This deposition of an expert witness was taken in a metoclopramide lawsuit that alleged an individual developed a movement disorder after extended use of the drug.
Q. Do you agree that there are no incidence studies on metoclopramide-enduced tardive dyskinesia?
A. Unfortunately that is the truth, sir.
Q. Do you agree that there are no incidence studies on metoclopramide enduced-tardive dyskinesia among diabetics?
A. That is unfortunately the truth, sir.
Q. Do you agree that there are no incidence studies on metoclopramide-enduced TD among the cognitively impaired?
A. That is unfortunately the truth and that is indicting your client terribly in my opinion.
Q. Do you agree that there are no incidence studies on metoclopramide-enduced tardive dyskinesia in the elderly?
A. Unfortunately that's true, sir.
Q. Do you agree that there are no incidence studies on metoclopramide-enduced TD for elderly females?
A. Unfortunately that's true, sir, that's why the investigators that we were talking about do the studies they try to do.
Q. Do you agree that there are no incidence studies on metoclopramide-enduced TD among those persons who had prior exposures to dopamine receptor blockers?
A. That's probably also unfortunately true.
Q. Doctor, you told me before that the only available studies are prevalence or cross-sectional studies, correct?
MR. MCGLYNN: Object to form.
A. That is unfortunately true, sir.
Date: February 21, 2008.
Pursuant to Notice the deposition of ROBERT C. NELSON, Ph.D., was taken on Thursday, February 21, 2008, commencing at 9:30 a.m. at the Loew's Annapolis Hotel, 126 West Street, Annapolis, Maryland, before Linda Harten, a Notary Public.
Q. Are you an internist?
A. No, sir.
Q. Do you hold yourself out as an expert in internal medicine?
A. No, sir.
Q. Are you a gastroenterologist?
A. No, sir.
Q. Do you hold yourself out as an expert in gastroenterology?
A. No, sir.
Q. Are you a psychiatrist?
A. No, sir.
Q. Do you hold yourself out as an expert in psychiatry?
A. No, sir.
Q. You've never prescribed metoclopramide, correct?
A. That is correct.
Q. Is the first time that you studied the metoclopramide label in any detail for the purpose of this litigation?
A. That is correct.
Q. Do you recall at any point before this litigation, being contacted about this litigation, during which you looked at the metoclopramide label?
A. I don't recall, but I am sure I have.
Q. In your capacity as an FDA employee?
A. Either that or as a person that functions in the field of drug safety. Oftentimes you get home at night and you sit there with your journals. It's more a professional exercise than it is a 9 to 5.
Q. Again, you told me it's possible you read the published literature about metoclopramide before?
MR. MCGLYNN: Object to the form.
A. I told you that's possible.
Q. Isn't it possible you read the metoclopramide label before this litigation?
[Note: Pages 67-88 missing in original document] diagnose all these other people. Obviously I have run across people with psychosis that have strange movements.
I would not be able to tell you whether that particular individual was moving strangely because of an underlying disease or from the therapy for that disease.
Q. Let me go back to my question. How often do you see people out in the general community with tardive dyskinesia?
MR. MCGLYNN: Object to form.
A. You don't see that very often.
Q. Have you ever wrote or published about metoclopramide?
A. I have not.
Q. Have you ever presented or spoken about metoclopramide?
A. No.
Q. Have you done any research about metoclopramide?
A. Not before this.
Q. That would include any bench work, do you do bench work?
A. Not for a long time.
Q. Any lab work on metoclopramide?
A. No.
Q. Have you done any research on metoclopramide beyond what you have done for this case?
A. No.
Q. Have you done any pharmacological or toxicological testing of metoclopramide?
A. No, I have not.
Q. Have you ever consulted with any companies about metoclopramide?
A. Not to the best of my knowledge.
Q. Have you done any clinical research of any kind about metoclopramide?
A. No.
Q. Have you participated in the drafting of the metoclopramide label?
A. Not to the best of my knowledge.
Q. While you were at the FDA you recall no involvement with metoclopramide?
A. I recall no involvement.
Q. Doctor, have you contacted the FDA about safety issues with drugs?
MR. MCGLYNN: Objection.
Q. Let me back up and address Mr. McGlynn's objection. You have been a consultant since 1998, correct?
A. That is correct.
Q. In private practice, so to speak?
A. That is correct.
Q. Since 1998 have you contacted the FDA about concerns about the safety of certain drugs?
A. Yes, I have.
Q. What concerns have you brought to the FDA's attention?
A. Well, they were discussions that I would consider confidential.
Q. This is a confidential deposition
[Note: Pages 92-93 missing in original document]
Q. Let me see if I can get at it this way, have you since 1998 approached the FDA and discussed concerns that you have about the safety of drugs?
MR. MCGLYNN: Objection to form.
A. I've talked with individuals about the safety of drugs, I have not approached FDA as an institution in a formal manner.
Q. Doctor, is it fair to say that the opinions that you have developed in this case were developed as the result of your research and investigation relating to this litigation?
A. Yes, on top of my experience and general expertise, yes.
Q. I understand, but all the work and knowledge about metoclopramide has been Fathered and developed in connection with this Litigation, correct?
A. Yes, I would say that's a good characterization for the vast majority of it.
MR. MCGLYNN: Object to form.
[Note: Pages 95-147 missing in original document] rate of tardive dyskinesia in metoclopramide users. Where in the study does it say that?
MR. MCGLYNN: I believe he's answering the question.
MR. PILKINGTON: He is not.
MR. MCGLYNN: He is trying to.
BY MR. PILKINGTON:
Q. Do you have the study? Why don't you get it.
MR. MCGLYNN: Which one do you want to refer to?
Q. Again, just so we're clear, point to me where in the Ganzini study there is quantification of either incidence rate or risk for tardive dyskinesia.
A. The question I believe you're actually asking me is do the authors in their write up of that study use incidence or a calculated incidence rate. If that's the question the answer is it's not in there.
Q. There is nothing in the study itself that says the incidence rate of tardive dyskinesia is “x” or the risk of tardive dyskinesia is “x”?
MR. MCGLYNN: Objection to form.
A. If those authors had made such a statement in prevalence cross-sectional studies they would be inaccurate because they couldn't calculate an incidence rate from the data.
But you can estimate an incidence rate from what they have done.
Q. These are cross-sectional studies, correct?
A. Yes.
Q. And cross-sectional studies determine the presence or prevalence, correct?
A. That's correct.
Q. Cross-sectional studies do not determine the development of disease or the risk of disease, correct?
A. In and of themselves they don't.
Q. But you're trying to tell me
done a study that identifies the incidence of tardive dyskinesia from metoclopramide use in diabetics?
A. I am sorry, I am going to have to find the reference before I can focus on what the question is.
MR. PILKINGTON: Let's go off the record. This isn't going to be on my time.
THE WITNESS: I am not trying to give you a hard time and not answer the question but I need to find something to answer the question.
(Brief recess taken off the record.)
BY MR. PILKINGTON:
Q. Do you agree that there are no incidence studies on metoclopramide-enduced tardive dyskinesia?
A. Unfortunately that is the truth, sir.
Q. Do you agree that there are no incidence studies on metoclopramide enduced-tardive dyskinesia among diabetics?
A. That is unfortunately the truth, sir.
Q. Do you agree that there are no incidence studies on metoclopramide-enduced TD among the cognitively impaired?
A. That is unfortunately the truth and that is indicting your client terribly in my opinion.
Q. Do you agree that there are no incidence studies on metoclopramide-enduced tardive dyskinesia in the elderly?
A. Unfortunately that's true, sir.
Q. Do you agree that there are no incidence studies on metoclopramide-enduced TD for elderly females?
A. Unfortunately that's true, sir, that's why the investigators that we were talking about do the studies they try to do.
Q. Do you agree that there are no incidence studies on metoclopramide-enduced TD among those persons who had prior exposures to dopamine receptor blockers?
A. That's probably also unfortunately true.
Q. Doctor, you told me before that the only available studies are prevalence or cross-sectional studies, correct?
MR. MCGLYNN: Object to form.
A. That is unfortunately true, sir.
Q. Those three studies would be Sewell, Ganzini and Madsen, correct?
A. Those would be the three best to shed light on what's going on here.
Q. When you say shed light on what's going on, none of those studies tell you what the incidence rate is of tardive dyskinesia among metoclopramide users?
MR. MCGLYNN: Object to the form.
A. I'll tell you what it tells me, that it's not rare which is what the label implies.
Q. No, that wasn't my question. Do the authors of those articles state what the incidence rate is of TD among metoclopramide users?
A. They do not because they cannot.
Q. Do the authors of those studies tell you what the incidence rate is of tardive dyskinesia — I am sorry, do the authors of those studies tell you what the risk is of tardive dyskinesia among metoclopramide users?
A. By calculation — yes.
Q. They tell you what the incidence rate is?
A. No, the last question you asked me was risk, calculated risk ratios; do they estimate risk, yes.
Q. Prevalence studies don't give you risk, do they?
MR. MCGLYNN: Object to form.
A. A prevalence study that has a non-exposed arm where you calculate a risk ratio will give you some estimate of risk.
Q. What's the estimate of risk that you take from the Ganzini study?
A. That it's elevated.
Q. Can you quantify that?
A. I think they have a 1.8 with a confidence interval that overlaps one.
Q. I think it's 1.67, does that sound right?
A. That could be.
Q. And that figure was not statistically significant, correct?
MR. MCGLYNN: Object to form.
A. Given the small sample size of this it's not surprising.
Q. But it had a P value of greater than .05?
A. That is correct.
Q. Which means it's not statistically significant?
A. That doesn't mean it's not real.
Q. But under epidemiologic principles using a 95 percent confidence interval it's not statistically significant?
A. Using a 95 percent confidence interval it's not statistically significant. Using a 90 percent it probably would be.
Q. You also said that the confidence interval included one?
A. A 95 percent confidence interval in that study included one.
Q. Which tells you also that it's not statistically significant?
A. It tells you that these results cannot rule out random chance. It also tells you that random chance is unlikely.
Q. Doctor, what you've done is you have taken those — neither of those studies enable you to precisely quantify risk or incidence rate, is that correct?
A. Precisely; you're absolutely correct.
Q. Have you tried to use these studies to get an estimate of what risk is or incidence rate may be?
A. In the absence of properly done studies by the people responsible, i.e., the applicant, researchers who have a suspicion, and there are certainly many signals out there, I mean your label confirms the fact that there is an increased risk in these populations.
We are just trying to find out what the magnitude of that increased risk is in light of the non-conduct of the proper research.
Q. That wasn't my question.
A. Well, that's my answer.
Q. That's the problem we're having. Let me go back to the question that I was trying to pose. Neither of these studies can be used to quantify risk or quantify incidence rate, do you agree with that?
MR. MCGLYNN: Object to form
A. They calculate a risk ratio so that's an estimation of the magnitude of risk. I agree with you, you cannot, I cannot, justify or calculate incidence from those studies. You had a compound question.
Q. So neither of those studies allow you to calculate the incidence rate?
A. That's why I didn't calculate it, I estimated it.
Q. How did you estimate it?
A. I explained that in my paper.
Q. Where at?
A. It's not a perfect way to do it but, again, my working hypothesis was is it as rare as the label implies or is it much more common.
In these analyses, as imperfect as they are, it tells me that it's much more common.
Q. That's predicated on your assumption that the one in 500 applies to tardive dyskinesia, correct?
A. One in 500, even though it's in the label preceding the addition of tardive dyskinesia, and it should apply only to intended use, it's worded in such a way that it includes all extra-parametal symptoms and tardive dyskinesia is an extra-parametal syndrome.
Q. Where is your estimate in your report? Point out to me where you estimate the risk of tardive dyskinesia for metoclopramide users.
A. On page 7, a couple of bullets above Roman numeral 4, I state that for the long-term metoclopramide exposed cohort, and I did not properly say we're talking about the elderly here, elderly or cognitively impaired, the risk for tardive is much closer to one in 20 than it is one in 500 and certainly not rare.
Q. How did you calculate that?
A. I did not. It's an estimation of risk, first of all.
Q. How did you estimate that?
A. I think that's an important distinction because of all the reasons we've talked about, the inability to precisely calculate it.
Q. Let me stop you. It's not a precise calculation, rather it's your estimation?
A. I came to the estimation by calculating, but I am not telling you that that number is an exact answer. I am telling you that that number is an estimation of the exact answer and in an order of magnitude fashion is what I'm trying to explain.
When you're working with the hypothesis that the labeling is telling you that this is exceedingly rare and you're trying to determine whether that is true or not, and in the aged, long-term exposed individuals my belief is that that's not true because it actually can be common in that cohort, physicians need to know that in order to manage it correctly.
[Note: Pages 172-176 missing in original document]
Q. You were done, right?
A. The point of the last, I think that's, in essence, generalized them up to the range that you would find in elderly females.
Q. You described your calculation or your methodology —
A. An estimation, sir.
Q. You described this estimation in Exhibit 10 as a ballpark?
A. Yes.
Q. Doctor, is there any published literature that you can point to that uses your Methodology or your estimation to estimate risk from prevalence studies?
MR. MCGLYNN: Objection to form.
A. I wouldn't call that a methodology, sir, because, again, what we're doing there, and I am going to repeat myself, is we're trying to unravel an issue where the proper studies aren't there.
Q. The proper studies aren't there?
A. The proper studies aren't there because the people responsible for that haven't explored it.
Given that you do the best you can with the studies that exist or mount your own, which I haven't been able to do nor can I do financially, you try to get an estimate based on what's available if you can't do the other two and that's what I have done.
Q. My question was this, you described what you put in Exhibit 10 as a methodology, is that an incorrect description or would you rather call it an estimation?
A. No, we can call it a methodology. It's not a statistical methodology, it's a logical methodology. It's the kind of thinking that the FDA has used for 30 years, 40 years, when you're trying to address safety issues.
When you're addressing an efficacy issue you use the statistical tasks and the P value and you're okay with that. When
[Note: Page 180 missing in original document] you're addressing a safety issue you're dealing with different kinds of data and you're using all kinds of different techniques to try to get close to the truth.
What I've done here is use techniques or methods that are not textbook, I know that's exactly what you want to hear, but they're the types of things that allow you to get a glean into what could possibly be the truth.
Q. Doctor, you told me that the estimation that's set forth in Exhibit 10 has not been used in any published textbook.
Let me rephrase the question. The estimation in Exhibit 10 by which you used prevalence data to estimate risk is nowhere to be found in any published textbook?
A. Textbooks on epidemiology would be there to show you how to do research correctly and how to make your estimates correctly.
If I had the ability or someone did that type of research we can go by the aware of published literature other than your own chapter that has used or determined an estimate of risk by using prevalence data as you have?
A. I don't believe that was your earlier question. You asked me if there are any authorities, I answered you the FDA and other regulatory authorities would find it acceptable.
Q. Let me ask a new question. Are you aware of any published literature, texts, studies or anything that followed your analysis or estimate in Exhibit 10 and used prevalence data to estimate risk?
A. I would have to search, but I am not surprised if I would find something.
Q. Do you know of any sitting here today?
A. Not sitting here today.
Q. Would you publish this?
MR. MCGLYNN: Objection.
A. Obviously I wouldn't publish that piece of paper. First of all, it's not original
would this be a prospective cohort study?
A. You can design it as a prospective cohort but, of course, you can set your time zero in the past and move forward so it's prospective even though it's done in the past.
Q. Would that be the best kind of study to examine the risk of tardive dyskinesia among metoclopramide users?
A. If you're trying to get some incidence rates and you think the data base is appropriate for that measurement you would want to do it prospectively within the appropriate controls.
Q. Prospective cohort studies?
A. Yes.
Q. Doctor, do you agree that an important question about any epidemiological study is whether the study design was appropriate to the research question?
MR. MCGLYNN: Objection to form. Esquire Deposition Services MD – 1-800-539-6398
MR. PILKINGTON: These are my questions.
MR. MCGLYNN: Your notes?
MR. PILKINGTON: These are my questions.
THE WITNESS: Yes.
BY MR. PILKINGTON:
Q. Cross-sectional studies determine the presence or prevalence of both exposure and disease in subjects?
A. Yes, they determine the point prevalence of each.
Q. Cross-sectional studies do not determine the development of disease or risk of disease incidence?
A. That is correct.
Q. If both exposure and disease are determined in an individual at the same point in time it is not possible to establish the temporal relation between the exposure and the disease?
A. That is correct.
something than it might have been a little crazier.
Q. Doctor, could the results of the Ganzini and/or Sewell study be generalized to the general population?
A. No, they can't. I believe they can be generalized to females of the same age group but they're obviously not. I mean, that's what I have been saying all day, that a risk is very dependent on a number of factors and in this case age is a very important factor.
So it's not generalized at all.
Q. Doctor, would it be correct to say that based on Sewell and Ganzini the risk of tardive dyskinesia is one in four for metoclopramide users?
A. No.
Q. Doctor, have you ever used Med analysis?
A. No, I haven't.
Q. Have you studied in detail