This study investigated the cost-effectiveness of leflunomide (LEF) compared to methotrexate (MTX) and sulfasalazine (SSZ) in the United Kingdom. A Markov model was constructed using health states defined by Health Assessment Questionnaire score. The model is based on a cohort of patients with recently diagnosed definite RA who were followed for up to 15 years at nine rheumatology clinics in the UK. The treatment effect was calculated based on clinical trials comparing LEF to MTX (one international and one United States trial) and to SSZ (one international trial). Transitions between health states for the first 2 years of treatment were calculated from the clinical trials, while the extrapolation beyond the trial was based on the Early Rheumatoid Arthritis Study cohort, using an ordered probit model. This makes it possible to predict transitions for patients with similar characteristics (age, time since disease onset) as in the trials. Separate analyses were performed for each trial, and all analyses covered a 10-year timeframe. Using the US trial, LEF had slightly lower costs and better effects (44,017 and 4.307 pounds QALYs, compared to 44,988 and 4.158 pounds QALYs for MTX), while for the international trial this was reversed (34,070 and 4.487 pounds QALYs for MTX compared to 36,351 and 4.372 pounds QALYs for LEF). Compared to SSZ, the cost of using LEF was slightly lower, with an increase in QALYs (35,855 and 3.896 pounds QALYs compared to 36,731 and 3.721 pounds QALYs for SSZ). The two trials comparing LEF to MTX gave differing results. One possible reason for this is that MTX patients in the US trial were given folic acid, whereas in the international trial folate supplementation was not mandated. This may have reduced the effectiveness of MTX. In the UK it is standard practice to use folic acid with MTX, and therefore the results from the US trial may be more relevant for the UK. Compared to SSZ, the use of LEF appears to be cost-effective in the UK.