Epistaxis in anticoagulated patients with atrial fibrillation in the ENGAGE AF-TIMI 48 trial

Abstract

Abstract Background Epistaxis is common with antithrombotic therapy yet under-investigated in cardiovascular clinical trials and observational studies. These bleeding events are often troublesome to patients and may lead to anticoagulant discontinuation. Purpose To describe the frequency, severity, and outcomes of epistaxis in patients with atrial fibrillation (AF) randomized to edoxaban vs warfarin. Methods ENGAGE AF-TIMI 48 randomized 21,105 patients with AF and CHADS2 ≥2 to a higher-dose edoxaban regimen (HDER; 60mg daily), a lower-dose edoxaban regimen (LDER; 30mg daily), or warfarin, with edoxaban doses reduced by 50% in patients meeting dose-reduction criteria. Location and severity of bleeding were adjudicated by a blinded Clinical Events Committee using ISTH criteria. Patients with intracranial hemorrhage were excluded from this analysis. Patients with more than one bleeding event were categorized according to the most severe event. The safety cohort with interval censoring during study drug interruption was analyzed. Proportions were compared using a Pearson's chi-squared test and treatment arms were compared using a Cox proportional hazards model. Results 5,247 patients with a bleeding event were included in this analysis, including 1,008 (19.2%) with epistaxis and 4,239 (80.8%) with exclusively non-epistaxis bleeding. Baseline characteristics were similar for patients with epistaxis as compared to non-epistaxis bleeding. Epistaxis events were overall less severe than non-epistaxis bleeds (ISTH major: 3.2% vs 20.7%; CRNM: 64.7% vs 60.1%; minor: 32.1% vs 19.2%; p<0.001; Fig. 1, Panel A). Two life-threatening and no fatal epistaxis bleeds occurred. Permanent study drug discontinuation was similar following epistaxis vs non-epistaxis bleeding in patients with major (59.4% vs 53.6%; p=0.52) or CRNM bleeding (32.5% vs 33.3%; p=0.70), but was significantly higher after minor epistaxis versus minor bleeding at other sites (33.3% vs 23.9%; p=0.001; Fig. 1, Panel B). Compared to warfarin, hazard ratios (HR) for epistaxis bleeding were: 1) major: HDER 0.47 (0.19–1.15), LDER 0.65 (0.29–1.45); 2) major/CRNM: HDER 1.00 (0.84–1.19), LDER 0.70 (0.58–0.85); 3) major/CRNM/minor: HDER 1.09 (0.95–1.26), LDER 0.73 (0.62–0.86) (Fig. 1, Panel C). Conclusion Epistaxis was frequent in patients with atrial fibrillation on anticoagulation. When compared to warfarin, LDER reduced the risk of epistaxis by 27–30% while HDER neither increased nor decreased these events. There were significantly higher rates of permanent drug discontinuation following minor epistaxis as compared to minor bleeding at other sites. These findings suggest that epistaxis is symptomatically important, may cause disproportionate interruption in antithrombotic therapy, and deserves increased attention in cardiovascular studies. Funding Acknowledgement Type of funding sources: Private company. Main funding source(s): Daiichi Sankyo Pharma Development Figure 1