Document e5mJXd4dYeEB5ZxJZnBG7nvEG

An Epidemiological Study of Lung Cancer in Asbestos Miners DANIEL C. BRAUN, M.D., and T. DAVID TRUAN, M.A., Pittsburgh Ever since the pronounced increase in the incidence of lung cancer among males became apparent, there have been attempts to associate it with one or another of the various elements in the environment of man. The approach used by some workers has been to suspect one or several sub stances and then set-about- in-an intensive search for lung cancer among persons who have had any exposure to those materials. In this connection, Smith82 writes: "The tendency of authors reporting the coinci dental occurrence-*) f primary-lung cancer with silicosis or with any other theoretical etiologic conditions, has been to emphasize the percentage relationship in extremely small series of cases, with control cases which are not in any way comparable." It would seem inevitable that asbestos should come under scrutiny in this manner, because prolonged exposure to this material is known to cause a specific type of pneu moconiosis, and because persons who show this form of pneumoconiosis often come to autopsy and provide a ready source of material for study. It was in this way that reports of the simultaneous occurrence of lung cancer and-asbestosis began to accu mulate after the report of a case by Lynch and Smith88 in 1935. Within the next 10 years,. abput_l5~ additional:; cases- were- re ported, and in 1954 Merewether 78 reviewed all deaths from asbestosis recorded in Eng land since March, 1924. Lung cancer oc curred in 16% of these cases. Gloyne,41 Accepted for publication Jan. 20, 1958. This study was made possible through a grant from the Quebec Asbestos Mining Association. Medical Director fDr. Braun) and Statistical Consultant (Mr. Truan), Industrial Hygiene Foundation. . 6.14 whose work is also frequently referred to as establishing a connection between asbes tosis and cancer of the lung, reported in 1951 that cancer of the lung was present in 14.1% of asbestosis cases examined by him. In 1941, Nordmann and Sorge7S claimed to have produced lung cancer in mice which they exposed to asbestos dust. Since 1951, additional cases of cancer of the lung coexisting with asbestosis have been reported, and, according to Hueper58 about 100 such cases had been reported up to 1955. As a result, an association between the two diseases appears to have been ac cepted by many authors, and several writers were using the term "asbestosis cancer" of the lung. Werber,8* in 1952, stated cate gorically that in 7% to 17% of cases of asbestosis, after a latent period of about lyi to 20 years, carcinoma becomes estab lished in the lung. On the other hand, not all authors ac cepted this alleged association without reservation. Saupe 80 in 1939 reported that he had discovered no cases of lung cancer among 620 cases of asbestosis which hehad examined; and in 1942, Holleb and Angrist48 expressed the opinion that the number of cases of asbestosis with lung cancer was too small for statistical evalua= tion. In 1947, Wegelius88 reported 126 radiologically diagnosed cases of asbestosis among 476 workers in Finland, and found no cases of lung cancer in this group. Goldblatt and Goldblatt in their section of Merewether's latest book," state: "But at no stage in all these impressive researches was any clue obtained which might have offered any support to the possibility that asbestos could act as a carcinogen. There 500^01.74 !.('NO CANCER IN ASBESTOS MINERS is no reliable criterion by which one can anticipate carcinogenicity and, as is well known, relatively minute changes in the structure of a chemical carcinogen are suffi cient to diminish or eliminate carcinogenic action. If asbestos is indeed to be regarded as a carcinogen, the need is felt to demonstrate some property which can be regarded as something more than inertness." These authors advance the theory that, until some more experimental evidence of direct carcinogenesis by asbestos or a de composition product of it can be obtained, asbestos might be considered as a "co-car cinogen" which only induces a further de velopment of a preneoplastic condition brought about by something.independent of the asbestos, such as an endogenous factor. Thus the literature, while tending to supl>ort the thesis that asbestosis is in some way related to the development of lung cancer, is by no means' unanimous. Alto gether, it is perhaps' more confusing than enlightening. A careful review shows that the majority of the reports are clinical and not epidemiological. They lack many ele ments necessary for the application of epidemiological techniques to their content, and most of the authors do not make claim to having done so. What has happened is that succeeding authors have drawn conclu sions and generalized beyond the scope of the works which they quote. Nowhere, for example, have we found references to a population of asbestos workers, although several authors who have quoted the ob served incidence of lung cancer in autopsies of persons who~also had asbestosis imply that this incidence applies to asbestos work ers. generally. We have- likewise- betn unable to find any study which actually calculated the incidence of lung cancer among a fiopulation of persons who had asbestosis, and not just those who came to autopsy. With the exception of a paper by Doll,28 none of those reviewed gave any data on exftosure and dust concentrations, and even Doll's paper merely mentions "scheduled" areas, by which is meant, "those areas where processes are carried on which were scheduled under the Asbestos Industry Regulations of 1931 as being dusty." There is, furthermore, a complete lack of definition of terms as used in the pub lished literature. For example, the term "asbestosis," as used, may refer to changes observable only by microscopic examination of the lung tissue, or it may mean a radio logically detectable condition. Most of the published reports obviously included women among their cases, but some of them do not give the number or proportion of women involved in the study. There is also a lack of uniformity as to what type of exposure most studies havedealt with. Of 99 cases enumerated by Hueper58 in I9S5, only 10 appear to have originated in the United States, and 7 in Canada. Some of the earlier reports ap parently included asbestos miners, but it can be assumed, since 82 of the 99 cases had originated in England, and since no asbestos mining operations are carried on in that country, that most of the reported cases have involved workers in the textile or fabricating industries. Such factors as smoking habits, family history of cancer, length of time in the industry, and age of the individual case are also notably absent in the majority of these reports. With this understanding of the limita- . tions of the existing literature with respect to epidemiological generalization, it may be of value to consider in somewhat more de tail some representative earlier publications, a few of which were referred to briefly above. - -- One of the most detailed studies and one which deserves the most serious considera tion is that reported by Doll28 in 1955. This study reviews causes of death among asbestos workers based on coroners' rec ords. It also attempts to estimate the risk by studying records of men who worked for at least 20 years in exposed situations. Broun--Truan 635 50030175 A. M. A. ARCHIVES OF INDUSTRIAL HEALTH Doll concluded that lung cancer was a spe cific industrial hazard of certain asbestos workers and that, after 20 years of expo sure, the risk is 10 times as great as for the general population. This article is important for several rea sons, in addition to the definite conclusions at which it arrives. For example, it begins by stating that "in view of the infrequency of asbestosis, this large number of cases (61 cases of lung cancer) suggests--but does not prove--that lung cancer is an occupational hazard of asbestos workers." Neither this article nor any previous one which we have examined presents any fig ures to prove that asbestosis is an infre quent occurrence. Estimates of the number of persons potentially exposed to asbestos dust in the United States alone vary from 10,000 to 35,000, and the incidence of as bestosis of any degree might be higher.than Doll imagines. This studyrlike so many others, involves autopsy records. The number of persons involved in the statistical analysis is only 113, representing only 1,042.25 man-years of life. It is also true that in selecting men who had been employed for at least 20 years, the study automatically excluded those who died from other causes after shorter employment. . Another reason why this publication is of importance is a statement which it con tains to the effect that "the strongest evi dence that it (lung cancer) may be a hazard (in asbestos workers) has been produced by Merewether and by Gloyne." In 1951 Gloyne i1 presented a review of 1205 autopsies on persons who had worked in various dusty occupations. This numbeF-included~l'32' asbestos~workers, of whom 121 showed "pneumoconiosis"--pre sumably asbestosis. Primary cancer of the lung occurred 17 times in this group, an incidence rate of 14.1% for lung cancer among asbestosis cases coming to autopsy. There were in his series 796 cases with sili cosis, and 6.9% of these also showed pri mary cancer of the lung. The incidence of lung cancer in other forms of pneumoconio sis was 6.7%, and in 169 cases which proved not to have any type of pneumo coniosis it was 8.3%. Gloyne considered "the mortality of the asbestos workers" to be "disturbing." First of all, it is obvious that the paper does not deal with the "mor tality of asbestos workers," and secondly, it must be borne in mind that all of Dr. Gloyne's cases were submitted to him for study because the findings were unusual for uncomplicated pneumoconiosis. It can reasonably be assumed that cases, including those of asbestosis, in which the findings were not considered unusual were not sent to Dr. Gloyne for examination. As a mat ter of fact, in the same paragraph in which he expresses concern over the incidence rate in asbestosis, Dr. Gloyne himself points out that the rate for lung cancer based on necropsies at the London Chest Hospital was 21.3% while the figures of the Registrar-General showed only 2.4%. He thus recognized that autopsies on a cer tain selected group of cases were not rep resentative of the general population. It would seem, then, that notwithstanding the value of Dr. Gloyne's work, its importance as an index of the prevalence of lung can cer in asbestotics has been misinterpreted by some who have quoted him. AIL that, it really shows is the fact that in a group of121 cases, selected for special study pri marily because they seemed abnormal by preliminary examination, 17, or 14.1%, had lung cancer. Merewether75 in 1947, in the report of the Chief Inspector of Factories, reviewedall cases reported between 1924 and 1946 in which asbestosis was the cause of death or a coexisting condition. This work*.was-u later extended to include all such cases re ported up to December, 1954, by which time there were 344 deaths, including 205 males and 139 females. Among them were 55 cases (16%) of cancer of the lung, 41 in males and 14 in females. It is quite possible that a large number of asbestotics who did not die of their asbestosis, or in whose 636 I'iiI, 17, Junr. IV5X 50010176 U'\\; C.tXc t.K IX .)s!il:ST<>\ MIXHKS death certificate it was not mentioned, may have been missed. The import of this 16% is enhanced by the simultaneous statement that the incidence of lung cancer in autop sies of the general population is only 1%. The danger of attempting to compare a rate found in 344 cases with the rate for the general population without respect to age, occupation, and many other variables, such as smoking habits, is obvious. Lynch,70 who with Smith 88 had reported the first case, in 1935, reported 4 cases of carcinoma of the lung in a series of 49 autopsies on workers in an asbestos manu facturing plant who were shown to have "demonstrable deposits of asbestos in the lungs." This, of course, is not necessarily identical with the disease asbestosis. Lynch, himself, points out that, although this is an incidence of 8.2%, "both figures are too small for very' serious statistical types of calculation." Nevertheless, later writers have used this paper to strengthen the case for an association of carcinoma of the lung with asbestosis. It is also of interest that Klotz83 found only the same number of cases of lung cancer in a series nearly 10 times as large, i. e., 4 in 47.8 cases of asbestosis. Behrens, as cited by Merewether,78 esti mated that, of 309 cases of asbestosis in the literature; 44 showed associated cancer of the lung--giving am incidence of 14.2%. This is an illustration of generalizing an incidence obtained in a group of cases which were undoubtedly reported only be cause some of them showed lung cancer, to possibly hundreds, of. asbestotics whose cases were never reported. The same ap plies to the conclusion of Telekv,85 who appears to have' reviewed reports of 39 autopsies on persons with asbestosis among which 6 cases of lung cancer occurred. In formation from sources such as these does not justify generalizations with regard to mortality rates. Perhaps no one has written so exten sively on the subject as has Hueper.32-59 In 1955 he reviewed the cases58 reported prior to that date and enumerated a u*ui of 99. Eleven of these were those dis cussed by Doll28 and may have been cases covered by other authors. Eight were dis covered by Kennaway and Kennaway62 in an analysis of death certificates, and, unless Merewether's study was incomplete, these cases should have been included in his re port. Of the remaining 80, it is quite pos sible that the 31 contributed by Merewether and the 17 by Gloyne contain some duplica tion with each other or with those of other English authors. Principles of the Epidemiological Method Dorn 30 has pointed out that much of what is now thought to be pertinent concerning the comparative frequency of lung cancer in different population groups has been de veloped from the analysis of clinical ma terial, particularly surgical and autopsy records, supplemented to some extent by the reported impressions of various clini'cians based upon their personal observa-, tions. More recently, however, attention has turned to the systematic investiga tion of this problem by the same methods that have proved so successful previously in the study of communicable diseases, that is to say, by epidemiological methods. In order to apply this method of investi gation to the problem under discussion, we were of the opinion that a study should be planned so as to provide (1) a well-defined, population group; (2) available data for all members of this population, including the. healthy as well as the ill; (3) a sample which is truly representative of the popula tion; (4) reliable and valid observations relating to the problem of the study. A serious defect, common to most of the studies which have been reported, is that little or no information concerning the healthy people in the group seems to have been available to the author. In order to draw a generalization regarding all asbestos workers, it is necessary for a study to in clude living persons as well as the dead. Braun--Truan 637 500-501.77 A. M. A. ARCHIVES OF INDUSTRIAL HEALTH Limiting the investigation to cases coming to autopsy, as has been frequently done in earlier studies, still further restricts its use in generalization. The problem with which we are concerned is whether asbestos be found in another group of the same size, but representative of the general population. It is true that investigation of cases from such a sample can furnish information valuable for research, but the use of this miners experience more lung cancer than information in drawing generalizations is does the general population. The answer necessarily restricted. It is the obligation necessitates the collection of reliable infor of both the investigator and of those who mation on asbestos miners as a group, as read his report to make proper comparisons well as on the general population. and to draw only those conclusions which It seems advisable to discuss the differ are valid and justified. A good statistical ences between the epidemiological approach study of cases of cancer of the lung occur and that used in the studies which have ring in a group of 'autopsies can lead to a been reported to date. A very important proper inference concerning the frequency consideration is the fact that lung cancer, of lung cancer among cases coming to in spite of its increasing numbers, is still a autopsy, but only to such cases. For in disease of low incidence; that is, in a given formation from such a-study to be pro-, population not many persons will contract jected to some larger group, it is necessary this particular disease. This fact requires that the autopsies represent a good sample that-large-samples or groups must be stud of that larger group. To assume that such ied to provide meaningful results. is the case in any particular series is dan Recognizing the difficulty of obtaining gerous and likely to be false. such large samples, most earlier writers deviated from the epidemiological method and sought to circumvent the requirement There is some danger that the figures reported by some authors may be miscon strued as applying to asbestos workers or of. observing well persons by (1) compar even asbestos miners, when,- in fact, the ing the relative frequency of cancer in vari authors in question do not make this gen ous sites; (2) comparing the relative eralization, nor can the generalization be frequency of cancer in a group of hospital made for the reasons stated. Close study of ized patients; (3) comparing the relative the reports reveals that the percentages frequency of cancer in a group of cases cited relate only to the group of autopsies coming to autopsy. covered by the particular investigation. Attempting to compare two population The present study, in contrast to the groups,* looking only at the relative fre earlier works, has been planned to utilize' ' ` y.\ jj'vsS quency of cancer in various body sites, the epidemiological method. A well-defined may ~resultrin finding a higher percentage group of asbestos miners~has" been estab- ' (relative frequency) in one of the groups, Iished in such a way that it constitutes n\ when,, in fact, the mortality rate of cancer good sample of the whole population, of y of a particular organ is exactly the same asbestos miners in Quebec. Data for all in both groups: This is very clearly dem members of this group have, been-collected onstrated in the excellent article by Dorn.30 and analyzed. Those concerning lung can-" The mortality rate from a particular cause cer have received most careful considera is the true measure of comparison. tion. Details of the methods employed will It is apparent that selected groups such as hospitalized patients or autopsy cases may not be in any way representative of be set forth later, but the type of approach is considered to permit of fair comparisons and valid generalizations. a larger group, and that in dealing with Collection and Analysis of Data such samples, the observer may easily find A preliminary survey of potential sources more cases of a given disease than would of information in February, 1956, involved fi.W Vol 17. June, ms 500 *.0178 r.rxa caxcer ix asbestos miners discussions with the physicians in charge of the asbestos companies' programs and with clinicians, pathologists, representatives of City and Provincial health departments and of the Canadian Cancer. Society, and other interested persons. It was found that morbidity data, although somewhat limited, were available from such sources as the. hospitals in Montreal and Quebec City, and the 13 cancer detection centers in the Prov ince. However, because of the high mor tality in lung cancer, it seemed advisable to depend upon data relating to deaths. These we found to be obtainable' at the vital statistics department of the Ministry ` of Health in Quebec City. From the pre liminary survey, it was apparent that exten sive- and- detailed information could be gathered with respect to both the persons employed in the asbestos mining industry and the mortality figures for the general population. Following this exploratory survey, the initial effort was directed to the collection of data relating' to all .workers who had been processed through the clinic at Thetford Mines since its inception in 1947, and similar information regarding all workers at Asbestos, Que. Data from the clinical records included the age, family and per sonal medical histories, smoking habits, number of years of exposure, an estimate of weighted exposure, and the course of the -individual's health status or the cause of his death. - From this information it was possible to formulate a "cohort" which could be well defined,' ^should- be representative of the whole group; and could be followed for a definite period of time. All of the available experience indicates that the development of asbestosis in less than five years of ex posure must be somewhat rare. Accord ingly, the cohort was defined as including every miner who had a total ex]>osure of five or more years, and who was on the employment rolls in 1950. Office and other nonexposed personnel, regardless of length of employment, were not included. This cohort was then followed by means of the annual physical examination records through a six-year interval, 1950 through 1955. All data regarding this group were then tabulated in order to determine the characteristics of the cohort. For those who survived the entire period, reference was made to the physical examination results and x-ray findings at the end of the period. Those who had died were tabulated sepa rately, and the cause of death was corroborated by examination of the death certificates. A further search was made concerning those in the original cohort who remained unaccounted for when the living and the known dead had been tabulated. They represent men who had left employ; ment through retirement or resignation. Eventually, all but a small number of these were accounted for as either living or dead, and in the latter event, the cause of death was substantiated in a similar manner, and the results added to the original list of deaths. Death certificates for the Province" of Quebec for the years 1952 to 1955, inclu-. sive, were reviewed in the department of vital statistics of the Provincial Health Ministry, together with statistical sum maries of the causes of deaths in the. Prov ince by counties. All cases in which death was certified as having been due to pri mary cancer of the lung were examined for such information as place of "residence, occupation, date of death, hospital in which death occurred, and whether or not an autopsy was performed. Cases in which lung cancer was given as a cause of death, but in which it was not specified as to whether the cancer originated in the. Iung^, were also reviewed in an effort to include all instances of primary carcinoma of the lung in the study. The statistics for the Province of Quebec relate to population, total deaths from all causes, total deaths from cancer of all types, and deaths from lung cancer. These were collected and tabulated by counties and by sex for the years 1950 to 1955. inclusive. Rrmtn--Trttan 500*0179 A. M. A. ARCHIVES OF INDUSTRIAL HEALTH From them, death rates for the general population of Quebec and of individual counties were calculated for specific years and analyzed by cause. Practically all employees of one company are covered by a group policy of life insur ance which, fortunately, nearly all of them continue to carry when they retire. A very few are not covered by this policy, and those who leave the industry for one reason or another except retirement usually are no longer covered, but this is likewise a small number. As an additional check upon the. information obtained from the clinical records on this group, the records of the life insurance company were examined for all death claims paid tinder the p>olicy, and particular notice was taken of the claims in which the proof of death was based on cancer of the lung. Deaths from lung cancer among asbestos miners. were thus determined from the clinical records in the medical service of the industry and checked by means of the death certificates,and insurance company records. The deaths were then verified individually by reviewing them with the physicians in charge of the medical services. In this man ner, there was established a list of cases in which primary cancer of the lung is con sidered to have been proved as the cause of death. A few cases in which lung cancer is strongly suspected but not proved as the cause of death were considered separately. Mortality rates have been calculated using both the "proved" and the total of "proved" and "suspected" cases during the years un der observation. Comparisons were then made between the death rates from the same cause among specific segments of unexposed persons. ' AiP lung cancer deaths, both suspected and proved, were carefully ana lyzed to determine possible relationship or correlations between the development of lung cancer and any factor known from the clinical records, such as family history of cancer, personal history of heavy smoking, coexistence of asbestosis. or exposure to asbestos. .' 640 In addition to this analysis of deaths oc curring in the cohort and during the years under observation, every known death from cancer of the lung, as well as every case diagnosed but still living, has been tabulated and analyzed. They will be discussed separ ately from those included in the-population and time-interval under study. A comparison of lung cancer mortality in the asbestos-producing counties has been made with that in counties which are far removed from the asbestos mines and in which, presumably, no asbestos miners live. Finally, in order to broaden the compari son of death rates in different population groups,- the rates have been collected for Canada generally, and for the United States, according to the most recent published and unpublished material. ' Results and Interpretation The cohort which was constructed accord ing to the criteria described in the preceding section has been considered individually and compared with the general population. Description of the cohort will be presented here as a preface to the results of the study: Original Cohort Loot Persons Deducted Final Cohort Living in 1955 (working and retired) Dead by 1955 Cancer of lung - *" * Questionable cancer of lung Other causes Unknown causes *` ... Smokers Nonsmokers Unknown ~. 6,091 133 5,958 5,771 187 " * `9 3 ' l'69 -g 4,673 1,265 20 Tables 1, 2, 3, and 4 present age, number of years of employment, weighted average' exposure, and smoking habits of the cohort. A comparison of the exposure- to^asbestos, dust is presented in Table 3. All members of the cohort were placed in one of three categories, representing increasing degrees of exposure based on a weighted average of the years spent at various levels of dustiness. The degree of dustiness for each job category was determined after consulta tion with persons familiar with the environ ment and conditions in the various work I'ii/ IT, !imr. W.V 500101.80 LUNG CANCER IN ASBESTOS MINERS Table 1.--Number and Percentage Distribution Table 2.--Number and Percentage Distribution by by Age Length of Employment An 20-44.................................... 46-64.................................... 65-64.................................... 65+..................................... Unknown...................... . Totals................... ................ Annie Ago_____ _____ _ Number 5,058 * Leu thin 0.4%. Per Cent 66 19 10 5 too Length of Employment 5-9................................................... ... 10-19..................................................... 20-29.................................................... 30-89..................................................... 40-49..................................................... 80+................................................... .. Totals................................. . Average years of employment___ .. Number 1,795 2,396 922 603 182 55 18 Per Cent 30 40 16 10 3 1 100 areas. For the purposes of calculation, the assumption has been made that the relation ship between these categories is linear, and that Category II is twice as dusty, and Category III three times as dusty as Cate gory I. ually smoking more than five cigarettes per day. Persons who smoke pipes or cigars exclusively were not considered to be smok ers for the purpose of this study. Table S presents the year-by-year ex perience of the -cohort and indicates - the Table 3.--Number and Percentage Distribution by Exposure Category Table 4.--Number and Percentage Distribution by Smoking Habits Exposure Category I......... II........... Ill........... Unknown.. Totals....... Average Exposure * Leu than 0.2%. Number Percent *,031 2,120 1,7732 . 34 30 30 ------2,028 -- 100 2.0 ______________________ Smoking Habits . Number Per Cent Smoken.......... ...................................... Nonsmokers........................................ Unknown............................................ Total*..................................... 4,073 1,262 ---2-0- 5,058 7* 21 *_ 100 -------------------------------------------------------------------------------- ------------------------ ------- ------------------- * Leu than 0.5%. The fourth variable, smoking habits, was similarly tabulated and is shown in Table 4. This was included because the informa tion was available and because smoking was regarded as one of the variables which, besides the environment, could conceivably influence the development of lung cancer. As used in this presentation, the term smoker refers to a cigarette smoker, habit- number of deaths each year from specified causes. In general, a case was considered to be "proved" as one of primary cancer of the lung when the records showed that the diagnosis had -been -supported by an autopsy or surgical resection of the lung with microscopic examination of the re moved tissue. In one. case so considered, however, diagnosis was confirmed by bronr Table S.--Year-by-Year Experience of Cohort and Death Rates per-100,000 Man-Years of Risk . Cauae of Death Year No. Alive At Beginning of Year Proved Primary Cm. of Luos Suspected Primary Ca. of Lung Other Cause* 1950........................... .... 1951................................ 1952.......................... ..... 1953.......................... 1954........................... 1955......... ................ . 5,958 5.942 5.922 1 2 0 2 l 3 2 12 0 17 0 37 1 34 0 33 0 36 Totals......... '9 g "Proved" rate per 100*000 man-year* of risk-- - ----- 25.5 3 169 Unknown 1 1 0 0 4 0 6 Man-Years of Risk 5,950 5,932 6,903.6 5,866.6 5,829 5,790,5 35,271.5 `Total" rate per 100,000 man*yearto(rl*k^r=r7---- 3441 Broun--Truan 509*Q%81 641 A. M. A. ARCHIVES OF INDUSTRIAL HEALTH Table 6.--` 'Proved" Cases of Primary Cancer of the Lung Cue No. 1 2 3 4 5 0 7 8 9 Age 66 66 GS 69 48 61 56 37 68 Smoker Yes Yes Yes Yes Yes Yes Yes Yes Yes Exposure 26 yr. In Cst. II 34 yr.In Cst. I 37 yr. In Cat. I 32 yr. in Cat. I 22 yr. in Cat. 11 33 yr. in Cat. Ill 30 yr. In Cat. II 16 yr. In Cst. I 28 yr. In Cst. Ill . Died 10-12-61 5- 3-65 7-30-65 8-30-60 -61 4-90-63 11-22-63 8- 6-64 6- -65 Autopsy Yes Yes ' No Yes Yes Yes Yes Yes Branch. Asbestosls Yes No No No Yes Yes Yes No No choscopy with visualization and biopsy. In another, although there was no autopsy, the diagnosis of primary cancer of the lung seems to have been beyond question. The term "suspected'' primary cancer of the lung was applied to those cases in which the diagnosis remains in doubt but some of the evidence points to cancer of the lung. There were three such cases. The term "man-years of risk" has been used to mean the number of men at risk for the year under observation. A person who lived throughout the year was counted as a full man-year of risk, but one who died during the year was counted as one-half a man-year. Deaths occurring in the cohort, and in which lung cancer is considered to have been proved as a cause, are shown in Table. 6. In Table 7 are shown three deaths which have been considered as "suspected" lung cancer cases. An indication of the importance of these "suspected" cases in interpreting the results of the calculations is desirable before fur ther discussion of the mortality rates which are derived in later tables. For example, it happens that the rate found for the proved cases is close to the "expected" rate based on the general population figures for the Province, as will be shown later, and on this basis we should find eight deaths from lung- cancer among the cohort. Actually, nine cases were observed. If, however, the 3 additional "suspected" cases were in cluded. increasing this figure to 12, the total would be very close to the 95% level of significance. However, having found just 12 cases, we are not above this level, and therefore the hypothesis that asbestos miners do not have a higher mortality from lung cancer than does the general population can not be rejected. Nevertheless, the occur rence of 12 cases in this sample would increase the rate to a point which approaches the significant level. Because of the tre mendous importance of the questionable cases in this respect, some detail regarding them will be given here. In one of these cases, the suspicion of cancer of the lung is based upon the x-ray interpretation, and, although no autopsy was performed, the death certificate indicates that death was due to lung cancer. It is well known that the x-ray appearance of fibrosis, especially if a localized density or a superimposed tuberculous lesion is pres ent, can simulate that of a tumor, and by itself, does not justify the inclusion of this case as one of-"proved"--cancer of the lung. A second case was certified as having died by reason of hydrothoraxT possibly due to lung cancer, but again there was no surgery and..no postmortem examination. In the third case, although it was subjected to autopsy, two-pathologists disagreed as to whether lung cancer was present.- The death was certified as having been due to chronicmyocarditis With nephritis arid pulmonary congestion, and possibly cancer of the lung. On the basis of these facts, it seems unwar ranted to include these three cases among Table 7.--"Suspected" Primary Cancer of the Lung Case No. 1 2 3 Age 49 46 65 Smoker Yen Ye* Yes Exposure 32 yr. In Cat. Ill 33 yr. In Cat. HI 42 yr. In Cat. II Died 10-25-50 7- -53 9-30-55 Autopsy Yes No Asbestosls No Yes No Vol. 17, June, 195ft 500101.82 LUNG CANCER IS ASBESTOS MINERS "proved" instances of lung cancer. On the other hand, they cannot, in fairness, be dis Table 9.--Lung Cancer Deaths by Length of Employment regarded completely. It is for this reason Number of Persons and Number of I.unu C ancer Deaths that mortality rates have been calculated both ways. Length of Employment * No. of Deaths No. of t'enons Proved Total Table 8 gives the rates by age groups. 5-*......................... 00 The rates by length of employment are 10-19......................... 20-29......................... .. 922 l I 3 i'M ' ,'-T shown m Table 9. During the first 40 years of employment, the rate rises, an observa 30-39......................... ... 40-49......................... .. M+........................... .. 003 185 55 A 0 0 7 l (i tion which seems plausible since the men Tout,........ .. S.M8 9 12 were growing older. However, after 40 years of exposure there are no "proved" cases reported for a total of 240 men dur Annual Lung Cancer Death Rates per 100.000 MacuYears of Exposure Length of Employment Proved Total ing the six years, or about 1440 man-years of exposure. When the "suspected" cases are added, one case does show up in this Table 8.--Lung Cancer Deaths by Age Groups ................................................ o 10-19..................................................... 7 20-29..................................................... 30-39..................................................... 13 40-49..................................................... 0 50+....................................................... 0 Over-nil.................................. 25 a m 9U0 Number of Persons and Number of Lung Cancer Deaths dicates that the members of the cohort A|i Group No. of . Persons No. of Deaths Proved Tout did not die from lung cancer at a younger age than the general population. 20-44......................... .. 45-54......................... 55+-64......................................... Unknown............... .. 3.901 3 1 1 3 4 0 l The rates by weighted exposure are 3 3 shown in Table 10, and it will be noted 5 0 that they present strong evidence against Tout,......... .. 5,958 9 12 asbestos being a carcinogenic agent, for. if exposure to asbestos is in any way con Annua] Lung Cancer Death Rates per 100,000 Man-Years of Exposure nected to lung cancer, we would expect that Age group Proved Tout the longer and heavier the exposure, the ' i\ 20-44....................................... 4 higher the rate that would be found. The 45-54....................................... 55-64....................................... fi5+......................................... Unknown.............................. 44 81 only possible error in this interpretation 265 0 could occur if the weighted exposures were Over-all.................... 34 inversely related to years of employment. period. This would produce a rate of 69 Table 10.--Lung Cancer Deaths by Exposure .. Category_ . per 100,000, again demonstrating the im portance of these questionable, but un proved, cases to the final conclusion, because, if there were no cases in this number of men with long exposure, and if asbestos is a carcinogenic agent, it must be Number of Persons and Number of Lung Cancer Death- Exposure Category No. of I'eraon, No. of Deaths Proved Total l.............................. II.............................. Ill.............................. Unknown.................. 2,031 2.150 , 1,772 5 4 3 2 0 4 4 4 0 concluded that these 240 men have demon Total,-. -i.S5S~- - - * 9 * 12 strated considerable resistance. This is a bio :?* zji logical phenomenon which has been observed previously and is consistent with the theory Annual I.ung Cancer Death Rates per 100,000 Man-Years of Exposure Exposure of an intrinsic or endogenous factor in can Category Proved Tom cer. The only other explanation would be that the susceptible members of this age group had died earlier of lung cancer. Table 18. which appears later in this section, in- l......................... ............................. II......................... ............................. Ill......................... ............................. Unknown............. ............................. Over-oil-- 33 23 19 0 33 31 37 0 34 Braun--Truan f4.t 5O0'i01 S3 .1 M. A. ARCIlll'liS Oh IXIH'.'TRIAI. I'h.Al.TII Table 11.--Xumber of Perseus w Carious Weighted Exposure Categories by Length of Employment Length of Employment Weighted Exposure Categories I II III * Un known 5-0 10-19 20-29 30-39 10-49 50+ TOT 663 314 247 76 24 Total! 23t Avenge years of exposure 17.8 598 883 363 218 67 21 i i ... 2,150 17.8 490 851 243 137 41 10 i u-- 1.772 16.6 0 1 2 1 1 0 -- 5 29.0 Total 1,795 2,396 922 603 185 55 1 5,958 17.5 in which case the heaviest weighted ex posure (Category III) would show the shortest length of employment. Table 11, which lists the number of persons in various exposure categories by length of employ ment indicates that this error has not oc curred. In fact, the average number of years of employment for each exposure category is almost identical. Table 12, which develops the rates for smokers and nonsmokers, is most striking. It shows that not a single case of lung cancer developed among the 1265 non smokers and that all cases of lung cancer, both "proved" and "suspected," occurred in smokers. Table 12 was so striking that it was felt that further verification was necessary. It was possible that some abnormal distribu tion may have occurred, e. g., the non Table 13.--Number and Percentage Distribution of Smokers and Nonsmokers by Age Groups Table 12.--Lung Cancer Death Smokers and Nonsmvkers Number of Persona and Number of Lung Cancer Heaths by Smoking Habits No. of Deaths Persons Proved Total Smokers................... .. Nontmokers...... . .. Unknown............ .. 4.673 1,265 * 20 9 0 0 12 0 0 Totals........ .. 5,958 9 12 Annual Lung Cancer Death Rates per 100.000 Man-Years of Exposure by Smoking Habits ""No. of Deaths Proved Total Smokers................................ ............ Nonsmokers......................... ............ Unknown.............................. ............. 32 0 0 43 0 0 Over-all................... ............. 25 34 smokers may have included a larger percentage of young men. Consequently, ad ditional Tables, 13, 14, and 15 were con structed to show the distribution of smokers and nonsmokers by age, length of employ ment,- and; degree-of exposure. Although there are - slight differences, they do not account for the fact that all observed cases of lung cancer'were in smokers. In respect to age (Table 13) the combined average age of the smokers was 4.9 years less than that of the nonsmokers. Table 14 shows that as far as length of employment is con- Table 14.--Number and Percentage Distribution of. Smokers-and Nonsmokers by Length of Employment Age Group 20-44......................... .. 46-64......................... .. 56-64......................... .. +........................... .. Unknown............... .. Total,........ .. krtni, ate........... .. Number Smoker Noosmoker 3,200 896 411 164 2 689 224 202 150 0 4,673 39.3 1,265 44.2 Unknown 12 4 2 1 1 20 40.7 Employments Smokers 5-9.............................. 10-19............................ 20-29............................ 30-39............................ 40-49.^..................... 50+.............................. .1,408 1,967 722 433 no 33 Totals............ Average length of employment.-.*..,..^- 4.673 17.0~ Nonsmokers . 377 426 199 166 75 22 1,265 , 19.3 Unknown 10 4 2 4 0 0 20 ir,.o Percentage Distribution Age Group Smoker Nonsmoker 20-44......................... .. 45-64......................... .. 56-64......................... ... 96+.............................. Unknown............... 68.5 19.2 8.8 3.6 54.5 17.7 16.0 11.8 0JO Total!......... .. 100% 100% Unknown 60J) 20.0 10.0 5.0 5.0 100% * Leo than 0J35%. 644 Percentage Distribution Length of Employment Smoker Nonsmokcr S-fl.............................. 10-19............................ 20-29............................ 30-39............................ 40-49.................................. 50+................................... 30.1 42.1 15.4 9.3 2.4 0.7 29.8 33.7 15.7 13.1 5.9 1.8 Totals............ 100% 100% Unknown SO.O 20.0 104) 20.0 0.0 0.0 100% I'ol. 17. June. 1958 50010181 l.CNG C.4NCEK JX .-ISfiliSTOS MINERS Table IS.--Number and Percentage Distribution of Smokers and Nonsmokers by Exposure Category Exposure Category Number Smokers Nonsmokers Unknown 1.............................. 1,553 474 4 IL.............................. 1,691 452 7 III.............................. 1.425 338 9 Unknown.................... 4 1o Totals............ 4,673 Average exposure category......................... 2.0 1,265 1.9 20 2.3 Percentage Distribution Exposure Category Smokers Nonsmokers Unknown 1.................................. II................................... III................................... Unknown.................... 33.2 36.2 30.5 0.1 37.5 35.7 26.7 0.1 20.0 35.0 45.0 0.0 1 Totals............ 100% 100% 100% cemed, the smokers had worked about 2.3 years less on the average than the non smokers. With longer exposure and greater age, one would expect the nonsmoking group to show a higher rate if lung cancer were due to asbestos. Table IS shows that the average exposure category was almost the same for the two groups. Therefore, this variable seems to be of no importance in accounting for this difference. The result of this additional analysis is that none of these factors appears to lessen the effect of Table 12. Comparison of the Cohort Experience unth that of the Province of Quebec, Do minion of Canada, and the United States.-- In order to make a comparison of the experience among asbestos miners with that of the general population of the Province of Quebec, statistics were gathered, as stated earlier, in the office of the Division of Demography in the Provincial Ministry of Health. The data on total deaths, deaths from all forms of cancer, and deaths from cancer of the lung were obtained by sex and by county for the years 1950 through 1955. In addition, all death certificates which specified primary cancer of the lung, and all those which indicated lung cancer but did not specify the origin, were exam ined for the years 1952 through 1955. Table 16 gives a tabulation of the number of deaths from lung cancer in the Province and in the cohort for the years 1950 through 1955, and shows the annual rate per 100,000 in these segments. It will be noted from the table that the mortality rate for the "proved" cases in the cohort is only slightly higher than the rate for the Province. When the "suspected" cases, are included in the calculation, the rate for the cohort rises to 33.8 per 100,000, which is about 50% higher than the rate for the Province. This, it will be recalled from the previous dis cussion of the effect of the "suspected" cases on the results, approaches but does not exceed the significant level. One further interesting observation from Table 16 is the rather marked increase in the total number of cases for the Province between 1950 and 1955. It is assumed that at least part of this increase is due to im proved recognition and ` reporting of lung cancer during the interval. For this reason, the years 1954 and 1955 'were thought to be more nearly representative of actual condi tions. - Even- so,- it is quite likely that the general population is not studied for lung cancer with the same diligence with which Table 16.--Comparison of Cohort with Province of Quebec Lung Cancer Deaths Annual Number of * * Rate per Persons 1950 1951 1952 1953 1954 1955 Total 100,000 Province 1,196.000 196 220 245 303 303 357 1,624 Cohort Total Proved Province 5.923 t 3 * 0 3 1 3 12 l 202 1 39 1,192,000 193 218 245 300 302 354 1,613 (excluding asbestoa workers) 22.6 33.8 25.3 22J In the Province figures, it has been assumed all male lung cancer deaths are (or men of 304- years. * Approximate midpoint o( the enumerated population for 1951, and the estimated population for 1954 (Rapport, Dir. de la De* tnographle). tNumber olive in cohort at beginning of 1952. Braun--Truan 645 50010185 Age Group 20-44 45-54 55-64 65+ TottU A. M. A. ARCHIVES OF INDUSTRIAL HEALTH Table 17.--Lung Cancer Deaths for the Province of Quebec * _ Population 777.000 202.000 137.000 121*000 1,237,000 Number of Deaths 1954 Total Certified Specified Primary Total Proved Total Proved 17 13 11 9 67 36 43 29 100 38 69 32 111 3S 68 29 295 120 191 99 1966 Total Certified Specified Primary Total 17 ti4) 117 148 Proved 35 64 Total 17 56 116 148 Proved 9 32 57 54 342 165 337 152 Age Group 20-44..................... ............................... 45-64..................... 55-64..................... M+....................... Over-all--............. Death Rates per 100,000 1954 Total Certified Specified Primary Total Proved Total Proved 2.2 1.7 1.4 1.2 17.8 21.3 14.4 7.7 50.4 23.4 27.3 56.2 24.0 9.7 15.4 8.0 J965 Total Certified Specified Primary Total Proved Total Proved 2.2 29.7 85.4 122-3 1.2 2.2 17.3 27.7 41.6 84.7 44.6 122.3 1.2 15.S 41.6 44.6 27.6 12.5 27.2 12.3 * Data from death certificates. 4 Ettlmate tor population obtained by applying 1551 percentages (or age group* (or males to the total population lor 1551. as given iu Rapport, DIt. de la Demographic. this disease is looked for in the miners, and This comparison, summarized in Table it seems probable that the mortality, rates 18, ' shows that the observed number of for the Province may be low. This would deaths in our sample is not significantly appear to be substantiated by the fact that -greater than the expected number of deaths, the reporting of cases in the cohort showed based on the average of the 1954 and 1955 no such increase over the same period.. . . . figures for the Province. It.is true that, in Table 17 was compiled to show the an the case of the age group of 65 and over, nual age-specific lung cancer rate of cases the five deaths provide a figure which is in which the death certificate merely read almost significant at the 95% level. How "cancer of the lung," as distinguished from ever, it should be noted that this number those in which the diagnosis was confirmed by autopsy, surgery, or biopsy. The term "specified primary" refers to those cases in these two categories in which the tumor was specified as having originated in the lungr It will be noted that, of the total cases reported in 1955, a much higher percentage includes one of the suspected but unproved cases previously referred to. Furthermore, it is rather likely that the rate for the general population i> understated in thF age group, for the obvious reason that the exact cause of death in the very old is not than in 1954 were specified as primary. The table also shows that a higher percentageof the total cases certified in 1955 were proved, again indicating increasing interest Table 18.--Comparison of the Actual ami Expected Number of Lung Cancer Deaths by Age Among Asbestos Miners in this disease. A comparison has been made between the age-specific rates shown in Table 17, and those for the cohort, shown in Table 8. An average of the 1954 and 1955 rates for the Province has been used, since the 1955 Age Group Province Total Specified Primary Rate per 100.000 20-44 45-64 55-64 65-4- Unknown 1.8 24.5 67.6 89.3 .. Observed No. Expected of Death? No. of No. of . Miner? Deaths* Proved Total 3,901 1.124 615 315 3 0t l 2t 23 24 .. 0 l 3 3 5 0 figure was higher and may have been ex ceptional. The expected numher Is based on the average of the 1954 and 1955 age-specific rates for the Province of Quebec, t Actually 0.4. 646 I'ol. 17, Junr. /W 50010186 LCSG CASCUR IS ASBESTOS MISERS Table 1*).--Age Distribution of Adult Males for Table 20.--Annual Death Rates per 100,000 for the Province of Quebec, 1951 * Cancer of the Lung in Canada * A Re Group 20-44.................... . 45-54.................... . 55-64.................... . *5+..................... . Tottl......... . Number 727,135 188,952 128,944 113,467 1,158,498 Percentage 63 16 11 10 100 Percentage in Cohort 66 19 10 5 100 Rapport, 1*54. a matter of the same intensity of interest as it is in younger persons. Table 18 also answers a question pre viously raised. It shows that the members of the cohort have not died from lung can cer at an age earlier than the general popu lation. and that such an explanation cannot le offered for the absence of lung cancer in 240 men with more than 40 years of employment referred to on jiage 643. - Before leaving this comparison of the Province with the miners, it should be shown that their age distributions are rea sonably the same. That this is the case can be observed from Table 19. ' It should be remembered that the miners retire and consequently, it can be expected that the oldest age group will be larger in the general population. The data presented in Table 20 indicate that the lung cancer rate generally decreases after age 70. There fore. we could expect the rate for all people over 65 to be smaller than the rate for the group between 65 and 75, which would apply to the oldest group of miners. It is felt that by using the whole adult male popula tion. we have developed rates for the gen eral population which are somewhat lower than if we had been able to exclude the people in the general population over, 75. A comparison between the asbestos miners and the population of the Dominion as a whole was made, using statistical material from several sources. In one source. Phil lips 70 gave age- and sex-specific rates for Canada for three periods between 1931 and 1952. The rates for males are given in Table 20. These figures show strikingly the increase in rates between 1931 and 1952. and this A}!* Group Under 30 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85+ All Ages 1931-1933 0.2 0.9 2.0 3.0 5.2 8.2 12.7 12.5 15.1 10.7 15.8 8.2 11.4 2.8 Gruups o(Years 1941-1943 0.3 0.9 3.2 5.4 12.5 18.9 27.0 33.7 34.5 ,*30.6 30.0 27.9 14.3 7.1 1950-1952 1.0 0.8 3.0 6.5 16.7 37.2 59.7 77.8 102.9 86.3 83.9 59.7 71.0 15.8 . * "Mortality from Lung Cancer in Canada," 1031 to 1833. increase is particularly marked after age 50, confirming an observation previously made, to the effect that until recently lung cancer has' probably been underdiagnosed in the older age groups in the general popu lation. To use these figures for purposes of comparison, it is necessary to combine the rates for certain age groups in order to conform to the age distributions used in this study. Since the exact populations in each age. group for the years indicated is not known, this must be an approximation. However, the rates would be somewhat as follows: Age Group 20-44 45-54 55-64 65 + Ratr 5 n7 69 OQ-95 These rates are. in general, lower than, those developed for the total (proved and suspected 1 cases of lung cancer among the asbestos miners. The only large difference, however, is in the age group of 65 years and over, and it is finite possible that the rate for- this---group*- may have increased for Canada between 1952 and 1954 as it did for the Province of Quebec (Table 17). A further comparison has been made with an fiver-all rate obtained from the American Cancer Society for respiratory cancer deaths in Canada ifi 1953. This rate, for males, is 20.8 per 100,000. or 5 more per 100,000 than Phillips* 1950-1952 rate, and compares with 25.3 per 100.000 for proved cases ami 33.8 per 100,000 for total cases among the Braun--Truan 509107.87 647 A. M. A. ARCH11'US OR INDUSTRIAL HEALTH Table 21.--Number of Deaths and Death Rates per lOOfiOO by Age Groups for the Adult Male Population of the United States* way, and other countries, at least 50,000 persons must be exposed throughout the world, and it can be assumed that this Ate aroup Population Cases Rat* per 100,000 number has been fairly constant in the 20 20-44 45-54 55-64 + Total 24*544,000 8,065,000 6,240,000 5,670,000 44,619,000 883 2.979 654 6,483 16,399 - 3.6 36 96.6 1MJ 37.2 years since 1935 when the first case of asbestosis with lung cancer was reported. At least 1,000,000 man-years of exposure has thus been accumulated, and this figure Data from "Vital Statistic* ot the United State*," Vol. 1 can be divided by the approximately 150 and 2, 1852. cases of lung cancer with asbestosis re asbestos miners in this study. It is there ported during the 20-yedr period. This fore obvious that there are no important gives a rate of 15 per 100,000, which is at differences between the rates for asbestos least indicative that any lung cancer rate miners and those for the general population which can be calculated for workers ex of Quebec and the Dominion of Canada. posed to asbestos dust is not much greater Since it is probable that figures for the than that for the unexposed population. United States are more complete and, there Comparison Between Eight Counties Ad fore, possibly more comparable to the data jacent to the Asbestos-Producing Areas and for the miners, age-specific rates were com Eight Selected Counties.--To compare lung puted from "Vital Statistics of the United cancer mortality rates in the counties sur States," Volumes I and II, for 1952. These rounding the asbestos-producing areas with rates have been tabulated in Table 21. another, group of counties in which no as It is apparent that these rates compare bestos miners are likely to reside, the rates favorably with those for the asbestos miners were computed on the basis of figures for as shown in Table 8. Still other rates for the years 1950 through 1955. The eight the United States were obtained from the counties selected for comparison were Ar- American Cancer Society, and for males, genteuil, Chateaugay, Montmagny, Port- these were 25.3 per 100,000 in 1953, and neuf, Richlieu, Riviere-du-Loup, St. 28.0 per 100,000 in 1955. They are not' Hyacinthe, and Terrebonne, mainly because identical with the rate calculated from the* they represent a wide geographic distribu figures of the office of Vital Statistics, but tion throughout the Province. The counties this is possibly because-the5 American-Can-- selected because of their proximity to the cer Society rates are for males of all. ages. asbestos mines include Arthabaska, Beauce, Nevertheless, they, too, compare favorably with the rates of 25 (or 34 for total cases) Drummond, Frontenac, Megantic, Rich obtaining among the asbestos miners. mond, Sherbrooke, and Wolfe. Table 22 Turning for a moment to a comparison shows the number of lung cancer deaths between the asbestos miners. and_persons_ for the years 1950 through 1955 for each who are exposed to asbestos-in-one-form - of these counties, and a mortality rate, or another (as distinguished from the gen based on the adult male population in 1952. eral population groups- jxfsr diseased; wlro- To- emphasize the comparison, Megantic have no exposure) an interesting observa County has been shown separately, as has tion can be developed by deduction. Hue- the Province of Quebec and also the Prov per 58 has stated that there are about 35,000 ince with the eight "asbestos-producing" I>ersons exposed in the United States, and counties subtracted. Because of its unique we have found that the Canadian mines lung cancer death rate, Montreal et Isle de employ about 8000. Elsewhere, it has been Jesus has also been listed in order to provide estimated that the workers in England who further comparison. have exposure total between 3000 and 5000. It is apparent from the table that the With workers in Africa, Denmark, Nor lung cancer death rate for the eight counties <48 Vol. IT, June, 1958 50010188 LUNG CANCER IN ASBESTOS MINERS Table 22.--Number of Lung Cancer Deaths and Rate per lOOJOOO Man-Years Counties Male Lung Cancer Deaths * Adult Male Population Kate per 1963 1930 1951 1032 1953 1954 1955 Total 100.000 Metarule County 13,100 3 l 3 3 1 4 15 Eight "Adjacent" Counties 97,NX) 6 3 9 16 4 16 54 Eight Selected Counties 83,000 3 10 3 18 5 0 49 Province of Quebec 1,198,000 196 330 243 303 303 357 1624 Province of Quebec less eight "adjacent" counties 1.100.000 190 217 236 287 399 341 1570 Montreal et Isle de Jmuj 394.000 3 7 158 193 185 235 770 18.9 9.4 9.8 22.6 23A 334 It Is assumed that all male lung cancer deaths occurred alter age 20. immediately surrounding the asbestos-pro ducing areas is practically identical with that of eight counties selected for comparison. While Megantic County has a rate nearly twice that of the combined eight selected counties, it is lower than the rate for the Province, and considerably lower than the rate for Montreal. The figure for Montreal would certainly be higher except for the very low numbers of deaths reported for 1950 and 1951, and it would appear that in those years some error in reporting has undoubtedly been made. On the basis of the other years, 1950 and 1951 deaths would be expected to be about 200 greater. This would result in a rate of 40 per 100,000.. . The only possible conclusion from this comparison is that there is no evidence'that the persons who live and work in the coun ties surrounding and adjacent to the as bestos-producing areas have any greater incidence of lung cancer than those who live elsewhere in the Province. ' Comment on All Recorded Lung CancerCases, Living and Dead, among the Asb'es-' tos Miners.--Although a simple enumera tion of all the known or suspected cases of cancer of the lung in these areas has no particular value from a'statistical point of view, it is of interest to summarize-such cases for the record. There were nine deaths prior to the beginning of the time period covered by the study, including one in which the diagnosis was mediastinal lymphosar coma. During the period covered by this investigation, there were nine proved cases and three suspected cases in the cohort. Through 1956 and to date in 1957, there were eight deaths, six of which were merely suggestive of cancer of the lung and in cluded such diagnose s'" as mediastinal lymphosarcoma, mesothelioma, cancer of the leg with metastases to lung, abscess of lung, and cancer of the pancreas. One other was diagnosed on the basis of x-ray only. In addition, there are now living four cases in which the diagnostic evidence is strongly suggestive of lung cancer. This is a total of 33 cases of all types, including 10 "sus pected" but unproved cases, and 4 that are still living. The remaining 19 constitute the total.of proved cases of cancer of the lung among the asbestos miners since 1940. The proved cases averaged 59 years of age at death, and varied between 37 years and 68 years. Their working span covered periods varying from a minimum of 14 years to a maximum of 37 years. Only three men had less than 25 years of em ployment in the industry. Seven among those on whom such information is avail able had a weighted exposure placing them in Category III, and six worked in an exposure represented by Category I. There were only 17 among these proved lung cancer cases in which we have in formation regarding the presence of asbestosis. Asbestosis was present in nine, although it was minimal in two. Two path ologists disagreed regarding its presence in another. At least seven of the 19 proved lung cancers, therefore, were not accom panied by asbestosis. Summary and Conclusions Interest in the question of whether there may be an association between lung cancer Braun--Truan . 6^ 50010189 A. M. .A. ARCHIVES OF INDUSTRIAL HEALTH and exposure to asbestos has been evident since the report in 1935 by Lynch and Smith of a case in which lung cancer and asbestosis were both present; As additional cases in which the two diseases coexisted were reported, a causal association appears to have been gradually accepted by many authors, although some workers considered the correlation to be inconclusive. The pres ent study was undertaken in an effort to de termine whether a causal relationship did, in fact, exist between exposure to asbestos and cancer of the lung. Since most earlier studies had been was not possible to trace 133 of these for the whole period, but 5771 of the remaining 5958 were found to be still living in 1955 or later. Of the 187 known dead, cancer of the lung was considered to have been reasonably proved in 9 and to be strongly suggested in 3. The members of the cohort were studied with respect to age, length of employment, a weighted average of their exposure, and their smoking habits. It was found that 4673 were smokers within the definition of that term as used in this study. Thirty-four per cent of the cohort were more than 45 limited to enumerating the lung cancers years of age, and thirty per cent had been found in certain selected samples, such as employed for longer than 20 years. Thirty cases coming to autopsy or death certificates per cent had a weighted exposure which in which asbestosis was mentioned, it was placed them in the category of highest ex apparent that they could not fulfill the re posure. quirements of an- epidemiological- and sta tistical approach to the problem. The present study was, therefore, designed to meet the requirements of this method. After a preliminary survey'to'explore the * availability of reliable information, data were gathered on workers in the asbestos mines in Quebec, based- on their- medical records. A cohort was defined as a group of asbestos miners having at least five years of exposure and who were in the industry in 1950. Data relative to their characteristics were collected and their.status- at the- endof a six-year period of observation was de termined. In the case of those who had died, an exhaustive search of death certif icates and insurance records was carried out The mortality rate for lung cancer, as computed on the basis of nine "proved" deaths among the cohort was 25.3 per 100, 000. When the three "suspected" cases were added, the "total" rate for the cohort rose to 33.8. The importance of the suspected but improved cases in determining these rates has been reiterated because it is likely that such cases would not be included in the statistics for the general population and because they influence the results so mark edly. According to the findings in this study, the mortality rate from lung cancer does not appear to increase with length of ex posure or with degree of exposure, a fact which presents strong evidence against the in order to determine as nearly as possible- carcinogenicity of asbestos. the exact cause of death; Mortality rates Comparison of the experience among the from lung cancer for the generaLpopulation . -asbestos miners with that of various seg of the Province of Quebec-and its various ments of the unexposed, comparable popu counties and for the Dominion of Canada, lation shows that the observed number of as well as the-United-States were calculated deaths among the miners is not significantly from statistics collected in the appropriate greater than the expected number. The rate places. Comparisons of the rates obtained for proved cases among the asbestos miners for asbestos workers and for the other popu (25.3 per 100,000) compares well with the lation groups were made according to ac rate of 22.5 per 100,000 for the rest of the cepted statistical methods. Province, and 20.8 per 100,000 for adult Records were obtained on 6091 persons males throughout the Dominion of Canada. who fulfilled the criteria of the cohort. It It also compares satisfactorily with rates 650 Vol. 17, June, 195S 50010190 LUNG CANCER IN ASBESTOS MINERS of 37.2,25.3, and 28.0 obtained from various sources for adult males in the United States. Finally, in this matter of comparison, it would appear that the world-wide experience of persons exposed to asbestos dust is not worse with respect to lung cancer than that of the unexposed population. The counties surrounding the asbestos- producing areas, in which it -is presumed most of the asbestos miners live, have al most identical mortality rates with those of eight counties widely scattered through the Province, and are lower than those for the remainder of the Province, and much lower than the rate for Montreal. Since 1940 there have been 19 cases in which the diagnosis of primary cancer of the lung may be considered to have been proved. Approximately half of these cases were associated with asbestosis. All but one died in the recognized "cancer-age" and at least one-third had only the lightest ex posure (Category I) to asbestos dust. . On the basis of what are believed to be complete and reliable data, it seems fair to conclude that the asbestos" miners in the Province of Quebec do not have a signif icantly higher death rate from lung cancer than do comparable segments of the general Imputation. '' " ' Furthermore, the death- rate from lung cancer in the areas contiguous to the asbes tos operations-is comparabie"to"that in areas widely scattered throughout the-Province of Quebec and is-lower than- in -some-urban ized areas within the Province., REFERENCES 1. Alascio Escobar, R.: Bronchial-Xarcinomar Review of 200 Cases, J. Intemat. Coll. Surgeons 26:375-379 (Sept.) 1956. 2. Allen, M. L.: Bronchiogenic Carcinoma As sociated with Pneunionoconiosis: Report of 2 Cases, J. Indust. Hyg. 16:346-347 (Nov.) 1934. 3. Cartier, P.: A Contribution to the Study of Asbestosis, Arch. mal. profess. 10:589-595, 1949. 4. Anderson, C. S., and Dible, J. H.: Silicosis and Carcinoma of the Lung, J. Hyg. 38:185-204 l March) 1938. 5. Baader, E. W.: Asbestosis, Deutsche med. Wchnschr. 65:407-408 (March 17) 1939. 6. Behrens, W.: Experimental Asbestosis, Schweiz. Ztschr. allg. Path. 14:275-297, 1951. 7. Behrens, W., Jr.: The Clinical Picture and Pathology of Asbestosis, Ztschr. Unfallmed. u. Berufskrankh. 45:129-140 (June 15) 1952. 8. Berblinger, W.: Increase of Lung Cancer and Diseases Due to Dust Inhalation, Med. Klin. 27:1337-1342 (Sept II) 1931. 9. Berenblum, I.: Irritation and Carcinogenesis, Arch. Path. 38:233-244 (Oct) 1944. 10. Bohne: Asbestosis, Deutsche med. Wchnschr. 62:928-930 (June 5) 1936. 11. Bohme, A.: Results of periodical Examina tions of Workers in an Asbestos Factory, Beitr. Silikose Forsch. 11:34, 1951. 12 Bowles, O.: Asbestos-Milling, Marketing and Fabrication, Information Circular No. 6869, U. S. Department of the Interior, Bureau of Mines, 1935, pp. 1-26. 13. Bresiow, L.; Hoaglin, L.; Rasmussen, G., and Abrams, H. K.: Occupations and Cigarette Smoking as Factors in Lung Cancer, Am. J. Pub. Health. 44:171-181 (Feb.) 1954. 14. Bristol, L. J.: Roentgenologic Aspects of Silicosis and Asbestosis, A. M. A. Arch. Indust. Health 11.-189-195 (March) 1955. 15. Cancer of the Lung: An Evaluation of the Problem, Proceedings of the Scientific Section, Annual Meeting, American Cancer Society, Inc., Nov 3-4, 1953, New York, American Cancer Society, Inc., 1956. 16. Cartier, P,: Asbestosis Cancer of the Lung, in discussion on Smith, W. E.: Survey of Some Current British and European Studies of Occupa tional Tumor, Problems, A. M. A. Arch. Indust Hyg. 5:262-263, 1952. 17. Cartier, P.: Some Clinical Observations of Asbestosis in Mine and Mill Workers, A. M. A. Arch. Indust. Health 11:204-207 (March) 1955. 18. Clerens, J.: Research into Pulmonary Asbes tosis in Belgium, Arch, beiges med. Sociale 8:557 565 (Nov.) 1951. 19. Clynes: Asbestosis and Silicosis, Brit. M. J. 1:379, 1931. 20. Coffin, G. J.; Duryee, H. C.; Maier, H. C.; Pardee, H. E. B., and Wymer, E. L.: The Effects of Tobacco Smoking, Panel Meeting, Bull. New York Acad. Med. 32:133-156 (Feb.) 1956. 21. Cohart E. M.: Lung Cancer and Economic Status, Cancer 8:1126-1129 (Nov.-Dee.) 1955. 22. Cooke, W. E.: Pulmonary Asbestosis, Brit. M. J. 2:1024-1025 (Dec. 3) 1927. 23. Cureton, R. J. R.: Squamous Cell Carcinoma Occurring in Asbestosis of the Lung, Brit. J. Cancer 2:249-253 (Sept.) 1948. 24. Cutler, S. J.; Schneiderman, M. A., and Greenhouse, S. W.: Some Statistical Considera- Bratal--Truait 651 500*0191