ABSTRACT
OBJECTIVES: Compare costs associated with all-cause healthcare resource use (HCRU), stroke/systemic thromboembolism (STE) and major bleedings (MB) between patients with non-valvular atrial fibrillation (NVAF) initiating apixaban or other oral anticoagulants (OACs). METHODS: We performed a retrospective cohort study using the French healthcare claims database, including NVAF patients between 2014/01/01 and 2016/12/31, followed until 2016/12/31. We used 4 sub-cohorts of OAC-naive patients, respectively initiating apixaban, dabigatran, rivaroxaban or VKAs. We matched patients initiating apixaban with patients initiating each other OACs using 1:n propensity score matching. All-cause HCRU and event-related costs by OAC treatment were estimated and compared between matched patients using generalised-linear models with gamma-distribution and two-part models. RESULTS: There were 175,766 patients in the apixaban-VKA, 181,809 in the apixaban-rivaroxaban, and 42,490 in the apixaban-dabigatran matched cohorts. Patients initiating apixaban had significantly lower HCRU costs than patients initiating VKA (1,105 vs. 1,578, p < 0.0001), dabigatran (993 vs. 1,140, p < 0.0001) and rivaroxaban (1,013 vs. 1,088 p < 0.0001). They have had significantly lower costs related to stroke/STE and MB than patients initiating VKA (respectively, 183 vs. 449 and 147 vs. 413; p < 0.0001), rivaroxaban (respectively, 145 vs. 197 and 129 vs. 193; p < 0.0001), and lower costs related to stroke/STE than patients initiating dabigatran (135 vs. 192, p < 0.02). Costs related to MB were not significantly different in patients initiating apixaban and those initiating dabigatran (119 vs. 149, p = 0.07). CONCLUSIONS: HCRU and most event-related costs were lower in patients initiating apixaban compared to other OACs. Apixaban may be cost-saving compared to VKAs, and significantly cheaper than other DOACs, although cost differences are limited.