The plasma pharmacokinetics of a second generation anthracycline derivative, idarubicin (Ida), have been studied in 17 patients with acute myelocytic leukemia (AML) and high risk features. The drug (10 mg/m2) was given in a randomized cross-over design as 3 min and 2 h infusions for three consecutive days. Cytosine arabinoside (Ara-C, 1 g/m2) was given on days 1-4. The plasma concentration time course of Ida was most properly described by the three-compartment pharmacokinetic model, independent of administration time. The maximum plasma concentration (Cmax) of Ida was reduced by a factor of 3 by increasing the infusion time from 3 min to 2 h. The pharmacokinetic pattern of the active metabolite idarubicinol (IdaOH) was only to a minor extent affected by the longer infusion time. The time course of IdaOH following each dose of Ida was accurately described by the one-compartment model with a first-order formation phase. The are under the plasma concentration versus time curves (AUC) of Ida and IdaOH were not affected by the administration time. Following Ida in combination with Ara-C, the medial duration of leukopenia (< 1.0 x 10(9)/l) was 14 days (range 5-56) and of thrombocytopenia (< 50 x 10(9)/l) was 22 days (range 7-120). The large majority of patients developed infectious complications. Two patients with MDS-AML showed a good response. The results of the present study give no evidence of reduced hematologic toxicity by increasing the administration time of Ida from 3 min to 2 h. However, minimizing Cmax, by administration of Ida as prolonged infusion during a 3 day course, might be clinically important in order to reduce cardiotoxicity and hopefully to increase anti-tumor efficacy through an increased accumulation of Ida and IdaOH in leukemic cells.