Tobacco Documents Online

~'~ ! ,i::~ ~- +' :~;~ iii+'• ii . %- STATEMENT OF IAN G. MACDONALD, M. D. CLINICAL PROFESSOR OF SURGERY UNIVERSITY OF SOUTHERN CALIFORNIA~ SCHOOL 0FMEDICINE LOS ANGELES3 CALIFORNIA +. presented before SUB-COMIML~!TEE ON LEGAL AND MONETARY AFFAIRS HOUSE COI$~'I~EE ON GOVER~IENT OPERATIONS Thursday# July 25, 1957 '~:,i : .,? My name is lan G. Macdonald. I am a Doctor of Medicine, having received ,++ ;++ my M.D. Degree from McGill University in 1928, and am Clinical Professor of Surgery.~?i~ ~i~ at the University of Southern California, School of Medicine+ where I have also+,, +~ ..... ,:.++~+++";+'+",+,.+., been coordinator of cancer teaching for the past i0 years. I hold membership in the American Association for Cancer Research, and in a number of national ..... :~: surgical and other medical societies. I have been engaged in the trea~nent of cancer and allied diseases since 1935, and am certified by the ~nerican Board of Radiology (in Radiation Therapy). I am a member of the National Board of Directors of the American Cancer , +. .... Society, and am past President of the California Division of that Society. am Chairman of the Sub-Committee on Cancer, Committee on Research of the Amerlcanili~ Medical Association. ~!'"+~'~,;+' - : The constant reiteration of the claim that cigarette smoking is one of ~. ;! the most important_, causes, if not the most important cause of lung cancer, has ~.O~;i:: persuaded many that a cause and effect relationship has actually been es~blished.! A poll of public opinion by the American Institute of Public Opinion reported in the newspapers for July 21, states that 50% of the general public believes that cigarette smoking is one of the causes of lung cancer. Those who advocate fer- vently the supposed causative importance of cigarettes in this disease se~mto have an almost evangelical attitude, and are remarkably oblivious of the fact that virtually the entire basis on which this belief rests is statistical. . p :~: +; , .. + t

-2- Although there is an apparent association between cigarette smoking and lung cancer, a review of the total evidence fails to establish a cause and effect relationship. . • r In discussing this particular problem# Berkson, biometrician at the Mayo Clinic has commented as follows: "(I am not affect by) the considerable number '~i~' of statistical studies published ..... showing an association between smoking and ~ • cancer of the lungs. On the contraryj undeviating consistency of statistical ..... i~ ~ results all in support of the same conclusion is in some circumstances the ..... ~! ~ hallmark of spurious statistical correlation, if correlation is produced by ~ ~,.~li~i?~.~ some elements of the statistical procedure itself, it is almost inevitable that • \~ the correlation will appear whenever the statistical procedure is used." It should be unnecessary, but probably is nQt, to say that things which are connected by a cause and effect relationship will show a high degree of correlation statistically, but that such high degree.of correlation between two things does not necessarily mean a casual relationship between them. - ' Cancer is a complex group of diseases of extremely disparate manifesta-i~. tions of abnormal growth of the tissues of the individual in whom the disease arises. A fundamental definition of cancer is an abnormality of growth of cells resulting from a disturbance in the extremely delicate system of check and balance which, under normal conditions, allows the body to replace worn out cells or to repair effectively the result of injury of various sorts. In general, the possibility of a given individual developing cancer depends upon two basic factors: i. The capacity of the individual to respond to unfavorable influences by developing cancer, and this capacity, or the lack of the capacity~ is now accepted by many as being genetically determined, or due to hereditary tendencies.

~ ~ ~i~i~ / 2. -S- The exposure of the individual to environmental factors which, in t~ the susceptible Individual3 may eventually result in the degree of abnormal growth whlch manifests itself' as cancer. Although we have limited information as to some of the genetic factors involved# and although we have considerable information concerning the predis- posing environmental influences which may lead to cancerj we are ignorant still ~'?~-> of the actual trigger mechanism which sets off the cancerous process in any " given individual. ~ ]~.~ ~.[~ A commonplace example is found in cancer of the skin. The genetie, i~~&]~./.~.~i~,>~]~..? influence of importance is the character of the individual's skin; the person ~ . -~',i-'~!,i~ with thin, delivate skin usually seen in blondes and redheads is extremely sensitive to the environmental agent, which in this instance is sunshine#.o~ at least the ultraviolet fraction of solar radiation. Another fact of importance is that people without sensitive skin structure if exposed over a long period of years to excessive amounts of sunshine have about the same hazard of developing • ~ skin cancer as does the person with "redheaded skin"who limits his exposure. ...... ~ ..... ,,~ ,. This will emphasize another important point, that an excessive amount of a ' > • potentially cancer producing substance over a long enough period of time may .. ..,.) make the individual who otherwise would have been resistant become as susceptible ~.i~ as a sensitive individual. In other words, three separate considerations are ' ~'! involved; sensitivity of the individual, exposure to the possible cancer pro- .... ducing substance, and the duration of exposure. There are hundreds of thousands of persons in our population who develop evidence of abnormal changes in the skin from exposure to the sun and wind, either by reason of sensitivity or long exposure, such as in farmers and sailors and ranchers; yet for every hundred people who develop the warty, scaling patches of abnormality there are only 4 or 5 who actually develop skin cancer even if they live to be well beyond the biblical three score and tex years. Thus we are aware of factors if individual

. ° • ~.~ ¸¸:¸I ~, .~" ;~,, -4- inherited sensitivity and of the most important single environmental influence which combine to create the hazard of skin cancer, yet we do not have any under- standing of the trigger mechanism which sets off actual skin cancer in one person. ~ .~ One further generalization appears to be worthy of mention in our effort Y.!~ :, to establish a general understanding of the cancerous process• Factors which /~-i'~:~~ are found rather consistently in the background of certain individuals with ~!~i~ certain types of cancer may contribute directly to the initiation of the cancerousS::!~i~/'~ ~ change, or theremay be evidence of some other factor which is actually the pre- /:i'~'°" disposing agent As an example, women who have their first ckild by the age of !~'~i~:~ 18 and who complete childbearing by age 25 have a dlstinc.tly greater hazzard of , !, cancer of the uterine cervix, but these factors are far more common in girls who grow up in less favorable economic situations and so the increased incidence ' of growths in the cervlxmay be a reflection of poor nutrition rather than early ~;i:L: . childbearing. In other words, the factor which is first discovered by statistical studies as being in excess in the background of the individual developing cancer may simply indicate the presence, or even the absence, of some other factor ,~ .- which is actually of predisposing or causative significance. There is a growing body of evidence in the past decade or more to ~: ;~ indicate that multiple factors are operative in setting the stage for a consider- able variety of cancers in man. There are a rather small group of human cancers in which only a single environmental agent is apparent as a predisposing cause of any significance; more commonly there are two or more separate factors frequently of entirely different nature which combine to. increase the liklihood of cancer developing in the individual. A This is the principle referred to as co-~ carcinogenesis, indicating the inability of ~ .single agent to produce cancer~ but when combined with another or other agents a cumulative effect is created which may

-5- lead to the development of cancer. Some of these factors are extremely weak carcinogens, or predisposing agents to cancer. As in experimental animals, some of them when administered in large excess cannot alone produce any semblance of cancer, while their combination in a relatively modest degree with other agents may produce cancer in the susceptible individual. There have been prior errors concerning causation of disease due to r , • statistical association. During the second and even the lhird decade of this century, medical textbooks referred to a certain form of cirrhosis of the liver as being "alcoholic" cirrhosis~ as practically 9 our of l0 individuals with • this type of liver disease could be demonstrated as having been excessive users of alcohol over long periods of years. However, it was also noted, even by those who were most convin~d of the causative role of alcohol in this disease, that it did occur at times in individuals who apparently had never been addicted to alcohol, and even in representatives of the clergy in whom a non-alcoholic history could be established• For some years now it has been established that what was formerly re- garded as "alcoholic" cirrhosis is not due to excessive alcohol consumption, but due to a nutritional deficiency. The high incidence of primary cancer of the liver in certain natives of South Africa and in inhabitabts of the Malayan Peninsula was originally thought to be due to environmental exposure; eventually it was shown with little rocm for doubt that the causative background of cancer of the liver in these arose from dietary deficiencies rather than from exposure to un- individuals favorable environmental influences. In the latter part of the last century dire warnings were broadcast *J concerning the rising incidence of cancer of the mouth, citing statistical ~'~ evidence that tobacco was the principal offender as a causative agent. Some of

• ~ °.. ~ • @ . • .J - • •:'2 - : ' -6- ..2 the indiViduals who corresponded to the present evangelistic statistical ex- ponents concerning lung cancer advocated a severe restrictive program on the use of tobacco in order that cancer of the mouth could be diminished. In the years that have followed it has been demonstrated that cancer of the mouth is attended by a severe nutritional deficiency in the great majority of individuals with this group of cancers. Tobacco has now been relegated, by all but the most feverish of the an~i-tobacco apostles, to an extremely minor role if# indeed, it occupies any position of causative significance. One of the principal arguments that smoking is responsible for an assorted increasing mortality from cancer of the lung is based• on the gradually .... rising mortality statistics for lung cancer. It is more than proper to offer some inquiry as to the validity, or at least the degree of the increase in cancer of the lung. It is a striking ~act that from 1900 to 1956 there has been a reduction in crude death rates from principal respiratory diseases from 430 to 57 per hundred thousand (U.S.P.H.S. Vital Statistics). When one considers that reasonably exact measures for diagnosis of lung cancer have became avail° < able only recently, it is obvious that many of the deatha~in the earlier decades of this century which actually were due to lung cancer were recorded in vital statistics as due to pneumonia, influenza, tuberculosis, ets. If one will accept that this error of past decades was as little as 5%, this adjustment would show that there has been no real increase in lung cancer during this century. A number of investigators have drawn attention to the questionable (~ validity of the figures usually quoted in support of the increase in cancer of the lung. One authoritative voice is that of Dr. Paul E. Steiner, Professor of ~-~ Pathology at the University of Chicago, who is generally regarded in the medical profession as one of the outstanding authorities in cancer research and epidemi- ology. Steiner remarks that an increase in frequency in lung cancer is reported

• _. ° @ ~L ~ ~ ~ ~ ~ ~ .i ~ ........ ~ • ......... "~ ~ ...... il. ~.~L I~"~ -?- in most clinical studies# in some mortality surveys, but in very few autopsy studies. His observations are such as to deserve quotation: "Great improvement has occurred in recent decades in the clinical diagnosis of lung cancer but accuracy in the autopsy recognition of the natural types has undergone little change~ having started from a much higher level. These facts cast doubt on whether the increase is disproportionate to other cancers. Even in the autopsy studies, reported increasesmay not be real but represent merely a shift in the type of material that the cliniJal departments send to the autopsy room. While much has been said of the increase in lung cancer relative to other typesI .... ~ .....~Q'~5~ .... Steiner reports that in St. Louis from 1935 to 1950 in 12,~43 autopsies lung .... cancer had increased from 1.5 to 1.9 to 2.5% of all autopsies in the three 5-year periods, but that two other types of cancer, those at the pancreas and breast had increased nearly as much as that of the lung, and that all cancers combined had increased from 16.8 to 20.6% of all autopsies during the period under study. In stmmmry, one may state that while most of the reports of studies based on the diagnoses of lung cancer made during life show an increase, studies based on autopsies vary, some showing an apparent increase and others showing no increase. Accuracy of diagnosis at post-mortem examination has changed very little since 1900, but the accuracy of clinical diagnosis has greatly improved. Considering these facts, an increase in lung cancer disproportinate to that for all or selected types of cancer has not been demonstrated in most geographical regions, and perhaps even in none. The main evidence claiming that smoking causes lung cancer is based on statistical surveys here and abroad. There are a number of geographic disparities related to tobacco con- sumption which seem to invalidate the suggestion that cigarette smoking is a causative factor. In Australia for example, from 1949 to 1951, the tobacco

-8- consumption averaged ~.7 to h.9 pounds per capita, compared with 5 pounds per head in the United Kingdom; in Australia the incidence of lung cancer was 13.3 per I00,000 as against 55.5 in the United Kingdom; in the 55-6~ age group the Australians had 30.5 cases per 100,000 as against the United Kingdom's figure of lll.6 per lO0,000. To take an example closer to home, in a recent year 2,003 tax-paid cigarettes were sold per person in Idaho compared to 2,319 tax~ paid cigarettes per capita in New York, but the lung cancer death rate in New York was four times greater than that in Idaho. The degree of difference in industrialization in New York and Idaho is well known. Further comment on the .}..,.";~.,,'~', factors of urbanization and industrialization will be made later. '~. • , It seems apropos to inject here a report from the Medical College Hospital in East Pakistan. During an 18-month study involving 362 cases of , pulmonary disease, 20 cases of cancer of the lung were found. The sex, age and occupational distribution corresponded to reports elsewhere; namelyj predominance in males, occurance in late adult llfe and preponderance of manual workers. All of the cases but two came from rural areas where alleged carcinogens from tarred roads and automobiles were conspicuously absent. Even more striking, of the 20 cases there were 14 who had never smoked, four were occasional smokers, and only two were heavy smokers. The causation of lung cancer is much more complicated than the over- simplified thesis that it is due to cigarette smoking. The data advanced by those who are mesmerized by statistics includes the English findings offered by Doll and Hill. Their controls (without lung cancer) actually contained more subjects in the moderate smoking group than there were lung cancer patients in that moderate smoking group. Their figures thus show that moderate smoking is actually commoner in persons without lung cancer. Even these data may indicate smoking to be a harmless pastime up to 24 cigarettes per day. One could modify an old slogan: a pack a day keeps lung cancer away.

D -9- Still another interesting fact in the data collected by English in- vestigators concerns the individual habits of inhalation of cigarette smoke. Many proponents of the cigarette - lung cancer theory have speculated that cigarette smokers inhale more than other types of smokers~ thus accounting for the excess of lung cancer in users of cigarettes. Yet Doll and Hill's figures showed that cancer of the lung was more common in cigarette smokers who did not inhale than in those who did inhale. Another curious finding advanced by the advocates of cigarette smoking producing lung cancer is the fact that the dilution of cigarette smoking by ~ ~ii~' some use of cigars and pipes materially decreases the probability of lung • ~ cancer. This seems to be without reference to the degree of moderation or the excessiveness with which the individual uses cigarettes along with the cigar and/or pipe. If cigarettes are car~inogenlc they surely should be so whether accompanied by tobacco in other forms or not. It is an insult to reason to believe that two men who indulge egually in cigarettes will have a different chance of lung cancer simply because one periodically adds to his cigarette smoking the use of tobacco in another form. Another fallacy in the theory that cigarettes have a causal relation- ship to cancer of the lung is shown by the difference in the relative incidence of lung cancer in males and females. Long before cigarettes were thought to produce the dramatic increase in lung cancer emphasized by the evangelistic statisticians, the ratio of lung cancer in males to females was in the range of 1.5 to 1 and in some areas almost 1.1. In the intervening period disparity in incidence of lung cancer has constantly increased until, for the U.S. at ~A large, it is approximately one female to five males, and in some areas such as Upper New York State, it is in a ratio of one female to seven males. During this period cigarette smoking by females has constantly increased, and a recent

-. o - lO - study of patterns of smoking by the U.S. Public Health Service indicated that slightly over ~0~ of women smoke or have smoked cigarettes since 1930. There has been at least a 25-year period during which women have been exposed to the possible causative effect of cigarette smoking, if there be any such effects, and nevertheless the disparity of incidence of lung cancer as compared with the male continues to widen. An opposite pattern of incidence of lung cancer is demonstrated by some racial groups, and one notable example is the high frequency of lung cancer in Mexican women dying in the Los Angeles area, as reported by Steiner. ~'~ • In autopsies performed at the Los Angeles County Hospital from 1918 to 1947,' ~' the Caucasoid and Maxlcan were the largest of various racial groups involved.~ Mexican men had only slightly more lung cancer than Caucasoid males, but Mexican females had significantly more lung cancer than Caucasoid females and nearly as much as their male counterparts. Nevertheless# a rough survey of patterns of cigarette smoking among older Mexican women in Los Angeles ~of lung cancer age) shows that their use of cigarettes is less than that of corresponding Caucasoid groups. If cigarette smoking were a real and causative factor in lung cancer, the relative incidence in men and women should be approaching parity instead of becoming more disparate, as is the actual situation. Knowing that lung cancer is predominantly a male disease, there should be a correlation of con- sumption of cigarettes by males with the "increase" in incidence of lung cancer. However, much of the increase in cigarette consumption since 1930 has been due (~ to the increasing use of cigarettes by women, and although there~re.mo accurate ~ figures available, our best attempts to evaluate this part of the problem indicate that aboutone-half of the U.S. increase in cigarette sales has been due to their increased use by women. Moreover~ the actual increased use of