Oral idarubicin (IDA) is an active drug in metastatic breast cancer, but its role in the management of this tumor is yet not established completely. To investigate a new modality of IDA administration, a dose-finding study was designed with hyperfractionated doses. The purpose was to determine the maximum tolerated dose (MTD), the dose-limiting toxicity (DLT), and the pharmacokinetics of this schedule. IDA was administered twice daily as outpatient therapy in cycles of 3 weeks followed by a 1-week rest. Thirty-one patients with progressive metastatic breast cancer and pretreated with chemotherapy (including epirubicin and doxorubicin) were enrolled. DLT was defined as G4 hematological toxicity or any other toxicity G3 or higher (Bloom and Richardson grading). Inter- and intrapatient dose increases were studied. Pharmacokinetics of IDA and its metabolite idarubicinol (IDOL) were evaluated. IDA dose was increased from 2 mg/day to 10 mg/day, by steps of 1 mg/day, with the larger dose given in the evening. MTD was reached at 10 mg/day. Overall, the therapy cycles were 69 (median/patient, 2; range, 1-6). DLTs were G4 neutropenia associated with leukopenia and thrombocytopenia in one patient and G3 diarrhea in another of the 5 patients in the 10 mg/day cohort. The two patients developing DLT at the daily dose of 10 mg received a dose normalized for body surface of 6.85 and 5.65 mg/m2/day, respectively. We considered 5.5 mg/m2/day to be the MTD. Other toxicities were nausea, vomiting, neutropenia, and diarrhea, grades G1 to G2. By univariate analysis, significant correlations were observed between absolute neutrophil count at nadir and IDA area under the curve (P = 0.022; r = -0.33), IDA Cmax (P = 0.0067; r = -0.38), IDOL area under the curve (P = 0.0009; r = -0.43), and IDOL Cmax (P = 0.0016; r = -0.41), respectively. By multivariate analysis, IDA Cmax was the strongest determinant for neutropenia (R2 = 0.14; P = 0.01). Among the 21 patients evaluable for response, 3 (14.3%) had partial response (lasting 3, 6, and 8 months, respectively), and 6 (28.6%) had a complete arrest of disease progression (lasting 2-6 months). In conclusion, the MTD of this schedule is 10 mg/day and the DLTs are neutropenia and diarrhea. Tolerance was good, and the treatment is feasible as home therapy. Some objective measurable responses were documented in this group of anthracycline-pretreated patients. IDOL could have a role for the pharmacological effect. Further evaluation of this schedule is warranted to assess the activity and toxicity of prolonged oral IDA administration.