Anaemia is frequently diagnosed in patients with cancer, and may have a detrimental effect on quality of life (QoL). We previously conducted a systematic literature review (1996-2003) to produce evidence-based guidelines on the use of erythropoietic proteins in anaemic patients with cancer.[Bokemeyer C, Aapro MS, Courdi A, et al. EORTC guidelines for the use of erythropoietic proteins in anaemic patients with cancer. Eur J Cancer 2004;40:2201-2216.] We report here an update to these guidelines, including literature published through to November 2005. The results of this updated systematic literature review have enabled us to refine our guidelines based on the full body of data currently available. Level I evidence exists for a positive impact of erythropoietic proteins on haemoglobin (Hb) levels when administered to patients with chemotherapy-induced anaemia or anaemia of chronic disease, when used to prevent cancer anaemia, and in patients undergoing cancer surgery. The addition of further Level I studies confirms our recommendation that in cancer patients receiving chemotherapy and/or radiotherapy, treatment with erythropoietic proteins should be initiated at a Hb level of 9-11 g/dL based on anaemia-related symptoms rather than a fixed Hb concentration. Early intervention with erythropoietic proteins may be considered in asymptomatic anaemic patients with Hb levels 11.9 g/dL provided that individual factors like intensity and expected duration of chemotherapy are considered. Patients whose Hb level is below 9 g/dL should primarily be evaluated for need of transfusions potentially followed by the application of erythropoietic proteins. We do not recommend the prophylactic use of erythropoietic proteins to prevent anaemia in patients undergoing chemotherapy or radiotherapy who have normal Hb levels at the start of treatment, as the literature has not shown a benefit with this approach. The addition of further supporting studies confirms our recommendation that the target Hb concentration following treatment with erythropoietic proteins should be 12-13 g/dL. Once this level is achieved, maintenance doses should be titrated individually. There is Level I evidence that dosing of erythropoietic proteins less frequently than three times per week is efficacious when used to treat chemotherapy-induced anaemia or prevent cancer anaemia, with studies supporting the use of epoetin alfa and epoetin beta weekly and darbepoetin alfa given every week or every 3 weeks. We do not recommend the use of higher than standard initial doses of erythropoietic proteins with the aim of producing higher haematological responses, due to the limited body of evidence available. There is Level I evidence that, within reasonable limits of body weight, fixed doses of erythropoietic proteins can be used to treat patients with chemotherapy-induced anaemia. This analysis confirms that there are no baseline predictive factors of response to erythropoietic proteins that can be routinely used in clinical practice if functional iron deficiency or vitamin deficiency is ruled out; a low serum erythropoietin (EPO) level (only in haematological malignancies) appears to be the only predictive factor to be verified in Level I studies. Further studies are needed to investigate the value of hepcidin, c-reactive protein, and other measures as predictive factors. In these updated guidelines, we explored a new question of whether oral or intravenous iron supplementation increases the response rate to erythropoietic proteins. We found no evidence of increased response with the addition of oral iron supplementation, but there is Level II evidence of improved response to erythropoietic proteins with the addition of intravenous iron. However, the doses and schedules for intravenous iron supplementation are not yet well defined, and further studies in this area are warranted. The two major goals of erythropoietic protein therapy are prevention or elimination of transfusions and improvement of QoL. The total body of evidence shows that red blood cell (RBC) transfusion requirements are reduced following treatment with erythropoietic proteins. This analysis also confirms that QoL is significantly improved in patients with chemotherapy-induced anaemia and in those with anaemia of chronic disease following erythropoietic protein therapy, with more robust evidence now available that QoL was improved in studies investigating early intervention in cases of chemotherapy- or radiotherapy-induced anaemia. There is only indirect evidence that patients with chemotherapy-induced anaemia or anaemia of chronic disease initially classified as non-responders to standard doses proceed to respond to treatment following a dose increase. None of the studies addressed the question in a prospective, randomised fashion, and so the Taskforce does not recommend dose escalation as a general approach in all patients who are not responding. There is still insufficient data to determine the effect on survival following treatment with erythropoietic proteins in conjunction with chemotherapy or radiotherapy. Our analysis of survival endpoints in studies involving patients receiving radio(chemo)therapy found that most studies were inconclusive, with no clear link between the use of erythropoietic proteins and survival. Likewise, we found no clear link between erythropoietic therapy and other endpoints such as local tumour control, time to progression, and progression-free survival. There is no evidence that pure red cell aplasia occurs in cancer patients following treatment with erythropoietic proteins, and the fear of this condition developing should not lead to erythropoietic proteins being withheld in patients with cancer. There is Level I evidence that the risk of thromboembolic events and hypertension are slightly elevated in patients with chemotherapy-induced anaemia receiving erythropoietic proteins. Additional trials are warranted, especially to define the optimal doses and schedules of intravenous iron supplementation during erythropoietic therapy. While our review did not address cost benefit evaluations in detail, the consensus is that studies taking into account all real determinants of cost and benefit need to be performed prospectively.