Friday, May 18, 2007

Hearing - Friday Goldberger

Because of scheduling issues, we're getting Landis witness Dr. Bruce Goldberger.

Dr. Goldberger is a Professor and Director of Toxicology in the Department of Pathology, Immunology and Laboratory Medicine in the College of Medicine at the University of Florida in Gainesville.

DIRECT BY JACOBS

q: qualifications.
a: ...Forensic toxicology.

q: experience in anti-doping.
a: ... Univ Florida professor of toxicology...

[more]


q: what do you do on a day to day basis.
a: director of a forensic laboratory, blood urine, tissue samples,...

q: do you use gc and gcms?
a: every day?

q: relevant certifications.
a: diplomate of american ... toxicologists.

q:
a: president of american society of forensic science in coloroado springs.

q: other?
a: helped draft the nccls.

q: what is the nccls?
a: national clinical chemistry laboratory standard.

SUH: Papp is talking to the press. I thought we were under a gag?

BRUNET: parties, not witnesses...

q: work with Dr. Bowers.
a: he was the chair.

q: journals
a: editor of the journal of forensic toxicology.

q: how many times have you testified in court?
a: 120-135 times.

q: what in?
a: state's witness in dui and other cases.

q: any other consultiong?
a: consultant to the NFLPA, since November 2006.

q: what on?
a: drug testing issues.

q: consulting business?
a: yes, allowing me to work on case.

q: have been a lab inspector.
a: yes, for the National Lab inspection, for drugs, and oncologists.

q: what did you look for?
a: most were for the national lab... chain of custody, assay quality, qualifications of people, very comprehensive, 6 months on all certified labs.

q: how long?
a: about 10 years. It takes about 3 days, can't do it now.

[ he's very qualified, I'll skip. ]

q: do you remember when you were retained here?
a: about 3-4 months ago.

q: you have an agreement, rate?
a: $300/hr.

q: time
a: 20-30 hours.

q: you were given documents?
a: yes.

q: you were given lab documents?
a: yes. large stack of 1500 pages.

q: what else?
a: WADA ISLs and a number of others.

q: general issues first. Why is lab documentatoin important.
a: the material we go through on a daily basis. I have no idea what I did yesterday a year ago, a year from now. everything is documented and retained.

q: when you were an inspector, how did you look at errors?
a: story, before we go to a laboratory, we get a data package. As an inspector, you get a feeling about what you are walking into. So those documents are very important to the inspector.

q: in your review of the documentatino here, did you notice errors that caused concern?
a: I was reluctant to become involved. I looked at some documents, late, and after my review, I saw glaring issues with how chemistry was performed, and signed on.

q: did you also notice mislabelling and documentation.
a: yes. forensic correction policy; there should be no obliteration, so it can be read in the future.

q: why?
a: because we're human and make mistakes.

[Ayotte looks Pained.]

q: let's look at some of these forensic corrections. USADA 008.
a: good example. it's not readable.

q: it may be a scanning error, but it's what we got.

[ Ayotte is working her gum hard. ]

q: you understand the sample number?
a: yes.

q: you see this other entry?
a: you can't see it on the screen.

q: so we have a misnumbering wasn't even numbering.
a: it's a fatal flaw, how do we know we have the right sample?

q: cause you concern?
a: tremendous. when I look at testing, it's the first thing I look at and it's incorrect.

q: there are a number of other pages with crossed out forensic corrections, right?
a: yes.

q: USADA 200, I counted 6.
a: theres one on column 1 3 lines down.

[ Floyd is enjoying this. ]

q: seing a page like this without any notations, concern?
a: yes. it's a consistent pattern. I had an employee with a 4 that started as a two; I counselled her to do an initialled strikeout.

q: if the response is how does that affect the results?
a: it might, we might not. Its the building of problems in the package that concerns me.

q: there are other pages?
a: more than several.

q: what effect did the totality of these corrections give you on the reliability of these results.
a: it's unreliable.

q: why?
a: it's your documentation. What I said about inspecting, the documents tell you the care.

q: what is a proper chain of custody?
a: start with collection through displosal. Include all transfers, all steps, every time handled by a human it should be documented.

q: what is the significance if there is an omission?
a: terrible. demonstrates lack of attention to documentatoin. I've seen it that the Coc is completed retrospectively, filling out forms in the afternoon after doing the work. That's wrong. You need to know exactly where it's at at any particular time.

q: did this document reveal CoC issues?
a: yes.

q: what about lab transfers. Is that something that is required for good lab practice?
a: yes. everything should be recored.

q: location at all times?
a: yes.

[ ayotte looks resigned ]

q: specific gcms issues starting with identification of E and T. Describe proper identification techniques.
a: involves retention time and mass spec criteria, the latter has full scan mode, need lots of ions; other is selected ion monitor, as done here. Typical process is 3 unique ions, with plotting using ion rations.

q: that is what WADA technical document requires?
a: yes, it says you must have three ions, and they must be analysed and evaluated, not thrown out the window.

q: anything unique about steroids?
a: no. the application of MS is just like the use for Cocaine and its metabolites.

q: on the confirmation analysis, did you review to see if it showed LNDD acquired and evalated three ions?
a: they acquired more, but their evaluation of the data has the display of a single ion 432.

q: lets look at USADA 0093. You understand this is the chromatogram for one of the A confirmations.
a: yes.

q: how many ions does it show?
a: one, 432.

q: the documentation shows they acquired others.
a: because they're operating in low resolution mode, they are acquiring 432 to 433, which may be responsible by the baseline problems in this chromatogram. So they are operating in low res mode, so it's not typical practice. That's why it's noisier than the beautiful chromatogram from Montreal.

q: USADA 86. what is shown here.
a: The calibration data for E and T, shows they evaluated one ion; they need more because it can effect the quantification.

q: USADA 0083, in their brief, said that they acquired the required 3 ions, and this page is proof.
a: yes. I don't know for sure, but there's no demonstration; there's only the printout. But the proof is in the paper, and we don't see the analysis.

q: Second A, USADA 213, this shows single ion, correct?
a: yes, same low res scale.

q: B confirmation, anything different there. USADA 277. same thing?
a: yes, this looks worse though.

q: how?
a: the chromatography is horrible.

q: get up and show if you like.

[ he gets up ]

q: what page are you holding?
a: USADA 278 which is the quantification of the E-T

gets technical.

a: I'd display all three ions at the same time to see co-elution. I'm more concerned with this Epi, admitting it is low. The chemstation has drawn integration marks.

a: I'd never do a manual integration.

q: given the issues you describe, do you have any confidence this is properly identifying E?
a: no. gives lots of reasons. This is bad chromatography. We looked at the montreal data, so we know it can be done.

q: they ran 3 of these in the B sample, right?
a: yes.

q: let's look, USADA 280. single ion?
a: yes.

q: any issues?
a: E looks the same, T peak has end peak is not strong.

q: any confidence properly identified?
a: no.

q: are there other substance that would share the 432 ion?
a: yes, I looked on my computer and got over 10 compound, and some steroid related, some not.

technical stuff.

q: which is why you need to look at more than a single peak.
a: yes. we don't know if we have T or T and some unknown compound.

q: USADA 282, single ion, any issues?
a: looks like the last one. Problems with E and T the way it's integrated.

q: based on this chromatogram, is the identication acceptable?
a: no.

q: what is the significance of the fact that LNDD did not provide the chromatograms of the ions collected?
a: they have not documented the T/E result they are claiming.

q: if you haven't adequately identified T and E you can't go forward with your analysis?
a: correct. Where cases can end up in litigation of arbitration, we have to be supportable. Minimum of 3. I know of no lab reporting quantification on a single ion in a setting like this.

q: do other wada labs give different documentation?
a: yes.

q: describe.
a: UCLA gives selected ion and full scan, the best of both worlds.

GC 524

q: is this the UCLA chromatogram?
a: yes, it has the internal standards on top, upper left; single ion; I prefer two minimum. T is in the panel below, showing 432, 433 and 209.

q: lnds does something different what is the effect?
a: don't know, we haven't studied they acquired 432 and 433 together, I don't know what that will do.

q: and UCLA shows all?
a: yes.

q: have you authored papers on this need to acquire analyse many ions?
a: yes, for years we struggle with the use of selected ion monitoring in forensic toxicology. We were comfortable, one of the first was the one was by Dr. Bowers.

Entered at GDC 58.1

q: this specifically addresses the need for three diagnostic ions.
a: yes.

others.

q: when you hire new people at your lab, what training do you provide?
a: it takes months or a year or two to get them competent. we cross train everyone. one thing they understand is chromatography. the ones from my lab are all classic, and if they are not, I'll send them back.

q: do they all go through a standardized training?
a: somewhat, but much is cookbook, but it takes a good amount of time.

q: continuing training?
a: do continuing education, weekly monthly meeting to bring them up to date, they go to conferences and meetings every year.

q: why?
a: otherwise their knowledge of the field becomes stagnant. To keep my diplomate, I need to go through continuing education.

q: Matrisx interference. what is it?
a: the chromatograms you saw before are matrix interference.

q: Any of them 272, 282.
a: yes.

q: what are the possible causes?
a: other steroids, other compounds, unknowns. that's what worries me. even doing a simple confirmation, we get them. Then we need to send them somewhere else that doesn't have it.

q: impact on identification?
a: faulty and incorrect quantification.

q: EX 24 USADA 100, control urine on A. looking a tthe peak at 19.33, what issues?
a: what's going on there? there is a shoulder. we may have co-eluting peaks, or injection ports wrong, ...

q: effect of coeluting peak on a control.
a: the quantification of the sample would be wrong.

q: USADA 93, athlete sample, internal standard chromatogram. Issues?
a: have a resolution problem. co-eluting at 21.1, not baseline separated. Looks like a 20%, 10% might be OK, but not 20%.

q: USADA 92 says the retention time is 20.98 what is the impact of this interfenc on an interla standard.
a: the target response would be incorrect, so ultimately the quantitation would not be correct. We can only guess where the peak starts and stops. The computer has done an estimate, and we need to rely on it.

q: Another A cong, USADA 213, internal standard. problem:
a: looks like an injection port problem... doesn't look like overload, something different.

q: affect of overload?
a: wrong quantitations.

q: USADA 93, LNDD 34. one is blown up. look at the E peak. issues?
a: we don't know where the integration started and stopped. sometimes we zoom, and get a wavering baseline, but shows a chromatographic problem.

q: what is the effect on the reliability of the quantification?
a: it would be unreliable, assuming it was the right stuff -- single ion.

q: USADA 277, any E peak issues?
a: same as before, except its worse. Leading baseline is higher. very uncertain about the baseline. multiple peaks mergine.

q: can you demonstrate on the screen what else might be coming in here?

example showing how drawing starts affects, and need for documentation.

q: USADA 280. any problems?
a: essentially the same. see both lines on the 18.57 peak. so it's not obliterated. I don't know if it's correct. I'd like to see the other ions, and we could identifying a co-eluting peak.

q: the blow up of the same peak, does it have any resolving power?
a: worst of the bunch. there are two or three co-eluting peaks before the epi-T peak.

q: illustrate?

a: since we don't have the other ions, we don't know where the E peak is. Might be here, here, or over here. don't know where to draw the baseline.

q: USADA 282, third replica of B conf. look at the E peak.
a: same problem, but baseline is wavering even more before and after than the other two, would be very unceratin about the baseline.

q: significance?
a: affects the area included in the quantification.

q: does the blow up help?
a: makes it worse. I don't know where the baseline is. on this one, if you go to the previous, wider view, you see the baseline is very noisy, much too high and would affect the ability to measure the E. It's only about 3X noise, making it near the Labs limit of detection. The data from UCLA and Montreal shows it can be done. This is obviously a problem.

q: what is the result of this bad chromatography on the TE results here?
a: no. It makes you question why this was sent to irms. There were problems with the screen too.

q: just to be clear, you are not offering opinions on the irms?
a: no, I am not.

q: this is represented as a reinjection with one standard, but you can see they are reporting Ad-4p, with a chormatogram.
a: this is funny. when you don't get multiple ions, the software is not smart enough to differnetiate from internal standards. It was told to look for a peak, and it found one.

q: good reason to qcuire and look at three ions?
a: yes.

q: do you have any confidence in this testing?
a: none at all?

q: in all of your 20+ years in the field, have you ever seen so many errors on a single sample.
q: No.

NO QUESTIONS.




[ Young is doing cross because Dunn isn't here, and Barnett might break some dishes. ]

YOUNG CROSS.

q: your CV. steroid mentioned only once?
a: correct.

q: of the 3000 a year, how many are uring.
a: almost all, 2700.

q: of those 2700 did you analyse for T or E?
a: very few. we send them to a reference lab?

q: you don't do that?
a: no.

q: cookbook toxicology?
a: SOP, very simple.

q: accreditations of your lab?

q: when you have a cocaine case, or an alcohol do you provide a documentation package?
a: yes, we call it a litigation file.

q: any EDFs?
a: no one has asked, but if they did for a test within the last 10 years, I could give it to them.

q: how many T/E's have you looked at?
a: I've just been solicited to apply for the job of director at the UCLA Olympic lab.

q: if I took you through the various standards...
a: yes, but there was no QC data in the packet.

q: you lab is accredited by the american colleged of pharmacology.
a: yes.

q: not ISO?
a: no need yet.

q: better?
a: not really?

q: the ISO standards are for the WADA labs,
a: yes.

q: so you haven't worked under ISO?
a: no.

q: you talked about two specific documents, USADA 200, with a number of non-forensic corrections.
a: yes.

q: deal with E and T
a: yes.

q: then USADA 008, and pointed to the fact that the sample was wrong.
a: yes.

q: and you see it has a problem with derivitiaztion, but you know it's the right sample.
a: don't know.

q: but this printout is the right number.
a: it is, yes.

q: so you confusiton is not the files generated by the instrument.
a: it's the symptom of a problem, like high blood pressure.

q: The applicable rules to a WADA laboratory are the technical documents?
a: yes.

q: EX 116(?) these are the 2006 forensic toxicology recc of SOF.
a: yes.

q: take a look at page 8.
a: ok.

q: The coc requirements are in section 7.2, right?
a: yes.

q: 7.2.5 all it says is that any transfer must be documented as part of the permanent lab record.
A: agree.

q: 7.2.6 coc for sample; separate bottle coc and aliquot coc?
a: depends on the lab. Can have cominations....

SUH: cut him off.

q: 7.2.6 it is reccomended that coc reflect...
a: correct.

q: the document your group has written is a reccomendation.
a: it's a guideline, not a basis for inspection.

q: the standard for LNDD is the ISL?
a: yes.

JACOBS:

q: the lab errors you mentioned you take as symptoms of sloppy work.
a: yes.

q: USADA 277, on the E and T peaks, you were not suggesting they could be confused,, you were saying this peak might be confuesd with this or this or anything in this whole area?
a: correct.

q: poor chromatography goes to accuracy of results?
a: yes.

q: affected by accreditation?
a: no.

NO QUESTIONS

SUSPENEDE D until 9:30 AM SATURDAY


10 comments:

James said...

Looks like the press has moved on from the GL episode...good headlines already today...

LiquidHuman said...

This testimony is very interesting. As it stands the Doping agency has been on the defensive with every witness. (except for Lemond).

This witness is crushing the doping case. Not that it matters. The Landis cannot win because the arbitrator choice makes a 2-1 victory for the prosecution predestined. Landis could only win with irrefutable evidence he did not dope. The best case scenario for him is probably a saved reputation and a reform of this clearly broken doping system. The Lemond escapade has certainly dimmed Landis' chances of preserving his reputation.

Anyway, I digress, it will be interesting to see what the prosecution can do with the witness on cross. It will be a true test of the skill of the attorneys and of their case. So far the prosecution (for lack of a better term) attorneys' have not done anything particularly well -- they lucked into the the Lemond debacle, there was no skill (or real evidence) involved there.

If there was a real trier of fact such as a jury or a neutral fact finder the case would be decided on the cross.

Chaz said...

I know that they haven't done cross yet, but that sounded like a grand slam!

Anonymous said...

ISO is for contract purposes...just to ensure the customer that you are meeting the minimum contract requirements and have a preventive and corrective action process in place.

~ Paul

Anonymous said...

The ISO standard is generic and the company who adopts the ISO standard will develop their own policies and procedures to conform to the ISO standard.

~ Paul

Anonymous said...

the doctor said he had never seen so many errors in a sample in his 20 year career, from such a credentialed guy this statement may carry a lot of weight. it sounds tough for usada to discredit, very tough

Anonymous said...

Don't even think every ISO on earth is simply one the ISO 900x.
WADA is requiring ISO17025 accreditation (and not simply a certification), which is a completely different amount of work.

Look for how many labs in the world are ISO17025 accredited and you will be very surprised.

Anonymous said...

“q: just to be clear, you are not offering opinions on the irms?
a: no, I am not.”


That kills most of what he said because T/E is just a screen. They don’t live and die by it. T/E is used to go to the second step (irms) of more precise testing. T/E crudely filters out samples with low precision.

My sense is that WADA/the lab is using one ion and this less accurate analysis method intentionally to catch a bigger net of samples for further analysis. The WADA testing procedure at the time called for T/E though since that time, they seem to have found irms so much more accurate that they’d prefer to just use it 100% of the time (though costs may discourage it) because T/E is missing some AAFs. When they used irms on Floyd’s other ‘B’ samples that had passed T/E, they found more evidence of high testosterone that the T/E screen missed.

Goldberger complained that he didn’t see more than one ion presented but at times seemed to indicate data for more than one ion was gathered during T/E test. Ayotte said something quite similar and suggested the quantification of their T/E ratio had 40% error. But it didn’t alarm her because the irms would deal with the accuracy issue. The more accurate irms second step with 3 ions sorts out or confirms if their catch with T/E was justified. The lab may counter that they used 3 ions during their testing but only reported or analyzed with one (to widen their net – even though it would widen their testing of samples that were ok and still wouldn’t catch all that were adverse).

Similarly Goldberger’s criticism of the chromatographs was only for T/E. He’s criticizing that they didn’t have a precision that he would like to see. Ayotte said something similar. But Ayotte acknowledged the screening purpose which Goldberger did not. Goldberger didn’t get into the more accurate irms. As Ayottte said, when the T/E peaks are as high as they were then much of the technical accuracy concerns are moot because the screening process has done it’s job. It’s identified samples for further analysis that will deal with the precision issue T/E can’t – even if it’s done well. As Ayotte pointed out, the peaks were so great in Floyd’s sample that you could eye ball they required further testing – they were so great that they would overcome any concern over precision during the T/E screening.

Ayotte also said that no one has been able to accurately determine the value of ‘E’. Therefore, no one has been able to accurately determine T/E. So fans need to understand that LNDD, Ayotte, WADA, USADA, etc are not hanging their results on T/E and never have. Goldberger blowing away T/E and it’s chromatographs sounds good but it’s told us little because Ayotte already told us it has a 40% error in precision. T/E is an imprecise screen. That’s all. Both sides knew that before Goldberger took the stand.

Goldberger made a big fuss about the errors but reported his reports are only 30 pages long – not the 1500 pages of evidence or the massive package provided by LNDD. He conceded on cross that the one specific error he raised of consequence was resolved with the parts of the documentation that would leave no one guessing. Goldberger also has no specific experience with lab work for testosterone – like what is going on with Floyd.

Goldberger’s chain of custody issue also was murky. He acknowledged that a standard he uses for chain of custody isn’t as stringent as what he talked about. Ayotte laid out what the rules from WADA are on chain of custody and they’ve followed the WADA rules on chain of custody.

I think Goldberger sounded good but when you roll up your sleeves and look closer at what he said and it’s relevance, there was much more smoke than substance because he didn’t go near the irms findings. It’s why the USADA lawyers didn’t waste much time cross examining him.

As bias seems to be of concern, I’ll state mine with the limits. I’d prefer that Floyd be found innocent. I don’t want cycling or Floyd to suffer the tragedy of being found guilty of doping. My limits are that I want to know the truth more. So I’m looking for proof that Floyd is innocent or that WADA/USADA/LNDD can’t prove their case or that Floyd can be proved innocent. While looking at that, I’m trying to be as objective as possible. I don’t regard LeMond’s or Papp’s testimony as providing much evidence one way or the other. I’m looking more to the science and solid facts. To be candid, so far, I haven’t seen much from Floyd’s team that firmly disproves the lab results. I’m hoping that I will but so far, I’m getting discouraged.

cleduc

Anonymous said...

Excellent analysis cleduc - thanks.

Anonymous said...

Cleduc,

You make some excellent points and I appreciate your observations very much. I do think in some ways you also highlight the important issue and that is the process. If one were to look at the case to this point they would, I think, find no revelation that says Mr. Landis did not dope because there is not one nor will there be. What does exist however is a process so flawed that it cannot possibly be used to support the assertion that an AAF exists. I guess if we need to hang our hat on one item that proves or disproves then case then it is the process. Is it the T/E or maybe the IRMS or maybe the SOP, could it be chain of custody, how about ISL, what about contamination or manual versus automatic subtraction? The process is what is on trial, if there is a flaw or a mistake in each step of the process then what will your end result be? With all due respect to Dr. Ayotte would she have given the following answer in the midst of her doctoral dissertation; “experience. If I say it's ok, it's based on scientific principles and its ok”. One example that helps a simpleton like me with the scientific process is when conventional wisdom was such that is was “known” that the Sun orbited the Earth (not relevant to Mr. Landis I now except when discussing the process). Everyone knew that Ptolemy was correct but when asked to explain and provide evidence for the motions of all the planets they could not but “they” still knew they were right, why? Because if I eyeball it I can clearly see the Sun revolving around the Earth. Galileo, building on the work of Copernicus, questioned conventional wisdom, even though Giordano Bruno had been burned at the stake for similar indiscretions earlier. Hence, what we know now is a lot different then what we believed then.

-pt