Objective During the first wave of the SARS-CoV-2 pandemic, management of anticoagulation therapy in hospitalized patients with atrial fibrillation (AF) was simplified to low-molecular-weight heparin (LMWH) followed by oral anticoagulation, mainly owing to the risk of drug–drug interactions. However, not all oral anticoagulants carry the same risk. Methods Observational, retrospective, and multicenter study that consecutively included hospitalized patients with AF anticoagulated with LMWH followed by oral anticoagulation or edoxaban concomitantly with empirical COVID-19 therapy. Time-to-event (mortality, total bleeds, and admissions to ICU) curves, using an unadjusted Kaplan-Meier method and Cox regression model adjusted for potential confounders were constructed. Results A total of 232 patients were included (80.3 ± 7.7 years, 50.0% men, CHA2DS2-VASc 4.1 ± 1.4; HAS-BLED 2.6 ± 1.0). During hospitalization, patients were taking azithromycin (98.7%), hydroxychloroquine (89.7%), and ritonavir/lopinavir (81.5%). The mean length of hospital stay was 14.6 ± 7.2 days, and total follow-up was 31.6 ± 13.4 days; 12.9% of patients required admission to ICU, 18.5% died, and 9.9% had a bleeding complication (34.8% major bleeding). Length of hospital stay was longer in patients taking LMWH (16.0 ± 7.7 vs 13.3 ± 6.5 days; P = .005), but mortality and total bleeds were similar in patients treated with edoxaban and those treated with LMWH followed by oral anticoagulation. Conclusions Mortality rates, arterial and venous thromboembolic complications, and bleeds did not significantly differ between AF patients receiving anticoagulation therapy with edoxaban or LMWH followed by oral anticoagulation. However, the duration of hospitalization was significantly lower with edoxaban. Edoxaban had a similar therapeutic profile to LMWH followed by oral anticoagulation and may provide additional benefits.