Anemia is very common in head and neck cancer patients, and seems to be correlated with intratumoral hypoxia. Anemia is one of the main prognostic factors of locoregional recurrence and, in some studies, of poor survival. Blood transfusions and human recombinant erythropoietin (rHuEPO) are the two main methods used in clinical practice to correct hemoglobin level during curative treatment. Blood transfusions were rarely evaluated, and did not influence locoregional control of patients treated with radiotherapy with or without chemotherapy. Retrospective studies evaluating combined treatment of rHuEPO and radiotherapy reported positive impact on locoregional recurrence and actuarial survival. Since the end of 2003, this approach is a matter for debate after the negative results of a prospective randomized study on progression-free survival concerning head and neck cancer patients treated with definitive or postoperative external radiotherapy with or without rHuEPO. Although many biases were reported against this publication, several questions are to be answered in the near future. Among them, erythropoietin receptor expression and activation on tumour cell seem to be the more appropriate explanation of these negative results. In October 2004, preliminary results of the RTOG 99-03 study have been presented at the Astro annual meeting in Atlanta. This prospective randomized trial was designed to determine if concurrent rHuEPO administration (40,000 units) with radiotherapy (with or without chemotherapy) could improve locoregional control in non-operative head and neck cancers. In the rHuEPO arm, haemoglobin level was significantly increased compared with control arm. However, the addition of concurrent rHuEPO to definitive radiotherapy did not improve locoregional control or survival for mildly/moderately anemic patients with head and neck squamous cell carcinoma. Future clinical trials using biological markers are thus imperative to target which patients could benefit from these molecules.