A retrospective cohort study of the association between triglyceride-glucose index and its derived indicators and all-cause mortality among community-dwelling adults with a history of stroke

Abstract Objective As a novel marker of insulin resistance, the triglyceride-glucose (TyG) index has been documented to correlate with mortality in coronary artery disease and heart failure. In recent years, growing attention has been paid to the relationship between the TyG index and all-cause mortality in stroke survivors. Moreover, composite TyG-derived indicators integrating multiple clinical factors have demonstrated superior predictive performance compared with the original TyG index alone. This study aimed to investigate the prognostic value of the TyG index and its derivative indices for all-cause mortality among adults with a history of stroke. Methods Data from 1,805 participants were extracted from the National Health and Nutrition Examination Survey (NHANES, 1999–2018). Multivariate Cox proportional hazards regression models incorporating sampling weights were constructed for statistical analysis. Restricted cubic spline (RCS) functions were applied to detect potential nonlinear relationships between exposure indicators and the primary endpoint. Kaplan-Meier survival curves were plotted to visualize the association between these indices and clinical outcomes in stroke patients. Interaction terms and stratified subgroup analyses were further performed to verify the robustness and heterogeneity of the research findings. Results A total of 1,805 participants with a history of stroke were enrolled in the final cohort, among whom 937 deaths were recorded during follow-up.In the fully adjusted regression model, elevated TyG-ABSI (HR = 1.218, 95%CI: 1.022–1.450, P = 0.027) and TyG-WWI (HR = 1.237, 95%CI: 1.023–1.497, P = 0.029) were significantly associated with increased all-cause mortality when compared with the lowest quartile, whereas no such significant correlation was identified for other indicators. Kaplan-Meier curves indicated that higher TyG, TyG-WWI and TyG-ABSI levels were positively associated with all-cause mortality, while TyG-BMI showed an inverse correlation with mortality risk.Restricted cubic spline analyses revealed a U-shaped association between all tested indices and mortality (all P for nonlinearity < 0.05), with the only exception of TyG-WHtR. Significant effect modification by diabetes status was only observed for TyG-WHtR, TyG-WWI and TyG-ABSI (all P < 0.05). Conclusion This study is the first to confirm that TyG-WWI and TyG-ABSI serve as robust predictors for all-cause mortality in stroke patients, with better predictive efficacy than the conventional TyG index. The results suggest that these two TyG-derived biomarkers have promising application value as auxiliary prognostic tools for mortality risk evaluation among individuals with stroke.