O3‐05‐03: GAINS FOR BRAINS: EVIDENCE FOR SEX‐DEPENDENT EFFECTS OF EXERCISE

Physical activity (PA) is a promising strategy to mitigate the deleterious effects of aging and disease on brain health. However, a large amount of variation exists in its efficacy. Given the greater prevalence and faster progression of Alzheimer's disease in women compared to men, and known sex differences in brain architecture, analysis of sex differences in the relationship between PA and cognition, including underlying mechanisms, is warranted. To address this, we conducted secondary analyses of data from two studies: 1) PROMoTE – a randomized controlled trial (RCT) of 6-month, thrice-weekly aerobic training (AT) in older adults with mild vascular cognitive impairment, and 2) Health ABC – a 10-year, longitudinal, epidemiological cohort study of 2873 older white and black adults with retrospective PA measured through self-reported walking behaviour. A subset of 313 Health ABC participants underwent magnetic resonance imaging at year 10. Analysis of covariance and latent growth curve modeling were utilized to examine the interaction between sex and AT or PA, controlling for age, study site, education, body mass index, diabetes, cardiovascular and cerebrovascular disease, brain atrophy, and total gray matter where appropriate. We focused on executive functions as meta-analyses suggest that aerobic exercise has larger benefits for this cognitive domain. In the PROMoTE RCT, we found that compared to usual care control, AT significantly improved executive functioning in females, but not males, an effect that was still evident 6 months later (p's<0.05). Further, AT increased levels of brain derived neurotrophic factor in females compared to males (p<0.01). In the Health ABC data, higher initial PA and increased PA over time were associated with less decline in executive functioning over time in females (p's<0.05), but not males (see Figure 1). Interestingly, maintenance of PA over time was related to larger left dorsal lateral prefrontal cortical volume in females only (p<0.05). Standardized results of latent growth curve model assessing longitudinal relationships among physical activity (walking) and executive functioning (DSST). To reduce model complexity, covariates (age, study site, education, body mass index, diabetes, cardiovascular and cerebrovascular disease) are not shown. Unstandardized estimates, except for path C, are shown for females in red and males in blue. Values must be less than 0.05 in both unadjusted and adjusted models to be significant. DSST = Digit Symbol Substitution Test. *p < 0.05; **p<0.01. Together, these findings provide evidence that sex differences exist in the effect of PA on cognition as well as in the underlying neurobiological mechanisms. This new knowledge can foster development of efficacious and sex-specific exercise recommendations to promote brain health.