The pharmacokinetics and pharmacodynamics of recombinant human erythropoietin (rh-EPO; CAS for EPO: 11096-26-7) after repeated intravenous and subcutaneous administrations in rats were studied. Administration of rh-EPO by both routes caused significant increases in hematocrit. The pharmacokinetics of rh-EPO after intravenous and subcutaneous administration exhibited nonlinearity. The pharmacodynamics of rh-EPO was analyzed using the maximum effect (Emax) and sigmoid maximum effect (sigmoid Emax) models. Both models involved the assumption that rh-EPO in plasma would stimulate the proliferation of erythroid progenitor cells. Akaike's information criterion for the Emax model was lower than that for the sigmoid Emax model, suggesting that the Emax model might be an optimal model. The rh-EPO concentration at which the effect is half of the maximum was 0.383 ng/ml. This pharmacodynamic analysis suggests that the maintenance of effective plasma concentration might be important for the efficacy of rh-EPO.