An In-Depth Analysis of The Shelby Houlihan Doping Case by Sports Scientist Ross Tucker

By Ross Tucker
April 21, 2022

LetsRun.com note: If you’ve followed professional track & field at all over the past year, you’re by now familiar with the Shelby Houlihan case. The broad strokes of it, at least. Houlihan tested positive for nandrolone, a banned steroid, in December 2020. She denied doping, claiming she had never heard of nandrolone and pointing to a tainted pork burrito ingested the night before the test as the source of the nandrolone. The Court of Arbitration for Sport did not buy this argument, and in June 2021, banned her from the sport for four years.

Three months later, CAS released its full Reasoned Decision on the Houlihan case, and if you read all 44 pages, you read both Houlihan and World Athletics try to make their case with scientific concepts such as GC/C/IRMS, delta-delta values, and carbon isotope signatures. Houlihan and her legal team were frustrated that the lab analyzing her sample did not conduct something known as a pharmacokinetics study; the lab director (and World Athletics) argued that such a test was not necessary. For non-scientists like ourselves it was quite confusing.

Article continues below player

We reached out to Ross Tucker, one of the preeminent sports scientists of the world (co-host of The Real Science of Sport Podcast), to make sense of the case for us. Ross has written two pieces on the case for LetsRun.com.

The piece below is long (over 5,000 words) and more technical. In it, Ross goes through the CAS decision step by step from beginning to end, clarifies some of the more esoteric aspects of the case and outlines in clear terms exactly what Houlihan argued and why he thinks CAS ultimately (correctly) ruled against her.

The second piece is where Ross gives his opinions on the case and answers the questions we had after reading this piece. If you’re a lay person or only have time to read one piece, we suggest you start with the second one. If you have lots of time and are really into science, definitely read this piece first and then go to the second one.

Houlihan’s opening argument

Houlihan when she broke the American 5000m record in 2020

Houlihan’s first point of attack is procedural.   In effect, she is saying to the lab: “You didn’t follow your own rules.  If you had, my sample would have been deemed inconclusive, and then flagged as atypical, not adverse”.  If that claim is won, everything else goes away, so it’s a pretty significant part of the defense, and worth considering in detail.

The important background to this claim is that WADA produces a number of Technical Documents (TDs) that guide how samples are processed, analysed and interpreted.  This is done to ensure a higher level of standardization between labs around the world. Initially, CAS gives Houlihan a small ‘win’, when they agree with her that the 2021 Technical Document should be used in the interpretation of her urine analytical result (this proves not to be too relevant once the details are discussed). 

Houlihan’s more important claim is this: according to the 2021 TD, the lab did not have enough information to issue an Adverse Analytical Finding (which triggers an automatic suspension) and instead should have conducted pharmacokinetic testing to establish the origin of the 19-NA in her sample (19-NA is short for 19-norandrosterone, which is the form in which nandrolone appears in urine after it is metabolized by the body).  Based on what they had, Houlihan argues, the lab should have issued an Atypical Finding instead.  Crucially, this does not trigger an automatic suspension.

In order to understand how CAS assessed this claim (and indeed, why it was made by Houlihan), we have to get into the weeds a little.  Let’s begin by looking at the section of the 2021 TD that Houlihan cites specifically in her defence.  I have highlighted what I believe to be the crucial sentence of the 2021 TD that Houlihan cites in her defence:

“Following consumption of the edible parts of non-castrated male pigs, concentrations of excreted 19-NA in urine are usually in the low ng/mL range (<10 ng/mL) although higher concentrations have been exceptionally reported.  The origin of the urinary 19-NA may not be established by GC/IRMS analysis, since the varying diets of migrating wild boars lead to dissimilar 13C values which may range between -15‰ and -25‰.  Therefore, if consumption of edible parts of intact pigs is invoked by an Athlete as the unlikely origin of a 19-NA finding, this may be established based on the pharmacokinetics of 19-NA excretion”

What this paragraph is saying is that when low 19-NA levels are detected in a urine sample (true for Houlihan), and when the athlete offers edible boar as an explanation (also true for Houlihan), there is a possibility that the lab method used (GC/IRMS) will produce an ambiguous finding, and so the lab can’t say with confidence that the 19-NA was the result of doping.  It could be a) naturally produced, or b) ingested in contaminated food.

At this point, it’s worth understanding exactly what the lab testing is doing, and how the GC/IRMS method works.  GC/IRMS is a test that measures the ratio of stable isotopes in a compound.  This is called the carbon isotope signature or fingerprint of the compound, and it is different depending on the source of the compound.  Compounds that are naturally occurring have a carbon isotope signature distinct from those that are injected or taken as pills.  GC/IRMS allows authorities to differentiate between possible sources of a drug, in this case 19-NA.  It is expressed as parts per thousand, or per mil, shown by this slightly modified percentage symbol: ‰.  

The GC-IRMS is important and necessary because we make steroids ourselves, and so they do exist in our bodies naturally (testosterone is the most obvious and well-known example, but nandrolone is another, albeit in very small amounts).  These are called endogenous steroids.

Boars also produce endogenous steroids, including nandrolone, which is why the pork burrito defense is even possible.  However, this creates a dilemma in anti-doping, because the carbon isotope signature of the nandrolone produced by boars falls into the same wide range as the carbon isotope signature of steroids that are made in the human body.  This range is mentioned in the TD above: -15‰ to -25‰.  

This means that GC/IRMS analysis of 19-NA may be unable to tell those sources (endogenous produced by the athlete, from endogenous produced by the boar) apart.  That’s really all the technical paragraph above is saying.

On the other hand, exogenous steroids are injected or ingested by dopers.  The challenge for anti-doping is that exogenous and endogenous steroids are chemically and pharmacologically identical – you can measure that they’re there, but if you rely on basic detection methods, you have no way of knowing whether that steroid originated outside or inside the body.  This is where GC/IRMS comes in, because this CAN tell them apart, because their carbon isotope signatures will be very different.  This is why most anti-doping labs now use isotope signatures (or fingerprinting, as it’s sometimes called), for final confirmation in adverse analytical findings.  

Embed from Getty Images

So the process is that the metabolite is first detected, and then possible sources or origins of that steroid are explored using the GC/IRMS test.  Basically, once 19-NA is detected in the sample, the lab must rule out two possibilities:

  1. Is this 19-NA endogenous?  That is, was it produced in the athlete’s body?
  2. Could this 19-NA have come from edible parts of pork?  That is, could nandrolone endogenous to a boar have been eaten to produce 19-NA in the athlete’s urine?

That technical paragraph from the 2021 TD says there are some situations – low 19-NA levels, and the edible boar explanation – where GC/IRMS may not be able to answer these two questions, because the isotope signature of boar meat may be similar to the isotope signature of human steroids.  The entire premise of GC/IRMS analysis is to compare one isotope signature to another, and if they’re very different, they can rule out endogenous sources, whereas if they’re quite similar, then exogenous sources (doping, basically) can be dismissed.  

Houlihan’s gambit is that her situation is one such, and so she should not have been issued with an AAF or an Anti-Doping Rule Violation (ADRV).

So how does CAS rule?

The initial win Houlihan obtained when CAS ruled that the 2021 Technical Document should be used in the interpretation of her urine analytical result would prove to be her only victory on this issue.  In this particular claim, CAS rules against Houlihan because of a section in the very same Technical Document she uses, which explicitly confirms that her finding should be treated as an AAF and ADRV.  That section, under the heading “Interpretation of GC/C/IRMS Results”, says the following (and again, bear with the technical nature of this, it’s important):

To reject the hypothesis of endogenous or in-situ 19-NA formation based on the application of GC/C/IRMS analysis (i.e. to report the finding as an AAF), the following criterion shall be met: The 13C values between two (2) ERCs and 19-NA, i.e. ∆δ13C= δ13CERC- δ13C19NA is greater than -3‰.

Embed from Getty Images

This is crucial, so it’s worth understanding exactly what the lab has done, and why.  Remember that the premise of GC/IRMS is to determine the carbon isotope signature of the compound (19-NA) and then to ask whether it’s different from the carbon isotope signature of other natural steroids and exogenous or synthetic versions of the steroid?  In order to do this, the lab will take other steroids in the athlete’s body, which are called Endogenous Reference Compounds (ERCs) and then compare these ERCs to the 19-NA in the sample.  

In plain English: “We take two compounds, one is the 19-NA and the other is a steroid  naturally found in the athlete’s body (the ERC), and we work out the difference between their carbon isotope signatures, which is called the delta-delta 13C value, shown by this symbol: ∆δ13C”.

That is what this part of the above section means:

∆δ13C= δ13CERC- δ13C19NA 

If that delta-delta 13C value is within -3‰, they’re deemed similar.  Obviously, Houlihan wants this to be the case, because that would mean that the 19-NA in her urine looks similar to her own endogenous steroids, so it rules out a doping origin and supports the idea of either endogenous production, or ingestion of boar meat, which would have a similar enough isotope ratio.  If, however, the delta-delta 13C value is greater than -3‰, then the 19-NA looks very different from her own steroids.  If this happens, the lab is instructed “to reject the hypothesis of endogenous or in-situ 19-NA formation”.

So which was it in Houlihan’s case?

The answer is revealed in point 76 of the decision, which says the following (again, I’ve highlighted in bold the key facts): 

The Panel notes that a GC/IRMS analysis was performed on the Athlete’s A- and B-Samples.  The Athlete’s A-Sample delta-delta values were 3.8 (i.e. -23 for 19-NA and -19.3 for pregnanediol) and the B-Sample delta-delta values were 3.8 (i.e. -23 for 19-NA and -19.2 for pregnanediol).  In brief, the Athlete’s delta-delta values were far out of range with that of human endogenous urinary steroids referred to in TD2021NA”.

So, what has happened here is that the lab HAS done the GC/IRMS test as instructed.  They have taken two other steroid hormones in Houlihan’s sample (pregnanediol, mentioned here, and androsterone, not mentioned here but in Point 73), and calculated the difference between the isotope signatures of those reference compounds and the 19-NA.  And that difference was -3.8‰.  

According to the Technical Document, a difference greater than -3‰ means they must reject the endogenous origin hypothesis, and issue the AAF (Section 4 of the TD).  So the lab was actually obliged to report the result as an AAF, because even though Houlihan’s 19-NA level was low, and even though she did raise the edible boar explanation, the delta-delta-13C was different enough that they could rule out endogenous formation.  CAS agrees – “The Athlete’s delta-delta values were far out of range with that of human endogenous urinary steroids.”

This ‘defeat’ then sets the stage for subsequent discussions.  Note that we haven’t yet fully explored the possibility that edible parts of boar might also cause a carbon signature similar to the 19-NA in Houlihan’s urine.  That dispute is still to be had, along with debates about the low concentration of 19-NA, the polygraph, the hair sample, and the likelihood of ingesting pork with high enough nandrolone levels.  

All that has been ruled on at this point is that the lab did correctly issue the AAF and subsequently, ADRV.  This is a significant ruling, because once an ADRV is issued, then the burden of proof to show that the ADRV was not intentional is given to the athlete.  This means that it is Houlihan who must convince CAS that her version of events is “more probable” than their non-occurrence (Point 92), which is the standard of proof.

And so begins the contaminated burrito defense.

Houlihan’s contaminated burrito defense

Having failed to have the ADRV effectively “downgraded”, Houlihan’s explanation for how the 19-NA ended up in her urine sample takes center stage.  By way of background, an athlete who is issued with an ADRV has to provide a reasonable and possible explanation for how the drug in question got into their body.  They do this either with the goal of being cleared entirely (if they can show no fault at all, and that they took reasonable steps to avoid an inadvertent doping violation), or being given a reduced sanction.  It’s not enough to make an appeal of innocence to the CAS panel – the athlete also has to provide a plausible alternative explanation for the source, one that CAS can assess against that standard of proof mentioned above.

Embed from Getty Images

Houlihan’s claim, which is known by most, is that she ate a pork burrito from a food truck the night before her urine sample was provided.  To support this, Houlihan must:

  1. Show that it could reasonably have occurred that she ate a pork burrito 
  2. Persuade the panel that the pork burrito could have had nandrolone levels high enough to cause her positive test
  3. Show that the concentration of 19-NA is what you’d expect after food ingestion, rather than oral ingestion or injections of 19-NA
  4. Convince the panel that the 19-NA in her sample is consistent with the nandrolone that would be expected to come from edible parts of a boar (we’re back to a GC/IRMS discussion here)

At the risk of simplification, Houlihan is saying the following:

The 19-NA in my urine couldn’t have come from doping.  The level is too low, and is more likely to be the case after accidental ingestion from food than from taking a drug on purpose.  Here’s my proof…

The way the CAS Reasoned Decision is written, it’s very easy to follow the sequence of things Houlihan has to show.  This is different from the previous section, where the CAS document is very difficult to follow and actually contains a few mixed up concepts and incorrect summaries of the various Technical Documents.

But here, we read a story, more than a technical puzzle.  The Decision is laid out in a linear way, even though the trial wouldn’t necessarily have proceeded this way, and so you can quite easily get a feel for how each of the above requirements from Houlihan’s side was a) responded to by World Athletics, and b) assessed by CAS.  We can look at each in turn.

a) It’s possible that I ate a pork burrito

This is not too difficult a claim to show is possible.  Houlihan does so by relying on her own testimony, supported by teammates, text messages, and even a written statement from the owner of the food truck.  Basically, she ordered a beef burrito, but is claiming that they were given pork instead.  There’s no way to confirm with absolute certainty that this happened, but CAS does rule that Houlihan ate a burrito (point 98), and that it is possible it was a pork burrito (point 101).  She clears the standard of proof here, and gains two small victories.

b) The pork burrito could have had nandrolone levels high enough to cause the positive test

This statement actually uncoils into a few different and sequential arguments.  Each is assessed by CAS, and they include:

  • The pork meat in that burrito had to be from an uncastrated boar.  This is necessary for Houlihan’s argument, because an uncastrated boar has higher levels of nandrolone, which is required for Houlihan’s argument, given the levels of 19-NA detected in her urine.
  • Because uncastrated adult boars are so rare in the food chain (according to the evidence heard), the only way this is possible is that the edible parts consumed by Houlihan would have had to be from a cryptorchid species.  This is because the removal of uncastrated boars from the food chain is the result of visual inspections along the food chain, but cryptorchid species have undescended testicles, and so they may pass through checkpoints undetected.  That is why CAS considers this possibility, as a kind of “what if” scenario to respond to the uncastrated boar likelihood.

The arguments around these issues presented in the CAS Reasoned Decision make for interesting reading.  They are relatively easy to follow, but difficult to critique, because the level of detail that was almost certainly offered by the various experts in their written submissions and probably in their testimonies and ‘hot tub’ sessions does not always make it into the final Decision.

The overwhelming picture, however, is one where World Athletics have found an expert, Prof John McGlone, professor of animal science at Texas Tech, to systematically go through a series of scenarios, those created or required by Houlihan’s explanation, and explain why each is highly unlikely.  Most tellingly, Houlihan’s team appears unable to mount significant challenges to almost all of this testimony.

First, McGlone explains that uncastrated boars very rarely make it into the food supply chain, because of the multiple checkpoints mentioned above.  Houlihan offers a report that describes that “thousands of boars with abnormally high levels of nandrolone are sold and slaughtered in the USA every day.”  However, World Athletics counters by arguing that this meat is used for inedible products and not for human consumption.  Prof McGlone provides documents in support of this, stating that boar meat constitutes only 0.33% of the US pork market, and that thanks to the multiple checkpoints for their removal, boar meat does not enter the food chain of ordinary pig meat.  CAS hears testimony that uncastrated boars have large testicles and so are easily spotted by inspectors at a number of points along this chain.  

McGlone next explains that cryptorchid species are rare (0.5%, or “extremely low, approaching zero”), and so the “what if” scenario that might explain how Houlihan ingested an uncastrated boar is also very unlikely.  McGlone’s final statement in this line of “plausibility of ingestion” is to point out that the plant where the Beaverton food truck obtains its pork meat does not operate a boar kill plant, which means there is a “near zero chance” that boar meat would be mixed up with normal pork products.  

The end result of all these low probabilities, according to McGlone, is a 1 in 10,000 chance that a human would consume meat from a cryptorchid, which made some headlines when the decision came out.  CAS agrees with this, pretty emphatically, too, concluding that it is “possible, but improbable” that uncastrated boar meat entered the pork supply and found its way to the burritos eaten at the food truck (point 106).  Most notably, they state that McGlone’s evidence on the food supply chain “remained basically uncontested.”  All Houlihan was able to show is that the pork stomach meat from the food truck in December 2020 came from a batch dating back to September 2020,  which may have been disrupted by COVID-19.  McGlone, however, argues that these disruptions would not have affected inspections, controls, and the slaughter of pigs in the normal food chain.  CAS were persuaded of this, and Houlihan’s necessary scenarios were dismissed as highly improbable (point 108).

This was of course a significant blow to her case, because if the process by which she consumed nandrolone in a pork burrito is improbable, then other further details related to the levels of 19-NA from that claimed ingestion would be very shaky indeed.  Nevertheless, CAS considers this also.

c) The concentration of nandrolone in the urine sample could have come from the pork burrito

Houlihan’s rationale here is that the concentration of 19-NA in her urine is low enough to suggest the ingestion of food rather than doping with either an injected or oral drug.  This claim depends on a few analytical prerequisites, some related to the likely levels of nandrolone in the ingested pork, others to the level that was detected in her urine.  CAS hears arguments related to:

  • The likely level of androgens in cryptorchid pork – is it plausible that they’re high enough to produce her results?
  • The research on food ingestion and whether Houlihan’s levels are consistent with it – does Houlihan’s sample look like that of someone who ingested nandrolone in pork?

With respect to the first question, Prof McGlone again delivers significant damage to the possibility of pork ingestion as the explanation for the 19-NA result.  He explains that cryptorchids have significantly lower androgen levels than fully adult boars, and only then in the muscle.  Even if uncastrated boars (like cryptorchid species) have elevated androgen levels, they’d only be found in the kidneys, testes and liver.  This is important because Houlihan’s claim is that she ate pork stomach, and McGlone then explains that the stomach has far lower androgen levels than other tissues, with reasons for this.

Embed from Getty Images

Once again, CAS notes here that this evidence is “in essence uncontested,” because even though Houlihan had two experts try to undermine McGlone’s contentions, they were “not very substantiated.”  CAS also reminds us here that the onus is on Houlihan to show that her explanation of the analytic results is more likely than not (the burden and standard of proof issue).  And since she did not, CAS once again finds that it is “highly improbable” that normal pork products in the US food supply chain would show elevated androgen levels.

With that mind, the scientific discussion about the expected levels of 19-NA following food ingestion is quite predictable.  Houlihan’s argument is that her 19-NA levels were low – 6.9 and 7.8 ng/mL before adjusting for specific gravity, which is a process where the lab corrects for the hydration status of the athlete (given that dehydration from exercise or dilution with diuretics could artificially elevate or decrease the the concentration of the substance).  Her anti-doping expert, Prof Strahm, argues that the 19-NA levels in her urine are consistent with food ingestion, because when taken orally or by injection, the values are way higher.  Prof Strahm uses the levels of another metabolite, 19-NE, to support this idea, saying that the ratio of 19-NA to 19-NE is too low to indicate doping, as well as the fact that tests either side of the positive 19-NA sample were negative, which would suggest that Houlihan was not doping.

For World Athletics, the expert in question is Prof Christiane Ayotte, who cites two studies (one is from her own lab) where the levels of 19-NA in the urine of people who eat pork meals are never as high as what was measured in Houlihan’s sample.  The main study cited and explained in the CAS decision is one in which 19 volunteers were given various sized pork meals, containing various mixtures of meat, kidneys, livers and heart (basically, combinations of the various edible parts).  That study, plus one other, Ayotte explained, are “loaded” so as to find high levels of androgens, because they use old boars whose androgen levels are higher, and they feed the volunteers with very large meals (300g of pork, in some cases).  And still, Ayotte argues, the highest value measured in the studies was 2.4 ng/mL and 2.1 ng/mL, under half of what was found in Houlihan’s urine sample.

As has usually been the case before, the CAS Panel finds in favour of World Athletics, saying that “it is possible but improbable that the ingestion of boar meat (cryptorchid) would have resulted in the urinary concentration” found in Houlihan’s sample (point 114).

This is the result of both the above arguments.  Prof McGlone’s testimony that young cryptorchid species have very low androgen levels, especially in the stomach, is combined with Ayotte’s testimony that lab studies rarely find levels of 19-NA even half that of Houlihan’s sample unless the person eats large amounts of kidneys, hearts, and liver, to dismiss this element of Houlihan’s explanation.

d) The 19-NA in Houlihan’s sample is similar to what would be expected from boar meat

By this stage, there’s not much left to Houlihan’s pork burrito defense, given that CAS have ruled that both the ingestion and the analytical outcome of ingestion are implausible.  But there is one final piece of information to consider, and that is the likely source of the 19-NA.  This also throws up perhaps the most interesting part of the whole case, which we’ll see shortly.

What is the source of the 19-NA?

Embed from Getty Images

For this, we go back to GC/IRMS testing.  Remember that GC/IRMS is a test that allows the lab to distinguish between the origin of compounds based on their carbon isotope signature.  In Houlihan’s case, the GC/IRMS result was used to conclude that the 19-NA in her urine was not endogenous (produced in her body), because it had a carbon isotope signature of -23‰, whereas two other steroids in her sample were measured at -19.2 and -19.3‰, respectively.  On this basis, the lab issued an AAF because the difference between the 19-NA and the endogenous reference compounds (Houlihan’s other steroids) was large.

That doesn’t entirely close the door on the issue, however, because that GC/IRMS result hasn’t ruled out the chance that the 19-NA may have its origins in the boar whose edible parts Houlihan is claiming to have eaten.

And so we return to a discussion of GC/IRMS, this time to include the question: “Is the isotope signature of US pork meat compatible with the analytical data in the Athlete’s sample?”  You’ll recall that earlier, we discovered that GC/IRMS is trying to answer two questions.  First, is the 19-NA in the urine sample endogenous?  Second, could the 19-NA have come from edible parts of pork?  World Athletics answered the first question in order to secure their AAF and ADRV, and they now attempt to answer the second question in trying to dismiss the possibility that Houlihan’s 19-NA had pork origins.

To do this, they rely partly on GC/IRMS, and partly on another lesson in USA food supply practice, this time a discussion of what the typical diet of boars in North America consists of.  McGlone testifies that almost 100% of commercial pigs in the USA are fed a consistent corn-soy diet, which is important because the World Athletics position is that when pigs are fed a diet that is corn-based, the steroids produced by that boar will have a carbon isotope signature between -18.5‰ and -21‰.  Ayotte’s testimony supports this, and she refers again to research she has been involved in, where the 19-NA produced by people who consumed edible boar parts was indeed in that range, as well as to “ample evidence supporting the likelihood of a carbon isotope in this range, and not in the range of -23‰, as was found for Houlihan’s 19-NA.

Ayotte also makes a claim here that the endogenous steroids used as reference compounds for Houlihan, which had an isotope signature of -19‰, supports her position.  One can appreciate why the isotope signature of the 19-NA in Houlihan’s urine (-23‰) stands apart from the endogenous steroids (-19‰) and refutes the idea of endogenous production, but quite why this is relevant for a discussion on the likely carbon isotope signature of the boar is unclear.  It may be another instance where the Reasoned Decision confuses some concepts.

Nevertheless, CAS finds Ayotte’s testimony in this regard convincing, despite Houlihan’s attempts to rebut it.  Those attempts take two forms.  One is that her expert Prof Strahm testifies that the carbon isotope ratio of wild boar in Germany or in any other regions of the world remains unknown.  His testimony, as written in the Reasoned Decision, appears to amount to one casting doubt on the certainty of ruling out boar meat based on the isotope signature of 19-NA, and he again repeats that pharmacokinetic testing would be required.  Presumably, as before, there is a great deal of supporting evidence behind this, but the section of the Reasoned Decision detailing his evidence is comparatively thin.

Houlihan’s second expert is Prof Anne Hope Jahren, from Norway.  She testifies that her analysis of edible pig parts produces a carbon isotope ratio between -21‰ and -24‰, which is exactly in line with Houlihan’s result, and refutes Ayotte’s statement that the range in pigs is between -18.5‰ and -21‰.  The crux of this issue is the pig diet.  Her research is on pigs who consume soy-based diets, whereas the testimony of McGlone has already informed CAS that almost 100% of commercial pigs in the USA eat a corn-based diet, with only 20% soy, dropping to 10% in the weeks leading up to slaughter.

It is this statement that pushes CAS into finding for World Athletics yet again.  They dismiss the evidence of Prof Jahren as “not relevant to this case,” because the soy-based diet is not typical of pigs in the USA.  Instead, it is McGlone’s testimony about corn-based diets, combined with Ayotte’s testimony that such a diet would be expected to produce an isotope signature between -18‰ and -21‰, and not the -23‰ of Houlihan’s sample, that leads CAS to conclude “The Panel finds that the carbon isotope signature of the Athlete’s A- and B-Samples is neither consistent with the carbon isotope signature of commercial pork in the US nor her own signature”.

And with that, almost every element of Houlihan’s explanation has been dismissed by CAS as implausible.  

One final point – this issue throws up what may be the most interesting testimony of the entire case. It comes from Prof Ayotte, who is describing that the carbon isotope signature of the 19-NA in Houlihan’s sample is -23‰, which is very different from Houlihan’s endogenous steroid compounds and from the edible parts of boar in the USA, as we have seen above.

However, having explained what it is not, she then offers some insight into what the 19-NA may be.  Her testimony, in Points 74 and 116, makes the fascinating observation 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 can be divided into two distinct batches.  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‰ , 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.

Houlihan’s final defense arguments

Two final defense arguments remain for Houlihan.  First, there is a polygraph, which anti-doping organizations tend not to take too seriously, and hair analysis.  CAS gives relatively short shrift to both, for the following reasons.

The polygraph is dismissed because the questions posed were ‘rather restrictive’ (point 134).  At this point, CAS also makes the observation that Houlihan stated that she was not aware of what nandrolone was before she received the charge sheet.  CAS then states that “it would have made more sense to ask the Athlete whether she had taken doping substances at the material time,” which is again an indication that this is not about doping with nandrolone, but rather something else that might lead to a 19-NA positive result.

As for the hair sample, it is diminished because it did not test for oral precursors of nandrolone like the 19-nor-DHEA that Prof Ayotte said would viably provide the level and carbon isotope signature found in Houlihan’s sample.  Also, the hair expert Dr Kintz was unable to say how large the exposure to nandrolone would need to be in order to detect it in the hair, which basically undermines the sensitivity of the test.  

Conclusion

And that does it. CAS has heard all of Houlihan’s arguments, and is not sufficiently convinced by any of them to come close to ruling in her favor. They rule, correctly, that the WADA-accredited lab in Montreal that analysed her finding was not required to perform pharmacokinetic testing because Houlihan’s delta-delta values were not within the endogenous range. The burden then shifted to Houlihan to prove how the nandrolone got in her sample. They agree she ate a burrito on December 14, 2020, maybe even a pork burrito. But CAS doesn’t find it likely that the concentration of nandrolone in her sample could have come from said pork burrito, not just because of the problem of boar meat showing up in the food supply but because the carbon isotope ratio of Houlihan’s sample does not align with what would be expected after consuming boar meat. Houlihan’s fight to prove her innocence is not over – she is appealing CAS’ decision to the Swiss Federal Tribunal – but she faces an uphill battle.

Discussion: We paid Sports Scientist Ross Tucker to break down the Shelby Houlihan case for us

*For more of Ross Tucker’s analysis of the Houlihan case, including his opinions on CAS’ decision, check out this article: Sports Scientist Ross Tucker Tackles Our Biggest Unanswered Questions About the Shelby Houlihan Case.

Previous Shelby Houlihan Coverage