Randomized clinical studies, carried out in patients with haematological malignancies and with solid tumours, have consistently demonstrated that treatment with recombinant human erythropoietin (Epo) increases haemoglobin levels, reduces blood transfusion requirements, and improves the quality of life. In addition, identification of erythropoietin receptor (EpoR) expression on many types of non-erythroid and cancer cells has spurred an interest in the extra-haematological activities of Epo itself and other erythropoiesis-stimulating agents (ESAs). Epo and its derivatives have emerged as major tissue-protective cytokines in ischaemic and degenerative damage of cardiovascular, neurological and renal diseases, while their angiogenetic and immunomodulatory properties indicate that their therapeutic potential may extend well beyond erythropoiesis alone. Both preclinical and clinical data, however, have suggested that they may contribute to tumour progression and prejudice survival when administered to anaemic cancer patients, though the results are equivocal and the assumed mechanisms by which tumour growth could be promoted are not fully understood. While these findings offer new perspectives, they nonetheless demand caution in the employment of ESAs. Further, well-designed experimental and clinical studies are warranted.