We are glad that you are all here today to listen to Dr. Haley's presentation, also the questions that will be coming from the personnel at the front table. I request that if the audience does have a question that you pass it to me and time permitting, we will attempt to get to it. However, the priority is to the learned individuals that we have gathered here today to pose questions to Dr. Haley.

With that, I would like to pass the meeting on to The Honorable Jesse Brown, former secretary of Veterans Affairs. Secretary Brown?

I would also like to modify something that was just said. I wanted more of an open process. Of course, we want to make sure that the people that are here have an opportunity to conduct their business. But at the same time, I want everyone in this room to feel perfectly free to ask any questions that they may have.

And we are not necessarily going to be limited by written questions. I would only ask that you be reasonable and give everybody an opportunity to participate. When you leave here today, I want everyone to feel that they have had a chance to not only voice their concerns, but to get answers to any questions that they may have if that is possible.

Having said that, I would like to just turn the proceedings over to ADM Zumwalt for an opening message.

As you know, we opened -- our committee meetings are held in the open so that everybody can see what we are doing. And this is all part of that process. And by the way, there is going to be another special oversight board public hearing here in Washington in July. The date is --

MR. KAPLAN: July 13th at George Washington University.

RADM STEINMAN: To which you are all invited, of course.

ADM ZUMWALT: Dr. Vinh Cam.

Those of us who went through the Vietnam War, grapple with the Agent Orange for many issues. During the beginning with a government policy to conclude that there was no correlation between Agent Orange and diseases.

And it took the battle of many years to bring out the truth and to get final justice in terms of a system that works and the progressive designation of diseases for which there is a reasonable likelihood that there is a causal relationship between exposure to Agent Orange and disease.

As a result of the lessons learned in that battle, the government and the public, I think, have come to a much more efficient process, in this particular case, with regard to the Desert Storm diseases and Syndrome.

The responsibility of this special oversight board is to report, at the end of an 18-month period, to the President in a way to give him confidence that everything that can be done has been done by the governmental process and the interaction between the public and the government to come up with good, sound answers.

Now I will ask if each of the people at the table, beginning with you, Dr. Green, will give their name and their brief background.

DR. GREEN: Thank you. My name is Dr. Allan Green. I am of counsel to the law firm of Mintz, Levin in Boston and Washington, but for today's purpose, I think the relevance to my background is that for many years I was medical director of New England Nuclear Corporation and involved with medical imaging in the development of a number of medical imaging agents, medical imaging techniques, and have been involved for over 25 years with development of new drugs and devices largely working through the university system and to some extent, through the Food and Drug Administration and the NIH.

So I have had an interest in the evaluation of these sorts of problems we are talking about and the technology that we will talk about today.

DR. SAMET: Even though I don't have a name tag I have a name. I am Jon Samet, the chair of the Department of Epidemiology at Hopkins. My background is in internal medicine and pulmonary disease and also in epidemiology. And I have a long-time interest in the health effects of the environment.

DR. ZEGER: My name is Scott Zeger and I am professor and chair of the Department of Biostatistics at Johns Hopkins. And I am also interested in the environmental exposures and in quantitative methods necessary to assess risks from those exposures.

DR. GUILARTE: My name is Tom Guilarte. I am a professor in the divisions of toxicological sciences, of radiation sciences, and the department of environmental sciences at Johns Hopkins University School of Hygiene and Public Health. And my area of interest is in neurotoxicology and neuroscience.

DR. LIN: Good morning. My name is Melody Lin. I am the deputy director for the Office for Protection from Research Risks, also known as OPRR, which is a part of NIH. Our office enforce and implement Department of Health and Human Services regulation for the protection of human subject. Also to implement public health service policy for the humane care and use of laboratory animals.

DR. HALEY: Robert Haley.

ADM ZUMWALT: Next over here. Tim.

DR. GARRITY: Yes. I am Dr. Tim Garrity, special assistant chief research and development officer for the Department of Affairs -- Department of Veterans Affairs here representing the Department of Veterans Affairs.

CPT MAZZELLA: I am Peter Mazzella from the Department of Health and Human Services. I direct the Office of Military Liaison of Veterans Affairs. And ours is mostly a policy office except we do have to weigh in on many, many Gulf War issues.

LTC FRIEDL: I am Karl Friedl from the U.S. Army and Medical Research and Materiel Command. I am the research area manager for the military operational medicine and research program and coordinator between the services for that kind of research. And I am also the -- I guess the technical manager for a grant that we have to Dr. Haley at this point.

COL ABREU: My name is COL Michael Abreu and I work at -- I am the director of investigations and analysis at the Office of the Special Assistant to Gulf War illnesses.

LTC RIDDLE: I am Dr. Rick Riddle. I am an epidemiologist, and I work for the Office of the Secretary of Defense for Health Affairs.

DR. SPENCER: Peter Spencer. I am professor of neurology, director and senior scientist of the Center for Research on Occupational and Environmental Toxicology at Oregon Health Sciences University in Portland. My interests are specifically within the area of neurotoxicology, and I have done some elaborate work on the Gulf War related illnesses.

ADM ZUMWALT: Well with this group of distinguished scientists, we have everything necessary to help us distill wisdom today. Let me introduce Dr. Robert Haley.

Robert Haley. I am from Southwestern Medical School and in the Department of Internal Medicine. I am an internist and epidemiologist.

My career that brought me here involved -- after an internal medicine residency at Parkland Hospital in Dallas, went to CDC and spent 10 years at CDC basically investigating epidemics, doing research relating to hospital acquired infections, and controlling epidemics in hospitals.

And then I directed the hospital infections division for several years in supervising epidemic investigations over the country and some oversees. Then came back to Dallas, been there for 16 years basically looking, investigating, thinking about epidemic investigations and doing studies academically on that subject.

The way I came to this subject was interesting. This was not a research interest that I was planning to get into. I was busy working on a study of Hepatitis C, which still sits on the shelf, and one day got a call from the president of our university, Kern Wildenthal, who asked me to come over the next morning to meet with Ross Perot.

I said, "Ross Perot. Why would I meet with Ross Perot? I don't know Mr. Perot." And he said, "He wants to tell us about the Gulf War Syndrome." And I said -- truly I said, "Look, let's don't involved in that. That is a psychological issue. It is stress. It has been decided. Let's don't waste our time." He said, "Well, just hear him out."

So I showed up and with Jay Sanford -- some of you all may know Jay. He had come back to Southwestern at the time, had just rejoined our faculty. And we met and Mr. Perot said, "I have been talking to veterans' groups for 30 years and something new is happening in the last couple of years." This was early 1994.

And he said, "Every time I go to a group now, I see some veteran or several veterans come forward with their wives and their company commanders and they say, 'This isn't the guy who went over. Something has happened to his personality. It is a different person. And nobody is looking into it anymore, he is not getting any help, and we don't know what to do.'"

And he said, "This is happening so regularly now -- I am a judge of character, I am not a scientist, but I am a judge of character, and I think this might be real. So we need an independent study and if you guys would do it, I will help fund it."

So we were really -- our interest was peaked and we decided to do an outing. We assured Mr. Perot that if this was psychological, if it was stress, we would include that as a very important part of our research and see if we could figure out what the psychological problem was so maybe we could help these guys and we would tell him truly what we thought. And he said, "That is what I want. I don't want to buy a solution, I want to know what is going on. So go for it."

So we started. What I am going to describe today is basically -- and this is an important term, an important characterization at the very beginning. What I am going to show you is an epidemic investigation. It is not a chronic disease epidemiology study of heart disease or cancer.

And this is a very important distinction because in epidemiology, there are different tools that we use to study heart disease or cancer -- an ongoing problem in a population -- as opposed to a one-time common source epidemic in which the exposure occurs in a time limited fashion followed by an uprise and fall in disease, what we call an epidemic curve, in the cessation of new illness.

That is what has happened. And all the way through this issue, as -- we have been involved now for five years. I have been working full time on this now for five years. And of all the things that I see that have confused the country on this, it is the failure to distinguish a one-time common source or at least time limited exposure situation followed by an epidemic curve and cessation of illness.

And failing to distinguish that from our usual epidemiologic pursuits that deal with chronic ongoing diseases that are continuing to occur that have long incubation periods, many multi-factorial causations, and require different much more sophisticated techniques to handle it.

Now let me tell you the difference. And I spent 10 years at CDC learning sort of the CDC fire drill, if you will, for handling epidemics. And this is a technique that has been used for 50 years now to solve, I don't know, countless epidemics, including Legionnaires Disease, Toxic Shock Syndrome, AIDS, the Four Corners Pneumonia epidemic, Hantavirus, et cetera.

Basically what you do is when you see an epidemic, you suspect a common source. You don't need to do a really sophisticated study where you prospectively enroll people and measure risk factors with sophisticated things. And the reason you don't have to do that is this.

In a common source epidemic, what you are dealing with is a limited set of causes that have a short incubation period and therefore, the disease that you get is a relatively homogeneous disease. So therefore the measure -- if you can come up with a case definition to characterize that disease, the measurement of the disease is very easy. The problem is sometimes you can't come up with a case definition. That is the difficulty.

The other thing is that the association between the causal factors and the disease are generally very strong because they are immediate and usually a fairly small number of causations. So therefore the technique -- so therefore you expect to find very large relative risks. Often in epidemics you get relative risks between 5 and 10, for example. Whereas in chronic disease epidemiology study you get a relative risk of one-and-a-half to three.

Therefore, since the disease has already occurred, the cause is already often over, you can't prospectively measure the risk factors. So therefore the time honored technique is a case control study. You create a case definition from examining a small number of the people. You pick some representative cases, define the case definition, go to a population that has the disease, define in that population who satisfies the case definition and who doesn't.

An example of a case definition would be a statement. The case of toxic shock syndrome is defined as people with high fever, low blood pressure, and red skin. That was the case definition for toxic shock syndrome. Then you go to a group of people who may have it, you divide them into those who satisfy the case definition and those who don't. You call them the cases and the controls. There are other names, but to make it simple, the cases and the controls.  

And then you deliver a questionnaire to them. Well, first, you probably examine the people to be sure this is really an illness and not some fictitious epidemic, what we call a pseudo epidemic. But if it really does look like a physical illness, you then administer a questionnaire to the cases and the controls and ask them what happened to them before they became ill. And of course, you have to include on the questionnaire the likely causes. You have to have a good hypothesis.

Those are then self-reported symptoms and self-reported risk factors. And that technique has been used for 50 years and has solved all the major operates in our country. However, to a chronic disease epidemiology that is heresy because you can't learn anything from self-reported symptoms, although all symptoms are self-reported, and you can't learn anything from risk factors that are self-reported, you have to measure. You see that is a chronic disease model as opposed to an epidemic investigation.

Now let me tell you, I know that this -- the way I am starting here is going to rankle some people because many people who are chronic disease epidemiologists don't like the epidemic investigation model. The problem is it works. It is a time honored technique that works.

Okay. Let me go into the slides now. In fact, let me say we are going -- I am going to organize this in sort of two segments. The first hour I am going to go through the epidemic investigation, the initial phases, to show you what we found. You have already read those papers, presumably, we passed them out. But I will go through the high points.

Then we will take a short break and then I am going to go into all new information that we have -- some we have just published in the last week or so and others are unpublished. But let me say I don't know who all is in the audience here, but I hope that people will be responsible and not take away the unpublished information and publicize it because we want to go through the peer reviewed scientific process.

And part of this not peer reviewed and should not be assumed as final information until it is. However, this is my one chance to talk to you. So we are going to pull out everything we know at this point. And the handout is really related more to this afternoon's discussion -- I mean, the second half of the discussion this morning.

All right. Let me go over now the strategy that we have followed. And the first part of this was laid out in the beginning. The second part has been an evolution since then. We started out with an exploratory, hypothesis raising epidemic investigation. Typical CDC study. We had an original epidemic investigation in the Seabees unit, which I will talk about.

And the purpose of this was, one, to come up with a case definition to see if we could define the syndrome, this Gulf War Syndrome or these Gulf War Syndromes, and then to define the risk factors and see if we could find an association between the two. A classic case controlled CDC study.

We then had planned that if we do find a disease, we would then do a secondary nested case control study in this little epidemic investigation. In other words, we would -- if we find an illness, we will fix them. Representative cases and some matched controls and do a case control study to verify whether we can verify any organic disease or not in this group.

Third, after we did this, we then decided to come back and revisit all of this, since we found some promising findings, and really pull out the stops. We then got funding by the Defense Department. Original studies were funded by Ross Perot. And did an extended case control study basically on the same design of the first nested case-control study, but looking for differences in genetics susceptibility to organophosphate injury, which we felt this was, and to try to quantify or measure the brain injury much more powerfully than we did in the first test.

Frankly, at this point, we still weren't convinced that there was anything real here. But after this was done, we felt this was very promising, we felt that this warranted a lot of resources.

Not shown here and not actually done by us, but Tom Kurt, who is my toxicology colleague, who was involved with us early, he was very early interested in the organophosphate hypothesis and interested Ross Perot in funding a study at Duke University with Dr. Abou-Donia, where Tom Kurt and Dr. Abou-Donia developed an animal study to look at the synergistic effects of organophosphates in similar chemicals that were indigo to see if combinations of those might cause neurological injury when individual -- in a situation where individual agents wouldn't.

So our hypothesis we also tested in animals, although we didn't do that at Southwestern, that was done at Duke. Now once we completed this, we then felt that we needed to try to replicate our findings. A lot of this involved large numbers of tests. And of course when you do large numbers of tests, there is a possibility that you can get false positive results. However, as you will see, the relative risks of these were so large that that is not an unlikely possibility.

However, we decided to replicate and we did it in two ways. First, we did a survey in the Dallas VA and found 336 ill veterans just advertising, brought them in and did a confirmatory factor analysis to see if our factor structure occurs in other U.S. veterans.

And then we did a small replication of our main -- what we think is our main finding, perhaps, of evidence of brain injury on MR Spectroscopy, Magnetic Resonance Spectroscopy. We picked a small number of the veterans from this survey and added them to this survey to see if we could replicate the brain injury results.

And finally, we propose now, on the basis of all of this, we now think we now have a very promising hypothesis. In fact, we think at this point had this been a product on the market and we had evidence of this degree, this product would now be off the market. CDC would have gone to FDA, they would have showed them results, the product would have been off the market.

Now let me make one more comment about this sort of duality between chronic disease epidemiology and epidemic investigation. In my early career at CDC in the hospital infections program, we had a long-standing conflict with the FDA.

Early on in the seventies, the FDA would not accept a case control study as evidence of a product injury. That is, statistical associations do not prove causality, and to the extent that they absolutely refuse to take toxic products and injurious products, infected products off the market or issue recalls, based on epidemiologic evidence alone.

Between 1972 and '76 or so we won that battle. We finally convinced FDA, with epidemic after epidemic, that this technique, because of the nature of epidemics, is really a powerful technique and does not often, if ever, if you get strong relative risks, does not really lead you to false positive results. It may lead you to false negative results, but usually not false positive.

Anyway, since we have done all of this work in a group of Seabees and a group of Dallas area veterans, some corroboration, we felt what is really needed next is a survey where we pick a random sample of the veterans who went over, about 3,500, and a random sample of veterans who didn't go over to the War, who were in the military at the same time, carefully -- the populations carefully adjusted so that they are comparable, eliminating what I have termed the healthy warrior effect. Some of you have heard about that literature.

And do a survey where we issue our same epidemic -- our same questionnaires that we used on our epidemic investigation, but in a population survey. So we can then measure the frequency or the prevalence of these syndromes, then pick random samples, small random samples of the sick and the well and both populations, and bring them to Dallas, repeat just those key imaging and genetic technologies that seem to work and see if it holds together.

If it doesn't, obviously it will fall apart here. But given the corroboration we have already got, we think it is a very high payoff. Okay. Well, that is the road map. That is where we are going.

Now let me digress a minute and talk a little about a condition which I am sure is known to some of our colleagues here, but is not well known to physicians in the United States. When we first started doing this, Tom Kurt, my toxicology colleague, came up with this while we were sitting at the NIH conference in April of 1994.

He was sitting in the audience and he looked at the risk factors that were being discussed and the disease manifestations and he punched me in the ribs and said, "Haley, I know what this is. This is OPIDN." And as an internist, a very erudite internist I thought, I said, "OP what? What are you talking about? What is this OP?" It took me three days to learn the acronym.

Organophosphate induced delayed neurotoxicity. And we have in the room one of the world's experts on this, Dr. Spencer, in the corner over there. He has written much of the basic literature. Basically an organophosphate has this generic structure. It has a phosphorous sort of a centerpiece, a phosphate with a -- with variable organic components here and a variable active component here.

For example, in sarin nerve agent and soman this is a fluoride here. But this is a very toxic molecule. This encompasses -- this, in similar structures, encompass most of the pesticides that we use in civilian life, as well as the nerve agents that are used in warfare. Basically a chemical nerve agent is nothing more than a pesticide for humans. It is just much more lethal, much more tenacious than we use commercially.

Now let me make this real with a case report. And those of you have my little grand rounds presentation I did at Parkin, this was the little case report. This actual case report came to me at the lunch table at the faculty club one day as I was getting ready to give grand rounds. One of my colleagues had just seen this lady.

This is a 66 year old north Dallas woman who presented with leg weakness, weakness in her legs, aching all over her body, particularly severe pain in her shoulders and hips, and burnings -- a tingling sensation in her extremities. Severe fatigue. And the weakness then progressed up her legs to her thighs, but there were no other objective neurological signs. No reflex changes really, no other neurological problems.

The physician could not make a diagnosis. He said chronic fatigue syndrome. He thought of a psychiatric problem. She gave a history, however, of her house being sprayed by an exterminator a few days to a week or so before the onset of her symptoms. The physician then called the exterminator and found out that there were two organophosphate pesticides, but the exterminator would not reveal what they were and shortly after that went out of business and left town.

The lady said that she often went barefoot in her home and thinks she probably did it around the time they sprayed her house. The fatigue and weakness lasted about six months and gradually then melted away and she got well. This, I think, is a classic case of OPIDN, organophosphate induced delayed neurotoxicity.

Now this is a well recognized medical condition. It was first -- really came to national -- world attention during a huge epidemic in the 1930s due to inadvertent consuming of an industrial chemical called TOCP, tri-ortho-cresyl phosphate.

That is irrelevant except here is what happened. This ultimately became known as Jake Paralysis or Ginger Jake Toxicity, 1930. In February 1930, suddenly a paralytic illness struck people throughout the southern United States.

And it was very geographically oriented, the southern and western part. Over 5,000 cases were documented in three months and an estimated 50,000 cases occurred in 17 states. One out of every 500 people living in the southern United States at that time was affected to some degree.

The onset of this -- and this is really important to understand. Some of the cases gave a very murky picture of an acute symptom, but in most people, this was inapparent or not present. Then there was a delay of, on average, six to seven days when the occurrence of leg weakness occurred -- began, which spread up the legs, up the arms, nobody died. But it presented with an initial flaccid paralysis indicating a peripheral nerve injury.

As the peripheral nerve injury healed over about a year, the paralysis regressed in the opposite fashion in which it came leaving in its place a spastic paralysis. This absolutely has been forgotten by everyone who knows this -- has heard about this. But the spastic paralysis indicates that there was a spinal cord injury, a central nervous system injury as well.

Many of these people also complained of fatigue that may have lasted the rest of their lives, suggesting a brain central nervous system thing. But notice the delay of several days to several weeks. In fact, the delay now in various studies has been documented up to six weeks after the exposure.

So you can have an exposure to an organophosphate and a delay of weeks before you get the onset of symptoms. And that is very thoroughly documented. Now it characteristically occurs within seven to ten days.

This outbreak was due to drinking TOCP in a drug store alcohol containing medicine called Extracted Jamaica Ginger or Ginger Jake, which during prohibition -- this was during prohibition -- was a common remedy for what ails you. It was one of the legal forms of alcohol at the time.

Now the injury, it was well studied back in that era. In fact, up through the sixties, it was well documented. This is a photograph of a longitudinal section of a vertical nerve showing typical ballooning of the nerve and loss of the axon -- this is the axon -- ballooning of the axon and interruption of the axon. You see the typical beading appearance, a degeneration of the peripheral nerve.

This is a schematic here showing the typical axon on a nerve, peripheral nerve, ballooning in the first stage, fragmentation and then healing slowly.

Now this is close-up view actually of a nerve showing the ballooning and the injured section and loss of the nerve filament there.

All right. Now that is not what this is. Gulf War veterans don't have OPIDN, a typical form. And it took us a while -- and I think Peter Spencer was maybe one of the first ones that really hit me with this distinction.

Gulf War veterans don't have much, if any, peripheral neuropathy. Perhaps some subtle peripheral neuropathy, but that is not a feature that is important here. TOCP or OPIDN, the original TOCP injury was primarily a peripheral neuropathy with a secondary CNS spastic condition that was unmasked later.

This is a very famous study in this field by Savage. Now there are other studies like this. There are at least half a dozen epidemic investigations of cognitive and other central nervous system damage from people who are exposed to pesticides, particularly seen in agricultural workers who spray pesticides.

In this study, they compared one hundred agricultural workers with accidental organophosphate exposure, pesticide exposure, with a hundred matched controls that they identified from the case registries from the Colorado and Rio Grande Valley in Texas, where they have registries to register injured agricultural workers, matched on age, sex, race, education, and socioeconomic level and did blinded examinations of them.

And what they found out is the cases were significantly more impaired on tests of short-term memory, the Haltstead Reitan battery of neuropsychological tests, which basically is a battery of about 50 tests of brain function.

This is not a psychological test, it is a test of sensory, motor, and all the different cortical functions and sub-cortical functions of the brain that are added up into a test that measures brain dysfunction. This is used by neurologists and neurosurgeons to follow their patients with a brain tumor, for example, to see if surgery helps, if the brain impairment decreases.

They were more impaired also on MMPI measures of psychological disfunction. Many of these people develop depression, obsessive compulsive characteristics, after the exposure of organophosphates. It is standardized self-reports of dysfunction, that is standardized questionnaires of what was wrong with them, and family reports of disfunction.

All of these were statistically much more common in cases than controls suggesting or indicating that the pesticide injury may well have caused this encephalopathy or brain disfunction. And there are other studies showing the same thing.

Here is one by Rosenstock comparing 36 Nicaraguan workers, agricultural workers, each with a single set of OP poisoning with 36 match controls. They did neuropsychological test batteries and tests of verbal attention and visual attention, visual memory, visual motor speech.

In other words, a group of neuropsychological tests looking at brain function and found them all to be more abnormal in the exposed symptomatic than the unexposed, suggesting encephalopathy. Now that is the background. In other words, the summary is, there really is a brain injury that is caused by pesticides.

All right. So now let me talk about our investigation. First of all, let me just tell you what our current provisional conclusions are.

First, let me point out, as I always do -- this is important to point out -- we believe from a very thorough reading of the literature, and we have read the literature specifically to answer this question, we think the decisions made before and during the Gulf War to protect our troops from chemical nerve agents, from insect born diseases occurring in the Gulf and all of the other hazards that our troops are going to face, these decisions that the military command made were medically sound, scientifically sound, given the knowledge of the day, of 1990.

There was no evidence we can find -- in fact, there was substantial evidence in the literature suggesting everything they were doing was a good idea and would have been unconscionable had they not done it.

We think in retrospect, however, evidence has come out after the war suggesting that some of those were a mistake, but not that anybody could have known. So we want to point out that there should no scapegoating, at least as far as the information we have.

Now our first main conclusion, though, is, contrary to the official word, we believe there is a Gulf War Syndrome. And I am going to define syndrome in a little bit and give you a case definition for what this is. But we think it has at least three variants, maybe three different divisions, but maybe the same, but with three different variants.

We think it is due to subtle brain and nerve damage and nerve damage in the central nervous system. Primarily the brain. And I will even say we think now that the damage is in the basal ganglia, possibly the thalamus, and the brain stem. And I will tell you what those are in a minute.

And we think it was caused by exposure to combinations of organophosphate like chemicals. Now not all these are organophosphates technically, but we think they act together in synergistic ways. To wit, sarin nerve agent, and I think it is sarin, not soman or VX or anything else. We think it was probably sarin.

We think pyridostigmine bromide, Mestinon, for physicians who use this to treat Myasthenia Gravis. It was given as an anti-nerve gas, anti-soman agent to increase survival in a soman attack, if one had occurred. DEET, the active ingredient in insect repellents and pesticides that were liberally sprayed in the environment and also were present in flea collars that many of the soldiers wore to, again, to protect themselves from insects.

There is evidence now that giving PB before a soman attack improves survival. That is well shown in animals. Extensive literature on it. However, there is now evidence occurring right after the war that if the PB is given following the exposure to a number of organophosphates, that it actually may promote brain damage.

Now the brain damage and the initial symptoms that can kill you; in other words, the acute immediate symptoms of organophosphate poisoning, and the long-term brain damage, these occur by different pathogenetic mechanisms. Okay?

So in other words, we think the PB -- and we may want to talk about that this afternoon, the different mechanisms of the acute effects of organophosphates and the chronic effects. These are two different mechanisms, another distinction that has really clouded this field.

Now let me reemphasize now the technique we followed. Standard steps in investigating an epidemic of a new disease. CDC tipicalous. It has been done a thousand times. First, examine a few of the people, develop a case definition, survey a population and divide them into cases and controls, verify that the cases have a real illness, then identify factors -- or questions in the cases and the control that differ between the cases and the controls. And that gives you a clue about what the cause is.

You then test that cause for plausibility in animal models. You then develop a screening test, a treatment and preventive measures, which we are going to talk about later, to, in epidemiological terms, to take the pump off the handle, to stop the epidemic if it is still going on or prevent it from occurring in the future. And then you continue surveillance to be sure it is controlled if it is an ongoing problem. That is the CDC formula.

Now is this an epidemic? Is it an epidemic? Some people would disagree over that, perhaps. Here is the epidemic curve. This is a graph showing the number of cases with onset by date by month during the war. The actual war was fought right in here. The number of people who -- by the date of appearance of their first symptom.

Now this doesn't necessarily mean that that was the date their -- whatever this Gulf War Syndrome occurred, it is the date of the first symptom that they attributed to it. And people may have falsely attributed. We don't know. But at least what you see a huge spike followed by a secondary peak and some little squiggles. In other words, within several months after the war, everybody who got sick in the control that we studied, had the onset of their disease.

This is a -- I won't say classic, but this is a -- has a very strong familiarity with the common source epidemic curve. This was distinguished by a bimodal peak and maybe a few little reverberations later. Those of you who are epidemic investigators would know that you look at the shape of the epidemic curve as a real clue to what is going on. And that suggests a common source epidemic with perhaps two or three different interactions of risk factors.

Yes?

DR. CAM: Dr. Haley, you mentioned earlier that PB post-exposure could cause brain damage. Did you really have data as to how many were in that situation and was that integrated into any of your graphs?

DR. HALEY: No. That is in the literature about -- those are basic chemistry studies.

DR. CAM: I see.

DR. HALEY: That show in animal studies that these different chemicals can promote -- they are agents that can promote and can increase the likelihood of occurrence of OPIDN in animals.

DR. CAM: So it would be very difficult to assess such a thing.

DR. HALEY: Well, let's go through it. In one sense yes, in another sense no. Okay. So that is the epidemic curve.

Yes, sir.

RADM STEINMAN: Excuse me. Could you back up, please, one slide. Tell me if I am wrong, but the major peak in appearance of first symptoms significantly predates the shooting.

DR. HALEY: Again, this is the date of onset of the first symptom --

RADM STEINMAN: Right.

DR. HALEY: -- that they had in the war. Yes, this is interesting, isn't it? So the question is, who is there and who is there? And I will come back to address that later.

Okay. Now when you find what you think is an epidemic, a common source epidemic, first examine a few of the ill people and develop a case definition. That is the cardinal rule. And notice if you look at all the other research out there, virtually no one has ventured a case definition. This was the point of the NIH conference in 1994, to come up with a case definition.

Jay Sanford proposed a case definition in 1993 to do studies like this. It was rejected. Over and over there has been a reluctance to come up with a case definition. What we tell our students is, first come up with a case definition. And if you can't, then develop a case definition because failure to do this means you have given up. You are not going to solve the problem. No epidemic will ever be solved without a case definition.

Now you say, and I have put this question to many of my old CDC colleagues over these last several years, "Well, what if you just can't develop one?" And they say, "Then you develop one. You come up with some way to do it." Because without doing that -- the problem is doing comparisons without a case definition, say comparing the deployed population and the non-deployed population. See only a small percentage of those people are sick.

And so when you compare those two, the effects are completely diluted out, but with a case definition, you focus on who is sick and compare them with the people who are not sick and you suddenly get very powerful differences. Well, let me tell you, this is a very important point. This is the key, the solution to this whole problem.

All right. Well, what is this Gulf War Syndrome? Here is the problem that we are trying to solve. We have to come up with a case definition. And the problem is, here is the Gulf War Syndrome. Now think of how many of these have you ever had.

Fatigue, fever, night sweats, muscle aches, joint aches, watery diarrhea, abdominal pain, breathing problems, skin rashes, hearing loss, swollen glands, headaches, memory loss, depression, emotional changes, personality changes, problems sleeping, problems concentrating, problems with thinking, problems speaking, becoming confused or disoriented, tingling, burning, numbness, muscle weakness, loss of balance and coordination and loss of bowel and bladder control, sexual impotence, chemical sensitivities, and panic attacks.

Now you know the average doctors, almost every doctor, sees people like -- that have multiple ones of these frequently, and frankly, what you do is you say -- in fact when I was an intern, you called that a crock. A person who is making up symptoms who we will never find a cause and it is a waste of time to deal with them. And doctors react to this like that, and that is because we usually never find anything and we can't do much about it. So it is very frustrating to doctors.

Well, how did we come up with a case definition. What we did was, as we usually do in an epidemic investigation, we found a group that -- we went out and found them. We didn't look for people to come to us. That is another key factor in epidemic investigations. You don't go find people that come to you because you get a strange mix. You go out and find a group and then get as many as you can in this group and get them together to do the investigation.

We picked a group called the 24th Reserve Naval Mobile Construction Battalion or Seabees. We picked them for several reasons. First, Jay Sanford had given me some reports produced by a Navy epidemiology group, Bill Berg was the head of that group. It was sent out initially in 1992, shortly after all this was recognized, and contacted a number of units where there were reports of a lot of this illness going on. This, he felt, was a typical one.

In this, he catalogued all of the symptoms -- it was a small number of these people, not as many as we looked at. I think there were only about 30 or 40 looked at. But what he found is these people have the Gulf War Syndrome that they were seeing elsewhere.

So it looked like a unit that seemed to be typical. Not necessarily representative, not statistically representative in terms of a random sample, but it looked like they had what everybody else had.

Also, it was a construction battalion and a Seabees battalion. They go everywhere all over the battlefield. They are not in one little location, boom you are there and the war is over. They are going and constructing things, building roads and bridges, and moving everywhere. So if there was a geographical risk, some of these guys would have been exposed.

Also, it was the only reserve Seabees unit and the reserves come back to their own hometown so you can find them whereas active duty people, they are spread out and you can't get them. December '94 and January '95 we went into the field, went out to the southeastern United States where they live and surveyed them.

We certainly -- we intended to include all Persian Gulf veterans in the 24th Reserve Battalion with addresses in those five states, which encompassed 98 percent of them. Alabama, Kentucky, North Carolina, South Carolina, and Georgia. There was 606 in all. There were 130 more attached to that unit, but they were from other units that were attached at the last minute. We didn't include them because they were elsewhere in the country.

We included both ill and well and we wrote letters to the entire battalion, many of which were returned undelivered, but we actually did have a roster and did cold calling trying to get everybody from one of these five sights out at their home locales.

Now whenever you do an investigation like this, you never get everybody in a -- you never get a representative population in an epidemic investigation. I guess you could, but you never do. We got 249 of the 606 to show up, which is 41 percent. Now many people feel that if you don't get 80 percent, you quit because you have got a -- you necessarily have a selection bias that would somehow invalidate your results. And it can happen.

So what you -- but you don't give up because there is an epidemic going on. What you do is you look at the characteristics of your sample and see if it is balanced enough to give you an answer. Most importantly, we felt, was the mix of people who were still in the unit and the people who had left because if you only look at people who are still in the service four years later, that is a fail selection bias because the sickest people have left. Okay. So we had a good balance there.

Looking at the balance on age, mean age is about the same; gender, they were all men in the Seabees unit in that era. The racial distribution of the participants and the non-participants was the same. The years of education were the same. Percent still in the reserve the same. Wartime rank is the same. We got this information from the in theater roster and we did a phone call survey of a small random sample of the non-participants to estimate these differences.

This was their wartime job rating. Remember it was the Seabees in construction battalion. So they had construction job duties. As you see, the percent distributions are approximately the same. The major difference was in the officers. Almost none of the officers participated. Apparently there was pressure from the command structure to have them not participate. The officers obeyed, but the enlisted didn't.

Now the one big difference is, and many people have suggested this is a flaw. Actually for a case control study, it is an advantage. But the follow-up survey idea, we had asked is there a serious -- have you had any kind of a serious health problem since the war. Now this isn't the case definition, just any kind of serious health problem.

Seventy percent of the participants said they had some serious health problem; forty-three percent of the non-participants. So you see, the ones who participated were bias towards having more of the sick people coming, whereas the non-participants tended to be more the non-sick.

Now if you are trying to estimate the prevalence of a disease, that is probably going to be a selection bias and you are going to therefore overestimate the prevalence of whatever health condition you are dealing with.

However, if you are trying to get a population to do a case control study, to get the cases so you can then pick some good well matched controls, what you want is to get all the sick people. You want to get some well people too, but you want to get as many of the sick people there so that you can pick good cases. A good representative of good typical cases.

We designed two instruments to go into this survey. One was a questionnaire booklet on exposures. This catalog and a series of questions in a self-administered format, questions that asked about the exposure that had been identified by the various expert committees that had reviewed this over the first several years after the war. Institute of Medicine Committees, Defense Science Board, NIH, Consensus Panel and others. They all came out with basically the same list.

So we incorporated those into a questionnaire with a couple of additions and we developed a symptom questionnaire booklet. I am going to go over these and show you how they were designed because the design of these particular system booklets are fairly unique and I think accounts for a lot -- for the fact that we were able to focus on this thing.

We also gave a personality assessment inventory. This is the PAI, similar to the MMPI, but we felt had better characteristics for this field survey. It was shorter, more interpretable, looking for psychological issues because as you know, that was the main hypothesis to begin with. We pretested this twice in Persian Gulf veterans to get the lingo, the military lingo right. And it went right into the field.

Here is the exposure questionnaire, the symptom questionnaire, and the standardized psychological inventory. Notice we had an assurance of confidentiality on the front. We felt this was important. In this era there was great fear among particularly the ones who were still in the unit to if they participated, one, they might lose their military status -- in fact, there were actually threats at this time, which was a very emotional situation, particularly for the ones still on active duty.

There were actually personal threats to some of these people in the small Alabama towns of the -- you know, the officer who was in charge of them saying, "I heard you went to the survey. You better watch out." That kind of thing.

So we gave them a very strong assurance of confidentiality we wouldn't release this to their military units or to the press or anyone. And this was signed by myself, as well as by Ross Perot. And we felt Ross Perot's name would be good as gold on that.

It is interesting, when we sent out a letter, I had Ross Perot write a letter to all these people to come to the survey. And then I got a frantic phone call from one of the leaders, a guy saying, "Gee, some of the guys aren't going to come because they don't like Ross Perot. They don't like his politics." So I quickly sent them all a letter and got it immediately turned around. It is interesting how those things work.

Now here is the example of a question on the exposure questionnaire. Okay. During Operation Desert Shield Desert Storm, our service personnel used many different types of insect repellent to avoid insects. Here are the instructions or questions. "Did you use Avon Skin So Soft. Yes or no." You know, this is the thing that contains no active DEET. It is toilet water that does repel insects to some extent.

The commercial varieties of OFF! was common in the Gulf. This is -- at that time was less than 30 percent DEET in an emollient that was not well percutaneously absorbed. And the military issued repellent. One of the main products used, particularly in this unit, was a 75 percent DEET and ethanol. It was the old Vietnam thing in a little gray squeeze bottle. Both DEET and ethanol are percutaneous absorption enhancers.

Those of you who know DMSO is. You know, you mix a drug with it, you put it on your knee, and it sucks the drug right through the skin into the subcutaneous, the DEET structures. So this would have been highly absorbed into the bloodstream.

And then others and then estimates, their semi-quantitative estimates of number of days. We actually gave them a calendar of the period of the war so they could -- and had them mark the day they arrived, the day they left.

And we had the beginning of the air war and beginning of the ground war, all the major events on there so they could then get -- remind themself, put themselves back in -- made them fill it out and do some things so they could then estimate and give us a semi-quantitative estimate.

Now we also gave them this map. This shows the Persian Gulf, little Kuwait here, Saudi Arabia and Iraq. Our troops during the bombing were here and down the coast. This is a grid map. And then on the facing page, we asked them to tell us the sector they -- all the sectors of that map they visited during the war and the approximate date they arrived and left that area so we could then computerize this and play and computer game, could we find a place and a date that coincides with disease suggesting an environmental exposure.

Now this is from the symptom questionnaire. Now in most health studies, what you do is you give a list of symptoms and then you say, "Which of these do you have? Yes/no, yes/no, yes/no." And sometimes we will see people just zip right down the yes column and right down the no column, and that is not good form.

We actually had a page for each symptom. We asked them first very clearly what this was. "In the past five years since the war have you experienced tingling, burning, or stinging pain in any part of your body lasting all day and continuing for at least a month. Do not count things that come and go quickly and are not present continuously."

Those of you know -- who interview people about tingling and numbness, you know everybody has a little thing here and there and everybody will say yes. So you have to really be more specific. If they say yes, then they have to go down here and tell us what part of their body is that involved in. So each part of the body and then yes or no I have -- this has been a problem. Approximately which month and day this began and is it still a problem.

So they had to go to a lot of trouble to answer, "Yes, I have this symptom." Also they gave us a lot of interesting follow-up information that we could analyze. Now one of the fundamental things that I think has come out of our investigation is in these vague, confusing illnesses like chronic fatigue syndrome, multiple chemical sensitivity, and Gulf War Syndrome, maybe breast implant disease, I don't know, one of the things we noticed is these symptoms that people use to describe their illness are highly ambiguous.

For example, tingling and numbness may occur in the extremities, the hands and feet. That might signal a sensory peripheral neuropathy. That would be a neurological thing perhaps. Also you might have tingling and numbness around the face, tongue, and lips. That could mean a hyperventilation episode or it could be some other neuropathy, a different kind of brain nerve neuropathy or something.

You could have tingling and numbness in your trunk and groin. I don't know what that is. That is very strange. That might be a central sensory problem of some kind. I don't know. But those distributions have different pathophysiologic explanations. And therefore, you need to distinguish among those, rather than just say I have tingling and numbness because then you would lump all three of those groups together.

So how do you distinguish those? Well, what we did was I took this and for every one of 22 questions, we had these follow-up items to allow us to disentangle the ambiguous components of them. Well, what we did is a factor analysis. The factor analysis, for those who don't do it all the time, is a black box statistical technique. It is actually a very useful data reduction technique.

If you look in the good statistics books, it is listed as a data reduction technique. And that is what we used it for to try to see if there are different components here. Are there three different factors here. Okay. I will describe that now.

The first order of business now, given that data set, is to define is there a case definition. Can we come up with a case definition that is plausible. Searching for syndromes. All right. First of all, what is a syndrome? What is a Gulf War Syndrome. What is any kind of a syndrome?

Well, Dorlands Medical Dictionary defines it as a set of symptoms that occur together. Okay. In addition, it sort of implies a common pathophysiology, although that is not necessarily true. Some syndromes have different -- have several different ideologies. But it is fundamentally a group of symptoms that occur together.

Now in this population where we have the symptoms measured, it is a simple mathematical question, is there a group of symptoms that hang together in a group of people. Well, here is basically the strategy we used to explore to that.

We used what is called -- and the statisticians will take on this more than others, but we used the thing called hierarchical exploratory factor analysis. Some would call it a two-stage, but technically it is a hierarchical factor analysis, an exploratory factor analysis.

We used the method of principal factor analysis and rotate it with very matched rotation, which is orthogonal rotation, so as not to assume some correlation among the factors. There is correlation among the factors, as we can talk about later. But first, we didn't assume that.

The level one of this two hierarchy factor analysis, we first disentangle the ambiguity of syndromes. So the follow-up questions for each of the 22 symptoms, for each of those we did a separate factor analysis to come up with two or three factor scales that we felt would be sort of unambiguous symptom factors gains, unambiguous symptom measures.

The are factor analysis of these follow-up descriptive features on each of the 22 ambiguous symptoms to produce 52 what we thing unambiguous symptom factors or symptom factors scales. So for each individuals, each of the 249 people we started out with 22 symptoms and just used the yes or no question, "Do you have tingling and burning and numbness," and we turned that into 52 symptom factor scales.

Now the symptoms were dichotomous, yes/no, do you have this yes or no. These are normally distributed factors, factor scales.

Yes, sir.

DR. ZEGER: I just have a question for clarification.

DR. HALEY: Sure.

DR. ZEGER: If somebody doesn't have a syndrome, so let's suppose only 50 percent of the people have a particular symptom, the first stage of exploratory factor analysis, you used only those people who reported the symptom?

DR. HALEY: No, we used everybody.

DR. ZEGER: So if somebody didn't report the symptom they had, then no for all of the --

DR. HALEY: That is right.

DR. ZEGER: -- subquestions.

DR. HALEY: That is correct.

DR. ZEGER: And then they went into a factor analysis in the exploratory first phase.

DR. HALEY: That is right. And then what we did we then took a factor of weights for this and produced a factor scale for each individual. We actually wrestled with that question, should we exclude the people. And we got a lot of opinions about that and it was pretty much divided. And we decided to leave it at that.

DR. ZEGER: So that in the end, the unambiguous factor score that you have would actually have a sort of a fraction, say 50 percent of the people with a --

DR. HALEY: Yes. Right. Right.

DR. ZEGER: Okay. Thank you.

DR. HALEY: Yes. Now the level two factor analysis, we input into the second level these 52 symptom factor scales and inputting into the second. We did a factor analysis of those using the same factor technique and we came up with six symptom factors or what we call syndrome factors.

Now let me say we didn't just put a black box. This took about four months of analyzing this data. This was a very, very long, tedious process because what we wanted was six -- we were staking our entire investigation on this. So we didn't just put it in a black box and churn it out in a couple of days, this was about four months.

What we did, we wanted to find syndrome factors where the loadings, and I will show you those in a minute, the loadings on those things, in other words, which symptoms contribute to which syndrome would be clinically plausible. In other words, we wanted to have a model where the things look like it might be something plausible.

So the final syndrome factor selection -- and let me say you can select a whole bunch of different models in something like this. There is no statistical rule that tells you this one is better than that one. So what we did is we had a group of clinicians that we looked at this and said, "Okay. Now how about a three factor model? How about a four factor model?"

Now there are some stopping rules and things, but ultimately you have got to have something that looks clinically plausible where you are lumping together things that could meaningfully go together. And we had a sixth syndrome factor model that looked really interesting clinically and we picked it.

Now remember what my colleagues at CDC say. If you can't develop a case definition, well, somehow you do it. You just come up with one because to fail to do that, you are not in the game. You have not started yet. This is a bit of a leap of faith, but you have got to start somewhere. So that was it.

So from this, we then calculated six syndrome factor scales, which are normally distributed things, all 52 factors for a given person. We then weighed all 52 of these symptom scales by the factor weights and come up with a score. A person has six syndrome factor scores now.

And then we dichotomized those at a convenient point, which I will show you in a minute, to come out with dichotomous or binary measures. So, "Do you have Syndrome 1, yes or no." "Do you have Syndrome 2, yes or no," et cetera.

Now let me just show you the first phase, the first hierarchial level. This is an unrotated factor analysis on this tingling and numbness thing. And you see I have simplified this down somewhat. So notice there are three symptoms that seem to go together here and four symptoms that seem to go together there.

So it looks like there may be two factors, but they are kind off the diagonals here. So what you do with rotation, you basically rotate the axis until these are as close together -- as close to the axis as possible. And then you are sort of measuring those more accurately.

For those who want to know what rotation is, that is basically it. And that is the rotation we had on this tingling and numbness in a simplified fashion. This actually came out with three factors. I am going to show two since it is easier.

So here is the rotated factor space. Here are the most important of those body places where they complained of tingling and numbness. Notice arms, feet, hands, and calves have very high correlations with factor one.

So we called this factor, this scale that was going to be produced, we called this the tingling and numbness of the extremities factor. Okay? And then this one, notice face, tongue, and lips is highly correlated here, but notice these aren't and see these are not correlated with this one, but these are correlated with this one.

So factor two we call tingling and numbness of the face, tongue, and lips. Possibly the psychological thing or the perennial nerve thing. So we have now two different scales, continuous measures, of these two factors. And we think that was a key thing in the fact that we found something later.

Now we put those measured into the second analysis, which is a -- the second phase, the same kind of factor analysis. This shows the rotated scree plot, if you will. This measures the strength of clustering, if you will. That is not exactly true, but I think you can picture that better.

How tightly are these -- this one group of guys that has this set of symptoms, how much did they resemble each other and how are they -- how much are they different from everybody else. The strength of clustering, if you will.

Here are the possible syndromes or the factors that were identified by the program from the highest, strongest cluster all the way down to the weakest. And then you draw a line here. This is at an igon value of about one. Here I am using percent of variants explained and rotated one. That is easier to explain.

But what we found was three highly clustered syndrome factors or three strong syndrome factors with high igon values or high clustering coefficients. Then we found three more -- four, five, and six -- that had lower levels. And then after that, it is below the level, just this garbage down here. We cut it off there.

And see, we could have cut it off after three and then recompiled this and did a three factor model. What we found was, though, separating these out, these used more clinically interesting, more clinically plausible. Where if you leave these in, these then jump in there and sort of mess up the whole thing.

So we felt a six factor model looks best, but the three -- the first three looked the most interesting. Now let me just make it short to cut to the chase here. Four, five, and six, when you dichotomize them, the people above the -- the ones with the extreme value, the most strongly clustered of these, also clustered on number 2. These appear to be some kind of variant or echo of number 2, which also turned out to be the most serious clinically.

Exactly what that means, you know, statistically and so forth, I don't think anybody could tell you. It is just this is a way of reducing very complex, large amounts of data to a small number of things that you might be able to work with. And the measure of it -- there are statistical measures here, but ultimately the measure of whether it works or not is does it discriminate clinically meaningful parameters later on.

All right. Now there are some statistical validations, which are in the paper. One, there was a large range of the commonality estimates. Now this is only speaking to the statisticians. The commonality estimates, the diagonal estimates on the diagonal went from .27 to .78, which justifies principal factor analysis, rather than principal components analysis. Most people just take principal components because that is what everybody does, but we think this one is more appropriate.

There is a measure, the Kayser's measure of sampling adequacy is very high. This means that our final factor model was a good fit to this common factor model. So it looks like it is a good factor analysis. It has high internal consistency of the syndrome factor scales, Cronbach alpha -- that is alpha there, it should be, is bigger than .8 -- and we replicated the two halves.

That was a bit of a hollow victory in that we had developed it on the whole sample, and then we went back and looked at split halves, which isn't exactly what you would do. Normally you would develop it on one half and replicate it on the other half, but we only had 249.

But we did go back and look at split halves and we looked at the correlations of the -- they were really correlated for a one and moderately correlated for two. And those all correlated for four, five, and six as you would get from the other.

So a statistical limitation, but this just shows the modeling kind of worked, but it does make this meaningful and the dichotomization here are the six symptoms syndrome factor scales. This is in the paper. Okay. In other words, all 249 people are in one of these, if that makes sense.

What you want to do is to come up with some level of dissent of these people in Syndrome 1. Now obviously there are people in the middle in this epidemiological study and there are those that were dichotomized by this dichotomization in different places. There is the distribution sort of constraint. And that was one alternative.

But we decided to just take an arbitrary one that would give us enough of each of these, feeling that we wanted the most typical, the most typical cluster in cases. Then in the controls it wouldn't really matter because there were so many of them, that would be a fairly large number of operator error.

Now there has also been a real misunderstanding of the -- the thing that really hit us all, the fact that when you pick the extremes, you pick the most extreme cases, the sickest people. That is not right. Thing of what we have got here. This is a factor scale. So the extreme up here is not the severity of illness, it is the legality of the clustering.

So we are measuring out whether it was clustered around this factor to see if it was it was internal or a construct. The higher you are, the more you have got everything, the more typical you are. So what are the positive advantages of this.

In the cases, the most difficult cases in a CDC epidemic, what you do is you go ahead and there would be some typical guess and you would just pick them up. Okay. He has got this. Well, let's pick you, you, you. That is what you do on this old toxic shock syndrome. Okay. So that is how we show the people above the line. Cases and controls on each of those six syndromes. Oh, shoot. I think that is worth backing up.

Now let's see if that works. No. Close. That always happens. Oh, okay. All right. This is the arithmetic in our 249. There were 606 in the battalion. Okay? We surveyed 249 that showed up. Remember we encouraged the sick and the well.

So we had 70 well people, no health problems at all. A hundred and sixteen -- now this is interesting -- a hundred and sixteen that said they had health problems, but no syndromes, they didn't cluster in the syndromes. And we had 50 that had Syndromes 1, 2, and 3.

Now there were also 13 -- 16 more -- let's see, 13 more that had syndromes -- only Syndromes 4, 5, and 6, but Syndrome 2 was right below the cutoff on those. So they were really syndromes too, but they are not reflected in here. They are considered negative here.

Now of these 50, what is the prevalence of this syndrome, these syndromes in the battalion? Well, we don't know because remember we have a biased sample. So what we do to get a conservative estimate, we divide 50 by 606 to get a low estimate of 8.3. So this is at least a prevalence of 8.3 in the population and probably somewhat higher, but we can't tell because we don't have the rest of those.

And that is in the paper. All of these things are in the paper. All right. Now let's -- let me give you a feeling of what these syndromes are by showing you what the symptoms were that loaded heavily on each of these three Syndromes, 1, 2, and 3.

Syndrome 1 we gave just gave the name impaired cognition, as you do in factor analysis. Impaired cognition. These people had distractibility, memory problems. What they call memory problems, and later we will talk about these, actually are not memory problems, they are concentration problems, but that is an interesting -- but they call it memory problems.

Depression, middle and terminal insomnia, which often goes along with depression. Fatigue. And here the factor analysis of the symptoms made a huge difference. Actually it made a big difference in all of these, but particularly this is the most noteworthy because of all the chronic fatigue syndrome literature.

Here, but loaded heavily in this syndrome was fatigue, meaning sleepiness during the day, but not body weakness. The body weakness thing came on -- loaded strongly on Syndrome 3. So you see, we think these distinctions on symptoms are very important for converting chronic fatigue syndrome into a clinical entity that we can study.

Slurring of speech. A lot of Gulf War veterans, they don't complain of this, but if you ask them, "Yes, my speech has changed. I don't talk the same way." Confused thought and severe migraine headaches. So just impaired cognition we call it, but it is obviously a lot more, as you can see.

Syndrome 2, confusion ataxia. This is the bad one. Thinking and reasoning problems, getting confused or lost. The wife will tell you, "I can't send John into town anymore. The police will bring him back." It is sort of like an Alzheimer's patient, some of these.

Getting disoriented, losing their balance, stumbling off and feeling like the room is spinning. We think now this is a vestibular ataxia. Not dizziness in any kind of just trivial sense. We think this is a real vestibular ataxia.

Sexual impotence. And then what is interesting, what also came in here was a physician's diagnosis of PTSD. I have gone back and studied each of these. What happens, they go to their personal physician or VA doctor and the doctors will say, "There is something really wrong with you. I think you are really sick, but I can't figure out what it is. So I think I am going to give you the diagnosis of PTSD." For one thing, PTSD gets some service connection.

Now they didn't do that for Syndrome 1 and they didn't do it for Syndrome 3, but they did it for Syndrome 2. So there is something about PTSD and Syndrome 2 that is really a tough -- a close connection. Now whether they meet the criteria for PTSD is another matter, and they don't. But they have got the diagnosis. And a lot of them had depression too. About two-thirds of them had some type of degree of depression.

Now the third one, arthro-myo-neuropathy, we gave it that name, you might call this fibromyalgia because I think that is probably the same thing. Generalized joint aches and muscle soreness, increased difficulty lifting heavy objects, fatigue, but fatigue meaning loss of muscle strength, not sleepiness during the day. And tingling and numbness of the extremities, but not the trunk and groin and not the face, tongue, and lips.

Notice this joint ache -- let me tell you about the joint aches because it is really interesting. This is not arthritis. You might call it arthralgia, pain in the joints, but it is not inflammation. In fact, almost all Gulf War veterans in any study we have seen, including ours, all their rheumatic workup, their sed rates -- the sed rates are all in a very low range. Nobody has got an elevated sed rate. A few of them have a positive ANA, but it is very equivocal.

In fact, that is an interesting subject. But there is no inflammation, there is no joint deformity. In most of them, there is no limited range of motion and there is not even a difference in the pain on motion. This is really atypical for arthritis. It is not arthritis, it is central pain.

What they tell you is, they have a toothache all over their body, particularly in their shoulders and hips. Thirty year old men with a toothache constantly boring. And that is what they mean by generalized joint aches and muscle soreness, at least in our battalion and in the Dallas veterans that we looked at now.

Okay. So that is number three. All right. Now are these syndromes different or are they really just the same thing, a statistical misadventure here. We looked at unemployment. What percent are unable to work. This is the 70 who are well and they have an unemployment rate of about 2 percent, 3 percent, lower than the national average. They are all working veterans. Those who have illness, but no syndromes, about the same.

Syndromes 1 and 3 have slightly higher rates of unemployment, but not really significantly different. But Syndrome 2, the bad one, look at this. Unemployment rate of 50 percent. And if you talk to these other guys who are employed, many of them will tell you, "I am still working, but they had to change my job after the war. I was the shop foreman and now I am working in the mail room because the company wanted to keep me employed because I am a war hero."

So these people are having a real hard time. These are functioning along, but are just very uncomfortable. And these seem to be -- maybe there are a few of some of these in here, but by and large, these don't appear to be ill.

It is interesting now -- I am sorry ADM Steinman left, somebody fill him in when he left -- there is an excess of these in that first hump of the epidemic curve. And these are primarily in the second hump of the epidemic curve.

Psychological tests. Remember we started out looking at psychological issues assuring Ross Perot that if that was there, we would find it. This is the PAI results and these are the psychological scales, you know, the somatic complaint scalings, anxiety, depression, anxiety related disorders, which is where PTSD is.

This pattern was found in every one of the syndromes. And that was a surprise. I was expecting to see each of our syndromes have a different psychological profile, one of them being depression, one of them being PTSD, one of them being a somatic form of disorder, whatever that is. It wasn't. This was none of those. This is clearly not the profile of PTSD, it is not the profile of hysteria or somatic form of disorder, it is not a profile of any psychiatric disease.

Now what we do -- that doesn't really say it is not for sure. It says it is unlikely that it is a psychiatric -- typical psychiatric disease. What this is, this is the pattern you see on this test in people who have -- who are in a neurology clinic. You give this to people who have a brain injury. Now that doesn't prove it is a brain injury, but it suggests that it is probably not PTSD or another psychiatric illness.

Okay. Now that was our case definition. The next step was to then do a nested case control study, pick some of the sick, some of the well, some of the people in these syndromes, some of the well people, bring them to Dallas and see if they have a real organic illness or is this just psychological and we are being fooled in some way. So a clinical case control study.

We picked 26 of the guys with the syndromes and 20 controls, age, sex matched. Notice here is where we get the real strength of our design. What we want is typical cases and some well matched controls. They are matched for age, sex, education levels And we matched the controls to Syndrome 2 and we oversampled Syndrome 2.

What we did is we got five Syndrome 1s, thirteen Syndrome 2s, and five Syndrome 3s because we really wanted to nail down Syndrome 2s because that is the bad disease. Okay. So actually I have noted here, I didn't show the 4, 5, and 6 because we just got one each of those just to kind of hedge our bets and see what might be there. But we weren't really interested in that at the time.

And here are the controls, 20 matched controls. Now the controls were actually two subgroups. Ten of them were men who served in the war in this battalion, but remained well, and ten were people who were in this battalion, but didn't go over, had not gone over to the war yet and the war ended before they could get over or they were logistical back at home or whatever. One of them had an appendicitis and couldn't get over, et cetera.

So we had a deployed subgroup and non-deployed. We did that because, hedging our bets again, because afraid that there might be subtle brain damage in everybody who went over there. I mean, who knows? And maybe our case control parameters wouldn't be different between cases in control. So we needed a group that had stayed home. It turned out the controls were all homogeneous. Both groups were the same on everything. So we now lumped them all together to show it.

This is the Halstead Pyramid index, the sum up of all of the neurological -- 50 or so neurological tests that measure different aspects of brain function. The controls here, as you see, are in the normal range, below .4, this goes from zero to one, and the higher it is the more likely there is brain impairment.

Syndrome 1 is slightly -- the main is slightly above, but there is a large variance and it is not significantly different. But look at Syndrome 2 and 3 are highly elevated, suggesting brain impairment, and they are significantly different from the controls.

Here is the general neuropsychological deficit scale, a different summary scale of neuropsychological results and you get a similar finding. They have the trailmaking test part B. Syndrome 3 looked worse on that. This is all in the paper. It is in the tables in the second paper.

Here is another test we did. We did a number of these tests. And looking for subtle parameters of brain dysfunction in people who have normal neurological exams -- I should have mentioned that at the beginning. All of these people when they brought them in, we did a thorough neurological exam by a neurologist and we did brain MRIs. All of that was normal.

So in a situation like that, you need to look for tests that look more sensitively, particularly at lower deeper brain structures. So we looked -- we did primarily all your vestibular tests, evoked potentials and that sort of thing, which is in the literature of neurotoxicity.

So here we show the velocity of eye movement from a reflex elicited by choleric stimulation in the ears. The eye movements measured by electromastagmography. It is basically an electronic way of measuring how fast the eyes vibrate when you put cold or warm air in the ears. It is measured on a computer and you can calculate the velocity.

Here is the velocity and the controls, cool air in the right ear, cool air in the left, warm in the right, warm in the left. And you see the controls are all up here pretty close to the same thing with a normal velocity of eye motion.

Syndromes 3s have a dramatic reduction in eye motion indicating some interruption or impairment of the vestibula ocular reflex, which goes from the eighth nerve -- acoustic vestibular branch of the eighth nerve, up the brain stem and out the ocular motor nerve. Somewhere along there, this is impaired. Not interrupted, but impaired because the eye does move.

Syndromes 1 and 2 are intermediate. Not individually statistically significant, but there appears to be an ordering and the ordering is highly statistical significantly.

Here is brain stem auditory evoked response. If you stimulate the ear with clicks, you get a different pathway up the brain stem and you can measure three spikes as this thing goes up the brain stem, and you can measure how long it takes the impulse to go up the brain stem and then compare the right and the left. This is a very sensitive measure because you are comparing the person's -- their own control. And you look at the difference between the speed on the right and left sides of the brain stem, here the controls are in the normal, but very little difference between right and left. Syndromes 1 and 2 have a big difference between -- actually, a larger difference.

This isn't a big difference, it is just a small difference, but a significantly different impulse. Syndrome 3 is normal. You will notice what we are seeing is that these syndromes has some difference between the controls, but the difference in terms of different patterns, which we felt went along with the fact that they have different symptoms, they have different neurological findings.

Of course, the opinion could be maybe this is all random. This is all random, this is a form of different patterns, but the controls are never higher than the cases and you can't explain that randomness. In other words, multiple testing can't explain this and the P value, looking at that, is less than the 0001, given the controls are always better off than the cases in all the tests.

So what we think, there appears to be at least something going on here in the brain stem or in -- you can't see it here, lateral to this, that is the basal ganglia, a very sensitive -- heavily controlled with influence by the basal ganglia. So that the lesion could be in either place. So that began our working hypothesis based on these preliminary exploratory neurological tests in the first round.

Now let me just show you the risk factors and then we will break. We now have a case definition, which we think looks plausible. We now want to know, okay if this is true, was it related to this disease. So we got back to our exposure question as to we had picked of the 249, we had lots of difference between the 63 or in syndrome and the ones who don't have the syndrome epidemiologic analysis.

Here is the list of risk factors that come up with all those committees. Chemical weapons, biological weapons, ciprofloxacin and chloroquin for malaria prophylactics, environmental pesticides, pesticides sprayed on uniforms, high potency DEET insect repellents, pesticides in flea collars that they wore around their necks.

Depleted uranium munitions, smoke exposure from the environment, burning jet fuel in tents, people coming with possible carbon monoxide poisoning or -- which also affects the basal ganglia by the way, and lead poisoning, they used leaded fuel, petroleum in drinking water, chemical absorption with a coating of paint. This can produce a pulmonary syndrome if you don't use protective gear. Inborn errors of metabolism, and this one was commonly mentioned back then. Stress, smoking, alcohol, cocaine use. This is a sort of a list.

When you are doing an epidemiological study, generally an epidemic investigation -- again, I am going to -- with four, five and six, you don't get that because of recall bias. In my liberal review of literature, I don't see a recall bias except for the Love Canal like studies which are very different from this. However, it was different a hypothesis.

For our hypothesis, this is an organophosphate like answer. We would hypothesize that these, the chemical exposures, would be highly associated with these measures whereas as these other were not plausible, we found those would not be associated maybe to a relative risk of one to one-a-half, one-and-a-half to two. And the others would have higher relative risk. Now if we follow it out to a relative risk of two, one or two, then we would attribute that to recall bias and then whatever.

However, if we get all this, these are not and that is more interesting. And that is what we found. Syndrome 1, impaired cognition. Remember this is the yes or no impaired cognition. Syndrome 1 had two respecters associated with it. This is on the next slide.

The first one was, did you wear a flea collar. Those who said yes, there were 20 of them. Five of them have Syndrome 1, or at a risk of 25 percent. Those who did not wear one, the risk was -- Syndrome 1 was 3 percent. You divide 25 by 3 and you get a relative risk of 8. That is Syndrome 1 was eight times more common in those who wore flea collars than in those who didn't.

Does it make sense? And this is highly statistically significant. It does not -- it could not be due to chance. We required a P value of less than -- less than or equal to .001 to get in the game. Again, because we are testing multiple hypotheses.

Now this is small numbers. And this is a very provisional finding. We looked for plausibility though against small numbers, and you don't want to overinterpret this, but we looked at those who wore flea collars, who never wore them, wore them, but never next to their skin and sometimes next to their skin and went 3 percent, 18 percent, 67 percent.

Again, small percent. You know, this may not be real, but it is the least of our trends, of the significant trend test and the trend test of the whole. So right away we suspected that maybe there is something about the flea collar, some impairment to those people.

Another risk factor that was significantly associated was does your job duties involve being a sentry. These were the guys that stood guard at night of the Seabees. They had what is called a yeoman person was made to stand guard. A little light on the lingo here.

The job duties involved -- and here 22 divided by 3, a relative risk of 6. Now these guys were exposed to ambient conditions at night, there was more spraying of pesticides initiated at night, also there were SCUD missile attacks more at night. Who know what this means. I wouldn't know to interpret it, but possibly more environmental exposure.

Now let me go to Syndrome 3 now. Arthro-myo-neuropathy. You might think about it as fibromyalgia or body pain, tingling, and numbness. Sort of a sensory problem, very central in nature. Here the two risk factors are shown here. Everything else is not associated.

I have developed a scale of the amount of pesticides and insect repellent used. I asked them on the questionnaire, "When you used insect repellent, how often did you put it on during the day? Once a day, twice a day, three, four, five, six, seven time a day. And when you put it on, how much did you put on each time? A little bit." The questionnaire said, "A dab here and there, covering the areas of skin that dried quickly or large areas that is taken up for a while. Or did it actually drip down the skin."

This quantitative scale and do a summary. Quantitative scales, zero to six, zero to eight, I never used any. Six many -- they used it many times. And it was just dripping down the skin. Now, here is the --

DR. ZEGER: Can you say a little bit more about that. So for example, if somebody used it very often late at night --

DR. HALEY: As shown in that paper three in that JAMA series, I actually show you a table that shows you what the numbers are and how I got them.

Now notice the rest, the percent that has Syndrome 3 in these groups were 4 percent, 7 percent, 5 percent, 16, 33. That is a dramatic response effect up to a relative risk of eight.

Highly statistically significant. Here the trend test is highly significant. The numbers are a real small -- I mean a small part. They are in the ballpark. That is a real interesting association. Now we also developed a scale.

Everybody has used pyridostigmine, but we couldn't test pyridostigmine. So couldn't test pyridostigmine or not. So what we looked at from having questioned people in our Dallas studies, we learned that the side effect of pyridostigmine occurred in -- 1 percent of people had minor mestirinic side effects at the beginning and a very minimal blood level. So very minor side effects.

The first to occur organophosphates, a little diarrhea maybe or a rumbling, or urinary urgency, that sort of thing. Very common. And that didn't relate to anything in our disease measures. However, about 10 percent had much more severe side effects, what appear to be nicotinic side effects; that is, muscle fasciculation, muscle twitching, muscle cramping, and clouded consciousness, pounding heartbeat, some other things, which tended to be very commonly endorsed items in some people who took pyridostigmine.

Now we didn't just make this up because this was also described by an Israeli study done during the Gulf War where they studied their troops that were taking pyridostigmine and actually did a survey of symptoms. They found the same dichotomy.

Fifty percent or so got minimal side effects and about ten percent got severe side effects and laurinstin -- linctonstine -- you know, some of you have seen that paper of a case report of an individual veteran who had a cholinesterase, genetic cholinesterase deficiency who developed severe cognitive side effects and muscle poisoning from pyridostigmine. Obviously a unique circumstance.

Anyway, so we developed a factor scale, used factor analysis again to disentangle the acute mestirinic side effects from the less common, more advanced side effects. The first scale, the mestirinic side effect scale had no relationship to anything, but the advanced side effects scale was strongly related to Syndrome 3. That is, none of those side effects or low level in that factor scale up to very high numbers of these symptoms, zero percent, four, eight, twenty-two, seventeen.

Of those response effect, not perfect, but pretty good, up to a relative risk of four highly significant. So this led us to hypothesize then that people who had high levels of pyridostigmine for some -- for whatever reason and who used large amounts of insect repellent, that these two perhaps interacted in some unknown synergistic fashion to produce Syndrome 3.

Now plausibility. There is one more important factor here. This scale of the insect repellents is kind of a -- I think a lot of people would be uncomfortable with that semi-quantitative scale. This is made a lot more plausible by the fact that those people who said -- who endorsed Avon Skin So Soft containing no DEET, this was 4, 3, 3, 2, 1.

And those who endorsed OFF!, the commercial variety, less than 15 -- 30 percent DEET in emollient -- it is not -- that will absorb in the skin, 4, 4, 4, 4, 4. This is in the people who endorsed the military issue insect repellent, 4, 7, 9, 16, 33. That is the 75 percent DEET and ethanol that is highly percutaneously to absorb. So you see this is plausible. So we propose then a synergistic effect between DEET and pyridostigmine.

Yes?

DR. GUILARTE: I have a question about pyridostigmine. You said everybody used it.

DR. HALEY: Yes, everybody used it.

DR. GUILARTE: Is there any assessment on the way that they used it? In other words, if people took a lot at one time and some other individuals took over a period of time. Was there any assessment of the exposure and frequency of taking the pyridostigmine?

DR. HALEY: In this questionnaire, we didn't expect this to come out. We only asked how many tablets did you take. Let's see, we asked, "How many times did you take it," on how many -- sorry, "On how many days did you take it, and approximately what is the total number of pills you took?"

Now what we should have asked that we got on our proposed national survey, we are going to ask, "When you took it," I forgot how this is phrased, but something like, "Did you take it regularly on the clock? Did you take it sometime and not other times or did you not take it until there was a scare and then you took a lot of it?" And there were a number -- and we think that is what happened.

In just talking to some of the -- when these people actually came to the survey, after it was over, we milled around in the auditorium with them and talked to a number of them. A lot of them wanted to stay and talk and we took a lot of personal history. A lot of people said, in this unit particularly, "I didn't take that stuff. Then when the chemical alarms went off, I took -- I made up for it."

Yes, Peter?

CPT MAZZELLA: The deployed controls also used insect repellent, but they also used PB. How did the controls fall out against your hierarchy of application of DEET versus amount of PB?

DR. HALEY: Let's see.

CPT MAZZELLA: These are cases.

DR. HALEY: Yes.

CPT MAZZELLA: But how did the controls have a comparable amount of --

DR. HALEY: Well, see the controls are on the other side here. That is a good question. I would have to go back and look at the prevalences of use. I don't know the answer to that.

Okay. Now let's go back to Syndrome 2 because this is the bad one. Remember these are the guys with confusion, balance disturbances, vestibular ataxia. First of all, the scale of advanced side effects of pyridostigmine, from nicotinic to central side effects of pyridostigmine. Zero to six.

Look at the dose response effect. Zero percent, one percent, eleven percent, forty-three percent, and the numbers are not small. This is a dramatic effect. So something -- the guys who got more -- who at least endorsed the questions, who suggested they got high blood levels of pyridostigmine, had a dramatically increasing risk of the bad ones, Syndrome 2, what we think is real central brain damage.

All right. There is one other risk factor. And that is, we actually had a set of questions to produce a combat exposure scale. You know, like in a lot of Vietnam Agent Orange research there is combat exposures scale. So you ask a number of items. You know, "Did you ride across the border in a military vehicle? Did you get shot at? Did you shoot somebody? Were you almost overrun?" And et cetera, et cetera. And that measures combat exposure.

And we added one. "Did you experience a likely chemical weapons attack," not expecting that to show anything, and wrote a risk of seven. Now this we really doubted. I can recall as a student and a resident at the Dallas VA, we used to rotate through the Dallas VA, still do.

But I remember the old World War I veterans who had chronic lung disease, you know, gasping for breath saying I was gassed in the Argonne. You know, we all just laughed that off, said, "Oh, I doubt that." You know. Well, I think maybe they were.

These are the guys who said, "I was gassed in the Gulf." Now what do they mean by this? You know, the validity of this would be in question. What they mean by this, there are other questions we asked about, "Were you near chemical alarms when the chemical alarms went off" and, "Did a SCUD missile land within so many kilometers of your position," and so forth. And those were highly correlated with this.

Well, we really didn't -- we were concerned about the validity of that. We then did the geographical time analysis. We looked in the computer, is there a place and date that the people who were there on that place and date have a high relative risk. And there was one hit only for Syndrome 2 and it was people who were located in what is on my analysis map, sector seven in northeastern Saudi Arabia on January the 20th. The 20th is the fourth day of the air war and the relative risk was four.

Let me show you where that is. Here is the Persian Gulf, Kuwait, Saudi Arabia, Iraq. These were the positions of our troops in the -- during the bombing period. Of course, then there was a big movement up here and in there during the ground war.

The sector seven on my map is here near Khafji. So it was the Seabees who were near Khafji that had the highest relative risk for Syndrome 2. Now what happened -- so then we went and started madly looking, what happened on the fourth day of the bombing, January 20, 19th and 20th. What happened near the town of Khafji.

And what we found is, that was the fourth day of the air war, that was the day on which our bombers struck the chemical weapon depots along the Euphrates river. There were four or five of them along there. Jim Tuite, those of you who know who -- he is the investigator for the Regal Commission, has weather satellite photos that I think he is going to try to publish soon that show on that day the cloud -- the bombing clouds went up to at least to about 10,000 feet and drifted right over our troops.

That was the same day in which the Czechoslovakian chemical weapon experts detected sarin and a mustard agent here at Haffa Al-Batin, which is somewhat of west of there, and on this day, the chemical alarms went off at Khafji. There is a case report in the Regal Commission report showing that the chemical alarms actually went off and people developed symptoms for 24 hours thereafter and so forth.

Now is that chemical weapons attack? Is that collateral fall-out from our bombing? Obviously, that is an issue that an epidemiologist can't address. The point is, the place and date is a plausible thing and it just fell out of our data. And there is no other place and date that comes close to being significantly related to any of the syndromes or to Syndrome 2 particularly.

Now experiencing chemical weapon exposure, whether it was an attack or a fall-out or a whatever, an alarm, yes/no, and the scale of advanced side effects from pyridostigmine, having a high blood level perhaps of pyridostigmine, yes or no, if those are truly biologically interactive, you would expect there would be a synergistic effect. That is, those exposed to both would have a higher risk of Syndrome 2 than those exposed to just one and much higher than those exposed to zero.

Those exposed to neither one, one out of a hundred and fourteen point nine percent had Syndrome 2. If you had only the pyridostigmine side effects, advanced side effects, 7 percent, 2 out of 27. Only the chemical weapon thing, but no pyridostigmine side effects, 1.6. If you had both, it was 38 percent. That is a significant -- those are synergistic effect by Rothman's synergy scale. There are two or three synergy scales we published. I think two of them. They are both statistically significant synergy.

So you see what we have now here, we have relative risks of four to eight -- actually to thirty-two if you look at the high end of some of those dose response effects. We have dose response effects and we have a synergistic effect, and we have a plausible hypothesis in terms of the war events that went on.

Now to summarize, impaired cognition, confusion ataxia, arthro-myo-neuropathy, Syndrome 1, impaired cognition associated with wearing flea collars. In that era, most flea collars contained Dursban, which is chlorpyrifos.

This has been reported to cause a brain injury, a central nervous system brain injury, and it has recently been taken off the consumer market by EPA because of a huge number of outpouring of case reports that actually were covered up by the chemical industry for several years. When they were discovered, there were sanctions and so forth of Dow Chemical.

And also worked in security. I am not sure what that means. Confusion ataxia, the bad one, these guys are occupationally impaired, severe symptomatology of what we think is a central nervous system injury, indicators of exposure to chemical nerve agent possibly and advanced side effects from pyridostigmine, and a dose response effect to this and a synergistic effect between the two.

The body pain syndrome, which we think is a central pain problem. Government issued insect repellent, 75 percent DEET ethanol. And a dose response effect. Ethanol and advanced -- in ethanol. And advanced side effects of PB in a dose response effect.

We couldn't test for synergy here because there was nothing on the off diagonal. There was nobody that had one and not the other of these two. But it looks like it would have been there because it is such a big difference similar to that.

Now let me just say one more thing. People who said -- two more things. People who said, "This study is no good. This study doesn't help us. Let's throw it out because it is too small. Twenty-six cases and twenty controls. What could you possibly know about 700,000 people with 20 cases and 20 controls?"

Well, first of all, those were developed -- there were sixty-three cases and a hundred eighty-six controls of sick and well and we picked twenty-six of these and twenty of those, but these were carefully age, sex, and education matched in the same battalion, same jobs, and that sort of thing.

Toxic shock syndrome, a nationwide epidemic. My wife, by the way, was the second case of toxic shock syndrome. She was at Bethesda Naval. We were up here on duty up here working from HCFA assigned from CDC and she got toxic shock. Got one of the tampons in the mail and almost died of toxic shock syndrome. I have a very active interest. In fact, her blood was the first case where they isolated the sleeper toxin.

But anyway, the case control that cracked it and related tampons, and particularly the Rely tampons, fifty cases, a hundred and fifty controlled. Selected, typical cases. No random selection, no two or three representative. The next fifty cases and a hundred and fifty controls. They then picked 28 of these and 32 matched controls and identified the toxin, which is now accepted and has been duplicated over and over again.

AIDS. The national AIDS case control study that told us everything we still know about the transmission of HIV, except for the virus. Fifty cases, a hundred and twenty-five controlled. Just 50 sporadically picked people who had this acquired syndrome. And of course there have been many studies since then to corroborate all of it.

Hantavirus, the four corners pneumonia that turned about to be Hantavirus. Seventeen cases, two hundred and forty-eight controls determined the opinion of that exposure, the relationship to rat urine and droppings and so forth.

Legionnaires disease, 59 cases and 59 controls. The second case control study, a hundred thirteen and a hundred and forty-seven. The big case control study number two, which is the definitive finding of the airborne, 56 cases and 50 controls.

The point I am making here is those people who have discounted our study have done so without knowing the history of investigating epidemics. Now granted, our study is not the end all and be all, but let me tell you, what we have found is something that is very important. And were this a commercial product, it would be off the market now.

Tim?

DR. GARRITY: Let's go back to the slide.

DR. HALEY: Yes.

DR. GARRITY: Could you tell me what the case definition was for toxic shock syndrome acquired in the efficiency study and Hantavirus and Legionnaires.

DR. HALEY: Yes. I mean, I have to go back for the exact wording, but here it was high fever, low blood pressure, and red skin. Okay. Here it was -- I have to go back and look, but it is immune deficiency and PCP. And they didn't want -- it might have all been PCP, I am not sure.

And here it was, you know, pneumonia in an otherwise healthy person. Here it was -- your point being that these were clear case definitions. That is what made this -- made those so non-controversial. Nobody -- there was no questioning about these when they came out. This was all accepted. And there are a thousand epidemics like this and usually the case definition is obvious.

What made this hard and the reason nobody could approach it is, we couldn't come up with a case definition. Well, what is the teaching though? If you can't come up with a case definition, then you got to come up with a case definition. Because failing to do so means these go unaddressed and they continue on and on in the population or they occur in the end over and over.

Now I am not underestimating the importance of corroborating this. So we are -- in the next hour, we are going to talk about what we are going to do to corroborate, replicate. And what we are proposing to do is a national survey to corroborate it in the best style.

Oh, one more point. Biological plausibility, animal studies, we have the dream team here of animal toxicology, Tom Kurt, our own toxicologist who designed the studies for not only Kansas states, but dean of all the journals in this field. He developed dosing the animals, Abou-Donia from Duke, has an EPA lab that does the OPIDN studies for the EPA constantly and had it all geared up ready to go, and Karl Jensen EPA did the neurotoxicology -- the neuropathology.

We did study for pyrifos, it was in the flea collars, Rabon in the flea collars, DEET from -- we were going to do chemical nerve agent. I don't know ever what happened. We actually got the serum, but Abou-Donia never did it. Or he did it, but he hasn't published the results. I think he is afraid he won't ever get another grant if he publishes the results. And then pyridostigmine.

And we looked at those individually with each agent individually and in combination. We exposed four more hens to a sub-lethal dose five days a week for sixty days. In your initial study, you don't do it gingerly, you do it for a long time to see is there an effect. Then you go back and try to do more plausible things.

Observe the clinical signs of neurotoxicity and then sacrifice the hens and look at the tissue for neuropathological. There is one of the hens who developed symptoms. Clinical pinpoints were gait disturbances, tremor, and paralysis developing in a delayed fashion, not acutely, post-mortem assay of neurotoxic esterase, which is the inside involved in producing the OPIDN nerve degeneration. Plasma and brain cholinesterase in histopathology, nerve, and sciatic nerve and spinal cord were examined.

Basically to summarize, what Abou-Donia and colleagues found was that controls had no clinical signs, no local motor dysfunction, tremor, no spinal cord legions, sciatic nerve legions. And so their overall rank was low in terms of outcome.

The pyridostigmine parathion and DEET alone had minimal effects. Some symptomatology, but basically no real things that you could be sure about. But in combination, all of these combinations, except the parathion was not as potent, but the other combinations, particularly the DEET and Dursban and the PB and Dursban, the PB and DEET, all of these caused delayed symptomatology, local motor dysfunction, and tremor. They all caused spinal cord and sciatic nerve degeneration that appeared to be statistically significant.

So this suggests the plausibility that these agents might work synergistically to produce promotion -- some kind of promotion effect among themselves to produce some kind of nerve damage or neurological damage. Now again, this is an OPIDN model. We are not sure that that is what is really going on in the brain syndrome in people with pesticide toxicity or Gulf War Syndrome, but it suggests a combination of these are neurologically active.

Yes, sir?

DR. GARRITY: That was the question I was going to have. Was the brain of these hens examined?

DR. HALEY: No. We are actually doing rat models now. We are doing these same studies in rats and we have a very good team of neurobiologists who are going to be doing neuron counting and all the various nuclei and we are going to find where this is because we have now got evidence.

DR. GARRITY: In rat models, OPIDN?

DR. HALEY: Yes. That has been described in the literature before, although not with these particular agents. We now have -- we have now tetraded so we can produce the spray legion or we can produce cognitive disturbances in animals that are long-lasting with single agents. And we are now in the process of trying this with dual agents and working out the doses to produce a minimally brain damaged rat. We get behavioral disturbances which are chronic and delayed, but not acute findings. Not verbal or neurotic.

DR. GARRITY: Are they behavioral disturbance in terms of accessory or memory recognition?

DR. HALEY: Yes.

DR. GARRITY: You know what type of tests?

DR. HALEY: I should have memorized that before I came. Watermaze --

DR. GARRITY: Moore's Watermaze?

DR. HALEY: The Moore's Watermaze test and the searching behavior test, searching around and counting.

Now here is a longitudinal section of the peripheral nerve of one of the Abadanian hens and you will see the characteristic legion of OPIDN. So they actually got OPIDN, which is not what the veterans have. At least they don't have the peripheral variety.

Okay. That is the morning -- the first session. Why don't -- you all want to talk a little bit or have a break?

ADM ZUMWALT: Can I just ask -- I would like to ask one question just to follow up on what Dr. Garrity had asked. I am a physiologist, not a statistician and, you know, we usually look for pretty clear outcomes in our experiments and a factor analysis is sort of, you know, a data dredge to salvage the data later. Can you come up with any example where CDC or anybody else has ever done an epidemic investigation where they discovered the disease through a factor analysis or three diseases?

DR. HALEY: Only one and that is Gulf War Syndrome because, you know, Racuda and his group at CDC did the same thing as I am going to show you this afternoon. They got two of our three factors and their factors are also covered in JAMA. They didn't mention that that they were the same for political reasons, but they did.

As far as I know, there is no other example. Again, the epidemics that CDC has focused on are ones where there is a clear disease. This one didn't have a clear disease. Now the question is, could we just go and never find it? Without a case definition, we are punting and you will never find it. So the question is, do we go to something like that?

ADM ZUMWALT: But is there a risk that we are constructing something that is artificial here?

DR. HALEY: Good question.

ADM ZUMWALT: You know, I mean, this goes to the heart of this.

DR. HALEY: The answer to that is two things. Once we construct this case definition of factors, we are all -- the question then is, when we go out and look at the brains of these people with very sensitive and very objective measures, the people with Syndrome 2 have demonstrable brain dysfunction and demonstrable brain damage on imaging or whatever else. And Syndromes 1 and 3 have minimal dysfunction and damage and the controls don't. That is your answer. In other words, that is why it works. If it works, then it is a case definition.

DR. BRIX: But then do you select just your Syndrome 2 as the case definition and you will discard the other two?

DR. HALEY: No. Here is our case definition.

DR. BRIX: You are giving yourself three chances here, aren't you?

DR. HALEY: No-no. Syndrome -- or out of our neurophysiological testing, the audiovestibular tests, you know, that I showed, we then for the next phase that I am going to talk about in the next hour, we constructed a hypothesis that two is more severe than one and three, is more severe -- more impaired than the controls.

In other words, we felt Syndrome 2 is a severe brain, central nervous systems brain injury, 1 and 3 are mild central nervous system brain injuries. And the controls, of course, should be normal.

Moreover, we looked at the symptoms of these and we saw what we think is a striking similarity with three diseases of the basal ganglia. There were three diseases that primarily affect the basal ganglia. These are Huntington's disease, it affects the caudate nucleus, one of the three nuclei of the basal ganglia; Wilson's disease, copper spores disease, hepatolenticular degeneration, affects the -- what is it? I will think of it.

DR. GUILARTE: Mendelian.

DR. HALEY: Mendelian. Thank you. And then Fahr's disease, which is a really idiopathic calcification of the globus pallidus. Those three diseases are well known in medicine. Now obviously, those people all developed severe neurological damage, which often proves fatal ultimately.

But in the initial presenting phase, they look like Gulf War Syndrome. They look like our three syndromes. So from that we are going to see in the later phase that Syndrome 2 is a severe damage to these structures and Syndromes 1 and 3 are mild damage to these structures and plausibly normal. Now look at -- this is a hypothesis. It is a guess.

Yes?

DR. ZEGER: I just might react to that -- Dr. Friedl's comment.

LTC FRIEDL: Friedl.

DR. ZEGER: That I think these sorts of methods are typical in the field of psychiatry where diagnoses have traditionally been defined in terms of a series of symptoms and that subsequent to early diagnosis, there are these sorts of analyses done to try to make a more rigorous diagnoses and to look at this -- if you look at the definitions of depression or schizophrenia or several of the major psychiatric disorders, which we now can effectively treat with pharmaceutical products, you start with symptoms that are defined sort of in the same technology -- using the same technology.

DR. LIN: You have two survey instruments, but you have also conducted neurological testing of brain scan. I was wondering that before you launch into the human experimentation, what kind of requirements do you as the investigator that your institution require. Do you meet "for review?"

DR. HALEY: Yes. We have an institutional review board that would use every step.

DR. LIN: And do you have an informed consent?

DR. HALEY: Yes. With all of these we have a consent form. In fact with later studies I am going to do in the next hour, we had a discussion with the Defense Department and Dr. Friedl.

DR. LIN: Also, to just follow-up, you have a slide that has the -- ensure the confidentiality to your other patient with your signature and Mr. Perot. How do you ensure absolute confidentiality. In other words, if court order you, how do you ensure?

DR. HALEY: Oh, in Texas there is a state law that protects medical records, which these would fall under. And they are only available -- they can only be discovered in a criminal case. And in a criminal case, all bets are off. And that is basically the standard in most of the states. I know in Georgia that was the case when I was at CDC. These things are protected except they are not -- in other words, they are not discoverable by the Open Records Act in Texas.

Yes.

DR. GARRITY: In Syndrome 1 and Syndrome 2, and I am not sure about Syndrome 3, but in 1 and 2 there were self-reported psychiatric conditions. In Syndrome 1 you said depression, and I wasn't clear whether that was a self-reported diagnosis of depression or whether that was a symptom that was --

DR. HALEY: Self-reported.

DR. GARRITY: And in Syndrome 2 self-reported -- that they had gotten a diagnosis of PTSD. And I was wondering two things. Number one was did you do any case to ascertain them, PTSD case. That is, did you either get their records, their medical records and go through it to see whether or not that diagnosis of PTSD was supported or did your own evaluation of PTSD?

DR. HALEY: The latter, but not the former.

DR. GARRITY: The latter, but not the former.

DR. HALEY: Yes. In other words, in the case control study when we brought these people in, the second time, which I am going to describe later, we did a very thorough SCID -- psychiatric interview by a psychiatrist.

DR. GARRITY: How many of those then was a self-reported diagnosis of PTSD turned out to be not or went the other way?

DR. HALEY: I have a slide I am going to show you.

DR. GARRITY: But I just want to conclude that -- there is something associated with this, and I want to emphasize that I am not suggesting that PTSD is the problem. Quite the contrary, I think, you know, most of us sitting here at the table do not believe that PTSD is the problem.

DR. HALEY: Right.

DR. GARRITY: But the concern that I have is that if somebody is suffering from depression, somebody who had a diagnosis of PTSD, there is a -- I would wonder whether or not they are on any medications and were they on any medications and if so, what is your control for the side inference of that?

DR. HALEY: None. We were able to take everybody off all meds when they came in for our study. We stuck to the R&Ds. That is a most sensitive issue, as you know. None of them were on meds they couldn't go off of and they were comfortable doing so. So none of them were and they were all off before. So this wouldn't be confounded by medication.

Yes, sir, Dr. Green?

DR. GREEN: Just a point of clarification. The definition of Syndrome 1, as I read the characteristics of that patient population, I have to admit that I am an increasingly elderly internist myself. And the definition really strikes me as those that either experience associate with a clinical diagnosis of depression. And I ask this to you as an internist, I mean, just for clarification, how do you distinguish these clinical symptoms from the general medical definition of depression?

DR. HALEY: Absolutely. That has been a source of lengthy discussion in our group. And when we first saw the syndromes and I made that list of the syndrome, we didn't have the psychological data examined yet.

So we saw those syndromes and I said, "Look. Number one is depression. Number two, that is more worrisome, but maybe that is somataform. You know, although, now that I have studied what somataform, that doesn't make -- I shouldn't have thought that. And three is fibromyalgia, whatever that is. Whe