Doping test reliability

jza86058

I run medical drug trials for a living. We test patients for illegal meds before enrolling them on studies and we monitor the test medicine at regular intervals during the trial. These studies represent massive investments for pharmaceutical companies and you can be sure we use the very best labs. Even so, many mistakes are made including mixed up samples, errors in dilutions, errors with reference standards and sample storage. Many of these types of human error can be accounted for with cycling doping tests by the secondary analysis of the B sample, but what about the validity of the assays themselves?



Can somebody direct me to information about how the tests for homologous / heterologous transfusions are performed, and how these tests have been validated. I saw something about examination of the shape of red blood cells, which seems suspect to me. I would have thought the most reliable way would be a genotyping (DNA) test.



I hear the need to come down hard on doping but if the penalties are going to be as severe as some wish for (life time ban), there is a need to ensure the validity of the laboratory based tests. Some of the big name pro’s will be tested several hundred times in their career. A false +ve rate of 0.1% would be considered pretty good for many clinical lab test (other than DNA based analyses). Given that each blood / urine sample has to be tested for multiple potential doping products, this sort of error rate could end up with half a dozen clean riders receiving life time bans per year in the pro tour.

jza86058

Sorry Dr, I can't subscribe to your view of the world. How do you expain the German TV station suspending coverage midway through this years TDF?

Anyway, I don't much care what the TV stations think, I'm more worried about the rest of the world who seem to think these tests are infallible and calling for life time bans.

- Can anybody tell me where to get more information about what analytes are tested in the urine and what endpoints are measured in the blood (other than haematocrit).

skeptical

Long time lurker, first time poster. Thanks for providing a very entertaining place to hang out during Tour time! After two years, I am finally motivated to comment. For the record, I believe all the top racers dope. I don't care, as long as they don't kill themselves - I have loved the blood, sweat, and drama of the Tour for over 2 decades. This year has been fascinating to follow, but then I don't have a horse in the race.

Anyway, I thought for what it's worth, I googled the rEPO testing and thought I'd post some thoughts on it. EPO is a glycosylated protein; the difference between endogenous human EPO and administered recombinant EPO is not in the amino acid content but in the sugars which are covalently attached and exposed on the surface ("glycosylation"). This difference is due to the different glycosylation machinery in the human cells which make endogenous EPO in our bodies and the glycosylation machinery in the hamster/mouse/monkey/yeast/insect cells which are used to make recombinant EPO, and in the additional processing of the sugars to aid in storage or pharmacological properties (e.g. it might extend the stability of the rEPO to clip off some of the sugars). The different brands of rEPO are slightly different from each other because they use different expression systems or post-expression processing. Anyway, according to a 2006 review (http://www.medicalnewstoday.com/articles/38295.php) of an article published in Blood by scientists at the University of Leuven (i.e. respectable university, highly reputable peer-reviewed international hemostatic journal), the IOC/WADA rEPO test is/was based on a 2-step process of gel electrophoresis to separate the proteins in the sample based on size and charge, and then detection using antibodies which bind to rEPOs. You may or may not care that I'm a biochemistry professor, but these tests ("Western blots") are run routinely in labs everywhere, and the specificity of the antibody recognition is crucial (i.e. in this case, how well it can discriminate between rEPO and non-rEPO). Since the difference between EPO and rEPO is subtle, the reliability of the test is highly dependent on the antibodies being able to selectively bind to rEPO. The Blood study shows that the rEPO antibodies cross-react to even non-EPO (let alone non-rEPO) proteins, and that many of the papers which IOC/WADA uses to support their test were missing crucial controls and could have been susceptible to false positives. This cross-reactivity should be a real deal-killer in rEPO testing. How is this relevant to the discussion? Well, if the rEPO antibodies recognize slight differences in sugar composition, then sample handling is really crucial and can alter antibody binding. If the sample isn't kept frozen at low enough temperatures, then chemical reactions such as hydrolysis or oxidation, or enzymatic deglycosylation, can alter the sugar composition and thus alter antibody recognition. Add to that the published observations that the rEPO antibodies can cross-react with non-EPO proteins and ... wow. I googled the rEPO test after reading in the tongue-in-cheek "Lance is the best" thread. So it is possible that Lance's years-old samples, if they were not kept properly frozen the entire time, could theoretically give a false positive. And according to the Leuven scientists, false positives can occur with the rEPO antibodies anyway. Personally, I don't really care about Lance's samples since I believe he, like everyone else, was artificially enhanced. It's just as a scientist, I find it bothersome that the testing procedures are neither well-tested nor well-documented. It's just incredibly poor methodology, and who gets "caught" and who gets off is entirely arbitrary and a crapshoot.

There's a nice quote here: http://www.iht.com/articles/2005/09/04/news/lance.php
""This is not like a pregnancy test, where you are either pregnant or you're not," said Dr. Nicolle Packer, executive vice president of Proteome Systems in Sydney, which has one of the research grants. "It has to be prepared carefully and interpreted by an expert, who can mostly call it, I believe," she said in a tele phone interview. "But it is definitely skill-based, and that is why WADA is looking for a more clear-cut test.""
B.S. A pregnancy test is also antibody-based. The difference is that the antibodies in a pregnancy test are highly specific. The antibodies IOC/WADA are using are apparently not. If interpreting a IOD/WADA Western is "skill-based" then the test is not reliable enough to use to end someone's cycling career. Scientists publish Western blots all the time in papers to support their conclusions. I think if IOC/WADA are going to use a Western to ban a cyclist, they should publish it. There are thousands of us out here who look at each others' published Westerns and judge each others' research every day. What are IOC/WADA afraid of?

I like some aspects of a proposal put forward at http://www.fatcyclist.com/2007/07/2...ing-a-solution/ - i.e. instead of checking for doping, TPTB should just set limits on acceptable physiological parameters and just not care whether those parameters are met naturally or artificially. Seems pretty pragmatic and logical to me.

Sorry for the essay. Back to lurking.

skeptical

Do you mean PCR-based? Theoretically this would be definitive, except if one is looking for a minor "contamination" from the transfusion, it could be argued that false positives could also result easily from minor contamination by other routes. Apparently the current method to detect homologous blood doping is to use flow cytometry (fluorescence-based cell sorting). This has recently become a commonly used research tool but I don't use it myself so I'm not an expert. Flow cytometry is typically used in research to follow the cellular changes upon some treatment/perturbation of the cells. It can be unambiguous when the cellular changes are distinct and easily discriminated, and less unambiguous when the changes are subtle. What the heck are WADA/IOC measuring by flow cytometry? From what little I could glean online, they are using fluorescently tagged antibodies. So again, the reliability of the rest is highly dependent on the specificity of the antibodies, the quality of the samples, and the nature of the controls. Who knows how good this test is? With careers at stake, they really need to make their controls and validation of methodology public. It's just how good science is done.

Are accused dopers given the raw data from the tests so that they can provide them to professional scientists for a second opinion? Are they allowed to request that duplicate tests be completed by a reputable, expert, impartial academic lab? Mistakes are made in science all the time, but we publish our methods and data so that others can reproduce and confirm, and in the long run the reliable information tends to persist and the unreliable tends to be forgotten. It's really poor science if IOC/WADA do not make their data available to be critiqued. I know their defense is that it makes it easier for dopers to evade, but I don't see how that's true. I don't see how showing a Western or cell population distribution determined by flow cytometry will give a doper any clues on how to evade the test. If the test is good/unambiguous, it would actually deter doping. The way things are now, we have no idea who to believe.

(Not quite an essay this time, thank goodness.)

Gregers

Professor Skeptical, an intriguing and informed analysis. It always seemed to be a given that Epo was only present in a sample if it had actually been introduced some time before the sample was taken. In other words, it cannot constitute itself in a test tube via a chemical reaction taking place. You have indicated that this is not necessarily the case, but there is a low order of probability of it happening. Is this a vanishingly low possibility or something that could realistically happen? Is this probability multiplied exponentially by the fact that EPO was supposedly discovered in 5-6 of LA's samples, or if it takes place is it more than likely to appear in all of them-given that all of the samples were kept in the same condition? I've always thought that Armstrong was a phoney in any number of respects too tediously rehearsed to bore you with here. However, this might shed some further light on a matter for which he is regularly (and with considerable justification) presumed to be guilty.

Klodifan

and i think that is so sad, on a human level. and yet, i tune in and watch, thereby feeding the system. its a vicious, vicious cycle...

Dream Plus

Red Blood cells lack Nuclei, so genotyping may be difficult. Maybe mitochondrila DNA could be used? Surface markers (I think a glycoprotein) are used to distinguish populations of red blood cells.