The Curious Case Of The IUML And The EPO Positive That Wasn't

Justin Sullivan/Getty Images

Warning: may contain subliminable references to Lance Armstrong.

In 2001, at least a dozen years after EPO was believed to have first begun to kill cyclists and destroy the reputation of cycling, the sport finally got an EPO test. Despite promising that any new EPO test would be applied to stored samples, the UCI decided to apply the test only to new samples. Still, something's better than nothing. Isn't it? And so long as the test was properly applied to all riders going forward, well, it'd catch the cheats. Wouldn't it?

The UCI, in its 40 Years Fighting Against Doping document, released in 2001, reported the arrival of the EPO test as follows:


"Success at last! During the Tour des Flandres, on 8 April 2001, the UCI proceeded with antidoping tests that can detect if exogenous EPO has been taken. The analyses of 40 riders who were tested were given to the laboratory of the University Institute of Legal Medicine (IUML) in Lausanne, and all produced negative results. No reliable method of detection existed previously. The blood tests set up by the SSCC measured the haematocrit value of riders, but were unable to detect its source - endogenous or exogenous (artificial). For health reasons, and not so much as a sanction, riders whose value exceeded 50% (47% for women) had an enforced rest period of 15 days imposed on them.

"The situation changed drastically at the beginning of 2001.Thanks to the laboratory of Châtenay-Malabry (F), led by Prof. de Ceaurriz, it is now possible to prove through a urine test that a person has taken EPO to artificially enhance their performance.

"In close co-operation with other federations, the UCI validated the new test on 1 April 2001.

"EPO, the star doping substance of the 1990s, captured the attention of the sporting world and of numerous laboratories for more than ten years. The sports federations were clearly unable to fight as effectively as they wanted to against a substance that remained undetectable for the researchers themselves. This problem has now been overcome, the 'monster', while not having been completely defeated, has at least been seriously injured!"

To those familiar with this document produced by the UCI, it will come as no surprise to learn that the above contains a number of mistakes, mis-statements and mis-directions. The EPO 'monster' was far from vanquished. The LNDD test was far from effective. And the 2001 Spring Classics were far from as pristine as the UCI tried to suggest by claiming that the Ronde was clean.

* * * * *

On the morning of April 18, at the start of the Flèche Wallone, Swiss cyclist Roland Meier (Team Coast) was subjected to a blood test. The results of this encouraged the testers to select him for a post-race anti-doping control.

The following day, April 19, Danish cyclist Bo Hamburger (CSC-Tiscali) was subjected to an out of competition anti-doping test.

Meier and Hamburger had been team-mates at TVM in the mid-nineties. But that was not the only thing they had in common.

On May 10, the Danish cycling union broke the news that Bo Hamburger was positive for EPO. A few days later, Hein Verbruggen hinted at the existence of a second EPO positive, while attending an international haematology congress in southern France. The Swiss Federation subsequently announced that the rider involved was Roland Meier.

Dr Jacques de Ceaurriz, director of the LNDD, had been bullish when the EPO test was introduced:

"Cyclists should be worried because they won't be able to take EPO the way they did in the Tour de France in 1998 and 1999. It's a major breakthrough, and gives us fresh impetus in the fight against drugs."

Dr Bengt Saltin, the Swedish head of the Danish anti doping agency, seemed to be echoing De Ceaurriz in his ebullience when he spoke to the media about the Hamburger case:

"I thought the cyclists were smarter. The test is so reliable that there will be no one that takes EPO without being caught. They might not have known how precise the test is, but now they do."

* * * * *

Saltin's confidence in the UCI's EPO test was mis-placed. The urine-based EPO test could only detect the drug if it had been used within seventy-two hours of the test. The test itself was easily defeated using detergent. And, in it's early use, was far from reliable.

One of the biggest problems with the test was that it produced false negatives. The science bit: the test used electrophoresis to compare the quantities of basic isoforms of certain proteins in the urine sample. More means more likely to have used EPO. You or I might show a percentage of basic isoforms in the high twenties to low forties, depending on who is telling you these things. The IOC set a threshold of eighty percent (one IOC-accredited laboratory in Paris - presumably Chatenay-Malabry - actually sometimes used a threshold of eighty-five per cent) which includes a safety margin of more than three standard deviations from a mean value, calculated from the validation studies for persons who tested negative, plus an additional precautionary allowance of 10%. The CAS were of the opinion that this produced the probability of a false positive in one in fifteen thousand case. They didn't compute a figure for false negatives.

Regular micro-dosing kept riders below the threshold without diminishing the effectiveness of the drug. This would become especially clear once scientists started explaining the workings of the EPO test to the riders and their entourages, particularly when they told them about that large middle ground between a clean sample and one they would have to declare dirty.

On the issue of false negatives, in 2008 the Copenhagen Muscle Research Centre in Denmark released results of a study it had conducted the previous year. Eight young men were dosed with EPO over a period of seven weeks. For the first two weeks they received injections of EPO every other day (the booster phase), followed by two weeks of weekly injections (the maintenance phase), followed by a three week period of no EPO use (the post phase). Urine samples were collected throughout the study period. These were submitted to two different WADA-accredited laboratories.

The first laboratory correctly identified the use of EPO in all samples during the initial 'booster' phase of the study. The second laboratory declared only one sample to be positive and the other seven samples to be suspicious, but not positive. During the maintenance and post phases, the first laboratory declared two samples to be positive, three to be suspicious and the remaining nineteen to be clean. The second laboratory declared all the samples to be clean.

Out of sixty-four samples submitted to the two labs, eleven were correctly identified as being positive, ten were declared to be suspicious and the remaining forty-three were declared clean. The better performing lab had a 68.8% chance of declaring a false negative, the worse 96.9%. And this was in 2007. Six years after the EPO test had been introduced.

During 2001, the UCI conducted 271 urinary EPO tests. Of those tests, 96% were not positive. The monster had been vanquished. Indeed. The eleven that did result in positives being declared were: Roland Meier (Team Coast), at the Flèche Wallonne; Bo Hamburger (CSC-Tiscali) in an OOC test the day after the Flèche; Sergio Barbero (Lampre) and Laurent Chotard (Mercury-Viatel) at the pre-Giro leg-loosener, the Tour de Romandie; Pascal Hervé (Alexia), Riccardo Forconi (Mercatone Uno) and Dario Frigo (Fassa Bortolo) during the Giro d'Italia; Txema del Olmo (Euskatel-Euskadi) at the Tour de France; the Swedish rider Niklas Axelsson at the World Championships; and Massimo Strazzer and the former USPS rider Juan Llaneras at the track World Championships. I'll leave you to discover for yourself how many of those cases failed to result in suspensions.

If you think the chances of beating the UCI's test were high, think about what the IOC were doing. At the Sydney Olympics in 2000 - and again at the Salt Lake Winter Olympics in 2002 - the IOC used a combined blood/urine EPO test. An athlete needed to be positive under both tests in order to be declared positive. At Sydney, the blood test - which had a greater window of opportunity than the urine test's three days - highlighted nine athletes, none of whom were positive under the urine test.

The UCI chose to use only the LNDD-developed urine test as proof of EPO use, but would also sometimes use a blood test as a screener. This is what happened in Meier's case. No blood was drawn in Hamburger's case.

According to the head of the IOC medical commission, Patrick Schamasch:

"The urinary test alone is not sufficient to be able to determine a final result. We must keep in mind the legal questions."

* * * * *

Testing of Meier and Hamburger's B samples commenced on June 5th. The testing was carried out by the same laboratory as had conducted the analysis of their A samples: the IUML. The B tests were witnessed by representatives of the riders. The results were passed back to the UCI and the respective federations.

In August 2001 it was announced by the Swiss Cycling Federation that Meier's B sample confirmed the A and accordingly he was to be suspended for eight months, with that ban suspended for the off-season. He would not be eligible to return to the pro peloton until July 19, 2002. He was also handed some financial penalties. He appealed his suspension to CAS.

Things were a bit more complicated for Bo Hamburger. Hamburger seemed to attract complications. Like his Salbutamol bust at the Tour de l'Avenir in 1993. Or an alleged confession to having used Ventoline and Pulmicort in 1998. Or his Corticoids bust at the Tour de France in 1999. In June it was announced that his B sample had confirmed the result of the A. Hamburger was initially found guilty of using EPO by the Danmarks Cykle Union (DCU), but appealed this to the Danmarks Idraets-Forbund (DIF), the Danish Olympic committee. In August they cleared Hamburger of EPO use, citing a procedural error in the testing of his B sample. DIF spokesman Finn Mikkelsen:

"In this case the problem was that there were two analytical results on the B sample. The lab has not given any indication of why there are two results - therefore there was doubt about the B result and thus the result from the B sample. I admit this is a strange case. The decision respected the EPO test, but they did not accept the circumstances around the B test."

The UCI appealed DIF's decision in the Hamburger case to the CAS.

Dr Laurent Rivier, of the Laboratoire Suisse d'Analyse du Dopage at the Institut Universitaire de Medicine Légale (IUML) in Lausanne, discussed the EPO test and the case of Bo Hamburger with the Danish newspaper, Ekstra Bladet:

"The tests are not one hundred per cent trustworthy but I think the UCI wanted to make one step forward to get things going and to diminish the massive use of EPO that has been observed in the cycling environment. The UCI definitely took some risks with introducing this test but now it will be interesting to see the legal effects it will have in the case of Bo Hamburger."

* * * * *

The CAS decision in the Hamburger case makes for interesting reading. The most important part - from our point of view today - relates to the non-existence of an EPO threshold:

"The UCI's rules do not contain any provision whereby a sample can be considered positive only if the rEPO exceeds a certain threshold. Evidence of even only minimal quantities of rEPO is sufficient for it to constitute a violation of the UCI's Antidoping Regulations."

The UCI's amended rules, which came into force on March 15, had this to say about the testing for EPO:

"the presence of an abnormal concentration of an endogenous hormone in class E or its diagnostic marker(s) in the urine of a competitor constitutes an offence unless it has been proven to be due to a physiological or pathological condition".

The CAS found that:

"Unlike the evidence required for certain substances such as nandrolone or caffeine, the UCI's rules do not provide that a result is positive only if the rEPO in the urine of a person who is tested exceeds a certain threshold. Rather, evidence of even only minimal quantities of rEPO is sufficient for it to constitute a violation of the UCI's Antidoping Regulations. The methods used by the IOC at the Sydney Olympic Games ultimately had the same objective.

The CAS went on to state that:

"the UCI's rules do not lay down any thresholds for the laboratory analysis. In particular, the threshold prescribed by the IOC, whereby a sample cannot be found to be positive unless more than 80% of the EPO findings are in what is called the 'basic range' of the test results, does not apply within the scope of application of the UCI's rules."

This threshold proved to be important. While the IUML testing of Hamburger's A sample had been above the IOC threshold, the testing of his B sample was more confusing. When first tested, it produced a value below the IOC threshold, 78.6%. The IUML staff had to carry out a second test on the B sample in order to produce a value - 82.4% - above the IOC threshold and consistent with the A sample, which showed 82.3%.

The IUML's director, Martial Saugy, was satisfied that, even though Hamburger's B sample had first produced a value below the IOC threshold, that value itself could still prove the use of EPO. Saugy expressed the view to the CAS that:

"there was no scientific justification for an 80% threshold and that a level marginally below 80% was reliable enough to allow one to assume a positive result."

Saugy also suggested to the CAS that if they wanted to insist on the 80% threshold then the average of the two tests on the B sample produced a result above that threshold.

The CAS decided that - effectively - Saugy and the staff at the IUML took a subjective 'I know it when I see it' approach to identifying EPO use. Worse, the CAS found that the IUML did not appear to apply their own rules consistently. In the case of Hamburger, Saugy was happy that an A sample above the IOC threshold could be supported by a B sample below that threshold. But had that A sample been below the IOC threshold he would have declared Hamburger clean, without reference to his B sample.

The CAS found in favour of Hamburger, finding that the double analysis of the rider's B sample did not confirm the results of the A sample:

"If the test results of the B sample have not been measured using the same standards as in the A sample, the A sample is not confirmed, rather a new analysis has been carried out pursuant to a different method of evaluation."

Roland Meier had also taken his case to the CAS, but his appeal was based on a variety of minor procedural errors of an administrative nature, along with the usual guff about the test itself being a piece of crap and the 'hey, could a face like this lie to you?' defence. In laughing Meier off the stage - his really was a slam-dunk case and the only grace the CAS showed him was to bring forward the commencement of his ban (meaning it would expire on May 17 2002) and reduce the fines imposed upon him - the CAS did make the following point with regard to the absence of a threshold for the detection of EPO in the UCI's rules:

"In connection with the evaluation of the test results, it should be pointed out that the UCI's rules and regulations - unlike those of the IOC - do not lay down any threshold which must be reached for the finding to be positive. Although the Panel does not consider such thresholds to be absolutely necessary in the rules and regulations of the federations, they are desirable to enable greater objective verification and, moreover, also for the purposes of acceptability."

In the absence of such a threshold, the CAS said

"the Panel must review, on a case by case basis, whether the levels, upon which the laboratory based its decision, justify a positive finding."

The CAS also again criticised the subjective 'I know it when I see it' approach to identifying EPO use:

"The witnesses [including the IUML's Martial Saugy] called by the Respondent [the UCI] gave the impression in their testimony that the finding of a positive doping result was based more on experience and subjective judgement than on objective reference points. The Panel does not, however, consider the standard, 'I know it when I see it', to be sufficient. The personal experience of a laboratory employee, no matter how much experience that employee has, is not sufficient to establish a positive doping result."

The CAS decisions in the cases of both Bo Hamburger and Roland Meier were delivered on January 28, 2002.

* * * * *

In our trawls earlier this year through the fight against doping in the nineties the IUML cropped up a couple or three times.

In 1996, it was the IUML that the Canadian researcher, Guy Brisson, was teamed up with in order to get the UCI's support, when he sought to develop a test for EPO. Brisson had wanted to carry out blood tests on riders at the Tour de Romandie, but the riders refused. Eventually, they agreed to allow themselves to be tested at the Tour de Suisse, after having received guarantees about the anonymity of the testing due to be carried out. Brisson's research, you will recall, came to nothing. Or, at least according to the UCI, it came to nothing. He himself, in 1998, claimed that he had actually developed a test for EPO which the UCI refused to implement, for fear of being sued by big name riders with deep pockets and guilty consciences.

In 1999, the IUML was also at the forefront of the UCI's fight against PFCs. 40 Years Fights Against Doping again:

"From the beginning of the 1999 season, PFC was put on the list of prohibited substances. The UCI, which in August 1998 commissioned the University Institute of Legal Medicine (IUML) in Lausanne, headed by Prof. Patrice Mangin, to checking for any presence of PFC in the blood of riders, has never recorded an offence."

Those tests, you will recall, started some months after Mauro Gianetti had been taken ill at the Tour de Romandie and was thought to have used PFC.

In 2001 Patrice Mangin and two IUML colleagues, Laurent Rivier and Martial Saugy, drew up a list of the principle doping products being used in the peloton, highlighting their side effects. As well as the usual suspects - EPO, steroids and corticosteriods - that list also included PFC. How the IUML knew PFC was being used if, as the UCI claimed, they had never once found any evidence of its use in the blood samples they tested ... well that's a question that has never been satisfactorily answered.

In 2004, while a newly developed test for homologous blood transfusions was being validated, the IUML were brought in. WADA at this stage were eager not to publicise the existence of such a test. The IUML applied the test to samples collected from cyclists, including some from the Tour de Romandie. The results were communicated to the UCI. For some reason, never satisfactorily explained, the UCI's Chief Medical Officer, Mario Zorzoli, discussed the results of those tests with at least one rider, effectively telling him to get his act together before the new test was introduced officially. Michael Ashenden, one of the developers of the new test, was particularly unhappy:

"I can understand why they would be reluctant to embroil themselves in controversy and bust too many cyclists. But when you know someone has done this [cheated], and you don't go after them, there's no excuse in my mind for that. At that point, I was disgruntled, to say the least, with their attitude. We bust our guts to get the test in place and they then don't see it through."

In 2005, the IUML's Martial Saugy was asked his opinion of the retrospective testing of urine samples from the 1999 Tour de France, which had apparently shown evidence of the use of EPO. Here's what he had to say then:

"You are looking at numbers and signals, but in the end what is most important here is the experience of the eyes of an expert. It's the 'now we see it - this looks like someone who has injected EPO."'

In more recent months, the IUML has been back in the news concerning the EPO test and the early months of its implementation. Martial Saugy had this to say to the Neue Züricher Zeitung:

"They [the four suspect samples] were taken at four different stages [during the Tour de Suisse in June 2001], so I don't know whether they were from four different riders or all of the same athlete. But the tests were not covered up, and it is also not correct that they could have been interpreted as positive. They were suspect, and you wouldn't stand a chance at all with that sole argument in front of a court.

"The Paris laboratory of Chatenay-Malabry fixed the criteria for a positive test result. An athlete was positive only if eighty percent of the signs typical for the use of synthetic EPO were found. [A suspect sample was one that] showed between seventy and eighty percent of the typical EPO parameters. That meant that the probability of doping was high. But because such a result can also be produced naturally, it was all about excluding false positives."

Remember, this is Saugy speaking in 2011, nine years after the CAS pointed out that the UCI had no EPO threshold in 2001. Ten years after he and his IUML colleagues tried to bust Bo Hamburger using test results which fell below the 'threshold.' This is the same Martial Saugy who told the CAS then that:

"there was no scientific justification for an 80% threshold and that a level marginally below 80% was reliable enough to allow one to assume a positive result."

Is it any wonder that the IUML is sometimes referred to as being the UCI's lab in Lausanne?