Far behind and suffering from a painful arthritic hip as he approached the end of last month's Tour de France, Floyd Landis made a miraculous comeback to win the race.
But wait - a urine test strongly suggested that he had taken testosterone, a forbidden anabolic steroid hormone. In fact, evidence that Landis took testosterone has now been found in two separate urine specimens obtained on the very day of his dramatic comeback. What could have happened?
Analytical techniques can easily detect the presence in urine of synthetic anabolic steroids like nandrolone, but it's far more difficult to identify injection of a hormone that's also made naturally by the body, such as testosterone or androstenedione.
Testosterone doping is detected by measuring the ratio of two substances, testosterone and epitestosterone, in a urine sample. Because both of these hormones are produced by the body in roughly equal amounts, similar amounts are normally excreted in urine.
When this happens, the ratio of testosterone to epistestosterone is close to 1. If testosterone is administered, the urine will contain more testosterone than epitestosterone and the ratio will be considerably greater than 1.
The World Anti-Doping Agency (WADA) has ruled that a ratio greater than 4 is evidence for administration of testosterone. Both of Landis' urine samples had a ratio higher than 4. Landis reportedly drank some alcohol the night before his urine test, and it's true that alcohol can cause the ratio to rise. While it's also possible that a sudden huge surge in the body's production of its own testosterone could explain these findings, that seems highly unlikely.
WADA is conducting a more sophisticated test using the powerful analytic technique of gas liquid chromatography combined with mass spectroscopy. This test measures the ratio of carbon-12 to carbon-13 isotopes in urinary testosterone to distinguish between naturally made and injected testosterone.
Because it is made from plant substances, synthetic testosterone contains a different amount of the heavy carbon-13 isotope than natural testosterone. As a result, a different ratio of these isotopes in Landis' urine would indicate whether he did or did not inject testosterone.
A similar test could have been used to determine whether home-run-hitting baseball star Mark McGwire injected androstenedione, another androgenic steroid made by the body, but the major league owners and players had not agreed to allow such testing before McGwire played his last game as a professional in 2001.
Credit (and blame?) goes to both sides, historically.
Since Landis passed drug tests on seven other occasions during this summer's Tour de France, any possible administration of testosterone must have taken place on the day before the abnormal sample was detected.
It's hard to understand why Landis might have injected testosterone at that time. Chronic administration of testosterone, not a single high dose, increases muscle mass over time, and probably muscle strength as well. But testosterone is not known to produce the rapid boost in endurance a cyclist might want for one day's winning edge.
There may be a long delay before there's a final decision on the legitimacy of Landis' victory.
