Floyd Landis and the chemistry of anti-doping
Thomas A. Fine
I've been doing a lot of research on the isotope test that was
used on Floyd Landis. A summary of my what I've found is here:
http://hea-www.harvard.edu/~fine/opinions/testosterone_d13C.html
The gist of it is, from the beginning, what I was reading about
the test seemed like it could be influenced by diet, and such influence
could be exaggerated in the metabolism of an elite athlete.
Some people at first accused me of hand-waving until I pointed out
that WADA sponsored a study to determine the influence of diet on
this (at the time proposed) doping test. I hadn't been
able to find the study results, other than an abstract that's
referenced on the above web page.
Until now. I've just come across a slide presentation from that study.
It's here:
http://www.bgc-jena.mpg.de/service/iso_gas_lab/gasir2005/presentations/ASI2005_U_Flenker.pdf
So one question I have based on the study is this: the
metabolites of testosterone are androsterone and etiocholanolone.
It's my understanding that these two compounds have the same
formula and structure, and differ only in the spatial orientation
of a hydrogen atom. If this is the case, why would their carbon-13
isotope ratios be anything but identical? And if you accept that
some process of metabolism tends to create a difference, why would
the difference be so variable?
A test will be considered a positive indication of doping if the
difference between a testosterone metabolite, and a reference
metabolite is greater than 3 per mil. I also assume it has to be 3 per
mil in the right direction - only a turn towards less carbon-13 in
testosterone should indicate doping.
So my question here is sort of the big one - does this study really
back up the notion that 3 per mil is a good cutoff? Because to me, it
looks like the test subjects came awfully close to this a few times.
From the graph on page 12, there was at least one difference greater
than 3 per mil in Female 2, although in the wrong direction to indicate
doping.
I find this high variability of the results to be particularly
disturbing considering that the latest news is that they tested Floyd
Landis testosterone against four different reference metabolites and
only found one failure. (For the record, prednanediol, 11
OH-Androsterone, 11 OH-Etiocholanolone, cholesterol, and prednanetriol
are listed by the World Anti-Doping Agency as reference metabolites).
But after all, I'm not a chemist or a biologist or a doctor. I've
learned a hell of a lot, but I'm hoping to find people who know more
than I do that can comment on this study.
Thanks,
tom