Anemia in cancer patients undergoing treatment is common and can cause debilitating symptoms such as fatigue and reduced exercise tolerance. The introduction of recombinant human erythropoietin represents a potential improvement in the treatment of this condition. Clinical studies in patients with solid tumors and nonmyeloid hematologic malignancies have convincingly shown an improvement in mean hemoglobin concentration, a reduction in transfusion requirement along with an improvement in quality of life scores, although an effect on survival is less clear. In myeloid disorders such as myelodysplasia, response to single-agent recombinant human erythropoietin is disappointing but significant synergism with granulocyte colony stimulating factor has been demonstrated and different dosing regimens may also improve response. Unfortunately, a significant proportion of patients remain refractory to treatment. Efforts have been made to identify treatable causes of erythropoietin refractoriness, such as functional iron deficiency, and concomitant intravenous iron supplementation does appear to improve response rates. The search for pretreatment factors that predict response has been largely disappointing, although a promising model for myelodysplasia has been developed that awaits large-scale evaluation. Recombinant human erythropoietin is well tolerated, although there were concerns in the late 1990s due to a rising incidence of pure red cell aplasia in chronic renal failure patients treated with subcutaneous Eprex® (Ortho Biologics) in Europe. Since potentially contributory manufacturing processes have been identified and corrected, the incidence of this complication has been falling.