Associations of the triglyceride-glucose index and its composite indices with mortality in adults with T2DM

The triglyceride-glucose (TyG) index and its composite indices have emerged as potential markers for metabolic risk, but their prognostic value for mortality in middle-aged and older adults with T2DM remains incompletely defined.

We analyzed data from 1,940 adults with T2DM (aged ≥45 years) who underwent routine health examinations. Baseline characteristics were described using mean (SD), median (IQR), and count (%). Cox proportional hazards regression models were constructed to evaluate associations of TyG-related indices (TyG-WC, TyG-WHR, TyG-WHtR, TyG-ABSI, TyG-WWI) with all-cause mortality, with adjustment for confounders. Restricted cubic splines (RCS) were used to explore nonlinear relationships, and time-dependent ROC curves assessed predictive performance. Subgroup analyses were stratified by age, sex, smoking, drinking, and hypertension status.

During follow-up, 312 (16.1%) participants died. In fully adjusted models, TyG-WC, TyG-WHtR, TyG-ABSI, and TyG-WWI were independently associated with increased mortality (HR range: 1.21-1.44, all p < 0.05), with TyG-WHtR showing the strongest association (HR = 1.44, 95%CI: 1.21-1.70). RCS analyses revealed significant J-shaped nonlinear relationships for these composite indices (all p for nonlinear <0.01). Time-dependent AUCs indicated moderate predictive performance, with TyG-ABSI and TyG-WWI achieving the highest values. Subgroup analyses confirmed consistent associations in participants aged >65 years, non-drinkers, non-smokers, and those without hypertension.

TyG composite indices, particularly TyG-WHtR, TyG-ABSI, and TyG-WWI, are independent predictors of all-cause mortality in middle-aged and older adults with T2DM, offering potential utility for long-term mortality risk stratification.