Document YrnLxr9zBJj02jaYd8RQY6M6O

REVIEWS Scand J Work Environ Health 1994;20:235--42 Mechanisms of the combined effect of asbestos and smoking in the etiology of lung cancer by Harri Vainio, MD, Paolo Boffetta, MD1 VAINIO H, BOFFETTA P. Mechanisms of the combined effect of asbestos and smoking in the etiol ogy of lung cancer. Scand J Work Environ Health 1994;20:235---42. The joint effects of exposure to two known lung carcinogens, tobacco smoking and asbestos, are reviewed. The variable pattern of interaction -- ranging from supramultiplicative to less than additive -- may reflect the fact that both asbestos and smoking are complex carcinogens which can affect more than one stage of lung carcino genesis. The joint effect of two such agents will depend on the relative magnitude of the effects at each stage. The epidemiologic evidence from studies of insulation workers with high exposures sug gests an interaction that approximates the multiplicative model, indicating that each of the two fac tors has an independent action on the multistage process of carcinogenesis. Very limited information is available on the interaction between these two agents in causing specific histological types of lung cancer. Both tobacco smoke and asbestos fibers can be genotoxic and cytotoxic and cause prolifera tive lesions in the lungs. Tobacco smoke is known to contain carcinogens that bind to critical genes in DNA (deoxyribonucleic acid) and cause mutations. Asbestos fibers may cause chronic inflamma tion of the lungs, which releases various cytokines and growth factors, and therefore may provide a possible selective growth advantage for mutated cells. Key terms -- adenocarcinoma, cancer, interaction, lung, review, tobacco smoking. Lung cancer is the leading cause of death from can cer throughout the world (1). The major histologi cal types of lung cancer are small-cell, squamous cell, and large-cell carcinomas and adenocarcinomas. Tobacco smoking causes all types of lung cancer, but the risk for squamous-cell and small-cell carcinomas of the lung is higher than that for adenocarcinoma (2). While tobacco smoking is the most important single cause of lung cancer, it is not a condition sine qua non. Increased risks for lung cancer have been demonstrated in nonsmoking populations with vari ous occupational exposures, and increases in risk occur for both smokers and nonsmokers exposed oc cupationally to asbestos fibers (3). The risk for lung cancer associated with tobacco smoking is substan tially increased in conjunction with exposure to as bestos, however, and therefore smoking and asbes tos seem to have a more than additive action in caus ing lung cancer. Lung cancer is not a single disease, and the vari ous histological types have unique characteristics and may have specific etiologies. In this review, we ex amine the evidence for interaction between tobacco smoking and exposure to asbestos in causing lung cancer in general and various histopathological types 1 International Agency for Research on Cancer, Lyon, France. Reprint requests to: Dr H Vainio, International Agency for Research on Cancer, 150 cours Albert Thomas, F-69372 Lyon Cedex 08, France. of lung cancer in particular and discuss the possible underlying mechanisms of that interaction. Tobacco smoke as a lung carcinogen Carcinogenesis is conventionally regarded as a se ries of events that begins with the initiation of ge netic alterations by an agent or agents that interact with deoxyribonucleic acid (DNA) and lead to her itable alterations. These "initiated" cells require fur ther stimuli to divide during an extended period of tumor "promotion" and "progression." The long la tency period between the initiation and diagnosis of cancer may reflect the latter stages. The epidemiologic data on tobacco smoking and lung cancer are more extensive than those for any other cause of neoplasms in humans. Smoking can affect both early and late stages of carcinogenesis. Smoking in early life has a substantial effect on the risk for cancer in old age and therefore evidently af fects at least one early stage. Giving up smoking in later life has a substantial effect on the risk five or ten years later, and therefore it also seems to affect at least one late stage (4). Tobacco smoke contains more than 3500 chemi cals (2). Smoking thus entails exposure to a variety of carcinogens, which include polycyclic aromatic hydrocarbons, such as benzo[a]pyrene, nitrosamines, and aromatic amines. Tobacco smoke is carcinogenic in animals. It induces micronuclei, sister chromatid exchanges, and cell transformation in vitro and has tumor-promoting activity in various test systems (2). 235 Scand J Work Environ Health 1994, vol 20, no 4 Many of the carcinogens in tobacco smoke are acti vated metabolically into DNA-binding intermediates, mainly by reactions mediated by cytochrome P450. DNA adducts derived from tobacco smoke have now been detected in lung and other tissues of cigarette smokers (5--9). In particular, adducts of benzo[a] pyrene diol epox ide to DNA have recently been detected in the lungs of smokers (10). This binding reaction is apparently activated via cytochrome P4501A1, the inducibility of which has been associated with a higher risk for lung cancer among smokers. (For a review, see ref erence 11.) There is some evidence that macrophages can re place the P450 system in activating benzo[a]pyrene (12). Benzo[a)pyrene-7,8-dihydrodiol can be oxi dized by activated macrophages to derivatives that are mutagenic, bind covalently to DNA, and induce sister chromatid exchange. (For a review, see refer ence 13.) Benzo[a]pyrene has been reported to cause oxidative damage in DNA about 20 times more fre quently than it forms adducts (14). Action of asbestos in multistage carcinogenesis Asbestos is not a single mineral. The term covers several fibrous inorganic minerals that share specific properties but differ in chemical composition, mor phology, durability, and, therefore, biological effects. All of the main types of asbestos fibers have, how ever, been shown to be carcinogenic in humans (3). Advances have been made recently towards under standing the molecular mechanisms of asbestos-in duced carcinogenesis. The cytotoxic, genotoxic, and proliferative effects of asbestos seem to be mediat ed in part by active oxygen species -- reactive me tabolites of oxygen that are produced from phago cytic cells or catalyzed by iron on the fiber surface (15, 16). (For a review see references 17 and 18.) Asbestos fibers can induce neoplastic cell transfor mation (19) and chromosome changes in vitro (20); workers exposed occupationally to asbestos have an increased incidence of double-strand breaks in lym phocyte DNA (21). Asbestos fibers can be phagocytized by macro phages, which then release a wide variety of cy tokines and mediators of inflammation that modu late growth and differentiation of the target cells. Damage to DNA in the inflammatory milieu sur rounding asbestos fibers may lead to genetic altera tions, such as the activation of protooncogenes or the inactivation of tumor suppressor genes, and result in the initiation of carcinogenesis and enhanced growth. Interactions of asbestos fibers with the DNA of their target cells can occur through various mechanisms, including oxidant activity and direct interactions with chromosomes. As a result, chromosome alterations and point mutations may occur in cellular protoon cogenes and the expression of genes such as those coding for growth factors may be altered and ulti mately lead to the development of tumors. (For a re view, see references 17 and 18.) Interaction between asbestos and tobacco smoking in causing lung cancer In 1968, Selikoff and his co-workers (22) showed that the exposure of cigarette smokers to asbestos during insulation work was associated with a lung cancer risk far higher than that associated with each agent separately, a finding indicating an interaction between the two agents. Since the time of that sem inal study, the asbestos-tobacco smoking-lung can cer paradigm has often been used as an example in methodological discussions of interactions. In an epidemiologic study, interaction can be clas sified according to the statistical model that most closely describes the relationship between the ob served relative risks among subjects exposed only to tobacco smoke (Rs), only to asbestos (Ra), and to both agents (Rsa), all relative risks being calculated in relation to subjects exposed to neither agent (23). For any Rs and Ra < 1, the two simplest models or relationships of interaction are the additive model (Rsa = Rs + Ra - 1) and the multiplicative model (Rsa = Rs Ra) (24, 25). In this review, the abso lute interaction magnitude, reconstructed from the figures available in the published reports, is classi fied with reference to these two models as follows: less than additive (< A) for Rsa < -25% of A; near additive (~A) for Rsa within 25% of A; additive (A) for Rsa within 10% of A; intermediate (I) for Rsa > 25% of A but below -25% of M; near multi plicative (~ M) for Rsa within 25% of M; multi plicative (M) for Rsa within 10% of M; and more than multiplicative (> M) for Rsa >25% of M. The statistical variability of the measure of inter action was not taken into consideration, but the measures of interaction are very imprecise as the small numbers of lung cancers among nonsmokers exposed to the two risk factors strongly reduce the statistical power. Even though several studies show a similar pattern of interaction, the available data are not sufficient to reject alternative models for mally. The largest epidemiologic studies of the interac tion between tobacco smoking and exposure to as bestos are summarized in table 1, which also gives the absolute measure of interaction derived from the original reports. The overall evidence indicates an interaction in the multiplicative region, although the pattern across studies was not uniform. In particu lar, an additive interaction was observed for Cana dian chrysotile miners and millers (26--28), while for Australian underground crocidolite miners a more than multiplicative interaction was seen (29). For workers exposed to amosite or a combination of 236 Scand J Work Environ Health 1994, vol 20, no 4 Table 1. Studies on the interaction between tobacco smoking and asbestos in the causation of lung cancer, (P = exposure in formation collected prospectively from study subjects, R = exposure information collected retrospectively from study subjects, 0 = exposure information collected from informants other than study subjects, for example, relatives, D = exposure informa tion collected from documents, A = additive, M = multiplicative, I = intermediate) Reference, design, location, years of observation, and exposure Exposure assessment Smoking Asbestos Relative risk3 * * 6 Smokers Non smokers Interaction*1 Berry et al (30), cohort, United Kingdom, 1960--1970, mixed asbestos factory Selikoff & Hammond (35), cohort, United States, 1963--1974, chrysotile and amosite in insulation work Martischnig et al (39), case-referent, United Kingdom, 1973--1974, any asbestos exposure Blot et al (36), case-referent, United States, 1970--1976, shipyards Hammond et al (37), cohort, United States and Canada, 1967--1976, chrysotile and amosite in insulation work Selikoff et al (31), cohort, United States, 1961--1977, amosite factory Blot et al (38), case-referent, United States, 1972--1976, shipyards Pastorino et al (40), case-referent, Italy, 1976--1979, any asbestos exposure Acheson et al (32), cohort, United Kingdom, 1947--1980, amosite factory Liddell et al (28), case-referent in cohort, Canada, 1950--1975, chrysotile mining and milling Baker (29), cohort, Australia, 1944 -- 1981, crocidilite mining and milling Hilt et al (34), cohort, Norway, 1966--1973, maintenance workers exposed to asbestos Berry et al (33), cohort, United Kingdom, 1971--1980, mixed asbestos factory Kjuus et al (41), case-referent, Norway, 1979--1983, any asbestos exposure De Klerk et al (42), case-referent in cohort, Australia, 1979--1986, crocidolite miners Cheng & Kong (43), cohort, China, 1972--1987, mixed asbestos factories D, O D P R R, O P P O R D O, P P P P R P D D R R, O D D O R D D D D D R D D 2.3d 7.4d 5.0 [6.0] [0] [3.2] 1.1 [1.6] [1.3] [5.3] 4.7 [1.6] [1.8] [1.6] [1.7] [5.0] [4.3] 2.0d 4.3d [2.1] 2.6 1.6 [5.7] 25 [1.9] [2.8] [2.0] [3.0] 0.7 0 6.3 12.5 [2.4] 1.9 1.6 [> M) [>M] [>M] >M M 1 ~M 1 [ -- M] -A >M >M [A] [1] --M >M M 3 Relative risk due to asbestos exposure; numbers in square brackets have been reconstructed from published data. b For definitions of the categories of interaction, see the text. The categories in square brackets have been based on the as sumption of a relative risk due to smoking = 10. c Men. 6 Women. amosite, chrysotile, and crocidolite in manufactur ing, variations were seen in the size of the observed interaction (30--34). However, studies of highly ex posed insulators (35--38) and case-referent studies of any type of asbestos exposure (39--41) showed a more uniformly multiplicative pattern. A case-referent analysis nested in the cohort of Australian crocidolite miners (42) confirmed the pat tern of a more than multiplicative interaction, while in a recently published study from China on work ers exposed to chrysotile asbestos in various manu facturing industries, a multiplicative interaction was seen (43). In conclusion, a variable pattern of interaction has been observed which may reflect both the fact that asbestos and smoking act at different stages of the carcinogenic process and the fact that there are dif ferences in the biological effects of different types of asbestos fibers (23). Overall, however, studies of workers exposed to high levels of asbestos, such as insulator workers, point to an interaction that approx imates the multiplicative model. Interaction between asbestos and tobacco smoking in causing different histological types of lung cancer It is often claimed that adenocarcinoma is the histo logical type of lung cancer most frequently associ ated with exposure to asbestos. However, Churg (44) provided evidence that this claim may not be true. 237 Scand J Work Environ Health 1994, vol 20, no 4 He examined the studies in which the distribution of cases of lung cancer with an association to asbestos exposure and those without such an association was reported by cell type. Table 2 adds a few recent stud ies to the data reported by Churg and gives more de tails on the exposed and reference populations used in each study. When cases of squamous-cell carci noma are taken as the reference group, a significant, positive association is seen between exposure to as bestos and small-cell carcinoma, but no association is seen with either adenocarcinoma or large-cell car cinoma. The data reported in table 2 do not allow any conclusion to be drawn with respect to specific types of asbestos fiber. A few recent studies also provide results on the risks for specific cell types of lung cancer associated with asbestos exposure. In a case-referent study in Japan, Minowa and his co-workers (50) found a sig nificantly increased risk for Kreyberg I lung cancer (small-cell and squamous-cell carcinoma) after ex posure to asbestos, as estimated from job titles [odds ratio (OR) 3.40], and a lower, nonsignificant risk (OR 1.72 total number of exposed cases 38) for Kreyberg II lung cancer (adenocarcinoma). In a co hort study of Danish asbestos-cement workers, who were mainly exposed to chrysotile, the standardized incidence ratio (SIR) was higher for adenocarcino ma [SIR 3.31, 95% confidence interval (95% Cl) 2.12--4.92, cases 24] than for squamous-cell carci noma (SIR 1.67, 95% Cl 1.18--2.31, cases 37); a trend was seen for adenocarcinoma, but not for squamous-cell carcinoma, according to the duration of exposure and latency (51). Finally, in a study on lung cancer cases in Finland, in which smoking was adjusted for, the odds ratio for adenocarcinoma as sociated with exposure to asbestos (measured as >3 106 fibers g lung tissue ') was 3.4 (90% Cl 0.7--15) when compared with squamous-cell carci- Table 2. Number of cases of lung cancer by exposure to asbestos and cell type in selected studies. Reference, location, time, exposed population Unexposed population Squamous-cell carcinoma Un Exposed cases exposed cases Small-cell carcinoma Exposed cases Un exposed cases Adeno carcinoma E2xp.o.s.ed exposed,, oases cases Kannerstein & Churg (45), United States (period and source of cases not specified), any type of asbestos Cases without history of asbestos exposure, matched by diagnostic procedure 11 12 11 14 11 9 Large-cell carcinoma Exposed Un; d oases cases 68 Martschnig et al (39). United Kingdom, 1972-- 1973, cases from one hospital, self-reported exposure to any type of asbestos Auerbach et al (46), United States, 1966--1979, three hospitals and a cohort of insulation workers, exposure to any type of asbestos derived from occupational history Ives et al (47), United States, World War II, shipbuilding workers3 Baker et al (48), Australia, 1958--1978. crocidolite miners Mollo et al (55), Italy, 1982-1986, exposure to any type of asbestos derived from occupational history Johansson et al (49), Sweden, 1953--1986, asbestos cement works (>95% chrysotile) Odds ratio for asbestos exposure15 95% confidence interval Cases without self-reported exposure Cases from same hospitals Cases from same hospitals Cases from same province, matched by age and year of diagnosis Definite or probable exposure versus possible or no exposure Cases from same hospital, matched by diagnostic procedure, age, year of duration and gender a Data provided by W Blot (personal communication). b Mantel-Haenszel odds ratio. 238 26 77 96 335 24 103 21 76 22 205 11 42 1 (reference) 22 27 48 102 7 27 9 46 1 21 5 28 1.33 1.00--1.76 18 29 122 5 31 9 64 15 97 9 13 0.93 0.69--1.25 20 89 6 32 4 33 4 13 0.83 0.56--1.23 Scand J Work Environ Health 1994, vol 20, no 4 noma, on the basis of nine and four exposed cases, respectively (52). The differences in the results on the association between asbestos exposure and specific cell types of lung cancer may be due to inadequate control for confounding factors. The source of cases may be an important variable, as shown by Whitwell and his co workers (53), who compared the proportions of dif ferent cell types in series of biopsy samples, surgi cal specimens, and necropsy samples. They found a lower proportion of adenocarcinomas in the biopsy series than in the necropsy series, and an opposite pattern for squamous-cell carcinoma. Other aspects that may play a role are the type of asbestos fiber and characteristics of the exposure pattern, such as duration and intensity. Tobacco smoking increases the risk of all main types of lung cancer. However, the risks of squa mous-cell and small-cell carcinomas of the lung are increased to a greater extent than that of adenocar cinomas (2). Very little information is available on the inter action between tobacco smoking and exposure to as bestos in causing different histological types of lung cancer. In a case-referent study of white male patients admitted to a large hospital in New York State in the United States, occupational exposure to asbestos was assessed from the job titles reported by the patients (54). Figure 1 shows the number of cases and the rel ative risks according to cell type and exposure to smoking and asbestos. On the basis of the catego ries used in table 1, the interaction is ~M for squa mous-cell carcinoma, ~A for small-cell carcinoma, and A for adenocarcinoma. A series of lung cancer patients seen at a hospital in Turin, Italy, was inter viewed with respect to asbestos exposure and smok ing habits (55), and asbestos bodies were counted in samples of normal lung tissue. A case-case analysis was then carried out in which cases of adenocarci noma were compared with cases of squamous-cell carcinoma. The term for interaction between tobac co smoking (more than 20 cigarettes per day versus 0--20 cigarettes per day) and exposure to asbestos (definite or probable exposure versus possible or no exposure) was negative (OR 0.94, 95% Cl 0.23-- 3.86), while that between smoking and asbestos body count (more than 10 000 bodies g 1 versus up to 10 000 bodies g_l) was positive (OR 1.82, 95% Cl 0.08--41.7). Neither interaction term was statistically significant and therefore suggested that there was no difference according to histological type in the inter action between exposure to asbestos and tobacco smoking. Both terms were, however, highly impre cise. The report of the study (52) on lung cancer cases from Helsinki, Finland, provided enough detail to allow an ad hoc analysis of the interaction between smoking (categorized as <50 pack-years versus more) and asbestos exposure (up to 3 106 fibers g lung tissue-1 versus more) by cell type. The age-ad- HR UsMK- ABS- SsMK- ASB+ BSMK+ ASB- BSMK+ ASB+j Figure 1. Numbers of cases and the relative risks accord ing to lung cancer cell type and exposure to smoking and asbestos. The results derive from the case-referent study of Vena et al (49). The analysis was based on crude data. Smoking (SMK) exposure is categorized as 0--39 pack-years (SMK-) versus >40 pack-years (SMK+). Asbestos (ASB) ex posure is categorized as never (ASB-) versus ever (ASB+) exposed. The reference category is SMK- ASB-. The num bers of referents were as follows: SMK- ASB- 718, SMKASB+ 72, SMK+ ASB- 304, SMK+ ASB+ 25. (ca = carcino ma, Adenoca = adenocarcinoma) justed odds ratios for adenocarcinoma, using squa mous-cell carcinoma as the reference, were 1.1 (95% Cl 0.1--9.3) for the subjects exposed to asbestos but not to smoking, 0.4 (95% Cl 0.1--2.8) for the sub jects exposed to smoking but not to asbestos, and 3.2 (95% Cl 0.6--20) for the subjects exposed to both, as compared with subjects exposed to neither agent, a finding suggesting a stronger interaction (ie, clos er to >M than <A) between exposure to asbestos and smoking in the occurrence of adenocarcinoma than in the occurrence of squamous-cell carcinoma. Biological basis of the combined effect In an attempt to explain the multiplicative increase in the risk for lung cancer associated with combined exposure to asbestos and tobacco smoke, hamsters were given intratracheal instillations of asbestos fib ers and benzo[a]pyrene. The substances induced tu mors only when given in combination (56). The mechanisms of the joint effect are not known. It is possible that fibers enhance the penetration of tobac co smoke constituents (such as benzo[a]pyrene) into cells. Cigarette smoke condensate in combination with crocidolite asbestos induces the formation of hydroxyl radicals and DNA strand breaks in isolat ed DNA (57). Asbestos fibers can cause inflamma tory reactions, oxygen radical bursts, and, after long term exposure, fibrosis of the lungs. Tobacco smoke contains several carcinogenic chemicals that bind to DNA and activate protooncogenes and tumor sup pressor genes. Activated oxygen species induce mu 239 Scand J Work Environ Health 1994, vol 20, no 4 tations in mammalian cells (58). Recently, there has been a surge of interest in the role of nitrogen oxide and its derivatives in the pathobiology of chronic in fection and its relationship to the carcinogenic proc ess (59). Macrophages produce inflammatory cy tokines and oxygen radicals when exposed to asbes tos fibers, and it was reported recently that macro phages also produce nitric oxide in response to as bestos (60). Tobacco smoke contains up to 600 pg of nitric oxide and other nitrogen oxides per ciga rette (2, p 95). Excess nitric oxide is mutagenic and induces DNA damage by deaminating nucleotide bases, such as 5-methylcytosine, and inducing strand breaks. Damage to DNA is critical to the process of initi ation in the multistage model of carcinogenesis, and the joint effect of exposure to tobacco smoke and asbestos depends on the relative magnitude of the effects on mutation rates and on the rate of clonal expansion of mutated cells (61). The DNA modifi cation must be sufficiently tenacious to escape effi cient repair processes but not so excessive that cell death results. Common sites for point mutations, in both the K-ras gene and the p53 suppressor gene, are guaninexytosine base pairs in deoxycytidine-3',5'deoxyguanosine dinucleotide sequences (62, 63), so that guaninexytosine base pairs in both tumor sup pressor genes and protooncogenes may represent vul nerable targets for mutation. As has already been described, exposure to asbestos is associated with an excess occurrence of adenocarcinomas in some stud ies, and adenocarcinomas have a higher frequency of K-ras mutations than squamous-cell carcinomas do (52). In a Finnish study of lung adenocarcinoma patients, mutations in the K-ras gene were present in 46% (17 of 37), and guanine-to-thymine transver sions were the predominant type of mutation (41%, 7 of 17) (Ridanpaa et al, unpublished manuscript). Smoking is associated with guanine-to-thymine transversions in codon 12 of the K-ras gene, which can be caused, for example, by benzo[a]pyrene. As bestos fibers can, under suitable circumstances, hydroxylate 2-deoxyguanosine to 8-hydroxydeoxyguanosine, mediated by hydroxyl radicals (64); the pres ence of 8-hydroxyguanine in DNA would also lead to guanine-to-thymine transversions. Guanine-to-adenine and cytosine-to-thymine transitions are also fairly common in lung tumors, especially in the p53 gene (65). The deamination of 5-methylcytosine at cytosine-guanosine dinucleotides by nitric oxide has been suggested to cause cytosine-to-thymine transi tions in vivo (66, 67). Exposure to asbestos may in crease the frequency of transition and transversion mutations indirectly via oxygen radical or nitric ox ide pathways (65) and therefore increase the likeli hood of K-ras mutations in adenocarcinomas (52, 68). Another mechanism could be enhancement of clonal expansion, improved recruitment of mutated cells, and a selective growth advantage to mutated cells. Concluding remarks The joint effect of two exposures, both of which af fect more than one stage of carcinogenesis, depends on the relative magnitude of the effects on early and late stages. Tobacco smoke and asbestos fibers may have interdependent effects on the multistage proc ess of lung carcinogenesis. Tobacco smoke can act at early stages, inducing genetic alterations, DNA adducts, and mutations in genes critical to oncogen esis, and the epidemiologic evidence suggests that tobacco smoke may also act at later stages of car cinogenesis. Asbestos fibers can be cytotoxic and genotoxic and cause proliferative lesions in the lungs, mediated in part by oxygen radicals and nitrogen oxides. Chronic inflammation of the lungs can re lease various cytokines and growth factors which may provide a selective growth advantage to mutated cells. The epidemiologic evidence for the interaction between asbestos and tobacco smoking is clearest in studies of workers exposed to high levels of asbes tos (ie, asbestos insulation workers), and that evi dence points to an interaction that approximates the multiplicative model. In other situations, a variable pattern of interaction has been observed. Tobacco smoking increases the risk for adenocarcinoma of the lung to a less extent than other major histological types. Asbestos exposure is associated with an ex cess occurrence of adenocarcinomas in some, but not all, studies. Unfortunately, the existing data for as sessing the interaction for a particle histological sub type of lung cancer are weak, and further epidemio logic studies are warranted. Acknowledgments We are grateful to H Oshima, PhD, for making help ful comments on the text, to E Heseltine, MSc, for editing the text, and to Ms J Mitchell for typing the manuscript. References 1. Pisani P, Parkin DM, Ferlay J. Estimates of the world wide mortality from eighteen major cancers in 1985: implications for prevention and projections for future burden. Int J Cancer 1993,55:891--903. 2. International Agency for Research on Cancer (IARC). Tobacco smoking. Lyon; IARC, 1986. IARC mono graphs on the evaluation of the carcinogenic risk of chemicals to humans, vol 38. 3. International Agency for Research on Cancer (IARC). Overall evaluations of carcinogenicity: an updating of IARC monographs volumes 1--42. Lyon: IARC, 1987. IARC monographs on the evaluation of carci nogenic risks to humans, suppl 7. 4. Doll R, Peto R. Cigarette smoking and bronchial car cinoma: dose and time relationships among regular smokers and lifelong non-smokers. J Epidemiol Com munity Health 1978;32:303--13. 5. Everson RB, Randerath E, Santella RM, Cefalo RC, Avitts TA, Randerath K. Detection of smoking-relat- 240 Scand J Work Environ Health 1994, vol 20, no 4 ed covalent DNA adducts in human placenta. Science 1986;231:54-7. ' 6. Everson RB, Randerath E, Santella RM, Avitts TA, Weinstein IB, Randerath K. Quantitative associations between DNA damage in human placenta and mater nal smoking and birth weight. JNCI 1988;80:567--75. 7. Phillips DH, Hewer A, Martin CN, Garner RC, King MM. Correlation of DNA adduct levels in the human lung with cigarette smoking. Nature 1988;336:790-- 2. ' " 8. Cusick J, Routledge MN. Jenmkinc D, Courner RC. DNA adducts in different tissues of smokers and non smokers. Int J Cancer 1990;45:673--8. 9. Geneste O, Camus A-M, Castegnaro M, Petruzzelli S, Macchiarini P, Angeletti CA. et al. Comparison of pul monary DNA adduct levels, measured by 32P-postla- belling and aryl hydrocarbon hydroxylase activity in lung parenchyma of smokers and ex-smokers. Carcino genesis 1991:12:1301--5. 10. Alexandrov K, Rojas M, Geneste O, Castegnaro M, Camus A-M, Petruzzelli S, et al. An improved fluori- metric assay for dosimetry of benzo(a)pyrene diol- epoxide-DNA adducts in smokers' lung: comparisons with total bulky adducts and aryl hydrocarbon hydrox ylase activity. Cancer Res 1992:52:6248--53. 11. Vahakangas K, Pelkonen O. Host-variations in carci nogenic metabolism and repair. In: Lynch HT, Hiraya- ma T, editors. Genetic epidemiology of cancer. Boca Raton, FL: CRC Press, 1989:35--54. 12. Trush MA, See JL, Kensler TW. Oxidant-dependent metabolic activation of polycyclic aromatic hydrocar bons by phorbol ester-stimulated human polymorpho nuclear leukocytes: possible link between inflamma tion and cancer. Proc Natl Acad Sci USA 1985;82: 5194--8. 13. Weitzman SA, Gordon LI. Inflammation and cancer: role of phagocyte-generated oxidants in carcinogene sis. Blood 1990:76:655--63. 14. Ide ML, Kaneco M, Cerutti P. Benzo[a]pyrene and ascorbate-CuS04 induce DNA damage in human cells by indirect action. In: McBrian DCH, Slater TF, edi tors. Protective agents in cancer. New York, NY: Ac ademic Press, 1983; 125--40. 15. Lund LG, Aust AE. Iron mobilization from crocidol- ite asbestos greatly enhances crocidolite-dependent formation of DNA single-straind breaks in 4>X 174 RFI DNA. Carcinogenesis 1992; 13:637--42. 16. Korkina LG, Dumev AD, Suslova TB, Cheremisina ZP, Daugel-Dauge NO, Afanas'ev IB. Oxygen radi cal-mediated mutagenic effect of asbestos on human lymphocytes: suppression by oxygen radical scaven gers. Mutat Res 1992;265:24-5--53. 17. Barrett JC. Mechanisms of action of known human carcinogens. In: Vainio H, Magee P, McGregor D, McMichael AJ, editors. Mechanisms of carcinogene sis in risk identification. Lyon: International Agency for Research on Cancer (IARC), 1992; 115--34. IARC scientific publications, no 116. 18. Mossman BT. Mechanisms of asbestos carcinogene sis and toxicity: the amphibole hypothesis revisited [editorial], Br J Ind Med 1993;50:673-6. 19. Hesterberg TW, Barrett JC. Dependence of asbestos- and mineral dust-induced transformation of mamma lian cells in culture on fiber dimension. Cancer Res 1984;44:2170-80. 20. Oshimura M, Hesterberg TW, Tsutsui T, Barrett JC. Correlation of asbestos-induced cytogenetic effects with cell transformation of Syrian hamster embryo cells in culture. Cancer Res 1984;44:5017--22. 21. Marczynski B, Czuppon AB, Marek W, Reichel G, Baur, X. Increased incidence of DNA double-strand breaks and anti-ds DNA antibodies in blood of work ers occupationally exposed to asbestos. Human Exp Toxicol 1994;13:3-9. ` 22. Selikoff IJ, Hammond EC, Churg J. Asbestos expo sure, smoking, and neoplasia. JAMA 1968;204:104-- 10. 23. Saracci R, Boffetta P. Interactions of tobacco smok ing with other causes of lung cancer. In: Samet J, ed itor. Epidemiology of lung cancer. New York, NY: Marcel Dekker, 1994;471--99. 24. Rothman KJ. The estimation of synergy or antagonism. Am J Epidemiol 1976:103:506--11. 25. Saracci R. Interaction and synergism. Am J Epidemi ol 1980; 112:465--6. 26. McDonald JC, Liddell FDK, Gibbs GW, Eyssen GE, McDonald AJ. Dust exposure and mortality in chryso- tile mining 1910--75. Br J Ind Med 1980;37:11-24. 27. Thomas DC. Statistical methods for analyzing effects of temporal patterns of exposure on cancer risks. Scand J Work Environ Health 1983;9:353--66. 28. Liddell FDK, Thomas DC, Gibbs JW, McDonald JC. Fibre exposure and mortality from pneumoconiosis, respiratory and abdominal malignancies in chrysotile production in Quebec, 1926--75. Ann Acad Med Sin gapore 1984; 13 suppl 2:340--4. 29. Baker JE. Lung cancer incidence amongst previous employees of an asbestos mine in relationship to cro- cidolite exposure and tobacco smoking [thesis]. Perth: University of Western Australia, 1985. 30. Berry G, Newhouse ML, Turok M. Combined effect of asbestos exposure and smoking on mortality from lung cancer in factory workers. Lancet 1972:2:476-- 9. 31. Selikoff IJ, Seidman H, Hammond EC. Mortality ef fects of cigarette smoking among amosite asbestos fac tory workers. JNCI 1980;65:507-13. 32. Acheson ED, Gardner JM, Winter PD, Bennett C. Can cer in a factory using amosite asbestos. Int J Epide miol 1984;13:3-10. 33. Berry G, Newhouse ML, Antonis P. Combined effect of asbestos and smoking on mortality from lung can cer and mesothelioma in factory workers. Br J Ind Med 1985;42:12-8. 34. Hilt B, Langird S, Andersen A, Rosenberg J. Asbes tos exposure, smoking habits, and cancer incidence among production and maintenance workers in an elec trochemical plant. Am J Ind Med 1985;8:565--77. 35. Selikoff IJ, Hammond ED. Multiple risk factors in en vironmental cancer. In: Fraumeni J, editor. Persons at high risk of cancer. New York, NY: Academic Press, 1975:467--83. 36. Blot WJ, Harrington JM, Toledo A, Hoover R, Heath CW. Fraumeni JF Jr. Lung cancer after employment in shipyards during World War II. N Engl J Med 1978; 299:620--4. ` 37. Hammond EC, Selikoff IJ, Seidman H. Asbestos ex posure, cigarette smoking and death rates. Ann NY Acad Sci 1979:330:473--90. 38. Blot WJ, Morris LE, Stroube R, Tagnon I, Fraumeni JF. Lung and laryngeal cancer in relation to shipyard employment in coastal Virginia. JNCI 1980:65:571 -- 5. ' 39. Martischnig KM, Newell DJ, Barnsley WC, Cowan WK, Feinmann EL, Oliver E. Unsuspected exposure to asbestos and bronchogenic carcinoma. Br Med J 1977;1:746-9. 40. Pastorino U, Berrino F, Gervasio A, Pesenti V, Ribo- li E, Crosignani P. Proportion of lung cancers due to occupational exposure. Int J Cancer 1984;33:231--7. 41. Kjuus H, Skjterven R, Langard S, Lien JT, Aamodt T. A case-referent study of lung cancer, occupational ex posures and smoking: II. role of asbestos exposure. Scand J Work Environ Health 1986;12:203-9. 42. De Klerk NH, Musk AW, Armstrong BK, Hobbs MST. Smoking, exposure to crocidolite, and the incidence of lung cancer and asbestosis. Br J Ind Med 1991 ;48: 412--7. 241 Scand J Work Environ Health 1994, vol 20, no 4 43. Cheng WN, Kong J. A retrospective mortality cohort study of chrysotile asbestos products workers in Tian jin 1972--1987. Environ Res 1992;59:271-8. 44. Churg, A. Lung cancer cell type and asbestos expo sure. JAMA 1985;253:2984-5. 45. Kannerstein M, Churg J. Pathology of carcinoma of the lung associated with asbestos exposure. Cancer 1972;30:14-21. 46. Auerbach 0, Garfmkel L, Parks VR, Conston AS, Galdi VA, Joubert L. Histologic type of lung cancer and asbestos exposure. Cancer 1984;54:3017--21. 47. Ives JC, Buffler PA, Greenberg SD. Environmental association and histopathological patterns of carcino ma of the lung: the challenge and dilemma in epide miologic studies. Am Rev Respir Dis 1983;128:195-- 209. 48. Baker JE, Reutens DC, Graham DF, Sterrett, GF, Musk AW, Hobbs MST, et al. Morphology of bron chogenic carcinoma in workers formerly exposed to crocidolite at Wittenoom Gorge in Western Austral ia. Int J Cancer 1986; 37:547--50". 49. Johansson L, Albin M, Jakobsson K, Mikoczy Z. His tological type of lung carcinoma in asbestos cement workers and matched controls. Br J Ind Med 1992;49:626-30. 50. Minowa M, Hatano S, Ashizawa M, Oguro H, Naruhashi H, Suzuki M, et al. A case-control study of lung cancer with special reference to asbestos exposure. Environ Health Perspect 1991 ;94:93--42. 51. Raffn E, Lynge E, Korsgaard B. Incidence of histo logical type among asbestos cement workers in Den mark. Br J Ind Med 1993;50:85-9. 52. Husgafvel-Pursiainen K, Hackman P, Ridanpaa M, Anttila S, Karjalainen A, Partanen T, et al. Kras mu tations in human adenocarcinoma of the lung: associ ation with smoking and occupational exposure to as bestos. Int J Cancer 1993;53:250--6. 53. Whitwell F, Newhouse ML, Bennett DR. A study of the histological types of lung cancer in workers suf fering from asbestosis in the United Kingdom. Br J Ind Med 1974;31:298-303. 54. Vena JE, Byers TE, Cookfair D, Swanson M. Occu pation and lung cancer risk: an analysis by histologi cal subtypes. Cancer 1985;56:910--7. 55. Mollo F, Beilis D, Delsedime L, Bernardi P, Ardissome F, Piolatto G, et al. Autopsy indicators of expo sure to asbestos and lung cancer. In: Riboli E, Delendi M, editors. Autopsy in epidemiology and medical research, Lyon: International Agency for Research on Cancer (IARC), 1991;141--7. IARC scientific publi cations, no 112. 56. Kimizuka G, Azuma M, Ishibashi M, Shimozaki K, Hayashi Y. Co-carcinogenic effect of chrysotile and amosite asbestos with benzo(a)pyrene in the lung of hamsters. Acta Pathol Jpn 1993;43:149--53. 57. Jackson JH, Schraufstatter IU, Hyslop PA, Vosbeck K, Sauerheber R, Weitzman SA, et al. Role of oxidants in DNA damage: hydroxyl radical mediates the syn ergistic DNA damaging effect of asbestos and ciga rette smoke. J Clin Invest 1987;80:1090--5. 58. Weitzman SA, Stossel TP. Mutation caused by human phagocytes. Science 1981 ;212:546--7. 59. Ohshima H, Bartsch H. Chronic infections and inflam matory processes as cancer risk factors: possible role of nitric oxide in carcinogenesis. Mutat Res. In press. 60. Prewitt TW, Chaudhri G, Pass H. Asbestos and inter- fon gamma synergistically induce murine macrophages to produce nitric oxide [abstract no 656]. Proc Am Assoc Cancer Res 1993;34:445. 61. Kodell RL, Krewski D, Zielinski JM. Additive and multiplicative relative risk in the two-stage clonal ex pansion model of carcinogenesis. Risk Anal 1991; 11: 483--90. 62. Nigro J, Baker SJ, Presinger AC, Jessup JM, Hostetter R, Cleary K, et al. Mutations in the p53 gene occur in diverse human tumor types. Nature 1989;342: 705--8. ` 63. Bos JL. ras Oncogenes in human cancer: a review. Cancer Res 1989;49:4682---9. 64. Kasai H, Nishimura S. DNA damage induced by as bestos in the presence of hydrogen peroxide. Gann 1984;75:841-4. 65. Rydberg D, Kure E, Lystad S, Skaug V, Strangeland L, Mercy I, et al. p53-Mutations in lung tumours: re lationship to putative susceptibility markers for can cer. Cancer Res 1994;54:1551--5. 66. Wink DA, Kasprzak CM, Maragos CM, Elespuru RK, Misra M, Dunams TM, et al. DNA deaminating abili ty and genotoxicity of nitric oxide and its progenitors. Science 1991;254:1001-3. 67. Nguyen T, Brunson D, Crespi CL, Penman BW, Wish- nole JS, Tannenbaum SR. DNA damage and mutation in human cells exposed to nitric oxide in vitro. Proc Natl Acad Sci USA 1992:89:3030--4. 68. Vainio H, Husgafvel-Pursiainen K, Anttila S, Kar jalainen A, Hackman P, Partanen T. Interaction be tween smoking and asbestos in human lung adenocar cinoma: role of K-ras mutations. Environ Health Per spect 1993; 101 suppl 3:189--92. Received for publication. 24 February 1994 242