Effective treatment of anemia in hemodialysis patients requires ongoing monitoring of iron status. The purpose of this study was to determine levels of commonly used iron indices predictive of iron deficiency in this population. Forty-seven patients with baseline serum ferritin levels < 600 ng/mL were treated with intravenous iron dextran (INFeD; Schein Pharmaceutical Inc., Florham Park, NJ), 1000 mg over ten hemodialysis treatments. Patients whose hematocrit value increased by 5% or who had a 10% decrease in their erythropoietin dose by 2 months were classified as having iron deficiency (N = 31; 66%). All other subjects were classified as having adequate iron (N = 16; 34%). There was no statistically significant difference in baseline serum ferritin, transferrin saturation, mean cell volume, mean cell hemoglobin content, or red cell distribution width between the two groups. Receiver operator curves demonstrated that none of the iron indices had a high level of utility (both sensitivity and specificity > 80%). Two tests had marginal utility, serum ferritin at a level of < 150 ng/mL, and transferrin saturation < 21%. It was concluded that because of the tests' marginal utility, they should only be interpreted in the context of the patient's underlying erythropoietin, responsiveness. In patients who are responsive to erythropoietin, a transferrin saturation value < 18% or serum ferritin level < 100 ng/mL should be used to indicate inadequate iron. When erythropoietin resistance is present, transferrin saturation of < 27% or serum ferritin < 300 ng/mL should be used to guide iron management.