Predicting long‐term ischemic events using routine clinical parameters in patients with coronary artery disease: The OPT‐CAD risk score

BACKGROUND The prognosis of patients with coronary artery disease (CAD) at hospital discharge was constantly varying, and postdischarge risk of ischemic events remain a concern. However, risk prediction tools to identify risk of ischemia for these patients has not yet been reported. AIMS We sought to develop a scoring system for predicting long-term ischemic events in CAD patients receiving antiplatelet therapy that would be beneficial in appropriate personalized decision-making for these patients. METHODS In this prospective Optimal antiPlatelet Therapy for Chinese patients with Coronary Artery Disease (OPT-CAD, NCT01735305) registry, a total of 14 032 patients with CAD receiving at least one kind of antiplatelet agent were enrolled from 107 centers across China, from January 2012 to March 2014. The risk scoring system was developed in a derivation cohort (enrolled initially 10 000 patients in the database) using a logistic regression model and was subsequently tested in a validation cohort (the last 4032 patients). Points in risk score were assigned based on the multivariable odds ratio of each factor. Ischemic events were defined as the composite of cardiac death, myocardial infarction or stroke. RESULTS Ischemic events occurred in 342 (3.4%) patients in the derivation cohort and 160 (4.0%) patients in the validation cohort during 1-year follow-up. The OPT-CAD score, ranging from 0-257 points, consist of 10 independent risk factors, including age (0-71 points), heart rates (0-36 points), hypertension (0-20 points), prior myocardial infarction (16 points), prior stroke (16 points), renal insufficient (21 points), anemia (19 points), low ejection fraction (22 points), positive cardiac troponin (23 points) and ST-segment deviation (13 points). In predicting 1-year ischemic events, the area under receiver operating characteristics curve were 0.73 and 0.72 in derivation and validation cohort, respectively. The incidences of ischemic events in low- (0-90 points), medium- (91-150 points) and high-risk (≥151 points) patients were 1.6%, 5.5%, and 15.0%, respectively. Compared to GRACE score, OPT-CAD score had a better discrimination in predicting ischemic events and all-cause mortality (ischemic events: 0.72 vs 0.65, all-cause mortality: 0.79 vs 0.72, both P < .001). CONCLUSIONS Among CAD patients, a risk score based on 10 baseline clinical variables performed better than the GRACE risk score in predicting long-term ischemic events. However, further research is needed to assess the value of the OPT-CAD score in guiding the management of antiplatelet therapy for patients with CAD.