In Different Columns yesterday, we passed on the breaking news that LNDD did not use the same GC column type on the Mass Spec that they used on the IRMS on every test done in the Landis case. They used an Agilent 19091s-433 on the MS, and a DB-17 on the IRMS. We'll call the one used for the MS the '433 from here on out.
Comments since have wondered whether this
- matters at all to results
- matters at all to the legal case
- late notice is a failure of the Landis defense team;
- late observance is indication even Landis thinks it doesn't matter.
(We'd assume it is raised in the filed appeal brief, which we'd like to get released. USADA has it by now, don'tcha think?)
As to point 2, legal implications, let's look at the page that identifies the DB-17 as the SOP column for the MS, LNDD 664 from Exhibit 84.
The original LDP (the USADA pages) do not ever contain the method description page for MAN-52, nor was it produced during discovery, though it was requested. It only appeared in the Exhibits for the B sample tests. Was this an intentional omission in the LDP, and an accidental inclusion in the exhibits? We'll probably never know.
The legal implication seems direct. Their own SOP produced as documentation of runs says they should use the DB-17, but the run results show them using a different column. This would appear to be an ISO and ISL violation.
Baker says in mail to Mr. Idiot that USADA 104 also documents MAN-52 as the SOP for the MS to be used in the CIR test:
Baker alludes to COFRAC accreditation documents for the EC31 test identified above as specifying the DB-17, but we do not have a copy of that document to show.
Now, on to question 1, does this make any difference to the results?
The claim is that columns of different types can change the order of elution of compounds. Baker offers in the mail to Mr. Idiot the following support for that claim.
The Agilent website documents the columns and gives reference chromatographs for each. We're shown acenaphthylene eluting before acenapthene with the '433, and in the other order for the DB-17.
Baker notes for nitpickers that acenaphthalene and acenaphthylene are synonyms, and summarizes for us idiots:
|(%-Phenyl)-methylpolysiloxane||5%||50% virtual |
|Acenaphthalene vs. acenapthene||Elutes before||Elutes after|
OK, that's one example, but does that prove anything for the testosterone test? We don't know we have any acenaphthalene or acenapthene, so maybe it doesn't matter.
Baker doesn't have indisputable proof of switches with these columns of compounds known to be present in the samples. He does offer numerous examples of switches using similar columns, in particular a set from a drug screening test. He says the DB-5 is about the same as the '433, and pulls this from Agilent
The ordering on the DB-17 vs the DB-5 is:
|DB-17 || 1-2-3-4-5-6-7-8-9-10-11-12-13-14-15-16-17-18-19-20-21-22-23-24-25-26-27-28-29-30-31-32-33-34 |
Is this ordering relevant? There are some huge differences in the ordering above: DB-17 peak 4 is DB-5 peak 16, and DB-5 peak 3 is DB-17 peak 13.
Baker finds a case from Skogsberg et. al with steroid analysis:
In this example, steroid analytes 2 and 3 get switched. Ayalyte 1 is norethisterone, 2 is norethisterone acetate, 3 is chlormadinone acetate and 4 is testosterone acetate.
Let's pretend, for argument, that the column switch did not affect the elution order of the 4 primary analytes in the Landis F3, the5aAC, 5bA, 5aA, and 5bP. We don't know that, but let's pretend.
Given the switches seen in the examples above, how can we be sure that any of the other compounds haven't been moved, perhaps a lot, between the columns? Remember the big changes in the DB-5 vs DB-17 example above.
How can we be sure that any of the peaks in the IRMS are properly identified, and that they do not contain co-elutes that have arrived solely from the column shift? It is not clear that even full-scan mass-spec of the MS portion of this test will tell us anything about the peak contents in the IRMS. Given this column change, it would appear necessary to get full-scan MS from the column before combustion in the IRMS, and this wasn't done.
If the change in column is an ISL violation that causes a burden flip, it looks difficult to prove that the reported measurements were taken correctly, unaffected by the ISL violation.
That is why Baker considers this a "case dispositive" error.
Now, he thought the retention time argument was good too, and we saw how that was dodged. We'll have to see if there are holes in this argument as well.