-threatening ventricular tachyarrhythmias in various pathological conditions such as digitalis toxicity, myocardial ischemia, or heart failure. On the other hand, VF itself directly and rapidly causes progressive myocyte Ca<sup>2+</sup> overload that maintains VF and renders termination of VF increasingly difficult. Accordingly, energy levels for successful electrical defibrillation (defibrillation thresholds) increase as both VF and Ca<sup>2+</sup> overload progress. Furthermore, VF-induced myocyte Ca<sup>2+</sup> overload can promote re-induction of VF after defibrillation and/or postfibrillatory myocardial dysfunction (postresuscitation stunning) due to reduced myofilament Ca<sup>2+</sup> responsiveness. The probability of these adverse events is best reduced by early detection and rapid termination of VF to prevent or limit Ca<sup>2+</sup> overload. Early additional therapy targeting transsarcolemmal Ca<sup>2+</sup> entry, particularly during the first 2 min of VF, may partially prevent myocyte Ca<sup>2+</sup> overload and thus, increase the likelihood of successful defibrillation as well as prevent postfibrillatory myocardial dysfunction.</p> | <contrib contrib-type="author"><name><surname>Zaugg</surname><given-names>Christian E</given-names></name><degrees>Ph D</degrees></contrib> | Indian Pacing and Electrophysiology Journal | <sec sec-type="" id="s1"><title> Myocyte Ca<sup>2+</sup> overload initiates ventricular fibrillation</title><p>Elevated myocyte Ca<sup>2+</sup> (Ca<sup>2+</sup> overload) has generally been accepted to be responsible for the initiation of potentially lethal ventricular tachyarrhythmias including ventricular fibrillation (VF) in various pathological conditions such as digitalis toxicity, myocardial ischemia, or heart failure [<xref ref-type="bibr" rid="R1">1</xref>-<xref ref-type="bibr" rid="R4">4</xref>]. Specifically, the accumulation of Ca<sup>2+</sup> in cardiomyocytes has long been suggested to cause delayed afterdepolarizations, triggered activity, and consequently life-threatening ventricular tachyarrhythmias [<xref ref-type="bibr" rid="R4">4</xref>,<xref ref-type="bibr" rid="R5">5</xref>]. Accordingly, myocyte Ca<sup>2+</sup> overload has been shown to be related to the initiation of tachyarrhythmic activity in isolated hearts or in cardiomyocytes of rats or ferrets using bioluminescence or fluorescence of intracellular Ca<sup>2+</sup> indicators (e.g. aequorin or indo-1) [<xref ref-type="bibr" rid="R1">1</xref>,<xref ref-type="bibr" rid="R6">6</xref>]. Further evidence of the importance of myocyte Ca<sup>2+</sup> for the vulnerability to VF arises from a close correlation between myocyte Ca<sup>2+</sup> levels and VF thresholds [<xref ref-type="bibr" rid="R2">2</xref>]. Controlled intracellular Ca<sup>2+</sup> accumulation by programmed rapid ventricular stimulation (minimizing the effects of other arrhythmogenic factors) led to a parallel increase of VF thresholds in isolated rat hearts under nonischemic conditions [<xref ref-type="bibr" rid="R2">2</xref>]. </p><p>In general, when Ca<sup>2+</sup> loading of cardiomyocytes becomes sufficiently high, the sarcoplasmic reticulum can generate spontaneous Ca<sup>2+</sup> oscillations that are not triggered by sarcolemmal depolarizations [<xref ref-type="bibr" rid="R1">1</xref>,<xref ref-type="bibr" rid="R3">3</xref>,<xref ref-type="bibr" rid="R8">8</xref>,<xref ref-type="bibr" rid="R6">6</xref>]. If sufficiently synchronized, these Ca<sup>2+</sup> oscillations may cause delayed afterdepolarizations and initiate VF or modulate the initiation of VF [<xref ref-type="bibr" rid="R3">3</xref>]. Additionally, myocyte Ca<sup>2+</sup> overload may facilitate the initiation of VF by Ca<sup>2+</sup>-induced cell-to-cell uncoupling [<xref ref-type="bibr" rid="R4">4</xref>], thereby slowing conduction and amplifying the tendency for reentrant arrhythmias. This tendency is particularly amplified in the hypertrophied heart where repolarization is heterogeneous and refractoriness prolonged [<xref ref-type="bibr" rid="R5">5</xref>]. Similarly, mutations in Ca<sup>2+</sup> handling proteins have been suggested to contribute to hereditary arrhythmias. For example, defective ryanodine type 2 receptors (calcium release channels at the sarcoplasmic reticulum) or reduced levels of calsequestrin (a high-capacity Ca<sup>2+</sup> binding protein expressed inside the sarcoplasmic reticulum), may cause increased Ca<sup>2+</sup> discharge from the sarcoplasmic reticulum, and consequently ventricular tachyarrhythmias and sudden cardiac death induced by exercise, stress, or heart failure [<xref ref-type="bibr" rid="R7">7</xref>-<xref ref-type="bibr" rid="R9">9</xref>].</p></sec><sec sec-type="" id="s2"><title>VF causes myocyte Ca<sup>2+</sup> overload</title><p>Furthermore, myocyte Ca<sup>2+</sup> and VF are mutually related. Myocyte Ca<sup>2+</sup> overload can induce VF and conversely, VF itself causes myocyte Ca<sup>2+</sup> overload [<xref ref-type="bibr" rid="R2">2</xref>,<xref ref-type="bibr" rid="R10">10</xref>-<xref ref-type="bibr" rid="R13">13</xref>]. Importantly, our studies using the Ca<sup>2+</sup> sensitive fluorescent dye indo-1 in isolated rat hearts suggest that myocyte Ca<sup>2+</sup> rises biphasically during VF (<xref ref-type="fig" rid="F1">Figure 1</xref>). In the first 2 min of VF, mean myocyte Ca<sup>2+</sup> rises steeply and rapidly reaches about double of normal levels. Thereafter, myocyte Ca<sup>2+</sup> continues to rise but at a slower rate [<xref ref-type="bibr" rid="R11">11</xref>-<xref ref-type="bibr" rid="R14">14</xref>]. Additionally, successful defibrillation (electrical or pharmacological) led to a sudden reduction of VF-induced myocyte Ca<sup>2+</sup> overload (<xref ref-type="fig" rid="F1">Figure 1</xref>) [<xref ref-type="bibr" rid="R10">10</xref>]. In contrast, failed defibrillation shocks did not alter Ca<sup>2+</sup> [<xref ref-type="bibr" rid="R10">10</xref>]. This demonstrates that VF directly and (dependent on VF duration) reversibly causes myocyte Ca<sup>2+</sup> overload. Because the Ca<sup>2+</sup> channel blocker diltiazem (1 μM) largely prevented VF-induced myocyte Ca<sup>2+</sup> overload in the initial phase of VF in our experiments, most of the Ca<sup>2+</sup>contributing to myocyte Ca<sup>2+</sup> overload presumably enters the cells through L-type Ca<sup>2+</sup> channels [<xref ref-type="bibr" rid="R15">15</xref>]. This is likely due to the rapid activation rate in VF as the pacing cycle length was inversely related to both the rate and the degree of myocyte Ca<sup>2+</sup>overload induced by rapid pacing [<xref ref-type="bibr" rid="R2">2</xref>]. As VF persists, the contribution of Ca<sup>2+</sup> entry through L-type Ca<sup>2+</sup> channels to myocyte Ca<sup>2+</sup> overload appears to decrease because diltiazem perfusion after 5 min of VF could not prevent myocyte Ca<sup>2+</sup> to increase further in perfused rat hearts [<xref ref-type="bibr" rid="R10">10</xref>]. At this stage, further myocyte Ca<sup>2+</sup> overload may arise from sarcoplasmic reticular Ca<sup>2+</sup> release, from reverse Na+/Ca<sup>2+</sup> exchange and/or from other sources (whereas individual contributions may vary species-dependently).</p></sec><sec sec-type="" id="s3"><title>VF-induced myocyte Ca<sup>2+</sup> overload maintains VF</title><p>Independent of the Ca<sup>2+</sup> source, VF-induced myocyte Ca<sup>2+</sup> overload contributes to maintain VF, leading to a self-maintaining vicious circle in which termination of VF becomes increasingly difficult (<xref ref-type="fig" rid="F2">Figure 2</xref>). Consequently, myocyte Ca<sup>2+</sup> overload can cause electrical defibrillation to fail and postshock re-induction of VF [<xref ref-type="bibr" rid="R10">10</xref>]. Accordingly, we could show that energy levels for successful electrical defibrillation (defibrillation thresholds) increase as both VF and Ca<sup>2+</sup> overload progress [<xref ref-type="bibr" rid="R10">10</xref>]. Manipulating myocyte Ca<sup>2+</sup> before defibrillation (increasing extracellular Ca<sup>2+</sup> during VF in perfused rat hearts) we could demonstrate a causal relationship between myocyte Ca<sup>2+</sup> concentrations and defibrillation success. Thus, the longer VF lasts, the higher both myocyte Ca<sup>2+</sup> concentration and defibrillation threshold rise. This relationship was not due to prolonged myocardial ischemia because the hearts were continuously perfused during VF (normal levels of coronary flow, of coronary effluent pH, and of myocardial O2 consumption) [<xref ref-type="bibr" rid="R15">15</xref>]. Moreover, with increasing duration of VF, modulation of intracellular Ca<sup>2+</sup> gets more difficult. Neither the Ca<sup>2+</sup> channel blocker diltiazem (in a negative inotropic concentration of 1 μM) [<xref ref-type="bibr" rid="R16">16</xref>] nor low extracellular Ca<sup>2+</sup> (reduction from 3.0 mM to 0.6 mM) could significantly decrease myocyte Ca<sup>2+</sup> in fibrillating rat hearts. [<xref ref-type="bibr" rid="R10">10</xref>] Accordingly, diltiazem or low extracellular Ca<sup>2+</sup> could not decrease defibrillation thresholds [<xref ref-type="bibr" rid="R10">10</xref>] as previously found for verapamil, another Ca<sup>2+</sup> channel blocker, in pigs [<xref ref-type="bibr" rid="R17">17</xref>] or human beings in vivo [<xref ref-type="bibr" rid="R18">18</xref>]. The mechanism by which VF-induced myocyte Ca<sup>2+</sup> overload increases defibrillation thresholds is probably related to a Ca<sup>2+</sup>-induced increase in the likelihood of defibrillation shocks to re-induce VF. We have previously shown that Ca<sup>2+</sup> modulates the induction of VF by an electrical stimulus applied during the vulnerable period of repolarization [<xref ref-type="bibr" rid="R2">2</xref>,<xref ref-type="bibr" rid="R19">19</xref>]. As some portion of the fibrillating myocardium is always repolarizing [<xref ref-type="bibr" rid="R20">20</xref>], myocyte Ca<sup>2+</sup> overload could increase the likelihood of a shock to re-induce VF. Thus, a shock applied to Ca<sup>2+</sup> overloaded myocardium may terminate VF but simultaneously re-induce it by stimulating myocardium that is in the vulnerable period of repolarization. Furthermore, the chances for re-induction of VF increase as VF persists because normalization of myocyte Ca<sup>2+</sup> becomes increasingly difficult. Incomplete reduction of myocyte Ca<sup>2+</sup> overload after initially successful defibrillation can be followed by synchronized spontaneous Ca<sup>2+</sup> oscillations from the sarcoplasmic reticulum and subsequent reinduction
of VF [<xref ref-type="bibr" rid="R10">10</xref>]. Because VF inevitably causes myocyte Ca<sup>2+</sup> overload, this vicious circle between myocyte Ca<sup>2+</sup> and VF might be a critical mechanism of failed defibrillation and postshock re-induction of VF. Moreover, this vicious circle concept suggests that the probability of these events is best reduced by early detection and rapid termination of VF to prevent or limit Ca<sup>2+</sup> overload, and of course to prevent cerebral ischemia.</p></sec><sec sec-type="" id="s4"><title>Myocyte Ca<sup>2+</sup> overload causes myocardial stunning after defibrillation
(postresuscitation stunning)</title><p>Even if the self-maintaining vicious circle of Ca<sup>2+</sup> and VF is interrupted and defibrillation succeeds, myocyte Ca<sup>2+</sup> overload continues to cause problems. This is because transitory Ca<sup>2+</sup> overload that occurs during VF can lead to reduced myofilament Ca<sup>2+</sup> responsiveness [<xref ref-type="bibr" rid="R15">15</xref>] and consequently to postfibrillatory myocardial dysfunction [<xref ref-type="bibr" rid="R11">11</xref>] [<xref ref-type="bibr" rid="R15">15</xref>], a condition that we have termed postresuscitation stunning [<xref ref-type="bibr" rid="R15">15</xref>]. We found that the degree of Ca<sup>2+</sup> overload during VF was inversely associated with the reduction of myofilament Ca<sup>2+</sup> responsiveness after pacing-induced VF in our experiments in isolated rat hearts [<xref ref-type="bibr" rid="R11">11</xref>,<xref ref-type="bibr" rid="R15">15</xref>]. Accordingly, as Ca<sup>2+</sup> overload progressed during VF, longer episodes of VF led to a more pronounced myocardial dysfunction than short episodes of VF. Moreover, increasing or decreasing Ca<sup>2+</sup> overload during VF led to parallel changes in myofilament Ca<sup>2+</sup> responsiveness (estimated as ratio of left ventricular developed pressure over myocyte Ca<sup>2+</sup> transient amplitudes). The molecular mechanisms whereby transitory Ca<sup>2+</sup> overload undermines contractile protein function seems to be related to proteolysis that is mediated at least partly by Ca<sup>2+</sup>-activated proteases (calpains) [<xref ref-type="bibr" rid="R21">21</xref>]. The substrates of calpains with respect to cardiac myofibrillar proteins include troponin I, troponin T, and others [<xref ref-type="bibr" rid="R21">21</xref>].</p><p>Postresuscitation stunning may be clinically relevant because defibrillation frequently results in a depressed myocardial function [<xref ref-type="bibr" rid="R22">22</xref>], particularly after prolonged VF [<xref ref-type="bibr" rid="R23">23</xref>]. Despite hemodynamic support, this dysfunction may be responsible for deaths after initially successful defibrillation [<xref ref-type="bibr" rid="R23">23</xref>]. In patients, however, VF is undoubtedly accompanied by myocardial ischemia. Therefore, global postischemic stunning most likely contributes to postfibrillatory myocardial dysfunction. Nevertheless, postfibrillatory myocardial dysfunction occurred even in the absence of ischemia or acidosis [<xref ref-type="bibr" rid="R11">11</xref>,<xref ref-type="bibr" rid="R15">15</xref>]. These results suggests that at least part of the postfibrillatory myocardial dysfunction is due to reduced myofilament Ca<sup>2+</sup> responsiveness following VF-induced myocyte Ca<sup>2+</sup> overload. Therefore, postfibrillatory myocardial dysfunction presumably is a consequence of reduced myofilament Ca<sup>2+</sup> responsiveness following myocyte Ca<sup>2+</sup> overload caused by both VF and consequent ischemia. Further contributions to postfibrillatory myocardial dysfunction may arise from ischemia-induced oxygen free radicals and/or from the impact of electrical defibrillation shocks. However, it seems that only high energy shocks of monophasic waveform may precipitate myocardial injury and dysfunction. Biphasic shocks of normal energy (up to 15 J/g wet heart weight, that is about 2-5 times the energy used in implantable defibrillator devices in relation to the myocardial mass) [<xref ref-type="bibr" rid="R15">15</xref>] appear not to significantly contribute to this dysfunction [<xref ref-type="bibr" rid="R15">15</xref>,<xref ref-type="bibr" rid="R24">24</xref>].</p></sec><sec sec-type="" id="s5"><title>Comparison to atrial fibrillation</title><p>The same vicious circle described above for myocyte Ca<sup>2+</sup> and VF may also apply to atrial fibrillation (AF). It has been proposed that with increasing duration of paroxysmal AF, the likelihood that it will become chronic increases in parallel[<xref ref-type="bibr" rid="R25">25</xref>]. This so called "AF begets AF" concept seems to be due to <italic>electrical remodeling</italic> of the atria involving myocyte Ca<sup>2+</sup> overload[<xref ref-type="bibr" rid="R26">26</xref>]. Atrial electrical remodeling seems also to be responsible for reinitiating AF after cardioversion because pretreatment with a Ca<sup>2+</sup> channel blocker [<xref ref-type="bibr" rid="R26">26</xref>-<xref ref-type="bibr" rid="R28">28</xref>] or ryanodine (sarcoplasmic reticulum Ca<sup>2+</sup> release blocker) [<xref ref-type="bibr" rid="R29">29</xref>] attenuated acute AF-induced electrophysiological changes [<xref ref-type="bibr" rid="R26">26</xref>-<xref ref-type="bibr" rid="R28">28</xref>] and reduced [<xref ref-type="bibr" rid="R27">27</xref>] or even eliminated [<xref ref-type="bibr" rid="R29">29</xref>] re-initiation of AF. Furthermore, similar to postfibrillatory ventricular dysfunction described above, depressed atrial contractile function persists after AF and cardioversion [<xref ref-type="bibr" rid="R30">30</xref>] even though atrial tissue is not (or only to a minor degree) ischemic during AF. Accordingly, atrial contractile dysfunction was reduced by verapamil but increased by the Ca<sup>2+</sup> agonist BAY K8644, [<xref ref-type="bibr" rid="R30">30</xref>] suggesting that transsarcolemmal Ca<sup>2+</sup> influx contributed to this dysfunction. All these observations fit in our vicious circle concept, whereby AF-induced Ca<sup>2+</sup> overload may be responsible for the progressive nature of AF and for the re-initiation of AF as well as atrial contractile dysfunction after cardioversion.</p></sec><sec sec-type="" id="s6"><title>Species considerations</title><p>Extrapolation of the experimental findings in this article to the clinical setting should consider inter-species differences of excitation-contraction coupling. Most of the findings that underlie the proposed concept of Ca<sup>2+</sup> and VF forming a vicious circle, arise from studies in cardiomyocytes or in isolated hearts of rats or ferrets. Importantly, the central role of Ca<sup>2+</sup> in excitation-contraction coupling involving Ca<sup>2+</sup>-induced Ca<sup>2+</sup> release in cardiac muscle physiology [<xref ref-type="bibr" rid="R31">31</xref>] suggests that VF leads to myocyte Ca<sup>2+</sup> overload in most species including adult human beings. However, important inter-species and developmental differences exist regarding Ca<sup>2+</sup>-induced Ca<sup>2+</sup> release from the sarcoplasmic reticulum and regarding Ca<sup>2+</sup> removal processes [<xref ref-type="bibr" rid="R31">31</xref>]. For example, Ca<sup>2+</sup>- induced Ca<sup>2+</sup> release was absent in frog or prenatal rat ventricle, intermediate in human ventricle, and most prominent in adult rat ventricular myocytes [<xref ref-type="bibr" rid="R31">31</xref>]. Consequently, activator Ca<sup>2+</sup> in cardiac muscle of various species depends on different contributions from the sarcoplasmic reticulum, from L-type Ca<sup>2+</sup> channels, and from forward Na+/Ca<sup>2+</sup> exchange [<xref ref-type="bibr" rid="R31">31</xref>]. The kinetics and the degree of VF-induced myocyte Ca<sup>2+</sup> overload may therefore vary among species and be part of the reason why some species are better protected against sustained VF than others (similar to myocardial mass). Based on the foregoing, VF in adult human beings most likely induces cardiomyocyte Ca<sup>2+</sup> overload and the Ca<sup>2+</sup> sources of this overload may be slightly but not fundamentally different from adult rat ventricles. Finally, similarities of myofilament Ca<sup>2+</sup> sensitivity [<xref ref-type="bibr" rid="R31">31</xref>] and molecular mechanisms for myocardial stunning among species [<xref ref-type="bibr" rid="R21">21</xref>] suggest that VF-induced myocyte Ca<sup>2+</sup> overload is an important and species-independent part of postfibrillatory myocardial stunning. This is why the proposed vicious circle of Ca<sup>2+</sup> overload is likely to be of importance in human VF (and may be in AF too).</p></sec><sec sec-type="" id="s7"><title>Conclusions and clinical implications</title><p>Cardiomyocyte Ca<sup>2+</sup> and VF are mutually related forming a self-maintaining vicious circle in the initiation, maintenance, and termination of VF. On the one hand, various experimental studies have shown that elevated myocyte Ca<sup>2+</sup> can cause delayed afterdepolarizations, triggered activity, and consequently life-threatening ventricular tachyarrhythmias. On the other hand, VF itself directly and rapidly causes progressive myocyte Ca<sup>2+</sup> overload that maintains VF and renders termination of VF increasingly difficult. Accordingly, VF-induced myocyte Ca<sup>2+</sup> overload can cause electrical defibrillation to fail. Furthermore, VF-induced myocyte Ca<sup>2+</sup> overload can promote re-induction of VF after defibrillation and/or postfibrillatory myocardial dysfunction (postresuscitation stunning) due to reduced myofilament Ca<sup>2+</sup> responsiveness. The same vicious circle may apply to AF during electrical remodeling to cause re-initiation of AF and atrial contractile dysfunction after cardioversion. The experimental studies reviewed here suggest that the probability of these adverse events is best reduced by early detection and rapid termination of VF to prevent or limit Ca<sup>2+</sup> overload. In other words, rapid termination of VF not only reduces the risk of cerebral injury but also the likelihood of failed defibrillation and postfibrillatory myocardial dysfunction. Early additional therapy targeting transsarcolemmal Ca<sup>2+</sup> entry, particularly during first 2 min of VF, may partially prevent myocyte Ca<sup>2+</sup> overload and thus, increase the likelihood of successful defibrillation as well as prevent postfibrillatory myocardial dysfunction. Thereafter, the vicious circle of myocyte Ca<sup>2+</sup> overload is established and therapeutic attempts have an inherently lower likelihood of success.</p></sec> |
Interesting Electrophysiological Findings in a Patient With Coincidental Right Ventricular Outflow Tract and Atrioventricular Nodal Reentrant Tachycardia | <p>Tachycardia induced tachycardias are not common in clinical practice, and it is believed that most cases of double tachycardia are coincidental. The existence of two different tachycardias in the same patient almost always poses problems in the electrophysiology laboratory. However, in rare instances, the emergence of a second tachycardia can actually provide invaluable information about the first one. In this report, we describe a 30-year-old woman who presented with palpitations. Electrophysiological study revealed that atrial programmed stimulation at baseline induced right ventricular outflow tract (RVOT) tachycardia and supraventricular tachycardia. The study also showed that each of the tachycardias was able to induce the other. A short run of RVOT tachycardia during supraventricular tachycardia was able to entrain the latter. This finding provided important information about the nature of the supraventricular tachycardia, which proved to be atrioventricular nodal reentrant tachycardia. Both of these tachycardias were successfully ablated, and the patient’s palpitations disappeared.</p> | <contrib contrib-type="author"><name><surname>Ozin</surname><given-names>Bulent</given-names></name><degrees>MD</degrees><role>Associate Professor in cardiology</role></contrib><contrib contrib-type="author"><name><surname>Pirat</surname><given-names>Bahar</given-names></name><degrees>MD</degrees><role>Resident in cardiology</role></contrib><contrib contrib-type="author"><name><surname>Muderrisoglu</surname><given-names>Haldun</given-names></name><degrees>MD</degrees><role>Professor in cardiology</role></contrib> | Indian Pacing and Electrophysiology Journal | <sec sec-type="" id="s1"><title>Introduction</title><p>Simultaneous occurrence of supraventricular and ventricular tachycardia in the same patient is very rare. Concomitance of both these forms is usually coincidental, except in cases of digitalis toxicity and catecholaminergic tachycardias [<xref ref-type="bibr" rid="R1">1</xref>,<xref ref-type="bibr" rid="R2">2</xref>]. We report a case of double tachycardia in which right ventricular outflow tract (RVOT) tachycardia and atrioventricular nodal reentrant tachycardia (AVNRT) were capable of initiating each other.</p></sec><sec sec-type="" id="s2"><title>Case Report</title><p>A 30-year-old woman was referred to our hospital for evaluation of non-sustained ventricular tachycardia. She complained of palpitations lasting 20 to 30 minutes, and said she had had this problem for 2 years. Episodes of non-sustained ventricular tachycardia were documented during an exercise stress test. She was treated with verapamil and metoprolol with no success. Her physical examination was normal and an echocardiogram revealed no structural abnormalities. Electrocardiography showed frequent monomorphic ventricular premature beats of left bundle-branch block and inferior axis configuration.</p><p>An electrophysiological study was performed with the patient in postabsorptive, non-sedated state. An EP Tracer electrophysiology system (Maastricht, The Netherlands) was used. During the study, sinus rhythm was very frequently interrupted by spontaneous repetitive monomorphic non-sustained ventricular tachycardias (NSVT) with left bundle-branch block and inferior axis configuration. Mapping studies performed using an ablation catheter (Blazer II, Boston Scientific, USA) revealed that the ectopic activity originated from the RVOT. However, attempts to precisely map these premature depolarizations were frequently interrupted by a supraventricular tachycardia of cycle length 280 ms, which was started by the NSVT. The supraventricular tachycardia had a ventriculoatrial interval of 45 ms, which suggested either AVNRT or atrial tachycardia as the pathophysiologic mechanism.</p><p>Due to the frequent ventricular premature beats and runs of NSVT, it was not possible to further analyze and identify the exact nature of this supraventricular tachycardia. In order to continue mapping for the ventricular tachycardia, it was necessary to terminate the supraventricular tachycardia by overdrive pacing each time. There were occasions when the supraventricular tachycardia initiated the NSVT and vice versa (<xref ref-type="fig" rid="F1">Figure 1 and B</xref>). An episode of NSVT during supraventricular tachycardia was observed to entrain the latter with a pattern compatible with AVNRT (<xref ref-type="fig" rid="F2">Figure 2</xref>) [<xref ref-type="bibr" rid="R3">3</xref>].</p><p>The ventricular tachycardia was eliminated after 14 radiofrequency applications (EPT, USA) were delivered to the RVOT. Ventricular programmed stimulation after the ablation revealed no inducible ventricular tachycardias and demonstrated that the ventriculoatrial conduction was decremental. Atrial programmed stimulation showed dual atrioventricular nodal physiology and inducible AVNRT. A single radiofrequency application to the right posterior septum eliminated the dual nodal physiology and AVNRT. At a follow-up check 1 month after the procedure, the patient was free of palpitations.</p></sec><sec sec-type="" id="s3"><title>Discussion</title><p>Tachycardia induced tachycardia is a rare phenomenon that is only discussed in a few case reports in the literature. Wagshal and colleagues described one patient with idiopathic left ventricular tachycardia and AVNRT [<xref ref-type="bibr" rid="R4">4</xref>]. In that case, each of the tachycardias transformed into the other spontaneously. Washizuka and co-workers reported another patient who had RVOT tachycardia and atrioventricular reciprocating tachycardia [<xref ref-type="bibr" rid="R5">5</xref>]. Similar to what we observed in our case, the ventricular tachycardia in this patient was able to entrain the supraventricular tachycardia.</p><p>Cooklin and McComb recently reported a case in which AVNRT was induced by RVOT tachycardia during infusion of isoprenaline in an electrophysiological study [<xref ref-type="bibr" rid="R6">6</xref>]. Our patient also exhibited both AVNRT and RVOT tachycardia; however, in contrast to the case reported by Cooklin and McComb, both these tachycardias were capable of inducing each other. In addition, there are a few other discrepancies between that case and ours. First, our patient exhibited excellent ventriculoatrial conduction, which could easily explain the mechanism of induction of AVNRT during RVOT tachycardia. Second, neither isoprenaline nor any other beta-mimetic agents were used during our electrophysiological study, and induction of RVOT tachycardia by AVNRT took place at the basal state. It is well-known that triggered activity is the mechanism behind RVOT tachycardia, and rapid atrial pacing in a patient with good atrioventricular conduction might induce this form of tachycardia. In our case, we believe that RVOT tachycardia was induced by the rapid atrial depolarizations conducted to the ventricle during AVNRT.</p><p>The consecutive occurrence of two different tachycardias during electrophysiological testing always presents a major challenge for the electrophysiologist. During our electrophysiological study, we aimed to ablate the RVOT tachycardia first since it was responsible for the patient’s main symptoms. However, our attempts to locate the earliest ventricular activating sites during RVOT tachycardia were frequently interrupted by intervening AVNRT. To address this, we changed our strategy and decided to ablate the supraventricular tachycardia first, but the frequent premature beats and ventricular tachycardias arising from the RVOT prevented analysis of the mechanism of the supraventricular tachycardia. Thus, we had to change our plan again, and finally ablated the RVOT tachycardia after 14 radiofrequency applications.</p><p>Apart from the problems mentioned above, we observed that double tachycardias actually helped establish the diagnosis during electrophysiological study. In our case, a three-beat run of RVOT tachycardia during the supraventricular tachycardia entrained the latter with an ‡atrial-ventricular’ pattern, and helped us diagnose AVNRT (<xref ref-type="fig" rid="F2">Figure 2</xref>).</p><p>In conclusion, this case is of particular interest because of the unusual finding of double tachycardia, with each form capable of inducing the other. The case is also noteworthy in that it highlights the advantages and disadvantages of dealing with two different tachycardias in the electrophysiology laboratory.</p></sec> |
Devices for Prevention of Atrial Tachyarrhythmias | Could not extract abstract | <contrib contrib-type="author"><name><surname>Lozano</surname><given-names>Ignacio Fernández</given-names></name></contrib><contrib contrib-type="author"><name><surname>Toquero</surname><given-names>Jorge</given-names></name></contrib><contrib contrib-type="author"><name><surname>Diaz</surname><given-names>José Antonio Fernández</given-names></name></contrib><contrib contrib-type="author"><name><surname>Ionescu</surname><given-names>Bogdan</given-names></name></contrib><contrib contrib-type="author"><name><surname>Moñivas</surname><given-names>Vanesa</given-names></name></contrib><contrib contrib-type="author"><name><surname>Ortiz</surname><given-names>Pilar</given-names></name></contrib><contrib contrib-type="author"><name><surname>Fuertes</surname><given-names>Beatriz</given-names></name></contrib><contrib contrib-type="author"><name><surname>Pulpón</surname><given-names>Luis Alonso</given-names></name></contrib> | Indian Pacing and Electrophysiology Journal | <sec sec-type="" id="s1"><title>Introduction</title><p>Atrial fibrillation (AF) is the most frequent sustained cardiac arrhythmia in clinical practice and, although its importance has been underestimated even in recent years, we are now becoming aware of its clinical transcendence [<xref ref-type="bibr" rid="R1">1</xref>-<xref ref-type="bibr" rid="R3">3</xref>]. The classical treatment is pharmacological, but its efficacy is limited and it does have side effects [<xref ref-type="bibr" rid="R4">4</xref>-<xref ref-type="bibr" rid="R5">5</xref>]. Therefore, in recent years, there has been an increasing interest in other types of non-pharmacological treatments [<xref ref-type="bibr" rid="R6">6</xref>,<xref ref-type="bibr" rid="R7">7</xref>].</p><p>Physiologic cardiac pacing has proven to be more effective than VVI mode pacing to prevent the occurrence of AF during the follow-up of patients who have had a permanent pacemaker implanted [<xref ref-type="bibr" rid="R8">8</xref>-<xref ref-type="bibr" rid="R10">10</xref>]. There are currently different lines of research that use different atrial pacing techniques to prevent and treat episodes of paroxysmal atrial fibrillation [<xref ref-type="bibr" rid="R11">11</xref>,<xref ref-type="bibr" rid="R12">12</xref>]. Techniques of multi-site pacing in the right atrium or both atria, new atrial pacing sites, prevention algorithms for paroxysmal atrial fibrillation episodes, and even high-frequency atrial tachyarrhythmia termination algorithms have all been proposed. In this article, we will try to synthesize the grounds for and findings of the different lines of research currently being developed.</p></sec><sec sec-type="" id="s2"><title>Pacing Activation Mechanisms</title><p>The occurrence of atrial fibrillation in a patient depends on the interaction of three factors: substratum, triggers and modulating factors. The substratum is the usually sick atrial myocardium with areas of fibrosis that result in heterogeneous refractoriness that facilitates the appearance of functional reentry areas. Shortening the refractory periods that take place in atrial myocardium as a result of electrical remodeling favors the perpetuation of AF [<xref ref-type="bibr" rid="R13">13</xref>]. The triggers are the atrial extra-systoles or episodes of atrial tachycardia or flutter that in many cases precede episodes of AF [<xref ref-type="bibr" rid="R14">14</xref>]. The modulating factors such as sympathicotonia or circulating catecholamines facilitate the occurrence and sustainability of AF.</p><p>Atrial pacing can prevent the development of AF through several mechanisms. It can prevent dispersion of the atrial refractory periods associated with bradycardia. This effect can be particularly beneficial in cases of vagotonic AF [<xref ref-type="bibr" rid="R15">15</xref>,<xref ref-type="bibr" rid="R16">16</xref>]. New approaches of interatrial septal, Bachmann or simultaneous dual-site pacing have demonstrated an enhanced speed of conduction of the electric pulse through the atrium. This higher speed of conduction is reflected in a shorter duration of the P wave of the superficial electrocardiogram [<xref ref-type="bibr" rid="R17">17</xref>-<xref ref-type="bibr" rid="R19">19</xref>] and can help to prevent AF.</p><p>Atrial pacing at frequencies higher than the basal frequency can prevent the occurrence of AE by overdrive suppression, or at least significantly decrease their number. It can also decrease the post-AE pauses that result in high dispersion in the refractory periods. In addition, it can suppress the increased automatism that is responsible for focal AF [<xref ref-type="bibr" rid="R20">20</xref>].</p><p>The new techniques of multi-site pacing can improve the hemodynamics of sick atria, and this hemodynamic improvement can in itself prevent episodes of AF [<xref ref-type="bibr" rid="R21">21</xref>]. Finally, the episodes of high-frequency atrial tachycardia that often precede the occurrence of AF can, on occasions, be terminated with atrial pacing [<xref ref-type="bibr" rid="R22">22</xref>].</p></sec><sec sec-type="" id="s3"><title>Patients with Permanent Pacing Indication</title><p>In 1994, Andersen and Cols [<xref ref-type="bibr" rid="R23">23</xref>] published the results of the first randomized trial that compared physiologic pacing with VVI mode. It included 225 patients with sinus node dysfunction and narrow QRS and an average age of 76 years who were randomized to receive an AAI versus VVI pacemaker. The long-term follow-up of these patients was published in 1997 [<xref ref-type="bibr" rid="R8">8</xref>]; the outcome, after 5.5 years of follow-up, was a lower rate of AF, a lower cardiovascular mortality and a significant reduction in overall mortality in patients randomized to the AAI modality. Subsequent publications of this same population showed a lower number of cases of progression to heart failure, improved echocardiographic parameters [<xref ref-type="bibr" rid="R24">24</xref>] and fewer thromboembolism events [<xref ref-type="bibr" rid="R25">25</xref>].</p><p>In 1998, the results of the PASE (“Pacemaker Selection in the Elderly”) study were published [<xref ref-type="bibr" rid="R26">26</xref>]. This study randomized 407 patients over 65 years of age and with bradycardia to receive a DDDR versus VVIR pacemaker. After an average follow-up of 2.5 years, the quality of life endpoint significantly improved in both groups (p<0.001) compared to the baseline situation, but no significant differences were detected between the two groups except for a slight difference in favor of using the DDDR pacemaker in patients with SND. No differences were found in number of deaths, ACVA or admissions for heart failure. However, a lower incidence of AF in the DDDR randomized group was demonstrated. The results of the study were partially clouded because a very high rate of crossover (26%) from the VVIR to DDDR group was recorded due to pacemaker syndrome during follow-up. This may have undervalued the benefit of DDDR pacing in this type of patient.</p><p>That same year, Mattioli published the results of a prospective study that analyzed the incidence of atrial fibrillation as a function of the pacing modality [<xref ref-type="bibr" rid="R27">27</xref>]. It included 210 patients, 110 with SND and 100 with AV block, who received a physiologic PM (AAI, DDD, DDDR or VDD) versus a ventricular PM (VVI or VVIR). Patients with a background of atrial fibrillation were excluded, and the incidence of AF was 10% during the first year, 23% after 3 years and 31% after 5 years. The patients randomized to a physiologic PM had a lower risk of developing AF during follow-up, and the greatest benefits were for patients with SND.</p><p>The CTOPP (Canadian Trial of Physiologic Pacing) [<xref ref-type="bibr" rid="R28">28</xref>] randomized 2568 patients with pacemaker indication, 1474 to ventricular pacing (VVI or VVIR) and 1094 to physiologic pacing (AAI, AAIR, DDD or DDDR). After an average follow-up of 3 years, no significant differences were found in the combined endpoint of cardiovascular mortality and ACVA between both groups. However, the annual incidence of atrial fibrillation was 5.3% in the physiologic pacing group versus 6.6% in the ventricular pacing group. A relative risk reduction of 18% was statistically significant (p=0.05). A subsequent analysis of this same study published in 2001 [<xref ref-type="bibr" rid="R29">29</xref>] demonstrated that physiologic pacing was very beneficial for the patients most dependent on pacemakers. Thus, those patients with an intrinsic cardiac frequency of less than 60 bpm during the first follow-up had a lower rate of cardiovascular death or ACVA and a lower total mortality (p<0.001).</p><p>The MOST trial, published in 2002, randomizes 2010 patients with SND to receive a VVIR versus DDDR pacemaker [<xref ref-type="bibr" rid="R30">30</xref>]. After completing an average follow-up of 2.7 years, the primary endpoint of ACVA or total mortality occurred in 22.2% of the patients, and no significant differences were observed between the two groups (p=0.32). However, a lower incidence of AF was observed in the DDDR randomized group (p=0.008). The occurrence of permanent AF was 26.7% in the VVIR group versus 15.2% in the DDDR group (p=0.001).</p><p>The preliminary findings of the UK-PACE trial [<xref ref-type="bibr" rid="R31">31</xref>] were reported during the 2003 Congress of the American College of Cardiology. This trial randomizes 2000 patients with AV block to a DDDR versus VVIR pacemaker, and no differences are found in total or cardiac mortality between the two groups.</p><p>In all, almost 7000 patients have been included in randomized trials, and there is indisputable proof of a lower incidence of atrial fibrillation and less progression to chronic AF when a physiologic pacemaker is used in patients with permanent pacing indication.</p></sec><sec sec-type="" id="s4"><title> Patients Without Associated Bradycardia</title><p>In 1999, the results of PA<sup>3</sup> [<xref ref-type="bibr" rid="R32">32</xref>] were published. This study assesses the efficacy of atrial pacing in patients with paroxysmal atrial fibrillation and without permanent pacing indication. In order to be included, the patients had to have had at least three episodes of paroxysmal AF during the preceding year in spite of having received correct antiarrhythmic treatment. A dual-chamber pacemaker was implanted in 97 patients, and they were randomized to DDIR programming at 70 bpm or DDI at 30 bpm. The primary endpoint was the time up to the first recurrence of an atrial tachycardia lasting more than 5 minutes. The counters of the device, which did not have intracavity electrographs, were used for this purpose. This endpoint did not differ significantly in the two groups.</p><p>Subsequently, AV node ablation was performed in 67 patients, randomizing the DDDR PM at 70 bpm minimum frequency or VDD PM at 60 bpm [<xref ref-type="bibr" rid="R33">33</xref>]. There were also no differences detected in the recurrence of AF in this second phase.</p><p>This study has some methodological limitations, the main one being a high rate - 25% - of group crossover from the DDI mode at 30 bpm to the DDIR mode at 70 bpm. In addition, the DDIR programming only achieved 67% atrial pacing, and other studies have demonstrated that the higher the percentage of pacing, the greater is the efficacy in preventing AF. Finally, the primary endpoint is based on analysis of the PM counters, without electrographs that confirm the recurrence of the arrhythmia.</p><p>In spite of all this, an analysis of the results of PA<sup>3</sup> puts in doubt the efficacy of atrial pacing in preventing episodes of AF in patients with associated bradycardia.</p></sec><sec sec-type="" id="s5"><title>Pacing at Special Sites</title><sec id="s5a"><title>A) Biatrial Pacing</title><p>The association between disorders of intra- and inter-atrial conduction and the occurrence of AF episodes has been known for some years [<xref ref-type="bibr" rid="R34">34</xref>]. Techniques of simultaneous pacing in both atria [<xref ref-type="bibr" rid="R35">35</xref>] have been proposed to correct this. This technique uses an electrode located inside the coronary sinus and another in the right atrium; both are jointly connected to the atrial port by a Y-connector. Some second-generation PMs have algorithms that resynchronize both atria in case an event sensed in the opposite atrium occurs [<xref ref-type="bibr" rid="R36">36</xref>].</p><p>The French group directed by Dr. Daubert publishes the experience of a single center over a period of 9 years [<xref ref-type="bibr" rid="R37">37</xref>]. It demonstrates a reduction of the P wave from 187 to 106 msec with biatrial pacing. After an average follow-up of 34 months, a third of the patients is free of arrhythmias, a third presents paroxysmal episodes but with maintenance of sinus rhythm, and the remaining third has permanent atrial fibrillation.</p><p>This technique has also been effectively used to prevent episodes of AF after heart surgery [<xref ref-type="bibr" rid="R38">38</xref>-<xref ref-type="bibr" rid="R40">40</xref>]. However, the technique is limited by the need for 2 electrodes, the high rate of displacement of the coronary sinus electrode and the use of Y-connectors.</p></sec><sec id="s5b"><title>B) Dual-Site Right Atrial Pacing</title><p>In this technique, an electrode is implanted in the right appendage and another in the coronary sinus ostium. This succeeds in reducing the conduction time and achieves a more homogeneous atrial activation, thus eliminating the displacement problems of electrodes placed inside the coronary sinus.</p><p>Early publications [<xref ref-type="bibr" rid="R19">19</xref>] demonstrated a beneficial effect of dual-site pacing versus single-site pacing or no pacing. In a prospective randomized study that includes 118 patients, dual-site pacing only demonstrates a tendency for longer times without arrhythmias. In the subgroup of patients that continues under antiarrhythmic treatment, dual-site pacing from the right atrium significantly reduces the number of atrial arrhythmia episodes.</p></sec><sec id="s5c"><title>C) Pacing at Special Sites</title><p>The atrial pacing site is of utmost importance to the total conduction time. To try to achieve the potential beneficial effects of dual-site pacing, different pacing sites in the right atrium have been proposed, specifically Bachmann's bundle [<xref ref-type="bibr" rid="R18">18</xref>] and the coronary sinus ostium [<xref ref-type="bibr" rid="R42">42</xref>]. In a randomized study, Padeletti analyzes the effects of pacing in 46 patients with a history of AF. During the pacing phase in the right appendage, the number of symptomatic AF episodes was reduced from 6 to 2 a month, whereas with interatrial septum pacing the number was reduced from 5 to 0.2 a month. In another study [<xref ref-type="bibr" rid="R44">44</xref>], septal pacing resulted in a subjective 68% improvement of symptomatic episodes and an objective 60% reduction of the incidence of AF.</p><p>The results with Bachmann's bundle pacing are promising [<xref ref-type="bibr" rid="R45">45</xref>]. The electrode is easily positioned by using a simple fluoroscopic reference, and this position is associated with a marked reduction in the P wave duration. In a prospective study that included 120 patients, pacing from Bachmann's bundle reduced the incidence of progression to permanent AF (45 versus 75%, p<0.05) after an average follow-up of one year.</p><p>In short, both techniques have proved to be somewhat useful, although in trials with short follow-up periods and with a small number of patients. Of the two positions, Bachmann's bundle offers the advantage of causing fewer problems in ventricular signal detection. The far field can be significant in this group of patients; it limits the utility of mode exchange algorithms and introduces incorrect information in the pacemaker event counters.</p></sec></sec><sec sec-type="" id="s6"><title>Prevention Algorithms</title><p>Different atrial pacing algorithms have been developed, all designed for atrial overdrive pacing and prevention of AF episodes. Basically they can be divided into four types [<xref ref-type="bibr" rid="R46">46</xref>]: atrial overdrive pacing algorithms (<xref ref-type="fig" rid="F1">Figure 1</xref>), response to atrial extrasystoles (to prevent short-long sequences), response to sinus rhythm recovery, and post-exercise relative bradycardia sequence prevention algorithms. The most commonly used algorithm is intended to stimulate the atrium just above the sinus frequency to achieve unified atrial refractory periods. The higher the atrial pacing percentage, the greater the effectiveness. However, excessively high frequencies are perceived by the patient as a disagreeable sensation. Therefore, the objective is to achieve a high percentage of atrial pacing without excessively increasing the patient's frequency.</p><p>The efficacy of an atrial overdrive pacing algorithm has been analyzed in the ADOPT-A trial [<xref ref-type="bibr" rid="R47">47</xref>]. A total of 288 patients with sinus node dysfunction and a background of paroxysmal or persistent AF received a DDDR pacemaker with a lower frequency programmed at 70 bpm. In a parallel design, the patients were randomized to activate or not activate the overdrive pacing algorithm. After a 6-month follow-up, it was confirmed that activation of the algorithm caused an increase in the pacing percentage of 92.9% versus 67.9% (p<0.001). The atrial fibrillation percentage dropped 25% with the algorithm activated (2.5% vs 1.87%, p=0.005). The number of rehospitalizations or the need for cardioversion did not differ between the two groups.</p><p>The AFT (AF Therapy) Trial included 372 patients with paroxysmal AF, with or without permanent pacing indication, in a complex protocol divided into 4 phases [<xref ref-type="bibr" rid="R48">48</xref>]. The first phase of the study was a monitoring phase intended to obtain information on the way AF begins in this population. The second phase compares the efficacy of conventional pacing. During the third phase of the study, the efficacy of four combined pacing algorithms is analyzed and compared to conventional DDDR pacing at 70 bpm. Unfortunately, it has only been possible to analyze data from 97 patients because information was lost during follow-up and because of detection of ventricular activity in the atrial channel. The prevention algorithms achieved a 34% reduction of the AF rate (p<0.05). A subsequent subanalysis of this study [<xref ref-type="bibr" rid="R49">49</xref>] showed how conventional pacing was very effective in patients with AF and bradycardia, whereas the algorithms tended to be more useful for those patients with prior AF and without bradycardia.</p><p>The ASPECT trial (Atrial Septal Pacing Efficacy Clinical Trial) included 298 patients who were randomized to receive the atrial electrode in the septum or another position in the right atrium. After one month, those patients with recurrence of AF were randomized to conventional pacing or algorithms activated during a 3-month phase. At the end of this phase, the PM of the opposite group was reprogrammed. The PM counters did not show evidence of any significant differences with activation of the algorithms, nor were any objective differences based on the electrode position found, except that the symptomatic episodes were less frequent in those patients who had an electrode implanted in the interatrial septum.</p></sec><sec sec-type="" id="s7"><title>PREVENT Register</title><p>Atrial fibrillation is a complex arrhythmia. Phase 1 of the AFT study [<xref ref-type="bibr" rid="R48">48</xref>] demonstrated that there are multiple mechanisms of AF initiation that differ in different patients. Nowadays, modern PMs have numerous diagnostic functions that help to characterize the way AF begins in a specific patient. In addition, the prevention algorithms generally vary, and it may be that not all of them are beneficial in all patients. The prospective studies have a strict protocol that makes it difficult to optimize programming for a specific patient. Therefore, to study the efficacy of the AF prevention algorithms in the real world by using all the advantages of a PM in a particular patient, we have designed the Prevent Register [<xref ref-type="bibr" rid="R51">51</xref>]</p><p>The Prevent-AF Register is prospective, non-randomized and multicentric. It includes candidate patients for permanent pacing due to sinus node dysfunction (type I or IIa indication of the AHA-ACC) [<xref ref-type="bibr" rid="R52">52</xref>] with or without prior paroxysmal atrial fibrillation.</p><p>Four preventive pacing algorithms are incorporated: Pace Conditioning™, Post-PAC Response™, PAC Suppression™, Post-Exercise Response™. The Pace Conditioning detects the patient's basal frequency and increases the pacemaker pacing frequency to increase the percentage of paced beats, in an attempt to maintain patients slightly above their basal frequency. The PAC Suppression is activated if atrial extrasystoles are detected, increasing the basal pacing frequency in an attempt to suppress them. The Post-PAC Response prevents the compensatory pause after an atrial extrasystole by eliminating the short-long sequences. The Post-Exercise Rate Control prevents abrupt frequency drops after episodes of physical exercise. The Register was implemented in 14 centers in Spain. A total of 68 patients (33 men, 35 women) were included from April 2000 to April 2001. The average age was 72±12 years and the pacing indication was sinus node dysfunction in all cases. A total of 53 patients (78%) had had documented paroxysmal AF episodes prior to implant.</p><p>The recurrence of at least one episode of paroxysmal atrial fibrillation was documented in a total of 32 patients during the monitoring phase, and at least one subsequent phase of preventive pacing was completed. The preventive algorithms were programmed on an individualized basis, based on information from the AF episode commencement mode.</p><p>The average atrial arrhythmic burden was reduced in the total group (32 patients) from 0.94 to 0.3 hours a day (Wilcoxon test: p=0.034) (<xref ref-type="fig" rid="F2">Figure 2</xref>). In relative values, the average atrial arrhythmic burden dropped from 3.9% to 1.3%, i.e. a 67% reduction. The mean atrial pacing percentage rose from 72% to 78%, i.e. an increase of only 13%.</p><p>The Register results confirm the usefulness of latest-generation pacemakers both for diagnosing episodes of paroxysmal atrial fibrillation and for preventing them by means of pacing algorithms. The diagnostic functions of the pacemaker provided relevant information on the occurrence and development of atrial fibrillation in each patient. In this way, the most suitable algorithms for preventing AF were programmed on an individualized basis.</p></sec><sec sec-type="" id="s8"><title>Atrial Tachycardia Termination Algorithms</title><p>Many episodes of AF begin preceded by a rapid, relatively regular atrial tachycardia. Some of these atrial tachycardias can be eliminated by overdrive pacing in a way similar to overdrive pacing and termination of other tachycardias by reentry. To end atrial tachycardias, algorithms similar to those used in implantable defribillators have been employed, and specific algorithms have been designed to act in the atrium as 50 Hz bursts (<xref ref-type="fig" rid="F3">Figure 3</xref>).</p><p>These therapies have been tested in 537 patients possessing a DAI with atrial therapies [<xref ref-type="bibr" rid="R53">53</xref>]. After an average follow-up of 11.4 months, the overall efficacy of the algorithms was 48%, with 59% in cases of atrial tachycardia and 30% in those cases classified as AF. The efficacy depends on the arrhythmia cycle length, and it is lower the faster the atrial tachycardia. The main limitation of this and similar work is that we know that many episodes of AF and atrial tachycardia are self-limited and brief, and therefore the quantification of efficacy may be clearly overestimated.</p><p>In the ATTEST study [<xref ref-type="bibr" rid="R54">54</xref>], the efficacy of atrial overdrive pacing in 368 patients possessing a pacemaker was evaluated. The overdrive pacing algorithms terminated 54% of the episodes, although the reduced percentage of atrial arrhythmias was not statistically significant: from 42 hours a month to 1.3 hours a month (p=0.2).</p></sec><sec sec-type="" id="s9"><title> Conclusions</title><p>As we have seen in this review of the state-of-the-art, several parallel lines of research have been developed in recent years to treat patients with AF, either associated or not with bradycardia. The situation may seem confusing, and even more so if we consider some limitations of this work.</p><p>The first limitation is that in each study, we are analyzing something different that has not been studied previously. The data that we have from classical prevalence studies [<xref ref-type="bibr" rid="R1">1</xref>-<xref ref-type="bibr" rid="R3">3</xref>] are based on records, either with ECG or by Holter, of AF episodes, most of them symptomatic. In most of the mentioned studies, the presence or absence of AF is analyzed according to records of the pacemaker itself. This has two limitations. The first is that, as the devices and detection criteria are different, the data from a study may not be comparable to data from other studies. The second is that, as most AF episodes are asymptomatic, we are overestimating the incidence and it is not easy to extrapolate the data from these studies, which use symptomatic episodes as diagnostic or efficacy criteria.</p><p>Another limitation is that atrial fibrillation is a complex, dynamic arrhythmia. For reasons of protocol, the programming of devices in most of the trials has been based on a rigid, inflexible protocol [<xref ref-type="bibr" rid="R55">55</xref>]. Perhaps if there had been a little more freedom on adapting the algorithms to each patient, the degree of efficacy of the algorithms would have been higher. In addition, we are increasingly more aware of the adverse effect of VD apex pacing on the hemodynamics of these patients [<xref ref-type="bibr" rid="R55">55</xref>]. This effect at least partially explains the high incidence of atrial arrhythmias in patients with a permanent PM that was not accounted for in some studies. Perhaps in the future, with new algorithms tending to reduce VD pacing to the necessary minimum and with the expansion of resynchronization techniques, the approaches for treating atrial arrhythmias may yield greater benefits.</p><p>Finally, the follow-up of these studies has been very short. In Andersen's classical study [<xref ref-type="bibr" rid="R8">8</xref>], the AAI pacemaker did not yield benefits until after 5 years of follow-up; however, the data we have on these new techniques are limited to 6-month or one-year follow-ups. We need a longer follow-up to be able to accurately evaluate the efficacy of these devices.</p><p>From the data presented to date, we can conclude that the benefits obtained by implanting two electrodes, either in the right atrium or in both atria, are insignificant and probably do not justify the increased complexity of the implant. These techniques should be limited to research protocols, to try to identify the type of patient that would benefit the most.</p><p>The prevention algorithms are sound and barely raise the PM price, although their utility is limited. Nowadays, they could be used for different categories of patients:
<list list-type="order"><list-item><p>Patients with PM indication and documented AF episodes.</p></list-item><list-item><p>Patients with a background of AF who develop secondary symptomatic bradycardia to medication.</p></list-item><list-item><p>Patients with paroxysmal atrial arrhythmias not suppressible by ablation and drug-refractory techniques for whom AV node ablation is considered as a therapeutic measure.</p></list-item></list></p><p>For the time being, the results of pacing do not justify its therapeutic use in patients with atrial fibrillation or atrial arrhythmias without associated bradycardia. One unanswered question is the optimum site of atrial pacing. Current data suggest a rather inconsistent benefit with septal pacing, whether it be from the coronary sinus ostium or Bachmann's bundle. But we still have very few data. Trials with more patients and longer follow-up periods are needed to change the current standard of implanting the atrial electrode in the appendage.</p></sec> |
Congenital Short QT Syndrome | Could not extract abstract | <contrib contrib-type="author"><name><surname>Antzelevitch</surname><given-names>Charles</given-names></name><degrees>PhD, FACC, FAHA</degrees><xref ref-type="aff" rid="aff1">*</xref></contrib><contrib contrib-type="author"><name><surname>Francis</surname><given-names>Johnson</given-names></name><degrees>MD, DM</degrees><xref ref-type="aff" rid="aff2">¶</xref></contrib> | Indian Pacing and Electrophysiology Journal | <p>Long QT intervals in the ECG have long been associated with sudden cardiac death. The congenital long QT syndrome was first described in individuals with structurally normal hearts in 1957 [<xref ref-type="bibr" rid="R1">1</xref>]. Little was known about the significance of a short QT interval. In 1993, after analyzing 6693 consecutive Holter recordings Algra et al concluded that an increased risk of sudden death was present not only in patients with long QT interval, but also in patients with short QT interval (< 400 ms) [<xref ref-type="bibr" rid="R2">2</xref>]. Because this was a retrospective analysis, further evaluation of the data was not possible.</p><p>It was not until 2000 that a short-QT syndrome (SQTS) was proposed as a new inherited clinical syndrome by Gussak et al [<xref ref-type="bibr" rid="R3">3</xref>]. The initial report was of two siblings and their mother all of whom displayed persistently short QT interval. The youngest was a 17 year old female presenting with several episodes of paroxysmal atrial fibrillation requiring electrical cardioversion [<xref ref-type="bibr" rid="R3">3</xref>]. Her QT interval measured 280 msec at a heart rate of 69. Her 21 year old brother displayed a QT interval of 272 msec at a heart rate of 58, whereas the 51 year old mother showed a QT of 260 msec at a heart rate of 74. The authors also noted similar ECG findings in another unrelated 37 year old patient associated with sudden cardiac death.</p><p>A review on the subject including a discussion of proposed mechanisms appeared in 2002 [<xref ref-type="bibr" rid="R4">4</xref>]. The review highlighted the lack of rate-dependence of QT interval in cases in which QT abbreviation was constant and the negative correlation of QT with RR in cases in which an abnormally short QT was evident only at bradycardic rates. Among the principal gene candidates proposed to underlie these syndromes were gain of function mutations of I<sub>Kr</sub>, I<sub>Ks</sub>, I<sub>K-ACh</sub> and I<sub>K-ATP</sub> [<xref ref-type="bibr" rid="R4">4</xref>] (<xref ref-type="fig" rid="F1">Figure 1</xref>). I<sub>K-ACh</sub> gain of function or other means by which the influence of the parasympathetic nervous system could be exaggerated was considered as the most likely mechanisms to explain the deceleration-dependent variant of the short QT syndrome.</p><p>The familial nature of this sudden death syndrome was confirmed by Gaita et al in 2003 [<xref ref-type="bibr" rid="R5">5</xref>]. They reported a study of six patients from two different families, with syncope, palpitations, resuscitated cardiac arrest and a positive family history for sudden cardiac death. The QT intervals never exceeded 280 msec or a QTc 300 msec. There was no evidence of structural heart disease in any family member. Electrophysiological evaluation yielded short atrial and ventricular refractory periods in all four patients who underwent testing. Three of them were also found to have increased vulnerability to ventricular fibrillation. Four patients received an automatic defibrillator (ICD) [<xref ref-type="bibr" rid="R5">5</xref>].</p><p>Factors that abbreviate the QT interval, including tachycardia, hyperthermia, hypercalcemia and digoxin should naturally be excluded before arriving at a diagnosis of congenital short QT syndrome. The mechanism of arrhythmogenesis in some cases of digitalis toxicity could be similar to those in congenital short QT syndrome [<xref ref-type="bibr" rid="R6">6</xref>].</p><p>The first gene responsible for the short QT syndrome was reported by Brugada et al in January of 2004 [<xref ref-type="bibr" rid="R7">7</xref>]. A candidate gene approach was used to screen for a causative mutation in the two families previously reported by Gaita et al [<xref ref-type="bibr" rid="R5">5</xref>]. Using direct sequencing techniques, two different missense mutations were uncovered in the two families, resulting in the same amino acid substitution in HERG (KCNH2), the gene encoding for the rapidly activating delayed rectifier channel, I<sub>Kr</sub> [<xref ref-type="bibr" rid="R7">7</xref>]. The substitution of lysine for asparagines at position 588 of KCNH2, was found to cause a loss of the normal rectification of the current at plateau voltages, thus resulting in a large increase of I<sub>Kr</sub> during the action potential plateau, leading to marked abbreviation of the action potential. The short QT syndrome is the first disease to be linked to a gain of function of I<sub>Kr</sub>. A third short QT family was not associated with a mutation in KCNH2, pointing to genetic heterogeneity of the disease. Interestingly, the N588K mutant channel proved to be insensitive to I<sub>Kr</sub> blockers such as d-sotalol both in the clinic and in heterologous expression systems. Quinidine, by virtue of its greater affinity for the open state of the channel, its additional I<sub>Ks</sub> blocking actions and anticholinergic activity was found to be more effective in reducing the repolarizing forces of the ventricle and prolonging the action potential [<xref ref-type="bibr" rid="R8">8</xref>].</p><p>It is noteworthy that substitution of aspartic acid for asparagine in the same position of HERG (N588D) has been linked to the LQT2 form of the long QT syndrome secondary to a loss of function of IKr [<xref ref-type="bibr" rid="R9">9</xref>]. This substitution leads to replacement of a neutral amino acid with a negatively charged one, whereas the N588K mutation responsible for short QT, secondary to gain of function of I<sub>Kr</sub>, replaces a neutral amino acid with a positively charged one.</p><p>Bellocq et al recently linked a second gene to the syndrome. A missense mutation in KCNQ1 (KvLQT1) giving rise to a gain of function in I<sub>Ks</sub>, the slowly activating delayed rectifier, was identified in a 70 year old male presenting with idiopathic ventricular fibrillation and short QT intervals in the ECG. Functional studies of the V307L KCNQ1 mutant revealed a -20 mV shift of the half-activation potential and an acceleration of the activation kinetics leading to an increase in I<sub>Ks</sub> [<xref ref-type="bibr" rid="R10">10</xref>].</p><p>A distinctive electrocardiographic feature of the short QT syndrome is the appearance of tall peaked T waves, similar to those encountered with hyperkalemia. ICD implantation is currently the treatment of choice for symptomatic patients with short QT syndrome and a family history of sudden cardiac death. Despite normal sensing behavior during device testing, inappropriate shocks due to T wave over sensing has been reported. This is due to the detection of short-coupled and prominent T waves [<xref ref-type="bibr" rid="R11">11</xref>]. The problem was corrected by reprogramming to lower sensitivity and decay delay. More specific diagnostic criteria will need to be developed as additional information about this syndrome becomes available in the coming months and years.</p> |
Radiofrequency Ablation for Post Infarction Ventricular Tachycardia | <p>Radiofrequency ablation has an important role in the management of post infarction ventricular tachycardia. The mapping and ablation of ventricular tachycardia (VT) is complex and technically challenging. In the era of implantable cardioverter defibrillators, the role of radiofrequency ablation is most commonly reserved as an adjunctive treatment for patients with frequent, symptomatic episodes of ventricular tachycardia. In this setting the procedure has a success rate of around 70-80% and a low complication rate. With improved ability to predict recurrent VT and improvements in mapping and ablation techniques and technologies, the role of radiofrequency ablation should expand further.</p> | <contrib contrib-type="author"><name><surname>O'Donnell</surname><given-names>David</given-names></name><degrees>MD</degrees></contrib><contrib contrib-type="author"><name><surname>Nadurata</surname><given-names>Voltaire</given-names></name><degrees>MD</degrees></contrib> | Indian Pacing and Electrophysiology Journal | <sec sec-type="" id="s1"><title>Introduction</title><p>Radiofrequency ablation (RFA) is the treatment of choice in the management of symptomatic patients with ventricular tachycardia (VT) in the absence of structural heart disease. The role of RFA in-patients with VT in the setting of chronic ischemic heart disease is less well defined. Ventricular tachycardia (VT) is a common complication of ischemic heart disease, with significant associated morbidity and mortality [<xref ref-type="bibr" rid="R1">1</xref>,<xref ref-type="bibr" rid="R2">2</xref>]. Traditionally, anti-arrhythmic medications have formed the mainstay of treatment, despite the low efficacy, the risk of pro-arrhythmia and long term adverse effects [<xref ref-type="bibr" rid="R3">3</xref>]. Anti-arrhythmic surgery is successful in abolishing VT, however, the operative mortality in most series was unacceptably high [<xref ref-type="bibr" rid="R4">4</xref>]. Over the past 10 years, implantable cardioverter defibrillators (ICDs) have become the treatment of choice for all but incessant VT. This treatment is based on the evidence of a number of studies which have shown that ICDs reduce overall mortality in certain subgroups of patients with ischemic heart disease with documented VT [<xref ref-type="bibr" rid="R5">53</xref>,<xref ref-type="bibr" rid="R6">6</xref>], or even in the absence of previously documented VT [<xref ref-type="bibr" rid="R7">7</xref>].</p><p>The development of anti-tachycardia pacing algorithms has significantly reduced the need for cardioversion [<xref ref-type="bibr" rid="R8">8</xref>] and resulted in much improved patient acceptance of such devices. Whilst ICDs are associated with significant mortality benefits in-patients with ischemic heart disease and VT, they do not actually prevent the onset of arrhythmia and remain relatively contraindicated in-patients with frequent arrhythmic episodes. In addition the implantation and function of an ICD can contribute to pathological anxiety and depression in some patients [<xref ref-type="bibr" rid="R9">9</xref>]. Furthermore, ICD therapy is relatively expensive, when considered over the expected lifetime of particularly younger patients. In-patients with ischemic heart disease RFA has developed predominantly an adjunctive role in patients with incessant or highly symptomatic, drug refractory VT.</p></sec><sec sec-type="" id="s2"><title>Defining the Substrate for ventricular tachycardia in ischemic heart disease</title><p>VT in the setting of ischemic heart disease is predominantly manifest as monomorphic VT, caused by stable reentrant circuits [<xref ref-type="bibr" rid="R10">10</xref>,<xref ref-type="bibr" rid="R11">11</xref>]. Whilst polymorphic VT and primary ventricular fibrillation are seen in patients with ischemic heart disease these more often relate to acute biochemical, ischemic or pharmacological insults; whilst ablation has been proposed for these arrhythmias [<xref ref-type="bibr" rid="R12">12</xref>], correction of the underlying abnormalities form the first line of treatment.</p><p>In simple terms the mechanism of reentry in ischemic heart disease relates to zones of heterogeneous conduction, typically at the edge of myocardial scars. Residual functioning myocytes at the edge of, or interspersed within, myocardial fibrous scar create zones of slow conduction and conduction block [<xref ref-type="bibr" rid="R13">13</xref>]. When appropriately located and with appropriate electrophysiological properties, these slow conduction zones can take the form of a discrete electrical channel or isthmus critical to the initiation and continuation of VT.</p><p>The re-entrant VT circuit utilizes the slow conduction properties of the critical isthmus. (<xref ref-type="fig" rid="F1">Figure 1</xref>) The electrical activity propagates relatively slowly along the isthmus during diastole and forms a silent zone on the surface ECG. When electrical activity reaches the exit zone it rapidly propagates throughout the ventricular myocardium, giving rise to mechanical systole and the QRS complex on the surface ECG. The electrical wavefront courses a loop through the myocardium before returning to the entrance zone and repeating the isthmus passage [<xref ref-type="bibr" rid="R14">14</xref>].</p><p>The diagrammatic representation of re-entry in <xref ref-type="fig" rid="F1">Figure 1</xref> is a simplistic description of a process, which can be highly complex. The reentry circuits can be multiple with shared and separate components; the circuits can be located in the endocardium, epicardium, transmuraly or throughout each of these zones. The isthmuses can vary significantly in length and electrophysiological properties. The regions of conduction block can be formed by scar tissue or by anatomical boundaries (typically mitral annulus) and can manifest absolute conduction block or functional conduction block developing only during VT. A comprehensive knowledge of the anatomy and electrophysiology is fundamental to approaching ablation of VT in ischemic heart disease.</p></sec><sec sec-type="" id="s3"><title>Mapping techniques</title><p>Electrophysiological mapping techniques are primarily focused to identify appropriate sites for ablation. In the mapping of VT in the setting of ischemic heart disease, the fundamental task is to identify and understand the critical VT isthmus. A number of different techniques have been developed to guide mapping.</p></sec><sec sec-type="" id="s4"><title>Mapping During VT</title><sec id="s4a"><title>Early systolic Activity</title><p>Traditional mapping approaches during VT have focused on identification of earliest systolic activation. The principle is that the electrical activation that precedes the QRS complex represents the breakout point of the VT into relatively normal myocardium. Fast moving discrete activity preceding the surface QRS by at least 30 msec is regarded as early activation. As seen in <xref ref-type="fig" rid="F2">Figure 2</xref>, the activity preceding the surface ECG will be recorded at the exit zone. Whilst identification of the exit zones is important, ablation here alone is rarely sufficient to eliminate the re-entry circuit.</p></sec><sec id="s4b"><title>Diastolic potentials</title><p>Slow moving activity during diastole can be identified in the critical isthmus during VT [<xref ref-type="bibr" rid="R11">11</xref>,<xref ref-type="bibr" rid="R15">15</xref>]. The timing of this activity will become progressively earlier as the mapping moves from the exit to the entrance. Whilst this is often a useful finding, alone it is inconsistent; blind loops and far field potentials can give rise to diastolic activity that is unrelated to the critical isthmus. The diastolic potentials which arise in the critical isthmus will remain associated with VT during entrainment mapping (described below), whilst diastolic potentials arising from bystander areas should be dissociated from the VT during entrainment [<xref ref-type="bibr" rid="R15">15</xref>]. The morphology of the diastolic activity has also been studied to predict the location of the critical isthmus but has limited additional value.</p></sec><sec id="s4c"><title>Entrainment Mapping</title><p>Entrainment refers to a specific response of an arrhythmia to a pacing stimulus at rates greater than the tachycardia cycle length [<xref ref-type="bibr" rid="R16">16</xref>]. Whilst no one technique alone can fully identify the re-entrant circuit in is entirety, entrainment mapping is perhaps of most use in localizing the critical VT isthmus [<xref ref-type="bibr" rid="R11">11</xref>] [<xref ref-type="bibr" rid="R17">17</xref>]. The principles of entrainment utilize the unique electrophysiological properties, which lead to the establishment of a reentrant arrhythmia with an excitable gap. Whilst a complete discussion of entrainment is beyond the scope of this paper, the principles warrant discussion.</p><p>Pacing that leads to classical entrainment will result in temporary QRS fusion. If pacing is conducted from within the critical isthmus of the arrhythmia, the resultant QRS morphology should match that of the VT being studied; this is known as entrainment with concealed fusion or concealed entrainment. During concealed entrainment, the duration from the pacing stimulus to the surface QRS onset can localize the pacing site within the isthmus. The shortest S-QRS will be recorded at the exit zone and this will progressively lengthen as the site of pacing moves along the isthmus towards the entrance zone.</p><p>When pacing at a site within the re-entry circuit, the post pacing interval, from the last pacing stimuli to the return electrogram seen at the pacing site should be similar to the cycle length of the VT. (Post pacing interval - VT cycle length less than 30 msec.) With the complexity of the reentry circuits, neither concealed entrainment nor the post-pacing interval are completely accurate in identifying the isthmus zone [<xref ref-type="bibr" rid="R18">18</xref>]. The utilization of entrainment mapping is further limited by a number of factors including: the dependence upon haemodynamically tolerated rhythms to enable mapping, the ability to capture the local myocardium, which can be difficult in zones of scar with high capture threshold and the often complex and fractionated signals within the scar border zone which complicate interpretation.</p></sec></sec><sec sec-type="" id="s5"><title>Mapping During Sinus Rhythm</title><p>Mapping during VT can be limited by haemodynamic instability, difficulties in inducing VT and uncertainties regarding the clinical significance of induced VTs. Mapping during sinus rhythm is designed to delineate the anatomical and electrophysiological substrate, which predispose to reentry and VT. Mapping in the area of previous infarction can identify these potential areas of abnormal conduction.</p><p>Different groups perform substrate mapping using a combination of different techniques. Analysis of electrograms during sinus rhythm can identify regions of fractionated electrograms and late potentials, which are markers of abnormal, slowed conduction. Voltage mapping identifies scar zones and the scar border zones in which most isthmuses are located. Pacing threshold maps add additional information to the voltage map to identify electrically heterogeneous areas. Pacemapping to produce a QRS morphology that is similar to that during VT can roughly localize the site of the reentry circuit exit. The duration from the pacing stimulus to the onset of QRS (S-QRS duration) can further identify abnormal substrate that predisposes to reentry.</p><p>None of these above techniques have adequate specificity to be used alone, but in combination can successfully identify regions of abnormal conduction. Substrate mapping does not identify a discrete point for ablation, but rather a zone or region of abnormal myocardium with the potential to facilitate reentry.</p></sec><sec sec-type="" id="s6"><title>Newer Mapping Tools</title><p>The newer mapping tools utilized in electrophysiology all have the potential to facilitate the mapping and ablation of VT. By providing reproducible anatomical information the mapping tools each have the ability to simplify mapping approaches and to guide the delivery of optimal radiofrequency ablation lesions.</p><p>CARTO (Biosense-Webster, Diamond Bar, USA) is the most studied mapping tool to aid ablation of ventricular tachycardia. CARTO is most useful in providing accurate substrate maps. The EnSite 3000 multielectrode array (Endocardial Solutions, St Pauls, USA) uses non-contact virtual electrograms which can not only provide myocardial substrate information, but can map ventricular activation with a little as one recorded cardiac cycle.</p><p>As technology improves, the mapping systems will become an integral part of the study of ventricular tachycardia. In the foreseeable future, however, the mapping tools will provide guidance and accuracy, but will not replace the need for an operator with a full grasp of electrophysiology and anatomy.</p></sec><sec sec-type="" id="s7"><title>Ablation Techniques</title><p>The ablation of VT in the setting of ischemic heart disease has been performed according to two basic strategies. The first strategy involves induction of ventricular tachycardia, mapping and targeted ablation [<xref ref-type="bibr" rid="R19">19</xref>-<xref ref-type="bibr" rid="R27">27</xref>]. The second strategy involves substrate mapping in sinus rhythm and linear ablation to eradicate regions of abnormal myocardium with the potential to facilitate reentry.28 It is difficult to comment on the relative success of the different approaches as most centers now use a combination of these two approaches.</p><p>Targeted ablation of a single morphology of haemodynamically stable VT can be successful in over 80% of cases [<xref ref-type="bibr" rid="R19">19</xref>-<xref ref-type="bibr" rid="R27">27</xref>]. Unfortunately, acute procedural success does not always predict long term recurrence of VT. Some studies have reported a recurrence of VT of close to 50% if a single dominant morphology is successfully ablated [<xref ref-type="bibr" rid="R25">25</xref>]. As such the trend in ablation of VT has been to target all inducible morphologies of monomorphic sustained VT. When multiple morphologies of VT are targeted the acute success rate diminishes proportional to the number of different morphologies and diminishes further when the VT is hemodynamically unstable [<xref ref-type="bibr" rid="R26">26</xref>,<xref ref-type="bibr" rid="R29">29</xref>]. However, even in-patients with more than 3 inducible morphologies of VT or with hemodynamically unstable VT procedural success rates are approach 60% [<xref ref-type="bibr" rid="R26">26</xref>].</p><p>The linear approach pioneered by Marchlinski [<xref ref-type="bibr" rid="R28">28</xref>] was acutely successful in eradicating or modifying VT in 7 of 9 patients with a recurrence in less than 20%. Using a combination approach large series report an acute success rate of 60-80%, with a longer-term freedom from VT of 75-95% in those successfully ablated. Recent developments in catheter technology have led to the use of irrigated tip catheters, which enable deeper lesions to be formed. These catheters have been shown to be more successful in the acute termination of VT [<xref ref-type="bibr" rid="R30">30</xref>], long term data is not yet available. Alternative approaches via the epicardium [<xref ref-type="bibr" rid="R31">31</xref>] and via the coronary arteries [<xref ref-type="bibr" rid="R32">32</xref>] have also been used but as yet are not applicable beyond a few expert centers.</p></sec><sec sec-type="" id="s8"><title>Complications</title><p>In the larger reported series, significant complications (cardiac tamponade, TIA, complete heart block) occurred in 5 - 10% of patients and minor complications (access site hematoma, nausea etc) in a further 5 - 10% of patients [<xref ref-type="bibr" rid="R19">19</xref>-<xref ref-type="bibr" rid="R29">29</xref>].</p></sec><sec sec-type="" id="s9"><title>Defining success</title><p>Previous series have reported the recurrence of VT following a successful procedure to vary from15 - 40% [<xref ref-type="bibr" rid="R19">19</xref>-<xref ref-type="bibr" rid="R25">25</xref>], depending upon the specific definitions used. The non-inducibility of clinical VT is often regarded as a marker of success following RFA. However, it has been shown that even following successful ablation of all clinical VTs, the presence of inducible non-clinical VT is associated with an increased incidence of recurrent VT [<xref ref-type="bibr" rid="R21">21</xref>]. Indeed, following ablation of all morphologies of clinical VT in one series, further morphologies, which had not been documented clinically, were inducible in 83% of patients [<xref ref-type="bibr" rid="R26">26</xref>].</p><p>Because of the difficulties with recognizing and defining clinical VT, we prefer to define success based on the level of stimulation required to induce VT rather than the particular morphology of VT induced [<xref ref-type="bibr" rid="R33">33</xref>]. Prior to ablation, we conduct a stimulation protocol from the right ventricular apex comprising an 8 beat drive followed by up to 5 extrastimuli introduced at decreasing cycle lengths until refractoriness was reached. At the end of the ablation procedure we repeat the stimulation protocol. Procedural success was defined by comparing the results of the two stimulation protocols.</p><p><list list-type="alpha-lower"><list-item><p><italic>Acute success</italic> - No monomorphic VT induced at refractoriness of 5 extra stimuli.</p></list-item><list-item><p><italic>Modified Result</italic> - Monomorphic VT only inducible by 2 extra stimuli more aggressive than at baseline.</p></list-item><list-item><p><italic>Failed Procedure</italic> - Monomorphic VT inducible at similar levels to baseline.</p></list-item></list></p><p>Using this protocol we were able to accurately predict the patients who will remain free of further VT following radiofrequency ablation. In patients with a minimum follow up of 12 months the rate of recurrent VT with either a complete success or a modified result was 4%, compared with 66% amongst patients with a failed procedure [<xref ref-type="bibr" rid="R33">33</xref>].</p></sec><sec sec-type="" id="s10"><title>Patient selection</title><p>At one end of the spectrum, the patient with incessant or intractable ventricular tachycardia, resistant to pharmacotherapy and overdrive pacing and requiring frequent cardioversion is an obvious candidate for ablation of ventricular tachycardia. At the other end of the spectrum there is presently no role for ablation in the sense of primary prevention.</p><p>Patients with frequent episodes of VT should be considered for ablation. The threshold for deciding to undergo ablation should consider the local expertise, the frequency of episodes, patient and economic factors. Comprehensive data has shown a substantial mortality benefit supporting the use of ICDs in-patients with post infarction VT. At present ablation should ideally be considered to have a role as an adjunct to an ICD in-patients with highly symptomatic VT.</p><p>As mapping and ablation techniques develop and the accuracy of predicting the long-term outcome improves, the role of RFA may expand, making it the therapy of first choice in a growing proportion of patients. Perhaps with time ICDs may be reserved as the treatment for failed ablation procedures or for prophylactic indications, particularly in parts of the world where widespread implantation of ICDs is financially impractical. Even if ICDs remain justified on the basis of residual uncertainty, the quality of life is likely to be improved by the reduction in discharges resulting from catheter ablation.</p></sec><sec sec-type="" id="s11"><title>Conclusions</title><p>Radiofrequency catheter ablation of patients with highly symptomatic, sustained, monomorphic post infarction VT can be performed with high success rate and acceptable procedural complication rate. The procedure can be successfully applied to a wide spectrum of patients including those with multiple morphologies of VT and hemodynamically unstable VT. The procedure at present requires extensive understanding of anatomical and electrophysiological principles and is prolonged and technically demanding. Its application beyond a few expert centers is dependent upon electrophysiological and technological advances.</p></sec> |
The role of Psychological Factors in the Aetiology and Treatment of Vasovagal Syncope | Could not extract abstract | <contrib contrib-type="author"><name><surname>Gracie</surname><given-names>Jennifer</given-names></name><xref ref-type="aff" rid="aff1">*</xref><xref ref-type="aff" rid="aff2">¶</xref></contrib><contrib contrib-type="author"><name><surname>Baker</surname><given-names>Christine</given-names></name><xref ref-type="aff" rid="aff1">*</xref></contrib><contrib contrib-type="author"><name><surname>Freeston</surname><given-names>Mark H</given-names></name><xref ref-type="aff" rid="aff2">¶</xref></contrib><contrib contrib-type="author"><name><surname>Newton</surname><given-names>Julia L</given-names></name><xref ref-type="aff" rid="aff3">§</xref></contrib> | Indian Pacing and Electrophysiology Journal | <sec sec-type="" id="s1"><title>Introduction</title><p>Syncope is a sudden transient loss of consciousness with loss of postural tone, followed by spontaneous recovery [<xref ref-type="bibr" rid="R1">1</xref>]. Around 30 percent of the general population have one syncopal event in their lifetime, with 3% having recurrent episodes [<xref ref-type="bibr" rid="R2">2</xref>]. Vasovagal syncope (VVS) is an exaggerated tendency towards the common faint that accounts for up to 29% of syncope3 and affects all age groups. VVS is characterised by profound hypotension with or without bradycardia. Those with VVS are at risk of injury during episodes and the long term implications of recurrent episodes of hypotension are unclear [<xref ref-type="bibr" rid="R1">1</xref>]. The underlying pathophysiology of VVS is uncertain and current treatments involve salt and fluid replacement and maintenance of blood pressure using mineralocorticoids or alpha agonists [<xref ref-type="bibr" rid="R1">1</xref>]. These treatments are largely symptomatic and may be associated with side effects that make their use in younger age groups inappropriate [<xref ref-type="bibr" rid="R4">4</xref>].</p><p>VVS can generally be differentiated from other causes of syncope through clinical evaluation. A definitive diagnosis of VVS is made by Head-up Tilt Table (HUT) test with the diagnosis of VVS being confirmed when a patient develops hypotension and/or bradycardia in association with syncopal or presyncopal symptoms [<xref ref-type="bibr" rid="R5">5</xref>].</p><p>One of the main methodological threats to current research into VVS is the appropriate measurement of VVS. A HUT is the most sensitive tool to diagnose VVS [<xref ref-type="bibr" rid="R5">5</xref>]. The majority of studies however have recruited participants diagnosed by medical evaluation and symptom pattern recognition only [<xref ref-type="bibr" rid="R6">6</xref>-<xref ref-type="bibr" rid="R10">10</xref>]. Although this method of diagnosis has some degree of reliability and validity, it is not as reliable as a diagnosis made by HUT.</p><p>There can be a wide range of reported illness and disability associated with the same level of disease [<xref ref-type="bibr" rid="R11">11</xref>]. The experience of VVS is at best inconvenient, and may be perceived as threatening and disabling.</p><p>In this review we will consider the evidence that psychological factors play a role in the development and maintenance of the symptoms of syncope, and more specifically vasovagal syncope, and whether psychological interventions might impact upon the manifestation and consequences of these symptoms.</p><p>Research into the psychological factors in syncope has mainly involved groups of participants who have syncope of mixed origin. There are very few studies that have focussed exclusively on VVS as a separate condition. Further, it is also important to appreciate that there are no measures specific to VVS and that in many studies the diagnosis of VVS is by clinical evaluation rather than by a positive diagnostic test, raising the question that all that is syncope may not be vasovagal in origin.</p></sec><sec sec-type="" id="s2"><title>Psychological factors associated with syncope</title><sec id="s2a"><title>1. Impact of syncope on reported quality of life</title><p>Patients with frequent syncope have been shown to have a markedly reduced quality of life, similar to that of patients with severe rheumatoid arthritis or chronic low back pain [<xref ref-type="bibr" rid="R8">8</xref>,<xref ref-type="bibr" rid="R12">12</xref>]. This provides some indication of the significantly disabling nature of this condition.</p><p>Quality of life appears to decrease with increasing frequency of syncope, and there is evidence of impairment in all reported dimensions - particularly in terms of mobility, usual activities, and self care [<xref ref-type="bibr" rid="R13">13</xref>]. This impact upon quality of life is reversible and improves within six months when the frequency of syncope is reduced, as exemplified when syncopal events are reduced on permanent pacemaker implantation [<xref ref-type="bibr" rid="R14">14</xref>].</p><p>Linzer’s group have developed a disease specific measure of impairment due to recurrent syncope (mixed aetiology) [<xref ref-type="bibr" rid="R12">12</xref>]. This tool measures psychosocial impairment due to syncope and has reported difficulties with activities of daily living (71%), driving (60%), physical activities (56%), and even walking (42%). Problems in relationships with friends, family and spouses have also been noted. Twenty one per cent of patients have a high degree of fear and worry about their condition which is not associated with injury due to syncope.</p></sec><sec id="s2b"><title>2. Psychological reactions to syncope</title><p>Patients with syncope have a high prevalence of psychological distress, especially anxiety and depression [<xref ref-type="bibr" rid="R6">6</xref>,<xref ref-type="bibr" rid="R8">8</xref>,<xref ref-type="bibr" rid="R15">15</xref>,<xref ref-type="bibr" rid="R16">16</xref>]. Individuals may fear their syncope and the negative consequences of fainting and can be severely disabled by their condition [<xref ref-type="bibr" rid="R17">17</xref>].</p><p>Insights into thoughts immediately before syncope come from a study involving US Air force active flying personnel who had experienced syncope [<xref ref-type="bibr" rid="R9">9</xref>]. Thoughts included predictions about potential bodily or psychological harm, particularly people’s reactions to their action, fear of others laughing at them, or being embarrassed due to being socially inappropriate. Virtually all subjects reported that they felt helpless in avoiding the threatening situation.</p><p>Vigilance of somatic symptoms linked with fainting, such as feeling hot and giddy, may also sustain apprehension and fear about fainting. The autonomic symptoms of syncope and anxiety are similar, so fear arousal may amplify the physiological signs associated with syncope and lead to an increase in the fear associated with syncope [<xref ref-type="bibr" rid="R18">18</xref>].</p><p> Some studies suggest that co morbid diseases, number of syncopal events, duration of syncope and injury due to syncope have no relationship to impairment due to syncope [<xref ref-type="bibr" rid="R12">12</xref>]. Psychological factors may mediate impairment associated with syncope, and this would mirror findings in other fields of chronic disease, for example chronic pain [<xref ref-type="bibr" rid="R19">19</xref>] and angina [<xref ref-type="bibr" rid="R20">20</xref>]. Studies suggest [<xref ref-type="bibr" rid="R6">6</xref>] that psychiatric illness is more likely in females, younger age groups (as would be expected in the general population), and in those with a higher number of previous syncopal events.</p><p>Despite the methodological problems that make it difficult to draw definite conclusions for these populations with VVS, there is a suggestion that the ways that people understand their condition and their experience and then make predictions about the consequences of fainting is associated with different styles of coping and adjustment, and subsequent quality of life. This raises some interesting areas of potential psychological investigation.</p></sec></sec><sec sec-type="" id="s3"><title>Psychological factors associated with vasovagal syncope</title><p>Studies specifically carried out in those with VVS are few, and in those with VVS diagnosed exclusively by HUT are rarer still. It could be argued that it is not appropriate to extrapolate data from heterogeneous groups of syncopal patients to those with VVS. Our experience would support the fact that VVS can lead to high levels of illness and impact profoundly upon the quality of life of the sufferer. For example VVS is associated with school absences in children and absence from work in adults [<xref ref-type="bibr" rid="R4">4</xref>].</p><p>Studies suggest that the measurement of psychological constructs in VVS is narrow and tend to focus on demonstrating there is psychological distress among VVS participants. The measures used however are not specific to VVS. For example, McGrady et al [<xref ref-type="bibr" rid="R17">17</xref>] used the Beck Depression Inventory (BDI) and the State Trait Anxiety Inventory (STAI). These give a level of depression and anxiety within the groups, but it does not allow for the question of cause or effect to be resolved.</p><p>One retrospective study [<xref ref-type="bibr" rid="R16">16</xref>] has gathered data from medical notes around the emotional impact of VVS and the reported stressful aspects of VVS. Over half (56%) reported a history of mood disturbance and twenty-one percent were taking a psychotropic medication. A range of psychological problems were reported including suicidal ideation, but in particular depression, panic and chronic anxiety. This is in line with levels of emotional distress reported in other groups with chronic disease. This study also found that patients often experience disabling and uncomfortable symptoms that persist for hours or days after an episode of syncope. The symptoms they reported as distressing include fatigue, dizziness and dyspnoea. Over forty percent were anxious, worried and frustrated by their situation. The thoughts reported by this group concerned the condition becoming worse, transportation problems due to driving restrictions, not being able to fulfil occupational or family roles, concern about mood swings, interpersonal relationships and the impact of the condition on spouses.</p><p>The more frequently reported personality characteristics of these individuals were of discipline, motivation, organisation, perfectionism, punctuality and sensitivity. In this group, those who struggled most to cope with their condition were those who tended to be somewhat self-stressing in their personal style.</p><p>The perceived negative consequences of fainting and assumptions of having little or no control together may result in fear and avoidance or restriction of activity leads to an impact on the individual’s quality of life. In anxiety disorder such as panic, avoidance and other self-protecting behaviours have been shown to reinforce the belief that particular situations are dangerous and prevent the development of more adaptive ways of coping with the condition [<xref ref-type="bibr" rid="R21">21</xref>].</p></sec><sec sec-type="" id="s4"><title>The potential role of psychological intervention in the treatment of VVS</title><p>Given the high prevalence of psychological dysfunction in those with syncope and VVS and the role of illness-related beliefs and predictions in this condition, psychological interventions could potentially have a part to play in the management of VVS. Indeed, case reports have described the successful use of applied tension (the use of physical manoeuvres such as muscle tensing) and cognitive behaviour therapy (CBT) in patients with VVS [<xref ref-type="bibr" rid="R22">22</xref>-<xref ref-type="bibr" rid="R26">26</xref>].</p><p>A recent case series of patients with a definite diagnosis of VVS (i.e. positive head up tilt and full reproduction of symptoms), in whom conventional treatments had not improved symptoms showed that CBT (incorporating applied tension) resulted in dramatic improvements in symptomatology, consultation behaviour and outcome. In nine patients there were significant reductions in syncopal episodes and consultation rates post CBT intervention with clear subjective improvements in quality of life and ability to return to work or school [<xref ref-type="bibr" rid="R4">4</xref>]. The CBT treatment components included identifying and restructuring unhelpful beliefs; addressing maladaptive somatic attention; reducing avoidance of certain activities and situations; the use of applied tension, and addressing idiosyncratic issues, for example, difficulties sleeping and coping with reactions of others.</p><p>Whether similar affects are seen in patients with milder symptoms of VVS is unclear. Likewise, whether psychological symptoms are present early in the natural history of the disease or whether psychological symptoms develop as the symptoms become more entrenched requires clarifying.</p><p>There is evolving evidence that the SSRI class of antidepressant is effective in the treatment of VVS and the neurotransmitter serotonin has been implicated in the aetiology of VVS. Serotonin inhibits central neurons that regulate sympathetic activity and there is an association between VVS and depression, in which serotonin is also involved [<xref ref-type="bibr" rid="R27">27</xref>]. Further work is needed in order to clarify whether the effectiveness of SSRI in VVS is due to syncope recurrence or whether it has affects upon the subjects’ mood and ability to cope with their symptoms.</p></sec><sec sec-type="" id="s5"><title>Conclusions</title><p>There are a number of psychological issues that appear to be important in the development and maintenance of disability and psychological distress associated with VVS. Current research suggests there is a role for psychological intervention in addressing the distress and disability caused by VVS. Further longitudinal research may help to develop the understanding of the experience of adapting to and coping with VVS, and provide a valuable insight into the long-term psychological consequences of having VVS. Researchers must aim to study pure groups of patients with VVS and develop and use measures that are specific to VVS. This process has started with the development of a measure of impairment due to syncope [<xref ref-type="bibr" rid="R8">8</xref>,<xref ref-type="bibr" rid="R12">12</xref>]. More work is needed to identify and match effective interventions to those who might most benefit.</p></sec> |
Automatic Capture Verification in Pacemakers (Autocapture) - Utility and Problems | Could not extract abstract | <contrib contrib-type="author"><name><surname>Kam</surname><given-names>Ruth</given-names></name><degrees>MBBS (Singapore), FRCP (Edin), M Med Int Med (Singapore) FAMS, Testamur NASPExAM</degrees></contrib> | Indian Pacing and Electrophysiology Journal | <sec sec-type="" id="s1"><title>Introduction</title><p>The concept of a closed-loop feedback system, that would automatically assess pacing threshold and self-adjust pacing output to ensure consistent myocardial capture, has many appeals. Enhancing patient safety in cases of an unexpected rise in threshold, reduced current drain, hence prolonging battery longevity and reducing the amount of physician intervention required are just some of the advantages.</p><p>Autocapture (AC) is a proprietary algorithm developed by St Jude Medical CRMD, Sylmar, CA, USA, (SJM) that was the first to commercially provide these automatic functions in a single chamber pacemaker (Microny and Regency), and subsequently in a dual chamber pacemaker (Affinity, Entity and Identity family of pacemakers).</p><p>This article reviews the conditions necessary for AC verification and performance and the problems encountered in clinical practice.</p></sec><sec sec-type="" id="s2"><title>What is Autocapture?</title><p>The AC algorithm comprises four fully automatic pacemaker functions:
<list list-type="order"><list-item><p>Capture confirmation</p></list-item><list-item><p>Back-up high voltage pulse in case of loss of capture</p></list-item><list-item><p>Threshold search and documentation</p></list-item><list-item><p>Output regulation</p></list-item></list></p><sec id="s2a"><title>Capture Confirmation</title><p>This is achieved by the ability to detect the presence or absence of an evoked response (ER) potential by the pacemaker circuitry. There exists 2 sense amplifiers in AC pacemakers - a PR sense amplifier to sense spontaneous R waves and an ER sense amplifier to sense evoked response. After the delivery of a pacing pulse, the ER sense amplifier is blanked for 14 ms (so as to disregard the residual polarization effects of the pacing pulse) and then open from 15 to 62.5 ms in order to detect the ER. Accurate discrimination between the polarization signal (PS), which is the result of the electrical charge that remains on the lead tip after the pacing pulse has been delivered and the ER, which is produced by local myocardial capture, is crucial for its determination. A high lead polarization signal can produce so much "background noise" relative to the ER signal that the pacemaker is unable to distinguish capture from non-capture.</p></sec><sec id="s2b"><title>Back-up pulse after loss of capture</title><p>Patient safety is enhanced by delivery of a 4.5 V/0.49 ms back-up pulse when no ER signal is detected within the 15 - 62.5 ms detection window after the pacing pulse.</p></sec><sec id="s2c"><title>Threshold search</title><p>An automatic threshold search is initiated under one of three conditions:
<list list-type="alpha-lower"><list-item><p>Whenever the pacemaker detects loss of capture in two consecutive beats, which it interprets as a rise in prevailing threshold.</p></list-item><list-item><p>Every eight hours in the absence of (a).</p></list-item><list-item><p>Manually using a programmer or a magnet.</p></list-item></list></p><p>In the Microny and Regency, the pacing output is reduced by 0.3V till two consecutive losses of capture and back-up pulses are emitted. The output is then increased by 0.3V until there are two consecutive capture confirmations. This is taken as the capture threshold.</p><p>For the dual chamber pacemakers like Affinity and Identity, during a threshold search, the AV interval is first shortened to ensure there is no fusion with intrinsic conduction (25 ms for sensed AV and 50 ms for paced AV delay).The pacing output is reduced stepwise by 0.25 V until loss of capture and a back-up pulse is emitted. After two consecutive losses of capture at the same output, the output is increased by 0.125 volts until there are two consecutive capture confirmations at the same output. This value is then taken as the stimulation threshold. The pacemaker also documents each threshold result in the form of a graph over time.</p></sec><sec id="s2d"><title>Output regulation</title><p>After determination of the threshold, the pacemaker adjusts its output to stimulate at 0.3V (for single chamber) and 0.25 V (for dual chamber) above the prevailing threshold. It is important to emphasise that capture confirmation is occurring on a beat-to beat basis and back-up pulse, threshold determination and output regulation then follow accordingly.</p></sec></sec><sec sec-type="" id="s3"><title>What can impact proper operation of the AC function?</title><p>These variables may be hardware-, implant-, patient-, or programming-specific.</p><sec id="s3a"><title>Hardware- and implant-related factors</title><p>The use of a bipolar lead is mandatory for AC function. The AC pacemaker paces in unipolar mode, from tip to case,(to increase the signal to noise ratio) and senses in bipolar mode, from tip to ring, (to minimize interference from muscle activity).</p><p>Choice of a pacing electrode with low lead polarization characteristics is essential. The most important factor influencing polarization appears to be the type of coating at the lead tip rather than the tip electrode size or the type of fixation. Titanium nitride (TiN) coated electrodes have the lowest and platinum helix (PH) electrodes give the highest polarization signals [<xref ref-type="bibr" rid="R1">1</xref>]. As a result TiN electrodes had a higher functional rate of AC (91.7%) versus PH (0%).</p><p>Intraoperative measurement of the ER signal during lead positioning has been recommended by SJM in order to achieve the biggest ER signal relative to the PS so that at the recommended ER sensitivity setting, the PS comprises less than 60% of it. Unfortunately these measurements cannot be made with the standard pacing system analyzer (PSA). The lead has to be connected to the pulse generator and the system has to be interrogated with a sterile wand connected to the APS II or APS μ-programmer. The ER signal must be at least 4.5 mV and the polarization signal must not be more than 1.5 mV. There is also no correlation between the spontaneous R wave as measured with the standard PSA and the ER signal [<xref ref-type="bibr" rid="R2">2</xref>,<xref ref-type="bibr" rid="R3">3</xref>] . If these intraoperative measurements are not made and the final lead position is decided by the usual criteria, approximately 5 - 7% of patients were found to have inadequate ER signals and AC could not be activated [<xref ref-type="bibr" rid="R3">3</xref>,<xref ref-type="bibr" rid="R4">4</xref>].</p></sec><sec id="s3b"><title>Patient-related factors</title><p>For AC to be activated, the pacemaker has to be able to initiate a threshold search. For single chamber pacemakers, this is done by temporarily increasing the pacing rate to 100 bpm or incrementally by 10 bpm up to a maximum of 120 bpm if the patient’s intrinsic rhythm inhibits the device. If the patient’s heart rate is above 120 bpm, e.g. atrial fibrillation with rapid ventricular response, the threshold test is postponed until the rate slows to the point when pacing is established. If the rate is persistently high, so that an automatic threshold search cannot be initiated, AC may have to be programmed OFF [<xref ref-type="bibr" rid="R4">4</xref>]. One solution to this would be to slow the rate with drugs. More importantly, good rate control would benefit the patient symptomatically in terms of palpitations and heart failure, besides ensuring AC function. For dual chamber pacemakers, this is usually not a problem as the AV delay is temporarily shortened during a threshold search, so as to avoid fusion beats.</p><p>Besides true loss of capture, the AC pacemaker may misinterpret certain conditions as loss of capture. This could be due to too low an ER signal (or inappropriate ER sensitivity setting relative to it), micro-dislodgement of the lead, R wave undersensing or pseudofusion beats.</p><p> If the ER signal is too small, repositioning of the lead can easily correct the problem, provided this is known at the time of implant. If the ER signal is not routinely measured during implant, and if increasing the ER sensitivity cannot compensate for the small ER signal, then AC function cannot be enabled subsequently.</p><p>Depolarisation of the heart in fusion beats result partly from intrinsic conduction and partly from the pacing pulse, so that an evoked response may sometimes be seen. In pseudofusion beats, the heart’s intrinsic conduction system depolarizes the myocardium entirely, so that when the pacing pulse arrives, the tissue is refractory and no ER signal is detected, causing delivery of a back-up pulse. After two consecutive pseudofusion beats, a threshold search is initiated. When capture is confirmed, the output is then increased to 0.25 V above the threshold. However if the output climbs to more than 4.2 volts, the pacemaker will stimulate at high output mode (HOM) or 4.5 V at 0.49 ms. It will then stay there until the next time a threshold search is initiated. The presence of frequent pseudofusion beats will result in many threshold searches being initiated. Some will give normal threshold values if pseudofusion beats were not present during the search and some will end up in HOM if capture losses were due to pseudofusion beats. This may not necessarily lead to excessive energy wastage if the pacemaker is predominantly inhibited, but it can cause confusion during follow-up especially if there are frequent oscillations of threshold due to pseudofusion beats.</p><p>A fusion avoidance algorithm, which is inherent in the AC pacemakers, resets the timing circuit from the back-up pulse to slow the pacing rate and allow intrinsic beats to come through. For dual chamber versions, this fusion avoidance algorithm consists of extending the AV delay in the next beat following a back-up pulse, by as much as 100 ms. An understanding of these algorithms is essential when interpreting Holter recordings of patients with these types of pacemakers, as these are often misinterpreted as pacemaker malfunction.</p></sec><sec id="s3c"><title>Programming related factors</title><p>R wave undersensing, similarly, may cause the pacemaker to stimulate in the refractory period of the ventricle and fail to capture, thus leading to many threshold searches and possibly HOM. This problem can be easily corrected by increasing the PR sensitivity appropriately.</p><p>In the case of pseudofusion beats causing mistaken "loss of capture", programming a hysteresis rate at least 10 beats per minute below the lower rate will prevent competition with the patient’s intrinsic rhythm and the presence of pseudofusion beats. In the case of a dual chamber pacemaker with intrinsically conducted beats, the "Autointrinsic Conduction Search" option can be turned on to automatically prolong the AV interval when the pacemaker detects an intrinsically conducted beat and thus avoid pseudofusion.</p><p>However, interactions between these algorithms and the inherent fusion avoidance algorithms occasionally lead to complex pacing behaviour that make ECG interpretation difficult for the clinician [<xref ref-type="bibr" rid="R5">5</xref>].</p></sec></sec><sec sec-type="" id="s4"><title>Clinical Experience</title><p>Extensive clinical experience has been obtained with the AC algorithm in several multicentre studies. In both the European [<xref ref-type="bibr" rid="R2">2</xref>] as well as the North American [<xref ref-type="bibr" rid="R6">6</xref>] studies, which followed patients up to one year, the algorithm was found to be safe and effective and did not result in exit block or other adverse events. If anything, the algorithm tended to err on the side of safety, providing back-up pulses in all cases where there was presumed loss of capture.</p><p>A case report appeared in 2001, concerning a patient, in whom failure of emission of back-up pulse after loss of capture, resulted in 8 seconds of ventricular standstill [<xref ref-type="bibr" rid="R7">7</xref>]. This was later discovered to be due to a transient threshold rise in the patient, which coincided with an automatic function in the pacemaker, which measured battery current drain every 11.25 minutes for 10 seconds, during which time the back-up pulse was not available. This was finally rectified in that patient, using a software correction to reduce the frequency of that measurement to once every 24 hours and provide an automatic increase in output by 1V during the measurement. While this may be an isolated report in the literature, it emphasizes the need for close post-marketing surveillance of any new technology or features in any pacemaker or defibrillator, before we embrace it fully or relax on our previous follow-up practices.</p></sec><sec sec-type="" id="s5"><title>Future Directions</title><sec id="s5a"><title>Use with other leads without low polarization characteristics</title><p>Modification of the pacing pulse has been shown in one experimental study to improve the PS, such that pacing leads without low polarization characteristics could have a better ER to PS ratio and hence, AC function could be enabled in a higher proportion of these leads [<xref ref-type="bibr" rid="R8">8</xref>]. The pacing pulse conventionally used is a biphasic pulse with an initial negative deflection, followed by a rapid discharge in the positive direction to counteract the effect of the polarization charge. Increasing the duration of the fast positive second phase from 6 ms to 10ms significantly reduced the polarization voltage in all leads but especially in leads with high polarization characteristics. This has favourable implications for pacemaker replacements so that AC function can be used with pre-existing leads from other manufacturers.</p></sec><sec id="s5b"><title>Use with epicardial leads</title><p>In paediatric patients, epicardial pacing electrodes are frequently necessary because of their small size and the need to preserve their veins for future use when they reach a suitable size for transvenous implants. The higher pacing threshold, when compared to endocardial leads, is of concern, because of higher current drain and hence more frequent battery replacements. The use of steroid-eluting bipolar epicardial leads from Medtronic with the Regency SR+ pacemaker was evaluated and found to be feasible in 12 of 14 children (86%)s [<xref ref-type="bibr" rid="R9">9</xref>,<xref ref-type="bibr" rid="R10">10</xref>].</p></sec><sec id="s5c"><title>Use with implantable defibrillators</title><p>An acute study [<xref ref-type="bibr" rid="R11">11</xref>], looking at the feasibility of using the AC algorithm in a group of patients undergoing ICD implantation, reported that evoked response and polarization signal amplitudes using standard, single coil, true bipolar ICD leads were sufficient to allow the AC algorithm to function in 20 of 21 patients (95.2%). However, with dual coil ICD leads and integrated bipolar sensing, AC could only be functional in 2 out of 9 patients (22.2%). Unfortunately, since the present AC algorithm requires unipolar pacing, this is not recommended for use with ICD systems in case it interferes with sensing of ventricular fibrillation.</p></sec><sec id="s5d"><title>Other threshold tracking algorithms</title><p>Although autocapture is the first algorithm to provide automatic threshold tracking and output regulation, other pacemaker manufacturers have also come up with competing algorithms. These are Capture Control in Biotronik, Ventricular Capture Management in Medtronic and Automatic Capture in Guidant pacemakers. There are important differences between these algorithms.</p><p>With Biotronik, ER detection also requires a bipolar lead with low polarization characteristics. There is no back-up safety pulse during loss of capture and output is increased in 2V steps if there is persistent loss of capture. After a programmable length of time, the output is decreased to its original value to test if capture can be obtained with a lower output. This kind of algorithm ensures safety in unexpected threshold rise but will not improve battery longevity.</p><p>Other manufacturers (Medtronic and Guidant) have devised alternative methods of capture verification, which can operate with both unipolar and bipolar leads without low polarization characteristics. These rely on modifications to the sense amplifiers to discriminate between the ER and PS, but clinical experience with these algorithms is far less extensive and has not been systematically reported to the same extent as the original AC algorithm.</p></sec></sec><sec sec-type="" id="s6"><title>Atrial autocapture</title><p>Currently, the AC algorithm is only available in the ventricle, but there is considerable interest in applying it to the atrium as well. There are potentially more difficulties with AC verification in the atrium because the atrial evoked response is considerably smaller than the ventricle and differentiation between the PS and ER signal is significantly more problematic. Early attempts were made by Curtis [<xref ref-type="bibr" rid="R12">12</xref>], using a triphasic pulse to minimize polarization signals. However the pacing circuitry itself would consume so much energy as to negate any potential savings from reduction of the pacing output. One practical approach may be to use the atrial evoked response integral ("area under the curve") and using bipolar sensing (tip to ring) and unipolar pacing [<xref ref-type="bibr" rid="R13">13</xref>]. Another approach, using different pairs of electrodes for pacing and sensing (e.g. unipolar pacing from atrial tip electrode and bipolar sensing from the atrial ring electrode to the can or an indifferent electrode) has been described [<xref ref-type="bibr" rid="R14">14</xref>] but the disadvantage is that detection of the atrial ER may be so delayed,due to the time needed for the impulse to propagate to the sensing electrode, as to preclude timely delivery of a backup pulse in the event of loss of capture.</p></sec><sec sec-type="" id="s7"><title>Conclusion</title><p>Autocapture has been shown to be safe and effective. However, to reap the full potential of all its benefits, one must understand its limitations and pitfalls and the ways to overcome these. "To implant and forget" is a dream not to be realized, even with this advanced technology.
</p></sec> |
Fetal Tachyarrhythmia - Part I: Diagnosis | Could not extract abstract | <contrib contrib-type="author"><name><surname>Oudijk</surname><given-names>Martijn A</given-names></name><degrees>MD, PhD</degrees><xref ref-type="aff" rid="aff1">*</xref></contrib><contrib contrib-type="author"><name><surname>Visser</surname><given-names>Gerard HA</given-names></name><degrees>MD, PhD</degrees><xref ref-type="aff" rid="aff1">*</xref></contrib><contrib contrib-type="author"><name><surname>Meijboom</surname><given-names>Erik J</given-names></name><degrees>MD, PhD</degrees><xref ref-type="aff" rid="aff2">†</xref></contrib><aff id="aff1"><label>*</label>Department of Obstetrics and Gynecology, University Medical Center Utrecht,The Netherlands</aff><aff id="aff2"><label>†</label>Division of Pediatric Cardiology, Central Hospital University of Vaud, Lausanne, Switzerland</aff> | Indian Pacing and Electrophysiology Journal | <sec sec-type="" id="s1"><title>Introduction</title><p>Fetal tachycardia, first recognized in 1930 by Hyman et al [<xref ref-type="bibr" rid="R1">1</xref>], is a condition occurring in approximately 0.4-0.6% of all pregnancies [<xref ref-type="bibr" rid="R2">2</xref>]. A subset of these cases with more sustained periods of tachycardia is clinically relevant. The necessity of therapeutic intervention in this condition is still a matter of discussion focused on the natural history of the disease. The spectrum of opinions varies from non-intervention [<xref ref-type="bibr" rid="R3">3</xref>-<xref ref-type="bibr" rid="R2">2</xref>] based on reports of deterioration of the fetal condition ultimately ending in significant neurological morbidity based on a number of cases in which the tachycardia subsided spontaneously [<xref ref-type="bibr" rid="R6">6</xref>], to aggressive pharmacotherapeutic intervention [<xref ref-type="bibr" rid="R9">9</xref>-<xref ref-type="bibr" rid="R11">11</xref>], or fetal demise [<xref ref-type="bibr" rid="R12">12</xref>-<xref ref-type="bibr" rid="R14">14</xref>]. Prenatal treatment through indirect, maternally administered drug therapy seems to be the preference of most centers [<xref ref-type="bibr" rid="R15">15</xref>] [<xref ref-type="bibr" rid="R21">21</xref>]. This matter will be discussed further in <italic>Fetal Tachyarrhythmia, Part II, Treatment</italic>.</p><p>The choice of the specific pharmacotherapeutic agents and the chances on success of therapy depend largely on the type of tachycardia. The determination of the type of tachycardia is therefore of the utmost importance [<xref ref-type="bibr" rid="R20">20</xref>] [<xref ref-type="bibr" rid="R22">22</xref>], however the available diagnostic armamentarium is limited. The most widely used method of diagnosis of fetal tachycardia, M-mode echocardiography, provides a time related documentation of function of the various cardiac structures. The current subdivision into supraventricular tachycardia (SVT), atrial flutter (AF) and ventricular tachycardia (VT) derived from M-mode echocardiography is not sufficient enough for differentiation according to the electrophysiologic mechanism. Several attempts have been made in the last years to increase the accuracy and reliability of M-mode echocardiography by measurement of atrioventricular (AV) and ventriculoatrial (VA) intervals [<xref ref-type="bibr" rid="R23">23</xref>,<xref ref-type="bibr" rid="R24">24</xref>], and the addition of Doppler echocardiography [<xref ref-type="bibr" rid="R25">25</xref>,<xref ref-type="bibr" rid="R26">26</xref>]. New methods registering the actual electrophysiologic events are in development including noninvasive techniques such as magnetocardiography(MCG) [<xref ref-type="bibr" rid="R27">27</xref>]. Characteristics of the most common types of fetal tachycardia are described and examples of M-mode echocardiography and FMCG are presented.</p></sec><sec sec-type="" id="s2"><title>Definitions of tachycardia</title><p>The normal fetal heart rate ranges are approximately 120-160 bpm at 30 weeks and 110-150 bpm at term [<xref ref-type="bibr" rid="R28">28</xref>,<xref ref-type="bibr" rid="R29">29</xref>]. Frequencies up to 170 bpm are considered mildly abnormal, whereas overt tachycardia is usually defined as a heart rate exceeding 170 bpm [<xref ref-type="bibr" rid="R28">28</xref>] or 180 bpm [<xref ref-type="bibr" rid="R22">22</xref>]. These rhythm abnormalities of the fetus are usually noticed at routine prenatal visits. Fetal echocardiography is used to exclude structural cardiac defects and used to position the M-mode sampling line to intercept both the atrial and ventricular walls. The relationship between peak atrial and ventricular systolic excursions currently provides a division into SVT, AF and VT.</p><p>Fetal magnetocardiography, a technique recording the magnetic field generated by the electrical activity of the fetal heart [<xref ref-type="bibr" rid="R27">27</xref>,<xref ref-type="bibr" rid="R30">30</xref>], offers a more precise delineation of the fetal electrophysiology. Detection occurs non-invasively by sensors cooled by liquid helium positioned several centimeters above the maternal abdomen in a magnetically shielded room. As the maternal heart generates magnetic activity as well, a maternal ECG is recorded simultaneously and substracted from the fetal MCG. This way, an averaged one lead fetal MCG is obtained and allows for a more detailed differentiation of the type of tachycardia.</p></sec><sec sec-type="" id="s3"><title> Premature Atrial Contractions</title><p>Premature atrial contractions (PAC’s) do not qualify as a form of fetal tachycardia, and are associated with good outcome. However, in approximately 0.4 % of cases, it may progress to runs of tachycardia and even become persistent. It is therefore recommended that these patients are monitored weekly by doptone to exclude the presence of runs of tachycardia [<xref ref-type="bibr" rid="R12">12</xref>,<xref ref-type="bibr" rid="R16">16</xref>]. M-mode echocardiography will show premature atrial contractions not followed by a ventricular contraction and a subsequent ‘drop’ in heart rate (<xref ref-type="fig" rid="F1">Figure 1</xref>).</p><p> The FMCG will show a premature P-wave not followed by a QRS complex (<xref ref-type="fig" rid="F2">Figure 2</xref>).</p></sec><sec sec-type="" id="s4"><title> Atrial Flutter</title><p>AF is a condition that accounts for approximately 21 – 50 % of fetal tachycardia and may be associated with structural abnormalities [<xref ref-type="bibr" rid="R14">14</xref>,<xref ref-type="bibr" rid="R21">21</xref>,<xref ref-type="bibr" rid="R31">31</xref>]. It is defined as an atrial rate ranging from 250 up to 500 bpm with a fixed or variable AV block, as the AV-node is not able to conduct every contraction of the atrium which results in a 2:1 or 3:1 conduction to the ventricles. Rarely in some cases paroxysmal 1:1 conduction is seen. It may be either paroxysmal or incessant in nature and is reported to be associated with fetal hydrops in 7 – 43 % of cases [<xref ref-type="bibr" rid="R14">14</xref>,<xref ref-type="bibr" rid="R18">18</xref>,<xref ref-type="bibr" rid="R31">31</xref>,<xref ref-type="bibr" rid="R32">32</xref>]. The most common electrophysiologic mechanism in AF is a re-entry circuit confined to the atrium. M-mode echocardiography will show typical atrial contractions that are followed by ventricular contractions every 2 or 3 atrial contractions (<xref ref-type="fig" rid="F3">Fig. 3</xref>).</p><p> FMCG will show typical flutter waves (P-waves) followed by QRS complexes every 2-3 P-waves (<xref ref-type="fig" rid="F4.1">Figure 4.1</xref>).</p></sec><sec sec-type="" id="s5"><title>Supraventricular tachycardia</title><p>SVT is reported to account for 47 – 68 % of cases of fetal tachycardia [<xref ref-type="bibr" rid="R14">14</xref>,<xref ref-type="bibr" rid="R18">18</xref>,<xref ref-type="bibr" rid="R21">21</xref>] and is associated with a low percentage of structural abnormalities in 2 % of cases. It is defined by a 1:1 atrioventricular conduction in which the atrial contraction precedes the ventricular contraction. Heart rates in SVT most commonly range from 200-300 bpm, is either paroxysmal or incessant in nature and associated with fetal hydrops in 36 - 64 % [<xref ref-type="bibr" rid="R14">14</xref>,<xref ref-type="bibr" rid="R33">33</xref>]. The condition has several different underlying electrophysiologic mechanisms, such as re-entry using an accessory atrioventricular connection (AVRT), which may be either apparent or concealed. Other possibilities include primary atrial tachycardias and re-entry within the AV node [<xref ref-type="bibr" rid="R34">34</xref>].</p><p>The diagnosis of these specific types of fetal tachycardia is difficult, if not impossible, with the most widely used current technology, M-mode echocardiography. As mentioned in the introduction, several studies have been conducted to allow for a more specified diagnosis using Doppler and M-mode echocardiography. These techniques are based on the relationship in time between atrial and ventricular wall excursions on M-mode, and flow patterns over the AV and semilunar valve orifices. In most re-entrant tachycardias, conduction through the accessory connection is fast, as in the Wolff-Parkinson-White syndrome (WPW). This results in a short RP interval on the postnatal ECG, comparable to a short VA interval on M-mode (<xref ref-type="fig" rid="F5.1">Figure 5.1</xref>). In several patients, slow conduction is observed in the accessory pathway, as in Persistent Junctional Reciprocating Tachycardia (PJRT) and Atrial Ectopic Tachycardia (AET) [<xref ref-type="bibr" rid="R26">26</xref>], resulting in a long RP interval on the ECG, comparable to a long VA interval on M-mode echocardiography (<xref ref-type="fig" rid="F5.2">Figure 5.2</xref>). In the literature there has been one case diagnosed as Junctional Ectopic Tachycardia (JET) by Doppler echocardiography [<xref ref-type="bibr" rid="R35">35</xref>]. It is extremely rare and we have not encountered JET in the prenatal situation at our center.</p><p>The detailed electrophysiological events however, are irretrievable by M-mode echocardiography. Specific types of fetal tachycardia have been detected and published using FMCG [<xref ref-type="bibr" rid="R36">36</xref>,<xref ref-type="bibr" rid="R37">37</xref>]. A more detailed configuration, comparable to the postnatal ECG is obtained. An example of this category of fetal SVT’s is shown in <xref ref-type="fig" rid="F6">Figure 6</xref>.</p><p>The electrophysiologic events are clearly defined in this situation, however, a precise diagnosis is still debatable. The SVT shown in <xref ref-type="fig" rid="F6">Figure 6</xref> for instance, may be AVRT using a concealed accessory pathway, but could also be attributed to represent a PJRT.</p></sec><sec sec-type="" id="s6"><title>Ventricular Tachycardia</title><p>VT in the prenatal situation is rare and has only been reported occasionally in the international literature [<xref ref-type="bibr" rid="R5">5</xref>,<xref ref-type="bibr" rid="R16">16</xref>]. It is usually paroxysmal in nature and outcome depends on the electrophysiologic mechanism. Of the utmost importance is the diagnosis of the congenital Long QT syndrome (LQTS). The origin of this arrhythmia lies in malfunctioning myocardial ion channels as a result of mutations in genes encoding these ion channels. It is characterised by prolongation of the QT interval and the occurrence of polymorphic ventricular arrhythmia (such as Torsade de Pointes). Patients with the LQTS are predisposed to ventricular fibrillation and sudden death [<xref ref-type="bibr" rid="R38">38</xref>,<xref ref-type="bibr" rid="R39">39</xref>]. The prenatal diagnosis is focused on fetuses of mothers with prolonged QT syndrome to document the possible presence of this syndrome in the fetus. In addition, fetuses presenting with (baseline) bradycardia or intermittent ventricular tachycardia of the torsade de pointe type, the diagnosis of LQTS must be considered as well, as spontaneous mutations have been reported. The suspicion of fetal LQTS may be raised in case M-mode echocardiography shows (baseline) bradycardia varying from 60-110 bpm and intermittent ventricular tachycardia in which there is atrioventricular dissociation. FMCG shows a prolonged QT interval (<xref ref-type="fig" rid="F7">Figure 7</xref>), and the possible presence of intermittent VT [<xref ref-type="bibr" rid="R40">40</xref>,;<xref ref-type="bibr" rid="R41">41</xref>].</p></sec><sec sec-type="" id="s7"><title>Conclusions</title><p>Fetal tachycardia is a rare disorder, in which difficulties are encountered in the diagnosis of the exact underlying electrophysiological mechanism and is therefore probably underreported. The spectrum of pathologic symptomatology varies from an infrequent paroxysmal tachycardia to a persistent form, which may deteriorate in fetal hydrops, neurological morbidity and even fetal demise. The presence of fetal tachycardia therefore deserves more attention and requires a specialized evaluation to define its electrophysiologic diagnosis.</p><p>Fetal atrial flutter is a frequently encountered form of fetal tachycardia and can be well diagnosed by either M-mode echocardiography or FMCG. Fetal SVT remains the most complicated form of fetal tachycardia, both in diagnosis as in treatment. The difficulty in delineating an accurate diagnosis of the specific form and the relative therapy resistance of the various subtypes of this tachycardia, for instance PJRT, makes the development of an optimal treatment protocol complicated.</p><p>The most infrequent encountered form of tachycardia is fetal VT, however, if present, it has very serious implications for the fetus. In utero management is complicated and yields a limited success but it is also predictive for an uncertain future after birth. Improved prenatal diagnostic techniques may support a better understanding of the pathophysiology, but does not necessarily result in an improved outcome. Studies on these new techniques will hopefully result in the ultimate goal of reduced morbidity and mortality in the field of fetal tachycardia.</p></sec> |
Idiopathic Fascicular Ventricular Tachycardia | <p>Idiopathic fascicular ventricular tachycardia is an important cardiac arrhythmia with specific electrocardiographic features and therapeutic options. It is characterized by relatively narrow QRS complex and right bundle branch block pattern. The QRS axis depends on which fascicle is involved in the re-entry. Left axis deviation is noted with left posterior fascicular tachycardia and right axis deviation with left anterior fascicular tachycardia. A left septal fascicular tachycardia with normal axis has also been described. Fascicular tachycardia is usually seen in individuals without structural heart disease. Response to verapamil is an important feature of fascicular tachycardia. Rare instances of termination with intravenous adenosine have also been noted. A presystolic or diastolic potential preceding the QRS, presumed to originate from the Purkinje fibers can be recorded during sinus rhythm and ventricular tachycardia in many patients with fascicular tachycardia. This potential (P potential) has been used as a guide to catheter ablation. Prompt recognition of fascicular tachycardia especially in the emergency department is very important. It is one of the eminently ablatable ventricular tachycardias. Primary ablation has been reported to have a higher success, lesser procedure time and fluoroscopy time.</p> | <contrib contrib-type="author"><name><surname>Francis</surname><given-names>Johnson</given-names></name><degrees>MD, DM</degrees><xref ref-type="aff" rid="aff1">*</xref></contrib><contrib contrib-type="author"><name><surname>K</surname><given-names>Venugopal</given-names></name><degrees>MD, DM</degrees><xref ref-type="aff" rid="aff1">*</xref></contrib><contrib contrib-type="author"><name><surname>A</surname><given-names>Khadar S</given-names></name><degrees>MD, DM</degrees><xref ref-type="aff" rid="aff1">*</xref></contrib><contrib contrib-type="author"><name><surname>N</surname><given-names>Sudhayakumar</given-names></name><degrees>MD, DM</degrees><xref ref-type="aff" rid="aff2">†</xref></contrib><contrib contrib-type="author"><name><surname>Gupta</surname><given-names>Anoop K</given-names></name><degrees>MD, DM, DNB, FACC</degrees><xref ref-type="aff" rid="aff3">‡</xref></contrib> | Indian Pacing and Electrophysiology Journal | <sec sec-type="" id="s1"><title>Introduction</title><p>In general ventricular tachycardias have wide QRS complexes. One of the earliest descriptions of ventricular tachycardia (VT) with a narrow QRS complex was by Cohen et al in 1972 [<xref ref-type="bibr" rid="R1">1</xref>]. Their description was a left posterior fascicular tachycardia with relatively narrow QRS. In 1979, Zipes et al [<xref ref-type="bibr" rid="R2">2</xref>] reported three patients with ventricular tachycardia characterized by QRS width of 120 to 140 ms, right bundle branch block morphology and left-axis deviation. These patients were young and had no major cardiac abnormalities. The arrhythmia could be induced by exercise, atrial and ventricular premature beats as well as atrial pacing and ventricular pacing. Belhassen et al observed that this tachycardia can be terminated by the calcium channel blocker verapamil [<xref ref-type="bibr" rid="R3">3</xref>] This observation has been confirmed subsequently by others as wells [<xref ref-type="bibr" rid="R4">4</xref>-<xref ref-type="bibr" rid="R7">7</xref>]. Belhassen et al proposed that this is a specific ECG-electrophysiological entity [<xref ref-type="bibr" rid="R8">8</xref>]. Fascicular tachycardia has also been called Idiopathic Left Ventricular Tachycardia (ILVT) by other authors, though left ventricular outflow tract VT also comes under the purview of this term [<xref ref-type="bibr" rid="R9">9</xref>,<xref ref-type="bibr" rid="R10">10</xref>]. Fascicular tachycardia is usually paroxysmal, but a case which was persistent, leading to cardiac enlargement and complete resolution following therapy with verapamil has also been reported [<xref ref-type="bibr" rid="R4">4</xref>]. Termination of idiopathic fascicular ventricular tachycardia by vagal maneuvers was noted in 4 cases by Buja et al. [<xref ref-type="bibr" rid="R11">11</xref>]. Successful radiofrequency catheter ablation was described by Klein et al. [<xref ref-type="bibr" rid="R12">12</xref>]. In this article we propose to review the current status of our knowledge regarding the genesis and treatment of idiopathic fascicular ventricular tachycardia.</p></sec><sec sec-type="" id="s2"><title>Mechanism and Classification</title><p>Zipes et al postulated that the origin of the tachycardia was localized to a small region of reentry or triggered automaticity located in the posteroinferior left ventricle, close to the posterior fascicle of the left bundle branch.2Response to verapamil suggested a role for the slow inward calcium channel in the genesis of the arrhythmia. Endocardial mapping during tachycardia revealed the earliest activation at the ventricular apex and mid septum [<xref ref-type="bibr" rid="R13">13</xref>]. The tachycardia can be entrained by ventricular and atrial pacing. Entrainment by atrial pacing suggests easy access over the conduction system into the reentry circuit and hence a role for the fascicles in the reentrant circuit [<xref ref-type="bibr" rid="R14">14</xref>]. Lau suggested the origin as reentry circuits involving the lower septum or posterior part of the left ventricle close to the endocardial surface in view of the response to radiofrequency ablation in these sites [<xref ref-type="bibr" rid="R15">15</xref>]. Purkinje potential recorded in the diastolic phase during VT at the mid-anterior left ventricular septum in rare cases with RBBB pattern and right axis deviation suggested origin near left anterior fascicle in those cases [<xref ref-type="bibr" rid="R16">16</xref>].</p><p>Recently Kuo et al has questioned the involvement of the fascicle of the left bundle branch in ILVT [<xref ref-type="bibr" rid="R17">17</xref>]. They studied two groups of patients with ILVT. One with left anterior or posterior fascicular block during sinus rhythm and the other without. They noted that the transition zone of QRS complexes in the precordial leads were similar during VT in both groups. New fascicular blocks did not appear after ablation. Therefore they concluded that the fascicle of the left bundle branch may not be involved in the anterograde limb of reentrant circuit in ILVT. </p><p>Fascicular tachycardia has been classified into three subtypes: (1) left posterior fascicular VT (<xref ref-type="fig" rid="F1">Figure 1</xref>) with a right bundle branch block (RBBB) pattern and left axis deviation (common form); (2) left anterior fascicular VT with RBBB pattern and right-axis deviation (uncommon form); and (3) upper septal fascicular VT with a narrow QRS and normal axis configuration (rare form) [<xref ref-type="bibr" rid="R18">18</xref>].</p></sec><sec sec-type="" id="s3"><title>Anatomical Substrate</title><p>Endocardial activation mapping during VT identifies the earliest site in the region of the infero-posterior left ventricular septum. This finding, along with VT morphology and short retrograde VH interval suggests a left posterior fascicular origin. Nakagawa and colleagues [<xref ref-type="bibr" rid="R19">19</xref>] recorded high-frequency potentials preceding the site of earliest ventricular activation during the VT and sinus rhythm. These potentials are thought to represent activation of Purkinje fibers and are recorded from the posterior one third of the left ventricular septum. Successful RF ablation is achieved at sites where the purkinje potential is recorded 30 to 40 ms before the VT QRS complex.</p><p>Some date suggest that the tachycardia may originate from a false tendon or fibro- muscular band that extends from the posteroinferior left ventricle to the basal septum [<xref ref-type="bibr" rid="R20">20</xref>]. Histological examination of false tendon disclosed abundant Purkinje fibers.</p></sec><sec sec-type="" id="s4"><title>Electrophysiological Study </title><p>Fascicular tachycardia can be induced by programmed atrial or ventricular stimulation in most cases. Isoprenaline infusion may be required in certain cases; rarely there may be difficulty in induction despite isoprenaline infusion. Endocardial mapping identifies the earliest activation in the posteroapical left ventricular septum in patients with posterior fascicular tachycardia.</p><p>A high frequency potential with short duration, preceding the QRS has been described as the Purkinje potential (<xref ref-type="fig" rid="F2">Figure 2</xref>). This has also been called P potential and diastolic potential. P potentials can be recorded both in sinus rhythm and during ventricular tachycardia. Pacing at sites manifesting the earliest P potential produces QRS complexes identical to that of the clinical tachycardia [<xref ref-type="bibr" rid="R19">19</xref>].</p></sec><sec sec-type="" id="s5"><title>Pharmacological Therapy</title><p>Intravenous verapamil is effective in terminating the tachycardia. However the efficacy of oral verapamil in preventing tachycardia relapse is variable. Good response and resolution of tachycardiomyopathy with verapamil treatment was noted by Toivonen et al [<xref ref-type="bibr" rid="R4">4</xref>], while Chiaranda et al commented on the poor efficacy [<xref ref-type="bibr" rid="R21">21</xref>]. Treatment with propranolol has also resulted in cure of arrhythmia and resolution of features of tachycardiomyopathy in another case with incessant fascicular VT [<xref ref-type="bibr" rid="R22">221</xref>]. Though fascicular tachycardias do not generally respond to adenosine, termination of VT originating from the left anterior fascicle by intravenous adenosine has been documented [<xref ref-type="bibr" rid="R23">23</xref>].</p></sec><sec sec-type="" id="s6"><title>Catheter Ablation</title><p>The young age of most patients with need for long-term antiarrhythmic treatment and attendant side effects prompted the search for curative therapies. Fontaine et al (1987) described successful treatment of ILVT by application of a high-energy DC shock (fulguration) between the catheter tip and a neutral plate placed under the patient's back [<xref ref-type="bibr" rid="R24">24</xref>]. Klein et al (1992) reported cure of ILVT by radiofrequency catheter ablation [<xref ref-type="bibr" rid="R25">25</xref>]. Since then radiofrequency has remained the procedure of choice.</p><p>Different approaches for radiofrequency ablation have been described by various authors. Nakagawa et al preferred careful localization of the Purkinje potential in guiding ablation. They selected the area where a Purkinje potential precedes the QRS complex during tachycardia [<xref ref-type="bibr" rid="R19">19</xref>]. Wellens et recommend pace mapping with a match between the 12 simultaneously recorded ECG leads during pacing and the clinical tachycardia for localizing the site of ablation [<xref ref-type="bibr" rid="R9">9</xref>]. They hypothesize that pathways within the Purkinje network that are not included in the reentry circuit responsible for the tachycardia may also become activated. Ablation of those regions may not result in interruption of the tachycardia circuit.</p></sec><sec sec-type="" id="s7"><title>Primary Radiofrequency Ablation</title><p>Since fascicular VT is sometimes difficult to induce despite pharmacological provocation, some workers (Gupta et al) prefer primary ablation. In a recent report, seven cases of incessant fascicular VT were successfully ablated with no recurrence [<xref ref-type="bibr" rid="R26">26</xref>]. They reported a shorter procedure time, significantly lower fluoroscopy time and lesser number of radiofrequency energy deliveries in the primary versus elective groups. The longer procedural time during elective ablation was mainly due to the time spent in induction of fascicular VT.</p></sec> |
The Automatic External Cardioverter-Defibrillator | <p>In-hospital cardiac arrest remains a major problem but new technologies allowing fully automatic external defibrillation are available. These technologies allow the concept of “external therapeutic monitoring” of lethal arrhythmias. Since early defibrillation improves outcome by decreasing morbidity and mortality, the use of this device should improve the outcome of in-hospital cardiac arrest victims. Furthermore, the use of these devices could allow safe monitoring and treatment of patients at risk of cardiac arrest who not necessarily must be in conventional monitoring units (Intensive or Coronary Care Units) saving costs with a more meaningful use of resources. The capability to provide early defibrillation within any patient-care areas should be considered as an obligation (“standard of care”) of the modern hospital.</p> | <contrib contrib-type="author"><name><surname>Martínez-Rubio</surname><given-names>Antoni</given-names></name><degrees>MD, FESC, FACC</degrees><xref ref-type="aff" rid="aff1">*</xref></contrib><contrib contrib-type="author"><name><surname>Barón-Esquivias</surname><given-names>Gonzalo</given-names></name><degrees>MD, FESC</degrees><xref ref-type="aff" rid="aff2">†</xref></contrib><aff id="aff1"><label>*</label>Departments of Cardiology . Hosp. de Sabadell (Fund. Univ. Parc Tauli), Sabadell, Barcelona, Spain</aff><aff id="aff2"><label>†</label>University Hospital Virgen del Rocio, Seville, Spain</aff> | Indian Pacing and Electrophysiology Journal | <sec sec-type="" id="s1"><title>Introduction</title><p>Cardiovascular disease is the major cause of death in the majority of countries [<xref ref-type="bibr" rid="R1">1</xref>,<xref ref-type="bibr" rid="R2">2</xref>]. In developed countries (e.g. United States), nearly the half of these deaths are unexpected and sudden [<xref ref-type="bibr" rid="R3">3</xref>,<xref ref-type="bibr" rid="R4">4</xref>]. Although the broad majority of sudden cardiac deaths occur outside the hospital and very poor survival rates have been reported, in-hospital cardiac arrest remains a major problem. The aim of this article is to review the actual knowledge of the utility of “fully automatic” external defibrillators for treatment of in-hospital cardiac arrest. The name “automated” external defibrillator (AED) must be distinguished from fully “automatic” external defibrillator. The automatic device is able to deliver a shock without intervention from an operator whereas the automated device only “advises” by a voice or a text to deliver the energy. Thus, the automated device requires operator intervention.</p></sec><sec sec-type="" id="s2"><title>Physiological considerations of cardiac arrest</title><p>Cardiopulmonary resuscitation maneuvers can sustain a patient for a certain time but is unlikely to restore an organized rhythm of the heart. To achieve this goal defibrillation and advanced cardiovascular care are necessary, which must be administered in the shortest possible interval after the arrhythmic event. Furthermore, the risk of ischemia and the defibrillation threshold increase with arrhythmia duration [<xref ref-type="bibr" rid="R5">5</xref>]. The time to defibrillation is the single most important determinant of survival from cardiac arrest [<xref ref-type="bibr" rid="R6">6</xref>-<xref ref-type="bibr" rid="R17">17</xref>]. In addition, ventricular fibrillation (VF) tends to evolve to asystole within few minutes [<xref ref-type="bibr" rid="R18">18</xref>-<xref ref-type="bibr" rid="R27">27</xref>]. Thus, survival rates after VF decrease approximately 7% to 10% with every minute that defibrillation is delayed [<xref ref-type="bibr" rid="R18">18</xref>-<xref ref-type="bibr" rid="R24">24</xref>]. In addition, the cerebral cortex is irreversibly damaged if cardiac arrest is not quickly interrupted [<xref ref-type="bibr" rid="R18">18</xref>,<xref ref-type="bibr" rid="R28">28</xref>-<xref ref-type="bibr" rid="R30">30</xref>]. Because of these facts, individual physicians as well as the American Heart Association, the American College of Emergency Physicians, the European Resuscitation Council and others have advocated the widespread dissemination of defibrillators for decreasing mortality and morbidity for victims of cardiac arrest [<xref ref-type="bibr" rid="R31">31</xref>,<xref ref-type="bibr" rid="R34">34</xref>].</p></sec><sec sec-type="" id="s3"><title>In-hospital cardiac arrest</title><sec id="s3a"><title>The classical reality</title><p>Sustained ventricular tachycardia (VT) and VF, even in hospitalized patients, are major causes of morbidity and mortality [<xref ref-type="bibr" rid="R28">28</xref>,<xref ref-type="bibr" rid="R35">35</xref>-<xref ref-type="bibr" rid="R37">37</xref>]. In the best setting (monitoring wards), continuous ECG monitoring allows identification of lethal arrhythmias and alarm systems alert nursing and medical staff. However, a time delay between the arrhythmic event and human intervention obviously exists. This time delay may be prolonged in certain circumstances even in monitoring wards. In addition, response time intervals for in-hospital resuscitation events are often inaccurate and must be corrected before documented times to defibrillation can be considered reliable [<xref ref-type="bibr" rid="R34">34</xref>].</p><p>Although major evidence exists supporting the need for rapid defibrillation and important advances in out-of-hospital cardiac arrest treatment have been achieved (e.g. out-of hospital cardiac arrest quick response programs) [<xref ref-type="bibr" rid="R15">15</xref>,<xref ref-type="bibr" rid="R16">16</xref>,<xref ref-type="bibr" rid="R34">34</xref>,<xref ref-type="bibr" rid="R38">38</xref>-<xref ref-type="bibr" rid="R43">43</xref>], in-hospital cardiac arrest is still a major problem without significant advances (e.g. changing strategies) during the last 30 years. This leads to important mortality, morbidity and social as well as economic costs [<xref ref-type="bibr" rid="R28">28</xref>,<xref ref-type="bibr" rid="R35">35</xref>-<xref ref-type="bibr" rid="R37">37</xref>]. The scarcity of data [<xref ref-type="bibr" rid="R28">28</xref>,<xref ref-type="bibr" rid="R44">44</xref>] related to deployment of AEDs in hospitals and its impact on patient outcome reflects the very limited existence of in-hospital early defibrillation programs.</p><p>However, as stated in major guidelines, early defibrillation is a high-priority goal in out-of hospital as well as in-hospital cardiac arrest [<xref ref-type="bibr" rid="R31">31</xref>-<xref ref-type="bibr" rid="R34">34</xref>]. Clearly the earlier defibrillation occurs, the better the prognosis in adults and children [<xref ref-type="bibr" rid="R18">18</xref>-<xref ref-type="bibr" rid="R21">21</xref>,<xref ref-type="bibr" rid="R23">23</xref>-<xref ref-type="bibr" rid="R25">25</xref>,<xref ref-type="bibr" rid="R27">27</xref>,<xref ref-type="bibr" rid="R45">45</xref>-<xref ref-type="bibr" rid="R47">47</xref>].</p><p>The capability to provide early defibrillation within any patient-care areas should be considered as an obligation (“standard of care”) of the modern hospital. Furthermore, cardiac arrests often occur outside monitored areas. Recently, Herlitz et al [<xref ref-type="bibr" rid="R28">28</xref> ]. reported that out of 557 patients suffering in-hospital cardiac arrest, only 292 patients (53%) were in monitored wards and from those only 43.2% of the patients could be discharged alive. They reported that the median interval between collapse and first defibrillation was 1 minute in monitored wards and 5 minutes in non-monitored wards. Only 31% of patients from non-monitored wards could be discharged alive and with a cerebral performance inferior to that of survivors of monitored wards. Other authors [<xref ref-type="bibr" rid="R48">48</xref> ] present similar data showing better in-hospital survival for witnessed arrest (25%) than for non-witnessed arrest (7%) but, in addition, they report a disproportionately high incidence of non-witnessed arrests during the night (12AM to 6 AM) resulting in a very poor survival rate (0%). Cardiac arrest occurs in 4.8% of hospitalized patients because of acute myocardial infarction [<xref ref-type="bibr" rid="R49">49</xref> ]. The survival rate to hospital discharge for these individuals was 29.4% [<xref ref-type="bibr" rid="R49">49</xref> ]. Although ventricular tachycardia or fibrillation was documented in 34.7% of patients, only 47.5% of those survived to discharge [<xref ref-type="bibr" rid="R49">49</xref> ]. Thus, the use of “automatic” defibrillators had probably improved outcome of those patients presenting in-hospital cardiac arrest. Furthermore, although survival to hospital discharge offers an objective evaluation point and is used in the broad majority of reports, several patients who survive a cardiac arrest present neurological damage, which is highly dependent of the response time to cardiac arrest [<xref ref-type="bibr" rid="R18">18</xref> ,<xref ref-type="bibr" rid="R28">28</xref>,<xref ref-type="bibr" rid="R29">29</xref> ]. Therefore, the neurological status should also be considered when reporting results of resuscitation procedures [<xref ref-type="bibr" rid="R30">30</xref> ].</p></sec><sec id="s3b"><title>The new reality</title><p>Very early response to lethal ventricular tachyarrhythmias is achievable now with fully “automatic” external defibrillators. As occurs with implantable cardioverter defibrillators, these devices may be individually programmed to intervene. Thus, as occurs with other medical management strategies (e.g. drugs prescription), physicians may and should decide under which circumstances and how the device will react in presence of arrhythmias.</p><p>In a multicenter European trial, Martinez-Rubio et al. [<xref ref-type="bibr" rid="R50">50</xref> ] tested the safety and efficacy of an automatic external cardioverter-defibrillator [Powerheart®; Cardiac Science Inc., Irvine (California, USA)]. Subjects (n=117) included patients undergoing electrophysiologic testing (n=66; all with suspected/documented VT/VF or undergoing implantation of implantable cardioverter defibrillator) as well as patients in monitoring wards (n=51) at risk of cardiac arrest. Patients with active implanted cardioverter-defibrillators were excluded from the study.</p><p>The automatic external cardioverter-defibrillator was connected to patients using self-adhesive electrodes. Placement of the electrodes (either antero-posterior or sternal-apex) was at the discretion of the investigators.</p><p>The automatic external defibrillator was used for monitoring and treatment of patients in the coronary care units or intensive care units. It was also used as a “rescue” defibrillator during ICD implantations. This device operates either in “advisory” or in “automatic” mode. In advisory mode, it will detect the arrhythmia, alarm, charge the capacitors, and prompt the user to deliver the shock. The energy (maximum of 360J monophasic) and time delay (10s to 600s) for each shock could be programmed (up to 9 shocks per episode). Shock delivery was non-committed such that therapy is automatically aborted should the rhythm spontaneously convert to a non-shockable rhythm. In addition, the device algorithm featured a modulation domain function which allows to discriminate between ventricular and supraventricular rhythms.</p><p>The performance of the device in detecting shockable and non-shockable rhythms was confirmed by review of the simultaneously recorded Holter data and the programmed parameters. A total of 1,240 hours of monitoring and 1,988 episodes of rhythm changes were documented. The device detected ventricular arrhythmias with a sensitivity of 100% and specificity of 97.6%. The mean response time to shock was 14.4 seconds [evaluated in those episodes (sustained) that were associated with a full capacitor charge cycle]. All false positives were caused by T-wave oversensing during ventricular pacing. There were no complications or adverse events. There were a total of 35 sustained arrhythmic episodes treated by the device (12 in the EP lab and 23 in monitoring wards). All 35 episodes were successfully (100% efficacy) converted to normal sinus rhythm with a first shock success of 94.3% (<xref ref-type="fig" rid="F1">Figure 1</xref>). This European study [<xref ref-type="bibr" rid="R50">50</xref> ] confirms the initial experience presented by Mattioni et al [<xref ref-type="bibr" rid="R51">51</xref> ] in a multicenter American trail. These authors reported a sensitivity and specificity of 100% and 99.4%, respectively, and a response time of 22 seconds [<xref ref-type="bibr" rid="R51">51</xref> ].</p><p>Although some differences exist (e.g. sample size, hours of monitoring, software version, etc.) both studies [<xref ref-type="bibr" rid="R50">50</xref>,<xref ref-type="bibr" rid="R51">51</xref> ] lead to the same conclusions. Thus, both groups (European and American) of investigators agree that the device is safe and highly effective in monitoring, detecting and treating spontaneous rhythms. Therefore, with its wide use a significant improvement in the treatment of in-hospital cardiac arrest should be expected. In addition, the now available version of the described automated external cardioverter-defibrillator has important improvements (compatibility to commercially available patient monitoring systems, biphasic wave forms, etc.).</p></sec></sec> |
Frequency Analysis of Atrial Fibrillation From the Surface Electrocardiogram | <p>Atrial fibrillation (AF) is the most common arrhythmia encountered in clinical practice. Neither the natural history of AF nor its response to therapy are sufficiently predictable by clinical and echocardiographic parameters.</p><p>Atrial fibrillatory frequency (or rate) can reliably be assessed from the surface electrocardiogram (ECG) using digital signal processing (filtering, subtraction of averaged QRST complexes, and power spectral analysis) and shows large inter-individual variability. This measurement correlates well with intraatrial cycle length, a parameter which appears to have primary importance in AF domestication and response to therapy. AF with a low fibrillatory rate is more likely to terminate spontaneously, and responds better to antiarrhythmic drugs or cardioversion while high rate AF is more often persistent and refractory to therapy.</p><p>In conclusion, frequency analysis of AF seems to be useful for non-invasive assessment of electrical remodeling in AF and may subsequently be helpful for guiding AF therapy.</p> | <contrib contrib-type="author"><name><surname>Husser</surname><given-names>Daniela</given-names></name><degrees>MD</degrees><xref ref-type="aff" rid="aff1">*</xref><xref ref-type="aff" rid="aff2">†</xref></contrib><contrib contrib-type="author"><name><surname>Stridh</surname><given-names>Martin</given-names></name><degrees>PhD</degrees><xref ref-type="aff" rid="aff3">‡</xref></contrib><contrib contrib-type="author"><name><surname>Sornmo</surname><given-names>Leif</given-names></name><degrees>PhD</degrees><xref ref-type="aff" rid="aff3">‡</xref></contrib><contrib contrib-type="author"><name><surname>Olsson</surname><given-names>S. Bertil</given-names></name><degrees>MD, PhD</degrees><xref ref-type="aff" rid="aff4">§</xref></contrib><contrib contrib-type="author"><name><surname>Bollmann</surname><given-names>Andreas</given-names></name><degrees>MD</degrees><xref ref-type="aff" rid="aff1">*</xref><xref ref-type="aff" rid="aff2">†</xref><xref ref-type="aff" rid="aff4">§</xref></contrib><aff id="aff1"><label>*</label> Department of Cardiology, Good Samaritan Hospital, Los Angeles, USA</aff><aff id="aff2"><label>†</label>Harbor-UCLA Medical Center, Los Angeles, USA</aff><aff id="aff3"><label>‡</label> Departments of Electroscience, Lund University, Lund, Sweden</aff><aff id="aff4"><label>§</label> Departments of Cardiology, Lund University, Lund, Sweden</aff> | Indian Pacing and Electrophysiology Journal | <p>In recent years, mechanisms leading to atrial fibrillation (AF) induction and maintenance have begun to be explored and rapidly evolving therapies including new antiarrhythmic drugs, antitachycardia pacing algorithms and catheter ablation techniques have been introduced in clinical practice. Nevertheless, outcome of AF patients is still unsatisfactory which - to the greatest extent - can be attributed to the progressing nature of this arrhythmia and high AF recurrence rates following restoration of sinus rhythm. Moreover, AF treatment (e.g. rate vs. rhythm control, choice of antiarrhythmic drugs or device therapy, curative ablation) may be viewed as being based on trial and error, since no test is able to predict the natural history of this arrhythmia or its response to treatment. Subsequently, current AF management guidelines [<xref ref-type="bibr" rid="R1">1</xref>] provide no treatment recommendations that “take the various mechanisms and patterns of AF into account”. Thus, it seems desirable to develop tests that quantify AF disease state and guide AF management [<xref ref-type="bibr" rid="R2">2</xref>].</p><p>Virtually in every patient with AF, a standard surface electrocardiogram (ECG) is acquired, the main purposes being confirmation of arrhythmia presence and determination of ventricular rate. Fibrillation waves have, moreover, various appearances ranging from fine to coarse and from disorganized to organized, flutter-like. Interestingly, the mechanisms behind the various appearances of the fibrillatory process on the ECG and the possible prognostic information contained within fibrillation waves have just very recently begun to be explored.</p><p>The aim of this article is twofold; (1) to describe novel frequency analysis techniques of the surface ECG for characterization of the fibrillatory process, and (2) to present possible applications of these methods, namely for (a) exploring autonomic modulation of AF, (b) monitoring and predicting antiarrhythmic drug effects, and (c) predicting AF recurrence following restoration of sinus rhythm. For this purpose, Medline-listed, peer-reviewed studies are included in this review.</p><sec sec-type="" id="s1"><title>Rationale for Using Atrial Fibrillatory Rate to Characterize Human AF</title><p>Shortening of atrial refractoriness and maladaptaion to rate are hallmarks of atrial electrical remodeling in AF [<xref ref-type="bibr" rid="R3">3</xref>].. During AF, re-excitation occurs during the repolarization phase of the preceding electrical wave, implying that local excitation almost always occurs without any obvious latency beyond the refractory period. Subsequently, the average atrial fibrillatory rate is likely to reflect averaged refractoriness at any part of the tissue involved. This assumption has been independently verified in both animal [<xref ref-type="bibr" rid="R4">4</xref>]. and human AF [<xref ref-type="bibr" rid="R5">5</xref>,<xref ref-type="bibr" rid="R6">6</xref>]. Hence, the length of the averaged atrial fibrillatory cycle (which is inversely related to fibrillatory frequency or rate) can be used as an index of the averaged atrial myocardial refractoriness and subsequently AF organization [<xref ref-type="bibr" rid="R7">7</xref>]. The importance of shortened refractoriness and subsequently its transformation into short atrial cycles or high rates as assessed by direct intraatrial measurements for AF progression and response to therapy has been well established. Results will be reported as cycle length in milliseconds (ms) if they were obtained from intraatrial recordings. If they were obtained from surface recordings, they will be reported as fibrillatory rate in fibrillations per minute (fpm, the rationale to use this expression is discussed in depth below).</p><p>Firstly, spontaneous arrhythmia behavior is related to baseline fibrillatory cycle length. Asano et al induced AF with rapid pacing in 30 patients undergoing electrophysiologic study [<xref ref-type="bibr" rid="R8">8</xref>]. Patients in whom AF terminated spontaneously (n=20) had an average fibrillatory cycle lenghth of 176 ms (341 fpm), significantly lower than the 157 ms (382 fpm) recorded in the group of patients where the arrhythmia persisted (n=10). Similar observations were made by Boahene et al who also measured the fibrillatory cycle length from the right atrium in 55 patients with Wolff-Parkinson-White syndrome [<xref ref-type="bibr" rid="R9">9</xref>]. These investigators also found that patients with sustained AF (n=45) had shorter mean cycle lengths compared to patients with non-sustained AF (n=10).</p><p>Secondly, response to antiarrhythmic drug therapy has been shown to be associated with baseline cycle length and drug-induced cycle length changes. Stambler and colleagues identified a mean right atrial cycle length of 160 ms (375 fpm) as a valuable cutoff-point for conversion to sinus rhythm with ibutilide [<xref ref-type="bibr" rid="R10">10</xref>]. No patient with shorter cycle length (higher rate) was converted by ibutilide, whereas conversion occured in 64 % of those patients with longer cycle length (lower rate). Similarly, Fujiki et al [<xref ref-type="bibr" rid="R11">11</xref>] by using spectral analysis of right atrial electrograms reported 100 % AF conversions with cibenzoline or procainamide if the baseline atrial cycle length was > 168 ms (< 357 fpm) as opposed to only 17 % in patients with shorter cycle lengths (higher rates). Moreover, AF conversion was associated with cycle length prolongation. AF terminated after class I drug administration in 88 %, if the atrial cycle length had been prolonged to > 210 ms (< 285 fpm), in contrast to only 10 % if the post-drug cycle length was shorter.</p><p>Finally, there exists a close relation between the degree of electrical remodeling expressed as refractoriness changes and AF recurrence following cardioversion. Manios et al found that patients who failed to shorten the monophasic action potential duration to less then 195 ms at a pacing cycle length of 350 ms were more likely to have AF recurrence [<xref ref-type="bibr" rid="R12">12</xref>]. This finding was explained by a more abnormal rate adaptation curve, which is in agreement to an early study performed by Attuel and colleagues showing that poor or absent rate adaptation of atrial refractory periods (that is, the normally close correlation between stimulation cycle length and refractory period was lost) is related with vulnerability to AF [<xref ref-type="bibr" rid="R13">13</xref>]. These findings have more recently been replicated in patients undergoing internal cardioversion [<xref ref-type="bibr" rid="R14">14</xref>]. In their study, Biffi et al identified both a shorter atrial effective refractoriness and abnormal rate adaptation as independent predictors for AF relapses, the latter being the strongest predictor.</p><p>Prompted by the likely usefulness of non-invasive test that assesses the average fibrillatory rate from the surface ECG for exploration of AF pathophysiology and AF managment, frequency analysis techniques were independently introduced by our groups [<xref ref-type="bibr" rid="R15">15</xref>,<xref ref-type="bibr" rid="R16">16</xref>], while two earlier studies proposed to use this technique to subdivide AF and atrial flutter [<xref ref-type="bibr" rid="R17">17</xref>] or automatically identify AF among different rhythms [<xref ref-type="bibr" rid="R18">18</xref>].</p></sec><sec sec-type="" id="s2"><title>Frequency Analysis Techniques</title><sec id="s2a"><title>Methodological Considerations</title><p>Using digital recording techniques, traditional ECG signals are recorded. In most studies atrial fibrillatory rate has been obtained by spectral analysis techniques of resting ECG recordings such as standard 12-lead [<xref ref-type="bibr" rid="R15">15</xref>,<xref ref-type="bibr" rid="R16">16</xref>] or (modified) orthogonal recordings [<xref ref-type="bibr" rid="R19">19</xref>-<xref ref-type="bibr" rid="R21">21</xref>]. The method has, however, also been applied to ambulatory ECG recordings using conventional ambulatory leads [<xref ref-type="bibr" rid="R22">22</xref>-<xref ref-type="bibr" rid="R24">24</xref>].</p><p>The accuracy of frequency analysis techniques strongly relies on the fact that the largest possible fibrillatory waves should be present for further signal processing. Therefore, it was suggested to analyze lead V1 when using a standard ECG or, as proposed previously by Waktare et al [<xref ref-type="bibr" rid="R19">19</xref>-<xref ref-type="bibr" rid="R25">25</xref>], a “bipolar modification of V1” covering the atria (low C1) when applying different ECG recording systems (e.g. Holter ECG) [<xref ref-type="bibr" rid="R19">19</xref>-<xref ref-type="bibr" rid="R20">20</xref>].</p><p>Since the atrial and ventricular activities overlap spectrally, linear filtering techniques are not suitable for extraction of the fibrillatory signal from the surface ECG. Instead, subtraction of averaged QRST complexes needs to be performed producing a remaining atrial fibrillatory signal for further analysis (<xref ref-type="fig" rid="F1">Figure 1</xref>). Originally, a fixed averaged QRST-complex was used for cancellation in individual leads. Further development of the method allowed an improved cancellation of the QRST-complex by taking changes in QRS morphology due to alterations in the electrical axis of the heart into account (spatiotemporal QRST cancellation) [<xref ref-type="bibr" rid="R26">26</xref>].</p><p>Following QRST cancellation, a power spectrum is obtained by using a windowing technique and Fourier analysis to process the remainder ECG. This as well as the associated windowing technique (window type, length and overlap) determine the appearance of the frequency power spectrum [<xref ref-type="bibr" rid="R15">15</xref>]. Variants of Fourier transform based methods including conventional Fourier analysis and spectral averaging techniques based on short overlapping segments have been applied to ECG segments ranging from 10 seconds to 5 minutes [<xref ref-type="bibr" rid="R15">15</xref>,<xref ref-type="bibr" rid="R16">16</xref>,<xref ref-type="bibr" rid="R27">27</xref>-<xref ref-type="bibr" rid="R29">29</xref>]. Typically a distinct spectral peak is obtained which corresponds to the most dominant fibrillatory rate of nearby endocardial sites [<xref ref-type="bibr" rid="R15">15</xref>,<xref ref-type="bibr" rid="R16">16</xref>], but bi- or multimodal peaks are also sometimes observed [<xref ref-type="bibr" rid="R15">15</xref>,<xref ref-type="bibr" rid="R30">30</xref>].</p><p>More recently, a time-frequency analysis approach has been developed by which the time-frequency distribution is decomposed into a number of descriptive functions reflecting second-to-second variations in fundamental frequency and waveform morphology [<xref ref-type="bibr" rid="R31">31</xref>,<xref ref-type="bibr" rid="R32">32</xref>]. The main advantage of using this approach is that the local signal quality and structure is also assessed. The spectral profile differs from the conventional spectrum in that the different local spectra have been frequency aligned before averaging. This results in more distinct peaks and a more clearly discernable harmonic structure. With its second-to-second resolution, the method can rapidly detect changes in frequency, regularity, amplitude, waveform and signal structure and quality (<xref ref-type="fig" rid="F1">Figure 1</xref>).</p><p>To date, the dominant spectral peak has been presented as either dominating atrial cycle length (in ms), fibrillatory rate (in fpm) or fibrillatory frequency (in Hz). Since it has become evident that the results of this method might be of practical clinical importance, we believe that the results should be expressed in a way which is closest to the nomenclature of other surface ECG rate-variables (e.g. sinus rate, atrial tachycardia or flutter rate) [<xref ref-type="bibr" rid="R33">33</xref>]. The expression “fibrillatory rate”, with its unit “ibrillations per minute - fpm” may therefore be most appropriate [<xref ref-type="bibr" rid="R20">20</xref>]. Moreover, the calculation of cycle length from the frequency power spectrum with its standard unit Hz (cycle length in ms = 1000/frequency in Hz) seems somewhat troublesome [<xref ref-type="bibr" rid="R34">34</xref>]. While this calculation is appropriate for single measurements, its application to multiple simultaneous or repeated measurements such as pre- and post-drug states may introduce significant statistical and subsequently scientific errors, especially if the baseline frequency exhibits large inter-individual variability. The reason is, that the same frequency difference results in larger cycle length differences for low frequencies compared to high frequencies.</p></sec><sec id="s2b"><title>Method Validation and Reproducibility</title><p>A direct comparison between endocardially recorded electrograms and body surface recordings clearly evidences the validity of fibrillatory rate obtained from surface ECG as an index of the average atrial fibrillatory cycle and subsequently atrial refractoriness. Fibrillatory rates calculated from lead V1 substitute the right atrial free wall [<xref ref-type="bibr" rid="R15">15</xref>,<xref ref-type="bibr" rid="R16">16</xref>] whilst rates from an oesophageal lead reflect atrial septal and left atrial activity [<xref ref-type="bibr" rid="R15">15</xref>]. <xref ref-type="fig" rid="F2">Figure 2</xref> illustrates the close agreement of right atrial and coronary sinus fibrillatory rates with rates obtained from simultaneously recorded surface ECG lead V1 in persistent AF. Interestingly, rate differences and variability between ECG and electrograms increase with growing anatomical distance (right atrium - coronary sinus - pulmonary veins) to V1. Approximately 50 % of V1 and superior pulmonary vein rates are also within close range. In the other cases, the latter has been found to be substantially faster but also slower than V1 (right atrium), indicating that superior pulmonary veins might not necessarily be the driving source of the fibrillatory process.</p><p>Although the gross atrial fibrillatory pattern, studied by multiple simultaneous epicardial recordings, is reproducible already in repeated measurements of only 8 seconds duration [<xref ref-type="bibr" rid="R35">35</xref>], calculation of fibrillatory rates during steady state conditions from even longer time intervals discloses a true variability. Thus, increasingly longer recording times of up to 30 minutes result in a successive decrease of this variability. The reproducibility of the method is thus enhanced by prolonging the recording time during steady state conditions, allowing integration of atrial fibrillatory activity over longer periods. For practical reasons, steady state recordings may be restricted to one to five minutes, yielding rate variation coefficients of 2.1 % [<xref ref-type="bibr" rid="R15">15</xref>].</p><p>In persistent AF, there is minor short-term rate variability [<xref ref-type="bibr" rid="R15">15</xref>,<xref ref-type="bibr" rid="R16">16</xref>,<xref ref-type="bibr" rid="R27">27</xref>] and considerable diurnal variability (for details see below) [<xref ref-type="bibr" rid="R23">23</xref>] [<xref ref-type="bibr" rid="R24">24</xref>], whilst repeated daily frequency measurements at identical medication at the same time under similar conditions discloses an insignificant fibrillatory rate variability [<xref ref-type="bibr" rid="R36">36</xref>]. In contrast, rate variability in paroxysmal AF seems to be related to its natural course with a rate increase within the first five minutes of an AF episode [<xref ref-type="bibr" rid="R22">22</xref>] and a rate decrease prior to termination [<xref ref-type="bibr" rid="R22">22</xref>,<xref ref-type="bibr" rid="R28">28</xref>].</p></sec></sec><sec sec-type="" id="s3"><title>Exploring Autonomic Modulation of AF</title><p>Vagal as well as sympathetic stimulation have been shown to reduce atrial refractory periods and increase their heterogeneity [<xref ref-type="bibr" rid="R37">37</xref>-<xref ref-type="bibr" rid="R39">39</xref>]. In the human atrium beta stimulation has been found to predominate over vagal stimulation [<xref ref-type="bibr" rid="R38">38</xref>]. Moreover, invasive studies performed in subjects with sinus rhythm have suggested a circadian pattern in atrial refractoriness with longer refractory periods during night time and refractory period shortening during daytime [<xref ref-type="bibr" rid="R40">40</xref>,<xref ref-type="bibr" rid="R41">41</xref>] which supports the role of the autonomic nervous system in modulating atrial electrophysiologic properties.</p><p>Subsequently, atrial fibrillatory rate seems to be ideal for monitoring the effect of autonomic tone changes - either spontanenous (circadian) or by autonomic maneuvers provoked - on atrial electrophysiology. Indeed, the circadian variability of atrial fibrillatory rate has been explored in independent studies [<xref ref-type="bibr" rid="R23">23</xref>,<xref ref-type="bibr" rid="R24">24</xref>]. Fibrillatory rate obtained from Holter ECG’s with persistent AF showed a significant decrease at night and an increase in the morning hours studies [<xref ref-type="bibr" rid="R23">23</xref>,<xref ref-type="bibr" rid="R24">24</xref>]. In 6 out of 30 individuals studied by our group studies [<xref ref-type="bibr" rid="R23">23</xref>], dominant nocturnal fibrillatory rate increased, however, concomitantly with a decrease in ventricular rate, while the opposite change occurred in the morning hours.</p><p>The second area of investigation concerns effects on atrial fibrillatory rate following vagal or sympathetic stimulation during experimental conditions. Carotid sinus massage, supposed to mainly induce vagal stimulation, resulted in a variable effect on fibrillatory rate in 19 patients studies [<xref ref-type="bibr" rid="R42">42</xref>]. A reproducible decrease was noted in 9 individuals, whilst a rate increase occurred in 8 and no change was observed in two patients. Interestingly, calcium-channel blocker treatment was the only variable effecting the rate response to carotid sinus massage. Calcium-channel blockers were more frequently used in patients with a decrease in fibrillatory rate compared to patients with a rate increase. The effect of adrenergic stimulation via head-up tilting on fibrillatory rate was studied in 14 patients with long-lasting AF studies [<xref ref-type="bibr" rid="R43">43</xref>]. In 12 patients head-up tilting increased fibrillory rate significantly, while there was no rate change in the remaining two.</p><p>With the availability of instantaneous fibrillatory rates obtained from time-frequency analysis, their second-to-second variation can be explored by spectral analysis techniques, similar to those used for analyzing heart rate variability. By applying this approach, Stridh et al studies [<xref ref-type="bibr" rid="R44">44</xref>] noted a spectral peak in 2 out of 8 patients with permanent AF at the breathing frequency of 0.125 Hz during controlled respiration that disappeared after atropine injection.</p><p>All these findings together clearly highlight the complexity of autonomous nervous system effects on atrial electrophysiology and justify further exploration.</p></sec><sec sec-type="" id="s4"><title>Monitoring and Predicting Antiarrhythmic Drug Effects</title><p>Class I and III antiarrhythmic drugs have been shown to increase atrial cycle length (decrease fibrillatory rate) which coincides with increased refractoriness and decreased conduction velocity [<xref ref-type="bibr" rid="R45">45</xref>]. These mixed effects might explain why atrial cycle length is closely related with baseline refractoriness [<xref ref-type="bibr" rid="R6">6</xref>] but not with refractoriness after antiarrhythmic drug administration [<xref ref-type="bibr" rid="R46">46</xref>]. Even though the individual contribution of refractoriness prolongation or conduction slowing might currently not be differentiated, monitoring drug action on atrial fibrillatory rate seems well fit to be explored by analysing the surface ECG. Indeed, a substantial reduction in atrial fibrillatory rates following several different intravenously or orally administered class I and III antiarrhythmic drugs as well as following verapamil or magnesium has been observed (<xref ref-type="table" rid="T1">Table 1</xref>).</p><p>Two examples of drug monitoring are presented in <xref ref-type="fig" rid="F3">Figure 3</xref> showing the transition from a high rate (less organized) to a low rate (more organized) fibrillation following acute intravenous sotalol infusion or 3-day oral flecainide administration.</p><p>Besides direct monitoring of antiarrhythmic drug effects, it seems also possible to identify suitable patients for pharmacological cardioversion. A baseline fibrillatory rate of 360 fpm was highly sensitive and specific for prediction of AF termination following intravenous ibutilde [<xref ref-type="bibr" rid="R16">16</xref>] or oral flecainide [<xref ref-type="bibr" rid="R19">19</xref>] (<xref ref-type="fig" rid="F4">Figure 4</xref>). In contrast, Fujiki et al noted no baseline fibrillatory rate difference between patients who converted to sinus rhythm and those who did not following oral bepridil administration. Instead, larger rate increases were associated with AF termination. While this concept is interesting and well established in experimental AF [<xref ref-type="bibr" rid="R45">45</xref>], it needs to be emphasized that these authors calculated atrial cycle length from the frequency spectrum, which may have introduced the aforementioned statistical errors [<xref ref-type="bibr" rid="R34">34</xref>].</p><p>Patients with a low fibrillatory rate may have a small number of wavelets (long wavelength), whereas those with higher rates have multiple wavelets (short wavelength) [<xref ref-type="bibr" rid="R7">7</xref>]. In the former group class I or III antiarrhythmic drugs by decreasing fibrillatory rate may have increased wavelength (or the excitable gap) and therefore reduced the number of wavelets that could coexist. This would have increased the statistical chance that all wavelets might extinguish simultaneously and terminate the fibrillatory process [<xref ref-type="bibr" rid="R47">47</xref>].</p><p>One previous study suggested that a stepped conversion regimen of first-line ibutilide followed by electrical cardioversion for patients who fail to convert is less expensive and has a higher conversion rate than first-line electrical cardioversion [<xref ref-type="bibr" rid="R48">48</xref>]. Given the expense of antiarrhythmic therapy and the risk of side effects including ventricular proarrhythmia, a test that differentiates responders from non-responders is likely to be even more cost-effective. In addition, drug monitoring using frequency analysis techniques may be useful for finding optimal drug dosages and timing of interventions in the individual patient.</p></sec><sec sec-type="" id="s5"><title>Predicting AF Recurrence</title><p>Previous investigations [<xref ref-type="bibr" rid="R49">49</xref>,<xref ref-type="bibr" rid="R50">50</xref>] have shown that most AF relapses occur within the first weeks after cardioversion with decreased but constant recurrence rates thereafter. Early vulnerability to AF re-initiation within this time period is related to electrophysiological abnormalities, while structural abnormalities seem to be primarily responsible for later AF recurrences [<xref ref-type="bibr" rid="R51">51</xref>]. This time course might be explained by the fact that reversal of the electrical remodeling process occurs rapidly once sinus rhythm is restored [<xref ref-type="bibr" rid="R12">12</xref>,<xref ref-type="bibr" rid="R52">52</xref>,<xref ref-type="bibr" rid="R53">53</xref>], while structural changes persist for longer periods [<xref ref-type="bibr" rid="R51">51</xref>]. Subsequently, characterization of atrial electrophysiology has been suggested for identification of patients at risk for early AF recurrence [<xref ref-type="bibr" rid="R2">2</xref>].</p><p>Atrial premature beats with short coupling intervals have been shown to promote early AF reinitiation following cardioversion [<xref ref-type="bibr" rid="R54">54</xref>,<xref ref-type="bibr" rid="R55">55</xref>]. AF reinduction by an atrial premature beat relies on the fact that a relatively short atrial wavelength (conduction velocity x refractory period) must be present [<xref ref-type="bibr" rid="R56">56</xref>]. As highlighted before, atrial fibrillatory rate reflects atrial refractoriness [<xref ref-type="bibr" rid="R5">5</xref>,<xref ref-type="bibr" rid="R6">6</xref>], and might subsequently represent a marker for early AF susceptibility following restoration of sinus rhythm. Indeed, previous studies reported higher rates in relapsed patients immediately prior internal [<xref ref-type="bibr" rid="R57">57</xref>] or external cardioversion [<xref ref-type="bibr" rid="R21">21</xref>] when compared with non-relapsed patients and also close relationships between fibrillatory rate and defibrillation thresholds [<xref ref-type="bibr" rid="R57">57</xref>,<xref ref-type="bibr" rid="R58">58</xref>].</p><p>Two previous studies [<xref ref-type="bibr" rid="R21">21</xref>,<xref ref-type="bibr" rid="R59">59</xref>] have investigated the combined predictive value of fibrillatory rate and echocardiographic left atrial parameters. In one study [<xref ref-type="bibr" rid="R59">59</xref>], the authors calculated an index combining shortest cycle length from oesophageal or V1 lead and standard left atrial diameter and were able to show that patients with recurring AF had significantly lower values than patients who remained in sinus rhythm. Another study [<xref ref-type="bibr" rid="R21">21</xref>] demonstrated that the combination of fibrillatory rate and systolic left atrial area predicted early AF recurrence after succesful cardioversion with a high accuracy and was able to provide individual risk estimates (<xref ref-type="fig" rid="F5">Figure 5</xref>).</p><p>These parameters seem therefore well suited to describe the individual atrial remodeling. The concept of ECG-guided cardioversion may gain even greater importance in the light of previous findings from two studies comparing rate- with rhythm-control strategies [<xref ref-type="bibr" rid="R60">60</xref>,<xref ref-type="bibr" rid="R61">61</xref>]. In these studies, rhythm-control seemed not to be superior than rate-control in asymptomatic, mostly eldery patients with recurring AF and structural heart disease. This highlights the need to select candidates for cardioversion not only based on clinical judgement but also on measures that are able to determine the likelihood of maintaining sinus rhythm such as rate parameters obtained from the surface ECG.</p></sec><sec sec-type="" id="s6"><title>Conclusions</title><p>Atrial fibrillatory rate and its variability can be reliable obtained from the surface ECG in AF patients using spectral analysis techniques. These parameters exhibit a significant interindividual variability allowing individual quantification of the atrial electrical remodeling process. Frequency analysis of AF might prove useful in identification of underlying AF pathomechanisms, and prediction of therapy efficacy (drug-induced conversion, maintenance of sinus rhythm, selecting antiarrhythmic drugs, identifying candidates for non-pharmacological AF therapy). Further larger studies are necessary to determine the role of these techniques in different AF management strategies in order to select and time the appropriate therapy for the individual patient.</p></sec> |
Sensors for rate responsive pacing | <p>Advances in pacemaker technology in the 1980s have generated a wide variety of complex multiprogrammable pacemakers and pacing modes. The aim of the present review is to address the different rate responsive pacing modalities presently available in respect to physiological situations and pathological conditions. Rate adaptive pacing has been shown to improve exercise capacity in patients with chronotropic incompetence. A number of activity and metabolic sensors have been proposed and used for rate control. However, all sensors used to optimize pacing rate metabolic demands show typical limitations. To overcome these weaknesses the use of two sensors has been proposed. Indeed an unspecific but fast reacting sensor is combined with a more specific but slower metabolic one. Clinical studies have demonstrated that this methodology is suitable to reproduce normal sinus behavior during different types and loads of exercise. Sensor combinations require adequate sensor blending and cross checking possibly controlled by automatic algorithms for sensors optimization and simplicity of programming. Assessment and possibly deactivation of some automatic functions should be also possible to maximize benefits from the dual sensor system in particular conditions. This is of special relevance in patient whose myocardial contractility is limited such as in subjects with implantable defibrillators and biventricular pacemakers. The concept of closed loop pacing, implementing a negative feedback relating pacing rate and the control signal, will provide new opportunities to optimize dual-sensors system and deserves further investigation. The integration of rate adaptive pacing into defibrillators is the natural consequence of technical evolution.</p> | <contrib contrib-type="author"><name><surname>Dell'Orto</surname><given-names>Simonetta</given-names></name><degrees>MD</degrees></contrib><contrib contrib-type="author"><name><surname>Valli</surname><given-names>Paolo</given-names></name><degrees>MD</degrees></contrib><contrib contrib-type="author"><name><surname>Greco</surname><given-names>Enrico Maria</given-names></name><degrees>MD</degrees></contrib><aff>Unità Operativa Cardiologia, Ospedale Uboldo, Azienda Ospedaliera di Melegnano; Cernusco sul Naviglio, Milano, Italy</aff> | Indian Pacing and Electrophysiology Journal | <p>The advances in pacemaker technology in the 1980s have generated a wide variety of complex multiprogrammable pacemakers and pacing modes. Pacing systems can be classified as (1) single lead ventricular, such as VVI or ventricular demand; (2) atrial based, single lead, such as AAI or atrial demand; or dual chamber, such as DDD, which senses and paces from atrial and ventricular chambers. These pacing modalities may or may not be rate adaptive or sensor-driven. In the pacemaker code, rate adaptive pacemakers are designated with a fourth letter R, for example a rate adaptive DDD becomes DDDR. Rate adaptive pacemakers are useful for patients who cannot increase their heart rate appropriately on exercise.</p><p>The goal of new technologies is to come as close as possible to sinus node electrical activity, combining the need of regular ventricular rate to a better quality of life. In this context the clinical benefit of increasing ventricular rate, and consequently cardiac output, during physical activity in patients with “chronotropic incompetence” consequent to Sick Sinus Syndrome or advanced AV block, has been established by different studies [<xref ref-type="bibr" rid="R1">1</xref>-<xref ref-type="bibr" rid="R7">7</xref>].</p><p>The ACC/AHA guidelines [<xref ref-type="bibr" rid="R8">8</xref>] define chronotropic incompetence as failure to achieve a heart rate of 100 beats/min at maximal exertion. From the practical standpoint heart rate at a given metabolic load should mirror the line that describes the correlation of heart rate to metabolic demand for age, sex, and weight matched control group [<xref ref-type="bibr" rid="R8">8</xref>]. This definition avoids a fixed cutoff heart rate as a criterion, and considers the trend of heart rate during different intensity of exercise as the frame of reference [<xref ref-type="bibr" rid="R9">9</xref>-<xref ref-type="bibr" rid="R11">11</xref>].</p><sec sec-type="" id="s1"><title>Fixed rate versus rate responsive pacing</title><p>Many studies have shown that during exercise, an increase in the pacing rate provided by VVIR, VDD, DDD or DDDR modes augments the cardiac output, achieved workload, and duration of exercise more than does fixed-frequency VVI pacing in patients with both normal or impaired LV function [<xref ref-type="bibr" rid="R2">2</xref>-<xref ref-type="bibr" rid="R7">7</xref>,<xref ref-type="bibr" rid="R12">12</xref>-<xref ref-type="bibr" rid="R15">15</xref>]. Beside superior hemodynamic effects, there is an increase in maximum oxygen consumption, a reduction in arteriovenous oxygen difference and an increase in subject well-being. Moreover, the acute hemodynamic advantage is retained on a long term basis: by 6 and 12 months after implantation, dual chamber and rate adaptive pacemakers may further augment LV function, reduce heart size and improve ventricular performance compared with results in the immediate postoperative period [<xref ref-type="bibr" rid="R16">16</xref>-<xref ref-type="bibr" rid="R17">17</xref>].</p><p>When patients with complete AV block exercise, the sinus rate is significantly higher with VVI pacing than during dual chamber or VVIR modes, a response possibly reflecting the increased activity of the sympathetic nervous system when pacing is set on VVI mode [<xref ref-type="bibr" rid="R18">18</xref>-<xref ref-type="bibr" rid="R19">19</xref>]. Indeed, coronary sinus norepinephrine is higher in patients with VVI pacemaker during exercise. The increase in catecholamine on exercise during VVI pacing is likely to be related to the need of improving contractility and therefore cardiac output to compensate the lack of rate response. This enhanced cardiac sympathetic activity may eventually produce an adverse effect on LV function, with the possible development of congestive heart failure and arrhythmias.</p><p>In patients with atrial chronotropic incompetence, VVIR and DDDR pacing modes are clearly superior to the DDD mode in terms of exercise performance, because the sensor increases the pacing rate according to metabolic needs. Most studies [<xref ref-type="bibr" rid="R20">20</xref>-<xref ref-type="bibr" rid="R21">21</xref>]. on patients with atrial chronotropic incompetence and DDDR pacemakers have shown superior hemodynamic performance on exercise, and patients prefer the DDDR mode to the VVIR mode.</p><p>Moreover, preliminary data suggest that rate adaptive AAIR and DDDR modes may be more efficacious in preventing atrial arrhythmias than their non rate adaptive counterparts in Sick Sinus Syndrome [<xref ref-type="bibr" rid="R22">22</xref>-<xref ref-type="bibr" rid="R24">24</xref>]. DDDR pacemakers may prevent arrhythmias by eliminating the relative bradycardia noted during exercise in patients with non-adaptive devices, when excessive catecholamine release may increase the likelihood of atrial arrhythmias.</p></sec><sec sec-type="" id="s2"><title>Rate responsiveness</title><p>Rate-adaptive pacing has been designed to increase heart rate according to metabolic needs during physical, mental or emotional activity. Rate responsive pacemakers control heart rate by sensing physiological or nonphysiological signals other than atrial rate.</p><p>Ideally the rate adaptive sensors should reproduce the sinus node as close as possible; therefore some definite properties must be accomplished: (1) the chronotropic output should respond as promptly as the normal sinus node. (2) Sensors should perform a highly specific and sensitive detection of the need of increasing heart rate. (3) These latter have also to be proportional to metabolic demand. (4) Rate decay during recovery after exercise should match metabolic needs (i. e. fast after short exercise but prolonged after longer and maximal exercise in response to an oxygen debt or in pathological conditions like heart failure); (5) it should ideally operate in a closed loop system, making rate adaptive pacing also insensitive to inputs not heart related. Finally, (6) dedicated sensors should avoid undesiderable overpacing and the need of complex programming.</p><p>Different parameters have been investigated for controlling the pacemaker rate: oxygen saturation [<xref ref-type="bibr" rid="R25">25</xref>], venous pH [<xref ref-type="bibr" rid="R26">26</xref>], QT interval [<xref ref-type="bibr" rid="R27">27</xref>,<xref ref-type="bibr" rid="R28">28</xref>], body motion [<xref ref-type="bibr" rid="R28">28</xref>], respiratory rate [<xref ref-type="bibr" rid="R29">29</xref>], stroke volume [<xref ref-type="bibr" rid="R30">30</xref>], central venous temperature [<xref ref-type="bibr" rid="R31">31</xref>-<xref ref-type="bibr" rid="R33">33</xref>], minute ventilation [<xref ref-type="bibr" rid="R34">34</xref>], peak endocardial acceleration [<xref ref-type="bibr" rid="R35">35</xref>], and changes of the right ventricular impedance during the cardiac cycle (CLS, closed loop stimulation) [<xref ref-type="bibr" rid="R36">36</xref>]. Clinical studies have outlined advantages and limitations of the different sensed parameters. Finally complexity of implanting and programming, the evidence of instability related to influence of external conditions or concomitant disease, have defined the inadequacy of some parameters to the required characteristics, and only some of these indicators are still used as single or dual sensor technology.</p></sec><sec sec-type="" id="s3"><title>Single sensor technology</title><p>Activity sensors are the older and more widely used. The working modality is based on the relationship between activity and heart rate. Activity may be acknowledged either by a piezoelectric crystal, which recognizes the muscular pressure waves, produced by physical activity and convert them to an electrical signal sent to the pacemaker, or by an accelerometer that identifies the postural changes and the body movements related to physical activity.</p><p>Activity sensors offer rapid response to exercise by assessing body vibrations or movements. A rapid response plays an important role in “burst activity” during daily life. Fast reaction to termination of short exercise and technical simplicity that allows for instance to tailor the rate response (RR) to the single patient with proper treadmill protocols, [<xref ref-type="bibr" rid="R37">37</xref>] represent further advantages of this sensor type. However, after longer exercise, an oxygen debt may require a sustained rate increase, which is not provided by activity sensors during recovery because these sensors are unable to recognize the oxygen debt. Moreover, low specificity with inappropriate rate increase in conditions like laughing, coughing, driving, the fact that activity sensors does not respond to activity not related to body movements (isometric exercise, mental stress, or metabolic inadequacy consequent to pathologic conditions), and the possible mismatch between exercise intensity and rate increase, represent the main limitations of activity sensors.</p><p>Metabolic sensors, based on QT interval, minute ventilation (MV) or peak endocardial acceleration, provide pacing rates more closely and specifically related to physical and mental stress requirements.</p><p>Minute ventilation, the product of respiratory rate and tidal volume, is a physiological indicator that has been shown to be correlated with metabolic demand [<xref ref-type="bibr" rid="R38">38</xref>,<xref ref-type="bibr" rid="R39">39</xref>]. This parameter, which also correlate linearly with heart rate [<xref ref-type="bibr" rid="R40">40</xref>,<xref ref-type="bibr" rid="R41">41</xref>], can be derived from variations in transthoracic impedance signal. RR pacemakers, using impedance MV sensors, change the pacing rate in response to the variations in the patients MV.</p><p>Limitations of the MV sensor include the lower reliability in patients with obstructive pulmonary disease, false positive reaction in hyperventilation or interference with cardiac monitors [<xref ref-type="bibr" rid="R42">42</xref>] and posture [<xref ref-type="bibr" rid="R43">43</xref>].
Sensors using QT interval variations [<xref ref-type="bibr" rid="R27">27</xref>,<xref ref-type="bibr" rid="R44">44</xref>] are based on the finding that physical activity and circulating catecholamine produce shortening of the QT interval. These sensors are highly specific; furnish sustained increase of sensor-driven heart rate during post-exercise recovery to compensate for an oxygen debt, and are responsive to mental stress. However, measurement of evoked QT interval may be unreliable in T wave undersensing; it can not be used in patients with acute myocardial infarction, is affected by drugs, electrolyte disturbances and increased circulating catecholamine, a common condition in patients with congestive heart failure. Because it requires ventricular pacing, it can not be used in AAIR mode.</p><p>More recently, a sensor that assesses mechanical vibrations generated by the myocardium during the isovolumetric contraction phase (peak endocardial acceleration [PEA]), has been developed. A micro accelerometer is housed inside a rigid, perfectly hermetic capsule inserted in the tip of a standard unipolar pacing lead. The rigidity of the capsule prevents the generation of artifacts that may arise from compression of the electrode by the cardiac muscle during contraction. Therefore, the sensor is only sensitive to the inertial forces generated by myocardium movements. An associated electronic circuit pre-processes the signal to ensure its correct transmission trough the catheter.</p><p>Experimental and clinical trials have shown that ∆PEA is correlated with dP/dT max [<xref ref-type="bibr" rid="R35">35</xref>,<xref ref-type="bibr" rid="R45">45</xref>,<xref ref-type="bibr" rid="R46">46</xref>] and is consequently related to contractile function. ∆PEA dynamic monitoring has been shown to provide fast pacing rate responses with long term performance of sensor lead and effective and rapid RR tailoring [<xref ref-type="bibr" rid="R35">35</xref>] [<xref ref-type="bibr" rid="R47">47</xref>], also in patients with heart failure and wide QRS [<xref ref-type="bibr" rid="R48">48</xref>]. Moreover Peak Endocardial Acceleration assessment allows AV delay automatic optimization in DDDR pacing that can be calculated directly from the device using a time saving procedure [<xref ref-type="bibr" rid="R49">49</xref>]. Experimental data in animals indicate that PEA monitoring is feasible also during atrial fibrillation [<xref ref-type="bibr" rid="R50">50</xref>].</p><p>PEA sensor is usually combined with activity sensors, however, in view of its rapid and appropriate heart rate response in different conditions, the possibility to use PEA sensor as a single sensor should also be considered.</p></sec><sec sec-type="" id="s4"><title>Dual sensors pacemakers</title><p>As previously described no single sensor can reproduce sinus node behavior in all the different activities of daily life. To overcome these limitations a possible option is to combine two different sensors. Rate modulation has to be considered at three different levels: (1) short term response for effort or emotions; (2) medium term adaptation for circadian variation of heart rate during day- and night-time; and (3) long term regulation to obtain different rate variations according to rest and activity periods [<xref ref-type="bibr" rid="R51">51</xref>].</p><p>Combining different sensors might more closely mimic intrinsic heart rate, if the chosen sensors are complementary. The most common combination include association of an activity sensor giving a rapid response for light or for short duration exercise, and a metabolic sensor, e.g. QT interval or minute ventilation (MV) that provides a delayed but proportional and stable acceleration to sustained exercise and deceleration during recovery [<xref ref-type="bibr" rid="R51">51</xref>]. Another option in rate response devices is to obtain circadian heart rate variation with two different hourly mean rates during day and night. Physiologic sensors and activity sensors could provide rate variations based on signal sensor solicitation. Two lower heart rates are programmed for daytime and nighttime. When the sensor is constantly solicited, the daytime lower rate is used. On the contrary, when the signal sensor level is low for a consistent period of time, the device switches on nighttime lower rate. Metabolic sensors could provide a modulation of the algorithm curve slope according to the long term activity.</p></sec><sec sec-type="" id="s5"><title>Sensors optimization</title><p>Independently of the choice of sensors and mode of integration, algorithms for sensors optimization determine the performance of dual sensor rate adaptive pacemakers.</p><p> Combining sensors with different rate responses requires adequate blending of respective sensor activities. Blending can be performed at signal production. The resulting signal transmitted to the algorithm is a mixture of a percentage of activity sensor signals (0%…10%) with an inverse percentage of non activity sensor signals (100%…0%). This blending modality is used for instance in the Vitatron device, combining QT interval and activity in 5 different possibility. This blended signal is transmitted to the algorithm working with a variable automatic slope [<xref ref-type="bibr" rid="R51">51</xref>].</p><p>Another possibility for sensor blending is priorization as in Medtronik Kappa 400: the activity sensor initially accelerates from the lower rate to a plateau (programmable 90-95 b/min). The rate returns to the lower rate if activity stops, or is proportionally increased from the plateau up to the sensor maximal programmed rate, if the minute ventilation sensor activates. MV sensor is then in charge during the recovery rate decrease.</p><p>Sensors cross checking are used to avoid inappropriate rate increase. During crosscheck both sensors can control each other and the pacing rate will only be changed if both or a predominant sensor agrees. For example, after administration of a drug that shortens the QT interval, a QT interval sensor would indicate the need for rate increase, but the pacing rate would not change because the activity sensor is not activated. Conversely, passively tapping on the device would activate the activity sensor and indicate a rate increase, but the pacing rate would not be modified because the QT-interval sensor would not be activated by this manouver.</p></sec><sec sec-type="" id="s6"><title>Algorithms for optimization of dual sensor performance</title><p>Algorithms for sensors optimization determine the performance of dual-sensor rate-adaptive pacing systems. Automatic setting has been developed in complex pacemakers to simplify programming and optimize time. It is necessary for physicians to be able to verify that rate adaptive pacemakers respond correctly, according to the patient’s need. Nevertheless, even though these systems are generally reliable [<xref ref-type="bibr" rid="R47">47</xref>], manual access to sensor programmability is important and should be performed by physicians with a thorough knowledge of the sensors capabilities. In addition, a specific apparatus for O<sub>2</sub> consumption measurement during physical exercise (cardiopulmonary stress test) should be used. Alternatively an algorithm computing the best correlation between heart rate and metabolic needs, such as Pacing Rate Profile Software (PRPS), ought to be used [<xref ref-type="bibr" rid="R47">47</xref>].</p><p>Future devices may provide the opportunity to use physiologic sensors to monitor cardiac function and to adapt pacemaker function to assist therapy for associated disorders. Multisensor devices can be used for cardiac rehabilitation in pacemaker dependent patients, particularly the elderly and affected by cardiopulmonary disease. These patients deserve a physically and psychologically autonomous life style, which may be accomplished by using two rehabilitation methods: (1) the set up of appropriate rate response; and (2) the institution of aerobic training programs [<xref ref-type="bibr" rid="R52">52</xref>] for in- and outpatients.</p><p>Moreover, the integration of rate adaptive pacing into biventricular pacemakers and implantable defibrillators is a natural consequence of technical evolution. The rationale for the use of rate adaptive pacing in implantable defibrillators is the same as for pacemaker. In normal heart, an increase in oxygen uptake from 3 to 50 mL/kg per minute in a healthy subject is due to an increase of oxygen extraction rate, stroke volume (by a factor of 1.5) and heart rate (by a factor of 3). This allows a 4-5 fold increase in cardiac output compared to resting values. In the large majority of patients receiving an implantable defibrillator, the contractility reserve is limited. Therefore, an increase in cardiac output is strictly related to the possibility of increasing heart rate. Concomitant therapy with beta blockers, amiodarone or other antiarrhythmic drugs may also impair the chronotropic response. This makes the issue of chronotropic competence crucial in patients with biventricular stimulation and implantable defibrillator despite the fact that only 20% of these latter need primarily a cardiac pacing [<xref ref-type="bibr" rid="R8">8</xref>].</p><p>The concept of closed loop pacing should be the next step in future technical developments. Sensors that could be used in a closed loop system, indicating whether heart rate is adequate to a given metabolic situation, are endocardial accelerometers and sensors using impedance derived ventricular signals. It might take some time to test technical feasibility and clinical reliability of those closed loop systems before they will be implemented in cardiac pacemakers and implantable defibrillators or biventricular pacemakers.</p></sec> |
Figure-8 Tachycardia Confined to the Anterior Wall of the Left Atrium | <p>Incisional atrial tachycardias have been described most frequently in patients with previous corrective surgery for congenital heart defects and mitral valve disease. Less information is available on atrial tachycardias appearing late after isolated aortic valve surgery. We report the case of a patient who developed a left figure-8 tachycardia after undergoing aortic valve replacement. During electrophysiologic study the entire cycle length of the tachycardia was mapped within a low voltage area confined to the left anterior atrial wall. However, during ablation a transmural lesion could not be attained. The mapping and ablation strategy along with the mechanism of the tachycardia are discussed.</p> | <contrib contrib-type="author"><name><surname>Liuba</surname><given-names>Ioan</given-names></name><degrees>MD</degrees></contrib><contrib contrib-type="author"><name><surname>Jönsson</surname><given-names>Anders</given-names></name><degrees>MD</degrees></contrib><contrib contrib-type="author"><name><surname>Walfridsson</surname><given-names>Hakan</given-names></name><degrees>MD</degrees></contrib><aff>Department of Cardiology, Heartcenter, University Hospital Linköping, Linköping, Sweden</aff> | Indian Pacing and Electrophysiology Journal | <sec sec-type="" id="s1"><title>Case report</title><p>A 57-year-old man with dilated cardiomyopathy and chronic heart failure, who had undergone aortic valve replacement 26 years ago, was referred to our department for a third attempt at radiofrequency (RF) catheter ablation of an atrial tachycardia. According to reports, during the previous two electrophysiologic (EP) studies both a typical atrial flutter and a left atrial reentrant tachycardia (cycle length 190 ms) could be induced with burst pacing. Nevertheless, RF catheter ablation had failed in both cases, either because it was not possible to attain bidirectional block across the cavotricuspid isthmus (despite a large number of RF applications), or due to transformation of left reentrant tachycardia into atrial fibrillation during entrainment mapping and the impossibility of reinducing the same tachycardia after DC cardioversion. The patient continued to have frequent, prolonged episodes of palpitations despite several antiarrhythmic drug trials. Twelve-lead ECG documented an atrial tachycardia with a p-p interval of 240-280 ms and a variable degree of AV block as well as atrial fibrillation. The patient underwent several successful DC cardioversions for both arrhythmias. Transthoracic echocardiography revealed left atrial and ventricular enlargement, left ventricular systolic and diastolic dysfunction, and mild mitral and moderate tricuspid regurgitation.</p><p>At the time of the third EP study, a decapolar catheter was placed in the coronary sinus (CS) and a bipolar catheter was positioned in the right ventricle. An 8-French deflectable quadripolar irrigated-tip catheter (Cordis-Webster, Thermocool, F-curve) was used for mapping and ablation. Mapping was performed using the Carto system (Carto; Biosense Inc). RF current was delivered through a Cordis Stockert generator between the 4-mm tip of the catheter and an adhesive patch on the posterior chest wall. Ablation was performed in temperature-controlled mode with a target temperature of 50°C and a power limit of 50 W. Saline was infused at a rate of 3 ml/min between ablations and 30 ml/min during RF current delivery. After obtaining a right atrial activation map with an electroanatomical reconstruction of the right atrium during proximal CS pacing, a regular tachycardia, with ECG morphology similar to the clinically documented tachycardia, was induced with burst pacing (<xref ref-type="fig" rid="F1">Figure 1</xref>). The CS electrograms were of low amplitude, with electrograms in CS 9-10 starting 23 ms before electrograms in CS 1-2.</p><p>Sequential mapping of the right atrium during tachycardia revealed a nonreentrant activation pattern with an early septal area. Mapping of the left atrium was performed through transeptal catheterization. Two discrete scars (areas with bipolar voltage ≤0.05 mV) (1-3) along with adjacent points displaying fragmented or double potentials were located on the left anterior wall. They were confined to a larger region of low voltage (≤0.5 mV). The entire cycle length of the tachycardia was recorded within this region and the activation map was consistent with a figure-8 circuit. We hypothesized that activation propagated craniocaudally, through a common channel bordered by two lines of double potentials in its upper portion (<xref ref-type="fig" rid="F2">Figure 2</xref>). In its inferior part this channel was bordered medially by a discrete scar. One loop then revolved medially, around a transverse line of double potentials and anterior to the ostium of the right inferior pulmonary vein. The second loop revolved laterally, through a channel bordered by the first line of double potentials and the second, lower, scar. A fourth possible channel appeared between the transverse line of double potentials and the medial scar. In order to avoid terminating or transforming the tachycardia or inducing atrial fibrillation, entrainment was not performed in this patient.</p><p>Ablation targeted the common isthmus (isolated RF applications) and both the common channel and the medial turnaround of the circuit (point-by-point and dragging technique, aiming also to transect the possible fourth channel). Both attempts failed due to poor catheter stability. With the aim of continuing the ablation procedure under coronary sinus pacing, we attempted to interrupt the tachycardia by overdrive pacing, but this maneuver led to transformation into a second tachycardia with a CL of 190 ms (not shown in this paper). This tachycardia was DC-converted to sinus rhythm and the procedure continued afterwards under CS pacing. Nevertheless, the same technical difficulties further precluded the successful ablation of the circuit. Remapping the ablation points confirmed persistence of electrograms with the same configuration and the procedure was interrupted (long procedure time).</p></sec><sec sec-type="" id="s2"><title>Discussion</title><p>Tachycardias with figure-8 activation pattern have been described both in the right and left atrium, most frequently in patients who have undergone surgery for congenital heart defects and valve disease [<xref ref-type="bibr" rid="R1">1</xref>-<xref ref-type="bibr" rid="R7">7</xref>]. As with other incisional tachycardias, the electrophysiological substrate is provided by the presence of areas of low voltage and depressed conductivity, electrically silent zones and lines of block that are attributable to an insufficient myocardial protection of cardioplegia, atriotomies and cannulating sites [<xref ref-type="bibr" rid="R3">3</xref>,<xref ref-type="bibr" rid="R4">4</xref>,<xref ref-type="bibr" rid="R8">8</xref>]. In the right atrium the loops are often disposed around the tricuspid annulus and lateral atriotomies, while in the left atrium the geometry seems more variable, with loops rotating around mitral annulus, electrically unexcitable areas, or the ostia of the pulmonary veins [<xref ref-type="bibr" rid="R1">1</xref>,<xref ref-type="bibr" rid="R2">2</xref>,<xref ref-type="bibr" rid="R4">4</xref>,<xref ref-type="bibr" rid="R5">5</xref>].</p><p>It is conceivable that a figure-8 activation pattern may be encountered in both reentrant and focal tachycardias [<xref ref-type="bibr" rid="R9">9</xref>]. The latter scenario supposes the existence of a discharging focus inside a protected isthmus having a zone of slow conduction at one of its extremities. Entrainment is therefore of particular interest in such tachycardias since its demonstration may relatively simply differentiate between a reentrant and a focal mechanism. Furthermore, a post-pacing interval equal to tachycardia cycle length indicates a site within the reentry circuit while supplementary demonstration of a short stimulus-to-P wave interval points out the exit of the protected isthmus. However, one of the main shortcomings of entrainment mapping is the possibility of interrupting or transforming the tachycardia during rapid pacing. This aspect is of particular concern in patients with left atrial tachycardias [<xref ref-type="bibr" rid="R4">4</xref>]. Therefore, some authors prefer to delineate the circuit by performing a high-density activation mapping of the entire cavity avoiding thus entrainment [<xref ref-type="bibr" rid="R1">1</xref>-<xref ref-type="bibr" rid="R3">3</xref>]. In this case, the reentrant nature of the tachycardia as well as the configuration of the circuit are finally proven by successfully interrupting the loops during ablation (Mines’ test). RF current may be delivered either in the common isthmus or separately for each of the two loops (in the latter case, sometimes with an instantaneous ECG transformation of the tachycardia after the interruption of the first loop) [<xref ref-type="bibr" rid="R1">1</xref>].</p><p>We have found in the literature 5 patients with atrial tachycardias and a remote history of isolated aortic valve replacement [<xref ref-type="bibr" rid="R2">2</xref>,<xref ref-type="bibr" rid="R5">5</xref>-<xref ref-type="bibr" rid="R7">7</xref>]. All had macroreentrant tachycardias (4 in the left atrium, 1 in the right atrium) and 1 presented a concomitant focal tachycardia. Among patients with left tachycardias, data regarding activation map and/or ablation were available only in two cases. In the first case, described by Jais et al. [<xref ref-type="bibr" rid="R2">2</xref>], although no complete electro-anatomical map was obtained, the tachycardia was successfully ablated by drawing a line between the ostium of the right superior pulmonary vein and the mitral annulus. In the second case, reported by Ouyang et al. [<xref ref-type="bibr" rid="R5">5</xref>], mapping revealed a figure-8 activation pattern with one loop situated in the anterior left atrial wall and the other one in the roof of the same atrium. Curative ablation was performed at the shared channel identified by activation and entrainment mapping.</p><p>In this report we describe a figure-8 tachycardia originating in the left anterior atrial wall in another patient with previous aortic valve replacement. The full cycle length of the tachycardia was mapped inside a low voltage zone containing electrically silent areas and multiple sites exhibiting double- or fragmented potentials. This fact suggests a severely affected anatomical substrate. In order to avoid terminating or transforming the tachycardia, as reported during the previous two EP studies in this patient, we did not perform entrainment. Ablation targeted what was regarded as the common and medial channels (including the possible channel between the medial scar and the transverse line of double potentials) and, presumably, in order to avoid other secondary circuits, it should have been continued in the lateral channel as well. Tachycardia interruption along with its noninducibility after severance of these channels would have supported our hypothesis concerning the electrophysiological mechanism and the configuration of the circuit. However the instability and difficult maneuverability of the catheter, probably explained by the anatomical position of the targeted area, precluded a transmural ablation. Additional factors contributing to this outcome might be the structural changes of the anatomical substrate and the fact that we targeted first the common and medial channels. Shah et al. [<xref ref-type="bibr" rid="R1">1</xref>]. noted that in the case of right figure-8 tachycardias, catheter stability in the region of the shared isthmus may be problematic.</p><p>This report confirms that patients with previous aortic valve replacement may have areas with abnormal electrophysiology in the left atrium that favor complex atrial tachycardias. Participation of the left anterior atrial wall in such tachycardias (as showed in the patient presented by Ouyang et al. and in this report) may be related to mechanical trauma during surgery (due to the proximity of this region to aortic root) as well as to intra- and postoperative atrial ischemia. In cases of figure-8 activation patterns, confirmation of the tachycardia mechanism and geometry of the circuit requires detailed activation mapping combined with entrainment and/or a successful ablation strategy (defined as modification/interruption of the tachycardia due to a transmural lesion with impossibility of reinitiating the same tachycardia).</p></sec> |
An Unusual Tachycardia | <p>The following article presents an unusual case of atrial tachycardia, initially misdiagnosed due to a lack of clear P waves. The diagnosis was eventually confirmed using the atrial electrogram from the patient’s pacemaker.</p> | <contrib contrib-type="author"><name><surname>Hanon</surname><given-names>Sam</given-names></name><degrees>MD</degrees></contrib><contrib contrib-type="author"><name><surname>Shapiro</surname><given-names>Michael</given-names></name><degrees>DO</degrees></contrib><contrib contrib-type="author"><name><surname>Schweitzer</surname><given-names>Paul</given-names></name><degrees>MD, FACC</degrees></contrib><aff>Division of Cardiology of Beth Israel Medical Center, University Hospital and Manhattan Campus for the Albert Einstein College of Medicine New York, NY</aff> | Indian Pacing and Electrophysiology Journal | <sec sec-type="" id="s1"><title>Case report</title><p>An eighty-four year old Dominican man presented to the Emergency Department complaining of “dizziness”. He had a past medical history significant for aortic stenosis, status post aortic valve replacement, paroxysmal atrial fibrillation and sick sinus syndrome, for which a DDD pacemaker was placed. Two days prior to admission he noted light-headedness, which began while walking.</p><p>Initial electrocardiogram (ECG) (<xref ref-type="fig" rid="F1">Figure 1</xref>), revealed a wide complex tachycardia at 170 bpm with LBBB morphology. The differential diagnosis included ventricular tachycardia with exit block, atrial tachycardia, atrial flutter, and A-V nodal reentrant tachycardia. Previous tracings had shown a preexisting LBBB.</p><p>The tachycardia terminated spontaneously but recurred. After the resolution of symptoms, the ECG was repeated (<xref ref-type="fig" rid="F2">Figure 2</xref>). This ECG showed a QRS morphology similar to the admission tracing with a ventricular rate of 85 bpm and no clear P wave activity.</p><p>Because the follow up ECG was at a rate exactly half the initial one, an underlying atrial tachycardia was suspected. The pacemaker was interrogated to assist in analysis of the rhythm (<xref ref-type="fig" rid="F3">Figure 3</xref>). The interrogation revealed atrial tachycardia with 2:1 block, with an atrial rate of 170 bpm and a ventricular rate of 85 bpm. The patient was treated with metoprolol and his symptoms resolved after normalization of his rhythm.</p></sec><sec sec-type="" id="s2"><title>Discussion</title><p>P wave size and morphology are determined by the location of the ectopic pacemaker and the sequence of atrial activation, which may be influenced by atrial disease [<xref ref-type="bibr" rid="R1">1</xref>]. As a result, the correlation between atrial activity and cardiac rhythm can be difficult to establish by surface ECG alone. Specifically, in cases of atrial tachycardia, identification of the underlying rhythm can be problematic, as P waves may be small or hidden in the QRS complex. Occasionally, as in the present case, no P waves may be detected from the ECG while the atrial electrogram demonstrates clear atrial activity. Consequently, the rhythm may be mistaken for junctional tachycardia or A-V nodal reentrant tachycardia. This case illustrates that the atrial wire can facilitate elucidation of atrial activity during an arrhythmia.</p></sec> |
Short- and long-term experience in pulmonary vein segmental ostial ablation for paroxysmal atrial fibrillation<xref ref-type="fn" rid="fn1">*</xref> | Could not extract abstract | <contrib contrib-type="author"><name><surname>Pürerfellner</surname><given-names>H</given-names></name></contrib><contrib contrib-type="author"><name><surname>Aichinger</surname><given-names>J</given-names></name></contrib><contrib contrib-type="author"><name><surname>Martinek</surname><given-names>M</given-names></name></contrib><contrib contrib-type="author"><name><surname>Nesser</surname><given-names>HJ</given-names></name></contrib><contrib contrib-type="author"><name><surname>Janssen</surname><given-names>J</given-names></name><degrees>MD</degrees><xref ref-type="aff" rid="aff1">†</xref></contrib><aff>Public Hospital Elisabethinen, Academic Teaching Hospital, Cardiological Department, Linz, Austria</aff><aff id="aff1"><label>†</label>Medtronic Inc. , Bakken Research Center, Maastricht, Holland</aff> | Indian Pacing and Electrophysiology Journal | <sec id="s1"><title>Introduction</title><p>Based on groundbreaking work by Michel Haissaguerre et al. a catheter-based procedure was recently introduced into clinical routine to treat drug-refractory paroxysmal atrial fibrillation (PAF) [<xref ref-type="bibr" rid="R1">1</xref>,<xref ref-type="bibr" rid="R2">2</xref>]. During segmental ostial pulmonary vein isolation (PVI), electrically conducting myocardial extensions bridging the pulmonary veins (PV) with the posterior wall of the left atrium (LA) are disconnected to prevent the initiation (and the maintenance) of PAF.</p><p>Success rates reported in recent literature differ to a big extent (between 50 and 90%) [<xref ref-type="bibr" rid="R3">3</xref>,<xref ref-type="bibr" rid="R4">4</xref>] and are not always comparable, as applied techniques may vary considerably and are changing with growing experience in the same group. Moreover, evaluation of success is not consistent within published reports, as clinical judgement of ablation outcome may be difficult to determine (missing ECG documentation in symptomatic periods, asymptomatic episodes). In addition, follow up periods are rarely comparable and sometimes are rather short, so that there is still a lack of data addressing the long-term outcome in ablation of the PV ostia to treat PAF nowadays.</p><p>We have reported on our early experience in the past [<xref ref-type="bibr" rid="R5">5</xref>]. In this paper we report on our own short and long-term results with special consideration of distinct procedural parameters which have been developed in a time period of three years from 2001 to 2004.</p></sec><sec sec-type="methods" id="s2"><title>Patients and Methods</title><sec id="s2a"><title>Lasso Procedure</title><p>Briefly, after puncture of one (or both) femoral and the left subclavian (in case the coronary sinus can not be entered by a femoral access) vein(s) a multipolar electrode catheter is introduced in the coronary sinus to record left atrial activity. Thereafter, a (usually double) transseptal puncture is performed followed by pulmonary venography. Next, two electrode catheters are placed in the LA via two long sheaths, one for ablation (Celsius THERMOCOOL 7F, Biosense Webster, Inc.) and one circular decapolar catheter (“Lasso”, Biosense Webster, Inc.) to record electrical activity of myocardial extensions connecting the LA with a PV. The lasso catheter is situated as proximal as possible inside the PV whereas the ablations catheter is positioned at the very os to segmentally isolate the PV, thereby preventing stenosis which may occur when delivering energy deep inside a PV. Electrical entrance block from the LA to the PV is considered as the endpoint of the procedure which is documented by the elimination of PV potentials on the Lasso catheter. We have been using irrigated ablation catheters from the very beginning of our PV ablation experience, delivering up to 30 Watts of energy in the superior PVs (left superior, LSPV and right superior, RSPV) and a maximum of 20-25 Watts in the inferior PVs (left inferior, LIPV and right inferior, RIPV). The approach of targeting a PV for ablation is “empiric” i.e. independent of its arrhythmic activity during the time of the procedure. In the time period from 2001 to the first half of 2002, 3 PVs were routinely isolated, whereas since then all 4 PVs (including the RIPV which is sometimes more difficult to reach) were targeted during a single procedure.</p></sec><sec id="s2b"><title>Follow up</title><p>In the following we report on our experience in ostial PVI in the time period between 12/2000 and 02/2004. The last follow up visit was performed in 04/2004. After the procedure, patients were monitored on an outpatient basis after one month and on an inpatient basis after 3, 6 and approximately 24 months. The following tests were routinely performed: Clinical examination, Holter-ECG, evaluation of quality of life using a standardized questionnaire, in addition (on an inpatient basis) transthoracic and transesophageal echocardiography, stress test, spiral computed tomography (CT) of the PVs, and (only in case of significant PV stenosis) magnetic resonance imaging of the PV and a lung scan. </p></sec><sec id="s2c"><title>Success criteria</title><p>Two distinct time periods were analyzed in order to evaluate the clinical outcome: a short term follow up period of 6 months (3 months for the last patients in this series) and a long-term follow up of approximately 24 months. The clinical outcome was classified as complete success (CS) if patients were without any documented arrhythmias and free of antiarrythmic medication. A partial success (PS) was defined by an absence of clinical symptoms while patients being still on an antiarrhythmic drug. These two groups comprised the patient cohort for which the clinical response rate (CRR) was calculated. A clinical failure (CF) was classified as such in the rest of patients who showed no benefit while usually still on drug.</p></sec></sec><sec id="s3"><title>Results</title><sec id="s3a"><title>Demographics</title><p>117 patients (96 male, 21 female) aged 51±11 years (range 25-73 years) were included for chronic follow up. 26 patients (22%) had an underlying heart disease (hypertension n=18, coronary artery disease n=4, diabetes n=4); 2 patients had a history of a transient ischemic cerebral attack. Patients had used a mean number of 3.1±1.4 antiarrhythmic drugs. In most instances several class I, II and class III drugs had been administered previously (propafenone 70%, betablocker 65%, amiodarone 56%, sotalol 51%, flecainide 26%); uncommonly, class IV drugs (verapamil in 18% and digoxin in 14%).</p></sec><sec id="s3b"><title>Arrhythmia characteristics</title><p>Arrhythmia history showed a wide inter-individual variation and lasted for 73±67 months (median 48 months), the maximal episode duration within the last 3 months had a median of 390 min, the mean number of arrhythmia recurrences acccounted for n=2/week. <xref ref-type="table" rid="T1">Table 1</xref> illustrates the classification of the preexisting arrhythmia as defined by the surface ECG. 115/117 patients (98%) exhibited AF as their primary clinical atrial tachyarrhythmia with 86 presenting only with PAF, 10 showing additional regularized atrial tachycardias, 14 typical atrial flutter and 5 patients presenting both types, respectively. The majority of patients (109/115, 94.7%) presented with PAF, 6/115 (5.3%) showed only persisting AF . Two patients (1.6%) exhibited nearly incessant “P on T” atrial runs suggesting a pulmonary venous origin which was reconfirmed during the ablation procedure. These 2 patients underwent pulmonary venous segmental ostial ablation and therefore were both included in this analysis.</p></sec><sec id="s3c"><title>Procedures</title><p>166 procedures (1.42/patient) in 2-4 PVs were performed in total, consisting of 117 (71%) primary, 44 (27%) secondary und 5 (2%) tertiary procedures. <xref ref-type="fig" rid="F1">Figure 1</xref> illustrates the respective numbers of primary, secondary and tertiary procedures per patient and year (of first procedure) between 2001 and 2003. In 63 patients (55.8%) a single procedure was carried out, as opposed to 45 patients (39.8%) with a second and 5 patients (4.4%) with a third ablation procedure. The duration of the primary procedure lasted for 228±55 min and was significantly longer in comparison with the secondary (206±56 min, p=0.004). The fluoroscopic time duration for the primary and secondary procedure accounted for 62±15min, and 53±20min, respectively.</p><p>In most of the patients the LSPV (96%), LIPV (96%) and the RSPV (95%) were isolated ostially, whereas the RIPV was targeted in only 55%. The reason for this is that this PV is usually more difficult to reach and was spared in our early ablation experience from 2001 to the first half of the year 2002. Consequently, the number of patients with PVI in 3PVs, 4 PVs or 2 PVs accounted for 40 % (n=47), 54% (n=63) and 6% (n=7), respectively.</p></sec><sec id="s3d"><title>Success rates</title><p>113 patients were evaluated for the determination of the clinical outcome after 6 months (after 3 months only in 8/113 patients with a limited follow up). CRR accounted for 78% (88 patients) consisting of 52% CS (59 patients), 26% PS (29 patients) und 22% CF (25 patients), respectively (<xref ref-type="fig" rid="F2">Figure 2</xref>). CS was higher in patients undergoing PVI in 4 PVs versus PVI in 3 PVs (54% versus 44%) (<xref ref-type="fig" rid="F3">Figure 3</xref>). However, the CRR was similar in both groups (77% in 4 PVs and 78% in 3 PVs, respectively).</p><p>For the determination of the long term succes rate after a mean of 21±6 months only those patients with PVI in 3 PVs (sparing the RIPV) were evaluated. For a total of 39 patients the CS, PS and CF rates accounted for 41% (n=16), 21% (n=8), 38% (n=15), respectively, leading to a CRR of 62% (n=24). Considering only this group of 39 patients for their specific clinical course after 6 months separately, CS, PS and CF rates account for 41% (n=16), 33% (n=13) and 26% (n=10), respectively, leading to a CRR of 74% (n=29) (Figure 2). From these results one may conclude that CRR in the longterm follow up is reduced by a reduction in PS, however, the CS rate seems stable even in the long run.</p><p>The clinical success rates per year of experience are depicted in <xref ref-type="fig" rid="F4">Figure 4</xref>. As patients had to have all procedures performed within a single year for this analysis, the number of patients evaluated decreases to n=103.</p></sec><sec id="s3e"><title>Complications</title><p>Significant PV stenoses are the leading cause of side effects in the longterm follow up ranging up to 7.7% (n=9), all other complications account for less than 1.5%, such as pericardial tamponade (n=2), pericardial effusion (n=2), stroke (n=2), pneumo-/hemothorax (n=2), groin hematoma (n=2) and pericarditis (n=1). In total, 20 patients (17.1%) had either a single or 2 (3 patients, 2.6%) complications independent of the number of procedures performed.</p></sec></sec><sec id="s4"><title>Discussion</title><p>We have reported repeatedly on our experience in ostial PVI recently, in particular on the occurence of PV stenosis [<xref ref-type="bibr" rid="R6">6</xref>,<xref ref-type="bibr" rid="R7">7</xref>], quality of life [<xref ref-type="bibr" rid="R8">8</xref>] and the evaluation of procedural success by use of longterm implantable monitoring provided by special implantable pacemakers with extended storage capabilities in chronically paced patients [<xref ref-type="bibr" rid="R9">9</xref>].</p><p>In this paper our data highlight the potentially curative therapeutic approach of segmental ostial PVI over a short- and longterm follow up in a patient cohort with highly symptomatic and medically drug-refractory recurrent PAF and without significant underlying heart disease. In total, 78% of patients show a clinical benefit over a follow up period of 6 months, which is reduced to 62% in the long-term after approximately 24 months in a patient cohort that underwent PVI in 3 PVs (sparing the RIPV). An ablation procedure that includes PVI in the RIPV as well exhibits a higher curative success rate after 6 months of follow up (54% versus 44%).</p><sec id="s4a"><title>Segmental ostial PVI</title><p>According to current knowledge, a segmental ostial ablation approach at the PV-LA junction represents a predominant trigger elimination of ectopic foci from within the PVs capable of initiating PAF by rapid focal discharges. To which extent an ongoing fibrillatory process in the LA may be maintained by such PV discharges remains an open issue at this point in time [<xref ref-type="bibr" rid="R10">10</xref>]. In addition, ectopic activity that may trigger PAF may also be generated at different sites in both atria (posterior and anterior LA, coronary sinus, terminal crest, superior caval vein). This may well explain why a 100% cure rate is unlikely when applying radiofrequency ablation at the ostial level of the PVs solely.</p></sec><sec id="s4b"><title>Procedural parameters</title><p>Despite the fact that the arrhythmogenic potential of the RIPV was somewhat in doubt years ago, our results are in accordance with other more recent reports in the literature demonstrating the importance of implementing this PV in the ablation procedure. As illustrated in <xref ref-type="fig" rid="F3">Figure 3</xref> the CS is higher when isolating all 4 PVs (versus 3 PVs). The RIPV is sometimes more difficult to reach technically, however, with improved catheter steerability this problems is solved most of the time nowadays. Other procedural parameters do not seem crucial in our series as our approach of isolating the PVs did not change over the years using the same diagnostic catheter (Lasso catheter) within the PV ostium and a stable energy source for ablation (radiofrequency ablation with irrigated catheters) with unchanged energy settings (30 Watts for the superior PVs and 20-25 Watts for the inferior PVs).</p></sec><sec id="s4c"><title>Longterm success rate</title><p>Our longterm success rates after approximately 24 months may be limited by the fact that these results represent the outcome in patients having undergone ostial PVI in 3 PVs exclusively. If we compare the longterm success rate of this separted group with its 6 month follow up data, it is obvious that the CS is literally idenitical (41%) whereas the PS and the CRR seem to be somewhat higher on a short term basis (33% versus 21% and 74% versus 62%, respectively). Our results demonstrating a stable cure rate over an extendend period are confirmed in a previously published larger report [<xref ref-type="bibr" rid="R11">11</xref>]. However, it remains interesting to reconfirm those data by our own series of patients having undergone a 4 PV ablation approach from the second half ot the year 2002 onwards.</p></sec><sec id="s4d"><title>Learning curve</title><p>To find an appropriate measure for our learning curve we have analyzed the number of secondary procedures and the success rates . As depicted in <xref ref-type="fig" rid="F1">Figure 1</xref> there is a constant decrease of secondary procedures per patient with growing experience: The percentage decreases from about 50% in patients with their primary procedure in the year 2001 to 40% and 25%, respectively in patients who underwent PVI in the year 2002 and 2003. The reason for this is based on the growing knowledge about the nature of early recurrences (within the first days after the primary procedure): As those recurrences may merely represent transient effects of leason healing we tended to treat those arrhythmias conservatively with growing experience by administering antiarrhythmic drugs instead of performing an early redo procedure. In addition , we were able to exactly quantify and publish the decrease of arrhythmic burden of PAF within the first 3 months after PVI in 12 patients who were previously implanted with a last generation implantable pacemaker offering extended storage capabilities (Medtronic Inc, AT 500) [<xref ref-type="bibr" rid="R9">9</xref>]. <xref ref-type="fig" rid="F4">Figure 4</xref> demonstrates the success rate based on the year of the primary procedure. It shows an increase in the CS from 40% in the year 2001 to about 50% in the next two years. The most likely cause for this is the fact that the procedure changed from a 3 PV to a 4 PV isolation approach (including the RIPV) . The CRR, however, remains fairly constant, reaching 70 to 80%.</p></sec><sec id="s4e"><title>Redo procedures</title><p>In general, concomitant atrial tachycardias or atrial flutter were not a primary target for ablation during the first ablation session. In case of a recurrence of PAF, one or more PVs were usually reconnected and had to be reisolated. In addition, sustained atrial tachycardias as well as right or left atrial flutter circuits were mapped as appropriate. However, no detailed ablation results for these substrates are available in this reported patient series.</p></sec><sec id="s4f"><title>Complications</title><p>With regard to diagnosis, management and outcome in PV stenosis follwing ostial PVI we may refer to our previously published results [<xref ref-type="bibr" rid="R6">6</xref>,<xref ref-type="bibr" rid="R7">7</xref>]. Despite the fact that there may be multiple factors operational for the occurence of PV stenosis (energy source, amount of delivered energy, localisation of the affected PV) the exact site of energy delivery at the ostial level seems to be of highest importance: The more energy delivered within a PV (versus ostial) the higher the chance of developing significant PV stenosis. In general, the symptom threshold (including dyspnea, hemoptysis and pneumonia) for significant PV narrowing exceeds 60% of luminal narrowing, however, not every single significant PV narrowing produces symptoms. Lung scanning is a useful diagnostic test to document hemodynamically significant PV stenosis by detecting a segmental perfusion deficit. Interventions in significant PV stenosis including PV dilatation and/or stenting may lead to a high reocclusion rate. In our published series of 6 patients with predominant single PV stenosis the clinical outcome is beneficial despite reocclusion in 2 patients during a longterm follow up.</p></sec><sec id="s4g"><title>Limitations</title><p>Longterm success is based on the results in patients with a 3 PV ostial isolation approach (sparing the RIPV). It may well be that chronic results in patients with a 4 PV (including the RIPV) approach are superior.</p><p>Although routine longterm ECG monitoring was performed at prespecified points in time during follow up, clinical success is primarily based on symptomatic episodes of PAF. It remains unclear who often and to what extent silent episodes of PAF may have changed the clinical results of PVI in our patient subgroups.</p></sec><sec id="s4h"><title>Future outlook</title><p>According to many other working groups worldwide we have changed our ablation strategy recently: Using a 3 dimensional mapping system (CARTO, Biosense Webster Inc.) an electroanatomic map of the LA including the mitral valve annulus and the PVs is reconstructed pre ablation. Thereafter, periostial circumferential atrial lesions are deployed around the septal and lateral PVs in order to modify the substrate of PAF (<xref ref-type="fig" rid="F5">Figure 5</xref>). In addition, an ablation line in between the LIPV and the mitral valve is drawn (so called "mitral isthmus"), which is at times accompanied by a roof line in the LA connecting the superior PVs. It seems critical to attain complete lesion sets designed to prevent the spread of focal discharges from withhin the PV to the rest of the LA (Pappone approach [<xref ref-type="bibr" rid="R12">12</xref>]). At this point in time we are still using a multipolar circular catheter to be located at the ostial level to monitor and guide PV isolation at the LA-PV junction. Recently published reports demonstrate higher success rates using circumferential PV atrial ablation in contrast to segmental ostial PVI both in PAF [<xref ref-type="bibr" rid="R13">13</xref>] and (even more) in persistent to permanent forms of AF. An additional advantage of this technique is based on the diminished occurence of PV stenoses by applying energy >1cm away from the ostial level. However, additional point lesions at the ostial level of a PV are often required to accomplish complete LA-PV block.</p><p>In the meantime a multislice CT imaging modality is available in our institution which enables the 3 dimensional representation of a patient’s individual cardiac anatomy (<xref ref-type="fig" rid="F6">Figure 6</xref>). In addition, worldwide intensive collaboration within different manufacturers of cardiac visualization systems is ongoing to enable "3 D image integration" which will allow a direct transfer of imaging files to the mapping system at work. By this, a direct transfer of the individual electrical activation to the individual patient’s anatomy seems feasible ("image guiding").</p><p>Moreover, interesting and promising efforts are made by one manufacturer (Stereotaxis, Inc) to conduct an ablation procedure by remote navigation of the ablation catheter which is directed by a strong magnetic field around the patient’s body.</p></sec></sec><sec id="s5"><title>Conclusion</title><p>The CRR in patients with medically refractory PAF undergoing ostial PVI accounts for 78% after a short term follow up of 6 months. A 4 PV procedure (including the RIPV) warrants a higher CS. In patients with a 3 PV procedure (sparing the RIPV) the CRR in the longterm follow up is reduced to 62%. However, the CS rate seems to be constant, whereas the PS is reduced. PV stenosis is the main cause for complications. In addition to an exact analysis of procedural parameters a longer follow up period seems necessary to better quantify the clinical benefit to risk ratio of segmental ostial PVI in contrast to other strategies in the treatment of atrial fibrillation.</p></sec> |
Predictors of appropriate ICD therapy in patients with implantable cardioverter-defibrillator | Could not extract abstract | <contrib contrib-type="author"><name><surname>Dehghani</surname><given-names>Mohammad Reza</given-names></name><degrees>MD</degrees><xref ref-type="author-notes" rid="fn1">*</xref></contrib><contrib contrib-type="author"><name><surname>Arya</surname><given-names>Arash</given-names></name><degrees>MD</degrees><xref ref-type="author-notes" rid="fn1">*</xref></contrib><contrib contrib-type="author"><name><surname>Haghjoo</surname><given-names>Majid</given-names></name><degrees>MD</degrees></contrib><contrib contrib-type="author"><name><surname>Emkanjoo</surname><given-names>Zahra</given-names></name><degrees>MD</degrees></contrib><contrib contrib-type="author"><name><surname>Alasti</surname><given-names>Mohammad</given-names></name><degrees>MD</degrees></contrib><contrib contrib-type="author"><name><surname>Kazemi</surname><given-names>Babak</given-names></name><degrees>MD</degrees></contrib><contrib contrib-type="author"><name><surname>Nikoo</surname><given-names>Mohammad Hosein</given-names></name><degrees>MD</degrees></contrib><contrib contrib-type="author"><name><surname>Sadr-Ameli</surname><given-names>Mohammad Ali</given-names></name><degrees>MD</degrees></contrib> | Indian Pacing and Electrophysiology Journal | <sec id="s1"><title>Introduction</title><p>Development of implantable cardioverter-defibrillators (ICD) has been a dramatic advancement in the management of patients with ventricular arrhythmias. A major issue in patients with ICD is the high incidence of ICD therapies[<xref ref-type="bibr" rid="R1">1</xref>-<xref ref-type="bibr" rid="R7">7</xref>], which have a major effect on morbidity and quality of life [<xref ref-type="bibr" rid="R4">4</xref>,<xref ref-type="bibr" rid="R5">5</xref>]. Knowing the predictors of appropriate ICD therapy could also help to better identify candidates for ICD therapy. This is a retrospective single centre study to identify potential predictors of appropriate ICD therapies.</p></sec><sec sec-type="methods" id="s2"><title>Methods</title><sec id="s2a"><title>Patients Population</title><p>Between January 2001 and January 2005, 162 patients with coronary artery disease (CAD) or dilated cardiomyopathy (DCM) underwent ICD implantation at our centre. Among these 94 (58%) received single-chamber and 68 (42%) received dual-chamber ICD. The left ventricular ejection fraction (EF) was measured by transthoracic echocardiography. All the patients gave written informed consent for the procedure of ICD implantation. The mean age was 58.2 ± 13.5 years. <xref ref-type="table" rid="T1">Table 1</xref> and <xref ref-type="table" rid="T2">2</xref> show the basic characteristics of the patients.</p></sec><sec id="s2b"><title>Implanted ICDs and programming</title><p>The implanted devices were manufactured by Medtronic ([GEM-VR, GEM-DR, GEM-II-VR, GEM-II-DR, GEM-III-VR, GEM-III-DR, Marquis-VR, Marquis-DR] Medtronic Inc., Minneapolis, MN, USA) and St. Jude ([Photon-VR, Photon-DR, Photon-μ-VR, Photon-μ-DR, Atlas-VR, Atlas-DR, Epic-VR, and Epic-DR] St. Jude Medical Inc. Sylmar, CA, USA). In implanted devices all the detection and discrimination criteria were activated with the nominal values. In all the devices we defined ventricular fibrillation zone (300 ms) plus one VT zone (400 ms). If the patient had an episode of spontaneous or induced sustained monomorphic VT slower than 370 ms we extended the VT zone to VT cycle length plus 40ms. In the VT detection zone the first therapy was three antitachycardia bursts pacing. We used the nominal values of the ICDs for the duration and tachyarrhythmia detection criteria.</p></sec><sec id="s2c"><title>ICD Data Storage and Retrieval</title><p>After ICD implantation the patients were followed on a regular basis (3 months) and upon receiving high voltage therapy in our outpatient ICD clinic. The devices were interrogated at each session and the complete set of data (including intracardiac electrograms) was recorded on floppy diskettes. The summary of the episodes were also recorded in the patient’s file. The floppy diskettes were used in this study to retrieve all spontaneous sustained arrhythmia episodes resulted in ICD therapy. These episodes resulted in ICD therapies, studied by two independent electrophysiologists (AA and MRD) to define the diagnosis. In case of discrepancy in diagnosis the final analysis of the arrhythmia episode was made by a consensus of three electrophysiologists (AA, MRD, and MH). Beside from diagnosis, the time of arrhythmia after implantation and the mode of therapy were recorded.</p></sec><sec id="s2d"><title>Definitions</title><p>Appropriate ICD therapy was defined as an antitachycardia pacing or shock therapy for ventricular tachycardia or fibrillation. Indication of ICD implantation was defined as secondary prevention (n=117) in patients who had experienced aborted sudden cardiac death, sustained ventricular arrhythmia, or syncope (whose electrophysiologic study [using three basic drive cycle lengths of 600, 500 and 400 with up to three premature extra-stimuli from right ventricular apex and/or outflow tract] who showed inducible sustained hemodynamically unstable ventricular arrhythmias). The indication of ICD implantation in all the other patients (33 patients with CAD without history of syncope had left ventricular ejection fraction < 40%, nonsustained ventricular tachycardia on Holter monitoring, and inducible sustained hemodynamically unstable ventricular arrhythmia during electrophysiologic study; and 13 asymptomatic patients with DCM with nonsustained ventricular tachycardia during Holter monitoring who had inducible sustained hemodynamically unstable ventricular arrhythmia during electrophysiologic study) was categorized as primary prevention [<xref ref-type="bibr" rid="R6">6</xref>].</p></sec><sec id="s2e"><title>Statistics</title><p>Variables are expressed as mean ± SD, and percentage. Differences in frequency of characteristics were assessed by independent sample student’s t-test for continuous variables. Chi-square statistics (or fisher’s exact test if applicable) used for discrete variables. We used binary logistic regression analysis with forward selection method to find the potential predictors of appropriate ICD therapy after device implantation. A two-tailed p-value < 0.05 was considered statistically significant. We used SPSS® 13.0 (SPSS Inc. Chicago, IL, USA) for data storage and analysis.</p></sec></sec><sec id="s3"><title>Results</title><sec id="s3a"><title>Baseline Characteristics</title><p>One hundred sixty two patients (123 men) with ICD were followed for a mean of 15±11 months. <xref ref-type="table" rid="T1">Table 1</xref> and <xref ref-type="table" rid="T2">2</xref> show the baseline characteristics of these patients. We compared patients’ characteristics between different underlying diseases (<xref ref-type="table" rid="T1">Table 1</xref>). We took into account these differences in subsequent statistical analysis to find potential predictors of appropriate ICD therapy. Among 49 patients who received ICD as primary prevention, 20 (41%) received appropriate ICD therapy. Among 113 patients who received ICD as a secondary preventive measure 34 (30%) received appropriate ICD therapy (unadjusted odds ratio [OR] = 1.74, 95% confidence interval [CI]: 0.8 - 3.6, Mantel-Haenszel P value P=0.14). We adjusted this analysis for left ventricular EF, QRS width, gender, and underlying heart disease which resulted in adjusted OR = 1.66, 95% CI = 0.7 – 4.0, Mantel-Haenszel P value P=0.355. During the same period 30.2% (n = 49) of patients received inappropriate ICD therapy. Twenty seven out of 54 patients (50%) who received appropriate ICD therapy received also inappropriate ICD therapy. The rate of inappropriate ICD therapy among patients who did not receive appropriate ICD therapy (n = 108) was 20.4%. Atrial fibrillation/tachycardia, sinus tachycardia, and oversensing were the most common causes of inappropriate ICD therapy in our patients.</p></sec><sec id="s3b"><title>Predictors of ≥ 1 Appropriate ICD therapy</title><p>During the follow up period, 54 patients (33%) received appropriate ICD therapy (<xref ref-type="table" rid="T2">Table 2</xref>). There were several differences between patients with and without ≥ 1 appropriate ICD therapy. We used these variables as covariates to find the predictors of appropriate ICD therapy during the follow up period. We also included all the other parameters which showed a P<0.3 during bi-variable correlation with probability of ≥ 1 appropriate therapy a binary logistic regression analysis model.</p><p>During binary logistic regression analysis QRS width > 100 ms (P=0.003), male sex (P = 0.021), and DCM as underlying heart disease (P = 0.001) were correlated with ≥ 1 appropriate ICD therapy in all patients (<xref ref-type="table" rid="T3">Table 3</xref>). We chose the subgroups of left ventricular ejection fraction and QRS duration for this analysis based on the median values in the study population. Other factors including age, left ventricular EF, baseline medical therapy (including amiodarone), and indication of ICD implantation failed to correlate with the probability of ≥ 1 appropriate ICD therapy during the follow up period (all Ps > 0.05).</p></sec><sec id="s3c"><title>Number of Appropriate ICD therapies</title><p>During our follow up period the above mentioned patients received mean number of 17±29 (range 1 – 132) appropriate ICD therapy. Among these the number of appropriate ATP was 11.9±28 (range 0 – 131) and the number of appropriate shock therapy was 5.1±9.9 (range 1 – 56). The success rate of ATP therapy among the episodes in the VT detection zone was 88%. However, as our study is retrospective it is possible that after first appropriate ICD therapy, electrical and/or medical treatment was adjusted or optimized to prevent new VT recurrences. Therefore, to minimize this effect we just assessed the predictors of ≥ 1 appropriate ICD therapy rather than total <italic>number</italic> of appropriate ICD therapies.</p><p>During subgroup analysis in patients with CAD and DCM, QRS duration > 100 ms was correlated with the probability of ≥ 1 appropriate ICD therapy during binary logistic regression (P = 0.006 and P = 0.003, respectively).</p></sec></sec><sec id="s4"><title>Discussion</title><p>The main findings of this retrospective study are 1) QRS width might be a useful risk stratifier (beyond ejection fraction) in patients with CAD and DCM; 2) adjunctive amiodarone therapy in our patients failed to decrease the incidence of ≥ 1 appropriate ICD therapy; 3) indication of ICD implantation did not influenced the probability of ≥ 1 appropriate ICD therapy during the follow up period. Several points merit consideration. The rate of appropriate ICD therapy was different in our patients compared to the other studies [<xref ref-type="bibr" rid="R1">1</xref>,<xref ref-type="bibr" rid="R4">4</xref>,<xref ref-type="bibr" rid="R7">7</xref>]. This difference can be at least partially explained by the different patients population and the follow up period.</p><p>No empiric antiarrhythmic therapy (including amiodarone) is <italic>currently</italic> indicated in patients who received an ICD. These patients frequently receive ICD shocks, which severely impair quality of life. Intravenous amiodarone followed by oral dose, in patients with electrical storm results in successful management of ventricular arrhythmias and possibly a long-term prognosis similar to patients who do not have electrical storm [<xref ref-type="bibr" rid="R9">9</xref>,<xref ref-type="bibr" rid="R10">10</xref>]. The OPTIC (Optimal Pharmacological Therapy in Implantable Cardioverter) study currently assesses the potential benefit of antiarrhythmic medications in reduction of ICD therapy and electrical storm. In OPTIC the patients are randomized to β-blocker, amiodarone plus β-blocker, or sotalol [<xref ref-type="bibr" rid="R9">9</xref>]. Amiodarone may have some other potential benefits in patients with ICDs including the prevention of supraventricular tachycardia, and so may decrease inappropriate ICD discharges; and the prevention of nonsustained but symptomatic ventricular arrhythmias. Further studies are warranted to clarify these issues [<xref ref-type="bibr" rid="R11">11</xref>].</p><p>Patients with ICD who received it for primary prevention of sudden cardiac death received frequent appropriate ICD therapy. Among our patients the probability of ≥ 1 appropriate ICD therapy among those who received ICD for primary and secondary prevention was comparable. Although this may be partially explained by our restricted criteria of ICD implantation for primary prevention (see definition) it also highlights the importance of implementing guidelines of primary prevention of sudden cardiac death in daily clinical practice. However, it should be mentioned that practically for the majority of the countries (especially the developing countries) and health care providers it is simply economically impossible to follow all the indications derived from the recent primary prevention ICD-trials in which the major risk stratifier is EF [<xref ref-type="bibr" rid="R12">12</xref>]. Therefore, further risk stratification beyond EF is highly desirable and necessary [<xref ref-type="bibr" rid="R12">12</xref>].</p><p>Several epidemiological studies showed that QRS duration (both its absolute value and its dynamic changes) predicts sudden and all-cause mortality [<xref ref-type="bibr" rid="R13">13</xref>-<xref ref-type="bibr" rid="R16">16</xref>]. These studies have shown that the rate of sudden and non-sudden cardiac mortality increases sharply as QRS duration rises above 120 ms. This effect is observed both in patients with and without bundle branch block. Our finding also showed that QRS width ≥ 100 ms (median value in our study population) is associated with increased incidence of appropriate ICD therapy (<xref ref-type="table" rid="T3">Table 3</xref>). However, a recent comparison of QRS duration with microvolt T wave alternans showed a false negative rate of narrow QRS complex (10.2%) among MADIT II like patients which could limit its value as a risk predictorc [<xref ref-type="bibr" rid="R17">17</xref>].</p><p>Finally, only 57% of our patients received beta-blocker therapy. There is substantial evidence that beta blocker therapy has positive effects on morbidity, and mortality, in patients who have been diagnosed with heart failure and/or CAD. Beta blockers should be considered a cornerstone of therapy for these patients [<xref ref-type="bibr" rid="R18">18</xref>-<xref ref-type="bibr" rid="R20">20</xref>]. Therefore although in our study beta-blocker therapy failed to reduce the incidence of appropriate ICD therapy, beta-blockers should be administered to all patients with CAD and DCM who have ICD unless an absolute contraindication is present. This will decrease the morbidity and mortality in these patients as the beneficial effect of beta-blockers in these patients is additive to the effect of ICD.</p><sec id="s4a"><title>Study limitation</title><p>Although we showed that (1) the adjunctive amiodarone therapy do not reduce the incidence of ≥ 1 appropriate ICD therapy and (2) beta-blockers failed also to reduce it, our study was retrospective and non-randomized. Before making any conclusion from our data we have to wait for results of randomized studies such as OPTIC.</p></sec></sec> |
Pharmacological cardioversion for atrial fibrillation and flutter<xref ref-type="fn" rid="fn1">*</xref> | Could not extract abstract | <contrib contrib-type="author"><name><surname>Cordina</surname><given-names>J</given-names></name></contrib><contrib contrib-type="author"><name><surname>Mead</surname><given-names>G</given-names></name></contrib> | Indian Pacing and Electrophysiology Journal | <p>Atrial fibrillation is the commonest cardiac dysrhythmia. It is associated with significant morbidity and mortality. There are two approaches to the management of atrial fibrillation: controlling the ventricular rate or converting to sinus rhythm in the expectation that this would abolish its adverse effects.</p><p>The objective of this review was to assess the effects of pharmacological cardioversion of atrial fibrillation in adults on the annual risk of stroke, peripheral embolism, and mortality.</p><p>We made a thorough search for existing evidence in the following databases: the Cochrane Controlled Trials Register (Issue 3, 2002), MEDLINE (2000 to 2002), EMBASE (1998 to 2002), CINAHL (1982 to 2002), Web of Science (1981 to 2002). We also handsearched the following journals: Circulation (1997 to 2002), Heart (1997 to 2002), European Heart Journal (1997-2002), Journal of the American College of Cardiology (1997-2002) and selected abstracts published on the web site of the North American Society of Pacing and Electrophysiology (2001, 2002). We selected trials based on the following criteria: randomised controlled trials or controlled clinical trials of pharmacological cardioversion versus rate control in adults (>18 years) with acute, paroxysmal or sustained atrial fibrillation or atrial flutter, of any duration and of any aetiology.</p><p>We identified two completed studies AFFIRM (n=4060) and PIAF (n=252). We found no difference in mortality between rhythm control and rate control - relative risk 1.14 (95% confidence interval 1.00 to 1.31). Both studies show significantly higher rates of hospitalisation and adverse events in the rhythm control group and no difference in quality of life between the two treatment groups. In AFFIRM there was a similar incidence of ischaemic stroke, bleeding and systemic embolism in the two groups. Certain malignant dysrhythmias were significantly more likely to occur in the rhythm control group. There were similar scores of cognitive assessment in both groups. In PIAF, cardioverted patients enjoyed an improved exercise tolerance but there was no overall benefit in terms of symptom control or quality of life.</p><p>There is no evidence that pharmacological cardioversion of atrial fibrillation to sinus rhythm is superior to rate control. Rhythm control is associated with more adverse effects and increased hospitalisation. It does not reduce the risk of stroke. These conclusions cannot be generalised to all people with atrial fibrillation as most of the patients included in these studies were relatively older (>60 years) with significant cardiovascular risk factors.</p> |