Analysis of Oral Anticoagulant Dosing and Adherence to Therapy Among Patients With Nonvalvular Atrial Fibrillation

IMPORTANCE: Although reduced doses of direct oral anticoagulants (DOACs) are approved for patients with nonvalvular atrial fibrillation (NVAF) at high risk of bleeding, little is known about dosing accuracy, particularly in patients with renal dysfunction. OBJECTIVE: To determine whether underdosing of DOACs is associated with longitudinal adherence to anticoagulation. DESIGN, SETTING, AND PARTICIPANTS: This retrospective cohort analysis used data from the Symphony Health claims data set. This national medical and prescription data set comprises 280 million patients and 1.8 million prescribers in the US. Patients included had at least 2 claims for NVAF between January 2015 and December 2017. The dates of analysis for this article were from February 2021 to July 2022. EXPOSURES: This study included patients with CHA(2)DS(2)-VASc scores of 2 or higher who were treated with a dose of DOACs who did and did not meet label-specified criteria for dose reduction. MAIN OUTCOMES AND MEASURES: Logistic regression models examined factors associated with off-label dosing (ie, dosing not recommended by US Food and Drug Administration [FDA] labeling), the association of creatinine clearance with recommended DOAC dosing, and the association of DOAC underdosing and excess dosing with 1-year adherence. RESULTS: Among the 86 919 patients included (median [IQR] age, 74 [67-80] years; 43 724 men [50.3%]; 82 389 White patients [94.8%]), 7335 (8.4%) received an appropriately reduced dose, and 10 964 (12.6%) received an underdose not consistent with FDA recommendations, meaning that 59.9% (10 964 of 18 299) of those who received a reduced dose received an inappropriate dose. Patients who received off-label doses of DOACs were older (median [IQR] age, 79 [73-85] vs 73 [66-79] years) and had higher CHA(2)DS(2)-VASc scores (median [IQR], 5 [4-6] vs 4 [3-6]) compared with patients who received appropriate doses (as recommended by FDA labeling). Renal dysfunction, age, heart failure, and the prescribing clinician being in a surgical specialty were associated with dosing not recommended by FDA labeling. Almost one-third of patients (9792 patients [31.9%]) with creatinine clearance less than 60 mL per minute taking DOACs were either underdosed or excess-dosed not consistent with FDA recommendations. For every 10-unit decrease in creatinine clearance, the odds of the patient receiving an appropriately dosed DOAC was lower by 21%. Treatment with underdosed DOACs was associated with a lower likelihood of adherence (adjusted odds ratio, 0.88; 95% CI, 0.83-0.94) and higher risk of anticoagulation discontinuation (adjusted odds ratio, 1.20; 95% CI, 1.13-1.28) by 1 year. CONCLUSIONS AND RELEVANCE: In this study of oral anticoagulant dosing, DOAC dosing that did not follow FDA label recommendations was observed in a substantial number of patients with NVAF, occurred more frequently in patients with worse renal function, and was associated with less-consistent long-term anticoagulation. These results suggest a need for efforts to improve the quality of DOAC use and dosing.