In previously untreated adult patients with acute non-lymphocytic leukemia (ANLL), idarubicin (IDA) and daunorubicin (DNR) were compared for efficacy and safety when used in combination with cytarabine (Ara-C). IDA 12 mg/m2/d and DNR 40 mg/m2/d were administered by intravenous (i.v.) bolus for 3 consecutive days (days 1 to 3), respectively, in combination with Ara-C 80 mg/m2 given by 2-hour i.v. infusion, every 12 hours for 7 consecutive days. The number of assessable patients was 32 for each group. The rate of complete remission (CR) was 59.4% (19/32) in the IDA group and 40.6% (13/32) in the DNR group. The clinical equivalence test with delta = 10% demonstrated that the remission rate in the IDA group was equal or superior (P = .010) to the DNR group. In addition, the Cochran-Mantel-Haenszel test for response means with scores of 3 (CR), 2 (partial response [PR], and 1 (no response [NR]) showed the IDA group to be significantly superior (P = .044) to the DNR group. The duration needed to attain less than 5% leukemic cells in the bone marrow tended to be shorter in the IDA group (P = .072), and in CR patients, the number of days needed to reach the nadir value for leukemic cells was significantly fewer in the IDA group (P = .037). The nadir value for WBC count was significantly lower in the IDA group (P = .022). As for adverse reactions, high incidences of diarrhea and stomatitis were observed in the IDA group, while the incidences of other adverse reactions were similar between the two groups. When the effects of the drug on the ECG were examined, significant changes in ECG parameters were observed after treatment in the DNR group but not in the IDA group. Based on these findings, it was surmised that the combination of IDA plus Ara-C is the treatment of first choice for adult ANLL patients.