Abstract
AIMS: This multi-cohort study evaluated whether the cardiometabolic index (CMI)-a composite of waist-to-height ratio and triglyceride-to-HDL cholesterol ratio-serves as an early predictor of cardio-kidney risk and whether its predictive value varies across glycaemic states. METHODS: We analysed 327,902 adults in the UK Biobank to examine the association of CMI with baseline cardio-kidney comorbidities, incident cardio-kidney events (CKE)-defined as the composite occurrence of cardiovascular and chronic kidney outcomes, and mortality. Baseline comorbidities was assessed using logistic regression, and Cox models with stratified analyses and restricted cubic splines (RCS) evaluated prospective associations. Findings were externally validated in CHARLS and NHANES. Machine-learning survival models further assessed predictive performance. RESULTS: Higher CMI was associated with baseline cardio-kidney comorbidities (OR 2.25, 95% CI 2.10-2.42). Among 303,113 participants free of cardiovascular and/or kidney disease at baseline, CMI predicted incident CKE (HR 2.18, 95% CI 2.01-2.38; median follow-up 14.3 years), all-cause death (HR 1.10, 95% CI 1.06-1.14; 15.8 years), and cardio-kidney death (HR 1.55, 95% CI 1.37-1.76; 15.8 years). The strength of associations was greatest in normoglycemia and progressively attenuated in prediabetes and diabetes. RCS analyses revealed nonlinear dose-response relationships, with steep increases in CKE and cardio-kidney mortality below CMI thresholds (∼0.70 and ∼0.95) and more gradual rises thereafter. Results were directionally consistent in external cohorts, particularly for cardio-kidney comorbidities and incident CKE. ML models demonstrated strong discrimination and consistently ranked CMI among the top predictors of incident CKE. CONCLUSIONS: CMI is a simple, robust predictor of cardio-kidney risk especially in earlier metabolic states, with particularly strong prognostic value in normoglycaemic individuals where excess risk appears at lower CMI levels.