BACKGROUND: Rheumatoid arthritis patients who develop refractoriness are left with no alternatives other than leflunomide and costly biological response modifiers. Leflunomide, though effective, was associated with adverse events and has not been extensively studied in the Indian population. AIMS: Determination of safety and efficacy of leflunomide alone and if not useful, in combination with methotrexate in patients refractory to conventional disease-modifying agents. SETTING AND DESIGN: Open labeled clinical trial with leflunomide [100 mg, OD x 3 days followed by 20 mg, OD x 6 months], if no improvement at three months, combined with methotrexate [5-7.5 mg, OD x 3 months] at a tertiary care hospital. MATERIALS AND METHODS: The primary endpoint in the improvement in EULAR criteria and secondary endpoints were patient and physician global evaluation, incidence of remission and biochemical and clinical adverse events. STATISTICAL ANALYSIS: Chi square test or Fisher 's exact test and parametric and non-parametric repeat measure ANOVA were used for analysis. RESULTS: Among 84 patients who were included in the study, leflunomide showed improvement and remission in 52 [62%] and 6 [7%] in six months, by intention to treat analysis. Adverse events were observed in 15, discontinuation in 5 and 24 dropped out. With combination in 11 patients, there was improvement and remission in nine [91%] and one [9%] after three months. Adverse events were observed in six and one discontinued. CONCLUSIONS: If regular monitoring of hepatic function and hematological parameters are performed, leflunomide is an effective and safe drug in the Indian population in resistant rheumatoid arthritis patients, especially if used alone.