1 JULY 22, 1994 2 * * * * * 3 DIRECT EXAMINATION (CONTINUED) 4 5 BY MR. SMITH: 6 Q Doctor Thompson, on 7 Monday we were discussing the study done by Doctor 8 Stuart Montgomery, do you recall that? 9 A Wednesday? 10 Q I mean Wednesday, I'm 11 sorry. 12 A Yes, sir. 13 Q And I had asked you if 14 you would seek to get some more information 15 concerning that study, the details of which we 16 discussed generally on Wednesday. Have you had an 17 opportunity to do that? 18 A I have to ask my 19 colleague. 20 MR. MYERS: I've looked 21 into that, Paul, and I believe that sometime later 22 this morning we'll have some additional 23 documentation of that study from the NDA 24 documents. 579 1 MR. SMITH: All right. 2 Q (BY MR. SMITH) Was this 3 something that you did and requested, Doctor 4 Thompson, or was this something that was taken 5 care of by the legal department at Eli Lilly? 6 A No, I asked my colleagues 7 to help me with that since by the time we got 8 through many people had already left work. 9 Q Are you talking about 10 your lawyer colleagues or your colleagues in the 11 scientific area at Lilly? 12 A Lawyer colleagues. I 13 mean, we're all colleagues. 14 MR. SMITH: And you think 15 we'll have some later this afternoon or later this 16 morning? 17 MR. MYERS: Later this 18 morning, I think. 19 Q (BY MR. SMITH) I believe 20 it was also on Wednesday that we were discussing 21 some work that you had done either during your 22 training or when you were in the active practice 23 of medicine concerning overdose and poisoning 24 cases? 580 1 A Yes, sir. 2 Q Do you recall that? Can 3 you give me some more details? Where was that 4 done, number one? 5 A Well, I have to divide it 6 into at least three different segments because I 7 both took care of a lot of poison patients, I've 8 done some research in poisoning, and in addition 9 I've done a lot of teaching in organized 10 institutions to manage poisoning. 11 Let me go back to the 12 very beginning. Besides my ordinary training when 13 I first developed the intensive care unit at Johns 14 Hopkins we decided to admit all poison patients, 15 and in fact all nonsurgical suicide attempts that 16 had a medical problem to the medical intensive 17 care unit that I was developing and directing, and 18 then when I moved to Case Western Reserve and 19 built my third ICU we had the same sort of rule 20 because in-hospital suicide actually is a fairly 21 significant loss of patients who actually reach 22 the hospital alive on a suicide attempt, and it's 23 harder to kill yourself in an ICU. 24 So, keeping records of my 581 1 own patients in the ICU environment, about 2 twenty-five percent of them were in fact admitted 3 because of a drug overdose or suicide attempt. So 4 that's my patient population experience. 5 In terms of research, I 6 actually had NIH grants and did a fair amount of 7 research looking at the alternatives of using 8 everything from emetics to charcoal. I developed 9 a new emetic myself, which is patented but never 10 marketed. I developed new charcoals, published 11 and lectured widely on different ways of managing 12 the poison patient, hemoperfusion, a whole variety 13 of techniques. And then I set up the Northeast 14 Ohio Poison Center as both a treatment center and 15 a reference center for people in northeast Ohio to 16 use, physicians primarily, when they had a patient 17 with poisoning. 18 Q All right. Are you 19 drawing a distinction between poisoning and 20 attempted suicide by overdose? 21 A Well, you can have 22 certainly accidental overdose. I mean, I've 23 treated some patients who without any intent to 24 harm themselves quite legitimately took say 582 1 several different types of salicylate and got 2 therefore an overdose of salicylate. 3 Q What's salicylate? 4 A I'm sorry, aspirin. 5 Q All right. 6 A And what will happen is 7 aspirin is in so many preparations that the 8 patient may not realize that they're actually 9 taking two or three preparations for pain or 10 whatever, all of which contain aspirin, and you 11 get into trouble that way. I've also, of course, 12 taken care of poisoning that was related to 13 industrial accidents or whatever that had no 14 self -- you know, self-injury intent, but the 15 predominant, the most of those patients in fact 16 were making suicidal gestures or real attempts at 17 killing themselves. 18 Q All right. Would you -- 19 could you characterize those people who had intent 20 as individuals who took a -- had intentional 21 overdose and were not trying to increase the 22 therapeutic effects of whatever they were 23 ingesting? 24 A That's the vast 583 1 majority. I mean, again, there are a few people 2 who make a mistake and take too much, but have no 3 real intent to injure themselves. 4 Q All right. For instance, 5 somebody might be anxious and take a Valium and 6 then say, well, I'm still anxious, take another 7 Valium, I'm still anxious and take another Valium 8 and have an intentional overdose, but not have an 9 intent to harm themselves, correct? 10 A Yes, sir, that's correct. 11 Q What I would like to 12 discuss your experience in is those situations 13 where there was an intentional overdose where the 14 patient intended to harm themselves, all right? 15 A Yes, sir. 16 Q Not accidental and not 17 those instances where a patient was seeking 18 greater therapeutic benefits of the drug they were 19 taking, all right? 20 A Yes, sir, I understand. 21 Q Now, can you give me a 22 number of patients that you have seen for 23 intentional overdose? 24 A Around eight hundred -- 584 1 well, eight hundred total overdoses, of which 2 probably eighty-five to ninety percent were 3 intending to harm themselves, so somewhere in the 4 seven hundred to seven fifty range that I -- those 5 are the ones I personally took care of as the 6 primary physician. I often got called in on 7 consults to see other people's patients. 8 Q Let's just limit it for 9 ease of discussing this situation those patients 10 that you were directly taking care of, and you saw 11 those patients at Johns Hopkins? 12 A Yes, sir. 13 Q As a critical care 14 specialist? 15 A Yes, sir. 16 Q And at Case Western 17 Reserve University as a critical care specialist? 18 A Yes, sir. 19 Q Did you see those 20 intentional overdose patients at any other 21 facilities? 22 A Well, in both of those 23 facilities there were multiple hospitals within 24 the university system, and for my own patients, I 585 1 mean that I was responsible for, that's the only 2 place that I saw them, although since I've 3 lectured fairly widely on it, I've seen -- as a 4 visiting professor I've seen patients many other 5 places. 6 Q All right. Of those 7 seven hundred and fifty patients that you 8 treated -- and I assume you rendered treatment for 9 each of those patients? 10 A Yes, sir. 11 Q Approximately what 12 percentage or number had taken overdoses of 13 prescription medications? 14 A At least probably 15 ninety-five percent. 16 Q Did you ever do any 17 survey or make any determination as to what 18 particular drugs were involved? 19 A Oh, yes, sir. I mean -- 20 remember earlier I said I had set up the drug 21 analysis laboratory in Cleveland, and I worked 22 very closely with the Cuyahoga County coroner's 23 office that was a referenced toxicology facility, 24 so we endeavored to measure the drugs in almost 586 1 all of those patients. 2 Q All right. What was the 3 primary prescription medication that was taken in 4 intentional overdose cases that you saw? 5 A It was all over the map. 6 We actually did an epidemiological study in the 7 thirty hospitals in the Northeastern Ohio region, 8 so somewhere I've got exact data on that. 9 Probably -- as a group, rather than a single 10 molecule, probably it would be a toss-up between 11 antidepressants and benzodiazepines. Now, ethanol 12 is often seen in people who take other drugs as 13 well, so -- but you were focusing primarily on 14 prescription drugs. 15 Q Yes, sir. 16 A I'd say maybe equal 17 numbers of antidepressants and benzodiazepines. 18 Q Do you recall the trade 19 names of any particular antidepressants that 20 constituted a majority of the antidepressants that 21 were taken in overdose? 22 A Probably Elavil was 23 perhaps slightly ahead of the others, but most of 24 the others that were on the market I've seen at 587 1 one time or another. 2 Q Have you got a number two 3 drug behind Elavil, if you recall? 4 A Trazodone was pretty 5 common. 6 Q And of the 7 benzodiazepines, was there a particular brand name 8 drug that was seen more often than others? 9 A Yes, sir, Valium was 10 clearly the number one drug, and my guess would be 11 librium might be the number two. 12 Q Of these seven hundred 13 and fifty patients that you treated, did you keep 14 any record of the survival rate of those patients? 15 A Yes, I actually 16 instituted a computer system to track every 17 patient that came into the intensive care unit, 18 including those. 19 Q And what percentage 20 survival were you getting? 21 A More than ninety-nine 22 percent. I think I can probably tell you each one 23 of the patients that died. 24 Q Of those fatalities, was 588 1 there one single drug that seemed to appear more 2 often in the fatalities? 3 A No, each one of them was 4 unique. I only had -- let's see, I lost one 5 cyanide. Again, if you get back to prescription 6 drugs, there was a young woman who died, which we 7 never understood, where we only were able to 8 measure sedatives, including Valium in her, and we 9 never understood why she died because she -- you 10 shouldn't die of a sedative overdose. I remember 11 that one very distinctly. I don't think there was 12 a predominant one at all. But again, the number 13 of people who died was pretty small. 14 Q Why is that? 15 A If you reach a hospital 16 alive and you don't take -- and you take a 17 prescription drug and you reach a hospital alive, 18 by and large you ought to survive. I mean, the 19 people that I lost were people that took Vacor, 20 which is a rat poison, a guy who came in brain 21 dead from cyanide, a brain dead heroine overdose. 22 This one young woman who we could only measure 23 sedatives in, she was the only one that was really 24 shocking in terms of we shouldn't have lost her, 589 1 and I still don't know why. I mean, she was just 2 brain dead, her brain never recovered. 3 Q Of those seven hundred 4 and fifty patients that you treated for 5 intentional overdose, how many of them -- or did 6 you ever monitor how many of them received 7 psychiatric care subsequent to the overdose? 8 A Well, close to all of 9 them because my practice would be to call a 10 psychiatric consult on virtually all of those 11 patients. I can't guarantee it was a hundred 12 percent, but it was very close to that. 13 Q Did you track whether or 14 not that patient was followed after that patient 15 was discharged from the hospital from the 16 psychiatric standpoint? 17 A Not systematically. I 18 mean, some of those patients would be transferred 19 from the medical ICU that I was responsible for to 20 inpatient psychiatric service. Some got 21 transferred to another inpatient psychiatric 22 service outside of our own hospital, and I know 23 the psychiatrics followed some of them, but I 24 honestly didn't do any systematic tracking of 590 1 that. 2 Q In your analysis, did you 3 track or did you make any determination concerning 4 whether or not of those seven hundred and fifty 5 patients you saw this was the first or a 6 subsequent intentional overdose that the patient 7 had made? 8 A Not in a scientifically 9 rigorous sense. I can give you a pretty good 10 estimate from the histories that I took from them, 11 but I didn't track it in the sense of doing a 12 study. 13 Q All right, what would be 14 your estimate based on your recollection of taking 15 their histories? 16 A Between a half and 17 two-thirds gave a clear-cut history of attempting 18 suicide by some means, not necessarily a drug 19 overdose, in the past. 20 Q So fifty to sixty-six 21 percent approximately would have had previous 22 suicide attempts? 23 A Yes, sir. 24 Q And did you make any 591 1 determination concerning those previous suicide 2 attempts with respect to whether or not those were 3 with prescription medications also? 4 A Well, I took a complete 5 history, but I -- and it was all over the map. I 6 mean, as I indicated to you, some of the patients 7 would have used a nonmedical means of trying to 8 kill themselves, so I would feel a little bit less 9 comfortable giving you a precise number, but most 10 adults who take drug overdoses in fact include 11 prescription drugs in their overdose. Most adults 12 take multiple drugs, if you include alcohol, and 13 often over-the-counter drugs are included. So I 14 would say that -- again, from my general memory of 15 all of those histories, probably most of them had 16 in fact used prescription drugs in the past. 17 Q Of those seven hundred 18 and fifty patients that you treated, in what 19 percentage did you find alcohol involved? 20 A Well, when we measured 21 it, because we didn't necessarily measure it in 22 every patient, when we actually tried to measure 23 it, it was at least fifty percent, and probably 24 more than that. Again, that wasn't something that 592 1 I studied specifically. 2 Q Beg your pardon? 3 A I'm sorry, it wasn't 4 something I did an actual study on. 5 Q But took a history of 6 these patients, once they regained consciousness, 7 if they had been unconscious? 8 A Yes, sir. 9 Q And you performed 10 laboratory analysis that would tell whether or not 11 the patient had consumed alcohol, is that correct? 12 A Yes, sir, but not in 13 every case because, for example, if we could 14 elicit a good history from the patient within a 15 few hours of admission, we wouldn't necessarily 16 survey for alcohol, where if the patient were 17 comatose we would do that as a routine part of the 18 workup. 19 Q How about illicit drugs, 20 did you come to any conclusion or do you have any 21 percentage of those seven hundred and fifty 22 patients who had consumed illicit drugs in 23 connection with their overdose? 24 A Yes, sir, when I first 593 1 opened the ICU in Hopkins, which was in 1970-71, 2 fifty percent of the drug overdoses were overdoses 3 with heroine or related opioid, but that actually 4 dwindled, so by the time I left -- the last time I 5 attended in the ICU was November of 1991, and at 6 that time fewer than ten percent of the people had 7 illicit drugs that we could measure, and it had 8 changed from being predominantly heroine or 9 another opioid to being all over the map, PCP, 10 cocaine -- didn't see too much cocaine at that 11 time, but PCP, cocaine, LSD, speed. 12 Q Would those individuals 13 who had made intentional overdoses that included 14 consumption of illicit drugs also had on board 15 prescription medication? 16 A Often. Occasionally, 17 especially early on in Baltimore, we were seeing 18 people who had just overdosed on heroine and 19 whatever other junk happened to be in the bag, but 20 later we were seeing a population of mixed 21 abusers, so they would often come in with 22 prescription drugs, alcohol and several illicit 23 drugs all together. 24 Q Do the percentages that 594 1 you've given me concerning your experience with 2 the intentional overdose patients that you've seen 3 comport with statistics that have been recorded 4 and reported generally throughout the medical 5 profession? 6 A Yes, sir, it would be 7 representative of large cities in America during 8 that time period, because then -- again, there's 9 been a big change, but over the time period of the 10 mid '60s to '81, my review of the literature, 11 because I was lecturing constantly on this 12 subject, were that those were pretty close to what 13 my colleagues in other cities were seeing and what 14 was in the literature. 15 Q When you were doing this 16 work on intentional overdoses, did this consume a 17 significant portion of your work, of your day for 18 instance? 19 A Well, it was a fourth -- 20 when I was seeing patients, which would be a month 21 at the time, I would be responsible for all the 22 patients admitted to the ICU. Again, those were a 23 fourth of my patients. In terms of research, most 24 of the time I had three major government grants 595 1 and a bunch of industry grants, and not 2 necessarily the biggest, but about a third of the 3 government supported research was in this area, I 4 told you we developed a new emetic, working with 5 Amoco we developed charcoals and so forth. 6 Q When you took these 7 histories from these patients and treated these 8 patients and made notations with respect to the 9 medical records, would your medical records 10 indicate intentional overdose or would it be 11 suicide attempt, suicidal gesture, how would that 12 generally be termed when you would refer to it in 13 the chart of the patients? 14 A Before I came to Lilly, I 15 didn't have any systematic lexicon and I wouldn't 16 know -- I mean, again just as in notes I type 17 about memos -- I mean about conversations, I try 18 to as close as possible stick to the words of the 19 patient. 20 Q If the patient said I 21 attempted suicide or if the ambulance driver 22 happened to walk in and say we got a suicide 23 attempt here, or the nurse happened to call you 24 into the emergency room and said we got a suicide 596 1 attempt, that would generally be what would be 2 recorded on the medical records? 3 A No, sir, not hearsay, 4 using the legal definition of hearsay. I'd try to 5 stick to the best information that I had, so if I 6 had absolutely no other information, I'd put 7 hearsay down, but generally I'd try to elicit the 8 history from the patient, the patient's 9 physicians, the patient's family, and of course 10 the psychiatrist would get an independent history 11 and probably got deeper into the motivation of the 12 patient than I would as an internist. 13 Q All right. But if the 14 patient said I attempted to kill myself, that 15 would be what you would record in the office notes 16 or the hospital chart? 17 A Yes, sir. 18 Q If the patient said I 19 wanted to commit suicide, you would say attempted 20 suicide? 21 A Yes, sir. 22 Q If the patient said I 23 wanted to get attention, you would say something 24 of that nature? 597 1 A But of course I'd draw my 2 own conclusions from all the data in terms of how 3 serious the attempt actually was; that's a pretty 4 subtle call, to be honest with you. 5 Q Well, I would assume that 6 those individuals that you saw in the critical 7 care wards in these hospitals, ninety-nine percent 8 of them were indeed attempting to cause harm to 9 themselves, in your judgment? 10 A I don't think I testified 11 it was ninety-nine percent at all; I thought I 12 said it was something like eighty-five percent. 13 Q Okay. 14 A But, you know, you have 15 to -- if you look at the intent of the patient, I 16 mean some patients vividly stand out in my mind 17 because, for example, they would take the pills 18 and call their boyfriend or girlfriend and say 19 I've just taken an overdose. So, you know, if you 20 asked them why did you do that, they'd say, well, 21 I was deeply upset and I wanted to hurt myself. 22 And then you'd say, well, how hurt did you want to 23 get? Oh, I didn't want to kill -- gee, I didn't 24 want to kill myself. So that's what I say, I 598 1 think it's a fairly fine line between what you 2 termed earlier as calling attention, which many of 3 these clearly are, and somebody who really wants 4 to die. Since all of us have the ability to for 5 sure kill ourselves, if somebody does something 6 like that, take a handful of pills and call 7 somebody right away and announce they've done it, 8 my own feeling would be often that, yes, they were 9 desperately in need and they were calling for 10 help, but they probably didn't really expect to be 11 dead. 12 Q And you say that while 13 you were a critical care physician, intentional 14 overdose attempted suicide comprised twenty-five 15 percent of the patients that you saw in the 16 intensive care unit? 17 A Yes, sir. 18 Q So this was, in your 19 opinion, an extremely serious medical problem, I 20 would assume? 21 A Well, let's define 22 extremely very carefully. Not all of those 23 patients in fact had a life-threatening overdose. 24 Remember, earlier I'd indicated that one of the 599 1 reasons we put them in the ICU is that in fact 2 when I surveyed patients at Case Western Reserve, 3 for example, in the previous ten years, half of 4 the deaths occurred in-hospital suicides, people 5 who were medically well but who were able to kill 6 themselves in the hospital, and that was why we 7 put them in the intensive care unit. There were 8 two real reasons: One is even if they looked 9 stable at the moment, we didn't know exactly what 10 they had taken and what would happen as they 11 progressed, so we were in the best position if 12 they got worse to take care of them. Occasionally 13 you can have a seizure occur right out of the blue 14 on arrhythmia. And then the second thing was we 15 were sure that for the time we had them in the ICU 16 that it would be very hard for them to kill 17 themselves; nobody ever did. So they weren't 18 necessarily an extremist in terms of being 19 critically ill except for the fact that their 20 admitting diagnosis was that they had had an 21 overdose. 22 Q Well, but if twenty-five 23 percent of the ICU beds were being occupied by 24 individuals who had taken an intentional overdose, 600 1 then I would suspect that that would be a -- can 2 we say significant medical problem that you were 3 addressing? 4 A Yes, sir. It also wasn't 5 the occupancy rate, because we also tracked the 6 occupancy minute by minute because many of these 7 people had relatively short term admissions to the 8 unit, and when I ended up my career in '81, we 9 were running average unit stays of, I think, 10 around five or five and a half days, whereas most 11 of the overdoses would probably be there for a 12 day, and some of them two days or three days, but 13 very, very few would be there for more than three 14 days, but overall occupancy must have been closer 15 to ten percent. 16 Q All right, but at any one 17 time, if you went in an intensive care unit, 18 twenty-five percent of the beds would be occupied 19 by individuals who had attempted suicide? 20 A No, I think that in terms 21 of admissions, what I meant by the twenty-five 22 percent is of the patients admitted to me, 23 twenty-five percent of the patients admitted would 24 be overdose, but if you went in at any time on the 601 1 average, you would find about ten percent of the 2 people resident at that time in the ICU would be 3 there because of an overdose. 4 Q But even at ten percent, 5 that's a serious medical problem, is it not? 6 A Oh, yes, absolutely; I 7 mean twenty-five thousand people a year at least 8 in America kill themselves every year. It's a 9 very serious medical problem. 10 Q Did you do any studies or 11 follow up to determine how many of the individuals 12 that you treated were suffering from the medical 13 diagnosis of depression? 14 A Yes, sir, that would 15 ordinarily be a diagnosis made by the psychiatric 16 consults, or it might be in their history from a 17 psychiatric thing, but -- 18 Q Previous admission. 19 A But I would say carrying 20 the actual diagnosis of major depressive disorder 21 or another psychiatric depression equivalent, 22 probably at least a third. 23 Q So in your experience, a 24 third of the individuals that you treated for 602 1 intentional overdose were depressed? 2 A Carried, yes, a formal 3 psychiatrically valid diagnosis of depression. 4 Q Did you make any 5 determination concerning whether or not the other 6 two-thirds were suffering from any mental 7 disorder? 8 A Oh, yes. I mean, many of 9 them had other kinds of mental illnesses, and the 10 reason I was trying to distinguish the psychiatric 11 major depression is that some of these people had 12 what I would call short-term exogenous depression 13 rather then the major depressive disorder defined 14 by DSM-3R. But I mean I had people who were 15 psychotic, I had people who were personality 16 disorders, people with obsessive compulsive 17 disease, a whole variety of people. 18 Q So in your experience of 19 the seven hundred and fifty patients that you 20 treated, only a third would have met the DSM-3R 21 diagnosis for major depression? 22 A That's my best memory. 23 Q And the other two-thirds 24 would have fallen into a variety of categories? 603 1 A Yes, sir. 2 Q Would there have been 3 some that would not have been diagnosed as 4 mentally ill or having a mental disorder? 5 A Yes, sir. You got to be 6 careful I guess with what mental disorder means. 7 Anybody who intentionally tries to hurt themselves 8 I would say can't be classified as normal, but 9 there would certainly be some that didn't carry a 10 diagnosis -- a specific diagnosis of OCD or 11 anxiety or even exogenous depression. 12 Q What percentage of those 13 patients would you see that wouldn't fit into any 14 established definitions of any particular 15 disorder? 16 A Probably around maybe as 17 many as a quarter who were signed out as having 18 sort of a situational reaction. That would 19 include exogenous depression or -- in other words, 20 people who didn't require long-term psychiatric 21 follow-up or care, just had had a bad situation 22 that led to this event. 23 Q Have you read anything 24 that confirmed or was different about these 604 1 findings concerning the percentage of individuals 2 who were depressed in intentional overdoses in 3 your patients as compared to what was found in the 4 general medical literature? 5 A Yes, sir, there would 6 have been in the general medical literature a 7 higher proportion of people carrying the 8 definition of depression, but I think that was 9 because of the intake criteria, because remember 10 we put everybody with an overdose or intent to 11 harm themselves that didn't require admission 12 elsewhere, we put them in the ICU, so I may have 13 had a more -- a broader definitions, if you would. 14 Q All right. Are you 15 saying that in your practice in the hospitals in 16 which you were affiliated, that there was a 17 requirement that if a patient came in even with an 18 intentional overdose, even though that patient's 19 vital signs might be stable, that they would be 20 transferred and admitted into the intensive care 21 unit? 22 A Exactly. If they didn't 23 go to surgery or psychiatry or some other -- or we 24 could have transferred them to another inpatient 605 1 facility, but they didn't go home from the 2 emergency room. 3 Q All right. 4 A As long as people thought 5 that it was legitimate, because some people come 6 in the emergency room and you don't believe their 7 story of having taken things, so those could go 8 home, but if the physicians in the emergency room 9 actually expected that they did take something, 10 then we admitted them at least overnight. 11 Q Now, did you study, 12 during that period of time, individuals who had 13 attempted suicide by means other than drugs? 14 A Yes, sir. 15 Q And by that I mean 16 prescription medicines and over-the-counter 17 medicines. 18 A Yes, sir,. 19 Q But I'm not including, 20 say, like cyanide or some type of insect repellent 21 or something of that nature. 22 MR. FREEMAN: What are 23 you including, like guns or shooting or -- 24 MR. SMITH: Yes, I'm 606 1 trying to distinguish between an individual taking 2 either a prescription medicine or a medicine such 3 as aspirin or some over-the-counter antihistamine 4 versus somebody taking cyanide or somebody taking 5 a gun to themselves. 6 A Well, I think there are 7 three or four studies that fall into that 8 category. First of all during the time that I was 9 both at Hopkins and at Case Western Reserve we 10 were tracking the patients admitted to the 11 hospital, the primary hospital I was in for 12 suicide attempts in general. Secondly, when I was 13 in Cleveland, I surveyed for a number of years the 14 Cuyahoga County Coroner's Office data in terms of 15 what they were seeing because obviously a lot of 16 people complete their suicide and never appear in 17 a hospital, so I was -- and that's a quite 18 different population of patients in terms of the 19 chemicals that you find. 20 And I also -- remember I 21 indicated that I had a government grant to do an 22 epidemiological study in Northeast Ohio where we 23 looked across all the hospitals in that region for 24 varying periods of time, reasonable periods of 607 1 time looking at everybody that came in the 2 emergency room who by any definition would have 3 been thought to have attempted to harm 4 themselves. I also used the data that was 5 collected by the government's program surveying 6 emergency room -- primarily emergency room nurses 7 for drug related incidents. Now, that's a whole 8 mixed bag, it includes overdoses and adverse 9 events and so forth, but I did survey those data, 10 and of course I was looking at all the stuff I 11 could find in the medical literature on the 12 subject. 13 Q All right. So were you 14 able to determine what -- if there was a 15 different -- what the different percentage was of 16 attempted suicides that were intentional overdose 17 versus other means, say non-drug means? 18 A Oh, yes, sir, if you 19 looked at Cuyahoga County Coroner's Office for 20 example, most of the people that they signed 21 out -- now, remember I didn't see them, but they 22 signed out as self-inflicted death, most of them 23 were carbon monoxide and some physical means, 24 hanging, shooting, et cetera. 608 1 Q What I'm wanting is a 2 difference in percentage. You know, can you -- 3 A You can look at my 4 chapters I've written on this and look up the 5 exact number, but -- I can give you a ball park 6 figure. 7 Q Right. For instance, 8 what I'm wanting to know really is, of attempted 9 suicides, what your experience of that percentage 10 that would fall into the overdose of drug 11 categories versus slashing wrist, shooting 12 themselves and things of that; I'm talking about 13 attempted as opposed to completed suicides. 14 A Oh, then you have to 15 leave out all the Cuyahoga County stuff. If you 16 look at attempts -- it varies both by gender and 17 by age, and if you look at attempts -- if you put 18 all those patients together, a big majority, a big 19 majority have taken drugs, either prescription or 20 over the counter, rather than some physical 21 means. One of the problems we have is that if you 22 use a physical means, you often succeed. 23 Q Now, of the completed 24 suicides, what is the percentage of drug overdose 609 1 versus other means? 2 A I think there it's closer 3 to -- and I don't know today; again, I'm speaking 4 primarily of the period in the '70s and 5 early '80s, I think it's closer to fifty/fifty, 6 but there's a problem in inferring intent because 7 when I was talking about the Cuyahoga County 8 Coroner's Office, for example, and carbon 9 monoxide, the vast majority of their deaths from 10 carbon monoxide were accidental rather than 11 purposeful, and they didn't -- their job was not 12 to make that distinction, so I just looked at all 13 the chemical related deaths and I'm not sure I can 14 give you an exact break-out on that. 15 Q What do you mean, 16 somebody might be intoxicated and pass out in an 17 automobile in a garage and -- 18 A More often space 19 heaters. I mean, those are real killers, and 20 you'd see six or eight people in the family, some 21 would be dead, some would have brain damage and 22 some would survive. 23 Q I would like to switch 24 subjects with you a little bit, Doctor Thompson, 610 1 and talk to you about side effects of 2 medications. Are there certain medications -- or 3 do all medications have some side effects? 4 A Yes, sir. 5 Q Is there any known 6 medication that doesn't have some side effect? 7 A No, sir. 8 Q Is there a difference, 9 medically speaking, concerning a medication -- a 10 medication side effect and an adverse effect of a 11 drug? 12 A Well, that gets into a 13 really arcane definition, and to be very precise, 14 an adverse event -- and I'm going to use primarily 15 an FDA regulatory definition -- is anything that's 16 untoward, therefore not desired, that occurs to a 17 patient after the intent to give them a drug. So, 18 for example, I remember we had an adverse event 19 report of a patient who had gotten a prescription 20 for Prozac and got hit by a bus on the way to the 21 pharmacy, and I've often lectured that if a person 22 has a divorce or an automobile accident, in fact 23 that has to be reported as an adverse event 24 because drugs can change your ability to drive or 611 1 make you dizzy and fall down or change your 2 personality and lead to a divorce. 3 Q But all medication is 4 going to have some side effects, is that right? 5 A Yes, sir. Now, when you 6 get into side effect, you get into -- you begin to 7 get into the question of causality, and most 8 scientists, and certainly the US regulators, do 9 not assign causality, at least in postmarketing 10 events at all because it's so difficult on an 11 individual case. So adverse event is anything 12 that happens, so every one of us has probably had 13 somebody intend to prescribe a drug and every one 14 of us will die, so there is a serious unexpected 15 adverse event which must occur in every human 16 being unless you have never seen a doctor. 17 Q Well, let me see if we 18 can get at it this way: My physician may 19 prescribe for me an antihistamine because I've got 20 a runny nose, all right? 21 A Yes, sir. 22 Q And then I begin taking 23 that antihistamine and notice that I have a 24 different feeling than I did before I started 612 1 consuming that antihistamine. My nose has quit 2 running within twenty-four hours. 3 A So you wanted that, so 4 that's not the adverse event. 5 Q That is, I would assume, 6 the effect that is desired by me as the patient 7 and the physician who prescribed the 8 antihistamine, is that correct? 9 A Yes, sir. 10 Q But I'm feeling real 11 drowsy and draggy, I just am -- you know, I just 12 can't concentrate, I just feel lethargic, I just 13 want to lay on the couch and sleep, and I go to my 14 doctor and say I didn't feel this way, Doctor, 15 before I started taking this antihistamine, should 16 I have an EKG, has something gone wrong with my 17 brain, am I fixing to have a stroke or something? 18 And my physician says, no, Mr. Smith, what you are 19 describing to me is probably you're having a side 20 effect from the antihistamine that I prescribed 21 for you. Correct? 22 A Yes, sir, I understand 23 what you're saying. 24 Q So it is -- physicians 613 1 and patients experience this frequently, that is a 2 patient reports some other effect as opposed to 3 the direct effect that was intended by the 4 physician, correct? 5 A Yes, sir, and then the 6 physician has to sort out whether it's more likely 7 due to that drug or some herb that the patient 8 took, or whether it's just the underlying illness. 9 Q All right. And if the 10 physician told me, Mr. Smith, I think that the 11 reason you're feeling sleepy and draggy is because 12 of the antihistamine that you're taking, and when 13 he says because, he's giving to me his medical 14 judgment of what is causing my drowsiness that 15 didn't exist before I started talking the 16 medication, correct? 17 A Yes, sir. 18 Q So in that situation, is 19 it reasonable for a physician to assign causation 20 to the antihistamine for the drowsiness and 21 lethargy that I feel? 22 MR. FREEMAN: On a 1639 23 or how are you using that? 24 THE WITNESS: I think he 614 1 means just in talking with him as a patient. 2 MR. SMITH: Yes. 3 A That's commonly done, but 4 it's different from scientific causation, 5 establishing it. It's not unreasonable. You 6 always have tentative diagnoses. I mean, when you 7 see a patient, you say, well, you know, here's 8 what I've observed and here are five things that 9 might be causing that, and so I've got these 10 things and this is the most likely one, so in 11 conversing with you as a patient, I would say, 12 well, in my opinion the most likely cause of your 13 lethargy is you shouldn't have taken that 14 antihistamine. 15 Q Well, but I wanted my 16 nose to stop running, Doctor Thompson. 17 A Quibbling, but that's not 18 good therapy, but that's okay. 19 Q Beg your pardon? 20 A It's not good therapy, 21 but that's okay. 22 Q But here it's perfectly 23 reasonable for the physician to assign the cause 24 of my drowsiness because of the consumption of my 615 1 antihistamine? 2 A If I want to distinguish, 3 I mean even in my own care of patients, I might 4 well use those same words in telling the patient 5 that rather than going through a long litany of 6 well, there are five things it might be, you might 7 have lupus, you might be dying of cancer and so 8 forth, but the most probable one is this, but in 9 the same case, for the same patient, if I were 10 asked scientifically is that causally related to 11 this, I would probably do something quite 12 different. 13 Q What would you do? 14 A Well, I would say in most 15 cases that on an individual patient basis it's 16 almost impossible to establish causality, 17 certainly from the first exposure to a drug. 18 Q So are you saying on a 19 case-by-case basis, it's almost impossible to 20 assign causality? 21 A Yes, sir, that's what our 22 studies have shown, that's what's in the 23 literature over and over again, it's generally 24 accepted among people who are expert in this area. 616 1 Q Now, are you talking 2 about assigning causality from a scientific 3 certainty? 4 A Well, no, I don't mean 5 certainty, I just mean with high probability. Is 6 that fair? 7 Q How high a probability 8 are you talking about, Doctor Thompson; are you 9 talking about that ninety-five percent factor that 10 we've seen in connection with statistics and what 11 is generally liked to be -- what is generally 12 comfortable by scientists in assigning a 13 particular factor to a judgment? 14 A Well, you're asking me a 15 hypothetical, so there's a whole range of 16 answers. If I could, let me give you a couple of 17 examples. If the patient were experiencing a 18 very, very rare event -- let me give you a very 19 specific example, and that's phocomelia with 20 thalidomide where the babies are born with 21 shortened arms and legs. The reason that was 22 picked up is although that had been seen, it was 23 very, very rare, and so when individual physicians 24 began to see a number of patients that had the 617 1 same defect, it clicked and said gee, the 2 likelihood that I'm just seeing background 3 incidence, if I myself have seen a couple of these 4 and my friend over there has seen a couple, gee, 5 the probability of that just being the background 6 instance is relatively low. And then they looked 7 into it and they found a common element among the 8 women who had given birth to those babies. So in 9 that case you would have to say gee, the 10 probability is awfully high that in a patient with 11 phocomelia -- not a hundred percent, not certain, 12 but it's awfully high, so there it would be in the 13 point nine eight, point nine nine range of 14 probability. 15 Now, on the other hand, 16 as you know, there's been an awful lot of cases of 17 people who smoke and who have lung cancer and 18 which the plaintiffs have tried to establish that 19 that patient's lung cancer was due to that patient 20 smoking, and again you can say, well, now, 21 epidemiologically I can tell you what the odds are 22 because he had some percentage of all the lung 23 cancers that are related to smoking, but in fact 24 in this individual patient I would never testify 618 1 that I'm a hundred percent certain that this 2 patient's -- that tumor is due to that smoking. 3 And so that you would have to say again is in a 4 probability range -- reasonably high, but not 5 anywhere close to certainty. 6 And then you have things 7 that are very common. For example, in our studies 8 of people with obesity, something just over thirty 9 percent of people, whether on placebo or Prozac or 10 whatever, report headache at some time in the 11 six-week period, which is higher than in the other 12 diseases that we've studied. So -- and it makes 13 no difference what dose of Prozac they're on, it's 14 exactly the same. So in that case, if you said, 15 well, the patient took Prozac and, you know, every 16 day they took Prozac they got a headache, is that 17 causally related, I would say gee, we have a pile 18 of evidence that Prozac doesn't cause headache, 19 and therefore the probability is very low. 20 Q How do you explain the 21 fact that these people were getting a headache? 22 A Well, again, since it 23 occurred at exactly the same rate on placebo -- 24 very interesting data, in fact -- 619 1 Q Okay, so you're saying 2 you're looking at -- and I didn't mean to cut off, 3 and I'll let you get back to it, but I just want 4 to make sure I'm clear with respect to what you're 5 saying, you're saying one way you can do it is 6 look at -- if you have the benefit of a control 7 clinical trial, you can look at the comparison 8 rate among those individuals who are taking the 9 medication and those who are not? 10 A Exactly. 11 Q And if there's a 12 difference of a certain percentage that headache 13 did occur on and didn't occur on, then you can 14 assign some causality to the medication, is that 15 right? 16 A In the population, but 17 not necessarily in the individual. 18 Q All right. 19 A That goes back to the 20 lung cancer smoking dilemma. I mean, there isn't 21 any question that smoking increases the odds that 22 you'll get lung cancer, but in the individual 23 patient with lung cancer, no matter how much 24 they've smoked, it's not a hundred percent certain 620 1 that that tumor came about because of smoking. So 2 when you look at a population of patients, for 3 example there's more nausea in people with 4 depression given Prozac than placebo, so if you 5 were to ask me is nausea something that's more 6 likely to occur when you take Prozac, the answer 7 is yes, on a population basis. If you say here is 8 Mrs. Jones, Mrs. Jones has been taking Prozac for 9 three days and Mrs. Jones says she's nauseated, is 10 her nausea related to Prozac, I have to tell you I 11 don't have any idea, but depending on the 12 difference between the placebo group and the 13 Prozac group, I could begin to give you some 14 probability ranges of that patient's nausea being 15 related to the administration of that drug. 16 Q Well, okay, can you say 17 that if you look at all the patients that take 18 Prozac, that it's more likely that a patient who 19 is taking Prozac and becomes nauseous is related 20 to the consumption of Prozac versus some other 21 cause? 22 A Well, the fact that 23 there's a statistically significant difference for 24 depression and Prozac, not for other illnesses and 621 1 Prozac, means that there's a very complicated 2 relationship between the drug, the dose, the 3 duration and the patient's underlying illness, 4 because as I recall the data on obesity, for 5 example, there's not a significant difference. So 6 it's not a straightforward thing that you could 7 say this chemical is more likely than a placebo to 8 produce nausea in any given patient. 9 If you're talking about a 10 patient with major depression, I'd say the answer 11 is yes, Prozac is more likely than placebo to be 12 associated with nausea. If you're talking about 13 obesity, I'd say the answer is no, because the 14 best data we have is that it's not different than 15 placebo. And if you asked about a specific 16 patient, I'd tell you I can't be certain, but I 17 can give you some odds as to whether it's probably 18 either related or not related. 19 Q Okay, so where do you 20 have to go in order to assign causality to a 21 particular patient's nausea to Prozac versus -- 22 let's say a depressed patient, to Prozac versus 23 something else, what does the difference have to 24 be before you would say it's more likely to be the 622 1 Prozac that's causing this individual to be 2 nauseated than something else? 3 A Okay, for me to say more 4 likely, I would want to see statistically 5 significant, meaning an alpha probably of point oh 6 five or less on a control study comparing Prozac 7 with a reasonable comparator like placebo in that 8 defined population, assuming that there are no 9 other factors. I mean, there are obviously 10 illnesses and other drugs that also produce 11 nausea. 12 Q All right. But what 13 you're saying is that from a scientific 14 standpoint, it's hard to know in an individual 15 case, on a case-by-case basis, whether or not the 16 medication is causing a particular reaction? 17 A It's almost impossible 18 because there have been some pretty good 19 publications, even looking at rechallenge, and if 20 you looked at it superficially you'd say, well, 21 gee, if you keep giving the drug and every time 22 you give the drug an event occurs and it doesn't 23 ever occur in that patient until you give the 24 drug, and then the temporal relationship is 623 1 exactly the same, you give the drug and ten hours 2 later they get nauseated, you know, that's a 3 fairly high level of probability that there's a 4 relationship, but in fact there have been very 5 good publications to show that that same thing has 6 occurred where it's not been causally related, so 7 it's not even a hundred percent certain when you 8 see that in an individual patient. 9 Q But that's one of the 10 benefits, though, of a rechallenge is that you get 11 a greater likelihood that you're going to be able 12 to assign causation? 13 A Yes, sir, you move the 14 probabilities up higher in that individual 15 patient. 16 Q In your knowledge of the 17 clinical trials and of the adverse event 18 postmarketing data on Prozac, is there any adverse 19 experiences that in your opinion are caused by 20 Prozac? 21 A Yes, sir, using the word 22 cause again as we've defined it -- 23 Q Yes, sir. 24 A -- that in a population 624 1 of patients given Prozac versus an alternative, 2 there would be a higher probability of the patient 3 having that event on Prozac than something else. 4 Q All right. 5 A Absolutely. 6 Q What are those, sir? 7 A Well, I haven't reviewed 8 this for some time, so I'm going to give you my 9 very best answer, but again it varies with -- in 10 fact, I've lectured on the fact that it varies 11 with diagnosis, dose, duration, age, gender and 12 probably race, so it's not an easy answer. If you 13 look at major depressive disorders, per se, in our 14 control trials, and this must be written down in 15 many of our reports, the ones that I would focus 16 on would be nausea, the patient reporting 17 difficulty in sleep at night, whether you call it 18 insomnia or something else. I think in the 19 depression trials, sweating and tremors come close 20 to statistical significance. As I recall in 21 obesity, for example, at sixty milligrams, I think 22 sweating does, but I'm not sure the others do. 23 Q Let's limit it to 24 depression, okay? 625 1 A Okay. I think those four 2 are the primary ones. 3 Q Nausea, sleep, sweating, 4 tremors; is that right? 5 A I think -- now, again, I 6 haven't reviewed this in some time, so that's the 7 best I can recall. 8 Q You don't include anxiety 9 or agitation or nervousness? 10 A Again, I'm not sure of my 11 memory. If I went back to notes we would find 12 explicitly because I'm really testifying what I 13 can remember from the data. The data are pretty 14 explicit. 15 Q All right, what is the 16 percentages of difference in patient populations 17 that causes you to attribute nausea to Prozac? 18 A Okay, I -- again, this is 19 off the top of my head because I haven't reviewed 20 this for a long time. As best I can recall from 21 the early studies, like in the NDA, the percentage 22 reported at any time during treatment was around 23 twenty percent of nausea with Prozac treated 24 patients at all doses versus something on the 626 1 order of fourteen or sixteen percent with 2 placebo. Now, I believe that in later trials of 3 depression, those -- the twenty percent number 4 became smaller, and it's certainly a lot smaller 5 with other illnesses. 6 Q Speaking solely of 7 depression. All right, so there was a six percent 8 difference there? 9 A Something on that order. 10 Q How about insomnia? 11 A I think that was -- I 12 think that was less of a difference, but maybe 13 four or five percent, something like that. 14 Q How about the sweating? 15 A The best I can recall it 16 was in that same ball park. 17 Q Four to five percent? 18 A Something like that. 19 Q And how about tremors? 20 A Again, we have to -- I 21 think that was significant, and I think it was in 22 the range of three or four percent, something like 23 that, difference. 24 Q And the reason you're 627 1 able to attribute those four general factors is -- 2 as to being Prozac related in depressed 3 individuals is because of a difference in 4 percentage that has been seen? 5 A Yes, sir. 6 Q And that percentage is up 7 to six percent in nausea, between four and five 8 percent in sleep and sweating, and three to four 9 percent in tremors, is that right? 10 A Well, it's not, as you 11 know, strictly the percentage, it's also related 12 to the number of individuals, meaning it's a 13 POISSON distribution and therefore it's related 14 both to the number that you see and the number 15 that could be seen. 16 Q All right. But that's 17 covered statistically by the design of the 18 protocol, is that right? 19 A Or the way it's analyzed. 20 MR. SMITH: Let's take a 21 little break. 22 (SHORT BREAK TAKEN.) 23 MR. BOUR: The time is 24 10:25. 628 1 Q (BY MR. SMITH) Doctor 2 Thompson, did I understand you to say before we 3 took our break that it's impossible to establish 4 causation of lung cancer to smoking? 5 A In an individual patient, 6 yes, sir. 7 Q How about heart disease? 8 A In an individual patient, 9 yes, sir. 10 Q How about complications 11 in pregnancy? 12 A Relationship to smoking? 13 Q Yes, sir? 14 A Well, there are some that 15 have been -- there's some that have in fact been 16 related epidemiologically to smoking, but in an 17 individual patient, I don't think it would be 18 possible to be certain. 19 Q Well, how about 20 epidemiologically, can it be an established that 21 long cancer is related to smoking? 22 A Oh, yes, sir, without 23 question. 24 Q How about heart disease, 629 1 can it be established that heart disease is 2 related to smoking from an epidemiological 3 standpoint? 4 A Yes, sir. 5 Q But you're saying you 6 can't prove any individual developed lung cancer 7 by virtue of smoking? 8 A That's right, you 9 couldn't say it with certainty. You could say 10 that there was more probability that a tumor in 11 someone who smoked was related to smoking, but you 12 couldn't say it with certainty. 13 Q All right. Could you say 14 that there is a risk of lung cancer associated 15 with smoking? 16 A Yes, sir. 17 Q Could you say there is a 18 risk of nausea in depressed individuals related to 19 use of Prozac? 20 A Yes, sir. 21 Q Could you say there is a 22 risk of heart disease in relation to smoking? 23 A Yes, sir. 24 Q And could you say there 630 1 is a risk of sleep disturbances as a result of 2 ingestion of Prozac? 3 A Yes, sir. 4 Q Could you say that there 5 is a risk of anxiety in connection with use of 6 Prozac? 7 A No, that you couldn't. 8 Q Because? 9 A Because when you 10 measure -- again, we're restricting this to 11 depression, right? 12 Q Yes. 13 A Well, there are two 14 pieces of evidence. First of all that in the 15 depressed patient population, when you measure 16 their anxiety scores by well validated anxiety 17 scales, the anxiety scores improve, so that 18 epidemiologically one would have to say anxiety in 19 the patient who is primarily depressed and treated 20 with Prozac is more likely to get better than to 21 get worse. 22 Q Can you say that there is 23 a risk that consumption of Prozac in individuals 24 will increase or exacerbate anxiety? 631 1 A No. Again, I mean, if 2 you look at the population of patients with 3 depression who have very well validated 4 measurements of their anxiety level, on average 5 their anxiety improves. So that 6 epidemiologically, if you pose the question in the 7 depressed patient given Prozac, is anxiety more 8 likely to get better or worse, the answer would be 9 we've got a lot of data that say it would be more 10 likely to get better. 11 Q Are there any 12 individuals, any depressed individuals that in 13 your opinion should not be given Prozac? 14 A Well, I've never heard of 15 an allergy to Prozac, but if there were such 16 individuals that are thought to have some sort of 17 immune reaction to it, I would say that would -- I 18 would not give Prozac to them. And I certainly 19 wouldn't combine it with certain other drugs. I 20 mean, I certainly would never use it together with 21 a monamine oxidase inhibitor or L-tryptophane. I 22 know that well respected psychiatrists sometimes 23 combine it with tricyclics, but I think you have 24 to be careful with that combination. So, I mean, 632 1 I think you can find individual patients who 2 hypothetically I would say this patient is not one 3 to use Prozac in, but I think I've encompassed 4 most of that. 5 Q That would be an 6 individual who had some type of allergy? 7 A If there's ever been one 8 reported; I'm not aware that there has. 9 Q All right. Or an 10 individual who is taking a monamine oxidase 11 inhibitor? 12 A Or L-tryptophane, or I 13 would certainly be careful if they were taking 14 tricyclics, and because of the drug-drug 15 interaction I would obviously be careful, not 16 because of Prozac but because of other drugs, if 17 they were taking other drugs. 18 Q If there -- if an 19 individual is not taking any other drug and is not 20 allergic to Prozac by virtue of some quirk of 21 their -- I guess it would be immune system? 22 A Yes, sir. 23 Q Then do you know of any 24 situation where it would be contraindicated to 633 1 give Prozac to a depressed individual meeting the 2 DSM-3R diagnosis? 3 A No, sir. 4 (THOMPSON EXHIBIT NO. 31 MARKED FOR 5 IDENTIFICATION.) 6 THE WITNESS: Yes, sir. 7 Q (BY MR. SMITH) Exhibit 8 31 is a document dated September 13th and 9 September 14th, in part authored by you and in 10 part authored by Doctor John Heiligenstein, is it 11 not? 12 A Yes, sir. 13 Q And that is a three-page 14 document with a Pz notation 542, 2101, 2102 and 15 2103, is it not? 16 A Yes, sir. 17 Q And this is what, Lilly 18 E-mail? 19 A Yes, sir. 20 Q To read it in 21 chronological order, you have to start with the 22 bottom of Page 1, do you not? 23 A Yes, sir. 24 Q And then go through all 634 1 of Page 2 and 3, and then come back to the top of 2 Page 1 to get your comments, correct? 3 A Yes, except the very 4 bottom of the entries on Page 3 would be my 5 signature to the forwarded message. So below the 6 end of forwarded messages, everything above that 7 line was what John Heiligenstein wrote and the top 8 of the first page and the Thompson, Leigh, RVAX 9 are my forwarding of that. 10 Q You're forwarding what? 11 A Forwarding it to somebody 12 else. In other words, this was a message that had 13 been sent to me, I had it electronically, I added 14 to it the comments that are at the top of the 15 first page, and the computer added the fact that I 16 was the author which is shown at the bottom of the 17 third page. 18 Q All right. To put this 19 in some context, apparently you had on the Monday 20 before September 13th, 1990 given to Doctor Dan 21 Masica and Bob Zerbe and John Heiligenstein some 22 information for them to review, had you not? 23 A I don't remember exactly 24 who of that group I had given it to; the best I 635 1 can recall it was Beasley, Heiligenstein, Masica 2 and Wheadon, and whether I actually gave it to 3 Gilad Gordon and Bob Zerbe or not I can't remember 4 specifically. 5 Q If you look at the 6 September 13th notation, Doctor Heiligenstein says 7 Charles, David and I have reviewed your 8 presentation and offer the following 9 thoughts/comments, correct? 10 A Yes, sir. 11 Q So that would be Charles 12 Beasley, David Wheadon and John Heiligenstein? 13 A Yes, sir. 14 Q And Heiligenstein is 15 authoring, I assume, their thoughts in connection 16 with the information that you've asked them to 17 review? 18 A Yes, sir, exactly. 19 Q And the information that 20 you had asked them to review was some information 21 that you were going to use to present to the board 22 of directors of Eli Lilly and Company, is that 23 correct? 24 A Yes, sir. 636 1 Q And additionally, 2 apparently you had shared some of that information 3 with Mr. Wood before the actual presentation to 4 the board, is that right? 5 A I don't know whether that 6 was before or after. I can't recall. I mean, I 7 obviously distinguished the verbatims, but I don't 8 recall whether that was done before or after. 9 Q After -- 10 A The board meeting. 11 Q You say the slides are 12 made, I showed them -- the board meeting, by the 13 way, was on September 17th. 14 A Oh, okay. 15 Q And I'll give you a copy 16 of the minutes of the board meeting to show that, 17 then I'll discuss it with you in just a second, 18 but for your help in reviewing and answering my 19 questions in connection with Exhibit 31, the board 20 meeting was on September 17th. 21 A Thank you. 22 Q So when you say the 23 slides are made, I showed them yesterday 24 afternoon, the proofs have been shown repeatedly 637 1 to Wood, et al., and I'm not sure that this 2 date -- at this date changes can be made in the 3 slides before Monday. 4 A Yes, sir. 5 Q Then that would indicate 6 that you had in fact shown some of this data to 7 Mr. Wood, who at that time was chairman of the 8 board of Eli Lilly and Company? 9 A Yes, sir. 10 Q Prior to the actual board 11 meeting itself, correct? 12 A Yes, sir, but the -- what 13 I'm not sure of is what the verbatims were in 14 terms of the informational content. 15 Q I understand, I 16 understand. But you had shown some slides -- 17 A I'd shown some slides, 18 showed some verbatims or some data or something to 19 Mr. Wood for sure. 20 Q Was this normally your 21 practice in making presentations to the board of 22 directors of Eli Lilly and Company, to review that 23 presentation that you were going to make with 24 Mr. Wood or some members of his staff? 638 1 A More often than not. 2 Q I guess Mr. Wood would 3 have the final responsibility for the agenda that 4 the board was going to review? 5 A Yes, sir, and Mr. Wood 6 was usually very concerned about how long it would 7 take because I usually took more minutes than he 8 wanted me to, so -- 9 Q So you wanted to make 10 sure that you and he were together, if for nothing 11 else to make sure that you wouldn't exceed the 12 time? 13 A Yes, sir. 14 MR. FREEMAN: It's my 15 same concern with somebody in the room. 16 Q The data and the material 17 verbatims and slides all had to do with Prozac, 18 did they not? 19 A I think so. 20 Q And all had to do with 21 the issue of Prozac and suicidal violent 22 aggressive behavior, did they not? 23 A In one way or another. I 24 mean, I don't have any really independent memory 639 1 of exactly what all the content was, and often 2 when I talked to the board, we'd go through all 3 the safety stuff, so I think you're correct. 4 Certainly it's a central focus of it, that was 5 correct. 6 (THOMPSON EXHIBIT NO. 32 MARKED FOR 7 IDENTIFICATION.) 8 Q Let me go ahead and give 9 you Exhibit 32 so you can see what was reflected 10 in the board meetings concerning specifically what 11 your discussions. By the way while they're 12 looking at that, have you seen Exhibit 31 before? 13 A Yes, sir. 14 Q And have you reviewed it 15 in preparation for your deposition here today? 16 A Yes, sir. 17 MR. FREEMAN: For the 18 record, there's nothing reflected in the document 19 about violent acts. 20 MS. ZETTLER: I object to 21 side bar comments and instructions by the attorney 22 by his comments. 23 A Yes, sir. 24 Q Why don't you go ahead 640 1 and identify, Thompson Exhibit 32 is apparently a 2 portion of the minutes of the board of directors 3 of Eli Lilly and Company, is it not? 4 A Yes, sir. 5 Q And I'll tell you, for 6 your benefit, that these are documents that also 7 bear Pz numbers which have been supplied to us by 8 the Lilly Company. 9 A Yes, sir. 10 Q All right. So that you 11 can know their authenticity, or lack thereof. 12 Back to Exhibit 31, 13 apparently you have asked Charles Beasley, David 14 Wheadon and John Heiligenstein to review this data 15 on Prozac and suicidality prior to your submitting 16 it to the board of directors, have you not? 17 A Yes, sir. 18 Q Do you recall why you did 19 that? 20 A I always do that. 21 Q Why do you always do 22 that? 23 A Because I haven't been 24 primarily responsible for the data, and in this 641 1 particular case I especially did that because 2 there were a large number of people that were 3 involved in looking at different aspects of the 4 data and I wanted to make sure that I reflected 5 not just my own comments, but a consensus when I 6 presented it to the board. 7 Q You wanted their input, 8 as I believe you put it, and their expertise into 9 the presentation that you were going to make to 10 the board? 11 A Yes, sir. 12 Q You were responsible for 13 making the presentation to the board, but they 14 were responsible for having specific information 15 concerning the data that underlined the 16 information that you were presenting to the board, 17 correct? 18 A Yes, sir. 19 Q And they had more 20 hands-on experience with this subject of Prozac 21 and suicidality and violent aggressive behavior 22 than yourself? 23 A Yes, sir. 24 Q Go back to Exhibit 31. 642 1 It appears that Doctor Heiligenstein has said in 2 connection with Heiligenstein's, Wheadon's and 3 Beasley's review of Verbatim 4, quote: We feel 4 that caution should be exercised in a statement 5 that, quote within a quote, suicidality and 6 hostile acts in patients taking Prozac reflect the 7 patient's disorder and not a causal relationship 8 to PZ, end the quote within a quote. 9 Doctor Heiligenstein goes 10 on to say: Postmarketing reports are increasingly 11 fuzzy and we have assigned, yes, reasonably 12 related on several reports. Our physician letter 13 summarizes experiences to say that, quote within a 14 quote, ongoing analysis to date of the 15 prospective, randomized, double-blind depression 16 clinical trial data taken as a whole do not 17 support the hypothesis of a causal relationship 18 between fluoxetine and suicidality, end quote. 19 Correct? 20 A Yes, sir. 21 Q Did you have a statement 22 in Verbatim 4 that said suicidality and hostile 23 acts in patients taking Prozac reflects the 24 patient's disorder and not a causal relationship 643 1 to Prozac? 2 A I think the minutes of 3 the board probably reflect pretty accurately what 4 I told the board. 5 Q No, my question was did 6 you have a verbatim that said that. 7 A Probably, at least close 8 to those wordings, yes. 9 Q All right, and that's 10 what they were reviewing, correct? 11 A Yes, sir. 12 Q That verbatim. And their 13 report to you concerning the accuracy of the 14 statement suicidality and hostile acts in patients 15 taking Prozac reflect the patient's disorder and 16 not a causal relationship to Prozac was that 17 caution should be exercised in making such a 18 statement, correct? 19 A Yes, sir. 20 Q And that in fact 21 postmarketing reports were becoming increasingly 22 fuzzy, correct? 23 A Yes, sir. 24 Q And that they had 644 1 assigned, yes, reasonably related on several 2 reports, correct? 3 A The we may not mean just 4 those three psychiatrists, but we as Lilly, yes. 5 Q All right. In other 6 words, there were -- there was someone within the 7 medical component that the psychiatrists 8 themselves had made or individuals over whom the 9 psychiatrists were responsible who had assigned a 10 causal relationship or a reasonably causal 11 relationship to Prozac ingestion and reports of 12 suicidality? 13 A Well, to be absolutely 14 precise, some of those relationships were assigned 15 by people that did not report to those 16 psychiatrists, but the gist of what you said is 17 correct; someone at Lilly in a responsible 18 position had in fact made that determination. 19 Q That there was a 20 reasonable relationship between causality of 21 Prozac and suicidality and hostile acts, correct? 22 A In the -- the meaning of 23 the words has to do with the outside the United 24 States regulatory definition of those terms, but 645 1 that determination had been made, somebody checked 2 the "yes" box. 3 Q Well, someone had checked 4 "yes" in response to what the United States 5 regulatory authorities were requesting in 6 connection with whether or not the event of 7 suicidality and hostile acts was related to -- 8 reasonably related to ingestion of Prozac, right? 9 A You're dead wrong. 10 Q I'm dead wrong? 11 A You're dead wrong. 12 Q Well, I hope I'm just 13 wrong as opposed to dead wrong, but -- 14 A Well, if you can ever get 15 to certainty, we're very close. 16 Q All right. Where am I 17 wrong, sir? 18 A Because postmarketing 19 events under the Food and Drug Administration laws 20 and regulations do not in fact ask for any 21 assessment of causality. 22 Q Well, then is it my 23 understanding that Lilly themselves, in connection 24 with this, had made some assessment of causality? 646 1 A Yes, we were working with 2 a group call CIOMS, which is a spin-off of WHO, to 3 develop a worldwide uniform way of reporting 4 adverse events and European regulators in some 5 cases requested some causality assessment because 6 their laws required the reporting of reactions 7 rather than events, and so it was for outside the 8 United States reporting in some countries, for 9 their local regulations, that that assessment was 10 ever made. 11 Q All right, and in some of 12 those instances where you were making an 13 assessment of causality between Prozac and 14 suicidality and violent -- and homicidal -- and 15 hostile acts, to use their quote, you were 16 assigning causality to Prozac and the suicidality 17 and hostile acts? 18 A Well, suicidality and 19 hostile acts is the phrase, that's correct. 20 Q I want to be careful, I 21 didn't want to say violent aggressive behavior 22 when what they're saying is hostile acts. Do you 23 follow what I'm saying? 24 A Right, we agree, 647 1 responsible people at Lilly had in fact checked 2 "yes" under that entry. I was only correcting 3 you that that was not an FDA regulation, that we 4 didn't do it for the US. 5 Q Well, those responsible 6 people at Lilly you considered responsible, did 7 you not -- 8 A Yes, sir. 9 Q -- in performance of 10 their duties? 11 A Absolutely. 12 Q And the physician, you 13 considered that they should exercise their medical 14 judgment in connection with this, didn't you? 15 A Yes, sir. 16 Q And that if they were 17 saying, yes, reasonably related on several 18 reports, you would have expected them to make that 19 reasonable relationship based on their medical 20 judgment, correct? 21 A Yes, sir, and our 22 underlying standard operating procedures. 23 Q But they were primarily 24 required to exercise, and they were employed to 648 1 exercise their medical judgment, were they not? 2 A And instructions that I 3 gave them in regard to that judgment was to always 4 be extremely conservative and err on the side of 5 saying something was causally related when it was 6 a judgment call. 7 Q Well, you expected them 8 to make good medical judgment, did you not? 9 A I did, but there's 10 obviously a huge gray area because I've already 11 testified that in fact good scientific evidence 12 exists that you can't make that kind of call 13 reliably, and we can go on for days about the 14 science that exists here, but I can tell you that 15 we've gotten into -- I personally have gotten into 16 a lot of trouble with regulators for being on the 17 side in that gray area of reporting more things by 18 making more conservative kinds of judgments than 19 the regulators felt were appropriate. 20 Q Whatever the regulators 21 thought, it's my understanding -- and correct me 22 if I'm wrong -- that in fact what was occurring 23 was that psychiatrists at Eli Lilly and Company 24 were assigning a reasonable relationship to the 649 1 use of Prozac and suicidality and hostile acts; is 2 that correct, sir? 3 A Yes, sir, some of these 4 individual reports they had checked "yes" under 5 reasonably related. 6 Q All right. Did you 7 report that to the board of directors at Eli Lilly 8 and Company in your presentation to the board on 9 September -- what is it, the 17th? 10 A You said it was the -- 11 yes, 17th. I don't remember. 12 Q You don't remember 13 whether you reported it? 14 A I don't remember whether 15 I reported it. 16 Q Well, is what you 17 reported, you think, accurately reflected by the 18 minutes of the board of directors meeting? 19 A I think the general gist 20 of it, absolutely. I haven't seen this before, 21 but I think it has it. 22 Q Is there a court reporter 23 present at the board of directors meetings? 24 A No, sir. 650 1 Q Is there a secretary of 2 the board -- 3 A Yes, sir. 4 Q -- there that takes 5 notes? 6 A There is a secretary of 7 the board, but I'm not aware that he takes notes 8 during the presentations. 9 Q Is there some individual 10 brought in to make notes of what is said and what 11 occurs at the board of directors meetings? 12 A Not in those segments of 13 the meetings to which I have appeared. 14 Q All right. Are the board 15 of directors meetings recorded in any manner? 16 A I don't have any idea. 17 Q You haven't seen any 18 recording devices such as we've got here, video 19 cameras, audio devices or anything of that nature? 20 A No, sir. 21 Q Did you discuss with 22 Doctor Heiligenstein or Doctor Beasley or Doctor 23 Wheadon this issue when or subsequent to authoring 24 the September the 14th portion of this exhibit and 651 1 prior to your meeting with the board of directors? 2 A I don't have specific 3 knowledge of what happened those three days. I 4 mean, I can't tell you exactly who I talked with. 5 Q But you don't have a 6 recollection of seeing this writing here and not 7 only going to your computer and typing up your 8 response to their response, which was some chagrin 9 at the amount of time that had elapsed between the 10 time you had given them this task and they 11 reported to you. Other than that, you don't 12 recall going -- leaving your office and going and 13 specifically discussing this issue with them? 14 A I don't remember. 15 Q Do you remember Doctor 16 Beasley, Wheadon or Heiligenstein coming to you 17 and apologizing for the delay in getting the 18 information back to you? 19 A No. 20 Q Saying, seriously, maybe 21 seeing you at the cafeteria or something and 22 saying, hey, Leigh, I'm sorry we were late in 23 getting that to you, I hope it didn't put you in a 24 bind or anything of that nature? 652 1 A No, I can remember a 2 whole lot of conversations on this issue, but you 3 asked me specifically between the Friday morning 4 when I typed this and the midday Monday -- 5 Q Yes, sir. 6 A -- and I can't remember 7 what happened on those three days. 8 Q Are you saying that there 9 may have been discussed -- some discussions, you 10 just don't recall -- 11 A Yes, sir. 12 Q -- during the three days? 13 A That's correct. 14 Q But you certainly don't 15 recall what they are? 16 A That's correct. 17 Q There may not have been 18 any discussions during that period of time? 19 A That's correct, I can't 20 remember those three days specifically. 21 Q Do you remember after 22 September 17th, after the board meeting, having a 23 discussion with either Doctor Beasley, Wheadon or 24 Heiligenstein, the subject of the fact that they 653 1 had assigned, yes, reasonably related on several 2 reports of suicidality and hostile acts? 3 A Oh, there were a number 4 of conversations about that. 5 Q Afterwards? 6 A See, that's my problem, I 7 can't remember if it was before this time, during 8 those three days or after this time, but there 9 were a whole series of conversations with a whole 10 lot of people. 11 Q Did you ever criticize 12 Doctor Beasley, Wheadon or Heiligenstein, or any 13 member of their staff, for assigning reasonably 14 related on several reports of causality of Prozac 15 usage and suicidality and hostile acts? 16 A Probably, but I can't 17 remember it specifically. 18 Q You criticized them for 19 making that assignment? 20 A Well, I need to put it in 21 a more general context, and that is that from the 22 time I first came to Lilly, when we first studied 23 this issue, I think everybody at Lilly will tell 24 you that my position was that it was virtually 654 1 impossible to assign causality on an individual 2 case, and therefore when regulators required us to 3 do that, we would do it very conservatively, but 4 this was a major issue because very good 5 scientists looking at the same material will not 6 agree on causality. That's been proven over and 7 over and over again. We did a study at Lilly to 8 prove that. So, again, it's an extremely 9 difficult issue and I'm engaged all the time in 10 discussions around the world about causality 11 assessments. 12 Q What study was done? 13 A When I first got to 14 Lilly, and it was probably either late '82 or 15 maybe it could have been in early '83, because of 16 concerns about causality assignments, both by 17 investigators doing trials and by Lilly 18 physicians, I pulled -- and me because I was a 19 principal investigator, but we had other people 20 working on it, we pulled a bunch of charts that 21 were on different drugs -- and I don't think 22 Prozac was one of the drugs, but I'm not a hundred 23 percent sure, I can remember some of them -- where 24 the event being determined was one that had in 655 1 fact occurred commonly with that drug, so we 2 didn't necessarily have a controlled trial with 3 statistical significance, but nonetheless we're 4 talking about an event that had been seen a number 5 of times with the drug. And then we had Lilly 6 physicians look at the same events and make a 7 causality judgment, and this is the same thing 8 that's been done over and over again in the 9 literature, and the agreement among us, good Lilly 10 physicians studying the same record, the agreement 11 among us was terrible, just as it's been every 12 time this has been studied. 13 Q Isn't what you're saying 14 is it basically there can be -- can be legitimate 15 disagreement concerning causality in connection 16 with a specific instance of a specific adverse 17 event? 18 A Exactly. 19 Q And that doesn't mean, 20 though, that one person is necessarily right or 21 one person is necessarily wrong in assigning that 22 causality, correct? 23 A That's correct. 24 Q So does that mean that 656 1 there might have been instances where one of the 2 psychiatrists at Eli Lilly had assigned, yes, 3 reasonably related to an instance of suicidality 4 or a hostile act and you would have agreed that 5 that assignment of, yes, reasonably related was 6 correct? 7 A My -- and this is very 8 well known, my feeling in general with adverse 9 events is that we should avoid causality 10 assignments totally on individual event reports, 11 and I would couch that kind of assessment in 12 probabilistic terms rather than saying it is or 13 isn't probably related. 14 Q Well, my question was did 15 you ever agree with one of the psychiatrists at 16 Lilly in their assignment of a particular adverse 17 event of suicidality and a hostile act as being 18 related to the use of Prozac? 19 A I don't recall ever 20 agreeing with that. 21 Q Did you ever disagree on 22 a particular instance of them recording, yes, 23 reasonably related to an instance where 24 suicidality and hostile act occurred while a 657 1 patient was on Prozac? 2 A Yes, sir, I'm sure I did, 3 but I can't recall the specifics of that event and 4 that drug because I took the same stand with all 5 events and all drugs. 6 Q That was just your 7 opinion on how the company should react in 8 considering postmarketing events with respect to 9 that particular situation? 10 A I thought previously we 11 were talking about scientific evidence as opposed 12 to fulfilling regulatory requirements, because 13 they diverge. 14 Q I understand. 15 A My feeling as chief 16 scientific officer is the company ought to comply 17 with all laws and regulations worldwide, but was a 18 scientist, if you asked me would I assign or agree 19 with an assignment of causality in an individual 20 case, my answer would be no. So, I am sometimes 21 forced to do things by regulations with which I do 22 not agree as a scientist. 23 Q But if something -- even 24 though you don't agree with it as a scientist, 658 1 regulatory bodies or legislative bodies of a 2 particular governmental entity, be it within the 3 United States or some foreign country, have in the 4 public interest required the manufacturer to do, 5 correct? 6 A That's right, they don't 7 want a whole bunch of adverse event reports, they 8 want to cull out those that aren't very important. 9 Q For whatever their 10 reasons, that's something that the company -- that 11 the governments have required manufacturers to do? 12 A Yes, sir. 13 Q Go to -- well, actually 14 you could have made the assessment to satisfy a 15 governmental regulation that an event was not 16 reasonably related, could you not? 17 A That's also fairly 18 difficult to do. For example, I mean, I -- 19 Q Is it equally difficult 20 to do? 21 A Yes, you know, I think 22 that either certainty or in either direction with 23 causality is awfully difficult. I mean, you can 24 say that this adverse event has never before been 659 1 seen in relationship to this drug, but I would 2 still maintain that that doesn't make it 3 impossible. The German drug law, for example, 4 says -- this is the best I can translate it -- 5 that unless there is scientific proof -- and I 6 think the word really translates to proof -- of 7 another cause that's more likely, then you have to 8 have presumptive evidence that the drug is 9 causally related. And when I've asked the BGA 10 what happens if the patient is on ten drugs, you 11 get different answers from different people as to 12 what you're supposed to do. But the French have a 13 beautiful algorithm of pharmacovigilance to decide 14 the level of relationship between a drug and an 15 adverse event. I mean, it varies from year to 16 year, from country to country. 17 Q And you're just saying, 18 from your scientific opinion, you don't think you 19 can assign or dispute causality to a particular 20 act in a particular drug in any particular 21 situation? 22 A With certainty. I think 23 you can assign probabilities to the relationship. 24 Q So you couldn't say with 660 1 certainty that Joseph Wesbecker's actions at the 2 Standard Gravure plant on September 14th, 1989 3 were related to the fact that he had consumed 4 Prozac? 5 A You know, the facts that 6 I know in that case are largely from the lay 7 literature from whenever it happened, so I don't 8 want to testify on what they were because I don't 9 remember very much about it. 10 Q Well, you talked to the 11 coroner, didn't you, Doctor Thompson? 12 A Yes, I did. 13 Q So I believe you can 14 answer my question, and we can have the court 15 reporter read the question back to you, if you'd 16 like. 17 A No, I remember it very 18 well. I would say that it would be impossible for 19 a scientist to say with certainty that it either 20 was related or was not related. 21 Q My question was: You 22 can't say with certainty that it was not related 23 to his use of Prozac, can you? 24 A You can, from an 661 1 epidemiological standpoint, assign -- may I 2 finish -- you can assign probabilities, but with 3 certainty you cannot. 4 Q Turn to Page 2 of Exhibit 5 31. 6 A Yes, sir. 7 Q And we talked about the 8 clinical trials. Doctor Heiligenstein says, Page 9 5 -- under -- I'm reading immediately on top of 10 that gray area in the middle of the page. 11 A Yes, sir. 12 Q Page 5: Under suicidal 13 thinking in clinical trials. You may want to note 14 that trials were not intended to address issue of 15 suicidality. Also in Paragraph 2, patients were 16 excluded who were serious suicidal risks. One 17 patient who made an attempt was randomized with no 18 further attempts on fluoxetine. Correct? 19 A Yes, sir. 20 Q I take it from reading 21 that, and Doctor Heiligenstein has sworn in his 22 deposition that it was his intent in making this 23 statement to advise you that the clinical trials 24 were not intended to address the issue of 662 1 suicidality. 2 A Yes, sir. 3 Q Is that your perception; 4 do you agree with that, that the clinical trials 5 were not intended to address the issue of 6 suicidality? 7 A Well, one, I agree that 8 that was the meaning of Doctor Heiligenstein's 9 message to me. Two, I agree that that was not the 10 primary intent of those trials. But three, there 11 were valid collections of data that addressed this 12 issue, as all other safety issues, in those same 13 trials. 14 Q But they only address 15 those issues indirectly? 16 A Well, the trial says it 17 was to collect data on efficacy and safety of the 18 product, since when you begin a new trial of a new 19 drug you don't have any idea what the safety data 20 might be. 21 Q Well, it wasn't intended 22 to assess -- the clinical trials were not intended 23 to assess -- address the issue of suicidal -- 24 suicidality, correct? 663 1 A That's what I said, the 2 primary purpose of the study was not to measure 3 suicidality, either positive or negative. 4 Q And in fact patients were 5 excluded from some clinical trials who were 6 serious suicidal risks, correct? 7 A Yes, sir. 8 Q Now, did you tell the 9 board of directors at Eli Lilly and Company that 10 the clinical trials were not intended to address 11 the issue of suicidality? 12 A I don't remember. 13 Q Did you tell the board of 14 directors that the clinical trials excluded 15 individuals who were a serious suicidal risk? 16 A I don't remember it 17 specifically. 18 Q Well, does Exhibit 32, 19 which is the board meeting, say anything about 20 your advising the board of directors that the 21 clinical trials were not intended to address the 22 issue of suicidality? 23 A No, but what I've said is 24 that I think this accurately reflects the overall 664 1 gist of what I said to the board, but I sure said 2 an awfully lot of things to the board, and don't 3 forget that I was reporting to board members, 4 Mr. Wood et al., all the time, so that many of the 5 members of the board had seen all of the data in 6 very great detail and been in meetings with far 7 more people than me. 8 Q Well, are you telling us, 9 Doctor Thompson, that you told some members of the 10 Board of Directors individually that the clinical 11 trials were not intended to address the issue of 12 suicidality? 13 A Well, whether I spoke 14 those words or somebody else from our group spoke 15 those words, members of the board who were meeting 16 with us very frequently, inside employees of Lilly 17 who were members of the board were in many 18 discussions about the design of the trials and 19 whether or not the trials had excluded certain 20 patients and so forth. Now, whether words came 21 out of my mouth or not, I can't remember. 22 Q How many members of the 23 board of directors at Eli Lilly and Company are 24 psychiatrists, medical doctor psychiatrists? 665 1 A I don't think any of them 2 are. 3 Q How many members of the 4 board of directors at Eli Lilly and Company on 5 September 17th, 1990 were psychiatrists? 6 A I don't believe any of 7 them. 8 Q How many of them 9 possessed a MD degree in September 1990? 10 A One of them. 11 Q And how many members of 12 the board were there, approximately, in September 13 1990? 14 A Thirteen to fifteen, 15 something like that. 16 Q Who was the one 17 individual member of the board at that time who 18 possessed a medical doctor degree? 19 A He is now president of 20 Purdue, at one time he was dean of the medical 21 school, and I'm blocking on his name. I'm sorry, 22 I'll think of it in a minute. 23 Q Was he a Lilly employee? 24 A No, sir, except as a 666 1 member of the board of directors. 2 Q Did you tell any members 3 of the board that some of the psychiatrists or 4 medical staff who were dealing with the 5 postmarketing reports of Prozac and suicidality 6 and hostile acts were determining -- were making 7 the determination, yes, reasonably related to some 8 of those reports? 9 A Well, whether the words 10 came out of my mouth or not, there were certainly 11 members of the board who were present in a number 12 of meetings where that was discussed directly with 13 the psychiatrists and other physicians that were 14 involved. 15 Q I need to know who the 16 members of the board were that knew this and who 17 the psychiatrists were that reported this, sir. 18 A Well, Doctor Perelman was 19 in a whole series of meetings, Doctor Herr was in 20 many of those same meetings where this issue was 21 discussed, and it was discussed by various members 22 of the staff, and I can't remember the specific 23 meeting or who was there. But the members of the 24 staff who would likely have participated in those 667 1 discussions were people like Doctor Zerbe, Doctor 2 Masica and the three psychiatrists we've been 3 talking about who were named on this memo. 4 Q Beasley, Heiligenstein 5 and -- 6 A And Wheadon. 7 Q -- Wheadon, is that 8 correct? 9 A Yes, sir. 10 Q Did you ever post the 11 question to any members of the board of directors 12 as to whether or not or how -- let's make it 13 whether or not an assessment should be made for 14 regulatory purposes of whether or not use of 15 Prozac was reasonably related to suicidality and 16 hostile acts? 17 A I don't remember whether 18 the discussion focused only on that drug and that 19 specific event, but there were a number of 20 discussions about how to fulfill the regulatory 21 requirements in European countries in general 22 and -- 23 Q I'm talking about that 24 drug in that event. 668 1 A Well, as a part of a set, 2 which would include all events and all drugs, it 3 absolutely was discussed many times with Doctor -- 4 Q I'm not talking about as 5 a part of a set of all drugs; I'm talking about 6 Prozac and suicidality and hostile acts and 7 reports of whether or not assigned causality for 8 purposes of regulatory requirements. Were there 9 any discussions with any members of the board by 10 you concerning whether or not that should be done? 11 A I don't remember any 12 discussions which focused specifically on that 13 drug and that adverse event in the terms of this 14 issue, but there were many discussions about this 15 in general, of which that's an example. And it 16 could have been one specific to it, I just can't 17 remember them. 18 MR. SMITH: I'd object to 19 everything after but as being not responsive to 20 the question asked. 21 Q Now, when you indicated 22 that you had given instructions to the Lilly 23 psychiatrists in connection with assigning 24 reasonably related or not reasonably related with 669 1 respect to causation of suicidality and Prozac and 2 had given them instructions to be -- what was your 3 term? 4 A Conservative in a gray 5 area, err on the side of assigning causality. But 6 if you're through with your question -- 7 Q No. 8 A Okay. 9 Q But you've been 10 successful in driving my brain from the question, 11 so I'm going to ask the court reporter to read the 12 first part of my question again. 13 REPORTER: (READING) When 14 you indicated that you had given instructions to 15 the Lilly psychiatrists in connection with 16 assigning reasonably related or not reasonably 17 related with respect to causation of suicidality 18 and Prozac and had given them instructions to 19 be -- what was your term? 20 Q Conservative, what 21 guidelines did you give them on how to be 22 conservative? 23 A Okay, first of all, I 24 don't recall specific instructions related to this 670 1 drug and those events as different from any other 2 drug and any other events. 3 Q All right. Are you 4 saying by that that you didn't give specific 5 instructions concerning Prozac and causality to 6 these individuals? 7 A No, I'm saying that I 8 don't recall that those instructions were any 9 different from or diverged in any way from the 10 instructions on all drugs and all adverse events. 11 Q I want to know what 12 instructions you gave in connection with Prozac 13 and its causality and the conservative guidelines 14 that you might have required that they use, 15 specifically concerning Prozac, specifically 16 concerning whether or not it was related to 17 suicidality or hostile acts. 18 A I had numerous 19 discussions, many, many discussions over many 20 years with probably every physician that worked in 21 my area, and the non-physicians who were assigning 22 causality, about the difficulties and the data of 23 doing this regulatory task as opposed to a 24 scientific task, and I coached them that we needed 671 1 to fulfill all of the regulations worldwide, and 2 we needed also to be sensitive to the scientific 3 dilemma of trying to assign causality, but that 4 when there was a gray area and medical judgment 5 was being exercised within a gray area, that in 6 some countries the regulations required us or 7 would best be served by our being very 8 conservative and therefore assigning more 9 causality than we would as scientists. 10 Q And those were 11 instructions that you gave these psychiatrists in 12 connection with Prozac and the issue of 13 suicidality and hostile acts? 14 A I'm saying I can't 15 distinguish in my memory any specific instructions 16 for that drug and those acts. 17 Q That's what I want to 18 know, what instructions or guidelines you gave 19 them with respect to Prozac and causality and 20 suicidal hostile acts. 21 A I don't remember anything 22 other than being in concert with what I had said 23 before, which was exactly the same guidelines as 24 with all drugs and all adverse events. 672 1 Q All right. Were there 2 any written guidelines in place? 3 A Yes. 4 Q All right. What -- where 5 are those written guidelines contained? 6 A Well, the first ones were 7 probably in the policies, practices and procedures 8 I wrote over the Christmas vacation between 1982 9 and 1983, and those policies, practices and 10 procedures were updated periodically and to my 11 knowledge they're still being used in whatever the 12 current version is. 13 Q Do you have a specific 14 recollection of what those guidelines state? 15 MR. DOWNEY: I missed all 16 of those years, you wrote them over the Christmas 17 vacations -- 18 THE WITNESS: I'm sorry, 19 between -- the Christmas vacation that spanned the 20 end of 1928 and the beginning of 1983 was the time 21 that I revised the standard operating procedures 22 of the Phase II/III medical group at Lilly and 23 wrote -- and I changed its name to policy, 24 practices and procedures, and that same document 673 1 was a living document that was -- after I was no 2 longer the specific author of it, it was updated 3 frequently by many people and is in fact the basis 4 of the standard operating procedure for the 5 medical group at Lilly, and I believe it's still 6 being used and I think it still has something like 7 that name. 8 Q (BY MR. SMITH) Well, did 9 it in 1990 give specific instructions and 10 guidelines concerning the assignment or 11 non-assignment of causation to a particular 12 adverse event? 13 A Yes, sir, I think so. 14 MR. SMITH: All right. 15 Want to take a quick break? 16 (SHORT BREAK TAKEN.) 17 MR. BOUR: This is the 18 start of Tape 8, the time is 11:37. 19 Q (BY MR. SMITH) Doctor 20 Thompson, does -- based on your experience as a 21 physician and your experience working with the 22 United States Food and Drug Administration -- 23 you've been asked to join them on two occasions, 24 correct? 674 1 A Yes, sir. 2 Q Does FDA approval of a 3 drug product insure that that product is safe? 4 A It insures that it has 5 sufficient evidence of safety and efficacy to 6 allow marketing in the United States. 7 Q But does that insure that 8 an individual might not have an adverse event as a 9 result of ingestion of an FDA approved drug? 10 MR. FREEMAN: We object 11 to the form of the question in that the standard 12 here is not insurance or insured anything, and 13 that's a legal concept. This witness is not 14 appearing as a legal expert. 15 MR. SMITH: Go ahead. 16 A Do you want me to try to 17 answer it? 18 Q Yes, he's just making a 19 legal objection. 20 A Yes, as I said -- 21 Q Your answer may not be -- 22 A As I said earlier, I'm 23 not aware that there's any drug that doesn't have 24 some adverse event associated with it, so whether 675 1 the FDA approves it or doesn't approve it, I would 2 be happy with my former assertion that every drug 3 ever known has some probability of causing some 4 adverse event in somebody. 5 Q And certainly FDA 6 approval of a drug does not mean that that drug 7 might be -- might not be dangerous to a particular 8 individual? 9 A I agree with your 10 statement, yes. 11 MR. SMITH: I apologize 12 for the delay. We may have a little noise in our 13 documents. Both from a statistical standpoint and 14 from a practical standpoint. 15 (THOMPSON EXHIBIT NO. 33 MARKED FOR 16 IDENTIFICATION.) 17 THE WITNESS: Yes, sir. 18 Q (BY MR. SMITH) Exhibit 19 33 is a document authored by Richard Huddleston, 20 is it not? 21 A Yes, sir. 22 Q It is directed to Hans 23 Weber, correct? 24 A Yes, sir. 676 1 Q Who is Richard 2 Huddleston? 3 A I think he's a CRA, I'm 4 not sure of his exact title, and I think at this 5 time he was working for Max Talbott. 6 Q In regulatory? 7 A I think so. 8 Q Exhibit 33 uses the term 9 CIOMS and you used the term CIOMS earlier. Could 10 you explain to us what CIOMS is, sir? 11 A Yes, sir, it's a 12 French -- it's an acronym for French words, and 13 I'm not sure of the exact words, but something 14 like the Centa International -- the O has 15 something to do with health, medical whatever; I 16 can't give you the exact translation. It is in 17 fact a body which is loosely affiliated with the 18 World Health Organization and in fact 19 headquartered near the World Health Organization, 20 and they do various things to improve world 21 health, and I think this reference and my previous 22 reference was to one specific activity that they 23 were doing that we were a part of from the 24 beginning, which was to get a uniform set of rules 677 1 for reporting an adverse event that occurs in one 2 country to another country. 3 Q All right. It's my 4 understanding that CIOMS -- it's not a regulatory 5 body itself, is it? 6 A No, sir, it's not. 7 Q CIOMS is in fact an 8 affiliation of individuals involved in health 9 issues throughout the world, is that right? 10 A To the best of my 11 knowledge. I only know a few people that have 12 been involved in this specific thing. 13 Q And one of the things 14 that CIOMS does is try to facilitate reporting of 15 adverse events of drug products from one country 16 to another? 17 A Yes, sir, that was the 18 specific thing we were involved in. 19 Q And isn't it correct that 20 the impetus for that in part came from the 21 pharmaceutical industry throughout the world? 22 A No, the actual two people 23 who began that particular initiative at CIOMS were 24 Win Castle, who was in fact at that time -- she's 678 1 Rhodesian, but she was a physician at ICI in 2 England and Gerry Faich who is a physician who at 3 that time was head of the Office of Statistics, 4 Epidemiology and Adverse Event Reporting at the 5 FDA. Those two people then got a group of us to 6 begin meeting to talk about this issue, and that 7 group was made up of more people from industry 8 than regulators, but included both industry, 9 regulators, and then I would call them neutral 10 observers from things like the World Health 11 Organization and the CIOMS staff themselves. 12 Q It's a voluntary -- 13 volunteer body, is it not? 14 A Yes, sir. 15 Q You don't have to 16 become -- be a member of CIOMS? 17 A Oh, no, sir. 18 Q And it's done with the 19 intent to improve, in this aspect, reporting of 20 adverse events worldwide? 21 A Yes, sir. 22 Q And Eli Lilly is one of 23 the pharmaceutical manufacturers that is a part of 24 that group? 679 1 A Yes, sir. 2 Q And they have together, 3 the pharmaceutical manufacturers and the 4 regulatory bodies and neutral observers as you 5 say, got together and set some standards for 6 reporting adverse effects in connection with drug 7 products? 8 A And continue to do so 9 today. 10 Q And were you a part of 11 the early formulation of CIOMS? 12 A I individually was part 13 of the original group. 14 Q When was that? 15 A I think it was around '85 16 or '86, but I can't be a hundred percent sure. 17 Q All right. And you were 18 an employee of Eli Lilly at that time? 19 A Oh, yes, sir. 20 Q And Lilly encouraged your 21 participation in that group? 22 A Yes, sir. 23 Q And I would assume that 24 you urged Lilly to participate in sharing 680 1 information with that group? 2 A Yes, sir. In fact, they 3 adopted our form; we sold them our adverse event 4 reporting form. 5 Q And the CIOMS adopted 6 Lilly's form -- 7 A Essentially -- 8 Q -- or Lilly adopted the 9 CIOMS form? 10 A No, they adopted the one 11 that we had evolved with the FDA. 12 Q All right. So Lilly and 13 CIOMS are closely associated in connection with 14 the manner in which adverse events were reported 15 internationally? 16 A We were a part of 17 evolving the CIOMS criteria, and in some countries 18 those in fact did get approved by regulators, so 19 they became part of local regulations. For 20 example, the FDA have approved that as one 21 mechanism of reporting, and Lilly was struggling 22 to be in concert with those CIOMS definitions. 23 Q But Eli Lilly in 1990 24 reported to CIOMS exclusively on a voluntary 681 1 basis? 2 A Let me make that clear. 3 To my knowledge, we've never reported an adverse 4 event to CIOMS. CIOMS created a standard for how 5 they were to be reported. But I'm not sure any 6 individual events ever got to CIOMS. This was a 7 method of reporting, not the individual events. 8 The events would go to individual regulators in 9 different countries. 10 Q But it's a standardized 11 method of reporting to international bodies? 12 A Yes, sir. 13 Q In other words, your 14 16 -- well, does your 1639 or does the FDA 1639 15 meet a CIOMS criteria? 16 A Well, it's complicated, 17 but we've modified and got permission from the FDA 18 to change the FD-1639 form itself, and we use a 19 form which is agreed both by the FDA and does meet 20 the CIOMS criteria as our reporting mechanism 21 worldwide. 22 Q All right. Now, with 23 that background, let's go to Exhibit 33. 24 Apparently Doctor Weber has directed some 682 1 questions to Richard Huddleston concerning Prozac, 2 suicide and CIOMS, correct? 3 A Yes, sir. 4 Q He says, number one: 5 CIOMS criteria includes reports that are serious, 6 unexpected, reasonably causally related. 7 Seriousness is defined by CIOMS serious criteria 8 and expectancy is defined by FDA criteria. 9 Literature reports qualify for CIOMS if the above 10 criteria are met. We classify suicide as expected 11 for Prozac. 12 A I think you mean 13 unexpected, don't you? 14 Q Yes, I'm sorry, thank 15 you. I'll continue with the quote: We classify 16 suicide as unexpected for Prozac, so if other 17 criteria are met, a suicide report would be mailed 18 via CIOMS. 19 Two, CIOMS -- suicides 20 are being monitored on a weekly basis by Doctor 21 Gilad Gordon in Indy. Increased frequency has 22 been monitored at the very least, quarterly, since 23 marketing, and the results published in the 24 Highlighted Events Compilation which is forwarded 683 1 to all affiliates each quarter. Correct? 2 A Yes, sir. 3 Q So do I take it by that, 4 that Doctor Weber is saying that -- well, actually 5 Rick Huddleston is saying to Doctor Weber that if 6 an event is serious, unexpected and reasonably 7 causally related, it will be reported via CIOMS? 8 A Yes, now, there -- to be 9 precise now, there are two qualifications you need 10 to understand: One is the initial CIOMS criteria 11 did not cover clinical trials, they only related 12 to spontaneous events. And the second thing is 13 the initial criteria only related to reporting an 14 adverse event that occurred at one country to 15 another country, rather than what you'd do within 16 your own country to report that event. 17 Q I understand that. 18 A Okay. 19 Q It's a standard for 20 reporting international reports intercountry? 21 A Yes, sir. 22 Q As opposed to 23 intracountry? 24 A At least initially, yes, 684 1 sir. 2 Q So if -- it's a 3 standardized procedure whereby an adverse event in 4 France can be reported to an entity in the United 5 States? 6 A Yes, sir. 7 Q And the requirements are, 8 in order that that event be reported, that it be 9 serious, unexpected and reasonably causally 10 related, correct? 11 A That fulfills the CIOMS 12 criteria, but of course different countries either 13 add to or subtract from those criteria. 14 Q I understand that. 15 A Okay. 16 Q But I'm talking about the 17 CIOMS criteria; I'm trying to understand what the 18 CIOMS criteria -- and they should be serious, 19 unexpected and reasonably causally related, 20 correct? 21 A Yes, sir. 22 Q And is the reason for 23 those three requirements so that you just won't 24 have all kind of adverse events at any time, 685 1 whether they bear any relation, whether they're 2 serious or nonexpected, reported? 3 A I think that's generally 4 correct, yes. 5 Q Wasn't the thought that 6 if -- that you needed to limit adverse events as 7 being reported via CIOMS -- 8 A Well, see, there's a 9 huge -- 10 Q -- in some way? 11 A There's a huge debate 12 about that because the FDA clearly doesn't agree 13 with some parts of that definition, and therefore 14 implies their own rules, and the German drug law 15 does as well. 16 Q I'm just talking about 17 CIOMS, though. Wasn't the rationale for the CIOMS 18 rules concerning reporting is that if you took 19 every report of every event, be it serious or 20 nonserious, expected or nonexpected, reasonably 21 related or not reasonably causally related, then 22 you would have so many reports that the 23 information transmitted via CIOMS would in fact be 24 meaningless; it was an effort to limit in some way 686 1 the amount of information concerning adverse 2 events that was going to be reported? 3 A Well, in there you have 4 the word meaningless, I don't agree with that. 5 And the second thing I don't agree with is that is 6 how the CIOMS criteria evolved, but the actual 7 discussion was to have one set of rules that would 8 work worldwide, and some of us were on one side of 9 the issue of how many to have and some were on the 10 other side, and we all agreed to a middle 11 position, which was as you've stated. 12 Q Was the three criteria? 13 A Yes, sir. 14 Q All right. Is what 15 Mr. Huddleston is doing is simply reflecting what 16 the criteria are? 17 A Under the CIOMS 18 agreements? 19 Q Yes. 20 A Yes 21 (THOMPSON EXHIBIT NO. 34 MARKED FOR 22 IDENTIFICATION.) 23 THE WITNESS: Yes, sir. 24 Q (BY MR. SMITH) It 687 1 appears that Exhibit 34 is a document dated 2 December 7th, 1990 from Doctor Hans Weber to 3 Doctor or Mr. Gordon Gilad? 4 A Doctor Gilad -- Gilad. 5 Q Doctor Gilad -- Gilad 6 Gordon, correct? 7 A Yes, sir. 8 Q Have you seen this 9 document before? 10 A Yes, sir. 11 Q When did you see this 12 document? 13 A Years ago. I don't think 14 this is one of the documents I reviewed in the 15 last month. 16 Q All right. This document 17 is talking about CIOMS and reporting suicides in 18 connection with CIOMS criteria, is it not? 19 A Yes, sir. 20 Q And there Doctor Weber 21 says that in quarter two of '90 we had forty-six, 22 four went to CIOMS. In quarter three, '90, we had 23 a hundred and twenty-two total, 19 of them CIOMS. 24 Correct? 688 1 A Yes, sir. 2 Q Does that mean, Doctor 3 Thompson, that those four that went to CIOMS in 4 quarter two of '90 were serious, unexpected and 5 reasonably causally related? 6 A Probably, I think that's 7 correct. 8 Q And does that mean that 9 the nineteen that went in the third quarter of 10 1990 had been determined to be serious, unexpected 11 and reasonably causally related to the use of the 12 drug? 13 A Sir, I think that box was 14 checked because that's how the computer would have 15 typed that out as a CIOMS report. 16 Q So is the answer to my 17 question yes? 18 A That that determination 19 had been made, yes. 20 Q That nineteen of the 21 hundred and twenty-two? 22 A I think that's correct. 23 Q So somebody at Lilly was 24 making a judgment in connection with that? 689 1 A Yes, sir. 2 Q And judged nineteen of 3 the one hundred and twenty-two as being reasonably 4 causally related to the Prozac? 5 A Yes, sir. 6 Q Is nineteen out of one 7 hundred and twenty-two a statistically significant 8 factor? 9 A Compared to what? 10 Q Nineteen compared to one 11 hundred and twenty-two. 12 A I don't know of any test 13 to -- I mean, that doesn't make -- I don't 14 understand how to do a statistical test on what 15 you're telling me. 16 Q If you had one hundred 17 and twenty-two reports of suicide on Prozac and it 18 was judged that nineteen of those were reasonably 19 causally related to the use of Prozac, as was done 20 with respect to these nineteen reports, would that 21 be a statistically significant factor -- number? 22 A I'm not -- I'm really not 23 playing a game, sir, there just isn't -- there 24 isn't a hypothesis being posed by which I 690 1 understand a test to be applied. I mean, if you 2 said if nineteen -- is nineteen out of a hundred 3 and twenty-two different than zero out of a 4 hundred and twenty-two or a hundred and twenty-two 5 out of a hundred and twenty-two, I'll try to 6 answer the question. But that nineteen had this 7 designation by whoever made it out of a hundred 8 and twenty-two simply isn't a hypothesis that I 9 can apply a statistical test to. I'm not trying 10 to be confusing -- 11 Q Can you say nineteen is a 12 statistically significant number as to one 13 twenty-two? 14 A Well, but the meaning of 15 statistical significance relates to the 16 probability that one event is the same or 17 different than another event, so I don't know how 18 to do a test on that. It's like saying that you 19 had a deck of cards and you're going to assert to 20 me that of a hundred and twenty-two cards, 21 nineteen of them are spades, is that expected or 22 unexpected? Well, without knowing what the deck 23 of cards is made up of, I have no way of answering 24 the question. 691 1 Q No, I guess my question 2 is, in your example, is it -- is the number of 3 spades in that deck statistically significant by 4 virtue of the fact that there's one hundred and 5 twenty-two in the deck and nineteen of them are 6 spades? 7 A We're having problems 8 because I can't -- I'm not aware that there is a 9 statistical test that could be applied to your 10 question as you've framed it. If you frame the 11 question differently, I can see a test that can be 12 applied. Want some help? 13 Q Yes. 14 A Okay, if you asked me is 15 nineteen out of a hundred and twenty-two 16 statistically different from zero out of a hundred 17 and twenty-two, I can say, with the usual kinds of 18 testing that you do, that would be a statistically 19 significant difference. So under the same 20 conditions with the same kind of tests applied to 21 the same kind of population, one time you find 22 nineteen out of an a hundred and twenty-two, the 23 next time you find zero out of a hundred and 24 twenty-two. If you say gee, is this just random 692 1 variation, that's not very likely, so it's 2 probably a statistically significant difference. 3 MR. SMITH: All right. 4 (THOMPSON EXHIBIT NO. 35 MARKED FOR 5 IDENTIFICATION.) 6 Q (BY MR. SMITH) Since 7 we're in the area of December 7th, 1990, I'll give 8 you another document. 9 A Yes, sir. 10 Q Exhibits 25 -- 11 A 35. 12 Q 35 is a document -- and I 13 can't tell who authored the document, can you? 14 A I think from the header 15 at the top, it's probably from Rick Huddleston. 16 Q All right, and it's to 17 Hans Weber? 18 A Yes, sir. 19 Q And it concerns suicide, 20 a BGA response, correct? 21 A Yes, sir. 22 Q The document is difficult 23 to read, but can you, by virtue of your knowledge 24 of the individuals at Lilly, give me who that 693 1 first word -- name might be? 2 A Claude Bouchy you mean? 3 Q Well, under that it says 4 A-N gave us -- 5 A Oh, I think that's can, 6 that would be my guess, because you can make out 7 the A-N, and it looks to me like one letter is cut 8 off on each line. 9 Q All right. 10 A Can give us unique -- 11 some unique identifier. 12 Q Can give us -- is it 13 saying Claude Bouchy can give us some unique 14 identifier -- no, that wouldn't work. 15 A I think it's a request to 16 Hans Weber to give an unique identifier for the 17 patient that Claude Bouchy had referenced. 18 Q And Claude Bouchy had 19 described, quote: A patient that you have that 20 committed suicide, had no concomitant meds, and 21 the reporting physician stated suicide was caused 22 by a surge of serotonin. According to Claude the 23 Indy monitor judged the report to be not related. 24 If you have the DEN -- something number? 694 1 MS. ZETTLER: Manu- 2 facturing number. 3 Q -- manufacturing number, 4 that would be great. Correct? 5 A Yes, sir. 6 Q Have you seen this 7 document before? 8 A I don't think so. 9 Q Have you heard of a 10 situation where a physician was reporting a 11 suicide without any concomitant medications that 12 was caused by a surge of serotonin that the Indy 13 monitor was judging to be not related to Prozac? 14 A I don't remember the 15 phrase "surge of serotonin" before, but there were 16 many instances of disagreements about reporting 17 related to those three criteria. 18 Q Well, have you ever seen 19 the term "surge of serotonin" before? 20 A I know what it means, but 21 I've not -- I don't recall ever seeing surge of 22 serotonin before as stated exactly that way. 23 Q Doctor Ray Fuller has 24 described to us that it's believed that one of the 695 1 actions of Prozac is that it causes serotonin to 2 remain present in the synaptic cleft at a greater 3 concentration than it would be ordinarily. 4 A Oh, I agree with that 5 statement. 6 Q Would that be 7 characterized as a surge of serotonin? 8 A I don't think so. 9 Q All right. Do you have 10 any idea why this event would be judged as not 11 related to Prozac when the reporting physician was 12 stating that the suicide was caused by a surge of 13 serotonin? 14 A I don't know what the 15 facts are in the case. 16 Q Have you ever read in the 17 medical literature or in your research of 18 individuals that have been seized by a surge of 19 serotonin before? 20 A Again, that particular 21 word I can't relate to; I know what you mean by 22 having excess serotonin effects, but I've not seen 23 surge used. 24 Q Has it been -- have you 696 1 seen reports of individuals who have had excess 2 serotonin? 3 A Yes, sir. 4 Q And in what instances has 5 that occurred? 6 A Well, it was actually 7 reported before Prozac was developed with other 8 drugs that were thought to work primarily on the 9 serotonergic system. I think that is a way to 10 characterize interaction between monamine oxidase 11 inhibitors and Prozac in some of those patients 12 that have had a bad -- an adverse event who took 13 both of those kinds of drugs. We tried to 14 reproduce that in animals, so that you can clearly 15 in rodents see what I would attribute to be fairly 16 called an excess of serotonin. Other people have 17 called it a serotonergic syndrome. 18 Q All right. And that's 19 where, for whatever reason, there is more 20 serotonin present in the brain than is desirable 21 or normal? 22 A That's the hypothesis, 23 sure. 24 Q And is that in fact -- 697 1 has that in fact been demonstrated in animals? 2 A I wouldn't want to 3 testify to that, I'm not sure of the exact data, 4 but it's a very good scientific hypothesis. 5 Q That that can occur? 6 A Yes, sir. 7 Q So you can have 8 situations where there is excess serotonin present 9 as well as situations where there's not enough 10 serotonin present? 11 A Yes, sir. 12 Q Is it your understanding 13 that serotonin levels vary among individuals? 14 A Yes, sir. 15 Q And that serotonin levels 16 vary within a particular individual during the 17 day? 18 A Serotonin levels measured 19 in the brain? 20 Q Or in the platelets, 21 either way. 22 A I'm sure there's some 23 variation, but I wouldn't want to testify 24 specifically what those data are. 698 1 Q In other words, your 2 serotonin level as it exists at this time may be 3 different than as it existed when you walked in 4 this building this morning? 5 A Oh, that's certainly 6 possible. 7 Q And what is your 8 understanding of why serotonin levels rise and 9 fall? 10 A I don't have any 11 understanding at all. 12 Q All right. You just know 13 that it's understood that it does fluctuate? 14 A You can certainly eat 15 precursors of serotonin, like L-tryptophane that 16 my guess would be would make it go up. You can 17 take drugs that will interfere with its 18 disposition or make it go down. I mean, I know a 19 few things that influence it, tumors that release 20 large amounts of serotonin and so forth. But in 21 terms of a circadian or daily or regular pattern 22 of it going up and down, I don't know the data on 23 that. 24 Q But there's not been any 699 1 measurement, as far as you know, of what a 2 standard serotonin level is in the brain, in the 3 human brain? 4 A Well, there is in people 5 who have died. I mean, there was a nice study by 6 Henry Bourne in '65 that looked at people who died 7 suddenly and compared depressed folk and people 8 that died without depression. It measured brain 9 serotonin levels, so there are data that speak to 10 that. 11 Q Yes, but there -- none of 12 that data spoke to the question of what is normal 13 versus what is normal and assigned a normal range 14 of serotonin? 15 A No, it did. I mean, in 16 people that were not depressed but who died 17 suddenly through trauma or whatever, presumably 18 those measurements would reflect what the normal 19 level of serotonin was in the brain. 20 Q What is the normal level 21 of serotonin, what's that level? 22 A I don't remember what the 23 concentration is. 24 Q But it's your testimony 700 1 that there has been determined to be a normal 2 concentration measurement in the brain? 3 A Yes, sir. 4 Q In other words, you could 5 send cerebrospinal fluid -- 6 A You've got to be very 7 careful there because cerebrospinal fluid 8 concentration of serotonin and its metabolites may 9 not be the same as the brain concentration. 10 Doctor Bourne's study, for example, actually 11 measured the concentration in the brain, and 12 people have been trying to measure spinal fluid 13 levels of 5HIAA and other serotonin metabolites 14 forever. 15 Q Well, what I think I'm 16 interested in, Doctor Thompson, is, is it your 17 testimony here that you can have a normal lab 18 value of serotonin in brain tissue? 19 A You could take brain to 20 the laboratory, measure the concentration and say 21 this falls within the range of people that are not 22 recognized to have a medical illness, but just 23 died suddenly, and so that would be regarded as 24 normal. 701 1 Q No, that's not what I'm 2 asking. If you send blood to a lab and you ask 3 for cholesterol levels in that blood, you'll get 4 back a report concerning the cholesterol levels, 5 will you not? 6 A Yes, sir. 7 Q And there are various 8 reports of various types of cholesterol in human 9 blood, are there not? 10 A Yes, sir. 11 Q And you'll get a 12 breakdown of each type of that cholesterol level? 13 A Yes, sir. 14 Q Each level of the various 15 types of cholesterol, right? 16 A Yes, sir. 17 Q And there is a scientific 18 normal range within which that numerical level 19 should fall, is there not? 20 A You were doing so well, 21 sir. If you read my publications on laboratory 22 values, you'll find that this is an exceedingly 23 difficult area, and if you want a yes or no to 24 that, I will answer it no, there is not a 702 1 scientific reference range except those that Lilly 2 has prepared. So, I mean, I'm sorry, you wandered 3 off into an area that's very arcane, but which in 4 fact there's an lot of information. 5 Q Well, I don't know, I 6 just have seen a lot of medical records as a lay 7 person over a large period of time -- long period 8 of time where I see cholesterol of two 9 thirty-three -- 10 A And it says normal, high, 11 low. 12 Q No, it will say normal 13 one sixty to two hundred, for instance, okay? 14 A Uh-huh. The FDA is 15 teaching other companies to use the Lilly way of 16 looking at laboratory reference ranges. 17 Q But physicians understand 18 and make medical judgment of these reference range 19 in laboratory values, do they not? 20 A Well, they have to make 21 some judgment from it, but the problem is that 22 what that actually means is very difficult, and 23 you picked one of the more difficult areas. If 24 you picked serum sodium there's a lot more 703 1 agreement. 2 Q Okay. 3 A But goodness gracious, if 4 you go to the heart association, you get one set 5 of things that are called normal; if you go to NIH 6 you get a different set; if you go to one lab, you 7 get one set; if you go somewhere else you get 8 another. But if you take serum sodium, then we 9 can have more agreement among physicians as to 10 what's normal and abnormal. 11 Q All right. So serum 12 sodium, there's more unanimity of what might be 13 normal, right? 14 A Yes, sir. 15 Q And there is a specific 16 numerical assignment of normal serum sodium 17 levels, correct? 18 A Not a specific one, but 19 we would have more level of agreement among 20 different physicians about different patients. 21 Q About what is the normal 22 range? 23 A Yes, sir. 24 Q What is the normal range 704 1 in that instance? 2 A Okay. I'm sorry, sir, 3 but now you've gotten off into an area that's 4 again very arcane. There is no meaning to the 5 word normal. However, what Lilly has done in a 6 pioneering fashion was to simply look at the value 7 of analytes like sodium and cholesterol in twenty 8 thousand patients while they were taking placebo, 9 and what we looked at was the actual probability 10 distribution or the frequency distribution of what 11 the values were. What we found was that there 12 were five characteristics of the patient that in 13 fact made a big difference in the assessment of 14 laboratory values: Age; we divided people by 15 greater than or less than fifty. Gender; male, 16 female. Race; the best division was black versus 17 non-black -- no, I'm sorry, I think we came to 18 Caucasian, non-Caucasian. Whether or not the 19 patient said that they were a smoker, and whether 20 or not patients said they were a drinker of 21 alcoholic beverages. Depending on those five 22 criteria then, if each one is divided into two 23 different possible answers, there are thirty-two 24 different ways of classifying patients. If you 705 1 look at those thirty-two ways of classifying 2 patients against forty different measurements from 3 the laboratory, you'll find that seventy percent 4 of the time those five criteria make a difference 5 in how you'd interpret the test. 6 So that the way we 7 analyze laboratory tests and the way FDA is asking 8 other companies to follow our lead is that for a 9 given patient, and you, if I knew those five 10 criteria, I could tell you the probability with 11 which an individual value of something like serum 12 sodium in you was different from a reference 13 population, those twenty thousand folk on 14 placebo. But in fact, without knowing those five 15 criteria in you, I couldn't give you a specific 16 numerical range, but I could go look up in the 17 book what the one percent high and the one percent 18 low or the three percent high and the three 19 percent low cutoff values are for each one of 20 those analyzed for each of those thirty-two 21 groups, and that's become the reference standard 22 in our industry of how you analyze laboratory 23 values. 24 Q But there are values 706 1 assigned, correct? 2 A Yes, sir. 3 Q Like -- how is the value 4 assigned in serum sodium? 5 A We use what's called the 6 international standard units, which are in moles, 7 and so it's some number of millimoles per liter, 8 and roughly would be between -- I mean, I'll give 9 you a rough number, it's between about a hundred 10 and thirty-five and a hundred and forty-five. 11 Q Millimeters? 12 A Millimoles per liter. 13 Q Millimoles per liter? 14 A Yes. 15 Q Now, is there any 16 assignment, either per liter, per cubic 17 centimeter, or any particular value assigned to 18 either normal or abnormal serotonin level in the 19 brain? 20 A Yes, sir. I mean, you 21 can go look at papers that have looked at a whole 22 bunch of people that are thought to be normal, but 23 dead, and you can measure what their 24 concentrations are and see what that range is. In 707 1 fact, you can diagnose a carcinoid tumor by 2 measuring serotonin in the blood, so there is a 3 blood level above which there is a probability of 4 having a carcinoid tumor. 5 Q I thought I was using 6 brain tissue by virtue of the fact that I thought 7 there was no relationship between the amount of 8 serotonin present in brain tissue and the amount 9 of serotonin present of blood platelets? 10 A There's not much, it's 11 mostly unrelated. 12 Q All right. 13 MR. FREEMAN: At some 14 convenient point let's go to lunch, it's 1:20 in 15 Atlanta. I know you all are used to this time. 16 MR. SMITH: Well, let's 17 stop now. 18 (LUNCH BREAK TAKEN.) 19 MR. BOUR: The time is 20 1:36. 21 Q (BY MR. SMITH) Doctor 22 Thompson, as I understand it, you were part of a 23 group of upper management that met every Tuesday 24 or Thursday morning to discuss Prozac related 708 1 issues? 2 A Yes, sir. 3 Q Was it on a Tuesday or a 4 Thursday? 5 A I can't remember because 6 it changed from time to time and the group 7 composition changed from time to time. 8 Q But it was upper 9 management, was it not? 10 A There were some people I 11 think at director level, but it included people 12 all the way up to Doctor Herr. 13 Q Doctor Herr, I believe, 14 Doctor Perelman -- 15 A Yes, sir. 16 Q -- has testified that he 17 participated in those meetings, Mr. J.B. King was 18 in some of those meetings -- 19 A After he joined the firm. 20 Q The -- some of the 21 marketing individual -- upper marketing 22 individuals were at the Thursday morning meetings, 23 I believe. 24 A Again, I think it changed 709 1 days and I can't remember. 2 Q But you tried to meet 3 once a week to discuss Prozac related issues? 4 A For some periods. I 5 mean, sometimes they'd get cancelled, sometimes 6 they were less frequent, but it was -- sometimes 7 it was once a week. 8 Q And there were many times 9 when those discussions were related to the 10 allegations that have been made concerning Prozac 11 and suicide and Prozac and violent aggressive 12 behavior? 13 A Yes, sir. 14 Q Were there any of those 15 meetings that occurred prior to the publication of 16 the Teicher article in February of 1990 that you 17 recall? 18 A I don't think so. 19 Q Was there ever any 20 consideration given in any of those meetings to 21 making some labeling change where there would be 22 some type of discussion in the Prozac labeling in 23 the United States concerning suicide risk and 24 Prozac other than what the labeling contained? 710 1 MR. FREEMAN: At 2 present? 3 MR. SMITH: Yes. 4 A Was there a discussion of 5 doing something with the label other than its 6 current wording? 7 Q Yes. 8 A Not that I recall 9 specifically, but there were a whole variety of 10 possible actions that the company could take on 11 this issue. 12 Q Was it ever suggested or 13 was it ever considered that an addition be made to 14 the Prozac labeling containing language to the 15 effect that in a small group of individuals Prozac 16 may cause or increase suicidal ideation or violent 17 aggressive behavior in some individuals? 18 A That was one of the 19 things that was talked about. 20 Q And was there anyone 21 there who proposed making such a change? 22 A If there were a 23 proponent, I can't remember who that was. 24 Q Were there individuals 711 1 who opposed that idea? 2 A I would have opposed it, 3 but I don't remember taking that position 4 formally. 5 Q And would it be your 6 testimony that you would have opposed it based on 7 scientific reasons? 8 A Yes, sir. 9 Q But as far as I 10 understand your testimony, it was considered that 11 labeling changes as I had suggested be made, but 12 that idea was rejected? 13 A Yes, I can't remember the 14 specific context, but we certainly had many 15 discussions about this issue and a whole variety 16 of possible actions, including changing the 17 labeling was discussed, so -- and that's such a 18 likely thing to have been discussed, I think 19 that's an accurate description. 20 Q And it was decided by 21 upper Lilly management that there should not be a 22 change in the Prozac labeling suggesting that in 23 some individuals Prozac might cause suicidality or 24 cause a violent aggressive behavior? 712 1 A I don't think that's a 2 fair characterization, Mr. Smith. First of all, 3 that wasn't the group that made the decision on 4 label changes in general; there was another group 5 that in fact was responsible for making the label 6 changes. And secondly, I'm not denying the fact 7 that that group could have overruled our primary 8 group, and I'm not denying the fact that in fact 9 if that group had said you will go and make this 10 change, it in fact probably would have happened, 11 but I don't recall any discussion at which, you 12 know, a vote was taken or people sort of lined up 13 pro and con and they said, well, we're not going 14 to make that change. Again, there -- a whole 15 bunch of possibilities came up in all of those 16 discussions, so I think that probably was one of 17 the possibilities. 18 Q Well, a label change was 19 a possibility? 20 A Yes, sir. 21 Q One of the 22 possibilities -- 23 A Yes, sir. 24 Q -- that was discussed, 713 1 and would my characterization of a suggested label 2 change be accurate concerning potential label 3 changes? 4 MR. FREEMAN: The first 5 one or the second one? Because you made it 6 absolutely causal in the second one, in the first 7 one you qualified it and said may in a small group 8 of people increase. 9 Q All right, let's take it 10 in both instances. Was it -- the first potential 11 change that I might suggest would be language 12 something to the effect -- and don't hold me to 13 specific language -- but something to the effect 14 that in some individuals Prozac may cause or 15 aggravate suicidality and violent aggressive 16 behavior? 17 A I think that was within 18 the context of a general discussion, but I guess 19 the problem I'm having is that your question 20 related to a decision made by this group, and that 21 would suggest that in fact there was a focused 22 discussion should we or should we not do this one 23 specific act, and the decision was made we should 24 not put that in the label. That -- I don't 714 1 believe I ever sat in on a discussion like that. 2 In terms of talking about a whole number of things 3 that could happen, we could do this, we could do 4 that, we could give this report, et cetera, some 5 of which in fact then were decided on to do 6 positively and that was left out, I will agree 7 that that probably occurred. 8 Q So was there ever a label 9 change as a potential for action that Lilly could 10 make? 11 A Yes, sir. 12 Q Now, would such a label 13 change have required United States Food and Drug 14 Administration approval, that I've suggested? 15 A Not the wording you 16 suggested. 17 Q What if I had -- what if 18 the label were to say Prozac causes suicidality 19 and violent aggressive behavior in some 20 individuals, would that be such a change that 21 would require United States Food and Drug 22 Administration approval? 23 A Prior approval, no. 24 Q Both, I would assume, 715 1 would be label changes that you would submit to 2 the Food and Drug Administration as a matter of 3 advising them of what you were doing? 4 A Yes, sir. 5 Q Would the United States 6 Food and Drug Administration have -- if you 7 know -- the ability to cause Lilly to make a 8 label change in any manner that I've suggested? 9 A Yes, sir. 10 (THOMPSON EXHIBIT NO. 36 MARKED FOR 11 IDENTIFICATION.) 12 MR. FREEMAN: We're going 13 to carefully read that on the record to see if the 14 court reporter can take it down. It's in German. 15 Q (BY MR. SMITH) Doctor 16 Thompson, I've handed you -- since you speak 17 German, you say, to some degree, a -- 18 A I said I didn't read it 19 fluently, but if you give me a dictionary, I'll do 20 my best. 21 Q It's going to be easier 22 than that, I hope. You should have an exhibit 23 before you that is the Berlitz translation of the 24 label that we've given you. 29. 716 1 A This is Exhibit 29, yes, 2 sir; I have Exhibit 29 and Exhibit 36. 3 Q All right, what I would 4 like you to do, since you are somewhat familiar 5 with German, is mark on that area that language we 6 were talking about in the German insert where it 7 says risk of suicide, Fluctin does not have a 8 general sedative effect on the central nervous 9 system, therefore for his/her own safety, the 10 patient must be sufficiently observed until the 11 antidepressant effect of Fluctin sets in. Taking 12 an additional sedative may be necessary. This 13 also applies in cases of extreme sleep 14 disturbances or excitability. 15 A Yes, sir, I've got that 16 on Exhibit 29. 17 Q All right. Can you mark 18 for us where that actually is in the actual 19 German, German package insert of Exhibit 36? And 20 I apologize for the quality of the print, that's a 21 copy of a fax, I believe. 22 A I think -- now, remember, 23 I don't read German -- that is in the middle of 24 the first page on the middle column of Exhibit 36, 717 1 and I think it begins with "Selbtsnordga", and 2 then I lose what the letters are. 3 Q All right, can you just 4 make a check mark by that dot where that's 5 located? 6 A Yes, sir. 7 Q And why don't you initial 8 it so -- 9 MR. BRENNAN: Are we 10 stipulating that Doctor Thompson's translation is 11 controlled, that he's qualified to make that? 12 MR. SMITH: Have you got 13 a problem with that, John? 14 MR. BRENNAN: I do if 15 from an evidentiary basis, yes. 16 A I can only make out -- 17 first of all, I can't even make out the letters 18 very well, but the words that I think are there, I 19 think that's my best guess as to the paragraph. 20 Q (BY MR. SMITH) That's 21 fine. I just -- it is a fax. 22 Now, under risk of 23 suicide on that package insert, there's a the word 24 epilepsy? 718 1 A Yes, sir. 2 Q Can you locate that and 3 maybe put an arrow where the word epilepsy is so 4 we'll know where the epilepsy portion starts? 5 A I can do better with that 6 one. 7 Q All right. Thank you. 8 (THOMPSON EXHIBIT NO. 37 MARKED FOR 9 IDENTIFICATION.) 10 THE WITNESS: Yes, sir. 11 Q (BY MR. SMITH) Exhibit 12 37 is a confidential minute meeting of the 13 International Marketing Committee dated February 14 21st, 1985, is it not? 15 A Yes, sir. 16 Q I don't see your name on 17 the list of members present or others present. Is 18 it -- were you there? 19 A No, sir. 20 Q Have you seen this 21 document? 22 A I don't think so. 23 Q Is this a document you've 24 reviewed in preparation for your deposition? 719 1 A No, sir. 2 Q Can you tell us what the 3 International Marketing Committee is? 4 A Not for sure, I've never 5 been a member of it. I presume it was a committee 6 in marketing that had to do with international 7 marketing, but, I mean, that's the best I can do 8 for you. 9 Q Well, there's a notation 10 on the first page that Mr. E.W. Lange was present, 11 correct? 12 A Yes, sir. 13 Q Is he Wally Lange? 14 A I think so. 15 Q Also it indicates that 16 Doctor A.J. Weinstein was present? 17 A Yes, sir. 18 Q And he was in the 19 International Marketing Division? 20 A No, sir, he was in -- 21 Q I mean International 22 Medical Division. 23 A He headed that, yes, sir. 24 Q All right, and he 720 1 reported to you at one time? 2 A At one time, but not at 3 this time. 4 Q All right. The document 5 says: Mr. Thompson reviewed with the group the 6 current status of premarketing activities with 7 fluoxetine, primarily focusing on the results of 8 the product profile test recently conducted in 9 Europe. The outcome of that study suggests 10 several key issues which would need to be 11 addressed in the near future. Correct? 12 A Yes, sir. 13 Q The first bullet point 14 under that indicates that one of the key issues 15 was the need to initially develop a message and 16 superior materials centered around the efficacy -- 17 A Support materials? 18 Q And support materials, 19 thank you. Let me start that again. The need to 20 initially develop a message and support materials 21 centered around efficacy and not the lack of side 22 effects. Correct? 23 A Yes, sir. 24 Q Do you interpret that to 721 1 mean that there was some discussion within 2 individuals at Lilly whether or not to center on 3 marketing Prozac as an efficacious medication or 4 one whose benefit is its lack of side effects as 5 opposed to those tricyclics and MAO's already in 6 existence? 7 A The words speak for 8 themselves, but I'm no better than you or anybody 9 else at guessing what it means. That's one 10 interpretation of it, but there are many others I 11 could give you as well. 12 Q No, maybe I didn't make 13 myself clear. My question is, were you aware that 14 there was some discussion concerning whether or 15 not Prozac should be -- whether the efficacy of 16 Prozac should be emphasized or the fact that it 17 had different or lack of side effects of existing 18 antidepressant therapy? 19 A Different marketing 20 people at different times came and asked us for 21 all kinds of -- you know, ways to analyze the data 22 to show the drug was worthwhile, and I'm sure that 23 was one of them, but I -- excuse me -- in regard 24 to this meeting and this statement, which country, 722 1 I don't know. 2 Q The second bullet point 3 indicates, quote: The need to be prepared to 4 handle the usage of fluoxetine in the agitated 5 depressed patient, given the fact that it is a 6 non-sedating antidepressant. Correct? 7 A Yes, sir. 8 Q Were you and anyone in 9 medical consulted on how to handle the usage of 10 fluoxetine in agitated patients? 11 A Yes, sir. 12 Q And tell me about those 13 discussions. 14 A Well, fortunately the 15 Hamilton depression scale, which had been given to 16 almost all the patients, can be used to look at a 17 sub-scale which measures how much agitation, 18 that's a word that's often used for what a patient 19 has, so that you can in fact partition the group 20 of patients at entry into those that have higher 21 scores on the agitated side, those that have 22 higher scores on the retarded side, so that we 23 could look at whether there was a difference 24 between those two groups of patients in terms of 723 1 their response to the drug. So, in fact, we did 2 that kind an analysis. 3 Q Was there a difference? 4 A No, not really; there was 5 very little difference. 6 Q Well, how little 7 difference? I mean, can that be quantified in any 8 way? 9 A If we go and get the 10 data, I'm sure it can be, but I can't remember 11 exactly. 12 Q And there was no need to 13 handle those patients differently? 14 A That's correct. 15 Q From a medical 16 standpoint, as far as you're concerned? 17 A Yes, sir, that's correct. 18 Q Do you agree that Prozac 19 is a non-sedating antidepressant? 20 A Yes, sir. 21 (THOMPSON EXHIBIT NO. 38 MARKED FOR 22 IDENTIFICATION.) 23 THE WITNESS: Yes, sir. 24 Q (BY MR. SMITH) Doctor 724 1 Thompson, Exhibit 38 is a set of documents that 2 was produced to us by Lilly, purportedly out of 3 your files. 4 A Yes, sir. 5 Q I can't identify exactly 6 why this set of documents is together, can you? I 7 mean, have you had somebody prepare for you some 8 data and that was something you retained in your 9 file? 10 A Yes, sir. I can't -- I 11 can't recall that all of this was prepared as one 12 distinct entity at the same time, but they all 13 certainly are related and it does have a couple of 14 marks that I made on it. 15 Q All right. Can you 16 indicate where your writing is? 17 A On the page that's 18 labeled 1131, down in the lower right-hand 19 quadrant where it says three seven three nine 20 serious with a line under it, two three two six 21 five, sixteen percent, that's my writing. 22 Q All right. Anything else 23 that you recognize as your writing? 24 A No, sir. I looked at 725 1 that carefully, I think that's all that's on here 2 that's mine. 3 Q There's a lot of 4 handwritten notes, like 1131 says from Gary 5 Martindale -- is that right, Martindale? 6 A Yes, sir. 7 Q DEN Query, Prozac 8 Spontaneous Reports, Annual, 3 June -- 9 A I think it's 13, but -- 10 Q 13 June 91? 11 A Nine something. 12 Q All right. Can you 13 identify what this page is? 14 A Well, I think that Gary 15 Martindale is reporting from the DEN system the 16 total number of adverse events that were 17 classified as originating in the US with a serious 18 code, OUS meaning international, not US, serious, 19 versus total, both in the US and OUS, in 20 '88, '89, '90 and '91. Now, what I don't know 21 from this is whether this was calendar '88 or 22 whether this was in the annual report to the NDA 23 and therefore might have been at some other date 24 within '88. 726 1 Q All right. Would this be 2 all serious reports within the DEN system 3 regardless of whether or not it had to do with 4 attempted suicide, violent aggressive behavior, 5 any serious report? 6 A That's the way I 7 interpret it. 8 Q Now, look at the next 9 page, and it looks like that's all -- on the 10 bottom of that is written Sarah McKenzie? 11 A Yes, sir. 12 Q 15 June 91? 13 A Yes, sir. 14 Q And can you tell me what 15 that document reflects? 16 A No, I had a problem with 17 that because I don't know what Prozac reports from 18 HEC documents means, and it looks to me like 19 there's a line above that that's cut off, and it 20 also says it's the second page, but the one that's 21 in front of it, which also says number one, 22 clearly to me looks like it's from Martindale, and 23 that one looks like it's from Sarah McKenzie. So, 24 there must have been two different people that 727 1 prepared the reports, but I don't know how to 2 connect them up. Some of the text, however, talks 3 about doing it on a quarterly basis, even on a 4 monthly basis, so I'm assuming -- I'm assuming 5 that these are the total DEN Prozac reports in 6 each of the quarters that are indicated, US, 7 international, and total. 8 Q Is there anything at 9 Lilly known as HEC that you can recall off the top 10 of your head, that would have to -- that would be 11 explanative of this HEC being used here? 12 A And it could be HEL. It 13 says HEC -- to me it reads HEC or HEL documents, 14 and I don't know what that means. 15 Q The next document -- next 16 page says "DEN Reports (Events) of Personality 17 Disorder," is that right? 18 A Yes, sir. 19 Q Is this data that you had 20 requested at some time apparently? 21 A I think so, but whether I 22 was the originator of that request or somebody 23 else was and I was just being copied on it, I was 24 getting reports like this. 728 1 Q Do you recall what the 2 reason for that would have been? 3 A Sure, because we were -- 4 we were very much interested in everything related 5 to suicidality and violent behavior, and these 6 were terms that we were searching to see if there 7 were any clues there as to whether we had a 8 problem. 9 Q All right. It looks like 10 this goes by quarter from 1988 through the second 11 quarter in 1991, does it not? 12 A Yes, sir. 13 Q And lists events of DEN 14 reports of personality disorders? 15 A I think that's correct. 16 Q Do you know of any reason 17 why we couldn't consider this as accurate data 18 concerning information contained in the DEN system 19 concerning personality disorders? 20 A I don't have any reason 21 to say whether that it's not accurate, but if you 22 wanted me to get these numbers for you, I'd go 23 back to DEN itself. 24 Q Of course, you could go 729 1 back to DEN and get current numbers also? 2 A No, you can get all of 3 these numbers. If a report goes into DEN, you can 4 always retrieve it. 5 Q Yes, I understand we 6 might be able to get the first quarter of '90, but 7 we could also get the third quarter of '93 now, 8 couldn't we? 9 A Yes, sir. 10 Q All right. The next 11 page, Page 1134 apparently is "DEN Reports 12 (Events) of Intentional Injury," is it not? 13 A Yes, sir. 14 Q And that shows a total in 15 1990 of forty-two in the United States and three 16 outside the United States, correct? 17 A Yes, sir. But it is a 18 little confusing in the fact there's some text 19 that addresses the difference between an event and 20 a report. 21 Q It says: The above data 22 represents the total number of spontaneous reports 23 in DEN as of 13 January 91. As of 13 January 91 24 there were no trial reports -- is that correct, 730 1 trial reports? 2 A That's the way I'd read 3 it. 4 Q Of intentional injury in 5 DEN. Correct? 6 A It's absolutely correct, 7 sir, but I don't know where those asterisks refer 8 because there isn't any asterisk above that I see 9 that they would logically refer to, not on this 10 copy. 11 Q It looks like that's 12 being written on every copy, does it not? 13 A Yes, sir. 14 Q I assume by trial, 15 they're talking about clinical trials? 16 A Yes, sir. 17 Q Well, it shows forty-two 18 reports from the United States in 1990, three 19 outside the United States, for a total of 20 forty-five, correct, of reports of intentional 21 injury of individuals consuming Prozac? 22 A Yes, sir. 23 Q And a total of fifty in 24 the first half of 1991, correct? 731 1 A Yes, sir. 2 Q The next page apparently 3 is a breakdown of the number of reports of 4 hostility in Lilly's DEN system, correct? 5 A Yes, sir. 6 Q By year? 7 A By quarter, even by 8 month. 9 Q And there was ten in -- 10 ten reports of hostility in the United States in 11 1988, correct? 12 A Yes, sir. 13 Q There were a total of 14 twenty-seven reports of hostility in the DEN 15 system for the year 1989? 16 A Yes, sir. 17 Q For the year 1990, the 18 DEN system had received reports of one hundred and 19 eighty-three instances of the adverse event of 20 hostility, is that right? 21 A Yes, sir. 22 Q And for the first half of 23 1991, there had been reports of a hundred and 24 sixty-five individuals who had become -- 732 1 reportedly become hostile while taking Prozac, 2 correct? 3 A Or after the intention to 4 treat with Prozac, yes, sir. 5 Q Beg your pardon? 6 A Or after the intention to 7 treat with Prozac, yes, sir. 8 Q The next page, Page Pz 9 1211 dash 1136, it appears to be a compilation of 10 the DEN reports of aggression, correct? 11 A Yes, sir. 12 Q Twenty-one reports of 13 aggression in 1988? 14 A Yes, sir. 15 Q Forty-eight in 1989? 16 A Yes, sir. 17 Q Two hundred and 18 eighty-three in 1990? 19 A Yes, sir. 20 Q And a hundred and -- two 21 hundred and eighty-nine in 1991? 22 A How about forty-nine? 23 Q Two hundred and 24 forty-nine in 1991? 733 1 A Yes, sir. 2 Q Of reports of individuals 3 who had experienced aggression while on Prozac or 4 being intended to be treated with Prozac? 5 A Aggression as defined by 6 the asterisks, no. 7 Q How is it they define -- 8 how do they define aggression there? 9 A If any one of the three 10 event terms, either intentional injury or 11 hostility or personality disorder, were associated 12 with that patient, then it was counted as, quote, 13 aggression, end quote. 14 Q All right, so would we 15 then have -- would actually that aggression 16 cluster, those numbers there would be the totals 17 that we add up, personality disorder, intentional 18 injury and hostility? 19 A It shouldn't be; that's 20 explained very nicely in the last two lines. 21 Q What does it say there? 22 A Well, it says: Since one 23 report may use more than one event term, these 24 totals will vary from simple addition of the event 734 1 terms. So that in one report, the term hostility 2 and the term personality disorder might both 3 appear, so they would be counted under each one of 4 those headings, but they would only be counted as 5 one report. 6 MR. SMITH: All right. 7 (THOMPSON EXHIBIT NO. 39 MARKED FOR 8 IDENTIFICATION.) 9 Q (BY MR. SMITH) While 10 they're looking at that, Doctor Thompson, there 11 are no duplicates in the individual listings of, 12 say, hostility and intentional injury -- 13 A No, one -- 14 Q -- and personality 15 disorders as listed in this exhibit? 16 A Well, the way DEN is 17 supposed to work is that one patient at one time 18 would only be counted once, so the only way that 19 patient would get counted again, except by a 20 mistake, would be if they had another event at 21 some other time. 22 Q Okay. So there would be 23 no duplication unless they had been -- had two 24 separate events on two different dates? 735 1 A Unless there was a 2 mistake. We often have duplicates because we hear 3 about the report from different people at the same 4 time. 5 Q Well, do you know of any 6 mistakes contained in that? 7 A I don't know of any, but, 8 I mean, that -- I was telling you that that's one 9 of the problems that we have, is that a 10 salesperson, a physician, and somebody else would 11 tell us all about the same individual patient, and 12 when we can find that out, then we try to go back 13 and correct it by annotating the file. 14 Do you want me to read 15 all of this? 16 Q No. Why don't you review 17 it real quickly just to get familiar with it; I'm 18 certainly not going to ask you about the whole 19 thing. 20 A I flipped through it, so 21 we won't say I've read it in detail, but I'm 22 generally familiar with it. 23 Q In answering my 24 questions, feel free to examine it in more detail, 736 1 Doctor Thompson, if you need to. Can you identify 2 Exhibit 39? 3 A Yes, sir. I think this 4 is a letter from Doctor Talbott to Doctor Leber 5 dated April 2, 1986, and I think this happened 6 after a teleconference, and I was -- this is 7 probably the one that I was sitting in, in which 8 Doctor Leber told us how to organize the safety 9 update to the Prozac NDA, and this -- the purpose 10 of this, I believe, is to show the FDA how we 11 proposed to report the adverse event and the 12 instance of adverse events and asked them for 13 comment. 14 Q All right, and I see your 15 name among individuals receiving blind carbon 16 copies of this document up at the top left-hand 17 corner of the page? 18 A Yes, sir. 19 Q Did you review this when 20 it was prepared? 21 A In general, yes, sir. 22 Q Certainly in more detail 23 than you did just a minute ago? 24 A Yes, sir. 737 1 Q Is this a document that 2 you reviewed in preparation for your deposition? 3 A No, sir. 4 Q Now, the first fifteen or 5 so pages talk about combination terms, do they 6 not? 7 A Yes, sir. 8 Q And how -- how is that 9 supposed to work? 10 A Well, this was at the 11 time we were still using the SSAI dictionary, 12 which was a dictionary of classification terms 13 that Lilly had prepared. So, to my knowledge, 14 this was unique to Lilly, at least some of it 15 was. And the way it was supposed to work was that 16 if one of the terms on the right were used in 17 describing an adverse event, then it would be put 18 into a cluster called -- for example, on Page 2, 19 the term agitation would include all of the events 20 that had the words akathisia, anxiety, et cetera, 21 et cetera, anywhere in the adverse event report. 22 Q All right. So if you got 23 a report from a physician that says my patient 24 became irritable on Prozac, it would be reported 738 1 as agitation, is that what you're saying? 2 A Not quite. 3 Q All right. That's one of 4 the purposes in this discussion is to flush this 5 stuff out. 6 A Well, I'm going to tell 7 you from the best of my knowledge because, again, 8 this system, the SSAI system was in place when I 9 came to Lilly and I changed it later, but I think 10 this was used very much like the Co-Start system, 11 which means that whatever the original 12 investigator described the event as would be 13 included in the database. So, you're correct, if 14 the investigator said irritable, irritable should 15 be a word that's somewhere in the database. 16 However, I think these 17 terms on the right were in fact dictionary terms. 18 In other words, the investigator could pick any 19 word in the English language to describe an event, 20 but for the purposes of being able to do 21 statistical evaluation, we grouped events so -- 22 into a number of buckets, less than two thousand 23 buckets as I recall for SSAI. So irritable, I 24 think, was one of the defined terms in SSAI, and 739 1 therefore could have been used, but it wasn't a 2 hundred percent certain that the Lilly person 3 coding that event would necessarily use the word 4 irritable, even if the investigator had used the 5 word irritable; that's not a hundred percent 6 certain. 7 Q Okay. 8 A But, if, in fact, the 9 coding term irritable had been used, the computer 10 would have then put it into the agitation bucket 11 as well. What we did was use a hierarchy of 12 buckets. So the lowest level is whatever the 13 terms are the investigator used, it could have any 14 English language in it, and then there were 15 defined classification terms, and you had to use 16 one of those terms, you couldn't go and make one 17 up without adding it to the dictionary. And then 18 those were progressively clustered into a smaller 19 number of terms, again primarily so you could sort 20 through the adverse events and find out where to 21 go to find one and count up how many there were. 22 Q All right. So if 23 somebody asked are individuals becoming agitated 24 on Prozac, are we receiving any reports of 740 1 agitation in general, you could type in agitation 2 to this SSAS -- 3 A SSAI. 4 Q SSAI, and you would get a 5 number that would include at least these terms 6 that are described on the right hand page -- 7 right-hand side of the page? 8 MR. FREEMAN: If there 9 had been a report. 10 MR. SMITH: Yes. 11 A That's correct, although 12 I think by the time we marketed the drug, we had 13 switched over to Co-Start. So the same philosophy 14 is true, we just had the FDA dictionary rather 15 than this dictionary. 16 Q Did that produce any 17 substantial change? 18 A Oh, yes. Oh, my 19 goodness, yes. 20 Q Why? 21 A Well, I mean, they're 22 different -- 23 Q Because you had some more 24 terms? 741 1 A They were just different 2 terms. Lilly, at this time, when I came, was 3 actually using three different dictionaries, and 4 so depending on what the drug was, they used one 5 dictionary or another. The one used for 6 antibiotics, as I can recall, was called DOT. And 7 this dictionary also grew, so that if I were 8 monitoring a drug, Fenestra was a drug I monitored 9 when I came to Lilly, and I wanted to add another 10 term to this, you know, I could just add another 11 term and it would get added to the dictionary. So 12 that all the events before I added that term could 13 not obviously have used that term, but the events 14 after that term could now have a new term. 15 Co-Start -- the reason we 16 went to Co-Start was that is the list of terms 17 that the FDA uses themselves. In fact, no matter 18 what we send to them, they don't pay any attention 19 to it, they go and code it under Co-Start. So we 20 said since they're our primary customer, at least 21 in the US, why not use their terms, let them worry 22 about managing the dictionary; when they add a 23 term, we'll add it to ours, and if they don't add 24 a term, then we'll do the best we can in 742 1 classifying under their terms, since those are the 2 terms they're going to use in classifying all of 3 the events we send to them. 4 Q I've also seen the term 5 ELECT dictionary. 6 A Yes, sir. 7 Q Is that something 8 different than Co-Start and SSAI? 9 A Yes, it -- but it's very 10 close to Co-Start. When we adopted the Co-Start 11 terms -- I'm sorry this is complicated, but let me 12 explain it to you, because I made up the word 13 ELECT, and the reason was that in the FDA Co-Start 14 thesaurus, they actually have a variable body 15 system to associate with the term. And let me 16 think of one example that I think I remember, 17 which is edema, a collection of fluid. Under the 18 Co-Start used by the FDA, edema could be 19 referenced to lungs, brain, or probably leg or 20 body as a whole. Lilly didn't want to do that, 21 and so what we decided to do was to use all of the 22 Co-Start terms but not have a variable body 23 system, so we had a fixed body system assignment. 24 We also had a couple of 743 1 terms that were slightly different than the FDA, 2 there were a couple out of roughly fourteen 3 hundred, and so because it wasn't precisely the 4 same, I thought we shouldn't, in fact, use the 5 word Co-Start since it wasn't exactly what the FDA 6 defined, so we made up the Eli Lilly Event 7 Classification Terms, which is what ELECT means. 8 Q All right, so what is -- 9 is ELECT being used now, or Co-Start? 10 A I couldn't tell you 11 exactly what people are referring to now. We've 12 gradually gone back to using only the Co-Start 13 terms, but we still don't do variable body system 14 assignments. 15 Q All right. You had 16 indicated that you felt that tremor was an adverse 17 event that is causally associated with Prozac use? 18 A I think I said that the 19 best of my knowledge was it was a statistically 20 significant difference in the incidence of patient 21 reports of tremor on Prozac versus placebo and 22 studies of depression, but not necessarily the 23 other diagnosis. I think that's right. 24 Q All right. If you look 744 1 at Page 2 under agitation of this exhibit, I see 2 tremor as part of what's being included as 3 agitation. 4 A Yes, sir. 5 Q So did you take the 6 reports of tremor out of that group in agitation 7 to examine tremor individually compared to placebo 8 to come to this conclusion? 9 A Yes, sir. 10 Q Did you do that with 11 nervousness, or could that be done with 12 nervousness? 13 A Yes, sir. 14 Q Was it done with 15 nervousness? 16 A Yes, sir. 17 Q And what were the results 18 of that? 19 A I don't remember it being 20 statistically significant, but remember there were 21 different times with different amounts of data and 22 so forth. I don't think overall that that was a 23 significant difference, I don't think. I gave you 24 four, I think, that I thought were. 745 1 Q Al wrote them down. 2 A Nausea, sweating, tremor 3 and sleep disorder, I think were the four that I 4 thought made the cutoff. 5 Q That's right. Speaking 6 of sleep disorder, insomnia is also included under 7 agitation here? 8 A Yes, sir. 9 Q Would that be what you 10 mean by sleep disorder, insomnia? 11 A One of them. 12 Q Why is insomnia included 13 under the term agitation? 14 A I have no idea. That's 15 one of the reasons I got rid of the SSAI 16 dictionary. 17 Q How would it be 18 classified under the Co-Start or ELECT dictionary? 19 A I think insomnia is one 20 of the classification terms, but I would have to 21 go and check that, and my guess is it's under CNS, 22 but, I mean, that's a guess, I don't recall the 23 classification. 24 Q Look at Page 9. The 746 1 combination -- the term is psychosis related, 2 right? 3 A Yes, sir. 4 Q And it includes such 5 items as fears, hallucinations, hyperactivity, 6 hypomania, hysteria, manic psychosis, panic 7 disorder, paranoia, psychosis, and violent 8 behavior, does it not? 9 A Yes, sir. 10 Q Do you know why violent 11 behavior wouldn't be included under agitation? 12 A You mean under psychosis 13 related? 14 Q Yes. 15 A I don't have any idea. 16 Q Who developed the SSAI? 17 A I don't know, because it 18 was in use when I came. Doctor Dobbs was one of 19 the people that was in fact maintaining it, but I 20 have no idea where it or when it began. 21 MR. FREEMAN: When you 22 get to a convenient place, can we take a break? 23 MR. SMITH: Yes. 24 (SHORT BREAK TAKEN.) 747 1 MR. BOUR: The time is 2 2:57. 3 (THOMPSON EXHIBIT NO. 40 MARKED FOR 4 IDENTIFICATION.) 5 THE WITNESS: Do you want 6 me to read this word for word in detail? 7 Q (BY MR. SMITH) No, 8 just -- it's the approvable letter of September 9 9th, 1987, is it not? 10 A Yes, sir. 11 Q It's marked as Exhibit 12 40? 13 A Yes, sir. 14 Q Signed by Doctor Temple? 15 A Yes, sir. 16 Q I believe this document 17 was provided to us out of your file. Is this your 18 handwriting -- 19 A At least a lot of it is. 20 Q -- in the margins? 21 A A lot of it is. 22 Q Do you see any of it 23 that's not your handwriting? 24 A On Page 2, the four okays 748 1 on the right hand side, I can't tell, and I'm not 2 sure about the underlining. 3 Q But it's definitely your 4 writing in the margins? 5 A Yes, sir. On Page 4 6 where it says bad word, iteratively, that could be 7 mine, but I'm not sure I remember that. 8 Q Where is that? I didn't 9 see that. 10 A I'm sorry, on Page 4, the 11 left-hand margin, about -- almost halfway down the 12 page, it says bad word, exclamation point, and 13 there's a line that goes up to iteratively. 14 Q I agree it's a bad word. 15 Do you agree it's a bad word? What does 16 iteratively mean? 17 A Repetitively, over and 18 over again. Again, I'm not a hundred percent sure 19 about the underlining, but I think all the rest of 20 the words are mine, the notes, the handwritten 21 notes. 22 Q You say on Page 2 -- the 23 approvable letter on Page 2 says -- the last 24 sentence of the first paragraph says: To support 749 1 specific claims for the use of fluoxetine in 2 treating anxiety symptoms, the efficacy of Prozac 3 in primarily anxious patients would have to be 4 demonstrated. Correct? 5 A Yes, sir. 6 Q Then there's a notation 7 apparently in your handwriting, should we do this, 8 and I assume what you're saying is do a study on 9 anxious patients? 10 A Yes, sir. 11 Q Was that done? 12 A I don't think we've ever 13 gotten around to it. 14 Q Why did you question 15 whether or not Lilly should do this? 16 A Well, the primary thing 17 that we were after is a statement that would say 18 that Prozac will treat anxiety symptoms in the 19 patient who has depression, and I thought -- 20 Q You never got that, did 21 you? 22 A We didn't get that in the 23 US label, although I thought we had data to 24 substantiate that. What I think Doctor Temple was 750 1 asking for was to go out and find people who were 2 primarily anxious and treat them with Prozac, and 3 at this time that wasn't -- that was our purpose, 4 we weren't after a label that says in people that 5 just have anxiety you should use Prozac. 6 Q And that still has not 7 been done to date, as far as you know? 8 A I don't think so. We've 9 talked about it several times, but I don't think 10 it's been done. 11 Q Does your document have 12 the proposed label on it that says -- 13 A I don't see the -- 14 although -- 15 MR. SMITH: Maybe we 16 didn't get it stapled together. 17 MS. ZETTLER: I'm sorry, 18 I didn't -- I misstapled it. Do you want to make 19 it another exhibit? 20 MR. SMITH: Let's staple 21 it. 22 MR. MYERS: Okay. 23 Q (BY MR. SMITH) The 24 approvable letter had the proposed labeling, did 751 1 it not? 2 A Yes, sir. 3 Q And we omitted to add 4 that to this -- well, this handwriting here, is 5 this your handwriting also? 6 A I can't even read what's 7 on page -- is it 10? Well, whatever the first 8 page is of the label, I don't know what that is. 9 The "oops" I think is mine on the next page. 10 Q Where are we? 11 A I'm sorry, Page 2 of the 12 proposed label, right-hand side towards the bottom 13 you see "oops", I think that's mine. 14 Q Why did you write that? 15 A Well, I don't -- I think 16 my concern was that we hadn't explained our data 17 well, because what they were saying was, you know, 18 the proportion of the dose actually excreted as 19 recognizable fluoxetine metabolite by the kidney, 20 as best I can recall, I thought we had, in fact, 21 those data, but we must not have explained it well 22 enough that they understood that. 23 Q All right. 24 A And again, I'm not a 752 1 hundred percent sure about the underlining. Let's 2 see, on Page 4, I think that's all my writing. 3 Page 5, I think that's my writing, but at 4 somewhere in the middle of the page on the 5 right-hand side there's some marks that I don't 6 know what that is. 7 Q It may just be something 8 that was picked up. What is it you say, very 9 reasonable? 10 A Yes, sir. 11 Q And what did you have -- 12 what did you refer to when you said very 13 reasonable there? 14 A The contraindications, 15 and the specific thing being talked about here was 16 the rash, which you remember was our primary 17 concern. 18 Q All right, Page 6? 19 A I think that those are my 20 writings. 21 Q Under seizures you've 22 got -- what is that, whew? 23 A Whew. 24 Q What is it that you've 753 1 got underneath that, we don't think -- 2 A Twelve had true seizure. 3 Q All right. Why did you 4 write that there, the whew? 5 A Because I didn't think 6 point two percent was an accurate reflection of 7 the data. 8 Q All right. Then there's 9 the language on suicide, and looks like you've got 10 another whew written out there. 11 A I guess I agreed with 12 their writing. 13 Q Why did you write whew 14 where the -- on the labeling concerning suicide, 15 Doctor Thompson? 16 A Well, we had seen all 17 kinds of different statements in antidepressant 18 labeling related to the risk of suicide in 19 therapy, and this was something that I thought was 20 a fair representation of the facts. 21 Q Well, you've got a line 22 specifically drawn to depression where it says the 23 possibility of suicide is inherent in depression. 24 A I think in all honesty it 754 1 means the whole paragraph, but we didn't know what 2 they were going to suggest as a discussion of that 3 risk since it had varied so much from label to 4 label, and they hadn't approved an antidepressant 5 in a long time. 6 Q So you were glad to 7 get -- whew generally means -- is an explanation 8 of relief. I assume that you were relieved to get 9 that? 10 A Comfortable with what 11 they said, just like the rash thing. 12 Q Right. Then you felt 13 that their -- the labeling that they were going to 14 require concerning interference with cognitive and 15 motor performance was fair? 16 A Yes, because we actually 17 had data that -- what scared me is we had data 18 that actually improves performance if you're drunk 19 and take Prozac, and I didn't want -- I didn't 20 want them not to put this in the label, I wanted 21 them to have a warning in that stage in the label 22 that you shouldn't do that. 23 Q Is that your writing 24 under tryptophan? 755 1 A I think I meant what is 2 this, because I didn't think it was tryptophane, 3 in fact, that had those -- that was associated, I 4 thought it was 5HTP, the metabolite in 5 tryptophane. 6 Q It appears there's also 7 in this label a chart indicating treatment 8 emergent adverse experiences incident in the 9 placebo-controlled clinical trials -- 10 A Yes, sir. 11 Q -- correct? And that 12 describes seventeen hundred and thirty patients on 13 Prozac and seven hundred and ninety-nine patients 14 who received placebo? 15 A Yes, sir. 16 Q Then it appears that you 17 wrote a notation we want tricyclics here? 18 A Yes, sir. 19 Q I'm assuming from that, 20 that you wanted to have the chart reflect the 21 performance of the tricyclics in connection with 22 these treatment emergent adverse experiences? 23 A Well, adverse events 24 reported in those same trials where that was one 756 1 of the control drugs. 2 Q I understand that. Was 3 that ever accomplished, that you got a line with 4 the tricyclics? 5 A You know, I don't 6 remember. We certainly had precedence where they 7 had allowed that kind of table with a group of 8 drugs, not with the individual drug by name, and 9 therefore I thought that was perfectly fair to put 10 in, and it would have added to the informational 11 content. But I don't think that was ever done, 12 I'm not sure. 13 Q Under the body system, 14 adverse event, under the heading nervousness or 15 nervous -- 16 A Yes, sir. 17 Q Do you see that? 18 A Yes, sir. 19 Q It reports nervousness on 20 Prozac in eighteen percent, eighteen point three 21 percent, correct? 22 A Yes, sir. 23 Q And in placebo, nine 24 percent, correct? 757 1 A Yes, sir. 2 Q Does that mean that there 3 were reports -- that there was twice as many 4 reports of individuals taking Prozac of 5 nervousness than those individuals taking placebo 6 during the clinical trials? 7 A No, it would be even more 8 than that, because it means that eighteen point 9 three percent of seventeen hundred and thirty 10 patients at some time during the clinical trial, 11 whatever its duration, would have reported an 12 event which got classified as nervousness if they 13 had the intent to treat with Prozac as opposed to 14 those with intent to treat with placebo that had 15 nine percent of the seven hundred and ninety-nine 16 patients. 17 Q So it's the percentage 18 that's twice as high? 19 A Yes, sir. 20 Q As opposed to the actual 21 number of patients? 22 A Yes, the number would be 23 closer to four times, four and a half times more. 24 Q And anxiety, nine point 758 1 nine percent of patients treated or intended to be 2 treated with Prozac reported anxiety symptoms, 3 where of those patients being treated with only 4 placebo, which I guess is no treatment, you have 5 five point nine percent of those patients 6 complaining of anxiety? 7 A Or something that got 8 classified as anxiety, yes, sir. 9 Q And again, we're going 10 back to that cluster group reporting system, are 11 we not? 12 A Yes, sir. 13 Q And here we have the 14 figures on tremor, eight percent of those 15 individuals treated or intent to be treated with 16 Prozac reported a tremor, and two point four 17 percent of those patients on placebo, or getting 18 no treatment, no active medication, were reporting 19 a tremor? 20 A Yes, sir. 21 Q And insomnia was reported 22 by thirteen point eight percent of the patients 23 taking Prozac, correct? 24 A Yes, sir. 759 1 Q And seven point four 2 percent of the patients taking placebo? 3 A Yes, sir. 4 Q Have those percentage 5 numbers changed? 6 A Well, they have certainly 7 changed in the trials; I don't know whether they 8 changed in the label or not. 9 Q All right. Well, this 10 label only indicates that there were seventeen 11 hundred and thirty people on Prozac and there were 12 seven hundred and ninety-nine individuals 13 receiving placebo -- 14 A Yes, sir. 15 Q -- is that right? 16 A (WITNESS MOVES HEAD UP 17 AND DOWN.) 18 Q Would that have been the 19 number of individuals on the double blind, 20 placebo-controlled pivotal studies? 21 A It might be more broad 22 than just pivotal. I mean, I think they included 23 all the placebo-controlled depression trials that 24 they had at this time. 760 1 Q Would that have also 2 included trials that used a comparator drug as the 3 third arm? 4 A Yes, sir. 5 Q But it just wouldn't 6 reflect the numbers on the comparator drug in this 7 chart? 8 A Yes, sir, that's correct. 9 Q Back to Page 7 of the 10 label. 11 A Yes, sir. 12 Q This statement concerning 13 suicide that you made the whew comment to is the 14 same statement that's contained in all 15 antidepressants, is it not? 16 A No, there's a variation 17 from antidepressant to antidepressant in the 18 actual wording. 19 Q Are you sure of that? 20 A Yes, sir. 21 Q In what way is there a 22 variance? 23 A I mean, the words. This 24 is not exactly the same as the words used in other 761 1 antidepressant drugs. The gist of it is similar, 2 but some of them have different words than this. 3 Q Why were you concerned, 4 then, if the gist of the same was -- the gist of 5 the words were the same in all of the 6 antidepressants? 7 A Well, I wanted the label 8 to reflect what my feeling was, I mean my personal 9 belief was, was the right information to give to 10 physicians, and since the FDA really didn't 11 discuss that part of the label with us, we didn't 12 know what they were going to propose. 13 Q Well, were you afraid 14 that they were going to propose or was there a 15 concern that they were going to propose a label 16 that indicated that in some individuals there 17 might be a possibility that suicidality could be 18 caused by ingestion of Prozac? 19 A No, sir, there wasn't -- 20 there wasn't any specific concern about that at 21 all, but they had changed the format of their 22 labels around this time, and we didn't have any 23 precedence for what they were going to put in, and 24 we were concerned about what they were going to 762 1 put in about drug interaction, about the kinetics, 2 what they were going to say about elderly people, 3 what they were going to say about liver 4 dysfunction -- 5 Q No, I'm talking about 6 suicide. 7 A I know, but I'm saying 8 across many different segments of the label we 9 were concerned about what they were going to put 10 in, and we were very concerned about the length of 11 this label, which was a lot longer than comparable 12 antidepressants. 13 Q But the statement 14 concerning suicide is essentially the same, isn't 15 it? 16 A Yes. So we were 17 comfortable with that language, but not knowing 18 where they were going when they were increasing 19 the size of the label, I would have felt 20 uncomfortable if they had used some other terms, 21 but we weren't specifically worried about that. 22 Q Well, hadn't there been 23 discussions between Lilly and the FDA specifically 24 concerning what language would be contained 763 1 regarding suicide? 2 A No, I don't -- not to my 3 knowledge. 4 Q None whatsoever? 5 A Not to my knowledge. 6 Q Had there been 7 discussions within the Lilly organization 8 concerning what language should be contained? 9 A This is the original 10 label; no, sir. 11 (THOMPSON EXHIBIT NO. 41 MARKED FOR 12 IDENTIFICATION.) 13 14 (SHORT BREAK TAKEN.) 15 MR. BOUR: Tape No. 9, 16 the time is 3:18. 17 THE WITNESS: I have gone 18 through this one quickly, but I haven't read every 19 word. 20 Q (BY MR. SMITH) All 21 right. Doctor Thompson, Exhibit 41 is a document 22 apparently directed to Doctor Dorothy Dobbs by 23 Doctor R.P. Bergstrom? 24 A Yes, sir. 764 1 Q And it states that: 2 Attached are Doctor Lemberger's and my comments on 3 the package insert for fluoxetine. We will be 4 glad to meet with you to discuss any of these 5 changes. Correct? 6 A Yes, sir. 7 Q And Doctor Lemberger 8 we've identified earlier -- you've identified him 9 earlier as an eminent pharmacologist? 10 A And psychiatrist. 11 Q And psychiatrist. We've 12 identified Doctor Dobbs as being a well-respected 13 psychiatrist? 14 A Yes, sir. 15 Q And who is Doctor 16 Bergstrom? 17 A He's a statistician. 18 Q All right. 19 A I'm sorry, Doctor 20 Bergstrom is a person who does pharmacokinetics. 21 He's a pharmacologist who does the time course of 22 drug concentrations and effect, I misspoke. 23 Q Turn to page -- well, 24 again there is some general statements on the 765 1 first one -- two pages, is there not? 2 A Yes, sir. 3 Q Turn to the page that is 4 marked 364 space 2866. 5 A Yes, sir. 6 Q Do you see where it says 7 precautions that they are talking about there? 8 A Yes, sir. 9 Q It says, Add: A 10 statement should be added that patients who are 11 considered possible suicide should be considered 12 for hospitalization. Correct? 13 A Yes, sir. 14 Q Were you aware that that 15 was a possibility of what the Prozac label should 16 include? 17 A I'm not sure I've ever 18 seen this document, but that certainly was 19 something that was discussed in my presence. 20 Q Certainly this is a 21 document that you haven't been shown in 22 preparation for your deposition, is it? 23 A No, sir. 24 Q This is the first time 766 1 you've ever seen this document? 2 A I don't recall having 3 seen it before. 4 Q Who was it that was 5 suggesting at Lilly that the package insert should 6 contain a statement that, quote, patients who are 7 considered possible suicide should be considered 8 for hospitalization? 9 A I assume from the cover 10 that it was either Doctor Lemberger or Doctor 11 Bergstrom. 12 Q All right. But was this 13 ever brought up to you by them prior to -- at this 14 time -- 15 A I don't think so. 16 Q -- in August of '83? 17 A I don't think so. 18 Q Did you know anybody at 19 Lilly was suggesting that the Prozac package label 20 should have a statement added to the package label 21 that, quote, patients who are considered possible 22 suicide should be considered for hospitalization? 23 A I don't -- I don't recall 24 any specific conversations on that issue in 1983 767 1 at all. 2 Q Do you have any objection 3 to a statement such as that being contained in the 4 Prozac package insert, Doctor Thompson? 5 A No, not at all. 6 Q The next few pages are 7 apparently a working draft of the actual package 8 insert that's going to be -- that's being 9 considered, correct? 10 A Yes, sir. 11 Q Turn with me to Page 5 of 12 that package insert. Do you see there under 13 precautions, last sentence -- 14 A Yes, sir. 15 Q -- of Page 5? It says: 16 Precautions. General: The possibility of suicide 17 in seriously depressed patients is inherent in the 18 illness and persist until significant remission 19 occurs. Therefore, to -- in order to minimize the 20 opportunity for overdosage, prescriptions should 21 be written for the smallest number of capsules 22 consistent with good patient management. Correct? 23 A Yes, sir. 24 Q Now, that's basically 768 1 what was approved -- was approved by the FDA for 2 the Prozac package insert -- 3 A Okay. 4 Q -- isn't it? I mean, you 5 can compare it with the approvable letter. 6 A And the final label is 7 the same as the approvable letter's copy? Because 8 there are often changes made after the approvable 9 letter, I don't -- 10 Q Yes, there were no 11 changes made as far as I know. 12 A Well, I hope they changed 13 tryptophane at least. 14 Q It's on the top of one of 15 those pages, near the last -- 16 A Towards the end? 17 Q Yes, it's on Page 7 of 18 the label. Remember, it's got your notation -- 19 A The whew, the whew. 20 Well, it's a little different, but, I mean, it's 21 got the -- essentially the same kind of content. 22 Q Right. Now, read the 23 notation made out in the corner there by the 24 suggested label, Exhibit 41. Doesn't it say 769 1 should we not state that patients with suicidal 2 ideation should be considered for hospitalization? 3 A There are a couple of 4 words I have trouble reading, but your 5 interpretation is as good as any I can come up 6 with. 7 Q Well, I don't want to put 8 words in your mouth. See if you can read what it 9 says. 10 A Well, I can't really read 11 suicidal ideation or whatever that word is above 12 the caret, but I don't have any better words than 13 the ones you chose. 14 Q Those words I've chose 15 seem appropriate -- I mean, or seem to be an 16 accurate interpretation of what was written? 17 A As accurate as I could 18 be. 19 Q And it fits in with what 20 the general statements were on Page 2866 of the 21 general comments; Add: A statement should be 22 added that patients who are considered possible 23 suicide should be considered for hospitalization? 24 A Yes, sir. 770 1 Q And in fact, the original 2 draft or one of the earlier drafts that they have 3 got marked out says a statement should be added 4 patients with suicidal ideation should be 5 considered for hospitalization, does it not? 6 A Where is that? 7 Q That's back on page 2866 8 of the general comments. 9 A Well, the original text 10 before that was handwritten in and the line was 11 struck out says, quote, patients with suicidal 12 ideation should be considered for hospitalization, 13 yes, sir. 14 Q All right. Now, in 1983, 15 Doctor Dobbs had the same information before her 16 as you did, did she not? 17 A Plus all the rest of the 18 information. 19 Q You think she probably 20 had more information than you did? 21 A Well, she was writing the 22 NDA, so she had all the information we had. 23 Q All right. Doctor 24 Lemberger had the same -- access to the same 771 1 information that you did? 2 A He had access to it. 3 Q So did Doctor Bergstrom? 4 A Yes, sir. 5 Q Doctor Lemberger, in 6 fact, actually conducted some of the earlier 7 clinical trials on Prozac. 8 A He gave the first dose of 9 Prozac to a human being -- 10 Q Oh, is that right? 11 A -- on May 11th, 1976. 12 Q Was there a package 13 labeling committee in August 1983? 14 A I don't think so, because 15 I -- I don't recall that we've ever had one that 16 was called that, but remember, I wasn't in 17 regulatory affairs at the time. That was largely 18 in regulatory rather than a clinical research 19 issue. 20 Q Well, I thought you said 21 that there was a committee whose specific role it 22 was to review package inserts and labeling? 23 A To change them. I 24 mean -- 772 1 Q Changes -- 2 A -- for adding new safety 3 information to the label there was a specific 4 committee, and that was later, though. That was 5 after I had become responsible for regulatory. 6 Q All right, there wasn't 7 such a committee in existence? 8 A Not -- not that I'm aware 9 of. 10 Q Who were the individuals 11 that were responsible for label changes in the 12 1990's when this Prozac suicide issue resurfaced? 13 A In the 1990's when it 14 resurfaced? 15 Q Yes. 16 A Well, it would have to go 17 Wood, Herr, Perelman, me, and then under me there 18 were various people in the chain of command. Max 19 Talbott at some times reported directly to me, 20 sometimes there was somebody in between. And at 21 various times on the CNS side were people -- Dan 22 Masica, Bob Zerbe, who were the directors of the 23 CNS division, and then there were a group of 24 physicians there, and then there were a group of 773 1 Ph.D. and non-Ph.D. regulatory scientists under 2 Doctor Talbott. 3 Q And all those individuals 4 would have made up the label change group -- 5 A Well, the -- 6 Q -- committee? 7 A There wasn't a formal 8 committee that carried that title. Nonetheless, 9 label changes could originate in at least two 10 different ways. One is anybody at any time could 11 suggest a label change which would have been -- 12 during that interval, it would have been brought 13 to my staff, because I had a staff meeting with 14 the senior people in medical. In addition to 15 which there was a formal quarterly meeting, which 16 was called the DEN Quarterly Review Meeting, at 17 which we reviewed all of the adverse events across 18 all of the products, and at that meeting 19 specifically one of the issues was should we 20 change the label, because that's when we did far 21 more statistical tests to see if there were 22 increased frequencies and so forth, or anything 23 new that had popped up. 24 774 1 (THOMPSON EXHIBIT NO. 42 MARKED FOR 2 IDENTIFICATION.) 3 THE WITNESS: Yes, sir. 4 Q (BY MR. SMITH) Can you 5 identify Exhibit 42? 6 A Exhibit 42 is a message 7 on E-mail initiated by Claude Bouchy and addressed 8 to me, Doctor Weinstein, Doctor Zerbe, with three 9 carbon copies. 10 Q Dated November 13th, 11 1990? 12 A Yes, sir. 13 Q Have you seen this 14 document before, Doctor Thompson? 15 A Yes, sir. 16 Q I assume you saw it at or 17 about November 13th, 1990? 18 A Yes, sir. 19 Q And have you reviewed 20 this document in preparation for your deposition 21 here today? 22 A Yes, sir. 23 Q The document is authored 24 by Claude Bouchy -- Bouchy? 775 1 A I think that's correct. 2 Q Who was he in November of 3 1990? 4 A I think he was the 5 general manager of the German affiliate. 6 Q And when you say German 7 affiliate, you're basically talking about Lilly in 8 Germany? 9 A Yes, sir. 10 Q And he was responsible 11 for all operations in Germany? 12 A Yes, sir. 13 Q Where is he now, sir? 14 A I have no idea. 15 Q When did he leave Lilly's 16 employ? 17 A I don't know. 18 Q Was he -- did he leave 19 voluntarily or was he terminated by Lilly? 20 A I have no idea. I don't 21 think I ever met him. 22 Q All right. You 23 communicated with him on a number of occasions? 24 A By electronic mail. 776 1 Q Yes. Did you ever meet 2 with him in person or talk with him over the 3 phone? 4 A The phone I can't 5 remember, but I can't place his visage, so I don't 6 think I ever met with him. I don't know what he 7 looked like. 8 Q Is he a medical doctor? 9 A I don't think so. 10 Q He mentions Hans Weber in 11 the first sentence, does he not? 12 A Yes, sir. 13 Q Hans Weber is a medical 14 doctor, is he not? 15 A Oh, yes, sir. 16 Q Is he a psychiatrist? 17 A No, sir, he's a superb 18 internist, pulmonary specialist. 19 Q Does Doctor Weber have a 20 private practice? 21 A You know, I don't -- I 22 don't think so, but I don't really know what the 23 Lilly practice is in Germany for Doctor Weber. We 24 do have physicians in Europe who in fact do -- I 777 1 know them who in fact do have private practices, 2 and all of us in the US are given ten percent of 3 our time to attend at the university, but I don't 4 know whether Doctor Weber does or not. He's a 5 very distinguished academician, so he certainly -- 6 at one time he was seeing a lot of patients. 7 Q Has he had some 8 university lectureships or professorships? 9 A I can't tell the details 10 of his CV, but he's had a very distinguished 11 record before he came to Lilly. 12 Q All right. And I assume 13 you hold him in the highest regards? 14 A Yes, sir. 15 Q Mr. Bouchy says: Re 16 adverse drug event reporting - Suicide 17 Fluoxetine. Hans Weber and I have had problems 18 with directions our safety people are getting from 19 the corporate group, drug epidemiology unit, and 20 requesting that we change the identification of 21 events as they are reported by physicians. Then 22 he gives a number of a particular event, correct? 23 A Yes, sir. 24 Q And he says: On this 778 1 one, our safety staff is requested to change the 2 event term "suicide attempt" as reported by the 3 physician to "overdose". Correct? 4 A Yes, sir. 5 Q Then he gives another 6 adverse event code number, does he not? 7 A Yes, sir. 8 Q And he says: On this 9 one, it is request that we change from "suicidal 10 ideation" to "depression". Correct? 11 A Yes, sir. 12 Q It goes on to say: Hans 13 has medical problems with these directions and I 14 have great concerns about it. I do not think I 15 could explain to the BGA, to a judge, to a 16 reporter or even to my family why we would do this 17 especially on this sensitive issue of suicide and 18 suicide ideation. At least not with the 19 explanations that have been given to our staff so 20 far. I am quoting -- that is Doctor -- Mr. Bouchy 21 is quoting -- I am quoting, when an overdose is 22 taken in a suicide attempt, our research 23 physicians prefer to list the event term overdose 24 even if when tracking suicides we always look at 779 1 all overdoses in suicide attempt reports, end 2 quote within a quote. 3 He goes on to say: This 4 issue has been argued back and forth for about a 5 month between Bad Homburg and Indy, therefore, I 6 am bringing it to your attention and await your 7 directions. Regards, Claude. Correct? 8 A Yes, sir. 9 Q Mr. Bouchy generally 10 didn't direct and report directly to you, did 11 he -- 12 A Oh, no, sir. 13 Q -- at that time? It 14 appears to me that he says the issue has been 15 argued back and forth, so I'm bringing it to your 16 attention, Doctor Thompson, is that right? 17 A Yes, sir. 18 Q I get the sense from 19 reading this that he was dissatisfied with the 20 results of the discussion up to that date and felt 21 like it needed your attention. Would that be 22 accurate? 23 A I don't know whether he 24 was particularly addressing it to me or Allan or 780 1 Bob or the three of us together, but you've got it 2 right. 3 Q He wanted some direction 4 from somebody with some responsibility in this 5 connection, did he not? 6 A Yes, sir. 7 Q Do you know who he had 8 been talking to previously prior to issuing this 9 document? 10 A No, I don't, because I 11 remember this debate going back and forth, and 12 some people from Germany talked with me. I'm not 13 sure whether they talked to the FDA directly or 14 not. I was talking to the FDA. I mean, this 15 was -- I thought it was even more than a month. 16 There was a lot of communication about this. 17 Q All right. What -- was 18 November 13th, 1990 the first time you had been 19 aware of this situation? 20 A Oh, I don't think so. 21 Q Okay. So you were aware 22 that there was some -- I don't know whether it's 23 controversy, debate or argument, I think argued is 24 the term Mr. Bouchy uses. There was some 781 1 discussions about this of which you were aware 2 before you received this? 3 A Yes, sir, I think so. 4 Q All right, so you were 5 not surprised that this issue was coming up from 6 Germany? 7 A No, there had been a 8 debate between the Germans and the FDA and us for 9 a long period of time about the use of the 10 Co-Start terms. 11 Q Is Mr. Bouchy accurate 12 when he says that there was a request made by Indy 13 that Germany change an event term from suicide 14 attempt to overdose? 15 A Probably. 16 MR. FREEMAN: Explain 17 what that means so we can clear it up. 18 MR. SMITH: Well, I'll 19 ask him for any explanation if I need it, 20 Counsel. 21 Q (BY MR. SMITH) Was there 22 a situation -- 23 MR. FREEMAN: If you 24 don't want to know, that's fine with me. 782 1 Q (BY MR. SMITH) Was there 2 a situation that had existed, that happened where 3 there were instructions given by Lilly in 4 Indianapolis to change a specific event term from 5 suicidal ideation, as reported by a physician, to 6 depression? 7 A Well, you're now 8 confounding the problem because the issue was how 9 to use the FDA's Co-Start terms, and to make sure 10 that all of the terms that were similar were 11 appropriately grouped under a single 12 classification term. I think it was at this time 13 that our standard operating procedure for DEN was 14 that changes in classification were made at the 15 country of origin rather than in Indianapolis. 16 And so I think the content of this was that 17 because they had misclassified those events, it 18 was a request made from Indy back to Germany, 19 since they had control of the actual text, to use 20 the correct classification term. 21 Q Okay. Are you -- are you 22 saying that in fact the event term "suicide 23 attempt" was a misclassification? 24 A Yes, sir, I don't think 783 1 there was such a term in the Co-Start dictionary, 2 the best as I can recall at that time. 3 Q But there wasn't any 4 prohibition in reporting suicide attempt as an 5 event term to the Food and Drug Administration, 6 was there? 7 A Yes, there was. Now we 8 have to expand on that one because I can't answer 9 that one yes or no at all. 10 Q Okay. The only times 11 I've wanted you to answer something yes or no was 12 when I was reasonably sure it could be answered 13 yes or no, but I'm not about that. 14 A There were a whole series 15 of discussions with the FDA about whether they 16 preferred us -- this is not regulation, whether 17 they preferred us to use the Co-Start terms 18 without adding or changing them. The gist of 19 those discussions had to do with a lot of things 20 other than suicide attempt and so forth, was that 21 there opinion was that we should be very, very 22 cautious in changing the terms because they would 23 then partition the data set -- 24 Q Let me interrupt you and 784 1 ask you, when you say they wanted us to be very 2 cautious about changing the terms, you mean 3 employing a different term than was being used by 4 Co-Start? 5 A Or just using a new term 6 in the midst of the marketing experience or 7 clinical trials with the drug. 8 Q All right. Go ahead, I 9 interrupted you. 10 A And the reason was that 11 if you -- if you added a new term, suppose you 12 added suicide attempt or whatever you wanted to 13 add, before that date, all of the adverse events 14 that were then going to go into that box would be 15 stuffed in some other box. After that date, when 16 you redefined what that box meant, presumably 17 those terms would go in that box, but it would 18 take some time before you trained investigators 19 and Lilly staff worldwide to classify it right. 20 So if you were to go back -- I mean, suppose you 21 had identified suicide attempt and say here's what 22 constitutes a suicide attempt, put it in this box, 23 you couldn't go to a single box and find all the 24 suicide attempts, you would have to before that 785 1 date go back to wherever they had been assigned 2 before that -- 3 Q I guess it would have 4 been overdose? 5 A Whatever, I think it was 6 overdose. After that, you would have to go to the 7 new box you created, plus you would still have to 8 look in overdose for a while to make sure people 9 hadn't just been using the old rules, because, you 10 know, we -- when we had investigators, we'd give 11 them the Co-Start dictionary and the glossary, but 12 we wouldn't update it every few months. So we'd 13 have a whole bunch of investigators worldwide that 14 had one set of rules, we couldn't go out and 15 modify those all over the world the next day. 16 So what the FDA's advice 17 was, was to be very, very cautious at changing the 18 classification terms, because what they told us 19 was that in fact it didn't matter what we called 20 them anyway, they were going to classify them 21 themselves and they weren't going to change the 22 Co-Start terms. 23 Q So, what if you had 24 reported to the United States Food and Drug 786 1 Administration the term suicide attempt, what's 2 your impression concerning how they would have 3 classified that? 4 A Oh, it would have gone 5 into overdose, because we did what they told us to 6 do. 7 Q Okay. 8 A I mean, I think what 9 you're saying is that we weren't obligated, 10 there's no rule that says you have to use 11 Co-Start. 12 Q Well, the general rule 13 is, is that you're to accurately report to the 14 United States Food and Drug Administration the 15 adverse experience that is being reported to you, 16 isn't it? 17 A Exactly. So that's why 18 the -- whatever the text is, is always reported to 19 them in full and -- 20 Q What do they do with that 21 full text? 22 A Well, they don't pay any 23 attention to our classification terms. What they 24 do is they go read the text and they pay a 787 1 contractor to assign the Co-Start terms the way 2 they feel they should be assigned, and they don't 3 match up for one or one either. 4 Q At that time did overdose 5 map to suicide attempt? 6 A Yes, sir. 7 Q Are you sure about that? 8 A Yes, sir, because that's 9 why we were using it. 10 Q And there was no mapping 11 of suicide attempt to anything? 12 A No. I may have misstated 13 because you may have gotten it backwards. I think 14 suicide attempt was assigned to the Co-Start 15 classification term overdose. 16 Q All right. So it could 17 be reported suicide attempt to the FDA, the FDA 18 already knew they had suicide attempt as a term 19 within the overdose category, is that right? 20 A I don't think they had 21 suicide attempt as a term, but their rules -- 22 their operating rules would have had their 23 contractor assign that to the term overdose. 24 That's what they told us. 788 1 Q So you knew then that if 2 you reported suicide attempt to the FDA, that it 3 would go into the appropriate box? 4 A Well, we knew what rules 5 were being used by their contractor; we didn't 6 always agree with their assignments because, as I 7 told you, we took a lot of trouble making what we 8 thought were the right assignments and they 9 wouldn't necessary agree with ours. 10 Q Well, your responsibility 11 is to report to the FDA as accurately as possible 12 what the adverse event was, isn't it? 13 A Yes, sir. But there's 14 absolutely no regulatory requirement to even use 15 any classification terms. That's something Lilly 16 decided to add to the modified 1639. 17 Q So why were you changing 18 suicide attempt to overdose, just to have it 19 automatically go into the more broad, generic 20 term? 21 A Because the official term 22 was overdose, to which suicide attempts were 23 mapped. 24 Q What did you do with 789 1 someone that blew out their brains, how did you 2 report that to the FDA? 3 A Well, it would have said 4 patient's physician says blew out their brains, 5 and -- I'm not sure what the term is, if you show 6 me the Co-Start dictionary from this era, I'll try 7 to pick out the right term for you. 8 Q Was there a Co-Start term 9 at that time for suicide? 10 A I don't think so. 11 Q Was there a Co-Start term 12 for injury, self-inflicted; intentional injury 13 self-inflicted? 14 A I can't remember. There 15 certainly was for injury, but I'm not sure what 16 the classified -- 17 Q If at that time the 18 United States Food and Drug Administration didn't 19 have an event term suicide and didn't have an 20 event term attempted suicide, it seems to me that 21 you automatically have a difficulty in getting 22 information from the Food and Drug Administration 23 concerning how many patients attempted or 24 completed suicide on any particular medication, 790 1 Doctor Thompson, is that correct? 2 A You might have difficulty 3 in getting the information from the FDA according 4 to their use of the Co-Start terms. I mean, you'd 5 have to understand exactly what the contractors 6 were told to do in terms of using those terms. 7 But there wouldn't be any difficulty in getting 8 the narrative account of the event itself. If you 9 wanted to look for -- I mean, our procedure was if 10 you wanted to look for people who died of 11 self-inflicted wounds, we would go, first of all, 12 and look under deaths, and we'd search under all 13 deaths, and then we'd search under whatever the 14 narrative terms were we were looking for. 15 Q Is there a Co-Start term 16 deaths? 17 A I don't think so. That 18 was another big controversy, because the FDA felt 19 strongly that we shouldn't be classifying events 20 as death, that death was an outcome, not an 21 event. That was a huge controversy. 22 Q Well, if you had suicide 23 attempt by overdose, I can see how it could be 24 argued that you could find the suicide attempt by 791 1 looking at overdose, but what if you had a suicide 2 attempt, attempted by hanging, how would you find 3 that if there's no Co-Start term for suicide 4 attempt? 5 A Exactly the way that 6 Doctor Beasley, et al., searched the clinical 7 trial data to try to find everything that was 8 related. We looked for actual English language 9 strings of characters that might be related, like 10 S-U-I, H-A-N-G, G-U-N-S, and tried to find in the 11 narrative descriptions of adverse events any 12 English text string that fit those strings, and 13 then we'd go read the text and see what it said, 14 just to make sure we hadn't misclassified them. 15 Remember, the 16 classification terms are just buckets in which you 17 can conveniently -- it's like a library indexing 18 system, and it's a place of putting all the 19 events. So it actually, theoretically, you could 20 say they don't even need to be English language 21 terms, you could say classification term nine six 22 three is where we're going to put all the suicide 23 attempts, and it would work just as well. 24 Q Well, is the reason for 792 1 reporting the event to the Food and Drug 2 Administration so that the Food and Drug 3 Administration can know what particular adverse 4 events are being reported? 5 A Yes, sir. That's why we 6 have a full narrative description and all the 7 other things in the 1639. 8 Q Do you agree with Doctor 9 Weber where he says -- or Doctor -- or Mr. Bouchy 10 and Doctor Weber where they say I do not think I 11 could explain to the BGA, to a judge, to a 12 reporter or even to my family why we would do 13 this, especially on the sensitive issue of suicide 14 and suicide ideation? 15 A Oh, I understand and 16 sympathize with their problem, but our primary -- 17 our primary use of adverse events was to fulfill 18 the FDA regulations, and we worked very closely 19 with the FDA in how we developed the DEN system 20 and how we used it. And the FDA wanted us to put 21 all of the things that were alike in one bucket 22 and to define that as best we could, and that was 23 their choice of bucket. I mean, given my own 24 purpose, I would have defined some more buckets, 793 1 but they said, hey, be careful doing that. 2 And again, I mean, if 3 Lilly had not been doing any research at all and 4 we were going to start doing new research, we 5 could have made up a whole different set of 6 terms. In fact, we're faced with that right now 7 because there's another very well established 8 dictionary called the WHO-ART dictionary, and 9 there's a -- 10 Q Based on the World Health 11 Organization? 12 A Adverse Reaction Terms, 13 and those are different terms than the Co-Start 14 terms. And so far people have only been able to 15 figure out how ninety percent of them map from one 16 to the other. And so some companies use one, some 17 companies use another. Most regulators don't even 18 use a consistent set of terms to do their own 19 analysis. The FDA at least maintains Co-Start. 20 Q Why were you reporting 21 suicidal ideation as depression? 22 A Because that's what we 23 were told to do. 24 Q But you're going to have 794 1 a bucket that has -- of depression, right? 2 A Yes, sir. 3 Q And you got a drug that 4 is an antidepressant, correct? 5 A But, you know, everything 6 else -- I mean, every one of our drugs, an 7 antibiotic in which a patient was reported to have 8 depression, would go into that same bucket. 9 Q I understand that. But 10 you wouldn't expect to see depression being 11 reported as an adverse event from an 12 antidepressant medication, would you? 13 A But Mr. Smith, you're 14 misinterpreting the use of the classification 15 terms. I want to emphasize again to you, one, is 16 Lilly is not under an obligation to use any 17 classification terms at all -- 18 Q They're obligated, 19 though, to report the adverse event in the most 20 accurate manner possible, are they not? 21 A And we actually added a 22 whole extra page to the 1639 to be sure that we 23 could, in fact, fulfill that obligation. So all 24 the narrative, whatever people told us, whatever 795 1 our analysis showed was all there in detail, and 2 that's what the FDA used in their 3 classifications. Now, what we did was we, as 4 closely as we could, we followed the FDA rules in 5 classifying things, because, remember, one of our 6 jobs also was to look at increased frequency, and 7 so we needed to have a consistent rule so that all 8 of the events of any kind -- any kind, but say 9 suicidal ideation, we needed to have one bucket 10 where you could go to and with some reliability 11 expect to find all of the suicidal ideation in 12 that bucket and not split them between two 13 buckets. 14 Q But don't you think -- 15 don't you agree that it makes it harder to find 16 suicide attempt if you put it in overdose? 17 A No, sir, I don't, because 18 whatever we want to partition that bucket -- first 19 of all, we wouldn't even stop looking with just 20 that bucket, but hopefully we had all of them in 21 that bucket. We would then go and either read 22 them or use a computer to look for English text 23 strings in the narrative descriptions of the 24 events to find out which ones were which. 796 1 Q What if I want to get 2 that information as a private citizen from the 3 United States Food and Drug Administration, what 4 if I want to know, Doctor Thompson, how many 5 people have attempted suicide on Prozac? 6 A If you asked them, in 7 this era, what they would tell you is to put an 8 FOI in using the classification term of overdose. 9 Q All right. 10 A And you could get all of 11 those 1639's and you could read them yourself. 12 Q Okay, wait a minute, I 13 don't think that would -- if I did that, if I just 14 got overdose, then I wouldn't be able to get those 15 individuals who had attempted suicide by slashing 16 their wrist? 17 A No, I think those were 18 classified the same. 19 Q As overdose? 20 A Yes, sir. That was one 21 of the issues. Yes, sir. That's exactly why we 22 followed what the FDA was doing. I'm not telling 23 you it may be the -- it may not be the most 24 logical thing in the world, but remember, the 797 1 classification terms are just like the Dewey 2 Decimal system in terms of where you're going to 3 go and look for books in the library. 4 Q Well, let's look at 43. 5 (THOMPSON EXHIBIT NO. 43 MARKED FOR 6 IDENTIFICATION.) 7 THE WITNESS: Yes, sir. 8 MR. SMITH: What's the 9 number? 10 MS. ZETTLER: 43. 11 Q (BY MR. SMITH) Exhibit 12 43 indicates the Eli Lilly and Company spontaneous 13 reporting of adverse events, postmarketing, 14 through March 31st, 1990, correct? 15 A Yes, sir. 16 Q And then it has the 17 events terms listed on the left-hand side? 18 A Yes, sir. 19 Q Then the number of events 20 on the right-hand side? 21 A Yes, sir. 22 Q There's intentional 23 overdose and overdose -- 24 A Yes, sir. 798 1 Q -- correct? And then 2 there's suicide attempts and suicide attempt, no 3 drug used? 4 A Yes, sir. 5 Q So apparently there was 6 event terms for suicide attempt and suicide 7 attempt, no drug used? 8 A I think at this time some 9 people had tried to add those terms to the ELECT 10 dictionary. 11 Q On March 31st, 1990 -- 12 A That's what it says. 13 Q -- there were those 14 terms? 15 A Well, but I think that we 16 in fact changed that because, when we consulted 17 with the FDA, they advised us to be very, very 18 cautious in changing the classification terms that 19 we were using. 20 Q That's just four months 21 after this controversy was brought to your 22 attention by Mr. Bouchy and Doctor Weber? 23 A Isn't this March 1990? 24 Q Yes. In fact, it's prior 799 1 to -- 2 A Yes, sir, this was -- 3 Q -- 1990. 4 A This was a continuing 5 question. You've just chosen one particular 6 adverse event term, but we had this discussion 7 ongoing with a whole bunch of them. 8 Q Well, didn't you say 9 earlier that there wasn't any event term suicide 10 attempt? 11 A In the Co-Start 12 dictionary, I believe that's correct. 13 Q Well, apparently there is 14 because it's listed. 15 A No, we'd have to pull the 16 Co-Start dictionary and see whether I'm right or 17 wrong in my memory. I think these were terms that 18 at this period somebody at Lilly had tried to add 19 to -- remember, I told you ELECT dictionary wasn't 20 precisely the same? 21 Q Uh-huh. 22 A And I can't remember when 23 the exact conversations occurred, but in a long 24 series of ongoing conversations with the FDA, when 800 1 we talked about the issue of adding terms and 2 redefining them, what they advised us was, one, we 3 could do it, but to be very careful with it 4 because they weren't going to change their terms. 5 Q But, Doctor Thompson, it 6 looks like it was being done prior to this. 7 A I agree with you. 8 Q You had situations where 9 suicide attempt was indeed being reported to the 10 United States Drug Administration -- Food and Drug 11 Administration as suicide attempt. 12 A I'm assuming that that 13 was what was typed on the 1639. But, remember, 14 they paid no attention to what our classification 15 terms were. 16 Q See what you're going to 17 get is you're going -- if somebody asked the Food 18 and Drug Administration in March 1990 how many 19 people had attempted suicide while on Prozac, you 20 would not get back overdose, you would get back 21 the data as described as an event term and suicide 22 attempt, and you'd only get twenty-two, when what 23 you're saying is in fact there were indeed one 24 hundred and ninety-three. 801 1 A There are two things 2 wrong with your statement. One is that if you 3 wrote to the FDA, you would get back their 4 classification terms, which I've already told you 5 don't agree with ours, and they specifically 6 didn't add any. So we could have typed anything 7 you wanted to in that part of our modified 1639 8 and they didn't pay any attention to it. 9 The second thing is I 10 don't know what all is in the bucket called 11 intentional overdose or overdose. So you were 12 making an assertion about that; for all I know, 13 all kinds of things might be defined as being in 14 that bucket. 15 Q Wouldn't you think that 16 in ninety-nine point nine percent of the time, an 17 intentional overdose would be a suicide attempt? 18 A No, we already talked 19 about that from my own clinical experience. 20 Q Yes. I think you said 21 about eighty-five percent? 22 A Yes, sir. 23 Q All right. 24 A That's not ninety-nine. 802 1 Q Okay. Well, it's a high 2 percentage, isn't it? 3 A I'll agree. 4 MR. SMITH: Let's take a 5 quick break. 6 (SHORT BREAK TAKEN.) 7 MR. BOUR: The time is 8 4:17. 9 * * * * * 10 EXAMINATION 11 12 BY MS. ZETTLER: 13 Q Doctor Thompson, we met 14 two days ago now. My name is Nancy Zettler, I 15 represent, along with Paul, a number of plaintiffs 16 down in Kentucky in the Fentress litigation, as 17 well as a couple of parties in Illinois. 18 A Yes, ma'am. 19 Q I'm a little bit confused 20 about the testimony you just gave Paul on the 21 Co-Start dictionary and the spontaneous reporting 22 system and the DEN database, so I would like to 23 clear a few things up, if that's okay. 24 A Yes, ma'am. 803 1 Q Let's start with the 2 spontaneous reporting system. That is the FDA's 3 computer system, correct? 4 A If you want to define it 5 that way. 6 Q Well, it's their system 7 where they collect adverse event reports and they 8 keep them in that -- on a computer database, and 9 if you wanted to retrieve them, you can ask them 10 to do that or they can do that if they need to, 11 correct? 12 A Yes, ma'am. 13 Q Okay. How is it that an 14 adverse event gets from the manufacturer to the 15 spontaneous reporting system? 16 A We mail them what we have 17 modified, but with an agreement with them, an FD 18 1639, which Lilly added a second page to and made 19 a few other changes, so we mail that to them. 20 When they get it, without looking at its content, 21 they bundle it up and mail it to a contractor, and 22 I don't know who the contractor is now, but three 23 or four years ago it was somewhere west, Kansas, 24 somewhere like that. The people in the 804 1 contracting house then would read everything that 2 was on that report and assign Co-Start terms and 3 do whatever else the FDA asked them to do, and 4 then sent back both -- my understanding is they 5 sent back both the paper as well as whatever the 6 electronic image was, which was then reviewed by 7 the Division of Epidemiology, Safety and Adverse 8 Events at the FDA, and they would have periodic 9 meetings to look at those data. But I think they 10 get back both paper -- the original paper as well 11 as getting back the electronic images from their 12 contractor. 13 Q Who at the FDA reviews 14 those? It says the Department of Epidemiology. 15 Is there one person in charge of making sure they 16 all get reviewed? 17 A There was at one time, 18 Doctor Gerry Faich was the head of it until -- I 19 think he was the head of it until 1990, for 20 example, and I think Chuck Anello is the current, 21 but I don't remember whether he's an acting 22 director or the director that replaced Jerry. And 23 they've got a whole staff of people under them, 24 they've got physicians under them, as well as 805 1 non-physicians, and they can be reviewed by the 2 divisions. The division can independently be 3 reviewing the same adverse event. 4 Now, I'm not sure what 5 all the divisions do with, but I am sure that when 6 Faich was running that division, that they would 7 have very formal meetings and they would look at 8 certain -- they would pick out certain signals, 9 combination of a drug and an adverse event or a 10 group of adverse events, and they would meet 11 periodically and go over that with or without the 12 people from the division. 13 Q And this is done for all 14 drugs, not just Lilly drugs? 15 A Yes, sir -- yes, ma'am, 16 as far as I know. 17 Q Okay. Other than the 18 1639's, are there other methods to report an 19 adverse event that will -- ended up in the SRS? 20 A I don't know what they do 21 with -- explicitly with the periodic reports, 22 because for three years after marketing you have 23 to file quarterly reports, and then annual reports 24 thereafter, and this is just on the postmarketing 806 1 events, and I honestly don't know whether any of 2 those get into that computer database or not. 3 Q Is -- to your knowledge, 4 is there another database for adverse events that 5 are reported in clinical trials as opposed to 6 spontaneously? 7 A I'm not sure database is 8 the right way to call it. The division is 9 primarily responsible for IND's, although with 10 biologics it's a whole group that does nothing but 11 IND stuff, and so they're getting reports. But if 12 in fact they have organized that into what you 13 meant by a database, I'm not aware of it. 14 Q I'm talking about other 15 than having to go through and look at every 1639 16 or every quarterly report; obviously, if the FDA 17 wants to do some sort of epidemiological review on 18 the adverse events that have been reported for a 19 certain drug, they are not necessarily going to 20 sit there and look through every piece of paper in 21 the IND or the NDA, correct? 22 A I wish that that were 23 true, but I'm not aware that they have any sort of 24 electronic collection of the clinical trial data. 807 1 I'm not aware that they have an electronic 2 collection of anything other than the 1639's; I 3 mean, what we dump in as the periodic report. So 4 if they've got it, I'm not aware of it. 5 Q Do you know of anybody at 6 the FDA who went and did their own physical search 7 of all the, say, clinical report forms that you 8 submitted to them on Prozac once the issue of 9 suicidal ideation and the use of Prozac came up, 10 or did they rely on you, meaning Lilly, to provide 11 them with that information? 12 A Well, one, I don't know 13 of anybody who did that specific search, although 14 I'm aware of people at the FDA or academia who 15 have gone to the FDA and done other searches for 16 other adverse events, but I'm not aware that 17 anybody at FDA that did that, but I also would not 18 necessarily be aware of that. 19 Q Okay, how about violent 20 aggressive behavior; are you specifically aware of 21 anybody who made a search of the IND, NDA 22 documents besides 1639's -- 23 A I'm not aware 24 specifically of that. 808 1 Q We have, through a FOIA 2 request, gotten a copy of the SRS data for 3 fluoxetine up to, I believe, it's June of 1993. 4 A (WITNESS MOVES HEAD UP 5 AND DOWN.) 6 Q In that database we ran a 7 bunch of different queries of our own to find out 8 what's in there. There are numerous reports of 9 suicide attempt. 10 A (WITNESS MOVES HEAD UP 11 AND DOWN.) 12 Q There are numerous 13 reports of overdose. There are reports of suicide 14 attempt, no drug used. There are reports of 15 intentional injury. 16 A Now, by that, do you mean 17 the FDA's classification term or -- 18 Q It's the Co-Start term 19 that they put into their computer database. 20 A It's their Co-Start term, 21 okay. 22 Q Okay. So your testimony 23 earlier that as far as you know this term suicide 24 attempt is not included in the Co-Start, at least 809 1 from my experience doesn't appear to be true. 2 A As of that time, that was 3 my best recollection. 4 Q As of 19 -- November of 5 1990? 6 A Yes, ma'am. 7 Q So it's your testimony 8 that since November of 1990, they have added the 9 term suicide attempt? 10 A I don't know, because I 11 really haven't been in the loop that much since 12 then. 13 Q I don't have a copy of it 14 with me, I'll admit that to you, Doctor, but I 15 have looked at the 1989 version of the Co-Start, 16 and there is the event term "suicide attempt" in 17 there. 18 A Then I misspoke. I gave 19 you the best I could remember. 20 Q Then can you explain to 21 me why you would not have mapped a suicide attempt 22 where no drug was used as the method to try 23 suicide under overdose? 24 A Because in my 810 1 conversations with the FDA staff, in that time 2 frame or right before that time frame, that's what 3 they advised us was their practice, and so we used 4 rules within Lilly that were as consistent as we 5 could make them in what the FDA told us their 6 practice was. 7 Q I've also taken 1639 8 forms that have been provided to us by Lilly with 9 various event terms such as suicide attempt, 10 things of that, and have run the manufacturing 11 numbers through our copy of the SRS, and they come 12 up with the same adverse event term that you do. 13 A Well, I can guarantee you 14 that in other cases -- I've never looked at the 15 data on Prozac and suicide, but on other cases we 16 have found that they have not used our 17 classification terms, and in asking that division 18 whether or not they would use them, we were told 19 specifically no, that the contractors would in 20 fact follow their own rules by reading the 21 narrative. 22 Q Who specifically told you 23 that at the FDA? 24 A Doctor Faich, among 811 1 others on his staff. 2 Q Is Doctor Faich still 3 with the FDA? 4 A No, ma'am. 5 Q Where is he now? 6 A I don't know. 7 Q Do you know if he's still 8 alive? 9 A Well, he was a few months 10 ago. He was head of one of the units of Corning 11 Glass, but I had been told -- and in fact I think 12 he told me that he was resigning from that 13 position, so I don't know what he's doing now. 14 Q Why didn't the FDA, to 15 your knowledge, just require that everybody follow 16 Co-Start as opposed to using their own event term 17 dictionaries; wouldn't that have made it a lot 18 easier? 19 A Yes, indeed. 20 Q Wouldn't that have -- not 21 have necessitated the use of a contractor on the 22 part of the FDA? 23 A It sounds like it to me. 24 Q Who else besides Doctor 812 1 Faich can you recall have told you that they were 2 going to -- 3 A Various members of his 4 staff. Anello worked for him for a long time, and 5 I met with some of their technical staff at the 6 FDA who were doing this. 7 Q Is it Mr. Anello or 8 Doctor Anello? 9 A I think it's Mister. 10 Q Is he still with the FDA? 11 A Yes, I think he is. He 12 may be a Pharm.D., I'm not a hundred percent sure 13 on that. 14 Q Have you ever seen 15 computer printouts from the SRS database, Doctor? 16 A Yes, but not on Prozac. 17 Q Okay. Do you consider 18 yourself fairly familiar with the way they're set 19 up? 20 A Long ago. 21 Q Have you ever seen one 22 where the description that is given by the initial 23 reporter is actually included on the printout as 24 opposed to just using the Co-Start term? 813 1 A I don't think so. The 2 ones that I've seen I think are fairly abbreviated 3 and they have large demographics of the patients 4 and their classification terms. 5 Q They have things like the 6 patient's sex and age, how the FDA gets the 7 report, if it's through the manufacturer as 8 opposed to a physician, correct? 9 A Yes, exactly. 10 Q They have the outcome of 11 the adverse event down there? 12 A Yes, for that little 13 panel. 14 Q And it's all set up in 15 graphs -- or not graphs, but in like a chart kind 16 of -- it all runs basically by the FDA's control 17 number, correct? 18 A That's what I've seen. 19 Q Okay. Do you know of any 20 way somebody such as Mr. Smith and I could contact 21 the FDA and find out in a computer listing type 22 form what the actual term or the initial term that 23 was used by the initial reporter was on the 1639 24 as it was reported? 814 1 A That would never be 2 changed. I mean, you can go to Lilly and get that 3 because there's an audit trial electronically. 4 You want to go to the FDA and look -- 5 Q I'm talking about the 6 FDA. 7 A Now, by term, you got to 8 distinguish between whatever the narrative was. 9 If you called us up and said this patient blew his 10 brains out, we should have in the narrative blew 11 his brains out. 12 Q Well, you've also -- I 13 have also seen 1639's -- 14 A On the second page at the 15 top we type our use of the ELECT terms. 16 Q Uh-huh. 17 A That should be in the 18 FDA's records, and when you ask to retrieve the 19 1639's, you ought to be getting a copy of that 20 two-page form. 21 Q I'm talking about the 22 computer database. 23 A Oh. 24 Q Okay, because when you 815 1 ask for the 1639's, they first tell you that you 2 need to look at the SRS database and ask them 3 specifically for 1639's. You cannot just call 4 them up or write them an FOI letter and say we 5 want to see all the 1639's, okay. There's a 6 procedure that they want you to go through at this 7 point. But in that -- and I take it that at least 8 the first page of the 1639's that you use is a 9 copy of the actual report? 10 A It's -- I think it's 11 slightly different. 12 Q Okay. The report that's 13 used by the FDA or the form that's used by the FDA 14 has a place where it says please describe event, 15 correct? 16 A Yes. 17 Q And that's either where a 18 narrative is or if the reporter says my patient 19 tried to commit suicide, that's where that goes, 20 correct? 21 A Exactly, and that's why 22 we added that second page, so we could just have a 23 long -- a bigger segment to put all the comments 24 on. 816 1 Q And on the second page is 2 also where you put the Co-Start term, correct? 3 A Exactly. 4 Q When did you start using 5 Co-Start as opposed to ELECT? 6 A Doctor Matsumoto was the 7 one who first suggested that, and it was either 8 late '82 or in '83. I think it was probably '83. 9 Q That you started using 10 Co-Start as opposed to ELECT? 11 A Oh, I'm sorry, I'm sorry. 12 Q Not SSAI. 13 A I'm sorry, I missed it. 14 Sometime in the last couple of years. 15 Q So back in 19 -- November 16 of 1990, when Claude Bouchy was talking with you 17 about what is being told by the physicians at 18 Lilly to do with adverse event terminology, you 19 were not using the Co-Start, you were using the 20 ELECT dictionary? 21 A That's my understanding, 22 because I think the change to Co-Start has 23 occurred since I was responsible for that group. 24 Q Okay. And the ELECT 817 1 dictionary at that point had suicide attempt as an 2 adverse event term, did it not? 3 A That's not the way I 4 remember it. 5 Q Well, I have some copies 6 of the ELECT that also -- that were dated 1991 7 where suicide attempt was an event term, Doctor. 8 A Well, I've testified to 9 the best of my ability, and because of this 10 document, Exhibit 43, I've told you that there was 11 in fact some change in the ELECT terms, and that 12 when I went and talked to the FDA about that, 13 their advice to me was -- not on this term 14 specifically, but in general, to be very, very 15 careful in doing that because it could lead to 16 partitioning an event and they would be concerned 17 that we would in fact be concealing, if they were 18 suspicious, the adverse event terms by dividing 19 them up into a bunch of different buckets rather 20 than putting them consistently in one bucket. 21 I mean, they had the same 22 desire that we did, to make sure that when they 23 went to a bucket and followed the rules, that all 24 terms of that kind were in the bucket. Now, there 818 1 might be a bunch of other terms, too, but when 2 they wanted to find headaches, that they could go 3 to one bucket, whatever it was, and all the 4 headaches would be there. And so they were very 5 concerned about the fact that if you split those 6 up across in a bunch of different buckets and then 7 did your increased frequency analysis, that you in 8 fact might not hit their trigger on increased 9 frequency analysis because half of them were in 10 one bucket and half were in another, where as if 11 you put them all in the same bucket, it would 12 trigger the statistical significant limit. 13 Q Would you agree with me, 14 Doctor, that if there had been -- assuming that 15 there wasn't, if there had been a bucket called 16 suicide attempt, it would have been less confusing 17 and more appropriate to put all the suicide 18 attempts, no matter what method was used, in that 19 bucket? 20 A Had there been a bucket 21 called suicide attempts, I would agree with you 22 that everything that was suicide attempt would 23 have gone in that bucket. But again, remember 24 that our own analyses, that I'm sure that you've 819 1 seen, when we wanted to ask the question did we 2 have any of the certain kind, we went back and did 3 string searches, that's what I did when I was 4 responsible for Fenestra when I first came to 5 Lilly. And so we again only used the 6 classification terms to look at frequencies and to 7 make a first cut on where to go to find something. 8 Q When Doctor Stadel did 9 his review of the SRS for the 1981 PDAC -- 10 A I'm sorry, I don't know 11 what the PDAC is. 12 Q Pharmaceutical Drug 13 Advisory Committee or Psychological Drug Advisory 14 Committee. 15 A And I don't know who 16 Doctor Stadel is. 17 Q Doctor Stadel is a 18 statistician and he did the epidemiology review. 19 A At Lilly? 20 Q No. 21 A Oh, I'm sorry. 22 Q At the FDA. 23 A Okay. 24 Q Okay, and he did his 820 1 review of the spontaneous reporting system 2 searches that they did, do you know if he was 3 privy from a search of that system, the 4 spontaneous reporting system itself, to all of the 5 actual terms or all of the descriptions of the 6 events that were provided to him by Lilly? 7 A I don't know what's in 8 their computer database beyond what I've already 9 testified to. It's my understanding, from what 10 they told me, that they don't have the Lilly 11 classification terms or any other company's 12 classification terms, and I'm not sure how much of 13 the actual narrative is in the computer, so I 14 don't know from my own knowledge. 15 Q So if your recollection 16 is true and there was no suicide attempt bucket 17 for at least a period of time during the time that 18 Prozac was being developed and had been marketed, 19 then if I went to the spontaneous reporting system 20 and typed in the words suicide attempt, not 21 knowing that suicide attempts were being mapped to 22 overdose according to Co-Start, regardless of the 23 method, I'm not going to see a lot of suicide 24 attempts, am I? 821 1 A Which database, FDA's or 2 ours? 3 Q FDA's. 4 A If what you say is true, 5 I think you've got it correctly. 6 Q If there is in fact a 7 suicide attempt bucket and Lilly was not mapping 8 suicide attempts to that bucket, then something is 9 seriously wrong with the spontaneous reporting 10 system's information on Prozac, isn't there? 11 A No, because we rely, just 12 like the FDA relies, on the full narrative 13 discussion of events. 14 Q I mean, you said yourself 15 you didn't believe that the full narrative 16 discussion of events was in the spontaneous 17 reporting system itself. I'm narrowing it down to 18 the spontaneous reporting system. 19 A In the computer records 20 thereof, but they've got the entire report. I 21 mean, even if we appended a hospital record of 22 fifty pages to the report, they've got all that. 23 Q Do you seriously believe 24 that the FDA has the manpower or the time to go 822 1 through every single 1639 as opposed to relying on 2 their contractors to report accurately what the 3 adverse event terminology is? 4 A Yes, I do. 5 Q Do you remember the name 6 of the contractor? 7 A No, I'm sorry, I don't. 8 Q Are you aware that in the 9 Co-Start dictionary that there is a section where 10 it sets out the WHO-ART terminology and how it 11 maps to the Co-Start terminology? 12 A It only is a ninety 13 percent map, and that's fairly recent. As a 14 matter of fact, we're still working -- 15 Q In the 1989 -- 16 MR. FREEMAN: Let him 17 finish. 18 MS. ZETTLER: I'm sorry, 19 I didn't mean to cut you off. 20 A Yes, we're still working 21 on it because there's so much disagreement about 22 that mapping that in fact the British authorities 23 were trying to create a third dictionary. 24 Q To your knowledge, did 823 1 Lilly look into any other suicidality scales 2 besides the HAMD or Item 3 of the HAMD to include 3 in their clinical trials? 4 A The MADRS scale, and we 5 already talked about the one that we were trying 6 to develop from Doctor -- I've forgotten his name. 7 Q Ivan Miller. 8 A Miller, okay. You showed 9 me one thing that had a twenty-one question scale, 10 and then you showed me something that had a 11 modified scale. I'm not sure I recall that in 12 detail. 13 Q It's our understanding 14 that Ivan Miller has conducted a clinical trial to 15 try to validate the modified scale for suicidal 16 ideation revised that he developed with Lilly to 17 include in clinical trials. 18 A Super. 19 Q Okay. So far you haven't 20 used it. Do you know if it's been validated? 21 A You said that he 22 conducted the trial. That's -- one, that's news 23 to me. Is it finished and validated? 24 Q I don't think it's 824 1 validated. It's finished, and Catherine Messner 2 was the CRA on the study. And you don't have any 3 information regarding that study? 4 A No. 5 Q You have no knowledge of 6 that study whatsoever? 7 A No, I think I told you 8 that we were working to try to validate it, but it 9 was news to me that you just told me that that 10 study had been finished. I'm delighted. 11 Q Do you recall the study 12 being done? 13 A I didn't remember Doctor 14 Miller. I remember that we were working on trying 15 to validate a new scale with more questions on it. 16 Q Is there a time limit 17 that Lilly -- strike that. Was there a time limit 18 that the FDA gave Lilly to conduct the rechallenge 19 study within? 20 A Well, they never -- they 21 got around to telling us we had to do the 22 rechallenge study, so that one, there was never 23 agreement that in fact we would do it, and there 24 certainly wasn't any discussion of the time limit, 825 1 no. 2 Q If Doctor Miller's scale 3 is in fact validated, would Lilly still consider 4 doing the rechallenge study? 5 A I can't speak for Lilly; 6 I would vote against it. 7 Q Why? 8 A Because it's unnecessary 9 with the wealth of data that's now available. 10 Q A wealth of retrospective 11 data, correct? 12 A And prospective data; 13 remember, I said that we probably had thirty 14 thousand people in controlled trials of Prozac 15 around the world, and I don't know how many people 16 have been on the drug now, ten million, eleven 17 million. 18 Q But none of those studies 19 were conducted with the specific objective to 20 study the incidence of suicidality in patients 21 taking fluoxetine, correct? 22 A I tell you, I don't think 23 that's necessary at this -- 24 Q That's not my question; 826 1 yes or no, none of those studies were specifically 2 conducted with the objective -- 3 A We've gone through that 4 so many times, let me just say no. 5 Q Okay. The Marie Ashberg 6 Scale or Asberg scale you talked about is the 7 MADRS scale, correct? 8 A M-A-D-R-A-S. It's the 9 Montgomery Asberg scale; Marie Asberg is one of 10 the authors. 11 Q I just wanted to make 12 sure it wasn't another scale that we weren't aware 13 of. 14 A No, I think that's it. 15 Q Why didn't you consider 16 using Doctor Miller's initial scale, the Modified 17 Scale for Suicidal Ideation, or Doctor Beck's 18 scale or the Adult Suicide Ideation Questionnaire 19 in the clinical trials as opposed to developing a 20 brand new scale? 21 A After the issue evolved? 22 Q Sure. 23 A Because we were talking 24 to a whole bunch of experts about how to do the 827 1 study, both within the FDA and our own experts 2 internally, plus our worldwide suicidality 3 experts, and their advice was to develop a new 4 scale because they didn't think anything that was 5 currently available would add to the HAMD and the 6 MADRS scale and the adverse event collections that 7 we had done. So they said if you're going to do 8 anything, develop a new better scale, and that's 9 why we tried to do that. 10 Q And you were looking 11 specifically for the Teicher phenomena of 12 suicidality, correct? 13 A Exact -- well, primarily, 14 yes. 15 Q You weren't looking 16 generally for suicidality in any form, or suicide 17 attempts in any form? 18 A We had already been 19 collecting data on ideation, acts and suicides. I 20 mean, you know, if you wanted to do something new 21 and novel, you would have to add something to have 22 a more finely granular description of what it is 23 that you were looking at, because anything that is 24 called suicidality would have been captured by 828 1 what we had already done. The question is do you 2 have four buckets to put it in or do you have ten 3 buckets to put it in. 4 Q Is it your testimony that 5 the FDA was satisfied with the use of only the 6 HAMD or the MADRS score items of suicidality to 7 look at suicidal ideation or suicide attempts in 8 fluoxetine clinical trials? 9 A Well, they certainly were 10 satisfied before the Teicher article was published 11 that we reported all of that to them, and they're 12 certainly satisfied now. In between we were all 13 struggling for how can we get more information, 14 and none of us understood the Teicher phenomenon, 15 and we didn't think it occurred often enough we 16 could easily study it. So there was a period in 17 there where both FDA and Lilly were struggling to 18 find a better way of studying that specific 19 question. 20 Q Are you aware that the 21 FDA has requested Lilly to include a suicidal 22 ideation questionnaire or another suicide scale 23 specifically in the clinical trials on top of the 24 HAMD? 829 1 A Other than in the 2 documents that you've shown me already in the last 3 three days, that they had made a request? 4 Q Yes. 5 A Because we have already 6 gone through that we had discussions and I said on 7 one occasion we will do a prospective study and so 8 forth. No, I'm not aware that we have a letter 9 that says you will in fact add this scale to your 10 studies. 11 Q Not just a letter, I mean 12 any request whatsoever, any suggestion by the FDA, 13 are you aware of? 14 A Sure, there were -- we've 15 already talked about a number of conversations in 16 which I and other people at Lilly agreed with 17 Doctor Leber and other people at FDA that we ought 18 to try to get more information in this area, and 19 we said in good faith that we would try to do 20 that. Now, if you want to characterize that as a 21 request, that's fine. 22 Q Okay. 23 A But when FDA requests 24 something of me, they send me a letter and say you 830 1 better do this right away. 2 Q I'm talking about the 3 HAMD or the MADRS scale. 4 A Yes; no, no, that's what 5 we were working on, because we already had a ton 6 of data using those, and all the plans for ongoing 7 trials were to continue using those. 8 Q Would it have been 9 difficult to include, say, the Adult Suicidal 10 Ideation Questionnaire on the clinical trials 11 before Doctor Teicher's article came out? 12 A Well, it would have only 13 added to the database, again, since, one, it 14 wasn't an issue; two, we are collecting the data 15 already that were relevant to that; and three, we 16 didn't think that was a fully validated scale. 17 The issue never came up, so -- 18 Q Are you aware that Lilly 19 is using that scale in clinical trials now? 20 A The Miller scale or the 21 modified Miller scale? 22 Q No, the Adult Suicidal 23 Ideation Questionnaire. 24 A No, I wasn't. 831 1 Q I believe it was Doctor 2 Beasley who testified that they were using it now. 3 A Good. 4 Q Partly because the FDA 5 had voiced a desire to have that information on 6 top of the HAMD. 7 A Super. 8 Q Why wasn't a scale like 9 that included in the clinical trials after the 10 issue was raised by the German government back in 11 1984? 12 A I'm not sure. One, I'm 13 not sure what trials were beginning at that time, 14 and two, since they were the only government 15 around the world that raised the issue, and 16 because the data, the way we analyzed it at least, 17 didn't suggest we had a problem, I didn't think 18 the questions from the BGA -- and I'm only 19 speaking for myself -- in fact were of great 20 concern. 21 Q It's your testimony that 22 the only government around the world, before the 23 Teicher article came out, that raised the issue of 24 suicidal ideation and the use of Prozac was 832 1 Germany? 2 A The only one that I 3 remember. 4 Q How about Sweden? 5 A You know, I was never a 6 part of the whole Swedish discussion. I know -- 7 don't think they have yet approved the drug, but I 8 wasn't really in that loop. I know they didn't 9 approve it, but I thought the issue was related to 10 dose. 11 Q There were issues that 12 were raised regarding suicidal ideation also. 13 MR. FREEMAN: Don't take 14 that as fact; she's making statements and trying 15 to testify. We'll get somebody from Sweden to 16 testify to that, if you want to. 17 MR. SMITH: I was 18 thinking about Ingrid Bergman or something like 19 that. 20 THE WITNESS: If you get 21 her, get Elvis too. 22 MR. SMITH: Did she die, 23 too? 24 Q (BY MS. ZETTLER) The 833 1 other day you said something that I have a 2 question about. You said that there was an 3 epidemiology study conducted by Lilly on the issue 4 of suicidality and violent aggressive behavior? 5 A Violent aggressive -- 6 well, it was an epidemiology study in the sense 7 that we searched all the literature that we could 8 find and -- 9 Q And also something about 10 employees calling reports in? 11 A Employees calling -- 12 Q Adverse event reports in? 13 A I can't recall that. The 14 rules under the DEN system that I put in place, I 15 think it was '83 or maybe '84 -- no, it was '83 we 16 began it, was that any Lilly employee anywhere in 17 the world that learns of any adverse event through 18 any source, and I've done it from cocktail parties 19 and overhearing conversations, should in fact 20 report by telecommunications to DEN within 21 twenty-four or seventy-two hours, depending on 22 where they are in the world. So, yes, we got a 23 lot of adverse events, including ones I report 24 because I've literally called up DEN and said, you 834 1 know, that I overheard a conversation in the 2 hallway just now that somebody's wife was taking 3 Ceclor and had an allergic reaction. 4 Q Do you usually stop and 5 ask for more information about that adverse event 6 or do you just report what you overheard in a 7 conversation in the hallway? 8 A Depending on who the 9 individual is, but, for example, you remember you 10 showed me the document about that meeting of the 11 expert opinion leaders that I had gone to, 12 wherever it was, I had another Lilly employee come 13 to that meeting and follow me around so that that 14 person could in fact report all the adverse events 15 that I was hearing, so that I wouldn't have to do 16 it myself. 17 Q Have there been times 18 where you reported things that you've overheard 19 without getting other information about it? 20 A Absolutely. 21 Q And you think that's a 22 scientific method of gathering data? 23 A No, but I think it 24 fulfills a regulatory requirement. 835 1 Q Have you ever heard of 2 the term noise used in statistics or in 3 information? 4 A Well, in a lot of my 5 publications, yes. But the regulations, if you 6 right now told me that you had heard of a patient 7 who had taken one of our drugs and who had their 8 hair turn blue, I will guarantee you that I would 9 call DEN up within two days and that would be in 10 the DEN database. And I may or may not ask you 11 for more information, but I would identify you. 12 MR. SMITH: What if their 13 hair turned loose? 14 A You're the lawyer, but if 15 you read the law and the regulations, I think 16 that's exactly what it requires. The NDA, for 17 example, requires all information. 18 Q Including suicide 19 analyses that are done to answer questions for 20 other regulatory agencies outside the United 21 States? 22 A There's a big debate on 23 that, it depends on which NDA regulation you want 24 to talk about because they changed, if you recall. 836 1 Q Do you feel that you have 2 an obligation to report the fact that the German 3 government raised the issue and that it was fully 4 analyzed back in 1984 or '85? 5 A I would probably call on 6 my legal colleagues to help me interpret it, but 7 under the current NDA regulation, I think the 8 current NDA regulation does request information on 9 negative findings by other regulators. But what 10 that means has never been tested in court in terms 11 of whether that is just a letter that you receive 12 from the BGA saying we're not going to approve 13 your drug, or whether it means everything that you 14 talk about in working up whatever their questions 15 are. I'm not aware that that's ever been 16 resolved. 17 Q So let me make sure I 18 understand. You feel it is your absolute duty to 19 report a conversation that you have overheard in 20 the hallway, but you don't necessarily feel it's 21 your duty to report an issue raised by another 22 regulatory agency and the analyses that were done 23 by the drug company in response to that issue 24 being raised? 837 1 A Under the FDA laws and 2 regulations, I think you've stated it correctly. 3 Q You said earlier that 4 there was an argument with the FDA on whether or 5 not to use 1639's to report clinical trial adverse 6 events; do you recall that? 7 A Yes, absolutely. 8 Q Tell me about that 9 argument. 10 A Well, Bob Temple 11 specifically did not want us to use the 1639 form 12 in reporting clinical trial adverse events. And I 13 wanted to use the 1639 form because what I was 14 trying to do was to move us towards a single 15 system for handling all adverse events across the 16 whole world. And so when we were beginning to 17 develop DEN in 1983, it was a matter of some 18 importance to us that in fact we use that same 19 two-page form for clinical trial or spontaneous 20 events, and I went down and had a big argument 21 with him in which he finally -- in fact, I don't 22 think he relented at the meeting, but I think we 23 finally resolved that the FDA would allow us to do 24 that. 838 1 At that time, they didn't 2 have any specifications about how to report, so 3 that my argument was that we had not -- we could 4 fulfill their obligations by using that form, even 5 though Bob Temple told me he didn't want us to use 6 the form. 7 Q So it's your testimony 8 that every single adverse event that occurs with 9 fluoxetine, regardless of whether or not it's 10 spontaneous or happens in a clinical trial, is 11 reported to the FDA on the 1639? 12 A No, that's not what I 13 said. 14 Q Okay, clarify it for me. 15 A Okay, everything that we 16 have to report to the FDA as an individual alert 17 event, whether it's from a trial or from 18 spontaneous -- and the definitions, as you know, 19 are different with those two. The mechanism of 20 reporting to the FDA is that modified 1639 21 two-page form, which may or may not have stuff 22 appended to it. I mean, that's the primary 23 mechanism. In addition to that, of course, we 24 also have a whole variety of other periodic 839 1 reports, et cetera, that are also sent to them. 2 So, it's not the only way every event gets there. 3 Q So which adverse events 4 were you fighting to have reported from the 5 clinical trials on a 1639, the serious fifteen-day 6 reports, things like that? 7 A Ten day under the IND 8 rules. Exactly, the ones that have to be reported 9 individually in an expedited fashion rather than 10 being summarized in the next annual report to the 11 IND. 12 Q Regulations require that 13 the sponsor notify the FDA of any reports within a 14 certain period of time, say ten days anyway, 15 correct? 16 A Under the IND regulations 17 currently, rather the NDA regulations, it's three 18 working days for a telephone report for -- I think 19 it's an unexpected death, and ten working days for 20 mailing the written report after Lilly learns 21 about it, although that again is the subject of -- 22 when the clock starts is another debate. 23 Q So you really couldn't 24 wait for the quarterly reports to notify the FDA 840 1 of those reports, could you? 2 A No, no, there are two 3 obligations. One is for what is usually called 4 expedited reporting, one by one. Then there's 5 another obligation which is to report stuff under 6 the annual report to the IND. 7 Q Okay, so which ones did 8 you want to use the 1639's with? 9 A The expedited or 10 individual reports. 11 Q And what was the FDA 12 asking you to do to report those prior to being 13 able to use the 1639's? 14 A Well, there were no 15 regulations, and I don't think to this day there 16 are any regulations about the format to be used in 17 the reports. The argument was that specifically I 18 wanted to use the Lilly modified 1639 and Bob 19 Temple didn't want it. 20 Q Didn't want the Lilly 21 modified or the 1639, period? 22 A Didn't want the 1639, 23 period. 24 Q What form did he want? 841 1 A Type out a description. 2 Q Why didn't he want to go 3 with the 1639? 4 A Two reasons: One, he 5 thought it would be confusing, because then his 6 staff -- and remember he was -- although he did 7 look at postmarketing events, they were primarily 8 looking at the clinical trial events, he didn't 9 want his staff to be confused because they would 10 see something that looked like a 1639 and they 11 might not look down to see the origin as to where 12 it came from. And the other thing is that he said 13 that we should be able to have the freedom to be 14 creative in how we would report those events. 15 Q That you should have the 16 freedom to be creative? 17 A Yes. 18 Q Interesting. You also 19 said yesterday that whether or not -- and I'm not 20 sure if you said to use the 1639's or to actually 21 report the adverse event with the -- under the IND 22 regulations, okay? And I think Paul was asking 23 you about some specific adverse events. Depended 24 on the characteristic of the adverse event? 842 1 A (WITNESS MOVES HEAD UP 2 AND DOWN.) 3 Q Now, are you talking 4 again about alert reporting as opposed to 5 quarterly reports? 6 A Yes. 7 Q Or are you talking about 8 there are some adverse events that are not 9 required to be reported at all? 10 A I think that -- I think 11 they're all reported eventually in the annual 12 reports, the IND. But there are explicit rules 13 about both the three-day and the ten-day report, 14 and it depends upon whether they're serious, 15 whether they are expected, and under the US FDA 16 IND regulations, it also relates to causality. 17 Q So a causal relationship 18 has to be determined? 19 A That's right. 20 Q Before it's reported on 21 an alert basis or before it's reported at all? 22 A An alert basis. 23 Q So you could have a 24 serious unexpected adverse event, but unless you 843 1 determined that it's causally related -- 2 reasonably causally related to the drug, you don't 3 have to report it on an alert basis? 4 A Under the IND 5 regulations, that's my understanding of the law 6 and regulations. But we've -- I mean, I haven't 7 looked at this for a number of years, and I have 8 written, however, the standard operating 9 procedures in the past, we'll get it out and see 10 what I had wrote. 11 Q Are you aware that 12 adverse events were collected by Lilly employees 13 during telemarketing sessions with doctors? 14 A On Prozac? 15 Q Yes. 16 A I wasn't aware we did 17 telemarketing sessions on -- well, maybe I am 18 aware of that, but I don't know what happened to 19 any of the reports of adverse events. 20 Q Would that qualify under 21 your interpretation of the 1639 reports for 22 spontaneous adverse events? 23 A No. 24 Q Why not? 844 1 A Because it's collected in 2 a systematic fashion, and in fact Upjohn had 3 negotiated that with the FDA, not so much for 4 telemarketing conferences, but for their 5 postmarketing surveillance studies. See, there's 6 a whole third bucket, unfortunately, which is what 7 happens when you have a marketed drug, but in fact 8 the origin of the event is not spontaneous in the 9 sense that somebody just calls you up, but in fact 10 you elicit the adverse event in some controlled 11 fashion. There's a whole new set of rules for 12 those. 13 Q Weren't Lilly employees 14 eliciting adverse events from physicians when they 15 were detailing to them? 16 A Yes. 17 Q What's the difference of 18 eliciting it one on one as opposed to eliciting it 19 in a group? 20 A Whether or not you have a 21 protocol, a systematic fashion of doing it. 22 Q And the detail people had 23 a protocol or did not? 24 A I don't think so. Not 845 1 that I'm aware of. 2 Q So you're saying because 3 you had -- you're saying in the tele-sessions, the 4 marketing sessions you had a protocol and a 5 systematic way to do it? 6 A I think so. That's what 7 was discussed when we first talked about it. 8 Q So you're saying that in 9 the formal setting or the semiformal setting of 10 the telemarketing session, where you would have a 11 group of doctors and you asked them questions 12 about their use of the drug and what their 13 experiences have been and, you know, whether they 14 like it, they had problems with using it, et 15 cetera, if an adverse event is reported through 16 being asked about it, then it's not spontaneous 17 and therefore it does not need to be reported in 18 the 1639 to the FDA? 19 A That's exactly what was 20 the issue between FDA and Upjohn, because they 21 didn't want overreporting of non-stimulated 22 events, and we followed exactly what that 23 agreement had been, I think. I mean, again, I 24 wasn't that close to that particular thing, but I 846 1 was in the discussions about how to handle those 2 events. It's like literature reports, that's 3 another little bucket. 4 Q Well, in the spontaneous 5 reporting system there is an indication on the 6 line listing that you get from the FDA as to 7 whether or not it was a literature report or not, 8 correct? 9 A That's correct. 10 Q I guess I fail to see 11 what the difference is between having a group of 12 doctors that are participating in a tele-session 13 that is conducted by marketing people and the 14 doctor who is having a one-on-one conversation 15 with the detail person where that detail person is 16 eliciting a spontaneous report on an adverse 17 event. 18 A Let me do the best I can 19 with that. I think that in fact the salespeople 20 do not have an obligation to go and ask the 21 physician if they have seen an adverse event, but 22 that if the physician spontaneously tells them 23 what do you know about nausea and Prozac, what I 24 have instructed to at least some of the 847 1 salespeople is, having been asked that question, 2 they should then ask the physician are you just 3 asking me in general or in fact do you have a 4 specific patient in mind, and if the answer to 5 that is, well, Doctor Jones told me that he or she 6 had a patient who had nausea and Prozac, that 7 should then trigger a spontaneous adverse event. 8 Now, if in fact the 9 salespeople are going out and just spontaneously 10 asking doctors have you ever seen an adverse event 11 on Prozac, I'm not aware of that, and I don't 12 believe that's a regulatory obligation. Now, what 13 you described, however, was a systematic search 14 for adverse events where, if you're a physician, I 15 call you up and say, you know, hello, I'm Doctor 16 Thompson at Lilly, have you ever seen an adverse 17 event with Prozac. That, you see, I think falls 18 into a different category, because rather than 19 being unstimulated, it's stimulated, and it would 20 therefore be treated quite differently. And you'd 21 have real problems scientifically if you mix those 22 two, because the incidence of adverse event 23 reporting would be expected to be quite different 24 under those two circumstances. 848 1 Q Wasn't a big problem with 2 adverse event reporting that it's under reported, 3 Doctor? 4 A Well, but you don't want 5 to mix apples and oranges. That's exactly the 6 point I was making. It's like which 7 classification term do you put the events in. 8 You're going to be highly consistent. If, for 9 example, we went out -- let me give you a very 10 good example, because we think some companies have 11 done this, that they send their people out to go 12 and elicit adverse events to our drugs which they 13 then report. 14 Q To your drugs? 15 A Yes, and therefore, if 16 you go and start counting numbers of adverse 17 events, you all of a sudden have apples and 18 oranges. What you're trying to do is to have a 19 consistent thing, and that's why we try to 20 distinguish between truly unprobed spontaneous, 21 and as you know the frequency of reporting from 22 some countries is very different from others, so 23 we try to do this country by country or region by 24 region, versus probed in the sense of a clinical 849 1 trial, whether or not it's a pre or a 2 postmarketing trial. 3 Q Doctor, I'm not talking 4 about another company going out and trying to in 5 effect sabotage your drug by eliciting adverse 6 events about your drugs, okay? I'm talking about 7 the company who is manufacturing the drug, the 8 company that has created and tested and marketed 9 the drug, finding out about adverse events through 10 whatever means possible. You went to great 11 lengths a few minutes ago to tell us that it is 12 your belief and your understanding that Lilly 13 employees should report adverse events that they 14 hear from any source, including walking by and 15 hearing a conversation in the hallway -- 16 A Yes. 17 Q -- regardless of whether 18 or not you have the opportunity to get any 19 information about that other than what you 20 overheard, correct? 21 A Well, I always report the 22 best I knew of who the originator was. 23 Q But if you walk down the 24 hall and didn't know somebody and heard them 850 1 talking about, you know, gee, Joe, my wife had an 2 adverse event of some sort on Prozac, you would 3 report that even though you didn't feel 4 comfortable going up and approaching those people 5 and saying, hey, I'm Leigh Thompson from Lilly, 6 can you give me some more information about that; 7 you would go back to the office and have a 1639 8 filled out on that adverse event, even though you 9 had no clue who any of those people were, correct? 10 A That's correct. 11 Q But you're also telling 12 us on the other hand that you either cannot or did 13 not report adverse events that you were told about 14 by physicians during telemarketing sessions where 15 those physicians who are sitting right there and 16 willing to answer any questions that you had about 17 their drug, that you would not report those to the 18 FDA? 19 A That's not exactly what 20 my testimony was. First of all, I don't know how 21 that study was done. In fact, I didn't even know 22 that the study had been done, to be honest with 23 you. But I know that in discussions about how to 24 handle this in general, and it wasn't specific to 851 1 Prozac, we went to Upjohn and asked them about how 2 they had done it and what their conversations were 3 with the FDA, and other people, not me, had 4 carried out that conversation with the FDA to 5 decide what our regulatory reporting requirement 6 was. But I will absolutely defend not just 7 dumping it into the spontaneous bucket, which is 8 what you suggested. You asked me if that was the 9 right thing to do, and I'll affirm that you should 10 not dump those kinds of things into the 11 spontaneous buckets. 12 MR. FREEMAN: It's after 13 5:00 o'clock. 14 MS. ZETTLER: Just a 15 couple more questions. 16 Q (BY MS. ZETTLER) If in 17 fact the detail men or detail people were sent to 18 the doctors after the Teicher article came out and 19 asked to elicit any adverse events that doctors 20 had experienced or had -- their patients had 21 experienced on Prozac, and they were then reported 22 to the FDA through 1639's as spontaneous reports, 23 what would be the difference between doing that 24 and the telemarketing? 852 1 A That's so hypothetical 2 that I'm not -- I mean, I see what your point is, 3 and you're getting down to increasingly fine 4 distinctions, but in all honesty, I don't -- I'm 5 not sure, one, that that occurred, and two, I 6 think you should define a different rule for 7 reporting. Not that all those things shouldn't be 8 reported, but that you shouldn't just dump them in 9 the bucket, because what you're trying to do under 10 increased frequency reports is to get the most 11 reliable estimate that you can against the 12 background of spontaneous reporting of whether 13 there's been a change or not. And if you elicit 14 adverse events and therefore increase the 15 numerator for whatever the denominator is, that 16 gives you false information. 17 If you do that, you ought 18 to at least be able to describe the numerator and 19 denominator of the population you're looking at. 20 You ought to say we went out and asked twenty 21 thousand doctors what they had in the way of 22 adverse events, and here are the things that we 23 heard. So, the numerator is this, the denominator 24 is that. That's different than saying, hey, the 853 1 spontaneous event pool comes from ten million 2 people that we think have taken the drug in the 3 world. 4 Q Are you aware that the 5 FDA itself feels that adverse events with drugs 6 that are grossly under reported, that in fact only 7 approximately ten percent of adverse events that 8 actually occur with the drug are reported 9 generally? 10 MR. FREEMAN: On Prozac? 11 MS. ZETTLER: Generally. 12 A No, the leaders of the 13 FDA, I believe, because we've been on many panels 14 together, think that it varies depending upon the 15 severity of the event and other things like 16 Seminole papers. In fact, one of the big problems 17 is when somebody publishes a case report and says 18 hair falling out is seen with Elavil, then what 19 you get is a whole flurry of reports of hair 20 falling out with Elavil and you have to be very 21 careful with that single event. So, it varies, 22 but I think that the ten percent is low for 23 serious adverse event reporting. I'm not aware of 24 anybody in the upper management of FDA who would 854 1 say that for the serious ones, the really serious 2 ones. But for the general adverse events, I 3 think it's a lot lower than ten percent, because 4 if you define adverse events the way I've defined 5 it here carefully, virtually everybody should in 6 fact have at least one adverse event on a drug, 7 and the way our clinical trial is run, and if you 8 look at the cancer trials, we probably report ten 9 or twenty adverse events for every patient in the 10 trial. 11 Q You didn't do that for 12 Prozac, though? 13 A I'm not sure what the 14 actual ratio is, but across general trials -- it 15 depends obviously on the length and the underlying 16 illness, but in general, if you look at a trial 17 that lasts a number of months in the US or in 18 Scandinavia or Canada, I would expect to see, if 19 not one to one, somewhere approaching that. I 20 mean, one event per person. But Bob Zerbe has 21 published a very nice article looking at regional 22 differences in adverse event reporting on -- I 23 think it was Lorobid, one of our drugs. 24 Q Well, my question was 855 1 generally with the FDA, they feel that the adverse 2 events are generally under-reported. 3 A I mean, I don't have 4 any -- I agree with you on that, in spontaneous 5 reporting. 6 MS. ZETTLER: In 7 spontaneous reporting. Thank you, Doctor. 8 (WITNESS EXCUSED.) 9 * * * * * 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 856 1 STATE OF KENTUCKY ) ) SS. 2 COUNTY OF JEFFERSON ) 3 I, MARY KATHLEEN NOLD, a Notary Public within 4 and for the State at Large aforesaid, do hereby 5 certify that the foregoing is a true, correct and 6 complete transcript of the deposition of W. LEIGH 7 THOMPSON, MD, taken at the time and place and for 8 the purpose set out in the caption hereof; that 9 the witness was duly sworn before giving said 10 deposition; that the said deposition was taken 11 down by me in stenotype and afterwards transcribed 12 on a computer under my direction; that the 13 appearances were as set out in the caption hereof; 14 and that a request was made by counsel that the 15 deposition be submitted to the witness for reading 16 and signature. 17 GIVEN my hand as notary aforesaid, this 18 __________ day of ___________________, 1994. 19 My commission expires March 10, 1998. 20 21 22 _______________________________________ MARY KATHLEEN NOLD 23 COURT REPORTER AND NOTARY PUBLIC STATE OF KENTUCKY AT LARGE 24 857 1 STATE OF___________________) ) SS: 2 COUNTY OF__________________) 3 I, W. LEIGH THOMPSON, MD, do hereby certify 4 that I have read the foregoing deposition given by 5 me on July 20, 21 and 22, 1994, and that the 6 answers contained therein are true and correct to 7 the best of my knowledge and belief. 8 9 10 ______________________________ LEIGH THOMPSON, MD 11 12 13 ______________________________ (DATE) 14 15 16 Subscribed and sworn to before me this day by 17 LEIGH THOMPSON, MD. 18 My commission expires______________________. 19 20 ___________________________________ 21 NOTARY PUBLIC 22 23 ______________________________ 24 858 1 STATE OF____________________) ) SS: 2 COUNTY OF___________________) 3 I, W. LEIGH THOMPSON, MD, do here certify that 4 I have read the foregoing deposition given by me 5 on July 20, 21 and 22, 1994, and that the answers 6 contained therein are true and correct to the best 7 of my knowledge and belief, with the following 8 corrections: 9 PAGE/LINE CORRECTION REASON FOR CORRECTION 10 __________________________________________________ 11 __________________________________________________ 12 __________________________________________________ 13 __________________________________________________ 14 __________________________________________________ 15 _________________________________ W. LEIGH THOMPSON, MD 16 _______________________________ (DATE) 17 18 19 Subscribed and sworn to before me this day by 20 W. LEIGH THOMPSON, MD. 21 My commission expires_____________________. 22 ________________________________ NOTARY PUBLIC 23 24 ________________________________ 859