The association between triglyceride-glucose index combined with obesity indicators and stroke risk: A longitudinal study based on CHARLS data

Background and objective The triglyceride-glucose (TyG) index, a surrogate marker for insulin resistance, has been proposed as a predictor of cardiovascular events. However, the combined impact of the TyG index and obesity indicators, such as body mass index (BMI), waist circumference (WC), and waist-to-height ratio (WHtR), on stroke risk was not fully understood. This study aimed to investigate the association between the TyG index combined with these obesity indicators and stroke risk in the Chinese population. Materials and methods Data on 17,708 participants aged 45 years or older for this study were collected from the China Health and Retirement Longitudinal Study (CHARLS) from baseline (2011) to the Wave 5 follow-up (2020). Cox proportional hazards model, restricted cubic spline (RCS) and receiver operating characteristic (ROC) analysis were employed to examine the association between the TyG index and its combined obesity-related indicators with stroke. Mediation analysis was conducted to explore the mutual potential mediating role of TyG and obesity indicators in the above relationships. Results A total of 8,207 participants with an average age of 58.2 years were investigated, of which 11.0% were stroke individuals, 44.80% were men and 84.6% were from rural areas. TyG, TyG-BMI, TyG-WHtR, TyG-WC were significantly higher in stroke subjects than in the non-stroke subjects (P < 0.001), and were significantly and positively associated with stroke in all 3 models (P < 0.05). Restricted cubic spline models revealed nonlinear associations between TyG and TyG-BMI with stroke (P-overall < 0.001, P-nonlinear = 0.003 for TyG, and P-overall < 0.001, P-nonlinear = 0.028 for TyG-BMI), while TyG-WC and TyG-WHtR (P-overall < 0.001 and P-nonlinear > 0.05) demonstrated linear associations with stroke after adjusting for covariates. TyG-WHtR, TyG-BMI and TyG-WC had more robust predictive power than TyG for risk of stroke. TyG-WHtR or TyG-WC had the highest predictive power for stroke (AUC:0.696, 95% CI 0.677-0.715), slightly higher than the other indicators. Associations between TyG, TyG-WC, TyG-WHtR, and TyG-BMI with stroke were found to be stronger among individuals who were >= 55 years of age, male. The relationship between TyG and stroke was significantly mediated by BMI, WHtR and WC (15.79%, 21.72%, and 24.06% respectively), while the relationship between these obesity measures and stroke was significantly mediated by TyG (18.48%, 14.45%, and 14.70% respectively). Conclusion The combination of TyG and obesity-related indicators was significantly associated with stroke risk, and could improve predictive power for stroke compared to the single TyG. Obesity indicators and TyG mediated each other in their respective associations with stroke risk.