In trial ALL-BFM 86, the largest multicenter trial of the Berlin-Frankfurt-Münster (BFM) study group for childhood acute lymphoblastic leukemia (ALL), treatment response was used as an overriding stratification factor for the first time. In the previous trial ALL-BFM 83, the in vivo response to initial prednisone treatment was evaluated prospectively. A blast cell count of > or = 1,000/microL peripheral blood after a 7-day exposure to prednisone and one intrathecal dose of methotrexate (MTX) identified 10% of the patients as having a significantly worse prognosis. In trial ALL-BFM 86 patients with > or = 1,000/microL blood blasts on day 8 were included in an experimental branch EG. Patients with < 1,000/microL blood blasts on day 8 were stratified by their leukemic cell burden into two branches, Standard Risk Group (SRG) and Risk Group (RG). SRG patients received an eight-drug induction followed by consolidation protocol M (6-mercaptopurine, high-dose [HD] MTX 4 x 5 g/m2) and maintenance. RG patients were treated with an additional eight-drug reinduction element. For EG patients protocol M was replaced by protocol E (prednisone, HD-MTX, HD-cytarabine, ifosfamide, mitoxantrone). All patients received intrathecal MTX therapy; only those of branches RG and EG received cranial irradiation. In branch RG, patients were randomized to receive or not to receive late intensification (prednisone, vindesine, teniposide, ifosfamide, HD-cytarabine) in the 13th month. During the trial reinduction therapy was introduced in branch SRG, because in the follow-up of trial ALL-BFM 83 the randomized low-risk patients receiving reinduction did significantly better. Nine hundred ninety-eight evaluable patients were enrolled, 28.6% in SRG, 61.1% in RG, 10.3% in EG. At a median follow-up of 5.0 (range 3.4 to 6.9) years, the estimated 6-year event-free survival was 72% +/- 2% for the study population, 58% +/- 5% in branch SRG for the first 110 patients without reinduction therapy, 87% +/- 3% for the next 175 patients with reinduction therapy, 75% +/- 2% in branch RG, and 48% +/- 5% in branch EG. Late intensification did not significantly affect treatment outcome of RG patients; however, only 23% of the eligible patients were randomized. Prednisone poor response remained a negative prognostic parameter despite intensified therapy. The results confirmed the benefit of intensive reinduction therapy even for low-risk patients. The strategy of induction, consolidation, and intensive reinduction may offer roughly 75% of unselected childhood ALL patients the chance for an event-free survival.