Abstract Aims Heart failure (HF) is a risk factor for major adverse events in atrial fibrillation (AF). Whether this risk persists on non-vitamin K antagonist oral anticoagulants (NOACs) and varies according to left ventricular ejection fraction (LVEF) is debated. Methods and results We investigated the relation of HF in the ETNA-AF-Europe registry, a prospective, multicentre, observational study with an overall 4-year follow-up of edoxaban-treated AF patients. We report 2-year follow-up for ischaemic stroke/transient ischaemic attack (TIA)/systemic embolic events (SEE), major bleeding, and mortality. Of the 13 133 patients, 1854 (14.1%) had HF. Left ventricular ejection fraction was available for 82.4% of HF patients and was <40% in 671 (43.9%) and ≥40% in 857 (56.1%). Patients with HF were older, more often men, and had more comorbidities. Annualized event rates (AnERs) of any stroke/SEE were 0.86%/year and 0.67%/year in patients with and without HF. Compared with patients without HF, those with HF also had higher AnERs for major bleeding (1.73%/year vs. 0.86%/year) and all-cause death (8.30%/year vs. 3.17%/year). Multivariate Cox proportional models confirmed HF as a significant predictor of major bleeding [hazard ratio (HR) 1.65, 95% confidence interval (CI): 1.20–2.26] and all-cause death [HF with LVEF <40% (HR 2.42, 95% CI: 1.95–3.00) and HF with LVEF ≥40% (HR 1.80, 95% CI: 1.45–2.23)] but not of ischaemic stroke/TIA/SEE. Conclusion Anticoagulated patients with HF at baseline featured higher rates of major bleeding and all-cause death, requiring optimized management and novel preventive strategies. NOAC treatment was similarly effective in reducing risk of ischaemic events in patients with or without concomitant HF.