Simplicity meets accuracy: TyG-Waist as a practical alternative to CVAI for type 2 diabetes prediction in high-risk Chinese adults

Early identification of high-risk individuals is crucial for type 2 diabetes mellitus (T2DM) prevention, particularly in populations with existing cardiovascular risk factors. While the Chinese Visceral Adiposity Index (CVAI) is a robust predictor, its calculation is complex and requires lipid sub-fractions. We aimed to evaluate whether the parsimonious TyG-Waist index—combining insulin resistance and central obesity—could serve as a practical alternative to CVAI for predicting incident T2DM in high-risk Chinese adults. This prospective cohort study included 21,074 adults (median age 60 years) from the high-risk sub-cohort of the China PEACE Million Persons Project (Changzhou), who were free of T2DM at baseline. Participants were followed for a median of 7.48 years (IQR: 6.35–7.61). We compared the predictive performance of TyG-Waist against CVAI, Visceral Adiposity Index (VAI), and Lipid Accumulation Product (LAP) using multivariable Cox regression, Harrell’s C-index, time-dependent ROC, and Net Reclassification Improvement (NRI) analyses, stratified by sex. During follow-up, 2,842 incident T2DM cases were identified. We observed a distinct sexual dimorphism in predictive performance. In women, CVAI (C-index: 0.628) and TyG-Waist (C-index: 0.626) demonstrated the highest discrimination, with a negligible difference (\:Δ\:C = 0.002), indicating clinical equivalence. In men, the simple TyG-Waist (C-index: 0.603) significantly outperformed the complex CVAI (C-index: 0.594) and other indices. Dose-response analyses revealed a J-shaped association in women and a linear association in men. Furthermore, adding TyG-Waist to baseline models significantly improved risk reclassification (NRI: 0.155 in women, 0.143 in men) and provided substantial net clinical benefit in decision curve analysis. The simple TyG-Waist index is a robust predictor of T2DM in high-risk Chinese adults. It demonstrates clinical equivalence to the complex CVAI in women and superior performance in men. Given its dependence on only basic parameters (glucose, triglycerides, and waist circumference) and independence from HDL-C, TyG-Waist represents an economically efficient and pragmatic alternative to CVAI for large-scale risk stratification in resource-limited settings. Not applicable.