South Africa's Caster Semenya was prevented from competing in athletics events for a year after tests showed she had unusually high androgen levels. Credit: David J. Phillip / AP Photo

Female athletes may not be eligible to compete as women if they have natural testosterone levels in the male range. That's the upshot of new guidelines on female hyperandrogenism, recommended by the International Olympic Commission (IOC) on 5 April and accepted by the International Association of Athletics Federations (IAAF) on 12 April. The rules have been welcomed by experts as a reasonable compromise, but there remain some doubts over how they would work in practice. Nature looks at the science behind the tests.

Why have the rules been brought in?

The regulations cover female athletes with hyperandrogenism, a condition in which the body produces higher than normal levels of hormones called androgens, particularly testosterone. This can cause the development of bulky muscles, perhaps giving athletes an unfair competitive advantage.

The issue hit the headlines in 2009, when the South African athlete Caster Semenya was asked to undergo sex testing after winning the 800-metre final at the Berlin World Championships in Athletics (see 'A question of sex'). She was unable to compete for nearly a year, but has now returned to the track. The results of her tests remain confidential.

Officials deny that the new rules are a direct response to Semenya's case, however. There have been other controversial cases, and they say there has long been a need to clarify the rules and procedures involved. This week's announcement is the culmination of an 18-month review carried out by the IOC Medical Commission and the IAAF, including a meeting held last October and attended by scientists as well as athletes, bioethicists, human-rights lawyers and a representative of the intersex community.

What are the rules?

A female athlete with hyperandrogenism who has testosterone levels in the male range (as measured by a blood test) will not be eligible to compete as a woman. The IOC has not yet decided on the cut-off levels, but the normal range of total testosterone for an adult premenopausal woman is typically defined as 15-70 nanograms per decilitre, compared with 260-1,000 nanograms per decilitre for a man.

There are exceptions, though. The most common cause of extreme hyperandrogenism is androgen insensitivity syndrome (AIS). In these cases, an embryo is genetically male but lacks a fully functioning receptor for testosterone, so does not respond normally to the hormonal signal to become male. In typical cases, she develops as a female — although with internal testes instead of ovaries. The IOC and IAAF concluded that, because such women are resistant to androgens, they gain no competitive advantage from their high testosterone levels and have exempted them from the ban.

"If you have AIS, you should still be able to compete," says Malcolm Collins, a medical biochemist specializing in sports medicine at the University of Cape Town, who was not involved in drawing up the guidelines.

Women with testosterone levels that are high but below the male range – as commonly occurs with conditions such as polycystic ovary syndrome – would also be unaffected by the ban.

The IOC says that an athlete banned from female competition under these regulations would not be eligible to compete as a male.

How will the tests work in practice?

The IOC says that high testosterone levels found during drug testing, or during a routine blood test, could trigger an investigation. A panel of international medical experts would then review each case anonymously and make a recommendation.

One potentially thorny issue is that many cases of AIS are 'incomplete', meaning that the person does have some sensitivity to the hormone. It is unclear how experts will assess the level of an athlete's sensitivity to testosterone, let alone the extent to which this might give them a competitive advantage, points out Myron Genel, an endocrinologist at the Yale School of Medicine in New Haven, Connecticut. Genel advises the IOC Medical Commission on a range of issues, including the eligibility of transsexual athletes.

But despite the uncertainties, the guidelines have been welcomed as a reasonable compromise. "It makes the playing field fair," says Collins.

How many athletes might be banned?

Disorders of sexual development, in which there is ambiguity between a person's chromosomal and anatomical sex, affect around 1 in 5,000 people in the general population, with AIS being the most common1. But the proportion among athletes is much higher. According to Collins, one in 421 athletes across five Olympic Games were identified as having AIS1. At the 1996 Atlanta Olympics, 8 of 3,387 female athletes tested positive for the male-determining SRY gene. All were ruled to have partial or complete AIS and allowed to compete1.

So the regulations are likely to clarify the testing procedure and ensure anonymity for women with AIS, rather than ban any significant number of athletes. "It ought to at least eliminate the stigmatization of certain women who people feel 'don't look quite right'," says Genel.

Why not accept different androgen levels as natural genetic variation?

Some experts, including Genel, argue that, as other kinds of physiological variation, such as height or oxygen-carrying capacity, are accommodated in sport, perhaps natural variations in hormone levels should be accepted too. Others, such as Collins, argue that androgen levels are the main reason for the difference in men and women's sporting performance, and so it makes sense to take these levels into account when deciding eligibility.

Collins and Genel agree, however, that despite the high numbers of female elite athletes with AIS, there isn't any direct evidence that such disorders give them an unfair advantage. "It's impossible to test," says Collins.

Credit: David J. Phillip / AP Photo