Cycling: The Tour de France winner Floyd Landis's defence against his initial positive test for testosterone was undermined yesterday when the New York Times reported that a source within cycling's governing body had confirmed reports that synthetic testosterone had been identified in the urine sample delivered by the rider after his victory in stage 17 of the race.

Tests on the B sample are expected to be completed by Saturday.

In the initial analysis of Landis's urine sample last week, testosterone and its "shadow", epitestosterone, were found to be present in a ratio of 11-1, when the World Anti-Doping Agency (Wada) guideline says a ratio of above 4-1 may indicate that testosterone has been ingested.

A source within the International Cycling Union (UCI), who had access to the test results, told the New York Times that, in addition to checking the ratio, France's national anti-doping laboratory in Paris had carried out an analysis of the carbon isotopes within the sample, which enables synthetic testosterone to be spotted.

This would contradict Landis's line of defence last Friday against his positive test, when he said that the abnormal ratio of testosterone in his urine sample was entirely due to natural causes: "The levels I have had during the Tour and all during my career are absolutely natural and are produced by my own (body). We will explain to the world why this is not a doping case but a natural occurrence."

Speaking on the Larry King Live show, he added: "What this test shows is an abnormal ratio of two naturally occurring substances. It's not a positive test in the same criteria of finding something exogenous (not produced in the body) in the body."

Landis strongly denies any use of doping procedures or banned substances.

His lawyer, Jose Maria Buxeda, strongly hinted last Friday that in his defence they might attempt to prove the isotope test used to identify exogenous testosterone (see panel) was fallible, claiming the test "is so unreliable, scientifically, that most labs do not use it".

There are parallels with the case of Landis's fellow American Tyler Hamilton, who also attacked procedures when he tested positive for blood doping in September 2004. But after an appeal by Hamilton to the Court of Arbitration for Sport his positive was confirmed.

Dr Christiane Ayotte, head of the anti-doping laboratory in Montreal, said yesterday that the isotope test has been used "routinely" by about half the Wada-accredited laboratories worldwide since its adoption in 1998.

The test is also widely relied upon for other applications than anti-doping and to isolate other naturally occurring steroids.

Landis was tested six times in the Tour before giving the sample after stage 17, but his personal doctor, Brent Kay, has confirmed to the New York Times that this test found a ratio between testosterone and epitestosterone of 11-1.

The UCI yesterday confirmed that it had applied on Monday for the second test on Landis's samples to be carried out to ensure the process was completed before the Paris lab closes this weekend for the summer holidays.

"The longer it goes until the B sample is tested, the more speculation and the more denial," said the UCI's Irish president Pat McQuaid. The test takes approximately two and a half days.

If Landis's B test is also confirmed as "abnormal", then the result will be forwarded to the US Cycling Federation which, subject to any appeal, will be expected to take disciplinary action.

Landis faces a two-year ban from the sport and the loss of his Tour de France title.

• Guardian Service

How test works

Synthetic testosterone replicates the naturally occurring steroid, produced in the body, from cholesterol of mixed origin from various plants and animals.

Synthetic steroids are derived from sterols, which come from a single plant source - soya for example - and thus have a different isotopic profile.

The Isotope Ratio Mass Spectrometry test used to identify synthetic testosterone measures the carbon isotopes in the sample, the most abundant being C12, with C13 and C14 also present. A urine sample that is low in C13 indicates exogenous testosterone. The test is also used to identify the synthetic versions of other naturally occurring steroids such as DHEA and nandrolone.

There is no financial incentive for pharmaceutical companies to make testosterone from more than a single plant but there is for laboratories that want to produce steroids that will get past drugs tests. Hence designer steroids.