Sports Scientist Ross Tucker Tackles Our Biggest Unanswered Questions About the Shelby Houlihan Case

Tucker: “The contaminated food explanation doesn’t stand up to basically any level of scrutiny.”

by Ross Tucker and LetsRun.com
April 21, 2022

By now you’re likely familiar with the Shelby Houlihan doping case. Houlihan, the American record holder at 1,500 and 5,000m, tested positive for nandrolone, a banned steroid, in December 2020. Houlihan denied doping and eventually argued that a tainted pork burrito was the likely source of nandrolone in her urine.

Houlihan, hoping to compete at the 2020 Tokyo Olympics, asked for an expedited hearing before the Court of Arbitration for Sport (CAS). However, in June 2021, the CAS arbitrators ruled against Houlihan’s burrito defense and she was banned from the sport for four years for a doping offense.

In September 2021, CAS released its full Reasoned Decision on the Houlihan case. It’s 44 pages long and full of scientific concepts such as GC/C/IRMS, delta-delta values, and carbon isotope signatures.

After reading the decision, we still had many questions about the decision, the case, and the scientific concepts underpinning it, especially since Houlihan continues to express her innocence.

To get answers, we turned to a man who knows running, science, the global anti-doping system, and most importantly, how to explain scientific concepts to the masses — Ross Tucker of Sportsscientists.com and the Real Science of Sport podcast — and asked him to analyze the ruling.

Ross has written two pieces on the case for LetsRun.com. In the first piece, which you can read at this link and is more technical, Ross does a scientific breakdown of the CAS ruling from start to finish. To summarize that piece, Ross finds the Athletics Integrity Unit ruled correctly in issuing an Adverse Analytical Finding for the nandrolone detected in Houlihan’s system. The burden then shifted to Houlihan to show the nandrolone use was not intentional. Ross agrees that CAS ruled correctly that Houlihan did not show either 1) the amount of nandrolone in her system or 2) the isotope signature of the nandrolone likely came from natural sources, so she was judged to have committed a doping offense.

In the piece below, Ross goes beyond the technical analysis, giving his opinions on the case and answers some of our most pressing unanswered questions. If you have time to only read one piece (they are both long), we suggest you read this one.

What does Ross think is the most likely explanation for the nandrolone in Houlihan’s system? Does he think she was intentionally doping? Is there any chance she is innocent? Should athletes be banned for the amount of nandrolone found in her sample? Ross shares his thoughts below.

Did CAS Get It Right?

Q: Ross, thank you for taking a look at this case for us! Sorry, we’re not scientists and you had to churn out a 5,000+ word piece (shortened from its original 9,000 words) on this. For the record, super big picture, you agree that CAS ruled correctly on this, right?

Yes, based on what’s presented in the CAS Reasoned Decision of the case, it’s difficult to argue for any decision other than the one CAS reached. Houlihan had a number of approaches to get that ADRV (Anti-Doping Rule Violation) set aside, but pretty much every single one she tried fell short. In fact, some of the conclusions made by CAS are pretty strong, they use quite definitive language that portrays in pretty clear terms that they feel that her claims are implausible, whereas World Athletics’ evidence and experts are credible, often unchallenged, and much more persuasive. Houlihan’s very first claim is that the lab didn’t follow its own instructions when issuing the Adverse Analytical Finding. CAS ends up making a point that not only did they follow it, but they were obliged to issue the AAF as a result of her sample results. It’s very strong. In that particular instance, Houlihan makes a claim based on one section of the WADA Technical Document when actually, another section said something entirely different, and it ends up being very bad for Houlihan’s case. 

We also see language like “highly improbable” when referring to some of Houlihan’s claims, compared to “convincing” applied to the experts for World Athletics. By the time you read how CAS assesses Houlihan’s polygraph evidence (hint: it’s not complimentary at all), it’s difficult to argue against that final decision, and the one-sidedness of the whole proceeding. Houlihan gets one or two very small “wins,” irrelevant in the larger scheme of things, but on every substantive issue, falls well short of the “balance of probabilities” standard required to prove her innocence.

So yes, based on what we see in the Reasoned Decision, it’s a pretty clear case. There are some things NOT in the Decision, where I was left wondering whether Houlihan simply couldn’t, or maybe didn’t, offer evidence in support of her claims, or disprove and rebut arguments made by World Athletics. But one can’t assess a decision based on hypotheticals of what might have been overlooked or not done.  

Embed from Getty Images

Q: Just because WADA’s rules were followed and Shelby should be banned, that does not mean she was intentionally doping, but does not rule it out, right? Having analyzed the CAS decision in such detail, do you have a strong working theory on what happened? 

Right. And in this regard, it’s instructive to pay attention to the language used in the CAS Decision. Once the Anti-Doping Rule Violation is determined, the burden of proof moves onto the athlete, who must now provide explanations for how that banned substance got into their body. And we must remember that standard of proof for the athlete is “balance of probability,” which means Houlihan can’t just say “I promise it was an accident.” She has to provide explanations that are “more probable than their non-occurrence.” This is one area where the CAS Decision makes it quite easy to follow the process. Even though it wouldn’t have been so “linear” in the actual hearing, the way the Decision is laid out reads more like a story (with some admittedly technical bits in between), but it’s easy to get a feel for how CAS rules on each of Houlihan’s arguments.

And when you look at the language used by CAS as it describes its findings for each of Houlihan’s arguments, they use phrases like “It is possible but improbable” (points 106 and 114), and “possible but highly improbable” (point 109). That clearly means that one can’t commit 100% to whether any given event occurred. Jokingly, it makes me think of that line from Dumb and Dumber, where Lloyd (Jim Carrey) says to Mary (Lauren Holly), “So you’re saying there’s a chance.” There’s always a chance, but CAS is so clear about how improbable that chance is that the decision looks really firm. Of course, there’ll always be people who disagree, and who believe in the “improbable” component of those phrases. It’s a strange tension, perhaps, but one that exists in these cases because only one (or maybe a few) people truly know if it was deliberate doping or not.

As for what happened, that is indeed the question. I genuinely don’t know. The contaminated supplement defense is easily the most common one used by athletes after a failed test. I’ve been involved in a handful in the last few years, where lawyers have asked for insights, the athlete has tested the supplement, and sometimes, shown that it is the source of the banned substance. So we can say with confidence that it IS POSSIBLE for supplements to cause doping violations. In fact, I’d go so far as to say that if an athlete is taking supplements without taking every precaution to avoid it, they may well be doping and shouldn’t get too much sympathy for it — we all know how poorly-regulated supplements are. So I do wonder whether Houlihan’s first thought upon notification of the positive test was to look at her supplements?  

That’s not enough, though, and she’d have needed to test her supplements and actually find the banned substance in them, and also show that she’d taken the necessary precautions to avoid inadvertent doping. Did that happen? Did they test, and not find anything? Had they already thrown out the supplements in use at the time, making it impossible to test them? We don’t know, but presumably the answer was no, because if it were “yes” they’d definitely have played that card rather than contaminated pork, which ends up creating for Houlihan a real plausibility problem (the 1 in 10,000 chance headline that the case gave us).

(Editor’s note: Houlihan says she sent her supplements and vitamins to a lab for testing in January 2021 but none of them came back positive for nandrolone).

So if Houlihan can’t show supplements, and couldn’t show pork ingestion as credible, what’s left? Well, doping. And there’s a really interesting section in the CAS Decision where we get a hint of what World Athletics might believe happened. Of course, World Athletics don’t have to prove anything — it’s Houlihan who has to show a plausible non-doping explanation – but Prof Christiane Ayotte, who testifies for World Athletics in the case, says, in Points 76 and 114 that in recent years, they have begun noticing a new pattern of carbon isotope signatures in these 19-NA doping cases. She says that since 2018, 31 conclusive Adverse Analytical Findings for 19-NA belong to one of two distinct groups or types. One batch has an isotope signature around -29‰, while the other is clustered around -23‰. 

Presumably, the -29‰ is injected nandrolone, but the -23‰ belongs to what Ayotte describes as oral precursors of nandrolone.  She even names two – 19-nor DHEA and nor-Andro, says they can be purchased on Amazon, and says that she has tested such a product and found that its isotopic signature was -23.8‰. Given that Houlihan’s 19-NA was measured at -23‰, and thus very similar to these precursors but very different from what would be expected from pork, this is as close as Ayotte comes to offering what they believe to be the doping act in the Houlihan case. But of course, they never have to explain the origins of the 19-NA – that burden is on Houlihan.

As for me, I really don’t know, but that would seem to be the likeliest explanation based on the evidence. If precursors have the same “fingerprint” as the 19-NA in Houlihan’s urine sample, that feels significant. Of course, one could argue that these precursors were present as contaminants in banned supplements, bringing that defence back around. But Houlihan couldn’t (or didn’t) prove the supplement argument, and couldn’t disprove the doping one with the pork burrito defence, so the precursor use remains the likeliest one.  

It would be fascinating to explore Ayotte’s findings more, and know details of those 31 cases, and also have a wider spread of testing of products to establish how likely it is that these oral precursors have isotope signatures so similar to what was measured for Houlihans’ 19-NA.

Q: Do you ever read CAS decisions and think “100% that person was doping?” Is this one of those cases?

No, never 100%. The system is not built on that kind of certainty. It’s pretty much always circumspect. The moment the athlete offers a non-doping explanation, CAS assesses their explanation through that standard of proof lens “balance of probabilities.” So you’ll struggle to find a decision where they say “Athlete X certainly doped.”

Though I will say that when you read the CAS Panel findings in this case, especially from Point 102 onwards, that they must have been pretty close to sure that her failed test is not the result of the contaminated burrito. As I said above, the language in the Decisions is always circumspect – “possible, but improbable,” and, in this case, “possible but highly improbable.” Position determines perspective, though, so some will look at that and say “There’s room for innocence in the word possible.” Others will say “highly improbable” means her explanation is not believable.

Could Shelby be Innocent?

Q: Is there any chance Shelby is an innocent victim in this?

Embed from Getty Images

Yes, but referring again to the above and Lloyd’s “So you’re saying there’s a chance” line. But of course it’s always possible – no “justice system” (if we can refer to anti-doping as such) is going to be absolutely perfect. The system is dependent on the inputs it receives from two sides of every issue, and then it makes decisions against a standard of proof (balance of probabilities in this case). So you can envisage a situation, let’s call it for Athlete X, who inadvertently ingests a banned substance either from a supplement, or in food, and then cannot find the necessary proof to present to a panel. Maybe the supplement was contaminated but discarded by the time the positive test was communicated to them. We know that the former happens (athletes use tainted supplements), and so thus the latter will happen. In that situation, you’ll have an innocent athlete found guilty of doping because they can’t meet the requirements to exonerate themselves.

That may seem harsh, but it’s necessary for the integrity of sport. If such a process didn’t exist, every doper would blame contamination, and the system would be impotent and powerless to sanction them. So it makes sense that the burden falls upon the athlete, provided the testing procedure and its “rules” are adhered to. Once that happened (and for Houlihan’s case, it was a crucial step, challenging the lab procedure), it’s right that the athlete must accept the burden.

But long answer short, of course it’s possible, but based on what was presented to CAS (and possibly, what was NOT presented, since it may be quite significant, as I explain in the longer document, and below), it’s difficult to argue against the CAS decision here. 

Q: One of Houlihan’s biggest arguments was that the lab should have conducted a pharmacokinetics study because the initial testing could not definitively determine the source of the nandrolone in her system. Lab director Christiane Ayotte argued that the pharmacokinetics study was unnecessary, and the CAS panel agreed with her. But is there any reason for them not to conduct this study? Could Houlihan hire a lab to conduct it herself? What is the downside in having more data?

Yeah, this is a crucial part of the case, and arguably the most complex, which is why my explanation ran so lengthy! Basically, Houlihan tries to get the AAF set aside, called instead an Atypical Finding, which would make the problem go away. She does this by saying that the laboratory didn’t follow its own rules, and should have done this pharmacokinetic test on her.  This is one of the arguments she makes that CAS dismisses emphatically. In fact, CAS goes to the length of actually making a point in the Decision that not only did the lab act appropriately by not doing this pharmacokinetic testing, but that they were compelled to issue the Adverse Finding without needing it. It was very strongly worded, quite definitive. The Technical Document in question does say that the pharmacokinetic testing “may” be conducted in some cases of ambiguous or unclear carbon isotope results, with low concentration of 19-NA. That’s the hook that Houlihan hangs her initial argument on.

However, she loses it because the analysis of her sample does not qualify it as such an ambiguous case. Rather, her finding is deemed to be quite clear — the isotope ratio of the 19-NA is different enough from that of her other steroids and steroids of boars that the lab must issue the Adverse Finding without even needing the pharmacokinetic test.

The reason it’s set up in this way is partly to avoid doing unnecessary tests that will in all likelihood not shed more light on the issue. In this instance, a pharmacokinetic test would require that Houlihan eat the same burrito, in the same conditions, with the same preceding training and daily routine, and then test her urine regularly in an attempt to recreate the positive result and show boar meat is a plausible origin for the 19-NA. Getting that recreation right is very difficult, if not impossible. It has time and financial implications, for what is probably, in this instance, a really low likelihood of an answer either way. So you can appreciate why the authorities would not want to do it unless it was absolutely necessary. In their view, it only becomes necessary when the sample analysis is unclear, and it wasn’t for Houlihan.

Could Houlihan have done it herself? Yes, she could have entered it as evidence along with hair analysis, the polygraph, her experts’ statements.  Remember that with a pharmacokinetic test, the athlete is trying to recreate the positive doping result, in an entirely new sample as part of a controlled process. Their urine contained a banned substance (19-NA), and they have offered an explanation for it (“I ate a pork burrito”). But the sample they provided cannot, by itself, confirm or refute that explanation, because all that sample represents is an “outcome”. What is missing is the “input.”  So what they do with this test is a “repeat experiment,” but this time, they are very deliberate and controlled, and they try to repeat the behaviours in the hope that a new sample will once again repeat the outcome.  If that happens, they can go to the courts and say “Look, I did nothing wrong, all I did was eat that pork burrito, and I failed a doping test again, and so therefore, I should be given the benefit of the doubt for my original sample too.”

Again, though, it’s a matter of practical barriers — can you recreate the result? I suspect it would fail at the first hurdle. How do you guarantee you’ve ingested a pork burrito that contains nandrolone in the same amounts as what Houlihan claims, given that it’s rarely in the food chain in the first place? If you can’t do that, what’s the point of the testing? So for both Houlihan and WADA/World Athletics, committing to do a test that has a low probability of definitive answer is an unnecessary step.  

We have of course seen athletes do these tests, most notably for asthma medication positives in cyclists. Sometimes they succeed, sometimes they fail, but at least with those drugs, you know the source of the drug is right there in your inhaler, and you’re testing out how different dosages appear in the urine under different exercise conditions. How much more difficult would this be for a “speculative source” like pork meat, given all we learned about agriculture in the CAS decision? For Houlihan to try this, and fail (which feels likely) would invite enormous downside. I’d be interested to know if they tried, and didn’t present it to CAS since it didn’t support the explanation.

Q: Ayotte explains that the concentration of nandrolone in Houlihan’s sample – 5.2 ng/mL in her A sample, 5.8 ng/mL in her B – was too high to have come from ingesting boar meat, citing a study in which none of the subjects had a concentration above 2.4 ng/mL. But the subjects in that study ate boar meat. In the same study, the levels from those eating other parts of the boar measured significantly higher, as much as 130 ng/mL for a 37-year-old female who consumed a kidney/heart/liver mix. Why did Ayotte (and, seemingly, Houlihan’s team) totally ignore this? Isn’t it possible Houlihan could have consumed an organ like that and produced a much higher level?

Embed from Getty Images

This is a very good question, an important one, and actually instructive because it reminds us that what we read in the CAS Decision, while lengthy, is still a summary of what actually went down throughout the hearing process. The process for the CAS hearing will have involved thousands of pages of submissions before the actual hearing happened, plus lengthy deliberations and debates among the experts in what CAS calls “Expert Hot Tubs.”  What we read is a summary, and that means we don’t really know if some of these finer, more technical points were debated or discussed in either the written submissions or the “in person” (virtual in this case) hearing.

However, based on what we can glean from CAS, we can reasonably answer this question.  There are a couple of places in the CAS report that mention the level of nandrolone that would be expected from different parts of the pig/boar.  For instance, in Point 107, the agriculture expert for World Athletics explains that pig stomach has one of the lowest androgen levels of any organ. Importantly, Houlihan’s claim is that she ate pig stomach (see points 40 and 99). McGlone explains that the only part of the stomach that is sold to the food industry is the outer stomach, which is metabolically inactive and has a very low steroid concentration compared to the whole stomach. This comes up again in point 114, which is the CAS panel finding and which says that McGlone’s evidence is not disputed, and that pork stomach does contain one of the lowest androgen levels of any organ.  

This is one of a few areas where expert evidence is unchallenged by Houlihan, and which really hurts her case. What I did find strange, given what you mention from that study, is that Houlihan would not at least have tried to cast some doubt on the possibility that eating other parts of a pig might cause high levels. What Houlihan’s team does is question whether the nandrolone levels in the stomach are truly known to be low? If you read Point 108, you’ll see that their contention is that there is no study available on 19-norsteroid concentrations in pig stomach. This doesn’t carry much sway with CAS, who still describe the World Athletics evidence in this regard as “in essence uncontested.”

That said, the high levels in the Guay paper were from kidney, liver and heart. The kidney in particular is what triggers the high values. Did Houlihan paint her defence into a corner by claiming stomach? It certainly allowed McGlone to make that case about the outer lining of the stomach. Had she claimed kidney, liver or heart, would her defence have been accepted, or at least stronger? I have to assume that they looked closely at this study, and saw the same thing you did. But they appear not to have played this card anywhere — I couldn’t see an occasion where Houlihan argued that it’s possible for kidney to make its way into a pork burrito. All we know, from Point 99, is that the food truck truck’s pork burritos were either chorizo (pork sausage) or buche/maw (stomach). But perhaps that information just didn’t make its way into the Reasoned Decision.

Q: Another question about the study Ayotte cited (which she herself helped conduct in 2006): she said that none of the subjects who consumed boar meat had a nandrolone concentration of more than 2.4 ng/mL in their urine. But only three subjects in the study ate 300g of boar meat and there aren’t many other studies on this topic. In the sports science community, is a three-person study enough to accept the finding as accurate? Or could/should Houlihan’s team have argued this is far too small a sample size from which to draw a meaningful conclusion?

Any study will always benefit from more participants and data points, but to be fair, that Guay study used 18 volunteers who ate a total of 150 meals. The explanation in the paper is a bit confusing, but my read of it is that 18 volunteers actually ate numerous meals of various combinations of edible parts, including 300g of meat, while only three of the 18 also ate the smaller 100g meat portion.  They’ll argue that they were “loading the dice” by doing it this way – in fact, Ayotte does exactly this in Point 111, when she explains that they used old boars to maximise the level of androgens, that they ate very large portions, and that most of them ate this 300g portion.  The rationale would be that if the 300g portions that were eaten multiple times by 18 volunteers didn’t trigger a positive test unless it had kidney in it, then a 100g portion would be highly unlikely to trigger a positive test. They’ll say why bother with a lower dose in many people when the higher dose didn’t do it anyway? Which I can understand. Then there is a second study, out of Germany, also cited at CAS, that used 12 subjects who ate between 187g and 491g of boar meat, and the highest concentration of 19-NA measured was 2.1 ng/ml (Houlihan’s values were 5.2 and 5.8 ng/ml). Based on the Decision, CAS found these two studies IN CONJUNCTION with the isotope signature (this is an important point — the level and type both argued against Houlihan’s claim) to be convincing.

Of course, there’s always an argument for more testing – they should test a thousand people, because maybe abnormal metabolic outcomes happen 1 in a 1000 times. But even more than this, what would probably be more important than testing more people would be to test more meat sources. After all, the accidental doping defence they’re trying to avoid is one where contaminated meat slips into the food supply, so 1000 people eating meat from say, five sources, would seem to me less important than letting five people eat meat from a 1000 sources! But there are obviously practical limitations to doing that. This is a permanent challenge in anti-doping.

Christiane Ayotte and Her Inaccurate Testimony in Jarrion Lawson Case

Q: Speaking of Ayotte, she played a key role in the Jarrion Lawson case a few years ago. Lawson tested positive for the banned substance trenbolone, and in Lawson’s AIU hearing, Ayotte stated that all positive trenbolone tests in her lab showed low levels, making it impossible to separate intentional cheaters and contamination cases. Yet evidence from her own lab eventually showed that this statement was false, and that was one of the main reasons CAS wound up overturning Lawson’s suspension.

Embed from Getty Images

Ayotte was not only the director of the lab that analyzed Houlihan’s sample, but she was also AIU’s expert witness in Houlihan’s CAS appeal. Does this strike you as problematic? Should she still be allowed to testify in a case like this? How big of a mistake did Ayotte make in the Lawson case? 

The authorities definitely should be held to a very high standard in these cases.  Given that the moment you have an ADRV, the burden of proof is on the athlete, it is only right to ensure that there is maximum confidence in why the ADRV exists in the first place. In the Houlihan case, this is literally in play right at the start, when Houlihan claims that the laboratory didn’t follow its own instructions as laid out in a Technical Document. That Technical Document of course serves to ensure high testing quality. You lose that, your entire system will crash. And there are instances where one would look at the subjectivity in the scientific process and wonder whether it’s fair for athletes. For instance, Steven Colvert’s EPO suspension was highly contentious.

Now, I don’t know the background of the Lawson case nearly as well as you do — I only read your summary, but it sounds as though the best case one can make is that Ayotte was caught making an inaccurate generalization with significant implications for the initial sanction.  Apparently the consequences of this error affected her minimally, and so she remains influential in the anti-doping world, and certainly going back as far as 2006, she’s a world leader in 19-NA doping cases, having published a number of papers on it, probably having been involved in the development of the technical protocols. In that regard, her involvement is understandable — the system must use its best and most qualified people. 

I guess the problem is that we have an inherent trust of those people and the system they serve, even if we recognize that even in analytical science, not every single thing is clear cut, black and white. Reasonable experts might well disagree on the interpretation of a finding. That seems to have happened in Houlihan’s case, at least with respect to the concentration of 19-NA in the sample, where Houlihan’s claim is that it’s too low to be deliberate doping, and World Athletics (through Ayotte) are saying it is clearly high enough.

All of this does create a potential for mistrust, which means the authorities need to be absolutely, 100% sure that they’re watertight in every possible respect. When that doesn’t happen, well, yes, there are going to be problems. All of which is a lengthy way of saying yes, this is potentially problematic.  

That said, if the merits of all the expert arguments are assessed, and the studies Ayotte presents are evaluated in the context of what World Athletics are claiming and what Houlihan is explaining, then it’s difficult to see that Ayotte’s testimony here is decisive. In the Lawson case, was her statement re: the difference between doping and contamination decisive to the initial verdict? Perhaps it was, but I don’t believe it is as decisive here, where I think Prof McGlone, the agriculture expert, does far more damage to the plausibility of Houlihan’s case than the actual analytics of her sample.  

As for the analytical results, which is what Ayotte testifies, CAS is not being asked to trust her opinion, but rather the procedures and findings on the sample, and her interpretation thereof. This is significant because both the concentration and the isotope signature of the 19-NA in Houlihan’s sample refute the pork burrito defence. So Ayotte is more the “messenger” of already published work in this case, and CAS believe her to be a credible messenger with strong supporting evidence.

Q: You noted multiple times that the AIU’s expert, Prof. John McGlone, essentially argued that mistakes are very rare in the US pork supply chain, which means the chances of the nandrolone in Houlihan’s sample coming from contaminated pork were “near zero.” Are you surprised Houlihan’s team didn’t push back on this or find an expert to discuss potential disruptions to the US pork supply chain, particularly during the COVID year of 2020?

Yes, very. In fact, this is the main area where I’ve found myself wondering what hasn’t made it into the CAS Decision. To begin with, did Houlihan use the pork burrito explanation because she had to offer something in her defence, and supplements weren’t panning out as a possible explanation? And if so, how much did she do, with her legal team, to find experts who could support the argument that edible pork parts with high 19-NA potential can make their way into the food supply? Remember that for her explanation to be accepted, she has to convince the CAS panel that it is plausible to ingest large enough amounts of nandrolone in meat. We mentioned above that kidneys and hearts have been shown to have the highest nandrolone levels, so I wonder whether Houlihan had explored the idea that these organs could contaminate the pork meat?  

The COVID disruptions get a passing mention, but no more. The carbon isotope ratio of boar meat in the USA is debated, which involves discussion of the typical diets of pigs (since these determined the isotope signature of the edible parts), and COVID induced disruptions to that diet, but only McGlone is heard on all these issues.  

McGlone also describes the agricultural practices, and the checks and balances that keep uncastrated boars out of the food supply chain, without any rebuttal.  McGlone gives CAS every possible reason to dismiss Houlihan’s explanation. His testimony really does significant damage to her case, and so it’s really striking in the CAS Decision how Prof McGlone’s testimony is uncontested, not in dispute, and unchallenged (CAS use variations on this theme throughout).  It makes me wonder whether what we’re not seeing is more significant than what we are? Houlihan must surely have looked into poor agricultural practices, human error, systematic failures etc, that might strengthen her position, but none of it makes it into CAS’ decision. If she could find even one farm or supply chain that disproves McGlone’s confidence in the system, surely she would have played it? But she doesn’t, and in fact, she doesn’t even have an expert from the field to rebut McGlone’s testimony. Does this mean that rebuttals simply did not exist?  That’s not necessarily a criticism of her approach, by the way – it’s possible that there’s just no counter-argument to be made. Is the system in the USA so robust that McGlone is unquestionably right about the “near zero” chance of errors causing a doping positive? 

Given Houlihan’s inability to find experts to even try to refute pretty much all of his claims (let alone do so successfully), we have to conclude that it is, that he is correct, and that his testimony is true. He was the star witness for World Athletics, because his testimony revealed that the sequence of events required by Houlihan were so unlikely as to be next to impossible, and he did it without any significant opposition.

Q: Hundreds of athletes throughout the world, including many Kenyan distance runners, have been banned for nandrolone positives, so nandrolone positives actually look to be pretty common in distance running. Do you feel comfortable with an athlete like Shelby being banned for four years for the amount of nandrolone found in her system with the isotope profile it had?

Embed from Getty Images

In terms of whether nandrolone is the best drug to use for performance, of course you could list half a dozen that would be better, because of detection time, effect etc. Yet we know it is used, and perhaps that is because of availability or ease of acquisition, or it may be part of a collection of substances used, some legal, some not. I abandoned any notion of interpreting athletes’ actions against some logical framework long ago!  

However, what I would love to know about those other nandrolone positives is whether their 19-NA looked the same as Houlihan’s? As I mentioned above, the CAS decision gives us a really interesting insight into what World Athletics believe is happening, when Ayotte explains how a lot of positive 19-NA cases have an isotope signature around -23‰, which is very different from what would be expected from pork meat, but consistent with oral precursors that we know are sold as recovery and performance aids (see points 76 and 114 of the Decision). So they’re quite widely available, often sold as “legal” (many people would be persuaded that a precursor is legal, when it may not be), and according to Ayotte, popping up regularly in doping controls.

So I wouldn’t dismiss as implausible the likelihood of an athlete using these substances without knowing the risk, or perhaps unknowingly, because the number of cases alone tells us that it’s not all that implausible!

Q: If someone has too high a level of nandrolone in their system it seems like the possibilities are 1) intentionally doping; 2) food contamination; or 3) supplement contamination. Is that correct? From your understanding of the science is there any way it is food contamination? 

Yes, that’s pretty much it, though as we discover in the CAS decision, the first option there, “intentional doping”, can take a few different forms. It might be injected, which is probably very unlikely because of the very high levels and long detection times. Or it might be ingested, and even then, it might be ingested as an oral precursor, as Prof Ayotte testifies they believe is happening relatively regularly around the world.

As for whether it could be food contamination, the answer must of course be yes, in theory – even the studies they cite at CAS show this potential, because in those studies, some people did produce detectable levels of 19-NA after eating boar meat. Admittedly, very low levels, and that was after very large meals, but then kidney and heart were found to cause much higher levels. So the existence of a Technical Document that acknowledges the risk is the result of that recognition. Perhaps this is what led Houlihan to make this claim in her defence.  

That said, and fully disclosing that I know literally nothing about US agriculture practice other than what the CAS Decision reports was said by Prof John McGlone, it seems to be very unlikely for a person in Portland to eat a pork burrito that contains sufficient nandrolone to cause relatively high 19-NA levels in the urine. Unless there’s something “wrong” with McGlone’s testimony, some way his account of the checks and balances can be disproven, you have to conclude that no, contamination can’t explain the presence of 19-NA with the specific isotope signature and the level of 19-NA in this particular case.

Q: Big picture is CAS has ruled against Houlihan, she’s banned for four years, and we don’t see any gross procedural errors in the case to get it overturned by the Swiss courts, which is where she is currently appealing the decision. If Houlihan wants to compete in the next four years, she almost needs to put on a public service campaign and show the science in the decision was flawed. Did anything stand out as flawed?

Not in the analytical results, no. The concentration in her urine wasn’t ridiculously high, but it needn’t be (there’s a published paper that explains how the level of 19-NA is not a good way to evaluate doping vs contamination). It was almost three times higher than the highest levels measured in those two studies where people ate much larger pork meals. More tellingly, the isotope signature didn’t meet the requirements for an “ambiguous case” that might require further testing. The WADA Technical Document sets up the criteria to rule on a sample being an Adverse Finding, and Houlihan’s meets that, pretty clearly. Her 19-NA looks quite different from her other endogenous (self-produced) steroids and also from the 19-NA that would be expected from boar parts. If it was only the concentration, or only the isotope signature, I think one could probe for a weakness. But taken together, that science looks pretty watertight.

But that science alone is not what drives the CAS Decision. There is also the evaluation of the sequence of events required to ingest nandrolone in the first place, which, you have to say, is incredibly unlikely, from start to finish (and that it should happen to an elite athlete the day before a test). That’s really the only place where I found myself thinking “I wonder whether this is definitely true?” Prof McGlone testifies about the supply chain that gets pork into human consumption, and he portrays it as a really robust system that almost never fails. I found myself wondering how true that is — it felt a bit like reading claims that businesses have checks and balances to prevent financial fraud, and thus no fraud ever occurs. We know that human error and deliberate acts can bypass checks and balances, and so I wondered whether that might be a thread to pull on? It’s not really flawed science, but flawed process. As I mentioned above, though, Houlihan doesn’t even have an expert to rebut this testimony, and that may be very telling.

Q: A four-year ban normally means an athlete misses one Olympic cycle. Because of the Tokyo 2020 delay, Houlihan is scheduled to miss two Olympics for a first-time drug ban. Do you think the sentence should be reduced to 3 years so that she can compete at Paris 2024?

Embed from Getty Images

I think the general principle of 4 years is good – remember it used to be 2, and that was thought to be an insufficient deterrent, precisely because it wouldn’t cost the athlete enough. As you note, though, the change in the global sporting calendars in 2020 and 2021 means Houlihan’s sanction is now disproportionately large, at least in terms of Olympic Games. That said, the only time they reduce sanctions is when the athlete can show a reasonable case for inadvertent doping (if not innocence), and so the fact that they kept it to four years would indicate just how strongly they believe Houlihan’s was a doping offence, or put differently, how far short she fell of persuading them otherwise. I’d be interested to know whether she asked for a one-year reduction in order to not lose two Olympic Games. Do we know whether they tried?

In any event, as a broad principle, I wouldn’t necessarily mind if the global anti-doping system did reduce sanctions by a few months in 2024 only, but it would need to be global and without exception, justified on the basis of not adding an extra Olympics to a ban cycle. I also wouldn’t take a full year off the ban – it should be as short a reduction as possible to allow an athlete to miss one Olympic Games.  But it would have to be for all,  not for Houlihan only. 

If Shelby is Innocent What Should She Do?

Q: If Shelby Houlihan came to you and said I want to prove my innocence, what would you suggest she do?

I’d hire a few private investigators to go into the commercial pig farming world in the USA and try to find places and cases where the confidence in its checks and balances can be eroded. I’d be looking to identify “loopholes,” ways in which boars with high nandrolone levels (uncastrated or cryptorchid species) do in fact find their way into the food supply, where the parts with the highest nandrolone levels can contaminate pork meat, and where the feeding practices of boars could change the isotope signature of the 19-NA that comes from boar meat.

Embed from Getty Images

If I could find this (and this may well be a big “if”), then I’d try to trace that pork meat to its commercial destinations, and I’d test those foods to see if they contain nandrolone. If I can find nandrolone in pork that is available to the public, I’d put that meat through laboratory testing to see if I could detect any nandrolone or its precursors. I’d be hoping to find nandrolone in say, 5% of all meat that can be consumed by the public, because then I could say to the courts that it does in fact happen, despite the assurances from Prof McGlone that it can’t.  

If I’ve managed to find pork meat that contains nandrolone, I’d have a bunch of people similar to me (athletic, small size) eat regular meals from these places, whether they’re restaurants, grocery stores, or food trucks like the one I’m claiming I ingested it from. Then I’d test their urine to see whether they produce urine samples similar to mine. That is, what’s the level of 19-NA in their urine? What is the carbon isotope signature of that 19-NA? If they end up with levels between 4 and 10 ng/ml, and with a carbon isotope signature of -23 like I did, then I’d present that as evidence and say, “look, 5% of all meals from these places produced a positive result that is very, very similar to mine.”

The reason I’d do this is to try to undermine the confidence of the anti-doping and agriculture system. Of course, this could backfire spectacularly. It’s quite possible, even likely if one reads and trusts McGlone’s testimony to CAS, that Houlihan could have an army of people eat every single pork meal in the Northwest of the USA for six months and not a single one would produce a positive result like hers! It might just be that in the USA, with their agricultural practices, it is impossible to trigger a positive test. In fact, Houlihan may well have tried all of the above in her defence already. Maybe they did buy up a whole bunch of pork products from that food truck and other places and try to find evidence for nandrolone contamination in the city of Portland (let alone one truck). We don’t know what was tried, but couldn’t produce the results they needed it to.

The other thing I’d do in parallel is try to verify Ayotte’s claims about the oral precursor having an isotope signature similar to mine. I’d give these oral precursor products to people who volunteer to help, in a controlled setting, and see if I can recreate the positive results with precursors. What I’d be hoping for is a failure to recreate my positive test – maybe I show that oral precursors cause much higher levels of 19-NA, or that their isotope signature is very different from my value of -23. This would help by punching a hole in the idea of ingestion of doping products. The two together might help me strengthen my case.

I can’t stress enough how much Houlihan would need to show. She’d need to show that it’s plausible for high enough levels to find their way into meat, which means uncastrated boards or cryptorchid species by way of a failure of normal agricultural processes. She’d have to find a way to explain why the carbon isotope ratio of the 19-NA from these boars is different from normal. She’d have to account for the level in her urine. There are a lot of dominoes to fall, but that, theoretically, is what I’d try to do. It’s a kind of mass pharmacokinetic experiment, in many people, but only after some private investigator work to establish plausibility of the origins.

Q: Do you have any idea on how much all that would cost?

I imagine it would be incredibly expensive.  I mean, I’m suggesting hiring of people to investigate the system and try to ‘dig up some dirt’ on the agricultural practices to undermine McGlone’s evidence.  I have no idea how much it would cost, but I guess you’d be hiring a couple of private investigators for a pretty significant undertaking.  Not cheap.  The lab tests will run into the thousands of dollars, because you might realistically need to test hundreds of pork samples from around Portland for steroid contamination.  Imagine that’s $200 a shot (I have no idea of these costs, by the way).  I could see that being mid five figure, just for lab testing.  Then conducting those multiple pharmacokinetic tests would probably run into the low six figures. Maybe high five.  I wouldn’t be surprised if it ends up in the mid- six figure range.  Again, though, it’s just so infeasible that pork would even contain nandrolone, it’s the proverbial needle in a haystack.  So unless you could find evidence of the nandrolone, all the other steps that involve volunteers eating this stuff and being tested would be a complete waste of time and money.

It would be really interesting to know how far they explored this, if at all.

Final Verdict

Q: You’ve spent the last month studying this case in-depth. Gut feeling: is Houlihan innocent or guilty?

With confidence, I’d say “not innocent on the basis of pork burrito ingestion.” Which of course, in doping cases, means guilty. There is much unknowable in between, but the contaminated food explanation doesn’t stand up to basically any level of scrutiny.

More Ross Tucker Analysis on this Case: Piece #1: An In-Depth Analysis of The Shelby Houlihan Doping Case by Sports Scientist Ross Tucker

Talk about this article and the Houlihan case on our world-famous messageboard / fan forum. MB: Sports Scientist Ross Tucker breaks down the Shelby Houlihan case for us.

Previous Houlihan Related News: