Predictive scores for major bleeding after coronary artery bypass surgery in low operative risk patients.

Background: Cardiac surgery is associated with perioperative bleeding and carries high risk of allogeneic blood transfusion. Recently new scores for prediction of severe bleeding have been developed. This study aims to compare the WILL-BLEED, CRUSADE, PAPWORTH, TRUST, TRACK and ACTION scores in predicting major bleeding after CABG in patients with low estimated operative risk.

Methods: A multicenter observational study included 1391 patients who underwent isolated CABG from July 2015 to January 2018. We tested the hypothesis that the WILL-BLEED score, specifically designed for CABG, would perform at least as well as the CRUSADE, PAPWORTH, TRUST, TRACK and ACTION scores in predicting postoperative major bleeding in low operative risk patients. The primary endpoint was the performance of known bleeding risk scores after CABG. The secondary endpoint was the evaluation of in-hospital mortality.

Results: Mean age was 68.2±9.4 years and median Euroscore II value was 1.69% (IQR 1.15-2.81%). Mean blood losses in the first 12 postoperative hours was 339.75 mL. Seventy-three (5.2%) subjects underwent administration of blood products. The rate of severe-massive bleeding according to UDPB grades 3-4 was 1.5%. WILL-BLEED, TRUST, TRACK and ACTION scores were significantly associated with severe postoperative bleeding. WILL-BLEED presented the best c-index (AUC: 0.658; 95% CI: 0.600,0.716). Reclassification analysis showed a worsening in sensitivity and significant negative reclassification of CRUSADE, PAPWORTH, TRACK and ACTION scores when compared with WILL-BEED. The combination of WILL-BLEED and TRUST scores improved the prediction ability (AUC: 0.673; 95% CI: 0.615-0.732). Overall in-hospital mortality was 1.65%. Early mortality in patients with severe versus no-severe bleeding was found to be 11.8% vs. 1.0% Severe bleeding (OR: 13.26; P value<0.001) was found to be significantly associated with early mortality.

Conclusions: Severe bleeding after CABG is a harmful event associated with adverse outcomes. WILL-BLEED Score has the better performance in predicting severe-massive bleeding after CABG. The TRUST Score, although suboptimal, represents a valuable alternative in this setting.