Document e7z0RjJ8jX9GG9XwRxkz8MnMM

i^fli:^,ji]c-jiija^^^iiyiiffliiaaiMiaiHWiiiiiaiiifc.anaiiiai The control limit for asbestos RECEIVED flov 4 1983 ED Acheson and MJ Gardner MRC Environmental Epidemiology Unit DU 038396 DUP 0947118.002 I The control limit for asbestos An up-date of the relevant sections of The Illeffects of Asbestos upon Health prepared by the authors for the Advisory Committee on Asbestos and published in Asbestos: vol 2: final report of the Advisory Committee Health and Safety Commission 1979 HMSO E D Acheson and M J Gardner, MRC Environmental Epidemiology Unit, Southampton General Hospital, Southampton S094XY DUP 0947118.003 Health & Safety Commiition _* -- i. U- C*innn/ fWir# DU 038397 iaai-ila3lFBii!i"Ti;M :kl__ ailffgnil J Acknowledgments ' " ................ We are grateful to the following people for reading earlier drafts of the manuscript and for contributing comments: Professor Sir Richard Doll Dr J C Gilson Dr M Greenberg Professor F D K Liddell Professor A D McDonald Professor J C McDonald Mr J Peto Dr J Steel Professor M E H Turner-Warwick The final draft was discussed by an ad hoc Expert Panel set up at the request of the Health and Safety Commission. The discussion in that Panel has been taken into account in this text, but this final version is the responsibility of the authors. The Panel members were: Dr K P Duncan (Chairman) Deputy Director-General, Health and Safety Executive Professor A Mair University of Dundee Dr T W Meade Medical Research Council Epidemiology and Medical Care Unit DrMKB Molyneux Shell UK Ltd Dr R Owen Trades Union Congress Mr J Peto University of Oxford Dr J Steel University of Newcastle upon Tyne Dr P J Taylor Faculty of Occupational Medicine Professor M E H Turner-Warwick Bromplon Hospital We would like to thank Mrs Bridget Wilde for typing the manuscript. E D Acheson M J Gardner June 1983 DUP 0947118.004 DU 038398 I 1 Contents____________________________________________ ______________________________ Introduction /... Trends in asbestos imports 2 Update 2 . ., 1 -~ The asbestos-related diseases 2 Update ,? ' ~ The relationship of health effects to fibre type 3 Update 4 Summary 8 ' Dose-response relationships 9 Update 9 Implications for the industrial control limit Crocidolite 13 Amosite 13 Chrysotile 14 Update 14 13 Conclusions 16 Appendix A Review of report by Dr J G Morris 16 Appendix B Note by Dr J Steel on dust estimates in the South Carolina chrysotile textile factory 17 Appendix C Note by Professor M Turner-Warwick on indices of adverse effects and asbestosis in the BOHS Committee on Asbestos study 18 Appendix D Problems of measurement of dose and response 18 References 19 Figures 22 Tables 23 DUP 0947118.005 DU 038399 igiHiSWHn Introduction _______________________ 1 Our brief is to prepare a paper for the Health and Safety Commission which will examine in relation to the control limit for asbestos: "l. Any material which has been suggested was available but not presented to the earlier Advisory Committee on Asbestos which would have - influenced significantly your joint report to the Committee........ . ' ' : j 2. Any published data since the time of the report that would cause you to think that significant. change should now be made to the report as then published."* .. . _ 2 The review is to be limited to the issue of the control limit in the workplace and the evidence bear ing upon it, including the question of the desirability of different limits for different types of fibre. We have not been required to bring up-to-date the sections of our report which dealt with asbestos outside the workplace--for example, inside schools and public buildings, out of doors, and in drinking water. 3 In the main part of this review we will discuss the relevant scientific literature published since 1979 in relation to each of the main conclusions of our previous report which are within our present terms of reference. To assist the reader a resume of each of these conclusions (in italics) will precede the dis cussion of the recent literature, the reference to the appropriate paragraphs in the 1979 report being given in parentheses. 4 A review of an unpublished paper by Dr J G Morris dated 24 February 1978 which was circulated to the Medical Working Group of the Advisory Committee on Asbestos in April of that year is given in Appendix A. Appendix B consists of a critique by Dr J Steel of the dust measurements in the paper by Dr J M Dement and colleagues describing the mor tality of South Carolina textile workers exposed to chrysotile. Appendix C contains some detailed comments by Professor M Tumer-Warwick on the applicability to asbestos-related diseases of the indices of adverse effects used in the morbidity survey by the British Occupational Hygiene Society (1983). Appendix D includes a summary of some of the main problems in the interpretation of the available data on dust measurements and their relationship to the occurrence of disease. 5 We have also examined various other papers pro vided for us by the Health and Safety Executive. These include excerpts of a transcript of a television programme broadcast by Yorkshire TV on 20 July 1982; a list of points which arose out of the pro Excerpt from letter from Dr K P Duncan to Protestor E D Acheson dated 15 November 19S2. gramme or out of research for it which were the ' subject of a meeting held by the Health and Safety Executive on 26 August 1982; papers and correspon- ' dence dealing with various cases of mesothelioma; a ' reprint of a paper by Drs V L Henderson and PE Enterline on mortality in a US asbestos factory; a draft of a study on The health experience of two UK ' asbestos factories by a sub-committge of the British > Occupational Hygiene Society prepared for a meeting in 1980; some material concerning six fatal cases of lung cancer in persons who had entered the asbestos industry subsequent to 1951; and some material relat ing to the cohort of workers studied by Dr J G Morris (sec Appendix A). 6 We summarise the points contained in this material which are relevant to our terms of reference as follows: (a) The incidence of pleural mesothelioma, its rel ationship to fibre type and the bearing this has on mesothelioma at Rochdale. This is discussed in the context of the recent scientific literature in paras 24 to 40, and with particular reference to Rochdale in para 30. The evidence that crocidolite has been a factor in the occurrence of the cases of pleural mesothelioma at Rochdale has strengthened since our last report. (b) The prevalence of "adverse effects" and their rel ationship to asbestos in persons employed in Rochdale since 1951. This is discussed in paras 59 to 64, para 95, and in Appendix A and Appendix C. (c) The mortality from lung cancer at Rochdale in persons first employed since 1951. The mortality from lung cancer is dealt with in relation to the recent literature in paras 65 to 80 and 89 to 103 and in respect to Rochdale in para 75. The finding of a higher mortality in men first exposed after 1950 than previously (Peto, 1980a) now seems to have been an anomaly due to small numbers (Peto J, personal communication). 7 Henderson and Enterline (1979) report a large number of mesotheliomas amongst the younger work force of a factory from which they had previously studied a group of pensioners (Enterline and Henderson, 1973), only one of whom was known to have developed a mesothelioma. The occurrence of a large number of mesotheliomas in younger men in this workforce had previously been noted by Borow et a! (1973), and was reported in our previous review (Acheson and Gardner, 1979a, p 29). Crocidolite is known to have been used in this factory, but nothing was or is known about the history of fibre exposure in these mesothelioma cases. The recent paper by Henderson and Enterline does not contain any new in formation with a bearing on the control limit for chrysotile, unless one takes the view that crocidolite was not a factor in the causation of these cases. 1 DU 038400 DUP 0947118.006 Trends in asbestos imports____________ 8 For practical purposes the types of asbestos fibre important in the UK were shown to be chrysolite, crocidolite and amosite. In 1975 chrysolite accounted for about 85 T* of the total imports offibre, with almost the whole of the remainder being amosite. The considerable expansion in imports of amosite since the end of World War II was noted, so that in spite of the discontinuation of crocidolite imports in 1970, the absoluje and proportional amount of amphibote fibre (in the form of amosite) imported in 1975 was con siderably greater than that in 1945 (paras 72, 74 and 75). Up-data 9 Table 1 brings up-to-date the information on imports of raw asbestos fibre into the UK published in our previous report (Acheson and Gardner, 1979a). The Table includes figures for the serpentine asbestos, chrysolite (white asbestos), and for the amphiboles, crocidolite (blue asbestos) and amosite (brown asbes tos). The most striking change is that the import of amosite has virtually ceased, as shown in Figure 1 (the estimate for 1982 is 16 tonnes). Of the 2700 tonnes of amosite imported during 1978, the large majority (2000 tonnes) was incorporated into insulation board, with almost all the remaining 700 tonnes being used for asbestos cement pressure pipes. In 1980 and 1981 all the imported amosite was used in the latter appli cation. The manufacture of insulation board is now carried out with other materials. Chrysotile imports have also declined, but at 76400 tonnes in 1981 are estimated still to be about 50% above the 1946 figure. The largest uses of chrysotile during 1981 were in as bestos cement (280S0 tonnes), friction materials (11 200 tonnes) and floor tiles (8000 tonnes), with a variety of other applications (including 3000 tonnes in textile products) making up the remaining 29150 tonnes (Table 2). 10 Although imports of amphiboles as raw fibre into the UK have at present effectively ceased we note that the import of manufactured products containing am phiboles continues. Also, substantial quantities of these materials remain in the insulation of buildings, ships and railway carriages, and in certain cement products, including pressure piping. Overall, more than half-a-million tonnes of amphiboles had been im ported into the UK up to 1982. From 1 January 1983 the utilisation, repair or demolition of materials con taining amphiboles are subject to control limits of 0.5 fibres/ml for amosite and 0.2 fibres/m! for croci dolite. 11 As far as production is concerned, chrysotile is now for practical purposes the only type of asbestos used in manufacture in the UK. However it will be necessary for many years to come to continue to con trol the servicing and demolition of previously pro duced articles containing amphiboles. w The asbestos-related diseases 12 ASBESTOSIS, which may be defined as fibrosis of the lungs caused by asbestos dusts which may or may not be associated with fibrosis of the parietal (outer) or pulmonary (inner) pleura, LUNG CANCER, and MESOTHELIOMA of the PLEURA or PERITONEUM were considered as having been universally accepted as diseases with a causal relation ship to the inhalation (and for peritoneal meso thelioma possibly the ingestion) of asbestos fibre. * 13 With regard to other putative asbestos-related diseases we concluded that there is a risk (small in absolute terms) of CANCER OF THE LAR YNX as sociated with mixtures of asbestos containing amphi boles. In respect of CANCERS OF THE ALIMEN TARY TRACT we again noted an association with amphibole, that the clinical and pathological data available about such cancers were often sketchy and that some at least were likely to have been undisclosed cases ofperitoneal mesothelioma. However, some of the excesses noted in men exposed to amphiboles were Benign pleural effusion and pleural thickening, noted as asbestosrelated in our previous report, are not considered further here. > .- so large that an association between cancers of the ali mentary tract and inhalation or ingestion of such fibres seemed likely (paras 8--12). Up-data 14 In recent publications up-dating the mortality experience of 17 800 North American insulators and workers in a New Jersey amosite asbestos factory, Selikoff has continued to observe excesses of deaths from cancer of the oesophagus, stomach, large bowel, mouth and pharynx and kidney (Selikoff et ol, 1979; Seidman et at, 1979) but has shown that an apparent excess of deaths ascribed to cancer of the pancreas in the insulators was likely to be due to mis-diagnosed peritoneal mesotheliomas and cancers of other parts of the alimentary tract (Selikoff and Seidman, 1981). The men in each of these studies also show remark ably high cancer rates for all other cancers combined, and this is true whether the original death certificate diagnosis is accepted or a 'corrected' diagnosis based on examination of hospital records is used. The latest DU 038401 DUP 0947118.007 published paper on the Barking asbestos factory of asbestos in certain sectors of industry. In respect of shows an excess of gastro-intestinal cancers (60 cancers of the upper alimentary tract, social factors observed deaths; 44 expected) (Newhouse and Berry, are particularly important, and differences between 1979) , as also does the most recent report on pro the workforces studied and the standard populations duction and maintenance employees in the United with which they have been compared need to be taken States (55 observed; 40 expected) (Henderson and into account. Enterline, 1979). No significant excess of deaths from alimentary tract cancer was reported in a study of 17 Two recent studies of women, involved in the London insulation board makers exposed to amosite manufacture of gas masks from crocidolite in World under dusty conditions, but only 7% of the cohort War II have reported a significantly increased mor have died and the likelihood of having detected a 3:1 tality from ovarian cancer (Wignall and Fox, 1982; relative risk such as was observed by Selikoff is small Acheson et al, 1982). These findings supplement an (Acheson et al, in press). The most recent report on earlier report of excess ovarian tumours among the mortality among men at Rochdale exposed principally most severely exposed group of female asbestos to chrysotile, but also to crocidolite, supported the factory workers manufacturing products containing earlier finding of no excess deaths from alimentary amphiboles and chrysotile (Newhouse and Berry, tract cancer as a whole (14 deaths observed; 12.6 1979). A group of women who manufactured civilian expected) (Pcto, 1980a). 15 McDonald and her colleagues have recently reported a slight excess of deaths from alimentary tract cancer (26 observed; 17.1 expected) in a South Carolina textile factory (McDonald et al, unpublished a). This study is important because chrysotile was the only fibre used (see also para 77). No excess of ali gas masks out of chrysotile experienced no such in crease in mortality (Acheson et al, 1982), nor was any excess found among a group of women involved in the manufacture of friction materials (Newhouse et al, 1982). It is not yet clear whether the excess mortality ascribed to this cause is in fact due to ovarian tumours or to misdiagnosed peritoneal mesotheliomas. mentary tract cancer (132 deaths observed; 134.6 18 Three further case-control studies of laryngeal expected) was found in an English factory manufac cancer have been published since our last report. A turing friction materials from chrysotile (Berry and Canadian study supports the role of asbestos in cancer Newhouse, 1983). Likewise no excesses were reported of the larynx, although the influence of this factor is from four asbestos cement factories--two in New small compared with tobacco and alcohol (Burch et Orleans using chrysotile (predominantly) and crocido al, 1981). A study by Newhouse et al (1980) is lite (25 deaths observed; 50.1 expected) (Weill et al, negative. It should be pointed out, however, that the 1980) , one in South Wales in which the fibre used was latter study had a 30% chance of missing a relative almost exclusively chrysotile (18 observed; 19.6 risk of 2 associated with asbestos exposure and that expected) (Thomas et al, 1982), and the fourth in the upper 95% confidence limit for the relative risk of Denmark using chrysotile, crocidolite and amosite (31 laryngeal cancer is 2.1. An excess of cancer of the observed; 29.9 expected), although in this study there larynx has been reported in a cohort study of asbestos was a slight suggestion of an excess of stomach cancer cement workers in Denmark (6 observed; 2.9 ex (Clemmesen and Hjalgrim-Jensen, 1981)*. No excess pected), where exposure was to chrysotile, crocidolite mortality from alimentary tract cancer has been and amosite (Clemmesen and Hjalgrim-Jensen, 1981). observed in a cohort of 6200 West Australian ex- crocidolite miners who suffered heavy dust exposure 19 Ross and colleagues (1982) have suggested that (Hobbs et al, 1980), or in two groups of women there may be an association between non-Hodgkin's engaged in the manufacture of gas masks in World lymphoma of the alimentary tract and exposure to as War II (Wignall and Fox, 1982; Achcson el al, 1982). bestos. A subsequent report of a case-control study of 16 It can be seen that the evidence published since non-Hodgkin's lymphoma estimated a relative risk of our previous report has tended to weaken the view 1.3 (but with wide confidence limits of 0.5 and 3.1) that exposure to asbestos causes alimentary tract due to asbestos after adjusting for other relevant ex cancer. Large excesses continue in the studies where posures (Benglsson et al, 1982). The debate on this they were found in 1979, but in other studies there are matter continues (Olsson and Brandt, 1983). no such excesses. The results remain inconsistent and leave open the question whether such an effect should be attributed to exposure to asbestos fibre per se or to some other factor(s) in the preparation and utilisation Our attention has been drawn to * study of an asbestos cement factory in Belgium in which an excess of alimentary but not of respiratory cancer was detected (Lacquet et at, 1980). This study appears to have serious methodological shortcomings. The relationship of health effects to fibre type 20 In animals, experimental work involving inha lation and pleural implantation shows that chrysotile can produce al least as many mesotheliomas and malignant lung tumours as samples of amphiboles. 21 For mesothelioma in man, evidence from miners, from process workers exposed to a single fibre type, from the distribution of neighbourhood and domestic cases, andfrom the geography of mesothelioma pre- 3 DUP 0947118.008 DU 038402 JigaalB, sents a powerful case from four different points of view that crocidolite has been more dangerous than chrysolite. The position of amosile may be inter mediate between crocidolite and chrysotile. It can be concluded that exposure to chrysotile alone so far has rarely been shown to cause mesothelioma. Evidence in man relating lung cancer and asbestosis to fibre type is inconclusive, but the small amount which exists is consistent with the view that exposure to amphiboles or mixtures rich in them have been more dangerous than to chrysotile alone (paras III--164). Up-date Animal experiments 22 A number of recent papers and reviews relating to experimental work in animals have upheld the hy pothesis of Stanton that durable fibres, of which as bestos is but one example, can cause cancer irrespec tive of their physicochemical nature, provided that their configuration falls within well-defined ranges of diameter and length with thin long fibres being most carcinogenic (Harington, 1981; Stanton et al, 1981; Craighead and Mossman, 1982). It has been suggested, however, by Bertrand and Pezeral (1980) that there may be no limit of length or diameter below which carcinogenicity disappears but that carci nogenicity may be a continuous increasing function of the aspect ratio of the fibres. Up to 14 fibrous materials are now known to produce malignant neo plasms following implantation in the pleural or perito neal cavities of animals: amosite, anthophyllite, chrysotile, crocidolite, and tremolite; borosilicate glass, aluminium silicate, glass, mineral wool, alu minium oxide, potassium titanate, silicon carbide and sodium aluminium carbonate; wollastonite and attapulgite (Harington, 1981; Stanton el al, 1981). 23 As far as the animal work is concerned, there fore, there is little change in the position we summar ised in 1979: namely that in the experimental situation specially prepared samples of chrysotile are at least as carcinogenic as are amphiboles both in implantation and inhalation experiments. Davis has suggested that the carcinogenicity of chrysotile may be increased due to the separation of large fibres into individual fibrils in the tissues (Davis, 1982). Mesothelioma in man 24 Peto and co-workers have drawn attention to the relationship between the incidence rates of both peri toneal and pleural mesothelioma and the third power (approximately) of the interval since first exposure to asbestos (Peto et al, 1982). They have calculated that the highest relative risks of peritoneal mesothelioma so far reported are in North American insulators, workers in a New Jersey amosile factory, and workers at Barking, all of whom were exposed to amphiboles. Peritoneal mesotheliomas have also been reported from Canada and Nottingham among gas-mask makers exposed to crocidolite (McDonald and McDonald, 1978; Jones et al, 1980; Wignall and Fox, 1982) , and recently from Western Australian crocidolitc miners (Hobbs, 1983). In the last two studies mentioned, however, the numbers of cases have been very small compared with the numbers of cases of pleural mesothelioma (Nottingham: 5 peritoneal, 34 pleural; Western Australia: 4 peritoneal, 60 pleural). Of five mesotheliomas reported as the underlying cause of death from a British factory using almost ex clusively commercial amosite in the manufacture of insulation board, one was a peritoneal tumour (Acheson et at, 1981). No peritoneal mesotheliomas have been reported from Quebec among chrysotile miners and millers who were not elsewhere exposed to crocidolite (McDonald, 1980), or from some factories manufacturing friction materials or asbestos cement from chrysotile. A single peritoneal mesothelioma has been reported from a textile factory in South Carolina using exclusively chrysotile (Dement et at, 1982; McDonald et at, unpublished a). With this exception all reported cases of peritoneal mesothelioma seem likely to have been exposed to one or both of the amphiboles, amosite or crocidolite. Difficulties in diagnosis and other factors, as yet unknown, may explain the variation in estimated life risks of perito neal mesothelioma amongst those exposed to amphi boles. These range from zero at Rochdale to a figure perhaps as high as 10% for American insulation workers and factory workers first exposed to amosite at age 20 (Peto et al, 1982). 25 A study of 144 cases of mesothelioma know n by Cape Industries to have occurred in their employees has recently been published (Browne and Smither, 1983) . It is not certain what proportion of the total number of cases that have occurred in workers and ex-workers these cases represent, and little numerical information is given about the populations from which the cases were derived. Within these limits the results support a relationship between both pleural and peritoneal mesothelioma and exposure to amphi boles, particularly crocidolite. The evidence suggests that workers who developed peritoneal mesothelioma may have been exposed to heavier doses of fibre than those who de\eloped pleura! mesothelioma. 26 Table 3 shows the excess lung cancer deaths over those expected and the numbers of deaths due to pleural and peritoneal mesothelioma for miners and millers, industrial process, insulation and shipyard workers by fibre type. A substantial difference also exists in the incidence of pleural mesothelioma between workers exposed to chrysotile alone and those who have been exposed to amphiboles, although the differences are not as great as those for peritoneal mesothelioma which have been described above. In men, 0.3% of the deaths amongst chrysotile miners (McDonald et al, 1980b) were due to pleural meso thelioma as opposed to 8.8% among crocidolite miners (Hobbs et al, 1980). Forty-four additional deaths from pleural mesothelioma have recently been DU 038403 DUP 0947118.009 reported among the West Australian crocidolite miners but these cannot be included in the Table as the up-dated total number of deaths is not known. In the remainder of the Table (manufacture, insulation, shipyards) only two (0.1%) deaths (out of 2752) from pleural mesothelioma occurred among men exposed to chrysotile only (excluding Dement el a! (1982) which is a sub-cohort of McDonald el al (unpublished a)) but 181 (1.7%) of the 10447 deaths which occurred in association with exposures to amphiboles alone or with chrysotile were pleural mesotheliomas. Among women all but one of 34 deaths from pleural meso thelioma (3.0% of the total of 1103) were associated with amphiboles or mixed exposures. It has previously been suggested that the mortality from mesothelioma in a workforce may be used to predict other asbestos related cancer mortality (McDonald and McDonald, 1981). However, as can be seen from Table 3, the ratio between the numbers of mesothelioma deaths and excess lung cancer mortality varies between wide limits. 27 Talent and his co-workers have undertaken a survey of a list of black ex-employees of the Cape crocidolite mines in South Africa, which is not in cluded in Table 2 because it measures the prevalence rather than mortality due to pleural mesothelioma (Talent el al, 1980). However, it supports a high risk of this disease associated with crocidolite. The follow up was limited to 1185 men who were citizens of Bophuthalswana, most of whom had worked in the mines for less than five years, and who had been recruited into employment in the mines at least 11 years before the survey. Two hundred and fifteen of the men were found to have died prior to the survey, but nothing could be discovered of the causes of death. Of 755 ex-employees who were located four were found on clinical examination and biopsy to be suffering from pleural mesothelioma (a prevalence rate of 0.5%). Two further cases of pleural meso thelioma were discovered shortly after the completion of the survey. Given an average duration of illness of about two years a prevalence of 0.5% suggests an annual incidence at the time of the survey of about 1 in 400. The prevalence of mesothelioma was found to increase with increasing time since first `7 exposure - from 0.4% at 16--20 years, to 1.3% at 21--25 years, and 2.5% beyond 25 years. In further studies of 947 ex-employees of these mines not on the list, 11 other cases of pleural mesotheliomas were identified (a prevalence of 1.2%). Some of the men in the last group may have attended for examination because they were ill. 28 Rossiter and Coles (1980) found 31 meso theliomas among 1043 deaths in a cohort of 6076 British naval dockyard workers who had been exposed to amphiboles and chrysotile. There was a deficiency (84 observed; 100 expected) of deaths from lung cancer in this group of men. This is an unexpected deficit. The reason for it is unknown but it has been suggested that it may be related to lower cigarette smoking habits in these workers or possibly to ex posure to asbestos fibres of an unusual composition in terms of size due to the use of respiratory protection equipment by the men. 29 The lack of pleural mesotheliomas among 872 deaths in the workforce of a factory manufacturing textiles from chrysotile in South Carolina, in which the men suffered a substantial excess mortality from lung cancer, is of particular interest (McDonald and Fry, 1982). In the same paper, McDonald and Fry also report an absence of mesotheliomas among 1630 deaths in a factory using only chrysotile in the manufacture of friction materials. The authors con trast these findings with the experience of another American factory which used amphiboles as well as chrysotile where 14 deaths from mesothelioma occurred among 895 deaths (McDonald et al, un published b). 30 Peto has suggested that the rate of mortality from pleural mesothelioma at Rochdale, where the principal fibre used was chrysotile, is similar to that at Barking where in addition to chrysotile large quan tities of amphiboles were also used (Peto, 1980b). His comparison was between three sub-cohorts from Rochdale, selected because they had at least 10 (and in one case more than 20) years' exposure, and the complete Barking cohort (excluding only laggers), w hich had large numbers of men and women exposed for less than two years. Since the incidence of pleural mesothelioma depends on duration of exposure, these figures could be interpreted to indicate a higher risk, for any given length of employment, at Barking than at Rochdale, rather than similar risks. Peto's implied conclusion was that the heavier crocidolite exposure of workers at Barking more than compensated for their shorter duration of exposure. Subsequently a study of lung tissue in 12 cases of mesothelioma and 12 controls, all of whom were men who had worked at Rochdale, has shown a substantial burden of croci dolite as well as chrysotile (Wagner et al, 1982). According to the authors the lung tissue of eight of the cases of mesothelioma reported by Peto and colleagues (1977) are included in this series. As no dif ferences in the nature of the fibre burdens were found between the mesothelioma cases and controls, it is dif ficult to attribute the tumours with certainty to either fibre type. The controls contained at least five deaths from asbestos-related causes, four from lung cancer and one from asbestosis. The crocidolite fibre counts in both cases and controls are similar to those found in the lungs of gas-mask workers by Morgan and Holmes (1982). The results of this study, therefore, suggest that substantial amounts of crocidolite must have been used at the Rochdale factory.* This is borne out by recent data which sujjest that about 10000 tonnes of crocidolite fibre and yarn were used between 1931 --1970 (Doll R, personal communication), not "at least 2,500 tonnes" as we indicated on p 26 of our previous report (Acheson and Gardner, 1979a). DUP 0947118.010 DU 038404 ilgfaM , 31 Newhouse, Berry and Skidmore (1982) have carried out detailed studies of a factory which has manufactured friction materials from chrysotile since 1920, but which utilised crocidolite in a particular area of the factory (Skidmore and Dufficy, 1983) for two periods from 1929 to 1933 and from 1939 to 1944 respectively. A total of 13460 subjects who have been emplo>ed since proper records were set up in 1941, and of whom over 99% were traced to the end of 1979, were included. Among these persons 10 deaths due to pleural mesothelioma were found. In order to study the relevant environmental factors involved four controls for each case were selected, matched for sex, year of starting work, year of birth, and survival up to the time of death of the case of mesothelioma. In addition each control had to have been employed at the factory during one or more of the periods when crocidolite was used for the same time as the case. The results indicate that eight of the ten mesothelioma cases had definite and one had fringe exposure to crocidolite, as compared with three definite and seven fringe exposures to crocidolite among the 40 controls (Table 4a). A strong and significant association was thus demonstrated with crocidolite exposure. An association with exposure to chrysotile of at least 5 fibres/ml was also demonstrated (Table 4b), but combined exposure to chrysotile and crocidolite may also be relevant (Table 4c). In an attempt to hold con stant the effect of chrysotile, those matched sets where the case and at least one of the controls had been exposed to 5 fibres/ml of chrysotile were examined. Within these sets it was possible to show an association with exposure to crocidolite (five out of six such cases and two out of ten such controls had been exposed to crocidolite; a two-sided significance test gives p = 0.06). A reverse analysis, in matched sets where the case and at least one control had definite exposure to crocidolite, is limited by small numbers, with all three such cases and two out of three such controls being exposed to at least 5 fibres/ml of chrysotile. Three further pleural meso theliomas are known to have occurred at the factory, but adequate reasons for their exclusion are given. They were not included in the study population as they had, in fact, left before the filing system was set up in 1941. A further mesothelioma death, which occurred in late 1980, of a man who had worked at the factory for only two weeks, is reported in a more recent paper (Berry and Newhouse, 1983). There is no evidence that this man was exposed to crocidolite in the factory. Examination of lung tissue from seven of the mesothelioma cases showed crocidolite and chrysotile in excess of 1 x 10* fibres per gram. Un fortunately, no samples of lung tissue from controls were examined. 32 The results of the case/control study are compat ible with both crocidolite and chrysotile exerting an influence, possibly jointly (Table 4c), on the incidence of mesothelioma. However, an analysis (Berry G, per sonal communication) to investigate this, taking account of the matching and disregarding fringe ex posure to crocidoli.e, finds no evidence of any syner gistic effect due to the mixture on top of the additive effects of crocidolite and chrysotile. Acheson and Gardner (1979b) suggested the possibility of a syner gistic interaction between chrysotile and amphiboles on the incidence of mesothelioma on the basis of studies of the quantities of fibres of different types in lung tissue and epidemiological data, but this was not borne out in later data (Acheson and Gardner, 1980a). 33 Langer and McCaughey (1982) have described a case of pleural mesothelioma in a man whose sole reported exposure to asbestos was to chrysotile during brake maintenance and repair. Fibrils of chrysotile but not of amphiboles were found in the thoracic tissues at autopsy. 34 McDonald and McDonald (1980) attempted to ascertain all fatal malignant mesotheliomas in Canada (1960--75) and the US (1972), and obtained records of 668 cases. A control matched for sex, age and year of death in which pulmonary secondaries from a nonpulmonary tumour were present was selected for each case. Occupational histories were obtained from cases and controls. By far the highest relative risk (46.0) was for insulation work, with asbestos production and manufacture (6.1) second, and heating trades other than insulation (4.4) third. Although this survey included all mesotheliomas that could be found in Canada during a 16-year period (1960--75), only seven of the subjects had been associated with the Quebec chrysotile mines, four having been employed in the mines and three being children of employees. No other subjects with mesothelioma who had lived in the Quebec mining area were found. This study supports the reports which we summarised previously which showed the very low incidence of mesothelioma in and about the Quebec chrysotile mines, although these have been established on a large scale since early in the century (Acheson and Gardner, 1979a). Although pleural and peritoneal mesotheliomas were considered together in this survey, a particularly strong relationship was noted between work in insulation and the occurrence of peritoneal meso thelioma. 35 In a study of the geographical distribution of deaths from pleural mesothelioma during 1968--78 in England and Wales, cases were concentrated in areas where asbestos, in particular amphiboles, was used in the relevant past (Gardner et al, 1982; Gardner, 1983a). Among men the major locations were naval dockyards and ports where shipbuilding and repairing were carried out, whereas among women a high occurrence of mesothelioma was found in areas where gas masks had been manufactured. For both sexes there were areas with high rates on the eastern side of London, and in Leeds and Rochdale. These findings arc consistent with the view that in this country a large proportion of mesothelioma deaths are asbestosrelated. DU 038405 DUP 0947118.011 36 Acheson and his colleagues compared the mor 39 Wagner, Pooley and their colleagues (1982) tality over a period of 40 years of two groups of compared the lung burden at autopsy of cases Lancashire women who had manufactured respectively submitted to the British Pneumoconiosis Medical service gas masks (containing crocidolite) at Leyland Panel with control autopsies and mesotheliomas from and civilian gas masks (containing chrysotilc) at various sources. Greater quantities (by a factor of Blackburn (Achcson el al, 1982). Mesothelioma was about 100) of both crocidolite and amosite were mentioned on five of the 219 death certificates from found in the Panel cases than in the controls, but the the former (three pleural, two peritoneal) and one quantities of chrysotile were similar-. The amount of (pleural) of 177 from the latter. The woman who died amphiboles was similar in each category of Panel from pleural mesothelioma in the civilian gas-mask cases, namely mesothelioma, lung cancer and asbes- factory at Blackburn had also worked in another gas tosis, but in the cases of asbestosis the amount of am mask factory in the same city where crocidolite was phiboles, but not of chrysotile, related well quantitat used. The mesothelioma cases from Blackburn ively to the severity of the disease. Morgan and referred to in the paper of Morgan and Holmes (1982) Holmes (1982) measured the concentrations of amphi also worked in the latter factory (Grimes P, personal boles fibres in the lungs of 26 mesothelioma deaths communication). among women who made gas masks in three English 37 If it is accepted that, from the point of view of the risk of mesothelioma, crocidolite and amositc are more dangerous than chrysolite it does not necessarily follow that a fibre of a particular configuration of one mineral is more dangerous than that of another. Walton (1982) has pointed out that crocidolite pro duces fine straight fibres more easily (i.e., with the application of less energy) than chrysotile, and that in the past people working with crocidolite have been exposed to substantially higher concentrations of fibres in "the carcinogenic size range" than those working with other varieties. He further points out that most of these fibres w ould not be visible by optical microscopy and would therefore not be counted in routine industrial hygiene measurements. cities. In 19 cases levels of greater than 1 x 10* fibres per gram of dried lung were found, but there were no controls. Amosite and mesothelioma 40 The occurrence of four deaths attributed to pleural mesothelioma and one to peritoneal meso thelioma among 431 deaths in London insulators exposed to amositc has already been noted (Acheson et a!, 1981). These results, the analysis of American insulators reported by Peto and his colleagues (Peto et al, 1982) and referred to in para 24, and the findings of the analysis of the fibre burden in the lungs of American and Canadian cases of mesothelioma sup port our previously expressed view that the risk of mesothelioma in association with amosite is greater than with chrysotile. However the nature of the lung Examination of human lung tissues fibre burden of the London insulators has not yet been examined. 38 Studies of blocks of lung tissue were carried out in 99 cases and matched controls from the North American survey of mesothelioma described above (McDonald el a!, 1982). Electron microscopy revealed equal quantities of chrysotile among the male cases and their controls. However, there were more cases than controls (26 cases compared to eight controls) with more than 1 x 10` fibres of amosite per gram, and more cases than controls (15 cases compared to five controls) with more than 1 x 10* fibres of crocidolitc per gram. In the females there were slight excesses of chrysotile and amphiboles among the cases compared to the controls. Although these findings are consistent with the view that the male mesothelioma cases inhaled more amphiboles than the controls, the fibre burden of lung tissue at autopsy does not necessarily reflect the nature of the fibres which penetrate to the pleura. Scbasticn and his colleagues (1982) have shown that chrysotile may be present in the pleura in quantities unrelated to the amount and type of asbestos in the lung. Pooley and Clark (1979) have shown that the configuration of fibres of the three common types of asbestos recovered from human lung tissue differs widely, and that the shortest fibres with the smallest diameter distribution on Lung cancer and fibre type 41 As was the case at the time of our previous report, no data have been published from which it is possible to relate dust-exposure measurements for crocidolite in the absence of other types of asbestos to the risks of lung cancer. Weill and his colleagues studied men engaged in the production of asbestos cement in two factories, in one of which there was exposure to crocidolite in addition to chrysotile during the manufacture of pipes (Weill et al, 1979). They showed significantly raised mortality from lung cancer in their highest cumulative dust exposure group among those exposed to crocidolite, but not for those exposed to chrysotile alone. This result may or may not have been influenced by the fact that it was possible to confirm the living/dcad status of only 75% of the workforce at the end of the follow-up period.* 42 A summary of all the published estimates of the slopes of the linear relationships between mortality from lung cancer and cumulative dose of fibre has been prepared in Table 6. According to Nicholson the lowest and highest values of the slopes for exposure to chrysotile alone are 0.06 to 5.3 and for mixed ex- average are chrysotile. The biological significance of these fibres is uncertain (Walton, 1982). The tiacc rate in this study is now reported to be 94*7, (Weill, personil communication). ____ 7 DUP 0947118.012 DU 038406 afiMiiittii mmiM posures 0.07 to 8.4. According to the calculations of lile and amosite,* but within the categories of workers Liddell the figures are 0.038 to 1.6 (chrysotile) and exposed to these fibres there are large variations in the 0.06 to 2.7 (mixed). There is thus overlap between the estimated life risk which reflect the influence of other estimated slopes for chrysotile and mixed exposures in factors, some of which are unknown. As far as both sets of calculations. In addition we have already pleural mesothelioma is concerned, although the as referred to the study of British naval dockyard sociation with amphiboles is less strong than for peri workers exposed to amphiboles and chrysotile in toneal mesothelioma and such tumour; have been whom, although no dose-response relationship has caused by exposure to chrysotile alone, all the recent been published, a deficiency of deaths from lung publications that we have been able to find support cancer was observed (Rossiter and Coles, 1980). our previous view that crocidolite is substantially more dangerous than chrysotile. A single relevant 43 Jn the studies of female gas-mask makers exposed study published subsequently (Acheson el at, 1981), respectively to crocidolite and chrysotile already and updating of previously published results (Seidman referred to (Acheson et al, 1982), the former el at, 1979; Selikoff el al, 1979), support the view experienced a significant excess of lung cancer mor that, in process work and among insulators, exposure tality (13 observed, 6.2 expected) while the latter did to commercial amosite may confer as great a risk as not (6 observed, 4.8 expected). Although these women crocidolite, or at least a risk intermediate between were comparable in terms of lapsed period, the those of chrysotile and crocidolite. duration of exposure may have been longer among those exposed to crocidolite and no measurements of dust levels are available from either factory. I.ung cancer 47 Where data are available which make possible a direct comparison within a single study (Weill el al, 44 As far as amosite is concerned an estimate of 1979) they support our previous view, that crocidolite (c 35 fibres/ml for the average dust level in the insu- and amosite at a given dust exposure level may be I lation materials factory in New Jersey studied by more dangerous than chrysotile. However, when com Selikoff and his colleagues has recently been published parisons are made between the slopes of the dose- (Nicholson, 1981). Using this figure Nicholson calcu- response curves of studies of various populations of ' lated that the mortality from lung cancer per fibre- asbestos workers, no clear distinction can be found year/ml may be substantially higher than for any between the experience of those exposed to chrysotile group of workers exposed to chrysotile or to mixtures and those exposed to mixtures (see Table 6). On the of chrysotile with amphiboles. It is worth noting that basis of the evidence now available it cannot be the dust estimate was derived during 1967, 1970 and assumed that dose for dose chrysotile is in all circum 1971, not from the factory in question, but from two stances necessarily safer than mixtures with amphi other factories operating the same process at a later boles in respect of lung cancer. date. Although Nicholson states that these plants "manufactured the same products from the same fibre type (amosite) to the same specifications and Asbestosis 48 No recent relevant data have been found. using the same type of machinery", it would seem un justified to use uncritically their dust measurements for another factory which was in operation over 20 years previously during wartime conditions. Other reservations about this study are made later (para 76). 49 In our previous report we concluded the relevant section with a proviso based on the animal work which we regard as sufficiently important to repro duce in full here. The early experience of workers at a London factory 50 "In view of the fact that it seems clear from where commercial amosite was used reveals a two-fold animal work that all fibre types have the potential to increase in mortality from lung cancer (57 observed, produce the three main types of asbestos-related 29.1 expected), but no dust measurements are avail disease to a similar extent if the physical conditions of able prior to 1969 (Achcson el al, in press). the dust cloud are appropriate, it follows that any change in industrial practice in the direction of the Asbestosis 45 In our previous report we commented on the paucity of data on the relationship of the frequency of asbestosis to fibre type for jobs of comparable dustiness. So far as we are aware nothing relevant to this issue has subsequently been published. production of more finely dispersed fibre may be sig nificant to health. This work suggests that, should commercial chrysotile be milled in such a way that it approaches the respirability of crocidolite and amo site, particularly if long (10--80p) and extremely thin 2.sf/fibres are produced, the morbidity gradient in favour of chrysotile seen up to the present and dis- q C cussed in this chapter may disappear in the future. In 0947118.013 Summary considering the possible future significance to health of such a trend it is necessary to balance the very Mesothelioma much greater consumption of chrysotile than of the 46 Peritoneal mesothelioma has an almost exclusive relationship with exposure to the amphiboles crocido- Wc are aware of one cue reported in association with chrysotile (Dement et al, J982; McDonald et al, unpublished a). a DU 038407 .'liisiaahm*. amphiboles (Table 1, Fig 2) against the considerable improvements which have undoubtedly been secured in dust control in many areas in recent years. Changes in the range of usage of asbestos and any effects these would hove on the accessibility of the fibre to human contact would also have to be taken into account" (para 168). 51 In the light of subsequent animal work, fibres of length < 10p with high aspect ratio may also be relevant (Bertrand and Pezerat, 1980). 52 The differences in the risk of mesothelioma that we attribute to the three inincralogical types of asbes tos do not necessarily reflect chemical differences. They may well reflect differences in the distribution of fibres of various configurations in the dust cloud not detected by current methods of measurement. 53 Walton (1982) has deduced from the data of Hwang and Gibbs (1981) that people who have worked with crocidolite have probably been exposed to substantially higher concentrations of fibres in the carcinogenic range of particle size than those exposed to other varieties of asbestos and that most of these fibres are not visible to the optical microscope. Dose-response relationships 54 We reviewed the main sources of the considerable uncertainty relating to historical dust measurements upon which dose-response relationships depend, stressed the dangers in converting particle counts to fibre counts (all British as welt as all North American data prior to about 1964 were recorded in particles), and noted that the introduction of the 'eye-piece graticule' and persona! as opposed to static sampling may have resulted in a DE FACTO tightening of the hygiene standard of 2 fibres/ml since its introduction in 1969. 55 With regard to asbestosis we drew attention to the special difficulties of definition in calibrating response to exposure in this condition. Although the data did not agree on this point we concluded that there was no evidence of a safe threshold in respect to chrysotiie (except for certified asbestosis), and that the data from the only available longitudinal study were compatible with a linear relationship between incidence and dose. No data were available to study the relationship of dose to the frequency of asbestosis for crocidolite or amosite. We concluded that the fibrogenic effect of the fibre used at Rochdale had been underestimated, but this had been mitigated at least in part by the fact that the hygiene standard is currently enforced with more sensitive instruments and from persona! samples taken in the breathing space of workers. 56 For LUNG CANCER, quantitative data from which the shape of a dose-response relationship could be determined were limited to Quebec chrysotiie miners and millers and retired New Jersey asbestos workers. In both sets of data we found no evidence for a threshold of dose of dust below which there was no increment in risk, and format statistical tests showed that the data were suitably described by straight lines. 57 For AMOSITE and CROCIDOLITE no quanti tative data from which the shape of the dose-response relationship could be deduced were available. 58 For MESOTHELIOMA, although quantitative dust measurements are limited to one group of workers, these and semi-quantitative data strongly suggest that the risk of these tumours increases with increasing dose (paras 169--214). Up-data Asbestosis 59 The British Occupational Hygiene Society (BOHS) Committee on Asbestos has recently pub lished a further report on morbidity (BOHS, 1983). It contains data on a reconstituted study population all of whom started work in a factory after 1950, and so have been employed during the years for which dust measurements, which started in 1951, are available. The cohort included 295 men with medical records, who had no known exposure to asbestos dust prior to employment in the factory, and who had completed 10 years' (not necessarily continuous) work with as bestos in the factory by 31 December, 1976. Efforts were made to trace 130 of the 295 men who had left the factory by the end of 1976 (excluding 28 known to have died) so that they could return for an X-ray and clinical examination. 60 Radiological and clinical data were used to estab lish the presence (or absence) of seven "adverse effects" according to the criteria shown in Table 5. These included the occurrence of parenchymal and pleural abnormalities, changes in lung function, and the appearance of "chest sounds" (including crepi tations). By far the most frequent adverse effect was a value of FEV,/FVC <0.70, which was found in 111 men, while the frequencies of the other adverse effects ranged from 22 to 34. Without considering dust exposure level, 163 (55*70) of the 295 men had at least one of the adverse effects, and 52 (18%) had an effect other than FEV,/FVC <0.70. Estimates of the probabilities of each (or any one) of these adverse effects occurring in relation to the cumulative dose of fibre were made. For a cumulative dose of 50 fibreyears/ml it was estimated that about 7% of persons may be expected to have at least one effect, with DU 038408 DUP 0947118.014 about 27o for effects B, E and F, under 2Vo for effect G (chest sounds including crepitations), and under Wt for effects A and D (no data are given on effect C). In this study the dust levels are presented in terms of modern instruments, after a very detailed comparison of the different measurement techniques used between 1951 and 1976 had been carried out, but related to static, not personal, sampling. 61 In our previous review we dealt in some detail with the differences between the two previously re ported morbidity studies (BOHS, 1968; Berry et at, 1979). In the new study more men were included who had been exposed to lower concentrations of dust. On the other hand the maximum duration of follow-up is at present less than 30 years since first employment. From the most recent of the two earlier studies, in which men who had left employment were also in cluded, the estimated proportion with crepitations by 25 fibre-years/ml (equivalent to 50 fibre-ycars/ml by modern counting methods with static sampling) was almost negligible. In the present study the estimated comparable figure (for adverse effect G) was under 2ro. If one takes the view that personal, rather than static, sampling doubles the measured exposure levels, then enforcing a control limit of I fibre/ml with personal sampling devices might result in less than I Vo of persons developing adverse effect G, and less than 2Vo of persons developing at least one of the seven effects over 50 years' employment. These proportions have been estimated from Figs 10 and 11 of the latest BOHS report (1983). 62 The authors of this study point out that the adverse effects are not necessarily attributable to as bestos. In particular, the inclusion of FEV,/FVC <0.70 as an index of adverse effect is debatable, and the authors themselves state that "it did not correlate w-ith dust exposure and this bears out the latest thinking that FEV,/FVC <0.70 is an in appropriate criterion of an adverse effect of asbestos. It is well known, for example, that smokers develop a reduced FEV ,/FVC and there were many smokers in the populations studied." To be certain of the extent to which these changes arc independent of age, smok ing and other confounding factors it would be necess ary to make comparisons with a control group of men not exposed to asbestos. An extended discussion of these adverse effects is given in Appendix C. 63 It is noted that the composition of the cohort for the new BOHS study (BOHS 1983) is similar in definition and size to the group investigated by Morris (sec Appendix A). Lung cancer 65 Since our previous report was published a large number of papers have appeared which include esti mates of individual dust exposure levels. We have also traced many studies among different groups of workers where linear dose-response relationships pro vide a satisfactory description of the association between mortality from lung cancer and cumulative dose of asbestos, compared to the three available in 1978. The concept of a linear relationship now seems to be generally accepted by most of the authors in de scribing their own data, and in the various summaries that have been made, although not without reser vations (Peto, 1978, 1979; Acheson and Gardner, 1979a; McDonald er of, 1980a; Enterline, 1981; Nicholson, 1981; Sclikoff, 1981; Dement et at, 1982; Liddell, 1982; McDonald et al, unpublished a and b). 66 Quantitative studies of the relationship of dust exposure to mortality from lung cancer have now been reported in miners and millers of chrysotile as bestos, workers in the production industry manu facturing textiles, friction materials and cement, and maintenance workers. In more detail, they are: 1 Quebec chrysotile miners and millers (McDonald et at, 1980b), 2 Thetford Mines chrysotile miners and millers (Nicholson et at, 1979), 3 New Jersey factory workers manufacturing pro ducts from chrysotile (Henderson and Enterline, 1979), 4 New Jersey factory workers manufacturing pro ducts from chrysotile (Nicholson et at, 1981), 5 New Jersey maintenance workers exposed to chrysotile, amosite and crocidolite (Henderson and Enterline, 1979), 6 UK factory workers manufacturing textiles from chrysotile and some crocidolite (Peto, 1980a), 7 South Carolina factory workers manufacturing textiles from chrysotile (Dement et a!, 1982), 8 South Carolina factory workers manufacturing textiles from chrysotile (McDonald et at, un published a), 9 UK factory workers manufacturing friction materials containing chrysotile and some crocidolitc (Berry and Ncwhouse, 1983), 10 New Orleans factory workers manufacturing as bestos cement products containing chrysotile and some crocidolite (Weill et al, 1979), 11 Pennsylvania factory workers manufacturing tex tiles using chrysotile, amosite and crocidolite (McDonald et al, unpublished b), 12 Canadian factory workers manufacturing asbestos cement containing amosite and chrysotile (Finklestein, 1982). 64 In a study of the mortality of workers certified by pneumoconiosis medical panels as having asbestosis, Berry (198!) has shown that the excess lung cancer rate and the mesothelioma and asbestosis rates all increased with percentage disability. no 67 Study 2 (which was not based on individual ex posures) is an investigation of a sub-group of study 1, with study 4 (again not based on individual exposures) being similarly related to study 3, and study 7 to study 8. DUP 0947118.015 DU 038409 68 The twelve studies mentioned above have been based on estimations of the cumulative exposures of individual workers (except for studies 2 and 4) either for the complete cohort or on a case-control basis. Some of the difficulties inherent in the available dust measurements and in this index as a measure of bio logical dose, in addition to the problems of appro priately measuring the response (mortality), were dis cussed in our previous report. These arc repeated in Appendix D, together with some further points which apply to some of the more recent studies, and should continue.Bo be borne in mind. The cumulative dose, however, is still the method of presentation of results with regard to dust levels used in the literature, and so our arguments are in these terms. However, because of different modes of accumulation in different studies, sometimes including exposures to age 45, or during the first 20 years of employment, or to death etc, comparison between results is further compli cated. 69 In addition, a number of other relationships have appeared which have incorporated estimated ex posures of groups of workers in a similar way to studies 2 and 4, using either an average dust level together with duration of work or assigning numerical dust values to previously used qualitative assessments (for example, "low to moderate", "severe") of degree of exposure. Studies treated in this way have been: 13 UK factory workers manufacturing a number of products containing chrysotile, crocidolite and amosite (New house and Berry, 1979), 14 US factory workers manufacturing insulation products containing amosite (Seidman et al, 1979), 15 US insulation workers exposed to amosite and chrysotile (Selikoff el al, 1979). 70 We used earlier published results from studies 1, 3 and 5 in our previous review. Study 6 was also in cluded, but only an overall SMR from lung cancer and an average cumulative dose were available, with linearity being assumed. fibre-years/ml of asbestos. By contrast those given by Liddell are usually from straight lines of the form: RR = 1 + D, where RR = the Relative Risk compared to un exposed persons. 72 The essential difference between these two approaches, and vital to their interpretation, is that in the former the line is constrained to pass through an SMR of 100 at zero exposure (D = 0). For each of the studies where SMRs were calculated based on an external standard, this constraint is tantamount to assuming that the chosen standard population was appropriate. In other words it is assumed that the standard population and the asbestos workers had a similar experience in respect of all factors other than exposure to asbestos (including smoking) with a bear ing on the incidence of lung cancer. It is unlikely that this will have been true in any particular study, and the relative risk approach allows that the SMR at zero exposure may be different from 100, say SMR,. The relative risk at any exposure level D is thus obtained as SMRp/SMR,, so that if there is a linear relation ship with dose of slope b', then: SMRd = SMR, + b' x D This leads to: RRp = 1 + Xjqq D, where b* = lOOb'/SMR, is similarly scaled for com parison purposes to the slope b above. Before each of our earlier slope estimates (Acheson and Gardner, 1979a) for studies 1, 3 and 5, we carried out a formal statistical test (using a weighted regression analysis) of the intercept on the SMR axis and found no reason to suggest that 100 was inappropriate. In his review Liddell (1982) found that 100 was inappropriate for several of the recently published dose-response relationships. For studies analysed on a case-control (case-rcferent) basis the slope of the relationship between relative risk and cumulative dose is deter mined more directly, without the involvement of an external standard population. 71 For some of the studies mentioned above the slopes of the dosc-rcsponse relationships have been estimated and tabulated by Nicholson (1981) and repeated by Selikoff (1981) and Liddell (1982). We have collated their figures into Table 6, and have given our previously estimated slopes together with those of the original authors if provided. It is import ant to note that the slopes given by Nicholson and our own previous values are from straight lines of the form: SMR = 100 + b x D, where SMR = the Standardised Mortality Ratio, b = the slope of the line, and D = the cumulative dose in 73 The slope estimates shown in Table 6 have size able sampling errors attached to them; for example, Liddell (1982) quotes a 90^o confidence interval of from 0.016 to 0.069 for the value of 0.038 in chryso tile miners (although this is based on the largest number of lung cancer deaths in all the studies) and this should be borne in mind. A possible reason why Liddell's estimated slopes arc much lower than those of Nicholson (c.g. studies 7, 13 and 14) is that the populations used for comparison had a more favour able experience than the workforce under consider ation in respect of other carcinogens with a bearing on the incidence of lung cancer, e.g. smoking habits. If this were so the effect would be to raise the calcu li DUP 0947118.016 DU 038410 ltd SMR at zero exposure and reduce the gradient 77 The interpretation of tne siu,,.. ,, f the slope (Figs 2 and 3). We prefer the relative risk response relationship in the South Carolina textile fac pproach because it does not involve any assumption tory is of particular importance because it is by far bout the appropriateness of the standard population the steepest attributed to the effects of chrysotile, and ind also describes the empirical data more closely. chrysotile is now for practical purposes the only type of asbestos fibre imported into the UK (Table 1). In 74 The studies shown in Table 6 have been listed so "both the first study from this factory, of a limited that those where exposure was only, or principally, to cohort of white males employed for at least six chrysotile are given first, and those where amphiboles months (Dement et al, i982)*, and in a second study were used arc given later. It is clear, whichever of the of the whole workforce (McDonald et al, unpublished sets of estimated-'lopes by cither Nicholson or Liddell a) steep slopes were reported by Liddell (1982). From are taken, that there are considerable differences in Fig 2 of McDonald et al (unpublished a), slopes of 2.3 the"slopes of the lines found in the various studies. and 1.7 respectively for the relative risks from the two The variation in Nicholson's (0.06 to 9.1) is greater studies can be derived by applying a conver'sion factor than in those calculated by Liddell (0.038 to 2.7). of 3 for million particles per cubic foot to fibres per Without exception wherever estimates of slopes were ml. Since Liddell estimates an SMR of 161 at zero ex made by Liddell they were lower than those of posure from the data of Dement et al (1982), the Nicholson. Among the studies of persons exposed to slope of the relationship between the SMR and cumu chrysotile only, there are low estimated slopes for lative dose (in fibre-years/ml) is thus 1.61 x 2.3 = 3.7. miners and millers in Quebec and steep slopes for tex This is the value we have used in Fig 3, and it is tile workers in South Carolina. Men who entered the similar to that (of 4.0) derived from Fig 2 of Dement Rochdale textile factory after 1951 and were exposed et al (1982) (see Table 6). to asbestos for at least 10 years (study 6L), women employed at Barking (study 13F) and men employed during and immediately after the second world war in the manufacture of amosite-containing insulation pro ducts (study 14) all experienced a steep relationship between dose of fibre and lung cancer mortality. 78 In their lowest exposure group at dust levels up to 27.4 fibres-year/ml, Dement and his co-workers found an SMR of 223. A plot of their data (see Fig 3) suggests that the use of the US national rate as a standard may have been inappropriate and that the expected numbers of deaths are therefore too low. 75 With regard to the Rochdale results, the finding However, it is only fair to state that the authors them of higher mortality in more recent employees (Peto, selves do not agree. If they had based their expected 1980a) has nov, lessened, and was probably due to the numbers on local mortality rates the SMRs for lung small number of deaths at that time (Peto J, personal cancer would have been much lower than the pub communication). The figures at Barking are based on lished figures, because the county in which the factory estimated group exposures (5 to 10 fibres/m! for `Mow was situated had a remarkably high mortality from to moderate" exposed groups and 20 fibres/ml or lung cancer in men; 75% above the national average. l higher before 1945 for the groups with "severe" Although Dement and his colleagues (1982) attribute exposure; Nicholson, 1981) in a situation where no this to the presence of a large number of ex-shipyard individual dust measurements were available, and the workers who had worked in wartime naval con number of employed women was small. It must be struction, the Atlas of Cancer Mortality for US seriously questioned as to whether or not these Counties (Mason et al, 1975) shows that the lung estimated figures have any real meaning. cancer mortality rates in women are also significantly high in this county, a finding unlikely to be due to 76 With respect to the exceptionally high slope in amosite factory workers reported by Nicholson (1981), at least two points are relevant. The first, relating to the relevance of the estimated dust levels, has already been pointed out (para 44). Secondly, the SMRs from lung cancer for men employed for very short durations of time (up to three months) are markedly raised (between 275 and 300) as we discussed in our earlier report (sec present Fig 2; and Fig 13 in Achcson and Gardner, 1979a). Nicholson apparently shipyard work. Even though women were employed in the shipyards, the numbers involved were small (Blot et at, 1979). Smoking docs not seem to be responsible, as the smoking habits of members of the workforce were similar to those of the adult population of the US. There remains the possibility that the population of this county have been exposed to an unknown car cinogen which explains in part the excess mortality from lung cancer experienced by the workforce when compared with the US population as a whole. takes the view that the external standard is nonethe less appropriate and fils a line with an SMR of 100 at 79 If one regards the standard population as suitable zero exposure (see Fig 2); Liddell has doubts about and forces the line through the origin (as does the appropriateness of the death rates of the general Nicholson) the slope becomes much steeper (5.3 con- q population of the State of New Jersey as a standard trasted with 2.3). Even if one were to reject the value C for these factory workers, and consequently reports a slope for the relative risk which is much less steep (see A further study of men employed in this factory for at least one month which docs not alter the conclusions has recently been Fig 2). published (Dement ft of, 1983). 0 9 4 r7 78 _ "Hu \7 038411 of 5.3, the lower figure is still appreciably higher than for other chrysotile (or predominantly chrysotile) applications, including mining and milling (studies I and 2), manufacture of textiles (study 6) and friction materials (study 9) in the UK, asbestos cement pro duction (study 10) and the manufacture of assorted products (studies 3 and 4). 80 In all the studies discussed in this section a major consideration to bear in mind is the validity of the ex posure data. Tliesc arc based almost completely on measurements made in terms of particles of dust collected fiom static sampling points (and, prior to about 1965, not on fibre counts) and it has been necessary to convert concentrations originally in millions of particles per cubic foot to ftbres/ml in order to make the data relate to modern hygiene stan dards (which arc based on fibrcs/ml). Both the use of data from non-fibre-spccific techniques based on out moded methods which differed in different countries, and the incorporation of conversion factors which are at the best arguable introduce major uncertainties, especially when similar assumptions arc applied to such different environments as mining and textile manufacture. A conversion factor of 3 fibres/ml to 1 million particles/cubic foot, a value in the middle of the range considered in our previous report, has been used for chrysotile miners and millers by McDonald et a! (1980a), for textile workers (except for the prep aration area where a conversion factor of 8 was used) by Dement et al (1982) and by Nicholson (1981). McDonald et al (unpublished a) consider an average figure of 6 to be more appropriate for the South Carolina textile plant, and quote a range of from 1.3 to 10.0 for different areas of the plant. On the basis of the high mortality from asbestosis in this plant, Weill (1982) has challenged the validity of the dust measurements in the South Carolina textile factory and has suggested that the early dust levels may have been underestimated. This is also the view of Steel (1983), whose comments are reproduced as Appendix B to this report. Cigarette smoking and asbestos 81 The evidence on the synergism between asbestos and cigarette smoking in lung cancer which has recently become available suggests that this may be intermediate between an additive and multiplicative effect, with a smaller than multiplicative interaction term (Berry, 1980; Sclikoff et at, 19S0; Saracci, 1981; Acheson et al, in press). An interaction of this form has a modifying influence on the slope of a linear dose-response relationship, namely, to increase the observed slope above that which would be found in the absence of an interaction (or in the absence of cigarette smoking). Moreover, if the smoking habits of the industrial cohort are greater than in the stan dard comparison population the effect will be similar and will exaggerate the slope of the dose-response curve (Achcson and Gardner, 1981fGardncr, 1983b). For these reasons a further note of caution should be struck in relation to many of the dose-response curves in Table 6, as smoking habits were studied in few of them. 82 The likely reduction in the incidence of lung cancer among asbestos exposed persons who cease smoking cigarettes is discussed by Sclikoff et al (1980) and Saracci (1981), and the latter suggests that over 90% of the excess lung cancer deaths may thus be prevented. Mesothelioma 83 As far as mesothelioma is concerned one of the important findings has been that the incidence appears to be related to between the third and fourth power of the lapsed time since beginning exposure (Peto et al, 1982). In specific terms: 1, = b.tk where 1, = incidence t years after initial exposure, k = a numerical constant between 3 and 4, and b is another constant which is possibly dependent upon a number of factors including fibre type, duration and intensity of exposure, age and site. Peto and his co workers have shown that for five different cohorts of asbestos workers, where the data had been published in a form suitable for analysis, a value of k = 3.2 gave a satisfactory fit in each case. 84 Apart from McDonald et aPs data quoted in our previous report (Table 3IX) no quantitative infor mation is available about the relationship of the inci dence of mesothelioma to dust exposure. However, subsequently published semi-quantitative data support the conclusion that the risk increases with increasing intensity and duration of exposure (Hobbs et al, 1980). At present nothing is known about the risk of mesothelioma as compared to lung cancer at a given level of exposure to asbestos. Implications for the industrial control limit Crocidotite Amosit* 85 H'e concluded that contact between man and crocidolite should be limited to the minimum practi cable, that its importation should be prohibited, and that stringent regulations should continue to be applied to protect workers and the public in respect of crocidolile-containing materials. 86 We concluded that although strictly quantitative data were lacking, the evidence suggested strongly that this material had been more dangerous than chryso tile. Therefore a stricter control limit might be appro priate. 13 DUP 0947118.018 DU 038412 Chrysotile Asbestosis 87 We compared two studies from Rochdale and concluded that, if modern techniques used to enforce the standard are two to five times more sensitive than those used to arrive at measurements in the 1950s and 1960s the evidence for a cumulative incidence of 1% asbestosis occurring after a life-time's exposure at 2 fibres/ml (us curienlly enforced) fell away. I.ung cincer 88 Bising our calculations on a linear dose/response relationship without threshold and using various con\ersion factors for particles/fibres and old/new . measurement techniques we produced an array of possible control limits at which a particular excess mortality of asbestos-related disease might occur. We concluded that, for example, an excess mortality from asbestos-related disease of 2% might be associated with any point in a range offrom 5 fibres/ml to 0.4 fibres/ml, and that bearing in mind the very con siderable uncertainties a figure towards the lower end of the array might represent an appropriate compro mise (paras 242--259). Up-data Crocidolite 89 N'o subsequent data have been published which suggest we should modify our conclusion. Amosite 90 Subsequent data strengthen our conclusion that amosite is probably more dangerous than chrysotile (see paras 40 and 44). In the light of the gross re duction in imports resulting from its replacement in the production of insulation board, formal prohib ition of new products might be considered. 91 Since the amphiboles are of less relevance in terms of imports and manufactured products now than they were at the time of our earlier report, our discussion of the control limit is focused on chryso tile. This docs not preclude the need for careful restrictions on the handling and uses of materials con taining crocidolile or amosite, including their mainten ance, repair or demolition, for which provision already exists. Chrysotile 92 Our earlier recommendations for chrysotile were based on the following assumptions: (a) that there is a linear relationship without threshold between cumulative exposure to chryso tile and lung cancer mortality; (b) that in populations exposed to chrysotile the numbers of excess deaths from lung cancer and from asbestosis have been similar; (c) that mesothelioma deaths were rare from exposure to chrysotile alone; 14 (d) that, from the latest morbidity data related to cumulative exposure front Rochdale then avail able, the evidence (after taking into consideration personal sampling with modern instruments) in favour of a material prevalence of asbestosis occurring after a lifetime's work at the control limit as enforced in 1979 was slight. 93 Nothing has been published -since our report which leads us to wish to vary our first three assump tions. In relation to the second assumption McDonald el al (unpublished a) report 21 dealhs from "pneumo coniosis" and 29.4 excess dealhs from lung cancer in the South Caroline chrysotile textile factory. 94 We wish to point out that this and other similar figures relate to past exposures to asbestos dust which were higher than those likely to be encountered today or in the future. As asbestosis is particularly associ ated with high exposure levels our second assumption is conservative and incorporates a safety factor into a control limit based on this approach. 95 The final assumption merits further discussion in view of the recent BOHS survey of morbidity in relation to dust measurements (BOHS, 1983). The probability of at least one of seven adverse effects developing in an individual was estimated to be 7% after the equivalent 50 years' exposure to the control limit of I fibre/ml, when static sampling was used. If it were to be accepted, as was not the case by the authors of the latest BOHS report, that measurements based on personal samplers are on average twice as high as those based on static samplers (Acheson and Gardner, 1979a; Steel, 1979), the estimated prevalence of at least one of seven adverse effects would be lowered to less than 2%. In cither case a view has to be taken on which of the seven adverse effects are indicators of serious morbidity and also to what extent they are asbestos-related. We tend to take the position that, although the present estimates of adverse effects at any level of cumulative exposure appear to be higher than in the previous BOHS studies, the extent to which some of these effects, notably the commonest, F, are related to exposure to asbestos is questionable. Revised estimates for lung cancer 96 The major difficulties concerning the use and interpretation of past exposure levels in the con struction of dose-response relationships, which we re viewed in our previous report and elsewhere (Acheson and Gardner, 1980b) and have referred to in para 68 and Appendix D of this paper, remain. We previously look a range of conversion factors (I, 2 and 5) from million particles per cubic foot to fibres per ml, and recent papers have follow-ed an average figure of 3 suggested as appropriate to chrysotile miners and millers (McDonald ct al, 1980a), although there was much variation of this factor between sites, and textile workers in South Carolina (Dement et al, 1982). The introduction of the eyepiece graticule method of counting around 1974 had the effect that more fibres DUP 0947118.019 DU 038413 v.ere counted per area of slide by t factor of about 2, and so implied that past counts should be approxi mately doubled to compare them with results from techniques used to enforce the control limit today (Steel, 1979). The determination of a suitable con version factor between personal and static sampling is undoubtedly complicated by the fact that some indi vidual exposures are lower when measured by per sonal sampling, but they are in the minority (Steel, 1979; BOHS, 1983). In general, it appears that con centrations obtained from personal samplers are highertftan for static samplers, and that a factor of 2 may be an accepted compromise. An exception to this comes from the results of a study at Rochdale in 1971, where dust levels from one set of personal samplers were found to be lower than average levels from static samplers collected over a period of one year for nearly all processes (Smither and Lewinsohn, 1973)*. 97 As far as lung cancer is concerned, recent publi cations imply that the range of excess mortality associated with exposure to chrysotile is greater than had previously been reported. Thus, studies from the South Carolina textile factory indicate a much steeper dose-response relationship than any earlier study, while at the other extreme a large British workforce making friction materials experienced no excess mor tality from lung cancer (149 observed, 151 expected) and there is doubt as to whether there is a positive dose-response relationship (see Table 6). A study of an asbestos cement factory in Wales also showed no excess mortality (30 observed, 33 expected), but no attempt was made in the analysis to look at the doseresponse relationship. More information is urgently needed about the occurrence of asbestos-related disease in the asbestos cement industry where chryso tile only has been used, particularly in view of the fact that this is the largest single application of chrysotile in the UK (Table 2). 98 Using the various slopes of the linear doseresponse relationships given in Table 6 it is possible, adopting the same approach as in our earlier report, to estimate fibre concentrations which after 50 years' exposure might be associated with varying levels of excess mortality from lung cancer. These are shown in Table 7 for various proportions (relative to all deaths) of extra deaths due to lung cancer. The lowest fibre levels associated with a given level of mortality are given by the textile factory in South Carolina. Ac cording to Nicholson's estimate of the slope, 1% extra deaths from cancer of the lung would occur after a working lifetime's exposure to a dust level of 0.04 fibres/ml (one twenty-fifth of the current UK control limit). A similar proportion of extra lung cancer deaths would result from a rather higher dust level (0.09 fibrcs/ml) if Liddell's method (which we prefer for the reasons stated in para 73) is followed. There are also thought to be problems with the manner in which (he personal samples were collected in (his study (Steel, J personal communication). If a conversion factor of 6, rather than 3, is used (as suggested for this factory by McDonald el al, unpub lished a) the relative risk method of calculation yields an associated dust level of 0.17 fibres/ml. Table 7 shows that the range of fibre concentrations that would produce an excess of lung cancer of l<7o is from 0.04 to 5.0 fibres/ml. The highest dust levels compatible with a given level of excess mortality are those derived from miners and millers and the manu facture of friction materials, the lowest from textile manufacture. 99 If it were agreed that the effect of the introduc tion of new counting techniques and personal sam pling instiuments had increased the measured fibre concentrations by a multiple of between 2 and 5 (Acheson and Gardner, 1979a), th.cn the figures shown in the body of Table 7 would be increased accordingly. Looked at another way, the excess deaths from lung cancer associated with exposure to 1 fibrc/ml for 50 years range from 0.2170 to 27To (see Table 24a in Advisory Committee on Asbestos, 1979), which would be reduced by a factor of between 2 and 5 to take account of modern counting and personal sampling. The weight given to the different figures depends upon various technical considerations relating to the calculations, such as the validity of the dust estimates and of the standard populations used in the comparisons which have already been discussed. 100 Other points that should be taken into account in the interpretation of these figures are that the con trol limit in practice is intended to be a maximum value and that the average exposure will be less, perhaps as little as one-tenth of the control limit (see Appendix M in Advisory Committee on Asbestos, 1979). Furthermore, with increasing mobility of labour it is becoming usual for the workforce to leave very much earlier than with 50 years' exposure, although they may move to other asbestos-using employment. There are also questions outside the scientific data, relating to policy, cost etc which need to be considered in setting a control limit (Duncan, 1981). 101 In view of the wide range of values displayed in Table 7 the question must arise whether it is rational to employ a single control limit for chrysotile throughout industry (McDonald, 1982). To employ a single control limit might either put certain categories of workers (for example, those in textiles) at risk if the limit were at one end of the spectrum, or if the control limit were at the other extreme place an un necessary and possibly insupportable burden of cost on others (for example, manufacturers of friction materials). A further complication is that for large sectors of the industry, such as asbestos cement and tiles (which together consume about half of all im ports of chrysotile), and for miscellaneous uses and for persons who utilise asbestos pioducts, no doseresponse data have been collected. Such data as are available suggest that the manufacture of asbestos cement products from chrysotile may have been as- 15 DUP 0947118.020 DU 038414 sociated with relatively little excess mortality, but more information is urgently needed on this point. 102 Epidemiological data which have been published since our last report suggest that the manufacture of textiles from chrysotilc may have been more danger ous than was previously recognised, and that it might therefore be prudent to recommend stricter control in this part of the industry than for the remainder. How ever, it should be borne in mind that uncertainties re main concerning both the validity of the historical dust measurements and the external mortality stan dard used in the relevant study (see paras 75 to 80). Textile production only accounts for about 47o of the British asbestos industry (Table 2). 103 Studies published from other parts of the in dustry do not suggest the need for a reduction in the control limit. Nevertheless, as was indicated in Advisory Committee on Asbestos (1979), in view of the fact that all forms of asbestos can be carcino genic, further reductions in the control limit should be made whenever advances in engineering make them reasonably practicable, and the use of all types of as bestos should be curtailed as safe and effective substitutes become available. - 104 Public policy should take-account of the fact that all types of asbestos are extremely durable and that products containing them may require further processing, including servicing and demolition, in cir cumstances where dust control may be difficult and far removed in place and timefrom the original pro duction process. Conclusions 105 In the light of our terms of reference (paras 1 and 2) we conclude as follows: (a) we have not found any material which was avail able to the Advisory Committee on Asbestos which on reconsideration would have altered the conclusions in our previous report; (b) data which have become available subsequently lead us to alter the conclusion of our previous report in so far as they relate to the control limit in the workplace as follows: (i) amosite imports have effectively ceased and chrysotile is at present for practical purposes the only type of raw asbestos fibre imported into the UK; (ii) the evidence that asbestos fibre causes alimen tary tract cancer in man is less convincing than in 1979; (iii) the case that amosite is more dangerous than chrysotile has strengthened in respect of both peri toneal and pleural mesothelioma. We recommend that formal prohibition of the manufacture and importation of new products made of amosite and crocidolite should be considered; (iv) the range in the slopes of the dose-response relationships for lung cancer and chrysotile ex posure has widened since our report of 1979. At one extreme the risk associated with the manufac ture of textiles in South Carolina may have been greater than has been previously reported in relation to any other chrysotile application, while at the other extreme the presence of any increased risk with increasing exposure in a large cohort of men who manufactured brake linings is question able. 106 We consider that subsequent evidence has sup ported the view expressed in 1979 that in respect of lung cancer the relationship of exposure to mortality is linear. 107 Subsequent evidence has supported our previous view that peritoneal mesothelioma for practical pur poses never, and pleural mesothelioma rarely, has occurred in man in relation to exposure to chrysotile alone. 108 In view of the fact that all forms of asbestos can be carcinogenic, further improvements in control should be made as advances in engineering make them reasonably practicable and the use of all types of asbestos should be curtailed as safer and effective substitutes become available. 109 Steps should be taken to ensure that measure ments of exposure to asbestos in the workplace are made and recorded in such a way that it is possible to determine what benefits (if any) have occurred as a result of improvements in control of dust. 110 We draw attention to the fact that little epidemi ological information is available about the effects of work in the asbestos cement industry where chrysotile only has been used. As this is now the largest single section of the asbestos industry in the UK, further re search should be carried out at the earliest opportunity. Appendix A Review of report by Or J G Morris 111 The report by Morris was dated 24 February 1978 and was circulated to members of the Medical Working Croup of the Advisory Committee on Asbes tos in April of that year. The papers comprised a 1R brief description of a further study of the incidence of disease in a group of men who had been exposed to asbestos dust at TBA (Rochdale), carried out by the Medical and Health Physics Department of the fac tory in 1977; an appendix (Appendix I) describing methods of dust measurement at TBA by Dr A L DUP 0947118.021 DU 038415 :;.KMsiaiu-i,,,iBBiBfiiJTj, I iftijiia Rickards; an appendix (Appendix II) describing criteria of adverse effects; some graphs; a specimen proforma for recording clinical data; and an enciphered summary of the clinical and dust infor mation for the 286 men in the cohort. 112 The cohort consisted of men who had been first employed at TBA Rochdale after 1 January 1951 and who had completed 10 years of service by 31 December 1976. Men with known industrial exposure to asbestos outside TBA cither pre- or post-1951 were excluded from the study. According to Morris the study was undertaken "because of the pressing need to know whether the present 2 fibres-ycar/cm' was in fact a safe standard or not in view of public, govern mental, union and management anxiety". Many of the men in the cohort had also been included in a study carried out on behalf of BOHS by Berry et aI (1979), which was reviewed in detail by the present authors in their 1979 report. Berry et a/'s data related to 379 men employed in the same factory who had worked for at least 10 years between 1 January 1933 and 31 December 1972. Of these 379 men, 197 were first employed after 1951 and will presumably have been included in the 286 men studied by Morris. 113 Morris's data differ from those of Berry et a! in the definition of adv erse effects and in a number of other particulars, notably that re-examination of as many of the men as possible was undertaken and a further attempt was made to date the occurrence of the first adverse effect by review of clinical records and X-rays. The dust measurements which Morris uses are those prepared by Rickards. For the period 1951--74 they are based on static sampling, sub sequently on personal sampling. Rickards has con verted the data "to modern membrane filler values expressed as fibres per ml". 114 Morris's results show during the period 1976--77 that 66 (23^o) of the 286 persons were identified as having an adverse effect defined as one or more of the following criteria; crepitations, radiological changes, lung function-tests (including gas transfer and K factor), clubbing, dyspnoea and other symptoms. An analysis of cumulative dust level against years of exposure suggested that the 66 men with adverse effects had experienced dust exposures at a rate greater than 2 fibre-years per ml. 115 In his report Morris offered to provide further information on the men he had studied. This offer was not taken up because it was decided that, in view of the decline in dust levels in industry, cancer mor tality was likely in future to prove a more sensitive indicator of risk of asbestos-related disease than asbestosis and because there was a substantial degree of overlap between Morris's data and Berry's post-1951 material which had been dealt with in detail. 116 The recent BOHS Report (1983) reviewed in paras 59 to 63 describes an almost identical cohort in terms of similar criteria of adverse effect. The authors (EDA and MJG) of this paper visited Dr Morris in February 1983 to disucss the progress of his study. It remains in unpublished form, and it was agreed then that because of the substantial overlap with the BOHS (1983) report the findings would be similar. Appendix B Note by Dr J Steel on dust estimates in the South Carolina chrysotile textile factory 117 If we examine the sampling records reported by Dement cl at (19S2), it appears that from 1930 to 1945 about 12 atmospheric samples were obtained on average each year for the whole plant. As they were required for insurance purposes or for use by govern mental health agencies, it is likely that the samples were collected annually by these organisations and the possibility exists that estimates of dust exposure within large areas of the plant arc based on single samples taken once a year. If unrepresentative sam pling locations were originally chosen, then large errors will have been introduced into the exposure estimates. Moreover, once the locations of the sam pling stations were fixed, they would be unlikely to change from year to year and the errors would be perpetuated. 118 From 1945 to 1960, roughly the same low mean number of samples was taken annually, but samples were collected almost entirely by company personnel. In the potentially dusty operation of preparation/ waste recovery, 1945 marks a watershed between high and much lower estimated mean exposure levels and while this change could be attributed to improved dust control, it might also be ascribed to a change in sam pling personnel. The people involved in sampling from 1945 to 1960 would initially have little experi ence and expertise, and their motives and bias would be completely opposite to those of the agencies sam pling during 1930 to 1945. 119 Relatively large numbers of air samples were taken over the period 1960 to 1975, but it is likely that most of these samples were obtained after 1971 when the membrane filter technique was used exclus ively. Thus, if the possible sampling errors described above tended to produce low concentration figures for the relatively few measurements taken during the period 1930 to 1970, then this would result in a systematic underestimation of exposure over the major period of the study with consequent effects on the slope of the exposure-response line for lung cancer. 17 DU 038416 DUP 0947118.022 -3iS:siaaiasigiiMBlai. = Hisr'iti-I. S-KJOStj... :;;iaaailiiayLJajMM{ Appendix C Note by Professor M Turner-Warwick on indices of adverse effects and asbestosis in the BOHS Committee on Asbestos study 120 Attempts to identify the earliest manifestations of asbestosis and thence their relationship to dust ex posure levels have been considered recently in a report by the BOHS Committee on Asbestos (BOHS, 1983) and in a report on a closely similar cohort by Dr Morris. The BOHS used seven radiological, clinical and lung function criteria as adverse effects (sec Table 5). These included five effects recorded as change between at least two measurements over the obser vation period; namely, the development or increase of parenchymal abnormalities, the development of pleura! chances and abnormal change in various lung function measurements. Two other adverse effects were recorded if they were observed on the initial or subsequent examination; namely an FEV,/FVC ratio of less than 0.70 or the presence of chest sounds (in cluding crepitations, rales or crackles). In the cohort studied chest radiographs were taken and clinical examination made regularly over the period between 1951 and 1970 but lung function studies were only introduced in 1967 (Factory A) and 1960 (Factory B). Thus, information on four of seven adverse effects were not available for the first 10 to 17 years of the study. By far the most frequent adverse effect in the BOHS report was Factor F (FEV,/FVC <0.70) which accounted for 40TO of all effect occurrences in Factor) A and 347o in Factory B. It is difficult to assess the relevance of this measurement as an adverse effect (a) in the absence of lung function tests during the fust 10 years of the study, (b) without infor mation on the numbers having a reduced ratio at the time of initial examination, and (c) absence of detail on smoking. 121 The authors of this study themselves point out that the adverse effects are not necessarily attributable to asbestos. In particular, the inclusion of FEV,/FVC <0.70 as an index of adverse effect is debatable and as the authors stated: "it did not corre late with dust exposure" and "this bears out the latest thinking that FEV,/FVC <0.70 is an inappropriate criteria of an adverse effect of asbestos. It is well known, for example, that smokers develop a reduced FEV,/FVC and there were many smokers in the population studied." To be certain of the extent to which these changes are independent of age and smoking and other confounding_factors, it would be necessary to make comparisons with a control group of men not exposed to asbestos or to study workers at the beginning as well as after exposure. 122 Without considering dust exposure levels, 52 (lS^o) of 295 men at Factory A developed only ad verse effects other than Factor F. Estimates of the probability of developing each or any one of these adverse effects in relation to the cumulative dose of fibre were made. For a cumulative dose of 50 fibreyears per ml it was estimated that about 2To of persons may be expected to develop effects B and E; under 2% factor G and under 1% for effects A and D. In this study, the dust levels are presented in terms of modern instruments after a very detailed compari son of the different measurement techniques used between 1951 and 1976 had been carried out. It related to static and not personal sampling. 123 It should be noted that in this report there is a good deal of variation demonstrated between the three readers of the chest radiographs. It should also be noted that in the absence of controls several assump tions have had to be made with regard to the physio logical data. It docs not appear from this report that there is any dear order of appearance of adverse effects in relation to asbestos exposure. 124 Overall, it is reasonable to conclude from these data that, with the improved hygiene standards now implemented, the occurrence of features that might reflect on adverse effect of asbestos on the lung (but which are also common in many other lung diseases) are likely to be found in only a very small number of workers after a lifetime of exposure. Appendix D Problems of measurement of dose and response 125 The difficulties of measurements in these two areas, and the interpretation of relationships between them, have been mentioned briefly in the body of this paper and were considered in more depth in our previous report (Achcson and Gardner, 1979a). A short summary is provided here. Measurement of dose 126 Ideally the measurement required is the dose of fibre reaching the target organ and remaining-there long enough to produce a deleterious effect. In man, IB however, the only information available, if any, con sists of estimates of dust in the workplace available to be inhaled. Methods of measuring dust and fibres have changed with time, and reconciliation of measurements made with different techniques is far from stiaightforward (Steel, 1979). Instruments have changed with the membrane filter sampler replacing the thermal precipitator, fibre counting replacing par ticle counting, the `eyepiece graticule' replacing the `whole field' method of counting, and personal sam plers attached to the workers replacing static samplers. DU 038417 DUP 0947118.023 a3SB^g^iaSia;3jai SJaS.ai.tliSssSs.,Sizzlm., :,,,.jamS! j^lS s& 127 Two other relevant aspects are that some pro portion of asbestos fibres taken into the body remain and may be active for prolonged periods, and that the disease may appear after exposure has ceased. These points can be taken account of when producing an index of dose by weighting the exposure by the esti mated residence time within the body, by making various assumptions about the rate of decay of activity of fibres in the body (Berry et al, 1979), and by making allowances for cases of disease occurring after workers have left the factory. Measurement of response 128 The comparison of observed mortality with that which would be expected on the basis of the death rates in an external standard population is the most common procedure for megsuring response. This en ables the study group to be assessed against a pre dominantly unexposed population for diseases of interest. The selection of an appropriate standard is often complex. Because of the extended time from first exposure to recognition of asbestos-related diseases a long period of follow-up is necessary to avoid misleading results. Appendix C has reviewed the problems in measuring morbidity. Assessment of dose-response relationships 129 In using estimates of risk based on any of the asailable studies, the inadequacies and problems should always be borne in mind. The relationships be tween dose and response vary greatly in slope between the studies reported in Table 6. These differences may be due, among other reasons, to: (a) exposures to different fibre types or mixtures; (b) different work conditions; (c) varying accuracies of dust measurements, arising from different methods and positions of dust sam pling, techniques of counting, use of measure ments from other locations or factories, and `guesstimates' made where no measurement data are available; (d) varying fibre: dust ratios; (c) differences in size distribution of the fibres; (0 different relationships between airborne fibre con centrations and the amount of fibre deposited in the lungs; (g) differing follow-up periods; (h) appropriateness or otherw ise of external standard population; (i) differences in background level of lung cancer in standard populations; (j) smoking habits of industrial population relative to standard population; (k) differences in smoking habits between men at varying levels of cumulative dose; (l) use of retired population compared with a full cohort. References Acheson E D, Bennett C, Gardner St J and Winur P D (19S!) Mesothelioma in a fac tory u>sng amosite and chrysolilc asbestos Lcr.cei 2 1403- 1405 Acheson E D and Gardner M J (1979a) The ill effects of asbestos on health Asbestos, Final Report of the Adusory Committee 2 London; HMSO Acheson E D and Gardner M J (1979b) Mesothelioma and exposure to mixtures of chrysolite and amphibnle asbestos Archives of Enxironmcntal Health 34 240 - 242 Acheson E D and Gardner M J (1980a) Po^ible syrerrhm between chr>votile and amplubolc asbestos J.amei 1 706 Acheson E D and Gardner M J (1980b) As bestos: scientific basis for environmental control of fibres Biological Effects of Mineral Fibres 2 737 - 754 IARC Scientific Publications No. 30 Acheson E D and Gardner M J (1981) Ex posure limits -- the scientific criteria Recent Ad'antcs in Otcu/iutmnal Health (ed, J C McDonald) 257 - 269 London: Churchill Livingstone Acheron E D, Gaidncr M J, Pippard E C and Grime L P (19R2) The mortality of two groups of women who manufactured gas masks from chrysotilc and croodolitc asbes tos, a 40-ycar follow-up British Journal of Industrial Medicine 39 344 - 348 Acheson. E D, Gaidner M J, Winter P D and Bennett C (1983) Cancer in a factory using amosite asbestos International Journal of Epidemiology' (in press) Advisory Committee on Asbestos (1979) Asbestos: Final Report of the Advisory Committee 1 London HMSO Asbestos Information Centre Ltd. Personal communication to the authors Bengtsson N O, Hardell L and Eriksson M (1982) Asbcstos_expovure and malignant lymphoma Lancet 2 1463 Berry G (1980) Dove-response in casecontrol st\id\es~Journal of Epidemiology and Community Health 34 217 -222 Berry G (1981) Mortality of workers certi fied by pneumoconiosis medical panels as having asbestosis British Journal of Industrial Medicine 38 130-137 Berry G, Gilson J C, Holmes S, Lcwinsohn H C and Roach S A (1979) Asbestosis: a study of dose-response relationships in an asbestos textile factory British Journal of Industrial Mednine 36 98 - 112 Betry G and Ncwhousc M L (1983) Mor tality of workers manufacturing friction materials usihg~asbcstos British Journal of Industrial Medicine 40 1 - 7 Bertrand R and Pe/crat H (19R0) Fibrous glass: carcinogenicity and dimensional characteristics Biological Effects of Mineral Fibres 2 901 - 911 IARC Scientific Publications No 30 Blot W J, Stone B J, Fraumcni J F and Morris L E (1979) Cancer mortality in US counties with shipyard industries during World War 2 Environmental Research 18 281-290 Borow M, Conston A, Livornese L and Schalet N (1973) Mesothelioma following exposure to asbestos: a review of 72 cases Chest 64 641 -646 British Occupational Hygiene Society: Committee on Hygiene Standards (1968) Hygiene standards for chrysotilc asbestos dust Annals of Occupational Hygiene II 47-69 British Occupational Hygiene Society: Committee on Asbestos (1983) A study of the health experience in two UK asbestos factories Annals of Occupational Hygiene 27 1 - 55 Brown D P, Dement J M and Wagoner J K (1979) Mortality patterns among miners and millers occupationally exposed to asbestiform talc Dusts and Disease 317 Pathotox, Park Crescent, Illinois Biowne K and Smither W J (1983) Asbestos related mesothelioma: factors discriminating between pleural and peritoneal sites British Journal of Industrial Medicine 40 145 - 152 19 DU 038418 DU P 0947118.024 Burch J Df Howe C R, Miller A B and Scmenciw R (19SI) Tobacco, alcohol, asbes tos and nickel in the aetiology of cancer of the larynx a case-control study Journal of the National Cancer Institute 67 1219- 1224 CIcmmcsen J and Hjalgrim-Jensen S <1981) Cancer incidence among 5686 asbestoscement workers followed from 1943 through 1976 Journal of Ecotoxicology and Environmental Safety S 15-23 Craighead J E and Mossman B T (1982) The pathogenesis of asbestos-associated diseases New England Journal of Medicine 306.1446- 1455 Davis J M G (19S2) Carcinogenic effect of fibres in inhaljtion studies. Abstract of paper presented to the VD1 Conference on Fibrous Dusts; Strasbourg, October 1982 Dement J M, Harris R L, Symons M J and Sh> C (I9S2) F'tinutes of dove-revponsc for respirator) cancer among chrysotile asbestos tevtile workers Annals of Occupational Hygiene 26 S69-887 Dement J M, Harris R L, S>mons M J and Shy C M (19S3) Exposures and mortality among ehr>sotile asbestos workers. Part II: Mortality American Journal of Industrial Mi dune 4 421 -433 Duncan K P (19S1) Exposure limits -- whose responsibility Recent ad\ances in 0<<upationai Health (ed J C McDonald) 247 - 255 London: Churchill Lixingstone Elines P C and Simpson M J C (1977) Insu lation workers in Belfast. A further study of mortalit) due to asbestos exposure (1940 -- 1975) British Journal of Industrial Mtdutne 34 174 - 180 Enicrline P E (19SI) Extrapolation from occupational studies: a substitute for cn\ironrr.<-n!al epidemiology Environmental Health Perspedt\es 42 39-44 Finklestein M (1982) Mortality in asbestoscerncnt factory workers. Paper presented to 2nd Internationa! Symposium on Epidemi ology in Occupational Health, Montreal, September 1982 Gardner M J (I9?3d) Tumour incidence after avbc-'tox exposure m Great Britain -- with special reference 10 the cancer risk of the non (N^upjtioi.al population. VDI -- Bericht Report of the Pick codings of a Conference on Fibrous Dusts, October 1982. In press Gardner M J (1983b) The effect of a second factor on a linear dose-response relationship Journal of the University of Occupational and Environmental Health 5 183-188 Gardner M J, Achexon E D and Winter P D (1982) Mortality from mesothelioma of the pKura during 1968--78 in England and Wales British Journal of Canter 46 81 - 88 Grimes P (19S2) Personal communication to the authors Haringion J S (1981) Fibre carcinogenesis: epidemiological observations and the Stanton hypothesis Journal of the Notionol Cancer Institute 67 977 - 989 20 Henderson V L and Enterline P E (1979) Asbestos exposure: factors associated with excess cancer and respiratory disease mor tality Annals of the New York Academy of Sciences 330 117-126 Hobbs M S T (1983) Personal communi cation to the authors Hobbs MST, Woodward S D, Murphy B, Musk A W and Elder J E (1980) The incidence of pneumoconiosis, mesothelioma and other respiratory cancer in men engaged in mining and milling crocidolite in Western Australia Biological Effects of Mineral Fibres 2 615 - 625 IARC Scientific Publications No 30 Hughes J and Weil! H (1980) Lung cancer risk associated with manufacture of asbes tos-cement products Biological Effects of Mineral Fibres 2 627 - 635 IARC Scientific Publications No 30 Hwang C Y and Gibbs G W (1981) The dimensions of airborne asbestos fibres: I. crocidolite from Kuruman area, Cape Province, South Africa Annals of Occupational Hygiene 24 23 - 41 Jones J S P, Smith P G, Pooley F D, Berry G, Sawle G W, Wignall B K, Madeley R J and Aggarwa! A (1980) The consequences of exposure to asbestos dust in a wartime gas-mask factory Biological Effects of Mineral Fibres 1 637 - 653 IARC Scientific Publications No 30 Kleinfeld M, Messite J and Kooyman O (!967a) Mortality experience in a group of asbestos workers Archives of En\ironmental Health 15 177- 180 Kleinfeld M, Messite J, Kooyman O and Zaki M H (1967b) Mortality amongst talc miners and millers in New York State Archives of Environmental Health 14 663-667 Kolonel L N, Hirohata T, Chappell B V, Viola F V and Harris D E (1980) Cancer mortality in a cohort of naval shipy ard workers in Hawaii; Early findings Journal of the National Cancer Institute 64 739-743 Lacquct L M, van der Linden L and I.epoutrc j (1980) Rocnigcnographie lung changes, ashcMntU and mentality in a Rxlglun avbeMos cement fuctory Biological Effects of Mineral Fibres 2 783 - 793 IARC Scientific Publications No 30 Langcr A M and MeCaughey W T F. (1982) Mesothelioma in a brake repair worker Lancet 2 1101 - 1103 Liddell F D K (1982) Some new and revised risk extrapolations from epidemiological studies on asbestos workers Proceedings of the BCA Symposium rebmary, I9&2 BGABcriehic (in press) Liddell F 6 K (I9K2&) Contribution after Schneiderman's address Prthredtngs of the BCA Symposium, February, 1982 BGA* Berichte (in press) Mancuso T F and Coulter E J (1963) Meihodology in industrial health studies -- the cohort approach, with special reference to an asbestos company Archives of Environmental Health 6 2J0 - 226 Mason T J, McKay F W, Hoover R, Blot W J and Fraumeni J F (1975) Allas of Cancer Mortality for U.S. Counties: 1950--1969 Washington: US Government ' Printing Office McDonald A D (1980) Malignant meso thelioma in Quebec Biological Effects of Sfincral Fibres 2 673 - 680 IARC Scientific Publications No 30 McDonald A D and Fry J S (1982) Meso thelioma and fibre type in three American asbestos factories -- preliminary report Scandinavian Journal of Work, Environment and Health 8 (Supplement 1) 53-58 McDonald A D, Fry J S, Woolley A J and McDonald J C (unpublished a) Dust ex posure and mortality in an American chrysolite textile plant. Submitted for publication McDonald A D, Fry J S, Wooley A J and McDonald J C (unpublished b) Dust ex posure and mortality in an American factory using chrysotile, amosite and croci dolite in mainly textile manufacture. Sub mitted for publication McDonald A D and McDonald J C (1978) Mesothelioma after crocidolite exposure during gas mask manufacture Environmental Research 17 340 - 346 McDonald A D and McDonald J C (19S0) Malignant mesothelioma in North America Cancer 46 1650-1656 McDonald A D, McDonald J C and Pooley F D (1982) Mineral fibre content of lung in mesothelioma tumours in North America Annals of Occupational Hygiene 26 417-422 McDonald J C (1982) Aspects of the asbes tos standard. Paper presented to the Inter national Conference on Occupational Lung Disease; Chicago, March, 1982 McDonald J C, Gibbs G W and Liddell F D K (1980a) Chrysotile fibre concentration and lung cancer mortality: a preliminary report Biohignal Effects of Mineral Fibres 2 811 817 IARC Scientific Publications No 30 McDonald J C, Liddell, F D K, Gibbs G W, Eysscn G E and McDonald A D (!9S0b) Dust exposure and mortality in chrysotile mining, 1910--75 British Journal of Industrial Medicine 37 11 - 24 McDonald J C and McDonald A D (1981) Mesothelioma as an index of asbestos impact Quantification of Occupational Cancer Banbury Report Numher 9, 73-85 Cold Spring Harbor, New York Meurman L O, Kiviluoto R and Hakama M (1974) Mortality and morbidity among the working population of amhophyllite asbes tos miners in Finland British Journal of Industrial Medicine 31 105 - 112 DUP 0947118.025 DU 038419 Morgan A and Holmes A (1982) Concen* (rations and characteristics of amphibole fibres in the lungs of workers exposed to crocidoluc in the British gas mask factories and eKewhere, during the Second World War British Journal of Industrial Medicine 39 62 - 69 Morris J G (I97S) Paper submitted to the Ad\isor> Committee on Asbestos Newhouse M L and Berry G (1979) Patterns of mortality in asbestos factory workers in London Annals of the Sc' York Academy of Sciences 330 53-60 Newhouse M L, Gregory M M and Shannon H (1980) Aetiology of carcinoma of the larynx Bidogua! Effects of Mineral Fib/es 2 6S7-6S9 IARC Scientific PubliCanons No 30 Ncuhou^e M L, Berry G and Skidmore J W (1982) A mortalii) study of workers manu facturing friction materials with chrysotile asbestos Annuls of Occupational Hygiene 26 99 - 909 Nicholson W J (1981) Criteria document for Swedish occupational standard: Asbestos end inorganic fibres Arbcte Och Halsa, Veterskaphg Skritserie, 1981: 17, Siockhoim Nicholson W J, Selikoff I J, Seidman H, Lihs R and Formby P (1979) Longterm n.oMd;c> experience of chrysotile miners and miners in Thetford Mines, Quebec Arncls of the Ac* York Acodemy of S^nn.es 330 11-21 N::ho1*-on W J et ai (19S1) Mortality experience of asbestos factory workers: effects of differing intensities of asbestos exposure Environmental Research in press Obson H and Brandt L (1983) Asbestos ex* po^ure and nonHodgkin's lymphoma Lancet 1 588 Peto J (1978) The hygiene standard for chrysotile asbestos Lancet 1 484-489 Peto J (1979) Dose-response relationships for asbestos-related disease: implications for hygiene standards Part 11, mortality Annals of the Sew York Academy of Sciences 330 195 - 203 Peto J (1980a) Lung cancer mortality in relation to measured dust levels in an asbes tos textile factory Biological Effects of Mineral Fibres 2 829 - 836 IARC Scientific Publications No 30 Peto J (1980b) The incidence of pleural mesothelioma in chrysotile asbestos textile workers Rio/ogicol Effects of Mineral Fibres 2 703 - 71 IARC Scientific Publications No 30 Peto J, Doll R, Howard S V, Kinlcn L J and Lewinsohn H C (1977) A mortality study among workers in an English asbestos factory British Journal of Industrial Medicine 34 169-173 Peto J, Seidman H and Selikoff 1 J (1982) Mesothelioma mortality in asbestos workers; implications for models of car* cinogencsis and risk assessment British Journal of Cancer 45 124-135 Pooley F D and Clark N (1979) Fiber dimensions and aspect ratio of crocidolite, chrysotile and amosite particles detected in lung tissue specimens Annuls of the Sew York Academy of Sciences 330 711 -716 Robinson C E, Lenien R and Wagoner J K (1979) Mortality patterns 1940--1975 among workers employed in an asbestos textile friction and packing products manu facturing facility Dusts and Diseases 131 - 143 Pathotox, Park Forest, Illinois Ross R, Dworsky R, Nichols P, PaganiniHil! A, Wright W, Koss M, Lukes R and Henderson B (1982) Asbestos exposure and lymphomas of the gastrointestinal tract and oral cavity Lancet 2 1118-1119 Skidmore J W and Dufficy B L (1983) Environmental history of a factory pro* ducing friction material British Journal of Industrial Medicine 40 8 - 12 Smither W J and Lewinsohn H C (1973) Asbestosis in textile manufacturing Biological Effects of Asbestos 169 - 174 IARC Scientific Publications No I Stanton M F, Layard M, Tegcris A, Miller E, May M, Morgan E and Smith A (1981) Relation of particle dimension to carcino genicity in amphibole asbestos and other fibrous minerals Journal of the National Cancer Institute 67 965 - 975 Steel J (1979) Asbestos control limits /fiberfos, Final Report of the Advisory Committee 2 Appendix 3 London: HMSO Rossilcr C E and Coles R M (1980) HM Dockyard, Dcvonport: 1947 mortality study Biological Effects of Mineral Fibres 2 713-72! IARC Scientific Publications No 30 Talent J M, Harrison W O, Soloman A and Webster J (1980) A survey of black mineworkers of the Cape crocidolite mines Bio logical Effects of Mineral Fibres 2 723 - 729 IARC Scientific Publications No 30 Rubino G F, Piolatto G, New house M L, Scansetti G, Aresini G A and Murray R (1979) Mortality of chrysotile asbestos miners at the Balangero Mine, Northern Italy British Journal of Industrial Medicine 36 187-194 Saracci R (1981) Personal-environmental interactions in occupational epidemiology Recent Advances in Occupational Health (ed J C McDonald) 119-128 London: Churchill Livingstone Thomas H F, Benjamin I T, Elwood P C and Sweetman P M (1982) Further follow up study of workers from an asbestos cement factory British Journal of Industrial Medicine 39 273 -276 Wagner J C, Berry G and Pooley F D (1982) Mesothelioma and asbestos type in asbestos textile workers: a study of lung contents British Medical Journal 285 603-606 Sebastien P, Janson X, Gaudiehet A, Hirsch A and Bignon J (1980) Asbestos retention in human respiratory tissues: comparative measurements in lung parenchyma and in parietal pleura Biological Effects of Mineral Fibres 1 237-246 IARC Scientific Publications No 30 Seidman H, Selikoff l I and Hammond E C (1979) Short term asbestos work exposure and longterm observation Annals of the Sew York Academy of Sciences 330 61-90 Selikoff I J (1981) Disability compensation for asbestos-associated diseases in the United States. Report to the US Department of Labour Selikoff I J, Hammond E C and Seidman H (1979) Mortality experience of insulation workers in the United States and Canada, 1943--76 Anndfs of the Sew York Academy of Sciences 330 91-116 Selikoff 1 J and Scidinan H (1981) Cancer of the pancreas among asbestos insulation workers Cancer 47 1469 - 1473 Selikoff 1 J, Seidman H and Hammond E C (1980) Mortality effects of cigarette smoking among amosite asbestos factory workers Journal of the Sational Cancer Institute 65 507-513 Wagner J C, Pooley F D, Berry G, Seal R M E, Munday D E, Morgan J and Clark N J (1982) A pathological and mineralogicaJ study of asbestos-related deaths in the United Kingdom in 1977 Annals of Occupational Hygiene 26 423 - 431 Walton W H (1982) The nature, hazards and assessment of occupational exposure to airborne asbotos dusts: a review Annals of Occupational Hygiene 25 117-247 Weill H, Hughes J and Wacccnspack C (1979) Influence of do>c nd fibre type in respiratory malignancy risk in a<be*tos cement manufacturing American Rcvic* of Respiratory Disease 120 345 - 354 Weill H (1982) In discussion of `Estimates of dose-response for respiratory cancer among chrysotile asbestos textile workers' by Dement et ol Annals of Occupational Hygiene 26 884 Weiss W (1977) Mortality of a cohort ex posed to chrysotile asbestos Journal of Occupational Medicine 19 737 - 740 Wignall B K and Fox A 3 (1982) Mortality of female gas-mask assemblers British Journal of Industrial Medicine 39 34 - 38 $1 DUP 0947118.026 DU 038420 jlaaiSiBiiiBi i .iif----* Figures Ttg 1 Tiends in annual imports of amosite and cio, hJolitc asbestos into the UK Source Asbestos Information Centre Ltd DUP 0947118.027 Cumulative dose (fibre-years/ml) i--_______ iii -- 0 12 3 Duration of employment (years) 4 Cumulative dose (fibre-years/ml) Tig 2 Mortality from lung cancer among amosile insulation ptodmtmn v.\tVvjs according to estimated cumulative dose of exposure to a'bcvtos and dural ion of employment . . !ndi\ idual data points from Scidman et al (1979) * dusc responsc relationship from Nicholson (1981) -- dose response relationship from Liddell (1982a) f ig 3 Mortality from lung cancer among chrysotfle textile production workers according to estimated cumulative dose of exposure to asbestos .. Individual data points from Dement et al (1982) dose-response relationship from Nicholson (1981) -- dose-response relationship due to Liddell from McDonald ct al (unpublished a) DU 038421 liainatHTnaE atn- ik mi amr\ utai.niirnl .li ifc^Mniar.faUn|| Tables Table 1 Imports of raw asbestos fibres to the UK by amount (tonnes) and percentage of 1946 figure Chrysolite CrocidoUte Amosite Year tonnes % tonnes 44 tonnes 44 1946 1955 1965 1975 ____ 1978 1980* 1981* : -- 50,700 123,000 147,000 120,000 123,000 89,500 76,400 100 243 290 237 243 177 151 1,000 6,800 3,400 -- -- -- -- -- 100 680 340 -- -- -- -- 2,700 12,300 22,600 19,200 2,700 700 261 100 456 837 711 100 26 10 Total tonnes 54,400 142,100 173,100 - 139,400 125,800 90,200 76,700 44 100 261 318 256 231 166 Ml Tf j-n foi and 1911 are cn mates Amwiic imports for 1912 estimated 10 be 16 tonne*. $o*'ct The rf,,'ei for 19*6. 1955. 1965 and 19?5 *ere given in Ttble 1 of our previous report. The figutes for 1971. 1910. 1911 and 1912 have been provided b> the Asbestos 1nfon*ion Centre Ltd Table 2 Uses of chrysotilc asbes tos imported into the UK during 19S1 Usage Asbestos cement a) building products b) pressure pipes Friction materials Floor tiles Jointings &. packings Textiles Other miscellaneous uses Chrysotile asbestos tonnes 28,050 26,350 1,700 11,200 8,000 7,250 3,000 18,900 percentage 37 35 2 15 10 9 4 25 Total 76,400 100 These figures a/e estimates In addition. 261 tonnes of amositc cre imported and used m the manufacture of pressure pipes Sourcr Figures provided by the Asbestos information Centre Ltd. DUP 0947118.028 DU 038422 P--- -- -- *1 n H 0 ? 1 I f! 3 ? {!>* :> DU 038423 oc TJ o 4SOk 09 bN) SO uzkMM.Iliili. - is.- i;I. i iliii'ji'ili' --. Table 4 Exposure of cases of pleural mesothelioma and matched controls to crocidolite and chrysotile in a factory manufacturing friction materials (*) Crocidolite Subject Case Control No known exposure i 30 Exposure to crocidolite Fringe exposure Definite exposure ii 7 3- ' Total 10 40 (b) Chrysotile -- Exposure to chrysolite Subject Under 5 ftbres/ml 5 fibres/mt or higher Case Control i 30 9 10 Total 10 40 (c) Crocidolite (Cr) and Chrysotile (Ch) Exposure (Cr, Chjf Subject Case Control N,<5 0* 26 N.S + 1 4 E.<5 1 4 E.S + 8 6 Tolal 10 40 +Cr, N No known exposure. E Etposed (defimie or fringe). Ch. <5 <i ftbres/ml, 5* - 5+ fibres/m) If fringe cspoiuie to crocidolite is disregarded, the numbers become I, 0. I, I and 29, 1, I. 2 respectively for t*'o and controls Source Nf house er/ (! 912) Table 5 Criteria employed to define the occurrence of adverse effects in the BOHS Committee on Asbestos morbidity study A At least two readers agreed that there were two or more steps of change on the profusion scale for combined small opacities over the interval between films. B At least two readers agreed that a pleural abnormality was present on the later but not on the earlier Him of a pair (where 'pleural abnormality' means at least one of pleural thickening, pleural calcification, or costophrenic angle obliteration). C Unusual* rate of change in FEV, D Unusual* rate of change in FVC E Unusual* rate of change in CtF F Any one measurement of FEV,/FVC <0.70 G Chest sounds which did not clear on coughing or on any subsequent examinations 'Unusual' it defined at any value in the tower 20 petermik of the distribution after adjustment for initial level, age, height, cifhi and smoking habits Sec paras S4 - SI m SONS (1913). Sourer ftOHS (I*)) DUP 0947118.030 DU 038424 FiMaraafii' L: t..: Table 6 Estimates of the slopes of the linear relationships* between mortality from lung cancer and cumulative asbestos exposure in various studies of asbestos workers Study number (for reference see text) Types of asbestos Acheson and Gardner (1979a) (SMR) Estimate 0/ slope by: Nicholson (1981) (Selikoff, 19SI) (SMR) Liddell (1912) Original authors \ 2 7 8- 9 6E 6L 3 5 4 15 10 nm 13 F 11 u Ch Ch Ch Ch Ch, Or Ch, Or Ch, Cr Ch Ch, Cr! A Ch Ch, A Ch, Cr, \ Ch, Cr, K Ch, Cr, K Ch, Cr, K A 0.06 -- -- - -- 0.J 0.125 0.42 -- -- -- -- -- -- 0.06 0.15 5.3 - -- 0.07 0.8 0.3 1.1 1.7 -- 1.3 8.4 9.1 0.038f (RR) 0.045 (SMR) 1.0 (RE) -1.6 (RR) 0.058 (RR) 0.12 (RR) 0.5 (SMR) 1.5 (SMR) 0.22 (RR) 0.4-1.1 (RR) 2.7 (RR) -- 1.1 (RR) 0.038 (RR) 0.15 (SMR) 4.0b (SMR) 1.7*c (RR) 2.0*d (RR) 0.058 (RR) 0.5 (SMR) 0.219 (SMR) -- -- -- -- -- 1.7* -- (RR) E = pre-195) entry, L = post-1950 entry; M * males, F * females; Ch * Chrysotile, Cr * Crocidolite, A Amosite. f = Liddell FDK, Thomas D C, Gibbs G W and McDonald J C Fibre exposure and mortality from pneumoconiosis, respiratory and abdominal malignancies in chrysotile production irTQuebcc, 1926--75. In preparation, a = Amended to 2.3 (Liddell FDK, personal communication), b Estimated from Fig 2 of Dement et al (1982). c = From Fig 2 of McDonald et at (unpublished a), using as far as possible the same methods as Dement et al (1982) -- incorporating an average conversion factor of 6, for million particles/cubic foot to fibres/ml, would give an estimated slope of 0.85. d = From McDonald et al (unpublished b) using same methods as in study II -- incorporating an average conversion factor of 6 would give an estimated slope of 1.0. e * Liddell (1982) unable to make independent estimate. 4- = Incorporating a conversion factor of 3 for million particles/cubic foot to fibres/m! into results of McDonald et at (unpublished a and b). * Either SMR = 100 + slope x cumulative dose in fibre-years/ml, for Standardised Mortality Ratio; or RR * I + x cumulative dose in fibre-years/ml, for Relative Risk. Table 7 Concentrations of chrysotile asbestos in fibres/ml to produce after 50 years* exposure stated excess proportions of deaths from lung cancer (as proportions of deaths from al! causes) according to linear dose-response models at current male lung cancer rates in Britain Excess lung cancer mortality Study Miners and millers Friction materials Production Workeri (E) Textiles Rochdale (A) Rochdale (L) South Carolina (M) South Carolina (N) South Carolina (M) South Carolina (N) Conversion foetor* 3 3 3 -t -t 3* 36" 6- Slope 0.04 0.06 0.125 0.5 0.8 2.3 5.3 1.2 2.7 0.5% 2.5 1.7 0.8 0.2 0.1 0.04 0.02 0.09 0.04 5.0 3.3 1.6 0.4 0.3 0.09 0.04 0.17 0.08 2% 10,0 6.7 3.2 0.8 0.5 0.17 0.08 0.35 0.15 * From m>ll>oM of pantclet/cub*c foot to flbret/m] * - Eicrr" in the arc* vfccrt chtyxxik ti prepared tot which con*m*oft figure of I uted (Dement ti of. IM2) McDonald ft (unpublished ) E Enirrl.nc (iv|). A Achevon and Gardner (l*73a). L pou-ISO entry rohofl, N NtcholwNi (IVdl). M McDonald tt ! (unpubtuhed ) t * Convettton alto made from million panickt/cvbtc foot to fihm/ml at earlier wafe of calculations Printed ui the UK for MMSO Od 7153)0 C SO IfS3 DU 038425 DUP 0947118.031 la.; M.auaLiui.irt*AJiAiia aaBj-- A catalogue of HSE publication! b available on tale from government bookshops. Indexed by subject headings, the catalogue b an invaluable source of reference for anyone who needs access to advice and information on the requirements of the 1974 Health and Safety at Work Act and related legislation and publications bsued prior to the formation of the Health and Safety Executive. * HER MAJESTVS STATIONERY OFFICE Government bookshops 49 High Holborn. London WC1V6HB 13a Castle Street. Edinburgh EH2 3AR Brazennose Street, Manchester M60 8AS Southey House, Wine Street, Bristol BSI 2BQ 258 Broad Street, Birmingham B1 2HE 80 Chichester Street, Belfast BT1 4JY Government publications are also available through booksellers 1 f i j DU 038426 DUP 0947118.032 Health & Safety Executive 25 Chapel Street London NW1 5DT Telephone 01-262 3277 ext 349/359 Telex 299950 From the Director of Hazardous Substances Division Dr. Leonard Vance,. Director, Health Standards Programs, Occupational Safety and Health Administration, 200 Constitution Avenue, Washington DC 20210, U.S.A. ' Your reference Our reference Date 1 August, 1983. Dear Dr. Vance, I am pleased to say that the Acheson and Gardner report has now been published so there are now no restrictions on referring to it or quoting from it. The Health and Safety Commission were only able to give it preliminary consideration at their meeting on 19 3uly. It is therefore too early to say what view they will take about its conclusions and recommendations. They will be reviewing their policy on asbestos at their meeting on 23 August. However, at their last meeting the Health and Safety Commission did decide that there should be a change in the recommended method for evaluating airborne concentrations of asbestos. The Health and Safety Executive will therefore be changing to the European Reference Method (i.e. the methods set out in the annex to the EC Workplace Directive on Asbestos) as soon as the necessary hardware has been obtained and retraining completed. Due to the time required to implement this change it is not likely to be formally introduced until 1984. The HSE Guidance Note will be amended accordingly in due course. Yours sincerely, DUP 0947118.03 (dictated by Mr. Burgess and signed in his absence.) DU 038427