OBJECTIVES: To perform a cost-effectiveness analysis of rivaroxaban (once-daily) in the prevention of stroke and systemic embolism of patients with non-valvular atrial fibrillation (NVAF) and in patients sub-groups from the perspective of the Italian healthcare system (SSN). METHODS: A Markov model was developed with a lifetime timeframe where a hypothetic NVAF patients’ cohort is treated with Vitamin-K antagonists (VKAs), antiplatelet drugs (ASA) or no therapy. Patients remain stable or progress towards other health states (ischemic or hemorrhagic stroke, myocardial infarction and bleedings) until death. The base case compares rivaroxaban with VKAs. In subgroup analyses, rivaroxaban is compared with patients at highest unmet medical need: 1. VKA patients with poor INR control, 2. patients under ASA or 3. not treated. Clinical data were derived from ROCKET-AF trial or a network meta-analysis. Utility data were retrieved from published literature. Healthcare resources consumption was valued using average regional tariffs in Italy. Since rivaroxaban price is not officially published, the price of the first novel oral anticoagulant approved in this indication in Italy was considered. Model outcomes are expressed in terms of incremental cost per quality adjusted life year (QALY) gained (ICER). Univariate and probabilistic sensitivity analyses were performed RESULTS: In the base case, rivaroxaban showed to be cost-effective compared to VKA with an ICER of €11,000/QALY which is below the threshold deemed acceptable from Italian payers (25.000€-40.000€). In the subgroups analyses, rivaroxaban demonstrated to be a dominant strategy (more effective and less costly). Sensitivity analyses showed the robustness of the results. CONCLUSIONS: Rivaroxaban is a cost-effective alternative to VKA and is cost-saving for the SSN in the treatment of NVAF patients at highest unmet medical need.