High-dose methotrexate (MTX) has been extensively used for treatment of acute lymphoblastic leukemia (ALL). To determine the optimal dose of MTX in childhood relapsed ALL, the ALL Relapse Berlin-Frankfurt-Münster (ALL-REZ BFM) Study Group performed this prospective randomized study. A total of 269 children with a first early/late isolated (n = 156) or combined (n = 68) bone marrow or any isolated extramedullary relapse (n = 45) of precursor B-cell (PBC) ALL (excluding very early marrow relapse within 18 months after initial diagnosis) were registered at the ALL-REZ BFM90 trial and randomized to receive methotrexate infusions at either 1 g/m(2) over 36 hours (intermediate dose, ID) or 5 g/m(2) over 24 hours (high dose, HD) during 6 (or 4) intensive polychemotherapy courses. Intensive induction/consolidation therapy was followed by cranial irradiation, and by conventional-dose maintenance therapy. Fifty-five children received stem-cell transplants. At a median follow-up of 14.1 years, the 10-year event-free survival probability was .36 (+/- .04) for the ID group (n = 141), and .38 (+/- .04) for the HD group (n = 128, P = .919). The 2 groups did not differ in terms of prognostic factors and other therapeutic parameters. In conclusion, methotrexate infusions at 5 g/m(2) per 24 hours, compared with 1 g/m(2) per 36 hours, are not associated with increased disease control in relapsed childhood PBC acute lymphoblastic leukemia.