Clinical outcomes of direct oral anticoagulant off-label dosing in nonvalvular atrial fibrillation

Direct oral anticoagulants (DOACs) used in nonvalvular atrial fibrillation (NVAF) are superior or non-inferior to warfarin in reducing the risk of stroke while at the same time having a similar or lower risk of bleeding. However, reduced doses are prescribed more often than expected from clinical practice and off-label underdosing is a frequent issue. The objective of this study was to compare effectiveness and safety between guideline and off-label dosing of DOACs. Auricula, a Swedish anticoagulation registry was used in identifying eligible patients from July 2011 to December 2017. The study cohort consisted of 47,355 patients with newly initiated DOAC (apixaban, dabigatran, or rivaroxaban) after exclusion of 92,316 patients due to concomitant venous thromboembolism, previous mechanical heart valve (MHV) or previous data entry in Auricula. The median durations of follow up were 403, 419, 373 and 209 days in the on-label standard dose cohort, off-label reduced dose cohort, on-label reduced dose cohort and the off-label standard dose cohort respectively. Endpoints (stroke and major bleeding) and baseline characteristics were collected from hospital administrative registers using ICD-10 codes or the Swedish Stroke register. Cohorts were compared using weighted adjusted Cox regression after full optimal matching based on propensity scores. Off-label underdosing of DOACs (n=6,187, 9.7%) was associated with higher risk of major bleeding HR 1.16 (95% CI 1.05–1.27), other bleeding HR 1.16 (1.04–1.30), myocardial infarction HR 1.47 (1.20–1.80), ischemic stroke HR 1.25 (1.04–1.50) and all-cause mortality HR 1.52 (1.37–1.69) compared to on-label standard dosing (n=35,065, 55.2%). Among off-label underdosed DOACs, dabigatran was associated with higher risk of all-cause stroke 1.86 (1.07–3.23), ischemic stroke HR 1.97 (1.10–3.52) and all-cause stroke and systemic embolism HR 1.92 (1.11–3.32) compared to apixaban. Rivaroxaban was associated with major bleeding HR 1.70 (1.41–2.03), gastrointestinal bleeding HR 1.92 (1.33–2.77), and other bleeding HR 1.97 (1.57–2.47) compared to apixaban. The study could not show any differences comparing off-label overdosing of DOACs and on-label reduced dosing, besides lower risk of all-cause mortality HR 0.69 (0.52-0.93) in the overdosed patients. In this large observational registry-based NVAF cohort, underdosing of DOACs is associated with higher risk of ischemic and all-cause stroke but also major bleeding when compared to on-label dosing. Underlying DOAC therapy may need to be tailored to the specific patient when choosing off-label reduced dosing since underdosed rivaroxaban is associated with higher risk of bleeding complications, and underdosed dabigatran is associated with higher risk of stroke complications, when comparing both with apixaban. • Underdosing of DOACs is associated with higher risk of ischemic and all-cause stroke but also major bleeding when compared to on-label dosing. Possibly, • Underdosed rivaroxaban is associated with higher risk of bleeding complications when compared with apixaban. • Underdosed dabigatran is associated with higher risk of stroke complications when compared with apixaban.