Document 2R8EQqz9yr0rE9oyMpvwg2YXN

1 PLAINTIFFS 1i XEfXlHF IBIlTZ^ . .A / AS EPIDEJ.UODOGICAL STUDY OF LUNG CANCER IS ASDESTC: MINERS Daniel C. Braun, M. D. / y Medical Director, industrial Hygier.e Touacation ' f' Ad T. David Truan. /'I. VStalistical Consultant, Industrial Jiy^isiie foundation Ever aiaca the pronounced laerease in the iacidcace of luag cs&ccr among males became apparent, there have been attempt* to atseclate It with oae or aaothsr of the varioua elcmcoto in the eavironrr.ea: of rr.ao. The ap proach uses by aeme 'worker* ha a been to auapect oae or several substances sad then act about in aa iateaaive eeareh for lur. cancer among person* 'ho tSi} have bed any expoiure to those material*. In tils eor.sccf.cn. Smith write*; "The tendency of author* reporting the coincidca:al occurrence o; pri mary lung eaneer with sllleosis or with nay other theoretical etlolo^ie eon* dittoes, ha* beta to emphatlxe the percentage relationship In extremely email erica of ca*ee, with eoatrel eaee* which arc not In any way eomparablc." It would ietm inevitable that aabetto* should come under acrutlny In thic menner, because prolonged exposure to this material U known to cause a specific type of pncumoeoalo*!*. aad because persona who show this form This study wm made possible through 4 grant from the Guiibcc Asbestos Mining Association. 0250 / 2. of pneumoconiosis oftea corn* to autopsy aad provide a reedy soures of me* tsrlal for study, U woe la thla way that report* of the simultaneous occur* reace of luag cancer cad asbestos!* began to accumulate after the report of a caee by Lynch and Smlth*68^ In 193$. Within tho next tea year*, about 25 addUtoaal case* were reported, sad 1b id54 Merewetber^78^ reviewed ail death* from eabeetotl* recorded la ZagUad *loc* March, 1924, Luag can* cor occurred la 16% of the*t case*. Gloyne/*^ whose work l* site frequently referred to a* establishing a eoaaecttos between aebestoele aad cancer of ibo luag, reported la 1951 that caacor of tho luag was present la 14.1% of asbes* to*l* case* examined by him, la 1941. Nordmaaa and'Sorgo (T&l claimed to have produced luag caacer la mice which they exposed to aebestoe duet. Slaco 1951, additional caeee of cancer of the luag coexlstlag with ($81 aebeetoeie have beea reported, aad. aceordlag to Hueper about 100 eueh caeee had beea reported up to 1955. As a result, aa association between the two diseases appears to havo boon accepted by many authors, aad several (89) writers wsre using the term "aebestoele caacer" of the lung, Werber la 1952, stated categorically that la 7% to 27% of caeos of aebestoste, after a latest ported of about 1 1/2 to 20 years,' caretaoma becomes established la tho luag, Qa tho other hand, not alt authors accepted this alleged association (80) without reservatioo. Soups la 1939 reported that he had discovered no eases of luag caacer amoog 520 cases of asbestos!* which hs bad examined: (48) aad la 1942, Jlolleb and Aogrist expressed the opinion that tht number 0251 1. of case* of asbeetoai* with lung caneor wae too small for statistical evalua tion. In 1947, Wegeliua*0* reported 116 radlologically diagnosed case* of asbestos!* among 476 worker* In rialaad, and found no aae of tung cancer la this group* Goldblatt and Goldblatt la tbelr section of Merewether'e latest book/*^ tuts: "But at no stage la all these impreeelvo researches was any due obtained which might have offered any support to the possibility that asbestos could act ae a carcinogen. There 1* no reliable criterion by which one can anticipate carcinogenicity and, at la wall known, relatively minute changes in the structure of a chemical carcinogen are sufficient to diminish or eliminate carcinogenic action. If asbestos is Indeed to bo regarded ae a carcinogen, the need is felt to demonstrate some property which can be regerded ae something more than Inertness. These authors advance the theory that, until seme more experimental evidence of direct carcinogenesis by asbestos or a dlcomposUion produce'of It can ho obtained, asbestos might be considered as a "co-carcinogen" which only Induce# a further development of a preneoplastlc condition brought ab>ut by something independent ef the asbestos, such as an endogenous factor. Thus the literature, while tending to support the thesis thst asbeetosis Is In seme way related to the development of lung cancer, is by no means unanimous. Altogether. U is perhaps more confusing than enlightening. A careful review shows that the majority of the reports are clinical and not epidemiological. They lack many elements necessary for the application of 4. epidemiological techniques to their content and moit of the author* do sot make claim to having done so. What ha* happened la that succeeding authors have drawn conclusion* aad genertllsed boyoad tho scope of the works which they quote, Nowhere, for example, have we found roforeacs* to a population of asbestos workers, although several authors who have quoted the observed iacldeace of lu&g cancer is autopeloe of pereoos who also had asbeetotie imply that this lac Id toes applies to asbestos workers, generally. We have likewise beta usable to find asy study which actually calculated the l&cldeoce of laag cascer sraoag a populatloa of per boos who had asbestosls. aad aot just those who earn* to autopsy. With the esceptioa of a paper by" Doll aene of theee reviewed gave asy data oa exposure aad duet concentrations, aad evea Soil*# paper merely mentions '`scheduled" areae, by which Is meant, "these srsae where processes are carried oa which wore scheduled usder the Asbss- tee Industry Regulations of 1931 as being dusty," Thsrs is, furthermore, s complete lack of definition of terms as seed is the published literature, For example, the term "asbestosle", as seed, may refer to changes observable only by microscopic examination of the lu&g tissue, or it may meas a radiological!/ detectable condition. Most of the published reports obviously i&cludod women among their cases, but seme of them do aot give the number or proportion of women In* volvod la the*study. There is alee * lack of uniformity as to what typo of oxposuro most studies have dealt with. Of 99 cases enumerated by Huepsr^^ in 1955, oni? 0253 5. ten appear to have originated in (he United States, and itvn in Canada. Sortie af the eajller reporta apparently Included aabeatee miner*, but It can be sssumed, since 82 of the 99 caaea had originated In England, and aiace oo aebeatoa mining operationa are carried on la that country, that moat of the reported caeea have Involved worker* In the textile or fabricating industries. Such faetora aa amoklng babltt, family hletory of cancer, length of time la the laduetry, and age of the Individual cate are alao notably abeent la the majority of thee* reporta. With thle uadaratandlng of the Umi^tien* of the exiatlng literature with reapect to epidemiological generalisation, It may be of value to conalder la aomowbat more detail aome representative earlier publications, a few of which were referred to briefly above. One of tbe meet detailed etudle* aad one which deeerves the most aertoua consideration ie that reported by Doll^a^ In 1955. This study reviews cautos of death among aabeetet workers based on Coroners' records. It Also Attempts to estimate tbo risk by studying records of men who worked for at least 20 years la exposed situations. DoU concluded that lung cancer wee a specific industrial haaard of certain aabeetoe workers aad that, after 20 years of exposure, the risk is ten timet as great as for the general popu. letlon. This article Is Importaet for several roeeoos. la addition to the definite conclusions at which U arrives. For example, U begins by stating that "In view of the infree.uency of atbeetotle, thle large number of ease* 0254 5. (61 eases e/ lung caactr) suggests--but doe* not prove--that lung esccsr is 4a occupational hazard of asbestos worker*. " Neither this artlclo or any pravioua oaa which wa hav examined praaanta any figures to prove that as- bestosls la an infrequent occurrenca. Estimates of the number of portent potentially expoeed to aabeatoa duet is the United Statae alone vary from 20,000 to 35, 000 and the iacideoce of atbettotla of any degree might be hlgbor thaa Doll imagines. This study, like ao many othera. Involvet autopey recorda. The number of peraone Involved la the etatiatieal aaalyala ie oaly U3 repreaentlng only 1,042.35 man years of life. It la alao true that Co selecting men who had been employed for at leaat 20 years, the study automatically excluded those who died from other causes after shorter employment. Another reason why this publication (a of tmportaaca ie a statement which it contains to the effect that "the strongest evidence that it (lung cancer) stay be a hasard (la aabeatoe workers) has been produced by Marswother and by Gioyae." la 1951 Gloyae presented a review of 120$ autopsies ea persons who had worked la various dusty occupations. This number Included 132 sebeetoe workers, of whom 121 showed "pneumoconiosis"--presumably as- bsstosis. Primary cancer of the lung occurred 1? times In this group, an incidence rate of 14.1% for lung cancer among autooslcd asbeetoela cases. There were in his series 796 caeee with silicosis, and 6,971 of those also showed printary eancer of the lung. The incidence of lung cancer la other 0255 7. farm* of pneumoconlosts wai 8. 7%, and to ll>9 cates which proved not to have any type*; pneumoconiosis U wu 8.3%. Cloyse considered "the mortality *f the asbestos workers" to be "disturbing". First of all. U la obvious that the paper does oot deal with the "mortality of asbestos workers", and secondly, it must bo borne In mind that all of Or. Cloyne's cases were submitted to him for study because the finding# were unusual (or uncomplicated pneumoconiosis. U can reasonably be assumed that cases, including those of asbestosis. in which the findings were not considered unuaunl were not sent to Or. Gloyne for examination. As a mattsr of fact, in the same paragraph in which ho ex* presets concern over the Incidence rate in asbestosis. Or. Gloyne himself poibts out that the rate for lung cancer based on aecropsiee at the London Chest Hospital was 21, 3% whlls ths figures of the Registrar-Csneral showed only 2,4%. He thus recognised that autopsies on a certain selected group of eases were not representative of the general population, It would seem, then, that notwithstanding the vain# of Or, Cloyne's work, (ts importance as an lades of the prevalence of lung cancer la asbestotlcs has been misinterpreted by some who have quoted him. All that It really shows Is the (set that in a group of 121 easea, selected for special study primarily because they seemed abnormal by prallminary examination, 1?, or 14.1% had lung canesr. Uorewcihar In 1947, in the report of the Chief Inspector of rectories, reviewed all cases reported between 1924 and 194b in which aa- bestesis was the caese of death or a coexisting condition. This work was lattr extended to include all such casas reported up to December, 1994. by 0256 i. which time there were *44 deaths Including 205 metes end 130 females. Among them were 55 case* (16?.) of eencer of the lung, 41 In melee ar.J U is females. It ie quite poeelble that * largo number of aebeetotiee who did not die of their asbestosl*. or in whose death certificate It wee sot mentioned, map have been mleaed. The import of thla 16? Is enhanced by the simul taneous statement that thw incidence of lung cancer in autopsies of the general population it only 1'7. The danger of attempting to compare a rate found in 344 cases with the rate for the general population without reepoet to age, occupation, and many other variables, such as smoking habits. Is obvious. Lynch ^^who wtth SmUa had reportod-the first case In 1935, reported four cases of carcinoma of the lung In a series of 49 autopsies on workers Is an asbestos manufacturing plant who wore shown to have "demon strable deposits of asbestos in the lungs. " This, of course, is not neces sarily identical with the disease asbeatosis. Lynch, himaali, point* out that, althoegh this is as incidence of 8, 2%, "both figures are too small for very orioue statistical types of calculation." Nevertheless, later writers have need this paper to strengthen the case for as association of carcinoma of the lung with asbeatosis. It is also of interest that KUtx found only the same number of cates of lung eencer in a series nearly 10 times as Urge, l. e.. four is 478 cases of asbeatosis. Behrens, ae quoted In Merewether l7^), estimated that, of 309 cases ef asbestosl# in the literature, 44 shewed associated cancer of the lung--giving an incidence of 14.27,. This is an Illustration of generalising an incidence 0257 9, obtained ir. a ^rcup which r.-a?: uudaubicdly rcpsrtco uni,- bee.'.^sc some of them shewed lung cancer, to possibly hundreds of asbcstotlcs whose tksca were never reported. The same applies to tb* conclusion of TcleJ:-/5) who appear* to hav. reviewed report* f i*i nutopsiea on persona with ssbos- tests among which six cases of lu03 cancer occurred. Information from sources such as these does not justify generalisation* with regard to mortalityrates. Perhaps no one has written to extensively on the subject ae has lluepcr^2"*^. In 1955 he reviewed the cases^^ reported prior to that data (28] and enumerated a total of .vi. Eleven ef thcoe were thoeo discussed by Toll and mt; havo been cases covered by other authors. Eight were discovered {bJl by Itcnnaway and Xennaway in an analysis of death certificates and. un* less Meriwether's study was Incomplete, these eases sbeuld have been in* eluded In hia report. Of the remaining 0, it ie quite possible that tho 11 contributed by Marewether and the 1? by Cloyse contain aome duplication with each ether or with these of other English authors. Principles of the Epidemiological Method Dore^*0) has poietod out that much of what is sow thought to be per* tinent concerning the comparative frequency of lung cancer in different popu lation groups has boca developed from tho analysis of clinical material, par ticularly surgical and autopsy records, supplemented to seme extent by the reported impression* (if various cllniclane baeud upon their personal obser vations. More recently, however, attention has turned to the systematic .ID (Aveaugatioa of this problem by the same methods that have proved so sue- asssful previously la tbo atudy of communicable claeaees, that is to aay, by epidemiological methods, la order to apply tbla method of investigation to the problem under flscuealon, we ware of .the opialoa that 4 atudy should ba planned ao aa to provida: 1. A well-defined population group, 2. Available data for all membera of tbla pop ulation, including tbo baaliby aa wall aa riba ill. 3. A enmple which la truly rapraaaatatlva of tba popwlatloa. 4. RollabU and valid obaarvatioaa relating to tbo problem of tbo atudy. . A aoticua defect, common to moat of tba etudlee which have bean reported, la that Hula or ao Information concerning tba healthy people la the )roup aaoma to have boon available to the euthor, la order to draw a generalisation regarding all aabaatoa worker#. It la noceeoary for a atudy to Include living pa none aa wall aa tbo dead. Limiting tba lavattlgatloa to automated eaaaa, aa haa been frequently dona in earlier etudlee, still further reatricta tta uae in gaaeraliaation. Tba problem with which we are concerned It whether aabeetoe mlaere experience more lung cancer thaa doce the general population, Tbo answer necessitates tbo collection of reliable Information on tebeetos miners aa a group, aa wall aa on the general population. 0259 ii. It dm* advisable to discus* the differences between tho epldemlological approach aad that used l& tbo studies which have bo roported to data. A very Important consideration Is tbs fact that lu&g cancer, la spite of Its Increasing numbers. U stUl a disease of low Incidence; that Is. In a given population not many persone will contract this particular dlseaee. Tils fact requires that Urge samples or groups must bo studied to provide meaningful results. Recognising tbo difficulty of obtaining such large samples, most arller writers deviated from the epidemiological method and sought to cir cumvent the requirement of observing well persone by:: 1, Comparing the relative^reouone^ of caneer in various eltee, i* Comparing tho relative frequency of esneer In a group of hnspUnllnsd patients, S, Compering tho relative frequency of caneer in a group of cases coming to autopsy. Attempting to compare two population groups, looking only at the reUtlve frequency of cancer la verioue body sites, may result in finding e higher percentage (relative frequency) in one of the groups, when, in fact, the mortality veto of cancer of a particular organ is oxaetly the same In both groups. This Is very eloat ty demonstreted in the excellent article by Dora('331 , The mortality rets from a particular eause le the true measure of comparl* son. ' 0260 u. It is ??*r*nt that selected groups such * hosi'icallsed patients or autopey'casss may not be Is any way representative of a larger group, and that is dealing with such samples, the observer may easily find more cases of a given disease than would be found is another group of the same sUo, but representative of the general population. It la true that Investigation of cases from such n sample can furnish information valuable for research, but the use of this Information In drawing generalisations Is necessarily rs* striated. It la the obligation of both the Investigator and of those who read his report to make proper comparisons and to draw only those conclusions which are valid and justified. A good statistical study of cases of eaacer of the lung occurring la a group of autopsies can lead to a proper inference concerning the frequency of lung cancer among coses coming to autopsy, but only to such cases. Far Information from such a study to be projected to seme larger group, U Is necessary that the autopsies represent a good sample of that larger group. To assume that such is the case la any particular serise is dangerous and likely to bo false. There is some danger that the figures reported by some authors may be misconstrued as applying to asbestos workers or even asbestos miners, when, In fact, the authors In question do not make this generalisation, nor can the generalisation be made f<-r the reasons stated. Cloae study of the reports reveals that the percentages quoted relate only to the group of autopsies covered by the particular investigation. 0261 The preterit study, In contrast to tho earlier works, has bees plan ned to utilize the epidemiological method, A well-defined group of asbestos miner* Lee bees established lo tuch a way that It constitutes a good temple cl the whole population of aibeetos miners la Quebec, Data for all members of (hie group have bees collected and analysed, Those concerning lung can cer have received most careful consideration. Details of the methods em ployed will be set forth later, but the type of approach le considered to per mit of fair comparisons and valid generalization*. Collection and A&alyate of Data A preliminary survey of potential sources of information in Febru ary of 1956 involved dlscuseioni with the physician* la charge of the aebesto* companies' programs and with clinicians, pathologists, representatives of CUy and Provincial health departments and of the Canadian Cancer Society, end other Interested persons. It was found that morbidity data, although some what limited, were available from such sources a* the hospitals la Montreal and (Quebec City, and tbs 13 Cancsr Detection Centers In the Province, How ever, because of the high mortality in lung cancer, it seemed advisable to depend upon data relating to deaths. Thsss we found to be obtainable at the vital statistics department of the Ministry of Health in Quebec CUy. From the preliminary survey, U was apparent that extensive and detailed information could be gathered with respect to both the persons employed la the asbestos mining Industry and mortality figures for the general population. 02S2 14. Following this exploratory survey, the Initial effort '*+* t\rectse to the collection of dot* relating to all worker* who had been processed through the clinic at Toetford Mine* slace It* Inception la 1947, end similar Worn*.*. Ilea regarding all workers at Asbestos, P.Q. Data from the clinical records Included the age, family aad personal medical histories, smoking habits, number of years of exposure, an estimate of weighted exposure, aad the courso of the Individual's health status or the cause of his dsoth. From this information It was possible to formulate a "cohort" which eeuld be well defined, should bo representative of thg whole group, aad could be followed for a definite period of time. All of the available experience indi cates that the development of asboetosis la Iocs than five years of exposure must be somewhat rare. Accordingly, the cohort waa defined as including every w*#*ev who bad a total exposure of five or more years, aad who was on Offttesh* tr*<* fc*i***, #fJ **,Le- th* employment rolls in 19$0.4Tbi cohort waa then followed by mean* of O*M/t<C //e.Xvy'fV. the annual physical examination records through a six-year interval, 1953 through 195$. MX data regarding this group were then tabulated la order to determtao the characteristics of the cohort. For those who survived the en tire period, reference was made (o the physical examination results snd x-ray findings at the end of the period. Those who bed died were tabulated separately aad the cause of death corroborated by sxamiaatlou of the death certificates. A further search wee made concerning those la the original cohort who re- melned unaccounted for when the llvtag aad the known dead had been tabulated. 02S3 They repreeent men who had left employment through retirement or resigna tion. Eventually, all hut a small number of these were accounted for as either 15. living or dead, And in (he letter event, the cause of death a< subsiastiatec la 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 1*55, Inclusive, were reviewed in the department of vital statistics of the Provincial Health Ministry, together with statistical summaries of the cause* 4< ef deaths in the Province by counties. All cases in which death was certified ' as havin| been due to primary cancer of tho lung were examined for such in formation as place of residence, occupation, date of death, hospital in which death occurred, and whether or not an autopsy was performed. Cases in *e' 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 lung, were also rsviewed la an effort to include all instances of primary carcinoma of the lung in the '* study. dsatha The statistics for the Province of Quebec relate to population, total e*s t.^ , r * * r *?. <*< i * f ,^i ^ "*i --1 and deaths from lung cancer. These were collected aad tabulated by counties and by sex for ths years 1*50 to 1*55, Inclusive. * Prom them, death rates for the general population of Quebec aad of Individual counties were calculated for specific years aad analysed by cause. i. Practically all employees of one company are covered by a group policy of life insurance which, fortunately, nearly all of them continue to carry when they rettre. A very few are not covered by this policy, and these mho leave the Industry for one reason or another except retirement usually ( are am longer covered, but this is likewise a small number. As an additional CC64 living or dead, and In the Utter event. the cause ot death was substantiated in a similar manner, and the results added to the original list of deaths. Death certificates for the Province of Queboc for the years 1951 to 1955, inclusive, were reviewed In the department of vital statUtlcs of the Provincial Health Ministry, together with statistical summaries of the causes ei deaths in the Province by counties. AU chess In which death was certified as having been duo to primary cancer of the lung were examined for such In* formation as placs of residence, occupation, date of death, hospital in whleh death occurred, and whether or not an autopsy was performed. Cases In which lung cancer was given as a cause of death, but la which it was net specified ss to whether the cancer originated in the lung, were also reviewed (a an effort to include ail instances of primary carcinoma af the lung in the study. The statistics for the Province of Oueboc relate to population, total e*j , , fc .. r s,, f?.. n c*#*n.* ** ry/,i ^ deaths (romjaU^s^at and deaths from lung cancer. These wore collected sad tabulated by counties and by sex for the years 2950 to 2955, inclusive, from them, death rates for the general population of Ouebec and of individual counties were calculated for specific years and analysed by cause. Practically alt employees of ono compesy are covered by e group policy of life Insurance which, fortunetely, nearly all of them continue to cerry when they retire. A very few are not covered by this policy, and those oho leave the Industry for one reason or another except retirement usually ere ao longer covered, hut this is likewise a small number. As an additional \ 16. i'it check upon th* information obtained iron the clialeal reeord* ei'thi* group, the record* of Lit life insurance company wj examined lor all coatt cLiirr.* paid under the policy, end particular notice i takes of the claim* Is which . u. th* proof of death wa* bated os case*? of the lung. a I Death* from lung eaaeor among asbestos miser* were thu* deter* alsed from the clialeal record* la the medleal service of the industry and . /! checked by mesa* of the death certificate* aad laeuraaco company record*. The death* were thaa verified Individually by reviewing them with th* phyelclan* te'i la charge ef the medical aervlcee* la thl* manner, there wa* *tabll*hed a ld*r M li*t ef east* la which primary caacer ef th* lung i* considered to have bees proved ae th* eauee ef death. A few caee* la which luag caacer l* etroagly - 41 suspected but not proved aa the cauee ef death were ceealdered eeparetely, Mortality rate* have been calculated uelag both the "proved" aad the total of "proved" aad "suspected" eases during the year* under observation. Com parison* ware then made between the death rate* from the same cause among 44 specific segment* of unexposed person*. AU luag cancer deaths, both sus .tl pected aad proved, were carefully analysed to determine possible relation .*9 ship or correlation* between the development ef luag caacer aad any factor known from the clialeal record* such a* family history of caacer, personal history of heavy smoking, cooxlotenco of asbosto*!*, or exposure to asbestos. 3 la addition to this analysis of deaths occurring la the cohort and .* daring the yeare under observation, every known death from cancer of the '.t tang, aa well *e every case dlagnoeed but still living, has beas tabulated and 0266 17. analysed.^. They will be discussed separately from those Included in cb population aad time-interval under study. A csmpiiUw oi lung cancer mortality la the asbesto^produclng couailee bs been made with that la counties wbleb are far removed from lb* asbestos mines aad la wbleb. presumably, ae asbestos minora live. z1 (. riaally, la order to broadea the comparlsea of death rates la dif ferent populatlos groups, tbe rates have been eolleetad lor Canada generally, tad for tbs United States, according to the moat reeeat published aad unpub itl lished material. m; w Results and lateroretstion Tbe cohort which was constructed according to the criteria described la the preceding section has been considered Individually and compared with Ike general population. Description of tbe cohort will be presented here as . a preface to tbe results el the study. ` `1 Original Cohort Lost Persons Deducted Final Cohort Living la 195$ (working and retired) Used by 1955 Cancer el lung Questionable cancer el lung Other caueee Unknown caueee Smokers Noasmokers Unknown . . 6091 133 5956 5771 U7 9 3 169 6 4673 1265 20 0267 Table* 1( 3, and 4 prtiaiu age, number of years of employment, weighted Average exposure, and smoking habit* of the cohort. A comparison of the exposure to aabeetoe duet 1* presented In Table 3. Ch**r AU r--~Vr. were placed In one of throe categorlee, representing increasing degree* of exposure baeed on a weighted average of the year* epeat at varloue level* of dustiness. The degree of dustiness for each job category wa* deter, mined after consultation with person* familiar with the environment and eon. dUioa* la the various work area*. For th* purpose* of ealcuiatloa, the a*, sumption ha* been mad* that th* relationship between th*** eategori** Is linear, and that category 12 l* twice a* dusty, and catttory 112 three tlcr.ee a* dusty ae category-1. The fourth variable, smoking habits, wa* similarly tabulated and i shown in Tahl* 4 This wa* included because the Information wae available and because tracking was regarded as one of th* variables which, besides th* environment, could conceivably laflueneo the development of lung cancer. As seed la this presentation, the term smoker refers to a cigarette smoker, habit* ally smoking mors than fiv* cigarette* per day. Persons who smoke pipe* or etgars exclusively were net considered to be smokers for the purpose of this study. 0268 19. TABLE i Nunbar tnd Percentage Dletributiom by A| sir'Y t a: j . < i!'. 5^ *..t t*"' < 'il 20-44 41-54 55*64 65 4 Unknown Totals Avertge Ago Let* than ,5% Nuwbir 3901 1124 615 315 3 5956 41 Ptr Coat 66 19 10 5 100% * ;< *.* .. ' TABLE 2 Oi .1 X. -<ii Ai,W . ni htt:. k "i**.. * :*>I3 ' i* si*. * Number and Percentage Distribution by Length of Employment Length of Cmoloynifnt 5-9 10-19 20.29 30-39 40.49 50 Totals Avert;. Years of CnjtiOymtnC Numbor 1795 . 2396 922 603 165 55 5956 18 Per Cent 30 40 16 10 3 1 100% 0283 / -* . 20 TABUi 2 Nutabcr t&d Pertcaug o Distribution by Sxpoeurt Category xpour* Cttcgory Number Par Cent I 2021 24 a 2150 2b hi VataowA 1772 20 S'. /* Total* S9S5 100% Average wpo*uro *0 Lets th&ty 5 % - TABUS 4 Number *fid Pereeatago Distribution by Smoking Habits Smoking Habit* lk*r XtBinokin tlafeaowa Total* Un U*o . S % '* \ Number 467* 126S 20 S958 . Per Cent 7* 21 100% O' .  * n' / \ J* -- Table 3 ^riieou the year-by-year experience el the cohort and iadlcstes the number ef deaths each yaar from specified causes. la general, s eats was considered to be'preved" as ooa of primary cancer of tht lung vbta tha records showed that the diagnosis had bees supported by an autopsy or surgical rataceloa of tha lung with microscopic anamination of tha removed i r 7! tissue. Is osa case so eosnldorod* however* dlagaosls was eaafirmad by <U ir.iin J V broachostopy with visualisation aad biopsy. Is aaothor, although there ws t*v A so autopsy* tha diagnosis of primary eaacor of the luag seems to have beea i-i beyond question. TABUS 5 Yenr-by-Yaar Experience of Cohort aod Death Rates par 100*000 Mas*Years of Risk a* a #0 Tear No. Alive At Beginning of Year Proved Primary Ca. of Lune Causa of deepened Primary Ca. of Luna Death Other Causae Unknown Mas* Years of Risk 19 SO 595ft 1951 3952 1951 5922 1953 556$ ms 3848 1955 5810 Totala 1 ; .`*2 12 . l 5950 2 0 17 1 3932 0 ' .0 >7 - 0 5903.5 2 1 34 0 5866.5 1 0 33 4 5629 3 0 36 0 57*>0. 5 9 3 7S7~ T 35271.5 'Proved'* rate par 100, 000 maa*years of rlak -Total rate par 100 * 000 msn>yoare of rlak * -- - 25. 5 35271*5 12 35271.5 -- 34.0 0271 a -ni *9 4 til* 16 /eft Ad The term "suspected" primary cancer f (he lung u-aa appHe to (hole < la which the diagnosis remain* la doubt but lomt oi the evidence point* to cancer o the lung. There were three such case*. The term "man-year* of rlak" ha* beea uied to mesa the number of mea el rlek for the year under oheervaeton. An Individual who lived through* eel the year wae counted a a full man-year of rlek, hut oae who died durlag the year wee counted ae oae-half a man-year. Death* occurring la the cohort, and la which lung cancer le eoaeldered ' 4 , " 'Z-t-t-K 7 *''-* *ad tebeve been proved at a cauae, are ahow la iabuiMte^i.fe;ui bilu.-. Also ahown ara three deaths which have beta considered aa "auapeeted" lung cancer caasa. L % Nam* M.; ?I .1 V* :t t 1. A. 3, 2. N. P. 3. X- M. 4. C. B. s. A. S. 4. w. r. 7. L. D. 1. . 1*. 1. A. R. *f.4 l/J ^ ''Proved'' Caaee of Primary Cancer of the Lung Ago Cmokor Enooeuro Died Autonsv Aabeatoaia 64 Ye* 26 yra.la Cat.II 10.22.S1 Yea Yee 65 Yee 34 yra. la Cat. 1 5- 3-55 Yea No 65. Yoa 37 yra.la Cat. 1 7.20.55* No No 59 Yee 32 yra. la Cat. X 6.20*50 Yea . No 46 Yoa 22 yra.la Cat.IX. .51 Yea Yoa 61 Yes 33 yra. la Cat. XXX 4-30.53 Yea Yea 56 Yoa 30 yra.la Cat.XX 21-22-53 Yea Ye* 37 Yea 16 ) ra. in Cat. 2 6- 5-54 Yes No 66 Yee 26 vrs. In Cat.XXX $. .55 Droncb. No . />4us 7 "Suspected" Primary Cancer of the Luag Nam* Age Smoker r.eooiture Died Autoosy Aabeatoal* I. 3. T. 49 Yea 32 yra.la Cat.HI 10-25.50 No 2. A. D. 46 Yea 33 yra.la Cat.Ul 7- -53 Yea 3. P. P. 6$ Toe 42 yra.In Cat. II 9-30-55 IY* No Yaa .No 21. An indication of the Importance of these "suspected" casee la Interpreting the result* 0/ the calculation* 1* deslrsble before further discussion oi the mortality rate* which *7* derived in later table*. For example, U happen* that the rate found for ih* proved case* l* cloae to the "expected' rate ba**d on the general population figure* for the Province a* will be *bown later, aad on which baie, w* ehould find eight death* from lung cancer among the cohort. Actually, nine canoe were observed. If, however, the three additional "suspected" ease* are Included, Increasing this figure to 12, the total would bo very do*e to the 95% level of signif icance. However, having found just 12 caees, we are not above this level, and iri 11 u j n*.fcftJC'tii vij l .'.LSg the hypothesis that asbestos miners do not have a higher mortality from lung cancer than doe* a* *9T**>rtr9. tbs general population Nevertheless, the occurrence of 12 case# In thle sample would increase the rate to a point which approaches th* significant leval. Because of the tremendous importance of the questionable eaeee in this respect, some detail regarding them will be given hero. V /, In one of these eases, (he suspicion of cancer of the lung l# b*#d upon the x-ray interpretation and, although no autopsy vu performed, the death certificate indicate# that death was due to lung caaeer, it Is well )uaon that the x-ray appearance of fibrosis, aapocially U a localised denelty ora superimposed tuberculou* leeion It present, can e(mutate that of a tumor, and by Itself, dote not justify the Inclusion of this caeo as one of "proved" cancer of the lung. A second case waa certified as having died by reason of hydrothorax, possibly due to lung c&neor, but again there was no surgery and oo post-mortem examination, la the third ease, although It wae subjected to autopsy, rwo pathologists disagreed as to whether lung caaeer was present. The death waa certified as having been due to chronic myocarditis with neph ritic cad pulmonary congestion, and possibly cancer of tho lung. On the basis of these facte, It seems unwarranted to leelude these three cases Among "proved1* instances of lung caacar, Ob the other hand, they cannot, in fairtees, be disregarded completely. St Is for this reason that mortality rates have been calculated both ways. i Table gives the ratae by age groups. -5. vc* *-S*f /tt Group 20*44 43*51 5 S*64 65s Unknown Total* TABLE $ Number of persons and Number of Lung Cancer Deaths by Ags Groups Persons 3901 1124- 615 315 3 5953 Lv.rrCanccr Deaths Proved Total 1 1 l 3 3 3 4 5 0 0 T 12 < w. -d*. Annual Lung Cancer Death Ratos per 100, 000 Man*enrs of Exposure by Ago Groups j*Sif Are Group 20*44 43*54 53*64 45s Unknown TetaC* 4*4, Provod 4 15 61 212 0 ~zT * . . Total 4 44 61 265 0 34 Ths rates by length of employment ere sho\m la Table During the first 40 years of employment, the rate rises, an observation which seems plausible slsce the men were growing older. However, after 40 years of ex* posers there are no "proved" eases reported for a total of 240 men during the sis years, orxhout 1440 man-years of exposure. When the "suspected" 0275 26 ***** are added, e css* does shew up In this period. This would produce ft rate oi 69 per 100, OOO.agfttn demonstrating the importance of these question able. but unproved* cases to the final conclusloo, because. If there were ao eases la this number of men with leaf exposure, aad If asbestos Is a carcln- gsnle agent. It must be concluded that these 240 men have demonstrated eon* slderable reslstaaes. ThUjja biological pbenomeaoa which has been observed previously and (s coaslstsat with the theory of as Intrinsic or endogenous factor la cancer. The only other explanation would be that the susceptible members ; d of this age group had died earlier of lung cancer* Table which appears later la this section Indicates that the members of tho cohort did not die from luag cancer at a younger age than the general population. TA3L ? Number of Persons aad Number of Lung Cancer Deaths by Length of Employment Luna Lancer DeiU-i* Leseth oi r.*r;ioymcnt Persons Proved Totsi !- 10*19 *3-29 JO-39 40*49 60* Totals 1796 -- 2396 922 603 165 66 T^TT 0 0 . - 1. 1 3 3 .".3 7 .... 0 1 0 0 ? 12 Annual Lung Cancer Death Rates per 100.000 Man-Year* of Exposure by Length of Emolovment Length of Cmolovment Proved 5*9 10-19 20.29 30-39 40-49 $0* Over-all 0 7 64 166 0 0 ~ Total 0 7 $4 193 90 0 3-i C~76 TbJ rate# by weighted exposure arc ahowa la Table /, aaj U will be eelad tha. they pretest auong evidence agaiaat axbaatoa being 4 carcinogenic ftgoal. for. If xpoaure u ache*;* ie la aay way connected to lung cancer, we would expect that the longer and heavier the expoeure, the higher the rate that would bo found. Tbujaaly poesible error la thie Interpretation could occur if the weighted exposure* wore iaverccly related to year* of employment. la which caee the bcavicat weighted oxpocure (category III) would chow the abort* It aat length of employment. Table /, wbieb Uota the number of peraono in varloua exposure catcgorlca by length of employment iadleatce that thia error h* not occurred* la tact, the average number of yeara oi employment for each expoaure category la almoat identical. Expoaure Catacorv 1 11 m Unknown Tetala TABLE / /0 Member of Peraono and Number of Lung Caacer Deathe by Expoaure Category Paraoaa 20)1 2150 1772 $ 5953 Lund Cancer Destha Proved Total 4 4 S 4 2 4 0 0 ? 12 (Teblo 4/P continued oa next page.) 0277 28 tf TABU 8 (eeatlnuod) Annual Lung Canear Doatb Ratos par 100, 000 Ma&*Yoaro of Zxpoouro by Sxpoouro Category . *b # JCxpooaro Cfttet 3TV Proved Total ' *A# 1 S .1 m /- Unknown ..i y Over-all 33 33 23 31 19 3? 0 0 23 34 'Cn l TABU Y / / Number of Persona la Various Weighted Sxposure Categorise by Length of Employment Ut|tk f Etnelorment -- J- 10*19 * 20*29 20*39 40*99 $0* Total e Amrigf yurt of ospooure .. Weighted Svoesuro Catenaries 1 U lit 707 462 314 247 * 76 24 2031 391 003 363 210 67 21 2130 490 831 243 137 41 10 mi 17.8 17.8 16.6 Unknown Toul 0 1 2 1 1 0 ? 29.0 1795 2398 922 603 185 55 5`>5* 17.5 0279 Z 9. Table j4, which develop# the rates (or smokers and noftsrr.ker, i meet eulklag. It shows that not * staglo case 0/ luag eaacer developed amsag the U65 aoasmokers aad that all cases 0f luag caacsr, both "proved" ed '`euspected", occurred la smokers# Table W was so strlklsg that It was Jolt that further verification TW ry. It was possible that some sbaoraal dlstributioa may have t> ia* ,r occurred, e. g., the aoasmokaro may have iaeluded a larger parceauge of /** ysusg msa. Consequently, additional Tables, H. 1^, aad >S were constructed ts show the dlstributioa of smokers and aoasaokers by age, length of em* pioymeat, sad degree of exposure. Although there ere slight difference#, they do net account for the fact that all observed eases of luag eaaeer were 13 la smokers, la respect to age (Table the eoabtaed average ago of the sotskere wae d.9 years lass thas that of tho aoasmokers. Table K shows .1 that as far as leagth of employment is coacoraed, tho smokers had worked a.t sheet l*i years leas oa the average thaa the aoasmokers. With loager ex* posers sad greater ege, oae would expect tho: aoasmoktag group .to show a hither rate U luag caacer were due to asbestos. Table >4 shews that ths average exposure category was almost ths same for tho two groups. There* fere, this variable atoms to bo of no Importance la aeeouattag for this difference. Tho result of this additional a*lysis is that eons of these factors \ eppeurs to Ussoa tho effect of Table 14* 0273 30. T/8I.S Number ot Pereo>> end Number of Lv\j Cer.<:T Destbt by Smoking Habits JK.C Perrons Lung Cancer Ucaths Proved Tota, iW Smoker* 4673 9 12 r I HUnokaneotnwonkere 111605 00 00 Totals 575* 9 li C\ * e* f l Annual Lung Cancer Seech Kates per 100, 000 Man* Veers of %>osure by naoktag Habits u m* listokers Nenimnkers Unknown Ovet-ell *n Proved 32 0 0 25 Toul 43 0 0 34 .it a * fib 0280 3). i3 TABLC NwnUr of Smoker# and Nonamokera by Age Oroupi ... ' At* Croup Smoker Nonemoker Unknown 20*44 5200 689 12 45*54 96 224 4 .s>i 55*64 4U 202 2 t.U.1) 65 164 150 1 /" j*y Unisown 2 0 l Total# 4673 1265 " 20 Average At* 59.3 44.2 40.7 Percentage Distribution of Smoker a and Nonamoke ra by Age Cirou?* Age Croup i. ; /A 20*44 45*54 rJ 55*64 65 4 Unknown Total# Smoker 61,5 19.2 U 3.S 0 looji Lee# Uua 0$ % Nonemoker 54.5 17.7 16,0 11.8 .0 ToS^ Unknown 60.0 20.0 10.0 5.0 5.0 100% ( 0281 TABLE U. Number of Smoker* end Nonemoker* by Length nl Employment Length of Employment ; . - t <i >.u*. U 'Itoi 4V(H $-9 10*19 20*29 30*29 404f 30* Toub Averse Length e( employment Smoker 1408 1967 722 433 - 110 33 4673 17.0 Non*moker 377 426 199 166 75 22 "U6S 19.3 ' 32. Unknown 10 4 2 4 0 0 20 16.0 Percentage Distribution of Smoker* end Neaemokers by Length el Employment i i.<4 :l ***JjiH .14 * Length el Employment 5*9 10*14 20.21 30*39 40*49 $0 4 Tottli Smoker 30.1 42.1 15.4- 9.3 2.4 .7 Too*/. Nonemoker- 29.0 33.7 15.7 13.1 5.9 1.0 100% Unknown 50.0 20.0 10.0 20.0 .0 .0 100% 0282 31. I* TASLJ H*. Number of Smoker* end Nonemoker* by fixpoeure Category Cxpoeuro CAte^ory Smoker l U Ut Uakeowa TeuU Am<|i Aipotart Category 1553 1691 142$ 4 4673 2.0 Nonemoker 474 452 33$ 1 126$ 1.9 Unknown 4 7 9 0 20 2.3 Percentage DUtributloo of Smoker* ad Noneaokers by xpoeure Category txpeeore CaUjory S a m Vokaoeo Toule Smoker 33.2 36.2 30.5 0.1 100*/* Nonemokor 37.5 * 35.7 26.7 0.1 100% Unknown 20.0 35.0 46.0 .0 100*/* 0283 Comparison of tho Cohort Experience with th*t oi the Province of Quebec, Dominion of Ctnada, aad the United State* Za order to make a comparison oi the experience among asbeetoe minors with that oi the general population of the Province of Quebec, etattitlc* were gathered. as stated earlier, is the office of the Division of Demography la the Provincial Ministry of Health. The data oa toul death*, death! from all forma of caaeer, aad deatbe from caaeer of the luag were Obuiaed by aex aad by couaty for the yeara 19S0 through.1955. la addition, all death certificate* which apecifled primary caacar of the luag. aad all those which Indicated luag cancer but did aot specify the erlgta. were examined for the year* 1952 through 1955. it* Table 1^ glvee a tabulation of the number of death* from luag caaeer la tbt Proviaca aad la the cohort for the yearn 1950 through 1955. aad show* the annual rat# par 100.000 ia the** segments. It will be noted from the table that the mortality rata for the "proved*' cases.In tho cohort lo only slightly higher than the rate for tho Province. Whoa the "suspected" cases are Included la the calculation, tho rato for tho cohort rises to 53.8 per 100. 000 which Is shout S0 higher thaa the rate for the Proviace. This, it will be recalled from the previous discussion of th* effect of the "euapeeted" case# as the results, approaches hut dees net exceed the significant level. 0284 l ".di . . .41% o .1. . .... .1 c ftjfl -KM U\H* v\ti *.o b I 0.0 Z O 5 < a-s2"* A #4 A *4 <M 2 A A Wt o* R an e O O to* 21 N f* -- oAt ^S A ttoo 0* ?to so 2 nto fOt a f a * ttoo U S afi u ft,? 2 J3 & * w 5 Sto ft a ! Huxnbtoc lU t e la c o h o rt o t b e g in n in g o f 1951 35. $255