In the ongoing trial ALL-BFM 90 for the treatment of childhood non-B cell acute lymphoblastic leukemia (ALL) 1468 unselected patients (pts) were enrolled from 84 centers in Germany and Switzerland from 4/90 to 12/93. Based on the results of the previous trial ALL/NHL-BFM 86 this treatment program focused especially on therapy modifications for average (MRG) and high risk (HRG) pts, on the evaluation of therapy response for prognosis, and on the identification of high risk pts by molecular genetics. For average risk pts consolidation therapy was intensified by the addition of L-asparaginase (L-ASP) on a randomized basis. In HRG induction and consolidation therapy was modified by introduction of early intensification elements that had proved to be effective in relapsed pts. This patient group was randomized for the evaluation of the effects of G-CSF administered in the intervals between the intensification elements. Distribution of the 1376 eligible pts into the three treatment arms SRG (standard risk), MRG, and HRG was as expected (17 pts not yet assigned): 385 pts (28.0%), 834 pts (60.6%), and 140 pts (10.2%), respectively. Treatment consisted of the 8-drug induction (Protocol I), consolidation (Protocol M), reinduction (Protocol II), and maintenance therapy (total therapy duration 24 months). The drug doses and combinations were only slightly modified compared to the previous study ALL-BFM 86 with the exception of the randomized L-ASP containing arm MRG-2 (Protocol M-A) and group HRG. Preventive cranial irradiation was reduced to 12 Gy and applied to MRG and HRG pts only. As in study ALL-BFM 86, the initial response to a 7-day exposure to prednisone and to the first intrathecal injection of MTX at diagnosis was evaluated at day 8 of treatment with regard to blast count in peripheral blood (PB). In addition, pts were now investigated for the presence of blasts in the bone marrow (BM) at day 15 of treatment to compare the prognostic power of both response parameters. Identification of translocation t(9; 22) and/or BCR-ABL rearrangement characterized a small subgroup of pts that were not detected by poor initial therapy response. These pts were enrolled in HRG for more intensive treatment including allogeneic bone marrow transplantation (BMT). After a median observation time of 22 months, the overall probability for event-free survival (p-EFS) is 82 +/- 2%. 11 pts (0.8%) died before complete remission (CR) was achieved, 15 pts (1.1%) died while in CR for reasons other than relapse.(ABSTRACT TRUNCATED AT 400 WORDS)