Purpose of the study: Abatacept, a selective co-stimulation modulator, represents a potential new treatment for patients with active rheumatoid arthritis (RA) who have had an inadequate response to methotrexate (MTX) and tumor necrosis factor (TNF) inhibitors. The objective of the study is to perform a cost-effectiveness analysis of treatment strategies for patients with MTX-resistant RA satisfying the indication for TNF inhibitors and abatacept. Methods: A Markov model was developed to estimate the health effects and costs associated with three treatment strategies for patients with active RA: (1) treatment with adalimumab plus MTX, in case of non-response switch to abatacept (2) treatment with etanercept plus MTX, in case of non-response switch to abatacept, and (3) treatment with infliximab plus MTX, in case of non-response switch to abatacept. The model was simulated for 10,000 hypothetical patients (55-year old women) over patients' remaining lifetime and was evaluated using Monte Carlo simulation. Data such as efficacy rates, treatment withdrawal rates, and number of adverse events were obtained from clinical trials and published literature. Costs associated with joint replacement surgery were modeled in patients who did not respond to these treatments. The study was conducted from a payer perspective and only direct costs were included in the analysis. The main outcome measures were net gains in quality-adjusted life expectancy and incremental cost-effectiveness ratios, (ICERs) in dollars per quality adjusted life year (QALY) gained. Costs and effects were discounted at 3%. To test the robustness of the model extensive sensitivity analyses were conducted, including a probabilistic sensitivity analysis. Results: The treatment strategy of etanercept plus MTX and, in case of non-response switching to abatacept, was the most cost-effective treatment option from a payer perspective with an ICER of $15,391/QALY. The strategy of adalimumab plus MTX followed by abatacept was the next most cost-effective option. One-way and probabilistic sensitivity analyses indicated that the conclusions were relatively stable to variations in model assumptions. Conclusions: The ICER etanercept plus MTX followed by abatacept remained within the acceptable range of $50,000/QALY in the simulated population under a wide range of assumptions, as compared to the other comparators. The addition of abatacept to existing therapies is associated with lifetime cost-savings in patients who have had an inadequate response to MTX and TNF inhibitors.