The optimal methotrexate dose differs between rheumatoid arthritis (RA) patients, and dose-escalation strategies also differ among rheumatologists. By taking advantage of the heterogeneous methotrexate dosing that occurs among Japanese rheumatologists, we analyzed the efficacy and safety of different methotrexate doses. A large observational cohort of RA patients, IORRA, was established in 2000. A dataset from April 2003 that included 4578 RA patients was used for a cross-sectional analysis, while a dataset of 1649 patients who received methotrexate from October 2000 to October 2005 was used for a longitudinal analysis. The cross-sectional analysis included 12 rheumatologists who prescribed methotrexate to more than 60 patients. Mean methotrexate dose ranged widely (4.8-9.0 mg/week) among rheumatologists with a significant positive relationship between average methotrexate dose and the percentage of patients with Disease Activity in 28 Joints (DAS28) scores below 3.2. During the longitudinal analysis, both methotrexate prescription frequency and the average dose prescribed by 16 rheumatologists increased. Overall disease activity as assessed by DAS28-area under the curve (AUC) and disability progression as assessed by Japanese version of the Health Assessment Questionnaire (JHAQ)-slope inversely correlated with the extent of methotrexate use. This study demonstrated that extensive methotrexate use effectively suppressed RA disease activity and inhibits disability progression. In addition, we have found that it is critical to pay attention to patient-reported adverse reactions.