Tobacco Documents Online

~C' 7W AA"'r'" LAj ' dA`P" I'`PP' VbIJ y,, z CIGARETTE SMOKING AND CORONARY HEART DISF.ASE : AN OUTSTANDIN(; MEDICAL FALLACY OF OUR ERA ? 13 v1%- c~-~- "'rc~m r-4ti D14 `~ ~igl C-0*%fft r-C~y- G"_)- ~r.As_C .

-la- INTRODUCTION "I beseech you, in the bowels of Christ, think it possible you may be mistaken - Oliver Cromwell (1650) to the Church Assembly of Scotland - Throughout history, sci~entific knowledge and progress have often been held captive by conventionali wi~sdom„ religiousprejudices, and social judgments. Examples abound. The know-nothings refused tollook through Calileo's telescope, andwhen they did refused to believe what they saw. Copernicus had to overcome established religious beliefs in presenting,his heliocentric theory, with planets goingaround the sun at the center. Darwin was faced with fervent religious and!sociali pressures in propounding his theory of evolution. The history of medicine, too,, contains many examples over the centuries evangelical fervor for both etiological andi therapeutic theories that were of later shown to be wrong. The names of Harvey, Pais.teur, Lamarck„ Freud, among many others come to mindlas examples of men of medicine who underwent long a!nd' painful struggles to overcome emotionally held views of the establishment of their times. Unfortunately, this same syndrome is being repeated today. The current controversy over the nature of the relationship between ci Qarette smak-inQ, nd cyojSr= he,= .4i~as is a present day example of the extent to whichipreconceived ideas, passionately held and advanced, have impeded the search, for scientific truth The nature of the two elements involved, cigarette smoking and coronary rj C) heart disease, have exacerbated the situation. This: is because heart disiease}a is'the greatest killer of all of'middle-aged and elderly populations in industrialized nations, andibecause seldom has any target evoked so many strident voices and so much mindless emotion as tobacco smoking.

There is hardly a segment in our society that does not firmly believe that cigarette smoking contributes or causes premature death from coronary heart disease (CHD). Most recently, the Surgeon G'enerali declared that: "...cigarette smoking,is a major cause of coronary heart disease for both men. an& women,. So what is the probl~em? Simply put, the problem is thab the Surgeon, General's declaration is without valid scientific foundati~on. A causal relationship between cigarette smoking and'CHD has not been establishedi. Except for the first Surgeon General's Report (1964) on CHD, the numerous: annual andibiennial subsequent reports on this subject are neither objjective nor authoritative. These Public Health Service reports on GHD are flawed!, inherently biased, selective in its prssientation ofavail~able data by omitting or impugning contradictory evidence, and fitting the interpretations of studies into alpreconceived and desirable conclusion. From these reports and from the declarations of the American Heart Association and others, a series of myths have evolvedinto what has now become the conventional wisdom. This article will examine these myths and claims in some detail. -lb-

HISTORICAL PERS'PE'CTIV'6 To understand how the present indictment of smoking and CHWcame into being,, the matter must be examined in historical perspective. The first major U.S. Public Health Service review ofthe relationship between smoking and heart disease was conductedby the Surgeon General's Ad!visory Cmmmittee on Smoking and Health in 1964. A special' feature of this C'ommittee wasin its composition.A11 the mebers were non-Qovernmental scientists, chosen by the Public Health Service with the tacit assent of the tobacco:industry and other interested groups. The purpose was to insure as f'ar as possible appointees uneommitted one way or another to the question oftobacco:and health. After examining the available evidence, th,i&Advisory Committee did 221 find a causal relationship between cigarette smokingand CHD: The Committee concluded that: "...ma1e smokers have a higher death rate from coronary artery disease than nonsmokingmales, but it is not clear that the association has causal signif'icance."' In other words, the Committee made.it explicit that this association did not automatically imply causation. Why not? Because the studies comparingsmoking and nonsmoking populations did not establish that they were ib ini-tia similiar except for the smokinghabit, a basic scientific requirement. In the words of the Committee: "The question of whether they are, in fact, similar except for smoking is, ofcourse, basic to the problem of whether cigarette smoking actually promotes the development ofcoronary disease or whether it is closely associated with some factor or factors which promote the development of coronary disease." In presenting the Committee's Report, Dr. l,uther Terry, the then Surgeon General, speaking of CHD on Jan 29', 1964 stated: "..the committee was unable to reach a firm conclusion as to the role smoking plays in causing,or precipitating,a death from this d!isease...we have no real clues as to what it is in tobacco that influences coronary artery disease, ifindeed it does." -2-

-3 - Nevertheless, anti-smokingforces, completely ignoringthe finding of the 1964 Report, immediiately launched a two-tiered campaign to overturn the non-causal findings of the SnrgaonlGeneral's Committee On one level, spearheadad by the American Heart Association, an intensive propaganda effort was put into action to imprint on the minds of everyone that cigarette smoking was a:culprit in the development of heart disease. All forms of the media,,TV, radio, newsprint, etc. were saturated with its message. Resolutions against cigarette smoking were sent to medical socities throughout the country urging their adoption„ and action committees (ASH, GASP, etc.) were formed to insure abroader based support for their anti-smoking views. So effect,ive was this campaign even in its early stages that the CHD findings of the 1964 Committee didn't even have a chance to permeate medical circles, let alone the public. The findings were so quickly overwhelmed by the media campaign ofthe anti-smoking forces that few persons even remember that the Surgeon General's 1964 Committee didinot find a causal relationship between smoking and'CHD. On the second level, a small coterie of anti-smoking epidemioliogists, led by Dr. Jeremiah Stamler of NorwesterniUTniversity, began a lobbying effort to establilsh their anti-smoking views:onithe Suregon General, the Public Health Service, and especially on the leadershiplof the National Heart Institute. This was done in recognition that the Public Health Service was charged by Congess with providing all future reviews of smoking and health. No longer would there be reviews by non-governmental uneommitted scientists such as the Surgeon General"s Advisory Committee of 11964. Thus,: the stage was set for a scieMtific reversal of the "non-causal" CHD findings of the 1964!Committee. From 1967 to the present, all the many reports of the Suregon General were prepared by governmental bureaucrats, staff memebrs of the Public Health,Service, with the assistance and advice from selected epidemiologis.ts.who were committed to the view that smoking was harmful to healith.

-4- So, beginning with the 1967 SiurgeoniCeneral's Report, there began a creeping escalation of the claimsof the Public Health Service. From "...it is not clear that the association has.causal signifitance" in the 1964 Committee's Report,, to "...cigaxette smoking is a significant risk factor contributing to the development of coronsry heart disease" in the 1975 Report, to an outright statement in the 1981Report that:~ "Cigarette smoking is a major cause of coronary heart diseasie in the Unyited States for both men and women." To be sure, since 1964 there have beenia mountain of studies dealingwith smoking and CHD and more than a dozen Surgeon Ceneral reports on this subject. But the ca!use/effzct linkage which many of my colleagues believe they 1u must be there, still has not been found. These new studies have failed to produce the "'har&evidence"of a breakthrouigh inithe sense of a proven establishment of the mechanismi(s) for a causal effect oflsmoking on CHD. This is admitted by our government inia publication from the National Heart, Lung, and Blood institutein 1982, stating:~ "A funidamenta!1gapinithe accumulated knowledge is that the exact mechanisms through which cigarette smoking increases the riskofCHUarest~i11 largely u~nknown. " With the "hard evidence" elusively undeterminedi, th Skirgeon General and his Public Health Service have perforced to rely on "soft evidence" to substantiate their claims. An examination of the nature and!validity of this "soft evidence"' will now form the major part of this article.

-5- THE CASE AGAINST CIGARETTE SMOKING The connection between cigarette smokingand heart disease propounded by our government is essentially epidemiological (statistical~). Epidemiological evidence is "soft evidence" since it relies on infencag from supposedly "'common-siense" interpretationof its statistical observations.The major declarations, to be found in the report~w oftheSurgeo~n GeneraL"s Public Health Service which have led to the claim that c4grette smoking is a'"major cause of coronary heart disease" are best exemplified in an editorial from the Framingham Heart Study -the government's own intensive long-term investigation - which stated: "Causal inferences are supported by the fact that the relationship;is strong, consistent, demonstrated prospectively, independent oflassociated risk factors, a#4d can be explained by the knowrleflfects~ofcigarette smoking on the cardiovascular apparatus. Finally, it has been shown that those who quit smoking have only halflthe risk of those who continue to smoke. Epidem4~oLogic data,...tended to!indicate an independent, transient, noncumulative, reversible, triggering effect of cigarette smoking." From such a dieclaration by, our government and!others, aided by an intensive propaganda campaign, and a blind "hype" by the media, a causal connection between cigarette smoking and CHD has become the conventional M wisdom. Within the restrictedispace of this article, let us examine the tru~th ofthese declarations.

THE AS'SpC'.IATICNi The case against smoking begins with the association of smokers with excess d~eaths amd disabilities from CHD. It is the primary reason why cigarette smoking is said to be implicated in the causation of CHD, andistems from the results.of studies in the United States and parts of'western Europe:. These statistical associations are not in question. What is in question, however, is whether or not these associations by themselves automatically imply a cause/effect relationship, for, it is aluniversal dictum clearly stated in the Surgeon Genezal's 19719 Report that "correlation is not synonymous with causation." When considering the issue of cause and effect, textbooks of epidemiology and biometry caution investigators first to regard associations or correlations as leeds or markers warranting further inquiry for sumIDviS evidezebefore reaching a firm conclusion. And, as noted previously,, this was the position taken by the Surgeon General's 1964 Committee when it notedl in its summary that they found'an association between smokers and CHID mortality "but it is not clear that the association has causal significance."' The reason given was that whether or not cigarette smoking actually promotes CHD is dependent on whether or not the two groups, smokers and nonsmokers, are alike to:begin with in all characteristics likely to affect the rate of CHD. To put it simply, if smokers and nonsmokers resemble each other at entry to aistudy.for all known relevant factors, this strengthens the concllusion that differences found in eventual disease outcomes are due to the act of smoking. On the other hand, if important baseline differences ("confounding factors"') between smokers and nonsmokers are observed, these,differences and no thg ac ojsm ki;ns,may be responsible for the greater CHD rates of smokers relative to nonsmokers. Thus, the prime determinant may be the nature of the s o r rather than the smokinst. -6-

-7- Let us clarify. this by an example. Suppose in an inves.t.igationi, it is found that smokers are found to have a higher rate of CHD mortality than nonsmokers. Suppose it is also found!thst the a particular characteristic found among subjects, let us say, Type A personaliity trait, is associated with a greater prevalence of CHD mortality, and'inladdition is more prevalent among, smokers than ininonsmokers. It is therefore possible that the higher rate of' CHD among smokers is not the result of the act of smoking but due to the fact that smokers tend to have among their members more persons with Type A personality. Thus, the association of smokers with CHD mortaliity may be indirect (non-causal) an&not direct (causal). Thecrucial query, then, are there important differences between smokers and nonsmokers apart from the smoking habit? Of this, there can be no doubt, for the literature is replete with a myriad of important differences between these sellf-selected groups. They differ in a number of morphological, physiological, biochemical and psychosocial characteristics, in social and economic status, in,occupation, education, marital status, personality traits,, etc., most.of theserel~ated to CHD. Have these potentially "confounding" factors been,routinely taken into considieration in those studies which demonstrated primary associations between cigarette smokingand CHD? Not really. To beginiwith, some of the very large. early stud!ies provided virtually no additional descriptive information on. their subjects beyond!age and smoking habits, i.e. Doll and Hill's British doctors, the U.S. VeteransiStudy, and the Canadian Veterans Survey. Since no data were available in.thesestudies to determine whether or not smokers and nonsmokers, were significantly different in other features apart from the smoking habit, inferencesoflcausality from the a~ssociations found in these investigations remain uncertain. ©therstudies such as the FtaminghamiReart Study, Hammond"s American Men and Wmmen,and the Tecumseh, MichigAn survey collected'hardly more than a

handful of additional characteristics descriptive oftheir subjects andi frequently attempts were made to remove their potential bias by complicated' statistical manipulations In this connection, Professor Alvan R.Feinstein of Yale Medical SchooLL pointed out: "Instead'of searching diligently for better scientif'ic methods to identify and' remove sources of major bias, the investigators have often*become diverted'by the allure of mathematical manipulations (and since) mathematical adjustments cannot identify important variables that have been omitted,,the adjjustment process may yield calculations that embellish the fundamental bias without affecting it." Since as many as 246 so-called coronary "risk factors" have been noted, the task of id'entifying the major sources of bias and the results suitably adjusted, remains a£ormidabl!e.but necessary task. More importantly, perhaps, is the fact that an association or correlation of smokers with excess CHD deaths is not a universal phenomenon. As wi11" be detailed later, there ispa association between smoking and CHD to be found in large parts of the world, an inconsistency which defies definitive expsanations. Thus, the very heart of the case.against cigarette smoking,, the "essociation", is not that solid or impelling. It cannot be denied that it exists in,some parts of the world, but its meaning in terms,ofcausation, requires definitive supportive evidence and removal of sources,of bias in order toldetermine whether the "association" is indirectly due to the nature ofthe smoker (non-causal) or to the act of smoking (causal). Until this is forthcoming, inferences of causality from the "associations" remain uncertaim.,