Iron supplementation has become an integral part of the management of patients receiving epoetin therapy, and clinicians have found it necessary to learn how and when to use it to the best advantage. Three routes of administration for iron are available: oral, intramuscular, and intravenous. Oral iron has the advantage of being simple and cheap, but it is limited by side-effects, poor compliance, poor absorption, and low efficacy. Intravenous iron is the best means of guaranteeing delivery of readily available iron to the bone marrow, but it requires greater clinical supervision. The i.v. iron preparations vary widely in their degradation kinetics, bioavailability, side-effect profiles, and maximum dose for single administration. Iron dextran is hampered by a small but significant risk of anaphylaxis, whereas all i.v. iron preparations can induce "free iron" reactions if the circulating plasma transferrin is overloaded. Intravenous iron may be given in advance of epoetin therapy, as concomitant treatment to prevent the development of iron deficiency, as treatment of absolute or functional iron deficiency, or as adjuvant therapy to enhance the response to epoetin in iron-replete patients. Markers of iron status that may indicate a need for i.v. iron include a serum ferritin of less than 100 microg/liter, a transferrin saturation of less than 20%, and a percentage of hypochromic red cells more than 10%. Various regimens are available for giving i.v. iron: low-dose administration of 20 to 60 mg every dialysis session in hemodialysis patients, medium-dose administration of 100 to 400 mg, and high-dose administration of 500 to 1000 mg. Iron sodium gluconate can only be given as a low-dose regimen because of toxicity, whereas the only preparation suitable for high-dose administration is iron dextran. Although concerns have been raised regarding iron overload and long-term toxicity with i.v. iron therapy in terms of increased risk of infections, cardiovascular disease, and malignancy, there is little evidence to substantiate this in patients receiving epoetin. Care should be taken, however, to prevent the serum ferritin rising above 800 to 1000 microg/liter and the transferrin saturation above 50%. Provided this is done, the benefits of i.v. iron almost certainly outweigh the risks in terms of optimizing the response to epoetin therapy.