Abstract
BACKGROUND: Cardiovascular disease (CVD) remains the leading cause of mortality worldwide, emphasizing the need for enhanced risk stratification tools. The triglyceride‒glucose-weight adjusted waist index (TyG-WWI), which integrates insulin resistance and central obesity, has emerged as a potential predictor, but its performance relative to traditional TyG-related indices (TyG, TyG-WC, TyG-WHtR) requires further validation. METHODS: We analyzed data from 24,255 participants in the National Health and Nutrition Examination Survey (NHANES, 1999-2018). Weighted Cox proportional hazards models were used to assess the associations between TyG-related indices and cardiovascular mortality. Restricted cubic splines (RCSs) with four knots were employed to explore dose‒response relationships. Traditional and time-dependent receiver operating characteristic (ROC) analyses, net reclassification improvement (NRI) analyses, and subgroup and sensitivity analyses were conducted to evaluate predictive performance and robustness. RESULTS: Over a median follow-up of 9.67 years, 854 cardiovascular deaths were recorded. According to the fully adjusted models, each increase in the standard deviation of the TyG-WWI was associated with a 45% greater risk of cardiovascular mortality (HR = 1.45, 95% CI 1.31-1.60), which was stronger than the associations observed for TyG (HR = 1.24, 95% CI 1.12-1.38), TyG-WC (HR = 1.39, 95% CI 1.27-1.53), and TyG-WHtR (HR = 1.43, 95% CI 1.30-1.58). When stratified by quartiles, the TyG-WWI exhibited a clear dose‒response relationship. RCS analyses revealed that the TyG-WWI had a linear association with cardiovascular mortality (P-nonlinear = 0.491), whereas the TyG index exhibited a U-shaped association, and the TyG-WC index and TyG-WHtR showed L-shaped associations (all P-nonlinear < 0.05). Traditional ROC analysis revealed that the TyG-WWI had the highest AUC (0.694, 95% CI 0.678-0.710). Time-dependent ROC analyses demonstrated that the AUC for the TyG-WWI ranged from 0.706 to 0.751 across different follow-up time points, which was consistently greater than those of the other TyG-related indices. NRI analyses indicated significant improvements in risk reclassification when the TyG-WWI was used compared with traditional TyG-related indices (10.4% vs. TyG, 9.4% vs. TyG-WC, 9.1% vs. TyG-WHtR). Subgroup analyses revealed stronger associations in younger adults (≤ 60 years, HR = 2.03, 95% CI 1.78-2.32). CONCLUSION: The current study is the first to validate that the TyG-WWI is a reliable risk prediction tool for cardiovascular death in the general population and has greater predictive value than traditional TyG-related parameters. The results support its potential as a supplementary tool among TyG-derived markers for assessing cardiovascular death.