Abstract

Background: The aim of this study was to investigate the associations between long-term weight change after age 20 and overall mortality and cause-specific mortality in the general Asian population.

Methods: From 1990 to 2005, the Japan Public Health Center (JPHC)-based prospective study conducted a follow-up of 42 242 men and 46 177 women aged 40–69 years with no history of cardiovascular disease (CVD) or cancer. Sex-specific multivariable-adjusted hazard ratios for cause-specific mortality were computed in accordance with weight change categories from age 20, as assessed by a self-administered questionnaire, and clustered by the JPHC communities and age groups, using Cox’s proportional hazard model.

Results: During the 12.9-year follow-up, there were 6494 deaths, including 2888 from cancer, 1011 from CVD and 2595 from other causes. In all, weight loss ⩾5 kg since age 20 increased hazard ratios for all-cause mortality in men (1.44, 95% CI 1.32 to 1.56) and women (1.33, 95% CI 1.17 to 1.52) compared with maintenance of a stable weight, and elevated risk was also found within each age group. The risk of weight loss was higher for individuals in the younger age group. Weight loss predicted cancer and CVD mortality only for men ⩾50 years of age. The increased risk was observed regardless of whether the individual was ill, a smoker or overweight at baseline or at age 20. There was an inverse association between weight gain and mortality risk.

Conclusion: Weight loss strongly predicted all-cause, cancer and CVD mortality, primarily for men. An unfavourable effect of weight gain was small at the population level.