Maturitas · Mar 19, 2026

Nonlinear association between triglyceride-glucose frailty index and new-onset stroke in adults at cardiovascular-kidney-metabolic syndrome stages 0-3: a CHARLS cohort study

By Wei J, Liu Y

BACKGROUND: The triglyceride-glucose frailty index integrates metabolic load and physiological vulnerability. Its association with incident stroke among individuals at cardiovascular-kidney-metabolic syndrome stages 0-3 remains unclear. METHODS: We performed an observational cohort analysis in the China Health and Retirement Longitudinal Study. Adults at cardiovascular-kidney-metabolic syndrome stages 0-3 who were free of stroke at baseline were included. The exposure of interest was the triglyceride-glucose frailty index, and the outcome was incident stroke during follow-up. Cox proportional hazards models estimated adjusted associations. Nonlinearity was assessed with restricted cubic splines. Effect modification by baseline cognitive function was tested using interaction terms. Regression-based mediation evaluated self-rated health and hypertension. RESULTS: A total of 6882 adults without stroke at baseline were included, among whom 583 incident strokes occurred during follow-up, an incidence of 8.47%. Stroke incidence increased across tertiles of the triglyceride-glucose frailty index, from 5.14% to 8.24% and 12.03%, with clear separation of Kaplan-Meier curves. In Cox proportional hazards models, each 5-unit increase in the triglyceride-glucose frailty index was associated with a higher hazard of stroke, with a hazard ratio of 1.46 and a 95% confidence interval of 1.31 to 1.64. Compared with the lowest tertile, the hazard ratios were 1.54 (95% confidence interval 1.22 to 1.95) and 2.16 (95% confidence interval 1.71 to 2.73) for the middle and highest tertiles, respectively. Restricted cubic spline analyses indicated a nonlinear association, with an inflection point at a triglyceride-glucose frailty index value of approximately 8.31. Below this level, stroke risk increased more steeply, whereas the association was attenuated above this threshold. Baseline cognitive function significantly modified the association, with stronger effects observed among individuals with better cognitive function. Exploratory mediation analyses suggested that self-rated health and hypertension accounted for approximately 23.0% and 7.6% of the overall statistical association, respectively. Subgroup and sensitivity analyses yielded results that were broadly consistent with the primary findings. CONCLUSIONS: Among adults at cardiovascular-kidney-metabolic syndrome stages 0-3, a higher triglyceride-glucose frailty index was associated with an increased risk of incident stroke. This association was nonlinear, varied by baseline cognitive function, and was partially mediated by self-rated health and hypertension.

doi.org/10.1016/j.maturitas.2026.108922PubMed 41864135Biomarkers