Second generation blood tests to detect erythropoietin abuse by athletes: effective but not preventive?

BACKGROUND AND OBJECTIVES We previously developed blood tests that were introduced at the Sydney 2000 Olympic Games to identify athletes injecting recombinant human erythropoietin (rHuEPO). The aim of this study was to re-analyse our existing database to develop models with heightened sensitivity, using wherever possible blood parameters measurable with appropriate standards of analytical performance. DESIGN AND METHODS The principal database for this study was derived from a double-blind trial in which 57 recreational athletes were administered either rHuEPO or placebo. Standard discriminant analysis was used to derive two ON models (ON-hes and ON-he) and two OFF models (OFF-hr and OFF-hre) sensitive to accelerated and decelerated erythropoiesis respectively, utilising concentrations of hemoglobin (h), erythropoietin (e) and serum transferrin receptor (s), as well as percent reticulocytes (r). The ability of our models to detect rHuEPO administration was assessed by comparing model scores of subjects in the administration trial with the model scores of 1152 elite athletes from 12 countries. RESULTS The ability of the new models to detect rHuEPO administration was generally higher than that of our previous models, particularly during phases when low doses of rHuEPO were used, and after injections had ceased. INTERPRETATION AND CONCLUSIONS. The increased stability of the new blood parameters facilitates transport of samples to central laboratories, and the heightened sensitivity of the new models makes them better than existing models for federations wishing to screen samples for urine testing and to identify and target suspect athletes for out-of-competition testing. However procedures should be incorporated that respect an elevated model score caused by genetic, health or environmental circumstances.