Thursday, March 13, 2003


8:00 a.m.








San Diego Marriot La Jolla

Newport-Irvine Room

4240 La Jolla Village Drive

La Jolla, California





Robert Harrison, M.D., Chairman

Leonard Schechtman, Ph.D., Executive Secretary


    Michael Stoto, Ph.D.

    Robert Sills, Ph.D.

    Paul Comacho, Ph.D.

    Ronald Trewyn, Ph.D.

    Kwame Osei, M.D.

    Ronald Varsaci




    Richard Ogershok

    Jay Miner, Ph.D.

    Joel Michalek, Ph.D.

    Lt. Col Julie Robinson

    Maurice Owens, Ph.D.

    Monica Perlman

    Bill Grubbs





Welcome                                              5

    Opening Remarks

Request for Approval of Minutes From 11/14-15/01 Meeting

    Status of New Members


Robert Harrison, M.D., Chair


Personnel Changes and Contract Actions            22


Cancer - Review/Discuss Manuscript                 53

    Joel Michalek, Ph.D.


Mortality - Review Findings from Latest Paper      111

    Joel Michalek, Ph.D.


Diabetes - Summarize latest Analysis of the      129

Insulin Sensitivity Study

    Joel Michalek, Ph.D.


Public Oral Presentation [None]                   146


Hypertension - Summarize the Latest Analysis,      190

Including Skin Exposure Index Results

    Joel Michalek, Ph.D.


Thyroid - Review Results of the Latest Paper       230

    Joel Michalek, Ph.D.


Statistics on Study Compliance to Cycle 6        238

    Joel Michalek, Ph.D.


Latest Results of Consent for Future Use of       244

The Data

    Joel Michalek, Ph.D.


Update on Study Shut Down and Transfer of Data       249

    Joel Michalek, Ph.D.


Publishing Strategies - Journal Choices Versus   276

Time Remaining Before Shut Down

    Joel Michalek, Ph.D.



      DR. HARRISON: Shall we get started?  With the permission of this August assemblage.

     I'm Robert Harrison.  And it's been my poor decision-making that has led me to chair this committee for the last several years.  A much smarter man that I --Mick Gough, sitting to my left--somehow or another managed to weasel out of doing this job back in the early '90s, and no one else has been silly enough to take it since then.


    Before we start, can we go around the room and introduce ourselves.  We've got a few new people here.  Again, I'm Robert Harrison.  I'm no professor emeritus of medicine at the University of Rochester, and doing mostly consultant work and running and internet-based weight management program called "Be Slender."

    Why don't we go this way.

    DR. GOUGH: I'm mike Gough.  I'm retired.  I guess I'm a retired bureaucrat--a retired Congressional bureaucrat.  That says it.  And as Bob said, I chaired the committee for five years, from '85 to '90.  And then I came back four or five years ago.  So I've been here a long time.

    DR. STOTO: I'm Mike Stoto.  I'm an epidemiologist, and also I [inaudible]--the Rand Corporation [inaudible]--use to be true.  I'm also on the faculty of Harvard School of Public Health [inaudible].

    DR. STILLS: I'm Robert Stills.  I'm a Veterinary Pathologist in the National Institute of Environmental [inaudible] Sciences.  And I'm also [inaudible] at UNC [inaudible].

    LT. COL. ROBINSON: I'm Julie Robinson, and I'm Branch Chief for Biomechanisms and Modeling at Brooks City Base.  And I'm also the Director of the [inaudible], and just recently re-associated with [inaudible].

    DR. HARRISON: Let's continue on around.


    DR. TREWYN: Ron Trewyn, Kansas State University.  I was on the committee for awhile, and then got thrown off for causing too much trouble--


    --and I think they figured out that it was too quiet without me, so brought me back the last time.  And so, you know--

    DR. CAMACHO: I'm Paul Camacho.  I'm at the William Joiner Center for the Study of Law and Social Consequences at the University of Massachusetts.  It's a research institute--one of several that are located in Boston.

    DR. HARRISON: Okay. And continuing in this--I'm sorry--

    DR. SCHECHTMAN: I'm Leonard Schechtman.  I'm the Executive Secretary of the Ranch Hands Committee.  I'm also the Associate Deputy Director of the FDA National Center for Toxicological Research.

    DR. HARRISON: Sounds important to me.

    DR. MICHALEK: I'm Joel Michalek, Principal Investigator of the [inaudible] study, and adjunct faculty at the University of Texas School of Public Health, University of Texas, San Antonio.

    DR. BABYAK: I'm Babyak [inaudible].

    DR. I'm Manuel Bancas, the [inaudible] contractor working for program management [inaudible].

    DR. MINER:  I'm Jay Miner, formerly on the technical side of the house.  I [inaudible] completed my active duty and came back to the dark side of the force, in program management.


    DR. OGERSHOK: I'm Richard Ogershok.  I am the program manager of the Ogershok Study.

    DR. BROWN: Program monitor from Washington, D.C. [inaudible].

    DR. PERLMAN: I'm Monica Perlman [inaudible].

    DR. OWENS: My name is Maurice Owens.  I'm the science applications [[inaudible] for my program managers.

    DR. GRUBBS: I'm Bill Grubbs [inaudible].

    DR. YEAGER: I'm Meghan Yeager [inaudible].

    DR. KEIHL: Bill Keihl [inaudible].

    DR. LANGER: Walt Langer [inaudible].

    DR. FORNEY: Joseph Forney [inaudible].

    DR. HARRISON: Welcome, everybody.

    Some additional information for everyone.  The Secretary's office has the Ranch Hand Advisory Committee nomination package.  Two new members have been nominated: Sanford Leffingwell, M.D., M.P.H., and Isabel Hoverman, M.D.  And it's also been requested that Dr. Stoto be reappointed.

    Dr. Leffingwell is an occupational and environmental medicine consultant, with HLM Consultants in |Atlanta; served as a medical epidemiologist for the chemical de-militarization program with the CDC, a group providing Congressionally mandated oversight of the Army's chemical weapons disposal activities.  Are those weapons of mass des--nah.


    Dr. Leffingwell also served as a member of the U.S. negotiating committee for a bilateral chemical weapons treaty, and was a member of the U.S. medical management of people injured in the 1995 subway nerve-gas incident--I assume that's Japan?

    He serves on the NRC Standing Committee on Program and Technical Review of the U.S. Army Chemical and Biological Defense Command, and the Committee on Chronic Reference Doses for Selected Chemical Warfare Agents.  It says "welfare agents" here.


    Dr. Hoverman is a practicing physician in internal medicine in Austin, with extensive experience in the development of clinical practice guidelines, and has participated in the design of health outcome studies and measures for out-patient practice.

    Prior to entering private practice, she was an assistant professor of medicine at the University of Texas Medical Branch at Galveston.  Dr. Hoverman has served on Previous Institute of Medicine committees, most recently on the Committee on a National Center for the Study of War-Related Illnesses and Post-Deployment Issues.

    Does this mean, Len, that you have some housekeeping items?  Should we interrupt here for you?  Okay.  I'll finish the rest.

    One of the things that's come up that's been an issue--I know, with me, and perhaps with others--is that the FDA has instituted new guidelines for advisory committees, requiring that prior to every meeting the agency is to review members' files for conflict of interest.  This means at least once a year you're going to be requested to submit updated and signed OGE-450 forms; those are the confidential financial disclosure reports; and Form HHS-697, which is a foreign activities questionnaire.

    Those at present are the only requirements.  In order for FDA, NCTR to be responsive to these requirements now in effect, and in order for members to attend and participate in Ranch Hand Advisory Committee meetings, and to qualify for reimbursement of committee-related expenses, it's essential that you have these things turned in in a timely fashion.

    Later on we'll be discussing the next meeting dates.  As far as travel voucher expenses are concerned, please submit to Ron Varsaci who, I think somehow or another, managed to avoid introducing himself during the round of introductions.

    Hotel receipts; mileage if you used a vehicle; airport parking receipts, taxi costs; receipts are required for $75 or more, and airline ticket receipts, where applicable.

    Now, Len has some other housekeeping stuff.

    DR. SCHECHTMAN: Just to reinforce the issue around these new FDA policies: it's critically important that everyone submit their conflict of interest forms, filled out and signed, as well as their foreign activities questionnaire.

    Unfortunately, because FDA is now under extremely tight scrutiny around these issues from the Department--DHHS--we have to comply, as a committee, with those new rules and regulations.  And because of that, unfortunately, if those forms aren't--if we don't get responses to those questionnaires, and those forms aren't filed out and signed in an appropriate manner, with complete identification of your financial holdings, we can't have you sit at the table, we can't have you participate in the meeting, we can't reimburse you for any of your expenses to get to the meeting.  And we apologize for that, but this is how they're trying to enforce these new--or tighter--regulations that weren't, perhaps, as tightly adhered to before.

    So, FDA is being held to the task, and that's passed right down through all the committee and subcommittee activities that FDA leads.

    So, please, when the requests come--and they come every October, while this committee is still operational, within two weeks of that request, please turn that information around to us.  It will be scrutinized before every meeting that the committee holds.  And if there are questions that arise, we need to have that all cleared up before we congregate and have our meeting.  And then there will be no questions; we won't have to be making phone calls back and forth to FDA from whatever meeting location we're at to get last-minute approval.  Because often they don't want to deal with it at that point in time.

    So, please--thank you for your cooperation, and we look forward to hearing from you two weeks into October.

    DR. HARRISON: Okay.

    [Discussion off mike.]

    Can we have just a short break?

    [Off the record.]

    DR. HARRISON: Sorry for the interruption.  I was concerned because one of the committee members, Kwame Osei, because of some of the business I just finished describing about financial reporting, is having difficulty being allowed to attend this meeting; or participate in this meeting.

    It's one of those trivial things, where you put down "Vanguard Fund," and then you have to put down precisely which Vanguard Fund.  And I'm concerned, because Kwame is not just a clinician, but is an active and prolific researcher in the area of diabetes, which is one of the areas that we've identified, and we'll be listening to a report later on.  And I'm basically tired of being the only person who is supposed to know anything about diabetes, being on this committee, listening to this report.

    So I've Kwame--talked to him and asked him to contact his broker, and see if the FDA staff can't get his approval expedited so that he can participate in this program by 10:00 a.m. this morning.  If he cannot, then we'll have to stop.  I'm not going to participate.

    MR. VARSACI: Could I add one thing?  There was some confusion as to whether the individual companies of the fund holds, and that's not that far.  You have a family of funds, and they have sub-segment funds--

    DR. HARRISON: I'm not really interested in that.  What I'm interested in is being able to apply the scientific method to this advisory committee's actions.  And to do that, we need certain people.  And if we can't have them, then I'm not sure that I want to be a participant.  It's just that simple.

    DR. STOTO: Bob, if he can't get the information by 10:30 this morning, are we--

    DR. HARRISON: I'm--we can rearrange the agenda.  But I'm just getting irritated.  I'm sorry.  For some reason, the older I get, the less tolerant I become of this kind of stuff--you know.  And so we'll see what happens.

    So, until then, let's proceed.  So I just wanted to--a couple of people have asked me why I asked for the break and everything, and I just wanted to explain to everyone what had happened.

    Okay.  Can we go ahead now and proceed with--oh, wait a minute.  We need to approve the minutes from the 11/14, 15 meeting. All of you have copies of the agenda in front of you.  Are there any corrections to the--I'm sorry, not agenda--minutes.  Al of you have copies of the minutes?  Do any of you have--thank you, thank you, thank you.  Try to keep them, you know.

    Okay.  I'm starting to come back down again. We're doing okay.

    DR. GOUGH:  [off mike.]

    DR. HARRISON: No, no.  Not at all.  You could never rile me up.

    DR. GOUGH: I have some corrections to the structure of some of these sentences, which I'll just pass on.  But I do have one significant thing, and that is on the second page, under "Technical Presentations," the second bulleted item.  That sentence, which says, "Stated another way, after adjustment--"--blah, blah, blah.

    I would like to have an insight to--that's what Dr. Michalek said?  Because I disagree with the statement.  I mean, Ranch Hand veterans have not experienced a significant increase in the risk of diabetes.

    DR. HARRISON: Well, why don't we do this.  At the beginning of the technical presentation section, it says, "Dr. Michalek presented a series of slides on his statistical analysis on diabetes and other effects."  These are Dr. Michalek's statements.

    DR. GOUGH: That's fine.  That's fine.

    I think all of the--in fact, I think every time it--

    DR. HARRISON: You could almost assume that it was.

    DR. GOUGH: I know.  But when you say, "in other words," is as though--

    DR. HARRISON: And then the last bullet in that section, then, should be pulled out and not made a bullet.  We can leave the statement in.  It just shouldn't be a bullet.  Then it's obvious, then, that the Ranch Hand Advisory Committee complimented Dr. Michalek--okay.

    Anything else?

    DR. CAMACHO: So, you wanted to pull that sentence.

    DR. HARRISON: Well, otherwise, if it's left as a bullet, Paul--

    DR. CAMACHO: No, no, no, no.  The sentence--"Ranch hand veterans have experienced significant increase in risk--"--

    DR. GOUGH: No, just leave it in, because Dr. Michalek said that.

    DR. HARRISON: Yes.  All of those--everything that's in there that's bulleted stays in.  It now is just preceded by a sentence that says, "These are Dr. Michalek's statements."

    DR. CAMACHO: Okay.

    DR. HARRISON: And then the last bullet, which isn't Dr. Michalek's statement--

    DR. CAMACHO: Gets the bullet pulled--

    DR. HARRISON: Gets the bullet pulled out, but the statement stays in that we complimented him on his stuff.

    DR. CAMACHO: All right.

    DR. HARRISON: Is that fair enough, everybody?  Okay.

    DR. GOUGH: Then can I ask for one other--

    DR. HARRISON: Yes, sir?

    DR. GOUGH: If it's possible, on page 3, under dioxin and insulin sensitivity--right within that paragraph, the penultimate sentence in that paragraph, it says, "Dr. Michalek's slides presented preliminary data--"--Joel, is it possible to summarize those data, and have them in the minutes?  I don't know if that could work.

    DR. MICHALEK: I have not seen the minutes.  [Off mike.]

    DR. HARRISON: That's reasonable enough.  Anyone disagree with that?  Or do you want to add to it in any way?

    DR. STOTO: Consistent with what we just did [inaudible].

    DR. HARRISON: Hmm?

    DR. STOTO: It's consistent with what we just did with the previous one, is to--you know, have the summary be prepared by the person who was doing it.

    DR. HARRISON: Oh, yes.  Yes.  But, in this case, it's adding Michalek's rendition of the results to the paragraph.

    DR. MICHALEK: [Off mike]--by e-mail?

    DR. HARRISON: By all means.  By all means.  At that point, what we can do then is we'll make the edit changes that have been suggested and we'll send it out--we'll send the minutes out to the committee for final approval, through you.  Can we do that?

    What we can do is approve it, with changes, subject to my review and--

    DR. GOUGH: I so move.

    DR. CAMACHO: Second.

    DR. HARRISON: Moved and seconded.  Any other questions or comments?  Any other corrections to the minutes?

    DR. CAMACHO: Was Mr. Weidman ever sent the public section that he spoke--

    DR. HARRISON: What page?

    DR. CAMACHO: It's page 4, bottom--"Public Session--Mr. Weidman--".  Does he have to be sent--

    DR. HARRISON: I don't know.  Maybe--

    DR. STOTO: I don't think it's appropriate for him--

    DR. HARRISON: Maybe I'm--I mean, Michalek's not a committee member either, but in this case, these are minutes of what we think happened, not minutes of what Weidman thinks he said.


    DR. HARRISON: That would be my thought.

    Okay.  Anything else?

    Okay.  Then I--sir?

    DR. OGERSHOK: [Off mike].

    DR. HARRISON: Oh, yeah.  No, we're almost there.  So, there's a consensus that that's how these minutes will be handled?  So we're almost exactly on time, and we'll proceed with personnel changes and contract actions.

    DR. OGERSHOK: [Off mike].  In the context of--what the--the website had certain data pulled off it, right?  Remember, there was a big issue about privacy?

    DR. HARRISON: Yes.

    DR. OGERSHOK: Is that reflected in the--I didn't see that reflected at all in this.  Is that something we should be concerned about, or I'm being--

    DR. HARRISON: That was in the minutes.

    DR. OGERSHOK: It was.

    [Discussion off mike.]

    DR. HARRISON: It came off e-mail.  I'm sorry, we've regressed--just momentarily here.

    DR. STOTO: My recollection is that it was resolved.  More happened after the meeting, but these are minutes of what happened at the meeting.  And I think from that perspective, they're correct.

    DR. GOUGH: I thought that all the data was pulled off the website, because of the legal determination by [[inaudible].  And that's not reflected in the minutes, but I think that happened after--

    DR. HARRISON: So then shall we include it as a part of the discussion at this meeting?  That as the minutes were being discussed, the question was being brought up about the display, or the availability of Ranch Hand data on the Air Force website.  And it was suggested and agreed by Joel Michalek that the material was pulled from the website because of legal concerns by the Air Force.

    DR. SCHECHTMAN:  Was that done advice of legal counsel at that time?

    DR. HARRISON: By the Air Force.

    DR. SCHECHTMAN: By the Air Force.

    DR. HARRISON: So that will now be a part of this meeting's minutes, in a--as a display of the discussion.

    DR. TREWYN: Probably generally, just having a topic of "Business Arising from the Minutes" would be good so those sorts of issues--because I think we generally wind up leaving something that needs to be handled afterwards, and if we could have those highlighted in the minutes, then that could be an update right after we've dealt with those.

    DR. HARRISON: It looks like Len has agreed with that statement, and I expect that we see that sub-heading when we see the next set of minutes.

    DR. CAMACHO: Yes, I just brought it up because--

    DR. HARRISON: That's a good point--

    DR. CAMACHO:  --the veterans community--causes a concern, among others.

    DR. HARRISON: Well, and I think there should be more discussion about that when we get to the section on the consent for further use of data that's in here.  I certainly have some questions about how that was done to achieve a 99 percent compliance.

    Okay.  Any other questions about the minutes?  Are we finished with the minutes?  Anyone who has any thought that we're not finished with the--any glimmer of--no?  We're finished with the minutes now.  All right.

    DR. GOUGH: Will you yield me three minutes?


    DR. HARRISON: If you have something, let's deal with it now.

    DR. GOUGH: I don't have it now.  No.  I have to think about it.  No, I had nothing, actually.  I just want to keep it open.

    DR. HARRISON: Okay.  Anybody got any Zantac?


    Okay.  The minutes are approved.  Let's go on now to program management update.

    Thank you.  Sorry to have kept you delayed, sir.

Personnel Changes and Contract Actions

    DR. OGERSHOK: Since our last meeting I thought we'd go over some of the events that transpired in the program since--I think it was, what?--November of '01 or so?.

    DR. HARRISON: Sounds right.

    DR. OGERSHOK: All right.  Next slide.


    This is our program team.  And as you see, we include the advisory committee as an integral part of the team.  One change--I believe it was changed about the time we last met, however we have our program element monitor was shifted from Air Force RD to the Surgeon General of the Air Force.



    To get ready for the physical exams, we had different types of documents that had to be prepared, between October and March of 2002, which included the biomedical test plan, schedule and tracking plan and a QA plan.

    And then we had to do some testing of some of the components of the examination.  We had a schedule and tracking system demonstration at the National Opinion Research Center in Chicago.  We did a CAPI demo down at Brooks Air Force Base.  And then we did a test of the examination itself, to see how the protocol was being adhered to by the examiners.  And that was done here in Scripps in March of last year.



    We started the exams in May of 2002.  We planned on doing 81 groups, approximately--hopefully--2,100 participants.  We're not going to achieve that.  We've scheduled about 1,969, I think it was, as of yesterday--which probably changed by today.  And we've completed 75 groups, with 1,831 participants.

    To give you an idea what it costs--on the cost of doing the examinations, when you include the exam and the reimbursables associated with it, it's about $6,350 per person.

    What we're planning on doing in the near term: SAIC has already submitted the stat plan to us.  We plan to get back with them at the end of this month with out comments.  Exams will be ending the first week in April.  And, let's see the next one up there.  Oh--delivery and acceptance of the exam data will be taking place in June of this year.



    What's going to happen in the following fiscal years, of course, is the stat analyses which take about--what?--13 months, I believe.  Then we have the report-writing and chapter reviews, which you will all be an integral part of.  And Dr. Miner's going to give you his layout of how that's to be accomplished.  And then, of course, our final cycle report delivery.

    Now, after that--after the data, the report--obviously we can do peer reviewed articles.  And Dr. Michalek will go into that.  We obviously have to publish a final mortality report; any special studies--depending on the findings and data that we get; and then digitize the rest of the exam data that we are getting right now; and, of course, complete the study according to the protocol.



    Now, special interest items.  I mentioned that we transferred our program element monitor to the Surgeon General of the Air Force.  We're short money in 2003 because of some cuts that were taking place.  However, we were fortunate to get a Congressional plus-up of a million dollars in 2002 that we were able to use this year.  So--we're copacetic.

    And, of course, our special interest item on the process for completion, disposition of data and specimens, which we have talked about in the past.  And, again, we will discuss the process for review of the chapters.

    Dr. Miner will take over.

    DR. MINER: You all may remember these slides from before, but the Air Force is very concerned that they want the committee's involvement in chapter review.  And just to kind of serve as a reminder, what we have--April, the exams end.  In about 60 days--but that might be 30, it might be a little bit more than that--we'll get our last data delivery.  Air Force then has 30 days to accept the data, and then our prime contractor, Science Applications International, will begin the analysis.

And that will be a 13-month effort, and the draft report is due in August of 2004.  We have 90 days, then, to look at the draft report, provide comments back to them.  Thirty days after that, a final report is due, and then in December of 2004, the final report will be delivered.

    It's this 13-month period that I would like to highlight; specifically, the involvement of the committee.  As you may remember, the Air Force has an iterative chapter review with SAIC, and they send us draft chapters; we resolve comments.  And you had requested that we then provide you those draft chapters with Air Force comments for review.

    So the committee review process--I don't want to say here's what--yes, sir?

    DR. HARRISON: Jay, can I interrupt you for a second?

    DR. MINER: Sure.

    DR. HARRISON: My question actually has to do with the selection of experts and writers for the draft chapters.  How is that done?  I assume that's done by the contractor, and the--

    DR. MINER: We had a presentation on that yesterday with our management meeting; individuals here--if you want to address that a little bit.

    DR. HARRISON: Because the concern I have is the quality of the authors and the fact that once the thing is written, you're basically reduced to editorial changes, as opposed to anything substantive.

    Can we have some comment on that?  Is that--

    DR. MINER: Yes.  Dr. Owens is here, and he can address that.

    DR. OWENS:  I'll just review our process and procedures for accomplishing this.

    The bulk of the material that's in a chapter--a clinical chapter, say "General Health," or "Hematology,"--in the report is statistical analysis that is presented according to plan.  And that's pretty straightforward procedure.  I think our practice has been very factual and objective.

    There are two areas where we have people coming in--expertise--in terms of preparing the report.  One is the review of the literature.  We had, in the past, we had one person that did that, who was a cardiologist.  This committee commented on that.  The Air Force changed its requirements, and we have requirements to provide experts in medical literature.  We have an epidemiology firm, experienced people--Ph.Ds in epidemiology--that will be doing those literature reviews. So that's one piece.

    The other piece is the clinical interpretations of the findings of the report, that hithertofore were written by one person.  The Air Force has changed its requirements, given some of the committee's feedback on that, and we're required to provide an expert in the respective clinical area.

    So we have made arrangements with Scripps Clinic to handle the medical reviews for the individual chapters.  Dr. Monica Perlman, here, will be--she's going to be heading up that process, and will be working with physicians, board certified in their respective areas, for the interpretation for the chapters.

    So that is the process that we have in place, and are ready to go for the 2002 study report.

    DR. HARRISON: You could almost gather from--so, out of all of that, the question that I'm asking is specifically about those experts.  And all that you've said so far is that they're board certified.

    Dr. Perlman, George Dailey will tell you that I'm a bit difficult at times--

    DR. PERLMAN: That's okay.

    DR. HARRISON: I knew him when he was a young lad.

    DR. PERLMAN: It won't just be a board certified physician in that area.  It will be an expert in that area who's done research in the area.  George Dailey, as an example, would be for diabetes endocrinology.

    DR. HARRISON: So George is going to write the diabetes section?

    DR. PERLMAN: Well, he's the one I was going to go--I have not specifically gone and asked anybody, but he's the one in our division--our division of diabetes endocrinology--that I would approach, unless--you know, I mean we have two other choices that are excellent.  And I haven't really done final selection at this point.

    DR. STOTO: I wonder if we could resolve this by just asking to see the list so people will have an opportunity to comment on the list.

    DR. HARRISON: That would be very reasonable.  Is that logistically achievable?

    DR. PERLMAN: To give you a list of who--yes, sure.

    DR. STOTO: I'd ask the same with the epidemiologists.

    DR. PERLMAN: Yes.

    DR. HARRISON: Great.

    DR. MINER: We do try to pay attention.

    DR. STOTO: Not that we have any, you know, concerns in advance.  But I just think that part of our oversight responsibilities, it would be appropriate.

    DR. HARRISON: Well, and that's a really critical process.

    DR. MINER: Right.

    DR. HARRISON: I mean, that puts the rubber on the road.

    DR. MINER:  Yes, sir.

    And then again, back to the advisory committee review, there's about 20 chapters that will require reviewing in this 13-month period.  And whatever process the committee decides on how they want to do that--that's why I put a little star there that says that's your process, you decide, and we'll do our very best to expedite and help you do that.

    DR. STOTO: On those three meetings to review chapters, I think the first one should be November '03.

    DR. MINER: Again, keep in mind that that little star says you decide how you want to do that.

I just tossed some--

    DR. STOTO: It's just a typo, though, I think.  It says, "November '04."  I think you--

    DR. MINER: Oh--a typo.  I'm sorry--yes.  Or you might work real fast and--yes, that should be November '03.

    Anyway, so by July of 2004, we will need to review and resolve comments.  August would be the draft report for 13 months.  And then, as the custom, we have an advisory committee meeting and review of that big, giant final thing.  But if everybody's reviewed chapaties, that should not be nearly the trauma that it was the last time--we hope.

    And then, again, we have 90 days from the draft report for Air Force comments, and the final report in December.

    DR. HARRISON: I'm not sure I understand one thing--and that is, the draft chapters--the ones before--that the Air Force is going to comment on, when do those chapters start to come out?

    DR. MINER: As soon as Air Force accepts the data, the statistical analysis will begin; some report-writing will have been done on the basic chapters at that time.  And as quickly as they get chapters to us--and that could be as early as--do you want to comment on--

    This is Dr. Bill Grubbs, our--SAIS's statistician.

    DR. GRUBBS: Hi.  It depends on the--as Dr. Miner said, it depends on the approval of the final data.  That kicks off the process.

    If the Air Force has comments on the data, need to be revised, it pushes the process back a little bit.

    DR. HARRISON: But according to the schedule, when should the draft--not the ones after Air Force approval, the ones before Air Force approval.

    DR. GRUBBS: To give you an idea from last time, the approval from the Air Force happened on August 26 of 1998--it happened near the end of August.  The 13-month period then ended in September of--late September of 1999.  Just to review from last time.

    DR. HARRISON: But if the Air Force approves the data in August of 2003--

    DR. GRUBBS: Right.

    DR. HARRISON:  --according to the schedule, when should the first drafts be back to the Air Force for Air Force's comments?

    DR. GRUBBS: the process is a rolling report.  The first chapter of the 12 clinical chapters would be there approximately October or November of 2003.

    Then undergoes the process where they review it, give it comments, we resolve those comments.  We get back the second draft to them, and then you get that.

    So, as an estimate, from when we submit the first draft of a chapter to the Air Force, it would be approximately two months before a second draft would be available for the committee to review, incorporating the Air Force comments.

    DR. MINER: I think you're asking when will the first chapter be in to the Air Force.

    DR. HARRISON: Well, and Bill is saying that that's approximately two months after the Air Force approves the data.  And--

    DR. GRUBBS: Not quite.

    DR. MINER: [Off mike].

    DR. GRUBBS: The clinical chapters--once they approve the data, we need to do the analysis, and submit a draft chapter.

    DR. HARRISON: But you said August?  You said the last time, it was approved in August, and the first chapter was out in October.

    DR. GRUBBS: Late October, early November --correct.  So that's two months.  But then the review process will probably take--

    DR. HARRISON: I'm not asking about the review process.

    DR. GRUBBS: Okay.

    DR. HARRISON: I want to know when the first draft is ready.

    DR. GRUBBS: Probably--the first draft of the first chapter, about two months after the approval of the final data.

    DR. HARRISON: Okay.  And obviously the reason why I'm asking when the first draft is ready is to ask whether the committee would like to discuss whether it's worth seeing the drafts before the Air Force's corrections, versus after.

    DR. MINER: Do not use the word "corrections."


    DR. MINER: We don't like that word.

    DR. HARRISON: After the Air Force's airbrushing--


    DR. MINER: That's worse.

    DR. HARRISON: After the Air Force review.

    DR. STOTO: So, in other words, that's three times, rather than two--adding upon, rather than--

    DR. CAMACHO: Well, already we're talking--I mean, this plan has us looking at them twice--chapter by chapter, and as a whole.  And you're--just clarifying what you have in mind.

    DR. HARRISON: I'm not--actually, I don't have a proposal in mind.  What I'm asking is when the least--huh?

    DR. CAMACHO: When the least work on the data.  So you want to see closer to the raw--

    DR. HARRISON: I'm not saying I even want to see it.  I'm just saying that that's when the stuff first comes out, then there's an intermediate stage, after it's been--any comments shave been incorporated, and then there's a stage after, say, some of us have seen it and made other comments.

    And the question is--for two reasons--number one, if I were asked if I knew--how I knew that there was nothing contrary to the final report in the initial report, I'd have to say I didn't know that in any way.  You know.

    So if you have a segment of the population that is suspicious of the reports, and if the advisory committee is--if a part of the advisory committee's function--then can you validate a report that's already had comments incorporated?

    DR. STOTO: I'm concerned about looking at it three times.

    DR. HARRISON: I understand that.

    DR. STOTO: So what I would propose is that when we see them for the first time, according to the schedule, that we be told what changes were made, beyond--

    DR. HARRISON: Or could we simply ask for both copies?

    DR. STOTO:  Well, I guess I'd prefer to.

    DR. HARRISON: I understand.  I understand.  I'm not opposed at all.  I'm just brining this up as we go along, that's all.

    DR. TREWYN: I think seeing the draft, when we see it, that has the changes highlighted or incorporated by the initial review--I think that would be useful.  Because then we've seen what they have incorporated.  So, presumably, what isn't highlighted--

    DR. STOTO: And what they've taken out, too.

    DR. TREWYN: What they've taken out--exactly.  I think that--

    DR. HARRISON: Okay.  All right.

    Any other comments?  Is that an official consensus-driven request by the committee?

    DR. TREWYN: I think that would address the issues that outside groups might have of, you know--were we aware of what was there, you know.

    DR. HARRISON: Can we word it this way?  Can we say that the advisory committee would like for its review, the revised draft, with indicators of deletions and changes that have been made.  Or, if that is not possible--it should be possible.  But if that is not possible, then both the initial draft and the revised draft.


    DR. MINER: As part of the process, when we are resolving comments with SAIC, they usually prepare a table for us of comment resolution and recommendations.  And perhaps we could just include that.  Say "Here's what it was, and here's what was suggested," and then a resolution--Air Force resolution, which will be incorporated.

    That makes use of a document that we currently have.

    DR. HARRISON: Yeah.  And we don't want to make any more work.  I mean, that's not the objective.  I understand.

    What's you all's feeling about this?

    Okay.  So we'll revise the request, that when the revised draft is given to us for review, that it should be accompanied by the table referred to by Jay Miner, that includes the comments and actions taken on comments within the draft.  And that's a comprehensive table.  That includes any changes that are made?

    DR. MINER: Yes.

    DR. HARRISON: Okay.

    DR. MINER: Every single--

    DR. TREWYN: We get the original draft, then we get all the changes--

    DR. HARRISON: No.  No.  We get the changed draft--

    DR. TREWYN: We get the changed draft.

    DR. HARRISON: --plus the table that tells us what changes were made.

    DR. TREWYN: What was changed.

    DR. HARRISON: Yes.

    DR. STOTO: And I presume if there were something--

    DR. MINER:  "We took out 10 pages here--"--

    DR. STOTO: We might not, but if we though it was something, we could ask for it.

    DR. HARRISON: Well, that's a good question, though.

    [Discussion, off mike.]

    DR. MINER: Well, send the original draft, the table of changes, and the revised paper.

    DR. HARRISON: Okay.  All right.

    DR. MINER: We will send it electronically, and you all can print it.

    DR. HARRISON: We can kill our own tree.

    DR. MINER: Right.

    DR. HARRISON: Everybody happy?  Any dissent?


    DR. MINER: That's it, sir?

    DR. HARRISON: That's it.  Thank you very much.

    So, next we have--

    DR. GOUGH: When are these going to start coming in?

    DR. HARRISON: Next week, you get your first draft.


    DR. STOTO: Probably November.

    DR. HARRISON: Yes, somewhere around October--no, actually, November--somewhere in that range.

    Now, can I--before we get off of this, can I ask a non-agenda-driven question?  And that is: how does this impact on when we meet next?


    DR. STOTO: Well, it sounds like the next business for us on this agenda is November--right?

    DR. TREWYN: That's for reviewing--we have to read it.

    [Discussion off mike.]

    DR. HARRISON: I'm just raising the question.  You know, we have a certain amount of work in front of us.  We have a tentative schedule.  How does this impact on when we meet the next time?


    DR. GRUBBS: [Off mike].

    DR. HARRISON: You need to get the microphone.  Otherwise--

    DR. GRUBBS: My best estimate, when you would see the second draft--have the second draft available, of the first clinical chapter that comes out, would be approximately the start of the year--start of January--

    DR. HARRISON: Start of what?

    DR. GRUBBS: Start of January 2004.  Because the first draft will come in November.  There will be the resolution time--the revision of the chapter to get to the second draft.

    So I would think you would have second drafts of chapters beginning January 2004, and going throughout the year until, I would guess, July or August.

    DR. STOTO: I thought the dates here on the chart we just looked at were for the second draft.

    DR. MINER: Those are just notional dates.  As an example, you would have to have, perhaps, three meetings, with six or chapters each meeting.  That's your process for deciding when you meet.

    [Discussion off mike.]

    DR. GRUBBS: Again, it's contingent on a final approval of the data by the Air Force.  And any additional data that they would need to provide to us, such as medical records verification.

    So, it's contingent on a particular approval date by the Air Force.

    DR. MICHALEK: I'd just like to remind you that we have been reviewing this data all along.  As SAIC delivers the data to us from Scripps, we get it every quarter.  And we've been checking the data and reviewing it constantly.  So this is not like in August of this year we're going to see the data for the first time.  That's not true.  We've been seeing it since the beginning of the Scripps process.

    And so, in other words, I'm telling you we're very highly confident that we'll provide an acceptance within a very short time of receiving the final data set from SAIC.  We have a process in place to assure that that happens, so that we do not have a slip.

    Everything is checked and rechecked--may times over.  And that's part of our process.  The same is true of the medical coding data, which is a necessary piece that isn't collected at Scripps, but which is derived from medical record review.  We have a staff that does that full time, all the time, and keeps up with it, so that when the final Scripps physical exam is complete, we're nearly ready to deliver data as needed by SAIC to do the report.

    So, in other words, I'm telling you we have very high confidence that when the deadline approaches in August that, yes, we will have an approval, and SAIC will launch their statistical analysis.

    DR. HARRISON: So--okay.  I'm still trying to get my hands around this thing, though.  All the chapters are going to be the chapters containing responses that are responsive to the Air Force's comments are going to be completed by November.

    DR. MICHALEK: There is a--SAIC is working under a fixed-price contract.  There are fixed timelines in place that Air Force must respond within, I think, 30 days.  Is that true?  Of every chapter?  Yes?  30 days?  As soon as they--15 days--15 days.

    As soon as SAIC delivers a draft chapter to us, we, the Air Force, have 15 days to review and comment, and send our markup back to them.  And then they are free to agree or disagree with our changes--our suggested changes.  And they resolve--we resolve together, as colleagues--these issues, and they provide a document which lists every single issue and as to whether we yield or they yield on a point.  And you will see all of that.

    DR. HARRISON: When?

    DR. MICHALEK: When.  The first chapter should be delivered first week of November--we think.

    DR. HARRISON: And the last chapter will be delivered--

    DR. MICHALEK: The last chapter, first draft, would be delivered when?  Roughly?

    DR. HARRISON:  This is with the corrections, Mike.

    DR. GRUBBS: Roughly July of 2003.

    DR. STOTO:  No, I thought he said the first draft.

    DR. MICHALEK:  The first clinical chapter, roughly the first week of November of this year.

    DR. STOTO: First draft?

    DR. MICHALEK: First draft to us--to the Air Force.  We have 15 days to turn it around and get it back to SAIC.  And how much time do you have to--they have 15 days to revise.  So we're talking about a one-month period before we have a second draft ready.

    DR. HARRISON: And the last chapter is not going to be done until the--the first draft is not going to be done until July of 2004?

    DR. MICHALEK: Approximately July of 2004 would be the last first-draft clinical chapter that we will see.

    DR. HARRISON: Okay.

    DR. MICHALEK: We're talking about--for those that don't know--it's about a 4,000 page report.

    DR. HARRISON: Yes.  Yes.  But--so we've got a period, essentially, from December of 2003, through July 2004 when these babies are going to be rolling out.

    DR. MICHALEK: That's true.

    DR. HARRISON: And there's approximately 12 chapters.  So that means that there's going to be, on average, two chapters per month.

    DR. MICHALEK: There are actually 20--

    DR. HARRISON: Twelve clinical chapters.

    DR. MICHALEK: Twelve clinical chapters, and you should know that some of those other chapters are just as important to you as the clinical chapters, such as "Methods"--

    DR. HARRISON: I agree.  I agree.

    DR. MICHALEK:  --you need to see the "Methods" chapters--and the concluding chapter.  There's about 21 chapters.

    DR. HARRISON: Okay.  All right.


    DR. CAMACHO: We're getting first raw, and the changes--

    DR. HARRISON: Well, yes.

    DR. CAMACHO: Which one?

    DR. HARRISON: Well, we're getting--

    DR. CAMACHO: It's as if we're getting different versions.

    DR. HARRISON: Well, we're getting them in a stream.

    DR. CAMACHO: Yes.

    DR. HARRISON: And the point I'm trying to get at is, that if the stream begins in December, we probably want to chop that into--three parts you say?  So we're talking about--

    DR. TREWYN: Meeting a lot in 2004.

    DR. HARRISON: We're talking about February, April--June.  June would be too early.

    DR. MICHALEK: Plus, I want to emphasize that the most important chapters are, finally, the concluding chapter and the executive summary, which are extremely important.

    DR. HARRISON: I understand.  And I didn't mean June in terms of cutting out before the end of it.  That was my quick--I'm not good at this stuff.

    I'm just trying to get a discussion right now about--at least in terms of months, when we need to meet.  It looks like we don't need to meet again in 2003.  But we need to meet early in 2004--which, for those of us coming from the northeast is always problematic.  To quote a friend of mine in Rochester, New York, in February you can't always be sure of getting out, nor can you be sure of getting back.


    DR. CAMACHO: So how many chapters do you think we'll be looking at in, say, January?

    DR. HARRISON: So, in January, we'll have four to six chapters ready--they'll have four to six chapters ready.

    DR. CAMACHO: Ready for us to look at.

    DR. MINER: We think March would be a better guess as to when you would have enough work to discuss.

    DR. HARRISON: Okay.  All right.

    DR. TREWYN: So if we don't get them before that time, or if we don't--

    DR. HARRISON: We'll get them before that--

    DR. MICHALEK: You'll have time to look at them.

    DR. HARRISON: If we're starting with two in--two delivered with--I'm sorry--the corrections in December, two delivered in January, two delivered in February, then that would be six for us to talk about in March--with, you know, a couple weeks for whoever--

    DR. MICHALEK: You're being a bit ambitious.  We think that's an over-estimate, a little bit.

    DR. HARRISON: Well, even if we are--

    DR. MICHALEK: Maybe five chapters.

    DR. HARRISON: Even if--don't, you know, don't--March.  March.

    DR. MICHALEK: March is a good time.

    [Discussion off mike.]

    DR. GOUGH: If it's not six, and if it's five--and then if it's not six and it's five, you're going to end up with four that aren't delivered by July.

    DR. HARRISON: Well, they've already--actually, the point that's already been made is the last meeting will have to be after July.  The last meeting will have to be August or September.

    So we're talking about a March meeting, let's say a September meeting.  So we're talking about 3, 9--and 6 is really going to be the middle meeting.

    Now, I agree with Mike, though.  What Mike wants is what I've been trying to get for the last 15 minutes, and that is your best guess as to when these chapters are going to roll act--written down.  Okay?

    DR. MICHALEK:  Can we put that as an action item, and we will deliver that by e-mail to Leonard?

    DR. HARRISON: Well, can you do that this morning?

    VOICE: I don't think we can do that on laptops.

    DR. HARRISON:  No, not on laptops, just handwritten.  Just--roughly--

    VOICE: So we can plan our meetings.

    DR. STILLS: So, Dr. Harrison, I really want to underscore that that's important, because a number of us have many--have commitments already for next year.  And so that will help us to plan.

    I think these reports are really critical, that we have the time to review them and do the best job that we could.  And, like you're saying, if we have a heads up as to the time frame for next year, then we can put it in our calendars, and we'll be here to do the work.

    DR. HARRISON: Also, you know, I think that if we set March, June and September at the times that we're going to do this, then I would hope that you all would--the Air Force and its consultants--or contractors--would make an effort to help us out a little bit, too, and not come back and say, "Well, you know, we weren't able to get the stuff done in time--"--you know, sort of thing.

    If we make a best guess as to when we--because what's been said is correct, and that is that we need to plan these meetings, essentially, now for several of the members of the committee to be able to incorporate into their schedules.

    Yes, sir--Jay?

    DR. MINER: Shown to us yesterday was what happened with chapters last cycle.  And here is that outline right here.  And perhaps during the break you can kind of look and see how many chapters you can expect, and so forth.  Would that be helpful?  There are dates on here.

    DR. HARRISON: So you're saying your best guess is that they're going to be the same dates this time as they were the last time?  Same months?  That's a reasonable assumption--give or take a little?

    DR. MICHALEK: Yes.

    DR. HARRISON: And the delivery that we're talking about, of the chapters with the Air Force comments responded to are the "X's," the "S's," the "AF's" or the double-X's?


    DR. HARRISON: The X's.

    LT. COL. ROBINSON: [Off mike].

    DR. HARRISON: Okay.  So, according to this, this is--final data approval was in August.  "General Health" was delivered in November, "Dioxin Assay" in December, "Hematology" in January, "Pulmonary" in February, "Physical Examination" in February"--so we've got--we can look at that--work this out.

    DR. TREWYN: But you're not going to meet the projected timeline on here, then, to have the final report due at the date indicated.

    DR. STOTO:  I think this would be beset worked out off-line here.  I mean, I think that--

    DR. HARRISON: Okay.  All right.

    DR. STOTO: We have--

    DR. HARRISON: Well, before the end of the meeting I want to have some months down, if not dates down.

    DR. STOTO: Let's ask them when they're reasonably confident that a third of the chapters will be done.

    DR. HARRISON: Well, that's what Jay's saying, is that's what's here.  So it's just a matter of doing it without taking up too much more time here.

    Okay.  Can I just--do you want to make a copy of this for me?

    Next.  Let's go on.  As Mike is intimating, we're behind time now.

    So, Joel is going to discuss the cancer findings.

Cancer-Review/Discuss Manuscript

    DR. MICHALEK: Well, good morning.  I am Joel Michalek, Principal Investigator in the study.  Dr. Fatema Akhtar is the statistician, and David Garebrant is an epidemiologist.

    This effort was begun shortly after the last committee meeting.  It was motivated by committee comments on our previous publication of describing cancer and dioxin levels in Ranch Hand veterans.

    What's different about this cancer analysis is that we are now comparing the observed with the expected cancer incidence in both cohorts, relative to national rates.  Yes sir?

    DR. HARRISON: No--I'm sorry.

    DR. MICHALEK: We are using the surveillance epidemiology end result rates of the National Cancer Institute as our standard--the SEER rates--as our standard.  And we're referencing against--in other words, the cancer experience in U.S. males.  And we're also tabulating and analyzing the observed and expected mortality due to cancer, using national rates.

    And, finally, we're conducting our internal analysis against dioxin levels, stratifying and adjusting for when these veterans were in Viet Nam, and for how long, and what their exposure opportunity may have been.  This is new, in other words.

    There are three levels of change in our cancer approach here, as opposed to previous reports.  First, we're using SEER definitions of cancer categories, which we never did before.  Secondly, we have the very latest cancer data, up through December 31, 1999, which has increased our prevalence, and increased our statistical power.  And, finally, we're making adjustments for concepts that were given to us by the committee.


    Somehow the computer is not working.  The machine just locked up.


    DR. HARRISON: Presentation equivalent of the smart bomb.


    DR. MICHALEK: Okay.  This summarizes what I just got through telling you.

    Slide, please.


    This was published in the American Journal of Epidemiology, 1998/99.  These are the latest cancer results published in a peer reviewed scientific journal.  They show now evidence of a dose-response pattern on dioxin versus cancer, on any site, in Ranch Hand veterans.  I've just put this up here to remind you that what you're going to see today is different from what's been published.

    Slide, please.


    The dioxin measurements remain, as by the Centers for Disease Control in Atlanta, the very first measurements were made for the study in 1987, and subsequently in 1992 and '97, by the Center for Environmental Health Laboratory Sciences in CDC.

    We have made an adjustment for dioxin levels below the limit of detection, and below the limit of quantitation, using a published method by Kornon, to divide these limits of detection by the square root of 2 to provide an unbiased estimate of the true dioxin level at these very low levels.

    Slide, please.


    Again, following ideas given to us by the committee, we used published information in the National Academy books on veterans and Agent Orange to stratify the cohort in accordance with when they were in Viet Nam, focusing, most important to us, the period of 1966 to 70, which was the period of heaviest Agent Orange spraying, and taking account of periods prior to that, and subsequent to that, where no spraying occurred, for example between '62 and '72.

    Slide, please.


    And we have also stratified by military occupation, because we know from our own measurements in these study subjects that the enlisted ground crew, for example, have the highest dioxin levels.

    Slide, please.


    Following comments from the committee a year-and-a-half ago, we have devised some measures of exposures opportunity which are approximate.  We attempted to isolate Ranch Hand veterans who spent the majority of their time in Viet Nam, as opposed to Southeast Asia, and for that we've adopted a two-year cut point.  And I have some graphics to show you where that two-year cut point came from.

    And, so trying to isolate comparisons who spent little or no time in Viet Nam.  And to drive that point home further, we carried out another analysis, in addition to this one, where we pulled off only rand hand veterans who spent all of their time in Viet Ham, and only controls that spent absolutely none of their time in Viet Nam--to address the concern that some controls, having been in Viet Nam, are disrupting our efforts to find a treatment effect, because they themselves were in Viet Nam, and possibly exposed to herbicides.

    Next slide, please.


    And so there's two rounds of analyses: external against national rates, adjust for age, calendar period and race, using SEER rates, which are available on the internet.

    Then we strategized those external analyses, again by the same factors that we were using in the internal analysis, by tour-date, military occupation, and exposure opportunity category, and using second order tabulations of tour date by military occupation, which I'm not going to describe here.

    In the internal analysis, we're comparing Ranch Handers with high dioxin levels against comparisons, stratified by these exposure opportunity category strata.

    Slide, please.


    To give you an idea as to why we chose a two-year cut point, this is a plot of the percent of time that Ranch Hand veterans spent actually in-country, against he number of years spent in Southeast Asia.  And you see there's a natural break point here of about two years.  If you isolate those individuals that were in Southeast Asia less that two years, you predominantly pick up a lot of Ranch Handers who were there a hundred percent of their time.  And so by cutting this at two years we're attempting to find a very simple dichotomy that would allow us to pull off Ranch Handers that may have had increased exposure opportunity.

    Slide, please.


    DR. HARRISON: Joel?

    DR. MICHALEK: Yes, sir.

    DR. HARRISON: I though Ranch Hand tours were one year--I mean, not Ranch Hand--Viet Nam tours were one year.  Was that just army?

    DR. MICHALEK: I'm not familiar with the Army.  Most Ranch Hand tours were about 11 months, but many men came back for more than one tour.  And some of the Ranch Hand veterans were in Southeast Asia for a long time--if you back up that previous slide.

    DR. HARRISON: I've read that one of the criticisms of how the war was conducted was that one-year tour.  So that by the time you actually knew what you were doing, it was time for you to rotate out again.

    DR. MICHALEK: Many Ranch Handers were there one tour, which was about 11 months.  Some of them re-upped--re-enlisted and came back.  But some of them were there a long time--as you see--many years in Southeast Asia, but only one year in Viet Nam.

    DR. HARRISON: Okay.  Sorry.

    DR. CAMACHO: I didn't notice it the first time through, that line at 100 percent there--how many people there are.

    DR. MICHALEK: Well, you'll see that in a minute, because we're going to isolate those people.

    Slide, please.


    And this is the pattern in the control group. You see a far different pattern here.  With the two-year cut point we're basically just controlling the two-year period.  Down here are a lot of veterans comparisons who were never in Viet Nam at all--spent zero percent of their time in Viet Nam.  There's about 300 comparisons like that.

    Slide, please.


    So, the external analysis is based on any study participant who was ever at Scripps Clinic, or even participated partially in the study and took only a questionnaire.  If an individual participated at baseline, and took only a questionnaire, and never went to Scripps Clinic, or--we followed that subject just like we do all the other with regard to disease and mortality.

    So this is the maximum possible sample size that could have been derived in the entire cohort: 2,981.  We excluded individuals who we knew from medical records had cancer during or before their tour of duty in Southeast Asia, leaving a sample size of 2,965 for the entire cohort, against national rates.

    In the internal analysis, we excluded individuals with missing dioxin levels, and most of these individuals that were excluded were the partially complying from baseline.  We never changed the subsequent cycles, and particularly the 1987 cycle, when we put our first dioxin measures, giving us a total sample size of 2,438--in the internal analysis--against dioxin levels.

    Slide, please.


    And here's an overall sample size table for our internal analysis, where we have broken out Ranch Hand veterans to three levels: high, low and background, depending on their dioxin level and their estimated initial dose in Viet Nam.  Then the 10 parts per trillion is the 99th percentile of the dioxin distribution in the comparison group, is the 40th percentile of the dioxin distribution in the control group.  Of those that are above 10 parts per trillion in the Ranch Hand group, the median dioxin level in Viet Nam was estimated to be 118.5 parts per trillion.  So this cut point was used to separate those individuals with values above background into a low and high category.  And it's a scheme that's been used in many of our publications.

    And this table--all sample sizes, by the way, are for white individuals.

    Slide, please.


    And here are the sample sizes for the "at most two years" and SEA exposure opportunity strata, and there you see 580 comparisons and about 600 or so Ranch Handers who were in that category.

    Slide, please.


    And here are the "100 percent" versus "zero percent."  So there were 291 comparisons who had absolutely no time in Viet Nam and the rest of their time in Southeast Asia, which would have included the Philippines, Taiwan, perhaps Cambodia--wherever they were--and we're attempting to find out exactly where they were, just now, by reviewing military records.

    And here are the Ranch Handers who spent 100 percent of their time in Viet Nam.

    Slide, please.


    And this is just a table of person-years calculations.  I want to emphasize that, as opposed to our previous reports and articles, we had to spend a lot of time very carefully reviewing race, because race is a factor in the SEER tabulations of national rates, and you must know who is white and who's black and who's neither.  So this is unprecedented for us.  We had never gone through this level of detail before, by reviewing records and photographs, and surnames to decide who's in any of these categories.  It took a lot of time.

    Yes, sir?

    DR. HARRISON: Keep your hands under the table.


    DR. MICHALEK: And here are the Ranch Hand cohort for the period 1966 to 1970.  You can see that's where the bulk of the activity was during the war, or the period of heaviest spraying.  And that's where most of your person-years are.

    Slide, please.


    And the same is true in the comparison group.

    DR. STOTO: May I ask what other [off mike.]

    DR. MICHALEK: Want to say that again?

    DR. STOTO: What "Other" means?  Does it mean Hispanic, or--

    DR. MICHALEK: That would include Hispanic, primarily--and Oriental.

    DR. STOTO: But would SEER--

    DR. MICHALEK: We attempted, the best we could to match the SEER definition of race.  I can amplify that later.

    DR. STOTO: In most federal statistical systems, "Hispanic" is not a race.

    DR. MICHALEK: Okay.

    DR. STOTO: And so I would be surprised as many "others" as Black.

    DR. MICHALEK: Well, we'll need to discuss that in detail, then.  Because I have [off mike.]

    DR. STOTO: The key thing is that you are consistent with SEER.

    DR. MICHALEK: Yes.

    DR. HARRISON: Except this is not self-identified race, this is guessed-at race.

    DR. STOTO: Well, no, for SEER, it is--

    DR. HARRISON: No, I'm talking about for the Air Force.

    DR. STOTO: Yes.

    DR. HARRISON: Because that's not a part of the--

    DR. MICHALEK: It is our best effort to determine race.  It is not easy to determine race.

    DR. HARRISON: But what I'm getting at is, that in the examinations--

    DR. MICHALEK: They were asked to describe their race.  Our exams do that during the exam, I believe.  We have that on the questionnaire.

    DR. HARRISON: They are?

    DR. MICHALEK: Yes.  They classify themselves.

    DR. HARRISON: So why did you go through this other process?

    DR. MICHALEK: Because an individual may call himself black or white, depending on his parentage, and he may--I don't know, there are all kinds of complications here, when you're trying to classify an individual.  It's not easy.

    DR. HARRISON: Well, but--

    DR. MICHALEK: Individuals who call themselves white may not be white, they may be Hispanic.  And if you're telling me that the "other" category in the SEER rates does not include Hispanics, then we need to rethink the racial characteristics.

    DR. STOTO: I think the key thing here is that your classifications are as similar as possible to what's done at SEER.

    DR. MICHALEK: We tried.  Yes.

    DR. STOTO: And I suspect that SEER is based on self-report.

    DR. MICHALEK: All right.  Then we went beyond self-report, because we looked at photographs.

    DR. STOTO: But that's not good.

    DR. HARRISON: Yeah--that's actually.

    DR. MICHALEK: We looked at every single case--one by one.

    DR. HARRISON: You'd possibly classify me as an Eskimo.


    No, I mean self-identification is--I mean, one of the problems with race classification is that it depends on self-classification.  And--

    DR. MICHALEK: It does.

    DR. HARRISON:  --you don't have any objective way--

    DR. MICHALEK: No, we don't.

    DR. HARRISON:  --of determining a person's race.

    DR. MICHALEK: That's right.

    DR. STOTO: So, consistency is what's important.

    DR. HARRISON: Yes.

    DR. MICHALEK: That's right.

    DR. STOTO: If they told you in their survey that they were green, then--

    DR. MICHALEK: Then we've got a problem.

    DR. STOTO:  --then they're "other."

    DR. MICHALEK: I can tell you that it makes little difference as to how we classified the Ranch Handers.

    DR. HARRISON: Okay.

    DR. MICHALEK:  Slide, please.


    Here is the bottom line.  The national rates, in the Ranch Hand cohort, overall, there is no significant increase in the risk of cancer at any site, using the SEER definition against national rates.  We have a standardized incidence ratio of 1.09, and that is not significantly different from 1.0

    I want to remind you that "cancer at any site," following the SEER definition, is not the same as "cancer at any site," following our published paper.  In the SEER definition, "cancer at any site" does not include basal or squamous cell carcinoma, whereas in our previous publications, and all of SAIC reports, "cancer at any cite includes all of those other skin cancers that the SEER does not.  So you're going to find differences here between our published reports written by SAIC, and even our own published papers, because now we're using SEER definitions.

    VOICE: [Off mike]--SEER's--

    DR. MICHALEK: Yes.  Yes.  And I also want you to know that that is one of the reasons why we're seeing things now that we never saw before--because we're using the SEER definitions.  When we use our "all cancer combined" analysis, we see nothing.  And yet we use the SEER definitions, we now see trends.

    DR. STOTO: Does the SEER definition classify melanoma as a "site" or does it--[off mike].  I'm just trying to get a better feel.  Because melanoma is a tumor, and--

    DR. MICHALEK: All right.

    DR. STOTO: You know--

    DR. MICHALEK: Yes.

    DR. STOTO: So this [off mike]--I was just trying to get your classification, versus SEER's classification.

    DR. MICHALEK: This slide is an attempt to put as few words and symbols on a PowerPoint slide as possible.  You will find the full, detailed definition of melanoma in the document, which was all the ICD codes that went into the melanoma site definition.  It's all taken directly from the SEER specifications that we received from the National Cancer Institute.

    So the actual nomenclature, as to whether the name that goes to this category is simply melanoma, or melanoma plus qualifiers is all in the article.  And we have distributed the article to the committee.

    DR. STILLS: [Off mike] I read the article last night when I was going through--and that was a question that I had.  Because I never--in the reports--melanoma.  It seems as though [inaudible].

    DR. MICHALEK: Yes, sir.  And that's--

    DR. STILLS: [Off mike] And as I looked through the article, I had--the rate of melanomas were, you know--some more description as to is it malignant or benign--

    DR. MICHALEK: Those are all malignant.

    DR. STILLS: And that was my assumption.  But I never could get at it in the paper.

    DR. MICHALEK: I'll show you where that is.

    We do see a significant increase in the risk of melanoma in the Ranch Hand cohort, as against national rates relative risk 2.33, the observed 17, we expected 7, and that's significant.

    You see a significant increase in the risk of malignancy of prostate.  We observed 36, expected 23, relative risk 1.58, and that's significant.

    The category called "genitalia"--it's almost entirely prostate.  It was 38 genitalia, and 36 prostate--36 of those 38 are cancer of the prostate.


    DR. GOUGH: You go from all sites, to systems, to specific sites--prostate.  How many comparisons did you make like that?

    DR. MICHALEK: The entire list is in the manuscript that was sent to the committee.  This slide is a reduction of the entire list to reduce the number of slides down to 60.

    The document is 92 pages--

    DR. GOUGH: I know that--

    DR. MICHALEK: The entire list gives all the sites described by SEER are in the--

I cannot recite to you--I can tell you that in one table we counted up the number of significant tests that were conducted--we conducted a lot of significant tests--a lot of statistical testing.  The number of results that were statistically significant was, like 36 percent and the expected number was 5 percent.

    DR. GOUGH: That's what I wanted to know.

    DR. MICHALEK: All studies of this kind are subject to the criticism that we did extensive multiple comparisons--statistical multiple comparisons.  And that is absolutely true, not only of this paper and this report, but all papers and all reports produced by the study.  The study--the report you're going to review will be approximately 4,000 pages.  SAIC will conduct thousands of statistical tests, looking for dose response of health against dioxin.  This is just one example of multiple statistical--

    DR. STOTO: Well, I can see what would make you want to do that, but it would help to acknowledge that.

    DR. MICHALEK: Yes.  And that's acknowledged in the discussion section.  We have a paragraph in there about multiple testing.

    DR. STOTO: Well, but you might be able to do a little more than that; say, give us that number, 36 percent versus 5 percent.

    DR. MICHALEK: It's in the discussion.

    DR. STOTO: Okay. I did read it, and didn't see that.

    DR. MICHALEK: It's there.

    DR. STOTO: Another alternative to consider, is to actually do a correction [inaudible] report both the uncorrected and the corrected--maybe for some selected ones or something.

    DR. MICHALEK: One point to make about the issue of multiple testing is that this was--the scenario as to so-called fishing expedition, where we're out there chasing a p-value, this is not the approach that was taken here.  We developed a strategy based on comments from the committee and we followed it--no matter where it led.  And we developed two levels of dioxin or herbicide exposure opportunity that we tried our best to do, and we used known classifications, by occupation and tour date, taken from published sources.  And we used national published rates.  And we used the prescribed method for internal analyses that we had used in other papers.

    So, in other words, this was not a chasing of a result down a road.

    DR. STOTO: I'm certainly not criticizing you for that.

    DR. MICHALEK: Okay.

    DR. STOTO:  I'm just saying that--


    --it would help to add a little bit more information to--about the comparison.

    DR. MICHALEK: Yes.  I agree.

    DR. GOUGH: The fact that you did--the 36 percent.

    Now then, the next question about that, were about half the statistically significant results increases and half decreases?  Or--

    DR. MICHALEK: I cannot answer that without studying the article.

    DR. HARRISON: Go ahead, Joel.

    DR. STILLS: Joel, the only question I had was--there's genitalia and prostate.  And so you're saying that prostate is a subset of the 38.  So you only have two--and what were the other two?

    DR. MICHALEK: I don't have those memorized.  I know that one of them is a carcinoma of the penis.  I don't know what the other one is.

    DR. STILLS: And it seems as though--you know, I mean, like in the studies that we do, where we report cancer data, we tend to, rather than--you know, obviously there are duplications here, and so we tend to try not to report duplications, especially when you're trying to summarize the data.

    And it seems as though the most important thing here is the prostate, because the prostate--without the prostate being in the genitalia group, you would not have--you may not have that significance.

    DR. MICHALEK: That's true.

    DR. STILLS: And so I think you really need to--I mean, this is very impressive data, and you want to be able to report the most important things.  And it seems as though it may be helpful to take out "genitalia."  You know--I mean, and just have prostate.  Because that's where the findings are.  I mean, "genitalia" takes away from--or confuses the reader, and so you really want to be careful with that.

    DR. MICHALEK: We are always in this dilemma, that we try to show everything, and by doing so, we confuse sometimes.

    DR. STILLS: Right.  And so, I think in this case, I think it may be best to just show what is important.  Because I was a little confused.  And, you know, you helped in terms of--

    DR. HARRISON: Before there are other comments made, let me make sure I understand something here, Joel.

    There's a paper, and this is basically a condensation of what's in the paper.

    DR. MICHALEK: Right.

    DR. HARRISON: So, as--when we're criticizing these tables, we're criticizing--we're not necessarily criticizing the paper, which is out being reviewed, or is ready--potentially ready to go out and be reviewed by whatever journal.  So keep in mind that we're not--it seems to me we'd be better off commenting on the process than worrying about some of the smaller details in some of these tables, otherwise--Joel just has 40 more slides to go, and--


    --and he's got five minutes to do it.

    So--go ahead. I'm sorry.  You were next, Ron.

    DR. TREWYN: I was just going to say, I mean, obviously, what you're going to synthesize for the paper is different from what you're going to put into the final report, where you want everything, you know--ad nauseam--that's probably not the way to put that--


    --but you want everything included there.  For the paper, I mean you want to make it as showy--to punch--without leaving out critical components, but want to be able to punch it up as much as possible, and not confuse the readers out there.

    DR. MICHALEK: The paper currently is 6,000 words.  The maximum any journal we've ever experienced is 4,000 words.  So it obviously has to be trimmed down.

    It's 92 pages.  Most papers submitted to journals are about 35 pages.  And I want to also tell you that the review we gave to you at the outset was held back from you on purpose.  That was a review by David Garabrant, University of Michigan.  All of the comments from him have been incorporated in the document we gave to you, except for one.  He said delete table 15.  We have included all of the tables.

    [Conversation off mike.]

    DR. MICHALEK: He is a co-author.

    DR. STOTO: First of all, I want to second Ron's point.  The other point about genitalia and prostate, is that there's a reason for hypothesizing prostate cancer, and there's other evidence suggesting it.  I'm not aware of evidence suggesting other cancer of the genitalia.

    DR. MICHALEK: The choice of the SEER categories was driven by SEER, not by us.  We used the SEER category definitions across the board, without making any pre-judgments as to which ones to show or not show.

    DR. STOTO: But you could have included melanoma and "other skin."

    DR. MICHALEK: "Other skin" is shown later in the internal analysis.

    DR. STOTO: Yes.

    DR. MICHALEK: There are no national rates for "other skin," only for melanoma.

    DR. STOTO: Okay.  So you do include both respiratory and lung blockages.

    DR. MICHALEK: Because there are--

    DR. STOTO: The 30 are nested in the 33.

    DR. MICHALEK: Probably.  And the definitions are given in the article.

    DR. STOTO: Right.

    DR. MICHALEK: So a lot of them are nested, that's true.  But that's the way the SEER data is presented.

    DR. STOTO: Well, I guess that for a summary table like this, I would go with prostate and lung, and drop off respiratory and genitalia.

    DR. MICHALEK: Okay.  Thank you.

    DR. HARRISON: That having been said, am I understanding this slide correctly that these are just the White Ranch Handers?

    DR. MICHALEK: These are Whites only, yes.  But the article shows everyone.

    DR. HARRISON: And what you're saying is--the specific question I have is: was there an effect on prostate cancer in African Americans?  And if it wasn't, do you think it was because of smaller number, or what?

    DR. MICHALEK: Five percent of the cohort is Black, so the number is extremely small.  The numbers are in the paper.

    DR. HARRISON: Okay.

    DR. MICHALEK: Whenever possible, we broke it out by race.  You'd see a table of zeros--many zeros--if I gave you a table on Blacks, for example.

    DR. HARRISON: Okay.  Got you.

    DR. MICHALEK: And here the same table for the comparison cohort against national race.  And you see--here you see an increase in melanoma not significant; and increase in prostate which is significant.  Again, the genitalia is driven by the prostate, because four of the 54 of the 55 are prostate cancers in this cohort.

    Slide, please.


    And once again--and now we're going back to the occupational strata derived from our own information snout exposures.  And we see--I'm going to now highlight--I'm not going to show every result in the paper.  In an effort to bring the number of slides under control, I'm showing you only prostate and melanoma now in subsequent slides.  I'm not showing you second order effects, which would be cross stratifications of tour date by military occupation, for example.  I'm showing you only what are called "main effects."  I'm concentrating only on prostate and melanoma.

    And here, for example, is the pattern in the Ranch Hand cohort.  You see an increase in the enlisted flyers, of relative risk 2.27, and that reaches  significance.

    Slide, please.


    In melanoma, we see increase in ranch-hand officers, a relative risk of 3.02.



    DR. HARRISON: Time at the golf course.


    DR. MICHALEK: Except that we see a trend against dioxin levels against melanoma, which suggests that it's not simply a matter of sun exposure.

    Slide, please.


    This is the comparison prostate, which is elevated in both enlisted flyers and officers.



    Comparison, melanoma--borderline significant increase in officers, but not in the other two occupational strata.

    The comparison cohort consisted of air and ground crew who serviced or flew C-130 aircraft in the Southeast Asia region but did not spray herbicides.

    And here's a breakout by the tour dates--the period of 1966 to '70 is the period where we have the most Agent Orange spraying is where we have the most statistical power, and it is the period where we see the increase in risk of prostate cancer in the Ranch Hand cohort.

    And in melanoma, the same pattern: increased risk of melanoma in Ranch Hand veterans against national rates.  Among those Ranch Handers who were in Viet Nam and--and on the Ranch Handing between 1966 and 1977, the period of heaviest Agent Orange spraying.

    Slide, please.


    In the comparison cohort, the period '66 to '70 exhibits an increased risk of prostate cancer against national rates.



    DR. STILLS:  Joel, did any of the Ranch Handers in the years 1966 to '97 have multiple tumors--some of them may have had melanoma and prostate cancer.  Did you ever see multiple--

    DR. MICHALEK: Yes, we have many veterans with multiple tumors, and all we're analyzing here is the occurrence of the first tumor.  And all these tumors are malignant.

    We have not made an attempt to describe or tabulate multiple tumors, but we could do that.

    DR. STILLS:  It would be interesting to look at these two and get a feel for the impact of exposure.  I mean, it would strengthen, the no, that that's what happened in terms of cancer.

    DR. STOTO: I had another comment about this breakout by year.

    Obviously, most of the cancers are in '66 to '70, because those were most of the experiences.

    DR. MICHALEK: Yes, it is.

    DR. STOTO: It ended in '72--right?

    So one thought would be to look at that period versus all other years.

    DR. MICHALEK: True.

    DR. STOTO: First and last years--prior to '66 and after '70.

    DR. MICHALEK: Right.  You could do that.

    [Comments off mike.]

    DR. OSEI:  Help me out.

    DR. MICHALEK: The comparison group--flew or serviced C-130 aircraft in Southeast Asia but did not spray herbicides, about more than half the comparison group were actually in Viet Nam, but the were not in the Ranch Hand unit.  It's been a criticism of the study that our comparisons are themselves Viet Nam veterans, and so we may have frustrated our own attempts to find a treatment effect by including Viet Nam veteran comparisons.  This was designed into the study in 1978.  Many veterans are frustrated that we don't have control cohorts: one that went to Viet Nam, and one that didn't.

    Well, we do have a Viet Nam veteran cohort.  So the analyses that you see here today are an attempt to address those concerns.

    DR. STOTO: This has been a recurring question.

    DR. MICHALEK: Yes, it has.

    DR. STOTO: And, of course, it gets to what the hypothesis is.

    DR. MICHALEK: And what is the direct comparison group.

    DR. STOTO: Right.  And we have had Ranch Hand versus non Ranch Hand.  We have had Viet Nam versus not Viet Nam, and now you--I think it's become now, Ranch Handers in Viet Nam versus--

    DR. MICHALEK: Comparisons who didn't go to--

    DR. STOTO:  --non Ranch Hand, non-Viet Nam.

    DR. MICHALEK: It's Ranch Handers who were in Viet Nam, versus comparisons who were not in Viet Nam in some of our analysis, that's true.

    We're going to attempt to address--

    DR. STOTO: Which is an interesting question, but a different one.

    DR. MICHALEK: It's a different one.

    DR. STOTO: Than the two we have been focusing on in the past.

    DR. MICHALEK: That's true.

    DR. HARRISON: Well, the problem is that anything that we do is inferred.  You know, this is the issue with surrogates.

    DR. MICHALEK: Right.

    DR. HARRISON: And we've just--so it's nor really different, it's just a different surrogate.

    You can argue that makes it different, but--

    DR. STOTO: I think it does make a difference.  I mean, in one case we're talking about the chemicals that were a part of the Ranch Hand experience.  In another one we're talking about all sorts of things that happened in the country--

    DR. MICHALEK: Yes.

    DR. STOTO:  --and now we're kind of putting both of those together.

    DR. MICHALEK: Yes, we are.

    DR. STOTO: We're putting both those together.

    DR. MICHALEK: We're attempting to do that.  That's right.

    DR. STOTO: So if it turns out something turns up positive here, we still can't isolate whether it's Ranch Hand or Viet Nam.  We can say it's one or the other.

    DR. MICHALEK: Yes.  We need to study the responses--

    DR. TREWYN: But for those who served, that's okay.

    DR. MICHALEK: I would suggest that we put that idea on the table--

    DR. TREWYN: I mean, because you are--if it was their service in Viet Nam that caused it--whatever these adverse health effects--that is useful information for them to have.

    DR. STOTO: But if that's the question, the primary comparison should be Viet Nam versus non-Viet Nam.

    DR. GOUGH: But you're attempting to do that right now.

    DR. MICHALEK: Yes, we are.

    DR. GOUGH: And with the data you have--

    [Multiple comments off mike.]

    DR. HARRISON: Well, let's--let me ask--the point's been brought up, and that is what the objective of this entire study is.  And the objective of the study--I'm asking this question.  Let me rephrase this.

    Is the objective of the study the health effects of service Viet Nam, or is the objective of this study the health effects of Agent Orange?

    DR. MICHALEK: Are you asking me to tell you?

    DR. HARRISON: Ahh-no.


    DR. GOUGH: Well, the original--the title of the study is about herbicides.

    DR. MICHALEK: Yes, it is.

    DR. GOUGH: But I think that the legitimate concern has come up about men who served--Ranch Handers who were exposed to the Viet Nam experience, whatever that may have encompassed, plus a more specific exposures of Agent Orange.  And Joel's trying to divide it up now and say, "This is my best attempt to separate out the effects of Agent Orange, and in order to do that, I also have to know what happens to the guys who were just in-country."

    DR. CAMACHO: What was in-country was where it was dropped.  It got dropped--if you'll look at the tables and where the stuff got dropped, and where troops go in, and when troops walked around and, Christ, people were walking and swimming in the stuff up on the DMZ or in certain--

    DR. MICHALEK: That's hard to cover.

    DR. CAMACHO: No, it's just applied research, because you're going back to a huge fight that took place over three years.

    DR. STOTO: I think it's fair to do this analysis--I think it's useful to do this analysis.

    I do think, though, that it should lead to an extensive discussion in the introduction to the paper and the discussion session.  This is not just all we're doing is--you're changing the hypotheses.  Which is fine.  But you need to be absolutely clear that that's what's going on.

    DR. MICHALEK: Thanks.  I'd appreciate further comments by e-mail from you or whatever you suggest.

    DR. OSEI:  We may have a further--the issue is dioxin.  Do you by any chance have data on when, you know--what industries that could use [off mike.]

    DR. MICHALEK: Yes, I do.  And those data have been published.  And they have found--those other studies have found increased risk of cancer in the factory workers that made herbicides in the United States.

    DR. OSEI:  Then how about a comparison there.  Because that may--some of the issues--

    DR. MICHALEK: And many of the patterns seen in those studies parallel these--yes.

    DR. TREWYN: Well, the charge was not "dioxin," it was "herbicides"--

    DR. HARRISON: Yes.

    DR. TREWYN: --okay--and all of those bases were saturated with herbicides.  There wasn't a blade of grass growing at most of them, because they were hand spraying herbicides all over everything around them.  And it may not have been Agent Orange in all those cases.  So, again, the exposure to those on the bases could have been different.  And just looking for dioxin may not help you sort that out.  So, just a--

    DR. STOTO: The industrial studies that you were asking about--some of them focus on dioxin, and some of them do focus on herbicides more generally.  In fact for prostate--one thing I noticed here, not just those ones you cited there.  There was a study in, I think it was--maybe in farmers, with some strong results in there.  And that was farmers exposed to herbicides, not just dioxin.

    DR. MICHALEK: That's right.

    DR. OSEI: The reason I'm bringing this up, because I was reading your intro.  It seems that the focus, the starting point, was dioxin in most of the instruction to your paper, instead of herbicides in general.


    DR. MICHALEK: We are in a bit of a quandary, in that we measure dioxin.  We don't have a biomarker for 2-4D, 2-45T or any of the components of the other herbicides. We have a biomarker for exposure to the noxi-herbicides, which are those that were contained in many of them.

    It is the only study in the world, by the way, with a biomarker like this.


    DR. HARRISON: I'm sorry.  Finish your sentence.

    DR. MICHALEK: It's the only Viet Nam veteran study with a measured biomarker.  And, by the way, this is the only Viet Nam veteran study of cancer incidence against national rates.  This is the only SIR analysis against national rates in any veteran study.


    DR. HARRISON: You guys that do this epidemi--whatever it is stuff--


    --isn't it--do you think it would be helpful--the difficulty with this surrogate is it's even more tenuous--it is less--it is more tenuously related to herbicide exposure than, say, dioxin measurements--dioxin levels.

    DR. STOTO: No, I don't think that's true.

    DR. HARRISON: The method-time, or the year that you were in Viet Nam does not have a direct relationship to herbicide exposure that a dioxin level has.

    DR. STOTO: Well, they stopped using Agent Orange after--

    DR. HARRISON: Let me rephrase this.  If I had a high dioxin level in my blood, there's no doubt that I was exposed.  If I was in Viet Nam when Viet Nam was dripping with Agent Orange, there is no assurance that I was exposed.

    DR. STOTO: Well, no.  No.  If you had a high exposure level, you could have come back in worked in the factory.

    DR. HARRISON: No, I'm just saying the fact that I was there is no assurance that I was exposed.  But if I have a positive dioxin level, there is assurance that I was exposed--whether it was there or anywhere else, there is assurance.

    So the one--even granted that the dioxin level is a more specific surrogate than the fact that I was in Viet Nam--

    DR. CAMACHO: The fight behind all of the issue, you're taking at that level that there's an implied research-that this all came from way back in the '70s.

    DR. HARRISON: I understand.  But it seems to me that that is the point that needs to be made about this presentation.

    DR. STOTO: I agree with you.

    DR. HARRISON: That's all--that it's a less direct surrogate.

    DR. STOTO: I would just say that it's different.

    DR. HARRISON: I think it's important that it's less direct.  You know, it's--it reminds me of how salmon calcitonin got approved for the treatment of osteoporosis based on an increase in total body calcium, measured by--oh, some kind of ion displacement technique.  And when the real surrogate is bone fractures, of which a more direct surrogate is bone calcium, of which a very indirect surrogate is total body calcium--which turned out not to be related, when it turned out that shortly thereafter people treated with salmon calcitonin actually had an increase in fracture rates in certain locations.

    DR. STOTO: Well, this is common in epidemiology that since you don't control who's exposed to what, you have to kind of live with what's here.  And I think that the point that needs to be made is there really are three different hypotheses here.

    One is that dioxin is a health risk--is a cancer risk, and we can address that because we have these measures.  The other one is that being part of Ranch Hand--whatever that means--is a risk, and we can look at that.  And the third one is being in Ranch Hand in Viet Nam, versus being a comparison who didn't serve in Viet Nam--that's an interesting question.

    DR. HARRISON: But, Mike, I've never thought of the question about dioxin.  To me, dioxin--I've never thought that the question was whether dioxin was a health risk.

    DR. STOTO: Well, that's what--

    DR. HARRISON: The question is whether herbicides are a health risk, and dioxin is the surrogate.  We're not testing dioxin directly.

    DR. STOTO: And, I think, the surrogate for one of the herbicides.

    DR. HARRISON: That's right.  And so it's--like all surrogates, it's imperfect.  But it's a surrogate.  We're not testing dioxin.  That's not the objective of the study.

    DR. STOTO: I understand that wasn't the original objective, but that has become an important topic in science.

    DR. HARRISON: Well--

    DR. STOTO: And this has some bearing on it.

    DR. TREWYN: Well, it's become a pertinent topic for the VA, because they want to use--if you can't prove that dioxin caused it, then they don't have to pay benefits to anybody who has adverse health outcomes.  And this gets--this opens up more areas.  If you take the dioxin away from that argument, well, all of a sudden, you've got the potential that the VA has to pay a lot more health benefits, and they don't want to do that.

    DR. STOTO: So it's a valid scientific question that we have some evidence that relates to it.

    DR. CAMACHO: We're at this imperfect study level, in a sense, because of the fights of 30 years ago.  When people went down to CDC and CDC was pressured not to do certain studies, and to view certain studies as everybody had--was in the military, but didn't go to Viet Nam, went to Viet Nam.  They could have had that--they still could have done that, as Massachusetts.  But we're not there.  We're here.

    [Comments off mike.]

    DR. HARRISON: Wait a minute.  Wait a minute.  Wait a minute.  Yes, we can.

    But Paul, let's stick with the science here.  Okay?  Mike, too.

    DR. CAMACHO: You're asking me not to say anything.

    DR. HARRISON: No, no.  No.  No, I'm not.  And I apologize if you feel that way. What I'm saying is, I'm still concerned about the science--

    DR. CAMACHO: Make it clear about the three pieces.  I can go along with that.

    DR. HARRISON: Maybe I can't make it clear about the three pieces.  I'm beginning to falter.


    DR. GOUGH: I want to echo what Paul just said.  I think despite the fact that you've got a paper that's three times longer than is going to get accepted, that it would be far better to truncate these tables--as your other author suggests--as much as possible, and to explain, in this paper, where you're shifting the weight; explain so that everybody in this room will understand that we're now doing something new, and we're doing it because if the results go this way we would--this happened, and the results go that way.   That's far more important, I think, than the exhaustive tables.

    And it's got to be done, because otherwise this fight's going to come up--it's not a fight, but this confusion's going to arise in every person's mind who reads this paper.

    DR. MICHALEK: Yes, I agree.

    DR. TREWYN:  --goes into the final report.

    DR. STOTO: And Paul is right.  This does reflect what happened 30 years ago.  But we have what we have, and I think that what Mike said is exactly the way to handle it.

    DR. HARRISON: Any other comments?  We still have another 30 or 40 slides to look at.


    There's plenty of opportunity.

    You're doing great, Joel.

    DR. MICHALEK:     Next slide, please.


    And the comparison group--melanoma, we see no significant increase in risk of melanoma.

    Slide, please.


    Here are the two so-called Viet Nam exposure index strata.  In the first, we have the Ranch Hand veterans with less than two years in Southeast Asia--which tends to isolate Ranch Handers with a lot of time in Viet Nam.

    Here are the Ranch Handers with a hundred percent of their Southeast Asia experience in Viet Nam.  And in both strata we see a significant increase in the risk of prostate cancer against national rates.

    DR. HARRISON: But those are all in 17, right?

    DR. MICHALEK: Those are the same numbers you saw in previous slides.  There were observed 17, expected 9.

    DR. HARRISON: Okay.

    DR. MICHALEK:      Slide.


    And here's the corresponding slide for melanoma.  We see in both strata increased risk against national rates of melanoma.  In the Ranch Handers who were there 100 percent of their time, we observed 12, expected 4.  The relative risk is 3.

    Slide, please.


    In the comparison cohort, looking at either strata, we see no significant increase in risk against national rates on the occurrence of prostate cancer.  And the same is true in the comparison cohort on the risk of melanoma against national rates.  In either of the two strata we see no significant increase against national rates.

    It's interesting to note that although we say significant increases in the risk of cancer incidence, we see incidence, we see no significant increase in the risk of cancer mortality against national rates in either cohort.  Just an example: in the Ranch Hand cohort, these same SEER categories now are used to tabulate observed and expected deaths from cancer at any site, for example, and you'll see-

    DR. HARRISON: Joel, that shouldn't surprise you. People don't die from prostate cancer.

    DR. GOUGH: But what about melanoma?

    DR. MICHALEK: No, again I'm trying to reduce the number of slides, and did not include everything.  I only included data for which there were some counts.  And many of these SEER categories there were no deaths, or the counts were zero or one, so I didn't include those.  They are in the article.  If you look at Table 12 or so in the paper, you'll see them.

    DR. GOUGH: But melanoma is a pretty fatal tumor.

    DR. MICHALEK: Again, the actual counts are in the paper I did not include all data in the slides.  It was just an attempt to give you a thumbnail sketch of what's in the article.

    DR. HARRISON: Does anybody know if melanoma is more common in males versus females?  Is there a sex-based difference in the incidence of melanoma?  Does anybody know that?

    [No response.]

    Dang.  Okay.

    DR. MICHALEK: There's an important reason why we don't.  Because all the veterans in this study are male.

    DR. HARRISON: No, no, no.


    DR. GOUGH: No, we're making comparisons with the male SEER data.

    DR. MICHALEK: Yes, sir, we are.

    DR. HARRISON: No, I'm just asking in general.  The reason I'm asking is the two things that have popped up--of the two things that have popped up, prostate cancer is certainly androgen dependent, and it would be nice--not nice, but it would be comforting to--as far as the data was concerned, if melanoma was more commonly found, and also because androgen levels have something to do with aggression and so on, it may be that those people who spent more time in Viet Nam, who voluntarily spent more time in Viet Nam might have been pumping a little more testosterone than those who were out.

    DR. MICHALEK: The issue of volunteerism has been raised many times over the beginning of the study.  You know, the current physical examination, we are displaying the tours of duty for each veteran and asking him which, if any, of these tours did he actually volunteer or attempt to manipulate his career so he would actually end up in Viet Nam.

    So, for the first time, we're getting that data on volunteerism, although we haven't summarized it yet.

    DR. HARRISON: I assume, though, that those people who spent eight years in Viet Nam had to have volunteered at some point, no matter how--

    DR. MICHALEK: That's an assumption.  We'll find that out when we actually review that data.

    DR. HARRISON: Okay.

    DR. GOUGH: You're making a very interesting point.  Because, I mean, it says a lot about the VA--


    --it's available to them.  If they go and they get good care, then we've got--with the rest of our cancers, and we have prostate cancers where there's a significant increase in one group compared to the national, and without a significant increase--not made a significant--but there's a decrease here.

    DR. MICHALEK: Well, you see, for example, in the Ranch Hand groups there were no deaths from cancer of the genitalia.  Of course those include mostly prostate.  So these Ranch Hand veterans are not dying from prostate cancer.  They're dying from something else.

    DR. OGERSHOK:  Don't forget, every five years they go through an exam here.  I mean, we picked up a lot of this stuff there, and they get treated.

    DR. HARRISON: That's true.  That's true.

    DR. MICHALEK: Well, we are comparing the mortality experience of those who attended Scripps exam, and we didn't see any significant difference at all.

    So, that's a very interesting hypothesis, but it doesn't seem to be held up by the data.

    Slide, please.


    In the comparison cohort, we see no significant in the risk of mortality from cancer at any site or any specific site.



    Now, I'm going to display for you some of the internal analyses that we conducted using the actual dioxin measurements, rather than simply comparing against the national rates.  We're now focusing on those individuals that actually have dioxin measurements.  And we're now within that line  strata of exposure opportunity: those that were there less than two years, and we see a significant increase in the risk of cancer at any site in the Ranch Hand low and high categories against the comparison in this analysis, with a relative risk and 2 and 2.2 in the low and high categories.  We've never had this before



    And looking at the Ranch Handers who were there a hundred percent of their time, and a comparison to "never in Viet Nam"--there were 291 of them--we see a significant increase--actually a stronger increase, in the risk of cancer at any site, using the SEER category definition in the Ranch Hand cohort, accommodating dioxin levels.

    Slide, please.


    In prostate cancer, we see a significant increase in the risk of prostate cancer in the high-dioxin category, with a relative risk of 7, against the comparison cohort.  This is among those individuals that were there less than two years.



    And in melanoma we see significant increases in the risk in the low and high category, dioxin exposure against the comparison cohort.



    Among the Ranch Hand veterans who were there a hundred percent of their time, and the comparisons that were not there at all, we see a significant increase in the risk of prostate cancer in the high-dioxin category.



    And the same is true of--well, not the same--not quite true in melanoma; relative risk 5.5, which does not reach significance.



    DR. HARRISON: Joel, those relative risks are pretty high, and the p-values are pretty--unimpressive.  Are those because of small numbers again?

    DR. MICHALEK: Yes.

    DR. GOUGH: I had the same concern about this analysis I had about the diabetes analysis after I thought about it last time.  You have stratified the Ranch Hands to be the highs and lows; you've taken the comparisons to get them homogeneous.

    If you stratify the comparisons between low and high, do these--

    DR. MICHALEK: On dioxin, you mean?

    DR. GOUGH: Yes, on dioxin--what happens?

    DR. MICHALEK: We have stratified the comparison group and we see differences within the comparison group.  I have not displayed those in this--because that was not the thrust of these analyses.

    DR. CAMACHO:  There's little--there's not much variance in the comparison group--

    DR. MICHALEK: The dioxin levels ranks--

    DR. GOUGH:  --with low levels?

    DR. MICHALEK: Yes, dioxin levels in the comparison groups range from zero to 10.  In the Ranch Hand group they range from zero to 600.

    All we're doing here is stratifying, not by dioxin level with the comparison group, but whether they were in Viet Nam or not.

    DR. HARRISON: But background and lowest dioxin levels?

    DR. MICHALEK: Background and low--these are Ranch Hand veterans who were there in Viet Nam for a second time, and who had the background levels of dioxin.

    DR. GOUGH: Yes, that's what I'm saying.

    DR. HARRISON: What you're saying, though, is that even at background levels, the relative risk is 4.

    DR. MICHALEK: Right.  That doesn't reach significance.

    DR. GOUGH: But given what happened with--given the relationship--Paul, you're exactly right, because if you plot this out on a log-base 2 table, and the comparisons go from essentially zero up to 10, and the Ranch Hands go from zero to way over here.  But then if you cut in the quintile which drove it the last time, diabetes goes up with increasing body level of dioxin in the comparison, it also goes up with increasing levels in--I mean, it goes up with increasing dioxin levels in both the comparisons and the Ranch Hands, which is a puzzle.  Because if it's dose response, you'd expect it to be much, much higher in the Ranch Hands.

    DR. MICHALEK: We did not analyze dioxin versus cancer in the comparison group in this study--in this particular manuscript.

    DR. HARRISON: We're going to need to conclude this discussion in about five minutes.

    Do your best, Joel.

    DR. MICHALEK:      Slide, please.


    So here are the conclusions of--


    --all of which I already stated, and you already know what they are.

    This study was not designed to do this, and we tried our best to do it, following comments from the committee.



    Increased risks against cancer rate an site, not increased against the national rates.  I'm not going to say these words.  You know what they are.  We just got through talking.



    Again, this is summarizing the same slides you just saw.



    We saw no evidence of a healthy-worker effect, as regards incidence.  Healthy-worker effect would be expressed if the incidence, for example, were much less than the national rates, and we are seeing increases against national rates in the incidence of prostate and melanoma in the Ranch Hand cohort.  However we're not seeing any increase in cancer mortality in either cohort, which suggests they're getting better than average health care.



    And here are the limitations that you all know about.  And these are the results you already mentioned, Mike, about the multiple testing.  In one particular table we have nine significant results out of 34 tests.  And I didn't make an attempt to break those out as to whether they're adverse or not adverse, but we can certainly do that in our revision.



    And those same caveats apply to all of our work.  We don't know the initial dose.  There was no dosimetry in Viet Nam.  We're using a first order kinetic model to estimate the initial dose to estimate the initial dose that they had while they were in Viet Nam.



    The strengths are that we think we have full ascertainment: every single study subject is followed up with our own staff by telephone to make sure they see their doctor as following Scripps physical exams, and we have record review to back up every single case that's been shown in every table.  And we've had good compliance.

    And this is the only Viet Nam veteran study with dioxin measurements in both cohorts.



    And these are the same bullets we just talked about.

    Thank you very much.

    DR. HARRISON: Okay.  Any other questions?

    DR. STOTO: I have a suggestion--to deal with the length, and that is to break it up into two papers, for the internal and the external.

    DR. MICHALEK: Yes, that's on the table as an option.

    DR. STOTO: Get's two publications instead of one.


    DR. HARRISON: Okay.  So what we've managed to do is not do the mortality paper before the 10:15 break.

    Let's see if we can't make up a little bit of the time as we go through, so we have a break now at 10:15.  Let's try to get back here--why don't we try to do this: why don't we try to get back here in 10 minutes, get started promptly on the mortality paper; see if we can't finish the mortality paper a little early, and squeeze the diabetes paper in and let the public presentation run over into lunch a little bit.  Okay?

    So--10 minutes.


    DR. HARRISON: Let's go.

Mortality-Review Findings from Latest Paper

    DR. MICHALEK: I have a statistician with me from Brooks Air Force Base.

    This is summary slide of a paper on post-service mortality in the Ranch Hand cohort, which was submitted and rejected by the American Journal of Epidemiology.  And I want to go through the results and the referee comments.

    In this analysis we are comparing the mortality experience in the Ranch Hand cohort against all 19,000 comparisons in the comparison population--which is different than what you just saw in the previous talk.

    In the previous talk we had isolated comparisons who had attended the Scripps exams, or the Kelsey-Sebold exams, questionnaires.  This involves all 19,000 comparisons. And so we have a lot of numbers, but we have fewer co-variants.  we are unable to adjust for known risk factors for certain kinds of death; for example, in these analyses we can only adjust for age, race and military occupation.

    We're looking at the same standard of causes and cause-specific assessments of mortality we have done in previous reports.  The cut-point for mortality was 31 December 1999, which was the same cut-point we used in the previous talk.

    And this updates a published paper in 1999 in the American Journal of Epidemiology.  And it was recently submitted and rejected by the same journal.



    So here is risk.  As I said, it was from the start of service in Viet Nam to 31 December 1999, all outcomes verified by record review, following ICD rules.  Twenty-two Ranch Hand were killed in action, and they're not included in this report or any of our reports, actually.

    Slide, please.


    We are breaking out by military occupation, which we already know correlates with dioxin body-burden in Ranch Hand veterans.



    The statistical analysis, now, is an internal comparison of Ranch Hand mortality against comparison mortality, using something called "proportional hazards model," adjusted for birth date and military occupation.

    The summary statistics are relative risks.



    At the time--by December 31, 1999, of the 1,262 Ranch Hand veterans who ever existed--that's minus the 22 killed in action--there were 440--I'm sorry--this doesn't give me a total number--this is numbers at risk.  Of the 1,262 Ranch Handers who ever existed, to the 19,078 of the comparison population, and there it is broken out by military occupation.



    And these demographics will show you that the groups are approximately comparable on dates of birth and military occupation and race.



    And so here are the mortality analyses.  We have approximately--we have 14.7 percent of the Ranch Hand veterans have died as of 31 December 1999, versus 12 percent in the comparison cohort.  And that relative risk is borderline significantly increased.  And that's a first in our study.  Prior reports, the relative risk was a lot closer to 1.0.

    There are reasons for that increase, which I'm going to bring out in the subsequent slides.  This is not a comprehensive of all the causes.  Once again I'm trying to reduce the number of PowerPoint slides.  The document shows all causes.  I'm only showing those for which there are some counts to show.

    Cancer risk, for example--cancer mortality, there is no significant increase in the risk of cancer mortality.  Previously reported in the published article in the American Journal of Epidemiology--again, 118 deaths, and there we used a similar methodology instead of a Cox model, we had an expected of 120, a relative risk of 1.0.  So you see, in the last couple years there's been a change in the mortality experience, relative to the comparison cohort in the adverse direction.



    And here you see one of the reasons why we have an increased risk--borderline significant increase in all-cause mortality, is being driven by circulatory deaths in the Ranch Hand cohort; relative risk 1.3, also borderline significant.  And there's a reason for that, which you'll see in subsequent slides.

    In previous analyses we had a significant increase in the deaths due to digestive diseases, which no longer reaches significance.

    And on asterisk on the 66 means that in the previous publication we have 39 deaths from circulatory diseases, whereas now we have 66 in the Ranch Hand cohort.

    Slide, please.


    DR. STILLS: Joel, what sort of circulatory diseases do you mean here?

    DR. MICHALEK: Primarily heart disease.  And I have another slide breaking that out by specific type of heart disease.

    DR. STILLS: Thank you.

    DR. HARRISON: Joel, in your discussion of this, do you relate that at all to the increased risk of diabetes?

    DR. MICHALEK: We don't have diabetes as a cause of death on death certificates.

    DR. HARRISON: Diabetes doesn't cause death.  Diabetes causes circulatory disease.

    DR. MICHALEK: Yes.  And we're attempting to explore that now.  We're trying to relate these deaths--obverse of what we see at Scripps, and that would include diabetes.  I'm not ready to show that.

    DR. HARRISON: Maybe let me ask the question another way: of those people who died with circulatory illnesses, were a disproportionate number of them diabetic?

    DR. MICHALEK: That analysis is in the works.  I'm not ready to show that.

    DR. HARRISON: Okay.  So that's the answer.  Okay.  Thank you.

    DR. MICHALEK:     Slide.


    And here are the three external causes: accidents, homicides and suicides.  And you see no significant increase in the Ranch Hand cohort on those three causes.

    And now we start to see what's really driving this increased mortality in the Ranch Hand cohort.  It's coming from the enlisted ground crew.  And you'll see soon that what's happening in the enlisted ground crew is that those--the overall mortality in that cohort is being driven by circulatory-disease deaths.

    However, we do have also a significant increase--although small numbers--in Ranch Hand administrative officers.  There are Ranch Hand officers who did not fly airplanes but--and certainly did not handle herbicides, but are experiencing increased mortality.  And I've listed them all down here.  There are seven deaths, and there are certainly no apparent pattern there: one accident, one suicide, one cancer, and so on.



    Cancer mortality by occupation--cancer and heart disease are the two causes of death that have enough numbers so we can actually break them out by military occupation, which is our best surrogate for herbicide exposure.  And you see here, with regard to cancer mortality, there is no apparent pattern here that would suggest an adverse relation between dioxin or herbicide exposure and cancer mortality.



    However, when we look at circulatory disease mortality, now we see the reasons for these increases are driven by 40 enlisted ground individuals who have died of diseases of the circulatory system; 6.8 percent, against 3.7 percent enlisted group comparisons, which is a relative risk of 1.7, and that reaches a significance.

    In our previous article, we had 24 enlisted ground who had died from diseases of the circulatory system, and expected 16, a relative risk of 1.5.  And so what's happened is that the result becomes stronger since we published our last paper.

    DR. STILLS: Are they included in the 40?

    DR. MICHALEK: Those 24 are included in the 40.  Yes, this is cumulative.

    DR. STILLS: Is there any co-that you could identify?

    DR. MICHALEK: Yes.  That's coming in a subsequent slide.

    DR. HARRISON: What was the question?

    DR. MICHALEK: Any particular kind of heart disease that would explain these--and the answer is yes, and that's coming in another slide.

    And here is--the analysis that we do, driven by referee request that we break out cancer deaths as to those that occurred within 20 years of service in Viet Nam, versus those that occurred subsequent to 20 years, looking for a latency effect.  And we see no increase in risk among veterans who had cancer within 20 years.  However--

    Next slide, please.


    --after 20 years, there is a suggestion of increased risk, but certainly nothing statistically significant.



    Among those enlisted ground personnel who died from disease of the circulatory system, the majority of those deaths were cause by atherosclerotic heart disease.  There you see 28 of the 40 were caused by atherosclerotic heart disease among Ranch Hand, or 4.8 percent versus 2.6 percent in the comparison cohort; p-value of 009.

    DR. HARRISON: Joel, atherosclerotic heart disease produces cardiomyopathy--atherosclerotic disease--I'm sorry--when applied to the heart can result in cardiomyopathy; in terms of the cerebrovascular system can produce strokes.  So how was atherosclerotic--"atherosclerotic" doesn't kill you.

    DR. MICHALEK: I'm reporting to you the underlying cause of death as recorded on the death certificates from these men.  We have not yet finished an analysis comparing these results with what we see at Scripps clinic.  We have the health records on many of these men from their repeated physical exams at Scripps.  They are not reflected in these numbers.  And that is coming in a separate paper to be presented to the committee at the next meeting, if they--

    DR. HARRISON: So you're saying on the death certificate it was just the single word "atherosclerotic?"

    DR. MICHALEK: This is an abbreviation of a specification that was given to us from the underlying cause of death on death certificates.  I can provide a more detailed description of that if you would like it.

    DR. HARRISON: So it may well have been "atherosclerotic XX," "atherosclerotic XY," "atherosclerotic XYZ."

    DR. MICHALEK: Yes.  There is an ICD code associated with that word, and I can tell you the exact specification on those codes if you would like to have that.

    DR. STOTO: The underlying cause, rather than--

    DR. MICHALEK: Yes, they're underlying causes--underlying cause of death.

    DR. HARRISON: Only in terms of whether it adds credulity to the statistical findings--

    DR. MICHALEK: I believe those codes are in the article.

    DR. HARRISON:  --it would be nice to be able--

    DR. MICHALEK: Yes.

    DR. HARRISON:  --to related things up.

    DR. MICHALEK: Those are in the article.

    DR. HARRISON: Question a little bit off--does anybody happen to remember if there is a significant difference in proteinuria between the Ranch Handers and the comparisons?

    DR. MICHALEK: Proteinuria would be in the renal functioning?

    DR. HARRISON: Yes.

    DR. MICHALEK: As I recall, there were no significant findings in renal.  We've never seen significant findings in renal in the entire study.  However that doesn't mean they're not there, and we can look at that again.

    DR. HARRISON: It may be that you just didn't have enough people.  But you know if you had a group with more diabetes than another group, you'd expect there to be some diabetic nephropathy.  And you know, if they had enough you'd expect to see some renal failure, and you'd expect to see deaths from renal failure--which might be in that "other" category that you've got down there.

    DR. MICHALEK: Okay.



    So--and the bottom line is what you just heard: that there's a significant increase in risk of death from circulatory diseases among enlisted Ranch Hand ground crew, which is increased over the last analysis published in the American Journal of Epidemiology; and no evidence of an increased risk of death from cancer, matching reference.  And the risk of death from digestive diseases is no longer significant.

    I need to underline once again that these analyses are unadjusted for known risk factors for these diseases: for example, in heart disease we can't adjust against the 19,000 comparisons for smoking or family history of disease, like we can in the analysis of data from Scripps.

    Slide, please.


    And these are same strengths that we've seen before.



    The limitations, I just said: the lack of adjustment, and the fact that the information from death certificates is limited, as compared to what we see from the extensive Scrips physicals.

    Now, this paper was rejected within the last two months by the American Journal of Epidemiology.  I want to just go through the referee comments.



    The first reviewer was concerned that these results are already published, and why are we trying to publish this again?  Well, the results are worse; I mean, they're adverse.  They're getting worse than they were two years ago, and that's important to us.

    And the second part of that is that the protocol prescribes a series of mortality analyses coming out of the study.  In previous years we would write an Air Force Technical report summarizing mortality, and that would get a blue cover and put on our shelf.  It doesn't receive as much attention as a published article does.  It's not peer reviewed.  So we're attempting to peer review all of our work, rather than write blue-covered Air Force tech reports.  And so that was why we submit to journals.

    Small sample size, exposure period and so on--all of these critiques were on the table in 1977 when the study was conceived.  We knew we had a fixed sample size, and therefore a limited ability and limited power of Viet Nam veterans comparisons.  And here the referee is saying exactly the reason why we did the analyses I just showed you on cancer, where we attempt to break out exposure opportunity and limit the comparison cohort.

    And some of these techniques I showed you on cancer incidence could be applied here, although we haven't done so yet.

    And he's making some of the statements that you just said at the table just a few minutes ago.



    He's saying the Ranch Hand cohort is not the best group to study for health effects of dioxin or health consequences of service in Viet Nam.  And here, I think, the reviewer is revealing his own bias towards industrial cohort studies: namely the NIOSH industrial cohort studies, where we think he's been basing his experience.

    "Not representative of Viet Nam veterans," sample size is too small."

    Slide, please.


    And in here he's arguing that the industrial cohort studies have experienced exposures 10 to 20 times greater than Ranch Hand veterans.  And all of that is true.  However, Ranch Hand veterans is the only identifiable cohort of Viet Nam veterans that has demonstrably increased dioxin levels.  And it does have this order of exposure period relative to industrial cohort members who worked in the factories for up to 20 or 30 years.  All of that is true.

    Slide, please.


    DR. STOTO: I think that this reviewer says that you did the wrong study.

    DR. MICHALEK: Yes.  He's saying the veterans should be happy with the industrial cohort studies, and you shouldn't even be doing the Ranch Hand study.

    DR. STOTO: Right--and that--you know, those are their comments if this were a review group that--whether or not the study should be funded or not.  But it's irrelevant.

    DR. MICHALEK: That's right.

    DR. HARRISON: Can we, instead of going through all the reviewers' comments, can we just stop here and discuss--in terms of time, unless someone protests vociferously?  Comments?

    DR. GOUGH: Well, Joel, since we're almost to the end of the study--when was the last collection of mortality data?

    DR. MICHALEK: We will continue mortality assessment up until the last day.

    DR. GOUGH: So wouldn't it be reasonable, then--I mean, I think it's legitimate.  They're saying they're coming too fast.  But that final mortality report I think would be published almost anywhere.

    DR. MICHALEK: Probably.  Except that, in the worst scenario, the program would be ended and the village would be shut down, and all of our computers archived, and we wouldn't be around to receive comments from the referees to get the thing published.  I would be working somewhere else.

    So, in other words, there has to be an end-game worked out for this project.

    DR. HARRISON: Well, first of all--as, Mike, you and I discussed this morning--Joel, you're to be commended on doing something that was brought up to you several years ago, and that is the desirability of publishing this material in peer-reviewed publications where there's some assurance of quality because of the nature of peer-reviewed publications.

    I also thought that this was kind of like a locomotive that finally got going--


    --you know, the juggernaut kind of theory.  I mean, papers flying all over the place.

    The other way that you might be able to add interest to this paper would be if you tried to relate it to your other major finding.  If there's a lot of cardiovascular death in this study, and there's a lot of diabetes in another study, those two are actually related--

    DR. MICHALEK: Yes.

    DR. HARRISON:  --and add interest because of their relationship.  So that might actually add relevance that would help you in terms of the review route.

    DR. TREWYN: Yes, and I think you don't want to just package these the same way each time.  You can use the learnings from the other things, but there are also--as we all know in the scientific area there are multiple journals.  And so you can utilize reviewer comments--even if you can't go back to that particular journal, but you utilize the good points, ignore the dumb stuff that they said--and they always have some of those--and you try to improve the paper and submit it somewhere else.

    And I think another point here is: is this the right time, or do you wait until you've got--you see where the trends are going, and so--but I think doing a different collection, and focusing it differently using your other positive findings is one way.  And they looking towards the end of targeting, "Okay, at such-and-such a time we try to say, all right, we're going to cut off now," so we've got enough time to package this all and get it published somewhere while we're all still--

    But I think looking at that--again, I commend you on pushing these things out the door.  It helps everyone.

    DR. HARRISON: Speaking of other findings, shall we go on to the insulin sensitivity?

    DR. MICHALEK: Yes.  Then we'll  skip Reviewer 2?

    DR. HARRISON: Reviewer No. 2 always has the same comments.

Diabetes- Summarize Latest Analysis of the Insulin

Sensitivity Study

    DR. MICHALEK: All right--insulin sensitivity.

    Diabetes has been an issue in the study since 1992, when we found a significant trend in the Ranch Hand cohort, of diabetic risk against dioxin category.  That paper was published in 1996, in Epidemiology, and since then we've continued to see increased risk.  The pattern is complicated by confounding, and the results require careful analysis to display.

    We have since then made two efforts to try to understand the diabetic finding at a biological level.  One is to conduct a matched-pair insulin sensitivity analysis, using the glucose clamp method at the Little Rock VA medical center.  The other was to extract adipose tissue specimens from 313 study subjects and have those assayed by the Department of Toxicology, University of California-Davis.

    This is a series of slides summarizing preliminary results from the Little Rock study on 29 matched pairs who received the glucose-clamp measurement of insulin sensitivity.  The first officer was Dr. Philip Kern.  And I'm sorry Dr. Kern is not here--emphasizing, once again, I am a statistician.  Dr. Kern is an M.D., and he is a specialist in this area and can describe the methodology that I cannot.

    Sensitivity index was derived by the euglycemic clamp in 29 matched pairs.  We have 29 Ranch Hand veterans matched to 29 comparisons, all non-diabetic and matched in great detailed, and will be described to you in coming slides.

    We also have a quick index of insulin sensitivity derived from data collected currently at Scripps clinic, using fasting glucose and fasting insulin, and the reference for that calculation is a paper by Katz, et al., which is given on the last PowerPoint slide.

    The dioxin measurements were made by the Centers for Disease Control using the same dioxin data you've seen represented in other slides.

    Slide, please.


    Inclusion criteria was these veterans had to have participated in the 1997 physical examination at Scripps.  They had to have all been non-diabetic, and the proscription for non-diabetic--in absence of a doctor's diagnosis are these specifications for fasting glucose and two hour postprandial.

    The comparison had to have a dioxin level less than 10 parts per trillion, and the Ranch Hand had to have four repeated dioxin measurements all greater than 10.  So all of these Ranch Hand veterans that qualified for inclusion in this study are in our half-life study, and they all have demonstrably high levels made four times over a 15 year period.

    Exclusion criteria--following a protocol written by Dr. Kern, we excluded individuals who had large weight gains or losses, or acute illnesses or medications that would have affected the measurement, or any other condition that would have affected the measurement.



    The matching criteria were more stringent than the matching that's used in the global study.  They're matched on age within five years, body mass index within two units, perfectly on race, and imperfectly on reported family history of diabetes in first order relatives.

    DR. HARRISON: Joel, was that race self-reported?

    DR. MICHALEK: Self-reported.



    This is a matched-pair study, so that the first statistic that people like to see is the pair t-test.  And that is the first analysis that you will see.  Subsequent to that we're going to reveal significant adverse trends in both cohorts relative to dioxin and insulin sensitivity.  And for that we're going to use a number of different regression models.  In one we're going to combine all 58 subjects together and display a single regression line of insulin sensitivity versus dioxin, using a single intersect and a single slope.

    In another analysis we're going to realize that we've got dose response patterns going on within both cohorts: within the 29 comparisons and within the 29 Ranch Handers, and so we're going to fit separate regression lines within each of those two cohorts of 29 veterans--with this for the common slope, because the slopes are parallel.

    And, finally, in other analyses we're going to introduce a separate regression line in each cohort, because we see differing patterns sometimes in the comparisons as compared to the 29 Ranch Hand veterans.

    This is the sample reduction, which gets us from the 567 Ranch Hands and 815 comparisons down to the actual 29 that were actually used--driving home the point that this is a very stringent sampling criterion.  Not only did you have to have them matched, if a Ranch Hander volunteered to go to Little Rock and his matched comparison didn't, that Ranch Hander can't go, because we had to have 29 matched pairs.  And that's a very stringent criteria which led us to the absolute--if we actually had more funding, we could not have exceeded 29 matched pairs.  We have used up all available resources, in other words, to do this little study in Little Rock.

    Slide, please.


    The sensitivity index was measured in a fasting state.  These individuals were invited to Little Rock and arrived the night before, and had a special hospital room set up for them where they slept the night and approached the examination room in a fasting state in the morning.  And these details here describe the actual clinical procedure to do the glucose clamp.

    DR. OSEI: Joel, you've got to be very careful if you're actually say this is euglycemic clamp.  This is not euglycemic clamp.  This is a MINMOD model, the FSIGT.  So you cannot equate them.  And so you really--what did he do?  Did you guys do MINMOD model?  This one, or you did a clamp.

    DR. MICHALEK: We did a MINMOD model, and that's described in the methods section.  So I'll make sure that the words are correct, as you're describing.

    DR. OSEI: Because you've got to, you know, get rid of the euglycemic clamp, because that's not what you did.



    More detail on the measurement.  This was done over a four-hour period, with frequently sampled insulin and glucose measurements made, both arms; insulin measured by radio-amino assay at the endocrinology laboratory at the university.



    And there's the MINMOD software that was used.  There are the units.  And we also measured something called "acute insulin response to glucose," which is not shown in these talks.  I mean, I'm only going to concentrate on the sensitivity index in this talk.  Although we have this data available, we have not analyzed it yet.



    We also have measured cytokines, which are also not available yet, or haven't been statistically analyzed.  They're available, but not analyzed, so we're not going to talk about that in this presentation.



    Using the Scripps data and data from Little Rock, we computed a so-called "quick Index" of insulin sensitivity, which is based solely on fasting glucose and fasting insulin, which is published in a recent issue of the Journal of Clinical Endocrinology, first author Katz.  We were able to compute this index using data collected in Little Rock, and the data collected at Scripps.  And you're going to see that data in this talk.

    And the Scripps data, you're going to see, is cumulative as of January of this year.  So as this paper evolves and we get our final data sets from SARC and Scripps, we're going to update these analyses to include everyone that went to Scripps and satisfies the inclusion criteria.  We're restricting to non-diabetic participants.

    DR. GOUGH: Joel, I hate to interrupt, but I have a really basic question, because I don't understand.  The measurement is in levels of insulin.  You did a glucose bolt to the subject, then that sample is taken and insulin is measured in blood, over a period.

    DR. MICHALEK: Dr. Harrison is a better person to describe the actual method than I am.

    DR. HARRISON: I defer to Dr. Osei.  I know how it's done, but I'm just not going to tell you.

    DR. OSEI: The way you do it, you take at zero time, it has to be a basal sample, the first sample that he did.  Then you give IV glucose, which--what's that 3.3 [inaudible] as a bolus.  Then in 20 minutes, you give IV [inaudible], like you've got to read the insulin now.  And we have a computer derived computer model that calculates the relationship of the insulin and glucose assay you can pull down.  And that's how you derive the sensitivity index.  Okay?  And this was described by Bergman.  It's called Bergman MINMOD model.  So the relationship is actually the [inaudible] disappearance of glucose into the tissue, and that is a reflection of sensitivity.

    DR. GOUGH: If a person is diabetic what way does it go compared to a normal person?

    DR. OSEI: The normal individual will have a high level.  If you have impaired glucose you will be little, and if you're diabetic very, very low.  So the higher the number, the more sensitive you are.

    DR. GOUGH: And that's good.

    DR. OSEI: That's good.

    DR. MICHALEK: That's important, because the slides you're going to see will be precisely in that direction.  With higher dioxin levels you have reduced insulin sensitivity.

    DR. STOTO: I'd like to ask a different question about the design.  I mean, you've take out the people who have diabetes.

    DR. MICHALEK: Yes, we did.

    DR. STOTO:   And so you're looking at people--you're comparing--you're looking for insulin sensitivity in people who don't have diagnosed--or even symptomatic--

    DR. MICHALEK: That's right.

    DR. STOTO:  --diabetes.

    DR. MICHALEK: There's a reason for that.

    DR. STOTO: That's what I want to know.

    DR. MICHALEK: The reason for that--

    DR. OSEI: The best study you can think of to look at, you know, this issue.  If you have diabetes already, that per se can impact on insulin sensitivity, just the hyperglycemia on the cell function.  So that's not the right study to do.  This is actually the best study to do--to really take healthy individuals and try to look at one variable, and that is exposure to dioxin.  So if you can isolate that, you have a case to make.

    DR. STOTO: So it's that these people who have decreased sensitivity to insulin will--are at risk for getting diabetes eventually?

    DR. OSEI:   Exactly.  Actually, if you go to families who don't have diabetes like he's doing, you just have the genetic predisposition to diabetes, they will have insulin resistance--so they have lower sensitivity.  And these are individuals who are predisposed to becoming diabetic in the future.

    So having insulin resistance is bad for you; or having lower insulin sensitivity is dangerous, actually.  So I think it's critical.

    DR. CAMACHO: Moving into the tissue faster, that's better?

    DR. OSEI: Yes.  Actually, the disposal--the reason you're blood sugar goes up, because the blood sugar cannot get into your tissue.  So you build up almost as a black of transported glucose from the circulation into the cells.  And that's called insulin resistance.  So if that herbicide causes insulin resistance, they are more prone to become diabetic.

    DR. STOTO: So we're trying to understand the mechanism by which dioxin could be a cause of diabetes?

    DR. OSEI: Right.  One way you can say the mechanism--this is not at a molecular level.  This is the global physiologic level.  But if you really wanted to do.

    DR. STOTO: But it could mediate between whether or not they develop or not.  It's why--it's how does this--

    DR. OSEI: Right.  Actually, to really know whether this is--dioxin is the way to--you have a tissue you can really look at the transport of glucose in vitro, outside the body, and as the question: does dioxin impair the ability to transport glucose, which is something I think global about this as you state.

    DR. HARRISON: Now that you've got them on the ropes, Kwame, tell them about syndrome-X.


    DR. OSEI: Give me time.

    DR. HARRISON: No, I'm serious.  Because it's relevant.

    DR. OSEI: One of the issues that have come up, you know, over the years is that the insulin resistance that you have--whatever the cause is, genetic or acquired--that is associated with other cardiovascular risk, such as hyperlipidemia, lipid disorders--all these things--obesity, and trunk obesity, and more fat can develop.  And these are what everybody's looking at in terms of what we call metabolic syndrome, or syndrome X, which are risk factors for cardiovascular death.

    So these individuals who can demonstrate, that have insulin resistance earlier, that may be relevant to the mortality data you showed, in terms of atherosclerosis and death.

    DR. STOTO: So it makes the whole thing consistent.

    DR. OSEI: Exactly.  And I think the point that you made, Ron, about taking hypertension, taking the diabetes into the cardiovascular mortality is relevant, it's very critical, because you may be able to tease out what is driving the horse at this point.

    DR. STILLS:  And I think you answered my question, when you were talking about atherosclerosis, and one of my question was, you know--in terms of the Ranch Hand population--you know, people with cardiovascular disease, do they have hyperlipidemia, and all the other factors--you know, diabetes.

    And so I think that's the critical aspect, in terms of writing these papers and making an impact, is pull in all the clues together that really tell the story.

    And I think, you know--

    DR. OSEI: And I hope Joel will take advantage of this because you have, you know, well characterized sub-groups to be able to look the importance of lipids in relationship to what Bob was talking about in insulin resistance.

    So you can have lipid disorder, which would predispose them to atherosclerosis in the future.  Yes.

    DR. STILLS: I just want to ask one other question of the committee.  Is there not an example where people are exposed to like herbicides or chemicals where we know that there's a link to cardiovascular disease?  Or this would be the first study that really shows that?

    DR. HARRISON: I don't know the answer to that.

    DR. OSEI: What I know that--I don't really trust about environmental factors as a cause of atherosclerosis, by pointing--by asking what specific environmental factor, what are you talking about.  It's very difficult to actually pinpoint exactly what factor in the environment that leads to all these things.

    And I think this will be one of the best studies to link an environmental factor to this.  But the variables here, the variable that you're dealing with is the whole experience of the Ranch Hand--the military itself, or the Air Force itself.  That experience may be another variable that we need to look into.

    DR. STILLS: And I just want to underscore--the importance of this study, Joel, is at NIHS, that's one of the questions we're asking, because we're an environmental institute, because we're really working hard to find examples where you could pull--you have this story to tell.  And I think your study is one, you know, if you pull out aspects it would be a nice example.  And, you know, what type of models we use for asking that question: environmental exposure, how does it play a role in terms of cardiovascular disease.

    So, I mean, I think it's a very impressive study, and it's something that at a Congressional level and at the intergovernmental level, that these questions have been asked.  So packaged correctly, like the committee is recommending, I think, you know, you have a highly visible paper.

    DR. OSEI: And just to finish on that, you're not going to look at ARR, which is a mistake at this point.  Either you don't have the data or--but it's so critical--

    DR. MICHALEK: We have the data, if somebody wants to do it.

    DR. OSEI: Okay.  All right.  Because that is an issue.

    DR. MICHALEK: To answer your question, the reason why we restricted to non-diabetics is that that is the cohort where we found the dose response on dioxin.  It was among the non-diabetics that we saw increased insulin levels.

    And so once you look at diabetics, you see no pattern whatsoever against dioxin.

    DR. STOTO: Because it's the diabetes itself.

    DR. MICHALEK: Apparently--it's disrupted, yes.

    DR. HARRISON: Yes, as Kwame was pointing out, there's a good reason why you don't see a pattern in diabetics.

    DR. OSEI: Yes, you know, it's a mistake.  We used to do this a lot--for 20 years that's what we've been doing, looking at diabetics.  At the end of the day, it's very difficult to look at secretions and ask specific question what your study's about.  And this particular one, since you want to know why that one agent is involved in this.  You cannot study a diabetic, because at the end of the day you can't interpret your data.

    DR. HARRISON: Joel?

    DR. STOTO: That was very instructive.  Thank you, Kwame.

    DR. HARRISON: There's been such a long, long interruption here of your presentation, I wonder if we could pause for a moment and ask if there's anyone here--because public statements were scheduled for 11 o'clock.  I wonder if we could pause for just a moment and ask if there is anyone that has a public statement that they would like to make and, if so, they would be welcome to do so now.  And then we could finish with the--because we're not to the results section yet.  So you've introduced us.  We're all primed and everything.  This is a nice place to make a cut, and have the public statements on time, which will satisfy lots of people.

Call for Public Oral Presentations

    DR. HARRISON:  Public statements.  Public statements.  Public statements.

    [No response.]

    Going once.

    Be it known that public statements were scheduled for 11:00.  We've requested public statements at five minutes after 11:00.  There's no one present to make one.  If it turns out that someone shows up at an odd time, if it's possible to make provision for that we'll do so.  But, otherwise--carry on, Joel.

    DR. MICHALEK: I'll just back up a few slides.  I missed the point that I want to show you--that these measurements were made between December 199 and March 2001--at Little Rock.

    Demographic summary of the 29 matched pairs--you see a very good concordance, of course, because of the matching, on everything except dioxin.  And there you see the 45 parts per trillion in the Ranch Hand cohort, among those 29, versus 3.6 median in the 29 comparisons.  We have Ranch Handers with a lot of dioxin in their bodies are being matched to comparisons.

    Slide, please.


    And this is the corresponding contrast of demographics on the individuals that were available to us from the Scripps data as of January of this year to be used in our Quick Index.

    And we see comparability across the board on these measurements, except dioxin, of course, where the Ranch Hand's are increased.

    Next slide, please.


    And here is the first cut analysis, where we found a mean sensitivity index among the 29 Ranch Handers is decreased--in other words, in the adverse direction--relative to the 29 comparisons, but did not reach statistical significance, either in original units or in log units.  But in both analyses, the Ranch Hand experience is in the detrimental direction.



    Now, we're going to march through a series of analyses to describe to you the patterns within each cohort which are statistically significant.

    In the Model 2 analysis, we're simply taking all 58 individuals and lumping them all together, and asking: is there a relation between dioxin and insulin sensitivity, and does it reach significance?  And the answer is no.  It's in the detrimental direction--insulin sensitivity decreases with dioxin.  It's a negative slope, but did not reach significance.

    And the numbers on this slide are unadjusted for covariance.  The numbers in the footnotes are--the point being that whether we adjust or don't adjust, we find the same results.

    Over here we have what we call Model 3, where we accommodate a regression line within each cohort, but with different intercepts, and we find a common slope, which is the detrimental direction, and it does reach significance within each cohort.  But we have significantly different intercepts.  So what I have to say is that when we do the pair t-test, we find no significance.  However, when we look within each of the 29 components of the matched pairs, we see significant and detrimental results relative to dioxin.

    Slide, please.


    Parallel to that, using the Scripps data derived from very current information from right here at Scripps clinic, we see the same pattern with an overall decrease in insulin sensitivity among non-diabetics who were at the Scripps clinic.  Using the Quick Index, in all non-diabetics combined, and yet also within each cohort, within the comparisons and he Ranch Hands, we see an even stronger detrimental slope of dioxin versus the Quick Index, with differing intercepts.



    And all of that--those words and those slides were unadjusted, and the footnotes show that the results remain the same--with adjustment for covariance.

    Here is the picture--and to back out the number and p-values I just described on the 29 matched pairs.  This is the Model 2 where we lump all of the 58 individuals together and ask is there a relation between dioxin level and insulin sensitivity in log units, and we see that lack of significance on the slope.  It's in the negative direction, but it's not significant.  And yet when we look within the 29 comparisons, and within the 29 Ranch Handers, we see significant decrease in insulin sensitivity within increased dioxin within both cohorts.

    And this vertical line here is used to remind you that of the 29 pairs, we had comparisons less than 10 ppt, and 29 Ranch Handers greater than 10 ppt.  And so we have this built in dichotomy to the Little Rock study where we had comparisons with low levels, and Ranch Hand veterans with high levels.

    DR. STOTO: Do I understand this correctly when I say that this suggests that dioxin alone does not explain the difference in insulin sensitivity?  There actually is a difference between--

    DR. MICHALEK: To accommodate the cohorts, yes.  Within each cohort you see a significant detrimental association between dioxin in insulin sensitivity.

    DR. STOTO: Well, our focus is on the difference between the cohorts.

    DR. MICHALEK: There was also a difference--well, there was no difference between the cohorts--significant difference between mean insulin sensitivity and log units.  That's the paired t-test.  And yet, when you look within the cohorts, you see these significant dynamics going on within each cohort.

    DR. STOTO: And why would that be--what does that mean?

    DR. MICHALEK: I have reasons for that.  It comes in the subsequent slides.


    DR. HARRISON: Nice save.

    DR. MICHALEK: You think just like me.  That's the problem.

    Let's back up one more slide.  I want to just emphasize a point here.

    DR. HARRISON: Joel, I have a question.

    DR. MICHALEK: Yes.

    DR. HARRISON: It may be Mike's question, but I've never understood Mike's questions, so--


    If you look at the comparisons at a level of 10, the slope give an insulin--whatever the vertical axis is--

    DR. MICHALEK: The vertical is log insensitivity.

    DR. HARRISON:  --is like .5.

    DR. MICHALEK: Yes.

    DR. HARRISON: And then you go over to the Ranch Handers, and at 20, their insulin sensitivity--

    DR. MICHALEK: Is increased.

    DR. HARRISON: Right.  Is that what you're--is that--was that Mike's question?

    DR. MICHALEK: Yes.

    DR. HARRISON: Okay.  And you've got a reason for that.

    DR. MICHALEK: We think we understand that.

    DR. HARRISON: Okay.

    DR. MICHALEK: Want to hear the bottom line?  The punch line?  I can tell you now.

    DR. HARRISON: Okay.

    DR. MICHALEK: It goes like this.  Even though we have matched, to the bet of our ability, based on all those factors that we though and know are important, such as family history, we cannot match for other things.

    As you'll see in other slides--not just in this talk, but on the hypertension talk--what's happening is that comparison veterans with high dioxin, for a comparison the high dioxin is somewhere between 5 and 10 ppt. Those individuals are older and heavier than their comparisons with low dioxin levels.  What's happening is that these comparisons are getting dioxin from the same sources in the United States that we get it from: from out diet, from pollution, from air pollution, from smoke from burning trash, from Styrofoam cups--from all kinds of sources, we get dioxin.  And they're building up a body burden as they get older.  These individuals in the higher range of dioxin in the comparison group, between 5 and 10, are heavier, and therefore they are--because they're heavier, they have higher BMIs and higher body fats, they are at risk of having less insulin sensitivity than the individuals who are younger and have low body fats, because BMI is adversely related to insulin sensitivity in both cohorts.

    DR. HARRISON: But you matched for BMI.

    DR. MICHALEK: I know we did.  Our statistical models are failing us.  Our statistical models most of the time work for us.  But this is an example where our statistical modeling has reached the limit--

    DR. GOUGH: A far more trivial recommendation--which--and I'll express it trivially, because I can't--but, we have to send the comparisons--the comparisons and the Ranch Hands both have the same background exposure.  They live in the United States.  I mean, that's what Joel is talking about right now.  And it seems to me that what these data say is that a person who is likely to get diabetes, or likely to become insulin insensitive, for some reason is more likely to retain dioxin in his body than the other person.  And I don't know how to explain that.

    DR. HARRISON: But the mechanism of that should be BMI, because dioxin is retained in the fatty tissue.

    DR. STOTO: Yes--and they don't.

    DR. HARRISON: No--and it's not.  Because Joel is putting the comparison--is saying that the comparisons have higher BMIs than the Ranch Handers.

    DR. GOUGH: No, he's not.

    DR. HARRISON: That's what he's just saying.

    DR. MICHALEK: There are many factors going on at the same time.

    DR. GOUGH: Let me get this straight, Joel.  Didn't you say that the comparisons who have the decreased sensitivities--

    DR. MICHALEK: Right.

    DR. GOUGH: --are heavier and older than the comparisons, who don't have it.  Isn't that also true of the Ranch Hands?

    DR. OSEI: But the BMI is only one index.  It doesn't give you how much fat you have in your body.  So as you get older, and you accumulate more fat, dioxin accumulates in fat. You can have the same BMI, but your total--

    DR. HARRISON: Higher percent--

    DR. OSEI:  --you have higher dioxin, because you're accumulating that in the fat.

    But the point that is missing here is the fat distribution.  The BMI just gives you the overall.  If you have more fat in your belly, you have trunk obesity, you are more likely going to have more insulin resistence.

    But do the numbers show--we can do this in so many ways, you can measure the total body fat content, which would give you some of the answers that you're talking about.

    DR. MICHALEK: We have measured all the BMI.

    DR. OSEI: Yes, one you measure the BMI, or you can actually, in a real substance kind of way, do the CAT scan of their belly, measure how much fat is in the belly.  So if you're on the match, which you did, the difference is coming from what you're talking about--that is this more distribution of fat content in the body.  And you need to have another way of looking at that.  And it can be done in so many ways.

    DR. GOUGH: But to return to my comment--but the same argument you make to explain the decreased sensitivity in comparisons could be made for the Ranch Hands.

    DR. MICHALEK: Yes, sir.  I'm not finished with my argument.

    DR. GOUGH: So it's not an explanation.

    DR. MICHALEK: I'm not finished yet.

    DR. GOUGH: Sorry.

    DR. MICHALEK: There's more to show.

    DR. HARRISON: Let's wait until Joel is fully committed before we jump on him again.

    We're having fun, though, Joel.

    DR. MICHALEK: All right.

    Next slide.


    In this, we are looking simply, just to slow you how the log of the SI relates to the QUICKI index.  We have strong correlations between the SI, as measured by Dr. Kern, and the QUICKI measurement derived from the Katz formula.  Here's the--among the--there's 58 individuals in the matched pair study.  Among those 58, using measurements made at Little Rock, and correlation between the log SI and the QUICKI's is very high.  It's .71.

    If you ask what is the relation between low SI measured at Little Rock and the QUICKI measured among those who went to Scripps clinic, of the 58, 37 were non-diabetic and available and had measurements from Scripps clinic at the time this was written, the correlation was .51.  Remember that the measurements at Little Rock were made between December 1999 and March 2001, and the measurements at Scripps were made more recently, and so there's a time factor involved here.

    DR. OSEI: And what you did was the same insulin assay--

    DR. MICHALEK: Yes.

    DR. OSEI:  --because if it was different you're going to have different results.

    DR. MICHALEK: Yes.  Here are the measurements of fasting glucose and insulin that were made at Scripps, and over here they were made on site at Little Rock.

    DR. OSEI: That's a problem.

    DR. MICHALEK: That's right.

    DR. OSEI: Okay.

    DR. MICHALEK: That would explain--

    DR. OSEI: Okay.  So it was a different assay--

    DR. MICHALEK: Correct--different time points, yes.  But they're in the same direction and highly significant.

    And here is the BMI when they were in Viet Nam is adversely related to the insulin sensitivity, as you would expect, in heavier individuals.  But it doesn't reach significance.

    Among the 37 who showed up at Scripps, using measurements made here at Scripps clinic measuring BMI, there is a negative and adverse relation between BMI and the log of the DR. STILLS:   measured at Little Rock--which is all in the direction of adversity.  Increased body mass index--

    DR. OSEI: When was the BMI--when they went in?

    DR. MICHALEK: BMI measured in 2-R was calculated from the height and weight that they had when they were in Viet Nam during the war.

    DR. OSEI: And they were much younger.

    DR. MICHALEK: Yes, they were.

    DR. OSEI: Okay.  So the variable--

    DR. MICHALEK: That's measured--yeah--

    DR. OSEI: --as you get older you put on more weight--

    DR. MICHALEK: Right.

    DR. OSEI: --you are more sedentary.  They're back home.  So what you're looking out, maybe you're fleshing out not only BMI but the lifestyle--

    DR. MICHALEK: Yes.

    DR. OSEI:  --for coming back.

    DR. MICHALEK: All kinds of intervening variables.

    That BMI was made in 1967.  This BMI was made in the year 2002.

    DR. HARRISON: Another way of saying that is that the variation between the low BMI--the distance between the low BMI and the high BMI in Viet Nam would have been shorter than the distance between the low BMI and high BMI now--

    DR. MICHALEK: Yes.

    DR. HARRISON: --because you had to meet certain physical fitness standards.

    DR. MICHALEK: Yes.  All of that's true.

    DR. HARRISON: And accordingly, you'd have more difficulty making a correlation.

    DR. MICHALEK: And in all of our statistical models we accommodate not just the BMI in Viet Nam but the change in BMI from Viet Nam to the present, as well as the BMI in the present.  The change itself is important--and detrimental.  People who gained weight are at increased risk of insulin resistance and diabetes.

    DR. HARRISON: Kwame, would you have been happy with a waist measurement as an assessment of visceral fat?

    DR. OSEI: Yes.  You know, if you don't have the money, you can do it in a very cheap way by just looking at--taking tape and measure the waist circumference.  And that's very easy to do.  You can do the waist hip ratio or so, and try to help you at.  But if you have the money, you have to do a CAT scan or MRI.

    DR. MICHALEK: We did the waist.  Waist measurements were done as a current physical, but not an MRI.

    DR. OSEI: Did you have a waist circumference?

    DR. MICHALEK: Yes there were.  They were done.  It's not been used yet, but could be used.

    DR. OSEI: It will help you out, you know, to really resolve some of the issues.

    DR. MICHALEK: Okay.

    And here is the corresponding pattern, as I showed you in previous slides--there's a parallel pattern of the QUICKI using Scripps data that we saw in the log insulin sensitivity data measurement at Scripps.  We have a weak detrimental trend overall, but stronger trends within both groups in the detrimental direction of dioxin versus the insulin sensitivity measured using the QUICKI index.

    Here, again is the definition of the QUICKI index.  In this so-called statistical Model 4, we have simply fit a separate regression line within each cohort.  And here I isolated comparisons with less than 10 ppt and Ranch Handers with greater than 10 ppt who went to Scripps and were non-diabetic and satisfied all the conditions of this particular study.  I'm trying to imitate, in other words, the Little Rock experience, using the Scripps experience.  In other words, these are not all comparisons, only those less than 10.  These are not all Ranch Handers, only those Ranch Handers greater than 10.  So I, the best I can, imitate Little Rock using Scripps data.  And we see a parallel pattern.



    DR. STOTO: So the fact that there's a greater slope on the left in the comparison--

    DR. MICHALEK: Yes.

    DR. STOTO:  --could that be because they're more sensitive--

    DR. MICHALEK: To dioxin.

    DR. STOTO:  --dioxin being retained because of their fat.

    DR. MICHALEK: Possibly.

    DR. STOTO: The guy's in the Ranch Hand group are really even--direct affect.  That's the hypothesis--

    DR. MICHALEK: That is a hypothesis, yes.

    Another hypothesis that we've entertained is that these Ranch Handers down here who are at the low end of the Ranch Hand scale are lucky guys.  They were able to eliminate dioxin more rapidly than other Ranch Handers, therefore they're healthier.  They could eliminate not just dioxins, but PCBs and feurans and other organic pollutants from their bodies.  And so that would maybe accommodate some of the differences we see in intercepts.

    DR. OSEI: But you are using serum levels, and you may have an opportunity to look at tissue levels, which would be actually the key.  So the comparisons on the left side, you may have a lower level--serum level--but it's all in the fat.  If it's already there you're going to see the same result.  So the opportunity--you have an opportunity to look at the tissue levels, and that may actually tell you the difference.

    DR. MICHALEK: Yes.  That's been done already. That was done by CDC.

    DR. OSEI: The tissue levels?

    DR. MICHALEK: Yes.

    DR. HARRISON: I'd argue that the dioxin that's in the fat is biologically inactive.

    DR. OSEI: I don't know that.

    DR. HARRISON: That the dioxin that's in the circulation, that's able to go to other tissues and interact with dioxin receptor would be the biologically relevant measurement.

    DR. OSEI: But maybe--

    DR. HARRISON: That the fat may retain the dioxin longer, and hence give you a longer dioxin effect, but it's still mediated through plasma dioxin levels.

    DR. OSEI: Yes, but, you know, separation of a hormone would have no meaning unless it acts on the tissue at some receptor enzymatic pathway.

    DR. HARRISON: Exactly.

    DR. OSEI: So that if you are accumulating that longer, you're more likely going to have the tissue fat at the muscle or the fat longer for it to have more biologically negative effect than just the circulation.

    So it may not be a direct relationship between the serum and the tissue.  And I think you maybe hit on something interesting if you have the numbers.

    DR. HARRISON: Well, but unless you're arguing that fat is a dioxin target tissue--

    DR. OSEI: Or muscle.

    DR. HARRISON: You take another hormone, do an injection with cortisone, and what you'll find very shortly is that a certain amount of cortisone may be retained in fat, but that's not--the cortisol effects on the liver and all the rest are mediated through the cortisol that's circulating and able to enter muscle and liver cells.

    DR. STOTO: Have you been showing a strong correlation between dioxin and fat--

    DR. HARRISON: What?

    DR. STOTO: A strong correlation between dioxin and fat--in the serum.  So that it could be a reservoir of the stuff comes and goes.

    DR. HARRISON: Well, that's my point though, is that Kwame's arguing that a low serum level is not--may not be the right thing to look at because it may all be in fat.  And I'm saying that the only reason that fat matters--yes, I'm saying the only reason that fat matters is because it provides dioxin to the circulation to go to other tissues and cause a dioxin effect.

    DR. OSEI: It may be that the same scenario, the dioxin also goes to the muscle.  And this [off mike] muscle mass.

    So, if it's a poison used for whatever reason used in Viet Nam--dioxin they could have a low insulin sensitivity, even though your levels may not be very dramatic.

    DR. MICHALEK: Well, relevant to your point, the CDC did the study of 50 individuals and measured dioxin in serum and in adipose tissue, and their best estimate on the partitioning coefficient was 1.0.  In other words, there is no difference between the method in adipose as opposed to the method and the basis in serum in 50 individuals from Missouri.  And that was published in 1989 or 1990, I believe.

    DR. STOTO: The thing that strikes me here is that you've left off the comparisons with dioxin greater than 10.

    DR. MICHALEK: Which number only about 10 people.

    DR. STOTO: Oh, is that right?

    DR. MICHALEK: Yes.

    DR. STOTO: Okay, and then similarly, the Ranch Handers with less than 10.

    DR. MICHALEK: I lopped off Ranch Handers with less than 10.  However, in the footnotes, if you include all Ranch Handers, including those less than 10, you see the same pattern.  You see a negative and detrimental slope.  I only did this to try to imitate Little Rock.



    Here I'm looking at correlations between the QUICKI and these covariates at the Scripps examination.  And you see negative and detrimental associations between insulin sensitivity, various measures of BMI and age--and BMI but not age.  Actually the correlation is positive with age, but negative with BMI.

    These are the non-diabetes who were extracted from the Scripps data to try to imitate Little Rock.



    However, we see a different pattern when we relate dioxin to BMI, using that same dichotomy of comparisons less than 10, Ranch Hands greater than 10.  We see a positive and significant slope among comparisons--there it is right there.  But in the Ranch Hand cohort we see nothing.  We see a relatively attenuated slope.  However these are Ranch Handers greater than 10, but when we include all the Ranch Handers, then the slope becomes positive.

    So we see this as kind of an interference of the exposures in Viet Nam that are sort of disrupting the natural relation between dioxin and BMI that we see in the comparison group is being disrupted by what happened to the Ranch Hand veterans in Viet Nam.

    The Ranch Hand veterans in Viet Nam experienced a slug of dioxin over a one-year period.  The comparisons received their body burden over many years in the United States and during their tours in Southeast Asia.  So there's a different exposure experience in the two cohorts.



    Well, there are many caveats associated with the dioxin measurement, as you've already pointed out.  The dioxin half-life is associated with BMI.  Heavier individuals have a longer half-life.  Heavier individuals have higher dioxin levels for that reason.  They retain it longer than thin individuals do.  There's perhaps--and we suspect very strongly--that there are differences between individuals and how they eliminate dioxin.  And so to avoid and to address those caveats, we re-analyzed using another measure which is free of many of those caveats.

    We interviewed all enlisted Ranch Hand veterans by questionnaire in 1989, after having received the first dioxin levels from Scripps, we sent them a questionnaire asking them what they did on the job in Viet Nam.  Did you get in the tanks?  Did you ever use herbicide as a hand cleaner?  Did you get it on your skin, did you get it on your clothing?

    We asked them all of this before we had told them their dioxin levels, so they were blinded to their dioxin levels when they responded to our questionnaire.  And we simply counted up the number of days of reported skin exposure, and this published later showed a significant and detrimental association between skin exposure on the questionnaire and subsequent dioxin measurements made by CDC.  This was the first direct validation of the dioxin measurement in a Viet Nam veteran study.  That was published in the Journal of Exposure Analysis and Environmental Epidemiology

    So we have this exposure index, and we're now going to apply the exposure index, based on this so-called enlisted questionnaire, to ask whether that index is related to any of the things we see in this study.  And we asked, in particular, does dioxin relate to the index and body mass index, for separate intercepts for enlisted flyers and enlisted ground crew.  And we're going to propose an argument that may explain some of the patterns we see in these data.



    DR. HARRISON: Joel, the questioner also did not know the dioxin level, right?

    DR. MICHALEK: The questioner was me.  I wrote the questionnaire.  And, of course, I wasn't study individual dioxin levels as I wrote the questionnaire to distribute to the veterans.

    DR. HARRISON: You said the veteran didn't know what his dioxin level was.  Did the person questioning the veteran--

    DR. MICHALEK: These were not in-person questionnaires.  They were mailed out to the individuals.

    DR. HARRISON: Okay. |Okay.

    DR. MICHALEK: What we've got here, based on the previous quadratic model that you just saw shown in the previous slide, we're able to--all of the coefficients of which are statistically significant--we were able to predict a curve of constant dioxin as a function of body mass index and skin exposure as derived from that mail-out questionnaire.

    The point here is an individual--this is a curve of constant dioxin.  Everyone on this curve has 28 ppt.  A person could have 28 ppt because he had a lot of skin exposure but a low BMI, or he can have 28 ppt because he had very little exposure and a high BMI, which is consistent with the idea that individuals can absorb more dioxin if they have more body fat--even though they had less skin exposure.  And individual with high skin exposure can have an elevated ppt for dioxin, even though they have reduced body mass index.



    That slide, together with the patterns you saw in previous slides, and together with known differences of comparisons in Ranch Hands by dioxin level on body mass index may explain why we have the parallel lines but different intercepts for comparisons with values near the high end of their distribution have more insulin resistance than the Ranch Hand veterans with low dioxin levels, because the Ranch Handers with low dioxin levels consistent with the slide I just showed you are going to have lower BMI than the comparisons with high dioxin levels.  And BMI relates adversely to insulin sensitivity.

    DR. STOTO:  That's comparing models 2 and 3.

    DR. MICHALEK: Yes.

    DR. STOTO: Was 4 demonstrably better than 3?

    DR. MICHALEK: Only with regard to BMI, where we had different slopes; when the thing became cockeyed.  But the other--against the sensitivity we have parallel slopes.  So that was why--the only reason we introduced model 4 was to accommodate the different patterns on BMI versus dioxin, as sort of a second- or third-order effect.

    We don't have a manuscript ready for use today at this meeting, but we will have one to be distributed to the committee very soon.  We gave you our manuscript on cancer, and this manuscript isn't ready yet for you to see, but it's very close.


    DR. GOUGH: Joel, could you go--I'm not--this is a question about my understanding.

    Why does this plot--the plot of the 28 ppt--the ones--could you--

    DR. MICHALEK: The point here is that there's a concept going on here that this plot is approximating.  And that is there's two ways you could have a high dioxin.  You could have high dioxin because you had a lot of exposure but low BMI.  Or you could have high dioxin because you had a relatively low exposure but a high BMI.  That's what the plot is telling us.

    DR. GOUGH: Okay.

    DR. MICHALEK: And that would explain, perhaps, the difference in the intercepts on those two plots.

    DR. GOUGH: Okay.  What was the correlation between the reported skin exposures and the measured dioxin levels.

    DR. MICHALEK: That reached statistical significance, and it's published--

    DR. GOUGH:  It was statistically significant?  Okay.

    DR. STOTO: Why is it that the curve has that shape?  I mean, I would guess that a straight line would fit equally as well.

    DR. MICHALEK: Well, because--back up a slide.

    DR. HARRISON: I would, too.

    DR. MICHALEK: We had a significant quadratic contribution to the model here.\

    DR. STOTO: Oh, okay.  So you tested it, and--

    DR. MICHALEK: Yes.  It's more than linear--perhaps hyperbolic.

    DR. OSEI: [Off mike.]

    DR. MICHALEK: Let's back up a slide.  This is on BMI.  BMI is already normally distributed. On the dependant variable--sorry, I did not put that in the slide, whether that's log dioxin or dioxin.

    DR. HARRISON: You said BMI is normally distributed?

    DR. MICHALEK: The histogram of BMI is relatively bell-shaped relative to dioxin, for example, which is highly skewed and has a logmoral distribution--yes.

    DR. HARRISON: But it doesn't go to zero.

    DR. MICHALEK: No, it's bell-shaped, though, and the minimum is about 17, or maybe 12.  And the maximum's about 50.

    DR. HARRISON: You just made me think of something.  You said that the CDC did a study of tissue and serum--

    DR. MICHALEK: Yes, I did.

    DR. HARRISON: And found that the partition coefficient was 1.

    DR. MICHALEK: 1.0--yes.

    DR. HARRISON: Which means that you would expect, if you did a partition coefficient using liquid--you know, a typical chemistry experiment, that dioxin is as soluble in water as it is in lipid.

    DR. MICHALEK: In this case it's as soluble in lipid as it is in adipose.

    DR. HARRISON: And the lipid fraction of the blood--that's what it's measured in.

    DR. MICHALEK: That's right.

DR. HARRISON: Lipid fraction of the blood.

    DR. MICHALEK: Fraction of the blood, as opposed to adipose itself.  They measured it in adipose tissue and they measured it in the lipid blood.

    DR. HARRISON: What I'm trying to figure out is if the partition was 1, and dioxin is excreted by the kidney or metabolized by liver--more likely--why it has such a prolonged half life?  Because it sounds like it doesn't have any trouble getting out of fat.  And once it's in the circulation, then why isn't it metabolized.

    It doesn't say it's in fat--it's in the fat fraction.  It's in the fat fraction the same way as cholesterol, the same way that steroid hormones are in the fat fraction.  It doesn't say that it's bound to fat any more than cortisol is bound to fat when it's in the circulation.

    DR. MICHALEK: I think you'll find the explanation in the first National Academy book.

    DR. HARRISON: First what?

    DR. MICHALEK: The first NAS book explains in great detail--I think--the concept you're trying to describe.

    DR. HARRISON: Okay.

    DR. MICHALEK: Why the half-life is so long.

    DR. HARRISON: I didn't mean to prolong this discussion.

    DR. OSEI: The problem is it's accumulated in the fat, just like vitamin D, so that you have to leak it out over a long-term period.  So--it's lipophilic, therefore it doesn't come out.  So it just leaks out over a long-term period.


    Slide.  Slide.  Slide.


    Okay.  So as you'll see--it's not as graphically displayed here as I'd like--but in the hypertension talk you'll see that comparisons with high dioxin levels are older and heavier than the Ranch Handers with low dioxin levels.  And that could explain the difference in the intercept.  I guess what we'd like to see is a single line coming down in the adverse direction, where the comparisons are here and the Ranch Handers are here, and you have this nice clean dose response.  We don't have that.  We have this.  And there are many intervening reasons why that could be the case.  And we've tried to touch on some of those here today.

    One would be differences in BMI and age and maybe the rate at which people eliminate dioxins from their body.

    Slide, please.


    DR. OSEI: Or you may have a--it may be a, you know, threshold effect that at some point, some dose, it doesn't matter.  What you need is just a small amount of dioxin in your body to do what it's supposed to do on insulin sensitivity.  If you go beyond that, you're flat because you already maxed out on what dioxin can do.  So you will have to transform your curves again to see where the threshold--what is the critical dioxin level that gives you the maximum effect of the EP-50 that can help you separate that.

    DR. MICHALEK: I made a note of your remarks, and we'll attempt to do that.

    And the concepts I just described are described by the plots we've just shown.



    So, finally, we have a relation between BMI--not finally--we also have a relation to BMI and dioxin that may have been disrupted by exposures in Viet Nam, which is a statement I already made, which is supported by the plot.



    So, in conclusion, we have no overall significant mean difference on the 29 Ranch Hand versus the comparisons, however we see adverse trends within both cohorts, supported by data from Scripps, that suggests that dioxin is adversely related to insulin sensitivity in both cohorts among non-diabetics.



    And there is the Katz reference that will give you the QUICK Index, published in 2000.

    Thank you very much.

    DR. STOTO: So, let's see--I think this may be consistent with what Kwame was saying--is that you had a negative slope in both groups.

    DR. MICHALEK: Yes.

    DR. STOTO: In the Ranch Hand group the slope is primarily due to dioxin affecting the insulin sensitivity.

    DR. MICHALEK: Yes.

    DR. STOTO: In the other group, you had that--you may not have that going on because you're below the threshold, but you have the opposite causal relationship going on, in that the--


    DR. STOTO:  --that the BMI is affecting how much dioxin is measured.

    DR. MICHALEK: It could be that there's a reverse causation going on.

    DR. STOTO: In the lower--in the comparison group.

    DR. MICHALEK: Yes.  The hypothesis there is that people who have dioxins in their bodies have triggered a mechanism to put on fat to thin it out.  For example, if I gave you a hydrophilic poison, you tend to end up drinking a lot of water to try to thin it out, and urinate it out.  Individuals with a chronic body burden--this is the hypothesis.

    DR. STOTO: But that's not what I'm--

    DR. MICHALEK: But that is a hypothesis that's been proposed by referees--tat individuals who have a chronic body burden of dioxin tend to put on weight because their body's trying to compensate for it, it's trying to thin it out.

    DR. STOTO: Well, I don't think you have to go that far.

    DR. MICHALEK: Okay.

    DR. STOTO: I'm not saying it's not true.  But it's just that the--we don't know how much dioxin people were exposed to when they were in Viet Nam.

    DR. MICHALEK: Who?  The comparisons or the Ranch Hands?

    DR. STOTO: Either one.

    DR. MICHALEK: Either one--we don't.  That's right.

    DR. STOTO: We only know that they were exposed and how much we measured 10 years ago.

    DR. MICHALEK: That's right--'87.

    DR. STOTO: And if, in fact, the fat guys would retain more of it--

    DR. MICHALEK: Yes.

    DR. STOTO:  --so we expect to see a relationship between BMIs and dioxin, just because of the fact that the more fat you are the more you retain.  And also the heavier guys--do they have a higher insulin--

    DR. MICHALEK: Yes.

    DR. STOTO: Lower insulin sensitivity.

    DR. MICHALEK: More insulin resistance.

    DR. STOTO: Right.  So in the left half of the graph you had two things going on.

    DR. MICHALEK: That's right.

    DR. STOTO: In the right half of the graph you only have one thing going on.

    DR. MICHALEK: Primarily dioxin.

    DR. STOTO: Right.

    DR. MICHALEK: Those are higher levels.

    DR. STOTO: And therefore--

    DR. MICHALEK: Somehow, after adjusting for BMI, we still have those parallel slopes.

    DR. STOTO: But just adjusting for them in a linear way isn't enough.

    DR. MICHALEK: Right.

    DR. STOTO: And this is the same--

    DR. MICHALEK: Perhaps some other model--

    DR. STOTO: And this is the same as a check markers--people often say about check-

    DR. MICHALEK: There might be some other statistical model we haven't considered.  And if you could help me with that, I'd love it.  I've used all available modeling that we could think of; all available covariates.

    DR. CAMACHO: Isn't what he said that it may be that the dioxin, all the evil effects of the dioxin only requires so much.  If you get more dioxin beyond that, you're not going to get that much more of an effect.  The damage has already been done.

    DR. MICHALEK: That's called the saturation effect, and that's--

    DR. OSEI:  Saturation effect--so you need to do--if you can look at what that saturation point is, because that will help you to go back and maybe for insulin sensitivity you may get this to do the work.

    DR. MICHALEK: To do an ED-50 you need a binary response, dead or alive, all or none.  And here we have a continuous insulin response among non-diabetics.  All of these people are healthy

    DR. HARRISON: Those responses max out, Joel.

    DR. MICHALEK: That's true.  They do.

    DR. HARRISON: In fact, there are two assumptions that may b e a problem here.  The first assumption is that there's no lower limit to dioxin action.

    DR. MICHALEK: There might be.

    DR. HARRISON: And, in effect, if it's a receptor mediated mechanism, the way it seems to be--there's a dioxin receptor and all this sort of business--then there is a concentration of dioxin under which the organism doesn't see anything.  There's no effect.

    And the same thing that Kwame is saying is true at the upper end; and that is that there is a concentration of dioxin that is 10 times the disassociation constant for dioxin receptor interactions, where you're at the end of this asyntopic curve, and there's no more measurable effect--in fact, there's no more measurable effect way before then for a biological system like this.

    So it may be that there's two problems.  One is the assumption of action all the way down to the very lowest dioxin levels.  That's not true.

    DR. MICHALEK: At the highest end, too.

    DR. HARRISON: And the other is that it just keeps going on and on.  That's not true either.

    DR. MICHALEK: Right.  We agree.  And some of those words are already in the discussion of the paper.

    DR. HARRISON: Okay.  So rectilinear hyperbola.

    DR. OSEI: The dioxin that we talking about with insulin sensitivity, insulin sensitivity is also a function of--okay, you can't just look at the SI alone.  And you say--trying to isolate SI.  How your body responds to insulin sensitivity on organism function because of a high insulin sensitivity log.  So you have to really look at that combination.  And what we have done, just to look at the insulin sensitivity--for you take the insulin--by pancreatic function and sensitivity, just like we did in the QUICKI, and do a model of that, which would really help you to define, is this person really insulin sensitive, or is not insulin sensitive.  What we need is insulin levels in this.  The data you showed us is just the SI.  We need an insulin level--the true insulin level in order to make any decision about SI.

    DR. HARRISON: Okay.  Any other comments?


    DR. STOTO: Regarding these models, 1, 2, 3 and 4--there's another alternative, one that it would make sense to think about.  And that is to fit two straight lines that are constrained to meet--

    DR. MICHALEK: Intercept.

    DR. STOTO:  --at--not the same intercept, but at the dioxin level at 10--

    DR. HARRISON: Constrain the continuum from low to high?

    DR. STOTO: Yes.  You could have different slopes, but they'd be constrained to be continuous.

Or quadratic, or something like that.

    DR. MICHALEK: But non-linear.

    DR. STOTO: Yes.

    DR. MICHALEK: Sure.  We could try that.

    DR. STOTO: It seems like that's a little bit easier to think about conceptually than having these two lines.

    DR. MICHALEK: Okay.

    DR. HARRISON: Okay.  Anything else?

    Thank you, sir.

    DR. MICHALEK: Thank you.

    DR. HARRISON: A couple of things.  How long is it going to take us from the time we walk out the front door to the time we walk into wherever we're supposed to be at Scripps--how long is that going to take?

    LT. COL. ROBINSON: Five, 10 minutes.  Very close.

    DR. HARRISON: So if we allowed 15 minutes, that would be plenty of time?


    DR. HARRISON: Okay.  So, we're supposed to be there at one o'clock for the closed session.  So if we could present ourselves outside at 12:45, we should be in good shape.

    LT. COL. ROBINSON: 12:45--[off mike.]

    DR. HARRISON: What?

    LT. COL. ROBINSON: If you tell the participants if they're not there on time we'll start calling rooms.

    DR. HARRISON: So we're at 11:40 now, so we've got almost an hour to do whatever we do and call it lunch.

    The other comment that I wanted to make--that last one before we leave--is that both Len and the Air Force agree that the months for our meetings, according to the schedules of chapters, should be April, August and October of 2004.  And my suggestion to Len and the Air Force is going to be that we actually put some days on there very quickly so that we can try and capture as many members of the committee as possible.

    April, August and October.  That will take us through 19 chapters, plus future directions and the draft final report.  That will leave the final report and journal articles, which will probably require yet another gathering of the eagles.

    DR. STOTO: August is a very bad month for many people, and for me in particular.

    DR. GOUGH: Nothing wrong with vacation.

    DR. STOTO: Well, I teach every day in August--the first three weeks of August.

    DR. CAMACHO: Has this committee judged--also worth considering, not just the report, but the closing down of the study in any way?

    DR. STOTO: We don't really want to go there right now.

    DR. HARRISON: But that's been a topic of discussion, and that's a topic of one of the presentations yet to come.  Sure. "Update on Study Shut Down and Transfer of Data"--that's what that really implies, Paul, is more--I mean.

    So August is already meeting with protest.  So--

    LT. COL. ROBINSON: End of July?

    DR. HARRISON: Well, I think the other thing that's true for many academic people is that July and August are just not good months period.  Those are the months when everybody's kind of buck naked and running free.

    So we may be better off looking into early September rather than August.

    DR. TREWYN: Not everybody, though--except for Rochester.  We have--they do those things.

    DR. HARRISON: Okay.  Anything else before we break?

    DR. SCHECHTMAN: A proposal was made for the members who will be involved in the closed session not to leave the hotel.  I mean, you can do your walk.  But we want to be able to gather promptly, as was pointed out, at 12:45.  Julie has a van, Ron has a van.  So we could probably accommodate the members of the committee, in terms of transportation.

    So we're just asking you not to wander off to the malls and all that, because we really have to--12:45.

    LT. COL. ROBINSON: Front of the hotel.

    DR. SCHECHTMAN: In front of the hotel--right.

    DR. HARRISON: 12:45 we load up.

    DR. SCHECHTMAN: We'll be returning to this meeting room by 3:00 p.m., using the same means of transportation so that we can continue with Joel's presentation beginning at 3:15, and further committee discussion.  Thank you.

    DR. HARRISON: Okay.

    [Luncheon recess.]

A F T E R N O O N  P R O C E E D I N G S

    DR. HARRISON: Joel now has a five minute presentation.

Shall we?

Hypertension--Summarize the Latest Analysis,

Including the Skin Exposure Index Results

    DR. MICHALEK: All right.  This is a summary of a paper on hypertension and dioxin that we have been working on for a number of years.  It was recently rejected by the American Journal of Epidemiology.  It as part of an invited talk at Canada's Center for Inland Waters earlier this year in Toronto, and it was presented, an earlier version, in Barcelona, Spain, in 2002.  It is an important paper because it illustrates the complexities of the study.



    The end point is diagnosed history of hypertension after service in Viet Nam.  We did not analyze blood pressures, the reason being that once an individual is diagnosed with hypertension he's generally on anti-hypertensive medication, and so that the blood pressures are artificially maintained in the normal range, making blood pressures themselves not very informative.  It is the diagnosis and the date of onset that is most important.



    So the methods are, as before, repeated physicals and questionnaires and record review.  We have 100 percent ascertainment, and more than 100 percent quality control.  And the study includes--this paper includes any veteran who's compliant to at least one physical exam--a total sample size of 2,682.



    We're using a Cox model, a proportional hazards model.  Because we found strong trends in both groups--as we saw in the earlier study on insulin--we've done both between and within-group analyses using a variety of models: dioxin quintiles, dioxin on a continuous scale.

    The dioxin quintile analyses are there to imitate the Plestagianes study of an industrial cohort in Germany, co-authored with James Dwyer, who is the first author on this paper.  We were trying to imitate the German study because the Germans found a significant dose response in dioxin quintiles.

    Slide, please.


    We have a long list of covariates, which

is unique to this study: drinking history, and smoking, and BMI and age and military occupation.

    Slide, please.


    The cholesterol--the lipids that were measured by CDC, and the specimen that was used to measure the dioxin are in the model, along with family history of hypertension and stroke.



    Here's the study reduction.  From the 2,682 down to the 2,417 that were actually included in the analyses by throwing out individuals with missing dioxins or hypertension prior to service in Viet Nam or Southeast Asia.



    Groups are comparable, as before, with respect to these covariates, except for dioxin.



    Actually, across the board they're fairly comparable.



    Overall--and this is the puzzle of this data--overall, there is no difference on prevalence of hypertension between groups.  About 41 percent of the Ranch Hand group and 40 percent--sorry, 40 percent of the Ranch Handers and 41 percent of comparisons have a documented history of hypertension; relative risk 1.0.



    However--same however as in the insulin study--if you break out the comparison by quintile, within the comparison group, and Ranch Hand by quintile of dioxin within their cohort--there are the sample sizes--



    --and consider the prevalence of hypertension by quintile of dioxin in the comparison group and within the Ranch Hand group, you see trends of increased risk of hypertension--these are unadjusted, of course--hypertension with increased dioxin level.



    And here's a plot showing the trends within both groups.  Everything on that plot is significant.  The trends are significant, the within-quintile contrasts are significant--

    Slide, please.


    --and here are the quintile contrasts, using the first quintile in the comparison group as the referent, and contrasting all the other quintiles with that first quintile, fully adjusted for all those covariates you saw on the earlier slides, we find significant increases in risk in the higher quintiles in both cohorts relative to the first quintile in the comparison group.

    Slide, please.


    If you combine the entire cohort together--all 2,400, and then ask whether there is a relation between log dioxin on a continuous scale, and diagnoses of hypertension, you find a significant dose response, with about 8 percent increase in risk for every doubling of the dioxin concentration in the lipid in the serum

    DR. STOTO: Is there a dummy variable in there for Ranch Hands versus comparison?

    DR. MICHALEK: Yes.  And there's no interaction.



    And here it is represented for the entire 2,400, broken out now into deciles of dioxin, with the hypertension prevalence on the vertical.  And we've got a log base 2 transformation of dioxin so that a unit change on the horizontal represents a doubling of the dioxin level.  And there's a significant trend there, in value 001.



    Now, the idea here was to further examine this to looking for inconsistencies, in other words, and to further examine those patterns by the skin-exposure index you saw in the previous talk.  And what we're headed for here is--we're headed for a final bottom line that there is going to be a trend in both groups that we can't make disappear.  And this was the paper we submitted to the Journal and the journal rejected.  And there are many reasons for the rejection which I'll show you in the referee comments.



    So what we did now was take the entire cohort, and rather do it in quintiles within each cohort, we took the entire cohort and broke it out into quintiles, and then compared comparisons in the first quintile with the Ranch Handers in the first quintile.

    Now these quintiles are in the entire spread of dioxins over all 2,400 people.

    Back up one slide, please.



    So in the first quintile we have a range of dioxins from .3 to 2.9 ppt, and in that first quintile there's 431 comparisons and 53 Ranch Hand.  And as you see in moving up to, let's say, the fourth quintile, between 6.3 and 13.1 ppt, it's about equally balanced--more Ranch Handers than comparisons.  And finally, from 13 to 600, of course, that that's where most of the Ranch Hand--or many Ranch Handers are, because they have higher levels.

    So what we're doing now is we're just taking the entire distribution and breaking them up into quintiles; again, trying to imitate the Germans.  And we're considering the prevalence of hypertension within each cohort by quintile of dioxin.  And we're comparing across quintiles.

    So we see here in the fourth quintile, for example, we have a higher prevalence of hypertension in the comparison group than in the dioxin group.

    Slide, please.


    And here's a graphic representation of that, where we have the comparisons dominating the Ranch Hands by quintile of dioxin in the combined cohort--a pattern that was certainly not contemplated in 1977 when we wrote the protocol.


    DR. STOTO: This is quintiles for the whole group.

    DR. MICHALEK: Yes.

    So the first quintile here--take the lowest 20 percent in the entire cohort.

    DR. STOTO: Well, could this be that, like for the other studies, just swing BMI into the equation--doesn't control?

    DR. MICHALEK: No, the BMI adjustment, using standard statistical methods doesn't make this go away.  If you look at the next slide--


    --here is BMI on the vertical and dioxin on the horizontal.  And so what we've got here is a picture of what I said in the earlier talk, that the comparisons at the high dioxin levels within the comparison group are at higher BMI than the Ranch Hander do, because we think that they are just older and heavier, and they got their dioxins over many years in the States, whereas the Ranch Handers got it in a heavier dose in Viet Nam as young men.  And so that's part of our problem.


    DR. HARRISON: Joel, I would expect the relationship to be cleaner between hypertension and insulin sensitivity.

    DR. MICHALEK: Because we're a little bit closer to the mechanism there.

    DR. HARRISON: Well, because of syndrome X.

    DR. MICHALEK: Because of syndrome X--yes.

    We have--syndrome X involves heart disease, hypertension and diabetes.  And--

    DR. HARRISON: Well.

    DR. MICHALEK:  --40 percent of the cohort has a diagnosed history of hypertension.  About 23 percent have a diagnosed history of diabetes.

    DR. HARRISON: Well--

    DR. MICHALEK: And about 65 percent--

    DR. HARRISON:  --you wouldn't want to--I don't think you'd want to include the people with diabetes.  But if you were to look for a relationship between--and you've got those data now.  You've got the QUICKI data, and then in the smaller group you've got the comparison in between the 29 versus 29, out of one of those I would expect you to be able to show that the loss of insulin sensitivity was related to the prevalence of hypertension.

    And if that were the case--see, the BMI in this case is--I hate to keep using this word--but the BMI is really a surrogate for insulin resistance.  And you already have a better measurement of insulin resistance now.  You have a direct measurement of insulin resistance now.

    DR. MICHALEK: Thank you.  That's our way out of the box on this.  I'll make a note, and--

    DR. HARRISON: I'm very familiar with boxes, and ways out of them.

    DR. MICHALEK: Way out--that's right.



    DR. HARRISON: I learned it from Dr. Osei.

    DR. OSEI: Going back to your last study on insulin sensitivity, do you have blood pressure and lipid profile in those?

    DR. MICHALEK: Yes--at Little Rock.

    DR. OSEI: That will help you.

    DR. MICHALEK: We also have history of medication use, so we know who's on anti-hypertensive medications.

    DR. OSEI: Do you have info on anti-hypertensives--

    DR. MICHALEK: Yes.

    DR. OSEI: In that study?

    DR. MICHALEK: Yes.

    DR. HARRISON: The--

    DR. MICHALEK: Because 40 percent of the group are hypertensive, or a history of, and many of those are on medications.

    DR. OSEI: But, see, that's the very serious problem for you.

    DR. MICHALEK: Yes, it is.

    DR. OSEI: Because somewhat--

    DR. MICHALEK: It's almost half the cohort.

    DR. OSEI: That half of the anti-hypertensive may either go either way.

    DR. MICHALEK: So it's possible that by adjusting for our anti-hypertensive medication, we can clarify--

    DR. OSEI: Exactly.

    DR. MICHALEK:  -picture on insulin.

    DR. OSEI: Exactly.  Because you have [off mike] by insulin sensitivity [off mike], neutral, so you may have to really look at that and tease out, and it may give you more answers.

    DR. MICHALEK: Thank you.

    DR. HARRISON: I didn't hear everything you said, Kwame, but one of the things that I would be concerned about, if you used blood pressure measurements, versus a history of hypertension, is that it's really difficult to get accurate blood pressure measurements.  And unless you standardized on something like a mercury manometer, and unless you standardize on three measurements over a certain period of--you know, it's very difficult to get reproducible, useful data out of blood pressure.

    DR. OSEI: Well, also trying to--the effect of anti-hypertensive medications on glucose regulation.

    DR. HARRISON: Ahh.  Okay.

    DR. OSEI: And half of his population, the last study, which apparently were on some anti-hypertensives.  So this is a serious confounding variable, in terms of interpretation of this data.  And part of his methodology demonstrates, you know, differences could be the effect of these drugs.  Somebody--one group will take and ACE inhibitor which we know will put insulin sensitivity--that it can ameliorate the differences.  So we need to go back and re-look at, you know--you don't have a--the n is not large enough to take 50 percent out, and just look at just those people on low medications and ask that question: is there any difference in terms of the total dioxin on insulin sensitivity.

    DR. HARRISON: Well, you could do that with the QUICKI group, right?

    DR. MICHALEK: Yes.

    DR. HARRISON: He can't do that with the rapid sampling group.

    DR. OSEI: Yes, the--I think the same problem, because the--you know, the drugs themselves can induce insulin resistance, regardless whether you measure fasting glucose or fasting insulin.  So the farther that we're taking the medications, you know, throw you off because of any reasonable deductions you can make.

    DR. HARRISON: What I'm saying, though, is they've got 2,000 QUICKI measurements--

    DR. OSEI: Mm-hmm.

    DR. HARRISON: So if they get rid of 40 percent of those--

    DR. OSEI: Oh, yes, sure.

    DR. HARRISON:  --they still have enough to make the correlation.

    DR. OSEI: Mm-hmm.

    DR. HARRISON: But for the rapid sampling study, they only have 58 total patients there, so they might not have enough numbers when they finish.

    But you don't know how many of those were hypertensive.

    DR. MICHALEK: Not offhand, no.  But I would suspect at least 30 percent.

    DR. HARRISON: Yes.  And--yeah.  Okay.

    DR. MICHALEK: Well, thank you.  That's why we're here.

    Slide, please.


    And here is the concept we had earlier, that there's two ways to get exposure of dioxin: one through skin and one through BMI, and that Ranch Handers, because they had more--a Ranch Hand comparison with the same dioxin level are going to be at different points on this curve--we think. A Ranch Hand with, say 28 parts per trillion will have to be here, as compared to a comparison with 28 parts per trillion.  And we think there's support for this idea, which may explain some of the anomalies we see in this data set.



    Well, we have these complications with dioxin and hypertension that you've just seen.  Now we ask--giving up the dioxin assay measurement, do these trends hold together if we look at what they told us on their questionnaire about what they did on the job in Viet Nam?

    So we took the individuals who had responded to the questionnaire, and we broke them out into quintiles by their skin-exposure index.  And here are the sample sizes, and there are the dioxin levels.  We see they're increasing--age--the point here being that individuals who report high skin exposure were generally younger than those who didn't, reflecting the senior master sergeant who gave the orders, versus the one and two-stripe or the guy on the tanks.  You know, there's a definite demographic difference going on here.

    BMI--and there's the percent hypertensive, and the pattern is certainly not clear by looking at that, certainly on these raw percents whether there's going to be a trend.



    And certainly when you do the adjusted relative risks, you don't see much going on here against the first quintile, and the entire cohort.



    However, if you look carefully at that you'll find that a good percentage of the people in the first cohort were those senior enlisted, and they were administrators in other words.  So, start again--



    --we started over again and excluded the administrators and only look at the non-administrative enlisted Ranch Hand veterans who responded to the questionnaire, broke them out into quintiles by skin exposure--and there are the sample sizes.  And now we begin to see a trend.  And it's obscured by changes in age here.  These individuals in the fifth quintile are a lot younger than the ones in the lower quintiles--still.  And so after you adjust for age, you see a significant increase in the risk of hypertension in the highest quintile of skin exposure in the Ranch Hand enlisted, which is consistent with the idea that there really is a relation between dioxin exposure, or herbicide exposure, and hypertension.

    DR. GOUGH:  Unless you're not exposed at all. I mean it seems to me if you have an essentially non-exposed group, like quintile one, then everything should be different from it, if you're going to have a difference between quintile two and quintile five.

    DR. MICHALEK: Are you objecting to the use of the first quintile as the referent?

    DR. GOUGH: No.  I'm saying--

    DR. HARRISON: He's objecting to the exclusion of the administrators.

    DR. GOUGH: I mean, the exclusion of the lowest exposed groups.  And actually the basis of the comparison--that's brutal.  When he does that--I mean, with them as a comparison, he doesn't find a difference.  He doesn't find a trend.

    DR. STOTO:  But there the reference group.  The lowest quintile is a reference group.

    DR. HARRISON: But the administrators are out of that group.

    DR. GOUGH: The administrators are out of that.

    DR. MICHALEK: These are non-administrative enlisted.

    DR. GOUGH: It just puzzles me that you would do that.

    DR. HARRISON: What, again, was your rationale for excluding the administrators?

    DR. MICHALEK: Administrators?  They're older, heavier and they make a good referent for the younger, heavily exposed enlisted.

    DR. STOTO: I agree with this.  I mean, you want the people to be as similar as possible--

    DR. MICHALEK: Thank you.

    DR. STOTO:  --in every respect except their exposure.

    DR. MICHALEK: Yes.

    DR. GOUGH: Well, the big difference in exposure is one group has none, or very little, and the other groups--

    DR. MICHALEK: Right.

    DR. HARRISON: Well--

    DR. GOUGH: I mean that is really cutting.

    DR. HARRISON: Mike, it is possible, though, to make the argument--

    DR. GOUGH: Yes, I can understand the argument.

    DR. HARRISON:  --that hypertension increases with age--

    DR. GOUGH: And weight.

    DR. HARRISON:  --and with weight, and weight increases with age, and hypertension increases with weight and age.  So, you might have a higher incidence of hypertension in the first group.

    I just think that the other problem that you're looking at here is that the stuff is far removed from a direct relationship, and that's why you're having to go through all these maneuvers to see anything.

    DR. OSEI: The other thing, Joel--I keep on trying to look at the numbers you're showing us: the 51 percent, the 29 percent.  I don't really know of any group that has that high rate of hypertension.

    DR. MICHALEK: You're looking at one right now.

    DR. OSEI: So, there has to be something that--you need to explain the higher rate of hypertension in this group here.  I mean, the general population has a rate of about, maybe 22 percent; the highest in African Americans, 28 percent.

    You have people with half of the population have hypertension.  So their background may be different before even they go--

    DR. MICHALEK: As the principle in statistics, if you're trying to look for a treatment effect, is to produce a group as homogeneous as possible--except on the variable of interest which, in this case, is skin exposure.

    These guys are all non-administrative, enlisted Ranch Hand veterans who were either on the ground loading tanks or sat in the back of the plan, operating spray equipment.  Forty-nine percent of the individuals--

    DR. HARRISON: But Joel, it's like saying that you're relating height to skin exposure, and the average height in the first quintile is six-foot-five, and the average height in the fifth quintile is seven-foot-two.  And we know that the general population has an average height of whatever.

    DR. MICHALEK: We're not asking about the general population.  We want to know--

    DR. HARRISON: The general Air Force population.

    DR. MICHALEK: We want to know of these enlisted Ranch Hand veterans, is there a trend?  And the answer is, yes, there is.

    DR. STOTO: Now, we know that these are older guys, and it's not surprising that what is it some--

    DR. MICHALEK: 49 percent.

    DR. HARRISON: I agree with Kwame.  It is surprising.

    DR. MICHALEK: Well, at least to emphasize the point, that every single case here has been verified by record review.  Every case--all of those 49--have a doctor's diagnosis of hypertension.

    DR. HARRISON: The disagreement is not over the diagnosis.  The problem is--the problem is that, you know, that the average height in the first group is six-five.  I mean, that's--

    DR. MICHALEK: The average what?

    DR. HARRISON:  --totally--it's--

    DR. MICHALEK: Say again?

    DR. HARRISON: It's--what you're saying is, statistically, I'm sure is correct.

    DR. MICHALEK: Yes.

    DR. HARRISON: It's just that what you're saying biologically is biologically implausible.

    DR. MICHALEK: I understood the purpose of this study was to determine whether exposure to herbicides is detrimental to your health.  Is that true?

    There's a hierarchy here of exposure to herbicides, as measured by the skin exposure index, which was given to the men, blinded to their dioxin levels, in 1989.  We ranked the men according to their skin exposure.  We tabulated the percent hypertensive.  The raw percentages are a bit misleading, because these are unadjusted for age.  The individuals in the high category are younger.  That's why this relative risk is 2.6, even though the raw percent down here is 51 and that's 49.

    DR. HARRISON: When you're saying "younger," though, you're still saying that they're in their forties or fifties now.

    DR. MICHALEK: They're at least five years younger than those individuals in the first quintile.

    DR. HARRISON: But they're still in their forties and fifties.  And--

    DR. MICHALEK: This is the entire prevalence from Viet Nam to the present.  Currently, the average age is about 62.

    DR. GOUGH: This is a diagnosis--

    DR. MICHALEK: Are you telling me that this analysis should not be done?

    DR. HARRISON: Well, let me ask you this--are you saying that you have participants in this group who were diagnosed as being hypertensive in the '50s, who do not have a diagnosis of hypertension now?

    DR. MICHALEK: No.  All these diagnoses were after service in Viet Nam, which meant since 1975.

    DR. HARRISON: Well, let me rephrase the question.  Do all of the people in this group--are all of the people in each of those groups presently hypertensive?  Do they presently have the diagnosis of hypertension.

    DR. MICHALEK: Well, if an individual is hypertensive in 1982, and is put on anti-hypertensive medication, and holds his blood pressure steady--

    DR. HARRISON: But is not on a an anti-hypertensive medication now.

    DR. MICHALEK: Well, I don't know if they're on or off hypertension.

    DR. HARRISON: That's right.  You don't know that.

    DR. MICHALEK: I didn't include it in this analysis.  All I included was date of onset, according to the diagnosis.

    DR. HARRISON: So, Kwame, that's why he's coming up with such high numbers.

    DR. MICHALEK: Yes.

    DR. HARRISON: Because you're talking about a cumulative--

    DR. MICHALEK: Cumulative--

    DR. HARRISON:  --diagnosis--

    DR. MICHALEK: --over the entire--

    DR. HARRISON: --where it may well be that half of the people that you've got in this group are not presently hypertensive, nor are they on anti-hypertensives.

    DR. MICHALEK: But they had a history of--they were diagnosed, they had a date of onset.  It could have been anywhere between 1969 and the year 2002.

    DR. HARRISON: Yep.

    DR. MICHALEK: That could explain the high percentages relative to the rates you're quoting.  That's true, I guess.

    DR. STOTO: Well, the other thing is that they're old.

    DR. MICHALEK: That's right.  The earliest birth year is 1908.

    DR. HARRISON: Yeah, I think I met him.


    DR. CAMACHO: We're just seeing the confounding variables.

    DR. HARRISON: It doesn't surprise me--

    DR. CAMACHO:  --confounding variable on these guys?

    DR. HARRISON: I'm not saying age is a confounding variable.  Mike is saying it's just they're old, it's okay for half of them to be hypertensive.  I don't know--

    [Simultaneous multiple comments.]

    DR. OSEI: I'm very confused, actually, on what you did.  I have no idea what you did.

    If you took them when they came in and you measured their blood pressure from day one, then just every so often.  So you're getting a cumulative data over 20 years.  What are you giving us?  What is that?

    DR. MICHALEK: That is the risk of diagnosis of hypertension by a doctor.

    DR. STOTO: Of ever having been diagnosed.

    DR. MICHALEK: Diagnosed.  By a doctor.

    DR. HARRISON: This is a chart review.  This a chart review, Kwame.  It's not a direct measurement of blood pressure.

    DR. CAMACHO: So what's the objection here?  The age is a confounding variable?

    DR. HARRISON: I don't have any--no.  I think--the majority of these people are white, so--the incidence of hypertension is different in African Americans--Kwame and I will differ a little bit on this--but I expect a third of adult African Americans to be hypertensive, and that can slide up a little bit.

     But for White Americans--20, 22 percent is probably a much more reasonable--

    DR. STOTO: These are not White Americans, these are White Americans over 50.

    DR. OSEI: Even still, you are now talking half of the White Americans are hypertensive.  Is that what you're saying?

    DR. STOTO: It's not half, either.  It's--

    DR. HARRISON: 49 percent.

    DR. STOTO: Well, that's 49 percent in one quintile, but that's 35 percent of the quintiles overall.  So it's about 40 percent.

    DR. MICHALEK: Overall, it's 40 percent.

    DR. STOTO: Yes.

    DR. MICHALEK: If I could enumerate every single case, I would.

    DR. HARRISON: But if you take into account that some of these are people who got a little bit overweight, got hypertensive, got put on a hypertensive medication--or simply got told to lose weight and did so.  And then you've got some other people, you know, who may have--you're not even distinguishing between systolic and diastolic hypertension here.  So you've got a few people in there that are--I mean, that's a mixed bag, as well.

    DR. CAMACHO: But all this data was taken at a point in time in the past.

    DR. HARRISON: No.  No.  All this--this is cumulative.  So if you were diagnosed as hypertensive in 1971, and no mention was made of it since then, you're in this group.

    DR. MICHALEK: That's right.

    DR. HARRISON: And that's why the numbers, I think, are higher.  And that's why I think they're inaccurate.

    DR. MICHALEK: Why would they be inaccurate?

    DR. HARRISON: Because you have no way of judging whether, for the different quintiles, that the occurrence of hypertension was distributed the same between the quintiles.  You know, you're looking at a phenomenon, but you're not--it's--

    DR. STOTO: That's the nature of this kind of work, because you have to look at something.  And then there's often not a perfect measure.  But the key thing is that it's consistent.

    DR. HARRISON: No, it's not.

    DR. STOTO: It's consistently defined between the people in the five different groups.

    DR. HARRISON: It's like extracting and saying that the fifth quintile has two noses.  I mean, it could be consistent statistically, but it just wouldn't make sense.

    Just because it's consistent statistically, doesn't make it--doesn't mean that you should.

    DR. STOTO: It makes it the proper epidemiological study.

    DR. HARRISON: It means that you can only to your mother about stuff like that.

    DR. CAMACHO: But what's the statement?  These people experienced hypertension at some point--at least once.  As long as that's said, what's the--

    DR. OSEI: I will tell you the major--[off mike] you said it--these are some of the issues.  I don't have enough breadth, the reviewer on one or two comments.  These are some of the things--if I had read your paper, this is what I'm going to ask.  Unless you really think through that, it's going to be rejected again, because at the end of the day, I have not read this paper very well--I can't even tell how this study--how you got this focus?  And [off mike].  And that's the point that Paul makes.  The fact that I went to a doctor and one day had hypertension, my blood pressure is high, doesn't give me a diagnosis of hypertension.

    DR. MICHALEK: No, no.  These are all diagnosed by physician.  We have a record--a medical record proof of that.  We have a signed statement by a doctor, "He was hypertensive," at a certain point in time.

    DR. OSEI: Okay.

    DR. STOTO: So how about if we asked--can you think of a better way to define hypertension for a study of this sort?

    DR. MICHALEK: How else would you define hypertension?

    DR. OSEI: I can't tell you.

    DR. STOTO: All right.  If you can't come up with something better, you can't criticize this.

    DR. GOUGH: But if it's a meaningless measure, you don't want to do the study.

    DR. STOTO: Well, it's not a meaningless measure.

    DR. HARRISON: One of the first cuts that you could make it at is you could ask these people--

all right, I'll try this on for size.

    I would make the definition of hypertension those people who are presently on hypertensive medication; not have a history of having taken it in the past, not a history of it having been diagnosed in the past, but those people who presently are on anti-hypertensive medication.

    DR. STOTO: What about the people who have been diagnosed, really have high blood pressure now, but can't stand the medication?

    DR. HARRISON: All right.  I'll add that group in: those people that, on this physical examination had an elevated blood pressure--or elevated diastolic blood pressure.  Let's get rid of some of the white-coat hypertension.

    DR. OSEI: If you have--you said that the doctor's specification that--nobody measured blood pressure.

    DR. MICHALEK: Yes.

    DR. OSEI: From where?

    DR. MICHALEK: That probably led to a diagnosis.

    DR. OSEI: All right.  So, if you are going back through the charts, you should be able to find out when--

    DR. MICHALEK: Mm-hmm.

    DR. OSEI:  --and what was the blood pressure?

    DR. MICHALEK: Yes.

    DR. OSEI: So you can assess change in that.

    DR. MICHALEK: Yes.

    DR. OSEI: That would strengthen this, because just the fact that somebody tells you that--

    DR. MICHALEK: Right.

    DR. OSEI: --it's not good enough.

    DR. MICHALEK: No, no.  It's not a self-report.

    DR. OSEI: Maybe define what hypertension is.

    DR. MICHALEK: It's been verified.

    DR. OSEI: We have GNC-6 numbers.  You can actually--this is standard definition of hypertension.

    DR. MICHALEK: Yes.  For every single case, we have document of the date and the doctor's signature that they're hypertensive.

    So we know the date, we know the--

    DR. OSEI: And what was the blood pressure--

    DR. MICHALEK: We know the blood pressures--yes.  It would probably be on the same document.

    DR. OSEI: Would it be grouped by date?

    DR. MICHALEK: That's what this does.  This is the Cox model analysis, the same model we used to track the diabetes.

    DR. CAMACHO: So this is grouping by dates--

    DR. MICHALEK: Yes.

    DR. CAMACHO:  --of when they were reported hypertension.

    DR. MICHALEK: Yes.

    DR. HARRISON: Okay.  Okay.

    You know, I'd like to take another tack on this.  It may be--I'd like to return to my earlier comment.

    It may be that the problem here is that it's kind of a bloodless statistical observation, without--it doesn't point you towards any mechanistic explanation for what's going on.  And if it were possible to go back and do some sort of an analysis that tried to relate the occurrence of hypertension to the present--frankly, I think I would much prefer those people who were presently hypertensive and relate that to their insulin sensitivity.  It that turned out to show a relationship, I think then you'd have a paper that would excite a lot of interest, and you wouldn't have much trouble selling it.

    The problem here is that it's a--I'll take Mike's word for it, I'll take your word for it--it's a valid statistical analysis of phenomena, that doesn't point to any mechanism.

    DR. CAMACHO:  Is that skin exposure--

    DR. HARRISON: It's not a mechanism.  That's just a phenomenon.  You know, you're saying this phenomenon and this phenomenon have a statistical relationship.

    DR. CAMACHO: Are correlated.  Yes.

    DR. HARRISON: And I suspect the reviewers are saying, "Well, yeah, okay."  It doesn't light my fire.

    DR. MICHALEK: I think that's helpful.  And I think we can do something along those lines.  Thank you.


    DR. GOUGH: I'm unclear--if there's a good correlation between the skin exposure index and current ppt, then why do these results turn out so different?

    DR. MICHALEK: Why are they different?

    DR. GOUGH: Well, when you do it by current ppt, you don't get this trend.  It's only when you do the skin--

    DR. MICHALEK: Well, you saw a trend with current ppt also.

    DR. GOUGH: Here?

    DR. MICHALEK: Yes.  We did the deciles--back up a few slides.


    DR. GOUGH: Okay.  But then what do we get different when you do the--all right.  But they go up together.

    DR. MICHALEK: Hypertensive risk increases with dioxin.

    DR. GOUGH: Yes.  Okay.  But then, didn't you do a comparison between the first through fifth quintiles, based on ppt's?

    DR. MICHALEK: Yes.

    DR. GOUGH: And wasn't the result from that different from doing it based on skin exposure index?

    DR. MICHALEK: No.  We saw a trend with increased risk in the increased quintile.

    DR. GOUGH: That's the one I'd like to see.  Because I missed it when you went through.


    DR. MICHALEK: Right there.  It's increasing in both groups.  Take a quintile dioxin in each group.

    DR. GOUGH: Okay.  Then why is it necessary to do this skin exposure index?

    DR. MICHALEK: Only because some people object to the dioxin measure because of all the caveats that are associated with it, such as differential elimination, such as confounding with body fat, such as it's being measured in 1987, and the exposure that occurred in 1967.

    DR. GOUGH: Okay.  All right.

    DR. MICHALEK: And the advantage of--

    DR. GOUGH: Okay.  Okay.  Well, if the results are the same--fine.

    DR. MICHALEK: Yes.

    DR. GOUGH: You said the results are correlated.  All right.

    DR. MICHALEK: But I think the comments we've heard so far will help us strengthen the paper.  Thank you very much.

    DR. HARRISON: Okey doke.

    DR. MICHALEK: So, that's just about the end.  Let's continue on and finish this.

    So-slide, please.


    This is the same slide you saw before.



    And so we have this puzzle of no great difference on a history of diagnosed hypertension, and yet we see significant trends within each group.

    Some of the same mechanisms would apply here that we thought applied in insulin; that is, the comparisons at high levels could be older and heavier, and our statistical modeling has failed to accommodate that--to adjustment on the Ranch Hands with low levels who just happened to be lucky, and they're able to eliminate dioxin quicker, and that's why they're low.  And that would make them healthier, and that would confound our dose response patterns.



    And these are all the comments that we've taken into account: possible uncontrolled confounding, and dioxin level may depend on a lot of things, including--as well as exposure and body fat and uptake.  Uptake may correlate with BMI, and it may correlate differently in the two groups.  These are all speculations.



    Strengths, we already know about.



    Limitations, we know.

    Now, it was rejected recently by the American Journal of Epidemiology.  One of the reviewers said this report will have a major impact on public policy regarding the regulation of dioxin use.  He liked the paper, basically, and his comments were fairly trivial.



    And he thought we should focus on the within-group analyses, and not at the size of the between-group--similarity between the cohorts, and he was fascinated by the dose response within the Ranch Hands and controls.  And he mentioned the EPA.

    However the second referee was far more critical, telling us we shouldn't have used the Cox model.  We should have used logistic regression, which is a point we've raised many times, and both methods give the same result, basically.



    "Maybe due to confounding or some other variable that operates at the low end;" "This is all speculation," but we've tried all available covariates and cannot make these patterns go away.



    You know, he doesn't like the fact that we used the change in BMI, but overlooked the fact that we used the absolute BMI.  He thinks we used the change only, and we didn't.  We used both the change and the current BMI, and a tour BMI.

    And he thought we should be using current alcohol consumption and smoking.  But we didn't.  We used the measures made in 1982, because we thought that however much he drank or smoked today should have no bearing whatsoever on a diagnosis that occurred 10 years ago.  And so we tried to use an earlier measurement of the covariate to avoid the problems associated with time-dependent covariates.

    Slide, please.


    And he was just trying to discount the findings, saying the trends are not strong, and that there are insufficient data to support a causal relation.  Well, that's not our job anyway.  We're here to demonstrate an association or not.



    "Discussion of possible mechanisms was vague."  Well, after you present the findings, in a discussion section the idea is to speculate, and that's what we did.  And he didn't like our speculations.

    So I think your ideas would help us revise the paper.

    DR. HARRISON: Okay.  Any last comments on this paper?

    DR. TREWYN: Okay.  Since one reviewer liked it and one didn't, but the editor just rejected the paper.

    DR. MICHALEK: Yes.

    DR. TREWYN: You did not re-contact, explaining how you could refute Reviewer #2?

    DR. MICHALEK: The editor's letter was clear.  He didn't want to hear rebuttal.  We need to start over with a different journal.

    DR. TREWYN: And--yes.  And that's always an option anyway.

    DR. HARRISON: Was it a definitive letter?

    DR. STOTO: Just had kind of a saturation effect--there are just so many epidemiology journals, that--

    DR. MICHALEK: One referee said we publish too much.


    DR. HARRISON: Blame us.  Shame, shame, shame.

    Any other comments before we go on to the next paper?  We're only 15 minutes late now.

    [No response.]


Thyroid - Review Results of the Latest Paper

    DR. HARRISON: Thyroid.  Thyroid.  Basically, this paper says there wasn't any thyroid--weren't any thyroid changes, right?

    DR. MICHALEK: There was no association between thyroid disease and dioxin, but there were associations between thyroid hormones and dioxin.

    DR. HARRISON: But you didn't measure TSH.

    DR. MICHALEK: We did measure TSH.  TSH changed with dioxin.  T4 and T3 did not.

    DR. HARRISON: Okay.

    DR. MICHALEK: This paper was accepted by the Annals of Epidemiology, and is in press.  It was presented in Barcelona.  I just show it here, because you may not have seen this before.  This is a summary of all of our thyroid data over the entire study, from 1982 through 1997, written under contract with Drs. Schecter and Pavuk at the University of Texas at Dallas.

    Slide, please.


    We all know the background.



    You know the goals.



    Okay.  We're talking about any veteran who participated in any of the five physical examinations from the baseline to cycle five.  And here is the same dioxin category analysis that we've been using in all of our published papers where we have broken out comparisons in Ranch Hand into these four categories.  And this slide is simply showing that three is some heterogeneity across these categories that the Ranch Hands in the high category are far more likely to be enlisted ground crew than those in the background category, and they're far more likely to be officers.



    And here are the dioxin levels across categories, in case you didn't know these.

    In the high category, the median dioxin is 46 parts per trillion, and in the background category on the Ranch Hand side it's 6 parts per trillion, and in the comparison group it's 4 parts per trillion.  And the maximum dose in Viet Nam is about 3,500 ppt.



    And here is a graphic showing the dioxin levels in the background, low and high categories, compared to the comparisons.



    And here is the finding of the paper.  If you look at log transformed TSH means by dioxin category, by physical exam year, you'll see a significant trend in many of the exam years with higher TSH means in the high dioxin category.  In spite of the fact there's been a change in the laboratory method for measuring TSH--here is the Kelsey Siebold data, it was obviously on a different scale form the measurements made at Scripps in 1985 through 1987.  And those were all accommodated in the statistical analysis.



    In a repeated measures analysis that accommodates all data over all cycles in all participates we see a significant increase in the mean TSH, using a repeated measures linear model.  Accommodating all data, we see this significant increase, whereas the other analyses were cross-sectional in every single physical exam year.



    And, however, we saw a mean shift.  We didn't see a significant increase in abnormally high TSH levels.  The odds ratios are all increased, but none of them reach significance in any of the examination years in the high Ranch Hand category.



    And we saw now corresponding pattern with T4--



    --or T3, which is not shown.  And when we considered diagnosis of hypo- or hyperthyroidism, we saw no significant increase in risk in the Ranch Hand high category.

    So this is the conclusion.  We found a significant dose response between dioxin and TSH, but not between dioxin and T4 or T3, and no significant increase in the risk of thyroid disease with increased dioxin level.

    Thank you very much.

    DR. HARRISON: Whew.  Neat.  Or fast.

    DR. STILLS: [Off mike.] What is the thought behind the increases in TSH?  The link between dioxin exposure and TSH?  Is it--

    DR. MICHALEK: There are only two thoughts I have.  First, the results are consistent with animal studies that show an increase in TSH with dioxin body burden, perhaps in mice, and that whatever effect is going on is sub-clinical.

    The same is true in our analysis of liver data.  We found increases in liver enzymes, but no increase in liver disease with increased dioxin--the interpretation being that perhaps a dose is at such a range that no adverse disease endpoint could be detected, yet measurements on enzymes did detect a mean shift.

    DR. OSEI: [Off mike.] Do you have antibodies--

    DR. MICHALEK: We measured what are called microsomal antibodies and we found no dose response.

    DR. OSEI: [Off mike] Okay.  How about--antibodies?

    DR. MICHALEK: That data was not available.  I'm not sure whether we're measuring that in cycle six or not.  I think the answer is no.

    DR. OSEI: [Off mike] And these are--in T4, what are you looking at?  The number shift between free T4, and just the total T4?

    DR. MICHALEK: I think we measured both--free and bound.  And that's described in the paper.

    DR. OSEI: Okay.

    DR. HARRISON: Didn't you measure what's called T-3 uptake?

    DR. MICHALEK: Yes.

    DR. HARRISON: Okay.  So and what you then calculated, using the T4 and the T3 uptake was a free thyroxin index--

    DR. MICHALEK: Yes.

    DR. HARRISON:  --but you didn't measure free T4.

    DR. MICHALEK: No.  We measured the free thyroxin index.

    DR. HARRISON: Okay.

    DR. MICHALEK: And found no dose response in free thyroxin.

    DR. HARRISON: This is subtle enough.  I'm not sure that that matters a whole lot anyway.

    Okay.  And it's accepted.  Hey.  Whoo.

    DR. MICHALEK: Yes.

    DR. HARRISON: Go, man, go.


    Before we go on to the next topic--just in case we lose anyone along the way--there was some comment earlier that August would be a bad month to plan a meeting next year.  And the idea was to shift forward to September.  And the only thing else that I wanted to raise was that if we shift from August to September for our second meeting, then we--as Len pointed out--we should switch from October to November for our third meeting, otherwise we'll be killing ourselves.

    So, does it sound reasonable to everyone to plan now on April, September and November for next year, and we should try and come up with some actual dates--proposed dates--pretty soon, and circulate those.  Hopefully, they'll be far enough in advance of everyone's schedule that we can make a meeting date and you all can adjust your schedule for those?  Does that sound?


    DR. CAMACHO: [Off mike]  April I know I've got one commitment on 12, 13, 14--is not good.

    DR. HARRISON: Does anybody else have commitments in April so far?

    DR. CAMACHO: Well, you're talking about April '04.

    Same date always on hold.  It's going to be that same time--12, 13, 14.

    DR. STILLS: [Off mike]  I'll have to check my--it may be around that time, too.  So I'll let you know.

    DR. HARRISON: So we may be looking at late April.  Either that, or we'll have two new members then. We might not need them.


    Okay.  All right.  But we're in general agreement that those months are okay.  April, September and November.  Okay.

    So let's go on into the next show.

Statistics on Study Compliance to Cycle

    DR. MICHALEK: Just to give you a little summary of how we stand currently on the physical exam at Scripps--and I may need some help here from Bill Grubbs--here's the total number of when this slide was made, which was up through--we have 81 groups that are going through Scripps.  There's two groups a week.  This slide was made as of group 52, which passed through the clinic on the 20th of November last year.

    At that time, we had 1,932 scheduled their exam, and the entire total from last cycle was 2,204 at this time, at group 52.  So you see we're running behind what we did in 1997.

    Currently, 45--or at the time this slide was made--45 were in process, 7 were being located.   These numbers change every day.

    A number of were excluded right up front because we had realized that they're not friendly, they're not comfortable, they don't want to be talked to, and we called them "hostile," or "can't locate"--and 216 were called hostile.

    And so I think our current best estimate is that we're going to arrive at about 1,960 individuals physically examined at Scripps this cycle, as compared to--what was the total last time?

--2,121.  So it's running around 200 less than what we had in 1997--for a variety of reasons.  There's apparently no single dominant reason--is there?--for this.  It's sort of an across-the-board spread of reasons.  And we're looking at those very closely now--or we will--specific group differences.

    DR. HARRISON: Joel, were the hostile--what you call the hostile people, were they always hostile?  Or did any of them become hostile?

    DR. OGERSHOK: Most of them--I think the vast majority of that number have always been hostile.  There have been a few that may have gotten that way after a written exam or two, for one reason or another; or people that showed up with basically problems--weapons--


    --and we didn't invite them back.

    DR. CAMACHO: Sounds hostile to me.

    DR. OGERSHOK: There was one gentleman that threatened to--he had an Uzi and he was going to use it on everybody.  So we didn't invite him back either.

    DR. HARRISON: Was he told anything about the identity of the advisory committee?


    DR. TREWYN: Or does he want to be a member of the advisory committee?


    DR. MICHALEK: We never analyzed the hostiles.

    DR. HARRISON: Okay.

    Well, you know, obviously the question, Joel, is: is there any increase in the hostility with time.  It sounds like you're saying no, there isn't.  But it also brings to mind an interesting difference between the attitude of the people that we talked to, who obviously aren't hostile and who are here participating, and they're all, you know, very enthusiastic about it.  And then you've got a smaller--much, much smaller group that, you know, feels differently.  And I--if they were like that way from the beginning, then you can absolve yourself of any responsibility for it.  But if they developed this with time, then--

    DR. MICHALEK: Well, you raise a good point.  Do we have the data that leads to when we declared them hostile, so we could analyze that?

    DR. TREWYN: It may have to do with that original Holiday Inn versus the Marriott.

    DR. HARRISON: Okay.

    DR. OSEI: [Off mike.] The get tired.  It's very fatiguing.  It's not uncommon to these this--and this is a long-term study, before they come here --they want to contribute to whatever--same boat--give up.  It's not that they're even hostile, but that they're just fatigued.

    DR. MICHALEK: Some of these men are genuinely hostile.  "Don't call me again."  You don't want to mess with some of these people.

    Thank you.

    Next slide, please.


    At every physical exam, the participants fill out a questionnaire rating us, rating the clinic, rating the investigators.  And you see the very high scores we get on the different components of the study.  And these are all real numbers. These are the men telling us what they think.

    And so the current study is very similar to the last study in many components, but you'll see a few differences.

    Slide, please.


    DR. HARRISON: What range of choice to they have, Joel?

    DR. MICHALEK: It's a scale.

    DR. HARRISON: Is this a five-scale?

    DR. MICHALEK: "Good, excellent, fair, poor," and this is the percent that scored us excellent or good.

    DR. HARRISON: Good, excellent--

    DR. MICHALEK: Fair or poor--something like that.

    DR. HARRISON: You don't have--

    DR. MICHALEK: Do you remember the choices?

    DR. GRUBBS: I think it's excellent, good, fair or poor.

    DR. MICHALEK: Okay.  And not applicable.

    I guess the only one I want to point out is--somewhere in here.  Maybe it's in the next slide--



    --psychological testing, we're doing better this time.  What's different this time is we actually had a psychological test being administered by a psychologist, works with memory scales, whereas in previous cycles it was all self-administered testing which, I guess, can get kind of dull.  So we seem to be having a favorable response to our psych testing, and we seem to be doing better with the cafeteria meals this time--


    --as opposed to last time.

    DR. HARRISON: Okay.  Questions?

    They like the psychological evaluator better than themselves--right?  That's the--


    Okay.  Moving right along.  Roll on, Joel.

He who hesitates is lost.

Latest Results on Consent for Future Use of the


    DR. MICHALEK: Okay.  The whole idea of--the end result, the disposition of the study after 2006 has been on the table now for a while, and we have given options to Congress, and to the Surgeon General about, you know, what are the options regarding new study.

    The whole concept of continuing the study is tied up with informed consent.  If the study is to continue in any way whatsoever, study subjects must provide their consent.  And so at the current physical, in addition to the consent forms that they have to sign for the procedures that Scripps is administering to them, they have another consent form that says something to the effect, "You have my permission to use my data for continuing Agent Orange Research," or "for continued study of other health endpoints," or you can't use my data at all.

    So they have something like--they have three questions--in fact, the next slide shows that right here.  These are the three options.  This is copied right from the consent form.

    "After September 30, 2006, your data may be used for Agent Orange research and other military health issues."  That was a choice.  The threw a box next to each one of these, and they would check off what they wanted to do.

    The second choice was "May be used for Agent Orange research only."  And finally, "May not be used for any purpose" at all.

    Three boxes, and they were to check one of them and sign it.  And this was one of the many consent forms they were presented with at Scripps on their first morning there.

    Is there a question about the consent form before I continue?

    DR. HARRISON: There's another one about samples?

    LT. COL. ROBINSON: There's three consent forms all together.  A general, HIV, and then the future use.

    DR. HARRISON: Just to cut to the chase, I was curious about the high percentage--very high percentage that agreed to allow--

    DR. MICHALEK: Yes.  We were really elated--