PII: S0003-4878(96)00100-7 Ann. occup. Hyg., Vol. 41, Supplement 1, pp. 475-479, 1997 1997 British Occupational Hygiene Society Published by Elsevier Science Ltd. All rights reserved Printed in Great Britain 0003-4878/97 $17.00 + 0.00 Inhaled Particles Vlll DOES OCCUPATIONAL SILICA EXPOSURE OR SILICOSIS CAUSE LUNG CANCER? D. F. Goldsmith Public Health Institute, 2001 Addison Street, Suite 210, Berkeley, CA, 94704-1103 U.S.A. INTRODUCTION The International Agency for Research on Cancer's (IARC) assessment of silica as a probable human carcinogen rested on three evidentiary bases: (a) positive experimental rat model of pulmonary carcinogenesis via inhalation and injection; (b) elevated cancer (mostly lung) among silica-exposed workers; and (c) elevated pulmonary cancer risk among workers with extant silicosis, often diagnosed and recorded in conjunction with compensation registers (IARC, 1987). IARC indicated that the following epidemiology issues were of concern to the scientists on the Monograph writing committee 10 years ago. (1) Absence of dose-response findings; (2) Lack of adjustment for smoking; (3) Lack of adjustment for other possible confounding factors (i.e. radon, polycyclic aromatic hydrocarbons, arsenic); and (4) Diagnostic bias of lung cancer among compensated silicotics. For these reasons, the evidence among humans was judged to be "limited". RATIONALE If the scientific community has not addressed these issues in the 10 years since IARC's report, then there would be no justification for changing the Agency's assessment. However, if research has been undertaken to evaluate these limitations, then it may be justified to change assessment for silica exposure [see Goldsmith (1996) for an expanded review of the underlying issues]. Thus, the rationale for this paper is to critically evaluate the evidence over the past 10 years and offer a review of the causation issues for silica's carcinogenicity status. Each of the four weaknesses will be listed and evidence in response to it will be described. In conclusion, the criteria for causation will be examined for both occupational silica exposure and for silicosis. Absence of dose-response findings Fu et al. (1992) in a study of 1113 underground Chinese miners from 1960 to 1988 showed that cumulative dust (mg-y) relative risk (RR) for lung cancer rose from 1.0 for < 500 mg-y to 4.2 for 500-1449 mg-y, to 5.2 for 1500-2999 mg-y, to 13.0 for 3000-4449 mg-y, to 16.1 > 4500 mg-y. Using a case-control analysis to adjust for age and smoking McLaughlin et al. (1992) examined the risk among Chinese male miners by degree of cumulative exposure to respirable silica. There was a 475
476 D. F. Goldsmith significant positive trend for cumulative dust exposure (P = 0.02) and for respirable silica (P = 0.004) among tin miners (87 cases, 371 controls). For cumulative respirable silica the lung cancer odds ratio (OR) rose from 1.5 for low cumulative exposure to silica (0.1-8.69 ug m~ 3 y~ x ; 15 cases; 67 controls), to 1.9 for medium cumulative exposure to silica (8.7-26.2 ug m~ 3 y" 1 ; 22 cases; 82 controls), to 3.1 for high cumulative exposure to silica (> 26.3 ig m~ 3 y" 1 ; 35 cases, 108 controls). The authors examined cumulative exposure to arsenic, polycyclic aromatic hydrocarbons and radon decay products, and only arsenic produced a significant trend among tin miners, suggesting that arsenic and silica dusts may interact. Muller and colleagues (1986) conducted a retrospective cohort study of 6972 underground gold miners from the Province of Ontario, Canada who were employed at least 60 months in mining. These men were followed from 1955 to 1977 using tracing information from the Ontario Workers' Compensation Board/Mining Master File, death records from Statistics Canada and Social Insurance files. When the years worked was defined, a dose-response emerged with the risk of < 5 years having a standardized mortality rate (SMR) of 139; 5-9 years a SMR of 143; 10-14 years SMR of 134; 15-19 years SMR of 163; and > 20 years SMR was 202. Lack of adjustment for smoking A cohort of 3971 white South African gold miners (exposed to an environment containing ~ 30% airborne silica) born between 1916 and 1930 and actively employed in 1970, was followed for 9 years. Wyndham et al. (1986) demonstrated that overall, there was an elevated risk for lung cancer (SMR = 161; 39 observed; 24.2 expected; 95% CI = 115, 220). The authors showed that cumulative dust exposure had a smoking-adjusted rate ratio (RR) of 1.77 per 10 particle years of underground exposure [95% CI = 0.94-3.31 (P = 0.06)]. Hnizdo and Sluis-Cremer (1991) presented a study of 2209 South African white male gold miners who began mining during the years 1936-1943 and whose mortality was traced from 1968 to 1986. The authors classified the gold mine jobs into 11 groups based on the average respirable silica dust concentrations developed by Beadle (1971). Cumulative dust measures (in respirable particles-years/1000) were calculated for each miner. Overall, the authors showed that the 77 cases of lung cancer had greater levels of silica particle-years (SPY) exposure, smoked more, had greater pack-years, had a greater percentage of ever smokers (96% vs 88%) than control miners. Using a Cox proportional hazards model that adjusted for age and pack-years, the RR rose from 1.00 for < 15 SPY/1000 to 1.54 for 16-30 SPY/1000, to 2.07 for 31-40 SPY/1000, to a RR of 2.92 for > 41 SPY/1000 (95% CI = 1.02-8.4). According to the model, the RR for lung cancer per 1000 SPY (adjusted for smoking, year of birth and age) was 1.023 (95% CI = 1.005-1.042). This means that the most veteran gold miners (all employed > 10 years) exposed to 50 000 SPY have 3.2 times greater likelihood of dying from lung cancer than those exposed to only 15 000 SPY. The data indicated the silica and smoking appeared to act synergistically and the synergy was greatest for miners with the most smoking and mining experience (Hnizdo and Sluis-Cremer, 1991). Hnizdo raised the possibility that radon exposure may have confounded the associations reported with silica dust (Hnizdo, 1994), though more recent research does not confirm a radon association (Hnizdo, personal communication).
Does occupational silica exposure or silicosis cause lung cancer? 477 Checkoway et al. (1993) have conducted an extensive study of the health of 2570 white male diatomaceous earth (DE) workers who were exposed to cristobalite, a silica polymorph. The men were employed at least 1 year in the industry and followed from 1942 to 1987. There was a SMR of 1.43 for lung cancer (59 observed; 41.4 expected; 95% CI = 1.09, 1.84). The authors found dose-related gradients for duration of employment and a semiquantitative silica exposure index for lung cancer with lagged exposure for 15 years. The risks were adjusted for age, calendar year, duration of follow-up and ethnicity. The lung cancer risk rose from 1.00 for < 50 exposure intensity-years, to 1.19 for 50 to 99 exposure intensity-years, to 1.37 for 100-199 exposure intensity-years, to 2.74 for > 200 intensity-years. Having limited smoking histories on the DE workers, Checkoway et al. reported that smoking could not account for all the link between silica exposure and lung cancer because the exposure-response was the same for the cohort as it was for smokers only. They concluded that high cristobalite exposure prior to the 1950s was the most etiologically significant contributor to the lung cancer risk. Lack of adjustment for other possible confounders (i.e. radon, PAH, arsenic) There have been positive lung cancer findings among workers in the ceramic industries (Forastiere et al., 1986; Winter et al., 1990); in granite industry (Koskela et al., 1987, 1990); in stone industry (Guenel et al., 1989); and in diatomaceous earth processing (Checkoway et al., 1993). All are industries characterized by little or no confounding from radon or arsenic (seen in mining) or polycyclic aromatic hydrocarbons (in foundries). Some residual confounding in uranium miners could indicate that silica and radon decay products may interact biologically in jointly producing excess of lung cancer (Goldsmith and Goldsmith, 1996). Diagnostic bias among compensated silicotics Rosenman et al. (1995) demonstrated the lung cancer risk was actually greater among hospitalized silicotics in Michigan and New Jersey who had not applied for Workers' compensation compared to those who had filed, PMR = 1.13 vs 1.60. Similarly, Merlo et al. (1995) found a SMR of 3.14 (37 observed; 11.8 expected; 95% CI = 2.21, 4.33) among hospitalized silicotics in Genoa, Italy without regard to compensation status. There have been five studies demonstrating lung cancer dose-response by severity of silicosis from chest X-rays (Chia et al., 1991; Miller et al., 1987; Ng et al., 1990; Hnizdo and Sluis-Cremer, 1991; [for silicosis of hilar lymph glands only]; and Goldsmith et al., 1995 [based on degree of compensation awarded]). Almost every study of silicotics (over 30 in the peer-reviewed literature since 1980) showed at least > 2 times lung cancer risk (see Smith et al., 1995), including Chiyotani et a/.'s (1990) RR of 2.22 among never-smoking Japanese silicotics. SUMMARY OF EVIDENCE OF HUMAN CANCERS RELATED TO SILICA EXPOSURE Table 1 shows that the modified criteria (after Hill, 1995) have been met for the most part for both silicotics and for silica-exposed workers (Goldsmith, 1996). Many studies have examined the cancer risk in silicotics and non-silicotics separately and found greater elevated lung cancer risks among silicotics. The
478 D. F. Goldsmith Table 1. Criteria for cancer causation for silica exposure and for silicosis Silica exposed Workers with Point of evidence workers silicosis Strong relative risk / Dose-response gradient // Consistent findings // Controlled confounding / / Biological plausibility y// ± Temporal cogency / J Specificity J J J J Overall coherence Yes Yes y = criteria met. ± = incomplete evidence. presence of silicosis itself may be an independent risk factor for pulmonary neoplasia, an indicator variable for duration and/or intensity of exposure, or a marker of genetic susceptibility to pulmonary damage by respirable size crystalline silica. The general trend is that silicotics show increased risk of pulmonary neoplasia as compared to both those exposed to silica without silicosis and the general population. Winter et al. (1990) among U.K. ceramic workers, Fu et al. (1992) and McLaughlin et al. (1992) among Chinese tin miners, Hnizdo and Sluis-Cremer (1991) among South African gold miners and Checkoway et al. (1993) among U.S. diatomaceous earth workers demonstrated excess lung cancer risk associated with occupational exposure to crystalline silica and dose-response relationships. Some studies also demonstrated greater than additive interaction between crystalline silica and tobacco smoke. Other epidemiologic studies demonstrated doseresponse gradients for the association between silicosis and lung cancer. CONCLUSIONS With two new positive animal studies since 1987 (Muhle et al., 1989; Spietoff et al., 1992) and silica's ability to bind with DNA (Daniel et al., 1995), IARC's prior assessment does not appear to be wrong. Furthermore, epidemiology evidence in favour of silica's carcinogenic role appears to be sufficient for both silica-exposed workers and for silicotics (Goldsmith, 1996). The shortcomings in human studies noted by members of the last IARC Working Group on silica have been addressed and there is now a large body of scientific evidence to conclude that occupational silica dust exposure is carcinogenic to humans. REFERENCES Beadle, D. (1971) The relationship between the amount of dust breathed and the development of radiological signs of silicosis: an epidemiological study of South African gold miners. In Inhaled Particles III. (Edited by W. H. Walton) pp.953-964. Pergamon Press, Oxford. Checkoway, H., Heyer, N. J., Demers, P. A. and Breslow, N. E. (1993) Mortality among workers in the diatomaceous earth industry. Br. J. ind. Med., 50, 586-597. Chia, S-E., Chia, K-S., Phoon, W-H. and Lee, H-P (1991) Silicosis and lung cancer among Chinese granite workers. Scand. J. Work Environ. Health 17, 170-174. Chiyotani, K., Saito, K., Okubo, T. and Takahashi, K. (1990) Lung cancer risk among pneumoconiosis patients in Japan, with special reference to silicotics. In Occupational Exposure to Silica and Cancer
Does occupational silica exposure or silicosis cause lung cancer? 479 Risk (Edited by L. Simonato, A. C. Fletcher, R. Saracci and T. L. Thomas), pp. 95-104. International Agency for Research on Cancer (IARC): Lyon, France. Daniel, L. N., Mao, Y., Williams, A. O. and Saffiotti, U. (1995) Direct interaction between crystalline silica and DNA a proposed model for silica carcinogenesis. Scand. J. Work Environ. Health 21(2): 22-26. Forastiere, F., Lagorio, S., Michelozzi, P., Cavariani, F., Area, M., Borgia, P., Perucci, C. and Axelson, O. (1986) Silica, silicosis and lung cancer among ceramic workers: a case-referent study. Am. J. ind. Med. 10, 363-370. Fu, H., Jing, X., Yu, S., Gu, X., Wu, K., Yang, J. and Qiu, S. (1992) Quantitative risk assessment for lung cancer from exposure to metal ore dust. Biomedical and Environ. Sci. 5, 221-228. Goldsmith, D. F., Beaumont, J. J., Morrin, L. A. and Schenker, M. B. (1995) Respiratory cancer and other chronic disease mortality among silicotics in California. Am. J. ind. Med. 28, 459-467. Goldsmith, D. F. and Goldsmith, J. R. (1996) Joint cancer risks among workers having silica, smoking, and radon exposures must be examined. International Conference on Radiation and Health, Beer Sheva, Israel, November 3-7. Goldsmith, D. F. (1996) Importance of causation for interpreting occupational epidemiology research: a case study of quartz and cancer. Occupational Medicine: State of the Art Reviews 11, 433-449. Guenel, P., Hojberg, G. and Lynge, E. (1989) Cancer incidence among Danish stone workers. Scand. J. Work, Environ. Health 15, 265-270. Hill, A. B. (1965) The environment and disease: Association or causation? Proc. Roy. Soc. Med. 58, 295-300. Hnizdo, E. (1994) Risk of silicosis in relation to fraction of respirable quartz (letter). Am. J. ind. Med. 25, 771-772. Hnizdo, E. and Sluis-Cremer, G. K. (1991) Silica exposures, silicosis, and lung cancer: a mortality study of South African gold miners. Br. J. ind. Med. 48, 53-60. IARC (1987) IARC monographs on the evaluation of the carcinogenic risk of chemicals to humans. Silica and some silicates, monograph #42, Lyon. International Agency for Research on Cancer, Lyon, France. Koskela, R.-S., Klockars, M., Jarvinen, E., Kolari, P. J. and Rossi, A. (1987) Cancer mortality of granite workers. Scand. J. Work Environ. Health 13, 26-31. Koskela, R.-S., Klockers, M., Jarvinen, E., Rossi, A. and Kolari, P. J. (1990) Cancer mortality of granite workers 1940-1985. Occupational Exposure to Silica and Cancer Risk (Edited by L. Simonato, A. C. Fletcher, R. Saracci and T. L. Thomas), pp. 43-53. Lyon, France, International Agency for Research on Cancer (IARC). McLaughlin, J. K., Chen, J.-Q., Dosemeci, M., Chen, R.-A., Rexing, S. H., Zhien, W., Hearl, F. J., McCawley, M. A and Blot, W. J. (1992) A nested case-control study of lung cancer among silica exposed workers in China. Br. J. ind. Med.49, 167-171. Merlo, F., Fontana, L., Reggiardo, G., Ceppi, M., Barisone, G., Garrone, E. and Daria, M. (1995) Mortality from lung cancer among 515 Genoa, Italy silicotics: results from the follow-up period 1961-1987. Scand. J. Work Environ, and Health 21(2), 77-80. Miller, A. B., Scarpelli, D. and Weiss, N. S. (1987) Report to the workers' compensation board on the Ontario gold mining industry of the scientific panel on mortality from cancer among Ontario gold miners 1955-1977 industrial disease standards panel Ontario Ministry of Labour, Toronto, Ontario. Muhle, H., Takenaka, S., Mohr, U., Dasenbrock, C. and Marmelstein, R. (1989) Lung tumor induction upon long-term low-level inhalation of crystalline silica. Am. J. ind. Med. 15, 343-346. Muller, J. and Kusiak, R. A. (1986) Study of mortality of Ontario gold miners 1955-1977. Ontario Ministry of Labour, Toronto, Ontario, July. Ng, P. N., Chan, S. L. and Lee, J. (1990) Mortality of a cohort of men in a silicosis register: further evidence of an association with lung cancer. Am. J. ind. Med. 17, 163-171. Rosenman, K.D., Stanbury, M. J. and Reilly, M. J. (1995) Mortality among persons with silicosis reported to two state based disease surveillance systems. Scand. J. Work Environ. Health 21 (Suppl. 2) 73-76. Smith, A. H., Lopipera, P. A. and Barroga, V. R. (1995) Meta-analysis of studies of lung cancer among silicotics. Epidemiology 6, 617-624. Spietoff, A., Wesch, H., Wegener, K. and Klimisch, H. (1992) The effects of thoratrast and quartz on the induction of lung tumors in rats. Health Phys. 63, 101-110. Winter, P. D., Gardner, M. J., Fletcher, A. C. and Jones, R. D. (1990) A mortality follow-up study of pottery workers: preliminary findings on lung cancer. In Occupational Exposure to Silica and Cancer Risk (Edited by L. Simonato, A. C. Fletcher, R. Saracci and T. L. Thomas), pp. 83-94. International Agency for Research on Cancer (IARC), Lyon, France. Wyndham, C. H., Bezuidenhout, B. N., Greenacre, M. J. and Sluis-Cremer, G. K. (1986) Mortality of middle aged white South African gold miners. Br. J. ind. Med. 43, 677-684.