In view of current uncertainty regarding the optimum route for iron supplementation in patients receiving recombinant human erythropoietin (EPO), a prospective randomized controlled study was designed to investigate this issue. All iron-replete renal failure patients commencing EPO who had a hemoglobin concentration < 8.5 g/dl and an initial serum ferritin level of 100 to 800 micrograms/liter were randomized into three groups with different iron supplementation: Group 1, i.v. iron dextran 5 ml every 2 weeks; Group 2, oral ferrous sulphate 200 mg tds; Group 3, no iron. All patients were treated with 25 U/kg of EPO thrice weekly subcutaneously. The hemoglobin concentration, reticulocyte count, serum ferritin, transferrin saturation, and EPO dose were monitored every two weeks for the first four months. Thirty-seven patients entered the study (12 i.v., 13 oral, 12 no iron). The three groups were equivalent with regard to age, sex, and other demographic details. Even allowing for dosage adjustments, the hemoglobin response in the group receiving i.v. iron (7.3 +/- 0.8 to 11.9 +/- 1.2 g/dl) was significantly greater than that for the other two groups (7.2 +/- 1.1 to 10.2 +/- 1.4 g/dl and 7.3 +/- 0.8 to 9.9 +/- 1.6 g/dl for Groups 2 and 3, respectively; P < 0.005 for both groups vs. Group 1 at 16 weeks). There was no difference between the groups supplemented with oral iron and no iron. Serum ferritin levels remained constant in those receiving i.v. iron (345 +/- 273 to 359 +/- 140 micrograms/liter), in contrast to the other two groups in which ferritin levels fell significantly (309 +/- 218 to 116 +/- 87 micrograms/liter and 458 +/- 206 to 131 +/- 121 micrograms/liter for Groups 2 and 3, respectively; P < 0.0005 for Group 1 vs. Group 2, and P < 0.005 for Group 1 vs. Group 3 at 16 weeks). Dosage requirements of EPO were less in Group 1 (1202 +/- 229 U/kg/16 weeks) than in Group 2 (1294 +/- 314 U/kg/16 weeks) or Group 3 (1475 +/- 311 U/kg/16 weeks; P < 0.05 vs. Group 1). The results of this study suggest that, even in iron-replete patients, those supplemented with i.v. iron have an enhanced hemoglobin response to EPO with better maintenance of iron stores and lower dosage requirements of EPO, compared with those patients receiving oral iron and no iron supplementation.