The Development of a Blood Test for EPO Abuse
Robin Parisotto and Mike Ashenden
Australian Institute of Sport
Sydney, Australia
Overview:
Anecdotal reports have suggested that endurance athletes have been blood doping with impunity for several decades. The original practice of autologous and/or homologous blood doping was superseded with the commercial availability of recombinant human erythropoietin (r-HuEPO), highlighted by the infamous 1998 Tour de France „drug busts.‰ However, efforts to detect r-HuEPO abuse remained elusive primarily due to the similar chemical structure of both the synthetic and natural forms of EPO. Subsequently much of the research was directed towards „indirect‰ detection in blood, accompanied by the introduction of the hematocrit (Hct) test by the UCI in the late 1990‚s. Many studies suggested an upper limit of a single hemopoietic marker as possible evidence of r-HuEPO abuse. The aim of our 1999 pilot study was to investigate whether multiple indirect markers of erythropoiesis were more effective in detecting use of r-HuEPO than a single marker.
The concept of using hemoglobin parameters to detect doping was first developed during the evaluation of the blood profiles of athletes at the Australian Institute of Sport. It was recognized at that time that hematological parameters were relatively stable over time, and that because these parameters were disturbed as the result of erythropoietin administration, they may be useful as indicators of doping.
Previous research in this area had usually suggested single indirect markers of r-HuEPO use. Findings had included that the serum concentrations of transferrin receptors increased with the administration of EPO,1 and that the hemoglobin content of reticulocytes decreased with EPO use.2 What emerged from looking at this previous research was the question of whether multiple markers of accelerated red blood cell production could provide greater discriminative power to detect r-HuEPO use than single markers.
The initial stage of research into this question was conducted in 1999, and involved three groups: placebo, EPO injection plus intramuscular iron injection, and EPO injection plus oral iron tablets. EPO injections were given three times per week during the administration phase, and blood samples were collected regularly. Additionally, hemoglobin mass was measured before, afterwards, and four weeks after administration ceased.
During the administration phase, participants‚ hematocrit increased significantly ˆ to such a level that would correspond to about a 4 percent increase in athletic performance terms. Percentage of macrocytic cells, reticulocyte hematocrit, and soluble transferrin receptor also increased during the administration phase, as did, of course, serum EPO levels.
It was also found that the group receiving oral iron tablets seemed to respond to EPO more than those given intramuscular iron injections, though the reason for this is not known ˆ possibly being attributable simply to individual responses.
To form the original experimental model, the five tested parameters (reticulocyte hematocrit, serum EPO, soluble transferrin receptor, hematocrit, and percent macrocytes) were each assigned a score. Data for the parameters from days 22 and 24 of the EPO administration phase were averaged for each group, and differences between the r-HuEPO and placebo groups were quantified. This resulted in a program in which, for example, an athlete who had elevated scores for only one parameter would have a relatively lower overall score than an athlete in whom four or five of the parameters were elevated.
Logistic Regression (logit) analysis used to evaluate all 31 possible combinations of parameters, with the valuation based on ability of logit to correctly distinguish between r-HuEPO recipients and placebo subjects at end of period of drug administration.
These models were then taken and applied to each of the blood samples collected during the study in order to identify the different levels of sensitivity and specificity of the different combinations of models we could derive. The models were also tested by application to data from 556 blood samples from athletes who had participated in studies involving natural altitude, simulated altitude, a six-day cycling race, or normal training.
This on-model score ˆ reflecting the combination of all five parameters ˆ correctly identified 94-100 percent of r-HuEPO recipients during the final two weeks of drug administration. There was one false-positive result in the placebo group, and one false-positive in the reference group (after the first night at simulated altitude).
Off-model analysis in this study, which was conducted using the hematocrit, reticulocyte hematocrit, and EPO parameters, was conducted 12-21 days after the cessations of EPO administration. This off-model correctly identified up to 72 percent of participants who had received EPO. There were no false-positive results in the off-model analysis in either the reference or placebo groups.
The overall conclusion, then, of this 1999 study was that multiple indirect hematological and biochemical markers used in combination are potentially effective for identifying current and recent users of r-HuEPO.
EPO 2000
During late August 1999, the International Olympic Committee (IOC) called for research to explore ways to detect EPO use. We at the Australian Institute of Sport submitted a proposal to the IOC, for which a research grant was approved in December 1999.
A major aspect of the EPO 2000 effort was to access international profiling study data. To this end, many blood centers which had studied the population we would be testing were contacted, and cooperative arrangements were made.
The international profiling portion of EPO 2000, which included data from North and South America, Africa, Asia, Southeast Asia, and Australia, had the goal of collecting blood samples from many countries to determine whether there were factors that could confound the test. These factors included gender, injury, sport, ethnic background, exercise and altitude/hypoxia. Blood profile signatures for this phase of the study were scored during a training period, during which three samples were taken from athletes over a two-week period.
Testing was continued with a repeat of the 1999 trial protocol in Australia (using more subjects than the original trial), and in Beijing, China, using members of a different ethnic group.
Two additional trials were undertaken, in Australia and Norway, to explore whether use of low-dose EPO could affect the study methodology.
Preliminary results from the EPO 2000 study at this point showed that responses were similar for all parameters across the 1999, Sydney 2000, and Beijing trials.
Next, to develop the test that was eventually submitted to the IOC, data from the Sydney trial were used to refine the models that were first developed in the 1999 study. Data from international profiling study were used to establish typical on- and off-model scores for the athlete population, and the models were integrated into a practical r-HuEPO testing procedure.
Also assessed were the effects of altitude, simulated altitude, repeated days of intense exercise, and various hematological abnormalities. There were no false positives associated with these factors.
The findings of the EPO 2000 validation study validated the trial results of the 1999 study, and also found that ethnicity, altitude exposure, intense exercise and/or competition did not significantly affect blood parameters.
Analysis of the on-model, which tested for current use of EPO, found that out of about 3500 subjects, only two false-positive results were returned. So, while this model was highly sensitive, it could not be used by itself for the detection of r-HuEPO abuse.
Analysis of the off-model, which tested for use of EPO up to four weeks after drug cessation, showed the test to be highly specific, with no false-positive results from the testing of about 3500 subjects.
Overall, this validation study showed that an Œindirect‚ approach was a viable approach for the detection of EPO use.
In July 2000, these results were taken to the IOC, which had called a special meeting of the doping commission. The researchers involved and the findings were subjected to quite intense questioning over the course of the two-day meeting, with the outcome being that the on-model protocol, in combination with the direct urine test for EPO, were accepted as a sanctionable test.
The EPO test was implemented in the Sydney Olympic Games, with approximately 300 blood samples taken in the two weeks prior to the opening ceremony. It is believed that this implementation had a deterrent effect. The testing resulted in one high on-model score among the tested athletes, but the urine test was not positive in this case, so it was not sanctioned. Of the 300 tests performed, there were also seven high off-model results.
Since the Sydney Games implementation, more research has resulted in the refinement of the models used. This research has resulted in a simplification of the models under which the hematocrit parameter was replaced by hemoglobin, reticulocyte hematocrit was replaced by percent reticulocytes, and percent macrocytes and serum transferrin were excluded, with serum EPO remaining. These changes have resulted in more sensitive models that are able to detect EPO abusers in low-dose maintenance phases. BLR
References:
1. Gareau et al, Horm Metab Res. 26:311-312, 1994.
2. Brugnara et al, J Lab Clin Med. 123:660-667, 1994.
 Copyright 1995-2010 - Carden Jennings Publishing Co., Ltd. All rights reserved. The material available at this site is for educational purposes only and is NOT intended for any diagnostic, clinically related, or other purpose. Carden Jennings Publishing Co., Ltd., assumes no responsibility for any use or misuse of this material and makes no warranty or representation of any kind with respect to the material available at this site.
|
![[BLlogo7.gif]](http://mmserver.cjp.com/images/blood/BLlogo7.gif)
