In January 1987 we started a multicenter study in order to evaluate in adult ALL patients the results of an intensive chemotherapy effected early after CR, and to compare the efficacy of allogeneic BMT vs autologous BMT vs prolonged intensive chemotherapy in the attempt to eradicate minimal residual leukemia. To September 1990 ninety-six patients entered this study; of the 87 evaluable for induction 25 were at low risk and 62 at high risk; 67 (77%) achieved CR by an induction chemotherapy including vincristine, adriamycin, cyclophosphamide, dexamethasone. Fifty-six out of 67 remitters were enrolled for the early intensification, which consisted of HDAra-C+amsacrine (or IDAra-C+mitoxantrone) followed by vincristine+adriamycin+cyclophosphamide and etoposide+Ara-C. During the early intensification an unexpectedly high number of relapses (10/56) was observed, showing that very intensive treatment with myelosuppressive agents is not useful at this point of the post-remission therapy. One patient suffered toxic death. Out of 45 patients who completed the early intensification 16 had a related well-matched donor and were selected for allogeneic BMT (performed in 11); of the remaining 29 patients, 14 were randomized for autologous BMT (performed in 9) and 15 for a second intensification. The overall DFS at 3 years is 35%. The high number of early relapses makes it difficult to draw conclusions from the comparison of the three eradication modalities. The best results, although without statistical significance, were obtained after allogeneic BMT; in high-risk patients this procedure should be effected as soon as possible after attainment of CR. Autologous BMT and prolonged intensive chemotherapy gave results similar to each other; both were sometimes followed by delayed relapses.