Body mass index and mortality from all causes and major causes in Japanese: results of a pooled analysis of 7 large-scale cohort studies

J Epidemiol. 2011;21(6):417-30. doi: 10.2188/jea.je20100180. Epub 2011 Sep 10.

Abstract

Background: We pooled data from 7 ongoing cohorts in Japan involving 353 422 adults (162 092 men and 191 330 women) to quantify the effect of body mass index (BMI) on total and cause-specific (cancer, heart disease, and cerebrovascular disease) mortality and identify optimal BMI ranges for middle-aged and elderly Japanese.

Methods: During a mean follow-up of 12.5 years, 41 260 deaths occurred. The Cox proportional hazards model was used to estimate hazard ratios (HRs) for each BMI category, after controlling for age, area of residence, smoking, drinking, history of hypertension, diabetes, and physical activity in each study. A random-effects model was used to obtain summary measures.

Results: A reverse-J pattern was seen for all-cause and cancer mortality (elevated risk only for high BMI in women) and a U- or J-shaped association was seen for heart disease and cerebrovascular disease mortality. For total mortality, as compared with a BMI of 23 to 25, the HR was 1.78 for 14 to 19, 1.27 for 19 to 21, 1.11 for 21 to 23, and 1.36 for 30 to 40 in men, and 1.61 for 14 to 19, 1.17 for 19 to 21, 1.08 for 27 to 30, and 1.37 for 30 to 40 in women. High BMI (≥27) accounted for 0.9% and 1.5% of total mortality in men and women, respectively.

Conclusions: The lowest risk of total mortality and mortality from major causes of disease was observed for a BMI of 21 to 27 kg/m(2) in middle-aged and elderly Japanese.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Adult
  • Aged
  • Aged, 80 and over
  • Body Mass Index*
  • Cause of Death
  • Cerebrovascular Disorders / mortality*
  • Female
  • Follow-Up Studies
  • Heart Diseases / mortality*
  • Humans
  • Japan / epidemiology
  • Male
  • Middle Aged
  • Neoplasms / mortality*
  • Proportional Hazards Models
  • Risk Assessment