OPINION: Semenya v IAAF: Implications for women’s participation in sport

This article first appeared on The Conversation.

When can a woman not compete with other women? This is, in essence, the question at the heart of the hearing currently before the Court of Arbitration for Sport (CAS), the international body that helps settle disputes related to sport. The ruling in the case of Mokgadi Caster Semenya and Athletics South Africa v International Associations of Athletics Federations (IAAF) is due at the end of April.

The case centres on the legality of a 2018 IAAF eligibility regulation for women with differences of sex development (intersex women), defined by the IAAF as women who have testosterone levels of over five nanomoles per litre of blood (nmol/l) and whose bodies can ostensibly use that testosterone better than other women can. Under these rules, women athletes with differences of sex development would have to reduce, and maintain, their testosterone levels to 5nmol/l or less in order to compete. The reasoning behind the regulation is that women with naturally high testosterone levels, and whose bodies are apparently highly sensitive to that testosterone, have a significant performance advantage over their peers in certain events.

However, this assertion has to be called into question. First, because research has been unable to prove a direct, causal relationship between testosterone levels and athletic performance – given that so many other factors play a role. And second, because there is no valid laboratory test to determine a woman’s degree of sensitivity to testosterone. Currently, the IAAF mandates physical, gynaecological, and radiological imaging to determine physical signs (such as an enlarged clitoris) as a proxy for testosterone sensitivity. However, this approach is not reliable, liable to false interpretation and subjectivity, and widely viewed as inappropriate and an invasion of privacy.


In 2015, following another case before the CAS, a previous iteration of the regulation was suspended and the IAAF was given a deadline to produce new evidence to uphold regulating women’s participation in this way. The IAAF subsequently developed the now contested 2018 regulation on the basis of a new study, which showed a supposed “performance advantage” for specific track events.

However, as my co-author and I argued in a BMJ editorial, this study was shown to be flawed and the authors subsequently acknowledged errors in the data used for the research. This is a red flag of the “science” that underpins the 2018 IAAF regulations. It draws into question the justification for the regulation for these, and indeed any, athletic event. It is against this background that we arrive at the 2019 CAS hearing.


Sports organisations have been grappling with the question of eligibility for years now. In reality, however, the science of this issue is quite clear. From a scientific and medical standpoint, we know that testosterone is not the only – or even primary – indicator of sports performance. Indeed, there are many other factors at play – including training, funding, and access to resources – in the development of a winning athlete.

Further, in non-athletes, testosterone ranges between 0.4-2.0nmol/l in girls and women. In elite women athletes, the testosterone range has been shown to be between 0.4-7.7nmol/l, and that women can and do have much higher levels than that, which can also overlap with men’s ranges. So an arbitrary 5nmol/l limit for women could have the effect of capturing and regulating a much larger group of female athletes than intended, including women with polycystic ovarian syndrome (who naturally have high levels of testosterone).

In a rebuttal to our BMJ editorial, Stéphane Bermon, a director in the IAAF Health and Science Department, wrote that the IAAF regulation testosterone cut-off point of 5nmol/l is not arbitrary, but rather aims to include women with polycystic ovarian syndrome but to exclude women with differences of sex development. The glaring inconsistency here is that we simply cannot distinguish between these groups based on blood testosterone alone, never mind determining who is sensitive to testosterone and so has a supposed “performance advantage”.


Just because regulations exist does not mean that they are evidence-based, ethical, or even effective. The crux here is that this kind of regulation has its legacy in the long and problematic history of “sex testing” women athletes. It is no accident that the vast majority of athletes affected by these regulations are black women and women of colour from the global south who do not conform to Western ideals of femininity.

We don’t yet know what full evidence, beyond a review from Bermon and colleagues before the current CAS panel is. But we do know that sex and gender are incredibly complex. Historically we have defined humans as binary “male” or “female”, based on what we knew then about genes and anatomy. This was a clear and useful way of categorising sports participation. But 21st-century medical and social sciences have since progressed, and we now know that both biological sex and sociocultural gender are much more complicated than that.

Dr Sheree Bekker is an applied health scientist, with a primary focus on injury prevention and safeguarding in sport settings and is a Prize Fellow in Applied Health at the University of Bath.

The Conversation