Combined associations of physical activity, BRI, and TyG index with incident stroke among middle-aged and old adults

To examine the independent and joint associations of physical activity (PA), body roundness index (BRI), and triglyceride-glucose (TyG) index with incident stroke, and to assess effect modification and longitudinal mediation by changes in TyG and BRI. This prospective study included 5,034 participants from CHARLS (2011–2018). PA was categorized, BRI and TyG were calculated from baseline data. Cox models evaluated associations, with restricted cubic splines (RCS) assessing non-linear relationships. Longitudinal mediation examined whether 4-year changes in TyG and BRI (2011–2015) mediated the PA-stroke link in a subsample ( n = 1,867). Each 1-unit increase in TyG was associated with higher stroke risk (HR = 1.17, 95% CI 1.00–1.38). High PA showed a non-significant association (HR = 0.78, 95% CI 0.59–1.02). No multiplicative interactions were found ( P ≥ 0.24). Non-linear dose-response relationship for TyG was significant only in the Low ‑ Moderate PA group ( P = 0.001), while BRI showed no evidence of non-linearity in either group. Hypertension modified the TyG×BRI product association ( P -interaction = 0.028). No mediation through 4-year metabolic changes was observed. TyG independently predicts stroke, while BRI and PA show no multiplicative interaction. However, short-term metabolic changes do not mediate the PA-stroke link. Exploratory mediation analyses failed to identify 4-year TyG/BRI changes as mediating pathways, though these results are limited by temporal measurement constraints and poor model fit and cannot rule out long-term metabolic mediation effects.