Abstract Background Balancing the risks of stroke and bleeding is necessary to optimise oral anticoagulation use in clinical practice. Concerns remain over the increased bleeding risk in patients receiving NOACs, and long-term safety data in this population are limited. Purpose To report annualised rates of haemorrhagic stroke in various sub-populations of patients with atrial fibrillation (AF) treated with edoxaban during the 4-year follow-up of ETNA-AF-Europe. Methods ETNA-AF-Europe, a post-authorisation, observational study conducted across 776 sites from 10 European countries, assessed the risks and benefits of edoxaban use in patients with AF. Here, we present the annualised event rates of haemorrhagic stroke during the 4-year follow-up, that occurred overall (on-/off-edoxaban) and on-edoxaban, including sub-populations stratified by age, renal impairment, hepatic impairment, chronic concomitant antiplatelet use and CHA2DS2-VASc score. Patient characteristics were collected at baseline and annualised event rates were reported for outcomes. Results A total of 13,164 patients were included in the full analysis set. Baseline characteristics are reported in Table 1. The overall and on-edoxaban annualised haemorrhagic stroke rates were low (number of events, % [95% confidence interval]: 36, 0.1 [0.05;0.11] and 31, 0.1 [0.05;0.10] respectively). The overall haemorrhagic stroke rates remained low in subgroups stratified by age (<65 years: 4, 0.1 [0.00;0.11]; ≥65–75 years: 13, 0.1 [0.04;0.13]; ≥75 years: 19, 0.1 [0.05;0.12]), renal impairment (yes: 23, 0.1 [0.05;0.13]; no: 12, 0.1 [0.03;0.11]), hepatic impairment (yes: 0, 0.0 [0.00;0.00]; no: 35, 0.1 [0.06;0.11]), concomitant antiplatelet use (yes: 2, 0.1 [0.00;0.16]; no: 34, 0.1 [0.05;0.11]) and CHA2DS2-VASc score (low [0–1]: 0, 0.0 [0.00;0.00]; moderate [2–4]: 28, 0.1 [0.06;0.12]; high [>4]: 7, 0.1 [0.03;0.18]) (Figure). Likewise, on-edoxaban haemorrhagic stroke rates were low irrespective of age (<65 years: 3, <0.1 [0.00;0.010; ≥65–75 years: 12, 0.1 [0.04;0.13]; ≥75 years: 16, 0.1 [0.04;0.12]); renal impairment (yes: 20, 0.1 [0.05;0.12]; no: 10, 0.1 [0.02;0.10]); hepatic impairment (yes: 0, 0.0 [0.00;0.00]; no: 30, 0.1 [0.05;0.11]); chronic concomitant antiplatelet use (yes: 2, 0.1 [0.00;0.18; no: 29, 0.1[0.05;0.10]) or CHA2DS2-VASc score (low [0–1]: 0, 0.0 [0.00;0.00]; moderate [2–4]: 25, 0.1 [0.05;0.12]; High [>4]: 6, 0.1 [0.02;0.17]) (Figure). Conclusions The overall and on-edoxaban annualised rates of haemorrhagic stroke were low and were not increased by non-modifiable risk factors such as age, renal/hepatic impairment, concomitant antiplatelet use and high CHA2DS2-VASc scores during the 4-year follow-up. The overall annualised rates on-edoxaban in ETNA-AF-Europe registry are quite comparable with incidence rates reported in the literature in age-stratified populations (incidence/100 person-years: 55-64 years: 0.06 [0.04-0.07]; 65-74 years: 0.1 [0.09-0.1]; 75-84 years 0.2 [0.1-0.2]).Baseline Characteristics