Background: The relationship between relative body weight and mortality has been well studied in Western populations and remains controversial. Little is known about the weight-mortality association in less well fed people in developing countries.
Methods: A cohort of 18,244 Chinese men aged 45-64 years in Shanghai, China enrolled in a prospective study of diet and cancer during January 1986 through September 1989. At recruitment, height and usual body weight were collected through interview. An active, annual follow-up of the cohort was conducted for cancer and death. Proportional hazards regression method was used to examine the relation between body mass index (BMI, weight in kg/height in m2) and overall and cause-specific mortality.
Results: By 28 February 1995, 1198 deaths (498 from cancer, 422 from cardio- and cerebrovascular disease, and 278 from other causes) had been identified. We found a U-shaped relation between BMI and total mortality among lifelong non-smokers. Compared with non-smokers with BMI 21.0-<23.5, the relative risk (RR) for all cause mortality was 1.73 (95% confidence interval [CI]: 1.23-2.42) for men with BMI <18.5 and 1.48 (95% CI: 1.07-2.03) for men with BMI > or =26 after adjustment for age, level of education, and alcohol drinking. The elevated risk of death in men with BMI > or =26 was largely due to fatal cardio- and cerebro-vascular diseases. There was a nearly twofold excess risk of death from cardio- and cerebrovascular diseases among lifelong non-smokers with BMI > or =26 compared with non-smokers with BMI 21-<23.5 (95% CI: 1.17-3.22). On the other hand, the increased overall mortality risk in men with BMI <18.5 was primarily due to causes of an infectious origin. Exclusion of the first 4 years of follow-up data did not materially alter the BMI-mortality associations.
Conclusion: Underweight and overweight both are associated with an increased risk of death in middle-aged Chinese men who never smoked cigarettes. The increased total mortality in overweight men is largely due to cardio- and cerebro-vascular diseases while the elevated risk of death in underweight men is attributed primarily to causes of an infectious nature.