Sex differences in the relationships between BMI, WHR and incidence of cardiovascular disease: a population-based cohort study

Transcription

1 (2006) 30, & 2006 Nature Publishing Group All rights reserved /06 $ ORIGINAL ARTICLE Sex differences in the relationships between BMI, WHR and incidence of cardiovascular disease: a population-based cohort study CLi 1, G Engström 1, B Hedblad 1, S Calling 1, G Berglund 2 and L Janzon 1 1 Department of Clinical Sciences in Malmö, Epidemiological Research Group, Malmö University Hospital, Malmö, Sweden and 2 Department of Clinical Sciences in Malmö, Internal Medicine Research Group, University Hospital, Malmö, Sweden Objective: Body mass index (BMI) is associated with increased incidence of cardiovascular disease (CVD). However, the risk could be very different for individuals with the same body mass. The present study explored whether regional fat distribution, as measured by waist hip ratio (WHR), could modify the impact of BMI on the risk of CVD in men and women. Design: Prospective population-based study. Subjects: A total of men and women, years old, from general population in Malmö, Sweden. Measurements: All subjects were followed over 7 years for the incidences of first-ever cardiac event (CE) and ischemic stroke in relation to BMI category (o25.0, , X30.0) and WHR. Results: The prevalence of overweight and obesity was 39.4 and 13.0%, respectively. During follow-up, 1280 subjects suffered a CVD event (750 CE, 530 ischemic stroke). The risk of CVD in women increased with increasing levels of WHR, irrespective of BMI category. In men, WHR (per 1 s.d. increase) was associated with increased incidence of CVD in those with normal weight (relative risk (RR) ¼ 1.24; 95% CI: ) after adjustments for confounding factors. However, WHR was not related to CVD in overweight men (RR ¼ 1.06; 95%CI: ) or obese men (RR ¼ 1.04; 95%CI: ). A significant interaction was observed between sex and WHR on the CVD risk. Conclusion: The effect of WHR on incidence of CVD is modified by the overall body weight and by gender. WHR adds prognostic information on the cardiovascular risk in women at all levels of BMI, and in men with normal weight. (2006) 30, doi: /sj.ijo ; published online 11 April 2006 Keywords: cohort study; body fat distribution; waist-to-hip ratio; body mass index; CVD risk; gender Introduction Body mass index (BMI) is an easily assessable and commonly used measure of overweight and obesity. Several prospective studies have shown that obesity, as measured by BMI, is associated with increased incidence of cardiovascular diseases (CVDs). 1,2 However, the risk could be very different for individuals with the same body mass. Recent studies show that the cardiovascular risk among obese subjects varies substantially depending on the levels of other risk factors associated with obesity. 1,3,4 Correspondence: Dr C Li, Department of Clinical Sciences in Malmö, Epidemiological Research Group, Malmö University Hospital, Entrance 59, 5th floor, Malmö, Sweden. cairu.li@med.lu.se Received 2 October 2005; revised 14 February 2006; accepted 5 March 2006; published online 11 April 2006 It has been shown that abdominal obesity, as measured by the waist circumference or waist hip ratio (WHR), is associated with an adverse metabolic profile as well as increased cardiovascular risk. 5 7 It has also been suggested that WHR is a better indicator of the cardiovascular risk in comparison to BMI. 5 7 However, it is unclear whether the impact of body fat distribution on CVD risk is similar at all BMI levels, and whether this association is similar for men and women. The present cohort study explored whether the cardiovascular risk for different levels of BMI was modified by the regional fat distribution as measured by WHR in men and in women. Materials and methods Study population Participants of the Malmö Diet and Cancer (MDC) cohort were eligible for the present study. Detailed information of the MDC study has been described previously. 8

2 1776 In brief, all men and women, born between 1923 and 1950 and living in Malmö city in the southern part of Sweden, were invited to the MDC study. During March 1991 to September 1996, the participants were invited to a health assessment program, including a self-administered questionnaire in combination with clinical examinations at the screening center. The cohort consisted of subjects ( men and women) from the eligible population of about individuals. 9 Out of them, 1442 were excluded from the present study due to missing values of anthropometric measurements, blood pressure (BP) values, lifestyle variables and history of stroke or myocardial infarction before the examination. The ethical committee at Lund University approved the MDC study (LU 51 90). Each participant gave an informed consensus. Baseline examinations A self-administered questionnaire assessed the information on medical history, current health, socio-demographic circumstance and lifestyle factors. 10 Subjects were categorized into current smokers or non-smokers. Alcohol intake was assessed by a diet method previously described in detail 11 and categorized according to daily consumption. 12 The categories for men/women were o20g/o15 g alcohol/day (low), g/15 30g alcohol/day (medium) and 440 g/430 g alcohol/day (high). Information on physical activities has also been described in detail elsewhere. 11 In brief, physical activities revealed through 18 questions covering a range of activities in the four seasons. The scores were divided into tertiles as low (tertile-1), moderate (tertile-2) and high (tertile-3) levels of physical activity. In addition, the questionnaire covered reproductive events, including prior use of oral contraceptives, oophorectomy, age at menopause, menopause status and administration of hormone replacement therapy (HRT). Postmenopausal were defined if women had no menstruation for at least 12 months or had a surgical menopause. Body weight, height, waist and hip circumferences and BP were measured at the screening centre. The examinations were performed by two trained nurses. Standing height was measured with a fixed stadiometer calibrated in centimeters. Weight was measured to the nearest 0.1 kg using balancebeam scale with subjects wearing light clothing and no shoes. BMI was calculated as weight in kilograms divided by height squared in meters. Waist midway was measured as the circumference (in centimetre) between the lowest rib margin and iliac crest and hip circumference (in centimeter) as the largest circumference between waist and thighs. WHR was defined as the ratio of circumference of waist to hip. By using a mercury sphygmomanometer, BP was measured in the right arm in a lying position after 5 min rest at the screening centre. Hypertension was defined as systolic blood pressure (BP) and/or diastolic BPX140/90 mm Hg or medical treatment for hypertension. Study categories Total body weight was grouped according to BMI category into normal weight (BMIo25.0), overweight (BMI ), and obesity (BMIX30.0). 13 Body fat distribution was classified by sex-specific tertiles of WHR. Cutoff points for tertiles of WHR were as follows: tertile-1 (meno0.917, womeno0.768), tertile-2 (men , women ) and tertile-3 (men40.962, women ). Retrieval of end points Every participant was followed for incidence of myocardial infarction, stroke or death until December 31, Information on morbidity and mortality were based on record linkage with Swedish Hospital Discharge Register, 14 the Swedish Causes of Death Register 15 and the Stroke Register in Malmö. 16 A cardiac event (CE) was defined as fatal or non-fatal myocardial infarction (International Classification of Diseases and Injuries (ICD)-9-codes: 410A-410X) or death due to ischemic heart disease (ICD-codes: 412, 414). Stroke was defined as rapidly developed clinical signs of local or global loss of cerebral function that lasted for 424 h or led to death within 24 h. The diagnosis of ischemic stroke was verified by computed tomography (CT) scan and /or necropsy. A first-ever CVD event was defined as fatal or non-fatal CE or ischemic stroke. Statistics Pearsons correlation was utilized to examine the correlation coefficients between anthropometric measurements. A general linear model or logistic regression model was applied to compare the age-adjusted baseline characteristics between the levels of BMI and WHR, respectively. Cox regression model was used to assess the relative risk (RR) of CVD events in relation to BMI or WHR. In the model assessing the continuous variable of WHR, four steps of adjustments were performed. The first step included age, the second step age and BMI, and in the third step additional adjustments were made for smoking, alcohol intake and physical activity. In the last step, education level and the confounders that could be parts of the causal chain between obesity and CVD (history of diabetes, hypertension and lipid-lowering treatment) were added. In women, menopausal status was included in the last step adjustments. Interaction terms were entered into the multivariate model to test potential interactions between WHR and sex, or WHR and age, on the risk of CVD events. All comparisons were two-sided and a 5% level of significance was used. The statistical analyses were conducted by the computer software SPSSt (11.5).

3 Results In the total of subjects ( men, women), the prevalence of overweight was 39.4% (49.6% in men, 33.1% in women) and obesity was 13.0% (12.5% in men, 13.4% in women). Background characteristics The sex-specific background characteristics in subjects with normal weight, overweight and obesity are described in Table 1. As compared to men and women with normal weight or overweight, obese subjects were lower educated, had a low rate of current smoking but higher 1777 Table 1 Baseline characteristics of studied participants according to BMI categories in men and women BMI categories P-value Normal weight Overweight Obesity (o25.0) ( ) (X30.0) Characteristics of men No. of subjects Age at screening, mean7s.d *,w High education (college/university), % *,z Lifestyle factors Current smokers, % **,ww,z High alcohol intake (440 g/day), % * Low physical activity (tertile-1), % **,zz Anthropometric measurements, mean7s.d. Waist circumference (cm) **,ww,zz Hip circumference (cm) **,ww,zz BMI **,ww,zz WHR **,ww,zz Systolic BP/mm Hg y **,ww,zz Diastolic BP/mm Hg y **,w,z Concurrent medical status Hypertension, % **,ww,zz Diabetes, % **,w,zz Use of BP-lowering drugs, % **,ww,zz Use of lipid-lowering drugs, % NS Characteristics of women No. of subjects Age at screening, mean7s.d **,ww,zz High education (college/university), % **,w Lifestyle factors Current smokers, % **,ww High alcohol intake (430g/day), % *,z Low physical activity (tertile-1), % **,w,zz Anthropometric measurements, mean7s.d. Waist circumference (cm) **,ww,zz Hip circumference (cm) **,ww,zz BMI **,ww,zz WHR **,ww,zz Systolic BP/mm Hg y **,ww,zz Diastolic BP/mm Hg y **,ww,zz Concurrent medical status Hypertension, % **,ww,zz Diabetes, % **,ww,zz Use of BP-lowering drugs, % **,ww,zz Use of lipid-lowering drugs, % **,ww Menopause status Age at menopause, mean7s.d NS Postmenopausal, % **,w Abbreviations: BMI, body mass index; BP, blood pressure; NS, nonsignificant; s.d., standard deviation; WHR, waist hip ratio;. P-values were adjusted for age: *Po 0.05, **Po between normal weight and obese groups; w Po 0.05, ww Po between normal weight and overweight groups; z o0.05, zz o0.001 between overweight and obese groups. y Blood pressure values were for untreated subjects only.

4 1778 prevalence of physical inactivity, diabetes and use of BPlowering medication. Women with normal weight were younger and nearly one-third were pre-menopausal. There were no significant differences in marital status, occupation levels and women s characters between groups (data not shown). In each BMI category, the baseline characteristics in relation to the sex-specific tertiles of WHR showed that the prevalence of current smoking, physical inactivity, diabetes and use of BP-lowering drugs increased linearly from the lowest to the highest tertile both in men and in women. Relative risk of cardiovascular disease events in relation to levels of body mass index or waist hip ratio The mean duration of follow-up was years. A total of 1280 first-ever CVD events (750 CE, 530 ischemic stroke) occurred during follow-up. The mean age of events was years (range: 30.0) in men and years (range: 32.4) in women. In men, there was an increasing age-adjusted risk for CVD by increasing levels of BMI (Po0.001 for trend) and WHR (P ¼ 0.01 for trend), respectively (Table 2). The increased risks remained significant when education, lifestyle factors, history of diabetes, hypertension, lipid-lowering medication were taken into account. Similar association was seen in women. When the analyses were made separately of CE or stroke, the results were compatible to the outcomes of total CVD events. One exception was that in women the multiple-adjusted relationship between BMI category and incidence of CE (P ¼ 0.22 for trend) and stroke (P ¼ 0.06 for trend) did not reach significance. Relative risk of cardiovascular disease events in relation to combined pattern of body mass index and waist hip ratio A significant positive correlation was found between BMI and WHR in men (r ¼ 0.62) and in women (r ¼ 0.42). A significant interaction (P ¼ 0.009) was observed between sex and WHR on CVD risk. No significant interaction was found between age and WHR in men or in women. The risk of CVD associated with sex-specific WHR (per 1 s.d. increase) in each BMI category is shown in Table 3. In women, the age-adjusted RR associated with 1 s.d. of WHR in normal weight, overweight and obese subjects were 1.29 (95% CI, ), 1.45 (95% CI, ), 1.33 (95% CI, ), respectively. The risks were attenuated but remained statistical significant after adjustments for lifestyle factors, BMI, education and other CVD risk factors. In men with normal weight, WHR was significantly associated with CVD risk (RR ¼ 1.26; 95% CI: ) after multipleadjustments. No similar risk association for WHR was observed in overweight and obese men. In Figures 1 and 2, the risk of CVD was assessed in relation to WHR tertiles in each BMI category. For women, the risk of CVD increased in parallel with increasing levels of WHR, irrespective of BMI levels (Figure 1). Normal weight and overweight women with high WHR level had two-fold CVD risk compared to women with low level of WHR. In obese women, this association was even stronger. For men, this association was observed only in men who had normal weight (Figure 2). Discussion Both BMI and WHR were associated with an increased incidence of CVD in this study, which is consistent with the Table 2 Relative risk (RR) and 95%CI for CVD events in relation to BMI category or WHR tertiles (T-1 to T-3) in men and women BMI category WHR tertiles o X30.0 T-1 T-2 T-3 Men No. of CVD cases (CE/stroke) 260 (164/96) 419 (253/166) 137 (84/53) 225 (136/89) 261 (172/89) 330 (193/137) Age at CVD event: mean (range) 65.9 (27.9) 66.6 (30.0) 65.8 (28.3) 67.0 (30.0) 66.7 (25.8) 63.4 (29.4) Relative risk (95%CI) Age-adjusted ( ) 1.67 ( ) ( ) 1.53 ( ) -Lifestyle-adjusted a ( ) 1.83 ( ) ( ) 1.50 ( ) -Adjusted b ( ) 1.57 ( ) ( ) 1.36 ( ) Women No. of CVD cases (CE/stroke) 179 (88/91) 187 (89/98) 98 (50/48) 87 (48/39) 132 (58/74) 245 (121/124) Age at CVD event: mean (range) 66.4 (32.3) 67.4 (28.0) 67.8 (27.7) 66.5 (32.4) 67.0 (27.1) 67.3 (27.7) Relative risk (95%CI) Age-adjusted ( ) 1.69 ( ) ( ) 2.48 ( ) Age-lifestyle-adjusted a ( ) 1.77 ( ) ( ) 2.29 ( ) Multiple-adjusted b ( ) 1.32 ( ) ( ) 1.95 ( ) Abbreviations: CE, cardiac event; CVD, cardiovascular disease. a RRs were adjusted for age, smoking, alcohol intake and physical activity. b RRs were adjusted for age, education, smoking, alcohol consumption, physical activity, history of diabetes, hypertension, and lipid-lowering drugs (menopausal status was included in the multiple model in women).

5 Table 3 Men Relative risk of first-ever CVD events according to sex-specific standard deviation (s.d.) of WHR in men and women BMI categories Normal weight Overweight Obesity (o25.0) ( ) (X30.0) 1779 RR (95%CI) Age-adjusted 1.31 ( ) 1.17 ( ) 1.12 ( ) Age-BMI-adjusted 1.32 ( ) 1.12 ( ) 1.07 ( ) Age-BMI-lifestyle-adjusted a 1.24 ( ) 1.06 ( ) 1.04 ( ) Multiple-adjusted b 1.26 ( ) 1.05 ( ) 1.02 ( ) Women RR (95%CI) Age-adjusted 1.29 ( ) 1.45 ( ) 1.33 ( ) Age-BMI-adjusted 1.28 ( ) 1.41 ( ) 1.31 ( ) Age-BMI-lifestyle-adjusted a 1.24 ( ) 1.31 ( ) 1.27 ( ) Multiple-adjusted b 1.22 ( ) 1.19 ( ) 1.20 ( ) One s.d. of WHR for men and women was and 0.053, respectively. a RRs were adjusted for age, BM, smoking, alcohol consumption and physical activity. b RRs were adjusted for age, BMI, education, smoking, alcohol consumption, physical activity, history of diabetes, hypertension and lipid-lowering drugs (menopausal status was included in the multiple model in women) Relative Risk of CVD Relative Risk of CVD T1-T2-T3 T1-T2-T3 T1-T2-T3 BMI<25.0 BMI BMI 30.0 Figure 1 Age-adjusted relative risk (RR) (95% confidence interval (CI)) of cardiovascular disease (CVD) events in relation to tertiles of waist-to-hip ratio (WHR) in normal weight (Body mass index (BMI) o25.0), overweight (BMI ) and obese (BMI X30) women. 0 T1-T2-T3 T1-T2-T3 T1-T2-T3 BMI<25.0 BMI BMI 30.0 Figure 2 Age-adjusted relative risk (RR) (95% confidence interval (CI)) of cardiovascular disease (CVD) events in relation to tertiles of WHR in normal weight (BMI o25.0), overweight (BMI ) and obese (BMI X30) men. results from other studies. 1,2,6,7 However, the patterns were more complex when the combined effects of BMI and WHR were assessed. The impact of body fat distribution on CVD risk was modified by the levels of overall body weight and by gender. WHR added to the cardiovascular risk in women, regardless of overall body weight. However, in men, WHR added prognostic information only in those with normal weight, but not in men with overweight or obesity. The results suggest that both BMI and WHR measurements and sex-specific relationship should be taken into account to classify the CVD risk. In this study, WHR was used as a proxy-estimate of intraabdominal fatness. Waist hip ratio showed a stronger relationship with CVD risk, particularly in women. Women with high WHR (40.811) had at least 2-fold higher risk of CVD than had those with low WHR (o0.768), regardless of BMI category (Figure 1). This association was confirmed when the continuous sex-specific variable of WHR (per 1 s.d. increase) was examined. The risks were somewhat attenuated when other factors (diabetes, hypertension and lipid-lowering medication) were added. However, it is likely that the addition of those risk factors into the analyses lead to

6 1780 overadjustments, as obesity is a main cause for developing those chronic diseases. By using more precise measurements (CT scan or magnetic resonance imaging), it has been shown that the volume of intra-abdominal fat only accounts for 6 20% of total adipose tissue volume. 17 However, this small volume of fat plays a key role in lipid availability by increased levels of free fatty acids and impaired adipocyte lipolysis, resulting in excessive tissue lipid accumulation, insulin resistance and vascular endothelial dysfunction. 17,18 Those impaired metabolic alterations could explain why individuals with increased visceral fat have high risk of CVDs. 19 In contrast to high WHR, women with low level of WHR had rather lower risk of CVD in this cohort. In addition, they also had a significantly lower prevalence of hypertension or diabetes. The low risks were even more noticeable in the obese women. These results imply that low level of WHR possesses a protective role on the risk of CVD. It has been suggested that the adipocytes in gluteo-femoral region (called female or gynoid type) are associated with minimal fatty acid flux, high lipoprotein lipase activity and high insulin sensitivity. 20 These bio-metabolic characteristics are highly efficient in fatty acid storage. Hence, gynoid obesity has been considered healthy obese pattern, at least from metabolic point. 21 The results were different in men. The association between WHR and CVD risk was only found in men with normal weight, but not in men with overweight or obesity. These findings are consistent with the findings reported from three other population-based cohorts from US, in which the risk of CVD events was examined by tertiles of BMI and WHR among men and women. 7,22,23 In the Nurses Health Study, a middle-aged women s cohort, a stepwise increase of incident coronary heart disease (CHD) was associated with increased tertile of WHR, independently of BMI tertile. 7 This association was also observed in elderly overweight or obese women when incidence of stroke was studied. 22 In the Physicians Health Study, the incidence of CHD in male physicians increased in line with ascending levels of BMI, but the relation between the CHD risk and WHR was only observed among men with BMI o The causes of the gender difference are unclear. Differences in anatomic, physiologic, metabolic and sex hormonal status between genders may provide certain explanations. The volume of visceral fat mass differs by sex. Men, on average, store 21% of total body fat in visceral region in comparison to 10% in women. 24,25 In addition, women have a wider hip circumference due to wider pelvis and larger gluteo-femoral muscle and fat depot. These anatomic distinctions result in different body fat distribution and cutoff points of WHR between genders. Body fat distribution has been linked to alteration of steroid hormones, in particular in women s menopause transition. Longitudinal studies have documented that women do not change BMI from pre-menopause to postmenopause, but that waist circumference increases and hip circumference decreases, resulting in increased WHR. 26 This suggests that the risk of CVD resulting from alteration of body fat distribution may become more apparent during women s middle-life. In addition, it has been suggested that women with increased abdominal adiposity may have increased androgenicity which results in high risk of CVD. 27 In this study, BMI was inversely related to smoking in men and women, and with alcohol consumption in women. The relationship between smoking and low body weight is well documented, 28 and the relationship with alcohol in women could be explained by a higher proportion of smokers in high consumers of alcohol. Because smoking is a major cardiovascular risk factor that is more prevalent in subjects with normal weight, this tends to reduce the relationships between obesity and CVD. Smoking was however taken into account in the multivariate analysis. The lack of follow-up anthropometric examination for all subjects is a limitation of this study. It is possible that body fat distribution in some cases changed during follow-up, but this is usually a slow process. Data from a longitudinal study has demonstrated that the variation of body weight and adipose tissue distribution is remarkably stable throughout the lifespan. 29 The participation rate was approximately 40%, and one question is if the study cohort was representative for the background population. A comparison between participants and non-participants showed that mortality was higher in non-participants. 9 For many baseline factors, including BMI, it was impossible to have information from non-participants. An attempt, therefore, was made to compare those baseline factors between the participants in MDC study and subjects from the same city who took part in a survey with a participation rate of 75%, which used a mailed questionnaire. BMI showed no substantial difference between two studies. 9 Finally, the absence of data on blood lipids, glucose and inflammatory markers is also a limitation as those examinations were only performed in sub-groups but not in whole cohort. However, these factors probably are links in the chain between obesity and CVD, taking these factors into account in the statistical model would lead to overadjustments. Conclusion The effect of body fat distribution on CVD risk was modified by the levels of overall body weight and by gender. High WHR adds prognostic information on the cardiovascular risk in women at all levels of BMI. Similar relationship was observed only in men with normal weight. Acknowledgements This study was supported by grants from the Heart and Lung Foundation, and the Segerfalk Fund. We thank Swedish

7 Cancer Society, Swedish Society for Hypertension, Swedish Medical Research Council and Scania Regional Government. References 1 Kannel WB, Wilson PW, Nam BH, D Agostino RB. Risk stratification of obesity as a coronary risk factor. Am J Cardiol 2002; 90: Zhou BF. Effect of body mass index on all-cause mortality and incidence of cardiovascular diseases report for meta-analysis of prospective studies open optimal cutoff points of body mass index in Chinese adults. Biomed Environ Sci 2002; 15: Jonsson S, Hedblad B, Engström G, Nilsson P, Berglund G, Janzon L. Influence of obesity on cardiovascular risk. Twenty-three-year follow-up of men from an urban Swedish population. Int J Obes Relat Metab Disord 2002; 26: Engstrom G, Hedblad B, Stavenow L, Jonsson S, Lind P, Janzon L et al. Incidence of obesity-associated cardiovascular disease is related to inflammation-sensitive plasma proteins: a populationbased cohort study. Arterioscler Thromb Vasc Biol 2004; 24: Dalton M, Cameron AJ, Zimmet PZ, Shaw JE, Jolley D, Dunstan DW et al. AusDiab Steering Committee. Waist circumference, waist-hip ratio and body mass index and their correlation with cardiovascular disease risk factors in Australian adults. J Intern Med 2003; 254: Lakka HM, Lakka TA, Tuomilehto J, Salonen JT. Abdominal obesity is associated with increased risk of acute coronary events in men. Eur Heart J 2002; 23: Rexrode KM, Carey VJ, Hennekens CH, Walters EE, Colditz GA, Stampfer MJ et al. Abdominal adiposity and coronary heart disease in women. JAMA 1998; 280: Berglund G, Elmståhl S, Janzon L, Larsson SA. Design and feasibility. J Intern Med 1993; 233: Manjer J, Carlsson S, Elmstahl S, Gullberg B, Janzon L, Lindstrom M et al. The Malmo Diet and Cancer Study: Representativity, cancer incidence and mortality in participants and non-participants. Eur J Cancer Prev 2001; 10: Li C, Engstrom G, Hedblad B, Berglund G, Janzon L. Blood pressure control and risk of stroke: a population-based prospective cohort study. Stroke 2005; 36: Wallstrom P, Wirfalt E, Janzon L, Mattisson I, Elmstahl S, Johansson U et al. Fruit and vegetable consumption in relation to risk factors for cancer: a report from the Malmo Diet and Cancer Study. Public Health Nutr 2000; 3: Royal College of Psychiatrists. Alcohol: Our Favorite Drug Tavistock: London. 13 World Health Organization. Obesity: preventing and managing the global epidemic. WHO: Geneva, Report of a WHO consultation on obesity, WHO/NUT/NCD The National Board of Health and Welfare. Centre of Epidemiology. Myocardial infarction in Sweden (In Swedish: Hälsa och Sjukdomar) Social-styrelsen 2003; The National Board of Health and Welfare. Centre of Epidemiology. Causes of Death Stockholm: Sweden, Jerntorp P, Berglund G. Stroke registry in Malmö. Stroke 1992; 23: Carey DG. Abdominal obesity. Curr Opin Lipidol 1998; 9: Brook RD, Bard RL, Rubenfire M, Ridker PM, Rajagopalan S. Usefulness of visceral obesity (waist/hip ratio) in predicting vascular endothelial function in healthy overweight adults. Am J Cardiol 2001; 88: Willett WC. Guidelines for healthy weight. N Engl J Med 1999; 341: McCarty MF. A paradox resolved: the postprandial model of insulin resistance explains why gynoid adiposity appears to be protective. Med Hypotheses 2003; 61: Despres JP, Lemieux I, Prud homme D. Treatment of obesity: need to focus on high risk abdominally obese patients. BMJ 2001; 322: Folsom AR, Prineas RJ, Kaye SA, Munger RG. Incidence of hypertension and stroke in relation to body fat distribution and other risk factors in older women. Stroke 1990; 21: Rexrode KM, Buring JE, Manson JE. Abdominal and total adiposity and risk of coronary heart disease in men. Int J Obes Relat Metab Disord 2001; 25: Kvist H, Chowdhury B, Sjostrom L, Tylen U, Cederblad A. Adipose tissue volume determination in males by computed tomography and 40 K. Int J Obes Relat Metab Dis 1988; 12: Kvist H, Sjostrom L, Tylen U. Adipose tissue volume determinations in women by computed tomography: technical considerations. Int J Obes Relat Metab Disord 1986; 10: Bjorkelund C, Lissner L, Andersson S, Lapidus L, Bengtsson C. Reproductive history in relation to relative weight and fat distribution. Int J Obes Relat Metab Disord 1996; 20: Haffner SM, Katz MS, Dunn JF. Increased upper body and overall adiposity is associated with decreased sex hormone binding globulin in postmenopausal women. Int J Obes Relat Metab Disord 1991; 15: Janzon E, Hedblad B, Berglund G, Engstrom G. Changes in blood pressure and body weight following smoking cessation in women. J Intern Med 2004; 255: Katzmarzyk PT, Perusse L, Malina RM, Bouchard C. Seven-year stability of indicators of obesity and adipose tissue distribution in the Canadian population. Am J Clin Nutr 1999; 69: