Background:
Direct oral anticoagulants (DOACs), including rivaroxaban, dabigatran, apixaban and edoxaban, have been approved as alternatives to warfarin for the primary treatment of venous thromboembolism (VTE). However, understanding of their comparative effectiveness in practice-based populations is limited.
Objective:
Among anticoagulant-naïve VTE patients, estimate the association of type of oral anticoagulant (OAC) with the rate of bleeding resulting in hospitalization.
Methods:
Patients with VTE and prescription for an OAC were identified from the US Truven Health MarketScan® Commercial and Medicare Supplemental databases for the period from 2011-2015. Hospitalization related to bleeding events (inclusive of intracranial, gastrointestinal and other) was defined using a validated algorithm. In head-to-head comparisons, initiators of a specific OAC were matched with up to 5 initiators of the comparing OAC by age, sex, and time since database enrollment. Cox proportional hazards regression was used to calculate hazard ratios (HR) and 95% confidence intervals (95%CI) for bleeding by OAC, adjusted for age, sex, and a comorbidity propensity score (created using prevalence of 20 common diagnoses and procedures).
Results:
The final analysis included 83,831 VTE patients who were 49.9% female and on average (standard deviation) 59.0 (16.0) years old. Of these, the initial OAC prescribed for 2,604 was apixaban, for 1,669 dabigatran, for 28,518 rivaroxaban, and for 48,514 warfarin. A total of 1,947 bleeding events occurred over an average of 13 months. Compared to new warfarin users, risk of bleeding was lower among patients initiating apixaban [HR (95%CI): 0.55 (0.36, 0.83)] and rivaroxaban [0.80 (0.72, 0.89)], but similar among new dabigatran users [0.96 (0.72, 1.27)]. In head-to-head DOAC comparisons, relative to rivaroxaban, risk of bleeding was lower among users of apixaban [0.57 (0.36, 0.89)] but similar for users of dabigatran [1.05 (0.73, 1.51)]. Due to low numbers we did not conduct analyses of edoxaban or a head-to-head comparison of dabigatran versus apixaban.
Conclusion:
In this practice-based population of 83,831 patients prescribed OACs for the treatment of VTE, subsequent risk of hospitalized bleeding was lowest among those prescribed apixaban, intermediate among those prescribed rivaroxaban, and highest among those prescribed warfarin and dabigatran. These data demonstrate that differences in bleeding risk exist by DOAC. While risk factors for bleeding might impact choice of warfarin versus the DOACs, the choice between DOACs may be less likely to be based on patient conditions.
Corrigendum to “Comparative effectiveness of oral anticoagulants in venous thromboembolism: GARFIELD-VTE” [Thromb. Res. 2020 Jul; 191:103–112]
