Relation of the HAS-BLED bleeding risk score to major bleeding, cardiovascular events, and mortality in anticoagulated patients with atrial fibrillation

Abstract

Background: Stroke risk in atrial fibrillation (AF) using oral vitamin K antagonists is closely related to bleeding risk. The HAS-BLED (hypertension, abnormal renal/liver function, stroke, bleeding history or predisposition, labile INR [international normalized ratio], elderly, drugs/alcohol concomitantly) bleeding score has demonstrated usefulness in assessing major bleeding risk in patients with AF. However, risk factors for warfarin-associated bleeding also predict stroke risk in patients with AF. We tested the usefulness of the HAS-BLED score for predicting both major bleeding and cardiovascular events in a cohort of anticoagulated patients with AF.

Methods and results: We recruited 965 consecutive anticoagulated outpatients with permanent or paroxysmal AF who were stabilized for at least 6 months on oral anticoagulation (international normalized ratio, 2.0-3.0). Medical history and HAS-BLED score were assessed. Cox regression models were used to determine the association between clinical risk factors and bleeding episodes, adverse cardiovascular events, and mortality. The median HAS-BLED score was 2 (range, 0-6; 29% with a score ≥3 [ie, high risk]). Median follow-up was 861 days (range, 718-1016 days). Independent predictors for major bleeding were age ≥75 years (hazard ratio [HR], 1.74; 95% CI, 1.05-2.87; P=0.030), male sex (HR, 1.70; 95% CI, 1.03-2.80; P=0.036), renal impairment (HR, 2.12; 95% CI, 1.20-3.73; P=0.010), previous bleeding episode (HR, 6.00; 95% CI, 3.73-9.67; P<0.001), current alcohol consumption (HR, 2.28; 95% CI, 1.03-5.06; P=0.043), and concomitant malignant disease (HR, 2.17; 95% CI, 1.13-4.18; P=0.020). Independent predictors for adverse cardiovascular events were age >75 years (HR, 2.20; 95% CI, 1.40-3.46; P=0.001), heart failure (HR, 1.78; 95% CI, 1.20-2.86; P=0.001), and previous stroke (HR, 1.85; 95% CI, 1.20-2.86; P<0.001). The HAS-BLED score was highly predictive for major bleeding events (HR, 2.04; 95% CI, 1.68-2.49; P<0.001) and adverse cardiovascular events (HR, 1.51; 95% CI, 1.27-1.81; P<0.001). The incidence of both bleeding and adverse cardiovascular events was higher as HAS-BLED score increased, and crude bleeding rates only exceeded thrombotic events at a HAS-BLED score >3. The HAS-BLED score also predicted all-cause mortality (HR, 1.68; 95% CI, 1.40-2.01; P<0.001).

Conclusions: The HAS-BLED score not only is useful in the assessment of bleeding risk, but also shows some predictive value for cardiovascular events and mortality in anticoagulated patients with AF, consistent with the relationship between thrombosis and bleeding. Nonetheless, the HAS-BLED score has been designed for predicting bleeding risk rather than thrombotic events per se, and specific risk scores for cardiovascular events and mortality should be applied for these events.