Document X7N5jp2rr2mqEj7NxjNky1LBB

Regulatory Toxicology and Pharmacology 34.170-177 (2001) ^ dol:10.10Q6/rtph2001.H91. available online at http-y/www.ideallbrarycom on iDEj^L Malignant Mesothelioma and Asbestos Exposure among Auto Mechanics: Appraisal of Scientific Evidence Otto Wong*ft *Applied Health Sciences Inc., P.O. Box 2078, San Mateo, California 94401; \Bepartment ofEpidemiology, Tulane University, New Orleans, Louisiana; and^Department ofCommunityand Family Medicine. Chinese UniversityofHong Kong, Hong Kong Received March 24,2001 In 1986 the U.S. Environmental Protection Agency (EPA) issued an official guideline on the prevention of asbestos disease among auto mechanics. In the EPA guideline, malignant mesothelioma was listed as a consequence of exposure to asbestos fibers from brake linings and clutch facings among auto mechan ics. EPA' formulated its 1986 opinion by relying solely on a few outdated case reports and not on epidemi ologic studies. A review of the literature indicates, that there are six epidemiologic studies providing relevant information on malignant mesothelioma among auto mechanics. Three of the six studies had already been published by 1986, the year in which EPA issued its guideline. The results ofthe six studies were remarkably consistent in that all six studies re ported no increased risk of malignant mesothelioma among auto mechanics. The relative risks reported in the six studies ranged from 0.62 to 1.00. Based on a meta-analysis of the combined data of all six studies consisting of approximately 1500 malignant mesothe lioma cases, the mesothelioma relative risk for auto mechanics is 0.90 (95% confidence interval 0.66 -1.23). An application of Hill's causation criteria to epidemi ologic data of malignant mesothelioma among auto mechanics clearly demonstrates that auto mechanics do not have an increased risk of malignant mesothe lioma as a result of exposure to asbestos fibers from brake linings and clutch facings. However, in spite of the scientific evidence, EPA has not modified or revised its 1986 guideline. Occupational regulatory policies and guidelines, when based on proper scien tific evidence, are invaluable and can prevent avoid able diseases in workers or other exposed individu als in the general public. On the other hand, it is the regulators' responsibility to develop, modify, and re vise policies and guidelines in accordance with the most relevant and the latest scientific data. In this instance EPA as a regulator has not fulfilled its re sponsibility of providing the most accurate and upto-date information to the workers or the general public. <0 2001 Academic Press Key Words: regulation; causation; case reports; epi demiology; case-control studies; asbestos; chrysotile; brakes; dutches; auto mechanics; malignant mesothe lioma. INTRODUCTION Regulatory poliefes and guidelines for occupational exposures are Intended to prevent avoidable diseases among workers and are supposed to be based first and foremost on sdentific evidence. Definitive scientific ev idence, however, may take years or even decades to ac cumulate. It is, therefore, the regulators' responsibility to develop, to update, and to revise, if necessary, poli cies and guidelines in accordance with the latest avail able scientific data. Unfortunately, this is not always the case. An example is the U.S. Environmental Pro tection Agency's (EPA's) guideline for the prevention of asbestos disease among auto mechanics who are poten tially exposed to asbestos fibers from brake linings and clutch facings. In 1986 EPA issued an official guideline, in which malignant mesothelioma was listed as a consequence of exposure to asbestos fibers generated from work ing with friction products such as brake linings and clutch facings among auto mechanics (EPA, 1986). In the 1986 EPA document several references were cited in support of the Agency's opinion on malignant mesothe lioma among auto mechanics. All references cited were case reports. Because not all malignant mesothelioma cases are related to occupational exposure to asbestos (Peterson eta!., 1984; Pelnar. 1988). using case reports as evidence for causation is hot only inappropriate but also misleading. For example, based on the Finnish Tumor Registry, malignant mesothelioma cases have been reported among the following occupations which do not entail asbestos exposures: judge, banker, book keeper. musician, school teacher, tailor, student, nurse, midwife, cook, seamstress, and maid (Nurminen, 1975). Based, on these case reports, one cannot conclude that ) 0273-2300/01 $35.00 Copyright <> 2001 by Academic Press AU rights of reproduction in any form reserved. 170 HWBUI0010236 MALIGNANT MESOTHELIOMA IN AUTO MECHANICS ' \ 's 171 these occupations are at an increased risk of develop ing malignant mesothelioma. Such a conclusion must be based on epidemiologic studies relevant to the occu pations In question. The 1986 EPA document did not cite any epidemiologic studies of malignant mesothe lioma among auto mechanics, even though three epi demiologic studies providing relevant information on malignant mesothelioma among auto mechanics had been published by 1986 (see below). Case reports are reports of isolated cases of diseases in individuals with certain characteristics or potential risk factors. However, the underlying population at risk is not defined In case reports and, therefore, no mea sure of disease frequency can be estimated. Further more, case reports do not Involve any comparison to individuals without the characteristics (or exposures) and, therefore, cannot provide any measure of risk rel ative to individuals with no exposure. In its 1986 docu ment EPA did not discuss the uncertaintiesofrelyingon case reports alone, even though the limitations of case reports are well recognized by scientists and regulators alike. For example, in its monographs on carcinogenic ity evaluation, the International Agency for Research on Cancer (IARC, 1989) discussed the uncertainties of case reports: "Case reports usually lack complete ascer tainment of cases in any population, definition or enu meration ofthe population at risk and estimation ofthe expected number of cases in the absence of exposure.'' Thus, case reports, in or by themselves, cannot be used as evidence for causation. Case reports, however, do serve some useful purposes: alerting researchers of the potential association be tween the exposure and disease in question and provid ing valuable information for priority setting in allocat ing research resources. However, it must be understood that case reports do not provide any definitive evidence for causation in or by themselves. A proper causation assessment must be based on appropriate epidemiologic studies. With regard to the use of epidemiologic studies in causation assessment, IARC (1989) remarked: "Co hort and case-control studies relate individual exposure to the agent under study to the occurrence of cancer in individuals, and provide an estimate of relative risk as the main measure of association." Epidemiologic studies have both confirmed as well as refuted associations previously suggested by case reports. Although it is prudent for regulators to issue early warnings at the first sign of danger suggested by case reports, it is equally prudent for regulators to re vise their positions ifsubsequent epidemiologic studies prove otherwise. In the case ofmalignant mesothelioma and asbestos exposure among auto mechanics, as will be demonstrated below, proper epidemiologic studies are available and they refute the association suggested by earlier case reports. The following example illustrates the danger of relying on case reports. In 1991, Woitowltz and Rodelsperger reported several cases of malignant mesothelioma in garage mechanics at a clinic in Giessen. Germany. No Information on the size of the underlying population of garage mechanics at risk was available and no measure of risk was provided. Nevertheless, based on the few cases alone. Woitowitz and Rodelsperger (1991) concluded: "In summary, we have found an increased incidence (emphasis added) of mesothelioma among car mechanics exposed only to chrysotile." Wong (1992) pointed out that the Woitowltz and Rodelsperger (1991) report was not an epidemio logic study, but a case report. As such, it would not have been possible to calculate an incidenceofmesothelioma among garage mechanics, since the underlying popu lation of garage mechanics at risk was undefined and unknown to the authors. Furthermore, no conclusion on the relationship between car mechanics and mesothe lioma should be drawn from such a case report. Sub---- sequent to this exchange in the literature, Woitowitz and Rodelsperger (1994) did complete an epidemio logic case-control study of mesothelioma and garage mechanics. The analysis Indicated that therewas no in creased risk of mesothelioma in garage mechanics (see below). Based on their 1994 epidemiologic investiga tion, Woitowitz and Rodelsperger reversed their earlier opinion based on the 1991. case report and concluded: "From these results there Is no evidence that car me chanics are exposed to increased risk of mesothelioma even if they do brake repairs," This example serves to demonstrate the treacherous nature of relying on case reports for causality assessment. Evidence for a causal relationship must be based on epidemiologic studies. EPIDEMIOLOGIC STUDIES OF MESOTHELIOMA AMONG AUTO MECHANICS Because malignant mesothelioma Is a relatively rare disease, the study design of choice is case-control study. The measure ofrisk in a case-control study is derived by comparing the proportion ofcases (individuals with the disease) with exposure to that among the controls (In dividuals without the disease). In a large case-control study of malignant mesothelioma In North America, McDonald and McDonald (1980) ascertained, through 7400 pathologists in North America, all fatal cases of malignant mesothelioma in Canada between 1960 and 1972 and all fatal cases of mesothelioma in the United States in 1972. Patients diagnosed with nonpuimonary malignant tumors at the same hospitals were selected as controls. A one-to-one case-to-control ratio was used. Cases and controls were matched with respect to sex, age, and year of death. Employment histories of both cases and controls were obtained through a blinded HWBUI0010237 172 OTTO WONG TABLE 1 Risk of Malignant Mesothelioma by Occupation In a Case-Control Study In Canada and the United States, 1960-1972 Occupation or industry Relative risk Insulation Asbestos products and manufacturing Heating trades Shipyards Construction industry Garage mechanics Source. McDonald and McDonald (1980J. 46.0 6.1 4.4 2.8 2.6 0.9 Interview. Analysis was based on a comparison ofthe oc cupational exposure historiesof480 cases and theircon trols. For insulators, a significant increase of mesothe lioma risk was found, with a relative risk of46-fold. For asbestos, production such as mining and milling, the rel ative risk was 6.1 . Workers in heating trades (furnaces, boilers, plumbing, steam pipes, and welding) were re ported to have a relative risk of4.4. For shipyard work ers the risk was 2.8 times that of the general population (Table 1). On the other hand, there were 11 mesothe lioma cases who had worked as garage mechanics, com parable to 12 controls who had done so. In a correspon dence, one ofthe authors stated: "The equal numbers of garage workers in cases and controls indicates (sic) that no increased risk ofgarage workwasfound in ourstudy" (letter from A. D. McDonald to O. Wong dated June 4, 1986). The relative risk ofmesothelioma for garage me chanics based on this study was estimated to be 0.91, similar to the background risk of 1.00 for individuals with no occupational exposure to asbestos. The corre sponding 95% confidence interval (95% GI) was (0.392.13). McDonald and McDonald (1980) concluded: "no increase in risk was found In garage workers, certainly exposed to chrysotile from brake linings." Incidentally, the authors also stated: "For nearly half the male cases and for about 5% of female cases, the tumor could be at tributed to occupational exposure to asbestos, of which a fifth were in shipyards." This observation emphasizes the large number of mesothelioma cases that are not related to occupational exposure to asbestos. In 1983 another case-control study of malignant mesothelioma and occupation that provided informa tion on auto mechanics was reported by Teta et al. (1983). Cases were identified primarily through the Connecticut Tumor Registry {Connecticut State Health Department), one of the oldest and most comprehen sive tumor registries in North America. The case as certainment was supplemented by a search of records at the largest Veterans Administration hospital in the state. The study consisted of 220 cases diagnosed be tween 1955 and 1977. Controls were randomly selected from the mortality registry in the state during the same time interval. Medical, demographic, and occupational exposure information was obtained for both cases and controls. Plumbers and pipefitters were found to have significantly increased risk (relative risk of 3.87). A sig nificantly elevated relative risk of 2.25 was also found for cabinetmakers and carpenters. The relative risk of mesothelioma for `automobile repair and related ser vice" was 0.65. The finding was based on 6 subjects who had worked as auto mechanics, and the correspond ing 95% Cl was rather wide (0.08-5.53). Even though the study was relatively small, the result indicated that there was no increase of mesothelioma associated with garage work, which was consistent with the result reported in the much larger study by McDonald and McDonald (1980). In a case-control study conducted by epidemiologists at the National Cancer Institute (NCI), the occupa tional exposures of 259 mesothelioma patients diag nosed between 1975 and 1980 in New York State (ex clusive of New York City) and Los Angeles County were compared to those of the Controls (Splrtas etal., 1985). Significantly elevated relative risks were reported for workers in several occupations: furnace or boiler Instal lation or repair, 4.5. (95% Cl 2.5-8.2); shipbuilding, ship demolition, or shipyard work, 4.3 (95% Cl 2.S-7.4); and insulation, 3.2 (95%Cl 2.1-5.0). In contrast, the relative risk for "brake lining installation and repair" was 1.0 (95% Cl 0.6-1.6). That is, mesothelioma risk for work ers engaged in brake lining installation and repair was the same as the general population. In a second report based on the same database. Splrtas et al. (1994) fo cused their analysis on attributable risk of asbestos ex posure in general, and no Individual risk estimates were calculated for specific occupations separately. However, the numbers and percentages of cases and controls in each occupation were reported. For several occupations, a higherpercentage ofcases than controls was reported: furnace or boiler installation or repair, 51 (24.5%) vs 39 (7.3%); insulation. 83 (39.9%) vs 83 (15.6%); shipbuild ing or repair. 67 (32.2%) vs 40 (7.5%); and plumbing or heating, 72 (34.6%) vs 125 (23.5%). On the other hand, there were 33 (15.8%) cases and-72 (13.5%) controls who had engaged in brake lining work or repair. The almost equal percentages of cases and controls Indicated that there was no increased risk associated with brake lining work, consistent with the more specific result presented in the earlier report (Splrtas et al., 1985). In a recent case-control study conducted at the Insti tute and Outpatient Clinic of Occupational and Social Medicine of the Justus-Llebig University in Giessen, Germany, 16 (4.9%) of the 324 cases of mesothelioma were classified as motor vehicle mechanics, compared to 16 mechanics (5.1%) among 315 hospital controls and 12 mechanics (6.6%) among 182 population controls (Woitowitz and Rodelsperger. 1994). According to the authors, the motor mechanics began their employment between 1925 and 1972. Based on a comparison to the combined hospital and population controls, the relative HWBUI0010238 MALIGNANT MESOTHELIOMA IN AUTO MECHANICS 173 TABLE 2 Bisk of Malignant Mesothelioma by Occupation in a Case-Control Study In British Columbia, 1990-1992 Occupation Relative risk 95% Cl Sheet metal workers 9.6 1.5-106 Plumbers and pipefitters 8.3 1.5-86.3 Shipbuilding workers 5.0 l.Z-22.7 (not classified elsewhere) Welders 3.9 0.8-21.8 Bricklayers, plasterers, and cement workers 3.5 0.9-14.0 Miners, drillers, and blasters 3.4 0.9-13.1 Construction foremen 3.1 0.9-11.0 Stationary engineers and boilermakers 1.8 0.5-5.9 Office derks and secretaries 1.3 0.5-3.5 Teachers and librarians 1.0 0.2-3.7 Sales clerks 0.9 0.4-2.1 Vehicle mechanics 0.8 0.2-2.3 Source. Teschke ef aS. (1997). risk for alt motor vehicle mechanics was 0.87 (95% Cl 0.46-1.64). More specifically, for motor vehicle mechan ics definitely engaged in brake service, the relative risk was 0.76 (95% Cl 0.28-2.05). The authors concluded: "From these results there is no evidence that car me chanics are exposed to an Increased risk of mesothe lioma even If they do brake repairs." In another recent case-control study In British Columbia, the investigators at the University of British Columbia and the University of Washington compared the asbestos exposure histories of 51 cases of pleural mesothelioma to those of 154 population-based con trols (Teschke et al,, 1997). The mesothelioma cases, diagnosed between September 1990 and August 1992, were identified through the British Columbia Cancer Agency. Interviews to obtain exposure histories were conducted by a registered nurse, and the exposure Infor mation was reviewed by an industrial hygienist. The fol lowing significant risk ratios were reported for several occupational groups: 9.6 for sheet metal workers, 8.3 for plumbers and pipefitters, and 5.0 for shipbuilding workers (Table 2). In contrast, there were 6 mesothe lioma cases and 20 controls who had worked as ve hicle mechanics, and the relative risk was 0.8 with a 95% Cl of (0.2-2.3). Thus, there was no Increased risk reported for vehicle mechanics in the University of British Columbia/University of Washington study. The investigators also examined the risk for mechanics who were specifically Involved in brake lining installation or repair. Again, no increased risk was found (relative risk of 0.3,95% Cl 0-1.4). Thus, the authors concluded: "As with vehicle mechanics in the occupational analysis, a history of brake lining Installation or repair had a risk estimate below 1.0." Finally, the Institute of Epidemiology and Clinical Research in Spain conducted a case-control study of mesothelioma and occupation (Agudo et al., 2000). The study consisted of 132 confirmed mesothelioma cases . .. diagnosed In two provinces. Barcelona and Cadiz. Age- ) and sex-matched controls were selected from other pa tients admitted to the same hospitals. The occupational exposurehistories of 132 mesothelioma cases were com pared to those of 257 matched controls. For occupa tions with at least 5 cases or a significant relative risk, the numbers of cases and controls as well as rel ative risks were reported. Significantly elevated rela tive risks were reported for machinery fitters (3.59), machinery fitters and assemblers (4.07), electrical fit ters (9.10), plumbers (7.49), and bricklayers (1.99). The authors reported 3 mesothelioma patients who had worked as vehicle mechanics, compared to 14 controls. Because the number of cases in vehicle mechanics was 3 (less than 5) and the relative risk was not significant, no risk estimate was presented in the report. However, based on the numbers presented by the authors, the relative risk was estimated to be 0.62 (95% Cl 0.17- 2.25). Thus, there was no Increased risk of mesothe lioma among vehicle mechanics in Spain. The authors also stated: "Overall, 37% of cases of mesothelioma in our population are attributed to a sure (or almost sure) occupational exposure to asbestos and this proportion comesup to 62%when occupations with any probability ofexposure are included " This observation underscores the high proportion of malignant mesothelioma cases not related to occupational exposure to asbestos. . n- META-ANALYSIS Based on these epidemiologic studies, clearly there is no evidence to support or even to suggest an associa tion between an increased risk ofmesothelioma and ex posure to brake linings or dutch facings among garage mechanics. On the contrary, the six studies reviewed above provide strong evidence that garage mechanics' mesothelioma risk Is similar to that tat the general pop ulation. None of the six studies reported any increased mesothelioma risk among auto mechanics. Some of the individual studies were relatively small, and the predsion of the risk estimates might not have bewi as high as we would prefer (as indicated by the wide 95% CIs). Because of the consistency of the results, the precision of the risk estimate can be improved by combining these studies through a meta-analysis. The primary purpose of the meta-analysis was to provide a summary mea sure of risk with Improved precision (i.e., a narrower 95% Cl). Meta-analysis is a well-accepted procedure in reviewing and summarizing epidemiologic studies of individuals with similar exposures (Wong and Raabe, 1996). For example, meta-analysis was used by the Sur geon General's Committee (1964) in summarizing epi demiologic studies on smoking and cancer. An obvious benefit of meta-analysis is the increase ) in sample size and the narrow confidence interval of ' -' HWBUI0010239 174 OTTO WONG TABLE3 Individual Study Results and Meta-analysis of Malignant Mesothelioma among Auto Mechanics Study (author and year) Relative risk 95% Cl London School ot Hygiene and Tropical Medicine, United Kingdom (McDonald and McDonald, 1980) Connecticut State Health Department (Feta eta/., 1983) US. National Cancer Institute (Spirtas 1985. 1994) Justus-Lteblg University, Germany (Woltowitz and Rodelsperger, 1994) University of British Columbia and University of Washington (Teschfce etal, 1997) Institute of Epidemiology and Clinical Research, Spain (Agudo etal., 2000) Combined data of all six studies (meta-analysls) 0.91 0.65 1.00 0.87 0.80 0.62 0.90 (0.39-2.13) (0,08-5.53) (0.60-1.60) (0.46-1.64) (0.20--2.30) (0.17-2.25) (0.66-1.23) the summary risk estimate based on the pooled or combined data. In reviewing studies for carcinogenic potential, IARC (1989) commented: "When several epidemiological studies show little or no indication of am association between an agent and cancer, the judg ment may be made that, fn the aggregate, they show ev idence of lack ofcarcinogenicity." Such a situation calls for the application ofmeta-analysis. With regard to the use of meta-analysis In this context, IARC stated: "In addition, all studies that arejudged to be methodolog ically sound should be consistent with a relative risk of unity for any observed level of exposure to the agent and, when considered together, should provide a pooled estimate of relative risk which is at or near unity and has a narrow confidence interval, due to sufficient pop ulation size." The results of the six case-control studies of mesothe lioma are presented in Table 3. The relative risks re ported in the six individual studies ranged from 0.62 to 1.00. The result of the meta-analysis (l.e., the summary relative risk) based on all six studies is also presented in Table 3. The weights used in the meta-analysis were Inversely proportional to the variance of the risk esti mates in the Individual studies (Armitage and Berry, 1994). In other words, larger studies received more weight and vice versa. According to the meta-analysis of the combined data ofall six studies consisting ofapprox imately 1500 mesothelioma cases, the mesothelioma relative risk for auto mechanics is 0.90 (95% Cl 0.661.23). Thus, Table 3 shows that all six studies, individ ually or in aggregate, demonstrate that auto mechanics are not at an increased risk ofmalignant mesothelioma, when compared to the general population. DISCUSSION More than three and half decades ago the criteria for assessing causation in chronic diseases, such as can cers, was formulated by Hill (1965) and utilized by the Surgeon General's Committee in assessing the relation ship between smoking and cancer (Surgeon General's Committee, 1964) as well as by IARC In the evalua tionofcarcinogenicity ofsubstances (IARC, 1989). Hill's criteria include strength of association, consistency -of association, specificity of association, dose-response (or exposure-response) relationship, and temporality or la tency. In evaluating the causal relationship, if any. be tween exposure to asbestos fibers from brake linings and clutch facings among auto mechanics and malig nant mesothelioma. Hill's criteria must be considered and applied directly to the relevant epidemiologic data. With respect to strength of association, the metaanaiysis ofthe combined data from all sixstudies shows a summary relative risk bf 0.90. meaning no increased' risk. Furthermore, the narrow 95% Cl (0.66-1.23) of the summary relative risk is indicative of the large un derlying database and the precision of the summary risk estimate. Furthermore, the results of the individ ual studies are remarkably consistent In that none of the studies shows an Increased risk. Specificity in both exposures and disease endpoints is one the most important criteria in causation assess ment (Wong, 1993). If the subject matter is exposure to asbestos from friction products among auto mechanics, we must base our evaluation on studies of auto mechan ics and not on other occupations. For example, It is not valid to apply results of studies of insulators or ship yard workers to auto mechanics, because workers in these vastly different occupations are exposed to differ ent types of asbestos as well as to different concentra tion levels of asbestos. In the studies reviewed above, mesothelioma risk of not only all garage mechanics but also those specifically engaged in brake repairs was as sessed. Exposure-response relationship is yet another im portantcausation criterion. Fiber measurement studies have Indicated that garage mechanics are exposed to extremely low levels of asbestos as a result of the thermal transformation of asbestos fibers during the braking process. The asbestos fibers in brake linings are entirely chrysotlle fibers in a bonded or encapsulated state. During braking, a temperature in excess of 700800C is reached, and the chrysotlle fibers break down to forsterlte (an anhydrous magnesium silicate), which is nonfibrous. Studies have shown that brake lining dust contains extremely little recognizable chcysotile fibers, which are mostly embedded in the plastic bond ing material, and that the amount of free chrysotlle fibers is less than 0.1 to 1% and most fibers are less than 0.5 pm in length (Hatch, 1970; Lynch. 1968; Davis and Conlam, 1973: Williams and Muhlbaier, 1982). HWBUI0010240 MALIGNANT MESOTHELIOMA IN AUTO MECHANICS 175 TABLE 4 Tune-Weighted Average Exposures to Asbestos at Garages Study (authors and year) Location Fiber concen tration (f/cc) Rodelsperger etal, 1986 Kaupinnen and Korhonen, 1987 Moore, 1988 Cheng and O'Kelly, 1986 NIOSH, 1980a,b Germany Finland Pennsylvania, U.S.A. Hong Kong Ohio, USA O.OZS <0.05 <0.002 0.05 0.04 Industrial hygiene studies have demonstrated that the typical 8-h tlme-weighted-average (TWA) exposures among auto mechanics at garages range from less than 0.002 to 0.05 fiber/cc (Table 4) (Rodelsperger et a}., 1986; Kauppinen and Korhonen, 1987; Moore, 1988; National Institute of Occupational Safety and Health, 1980a,b; Cheng and O'Kelly, 1986). Thus, the 8-h TWA exposure to asbestos among auto mechanics Is well below the safety standards mandated in most countries (for example; 0.1 fiber/cc in the United States), which should provide adequate protection to auto mechanics. In fact, according to the Committee on Hygienic Standards of the British Occupational Hygiene Society (BOHS), such fiber concentrations would have been categorized as "negligible" (Levadie, 1984). Given such extremely low exposures, it is not surprising that none of the epidemiologic studies reported an Increased mesothelioma risk among auto mechanics. In investigating occupational cancers, an adequate time interval between exposure and the development of disease must be allowed to account for the long latency of the disease. Since brake linings and clutch facings have been in commercial use for decades, latency is not a concern in case-control studies. For example, Woltowitz and Rodelsperger (1994) reported that the motor ve hicle mechanics in their study began employment as mechanics between 1925 and 1972, which would have provided a more than sufficient time Interval between exposure and disease development. Thus, the lack of an increased mesothelioma risk among auto mechan ics cannot be attributed to an inadequate latency in the studies reviewed. The application ofHill's causation criteria to epidemi ologic data ofmesothelioma among auto mechanics has clearly demonstrated that auto mechanics do not have an increased risk of malignant mesothelioma as a re sult of exposure to asbestos fibers from brake linings and clutch facings. It may be of interest to point out that there Is no in creased risk of lung cancer, an asbestos-related cancer, among auto mechanics. In its 1986 document on auto mechanics, EPA cited one study in reference to lung can cer. The study cited was a cohort study ofworkers at an amoslte factory in Paterson, New Jersey, that supplied the U S. Navywith asbestos insulation forpipes, boilers, and turbines during World War II (Seidman eta/.. 1979). The authors characterized the workers' exposure as "in tense work exposure to amoslte asbestos." This study, which was based on workers with intense exposure to amosite, was totally inappropriate for assessing risk in auto mechanics, who are exposed to extremely low lev els of chrysotile. Wong (1993) reviewed cohort studies of auto mechanics in relation to lung cancer. Individ ual cohort studies consistently indicated that there was no increased lung cancer risk: the standard mortality ratios (SMRs) were 0.92 among bus garage mechanics in London (Rushton et al., 1983; Rushton. 1987), 1.15 among garage mechanics in Stockholm (Gustavsson et al.. 1990), and 1.01 among auto mechanics in Denmark (Hansen, 1989). The summary SMR based on a meta-analysis of all three cohort studies was 1.01 (95% Cl 0.80-1.26) (Wong, 1993). Similarly, case-control studies reported no significantly elevated lung cancer risk ratios among auto mechanics from different loca tions: 1.06 in France (Benhamou etal., 1988); 0.9 inNew Mexico (Lerchen etal.. 1987); 1.2 in Louisiana, Florida, Pennsylvania, Virginia, mid New Jersey (Vineis et al., 1988); and 0.7 in nine metropolitan areas in the United States (Detroit, Chicago, Philadelphia, Pittsburgh. New York. Long Island, San Francisco, Birmingham, and Atlanta) (Morabia et al., 1992). Thus, epidemiologic studies Indicate that auto mechanics are not at an in creased risk of lung cancer as a result of working with ' brakes and clutches. However, it should be pointed out that these studies on lung cancer In auto mechanics were published after EPA had Issued its 1986 docu ment. CONCLUSION In 1986 EPA issued its guideline on the prevention of asbestos disease among auto mechanics. In the guide line, EPA's claim that auto mechanics were at an in creased risk of developing mesothelioma was based on a few case reports published between 1965 and 1982. No epidemiologic studies were Included in EPA's guideline. As reviewed above, three epidemiologic studies of ma lignant mesothelioma among auto mechanics had been published by 1986 (McDonald and McDonald, 1980; Teta etal., 1983; Spirtas et al., 1985). All three studies Involved auto mechanics in the United States, and none of the three studies reported an increased risk of ma lignant mesothelioma among auto mechanics. It Is not clear why EPA did not include these three studies in its guideline, particularly in view ofthe fact that one of the studies was conducted by another federal health agency, the National Cancer Institute (Spirtas etal., 1985), and a second one was funded by the National Cancer In stitute and conducted by the Connecticut State Health Department (Teta etal.. 1983). vJ HWBUI0010241 176 OTTO WONG Furthermore, subsequent to the publication of EPA's 1986 guideline, a second analysis of the NCI study had been published (Spirtas et al., 1994) and three additional new studies had appeared (Woitowitz and Rodelsperger, 1994; Teschke et al.. 1997; Agudo et al, 2000). [The Woitowitz and Rodelsperger (1994) paper was presented at the 1993 International! Workshop on Health Associated with Chrysotlle Asbestos held in Jersey, Channel Island, whose list of partici pants/attendants included an official from U.S. EPA.J Again, none of the additional new studies reported an increased risk ofmesothelioma. In spite of the remark ably consistent findings of no increased risk of malig nant mesothelioma among auto mechanics from these studies, EPA has not revised its 1986 guideline. [A re cent (May 2001) search of EPA's Web site reveals that the 1986 EPA guideline "Gold Book" Is still on EPA's current list of asbestos publications available to the public.] Regulatory agencies, such as EPA, are charged with the responsibility of protecting public health. It is, therefore, prudent for them to issue early warnings if a danger Is suspected, even in the absence of adequate scientific data. However, as discussed above, this was not the case for EPA's 1986 guideline. In this Instance, adequate scientific data were available but ignored. As such, EPA as a regulator has not fulfilled its respon. sibility of providing the most accurate and up-to-date information to the workers or the general public. Occupational regulatory policies and guidelines, when based on proper scientific evidence, are invalu able and can prevent avoidable diseases In workers or other exposed individuals in the general public. On the other hand, regulatory polices and guidelines that are not based on the most complete and up-to-date scientific evidence are not only incorrect but also misleading. It is the regulators' responsibility to develop, modify, and re. vise policies and guidelines in accordance with the most relevant and the latest scientific evidence. In particu lar. the reluctance of regulators to revise policies and guidelines when new scientific data become available is Inexplicable. REFERENCES Agudo, A., Gonzilez, C A.. Bleda. M. J., Ramirez, J,, Hernandez, S., Lopez, F., Calleja. A.. Panades, R,, Tumguet, D., Escolar, A., Beltran, M-, and Gonzalez-Moya, J. E. (2000). Occupation and risk ofmalignant pleural mesothelioma: A case-control study In Spain. Am. J. bid. Med. 37,159-168. Armltage, P,, and Berry, G. (1994) Statistical Methods in Medical Research, third ed. Blackwell, Oxford. Benhamou, S,, Benhamou. E., and Fiamant, R. (198. Occupational risk factors oflung cancer In a French case-control study. Be. J. bid. Med. 45, 231-233. Cheng. V. K. I., and O'Kelly, F. J. (1986). Asbestos exposure In the motor vehicle repair and servicing Industry in Hong Kong. J. Soc Occup. Med. 36, 104-106. Davis, J. M.. and Conlam. S. W. (1973). Experimental studies on the effects of heated chrysotlle asbestos and automobile brake lining dust Injected into the body cavities of mice. Exp. Mol. Pathol. 19, 339-353. Environmental Protection Agency (1986). Guidance for Prevent ing Asbestos Disease among Auto Mechanics. EPA. Washington, DC. Gustavsson. P.. Plato, N., Lidstram, E. B,, and Hogstedt, C. (1990). Lung cancer and exposure to diesel exhaust among bus garage workers. Sr.and. J. Work Environ. Health 16,348-354. Hansen. E. S. (1989). Mortality of auto mechanics, a ten-year follow up. Scand. J. Work Environ. Health 15, 43-46. Hatch, D. (1970). Possible alternatives to asbestos as a fraction ma terial. Ann. Occup Hyg. 13, 25-29. Hill, A. B. (1965). The environment and disease: Association or cau sation? Proc. R. Soc. Med. 58,295-300. international Agency for Research on Cancer (1989). IARC Mono graphs on the Evaluation of Carcinogenic Risks to Humans: Oc cupational Exposures in Petroleum Rehnbig; Crude Oil and Ma jor Petroleum Fuels, VoL 45. World Health Organization, Lyon, France Kaupplnen. T,, and Korhonen, (1987). Exposure to asbestos during brake maintenance of automobile vehicles by different methods. Am. Ind. Hyg. Assoc. J. 48,499-504. Lerchen, M. L., Wiggins, C. L. and Samec, J. L- (1987). Lung cancer and occupation In New Mexico. J. Natl. Cancer Inst. 79, 639-645. Levadle. B. (1984). DefinitionsforAsbestos and OtherHealth-Related Silicates, ASTM Special Technical Publication 834. American So ciety for Testing and Materials. Philadelphia. Lynch. J. R, (1968). Brake lining decomposition products. J. Air Pollut Cont Assoc. 18,824-826. McDonald, A. D., and McDonald. J. C. (1980). Malignant mesothe lioma in North American. Cancer 46, 1650-1656. McDonald, A. D. (1986). Letters to Otto Wong dated May 27 and June 4. 1986. Moore, L. L. (1988). Asbestos exposure associated with automotive brakerepair in Pennsylvania. Am. Ind. Hyg. Assoc. J. 49, A12-Al 3. Morabia, A., Markowitz, S., Garibaldi. K., and Wynder, E. L. (1992). Lung cancer and occupation: Results of a multicenter case-control study. Br. J. Ind. Med. 49,721-727. National Institute for Occupational Safety and Health (1980a). In dustrial Hygiene Report Asbestos at Allied Brake Shop, Cincin nati, Ohio.\ National Technical Information Service PBS 1-241879. NIOSH. Washington. DC. National Institute for Occupational Safety and Health (1980b). In dustrial Hygiene Report Asbestos at Reading Brake and Align ment Service, Reading Ohio, National Technical Information Ser vice PB81-24519. NIOSH. Washington. DC. Nurminen, M. (1975). The epidemiologic relationship between pleu ral mesothelioma and asbestos exposure. Scand. J. Work Environ. Health 1, 128-137. Pelnar. P. V. (1988). Further evidence of nonasbestos-related mesothelioma, a review of the literature. Scand. J. Work Environ. Health It, 141-144. Peterson, J. T., Greenberg. S. D.. and Buffler, P. A. (1984). Nonasbestos-related malignant mesothelioma. Cancer54,951-960. Rodelsperger, K,, Jahn, H.. Brocket B,. Manke. J., Paur, R,, and Woitowitz, H. J. (1986). Asbestos dust exposure during brake re pair. Am. J. Ind. Med. 10, 63-72. Rushton. L, Alderson. M. R.. and Nagarajah, C. R. (1983). Epidemi ologic survey of maintenance workers in London Transport Execu tive bus garages and Chiswick Works. Br. J. Ind. Med. 40,340-345. Rushton, L. (1987). better to Otto Wong, dated November 17. 1987. HWBUI0010242 MALIGNANT MESOTHELIOMA IN AUTO MECHANICS 177 ScIdman. H-. Selikoflf, I. J.. and Hammond, C. (1979). Short-term as bestos work exposure and long-term observation. Ann N. K Acad. Sci 330. 61-90. Spirtas. R.. Keehn, R, Wright, W,, Stark, A.. Beebe, G.. and Dickson, E. (1985). Mesothelioma risk related to occupational and other as bestos exposure: Preliminary results from a ease-control study. Am. J. Epidemol. 122, 518. Spirtas, R, Helneman. E. E, Berstein, L., Beebe. G. W,, Keehn, R. J.. Stark, A., Harlow, B. L., and Benichou, J. (1994). Malignant mesothelioma: Attributable risk ofasbestos exposure. Occup En viron. Med. 51. 804-811. Surgeon General's Committee (1964). Smoking andHealth. US. Pub lic Health Services. Washington, DC. Teschke. K,, Morgan, M. S:. Checkoway, H., Franklin, G., Splnelll. J. J..van Belle, G.. and Weiss: N. S. (1997). Mesothelioma surveillance to locate sources of exposure to asbestos. Can. J. Pub. Health 88, 163-168. Teta. M. J., Lewlnsohn. H.C., Meigs, J. W.. Vlodone, R A., Mowad, L. Z..and Flannery, J. T. (1383). Mesothelioma In Connecticut. 1955- 1977: Occupational and geographic association. J. Occup. Med. 25, 749-756. _ Vlnels. P-. Thomas. T.. Hayes. R B,, Blot. W. J., Mason, T. J.. Pickle, ' L. W.. Correa. P, Fontham, E. T. H., and Schoenberg, J. (1988). .. - Proportion oflungcancers In males, due to occupation. In different areas of the USA InL J. Cancer t2, 851-856. Williams. R L., and Muhlbaler. J. L. (1982). Asbestos brake emis sions. Environ. Res. 29, 70-82. Woitowltz, H. J.. and Rodelsperger, K. (1991). ChrysotUe asbestos and mesothelioma. Am. J. Ind. Med. 19, 551-553. Woitowltz, H. J., and Rodelsperger, K. (1994). Mesothelioma among car mechanics? Ann. Occup Hyg. 38, 635-638. Wong, O. (1992). ChrysotUe asbestos, mesothelioma and garage me chanics. Am. J. Ind. Med. 21,449-451. Wong, O. (1993). Consideration ofspecificity Inassessing the relation ship between asbestos and cancer. Am. J. Ind. Med. 23, 521--523. Wong, O., and Raabe, G. K. (1996). The application ofmeta-analysls In reviewingoccupational cohort studies. Occup Envlrnn. Med. 53, 793-800. ) HWBUI0010243