Association between coronary heart disease and cardiometabolic index: A study from NHANES 1999-2018

Background : The cardiometabolic index (CMI) is a straightforward and gender-specific marker that combines anthropometric measurements with lipid profiles. We aim to explore the correlations between CMI and the risk of coronary heart disease (CHD). Methods: The data from the National Health and Nutrition Examination Survey (NHANES) covering survey cycles from 1999 to 2018. We employed a multistage analytical approach to elucidate the relationship between CMI and CHD risk. Variable selection was initially performed using LASSO regression complemented by univariate and multivariate logistic regression analyses. Weighted multivariable logistic regression models incorporating restricted cubic splines (RCS) were subsequently developed to quantify independent associations and characterize potential non-linear dose-response patterns. Threshold effects were systematically evaluated through segmented regression modeling, while stratified analyses with interaction testing were conducted across clinically defined subgroups. The predictive performance of established models was rigorously assessed using receiver operating characteristic (ROC) curve analysis, with nomograms constructed to visually represent individualized risk stratification. Results: The final analysis included 20,888 participants from the NHANES database. After comprehensive adjustment for confounding factors, CMI demonstrated a significant positive association with CHD risk. This correlation was prominent in the highest CMI quartile (Q4) group (OR = 1.58, 95% CI: 1.06–2.37, P = 0.027). RCS analysis revealed a distinct, nonlinear, positive relationship between CMI and CHD risk (P for nonlinearity < 0.001). Threshold effect analysis identified an inflection point at 1.175, with stratified analysis showing that each unit increase in CMI below this threshold corresponded to a 133.1% elevation in CHD likelihood (OR = 2.331, 95% CI: 1.830–2.971, P <0.001). The prediction model developed through LASSO regression and logistic regression selection demonstrated robust discriminative capacity, achieving an area under the curve (AUC) of 0.861 (95% CI: 0.851–0.872). Conclusions : This study provides population-based evidence establishing CMI as an independent predictor of CHD risk through robust positive associations. These findings advocate for integrating CMI into precision prevention paradigms, where targeted monitoring of elevated CMI levels could enable cost-effective risk stratification and timely clinical interventions to mitigate cardiovascular disease burden.