Author's response to reviews - PDF Free Download

Author's response to reviews Title: Abdominal obesity vs general obesity for identifying arterial stiffness, subclinical atherosclerosis and wave reflection in healthy, diabetics and hypertensive. Authors: JOSE I RECIO-RODRIGUEZ JIR (donrecio@gmail.com) MANUEL A GOMEZ-MARCOS MAG (magomez@usal.es) MARIA C PATINO-ALONSO MCP (carpatino@usal.es) CRISTINA AGUDO-CONDE CA (cagudoconde@yahoo.es) EMILIANO RODRIGUEZ-SANCHEZ ER (emilianorodriguezsanchez@yahoo.es) LUIS GARCIA-ORTIZ LG (Lgarciao@usal.es) On behalf of the Vasorisk group VG (lgarcia@ciclorisk.com) Version: 6 Date: 28 December 2011 Author's response to reviews: see over

Dear Editor of the BMC Cardiovascular Disorders: Thank you for your help in reviewing this manuscript for consideration of publication in the BMC Cardiovascular Disorders. Following the suggestions of the reviewers, we enclose a new version of our manuscript entitled: Abdominal obesity vs general obesity for identifying arterial stiffness, subclinical atherosclerosis and wave reflection in healthy, diabetics and hypertensive. MS: 1794194876273877, together with replies to all the issues raised. GENERAL COMMENT: 1. All the changes made in the manuscript (text, tables and figures) are underlined. 2. Following the suggestions of the reviewers, we think that the manuscript has improved in terms of understanding and clarity. We therefore think that its interest has increased considerably. 3. The current version of the manuscript has been fully checked by a native English speaker with expertise in scientific translations. A certificate of quality is enclosed. We look forward to hearing from you. If you have any additional request or need any information, please contact us. Sincerely: José Ignacio Recio Rodríguez, Primary Care Research Unit, La Alamedilla Health Center Avda. Comuneros 27, 37003 - Salamanca, Spain. Tel:+34 923 124465; fax +34 923 123644; e-mail: donrecio@gmail.com 1

Reviewer: Ryuichi Kawamoto (1) Major critiques Point 1: The major problem of this study is its descriptive nature where conclusions are based essentially only on correlations. Following the recommendations of the reviewer, we have modified the conclusions, which now reads as follows: Based on the results obtained in our study, it can be concluded that, independently of the presence of diabetes or hypertension, the measures of abdominal obesity (WHtR and WC) correlates better than BMI and BFP with arterial stiffness and subclinical atherosclerosis evaluated by PWV and C-IMT respectively. However, the association with wave reflection (evaluated by CAIx) was better with general adiposity measures. In addition, we also highlighted in the discussion section the main limitation of the study, and the need for further studies designed to resolve the issue. The discussion section now reads as follows: The main limitation of our study is its cross-sectional design, which precludes the definition of causality and analysis of the behavior over time of the different stiffness parameters in relation to the different anthropometric indices considered. On the other hand, with a study design of this kind it would also be necessary to increase the sample size, since a larger number of individuals in each group (healthy, diabetic and hypertensive) could help clarify which anthropometric parameters are best related to arterial stiffness in each of them. INTRODUCTION Point 2: Several other studies described the association between abdominal obesity and cardiovascular disease. Why is this study important and needed?. As noted by the reviewer, other studies described the association between abdominal obesity and measures of arterial stiffness. However, we found no study describing if the association is different in healthy subjects, hypertensives and diabetics. For this reason, the following is included in the introduction: However, to our knowledge, no studies have examined whether this relationship differs in healthy subjects, diabetics and hypertensive individuals. Point 3: Is intima-media thickness of the common carotid artery (C-IMT) not evaluated as arterial stiffness? Enlarged C-IMT is thought to be evaluated as morphological abnormality rather than arterial stiffness as arterial dysfunction.. As indicated by both reviewers, the intima media thickness (IMT) is a measure of subclinical atherosclerosis. The only measure of arterial stiffness used in the article is pulse wave velocity. Other markers cannot be used interchangeably for the assessment of arterial stiffness. We have included the following reference: Arterial stiffness in relation to subclinical atherosclerosis. Wykretowicz A, Gerstenberger P, Guzik P, Milewska A, Krauze T, Adamska K, Rutkowska A, Wysocki H. Eur J Clin Invest. 2009 Jan;39(1):11-6. PMID: 19087126 2

We have rewritten the title, and now reads as follows: Abdominal obesity vs general obesity for identifying arterial stiffness, subclinical atherosclerosis and wave reflection in healthy, diabetics and hypertensive. We have changed the introduction and now reads as follows: The vascular structure and function can be assessed through the indices of subclinical atehrosclerosis, arterial stiffness and wave reflection [1]. A relationship has been found between measures that assess excess body weight or obesity to certain parameters that measure arterial stiffness and subclinical atherosclerosis, such as the pulse wave velocity (PWV) and the intima-media thickness of the common carotid artery (C-IMT), though their correlation to the augmentation index is not clear[2-4]. We have rewritten the aim of the study: The present study explores the relationship between anthropometric indices that assess abdominal obesity (WC, WHtR) and general obesity (BMI and body fat percentage (BFP)), with parameters that measure arterial stiffness (PWV, central and peripheral pulse pressure), subclinical atherosclerosis (C-IMT and and ankle-brachial index (ABI)) and wave reflection (central augmentation index) in healthy, diabetics and hypertensive subjects. We have changed the results section and now reads as follows: Table 1 shows the clinical and demographic characteristics of the 305 subjects included in the study (diabetics 32.8%, hypertensive subjects 37.0% and healthy individuals 30.2%), together with the anthropometric measures, inflammatory markers, central and peripheral arterial pressure, measures of arterial stiffness, subclinical atherosclerosis and wave reflection. We have rewritten the first paragraph of the discussion section, and now reads as follows: The present study shows that the measures of abdominal obesity (WHtR and WC) are better than the general obesity indicators (BMI and BFP) in predicting arterial stiffness as evaluated by PWV and subclinical atherosclerosis evaluated by C-IMT, independently of the presence of diabetes or hypertension. Therefore, WC and WHtR are the most useful parameters for estimating arterial stiffness in clinical practice. We have modified the conclusions, which now reads as follows: Based on the results obtained in our study, it can be concluded that, independently of the presence of diabetes or hypertension, the measures of abdominal obesity (WHtR and WC) correlates better than BMI and BFP with arterial stiffness and subclinical atherosclerosis evaluated by PWV and C-IMT respectively. However, the association with wave reflection (evaluated by CAIx) was better with general adiposity measures. We have changed the abstract (background section), and now reads as follows: Our aim was to analyze the relationship between abdominal obesity and general obesity, with subclinical atherosclerosis, arterial stiffness and wave reflection in healthy, diabetics and hypertensive subjects. We have changed the abstract (conclusions section), and now reads as follows: 3

The measures of abdominal obesity (WHtR and WC) correlates better than BMI and BFP with arterial stiffness evaluated by PWV, and with subclinical atherosclerosis evaluated by C-IMT, independently of the presence of diabetes or hypertension. METHODS Point 4: Were study participants randomly selected? However, the selection of the participants is not clearly indicated. Authors should clearly describe this point. What kind of reason did this study participants consult for?. This is a consecutive sampling. Patients were selected consecutively when they went to their family doctor for a routine visit. In the methods section, we added the following: A cross-sectional study was performed in a primary care setting. We consecutively included all the hypertensive, diabetics and healthy patients, that visited their family doctor, aged 20-75 years, from January 2010 to January 2011. After dealing with the reason for consultation, the patients were referred to the research unit for the assessment of cardiovascular risk. Point 5: How did authors measure office blood pressure? Blood pressure should be measured in the right upper arm of participants in a sedentary position while they were seated after having rested for at least 5 min. Was the appropriate cuff size used for obese individuals?. We have rewritten the methods section, in relation to the measurement of the office blood pressure, and now reads as follows: Office or clinical blood pressure was measured involving three measurements of systolic and diastolic blood pressure, using the average of the last two measurements, with a validated OMRON model M7 sphygmomanometer (Omron Health Care, Kyoto, Japan). Measures were taken in the right upper arm of participants in a sedentary position while they were seated after having rested for at least 5 min. with an appropriately sized cuff based on the measurement of arm circumference and following the recommendations of the European Society of Hypertension [5]. Point 6: In methods, authors should describe about measurements such as serum glucose, total cholesterol, triglycerides, HDL-cholesterol, LDL-cholesterol, and HOMA-IR.. Following the recommendations of the reviewer, we have modified the methods section, which now reads as follows: Blood samples were collected in the morning, after patient fasting 8 h. Basal glucose, HDL cholesterol, LDL cholesterol, total cholesterol, triglycerides and fibrinogen were determined. High sensitivity C-reactive protein was determined by the nephelometric method (Beckman Instrument APS; Beckman Coulter Inc., Fullerton, CA, USA) [6]. The parameters were measured on a blind basis in the Hospital Biochemistry laboratory using standard automatized techniques. The HOMA index was calculated as fasting insulin concentration (µu/ml) fasting glucose concentration (mmol/l)/22.5. 4

Point 7: I thick that glucose, TG, hscrp, and HOMA-IR are parameters with non-normal distributions, and then the data are shown as median (interquartile range). Also In all analyses, parameters with non-normal distributions were used after log-transformation. As noted by the reviewer, glucose, triglyceride, hscrp and HOMA-IR are parameters with non-normal distributions in our study. For this reason, we have changed the statistical analysis section, and now reads as follows: Non-normally distributed variables were evaluated by the Kruskal-Wallis test, and they were log transformed for further analysis. Because of their skewed distribution, glucose, triglyceride, hscrp and HOMA were presented as median and interquartile range. Pearson s or Spearman correlation coefficient were used to estimate the relationship between the quantitative variables, while the chi-square test was used to associate the qualitative variables. We have changed results in table 1 and table 2 in relation to the non-normal distributions variables. Point 8: In Table 3, multivariate adjustment should be performed in a model including office SBP (DBP), the presence of antihypertensive medication, glucose, presence of anti-diabetic medication, HDL-C, TG, LDL-C, and hscrp as well as gender, age, presence of anti-lipidemic medication, smoker, anthropometric measurements such as BMI, waist circumference, %Fat, and WHtR.. We have added a fourth model in the multivariate adjustment including systolic blood presure, total cholesterol, serum glucose, antidiabetic drugs, atherosclerotic plaques and hscrp, in addition of the rest of variables of the third model. We have included total cholesterol but not HDL-C, LDL-C and TG to avoid the colinearity. We have included in the statistical analysis section the following in relation of the new multivariate adjustment....and finally in the fourth model: systolic blood presure, total cholesterol, high sensitivity C-reactive protein, serum glucose, antidiabetic drugs and atherosclerotic plaques. We have added in table 3 (multivariate adjustment) the new model including systolic blood presure, total cholesterol, serum glucose, antidiabetic drugs, atherosclerotic plaques and high sensitivity C-reactive protein, in addition of the rest of variables of the third model. We have changed the results section and now reads as follows: After adjusting for age, gender and cardiovascular risk factors, the multiple regression analysis found that for every 0.1 point increase in WHtR, the PWV increased 0.041 m/sec and C-IMT increased 0.001 mm. Likewise, a 1 cm increase in WC implied an increase of 0.029 m/sec in PWV, 0.001 mm in C-IMT, and a decrease of 0.141 in CAIx. A 1 kg/m² increase in BMI in turn implied an increase of 0.052 m/sec in PWV, and a decrease of 0.270 in CAIx. Lastly, every 1% increase in BFP implied and a decrease of 0.284 in CAIx (Table 3). 5

We have changed the abstract (results section), and now reads as follows: WC and WHtR showed a positive correlation to PWV and C-IMT in the studied groups. After adjusting for age, gender, high sensitivity c-reactive protein, serum glucose and the presence of diabetes, hypertension, smoking, dyslipidemia, antidiabetic drugs, lipidlowering drugs, and atherosclerotic plaques, it was seen that for every 0.1 point increase in WHtR, and for every cm increase in WC, the PWV increased 0.041 and 0.029 m/sec, and C-IMT increased 0.001 mm and 0.001 mm, respectively. RESULTS Point 9: How about the medication, especially anti-hypertensive drugs, anti dyslipidemic drug, and anti-diabetic drug? The author should describe the prevalence of subjects with antihypertensive, anti-lipidemic, and anti-diabetic medication also in Table 1. The author should describe the prevalence of subjects with cardiovascular disease also in Table 1.. We have included in table 1 the number and percentage of subjects in each group on antihypertensive, antilipidemic and antidiabetic medication and the number of subjects with carotid plaque. The patients with cardiovascular disease are not included. We added the following in the methods section: Exclusion criteria were: patients with intermittent claudication, and previous cardiovascular events, patients unable to comply with the protocol requirements (psychological and/or cognitive disorders, failure to cooperate, educational limitations and problems for understanding written language, failure to sign the informed consent document), patients participating or who will participate in a clinical trial during the study. Point 10: In Table 3, multivariate adjustment for carotid plaque should also be performed in a model including office SBP (DBP), the presence of antihypertensive medication, glucose, presence of anti-diabetic medication, HDL-C, TG, LDL-C, and hscrp as well as gender, age, presence of anti-lipidemic medication, smoker, anthropometric measurements such as BMI, waist circumference, %Fat, and WHtR because these measurements are also important risk factors of carotid IMT and arterial stiffness.. We have included in table 1 the number and percentage of subjects in each group on antihypertensive, antilipidemic and antidiabetic medication and the number of subjects with carotid plaque. We have added a fourth model in the multivariate adjustment including systolic blood presure, total cholesterol, serum glucose, antidiabetic drugs, atherosclerotic plaques and hscrp, in addition of the rest of variables of the third model. We have included total cholesterol but not HDL-C, LDL-C and TG to avoid the colinearity. We have included in the statistical analysis section the following in relation of the new multivariate adjustment. 6

...and finally in the fourth model: systolic blood presure, total cholesterol, high sensitivity C-reactive protein, serum glucose, antidiabetic drugs and atherosclerotic plaques. We have added in table 3 (multivariate adjustment) the new model including systolic blood presure, total cholesterol, serum glucose, antidiabetic drugs, atherosclerotic plaques and high sensitivity C-reactive protein, in addition of the rest of variables of the third model. We have changed the results section and now reads as follows: After adjusting for age, gender and cardiovascular risk factors, the multiple regression analysis found that for every 0.1 point increase in WHtR, the PWV increased 0.041 m/sec and C-IMT increased 0.001 mm. Likewise, a 1 cm increase in WC implied an increase of 0.029 m/sec in PWV, 0.001 mm in C-IMT, and a decrease of 0.141 in CAIx. A 1 kg/m² increase in BMI in turn implied an increase of 0.052 m/sec in PWV, and a decrease of 0.270 in CAIx. Lastly, every 1% increase in BFP implied and a decrease of 0.284 in CAIx (Table 3). We have changed the abstract (results section), and now reads as follows: WC and WHtR showed a positive correlation to PWV and C-IMT in the studied groups. After adjusting for age, gender, high sensitivity c-reactive protein, serum glucose and the presence of diabetes, hypertension, smoking, dyslipidemia, antidiabetic drugs, lipidlowering drugs, and atherosclerotic plaques, it was seen that for every 0.1 point increase in WHtR, and for every cm increase in WC, the PWV increased 0.041 and 0.029 m/sec, and C-IMT increased 0.001 mm and 0.001 mm, respectively. Point 11: The authors discuss sufficiently the role of waist circumference, %Fat, and WHtR in increasing the cardiovascular risk and its association with other risk factors.. We do not understand what the reviewer means when he writes: The authors discuss sufficiently the role of waist circumference, %Fat, and WHtR in increasing the cardiovascular risk and its association with other risk factors. We assume that it is a mistake because we do not discuss in the manuscript at any time about the association with the different anthropometric indices and other risk factors. Anyway, we have added in the discussion section the following about the relationship about the different anthropometric indices and other cardiovascular risk factors. Since the arterial stiffness is a surrogate marker between the appearance of risk factors and cardiovascular outcomes, our findings support the importance of anthropometric measures as a risk factors [7]. Point 12: Some limitations of this study must be considered. First, based on its cross-sectional study design, the present findings are inherently limited in the ability to eliminate causal relationships between uric acid and metabolic syndrome. Second, since some of the study population had several risk factors including hypertension, and dyslipidemia, the authors could not eliminate the possible effect of underlying diseases and medications used for these diseases on the present findings. Further prospective population-based studies are needed to investigate the mechanisms in order to answer these questions.. 7

The major problem of this study is its descriptive nature where conclusions are based essentially only on correlations. Following the recommendations of the reviewer, we have modified the conclusions, which now reads as follows: Based on the results obtained in our study, it can be concluded that, independently of the presence of diabetes or hypertension, the measures of abdominal obesity (WHtR and WC) correlates better than BMI and BFP with arterial stiffness and subclinical atherosclerosis evaluated by PWV and C-IMT respectively. However, the association with wave reflection (evaluated by CAIx) was better with general adiposity measures. In addition, we also highlighted in the discussion section, the main limitation of the study, and the need for further studies designed to resolve the issue. The discussion section now reads as follows: The main limitation of our study is its cross-sectional design, which precludes the definition of causality and analysis of the behavior over time of the different stiffness parameters in relation to the different anthropometric indices considered. On the other hand, with a study design of this kind it would also be necessary to increase the sample size, since a larger number of individuals in each group (healthy, diabetic and hypertensive) could help clarify which anthropometric parameters are best related to arterial stiffness in each of them. We have not analysed causal relationships between uric acid and metabolic syndrome. As noted by the reviewer, the second limitation is the study population because had several risk factors including hypertension, and dyslipidemia. To eliminate the possible effect of underlying diseases and medications used for these diseases, we have added a fourth model in the multivariate adjustment including systolic blood presure, total cholesterol, serum glucose, antidiabetic drugs, atherosclerotic plaques and high sensitivity C-reactive protein, in addition of the rest of variables of the third model. We have included in the statistical analysis section the following in relation of the new multivariate adjustment....and finally in the fourth model: systolic blood presure, total cholesterol, high sensitivity C-reactive protein, serum glucose, antidiabetic drugs and atherosclerotic plaques.. We have added in table 3 (multivariate adjustment) the new model including systolic blood presure, total cholesterol, high sensitivity C-reactive protein, serum glucose, antidiabetic drugs and atherosclerotic plaques, in addition of the rest of variables of the third model. We have changed the results section and now reads as follows: After adjusting for age, gender and cardiovascular risk factors, the multiple regression analysis found that for every 0.1 point increase in WHtR, the PWV increased 0.041 m/sec and C-IMT increased 0.001 mm. Likewise, a 1 cm increase in WC implied an increase of 0.029 m/sec in PWV, 0.001 mm in C-IMT, and a decrease of 0.141 in CAIx. A 1 kg/m² increase in BMI in turn implied an increase of 0.052 m/sec in PWV, and a decrease of 0.270 in CAIx. Lastly, every 1% increase in BFP implied and a decrease of 0.284 in CAIx (Table 3). 8

Reviewer: Andrzej Wykretowicz (2) Point 1: This is an interesting study which in my opinion deserves publication, however some problems needs to be addressed. The main criticism concerns the issue underlined in the title. Namely the association of different measures of obesity with arterial stiffness. However the only measure of arterial stiffness (segmental) used in the article is pulse wave velocity. Other markers cannot be used interchangeably for the assessment of arterial stiffness. Intimia media thickness (IMT) is a measure of subclinical atherosclerosis while Augmentation Index (AIx) characterized wave reflection. These issues were discussed by others ( Cecelja M, Jiang B, Bevan L, Frost ML, Spector TD, Chowienczyk PJ. Arterial stiffening relates to arterial calcification but not to noncalcified atheroma in women. J Am Coll Cardiol 2011;57:1480 6; Tu ST, Wang IW, Lin HF, Liao YC, Lin RT, Liu CS, Juo SH. Carotid intima-media thickness and stiffness are independent risk factors for atherosclerotic disease J Investig Med 20101;58:786-790; Wykretowicz A, Gerstenberger P, Guzik P, Milewska A, Krauze T, Adamska K, Rutkowska A, Wysocki H. Arterial stiffness in relation to subclinical atherosclerosis. Eur J Clin Invest 2009; 39: 11 16 etc.) Therefore Authors evaluated abdominal obesity vs general obesity for identifying arterial stiffness, subclinical atherosclerosis and wave reflection (or vascular status for short) in healthy, diabetics and hypertensive. This should be discussed properly. Thank you for your kind remarks. As indicated by both reviewers, the intima media thickness (IMT) is a measure of subclinical atherosclerosis. The only measure of arterial stiffness used in the article is pulse wave velocity. Other markers cannot be used interchangeably for the assessment of arterial stiffness. We have included the following reference in the introduction: Arterial stiffness in relation to subclinical atherosclerosis. (Wykretowicz A, Gerstenberger P, Guzik P, Milewska A, Krauze T, Adamska K, Rutkowska A, Wysocki H. Eur J Clin Invest. 2009 Jan;39(1):11-6. PMID: 19087126) We have rewritten the title, and now reads as follows: Abdominal obesity vs general obesity for identifying arterial stiffness, subclinical atherosclerosis and wave reflection in healthy, diabetics and hypertensive. We have changed the introduction and now reads as follows: The vascular structure and function can be assessed through the indices of subclinical atehrosclerosis, arterial stiffness and wave reflection [1]. A relationship has been found between measures that assess excess body weight or obesity to certain parameters that measure arterial stiffness and subclinical atherosclerosis, such as the pulse wave velocity (PWV) and the intima-media thickness of the common carotid artery (C-IMT), though their correlation to the augmentation index is not clear[2-4]. We have rewritten the aim of the study: The present study explores the relationship between anthropometric indices that assess abdominal obesity (WC, WHtR) and general obesity (BMI and body fat percentage (BFP)), with parameters that measure arterial stiffness (PWV, central and peripheral pulse pressure), subclinical atherosclerosis (C-IMT and and ankle-brachial index (ABI)) 10

and wave reflection (central augmentation index) in healthy, diabetics and hypertensive subjects. We have changed the results section and now reads as follows: Table 1 shows the clinical and demographic characteristics of the 305 subjects included in the study (diabetics 32.8%, hypertensive subjects 37.0% and healthy individuals 30.2%), together with the anthropometric measures, inflammatory markers, central and peripheral arterial pressure, measures of arterial stiffness, subclinical atherosclerosis and wave reflection. We have rewritten the first paragraph of the discussion section, and now reads as follows: The present study shows that the measures of abdominal obesity (WHtR and WC) are better than the general obesity indicators (BMI and BFP) in predicting arterial stiffness as evaluated by PWV and subclinical atherosclerosis evaluated by C-IMT, independently of the presence of diabetes or hypertension. Therefore, WC and WHtR are the most useful parameters for estimating arterial stiffness in clinical practice. We have modified the conclusions, which now reads as follows: Based on the results obtained in our study, it can be concluded that, independently of the presence of diabetes or hypertension, the measures of abdominal obesity (WHtR and WC) correlates better than BMI and BFP with arterial stiffness and subclinical atherosclerosis evaluated by PWV and C-IMT respectively. However, the association with wave reflection (evaluated by CAIx) was better with general adiposity measures. We have changed the abstract (background section), and now reads as follows: Our aim was to analyze the relationship between abdominal obesity and general obesity, with subclinical atherosclerosis, arterial stiffness and wave reflection in healthy, diabetics and hypertensive subjects. We have changed the abstract (conclusions section), and now reads as follows: The measures of abdominal obesity (WHtR and WC) correlates better than BMI and BFP with arterial stiffness evaluated by PWV, and with subclinical atherosclerosis evaluated by C-IMT, independently of the presence of diabetes or hypertension. 1. Wykretowicz A, Gerstenberger P, Guzik P, et al: Arterial stiffness in relation to subclinical atherosclerosis. Eur J Clin Invest 2009, 39:11-16. 2. Ko MJ, Kim MK, Shin J, Choi BY: Relations of pulse wave velocity to waist circumference independent of hip circumference. Epidemiol Health 2010, 32:e2010004. 3. Rider OJ, Tayal U, Francis JM, et al: The effect of obesity and weight loss on aortic pulse wave velocity as assessed by magnetic resonance imaging. Obesity (Silver Spring) 2010, 18:2311-2316. 4. Wildman RP, Mackey RH, Bostom A, Thompson T, Sutton-Tyrrell K: Measures of obesity are associated with vascular stiffness in young and older adults. Hypertension 2003, 42:468-473. 11

5. O'Brien E, Asmar R, Beilin L, et al: Practice guidelines of the European Society of Hypertension for clinic, ambulatory and self blood pressure measurement. J Hypertens 2005, 23:697-701. 6. Montagne P, Laroche P, Cuilliere ML, Varcin P, Pau B, Duheille J: Microparticle-enhanced nephelometric immunoassay for human C-reactive protein. J Clin Lab Anal 1992, 6:24-29. 7. Kizer JR, Biggs ML, Ix JH, et al: Measures of adiposity and future risk of ischemic stroke and coronary heart disease in older men and women. Am J Epidemiol 2011, 173:10-25. 12