Document LKXdk2EM75z20vpQMewp6KOz7

269 A FURTHER STUDY 0? THU INCIDENCE OF CANCER OF THE LUNG- AND LARYNX. E. L. AND X. M. KEXNAWAY. (From the Pathological Department, St. Bartholomew's Hospital, London.) Received for publication. July 2-r>, 1047. MATERIAL EXAMINED. Tint data given below are derived from an examination of the death certifi cates of cases of cancer of the lung and of the larynx in males from England and Wales for the years 1921 to 1038 inclusive. The figures for the first 12 years of this period (1921 to 1932) were dealt with in an earlier paper (Kennaway and Kennaway, 1936) ; the material from the subsequent six years lias now been examined in the same way, aiul the data for the whole 18-year period 1921-38 cati no.v be presented together. The whole number of certificates examined individually was 38,418, while the annual total figures for cases in women have been utilized for comparison. This total was made up as follows : " Males. Cancer of lung Cancer of larynx . . Certificates Examined. 1021-32. 1933-33. . SSOS . 14,741 . 9472 . 5,397 . . Total. 23,549 14,869 3S,41S The investigation was carried out on the general lines followed in an earlier study of the occupational incidence of cancer of the bladder and prostate (Henry, Kennaway and Kennaway, 1931). The chief objects of the enquiry were: (1) to ascertain \i nether cancel of the lung showed any special incidence upon particular occupations, and (2) to seek for evidence to show whether the increase in death? attributed to cancer of the lung, now of the order of over 16-fold (from 361 death? of males in 1921 to 5982 in 1945) indicates any veal increase in the prevalence of this form Of cancer (Fig. 1). The figures for cancer of the larynx serve as a basis of comparison with those for cancer of the lung, in that the larynx Is a part of the same (respira tory) tract which is more accessible to inspection, and is not examined by di.-vnoetic methods width have changed greatly in recent years. The death certifi cates of person? under 2 years of age were omitted, as the investigation aconcerned primarily \\ ith the effect of occupation. Method of Calculation of Ratios. The comparative incidence of cancer of tho lung and larynx upon various occupations was e-timatod in the manner described in detail in tlie earlier paper (Kennaway and Kennaway, I'M}}. In judging of the occupational incidence of TXTINER RMC0082355 } ixcjdexcf of oxcer of hVSG .\XD LARYXX r ifmi of cancer it is of e.-our.-e necessary to correct for age distribution, for an > - I'aovinu' a, ' re r.rnooitiou of older -ivn till '.'iefI ca-ts oi 000o-- LuiiiS / /* / 4000- / r 3000- { \ i ? :oooji 1000 -- Larvnx --t - is:i : -i-------------------- i 1 . i i it-:* Paso ms _j_____ !_!_t io-n ioi3 Far. 1.--Dcatks "n,::L ic:. ~r of lung and larynx. 3Iak~. 1921-191-5. lI i er epnit from, any a-vcioal factor. Tlio census returns give the age ? -'.id.iitioii. in *- and In \vnvly periods, of the whole population of males, and I t 1 ' -A-S *. Vi*-'* i IVS. I. I K J i i .. ; tK_\. r*l \ t. i > ii I f i 1 s. { t 5 i 8 S f i ' rt i i5 :4 *:r l TXT1NER RMC0082356 - *rr? : 4 v.f ; 4 * :'3s.> ml K- ' -, .. -> - /4: ' ^l | f 262 E. L. .VXD S. M. KKXXAWAY il-y of `hose following e-ir-h nnc of the r-oo-'nized oeeupa! ions. at- fh fim<> of the The number of vU-ulis attributed. fa cancer of Due lung and larynx occurring in the whole male population in each of these age groups during'the years in question was obtained from material at the General Register Office. The comparison of the various occupations uith the general population is then a sum in pioportion, winch gives the ratio of the number of deaths hi any given occupation to those, reckoned as 100, which occurred in the same number of males in the same age groups in the whole population. An example of this calculation is given in the earlier paper. Sources of Error in Statistical IVo/L on Death Certificates. These were examined in the earlier paper (Kennaway and Kennaway, 10:03). Tlie use of death certificates for statistical purposes is liable to errors which arcinherent in the nature of the material and cannot be wholly eliminated. One must either recognize and admit these errors and correct them in any way possible, or abandon any attempt to obtain information from these data. (1) Correction for period between census.--In order to compare the incidence of cancer upon different occupations one must know the number of persons following each occupation. This is learned from the census returns. The longer the inter-census interval the greater is the risk of important alteration in these numbers. The figures for populations used for the earlier period {1921--32) were those of the 1021 and the 1931 census, which were combined by means of a formula described in the earlier paper (Kennaway and Kennaway, 1936); those in the present paper are from the 1931 census, which is the last taken. No census was made in 1941, on account of the war. The ratios for the whole IS-year period are calculated from the sums of the registered, and the calculated, deaths for the 12-year and 6-year periods as in the following examples: Table I.--Combination of 12-year and 6-year Periods. Cancer of the Lung. 1921-1932. ' Calculated death.*. ~ Total rtvi**t.-r.-U death?. Carpenters. 10? Printers 59-T 123 61 Ratio of 10*1 cal culated deaths. 73 102 1933-193*. Total Cal< iiJatfii r> dcsth.v t.V-'d d'-aiti*. 27tr6 8S-S 200 116 Rlltjr. of reud.to JOo ch!* death*. t L. 131 1921-19: VI -C 10- jit -Jh.'S. (.'almlMrd death*. tt-rM death.*. . 443-6 14S-5 323 177 1(0! de A'n*. 72o 119 (2) Sampling error.--In some of the most interesting occupations from the present standpoint (e.g. patent fuel workers, tar distillers) a few thousand men only are employed, and these will yield but very few cases of cancer of any ore organ even if the incidence is high (Table XIV). Hence the sampling error will be large even over the period studied here of 18 vears. The figures given in Table XVIII s! low the range of fluctuations which occurs. Kven Hie large group of agricultural worker's, numbering over 900,000 men, does not give .-o smooth a curve representing deaths from cancer of the lung as does the whole inale population (Fig. 2). TXTINER RMC0082357 I INCIDENCE OF CANCER OF LUND AND LARYNX 26:'. Factors which may affect the value of the individual death certificate are : (.") J>t`n'i!0'i of employment.--The individual entry in the death register li'O not give, and is not intended to give, any information about the length of \3 / * * t "AcC* ..VA7 .fv i '-'.i '-xyAt: '' . Af. : j : ` fr-'isSijS - -i M 1 *' . I I'i'T /!$ yt -f -` i f\ i.:T : ; -- ; b-te'-fh i i j r \.-***m *k -rk'~rW > r. * ; f ;'^C ; - Fig. 2.--Deaths from can-cr of lung. Males. A. General poj.-.:!aMou. lf>2l-l'.M4. B. Coal miners. \ y ''O 1 q-*i _i c*SS C. Agricultural workers, j time during which the deceased followed the occupation named, nor about the i nature ..r duration of any previous employment ; the uncertainty could be ii remove ) only by enquiry into eacli individual case, and that is impracticable j a> . H' 4t i i i i. .i C* -** ?5 .'* - ,-r T a*' L* --V trv~ ' iV * ' ' #* V* ' * --*** ^ **: Jtf . 'U a.' ^-J * - .t* * vJ+ * -*<?' =**vV ?' ' .- -f A ?Vr. TXTINER RMC0082358 204 E. L. AXD X. .V. KEXXX.WAY when large numbers are concerned. The immense amount of labour involve in enquiries can be .ecu In th-- 'uoi'i.nv nb no euii-rr <h the. -or'icion Hr Ih- :;- t lr'4n). 'Uur as nuue of the occupations included in tlie ` present study, with one exception, shows a very high incidence of cancer of the lung, comparable to that seen in the occupational cancers of the skin and btadder, this source of error is not a very serious one. (4) Incorrect diagnosis is of course a source of error in these, as in all other. medical data. In the case of death certificates received in bulk, perhaps some years after the dates of the deaths recorded, one can do nothing but exclude any certificate of which the wording suggests (1) that the cause of death did not come within the category under examination, or (2) that the diagnosis was uncertain or ill-founded (see next section ; such certificates are made the subject of individual inquiry by the General Register Office). When such exclusions have been made, one must treat the remaining certificates as correct if any use is to be made of the material. v j j ! '; ! Death Certificates Taken for Examination. ' The material utilized in this paper has been restricted as far as possible to cases of primary malignant growth of the lung, and of the larynx. The forms of words in the certificates taken for examination, and in those rejected, may be stated thus: Canctrjrf lung. (a) Retained.--Cancer, carcinoma, or sarcoma of lung, bronchus, pleura, root of lung, hilurh of lung, lung and mediastinum, or lung and pleura. Pul monary, or bronchial cancer, carcinoma or sarcoma. Cancerous pleurisy. (b) Rejected.--Any new growth of mediastinum, of mediastinal or of bronchial glands. New growths of chest or thorax. Peribronchial new growth. Any new growth qualified by " probable," " doubtful," or " query." Where another organ is mentioned, in addition to the lung, as the site of new growth, the practice has been to retain those cases where this organ is unlikely to be the primary site (e.g. cancer of lung and brain, or of lung and liver) and to reject the rest (e.g. sarcoma of arm and lung, sarcoma of.lung and testicle, cancer of lung and oesophagus). Cancer of larynx. (a) Retained.--The category of cancer of the larynx in the tabulation used in the General Register Office includes malignant growths of which the-situation is described by the following terms : Aryteuo-epiglotridean folds. Arytenoid glands and cartilage. Cricoid. Epiglottis. Glosso-cpiglottic fold. Glottis. Post-cricoid, retro-cricoid. Pyriform fossa. Sinus pyrifonnis. Thyroid cartilage. Vallecula epiglottica. Vocal cords. Trachea, windpipe. All these were retained with the exception of vallecula. t 1 'V v' IXCIBEXCE OF CAXCEP, OF LO'XO AXO r.ARVXX 2C>"> __._\nv ne-v gmwh -,;t| ififted X.- " ntrj':" " leu: tr !." or xuine c isps of which the nature is cloubttul (e.g. " tumour of cricoid cartilage ,v.d thv mid gland ''). A utopisies. The opinion is sometimes expressed that statistics of lung cancer are of no i die which comprise cases where no autopsy has been made, and some authori!'(see below, p. 2S9) maintain further that, even when autopsies are performed, m-niy cancers of the lung escape notice. Table II shows the numbers of autopsies iv cidvd in the death certificates for cancer of the lung* discussed here (i.e. after the- exclusions stated in the preceding section). Similar data for 1021-32 arc given in the earlier paper (Kennavay and Kennaway, 1930). These cases are clas,ili.`d as follows : NoRM.: P.M. : (I)NoP.M. (2) Inquest. No P.M. (3) P.M, (4) Inquest and P.M. (5) Coroner's P.M. without inquest. Taule II.--Autopsies. Cancer of L>.n>'j. (Males, 1933-38.) v-... IMS I'.cu r.'i' 1 If',* . .. . . . r. Xo PAI. II. Xo PAI. 1,24* 1,784 1.040 l.yss 2.23-> iVW* . . . . . . 6 3 -- 4 5 III. rsum wf I 2if;d II. 1.233 1,4*7 itr.4 1,002 ih-MO 2,529 . . . . . . IV. PAI. 432 404 374 53 3 522 o'Ji V. aiA PAL. 34 .) 30 37 61 42 . VI. Cor*jii?r,? PAT. No int|Ur?t. ' 2S 3(1 . 34 34 . 40 53 VII. Nim >f IV. 4 and VI. 501 055 044 610 023 033 . . . . . Total. 1,754 2.042 2,200 2,572 2,603 3,222 Total I'.'J 1 -32 . . 11,004 . . 6,5uU 23 . 11.117 . 31 '**> 34 Ii.0i.i3 . 21 ' -> 232 132 . . 228 68 3020 . 14,743 2205 . 8.S08 1321-38 . . 17,003 . 5 17.72U . 3171 304 2;m> , 5s3I 1*3.551 Table III.--Proportion of Autopsies. (Males, 1921-38.) Mean f Car.cc-r of lung, 1921-32 Total deutlis. S.S03 Au-.'-.psics. 221,1.3 Ter annum. 1S4 Per Cent death?. 25'0 1933-38 - of larvux, 1921-32 % 1933-38 14,713 9,472 5,398 392(1 CO 4 995 50 44S . 75 24-6 03 8-3 Tabic III shows \1) that the number of autopsies has increased much more 1(* the h-.ng 'Cik-s than in the lai\nx scries ; and (2) that the average proportion of tows of c..;:i.cr of the lung examined post mortem is the same in the earlier * A rr/rr-r-p'or.-ii;.,' ih.W f-*r iv.t.i.-er of th.- larynx is omitted to sp'.v'c. ..;r- si,rsirffTTSi."J . t!' .V---V F-CXc-s-j ... .V Vsf-.. . '5-i4 s" :v "! j ,i i * `-P& t T. S'jjf ! ,< p- *M !A iW I 1W i ?$$? i i,' r. i 1' | ; . I1: 1; !?-. r !.- .:,i : i , : i U- Wm. i j Sr -"3 fa. ; i; i1 r ". hx.K5j* * -'**>-** . i. ' .* V > v .i-tcjb,i t-. TX TIMER RMC0082380 : lSf.lDF.SCE OF CANCER OF LOTO AND LARYNX ; .Vr;- ..tov''t:.*V il^.' - -- - , -,.r 207 TXTINER i RMC0082362 T a iii.k V (tonhi.)  *** i .. .. ^ 270 E. I,. AXD X. M. KKXXAtVAY ............... cnnuinc; ira ;'r'is<* i.i H'.m!.!'". i ; ` ib'1 i r<> iv-uver of the lung is not due to any increase in this proportion. t i i COMPARATIVE INCIDENCE UPON VARIOUS OCCUPATIONS. In the sections below, the periods 1921-32, 1933-38 and 1921-38 arc referred to as A. B and C respectively. The 50 occupations named in Table IV employed during the period B 3,880,001 out of the 13,305,608 males aged 20 and upwards (census of 1931), and they include most of the men engaged in the two largest industries in the country, namely agriculture and coal mining. The tables include 3550 out of the 14.741 cases of cancer of the lung, and 1451 out of the 5397 cares of cancer of the larynx, occurring in period B, in the whole population in question. Tables IV and V show the data referring to the occupations which have been studied in detail. In Table IV the occupations are arranged in the order of magnitude of the ratio for cancer of the lung, and in Table V in the corresponding order for cancer of the larynx.* The population engaged in each occupation, and the numbers of calculated, and of actual, deaths which are required for the estimation of the ratios are included. The number of deaths in each year of period B in each of these occupations is given in Table XVIII. (1) Agriculture. A group of open-air occupations (gardeners, farmers, agricultural labourers, including shepherds, farm bailiffs and foremen), which together include about 954,000 men. has 505 cancers of the lung and 342 cancers of the larynx in B. This group shows very low ratios for lung (from 26 to 58 iu A and from 22 to 61 in B) and low ratios for larynx (from 41 to 75 in A and 29 to 79 in B). In this open-air class the highest figure for both lung and larynx is given by gardeners. of whom many live in or near large towns. j J (2) Con! mining. The chief coal-mining occupations (workers above ground, hewers, road- makers, persons conveying material to the shaft, other worker's below groundh which employed nearly 770,000 men, show low ratios (lung, 38 to 71 in A and 37 to 72 in B ; larynx, 44 to 58 in A and 42 to 73 in B), which are very like those of the open-air group considered in the preceding paragraph. The air in a coal mine is pumped in from outside, and collieries are not always situated in urban districts (e.g. the Kent coalfield). The close agreement in the ratios given in A and B by the agricultural and coal mining groups is noteworthy. There two industries (agriculture and coal mining) have also a low incidence of cancer of the bladder (Homy, Kennaway and Kennawav, 1931). The rate of increase of deaths attributed to cancer of the lung in these two groups is shown in Fig. 2. The difTtfiujiC..* between th* total numbers of occupation* in Table TV (5tl) owl Table V (*-) is du* t * f *ur fpatent fu l workers^ potters* mill worker*, bfost furnace engine me:;. anti'lb' -ter*) in which no death* from cancer of the larynx occurred in either period. j j j } ' ; (` j ' \ J } IX TWER RMC0082385 INCIDENCE OF CANCER OF LUNG AND LARYNX 27: (). !(''/ exposure to road dn.it. A of open-air occupations, where there is exposure to the dust of roods. 1 " .Mr \U''_ hu'ynx. 'an'.i "V :.ri >. d- :v;-.' have nearly normal figures fur larynx (Table Vlj. Xoim of these occupa- ' ions shows a high incidence of cancer of the lung, and the paviours, street masons, c iiicrc-io!' and asphalters show a distinct fall in the later period. Ml Gnm,-'t, horsekeepers, corpentera. Glooms ami horsekeepeis, who are exposed to various forms of dust arising fo-.m horses and fodder, have an incidence of cancer of the lung (A 106, B 71) r.!"J hnvnx (A 134, B 100) which is not very far from normal. Carpenters il ahle I and Fig. 7) also show low ratios (lung A 73, B 72 ; larynx A 82, B 88). These instances suggest that exposure to dust per se, irrespective of its chemical nature, docs not conduce to cancer of the lung. Table VI.--Copier of Lury and Larynx. Occupations with Exposure to Rood Dust. Ratio. Lung. Larynx. 1021-32. Paviours, street masons, eoiicretors and asphalters .... 20!* Council labourers, road sweepers, dust men 173 Drivers of horse-drawn vehicles . 144 Motor drivers (goods and passengers) . 142 1033-33. 137 167 131 153 1021-32. 154 10.73-33. no 172 267 184 143 m100 (5) Silicosis. . te greatest liability to silicosis are named in Table VII. Of these occupations, those mentioned in Table VIII, and the various subdivisions of coal-mining, were included in the present study. Table VIII shows that in the period 1033-38, the highest incidence of cancer of the lung in in potters' mill workers, slip makers and arkmen (ratio 218), but in tljis occupation the numbers are small (see Table IV), and hence the sampling error is considerable. In the other occupations in Table VIIT the tendency is towards a dccrea-e in the ratio. The stonemasons, and stone miner* and quai-riers of course include many workers with non-siUceous rocks. The figures for deaths from dlicuds in occupations of this class are falling ; in the metal grinders this is due to the adoption of non-siliceous abrasive wheels, which lias been advocated by the Senior Medical Inspector of Factories (Factories Report, 103S). In the various groups of coal miners (Table IV) the low ratios show no uniform tendency to rhe or fall.* The general indication of these results is that the factors which !*ad to silicosis are not very active in producing cancer of the lung or luiynx. There fie. however, in the literature various accounts of cases in which * C. .1 O r fy,! i f-!, I'i4". - '"I. s'i] in a -:!u !y of pneumoconiosi- in Routji Wales antlirarito :f rC'.- r f i!,.> Ion- in 2_'7 no'-rop-ies on cerlifit-U silGntic.*, and one cast, in a furt'.i-. " ,, :i, <, nearly 4"0 necropsies iu alt. ':'' l ' '.V .,y -r *.*? T ; : -_CTT x*'- z *> ?.l:S I : j1; , l| ': !; .' : ij: ,1 1 1 "'f.! , .i tcr-- _ " ; >; - 1 * , ,, - 1) \: . * . * 5" 1k : i if ' -~A txtiner RMC0082366 2(2 E. ].. AND -\. M. K KXXAV.'A V cancer of the lung wa> .-u-sor-iated with si'k-o.sis. Such individual observation' my ddlo-dt t<> h"'S ally as the- v.t> f-c*:i<!iti<jn- n;us'. .'.>r.ir-im*-' -Mr ` : 1r`L`. ,1: at ti'r* (.!'.*' a :!(-! -uioi o< "lACu1*: At rq'.;.,* . A * r>'. itiitt a favours the detection of a high proportion of carcinomas. Iaulf. VIT.--Deaths front Silicosis in England and W ales, 1939 to 1945 indusice. [From the `Annual Depart of the Chief Inspector of /'ca lories for the Year 1045,' p. 79.) 1. Refractories industries 2. Pottery, manufacture of . 3. Sandstone quarrying and dressing 4. Sandstone masons .... 5. Metal grinding .... 6. Sandblasting . . . . . 7. Steel dre-siug and cleaning of castings 8. Stone, pebble, Hint and sand crushing 9. Scouring powders, manufacture of lU. Abrasive wheel manufacture 11. Glass cutting and hovelling 12. Millstone dressing . 13. Slate quarrying and dressing 14. Granite quarrying and dressing 15. Tunnel mining (sewage works, etc.) . 16. Coal mining ..... 6S 310 173 303 9S 53 45 15 5 4 3 1 45 13 21 1731 Table Mil .--Cower of Lvng in Occupations with Liability lo Silicosis. Patios of registered to calculated deot ha. Potters' mill workers ; slip makers and arkmen .... Potters ; ware-makers, casters and finishers ..... Metal grinders .... Stonemasons, cutters and dressers . Stone miners and qua triers Sand-blasters .... Coal-miners (see Table IV) 1921-3-' 96 152 223 149 105 201 58-71 1933-38. 218 132 144 140 49 103 37-72 1021-3.8. 174 139-5 176 143 70 136 43-69 , Number of deaths 1921-3S. 5 22 39 SC 39 2 774 Some of the literature on the relation of silicosis to cancer of the lung i> summarized below. Bible (1934), in the course of 14 autopsies upon persons of both sexes Buffering from silicosis, found malignant disease in three cases in stonemason.', as fellows : Case 1. Carcinoma of bile-duct. Spheroiclal-cell carcinoma of.prostate. ,, 2. Oat-cell carcinoma of lung. Columnar-cell carcinoma of colon. ,, 3. Anaplastic carcinoma of lung. i' * TXTINER RMC0082367 INCIDENCE OK CANCER OF EUNO AND LARYNX 2 !> A A-'Y' r^i.'UArf'.l alllCR :n + . limits (A; il-J'-l V .' gs showing primary cancer I B). Twelve (C) of T> showed abnormal amounts _ a. the mean content of the-e being nearly 10 times that of A, while the v- content of the remaining oS (P) of B was the same as that of A. A : 'l opicaily. C showed typical silicotic- nodules in 7, areas of confluent : .!-:! in and fibrosis not more than suggestive of silicosis in 2 of the 12 cases. T '- . -i-: j;1 ,\ and D " did not in any sense reproduce the picture found in those m a - in which silica was present in excess." The authors conclude that "a of cases of pulmonary carcinoma exists in which the organs contain an < X'- - of silica and show histological evidence of silicotic fibrosis. The conclusion, r. c t'.iiJc, is that the role of the silicosis is aetiolcgicul." Table IX.--Silicon Cases of silicosis Un-eiected cases oo/l Cancer of Lung (Klotz, 19:10). Number of autopsies. 50 Carcinoma of lung. Number. 4 Per cent of autopsies. SO 4500 53 1*17 K'otz (1939) gives some data from Toronto (Table IX) of which the '-"t'-rical character, and indication', are similar to those of the material of And-.'son and Dible. In the report for 103s of the Chief Inspector of Factories (p. SI) the Senior Medkal Iii'peetor (Dr. J. C. Bridge) gave the following data derived from 1290 -j Acs carried out to ascertain whether silicons was present (Table X). Her-.c presumably the subjects had shown in life some respiratory symptoms. Table X.--Silicosis a,/? Cancer of Lung (/. C. Bridge). Number of autopsies. 943 347 iSilu'C"*:.*. Present Absent Car.cor of lung. Number of case??. 23 17 Per cent of autopsies. 2-4 4-9 1299 Table XI.--Cancer of the Lang in Wiluatersraml Miners (Miners' Phthisis Medical Bureau). t nr. rr. i! * . h-w ri-.t 'I.- .! H Noa-O-.-*--* E.r'-p-an n::ar -1 H.trru-J. ^Hit-otic Enr^-an miner* v< *. r* .: U'.> C**i f f I'rj:r, . r 13 P-rnt.vg* of .'ll! . p*'i*T lu alrinb o l* --: 1179 M3S *-f trii:^n* of li PiT' -- Jit of all iriur*.*ni-. 0-71 Pu*f lo-.n-m.'. . 143 S of priin\ry oCfiUluicni*gr. 10 IVieentage of all poet- U)ort-nH. 0-7 3117 The data O' a -davd dm irig In years from 3117 post-mortems on Europea L-ifico on the Vfitv.atertniiirl (Mi.s' Phthisis Bureau. 1930) indicate that- the 274 F;. 1.. AXD X. M. KKXXAU'AY incidence- of c-mwer of the inn" is the same in silicotic and uon-silieotic miners (Table XI). T1 mv-nnnr riven he.*- of tS.; l:t.r,iture t.n :he r"hti'>M <.,< ....... i: ' - .'m2 '-It-' *:or- di-.nn t-> h-e complete (see the review by Klorz, 193y). (6) AtbcxttS workers. In the years 1930 to 1944 in England and Wales the fatal cases of silicosis and aste-to-is numbered 2699 and 7.3 respectively. The average duration of employment in fatal cases was, in silicosis, 34 years, and in asbestosis 15 years (Factories Report. 1944). In his report for 1938, the Senior .Medical Inspector of Factories states that " among 103 fatal cases in which asbestosis or asbestosis with tuberculosis was present, cancer of the lung was associated in 12 cases (11-C per cent)." Table XII records the cases of cancer of the lung in asbestos workers which hare been found in the materials dealt with in this paper. f i i Table XII.--C'aacer of Lung. Asbestos Workers. Males. Death Certificates 1032-1938. Age. 1935 Asbestos weaver . . . . . t . 62 1933 ,, cloth weaver . . . ... 54 103S ,, weaver ...... 46 1935 Boiler and pipe asbestos c-overer ... 58 1938 Asbestos packer ....... 50 1936 Disintegrating room hand in asbestos works . 65 1934 Departmental Manager in asbestos company . 45 1936 Asbestos works manager . . . . 5S These eight cases of cancer of the lung occurred in the years 1934 to 1938 in men of ages from 43 to 63. If one takes the number of deaths from cancer of the lung in these years among the general population of males of approximately the same age groups (45-04) one can make the following comparison : * 4,088,382 males (ages 45-64) produced 8191 deaths from cancer of the lung. = approximately 4 million males (ages 45-64) produced SOuO deaths from cancer of the lung. J y = 4090 males produced S deaths from cancer of the lung. j The number of male asbestos workers in this age-group appears to be con siderably less than 4.000 but as we have been unable to obtain any exact figure for tire workers at risk one cannot carry the calculation further. j Lynch and Smith (1935) report from South Carolina a bronchiogenic carcinoma in a man who had been for 21 years an asbestos wear er. Egbert and Geiger (193G) record a bronchiogenic carcinoma, with the characteristic pulmonary fibrosis and asbestosis bodies, in a man who had been >V 4 `ii for 18 years an asbestos weaver at New Haven, Connecticut. Gloyne (1935) records two cases of fairly advanced asbestosis, with squamous carcinoma of the lung, in women workers in England. " One survived nine rears after an eight years' exposure to asbestos dust as a spinner; the -A n . >) : <* TXTINER RMC0082369 * - ners ncer tosis n of oars ctor 0-is 1-es *t0S 8 ill r of ( l-!v t lie t!:s '! 1 1* X nv I INCIDENCE OF CANCER OF LUNG AND LARYNX 275 oilier lived for fifteen years after two short periods of six months and thirteen months'exposure in the mattress and opening departments of the factory. . The malignant lesions of the lung were, in each ease, very small and not recog nised during life." Subsequently Gloyne met with asbestosis and ail oat-cell carcinoma of.the lung in a man who had been for 20 years foreman in the stores department of an asbestos works. (7) Alcohol and tobacco. (There is a eou.-ideraMe difference between occupations associated with the - supply of alcohol, and of tobacco ; the former show a very high incidence upon . the larynx, and a normal incidence upon the lung (Table XIII). The ratio for i the lung is the same for licensed victuallers as for physicians and surgeons j (Table XV). The occupations of barman and cellarman show by far the highest I figures of any in Table V for cancer of the larynx. The incidence of cancer of J the lung upon the two occupations connected with tobacco has decreased distinctly j in the later period. The increasing use of cigarettes (Fig. 10, 11) has caused the | tobacconist to become more and more a vendor of closed packets which can j hardly have any occupational effect. The decline of pipe-smoking makes the mixing by the tobacconist by hand of a blend to suit a customer's taste to be { less frequent. T.U5LB XIII.--Cancer of Lung and Larynx. Occupation-} Atiociated with the Supply of Alcohol and Tobacco. Ratio. ~~ L mg. Larynx. Barmen .... Cc-llarxnen .... Licensed victuallers Tobacco manufacturers . Tobacconists and tlieir assistants 1921-32. 104 92 128 196 . 175 1933-38. 122 87 124 110 82-5 1991-32. . 469 317 . 1S5 05 . 142 1933-38. 275 502 199 129 171 (8) Coal-gas and tar. The occupations associated with coal-gas and tar are important on account of the known liability of the workmen to cancer of the skin, and of the bladder (Henry, Kennaway and Kennawav, 1931), but they offer unsatisfactory statis tical material because some of the most important occupations employ very few men, and hence, produce hut few cases of cancer, so that the sampling error is high even over an 18-year period. This difficulty can only be met by omitting some of the smaller cla.-xes. Thus, in Tab'c XIV there are ten occupations in which there is especial exposure to coal-gas and tar. If from *hose ton one excludes the four which have produced less than im deaths from cancer of the hinu. one obtain*, by this quite arbitrary procedure, the results -i'oivn in Table XIV. Tail* Vi 1 X ,' 'pul I I ' U.cV h Vi*-: (iV* 1 C'-::sus) and t tcld 2'iT cu-es of cancer of t he lung, have an increased (up to 284) * ratio for tills iUm so. the lowest figure being that for chimney .-.weeps (119). F ) i s ' 1i if a ft ti if h( TX TIMER RMC0082370 ''* < - *-A,. V- -* '~ ' ^ I - r. L. AND N. XI. KKNXAWAY Taulf. XIV.- Cnnrtr of Ln, >'J. (Jo.Hp'lti >/!! -'l-v. -V !/ ; ' - . Gas stok-ris and coke-oven chargers . CMs producermen .... Gas-works foremen and inspectors Gas fitters ..... Gas-works' labourers Chimney sweeps .... id -.;st fUatljs. 8.7 12 2.5 41 90 21 U:!l-32. 342 102 197 182 102 170 frith K.iiiu 1J33-3S. 251 232 161 li>& 109 90 l'.rjl-ys. 2S4 202 174 107 12!) 119 Excluded : Labour ets, patent fuel works . Gas-works engine and crane drivers . Gas-works managers Tar distillery workers and coke-oven workers ..... 207 3 0 4 8 433 60 244 29 1529 ISO 5S 70 571 138 130 52 21 Jn five of the six cases, and especially in sweeps, the ratio has fallen in the later period. The ratios for cancer of the larynx are irregular throughout this class. (0) Profusion'll occupation*. A group of professional workers (stockbrokers and stockjobbers ; clergymen : Roman C.-tholic priests and monks ; ministers of other religious bodies ; judges. stipendiary magistrates and banisters ; solicitors ; physicians, surgeons and registered medical practitioners ; dental practitioners) numbering over lOs.OOO men and snowing 1 TO deaths from cancer of the lung and 39 from cancer of the larynx, give altogether ratios rather below the average (lung A 94, B 97 ; larynx A $4, B 39; Tables IV, V and XV). These ratios for cancer of the lung, being near 100, are in accordance with Stevenson's (1923) discovery of the ahsence of social gradient in this form of cancer. When tho individual occupa- tior.s ate examined considerable variations arc found. In the earlier paper attention was drawn to the high vatio for cancer of the lung in judges, stipendiary magistrates and barristers (173) and in stockbrokers and stockjobbers (1ST) ; in the later period (1933-33) the former fell to 112 and the latter rose to 2">::. The stockbrokers and jobbers also have the highest ratio among the professional workers for cancer of the larynx. The figures for Homan Catholic priests and monks arc- too small to bo significant. The ratio for ministers of other religion-: bodies }v- risen (Table XX). In regard to the view that the rapid increase in recorded deaths from cancer of the lung is due to the detection of more cases by improved diagnosis, the very moderate and consistent ratio (A 129, B 123) for cancer of tire lung in medical men is noteworthy, for this is an occupation whete the availability of the existing .methods for the detection of cancer is pivv.u.seb'.s- at a maximum. Dental practitioners give verv similar figmes (A 103, i`/l24). * < * \ t ) } < 1 ( 1 1 i ! j I 1 j | j j ? ; ? ' i * 1 TXTtNER RMC0082371 INCIDENCE OF CAN'CER OF LONG .AND LARYNX The ratio-: for cancer of the larynx in these occupations are below, and in sonic cases considerably below, 100, except in stockbrokers and stockjobbers (laS). - Table XV.--Cancer of Lung. Professional Occupations. ! 321-32. >73 38. 1321-S*. If-: Occupation. Popula tion. Stockbrokers and ctoekjobhers Roman Catholic Priests and M.nks Ckiyvmrn (Anglican rinirch) .Minister* of at her ReBodies J udizos, St ipendutry Magistrates and Barristers S'/wilors si* Surgeons oik) * Registered Medical Tracti- t loners P--lital Pra*-titloners 0.S93 3,604 23.422 11,709 3,658 17.30G 27,051 10,494 Total ltv.it lt>. Ratio of to I 'hi <ul- onlaU-d tl- nth?. P-Mit'a- ti'-n. 14 1ST . 7,123 1 31 . 4,050 15 53 . 22,754 5 33 . 11,960 i 173 . 3,632 19 100 . 17,353 35 139 . ;n.'.os S 103 . 11,633 Kat-o of Total. rv-.l. ro rrsti P-'i ir*. Bvi-i of r-iil. t* D-i ' lit lt^l Ctnr-r of l<ii> UK. Jvartj.. Ratio. 32 253 . 229 . 8 153 - ---- 11 5 24 4_) 49 30 i 45-5 15 HO . TM 6 64 7 112 . 136 26 S4 90 1 34-5 9 <'5-5 56 123 . 127 13 124-3 . 117 38 1 44 (10) Colton workers. 39 Cotton strippers, grinders and card-room jobbers, among w hom severe bronchitic and asthmatic affections occur, show a very low ratio (CO) for cancer of the lung in A, which rose to the normal level (104) in B (see p. 286). Cotton spinners and piecers have a low but increasing ratio for lung (A 02, B 75) and medium ratio for larynx (A 148, B 91). The total population of 43,991 (1931 census) includes ring spinners as'well as mule spinners, but the latter occupation certainly makes up a very large proportion of the total, for the number of mule spinners aged 20 and over was estimated in 1926 as 41,000 by the Departmental Committee appointed to consider evidence as to the Occurrence of Epitheiiomatous Ulceration among Mule Spinners. Tn view of the exposure to oils which may be carcinogenic the low incidence of cancer of the lung is r-.markable, but the very high incidence of cancer of the skin, and the increased liability to cancer of the bladder (Henry, Ktmnawav and Kennawav, 1931) in tins occupation, must be taken into account. The question of the penetration into the lung of oil droplets in the air of the mule room will be dealt with in another publication. BATE nr INCRVASE OF CANCER OF THE LCXG AND LARYNX >: .MEN AND WOMEN. Cancer of lung. Cancer of the lung has increased from 361 to 5982 ease-; -per annum in men (Fitr. 1) between I'l:'! and 1915. and from JSrt to 14 no in wo me". T!i-cs ilumv's in i i:. /." 1 : I1--1) .t i -:) a1,.': i : iwo.-.ienj j'l became more iapid in i9.9j and again in 1930 in men (Fig. 1) and about '929 in ''omen. The annual increase, in men, between 1932 and Iu37 was at the rate of from 200 to 3n0 cases per annum ; since then it lias been very irregular B- ' >" 4. jet* __ - J. ,s 2?Ki> _-0O^P. CCS 7if \ -b i TXTINER RMC0082372 -> : v. - . * A'; \V*% tff; m. $ : -ri 1 "< >?. VV r-- '* - ntm-AUlilrt tm*|T< I ifI fcjV~Ti*tfr Mil 27a E. L. AXX) X. \f. KEXXAWAY (Table XVI). The ratio between death.? from cancer of the lung in men, and in women, is now (i.e. in 1913. the lalest year for which figures are available) 1 vhieh is'apimwcHri: ihf >tpst .vai! !'- (t9391 ratio Coe oa-u-er of the j.anini (Tabic XVII). Cancer oi the rung diners iroin the ot,\rr iiirms of cancer in Table XVII in that the propoition of males is increasing. During the whole oooo Status / / 5000 / / 4000 / / / 3000 / / 2000 Females 1000 1935 1910 19-13 Fig. 3.--Deaths from cancer of lung. Male* <iul Females. 1930-1945. period under consideration here (1921-1943) the ratio of female to male deaths from cancer of the lung has change'! from 1 : 1-9 to 1 : 4-0. A similar change has occurred in the United States ; the ratio of the standardized death rates for cancer of the lung and pleura in white men and women was 3 : 1 in 19.4S, and 20 years earlier was less than 1-5:1 (Statistical Bulletin, 1939). Fig. 3 shows that the rate of increase in recent years of deaths attributed to cancer of the lung is much greater in men than in women, for the curve of deaths W1 IXCIDE.VCE OF CAXCEIi OF t.VXG AND I.AKYXX 279 of males rises much more steeply than does that of deaths of females. But this apparent great difference between the sexes is due in largo part to the fact that, in a graph of this type, a given increase in a larger quantity is much more conspicuous than is the same percentage increase in a smaller quantity. If one reckons the annual numbers of deaths as a percentage of those in any given initial year (Table XVI), the difference between the two sexes is not nearly so conspicuous (Fig. 4). 'a* ^ '** r'r j:' f f f- i .* " Fm. 4.--Cancer ofl-.ing. 'ii1 1935 'ii!iii 1940 1945 Males and Females. 1935-1045. Deaths per annum as percentage of death* in 1935. Table XVI.--Totals of Death Certificates (England and Wales) and Percentage Increase in Cancer of the Lung. Cnncer of lung. Cancer of laiytix. 1935 1936 1937 193S 1939 1940 1941 1942 J 11 j t _ 194.7 Men. 2337 = 100 2000 = 112 2914 = 120 3273 = 140 3391 = 145 380,8 = 104 4u09 = 174 4579 = 100 .-si ' L ! * ' 59S2 --= 250 Women. S58 = 100 91" = 107 927 = 108 1049 = 122 1149 = 134 J ISO = 137 11.50 = 135 1207 = 148 1290 = 1.51 13'S' =- let? 14M) ---= 172 Men. 903 916 918 887 929 003 929 $37 <r;> 852 841 Women 224 275 232 259 280 207 209 25.5 9s I > 279 f ' V r'/n TXTWER RMC0082374 2x0 K. L. ANT) X. M. KF.XX.UVAY TvliLE NATL -J.'ri'c-r of I Or 't't n i i.i/','iri l Site `if rancer. 192! Lung .... 1932 1' .... >> Oesophagus Mouth region (lip. tongue, mouth, tonsil, jaw, pharynx) 1937 Lung .... 1939 >> 1945 Oesophagus Mouth region Lung .... >>/ if ii Number of < Uath*. Women. 1S6 6C1 Men. 301 1553 1774 536 927 1029 729 585 1480 3040 2914 3544 1590 2538 5982 . . . f'ettio. 1 l*`> l : 2-7 l : 2-7 1 : 5*7 1 . 3-1 1 : 3-4 1:2-2 1 : 4-3 l : 40 Cancer of larynx. Cancer of the larynx has increased between 1921 and 1945 from 641 to S41 cases per annum in men, with a maximum of 9T2 in 1943, and from 138 to 27!) cases in women (Table XVI). These figures are in the ratio of 1 : 1- 3 in men and 1:2-0 in women. But the figures for men have been fluctuating in the region between Sod and 97 o since 1930, and in women there has been no uniform increase since 1928. Hence if the adoption of smoking by women has had any effect Upon their liability to the disease, this factor has reached equilibrium. COMPARISON OF THE RATIOS IX THE EARLIER AXD LATER PERIODS. The calculation of these ratios (Tables IV and V) for the two periods 1921-32 (A) and 1933-38 (B) provides material for a comparison. If the data from which the ratios are derived are reliable, this comparison should show whether tiie incidence of cancer of the lung or larynx upon any given occupation has under gone changes differing significantly from any which have occurred in the general imputation, and the general degree of agreement between tire ratios for the tv.o periods is some indication of the value of conclusions based on death certificate--. In Fig. 5 the comparison is made for cancer of the lung by calculating the uttio between the earlier and later ratios (B,A X 100) and setting down the figures in order of magnitude. Occupations with not more than one death in either period are omitted here, and are considered elsewhere (p. 2St>). The results shown in the grph may be summarized thus : Ratios below G6 . .. ,, 66 to 133 . .. ,, above 134 ... 11 35 3 49 Thus 35,'49 of the ratios of one period, or 71 -4 per cent, are within 33 per cent of those of the other, which is perhaps as close an agreement as one could look TXTINER RMC0082375 * ! INCIDENCE nr CAXCE11 OF U'NH AND LARYNX ~Nl Fic. 5. -- Caaocr of Male-: In 49 occupation*. .A. Ratio of rv^i-t^T*\\ t> ]> uakuUYed deaths, 1921-1932. B. R*tii> of nvktot-.-d to caY.dated deaths, 1933-193$. Dots r^prc^ent li A .. 100. Occupations with ivt more than oar death in either period are omitted. toi in siu-Ii matotial. AVhon the individual points on the graph are examined, the larger occupations arc found to lie chiefly in the middle part of the series, as one would expect from the lower sampling error. } A similar treatment of the figures for cancer of the larynx (Fig. 0) gives the fi.'.lowing result : i Ratios below 66 .7 l ,, 66 to 133 2S ,, above 144 14 49 In this ease 2S(49 or 37 per cent of the earlier ratios are within ~ 33 per cent uf the later one-. There are thus more occupations in which the ratio lias iii'-u-ased than in the ease of cancer of the lung. OCCCJ'ATIOXAL INCinE-NCE OF THE INCHLcSK OF C.VXCEU OF T1IE LUNG. If there is a real increase in cancer of the lung, and if this is due to some < mci mil factor, one might lind diSereuces in the rate of increase in classes of -on.s exj O'ed to different environments. The annual numbers of deaths ir in cancer of the lung, and of the larynx, in each one of the occupations studied in detail in this paper can be seen in Table XVIII. The increase in the total of n - tiled cases has ! n <o great (Fig. 1) that, if this increase were due to a con'Hniti'.-n fro:", a tty of the -mailer occupations, the incidence upon these would . i to be c-ry high indeed. e. } -. i; ii. i !:; I ' i |: ifL i i t hi i. L txtiner RMC0082376 . *r`: . f ' -y * *T : ri- F. L. AXI> X. M. IiLXXAWAV L \i;l.K XVlir.--S Huber oj D'Vitit< jront C'titcer oj lilt Ll'tt'J and Lar;/n.c i ~jt) 0~rjintion-s iii each Year 1933-1 !W$. The occu]i<ili^ni are pl.u'i-cl iu the sauie order as in Table IV. Labourer*, patent fuel work* . adok'-r* and coke-oven. charger* . Gas producer men .... Mctnl grinders .... Ges-works foremen ar.d u-.-peetors . Potter-' min workers ; makers and Arkmen. Mainly council labourer-. r. vl sxreepets and dustmen Ga* fitters ..... ravu.urs, street ma^-on-. con''retorand a^pholtcrs Motor din ers, goods ar.d pa-s\oger.s Blast furnace engine m--n Cabinet makers .... Tolxueo manufacturers . Stone ma=ous, cutter* and ilr->?ers . Tanner*, tenther dressers, curriers, skilled workers Potters; ware-makers, carters and fini.sla-rs Sandblasters ..... Gas-works managers Drivers of liorse-draw-i vehicles ~. Gas-works labourers French polishers .... Painters ..... Licensed victuallers Printers ..... Chimney sweeps .... PhiTiit-ers ..... Barmen .... Lurz I-ai> nx Lung Larynx Lung Larynx Lunsr Larynx Larynx bung baiy n. Lung Larynx Lung Larynx Lur.g Larynx Lung Larynx Lung Larynx Lung Larynx Lung Larynx Lung Larynx Lung Larynx Lung I-arvnx Larynx Lut g Larynx Lung Laiynx Lung Larynx Lung Larynx Lung Laiynx Lung Laiynx Lilt. z Larynx Lung Larynx Lui'z Laiynx Lung Laiynx Lung Larynx 10:;4. I M3. W30. 1J7. 193*. IVvUl. -- 1. 1 4 8 11 ii ! 4S 3 4 2 3 . 17 4 --3 41 3 12 3 r, 11 --111 1 in 9 i:> 18 10 16 15 16 23 6 7 9 16 12 13 44 8 36 1 12 2 430 1 2 112 30 42 43 72 S3 8 8 8 11 1 in:; 63 23 li 14 4 3" l t> l 10 9 14 14 13 5256 2 32 1 2 31 11 13 13 16 17 13 1 14321 433 5 4 12 2 21 15 11 1i __ 1 __ ,, 1 ---- 1 ____ .__ __ __ -- -- __ __ __ __ __ __ __ ___ 1 __ 1 __ __ ,__ ___ 23 40 3.3 51 45 44 is 17 23 17 is 7 4 8 8 9 9 12 I 1 3 3 1 __ 3244 24 2 2 4---- 1 29 47 43 51 70 77 27 24 20 11 IS 20 18 33 21 29 33 33 17 21 14 22 14 14 14 13 23 14 20 26 0466 74 ---- 3 1 24 1-- 1 1 2 l 5 10 34 8 12 14 17 ) 5 4 4 2 1 64 25 00 2" 12 3" 3 13 4 3 323 12" 172 1"2 li" 33 13 TX TINER RMC0082377 * Vv V* . ' - 'x '-r-'**v-- ' i:\CTDEXC].: of CAXCEP. OF T.UXO AXD LAP.VXX s:j Table XCIII (conk!.). 7 or.i'ts tn<l thf?ir A-Sdistanfs Uan'dre-ser* . P.iUvis and paltry cv>oks i'iufV-^ionnl men .... f-Harmon Pottery. etc.. kiln nud ovenruen, kiln setters an 1 placers Orootns and horsekt-epws Brick kiln and overunen Mattunary engine and crane drivers . f'cttun 'trippers, grinders mul card room jobbers Locomotive engine drivers, firemen and Heanera * rarponters ..... Mone miners and quarriers 1 oal miners--hewer* an*l getters Lithographic and process engravers . Hiast furnacemen and Labourers < iardeners , . , , 'ottvn spinners dud piecrs (mule, ring, cap or flyer) heal miners--workers above ground Tar distillery workers and coke-oven workers brick and plain tile m-tkers 1 u*d miner*--other workers below grmmd d miners--conveying material to the shaft Loal miners--making roads -. Cv#ttou weavers .... farmers . . .V.ri'-uhur.d hibourers, including d-ephtru bailiffs and foremen Lung Larvnx Lung Larvnx Lung Larvnx Lang Larvnx Lling Larvnx Lung I.arvnx Lung Larvnx Lunz Lar\*nx hung Larvnx 1 ' H'g Larvnx I.ung I.aim.x I.'mg Larvnx Lung Larynx Lung Larvnx Liing Larvnx Lung Larvnx Lung Larvnx Lung Larvnx Lung Larvnx Lunz Larvnx Lurz Larvnx Lung Larvnx I...!.-- Larvnx Lung Larvnx X. llt. r Larvnx Lung Larynx Luaz Larvnx Lunz Larynx 191*. 1 3 0 3 in 3 25 8 2 2 I -- 3 4 3 __ 9 7 __ -- ii 3 25 21 0 51 2t> 2 -- __ i 21 16 4 i 6 7 1 ----- G 2* 3 2 6 2 3 3 17 20 17 21 2 i 4 3 8 > \\ 7 29 3 -- 7 i 5 s 2 ___ 2 9 4 -- __ 13 5 31 14 > 2 49 14 1 2 -- __ 31 23 4 A ii A 7 __ -- 7 i A -- 11 4 5 __ ,, 24 12 21 19 -- 19J-V i 3 4 3 12 .> .77 4 __ 3 .7 __. 5 4 .7 ___ 13 C ___ -- 10 1 35 17 1 7 51 24 -__ 1 ---- 3S 11 3 2 5 3 .7 19:10. f) __ 11 __ 15 4 34 9 1 1 2 2. 5 __ 1 __ 11 2 2__ -- 13 2 31 18 5 1 40 14 -- 1 1 1 40 25 6 3 11 5 -- 1937. 3 3 6 -- 12 5 26 10 1 a 3 1 2 3 -- 1 14 2 i 1 13 5 35 15 3 1 62 20 -- 2 i 1 41 20 -- 7 3 -- 103$. 3 2 13 2 18 3 35 5 2 2 3 2 7 2 i -- 17 3 3 1 13 2 40 12 4 -- 66 15 1 -- 1 1 42 28 3 3 9 6 1 2 1 2-- -- ------ 10 10 45 7 12 24 53 ---- 4 11 14 s2 7 4 7 4 33 3 3 3-- 1 224 21 29 27 30 8 14 16 11 17 31 23 28 28 23 22 22 111I 2 Lung Larynx . Total. 17 14 47 10 76 24 17t> 39 6 12 12 10 30 17 9 3 75 24 6 2 66 18 200 97 17 12 319 107 4 5 4 4 213 123 27 11 49 26 6 5 -- 52 IS 28 7 37 24 17 12 143 81 137 135 6 3 3530 1131 i I - . r. ?v {* ?. f jm ! I i; - i, if, Mi :i *f It cr *r \: jp ,/* TXTWER f RMC0082378 . V iiS4 r!ii~ rtfnrfiiffi^i>i K L. .VXD X. M. KF.XNAWAV i .1 A. Ratio of rngistered to K0 oalonlatefl deaths, 1921-IO't:?. B. Ratio of iegi*tercd to lnf> calruhtted deaths. 1033-19*48. 1 Dot* represent B A :< 100. j Occupations with not more than one death in either period are omitted. Table XIX.--Increase of Deaths Attributed to Cancer of Lung in Certain Occupations (1927-193S). Death* from can'"vr of lung. See Table VI - ' Paviours. street masons, concre- tors, asphalters Council labourers, road sweepers, ` dustmen .... Drivers of horse-drawn vehicles . Motor dlivers . 1027-32. 7 47 11) 102 1933-3S. 14 . 103 . 240 . 301 Difference. .7 . 56 129 199 207 65S 391 Genera! population, males .... 0290 . 15,04o 8749 For example, if one takes the numbers of deaths from cancer of the hmg in those occupations where there is especial exposure to road dust (Table VI) and divides them between two successive periods of 0 years (J927-32 and 1933-3S) and compares these figures with those given by the whole male population, it becomes quite dear that even these occupations account for only a small fraction (391 cases or 4"> per cent) of the whole increase of 8749 cases (Table XIX). The following example shows the effect of the increase in deaths attributed to cancer of the lung upon the data for a single occupation. t 1 i ti < 1 i t TXTINER RMC0082379 Kfe.. r.XCIDKXC'E OF CA-VCEU OF LUXG AXD LAF.YXX 285 219.431 carpenters produced 123 deaths from c-anoer of the lung in 12 years A : > A -- <:!is . -r A!- 1 , rhr Fe -J "a i\s ; ! ill'h-AS) 2 !'> 92s .' j . i' l i (ji ui Uct-Li 2-no ^ii;i i i e.ii ii.'i, or >.*..> .--M l ii.*, <' i*i-a*i! , i.i;he IV The munher of carpenters producing 1 death in 1 year was therefore 210,431 -y ! 2 - 2o,f40 in the first period and 243.92$ -y 33-3 -- 7324 in the second, fee l.itio between the two figuu-s, namely 3-3 : 1, is reasonably near the inverse f th ratio (I : 2-8) between the mean annual deaths (755 and 25>o7) in these ;mj periods from cancer of the lung in the whole male population. The result indicates that the increase of cancer of the lung has fallen upon carpenters, who undergo considerable exposure to some kinds of dust, in much the same measure .I' it has upon men in general (Fig. 7). , 1921 1925 1940 I9J5 I93S Flo. 7.--])eRrh-< front saucer of lung. Carpenters 1021-1033. Occupational exposure to three agents winch are perhaps.liable to induce cancer of the lung (asbestos, arsenic, nickel [Factories Report, 1038, p. 72]) occurs in employments which compri'e very few workers, and hence have no appreciable effect upon the total. Of these three substances, arsenic seems to be the one to which the general population is most likely to he exposed. No attempt is made here to deal with all the data on the carcinogenic action of arsenic, as this subject is under investigation by a Committee of the Industrial Health Board, but the following quotations from recent Annual Reports of the Semor Medical Inspector of Factories show the possible importance of arsenic in relation to eunec-r of the lung. " In the ca>e of arsenical poisoning due to sodium arsenite, which was fatal, the post-mortem examination revealed that in addition to pigmentation of the trunk and limbs, warty growths all over the body, and perforation of the nus-il septum, there was a primary cancer of the right lung' with metastatic growths in neighbouring glands and in the liver'' (Factory Report, 1939, p. 22). " The other (case) occurred in a filing machine operator aged 57, for 43 years in a factory manufacturing sheep dip containing sodium arsenite, the ' au-c of death being due to carcinoma of the light lung." " Three similar cases of pulmonary carcinoma occurring in arsenical sheep,.ir workers have been notified since 1939 " (Factory Report, 1943, p. 45). Xcuhaiic-r (1947) collected the literature relating to 142 cases of hud a:>-:.k-ul epithelioma : in one of these cancer of the lung occurred Montgomery. 1935; Montgomery and Waisman, 1941). The patient had taken 2<i -I .1 ! '-I TXTINER RMC0082380 . *> Y* - .-,4* " v' A *. . : Z.. t 1 -" ** 2 SO E. I.. AXD X. V!. K.EXX.VAVAY a medicine for eczema for dx years in childhood, which was assum-d to contain ar.-enie on account of tiio appearance front the age of 42 onwards of numerous c>i! '!><>vn-'> of the 'kin a . senic.d ', r- r to :r.ia:lira! oiu.iOm >' v.'.'ped, and dually -.u tpuheiioma- ol i lie broucbti* with "malignant dvs- keratoVis characteristic of the arsenical type of epithelioma," and extensive metastases. Semcm <10 45) records a case of bronchial carcinoma in a man who, during a period of about 20 years, had taken large doses--up to 1 drachm 3 times daily--of Fowler's solution (equivalent to 1 per cent arsenic trioxide) for recurrent attacks of severe dermatitis herpetiformis. His skin was unaffected. There is, of course, nothing to show that the bronchial carcinoma was due to the arsenic, but unless attention is drawn to such a case others may not be noticed. Table XX.--Cancer of the Lung. Occupation* Showing a Considerable Rise in the Ratio in the Later Period. 1921-32. A. 2iWDier Population. of death*. Potter's mill workers, slip. makers and arkmen I,4S<> i Onnorki engine ami crune drivers .... 1,S34 i Cotton spinners and piecers 3S,817 7 Coton strippers and srrinders and caul room jobber? . 3,242 i Tar distillery workers and coke-oven workers . Briek kiln and oveiunen 9,697 6,225 2 i Jliui.decs of other religious bo<lies .... 11,769 5 jutio. 1933-3*. B. X umber Population. of death*. 96 1,766 4 6u 1,800 32 . 43,901 5 27 30 5,342 6 20 7,824 6 23 7,112 9 38 11,960 15 - Eatlo. Kutk B,A x 300. 218 . 188 . 75 104 . 70 121 . 69 . 227 310 234. 347 242 Ol'S 1S2 (1) Cancer of the lung. In 7 occupations the ratios for 13 are higher than those for A by more than 50 per cent (Table XX and Fig. 5). In all of these the ratios in A are lower, and in most cases much lower, than 100, and hence the considerable rise raises the ratio in B above 150 in two cases only (potter's mill workers, slip makers and arkmen 218; gas-works engine and crane drivers ISO). In 4 of the 7 occupations there is only one death in A, and hence the comparison of the two periods is liable to considerable sampling error. The fact that the rise starts from a low level does not, of course, prove that it is-unimpovtant, and these occupations must be Hatched in the subsequent years to see if the rise continues. Cases of cancer of the lung have occurred among stokers of oil-fired boilers, but no figure for the population of these workers is available. (2) Career of the lari/iu. In 9 occupations the ratios for B arc higher than those for A by more than 50 per cent (Table XXI and Fig. 6). In 6 of these there are not more than 3 deaths in the Gist period, so that the sampling error is high, and the ratio in the first period is low. TXTWER PMC0082381 V"o ; 4^ :.t*f-^*>:-^`":-!C: K ' i**"KV4' :.- *-.Jp:'-v. .-%j?i,iita&f*o,^>t-i;*;p^-"" \^.1.1 ;"**-< *r>* Ti"/ :h't- ~i,iA^i^iT ~i * itriifTi ~i \ IXCTDEXCX OF CANCER OF LUXf! .VXD LAl'.VXX 2S7 7/<-se oa /< Jiatio tu tins toiler 1 vtj.-j. 10' l-.ii. A. - >*ut:ib*r Prvu'.itiou. cL-jrl ?. (Vitamu'ii 6,333 14 JVt.il criiiUi.-r> . 15,220 IkicU-kiin nnd oven men . 6,223 Id 3 Potters', etc., kiln and oven men .... C otton strippers and grinders 12,18$ 9 and card room jobbers . Tobacco mniitifrtoturers Gas works managers . 3,242 7, Oft* 1,363 3 1 Lithographic and proofs engravers 3,704 3 furnaeemen aud labourers 0,391 3 Ratiu. 317 m 71 116 60 63 76 84 40 u ;-u?. b. -1 PcpiUjti'>/i. or dt-itii.s. 7.05n 17.0i*!> 7. iii 12 0 o . 14,442 5.342 8,192 1,649 10 ) 4 1 7,289 5 e,i35 4 ltntio. 501 . 234 123 220-5 US 129 r>7 214 138 . Ru'W 15$ 160 173 190 197 198-5 207 255 346 ERRORS IX THE DIAGNOSIS OF CANCER OF THE I.CXG. Some hold that cancer of the lung is not increasing, and that the ascent of the curve (Fig. 1) is due to improved diagnosis. This statement is easy to make and difficult, perhaps impossible, to disprove, and is apt to serve as a substitute for any investigation of the question. If one accepts it, at least one conclusion follows which is of some interest. Let us suppose that the steeply rising line (Fig. 1) represents the detection of an increasing fraction of a pool of cases awaiting discovery. When the means of diagnosis, and their availability, can be improved no further, the curve must flatten. We cannot say when this will happen, so for the present we must take the latest, and highest, figure (59X2 deaths of males in 1915) as representing the total. If the actual number has altered little during recent years a similar number of deaths took place in, say, 1921, but only 361 of these deaths (Fig. 1) were assigned to the right cause. Hence in that year in at least 59S2 -- 361 = 5621 fatal cases of cancer of the lung the cause of death entered upon the death certificate was wrong. This is of course possible, .hut it would be interesting to know what were the 5621 erroneous descriptions oifatal disease. Some conditions which might give rise to false diagnoses in cases of cancer of the lung are* : (1) llc'b.nitinal tumour.--Tiie increase of cases of cancer of the lung cannot depend wholly, or even largely, upon a diversion to that category of cases formerly Kt down as mediastinal tumour because the number of these has never been 'efficient. Thus the death certificates for males show no fall in deaths attributed to cancer of the mediastinum during the years following 1931 (when this form of cancer is first mentioned in the Registrar-General's Statistical Reviews), which " vai 1 account for the increase of 7024 in cases of cancer of the lung which took hi we in tlie years 1931-37 (Table XXII). * I am i-. ! btcu tj Profe?x'r D. W. .Suiithers, Professor R. A. WiJli* and Dr. Eric Scoweu for io;* ?;bjept. '' ' t .... i ` ' . ' j i ' ' ; !/. . j | ! i': : i !!"'. i j: ^ i iiid i ... - - . . ' 1..: ,, 1 -i ; \f` ,U ; *.:- ;| .' ii , ;j ; ; | '; '` !i . ,' i '1 ; i j j ij. : '! :i I 1 ;| 'v ' TXTINER i RMC0Q82382 i 'V:. *.. I Vs-e",."* ;i. : . -l&A >{ - , - y- - .s-\rZ E. L. AXH X. M. KEXN'AWAY -TAr.r.E X XT r.--C 103) .... 1032 .... 1033 .... 1033 .... 1935 .... 193(3 .... 1937 .... of Lava ami 'l M d:n <iin i, n.. Jj<M th-* of males. England and Wales. Cancer of lung. 1,055 1,553 1,820 2,095 2,335 2,611 2,930 Cancer of mediastinum.' 135 2(32 202 244 24& 222 308 Total 7 X 1055 . 14,409 7,385 16S1 Increase . 7,024 * "flies? figures are taken from the Registrar-General's Statistical Reviews for the years- in question. (2) Pulmonary infcclidns.--A long list could be compiled of infective processes mid tbeir results which might have to be considered in the diagnosis of cancer of the lung in the living subject, but those most likely t-o be named in error on death certificates are tuberculosis, lung abscess, pneumonia, and bronchiectasis. In theory, the numbers of deaths attributed to these conditions should be falling if increasing numbers of cancers of the lung are being detected among such cases. But actually the numbers of deaths attributed to respiratory tuber culosis, and to pneumonia, are so large, ranging from 2 to 10 times those assigned to cancer of the lung, that the proportional fluctuations prevent any diminution which could be due to this cause from being identified (Fig. 8). In this figure the two curves representing pulmonary infections show no uniform relationship to the third representing cancer of the lung. The numbers of deaths from abscess of the lung (Table XXIII) are too small (mean 100 per annum) to be of importance, and moreover they show no continuous fall during the period. Bronchiectasis is not separated from bronchitis in the Registrar-Generals Statistical Reviews. Table XXIII.--Death* of Males front Abscess of Lung. England and Wales. 1932 1933 1934 1935 1936 . . . . 92 82 91 111 . 102 1942 . 1937 193S 1939 1940 1941 100 . . 96 132 119 84 92 wr?:rw: txtiner RMC0082383 :vr-^ INCIDENCE OF CANCER OF U'XO AND LARVNX i l Fir, 8.--Deaths of Males. England and "Wales. A. Tuberculosis. Respiratory system. B. ^Pneumonia, all forms, -r 2. C. Cancer of lung. iData- from the Registrar-Ger.r-ud`s `Statistical Review of England and Walts f^r the Vear k`42. Tables. Part 1. Medi-.al.'; P-'.rccniage of trroneovs diagnose*. Some- pathologists have emphasized the very high proportion of error in death certificates in general and have produced evidence that the percentage of erroneous diagnoses, as revealed at autopsy, is considerable even in hospitals, and must be greater still in general practice. Thus Willis (1941) found the diagnosis to have been "erroneous or indefinite" in 15 out of 27 cases of cancer of the lung coming to autopsy at the Alfred Hospital, Prahram. Melbourne, and in a larger series (private communication) of Si autopsies performed by him on cases of pulmonary carcinoma dying in a major teaching ho-pital, where Table TCXIV.--Analysis of Diagnoses in S4 Cases of Cancer of the. Lung Examined at Autopsy (Willis). Diagnosis correct Tumour, but primary site wrong or uncertain Dia; Xon-neoplastie disease (mostly tuberculosis, abscess, .pneu monia. bronchiectasis) 10 49 = 5S per cent 35 =--- 42 II 84 - .I * 1 -1 ** it; u l* \i i ! >t I \*i'. , IX TWER RMC0082384 ; *7 - '-V -- ; V.- '- * i * -f '-A A- *; 290 E. L. AXD X. M. KEXXAWAY then- was a special surgical t-linic dealing with Umg ruses, and where full X-ray i;u 'r. Ik.-i i-Mmc ill-. :r .i-.l! -2 r '-.'i1 uf t';.- dnicat (iiivj. news were wrong ('table XXIV). Hence the diagnosis of tant-ec of cue Jung appears to be still very incomplete, even under favourable conditions, in spite of the great improvements which are generally stated to have been made. At any rate one cannot well emphasize at the same time both the high proportion of cases (e.g. 42 per cent) which escape detection through inaccurate diagnosis, and the adequacy of the improvements- in diagnosis to explain the apparent increase in cancer of the lung, though the two statements are of course com- patible. Willis does not give any information about false positive diagnoses of pulmonary cancer. Professor Willis (private communication) has recorded a very instructive example of the statistical effect of diagnostic errors. In a series of 1000 con- secutive autopsies which showed the presence of cancer, 155 were found to be cases of cancer of the stomach, and of these, 109, or 70 per cent, had been diagnosed correctly and 46 had therefore received incorrect negative diagnoses. There were also 40 eases diagnosed as cancer of the stomach in which tliis condition was not present. In this group of 1000 cases, therefore, 109 + 40 = 149 cases had been diagnosed as cancer of the stomach, and the actual number present was 155. In so far as the total is concerned therefore the clinical answer to the statistical question--how many cases of cancer of the stomach were there in these loOO cases ?--was very nearly correct. The far greater error in the diagnoses of individuals, of whom 80 were described wrongly, is concealed statistically by the approximate cancellation of the 46 false negatives by the- 40 false positives. This process of attaining accuracy by the balance of positive and negative errors is, of course, neither desirable nor reliable, but it is unavoidable in such material, and is bound to happen to a greater or less extent whether we like it or not. When any decisions are reached on inadequate evidence in any subject, there will be the same tendency for errors in opposite directions to cancel each other. FACTORS WHICH MIGHT CAUSE AX IXCFEASE IX THE DEATHS ATTRIBUTED TO CAXCER OF THE LUXC. The increase in recent years in the number of deaths attributed to cancer of the lung might be due to one, or more than one, of the following factors : (1) An increase in the actual nuuiber of cases of the disease. It is obviously difficult to obtain any direct evidence of tins factor; one is driven to seek for indirect evidence by means of enquiry how far the second and third factors can be excluded. Some possible causes of a real increase are con sidered in the next section. In the earlier paper (Kennawav and Kennaway, 1936) we drew attention to the evidence afforded by cases of cancer of the lung in which an erroneous diag nosis was made based on nervous or mental symptoms due to metastases. Thus Fried and Buckley (1939) record that metastases in the central nervous system were found in 15 of 33 proved cases of cancer of the lung. In 11 of the 15 a diagnosis of primary tumour of the brain was made and the luug tumour overlooked : an intracranial operation was performed in 11 and tho metastases ; , | ; j j I 1 . r ' . - Vr * . -- TXTINER ;RMC0082385 INCIDENCE OF CANCER OF LUNG -AAD LARYNX 291 t.-'.-'o'i -'1 in 10 ease:*. Obviously nr it her improvement nor fnshiou in the x ,v* i nr of the luns Gee below) come into play here, and hence if it can. w.-'i: c.us-^ h.v, jj'iy - Ios. ~ i ci ir-.ig. -t ;tv! ip cr.'i1 `-o in C "i' r-r of the lung ia indicated. ;i'i improvement in diagnosis, whereby a larger proportion of the actual number of cases is detected. The methods of examination of the lungs, and the availability of apparatus for such purposes have undergone very great development in the last twenty \ t ,u-. and it would be remarkable if these changes had not increased the number of cases of cancer of the lung discovered. This matter is discussed under `` Errors in diagnosis '' above. According to some authorities (e.g. Willis) these improved methods, even when fully available, still leave about half the cases undetected, so ue may look for a further considerable increase from this source. (3) A tendency to identify as cancer of the lung cases which are really of a different nature ; that is, a fashion in diagnosis. Wc have not found any data which would enable one to judge whether this factor is of any importance ; if it is to account for the fovm of the curve shown in Fig. 1, tw o or three hundred more of such errors must occur every year. Willis (p. 2S9) dues not mention such positive errors in the diagnosis of cancer of the limy. SOME POSSLBLE CAUSES OF CANCER OF THE LUNG IX MAX. Experimental pathology lias not yielded any conclusive evidence on this fjuedion. as it has done irrthe case of the occupational cancers of the skin. Most of the laboratory work, such as that of Argyll Campbell, has been done on the mouse, in which species the neoplastic factors in the luijg are certainly different from those in man. (1) In default of experimental evidence, indications of at any rate one direction in which to seek are derived from the material of the General Register Office. The>e data are compatible with the importance of some factor in town air, e.g. coal smoke. () C'ancor of the lung differs from that of the upper alimentary tract. larynx and skin in showing no inverse relationship to ascent in the social scale (Steven son, 1923). Owing to the mixing action of the wind* there is less difference in the outdoor air breathed by different classes than in other social conditions such a~ food and cleanliness. () The higher incidence in towns (Stocks. 1936). Table XXV shows that. " there is at every age a steep downward gradient of rates front London through * An elaborate study of the mixing of air-borne particles was made at The effect of wind appear* be not so great as one would expect. " The maximum effect of wind is to displace ti* point of maximum concentration by l raiie ** (Leicester, 4 Atmospheric Pollution,' 1945, p. 1*23). B-it it is difflcu!t to believe that, say. an east or west wind Mowing across a city does not havo a '*n-i Mrable effect in equalizing the condition of the air in the cast and w*$t end. Groat importance ,* attached uUo turbulence as a mixing agent. 41 Xo exact definition of turbulence has ret been .ua le, but the word ha$ been used, in this report and elsewhere. to describe tlie readiness with which dir with itself. During the lad nin* months of the Leicester Survey* special attention was rr. measurement of variable connected with turbulence, because of the importance of T..:b,.V o ir. dt***mlnirz the mncer.tre.tirtn of suspended pollution. It was thought that one of *.i v*e v.iriib:os rnijht pr.'iv -suitaM:*' for more general use in connection with measurements of -pht-rie p r.--.icn ** tp. 111}. App.ir<*ntly turbulence includes all forms of disturbance due. to heating. : < ' "v V i. \ 1: . Ij I : :i a d: !l :l n: i:: j, TXTINER RMC0082388 292 E. L. .VXD X. M. KEXX.UV.W loy,re -,p .! muhI! towns o tuval Ii-rTi.f- " -il". ' :'v vVrr. r` [ ' iOil<i l towns vi.ii'. a low mte ve.g. ilmtoi). F;ni|)iriviiiy, smoke concentration in a town is in proportion to the .square root of the population (Leicester, ` Atmospheric Pollution,' 1945. p. 124). (r) The low incidence in agricultural workers (lvenuaway and Kennaw.ty, 1930). There is an equally low incidence in coal miners. (6) ancl (c) might be explained by the better conditions for diagnosis in urban areas. Coal smoke does not account well for an increase in cancer of the lung during a period when, as one would think, atmospheric pollution has not increased owing to greater use of gas and electricity. The domestic fire is the chief source of soot in the atmosphere,* and it is in the domestic field that the substitution of gas and electricity for coal has been most general. One must, however, note that ten years ago the authorities on atmospheric pollution were not unanimous on the question whether there had been any measurable diminution.! i | j { . i Table XXV.--Cancer of Lung. Mates. Englawl awl Wales. 1921-1930. (Stocks.) Rat es per million at ages shown. - ' 25- 35- 45- 50 - 55- 00- 05- 70- IO--bi. London . . 15 . 55 . 121 . 193 . 253 ., 263 ., 236 .. 199 . 215 C'ountv Boroughs Other Urban District .$ 7 . 35 . 5 . 23 . 73 45 . . 1f0"J 3i . . 135 90 ., . 163 12S . . 148 150 .. .. 144 118 . . 03 93 Rural Districts . . 7 . IS . 30 . 34 . 03 . 73 . 87 . 92 . 71 (2) One must look, therefore, for some other factor to explain the increase, if there is a real increase, in cancer of the lung in recent years. Many such factors have been suggested, e.g. influenza, tobacco, popular drugs, tarred roads, i exhaust gases and other emanatin.- , including lead ethyl, from motor vehicles driven by petrol or Diesel engines. No attempt is made here to discuss all these possible factors. * " . . . the percentage of tar in domestic smoke i> very high, reaching 30-40 per cent. wlJle in industrial smoke from boiler fuma ?cs and *uvh there is practically no tar *' (' Atmospheric Pollution/ I2th Report, j>. 4*i). '* A high deposit of tar can only be due to pollution from domedi-' chimneys or very inefficient industrial furnishings, since in a properly constructed and worked furnace a hirrh temperature i* mjiotained for a sufficient dUtar.ce from the main place of combustion for all combustible vapour* to be fully burnt ci before the furnace gases leave the chimney '* {` Atmo spheric Pollution.' 23rd Report, p. 0). .Soot from the chimney of a dwelling house may contain 40 per cent of tar (Cohen and Ru-ton, 1025). and the dust suspended in the air of a town (` Leicester Atmospheric Pollution/ 1945; may contain 11 per cent of tar. t In the * 22nd Report on the Investigation of Atmospheric Pollution' (1035 193d, p. 5) th* Chairman snuia up some rather equivocal evidence thus: ** . . . on the whole there has been a definite reduction in the extent of pollution of the atmosphere in Britain during the past twenty yea/' On p. !l of the same report, honever, the conclusion is drawn that no general statement can be dcriv**l from the increases and deeiva.se* in different areas during 1925-1035. The special study of atmospheric pollution in Leicester Leicester, Atmospheric Pollution/ 1915) showed no difference between the two period* 1927-1932 and 1933-1939 in the amount of tar in deposited matter ; the report makes the ujge.*.ton that a decrease in tar might be masked by un increase in material from motor cars. But the inconclusive nature of the data shows that there has been no increase in atmospheric pollution in any way proportional to the increase in cancer of the lung. TXTINER RMC0082387 "let - *- -< ,S. -r " - .4 .-4'. % rv~*'v?*i -a - '* rT-.*'*ff '* ' V - . . *'\--',,'" *r"*>:'..'' fV^rCg-*'*-*^J,*' !*.< "1_''*. - mzr.*nii*fl ':vr. W.rti'.w.-s.^w*-.,-,-y.^ -ji.^ ligaiM 1I1Wil<y;naaiSiT))tiM"?*. ~nn J ' . .' 5 *'^ 1 f-.-: LSTIDEXCE OF CAXCER, OF 1ANU A>'0 LARYNX 2 93 (" i The subject of tarred roads was dealt with in the earlier paper (Kennaway ... ! Keimiuvav. 1 93K). . ' e inchteVd'e '' lung \.ii-er on ch-.,> i-cnpatkrti* tvhero uscre is pome to road dust- (see above, Table VI) is rather above the general level, it ' rnis very questionable whether the recent increase of this disease among the jcueral population can be attributed to the tarring of roads, for scores of year* before this increase began coal tar was being discharged into the atmosphere in the form of soot by the domestic fire in quantities vastly greater than any re it could now be derived from roads. Tarring of the roads has. undoubtedly increased the amount of dust derived from tar in the air, but this increase appearsnegligible in comparison with the amount of soot present already. One sees clouds of smoke drifting over any large town, but one does not see any similar clouds of tar-laden vapour arising continually from tarred roads. The annual consumption of coal for all purposes in Great Britain in 15)36-1938 was about IsS million tons, of which 40 to 45 million tons was bunted by domestic users, w ho are by far the greatest producers of smoke and soot.* It is difficult to believe that tarred roads made any important addition to this total. One must, however, bear in mind that the total amount of soot per unit vulume of air may not be the only important factor. The size and physical state of tar particles derived from the domestic fire, and from a tarred road, are no doubt different, and this difference may determine the amount of penetration into the deeper air passages (see, for instance, Bo viand. Gaddum and .MacDonald, KMT ; Oweus, 1923). (6) Tobacco.--A possible connection between tobacco, and especially cigarettes, and cancer of the lung, haybeen suggested many times, perhaps most recently in the case of Turkey (Lancet, 1946, i, p. 435). Peacock (1943) has pointed out that cancer of the stomach is far more common in manf than, so far as we know, in any other species, and- has suggested that this js due to his use of heated foods. A similar argument might be applied in the case of cancer of the lung, which is not known to be prevalent in any of the lower animals. The adenoma of the lung of the mouse, and certain affec tions of the lung of sheep in South Africa and Iceland, are neoplasms of which the exact nature is uncertain. We know one instance at any rate of the su-eeptibility of the lung of an animal to a carcinogenic agent, namely the lung of the cat in relation to 2-acetylaminofluoreue given by the mouth (Harding, "46) ; hence there i< no reason to think that animals are immune to any such agents. One obvious factor, possibly carcinogenic, to which the lung of man alone is exposed Is tobacco smoke. During the present century very considerable changes have taken place in thi< country in the social distribution of the various methods of smoking. There "ere. of course, exceptions to any rule, but roughly one might sav that in.the ea.lier part of this period men of the richer classes smoked pipes, cigarettes and ci-.us, and men of the poorer classes smoked pipes ; cigarette smoking was increasing among the richer women, while women of the poorer classes did not Jn. )ke. One never saw a woman scrubbing her front door-step, or hanging out the washing, with a cigarette in her mouth. The great change which has taken ' A. Rariti.. V. 5'-!'.!< Fub!ir Lecture. March lath, 1045. * In loau in ..c J '.Vales, carcinoma of the stoieael'. wax assumed as the cause of 7070 act -,f deaths from cHremi:n.i of all sites in males (` Kegistrar Geueral's Statistical Review '). .i t ;/'*- ! \ 1a. ,* , r-t ! i' ; iu ( *~ i 1' f (i ' ; iI -`- rt I !: !- !j <' i 4 t! TX TIMER RMC0082388 >. ?>; / c~"?- i*.\ r..294 E. AX'D X. M. KEXrXTAWAY place in the later years is the general increase of cigarette smoking,* and it; adoption hr' women of chi-ses which fnimorlv did not smoke at all. Men Ln i I Fig. 9.--Heaths. England ann Wales. 1932-1942. A. Males \ B. Females j Diseases of coronary arteries, angina pectoris. C. D. Males \ Females J Cancer of lung. (Data from the Rogistrar-Ceneral's 'Statistical Review of England and Wales for tho Year 1942. Tables. Part 1. Medical.') Tlie $v retain- of State for War state.! in the House of Commons on January 23, 1945 (` Parlia mentary Debates (Hansard),' 407, C2H) that " In making up tho stocks of tobacco and cigarettes for the Army overseas it i.s assumed that there nr- nine cigarette smokers to every one pipe smoker. Th proportion is ba^ed on cxpvrionoe . . -x w.4---*:__ -j: ' u-.; ; INCIDENCE OF CANTER OF LUNG AND LARYNX 205 .upatinns which do not subject them to any restrictions in this matter, e.g. . - - nknicn. pxi:ttorsi witiiiow .:?< uers. dustmen and iwul sw(v;>ers, who i~'-'i to sinoKe a pip* at meal times, now-moke cigarette while at work. Cig.iietfa -tuckers arc said to inhale more than do smokers of pipes, but it is very difficult get any conclusive evidence upon this matter, which might be important. There is some American literature, which has been summarized (Brit. rn&l. J., ih-pi, i. 91) upon the arsenic content of tobacco and tobacco smoke. The occurrence of cancer of the lung in makers of arsenical sheep-dip (p. 285) indicates the possible importance of this factor. Some writers have contended that smoking cannot be associated with cancer of the lung, because this form of cancer has increased more among men, while riie use of tobacco has increased more among women. But the sexual distribu tion of cancer involves unknown factors (Table XVII), and does not provide a Tobacco 100 Cigarettes 50 J_J___L 1924 _L_L 1930 < i I l l l__ 1_J___I I .,1___1--1 1935 1940 1945 Fio. 11.--Cigarettes ns percentage of tobacco consumed in the United Kingdom, 1924-1813. \ ery secure basis for an argument of this kind. Of course no claim is made here that the simultaneous increases in the consumption of tobacco, and in cancer of the lung, proves any etiological connection between the two. Other changes which have taken place in the same period, which no one proposes to associate with cancer of the lung, e.g. the increase in the issue of wireless licences, show a very similar curve, and such correlations are a common subject for statistical witticisms. Thus wireless licences have increased at a rate (about t"-folc! in the last 20 years) similar to that shown by deaths from cancer of the lung, or from coronary disease (Fig. 9). The annual consumption of tobacco in the United Kingdom has increased from 128 million pounds in 1924 to 250 million pounds in 1946, and the percentage of these amounts smoked in the form of cigarettes has risen from 56 in 1924 to over SO in 1913-5* (Fig. 10 and 11). Thus the consumption of cigarettes shows a considerable increase, both absolute and relative. * We are indebted to Mr. J. Tt. Willis, M.C., of tbo Board of Trade for these data. The properT::a of ci.irettc-s to total tA-vco consumption is availablo only in those years for which a Census of hroduotijn is takes. The monthly net clearances of tobacco mainly by manufacturers for home .-/.Ioption in the United K'l.v lom of Great Britain and Northern Ireland arc given in the Monthly of XtctwtKJi. *r.***>' ) *r jl? :;f r,* V i foil ;-i, - !: ! i:. i!; 4 4[ l. j n. V;pr ''' " TtXxTtIiNnEeRr RMC0082390 V* V..i r-.r. W 4*'; - v* `A'.**i r' - fft ..-.-rTe** t V> is' 2'JO K. I,. .VXD X. M. KF.XXAWAV To summarise the ev idence brought together in this section on the incidence ..f c<.:.ct-r (T trie li.sig : Tito, higher iuoit-.d:dv in towns, the !u>v -nurtality in aericultural occupations. and the absence of social gradient, are compatible with a factor in the air. such as coal smoke. In any such comparison of urban anil rural areas, the question of facilities for diagnosis must be considered. Hut coal smoke, which is probably a decreasing contaminant of the air, does nor account well for the increase of cancer of the lung. Among various possible factors is tobacco smoke ; the consumption of tobacco has risen, and so lias the percentage of it smoked in the form of cigarettes, of which the smoke is often inhaled. An effect of tobacco would accord well with the absence of social gradient. SOBURY. (1) The death certificates for cancer of the lung and of the larynx in males from England and Wales for the years 1921-38 inclusive, numbering 38,41s. have been investigated and the periods 1921-32 and 1933-38 are compared. The 63 occupations examined employ about 30 per cent of the male population aged 20 and upwards. (2) Sources of error in statistical work on death certificates are discussed. (3) The increase in the recorded cases of lung cancer cannot be attributed to any increase of data obtained by autopsy. (4) The agricultural and coal-mining industries show a low incidence of cancer of the lung and of the larynx. -f5) A group of open-air occupations, where there is exposure to the dust of roads, has ratios above 100 for cancer of the lung and of the larynx, with the exception that motor drivers have a normal liability to cancer of the larynx. But the comparative incidence of cancer of the lung is not increasing distinctly in any of these occupations, and in the paviours, street masons, concretors and asphalters there has been a distinct fall in the ratio. (6) The occupations in which there is a liability to silicosis do not show a high incidence of cancer of the lung, but there are in the literature some studies of small numbers of cases in which the two conditions were associated. (7) Cases of cancer of the lung have occurred in some occupations involving exposure to asbestos. (8) In the death certificates examined, and in the Reports of the Chief Inspector of Factories, no occupations involving, exposure to any kind of dint. except those concerned with asbestos, arsenic and nickel, which employ very small numbers, have been found in which there might be an increased incidence of cancer of the lung. (9) Workers exposed to coal-gas and tar tend to show an increased prevalence of cancer of the lung, but in the later period studied the incidence does not exceed two-and-a-half times that on the general population. (10) Occupations concerned with the supply of alcohol have a high incidence of cancer cf the larynx. (11) The later period studied shows a considerable decrease in the occurrence of cancer of the lung in those engaged, in the preparation and sale of tobacco. (12) The very moderate ratio (125) for cancer of the lung in medical men U important in regard to the view that the recent rapid increase in recorded deaths from cancer of the lung is due to the detection of more cases by improved diag- / I | ! ' f j ! > , * i , j t *. i.- ' '. ,***.: <7. *: lNCI-DEXCE OF CANCER OF LUNG AND LARYNX 207 for thi* is an occupation where the availability of the existing methods for - .-Mner i-i pn-jiiina'ily .it A maximum. * ['') Xo special occupations have been found, among the 03 examined, to which flic increase in the total of cases of cancer of the lung can be attributed. T)ii increase is now so great that the incidence upon any such occupations o. cult! have to bo- very liigh indeed. (14) Xo evidence has been found that tarring of roads has affected the inci dence of cancer of the lung. .Such data as are available suggest that coal-tar i:i the atmosphere, whether derived from roads, domestic chimneys or any other Miurce, does not cause an exceptionally high incidence of cancer of the lung. * Cotton mule spinners show an especially Small liability to cancer of the lung, I although they inhale air sprayed with an oil which produces cancer of the skin. Much further work is required on the factors which regulate the penetration of jmi tides and droplets of various shapes and sizes into the air passages. , (13) The higher mortality front cancer of the lung in towns (Stocks), the low '. mortality in agricultural occupations, and the absence of social gradient (Stevent sou) arc compatible with an etiological factor in the air such as coal smoke, t But in any comparison of urban and rural areas, the question of facilities for j diagnosis rm.-t be considered. j (10) Soot is probably a decreasing contaminant of the air owing to the sub stitution of other sources of heat for the domestic fire, which is the chief source ` of soot-containing smoke. Hence coal smoke does not account well for any , recent increase in cancer of the lung. Among various possible factors which [ luu e been suggested to account for the increase is tobacco smoke ; the con{ sumption of tobacco has risen, and so has the percentage of it smoked in the form of cigarettes, of which the smoke is often inhaled ; such au effect of tobacco would accord well with the abs.ence of social gradient. b We are indebted to the Registrar-General for the data considered in this ! paper, and also for kindly permitting us to use some figures which have not yet , been niade public officially. We have received much assistance in the classifica!tion of occupations, and on various other questions, from some members of the Yuff of the General Register Office, to whom we wish to express our gratitude. | W\ wish to thank Dr. E. R. A. Merewether, Senior Medical Inspector of Factories, l tor information upon various matters, and especially upon asbestosis, and also Professor R. A. Willis for unpublished material. We are indebted to the Anna Fuller Fund for secretarial assistance, to the *a Halley Stewart Trust for a grant, and to Miss Blanche Goode, who has taken parr in the classification of death certificates. REFERENCES. Note.--Xo attempt is made in this paper to review the statistical literature relating to cancer of the lung in this or other countries. Am ersiix, C'. S., and Dible, J. H.--; 193>) J. Hyg. Camb., 38, 1S5. 'Atiiii.i-plieric Pollution, The Investigation of.' Reports 1-23. London: H.M. Stationery Office. Boylam>. E., Gaumm, J. H., and MacDonald, F. F.--(1047) J. Hyg. Camb., 45, 290. * 'Kkx, J. 1>., and Rtston, A. G.--(1923) `Smoke. A Study of Town Air.' London. Lu-i.f.. J. H.--' H'3-4) Lnno-.i, ii, 9>2. * r w k- S > *- i s*** :* v \ !s ; .: : * > --su* -1 i r-'Yj* L- I *-rl i* \ t. --e * 1 U:*r* j t- t ir--; r - '; i .<? i-, i; i J.!; i> , - r<*t i"i' I .V!|,U. H*#;**^*, *:* *: - ; ;- -i * -ryj, V -..y- ii - - ?-. ***.." TXTINER RMC0082382 298 C. C. SPIC'ER Egbert, D. S., and Oeicek, A. J.--(Ift36) Amer. Her. Tuber*-'.. 34. 113. ;. | r ictoriea, Office. Report ot thr '. L\.A L.-poi-ior or.' Lacfioii-: 11.51. St,Ki>'nvrv-. '- ? j * ` ;j Fried, B. 51., and Bcckley, R. C.--(1930) Arch. Path., 9, 483. 4* Gloyne, S. R.--(1935) Tubercle, 17, 5. , Harding, H. E.--(1949) 23rd Ann. Rep. Brit. Emp. Cancer Campaign, p. 92. j Henry, S. A.->-(1946) `Cancer of the Scrotum in Relation to Occupation.' Oxford- j Univ. Press. . .. J Idem, Kenya way. N. 51., and Kennaway, E. L.--(1931) J. Hyg. Camb., 31, 125. Kennaway, N. 51., and Kennaway, E. L.--(1936) Ibid., 36, 236. Klotz, M. 0.--(1939) Amer. J. Cancer, 35, 3S. v-* ' '. I Leicester, Atmospheric Pollution in. A Scientific Survey.' London: HAT. St?.-' . tionery Office, 1945. Lynch, K. 51., and Smith, W. A.--(1935) Amer. J. Cancer, 24, 56. ` Miners' Phthisis 5Iedical Bureau, Report of the, for the 3 years ended July 31, 1935.' Pretoria : Union of South Africa Government Printer, 1936. Montgomery, H.--(1935) Arch. Derm. Syph., 32, 218. Idem, and Waisiian, 51.--(1941) J. Imeet. Derm., 4, 365. NECBarER, 0.--(1947) Brit. J. Cancer, 1,192. - Owens, J. S.--(1923) Trane, med. Soc. Lond., 45, 79. Peacock, P. R.--(1943) Glasg. med. J., 139, 117. ` Registrar-General's Statistical Renew of England and Wales.' Tables. Text. London: H.5I. Stationery Office. Semon, H. C.--(1945) Proc. Roy. Soc. Med., 38, Section of Dermatology, p. 128. ` Statistical Bulletin,' 5Ietropolitan Life Insurance Co., New York (1939), 20, 7. Stevenson, T. H. C.--(1923) Biomelrika, 15, 382. Stocks, P.--(1930) 13th Ann. Rep. Brit. Emp. Cancer Campaign, 239. Wellis, R. A.--(1941) Med. J. Austral., Sept. 6, p. 258. dUeussef its value where A r 1, a will be I modified The i makes al estimate! at which x, not tl tinuous c the rate other har of the ar< Using cancer w life table The threi mice wer 0-6 THE ESTIMATION OF TUMOUR SUSCEPTIBILITY IN PURE LINES. C. C. SPICER. -- From the Imperial Cancer Research Fund, London, and- the Department of Biometry, University College, Ijondon. .. L Received for publication August 23, 1947. It is the purpose of this paper to discuss the problem of assigning measures of cancer susceptibility to inbred strains of mice, selected for their proneness to mammary cancer. The question has been previously discussed by Murray and Hoffman (1941), and the related problem of the assessment of potency of carcino genic compounds by Irwin and Goodman (1946), Twort and Twort (1930, 1933), Bryan and Shimkin (1941) and Lea (1945). It is suggested that the problem is most clearly examined by using an actuarial function, the " force of mortality," on whose characteristics. the mortality of a strain depends. This function, which Is usually denoted by the symbol ji(.j), measures the rate at which the members of a life table are djing out. It can be calculated for a single cause, 05 0-4 0-3 3 0-2C OK fIO. |."Odl t i V^:vvV