On-label versus off-label direct oral anticoagulant dosing patterns and associated outcomes in atrial fibrillation: a real-world cohort analysis

Background Atrial fibrillation (AF) is the most common heart rhythm disorder worldwide. Direct oral anticoagulants (DOACs) are recommended for stroke prevention in patients with AF. The 2024 European Society of Cardiology Guidelines note that off-label use of anticoagulation may contribute to worse clinical outcomes, reinforcing the importance of on-label dosing. Evidence on real-world dosing patterns and associated clinical outcomes is needed to support on-label DOAC use in routine clinical practice. Purpose To describe characteristics and dosing patterns among patients with AF receiving DOACs and evaluate the impact of on-label vs off-label DOAC dosing on effectiveness and safety outcomes. Methods We conducted a retrospective observational cohort study using the UK Clinical Practice Research Datalink (Gold and Aurum). Adults (≥18 years) with AF initiating any currently approved DOAC between January 2016–June 2024 were eligible. Patients with prior oral anticoagulant use, venous thromboembolism, end-stage renal disease, valvular AF, mechanical valves, pregnancy, and invalid baseline/follow-up data (e.g., pre-index death or 0-day follow-up) were excluded. Eligible patients with determinable dosing information were followed from first DOAC prescription until discontinuation, treatment switch, death, emigration, or study end (December 2024). Clinical outcomes included effectiveness (ischemic stroke and systemic embolism) and safety (major bleeding), assessed in two cohorts: on-label dosing and off-label dosing. Propensity score overlap weighting (PSOW) addressed confounding and cohort size imbalance, with covariate balance assessed post-weighting. Associations between dosing group and clinical outcomes were evaluated using Cox models before and after PSOW. Results A total of 306,846 patients met the study criteria (171,218 [55.8%] apixaban, 68,620 [22.4%] rivaroxaban, 6,349 [2.0%] dabigatran, and 60,659 [19.8%] edoxaban). Among them, 115,040 had determinable dosing (61,501 apixaban, 20,078 rivaroxaban, 6,349 dabigatran, and 27,112 edoxaban). Overall, off-label dosing occurred in 14.0% of patients, most frequently those receiving dabigatran and apixaban. Patients receiving off-label doses were more often younger (<65 years; 41.9% vs 25.0%) and male (59.7% vs 50.7%). Event rates per 100 person-years of effectiveness were lower with on-label vs off-label groups (1.2 vs 1.5); similarly, safety event rates favored the on-label group (0.7 vs 1.0), and mortality rate was lower at 3.6 vs 5.28 per 100 person-years. Compared with on-label dosing, off-label dosing was associated with higher risks of ischemic stroke, systemic embolism, major bleeding, and mortality. Sensitivity analyses with 1-year follow-up showed similar results. Conclusion In this real-world analysis, on-label DOAC dosing was associated with more favorable clinical outcomes than off-label dosing, highlighting the importance of proper dose selection.