Body mass index, waist circumference, and mortality risks over 27 years of follow-up in old age

Type Article

Journal Article


P. S. O'Súilleabháin; A. R. Sutin; D. Gerstorf

Year of publication



Ann Epidemiol







PURPOSE: This study investigates the predictive effects of body mass index (BMI) and waist circumference (WC) for all-cause mortality in old age over 27 years of follow-up. METHODS: Participants were from the Berlin Aging Study (n = 444, M ± SD = 84.55 ± 8.38 years). Reported significance values for hazard ratios were adjusted for age, sex, socioeconomic status, smokers, depressive illness, functional status, cholesterol, and objectively assessed physical diseases. RESULTS: BMI emerged as a significant predictor of all-cause mortality after adjustment. WC was not a significant predictor of mortality either within the unadjusted model or when fully adjusted, including BMI. After the introduction of WC alongside all covariates, the effect for BMI remained significant. A significant quadratic effect for BMI and mortality within the fully adjusted model also emerged. CONCLUSIONS: Our results suggest that lower BMI in the oldest old is associated with increased mortality hazards, and risks were particularly elevated for people who are underweight. This study found no evidence that higher BMI in old age is associated with increased mortality hazards relative to normal weight ranges.