Tobacco Documents Online

PRESENTATION May 25, 1989 ~ :ist, ci

-2- To reset the stage I will define the mission as trying to contribute to the scientific defense of the industry. The assumption is that we remain vulnerable unless we can break down the near unanimity in the medical research community that cigarette smoking causes several hundred thousand premature deaths a year in the United States alone. While the only certainty is that there will be no quick and easy victory, we can defend ourselves along two general lines. The first is to pound away at the flaws in the case against us, that we have known of for years. Good science demands that an explanation fit all the known facts. Unfortunately for us scientists and physicians do not always live up to this standard. Thus the importance of the other line of defense - establishing non-smoking explanations for the average premature mortality of smokers, and, to the extent possible, for the skewing of their terminal illnesses towards respiratory diseases, particularly lung cancer. d'6x ternal_projects th t involve funding outsiders. Our activities can be dividec ~into internal projects that are our own attempts to originate publishable material Paradoxically our first internal project resulted in the publication by Ernst Wynder and two associates of an article in the Journal of the National Cancer Institute entitled Association of Dietary Fat and Lung Cancer. It reported international correlation data I had begun developing when I was still the Planning officer, but had long been unsuccessful in getting even our friends to N O AN CJ 0 W M CJ N CO

r -3- listen to. We thought we had spotted a strategy to get it published over our own names but Hugh Cullman then suggested the material would have more impact if Dr. Wynder could be persuaded to present it. The result was not nearly as kind to tobacco as it could have been. It was written to give the typical reader the impression smoking and dietary fat contributed equally to accounting for the international variations in male lung cancer, when in truth, dietary fat was a significant independent contributor and smoking was not. However, I'm sure Hugh was right. It was worth the sacrifice to have Wynder be the one to put dietary fat into play in regard to lung cancer. Furthermore his more honest and statistically knowledgeable henchman included (possibly without Dr. Wynder realizing their significance) statements and data sufficient to reveal the truth to the minority of readers who were familiar with the most appropriate statistical methods. The Wynder article is the first item in the folders. When another published study showing a protective relat~onship between vitamin A and lung cancer also revealed data indicating a predictive relationship between fat and lung cancer, I persuaded Dr. Wynder's associate to write a Letter to the Editor pointing out that the study's authors had not given appropriate emphasis to the fat-lung cancer relationship. That's item #2 in your folder. Also in the folder is a Letter from us to the Editor of the International Journal of Epidemiology correcting an ~ ~ w 0 Cj ga° w 0

...... . .. . ., .......a........~«_... _.a.....:..:.t ..~_....t.. . .......... ... .~,.J .:.t+:.,........2.t:i.:...[.t.i.eS .... i.e.....i.. .t.aS{;.... ,t. ..,., ...u. .__...,. . ,1,1,t..,.,~j.... -4- understatement of the relationship of fat to lung cancer in a population study they had published. Two passive smoking Letters have published over our signatures. Incidentally, our submissions are cleared in advance with Alex Holtzman.and whoever is mv current boss Full scale manuscripts are also reviewed by Ken Houghton. The first letter was to the Editor of the American Journal of Public Health, criticizing a passive smoking study it had reported. We claimed that stress could be the underlvincr ~ - _ cause of an association between passive smoking and lung cancer. The editor took a dim view of our position telling us, and I quote, "'it is a long leap from coronary diseases and hypertension to cancer" and "Even in the case of heart disease, studies on its relation to stress have yielded conflicting results". When he could not talk us into withdrawing, he did publish along with the authors' reply. They agreed with our contention that in the current climate the non-smoking marriage partner was stressed by fears the smoking partner was engaging in self-destructive behavior. As the editor had predicted they aimed their fire at our connection of stress with lung cancer. The correspondence is in your folders. (Item 4) In a Letter to the Editor of the American Journal of Epidemiology, criticizing on the same stress basis another passive smoking study, one that got the same silly answer recently printed about cervix cancer, we went beyond passive smoking in our final two paragraphs and defended active

-5- smoking, as well. We included the industry's standard point that their smoking habits are not the only relevant differences between smokers and non-smokers. We also introduced the thought that it is harmfully stressful to the active smoker to be constantly told smoking is self destructive; and the thought that smoking cessation may have a placebo effect. These latter points were original thinking with me, although experience has taught me that what one believes to be original is sometimes unconscious recall of unoriginal ideas. The authors declined to reply to this letter and the editor added a footnote pointing-up their failure to accept that opportunity. Our letter in the folders. With regard to the placebo effect of smoking cessation, I should mention another point we have not even tried to publish, because I have not yet figured out how to document it. It could be part of the real explanation of the benefits smoking cessation appears to confer. Experienced doctors and nurses will tell you that some patients have a to~live than others that is not always stroncer wil]t_ recognizable in interviews. I believe a climate has been created in which i~ixidtjals xjth a strona wi_ 11 to live are more likelv to give up smoking than those whose will to live is weaker. We have also published a Letter in the New England Journal of Medicine, seizing another opportunity to present the same defense for active smoking that appeared in our

-6- American Journal of Epidemiology letter. This time the authors replied that our arguments were "absurd" and justified this dismissal by referencing a 1987 Government publication that reiterated the conventional wisdom about smoking. While we have no way of knowing if our arguments impressed any of the journal's readers, we think that reply was unlikely to affect their appraisal of our points. (item 6) We have published one Letter relating to what I call the "reverse hypothesis", namely that individuals with lung problems may be disproportionately likely to en`oy the satisfactions of smoking; thus producing an association between smokina and luna problems that is not causa] Our best data in support of that hypothesis is quite old so we played our Letter off a recent article that provided much weaker support. The authors' reply jumped all over our weak peg but did not challenge the underlying idea or the strong data that supported it. (Item 7) We have had two Letters rejected, making us five for seven, to date. On free-standing articles we are 0 for 1. The reiected manuscript was a wir3~ rangincr summary of points favorable to the industry. The emphasis in the rejection was that our piece was one-sided. There were also a number of other criticisms, many of which were factually incorrect. The manuscript, the rejection letter and our response to the rejection letter, are all in the folders. (Items 8-10) We were more hopeful of continuing the debate than of getting N N W O W 00 W Ci Ci them to reverse themselves and print the article, but they

-7- declined declined even that opportunity. We intend to rework the broad aaua manuscript and submit it to another iournal - - ~~ We have submitted a second, more narrowly focused free-stander to another journal. It puts forward our version of the comparative importance of dietary fat and tobacco usage in accountin for int rna ional variations in 1unQ cancer. Based on the criticism of the first free-stander, we built a framework of academic detachment around our conclusion. No verdict yet. It is the last item in the folders. One point about the rejection letter on the first free-stander, that I now want to emphasize, is that we saw again that the battleship admirals of medical research are not yet ready to accept psychosocial stresses as significant contributors to mortality. Our contention, that psychosocial factors would turn out to be key determinants of longevity was dismissed by one of the reviewers as "trite pop-psychology", even though we supported it by quoting an epidemiologist from a major university. This leads me smoothly, I hope, to our external funding which is aimedt almost entirel at promoting_they_ role of psvchosociaI sssPg in mortality. This seems by O N 0 W far the most promising area for us at this time. Genetic 0 W determinants of survival are also important but twin studies have forced abandonment of the retrospectively unrealistic ~ G? W hope they would provide full exoneration. ~

-8- Much of our funding is related to establishing the independent importance of the psychosocial factors. Until that is more widely accepted, it will be difficult to gain acceptance for evidence that controlling for such factors reduces or eliminates associations between smoking and health. Another way of looking at the value of researching the psychosocial effects, independent of smoking, is that mortality is a zero sum game. As more is attributable to one factor, less remains to be attributed to others. We have also supported projects that have the possibility of combining psychosocial factors and the conventional "'risk" factors ~n ~ YDlanatory models that will allow the psychosocial effects to reduce or eliminate the associational relationship between smokinQ and_ mortality. One such project that came from Harvard Medical School some years ago, and that we had nothing to do with, showed that controlling for psychologic health, left an insignificant relationship between smoking and subsequent physical health. Admittedly it was a small project and the author.claimed that if it had been larger andlor gone on longer, the results would have been different. Nevertheless, it was an encouraging result, and I should add it did have some CTR support. There is also data from the Framingham study, unfortunately confined to cardiovascular mortality, indicating that controlling for psychosocial variables eliminates the significance of smoking.

-9- To be specifi ~ut--rou de funding we have given general sup rt to Dr._Robert__Ade the University of - __ ~---. . __---------- -- ~~~ Rochester - ad to_Dr- Marvin ' Mt Sinai, two of the pioneering figures in documenting the role of the psychosocial factors. Dr. Ader will soon publish an article demonstrating that just the way animals are handled affects the metastasizing of tumors to the lungs. Metastasizing, as you know, is the most lethal aspect of cancer. Dr. Stein's recent research does not appear to have produced any exciting end products. Recently Bill~a 'orized about $1 million over three years, f,, r Dr.William~ Gutste~f New York Medical College. "L __ ~-_ Dr. Gutstein has already published evidence that simulated stress damages coronary arteries in rats. He has also published indications that cell proliferation may be the first stage in that damage process. He will now work with humans to try to establish that stress can be related to a blood protein involved in this cell proliferation. He wi11 ,. - also investigate whether the stress-cell proliferation connection can be related to lung tissue. -~ Johns Hopkin~_js conducting a continuing study of the ` -.... - health and survival of its classes of 1948 through '64. Much physical and mental health data was taken while the subjects were in school. They have about a thousand subjects, over five times as many as the Harvard study. A number of papers have already been based upon it. We have funded two additional projects. The first is titled

4/ -10- Youthful Predictors of Premature Mortality. The second is tentatively called Temperament and Other Factors Relating to Premature Mortality--. Unfortunately the principal investigator on the first study is known to be a slow publisher even under ordinary circumstances; and two special events have had a retarding effect. One of his co-investigators died in an accident; and he left Johns Hopkins to accept a full professorship at Columbia. I cannot tell when the write-up of this study will be submitted. Hopefully it will be before the end of this year and publication before the end of next year. The second study will differ from the first mainly by incorporating information acquired after the subjects' left school. While it was started a year later than the first, it has a different principal investigator who is not distracted by a new post. It may be ready to be submitted for publication before the end of this year. Our largest funding to date, $5 million of endowment grants over five years, has gone to the Behavioral Medicine ~sQarch Center of Duke University,_headed by Dr. Redford Williams. Dr. Williams is an aggressive champion of the idea that psvchosocial factors have a profound influence on surv _1. He emphasizes that individuals who react to life experiences with hostility and cynicism are prone to premature mortality. He is now in the early stages of a major mortality study that will include a lot more psychosocial data than the Cancer Society studies have

-11- included. There can be no guarantee we will come out clean, but our relationship will encourage them to look beneath the simple associations of smoking with mortality for more fundamental explanations. Their sample is about 4,000 members of through '66. psychological data from their college days is available. The study took more time than expected to get started but the first follow-up questionnaire went out in May of 1987, including a detailed smoking history question. A second questionnaire went out in July of '88; a third will go out this year and follow-up will continue approximately at yearly intervals. I once thought five years after the first follow-up would be sufficient to ascertain the degree, the University of North Carolina classes of 1964 if any, of the study's exonerative potential. I n~ow t_hink it may take nearer ten vears to accumulate enoucth po~t ~ first-response-deaths to make a first estimate f the relative importance of the various possible contributing acto However, we should learn within the next year or two whether the study has identified large average differences between smokers and never smokers; if not, the mortality follow-up is much less likely to be helpful. Last year Bill approved $567,000 over three years to a group headed by Dr. Lindon Eaves of the Genetics Department of the Medical College of Virginia. This is for anothex- mortali dde larcLe- _part of the gample will be twins One reason for our support was that they had acquired a data base, including smoking behavior,

-12- while conducting a passive smoking study for Reynolds. It was plain to me, they were going to use it for a future mortality study of active smoking, even if they had to bear the costs themselves. However, it did not include coffee usage. This was potentially important because two other studies have shown heavy coffee usage has become a risk factor for heart disease. I'm not picking on coffee. I think the explanation is that a certain type, who once would have been a smoker and a heavy coffee drinker has been frightened out of smoking but is still a heavy coffee drinker and still the type that is prone to premature mortality. Another motive for the MCV grant was that its twin content would permit testing the hypothesis that the site and nature of the terminal illness is under stronger genetic control than longevity. Support for this would contribute to the broader proposition that the aggregate determinants of the nature and site of the terminal illness can differ from the determinants of longevity. It would be very helpful to our scientific defense if the ground could be shifted from smoking's strong association with lung cancer to its much more modest association with premature mortality. The weaker association with all-cause mortality is much more in the range of the excess risk that might be attributable instead to the psychosocial factors. Also, the historical trends of all-cause mortality rates are not distorted by the improvements in diagnosing lung cancer --

-13- so you can make legitimate comparisons with smoking trends. We should know in about seven years if this study will be helpful on the mortality question. Possibly another five years will be required to compare the genetic control on mortality with its control over the terminal illness. We also granted Dr. John Thomas of Mehari Medical Oollege, $112,000 to continue a prospective mortality study of black medical students. One particular motive here was to break through Dr. Thomas's conviction that high mortality among blacks is due to a higher prevalence of smoking. Also, I wanted him to put coffee usage into his data base, which he had previously failed to do. A grant for $165,000 that falls outside our mainstream was given to a group headed by Dr. Michael O'She.g-, then a Professor of Pediatrics at Duke. The subject is a study, including smoking habits, of teenagers who had respiratory problems in infancy. It could be helpful to our hypothesis that individuals with lung problems are preferentially likely to obtain satisfaction from smoking. This study is expected to last another year or two. I have saved for last intervention testing, also known as the clinical trial. This is the end game of medical research. In its purest form you randomize a group of subjects into two halves. One half gets the treatment prior research indicates ought to be effective. The other half gets a placebo. The technique has been imperfectly adapted to smoking by giving the treatment group extra advice on giving-up <0,

-14- smoking that the control group does not receive. As you know, smoking has made out rather well so far when the data is analyzed properly. The aggregate reporting to date of all the intervention testing that included smoking cessation advice shows no advantage for the treatment groups in lung cancer or all-cancer mortality, though it does give the treatment groups a minor advantage in all-cause mortality. That cannot be credited entirely to smoking cessation because most of those studies also intervened against blood pressure and cholesterol. Also, we can point out that conventional placebo treatment for the comparison groups has been absent, and additionally raise our point that the trials are biased by the harmful stress placed by the health warnings on those who continue to smoke, and by the placebo effect of smoking cessation. Our funding activity in this area is to support the clinical trial of stress management counseling. Such trials can demonstrate the role of stress in mortality, while avoiding the ethical and practical problems that would arise from a deliberate attempt to add enough additional stress to the lives of subjects to affect their longevity. There have already been results that are far more impressive than the trials th~ncluded smoking cessation advice. A European name 12 as reported 50% mortality rate Grossarth-Maticek reductions for his psychological counseling. He has become a protege of Hans Eysenck. Through Dr Ey~en~~k .c_ - funded some of the follow-up analysis of the Grossarth-Maticek

-15- counseling. The additional publication outlook here is now cloudy because some of the early results were published in minor journals and Dr. Eysenck commented extensively on both the early and the more recent results in Psychology Today, h" s not a tru ademic journal. M~er Friedma of Type A fame, has given what he calls his Type A behavior modification counseling to heart attack survivors, achieving nearly a 50% reduction in subsequent mortality. That project was to be his last. However, we talked him into beginning again, this time with individuals who have not yet had a heart attack. Bill has approved approximately a $5,000,000 commitment to cover the first five years of a study planned for 13 years. Success with subjects who have not already had heart attacks will confirm that the counseling is effective among individuals who have not been frightened into good behavior by their first attack. Also, because we can expect a much_ higher proportion of cancer in this group than in heart attack survivors, the ro ect will provide a chance to docuonclusivelv the role of stress in cancer, __ I think the crying need now is to start another primary prevention study. I believe success here has tremendous potential for us, particularly after the indifferent results of the primary prevention studies employing advice about diet, smoking and exercise. However, the battleship admirals will demand independent replication and government funding is very doubtful because NIH had its fingers burned

-16- by MRFIT. We can't fund another of these large projects unless we are prepared to lift the annual outlays, in my budget, above the present $2 million level. Funding another intervention study could take us to $3 million a year or even $4 million. On that grasping note this presentation concludes. N 0 N C.~ 0 W ~ W ~