Higher ambient air pollution, such as particulate matter 2.5 (PM2.5), has been linked to increased stroke risk and mortality in prior research. It may also lead individuals to decrease physical activity, creating an indirect behavioral path linking PM2.5 to stroke. Our overall objective was to evaluate whether physical activity mediates the path between PM2.5 and stroke. As an initial step in mediation analysis, we estimated the total effect of PM2.5 on incident stroke and mortality. We included participants from the REasons for Geographic and Racial Differences in Stroke (REGARDS) Physical Activity Accelerometer Ancillary Study (2009–2013; n=7,663; mean age 57.6 years; 57% women, 31% Black) followed through 2022. Residential exposure to annual PM2.5 was dichotomized as upper quartile vs. lower three quartiles. We fit pooled logistic regression models adjusted for baseline covariates: age, sex, race, education, household income, marital status, urbanicity, neighborhood SES, smoking, and health conditions (diabetes, hypertension, left ventricular hypertrophy, atrial fibrillation, history of heart disease). We estimated adjusted cumulative incidence curves and derived risk ratios (RR) and risk differences (RD) at 5 and 10 years for stroke and mortality separately; 95% confidence intervals (CIs) were obtained via bootstrapping. We observed 350 cases of stroke and 1,881 deaths over 13 years of follow-up. Adjusted models unexpectedly suggested an inverse association between PM2.5 and incident stroke at 10 years of follow-up (RR=0.65, 95% CI: 0.49, 0.84). Participants residing in areas with PM2.5 concentrations in the highest quartile had a 1.11 (95% CI: 1.02, 1.22) times the risk of all-cause mortality over 10 years compared with those in the lower three quartiles (Figure). Ongoing analyses will address competing risks between stroke and death, investigate potential uncontrolled confounding, and conduct mediation analysis to formally assess the role of physical activity.