Mortality and cardiovascular risk of triglyceride-glucose and derived indices in cardiovascular-kidney-metabolic syndrome stages 0-3: a dose-response meta-analysis
Background The associations of the triglyceride-glucose (TyG) index and its derived indices with the progression of cardiovascular-kidney-metabolic (CKM) syndrome remain poorly elucidated. Methods We conducted a systematic review and dose-response meta-analysis of 80 observational studies encompassing populations within CKM stages 0–3, evaluating TyG, TyG-BMI, TyG-WC, TyG-WHtR, and CTI for all-cause mortality, cardiovascular disease (CVD) mortality, and CVD incidence (including stroke, coronary heart disease, heart failure and major adverse cardiovascular events). A one-stage mixed-effects approach was used, supplemented by subgroup, meta-regression, leave-one-out, and trim-and-fill analyses. Results In both continuous (per 1-SD increment) and categorical (highest vs. lowest) analyses, TyG-derived indices generally demonstrated stronger risk associations with target outcomes than TyG alone. Notably, CTI, TyG-WC, and TyG-WHtR exhibited higher risk associations. In the highest-category analysis, TyG-WC yielded the most pronounced risk for CVD mortality (RR = 1.52, 95% CI: 1.35–1.70, 95% PI: 1.26–1.82). CTI was associated with a 57% increase in incident CVD (RR = 1.57, 95% CI: 1.33–1.85, 95% PI: 1.00–2.46). Conversely, TyG-BMI showed non-significant associations with mortality-related outcomes (P > 0.05), though it predicted a 51% risk increase for incident CVD (RR = 1.51, 95% CI: 1.36–1.67, 95% PI: 1.05–2.16). Subgroup analysis indicated that these positive associations were pronounced in CKM stages 1–3 but entirely non-significant in stage 0 and the obesity subgroup had a significant impact on the risk of mortality. Dose-response analyses revealed a significant non-linear U-shaped association of the TyG index with all-cause mortality (Pnon-linearity < 0.0001; nadir = 8.90). For CVD mortality, the overall association was significant (Poverall = 0.005; nadir = 8.91), whereas a monotonic increasing trend was observed for CVD incidence (Poverall = 0.0014). In sensitivity analyses excluding participants with cancer or consumptive diseases at baseline, as well as those in CKM stage 0, the U-shaped patterns remained stable (mortality nadir shifted to 9.1). Conclusions TyG-derived indices, particularly TyG-WC, TyG-WHtR, and CTI, show stronger risk associations than the TyG index for cardiovascular risk assessment in CKM populations, whereas the clinical utility of TyG-BMI is relatively limited. Moderate elevations in TyG were associated with lower mortality risk within a certain range, though the non-linear association was confirmed only for all-cause mortality. Systematic review registration https://www.crd.york.ac.uk/PROSPERO/view/CRD420251111362, identifier CRD420251111362.