New rules to reduce naturally high testosterone levels in female athletes have been branded "unscientific".

Last year, athletics chiefs ruled women with levels of five nanomoles per litre or more must have hormone treatment before being allowed to compete.

But experts, reporting in the British Medical Journal, say there is a lack of evidence about testosterone's effects and the cut-off figure is arbitrary.

A decision on the legality of the rules is expected later this month.

The International Association of Athletics Federations (IAAF) delayed implementing its regulations after South African runner Caster Semenya contested the legality of the new rules.

She was banned from international competitions for nearly a year for having testosterone levels above the athletics body's limit for female athletes.

World athletics bosses have previously said they want to protect the sanctity of fair and open competition.

'Unscientific precedent'

Writing in an editorial in the BMJ, Dr Sheree Bekker, from the University of Bath, and Prof Cara Tannenbaum, from the University of Montreal, say the IAAF's regulations risked "setting an unscientific precedent for other cases of genetic advantage".

"The medical profession does not define biological sex or physical function by serum testosterone levels alone," they say.

And they warned that the proposed rules could have "far reaching implications" on individuals and societies.

Dr Bekker and Prof Tannenbaum argue that testosterone levels vary naturally in men and women, with higher averages among elite athletes.

But there is also a big crossover between men and women, with 16% of men classified as having low testosterone and 14% of women having high, according to some definitions.

They say testosterone is just one indicator of sports performance and many other factors also play a role.

"If more science is needed... then call for health research organisations to deliver on this mandate," they say.

"History compels us to ensure that decisions about genetic superiority are supported by objective, rigorous and reproducible data."

The writers take issue with an analysis commissioned by the IAAF to quantify the relationship between high levels of testosterone and performance, because the results "could not be independently reproduced".

'Brilliant athlete'

Prof Peter Sonksen, professor of endocrinology at St Thomas's Hospital and King's College and visiting professor at the University of Southampton, agreed that the IAAF's proposed new rules were not "fit for purpose".

"It is not compatible with the science behind the issues and greatly overestimates the role of endogenous testosterone," he said.

"It is also personal and unfairly targets a brilliant athlete."

However, Prof Chris Cooper, emeritus professor of biochemistry at the University of Essex, said it was important not to get hung up on criticising the fine details of science that was never going to be conclusive.

"In my opinion, the current evidence is as good as we are going to get to show that both endogenous [natural] and exogenous [doped] testosterone levels enhance female sports performance," he said.

"Of course a lot of other factors, genetic and environmental, also affect sports performance but that is a separate story."