Association of TyG-related indices with cardiometabolic multimorbidity across early cardiovascular-kidney-metabolic syndrome stages 0–3: a prospective cohort study

Abstract

Abstract Cardiovascular–kidney–metabolic (CKM) syndrome is a systemic condition characterized by interrelated cardiac, renal, and metabolic disturbances. The triglyceride–glucose (TyG) index and several of its anthropometry-based derivatives have been increasingly recognized as surrogate markers of insulin resistance and potential predictors of cardiometabolic multimorbidity (CMM). However, whether their predictive performance differs across CKM stages 0–3 remains unclear. This prospective cohort investigation enrolled 646 participants without established CMM at baseline. CKM stages were classified according to AHA criteria based on obesity, metabolic abnormalities, kidney dysfunction, and subclinical cardiovascular disease. Five surrogate indicators of insulin resistance were examined, including the TyG index and four anthropometry-based TyG-derived indices. To evaluate model discrimination over time, Harrell’s time-dependent C-statistics were calculated. Improvements in predictive performance were further assessed by estimating the net reclassification index (NRI) and the integrated discrimination index (IDI). Restricted cubic spline functions were additionally fitted to depict potential dose–response relationships between each TyG-related measure and incident CMM. During a four-year follow-up, 21.36% of participants developed CMM. All TyG-related indices were significantly associated with increased CMM risk, displaying clear linear dose–response patterns. Among participants classified as CKM stage 3, TyG-WWI (TyG–weight-adjusted waist index) showed a robust association with CMM. Among all indicators, TyG-WWI and TyG-WHtR (TyG–waist-to-height ratio) showed relatively better discrimination, with meaningful improvements in NRI and IDI compared with the TyG index alone. Across CKM stages 0–3, the TyG index and its derivative markers showed consistent and independent associations with incident CMM. TyG-WWI and TyG-WHtR exhibited relatively better discrimination, highlighting their potential utility for early identification of high-risk individuals in community screening contexts, though their direct clinical applicability requires further validation.