69 patients (median age 53 years, 19-79 years old) with untreated acute non-lymphoblastic leukemia (ANLL) were randomized to receive either a regimen of amsacrine, cytarabine, thioguanine (AAT) or daunorubicin, cytarabine, thioguanine (DAT). AAT consisted of amsacrine 200 mg/m2/day x 5, thioguanine 100 mg/m2/12 h p.o. x 10; DAT was daunorubicin 50 mg/m2/day x 3, cytarabine 200 mg/m2/day x 5, thioguanine 100 mg/m2/12 h p.o. x 10. After one or two induction courses the patients subsequently received 2 consolidation courses. 17 patients were not assessable for response to therapy due to exitus during induction treatment. Complete remission could be obtained in 14/24 (58%) of DAT patients respectively. Patients less than 60 years of age achieved CR in 63% (AAT) vs 65% (DAT), whereas patients greater than or equal to 60 years obtained a CR in 50% (AAT) vs 13% (DAT). Toxicity appears not to be increased significantly with amsacrine. These data indicate that amsacrine could replace daunorubicin in remission induction regimens of ANLL containing cytosin arabinoside and thioguanine without decreasing the response rate.