OBJECTIVES: To conduct a systematic review of cost-effectiveness studies of newer oral anticoagulants for stroke prevention in atrial fibrillation versus warfarin and understand its implications for policy and for further research. METHODS: We searched DARE, Cochrane, NICE, Tufts, NHS EED, Science Direct and PubMed through May 2013 to identify studies of oral anticoagulants dabigatran, rivaroxaban and apixaban versus warfarin for stroke prevention in non-valvular atrial fibrillation in patients at a moderate-to-high risk of stroke initiating anticoagulation for/near lifetime using either a societal or healthcare perspective. A separate analysis was performed for ISPOR abstracts. RESULTS: 10 studies were identified, most based on one randomized trial per therapy, 3 reported an ICER above commonly accepted WTP levels. ICERs ranged from $9,099–$86,000 for dabigatran and $3,190–$55,757 for rivaroxaban. Apixaban was found to be cost-effective $11,400–$24,312. Upon PSA, dabigatran was cost-effective 40% to 98%, apixaban 44.1% to 60%, and rivaroxaban 2.1% to 80% of the time, for the lowest reported WTP. Variations in ICERs occurred between studies that used the same efficacy data. Key variables influencing variations include differences in costs; assumptions for INR monitoring and utilization of healthcare resources; and the probabilities of adverse event. Effectiveness, compliance, adverse events and INR management were derived from clinical trials. Additional data from recent abstracts reported similar trends. CONCLUSIONS: For this indication, novel agents seem to be cost-effective alternatives to warfarin, despite variations across countries. However, cost-effectiveness may depend on: prices of medicines in each country; the proportion of actual patients’ time within the INR therapeutic range; and the actual real-world effectiveness, safety, and utilization of these therapies. More head-to-head RCTs and pragmatic trials are required. Future cost-effectiveness studies of these new therapies should be periodically repeated during the lifecycle of medicines under real-world utilization and should also incorporate budget impact analysis.