The relation of coronary disease, stroke, and mortality to weight in youth and in middle age.

During ten years of follow-up 701 deaths occurred among 8006 45-68-year-old Honolulu Japanese men who were originally examined in 1965-68. Mortality was highest in both the leanest and fattest men and lowest in the fourth quintile of body mass index (BMI). The excess deaths in the top quintile were due primarily to coronary heart disease which was directly and positively related to BMI. In the bottom two quintiles excess mortality was due to cancer and "other" causes and was confined to men who had lost weight between age 25 and examination. This finding suggests that excess mortality in thin middle-aged men is due to occult antecedent disease rather than leanness per se. When BMI at age 25 was considered, men in the bottom quintile had the lowest mortality in middle age. It is concluded that upward revision of tables of desirable weight for young men would be unwarranted. A modest increase in the upper limit in middle-aged men may be justifiable, but the evidence is incomplete.