Background: The Bleeding Academic Research Consortium (BARC) definition was proposed to overcome the heterogeneity among the many bleeding definitions. The aim of this study-level meta-analysis was to explore the incidence of BARC-assessed bleeding in acute coronary syndrome (ACS) studies and to ascertain the relation between these events and variables related to bleeding risk.
Methods and results: We searched the literature for studies that reported bleeding events according to BARC criteria in ACS patients. An analysis on heterogeneity between studies in bleeding reports was performed with I test. A meta-regression was conducted to explore the relation between different types of BARC bleedings and patient and procedural features. Nine studies were included in the analysis. Overall, BARC 2 rates were higher than BARC 3 or 5 rates (6.3 versus 2.6%). An extremely high level of heterogeneity was detected both for BARC 2 (I 99.3%) and BARC 3 or 5 (I 97.5%) bleedings. Increasing age [β coefficient 0.4% (0.2-0.6%); P < 0.001] and renal impairment [β coefficient 1 6.5% (1-32.1%); P = 0.037] were associated with increased BARC 3 or 5 rates, whereas the use of glycoprotein IIb/IIIa inhibitors was the only factor related to an increased incidence of BARC 2 bleeding [β coefficient 2 2.3% (5.5-39%); P = 0.009].
Conclusion: The high level of heterogeneity in BARC bleeding reports only partially explained by bleeding risk profile suggests that a regulatory guidance to properly evaluate bleedings and to estimate the risk--benefit in clinical trials investigating different antithrombotic treatments in ACS patients is needed.