Effectiveness and Safety of the 15 Mg Dose of Rivaroxaban Compared with Vitamin k Antagonists in Patients with Atrial Fibrillation: Results from a Cohort Study in the Nationwide French Claims and Hospitalization Database (SNIIRAM)

Background: Large randomized trials have established that non-vitamin K antagonist oral anticoagulants (NOACs) have a similar efficacy to warfarin but generally have a superior safety profile for stroke prevention in atrial fibrillation (SPAF). However, evidence on reduced dose regimens of NOACs from "real world" clinical practice is scarce. A recent Danish analysis found mixed results with higher mortality in patients treated with reduced doses of some NOACs. In response to this lack of firm evidence about reduced dose regimens for SPAF, a nationwide cohort study was conducted to examine effectiveness and safety for new users of reduced doses of rivaroxaban (R) vs vitamin K antagonists (VKA) for SPAF, rivaroxaban 15 mg being indicated for SPAF in patients with moderately impaired renal function.

Methods: Cohorts of new users of R 15 mg once a day or VKA for SPAF in 2013 or 2014 were identified in the 66 million persons nationwide French claims and hospitalization database (SNIIRAM) and followed-up for at least one year and up to 2 years. AF was defined from full coverage, hospitalization or procedure for atrial fibrillation without valvular disease using 3-year database history. For each comparison, patients were matched 1:1 on gender, age, date of the first anticoagulant dispensing, and high-dimensional propensity score, including major arterial thrombosis and bleeding risk factors; Hazard ratios (HR) [95% confidence interval] were determined over follow-up during first prescribed anticoagulant exposure, using the Cox proportional hazard risk model or the Fine and Gray model.

Results: Of 734,599 incident anticoagulant users in 2013 or 2014, 133,181 were NVAF patients treated with VKA or 15 mg R (108,641 VKA and 24,540 R). Mean age was 78.7 years, 51.0% were male, 81.6% had CHA2DS2VASc score ≥ 2 and 15.2% HAS-BLED score > 3 with significant differences between anticoagulants. For R vs. VKA, 23,314 patients were matched per arm; overall incidence during exposure was 2.5 % person-years (%PY) for stroke and systemic embolism (SSE), 3.0%PY for major bleeding (MB), 5.4%PY for clinically relevant bleeding (CRB), 1.9% for acute coronary syndromes (ACS), 11.1% for death and 15.3% for composite of stroke and SE, major bleeding and death. HR were 1.05 [0.92-1.21] for SSE, 0.84 [0.74-0.96] for MB, 0.89 [0.81-0.98] for CRB, 0.85 [0.73-1.00] for ACS, 0.85 [0.79-0.90] for death and 0.89 [0.84-0.94] for composite.

Conclusions:

In this large propensity score matched nationwide comparative cohort study, rivaroxaban 15 mg was associated with similar rates of stroke/systemic embolism but lower rates of bleeding and death compared with vitamin K antagonists. A recently reported higher mortality with reduced doses of some NOACs was not confirmed with rivaroxaban in this larger study.