Comparison of the Predictive Power of the CRUSADE, MEHRAN, and ACTION Bleeding Risk Scores in Patients With the Acute Coronary Syndrome

Background: Various risk scores are used to predict the bleeding risk in patients with the acute coronary syndrome (ACS), including ACTION, CRUSADE, and MEHRAN. The purpose of the present study was to compare the accuracy of these 3 risk scores in the prediction of the bleeding risk in patients with the ACS. Methods: We studied 745 consecutive patients with the ACS undergoing coronary arteriography and calculated the ACTION, CRUSADE, and MEHRAN bleeding risk scores for all the patients. Then, we compared their prediction accuracy for major bleeding events and serious (major or minor) bleeding episodes with C-statistics. Results: The majority of the patients (72.6%) had non-ST-elevation myocardial infarction (NSTEMI), and the others (25.4%) had STEMI. The mean age of the patients was 62.55 +/- 12.12 years, and 62.4% were male. Bleeding complications were reported in 141 (18.9%) patients, with the catheterization site being the most frequent site of bleeding. The major bleeding rate was predicted in 29.1%, 28.4%, and 4.8% of the patients according to the CRUSADE, MEHRAN, and ACTION risk scores, respectively. The C-index values (AUC) for the ACTION, MEHRAN, and CRUSADE risk scores were 0.6182, 0.5413, and 0.6185, respectively. Pairwise comparisons between the scores showed no statistically significant differences in the discriminatory power between the ACTION and the CRUSADE (P=0.970); however, the differences between the CRUSADE and the MEHRAN (P=0.051) or between the ACTION and the MEHRAN (P=0.053) were near to significant. Conclusions: The bleeding risk score was predicted accurately by the ACTION, CRUSADE, and MEHRAN risk scores without significant differences among the 3 risk scores in our patients with the ACS undergoing coronary arteriography. These results showed that the predictive power was not excellent in any of the 3 risk scores.