Tobacco Documents Online

Z A BRIEF REVIEW of the ::I SMOKING-LUNG CANCER THEORY An Address by Clarence Cook Little, Sc.D. Scienii fic Direcsor Tobacco Industry Research Cornmittee Delivered before the blonrot County Cancer A.sociatioa, Inc_, Rocheater, New York April 28, 1960 I~ CTR• 1303 C If~ ~r4 ~'"~t`I 0 0 °~ ~ ~ '.~

ildditional copiea of this paper may be obtained upon request from: TO<ACCO IHDUSTRT RuuRCx Co~rrvTTi= 1SO East FOrty-secoOd Street New York 17, New York ~ ~j~- 1304 11 ~T~~~ N~•I ~¢.~ ~:~~'~~

0 A BRIEF REVIEW of the SMOKING-LUNG CANCER THEORY by Cuas:.NCe CooK Lrrns, Sc.D. Scitnti fic Director Tobacco Industry Rwearch Comrr.ittee For the past six years my colleagues on the Scientific Advisory Board to the Tobacco Industry Research Committee and I have been rr-sponsible for the placing of about =3,500,000 in it-search grants with independent scientists undertaking to study problems, both specific and basic, relating to questions raised about tobacco smoking and health. In the course of this activity, we have been exposed to and have had to consider an ever-growing mass of research reports relating par- ticularly to the complex problem of lung cancer and all possible suspect factors in its etiology. Some of these reports are directed at tobicco, principally cigarette smoking. In addition, there has been an attempt to involve tobacco in a number of other human ailments and causes of death. My object here today is not to attempt to convince any of you con- cerning the relative merits of claims or counter-claims about the effects of tobacco use on human health, particularly in relation to lung cancer. It is rather to discuss fundamental aspects of the cigarette-lung cancer charges, as published or stated from various platforms, in order to show you why many of us believe that the problems of lung cancer causation are not solved, and why we believe medical research will be rendered a dis- service, if the case is considered decided or closed, as some insist. There are three main lines along which data on the relation of to- baccm use, particularly cigarette smoking, to lung cancer have been col- lected and presented. They are statistical, pathological, and animal ezperimentation. Statistical Data Inconclusive and Inexact Statistical data were the first to be presented_ 'Ibey still lead numeri- cally and form the ovenvhelming majority of published material that besrs on the general subject of tobacco and health. The interest that chiefly stimulxted the intensive study of the whole avbject was itself a statistical one: The increuing number of recorded deaths from lung cancer during the past 30 or 40 years, particularly among white males. There was and still is wide divergence of opinion among statisticiatu 3 I CTR. 1305 I C ~~~~'~ ~~I ~~ ~~~~ 3 ~~~~

I and other scientists concerning what proportion of ~he recorded increase in lung cancer mortality is rcal and what part is due to improved detec- tion, diagnosis, aging of the population, and procedui +l factors applied in the classification and reporting of causes of death. -We can be sure of one important fact: The numerical increase in reported lung cancer deaths is not an exact measu a of change in the real attack rate. \Ve tannot be sure how much, if any, f' e actual rate of lung cancer mortality has been increasing. We do not, however, need to concern ourstlves toa much on this point, other than being sure we recognize its importance. It will, I am sure, require and rer.ive continuing attention. Tbe involvement of tobacco use in mortality i• based on the statis- tical association of smoking habits with the rate of death. Such an associa- tion has been reported for 'exceuive' cigarette smoking by a number of epidemiological studies. This is the keystone in the arch of accusations against smoking by those who call it "the major cause' of lung cancer and a large number of other diseases. Cigarettes Blamed for Centuries-Old Diseases Causes of death that have been related by statistical studies to ex- cessive cigarette smoking include general mortality; lung cancer, laryngeal cancer, oral cancer, esophageal cancer, stomach cancer, bladder can- eer, prostate cancer, malignant lympbomas, and all cancer combined; bronchitis, emphysema, tuberculosis and other respiratory diseases; cor- onary thrombosis, angina pectoris, coronary artery disease, general arter- iosclerosis, hypertension, cerebral vascular disease, peripheral vasculat diseases, and general mortality from diseases of the heart and circulation; peptic and duodenal ulcer, and cirrhosis of the liver. You may be in- terested to hear that other recent medical literature also relates cigarette smoking to eye troubles; nose and ear diseases; miscarriages, sterility, and other reproducti•re disturbances; and a host of alleged actions on the nervous, endocrine and digestive systems of man. On the face of it, one's credulity has to be stni.rxd to believe that these diseases, most of whicb are as old as the human race, are now being caused by the cigarette which has beeome a widespread custom in the past half-centvry. This impressive list of diverse and unrelated causa of death thus brings us to the first major area of disagreement among the statistical interpreters of statistics themselves. One point of view holds that excessive cigarette smoking introduces one or more specific carcinogenic, or cancer-producing, substances. But if this is true, how does one account for the association of cigarette smok- ing with other diseases? To answer this, another point of view holds that excessive smoking produces that old medical umbre]la= general debility' 4 I CTR, 1306 0 0 ~~ __ +J

-un,er which so many variable and unexplained af$ictions have crouched :or years, protected from the chilling rain of scientific definition and analysis. No Nurnan Carcinogens Found in Tob.cco Smoke The first point of view is disturbed by the failure of all efforts to detect in tobacco smoke any substance known to be carcinogenic to tan. The second point of view is disturbed by many things among which olle may Fe mentioned, as follows: General debility Is not charactesistic of over-weight and over-energetic people. Cardiooascular diaease, which incidentally has been charged statistically with the highest nurnericd e:- ccss of deaths among excessive cigarette smokers, is associated clinically with those who are advised to "lose weight" and to "slow down" because of their rapid tempo of living. There are many other inconsistencies between the actual findings and the kinds of interpretation of statistical results reported by various workers. Some of these will be discussed briefly for, as I have said, the statisti- cal data are the keystone in the often publicized arch of material accepted by enthusiasts as `proof' of the charges that cigarette smoking is causing many ills. Any habit that is believed to be the major cause of lung cancer and that is blamed as a significant factor in a greatly increased incidence of the wide range of other diseases would indeed be a dragon of such magnitude that Saint Ceorge himself might hesitate to tackle it. However, let us consider certain of the unresolved conflicts of opinion which exist among statisticians themselves. Con/licting Findings On 1nhalers, Non-InhnIeri One of these involves the comparison between inhalers and non- inhalen of cigarette smoke. Two British investicators found in a reasonably extensive population of doctors no diEerence in lung cancer incidence of any gzcat magnitude between the hvo categer'es. tVhat difference they did End w-as in favor of inhalers. Thcy seem to have less lung cancer than the non-inhalers. A few American investigators believe that there is a greater propor- tional incidenee of lung cancer amoag inhalers, but have not produced data to establish this theory. Obviously it seems that someone is u-rong. They can't both be right. Since replies from the Brih'sh physicians eonccming their habits of inhalation were collected in ezactly the same way and by the same peoplc who also collected 'mformation as to the number of cigarettes smoked, as to the duration of the smoking habit, and as to the periods of interruption of the habit, the decrce of accuracy of this information needs verification by further studies. This is all the more essential since a high degree of 5 . t

quantitatice and comparative significance has been given to these figures by many statisticians and propagandists. 1\oother difference in observational results it, found in the effect which continued smoking of cigarettes is supposed to have. - Some data show an unexplained but statistically significant protective effect of cigar and pipe smoking, if these types of tobacco use are super- imposed on or added to cigarette smoking, regardless of whether the Iatter be light, medium or heavy. 'The variation in the claimed 'excess risk' of lung cancer among cig- arette smokers as calculated from statistical studies is another example of conftict Contrary to many generalizations made about these reports, they do not all show the same thing. Depending upon which study you examine, you will find that the relative risk of lung cancer among cigarette smokers may be no different than that of non-smokers, may be fractionally higher than that of non-smokers, may be three times, four times, five times, six times, n.ine times-and so on up to 36 times the risk of non-smokers. Sarne Stotiiticat Evidence - Lots of Guesiea Similarly, there is a wide difference of opinion concerning the rela- tive quantitative role of cigarette smoking in the etiology of lung cancer, even among those who are crystallized and missionary in their behef in the guilt of smoking. These estimates, all based on statistical evidence as to how much of the lung cancer incidence can be blamed on smoking, vary from as high as 90 percent or almost totality, downward by degrees to less than 10 per- cent, or almost nothing. Since the same statistical evidence is available to all contestants, it is evident that such guesses are more a reflection of the degree of interpretive enthusiasm which each individual possesses than they are of scientific significance. All the guesses can't be right, and if all but one are wrong, who is to say now which one is right? 7bese selected essa_mmples of conflicting data and interpretations will, I believe, be helpful in establishing a balaneed and comprehensive basis for present evaluation of many aspects of statistical investigation. So will certain others next to be considered. One of these is the relatively higher mortality among males from all the con:^san respiratory diseases. Some statisticians and epidemiologists accept this as a const5tutional and.ior genetic difference between the sexes. The largely unbalanced X chroniesoaze in males would provide an increased opportunity for dircct expression of certain genes which may influence susceptibility. The balanced X chromosomes of the female would decrease this opportunity. Others believe that no real difference exists between the sexes and that when women have smoked as long and as intensively as men they will show an equal mortality from lung canccr. 6 I CTR. 1308 - ~ _~ e~.4 f'''Q ~. C ~Y~'N 111H 00 4 in L,.~..40 134

Two lines of statistical evidence are strongly in favor of the existence . of a real sex difference in susceptibility. One is direct and consists of the fact that the lung cancer mortality difference between the sexes has betn widtning in recent years instead of diminishing. On the theory that rela- tiveiy more and more women are completing the hypothesized cancer- latency period of 20 to 30 or more years of smoking, the gap should narrow. The second line of evidence, which is indirect, is the persistently greater susceptibility of men to other respiratory diseases where exposure of both sexes to infection and other causes is more nearly equaL I Unexplained Di;(Jerencu in Lung Cancer Rores • Further, there are unexplained differences in cigarette smoking-lung cancer calculations as one goes from country to country, from region to region, or from city to city in any given country where such data have been collected. For instance, people in the United States, who smoke 30 percent more cigarettes per capita than the British, have a lung cancer death rate less than half that of England and Wales. The American lung cancer death rate is about the same as that in Denmark and Switzerland, where per capita cigarette smoking is about one-half that in the U. S. In this eauntry, governmental studies of lung c2ncer incideace in several major cities show wide variations that cannot be explained by di$erences, if any, in cigarette smoking usage. To attempt to rrconcile such dif?'erences, which disturb them, some statisticians and their assistants have zealously collected cigarette butts and measured them carefully. ney have reported sborter butts in Britain than in the United States and feel that this means that our British cousin obtains especially harmful substances from the last part of his c3garette. They thus explain the statistical discrepancy to their own satisfaction. There still revnains the open question, however, as to why migrants from Britain to South Africa and New Zealand have a higher lung cancer incidenct than native whites in those countries, even though cigarrtte smoking is as high or higher among the native-bom. Two diaerent and independent reports have shown significant differences in lung cancer- rates between the native-born white men and those who immigrated, and these differences are not related to the rate of smoking among these groups. Both authors do relate these differences in lung cancer mortality to other environmental e_.posures sustained in the country of origin.. Turning to animal e_cperimentation, we &nd an equal or even more finely applied degree of quantitative interpretation. The various relative amounts of residue distilled from tobacco smoke obtained frorn machine combustion of thousands of cigirettes and then painted on the shaved backs of mice have been used to build up a whole superstructure of es- tirnated degrees of risk of lung cancer in human beings. The animal ra 7 I CIR- 1309 t r'r;r"`ti ~ ~=~ 3_1

( sults were obtained by use of concentrated chemical uterial in which no substance known to be carcinogenic to man has 'ueen demonstrated. Furthermore, no substances have been detected in this smoke residue which are present in amounts, either singly orin coms.•;nation, to account for biological activity on the skins of experimental animals. The tnatcrial has been obtained by methods of smoking which diSer in many ways from the process of human smoking. The material j.as been tested over the entire lifetime of animals often known to be ge :: ically susc-eptible to cancer. Even so, only a minority of the animals c. . -,eloped skin cancer. Th;s type of animal test for relative pathogenic pc,tenVtlit; frequently has been shown to give oositive neoplastic reactions to substances that have been used by and applied to man with impunity for ;-ean. !n?,atation Tuta with Animalr Negative Uniformly negative results have been reported from smoke inhala- tion experiments with animals. Those who hold to the tobacco guilt hy- pothesis ignore or soft pedal this evidence, which involves the use of actual cigarette smoke itself as the challenging agent and the animal lung itself as the target organ. Yet they accept an.i publicize as supporting evidence the results of applying a machine-made cone ntrate of smoke to the shaved skin of animak The scientific status of such selective and discriminative emphasis is open to qnestion. There are increasing numbers of experimental investigators who are doing just that, and who are openly skeptical as to techniques and interpretation. . There is also marked disagreement as to the nature and significance of the metaplastic pulmonary lesions which a minority of pathologists call carcinoma in ritu. T'he majority of pathologists who have been in- terested in lung pathology do not so define these metaplastic changes. Furthezmore, these changes are freqtztntly observed in areas of the res- pi.ntory system where little or no carcinoma is observed. This shows that they arc not definitive of cancer. Second, they ocrur frequently in cases of pneumonia among adult smokers and non-smokers alike and in young children in whom bronchogenic carcinoma is a rarity. They are, thezefort, not specific for smoking. There aze further disagreements among statisticians and pathologists as to the relation of adenocarcinoma of the lung to the smoking problem. Some believe that adenocarcinonia is associated; some that it is not. Some believe that it is possible to identify hvo clearly distinct histological groups of lung malignancies; others believe that this is not possible. This melange of con8icting interpretation is again selectively screened by ad- vocates of the tobacco guilt hypothesis and the selected favorable ma- terial is used as supporting evidence; the 'unfavorable" or dissenting mults are ignored. There is also a very definite difference of opinion as to the major or a I . ~ ~ rR. 1310 ~ a~- ~,b 4, .~~

sninor nature of any hypothisized role of tobacco in the etiology of lung cancer. A• before stated, some believe that there is a direct, major car- cinogenic efie_t of tobacco. Others, noting the large number of smoke irihalation tests in animals, none of which, in spite of massive exposure, resulted in produc:ing lung cancer, are convinced that if tobacco has any role it is a very minor one. T'he fact that lung malignancies have been observed in animals following inhalation of oil and gasoline combustion products shows that the animal lung as such can be a site of lung cancer. The consistently higher statistical incidence of human lung cancer in urban populations compared with rural is considered as supporting evi- dence of the need for further study of air pollutants as a potential factor of importance. I Latency Pe"iod A Statistical Convenience Another area et statistical disageement is the duration of the hypo- thesized Latent period before the carcinogenic changes attributed to smok- ing are supposed to occur. Estimates vary over a range of from 10 or 15 to 30 or 40 years. Since 90 percent of heavy smokers at age SO or over do not have lung cancer, and since many individuals report that they started smoking at age 12., there seems to be no valid reason why the latent period should not be extended from early adolescence throughout the life span. This is a great statistical cnnvenience, for it allows the sen- sational but meaningless statement attn-buted to one of surgery's most enthusiastic advocates of tobacco guilt: `Every cigarette smoker will die of lung cancer if he doesn't die of something else first." There are differences also in defining what may properly be classed as -heavy," modente' or "light" smoking. The "pack a day" criterion, as recorded from the recollection of smokers or their families and friends, is an approximate estimate and not a scientific measurement. It is obvious that an individual who smokes 20 cigarettes per day to the shortest con- venient butt length may draw into his mouth an amount of smoke equal to that of a smoker who daily smokes 60 cigarettes down to one-third of their length. The only thing that smokers of a given number of packs per d2y have in common-granted that the figures are accurate-is that they lighted the same number of cigarettes. This brings up another device used in sorrme of these studies, that of translating quantities of cigar and pipe tobacco used into terms of packs of cigarettes. This type of non-scientific conjectare merely helps to complicate the differences reported in all the statistical studies between the effects of cigarrttes, pipes and cigarsl If one is to understand scientifically the relative activity and phy- siological importance of the numerous factors involved in the etiology of lun¢ cancer, he is faced with one of the most difficult areas of investiga- tion in the whole field of cancer research. 9 I CjR. 1311 I C

1 ZZ= I Other Suspect Factors Exist Besides Smoking Factors other than smoking per se begin to suggest themselves as possible suspects to be considered. immediately we End many interesting problems worthy of continued attention and study. A socio-economic difference has been obsetved in lung cancer in- cidence. The lower the economic level, the higher the rate of lung cancer. Malnutrition and dietary deficiency, already apparently operative to sorne degree in the etiology of cancer of the buccal cavity and tongue, comprue one matter to be studied. Stress and strrin as an unbalancing and continuing element in the life of the individual form another possible element to be kept in mind. Another is the possible role of previous or current respiratory in- .fectious processes with their accompanying lesions, and d'uturbancrs or disruption of the continuity of function of certainareas of the lung. There is an amazingly s}-rrtmetrical divergence between the curve of recorded increase of lung cancer mortality and the curve of decreasing death rate from respiratory infections. Thic rnay- or may not prove to be a coinci- dence. It is certainly worthy of further study. This is especially empha- sized by the increase in the lungs of patients with influenza and other lung diseases of the kind of lesions believed by certain pathologists to be potentinl precursors of lung eancer. It can be seen that these are all complex matters for definition, measurement and analysis. The human individual is complex, however, and so are his motivations and reactions, including his habits. We cannot change these facts. They are an unavoidable part of the problem Cloud Mind Attitude is Unfortunate Those who accept the reported statistical association between smok- ing and lung cancer as proof of a cause and effect relationship often be- come irked and even abusive when others require experimental evidence and more carefully-controlled clinical and statistical studies. This at- titude seems to be unfortunate from at least two points of view. First, it accepts a superficial standard of scientific proof which can- not safely be applied in determining etiology in such a complex and delicate process as carcinogenesis. Second, the dogmatic attitude toward smoking produces a feeling of definiteness and conclusiveness which an individual cannot safely apply to his own case even though the theory is advanced with much the same degree of assurance that quite properly establishes the use of a vaccine or other specific for a reeognized disease. Some 20 to 30 years ago, cancer research and therapy were bedeviled and endangered by widespread and irresponsible announcements of 11 10 I CrR• 1312