Distribution of Silicotic Collagenization in Relation to Smoking Habits1 ,2
PATRICK A. HESSEL,3 GERHARD K. SWIS-CREMER, and SZE LAP LEE
Interest in the carcinogenicity of crystalline silica dates to the first part of this century (l). A recent review of human and animal studies that argued strongly for a carcinogenic role for crystalline silica has prompted renewed interest in this question (2). In response to this review, Heppleston (3) noted a number of shortcomings in the human studies, including the presence of other airborne contaminants in the occupational environment and the general lack of smoking data in these studies. An International Agency for Research on Cancer (IARC) working group issued a statement in 1987concluding that there wassufficient evidence to conclude that crystalline silica was carcinogenic in laboratory animals and that there was limited evidence for the carcinogenicity of crystalline silica in humans (4). Limited evidence in humans implies that a causal interpretation is credible but that alternative explanations, such as chance, bias, or confounding could not adequately be excluded. MeDonald (5) responded to the conclusions of the IARC working group, stating that "the epidemiological findings are not consistent, risk estimates are generally low,exposure response has not been studied, and the possibilities for confounding by other carcinogens, including tobacco, are many." Our own group has conducted two case-control studies examining the issue in white South African miners (6, 7). One of the studies used the mine worker's pension fund as the sampling frame; the other study used necropsy reports of the same population, which has an 86% autopsy rate. Both studies included fairly precise smoking data, and neither showed an association between lung cancer and either silica dust exposure or silicosis after controlling for smoking. A mortality follow-up in this population found a standardized mortality ratio of 161 for lung cancer relative to white South African men (8). A nested case-
SUMMARY Contl'OV8rsy regarding the association between silicosis and lung cancer has been clouded by the fact that studies examining this association generally do not Include Information on smokIng. A causal association between smoking and silicosis would seriously confound the association between silicosis and lung cancer. The current analysis ane888d the auoclatlon between silicosis and smoking using data on deceased white gold miners who underwent postmortem examination between 1976and 1981.Smoking histories and exposure Information were available. A subset analysis In a group of deceased miners for whom more detailed smoking and exposure Information was available confirmed the findings of the larger analysis. Both analyses showed a slight Inverse relationship between smoking and silicotic collagenlzatlon of the parenchyma and a stronger negative relationship between smoking and silicotic collagenlzatlon of the pleura, controlling for age and cumulative exposure to silica dust. No association between silicotic collagenlzatlon of the hllar glands and smoking was detected. The data should not be construed as advocating that workers exposed to silica dust should smoke. The hazards of cigarette smoking are likely to far outweigh the hazards of silicosis. The data suggest that the lack of smoking histories In studies of the association between silicosis and lung cancer probably does not seriously confound risk estimates. The distribution of silicotic collagenlzatlon In the lungs of smokers and nonsmokers Is consistent with enhanced mucus Interception, more central deposition In the lungs, and proportionally more lymphatic clearance to the hilum (as opposed to the pleura) of silica particles In smokers compared AM REV RESPIR DIS 1991; 144:297-301 to nonsmokers.
control study showed a nonsignificant positive association with dust exposure. In discussing the results of their recent case-control study in this population, Hessel and colleagues (7) summarized possible explanations for the discrepancies in the literature. These included: (1) the lack of data on smoking in these studies; (2) the possibility of an association between silicosis and smoking; (3) other possible concomitant exposures, including asbestos, radiation, and polycyclic aromatic hydrocarbons; (4) the criteria for certifying silicosis, which often include documented disability and therefore favored heavysmokers (who are at a relatively high risk of lung cancer). This last point is important because many of the studies havebeen performed using populations of certified silicoticpatients. The literature therefore contains a mixture of studies on the association between lung cancer and silica exposure and studies of the association between lung cancer and silicosis. These are two different issues. Becausethe studies relating lung cancer and silicosis have more often shown positive associations than the studies of lung cancer and silica ex-
posure, it is necessary to determine whether smoking is related to silicosis. Methods Current legislation in South Africa states that any attending physician who knows or suspects that a deceased patient was a miner is obliged to remove the cardiorespiratory organs and send them to the Medical Bureau for Occupational Diseases if permission is granted by next of kin. Pathologists at the National Centre for Occupational Health perform the necropsies. Although adherence to the law is not complete, approximately 86070 of white gold miners undergo postmortem examination. Since 1975the results ofthese ex(Received in original form October 18, 1990 and in revised form February 1, 1991) 1 From the Department of Health Services Administration and Community Medicine, University of Alberta, Edmonton, Canada, and the Epidemiology Research Unit, Medical Bureau for Occupational Diseases, Johannesburg, South Africa. 1 Correspondence and requests for reprints should be addressed to Patrick A. Hessel, Department of Health Services Administration and Community Medicine, 13-103 Clinical Sciences Building, Universityof Alberta, Edmonton, Alberta TOO 203, Canada. 3 National Health Research Scholar.
298 aminations have been computerized (9), and the records contain detailed descriptions of the macroscopic and microscopic examination of the necropsy material (10). The hilar lymph glands, the parenchyma, and the pleura were examined for evidence of silicotic coUagenization. Diagnosis usually depended on both macroscopic and microscopic evidence. The diagnosis of silicosis of the hilar glands depended on microscopic evidence of foci of concentric fibrosis. Severity was based on the macroscopic determination of enlargementand pigmentation of the glands and the degree of fibrosis present and was categorized as slight, moderate, or marked. The diagnosis of silicosis of the parenchyma was based on the presence of discrete silicotic nodules. Severity was based on the macroscopic assessment of the number of nodules that were palpable (occasional, a few, a moderate number, or a large number). Silicosis of the pleura was assessed solely on the basis of the macroscopic evidence if palpable plaques were present in the pleura and silicosis was present in the parenchyma. If there was no silicosis in the parenchyma, microscopic confirmation of the plaque was necessary. Severity was based on the number of palpable plaques detected (occasional, a few, a moderate number, or numerous). To simplify the presentation, both silicosis of the parenchyma and silicosis of the pleura were classified as absent, slight, moderate, or marked. The two highest categories distinguished pathologically, werecombined to create the "marked" category. Smoking information was obtained from the records of periodic (usually annual) examinations performed by the Medical Bureau for Occupational Diseases to assess fitness for work. Smoking status at death was determined (current smoker, exsmoker, or lifelong nonsmoker). The extant examination forms were reviewed, and the average number of cigarettes smoked per day was estimated. Smoking information has been recorded annually since 1960.In the categorical analyses, smoking wasconsidered as the number of cigarettes per day (categorized as none, 1to 14,15to 24, and ~ 25, and as ever versus never smoking). For gold miners, information on the number of months spent in high-dust, low-dust, and surface-dust jobs were recorded. A dust index was created by adding the number of months worked at low-dust jobs plus the number of months worked in surface-dust jobs plus twice the number of months spent in high-dust jobs. This index, which was developed for quick computation, is based on dust measurements performed in the late 1960s (11). These measurements demonstrated a strong association between exposure to respirable silica and risk of silicosis (12). The average level of free respirable silica in the high-dust jobs is approximately twice the level in the low-dust and surface-dust jobs, although there is wide variation about these mean values. The average concentration of respirable silica (0.5 to 5 JIm in diameter) for the industry is approximately 0.2 mg/m", Av-
HESSEL, SWI5-CREMER, AND LEE
erage dust levels in South African gold mines have remained fairly constant from the 1940s through the present (13). The dust exposure data were recorded for deaths occurring from 1976 to 1981. All white miners whose longest mining servicetype wasgold mining, who died from 1976 to 1981,and who had information on length of gold mining, intensity of dust exposure, and smoking data wereincluded. Miners who were not exposed to dust in the gold mines were excluded from the analysis even if they met the other criteria because they would not have been at risk of developing silicosis. That some of these excluded miners showed evidence of silicosis at necropsy indicated that they may have worked in other types of mines or in occupations not officially scheduled as dusty but in which dust was present. This included certain surface jobs. Data on age and dust exposure wereavailable for 4,456 miners. Of these, smoking status was unknown for 242 and the number of cigarettes smoked per day was unknown for 970. For pipe smokers, the number of pipes per day was not recorded. For some of the cigarette smokers and exsmokers, the number of cigarettes smoked per day was not available, although it was possible to determine that they had smoked. The data wereanalyzed using unconditional logistic regression, fitting silicosis as the dependent variable and age, smoking, and dust exposure as independent variables. In some analyses, silicosis was considered present or absent; for other analyses, those with silicosis at each level of severity were compared to those with no silicosis. Smoking (cigarettes per day) and dust exposure were analyzed as continuous variables in the logistic regressions. To further explore the association between silicosis and smoking, data from one of the previous case-control studies of silica, silicosis, and lung cancer mentioned previously were examined (7). In this study dust exposure and smoking were abstracted more precisely. Only the data for controls were used for the present analysis. The main purpose of the analysis was to validate the risk estimates from the larger analysis in which the data on risk factors were recorded with less precision. These results are not independent of the results of the main analysis already described because the source of subjects was the same.
The study population is described in table 1. The average age at death was rather low (59.8 yr). Because the recording of smoking information was introduced only in 1960, the miners who were older during the period 1976to 1981 were less likely to have information on smoking than the younger deceased miners. They would therefore have been more likely to have been excluded from the analysis. The average dust index was 309, which
TABLE 1 DESCRIPTION OF THE STUDY POPULATION*
1-14 15-24 ~ 25 Silicosis: hilar glands None Slight Moderate Marked Silicosis: parenchyma None Slight Moderate Marked Silicosis: pleura None Slight Moderate Marked Total
573 1,018 1,425 470
16.4 29.2 40.9 13.5
1,375 614 1,589 878
30.9 13.8 35.7 19.7
2,569 891 539 457
57.7 20.0 12.1 10.3
BO.6 12.1 4.4 2.8
197 128 4,456
• Age, mean ± SO (range),59.8 ± 13.9 (17-100); dust exposure, mean ± SO (range), 309.0 ± 209.7 (1-991).
equates to approximately 26 yr at low dust. Silicosis was found most commonly in the hilar lymph glands (69.1 070 were positive). Some evidence of silicosis in the parenchyma was seen in 42070, and in 19070 silicosis was detected in the pleura. The bivariate comparison of smoking and silicosis of the hilar glands indicated no association (table 2). A negative association was found between silicosis of the parenchyma and smoking. Heavier smokers were generally less likely than nonsmokers or lighter smokers to have evidence of silicosis, especially the higher grades of silicosis(not shown). A negative association was also seen for silicosis of the pleura and smoking. As expected, dust exposure was related to age and the presence and degree of silicosis of all three sites. There was a slight association between dust exposure and smoking (table 3). This can be seen by comparing the distribution of dust exposure for light smokers (1 to 14 cigarettes/day) and moderate smokers (15 to 24 cigarettes/day). Light smokers had relatively more dust exposure than moderate smokers. The distributions of dust exposure for nonsmokers and heavy smokers were similar. Thus, it was necessary to control for dust and age in examining the relationship between silicosis and smoking. Silicosis of the hilar glands was slightly more common among smokers when age and dust exposure were included in the logistic regression model (table 4).
SILICOSIS AND SMOKING
299 TABLE 2
COMPARISON OF THE DEGREE OF SILICOSIS BY SMOKING CATEGORY·
ODDS RATIOS AND CONFIDENCE INTERVALS FOR PRESENCE OF SILICOSIS OF THE PLEURA AND A HISTORY OF EVER HAVING SMOKED BY DUST EXPOSURE LEVEL, CONTROLLING FOR AGE AT DEATH AND DUST EXPOSURE
Hi/ar Glands NS
Degree of Silicosis
None Slight Moderate Marked
173 80 205 115
30.2 14.0 35.8 20.1
1,097 504 1,308 732
30.1 13.8 35.9 20.1
305 116 86 66
53.2 20.2 15.0 11.6
2,098 735 437 371
57.6 20.2 12.0 10.2
418 93 36 26
72.9 16.2 6.3 4.5
2,966 427 152 96
81.5 11.7 4.2 2.6
Definition of abbreviations: N8 lifelong nonsmokers, 8M ever smokers . • For silicosis of the hilar glands, x2 (3 OF, degrees of freedom) = 0.009, P = 0.9998; for silicosis of the parenchyma, (3 OF) = 6.13, P • 0.11; for silicosis of the pleura, X2 (3 OF) = 23.7, P = 0.00003.
COMPARISON OF DUST EXPOSURE BY SMOKING CATEGORY· Smoking Category (Cigarettes/Day) 1-14
22.2 17.8 18.4 17.3 24.3
195 168 209 206 240
19.2 16.5 20.5 20.2 23.6
336 275 297 248 269
23.6 19.3 20.8 17.4 18.9
1-120 121-240 241-360 361-480 ~ 481
127 102 105 99 139
• "X} (12 OF) = 22.4, P months high dust).
Dust index = months low dust
n 114 83 91 83 99
24.3 17.7 19.4 17.7 21.1
+ months surface dust + (2 x
TABLE 4 ODDS RATIOS AND CONFIDENCE INTERVALS FOR PRESENCE OF SILICOSIS OF THE HILAR GLANDS AND SMOKING CONTROLLING FOR AGE AT DEATH AND DUST EXPOSURE Hi/ar Glands
Smoking Comparison •
Ever/never 1-14/never 15-24/never ~ 25/never
1.16 1.18 1.18 1.08
0.92-1.47 0.90-1.56 0.92-1.53 0.79-1.48
0.97 1.00 0.91 0.77
0.79-1.20 0.79-1.28 0.72-1.15 0.58-1.03
0.69 0.69 0.73 0.70
0.55-0.86 0.53-0.90 0.57-0.94 0.51-0.97
Definition of abbreviations: OR = odds ratio, CI = 95% confidence interval. • For smoking comparisons, numbers refer to cigarettes per day.
The slight, nonsignificant, positive association between silicosis of the hilar glands and smoking decreased for heavier smokers. Silicosis of the parenchyma showed a slight, nonsignificant, negative association with ever versus never smoking. The odds ratios decreased with increasing smoking. For silicosis of the pleura, a stronger and statistically significant negative association was seen with ever versus never smoking. The magnitude of the odds ratios did not vary systematically with the number of cigarettes smoked per day. Severity-specific analyses were performed comparing subjects with each grade of silicosis to those with no silicosis, controlling for age and dust exposure. For silicosis of the pleura the odds ratios
95% Confidence Interval
1-120 121-240 241-360 361-480 ~ 481
0.952 0.972 0.694 0.607 0.588
0.398-2.276 0.496-1.905 0.429-1.124 0.384-0.958 0.406-0.852
Dust Exposure Category
Dust Exposure Category
for smoking decreased with severity (0.71 for slight, 0.66 for moderate, and 0.62 for marked). The associations between silicosis of each site and smoking were examined by levelof dust exposure. For silicosis of the hilar glands and the parenchyma there were no systematic trends in the odds ratios with increasing dust exposure. For silicosis of the pleura, the odds ratios for smoking decreased with increasing dust exposure (table 5). It was necessary to control for dust exposure in these analyses, in addition to stratifying by dust exposure to account for residual confounding. Risk estimates for silicosis and smoking were obtained using the data for controls from a recently completed case-
control study of silica, silicosis, and lung cancer (7). These subjects were a subset of the main study group. If genuine associations exist between silicosis and smoking, the odds ratios for the subset with more precise exposure data should be more extreme (further from 1) (14). There were very small differences between the odds ratios for the main study and those for the subset for silicosis of the hilar glands. For silicosis of the parenchyma and pleura, the odds ratios for all comparisons were smaller (more extreme) for the subset, possibly reflecting a lack of precision in the dust exposure and smoking estimates in the main study. The trend of decreasing odds ratios with increasing smoking seen for silicosis of the parenchyma was not apparent in the subset. The trend of decreasing odds ratios for silicosis of the pleura and smoking with increasing dustexposure was not apparent in the subset. Discussion
Based on the results of this analysis, it appears that silicosis of the hilar lymph glands is unlikely to be associated with smoking. The slightly positive association seen in the main analysis was small and not statistically significant and was not confirmed in the subset analysis. There appears to be a slight negative association between parenchymal silicosis and smoking that may be more pronounced at higher levelsof smoking. Although the results of the main analysis could be explained by chance, the subset analysis indicated that the odds ratios may have been overestimated in the main study (closer to 1),possibly owing to random error in the exposure estimates in the main study. A negative association between silicosis ofthe pleura and smoking is more evident, although apparently unaffected by the amount smoked. The association may be more pronounced at higher dust levels, although this cannot be said with assurance given the conflict-
ing results of the main study and the subset analysis. The necropsy rate among white South African gold miners is high (86070), but the next of kin of miners who are fully certified in life receive no further compensation on the basis of the results of the postmortem examination. It is suspected that miners fully certified in life are underrepresented (but certainly not unrepresented) in the necropsy series. Because full compensation usually depends on some demonstrated respiratory impairment and a suitable exposure history, it is possible that those with silicosis who have high dust exposure and impairment (often due to cigarette smoking in addition to dust exposure) would be underrepresented. The exclusion of these miners would have decreased the odds ratios somewhat. However, the relationships between smoking and silicosis of the parenchyma and pleura were consistent even at low levels of dust exposure at which full compensation is less likely. This possible bias must be considered, but it is unlikelyto have altered the results significantly. It has also been suggested that heavier smokers may be more likely to have respiratory impairment and be selected out of the dustier jobs (which may also be more strenuous). However, a recent analysis found that dust exposure in these miners did not change systematicallywith percentage of predicted FEV. (Hnizdo, personal communication). 1\\'0 additional points should be made. First, only the data regarding silicosis of the parenchyma are relevant to studies of silicosis and lung cancer. This study did not prove a negative association between silicosis of the parenchyma and smoking, and it is unlikely that the (possible) negative association between silicosis of the parenchyma and smoking is strong enough to substantially affect studies of silicosis and lung cancer. If there were any effect on these studies, it would be to slightly underestimate the association between silicosis and lung cancer as silicotic individuals would, as a group, be at an otherwise lower risk of lung cancer because they smoke less. The data do not necessarily relate to studies of silica exposure and lung cancer since the smoking habits of populations exposed to silica (and therefore lung cancer risk) may well differ from those of a reference group (most often a national or regional population). To date, however, most such studies have looked at populations of silicotic subjects rather than exposure to silica dust per se.
HESSEL, SWIS-CAEMEA, AND LEE
The second and more important point is that these data do not in any way indicate that dust-exposed individuals should smoke. The negative health effects of smoking are likely to far outweigh the health effects of silicosis. Our own data comparing silicotic and nonsilicotic individuals exposed to approximately equal levels of dust and smoking have shown very slight differences in lung function and no differences in exercise tolerance (Wiles, personal communication). The effect of silicotic pleural plaques on lung function has not been investigated, but it is unlikely that they have a major effect (15). Gough stated in 1960 (16): "There is an important relationship between the presence of bronchitis and the accumulation of dust in the lungs. The relationship is an inverse one, that is to say if there is chronic bronchitis less dust will be accumulated in the lungs. This is presumably due to the dust being swept out in the increased secretions."This relationship was observed in coal workers. A similar relationship was seen in farmers in relation to extrinsic allergic alveolitis (BAA). Warren (17) compared patients with EAA in a Manitoba hospital to several other patient groups and to the general population and found that nonsmokers were more heavily represented among the EAA patients, and Morgan and coworkers (18) noted that smoking farmers less often showed evidence of EAA than nonsmoking farmers. Lin and Goodwin (19) lent experimental support to these results when they demonstrated greater deposition of a respirable aerosol in the ciliated airways compared to the nonciliated airways among patients with more severeobstruction. They noted that the greatest focal hyperdeposition among patients with obstruction occurred in the hilar and perihilar large airways, although they did not mention whether there was a similar central distribution of particles in the nonciliated airways.Pavia and colleagues (20) found that airways obstruction as measured by FEV. caused decreased penetration of an aerosol. They explained this finding on the basis of increased impaction caused by disturbed airflow in subjects with airway obstruction. Newhouse and Ruffin (21)found poor distribution of aerosol to the periphery in subjects with chronic obstructive lung disease. Chamberlain and colleagues (22) found more central deposition of particles in smokers compared to nonsmokers and explained this by stating that excess mucus decreased the cross-sectional
area of the airways and increased the impaction of particles. It has been noted that lung cancers in asbestos-exposed individuals who smoke more often occur centrally in the lungs compared to lung cancers in nonsmoking asbestos workers (23). It may therefore be possible to explain the results presented here on the basis of increased deposition of particles in the large, ciliated airways in smokers as a result of narrowing of the airways and increased turbulence of flow. This would serve to reduce the total burden of occupational dust in the parenchyma of smokers. The differing odds ratios for the various sites of silicotic collagenization might be explained by enhanced central deposition of particles in smokers with preferential lymphatic clearance of these particles directly to the hilum as opposed to the pleura. Acknowledgment The authors are indebted to the pathologists and staff of the National Centre for Occupational Health whose careful work has made this research possible. The contributions of Dr. Jill Murray, Ms. Eva Hnizdo, and Dr. KeithMorgan in evaluatingthe resultsare also recognized with appreciation. References 1. Irvine LG. Report of the Miners' Phthisis Medical Bureau 1935-1938. Pretoria: Government Printer, Union of South Africa, 1939. 2. Goldsmith DF, Guidotti TL, Johnston DR. Does occupational exposure to silicacause lung cancer? Am J Ind Med 1982; 3:423-40. 3. Heppleston AG. Silica, pneumoconiosis, and carcinoma of the lung. Am J Ind Med 1985; 7:285-94. 4. International Agency for Research on Cancer (1987). Evaluation ofthe carcinogenic risk of chemicals to humans: silica and some silicates. IARC Monographs, Vol. 42. Lyon: IARC, 1987. 5. McDonald JC. Silica, silicosis and lung cancer. Br J Ind Med 1989; 46:289-91. 6. Hessel PA, Sluis-Cremer GK, Hnizdo E. Casecontrol study of silicosis, silica exposure and lung cancer in white South African gold miners. Am J Ind Med 1986; 10:57-62. 7. Hessel PA, Sluis-Cremer GK, Hnizdo E. Silica exposure, silicosisand lung cancer: a necropsy study. Br J Ind Med 1990; 47:4-9. 8. Wyndham CH, Bezuidenhout BN, Greenacre MJ, Sluis-Cremer GK. Mortality of middle aged South African gold miners. Br J Ind Med 1986; 43:677-84. 9. Soskolne C, Goldstein B, Haffajee I. An integrated cardiorespiratory pathology information system. S Afr Med J 1976; 50:1832-6. 10. Hessel PA, Davies JCA, Goldstein B, Webster I, Hnizdo E, Landau S. Pathological findings in mineworkers. I. Description of the PATHAUT data base. Am J Ind Med 1987; 12:71-80. 11. Beadle DG, Bradley AA. The composition of airborne dust in South African gold mines in pneumoconiosis. In: Shapiro HA, ed. Pneumoconio-
SILICOSIS AND SMOKING
sis: proceedings of the International Conference, Johannesburg, 1969. London: Oxford University Press, 1970; 462-6. 12. Beadle DG, Harris E, Sluis-Cremer GK. The relationship between the amount of dust breathed and the incidence of silicosis. In: Shapiro HA, ed. Pneumoconiosis: proceedings of the International Conference, Johannesburg, 1969. London: Oxford University Press, 1970; 473-7. 13. Sluis-Cremer GK. Silica exposure and silicosis in Witwatersrand gold miners in South Africa. In: Goldsmith OF, Winn OM, Shy CM, eds. Silica, silicosis, and cancer: controversy in occupational medicine. New York: Praeger, 1986; 67-70. 14. Rothman KJ. Modern epidemiology. Boston: Little, Brown, 1986; 108.
301 15. Sluis-Cremer GK, Solomon A. Radiological features of silicosis. In: Solomon A, Kreel L, eds. Radiology of occupational chest diseases. New York: Springer-Verlag, 1989; 101-42. 16. Gough J. Emphysema in relation to occupation. Ind Med Surg 1960; 29:283-5. 17. Warren CPW. Extrinsic allergic alveolitis: a disease commoner among non-smokers. Thorax 1977; 32:567-9. 18. Morgan DC, Smyth JT, Lister RW, et al. Chest symptoms in farming communities with special reference to farmer's lung. Br J Ind Med 1975; 32:228-34. 19. Lin MS, Goodwin DA. Pulmonary distribution of an inhaled radioaerosol in obstructive pulmonary disease. Radiology 1976; 118:645-51.
20. Pavia 0, Thomson ML, Clarke SW, Shannon HS. Effect of lung function and mode of inhalation on penetration of aerosol into the human lung. Thorax 1977; 32:194-7. 21. Newhouse MT, Ruffin RE. Deposition and fate of aerosolized drugs. Chest (Suppl) 1978;73:936-43. 22. Chamberlain MJ, Morgan WKC, Vinitski S. Factors influencing the regional deposition of inhaled particles in man. Clin Sci 1983; 64:69-78. 23. Kannerstein M. Recent advances and perspectives relative to the pathology of asbestos related diseases in man. In: Wagner JC, ed. The biological effects of mineral fibres. Lyon: IARC Scientific Publication No. 30, 1980; 149-62.