Background: Several studies have compared the cost effectiveness of non-vitamin K antagonist oral anticoagulants (NOACs) and warfarin using results from clinical trials evaluating NOACs. However, the time in therapeutic range (TTR) of warfarin groups ranged across clinical trials, and all were below the therapeutic goal of 70%. We compared the cost effectiveness of edoxaban 60 mg, apixaban 5 mg, dabigatran 150 mg, dabigatran 110 mg, rivaroxaban 20 mg, and well-managed warfarin with a TTR of 70% in preventing stroke among patients with atrial fibrillation at high risk of bleeding.
Methods: For the six treatments, we used a Markov state-transition model to quantify lifetime costs in $US and effectiveness in quality-adjusted life-years (QALYs). We simulated relative risk ratios of clinical events with each NOAC versus warfarin with a TTR of 70% using published regression models that predict how the incidence of thrombotic or hemorrhagic events changes for each unit change in TTR. We re-ran our analysis for two other estimates of TTR: 65 and 75%.
Results: Treatment with edoxaban 60 mg cost $US127,520/QALY gained compared with warfarin with a TTR of 70% and cost $US41,860/QALY gained compared with warfarin with a TTR of 65%. However, warfarin with a TTR of 75% was more effective and less expensive than all NOACs. For three levels of TTR, apixaban 5 mg, dabigatran 150 mg, dabigatran 110 mg, and rivaroxaban 20 mg were dominated strategies.
Conclusions: The comparative cost effectiveness of edoxaban and warfarin is highly sensitive to TTR. At the $US100,000/QALY willingness-to-pay threshold, our results suggest that warfarin is the most cost-effective treatment for patients who can achieve a TTR of 70%.