Document MG7DYKBjQB5X2BRBbR0pBV6Kz

Reprinted fromBritish Journal ofIndustrial Medicine 1986;43:436-444 Copyright 1986 British Journal of Industrial Medicine _ All rights of reproduction of this reprint are reserved in all countries of the world Cohort study of mortality of vermiculite miners exposed to tremolite JC MCDONALD, AD MCDONALD, B ARMSTRONG, AND P SEBASTIEN British Medical Association, Tavistock Square, London WC1H 9JR British Journal of Industrial Medicine 1986;43:436-444 Cohort study of mortality of vermiculite miners exposed to tremolite JC MCDONALD, AD MCDONALD, B ARMSTRONG, AND P SEBASTIEN From the School ofOccupational Health, McGill University, Montreal, Quebec, Canada abstract A cohort of 406 men employed before 1963 for at least one year in a vermiculite mine in Montana was followed up until July 1983. The vermiculite ore as fed to the mill contained 4-6% of amphibole fibre in the tremolite series. Vital status was established in all but one of the 406 and death certificates were obtained and coded for 163 of the 165 men who died. Compared with white men in the United States, the cohort experienced excess mortality from all causes (SMR 117), respiratory cancer (SMR 2-45), non-malignant respiratory disease (SMR 2-55), and accidents (SMR 2-14). Four deaths were from malignant mesothelioma (proportional mortality 2-4%). Compared with Montana death rates, the SMR for respiratory cancer was somewhat higher (3 03). Man-year analyses of respiratory cancer and estimated cumulative exposure gave a relation that did not depart significantly from linearity. The results of this and case-referent analyses indicate an increased risk of mortality from respiratory cancer in this cohort of about 1 % for each fibre year of exposure. In relation to estimated exposure the mortality experienced by the cohort from both lung cancer and mesothelial tumours was higher than in chrysotile mining. In April 1983 we were invited by WR Grace & Co, owners of a vermiculite mine in Libby, Montana, to investigate the health of their employees. Morbidity and mortality studies in the same population had been initiated by NIOSH in February 1982 but, because of the urgent decisions facing them, the com pany wished to have an independent inquiry as quickly as possible. Having assured ourselves that NIOSH and the employees' union had no objection to parallel and complementary studies, and indeed wel comed them, our research proposals were finalised and accepted in August 1983 and we began work in September. Our studies had four components: a cohort mortality study (described in this paper), a cross sectional radiographic survey, an investigation of asbestos bodies in sputum, and an electron micro scopic analysis of lung tissue samples at necropsy from ex-employees. Findings from the last three stud ies are reported separately,1'3 as are those of NIOSH. Although the NIOSH inquiries and our own were designed, conducted, and analysed quite sepa rately, we discussed and compared our data and methods at every stage. This was a valuable exercise; unnecessary differences and discrepancies were elimi- Accepted 30 September 1985 nated and the reliability of both sets of results increased. Concern for the health of workers in this mine stemmed not from exposure to vermiculite, a micaceous mineral for which no adverse effect is sus pected, but because the ore body is contaminated with fibrous amphibole deposits in the tremolite series. Fibrous tremolite and related minerals fall within the definition of asbestos4 but have limited value and are produced only in some small mines in northern Italy. On the other hand, they are common contaminants in various mining operations, notably the chrysotile mines of Quebec and certain talc mines in New York State. Other members of the amphibole family include crocidolite and amosite which cause pul monary fibrosis, lung cancer, and malignant meso thelial tumors. Existing evidence suggests indeed that most occupationally related mesotheliomas are caused by amphibole fibres and few by chrysotile.5'7 The experience of men in the Libby mine exposed only to the tremolite group of fibres was therefore of practical concern and scientific importance. Few observations on the effects on health of trem olite have been made. Cohort studies of employees of talc mines in New York State, some of which may contain fibrous tremolite, showed an excess of pul monary disease and lung cancer and one or two cases of malignant mesothelioma.8 9 In a postmortem study 436 10003333 444 10 McDonald AD, McDonald JC, Pooley FD. Mineral content of lung in mcsothdial tumours in North America. Am Occup Hyg 1982;26:417-22. "Pooley FD. An examination of the fibrous mineral content of asbestos in lung tissue from the Canadian chrysotile mining industry. Environ Research 1976;12:281-98. ud* Rowlands N, Gibbs GW, McDonald AD. Asbestos fibres in the lungs of chrysotile miners and millers. A preliminary report. Am Occup Hyg 1982:26:411-5. 11 Churg A, Wiggs B, Depaoli L, Kampe B, Stevens B. Lung asbestos content in chrysotile workers with mesothelioma. Am Ret Respir Dis 1984;138:1042-5. 14 Wagner JC, Chamberlain M, Brown RC, elal. Biological effects of tremolite. Br J Cancer 1982:45:352-60. I 5 Rousseaus JM, Rouxhet PG, Vielvoye LA, Herbillon AJ. The vermicuiization of Irioctahedral micas. The K level and its cor relation with chemical composition. Clay Minerals 1973;10:1-16. 14 Atkinson GR, Rose D, Thomas K, Jones D, Chatfield El, Going JE. Collection, analysis and characterisation of termiculite sonpies for fibre content and asbestos contamination. Washington DC: US Environmental Protection Agency, 1981. (Contract No 68-01-5915.) II US Department of Labor. Asbestfform and/or fibrous minerals in mine, mills andquarries. Pittsburgh: Mine and Safety and Health Adminstration. 1980. " Walton WH. The nature, hazards and assessment of occupational exposure to airborne asbestos dust: a review. Am Occup Hyg 198235:117-247. McDonald, McDonald, Armstrong, and Sebastian "Sebastian P, Gaudichet A, Billon-Galland MA, Janson X. Spectrometric X par dispersion d'enerjpe en microscopic electronique a transmission: application a la caracterisation des fibres miner als. Journal de Microscopic el de Spectroscopic Electronique 1980;3:83-97 20 International Mineralogtcal Association. Nomenclature of amphiboles. Canadian Mineralogist 1978:16:501-20. 11 Edwards GH, Lynch JR. The method used by the US Public Health Scrivce for enumeration of asbestos dust on membrane filters. Am Occup Hyg 1968;11:1-6. u Cora M, Esmen NA. Work place exposure zones for classification of employee exposures to physical and chemical agents. Am Ind Hyg Assoc J 1979:40:47-57. "Oldham PD. On estimating the arithmetic means of lognormallydistributed populations. Biometrics 1965;213:235-9. " Monson R. Analysis of relative survival and proportional mor tality. Computers and Biomedical Research 1974;7:325-32. 35 Hanley J, Liddell FDK. Fitting relationships between exposure and standardized mortality ratios. J Occup Med 1985;27:555-60. "Thomas DC. General relative risk models for survival time and matched case-control analysis. Biometrics 1981;37:673-86. 37 Doll R, Pelo J. Effects on health ofexposure to asbestos. London: Health and Safety Commission, HMSO, 1985. "Liddell FDK, Thomas D. Gibbs GW, McDonald JC. Fibre exposure and mortality from pneumoconiosis, respiratory and abdominal malignancies in chrysotile production in Quebec, 1926-75. Am Acad Med Singapore 1984;13:340-4. "Oakes D, McDonald JC. Restricted cohort study designs. Scand J Work Environ Health 19824:30-3. Destruction of manuscripts From 1 July 1985 articles submitted for publication will not be returned. Authors whose papers are rejected will be advised of the decision and the manuscripts will be kept under security for three months to deal with any inquiries and then des troyed. 10003341 - 442 Fig 2 Respiratory cancer SMRs 20 or more yearsfromfirst employment exposure ff/mlyears). ogist did not consider this a mesothelioma. Using Montana death rates instead of United States national rates reduced the number of expected deaths from respiratory cancer from 9-41 to 7-60, increasing the SMR from 2-45 to 3-03. Table 4 gives the SMRs for respiratory cancer (using Montana death rates) by cumulative exposure. It is reasonable to focus attention on the SMRs 20 or more years from first employment, when risk due to exposure to fibres would be expected to be most evident. A linear relation between these SMRs and cumu lative exposure (see fig 2) is given by the equation SMR -- 1-52 + 0-011 f/y or, expressed as a relative risk, RR = 1 + 0-007 f/y. Alternatively, by forcing the line through SMR = 1 at zero exposure, the equa tion SMR = 100 + 0-013 f/y was obtained. The observed deviation from linearity could be explained by chance Of2 (3df) = 3-47) and the trend has high statistical significance Of2 (Idf) = 18-9, p < 0-001). McDonald, McDonald, Armstrong, and Sebastien The constrained line may be interpreted to imply an increase in SMR for respiratory cancer of about 1 -3% for each fibre year of cumulative exposure. There is also evidence-fronrthe data in table 4 of an excess mortality from lung cancer less than 10 years from first employment. The two cases in the lowest exposure category could not have been due to work at Libby. One was a man who was first employed at age 66 and died at 69; the other joined at 25 and died at 28. The two other deaths in the short latency category were 8-9 and 91 years from first employment. Both had substantial levels of cumulative exposure (470 and 522 f/y) so a causal hypothesis is less easily dis missed. Case-control analyses of exposure-response were carried out for the 23 deaths coded to respiratory can cer (ICD 160-163), the eight coded to pneu moconiosis (ICD 514-515), and (with some overlap) the four mesotheliomas. Results are presented in table 5. A statistically significant linear relation was found between relative risk for respiratory cancer and cumulative exposure which is clearest in cases occur ring 20 or more years from first employment Of1 (ldf) = 7-04, p < 0-01). In this group the linear increase in relative risk is estimated at 1-0% per fibre year (95% confidence interval 0-1-9-7). This relation is described by the equation RR = 1-00 + 0-010 f/y. Although positive relations with cumulative exposure were also observed with pneumoconiosis and with meso thelioma, numbera were small, confidence intervals wide, and conventional levels of statistical significance were not reached. Discussion The cohort studied was not large but sufficient to show that the workers in this mine experienced a seri ous hazard from lung cancer, pneumoconiosis, and mesothelioma. This was presumably attributable to the fibrous amphibole contaminant of the vermiculite ore and mainly to dust conditions before 1974, partic ularly in the dry mill. No precise comparison with the New York State studies can be made but the crude SMRs for lung cancer were similar, 2-4 in this study, 2-8 in the study by Kleinfeld etal,8 and 3 2 in that by Table 5 Estimates ofexposure: responsefrom case-referent analysis Cause of death Respiratory cancer (ICD 160-163) Pneumoconiosis (ICD SIS) Mesothelioma (ICD various) No of eases 4 4 15 23 8 4 Time fromfirst employment (years) 0-9 10-19 >20 All All All (>20) fi % increase in relative risk (OR) per find year (95% Cl) 2 1 (0 0, 66) 0-4(-0 1, 8 5) 10(01,97) 0-7(0 1, 3-0) 0-3(00,4 1) 0-2 (-02, >100) X1 (ldf) for H.fi - 0 3-08 1 23 7-04 749 2-42 0-10 1000333? 440 brane filter method between 1970 and 1974 showed a mean concentration of 221 f/cc. It was therefore assumed that the concentration in f/cc before 196S was 4-6 times higher--that is, 101.5 f/cc. Only two operation locations were considered in the dry mill: dry mill working (except sweeping) and dry mill sweeping. Personal samples indicate that sweepers in the dry mill were exposed to concentrations about 20% higher. For the other operation locations fibre mea surements were available only for the recent periods. When the data were considered inadequate to describe past conditions, because of changes in pro cess or control practice, arbitrary correction factors were applied. This was done after discussion with the company's representatives and especially with a pre vious manager who had spent almost all his career with Libby and who had extensive knowledge of the operations. Table 2 shows the estimated average levels of exposure at the 28 operation locations from 1945 to 1982. Recent values (1957-82) were calculated in the same manner from data provided by NIOSH, where all available air measurements had been collected. As the company had introduced a systematic air sam pling programme in 1975, the number of mea surements for this period was much higher than in the past. As these recent measurements had been made by long term personal sampling, the operation location model may be less appropriate since the same sample may cover more than one operation location. Recent - McDonald. McDonald, Armstrong, and Sebastien fibre counts, however, were low so it is unlikely that this materially affected the accuracy of our cumu lative exposure estimates. The work histories were available in terms of job category, not operation location. We therefore used job descriptions to estimate for each job the duration and the number of hours spent by workers in different operation locations; the cumulative exposure was then calculated. Again, the help of company represen tatives made this possible. STATISTICAL ANALYSIS The mortality of the total cohort was compared with that of white men in the United States using the person-years at risk method to compute expected numbers of deaths and hence standardised mortality ratios (SMRs). These computations were made using Monson's computer program,24 from which United States national age specific death rates, available to 1975, were also taken. Additional analyses of mor tality from respiratory cancer were made using Mon tana death rates from published figures for numbers of deaths, together with appropriately interpolated estimates of population from censuses. Death rates for respiratory cancer for Montana were computed for the period up to 1978, and the United States national rates for this cause updated to this year for comparability. Analysis of the relation between mortality and exposure to fibrous dust was performed in two ways. Firstly, the exposure cumulated to the end of each Table 2 Estimated average prevailingfibre concentrations (flml) at main work location operations 1945-82 1950 I960 WO 1980 Mine pit: Workins Drilling Tails belt operation Transfer point Dry mill: forking Sweeping New wet mill Mill lunch room Service area Skip boot River station: Dock Office Railroad by-standing Concentrate: Loading Bagging Hauling Testing: Laboratory Van 36 0 24 0 * 2-3 125 7-9 2-8 101 5 1249 70 10 68-8 180 120 4-7 * 15-0 240 5-4 2-9 1-0 20 60 90 5-2 0-6 08 0-7 0-6 22-1 27-2 2-4 0-9 15-0 3-9 1 5*0-8 0-5 * 0-2 2-0 06 0-7 02 12 0-5 4-8 5-5 * 02 1-2 0-4 0-6 05 Fibre concentrations were taken as constant for the relevant periods at the following locations: bus transfer (1 2), mine lunch room (0*9), mine office (0-5), screen plant (0-5), verxite plant {6 0), office building (< 0-04), hygiene laboratory (0 2), old wet mill (5-1), river station conveyor tunnel (118-5). No measurements available for figures in italics. Time periods over which the estimates apply. 10003337 438 the material expands about 13 times in volume and acquires useful properties. Expanded vermiculite is mainly used as an insulator against heat and noise in the construction industry and also as an absorbent low density filler and chemical carrier. Of the 323000 4 tons of vermiculite concentrate (before expansion) used in the United States in 1984, 54% came from the Libby mine. According to legend, the vermiculite mountain in Libby was discovered in 1913 by Edward N Alley: walking in a short tunnel in search of metal deposit, the flame of his mining candle came in con tact with vermiculite and the immediate trans formation by heat attracted his attention. Commer cial exploitation started in 1923. Libby production, initially low, increased from 20000 tons in 1940 to 130000 in 1930 and 200000 in 1970. At present mining is conducted on benches or levels that extend laterally from the surface to more than 300 vertical feet below the original surface of the mountain. Although equipment has changed over the years, ore has always been extracted by the classic technique of drilling, blasting, loading (with shovels or front end loaders), and hauling to the mill or trans fer point. Drilling was considered the most dusty operation and, in 1970, nev' drills with baghouses were introduced. After removal of coarse rock, the ore is blended and fed at controlled rates to the top of the mill. At this point the granules are about 1 cm in diameter and contain approximately 20% vermiculite. Accom McDonald, McDonald. Armstrong, and Sebastien panying minerals are augite, biotite, calcite, diopside, hornblende, magnetite, quartz, sphene, and tremoliteactinolite.16 Milling takes advantage of differences in particle shape to separate the thin flakes of ver miculite from the blocky waste particles. The finished concentrate in 1984 contained 80-93% vermiculite. Before 1934, the milling was basically a dry screen process. A second mill, operating a wet process, was added in 1933 to accomplish the first stages of sepa ration. Work in the dry mill was very dusty but became much cleaner after 1963 when ventilation equipment was installed. In 1974, for hygienic and technical reasons, both dry and wet mills were shut down and ore was processed in a new mill built nearby which operated on an entirely wet process in which separation was made by vibrating screens, roll screens, Humphrey separators, and flotation. The new wet mill operates continuously, producing 1100 tons of concentrate each day in 1984; this is then de watered, dried, and hauled to the screening plant. All the mill water is reclaimed after clarification in a tail ings pond. The screening operation separates grades of ver miculite concentrate, differing in particle size and industrial application. The techniques for handling and shipping concentrate at Libby have been changed several times with the object of controlling exposure to dust. At present most of the concentrate is shipped by rail, the loading station being equipped to reduce dust exposure; bagging is still used to a small extent. Fig 1 Dust particles endfibres collected on membranefilter in new wet mill. Transmission electron microscope view (1 mm = 0-24 pm). 10003335