Adult Obesity and Number of Years Lived with and without Cardiovascular Disease

Transcription

1 Risk Factors and Chronic Disease Adult Obesity and Number of Years Lived with and without Cardiovascular Disease M. Carolina Pardo Silva,* Chris De Laet,* Wilma J. Nusselder,* Abdulah A. Mamun, and Anna Peeters* Abstract PARDO SILVA, M. CAROLINA, CHRIS DE LAET, WILMA J. NUSSELDER, ABDULAH A. MAMUN, AND ANNA PEETERS. Adult obesity and number of years lived with and without cardiovascular disease. Obesity. 2006;14: Objective: To determine the differences in number of years lived free of cardiovascular disease (CVD) and number of years lived with CVD between men and women who were obese, pre-obese, or normal weight at 45 years of age. Research Methods and Procedures: We constructed multistate life tables for CVD, myocardial infarction, and stroke, using data from 2551 enrollees (1130 men) in the Framingham Heart Study who were 45 years of age. Results: Obesity and pre-obesity were associated with fewer number of years free of CVD, myocardial infarction, and stroke and an increase in the number of years lived with these diseases. Forty-five-year-old obese men with no CVD survived 6.0 years [95% confidence interval (CI), 4.1; 8.1] fewer than their normal weight counterparts, whereas, for women, the difference between obese and normal weight subjects was 8.4 years (95% CI: 6.2; 10.8). Obese men and women lived with CVD 2.7 (95% CI: 1.0; 4.4) and 1.4 years (95% CI: 0.3; 3.2) longer, respectively, than normal weight individuals. Discussion: In addition to reducing life expectancy, obesity before middle age is associated with a reduction in the number of years lived free of CVD and an increase in the Received for review April 19, Accepted in final form March 7, The costs of publication of this article were defrayed, in part, by the payment of page charges. This article must, therefore, be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. *Department of Public Health and Department of Epidemiology and Biostatistics, Erasmus Medical Center Rotterdam, Rotterdam, The Netherlands; School of Population Health, University of Queensland, Queensland, Australia; and Department of Epidemiology and Public Health, Monash University, Victoria, Australia. Address correspondence to Anna Peeters. Department of Epidemiology and Preventive Medicine, Monash University Central and Eastern Clinical School, Alfred Hospital Commercial Road, Melbourne, Victoria 3004, Australia. anna.peeters@med.monash.edu.au Copyright 2006 NAASO number of years lived with CVD. Such information is paramount for preventive and therapeutic decision-making by individuals and practitioners alike. Key words: cardiovascular disease, life expectancy Introduction The prevalence of overweight has increased continuously over the last 20 years (1 3). It has been shown that being overweight is associated with risk factors for cardiovascular disease (CVD) 1 such as hypertension, hypercholesterolemia, and diabetes (4) and the risk of CVD itself (4 9). However, to provide comprehensive information for public and individual health care planning, it is necessary to move from measures such as relative risk to measures of the lifetime consequences of excess weight. While it is known that overweight increases the risk of CVD, it is unknown whether overweight people would live fewer or more years with CVD compared with those of normal weight. Studies evaluating the association between obesity and life expectancy (LE) have shown that obesity in adulthood is associated with decreases in LE of 6 to 7 years (10) and that severe lifetime obesity (BMI 45 kg/m 2 ) is linked to a maximum of 13 years of life lost in white men (11). One study has assessed the lifetime burden of CVD, estimating that men 45 to 54 years of age with a BMI of 37.5 kg/m 2 had a lifetime risk of coronary heart disease of 46.4%, compared with 34.9% for subjects with BMI of 22.5 kg/m 2 (12). Despite the contribution of these studies to the current knowledge of the lifetime implications of obesity, their approaches had limitations. On one hand, the reductions in LE associated with increased BMI did not distinguish between years of life spent without and with CVD (10,11). On the other hand, the assessment of the lifetime risks of CVD was done using a mathematical model based on short-term 1 Nonstandard abbreviations: CVD, cardiovascular disease; LE, life expectancy; MI, myocardial infarction; HR, hazard ratio OBESITY Vol. 14 No. 7 July 2006

2 Table 1. Baseline characteristic according to BMI groups for men and women Normal weight (18.5 <BMI<25) Pre-obese (25 <BMI<30) Obese (BMI >30) Men* 434 (38.4%) 545 (48.2%) 151 (13.4%) Mean BMI (kg/m 2 ) Mean weight (kg) Mean height (m) Never smokers 40 (9.2%) 71 (13.0%) 23 (15.2%) Current smokers 359 (82.7%) 411 (75.4%) 108 (71.5%) Mean SBP (mm Hg) Mean DBP (mm Hg) Free of hypertension 349 (80.4%) 347 (63.7%) 65 (43.0%) Mean SCL (mg/dl) Free of hypercholesterolemia 266 (61.3%) 280 (51.4%) 68 (45.0%) Free of diabetes 431 (99.3%) 541 (99.3%) 150 (99.3%) Graduated from high school or higher 287 (66.1%) 343 (62.9%) 88 (58.3%) Women* 869 (61.2%) 397 (27.9%) 155 (10.9%) Mean BMI (kg/m 2 ) Mean weight (kg) Mean height (m) Never smokers 305 (35.1%) 211 (53.1%) 85 (54.8%) Current smokers 511 (58.8%) 164 (41.3%) 60 (38.7%) Mean SBP (mm Hg) Mean DBP (mm Hg) Free of hypertension 754 (86.8%) 292 (73.6%) 66 (42.6%) Mean SCL (mg/dl) Free of hypercholesterolemia 608 (70.0%) 256 (64.5%) 87 (56.1%) Free of diabetes 867 (99.8%) 395 (99.5%) 152 (98.1%) Graduated from high school or higher 605 (69.6%) 241 (60.7%) 82 (52.9%) SBP, systolic blood pressure; DBP, diastolic blood pressure; SCL, serum cholesterol level. * Percentages refer to the total population. Percentages refer to the number of individuals within every BMI category. follow-up data without taking into account the competing risks of increased non-cvd mortality associated with obesity, leading to important differences in the estimations compared with other studies (11,13). One study found that obesity was associated with a small increase in the years lived with coronary heart disease, but this was within a relatively short follow-up of 15 years, and the years lived free of disease were not measured (14). The objective of this study was to determine the differences in the average number of years lived free of CVD (i.e., before the onset of CVD or death) and the average number of years lived with CVD for subjects who were normal weight, pre-obese, or obese at 45 years of age. We constructed multistate life tables using the mortality and CVD experience from 46 years of follow-up of the original Framingham Heart Study. Research Methods and Procedures To estimate the effect of obesity during adulthood on the number of years lived free of CVD and the number of years lived with CVD, we used data from the Framingham Heart Study. This is a cohort study involving 5209 respondents 28 through 62 years of age at enrollment, residing in Framingham, MA, between 1948 and 1951 (15). Data on occurrence of CVD and mortality were gathered by standardized biennial examinations, regular surveillance of hospital admissions, death registries, and medical records over a follow-up period of 46 years up to the year OBESITY Vol. 14 No. 7 July

3 Table 2. Hazard ratios for the different transitions for men and women Age 45 at start of follow-up Transition Disease BMI group* HR (95%CI) HR (95%CI) HR (95%CI) Men Incident disease CVD Pre-obese Age (1.19 to 1.90) 1.57 (1.30 to 1.90) 1.15 (0.73 to 1.80) Age (0.92 to 1.47) 1.19 (0.99 to 1.44) 1.30 (0.83 to 2.04) Obese 1.77 (1.41 to 2.22) 1.94 (1.54 to 2.43) 2.19 (1.36 to 3.52) MI Pre-obese Age (1.02 to 1.98) 1.48 (1.14 to 1.92) 1.04 (0.57 to 1.89) Age (0.74 to 1.44) 1.06 (0.82 to 1.37) 1.23 (0.68 to 2.23) Obese 1.40 (1.01 to 1.92) 1.46 (1.06 to 2.02) 2.04 (1.09 to 3.84) Stroke Pre-obese 1.55 (1.08 to 2.21) 1.62 (1.13 to 2.32) 1.97 (1.04 to 3.72) Obese 2.11 (1.32 to 3.38) 2.32 (1.45 to 3.72) 1.87 (0.54 to 6.44) Diseased to death CVD** Pre-obese 0.89 (0.74 to 1.08) 0.92 (0.76 to 1.11) 0.92 (0.65 to 1.29) Obese 0.92 (0.71 to 1.19) 0.96 (0.74 to 1.25) 0.97 (0.55 to 1.71) MI Pre-obese 0.88 (0.68 to 1.15) 0.89 (0.68 to 1.16) 0.96 (0.61 to 1.52) Obese 0.81 (0.56 to 1.17) 0.79 (0.54 to 1.15) 0.70 (0.32 to 1.52) Stroke Pre-obese 0.60 (0.40 to 0.91) 0.60 (0.39 to 0.90) 0.83 (0.36 to 1.93) Obese 0.68 (0.40 to 1.13) 0.69 (0.41 to 1.16) 1.10 (0.27 to 4.48) Non-diseased to death CVD Pre-obese 1.15 (0.99 to 1.34) 1.03 (0.79 to 1.33) 1.08 (0.74 to 1.58) Obese 1.53 (1.23 to 1.91) 1.33 (0.88 to 2.02) 1.80 (0.91 to 3.55) Women Incident disease CVD Pre-obese 1.17 (0.98 to 1.40) 1.23 (1.02 to 1.48) 1.07 (0.82 to 1.40) Obese Age (1.53 to 5.87) 3.07 (1.58 to 5.93) 1.08 (0.06 to 19.14) Age (1.05 to 4.02) 2.14 (1.10 to 4.13) 1.27 (0.07 to 22.66) MI Pre-obese 1.00 (0.72 to 1.39) 1.12 (0.80 to 1.56) 1.04 (0.64 to 1.68) Obese 2.72 (1.90 to 3.90) 2.99 (2.07 to 4.32) 1.03 (0.32 to 3.31) Stroke Pre-obese Age (0.12 to 1.52) 0.47 (0.14 to 1.63) 0.30 (0.02 to 3.63) Age (0.32 to 4.03) 1.26 (0.36 to 4.38) 1.04 (0.09 to 12.65) Obese 1.95 (1.25 to 3.06) 2.18 (1.38 to 3.43) 1.54 (0.47 to 5.00) Diseased to death CVD Pre-obese 0.89 (0.71 to 1.13) 0.97 (0.76 to 1.23) 0.98 (0.65 to 1.46) Obese 1.53 (1.17 to 1.99) 1.73 (1.32 to 2.27) 2.41 (1.19 to 4.92) MI*** Pre-obese 0.95 (0.64 to 1.43) 0.99 (0.66 to 1.50) 0.92 (0.47 to 1.82) Obese 1.88 (1.24 to 2.84) 1.95 (1.28 to 2.97) 2.93 (0.82 to 10.49) 1266 OBESITY Vol. 14 No. 7 July 2006

4 Stroke Pre-obese 1.69 (1.13 to 2.51) 1.85 (1.22 to 2.81) 2.46 (1.18 to 5.11) Obese 1.12 (0.68 to 1.86) 1.51 (0.88 to 2.59) 1.36 (0.28 to 6.48) Non-diseased to death CVD Pre-obese 1.09 (0.92 to 1.29) 1.24 (0.97 to 1.58) 1.17 (0.82 to 1.67) Obese 2.14 (1.73 to 2.64) 1.46 (1.00 to 2.14) 2.76 (1.59 to 4.77) HR, hazard ratio; CI, confidence interval; CVD, cardiovascular disease; MI, myocardial infarction. * Normal weight (18.5 BMI 25) category is the reference group. Crude hazard ratios. Hazard ratio adjusted for smoking status at baseline and number of cigarettes smoked/day. Hazard ratio adjusted for smoking smoking status and number of cigarettes smoked/day, calculated among participants free of hypertension, hypercholesterolemia, and diabetes at start of follow-up. Calculations made with 1130, 1011, and 421 participants age 45 at start of follow-up for the analyses, respectively. ** Calculations made with 657, 604, and 214 participants age 45 at start of follow-up for the analyses, respectively. Calculations made with 351, 307, and 118 participants age 45 at start of follow-up for the analyses, respectively. Calculations made with 157, 133, and 43 participants age 45 at start of follow-up for the analyses, respectively. Calculations made with 1421, 1290, and 762 participants age 45 at start of follow-up for the analyses, respectively. Calculations made with 627, 561, and 230 participants age 45 at start of follow-up for the analyses, respectively. *** Calculations made with 203, 150, and 89 participants age 45 at start of follow-up for the analyses, respectively. Calculations made with 178, 160, and 72 participants age 45 at start of follow-up for the analyses, respectively. Cox models were not proportional. HR for ages 50 and 70 years were calculated adding the interaction term age BMI category. Study Sample In this study, we followed participants from age 45 years. The participants may have reached this age at any point during the follow-up of the Framingham Heart Study. Start of follow-up in our study, therefore, did not correspond necessarily to Framingham Heart Study enrollment. Mean follow-up time was (standard deviation) years. Eligible participants were those enrollees in the Framingham Heart Study who were in the study and free of CVD at age 45 years (1215 men and 1532 women). Additionally, eligibility for this study required having at least one estimated measurement of weight and height and an estimation of smoking status at age 45 (1149 men and 1452 women). Weight, blood pressure, and serum cholesterol level at age 45 were estimated from the information collected through the biennial examinations. We estimated risk factor status at age 45 from linear regression on age using data from all of the exams within the 5 years before this age. Because physical activity was recorded at only 5 exams in the original Framingham Heart Study and in 859 (33.7%) participants in this study, it was not taken into account for this analysis. To assess smoking status at age 45 (never-smoker/ever-smoker/current smoker) and number of cigarettes smoked per day, we updated the information on smoking history at enrollment with the last available information on smoking status before age 45. Height was estimated from the closest measurement available, favoring those that were taken before age 45. Level of education was assessed according to the highest degree achieved at Framingham Heart Study enrollment. BMI was calculated as weight (in kilograms) divided by height (in meters squared). Subjects were classified as normal weight (18.5 BMI 25 kg/m 2 ), pre-obese (25 BMI 30 kg/m 2 ), or obese (BMI 30 kg/m 2 ) (16). Because the aim of this study was to assess the effect of overweight vs. normal weight, underweight participants (BMI 18.5 kg/ m 2 ) were excluded (10 men and 27 women). Participants who died in the first 2 years of follow-up were also excluded because their BMI may have changed recently because of impending mortality (9 men and 4 women). For the final analyses, 2551 (1130 men) participants were included, of whom 2301 (1011 men) had information on all potential confounders. Data Analysis To calculate number of years lived free of CVD and number of years lived with CVD, we constructed multistate life tables of any CVD, myocardial infarction (MI), and stroke, where any CVD includes coronary heart disease (acute MI, angina pectoris, and coronary insufficiency), cerebrovascular accident (stroke and transient ischemic attack), congestive heart failure, and intermittent claudication (17). In these models, we described three different states: OBESITY Vol. 14 No. 7 July

5 CVD free, with CVD, and death. The possible transition directions were from CVD free to death or to CVD and from CVD to death. Subjects were considered cardiovascularly diseased from the onset of CVD until death (18). Numbers of years lived free of CVD and with CVD are driven by the incidence of the disease, mortality after the cardiovascular disease, and mortality without prior cardiovascular disease. First, we assessed the increased risk associated with pre-obesity and obesity for each of the three transitions. Hazard ratios (HRs) for individuals with BMI between 25 and 30 kg/m 2 and BMI 30 kg/m 2 compared with normal weight subjects were calculated for death without prior cardiovascular disease, CVD incidence, and death among individuals with a history of cardiovascular disease. The resulting HRs are the basis of the multistate life tables. It is important that they represent the causative risk associated with pre-obesity and obesity as closely as possible. For this reason, we analyzed the HRs after adjusting or selecting for various potential confounders and intermediate variables. We performed Cox proportional hazards analyses with age as the time scale and stratified by sex. Next we adjusted for smoking status and number of cigarettes smoked per day at age 45 and for level of education. We also estimated HRs for preobese and obese subjects compared with normal weight individuals among participants without hypertension (systolic blood pressure 140 mm Hg and diastolic blood pressure 90 mm Hg) (19), with serum cholesterol 240 mg/dl (20), and without diabetes at age 45 to estimate the effect of obesity and pre-obesity before potential development of these risk factors. For all transitions, the hazards proportionality assumption was verified by testing the Schoenfeld residuals (21,22), with statistical significance set at 5%, using the statistical package Statistics/Data Analysis STATA 7 (StataCorp, College Station, TX) (23). These analyses were done for cardiovascular disease, MI, and stroke. Second, we used Poisson regression models to calculate smoothed age-specific transition rates for each of the transitions involved (10). We assessed graphically the goodness of fit of the model to the empirical rates for the different transitions. For those transitions where the hazard proportionality assumption held, we calculated the transition rates for the different BMI categories and ages, adjusted for smoking status and number of cigarettes smoked per day at age 45. Adjustment was performed only for smoking because we found that neither adjustment for education nor selection of a population free of hypertension, hypercholesterolemia, and diabetes appreciably decreased the relative risks associated with pre-obesity and obesity (see Results). When the HRs were not constant with age (in four of the cases of disease incidence), the transition rates were estimated by adding an interaction term between age at follow-up and BMI category in the model. Finally, we transformed the transition rates into probabilities and created multistate life tables for the total male and female populations, adjusted for smoking status. LE at age 90 was assumed to be a constant 4.53 years for men and 5.05 for women for each BMI category based on the life expectancies of the total Framingham Study sample (18). Calculations were done in S-plus 2000 (24), and confidence intervals were produced using a bootstrap procedure of 5000 replicates. In essence, the multistate life table represents men or women with the cohort s average smoking status at age 45 (men: 77.7% current smokers, mean 15.5 cigarettes/day; women: 51.7% current smokers, mean 7.64 cigarettes/day). While in a previous analysis we found that smoking status modified the association between pre-obesity (BMI 25 to 29.9 kg/m 2 ) and total mortality risk, the current study was underpowered to detect such interactions. To exclude any potential bias caused by smoking, we additionally calculated multistate life tables for never smokers. Results Men had a higher prevalence of overweight than women. At the start of follow-up, 78% of men and 52% of women were current smokers. Smoking status, blood pressure, serum cholesterol level, and education had the expected relationships with BMI (Table 1). We found that overweight was associated with an increased risk of incidence for all cardiovascular diseases studied and for death without previous CVD in analyses taking into account age and sex (Table 2). Obesity in women was also associated with an increased risk of death after cardiovascular disease. However, overweight in men did not increase the risk compared with those of normal weight, except for stroke, where pre-obesity seemed to have a protective effect. Adjustment for smoking status tended to slightly increase the HRs for incident disease. For women, but not men, adjustment tended to increase the HRs for the transitions from the diseased state to death (Table 2). The risk estimations for the transitions did not change appreciably after adjusting for education (data not shown). The relationship between BMI and incident disease declined somewhat after adjustment for baseline blood pressure, serum cholesterol level, and diabetes (data not shown). However, among subjects without hypertension, hypercholesterolemia, and diabetes at baseline, there was generally a slight increase or no change in the relative risks compared with the total population (Table 2), suggesting that the HRs related to preobesity and obesity were not caused by concurrent hypertension, hypercholesterolemia, or diabetes. Consequently, multistate life tables were calculated for men and women 1268 OBESITY Vol. 14 No. 7 July 2006

6 Table 3. Differences in life expectancy, in years, at age 45 for those who were pre-obese or obese compared with those who were normal weight (95% confidence intervals) Differences in total LE Differences in number of years lived free of disease Differences in number of years lived with disease Men CVD Pre-obese 0.9 ( 2.4; 0.7) 2.5 ( 4.0; 0.7) 1.5 (0.3; 2.8) Obese 3.3 ( 5.5; 1.3) 6.0 ( 8.1; 4.1) 2.7 (1.0; 4.4) MI Pre-obese 0.9 ( 2.4; 0.7) 1.5 ( 3.1; 0.2) 0.6 ( 0.3; 1.5) Obese 3.4 ( 5.6; 1.4) 4.7 ( 6.7; 2.4) 1.3 ( 0.1; 2.9) Stroke Pre-obese 1.1 ( 2.7; 0.3) 1.9 ( 3.5; 0.4) 0.8 (0.3; 1.4) Obese 3.4 ( 5.4; 1.3) 4.4 ( 6.4; 2.3) 1.0 (0.3; 1.8) Women CVD Pre-obese 1.3 ( 2.7; 0.3) 2.1 ( 3.7; 0.4) 0.8 ( 0.4; 2.1) Obese 6.9 ( 9.3; 5.0) 8.4 ( 10.8; 6.2) 1.4 ( 0.3; 3.2) MI Pre-obese 1.3 ( 2.8; 0.3) 1.3 ( 2.8; 0.2) 0.0 ( 0.6; 0.6) Obese 7.1 ( 9.1; 4.8) 7.7 ( 9.7; 5.6) 0.6 ( 0.1; 2.0) Stroke Pre-obese 1.4 ( 2.7; 0.2) 1.5 ( 2.8; 0.1) 0.1 ( 0.4; 0.7) Obese 7.0 ( 9.4; 5.1) 7.2 ( 9.5; 5.2) 0.2 ( 0.5; 1.1) LE, life expectancy; CVD, cardiovascular disease; MI, myocardial infarction. representing the population distribution of smoking using transition rate estimation models adjusted for smoking status and number of cigarettes smoked and for never smokers alone. The association between BMI and the risk of each transition was translated into differences in number of years lived without and with disease (Table 3; Figures 1 and 2). Obese men at age 45 expected to live a total of 3.3 years (95% CI, 1.3; 5.5) fewer than normal weight subjects. Obese women expected to have 6.9 years (95% CI, 5.0; 9.3) shorter LE. Overweight was also associated with fewer number of years free of CVD (Table 3; Figures 1 and 2): 6.0 years (95% CI, 4.1; 8.1) fewer for obese men and 8.4 years (95% CI, 6.2; 10.8) fewer for obese women free of CVD at age 45. Differences in number of years lived without CVD were larger for women than for men. On average, obese men developed CVD or died at age 65.0 (95% CI, 63.4; 66.7) years, whereas normal weight subjects became diseased or died at age 71.0 (95% CI, 69.8; 72.2) years. Obese women expected to get CVD or to die at age 69.0 years (95% CI, 66.7; 71.1), whereas for normal weight women. Number of years lived without CVD was 77.3 years (95% CI, 76.4; 78.3). Overweight at age 45 was associated with greater number of years lived with all CVD types, although these were statistically significant only for CVD and stroke in men (Table 3; Figures 1 and 2). The additional years lived with CVD were 2.7 (95% CI, 1.0; 4.4) years for obese men and 1.4 (95% CI, 3.2; 0.3) years for obese women (Table 3). Obesity was associated with a greater number of years lived with CVD in men than women. Differences in total LE, number of years lived free of CVD, and number of years lived with CVD between BMI categories for male and female never smokers are presented in Table 4. For male never smokers, differences in total LE among the BMI categories were smaller compared with those found for average male smokers, whereas for number of years lived free of CVD and number of years lived with CVD, those differences were greater among never smokers OBESITY Vol. 14 No. 7 July

7 Figure 1: Number of years lived from age 45 with and without CVD, MI, or stroke for men without CVD at baseline. than average smokers. For female never smokers, all differences among the BMI categories were similar to those found for the average smokers. Discussion Overweight at age 45 was associated with fewer number of years lived free of any cardiovascular disease, MI, and stroke and with an increase in the number of years lived with CVD compared with those with normal weight. On average, obesity was associated with 6.0 fewer years lived free of any CVD for men and 8.4 years for women. Furthermore, obese men and women lived an extra 2.7 and 1.4 years with any cardiovascular disease, respectively. Number of years lived free of CVD is a composite health measure resulting from a combination of two major effects: incidence of CVD and mortality in those without previous cardiovascular disease. In this analysis, the increased risk of incident disease for obese and pre-obese subjects compared with their normal weight counterparts was reflected in an earlier occurrence of disease and a higher proportion of cases throughout life. Additionally, the increased risk of mortality in those without previous CVD resulted in a decreased number of years lived and, consequently, an even shorter number of years lived free of CVD. Number of years lived with CVD is also a composite measure. In isolation, a higher incidence of CVD would lead to a greater number of years lived with disease, whereas a higher mortality after CVD and higher mortality in those without previous CVD would reduce the number of years lived with cardiovascular disease. We showed that obesity at age 45 increased the risk of any cardiovascular disease, MI, and stroke and the risk of mortality in those without previous cardiovascular disease, but had little effect on mortality after subsequent onset of CVD in men. The finding that obesity was associated with a greater number of years lived with CVD suggests that differences in number of years lived with CVD were driven largely by the increased risk of incident disease associated with obesity, exceeding the effect of the higher relative risks of mortality. The HRs for disease incidence were quite comparable with other studies (4,5,25 27). Unfortunately, to our knowledge, no other studies have examined the association between obesity and all-cause mortality in a population with cardiovascular disease. In our study, this transition was associated with the smallest numbers, and, therefore, with significant uncertainty. Consequently, it is not possible to determine whether apparent findings such as the sex differences found in the relative risk of progression from CVD to death, and particularly the possible protective effect of BMI between 25 and 30 kg/m 2 in men, represent true associations until further research is done in other populations. However, the impact of this uncertainty on the final LE results is currently reflected by the associated confidence intervals. It is not possible to fully establish to what extent the associations between obesity and number of years lived free 1270 OBESITY Vol. 14 No. 7 July 2006

8 Figure 2: Number of years lived from age 45 with and without CVD, MI, and stroke for women without CVD at baseline. of CVD and with CVD are causal. The existing association between obesity and other cardiovascular risk factors could introduce a confounding effect. To describe better the temporal relation between obesity and potential intermediate states, we assessed the risk associated with increasing BMI for all transitions among individuals free of hypertension, hypercholesterolemia, and diabetes at baseline. Restricting the analyses to this population free of other risk factors at baseline did not generally decrease the relative risks associated with obesity. We cannot rule out that some of the differences in number of years lived with and without CVD between those who were obese and those who were normal weight at age 45 are caused by potential confounding by factors such as physical activity, which we were unable to include in our analysis. The differences in number of years lived free of CVD and number of years lived with CVD presented here were estimated from multistate life table analyses using the disease and mortality experience from the Framingham Heart Study. As a result, our calculations were done on highquality data on the occurrence of CVD and mortality combined with accurate measurements of body weight and height. In addition, our outcome measure took directly into account the competing risks between death and disease incidence. The use of information from 46 years of follow-up instead of short-term follow-up data allows more accurate conclusions on the association between increased BMI during adulthood and LE (11 13). Nevertheless, our analyses had some limitations. First, the long-term implications of obesity on CVD were based on the experience of a particular cohort over a long period of time. As a consequence, the disease incidence and mortality rates differ from those experienced in the general population today. However, as stated above, our relative risks of incident CVD and mortality were similar to those reported in other studies (5,25 27). Second, we found that exponentially modeling the transition rates from history of cardiovascular disease to death somewhat underestimates the case fatality at younger ages. However, this did not have a large effect on the final outcomes because of the small number of events at younger ages (data not shown). Finally, our study was underpowered to either detect an interaction or perform a stratification by smoking status (10). While our analysis in the never smoker population showed a greater effect of overweight on the number of years lived with and without CVD in men, the small numbers of never smoking men in each health state make such comparisons relatively unreliable. Differences in LE associated with pre-obesity and obesity in women were very similar for never smokers compared with average smokers. OBESITY Vol. 14 No. 7 July

9 Table 4. Differences in life expectancy, in years, at age 45 for those who were pre-obese or obese never smokers compared with those who were normal weight never smokers (95% confidence intervals) Differences in total LE Differences in number of years lived free of disease Differences in number of years lived with disease Men CVD Pre-obese 1.8 ( 5.5; 1.5) 4.6 ( 9.0; 1.2) 2.9 ( 1.8; 3.4) Obese 2.1 ( 8.4; 1.4) 10.0 ( 16.8; 5.6) 7.9 (5.4; 12.9) MI Pre-obese 2.3 ( 2.2; 0.8) 3.1 ( 3.2; 0.1) 0.8 ( 0.4; 1.5) Obese 1.7 ( 4.7; 0.3) 6.1 ( 7.2; 2.6) 4.3 (6.4; 9.3) Stroke Pre-obese 3.1 ( 7.3; 0.3) 3.1 ( 8.1; 0.0) 0.0 ( 0.8; 0.7) Obese 1.3 ( 5.9; 0.7) 3.1 ( 6.2; 1.1) 1.8 (0.0; 6.3) Women CVD Pre-obese 1.9 ( 3.8; 1.1) 2.8 ( 4.0; 0.1) 0.9 ( 0.4; 2.3) Obese 6.6 ( 8.4; 5.3) 8.1 ( 9.5; 6.2) 1.5 (0.2; 3.0) MI Pre-obese 1.5 ( 14.3; 17.3) 1.8 ( 2.9; 0.4) 0.4 ( 0.6; 0.7) Obese 6.8 ( 3.5; 7.3) 7.8 ( 8.7; 5.2) 1.2 (0.1; 3.2) Stroke Pre-obese 1.2 ( 2.5; 0.2) 1.4 ( 2.6; 0.0) 0.2 ( 0.3; 0.7) Obese 6.4 ( 8.4; 4.4) 6.7 ( 8.7; 4.5) 0.2 ( 0.3; 1.1) LE, life expectancy; CVD, cardiovascular disease; MI, myocardial infarction. These findings suggest that the impact of weight management could be of particular relevance for male never smokers. This study suggests that, among men without CVD at baseline, obese individuals lived 3.3 (95% CI, 1.3; 5.5) years fewer from age 45 than those who were normal weight. A previous study using a slightly different sample from the same cohort showed an LE difference from age 40 between obese and normal weight of 5.8 (1.4 to 10.5) years in non-smoking men and 6.7 (4.6 to 8.8) years in smoking men (10). The total LE differences for women were similar in the two studies. The differences between the two studies for the men seem to be caused by the different samples selected, probably reflecting both better health in the preobese and obese in this study because of including only survivors free of CVD at age 45, and, presumably, also improvements in the treatment of CVD over time, and a slight attenuation in the effect of BMI when measured at an older age. In this case, the previous results may be an overestimation of the effect of obesity in adulthood in LE for men today. The obesity epidemic continues, despite substantial research in the area. While a range of strategies, from social policy to patient care, is required to halt this epidemic (16,28), accurate and meaningful measures of risk are required to underpin all strategies. It is known that different methods of communicating risk result in different perceptions of risk (29). Here we present a relatively new method of risk presentation: that of number of years lived with and without CVD. For risk factors such as obesity, whose downstream effects are multiple and extend into old age, it will be increasingly important to understand their effects across the life course, presenting both the time to disease and the time lived with disease. The prevalence of obesity has increased in the last two decades. It has been successfully identified as a major modifiable risk factor for CVD and a predictor of substantially reduced LE. We have now shown that obesity and pre-obesity before middle age are associated with an even greater reduction in number of years lived free of CVD and an increase in number of years lived with CVD. These 1272 OBESITY Vol. 14 No. 7 July 2006

10 findings constitute a solid basis for well-informed public and clinical prioritization of obesity prevention. Acknowledgments The authors thank the Framingham Heart Study coordinators for access to the original data set. The Framingham Heart Study is conducted and supported by the National Heart, Lung, and Blood Institute in collaboration with the Framingham Heart Study Investigators. This manuscript has been reviewed by the National Heart, Lung, and Blood Institute for scientific content and consistency of data interpretation with previous Framingham Heart Study publications, and significant comments have been incorporated before submission for publication. This study was performed within the context of the Netherlands Epidemiology and Demography Compression of Morbidity Research Group, whose other members include J. Barendregt, L. Bonneux, F. Janssen, A. Kunst, J. Mackenbach, and F. Willekens. This study was supported by grants from the Netherlands Heart Foundation and the Netherlands Organization for Scientific Research, and was part of M.C.P. s MSc Training Program in Clinical Epidemiology at The Netherlands Institute for Health Sciences. A.P. was partly funded as a VicHealth Research Fellow during this study. References 1. National Task Force on the Prevention and Treatment of Obesity. Overweight, obesity, and health risk. Arch Intern Med. 2000;160: Mokdad AH, Bowman BA, Ford ES, Vinicor F, Marks JS, Koplan JP. The continuing epidemics of obesity and diabetes in the United States. JAMA. 2001;286: Mokdad AH, Ford ES, Bowman BA, et al. Prevalence of obesity, diabetes, and obesity-related health risk factors, JAMA. 2003;289: 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: Manson JE, Colditz GA, Stampfer MJ, et al. A prospective study of obesity and risk of coronary heart disease in women. N Engl J Med. 1990;322: Kenchaiah S, Evans JC, Levy D, et al. Obesity and the risk of heart failure. N Engl J Med. 2002;347: Kurth T, Gaziano JM, Berger K, et al. Body mass index and the risk of stroke in men. Arch Intern Med. 2002;162: Hubert HB, Feinleib M, McNamara PM, Castelli WP. Obesity as an independent risk factor for cardiovascular disease: a 26-year follow-up of participants in the Framingham Heart Study. Circulation. 1983;67: Kannel WB, Wilson PW, Nam BH, D Agostino RB. Risk stratification of obesity as a coronary risk factor. Am J Cardiol. 2002;90: Peeters A, Barendregt JJ, Willekens F, Mackenbach JP, Al Mamun A, Bonneux L. Obesity in adulthood and its consequences for life expectancy: a life-table analysis. Ann Intern Med. 2003;138: Fontaine KR, Redden DT, Wang C, Westfall AO, Allison DB. Years of life lost due to obesity. JAMA. 2003;289: Thompson D, Edelsberg J, Colditz GA, Bird AP, Oster G. Lifetime health and economic consequences of obesity. Arch Intern Med. 1999;159: Peeters A, Bonneux L, Barendregt J, Nusselder W. Methods of estimating years of life lost due to obesity. JAMA. 2003;289: Visscher TL, Rissanen A, Seidell JC, et al. Obesity and unhealthy life-years in adult Finns: an empirical approach. Arch Intern Med. 2004;164: Dawber TM, G. Moore F. Epidemilogical approaches to heart disease: the Framingham Study. Am J Public Health. 1951;41: World Health Organization. Obesity: Preventing and Managing the Global Epidemic. Report of a WHO Consultation. Geneva, Switzerland: World Health Organization; Shurtleff D. Some characteristics related to the incidence of cardiovascular disease and death: Framingham study 16 years follow-up. In: Kannel WGT, ed. The Framingham Study: An Epidemiological Investigation of Cardiovascular Disease. Washington, DC: Government Printing Office; Peeters A, Mamun AA, Willekens F, Bonneux L. A cardiovascular life history. A life course analysis of the original Framingham Heart Study cohort. Eur Heart J. 2002;23: Chobanian AV, Bakris GL, Black HR, et al. The seventh report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure: the JNC 7 report. JAMA. 2003;289: National Cholesterol Education Program. Executive Summary of the Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on detection, evaluation, and treatment of high blood cholesterol in adults (Adult Treatment Panel III). JAMA. 2001;285: Grambsch PTT. Proportional hazards tests and diagnostics based on weighted residuals. Biometrika. 1994;81: Schoenfeld D. Partial residuals for the proportional hazards regression model. Biometrika. 1982;69: STATA, version 7.0. College Station, TX: Statacorp LP; S-Plus, Professional Edition, Version Release 1 for Windows. Seattle, WA: Mathsoft, Inc.; Seidell JC, Visscher TL, Hoogeveen RT. Overweight and obesity in the mortality rate data: current evidence and research issues. Med Sci Sports Exerc. 1999;31(11 Suppl):S Calle EE, Rodriguez C, Walker-Thurmond K, Thun MJ. Overweight, obesity, and mortality from cancer in a prospectively studied cohort of U.S. adults. N Engl J Med. 2003;348: Calle EE, Thun MJ, Petrelli JM, Rodriguez C, Heath CW Jr. Body-mass index and mortality in a prospective cohort of U.S. adults. N Engl J Med. 1999;341: World Health Organization. Obesity: Preventing and Managing the Global Epidemic. Report of a WHO Consultation. Geneva, Switzerland: World Health Organization Tech Rep Ser, 2000; pp. i xii, 1 253, Edwards A, Elwyn G, Mulley A. Explaining risks: turning numerical data into meaningful pictures. BMJ. 2002;324: OBESITY Vol. 14 No. 7 July