The aim was to assess the cost-effectiveness of erythropoietin (EPO) to reduce patients' exposure to perioperative allogenic blood products in orthopaedic surgery. The use of EPO was assessed for EPO used alone and for EPO, to augment preoperative autologous donation (PAD). A decision analytical model was designed incorporating (i) the risk of receiving allogeneic blood, (ii) the costs of blood products, (iii) the likelihood of developing transfusion-related diseases, (iv) the costs of transfusion-related diseases, (v) the impact of transfusion-related diseases on patient morbidity and mortality and (vi) the effect of EPO upon the probability of transfusion. The efficacy of EPO was derived from data from a meta-analysis of published randomized trials. Estimates for the other parameters were obtained by a systematic review of the literature. EPO alone led to only modest incremental benefit compared to no intervention for orthopaedic surgery (0.000024 life-years gained per patient). As an augmentation to PAD, EPO also led to modest benefits (0.000006 life-years gained per patient). For EPO compared to no intervention, the incremental cost per life-year gained was $66 million (Canadian). For EPO to augment PAD, the incremental cost per life-year gained was $329 million (Canadian). Detailed sensitivity analysis did not reveal any circumstances in which the cost-effectiveness ratios reached a level generally considered attractive. On the basis of cost-effectiveness, the use of EPO to reduce perioperative allogeneic transfusions in orthopaedic surgery did not meet criteria conventionally considered acceptable.