Chemotherapy-induced anemia in children with cancer is usually of acute onset. To investigate an alternate treatment to transfusion (Tx), we undertook a phase I-II clinical trial of daily administrations of recombinant erythropoietin (rHuEPO). Patients with a hemoglobin (Hgb) value < 75 g/l were treated for 14 days in cohorts of 3 at escalating daily doses of 25, 50, 70, 80, 90, and 100 U/kg respectively. The maximum-tolerated dose was not encountered. Of 18 courses given to 15 children aged 0.5-18 years, 7 (39%) were associated with increased or stable Hgb levels (courses without Tx), while 11 (61%) were terminated by a Tx, without evidence of a dose-response relationship. Changes in mean Hgb levels and absolute reticulocyte counts were paralleled by those of mean white blood cell, platelet, and absolute neutrophil counts during the first 7 days and when the end-points of the study were reached. Numbers of circulating burst-forming units-erythroid remained low throughout courses without Tx. No cumulative increase of serially determined serum EPO levels was observed and serum ferritin levels were elevated in both groups of courses. We conclude that daily administration of rHuEPO were safe but ineffective in our trial. Recovery of chemotherapy-induced myelosuppression appeared to be the rate-limiting factor for the outcome, without evidence of an enhanced stimulation of erythropoiesis. The lack of a proliferative response of specific progenitor cells suggested a mechanism of transient primary resistance to rHuEPO.