Effectiveness and safety of rivaroxaban versus warfarin in patients with unprovoked venous thromboembolism: A propensity-score weighted administrative claims cohort study

Background: In phase III trials, rivaroxaban demonstrated non-inferiority over enoxaparin/warfarin to prevent recurrent venous thromboembolism (VTE), with a reduction of major bleeding. However, compared to provoked VTE, the risk-benefit ratio of rivaroxaban may be different for patients with unprovoked VTE. Methods: In a retrospective claims data analysis using US MarketScan claims from 1/2012 to 12/2016, we included adults with a primary diagnosis of VTE newly-initiated on rivaroxaban or warfarin within 30-days of the incident VTE and with >= 12-months of continuous insurance benefits prior to the VTE (baseline). Patients with provoked VTE, a claim for anticoagulation during baseline or who redeemed prescriptions for >= 1 oral anticoagulant were excluded. Our primary outcomes were recurrent VTE and major bleeding at 6-months using an intention-to-treat (ITT) analysis. Three-month ITT and 12-month on-treatment (30-day permissible gap) analyses were also performed. Inverse probability-of-treatment weights based on propensity-scores and Cox regression were used to compare outcomes. Findings: We identified 10,489 rivaroxaban users and 26,364 warfarin users with incident unprovoked VTE. At 6-months, rivaroxaban was associated with a hazard ratio (HR) of 0.60 (95% confidence interval [CI] = 0.54-0.67) for recurrent VTE (number-needed-to-treat: 59; 95%CI 49-76) and a HR = 0.80 (95% CI = 0.66-0.98) for major bleeding versus warfarin. Our findings remained consistent in the 3- and 12-month analyses. Interpretation: Consistent with the results from the EINSTEIN phase-III trials, findings of our routine practice study suggest that, in patients with unprovoked VTE, rivaroxaban has the potential to reduce both the risk of major bleeding and recurrent VTE compared to warfarin.