Off-Label Dosing of Direct Oral Anticoagulants Among Inpatients With Atrial Fibrillation in the United States

BACKGROUND: Among patients hospitalized for atrial fibrillation, the frequency of off-label direct oral anticoagulant (DOAC) dosing, associated factors, hospital-level variation, and temporal trends in contemporary practice are unknown.METHODS: Using the Get With The Guidelines-Atrial Fibrillation registry, patients admitted from January 1, 2014, to March 31, 2020, and discharged on DOACs were stratified according to receipt of underdosing, overdosing, or recommended dosing. Factors associated with off-label dosing (defined as underdosing or overdosing) were identified using logistic regression. Median odds ratio (OR) and time-series analyses were used to assess hospital-level variation and temporal trends, respectively.RESULTS: Of 22 470 patients (70.1 +/- 12.1 years, 48.1% female, 82.5% White) prescribed a DOAC at discharge from hospitalization for atrial fibrillation (66% apixaban, 29% rivaroxaban, and 5% dabigatran), underdosing occurred among 2006 (8.9%), overdosing among 511 (2.3%), and recommended dosing among 19 953 (88.8%). The overall rate of off-label dosing was 11.2%. Patient-related factors associated with off-label dose included age (underdosing: OR, 1.06 per 1-year increase [95% CI, 1.06-1.07]; overdosing: OR, 1.07 per 1-year increase [95% CI, 1.06-1.09]), dialysis dependence (underdosing: OR, 5.50 [95% CI, 3.76-8.05]; overdosing: OR, 5.47 [95% CI, 2.74-10.88]), female sex (overdosing: OR, 0.79 [95% CI, 0.63-0.99]), and weight (overdosing: OR, 0.96 per 1-kg increase [95% CI, 0.95-1.00]). Across hospitals, the adjusted median OR for off-label DOAC dose was 1.45 (95% CI, 1.34-1.65; underdosing: OR, 1.52 [95% CI, 1.39-1.76]; overdosing: OR, 1.32 [95% CI, 1.20-1.84]), indicating significant hospital-level variation. Over the study period, recommended dosing significantly increased over time (81.9%-90.9%; P<0.0001 for trend) with a corresponding decline in underdosing (14.4%-6.6%; P<0.0001 for trend) and overdosing (3.8%-2.5%; P=0.001 for trend).CONCLUSIONS: Over 1 in 10 patients hospitalized for atrial fibrillation are discharged on an off-label DOAC dose with significant variation across hospitals. While the proportion of patients receiving recommended dosing has significantly improved over time, opportunities to improve DOAC dosing persist.