Associations Between Physical Activity Profiles Over 46 Years And Cognitive Health At Age 90

Population aging and related cognitive declines at old age are major public health issues. Based on cross-sectional and short follow-up studies, physical activity (PA) could potentially delay the decline of cognitive function, but more longitudinal studies are needed. PURPOSE: To identify PA profiles from midlife to age 90 and examine whether cognitive health characteristics at age 90 are associated with the PA profiles identified. METHODS: In total, 125 participants (81 women, 44 men) from the population-based older Finnish Twin Cohort were included in the study. The participants responded to health surveys at the mean ages of 45, 52, 59 and 91. The self-reported PA was provided by the surveys. Cognitive health was assessed at the mean age of 91 (SD 1.54); the modified Telephone Interview for Cognitive Status, including three learning trials of word list (TICS-m3), specific measures of episodic memory (immediate and delayed recall of word list) and semantic fluency (1-min animal naming) were used. K-means clustering for longitudinal data was used to identify PA profiles and generalized linear mixed-effects models to investigate cognitive differences between the PA profile groups. RESULTS: We found three PA profiles from midlife to age 90: 1) Low stable (51%), 2) Low increasing moderate (25%), and 3) Moderate decreasing low (23%). Cognitive characteristics at age 90 did not differ between the PA profiles, except the total TICS-m3 scores. Participants in the Low stable PA profile group had significantly lower total TICS-m3 scores (mean 37.1) than participants in the Low increasing moderate profile group (mean 39.1) (BH corrected p < 0.016), but the association did not remain significant after we adjusted for education. CONCLUSION: Three distinct 46-year PA profiles were discovered. The largest proportion of participants belonged to the Low stable profile, characterized by a stable low level of PA throughout the follow-up. When age and sex were considered, the consistently physically inactive participants were cognitively more impaired than those who increased their PA level over the years from low to moderate. After education was taken into account, no association was detected. However, this result may be due to low statistical power and, thus, larger sample size studies on the topic are warranted.