Hydroxyurea inhibits ribonucleotide reductase, resulting in depletion of intracellular deoxynucleotide pools and inhibition of DNA repair. It has been used in a variety of malignancies and is usually given orally. Deoxynucleotide depletion is directly related to the concentration of and duration of exposure to hydroxyurea; thus, prolonged continuous infusion may result in increased therapeutic efficacy. A total of 30 patients were treated on this trial, designed to determine the maximum tolerated doses (MTD) of intravenous hydroxyurea given as a 24- or 48-h continuous infusion. The MTD for the 24-h infusion was 13,520 mg/m2 following a bolus of 1,690 mg/m2, and the mean (+/- SD) plasma steady-state concentration was 1.93 +/- 0.52 mM. For the 48-h infusion, the MTD was 17,576 mg/m2 following a bolus of 2,197 mg/m2 and the mean steady-state level was 1.43 +/- 0.31 mM. The dose-limiting toxicity on both schedules was marrow suppression manifesting as neutropenia and thrombocytopenia. Pharmacokinetic analysis revealed decreasing clearance with increasing dose, implying that drug elimination is saturable. Pharmacodynamic analysis showed a slight correlation between steady-state plasma levels and the degree of marrow suppression.