The Prognostic Significance of Different Bleeding Classifications in off-pump coronary artery bypass grafting

BACKGROUND: Perioperative bleeding during cardiac surgery are known to make patients susceptible to adverse outcomes and several bleeding classifications have been developed to stratify the severity of bleeding events. Further validation of different classifications was needed. The aim of present study was to validate and explore the prognostic value of different bleeding classifications in patients undergoing off-pump coronary artery bypass grafting (OPCAB). METHODS: Data on baseline and operative characteristics of 3988 patients who underwent OPCAB in Beijing Anzhen Hospital from February 2008 to December 2014 were available. The primary endpoint was a composite of in-hospital death and nonfatal postoperative myocardial infarction (MI). The secondary endpoint was postoperative acute kidney injury (AKI). We explored the association of major bleeding defined by the European registry of Coronary Artery Bypass Grafting (E-CABG), Universal Definition of Perioperative Bleeding (UDPB), Bleeding Academic Research Consortium (BARC) classification and Study of Platelet Inhibition and Patient Outcomes (PLATO) with primary endpoints by multivariable logistic regression analysis and investigated their significance of adverse event prediction using goodness-of-fit tests of − 2 log likelihood. RESULTS: In-hospital mortality was 1.23% (n = 49) and postoperative MI was observed in 4.76% (n = 190) of patients, AKI in 24.69% (n = 985). The incidence of the primary outcome was 5.99% (n = 239). Multivariable logistic regression analysis showed that BARC type 4 (OR = 2.64, 95% CI: 1.66–4.19, P < 0.001), UDPB class 4 (OR = 3.52, 95% CI: 2.05–6.02, P < 0.001) and E-CABG class 2–3 (class 2: OR = 2.24, 95% CI: 1.36–3.70, P = 0.001; class 3: OR = 12.65, 95% CI: 2.74–18.43, P = 0.002) bleeding but not PLATO bleeding were associated with an increased risk of in-hospital death and postoperative MI. Major bleeding defined by all the four classifications mentioned above was an independent risk factor of AKI after surgery. Inclusion of major bleeding defined by these four classifications improved the predictive performance of the multivariable model with baseline characteristics. CONCLUSIONS: Bleeding assessed by BARC, E-CABG and UDPB classifications were significantly associated with poorer immediate outcomes. These classifications seemed to be valuable tool in the assessment of prognostic effect of perioperative bleeding.