Triglyceride-Glucose Index Mediated by Metabolic Burden Predicts Cardiovascular Events in Patients With Complex Coronary Lesions: A Prospective Cohort Study

Abstract

This prospective cohort study evaluated the prognostic value of the triglyceride-glucose (TyG) index and its metabolic mediation pathways in 16 499 patients with complex coronary lesions (American College of Cardiology/American Heart Association type B2/C) undergoing percutaneous coronary intervention (PCI). Over a median 3-year follow-up, 535 patients (3.2%) experienced cardiovascular (CV) events. Patients in the highest TyG tertile (T3: >9.10) demonstrated a 2.442-fold increased CV risk compared with the lowest tertile (T1: <8.63; 95% CI 1.923-3.102, P < .001), with a significant dose-response relationship (P-trend < .001). Structural equation modeling revealed that metabolic burden (hypertension, dyslipidemia, type 2 diabetes mellitus) mediated 50.0% of TyG-associated risk (indirect effect P = .038). TyG index exhibited superior predictive accuracy for CV events compared with the TG/HDL-C ratio (area under the curve [AUC] 0.617 vs 0.577, P = .004), though their combination did not significantly improve discrimination over TyG alone (△AUC = 0.001, P = .694). Type 2 diabetes mellitus accounted for 62.3% of the mediation effect, surpassing hypertension (20.8%), and dyslipidemia (16.9%). These findings establish the TyG index as an independent predictor of CV outcomes in complex coronary lesions, with metabolic comorbidities explaining half its prognostic impact. Integration of TyG-driven metabolic phenotyping may optimize risk stratification and targeted interventions in high-risk PCI populations.