From January 1970 through December 1991, 94 girls and 130 boys with a median age of 4.8 years were diagnosed with non-B cell acute lymphoblastic leukaemia (ALL) at the University Hospital, Rigshospitalet. Intensive risk-group adapted therapy based on age and white-cell count (WBC) at diagnosis, the presence of a mediastinal mass, central nervous system (CNS) or testicular leukaemia, T-cell disease, and certain cytogenetic translocations have been used since July 1981. Ninety-seven percent of all patients achieved complete remission (all patients diagnosed since July 1986). Ninety patients relapsed, all within five years from diagnosis: 59 in bone-marrow (BM), 20 in CNS (no BM-involvement), nine in testes (all isolated), one in the eye, and one in a lymph node. The five-year event-free survival increased from 0.20 to 0.72 from the first to the last five-year period (p < 0.0001). Age and WBC at diagnosis were of prognostic significance during the period 1970-1981, but not during the last 10-year period, when risk-adapted therapy was applied. Following a relapse, patients with an isolated testicular focus had the best five-year survival as compared to patients with CNS- and/or BM-involvement (1.00 vs 0.19; p = 0.003). Patients relapsing following cessation of therapy had a better prognosis than did patients relapsing on therapy (five-year survival 0.58 vs 0.17; p = 0.002). Identification of new risk factors, more individualized therapy and monitoring of minimal residual disease is expected to have increasing influence on the management of children with ALL.