OBJECTIVES: Myasthenia gravis (MG) is a rare autoimmune disorder characterized by exacerbations and remissions. The costs of managing MG are higher than those of many other chronic neurological diseases and there are no studies evaluating the cost-effectiveness of treatment alternatives for refractory MG in patients who failed to respond to at least two successive immunosuppressive drugs. This study aimed to assess the incremental costs and benefits of rituximab, cyclosporine, and tacrolimus in the treatment of refractory MG. METHODS: We used a Markov model from the perspective of a US private insurance payer to evaluate the cost-effectiveness of rituximab, tacrolimus, and cyclosporine in a hypothetical cohort of 1,000 patients with refractory MG, aged 20 years and above. We obtained disease transition probabilities, costs and outcomes data from the published literature. We calculated the incremental cost-effectiveness ratios (ICERs) as cost per quality-adjusted life-year (QALY) gained and cost per myasthenic crisis averted after the first two years of treatment and over a patient’s lifetime. RESULTS: In the first two years after treatment rituximab is not cost-effective compared with cyclosporine, given an ICER of $368,823 per QALY gained. However, over a patient’s lifetime rituximab has an ICER of $41,947 per QALY gained making it more cost-effective than cyclosporine in the context of the commonly accepted US threshold of $50,000 per QALY gained. Tacrolimus is more costly and less effective than cyclosporine and rituximab both after two years of treatment and over a patient’s lifetime. CONCLUSIONS: Assuming the benefits of treatment persist over time, rituximab is more cost-effective than cyclosporine over a patient’s lifetime but not after two years of treatment under both the standard US threshold of $50,000 per QALY gained and an alternative higher threshold of $100,000 per QALY gained. Additional research is needed to evaluate the long-term benefits of rituximab.