Differential associations between novel triglyceride-glucose-obesity indices and incident stroke in individuals with early-stage cardiovascular-kidney-metabolic syndrome: a longitudinal cohort study

The triglyceride-glucose (TyG) index combined with advanced anthropometric measures has emerged as a promising tool for cardiovascular risk assessment. However, evidence comparing the predictive utility of various novel TyG-obesity indices for incident stroke in individuals with early-stage cardiovascular-kidney-metabolic (CKM) syndrome remains limited. This study aimed to investigate and compare the associations of six novel TyG-obesity indices (TyG-Chinese visceral adiposity index [CVAI], -body roundness index [BRI], -conicity index [CI], -weight-adjusted waist index [WWI], -a body shape index [ABSI], and -relative fat mass [RFM]) with stroke in individuals with CKM stages 0–3.

This prospective cohort study included 3400 participants from the China Health and Retirement Longitudinal Study. K-means clustering was used to identify longitudinal change patterns, and the Boruta algorithm ranked feature importance. We used Cox proportional hazards models to evaluate the strength of associations, restricted cubic splines (RCS) to explore dose-response relationships, receiver operating characteristic (ROC) curves to assess discriminative ability, and net reclassification improvement (NRI) and integrated discrimination improvement (IDI) to quantify incremental predictive performance.

During a median follow-up of 8.6 years, 232 incident strokes were documented. Using a multidimensional analytical framework incorporating baseline, cumulative, and longitudinal change-based approaches, six indices were differentially associated with stroke risk, with TyG-CVAI and TyG-BRI showing stronger and more consistent associations. Each standard deviation increase in baseline TyG-CVAI and TyG-BRI corresponded to hazard ratios (HRs) of 1.36 (95% CI 1.17–1.57) and 1.32 (1.14–1.53). The corresponding cumulative HRs were 1.31 (1.14–1.52) and 1.29 (1.11–1.49). Compared with the stable low-level group, sustained high-level group yielded HRs of 2.28 (1.54–3.39) for TyG-CVAI and 2.01 (1.35–2.99) for TyG-BRI. Stage-specific analyses revealed stronger associations in CKM stage 2. Both indices demonstrated positive linear dose-response relationships. ROC analyses showed that TyG-CVAI and TyG-BRI had relatively better discrimination, and NRI/IDI analyses demonstrated incremental predictive value over the baseline model.

Among the six novel TyG-obesity indices evaluated, TyG-CVAI and TyG-BRI emerged as stronger predictors of incident stroke in individuals with early-stage CKM syndrome, particularly in stage 2. These findings highlight the potential utility of these two indices for stroke risk stratification and may inform early detection strategies in this vulnerable population.