Comparison of ischemic and bleeding risk scores in non-ST elevation acute coronary syndromes

OBJECTIVES: Compare the discriminatory performance of two validated ischemic risk scores and a bleeding risk score for in-hospital adverse events across the spectrum of non-ST elevation acute coronary syndromes (NSTE-ACS). MATERIAL AND METHODS: Single center, 516 consecutive patients admitted with the diagnosis of NSTE-ACS. The following risk scores were calculated for each patient: TIMI, GRACE and CRUSADE. The following in-hospital endpoints were used: major bleeding (as defined by the CRUSADE criteria); recurrent ischemia (re-infarction or recurrent angina); and death. Discriminatory performance was measured by the c-statistic and compared. RESULTS: There were 36 major bleeding events, 34 recurrent ischemic events and 10 deaths. TIMI RS, GRACE RS, and CRUSADE RS demonstrated fair discriminatory accuracy for major bleeding (c-statistic = 0.64, 0.58, and 0.61, respectively). GRACE and CRUSADE risk scores demonstrated a better performance than TIMI RS for predicting in-hospital death (c-statistic = 0.92 and 0.86, respectively versus c-statistic = 0.63, P < 0.001). For the combined endpoint of in-hospital death or recurrent ischemia there was no statistically significant difference between CRUSADE RS, GRACE RS, and TIMI RS (c-statistic = 0.58, 0.57, and 0.62, respectively). CONCLUSION: Both ischemic and bleeding risk scores are able to predict in-hospital bleeding, ischemic and fatal events.