The effect of common hematologic abnormalities on the ability of blood models to detect erythropoietin abuse by athletes.

BACKGROUND AND OBJECTIVES Algorithms that combine scores from multiple blood parameters are demonstrably effective in highlighting recombinant human erythropoietin (rHuEPO) administration, and have been used to deter rHuEPO use by athletes. These models are sensitive to atypical levels of blood parameters encountered during altered states of red cell production. Because hematologic abnormalities can also result in unusual blood profiles, the aim of this study was to document the incidence and magnitude of such abnormalities in an elite athlete population. DESIGN AND METHODS We screened blood samples obtained from 413 female and 739 male elite athletes from 12 countries for known hematologic abnormalities, and compared the algorithm scores for these athletes with those of their healthy counterparts. We also established the magnitude of blood parameters required for model scores to exceed cut-offs associated with rHuEPO use. RESULTS We found that 0.7% of male and 2.4% of female athletes were iron deficient either with our without anemia. An additional 1.4% of males and 1.0% of females had hemoglobinopathies. On average these athletes' model scores were at or below the score of their healthy counterparts. The greatest influence on our models was hemoglobin concentration. Values of other parameters must exceed normal ranges by a substantial margin in order for model scores to approach levels associated with rHuEPO use. INTERPRETATION AND CONCLUSIONS The hematologic disorders we encountered in elite athletes were not associated with model scores that exceeded the nominal cut-offs that we have previously recommended to delineate rHuEPO use. We did not find any abnormalities among elite endurance athletes that were associated with high model scores.