Objective:
The study aim was to compare major bleeding risk and health care costs after initiating oral anticoagulants (OACs) for treatment-naïve non-valvular atrial fibrillation (NVAF) patients.
Methods:
Patients in the Medicare advantage population prescribed apixaban, rivaroxaban, dabigatran or warfarin were selected from the Optum Research Database 01JAN2013-31DEC2014. The first OAC prescription date was designated as the index date. Patients were required to have a NVAF diagnosis, continuous health plan enrollment for 6 months and no OAC claims before the index date. Patients were classified into four cohorts based on their index OAC prescription. Major bleeding events, identified by the Cunningham algorithm plus additional bleeding sites, were compared using a Cox proportional hazards model. Health care costs were calculated per patient per month and compared using generalized linear models.
Results:
The study included 36,260 patients: 3,762 apixaban, 2,677 dabigatran, 8,740 rivaroxaban, and 21,081 warfarin patients. CHA2DS2-VASc score was higher in apixaban patients (4.2) compared to dabigatran and rivaroxaban (both 4.0; p<0.001), but lower than in warfarin patients (4.3; p<0.001). After adjusting for baseline characteristics, apixaban patients were significantly less likely to have a major-bleeding event within one year of treatment initiation compared to rivaroxaban (HR=0.69; 95% CI=0.59-0.81) and warfarin (HR=0.71; 95% CI=0.61-0.82) patients and trended towards numerically lower major bleeding compared to dabigatran patients (HR=0.87; 95% CI=0.72-1.06). Major bleeding-related medical costs were lower in apixaban patients ($53) compared to rivaroxaban ($111) and warfarin ($138) patients (p<0.001) and similar to dabigatran patients ($44, p=0.370). Furthermore, apixaban patients incurred lower all-cause medical costs ($1,646) compared to dabigatran ($1,974, p=0.02), rivaroxaban ($1,909, p=0.002) and warfarin ($2,162, p<0.001) patients.
Conclusion:
In a large national Medicare advantage population, treatment-naïve NVAF patients treated with apixaban were significantly less likely to have a major-bleeding event compared to those prescribed rivaroxaban or warfarin and had significantly lower medical costs.
How Does Risk Selection Respond to Risk Adjustment? New Evidence from the Medicare Advantage Program
