Hobgoblins and Little Minds

Back when I lived in Ann Arbor there used to be a morning show that had a character named “Dick the Bruiser” (real name: George Baer) who used to do a segment around 8 a.m. called “Meet The Bruiser” which was a parody of certain news shows, like Nightline.

Bruiser had some hilarious lines, oftentimes puns and double-entendres. And you knew you were going to get a good skewering of someone or something whenever Bruiser asked, “What does this mean?”

Looking at the numbers from some documents emailed to TBV by one M. Ferret an interesting picture emerges. What, you might ask? That Floyd Landis is telling the truth. (You can skip the “science” stuff, if you want to jump to the conclusion. — Pun intended.)

If you’re not inclined to look at the documents (or just don’t have time), I’ll break it down for you:

On page 2, there are two T/E ratios written: 4.9 and one that’s either 14.4 or 11.4

On page 3, we get the raw data behind the second of the two possible values. Testosterone = 45.4 and Epitestosterone = 3.9 for a ratio of (let me bang this out on my handy dandy HP-25) 11.6 and some change. Interestingly, if the T = 44.5, then the ratio would work out to (what do you know?) 11.4 and some change.

So now we have an idea of what the actual data might be and where the T/E ratio comes from. So … what does this mean? (You can probably guess who’s going to get skewered, but I’ll leave that for later.)

There’s another important discussion to have here rather than the math of it all, and that’s what the numbers mean. By the rules, this T/E ratio looks pretty damning. After all if the max allowed is 4:1, we have a result here that’s almost three times that. Certainly that’s clear evidence of wrong-doing, right?

Not so fast. Let’s take a look at what a normal level of T in a urine test would be. According to Dr. Olivier Rabin, the Chief Science Officer for WADA in an interview published at Medscape,

You can look at testosterone itself, and if you find values way above, say, 100 to 200 ng/mL, then it’s quite suspicious.

So now we have an idea of where the upper limit for normal in WADA’s eyes is: somewhere less than 100 ng/ml. So Floyd’s T is less than half that. Let’s delve further.

Taking a look at WADA’s own document that describes reporting procedures for the T/E ratio, these statements can be found:

It should be borne in mind that there is significant variation between individuals. A normal level for one individual may in another be elevated and be consistent with doping. … The concentration of urinary steroids such as testosterone and epitestosterone varies greatly between individuals and also depends upon the specific gravity of the urine Sample; only values corrected for a specific gravity value of 1.020 can be compared.

The procedure goes on to say:

It is recommended that a urine Sample in which any one of the following criteria is met during the Screening Procedure, be routinely submitted to the IRMS analysis:
i) T/E value equal or greater than 4;
ii) concentration of testosterone or epitestosterone … greater than 200 ng/mL;
iii) concentration of androsterone or etiocholanolone … greater than 10,000 ng/mL;
iv) concentration of DHEA … greater than 100 ng/mL

And further:

Any result that will be used to support an Adverse Analytical Finding shall be confirmed and quantified.

Confirmation of elevated T/E values, concentration of testosterone, epitestosterone or any other steroid metabolite under consideration is to be performed in triplicate.

The documents M. Ferret provided don’t show the confirmation in triplicate, at least as best as I can tell. But perhaps that’s detailed in other pages of the lab’s report.

So let’s circle back to Dr. Rabin’s statement in the Medscape interview. According to the actual protocol, someone whose testosterone level is between 100 and 200 ng/ml wouldn’t automatically get tested unless the T/E ratio exceeds 4:1. This implies that an E value of up to 200 could be considered normal, as it is is possible to have a T/E ratio of 1:1 where both values are equal to 200 ng/ml.

Let’s talk about the amount of mass being measured here. A nanogram (ng) is one billionth of a gram. That’s a pretty small amount. Landis’ E value is 3.9 ng/ml, or 3.9 billionths of a gram per ml of liquid. That is exceptionally small, which perhaps accounts for the greater range of uncertainty in the forms, but also makes me wonder: Just how sensitive are these tests? Can they reliably detect such a small concentration? And, of course, it raises the question of what, physiologically, would cause such a low reading.

Here’s an interesting tidbit from another WADA document, describing how they chose to develop the testing protocol methods:

This will enable us to screen large number of samples within a short time at relatively low cost and without the need for highly skilled personnel.

So we have a lot to chew over, including this last little tidbit, that the test they developed was designed to be cheap and easy to perform. But the idea that it wouldn’t need to be performed by highly skilled personnel is disturbing. People’s reputations and livelihoods are on the line here. If it were me being tested, I’d want the person doing the work to be very highly skilled, to ensure the accuracy of the results.

In researching what a normal value for T in a urine sample would be, I’ve not found a simple statement that it’s X or Y or Z. What I have seen in various charts and graphs of a number of manufacturers of these types of tests (and I have someone who’s a microbiologist very well versed in endocrinology looking further — I’ll update this post when the information comes in) suggests that a normal reading for testosterone in urine falls at about 60 – 80 ng/ml. That means a normal E (for a 1:1 ratio) would fall in the same range.

OK, so after all this science stuff, let’s put it into common English. Landis’ T was 45.4 according to the documents, and is not suspicious in and of itself. From what I gather, normal falls from around 60 – 80. And WADA itself doesn’t consider T by itself to be suspicious until it’s at least 100 or more (200 as listed in their procedure).

So, on this count, Floyd is telling the truth. His T was normal (actually, perhaps even a little below normal). And the E value is much lower than normal. To remain within the 4:1 range, a sample with a T of 200 (remember, this isn’t by itself suspicious), E would be as high as 50 and be considered normal. Landis’ E value is much lower than the expected value. So Landis was telling the truth on that one, too.

But WADA leaves themselves some wiggle room in the statement about reporting procedures, above. Otherwise, someone using a bit of critical thinking might look at the Landis numbers and determine that he’s not cheated. Clearly, those who wrote the procedures and those who are charged with implementing them aren’t using any critical thinking in how these tests are supposed to be applied. Landis’ results suggest that they need to look at why the E was low in order to determine why the ratio was high. What’s going on now is a perfect illustration of the phrase, “A foolish consistency is the hobgoblin of little minds.”

WADA’s statement says, in effect, that the science behind the T/E standards is not good enough to determine who is a cheat and who isn’t. And, although it’s been well documented that the natural variation in T/E values in men can go as high as 10:1, they’ve chosen to tighten their threshold to 4:1 (down from 10:1 and then 6:1) to use for further testing. So they use the IRMS test in order to determine whether someone has really doped with testosterone. Now, we haven’t seen the data on the IRMS test, but I wouldn’t be terribly surprised if the results are a little less certain than has been portrayed up to now.

And I wouldn’t be surprised if the IRMS test itself has more uncertainty to it than has been commonly portrayed up to now. But I’ve got to ask, what kind of Mickey Mouse organization is this, if they can’t find tests that are more reliable than the ones they’re currently using? The statement that there’s a wide range of variation in results says that what they’re testing for — or the way they’re testing for it — isn’t a good indication of whether or not a person is cheating. They need to do better.

If the Ferret documents are real (M. Ferret left a comment over at TBV that the interpretations of his material were accurate) and if the data within them is real, then Floyd Landis’ T value really does fall into the normal range for a test result from a urine sample. And he was telling the truth when he said that the E value was abnormally low.

Now, we don’t know who M. Ferret is, or what his agenda is, or whether these documents are in fact part of the lab report that came from LNDD. And we don’t have any documentation from the IRMS studies done on either the A or B sample. Assuming the documents are real, then from where I see it, Landis’ case is much stronger than certain wind-bags and certain organizations would lead us to believe.

What all this points to, is that the procedure itself needs some revision. And WADA, itself, may need an overhaul. Or perhaps they just need change of leadership. It’s too late for Floyd Landis, but it’s not too late to prevent further travesties from being committed.

It’s a damn shame that none of this is being reported in the mainstream media. They’re dropping the ball on this story, big time.