The real-world effectiveness and safety of edoxaban treatment in 27,333 Global ETNA-AF programme patients with and without a history of heart failure

Abstract

Background Heart failure (HF) occurs in approximately 26% of patients with atrial fibrillation (AF). Real-world data of oral anticoagulation with edoxaban in AF patients with HF history are limited. Purpose To compare edoxaban effectiveness and safety in AF patients with or without HF history. Methods The Global ETNA-AF programme (EU: NCT02944019, Japan: UMIN000017011, South Korea/Taiwan: NCT02951039) integrates data from multiple prospective, observational, noninterventional regional studies of AF patients receiving edoxaban for stroke prevention. This snapshot analysis summarises baseline characteristics and 2-year annualised rates of all-cause death, cardiovascular (CV) death, stroke (haemorrhagic, ischaemic, any), and bleeding (major bleeding [MB], major gastrointestinal [GI] bleeding, intracranial haemorrhage [ICH], clinically relevant nonmajor bleeding [CRNMB], and any bleeding) in patients with or without HF history. Univariate Cox regression models assessed clinical outcomes. Results Data from 27,333 patients (5258 with HF history) from Europe, Japan, South Korea, and Taiwan were analysed. Patients with HF history were significantly older and had lower mean body weight and creatinine clearance (P<0.0001 all; Table). Patients with HF history had significantly higher baseline stroke (CHA2DS2-VASc) and bleeding (HAS-BLED) risk scores (P<0.0001 both; Table). Significantly more patients with HF history reported previous experiences with MB (P=0.001) and major GI bleeding (P=0.007); these patients were also more likely to receive 30 mg edoxaban vs 60 mg edoxaban (P<0.0001; Table). Patients with HF history had significantly (P<0.0001 both) higher rates of all-cause (6.1% vs 2.5%; hazard ratio [HR] (95% confidence interval [CI]), 2.41 [2.17–2.68]) and CV death (2.8% vs 1.2%; HR [95% CI], 2.39 [2.05–2.80]), and fatal bleeding (0.3% vs 0.2%; HR [95% CI], 1.86 [1.20–2.89]; Figure). The proportion of all-cause deaths that were fatal bleeding events was 6% and 7% for patients with and without HF, respectively. Additionally, patients with HF history had significantly (P<0.0001 both) higher rates of MB (1.7% vs 0.9%; HR [95% CI], 1.87 [1.53–2.28]) and major GI bleeding (1.1% vs 0.4%; HR [95% CI], 2.69 [2.07–3.49]), with a greater proportion of MB events classified as major GI bleeding (64.5% vs 44.8%; P<0.0001). Patients with HF history also had significantly (P<0.0001 both) higher rates of CRNMB (HR [95% CI], 1.87 [1.58–2.21]) and any bleeding (HR [95% CI], 1.49 [1.34–1.65]). Rates of ICH and haemorrhagic stroke were similar in both groups. Conclusions In AF patients receiving edoxaban, the rates of MB, major GI bleeding, and CV or all-cause death were higher when comparing those with versus without HF history. The higher incidence of MB and major GI bleeding in patients with HF history did not lead to proportionally higher fatal bleeding rates among all-cause deaths. Funding Acknowledgement Type of funding sources: Private company. Main funding source(s): Daiichi Sankyo