Individual and combined associations of physical activity and cognitive function with all-cause mortality in older men and women: a prospective analysis of the German National Cohort (NAKO)

Low physical activity (PA) and poor cognitive function are associated with higher mortality risks. However, little is known about their interaction, including whether PA may moderate cognition-related mortality risks. This study examines the combined associations of PA and cognition with all-cause mortality, with attention to sex differences. Using data from the German National Cohort and its mortality follow-up, we analyzed mortality risk based on: a) baseline low vs. sufficient PA (assessed via the global physical activity questionnaire using a threshold of < vs. ≥ 600 MET-minutes/week), b) baseline low vs. medium vs. high semantic memory (SM) and executive function/processing speed (EF/PS), assessed through factor analyses of a neurocognitive test battery, and c) their interaction on mortality in individuals aged 65 + up to 10 years of follow-up (N = 28,892). Cox models were estimated both in the total sample and stratified by sex, adjusting for relevant confounders and reporting both distinct and combined associations. During follow-up, 1,605 individuals (5.6%) died: 1,097 men (7.5%) and 508 women (3.6%). Compared to individuals with low cognitive function, those with high SM (Hazard Ratio (HR) = 0.83 [95%CI: 0.67–1.02]), as well as high EF/PS (HR = 0.66 [0.53–0.83]) and medium EF/PS (HR = 0.68 [0.60–0.78]) had lower mortality risks. PA was associated with a 29% decreased mortality risk (HR = 0.71 [0.62–0.82]) compared to low PA. PA moderated the elevated risk from low cognition, with regard to EF/PS (low EF/PS*PA: HR = 0.65 [0.50–0.84] vs. low EF/PS*low PA: HR = 1 (ref.)) and SM (low SM*PA: HR = 0.60 [0.46–0.77] vs. low SM*low PA: HR = 1 (ref.)). The associations did not differ between men and women. Maintaining cognitive function and PA in older age is relevant for reducing mortality risk in both men and women. PA may offset risks linked to low cognition in both sexes, though mechanisms require further study. Clinical trial number: not applicable.