Document zozV81nQY34L730NDZY6GZaLa

Cancer Produced by Nonoccupational Asbestos Exposure in the United States Philip E. Enterline University of Pittsburgh There Is considerable evidence of asbestos fibers In the general en vironment, and asbestos fibers can be found In the lungs of most adults in urban areas of the western world. Concentrations In large urban areas appear to average around 3 ng/m3 of air and In rural areas around 0.1 ng/m3 of air. For the entire U.S. population it can be es timated that the average population exposure Is 1.5 ng/m3. Based on the results of case control and other studies, it is estimated that about a third of the 1000 or so cases of malignant mesothelioma that occur in the U.S. each year appear to be related to this type of nonoccupational asbestos exposure. This Is a lifetime risk of 100 per million population. Lung cancer caused by asbestos in the general environment can be estimated from linear extrapolations of doseresponse data arising from occupational studies. Using data from a study of retired asbestos products workers It Is estimated that the lifetime risk of lung cancer due to continuous asbestos exposure at 1.5 ng/m3 is 2 per million. Possible causes for a fifty-fold difference between lung cancer and mesotheliomas caused by nonoccupational exposure to asbestos are discussed. These estimates are compared with estimates derived Irom a recent report commissioned by the West German government. Coated fibers have been found in the lungs of residents of urban areas of the western world for nearly 20 years now. The meaning of this for human health remains as unclear today as it was when the phenomenon was discovered. At first, it was believed that all of the fibers were asbestos hut it was soon realized that many were not. Nevertheless, a sufficiently large proportion have been positively identified as asbestos to warrant the descriptive term "asbestos bodies". The discovery of asbestos bodies in human lungs has led to a careful examination of the air in urban areas and the con clusion that all urban areas are polluted with asbestos fibers.1 The source of these fibers appears to he primarily asbestos building products. Table I summarizes levels of asbestos found in U.S. cities based on 24 h samples from each city.2 On av erage, the concentration in large urban areas appears to be about 3.6 ng/m3, ranging from 0.09 to 70 ng/m3. In rural and in small urban areas of the U.S., asbestos concentrations are much lower--around 0.1 ng/m3.3 Weighting these concen trations by number of people in large urban as compared with other areas gives an average outdoor asbestos exposure level of around 1.5 ng/m3. Copyright, Air Pollution Control Aiwncioticn 318 This is an attempt to estimate the number of cancers that are likely to be produced by asbestos among the nonoccupationally exposed populations in the U.S., that is, the popula tions exposed to asbestos at the estimated average outdoor background level of 1.5 ng/m3 of air. Not included in this es timate is cancer caused by nonoccupational indoor air pollu tion. Only two of the health effects of asbestos will be con sidered: malignant mesotheliomas and lung cancers. Malig nant mesotheliomas produced by background levels of as bestos will be estimated directly from reports on the numbers of these cases and their probable asbestos exposures, while lung cancers will be estimated by extrapolation from lung cancers produced by asbestos in occupational environments. Studies which attempted to estimate the overall impact of occupational asbestos exposures on the cancer problem in the U.S. will first be reviewed, as these demonstrate some of the problems encountered in the estimation of healt h effects at background levels. Table I. Concentrations of asbestos in urban areas of the U.S. Asbestos concentration ng/m3 Mt. Sinai Number of samples Battclle Institute Number of samples <1 / 1-2 2-5 5-10 10-20 20-50 50-100 Mean Median 61 58 45 12 8 1 2 3.58 1.57 27 33 42 22 1 2 0 3.62 2.29 Table II and Figure 1 summarize six studies that estimated the contribution of occupational asbestos exposures to annual cancer incidence in the U.S.4-9 Five out of the six reports are in relatively close agreement, whereas the U.S. government's estimate that 13-18% of all cancer is the result of occupational asbestos exposure is quite different. The only other estimate of this magnitude was by Dr. Irving Selikoff, who estimated 40,000 cancer deaths pier year caused by occupational exposure to asbestos, or about 10% of all cancer deaths.10 The U.S. government report has had a great impact mi world opiinion about asbestos since it was announced in 1978 Journal of the Air Pollution Control Association Tablerll- Annual number of cancer deaths in the U.S. due to occupational exposure to asbestos--various estimates. Source Exposed Annual Percent of population number of all U.S. alive 1980 cancer deaths cancer deaths McDonald* (1981)4 Higginson (1979)5 Enterline (198111' 7,600,000 Nicholson et al1 (1981) 9,200,(XX) Hogan and Hoel 7-8,000,000 (1981)8 U.S. Govt. (1978)9 7-10,000,000 3.330 4,000 1,084 8,500 12,000 !>8-'/t),UU0 <1% 1% 1% 2% :n 13-18% and Canada. and received extensive press coverage throughout the world. In its 1979 report to the Congress, HEW listed this report as a significant accomplishment during the year 1978.11 If the finding that 13-18% of all cancer is due to occupational ex posures to asbestos were credible, clearly the banning of all asbestos should be considered. There is lit tle support for the government document, however, as Table II suggests. In fact, partly as the result of this estimate, Congress commissioned a study that included an evaluation of the importance of oc cupational exposure to asbestos. The study concluded that asbestos may cause about 5% of all present day lung cancer.12 Inasmuch as lung cancer constitutes about 20% of all cancer, this suggests that asbestos caused lung cancer accounts for about 1% of all cancers---a finding which tends to support most of the estimates shown in Table 11. f Author example, in 1976 a group of government investigators reported on the effects of very low human exposures t o asbestiform fi bers among workers at a gold mine at Lead, Si) 1:1 They re ported a threefold excess in respiratory cancer and suggested that this was due to very low fiber exposures. A second study of workers from the same mine by researchers from Mcdill University showed essentially no excess in respiratory can cer.1* The U.S. government commissioned a third study by a private research group. This study found no excess in res piratory cancer among these same workers.15 One method for estimating lung cancer caused by low dose environmental exposure is by extrapolation from effects of exposure at higher levels.IK An estimate of lung cancer caused by nonoccupational exposure was prepared (using the 1981 Enterline estimate from Table II), assuming a general popu lation exposure of 1.5 ng/rrr and that 1 ng contains 40 fibers of asbestos greater than 5/jm in length, seen by the optical microscope.1' The number of lung cancers at 1.5 ng/rn was estimated by a simple linear extrapolation from the experience in a single British study by Peto, ct al. of the dose-response relationship between asbestos and lung cancer in a population of asbestos textile workers.17 These estimates of asbestoscaused lung cancer are shown in Table III. Table III. Annual number of cancer deaths in the 1 ,S. rinsed hy exposure to asbestos. Type of exposure and cancer Annual number of eaneerdeaths 1 Vrcenl ol all 1 S. ca ref dual h , Occupational exposure hung cancer .Mesothelioma ()ther cancer Nonoccupational exposure I.ung cancer Mesothelioma 4084 2501 333 215!) 301 28 333 i <0.1 Figure 1. Estimates of the percentage of all cancer due to occupational exposiffe to asbestos in the U S in 1980 (McDonald estimate reflects U S. and Canada). While there are very large differences in the cancer deaths jhiwn in Table II, there is not much disagreement on the size i the occupationally exposed population. The reasons for the references in the estimates of heall h effects are twofold: (1 > plifference of opinion as to the intensity of exposures received ed | fvvarious groups of workers, and (2) a difference of opinion tal s to the appropriate dose-response relationship for asbestos re ,nd cancer. t's When dealing with the health effects of outdoor background ial jbestos concentrations, there are too few data on the intensity tie ;exposure to cause much disagreement. The main problem ed ; to calculate the response likely at very low exposure ire eels. Attempts to directly measure the amount of lung cancer on -reduced by low doses of asbestos have been unsuccessful. For >78 11983 Volume 33, No. 4 ion Alsu shown is malignant mesothelioma caused by asbestos exposure. This was calculated as follows: There are about 1<XX) new malignant mesotheliomas in the U.S. each year with a male-female ratio of 3:1, or about 750 males and 250 females.18 McDonald and McDonald estimate that for maDs with mesothelioma, 53% are the result of an occupational asbestos exposure.4 From their study it also appears that no inure Ilian 1 5% of females with a malignant mesothelioma have this a - the result of an occupat ional asbestos exposure. Thus, ' her;- are 397 males and 13 females out of the 1000 malignant meso theliomas reported each year whose mesotheliomas were caused by occupational asbestos exposure. This is a slight upward revision of the 333 reported in 1981 (Table 111), since it is based on data not available in 1981. By subtraction, there are 353 males and 237 females whose mesotheliomas were caused by something other than occupational exposure to asbestos. What percent of this total 590 are caused by the average outdoor asbestos level of 1.5 ng/nv1 of air? There is evidence that fibers other than asbestos cause malignant mesotheno mas.15 Then? is also evidence that nonoccupational exposures to asbestos cause malignant mesotheliomas.20 The previously estimated 333 malignant mesotheliomas caused by nonoc cupational asbestos exposure--a little over half the 590 nonoccupationally caused malignant mesotheliomas0--is consistent with an estimate made by McDonald and Mc Donald in 1981.4 This will continue to be used as a best esti mate. Since the estimate of nonoccupational lung cancer shown in Table III was made, some assumptions in the study upon which it was based have been revised. Apparently, exposure 319 wmmmmmm :: i; " levels reported for that study were too low.21 That study was chosen because it was the only study for which historic expo sure data were available expressed in fibers, the currently acceptable method of reporting asbestos concentrations. Moreover, it appeared to be well known and well accepted, and was one of three selected for inclusion in the dose-response section of the report by the Advisory Committee on Asbes tos.22 It has now been reported, however, that the fiber ex posures in this study were probably about 3 times higher than originally thought. If so, this would make the estimate of lung cancer deaths in Table III around 9 or 10 rather than 28. This has little effect on the overall estimate of 361. In 1981, the effects of asbestos fibers emitted by hair dryers with asbestos-containing heat shields on lung cancer mortality were reported.16 This estimate was based on a linear extrap olation from a study of lung cancer among retired U.S. as bestos products workers.23 Figure 2 shows the dose-response relationship between asbestos concentration expressed as million part icles per cubic fix)t (mppof) and excess lung cancer expressed as Standardized Mortality Ratios (SMR). SMR are the ratio of observed to expected deaths times 100 and are a commonly used measure of excess deaths in epidemiologic studies. From these data, lung cancer deaths that would result from continuous exposure to very low' levels of airborne as bestos can he estimated. Some results from the 1981 report are shown in Table IV, expressed as lifetime risk of lung cancer resulting from con rej of For retired workers from a Johns Manville asbestos product1{ th tinuous asbestos exposure per million people exposed. In plant in the U.S., average worker exposure was 10 mppcf T iss preparing this table, asbestos concentrations expressed as If this is converted to f/cc by assuming 1 mppcf = 3 f/cc the ? tv) mppcf were converted to fibers per cubic centimeter (f/c-c) of air by assuming 1 mppcf = 3 f/cc. Also, as in Table III, 1 ng of (asbestos was assumed to contain 40 fibers of asbestos greater Vhan 5pm in length as counted by the optical method. The i rationale for these conversions appears in the 1981 report. this gives an average exposure of 30 f/cc. In a textile operatic | reported on by Nicholson, historic fiber counts were estimated at 80 f/cc.24 Tt seems likely, therefore, that exposures in the^ British study could have averaged 36 f/cc rather than 12 f/cc| and that the estimates of U.S. lung cancer deaths shown i| fo: Si! im ret Lifetime lung cancers were calculated assuming that lung cancer starts 20 yr after the beginning of exposure and that an average lifetime is 70 vr, so that the effective exposure Table II should be lowered by a factor of 3. ; One important assumption in Tables III and IV concerns the conversion of ng to fibers seen by the optical method * O'O v/1" sir period is 50 yr. The effect of hair dryers appears on the first Many fibers cannot be seen by the optical method. Moreover, Ti line followed hv the effect of all nonoccupational expo fiber size determines the conversion factor. The conversion WF sures. factor of 40, used in estimating nonoccupationally caused lung fo: cancer in Tables III and IV, is perhaps high. For example,* C X Table IV. Lifetime lung cancer deaths resulting from various levels of asbestos exposure per million exposed. Level of exposure (ng/m3) Kquivalent fibers/co 05 fnn long) Lung cancer deaths 1 1.5 5 50 500 5.000 50,000 0.000037 0.00006 0.0002 0.002 0.02 0.2 2 0.85 2,08 4.6 16,06 160.6 4,606 46.060 conversion factor of ,30 was used in EPA's water quality doc ument.25 Bruckman recommends 20.26 To the extent that tL conversion factor is high, it overstates lung cancer cases caused by nonoccupational asbestos exposure. It is difficult to judge the validity of the estimates show in Tables III and IV. There have been only a few attemptsto^ est imate the health effects of asbestos exposure on the general population. In 1977, Bruckman reported on the development of criteria for setting an asbestos air standard for the state of Connecticut and concluded that 30 ng/m3 would be a safe level.26 It was estimated that at this level, a minimum of 1^ mesotheliomas would be produced on a nationwide basis. This level was justified as one that results in about the same number of deaths as train mishaps and about l/10 the year!) fatalities from airplane accidents. Only a minimum number mnci ex an do do dv fo. ce ar of mesothelioma deaths was calculated on the grounds that tk For a U.S. population of 225 million. Table IV suggests 225 this would be difficult to criticize as being too strict. An upper limit can be calculated from the date and appears to be about po ex X 2.08, or 468 lifetime deaths from lung cancer due to 10,000. Extensive details on the manner in which these esti nonoccupational asbestos exposure at 1.5 ng/nv1, whereas 'Fable III suggests 28 X 70, or i960 lifetime asbestos lung mates were made are not given. In 1981, Schneiderman, et al. were commissioned by the K*' ex cancers related to nonoccupational asbestos exposure at 1.5 West German government to prepare a report on the risks by a mi ng/m3. If the exposure estimate upon which Table III was exposure to low levels of asbestos in the general environ-1 or based are off by a factor of 3, however, these estimates are ment.27 Their report provides a basis for calculating the lit*- et much closer (653 vs. 468). For malignant mesothelioma. Table time risk of asbestos exposures at low levels, provided son* Iff suggests a lifetime risk of 333 X 7(1 or 2.3,310 cases for a LLS. conversion of ng to fibers can be made. The concluding table ex population of 225 million. in their report, labeled "Risk assessments for the general w] It is plausible that the original estimates of fiber exposure population ...", apparently is related to the risk for a working m: in the British study of asbestos textile workers were indeed too low. 'Flic British study estimated an average fiber count population exposed 8 h/dav, five days/week. These risks can; in he converted to continuous exposure, however, by dividing by | ea Ii lor asbestos textile operations of around 12 f/cc [trior lo 1950,17 0.24, since working time is 24% of total time. It can then be U:.: f k320 Journal of the Air Pollution Control Association j A; > calculated that the lower limit of lifetime risk of lung cancer for a population exposed to 1 f/cc for ! yr is 0.0015. The lower limit of risk for exposure to 1.5 ng/m:i with 36 yr effective ex posure (assumed by Schneiderman, et al.'r`) can be calculated if ng are converted to fibers. Using the conversion factor of 40 gives0.00006 f/cc. The risk is therefore (0.(*0006)(36H0.001f>), or 0.00000324. Since the U.S. population is about 225 million this means a lifetime excess of 729 lung cancer deaths due to nonoccupational asbestos exposure (0.00000324)(225 million). Tables in the Schneiderman, et al. report also show upper limits, with the upper limit of their lung cancer estimates simply 10 times the lower. Therefore, the upper limit of lung cancer deaths due to nonoccupational asbestos exposure is 7290. These estimates of 729 and 7290 contrast with the above estimates of 428 and 653. The Schneiderman, et al. report also permits an estimate of lifetime risk for malignant mesotheli oma. This is 0.0033 (lower limit) for 1 yr exposure at 1 f/cc. For exposure at 1.5 ng/m3, assuming 20 yr effective exposure (assumed by Schneiderman, et al.), this is a lifetime lower limit of 891 cases. The upper limit is 5,400 cases. This con trasts with a lifetime excess of 70 X 33;), or 23,310, calculated from Table III. / One of the reasons for the higher lung cancer estimates in the Schneiderman, et al. report is that they found reasons to reject the results of four epidemiologic studies which gave risks of less than 1% for 1 yr exposure at 1 f/cc. This included all three studies used in the dose-response section of the report issued by the Advisory Committee on Asbestos.'2* Rather, they relied on dose-response data compiled in 1981 by Nicholson, some of which were previously unpublished.2'* Of the five studies they found acceptable, three were from Selikoffs laboratory in New York City, one of which yielded a dose- response relationship based on only 2 lung cancer deaths.24 Neither of the others was designed to study dose-response. For one, only duration of exposure was measured for each worker, while for the other, workers could only be subdivided by time since first exposure, since duration of exp< >sure was unknown. The fourth study28 provides very few data on exposure and was described by Nicholson as containing "only sketchy in formation".24 The authors of that study only discuss "possible exposure levels" with no details on how these were deter mined. The fifth study, a study of asbestos textile workers by Dement, et al., appears to be a good one with well documented exposure levels.*-' Numbers of lung cancer cases were relatively small, however, and the results of this st udy are quite unlike anything observed elsewhere. Lung cancer response per unit aose was very high. It was 88 times greater than a similarly documented but much larger study of asbestos miners/10 Discussion The above calculations lead to estimates of lifetime risks due to nonoccupational asbestos exposure of 100 per million fur malignant mesotheliomas and 2 per million for lung can(ers. This is a fifty- fold difference in favor of mesothelioma and is quite unlike contrasts observed in studies of occupa tionally exposed populations. In occupationally exposed populations, the excess favors lung cancer, with the ratio of excess lung cancer to mesotheliomas around 2 or 3.4 Using a different set of assumptions, the Schneiderman, et al. study gives estimates of a lifetime excess due to nonoccupational exposure of 4-24 per million for mesothelioma and 3 30 per million for excess lung cancers.27 'This also differs from the 2 or 3 ratio of lung cancers to mesotheliomas in occupationally exposed populations. Is it reasonable to suppose that nonoccupational asbestos exposures produce more mesotheliomas than lung cancers while occupational exposures produce more lung cancer than mesotheliomas? One reason for an excess of mesotheliomas in nonoceupationally exposed populations may be the much earlier age at first exposure and an unusual mesothelioma risk associated with exposure at an early age. This is because the April 1983 Volume 33, No. 4 process by which asbestos produces mesotheliomas seems to be different from the process by which asbestos causes lung cancers. I'eto, et at. feel that in the production of mesotheli omas, asbestos acts at an early stage in a multistage process believed to produce cancer.:u As a result, they feel that mesotheliomas are manifest only many years after first ex posure, and that where exposure to asbestos starts at an early age, mesotheliomas will constitute a high proportion of as bestos related cancer deaths. On the other hand, in the pro duction of asbestos lung cancers there is considerable evidence to suggest that asbestos acts at a late stage in a multistage process. As a result, lung cancers occur more quickly than mesotheliomas, so that even where exposure starts at age 30 or 40 there is time for lung cancers to be manifest before deaths due to other causes intervene. Mesotheliomas may, therefore, be the predominant asbestos cancer related to nonoccupational exposures simply because of the early age at which these exposures start, whereas lung cancer is the predominant cancer in occupationally exposed populations due to the older age at which exposure starts and a lack of t ime for mesotheliomas to be fully expressed. Whether this is a sufficient explanation for a fifty-fold ex cess in mesotheliomas relative to lung cancers may lie ques tioned. Given the well established role of asbestos in the eti ology of mesotheliomas, and the exposure of nearly the entire U.S. population to asbestos, it seems likely that nonoccupa tional exposure causes at least a third of the mesotheliomas that occur in the U.S. each year and perhaps more. Therefore, the estimate of a lifetime risk of 100 per million due to nonoccupational asbestos exposure seems conservative and the true risk may, in fact, he higher than this. On the olher hand, the estimate of a lifetime risk of 2 per million for lung cancer could he low The dose response relationship may not he linear at low doses, as supposed, or perhaps the response per unit dose is greater than was assumed. No one has seri ously proposed a dose response relationship that is curvilinear downward, as would he needed to increase response ai lowdose. There are, however, studies that give a larger response for lung cancer per unit dose than the one used. The highest re sponse yet reported was in the study by Dement, et al. M If this had been used, instead of the study of retired asbestos work ers,*8 then the lifetime risk of asbestos lung cancers resulting from nonoccupational asbestos exposure would have been 40 per million instead of 2. This reduces the apparent mesothe lioma-lung cancer discrepancy but does not eliminate it. T here does not appear to be any reasonable set of assumptions t hat would yield, for the nonoceupationally exposed population, the 2-3 excess lung cancers per mesothelioma observed in occupationally exposed populations. The issue, therefore, doe's not appear to he whether the ratio of mesothelioma: to lung cancers caused by background levels of asbestos exceed.- I, but rather, by how much it exceeds 1. References 1. 1'. Sebastleu, M. A. Billon, G. Dufour, A. Gaudichet,C B.mnuud. 4. Bignon, "Levels of asbestos air pollution in some envir - am it it al situations,''Ann /V Y., Acad. Nri, 3.40: 41)1 (1979). 2. W. 4. Niefiofson, A. N. Hohl, It. N. Sawyer. K. 4. Swos -ov.ski, 4. I). Totlaro, `'Control ot Sprayed Asbestos Xuriaces m School Buildings: A Feasibility Study," Report to the National Institute of Bnvironmental Health Sciences, New York, -June I .7. I 97 8. 3. If. 4, Thompson, G. B. Morgan, "Determination of A: be~ins in Ambient Air," in I'roceedintts of Internationa! Symposium on Identifieatiun and Measurement of Environmental Putin hints, I). Hoffman, ed., National Research Council of Canada, ( Mtawa, Ontario, 1971. 4. 4, C. McDonald, A 1). McDonald, "Mesothelioma as an Index of Asbestos Impact," in liunhury Report 9, Quantification oj Oc cupational t'anver R. I'eto, M, Schneiderman, (ids., ('old Spring Harbor Laboratory, Coid Spring Harbor, NY, 1981, p. 73. ,7 4. Higginson, "Proportion of cancer due to occupati in." Pre ventive Medicine, t): 180 (1980). (I. I'. K. Knterline, "Proportion of Cancer Due to Kxposurc In As bestos," In Hanhury Report 9, Quantification of (h cupa'. nmul 321  Cancer R. Peto, M. Schneiderman, eds., Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, 1981, p. 19. 7. W. J. Nicholson, G. Perkcl, I. J. Selikoff, H. Seidman, "Cancer from Occupational Asbestos: Exposure Projects 1980-2000," in Banbury Report 9, Quantification of Occupational Cancer R. Peto, M. Schneiderman, eds., Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, 1981, p. 87. 8. M. D. Hogan, D. G. Hoel, "Estimated cancer risk associated with occupational asbestos exposure," Risk Analysis 1: 67 (1981). 9. K. Bridbord, P. DeCoufle, J. F. Fraumeni, D. G. Hoel, R. N. Hoover, D. P. Rail, IJ. Saffiotti, M. A. Schneiderman, A. C. Upton, "Estimates of the Fraction of Cancer in the United States Related to Occupational Factors," National Cancer Institute, Washington, DC, September 1978. 10. 1. Selikoff, Generic Cancer Pol icy Hearings, U.S. Occupational Safety and Health Administration, Washington, DC, June 1, 1978, testimony. 11. "Health United States," publication no. (1`HSl 80-1232, U.S. Dept, of Health, Education, and Welfare, 1979. 12. R. Doll, R. Peto, The Causes of Cancer, Oxford University Press, Oxford, England, 1981. 13. J. D. Gillam. J. M. Dement, Ft. A. I,emen, J. K. Wagoner, V. E. Archer, H. P. Blejer, "Mortality patterns among hard rock gold miners exposed to an asbestiform mineral," Ann. N. Y. Acad. Sci., 271:336 (1976). 14. J. C. McDonald, G. W. Gibbs, F. D. K. Liddell, A. D. McDonald, Mortality after long exposure to cummingtonite-grunerite, .4m. Rev. Respirat. Diseases 118:271 (1978). 15. S. D. Kaplan, W. R. Gaffey, "Miners Exposed to Amphibole Mineral: A Retrospective Cohort Mortality Study," report: Na tional Institute for Occupational Safety and Health, Cincinnati, 011,1981. ' 16. P. E. Enterline, "Extrapolation from occupational studies: a substitute for environmental epidemiology," Environ. Health Perspectives 42: 39 (1981). 17. J. Peto, R. Doll, S. V. Howard, L. J. Kinlen, H. 0. Lewinsohn, "A mortality study among workers in an English asbestos factory," Br .1 hid Med. 31: 169 (1977). ` 18. A. D. McDonald, J. C. McDonald, "Malignant mesothelioma in North America," Cancer 40: 1650 (1980). 19. Y. I. Baris, M. Artvinli, A. A. Sahin, "Environmental mesotheli oma in Turkey, health hazards of asbestos exposure," Ann. N Y. Acad. Sci. 330: 423 (1979). 20. J. C. Wagner, C A. Sleggs, P. Marrhan, "Diffuse pleural meso thelioma and asbestos exposure in the Northwestern Cape Province," Br. J. hid. Med. 17: 260 (1960). 21. J. Peto, "Lung Cancer Mortality in Relation to Measure Dust Levels in an Asbestos Textile Factory," in Biological Effects of Mineral Fibers, Voi. ii, J. C. Wagner, ed.. International Agency for Research on Cancer, Lyon, France, 1980, p. 829. 22. Advisory (lommittee on Asbestos, report to the Health & Safety Commission, Her Majesty's Stationery Office, London, 1979. 23. V. L. Henderson, P. E. Enterline, "Asbestos exposure: factors associated with excess cancer and respiratory disease mortality," Ann. N. Y. Acad. Sci. 330: 117 (1979). ' 24. W. J. Nicholson, "Dose-Response Relationships for Asbestosai Other Inorganic Fibers," Report to the National Institute | Environmental Health Sciences, February 15, 1981. j 25. "Ambient Water Quality Criteria for Asbestos," U.S. Environ mental Protection Agency, Office of Water Regulations aK: Standards, October 1980. : 26. L. Bruckman, "A Study of Airborne Asbestos Fibers in Con necticut," in Proceedings of the Workshop on Asbestos: Defi'- y tions and Measurement Methods, National Bureau of Standard- 1978, pp. 179-190. 27. M. S. Schneiderman, I. C. T. Nisbet, S. M. Brett. "Assessmeii of Risks PoBed by Exposure to Low Levels of Asbestos in tb General Environment," in Health Hazards Posed by Asbesto K. Aurand, W.-S. Kierski, eds., Federal Bureau of Health, Fedefi Republic of Germany, 1981. j 28. M. L. Newhouse, G. Berry, "Patterns of disease among long-tern! asbestos workers in the United Kingdom," Ann. N.Y. Acad. Sci| 330:53 0 979). j 29. J. M. Dement, R. L. Harris, M. J. Symons, C. Shy, Estimates ol;. Dose-Response for Respiratory Cancer Among Chrysotile As f bestos Textile Workers," Ann. Occupational Hyg., to be pubjy lished. i1 30. J. C. McDonald, F. D. K. Liddell, "Mortality in Canadian mines | and millers exposed to chrysotile," Ann. N.Y. Acad. Sci. 330:D (1979). _ _, 31. J. Peto, H. Seidman, I. J. Selikoff, "Mesothelioma mortality it asbestos workers: implications for models of carcinogenesis and risk assessment," Br. J. Cancer 45:124 (1982). Tra age Sir. per: co. ( rat: dltf lea arc spr par tra Dr. Enterline is a Professor and Chairman, Department of Biostatics, and Associate Director, Center for Environmental Epidemiology, Graduate School of Public Health, University of Pittsburgh, Pittsburgh, PA 15261. This paper was submit ted for editorial review on June 2, 1982; the revised manu script was received on February 14, 1983. / Fn CO! fo, no: syr yle civ. a:; Hi po st h a!? to: 18 tor. nr, by te; m: CD th; zo: pc wi a: ch wv I P" I 322 Journal of the Air Pollution Control Association A -