Abstract
In elite sport, the Athlete Biological Passport (ABP) was invented to tackle cheaters by monitoring closely changes in biological parameters, flagging atypical variations. The hematological module of the ABP was indeed adopted in 2011 by World Athletics (WA). This study estimates the prevalence of blood doping based on hematological parameters in a large cohort of track and field athletes measured at two international major events (2011 and 2013 WA World Championships) with a hypothesized decrease in prevalence due to the ABP introduction. A total of 3683 blood samples were collected and analyzed from all participating athletes originating from 209 countries. The estimate of doping prevalence was obtained by using a Bayesian network with seven variables, as well as "blood doping" as a variable mimicking doping with low-doses of recombinant human erythropoietin (rhEPO), to generate reference cumulative distribution functions (CDFs) for the Abnormal Blood Profile Score (ABPS) from the ABP. Our results from robust hematological parameters indicate an estimation of an overall blood doping prevalence of 18% in 2011 and 15% in 2013 (non-significant difference) in average in endurance athletes [95% Confidence Interval (CI) 14-22 and 12-19% for 2011 and 2013, respectively]. A higher prevalence was observed in female athletes (22%, CI 16-28%) than in male athletes (15%, CI 9-20%) in 2011. In conclusion, this study presents the first comparison of blood doping prevalence in elite athletes based on biological measurements from major international events that may help scientists and experts to use the ABP in a more efficient and deterrent way.