Several investigators reported that individuals with diabetes and women on hemodialysis treated with recombinant erythropoietin (EPO) attained lower hematocrits than individuals without diabetes and men. It is unclear whether these observed differences in achieved hematocrits are caused by inherent biological differences In responsiveness to EPO or undetected differences in modifiable factors that affect response to EPO. Also potentially modulating response to EPO is diurnal variation in the bioavailability of serum iron. To address these issues, we studied 309 patients undergoing hemodialysis in two large facilities in New York City. Retrospective data collected monthly for 3 months included patients' hematocrit, dose of EPO, urea reduction ratio (URR), total amount of intravenous iron administered, serum albumin concentration, transferrin saturation, and time of day patient underwent dialysis. The 309 study subjects (165 women, 144 men) included 207 blacks (67%), 74 Hispanics (24%), 23 whites (7%), and 5 Asians (2%) with a mean age of 55.4 +/- 15.6 (SD) years. Despite a greater mean URR (74% +/- 6.4% versus 71% +/- 6%; P = 0.001) and a 39% greater dose of EPO (97 +/- 65 versus 59 +/- 53 U/kg; P = 0.001), women (36% +/- 3.5%) had hematocrits equivalent with men (36.5% +/- 3.7%; P = not significant [NS]). There was no difference in the amount of intravenous iron administered to men (375 +/- 389 mg) and women (377 +/- 413 mg; P = NS). Diabetes mellitus (P = 0.48) did not significantly affect the odds of attaining a hematocrit greater than 33% after adjustment for URR, EPO dose, and amount of intravenous iron administered. The time of day a patient underwent dialysis (P = 0.93) had no effect on their response to EPO. We conclude that gender, but not diabetes status or time of dialysis, modulates response to EPO in hemodialysis patients. (C) 2001 by the National Kidney Foundation, Inc.