I was asked by the Science Media Centre to provide an expert reaction to an editorial about testosterone, women athletes, and rules in elite sport, as published in the British Medical Journal. The article can be found at this link.
As well as my “biochemical” comment, the Science Media Centre received two other expert comments from endocrinologists. Interestingly, whilst one endocrinologist wrote a very short note supportive of the article, the other was highly critical and wrote a lengthy piece outlining the editorial’s scientific shortcomings. I’ll see if I have permission to post that article later on this blog as it is really interesting and chimes with what I think.
Interestingly my opinion was quoted very differently (but still correctly) by both the BBC and the Daily Mail!
As I say at the end of my piece, there are some good reasons to drop the IAAF regulations on differences in sexual development (DSD). But I don’t think bad supporting science is one of them. In the words of TS Eliot from the play Murder in the Cathedral:
“The last temptation is the greatest treason: To do the right deed for the wrong reason”.
My full expert opinion on the BMJ editorial follows below:
First the closer overlap between male and female plasma testosterone levels in elite sport noted in this paper is perhaps not surprising for a number of reasons:
(a) Increases in plasma testosterone in female athletes (in part due to differences in sexual development, DSD) are likely to improve performance more than similar increases in men. Therefore women with higher testosterone levels will probably be overrepresented in elite athletes. It is well known that the proportion of conditions leading to high testosterone levels is higher in elite female athletes compared to the male population. For example at the Atlanta 1996 Olympics several female athletes tested positive for the SRY gene on a Y chromosome. SRY induces testosterone production during puberty so this likely led to increased levels of plasma testosterone. This testosterone was considered not to be having an effect on their body (androgen insensitivity syndrome, AIS). So an XY apparently “male” genotype led to a female phenotype. The prevalence of AIS in the non elite athlete population is much lower than the 7 in 3387 found in this sample of Olympians. This suggests that increased testosterone levels either do still slightly increase performance in some AIS females OR make them more likely to choose to become athletes. Either way, this is ONE example of why testosterone levels in female elite athletes may have a closer overlap with male elite athletes than in the population at large.
(b) Taking artificial anabolic steroids will likely lower the levels of the natural anabolic steroid (testosterone) in athletes. These effects can potentially be quite long term. Without knowing the number of athletes previously steroid doping in any sample this is difficult to correct for. I suspect because of the greater adverse side effects in female athletes rather than males, males take higher steroid doses than females. This would result in anabolic steroid use lowering the natural male testosterone levels more than the female levels, again closing the gap and increasing the overlap.
(c) The major physiological effects of plasma testosterone levels occur during puberty. However, it is not possible to test at puberty in future elite athletes. By its very nature any later testing is of somewhat secondary value. So an elite male athlete might have low testosterone now, but at the crucial time in their development it was significantly different from the future elite female athlete. It is hard to see how any data can be collected in this area. So the IAAF make do with what they can actually measure.
Second, gathering hormone data in elite sport is difficult. It is even harder (and potentially unethical) to gather performance data under conditions where these hormone levels are made to drop or rise due to external factors. Therefore by its very nature correlation data of the type described in this article will become important. But it is individual data that really matters, and this is even harder to get access to for the obvious grounds of medical confidentiality. As the authors mention Caster Semenya, it is perhaps instructive to look at her 800m times* at global championships when the IAAF testosterone levels were in place or not. This is worth mentioning as you can be sure the same discussion has been had amongst other female elite athletes (and also athletics fans). In periods when the testosterone regulations were not in place Semenya won all the global championships she competed in. When they were in place she ran slower and was second to cross the line at global championships. The winner those times (Mariya Savinova) was banned for doping (possibly by taking substances including anabolic steroids) and thus possibly giving her the same advantage by cheating that Semenya (who was not cheating) had been denied by the IAAF regulations.
Third, the IAAF testosterone regulations regarding DSD (differences in sexual development) – although not mentioned explicitly – need to be seen in the light of their (and the IOCs) anti doping and transgender regulations. This is not mentioned by the authors of this article, but seems relevant. This is NOT to say that females with DSD or male to female transgender athletes are cheating of course. But the science underpinning the possible performance benefits of their situation is related. And in many cases – perhaps surprisingly – the doping data is easier to access.
On doping regulations: there is good evidence that increasing exogenous testosterone levels (or taking artificial anabolic steroids) enhances sport performance. This effect is greater in females than males. It is (partially) reversible if anabolic steroids stop being taken. It is no surprise that anabolic steroids preferentially enhanced performance in female athletes in the former East Germany (where they were likely given to junior athletes). Nor that the Russians state doping system had some of its greatest ‘successes’ when giving anabolic steroids to female middle distance runners. Therefore doping is banned.
On transgender regulations: lowering plasma tesosterone levels is the main goal of male to female transgender hormone therapy. Transgender women, who are transitioning or have transitioned from male to female, are treated with hormones (by choice) to lower their plasma testosterone levels to similar levels as required by the IOC regulations. This decreases their performance. Therefore they are allowed to compete in a female category after hormone therapy to lower their testosterone.
Nothing about elite sports performance is ever going to be amenable to scientific conclusions “beyond reasonable doubt”. This is because the fractional differences in competition at the highest level are so small and the population so difficult to study ethically and practically. In my opinion the current evidence is as good as we are going to get to show that both endogenous (natural) and exogenous (doped) testosterone levels enhance female sports performance. In the terms of the Court of Arbitration of Sport, I am personally “comfortably satisfied” that both exogenous and endogenous testosterone increases sports performance in elite female athletes. Of course a lot of other factors, genetic and environmental, also affect sports performance, but that is a separate story.
Summing up, in my personal opinion, the IAAF regulations have as good a scientific basis that they are going to get in the foreseeable future. HOWEVER, this does not mean that they are ethically “right”. That is a separate matter. I write as someone who personally cheered Caster Semenya on to victory in the 800m in the 2017 World Athletic Championships in London (when the DSD regulations were not in force). However, we should focus on the ethics of what we want to encourage as participation in female sport, and not get hung up on criticising the fine details of science that is never going to be as conclusive as we would like.
Finally, I caution against the authors making too much of the effects sport has on the real world. They say “History compels us to ensure that decisions about genetic superiority are supported by objective, rigorous, and reproducible data” and “These issues highlight the fact that although sports policies exist to serve the organisations that develop them, the effect of these policies on individuals, societies, and even medical science has far reaching implications.” Sport – and especially elite sport – as we know it is a social construct created by (largely) males at the end of the 19th Century. We have chosen in society to place a huge value on it and honour its participants with wealth and fame. But it is inherently sexist. Biological females have no chance of competing on an equal footing with males – in terms of physical sports performance they demonstrably can’t compete. Equality of opportunity in sport does NOT equal equality of outcome. In fact it does not come close. In nearly all adult sports there would be NO female winners if we did not have a separate female category. This is unlike any other activity we value in our society. Therefore elite sport has created a special protected space for females (female sport). Inevitably this creates tension at the interface as noted in this article. But it is important that writers – such as the authors – do not try and put sport on a pedestal. Decisions, such as those of the IAAF, relate to the rather special social construction of the sporting world. We should note give them added importance by trying to translate them into more important parts of life and society. The Olympic Charter claims that the practice of sport is a human right. Well this may be the case, but in my opinion it is low on the list of ones that we should use as an example of how to construct how we run our society.”
* Caster Semenya times:
Pre regulation requiring testosterone levels to be dropped (2009 World Championship Berlin): Winner, 1:55.45
Post Regulation (2011 World Championships Daegu): Second*, 1:56.35
Post Regulation (2012 Olympics London): Second*, 1:57.23
Regulations dropped (2016 Olympics Rio): Winner, 1:55.28
Regulations dropped (2017 World Championships London): Winner, 1:55.16.
* upgraded to Winner after Mariya Savinova disqualified for doping offences (likely including anabolic steroids).