Not by gender, not by sex, but by testosterone saith the IAAF: International athletics and the new Female Eligibility Regulations

Published 22 May 2018 By: Dr Ben Koh, Daryl Adair, Liam Elphick

Athletics runner on track

On 23 April 2018, the International Association of Athletics Federations (IAAF) released its “Eligibility Regulations for the female classification: Athletes with differences of sex development“ (DSD Rule) which comes into effect on 1 November 2018.1 The stated purpose of the DSD Rule is for athletics to be inclusive and to respect the dignity of individuals with differences of sex development (DSD),2 through, paradoxically, the creation of discriminating criteria that affects only female athletes.

Although the IAAF claims that the DSD Rule is in no way intended to be a judgement on the sex or gender identity of any athlete, but is instead intended to preserve the dignity of athletes with DSDs, the new IAAF Rule nevertheless requires that otherwise healthy female athletes with DSD must reduce their naturally elevated testosterone levels by medical means. This intervention specifically excludes surgical anatomical changes, but chemotherapy, hormonal castration, and iatrogenic irradiation are all presumably acceptable means for reducing natural testosterone production in women subject to the DSD Rule.

This article explores the history leading to the DSD Rule, the apparent scientific basis for the criteria, as well as the justification and proportionality of the DSD Rule. Based on this review, the authors question what they view as the continued discrimination towards females in its many guises, and indeed specifically of intersex athletes, in an age when sex is understood to exist along a spectrum, and when gender identity is fluid.


The IAAF has long been engaged with processes to establish the "authenticity" of female competitors. As early as 1946, the IAAF required medical certificates from women in order to validate their sex status. The International Olympic Committee (IOC) followed the IAAF’s lead two years later. By 1966, international athletics featured a combination of "nude parades" or even gynecological examinations to establish, to the satisfaction of sport authorities, that the competitors were indeed women, rather than men masquerading as women.3

By 1968, sex testing via chromatin analysis was introduced at the Olympics with the support of the IAAF, along with drug testing of both male and female athletes. At this point the primary interest was with the impact of exogenous androgens (i.e. doping) on women: there was anecdotal evidence that Eastern Bloc female athletes were being virilised by the imposition of Anabolic Androgenic Steroids (AAS). Sport authorities were, by this time, more anxious than ever that the women’s category of competition needed protection, whether from faux females, people who were not "quite" women, or women who, in an effort to enhance athletic performance, were prepared to virilise using AAS.4

The chromatin test was abandoned in 1988 after it dawned on the IAAF that there are variations beyond the typical chromosomes of XX female and XY male. A Spanish female athlete, Maria Jose Martinez-Patino, who had been provided with the mandated "Certificate of Femininity" was later found, on the basis of a subsequent chromatin test, to have XY chromosomes. Ms Patino was androgen insensitive such that her endogenous testosterone levels have no impact on her body and everything about her life, gender and sex organs was typical of a woman. Despite this, her chromosomes did not conform to the XX female mould.

Ms Patino’s case led to a growing realisation that biological sex is on a spectrum, with a very small proportion of people (in this case women) who do not fit conventional expectations of neat binary categories of female and male. As such, the IOC and the IAAF abandoned sex testing just before the 2000 Olympics. The surveillance and intrusion on women seemed disproportionate to concerns about the "authenticity" of female athletes.5

By 2006, the IAAF – while assuring women that there would be no compulsory or regular sex testing during track and field events – announced that if a female athlete had "suspicion" about the "authenticity" of a fellow competitor they could institute a "challenge", this allowing the IAAF’s medical panel to evaluate the individual. This process could involve gynecologists, endocrinologists, internal medicine specialists, psychologists and others.6 That policy was applied three years later in what was arguably the most controversial moment in the history of women’s sport. Caster Semenya, a teenage athlete from South Africa, had performed beyond expectations at the IAAF World Championships, with her time prompting concerns of doping. After repeat excellent performances, it became obvious that Semenya was an authentic champion – there was no hint of doping. In the meantime, though, commentary by fellow competitors, combined with salacious media reporting, produced a discourse that Semenya was not an "authentic" woman. Put crudely, the charge was that she was virilised compared to a typical female athlete, this giving her an advantage over track competitors. It may be a natural part of who Semenya is, but critics insisted that this was unfair on other women athletes.7

During this cacophony of speculation and accusations, Semenya remained outwardly calm, even if inwardly distressed. The IAAF, meanwhile, was obliged to investigate the "suspicion" based on a "challenge" by one or more athletes. By mid 2010, the IAAF’s "panel of experts" had concluded that Semenya was free to compete in athletics with immediate effect.8 The case seemed closed. However, ten months later it became apparent that the IAAF was re-evaluating its position. A new "Hyperandrogenism" (H) Policy was introduced in 2011 with the focus of "authenticity" for the female category based on the levels of endogenous (naturally occurring) Testosterone (T).9

In the general population, women of reproductive age typically have T levels within a range that is less than 1.0 nmol/L,10 whilst healthy men have a higher level of around 10-30 times that of women.11 The inference appeared simple enough: a fundamental physiological difference between women and men, as well as a prime driver of athletic performance, is endogenous T. Women who participate in IAAF competition should, as a consequence, have a level of natural T that is commensurate with the broader female population. The assumption is that high levels of endogenous T provide female athletes with an "unfair" (albeit natural) advantage over competitors in what the IAAF deems the "normal" range.12

The 2011 H Policy, to the delight of some female athletes, was set to impact Caster Semenya who, it was widely understood, had high endogenous T (reportedly as a consequence of a congenital condition). Like other athletes deemed hyperandrogenic, Semenya had two choices: medical (including surgery) intervention or drop out of IAAF events. Semenya chose to continue, though it is unknown what intervention she adopted. Competing at the London 2012 Olympics she won silver, though later this was upgraded to gold after the Russian winner was found - ironically enough - to have doped. Semenya was still racing well, though observers noted that her times were a little slower than her best.13

Perhaps surprisingly, the IAAF’s H Policy was not officially challenged by Semenya. Instead, little known Indian sprinter, Dutee Chand, was the catalyst for an appeal to the Court of Arbitration for Sports (CAS).14 Chand had been disqualified from competing by the Athletics Federation of India, which ruled that she contravened the H Policy. A demure and non-striking physical figure, Chand was contesting a ruling that – as far as most commentators were concerned – was intended to "reign in" female athletes who gave the "appearance" of virilisation and an unfair (albeit natural) advantage.15

In 2015 CAS determined that the H-Policy was discriminatory and that the IAAF had not provided adequate evidence to support or justify such discrimination. In particular, the CAS held that the IAAF had not established that the the “degree of advantage that androgen-sensitive hyperandrogenic females enjoy over non-hyperandrogenic females16 was sufficient to justify the H-Policy. In other words , the IAAF had not established that the H-Policy was appropriate or necessary to ensure fair competition within the female category.17 CAS also noted that the difference in competition timing between the best male and female athletes in common athletic events was around 10-12%, and that the timings achieved by H women were nowhere near that of men. CAS therefore ruled the IAAF H Policy was invalid for a period of two years.18 The IAAF could present further evidence to justify the H-Policy within that two-year period, but if it failed to do so the H-Policy would be declared permanently invalid.19 Dutee Chand was therefore able to compete at the Rio Olympics, where she enjoyed the thrill of participation even though she did not get past the early heats. In the meantime, Caster Semenya enjoyed immense success, winning the 800m comfortably with the 5th fastest time among women in Olympic history.20

Sport policy and body politics

The IAAF did not come back to CAS within the two-year time frame, but the tribunal generously offered a six months extension. Curiously, though, the IAAF chose not to challenge Dutee Chand’s position: she and other high T female sprinters thus remain free to compete in the 100m and 200m without restriction. Instead, the IAAF identified a limited number of events for which it claims there is demonstrable proof of "unfair" performance advantage derived by women with naturally high testosterone levels. According to the IAAF, its scientific researchers were able to establish substantive differences in key events: the 400m (including hurdles), the 800m, 1500m, one mile, pole vault and hammer throw.21

Surprisingly, though, the IAAF deemed only track events to be of interest: it now mandated, through the April 2018 DSD Rule, that the 400m, (including hurdles), the 800m, 1500m and one mile are "restricted" events. Whether by coincidence or design, these are distances in Caster Semenya’s performance range. In these "restricted" events, female competitors with DSD will under the new rule need to demonstrate that their endogenous T is 5.0 nmol/L or lower22 at all times.23 The other requirement is that a competitor in the female category must be deemed, under law, to be either female or intersex.24 In essence, therefore, the IAAF has reconfigured the categories of male and female in athletics to be about levels of endogenous T, but only in the female category.25

The litmus test to be accepted as a female participant in 2018 is a maximum level of endogenous T.26 Athletes who are identified or present with high T will in real terms have three choices: (1) undergo treatments to lower their endogenous T and thus be eligible to compete at international levels of athletics in the female category; (2) refuse to alter their endogenous T levels and be eligible to compete only at national (or lower) levels of athletics; or (3) request participation in men’s athletic events.27

Science-based policy and "fair play" in scholarship

The 2018 DSD Rule claims to be based on broad medical and scientific consensus that high levels of endogenous testosterone circulating in athletes can significantly enhance their sport performance.28 However, the citations provided in support of this purported scientific consensus excluded independent empirical research suggesting the contrary,29 used unpublished research where its methodology and validity cannot be determined,30 and included a review paper where the authors specifically stated that the effects of elevated endogenous androgen concentrations on athletic performance among female athletes remains controversial and unresolved.31

The DSD Rule relies ultimately on a single published study (2017 Study)32 authored by Dr. Pierre-Yves Garnier and Dr. Stéphane Bermon, both affiliated with the IAAF. It was funded by the IAAF and was part of the evidence the IAAF prepared for its follow-up defence33 to its initial defeat of the 2015 Dutee Chand case at CAS.34 The stated purpose of the 2017 Study was to explore the possible relationship between serum androgen concentration, particularly serum testosterone concentration, and athletic performance. To do this, the authors performed a cross-sectional (i.e. a snapshot in time) retrospective statistical analysis of blood samples taken from female athletes at the 2011 and 2013 IAAF World Championships and compared for each athlete the best performance (heats or finals) achieved during the corresponding World Championships.35

In total, there were 1332 samples from female athletes and of these 17.3% of female athletes were counted twice - in 2011 and 2013. The 2017 Study found that twenty-four female athletes had a high testosterone level (>3.08 nmol/L) of which nine had DSD and thus presumably naturally occurring endogenous testosterone levels; nine were found to have been doping and presumably used exogenous androgens; and six were found to be impossible to classify.36 The failure to differentiate exogenous and endogenous androgens makes any conclusion problematic.37 No attempt was made by the authors to compare the endogenous DSD group with the exogenous doping group.

To test for the association between the serum androgen concentration and the type of athletic event, all athletes were firstly clustered into tertile38 groups based on their free testosterone (fT) level. No explanation was given as to why the groups were divided into three instead of two groups (a low-fT and high-fT), or otherwise simply leaving them as a continuous variable.39 This lack of explanation in its methodology may easily lead some to question the validity and practical value of any statistical effect recorded.40

Next, the samples were then assigned to one of the seven following groups:

  • throwing events (discus throw, javelin throw, hammer throw, and shot put),

  • jumping events (high jump, long jump, triple jump, pole vault),

  • sprinting events (100 m, 110 m hurdles, 200 m, 400 m, 400 m hurdles),

  • combined events (heptathlon for females and decathlon for males),

  • middle distance running (800 m, 1500 m, and 3000 m steeple chase),

  • long distance running (5000 m, 10 000 m, and marathon), and

  • race walking (20 km race walking in females and 20 km and 50 km race walking in males).

It is unclear how an athlete (if any) who has participated in more than one of these events was classified, and whether blood samples were taken from only one, from some, or from all the events that athlete participated in. For example, if a female athlete who only participated in the 2011 championship and competed in the 400m and 800m events had only one blood sample taken, would that data be classified into the sprinting group or middle distance group or both (double counting of the data)?

Finally, the serum androgen concentration tertiles were statistically compared with the type of athletic event groups. From this analysis the authors found that when the athletes in the highest-fT and lowest-fT tertiles were compared, a statistically significant but small improvement was seen only in the following:41

  • 400 m sprinters 2.73% improvement

  • 400 m hurdlers 2.78% improvement

  • 800 m sprinters 1.78% improvement

  • hammer throwers 4.53% improvement

  • pole vaulters 2.94% improvement

The authors did not report if the highest-fT tertile or lowest-fT tertile were any different from the middle-fT tertile (i.e. the middle group that was neither high nor low in fT).

Given that the study was cross-sectional in design, it cannot attribute causation, nor correlation. Additionally, since there was at least a double-counting in 17.3% of cases, this risks overstating any evidence,42 despite the authors’ assertion that it is a non-issue. It is also important to note that because the granular data were not reported for individual athletes who had both 2011 and 2013 samples, it is unclear if the levels in those athletes had changed in that time, and whether those changes (if they occurred) correlated with a change in performance of that athlete.

Even if we put aside the numerous methodological inadequacies43 and assume a true effect due to the naturally elevated testosterone that exists in DSD females, there was only a statistically modest 1.78%-4.53% increase in performance, and in just a handful of events. Certainly, the improvement was not in the order of 10-12% between males and females - a level CAS has determined to be the degree required for an unfair advantage and a legitimate separation for competition between males and females.44

The level of methodological flaws evident in the 2017 study and the DSD Rule’s reliance on it makes one question if confirmatory bias was at play and whether the conclusions drawn were potentially a strained attempt to force a square data peg into a round policy hole. The assertion of subjective scientific conclusions unrelated to objective empirical data provides a challengeable basis, either scientifically or legally, on which the DSD Rule now sits. Furthermore, robust scientific evidence of causation such as that required to satisfy the CAS Panel in the 2015 decision usually involves multiple and independent studies conducted over many years. Only three years ago, CAS formed the view that there was insufficient scientific evidence to show that female athletes with elevated T levels benefited from such a performance advantage to require that they be excluded from female competition.

According to skeptics, the onus is now on the IAAF to prove that any DSD Rule is a robust and proportionate justification for requiring high T women to medicate in order to compete in female events. No new emerging scientific evidence over the last three years, nor in the 2017 study, appear to have added anything to alter the 2015 CAS decision.

Many social science critics and human rights activists assert that the IAAF scientific research is, rather, a smokescreen for a prime objective: to curtail the athletic performances of Caster Semenya.45,46 Indeed, there has been vocal discontent among (mainly non-African) athletes who have lost against Semenya in competition, some decrying that it is "unfair" they be forced to compete against someone they believe has an unreasonable, even if natural, advantage by way of physical performance.47 If the science underpinning the DSD Rule is unconvincing, the IAAF risks being perceived as unfairly discriminatory, yet again.48

Legal, social and health-related concerns

Considering the significant concerns surrounding the research underpinning the IAAF’s DSD Rule, it is likely that a legal challenge will come. There are various legal arenas in which the DSD Rule could be challenged: namely the IAAF’s own tribunal, a national court or, most likely, CAS.49 While this breadth of potential jurisdictions makes it difficult to pinpoint the arguments that might be raised in any legal challenge, several key points can be made at this early stage by placing potential legal arguments in their social and health-related contexts.

First, the right to freedom of non-discrimination is a well-recognised and universal human right. Article 14 of the European Convention on Human Rights and Article 26 of the International Covenant on Civil and Political Rights both provide that all persons should be able to exercise their rights free of any discrimination on a number of grounds, including sex, religion, race, and "other status". The European Court of Human Rights has held that ‘other status’ protection extends to non-discrimination on the basis of gender identity,50 but no European case has yet considered whether this protection extends to intersex status.51 As the European Court of Human Rights has found that "other statuses" protected include any "differences based on an identifiable, objective, or personal characteristic, or “status”, by which individuals or groups are distinguishable from each other",52 it is likely that intersex status is protected.

The IAAF Constitution and the Olympic Charter also prohibit discrimination. The IAAF Constitution reads at article 4.4: “The Objects of the IAAF are […] To strive to ensure that no gender, race, religious, political or other kind of unfair discrimination exists, continues to exist, or is allowed to develop in Athletics in any form, and that all may participate in Athletics regardless of their gender, race, religious or political views or any other irrelevant factor.53 Similarly, the Olympic Charter, to which the IAAF is also bound,54 prohibits discrimination. Principle 6 of the Olympic Charter provides that "The enjoyment of the rights and freedoms set forth in this Olympic Charter shall be secured without discrimination of any kind, such as race, colour, sex, sexual orientation, language, religion, political or other opinion, national or social origin, property, birth or other status."55

The right to non-discrimination has also been enshrined in domestic law in numerous countries through anti-discrimination legislation. In Australia, for example, intersex status is a protected ground under federal anti-discrimination legislation.56 This means that treating an intersex athlete less favourably by reason of their intersex status or associated characteristics,57 for example by excluding them from female competition unless they alter their natural physiology, would likely be unlawful discrimination.58

There is also a human rights issue in regards to sex discrimination. The DSD Rule sets an upper testosterone limit for female international events, but no limits or bounds are set for male international events. This may raise the question of whether the IAAF is also discriminating against women on the ground of sex, considering their limitations are stricter than for men. If testosterone is to be the decisive factor for determining sex in international athletics events, it may even be discriminatory against men to not allow those men with low testosterone to compete in women’s events.59

Secondly, if the IAAF is indeed discriminating against intersex and/or women athletes, any court or tribunal determining the substantive legality of the DSD Rule would likely need to consider whether this policy is necessary and proportionate in pursuing legitimate aims - as is the standard legal approach when human rights are undermined or clash.60 As already noted, this type of proportionality approach underpinned the 2015 Dutee Chand decision by CAS. The IAAF would likely argue that the aim of the DSD Rule is to ensure fairness for competing female athletes at international athletics events, such that any "unfair advantage" is removed. While this is a legitimate aim, the means used to achieve this end may not be convincing nor proportionate. The IAAF would be questioned as to why 5.0 nmol/L is the specific benchmark set, why the DSD Rule applies only to track events from 400m to one mile and not in hammer throw or pole vault where testosterone has an even greater performance-related advantage,61 and indeed as to whether high endogenous testosterone is scientifically proven to create a substantial athletic advantage in female events. The significant individual harm62 the DSD Rule causes to athletes such as Caster Semenya renders the proportionality requirement even harder to satisfy.

If the DSD Rule is, though, proven proportionate, the broader question of whether the rule is necessary in pursuing a legitimate aim will be the next difficult hurdle to jump over. In considering the legitimate aim of removing "unfair advantage" from competitive athletics, the concept of "advantage" would be distinguished from "unfair advantage". Taller athletes tend to have an advantage over shorter athletes, yet few people would argue this is "unfair" - perhaps because height differences have been an ever-present sporting fixture, while intersex and transgender athletes competing openly has been a relatively recent phenomenon. In contrast, weight categories are common in sports such as boxing in order to ensure a "fair" fight and the safety of participants. To prove the DSD Rule is necessary to achieve the aim of "fairness", therefore, the IAAF would need to establish why higher naturally-occurring testosterone levels, which are as physiologically intrinsic to a person as most other athletic features, are "unfair". As one of the key features of a successful sporting career is using physiological attributes, whether acquired genetically or through rigorous training, to one’s advantage, such an argument would be nuanced. As Buzuvis notes, "[s]aying that no one can use natural advantage is antithetical to sport."63 The IAAF’s argument may well be, then, that natural advantages apply only within each sex category, and that high testosterone levels unfairly cross this boundary for the female sex category. However, drawing again on norms of non-discrimination on the ground of sex, if naturally-occuring physiology is used as a distinguishing feature for women such that they become "unfairly" advantaged, why would naturally-occurring physiology64 not be used as a distinguishing feature for men? Such questions may well turn on scientific data that, as noted above, simply are not yet available to us.

Finally, while international athletics remains segregated into only two categories characterised by sex for most events,65 excluding female-identifying athletes from women events is unjust and could lead to further liability under torts law for health-related concerns. While the DSD Rule does not explicitly exclude intersex athletes from being able to compete in "restricted events", health risks associated with testosterone reduction may well have the practical effect of excluding many intersex athletes from competing at international women events. This is particularly so in the absence of a separate category under which they could compete, with the IAAF flagging a potential new "intersex" category in the DSD Rule.66 While athletes can undergo treatment to lower their testosterone levels to below 5.0 nmol/L, this crude and dogmatic approach to a complex medical issue can cause significant issues to individual athletes and their health.67 Indeed, many researchers have discovered, and expressed significant concern at, adverse health effects caused by treatment where testosterone levels are reduced below those naturally produced and required by an individual’s body.68 Forcing athletes to do so in order to compete in female international events could therefore open up the IAAF to negligence suits by such affected athletes, were their health to deteriorate or suffer as a result of testosterone-reducing treatment.69

Concluding remarks

While any female athletes with testosterone over 5.0 nmol/L could seek treatment to reduce this level in order to compete under the IAAF’s DSD Rule, the health consequences of doing so may exclude many from being able to undertake this course of action. Exclusion from competition is arguably the most severe punishment that any athlete could face in a sporting context; such an intervention must only occur for the strongest of justifications and in the most extreme of cases. Moroever, in an era where we repeatedly combat doping in sport and seek to exclude those who enhance their performance by pharmacological means it is highly ironic that some athletes could be required to pharmacologically alter their body in order to be permitted to compete.

Far stronger reasoning than that presented by the IAAF and the research it has relied upon is required before we allow the IAAF to bar a group of athletes from being able to compete in their chosen events. In the absence of further justification by the IAAF as to why the DSD Rule applies in such a targeted way, and in a developing world where social norms now reflect a more fluid definition of sex and gender, the fall-back position of international athletics must be non-discrimination. While the Olympic motto reads "citius altius fortius", meaning "faster, higher, stronger", it is another Latin maxim that must prevail when assessing intersex eligibility in international athletics: primum non nocere, "first, do no harm".

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.

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.

Liam Elphick

Liam Elphick is a Lecturer at the University of Western Australia Law School, and has worked in Australian Football League (AFL) scouting, recruitment and list management for the past decade. His research focuses on anti-discrimination law, particularly in the sporting context, and he is a member of the Discrimination Law Experts Group.