Objective To investigate the factors associated with the occurrence of premature coronary artery disease (PCAD), and to construct a column graph to predict the risk of PCAD.

Methods Clinical data of 202 patients suspected of PCAD were retrospectively collected (training set). A total of 26 demographic and blood biochemical indicators were collected, and the subjects were divided into PCAD group (n=139) and non-coronary heart disease group (NCAD group, n=63) according to the occurrence of PCAD. Univariate analysis and Logistic forward stepwise regression were used to analyze the prediction indicators, screen the independent correlation factors of PCAD and constructed the nomogram. The model differentiation, calibration and clinical effectiveness were evaluated by calculating C index, drawing calibration curve and decision analysis curve. The 1 000 times Bootstrapping method was used for internal verification.

Results Impaired fasting glucose regulation/diabetes mellitus, hypertension, leukocyte-high-density lipoprotein cholesterol ratio (LHR) ≥ 7.58, overweight/obesity and smoking were significantly correlated with the occurrence of PCAD (P < 0.05). The column chart constructed based on the above factors had good differentiation ability (C index=0.815, 95%CI, 0.757 to 0.873), calibration degree and clinical effectiveness. The model also showed good differentiation during internal validation (C index=0.798).

Conclusion The nomogram constructed based on LHR, smoking, hypertension, body massindex (BMI), fasting blood glucose and other factors can be used to predict the risk of PCAD in young patients suspected of coronary heart disease, which has potential clinical application value.