Associations between a body shape index and coronary heart disease: a case-control study in southern China

Coronary heart disease (CHD) burden is increasing, and traditional obesity measures inadequately capture fat distribution and associated CHD risk. A body shape index (ABSI) is an emerging anthropometric metric of fat distribution, but evidence linking ABSI to CHD is limited, particularly in the Chinese population. This case-control study in southern China investigated the association of ABSI and related factors with CHD risk, aiming to facilitate early identification of high-risk individuals.

We retrospectively studied 996 patients who underwent coronary angiography in a southern Chinese hospital. After strict screening and propensity score matching (PSM), 125 patients with CHD (>50% coronary stenosis) and 125 controls (<50% stenosis) were selected. Key CHD risk predictors were identified using feature-selection techniques (LASSO regression, recursive feature elimination, random forest importance). Univariate and multivariate logistic regression models were constructed for CHD prediction. Model performance was evaluated by receiver operating characteristic (ROC) analysis and compared to individual predictors using the DeLong test. A nomogram was developed for individualized risk estimation.

Baseline characteristics were well matched between CHD and control groups after PSM. Across feature-selection methods, the most influential predictors for CHD included ABSI, prealbumin (PA), direct-to-total bilirubin ratio (DB/TB), apolipoprotein B (ApoB), globulin (GLO), apolipoprotein A-I (ApoA-I), and essential hypertension (EH). Each of these factors showed a significant univariate association with CHD (P < 0.05) but only modest predictive power individually (AUCs 0.57-0.66). ABSI exhibited the highest sensitivity (86.4%) among single predictors, while ApoB had the highest specificity (78.4%). The multivariable logistic model incorporating these variables achieved an AUC of 0.809, significantly outperforming any individual predictor (P < 0.001). At the optimal probability cutoff, the model's sensitivity and specificity were 69.6% and 82.4%, respectively. The nomogram combined ABSI with other key variables to provide a quantitative CHD risk estimate for individual patients.

This study identifies ABSI as a potential predictor of CHD risk among southern Chinese populations. Integrating ABSI with other candidate predictors improves the model's predictive performance. A multifactorial approach may better characterize CHD risk in this population and could inform prevention strategies.