The influence of three different dosage schedules of anthracycline (idarubicin or daunorubicin)-intensified preparative therapy prior to T-cell-depleted allogeneic bone marrow transplantation (BMT) on (1) the severity and duration of oral toxicity (mucositis), (2) the duration of bone marrow aplasia, and (3) overall survival, relapse, and disease-free survival was studied in 99 BMT patients with standard- or high-risk hematologic malignancies. A further 146 patients who did not receive the anthracycline-intensified conditioning served as (historic) controls. All patients received cyclophosphamide (total dose, 120 mg/kg) on days -6 and -5 and total body irradiation on days -2 and -1 prior to BMT. The 99 patients who received the anthracycline-intensified preparative regimen were given either idarubicin (total dose, 42 mg/m2; n = 88) or daunorubicin (total dose, 156 mg/m2; n = 11) by continuous intravenous infusion between days -7 and -1 prior to BMT in 59 cases (cohort 1), on days -8 and -7 in 17 cases (cohort 2), and on days -12 and -11 in 23 cases (cohort 3). The occurrence of severe oral mucositis and delayed bone marrow recovery was schedule dependent, being substantially lower with earlier administration of the anthracycline-intensified regimen on days -12 and -11 before BMT (cohort 3), in comparison with later administration (cohorts 1 and 2). Plasma drug and metabolite concentrations were measured in 11 patients who received idarubicin. At the time of allogeneic bone marrow infusion (day 0), patients in cohorts 1 and 2 had plasma concentrations of idarubicin and idarubicinol (its active metabolite) in the range of in vitro cytotoxicity. However, in cohort 3, plasma concentrations on day 0 were much lower, which correlated with the lower maximum intensity and shorter duration of mucositis in these patients. In terms of overall survival, relapse rate, and disease-free survival in standard-risk patients, the anthracycline-intensified regimen proved to be very effective. Transplant-related mortality was 25% in the anthracycline group compared with 32% in the controls. The probability of relapse also was significantly less in the anthracycline group in comparison with controls (17% v 46%; P < .001), and the probabilities of long-term overall and disease-free survival were significantly greater (71% v 37% [P < .01] and 63% v 32% [P < .01], respectively). Only three patients in the idarubicin group experienced cardiotoxicity (one in each cohort); the causative relationship with anthracyclines was considered likely in one, possible in one, and doubtful in the third.