Document V3ve5bQGLog03e4k4Kp9V58Rj

From (613)' 991-5726 ' ^< : - Order # 02983125DP02353685 Tue 3ul 23 20:52:50 2002 Page 2 of 32 UUnistrerxwrel Pttlbulogy, 26:55-65. 2002 CejwrijtUt fc 2092 Taylor & Francis 0l9l"-313,TO 5I2.W I .00 f. Taylof&frands " hcaftlisdenccs Malignant Mesothelioma and Occupational Exposure to Asbestos: A Clinicopathological Correlation of1445 Cases Victor L. Roggli, MD Anup&mai Sharma, MB Kelly 4. Butnor, MD Thomas Sporn, MB BoMn T. Volljner, MB Department ofPathology, Durham. Veterans Administration and Duke Untverftiiy Medical Center, Durham, North Carolina. USA ) Asbestos exposure is indisputably associated with development of mesothelioma. However, relatively few studies have evaluated the type of occupational exposure in correlation with asbestos fiber content and type. This study reports findings in 1445 eases of mesothelioma with known exposure history: 268 of these also had fiber burden analysts. The 1445 cases of mesothelioma were subclassified into 23 predominant occupational or exposure categories. Asbestos body counts per gram of wet lung tissue were determ ined by light microscopy. Asbestos fibercontent and type were determined by scanning electron microscopy and energy dispersive x-ray analysis. Results were compared with a control group of 19 lung tissue samples. Ninety-four percent of the eases occurred among 19 exposure categories. Median asbestos body counts and levels of commercial and noncommercial amphibole fibers showed elevated levels for each of these 19 categories. Chrysolite fitters WHte detectable in 38 of 268 castes. All but 2 of these also bad abovo-background levels of commorclal amphibolss. When compered to commercial amphibolss, the median values for noncommercial amphtbote libers ware higher in 4 of the 19 exposure groups. Most mesotheliomas in tiro United States fait into a limited number of exposure categories. Although a predominant occupation was ascertained for each of these cases, there was a substantial overlap in exposure types. All but 1 of tha occupational categories analyzed had above-background levels Of comrisareiol amphiboles. Commercial amphibolss are responsible for most of the mesothelioma cases observed in the United Slates. Keywords masatheSoma. asbestos, fiber analysis, occupation, amosite, trmiwKtB, chrysolite The association Jjetween mesothelioma and prior asbestos expostlhRis undisputed. A wide variety of occupational and environmental exposures have been implicated as the source of this exposure. In the past, insulation workers and shipyard workers were exposed to high levels of asbestos dust and decades later were found to have a markedly elevated rate of mesothelioma. Bystanders working in the vicinity of those directly handling asbestos products were also found to be at risk, in addition, household contacts of asbesfos workers and those living in the vicinity of an asbestos manufacturing plant also developed mesotheliomas [1-3]. Such studies have led to the Address ccsnspnrrienca to Victor L HoggS. MU Department ol Pathology. DurhamVbtemns Adm!nis!raBon and Duke UniversityMecfcal Center; Durham, NC2771HUSA. conclusion that mesothelioma may be related to brief, low-level, or indirect exposures to asbestos. Although numerous reports document meso thelioma risk associated with particular exposures, there are relatively few studies that have examined the distribution of occupations in a large series of mesothelioma cases. Otto reported the occupational categories in 71 patients with asbestos-related mesothelioma and 37 "spontaneous" or idiopathic cases [4]. The Australian Mesothelioma Register provides biennial reports on the occupational cat egories of the mesothelioma cases collected by the registry in Australia [5]. It is difficult hovvever. to: find detailed analyses of lung [Tiineral fiber content in a series- of mesothelioma patients that-also'contain detailed information regarding occupational or environmental exposures to asbestos. We have previously shown that amosite asbestos is the most common fiber type identified in the lungs of 55 HWBUI0010165 Freni CG13) 991-5726 Order # 02983125DP02353685 Tue 3u1 23 20:52:50 2002 Page 3 of 12 V. L. Rogidi efc nt- patients with mesothelioma from the United States [6]. Furthermore, there is a significant correlation between lung fiber burden of asbestos bodies and commercial amphibole fibers with duration of exposure to asbestos [7], For equal durations of exposure, individuals who directly worked with asbestos products have a higher lung burden than those who were Indirectly exposed (i.e., bystander exposure), and shipyard workers have on average a higher burden than non-shipyard workers. The present study extends these observations by examining exposure information In more than 1400 cases of histologically confirmed mesotheliomas. In 268 of those cases, the exposure information could be correlated with the results of tissue mineral fiber analysis. MATERIALS AND METHODS The consultation files of one of the authors (VLB) were reviewed for all cases of mesothelioma for which information was available regarding possible exposure to asbestos. The diagnosis of mesothelioma was based on the gross distribution of tumor, histologic pattern, and the results of immunohlstocheniicsl .studies as previously described [8]. In soma cases, histochemical stains and ultrastructural studies were performed when indicated. Information was also obtained regarding the age, sex, primary site of the tumor (pleura vs. peritoneum), smoking history* presence of pleural plaques and/or calcification, and presence of histologic asbestosts [9,10], Patients were classified into one of the 23 exposure groups shown in Tables 1-3. Industries with known asbestos exposure are listed in Table 1, occupations in TABLE 1 Mesothelioma Cases by Industry Single Multiple Total Shipbuilding* U.S. Navy1* Construction' Insulation* OH and chemical Power plant Railroad Automotive' Steal/matat* Asbestos mlg,0 Papertnill Ceramics/glass 203 91 99 -,-sas -78' 50 37 24 33 34 7 6 . 86 84 35 11 10 10 IB 27 10 6 0 0 289 175 134 103 88 SO 53 51 43 39 7 8 "Includes fohet shipwright rigger, sandblaster, shipfittet electrician painter, and weldor. blnclu<los merchant marine seamen. 'Includes construction worker laborer, carpenter, painter, dry wall/piastemr. "Includes pipecoveter. insulator, asbestossawyet asbastos sprayer "Includes auto mechanic, brake repair worker brake line worker `indudes steel, alumWum, and iron Foundry workers; furnace worker; potroom worker. "Includes asbestos torsito, asbestos manufacture. asbestos plant worker. TABUS 2 Mesothelioma Cases by Occupation Smgis Multiple Total Pipefitter" Boilermaker** Maintotrsnca0 Machinist Electrician Shee&meiai Other asbestos*1 159 81 75 62 59 17 23 28 31 IS 27 22 5 0 187 112 SO 89 81 22 23 "tpductes welders. `'Includes boiler worker. tioHar maksr, bailor mechanic, boiler repair man, steamlitter. 'Includes mechanical engineer. ''feidurfss millwright, brick mason/masoa bagging machine operator.'bagger, asbestos worker (not otherwise specified). Table 2. and nonoccupational exposures in_Table 3. The occupations listed in Table 2 are distributed across many of the industries listed inTable 1. For example, pipefitters may be employed in shipyards, oil refineries, power plants, etc. Many individuals worked in more than one of the categories indicated in the tables. However, for purposes of analysis, an attempt was made to place patients in a predominant category where the most intense exposures (based on previous analyses from our laboratory) were likely to occur (e.g., insulators or shipyard workers). When more than one category of exposure occurred, these separate exposures were recorded as well. Whan available, data were also collected regarding the duration of exposure for each of the categories indicated in Tables 1-3. Fiber analyses were performed on formaiin-fixed or paraffin embedded lung tissue specimens by using previously described methods [11]. Lung tissue was processed for digestion by using the sodium hypo chlorite technique. The residue was collected on 0.4-gm pore-sfee Nuclepore filters. For light micro scopic analysis, the filter was mounted on a glass slide for asbestos body quantification. Filters were counted at a magnification of x400, and only bodies with typical morphology and thin, translucent cores were TABLE 3 Mesothelioma Cases with Nonoccupartionai Exposures Single Multiple Total Household contacts Building occupants Other* Environmental 86 17 46 5 3 1 0 0 89 18 46 5 "Includes wcKteqt, baker, bookkeeper, businessman computer con sultant, drives dry clsening/laundry business, dyo and press setter, ele vator repamnsa forestsetvicev furniture co.. grinder, hairdresser, jewelry repair, Kentsmokers, landscaper, industrial manager, minister, no known exposure nuree. paScemoa pressman, professor: resident construction, salesman, saltcakeartd carbon plant operator, school adminrstrator.sitica plant and pattern shop status postradiation/chemotherapy, telephone instaSet tonlte worioar, tire monufecura. / HWBUI0010166 From (613)' 991-5726 ' mmsz: Order # 02983125DP02353685 Tue 3ul 23 20:52:50 2002 Page 4 of 12 Mesothettotna and Occupation f counted as asbestos bodies [12]. Results were reported as asbestos bodies per gram of wet lung tissue (AB/g). The detection limit for a 0.3-g sample size is approximately 3 AB/g. For cases in which no asbestos bodies were detected in the sample, the value was recorded as less than the detection limit for that sample. For scanning electron microscopic (SEM) analysis, the filter was mounted on a carbon disk, with colloidal graphite, sputter-coated with gold or platinum, and examined in a JEOL JSM-6400 scanning electron microscope at a screen magnification of xIOOO and with a screen size of 22.7 x 17.3 cm. Fibers 5 pm or greater in length ware counted by using a protocol in which 100 consecutive fields or 200 fibers were counted, whichever came first; Fibers were defined as particles with an aspect ratio (length to width) of at least 3:1 and roughly parallel sides. The concentration of fibers for each case was calculated based on the fiber density (i.e., per mmz) on the filter surface times the effective area of the fitter, divided by the weight of the lung sample that was analyzed. The results were reported as total asbestos fibers (coated and uncoated) 5 pm or greater in length per gram of wet lung. The detection limit for a 0.3-g sample size is approximately 440 fibors/g. For cases in which no asbestos fibers were detected in the sample, the value was recorded as less than the detection limit for that sample. Fiber types were determined by using a combination of fiber morphology as determined by SEM and elemental composition assessed by energy dispersive x-ray analysis (EDXA). Fibers were classified as amosite, crocidolita, tremolits, anthophyilito, actinolite, or chrysotile as previously described [11 ]. For purposes of analysis, amosite and crocidolite were grouped together as commercial amphiboles (AC). Tremolite. anthophyllite, and actinolite were grouped together as noncommercial amphiboles (TAA). The proportion of each fiber type and the total asbestos fiber concen tration were used to determine the tissue concentration of AC, TAA, and chrysotile for each case. All other fibers were counted as nonasbestos mineral fibers (NAMF). For cases^where no fibers of a particular type were detected; IffET'galue was recorded as less than the detection limit for that case. The results of fiber analysis for the mesothelioma cases were compared with 19 controls from our laboratory with macroscopically normal lungs, no evidence of asbestos-related disease, and no history of asbestos exposure. For statistical analysis we used the log rank test to explore the relationship between age of diagnosis as a time of failure and industrial exposure and the Cox proportional hazard model for the final analysis of age. We also used Kaplarv-Meier plots to illustrate age effects. We used a general linear model to relate the logarithmic transformation of fiber counts to industrial exposure, and fiber counts below the detection limit were arbitrarily set at half the detection limit. We used analysis of variance for exploratory analysis of con tinuous variables and chi-square tests for analysis of categorical variables. We used logistic regression analysis to relate binary variables to other, independent 57 variables. Finally, all P values were for 2-sided test, and all analyses were done with S-PLUS software (MathSoft, Seattle, WA). RESULTS A total of 1445 cases of histologically confirmed mesothelioma were retrieved from the files of one of the authors (VL0). These included 1322 men and 123 women. The median age for the entire group was 67 years (range: 17-94 years). The site of origin was the pleura in 1309 cases and the peritoneum in 136 cases. The histologic types included 698 epithelial, 442 biphasic, and 292 sarcomatoid cases. For 13 cases, information was not available regarding the specific histologic pattern of the tumor. Information regarding smoking was available in 1112 cases: 830 were smokers or ex-smokers and 292 were nonsmokers. Information was available regarding the presence or absence or pleural plaques (PPP) in 778cases. Plaques were present in 608ca%<asj[78%). Information regarding the presence or absence of asbestosis was available in 577 cases. Asbestosis was present in 135 cases (23%). Exposure Categories Cases were classified into the 23 exposure cat egories indicated in Tables 1 --3. The industry with the largest number of cases was shipbuilding, followed by U.S. Navy, construction industry, and insulation industry, each of which included ICO or more cases (Table 1). Among the 12 industries listed in Table 1, an exposure in some additional occupational setting was noted in 26% of the cases. The occupation with the largest number of cases was pipefitter, followed by boilermaker (Table 2). Among the 6 specific occu pations listed, an exposure in some additional occu pational setting was noted in 21% of the cases. The nonoccupational group with the largest number of cases was household contact of an asbestos worker (Table 3). Among the 4 nonoccupational groups, an exposure in an occupational setting was noted in 2% of the cases. Two of the categories were fairly nonspecific, with "other asbestos" (Table 2) including several occu pations for which there were few cases in any one subgroup, and "other" (Table 3) including a variety of occupations with no known exposure to asbestos, in addition, the "building occupant" group is known to be associated with tissue asbestos burdens that are oftentimes indistinguishable from background levels. Among the 4 cases with "environmental" or neigh borhood exposures, only 1 was from the United States. When these 4 groups are excluded, the remaining 19 categories (12 industries, 6 occupations, and 1 nonoccupational exposure) accounted for 94% of the 1445 cases of mesothelioma included in this study. Two of these exposure categories often have been considered as low-level exposures. These include household contacts of asbestos workers and environ mental (neighborhood) exposures. There were 89 cases that were household contacts of an asbestos worker. HWBUI0010167 From C613>'991-5726 Order # 02983125DP023S3685 Tue 3u1 23 20:52:50 2002 Page 5 of 12 J 58 V. r,, Rvg&i efc at. and 10 of those had an additional occupational expo month to as long as 56 years. Shorter duration wax >: sure to asbestos Of these, 79% were women with an observed for shipbuilding, U.S. Navy, asbestos average age of 59 years. There were 4 cases with manufacture, and building occupants. The former 2 arts environmental exposure to asbestos. Ail were men with related to large numbers of workers with shipyard an average age of 67 years. None had any additional employment or navy service limited to World War II, known exposure to asbestos. Three of these cases were bringing down the median value for these groups. i- from a region of Turkey where there was intense environmental exposure to tremolite-ectinolite. The Shorter duration for asbestos manufacture may have been related to the very dusty environment with many fiber burdens in these 2 groups are discussed below. workers quitting after a relatively short period of 1 employment [12], Table 4 provides the median and range of age for Gender, Exposure Duration, and Age each category of exposure. A log rank test demon Over 90% of our cases, of mesothelioma occured in strated thatthere was a significant relationship between men. In some occupational categories, all were men the age at development of mesothelioma and type of (Table 4). These included insulators, railroad workers, industry or occupation (chi-square=112, P < .0001). power plant workers, steel workers, brake repair Table 5 provides detailed results of a Cox model workers, boiler workers, electricians, and machinists. A analysis for the relationship between age and the key large percentage of the women with mesothelioma industries or occupations that had unusually high or occurred in the household contact group, which, as low hazard rates. In the table, the relative hazard rates noted above, consisted of 79% women. The building (column HR) provide the key result Whereas values of occupant group included about an equal number of HR above 1 (die reference for remaining undesignated men and women. cases) indicate a higher hazard and younger age for The median duration of exposure to asbestos is also developing mesothelioma, values of HR below 1 provided in Table 4 for those cases where this infor indicate a lower hazard and older age for developing mation was available. For most categories, the median mesothelioma. For example, the highest relative hazard duration was 2 or 3 decades, with ranges as short as 1 of 2.19 occured in those exposed in a nonindustrial TABLE 4 Age, (Sender, and Duration by Exposure Category <i }! Age* M:Fb it industry i Shipbuilding ; { U.S. Navy i ; Consfeuction : j Insulation . ! Oil and chemical j` Power plant :i1 Automotive Railroad j Stoel/mstal : Asbestos mfg. Papetmill --"""t. Ceramtes/gtass ;i Occupation Pipefitter Boilermaker Maintenance Machinist Electrician Sheermsts) j` Other asbestos Nonoccupatianal exposure Household contacts . Building occupants j Other Environmental 69 (38-91) 67 (38-84) 65 (39-85) 59.5 (37-78) 70 (45-89) 68 (39-81) 64 (31-83) 74 (45-94) 71 (44-81) 64.6 (48 86) 68 (61-77) 64.6 (40-79) 71 (38-91) 67 (39 -86) 70 (39-88) 68 (47-87) 69 (48-83) 70 (82-76) 60 (45-84) 69 (25-93) 45 (17-78) 635 (28-78) 87 (38-68) 37:1 55:1 96:1 All men 40:1 All men AH men All men Ail men 88:1 25:1 6:1 61:1 Alt men 89:1 All men All men 19:1 All men 13:67 8:9 27:11 All men *Woes represent mediafi with range h parentheses. ''Ratio of man to women In given group "Values represent msdhn duration of exposure in years wlttt range In parentheses. A&ffit NA, ncrtappEcable. Exposure duration* ) 10.5 (1 mo-45 yr) 9 (3 mo-S3 yr) 30 (1-48) 26 (1 mo-50 yr) 30 (6 mo-45 yr) 30(2-60) 23 (3-60) 30(2-47) 31 (5 mo-45 yr) 8 (7 mo-40 yr) 35.6 (13-42) 30 (17-62) 29 (6 mo-50 yr) 21 (6 mo-47 yr) 26 (2-56) 17.5 (1 61) 28.5 (1-48) 27.5 (1-43) 27(540) 20 (1-45) 14 (1-26) NA 28 (18-38) l HWBUI0010168 Froni (613)' 991-5726 Order # 0298312S0F023S368S Tue 3ul 23 20:52:50 2002 Page 6 of 12 S(aoUtsUt,II"i and Occupation TABLE 6 Cox Model Analysis of Age of Development of Mesothelioma Vtnable HR SE P value Nonindustrial Insulation Pipefitters Sail workers Duration Female 2.19 1.33 0.66 0-52 0.99 0.70 0.134 0.110 0.129 0.162 0.002 0.143 4.6 xIO-3 .010 .0010 5.9 x 10~E 23x10"1u .014 tfista Bere, HR is \ha relative hazard, and SE provides the standard er ror for the rogrossioti coefficient which can be calculated as tog (HR). setting (i.e., household contacts, building occupants, or other nonindustrial exposure). This result implied that this group experienced mesothelioma at an earlier age. Workers in the insulation industry also experienced a higher hazard rate, but pipefitters and railroad workers had lower hazard rates. All these results were controlled for duration of exposure and female gender, because both were covariates in the analysis, interestingly, longer duration was associated with increased rather than decreased age at diagnosis, and after controlling for category of exposure female gender was associated with a lower hazard and later age at diagnosis. The P values listed in the table provide the significance for the effect of each variable or group on the hazard for developing mesothelioma. Figure 1 illustrates these results further. FIG. l Kaplan-Meier plots of the probability of being tumor free versus age in years. Here age at diagnosis was treated as a failure time. Curve 1 provides the plot for pipefitters and railroad workers; curve 3 provides the plot for those in the insulation industry and those with nonindustrial exposure; and curve 2 provides the plot for the remaining groups of industrial/occupational exposures. 59 In the figure are 3 Kaplan-Meier plots of the probability of being tumor-free versus age. Curve 1 combines pipefittes and railroad workers, who had the lowest hazards end developed mesothelioma at the oldest ages. Curve 3 combines those with nonindustrial exposure and those in the insulation industry, and this group developed mesothelioma at the youngest age, Curve 2 is for the remaining cases. Tumor Location, Pleural Plaques, and Asbestosis Approximately 90% of our cases arose In the pleura. The other 10% arose in the peritoneum. The highest percentage of peritoneal mesotheliomas occurred among insulators and absestos manufacturing plant workers (Table 6). In each of these groups, the ratio of pleural to peritoneal tumors was about 2:1. In the remaining groups, this ratio ranged from 8.6:1 in construction workers to more than 50:1 in shipyard workers. The low ratio- oW .8:1 in building occupants suggests that mesotheliomas in this setting are at or near the background rate of occurrence for these tumors. TABLE 6 Site. Plaques, and Asbestosis by Exposure Category Site5 pppb Asbestosis11 . industry Shipbuilding U.S. Navy Construction Insulation Oil and chemical Power plant Automotive Railroad Stael/metal Asbestos mfg. Paparmil! Cdramics/glass Occupation Pipefitter Boilermaker Maintenance Machinist Electrician Shoetmetsl Other asbestos Nonoccupattona? exposure Household contacts Building occupants Other Environmental 52:1 64:1 8.6:1 2.1:1 82:1 17:1 8:1 38.0 9.3:1 2.2:1 6:1 8:0 50.1 30:1 26:1 22:1 74:1 20:0 5:1 4.3:1 1.8:1 8.6:1 4:10 144/177 (81%) 12/57 (21%) 14/41 (34%) 64/75 (85%) 36/46 (78%) 23/27 (86%) 8/12 (67%) 20/24 (83%) 15/18 (93%) 21/24(87%) 6/6 (83%) 1/2 (50%) 103/119 (87%) 60/74 (31%) 45/56 (80%) 32/44 (78%) 33/40 (83%) 9/11 (82%) 2/6 (33%) 20/35 (67%) 3/7 (43%) 11/24 (46%) 0/1 (0%) 37/132 (26%) 4/37 (11%) 6/35 (17%) 34/59 (58%) 5/29 (17%) 3/16 (19%) 0/12 (0%) 2/17 (12%) 3/11 (27%) 1/17 (65%) 1/5 (20%) 0/2 (0%) 20/83 (24%) 11/46 (24%) 7/35 (20%) 4/28 (14%) 8/30 (27%) 1/7 (14%) 0/6 (0%) 3/38 (7.9%) 0/10 (0%) 2/21 (9.5%) o/i (0%) Ratio of pleural to peritonea! tumors. 'Number of cases with plaques divided by number of informative coses (percentage). "Number of cases with asbestosis divided by number of informative cases (percentage). HWBUI0010169 Froth (613)- 991-5726 Order # O2983125DP02353685 Tue 3ul 23 20:52:50 2002 Page 7 of 12 60 V. L. Boggti ct at. When information was available regarding the median asbestos body count in each of the 23 expos;. ^; presence or absence of parietal pleural plaques (PPP), categories exceeded our normal range of 0-20 AB/g,- it was found that a high percentage of cases did in fact except for 2 categories: automotive brake repair have plaques (Table 6). Chest radiographs and even workers and building occupants. The industry with the computed tomographic (CT) scans are somewhat highest median asbestos body count was the insulation insensitive for the detection of PPP[13], so cases industry, and the occupation with the highest count where PPP were not mentioned in the radiographic was pipefitter. reports and were not observed in surgical specimens Analysis of variance demonstrated that the category were considered to be uninformative. Although logistic - of industry or occupation was significantly related to regression analysis demonstrated that there was a the number of asbestos bodies (P < .0001), to the significant association between category of exposure number of commercial amphibole fibers (P < .0001). to and the presence of PPP (P = .0015), these categories the number of noncommercial amphibolo fibers explained Just 5% of the noise in the data, and it was (P < .0005). and to the number of chrysotile fibers difficult to identify particular categories associated with (P < .0001). Furthermore, much of the industry/ PPP. In general, workers in insulation, pipelining, and occupational differences were due to. higher fiber levels steel had increased prevalence of PPP. and those in in the insulation industry coupled with lower fiber nonindustrial-exposure situations had lower preva levels in 3 industrial groups: those with nonindustrial i i lence of PPP.. On the subset of 199 cases with asbestos body counts, logistic regression analysis demonstrated exposure, those In the auto industry, and those who were machinists. This is illustrated in-Figure 2, which that the prevalence of PPP was closely related to log provides box plots of the natural logarithm of fiber (asbestos body counts) (P < .0001). In this subset of counts in these 4 categories versus an "other'' group of f|. cases, industrial category explained an additional 11% all remaining cases. Simple visual comparison shows \i\I of the noise In the data with a P value of .044. Having that the pattern for fiber counts is roughly the same for accounted for the effect of fiber burden, the auto all 4 types of fibers, and general linear model analyses industry seemed to have an increased prevalence of demonstrated that after controlling for gender and PPP, but the variance associated with the remaining duration of exposure, the associations between fiber exposure categories was too great to reach any con counts and these 4 industrial groups were significant (P clusions about the remaining categories of exposure. values ranging from approximately O to .02) with one Neither exposure duration nor gender was significantly exception--the level of noncommercial amphibo' related to the prevalence of PPP. fibers in the nonindustrial group (P > .1). ) When information was available regarding the The predominant fiber type identified in each of the presence or absence of asbsstosis, it was determined categories was commercial amphibole (mostly amosito that a relatively low percentage of cases had with some crocidoiite), except for 4 categories: histologically confirmed asbestosis (Table 6). Among automotive brake repair workers, household contacts, occupationally exposed individuals, the percentage of building occupants, and environmental (neighbor cases with asbestosis ranged from 0% of glass and hood) exposures (Table 6). In these instances, ceramic workers and friction product workers to as high noncommercial amphiboles (mostly tromolite with as 67% of asbestos manufacturing plant workers. some actinolite and anthophyllite) predominated over : !; Asbestosis was rarely observed among patients wiib commercial amphibole fibers. Chrysotile fibers were !; } mesothelioma and nonoccupational exposures. None detected less frequently, and for 7 categories were not of the environmental exposures or building occupants detected at all. Nonasbestos mineral fibers were present had asbestosis, but asbestosis was confirmed in all groups In roughly similar amounts. histologically--8.3% of household contacts of Further analyses were performed on the- 12 asbestos 'workers. Logistic regression analysts dem exposure categories for which there were at least 10 onstrated that although the category of exposure analyses within the category (Figure 3). Commercial was significantly associated with akrestosis (P < amphibole fibers were present in concentrations that .0001), there was no pattern of one category or another exceeded the background level in all 12 categories, that could be identified as having either a high or low ranging from 57% (other) to 100% (insulators) of the prevalence of asbestosis. Furthermore, when the individuals in any one category. Noncommercial burden of asbestos bodies was accounted for by amphibole fibers were also present in excess con including log (asbestos body count) in the model, the centrations in all 12 categories, ranging from 13% industrial category was no longer associated with the (maintenance workers) to 68% (shipyard workers) of prevalence of asbestosis. In other words, the influence individuals in any particular category. However, the 1 of industrial category on prevalence of asbestosis percentage of individuals in a category with elevated i appeared to be mediated entirely by differences in fiber levels of noncommercial amphiboles exceeded 50% in ! burden. only 2 categories: shipyard workers and electricians. Chrysotile was present in excess concentrations in 9 ! ; of the 12 categories (not in 'pipefitters, construct!' \ Ijimg fiber Burden Analyses The results of tissue fiber analysis for each of the workers, or other), ranging .from 5% to. 23% of v J-' individuals in any:- one :categary. . In none of the' exposure groups are summarized in Table 7. The categories was chrysotile; found in elevated concen- HWBUI0010170 From (613) 991-5726 Order # 02983125DP02353685 Tue 3u1 23 20:52:50 2002 Page 8 of 12 JfeMjthcliQma end Occupation 61 /, i::" TABLE 7 Lung Fiber Burden Analyses in 268 Mesothelioma Cases AB/gm AC1* TAAh Chrys** NAMF** Industry Insulation Asbestos mfg. Shipbuilding Power plant Construction Steel/metal Oil and chemical U.S. Navy Railroad Automotive PapermfH Ceramcs/glass* Occupation Maintenance Pipefitter Brateflnrfcer Machinist Sheetroetal Bectrician Other asbestos Nonoccupational exposure Other Household contacts Building occupants frnvkonmental Reference cases*1 20.1 (0.12-1600) 4.1 (0.9-322) 1.08 <0.006-436) 0.34 (0,013-21.9) 0.19 (0.002-83.6) 0.43 (6:13-2.3) 0.27 (0.018-2.8) 0.17 (0.002-44) 1.43 (0.005-5.7) 0.02 (0.008-1.6) NA . 3.06 0.4 (0013-155) 2.2 (0006-174) 1.68 (0.24-8.6) 0.1 (0.002-1.8) 0.78 (0.4-0.83) 1.4 (0008-6.3) 1253 (0.1-3.9) 0.043 (0001-11.5) 0.13 (0.002 14.1) 0.002 (0.002-1.1) 0.7 (0.005-02) 0.003 {0.002-0.022} 240 (0.81-11900) 33 (14-755) 23-9 (0.24-2160) 18,8 (0.92-200) 14.7 (0.48-268) 8.4 (2.3-129) 4.9 (1-393) 4 (0.54-202) 38.6 (2.95-48) 1.4 (0.12-17.3) NA 28.2 243 (032-316) 24.2 (1.05-49.4) 21.3 (4.14-149) 19 (0.98-48) 6.2 (0.8-34.1) 5.5 (2.3-94.5) 33.6 (2.0-33.7) 3.S (1.02-64.9) 3.4(0.45-118) 0.87 (0.38-43.8) ND <0.6 (<0.1-<2.64) 5.19 (1.11--19.7) 6.7 (1.7-90) 4.56 (0.026-79.8) 3.84 (0.73-2) 4.6 (0.48-16.7) 2.6 (0.43-6.6) 2.4 (0.67-0.9) 3.3 (0.98-17.1) 9.68 (8.85--27.3) 2 (0.24-9.8) NA 3.4 2.06 (0.87-124) 66.5 (43-90) 4.8 (0.91-11) 7.71 (1.13-14.2) 0.5S (0.48-0.6) ND ND 3,1 (2-4.3) NO 2.1 (0.72-2.18) NA 1.1 1.8 (0.63-17.1) 5.2 (0.98-19.2) 8.49 (0.78-24.2) 15 (2.9-27) 5.8 (4.1-7.4) 5 (0.51-45) 3.54 (0.66-17.3) 1.2 (1.13-1.26) ND 1.45 (1.44 1.45) ND "NO-------5 (1.6-10.4) ND 2.3 (1.05-43) 5.2 (0.98-22.4) 2.3 (1.1-4.1) 158 (1.7 455) <0.0 (<0.17 <2 64) 0.32 1.8 0.64 ND <0.6 (<0.1-<2.84) 11.6 (2.24-19.7) 27 (3.2-54.1) 10.3 (1.18-138) 6.2 (2.4 19.5) 6.3 (0.6-85.2) 6.4 (0.49-23.9) 6.4 (1.3-22.7) 8.3 (0.58-27.6) 17.7 (5.13-33.8) 1.4 (0.34-18.5) NA 6.3 6.4 (1.3-18.5) 10 (0.85-31.7) 7.3TI 84-19.1) 5.1 (0.49-34) 11 (7.4-18) 6.7 (3.1-54) 4.0 (2.4-45) 6.5 (0.48-28.1) 3.3 (0.24--14.6) 4.9 (1.5-36.8) 5.1 (3.1-14.5) <0.6 (<0.17--<2.54) "Asbestos bodes per gran of wet lung tissue (xTO3) as determined by light rnicraseopy.Values shown are medians with range in parentheses. `'Fibers 5 pm or greater in length per gram of wet lung tissue (x IBP) as determined by scanning electron microscopy and energy sfepurave x-ray analysis. Values shown are medians far eases whom fiber type was detected, with range in parentheses. When ranges are not shown only one casa had detectable levels of indicated fiber type. "Tissue was available for liber analysis in only one case of cetamtes/giass plant worker. ^Nimsefi cases with normal lungs, no Itistory of asbestos exposure, and no avkfance of asbestos related tissue injury at autopsy. Note. AB. asbestos bodies; AC, commercial amphibotos (amosite+crocidolite); Chrys. chrysotilt? grn gram of wet lung: HA, no tissue available for analysis; NAMF, nonasbestos mineral fibers: NO. none detected;TAA, noncommercial amphibotes fcremolito i actinalito -- anthophyUrte)- trations in more than half of the individuals in that category. TWo of the .categories deserve further comment Household co'ntaEi of asbestos workers had tissue asbestos burdens that were similar to the median value for some occupational groups. For example, the median asbestos body count of household contacts (130 AB/g) was of the same order of magnitude as construction . workers (190 AB/g). Hence, household contacts have . tissue asbestos burdens that are on the average equivalent to mild to moderate occupational exposure {14,15], Wives tended to have higher lung asbestos burdens than daughters or sons. In contrast the median . asbestos bodycountfor automotive brake repairworkers 0 was within our normal range. In many of these cases, the tissue asbestos content was completely within the limits of background. In those cases with an elevated -i tissue asbestos content, excess commercial amphibole 11 fibers (mostly amosite} were invariably detected. Lung -/ burden analyses in autofnotive brake repair workers reflect either a normal range tissue asbestos content or elevated commercial amphibotes (11. 17]. To further examine how the prevalence of plaques and asbestosis related to fiber burden, we used a logistic regression model to relate prevalence to the Fiber burden of commercial amphibotes. The analysis demonstrated that both plaques and asbestosis were closely related to the natural logarithm of commercial amphibole counts (P < .0001). and the result allowed us to fit these relationships with 2 equations, one for the probability of plaques (Pplaq) and the second for the probability of asbestosis (Pasb). If we symbolize the logarithm of commercial amphibotes as Lea, then Pptaq = 1 + exp{3.36 - 0.393 * Lea) * pasb - 1 + exp(30 2 _ 12.4* log(Lca)) (2) Figure 4 shows bow these 2 compare over the range of Lea found in our cases. Clearly, this plot demonstrates that piques appear at lower liber burdens HWBUI0010171 Front C613>-991-5726 Order # 02983125DP023S3685 Tue 3u1 23 20:52:50 2002 Page 9 of 12 82 Asbestos Bodies Commercial Arnphibofes V. L. BoggH at al. .4 Non-Commercial Amphibolss Chrysolite FIG. 2 Boxpiots of fiber burden for asbestos bodies (upper left), commercial arnphibofes (upper right), \ noncommercial arnphibofes (lower (eft), and chrysolite fibers. The vertical axis provides toe fiber burdetn. J fog (number per gram), and each plot provides die details for 5 industrial or occupational groups: v-'~ nonindustrial exposure (Non), insulation industry (Insul), automobile industry (Auto), machinists (Mach), and all remaining groups (Other). The results inside the solid dark box provide the range from the 25th to 75th percentiles, and the white bars indicate the median value. The more distant brackets and lines above and below the solid boxes provide the full range and most peripheral values. i If than ssbestosls. Asbestosis begins to become prevalent examining a large series of cases for which information i' when the commercial amphibole fiber burden exceeds 22000 per gram of tissue (Lca=10). By contrast at was available regarding the patient's occupation or other exposure to asbestos. We have found that 94% of this level of arnphiboles already 64% of cases have our cases had exposures in one or more of 12 different plaques. Thusjplagues are a sensitive marker for lower industries, 6 occupations, or 1 nonoccupationa) levels of fiEef burden, and asbestosis does not occur exposure. Exposure in more than 1 category was a until much larger fiber burdens. As a result of the above common finding, observed in at least 26% of the relationships among occupationally exposed indivi industrial categories and 21% of the occupations. The duals, the percentage of mesothelioma cases with most Important nonoccupationa! exposure was as a asbestosis remained low at around 20% until the household contact of an asbestos worker. The duration highest exposure category (insulators) was reached. At of exposure was typically on the order of decades, this poiht the percentage increased to more than 60% although we made no attempt in this regard to dis of cases with histologicallyconfirmed asbestosis. These tinguish between intermittent and continuous Findings indicate that there is a much higher threshold exposures. of asbestos exposure for asbestosis than for PPP, and ft could be argued that these cases are not rep this threshold is uncommonly met (about 20% of resentative of the types of exposures occurring in ii cases) for most exposure categories. The above models mesothelioma cases in the United States, since more and equations suggest that with increasing fiber than 90% of them are medicolegal cases. Therefore, it is burden, all will eventually get asbestosis. i of Interest to compare our exposure groups with those obtained by the Australian Mesothelioma Surveillar'"' Program [5J. The latter is a survey of all mesotheli ; DISCUSSION cases in Australia, and therefore does not suffer fi ' The current study extends our previous observations any medicolegal selection -bias. The top 14 groupings regarding mesothelioma and asbestos exposure by are shown arranged from greatest to least number of HWBUI0010172 From (613)' 991-5726 tsreRtetewft :: . Order # 02983125DP02353685 Tue lul 23 20:52:50 2002 Page 10 of 12 . MG. S Histogram showing the percentage of cases in each category with elevatedlung fiber burdens for commercial amphibofes (AC), noncommercialamphiboles (TAA), andchrysotile. Categoriesrepresentedare those for which at least 10 cases with lung analyses in the given category were available. HHC, household contact cases in Table 8. Carpenter/joiner would be included in our construction industry, as would builder/laborer. Waterside would be included In our shipbuilding industry, and plumber is included in our pipefitter occupation. Finally, asbestos dwelling is equivalent to j our building occupajjt category, and brake lining to our ;i: automotive industry It can- be seen that with the 7. exception of mesothelioma cases related to working in p;: the Wittenoom mine in Australia, the major categories )f:pf exposure are remarkably similar to those noted in our '.IvvSeries of cases. ,:Bs: More than 90% of our cases occurred in men, :;T Reflecting the predominance of mates in those occugOpations and industries most commonly associated with g|asbestos exposure. Women predominated among ^household contacts of asbestos workers, related to the g:j|durdsring of work clothes by the spouses of workers ipglHXJupationally exposed to asbestos [15, 16], The age for most groups was the seventh or eighth 3~3ebade, consistent with the long latency period for the ^Ullllivelopment of mesothelioma and initial exposure ' fil'gdtlring adulthood [8]. Exceptions to this Included ^'ygptOusehoId contacts and insulators. The former is due to '''^.exposure of some household contacts during 7, Childhood. The latter suggests the possibility of an inverse relationship between dose and latency for mesothelioma. In addition, pipefitters and railroad workers were significantly older (Figure 1). The latter may be due to the fact that most railroad exposures occurred prior to 1958 (i.e,, during the steam era). The highest percentage of peritoneal mesotheliomas occurred among workers involved with the manufacture of asbestos products and insulation workers (Table 6). The percentages, however, did not reach the level reported by Seiikoff et at., in which the ratio of peritoneal to pleural tumors among insulators was 3:1 [1,2]. Insulators and asbestos plant workers also had the highest lung asbestos fiber burdens, which is consistent with the observation that on the average, peritoneal mesotheliomas are associated with higher doses than are pleural mesotheliomas [7j. The obser vations are consistent with a model in which the pleural and peritoneal mesothelial ceils are equally susceptible to the development of mesothelioma, and the prob ability is, in turn, related to the dose. At sufficiently high exposure levels, the dose to the peritoneum is high enough that the probability of a peritoneal origin approaches that of a pleural origin [18]. Pleural plaques were found irr a high percentage of cases (78% of informative cases for the whole series). HWBUI0010173 Order # O2983125DP02353685 Tue 3u1 23 20:52:50 2002 Page 11 of 12 V. L. Tlogfgli et al. (Figure 4). In contrast, asbestosis was obsorved It commonly (24% of informative cases for the whole series), again with evidence for a dose response (Figure 4). These findings are consistent with our prior observation that pleural plaques and mesothelioma require less exposure to asbestos than is typically associated with the development of asbestosis [11], The results of fiber burden analysis are consistent with our prior observation that amosite is the most common fiber type associated with mesothelioma among U.S. workers (6). Furthermore, the present study shows that amosite predominance extends over most occupational exposure groups (Figure 3), con sistent with the widespread use of amosite-containing insulation products in these industries in the past. However, among nonoccupational exposures, noncom mercial amphiboles predominated (Table 7). Chrysotile was infrequently identified in any exposure group, consistent with its tendency to breakdown into smaller fibrilsthat are more readily removed fromthe lungs [19]. PIG. 4 Plots of the probability of either observing pleural plaques or asbestosis on ike vertical scale versus log (commercial amphiboles) found per gram of tissue on the horizontal scale. The lines provide a graphical picture of the results of two logistic regression equations (see equations 1 and 2), one for the probability of observing plaques (left curve, as noted) and me for the Environmental or neighborhood exposuresappear to be a very uncommon cause of mesothelioma, based on our observations. The 4 cases we studied in this category included 3 from the southern Anatolian region of Turkey that were exposed to high levels of tremolite and actinolite in the local environment[20]. The only U.S. case we studied had a fiber burden indistinguishable from background. Similarly, exposure., to asbestos as a building occupant is also an infrequr' \ probability of observing asbestosis cause of mesothelioma. Among the 6 cases of bulldi>._ ,J (right curve, as noted). occupants for which a fiber analysis was performed, 3 had a modestly elevated content of noncommercial amphiboles and 3 had levels Indistinguishable from lending further support to the proposition that a high percentage of our cases are asbestos related. Furthermore, there was evidence for a dose response, with a higher percentage of plaques observed with background. The finding of low levels of noncom mercial amphiboles is consistent with an exposure to predominantly chrysotile asbestos in this setting [21]. Only 1 occupational exposure group contained on average more noncommercial amphtbole fibers than increasing levels of commercials amphtbole fibers commercial amphtbole fibers. This was the category of automotive brake repair workers, in this group, fiber burden analyses performed on 11 cases showed either an elevated amosite content or an asbestos concen TABUS 8 Australian Mesothelioma Registry Report--1999 tration indistinguishable from background. Workers in this category frequently were exposed to asbestos in other categories as well (Table 1), and none had -. i1 Single Multiple Total asbestosis (Table 6). Among 15 workers exposed to brake dust who also had plaques, 10 had additional Carpantar/jomer 187 33 220 exposures in jobs or industries where commercial WhtenoomBuild er/laborer Navy Asbestos mfg. Shipbuilding 189 160 100 99 79 25 27 43 33 53 214 177 143 amphiboles predominated, and 3 additional cases had elevated commercial amphiboles in their lung tissue 132 samples. Lung tissue was not available for analysis in 132 the other 2. These findings are consistent with our prior Railways 89 36 125 observations of brake repair workers with ii Bollemnaker Power plant 77 40 76 39 117 mesothelioma [11, 17. 22], and with the nature of 115 brake dust, which contains low levels of short Waterside 70 S i ; Plumber 62 13 i Asbestos dwelling 66 6 Brake lining 58 12 Flnor/turnar 51 16 78 chrysotile fibers [23], These observations in combi 75 nation. with a negative case control study [24] indica*'- 72 70 67 that brake dust is unlikely to cause mesothelioma. : ) Another group of interest is individuals exposed ctoetdplftey fibers from smoking Kent cigarettes, the Scurte. From [5J mierq.rt|%filters of which contained crocidolite for a HWBUI0010174 From C613)'991-5726 Order 02983125DP02353685 Tue 3u1 23 20:52:50 2002 Page 12 of 12 % . and Occupation 4-year period during the early 1950s [16], We have performed fiber analyses on lung tissue samples from 4 such cases. Far 3 of these, smoking Kepis (or r pretending to smoke Kent cigarette butts as a child) 3 ' vvas the only known exposure to asbestos (these cases t pi included in the "Other" category in Table 3). One . fcase that also had an exposure as a household contact r fitted tow levels of amosite in her lung samples. None of t j: the 4 cases had detectable crocidolite. Based on these 3 '? analyses, smoking Kent cigarettes seems to be an t , dfilikely cause of mesothelioma. rf Nanasbestos mineral fibers were found in roughly similar amounts among each of the exposure categories 5 (Table 7). These include talc, silica, rutile, aluminium t. silicates, metallic fibers, apatite, and manmade mineral fibers [11. 25, 26], There is no evidence that these fibers contribute to the development of mesothelioma }, [27]. A possible exception is refractory ceramic fibers, ir which have produced a high rate of mesothelioma in i- inhalation exposure studies and have been associated o with the development of pleural plaques in humans n [28, 29]. These fibers have a greater biologic per s sistence than most other manmade mineral fibers [30]. n We have identified refractory ceramic fibers in lung if tissue samples from 6 patients with mesothelioma, and ] in 2 cases, they were the most frequent fiber type identified. In both cases, however, commercial amp- hibole fibers were also present in excess amounts. Insufficient information is presently available to implicate manmade mineral fibers (including refrac 3 tory ceramic fibers) as a cause of mesothelioma in a! humans [31]. r l- REFERENCES :o 1. SeHtoff U, Lee OH It AsbestosandDisease. Now York: Academic; ]- 1978. n 2. Seftoif IJ. Hammond EC. edb. Health hazards of asbestos in exposure. Am NY Acad Sd. 1979:330:1-S11. af er 3. Landrigan PJ, Kazemi H, etfe. Tfco third wavs of asbestos disease: exposure to asbestos in place public health control.Ann NYAcad Set. 1391 ,-643:1 026. sr 4. Otto H. Das berufebendingto Masothelivm in der BRD. Pathologe t- 1980;2:8-18. in 5. Asrstrails Meso#eofta Register. The mctd&tce of Mesothelioma in in Austmlmi'Austra'tmnMosoSidiofim Register Report, 1995, 1998, 1399. Camperdown, NSW: National Occupational Health and id Safety Commission. to 6. Roggli VL Pratt PC. Brody AR. Asbestos fiber type in malignant a! mesothelioma: an analytical scanning electron microscopic study al Of 94 cases. AmJtnd Med. 1993:23:605 614. 7. Roggli VI, Malignant mesothelioma and duration of asbestos id exposure: correlation with tissue mineral content. Am Occup tlyg. je 1996;39:303-374. in 8. Roggli VL, SanfUippo F. Shelburne JD. Mesothelioma. In: Roggli or VL Greenberg SD. Pratt PC. eda Pathology of th Asbestos-Associated Diseases. Boston: Utils, Brown: 1992109-164. of 9. Greenberg SD. Benign asbestos-related pleural diseases. In: )rt Roggli VL, Greenberg SO, Pratt PC, eds. Pathology of - Asbestos-Associated Diseases. Boston: Little. Brown; tr 1992165-187. 65 1 >. Roggli VL, Pratt PC. Asbestosis. In: Roggli VL. Greenborg SD, Pratt PC, axis. Pathology ofAsbestos-Associated Diseases. Boston: Little. Brown: 1992:77-108. 11. Roggli VL, Pratt PC, Brody AR. Analysis of tissue mineral fiber content In: Roggli VL. Greenberg SO, Pratt PC, eds. Pathologyof Asbestos-Associated Diseases. Boston: Little, Brown; 1992:299-345. 12. Roggli VL Asbestos bodies and nonasbestos farraginous bodies. In: Roggli VL Greenberg SD. Pratt PC. eds. Pathology of Asbestos-Associated Diseases. Boston: Little, Brown; 199239-75. 13. Roggli VL Greenberg SD, Scitzman LH, et al. Pulmonary fibrosis, carcinoma, aetd farraginous body counts in amosite asbestos worfcors: a study of sbe cases.AmJCUnPethef. 1980;73:498-503. 14. Wain SL, Roggli VL Foster WL Parietal pleural plaques, asbestos bodies, and neoplasia: a clinical, pathologic, and radiographic correlation of 25 consecutive cases. Chest. 1984;86:707-713. 15. Gibbs AR. Griffiths DM, Pooley FD, et al Comparison of fiber types and size distributions in lung tissues of pepaoccupational and occupational cases of malignant mesothelioma. BtJfnd Mod. 1990:47:821-626. 10. Roggli VL Oury TD, Moffatt EJ. Malignant mesothelioma in women. In: Fechner RE, Rosen PP, eds. Anatomic Pathology. Chicago: ASCP; 1997:147-463. 17. Srobro SH, Roggli Til, Ssmsa GP. Malignant mosothe&oms associated with low pulmonary tissue asbestos burdens: a light and scanning electron microscopic analysis of 18 cases. Mud Pathol, 1995;8:614-621. 18. Hodgson JT, Damton A. The quantitative risk of mesothelioma and lung cancer in relation to asbestos exposure, Ann Occup Hyg, 2006.44:565-601. 19. Roggli VL Brody Aft. Changes in numbers and dimensions of chrysolite asbestos fibers in lungs of rats following short-term exposure. Exp Lung Res. 1984;7:133-147. 20. Zeren EH, Gumuidutu D, Roggli VL Tuncer I, Zortudemir S, Erkisi. M. Environmental malignant mesothelioma in Southern Anatolia: a study of 50 cases. Environ. Health P&sp. 2000:108:1047-7 050. 21. Gaensler EA. Asbestos exposure in buildings. Clin Chest Med. 1392,13:231-242. 22. Roggli VL Fiber analysis. In: Ram WN, od. Environmental end Occupational Medicine. Philadelphia. Uppincott-Rsven;1998: 336-347. 23. Williams RL Muhlbaier JL Asbestos brako omissions. Environ Res. 1982:2670-82. 24. Woitowitz H-J. Rodelsperger K. Mesothelioma among car mechanics. Ann Occup Hyg- 133438:635-638. 25. Chung A. Nonasbestos pulmonary mineral fibers in the general population. Environ Rss 1983;31:189-200. 26. Roggli VL Nonasbestos mineral fibers in human lungs. In: Russell PE, cdl Microbeam Analysis--fS9$. San Francisco: Son Francisco Press; 1989:57-69. 27. McDonald JC, Armstrong 8, Case B, et al. Mineral fibre content of lung in mosothelial tumours in North America. Cancer. 1989;63:1544-1547. 28. Miller WC, Hostsitrerg TW, Hamilton RD. Oncogenicity study of respirable refractory ceramic fibers (RCF) in hamsters. MTC Report 092518. Littleton. CO: Mountain Technology Center: 1991:1-69. 29. Lackey J, Lemssters G. Rico C, et al. Refractory ceramic fiber exposure and pleural plaques. Am J Resprr Crit Care Med. 1936:154:1405-1410. 30. McConnell EE, Mast RW, Hesterberg TW, et al. Chronic inhalation toxicity of a ICvafin-based refractory ceramic: fiber in Syrian golden.hamsters. Inhalation Toxicol. 1995:7:503 -532 31. Rodelspenger K, Jacket J-H. Pohlabeln H, et al. Asbestos and manmade vitreous fibers as risk factors lor diffuse malignant mesothelioma: Results from a German hospital-based caso-conttol study. Am J Ind Med. 2EX)1;38:1--14. HWBUI0010175