Objective: Controversy regarding the relationship between body mass index (BMI) and mortality continues to exist. Most of the previous studies have not comprehensively accounted for major sources of bias. We examined the association between BMI and all-cause mortality according to pre-existing disease and smoking status in a large prospective cohort.
Methods: Participants were 99 253 male physicians in the Physicians' Health Study enrollment cohort (40-84 years) who provided self-reported information in 1982. We used Cox proportional hazards regression to examine the association between baseline BMI and mortality.
Results: A total of 5438 men died (median follow-up, 5.7 years). Although a U-shaped association between BMI and all-cause mortality was seen among all men, we found a linear relationship when accounting for potential sources of bias. In the optimal model excluding men who died within 2 years, and adjusting for age, smoking, alcohol consumption, physical activity, prior disease and interactions between BMI and prior disease, and between BMI and smoking, those with BMI <20.0 kg/m(2) had a relative risk (RR) of death of 0.88 (95% confidence interval (CI), 0.56-1.40), as compared to men with BMI 22.5-24.9 kg/m(2). By contrast, men with BMI 30.0-34.9 kg/m(2) had an RR of 1.45 (95% CI, 1.10-1.91) and those with BMI >or=35.0 kg/m(2) had an RR of 1.62 (95% CI, 1.12-2.35; P for linear trend, <0.001). According to WHO categories, men in the 'overweight' range (BMI 25.0-29.9 kg/m(2)) had an RR of 1.20 (95% CI, 1.05-1.38) as compared to men in the 'normal' range (BMI <25.0 kg/m(2)).
Conclusions: In this large, prospective cohort, we found a consistent linear association between higher BMI and increased risk of mortality after accounting for several potential sources of bias, even among those within the 'overweight' range of BMI. Public health messages should emphasize the preponderance of evidence supporting the adverse health effects associated with higher body weight.