To investigate the association between trigyceride glucose index - body roundness index (TyG-BRI) and the risk of stroke in middle-aged and older adults using the CHARLS database.

This study conducted a longitudinal study using CHARLS data from 2011 to 2020. The TyG index was calculated from baseline fasting blood glucose and triglyceride levels. TyG-BRI was developed by integrating the TyG index with baseline waist circumference and height. Stroke was identified through follow-up questionnaires. After imputing missing data using MICE, we analyzed the data by stratifying TyG-BRI according to quartile levels and conducted Kaplan-Meier survival curve analysis. Meanwhile, a multivariable Cox proportional hazards regression model was used to assess the association between different TyG-BRI levels and the risk of stroke. Restricted cubic spline (RCS) analysis was employed to examine the dose-response relationship between TyG-BRI and the incidence of stroke. Finally, subgroup analyses and interaction tests were performed to explore the heterogeneity of the association between TyG-BRI and stroke among different populations.

A total of 7 709 individuals were included in the longitudinal study throughout the entire follow-up period, and 536 developed stroke. Both Cox regression and Kaplan-Meier analyses showed that higher TyG-BRI levels were associated with increased stroke risk. Using the first quartile (Q1) as reference, the fully adjusted model (Model 3) showed significantly elevated risks for Q2 (HR=1.39, 95% CI: 1.04-1.86), Q3 (HR=1.70, 95% CI: 1.28-2.27), and Q4 (HR=1.75, 95% CI: 1.28-2.40). RCS analysis revealed a linear relationship between TyG-BRI and stroke risk with an overall positive trend. Subgroup analyses did not identify any factors independently associated with stroke. Sensitivity analyses confirmed the robustness of these findings.

TyG-BRI shows potential as a useful indicator for early detection and prevention in stroke risk management.