Performance of bleeding risk-prediction scores in patients with atrial fibrillation undergoing percutaneous coronary intervention

Purpose: The HAS-BLED (Hypertension, Abnormal Renal/Liver Function, Stroke, Bleeding History or Predisposition, Labile International Normalized Ratio, Elderly, Drugs/Alcohol), ATRIA (Anticoagulation and Risk Factors in Atrial Fibrillation), and mOBRI (modified Outpatient Bleeding Risk Index) schemes are validated bleeding risk-prediction tools, but their predictive performance in patients with AF receiving multiple antithrombotic drugs after percutaneous coronary intervention (PCI) is unknown. We sought to compare the predictive performance of bleeding risk-estimation tools in a cohort of patients with atrial fibrillation (AF) undergoing PCI. Methods: AFCAS (Management of patients with Atrial Fibrillation undergoing Coronary Artery Stenting) is a multicenter European prospective registry enrolling patients with AF undergoing PCI. We calculated HAS-BLED, ATRIA, and mOBRI bleeding risk-prediction scores and assessed the rate of bleeding complications as defined by bleeding academic research consortium (BARC) as well as all cause mortality at 12 months follow-up in 929 consecutive patients undergoing PCI. Results: According to low, intermediate and high bleeding risk scores, the rates of major bleeding events were 10.0% vs. 11.0% vs. 10.2% (p=0.94) for HAS-BLED; 9.0% vs. 13.9% vs. 12.2% (p=0.27) for ATRIA; and 8.3% vs. 10.5% vs. 10.5% (p=0.85) for mOBRI, respectively. Higher ATRIA scores were significantly associated with increased mortality (7.7% vs. 8.9% vs. 18.7% (p=0.001) for low, intermediate and high scores, respectively), but this was not evident for HAS-BLED and mOBRI. Patients with high ATRIA scores had increased risk for mortality (OR 1.7, 95% CI 1.3-2.2, p<0.001) compared to those with low scores. Conclusions: The performance of low/moderate/high ATRIA, HAS-BLED and mOBRI scores in discriminating bleeding complications in AF patient undergoing PCI was limited. The ATRIA score appears to be a better predictor for mortality rather than for bleeding.