Association between the triglyceride-high-density lipoprotein-glucose-body index and new-onset stroke risk: a prospective cohort study

BACKGROUND: Stroke is a leading cause of mortality and disability in China. Insulin resistance (IR) is a crucial modifiable risk factor. While the triglyceride-glucose (TyG) index is a recognized surrogate for IR, it overlooks the protective role of high-density lipoprotein cholesterol (HDL-C) and the impact of obesity (BMI). The novel TyHGB index integrates triglycerides, fasting blood glucose, HDL-C, and BMI, offering a comprehensive metabolic profile. However, the association between its longitudinal patterns and new-onset stroke risk remains unclear. METHODS: Data from 4,823 participants free of stroke at baseline were analyzed using the China Health and Retirement Longitudinal Study (CHARLS, 2011–2020). Longitudinal patterns of the TyHGB index were identified using K-means clustering. Cox proportional hazards models evaluated the association between TyHGB metrics (patterns, baseline, and cumulative) and stroke risk. Restricted cubic splines (RCS) explored non-linear relationships, and time-dependent ROC curves compared the predictive value of TyHGB versus the TyG index. RESULTS: During the 5-year follow-up, 417 (8.6%) incident stroke cases were recorded. Three distinct TyHGB patterns were identified: “Low,” “Medium,” and “High.” After full adjustment for confounders, participants in the High (HR = 1.33, 95% CI: 1.01–1.78) and Medium (HR = 1.34, 95% CI: 1.07–1.67) patterns had a significantly higher stroke risk compared to the Low group. A non-linear dose-response relationship was observed between cumulative TyHGB and stroke risk (P-non-linear < 0.05), with a risk saturation threshold at 25.76. Furthermore, the TyHGB index demonstrated a modest but statistically significant improvement in predictive ability compared to the traditional TyG index for 5-year stroke risk (AUC: 0.591 vs. 0.574). CONCLUSIONS: Long-term elevated patterns and cumulative burden of the TyHGB index are independent risk factors for new-onset stroke. The identification of an RCS-derived inflection point (25.76) serves as a useful population-level reference value.TyHGB provides incremental value over the single TyG index in risk stratification, highlighting the importance of dynamic metabolic monitoring for stroke prevention. TRIAL REGISTRATION: Not applicable.