Effectiveness of therapy with individual disease-modifying antirheumatic drugs (DMARDs) in rheumatoid arthritis (RA) is limited, and the number of available DMARDs is finite. Therefore, at some stage during the lengthy course of RA, institution of traditional DMARDs that have previously been applied may have to be reconsidered. In the present study we investigated the effectiveness of re-employed methotrexate in patients with a history of previous methotrexate failure (original course). A total of 1,490 RA patients (80% female, 59% rheumatoid factor positive) were followed from their first presentation, yielding a total of 6,470 patient-years of observation. We identified patients in whom methotrexate was re-employed after at least one intermittent course of a different DMARD. We compared reasons for discontinuation, improvement in acute phase reactants, and cumulative retention rates of methotrexate therapy between the original course of methotrexate and its re-employment. Similar analyses were peformed for other DMARDs. Methotrexate was re-employed in 86 patients. Compared with the original courses, re-employment was associated with a reduced risk for treatment termination because of ineffectiveness (P = 0.02, by McNemar test), especially if the maximum methotrexate dose of the original course had been low (<12.5 mg/week; P = 0.02, by logistic regression). In a Cox regression model, re-employed MTX was associated with a significantly reduced hazard of treatment termination compared with the original course of methotrexate, adjusting for dose and year of employment (hazard ratio 0.64, 95% confidence interval 0.42–0.97; P = 0.04). These findings were not recapitulated in analyses of re-employment of other DMARDs. Re-employment of MTX despite prior inefficacy, but not re-employment of other DMARDs, is an effective therapeutic option, especially in those patients in whom the methotrexate dose of the original course was low.