The need for a renewable source of erythropoietin to treat the anaemia of chronic renal failure was first recognized in the 1960s, but cloning and expression of the human gene was not achieved until 1983. Clinical testing of recombinant human erythropoietin (r-HuEPO) began in 1985, leading to the first licence as a therapeutic agent in 1988. The first clinical trials showed that an intravenous dose requirement of about 200 IU/kg/week would increase haemoglobin concentrations to 10-12g/dl in >90% of haemodialysis patients. Subcutaneous administration has subsequently been found to be effective, and may allow lower maintenance doses. It is now the route of choice in Europe, but not the USA. The best marker of benefit of the introduction of r-HuEPO is the reduction in need for regular blood transfusions. A marked improvement in anaemia-related symptoms has been clearly demonstrated. The most important factor in optimizing the response to r-HuEPO is iron supply. The marrow should be stimulated slowly, to allow mobilization of iron stores. Functional or absolute iron deficiency should be pre-empted by regular iron supplementation. It is also important to recognize resistant states induced by inflammation and bleeding, and to exclude severe hyperparathyroidism, aluminium overload and other haematological diseases. The most important adverse events associated with r-HuEPO are increased blood pressure and a possible increased risk of access failure. These are, however, challenges to improve practice, not reasons to avoid the use of r-HuEPO.