Intended for healthcare professionals


Sex, gender, and sports

BMJ 2019; 364 doi: (Published 20 March 2019) Cite this as: BMJ 2019;364:l1120
  1. Cara Tannenbaum, professor of medicine and pharmacy12,
  2. Sheree Bekker, prize research fellow3
  1. 1Université de Montreal, Montreal, Canada
  2. 2Institute of Gender and Health, Canadian Institutes of Health Research
  3. 3Department for Health, University of Bath, Bath, UK
  4. Correspondence to: C Tannenbaum cara.tannenbaum{at}

New regulations challenge the evidence based, benevolent ethos that underlies medical practice

In 2018, the International Association of Athletics Federations (IAAF) introduced new eligibility regulations for female athletes with differences of sex development, including those with congenital adrenal hyperplasia, androgen insensitivity syndrome, and 5α-reductase deficiency (box 1).1 The regulations are based on the contention that women with high levels of endogenous testosterone and androgen sensitivity have a performance advantage over their peers. Athletics South Africa and Mokgadi Caster Semenya, the South African runner who won Olympic gold medals in 2012 and 2016, are contesting the legality of the new regulations.

Box 1

IAAF eligibility regulations for female athletes with differences of sex development1

  • The regulations apply to athletes in events from 400 m to 1 mile, including 400 m, hurdles races, 800 m, 1500 m, 1 mile races, and combined events over the same distances

  • Any athlete who has a difference of sexual development, defined by a circulating testosterone concentration ≥5 nmol/L, and who is androgen sensitive must meet the following criteria to be eligible to compete in restricted events in an international competition or to set a world record in a competition that is not international:

  • She must be recognised in law as either female or intersex (or equivalent)

  • She must reduce her blood testosterone concentration to below 5 nmol/L for a continuous period of at least six months (for example, by using hormonal contraceptives)

  • Thereafter she must maintain her blood testosterone concentration below 5 nmol/L continuously (whether she is in competition or out of competition) for as long as she wishes to remain eligible


The basis for their claim is a lack of scientific evidence showing that endogenous testosterone concentrations substantially enhance sports performance. This is in alignment with the 2015 judgment of the Court of Arbitration for Sport on the case of Dutee Chand v Athletics Federation of India and the IAAF.2 The IAAF has delayed the implementation of the 2018 regulations until the outcome of the Court of Arbitration for Sport hearing on their legality is determined; this is expected on 26 March 2019.

New eligibility criteria

The new eligibility criteria for female athletes with differences of sex development—or intersex women—emphasise two unresolved sex hormone controversies: normal serum testosterone levels and physiological androgen sensitivity. The threshold that defines raised concentrations of circulating testosterone is arbitrarily set at 5 nmol/L. However, serum concentrations of testosterone vary naturally in men and women, and across athletic states. In non-athletes, serum testosterone values range bimodally from 8.8 to 30.9 nmol/L in men and boys and from 0.4 to 2.0 nmol/L in girls and women.3 But there is greater overlap among elite track and field athletes after competition, with mean (standard deviation) values 12.8 (7.9) nmol/L in men and 4.1 (4.9) nmol/L in women.4 Based on reference ranges above 8.4 nmol/L for men and below 2.7 nmol/L for women, testosterone concentrations would be categorised as low in 16.5% of men and high in 13.7% of women.4 Causal associations between serum testosterone and medal winning based on these reference cut-offs cannot be determined because these types of analyses stipulate anonymity to protect the participants.

The requirement to prove androgen sensitivity in athletes is also problematic since reproducible, valid laboratory tests to detect androgen sensitivity do not exist. To diagnose female athletes with androgen sensitivity, the IAAF mandates a physical examination, gynaecological assessment, and radiological imaging for women who have high testosterone levels on antidoping tests. There has been much criticism of this approach, particularly around examinations of clitoral size, which many view as inappropriate, subject to false interpretation, and an invasion of personal privacy.56 What is equally concerning is the lack of evidence that these physical signs are related to circulating testosterone levels, androgen receptor signalling, or the genetic regulation of muscle and red blood cell activity, particularly in women with partial or complete androgen receptor insensitivity or 5α-reductase deficiency.

The paucity of reproducible scientific data on the effect of testosterone on speed during track and field events poses an additional challenge. Given the many other variables that affect athletic performance, the 2015 Court of Arbitration for Sport panel said that although a 1% difference in event times may not justify a separation of categories of female athletes, a 10% difference in athletic performance justifies having separate male and female categories because of unfair advantage.7 No guidance exists on the grey zone between 1% and 10%.

An analysis commissioned by the IAAF to ascertain the quantitative relation between endogenous testosterone and enhanced athletic performance indicated that women whose serum testosterone levels were in the top third performed significantly better (P<0.05) than women with levels in the lowest third. The differences in 400 m, 400 m hurdles, 800 m, hammer throw, and pole vault were 2.73%, 2.78%, 1.78%, 4.53%, and 2.94%, respectively.8 However, these results could not be independently reproduced, and the data do not reliably mirror the source track times of athletes from the 2011 and 2013 world championships.7 The authors of the analysis subsequently acknowledged data errors, driving accusations of alleged data mongering and a potential breach of scientific integrity by the IAAF.7

Far reaching implications

At their core, the 2018 regulations seek to identify athletes whose lifetime sex assignment as female—and gender identity as a girl or woman—does not match an imposed female serum testosterone level ≤5 nmol/L. Beyond issues of medical governance by sports federations, the 2019 Court of Arbitration for Sport case is garnering growing media attention around gender and race discrimination, binary stereotypes of men and women according to the presence or absence of testosterone, and implications for trans women in sports.6 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.

Sitting at the intersection of biological sex and gender identity, the 2018 IAAF regulations challenge the evidence based, benevolent ethos that underlies medical practice. The medical profession does not define biological sex or physical function by serum testosterone levels alone. If more science is needed to develop an objective measure of androgen sensitivity, then call for health research organisations to deliver on this mandate. In the meantime, complacence around the IAAF 2018 testosterone regulation for women with differences of sex development risks setting an unscientific precedent for other cases of genetic advantage.

History compels us to ensure that decisions about genetic superiority are supported by objective, rigorous, and reproducible data. Although this is purely conjecture, we venture that the Olympian gods smile down on winners like Mokgadi Caster Semenya when they perform extraordinary feats of human endeavour.


  • Competing interests: We have read and understood BMJ policy on declaration of interests and declare that CT receives funding and salary support from the Canadian Institutes of Health Research. The views expressed do not necessarily reflect those of CIHR or the Canadian government.

  • Provenance and peer review: Commissioned; not externally peer reviewed.