To examine the longitudinal association between body mass index (BMI) and waist circumference (WC) with mortality and incident disability in Lc65+ cohort.

Population-based cohort of non-institutionalized adults with up to 8.9 years of follow-up.

City of Lausanne, Switzerland.

1,293 individuals aged 65 to 70 at baseline (58% women).

BMI, WC and covariates were measured at baseline in 2004-2005. Vital status was obtained up to the 31st December 2013 and difficulty with basic activities of daily living (BADL) was reported in a self-administered questionnaire sent to participants every year. Main outcomes were total mortality and disability, defined as difficulty with BADL for ≥2 years or institutionalization. Cox regression was used with BMI/WC quintiles 2 as the reference.

130 persons died over a median follow-up of 8.47 years (crude mortality rate, men: 16.5/1,000 person-years, women: 9.7/1,000 person-years). In Cox regression adjusted for age, sex, education, financial situation, smoking and involuntary weight loss (IWL) at baseline, mortality was significantly associated with neither BMI nor WC, but there were trends towards non-significant J-curves across both BMI and WC quintiles. Disability (231 cases) tended to increase monotonically across both BMI and WC quintiles and was significantly associated with BMI quintile 5 (HR=2.44, 95% CI [1.65-3.63]), and WC quintiles 4 (HR=1.81 [1.15-2.85]) and 5 (HR=2.58, [1.67-4.00]).

Almost half of the study population had a substantially increased HR of disability, as compared to the reference BMI/WC categories. This observation emphasizes the need for life-long strategies aimed at preventing excess weight, muscle loss and functional decline through adequate nutrition and regular physical activity, starting at early age and extending throughout life.