Testosterone, sex and gender differentiation in sport – where science and sports law meet

Published 14 October 2014 By: Dr Ben Koh, Peter Sonksen OBE, Daryl Adair

Caster Semenya and Dutee Chand

In historical terms, modern, codified sport was invented in Britain by men, for men and thus in their own interests. Games therefore centred on Victorian notions of athletic manliness and associated ideals of masculinity. This involved physical and biological appearance (i.e. categorisations by sex) and sociocultural norms (i.e. categorisations by gender).

That process stratified sport, placing men at the core and women at the periphery of an emerging ‘cult of athleticism’. A ‘natural’ symbiosis seemed to exist between sport and manliness, while for women anything more than gentle exercise was roundly condemned as ‘unnatural’ for the female body and contrary to ideals of femininity.

The 20th century witnessed a sustained struggle for female participation, with the Olympic Games something of a litmus test for women’s rights in sport. Reflecting the wider constraints of a patriarchal society, women were often disallowed from competing in several sports, or constrained from taking part in particular events within them. This was especially so for contests involving endurance or strength, neither of which women – as a consequence of their ‘biology’ – were presumed to possess. The rule of thumb was that men were ‘naturally’ equipped to be athletic; women, by contrast, faced physical limitations by virtue of their sex.

Women were, however, extremely important to Olympic national teams; after all, a medal for a female athlete counted the same towards the medal tally as that for a male athlete. So, in very crude terms, even though women were too often trivialised as athletes of calibre, they very much counted on the patriotic scoreboard. Indeed, by the 1960s the performance of women at the Games was a hot topic. This stemmed from suspicions about the use of Androgenic Anabolic Steroids (AAS), a class of drug that could be used to help “masculinise” the athletic body, whether male or female, with assumptions thereof about improved athletic performance. As is now known, systematic doping was especially pronounced in several East European countries.12

While sport administrators were wrangling with the question of doping as an unfair advantage, they had a particular fear about the androgenising impact of AAS on women. This concern extended beyond mere biological considerations of how a changed physiology afforded sporting prowess; AAS was also likely to have implications for idealised conceptions of the female gender in sport. In that respect it seems telling that sex testing (1966) was introduced before drug testing (1968) in sport. It also showcased how biology (sex testing) was used as a front to support an underlying commitment to prevailing gender norms and associated inequities. In the case of high performance sport, women could be asked to ‘prove’ that they were female, while there was no requirement for men to do the same. By convention, men and women were almost always separated in athletic contests; this was underpinned by an assumption that male athletes had performance advantages over female athletes by virtue of sex-based physiological differences.


Gender and Sex

Gender identity is a combination of physiology and social influence. Based on current acceptance, gender identity is the sense one has of being either male or female. Whereas sex is narrowly biological, gender is a combination of physiology, psychology and socio-cultural factors. Discussions about sex and gender are therefore complex, resulting in highly charged and difficult debates.

If gender identity discussions confined to known phenotypes (observable biological sex characteristics) of male and female are complex, even greater confusion arises when the phenotype itself is in question. This is because the physical phenotype of an individual is the result of sex genes (chromosomes; genotype) as well as the hormonal effects of those genes during the person’s development. A series of biological events need to occur for that genotype to manifest into physical form to create the phenotypic sex we see. The long chain of biological steps from the genes to the person’s observable physicality has meant that a natural spectrum of biological (sex) variability exists.

Gender identity is therefore a complex and socially nuanced matter that exists on a biological continuum, which has long been retrofitted in the context of sport into a social binary that uses sex differences to separate men and women. Without expert knowledge of biological science, sports administrators have historically made naive assumptions10 about ‘normality’ both between and within the sexes.


Verification and Classification

The International Association of Athletics Federations (IAAF) was the first to introduce a ‘chromosomal verification’ test, followed by the International Olympic Committee (IOC) in 1968, at which point drug testing also began at the Olympics. The chromosome test was based on the assumption, subsequently shown to be flawed, that sex could always be determined by the biological information derived from a buccal smear (e.g. the Barr body sex test).

In the mid-1980s, Spanish hurdler Maria Martinez-Patino’s high-profile campaign for reinstatement after being disqualified for failing a Barr body test was used to pressure the IOC and other organisations into changing (or eliminating) their sex tests. Patino ‘failed’ the Barr body test only because of a medical condition; a condition that gave her no physiological sporting advantage.5 Her successful appeal was largely influenced by the fact that human rights activists and geneticists believed that the test was unfair and took up her cause as a test case through which they could make profound points about equality, scientific objectivity, and the complexity of human gender identity. Their main argument, then, was not that testing for sex was problematic in and of itself, but rather than this specific test, using chromosomes as a proxy for sporting ability, was inappropriate.5

In 1988 the IAAF dropped chromosomal and genetic testing in favour of a manual/visual ‘health check’ by the team doctor, but then abandoned all forms of systematic sex testing in 1991. The IAAF argued that these were no longer necessary because doping regulations required athletes to pass urine in front of witnesses, thereby allowing a de facto ‘check’ of genitalia. Moreover, modern sportswear was now so revealing that it seemed unfeasible that a man could masquerade as a woman.

The IOC was more resistant to change, instead introducing a genetic test in 1992 that identified a specific region of genetic code usually found on the Y chromosome, known as the ‘sex determining region Y’. Even this test continued to throw what, for the purposes of sex segregation, seem to have been false positives, for although at the Atlanta Olympics in 1996 eight women ‘failed’ this test, all were allowed to compete after further examinations were carried out.5 Finally, in 1999 the IOC agreed to follow the IAAF and remove the requirement for blanket sex testing, so that the 2000 Olympics in Sydney was the first Games in three decades where the genetic make-up of female athletes was not scrutinised.6, 9

Of course, as we have seen recently in the case of Caster Semenya,11 if the gender of a female athlete is actually challenged, she can still be required to undergo a full gamut of tests: physical, physiological, genetic, hormonal, psychological. More generally, too, and notwithstanding the ‘end’ of sex testing, sports administrators soon turned to another biological variable – testosterone, a substance that men typically have in greater volumes than women, and whose presence is widely assumed to translate into performance capabilities in terms of strength, speed and endurance.



As the sex testing of the late 20th century indicated, there is much at stake for those who are firmly committed to a male-female binary. With that in mind, it is perhaps unsurprising that sex testing has been reintroduced in a new guise. A catalyst was the IAAF’s reaction in 2009 to outstanding performances by South African runner Caster Semenya, who bettered her personal best by an astonishing margin. While such an outstanding effort might reasonably call for target testing on the part of anti-doping authorities, the IAAF took the extraordinary step of seeking medical evidence to verify that Semenya was indeed ‘female’. This news leaked to the media, after which there was perverse speculation that she was ‘really a man’, ‘half man and half woman’ and so on. This was not helped by comments from fellow competitors in her 800-metre victory in Berlin. The Italian Elisa Cusma did not mix words: “For me, she’s not a woman. She’s a man.2 The complex details of the Semenya story and others like it have been told elsewhere; the key point here is that the IOC, with the support of the IAAF, was about to revisit sex testing under a new guise.

As Pieper8 explains, a new policy7 was announced just weeks before the 2012 London Olympics:

The ‘IOC Regulations on Female Hyperandrogenism’ ...[is] to ensure equitability, [so that] women with higher-than-average levels of androgens [are] deemed ineligible to compete. Androgenic hormones control muscular development and women with hyperandrogenism typically produce an excess of (naturally produced) testosterone. Concerned with ‘strength, power and speed, which may provide a competitive advantage in sport’, the IOC required all National Olympic Committees to ‘actively investigate any perceived deviations in sex characteristics’.

So we have come back to the future, as it were. Women athletes are again required to ‘prove’ they are ‘female enough’ to compete ‘fairly’ against others of the same sex.

There are various problems with using androgen (specifically testosterone) as a measure of eligibility in sport. Firstly, the assumption is that all women and men can be categorised neatly into one sex or the other. Where does this position people who have characteristics associated with an intersex status? For male athletes this seems no problem; after all, it is only women who are expected to prove themselves as ‘real’ women.

A second problem, which this article focuses on, is the assumption that testosterone levels in the human body have limited variability within the sexes, and whilst it may appear logical to have threshold testosterone measures that validate participation in sport according to sex, this is confounded by recent research. Although information on testosterone levels in the blood of elite athletes is limited,1, 4 the available data in fact shows that there is not a complete separation between the sexes, but rather an overlap between the levels seen in elite male and female athletes. This finding also has significant implications for the IOC’s Female Hyperandrogenism policy, which attempts to limit ‘natural advantage’ outside of what it considers to be normative levels of testosterone among women.


Testosterone distribution in male and female athletes

It has long been believed that all men have a higher level of testosterone in their blood than women. In part this may be due to the fact that 95% of men in the general population typically have levels between 10 & 30nmol/l while 95% of women typically have values between 0 & 3nmol/l. There are, however, many biologically normal women with so-called ‘hyperandrogenism’ as well as those with hereditary and medical conditions that can lead to much higher levels of testosterone in women. Some conditions, for example Androgen Insensitivity Syndrome (AIS), are a result of a deviation to the usual chain of biological events from genotype to phenotype described earlier. Others are due to acquired conditions after birth; for example, Polycystic Ovary Syndrome (PCOS) in women where the body often produces greater amounts of testosterone.

Outside of the hereditary and medical conditions that can alter a person’s testosterone levels, as has been reported by studies in the general population, it is only very recently that data on testosterone levels in elite female athletes have been available.1, 4 Two published papers found that there were a number of female athletes with testosterone values in the so-called ‘normal male’ range. Healy et al found a complete overlap between the testosterone levels in elite men and women, with 16.5% of men with a low testosterone value and 13.7% of women with a high one.

What the recent scientific evidence suggests, therefore, is that the discrimination of athletes into gender categories using testosterone levels alone is fundamentally flawed. Firstly, women with a higher than ‘normal’ testosterone, who are otherwise healthy, should not have to undergo medical (and sometimes surgical) treatment just in order to compete in the female sport category. Otherwise, the rationale of requiring female athletes to lower naturally high testosterone to abrogate any perceived performance advantage would also mean requiring athletes with higher than ‘normal’ height to undergo height-shortening surgery to ensure that they negate any height advantage for sports performance. Secondly, if women have to alter their biology to reduce their natural testosterone levels, then male athletes with lower levels should be able to increase their testosterone up to the maximum allowed for men by sport authorities.


Sports Policy and Law

The civil courts in most countries have traditionally been reluctant to intervene in matters of sports rules. There are only very limited grounds upon which athletes can appeal to the courts for consideration of sports governing decisions. For example, if there is restraint to the economic freedom of the individual (e.g. David Meca-Medina and Igor Majcen v Commission of the European Communities, 2004), a violation of human rights (e.g. Andriy Kashechkin v UCI, 2007; Doug Barron v PGA, 2009), or under a general duty of care and good faith, including, inter alia, protecting the athlete from exposure to public humiliation and ridicule (e.g. Vijay Singh v PGA, 2013). Whether an athlete can be successful in such legal proceedings is a secondary consideration, since sports and its governance is not just judged by the court of law, but also by the court of public opinion.

Female athletes should not have to apologise for ‘being’: some have physiological attributes that endow them with qualities optimal for certain sports, while others come to the contest without noticeable characteristics, such as abundant height. If sport is a competition between individuals with both natural (genetic) and contrived (behavioural) virtues, then it makes no sense to impede what has been provided at conception and during physiological development. In that sense, discriminating against female athletes on naturally occurring testosterone levels is contrary to equity in sport, as well as being inimical to a broader, more fundamental principle of the human right ‘to be’. Indeed, in the opinions of the authors the requirement set by sports administrators that female athletes undergo body mutilating surgery or endure the adverse effects of drug treatments just so as to compete in a category of sports violates a basic duty of care to the sportswomen.

This is not merely a discussion of hypotheticals. There have been several cases, beyond Caster Semenya, where female athletes have been constrained from participating on the assumption of ‘unfair’ levels of testosterone. The most recent is Indian sprinter Dutee Chand, who has been banned from participating by the IAAF under its ‘Hyperandrogenism policy’. She is appealing to the Court of Arbitration for Sport (CAS), crucially with the support of the Sports Authority of India.3 As a woman, Dutee is in effect asking to be accepted for ‘being’; that she should not have to change how she is naturally. Given that some female athletes have been coerced into hormonal treatments and, more seriously, surgery to ‘normalise’ themselves as women, the case before CAS is profoundly significant.

Dr Elizabeth Ferris

Liz Ferris MB BSThe authors dedicate this article to Dr Elizabeth Ferris, who died in 2012. Peter Sonksen was a close personal friend and colleague of Dr Ferris, an Olympian who won a Bronze Medal for diving at the Rome Olympic Games in 1960 and who was, as The Times obituary said: “… an unremitting and successful campaigner for the rights of women in sport, being notably involved in the controversy of gender eligibility for the Olympics, which recently enveloped the South African runner Caster Semenya.” (https://www.thetimes.co.uk/tto/opinion/obituaries/article3396702.ece). Liz sat on the IOC/IAAF working party that developed the ‘hyperandrogenism’ and was strongly opposed to its implementation.


Related Articles


Dr Ben Koh

Dr Ben Koh is a medical doctor with a Masters in Sports Medicine and a Masters in Psychology and has clinical and educational training in surgery, sports medicine, emergency medicine and critical care.

Peter Sonksen

Peter Sonksen OBE

Peter Sonksen OBE MD FRCP FFSEM (UK) is a 78-year-old retired Professor of Endocrinology from St Thomas’ Hospital and King’s College, London and a Visiting Professor at Southampton University. He was recently awarded an OBE for his contribution to the fight against doping in sport.

  • This email address is being protected from spambots. You need JavaScript enabled to view it.
Daryl Adair

Daryl Adair

Daryl Adair is a historian and Associate Professor of Sport Management at UTS Business School, University of Technology Sydney, Australia. His recent research focuses on policy challenges in sport, such as integrity processes and the politics of elite participation.

  • This email address is being protected from spambots. You need JavaScript enabled to view it.