Objective: How well cardiovascular risk models perform in selected atherosclerosis patients for predicting outcomes is unknown. We sought to compare the performance of cardiovascular risk models (Framingham, Globorisk, SCORE2 & SCORE2-OP, and an updated new model) in predicting the 4-year outcome of patients with obstructive coronary artery disease (CAD).
Methods: Patients with suspected CAD who underwent coronary computed tomography angiography (CCTA) were recruited. Obstructive CAD was defined from CCTA as ≥ 50% stenosis. Computed tomography images, the scores of the cardiovascular risk models, and 4-year composite endpoints were assessed. Whether the patients underwent revascularization within 60 days after CCTA was also recorded. Multivariate regression analysis and receiver operating characteristics (ROC) curve analysis were performed.
Results: A total of 95 patients (mean age: 69.5 ± 10.33 years; 69 males) with obstructive CAD were included in this study. After the ROC analysis, the Framingham, Globorisk, SCORE2 & SCORE2-OP risk score showed prediction values with AUC 0.628 (95% CI: 0.532– 0.725), 0.647 (95% CI: 0.542– 0.742), 0.684 (95% CI: 0.581– 0.776), respectively. Multivariate regression analysis showed that, among the three risk models, only SCORE2 & SCORE2-OP risk score was associated with composite endpoints (hazard ratio: 1.050; 95% CI: 1.021– 1.079; p = 0.001) after adjusting for confounding factors. The AUC of the new risk model by combing SCORE2 & SCORE2-OP risk score with revascularization and the number of obstructive vessels in predicting composite endpoints reached 0.898 (95% CI: 0.819– 0.951).
Conclusion: The SCORE2 & SCORE2-OP risk score combined with the number of obstructive vessels and revascularization is predictive for adverse outcomes in patients with obstructive CAD.
Keywords: coronary computed tomography, obstructive coronary artery disease, cardiovascular risk factors, prediction model