Document bBGrbpe0x6qo5QgvG248qeLo1

09/10 '00 15:44. FAX 41 22 7351463 D.M.BERNSTEIN @002 inhalatlonTooacotogy, ^Supplement*);*! 1-418,2000 Copyright .2000 Taylor 4 Francis 08 95-8378/00 SI 2:00 -F .00 ASBESTOS FIBER TYPE AND LENGTH IN LUNGS OF CWRYSOnLE TEXTILE AND PRODUCTION WORKERS: Fibers Longer Than 18 p,m Bruce W. Case, Andre Dufresne, A. D. McDonald, J,C. McDonald, Patrick Sbbastien Departments of Pathology, Epidemiology, Biostatistics, and Occupational Health, McGill University, Montreal, Quebec, Canada Excess hing cancer risk for a cohort of chrysotlte textile plant workers was many times the risk observed in a cohort of chrysotile miners/milters. The latter had greater exposure to chrysovle/tremaiite. A previous lung burden study confirmed this excess exposure in miners/milters and showed little difference in fiber length. Selection of too short a fiber length cut-off (5 urn or more) in the previous study could have masked differences in lung-retained fiber length. In this follow-up, we counted only those Intrapulmonary fibers exceeding 18 pm in length. Long fiber concentration and dimension were assessed by transmission electron microscopy (TEM) and energy-dispersive x-ray spectrometry (EOS) for autopsy samples from 64 textile workers and 43 ohryaotlle miners and millers. These long fibers were significantly more concentrated in the lungs of chrysotile miners and millers, consistent with their greater exposure. However, when only these longest fibers were compared, there was a somewhat greater mean and median intrapulmonary fiber length for chrysotile textile workers (mean fiber length, all fiber types combined, 25.2 10.2 pm vs. 22.3 <5.<5 pm in miners/miilers, p < .001; medians 2 1.8 vs. 20,p < .051. Despite their lesser apparent lung cancer risk, chrysotile, tremolite, total amphlbole, and total long fiber asbestos concentrations were all highest In the lungs of miners/miilers. Twenty-two of 64 textile workers had lung content of crocidolite and/or amosHe (32.5% of508). These amosite/crocidollte fibers werepresent in the lungs of workers who ceased employment prior to the first use ofsuch fibers recorded in this industry. The results suggest that H) asbestos fiber length differences cannot explain the difference in lung cancer risk excess and slope between cohorts and (2) the experience of textile workers should not be used to assess risk of lung cancer in miners, cement workers, and friction product workers, regardless of fiber type. Animal inhalation studies (Miller et a!., 1999) and previous theoretical cal culations (Lippmann, 1990) have suggested a role for fiber length in the genesis of lung cancer. Indeed, recent practice for the assessment of carcinogenic ity of both natural and synthetic fibers has concentrated analytical efforts on fibers longer than 20 pm and of diameter less than 1 pm. Human studies of lung-retained fibers have been less convincing (Becklake & Case, 1994), but can be criticized for their universal concentration on shorter fibers, usually defined either as all fibers longer than 5 pm (Sebastien et ai., 1989) or most often all fibers of any length resolvable by the electron microscopic magnrfica- This study was supported by the Medical Research Council of Canada, grant 1373S. Address correspondence to Bruce W. Case, Department of Pathology, McGill university, 3775 Uni versity Street. Montreal, Quebec, Canada H3A 2B4. E-mail: bcase@po-boxmcgill.ca 411 HWBUI0010279 09/10 Q0 15:44 FAX 41 22 7351483 D.M.BERNSTEIN 003 412 a W CASE ETAL. rion in use (e.g., Green et aL, 1997). This approach fails to assess adequately the longest fihers, since they comprise a relatively small proportion of fibers present in the lung, and length distributions In any analysis are heavily skewed toward the lower end. Of particular interest in this regard is the difference in lung cancer at tributable risk and exposure-diseaseslope for asbestos-exposed cohorts where fiber length might also differ. One example is the large difference in lung can cer risk between two well-studied chrysotlle asbestos industries. Miners and millers of chrysotile in Quebec have had very heavy exposures hut a shal low slope of lung cancer risk, with lung cancer excess mortality becoming apparent only at exceptionally high exposures, estimated recently as over 300 million particles/foot3/yr (over 1000 fiber-years, implying exposures such as an average of 50 fibers/cm3 for 20 yr) (Liddell et al., 1997, 1998; McDonald et al., 1980, 1993). One of the industries to which these milled fihers were shipped was a chrysotile textile plant in South Carolina studied hy McDonald et al. (i 983a, 7983b) and hy Dement et al. (1983, 1994).The textile workers had much lower levels of exposure to asbestos according to both groups of investigators. These workers nonetheless had exceptionally high lung cancer risk: at least one order of magnitude higher than that for the Quebec miners and millers. This textile plant has been cunsidered "an almost pure chrysotile operation" (WHO, 1998), based on observations that "Chrysotile asbestos re ceived from Quebec, British Columbia, and Zimbabwe was the only type of asbestos processed as raw fiber" (Stayner et al., 1997). Several authors have also suggested that the only use of commercial amphibole (amosite and crocidolite) in this textile manufacturing cohort was extremely small quantities of crocidolite yarn generating very little exposure in a solitary location in the plant and only after 1950 (McDonald etaL, 1983a). Also, "a very small quantity cf amosite (was used) for experimental purposes in the late 1950's" (WHO, 1998). In two other chrysotile textile operations in which similar excesses were recorded, there was also a record of exposure to crocidolite (McDonald et al., 1983b; Peto et al., 198.5). The first explanation for the divergence between results in these two groups f workers (low lung cancer rates in the more highly exposed min ers and millers) was that exposure must have been misdassified: too high an estimate for miners/millers; too low an estimate for textile workers, or both. This was refuted by a previous lung-retained fiber study (Sebastien et al., 1989). Lung samples taken at autopsy from 89 chrysotile miners/millers and 74 chrysotile textile manufacturing workers at the plant in question were com pared for lung fiber content for all fiber types, longer than 5 ym. Chrysotile and tremolite concentrations were highest in mlrters/milfers, and fiber dimen sions were closely simitar in the two groups. Exposure data were also available for each subject, and analyses matched for duration of employment and time from last employment to death demonstrated that ratio of chrysotile fiber con centrations in lungs from minOrs/millers to those of textile workers were "even HWBUI0010280 09/10_ 00 15:44 FAX 41 22 7351463 D.M.BERNSTEIN LONG FIBERS m LUNGS CF CHRYSOTILE WOmmS @004 () 1 ` - * Kni'.'i V. , higher than the corresponding ratios of estimated exposure intensl trivia! concentrations of amosite and crocidolite were present in 32rf<| workers' lung samples versus only 9% for miners/millers-. The latter reconfirmed by Green et al. (1997K who found amosite and/or crocidolite' in ' the lungs of 28% of a smaller group (n = 35) of textile workers from the same plant. Neither group of investigators assigned much significance to this find ing, however, as concentrations were low relative to those for chrysctile and tremolite (geometric mean 0.14 fibers of amosite or crocidolite/'ug dry lung as compared with 0.63 fibers/pg Cfcr chrysotile and 0.38 fibers/pg for tremolite; Sehastien et al., 1989). The authors speculated that the possible presence of cocarcinogens in the textile plant--specifically mineral oil, sprayed onto asbestos stock To facilitate spinning and reduce breakage-couid have played a role. By 1998, none of the suggested explanations (misclassification of exposure, greater fiber length in textile manufacture, mineral oil use) had enough support tu explain the huge risk difference, and it was suggested that this important paradox should he investigated further by looking at that fraction of lung-retained fibers longer than 20 pm (McDonald, 1998). This follow-up study addresses that issue. METHODS Samples in the current study were derived from those obtained in the previous work (Sebastien eta!., 1989). Subject selection, sample collection, and preparation are described in that study. Chrysotile mining subjects were drawn from the Thetford Mines portion of the Quebec cohort as constituted in 1976 (McDonald et at, 1980; Sehastien et al., 1989). Chrysotile textile workers were drawn from the cohort described by McDonald et al. (1983a). Usable electron-microscopic grids were obtained for 64 of the original 74 textile workers (SOmembers of the original cohort of McDonald et al. (1983a) as well as 14 women fulfilling the same cohort definition and having full expo sure histories available] and for a random selection of 43 of the 89 Thetford Mines male chrysotile miners/millers reported by Sebastien eta!. (1989). Lung samples consisted either of paraffin blocks (which were deparaffinized) or formalin-fixed tissue. Samples were chemically digested in fresh, filtered com mercial bleach, and the volume of bleach was adjusted to achieve a digest concentration of 1 mg dry wetght/mi. Then 15 ml was filtered through a stan dard Milllpore 25-Trim, 0.4'5-j.im pore size membrane filter, which was plasmaashed overnight. Suspended residue was filtered through 25-mm Nucleopore membrane filters (0.2 pm pore size), with final mounting on transmission elec tron microscopy (TEM) number 200 copper mesh grids using a carbon replica technique. In the current study, these grids were examined in a JEOL 100CX TEM at 10,000*. For each case, four grids were examined. For each grid, grid open ings having an area of 6400 ixnr4 were randomly selected for counting of HWBUI0010281 09/10 00 15:45 FAX 41 22 7351483 D.M.BERNSTEIN @005 414 8. W, CASE JET AL fibrous particles. From the four grids, up to 90 such openings (total), or 30 fibers (whichever occurred first), were counted. The length and diameter of all fibers were measured to the nearest 0.045 pm (0.5 screen mm) directly on the fluorescent screen, using 2 concentric circles (10 mm and 50 mm screen diameter). To be counted, fibers had to have a minimum length of 18 pm, min imum diameter of ,045 pm, and aspect ratio at least 3:1. Because such long fibers may overlap TEM grid bars, only those overlapping fibers in the upper and right quadrants were counted. Fiber type was identified from both mor phological features and energy^dispersive spectrometry (EDS), with selected area electron diffraction where necessary. Over 15 types of fiber were identi fied, but the current analysis considers only 4 individual asbestos fiber types; other fibers, which were rare, are grouped together as "nonasbestos." For ge ometric mean fiber concentration calculations, zero values were transformed to one-half the detection limit (0.018 ftbers/pg dry lung) prior to log trans formation. RESULTS Results of the study are summarized in Tables 1 -3. Overall, 1186 fibers were counted in this study: 679 in the lungs of 43 chrysotiie miners/mlllers (chrysotile 356, tremolite 286, amosite 22, crocidolite 8, all others 7) and 507 fibers in lung samples from 64 chrysotile textile manufacturing workers (chrysotile 256, tremolite 47, amosite 122, crocidolite 64, all others 39). In textile workers, geometric mean lung chrysotile fiber concentrations were highest and most closely related to cumulative exposure measured in million particles per cubic foot-years (MPCFY),with a correlation coefficient of .50 (Pearson r,p < .001). For miners and millers, exposure was most closely related to tremolite fiber concentration (Pearson r =39; p < .01). Fiber length distributions, which were rionriormal and skewed toward the lowest (18 pm) value, showed increasing percentages of amosite and crocfdolite with increasing fiber length in textile workers' lungs but not in those of miners and millers. Overall, in textile workers, of 95 fibers over 30 pm. TABLE 1. GJmparjsrjn mitheTwa ban Tissue QtofJtips Study auh;ccts M'mbrs/cnillers 43 oi 89'repotted by Sebastten .es al.. 1989: atl frcrtn TIielforrt Mines! Median birth yfear Median year oi l.vrv Median year termination Mncton, death year Mean years wockecKSbi Time since end esposure GeOrntriric mean exposure 1913. 1931 1971 197830 dr 1-6 ip .0011 Median 0 yr (jp = ,0H) 186 MPCFY (p < .001) Textile manufacture (M of 74'reported S#bASlien ei at., JfWl) (9T5 T94;l 195(1 T9 72 1-3 ~ 14 Median 20, vr 3.63 MPCPT HWBUI0010282 09/10 00 15:45 FAX 41 22 7351463 D.M.BERNSTEIN 006 LONG FIBERS IN LUNGS OF CHRYSOTILE WORKERS 415 TABLE 2. Qeontetf'nr -M.ean Fiber Coiicemratltin/Vg, Dr.y Lung Ani'l-Gearbeirtr wrap Fiber Length-(Fibers >18 um: Aspect Ratio >3:1) i'Alr`itx$.Jiu3liJfs in -- 41) Textile manuTaciure in- =* 64) Concern fiftion. RtJurfeijgth Concertuatiori Fiber frttglb Ghrysolife Trenuilitf Amoslle/crofTdoliio Total ainpHfofsieft Total asbestosTotal nonasbostos Total .all fibers 0.33 V' 0.323 0.024 0.728;< 0.036 22.Ax>)*mr, .2-1,7-ubS AmosJre 24.0:prn Cfbiidb)to?.2i6 m Sl.g-nm 22,2 Jm* 22.S'pm .054 .027. 437* .053: A'b .027 2X3 nm 2M pur ^ mesife Gcbcktollte. 24.-3.>wi 24:4 par? 23.8 Wry 23:0 pm 23:$ yAl,4 `'Slsnificatn alp < .00,1- (two-sample tusi).. ^Signiitcant at p < .01 :(iwo-sawplr test). TABLE 3, Cenmffjrtc Wean Fiber OianiOier And Meehan Aspecl Ratio (Fibers > 18 im) Miner.'.imlUeps. TeytOe unanxjfociuftf Cbrysoiile Trrno[ito Total asbestos ToUt-njcwasbestas '.DLaraeler ,08" Diamerer ,4-1 btdrftebelr.16 27<kl4 jS.-l 205:1 2 Hit "SJghfRtantai p < ,001 [two-sampletest). ^Significant at p < ,0-1 (two-sample tetit-j'. piametfir .T1 42.P5-1 OiiraiceF.:40- .47-:l Etiatwter .Ijfe .205:1 OtetneiOr A3- 42 were amosite or crocidolite and 31 chrysotile; .of 37 fibers over 40 pm in length, 20 were commercial amphibole and 12 chrysotile; and of the 19 fibers longer than 50 pm, 11 were amosite or crocidolite and 5 chrysotile. In the same length categories in chrysotile miners and millers, 100% of these longest fibers were chrysotile (32 fibers from 30 to 40 pm; 8 from 41 to 50; 2 over 50) or tremolite (12. from 30 to 40 pm, 7 from 41 to 50 pm, and 1 over 50 pm). DISCUSSION Lung-retained fiber study Is limited in inference, as the results represent only the fraction of internal dose that is retained until death. Further, compar ison of groups cf individuals using this technique is valid only insofar as those studied are representative of the larger groups (epidemiological cohorts, in this case) from which they are derived. We cannot be certain to what degree our groups of chrysotile miners/miIfers and textile workers are representative of the cohorts from which they are derived; there is some evidence that min ers and millers with lung tissue available may have a higher disease incidence HWBUI0010283 09/10 '00 IS:45 FAX 41 22 7351463 D.M.BERNSTEIN @007 41b B. W: CASE ET AL than the overall mining/milling cohort (Sebastien et at, 1989; Case, 1994).The situation is less certain for the textile workers, although lung cancer deaths in our study subjects (7 of 64 cases) were similar to proportional mortality for the whole cohort (Dement et al., 1983, 1994). In addition, the two groups differ in interval between cessation of work and death, and substantially in cumulative exposure, mean years worked, and year of termination. Each Of these differences could work in the direction of underestimating exposure in the textile workers' group. One hypothesis for the excess lung-cancer slope in textile workers, the idea that mineral oils could have played some role, cannot he addressed by the present study and remains to he tested either epidemiologically or experimentally. Lung-retained fiber analysis is most useful in exposure assessment and in cross-disciplinary validation of that assessment as it has been used in epidemi ological protocols. In the current study, the results may help us answer three questions. First, are the findings supportive of those in previous studies, both epidemiological and of lung tissue, in showing greater exposure for the cohort with less lung cancer (miners and millers)?Second, are the findings consistent with a role lor greater fiber length as an explanatory factor for the higher risk slope in textile workers (McDonald, 1998)? Finally, to what degree do the findings support or refute the claim that the textile cohort was exposed to "pure chrysotile" and is therefore an appropriate reference for chrysotile risk assessment (Stayner et al., 1997; WHO, 1998)? Our results closely parallel those reported by Sebastien et al. in 1989. Any other result woufd he surprising, since subjects were drawn from the latter study, although the population of fibers examined is not only longer hut of a completely different size range with almost no overlap. Indeed, direct correlation of filher concentrations in the two studies was very high: For the textile workers the two studies have a Pearson r of .94, .97. and .67 for amosite/crdeidolite, chrysotile, and tremolite, respectively (allp < .OOP). We cannot directly compare our results on an individual basis to those of Green et al. (1997), who examined an even shorter iiher length range than that assessed by Sebastien et al. It is nonetheless clear from the two previous studies as well as this one that exposure to chrysotile and tremolite is Indeed greater in the miners and millers, although the difference is perhaps less than that which would he suggested by the much larger difference in both cumulative exposure and total years of employment (Table 1). Our results do not support a role for fiber length alone in the genesis of greater lung cancer risk in textile workers. Direct comparison of fiber con centrations in this analysis of 1186 fibers longer than 18 pm for all asbestos, chrysotile, and tremolite fibers show all of these to he significantly higher in the miners and millers. Furthermore, chrysotile fibers were significantly thin ner and aspect ratios were significantly greater in the miners and millers' lungs, again consistent with the previous findings by Sebastien et al. When all fibers are considered together (including ndnasbestos iiher types), geometric mean, median, and mean iiher length are greater (p< .001) in the textile workers' HWBUI0010284 09/12. 00 15:46 FAX 41 22 7351463 D.M.BERNSTEIN @003 LONG FIBERS IN LUNGS OF CHRYSOT1LE WORKERS 417 lungs, even though concentrations of these longest fibers are lower in that group. Total amphibolies also had significantly greater geometric mean fiber length In textile workers, due largely to the excess of the shorter tremolite fibers In the mining population. This nonetheless confirms the suggestion by many investigators that environmental exposures in the textile plant were, at leastoverall and for total amphiboles, to somewhat longer fibers. Unless there is a very precise and specific "critical length" for lung-cancer genesis (rather than the total length cf all long fibers combined), fiber length alone must he questioned as the explanation for the high lung-cancer risk excess in this group of textile workers. Finally, as in two previous studies, we found that the "chrysotlle only" tex tile workers had a high proportion of individuals with lung tissue containing amosite and/or crocidoiite. Sebastien et al. had noted that these fibers were present in "nontrivial" concentrations in 32% of workers examined. Green et al. found commercial amphibole in 28% of 35 textile workers' lungs exam ined. Our findings are similar for the proportion of workers exposed: 19 of 50 males and 3 of 14 females, and 32% of all fibers counted in the lungs of textile workers were amosite or crocidoiite. Finally, 12 of the 22 textile workers having commercial amphibole in their lungs stopped working between 1938 and 1947, long before any such documented exposure in the textile plant. For 15 of the 22, amosite and crocidoiite formed the majority of all fibers present. This subset of the Charleston textile workers does not support the hypothesis that this is a "pure chrysotile" cohort (WHO, 1998). More generally, the ex posure experience of textile workers is clearly unique and should not be used to assess risk of lung cancer in miners, cement workers, or friction product workers, regardless of fiber type. REFERENCES EiptkUkfe. M.. nnd' Cast, W. 1994. /Tiher .'hucrlen and asljtssteiS. related king disease--delen'njnanis pf dose-response retntlpnrfjips. Am. (. fiesgifr, Oil. Care* Med'. J'S&)?t488--'S492. Case, B< W: '! 994. Biological 'Ann, .Oc.cpp. 7'tygi SI8. Demerit. J. M,, '[..Urm./it, L., 1 r, Svrjio.ns,;M. J-,, and Shy. C. M MBJ.-Espeisfires Arretmarfcrlify anwog'drysblfle .asb&stos workers. Part / hl /Vfc'qt O.emeM, ,1. M.. Brawn, pnd'OUuis. AL j,99d'..:J%ltwi?-up. study 4j'.dirysetrle asbestos teBR* .workers: cbhort mortality tfaAe<oenrol analyses. Am-1-trad, bteef. 36:431 -447'. Cirepn, P.. Plorfey, R-, Vnltvnrtym, V.. Allbuussy R., Kick, Cl.. Behire.nL j.. Millw, tt.. and Popley, F. 1997. Exposure and niioeraJogieal correlates ui pulmonary /jbrosis in ehrvsotife asbestos workers. Occup. .Bnvima. Med; 54:549-33.9: Liddell, F. O., McDonald, A.. 0., anpi MeDpijald, ], C. Hit? 1891-1920 birth cohort &i Cjuphec Chrysolite .miners ancl millers.- 'DevetepmenP from 19&4- add modality. to 1992. Ann. Occup,: Hyp. 4!:t3~36. Liddell, F. ;t>.f McDonald A. 13., and MrDtiiMd. f.,C. 19<J8. anti t.ung cancer in Quebec cbcysowle miners and millers.,Aon', Occup. tiyg. 4Z`.T-2Q. Uppinnnn, M. 1990. Effects nf liber dwacieristics on. iu,ng deposition, reiomlon. arc) disease, tnuiron. HcoW) Pmpecl. B8:3I1-.117. McDonald, A. O.. Frv, ) S.. Woolley. A. J., and McDwatd, |. C. 1,983ft. Oust exposure Jtxl niorlalify in up American chrysolSJ&ttsxtito.'pIstrH. Sr. f. irtcl. Med. 40:36,1--.87. HWBUI0010285 0.9/.10 ' 00 15 : 46 FAX 41 22 7351463 D.H.BERNSTEIN @009 418 B. W. CASE ET At. McDonald, A. O-, Fry, |. S,, Woolley, A. I., artef'McDonald. :.f. C. i983b. Dtrsi exposure. and mortality ?n an American factory ufdnjf chrysolite, acncrtlte.-and csocidnlitr in mainly textile- maput'acturfr. H* 1. tnrl. MpcI. 40:368-.3"4. Mcl lonaId. J. C I'JViH; Unfinished'business; The asbestos textiles, mysterv. Ann. Or cop. f-iyg. 43:3* 3. McDonald. 1. C,, Udclftlf, F. D, K., Gibbs, C. W., Gy&sfcrn, G. E1 anti McCtonald, A. D. 1*580, Dust exposure and.'morialHy to chrysolite mining, 19-W'-7%. Bf. J. Intf. 'Med. 3$: 1T '24. McDonald, .1- C-, Liddell K D. K.. Dufresnei.A,,,and'.{>4cDoeiald, A. fc>, IWJi Tl'fe 1-89 VI92Q blrthxdhoFi of Quebec chrysolite miners ari#millers: Mottjaliiy r976f-W?8- & jt 'todn filed: SKI 073- Miller. B, G,, tones, A. ,/S.etoI, ?V,, frueftadsit*, Dsj'Cpflervfii. T,, Sou.titr.'C. s-V,. DayTs,)J. M., Oorwldsoiji X. 1-999. Iriltoencetof cfiuraclorfertcs ai-in fullest flberf oo ffevakaprtienJ QFtornotJid In .the rat lurtg.-Aon. Occu/X ttys. 43:167-1-79. Petti. |,, Doll.. R:. Hertnorr, CL. Binn-s, W.r Clayton,, R.. arid-Gdffe;T. T983, Relationship ol mortality'to ntetir sures of onvirowmeni-al asbestos polluikip'.lo an: asbestos textile tortbry. Ana. Otcvp. i'-iyg. Z9;iiV>3nS. Sebastian, I5., McDonald. J, C., McDopalct; A. d~ Case, J. W_ and Hailey, R. 198$. Respirators'.cancer ir chrysotlle te\tile and winlh4hUostfte; Exposure InfluSnctes ft6m Iting analysts. Or. inti. Med. 46:180-187. Stayner. l_, Smith. !?.. Haller. J., Gilbert, S,, StSRnWmd; KL, Demers, L. Brown, f).. .and Le-iwer, R. Exposure-response Analysis of risk or respiratory disease associated wjth accupltftenal exposure to chrysolite asbestos. Occup. Environ. We*/, 54:64b~65Z World Health Organization. 1998. ChiVsorite Asbestos: Effects ran Homans, tnternattnnat.Rrugfijmme on Chemical Safety. Environmental Health Criteria 20}, {)} 1 CU-13Ef. Geneva: WHO. HWBUI0010286