Direct oral anticoagulants (DOACs) are effective to prevent strokes; however, bleeding risk of DOACs in very old and frail geriatric patients is poorly known.

: To assess bleeding risk in geriatric patients aged ≥80 years with AF newly treated with rivaroxaban.

Subjects from 33 centers, with non-valvular AF newly treated with rivaroxaban were enrolled and followed-up every 3 months for 12 months. Clinical and biological data were recorded. Major bleedings were recorded at each visit. The cohort was compared with a cohort treated with vitamin K antagonists (VKA) from the same centers (n=924).

In total, 995 subjects had a 1-year follow-up. The mean (standard deviation (SD)) age was 86.0 (4.3) years, 23% were aged 90 years and older. 61% were women and 48% had an estimated glomerular filtration rate (eGFR)<50mL/min. The main comorbidities were hypertension in 77% of subjects, malnutrition 49%, anemia 41%, dementia 39%, and falls 27%. The mean (SD) score for CHA2DS2-VASc was 4.8 (1.4), HAS-BLED 2.3 (0.9), Mini-Mental State Examination 21.5 (6.9), Activities of Daily Living 4.4 (1.9), and Charlson Comorbidity Index (CCI) 6.7 (2.0). The one-year rate of major bleeding events was 6.4% of which 0.8% were fatal and 1.1% intracranial hemorrhages (ICH), whereas the one-year rate of ischemic stroke was 1.4% and all-cause mortality 17.9%. Compared with VKA cohort and adjusted for age, gender, eGFR, weight, anemia, falls and CCI, antiplatelet, amiodarone, proton pump inhibitor there was a significant lower rate of major bleeding among rivaroxaban users hazard ratio of 0.67 (95% confidence interval, 0.45–0.99) and lower ICH HR=0.52 (0.21–1.28). No difference was observed for ischemic stroke HR=0.76 (0.32–1.80), and all-cause mortality HR=1.05 (0.82–1.34).

In this first large-scale prospective study in geriatric population with AF, major bleeding rates were significantly lower with rivaroxaban than with VKAs.