Once- versus twice-daily direct oral anticoagulants after ischemic stroke in atrial fibrillation - A post-hoc analysis of the ELAN trial.

Introduction
Whether the risk-benefit profile of once-daily versus twice-daily direct oral anticoagulants (DOAC) differs after atrial fibrillation(AF)-associated ischemic stroke is unclear. We explored this in a post-hoc analysis of ELAN trial data (NCT03148457).Patients And Methods
We compared the risk of the primary outcome (recurrent ischemic stroke, systemic embolism, intracranial hemorrhage (ICH), major extracranial bleeding, vascular death) from treatment initiation to the trial's 90-day follow-up in participants treated with once-daily or twice-daily DOAC after AF-associated stroke using Firth's logistic and Cox proportional hazards regression in unadjusted, inverse-probability-of-treatment-weighted and augmented-inverse-probability-weighted models to address confounding. Secondary outcomes were the primary outcome components and non-major bleeding. We calculated the net clinical benefit (NCB) of twice-daily over once-daily DOAC by subtracting the weighted rate of excess bleeding attributable to twice-daily DOAC from the rate of excess ischemic events possibly prevented by twice-daily DOAC.Results
We analyzed 1890/2013 (94%) participants (median age 77 years, 45% female), of whom 384 (20%) received once-daily and 1506 (80%) twice-daily DOAC. The primary outcome occurred in 64 (3.4%) participants, and did not differ between DOAC types in logistic (ORunadjusted 0.89 (95% CI 0.50-1.66); ORweighted 1.34 (0.71-2.79); ORaugmented 1.45 (0.81-3.21); twice-daily vs once-daily DOAC) nor in Cox models. We identified no clear differences in any secondary outcome. NCB analysis revealed a near-neutral net effect of twice-daily versus once-daily DOAC (+0.28 to +0.67 weighted events possibly prevented/100 person-months for ICH weights 1.5-3.3).Discussion And Conclusion
The risk-benefit profile of once-daily versus twice-daily DOAC after AF-associated ischemic stroke does not seem to differ.