[Effect of body mass index on morbidity and mortality in a general Japanese population--the Hisayama study].

In order to investigate the effect of body mass index (BMI, kg/m2) on life expectancy, 2053 Hisayama residents, aged 40 years or older were studied for 13 years from 1974. During the follow-up period, 419 subjects died; of these, 39 deaths due to accident or suicide were excluded from further analysis. On initial examination, male subjects with BMI > or = 27 had significantly higher age-adjusted prevalence rates of hypertension, glucose intolerance, hypercholesterolemia, electrocardiogram abnormalities and proteinuria, as compared with those with 23-25 BMI. In contrast, the frequency of male smokers was inversely associated with BMI levels. Female prevalence rates of glucose intolerance, hypercholesterolemia and proteinuria were significantly higher in 25-27 BMI than in 23-25. Body Mass Index showed a U-shaped relationship with all cause mortality rates with the lowest rate in 23-27 BMI for men and in 23-25 BMI for women. These associations remained substantially unchanged, even after controlling for age, systolic blood pressure, glucose intolerance, serum cholesterol, proteinuria, electrocardiogram abnormalities, alcohol consumption, and smoking habits. When analyzing the BMI mortality relationship by cause of death, age- and sex-adjusted mortality rates from myocardial infarction and stroke significantly increased in subjects with BMI > or = 27 compared with those with 23-25 BMI. In contrast, there was a decreasing risk of death from malignant neoplasms with rising BMI levels, but the relationship was not significant. Mortality from pneumonia and other causes showed a U-shaped relationship with significantly higher rates seen in BMI < 19 than in 23-25 BMI. These data indicate that BMI has a U-shaped relationship with total mortality in the general Japanese population, which results from various associations between BMI and cause-specific mortality rates.