Document 63R4ORrqJk8L8nzaOGakJ0OR

Eur Retpir J, 1996, 9. 1032-1942 Printed in UK - all right* reserved Copyright CERS Journals Ltd 1996 European Respiratory Journal ISSN 0903 - 1936 REVIEW The epidemiology of mesothelioma in historical context J.C. McDonald, A.D. McDonald The epidemiology ofmesothelioma in historical context. J.C. McDonald, A.D. McDonald. OERS Journals Ltd 1996. ABSTRACT: Primary malignant mesothelial tumours were recognized by pathol ogists before asbestiform minerals (chrysotile, crocidolite and amosite) were mined commercially. The discovery, 40 yrs ago, of a causal link with crocidolite and the wide-ranging epidemiological studies which followed are the subject of this review. Early case-control and descriptive surveys, supplemented by cohort studies in insulation workers and chrysotile miners, quickly demonstrated major occupation al and geographical differences, with high risk in naval dockyard areas and in the heating trades. In the 1980s, reliable cohort surveys showed that in mining and in the manufacture of asbestos products the mesothelioma risk was much higher when exposure included crocidolite or amosite than chrysotile alone. However, qualita tive and quantitative information on exposure was too often inadequate for this evi dence to be conclusive. Well-controlled lung fibre analyses have reduced these deficiencies and demonstrated the probable implications of the greater biopersis tence of amphibole fibres. Chrysotile for industrial use often contains low concen trations of fibrous tremolite, which may well explain the few cases of mesothelioma associated with this type of asbestos. Progress in this field has been much retarded by controversy, for which the 20 year gap between the availability of reliable estimates of risk for the mining of chrysotile and that for crocidolite or amosite may have been largely responsible. Eur Respir J., 1996. 9, 1932-1942. Dept of Occupational and Environmental Medicine. National Heart &. Lung Institute. Imperial College, London, UK- Correspondence: J.C. McDonald Dept of Occupational and Environmental Medicine National Heart A Lung Institute Imperial College London UK Keywords: Amphiboles biopersistence chrysotile lung analyses mesothelioma tremolite Received: March 19 1995 Accepted after revision May 3 1996 A new industry, a new disease Asbestos, the "magic mineral", was known to the ancient Egyptians and mentioned by Pliny, but only mined and manufactured in any quantity since about the 1890s. At that time, exploitation of large deposits of white asbestos (chrysotile) in Canada and Russia and blue asbestos (cro cidolite) in South Africa began to develop rapidly. Early in this century, a brown asbestos, named amosite after the village of Amosa and the company which discovered it, together with deposits of chrysotile were also mined in South Africa. Crocidolite was later found and mined in Western Australia after the second world war and in lesser amounts elsewhere. Primary malignant pleural tumours have been recog nized at least since 1870, when Wagner [1] published a report on a case of "tubercle-like" lymphadenoma of the pleura. In 1943, Saccone and Coblenz [2], in a review of cases published before 1940, referred to the tumour as an "endothelioma", and mentioned a report by Lieutaud in 1767 of two such tumours in a series of 3,000 autop sies. They went on to say that, as knowledge of the tum our's existence spread, reports appeared with increasing frequency. These authors identified 41 cases in seven se ries published between 1910 and 1938 in a total of some 46,000 autopsies (0.9%), and added two cases of their own among 1,000 autopsies. The male to female ratio overall was 1.8, with 50% of subjects aged 40 yrs or more. Saccone and Coblenz [2] discussed the confusion which existed over pathology due, in their opinion, to the tumour's rarity and lack of characteristic histology. From descrip tions and photomicrographs of the 41 cases mentioned, they concluded that some at least were of other diseas es, such as bronchogenic carcinoma. In view of diverging opinions at that time on the origin of cells lining serous cavities from which the tumour originated, they suggest ed that the name "pleuroma" be used. Later, however^ the term "mesothelioma" became more generally accepted. The link with asbestos The suggestion that mesothelioma resulted from occu pational exposure to asbestos was first made by Gloyne [3], in Britain in 1935. In Germany, Wedler [4] described two cases in men with asbestosis and Weiss [5] one case in a naval dockyard worker. Leicher [6] was the first to report a case of peritoneal mesothelioma, in a textile spinner. At a scientific meeting in 1952, Cartier [7], then in charge of the industrial medical clinic at Thetford Mines, Quebec, Canada, reported eight cases of respira tory cancer, two of which he described as pleural tum ours. In his opinion, two such rare cancers in a small series of only eight cases suggested an occupational ori gin. A report in 1960 by Wagner et al. [8] of 33 cases of mesothelioma mainly from the crocidolite-mining area in the north west Cape 'Province of South Africa put the association beyond reasonable doubt. Of the 33 cases, 28 were in persons who had worked in the mines or liv ed close to them. The pathological material available was limited to thoracic contents, and peritoneal tumours were n` SCen' t.DEFENDANTS. EXHIBIT HWBUI0008022 epidemiology of mesothelioma 1933 Table 1. - Early case-control studies of mesothelioma giving definite or probable occupational exposure to asbestos First author Year [Ref.] Place Years diagnosed Cases/ controls Male Occupationally RR % exposed % Elmes Newhouse McEwen McDonald Rubino Ashcroft Hain ZlELHUIS 1965 1965 1970 1970 1972 1973 1974 1975 [10] i. Belfast. UK [11] London, UK [12] Scotland, UK [13] Canada [14] Piedmont, Italy [15] Tyneside, UK [16] Hamburg, Germany [17] The Netherlands 1950-1964 1917-1964 1950-1967 1960-1968 1960-1970 1948-1967 1958-1968 1969-1971 42/42 76/76 80/80 165/165 50/50 27/56 150/150 67/67 95 49 91 65 64 88 71 94 76 3.6 41 3.9 58 4.2 21 7.0 12 6.0 93 2.3 58 6.3 72 4.0 RR: relative risk; [Ref.): reference number. The 1964 New York Conference At a conference on the "Biological Effects of Asbestos", a study by the Selikoff group of New York and New Jersey members of the International Association of Insu lators and Asbestos Workers was a key event [9]. Of 632 males at work in 1942,255 had died by 1962, three from pleural mesothelioma, and by 1964, 10 of 307 deaths were from mesothelioma - four pleural and six peritoneal. Two important case-control studies were also presented at this conference and later published (table 1). In the first of these, Elmes et al. [10] studied 42 cases of me sothelioma and 42 controls in Belfast, UK, matched for ^^ddaatte, sex and age. Occupational histories were obtained ' m the living and from relatives of those who had died. Thirty six cases had a history of occupational exposure to asbestos, mainly in shipyards, compared with nine con trols. The second study was by Newhouse and Thompson [11] of 83 patients with mesothelioma diagnosed at a hospital close to the large Cape Asbestos factory in the east end of London, UK, which had opened in 1913. The control series comprised patients admitted later with other diseases, matched for sex and age. The authors acknow ledged that neither these nor the interview methods were ideal but concluded that the case-control comparisons of occupational and residential histories were probably val id. Of 76 pairs, 18 cases (24%) had been employed at the asbestos factory and eight (11%) as insulators or laggers, compared with one (1%) and four (5%) controls, respec tively. A further nine cases (12%) were in persons who had lived in the same house as an asbestos worker and were indirectly exposed, compared with one control (1 %). Only crocidolite was used in the factory until 1926, when small quantities of chiysotile and amosite were introdu ced. At the end of the 1964 conference, a Working Group under the auspices of the International Union against Cancer (UICC) reviewed the papers which had been presented and made a series of recommendations, with considerable emphasis on the need for more extensive epidemiological studies of mesothelioma and on the importance of asbestos fibre type [18]. As fibre of more than one type of asbestos was almost always used in m anufacture it was suggested that research should con- ntrate on countries where it was mined or milled as the problem might then be less serious. Initially, this ap proach was taken only in the extensive chrysotile mines and mills of Quebec, Canada, and in the smaller chryso tile industry of northern Italy. Starting in 1966, all 30,000 current and previous employees of the Quebec industry who had worked for at least one calendar month were listed. In addition to comprehensive studies of respira tory morbidity [19], mortality was investigated in a cohort of some 12,000 workers, born 1891-1920. Among 2,413 male deaths in the cohort to the end of 1966, there was a modest, exposure-related excess of lung cancer but only three deaths were from mesothelioma [20], This prompt ed the need for studies of mesothelioma on a wider geo graphical scale. Almost 20 years passed before comparable studies were made in the production of crocidolite or amosite in South Africa and Australia, with a resulting controversy, which will be discussed at the end of this review. Widening the epidemiological net Case-control studies From the end of 1967, all pathologists in Canada (over 400) were approached periodically concerning any fatal case of primary malignant mesothelial tumour diagnos ed at autopsy or biopsy and, in 1972, the survey was ex tended to all pathologists (almost 7,000) throughout the USA. On each occasion, a response was obtained from nearly all of them [21], The pathologists were then vis ited to discuss the diagnoses and to collect reports and material for panel review. A control with metastatic lung disease from a primary tumour outside the chest, matched for date, sex and age was selected from the same pathol ogy file as the case. Relatives were interviewed, gener ally by a public health nurse ignorant of the case/control status, and detailed residential and occupational histo ries recorded. Jobs were coded blind, using a list classi fied by four different expert groups according to the probability of asbestos exposure. Of 344 male cases of mesothelioma, 188 (55%) com pared with 78 (23%) controls fell into one of the five defined exposure groups, presented in table 2. Insulation work - an infrequent occupation in controls - showed the highest relative risk (46.1). Asbestos production and man ufacture was next in relative risk (6.1), almost wholly due to factory work. Employment in heating trades, ship yards and construction, after excluding insulation work, gave a lower combined risk (3.4). Occupational expo sure to asbestos was recorded in only two of 162 female cases and no control. In three female cases and one con trol, exposure had been in the home to the clothing of a chrysotile production worker and in five cases and one HWBUI0008023 1934 3.C. MCDONALD, A.D. MCDONALD Table 2. - Case-control study of 344 male primary malig nant mesothelial tumours of pleura and peritoneum, Canada 1960-1972, and USA 1972 [21] Occupational group with definite Cases Controls RR or probable asbestos exposure (n=344) (n=344) A Insulation B Asbestos production & manufacture (excl. A) Mining and milling Manufacture C Heating trades (excl A. B) D Shipyards (excl. A, B, Q E Construction (excl. A, B, C, D) F Other (excl. A, B, C, D, E) G None 27 25 4 21 70 21 45 55 101 1 46.1 7 6.1 2 5 27 4.4 13 2.8 30 2.6 90 1.0 176 1.0 RR: relative risk; excl.: excluding control to that of an employee engaged in insulation or manufacture. Neighbourhood exposure was investigated by comparing the number of cases and controls who had lived within 20 miles of a chrysotile mine, excluding those occupationally or domestically exposed. No case with meso-thelioma but two controls had lived within 20 miles of a chrysotile mine in Quebec, Canada, and one case and two controls within 20 miles of a mine in California, USA. Further evidence of an increased risk in shipyard areas was afforded by four of five other case-control studies published in the 1970s (table 1). Except for the report from northern Italy, where most of the occupational ex posures were in or near Turin, shipyard exposures were mainly to blame. Analyses of incidence and mortality In the Canadian surveys described above, the annual incidence for 1960-1966 was one case per million per sons - about 1.5 in males and 0.8 in females; however, there was probably under-reporting during these early years. In 1966-1972, the incidence in Canada was 2.9 per million males and 1.4 per million females; and in the USA in 1972, the corresponding rates were 2.7 and 0.8 per million. Cases in which a biopsy or autopsy speci men could be obtained were later reviewed by the Cana dian and American mesothelioma panels of pathologists on a probability scale, first on histology, and then with clinical but not occupational information. In Canada, the incidence in Quebec was higher than that in Ontario, but fewer Quebec cases were accepted by the panel than for Ontario, and the corrected incidence in the two provinces was similar. These estimates were used in 1975 in a geographical analysis of all known cases of mesothe lioma worldwide in areas where reported cases could be linked to population estimates. By applying age- and sexspecific rates found in Canada, the number of mesothe liomas expected on this basis was compared with the number observed. High ratios were found in many European shipyard cities, notably Walcheren, The Netherlands (23.3), Wilhelmshaven, Germany (21.5) and Plymouth, UK (14.3). In two locations with large asbestos manu facturing industries, there were also high ratios: Dresden, Germany (16.8) and the Manville-Somerville area of New Jersey, USA (26.5) [22], It was evident that, even in the early 1970s, mesothe lioma mortality in North America was already two or three times higher in males than females. This pattern be came apparent in most industrialized countries and was followed by a steady upward trend in male mortality, which still continues. The implications of the much lower annual increase in females will be mentioned in a later section as evidence on the question of nonoccupational asbestos exposure. The steep rise in males, which prob ably began in the 1940s, is well explained as reflecting a parallel increase in the industrial use of asbestos, from about 1910, having taken account of a 30-40 year latency [23]. As a result of this increase, mesothelioma is curren tly responsible for some 20 deaths per million male pop ulation in Western Europe and North America compared with an estimated 1 to 2, 30-40 yrs ago. In early stud ies, only a minority of male cases were attributable to occupational exposure to asbestos, whereas, depending on location, up to 90% now are. As the epidemic has evolved in the UK, the overall distribution of occupations which have been responsible has remained much the same but there is evidence that the contribution of work in shipyards has fallen but has increased in construction [24], How long the mesothe lioma epidemic can be expected to last is an important but difficult question. Estimated crudely from the indus trial use of asbestos, it seemed possible that the peak might be reached in about year 2000; falling thereafter over a period of about 40 yrs [23]. Results of a recent and more sophisticated analysis of UK data are less op timistic and suggest that the peak may not be reached until 2010-2020 [25]. If true, this would point to the greatly increased importation of amosite for construc tion use in the UK during the 1960s and 1970s. However, these projections are highly dependent on the extent to which mesothelioma is better diagnosed and ascertain ed now than it used to be. Other causal agents [ Xj Some 20 yrs after the pioneering work of Wagner et al. [8] in South Africa, a discovery in some ways ipore dramatic was made by Baris et al. [26] in Turkey, In some small villages in a circumscribed volcanic area of central Cappadocia, an extraordinarily high mortality from mesothelioma was observed in the local popula tion. In Karain, a village of 575 inhabitants, in a 5 year period, 1970-1974,24 from a total of 55 deaths were attri buted to pleural mesothelioma. In Tuskoy, a larger village, both pleural and peritoneal tumours were in considera ble excess. The rocks and soil of the affected area are rich in zeolite minerals, a common constituent of vol canic tuff. Erionite, a specific zeolite, was found in fibrous form in soil and rock samples and in respirable concen trations in both villages. Experimentally, these fibres are highly carcinogenic and readily produce mesothelial tum ours in rats and mice [^7], The epidemiological investiga tions, although mainly descriptive rather than analytical, leave little doubt that these fibres are the cause of this unparalleled incidence of tumours in man [27], Deposits of fibrous erionite are common in many areas of the world but to date there is no certain evidence, other than in Turkey, of an association with mesothelioma. HWBUI0008024 epidemiology of mesothelioma 1935 In a report from India, five cases of malignant mesothe lioma were reported in sugar cane workers, with the sug gestion that organic fibres might have been the cause [28], Some support was obtained from Louisiana, USA, where two cases of mesothelioma were observed in simi lar circumstances [29]. In an electron microscopic study of ashed sugar cane leaf, silicon was found deposited along the hypoderm resulting in fibres 0.85 pm in diam eter and 10-100 pm in length. The possibility that such "biogenic silica" fibres might be causally related to the tumour awaits further epidemiological study. That mesothelioma might occur in workers exposed to man-made mineral fibres (MMMFs) was investigated in three large cohorts totalling some 41,185 workers employed in their manufacture in Europe and North America. Only four deaths were ascribed to the disease in a total of 7,862 deaths from all causes, one in a man also exposed to amosite [30]. Exposure levels in MMMF manufacture are extremely low, however. Systematic cohort mortality studies l The 1980s saw the publication of results from a con siderable number of cohort studies in specific occupa tional groups, which differed in industrial process and in type of asbestos fibre to which the workers were exposed. The main and most recent findings from the more impor tant of these investigations, classified by predominant industry, are summarized in three tables: mining and milling in table 3; manufacture of asbestos/cement prod ucts, asbestos textiles and asbestos friction products in table 4; and miscellaneous industrial groups in table 5. In a few of these studies, attempts were made to esti mate exposure in terms of intensity and duration but in none could the level of exposure to individual fibre types be reliably assessed. Mining and milling The large cohort of chrysotile miners and millers in Quebec, Canada, first reported in 1971 has been followed ever since with updated results published periodically, the most recent in 1993 [31]. By this time, almost 80% of the cohort had died and the youngest survivors were in their mid 70s. From a total of over 8,000 deaths, 38 were probably due to mesothelioma - all pleural with one possible exception - a proportional mortality of just under 0.5%. A cohort of approximately one tenth the size from Balangero in northern Italy gave similar results. No comparable data on amphibole miners were available Table 3. - Cohort mortality studies of male asbestos miners and millers* kFirst uthor Year [Ref.] Country Subjects n All causes n% Deaths Lung cancer n SMR Predominant Mesothelioma fibre n PMR/1000 McDonald 1993 [31] Canada PlOLATTO 1990 [32] Italy Sluts-Cremer 1992 [33] RSA (1) (2) Armstrong 1988 [34] Australia McDonald 1986 [35] USA 10918 952 3212 3430 6505 406 7312 427 648 423 820 165 67 45 20 12 13 41 545 1.33 22 1.11 26 1.38 27 2.03 91 2.64 21 2.45 33 4.5 2 4.7 4 6.2 20 47.3 32 39.0 4 24.2 Chrysotile Chrysotile Amosite Crocidolite Crocidolite Tremolite *: including vermiculite miners exposed to fibrous tremolite. [Ref.]: reference number; SMR: standardized mortality rate; PMR: proportional mortality rate. Table 4. - Cohort mortality studies of male workers in asbestos manufacturing industries First author Year [Ref.] Country Subjects n All causes n% Deaths Lung cancer n SMR Amphibole Mesothelioma content n PMR/1000 Asbestos-cement Thomas 1982 [36] UK 1592 Ohlson 1985 [37] Sweden 1176 Gardner 1986 [38] UK 1510 Hughes 1986 [39] USA (plant 1) 2565 FtNKELSTEtN 1984 [40] Canada 535 Ajues-Patin 1985 [41] France 1506 Hughes 1986 [39] USA (plant 2) 4366 Magnani 1987 [42] Italy 2608 Raffn 1989 [43] Denmark 7996 ALBtN 1990 [44] Sweden 1929 Neuberger 1990 [45] Austria 2816 Textiles McDonald 1983 [46] USA 2543 j>MENT 1994 [47] USA 1247 McDonald 1983 [48] USA 4137 Teto 1985 [49] UK 3211 Friction products McDonald 1984 [50] USA 3641 Newhouse 1989 [51] UK 9104 351 22 220 19 384 25 477 19 108 20 206 14 874 20 728 28 1305 16 592 31 540 19 570 22 607 49 895 22 727 23 803 22 2055 23 24 0.93 11 1.23 35 0.92 48 1.17 26 4.80 9 1.63 107 1.44 110 2.68 162 1.80 35 2.50 50 1.72 59 2.0 72 2.25 53 1.05 93 1.44 73 1.49 229 1.03 2 0 1 3.5 2 19 4 8 28 20.7 13 13 5, 1' 2 14 J 2.5 10 14.8 00 11 5.4 Minimal Substantial Minimal Substantial Minimal Localized For definitions see legend to table 3. HWBUI0008025 1936 J.C. MCDONALD, A.D. MCDONALD Table 5. - Cohort mortality studies of male and female asbestos-exposed workers in miscellaneous industries Fust author Year [Ref.] Country Subjects n All causes n% Deaths Lung cancer n SMR Mesothelioma n PMR/1000 Predominant fibre Insulation work Selkoff 1979 [52] USA & Canada Insulation products manufacture Shdman 1979 [53] USA Acheson 1984 [54] UK Filter assembly McDonald 1978 [55] Pntiarfa Jones 1980 [56] UK Talcott 1989 [57] USA Acheson 1982 [58] UK (a) UK (b) Dockyard work Rossiter 1980 [59] UK 17,800 unclear 4280 199 1088 33 757 570 6292 2271 13 528 333 8 56 28 166 15 28 85 219 29 177 31 1043 17 *: also exposed to crocidolite. For definitions see legend to table 3. 397 4.24 76 5.78 38 1.31 8 2.0 12 2.14 8 15.71 15 2.41 7 1.45 84 0.84 175 77.1 14 26.5 5 15.0 0 160.7 29 174.7 5 178.6 5 22.8 l1> 5.6 31 29.7 Chrysotile & amosite Amosite Amosite Crocidolite Crocidolite Crocidolite Crocidolite Chrysotile Not stated* until the late 1980s, when findings for crocidolite and amosite were published from South Africa and for croci dolite from Australia. At about the same time, two small cohorts of vermiculite miners in the USA, one of which had experienced substantial exposure to fibrous tremolite provided important evidence on its effects [35]. In interpreting the figures on proportional mortality in table 3, it should be noted that the crocidolite and amosite rates were derived from cohorts of which only 12-21% had died. At that stage in the evolution of the chrysotile cohorts, the proportional mortality from mesothelioma was less than 2 per 1,000. Proportional mortality rates are a very crude indicator of risk, since they do not take levels of exposure or competing causes of death into account. Nevertheless, the data in table 3 leave little doubt that crocidolite carries a much higher risk than chrysotile, with amosite and tremolite probably some where in between. Manufacturing processes The 17 cohorts shown in table 4 were informative in that all were engaged in the manufacture of asbestos-con taining products, predominantly from chrysotile but some with the planned inclusion of relatively small quantities of crocidolite or amosite. Interpretation of the results is facilitated by the fact that the proportions dead at time of analysis in each cohort were of the same order, all but one (49%) ranging 14-31% (median 22%). With each of the three types of product, there was a fairly consistent difference in the proportional rates for mortality from mesothelioma between cohorts with and without exposure to amphiboles. In cement workers, the relevant rates per 1,000 were 20,7 and 3.5, respectively; in textile workers 14.8 and 2.5; and in friction product workers 5.4 and 0. To these admittedly crude compari sons, made without reference to intensity or duration of exposure, some limited detail can be added. For exam ple, the only death in the cohort of Gardner ex al. [38] was in a man first employed in the factory less than 7 years earlier. In the friction products cohort of Newhouse and Sullivan [51], all but one of the 11 cases observed were from a small group of employees who worked for a short time on a special crocidolite contract, and there was no case attributable only to chrysotile in the remain-' der of the cohort. Miscellaneous occupations Although all but one of the cohorts shown in table 5 were exposed wholly or largely to amosite of crocido lite, the data are of considerable interest. The very large cohort of American insulation workers investigated by Seukoff et al [52] is one of die classic studies of asbestos epidemiology. At a stage when only 13% had died, al most 8% of deaths were attributed to mesothelioma, two thirds of which were peritoneal. Initially, it was thought that the exposure of these workers was mainly to chrysotile but it soon became dear that at least from 1930 onwards, amosite was the predominant type of asbestos used in American insulation materials. The suggestion was made at the Johannesburg Conference in 1977 that amosite rather than chrysotile might be responsible for the prob lem [60], but this has only recently beqn confirmed by lung butden analyses [61]. Two cohorts of workers expos ed only to amosite in the manufacture of insulation mate rials, one in the USA [54] and the other in the UK ^53], have provided further evidence of the capacity of this ritineral fibre to cause mesothelioma. However, at face value, their levels of risk did not reach that of insulation work ers employed in applying and removing these materials. The highest recorded proportional mortality rates from mesothelioma were observed in two remarkable cohorts of males and females who worked for quite short periods of time during the early years of the second world war in England and in Canada on the installation of filter pads in military gas-masks. The pads were made from pure crocidolite believed to have come from the Wittenoom mine in Australia, where a cohort of miners and millers was studied some years later (table 3). An identical pat tern in the incidence of cases began both in England and Canada 18 yrs after first exposure, and continued to grow for at least 10 more years; in the UK, new cases are still being seen almost 60 yrs later (J.S.P. Jones, personal communication). The nearest equivalent to these disas trous events was seen in a small group of employees in the manufacture of filters for cigarettes in the USA from, HWBUI0008026 epidemiology of mesothelioma 1937 of all things, crocidolite [57]. Five cases of mesotheli oma from a total of 33 deaths (15%) were reported in that group. Finally, a study by Rossiter and Coles [59] of over 6,000 males employed in the Devonport (Plymouth, UK) naval dockyard is worth noting. Asbestos exposures were not well-documented but certainly included an appreci able proportion of crocidolite. This was not sufficient to produce excess mortality from lung cancer but it result ed in a large number of cases of mesothelioma. Overview Few, if any, environmental hazards have been the object of so many cohort studies as asbestos, only some of which have been mentioned in this review. Although it would be difficult to conclude from these data that the capacity of amphibole fibres, especially crocidolite, to produce mesothelioma is not considerably greater than that of chrysotile, uncertainties remain. The problem lies in the evident complexity of the causal relationships and the many factors other than fibre type with which the crude information available on exposure in these cohorts has failed to deal. Whilst duration of exposure can be estimated fairly accurately, fibre concentrations by size And type cannot This is a serious problem when expo sures are known to have been to more than one type of asbestos, and of increasing importance where one type is sometimes contaminated by another as, for example, chrysotile by tremolite or amosite by crocidolite. There is also reason to suspect that the nature of the industrial process may affect the outcome, as the textile industry does for lung cancer but not apparently for mesothelioma It is, therefore, hardly surprising that, with the exception of the studies by Hughes and Weill [39] in the asbestos/ cement industry, there still exist virtually no reliable data on exposure-response for mesothelioma which take acc ount of any of these variables. To avoid the problems of environmental exposure assessment, more recent epide miological research has turned increasingly to the use of lung fibre burden measurements. Hus approach and im portant new insights into the disease potential of chryso tile are discussed in a later section. Nonoccupational mesothelioma It is clear enough that in most cases of mesothelioma, particularly in industrialized countries, there is a fairly definite history of direct or indirect exposure to asbestos at work. This has led many to suggest that all cases may be caused in this way. This hypothesis raises two sepa rate questions: firstly, is there evidence of a background incidence of the disease before, and presumably also since, asbestos was first exploited for industrial use? and, sec ondly, are there some cases attributable to asbestos but resulting from exposure in the domestic, neighbourhood or general environment rather than at work? These two aspects will be considered in turn. The background hypothesis In a recent review of this question [62], five types of evidence were examined, all of which pointed to the probable occurrence of mesothelioma as a rare malig nancy unrelated to asbestos exposure both before and since the industrial use of these fibrous minerals began at the end of the last century. The simplest evidence that cases are not all due to asbestos fibres is the high mor tality from the disease in certain villages of central Turkey, to which reference has already been made. These cases were clearly caused by local deposits of fibrous erionite, a mineral with many physical and biological properties similar to crocidolite and tremolite. Deposits of fibrous erionite are widespread on the earth's surface and there is, indeed, some suggestion of related cases in proximi ty to deposits in the Rocky Mountain states of North America [63] but not elsewhere. Other suggestive evidence, historical rather than geo graphic, has also been mentioned. It is fairly clear that primary malignant tumours of the pleura were recog nized by pathologists at autopsy at the end of the last century, before the industrial use of asbestos could have been responsible and in the absence of any link with occupation. Although these cases were somewhat more common in males than females, their distribution sug gests either a genetic aetiology or an environmental fac tor common to both sexes. The latter could conceivably include waterborne or airborne fibres originating from a wide range of naturally occurring minerals. Hie occur rence of mesothelioma in childhood leads to similar con clusions. Evidence of such cases was found in three surveys, the largest of which by Fraire et ai [64] record ed 80 cases in childhood, in only two of which was ex posure to asbestos at all likely. Unless the usual latency with asbestos-related cases is much shorter in childhood, the existence of these cases suggests that there must have been some other cause. Less conclusive support for a background incidence is afforded by the results of lung burden analyses in three case-control studies in North America and the UK. which are described more fully in the next section. In all three, a proportion of cases could not be attributed either to amphibole or chrysotile fibres. However, the small num ber of cases examined and the limits of fibre detection by electron microscopy reduce the confidence that can be put on negative findings. More robust epidemiological evidence is given by mor tality statistics for mesothelioma over the past 50 yrs or so. In most industrialized countries, the disease has incr eased much more rapidly in males than females, reflect ing the impact of occupational asbestos exposure 30-40 yrs earlier. Backward extrapolation of these trends sug gests that, before the diverging pattern began, mortality was about 1-2 per million population in both sexes. This conclusion is supported by data from countries or regi ons with low mesothelioma mortality, where both male and female rates are at about this level and by data for California, USA, after exclusion of occupationally re lated cases [65]. If indeed there always has been a low background inci dence of malignant mesotheUai tumours unrelated to the industrial exploitation of asbestos, there remains the ques tion of its aetiology. There are no indications that asbestos as it occurs naturally has been responsible, although this possibility would be difficult to refute. Similarly, although there are widespread deposits of other fibrous minerals, such as erionite and tremolite, which could conceivably % i t a i 5 i $ , HWBUI0008027 1938 J.C. MCDONALD, A.D. MCDONALD cause sporadic disease, there is very little evidence that they do. In special circumstances, however, where local rock has been used for domestic whitewash, as in New Caledonia and several Mediterranean countries [6668], exposure to fibrous tremolite from childhood may well have led to a substantial number of cases. A num ber of other agents, such as biogenic silica fibre from burned-off sugar cane, some heavy metals (especially beryllium) and ionizing radiation have also been cited but none confirmed. Neighbourhood and domestic exposure The question thus comes down to whether the indus trial use of asbestos - amphibole fibres in particular can lead to sufficient environmental pollution outside the workplace to cause nonoccupational cases. It is fairly clear that such cases do occur, but rarely. The evidence is strongest in cases presumably resulting from domes tic exposure among the household contacts of asbestos workers. Such cases were documented in several papers mentioned by Gardner and Saracci [69] in a recent review, in most of which crocidolite was probably res ponsible. The occurrence of neighbourhood cases has also been described in the immediate vicinity of croci dolite mines in South Africa and Australia [69], and of factories which used these fibres in London, UK, and Hamburg, Germany, [11,16]. Most other studies of neigh bourhood exposure have been negative. More difficult to investigate is the possibility that cases of mesothelioma may arise as a result of general urban air pollution with asbestos fibres - almost all chrysotile and usually very short - in industrial cities of North America and Western Europe. There is little or no direct evidence one way or another for this hypothetical risk. That it is probably small and almost certainly beyond the limits of detection is suggested both by exposureresponse findings in chrysotile miners and millers and by the fact that mortality in females in North America and Western Europe has shown little or no increase dur ing the last 20-30 yis. The latter holds despite the contri bution of occupational and household exposure and the greater awareness by physicians and pathologists of me sothelioma during the same period [25]. Lung burden studies Since the pioneering work of Lancer and Pooley [70] in the early 1970s, the use of analytical transmission elec tron microscopy to identify and quantify mineral fibres in lung tissue has introduced a technique of great poten tial value for epidemiological research. In the assessment of past exposure, its specificity and cumulative capaci ty are far greater than was possible from work histories and scanty environmental measurements. On the other hand, this approach is limited in practice by the highly se lected availability of lung tissue, and by the varied pen etration, persistence and distribution of mineral fibres in the respiratory tract. Also, the significance of fibres in lung tissue at death is affected by whether or not the dis ease mechanism is related to these same fibre qualities. It was shown by Pooley [71] and Rowlands etal. [72] that there were substantial differences in the ability of chrysotile and amphibole fibres to penetrate and persist in lung tissue. Thus, only studies which are idequately controlled for time variables and in other important respects, and where the analyses are conducted blind and in parallel, provide reliable information. The six surveys listed in table 6 are those in which these requirements were largely met, but even so with results which cannot readily be compared or tabulated. Whilst it is evident from all six studies that amphibole fibres over 8 pm in length could have explained nearly all the cases and shorter fibres few if any, there are seve ral arguments for and against the possibility that chryso tile may also have played some part. Certainly, the power of the multivariate analysis used by McDonald et al. [77] was not sufficient to negate findings from their uni variate analysis or to imply incompatibility with the ob servation by Rogers et al. [78] of a chrysotile excess in the few cases where amphiboles were absent In the lat ter study, the selection of all controls from a single hos pital in Sydney, Australia, over a short period of time and the lack of matched pairs threw some doubt on the comparability of the 25 cases and 31 controls without amphiboles. More substantial questions have been raised over the validity of conclusions on the role of chrysotile obtained from lung analyses at autopsy. Some have argued that only analyses of pleural tissue would have any relevance, perhaps without appreciating that the lung is simply be ing used in epidemiology as a sampling device to reflect past airborne exposure and not a direct pathological eff ect The fact remains that as chrysotile is of low persist ence, concentrations found at death may be unrelated to what was inhaled over a lifetime and, indeed, may predo minantly reflect only very recent exposure. These ques tions have been reviewed by McDonald [79], and more fully by Churg [80]. In essence, the validity of lung bur den measurements in epidemiology depends on: 1) how well they correlate with best estimates of past exposure; and 2) whether the investigations were controlled [for time-related and other potentially confounding factors. For neither of these questions is the answer entirely clbar. There is, indeed, evidence from a fairly small study by Rowlands et al. [72] and a larger one by Sbastien et al. [81] of a reasonably good correlation with past eiivironmental exposure both for chrysotile and for tremolite in miners and millers in Quebec, Canada, and also in American textile workers, but further confirmation of this kind is needed. On the more difficult matter of ade quate control for confounders, even among the six stud ies shown in table 6, not one was wholly adequate in this regard. It would seem unreasonable, even so, to rej ect entirely the consistent evidence which they present The association of fibrous tremolite with many chry sotile deposits, leading to low but varying contamination of the final product, was first identified by lung burden analysis in 1976 [71]. This discovery had two possible and conflicting implications: firstly, that these amphibole fibres might be disproportionately responsible for dis ease apparently due to chrysotile - particularly mesothe lioma - but, secondly, that as a much more persistent fibre, it might simply serve as a marker of chrysotile no longer present. Before dismissing chrysotile as a cause of mesothelioma, therefore, note must be taken of the extent to which tremolite is found in lungs at autopsy in HWBUI0008028  EPIDEMIOLOGY OF MESOTHELIOMA 1939 Table 6. - Analysis of mineral fibres in lung tissue from mesothelioma cases and controls First author Year [Ref.] Country Cases Controls Odds ratio for Evidence on chrysotile amphibole fibres* Jones 1980 [73] UK i 86 cases notified by coroners & pathologists, 1976 McDonald 1982 [74] USA & 99 cases from Canada survey of pathologists Mowe 1985 [75] Norway 14 cases, county cancer registry, 1970-1979 Gaudichet 1988 [76] France 20 cases from Nantes district, 1980-1982 i McDonald 1989 \ i 1 1 ROGERS 1991 [77] [78] m________ Canada Australia 78 cases from survey of pathologists, 1980-1984 221 cases from national surveillance, 1980-1985 56 cases (lung cancer 27, cerebrovascular disease 29); matched for age, sex & place Secondary lung cancer, matched for age, sex, date and hospital 28 cases excluding malignant and chronic pulmonary disease; matched for age, sex, year and residence 20 each of adenocarcinoma and squamous carcinoma, secondary lung cancer and cardiovascular disease; matched for age, sex and hospital ' NonmaUgnant, nonrespiratory disease; matched for age, sex, date, hospital and type of sample 359 tissue samples from a hospital in Sydney excluding nonmalignant respiratory disease and abdominal cancer, unmatched 7.4 Chrysotile present in 2 of 4 cases without amphiboles 3.8 In pairs where amphibole content was <106 fibres-g-1 closely similar distributions of chrysotile 8.5 Fibre type not identified (based on all types of amphibole fibres) amphibole fibre Similar concentration in concentration cases and controls "2-3 times higher than in controls 6.6 for fibres ^8 pm in length 16.6 for fibres >10 pm in length Low level risk in univariate analysis and none in multivariate analysis 7 of 25 cases and 3 of 31 controls without amphibole fibres had ^lO5 fibres-g-1 chrysotile *: calculated by Rogers et al. [78], from comparison of cases and controls above and below 106 amphibole fibres-g-1. this disease. Light is thrown on this question by sever al studies, of which that by McDonald et al. [77], being well-controlled, indicated that of 78 cases of mesothe lioma from across Canada, perhaps 29% were due to tremolite. An investigation in the USA by Roggu et al. [82], although without controls and therefore less inter pretable in terms of cause and effect, identified tremo lite fibres in unstated concentration in 55% of 94 cases. Thus, some of the mesothelioma cases apparently attrib utable to tremolite may have resulted from low-level con tamination of industrial chrysotile, for which tremolite is not only a marker but, as shown below, may indeed be the cause. It must also be remembered that fibrous tremo lite is a common contaminant of several industrially ex ploited minerals other than chrysotile. The tremolite question The extensive programme of epidemiological research in the mines and mills of Quebec, Canada, which began 30 yrs ago, was undertaken in the belief, wrong as it tur ned out, that the results would reflect exposure to pure chrysotile. Any impurities in the ore body were conside red to be of minor importance and unlikely to cause dif ficulties in interpretation. The geological data available tended to support this view. The first indication that the situation might be more complicated was revealed by the 'results of electron microscope analyses of lung tissue from ex-mine workers reported by Pooley in 1976 [71] and Rowlands etal. [72] in 1982. Not only did this show that chrysotile fibres were not alone but that amphibole fibres in the tremolite series were also present and usually in even higher concentration. In part, this reflected a lesser ability of chrysotile to penetrate the smaller airways but the major factor was undoubtedly the greater durability of the amphibole fibres. Whatever the explanation, the pos sible importance of persistence in pathogenesis was im mediately evident. There are various reasons for taking this finding se riously in relation to mesothelioma. The most direct in dication of the carcinogenic potency of mineral fibres in the tremolite series was seen in the small cohort of 406 Montana, USA, vermiculite miners and millers, results of which were published in 1986 (table 3). Among only 165 deaths, 21 were from lung cancer (standardized mor tality rate (SMR) 2.45) and four from mesothelioma (proportional mortality rate (PMR) 24.2 per 1,000). At a comparable stage in the evolution of the chrysotile co hort in Quebec. Canada, among some 4,000 deaths, 10 were from mesothelioma (PMR 2.5 per 1,000). The aver age cumulative exposures experienced by the two coh orts were 145 fibresmL'-yr1 in Montana and over 1,000 fibres-mL-'-yr1 in Quebec. A very rough calculation would, therefore, suggest that if about 1% of the Quebec expo sure was to tremolite, this amphibole fibre might expl ain the mesothelioma risk at both locations. At that time, however, there seemed to be no way in which this high ly speculative hypothesis could be tested, let alone vali dated. Further support for the general concept came to light a few years later in the case-control study based on lung tissue analyses of 78 mesothelioma deaths in Canada, 1980-1984 and matched referents [77], Concentrations of amosite, crocidolite and tremolite, but not of chryso tile or any other identified mineral fibre, discriminated sharply between the two series. The attributable risk * HWBUI0008029 1940 J.C. MCDONALD, A.D. MCDONALD associated with tremolite was estimated by multivariate analysis, suggesting that perhaps 23 of the 78 cases were due to tremolite, including all nine in the Quebec min ing region. After excluding these cases, there remained 14 of 69 (i.e. about 20%) in the rest of Canada attrib utable to tremolite. Given that in the past crocidolite and amositc comprised less than 10% of all asbestos used commercially, it thus appeared possible that tremolite as a contaminant of chrysotile might explain most of the remaining cases. A better opportunity to examine the hypothesis came a few years later when mortality in the Quebec cohort was updated [31]. By the end of 1988, 33 fatal cases of mesothelioma were identified in a total of 7,300 deaths from all causes (PMR 0.45 per 1,000) and by the end of 1992, 38 cases in a total of over 8,000 deaths (PMR 0.47 per 1,000). Among the 33 cases ascertained before 1989, 20 were from Thetford Mines, eight from the town of Asbestos, and five from a small asbestos products fac tory in Asbestos where crocidolite had also been used. Statistical analysis showed that the mesothelioma inci dence was about 2.5 times greater among miners and millers at Thetford Mines than Asbestos [83]. Data on lung tissue analyses by Sdbastien from our earlier sur vey [81], although scanty, showed that tremolite fibre concentrations were 2-4 times greater at Thetford Mines than Asbestos (ratio of medians 2.4; ratio of means 3.3) [83]. More demanding analyses are currently being applied to the most recent cohort data, focusing entirely on the 24 mesothelioma cases from Thetford Mines, 22 of which were in males employed by the largest company in that area. Advantage is being taken of two important facts: firstly, that this company had originally comprised many smaller companies distributed over a fairly wide geo graphical area; and, secondly, that lungs from 83 former cohort members from the same company had been analy sed by electron microscopy for another purpose some 5 yrs earlier [81]. A preliminary analysis, now published [85] has demonstrated a remarkably close correlation be tween the high incidence of mesothelioma in a localized area of five mines and the concentration of tremolite fibre in the lungs of men who had worked in them [77], These findings suggest that the relatively low risk of mesotheli oma associated with chrysotile mining and milling may be largely determined by tremolite fibre contamination. Fur ther analyses, which are now in progress, are needed to test this hypothesis and its wide implications. Conclusion After nearly 40 yrs of epidemiological research, the main determinants of malignant mesothelial tumours in man are reasonably clear. It could be said that this was apparent from the start, when Wagner et al. [8] described a disease of long latency in workers who mined, milled and used crocidolite, and in their family contacts, but rarely in miners of other types of asbestos. Sadly, the picture soon became confused and our understanding clouded by controversy. It is worth considering the rea sons for this. A major factor was that, although it was evident from cohort studies of chrysotile miners and millers in the early 1970s that this type of fibre rarely caused mesothe lioma, it was nearly 20 yrs before comparable informati on became available for miners and millers of crocidolite or amosite. It is understandable that investigators famil iar with the disastrous experience of insulation workers in North America, thought to have been exposed to chrysotile and possibly amosite, found it difficult to believe that all types of asbestos were not equally harm ful. Their regrettable conclusion was that the data from Quebec, Canada, were wrong, or worse - a view sup ported by laboratory experiments which showed that all fibre types were equally carcinogenic for rats. Against a background of suspicion and recrimination, the results of the several important cohort studies publi shed in the 1980s failed to have much effect on entrenched and conflicting views. For those who saw chrysotile as a mineral fibre of low carcinogenicity, the findings sum marized in tables 3, 4 and 5 confirmed this opinion. For those of the other persuasion, no great difficulty was found in maintaining their disbelief: the uncertainties associated with mixed exposures, lack of information on exposure intensity, and statistical chance were often cited, but less flattering reasons were not far below the sur face. The aphorism that "what a man would like to be true, that he more readily believes" probably applied to both sides. Some resolution of this unhelpful controversy came with the use of lung tissue analyses in epidemiological re search. Despite difficulties in interpretation of results and the absolute need for properly selected controls [79], these studies have shown two things and are on the way to tes ting a third. Firstly, the clear evidence of an overwhelming predominance, with dose-response, of amphibole fibres in mesothelioma cases. Secondly, that amphibole fibres per sist in lung tissue, whereas chrysotile does not; thus, the short lifespan of laboratory animals could not deal ade quately with tumours with a characteristic latency of 30-40 yrs in men. Thirdly, the fact that it has been only by analysing lung tissue that the varying presence of fib rous tremolite has been demonstrated in chrysotile {as produced commercially, and the growing probability that this previously unrecognized amphibole may be respon sible for most cases of mesothelioma associated with heavy chrysotile exposures. \ When peace finally returns to asbestos epidemiology, it may well be too late to restore chrysotile's soiled repu tation. Like Caesar's wife, such materials must be beyond suspicion, and neither political nor administrative poli cies are easily reversed. The more important lessons from this saga probably relate to the screening and manufac ture of asbestos substitutes - the man-made mineral and organic fibres. Given what we now know, it would sure ly be foolhardy, without extraordinary justification, to al low the widespread use of fibres which resemble crocidolite and tremolite physically and the amphiboles generally in their biological persistence. ' References 1. Wagner E. Das tuberkeiahnliche Lymphadenom. Arch Heilk 1870; 11: 495-525. 2. Saccone A, Coblenz A. Endothelioma of the pleura. Am J Clin Pathol 1943; 13: 188-207. HWBUI0008030 EPIDEMIOLOGY OF MESOTHELIOMA 1941 3. Gloyne SR. Two cases of squamous carcinoma of the lung occurring in asbestosis. Tubercle 1935; 17: 5--10. 4. Wedler HW. Obcr den Lungenkrtbs bei Asbestose. Dtsch Arch Klin Med 1943; 191: 189-209. 5. Weiss A, Pleurakrebs bei Lpngenasbestose in vivo morphologisch gesicbert Medizinische 1953; 1: 93-94. 6. Leicher F. Priroirer Deckzelltumor des Bauchfells bei Asbestose. Arch Gewerbepath Gewerbehyg 1954; 13:382-393. 7. Carder P. Abstract of discussion. Arch Indust Hyg Med 1952; 5: 262. 8. Wagner JC, Sleggs CA, Marchar.d P. Diffuse pleural mesothelioma and asbestos exposure in the North Western Cape Province, hr J Ind Med 1960; 17: 260-271. 9. Hammond , Selikoff U, Churg i. Neoplasia among insulation workers in the United States with special ref erence to intra-abdominal neoplasia. Arm NY Acad Sci 1965; 132: 519-525. Elmes PC, McCaughey WTE, Wade OL. Diffuse meso thelioma of the pleura and asbestos. Br Med J 1965; 1: 350-353. Newhouse ML. Thompson H. Mesothelioma of pleura and peritoneum following exposure to asbestos in the London area. Br J Ind Med 1965; 22: 261-269. 12. McEwen J, Finlayson A, Mair A, Gibson AAM. Meso thelioma in Scotland. Br Med J 1970; 4: 575-578. 11 McDonald AD. Harper A. El Attar OA. McDonald JC. Epidemiology of primary malignant mesothelial tumours in Canada. Cancer 1970; 26: 914-919. 14. Rubino GF, Scansetti G, Donna A, Paiestro G. Epidemi ology of pieurai mesothelioma in north-western Italy (Piedmont). Br J Ind Med 1972; 29: 436-442. 15. Ashcroft T. Epidemiological and quantitative relation ships between mesothelioma and asbestos on Tyneside. J Clin Pathol 1973; 26: 832-840. 16. Hain E. Daiquen P, BohJig H, Dabbert A, Hinz 1. Retro spective study of 150 cases of mesothelioma in Hamburg area. Ini Arch Arbeitsmed 1974; 33: 15-37. 37, Zielhuis RL. Verstecg JPJ, Planteydl HT. Pleural meso thelioma and exposure to asbestos: a retrospective casecontrol study in The Netherlands. IntArck Occup Environ Health 1975: 36: 1-18. 18. U1CC. Report and recommendations of the working group on asbestos and cancer. Ann NY Acad Sci 1965; 132. 706-721. 19. McDonald JC. Bccklake MR, Gibbs GW, McDonald AD. Rossiter CE. The health of chrysotile mine and mill workers of Quebec. Arch Environ Health 1974: 28: 63-68. 20. McDonald JC, McDonald AD. Gibbs GW. Siemiatycki J, Rossiter CE. Mortality in the chrysotile asbestos mines and mills of Quebec. Arch Environ Health 1971; 22: 677-686. 21. McDonald AD. McDonald JC. Malignant mesothelioma in North America. Cancer 1980, 46: 1650-1656. 22. McDonald JC. McDonald AD. Epidemiology of mesothe lioma from estimated incidence. Prev Med 1977; 6: 426-446. 23. McDonald JC. Health implications of environmental ex posure to asbestos. Environ Health Perspect 1985; 62: 319-328. * Ross DJ. Sallie BA. McDonald JC. SWORD V4: sui- veillance of work-related and occupational respiratory disease in the UK. Occup Med 1995: 45: 175-178. 25. Peto J. Hodgson JT. Matthews FE, Jones JR. Continuing increase in mesothelioma mortality in Britain. Lancet 1995; 345: 535-539. 26. Baris Yl. Artvinii M. Sahin LAA. Environmental mesothe lioma in Turkey Ann NY Acad Sci 1979,330: 423-432. 27. IARC monographs on the evaluation of carcico&enic risks to humans. Vol. 42. Silica and Some Silicates. Lyon, France, IARC. 1987; pp. 225-239. 28. Das PB. Fletcher AG, Deodhare SG. Mesothelioma in an agricultural community of India: a clinicopathological study. Aust NZ J Surg 1976; 46: 218-226. 29. Rothschild H, Mulvey JJ. An increased risk for lung canes' mortality associated with sugar cane farming. / Natl Cancer Inst 1982; 68: 755-760. 30. McDonald JC. Epidemiology of pleural cancer. In: Hirsch A, Goldberg M, Martin J-P, Masse R, eds. Prevention of Respiratory Diseases. New York, Marcel Dekker. 1993; pp. 65-79. 31. McDonald JC, Liddell FDK, Dufresne A, McDonald AD. The 1891-1920 birth cohort of Quebec chrysotile min ers and millers: mortality 1976-1988. Br J Ind Med 1993; 50: 1073-1081. 32. Piolano G, Negri E, La Yecchia C. Pira E. Decarli A, Peto J. An update of cancer mortality among chrysotile asbestos miners in Balangero, Northers Italy. Br J Ind Med 1990: 47: 810-814. 33. Shxts-Cremer GK, Liddell FDK, Logan WPD, Bezuidenhout BN. The mortality of amphibole miners in South Africa, 1946-1980. Br J Ind Med 1992; 49: 566-575. 34. Armstrong BK, De Klerk NH. Musk AM, Hobbs MST. Mortality in miners and millers of crocidolite in Western Australia Br J Ind Med 1988; 45: 5-13. 35. McDonald JC, McDonald AD, Armstrong B, Slbastien P. Cohort study of mortality in vetmiculite miners exposed to tremolite. Br J Ind Med 1986; 43: 436 444. 36. Thomas HF, Benjamin IT, Eiwood PC, Sweetnam PM. Further follow-up study of workers from an asbestos/ cement factory . Br J Ind Med 1982; 39: 273-276. 37. Ohlsoo C-G. Hogstedi C. Lung cancer among asbestos/ cemem workers: a Swedish cohort study and a review. Br J Ind Med 1985; 42: 397-402. 38. Gardner MJ, Winter PD, Pannett B. Powell CA. Follow up study of workers manufacturing chrysotile asbestos/ cement products. Br J Ind Med 1986; 43: 726-732. 39. Hughes JM. Weill H. Asbestos exposure - quantitative assessment of risk Am Rev Respir Dis 1986; 133: 5-13. 40. Fmkelstein M. Mortality among employees of an Ontario asbestos-cement factory. Am Rev Respir Dis 1984; 129: 754-761. 41. Aiies-Patin AM, Vaileron AJ. Mortality of workers in a French asbestos cemem factory 1940-1982. Br J Ind Med 1985; 42; 219-225. 42. Magsani C, Terracini G, Bertolone GP, et al. Mortal!ta per turned e altre malsttie del sistema respiratorio tra i lavorarori del cememo-amiartto a e&sale Monferrato: una studia di coone storico. Med Lav 1987; 6: 441-453. 43. Raffn E, Lynge E, Jud K. Korsgaard B. Incidence of can cer and mortality among employees in tite asbestos-cement industry in Denmark Br J Ind Med 1989; 46: 90-96. 44. AJbin M. Jakob&son K, Attewell R, Johansson L Welindcr H. Mortality and cancer morbidity in cohorts of asbestoscement workers and referents. Br J Ind Med 1990; 47: 602-610. 45. Ncubcrger M. Kundi M. Individual asbestos exposure: smoking and mortality - a cohort study in the asbestoscement industry. Br J Ind Med 1990; 47; 615-620. 46. McDonald AD. Fry JS, Woolley AJ. McDonald JC. Dust exposure in an American chrysotile textile plant. Br J Ind Med 1983. 40: 361-367. 47. Dement JM, Brown DP, Okun A. Follow-up study of chrysotile textile workers: cohort mortality and case-con trol analyses. Am J Ind Med 1994; 26: 431-447. t HWBUI0008031 - .*--". rr.sM 1942 epidemiology of mesothelioma 4a. McDonald AD. Fry JS. Woolley AJ. McDonald JC. Dim exposure and mortality in an American factory using chrysotile, amosite, and crocidolite in mainly textile man ufacture. Br J Ind Med 1983; 40: 368-374. 49. Peto J, Doll R. Herrooo C, Binns W, Clayton R. Goffe T. Relationship of mortality to measures of environmental asbestos pollution in an asbestos textile factory. Ann Occup Hyg 1985; 29: 305-355. 50. McDonald AD, Fry JS, Woolley AJ, McDonald JC. Dust exposure and mortality in an American chrysolite asbes tos iricrioa products plant. BrJ lndMed\9%4; 41:151-157. Newhouse ML. Sullivan KR. A mortality study of work ers manufacturing friction materials: 1941-1986. Br J ind Med 1989; 46: 176-179. 52. Selikoff U, Hammond EC, Seidman H. Mortality expe rience of insulation workers in the United States and Canada. 1943-1976. Ann NYAcadSci 1979; 330: 91-116. 53. Seidman H, Selikoff U, Hammond EC. Sbort-term asbestos work exposure sad long-term observation, Ann NY Acad Sci 1979; 330: 61-89. 54. Acheron ED, Gardner MJ, Winter PD, Bennett C. Cancer in a factory using amosite asbestos, ini J Epidemiol 1984; 13: 3-10. 55. McDonald AD. McDonald JC. Mesothelioma after crocidolite exposure during gas-mask manufacture. Environ Res 1978; 17: 340-346. 56. Jooes JSP. Smith PG, Pooley FD, et al. The consequeoces of exposure to asbestos dust in a wartime gas mask factory. In: Wagner JC, ed. Biological Effects of Mineral Fibres. 2. 1ARC Scientific Publications No. 30. Lyon, France IARC, 1980; pp. 637-653. 57. Talcott JA, Thurber RN, Kantor AF, et al. Asbestosassociated diseases in a cohort of cigarette-filter workere. N Eng J Med 1989; 321: 1221-1223. 58. Acheson ED, Gardner MJ, Pippard EC, Grime LP. Mortality of two groups of women who manufactured gas-masks from chrysolite and crocidolite asbestos: a 40 year follow-up. Br J Ind Med 1982; 39; 344-348. 59. Rossiter CE, Coles RM. HM Dockyard, Devcmport: 1947 mortality study. In: Wagner JC, ed. Biological Effects of Mineral Fibres. 2. IARC Scientific Publications No. 30. Lyon, France, IARC, 1980; pp. 713-721. 60. McDonald AD. Abstract of discussion. In: Glen HW. ed Proceedings of asbestos symposium. Johannesburg, South Africa, October 1977. National Institute for Metallurgy. Randburg. 1978; p. 82. 61. Langer AM, Nolan RP. Fibre type and burden in parenchy mal tissues of workers occupationally exposed to asbestos in the United States. In: Bignon Jx Peto J. Saracci R, eds. Mon-occupational Exposure to Mineral Fibres. IARC Scientific Publications No. 90. Lyon, France, IARC, 1989; pp. 330-335. 62. McDonald JC. McDonald AD. Mesothelioma: is there a background? Eur Respir Rev 1993; 3 (11): 71-73. 63. Enterline PE, Henderson VL. Geographic patterns for pleural mesothelioma deaths in the United States, 196$1981, J Natl Cancer Inst 1987; 79: 31-37. 64. Fraire AE. Cooper S. Greenberg SD, Baffler P, Langston C. Mesothelioma m childhood Cancer 1988; 62:838-847. 65. Peto J, Henderson BE, Pike MC. Trends in mesothe lioma incidence and the forecast epidemic due to asbestos exposure during World War D. In: Peto R, Schneiderman M. eds. Quantification of Occupational Cancer. Banbury Report 9. ColdSpring Harbor NY, 1981; pp. 51-72. 66. Luce D, Brocbard P. Qudnel P, et al. Malignant pleur al mesothelioma associated with exposure to tremolite. Lancet 1994; 344: 1777. Boutin C, Viailat JR, Steinbauer J. Dufour G. Gaudichet A. Bilateral pleural plaques in Corsica: a marker of nonoccupaiional asbestos exposure. In. Btgnon J. Peto J, Saracci R. eds. Nonoccupaiional Exposure to Mineral Fibres. IARC Scientific Publications No. 90. Lyon. France, IARC. 1989; pp. 406-410. 68. McConnochie K. Simonato L, Mavrides P. et al. Meso thelioma in Cyprus. In: Bignon J. Peto J, Saracci R, eds. Nonoccupaiional exposure to mineral fibres. LARC Scientific Publications No. 90. Lyon, France. IARC, 1989; pp. 411-419. 69. Gardner MJ, Saracci R. Effects on health of noo-occu- patiooal exposure to airborne mineral fibres. In: Bignon J. Peto J, Saracci R, eds. Nonoccupational Exposure to Mineral Fibres. IARC Scientific Publications No. 90. Lyon, France, IARC, 1989; pp. 375-397. 70. Langer AM. Pooley FD. Identification of single asbestos fibres in human tissues. In: Bogovski P. Gilson JC, TimbreU V, Wagner JC. Biological Effects of Asbestos. IARC Scientific Publications No. 8. Lyon, France, IARC, 1973; pp. 119-125. 7 i. Pooley FD. An examination of the fibrous mineral con tent of asbestos lung tissue from the Canadian chrysotik mining industry. Environ Res 1976; 12: 281-298. 72. Rowlands N, Gibbs GW. McDonald AD. Asbestos fibres in the lungs of chrysolite miners and millers. Ann Occup Hyg 1982;26:411-415. 73. Jones JSP, Roberts GH, Pooley FD. et al. The pathology and mineral content of lungs in cases of mesothelioma in the United Kingdom in 1976. in: Wagner JC, d. Bio logical Effects of Mineral Fibres. Vol 1. IARC Scientific Publications No. 30. Lyon. France, IARC, 1980; pp. 187-- 199. 74. McDonald AD, McDonald JC, Pooley FD. Mineral fibre contest of lung in mesothdial rumours in North America. Ann Occup Hyg 1982; 26: 417-422. 75. Mowe G, Gyiseth B, Hartveit F. Skaug V. Fibre con centration in lung tissue of patients with malignant mesotheboroa: a ease-control study. Cancer 1985; 56:1089-1093. 76. Gaudichet A. Janson X, Monchaux G, et al. Assessment by analytical microscopy of the total lung fibre burden in mesothelioma patients matched with four other patholog ical series. Ann Occup Hyg 1988; 32 (Suppl. 1); 213-223. 77. McDonald JC, Armstrong B. Case B, et al. Mesothelioma and asbestos fibre type. Cancer 1989; 63: 1544-1547. 78. Rogers AJ, Leigh J, Berry G. Ferguson DA. Mulder HB. Ackad M. Relationship between lung asbestos fibre type and concentration and relative risk of mesothelioma: a case-control study. Cancer 1991; 67: 1912-1920. 79. McDonald JC. Tremolite, other amphiboles, and mesothe lioma. Am J Ind Med 1988; 14: 247-249. 80. Churg A. Analysis of lung asbestos content Br J Ind Med 1991; 48: 649-452. 81. Sehastien P. McDonald JC, McDonald AD, Case B, Harley R. Respiratory cancer in chrysolite textile and mining industries: exposure inferences from lung analy sis. Br J Ind Med 1989; 46: 180-187. 82. Roggli VL. Pratt PC. Brody AR. Asbestos fiber type in malignant mesothelioma: an analytical scanning electron microscopic study of 94 cases. Am J Ind Med 1993; 23: 605-614. 83. McDonald JC. Epidemiological significance of mineral fiber persistence in human lung tissue. Environ Health Perspect 1994; 103 (Suppl. 5): 221-224. 84. McDonald AD, Gibbs GW. Rowlands N. Chrysolite and tremolite lung content of Quebec miners. Ann Occup Hyg 1988; 32 (Suppl. I): 1182-1183. 85. McDonald JC. McDonald AD. Chrysolite, tremolite. and mesothelioma. Science 1995: 267: 775-776. :mma HWBUI0008032