TyG-WC outperforms 15 metabolic markers for cardiometabolic multimorbidity risk

Researchers analysing the CHARLS cohort have compared 16 metabolic markers for their ability to predict incident cardiometabolic multimorbidity (CMM), the co-occurrence of multiple cardiometabolic conditions. In the cohort of 7,646 participants, 633 developed CMM during follow-up. Every marker examined showed a significant association with CMM in quartile analyses, but the Chinese visceral adiposity index (CVAI) stood out with the strongest association: participants in the highest quartile had a hazard ratio of 5.15 (95% CI: 3.52–7.54) compared with those in the lowest quartile.

The study, published open access in BMC Cardiovascular Disorders on 28 October 2026, used Cox regression and competing-risk models as its primary analyses, complemented by ROC, IDI and NRI assessments of incremental discrimination. XGBoost, quantile g-computation and structural equation modelling (SEM) served as secondary, exploratory analyses of feature importance, mixture effects and mediation pathways. Notably, the article was shared early as a peer-reviewed, accepted version of record in progress, citable with a permanent DOI but subject to further edits.

While CVAI showed the strongest association, a different marker delivered the greatest practical gain in risk discrimination: TyG-WC, a composite combining the triglyceride-glucose index with waist circumference. Adding it improved discrimination by an integrated discrimination improvement (IDI) of 0.026 and a net reclassification improvement (NRI) of 0.255, with P < 0.001. The authors emphasised that these incremental gains were modest and require external validation before any clinical implementation.

Composite insulin resistance (IR) and atherogenic lipid indices together accounted for 87.8% of the joint risk weight in the mixture analysis. The exploratory SEM then revealed an unexpected suppression pattern: biochemical IR statistically accounted for the association between structural (anthropometric) adiposity and CMM. The indirect effect was 0.34 (P < 0.001), and the mediation proportion reached 121.9% — a value the authors note reflects inconsistent mediation rather than complete mediation. Meanwhile, the direct effect of structural adiposity was negative and non-significant (P > 0.05).

In practical terms, the exploratory SEM results suggest that structural obesity without metabolic dysfunction was not independently associated with CMM in this cohort, because biochemical insulin resistance statistically carried the association. The authors caution that whether this points to a genuinely metabolically healthy phenotype remains an open question warranting further study. Their conclusion is measured: integrating biochemical and anthropometric markers, particularly TyG-WC, improves CMM risk stratification, but the modest incremental gains and the exploratory nature of the mediation findings mean external validation is needed before these composites could inform clinical practice.

Subject of Research: Comparative performance of 16 metabolic markers for predicting cardiometabolic multimorbidity in the CHARLS cohort

Article Title: The comparative performance of insulin resistance and visceral adiposity composites over traditional indices for cardiometabolic multimorbidity risk stratification: a multi-dimensional analysis of 16 markers and 5 pathophysiological pathways in the CHARLS cohort

Article References: Zhang, L., Zhang, Y., Deng, F., jiang, W., & Chu, C. (2026). The comparative performance of insulin resistance and visceral adiposity composites over traditional indices for cardiometabolic multimorbidity risk stratification: a multi-dimensional analysis of 16 markers and 5 pathophysiological pathways in the CHARLS cohort. BMC Cardiovascular Disorders. https://doi.org/10.1186/s12872-026-06827-0

Image Credits: AI Generated

DOI: 10.1186/s12872-026-06827-0

Keywords: cardiometabolic multimorbidity, insulin resistance, TyG-WC, visceral adiposity, CVAI, CHARLS cohort, risk stratification, structural equation modelling, NRI, anthropometric markers, comparative, performance