Document k69yzrrjLRyv3dqxGOvbb51GB

Brit. J. i/idustr. Afed., 1955, 12, 81. <y MORTALITY FROM LUNG CANCER IN ASBESTOS WORKERS BY RICHARD DOLL From the Statistical Research Unit, Medical Research Council, London (RECtIVED FOR PUBLICATION AUGUST 10, 1954.) Stxty-one cases of lung cancer have been recorded in persons with asbestosis (Boemke, 1953 ; Hueper, 1952) since Lynch and Smith (1935) reported the first case. In view of the infrequency of asbestosis, this large number of cases suggests--but does not prose--that lung cancer is an occupational hazard of asbestos workers. The strongest evidence that it may be a hazard has been produced by Nlerewether and by Gioyne.! Merewether (1949) found that lung cancer was reported at necropsy in 13-2% of cases of asbestosis (31 out of Z35) but rn only 1-3% of cases cf silicosis (91 out of 6.8841 and Gioyne (1951), on personal examination, found lung cancer in Hueper, 1952) consider that a causal relationship between asbestosis and lung cancer is either proved or is highly probable and the reality of the relation ship was agreed at the recent International Sym posium on the Endemiology of Lung Cancer (Council of the International Organizations of Medical Sciences, 1953). A minority, however, remains sceptical (Cartier, 1952 ; Warren, 1948), and, according to Hueper (1952), Lanza and Vorwald, so that it was thought desirable to undertake a fresh investigation. Necropsy Data l4l0o of necropsies on subjects with asbestosis (17 out of 121) against 6-90 in sdicotics (55 out of 796). Neither author gave full details of the se.x compo Since 1935, records have been collected of all the coroners' necropsies on persons known to have been employed at a large asbestos works.* Patho sition of the groups examined, but since women logical diagnoses in 105 consecutive cases are form a higher proportion of asbestos workers than of persons employed in occupations liable to give rise to silicosis (coal-miners, stonemasons, pottery workers, foundrymen, metal grinders) and since lung cancer is less common among women, the differences in ihe proportions of cancer cases cannot be accounted for by differences in sex distribution. In fact the proportions which are more properly com parable with the findings in silicotic subjects are the proportions of lung cancer found among men with asbestosis, 17-20 in Merewether's series and 19-6% in Gloyne's. summarized in Table I. Details of the cases in Table 1 CAUSES OF DEATH DIAGNOSED AT NECROPSY AMONG PERSONS EMPLOYED AT AN ASBESTOS WORKS (19JJ-J2F Cause of Death Asbestosis Asbestosis . Present Absent " Heart failure " Pulmonary tuberculosis Lung cancer Other diseases of the respiratory ssstem Other diseases }4 12 n 10 4 11 9j !I i 4! 31 All causes 75 30 All Cases 45 21 18 14 7 105 Animal experiments are inconclusive. A positive result was reported by Nordrnann and Sorge (1941) who found that of 10 mice which had been exposed to asbestos dust and survived for 240 days, two developed lung carcinoma. Smith (1952), however, considers that one of the " carcinomas " was, in fact, an example of squamous metaplasia and that the other, an adenocarcinoma, may have developed spontaneously from the common mouse adenoma. which lung cancer was found are shown in Table 2. During the first half of the period eight deaths occurred in which lung cancer was found in asso ciation with asbestosis, while in the second half of the period there were seven such cases and a further three in which lung cancer was found without asbestosis. The number of asbestos workers employed at the works increased steadily fiom 1914. A negative result has been reported by Vorwald and Karr (1938). The majority of workers (cited by * Necrops.es on asbestos orVcrs are ordered by the coroner when m hts opmu-n, there m.i\ be a question of a'.heNiosts being a contri butors c.ijyc of death. 81 ^ > .*-- V-.A~ 82 BRITISH JOURNAL OF INDUSTRIAL MEDICINE Table 2 OCCUPATIONAL HISTORY AND NECROPSY DATA OF ASBESTOS WORKERS WITH PRIMARY LUNG CANCER 1 Ycir , of Death I 1 Sen and Age Occupation Period of Exposure 1935 M. 62 Weaver 1919-32 1935 j M. 54 Weaver 1909-32 (936 ; M. 65 1933 1 M. 47 (939 : M 49 IWO 1 M. 52 1941 : M. 52 194; , M. 59 1948 1 M. 59 194S ; M 53 Fiberizer Weaver Disnfegratrr Disintegrator Weaver Bag carrier Weav er Weaver 1913-36 / 1910-12 \ 1920-37 /1910-14 \ 1919-39 ' 1911-1! ' 1919-21 ' 192 3-.'9 f 1913-19 \ 1924-33 19!1-41 / 19(2-14 \ 1916-48 1922-35 1948 | M. 45 Spinner 1922-48 1948 | M. 65 ( Maintenance man 1919-48 1950 j F. 51 . Spinner 1915-42 1951 19! M. 74 M. 60 Fiberizer NVeaver 1917-43 /1919-25 \ 1929-50 1 Years i of , Exposure ' 13 ; 25 I 23 19 24 22 20 I 28 32 13 26 29 27 26 27 Years of Exposure before Jan 1, 1933 Years from First Exposure to Death 13 16 23 26 19 23 14 28 17 29 15 29 14 19 16 10 , 10 13 ! 17 15 | 9 28 29 36 26 26 29 j 35 34 32 from Last E xposure to Death Pathological Report Ashe stosis Histological Type of Primary Lung Cancer 3 Present Less than 1 1 Lev than 1 * Carcinoma '' Epithelial carcinoma Endothelioma of pleura ** Carcinoma '* " Carcinoma " Less than l " Carcinoma " 3 1 Les* than 1 13 Less than I Less than 1 8i 8! 1 - Oat-celfed carcinoma .. Oat-celled carcinoma - Anaplastic carcinoma ** Carcinoma " M Carcinoma " Oat-celled carcinoma j " Carcinoma ** Adenocarcinoma - ** Carcinoma " 1944 1951 19?: M. 36 M. 43 M. 51 Weaver Fiberizer Wcjv er 1942 44 1939-48 1 1941 -5 12) ; 9 0 2 Lcsn than l Absent Oat-crlloi carcinoma 0 12 3 Anaplastic carcinoma 0 11 Levs tha/l 1 ** Carcinoma '' Also pulmonary tubejcuiosi*. and a great increase in the number of lung cancer deaths was also recorded among the whole popu lation of England and Wales over the same period. It might, therefore, have been anticipated that a larger number of cases in which the two conditions were associated would have been found in the last 10 years. National regulations for the control of asbestos dust were, however, introduced in 1931 (Asbestos Industry Regulations, 1931) and the precautions taken to prevent dust dissemination in the works had become effective by the end of the following year. All the subjects in whom the two diseases were found together had been employed for at least nine years under the old conditions, and although II of the 15 men and women died within 30 years of their first exposure, the association of the two conditions has not yet been found in any person taken into employment during the last 31 years (1923-53). It is, therefore, possible that the reason more cases were not found in the second half of the period.is that reduced exposure to dust has already begun to lessen the incidence and severity of asbestosis. Method of Estimation of Risk Although the necropsy data shown in Tables 1 and 2 suggest (1) that some groups of asbestos workers have suffered an increased risk of lung cancer, anJ (2) that the risk may now have decreased, it is not possible to be certain of either of these propositions without a more detailed knowledge of the whole mortality experience of the workers. The first proposition has, therefore, been tested by comparing the mortality experienced by that section of the male employees of the works referred to above, who had worked for at least 20 years in " scheduled areas with the mortality recorded for all men in England and Wales ; and the second proposition by comparing the incidence of lung cancer among men employed for different periods under the pre-1933 conditions. The investigation was limited to the small group of men who had been employed for at least 20 years, since the labour involved in searching out the individual records ot By " schcJukJ areas'' is meant fltoc arCis where p'octy-i'' were cirr.ed on v. Ireh wore scheduled under the .A^nixfO' Industry Regulation-. of |I .1' b.'ioj Jiistv. $-CoSS5S5 --fA' LUNG CANCER IN ASBESTOS WORKERS 83 men employed for shorter periods would be dis England and Wales aged 55 to 59 in 1951 was 2-77S proportionately great and. so far as was known per 1,000. The numbers of years lived in this age fiom Table would be compaiatively unrewarding. group in the five years 1949-53 were respectively The date of birth, date of completing 20 years' 15 years, 15 years, 171 years, 19 years, and 19 years. work in the " scheduled areas ", and, where applic The number of deaths expected in the period was, able, date of ceasing employment and date and cause therefore, estimated to be (15 -I- 15 -f 17.1 -- 19 + of death were obtained, for each man, from the 19j x 2-778,1,000 = 0-238. The total number of records of the firm's Personnel Officer. Full details deaths expected from each category of diseases was were, in most instances, already available for the men obtained by adding the numbers thus calculated for who had ceased employment as well as for the greater each age group for each of the five periods. number who continued to be employed, since some The great majority of the men lived and, when they of those who had left were registered as having died, died in the town in which the works was asbestosis and the attention of the firm had been situated, so that it would have been preferable to drawn to the death of others, in view of the have based the calculation of the expected deaths on possibility of the cause of death being industrial in the death rates observed in that town rather than on origin. All the remaining men were successfully the rates fora!! England and Wales. These, however, traced and the relevant details obtained. This was were not known in sufficient detail. Little error in not difficult since, by limiting the study to men who the expected number of deaths from king cancer is had been employed in one place for 20 years, few likely to have been introduced on this account since, were found to have changed their job or to have according to Stocks (1952), the age-adjusted death moved out of the region. rate for lung cancer among men in the town con From the data the numbers of men alive in each cerned was 96% of the rate for England and Wales. five-year age group were counted separately for each Stocks's figure was calculated only for the period of the years from 1922 (the first in which a man was 1946--49. but the propornon is unlikely to have recorded as having had 20 years'service) to 1953. A varied greatly over the longer period of the investi man who had completed the 20 years before the begin gation. The expected number of deaths from ail ning of a year and who was aloe at the end of it was causes is. however, likely to be somewhat under counted, for that year, as one unit ; a man who* estimated since the age-adjusted death rate from all completed the period before the beginning of a year causes for the town is about 25% higher than the but who died during it, and a man who completed the England and Wales rate (/.<?. the excess was 22% in period during a year and who survived to the end of 1950, 28% in 1951, and 22?0 in 1952). it, were each counted, for that year, as half a unit: the one man who died the same year as he completed his Results 20-year period was counted as a quarter of a unit. The number of men studied was 113; the numbers The causes of death were recorded as they were of man-years lived in each of the five periods in each given on the death certificate or, when available, as age group are shown in Table 3. The total number they were finally determined by necropsy. The causes were classified in five categories (see Table 4), and the numbers in each category were then compared with Table 3 NUMBER of MAN-YEARS LIVED BY MEN WITH 20 OR MORE YEARS OF WORK. IN A "SCHEDULED AREA " those which might have been expected to occur by multiplying the numbers of men alive in each fiveyear age group by the corresponding mortality rates for men in England and Wales over the same period. Because of the small numbers, however, the popula tions were not considered separately for each year, but were added together to form five groups living in the periods 1922-33, 1934-38, 1939-43, 1944-48, and 1949-53, and the mortality rates used for each Age (n cars) 50354C455055606570-- 75-7$ Period i 1922-33 1934-35 1939-43| 1944-43 1949-53' 0 0 5 15 : 0 0 , 4 5 2 M 17 5 9 , 9 5 16 33 5 : 43 55 i 95 19 5 50 1 78 5 34 i 6-5 25 5 39 5 85 96-5 ' 12 6 30 . 52 85 5 ; 15 3 5 25' 25 5 J6 ! 1 13 5 3 ; 10 21 5 ! 0 0 2 0 9 i i1 t-5 0'5 ! All PenxH ** 44 162 241-5 253 185 5 84-75 49 17-5 3 group were those for the years 1931, 1930, 1941, All agc< 1940, and 195!. The rates lot 1931 v.eie used for the ss ss 133 75 321 391 5 ' 1042 25 period 1922-33, rather than those for the mid-years, of deaths from all causes and the number of deaths since dispioportionately few men were under ob-.eived in each of the five disease categories, obscivation during the early part of the period. As together with the expected number of deaths, are an evunipV of the method, the mortality rase lor all show n in Table 4. Erom Table 4 it appears that the neopT.ui,;, other than lure cancer ntivr.g men in men who had been exposed to asbestos dust suffered 84 BRITISH JOURNAL OF INDUSTRIAL MEDICINE Tadle 4 CAUSES OF DEATH AMON'/j MALE ASBESTOS WORKERS COMPARED WITH MORTALITY EXPERIENCE OF ALL MEN IN ENGLAND AND WALES I No. of Deaths j Test cf Cause of Death i Expected ofDit!;rence on between | No. England1 Observed jObserved and | and 1 Wales | Expected Rates ! (Value of P) Lung cancer* : ! with mention of asbeuo-ifts . without mention of asbev:v>ss Other respiratory diseases (mc-od- i mg pulrr.on.ir> tuberculosis ?nd cardiovascular di'eases : with mention of as*vstosrv . . wtihout mention of asbestosis ,, 0 14 6 ______ 1 \ <0 000001 0-8 i | <0 001 76 Neoplasm^ other than lung caracer All other diseases* .. All causes . . A A 39 23 47 15 4 I >0 1 i ' <0 000001 fncluJi/Jg one case with pojmonary rubercuJosii. IncluJing two cases (berv>n stricture of oesophagus amJ septi caemia) m which ashestu-w- was present hut was not thought to have been a contributory causw of dejih. an increased monahjty from lung cancer, other respiratory diseases and cardiovascular diseases, in association with asbexrosis, but that their mortality from other diseases v.ns close to that expected. Four explanations cf the findings are possible : (1) that all the men wltvo had died of lung cancer were recorded because of interest in the condition, but that some of the records of other men dying of other diseases or still alive v*ere omitted, with consequent underestimation of the expected number of deaths ; (2) that lung cancer was incorrectly and excessively diagnosed among the asbestos workers ; (3) that lung cancer was insufficiently diagnosed among the general population of England and Wales ; or (4) that the asbestos workers studied suffered an excess mortality from lung cancer. It certainly cannot Ibe claimed that the records of the Personnel Office x*ere necessarily complete, but they were believed to be complete and no deficiency on this score would account for the total excess of deaths unless it were so gross that more than half the defined population had been omitted. Moreover, the number of deaths due to conditions unrelated to asbestosis was close to the estimated number and this is unlikely to have happened unless the population had been estimated approximately correctly and the deaths from alt causes fully reported. All the 11 deaths attributed to lung cancer were confirmed by nccrop-y and histological examination so that the excess n.-mber cannot be attributed to incot rect diagnosis .umong the group of asbestos workers. .Sonic of th-.. excess may well be tlue to an underestimation of th e expected deaths sir.ee part of the increase in mortality attributed to lung cancer over the past 30 years is certainly due to improve ments in diagnosis and in therapy (Doll, 1953). Even, however, if it were postulated that the whole of the recorded increase between 1931 and 1951 was spurious and that the real mortality from the disease throughout was that ascribed to it in 1951, the expected number of deaths is increased to only ![ and the observed excess is still grossly significant. For the actual number of lung cancer cases to be so little in excess of the expected as to be reasonably attributable to chance, it would be necessary for the expected cases to be 6-2, that is 5-6 times the number estimated on 1951 rates. In other words, it would be necessary to postulate that in 1951 (and throughout the previous 20 years) there was 5.6 times as much cancer of the lung as was recognieed in 1931, which would mean that the condition would have to have been present and capable of detection in over 20% of all men at death. Moreover, even if this were so, it would still not account for the fact that all the cases of lung cancer were found in association with asbestosis. It is, therefore, concluded that the fourth explan ation is the most reasonable one and that the asbestos workers who had worked for 20 or more years in the " scheduled areas " suffered a notably higher risk from lung cancer than the rest of the population. To test if the risk has altered since the 1931 regulations were introduced, it is not only necessary to make allowance for duration of employment before the end of 1932, but also to allow for the men's ages and for the total durations of their employment in the " scheduled areas ", since the nten employed in the earlier periods can also have been employed longer and lived to be older. On the other hand, there is no need to consider the changing incidence of lung cancer in the total population of England and Wales since the non-industrial risk has been shown to be small in comparison with the industrial one. The data required forcomparing the risks among men employed for under 10 years, for 10 to 14 years, and for 15 years and over in the pre-1933 conditions are shown in Table 5. The ages shown are the ages at death of the nten who have died and the ages in mid1953 for the men who are still alive. The expected numbers of men in each pre-1933 employment group found to have asbestosis or asbestosis and lung cancer are estimated by multiplying the numbers in each age, total employment, and pro-1933 employ ment subgroup by the proportions of men with asbestosis or with asbestosis and lung cancer in the same age and total employment group for all lengths of pre-1933 employment combined. For example, three out of the nine men aged 50 to 54 years who had MtX' cancer iprove- 1933). : whole /;1 was disease 51, the .illy l-l > j r scant. :o be so sonably . for the number ould be oughout -rues as in 1931, ..Id have ciion in . even if the fact uund in cxplanasbestos . rs in the . her risk '.uion. the 1931 necessary nloyment the men's doyment . mployed mployed ei hand, ncidence :!.md and n shown ' ial one. long men /ears, 3nd iitiorts are "c ages at . es in niid- expected ' ent group and lung umbers in ;3 einployircn with ccr in the :.'! lengths c .ample, v w ho had LUNG CANCER IN ASBESTOS WORKERS 85 Table 5 NUMBERS OF MEN EMPLOYED FOR DIFFERENT PERIODS DPTORE 103J AND NUMBERS KNOWN TO HAVE ASBESTOSIS AND LUNG CANCER IN ASSOCIATION WITH ASBLSTOSIS DrVtDED BY TOIAL DURATION OF EMPLOYMENT IN A SCHEDULED AREA AND BY AGE Total Length of Emplo> neat in " Scheduled Area (years) ' \%e at Jane 30, 1953, or a: Death (years) :o-2* 3540- 45505t_ 6->65-- 7075-9 0-9 Years Length of Em ployment before January 1, 1933 : 10-14 \ ear 15 4- Years All Lengths of Employ ment before January l, 1923 No. of Men 4 7 3 5 n 0 1 No. of , No. of Men swith Men wuh Asbes- < Cancer tons ; of Lung Vn ` ,f >lcn __ 1 __ 'i -- 1 -- 1i 1 1-- 0 2 1 -- i3 3 --2 t --l 0 --- 0 0 --0 1 i-- 0 No. of No. of M;a such Men with Asbes- Cancer tosts of Lung 1 __ 1 __ 0-- 31 l-- l-- ---- -- __ ---- No. of 1 No. of No. of 5 fen Men with Men with ' Asbes- ; Cancer l tosis 1 of Luag 0: 0' 1i 3: 0, 1 l 1 0; __ _ 1 3 o l i i -- __ __ 1 2 0 0 1 __ -- No. of Men | No. of No. of Men ith Men wuh Asbes Cancer tos is of Lung -> I 51 . 82 1 98 3 74 0 53 0 31 i 11 1I -- 25-29 40- 3 45- 10 50- 6 55- 6 60- 3 65- 1 l 70- ; 0 0 30-4 0 , __ 2 1 -- -- -- -- -- 1 __ l -- -- -- -- -- -- ,0 2 ,0 8 i 1 0 0 0 __ __ 0 1l 0.0 0 0 4(0 I i0 10 -- __ J __ __ -- --I __ __ a __ __ 12 -- .__ 6 1 -- 15 3 17 0 02 0 1 0 11 1 I 0 '-- 1 0 3 2 6 4 0 ] 0 l t 0 0 I 3^-34 4550556065- 70-4 2 2 1 l 0 0 __ -- -- -- -- -- __ 0 __ -- 5-- -- l-- -- 0-- -- 0-- -- 0-- -- -- 0 1 __ 2-- | __ i-- 2 9 o0 1! __ _ __ 3 --2 2[ 5 1 _ 13 1 11 --0 0 __ 0 0 1 __ --l 1 --l I -- 35- 55CO-- 65-9 0 ) 0 Alt lengths of employcnent ma " scheduled area " CO >xs.r) 35- l 40- 7 45- 19 50- 11 55- 12 60- 7 65- ' 3 70- ' 0 75- l 40-4 , 0 AJ1 ages 61 , J | __ -- -- _ -- 3 4 4 l 0 0 -- 13 0_ -- 0-- -- 0-- __ I --1 --2 _ I* __ 0 --1 --0 --2 11 1 8[ 3 1 1 5 --U --2 5 | 0 ,5 2 07 -- 1 13 -- .0 -- --- 3 -- 0-- -- l -- 0-- -- 1 1 26 14 1 3 26 __ __ __ _ ! 1 i , 0 0 __ __ ---- 0-- ___ __ 2) 1 -- _ 8 1 ! __ 1 3 (1 1 2 22 1 24 3 * 1 1 28 52 10 } 4 12 j 1 51 1iI 16 7 8t 2C*> 1 J 2 03 I |1 2 0 -- !I 25 2! i 0) 16 7 1113 43 j n been employed for 20 to 24 years in the areas in which they might be exposed to asbestos dust were found to have asbestosis and lung cancer. Since three men had worked for under 10 years in the pre1933 conditions, three had worked for 10 to 14 years, and three had worked for 15 or more years, the expected number of cases in each of the pre-1933 employment groups would have been the same, />., 3 x 3/9, or 1. In fact, the numbers of cases found uereO, I, and 2. The total numbers expected in each pre-1933 employment group are obtained by adding the numbers calculated for each of the age and total employment groups within it. The resutts are as follows ; The differences between the numbers of men observed and the numbers expected in each employ ment group, haJ the incidence of the conditions remained steady throughout, arc statistically signi- ficant (total asbestosis, /. = 7-52,n == 2, P = 0-025; asbestosis and lung cancer, /} = 8-74, n = 2, P = 0-01`). They are highly so if the trend, that is. Total number of merO1 observed w itb. asbestosis * expected Number of men v-ithl ob'cr'ed axlicitostv end lung j f rpected cancer Length of Employment before January 1, 1933 Under [ 10-14 1C Years j Years 15 Years and O'er 13 *J- , s-s j | j i 14 Ifl-J 3 2-4 16 IV-i 7 J-J The expected nu< ikrt of Un j cancer arc smalt and the probebtlip.. iftjl the O-iTerences c.iu'd .irtNC b> chance has con'equcntN been t!, but not < :r:c'j,t>T tindercjnVjatcd. IT a// men vrh more tl.m 10 v c:< rx pre-1 'J . t e-npro>inen: ar_* grouped together and ^ a*es* cn> t -c^riv; h rt.icv fur null number*. // -- 5-82, n - I, P OOd S~05S6568 86 BRITISH JOURNAL OF INDUSTRIAL MEDICINE the biologically important reduction in the propor One hundred and thirteen men who had worked tion between observed and expected numbers as the for at least 20 years in places where they were liable length of pre-1933 employment is reduced, is also to be exposed to asbestos dust were followed up and taken into consideration. It is clear, therefore, that the mortality among them compared with that which the mcidences both of asbestosis and of lung cancer would have been expected on the basis of the associated with asbestosis have become progressively mortality experience of the whole male population. leS-S, as the number of years during which men were Thirty-nine deaths occurred in the group whereas exposed to the pre-1933 conditions has decreased. 15-4 were.expected. The excess was entirely due The extent of the risk of lung cancer over the to excess deaths from lung cancer (I I against whole period among the men studied appears to have 0-8 expected) and from other respiratory and been of the order of 10 limes that experienced by cardiovascular diseases (22 against 7-6 expected). otfcvcr men. This agrees well with the data reported All the cases of lung cancer were confirmed histo by Merewethcr (1949), but it is somewhat greater logically and all were associated with the presence tha.o that suggested by Gloyne's data (1951). The of asbestosis. grerat reduction in the amount of dust produced in asbestos works during the period has been accom panied by a reduction in the incidence of king cancer arruong the workmen so that the risk Isefore 1933 is like.lv to have been considerably greater--perhaps 20 From the data it can be concluded that lung cancer was a specific industrial hazard of certain asbestos workers and that the average risk among men employed for 20 or more y ears has been of the order of 10 rimes that experienced by the general popula tutors the general nsk. Whether the specific industrial tion. The risk has become progressively less as the risk of lung cancer has yet been completely eliminated catanot be determined w ith certainty ; the number of mean at risk, who have been exposed to the new duration of employment under the old dusty con ditions has decreased. conditions only and who have been employed for a I would like to offer my thanks to the management of sutTirient length of time, is at present too small for con-.fidence to be placed in their experience, it is cie.-jr. however, that the risk has lor some rime been greiatly reduced. The extent of the reduction is particularly ,triking when it is recalled that between the firm concerned for permission to carry out this work and to the Medical Officer and members of the itac of the works where the men were employed, who carried out the greater part of the work oil which this report is baseJ. 19 "3 and 1953 the incidence of the disease among men in the country at large has increased sixfold. REFERENCES Summary The cause of death, as determined at necropsy, is repsorted for 105 persons who had been employed at onte asbestos works. Lung cancer was found in 18 instances, 15 times in association with asbestosis. All the subjects in whom both conditions were found hait started employment in the industry before 1923 anti had worked in the industry at least nine years before the regulations for the control of dust had become effective. Asbestos Industry Regulation* (1931). Statutory Rules and Orders, 19J1, No. 1140. H.M.S.O., London. Boemke, F. (I95J). Sled. Sfscftr.. 7,77. Cartier, P. (19521. Arch. inJusir.^ny^^ 5, 262 (contribution to discussioa). Council of the International Organizations of Medical Sciences (195)1. Acta L'n. ini. Carter., 9, 44J. Doll, R (1953). Bril. med. J., 2, 521. Glo>ne, S. R (1951). Lance:, I. 810. Huepcr. W. C (1952). >ceedi"ri of the Seventh Saranac Sympo* simn. To be published. Lynch, K. M.. and Smith, W. A. (1935). Amer. J. Cancer, 24, 56. Merewethcr, E. R. A. (1949). Annual Report of the Chief IntpeUor of rnotaries foe the 5ear 19 H. H.M.S.O . London. Nordmar.n, M.. and Sorge, A. (1941). Z. K'tbsforsch., 51, 163. Smith, \V. E. (1952). Ar^h. inJustr. Hvg., 5, 209. Srocks. P. 11952). Brit. J. Cancer, 6, 99. Vorvkjld, A. J.. and Karr, i. W, (1938). Amer. J.Path., 14, 49. Warren, S. (1943). Occup. Med., 5, 249. in <D CH uH T5 0) nj H 4J y: Tc3 0 fcl3 3 *01 -C in H 4HJ u CL 'J) U 0) u 0 2 in ou w a; yj cH u oj u u c 3 J e a0 ou U-l TU3 03 >1 4J u 03 -P P 0 >1 2 CQ