The incidence of pure red cell aplasia (PRCA) in patients with chronic kidney disease associated with the subcutaneous (sc) administration of epoetin alfa (EPREX®) began to increase in 1998. As part of an intensive investigation into the reasons for this increase, an in vivo model was developed to assess the ability of potential causative factors to stimulate an immune response to recombinant human erythropoietin (rHuEPO). Due to species differences in protein sequence and as well as antigen processing and/or presentation, animal models cannot fully predict immunogenicity in humans. However, changes in relative immunogenicity have been readily detected in animal models. Moreover, transgenic animals have been utilized to negate the effects due to sequence differences between species. The EPO sequence is highly homologous between mouse and man. We found it difficult to generate anti-EPO antibodies in both human EPO transgenic mice and two normal mouse strains (Balb/c and BD-F1). To mitigate the complicating factors associated with the transgenic model (e.g. difficulty breeding, improper regulation of EPO production), normal mouse strains were selected for the development of animal models. Mice were injected sc with rHuEPO weekly for 4–6 weeks. Serum samples were analyzed 1–2 weeks following the last injection to ensure complete clearance of the injected protein. Serum samples were screened by an ELISA and/or by a surface plasmon resonance (BIAcore) method. In animals injected with rHuEPO alone, anti-EPO antibodies were either absent or present at very low levels. The addition of an adjuvant to the immunization protocol was able to boost both the rate of occurrence and titer of the immune response and resulted in the generation of anti-EPO antibodies that, in most cases, recognized both human and mouse EPO. Some of the mice exhibited a dramatic reduction in hematocrit, suggesting that the anti-EPO antibodies neutralized endogenous EPO. The development of an anti-EPO antibody response in humans is an extremely rare event that cannot be reproduced in an animal model without significant manipulation of the system (e.g. use of an adjuvant with doses of EPO considerably higher than therapeutic levels). Nevertheless, the fact that anti-EPO antibodies can develop with an appropriate immunization protocol allows for this model to be used to evaluate factors that may potentiate an immune response to rHuEPO.