Race Radio said:
http://www.velonation.com/News/ID/1...d-to-bio-passport-experts-after-May-2009.aspx
This is very disturbing. Armstrong's most questionable blood values, the ones that USADA say show he doped during his comeback, were never submitted to the BioPassport panel
Indeed, very disturbing. It strongly suggests what I have thought for a long time: that the passport software—because, as usual, it’s set to minimize the possibility of false positives—allows a lot of highly suspicious values to slip through.
Hoggie posted a link to a very interesting video yesterday, in which LA’s 2009 Giro and TDF values were analyzed and compared. The gist of the presentation was that his Giro values appeared more or less normal, with HT dropping during the three weeks, presumably due to plasma expansion; while HT did not drop during the TDF. The latter results, coupled with very low reticulocytes, were interpreted as strong evidence of blood transfusion. This is what all the analysts mentioned in the Velonation story have been pointing out.
The nub of the problem, I think, is this. The passport is basically an extension and further development of the off-score, which is a ratio based on HT and reticulocytes. To trigger a warning, this ratio has to exceed a certain level, which usually means reticulocytes have to fall below a certain level. As the analyst in the video RR linked to yesterday pointed out, while LA’s 2009 TDF retics were very low, they were above the level that would normally result in a ratio above the critical level. In fact, IIRC, none of LA's posted blood values following his comeback indicated an off-score > around 110, well below the 132+ that used to be considered a warning flag.
This is where EPO use comes in during transfusion. Transfusion suppresses reticulocytes, but by micro-dosing EPO (probably at doses, I add in passing, that would NOT be detected by that new EPO test discussed in another thread), one can keep the retics out of the danger zone. And thus not get flagged in the passport.
Another way of putting it is this: the passport compares ongoing parameters with a baseline previously constructed from older data. You get flagged if your values deviate too much from this baseline. But it’s possible to stay close to the baseline and still transfuse during a GT, because what transfusion during a GT is designed to accomplish is not raise your HT, but just stabilize it. Just keep it from falling as it normally does over the three weeks. And with the help of EPO, you also stabilize your retics, keep them from rising as much as they otherwise would when you make these HT-stabilizing transfusions.
This is very clear in that video. LA’s HT did not rise very much during the 2009 TDF, nor did his retics fall very much. So they stayed within the baseline. But they should have changed over the three weeks. This kind of change, as far as I can see, is not built into the passport program.
So who’s at fault here? I have trouble believing UCI suppressed a positive, because if the software had flagged his values, it would have been picked up by the Passport Management Unit, which supposedly is independent of the UCI. It seems that the passport software simply is not capable of detecting certain kinds of suspicious values.
Much has been made of the analysis by Christopher Gore that reportedly showed that LA's 2009 TDF values had a one in a million chance of resulting from normal fluctuations, not involving doping. I have always questioned this analysis, because I never saw any actual statistics demonstrating this published. According to VN:
When Prof. Gore compared the suppressed reticulocyte percentage in Armstrong’s 2009 and 2010 Tour de France samples to the reticulocyte percentage in his other samples, Prof. Gore concluded that the approximate likelihood of Armstrong’s seven suppressed reticulocyte values during the 2009 and 2010 Tours de France occurring naturally was less than one in a million
The implication of this is that there was a significant deviation from the baseline. Then why didn't the passport software pick this up? I can only guess that Gore used a different kind of approach. The passport software, as I understand it, compares each new value with the baseline, and determines whether it exceeds it. If no individual value exceeds the baseline, the profile is deemed normal or not suspicious.
Gore, I think, determined the possibility that a set of values could deviate from the baseline. So even though each value in isolation did not significantly deviate from the baseline, the probability that every individual value in a group of values, all during a GT, would come close to the baseline was determined to be highly significant.
This is my speculation. But in any case, Gore almost definitely used a different statistical approach. This raises an obvious question: why isn't this kind of approach used in the passport program? Why isn't UCI modifying its software? And why aren't Gore, WADA, Ashenden, et al. pointing out the need for this modification?