The predictive ability of “TyG_CVAI” for incident stroke in individuals with different glycemic metabolic status: A national cohort study

The triglyceride-glucose-Chinese visceral adiposity index (TyG_CVAI) has been validated as an effective predictor of early stroke. However, its predictive efficacy across different glycemic metabolic states remains unexplored. This study utilized data from the China Health and Retirement Longitudinal Study, including 7744 stroke-free participants. Participants were categorized into 3 glycemic groups: diabetes mellitus (DM), prediabetes mellitus (Pre-DM), and normal glucose regulation (NGR). K-means clustering further divided participants into 2 clusters based on TyG_CVAI fluctuations: Cluster 1 (lower TyG_CVAI) and Cluster 2 (higher TyG_CVAI). The primary outcome was the first stroke occurrence. Statistical analyses included Kaplan-Meier survival curves, Cox proportional hazards models, restricted cubic spline analysis, and receiver operating characteristic curve analysis to assess associations between baseline TyG_CVAI and stroke risk. Over 9 years (2011-2020), 540 participants (7.0%) experienced a first stroke. Participants in Cluster 2 (higher TyG_CVAI) exhibited a significantly elevated stroke risk compared to Cluster 1. Kaplan-Meier analysis revealed significant differences in stroke incidence between clusters among Pre-DM, NGR, and the overall population (P < .001), but not in the DM group (P = .071). Cox regression models, adjusted for confounders, confirmed a significant positive correlation between TyG_CVAI and stroke risk in the overall population, Pre-DM, and NGR groups. The hazard ratios and 95% confidence intervals for Cluster 2 versus Cluster 1 were 1.40 (1.15-1.71), 1.42 (1.07-1.88), and 1.68 (1.19-2.36), respectively. In contrast, no significant association was observed in the DM group (P > .1). Restricted cubic spline analysis further supported that higher TyG_CVAI levels correlated with increased stroke risk. Elevated baseline TyG_CVAI is significantly associated with stroke risk in middle-aged and older adults. However, its predictive capacity is attenuated in individuals with diabetes compared to those with normal glycemic metabolism. These findings underscore the importance of considering glycemic status when evaluating TyG_CVAI as a stroke predictor.