Comparison of Events Across Bleeding Scales in the ENGAGE AF-TIMI 48 Trial.

BACKGROUND: Numerous scales exist for the classification of major bleeding events. Limited data compare the most commonly used bleeding scales within a single at-risk cohort of patients with atrial fibrillation. Here, we analyze bleeding outcomes according to the ISTH (International Society on Thrombosis and Hemostasis), TIMI (Thrombolysis in Myocardial Infarction), GUSTO (Global Usage of Strategies to Open Occluded Arteries), and BARC (Bleeding Academic Research Consortium) bleeding scales in the ENGAGE AF (Effective Anticoagulation With Factor Xa Next Generation in Atrial Fibrillation)-TIMI 48 trial (NCT00781391) of edoxaban versus warfarin. METHODS: A total of 21105 patients with atrial fibrillation at risk for stroke (CHADS(2) score >= 2) were enrolled in the ENGAGE AF-TIMI 48 trial comparing warfarin with a higher- (60/30 mg daily) or lower- (30/15 mg daily) dose edoxaban regimen. Median follow-up was 2.8 years. Bleeding events occurring among on-treatment patients were examined. Annualized event rates were calculated by the ISTH, TIMI, GUSTO, and BARC scales and compared across treatment arms. Cox proportional hazards for a first bleeding event of each type were calculated for higher-dose edoxaban regimen vs warfarin and lower-dose edoxaban regimen versus warfarin. RESULTS: A total of 10311 bleeding events were reported. In a comparison of the most severe events in each scale, ISTH major bleeding was the most common (n=1289), followed by TIMI major (n=548), GUSTO severe/life-threatening (n=347), and BARC 3c+5 (n=276) bleeding. Lower bleeding risk with edoxaban compared with warfarin was seen regardless of bleeding scale (higher-dose edoxaban regimen range: hazard ratio [HR], 0.47 [95% CI, 0.35-0.62] for BARC 3c+5 versus HR, 0.80 [95% CI, 0.71-0.91] for ISTH major; lower-dose edoxaban regimen range: HR, 0.32 [95% CI, 0.23-0.45] for BARC 3c+5 versus HR, 0.47 [95% CI, 0.41-0.55] for ISTH major). Furthermore, a gradient of more pronounced risk reduction with edoxaban was observed with greater severity of first bleeding event (higher-dose edoxaban regimen: HR, 0.47 [95% CI, 0.35-0.62] for BARC 3c+5 bleeds versus HR, 0.86 [95% CI, 0.81-0.91] for any BARC bleed; lower-dose edoxaban regimen: HR, 0.32 [95% CI, 0.23-0.45] for BARC 3c+5 bleeds versus HR, 0.68 [95% CI, 0.63-0.72] for any BARC bleed). The direction of this trend was consistent for both gastrointestinal bleeding and nongastrointestinal bleeding. CONCLUSIONS: Among patients with atrial fibrillation at risk for stroke, there was a >4-fold difference in the frequency of the most severe bleeding events across commonly used bleeding scales. Furthermore, the relative safety of edoxaban compared with warfarin tended to increase with greater severity of bleeding.