The evolution of body iron stores was prospectively analyzed during a stable erythropoiesis period in 27 subjects (14 males and 13 females) on hemodialysis for more than 2 years in order to clarify the iron requirements of these patients and the effectiveness and safety of the administration of sodium ferric gluconate as a method to maintain adequate body iron stores. All patients had a stable hemoglobin level (variation < 1 g/dl). Sixteen subjects were on maintenance recombinant human erythropoietin therapy at regular doses. All patients received intravenous sodium ferric gluconate for 6 months (62.5 mg/month). The iron requirements were estimated as the difference between the amount of iron administered and the variation of body iron stores (calculated by the empirical formula of Cook and coworkers). The hemoglobin remained stable (basal 10.7 +/- 1.1, at 6th month 10.6 +/- 1 g/dl). Considering all cases, there were no significant variations in body iron stores (basal 457 +/- 273, at 6th month 451 +/- 316 mg). The patients were classified into three groups according to whether their body iron stores decreased (group A, n = 8), remained stable (group B, n = 11), or increased (group C, n = 8). There were no differences among groups concerning sex, age, time on hemodialysis, or erythropoietin therapy. However, there were statistically significant differences concerning their basal body iron stores (group A 457 +/- 228 mg. group B 563 +/- 146, and group C 230 +/- 297 mg; p < 0.05, analysis of variance). The iron needs, considering the total group, were 2.12 +/- 2 mg/day. There were no differences in iron requirements according to sex, but menstruating women had higher iron needs than the nonmenstruating ones (4.29 +/- 2 vs. 2.08 +/- 1.45 mg/day; p < 0.01). The iron requirements in patients on erythropoietin therapy were higher than in those without (2.63 +/- 1.59 vs. 1.88 +/- 1.68 mg/day; p < 0.05). However, excluding the menstruating women, the iron need in patients on erythropoietin were similar to those in subjects without this treatment (2.16 +/- 1.13 vs. 1.88 +/- 1.68 mg/day). All patients showed good compliance with an excellent tolerance. We have observed that in subjects on maintenance erythropoietin therapy, the iron requirements are stable. The administration of sodium ferric gluconate is safe and efficient in maintaining adequate body iron stores.