Coronary heart disease risk in men and the epidemic of overweight and obesity

Transcription

1 (2005) 29, & 2005 Nature Publishing Group All rights reserved /05 $ PAPER Coronary heart disease risk in men and the epidemic of overweight and obesity K Nanchahal 1 *, JN Morris 1, LM Sullivan 2 and PWF Wilson 3 1 Public & Environmental Health Research Unit, Department of Public Health & Policy, London School of Hygiene & Tropical Medicine, Keppel Street, London, UK; 2 Departments of Biostatistics, Mathematics & Statistics, Boston University, USA; and 3 Department of Endocrinology, Diabetes, and Medical Genetics, Medical University of South Carolina, Charleston, USA OBJECTIVE: To evaluate the contributions of socioeconomic, lifestyle, and body weight factors to predicted risk of coronary heart disease (CHD) in the population and thus provide a focus for policies on prevention. DESIGN: Prospective study and cross-sectional population health survey. SUBJECTS: In all, 3090 men in the Framingham study and 2571 men in the 1998 Health Survey for England (HSE) aged y with no history of cardiovascular disease participated in the study. MEASUREMENTS: Data on sex, age, systolic blood pressure and antihypertensive medication, total and high-density lipoprotein cholesterol levels, diabetes, and their association with the incidence of myocardial infarction and fatal CHD in the Framingham study population were used to derive functions for predicting individual 10-y risk of CHD. These functions were applied to the same data on participants in the HSE. High risk was defined as 10-y CHD risk Z15%. The proportion of high risk in the English population attributable to each of the risk factors examined was assessed. RESULTS: In all, 32% of men in England had predicted 10-y CHD risk Z15%. Such high risk was significantly associated with body mass index (BMI, kg/m 2 ), waist:hip ratio (WHR), smoking, and levels of physical activity, educational attainment, and income (all Pr0.007). In this population, 47% of high CHD risk was attributable to excess body weightfbmi Z25 kg/m 2 and/ or WHR Z0.95Fand 31% to the sum of the four other significant factors: lack of educational qualifications, low income, smoking, and physical inactivity. CONCLUSION: Overweight and obesity now dominate the standard risk factors of CHD in men and should be the focus of national policies for prevention. (2005) 29, doi: /sj.ijo Published online 14 December 2004 Keywords: coronary heart disease; overweight; epidemiology; population attributable risk Introduction The UK government s National Service Framework for coronary heart disease (CHD) aims to reduce CHD mortality by 40% in the first decade of this century by setting national standards for improving prevention, diagnosis, and treatment. 1 Identification of individuals at very high risk of CHD and the offer of drug therapy 2 is one approach to reduce CHD incidence over the long term. An alternative is to develop public health policy prevention programmes. Aspects of socioeconomic status such as occupation, 3 educational *Correspondence: Ms K Nanchahal, Public & Environmental Health Research Unit, Department of Public Health & Policy, London School of Hygiene & Tropical Medicine, Keppel Street, London WC1E 7HT, UK. Kiran.Nanchahal@lshtm.ac.uk Received 11 May 2004; revised 22 September 2004; accepted 30 September 2004; published online 14 December 2004 attainment, 4 and income; 5 lifestyle factors such as smoking, 6 adverse dietary habits, 7 and physical inactivity; 8 and overweight 9 are associated with CHD. Evaluation of the contribution of each of these factors to the high risk of CHD in the population would help focus the policies and interventions for prevention. National 10 and international 11 recommendations on preventing CHD employ the Framingham Heart Study risk functions 12 to identify individuals at high risk for CHD. The published Framingham risk functions are a valuable clinical tool and include an important lifestyle factor, smoking, in the assessment of risk. In order to assess the contribution of smoking, as well as that of other lifestyle factors to predicted risk in the population, we derived new Framingham functions that excluded smoking. We applied them to data on a representative sample of men in England and assessed the proportion of high CHD risk in this population attributable to socioeconomic, lifestyle, and anthropometric

2 318 risk factors for CHD that are amenable to change through interventions at societal and individual levels. (WHR) was grouped into five categories based on cut-points at 0.85, 0.90, 0.95 and Methods Study population The Health Survey for England (HSE) is a large, annual, nationwide, household-based cross-sectional survey of a representative sample of the population. The present study uses the 1998 HSE that focused on cardiovascular disease. Details of the methods are provided elsewhere. 13 Briefly, computer-assisted interviewing was used to collect information on a range of lifestyle factors and health outcomes. Standardised anthropometric measures and blood pressure were determined, and a blood sample was obtained at a subsequent nurse visit. The overall response rate among men aged 16 y and over was 66%. Risk factors Socioeconomic. Social class was based on the Registrar General s standard classification by occupation of head of household. Total household income was adjusted for the number of people in the household and divided into quartiles. The perceived social support scale was derived from questions on the amount of support and encouragement received from family and friends, and participants were classified as having no lack, some lack, and severe lack of social support. Education was coded according to the highest qualification obtained. Lifestyle. Diet was assessed by questionnaire to estimate the total amount of fat and fibre intake, and consumption divided into low, medium, and high categories. Physical activity was assessed by questions on occupational activity, walking, and sport and recreation. Participants were asked about frequency, duration, and intensity of the different types of activity in the 4 weeks prior to interview. They were classified as active (meeting the current international guidelines of 30 min or more of moderate intensity activity at least five times a week), less active (undertaking light activities), or inactive. Participants were also categorised as neversmokers, ex-smokers, or current-smokers. Anthropometry. Weight, to the nearest 0.1 kg, was measured using Soehnle electronic scales after participants had removed outer garments and shoes. Height was measured using a portable stadiometer and recorded to the nearest millimetre. Waist circumference was measured to the nearest millimetre at a point midway between the iliac crest and the costal margin (lower rib) using an insertion tape with a metal buckle at one end. Hip circumference was measured at the widest point over the buttocks below the iliac crest. Body mass index (BMI), calculated as weight in kilograms divided by height in metres squared, was categorised into five groups using cut-points at 23, 25, 27.5 and 30 kg/m 2. Waist:hip ratio Framingham Heart Study functions We used data from the Framingham Heart Study on 3090 men aged y who were free of CHD at baseline examination carried out between 1968 and Participants were followed for 12 y and the incidence of hard CHD (defined as myocardial infarction or coronary death) was recorded. Cox s proportional-hazards regression functions were derived, relating age, systolic blood pressure, antihypertensive medication, total and high-density lipoprotein cholesterol levels, and presence of diabetes to the development of CHD. Statistical analysis All analyses were performed using STATA version 8.1. We calculated 10-y CHD risk for men aged y who participated in the 1998 HSE, applying the Framingham functions that were developed for the present study. Ageadjusted prevalence was calculated using the method of direct standardisation. We used logistic regression models to assess the association between high CHD risk (defined as 10- y risk Z15%), and potentially modifiable socioeconomic, lifestyle, and anthropometric risk factors. The standard way of estimating the public health impact of a particular risk factor is to use its prevalence in the population and relative risk of the event associated with it, to yield a population attributable risk fraction (PARF). The combined PARF for a number of risk factors cannot be obtained by simply adding individual PARFs, because of the overlap between the risk factors. We therefore estimated the unadjusted and adjusted measures of the PARF for each risk factor, based on logistic regression models 14 for the 2165 participants with complete data on all the significant factors, that is, BMI Z25 kg/m 2 ; WHRZ0.95; physical inactivity; ever-smoked; lack of educational qualifications; and household income in the lowest quartile. Results Data were available for 2571 (64.6% of 3981) men aged y who participated in the 1998 HSE and had complete information on factors required for the assessment of coronary risk, after excluding 9.9% (437/4418) men reporting angina, heart attack, or stroke diagnosed by a doctor. The participants had a mean age of 51.5 y, 8.3% were on antihypertensive medication and 2.6% were diabetic (Table 1). Coronary risk The median predicted 10-y CHD risk was 11.6 per 100 men based on the risk functions developed for this study. There

3 Table 1 Characteristics of men aged y in the two study samples Table 2 (continued) 319 Health Survey for England (1998) Framingham study (baseline ) Number Age (y) 51.5 (11.0) 52.6 (9.5) Total cholesterol (mmol/l) 5.71 (1.03) 5.60 (1.02) HDL-cholesterol (mmol/l) 1.28 (0.36) 1.16 (0.32) SBP (mmhg) (17.3) (18.9) Treatment for hypertension 8.3% 12.4% Diabetes 2.6% 5.8% Mean (s.d.) or prevalence (%). Table 2 Predicted 10-y CHD risk in men aged y according to socioeconomic, lifestyle, and anthropometric factors Number Predicted 10-y CHD risk (%) Minimum Maximum Median (lower, upper quartile) Educational attainment A level and above (7.1, 14.8) O level/cse and equivalent (7.6, 16.8) None (9.1, 19.1) Household income (quartiles) Top (7.0, 14.8) 2nd (7.8, 15.8) 3rd (8.0, 17.5) Bottom (9.5, 20.1) Social class ifprofessional (7.2, 15.9) iifmanagerial/technical (7.8, 16.0) iiinfskilled nonmanual (7.3, 18.1) iiimfskilled manual (8.0, 17.2) ivfsemiskilled manual (8.0, 17.1) vfunskilled manual (8.6, 16.4) Social support No lack (7.6, 16.2) Some lack (7.9, 17.5) Severe lack (7.9, 16.6) Smoking Never-smoker (7.0, 14.7) Ex-smoker (8.7, 17.8) Current-smoker (7.6, 16.5) Physical activity Active (7.0, 14.8) Less active (7.9, 17.3) Inactive (9.7, 19.3) Fat intake Low (7.8, 16.5) Medium (7.5, 16.6) High (8.4, 17.3) Fibre intake Low (7.6, 16.1) Medium (7.7, 17.1) High (8.1, 17.5) Number Predicted 10-y CHD risk (%) Minimum Maximum Median (lower, upper quartile) BMI (kg/m 2 ) o (5.5, 10.5) (6.4, 12.6) (7.8, 16.9) (10.0, 19.1) (10.3, 19.7) WHR o (5.4, 9.6) (6.3, 13.1) (8.4, 15.9) (10.7, 20.1) (12.2, 21.6) Waist (cm) o (5.0, 9.5) (6.5, 12.6) (7.6, 15.3) (8.8, 17.3) (10.4, 19.3) (10.3, 19.7) (11.2, 20.8) (12.1, 22.6) was considerable overlap in the distributions of CHD risk across the categories for each risk factor examined (Table 2). Among socioeconomic factors, risk was associated with educational attainment and income but not with social class or social support. Among lifestyle factors, risk was associated with smoking and level of physical activity but not with fat or fibre consumption. The strongest associations were observed with weight-related measures. High coronary risk About a third of men had a predicted 10-y risk Z15%. Such high risk was independently associated with educational attainment, income, smoking, physical activity, BMI, and WHR (all Pr0.007) (Table 3). The prevalence of high risk increased markedly with increasing WHR within each category of BMI, but less so with increasing BMI at high levels of WHR (Figure 1). Population attributable risk A total of 59.3% of high risk in this population was attributable to BMI Z25 kg/m 2, and 32.0% to WHR Z0.95 (Table 4). A reduction of 46.6% ( %, Table 4) in the number of men with predicted 10-y CHD risk Z15% could theoretically be achieved if all men attained BMI o25 kg/m 2 and WHR o0.95, after allowing for the contributions of the four significant socioeconomic and lifestyle factors. Smoking, physical inactivity, lack of educational qualifications,

4 320 Table 3 Predicted 10-y risk of myocardial infarction or death from CHD Z15% in men, according to their socioeconomic status, lifestyle, and body weight Number Percent with risk Z15% Odds ratio (95% confidence interval) a Unadjusted Adjusted b Education Po P ¼ A level and above O level/cse and (1.15, 1.79) 1.20 (0.93, 1.54) equivalent None (1.90, 2.96) 1.54 (1.18, 2.02) Income (quartile) Po P¼ Top nd (0.94, 1.58) 1.04 (0.79, 1.39) 3rd (1.19, 2.01) 1.22 (0.91, 1.64) Bottom (2.09, 3.53) 1.83 (1.34, 2.51) Smoking Po P ¼ Never-smoker Ex-smoker (1.54, 2.35) 1.46 (1.16, 1.85) Current-smoker (1.07, 1.79) 1.28 (0.96, 1.71) Physical activity Po P¼ Active c Less active (1.31, 1.99) 1.52 (1.21, 1.92) Inactive (1.80, 3.00) 1.54 (1.16, 2.05) BMI (kg/m 2 ) Po Po o (1.10, 3.10) 1.54 (0.89, 2.67) (2.82, 7.20) 3.04 (1.83, 5.06) (5.20, 13.4) 4.40 (2.61, 7.40) (4.92, 12.8) 2.74 (1.60, 4.69) WHR Po Po o (2.40, 9.20) 3.18 (1.60, 6.33) (4.40, 16.3) 4.38 (2.22, 8.63) (10.3, 38.6) 8.53 (4.25, 17.1) (15.3, 60.2) 11.9 (5.69, 24.7) a Odds ratios are based on 2165 men with complete data on all risk factors included in the table. b Odds ratios for each risk factor, adjusted for the other risk factors in the table. C 30 min or more of moderate intensity activity at least five times per week. and low income together accounted for an additional 31.4% of men at high risk. Prevalence (%) <23 Body Mass Index (kg/m sq) <0.85 Waist:Hip ratio Figure 1 Prevalence (%) of predicted 10-yCHD risk Z15% according to WHR and BMI in men aged y. Table 4 Prevalence of the risk factors, odds ratios, and PARFs of predicted CHD risk Z15% during the next 10 y Prevalence (%) Odds ratio a PARF (%) All men (n ¼ 2165) Men at high risk (n ¼ 685) BMI and WHR Complete data on all anthropometric measures were available on 2440 men. In all, 30% (745/2440) of men had both BMI Z25 kg/m 2 and WHR Z0.95, while 26% (636/2440) had both BMI o25 kg/m 2 and WHR o0.95. There was little difference in the age-adjusted prevalence of overweight/ obesity with educational attainment, income, or level of physical activity, and a marked reduction among current Unadjusted Adjusted b PARF(1) c PARF(2) d Education O level/cse and above No qualification Income Upper three quartiles Bottom quartile Smoking Never-smoker Ever-smoker Physical activity Active e Inactive BMI (kg/m 2 ) o Z WHR o Z a Odds ratios are based on 2165 men with complete data on all risk factors included in the table. b Odds ratios for each risk factor, adjusted for all the other risk factors in the table. c PARF(1)Fproportion of high risk in the population that is attributable to individual risk factors, based on separate logistic regression models for each risk factor. d PARF(2)Fcumulative proportions of high risk in the population attributable to these risk factors, summed consecutively, based on a single logistic regression model that included all the risk factors in the table. e 30 min or more of moderate intensity activity at least five times per week.

5 a Prevalence (%) b Prevalence (%) A level + A level + O level/ CSE None Highest 2nd 3rd smokers compared to non/ex-smokers (Figure 2). The relationship of high WHR with each of these four factors was much stronger than for BMI. The age-adjusted prevalence of WHR Z0.95 increased with decreasing levels of educational attainment and income, was much higher among the physically inactive compared to the two active groups, and was higher among current and ex-smokers than in never-smokers. Systolic blood pressure and total cholesterol increased sharply with increasing BMI among men with WHR o0.95 and was high at all levels of BMI in men with WHR Z0.95 (Figure 3). HDL-cholesterol decreased with increasing BMI, the rate of decrease being less among men with WHR Z0.95 than in those with WHR o0.95. Discussion Our results provide the strongest evidence to date concerning the effects of overweight and obesity on predicted CHD risk. A reduction of 47% in the number of men with predicted 10-y CHD risk Z15% could theoretically be achieved if all men attained BMI o25 kg/m 2 and WHR o0.95, after allowing for the 31% of high risk that is attributable to the other significant factors, that is, lack of Lowest Education Income Physical activity Smoking O level/ CSE None Highest 2nd 3rd Lowest Education Income Physical activity Smoking Figure 2 Age-adjusted prevalence of high BMI and high WHR according to educational qualifications, income, physical activity, and smoking in men aged y. (a) BMI and (b) WHR. Active Active Less active Less active Inactive Inactive Never Never Ex Ex Current Current educational qualification, low income, smoking, and physical inactivity. This is the first direct evidence on the impact of the ongoing epidemic of excess weight on CHD risk and suggests that this should therefore be the focus of preventive efforts. This study is strengthened by using the current national data on CHD risk and thus on the effects of long-term exposure to the overweight/obesity epidemic. A possible limitation is that application of Framingham CHD risk functions may not be altogether appropriate for the population of England. Previous Framingham risk functions have distinguished low- from high- risk individuals in diverse populations 12,15 and accurately predicted risk in northern Europe. 16 However, as Framingham risk functions have recently been reported to overpredict risk in England, 17 we raised the definition of high risk from 15 to 20%. This had little effect on the results for PARF, 85% of CHD risk Z20% was attributable to the factors examined (figures not shown) compared to 78% of CHD risk Z15%. Of greater importance is that our results show adiposity as the predominant factor underlying increased coronary risk. PARFs depend on the risk factors included, methods of assessment, and thresholds chosen for continuous factors. Furthermore, PARFs partly depend on the prevalence of risk factors in the population, and estimates may not apply to other populations where distributions of risk factors differ. The HSE 1998 was not designed to assess PARFs, and the measurements of factors such as physical activity and diet, which are particularly difficult to assess in large samples, may not adequately assess their importance in the development of CHD. Selection of different cut-points for the continuous factors will yield different PARFs. There are well-established definitions for overweight and obesity based on cut-points of BMI that we used, 18 but there is no consensus on appropriate thresholds for WHR. 19 An alternative cut-point of 27.5 kg/m 2 for BMI, with cut-points for all other factors as before, could achieve a total reduction of 60.2% rather than 78.0% in men at high risk of CHD. For WHR, we used the cut-point of the HSE, 13 where high WHR in men was defined as 0.95 or over. Finally, we may have underestimated the true size of the PARFs for the risk factors examined because assessment was made at only one time point, which can lead to an inaccurate assessment of exposures for factors such as physical activity. The association of adiposity and CHD is well documented. 9,20 22 BMI is the most commonly used indicator of relative body weight in population studies. It does not, however, take account of fat distribution, so WHR or waist measurements have been increasingly used. We prefer the WHR as it has a smaller correlation with BMI than waist circumference (correlation coefficients of and 0.858, respectively) and may provide more independent information on regional fat distribution. We know of no other study that has estimated PARFs for high predicted 10-y CHD risk. Using data from the Framingham study, Wilson et al 21 reported population attributable 321

6 322 Figure 3 Relationships between blood pressure and lipid levels, included in the risk functions, and BMI according to WHR in men aged y. The lines are based on predicted values from linear regression models including BMI as a continuous variable, WHR as a categorical variable, and a term for the interaction between them. risks in the region of 20 25% related to excess weight (BMI Z25 kg/m 2 ) for hypertension, angina pectoris, and heart disease development in the short term (2 y). Population attributable risk of overall mortality associated with overweight and obesity in Canada among y old increased from 5.1% in 1985 to 9.3% in These results compare to our unadjusted estimate of 59.3% for predicted 10-y risk Z15% of myocardial infarction and CHD mortality attributable to BMI Z25 kg/m 2. Much of the CHD risk associated with overweight is mediated by its association with hypertension, lowered HDL, and increased triglyceride concentration and insulin resistance. 23 Moreover, weight gain increases risk of CHD and weight loss significantly improves risk factors for CHD. 24 Every kilogram of weight gain after high school increased risk of CHD by 3.1% in men, while every kilogram of weight loss resulted in a reduction of 0.7% in LDL-cholesterol, 0.5% in systolic blood pressure, 0.2 mm in blood glucose, and an increase of 0.2% in HDL-cholesterol. The loss of 5 10% in initial body weight among overweight and obese adults can lead to reductions in various chronic disease risk factors, clinical benefits, and improvements in health. 22 Currently used risk functions for the prediction of CHD in clinical practice exclude excess body weight because it is considered to affect risk indirectly, through the more proximal physiological and metabolic factors such as blood pressure, lipid levels, and diabetes, which are included in individual risk assessment. However, obesity also appears to have an independent direct positive association in the long term on cardiovascular mortality, even after controlling for variables on the causal pathway. 25 Addressing the growing epidemic of excess body weight can be anticipated to have a substantial effect on the factors included in risk prediction and may also result in beneficial changes in other weightrelated factors associated with CHD such as insulin resistance, C-reactive protein, and fibrinogen. The increase in prevalence of obesity, diabetes, and physical inactivity between 1981 and 2000 in England and Wales is estimated to have contributed to about 8000 additional deaths due to CHD in The current high prevalence of excess body weight and the dominance of its role on predicted CHD risk demonstrated here, together with

7 the established role of weight reduction in the treatment of hypertension, dyslipidaemia and diabetes, makes it an obvious choice for intervention across the population. In addition, development and implementation of effective weight reduction and weight maintenance programmes for individuals has to be a priority. Intervention programmes could initially be aimed at those with high WHR, particularly those without educational qualifications, on low incomes, the physically inactive, and smokers. Such a focus could have a substantial impact on healthfincluding on the incidence of CHDFas well as implications for the economic costs to individuals, the health service, and society. Acknowledgements This research was partly supported by NIH N , NIH R01-HL , and a grant from the Donald W Reynolds Foundation Programme in Clinical Cardiovascular Research. References 1 Department of Health. National Service Framework for Coronary Heart Disease. Department of Health: London; Nanchahal K, Duncan JR, Durrington PN, Jackson RT. Analysis of predicted coronary heart disease risk in England based on Framingham study risk appraisal models published in 1991 and BMJ 2002; 325: Kunst AE, Groenhof F, Mackenbach JP, Health EW. Occupational class and cause specific mortality in middle aged men in 11 European countries: comparison of population based studies. EU Working Group on Scioeconomic Inequalities in Health. BMJ 1998; 316: Kilander L, Berglund L, Boberg M, Vessby B, Lithell H. Education, lifestyle factors and mortality from cardiovascular disease and cancer. A 25-year follow-up of Swedish 50-year-old men. Int J Epidemiol 2001; 30: Davey Smith G, Neaton JD, Wentworth D, Stamler R, Stamler J. Mortality differences between black and white men in the USA: contribution of income and other risk factors among men screened for the MRFIT. MRFIT Research Group. Multiple Risk Factor Intervention Trial. Lancet 1998; 351: Doll R, Peto R, Wheatly K, Gray R, Sutherland I. Mortality in relation to smoking: 40 years observations on male British doctors. BMJ 1994; 309: Hu FB, Willett WC. Optimal diets for prevention of coronary heart disease. JAMA 2002; 288: Morris JN. Exercise in the prevention of coronary heart disease: today s best buy in public health. Med Sci Sports Exerc 1994; 26: Jousilahti P, Tuomilehto J, Vartiainen E, Pekkanen J, Puska P. Body weight, cardiovascular risk factors and coronary mortality. 15- year follow-up of middle-aged men and women in eastern Finland. Circulation 1996; 93: Prevention of coronary heart disease in clinical practice. Recommendations of the Second Joint Task Force of European and other Societies on coronary prevention. Eur Heart J 1998; 19: Grundy SM, Paternak R, Greenland P, Smith Jr S, Fuster V. Assessment of cardiovascular risk by use of multiple-risk-factor assessment equations: a statement for healthcare professionals from the American Heart Association and the American College of Cardiology. Circulation 1999; 100: Anderson KM, Wilson PW, Odell PM, Kannel WB. An updated coronary risk profile: a statement for health professionals. Circulation 1991; 83: Erens B, Primatesta P (eds). Health Survey for England: Cardiovascular Disease. Volume 1: Findings. Volume 2: Methodology and Documentation. HMSO: London; Greenland S, Drescher K. Maximum likelihood estimation of the attributable fraction from logistic models. Biometrics 1993; 49: D Agostino Sr RB, Grundy S, Sullivan LM, Wilson P. CHD Risk Prediction Group. Validation of the Framingham coronary heart disease prediction scores: results of a multiple ethnic groups investigation. JAMA 2001; 286: Haq IU, Ramsay LE, Yeo WW, Jackson PR, Wallis EJ. Is the Framingham risk function valid for northern European populations. A comparison of methods for estimating absolute coronary risk in high risk men. Heart 1999; 81: Brindle P, Emberson J, Lampe F, Walker M, Whincup P, Fahey T, Ebrahim S. Predictive accuracy of the Framingham coronary risk score in British men: prospective cohort study. BMJ 2003; 327: World Health Organization. Obesity, preventing and managing the global epidemic. Report of a WHO Consultancy on Obesity. WHO: Geneva; Molarius A, Seidell JC. Selection of anthropometric indicators for classification of abdominal fatnessfa critical review. Int J Obes Relat Metab Disord 1998; 22: 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: Wilson PW, D Agostino RB, Sullivan L, Parise H, Kannel WB. Overweight and obesity as determinants of cardiovascular risk: the Framingham experience. Arch Intern Med 2002; 162: Katzmarzyk PT, Ardern CI. Overweight and obesity mortality trends in Canada, Can J Public Health 2004; 95: National Heart, Lung, and Blood Institute and North American Association for the Study of Obesity. The Practical Guide to the Identification, Evaluation, and Treatment of Overweight and Obesity in Adults, Bethesda, MD, NIH publication no ; 2000 (also available from practgde.htm ). 24 Anderson JW, Konz EC, Frederich RC, Wood CL. Long-term weight-loss maintenance: a meta-analysis of US studies. Am J Clin Nutr 2001; 74: Dyer AR, Stamler J, Garside D, Greenland P. Long-term consequences of body mass index for cardiovascular mortality: The Chicago heart association detection project in industry. Ann Epidemiol 2004; 14: Unal B, Critchley JA, Capewell S. Explaining the decline in coronary heart disease mortality in England and Wales between 1981 and Circulation 2004; 109: