Document NGjB7vVd06OxXOLLpXJ69wa2g
Asbestos Exposure and Laryngeal Cancer: An Analysis of the Epidemiologic Evidence
Charles E. Chan, MD; and J. Bernard L. Gee, MD
We conducted a critical analysis of the available epidemio
regard to other known potent carcinogenic factors, spe
logic investigations on the causal relationship between asbes tos exposure and laryngeal cancer. A review of pine casecontrol studies indicates that the estimated risk (odds ratio) attributable to asbestos exposure alone is negligible when smoking and ethanol intake are appropriately controlled for. Six of the 18 cohort studies demonstrated no significant in crease in the standardized mortality ratio due to asbestos exposure. The remaining six longitudinal studies showed an
cifically cigarette smoking and ethanol consumption, and second that the evidence implicating asbestos is minimal or absent. Aside from the inappropriateness of the oft-quoted conclusions and some misplaced litiga tion, one result may be that physicians do not always advise asbestos workers on the proper personal preven tive measures that they should take.
increased standardized mortality ratio from 1.91 to 5.41 but
no adjustment was made for the confounding effects ofsmoking
and ethanol consumption. In conclusion, the available epide
Effects of Smoking and Alcohol
miologic evidence does not support a causal association be
tween asbestos exposure and laryngeal cancer.
There has been a gradual increase in laryngeal cancer
rates in the United States over the last four decades.14
Asbestos, in spite of its technological and economic value to society at large, has been among the great
Part of the increase may be attributable to more accu rate reporting of the disease or increased detection14
est occupational hazards of this century.1 Asbestos has but smoking and alcohol drinking have been consistently
been clearly shown to be responsible for many cases of
shown to be the two most important environmental
lung cancer, mesothelioma, and also for asbestosis.2^ 6^
factors for this neoplasm.14 20 The dose-response rela
11 Asbestos has also been inculpated in other malignan
tionships which support causality between smoking, al
cies, including laryngeal tumors.3-11 The notion that
cohol. and laryngeal cancer (Table 1) have been ele
asbestos could cause laryngeal cancer is suggested both
gantly demonstrated by Wynder et alis and Rothman et
by limited evidence of asbestos fibers being present in
al.19 Flanders and Rothman20 also reported a synergistic
laryngeal tissue12'13 and by the larynx being part of the
effect between smoking and alcohol consumption for
respiratory tract. However, the larynx is, unlike lower
laryngeal cancer; the interaction being such that a one
airways, lined by squamous epithelium. We offer here a
pack-a-day smoker who also consumes a moderate
review of this latter relationship, which will suggest
amount of alcohol (more than 6 ounces of liquor per
first that much of the evidence derives from epidemio
day) has a greater than twenty-fold increased risk of
logic studies which have simply examined the relation
developing laryngeal cancer.1820 A particularly detailed
between asbestos and laryngeal cancer without any
study by Elwood et alzl reported on 374 cases of head
and neck cancer and specifically separated the sites to
include data on extrinsic and instrinsic laryngeal can
cers. They found that both smoking and alcohol were
From the Pulmonary Section, Department of Medicine, and the Robert Wood Johnson Clinical Scholars Program. Yale University
separate risk factors for 46 extrinsic cancers, whereas smoking was the significant factor for 108 intrinsic
School of Medicine, 383 Cedar St, New Haven, CT 06510 (address correspondence to Dr Gee, Professor of Medicine and Director, Win chester Chest Clinic).
0096-1736/86/3001-0023802.00/0 Copyright by American Occupational Medical Association
laryngeal cancers. These workers were not able to dis tinguish between additive and multiplicative interpre
tations of the data for alcohol and smoking. All these studies demonstrate the necessity of adjusting for smok-
Journal of Occupational Medicine/Volume 30 No. 1/January 1988
23
TABLE 1 Relative Risk Estimates tor Laryngeal Cancer"
Relative Risk
Smoking (cig<irettes/day) 0 1-15 16-34
35+ Alcohol intake (oz liquor/
day) <1.5
1.5-9 >10
adjusted for alcohol 1.0 4.6
11.0 14.0 adjusted for smoking
1.0 1.2 2.2
* Modified from Wyttder et al'6 and Rothman et ai.'9
ing and ethanol consumption when delineating other risk factors for laryngeal cancer. The need for such adjustment is underscored by the high prevalence (>90%) of smoking, frequently heavy, among asbestos workers.82
Effects of Asbestos
The available evidence on the causal relationship
between asbestos and laryngeal cancer in the English
literature can be divided into three main categories:
retrospective case-control design,2183-30 cohort or Ion-
gitudinal follow-up survey,
and case senes.
Case-control Studies
The pertinent results of the nine case-control studies21'83-30 are summarized in Table 2 and are consid ered in two groups.
Studies Adjusted for Smoking' and Alcohol. There were five well-executed case-control studies23-27 that adjusted the crude estimates of relative risk (odds ra tios) for cigarette smoking and alcohol consumption (Table 2).
Hinds et al23 stratified their cases and controls and according to cigarette and alcohol consumption and found no increase in risk from asbestos exposure alone. An impressive and statistically significant dose-depend ent increase in the odds ratios comparable to that reported in the general population16,20 was demonstrated for both cigarette and alcohol intake in the asbestos population. The unadjusted odds ratio for laryngeal cancer associated with asbestos exposure was a statis tically insignificant value of 1.75 (Table 2). When ad justments were made for cigarette and ethanol con sumptions,23 the ratio became one!
A similar but smaller case-control study by Blot et al24 showed (Table 2), after adjustment for cigarette smoking, no significant increase in the odds ratio among asbestos workers for either black or white persons (Ta ble 2). In spite of small numbers and with their inherent errors in relative risk estimates, clear relations between laryngeal cancer and both smoking and alcohol con sumption were again observed.24
In a search for "occupational causes" of laryngeal
TABLE 2 Case-control Studies on Asbestos and Laryngeal Cancer
Authors
Elwood et al21 Hinds et al23 Blot et al2*
Blacks Whites Olsen and Sarbroe25 Burch et al29 Shettigara and Morgan27 Morgan and Shettigara28 Stell and McGill29 Stell and McGill30
No. of Cases
154 47
11 2
17 14 43
54t
100 119*
Odds Ratio
_
1.75
2.4 0.23 1.8 1.6
X
13
14.5 14.8
P
NS' NS
NS NS NS NS <.001
<.001
<0bl <.001
NS, not significant. t Of the 54 cases, 43 were reported in Ref 27. * Of the 119 cases, 100 were reported in Ref 29.
cancer, Olsen and Sarbroe23 found a slightly increased odds ratio for asbestos exposure (Table 2). However, since the 95% confidence limit for the odds ratio when adjusted for tobacco and alcohol consumption was 1.0 to 3.4, the conclusion is moot.
The case-control study of Burch et al26 on 204 cases of newly diagnosed laryngeal cancer in Ontario, Canada, again showed strong associations between tobacco prod ucts and alcohol. A subanalysis of 14 cases and 9 controls with definite asbestos exposure showed a statistically insignificant odds ratio of 1.6 (Table 2).
The Elwood study81 referred to earlier also failed to detect an influence of occupational factors including asbestos, but the numbers of asbestos workers were small.
Unadjusted Studies. Four case-control studies report significant increases in the odds ratio for laryngeal cancer due to asbestos exposure (Table 2). Lamentably, these studies lack adjustments for either smoking or ethanol consumption. Shettigara and Morgan27 report an odds ratio of infinity for asbestos exposure in their original study of 43 cases with laryngeal cancer because none of their controls had an asbestos exposure. In a follow-up study, Morgan and Shettigara28 added 11 cases and found an odds ratio of 13 for asbestos exposure and laryngeal cancer (Table 2). In spite of their own observed striking dose-response relationship for smok ing and laryngeal cancer in both of their studies,27'28 the authors made no adjustments but they indicated that "the risk imposed by asbestos is confined to smokers"28 since no case of laryngeal cancer was found among nonsmokers with asbestos exposure only.27'28
Stell and McGill28'30 twice reported the increase in odds ratio due to asbestos exposure in a contrived population of 119 patients with laryngeal cancer to be about 15. The crucial information on the smoking habit of a subgroup of 33 cases with asbestos exposure was not available. Although the authors pointed out that smoking might be a "cofactor in the development of laryngeal carcinoma,"29 they made no adjustment since they believed that "smoking is unlikely to be very im portant."29 Ethanol was not considered.
Thus these four studies27-30 are seriously flawed.
24 Asbestos Exposure and Laryngeal Cancer/Chan & Gee
HWBUI0001170
Cohort Studies and Longitudinal Surveys
A number of retrospectively or prospectively assem bled exposed asbestos populations have been carefully observed for an average of 20 or more years.3"11'31"34 These longitudinal studies, in contrast to the casecontrol studies discussed earlier, provide a more precise assessment of exposure, including dose, time, and la tency, thus permitting a more accurate estimate of the risk attributable to asbestos. Also, the confounding ef fects due to biased interviews or questionnaires, and the "recall bias" provided by the interviewer33 frequently encountered in case-control studies are avoided in these prospective investigations. Such studies, if performed
prospectively, would take at least 20 years, are very expensive, and require cohorts of several thousands of exposed individuals since laryngeal cancers14 sure rela tively infrequent. Their results are summarized in Table 3.
In the birth cohort of 11,379 miners exposed to chrysotile fiber in Quebec, McDonald et al3 found no excess deaths from laryngeal cancer (Table 3). However, a subanalysis of deaths due to laryngeal cancer and smok ing revealed a convincing dose-response relationship. Specifically, the standardized mortality ratio (SMR) in nonsmokers was 0.46; in moderate smokers the SMR was 0.93, and in heavy smokers the SMR went up to 4.85. Similar linear exposure-response relationship was also found for asbestos with lung cancer and for pneu moconiosis.3
A cohort of 2,543 textile plant workers exposed to chrysotile in South Carolina was also studied by Mc Donald et al.4 Life table analyses and "log-rank" statis tical comparisons both indicated a much steeper expo sure-response slope for lung cancer in the textile work ers than that observed in miners.3 The overall SMR was also elevated to 1.27.4 However, over the 20 years, the major causes of death in these textile workers were lung cancer, asbestosis, and abdominal malignancies, whereas only three deaths were due to laryngeal cancers (Table 3).
In a follow-up survey of 1,970 workers employed at
an asbestos cement factory with predominant exposure to chrysotile, Thomas et al3 failed to find a single case of cancer of the larynx. However, over three quarters of this papulation had only worked in this factory for under 4 years.5
Hodgson and Jones reviewed the British national survey of over 30,000 asbestos insulation workers in 1981. The length of time after exposure was highly variable in this cohort. Although they observed an SMR of 2.56 for lung cancer, no excess deaths from laryngeal cancer occurred.
In a quasi-cohort survey, Botha et al7 calculated the SMRs in South African crocidolite mining districts from 1968 to 1980. Eleven contiguous districts with little crocidolite exposure were used as controls. In spite of elevated SMRs for asbestosis, mesothelioma, and cancer of the lung and stomach, no excess of laryngeal cancer (four deaths) occurred in this group of 18,000 persons of mixed race.
In a survey of cancer mortality between 1950 and 1969 in 49 US counties where shipyards were engaged in the construction and repair of large naval and cargo vessels during World War II, Blot et ala found "no upward trends in mortality for laryngeal cancer." The age-adjusted mortality rates for those counties in that period among whites was 1.18 for men and 1.16 for women. The obvious limitation in this type of large population survey is that major changes in the incidence of disease are needed before there is any noticeable increase in the relative risk in huge populations. There fore, one cannot use these results to exclude confidently any causal relationship.
Selikoff et al9 reported on a contrived cohort of 17,800 insulation workers in the United States and Canada. The length of time after exposure was also variable, ranging from less than 20 years to well over 35 years from onset of employment. An increase in the overall SMR to 1.37 was observed, and the major contributors to this increase in SMR were lung and abdominal can cers and asbestosis. Laryngeal cancer was reported in nine deaths, which elevated the SMR to 1.91 (Table 3), but no adjustment was made for cigarette or ethanol consumption.
TABLE 3 Cohort Studies on Asbestos and Laryngeal Cancer
Authors
Cohort Size
Years after Exposure
No of Cases of Laryngeal Cancer
McDonald et al3 McDonald et al` Thomas et al5 Hodgson and Jones3
Botha et al7 Blot et al8 Selikoff et al5 Puntoni et al' Puntoni et al" Newhouse3' Newhouse et al33 Rubino et al33
11.379 2,543 1,970
31.150 18,278
--
17.800 2,348 34.746T 4.000 5,100t
900
>29 >20 Variable Variable Variable Variable Variable >10 Variable >20 >20 Vanabte
16 3 0
--
4
--
9 8 15 2 3 6
RR, relative risk. t Person-years of observation. t Of the 5,100 workers, 4,000 were included in Ref 31.
SMR/nn*
1.07
--
--
-- --
1.16-1.18 1.91 1.57 1.96 5.41 3.70 3.16
Journal of Occupational Medicine/Volume 30 No. 1 /January 1988
25
Puntoni et al`'" analyzed the causes of deaths among the shipyard workers in Genoa, Italy. The first report in 197710 included 2,348 dockyard workers observed for 10 or more years after asbestos exposure for a total of 22,188 person-years of observation. The 1.57 relative risk for laryngeal cancer was not statistically signifi cant. In the follow-up report in 19791! of several thou sand dockyard workers employed between I960 and 1970, the investigators noted 15 deaths due to laryngeal cancer with a corresponding relative risk of 1.96. Their conclusion that asbestos exposure was a risk factor for laryngeal cancer is vitiated by failure to control for smoking and alcohol consumption.
Both studies of Newhouse et al31'32 (Table 3), although useful in other ways, contain too few cases of laryngeal cancer (two and three cases) to provide stable estimates of the relevant SMR.
The report by Rubino et al33 on the mortality of more than 900 North Italian chrysotile miners is particularly interesting. They report six deaths from laryngeal can cer (SMR of 3.16). Although all these six miners were smokers, no adjustment of the SMR was made for smok ing. Surprisingly, no excess deaths from asbestosis or lung cancer were found, but th*' SMRs for nonmalignant respiratory diseases and cirrhosis were, respectively, 2.53 and 3.13. One may well ask what were the smoking and drinking practices of these miners?
Why should 12 cohort studies3*1131~33 arrive at SMRs or relative risk that ranged from 1.07 to 5.41? Kven if we exclude Newhouse's data31'32 because of the small number of cases and discard Rubino's anomalous results33 the discrepancy in the SMR of 1.07 in Mc Donald's study3 and 1.91 in SelikofFs cohort9 and the conflicting results of Puntoni's two studies1011 needs comment. Aside from the failure to address the effects of smoking and ethanol consumption, two other potential confounding variables are the variable length of follow up38 and differences in the types and size of asbestos fibers exposed in these cohorts.37'38
Indeed, if we reanalyze Selikoffs data9 for their older cohort of 623 insulation workers with 20 or more years after onset of work, the SMR in this subgroup was 1.18.9 Since the average follow-up in McDonald's cohort3 was over 20 years, the adjusted SMR of 1.18 in the "older" Selikoff cohort would be a more appropriate comparison, and supports the lack of association between asbestos exposure and laryngeal cancer shown in the case-control studies.23"97 Furthermore, there are substantial differ ences in the occurrence of lung cancer among workers in various asbestos industries.4'8 Such differences may reflect variation in the type, quantity, and size of the asbestos fiber in the various trades.3"11'31'33"34
Case Series
Conclusion
In sharp contrast to other incontrovertible hazards of asbestos,12 the foregoing evidence inculpating asbes tos in laryngeal cancer is weak or non-existent.3-13 21 34 With a few exceptions,9"1127 33 discussed above, the SMRs for laryngeal cancer in asbestos workers are all small.3"8-23"2
The extraordinarily high odds ratios found by Shettigara and Morgan27-28 and Stell29'30 were not repro duced by the perhaps more accurate and somewhat less biased33 SMR estimates obtained in the cohort studies.3" u.31-33 other than the major flaw of not adjusting for the effects of smoking and ethanol, the odds ratios in these studies27"30 might also have been significantly confounded by the interviewer and interviewee bias, or ascertainment bias,33 which is an innate weakness of case-control methodology.
If only 10% of an asbestos worker population were to smoke more than one pack a day or to smoke one pack a day and drink 6 oz of ethanol a day, then the SMR for laryngeal cancers in the whole population would ap proach three times that of a nonsmoking, nondrinking population. Although this is a theoretical estimate, it does highlight in numerical terms the limited value of the unadjusted SMRs. This same effect makes it impos sible to exclude totally some interactive effects between smoking and asbestos in laryngeal cancer. However, we know of no case in the foregoing literature of asbestos workers with laryngeal cancer who did not smoke and there are no published data on their ethanol intake.
Other considerations include socio-economic factors and the beneficial influence of tea drinking21--the latter perhaps by the absence of other factors! An additional variable rarely considered in the asbestos cancer liter ature is diet, specifically the content of vitamin A and related materials. An inverse relation between vitamin A intake and lung cancer incidence has been reported in other populations.39 Nutritional status may conceiv ably be important in the pathogenesis of the laryngeal cancer40 because retinols have been shown to diminish asbestos-induced tracheal metaplasia in experimental animals.41 A preliminary study42 reports low serum i3carotene in asbestos workers. Dietary factors have not been otherwise examined in the asbestos worker, and this relationship to laryngeal cancer has not been crit ically assessed.
It is improbable that further evidence will be available on this asbestos-laryngeal cancer issue since the pre ventive measures are now generally in place, at least ir most Western society. However, we suggest that the foregoing evidence will focus preventive measures or life style factors.
Libshitz et al34 reported laryngeal cancer in three persons with asbestosis and so proposed a possible as sociation between asbestosis and laryngeal cancer. How ever, it is worth noting that all three patients were cigarette smokers!34
Acknowledgements
The authors wish to thank Dr Aivan E. Felnstein for reviewing thi manuscript and Ms Lynn Stiiimank for her assistance in preparim this manuscript.
26 Asbestos Exposure and Laryngeal Cancer/Chan & Get
HWBUI0001172
References
1. Beckiake MR: Asbestos-related diseases of the lung and other organs: their epidemiology and implications for clinical practice. Am Rev Respir Dis 1976:1X4:187-826.
2. Selikoff IJ. Hammond EC: Health hazards of asbestos exposure. Ann NY Acad Sci 1979;330:1-814.
3. McDonald JC. Liddell FDK, Gibbs GW. et al: Dust exposure and mortality in chrysotile mining, 1910-75. BrJlad Med 1980:37:1124.
4. McDonald AD. Fry JS, Woolley AJ. et al: Dust exposure and mortality in an American chrysotile textile plant. Br J lad Med 1983:40:361-367.
5. Thomas HF. Benjamin IT, Elwood PC. et al: Further follow-up study of workers from an asbestos cement factory. Br J lad Med 1982:39:273-276.
6. Hodgson JT. Jones RD: Mortality of asbestos workers in Eng land and Wales 1971-8L. Br J Ind Med 1986:43:158-164.
7. Botha JL, Irwig LM, Strebel PM: Excosb mortality from stom ach cancer, Lung cancer, and asbestos and/or mesothelioma in crocidolite mining districts in South Africa. Am J Epidemiol 1986:123:3040.
8. Blot WJ. Stone BJ, Fraumeri JF Jr. et al: Cancer mortality in US counties with shipyard industries during World War II. Environ Res 1979:18:281-290.
9. Selikoff IT, Hammond EC. Seidman H: Mortality experience in insulation workers in the United States and Canada 1943-1976. Ana AT Acad So* 1979;330:91-116.
10. Puntoni R. Russo L, Zarmini D. et al: Mortality among dockyard workers in Genoa, Italy. Tumori 1977;63:91-96.
11. Puntoni R, Vercelli M, Merlo F, et ai: Mortality among shipyard workers in Genoa, Italy. Aaa NY Acad Sci 1979:330:353--377.
12. Hirsh A, Bignon J, Sebastian P. et al: Asbestos fibers in laryngeal tissues. Chest 1979:76:697-699.
13. Roggli VL, Greenberg SD, McLarty JL, et ai: Asbestos body content of the larynx in asbestos workers. Arch Otolaryngol 1980:106:533-535.
14. Conn Cl, Fried MP: Determinants and prognosis of laryngeal cancer. Otolaryngol Clin North Am 1984;17:139-150.
15. Vincent RG, MarchettaF: The relationship of the use of tobacco and alcohol to cancer of the oral cavity, pharynx or larynx. Am J Surg 1963;106:501-505.
16. Wynder EL, Covey LS, Mabucki K, et ai: Environmental factors in cancer of the larynx. Cancer 1976;38:1591-1601.
17. Herity B. Moriarty M. Daly L. et al: The role of tobacco and alcohol in the aetiology of lung and larynx cancer. Br J Cancer 1982;46:961-964.
18. US Public Health Services: The Health Consequences of Smok ing. Cancer. A Report of the Surgeon General. Rockville. MD. 1982. pp 63-169.
19. Rothman KJ. Cann Cl. Flanders D, et al: Epidemiology of laryngeal cancer. Epidemiol Rev 1980;2:195-209.
20. Flanders WD, Rothman KJ: Interaction of alcohol and tobacco in laryngeal cancer. Am J Epidemiol 1982;115:371-379.
21. Eiwood JM. Pearson JCG, Skippen DH, et al: Alcohol, smoking, social and occupational factors in the aetiology of cancer of the oral
cavity, pharynx and larynx. Int J Cancer 1984:34:603-612.
22. Berry G, Newhouse ML. Antonis P: Combined effect of asbestos and smoking on mortality from lung cancer and mesothelioma in factory workers. Br J Ind Med 1985;42:12-18.
23. Hinds MW, Thomas DB, O'Reilly HP: Asbestos, dental x-rays, tobacco, and alcohol in the epidemiology of laryngeal cancer. Cancer 1979:44:1144-1120.
24. Blot WJ. Morris LE. Stroube R. et al: Lung and laryngeal cancers in relation to shipyard employment in coastal Virginia. J Natl Cancer Inst 1980;65:571-575.
25. Olsen J. Sarbroe S: Occupational causes of laryngeal cancer. J Epidemiol Community Health 1983;38:117-121.
26. Burch JD, Howe GR, Miller AD. et ad: Tobacco, alcohol, asbestos and nickel in the etiology of cancer of the larynx: A case-control study. J. Natl Cancer Inst 1981:67:1219-1224.
27. Shettigara PT, Morgan RW: Asbestos, smoking and laryngeal carcinoma. Arch Environ Health 1975:30:517-519.
28. Morgan RW. Shettigara PT: Occupational asbestos exposure, smoking and laryngeal carcinoma. Ann NY Acad Sci 1976:271:308310.
29. Steil PM. McGill T: Asbestos and laryngeal carcinoma. Lancet 1973:1:416-417.
30. Steil PM. McGill T: Exposure to asbestos and laryngeal carci noma. J Laryngol Otol 1975;89:513-517.
31. Newhouse ML: A study of the mortality of workers in an asbestos factory. Br J Ind Med 1969:26:294-301.
32. Newhouse ML, Berry G. Wagner JC: Mortality of factory workers in east London 1933-80. Br J Ind Med 1985;42:4-11.
33. Rubino GF. Piolatto G, Newhouse ML, et al: Mortality of chrysotile asbestos workers at the Balangero mine. Northern Italy. Br J Ind Med 1979:36:187-194.
34. Libshitz HI, Wershba MS, Atkinson GW, et al: Asbestosis and carcinoma of the larynx. JAMA 1974;228:1571-1572.
35. Feinstein AR: Ascertainment of data for focal variable, in Feinstein AR (ed): Clinical Epidemiology. Philadelphia, W. B. Saun ders Co, 1985, pp 503-511.
36. Browne K: A threshold for asbestos-related lung cancer. Br J Ind Med 1986:43:556-558.
37. Liddell FDK. Hanley JA: Relations between asbestos exposure and lung cancer SMRs in occupational cohort studies. Br J Ind Med 1985:42:389-396.
38. Henderson VL. Enterline PE: Asbestos exposure: Factors as sociated with excess cancer and respiratory disease mortality. Ann NY Acad Sci 1979:330:117-126.
39. Willett WC: Chemoprevention of occupational tumors, in Gee JBL. Morgan KC. Brooks EM (eds): Occupational Lung Disease. New York, Raven Press. 1984. pp 79--86.
40. Graham S. Mettlin C, Marshall J, et al: Dietary factors in the epidemiology of cancer of the larynx. Am J Epidemiol 1981:113:675680.
41. Mossman BT, Craighead JE. MacPherson BV: Asbestos-induced epithelial changes in organ cultures of hamster trachea: Inhibition by retinyl methyl ether. Science 1980:207:311-313.
42. McLarty JW. Yanagihara R. James H. et al: Baseline data analysis of serum beta-carotene in a randomized clinical trial: Tim Tyler cancer prevention program. The III International Conference on Environmental Lung Disease. Chest 1987:91:301.
Journal of Occupational Medicine/Volume 30 No. 1/January 1988
27