Which patients with unprovoked venous thromboembolism should receive extended anticoagulation with direct oral anticoagulants? A systematic review, network meta-analysis, and decision analysis

Introduction: Direct oral anticoagulants (DOACs) effectively prevent recurrent venous thromboembolism (VTE). However, it is unknown which agents should be used to prevent recurrent VTE and which patients with unprovoked VTE should receive extended anticoagulation. We therefore sought to compare the efficacy and safety among DOACs for secondary prevention of VTE. We also determined a risk-adapted threshold for initiating extended anticoagulation based on the likelihood of VTE recurrence (without treatment) and bleeding (with treatment) in patients with unprovoked VTE. Methods: Our systematic review of randomized controlled trials compares extended anticoagulation with DOACs to another DOAC, aspirin, or placebo for the prevention of recurrent VTE. We searched PubMed, EMBASE, and Cochrane Registry of Controlled Trials (CENTRAL) in October 2018. Our outcomes of interest were VTE recurrence, major bleeding, and all clinically relevant bleeding. We used network meta-analysis to make indirect comparisons among DOACs. We populated the threshold decision-analytic model with data from our meta-analysis to determine the risk of VTE recurrence above which the benefits of extended anticoagulation outweigh the harms compared with no treatment. Results: We included four, high-quality, randomized trials comprising 8386 participants. Low-dose apixaban, full-dose apixaban, low-dose rivaroxaban, full-dose rivaroxaban, and dabigatran reduce VTE recurrence compared with placebo (RR = 0.19, 95% CI, 0.12-0.31; RR = 0.20, 95% CI, 0.12-0.32; RR = 0.08, 95% CI, 0.03-0.27; RR = 0.14, 95% CI, 0.06-0.35; RR = 0.19, 95% CI, 0.09-0.40, respectively). No DOACs increased major bleeding risk compared with placebo. A VTE recurrence risk above 0.3% to 0.4% at approximately 1 year is the threshold to treat a patient with unprovoked VTE with extended anticoagulation (with any DOAC). Conclusions: All DOACs exhibit comparable efficacy for the prevention of recurrent VTE. Given that the risk of VTE recurrence is much higher than the calculated threshold for treatment, extended thromboprophylaxis should be considered in all patients with unprovoked VTE who do not have increased bleeding risk.