BACKGROUND AND OBJECTIVE The use of recombinant human erythropoietin (r-HuEPO) to enhance athletic performance is prohibited. Existing tests cannot readily differentiate between exogenous and endogenous EPO. Therefore the aim of our study was to investigate possible indirect detection of r-HuEPO use via blood markers of altered erythropoiesis. DESIGN AND METHODS Twenty-seven recreational athletes were assigned to three groups prior to a 25 day drug administration phase, with the following protocols: EPO+IM group (n = 10), 50 Ukg(-1) r-HuEPO at a frequency of 3wk(-1), 100 mg intramuscular (IM) iron 1wk(-1) and a sham iron tablet daily; EPO+OR group (n = 8), 50 U.kg(-1) r-HuEPO 3wk(-1), sham iron injection 1wk(-1) and 105 mg of oral elemental iron daily; placebo group (n = 9), sham r-HuEPO injections 3wk(-1), sham iron injections 1wk(-1) and sham iron tablets daily. Each group was monitored during and for 4 weeks after drug administration. RESULTS Models incorporating combinations of the variables reticulocyte hematocrit (RetHct), serum EPO, soluble transferrin receptor, hematocrit (Hct) and % macrocytes were analyzed by logistic regression. One model (ON-model) repeatedly identified 94-100% of r-HuEPO group members during the final 2 wk of the r-HuEPO administration phase. One false positive was recorded from a possible 189. Another model (OFF-model) incorporating RetHct, EPO and Hct was applied during the wash-out phase and, during the period of 12-21 days after the last r-HuEPO injection, it repeatedly identified 67-72% of recent users with no false positives. INTERPRETATION AND CONCLUSIONS Multiple indirect hematologic and biochemical markers used simultaneously are potentially effective for identifying current or recent users of r-HuEPO.