Adult patients with acute lymphoblastic leukemia (ALL) were treated according to the ALL90 study, the second prospective study for ALL of the Japan Adult Leukemia Study Group (JALSG). Its characteristics included response-oriented individualized induction therapy with six drugs (doxorubicin, mitoxantrone, vincristine, prednisolone, [corrected] cyclophosphamide and L-asparaginase), and a prospective comparison between allogeneic bone marrow transplantation (allo-BMT) and chemotherapy alone in patients below 45 years of age. The protocol consisted of one or two courses of induction, four courses of consolidation, and three courses of intensification including 12 month maintenance and six times of central nervous system (CNS) prophylaxis. Of 180 evaluable patients (median age, 43), 125 (69%) achieved complete remission (CR). Predicted overall survival (OAS), event-free survival and disease-free survival (DFS) were 15, 10 and 14%, respectively at the median follow-up period of 62 months. No specific toxicities were observed. Leukocytes < 30,000/microliter, normal karyotype, and blasts < 10% in bone marrow at day 15 of induction therapy were significantly favorable prognostic factors for the achievement of CR, DFS and OAS by univariate analysis. Multivariate analysis showed leukocytes < 30,000/microliter and blasts < 10% on day 15 was a significant factor for the achievement of CR, DFS and OAS. Ph-chromosome was found in 28% (36/130) of patients examined and was one of the worst prognostic factors. All Ph positive patients were predicted to die within 600 days. Allo-BMT was not significantly superior to chemotherapy with respect to DFS (P = 0.226). The overall results were inferior to those of the former ALL87 protocol. As reasons, the older median age of 43 years old (vs. 38 years old) and lower dose intensity, especially of l-asparaginase, etc. were suggested. However, patients with good prognostic factors (leukocyte < 30,000/microliter and age < 30 years old) showed better survival than others (P < 0.0001), and the result was similar to that of older children, the high risk group of childhood ALL, suggesting that ALL could be a disease of single entity, showing higher resistance to chemotherapy as patients become older.