Diabetology & Metabolic Syndrome · Aug 19, 2026

Insulin resistance-related indices and incident cardiovascular disease in individuals with PREVENT-estimated 10-year ASCVD risk < 5

By Jiang H, Zhao Y, Zhao Y et al.

BACKGROUND: The PREVENT equations have been recommended by recent AHA guidelines for estimating 10-year ASCVD risk to inform lipid-lowering strategies in primary prevention. Among individuals with low-to-borderline 10-year ASCVD risk according to the PREVENT equations, treatment decisions should be further refined by considering risk-enhancing factors. Although insulin resistance (IR) is closely associated with ASCVD, it remains unclear whether IR-related indices capture residual risk in this low-to-borderline risk population and whether commonly used indices differ in their associations with incident ASCVD or incremental predictive performance. METHODS: This prospective cohort study included 2,664 participants from CHARLS with low-to-borderline risk. Six IR-related indices were assessed, including the TyG, its derived measures, and estimated glucose disposal rate (eGDR). The primary outcome was incident ASCVD. Associations with incident ASCVD were examined using Cox proportional hazards models. Kaplan-Meier analyses and restricted cubic spline (RCS) models were used to evaluate cumulative incidence patterns and dose-response relationships. Standalone discrimination was assessed using receiver operating characteristic curves (ROC), and incremental performance beyond the base PREVENT model was evaluated using changes in AUC, calibration, net reclassification improvement (NRI), and integrated discrimination improvement (IDI). Calibration was assessed using bootstrap-corrected calibration curves with 500 resamples. RESULTS: Over the observation period, 324 participants developed ASCVD. In Model 3, TyG-BMI, TyG-WHtR, and TyG-CVAI were positively related to ASCVD occurrence, whereas eGDR was linked to lower risk. In ROC analyses, TyG-CVAI, eGDR, TyG-BMI, and TyG-WHtR yielded higher AUCs than TyG and CTI. TyG-CVAI had the highest AUC (0.578, 95% CI: 0.546-0.611), but did not differ significantly from TyG-BMI, TyG-WHtR, or eGDR in DeLong tests. When added separately to the PREVENT-based reference model, several indices significantly improved at least one performance measure, whereas only eGDR significantly improved the AUC, continuous NRI, and IDI simultaneously. Nevertheless, the absolute improvements were modest: the AUC increased from 0.554 to 0.579, the continuous NRI was 0.135, and the IDI was 0.004. The addition of the IR-related indices did not materially improve calibration. CONCLUSIONS: In the low-to-borderline risk population, TyG-BMI, TyG-WHtR, and TyG-CVAI were positively associated with incident ASCVD, whereas eGDR was inversely associated with ASCVD risk. Among the six commonly used IR-related indices evaluated, eGDR appeared to provide the most consistent complementary risk information within the PREVENT framework, although the observed improvements were modest and require external validation. The relatively low AUCs of all IR-related indices indicated limited standalone discrimination.

doi.org/10.1186/s13098-026-02265-2Open accessPubMed 43035757Biomarkers

AgingResearch.News