February 13, 1998
STATE OF MINNESOTA
DISTRICT COURT COUNTY OF RAMSEY
SECOND JUDICIAL DISTRICT
File No. C1-94-8565
The State of Minnesota, by Hubert H. Humphrey, III, its attorney general, and Blue Cross and Blue Shield of Minnesota,
Plaintiffs,
vs.
Philip Morris Incorporated, R.J. Reynolds Tobacco Company, Brown & Williamson Tobacco Corporation, B.A.T. Industries P.L.C., Lorillard Tobacco Company, The American Tobacco Company, Liggett Group, Inc., The Council for Tobacco Research-U.S.A., Inc., and The Tobacco Institute, Inc.,
Defendants.
---------------------------------------------------------
THE CLERK: All rise. Ramsey County District Court is again in session, the Honorable Kenneth J. Fitzpatrick now presiding.
(Jury enters the courtroom.)
THE CLERK: Please be seated.
THE COURT: Good morning.
(Collective "Good morning.")
MR. HAMLIN: Good morning.
JONATHAN M. SAMET called as a witness, being previously sworn, was examined and testified as follows:
BY MR. HAMLIN:
Q. Good morning, Dr. Samet.
A. Good morning.
Q. Dr. Samet, at the close of your testimony yesterday afternoon, we were talking about a lot of data and studies. What I'd like you to do now is summarize the conclusions that you have regarding the disease risks of lower tar and lower nicotine cigarettes.
First of all, what has happened to the tar content in cigarettes over the past 40 years? And in giving me that answer, I'd like you to refer to Trial Exhibit 30146, which is on the easel.
A. Referring to the trial exhibit on the easel, "The Changing Cigarette," I'd just like to say again that, as we look at that, we see the rise in use of filter cigarettes beginning in the early 1950s up to today where all but a few percent of smokers in the United States smoke filter cigarettes. The green line, the tar content, the tar yield, begins at about 35 milligrams or so in 1955, and if you notice over time to the end of the line at about 1990, that average of the tar in the cigarettes smoked has dropped to about 12 or so; in other words, it's dropped by over 20 milligrams in tar, so almost a two-thirds reduction in the FTC-measured tar yield, the machine-measured tar yield.
The red line shows the nicotine drop, which is going from a bit above two milligrams down to around one milligram.
Q. And what does the blue line show?
A. The blue line is the percentage of all U.S. cigarettes with filter tips.
Q. And what does that indicate in terms of the trend?
A. Again, there's a very steep trend until almost all cigarettes sold have filters.
Q. Now what has happened to the disease risks of smoking over the past 40 years?
A. Well we've seen, certainly, no drop in the disease risks associated with smoking, and evidence of increasing risks for some diseases.
Q. What do we see when we measure the disease risk in specific studies, doctor?
A. We saw some --
I talked about some information yesterday, for example the CPS-I study, that was the 1959-1960 to 1972 study that showed some drop, about a 20 percent drop in risk of dying from lung cancer for those who were smoking the lower tar cigarettes of the time '59 to '72, compared to the higher tar cigarettes of the time. So just remember that the higher tar cigarettes in that study had -- were in the range of 25 and higher in terms of the milligrams of tar, the lower tar were about -- under 17.6 milligrams of tar, if I remember correctly, so with that degree of reduction, there was about a 20 percent drop in the lung cancer risk.
For some other -- for some other consequences of smoking, we've also seen some reduction of risk going from higher tar to lower tar cigarettes. Respiratory symptoms, for example, in some studies, including a study of my own, have shown some drop comparing lower tar smokers to higher tar smokers. The paper I mentioned yesterday by Drs. Shenker, Speizer and myself, for example, smokers of lower tar cigarettes were less likely to report cough, sputum, phlegm we talked about, wheeze, and shortness of breath. And that type of findings have come up in other studies.
For coronary heart disease and for chronic obstructive pulmonary disease in individual studies, in general we've seen no evidence of reduced risk for smokers of the higher tar compared to lower tar cigarettes.
Q. Now were these specific studies studies of risk at a point in time?
A. Yes. These were studies done at one interval in time across that graph of -- the cigarette types that's in front of us now.
Q. But what do we see when we measure the risks of smoking over the past 40 years?
A. Well there we have some information that I discussed yesterday. We have the two studies of the American Cancer Society, the two studies of one million Americans, the earlier study across the '60s, '59 to '72, the later study beginning in the '80s. And what I showed yesterday was that in that study, the disease risk, lung cancer and the other major smoking-caused diseases, had actually gone up comparing the two studies, one looking at the risks in persons who were enrolled towards the left side of that exhibit and the other more towards the right side of that exhibit.
And also, in terms of the results of the individual epidemiological studies in the animations as I showed the findings of those studies over time, there was no evidence that risks had begun to drop off in the later studies.
Q. Doctor, if lower tar and lower nicotine cigarettes were reducing disease risks of smoking over the past 40 years, what would you have expected to see in that data?
A. Well over 40 years I would have expected to see some evidence of a drop in those risks. We have simply not seen it. We can see on the exhibit that tar and nicotine values have dropped substantially, but yet disease risks show no evidence of dropping. And if anything, we have evidence that for some diseases they have gone up over time.
Q. Doctor, have you reviewed literature regarding the number of new young smokers that start each day?
A. Yes, I have.
Q. And could you direct your attention to Trial Exhibit 26008. And it's in your testimony notebook.
Could you identify that, doctor?
A. Yes. This is a publication in the Journal of the American Medical Association, 19 -- January 6th, 1989, the title is "Trends in Cigarette Smoking in the United States, Projections to the Year 2,000," by John Pierce, Michael Fiore, Thomas Novotny, Evridiki Hatziandreu and Ronald Davis.
Q. Have you reviewed this document in connection with your investigation in this case?
A. Yes, I have.
Q. And have you relied, in part, on this document as a basis for your opinions in this case?
A. Yes.
Q. Doctor, do you consider this document to be a reliable authority in the published scientific literature?
A. Yes.
MR. HAMLIN: Your Honor, we offer Trial Exhibit 26008.
MR. GARNICK: No objection.
THE COURT: Court will receive 26008.
BY MR. HAMLIN:
Q. Doctor, according to this article, how many new smokers, young smokers, begin smoking each day?
A. This article provides an estimate that there are about 3,000 new smokers each day.
Q. And is the Journal of the American Medical Association a respected journal?
A. Yes, it is.
Q. Now in the course of your investigation in this case, have you reviewed literature regarding annual mortality from smoking?
A. Yes, I have.
Q. And could you direct your attention to Trial Exhibit 26006. Could you identify that, doctor?
A. Yes. This is an article published in the Morbidity and Mortality Weekly Reports published by the Centers for Disease Control. It's 1993, Volume 42, pages 645 to '9, it's an article entitled "Cigarette Smoking-Attributable Mortality and Years of Potential Life Lost -- United States, 1990."
Q. What is the MMWR?
A. The MMWR is a publication of the Centers for Disease Control that provides regular information on the occurrence of diseases, both infectious diseases and other diseases, as well as, for example, the effects of cigarette smoking in the United States population. It provides weekly and updated information on health issues.
Q. Can you explain briefly what the Centers for Disease Control is?
A. Well the Centers for Disease Control is, of course, part of the Department of Health and Human Services. It's charged with exactly what its name says, dealing with disease-control issues. You probably are most familiar with it in terms of infectious-disease outbreaks, but the Centers for Disease Control is also concerned with cigarette smoking and other causes of disease beyond dramatic infectious-disease outbreak.
Q. Doctor, have you reviewed this article as part of your investigation in this case?
A. Yes, I have.
Q. Does this article form part of the basis of your opinions in this case?
A. Yes.
Q. Is this article a reliable authority in the published scientific literature?
A. Yes.
MR. HAMLIN: Your Honor, we offer Trial Exhibit 26006.
MR. GARNICK: Defendants object on the grounds that it's being admitted for purposes of actual attributable-risk figures. That is beyond the scope of the expert report.
MR. HAMLIN: Your Honor, it hasn't -- it certainly is not beyond the scope of the expert report, and moreover, it is clearly a learned treatise. We've laid the proper foundation.
THE COURT: Court will receive 26006.
BY MR. HAMLIN:
Q. Doctor, what if anything is reported in this article regarding annual deaths from smoking?
A. The article --
MR. GARNICK: Objection.
A. -- states --
THE COURT: Excuse me, please.
MR. GARNICK: Objection. Same grounds as before, outside the scope.
THE COURT: Overruled. You may answer.
A. The article states, "During 1990, 418,690 U.S. deaths (approximately 20 percent of all deaths) were attributed to smoking."
Q. Doctor, could you turn now to Trial Exhibit 26007. Could you identify that article.
A. Yes. This is an article published in the Journal of the American Medical Association, November 10th, 1993, it's entitled "Actual Causes of Death in the United States," it's by J. Michael McGinnis and William Foege.
Q. And when was it published?
A. 1993.
Q. Have you reviewed this article as part of your investigation in this case?
A. Yes.
Q. And does this article form part of the basis of your opinions in this case?
A. Yes.
Q. Is this article a reliable authority in the published scientific literature?
A. Yes.
MR. HAMLIN: Your Honor, we offer Trial Exhibit 26007 as a learned treatise.
MR. GARNICK: Same objection, beyond the scope of the report.
THE COURT: Court will receive 26007.
BY MR. HAMLIN:
Q. Your Honor -- excuse me.
Dr. Samet, can you tell us what the Journal of the American Medical Association reports regarding annual deaths from smoking in this article.
MR. GARNICK: Same objection, Your Honor.
THE COURT: You may answer that.
A. In the article the authors state that the most prominent contributors to mortality in the United States in 1990 were tobacco, an estimated 400,000 deaths, and then the article goes on -- the sentence goes on to list some other causes of death.
Q. Doctor, could you direct your attention now to Trial Exhibit 26012, which is right in front of you in the folder, and could you identify that document.
A. This is a report entitled "State Tobacco Control Highlights-1990," it's from the U.S. Department of Health and Human Services, from the Centers for Disease Control, the Office on Smoking and Health.
Q. And what was the date?
A. Nineteen -- oh, I'm sorry, 1996.
Q. Okay. Have you reviewed this article as part of your investigation in this case?
A. Yes, I have.
Q. And does this article form part of the basis of your opinion in this case?
A. Yes, it does.
Q. Is this article a reliable authority in the published scientific literature?
A. Yes.
MR. HAMLIN: Your Honor, we offer Trial Exhibit 26012.
MR. GARNICK: Objection, beyond the scope.
THE COURT: Court will receive 26012.
BY MR. HAMLIN:
Q. Dr. Samet, could you turn to page 56 of the document.
A. Yes.
Q. Okay. What does the Center -- Centers for Disease Control report regarding the annual deaths from smoking in Minnesota?
A. There's an estimate for 1990 given as 6,127 deaths related to smoking.
MR. HAMLIN: Your Honor, that concludes our direct. We have no further questions.
MR. GARNICK: Your Honor, we'll need a minute or two to set up.
THE COURT: All right. Counsel, I wonder if you could lay that (referring to the easel) down --
MR. HAMLIN: Yes.
THE COURT: -- if it's not being used, so people can see.
Why don't we take a short recess.
THE CLERK: Court stands in recess.
(Recess taken.)
THE CLERK: All rise. Court is again in session.
(Jury enters the courtroom.)
THE CLERK: Please be seated.
THE COURT: Counsel.
MR. GARNICK: Thank you, Your Honor.
BY MR. GARNICK:
Q. Good morning, Dr. Samet.
A. Good morning.
Q. We had introduced ourselves the other day. My name is Murray Garnick, I represent Philip Morris, along with Peter Bleakley.
(Discussion off the record.)
Q. Dr. Samet, you're not now a treating physician; are you?
A. I'm sorry?
Q. You don't currently see patients; do you?
A. I'm licensed in the state of Maryland. At the moment I'm not seeing patients.
Q. Okay. But while you were in New Mexico you were seeing patients; is that correct?
A. New Mexico and prior years, yes.
Q. And that was in 1994?
A. I left New Mexico in 19 -- mid-summer of 1994, correct.
Q. So when was the last time that you saw a patient?
A. At the time I left the University of New Mexico.
Q. Okay. And you are a specialist in pulmonary disease; is that correct?
A. That's correct.
Q. Now when you used to see patients, did you use to take histories?
A. Yes.
Q. And what is a history?
A. A history involves obtaining information about the patient, including their symptoms, why they've come to see the physician. It's an exploration of what has brought the patient to the physician, obtaining information to lead to a diagnosis.
Q. And when you took a history, you asked many, many questions depending on the particular problem involved; is that fair to say?
A. The number of questions would depend on the problem.
Q. And the questions were themselves tailored to the specific situation; is that correct?
A. Some would be and some would not be.
Q. And you would usually ask about the demographics of a person; correct?
A. I would ask the person -- the people who they are, yes.
Q. You'd ask where they came from, place of birth?
A. Typically.
Q. You'd ask where they live?
A. Yes.
Q. You asked how old they were?
A. Yes.
Q. You asked what their occupation was?
A. Yes.
Q. And you would ask their social situation; is that right?
A. To a varying degree, yes.
Q. With respect to their social situation, what kind of questions would you ask?
A. I might ask questions about where they live, the environment in which they live, their housing, their family, who was in their family. The range of questions would really depend on the situation.
Q. And you would ask about their medical history, of course?
A. Yes.
Q. You would ask them about various risk factors for major diseases?
A. Again, depending on the context.
Q. Would you get a medical file for that person?
A. It would depend on whether the person had a medical file at that institution and what the situation was. Sometimes yes, sometimes no.
Q. Sometimes you would go and get the person's prior medical records; is that correct?
A. If such records were available and relevant, certainly.
Q. Well let's suppose that --
You used to treat people for heart disease; correct?
A. I've treated people for heart disease.
Q. Right. Well let's suppose that you were seeing a patient and you suspected possible heart disease and you were taking a history. Okay?
A. Yes.
Q. Okay. I want to write this down.
Well that's pretty bad. Let me try again. I apologize for the writing, but I would be even worse if I used the flip chart.
Okay. So you would ask a person where they lived; correct?
A. Yes.
Q. Where they were born; correct?
A. Most likely.
Q. Okay. Place of birth; correct?
A. Again, most likely.
Q. You would presumably know their sex, but that would be relevant as well; right?
A. Usually relevant, and obvious, as you point out.
Q. Yes. You would ask about their marital status?
A. I likely would, yes.
Q. And why would you ask about their marital status?
A. I would ask about their marital status to understand the family situation, potential support available to this individual.
Q. Would you ask about the family history for heart disease?
A. Yes.
Q. Why would you ask that?
A. Because family history of heart disease is associated with heart disease risk.
Q. Would you assess their weight?
A. That would likely be on the chart when the patient presented.
Q. If it wasn't on the chart, would you want to assess their weight?
A. I would.
Q. And why is that?
A. Again, weight is relevant to heart-disease risk.
Q. Would you ask questions about their level of exercise?
A. I might.
Q. Well why is that?
A. Well, level of exercise is one of the factors associated with heart- disease risk.
Q. Exercise is a risk factor for heart disease; is that correct?
A. Well I think lack of exercise.
Q. Lack of exercise.
Sedentary lifestyle is a risk factor for heart disease; correct?
A. Yes.
Q. Okay. Would you ask about their alcoholic consumption?
A. I likely would ask that of almost every patient that presented.
Q. Why is that?
A. Again, just part of the information that I would routinely obtain about the patient.
Q. Running out of room.
What about the amount of stress in their lives?
A. I, in this circumstance, would probably not directly try and address that.
Q. You wouldn't try to address stress, psycological stress?
A. I don't have any readily-used tools in a clinical encounter to assess stress.
Q. And I assume you would also ask about their smoking, if they were a smoker?
A. I'd ask about their smoking.
Q. What other risk factors for heart disease might you ask about?
A. I would likely ask about hypertension, high blood pressure.
Q. What else?
A. Perhaps another factor I would ask about is history of diabetes.
Q. Anything else?
A. These are, I think, the principal clinically- relevant risk factors.
Q. Okay. Well let's do the same thing for lung cancer.
I assume you would ask where they lived.
A. Well this would be part of, I think, any clinical encounter.
Q. Place of birth?
A. Yes.
Q. You would assess their sex?
A. Yes.
Q. Marital status?
A. Again, just part of the routine information.
Q. Family history?
A. Possibly.
Q. Now would you assess their occupation?
A. I would ask about their occupation.
Q. Why would you ask that?
A. Because they're occupational causes of lung cancer.
Q. Would you ask about possible radon exposure?
A. I don't know a question to ask other than occupation for assessing radon exposure.
Q. Would you be interested in their possible radon exposure?
A. The only group in which I would find that to be clinically relevant are the uranium miners that I treated who had very high levels of radon exposure.
Q. Okay. What about their diet?
A. No, I would not assess diet for lung cancer risk.
Q. Is diet a risk factor for lung cancer?
A. Diet has been examined as a risk factor for lung cancer, but certainly at this point in our understanding of the diet/lung cancer association, it certainly wouldn't be relevant to clinical care.
Q. Is vitamin A deficiency a risk factor?
A. That's not very clear right now. We've just seen, in fact, the beta carotene studies, that giving beta carotene experimentally, a vitamin A precursor is actually perhaps associated with increased lung cancer risk. So the vitamin A story is not very clear at the moment.
Q. What about saturated fat?
A. I would not assess that in a clinical encounter.
Q. Is saturated fat a risk factor for lung cancer?
A. There are some epidemiological studies that have suggested that fat intake is associated with lung cancer risk, but again, that information is not at a stage of clinical relevance.
Q. And why is that?
A. Again, there are a few studies, some associations, that would simply not provide information relevant to the clinician.
Q. Asbestos, possible asbestos exposure?
A. I mentioned that I would assess occupation.
Q. That would be part of occupation?
A. Yes.
Q. What about depression?
A. For lung cancer risk, no, I would not assess depression, except as a consequence of having lung cancer.
Q. Now there are some studies that suggest that depression is a risk factor for lung cancer; is that correct?
A. I have seen at least one paper suggesting that depression is associated with lung cancer.
Q. Okay. Could you turn to tab 38 -- we gave you two notebooks that I hope you have there -- and that's Exhibit AT000848. This is an article that appeared in the American Journal of Epidemiology; is that correct?
A. Yes, it is.
Q. And the American Journal of Epidemiology is the Johns Hopkins medical journal?
A. It's the official publication of the Society for Epidemiologic Research. The journal is owned by Johns Hopkins University.
Q. Okay. And it was published in nineteen ninety -- 1996; is that correct?
A. That's correct.
Q. And it was published --
Well let me ask you this: Is the American Journal of Epidemiology a peer- review journal?
A. Yes, it is.
Q. Is it well respected in the field?
A. Yes.
Q. An authority in the field?
A. Yes.
Q. And this study was conducted by -- I'm going to --
Why don't I ask you to read off these names.
A. Well I may no -- do no better than you would have, but these would appear to be Finnish names: Paul Knekt, Raimo Raitasalo, Markku Heliovaara, Ville Lehtinen, Eero Pukkala, Lyly Teppo, Jouni Maatela and Arpo Aromaa.
Q. These are reliable authorities in their field?
A. I don't know them personally.
Q. You don't -- you don't know them personally, but the journal --
Are you an editor on this journal?
A. Yes, I am.
Q. Certainly the journal believes that they were reliable experts in the field; correct?
A. The journal accepted the article for publication.
Q. And they must have thought that they were reliable experts in their field.
A. Again, the journal review process evaluates articles.
MR. GARNICK: Okay. I move the admission of this as a learned treatise, Exhibit AT000848.
MR. HAMLIN: No objection, Your Honor.
THE COURT: It will be received into evidence.
BY MR. GARNICK:
Q. Let me read it. These authors concluded, "The relative risks of lung cancer between smokers and non-smokers were 3.38 at normal depressiveness score levels...."
Let me pause there. Does this, what I just read, does that mean that if you take smokers and non-smokers at the same normal depressiveness levels, they're -- they're not particularly depressed, they have the same level of depression, that the relative risk of smoking and lung cancer is 3.38? Is that what it means?
A. I think you're restating what that piece of the sentence says.
Q. So if I --
And this was 1996; correct?
A. The publication?
Q. The publication.
A. Yes.
Q. Okay. And then it goes on and it says, "...and 19.67 at strongly elevated levels, respectively." And that means that if you compare smokers and non-smokers at strongly elevated depressive levels, people who are very depressed, their relative risk jumps to 19.67; is that correct?
A. Well that's what this sentence says. But again, that would depend on whether there was any bias or improper analysis leading to that conclusion.
Q. So you have some concerns about the validity of this study; is that what you're saying?
A. I was given this study to review during my deposition, and I noted that the analysis that led to that sentence in fact did not take into account the quantitative dimensions of smoking. It was simply a comparison of smokers versus non-smokers and did not take into account the actual amount smoked.
Q. Do you know of any other studies on this particular subject?
A. I have not made a systematic review of this issue.
Q. Is it true that by the time some possible causal factor emerges as a risk factor in the literature, it would have been found statistically significant in a number of studies?
A. That would be the usual case. Statistical significance is often one of the criteria for publication of articles.
Q. Well you would not call a possible exposure a risk factor if it was just -- if there was just one statistical study finding a statistical association; would you?
A. I think it would depend on the study and the context. I can't generalize.
Q. Well do you remember your deposition was taken in this case?
A. Yes.
Q. And do you remember being asked this question and giving this answer --
MR. HAMLIN: Your Honor, I object to having it displayed on the screen.
THE COURT: Do not display it at this time, counsel.
MR. GARNICK: All right.
BY MR. GARNICK:
Q. "Question: So is statistical significance necessary to be a risk factor? For a factor to be regarded as a risk factor, must it be statistically significant?
"Answer: Well typically, by the time one has identified a risk factor, it would have emerged as statistically significant in a number of studies."
Page 64.
MR. HAMLIN: Your Honor, improper use of the deposition. I think at the very least Dr. Samet should be shown the deposition so that he can review the testimony.
THE COURT: Give him a copy of the deposition.
MR. GARNICK: You have a copy there, page 64, volume one.
Q. Do you remember giving -- being asked that question and giving that answer?
A. I think the answer says "typically," and I just answered the question that said it would depend on the context.
Q. And most often, by convention, if a -- in order for a risk factor to be statistically significant, it would have to be found statistically significant at the 95 percent confidence level; is that true?
A. That's convention.
Q. Now in the case of smoking and relative risk, if the relative risk is one, that means that smokers and non-smokers have the same risk; is that correct?
A. That's the comparison that's being made.
Q. Right. And if the risk factor is above one, that would mean that the smoker has a greater risk; correct?
A. That's correct.
Q. And if the risk factor is below one, that means that the smoker has a lesser risk; is that correct?
A. Correct.
Q. Okay. Now when the relative risk is five, that means that the smoker has five times the risk of the non-smoker; correct?
A. Correct.
Q. Or another way of saying it is that the smoker has five hundred percent the risk of the non-smoker; right?
A. Correct.
Q. It doesn't mean that the smoker has five hundred times the risk of the non-smoker; correct?
A. Well, the comparison is to the baseline risk of one.
Q. It's five hundred percent, but not five hundred times the risk of the non-smoker.
A. It's five hundred percent.
Q. Okay. Now I want to explore a little bit how relative risk works. Let's say that you have population A, and population A consists of 200 people, and to make the hypothetical easy, let's say that we have a hundred smokers and a hundred non-smokers. And let's suppose that we're talking about heart disease. Okay?
A. Fine.
Q. All right. And we conduct a study comparing smokers and non-smokers for heart disease and we find that heart -- that smokers -- over a period of time we see ten cases of heart disease. Okay? And in the non-smokers we see five cases of heart disease. Okay?
The relative risk, then, would be two; right?
A. That would be ten over 100 and divided by five over 100, which would be two.
Q. Right. And because we're using --
Because in our hypothetical there's the same number of smokers and non- smokers, it's ten over five, which is two; correct, if the population of smokers and non-smokers are the same?
A. Fine. Correct.
Q. Okay. Now isn't it true, though, that smokers tend to follow a less- healthy lifestyle than non-smokers?
MR. HAMLIN: Objection, Your Honor, to the form of the question. Vague.
THE COURT: You can answer that.
A. I can only answer that question in the context of which smokers and at what -- at what time interval.
Q. You can't tell me generally if smokers tend to have a less-healthy lifestyle than non-smokers?
MR. HAMLIN: Objection, asked and answered.
THE COURT: It's been asked and answered.
Q. Well can you tell me if smokers today tend to have less education than non-smokers?
A. Well now you've given me a context, you said "today." And in the context of today, smokers tend to have less education, as you suggested.
Q. And today, do smokers have a less-healthy lifestyle?
MR. HAMLIN: Objection to the form, it's vague. That hasn't cured it.
THE COURT: You may answer.
A. Again, there have been some studies that have compared the, if you will, the general lifestyle, some disease risk factors in smokers and non-smokers today for current time with the suggestion -- the finding that profiles indicate somewhat less-healthy lifestyles in the current smokers compared to the current never smokers.
Q. In fact, smokers are more likely today to be overweight than non- smokers; isn't that true?
A. I don't think that's correct. I would have to see some data showing me that.
Q. Smokers tend to exercise less than non-smokers?
A. I believe that's correct, today.
Q. Smokers tend to drink more alcohol than non-smokers?
A. Again, I believe that's correct, today.
Q. Okay. In point of fact, as the prevalence of smoking -- as the rate of smoking has decreased, the differences between smokers and non-smokers have increased; isn't that right?
A. I'm sorry. Which differences?
Q. The differences in the -- in the nature of their lifestyle. Smokers tend to have a less-healthy lifestyle than non-smokers.
A. Well we talked about data for today. I'm not sure that comparison is available for a substantial period of time, but --
Q. Could you turn to tab 39, please. This is an editorial commentary that you wrote for the American -- I'm sorry. It's --
MR. HAMLIN: What is the exhibit number?
MR. GARNICK: BYT000271.
Q. This is an editorial commentary that you wrote for the American Journal of Epidemiology called "Editorial Commentary: New Effects of Active and Passive Smoking on Reproduction?" And as I said, you wrote this; correct?
A. Yes, I did.
Q. You wrote this in 1990?
A. Yes.
Q. And you are a reliable expert in the field?
A. Yes.
MR. GARNICK: I move for the admission as a learned treatise.
MR. HAMLIN: No objection, Your Honor.
THE COURT: The court will receive BYT000271.
BY MR. GARNICK:
Q. Now in this editorial you wrote, on page 349, at the top, "As the prevalence of smoking has declined, smokers have become increasingly distinct from non-smokers in other aspects of life-style as well. Smokers today tend to have less education and lower income, to drink more alcohol, and to generally follow a less healthy life-style than non-smokers." And you wrote that; correct?
A. I wrote that. The references for that are, I think, numbers one and four, the '89 Surgeon General's report and the '90 Surgeon General's report, and those were in reference to relatively recent data.
Again, I thought the question I was asked before had to do with the time period over which changes had occurred.
Q. Well you agree that as the prevalence of smoking has declined, smokers have become increasingly distinct from non-smokers, correct?
A. That's based on data cited from the '89 and 1990 Surgeon General's reports.
Q. Okay. Now going back to population A, if we just looked at smoking and heart disease and didn't take into account exercise or alcohol use, it is certainly possible that some of the five extra deaths that we see on the smoking side of the ledger might be due to lack of exercise or alcohol consumption; correct?
A. As you said, it's possible, but it would only depend on the patterns of the association in this particular population between those factors and smoking or not smoking.
Q. Right. You can't -- you can't say, without knowing more information, whether some of these five extra deaths are caused by these other risk factors; correct?
A. As you said, it's possible.
Q. Okay. And this is -- this is called confounding; correct? If some of these five extra deaths are really caused by alcohol or lack of exercise, then this relative risk of two would be confounded; is that right?
A. That would be the term that would be used.
Q. Okay. And if a person --
If you were seeing a person or treating a person for heart disease and the person didn't exercise and drank excessive quantities of alcohol and smoked, you as a physician could not determine which of those risk factors, if any, caused the specific heart disease in the individual; could you?
A. In a particular individual, using that information, no.
Q. Okay. Let's take another example of confounding. Do you agree that alcohol causes cirrhosis of the liver; correct?
A. Alcohol causes cirrhosis of the liver.
Q. And would you also agree that smoking does not cause cirrhosis of the liver?
A. To my knowledge, smoking does not cause cirrhosis of the liver.
Q. And would you agree that smokers drink more more often than non-smokers?
A. At present, data would indicate that.
Q. Right. Now because smokers drink more than non-smokers, it therefore would follow that smokers would have more cirrhosis of the liver than non- smokers; correct?
A. If that were true in the group you were referring to in this hypothetical, yes.
Q. Okay. And if one did a statistical study of smoking and cirrhosis in that population, one would find a statistical association between smoking and cirrhosis; correct?
A. Well that would depend on how the study was done and whether the alcohol consumption was taken into consideration, and, as -- as I discussed, confounding was considered and properly dealt with.
Q. But if you didn't take alcohol into account, you just looked at smoking and you looked at cirrhosis --
A. I -- I -- I can't imagine doing a study now of cirrhosis and not taking alcohol into account. I just can't imagine it.
Q. And the reason is?
A. Because we know that alcohol is a cause of cirrhosis.
Q. But if you did do that study --
First of all, that study has been done in the past; correct?
A. I don't know.
Q. Did any of the studies you talked about yesterday do that, measure the relationship between smoking and cirrhosis of the liver?
A. I'm sure that there were mortality figures reported for various diseases.
Q. Including cirrhosis of the liver?
A. I'll assume so, in these all-cause mortality studies.
Q. If you did such a study -- I know you can't imagine it, but if you did such a study, you would find a statistical association between smoking and cirrhosis; correct?
MR. HAMLIN: Objection, asked and answered, Your Honor.
THE COURT: It's been asked and answered.
Q. Doctor, please turn to tab 40 of the notebook. This is Plaintiffs' Exhibit 16660.
Now you have read this study; is that correct?
A. I've not read this study in depth, no.
Q. You have not read this study?
A. I was given this study several days ago.
Q. Well this is one of the studies in your database; isn't it?
A. Yeah, I said I've not read it in depth recently.
Q. Well have -- have you read it at all?
A. I'm familiar with this study.
Q. Have -- have you read each page of the study?
A. In the last two days I was handed approximately 70 studies, I have looked through the pages of this. I have not re-read this study --
Q. But you read it --
A. -- in depth.
Q. Prior than two days -- prior --
Before two days ago, had you read this study?
A. I've read this study in the past.
Q. And you've read it page -- page to page; right?
A. That's what reading a study means.
Q. Okay. And this study looked at alcohol, smoking, various occupational factors, and looked at cancer of the oral cavity, pharynx and larynx; right?
A. Correct.
Q. And at one point the study compared and tried to measure the association between smoking and cancer in the oral cavity, pharynx and larynx; correct? All three sites together.
A. Could you direct me to --
Q. Certainly. Page 606.
Now it says, "With regard to smoking, the crude relative risk rose irregularly from a reference value in non-smokers" --
That would probably be one; correct? The reference value of non-smokers?
A. Yes.
Q. -- "to a value of 3.7 in those smoking more than 50 cigarettes a day. Controlling for alcohol consumption reduced this association somewhat, although it remained significant. However, control for socioeconomic and marital status, dental care and a history of tuberculosis reduced the association with smoking still further and it did not remain statistically significant." Do you see that?
A. Yes, I do.
Q. And this would be an example of confounding; is that correct?
A. I don't think this is a very good example of confounding. This may be a good example of over- adjustment in fact. For example, to my knowledge, history of tuberculosis is not a risk factor for these cancers, so I don't understand why the authors controlled for it. And again, the dental care, perhaps relevant, but certainly not to cancer of the larynx. So I -- I think that the investigators here may have put too many factors into this particular adjustment.
Q. The investigators thought that those adjustments were necessary to unconfound the association; correct?
MR. HAMLIN: Objection, foundation, Your Honor.
THE COURT: You may answer.
A. Well again, I gave my view of, for example, the inclusion of history of tuberculosis, which is simply not a cause of cancer of the larynx or the oral pharynx.
Q. Well do you have any reason to believe why these researchers controlled for these variables?
A. I have no basis for knowing why they controlled for tuberculosis.
Q. So you believe that they -- that they controlled for too many factors, and as a result, the risk went away?
A. Well I'm suggesting that they controlled for some factors that are not potential confounding factors, because at least in the case of history of tuberculosis, the definition of a confounding factor was not met.
Q. But it still reduced the risk; correct?
A. Well that's because they included another variable in the adjustment, or several more variables.
Q. Would you agree that to the extent that one can, one would like to have as unconfounded an estimate as possible?
A. Yes.
Q. And would you agree with me that confounding can artificially increase or decrease the magnitude of a statistical association?
A. The magnitude of a risk, yes.
Q. And would you agree with me that if one was going to study heart disease and smoking, if one had all the available data, there would be a number of risk factors one would want to control for?
A. Again, it would depend on the associations in the data at hand, but in estimating the effects of any factor, whether smoking or some other factor, one would like to have unconfounded estimates.
Q. What do you mean by "associations in the data at hand?"
A. Confounding is a matter of associations in the data at hand.
Q. Can you tell me what that means?
A. Yes. A confounding factor by definition is associated with the disease itself, and in the data at hand, associated with the exposure that one is trying -- making estimates of its effect.
Q. So what --
Are you saying that what risk factors you may want to take into account depends upon the unique nature of the data, the characteristics of the people in the sample?
A. Well let me give an obvious example. If one is studying heart disease and the sample is only considering -- includes men, then sex is not confounding and one would not consider sex as a potential confounder.
Q. So it does depend upon the unique -- the nature of the sample you're studying; is that correct?
MR. HAMLIN: Objection, asked and answered.
THE COURT: It's been asked and answered.
Q. Now would you agree with me that every study does not find the same relative risk?
A. I'm sorry, I can't answer that question as posed.
Q. Well if one is looking at -- I think that's fair.
If one is looking at smoking and heart disease --
THE COURT: Counsel, counsel, no commentary.
MR. GARNICK: I apologize.
THE COURT: If you have a question, just direct it to the witness, please.
BY MR. GARNICK:
Q. If one is looking at smoking and heart disease, one would not expect that every study would find the same relative risk; is that correct?
A. If by "the same relative risk" you mean exactly the same value to a decimal point, no.
Q. Well in fact, medical specialists have found different relative risks for populations in different geographical locations; isn't that correct?
A. For --
Q. Let's say smoking -- smoking and heart disease.
A. I don't have data at hand, but I suspect that there's some variation across countries.
Q. And even within the country, if one was comparing Boston to New Mexico -- or Massachusetts and New Mexico, one could find a difference in relative risk?
A. If you can show me some data, I can comment on it.
Q. Would you agree that most chronic diseases show a marked geographic difference in prevalence?
A. No, I can't agree that most chronic diseases show a marked geographic difference in prevalence.
Q. Would you agree that most chronic diseases show a marked graphic -- I mean geographic variation in prevalence? That might be the same thing as I said, but maybe that's --
A. I -- I can't answer the question because I don't know what most are and I don't know what you mean by "marked."
Q. Well let's go on then. Let's go back to --
THE COURT: Why don't we take a short recess.
MR. GARNICK: Sure.
THE CLERK: Court stands in recess.
(Recess taken.)
THE CLERK: All rise. Court is again in session.
(Jury enters the courtroom.)
THE CLERK: Please be seated.
MR. GARNICK: Thank you, Your Honor.
BY MR. GARNICK:
Q. Dr. Samet, let's go back to our hypothetical population A. And there, as you'll recall, we had 200 people together, 100 smokers, 100 non-smokers, and ten of the smokers developed heart disease, ten -- oh, I made a mistake -- five of the non-smokers developed heart disease; correct? And we found the relative risk was two.
A. I recall.
Q. Okay. Now let's suppose that in addition to population A we have a hypothetical population B, and let's suppose that this population B consists, again, of 200 people, 100 smokers, 100 non-smokers. And let's further suppose that these -- that the 100 smokers had the same characteristics as the same -- as the 100 smokers in population A. Smokers were the same. And then again we saw that ten of the smokers got heart disease. Do you follow me?
A. Yes.
Q. Okay. And let's suppose that the non-smokers in our population B were healthier, had generally healthier lifestyles, exercised more, drank less than the non-smokers in population A, and so instead of observing five heart-disease incidents in population A, we observed three heart-disease incidents in population B. Okay?
A. Yes.
Q. Okay. The relative risk, then, for population B would be approximately 3.3; is that right?
A. Yes. I follow.
Q. Okay. And so what we see here, that even though the smokers were basically the same in population A and population B, that even though the same percentage of smokers in population A and B developed heart disease, the relative risk was different because of the actions and behaviors of non- smokers; correct?
A. That's true in the hypothetical example you've constructed.
Q. Okay. And in fact, in this hypothetical, as non-smokers became healthier and had less incidence of heart disease, the relative risk for smoking went up; correct?
A. I don't know if they became healthier. You just gave me a hypothetical comparing two populations.
Q. Well certainly they -- they developed less incidence of heart disease; correct? That's inherent in the hypothetical; correct?
A. Comparing population A and population B.
Q. Right. And as their rate of disease decreased, the relative risk of smokers increased even though the absolute rate of incidence of heart disease in the smoking population remained the same; is that right?
A. I can't quite follow, because it sounds like you're talking about something changing over time, but I'm just seeing a snapshot in population A and in population B.
Q. Well would it matter if it changed over time? Let's say the population A was 50 years ago and population B was today, but we're dealing with different people, not the same people in population A and population B, but it's the same geographic location. Okay? Am I clear?
A. Let me restate so I'm sure I'm clear. You're saying that population A is one population, and then population B follows by 50 years?
Q. Yes.
A. Okay. That's the hypothetical.
Q. Yes.
A. Okay.
Q. Okay. And we are looking at the relationship between smoking and heart disease. Okay?
A. Okay.
Q. And we find that for smokers, ten percent of the smokers in both population A and population B develop heart disease.
A. Okay.
Q. Right? That's what this hypothetical shows; right?
A. Yes.
Q. Okay. But the percentage of non-smokers who contract heart disease decrease from population A to population B. Okay?
A. Yes.
Q. The result would be a higher relative risk in population B than in -- than in population A for smoking and heart disease; is that correct?
A. If one did not take into account the factors that you suggested are confounding this comparison.
Q. Okay. And let's say we didn't know what factors were confounding this comparison. Let's say that we just looked at smoking and heart disease and we don't know why population B sees less heart disease in the non-smoking population. The result remains the same; right? The relative risk would still go up.
A. But this is a hypothetical, and we do know something about causes of heart disease that might confound such comparisons in the real world.
Q. But we see here, though, that -- certainly as a proposition, that the relative risk might well be affected by causes of the disease other than the one that we are studying; correct?
A. Again, if those other causes were not taken into account and the estimate of relative risk was left confounded.
Q. You would agree that the proportion of an illness attributable to smoking might vary from population to population depending upon the amounts of that illness caused by other diseases?
A. Do you have something you're reading from that I should see or that --
Q. Well generally speaking do you think -- is that --
Is that true?
A. I'm sorry, can you state the question again?
Q. That the proportion of illness attributable -- let's say the proportion of heart disease attributable to smoking will vary depending upon the amount of other risk factors that are in that population, risk factors for heart disease.
A. It's not a -- it's not a simple question to answer. It depends on what those risk factors are and how those risk factors act in combination with -- with smoking. The answer might be yes and might be no.
Q. But it's something you'd want to take a look at; right, to know the answer in a specific data set?
A. Well if I had a specific data set, I might explore that. I thought the question referred to comparisons.
Q. Well let's go back to your Exhibit 30170.
Now if one looks at men ages 55 to 59, 40 cigarettes a day for 40 to 44 years, one sees a huge increase in the relative risk, from 23.7 to 136.6; correct?
A. Correct.
Q. But yet one sees an increase in actual death rate per 100,000 persons of eight, between seven and eight; correct?
A. You're referring to the difference between 341.6 and 333.9, or --
Q. Yes. Between seven and eight deaths; right?
A. That difference is approximately eight.
Q. Right. And if one is going to do two studies, two million-person studies, one would expect to see some variation between one study and another. Do you not --
You don't expect to see the exact same results to the decimal point; do you?
A. One would not necessarily expect that, no.
Q. There's going to be statistical fluctuation; correct?
A. Well there may be fluctuation.
Q. Okay. So here you have a difference in the absolute rate of seven or eight individuals, but a difference in relative risk of over a hundred. That's what we see; correct?
A. Although I see in the line above an increase in absolute risk of approximately 100.
Q. Right. But the line below, my statement is correct; right?
A. The line below, your statement is correct.
Q. And what --
And so if one looks at the never smokers, though, one sees that the incidence of disease in the never smokers have dropped dramatically; haven't they?
A. Dropped from 14.1 to 2.5.
Q. That's a dramatic drop; isn't it?
A. Those are your words. It's a drop of about 12.
Q. You would not agree that that's a dramatic drop?
A. I don't have criteria for it, "dramatic."
Q. You don't use "dramatic" in your everyday language?
A. I'm sure there are times when I do.
Q. And even among men ages 60 to 64, the never smoking rate drops from 19.2 to 9.1; correct?
A. Correct.
Q. And the reduced incidence of disease in the never smokers is going to have an impact on the relative risk figures of the smokers; correct?
A. If the comparison is to never smokers, you can calculate the relative risk.
Q. And at least with respect to lung cancer, the never smokers were healthier in '59 to '66 as opposed to '82 and '88; correct?
A. I'm sorry, can you restate the question?
Q. At least with respect to lung cancer, we know that non-smokers were healthier in CPS-II as compared to CPS-I; correct?
A. I'm sorry, I thought I heard the question the other way around.
Q. I might have misstated the question. Let me state it again.
With respect to lung cancer, the never smokers were healthier in the CPS-II study as compared to the CPS-I study; correct?
A. Again, these two age groups, men ages 55 to 59, men ages 60 to 64, the lung cancer mortality rates were lower in the second study compared to the first.
Q. And this is consistent with the notion that as the prevalence of smoking has declined, smokers have become increasingly distinct from non-smokers in other aspects of lifestyle; correct?
A. I really can't comment on what has changed the rates in the never smokers comparing CPS I and II.
Q. You don't know?
A. I don't have information on why those mortality rates have changed comparing the two studies.
Q. Now Dr. Samet, it's not your testimony and you didn't mean to tell the jury that the reason why the relative risk of smokers increased was because of low tar cigarettes; is it?
A. I described --
MR. HAMLIN: Objection. First of all, let -- let me just object. It mischaracterizes his testimony if in fact Mr. -- if Mr. Garnick is suggesting that that was Dr. Samet's testimony.
THE COURT: Counsel, just state your legal objection, please.
MR. HAMLIN: Objection to form.
THE COURT: Sustained.
BY MR. GARNICK:
Q. Dr. Samet, do you believe that --
It is not true, is it, that the reason why the relative risk of smokers increased from CPS-I to CPS-II is the increasing use of low nicotine -- low tar and low nicotine cigarettes?
MR. HAMLIN: Objection, asked and answered.
THE COURT: No, you may answer that.
A. Can you restate the question?
Q. That the relative risk for smokers for lung cancer increased from CPS-I to CPS-II because of the increasing use of low tar cigarettes.
A. I described the increase in relative risks comparing the two studies, I also provided an exhibit that showed what was happening across the years 1955 through approximately 1990 in terms of tar and nicotine yields of cigarettes sold in the United States. That's the information I provided.
Q. And do you have an opinion -- is it your opinion that the reason that the relative risk increased for smokers from CPS-I to CPS-II is because of the increasing use of low tar cigarettes?
A. I think I offered the opinion that the lung cancer risks comparing CPS-I and CPS-II, the relative risks, and also in fact most of -- many of the absolute risk values, did not follow the trend of an approximate 60 percent reduction in tar yields of the cigarettes smoked across the time period 1955 through approximately 1990.
Q. And so --
MR. GARNICK: Your Honor, I move to strike the non-responsiveness.
THE COURT: The answer was responsive.
Q. Dr. Samet, can you tell me why the relative risks increased from CPS-I to CPS-II?
MR. HAMLIN: Objection, Your Honor, asked and answered.
THE COURT: No, you may answer that.
A. In terms of a specific reason that is testable by comparing the two studies? No.
Q. Thank you.
Now this chart does not separate out low tar cigarette smokers from high tar cigarette smokers; is that right?
A. That's correct.
Q. And this chart really doesn't tell us whether low tar cigarette smokers have a reduced lung cancer risk as compared to high tar cigarette smokers; correct?
A. Correct.
Q. Okay. And it doesn't tell us the reverse, it doesn't tell us if high tar cigarette smokers have an increased risk -- have a -- have a lower risk than low tar cigarette smokers; correct?
A. There's no information about tar.
Q. Okay. And in fact there is separate epidemiology on that issue; correct?
A. There are studies of that question.
Q. Now I believe you testified yesterday that manufacturers -- tobacco manufacturers over the years have modified the design of their cigarettes to decrease tar and nicotine yields; is that correct?
A. I described a changing --
Well I described the changing tar and nicotine yields.
Q. Right. And the result of those modifications is that the tar and nicotine yields in fact have dropped dramatically, wouldn't you agree?
A. I described the reduction by approximately 60 percent.
Q. Let me turn your attention to tab 31, which is defense Exhibit BYT000273, but I believe it was admitted into evidence as -- I don't know. But this is a -- this is part of your monograph seven of the NCI; correct? Chapter six, the chapter you wrote.
A. Yes.
Q. Okay. And this is a chart from page 77 of your chapter; correct?
A. Yes.
Q. And this chart shows the declining nicotine and tar ratios -- nicotine and -- and tar delivery levels over time; correct?
A. That's correct.
Q. And we see, according to this chart, that in 1957, the advent of reconstituted tobacco reduced --
A. Correct.
A. -- tar levels; is that correct?
A. This is a -- that's correct -- a chart taken from the Surgeon General's report of 1989, I believe.
Q. And in 1959, the use of porous paper also reduced tar levels; is that correct?
A. These are changes in design that are listed in this figure, correct.
Q. Okay. 1967, the use of expanded tobacco reduced tar levels; is that correct?
A. Again, these items -- dates refer to these changes as identified in the 1989 Surgeon General's report, I believe.
Q. And in 1971, the use of ventilation also reduced tar levels; correct?
A. Again, this is a point on the graph, yes.
Q. And all of these modifications of the cigarette by the manufacturers resulted in a reduction of tar levels; correct?
A. As tested by the FTC method, correct.
Q. And isn't it fair that public health officials in the 1950s and 1960s encouraged the tobacco companies to modify their products to reduce tar and nicotine delivery levels?
A. I think you'd have to show me which public health officials and which companies.
Q. Sitting here, you do not know of any public health official that made that recommendation?
A. Which type of public health official are you referring to? I --
Q. What are the different types of public health officials?
MR. HAMLIN: Your Honor, I'm going to object, this is beyond the scope of his direct examination.
THE COURT: Well, the question is too vague. Ask something more specific, counsel.
MR. GARNICK: Okay.
BY MR. GARNICK:
Q. Isn't it true that as early as the 1950s, Dr. Wynder suggested that tobacco manufacturers take steps to reduce tar levels?
A. Can you show me a specific writing of Dr. Wynder's so that I can see his recommendation?
Q. Well let me refer you to tab 32, the Defendants' Exhibit BYT000207, and let me refer to you -- you to page 1199 --
Well first of all, Dr. Wynder is an authority on smoking and health; correct?
A. Yes.
Q. He's a reliable authority?
A. Yes.
Q. And this article appeared in Cancer, the journal Cancer?
A. Yes.
Q. And that is a peer-review medical journal; correct?
A. Yes, it is.
MR. GARNICK: Move the admission as a learned treatise.
MR. HAMLIN: No objection, Your Honor.
THE COURT: That will be received into evidence.
BY MR. GARNICK:
Q. Now on page 1199, Dr. Wynder said that "It may be predicted that if the average smoker were exposed to only one half the amount of tobacco tar to which the smoker of regular-sized cigarettes is now exposed, his cancer risk would be significantly reduced. Any measure designed to thus reduce man's exposure to tobacco tar, whether through modification of the tobacco or the cigarette, or through more effective filtration, can significantly contribute to the decrease in risk."
You see that?
A. I actually am still having a little trouble with what you put in my --
Did you say tab 32?
Q. I did.
A. And that should be --
I have an article entitled "A Study of Tobacco Carcinogenesis, The Primary Fractions," and I don't have a page --
It was published in Cancer, but mine is from 1957 and goes up through page 271.
Q. Well let me give you -- let me give you my copy, doctor. I think we put the wrong copy in your notebook.
MR. GARNICK: May I approach, Your Honor?
(Document handed to the witness.)
BY MR. GARNICK:
Q. As I said, if you look at page 1199, that is the portion that is highlighted, and there, as I said, doctor --
THE COURT: I think -- I think I have the same problem, counsel.
MR. HAMLIN: Yeah. And so do I, Your Honor.
MR. GARNICK: Do I have the wrong one? Well let's just go on then. I apologize.
Q. You would agree, though -- and we talked about it before -- there is epidemiology, a body of epidemiological studies that compare the risks of lung cancer of high tar cigarettes with low tar cigarettes; correct?
A. There are such epidemiological studies, yes.
Q. And these studies compare the risks of smoking the actual cigarettes. They -- they do not compare -- let me -- let me strike -- strike that.
Because these studies look at the actual impact of low tar cigarettes, they already inherently take into account any compensation that may exist; is that fair?
MR. HAMLIN: Objection, Your Honor. This is beyond the scope. Dr. Samet did not get into compensation. Other experts have testified about that.
MR. GARNICK: I'm not asking him whether any does exist, I'm asking about what these studies measure.
THE COURT: Okay. I think the objection is valid, counsel.
BY MR. GARNICK:
Q. Let's try this again. Dr. Samet, please turn to tab 34, Exhibit 16344.
Have you read this study?
A. Yes, I have.
Q. This is one of the studies in your data set; correct?
A. That's correct.
Q. And in this study --
This is an early epidemiological study comparing high tar cigarettes with low tar cigarettes?
A. As I recall this study, it was a study of the effect of filter cigarettes on the risk of lung cancer, which I think is the title.
Q. In the introduction Dr. Bross says, "Some epidemiological data are presented on the question: Does switching -- Does switching to filter cigarettes reduce the risk of lung cancer? On the basis of these data, the answer is: The risk seems to be reduced to about 60 percent of what it would have been if the smoker had not switched." Correct?
A. That's what it says, yes.
Q. And in 1968 --
In fact, this was the first epidemiological study comparing filter cigarettes with non-filter cigarettes; is that right?
A. My understanding is that Dr. Bross's work was the first on this issue.
Q. And in 1968 he found that filter cigarettes did reduce the risk by 60 percent; correct?
A. Compared to non-filter cigarettes.
Q. Compared to non-filter cigarettes.
And so in 1968, this was the state of medical knowledge on the issue of whether switching from filter cigarettes -- from non-filter cigarettes to filter cigarettes would reduce one's risk of lung cancer; correct?
A. I think this may have been the only body of work available at the time, correct.
Q. And this is a reliable authority? He's -- Dr. Bross is a reliable authority?
A. Yes.
Q. Now of course since 1968 a number of other studies have appeared; correct? And you went through some of those yesterday; right?
A. Yes, I did.
Q. Okay. In 1979 the Surgeon General's report addressed this issue; correct?
A. There was comment upon it, yes.
Q. Comment on it.
This is Exhibit GJ000113, Surgeon General's '79 report. You have it there, I believe, doctor. I'm not sure if this has been admitted into evidence. I know there's been a stipulation.
It has been. Okay.
And in 1979 the Surgeon General said that "Overall mortality increases with the tar and nicotine content of cigarette smoke." Correct?
MR. HAMLIN: Counsel, could we have a page?
A. Yeah. Can you direct me to a page, please?
Q. Page 2-22.
A. This sentence would appear to be -- would appear to be reference to the study we addressed yesterday, CPS-I.
Q. Right. And the way the Surgeon --
The Surgeon General interpreted that study as finding that overall mortality increases with the tar and nicotine content of cigarette smoke; correct?
A. Yes. And I believe those were the data that I showed yesterday.
Q. Okay. And then on page 5-31, the Surgeon General found that "The long- term use (10 years or more) of filter cigarettes is associated with lower death rates from lung cancer than those experienced by people who smoke an equal number of non-filter cigarettes." Correct?
A. That's what it says, yes.
Q. And in fact, as you have testified, in 1981 the Surgeon General found that the filter-tip lower tar cigarettes produced lower rates of lung cancers than do the higher tar cigarettes; correct?
A. Can you again give me a reference for that '81 report?
Q. Okay. This is plaintiffs' Trial Exhibit 26005, page 18. He states "Today's filter-tipped lower 'tar' and nicotine" --
THE COURT: Counsel, counsel --
MR. GARNICK: Oh, I'm sorry.
(Witness retrieves book.)
THE WITNESS: I'm sorry, I have it now.
MR. GARNICK: I apologize, doctor.
THE WITNESS: And which -- which page?
MR. GARNICK: Page 18.
THE WITNESS: Okay.
Q. The Surgeon General concluded that "Today's filter-tipped, lower 'tar' and nicotine cigarettes produce lower rates of lung cancer than do their higher 'tar' and nicotine predecessors." Correct?
A. Correct. And then follows with the sentence beginning with "Nonetheless."
Q. Right. "Nonetheless, smokers of lower 'tar' and nicotine cigarettes have much higher lung cancer incidence and mortality than do non-smokers."
It was the state of medical knowledge in 1981 that the filter-tipped lower tar/nicotine cigarettes produced lower rates of lung cancer than do higher tar/nicotine predecessors; correct?
A. That was the conclusion.
Q. And that was the state of medical knowledge at the time; correct?
A. At the time.
Q. Item four says that "The occurrence of laryngeal cancer has been reported to be reduced among smokers who use filtered cigarettes, compared with those who use non-filtered cigarettes." Correct?
A. That's what it states, correct.
Q. And that was the state of medical knowledge at the time; correct?
A. Albeit limited, correct.
Q. Now in your chapter, chapter six of monograph seven of the NCI report on tar and nicotine levels, you concluded that as of 1996, the Surgeon General's conclusion with respect to low tar cigarettes and lung cancer were still valid; correct?
A. That's correct.
Q. And so in 1996, the state of medical knowledge was that low tar cigarettes reduced the risk of lung cancer as compared to higher tar cigarettes; correct?
A. That comparison was based on -- that --
Correct. And that conclusion was based on essentially cross-sectional epidemiological studies that compared higher tar and lower tar smokers at specific time intervals.
Q. And you estimated that that reduction in risk between high tar and low tar cigarettes were about 20 -- was about 20 percent; correct?
A. Yes.
Q. Now the estimate of 20 percent does not take into account more recent brands of cigarettes that were ultra low in tar; correct?
A. Well these estimates can only take into account the products that have contributed to causation of lung cancer in the people studied.
Q. Which would not yet include the very low tar cigarettes that are currently on the market; is that right?
A. As of what date should I consider in answering your question?
Q. Well let me ask you that question: What would be the cutoff date, in your view, that would --
Cigarettes sold before what date would be captured by these epidemiological studies?
A. Well I don't use the definition --
I don't have a definition of ultra low tar, so if you can provide when I should think about that time interval, I'll try and answer the question.
Q. Well what is the lowest tar delivery cigarette on the market today?
A. I don't track that personally.
Q. Well let's go back to this chart. Would tar levels of ten milligrams be captured by the epidemiology?
A. Well again, I think if a --
It would depend on when the epidemiological study was -- was done. And if you can give me a reference to a specific date when a study was done, I'll try and tell you whether that could have been studied yet.
Q. Okay. Well what -- what studies do you rely upon for your estimate of 20 percent?
A. Well I'm relying on the CPS-I study; there was my own study in New Mexico; there was a study in New Jersey, Wilcox, cited here; the more recent studies by Wynder; as well as past studies by Wynder.
Q. And what was the most recent date of the most recent study?
A. The most recent study that I'm familiar with is probably in 1997.
Q. So is it your testimony that cigarettes sold after 1997 are taken into account by your estimate?
A. No.
Q. You rely on a 1997 study, though, for your estimate; correct?
A. Well that study was published in 1997. And at the moment, without a study in front of me, I can't remember the exact dates covered by Dr. Wynder and his colleagues in these studies. It's probably through approximately 1990.
Q. Well let's turn to Exhibit 15914. This is your New Mexico study. Now I believe you testified yesterday that for non-Hispanics, the observed reduction in risk from high tar to low tar cigarettes was about 20 percent; correct?
A. Well the comparison was based on filter smoking in this study. The comparison was by percentage of the smoking history moving from a hundred percent non-filter to 100 percent filter.
Q. And you told us that there was a greater decline in risk for Hispanics; correct?
A. Yes, I did.
Q. What was that decline in risk?
A. Well I'll read the numbers. Non-filter only was set equal to one; one to 33 percent filter, .56; 34 to 66 percent filter, .39; 67 to 99 percent filter, .26; filter only, .04.
Q. So would that be a 96 percent reduction in risk?
A. Following these numbers, yes.
Q. And was that found to be statistically significant?
A. The decline in risk was, yes.
Q. Now doctor, medical specialists have examined the smoke chemistry of tar that comes from low tar cigarettes as compared to tar that comes from high tar cigarettes; correct?
A. I'm familiar with publications in that area. I don't consider myself to be a chemist.
Q. Let me direct your attention back to the 1991 Surgeon General's report, page 20.
A. I'm sorry, 1991?
Q. 1992. I'm sorry, doctor.
A. I don't have a 1992 report.
Q. We'll do it after lunch.
Do you know that Dr. Doll --
Dr. Doll is a reliable expert in the field; correct?
A. Referring to Sir Richard Doll?
Q. Sir Richard Doll, yes.
A. Yes, he is.
Q. And do you know whether he believes it's been proven beyond a reasonable doubt that low tar cigarettes are safer with respect to lung cancer than high tar cigarettes?
A. I know that Dr. Doll and his colleague, Richard Peto, have written on this topic.
Q. Do you know what he has written?
A. I'm familiar with some of his analyses.
Q. Well he would agree --
Isn't it true that it's his view that it's been proven beyond a reasonable doubt that low tar cigarettes pose less cancer risk than high tar cigarettes?
A. He may have said that. I would really need to see the specific statement.
Q. And you would agree that a smoker is better off smoking a low tar cigarette than a high tar cigarette?
A. I'm not sure that I can agree with that, no.
Q. Well certainly with respect to lung cancer you would agree with that; wouldn't you?
A. The evidence shows a reduction in lung cancer risk cross-sectionally in time, as we've said.
Q. Now let's suppose that we wanted to take the relative risk from one population and apply it to another population. Okay? You would attempt to use a relative risk -- well let's -- let's go back to our example.
If we are interested in a relative risk in population B and we didn't know the relative risk in population B, you would agree that before we used the relative risk from population A, we would want to study to make sure that the two populations were comparable; correct?
A. I'm sorry. You're referring to the hypothetical example of --
Q. Just hypothetically. Hypothetically.
A. And --
Q. We're interested in what the relative risk is in a -- in a specific population, which we'll call population B, and we want to take the relative risk that we have derived from population A and use it in population B.
Wouldn't we first want to study and to ensure that the two populations were comparable?
A. I'm sorry. Tell me how I'm using the relative risk value extending from --
Q. To derive attributable risk, let's say. You want to find out the attributable risk.
What is attributable risk again?
A. Well the attributable risk as I defined it, I guess two days ago, was the difference in risk in one group, the exposed group, less the risk in the non-exposed group.
Q. And you use relative risk to determine attributable risk; correct?
A. Well the attributable risk I just discussed was the difference in risk comparing the exposed and non-exposed.
Q. But you need the number, the relative risk number in order to derive the attributable risk number; correct?
A. You may be referring to the calculation population attributable risk.
Q. Okay. Population attributable risk. And let us say that we wanted to find out the population attributable risk in population B. We would need the relative risk number; correct, for population B to do that? We would need a relative risk number; correct?
A. Relative risk is used to calculate what I call the population attributable risk. This is not what I described before when I talked about attributable risk.
Q. Okay. But if we didn't have the relative risk for population B and we wanted to use the relative risk from a different population, before doing that, wouldn't you want to compare the two populations to make sure that they were similar?
A. I would have --
I can only answer that in the context of which disease and which factors might be of concern in the comparison. I can't offer a sweeping answer for all -- all diseases and all factors.
Q. Well to the extent that a relative risk can be generalized to another population, before doing that, you would want to compare the two populations; right?
A. One might. I mean again, it would just -- it would simply depend on the context and the other factors that are being considered.
Q. Well let's say that you wanted to -- well what -- okay. Let's go through that way. What --
What would it depend upon? It would depend upon the other factors --
A. It would depend on the exposure for which the relative risk was estimated, and the other factors that might be relevant to generalizing from one population to another.
Q. Okay. And in the context of, let's say, smoking and heart disease, what might some of those other factors be?
A. Well I think before, I listed many of those risk factors when you asked me about information that might be collected during a history.
Q. And you would want to compare those lists of risk factors between the two populations before using -- before generalizing the relative risk from population A to population B; is that correct?
A. You might. I mean it really would depend on the -- the purpose for which this is being done.
Q. Well let's say the purpose is that you want to know what the relative risk is in population B.
A. Well if I wanted to know the relative risk in population B, I would do a study to properly measure it and take those other factors into account.
Q. You would do a study on population B.
A. In this hypothetical.
Q. And in fact, everything being equal, if you wanted to know what the relative risk is in a certain population, you would do a study on that population; wouldn't you?
A. Again, it would depend on the question and context.
Q. Could you explain that, please?
A. Well --
Q. What context would you have to know? Everything being equal, you -- if you wanted to derive the population -- the relative risk of a population B, you would want to do a study of population B.
MR. HAMLIN: Objection to form, vague.
THE COURT: Do you understand the question?
THE WITNESS: Perhaps if I could hear it again, I --
THE COURT: Maybe rephrase it, counsel.
MR. GARNICK: Okay.
BY MR. GARNICK:
Q. If you wanted to generalize the relative risk from population A to population B -- no, strike that.
If you wanted to derive the relative risk of population B, the best way to do that would be to conduct a statistical study of population B; correct?
MR. HAMLIN: Objection, asked and answered.
THE COURT: You may answer that.
A. If I wanted to estimate the relative risk of a disease in a specific population, I would not do a statistical study, I would do an epidemiological study to make that estimate.
Q. Well the epidemiological study would certainly use statistics; right?
A. Statistics might be used in the analysis, yes.
Q. Now would you agree that the proportion of illness attributable to smoking differs from country to country?
A. It likely does. I don't have information at my fingertips, but likely does.
Q. And would you agree that the proportion of illness attributable to smoking likely differs from the United States as a whole to certain -- to various subpopulations?
A. It might, depending on how you defined the subpopulations. I mean again, leaving "subpopulations" undefined is a hypothetical.
Q. Well with respect to certain subpopulations, it might very well be different; correct?
A. Perhaps certain subpopulations.
Q. And would you agree that empidemiologists have done a poor job of researching the risks of smoking in special populations?
A. Well "special populations" is a term used now --
Let me make sure we understand what you mean by "special populations." I'll say that as used now in the context of research, this is a term used by the National Cancer Institute and others to refer to minority populations and often socioeconomically disadvantaged populations.
Q. And the Medicaid population is a socioeconomic special population; correct?
A. That's correct.
Q. And would you agree that empidemiologists have done a poor job of researching the risks of smoking in special populations?
A. Well today such populations have not been well-studied for many, many diseases and many causes of diseases.
Q. The GAMC population is a special population; correct?
A. It would fit within the broad definition I gave.
Q. Now have you studied the characteristics of the Medicaid -- of the Minnesota Medicaid population?
A. Have I personally?
Q. Yes.
A. No.
Q. Do you know whether Medicaid recipients tend to be less healthy than the general population in Minnesota?
A. I can't answer that.
Q. You don't know?
A. I can't answer that, no.
Q. Do you know if Medicaid recipients tend to be -- are more likely to be disabled than in the general -- than people in the general population?
A. Are you referring to in Minnesota or --
Q. Yes, Minnesota.
A. I don't have an answer.
Q. Do you know what the criteria for eligibility are in Minnesota?
A. I don't know your specific criteria, no.
Q. Do you know what generally the criteria is as far as the federal standards are concerned?
A. I don't know them at hand at the moment. In the past I've known them in specific state where I practiced, New Mexico.
Q. Well it's true, isn't it, that Medicaid covers the disabled?
A. Yes.
Q. And it follows, then, that it is likely that there's a higher proportion of disabled people in the Medicaid population than in the general population; correct?
A. That's true.
Q. And have you shown us any epidemiological study yesterday that specifically addresses the Medicaid population?
A. In general or --
Q. Any Medicaid population.
A. No, I did not.
MR. GARNICK: Your Honor, this is a good point, if you want to break for lunch.
THE COURT: All right. We'll recess for lunch and reconvene at 1:30.
(Recess taken.)
THE CLERK: All rise. Court is again in session.
(Jury enters the courtroom.)
THE CLERK: Please be seated.
MR. GARNICK: Thank you, Your Honor.
BY MR. GARNICK:
Q. Good afternoon, Dr. Samet.
A. Good afternoon.
Q. I want to follow up on one point that we talked about this morning.
Before the break I asked you if in 1996 it was the state of medical knowledge that low tar cigarettes reduced the risk of lung cancer as compared to high tar cigarettes. Do you remember that?
A. Yes.
Q. And I believe you answered that the conclusion was based essentially on cross-sectional epidemiological studies; is that correct?
A. Well the studies are cross-sectional in the sense that they're looking at the case-control studies, and in a sense cohort studies, looking at those cases of lung cancer occurring during some specific point in time. Like my study in New Mexico, the case-control study was done in 1980 to '82.
Q. Well you had the results of the CPS-II study at the time that you rendered your expert report -- that you prepared your expert report in this case; correct?
A. Correct.
Q. And in preparing your expert report, you were careful to review all pertinent literature on the subject matters covered in that expert report; right?
A. Correct.
Q. And in your expert report you stated -- well you gave the opinion that smokers of filtered cigarettes, in comparison with smokers of non-filtered cigarettes, have an approximately 20 percent lower risk of lung cancer; correct?
A. Yes.
MR. HAMLIN: Your Honor, I think this is improper. Could we at least have Dr. Samet have his expert report in front of him so that he could follow along?
THE COURT: Do you have any objection to that, counsel?
MR. GARNICK: None, none at all.
A. Just if you can tell me the page, please.
Q. Page 14. You stated in the expert report that smokers of filter cigarettes in comparison with smokers of non-filter cigarettes have an approximately 20 percent lower risk of lung cancer, although the absolute risk remains substantially above that for never smokers; correct?
A. Yes.
Q. And in fact, the state of medical knowledge today is that low tar cigarettes or -- let -- let me strike that.
The state of medical knowledge today is that smokers of filter cigarettes, in comparison with smokers of non-filter cigarettes, have an approximately 20 percent lower risk of lung cancer; correct?
A. I think the distinction I was trying to draw this morning, that at any one particular time, as in my study or other studies, let's say 1980-'82 or later, looking at smokers, higher tar versus lower tar, or filter versus non- filter, there's been evidence of a reduction in risk, approximately 20 percent or so as -- as mentioned here. That speaks to what I meant by this cross- sectional comparison of the risks in smokers of, in this case, filter compared with non-filter, or as we saw in the CPS-I study, that span of time, cases occurring '59 through '72, given that study, a reduction of risk. And that's what I meant early this morning when I said these works, they're snapshots in time of what the risks are of smoking these different types of products.
Q. Let me see if I understand. What you're saying is that if you compare smokers of filtered cigarettes with smokers of non-filtered cigarettes at the same time, the smokers of filtered cigarettes will have a 20 percent lower risk of lung cancer; is that it?
A. I think that's a reasonable interpretation of what I said is cross- sectional snapshots.
Q. Now Dr. Samet, you participated in developing the plaintiffs' statistical models in this case; correct?
MR. HAMLIN: Objection, Your Honor. Actually that's a vague question, so it's an objection to form.
THE COURT: Okay. Rephrase the question, counsel.
BY MR. GARNICK:
Q. Would it be fair to say that you were a consultant in the development of the model -- of the statistical model in this case?
A. I believe the report describes me as a consultant to Drs. Zeger, Wyant and Miller, as I described earlier.
Q. And you met with the authors of the model a number of times as the model was being developed; correct?
A. Correct.
Q. And in fact you met with the authors of the model on at least ten separate occasions, correct?
A. Possibly around that number, yes.
Q. And you made certain specific recommendations concerning the model; correct?
A. I described recommendations that I made yesterday.
Q. Now were one of the recommendations that you made concerning -- strike that.
Did you provide any recommendation as to which potential confounder the authors of the model should take into account with respect to heart disease or lung cancer?
A. I described my recommendations concerning the models yesterday. Those recommendations do not include confounding.
Q. So you did not make a recommendation as to which covariates, which variables the model should take into account; is that right?
A. I described my contributions to the model and listed them yesterday in my testimony.
Q. I know. But instead of having -- strike that.
And it did not include recommendations concerning identifying which covariates the authors of the model should take into account; correct?
A. That was not a --
Correct.
Q. Okay. What is the difference between attributable risk and population attributable risk?
A. Well I defined attributable risk as the difference between the disease rate in people exposed less the disease rate in those not exposed. For example, I gave the theoretical example two days ago of the study of 100 smokers, 100 non-smokers, and showed how one could subtract the disease rates to obtain the risk for -- the extra risk for the smokers.
Population attributable risk is a different statistic that -- or measure that empidemiologists calculate that describes the percentage of disease in a population attributable to some factor.
Q. Would you agree that a population attributable risk estimate is the theoretical possible number of cases of disease avoidable if the exposure were eliminated?
A. Correct.
Q. Now you did talk about attributable risk, smoke -- smoking-attributable risk in both the United States -- risk of death in both the United States and in the state of Minnesota; correct?
A. I'm sorry?
Q. Did you --
You testified about the number of persons who died, in your view, because of cigarette smoking in the United States in a certain year; isn't that right?
A. I cited an estimate for 1990 from the Morbidity and Mortality Weekly Reports.
Q. All right. Now was that an attributable risk, a population attributable risk, or something else?
A. That's based on a population attributable risk calculation.
Q. Now the population attributable risk measures deaths attributable to cigarette smoking; is that correct?
A. I'm sorry. It's a general measure. It could be used for that purpose.
Q. Okay. But it cannot be used and you didn't use it -- the numbers that you gave earlier today were not deaths attributable to any wrongful conduct of the defendants; were they?
A. The numbers as I described were the estimates of the numbers of deaths attributable to cigarette smoking in the United States. Misconduct or conduct was not mentioned.
Q. And it was not an issue; correct?
A. It was not considered in those estimates.
Q. I'd like to talk about the 1964 Surgeon General's report, which I think you have there, and I'd like to show you plaintiffs' demonstrative Exhibit 30156, or at least a copy of it.
Now it says here on page 20 of the 1964 Surgeon General's report that "Statistical methods cannot establish proof of a causal relationship in an association." Correct?
A. Yes, that's what it says.
Q. And you agree with that; don't you?
A. Yes.
Q. And it also says that "The causal significance of an association is a matter of judgment which goes beyond any statement of statistical probability." Correct?
A. Yes.
Q. And you agree with that, too; correct?
A. Yes, I do.
Q. And I believe your testimony was that the link between smoking and lung cancer satisfied four of these five criteria; correct?
A. That's correct.
Q. And you would agree with me that whether causation is proven, given the fact that four of the five criteria are met, is also a matter of judgment; correct?
A. Well I think in this case I provided the reasons why I did not find specificity to be applicable. I acknowledge fully that there are causes of lung cancer other than smoking. That does not mean that smoking is not a cause of lung cancer.
Q. But it is a matter of judgment. Your conclusion is a matter of judgment; correct?
A. I think this is a judgment that was shared in the '64 Surgeon General's report, and by all other reviewers subsequently, including the International Agency for Research on Cancer and so forth.
Q. Now you talked about the British doctors studies; correct? And I'm going to refer you to the female British doctors survey, which is Trial Exhibit 20203. And this paper was published in 1980; correct?
A. Yes.
Q. It states, "Extensive data are available relating mortality to smoking habits in men, but much less information is available for women. Such data as there are suggest at first sight that women are less susceptible than men to the effects of smoking. This is not intrinsically implausible because the effect of smoking is modified by the environment in which it occurs." Correct?
A. That's what it says, correct.
Q. And does this mean that the effect of smoking is modified -- well let me -- let me strike that.
The use of the word "environment," what does -- what does that mean? What does the author mean when -- what do the authors mean when they say that this is not intrinsically implausible because the effect of smoking is modified by the environment in which it occurs?
MR. HAMLIN: Objection, foundation.
THE COURT: You may answer if you know.
A. I can't say that I know what the authors meant by "environment" here. If there's any lead, it's perhaps in the sentence that follows where they address some of the factors.
Q. Well let's take a look. It says, "We note, for example, that the risk of cancer attributable to cigarette smoking is greatly increased if men are also exposed to asbestos dust, radon or alcohol, while the risk of coronary thrombosis attributable to cigarette smoking is greatly increased if women are also taking oral contraceptives."
Does that help you explain the meaning of the word "environment?"
A. Well here they refer to some factors that are occupational exposures or alcohol, it's a personal exposure.
I would note that what they're commenting on is how in fact people who have asbestos exposure or radon exposure, for example, are placed at greater risk from smoking.
Q. So is what they're saying is that the presence of other risk factors in the environment might affect -- might have an impact on the risks of smoking?
A. That would appear to be what they're suggesting.
Q. Doctor, I'd like to refer you to the Wynder 1950 study that's Trial Exhibit 15911. And this was a case-control study; correct?
A. Case-control study.
Q. And by that it means that you look at one point in time, especially in this study, at patients with lung cancer, and you compare them to patients without lung cancer; correct?
A. Well comparisons were made here between people with lung cancer and people without, correct.
Q. And a case-control study is also called a retrospective study; correct?
A. That's discarded terminology at this point.
Q. In 1998.
A. Correct.
Q. In 1950 this was called a retrospective study?
A. It may have been by some.
Q. I'd like to show you Exhibit -- Plaintiffs' Exhibit 30162.
Now Dr. Wynder in his 1950 study did not provide the risk estimate; did he?
A. No, he provided tables of -- of the data.
Q. He did not provide the risk estimates because as of 1950 the statistical formula used to derive the risk estimates from case-control studies were not yet invented -- was not yet invented; isn't that true?
A. I would have to think about when the odds ratio was first calculated. That's possibly true.
Q. It was not invented till later on in the 1950s; correct?
A. Probably. I -- again, I'd have to check my memory, but probably the first use was Cornfield somewhere in the '50s.
Q. You did not put this --
Would you agree with me that you did not even attempt to put this study into historical context?
MR. HAMLIN: Objection to form, vague.
THE COURT: You may answer.
A. I'm not sure I know what you mean by "historical context" here.
Q. Well you did not portray this study as scientists living at the time -- strike that.
You did not interpret this study as scientists living at the time would have interpreted the study; correct?
MR. HAMLIN: Objection to the form, mischaracterization.
THE COURT: Sustained.
Q. You provided information about Dr. Wynder's study that was not available to scientists living in 1950; isn't that true?
MR. HAMLIN: Objection to form, mischaracterization.
THE COURT: You may answer that.
A. The information on -- from which these calculations were done is included in the textual information here. Wynder does describe the prevalence of smoking of different amounts across the cases and the controls. Was -- was this exact table provided in a 1950 publication? No.
Q. And it could not have been because the formula had not been invented; true?
A. If your account is right and my remembrance is right, the odds ratio was not being calculated then.
Q. And just to clarify a point, in this Wynder study the author compared patients with lung cancer to patients without lung cancer; correct?
A. That's correct.
Q. Now in point of fact, in the same issue of the Journal of the American Medical Association in which the Wynder 1950 study was published, there was another case-control study studying smoking and lung cancer; isn't that true?
A. That's correct. The study of doctor Levin.
Q. Okay. Could you turn to tab six, this is Defense Exhibit AT000853. This is also Plaintiffs' Exhibit PX15923, which I believe was admitted previously.
Now Dr. Levin also found a statistical association between smoking and lung cancer; correct?
A. Correct.
Q. And he concluded the following based on this association, "The data suggest, although they do not establish, a causal relationship between cigarette and pipe smoking and cancer of the lung and lip, respectively. The statistical association may, of course, be due to some other unidentified common factor between these types of smoking and lung and lip cancer." Then he goes on.
That was the state of medical knowledge in 1950; isn't that right?
A. Well these -- this paper,, along with Dr. Wynder's and others', represented some of the first evidence on smoking and lung cancer.
Q. And Dr. Levin didn't believe that his statistical study proved that smoking caused lung cancer, did he?
HAMLIN: Objection, foundation, Your Honor.
THE COURT: No, you may answer that.
A. Well I think the words speak -- speak for themselves. He said the data suggest, although they do not establish the causal relation.
Q. And in 1950 there were still other statistical studies of smoking and lung cancer; correct?
A. Correct.
Q. And one of these was a study by Robert Shrek in Cancer Research; correct?
A. Yes.
Q. And that's tab seven, Plaintiffs' Exhibit PX15939.
And Dr. Shrek found a statistical association as well; correct?
A. Yes.
Q. And this is what he said: "The correlation is definitely statistically significant, but is it biologically significant? A statistical study cannot prove whether there is a cause-and-effect relationship between two factors. At best, the statistical study can provide circumstantial evidence that a correlation is biologically significant."
Dr. Samet, do you agree that at best the statistical study can provide circumstantial evidence that a correlation is biologically significant?
A. I'm sorry, can you repeat the question?
Q. Certainly.
Do you agree that at best, a statistical study can provide circumstantial evidence that a correlation is biologically significant?
A. I'm afraid that the language here, both in saying the "statistical study," a "correlation," and "biologically significant," I have some difficulty interpreting in modern context. I'm not sure I know what the author means.
Q. You don't understand what he was saying?
MR. HAMLIN: Objection, form, asked and answered.
THE COURT: Sustained.
Q. Do you agree that in 1950, medical specialists regarded epidemiology differently from how they regard it today in 1998?
A. I really can't speak to that.
Q. Now you also talked about the Doll and Hill study of 1950; correct?
A. Yes.
Q. And that is tab -- well, that's Plaintiffs' Exhibit PX 16769.
And this was another case-control study; correct?
A. That's correct.
Q. And it was another case-control based on a hospital population; correct?
A. Correct.
Q. Doll and Hill found that smokers with lung cancer inhaled less often than smokers without lung cancer; correct?
MR. HAMLIN: Can I just enter an objection? If counsel is going to read from the document, could he advise the witness of the page so the witness could follow along. I see Dr. Samet sort of haveing trouble finding it.
THE COURT: Advise the witness where you're reading from, counsel.
MR. GARNICK: Well I'm not sure I was reading, but I was getting the conclusion from page 744.
THE COURT: Then rephrase your question, counsel.
Q. Doll and Hill found, based on their study reflected in -- on page 744, that smokers who had lung cancer inhaled less often than smokers without lung cancer; correct?
A. That's -- that's not correct. What Doll and Hill studied was what people said about their inhaling. This is not necessarily a direct reflection on how they inhale. I'm -- I'm not aware of information that validates that if someone says they inhaled more deeply, that in fact one knows that they've taken a deeper breath or a different kind of breath. The data here refers to self- reported inhaling and the depth of that.
Q. So Doll and Hill found that smokers with lung cancer said they inhaled less often than smokers without lung cancer said they inhaled.
A. That's correct.
Q. Now it's recognized that the case-control studies of the 1950s had certain limitations; correct?
A. You'd have to list them so I could see if I agree.
Q. Do you agree that they have any limitations?
A. They -- they have limitations.
Q. Okay. What limitations do the case-control studies of the 1950s have?
A. Well --
MR. HAMLIN: Objection to form, vague.
THE COURT: Sustained.
Q. Well one problem that the case-control studies had, at least those that were based upon hospitalizations, was that it was recognized as early as the 1950s that case-control studies based on hospital populations suffered from a form of selection bias; correct?
A. No, that's a potential bias that hospital-based studies are subject to. Not every study is necessarily subject to that form of bias.
Q. Okay. Well one -- one criticism of the case-control studies in the 1950s was that they may be subject to a form of selection bias that derives from the fact that they were drawn from hospital populations; correct?
MR. HAMLIN: Objection, Your Honor. Counsel is testifying now. If he's got something in a document to show the witness, then I think he should show the witness. It's an objection to form.
THE COURT: Sustained.
Q. Doctor, what is Berkson's bias?
A. Berkson's bias is a bias that you have previously referred to as selection bias.
Q. Is it a certain form of selection bias?
A. Well Berkson originally gave his name to a bias that might arise in conducting research among hospitalized patients as opposed to those -- total population from which those individuals came.
Q. And what is the consequence if a study is subject to Berkson's bias?
A. Well first I think it's important to make clear that not every study is subject to Berkson's bias. It may or may not be. Berkson's bias could increase or decrease a risk depending on how it operated. But again, not every case- control is necessarily subject -- study is necessarily subject to this type of bias.
Q. Were the early 1950 case-control studies subject to Berkson's bias?
A. I don't know of any evidence one way or the other along that point.
Q. Well let me direct your attention to tab nine, Defense Exhibit MD000353. This is an article by Sir Richard Doll.
A. Correct.
Q. A reliable authority in the field?
A. Yes.
Q. And it's published in the Journal of the American Medical Association; correct?
A. Yes.
Q. And that's a respected peer-reviewed journal?
A. Yes.
MR. GARNICK: I move the admission as a learned treatise.
MR. HAMLIN: No objection, Your Honor.
THE COURT: Court will receive MD000353.
BY MR. GARNICK:
Q. Now in this study --
In this paper, Dr. Doll talked about the early 1950 case-control studies; correct?
MR. HAMLIN: Again, Your Honor, if he's going to refer to the paper, I would at least appreciate him giving a page number.
MR. GARNICK: Page one.
THE WITNESS: Okay.
Q. Dr. Doll talked about the early 1950 case-control studies; correct?
A. In this paragraph, yes.
Q. Okay. He said, "These early articles varied greatly in quality, but all shared the characteristic that they were based on information collected retrospectively from patients in hospital and that it was difficult to be sure that the patients with and without cancer (and particularly that control patients suffering from many other diseases) were representative, respectively, of all patients with the disease under study and of the general population of normal persons from whom the patients were drawn."
And he goes on and talks about the need for a cohort study; correct?
A. Correct.
Q. Now do you agree with Dr. Doll's views concerning the early 1950 case- control studies?
A. I think the characterization is not unreasonable. However, again, with regard to this representativeness issue, he does not offer any data to say that if the controls were not representative, that any bias was introduced. He's discussing potential limitations.
Q. Who was Dr. Berkson?
A. My understanding is that Dr. Berkson was a statistician employed at the Mayo Clinic.
Q. Was he a well-regarded statistician?
A. I came --
My career began after his ended, but I'm familiar with his work and assume that it's well-regarded.
Q. Please turn to tab 14, Defense Exhibit AZ000637. Doctor, this is an article by Dr. Berkson in the Proceedings of the Staff Meetings of the Mayo Clinic; correct?
A. Yes.
Q. And are you familiar with the journal Proceedings of the Staff Meetings of the Mayo Clinic?
A. I'm familiar with the journal the Proceedings of the Mayo Clinic, yes.
Q. Is it a well-respected journal?
A. Yes.
MR. GARNICK: I offer the admission as a learned treatise.
MR. HAMLIN: No objection, Your Honor.
THE COURT: That will be received into evidence, AZ000637.
BY MR. GARNICK:
Q. Let me direct your attention to page 331. To get the context, you may want to start at page 330. I'm sorry, that's a chart. 329.
Now in this portion of the journal, Dr. Berkson was talking about the American Cancer Society -- or the Hammond and Horn study. That was one of the first prospective or cohort studies that appeared in the 1950s; correct?
A. Correct. This is not CPS-I, but its predecessor.
Q. Right. And he stated the following: "I have included in the tabulation the death rates for the non-smokers of cigarettes of the American Cancer Society study. These -- these death rates are seen to be strikingly low compared with those of the United States white male population. The non-smokers of cigarettes in the population of the American Cancer Society study are evidently a lot of phenomenally hardy men. One may reasonably conclude from these comparisons that in the response to the call for individuals to enter the survey, men in relatively poor health tended to be excluded, so that the investigated population is selected favorably as respects death rates from all causes, and from specific causes including deaths from cancer.
"Thus the data of the American Cancer Society study taken as a whole exhibit prima facie evidence that they have been subjected to a kind of selection which can produce association in the data studied, such as in fact was found in these data, even if the association does not exist in the primary reference population."
Now Dr. Berkson believed that this early American Cancer Society study was subject to a form of selection bias; is that right?
MR. HAMLIN: Objection to form.
A. Well --
MR. HAMLIN: Foundation.
THE COURT: You may answer that.
A. Dr. Berkson -- I --
I think it's obvious that he was wrong. And let me just take you back through the paragraph you just read to me. You say -- he says, "I have included from the tabulation the death rates of the non-smokers of cigarettes in the study." He then compared those to the white U.S. male population. I would just point out that the U.S. white male rates would include both the non-smokers in the population as well as all the smokers, so I would not expect any comparison of rates of the non-smokers in this study to be comparable to those of the general population rates, which include both smokers and non-smokers. So I --
it seems to me that the basis for his dismissal or arguments about the American Cancer Society study are based on an incorrect interpretation of the rates in the non-smokers compared to the general population rates which include the combination of the lung cancer rates in smokers and non-smokers together.
Q. Did medical specialists in the 1950s and 1960s believe that it was obvious that Dr. Berkson was wrong?
MR. HAMLIN: Objection to form, again counsel is testifying.
THE COURT: It's argumentative. Sustained.
BY MR. GARNICK:
Q. Have you studied whether scientists living in the 1950s and 1960s believed that Dr. Berkson was obviously wrong?
A. I've done no direct study of that, no.
Q. Isn't it true that a number of researchers in the 1950s and 1960s believed that Dr. Berkson was right?
A. I would need to see examples for which specific researchers.
Q. But you don't know one way or the other sitting there today?
A. Again, if you can show me examples, I will try and interpret them.
Q. Now Dr. Berkson also believed that the early American Cancer Society study was at least suspect because it statistically associated smoking with a wide variety of diseases; isn't that right?
A. Show me the statement and again I will --
Q. All right. Page 334, carrying -- the last line, carrying over to page 335.
Dr. Berkson in 1955 stated, "It does not seem unfair to say that, so far as the American Cancer Society study is concerned, the hypothesis of causation of cancer of the lungs by smoking stands or falls with the conclusion that smoking causes also other cancer and also coronary heart disease and also other diseases that either cancer -- than either cancer or coronary heart disease. Indeed the question raised by the findings in the American Cancer Society study of higher death rates among cigarette smokers is not, 'Does cigarette smoking cause cancer of the lungs?' so much as it is, 'What disease does cigarette smoking not cause?"'
Now do you agree that Dr. Berkson believed that the American Cancer Society study was at least suspect because it found a statistical association between smoking and a wide variety of diseases?
MR. HAMLIN: Objection to form, mischaracterization of the statement.
THE COURT: You may answer that.
A. Well I can -- I can only interpret what Dr. Berkson wrote, and I don't know what he believed. He seems troubled by the issue of specificity, and I've discussed that later -- I discussed that issue already.
Q. Would it be fair to say that Dr. Berkson had a different judgment as to whether smoking caused lung cancer?
A. Different from --
Q. Different from yours.
A. Well he's expressing doubt in -- in this text written in 1955, so to the extent that this text written in 1955 differs from my judgment, yes, there's a disagreement.
Q. Now let's go back to the 1964 Surgeon General's report for a moment, page 20, and this is Plaintiffs' Exhibit 30156. I want to talk a little bit about the last criteria, the coherence of the association.
Now I understand that you believe that the coherence criterion has been met with respect to smoking and lung cancer; correct?
A. Yes.
Q. The coherence criterion asks, at least in part, whether causation makes sense, given everything we know about the circumstances; correct?
A. In a sense, yes.
Q. It asks whether it's biologically plausible; correct?
A. Yes.
Q. And it also asks what other biological evidence is available to support a link between -- a causal determination; correct?
A. Correct.
Q. Now I want to focus first on what was in front of the Surgeon General in 1964. And in the 1950s and 1960s, researchers attempted to produce lung tumors in laboratory animals exposed to tobacco spoke; correct?
A. Such studies were undertaken.
Q. Okay. And isn't it a -- (clearing throat) excuse me. Isn't it a fact that during this period of time -- I'm going up to the '64 Surgeon General's report -- not a single study produced such lung cancers in animals exposed to whole smoke?
A. That's my recollection, although I'd have to look at the exact statement of the report to verify that that's what the report says.
Q. Is it fair to say that -- again prior to the 1964 report -- the failure to obtain lung cancers in laboratory animals exposed to whole smoke was one reason why some medical specialists, toxicologists, pathologists, other medical specialists, believed that smoking had not been definitively proven to be a link -- to be a cause of lung cancer?
MR. HAMLIN: Objection to foundation and form.
THE COURT: No, you may answer that.
THE WITNESS: Can you repeat the question?
MR. GARNICK: Can I have the question read back, please.
(Record read by the court reporter.)
A. There may have been such writings. Although I -- I think, again, the need for animal studies when in fact there's so much human evidence available even through '64 questions that approach to doubting a causal connection.
Q. And you're questioning that approach today; correct?
A. Well with literally millions of people dead up to now from lung cancer who have chosen to smoke, I have no reason to ask for proof from animals.
Q. In 1964 --
Before 1964 there were medical specialists that were greatly troubled by the lack of lung cancer in a laboratory animal exposed to cigarette smoke; correct?
MR. HAMLIN: Objection, asked and answered.
THE COURT: Asked and answered. Sustained.
Q. Now you have not conducted a comprehensive review of the literature on long-term animal inhalation tests; have you?
A. No, I've not.
Q. I'd like to refer you to the 1982 Surgeon General's report. I -- I believe you have it there.
A. I do.
Q. I'd like to refer you to page 218. It says, "Attempts to induce significant numbers of bronchogenic carcinoma in laboratory animals were negative in spite of major efforts with several species and strains."
Now do you agree that as of 1982, attempts to induce significant numbers of bronchogenic carcinoma in laboratory animals were negative in spite of major efforts with several species and strains?
A. I agree, and I'm also not surprised because we know there's a great deal of species-to-species variation in responses of the lung to inhaled agents, including carcinogens.
Q. Now of course during this time there were a series of studies in which tar was painted on the backs of laboratory animals; correct?
MR. HAMLIN: Objection to form, vague.
THE COURT: You may answer that.
A. I'm aware that skin-painting studies were done.
Q. The first such skin-painting done was done by Dr. Wynder in 1953; correct?
A. I know that Wynder did skin-painting studies. I don't know if it was the first.
Q. Please turn to tab 10, and that's Defense Exhibit AZ000631. Doctor, I think you already testified that Dr. Wynder was a reliable authority in the field; correct?
A. Yes.
Q. And this article was published in Cancer Research; correct?
A. Yes.
Q. That's a respected peer-review journal; correct?
A. Yes.
MR. GARNICK: I move the admission as a learned treatise.
MR. HAMLIN: No objection, Your Honor.
THE COURT: Court will receive AZ000631.
BY MR. GARNICK:
Q. And Dr. Wynder painted, as -- as we just spoke about, tobacco tar on the backs of laboratory animals; correct?
A. I -- yes. I guess he called it tar, yes.
Q. And he produced a number of tumors as a result; correct?
A. Correct.
Q. And these tumors were cancerous; correct?
A. Correct.
Q. Now this is what he said about his results. "Animal data do not necessarily confirm or deny human data, although historically much of our present understanding of carcinogenesis is based on correlary studies between clinical and laboratory research."
MR. HAMLIN: I'm sorry, Your Honor, can we have a page number?
MR. GARNICK: I'm sorry, yes, 862.
MR. HAMLIN: Thank you.
Q. And he goes on, "The most far-reaching effect of this observation, perhaps, does not lie in its immediate relationship to human carcinogenesis, but, rather, in that the proven susceptibility of animals furnishes us with a working tool which may enable us to identify and isolate the carcinogenic agents within the tars;" correct?
A. That's what it says.
Q. Dr. Wynder believed that his animal experiments -- strike that.
Dr. Wynder believed that the importance of his animal experiments did not lie in proving causation, but rather it lied in providing medical researchers with a tool to enable them to try to identify carcinogenic agents within the tobacco product; correct?
MR. HAMLIN: Objection to form. Counsel is testifying.
THE COURT: You may answer that.
A. Well I can't speak to what Dr. Wynder believed. I mean the sentence speaks for itself. He says, "The most far-reaching effect of this observation, perhaps, does not lie in its immediate relationship to human carcinogenesis," and then he goes on to further elaborate.
Q. Now you agree, doctor, that mouse skin-painting experiments do not necessarily confirm or deny causation, whether smoking causes lung cancer in humans; correct?
A. Well I've given the basis for the determination that smoking causes lung cancer in humans. This piece of information on the mouse skin-painting would be part of the information considered under the coherence criterion.
Q. Well it's part of the information, but would you agree with me that it's -- strike that.
Certainly Dr. Wynder did not believe that very -- it was very important in proving causation; correct?
MR. HAMLIN: Objection to form, characterization.
THE COURT: Sustained.
Q. A fair interpretation of this article is that Dr. Wynder did not believe that animal skin-painting experiments were important in proving causation; correct?
A. Well I don't see anything in this paragraph about causation. Is there another paragraph where he addresses that more directly?
Q. Not that I was thinking about. I mean I think the whole -- I don't think it said that, doctor.
THE COURT: Counsel, I don't think it's appropriate for you to answer the questions.
MR. GARNICK: I apologize.
Q. I'm asking for your interpretation.
A. Well again, I -- I can only interpret what Dr. Wynder has written in this paragraph where he again points out that this might be useful tool, perhaps, and -- and elaborates.
Q. Is it fair to say that the fact that a substance causes cancer in one species does not mean that it causes cancer in another species, --
A. Well --
Q. -- as a general rule?
A. I don't know what you mean by "a general rule." There are certainly examples where something -- an agent causes cancer in one species but not in another.
Q. Is it fair to say that the fact that a substance causes cancer in one species does not necessarily guarantee that it will cause cancer in another species?
A. Well it's -- it's fair to say. I mean I guess we've seen that smoking causes cancer in humans, but it's been difficult to replicate that in animals.
Q. Is it fair to say that the fact that something causes cancer in an animal does not necessarily mean that it will cause cancer in human beings?
A. That could be, again, depending on the agent and how it causes cancer.
Q. Is it fair to say that the fact that a substance may cause cancer on the skin does not necessarily mean that it will cause cancer in the lung?
A. That again might be reasonable, depending on the agent.
Q. You also -- you also testified that medical scientists had identified carcinogens in tobacco smoke; correct?
A. Correct.
Q. I believe you singled out benzpyrene or forms of benzpyrene; is that correct?
A. That was one of the carcinogens that I mentioned.
Q. Now isn't it true that benzpyrene was found in tobacco smoke in the mid- 1950s?
A. Mid-1950s or perhaps earlier, I think. I don't have the exact date in mind.
Q. Perhaps earlier.
Now isn't it also true that at least in the mid- 1950s, scientists believed that the amount of benzpyrene in tobacco smoke was too small to account for the biological activity that they are seeing in their laboratory experiments?
A. I don't have a specific reference for that. If you can direct me to a scientist that said that --
Q. Okay. Let me direct you to tab 13, Defendants' Exhibit AB000299. Now I believe this is a portion of a book by Dr. Wynder; correct?
A. Well I have a chapter from a book. It doesn't list the author on the cover page.
Q. Is the --
Well, is the chapter by Dr. Wynder?
A. Chapter is by Dr. Wynder.
Q. And he is a reliable authority in the field?
A. Yes.
MR. GARNICK: I move the admission as a learned treatise.
MR. HAMLIN: No objection, Your Honor.
THE COURT: Court will receive AB000299.
BY MR. GARNICK:
Q. Now in 1955 Dr. Wynder says this: "As far as the benzpyrene content of cigarette smoke is concerned, its concentration is of more than academic importance. Wright estimates that if benzpyrene is indeed present in cigarette smoke it is present in a concentration of less than one part per million. A similar estimate has been made by Lindsey. Present animal experiments indicate that this amount of benzpyrene is insufficient to induce cancer in the experimental animal. No one can prove whether this amount of benzpyrene might have any activity in man, but on the basis of the animal data this concentration is not likely to have a carci