BACKGROUND/OBJECTIVES: the association between late-life obesity and late-life cognitive abilities is poorly understood. We studied the association between body mass index (BMI) and cognitive change in longitudinal population-based study spanning over the ninth decade of life.

SUBJECTS/METHODS: in total, 475 participants free of dementia at baseline from the Lothian Birth Cohort 1921 (mean age: 79.1 years, SD: 0.6) were included. Height and weight were assessed at baseline. BMI was calculated as kg/m(2). Cognitive abilities were assessed at age ∼11 years and at age ∼79, ∼83, ∼87 and ∼90 years.

RESULTS: latent growth models showed that men being overweight and obese had a 0.65 (SD: 0.3) and 1.10 (SD: 0.5) points less steep decline in general cognitive ability (as measured by the Moray House Test) for each year than people of normal weight. These associations were to some extent confounded by childhood intelligence. No other association between BMI and cognition was significant, either for men or women. People who were obese in old age had significantly lower childhood intelligence (m = 43.6, SD: 1.3) than people who were normal in weight (m = 47.0, SD: 0.8) and persons being overweight (m = 47.5, SD: 0.8), F (472, 3) = 3.2, P = 0.043.

CONCLUSIONS: the current study shows weak or no evidence for an association between BMI in old age and cognitive function, especially not when childhood intelligence is controlled for. Lower intelligence at the age of 11 years predicted obesity at the age of 79 years.